ALISSON LEONARDO MATSUO

Caracterização de proteinases, identificação de proteínas

associadas à parede celular diferencialmente expressas

e estudo preliminar de vesículas extracelulares de

Paracoccidioides brasiliensis.

Tese apresentada à Universidade Federal de São Paulo – Escola Paulista

de Medicina, para obtenção do Título de Doutor em Ciências

São Paulo 2007

ALISSON LEONARDO MATSUO

Caracterização de proteinases, identificação de proteínas

associadas à parede celular diferencialmente expressas

e estudo preliminar de vesículas extracelulares de

Paracoccidioides brasiliensis.

Orientadora: Profa. Dra. Rosana Puccia Co-orientadora: Profa. Dra. Adriana K. Carmona

Tese apresentada à Universidade Federal de São Paulo – Escola Paulista

de Medicina, para obtenção do Título de Doutor em Ciências

São Paulo 2007

Matsuo, Alisson Leonardo

Caracterização de proteinases, identificação de proteínas associadas à parede

celular diferencialmente expressas e estudo preliminar de vesículas extracelulares

de Paracoccidioides brasiliensis / Alisson Leonardo Matsuo – São Paulo, 2007

xii, 165 fls

Tese (Doutorado) Universidade Federal de São Paulo. Escola Paulista de

Medicina. Programa de Pós-Graduação em Microbiologia e Imunologia.

Título em inglês: Caracterization of proteinases, identification of differentially

expressed associated cell wall proteins and preliminary studies of extracellular vesicles

from Paracoccidioides brasiliensis

1. Paracoccidioides brasiliensis; 2. proteinases; 3. serino-tiol proteinase; 4. Lon

proteinase; 5. parede celular; 6. vesículas/exossomos

Trabalho realizado na Disciplina de Biologia Celular do

Departamento de Microbiologia, Imunologia e parasitologia da

Universidade Federal de São Paulo – Escola Paulista de Medicina

com auxílio financeiro concedido pela

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP).

iv

"Na maioria das vezes, um pepino é somente um pepino"

Freud

v

Dedico este trabalho às pessoas mais importantes da minha vida...

...à minha querida esposa Leda e ao nosso filhinho Rafael, pela presença constante,

momentos alegres, paciência e apoio nas horas difíceis. Quando estou “incubado” no

laboratório sempre lembro de vocês, principalmente dos sorrisos e das travessuras do

Rafa. Amo vocês.

...aos meus pais Francisco e Lucia pela dedicação, apoio, educação, ensinamentos,

compreensão e carinho. Se hoje cheguei até aqui é porque vocês me deram este

privilégio. Admiro-os muito! Obrigado por tudo.

...à toda minha família em especial ao meu sobrinho Gustavo, aos meus irmãos Cleber e

Daniel, meus cunhados Claudia, Elaine, Henrique e Ieda e aos meus sogros Kiiti e

Alice, por comporem o tripé de nossa família e ajudar em momentos de necessidade.

Tenho certeza que são pessoas com quem sempre poderei contar.

vi

Agradecimentos

Meu primeiro agradecimento é dedicado à minha orientadora Profa. Dra. Rosana

Puccia, também conhecida como Chefa, ou simplesmente Rô. Pela confiança, amizade,

paciência, honestidade e principalmente pelo meu desenvolvimento científico e correção

dos meus erros português. Acho que hoje já consigo escrever “um pouco” melhor.

À Profa. Dra. Adriana K. Carmona, minha co-orientadora, pela sua contribuição

na minha formação científica, confiança, alto astral e simpatia.

Ao Prof. Dr. Igor Correia de Almeida, pela oportunidade de intercâmbio na

UTEP, pelas sugestões científicas e bom humor.

Ao Prof. Dr. Ivarne Tersariol, pelos ensinamentos sobre enzimologia,

contribuição científica e idéias malucas.

A todos os professores dos departamentos de Microbiologia, Imunologia e

Parasitologia, Biofísica e Bioquímica da UNIFESP, pelos ensinamentos, equipamentos e

sugestões que ajudaram na minha formação. Em especial aos professores Travassos,

Elaine, Zoilo, Sérgio, Cida, Olga, Clara e Juliano, os quais admiro e são excelentes

pesquisadores.

Aos meu amigos de laboratório Luizão, Kátia, Luciane, Wagner, Antonio,

Milene, Graziela, Ernesto, Lívia, Milca, Thiago, Tânia, Flavia e Aguinaldo. Estes são

meus irmãos científicos, pessoas com quem convivi e compartilhei a maior parte do meu

tempo. Por isso agradeço pelos bons momentos, ajuda na hora do desespero, piadas com

ou sem a menor graça, seminários e é claro por me aturarem.

vii

Aos secretários Marcelo, Marcinha e Mércia, que estão sempre nos ajudando

com as burocracias, são muito eficientes e sempre sabem de tudo, até a matéria da capa

de jornal do dia seguinte.

Aos amigos de diversos laboratórios Rodrigo, Andrey, Thaysa, Xandão,

Hebeler, Geisa, Fabi, Luiz, Fizão, Bel, Ellen, Bianca, Mário, Ricardo, Simone,

Júnior, Cris, Roti, Patrícia, Rafael, Júlia, Teresa, Alberto, Fernando e Izaura por

tornar a nossa vida dentro de um laboratório muito mais alegre, ás vezes conversar sobre

outros assuntos além de resultados de experimentos faz muito bem. A compania de vocês

nestes anos com certeza foi muito importante.

A todos os funcionários da disciplina de Biologia Celular, em especial à Maria,

Américo, Claudeci, Claudio, Antonio e Sebastiana pela dedicação, esforço e

competência. Sem vocês o “andar não anda”. Obrigado por tudo.

Aos amigos da minha turma de graduação Bio 32 pelos 4 anos de convivência e

troca de experiência. Espero que se tornem grandes pesquisadores para um dia eu contar

para meu filho que eu tive a oportunidade de estudar com vocês.

Aos amigos do colegial, do clã e da faculdade Candiru, Danylo, Okano, César,

Yassumassa, Allan, Ogro, Fabrício, Bariri, Dragão, Kpone, Xarope, Ivan, Carlitos,

Luciano, Zé, Emperor, Henrique, Rodrigo, Marcelo, Shyvaia, Reclamador, Truz,

viii

Marcinho, Vagnão e Guto pelos momentos engraçados, viagens, bebedeiras, momentos

inusitados e alegria. Esses são meus amigos e com certeza contribuíram para a formação

da minha pessoa, meu caráter e até mesmo influenciaram na minha personalidade. A

segunda família.

Às bandas Pedreiros do Apocalipse e Rockcídio das quais sou integrante.

Através da música esqueço de todos os meus problemas, amarguras, tristezas e me sinto

muito mais confiante para enfrentar tudo. Também gostaria de agradecer ao Palmeiras, o

melhor time do mundo, pelas alegrias e títulos conquistados que me acompanham desde o

nascimento. Certamente ele não anda “bem das pernas” nos últimos anos, mas tenho

certeza que voltará a ganhar dezenas de títulos mundo afora.

À FAPESP e CNPq pelo financiamento deste trabalho.

ix

LISTA DE ABREVIATURAS

BSA: albumina sérica bovina

ELISA: “enzyme-linked immunosorbent assay”

Con-A: concanavalina A

DO: densidade ótica

PMSF: fluoreto de fenil-metil-sulfonil

F12: meio mínimo de cultura

IFN-γ: intérferon gama

IL: interleucina

NO: óxido nítrico

PBS: tampão salino tamponado com fosfato

PCM: paracoccidioidomicose

ATP: 5’-trifosfato de adenosina

Npys: “S-3-nitro-2-pyridinesulfenyl”

Abz: ácido orto-aminobenzóico

EDDnp: 2,4-dinitrofeniletilenodiamina

GalMan: galactomanana

HSP: “heat shock protein”

pHMB: para hidroximercuriobenzoato

EDTA: ácido etilenodiamino tetra-acético

E-64: “4-[(2S,3S)-3-carboxyoxiran-2-ylcarbonyl-L-leucylamido(4-guanidino)butane”

IPTG: isopropil-tiol-galactopiranosídeo

x

LacZ: β-galactosidase

min: minuto

nt: nucleotídeo

mRNA: RNA mensageiro

PCR: reação de polimerização em cadeia

SDS: dodecil sulfato de sódio

SDS-PAGE: eletroforese em gel de poliacrilamida

sec: segundos

TBE: tampão tris-borato-EDTA

TEMED: N,N,N,N´-tetrametilenodiamina

UV: ultravioleta

PbST: serino-tiol proteinase extracelular de P. brasiliensis

xi

ÍNDICE

LISTA DE ABREVIATURAS ix

INTRODUÇÃO 1

1.1 Aspectos gerais sobre o Paracoccidioides brasiliensis 1

1.2 Paracoccidioidomicose humana 4

1.3 Parede celular, transição e virulência 10

1.4 Proteinases de P. brasiliensis 17

OBJETIVOS 29

CAPÍTULO 1: Proteinases de Paracoccidioides brasiliensis: serino-tiol 30

proteinase extracelular e Pb Lon mitocondrial.

1.1 Artigo: Modulation of the exocellular serine-thiol 32

proteinase activity of Paracoccidioides brasiliensis by

neutral polysaccharides

1.2 Artigo: C-Npys (S-3-nitro-2-pyridinesulfenyl) and 34

peptide derivatives can inhibit a serine-thiol proteinase activity

from Paracoccidioides brasiliensis

1.3 Artigo: The PbMDJ1 gene belongs to a conserved 37

MDJ1/LON locus in thermodimorphic pathogenic fungi

and encodes a heat shock protein that localizes to both the

mitochondria and cell wall of Paracoccidioides brasiliensis

xii

CONCLUSÕES: Capítulo 1 39

CAPÍTULO 2: Moléculas secretadas de Paracoccidioides 42

brasiliensis: proteínas associadas à parede celular diferencialmente

expressas e resultados preliminares sobre vesículas extracelulares INTRODUÇÃO 43

2.1 Manuscrito: Proteomic analysis of differentially 49

expressed cell wall-associated proteins from the human

pathogen Paracoccidioides brasiliensis

2.2 Moléculas secretadas de Paracoccidioides brasiliensis: 80

resultados preliminares sobre vesículas extracelulares

Materiais e métodos 81

Resultados e Discussão 88

CONCLUSÕES: capítulo 2 102

REFERÊNCIAS BIBLIOGRÁFICAS 104

ANEXO 1 117

ANEXO 2 121

1

Introdução

1.1 Aspectos gerais sobre o Paracoccidioides brasiliensis

Paracoccidioides brasiliensis é o fungo dimórfico térmico causador da

paracoccidioidomicose (PCM) humana. A PCM é uma micose sistêmica granulomatosa

que pode ser letal se não tratada adequadamente. O fungo foi originalmente descrito por

Adolpho Lutz (Figura 1) em 1908, quando isolou o fungo de lesões orais e de linfonodo

cervical no Instituto Biológico de São Paulo (Brasil). Observando os exames histológicos

de um de seus pacientes, declarou que a ausência de esférulas com esporos no seu interior

diferenciava-os de especimens característicos de coccidioidomicose, descrita previamente

na Argentina por Posadas em 1892 (Posadas A., 1892). Mesmo alguns anos após a

descoberta, o P. brasiliensis era comumente confundido com o fungo Coccidioides

immitis, justificando seu nome, já que o prefixo para significa “similiar a”. Análises de

sequências do gene do RNA ribossomal revelaram que os fungos causadores de micoses

granulomatosas Histoplasma capsulatum, Coccidioides immitis, Blastomyces dermatitidis

e P. brasiliensis estão filogeneticamente relacionados e pertencem à mesma família

(Bialek et al., 2000; Peterson et al., 1998). Atualmente o P. brasiliensis é classificado

como Eukaryota do reino Fungi, filo Ascomycota, subfilo Pezizomicotina, classe

Eurotiomycetes, ordem Onygenales, família Onygenaceae, gênero Paracoccidioides e

espécie brasiliensis.

2

Figura 1. Retrato de Adolfo Lutz

Em condições ambientais (25oC), o fungo cresce na fase miceliana, na qual se

apresenta como hifas delgadas de 1 a 3 µm, possuindo como formas reprodutivas os

aleuroconídeos, artroconídeos, artroaleuroconídeos e clamidosporos intercalares

(Restrepo-Moreno, 1994). Os conídeos são unicelulares, porém quando incubados a 37oC

ou alojados nos pulmões (McEwen et al., 1987) transformam-se em leveduras

multinucleadas (Figura 2). As leveduras do fungo são esféricas ou ovais de 4 – 30 µm de

diâmetro, com paredes refringentes, múltiplos brotamentos (Lacaz, 1994), múltiplos

vacúolos, corpos lipídicos e um grande número de mitocôndrias (Kashino et al., 1987). A

forma sexuada ainda não é conhecida e recentemente o P. brasiliensis foi classificado

como haplóide ou aneuplóide (Almeida et al., 2006). Estimativas mostram que o genoma

haploide varia entre 26 a 35 Mb, distribuído em 4 a 5 bandas cromossômicas de 2 a 10

Mb (Feitosa et al., 2000; Almeida et al., 2006).

3

Figura 2. Dimorfismo do P. brasiliensis

Evidências experimentais indicam que o tatu é um reservatório natural do P.

brasiliensis (revisão em Restrepo et al., 2001). O fungo foi isolado de duas espécies

diferentes (Figura 3), o Dasypus novemcintus (Bagagli et al., 1998) e o Cabassou

centralis (Corredor et al., 2005). Em áreas hiper-endêmicas da PCM, o fungo foi isolado

em cerca de 75 a 100% dos tatus de nove bandas (Bagagli et al., 1998; 2003; Restrepo et

al., 2000). Evidências experimentais confirmaram que o P. brasiliensis isolado de

pacientes e tatus possuem o mesmo perfil molecular, antigênico e de virulência,

sugerindo que os mesmos isolados têm a capacidade de infectar o homem e o animal

(Sano et al., 1999; Hebeler-Barbosa et al., 2003). Infecções em animais domésticos e

selvagens também já foram descritas (Ricci et al., 2004; Garcia et al., 1993; Grose e

Tramsitt, 1965; Johnson e Lang, 1977).

4

Dasypus novemcinctus Cabassou centralis

Figura 3. Espécies de tatus de onde foi isolado P. brasiliensis.

Um amplo estudo de multilocus, para o qual foram comparadas seqüências de oito

regiões de 5 locus diferentes em 65 isolados de P. brasiliensis provenientes de diversas

áreas endêmicas da América Latina, classificou-os em 3 espécies filogeneticamente

distintas, a saber S1, PS2 e PS3 (Matute et al., 2006). O grupo S1 foi composto por 38

isolados do Brasil, Peru, Venezuela, Argentina e Paraguai, para os quais foram apontadas

diversas evidências de reprodução sexuada. O grupo filogenético PS2 compreendeu 6

isolados, 5 brasileiros e 1 venezuelano, e sua especiação ocorreu mais provavelmente

pela incapacidade da transmissão genética com o grupo S1 do que pelo isolamento

geográfico, uma vez que se situam nos mesmos países. O último grupo, PS3, constituído

por 21 isolados colombianos, foi considerado uma linhagem evolucionária independente

provavelmente causada pelo isolamento geográfico.

1.2 Paracoccidioidomicose humana

5

Durante as décadas de 30 e 40 houve um aumento de novos casos no Brasil de

PCM, principalmente com lesões cutâneas e mucosas, quando ficou conhecida como

blastomicose brasileira ou blastomicose sul-americana. Observações mais detalhadas

confirmaram que a PCM não era limitada a lesões mucocutâneas e de pele, mas sim uma

doença com amplo espectro de manifestações clínicas e patológicas (revisão em Lacaz,

1994). A PCM ocorre principalmente no Brasil, Argentina, Venezuela e Colômbia

(Figura 4), onde é endêmica em áreas rurais, com temperaturas em torno de 18 a 25 oC,

elevado índice pluviométrico (800 a 2000 mm por ano), altitude entre 50 a 1.300 metros e

predomínio de florestas nativas (Bagagli et al., 1998). Também há relatos de PCM na

Europa, nos Estados Unidos, África, Ásia e Oriente Médio, porém nesses casos todos os

pacientes viveram por anos em áreas endêmicas da América Latina. Não há registros de

casos em Belize, Nicarágua, Guiana, Guiana Francesa, Suriname, Chile ou na maioria das

ilhas caribenhas (Wanke e Londero, 1994). Estima-se que cerca de 10 milhões de pessoas

estejam infectadas e que a incidência anual da doença ativa em áreas endêmicas seja de 1

a 3 indivíduos para cada 100.000 habitantes (San-Blas et al., 2002). Entre 1980 e 1995, a

PCM destacou-se como a 8a causa de mortalidade por doenças infecciosas e parasitárias

de caráter crônico e apresentou a mais elevada taxa de morbidade entre as micoses

sistêmicas no Brasil (Coutinho et al., 2002).

6

Figura 4. Regiões endêmicas da paracoccidioidomicose (retirado de Shikanai-Yasuda et al., 2006)

A infecção do hospedeiro humano ocorre através da inalação de conídeos do

fungo que se instalam nos pulmões, onde se transformam em leveduras que causam a

infecção (McEwen et al., 1987). As lesões podem regredir com a destruição do fungo,

regredir com focos de fungos quiescentes que podem causar a doença posteriormente, ou

progredir imediatamente com o desenvolvimento dos sintomas da PCM. Na maior parte

dos casos há uma infecção sub-clínica com cura espontânea (revisão em Lacaz, 1994),

sugerindo uma predisposição genética para a progressão da doença (Brummer et al.,

1993). No Brasil sabe-se que o risco de desenvolver a PCM é 4,3 vezes maior em

indivíduos portadores de antígeno HLA-B40 (Lacerda et al., 1988); na Colômbia,

predominam pacientes com antígenos HLA-A9 e HLA-B13 (Restrepo et al., 1983).

De acordo com registros médicos, a doença ativa manifesta-se na forma aguda

(juvenil ou linfática) em 5 a 10% dos casos, ou na forma crônica (tipo adulto ou

7

pulmonar) na maioria dos indivíduos. A PCM juvenil progride rapidamente, atingindo o

sistema retículo endotelial de adultos jovens (baço, fígado, linfonodos e medula óssea). A

PCM adulta progride lentamente, é prevalente em homens de 30 – 60 anos, e pode ser uni

ou multifocal. A disseminação para outros órgãos ocorre pela via hematogênica e/ou

linfática. A baixa incidência da PCM em mulheres (13 vezes menor) está possivelmente

ligada com a presença do 17 β-estradiol, que é um inibidor da transformação de micélio

para levedura in vitro (Clemons et al., 1989).

Diferenças na evolução da PCM devem ter relação com a existência de isolados

do P. brasiliensis apresentando variação no grau de virulência e de sua interação com

fatores inerentes ao hospedeiro, como idade, sexo, raça, grau de nutrição e especialmente

o status imunológico. Modelos experimentais de camundongos isogênicos foram

selecionados quanto à susceptibilidade para infecção com P. brasiliensis isolado Pb18

(revisão em Calich et al., 1998). Em infecções intratraqueais e intraperitoniais, os animais

da linhagem A/Sn foram os mais resistentes, já que apresentaram um padrão de

imunidade predominante para o tipo Th-1. Camundongos B10.A, e em menor grau

Balb/c, foram susceptíveis à infecção, observando-se neles uma resposta imune Th-2

orientada, onde prevalesceram altos títulos de anticorpos, imunidade celular

comprometida e produção insuficiente de IFN-γ (Calich et al., 1985, 1994; Cano et al.,

1995).

A depressão da imunidade mediada por células está fortemente relacionada à

progressão da doença (Mota et al., 1985). A defesa do organismo é mediada pela

formação de granulomas compostos de células gigantes e epitelióides com a presença de

polimorfonucleares e leucócitos próximos ao fungo (Franco et al., 1987). Em pacientes

8

com infecções tênues, observam-se granulomas compactos com reduzido número de

fungos, enquanto em pacientes com sintomas graves, pode-se observar granulomas

frouxos e um influxo celular de polimorfonucleares, macrófagos e células mononucleares

com capacidade reduzida de combater o fungo (Bernard et al., 1994).

Estudos realizados in vitro com macrófagos mostraram que o P. brasiliensis é

capaz de se multiplicar intracelularmente, destruir o macrófago e liberar inúmeras células

leveduriformes (Brummer et al., 1989). Porém macrófagos ativados com IFN-γ possuem

a capacidade de inibir a replicação do fungo em 65 a 95% dos casos (Moscardi-Bacchi et

al., 1994). Estudos mais recentes mostram que macrófagos ativados com IFN-γ e TNF-α

são capazes de produzir óxido nítrico (NO), que pode levar à morte do fungo em

concentrações ideiais (Nascimento et al., 2002). Além disso, os macrófagos e monócitos

são capazes de produzir TNF-α, que é uma citocina envolvida no recrutamento e ativação

de células inflamatórias e na formação do granuloma (Kindler et al., 1989).

Recentemente foi demonstrado que monócitos ativados com essa citocina são capazes de

liberar H2O2 e causar a destruição do fungo (Carmo et al., 2006). Os polimorfonucleares

constituem outro tipo celular que tem efeito fungicida e fungistático aumentado por IFN-

γ e GM-CSF (Kurita et al., 2000).

A hiperreatividade da imunidade humoral em pacientes com PCM ocorre

principalmente nas formas disseminadas e severas. Cerca de 90% dos pacientes com

sintomas clínicos possuem anticorpos específicos das classes IgG, IgM e IgA em 98, 45 e

33% dos casos, respectivamente (Biagioni et al., 1984). Pacientes com PCM apresentam

ativação policlonal de células B, como ocorre normalmente em outras doenças

infecciosas (Ortiz-Ortiz et al., 1980; Galvão-Castro et al., 1994; Chequer-Bou-Habib et

9

al., 1989). O papel das imunoglobulinas na PCM ainda não está totalmente esclarecido, e

elas possivelmente não são relevantes para proteção nos estágios avançados da doença,

quando os títulos de anticorpos contra antígenos fúngicos são elevados (Restrepo et al.,

1982). Porém foi verificada uma alta disseminação do fungo em animais que produzem

baixas taxas de anticorpos, indicando que a imunidade humoral estaria participando

ativamente na contenção da doença (Carvalhaes et al., 1986). Também é importante

ressaltar que experimentos de imunização passiva utilizando anticorpos monoclonais

anti-gp70 (Mattos Grosso et al., 2003), em estágios iniciais da infeção, ou anti-gp43

(Buissa-Filho et al., manuscrito em preparação) em camundongos com a doença

estabelecida, impediram quase completamente a formação de granulomas no pulmão.

O tratamento da PCM é prolongado por anos de terapia. São utilizados os

derivados de azol, além das sulfonamidas, e anfotericina B (Mendes et al., 1994). O

primeiro derivado azólico, cetoconazol, demonstrou uma eficácia de 90% e seu

mecanismo de ação está relacionado à inibição da síntese de ergosterol de membrana ao

ligar-se em enzimas relacionadas ao citocromo P450, porém apresenta efeitos colaterais

como anorexia, náusea, vômito e diminuição dos níveis de testosterona. Outros derivados

como itraconazol ou fluconazol parecem ter uma eficácia semelhante ao cetoconazol,

porém com menos adversidades. A anfotericina B foi a droga mais utilizada durante

muitos anos. É um composto poliênico capaz de se ligar ao ergosterol da membrana

fúngica alterando sua permeabilidade, mas atualmente só é utilizado em casos graves ou

de resistência a outras drogas, porque os efeitos colaterais são muito acentuados,

notadamente nos rins e fígado. As sulfonamidas foram utilizadas pela primeira vez em

1940, quando a PCM era considerada uma doença incurável. Essa classe de antibióticos

10

age como antagonista metabólico competitivo do ácido para-aminobenzóico, interfere na

síntese do ácido fólico e consequentemente na produção de DNA. Seus derivados ainda

são utilizados quando combinados com outras drogas por causa do baixo custo e eficácia.

A cura da PCM é caracterizada pela falta de anticorpos circulantes contra antígenos

específicos de P. brasiliensis. Deve-se estender a manutenção do tratamento por um ano

após a negativação dos testes sorológicos e o paciente deve ser reavaliado

periodicamente. A remissão e as sequelas de fibrose pulmonar são frequentes na PCM

(revisão em Brummer et al., 1993).

1.3 Parede celular, transição e virulência

A transição do P. brasiliensis de micélio para levedura (Figura 5) é um

requerimento essencial para a ocorrência da PCM e de outras doenças causadas por

fungos dimórficos (Namecek et al., 2006). A diferenciação morfológica que o fungo sofre

no hospedeiro, por outro lado, é consequência de um conjunto de alterações bioquímicas

e fisiológicas desencadeadas pelo aumento de temperatura, a qual estimula uma

adaptação no padrão transcricional (Nunes et al., 2006) e/ou regulatório em geral de

genes e proteínas.

11

Figura 5. Transição do P. brasiliensis de micélio para levedura após aumento de temperatura de

incubação de 26 oC para 36 oC pelos tempos indicados (adaptado de Goldman et al., 2003).

As proteínas de choque térmico (Hsps) formam uma grande classe de moléculas

relacionadas à variação de temperatura, porém também podem eventualmente ser

induzidas por estresse nutricional, osmótico, oxidativo e por substâncias tóxicas (revisão

em Walter e Buchner, 2002). Em P. brasiliensis já foram caracterizadas algumas

proteínas de choque térmico, a saber, Hsp70 (Diez et al., 2002), Hsp60 (Izacc et al.,

2001), Hsp100 ou ClpB (Jesuíno et al., 2002), Mdj1 (Batista et al., 2006) e Lon (Barros e

Puccia, 2001).

Os eventos moleculares que podem afetar a transição são pouco conhecidos até o

momento, principalmente pela falta de um sistema padronizado de transformação do

fungo, que ocorreu apenas recentemente (Almeida et al., 2007). Contudo estudos

abrangentes tentando detectar proteínas e genes envolvidos no dimorfismo ou

relacionados à virulência estão sendo realizados na última década. Trabalhos prévios

identificaram algumas proteínas (Cunha et al., 1999; Salem-Izacc et al., 1997) e genes

(Silva et al., 1999; Venâncio et al., 2002) diferencialmente expressos nas duas fases do

fungo. Atualmente há 2 grandes bancos de sequências expressas (ESTs) de P.

12

brasiliensis. O primeiro foi desenvolvido pelo grupo paulista liderado por Goldman e

colaboradores (2003), que identificaram 4.692 genes do isolado Pb18 recém-isolado de

baço de camundongo B10.A infectado intraperitonealmente. O segundo foi produzido por

grupos do centro-oeste brasileiro liderados pelas pesquisadoras Maria Sueli Felipe e Célia

Maria A. Soares, que detectaram 6.022 genes expressos nas fases miceliana e

leveduriforme do isolado Pb01 (Felipe et al., 2003 e 2005). Uma utilização imediata das

sequências identificadas de P. brasiliensis foi detectar transcritos diferencialmente

expressos na transição de micélio para levedura, que é fundamental para o

estabelecimento da infecção. Para tal, foram utilizadas genotecas de subtração, macro e

microarray, além de comparações in silico (Marques et al., 2004; Felipe et al., 2005;

Nunes et al., 2005; Ferreira et al., 2006).

Numa tentativa de identificar genes envolvidos com o processo de infecção,

Tavares e colaboradores (2007) estudaram a transcrição gênica de 1.152 genes

selecionados após 6 horas de co-cultura do Pb01 com macrófagos peritoneais murinos.

Nesse período, 90% das células estavam internalizadas ou aderidas aos fagócitos.

Utilizando hibridização em microarray, verificou-se que 152 transcritos aumentaram pelo

menos 2 vezes quando comparados ao controle crescido in vitro. Diversos genes

relacionados ao estresse oxidativo sofreram aumento do acúmulo de mRNA, como por

exemplo a superóxido dismutase 3, Hsp60 e QCR8 (subunidade do citocromo oxidase c.)

Em outro estudo sobre a expressão diferencial de genes de P. brasiliensis em condições

de infecção, Bailão e colaboradores (2006) analisaram duas populações de cDNA por

subtração. Uma delas, proveniente de fígado de camundongos infectados, revelou que

mRNAs de genes ligados à defesa celular, produção de melanina e aquisição de ferro

13

estavam aumentados. Em outra análise, o fungo foi colocado brevemente na presença de

sangue humano, onde se observou que genes principalmente relacionados ao

remodelamento e síntese da parede foram detectados em maior abundância.

Como consequência da transformação dimórfica, um importante compartimento

celular que sofre diversas modificações, das quais algumas são conhecidas, é a parede

celular. A parede celular era considerada um compartimento estático e basicamente

estrutural, porém hoje se sabe que é uma estrutura dinâmica composta de glicoproteínas,

proteínas, lipídeos e principalmente polissacarídeos do tipo quitina e glucanas (de Groot

2005). No P. brasiliensis, grande parte das 1,3-glucanas da parede fúngica sofrem uma

alteração da conformação β, na fase miceliana, para α, na levedura. Além dessa

conversão, a levedura possui maior quantidade relativa de quitina (média de 43%), e

menor concentração de proteínas (média de 11%), quando comparada ao micélio que

contém em média de 13 e 33% respectivamente. Em ambas as formas, os lipídeos

respondem por uma média de 8% e as glucanas por média de 42% da parede celular

(Kanetsuna et al., 1969). Em termos glucanas, estima-se que a fase leveduriforme possui

cerca de 83% de α-1,3-glucana e 17% de β-1,3-glucana, enquanto a fase miceliana

contém 100% de β-1,3-glucana. As β-1,3-glucanas são receptores de dectina-1 (Brown e

Gordon, 2003) e são, portanto, inflamatórias (Silva e Fazioli, 1985; Figueiredo et al.,

1993). Em H. capsulatum, cepas contendo o gene que codifica a α-1,3-glucana sintase

mutado ou seu RNA interferido tiveram sua virulência atenuada (Rappleye et al., 2004).

Nesse fungo, foi demonstrado que a α-1,3-glucana localiza-se na camada mais externa da

parede recobrindo a camada de β-1,3-glucana, o que bloqueia a ligação da dectina-1 e

inibe em 5 vezes a produção de TNF-α por células fagocíticas, com a consequente

14

redução da eficácia da resposta imune (Rappley et al., 2007). No P. brasiliensis, pelo

menos parte da β-1,3-glucana pode estar acessível na superfície (Diniz et al., 2004).

No P. brasiliensis, além das alterações nas glucanas, foi observado um aumento

significativo de 3 vezes mais quitina na fase leveduriforme que no micélio. A quitina é

um importante componente estrutural da parede celular dos fungos. Em P. brasiliensis já

foram identificadas 4 quitinas sintases (Niño-Vega et al., 2000). Em C. albicans existem

4 genes para quitina sintase, sendo um deles essencial (Munro et al., 2001). Mutantes de

C. albicans contendo os outros 3 genes de quitina sintase deletados foram mais

susceptíveis a agentes que estressam a parede celular, como por exemplo o calcofluor, e

apresentaram divisão celular deficiente, especialmente na formação de um septo íntegro

(Munro et al., 2007). Recentemente foi demonstrado que uma lectina de P. brasiliensis,

conhecida como paracoccina, foi capaz de interagir com a quitina da parede celular.

Anticorpos reativos com essa proteína foram capazes de reduzir o crescimento do fungo

in vitro, que apresentou colônias menores e células menos agregadas quando comparadas

ao controle. A presença destes anticorpos também alterarou o padrão da distribuição da

quitina na parede celular, sugerindo que a interação da paracoccina com quitina é um

evento importante na organização da parede e consequentemente no crescimento fúngico

(Ganiko et al., 2007).

Poucas moléculas do P. brasiliensis têm sido apontadas como fatores de

virulência com alguma base experimental, a saber, a glicoproteína gp43 (Puccia et al.,

1986), melanina (Gomez et al., 2001), uma serino-tiol proteinase extracelular (Carmona

et al., 1995) e adesinas de 19 e 32 kDa (Gonzales et al., 2005), 30 kDa (Andreotti et al.,

2005) e a gliceraldeído-3-fosfato desidrogenase (GADPH) (Barbosa et al., 2006), que

15

possuem habilidade de ligação a proteínas associadas à matriz extracelular. Ensaios

utilizando células de P. brasiliensis pré-incubadas com anticorpos policlonais anti-

GADPH ou com GADPH recombinante inibiram a aderência e internalização do fungo

por pneumócitos.

A lacase é a enzima responsável pela síntese de melanina, a qual é um fator de

virulência reconhecido em Cryptococcus neoformans (Kwon-Chung et al., 1982) e foi

também encontrada em P. brasiliensis (Gomez et al., 2001). Estudos recentes

demonstraram que células de P. brasiliensis melanizadas são 3,5 vezes menos

internalizadas por macrófagos e possuem menor susceptibilidade a várias drogas

antifúngicas como anfotericina B, fluconazol, cetoconazol e itraconazol (da Silva et al.,

2006).

A gp43 é o antígeno majoritário do fungo (Puccia e Travassos, 1991) e protetor da

PCM murina (Taborda et al., 1998; Pinto et al., 2000). Além de induzir resposta imune

humoral a gp43 contém epitopos que suscitam uma resposta imune celular de

hipersensibilidade tardia (Rodrigues et al., 1994; Saraiva et al., 1996). Taborda et al.

(1998) mostraram que camundongos imunizados com a gp43 nativa produzem uma

resposta T-CD4+ do subtipo Th1 capaz de liberar IFN-α e IL-2, porém não IL-4 ou IL-5.

Através de ensaios de linfoproliferação com animais imunizados com a gp43, os autores

identificaram o epitopo imunodominante da resposta imune celular em um peptídeo

interno de 15 aminoácidos, denominado P10. Animais imunizados com a gp43 nativa, o

gene PbGP43 e o peptídeo P10 mono ou tetravalente (M10) foram protegidos contra

subsequente desafio intratraqueal por leveduras virulentas do P. brasiliensis (Taborda et

al., 1998 e 2003; Pinto et al., 2000), sugerindo sua utilidade como vacina. Além disso, o

16

peptídeo P10 tem utilidade terapêutica comprovada em modelo animal (Marques et al.,

2006).

A gp43 é capaz de modular a infecção experimental em camundongos. Estudos in

vivo monstraram um aumento da infecção intratesticular de hamster quando os animais

foram infectados com P. brasiliensis previamente recoberto com laminina murina

(Vicentini et al., 1994; Lopes et al., 1994). Essa ligação in vivo e in vitro foi inibida com

anticorpos monoclonais anti-gp43 (Gesztesi et al., 1996). Recentemente foi evidenciada a

interação da gp43 com a fibronectina, um outro componente asssociado à matriz

extracelular (Mendes-Giannini et al., 2006).

Nosso grupo constatou um extenso polimorfismo no exon 2 do gene PbGP43

(Morais et al., 2000), principalmente entre as posições 578 e 1160. As seqüências mais

polimórficas foram encontradas no Pb2, Pb3 e Pb4, as quais codificam uma proteína

básica com pI em torno de 8,0. Essas seqüências são filogeneticamente distantes das

demais, incluindo aquela de Pb18, o que contribuiu para a classificação desses isolados

no grupo filogenético PS2 (Matute et al., 2006). Um trabalho posterior do nosso grupo

investigou a virulência de isolados pertencentes a grupos genotípicos de PbGP43

distintos em camundongos B10.A pelas vias intraperitoneal, intratraqueal e endovenosa

(Carvalho et al., 2005). Os camundongos infectados com Pb2, Pb3 e Pb4 sofreram uma

infecção branda e regressiva, com baixo índice de mortalidade, em contraste com os

demais, inclusive Pb18; Pb12 foi o mais agressivo. Em infecções posteriores com Pb3,

Pb18 e Pb12 adaptados in vivo, verificou-se que Pb3 induziu um tipo de resposta imune

com crescente produção de IFN-γ e predomínio de IgG2a e 2b, e IgG3 anti-gp43, ao

contrário de Pb18, que induziu anti-gp43 dos tipos IgG1 e IgA, predominantemente, e

17

níveis decrescentes de IFNγ. Camundongos infectados intratraquealmente com Pb12 não

produziram IFN-γ detectável (Carvalho et al., dados não publicados).

1.4 Proteinases de P. brasiliensis

Enzimas proteolíticas podem desempenhar um importante papel na virulência de

bactérias (Finlay et al., 1989), protozoários (McKerrow et al., 1993) e fungos patogênicos

(Ogrydziak, 1993). Proteinases extracelulares de fungos saprófitas como A. niger e N.

crassa são secretadas principalmente para a nutrição. Fungos patogênicos, no entanto,

parecem ter adaptado funções especializadas para essas hidrolases durante a infecção,

como por exemplo, desestruturar a membrana celular para facilitar a adesão e invasão do

hospedeiro como demonstrado em plantas (Chaffin et al., 1998) e em insetos (Smithson

et al., 1995), ou causar danos a células e moléculas do sistema imune (Salyers et al.,

1994). A família das SAPs (aspartil proteinases secretadas), constituída de 10 genes de C.

albicans, é a melhor caracterizada entre os fungos patogênicos. Através de estudos de

diversas naturezas, demonstrou-se que as SAPs 1, 2 a 3 estão ativamente envolvidas em

infecções nas mucosas, enquanto as SAPs 4, 5 e 6 são ativas especialmente nas infecções

sistêmicas (revisão em Naglik et al., 2003).

Todas as proteinases catalisam a hidrólise de ligações peptídicas (CO-NH) de

proteínas, todavia apresentam diferentes mecanismos de ação e especificidade (Barrett et

al., 1991). As proteinases são classificadas com base no seu mecanismo catalítico e não

pela estrutura tridimensional, especificidade por substratos ou função fisiológica. De

acordo com a natureza dos aminoácidos do sítio catalítico as proteinases pedem ser

divididas em 5 classes: serino, aspartil, cisteíno, treonino e metalo proteinases, além de 9

18

proteinases com mecanismo catalítico desconhecido (Rawlings et al., 2006). Os

inibidores específicos mais comuns utilizados para classificar as proteases são: PMSF

(fluoreto de fenilmetanosulfonila) para as serino, pepstatina para as aspartil, E-64 (trans-

epoxisuccinil-leucilamido(4-guanidino)butano) para as cisteíno e EDTA (ácido

etilenodiaminotetracético) para as metalo proteinases.

As serino proteinases são as mais abundantes, e na maioria dos casos seu

mecanismo catalítico caracteriza-se pela presença da tríade catalítica no sítio ativo da

enzima, composto por histidina, serina e ácido aspártico. O grupamento –OH da serina

age como nucleófilo e ataca o grupo carbonila na ligação peptídica do substrato. O par de

elétrons do nitrogênio da histidina faz a captação do hidrogênio do grupo –OH da serina

(Figura 6). Em conjunto, o grupo carboxila do ácido aspártico forma pontes de

hidrogênio com a histidina, fazendo o par de elétrons muito mais eletronegativo para que

ocorra a hidrólise da ligação peptídica (Hedstrom, 2002). Dentro da classe das serino

proteinases há 68 famílias, das quais a quimotripsina (S1) e a subtilisina (S8) são as mais

numerosas (Rawlings et al., 2006).

19

Figura 6. Mecanismo catalítico geral das serino proteinases (retirado de

http://employees.csbsju.edu/hjakubowski/classes/ch331/catalysis/serprotease3.gif)

As proteinases da família da quimotripsina são endopeptidases subdivididas em

tripsina-like, quimotripsina-like e elastase-like, as quais clivam os substratos após os

seguintes aminoácidos na posição P1 (nomenclatura de Schechter e Berger, 1967):

arginina/lisina, hidrofóbicos e alanina, respectivamente. A maioria dessas proteinases

possui um peptídeo sinal para exportação e uma região N-terminal que necessita ser

clivada para a enzima ficar ativa. Inúmeras funções biológicas de proteinases já foram

identificadas, como por exemplo, digestão de alimentos, resposta imune mediada por IgA

20

(enzimas como a triptase e quimase presentes em mastócitos) e até mesmo na sinalização

dorso-ventral de drosófila (Rawlings et al., 2006).

Pensava-se que as subtilisinas fariam parte da família da quimotripsina, porém

análises de sequência (Smith et al., 1966) e estruturais (Wright et al., 1969) mostraram

diversas diferenças. A maior parte das subtilisinas são endopeptidades com pH ótimo

neutro ou alcalino, termoestáveis e não-específicas. A clivagem ocorre após um

aminoácido hidrofóbico, sendo a caseína comumente utilizada como substrato em ensaios

de clivagem (Rawlings et al., 2006). Em fungos e bactérias, grande parte das subtilisinas

estão envolvidas com nutrição, porém já foram descritos membros dessa família

implicados em virulência, por exemplo a cspA de Streptococcus (Harris et al., 2003).

Dentro das subtilisinas encontra-se um grupo com uma característica peculiar,

especificamente a susceptibilidade à inibição por mercuriais. Nesse caso, as moléculas de

Hg ligam-se a um resíduo de cisteína próximo ao sítio catalítico, como é o caso da

proteinase K (Muller e Saenger, 1993) e da cerevisina (Kominami et al., 1981). Foi

demonstrado por cristalografia de raio-x que a ligação do mercúrio ao grupamento tiol da

cisteína ocasiona uma mudança estrutural no anel imidazólico da histidina do sítio

catalítico (como indicado na Figura 7) provocando alta inibição da atividade hidrolítica

(Muller e Sanger, 1993).

21

Figura 7. Mudança estrutural no anel imidazólico da histidina no sítio catalítico da proteinase K

provocado por mercuriais (Muller e Sanger, 1993).

Em P. brasiliensis, foram detectadas proteinases citosólicas descritas nas duas

fases do fungo (San-Blas et al., 1998). Verificou-se que a atividade enzimática de

extratos totais foi maior na fase miceliana, aparentemente com preponderância de serino

proteinases devido à alta inibição por PMSF e pH ótimo alcalino. Todavia a preparação

enzimática também foi parcialmente inibida por iodocetamida, EDTA e orto-

fenantrolina, indicando a presença de cisteíno e metalo proteinases. Em 2005, Parente e

colaboradores analisaram o transcriptoma de P. brasiliensis quanto à presença de genes

de proteinases. Foram detectadas 15 proteinases ATP-independentes, 12 ATP-

dependentes, 22 subunidades do proteassomo e 4 proteinases relacionadas com a

desubiquitinação. As proteinases ATP-independentes subdividiram-se em 5,6% de

aspartil proteinases, 11,3% cisteíno proteinases, 22,6% de metaloproteinases, 18,8% de

serino proteinases e 41,5% de subunidades do proteossomo. Apenas uma protease da

família S8 das subtilinas de P. brasiliensis foi sequenciada e depositada no GenBank

22

(AAP83193) pelo grupo da Professora Célia M. A. Soares. Essa proteinase chama a

atenção porque seu mRNA foi diferencialmente expresso em leveduras expostas ao soro e

em células isoladas de fígado infectado (Bailão et al., 2006). Não parece, no entanto,

corresponder à PbST (Matsuo et al., 2007, artigo 2).

Uma serino-tiol proteinase extracelular de P. brasiliensis tem sido caracterizada

em nosso laboratório em colaboração com o Departamento de Biofísica da UNIFESP,

especialmente com a Professora Adriana K. Carmona. Pensava-se que a atividade

enzimática em sobrenadante de cultura de P. brasiliensis crescido na fase leveduriforme

era proveniente da gp43 (Puccia et al., 1991), a qual é a proteína secretada mais

abundante. Porém cromatografias de afinidade em coluna contendo monoclonal anti-gp43

revelaram que a atividade não era retida na coluna (Carmona et al., 1995; Puccia et al.,

1999). Puccia e colaboradores (1998) observaram que a atividade proteolítica estava

diretamente ligada ao aumento do pH extracelular da cultura, onde o pico de atividade

ocorreu em pH 7.7 com o fungo crescendo em fase logarítmica. Para caracterizar essa

atividade, Carmona e colaboradores (1995) testaram diversos substratos peptídicos

sintéticos com supresssão intramolecular de fluorescência do tipo Abz/EDDnp (ácido

orto-aminobenzóico e 2,4-dinitrofeniletilenodiamina) (Chagas et al., 1992). O substrato

originalmente clivado foi Abz-MKRLTL-EDDnp, para o qual houve uma clivagem única

entre os aminoácidos leucina e treonina. A partir deste, outros foram sintetizados para

determinar a importância de P2 e P3 (nomenclatura de Schechter e Berger, 1967). O

melhor substrato foi Abz-MKALTL-EDDnp, para o qual a troca da arginina pela alanina

aumentou a atividade catalítica 5 vezes, com pH ótimo em torno de 9. A classificação da

proteinase foi realizada através de inibidores especificificos: houve inibição irreversível

23

por PMSF (serino proteinases) e inibição reversível utilizando agentes redutores por p-

HMB e acetato de mercúrio. Outras classes de inibidores como EDTA (metalo),

pepstatina (aspartil) e E-64 (cisteíno) não tiveram efeito. Por isso a atividade enzimática

foi atribuída a uma serino proteinase contendo um grupo tiol livre (Carmona et al., 1995),

a qual denominamos PbST.

Para estimar a massa molecular da proteinase, foram realizados ensaios de

zimograma com gelatina e “overlay” com gel de agarose contendo o substrato sintético

Abz-MKALTLQ-EDDnp. Em ambos os casos foi observada uma atividade hidrolítica

difusa em torno de 43 a 69 kDa, a qual foi inibida totalmente por PMSF e parcialmente

por p-HMB, provavelmente pela presença de um agente redutor no tampão de amostra

(Puccia et al., 1999). A importância desta proteinase deve-se à clivagem seletiva de

substratos proteicos. Enquanto a maioria das subtilisinas é promíscua, a PbST possui a

capacidade de clivar componentes asssociados à matriz extracelular como laminina,

fibronectina, colágeno tipo IV e proteoglicanos. Porém não foi detectada atividade sobre

colágeno tipo I, fibrinogênio bovino, imunoglobulina G humana, albumina bovina, gp43

e nem mesmo caseína (Puccia et al., 1998). Por esse motivo especula-se que a PbST

possa ser importante na disseminação do P. brasiliensis para outros órgãos durante a

infecção.

A purificação da PbST ativa tem sido impedida por ocorrer em baixas

concentrações protéicas e ter tendência à autólise e à agregação com componentes

polissacarídicos de alto peso molecular (Carmona et al., 1995). Esse componente de alta

massa molecular refere-se à peptidogalactomana cuja estrutura foi analisada em nosso

laboratório (Puccia et al., 1986). Por muito tempo a atividade proteolítica foi concentrada

24

e fracionada a partir de filtrado de cultura saturado com sulfato de amônio, por interação

hidrofóbica em coluna de Phenyl Superose (Amersham Pharmacia), porém a banda da

protease nunca foi detectada com exatidão. Também tentou-se obter informações sobre

sequências internas da PbST pois são de fundamental importância para a aplicação de

uma estratégia direta de síntese de oligonucleotídeos e obtenção de produtos de PCR

correspondentes ao seu gene. Previamente no laboratório foram obtidas informações

sobre a seqüência interna de aminoácidos e do N-terminal de um componente de 55 kDa,

supostamente a PbST (Puccia et al., 1999), enriquecida por cromatografia em

concanavalina A. O perfil peptídico e a comparação com bancos de dados indicou a

presença de sequências homólogas à catalase, além de peptídeos da gp43 e concanavalina

A (dados não publicados). Ou seja, as condições de purificação e eletroforese utilizadas

propiciaram a formação, em preparações concentradas, de um agregado de catalase, gp43

e da própria concanavalina A, migrando em 55 kDa. A atividade proteolítica estava

incluída nesse agregado ou co-migrou com ele, porém a proteinase estava aparentemente

em quantidade insuficiente para ser detectada.

Na tentativa de identificar um fragmento do gene da PbST por PCR utilizando

oligonucleotídeos degenerados para domínios conservados dos sítios catalíticos de

subtilisinas, foi clonado e sequenciado ao acaso um gene homólogo ao da proteinase Lon

(Barros e Puccia, 2001). O gene em P. brasiliensis, denominado PbLON, é constituído de

3.369 pb com 2 introns na porção 3’, e origina uma proteína predita de 1.063

aminoácidos que contem um sinal de importe mitocondrial na região N-terminal. Possui

identidade com os genes LON de S. cerevisiae (73%), Homo sapiens (62%) e E. coli

(56%). O gene completo foi clonado a partir de um fragmento genômico SmaI de

25

aproximadamente 6,5 kb, o qual contem adicionalmente um fragmento 5’ do gene

homólogo ao MDJ1 de S. cerevisiae em direção oposta. Os genes PbLON e PbMDJ1

compartilham a região 5’ intergênica, na qual Batista e colaboradores (2006) mapearam 3

regiões putativas para elementos de choque térmico e um domínio ARE ligante de AP-1,

que é um fator de transcrição envolvido na indução de genes relacionados ao estresse

oxidativo em fungos (Wu et al., 1994; Grant et al., 1996). Em P. brasiliensis, a proteinase

PbLon está localizada exclusivamente na mitocôndria (Batista et al., 2006, artigo 3), onde

as proteínas estão sendo constantemente atacadas por radicais livres devido ao

metabolismo aeróbico, drogas ou condições adversas do hospedeiro. Estudos recentes

demonstraram que a transcrição do gene PbLON aumentou 5 vezes em condições de

estresse oxidativo e provavelmente está relacionada com a homeostase mitocondrial

promovendo a degradação seletiva de proteínas danificadas (Batista et al., 2007).

A proteinase Lon está inserida em uma família própria (S16) constituída por

endopeptidases ATP-dependentes. A primeira Lon foi descoberta em E. coli por Swamy e

Goldberg (1981) e denominada proteinase La. Está presente em todos os tipos de

organismos e o sítio catalítico é formado por uma serina e uma lisina (Besche e Zwickl

2004; Botos et al., 2004). Em E. coli, Lon é formada por 6 unidades idênticas de

aproximadamente 87 kDa e em S. cerevisiae é formada por 7 unidades de 117 kDa com

localização intramitocondrial (Stahlberg et al., 1999). Cada unidade possui 4 domínios

(Figura 8): na região N-terminal há o sítio para importe mitocondrial; na região central

ocorre a ligação e hidrólise do ATP (Fischer et al., 1994) e a discriminação do substrato

pelo domínio SSD (Smith et al., 1999); na região C-terminal localiza-se o sítio catalítico

26

(Amerik et al., 1991). A proteinase é expressa constitutivamente, mas sua expressão pode

aumentar após choque térmico (Van Dyck et al., 1994).

Figura 8. Desenho esquemático dos “motifs” da proteinase Lon

A principal função da Lon é degradar proteínas de meia-vida curta que estão

envolvidas em diversos processos biológicos e proteínas anormais que agregariam na

ausência da proteinase ou do sistema de chaperoninas DnaK (revisão em Tsilibaris et al.,

2006). O mecanismo pelo qual Lon reconhece os substratos ainda não é totalmente

entendido. Apesar de reconhecer preferencialmente certos amino ácidos ou domínios

(Griffith et al., 2004; Ishii et al., 2001), nenhum “motif” consenso ou combinações de

“motifs” foram identificados. A discriminação de substratos parece ocorrer através da

estrutura, ou seja, proteínas alvo de meia-vida curta e anormais não estariam em sua

conformação globular (Jubete et al., 1996; Van Melderen et al., 1996). Em eucariotos, a

deleção ou desregulação do gene LON leva a defeitos celulares. S. cerevisiae mutante

para o gene apresenta deficiência quando crescido em condições aeróbicas e não cresce

em meio de cultura com fonte de carbono não-fermentável; apresenta inclusões elétron-

densas na matriz mitocondrial devido ao acúmulo de proteínas mitocondriais e deleções

no mtDNA (Van Dyck et al., 1999). Em células de mamíferos a proteinase degrada

proteínas oxidadas e está implicada na disfunção mitocondrial relativa à idade (Bota et

al., 2002). Além disso, reduções nos níveis de Lon prejudicam a divisão celular e levam à

27

necrose, sendo que a depleção total leva à apoptose (Bota et al., 2005). A proteinase

também pode estar envolvida com regulação gênica (Langer e Neupert, 1996; Fu et al.,

1997) e virulência em bactéria (Robertson et al., 2000; Boddicker e Jones et al., 2004).

A PbMdj1 faz parte da família das proteínas com domínio J, que compartilham

identidade entre 35 a 59%. Em E. coli, a molécula típica possui um domínio J constituído

de 70 aminoácidos, seguido de um segmento ligante de zinco rico em glicina e uma

região C-terminal pouco conservada (Szabo et al., 1996). As proteínas J são subdivididas

em 3 tipos: o tipo I contém todos os domínios conservados em E. coli, o tipo II não

contém a região ligadora de zinco, e o tipo III possui apenas um domínio J em qualquer

região da proteína (Fliss et al., 1999). Seu papel principal é regular a atividade das Hsp70

cognatas, localizadas em diversos compartimentos celulares (Revisão em Walsh et al.,

2004). Em S. cerevisiae, Mdj1 é a única DnaJ mitocondrial do tipo I e é essencial na

biogênese da mitocôndria funcional (Rowley et al., 1994). O sistema Ssc1 (Hsp70

mitocondrial) / Mdj1 (Hsp 40 mitocondrial) está envolvido na manutenção da

configuração correta das proteínas mitocondriais ou da solubilidade das proteínas

desnaturadas que serão degradadas pela proteinase Lon e outras proteinases ATP-

dependentes (Figura 9) (Wagner et al., 1994; Savel’ev et al., 1998).

28

Figura 9. Rede de chaperones da matriz mitocôndrial de S. cerevisiae (retirado de Voos et al., 2002)

29

Objetivos

• Em relação à proteinase serino-tiol extracelular de P. brasiliensis (PbST), os

objetivos foram purificar e identificar a proteína, estudar sua modulação com

açúcares e estudar a inibição por compostos (2,4-dinitrophenyl)ethylenediamine

(Npys) acoplados com substratos peptídicos.

• Em relação à PbLon, objetivamos fazer a expressão heteróloga, purificação da

proteína PbLon recombinante, obtenção de anticorpos e localização celular. Um

estudo sobre a especificidade da enzima com substratos peptídicos foi previsto se

a expressão heteróloga resultasse em proteinase ativa.

• Objetivos posteriores do projeto incluíram:

a) a caracterização preliminar de vesículas extracelulares do fungo, as quais poderiam

conter a PbST.

b) a análise proteômica da expressão diferencial de proteínas associadas à parede celular.

Uma introdução específica e um racional para esses objetivos estão apresentados

no capítulo 2 desta tese.

30

Capítulo 1

Proteinases de Paracoccidioides

brasiliensis: serino-tiol extracelular

(PbST) e PbLon mitocondrial

31

Proteinase serino-tiol extracelular

de P. brasiliensis (PbST)

32

Artigo 1: Modulation of the exocellular serine-thiol proteinase activity of

Paracoccidioides brasiliensis by neutral polysaccharides.

Alisson L. Matsuo, Ivarne L. Tersariol, Silvia I. Kobata, Luiz R. Travassos, Adriana K.

Carmona e Rosana Puccia

Microbes and Infection 2006, Vol 8 (1) 84-91

33

Racional

A serino-tiol proteinase de P. brasiliensis (PbST) tem sido caracterizada em nosso

laboratório em colaboração com o Departamento de Biofísica da UNIFESP, como

detalhado na Introdução geral. Desde as primeiras tentativas de purificação da enzima a

partir de sobrenadantes de cultura (Carmona et al., 1995), ficou nítido que a proteinase

agrega-se com componentes glicosilados secretados pelo fungo. Por outro lado, a

atividade proteolítica é retida em coluna de concanavalina A, possivelmente carregada

por componentes glicosilados com os quais se associa. No trabalho apresentado a seguir,

foi estudado o efeito modulatório de polissacarídeos sobre a PbST. Primeiramente tentou-

se observar um efeito por glicosaminoglicanos, visto que este composto está contido na

matriz extracelular e tem a capacidade de modular cisteíno proteinases, porém o resultado

foi negativo. Todavia quando se testou dextrana, na verdade como controle, foi observada

uma intensa modulação da afinidade da enzima pelo substrato, o que ocorreu

posteriormente também com açúcares neutros derivados de uma preparação de mannana

de S. cerevisiae e de galactomanana do sobrenadante de P. brasiliensis. Os dados de

modulação de atividade proteolítica com açúcares neutros são originais e têm uma

possível relevância no controle da atividade da PbST in vivo.

Original article

Modulation of the exocellular serine-thiol proteinase activityof Paracoccidioides brasiliensis by neutral polysaccharides

Alisson L. Matsuo a,1, Ivarne I.L. Tersariol b,1, Silvia I. Kobata c, Luiz R. Travassos a,Adriana K. Carmona c, Rosana Puccia a,*

a Departamento de Microbiologia, Imunologia e Parasitologia, Disciplina de Biologia Celular, UNIFESP, Rua Botucatu, 862, oitavo andar,São Paulo, SP 04023-062, Brazil

b Centro Interdisciplinar de Investigação Bioquímica da Universidade de Mogi das Cruzes, Mogi das Cruzes, SP, Brazilc Departamento de Biofísica da Universidade Federal de São Paulo, São Paulo, SP, Brazil

Received 21 March 2005; accepted 31 May 2005

Available online 08 August 2005

Abstract

Our group characterized an exocellular serine-thiol proteinase activity in the yeast phase of Paracoccidioides brasiliensis (PbST), adimorphic human pathogen. The fungal proteinase is able to cleave in vitro, at pH 7.4, proteins associated with the basal membrane, such ashuman laminin and fibronectin, type IV collagen and proteoglycans. In the present study, we investigated the influence of glycosaminoglycans(GAGs) and neutral polysaccharides upon the serine-thiol proteinase activity by means of kinetic analysis monitored with fluorescence reso-nance energy transfer (FRET) peptides using the substrate Abz-MKALTLQ-EDDnp (Abz = ortho-aminobenzoic acid; EDDnp = ethylenedia-minedinitrophenyl). Only neutral polysaccharides exhibited patterns of interaction with the proteinase, while sulfated GAGs had no effect.Incubation with neutral polysaccharides resulted in a powerful modulation of the enzyme activity, intensely changing the enzyme kineticparameters of catalysis and affinity for the substrate. Commercial dextran at the highest concentration of 20 µM increased 6.8-fold the enzymeaffinity for the substrate. In the presence of 8 µM of purified baker’s yeast mannan, the apparent KM of the enzyme increased about 5.5-fold,reflecting a significant inhibition in binding to the peptide substrate. When an exocellular galactomannan (GalMan) complex isolated fromP. brasiliensis was added to the reaction mixture at 400 nM, the apparent KM and VMAX decreased about threefold. Moreover, GalMan wasable to protect the enzymatic activity at high temperatures, but it caused no effect on the optimum cleavage pH. Our results show a novelmodulation mechanism in P. brasiliensis, where a fungal polysaccharide-rich component can stabilize a serine-thiol proteolytic activity,which is possibly involved in fungal dissemination.© 2005 Elsevier SAS. All rights reserved.

Keywords: Paracoccidioides brasiliensis; Serine-thiol activity; Neutral polysaccharides; Modulation

1. Introduction

Paracoccidioidomycosis (PCM) is a systemic mycosiscaused by the dimorphic fungus Paracoccidioides brasilien-

sis. Acute and sub-acute PCM affect both sexes, progress rap-idly, and the fungi disseminate through the lymphatic sys-tem. Chronic forms are the most common; they predominantlyaffect male adults, starting as a lung granulomatous infectionand eventually involving other organs [1]. The disease isrestricted to Latin America, where P. brasiliensis has alsobeen isolated from the soil and from organs of nine-bandedarmadillos [2]. The fungus penetrates the human body throughinhalation of propagules, which can only establish infectiononce they undergo phase transition to yeasts in the pulmo-nary alveolar epithelium [3,4]. The disease can be fatal if nottreated adequately.

We have previously characterized a subtilisin-like, SH-dependent serine proteinase activity in culture fluids of P. bra-

Abbreviations: Abz, ortho-aminobenzoic acid; AUF, arbitrary units offluorescence; EDDnp, ethylenediaminedinitrophenyl; FPLC, fast perfor-mance liquid chromatography; GAGs, glycosaminoglycans; GalMan, galac-tomannan; PbST, exocellular serine-thiol activity from P. brasiliensis; PCM,paracoccidioidomycosis; p-HMB, sodium 7-hydroxymercuribenzoate;PMSF, phenyl-methylsulfonyl fluoride.

* Corresponding author. Tel.: +55 11 5084 2991; fax: +55 11 5571 5877.E-mail address: rosana@ecb.epm.br (R. Puccia).

1 Alisson L. Matsuo and Ivarne L.S. Tersariol contributed equally to thiswork.

Microbes and Infection 8 (2006) 84–91

www.elsevier.com/locate/micinf

1286-4579/$ - see front matter © 2005 Elsevier SAS. All rights reserved.doi:10.1016/j.micinf.2005.05.021

siliensis grown in the yeast phase [5]. The enzyme is able tohydrolyze fluorescence resonance energy transfer (FRET)peptides flanked by ortho-aminobenzoic acid (Abz) and eth-ylenediaminedinitrophenyl (EDDnp), homologous toMKRLTL. Cleavage occurs on the Leu–Thr bond, at an opti-mum alkaline pH, and is irreversibly inhibited by PMSF, mer-curic acetate and p-HMB (sodium 7-hydroxymercuri-benzoate). Hydrolysis of synthetic peptides and proteinsinvolves an enzyme of high specific activity, which is how-ever unstable and lost after several chromatographic steps aim-ing at its purification. Additionally, it tends to aggregate withglycosylated extracellular fungal components [5]. Early chro-matographic analysis of culture filtrates showed that theenzyme activity was eluted in a single peak from an anion-exchange Resource Q column followed by gel filtration [5].The active fractions, however, yielded several bands in SDS-PAGE stained gels. We then observed that chromatographyof culture fluid supernatants pretreated with (NH4)2SO4 at40% saturation in Phenyl Superose resulted in concentrationof the proteolytic activity in two eluted fractions [6]. Whenthese fractions were analyzed by electrophoresis and in situhydrolysis of both gelatin (zymogram) and Abz-MKALTLQ-EDDnp incorporated into buffered agarose overlays, theserine-thiol exocellular activity appeared as a diffuse and het-erogeneous band in SDS-PAGE gels. It localized to a regionbetween 69 and 43 kDa, but could not be correlated with anysilver-stained component.

The most relevant characteristic of the serine-thiol protein-ase activity of P. brasiliensis is its capacity of selectivelycleaving, in vitro, laminin, fibronectin, type IV collagen andproteoglycans [7]. Since these components are associated withthe basal membrane, such hydrolytic activity could play asignificant role in tissue invasion by the fungus. There are afew reports on the modulation by complex and negativelycharged carbohydrates, more specifically glycosaminogly-cans (GAGs), of the activity of proteolytic enzymes [8]. Inthe present work we describe a novel type of proteinase regu-lation by neutral polysaccharides using the Abz-MKALTLQ-EDDnp substrate.

2. Materials and methods

2.1. Fractionation of the serine-thiol proteinase activity

P. brasiliensis strain 339 (originally provided by Dr.AngelaRestrepo-Moreno, Medellin, Colombia) was cultivated withshaking at 36 °C in a modified YPD medium (0.5% of dia-lyzed, < 10 kDa, Bacto-yeast extract, 0.5% casein peptoneand 1.5% glucose, pH 6.5). The culture supernatant was col-lected by paper filtration at the peak of the proteinase activityagainst Abz-MKALTLQ-EDDnp, when the culture fluid pHwas about 7.0 [7]. Ammonium sulfate was gradually addedto 40% saturation (1.7 M) and the supernatant centrifuged(16,300 × g, 30 min) to discard eventual precipitates. Ali-quots of the supernatant containing ammonium sulfate (50 ml)

were fractionated by fast performance liquid chromatogra-phy (FPLC) (FPLC System - Pharmacia) in a Phenyl Super-ose HR 5/5 (Pharmacia/LKB) column equilibrated with 1.7 Mammonium sulfate in 50 mM sodium phosphate buffer, pH7.0 [6] and eluted with 50 mM sodium phosphate into 0.5 ml-fractions. The proteolytic activity was measured (10 µl-aliquots) against Abz-MKALTLQ-EDDnp, as described pre-viously [5]. The fractions were analyzed in silver-stained SDS-PAGE gels [9,10].

2.2. Purification of yeast mannan

The procedure of Gorin and Spencer [11] described indetail by Travassos et al. [12] was used. Briefly, Saccharomy-ces cerevisiae cultures were autoclaved for 30 min. Superna-tants were precipitated with 3 volumes of ethanol after add-ing 1% of sodium acetate. The precipitate and the mass ofcells were suspended in 20% aqueous KOH and extracted for2 h at 100 °C in a water bath. The suspension was neutralizedwith glacial acetic acid and the supernatant was evaporatedto 100–150 ml, then precipitated with ethanol. The solubi-lized precipitate was precipitated with Fehling’s solution at4 °C. The copper complexes were treated with Amberlite1R-120 for 1 h at room temperature and the solution pooledwith distilled water washings of the beads were precipitatedwith 4 ml of ethanol with drops of HCl. The final mannanpreparation contained more than 98% carbohydrate with lessthan 1% protein, and no contamination with alkali-solubleglucan. Yeast mannan obtained by hot alkali extraction andFehling precipitation has a molecular size of 20–60 kDa, aver-aging 40 kDa [13]. Milder methods used to prepare manno-proteins provide complexes of 133 kDa. Molecular sizes aredetermined in filtration columns calibrated with dextran ofdifferent sizes. Our calculations to determine the kineticparameters used the average mannan size of 40 kDa.

2.3. Purification of the P. brasiliensis galactomannancomplex

Phenyl Superose fractions containing the GalMan com-plex previously described [14] were pooled and fractionatedby FPLC (ÄKTA purifier – Amersham Pharmacia Biotech)through a reverse-phase Sephasil Protein C4 (Amersham Bio-sciences) equilibrated with 0.1% trifluoracetic acid (TFA) andeluted with 30 ml of a 0–90% acetonitrile gradient in 0.1%TFA into 0.5-ml fractions. Fifty-one fractions were dried in aspeed-vacuum and resuspended in 100 µl of 50 mM Tris, pH8.0, for SDS-PAGE analysis. The GalMan complex was pre-dominantly eluted in fraction 14, as shown by the presence ofa broad and heavily silver-stained component migratingslowly in SDS-PAGE gels [14]. This procedure was repeatedseveral times, and the GalMan was pooled, lyophilized andfurther purified by FPLC (ÄKTA purifier) in a Superose 12(Amersham Biosciences) column equilibrated and eluted with50 mM Tris–HCl, 0.15 M NaCl, pH 7.0. Total carbohydratecontents of the peak fractions were quantified using thephenyl-sulfuric acid method described by Dubois et al. [15].

85A.L. Matsuo et al. / Microbes and Infection 8 (2006) 84–91

2.4. Proteolytic assays

The effect of polysaccharides on the exocellular serine-thiol proteinase activity of P. brasiliensis against Abz-MKALTLQ-EDDnp was determined in a thermostatic Hita-chi F-2000 spectrofluorometer (kex = 320 nm; kem 420 nm).Kinetic parameters were determined by measuring the initialrate of hydrolysis at various substrate concentrations in thepresence or absence of different polysaccharide concentra-tions. We tested the effect of GAGs (heparin, heparan sufateand chondroitin) and neutral polysaccharides (dextran, yeastmannan and P. brasiliensis GalMan complexes). The reac-tions were carried out at 37 °C in 10 mM Tris–HCl (pH 8.0).Progress of the reaction was continuously monitored by fluo-rescence record of the released product. Analysis of the dataobtained was performed by nonlinear regression using theGraFit 3.0 computer program (Erithacus Software Ltd.).

The modulatory effect of dextran and GalMan complexeson the proteolytic activity of the enzyme was calculated fol-lowing Eq. (1), where S is Abz-MKALTLQ-EDDnp, NP isthe polysaccharide, KS is the substrate dissociation constant,KN is the apparent polysaccharide dissociation constant, � isthe parameter of KS perturbation, and b is the parameter ofVMAX (kcat) perturbation.

(Eq. 1)v =Vmax × [S]

Ks�1 +

[NP]

Kn ��1 +

b × [NP]

� × Kn �+ [S]

�1 +[NP]

� × Kn ��1 +

b × [NP]

� × Kn �Since the interaction between mannan and PbST resulted

in a powerful inhibition of the enzyme activity, the modula-tory effect of yeast mannan on the proteolytic activity of theenzyme was analyzed by nonlinear regression according toEq. (2) and by linear regression according to Eq. (4), whereK′ is the observed substrate dissociation constant in the pres-ence of mannan and KMan is the true mannan-protease disso-ciation constant; M is mannan and � is the cooperative param-eter for binding of the second mannan to enzyme.

(Eq. 2)v =Vmax · [S]

KS�1 +2 · [M ]

KMan

+[M ]2

� · KMan� + [S]

(Eq. 3)K′ = KS ·�1 +2 · [M]

KMan

+[M]2

� · KMan�

(Eq. 4)1

K′=

1

� · KMan2 · [M ] +

2

[M ]

2.5. Effect of GalMan on the temperature-inducedinactivation of the PbST activity

The kinetics of thermal inactivation of the PbST activityover Abz-MKALTLQ-EDDnp was evaluated in the absence

or in the presence of 300 nM of GalMan at 50 °C in 10 mMTris–HCl buffer (pH 8.0). The reactions were carried out inthe presence of 0.24 µM of Abz-MKALTLQ-EDDnp, whichis 10-fold below the KS value. The progress of the reactionwas continuously monitored by fluorescence measurementof the released product. The obtained exponential decaycurves were best fitted to the first-order relationship shown inEq. (5),

(Eq. 5)P = P∞� 1 − e-kobs.t �

where P and P∞ are the product concentrations at a giventime and at infinite time, respectively; kobs is the observedfirst-order rate of temperature-induced enzyme inactivation.

2.6. Protein and carbohydrate contents

Protein contents were estimated by a modification of themethod described by Bradford [16], using bovine serum albu-min as standard. Total carbohydrates were quantified by themethod of Dubois et al. [15], using mannan as standard.

3. Results

We studied the influence of polysaccharides upon theendopeptidase activity of a fungal serine-thiol proteinase usingthe Abz-MKALTLQ-EDDnp substrate. When culture super-natants of P. brasiliensis B-339 were fractionated in a PhenylSuperose column, the peptidase activity capable of hydrolyz-ing Abz-MKALTLQ-EDDnp was predominantly eluted infraction 3. This fraction showed few protein/glycoproteinsilver-stained bands, as seen in the profile of Fig. 1, which

Fig. 1. Serine-thiol proteinase preparation. Silver-stained 10% SDS-PAGEgel showing the fractionation of a P. brasiliensis culture supernatant (sup, in40% ammonium sulfate) through a Phenyl Superose column eluted with50 mM sodium phosphate, pH 7.0. The hydrolytic activity of each fraction(3–6) was measured in a 10 µl-aliquot against Abz-MKALTLQ-EDDnp andthe results are shown as arbitrary units of fluorescence (AUF). Fraction 3 ofthis experiment, which contained the activity peak, was used as enzymesource (PbST). The position of molecular mass markers (in kDa) is shownon the left.

86 A.L. Matsuo et al. / Microbes and Infection 8 (2006) 84–91

also shows that most of the supernatant components wereeluted in fractions 4 and 5. Therefore, we used fraction 3 asenzyme source (PbST) in the experiments shown below. Enzy-matic cleavage of Abz-MKALTLQ-EDDnp was due to theexocellular serine-thiol proteinase activity of P. brasiliensiscontained in fraction 3, since the hydrolysis was specificallyinhibited by p-HMB and PMSF, as previously reported [5]. Itis of note that total purification of the active proteinase hasnever been achieved due to aggregation to other fungal com-ponents and/or loss of enzymatic activity after several chro-matographic steps.

Negatively charged GAGs, namely heparin, chondroitinand heparan sulfate, did not affect the rate of Abz-MKALTLQ-EDDnp hydrolysis when added at a concentra-tion range of 0–100 µM (not shown). In addition, PbST didnot bind to heparin-Sepharose (not shown), suggesting thatthe enzyme does not interact with GAGs.

Surprisingly, in the presence of dextran (Sigma, averagemolecular mass of 100–200 kDa, where Mr = 100 kDa wasused in the calculations), a neutral polysaccharide initiallyused as control, there was a significant alteration in the kineticparameters of PbST (Fig. 2). The VMAX value for the hydroly-sis of Abz-MKALTLQ-EDDnp decreased significantly as afunction of dextran concentration (Fig. 2A). Dextran alsocaused a marked increase in the enzyme affinity for this sub-strate, evidenced by a sharp decrease in the KS value, which

dropped from 2.4 ± 0.2 to 0.35 ± 0.04 µM at the highest dex-tran concentration of 20 µM (Fig. 2B). The effect of dextranon PbST can be described as a hyperbolic mixed type inhibi-tion according to Eq. (1) (Section 2), where the efficiency ofsubstrate hydrolysis was estimated by changing either the KM

(parameter �) or the VMAX (parameter b). These data werefitted to the equation using nonlinear regression, and the val-ues for the constants were determined. The results revealedthat dextran bound to free PbST (E) with a dissociation con-stant of KDX = 2.7 ± 0.3 µM, whereas the parameter obtainedfor binding to the enzyme–substrate complex (ES) was�KDX = 0.35 ± 0.04 µM (parameter � = 0.13 ± 0.01). Thedata also showed that the interaction between dextran and theserine-thiol proteinase resulted in a 3.6-fold decrease in theenzyme VMAX (b = 0.28 ± 0.03) and in an almost 2.2-foldincrease in its catalytic activity (b/� = 2.2).

We then tested the effect of purified baker’s yeast man-nan, since mannan structures are commonly found in fungalglycoconjugates. Interestingly, the interaction between man-nan and PbST resulted in a powerful inhibition of the enzymeactivity. The effect of mannan upon PbST can be describedas a parabolic competitive type inhibition according to Eqs.(2)–(4) of Section 2. The data show that two molecules ofmannan bound to the enzyme and that the first inhibitorchanged the intrinsic dissociation constant of the vacant oneby factor �. The influence of mannan upon the substrate dis-sociation constant can be observed in Fig. 3A. Basically, themannan–proteinase interaction did not affect the catalytic con-stant of Abz-MKALTLQ-EDDnp; the presence of mannanexcluded substrate binding at the active site, suggesting thatthe enzyme is capable of binding two molecules of mannancompetitive inhibitor in a parabolic manner. It was observedthat binding of the first mannan induced structural changes inthe enzyme that facilitated binding of the second mannaninhibitor (Fig. 3B). These data were fitted to Eq. (2) usingnonlinear regression, and the values for the constants weredetermined. The results show that the first molecule of man-nan bound to free PbST (E) with a dissociation constant ofKMan = 11.4 ± 0.9 µM. The second mannan molecule boundto the mannan–proteinase complex (EI) with a dissociationconstant of �KMan = 2.1 ± 0.2 µM (� parameter =0.18 ± 0.02). Therefore, binding of two molecules of man-nan to the enzyme was not random and independent, butoccurred in a cooperative manner: binding of the first mol-ecule of mannan (KMan = 11.4 ± 0.9 µM) resulted in a 5.5-fold increase in the affinity of the enzyme for ligation of thesecond mannan (�KMan = 2.1 ± 0.2 µM).

It is noteworthy that high performance liquid chromatog-raphy (HPLC) and mass spectrometry analyses previouslyshowed that Leu–Thr is the only peptide bond cleaved inAbz-MKALTLQ-EDDnp by PbST [5]. We presently observed byHPLC analysis that the presence of either mannan or dextrandid not change the cleavage pattern of this peptide by theproteinase activity (not shown).

The modulatory effect of dextran and mannan upon theserine-thiol proteinase activity of P. brasiliensis prompted us

Fig. 2. Modulatory effect of dextran on the Abz-MKALTLQ-EDDnp hydro-lysis by PbST. The apparent VMAX (VMAX app), recorded as AUF per min(A), and the apparent affinity constants (Ks app) (B) were calculated for thereactions carried out in the presence of increasing amounts of dextran.

87A.L. Matsuo et al. / Microbes and Infection 8 (2006) 84–91

to investigate the effect of an endogenous fungal galactoman-nan (GalMan) complex originally described by Puccia et al.[14]. This component is commonly found in the supernatantfluids of P. brasiliensis cultures and migrates slowly in SDS-

PAGE gels, forming a diffuse smear due to its high degree ofglycosylation (see fractions 4 and 5, Fig. 1). In order to purifythis component, we pooled Phenyl Superose fractions 4 and5 and fractionated them in a Sephasil C4 reversed-phase col-umn.After this kind of chromatography (not shown), the larg-est peak (fraction 14) concentrated a slow-migrating, broadcomponent corresponding to the GalMan complex (see pro-file in “F14”, Fig. 4B). We then submitted a pool of fractions14 from independent chromatographic runs to gel filtrationin a Superose 12 column (Fig. 4A) in order to isolate GalManfrom a minor 50-kDa component (Fig. 4B). The GalMan com-plex was mostly excluded in the void volume, which corre-sponded to fractions 8–11, as verified by calibration of thecolumn with dextran blue (Sigma) and dextran T70 (Pharma-cia). GalMan was eluted from Superose 12 in fraction 10,which also contained the total carbohydrate peak(0.375 mg ml–1). This fraction was therefore used in the sub-sequent experiments, at the estimated average molecular massof 100 kDa, which was the value used to calculate the kineticparameters.

Fig. 5A shows that the P. brasiliensis GalMan complexcaused a marked increase in the PbST affinity for Abz-MKALTLQ-EDDnp, which was evidenced by the decreasein the Ks value. The presence of GalMan caused a threefolddecrease in the enzyme KM (� = 0.33 ± 0.03), i.e. from2.4 ± 0.2 to 0.78 ± 0.08 µM. On the other hand, the interac-tion of GalMan with PbST resulted in a 3.1-fold decrease inthe enzyme VMAX (b = 0.32 ± 0.03), as seen in Fig. 5B. Thesedata were fitted to Eq. (1) (Section 2), and the values for theconstants were determined using nonlinear regression. Theresults showed that GalMan bound to free PbST (E) with ahigh affinity dissociation constant of KGM = 94 ± 7 nM, whilebinding to the enzyme–substrate complex (ES) resulted in an�KGM = 30 ± 2 nM (� = 0.33 ± 0.03).

The effect of GalMan on the optimum pH of PbST wasanalyzed by monitoring the enzyme-catalyzed hydrolysis ofthe fluorogenic substrateAbz-MKALTLQ-EDDnp.As shown

Fig. 3. Inhibitory effect of baker’s yeast mannan on the Abz-MKALTLQ-EDDnp hydrolysis by PbST. The apparent affinity constants (Ks app) (A)and 1/K′ were calculated for the reactions carried out in the presence ofincreasing amounts of dextran.

Fig. 4. Purification of P. brasiliensis GalMan. A, Elution profile (A215) of a gel filtration chromatography in Superose 12 (FPLC) of a pool of fractions 14 (F14,4B) recovered from a Sephasil Protein C4 reverse-phase column (details in the text). B, Silver-stained 10% SDS-PAGE profile of the peak fractions seen in A(8–12). Total carbohydrate (Tcarb) corresponding to each fraction is indicated in mg ml–1. The position of molecular mass markers (in kDa) is shown on the left.Fraction 10 was used in the subsequent experiments.

88 A.L. Matsuo et al. / Microbes and Infection 8 (2006) 84–91

in Fig. 6, when the enzyme was assayed in the presence of300 nM GalMan, no significant effect on the enzyme pHresponse was observed. On the other hand, the presence ofGalMan caused a threefold decrease in the first-order rate ofthe serine-thiol proteinase inactivation at 50 °C, since the kobs

value dropped from 11.1 ± 0.6 to 3.6 ± 0.2 ms−1 (Fig. 7).

4. Discussion

The results of the present work show the modulation byneutral polysaccharides of a fungal endopeptidase activity invitro. Binding of neutral polysaccharides to the extracellularserine-thiol proteinase (PbST) secreted by the human patho-gen P. brasiliensis [5] changed its kinetic parameters ofhydrolysis of the fluorogenic substrate Abz-MKALTLQ-EDDnp. We found two different effects: commercial dextranand an endogenous galactomannan (GalMan) complex mark-edly increased the enzyme affinity for the substrate, whilebaker’s yeast mannan inhibited the PbST activity. The Gal-Man component did not affect the optimum pH of the enzyme,whereas the presence of mannan or dextran did not changethe enzyme cleavage point of the substrate Abz-MKALTLQ-EDDnp (L/T), suggesting that the interaction with thesepolysaccharides does not alter the enzyme specificity. Anyeffect on the enzyme–carbohydrate interaction caused by con-taminants present in the enzyme preparation would be irrel-evant due to the range of carbohydrate concentrations used inthe reactions.

Previous unpublished results from our group suggested apossible protective role of the GalMan complex on the PbSTactivity. A concanavalin A-bound preparation of the enzyme,which also co-purified GalMan and a 55-kDa glycoprotein,seemed to be less labile than the active fractions obtainedafter anion-exchange and gel filtration chromatography [5].The 55-kDa glycoprotein was later sequenced and found tobe homologous to catalases (unpublished results). We there-fore empirically associated stability of PbST with either Gal-Man or catalase, or with both.

Our present data show that the P. brasiliensis GalMan com-ponent indeed protects the peptidase activity of PbST at hightemperatures. At 37 °C, GalMan modulated the enzyme atthe nanomolar range (KGM = 94 nM and aKGM = 30 nM), pre-serving its catalytic activity (b/� = 1.0) by increasing affinity(threefold) for the substrate, but decreasing the catalysis veloc-ity (3.1-fold). Dextran affected the PbST activity similarly,however at the µM concentration range (KDX = 2.7 µM) andwith an increase in the catalytic activity (b/� = 2.2).

The GalMan component was previously characterizedamong other concanavalin A-eluted components of the super-

Fig. 5. Effect of P. brasiliensis GalMan complex on the Abz-MKALTLQ-EDDnp hydrolysis by PbST. The apparent VMAX (VMAX app), recorded asAUF per min (A), and the apparent affinity constants (Ks app) (B) were cal-culated for the reactions carried out in the presence of increasing amounts ofdextran.

Fig. 6. Lack of effect of P. brasiliensis GalMan preparation (300 nM) on theAbz-MKALTLQ-EDDnp hydrolysis by PbST at different pH values.

Fig. 7. Protective effect of P. brasiliensis GalMan (300 nM) on the Abz-MKALTLQ-EDDnp hydrolysis by PbST at 50 °C.

89A.L. Matsuo et al. / Microbes and Infection 8 (2006) 84–91

natant fluids of P. brasiliensis yeast phase cultures [14]. Fur-ther analysis using a chemically defined medium to grow theyeast phase rendered a carbohydrate rich (88%) exocellularcomponent containing two types of residues: a-Manp andb-Galf (unpublished results). Nuclear magnetic resonance(NMR) studies showed that the polysaccharide has an a-1,6-linked mannopyranan main chain and a-1,2/a-1,3-linkedManp side chains with terminal nonreducing units of a-Manpand b-Galf (50% of the side chains), (1,3), (1,6), and (1,4)-linked to mannopyranosyl units [17]. Structural differencesin the P. brasiliensis GalMan and baker’s yeast mannan mightexplain the opposite, inhibitory effect on enzyme activitycaused by the latter. Baker’s yeast mannan is composed solelyof mannopyranosyl units both in the backbone and in the sidechains, which contain up to four residues linked either a-1,2 ora-1,3. The size of the chains can vary with the glycoprotein:high mannose short chains bear 8–14 mannose residues, whilelonger chains will carry 20–200 residues [18]. The mannanpreparation used in the present work was extracted from wholecells of S. cerevisisae and was therefore rather heteroge-neous.

We have previously observed that the serine-thiol protein-ase activity secreted by the human pathogen P. brasiliensis isable to cleave tumor and cartilage decorin, aggrecan and ver-sican. In the present work we observed that heparin, heparansulfate and chondroitin do not interact with this proteinase.On the other hand, there is accumulating data on the interfer-ence of highly sulfated GAGs, especially heparin and heparin-like, in the activity of various classes of proteinases (reviewedin [8]). They include papain [19], cathepsin B [20], cathepsinG [21], neutrophil elastase [22], chymase [23], collagenolyticactivity of cathepsins [24,25] and tryptase [26]. The mecha-nisms involved in modulation of the protease activities byGAGs have been attributed to their ability to a) change pro-tein conformation upon binding, which leads to a change inthe catalytic activity or to a decreased rate of inhibition bynatural inhibitors, b) protect against pH inactivation or c)expose binding regions in the target proteins (reviewed in [8]).On the other hand, electrostatic interactions are probablyessential but not sufficient to govern binding specificity ofGAGs with proteinases and its consequences on their activ-ity: the glycosidic linkage, type of saccharide and molecularweight also determine the fine specificities of the interaction[25,27].

The protection of proteins against dehydration and dena-turation has been achieved in several instances through theinteraction with carbohydrates. A notorious example is thatof neutral disaccharides, mainly sucrose and trehalose. Inaqueous solution the major mechanisms may involve water-trehalose H-bond replacement, protein coating by a trappedwater layer and inhibition of conformational fluctuations [28].A study on the trypsin interaction with trehalose showed thatprotein preservation was directly related to the number ofhydrogen bonds formed [29]. Several sugars are able to pre-vent protein unfolding but its functional inactivation dependson the nature of the sugar [30]. Dextran, for instance, fails to

inhibit dehydration-induced lysozime unfolding because itdoes not H-bond adequately to the protein [31]. Seemingly,the interaction of carbohydrates with enzymes can protect theintegrity of protein structure even in adverse conditions ofheat shock, dehydration and presence of chaotropic agents,but also it may modulate their activity by restricting confor-mational states directly related with the catalytic reaction.

In the present work we have examined the modulation byneutral polysaccharides of a fungal protease. At first hand,we would expect H-bonding as the generalized interactionbetween carbohydrate and protein to be quantitatively differ-ent while using dextran, yeast mannan and galactomannan.They could either stimulate or inhibit protease activity, or evenhave no effect. In principle, specific recognition of sugar unitsby a lectin type interaction was not foreseen for the serine-thiol protease. In the case of mannan, there are several mam-malian lectins and a mannan binding protein (MBL), whichare involved in the host innate defense mechanisms (reviewedin [32,33]). The specificity of carbohydrate recognition ofMBL relies on the identity of the monosaccharide (man-nose), and on the fine structure and length (pattern) of theoligosaccharide side chains.

The serine-thiol proteinase interacts with polysaccharidesalthough there is no evidence for a lectin-like recognition.They could form H-bonding with the enzyme varying quan-titatively and even involving different sites according to theirstructure and hydration. In the case of the galactomannan, wecould speculate that the mechanism of secretion of the serine-thiol proteinase in P. brasiliensis and its cellular traffickingmay depend on the fungal polysaccharide for presentation atthe cell surface and release into the medium. Before and aftersecretion, the PbST and GalMan could aggregate for a morestable conformation allowing the enzyme to increase its affin-ity for the substrate. A threefold decrease in VMAX would beimportant to prevent rapid autolysis until the protease foundthe specific substrate. In consequence, the enzyme coulddegrade components of the membrane more successfully,leading to a more efficient fungal dissemination.

Acknowledgments

Research and fellowships were financed by FAPESP,PRONEX/CNPq and CNPq. The peptide substrate Abz-MKALTLQ-EDDnp used in this work was produced at theDepartment of Biophysics at UNIFESP and kindly providedby Dr. Maria Aparecida Juliano.

References

[1] M. Franco, M.T. Peracoli, A. Soares, R. Montenegro, R.P. Mendes,D.A. Meira, Host–parasite relationship in paracoccidioidomycosis,Curr. Top. Med. Mycol. 5 (1993) 115–149.

[2] A. Restrepo, J.G. McEwen, E. Castañeda, The habitat of Paracoccid-ioides brasiliensis: how far from solving the riddle? Med. Mycol. 39(2001) 233–241.

90 A.L. Matsuo et al. / Microbes and Infection 8 (2006) 84–91

[3] J.G. McEwen, V. Bedoya, M.M. Patino, M.E. Salazar, A. Restrepo,Experimental murine paracoccidiodomycosis induced by the inhala-tion of conidia, J. Med. Vet. Mycol. 25 (1987) 165–175.

[4] G. San Blas, G. Nino-Vega, T. Iturriaga, Paracoccidioides brasilien-sis and paracoccidioidomycosis: molecular approaches to morpho-genesis, diagnosis, epidemiology, taxonomy and genetics, Med.Mycol. 40 (2002) 225–242.

[5] A.K. Carmona, R. Puccia, M.C. Oliveira, E.G. Rodrigues, L. Juliano,L.R. Travassos, Characterization of an exocellular serine-thiol pro-teinase activity in Paracoccidioides brasiliensis, Biochem. J. 309(1995) 209–214.

[6] R. Puccia, M.A. Juliano, L. Juliano, L.R. Travassos, A.K. Carmona,Detection of the basement membrane-degrading proteolytic activityof Paracoccidioides brasiliensis after SDS-PAGE using agarose over-lays containing Abz-MKALTLQ-EDDnp, Braz. J. Med. Biol. Res. 32(1999) 645–649.

[7] R. Puccia, A.K. Carmona, J.L. Gesztesi, L. Juliano, L.R. Travassos,Exocellular proteolytic activity of Paracoccidioides brasiliensis:cleavage of components associated with the basement membrane,Med. Mycol. 36 (1998) 345–348.

[8] I.L. Tersariol, D.C. Pimenta, J.R. Chagas, P.C. Almeida, Proteinaseactivity regulation by glycosaminoglycans, Braz. J. Med. Biol. Res.35 (2002) 135–144.

[9] U.K. Laemmli, Cleavage of structural proteins during the assembly ofthe head of bacteriophage T4, Nature 227 (1970) 680–685.

[10] W. Ansorge, Fast and sensitive detection of protein and DNA bands bytreatment with potassium permanganate, J. Biochem. Biophys. Meth-ods 11 (1985) 13–20.

[11] P.A. Gorin, J.F.T. Spencer, Proton magnetic resonance spectroscopy.An aid in identification and chemo-taxonomy of yeasts, Adv. Appl.Microb. 13 (1970) 25–89.

[12] L.R. Travassos, P.A. Gorin, K.O. Lloyd, Comparison of the rham-nomannans from the human pathogen Sporothrix schenckii with thosefrom the Ceratocystis species, Infect. Immun. 8 (1973) 685–693.

[13] T. Nakajima, C.E. Ballou, Characterization of the carbohydrate frag-ments obtained from Saccharomyces cerevisiae mannan by alkalinedegradation, J. Biol. Chem. 249 (1974) 7679–7684.

[14] R. Puccia, S. Schenkman, P.A. Gorin, L.R. Travassos, Exocellularcomponents of Paracoccidioides brasiliensis: identification of a spe-cific antigen, Infect. Immun. 53 (1986) 199–206.

[15] M. Dubois, K.A. Gilles, J.K. Hamilton, P.A. Rebers, F. Smith, Colo-rimetric method for determination of sugars and related substances,Anal. Chem. 28 (1956) 131–137.

[16] M.M. Bradford, A rapid and sensitive method for the quantitation ofmicrogram quantities of protein utilizing the principle of protein-dyebinding, Anal. Biochem. 72 (1976) 254.

[17] L.R. Travassos, C.P. Taborda, L.K. Iwai, E. Cunha Neto, R. Puccia, in:J.E. Domer, G.S. Kobayashi (Eds.), Human Fungal Pathogens,Springer-Verlag Berlin, Heidelberg, 2004, pp. 279–296.

[18] M.A. Kukuruzinska, M.L. Bergh, B.J. Jackson, Protein glycosylationin yeast, Annu. Rev. Biochem. 56 (1987) 915–944.

[19] P.C. Almeida, I.L. Nantes, J.R. Chagas, C.C. Rizzi, A. Faljoni-Alario,E. Carmona, L. Juliano, H.B. Nader, I.L. Tersariol, Cysteine protein-ase activity regulation. A possible role of heparin and heparin-likeglycosaminoglycans, Biol. Chem. 276 (2001) 944–951.

[20] J. Ermolieff, J. Duranton, M. Petitou, J.G. Bieth, Cathepsin B activityregulation. Heparin-like glycosaminogylcans protect human cathep-sin B from alkaline pH-induced inactivation, Biochem. J. 330 (1998)1369–1374.

[21] B. Faller, M. Cadene, J.G. Bieth, Heparin accelerates the inhibition ofcathepsin G by mucus proteinase inhibitor: potent effect ofO-butyrylated heparin, Biochemistry 32 (1993) 9230–9235.

[22] L. Lindstedt, M. Lee, P.T. Kovanen, Demonstration of a two-stepreaction mechanism for the inhibition of heparin-bound neutrophilelastase by alpha 1-proteinase inhibitor, Atherosclerosis 155 (2001)87–97.

[23] Z. Li, W.S. Hou, D. Bromme, Chymase bound to heparin is resistant toits natural inhibitors and capable of proteolyzing high density lipopro-teins in aortic intimal fluid, Biochemistry 39 (2000) 529–536.

[24] Z. Li,Y.Yasuda, W. Li, M. Bogyo, N. Katz, R.E. Gordon, G.B. Fields,D. Bromme, Collagenolytic activity of cathepsin K is specificallymodulated by cartilage-resident chondroitin sulfates, J. Biol. Chem.279 (2004) 5470–5479.

[25] J. Hallgren, S. Lindahl, G. Pejler, Regulation of collagenase activitiesof human cathepsins by glycosaminoglycans, J. Mol. Biol. 345 (2005)129–139.

[26] D. Ledoux, D. Merciris, D. Barritault, J.P. Caruelle, Structuralrequirements and mechanism for heparin-dependent activation andtetramerization of human betaI- and betaII-tryptase, FEBS Lett. 537(2003) 23–29.

[27] D. Ledoux, D. Merciris, D. Barritault, J.P. Caruelle, Heparin-likedextran derivatives as well as glycosaminoglycans inhibit the enzy-matic activity of human cathepsin G, FEBS Lett. 537 (2003) 23–29.

[28] R.D. Lins, C.S. Pereira, P.H. Hunenberger, Trehalose–protein interac-tion in aqueous solution, Proteins 55 (2004) 177–186.

[29] E.C. Lopez-Diez, S. Bone, The interaction of trypsin with trehalose:an investigation of protein preservation mechanisms, Biochim. Bio-phys. Acta 1673 (2004) 139–148.

[30] M. Sola-Penna, A. Ferreira-Pereira, A.P. Lemos, J.R. Meyer-Fernandes, Carbohydrate protection of enzyme structure and functionagainst guanidinium chloride treatment depends on the nature ofcarbohydrate and enzyme, Eur. J. Biochem. 248 (1997) 24–29.

[31] S.D. Allison, B. Chang, T.W. Randolph, J.F. Carpenter, Hydrogenbonding between sugar and protein is responsible for inhibition ofdehydration-induced protein unfolding, Arch. Biochem. Biophys. 365(1999) 289–298.

[32] J.S. Presanis, M. Kojima, R.B. Sim, Biochemistry and genetics ofmannan-binding lectin (MBL), Biochem. Soc. Trans. 31 (2003) 748–752.

[33] J.K. van de Wetering, L.M. van Golde, J.J. Batenburg, Collectins:players of the innate immune system, Eur. J. Biochem. 271 (2004)1229–1249.

91A.L. Matsuo et al. / Microbes and Infection 8 (2006) 84–91

34

Artigo 2: C-Npys (S-3-nitro-2-pyridinesulfenyl) and peptide derivatives can inhibit a

serine-thiol proteinase activity from Paracoccidioides brasiliensis.

Alisson L. Matsuo, Adriana K. Carmona, Luiz S. Silva, Carlos E. Cunha, Ernesto S.

Nakayasu, Igor C. Almeida, Maria A. Juliano e Rosana Puccia

Biochemical and Biophysical Research Communications 2007, Vol 355 (4) 1000-1005

35

Racional

Diversas proteinases extracelulares de microorganismos patogênicos estão

relacionadas com invasão e disseminação pela degradação de componentes da matriz

extracelular. A importância no estudo da proteinase PbST de P. brasiliensis deve-se

principalmente à sua seletividade para proteínas associadas à matriz extracelular.

Determinar inibidores específicos constituiu-se em uma ferramenta importante para o

estudo de proteinases e também para uso terapêutico. Nesse sentido, a inibição da PbST

por compostos mercuriais levou-nos a testar o poder inibitório do composto C(Npys)

acoplado com o peptídeo MKRLTL, que é substrato da enzima. C(Npys) foi inicialmente

usado como protetor na síntese de peptídeos e posteriormente como inibidor de cisteíno

proteinases, pois tem a capacidade de reagir com grupos tiois livres. O trabalho que se

segue mostra que notadamente o composto Bzl-C(Npys)KRLTL-NH2 apresentou uma

elevada capacidade inibitória da PbST, a qual foi completamente revertida com

ditiotreitol. O inibidor não foi tóxico em cultura de células, oferecendo uma alternativa de

inibição in vivo, notadamente se a especificidade para PbST for comprovada.

www.elsevier.com/locate/ybbrc

Biochemical and Biophysical Research Communications 355 (2007) 1000–1005

C-Npys (S-3-nitro-2-pyridinesulfenyl) and peptide derivatives can inhibita serine-thiol proteinase activity from Paracoccidioides brasiliensis

Alisson L. Matsuo a, Adriana K. Carmona b, Luiz S. Silva a, Carlos E.L. Cunha b,Ernesto S. Nakayasu c, Igor C. Almeida c, Maria A. Juliano b, Rosana Puccia a,*

a Department of Microbiology, Immunology and Parasitology, Cell Biology Division, Federal University of Sao Paulo, Sao Paulo, SP 04023-062, Brazilb Department of Biophysics, Federal University of Sao Paulo, Sao Paulo, SP 04044-062, Brazilc Department of Biological Sciences, University of Texas at El Paso, El Paso, TX 79968, USA

Received 30 January 2007Available online 23 February 2007

Abstract

The inhibitory capacity of C-Npys (S-[3-nitro-2-pyridinesulfenyl]) derivatives over thiol-containing serine proteases has never beentested. In the present work we used an extracellular serine-thiol proteinase activity from the fungal pathogen Paracoccidioides brasiliensis

(PbST) to describe a potent inhibitory capacity of Bzl-C(Npys)KRLTL-NH2 and Bzl-MKRLTLC(Npys)-NH2. The assays were per-formed with PbST enriched upon affinity chromatography in a p-aminobenzamidine (pABA)-Sepharose column. Although PbST cancleave the fluorescence resonance energy transfer peptide Abz-MKRLTL-EDDnp between L–T, the C(Npys) derivatives were not sub-strates nor were they toxic in a cell detachment assay, allowing therapeutic use. The best inhibitor was Bzl-C(Npys)KRLTL-NH2

(Ki = 16 nM), suggesting that the peptide sequence promoted a favorable interaction, especially when C(Npys) was placed at a furtherposition from the L–T bond, at the N-terminus. Inhibition was completely reverted with dithioerythritol, indicating that it was due to thereactivity of the C(Npys) moiety with a free SH– group.� 2007 Elsevier Inc. All rights reserved.

Keywords: Paracoccidioides brasiliensis; Serine-thiol proteinase; Npys

Extracellular proteinases can facilitate the process ofinvasion and dissemination of microorganisms throughthe human tissues by degrading extracellular-associatedproteins [1]. Our group has previously characterized an

0006-291X/$ - see front matter � 2007 Elsevier Inc. All rights reserved.

doi:10.1016/j.bbrc.2007.02.070

Abbreviations: Abz, ortho-aminobenzoic acid; AUF, arbitrary units offluorescence; Bzl, benzoyl; Boc, tert-butyloxycarbonyl; C(Npys), S-(3-nitro-2-pyridinesulfenyl); E-64, (L-trans-expoxysuccinyl-leucylamido[4-guanidino butane]); EDDnp, N-(2,4-dinitrophenyl)ethylenediamine;FRET, fluorescence resonance energy transfer; pABA, p-aminobenzami-dine; PbST, exocellular serine-thiol activity from Paracoccidioides

brasiliensis; PCM, paracoccidioidomycosis; p-HMB, sodium 7-hydroxy-mercuribenzoate; PMSF, phenylmethylsulfonyl fluoride.

* Corresponding author. Present address: Disciplina de Biologia Celular,UNIFESP, Rua Botucatu, 862, oitavo andar, Sao Paulo, SP 04023-062,Brazil. Fax: +55 11 5571 5877.

E-mail address: rosana@ecb.epm.br (R. Puccia).

extracellular subtilisin-like serine proteinase activity inthe yeast phase of the human pathogen Paracoccidioides

brasiliensis [2]. This activity is able to selectively degrade,in vitro, murine laminin, human fibronectin, type IV-colla-gen, and proteoglycans, while common proteins likebovine serum albumin (BSA) or casein are not cleavablesubstrates [3]. The proteinase is able to hydrolyze fluores-cence resonance energy (FRET) peptides homologous toMKRLTL, which are flanked by Abz (ortho-aminobenzoicacid) and EDDnp (N-[2,4-dinitrophenyl]ethylenediamine).Cleavage occurs between the Leu–Thr bond at an opti-mum alkaline pH and is inhibited by PMSF, mercuric ace-tate, and p-HMB (sodium 7-hydroxymercuribenzoate), butnot by E-64 [2]. Therefore, the enzymatic activity has beenclassified as a thiol-containing subtilisin-like serine protein-ase (PbST, for P. brasiliensis serine-thiol proteinase)

A.L. Matsuo et al. / Biochemical and Biophysical Research Communications 355 (2007) 1000–1005 1001

belonging to the group of proteinase K. In addition, wehave recently described a novel type of regulation ofthe PbST activity against Abz-MKALTLQ-EDDnp byneutral polysaccharides, including a fungal extracellu-lar galactomannan, which might help stabilize the enzyme[4].

Paracoccidioides brasiliensis is a dimorphic fungus thatcauses paracoccidioidomycosis (PCM), a potentially lethalsystemic mycosis with multiple clinical forms that is preva-lent in South America [5]. The fungus grows in the infec-tive mycelial phase at environmental temperatures below26 �C and in the yeast pathogenic phase at body tempera-tures of about 37 �C. PCM is a granulomatous mycosisthat starts in the lungs and tends to disseminate rapidlythrough the lymphatic system in juvenile (acute and suba-cute) forms. In adult (chronic) forms, active disease affectsespecially the lungs, however dissemination to the cutane-ous, mucocutaneous tissues or other organs is likely tooccur through the blood and lymphatic route. In thiscontext, PbST is a potential virulence factor that couldcontribute to fungal spread to surrounding and/or distanttissues.

The S-(3-nitro-2-pyridinesulfenyl) group—C(Npys)—has originally been used for protection in peptide synthesis[6] and later demonstrated to be effective in the design andsynthesis of thiol proteinase irreversible inhibitors [7].These compounds have been found to be highly specificto cysteine proteinases due to their capacity of selectivelyreacting with free thiol groups to form unsymmetricaldisulfide bonds [8] and their inability to inhibit serine oracid proteinases [7]. A variety of cysteine proteinase inhib-itors have been developed such as diazomethyl ketones,halomethyl ketones, and epoxides [9]. However, they arestrong electrophilic reagents and may alkylate non-targetedenzymes and other biomolecules in vivo [10], preventing anytherapeutic use. The C(Npys) compounds represent analternative to overcome this problem, considering that theypromote the formation of disulfide bonds without non-spe-cific alkylation, but still keep the ability of being specificinhibitors.

The inhibitory capacity of C(Npys) derivatives overthiol-containing subtilisins has never been tested. In thepresent work, we used PbST to describe a potent inhibitorycapacity of two peptides derived from MKRLTL coupledwith the C(Npys) group.

Materials and methods

Isolation of the serino-thiol proteinase activity. Culture supernatantscontaining PbST activity against Abz-MKALTLQ-EDDnp wereobtained from P. brasiliensis B-339 as previously described [3,4] andbuffered to a final concentration of 50 mM Tris–HCl, pH 8.0, containing1.0 M NaCl, which did not alter the initial proteolytic activity. Aliquots(40 ml) were chromatographed, using a peristaltic pump, in an affinitycolumn of p-aminomethylbenzamidine (pABA)-Sepharose (Pharmacia/LKB) previously equilibrated in the same buffer. After a washing stepwith 10 volumes of the same buffer, bound compounds were eluted with50 mM glycine, pH 3.0, into 1.0 mL fractions immediately neutralized

with Tris–HCl. The fractions were analyzed for proteolytic activityagainst Abz-MKALTLQ-EDDnp, as described below, and againstfibronectin (FN, 2 lg) as described by Puccia et al. [3]. Activity wasassayed in 50 mM Tris–HCl, pH 8.0. Fibronectin was purified from seradonated from healthy individuals by chromatography in gelatin-Sepharose (Amersham Biosciences, GE Healthcare), as described [11].Protein contents were visualized in Coomassie blue and/or silver-stainedSDS–PAGE gels [12].

Peptide synthesis. The FRET substrate Abz-MKALTLQ-EDDnp wassynthesized using the solid phase synthesis method [13] according to theFmoc procedure. An automated bench-top simultaneous multiple solid-phase peptide synthesizer (PSSM 8 system; Shimadzu) was used in thesynthesis. The inhibitor peptides Bzl-MKRLTLC(Npys)-NH2 and Bzl-C(Npys)KRLTL-NH2 were prepared by the solid phase methodologyusing Boc-amino acids according to the general procedure describedpreviously [14]. All the peptides were purified by semi-preparative h.p.l.c.using an Econosil C-18 column. The molecular mass and purity werechecked by amino acid analysis and by mass spectrometry using a matrixassisted laser-desorption ionization-time-of-flight-mass spectrometer(MALDI-TOF-MS) TOFSpec E instrument (Micromass, Manchester,Waters, UK).

Enzymatic assays. The proteolytic activity of PbST against Abz-MKALTLQ-EDDnp was determined at 37 �C in 50 mM Tris–HCl, pH 8.0,in a Hitachi F-2000 spectrofluorometer (excitation = 320 nm; emis-sion = 420 nm), as described previously [2]. The effect of Boc-C(Npys), Bzl-MKRLTLC(Npys)-NH2 and Bzl-C(Npys)KRLTL-NH2 in the enzymaticactivity was determined using the same procedure, after 5 min of pre-incu-bation with the enzyme preparation. Kinetic parameters were determined bymeasuring the initial rate of hydrolysis at various inhibitor concentrationsand the residual activity was evaluated. Fluorescence emission was contin-uously measured and inhibition constant values (Ki) were obtained using theGraFit 3.0 computer program (Erithacus Software Ltd.).

Cell detachment assays. LLC-MK2 cells were expanded overnight in24-well culture plates (2 · 104 cells/mL/well) at 37 �C, in a 5% CO2

chamber. The cells were maintained in DMEM (Dulbecco’s modifiedEagle’s medium, Gibco) containing 0.37% NaHCO3 and 10% FBS. Theplastic-attached cells were washed five times with 0.5 mL of phosphate-buffered saline (PBS), then incubated for 5, 10, 15, 20, 30, 45 or 60 minwith 200 lL of PBS/0.05 M EDTA containing 10 lL of PbST prepa-ration (fraction 4, Fig. 1) previously inactivated or not by incubationwith C(Npy) inhibitors for 10 min at 37 �C. Controls were assayed inthe absence of enzyme. Classical inhibitors were used to characterize theactivity at the following excess concentrations: 2 mM PMSF, 1 mM p-HMB, 0.13 mM E-64, and 13 lM pepstatin. The tests were performedin the presence of EDTA, which inhibits metallo proteinases. Cells wereharvested and counted in a Neubauer chamber using Trypan bluestaining for viability. The experiments were carried out in triplicatesand statistical analyses were performed using the one-tailed distributionStudent’s t-test.

In-gel protein digestion and LC–MS/MS analysis. Proteins fraction-ated by SDS–PAGE were silver-stained using the Vorum method [15].In-gel protein digestion was performed as described [16]. Resultingpeptides were recovered from the gel and desalted in a Poros R2-50(Applied Biosystems) zip-tip, dried in a vacuum centrifuge, dissolved in30 lL of 0.1% formic acid (FA), and subjected (1 lL) to liquid chro-matography-mass spectrometry (LC–MS) analysis. LC was performedin a PepMap reversed-phase column (15 cm · 75 lm, 3-lm C18, LCPackings, Dionex) coupled to an Ultimate (LC Packings, Dionex)nanoHPLC. Bound peptides were eluted with 5–42.5% acetonitrile/0.1%FA over 30 min and directly analyzed in a Q-tof 1 (Micromass, Waters)mass spectrometer. Spectra in positive-ion mode were collected in the400–1800 m/z range, and each peptide was fragmented (MS/MS) for 3 sin the 50–2050 m/z range. MS data were converted into peak lists (PKLformat) and analyzed by the Phenyx (GeneBio) and Mascot (MatrixScience) softwares through the NCBInr and P. brasiliensis protein andnucleotide sequences from GenBank. Peptides with no match in thesedatabases were subjected to de novo sequencing using the MassLynxv4.0 software (Waters).

2 3 4 5 660 0 0 10 81

0

750

20(AUF/min)

In FT

47

73

29

0 5’ 10’ 15’ 30’

FN

200

200

Fig. 1. Isolation of PbST by chromatography in a pABA-Sepharosecolumn. Upper panel: silver-stained 10% SDS–PAGE gel showing theprofile of culture supernatant (10 ll) from P. Brasiliensis yeast cells before(In) and after (FT) chromatography. Eluted fractions are numbered. Thehydrolytic activity against Abz-MKALTLQ-EDDnp produced by a 10 ll-aliquot of each fraction is shown as arbitrary units of fluorescence perminute (AUF/min). The results are not presented in terms of specificactivity due to the low protein contents. The hydrolytic activity againstfibronectin (FN) correspondent to incubation with each fraction at 37 �Cfor 30 min is shown in a Coomassie blue-stained 8% SDS–PAGE gel. Atime-course of FN cleavage after incubation for 5, 10, 15, and 30 min withpeak fraction 4 (10 ll) is also depicted. The positions of molecular massmarkers (kDa) and of one of the FN chain are shown on the left.

1002 A.L. Matsuo et al. / Biochemical and Biophysical Research Communications 355 (2007) 1000–1005

Results and discussion

In previous publications we characterized the PbSTactivity using preparations of P. brasiliensis supernatantsenriched by either ion exchange/gel-filtration chromatogra-phy [2,3] or hydrophobic phenyl-Superose columns [4,17].To date, we have tried to purify the proteinase using manydifferent chromatographic methods (not published), how-ever total purification of the active proteinase has neverbeen achieved due to aggregation to other fungal compo-nents and/or loss of enzymatic activity after several chro-matographic steps.

Presently, a cell-free supernatant from P. brasiliensis B-339 yeast-cell culture bearing PbST proteolytic activity wasfractionated in a pABA-Sepharose column. The p-amino-benzamidine (pABA) compound is a commercially avail-able competitive inhibitor of trypsin that binds to thespecificity cleavage pocket of the enzyme [18]. We observedthat the PbST activity capable of hydrolyzing the FRETpeptide substrate Abz-MKALTLQ-EDDnp was bound tothe column and predominantly eluted in fractions 4 and5, as seen in Fig. 1. These fractions also concentrated theproteolytic activity towards fibronectin found in total

supernatants, as suggested by extensive degradation ofthe protein resulting from incubation with fractions 4 or5 (Fig. 1). A time-course assay showed that the proteolyticactivity contained in fraction 4 (PbST-4) was so strong thatafter only 5 min of incubation protein degradation wasalready evident (Fig. 1, bottom). Enzymatic cleavage ofAbz-MKALTLQ-EDDnp or human fibronectin was dueto PbST, since hydrolysis was specifically inhibited byp-HMB and PMSF (not shown), as previously reported[2]. In our experimental conditions, the activity againstAbz-MKALTLQ-EDDnp was fully bound to the columnand was recovered in the eluted fractions at a yield of about65%. In addition, PbST-4 was not active against casein orBSA (not shown), as previously described [3]. The chro-matographic profile in SDS–PAGE silver-stained gelsshowed that fractions 4 and 5 had a predominant 55 kDacomponent (Fig. 1), among others that appeared afterlonger development times of silver-stained gels. Prepara-tion PbST-4 could be kept at 4 �C for about 7 days beforeactivity started to gradually decrease.

Previously, we have been able to detect in situ activity ofPbST in SDS–PAGE gels [17]. The PbST activity wasrevealed as a broad smeared band migrating between 43and 68 kDa in gelatin zymograms or using an Abz-MKALTLQ-EDDnp-agarose overlay. It is uncertainwhether the proteinase molecular mass is around 50 kDaor if migrates as such due to aggregation in our electropho-retic conditions, where the samples were denatured at lowSDS concentration and without boiling. Presently, pABApreparation PbST-4 was tested in gelatin zymograms, buta proteolysis band has not been detected, possibly becausepABA-isolated PbST was unstable after electrophoresis.We then chose to carry out an LC–MS/MS analysis ofthe whole PbST-4 preparation and also of individual maingel bands present in overloaded gels. We consistently iden-tified NADP-dependent glutamate dehydrogenase (corre-spondent to the 55-kDa band), LPD1 (lipoamidedehydrogenase), and chitinase based on the high numberof tryptic peptides bearing over 10 amino acids identicalto peptides found in the NCBInr and GenBank databasecontaining fungal proteins. The identified proteins shouldcontain motifs that were recognized by pABA or they wereunspecifically bound to the resin as aggregates. Although acouple of peptides would match those of subtilisins, thenumber of identical amino acids was not enough to guaran-tee reliable protein identification. One difficulty to analyzegenerated peptides is the lack of complete genome informa-tion for P. brasiliensis, although there are two GenBankdatabases of partial translated sequences [19,20]. It isimportant to point out that there is a 481-amino acidP. brasiliensis subtilase of the S8 family deposited in theGenBank (Accession No. AAP83193). However, our pres-ent analysis suggests that PbST does not correspond to thisproteinase, since identical peptides have not beenidentified.

We used pABA-eluted PbST-4 to evaluate the inhibitoryeffect of peptides derived from MKRLTL coupled with

Table 1Inhibition of PbST activity by Boc-C(Npys) and its derivatives bound tothe sequence MKRLTL

Compound Ki (M)

Boc-C(Npys) 1.7 · 10�6

Bzl-MKRLTLC(Npys)-NH2 1.6 · 10�7

Bzl-C(Npys)KRLTL-NH2 1.6 · 10�8

A.L. Matsuo et al. / Biochemical and Biophysical Research Communications 355 (2007) 1000–1005 1003

C(Npys). A general mechanism of action of C(Npys)-pep-tides is schematically represented in Fig. 2, which showsthat a peptide sequence containing C(Npys) can block afree thiol group, liberating Npys. We presently synthesizedtwo analogous peptides, namely Bzl-MKRLTLC(Npys)-NH2 and Bzl-C(Npys)KRLTL-NH2, which bear theC(Npys) group, respectively, at the C- and N-terminus. Athird compound, Boc-Cys(Npys), was used to test theinhibitory capacity of C(Npys) devoid of a peptidesequence.

We observed that the three compounds inhibited theenzymatic activity against Abz-MKALTLQ-EDDnp [17]at different levels (Table 1). The best inhibitory activitywas seen with Bzl-C(Npys)KRLTL-NH2 (Ki = 16 nM),where the C(Npys) group was placed at the N-terminussubstituting the original methionine residue. The peptidebearing C(Npys) at the C-terminus was one order of mag-nitude less effective (Ki = 160 nM), whereas the Boc-C(Npys) compound was a weaker inhibitor (Ki = 1.7 lM).These results indicated that the peptide sequence promotesa favorable interaction increasing the magnitude of inhibi-tion, especially when C(Npys) was placed at a further posi-tion from L–T, at the N-terminus. The inhibitory effect ofall tested peptides was completely reverted with 5 mM dith-ioerythritol, indicating that it was due to the reactivity ofthe C(Npys) moiety with a free SH– group of the protein-ase (Fig. 2). It is noteworthy that the inhibitor peptideshave not been cleaved by the proteinase activity, asrevealed by h.l.p.c. analysis (not shown).

We have previously reported on the PbST capacity tocleave components associated with the basement mem-brane [3]. Presently, we have not been able to test the inhib-itory capacity of C(Npys) peptides on the cleavage offibronectin visualized in SDS–PAGE gels (as in Fig. 1)because they altered fibronectin resolution. Therefore, wedeveloped a cell detachment assay for this purpose. Wetested PbST ability to detach cells of the LLC-MK2 lineagefrom plastic dishes, which is generally achieved using tryp-sin. Cell detachment depends on cleavage of extracellularmatrix components or of surface integrins that directlyinteract with them. Fig. 3A shows that PbST-4 was ableto detach LLC-MK2 at increased rates over time, whichreached a plateau corresponding to approximately 50%of total cells detached after 20 min. Consequently, thisincubation time was chosen for the inhibition experimentsshown below. All experiments were carried out in tripli-cates and harvested cells were 70–80% viable. It is worthmentioning for comparison that 2.5 mg/mL of trypsin

Fig. 2. Schematic representation of the mechanism

can detach all the cells (90–95% viable) in 5 min (notshown). Fig. 3B shows that the detachment activity wasdue to PbST, considering that pre-incubations with E-64and pepstatin (or EDTA contained in the PBS) did notaffect the activity, whereas p-HMB and PMSF totally abol-ished it (Fig. 3B).

In order to evaluate the inhibitory capacity of C(Npys)peptides in the cell detachment activity of PbST-4, theexperiments were performed with the enzyme preparationpreviously incubated or not (control) with 6 or 12 lM ofeach peptide for 20 min at 37 �C. Fig. 3C shows that bothBzl-C(Npys)KRLTL-NH2 and Bzl-MKRLTLC(Npys)-NH2 exhibited statistically significant inhibitory effects, ina dose-dependent fashion. Bzl-C(Npys)KRLTL-NH2 wasa better inhibitor than Bzl-MKRLTLC(Npys)-NH2, inagreement with the Ki results seen in Table 1. Boc-Cys(Npys) showed no statistically significant effect at theconcentrations tested, probably due to its lower bindingaffinity, as suggested from the results in Table 1.

This is the first time C(Npys) compounds are used toinhibit a thiol-containing subtilisin. It remains to be testedwhether these compounds would effectively decrease P.

brasiliensis virulence in vivo. For that purpose, C(Npys)might be bound to peptides of the D-anomericity, fore.g., to avoid proteolysis by local endopeptidases.

The results presented here elegantly show that PbSTactivity can be blocked with compounds interfering witha free cysteine, confirming that it belongs to the subtilisinS8 family of serine proteinases, from which proteinase Kfrom Tritirachium album is the best studied member [21].By analogy with the members of this group [22], we specu-late that there is a thiol group lying near the imidazole ofthe histidine from the Ser-His-Asp catalytic triad. The pres-ence of such a residue does not change the catalytic activityof subtilisin Savinase, although it makes it susceptible tomercurials [23]. In proteinase K, the thiol group (Cys78)is buried in the molecule and covalently binds to the firstHg+ molecule, which disrupts the catalytic triad by chang-ing the imidazole conformation [24]. A second Hg+ com-plexes non-covalently with His72, Cys78 and Thr76,

of inhibition of thiol proteinases by Cys(Npys).

0123456789

10

0 10 20 30 40 50 60 70

Cel

ls/f

ield

time (min)

A

B

0

20

40

60

80

100

120

* *

*

* **

C- N- Boc-

Control PMSF p-HMB E-64 pepstatin 6 12 126 126 µM

Fig. 3. Evaluation of PbST inhibition in a cell detachment assay. (A) Time-course of cell detachment when incubated with PbST-4 (10 lL, fraction 4,Fig. 1). (B) Percentage of cell detachment when compared to the control (100%), as counted after 20 min of incubation with PbST-4 in the presence ofeither standard proteinase inhibitors or Bzl-MKRLTLC(Npys)-NH2 (C-), Bzl-C(Npys)KRLTL-NH2 (N-) and Boc-Cys(Npys) (Boc-) at the indicatedconcentrations. Incubation of cell cultures with equivalent concentrations of the inhibitors alone has not resulted in cell detachment or loss of cell viability(not shown). Statistically significant values (p < 0.05) are indicated with an asterisk.

1004 A.L. Matsuo et al. / Biochemical and Biophysical Research Communications 355 (2007) 1000–1005

decreasing substrate-binding affinity. However, 100% inhi-bition by mercurials is never achieved, possibly due to flex-ibility of the molecule [24]. In contrast to proteinase K,PbST is selective regarding substrate specificity and isinhibited 100% by p-HMB [2]. On the other hand, proxim-ity of the thiol group to the catalytic triad would explainwhy the C(Npys)-peptide derivatives are better inhibitorsthan Boc-C(Npys), considering that the peptide used herefits in the substrate pocket to target the inhibitor.

Acknowledgments

We thank Dr. Ivarne Tersariol for critical review of themanuscript. This work and fellowships involved were sup-ported fully or partially by grants from FAPESP, PRO-NEX/CNPq, CNPq, and NIH/NCRR (Grant No.5G12RR008124).

References

[1] M. Monod, S. Capoccia, B. Lechenne, C. Zaugg, M. Holdom, O.Jousson, Secreted proteases from pathogenic fungi, Int. J. Med.Microbiol. 292 (2002) 405–419.

[2] A.K. Carmona, R. Puccia, M.C.F. Oliveira, E.G. Rodrigues, L.Juliano, L.R. Travassos, Characterization of an exocellular serine-thiol proteinase activity in Paracoccidioides brasiliensis, Biochem. J.309 (1995) 209–214.

[3] R. Puccia, A.K. Carmona, J.L. Gesztesi, L. Juliano, L.R. Travassos,Exocellular proteolytic activity of Paracoccidioides brasiliensis: cleav-

age of components associated with the basement membrane, Med.Mycol. 36 (1998) 345–348.

[4] A.L. Matsuo, I.L. Tersariol, S.I. Kobata, L.R. Travassos, A.K.Carmona, R. Puccia, Modulation of the exocellular serine-thiolproteinase activity of Paracoccidioides brasiliensis by neutral poly-saccharides, Microbes Infect. 8 (2006) 84–91.

[5] G. San Blas, G. Nino-Vega, T. Iturriaga, Paracoccidioides brasiliensis

and paracoccidioidomycosis: molecular approaches to morphogene-sis, diagnosis, epidemiology, taxonomy and genetics, Med. Mycol. 40(2002) 225–242.

[6] R. Matsueda, R. Walter, 3-nitro-2-pyridinesulfenyl (Npys) group—anovel selective protecting group which can be activated for peptide-bond formation, Int. J. Pept. Protein Res. 16 (1980) 392–401.

[7] R. Matsueda, H. Umeyama, E. Kominami, N. Katunuma, Designand synthesis of cathepsin-B inhibitors by an affinity labelingapproach, Chem. Lett. (1988) 1857–1860.

[8] M.S. Bernatowicz, R. Matsueda, G.R. Matsueda, Preparation ofBoc-[S-(3-nitro-2-pyridinesulfenyl)]-cysteine and its use for unsym-metrical disulfide bond formation, Int. J. Pept. Protein Res. 28 (1986)107–112.

[9] D.H. Rich, in: A.J. Barret, G. Salvesen (Eds.), Proteinase inhibitors,Elsevier Science Publishers, 1986, p. 153.

[10] R. Matsueda, H. Umeyama, R.N. Puri, H.N. Bradford, R.W.Colman, Potent affinity labeling peptide inhibitors of calpain, Chem.Lett. (1990) 191–194.

[11] E. Dejana, M. Corada, Adhesion protocols, in: J.M. Walker (Ed.),Methods in Molecular Biology, Editora Humana, 1990, pp. 119–124.

[12] U.K. Laemmli, Cleavage of structural proteins during assembly ofhead of bacteriophage-T4, Nature 227 (1970) 680–685.

[13] M. Hirata, K. Nemoto, A. Hirata, Solubility of N-(benzyloxycar-bonyl)-L-aspartyl-L-phenylalanine methyl-ester forming a complexwith L-phenylalanine methyl-ester in aqueous system, J. Chem. Eng.27 (1994) 72–76.

A.L. Matsuo et al. / Biochemical and Biophysical Research Communications 355 (2007) 1000–1005 1005

[14] G. Barany, R.B. Merrifield, The Peptides: Analysis, Synthesis andBiology, Academic Press, New York, 1980.

[15] E. Mortz, T.N. Krogh, H. Vorum, A. Gorg, Improved silver stainingprotocols for high sensitivity protein identification using matrix-assisted laser desorption/ionization-time of flight analysis, Proteomics1 (2001) 1359–1363.

[16] A. Shevchenko, M. Wilm, O. Vorm, M. Mann, Mass spectrometricsequencing of proteins from silver stained polyacrylamide gels, Anal.Chem. 68 (1996) 850–858.

[17] R. Puccia, M.A. Juliano, L. Juliano, L.R. Travassos, A.K. Carmona,Detection of the basement membrane-degrading proteolytic activityof Paracoccidioides brasiliensis after SDS–PAGE using agaroseoverlays containing Abz-MKALTLQ-EDDnp, Braz. J. Med. Biol.Res. 32 (1999) 645–649.

[18] A. Kanamori, N. Seno, I. Matsumoto, Preparation of high-capacityaffinity adsorbents using formyl carriers and their use for low-performance and high-performance liquid affinity-chromatography oftrypsin-family proteases, J. Chromatogr. 363 (1986) 231–242.

[19] G.H. Goldman, E.D. Marques, D.C.D. Ribeiro, L.A.D. Bernardes,A.C. Quiapin, P.M. Vitorelli, M. Savoldi, C.P. Semighini, R.C.Oliveira, L.R. Nunes, L.R. Travassos, R. Puccia, W.L. Batista, L.E.Ferreira, J.C. Moreira, A.P. Bogossian, F. Tekaia, M.P. Nobrega,F.G. Nobrega, M.H.S. Goldman, Expressed sequence tag analysis ofthe human pathogen Paracoccidioides brasiliensis yeast phase: Iden-

tification of putative homologues of Candida albicans virulence andpathogenicity genes, Eukaryot. Cell. 2 (2003) 34–48.

[20] M.S.S. Felipe, R.V. Andrade, F.B.M. Arraes, A.M. Nicola, A.Q.Maranhao, F.A.G. Torres, I. Silva-Pereira, M.J. Pocas-Fonseca, E.G.Campos, L.M.P. Moraes, P.A. Andrade, A.H.F.P. Tavares, S.S.Silva, C.M. Kyaw, D.P. Souza, P. Network, M. Pereira, R.S.A.Jesuino, E.V. Andrade, J.A. Parente, G.S. Oliveira, M.S. Barbosa,N.F. Martins, A.L. Fachin, R.S. Cardoso, G.A.S. Passos, N.F.Almeida, M.E.M.T. Walter, C.M.A. Soares, M.J.A. Carvalho, M.M.Brigido, Transcriptional profiles of the human pathogenic fungusParacoccidioides brasiliensis in mycelium and yeast cells, J. Biol.Chem. 280 (2005) 24706–24714.

[21] A. Pahler, A. Banerjee, J.K. Dattagupta, T. Fujiwara, K. Lindner,G.P. Pal, D. Suck, G. Weber, W. Saenger, Three-dimensionalstructure of fungal proteinase K reveals similarity to bacterialsubtilisin, EMBO J. 3 (1984) 1311–1314.

[22] R.J. Siezen, J.A. Leunissen, Subtilases: the superfamily of subtilisin-like serine proteases, Protein Sci. 6 (1997) 501–523.

[23] A. Muller, W. Saenger, Studies on the inhibitory action of mercuryupon proteinase K, J. Biol.Chem. 268 (1993) 26150–26154.

[24] L.M. Bech, S. Branner, S. Hastrup, K. Breddam, Introduction ofa free cysteinyl residue at position 68 in the subtilisin Savinase,based on homology with proteinase K, FEBS Lett. 297 (1992)164–166.

36

Proteinase Lon mitocondrial

37

Artigo 3: The PbMDJ1 gene belongs to a conserved MDJ1/LON locus in

thermodimorphic pathogenic fungi and encodes a heat shock protein that localizes to both

the mitochondria and cell wall of Paracoccidioides brasiliensis.

Wagner L. Batista, Alisson L. Matsuo, Luciane Ganiko,Tânia F. Barros, Thiago R.

Veiga, Edna Freymuller e Rosana Puccia

Eukaryotic Cell 2006, Vol 5 (2) 379-390

38

Racional

Em nosso laboratório, o gene PbLON foi clonado acidentalmente quando

buscávamos clonar o gene da PbST. Durante a caracterização do seu locus, foi

encontrado em direção oposta o gene da PbMDJ1 compartilhando com PbLON a região

5’ intergênica. Ambos os genes codificam proteínas com direcionamento mitocondrial, as

quais em leveduras têm uma relação funcional, já que a Mdj1 participa do complexo

formado pela Lon e a proteína desnaturada alvo de proteólise. Neste trabalho verificou-se

que o locus LON/MDJ1 é conservado entre os Ascomicetos e que a PbMdj1 tem uma

localização extramitocondrial na parede celular de leveduras do P. brasiliensis, enquanto

a PbLon fica restrita à mitocôndria.

EUKARYOTIC CELL, Feb. 2006, p. 379–390 Vol. 5, No. 21535-9778/06/$08.00�0 doi:10.1128/EC.5.2.379–390.2006Copyright © 2006, American Society for Microbiology. All Rights Reserved.

The PbMDJ1 Gene Belongs to a Conserved MDJ1/LON Locusin Thermodimorphic Pathogenic Fungi and Encodes a Heat

Shock Protein That Localizes to both the Mitochondriaand Cell Wall of Paracoccidioides brasiliensis

Wagner L. Batista,1 Alisson L. Matsuo,1 Luciane Ganiko,1† Tania F. Barros,1‡Thiago R. Veiga,1 Edna Freymuller,2 and Rosana Puccia1*

Department of Microbiology, Immunology and Parasitology1 and Center of Electron Microscopy,2

Federal University of Sao Paulo, Rua Botucatu 862, 04023-062 Sao Paulo, SP, Brazil

Received 30 September 2005/Accepted 18 November 2005

J-domain (DnaJ) proteins, of the Hsp40 family, are essential cofactors of their cognate Hsp70 chaperones,besides acting as independent chaperones. In the present study, we have demonstrated the presence of Mdj1,a mitochondrial DnaJ member, not only in the mitochondria, where it is apparently sorted, but also in the cellwall of Paracoccidioides brasiliensis, a thermodimorphic pathogenic fungus. The molecule (PbMdj1) was local-ized to fungal yeast cells using both confocal and electron microscopy and also flow cytometry. The anti-recombinant PbMdj1 antibodies used in the reactions specifically recognized a single 55-kDa mitochondrialand cell wall (alkaline �-mercaptoethanol extract) component, compatible with the predicted size of the proteindevoid of its matrix peptide-targeting signal. Labeling was abundant throughout the cell wall and especially inthe budding regions; however, anti-PbMdj1 did not affect fungal growth in the concentrations tested in vitro,possibly due to the poor access of the antibodies to their target in growing cells. Labeled mitochondria stoodpreferentially close to the plasma membrane, and gold particles were detected in the thin space between them,toward the cell surface. We show that Mdj1 and the mitochondrial proteinase Lon homologues are heat shockproteins in P. brasiliensis and that their gene organizations are conserved among thermodimorphic fungi andAspergillus, where the genes are adjacent and have a common 5� region. This is the first time a DnaJ memberhas been observed on the cell surface, where its function is speculative.

Paracoccidioides brasiliensis is the fungal species responsiblefor paracoccidioidomycosis (PCM), which is endemic in cer-tain areas of Latin America. PCM is one the four deep-seatedgranulomatous mycoses caused by thermodimorphic fungi thatalso include Histoplasma capsulatum, Blastomyces dematitidis,and Coccidioides immitis. These fungal species are phyloge-netically related, according to sequence analysis of the ribo-somal RNA gene locus (3, 36), and have been classified asAscomycetes of the Onigenaceae family. P. brasiliensis grows asa multibudding yeast when in parasitism or cultivated at 37°Cand as a mycelium when incubated at temperatures below28°C. The fungus is multinucleated, and its sexual form is stillunknown. Genetic manipulation in this species is only startingto be attempted (23).

The human host most frequently acquires the fungus byinhalation of mycelial particles, but infection can be estab-lished only after transition to the yeast parasitic phase takesplace (28). A number of biochemical processes are associatedwith fungal adaptation to the host’s environment and temper-ature, and many of them are probably responsible for phasetransition and/or phase maintenance. Different groups haverecently addressed the overall scenario of gene expression in P.

brasiliensis yeast versus mycelium, or undergoing phase transi-tion in vitro, making use of microarrays, suppression subtrac-tion hybridization, real-time reverse transcriptase (RT) PCR,and in silico analyses (14, 15, 27, 33). Considering that theseanalyses have been undertaken with P. brasiliensis transformingto the yeast phase upon temperature increase, differentiallyregulated heat shock protein homologues have indeed beendetected, especially in the initial hours of temperature change(15, 33). Some of them, however, seem to be necessary forgrowth in the yeast phase, e.g., the HSP70 and HSP60 genes inP. brasiliensis that have been previously characterized (18, 41).

We have characterized a LON gene homologue from P.brasiliensis (PbLON) (2). Lon proteins are conserved ATP-binding, heat-inducible serine proteinases. They form high-molecular-mass complexes of homo-oligomers, which inSaccharomyces cerevisiae contain the stoichiometry of sevensubunits of 117 kDa (46). The yeast Lon, also called PIM1, issynthesized as a pre-proenzyme precursor in the cytosol andthen is sorted to the mitochondrial matrix where it is processed(54). It controls proteolysis by mediating cleavage of misfoldedor unassembled matrix proteins and has an essential role in themaintenance of mitochondrial DNA integrity and mitochon-drial homeostasis (51). Lon is also a DNA-binding protein (26)and is essential for cell survival in humans (4).

In P. brasiliensis, PbLon consists of 1,063 residues containinga mitochondrial import signal, a conserved ATP-binding site,and a serine catalytic motif (2). Its open reading frame (ORF)is within a 3,369-bp fragment interrupted by two introns lo-

* Corresponding author. Mailing address: Rua Botucatu 862, oitavoandar, 04023-062 Sao Paulo, SP, Brazil. Phone: (55) 11 5084 2991.Fax: (55) 11 5571 5877. E-mail: rosana@ecb.epm.br.

† Present address: University of Texas at El Paso, El Paso, Tex.‡ Present address: Department of Clinical and Toxicological Anal-

yses, Federal University of Bahia, Bahia, BA, Brazil.

379

cated in the 3� segment; an MDJ1-like gene was partially se-quenced in the opposite direction but sharing with PbLON acommon 5� untranslated region (2). This chromosomal organi-zation might be functionally relevant, since Mdj1p is a type IDnaJ molecule located in the yeast mitochondrial matrix and isessential for substrate degradation by Lon and other stress-inducible ATP-dependent proteinases (39, 53). In bacteria,DnaJ alone and/or the DnaJ/DnaK complex determine the fateof nonnative substrates to be cleaved by Lon (17).

DnaJ members (more recently referred to as J-domain pro-teins) belong to the Hsp40 family of molecular chaperones andlocalize to various cellular compartments, where their primaryrole is to regulate the activity of their cognate Hsp70 homo-logues (55). In the bacterial DnaK (Hsp70)/DnaJ/GrpE com-plex, a transient association of DnaJ stimulates the ATPaseactivity of DnaK, prompting substrate binding, while GrpEexchanges nucleotides (ADP/ATP), with the consequent sub-strate dissociation (12). This complex prevents aggregation ofnonnative proteins, thus facilitating their folding by Hsp60(22). In S. cerevisiae, Mdj1 is essential for the biogenesis offunctional mitochondria, but it is not involved in protein trans-location into the matrix (reviewed in reference 52). Typically,DnaJ members are constituted of a structurally conserved Jdomain located near the N-terminal end, followed by a glycine/phenylalanine-rich segment (G/F domain) and four repetitionsof a zinc-binding CXXCXGXG motif (55). The J domain andthe G/F segment are essential for interaction with DnaK, whilethe zinc finger-like motif contains elements for binding non-native substrates (49) and also displays thiol-disulfide oxi-doreductase activity (50). In S. cerevisiae, 22 J-domain proteins(types I, II, and III) have been found in the genome; theirsubcellular localizations are attributed to the cytoplasm, nucle-oplasm, mitochondria, and endoplasmic reticulum (55).

We have presently characterized the P. brasiliensis PbMDJ1gene and verified that the PbLON/PbMDJ1 locus is conservedamong the dimorphic pathogenic fungi besides Aspergillus nidu-lans and A. fumigatus. We show that PbLON and PbMDJ1 encodemitochondrial heat shock proteins; however, PbMdj1 has alsobeen found in the cell wall of P. brasiliensis. The finding of aDnaJ member on the cell surface is a novel observation.

MATERIALS AND METHODS

Fungal strains and culture conditions. We used P. brasiliensis isolates B-339and 18, provided by Angela Restrepo (Colombia) and Zoilo P. Camargo (Brazil),respectively. More details about the fungal isolates can be found elsewhere (29).Cultures were maintained at 36°C (yeast phase) in solid modified YPD (0.5%casein peptone, 0.5% yeast extract, 1.5% glucose, pH 6.3).

Isolation of total DNA and RNA. For DNA extraction, strain B-339 wasexpanded for 10 days at 36°C in liquid modified YPD under agitation. TotalDNA was extracted from a 10-ml pellet of nitrogen-frozen cells that were dis-rupted in a mortar and then briefly in a pestle, as described previously (11). TotalRNA was isolated from fresh cells mechanically disrupted by vortexing with glassbeads for 10 min in the presence of TRIzol reagent (Invitrogen), according to theinstructions provided by the manufacturer. Contaminating DNA in these prep-arations was digested with RNase-free DNase I (Promega), as described previ-ously (15). The efficiency of hydrolysis was tested by PCR with primers of thePbGP43 gene (11). For analysis of gene expression in cells undergoing heat shockat 42°C, yeast cells growing logarithmically at 36°C under shaking in modifiedYPD were transferred to 42°C for 30 and 60 min.

Cloning of the 3� end of PbMDJ1. We followed the strategy of 3� rapidamplification of cDNA ends to clone the 3� end of the PbMDJ1 cDNA, using theSuperScript II RNase H RT kit (Invitrogen), because at that time the PbMDJ1expressed sequence tag (EST) had not yet been found in the P. brasiliensis EST

bank (15). The positions of the primers can be seen in Fig. 1A. The first strandof the total cDNA pool was obtained from total RNA (2.5 �g) extracted fromstrain B-339 yeast cells previously heat shocked for 60 min at 42°C. The templatewas incubated (10 �l, final volume) with a 1 �M concentration of a poly(T)-containing cassette primer (511, 5�-GACTCGAGTCGACATCGT17-3�) at 65°C(5 min), cooled in ice, mixed with synthesis buffer (final concentrations of 0.05 MTris-HCl [pH 8.3], 0.075 M KCl, 3 mM MgCl2, 0.01 M DDT, 1 mM [each]deoxynucleoside triphosphate, 2 U/�l of RNAseOUT, and 10 U/�l of Super-Script II RT), and incubated at 55°C (50 min) and then at 85°C (5 min).Template RNA was removed with 0.1 U/�l of RNase for 20 min at 37°C.Negative controls were incubated in the absence of RT. The second strand wasobtained from total cDNA (1 �1 of the above reaction product) with an internalsense primer (B4, 5�-TGGTGGTGGGTTCTCTGC-3�) under standard PCRconditions at an annealing temperature of 52°C. We used Platinum Taq poly-merase (Invitrogen) and an antisense cassette primer (512, 5�-GACTCGAGTCGACATCG-3�) for amplification of the double-stranded PbMDJ1 fragment.The final product was used as a template in a second round of PCR primed withantisense primer 512 and nested sense primer G5 (5�-TGCCTTTGCGGGTGGTTC-3�). Sequencing information from the resulting 1,161-bp amplicon was usedto design primers located at the 3� end (antisense 2A7, 5�-TACCATTTTTTGCCTGCC-3�), which enabled PCR amplification of the whole genomic PbMDJ1gene using genomic DNA as a template. A cDNA segment of the 5� region, usedlater for heterologous expression, was amplified by RT-PCR carried out withtotal RNA (1 �g) and the ThermoScript RT-PCR System (Gibco BRL). The firstcDNA strand was amplified with the internal antisense primer E4 (5�-GCCGCACGAGGAACAGG-3�), according to the supplier’s recommendations. A PCRproduct of 757 bp was then obtained with primers B3 (sense, 5�-CTCCCGGCTCGTCTCCTG-3�) and A7 (antisense, 5�-CCGCTTTGTACCCTGTTT-3�),with the first-strand reaction product as a template. Sequence comparison be-tween DNA and cDNA fragments enabled the precise localization of two introns.Negative controls for 3� rapid amplification of cDNA ends and RT-PCRs wereincubated in the absence of RT. DNA fragments were recovered from agarosegels using either the Nucleiclean kit (Sigma) or the Concert gel extraction system(Gibco BRL). Purified amplicons were cloned into the pGEM-T easy vector(Promega) for maintenance and sequencing. Nucleotide sequencing was carriedout in the facilities of the Center of Human Genome at the Sao Paulo University.Sequences were compared with those of the National Center for BiotechnologyInformation through the GenBank database and the BLAST network service andanalyzed with the EditSeq, Seqman, Megalign, and Protean programs of theLasergene System (DNAstar Inc.).

Northern blotting. Standard conditions were used for electrophoresis andNorthern blotting (38), carried out with 25 �g of total RNA per lane. Hybrid-izations with [�-32P]dCTP-labeled probes were done in nylon Hybond-N mem-branes (Amersham) under high-stringency conditions with a B3/2A7 DNA frag-ment of PbMDJ1 (Fig. 1A) and a PbLON amplicon of 850 bp (from nucleotide[nt] 536 to nt 1386). The probes were labeled using the RadPrime DNA labelingsystem (Gibco BRL) as specified by the manufacturer and separated from freenucleotides in a molecular biology-grade Sephadex G-50 column (Pharmacia).The membranes were hybridized with the labeled probes (1 � 106 cpm/ml) at42°C overnight in 50% formamide, 6� SSC (1� SSC is 0.15 M NaCl, 0.015 Msodium citrate), 5� Denhardt’s solution, 0.5% sodium dodecyl sulfate (SDS),and 100 �g/ml of salmon sperm DNA; washed one time with SSC–0.1% SDS (20min, 42°C) and three times with 0.2� SSC–0.1% SDS (20 min, 68°C); coveredwith Vitafilm (Goodyear); and exposed to an X-ray film (X-Omat; Kodak) at�70°C with an intensifying screen.

Protein extraction. Total cell extracts of P. brasiliensis were obtained throughglass bead disruption (8). For isolation of total mitochondrial fractions, we adaptedfor P. brasiliensis the method described by Suzuki et al. (48) for S. cerevisiae. In orderto accomplish cell lysis under mild conditions, nitrogen-frozen yeast cells weremechanically disrupted in a mortar until they formed a fine white powder, which wassubsequently thawed in BB buffer (0.6 M sorbitol, 20 mM HEPES, pH 7.4) andsonicated for 5 min, alternating 15-s pulses with 15 s in ice. Cell debris were pelletedby centrifugation (1,500 � g, 5 min) and the ice-cooled supernatant was centrifugedat 1,200 � g for 10 min to precipitate the mitochondrial fraction (brownish pellet),which was then rinsed with BB and centrifuged (1,500 � g, 5 min). The supernatantwas subjected to the same procedure two times, the resulting brownish pellets werepooled and washed two times with BB and suspended in 0.2 ml of BB, and theprotein concentration was estimated by spectrometry in an aliquot diluted 100 timesin 0.6% SDS (an A280 of 0.2 corresponds to 10 mg/ml of mitochondrial proteins).Total mitochondrial protein preparations were kept at �70o in 20% glycerol–1mg/ml bovine serum albumin (BSA).

A �-mercaptoethanol cell wall extract was obtained from intact cells as de-scribed previously (9). P. brasiliensis yeast cells grown under shaking for 10 days

380 BATISTA ET AL. EUKARYOT. CELL

(cell viability, �85%) in modified YPD were pelleted by centrifugation (1,500 �g, 15 min) and rinsed three times with an excess volume of phosphate-bufferedsaline (PBS). Washed cells (12 ml of wet pellet) were suspended in 30 ml of 20mM ammonium carbonate (pH 8.6)–1% �-mercaptoethanol–1 mM phenylmethyl-sulfonyl fluoride and agitated for 1 h at 36°C. The released soluble componentswere separated from the cells by centrifugation, dialyzed in H2O, and concen-trated to 1.5 ml by lyophilization (about 8 mg/ml of total protein). The treatmentdid not affect cell integrity, as verified by microscopy of the treated cells. Proteincontents were estimated using BSA as a standard (5).

Expression of PbMdj1 and PbLon in bacteria. Expression of PbMdj1 andPbLon in bacteria was achieved with the pHIS1 and pHIS3 vectors (40). A757-bp cDNA fragment of the 5� region of PbMDJ1, obtained by RT-PCR withprimers B3/A7 (Fig. 1A; see description above), was excised from pGEM-T withEcoRI/SpeI and subcloned in the same sites of pHIS3 in frame with the sequenceencoding a His6 tag. The resulting construct was called pHISMdj1. For expres-sion of PbLon, an 845-bp BamHI/HindIII intron-free 5� gene fragment wasexcised from a pUCSmaI plasmid containing the entire PbLON genomic se-quence (2) and subcloned into the same sites of pHIS1 to generate a pHISLonexpression plasmid. Vector pHIS1 expresses heterologous proteins bearing His6

in both the N and C termini. Plasmids pHISMdj1 and pHISLon were used totransform Escherichia coli DH5� for maintenance and amplification, where pos-itive clones were selected for resistance to ampicillin. Correct in-frame ligationwas checked by endonuclease restriction and sequencing. Selected plasmids wereinserted into E. coli BL21/pLysS (Novagen) for IPTG (isopropyl-�-D-thiogalacto-pyranoside)-induced protein expression. For purification of recombinantPbMdj1 (rPbMdj) and rPbLon, individual bacterial clones were cultivated over-night at 37°C under shaking in LB medium containing 100 �g/ml of ampicillinand 37 �g/ml of chloramphenicol. An aliquot of this initial culture was diluted1:100 in fresh medium and grown until it reached an A600 of 0.6. At this point,

protein expression was induced for 3 h with 0.5 M IPTG. Bacterial cell pelletsobtained by centrifugation (10,000 � g, 5 min) were suspended in 50 mMTris-HCl (pH 7.0)–0.2 M NaCl–10% sucrose–1 mM phenylmethylsulfonyl fluo-ride and disrupted by sonication for 5 min (15-s pulses alternated with 15 s inice). Precipitates were pelleted by centrifugation (10,000 � g, 5 min), suspendedin 100 mM NaH2PO4–10 mM Tris-HCl, pH 8.0, containing 8 M urea, andcentrifuged, and the supernatant was chromatographed in Ni-nitrilotriacetic acid(NTA) columns (QIAGEN), according to the manufacturer’s instructions.rPbMdj1 and rPbLon fractions eluted at pH 4.5 were pooled for further use.Fractions eluted at pH 6.3 and 5.9 were discarded.

Rabbit immunization, antibody purification, and Western immunoblotting.Approximately 300 �g of Ni-NTA-eluted rPbMdj1 and rPbLon were fraction-ated in preparative SDS-polyacrylamide gel electrophoresis (PAGE) gels (21)and stained briefly with Coomassie blue. The correspondent bands were cut offthe polyacrylamide gels, homogenized in PBS, and injected subcutaneously atseveral spots on the shaved backs of rabbits. The same procedure was followed40 days later; 25 days after this booster, the rabbits were bled and the serum wasseparated, tested, and aliquoted at �20°C. The rabbits were also bled beforeimmunization to obtain preimmune sera. Polyclonal rabbit immunoglobulin G(IgG) anti-rPbMdj1 and anti-rPbLon were purified from both preimmune andimmune sera in protein A-Sepharose (Pharmacia), checked for purity in SDS-PAGE gels, pooled, dialyzed in PBS, aliquoted, and frozen at �20°C. The IgGconcentration was estimated by spectrometry at A280. For purification of affinity-purified antibodies, rPbMdj or rPbLon eluted from Ni-NTA columns (150 �g, asestimated in SDS-PAGE gels) were immobilized onto nitrocellulose membranes(0.45 �m) (Amersham) over an area of 12 cm2, which was then quenched with5% skim milk in PBS (overnight at 4°C) and rinsed with PBS for 10 min undershaking. The membranes were incubated for 1 h in ice with total immune sera(1 ml) and/or immune IgG diluted 1:1 (vol/vol) in PBS. The membranes were

FIG. 1. (A) Schematic representation of PbMDJ1, highlighting the localization of the primers and probes used in this work. Introns werelocalized from nt 205 to 313 and nt 1552 to 1683. (B) Schematic representation of PbMdj1 showing the localization of conserved domains.Kyte-Doolittle hydrophilicity plots (Protean module; DNAstar Inc.) of fungal Mdj1-like sequences are as follows: Pb, P. brasiliensis (AF334811);Bd, B. dermatitidis (http://genome.wustl.edu/BLAST/blasto_client.cgi); Hc, H. capsulatum (http://genome.wustl.edu/BLAST/histo_client.cgi); Ci, C.immitis (http://www.broad.mit.edu/annotation/fungi/coccidioides_immitis); An, A. nidulans (EAA57980); Cn, C. neoformans (EAL21819); Ca, C.albicans (EAK92195); Sc, S. cerevisiae (CAA82189); Ec, Escherichia coli (BA000007). The boxes indicate the predicted mitochondrial targetsequence (MT) and the J domain. The dots localize the Zn2� finger domains. Horizontal bars indicate the location of mouse T-cell epitopes witha minimum of 12 amino acids, as predicted by the Sette major histocompatibility complex motif (Protean module; DNAstar Inc). Percentages ofidentity with PbMdj1, as calculated with the MegAlign module of the DNAstar program, are given at right.

VOL. 5, 2006 PbMdj1 LOCALIZES TO THE MITOCHONDRIA AND CELL WALL 381

thoroughly rinsed with PBS–0.5 M NaCl, and the bound antibodies were elutedwith 50 mM glycine, pH 3.0 (500 �l), which was immediately neutralized with 1M Tris-HCl, pH 9.0. The protein profile of the eluted material in silver-stainedSDS-PAGE gels indicated the presence of a major IgG band. Preimmune seraand/or IgG fractions were processed similarly. IgG and affinity purified aliquotswere kept at �20°C.

Western immunoblot reactions with sera from hyperimmune rabbits or pa-tients with PCM were carried out for 3 h at room temperature under shaking,after which the membranes were washed six times with 0.1% Tween 20 in PBSand incubated for 1 h at room temperature with goat anti-rabbit or horseanti-human Ig conjugated to peroxidase (Sigma). The reactions were developedwith an enhanced chemiluminescence kit (Amersham Pharmacia).

Confocal analysis. P. brasiliensis yeast cells growing in aerated liquid culturesin logarithmic phase (viability, �90%) were collected, washed twice in modifiedYPD, and adjusted to a concentration of 2 � 106 to 3 � 106 viable cells/ml.Washed cells were initially incubated with a 20 nM concentration of MitoTrackerRed CMXRos probe (Molecular Probes) for 20 min at 36°C, in order to stain foractive mitochondria, and then rinsed three times with PBS and fixed in coldmethanol for 30 min in the dark. The MitoTracker Red-labeled cells wereincubated with 3% (wt/vol) BSA in PBS (blocking buffer) for 16 h at 4°C, washedthree times with PBS, and then incubated with either 100 �g/ml of IgG or 60�g/ml of affinity-purified antibodies from polyclonal rabbit anti-rPbMdj andanti-rPbLon for 4 h at room temperature. Control reactions used preimmuneIgG. The cells were then incubated for 1 h in the dark with 7.5 �g/ml offluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit IgG (JacksonImmunoResearch, West Grove, PA) in blocking buffer. Between each incubationstep with antibodies, the fungal cells were washed six times with PBS. A drop of10 �l of each suspension was mounted on glass slides with Vectashield mountingmedium (VECTOR Laboratories, Inc., Burlingame, CA), and the slides weresealed with nail polish. Labeling was analyzed using a laser scanning confocalmicroscope, LSM-510 NLO (Carl Zeiss, Jena, Germany).

Electron microscopy. P. brasiliensis yeast cells growing in logarithmic phase(viability, �90%) under aerated conditions were collected by centrifugation,washed twice with 0.1 M sodium cacodylate buffer, pH 7.2, and fixed with 2%(wt/vol) formaldehyde and 2.5% (vol/vol) glutaraldehyde in cacodylate buffer for3.5 h at room temperature under agitation (47). Yeast cells were stored in fixativeat 4°C until use. Cell pellets were solidified in 2.5% (wt/vol) agarose and slicedat a thickness of approximately 1 to 2 mm. Cell slices were postfixed with 1%potassium ferrocyanide-reduced osmium tetroxide (1%) in cacodylate buffer for1 h at room temperature under shaking (56). After two washes with ultrapurewater, the cell slices were transferred to 1% aqueous uranyl acetate for 1 h in thedark. Finally, the cell pellets were washed twice in water, dehydrated, andembedded in Spurr resin (Electron Microscopy Sciences). Ultrathin sections (70to 120 nm) were collected on Formvar/carbon-coated nickel grids, treated for 15min with saturated sodium meta-periodate, freshly prepared, and washed fivetimes with ultrapure water. The preparation was quenched with 5% (vol/vol)acetylated BSA (cBSA) (Aurion) in PBS for 1 h and incubated for 30 min witha drop of 0.1 M glycine in PBS. The grids were washed three times with PBScontaining 1% cBSA and 0.5% Tween 20 (buffer A) and incubated overnightwith 60 �g/ml of rabbit anti-rPbMdj1 IgG in 1% cBSA-PBS (buffer B) in a moistchamber at 4°C. After five washes with buffer A, the grids were incubatedwith 12-nm colloidal gold–AffiniPure goat anti-rabbit IgG (Jackson Immuno-Research) at 1:30 in buffer B for 1 h. The washing step was repeated, followedby three rinses in PBS and a fixing step with 2.5% glutaraldehyde in PBS for5 min. The grids were washed five times with ultrapure water and dried. Theultrathin sections were then contrasted with 2% uranyl acetate and lead citrate.Control grids were incubated with rabbit preimmune IgG (60 �g/ml). The sec-tions were examined in a JEOL 1200 EX-II electron microscope.

Flow cytometry (FACS) analysis. For fluorescence-activated cell sorter(FACS) analysis, P. brasiliensis yeast cells were prepared as described by Soareset al. (43), with modifications. Fungal cells were cultivated for 5 days at 36°C inmodified YPD under shaking (viability, �90%), collected by centrifugation,rinsed, suspended in PBS, and left to rest for 3 min, when clumped cells tend tosediment by gravity. The supernatant was collected, adjusted to 1 � 106 cells/ml,and fixed for 1 h at room temperature with an equal volume of 4% (vol/vol)paraformaldehyde in PBS, pH 7.2. Fixed cells were precipitated by centrifugation(1 min at 5,600 � g), rinsed three times with filtered PBS, and incubated for 30min in 150 mM NH4Cl to block free amine groups. Quenching was carried outfor 1 h at room temperature, with shaking, in PBS containing 1% BSA (blockingbuffer). The cells were then incubated overnight at 4°C with rabbit IgG anti-rPbMdj1 (200 �g/ml in blocking buffer) or with the same concentration of rabbitpreimmune IgG. Following five rinses with PBS, the fungal cells were incubatedin the dark for 1 h at room temperature in FITC–polyclonal anti-rabbit IgG

(Jackson ImmunoResearch) at 15 �g/ml (1:100) in blocking buffer. Labeled cellswere rinsed five times in PBS, suspended in the same buffer (1 ml), and analyzedin a Calibur FACS (Becton Dickinson, Mountain View, CA). A total of 10,000cells were analyzed for fluorescence at 492 to 520 nm. Unlabeled control cellswere previously analyzed for autofluorescence, relative cell size, and granularity.

Effect of anti-rPbMdj1 on in vitro growth of P. brasiliensis. The effect ofpolyclonal anti-rPbMdj1 IgG on in vitro growth of P. brasiliensis yeasts was testedin strain 18 cells growing at 36°C for 5 days (viability, �90%) in modified YPDunder agitation. The cells were collected, rinsed in culture medium, and indi-viduated by being passed through a 25-mm needle. The assay was performed asdescribed by Rodrigues et al. (37), with modifications, in a 96-well culture plateusing 100 to 500 viable yeast cells per well in modified YPD containing 0.5 mg/mlof ampicillin. The cells were incubated for 48 to 72 h at 36°C in the presence orabsence of different concentrations (25 to 200 �g/ml) of either anti-rPbMdj1 orpreimmune IgG. After this incubation period, the same concentration of anti-bodies was added again and the cultures were further incubated for 3 to 4 days.The number of CFU and the cell morphology were observed in an invertedmicroscope (Olympus CK40).

RESULTS

The PbMDJ1 ORF lies within a 1,897-bp fragment (GenBankaccession number AF334811), where two introns were local-ized (Fig. 1A). The deduced PbMdj1 protein contains 551amino acids and has a deduced molecular mass of 58.7 kDaand a basic isoelectric point of 8.9. In the sequence we canrecognize the J domain and the glycine/phenylalanine-rich andzinc finger domains (Fig. 1B), which are characteristic of typeI J-domain proteins. Although we found six potential glycosyl-ation sites (Asp-X-Ser/Thr) in PbMdj1, apparently none issubstituted with endoglycosidase H-sensitive oligosaccharidechains (not shown), suggesting that the molecule does notgo through the secretory pathway. The computer programTargetP (http://www.cbs.dtu.dk/services/TargetP/ [30]) pre-dicted a mitochondrial targeting peptide within the 28 firstamino acid residues (2.94 kDa). A comparison among fungalMdj1 homologues indicated a high percentage of identity(85%) with sequences from the dimorphs B. dermatitidis andH. capsulatum, which can be visualized by the similarities in thehydrophilicity profiles (Fig. 1B). The J domain is the mostconserved region among the Mdj1 homologues analyzed. Itcontains four characteristic helixes (reviewed in reference 20)and a highly conserved HPD tripeptide between helixes II andIII, which is essential for the DnaJ/Hsp interaction (35). In theC-terminal half of PbMdj1 there is one predicted T-cell mouseepitope with a minimum of 12 residues (LYTAQIPLTTALL)between amino acids 379 and 391 (Fig. 1B), which is conservedin the homologues from B. dermatitidis and H. capsulatum.

We observed an extremely conserved gene organization ofMDJ1 and LON among members of the Eurotiomycetes family,i.e., P. brasiliensis, B. dermatitidis, H. capsulatum, C. immitis, A.nidulans, and A. fumigatus. The number and position of theintrons are conserved, although their sizes and sequences maydiffer (Fig. 2). MDJ1 from Neurospora crassa and Fusariumgraminearum have the same intron numbers and positions as inP. brasiliensis, while their LON gene has only one 3� intron. InCryptococcus neoformans, both genes bear several small in-trons. The most interesting finding of our comparative analysiswas the fact that the whole MDJ1/LON locus is conserved inEurotiomycetes species, where the genes are adjacent, inverselyoriented, and separated by a common 5� region ranging be-tween 400 and 485 nt (Fig. 2). In these species, a BroA (Bro1)gene homologue is found adjacent to MDJ1, in the same di-

382 BATISTA ET AL. EUKARYOT. CELL

rection. In S. cerevisiae, BroA encodes a cytoplasmic proteininvolved in the sorting of membrane proteins into the lumenalvesicles of endosomal multivesicular bodies (34).

We expressed a His-tagged N-terminal region of PbMdj1(252 amino acids, 25.8 kDa), which included the entire J do-main, the Gly/Phe-rich region, and two zinc finger domains.This fragment contains a number of hydrophilic regions (Fig.1B) and antibody epitopes, as predicted by the Jameson-Wolfantigenic index of the Protean module of the Lasergene pro-gram (DNAstar Inc.). We additionally expressed a 282-amino-acid His-tagged N-terminal fragment of PbLon (31.7 kDa, be-tween residues 179 and 463), which also contains several peaksof predicted antibody epitopes. Both truncated recombinantproteins (rPbMdj1 and rPbLon) were expressed as major in-soluble cytoplasmic proteins of calculated molecular masses ofapproximately 31 kDa for rPbMdj1 and 36 kDa for PbLon,including their vector sequences. These values are compatiblewith their SDS-PAGE mobilities (Fig. 3A). The correspondentpH 4.5-eluted SDS-PAGE bands were cut off the gels to im-munize rabbits. Anti-rPbMdj1 and anti-rPbLon rabbit immunesera had high immunoblot titers, of 1:8,000, when tested withsmall amounts of the correspondent recombinant proteins, asseen in Fig. 3B. We therefore used these antibodies in subcellularlocalization experiments.

Immunoblot reactions were carried out with both cell lysatesand total mitochondrial proteins from P. brasiliensis yeast cellsgrowing at 36°C or heat shocked for 60 min at 42°C (Fig. 3C).Anti-rPbMdj1 immune serum specifically reacted with a singlemitochondrial component of approximately 55 kDa (Fig. 3D),which is comparable to the molecular mass predicted for theprocessed protein lacking the mitochondrial targeting region.Expression of this component increased in heat-shocked cells,demonstrating that the PbMDJ1 gene encodes a heat shockprotein in P. brasiliensis. A single 55-kDa band was also seenwhen total mitochondrial extracts were captured with immo-

bilized anti-PbMdj1 IgG and then tested by immunoblotting(not shown). When the same strategy was applied with whole-cell extracts, the reaction was negative (not shown), showingthat these antibodies were not cross-reacting with other intra-cellular components. The anti-rPbLon immune serum recog-nized a mitochondrial component of approximately 110 kDa,whose expression also increased in heat-shocked cells, suggest-ing that PbLON encodes a high-molecular-weight heat shockmitochondrial proteinase. The estimated molecular mass cor-responds well to the PbLon monomer (117.6 kDa) deprived ofthe 6.1-kDa mitochondrial targeting region of 54 amino acids.In nonreducing SDS-PAGE gels (not shown), anti-rPbLon re-acted with a mitochondrial component migrating slowly nearthe gel top, which is in accordance with the polymeric nature ofthe Lon-like proteases. The migration of the component rec-ognized by anti-rPbMdj1 was not changed in the absence of areducing agent (not shown). Neither PbLon nor PbMdj1 couldbe detected when 20 mg of total cell extracts was tested, pos-sibly because of their small relative concentration in the cyto-plasm. The heat shock nature of the mitochondrial proteinswas confirmed at the transcription level. PbMDJ1 and PbLONRNA bands could be seen in Northern blots with total RNAextracted from P. brasiliensis yeast cells only after heat shock of30 or 60 min at 42°C, but not with genetic material isolatedfrom the starting culture growing at 36°C (Fig. 3E). The esti-mated sizes of the transcripts (2.05 kb for PbMDJ1 and 3.4 kbfor PbLON) are in agreement with the correspondent genesand protein monomers labeled with anti-PbMdj1 and anti-PbLon (Fig. 3D).

Using confocal microscopy of P. brasiliensis yeast cells grow-ing under aerobic conditions, we observed that anti-rPbMdj1and anti-rPbLon antibodies reacted with the correspondentproteins in the mitochondria, as suggested by colocalizationwith MitoTracker Red (Fig. 4). Both antibodies (green) andMitoTracker (red) stained the cytoplasmic compartment with a

FIG. 2. Phylogenetic tree of fungal Mdj1 sequences according to the Clustal W program. Besides the fungal sequences indicated in the legendto Fig. 1, we included those of A. fumigatus (EAL93469), N. crassa (EAA32959), F. graminearum (EAA76137), S. pombe (CAB09769), C. glabrata(CAG60838), and Homo sapiens (AF061749). A phylogenetic tree of the correspondent Lon sequences showed similar distributions. On the right,schematic representations of the chromosomal organization and gene direction in the MDJ1/LON locus of the species that bear these genes in thesame chromosome. The distances separating the genes (interrupted line) in N. crassa, F. graminearum, and C. glabrata are, respectively, 236,960kb, 1,982,719 kb, and 19,091 kb.

VOL. 5, 2006 PbMdj1 LOCALIZES TO THE MITOCHONDRIA AND CELL WALL 383

punctuated pattern, which merged in tones of yellow and or-ange. Surprisingly, however, anti-rPbMdj1 antibodies also la-beled the whole cell surface, where patches of stronger greenstaining could be seen at the budding sites (Fig. 4A). This isparticularly evident in the images of Fig. 4D, where the budneck is intensely bright. The longitudinal slice of this imagefacilitated a view of a large fluorescent green surface on thedaughter cell. Control images obtained with preimmune anti-bodies showed only background fluorescence (Fig. 4C). Wetried to label S. cerevisiae, C. albicans, and C. neoformans withanti-PbMdj1 immune serum, but the reactions were negative,

as were immunoblotting reactions with S. cerevisiae cell ex-tracts (not shown).

Details of the cellular localization of PbMdj1 were furtherexploited by transmission electron microscopy using immuno-gold labeling (Fig. 5). Our preparations had preserved cellsurface and organelle structures (Fig. 5A), and the images showintense labeling with anti-PbMdj1 antibodies of the double-layered cell wall. A large number of mitochondria are seensurrounding the cell membrane (Fig. 5A) yet are not in contactwith it (Fig. 5B and C). The immunogold particles seem to beclustered inside the mitochondria, but in Fig. 5C they can also

FIG. 3. (A) Coomassie blue-stained 10% SDS-PAGE gels showing insoluble total bacterial extracts from recombinant bacteria expressingPbMdj1 or PbLon (extract) and the respective proteins eluted from Ni-NTA columns with pH 4.5 buffer. (B) Titration of anti-rPbMdj1 andanti-rPbLon rabbit immune sera (1:500 to 1:16,000) in immunoblotting. As a substrate, 100 ng of rPbMdj1 (left) or rPbLon (right), as shown inpanel A, was used. NC, negative control with preimmune serum. The same lack of reactivity was observed when testing insoluble bacterialcomponents from recombinant bacteria with vector alone eluted at pH 4.5 from a Ni-NTA column. (C) Coomassie blue-stained 10% SDS-PAGEgel profiles of total cell (T) and mitochondrial (M) extracts (20 �g per lane) from a P. brasiliensis yeast culture growing at 36°C before and afterheat shock at 42°C for 60 min. (D) Western blot of the gels shown in panel B, probed with the indicated rabbit sera at a concentration of 1:1,500.SDS-PAGE molecular markers are indicated (in kilodaltons) in panels A through D. (E) Northern blot analysis of total RNA isolated from P.brasiliensis cells growing at 36°C before and after heat shock at 42°C for 30 and 60 min, using PbMDJ1 and PbLON probes (upper panels). Theband sizes are indicated (in kilobases). Lower panels show ethidium bromide-stained agarose gel replicas of the blotted gels in which the 18S and28S rRNA components are visible.

384 BATISTA ET AL. EUKARYOT. CELL

be seen near/at the mitochondrial surface and within the thinspace between the mitochondrial surface and the cell mem-brane. Mitochondria that were not near the cell wall were lesslabeled, and the gold particles in the cytoplasm (far from thecell wall) were scarce. The bud neck in Fig. 5D has a greatconcentration of gold particles, in accordance with the intensefluorescence of this region seen in Fig. 4D. Nonspecific label-ing with preimmune rabbit antibodies was reduced to scatteredparticles in the cytoplasm and the cell wall (Fig. 5E).

Surface labeling of P. brasiliensis yeast cells was also sug-gested by FACS. We observed a typical dose-response curveupon incubation of paraformaldehyde-fixed cells with 50, 100,or 200 �g/ml of anti-rPbMdj1 IgG. At a concentration of 200�g/ml, over 50% of the cells were labeled, against 7% ofpreimmune IgG (Fig. 6A). In order to characterize the cell wallcomponent that was reacting with anti-rPbMdj1 antibodies,cell wall components were released under mild conditions with�-mercaptoethanol at an alkaline pH (Fig. 6B). When thisextract was assayed by immunoblotting with anti-rPbMdj1, a

55-kDa component was revealed, which was comparable in sizeto the mitochondrial band recognized by the same serum (Fig.6B). This result adds evidence to the dual mitochondrial andcell wall localization of PbMdj1 in P. brasiliensis. It is of notethat we have not detected measurable amounts of PbMdj1 inculture supernatants.

Since the structure of PbMdj1 predicts several antibodyepitopes and one potential T-cell epitope, we questionedwhether paracoccidioidomycosis patients produced antibodiesrecognizing the molecule and whether anti-rPbMdj1 would beable to interfere with fungal growth. We addressed the firstquestion by testing the reactivity of rPbMdj1 with three pa-tients’ sera by immunoblotting. As seen in Fig. 7, two of themrecognized rPbMdj1 at a 1:200 dilution. The effect of poly-clonal anti-rPbMdj1 IgG on in vitro growth of P. brasiliensisyeasts was tested in strain 18 cells, as described in Materialsand Methods. Under these conditions, there was no inhibitoryeffect on the growth pattern or CFU counts compared withcontrols (Table 1).

FIG. 4. Localization of PbMdj1 and PbLon by confocal microscopy of P. brasiliensis yeast cells. Red images are active mitochondria stained withMitoTracker Red. Green images are reactions with affinity-purified anti-rPbMdj1 or anti-rPbLon conjugated with FITC. Control reactions werecarried out with preimmune IgG (panel C, same result for both preimmune sera). Merged images show colocalization of the labels (yellow-orange).Panel D shows two inner sections of budding cells reacting only with anti-rPbMdj1 (lower panel) or as a merged image (upper panel). Similarresults were obtained with anti-rPbMdj1 IgG or affinity-purified antibodies, while only affinity-purified anti-PbLon showed interpretable images.

VOL. 5, 2006 PbMdj1 LOCALIZES TO THE MITOCHONDRIA AND CELL WALL 385

DISCUSSION

The present work demonstrates the presence of a mitochon-drial member of the J-domain protein family, Mdj1, not only inthe mitochondria but also in the cell wall of the pathogenicdimorphic fungus P. brasiliensis. To our knowledge, this is thefirst description of a DnaJ member on the cell surface. Theunexpected cell wall localization of PbMdj1 was visualized inthe fungal yeast cells using both confocal and electron micros-copy and was confirmed by FACS analyses. The protein waslabeled with rabbit polyclonal antibodies (IgG and affinity-purified fractions) raised against a recombinant PbMdj1 con-taining the N-terminal half of the protein. Evidence for the spec-ificity of the immunological reactions with PbMdj1 comes from

the following observations: (i) the anti-rPbMdj1 antibodies spe-cifically recognized in immunoblotting, before or after anti-body capture of soluble proteins, a single 55-kDa componentfrom total P. brasiliensis mitochondrial extracts; (ii) the anti-rPbMdj1 antibodies specifically recognized a single 55-kDaimmunoblotted component from a cell wall extract obtainedfrom yeast cells of P. brasiliensis with mildly alkaline �-mer-captoethanol treatment; and (iii) confocal reactions carriedout with affinity-purified antibodies resulted in a labeling pat-tern of both the cell wall and mitochondria.

Although the presence of surface J-domain proteins has notbeen reported before, the occurrence of extramitochondrialDnaJ homologues had been predicted by Soltys and Gupta

FIG. 5. Ultrastructural localization of PbMdj1 in P. brasiliensis yeast cells incubated with IgG fractions (60 �g/ml) of either anti-rPbMdj1(A through D) or preimmune control (E). Each panel shows a different cell. (A) Panoramic photomicrograph. (B and C) Details. Note labelinginside the mitochondria (M, arrows) and in the cell wall (CW). In panel C, labeling is also visible in the region between the cell membrane andthe mitochondrion. (D) The bud neck is intensely labeled.

386 BATISTA ET AL. EUKARYOT. CELL

(45). The presence of discrete amounts of mitochondriallysorted Hsp70 and Hsp60 on the plasma membrane, cytoplas-mic vesicles, and cytoplasmic granules of mammalian cells hasbeen well documented (42, 44; reviewed in reference 45).

Hsp60 has been detected in small clusters at discrete points onthe H. capsulatum cell wall, and it has been shown to mediateattachment of the fungus to macrophages via CD11/CD18 re-ceptors (24). Both Hsp70 and Hsp60 need specific cochaper-ones for optimum chaperone function, which is thus likely tooccur extramitochondrially. Therefore, outside the mitochon-dria Mdj1 could be assisting Ssc1 (mitochondrial Hsp70), but itcould also be functioning as an independent chaperone. Al-though an Ssc1 homologue has not been described in fungalcell walls, an Hsp70-like protein has been detected at the outersurface of the plasma membrane, throughout the cell wall, andat the cell surface of C. albicans (25). It will be interesting tofind out if the P. brasiliensis Ssc1, whose gene homologue hasbeen identified in the database, will also be found in the cellwall and if it colocalizes with Mdj1.

The fungal cell wall is a dynamic compartment that deter-mines and maintains cell shape but is also actively involved inmany other biological events. Although glucans and chitin arethe main scaffold components of the cell wall, it has a fasci-nating, complex structure with a number of other constituents(glycoproteins, proteins, and lipids), such as enzymes, ad-hesins, and structural and heat shock proteins, whose featuresdepend on multiple intrinsic and external factors (reviewed forC. albicans in reference 10). Some of these components frompathogenic fungi are promising targets for drugs and immuno-therapy (31). The interactions among these molecules seem toinvolve not only hydrogen and hydrophobic bonds but alsocovalent and phosphodiester (mediated by glycosyl phosphati-dylinositol) linkages with polysaccharides. In the present study,we detected abundant labeling of the cell wall with anti-rPbMdj1 antibodies; however, low yields of the reactive 55-kDa protein were obtained upon alkaline extraction with�-mercaptoethanol, suggesting that the molecule is not looselybound to this compartment. An additional observation wasthat when yeast cells were first treated with �-mercaptoethanoland then labeled for FACS analysis, the percentage of fluores-cent cells increased instead of going down to background lev-

TABLE 1. Lack of inhibition of P. brasiliensis growth byanti-rPbMdj1 IgG

Exptl condition(�g/ml)a No. of CFUb

Culture medium ....................................................................... 37 5.2

Anti-rPbMdj1200..........................................................................................38.33 4.7100..........................................................................................46.75 9.050............................................................................................37.75 2.725............................................................................................ 38 0.812.5......................................................................................... 39 0.8

Preimmune serum200.......................................................................................... 46.5 3.5100..........................................................................................41.75 4.950............................................................................................42.25 5.725............................................................................................43.25 3.112.5......................................................................................... 41.6 5.8

a Yeast cells (102) were cultivated in the presence of 200, 100, 50, 25, or 12.5�g/ml of IgG. After incubation for 48 or 72 h at 36°C, the cultures were supple-mented with the same amount of IgG.

b Numbers of CFU were determined on day 6 of culture. Differences innumbers of CFU were not statistically significant (P � 0.05).

FIG. 6. (A) FACS analysis of P. brasiliensis yeast cells incubatedwith 200 �g/ml of either anti-rPbMdj1 or preimmune IgG. (B) Coo-massie blue-stained SDS-PAGE profile of an alkaline �-mercaptoetha-nol cell wall extract (�-ME, 40 �g), which was also Western blottedand assayed with anti-rPbMdj1, anti-PbLon, and preimmune (PI) IgGat a dilution of 1:1,500. A mitochondrial extract (Mito) was run inparallel and assayed with anti-rPbMdj1. Molecular masses are indi-cated (in kilodaltons).

FIG. 7. Immunoblot reactivity of three PCM patients’ sera withrPbMdj1 (note visible bands for patients 1 and 2). Controls were serumfrom a healthy individual (�) and anti-rPbMdj1 (� [at 1:16,000]).Human sera were tested at a 1:200 dilution. The PCM sera reactedwith gp43 with immunodiffusion titers of 1:64 and 1:32.

VOL. 5, 2006 PbMdj1 LOCALIZES TO THE MITOCHONDRIA AND CELL WALL 387

els, as would be expected if PbMdj1 had been totally extracted(not shown). Therefore, it is likely that the initial peelingcaused by mild alkaline and �-mercaptoethanol treatment im-proved the access of anti-rPbMdj1 antibodies to formerly in-accessible PbMdj1 target molecules. That might also explainwhy, although PbMdj1 seems to be actively participating inyeast cell budding, anti-rPbMdj1 antibodies could not interferewith fungal growth in vitro within the range of concentrationstested: other cell components might be blocking antibody ac-cess to their target molecules. An example of such a blockagemechanism has been recently detected in Fonsecaea pedrosoi,for which anti-monohexosylceramide can react with the cellwall only when melanin is absent (32). Another possible expla-nation is that the antibodies cannot recognize PbMdj1 well ingrowing cells. On the other hand, genetic manipulation in P.brasiliensis is still an obstacle to research of P. brasiliensis; untilwe have mutants for both Lon and Mdj1, one can only specu-late about their role in the fungal cell wall. Another aspect tobe investigated is the functionality of the potential T-cell re-ceptor that is conserved among dimorphs.

A Southern blot of P. brasiliensis DNA restricted with endo-nucleases and assayed with an E5/E4 PbMDJ1 probe (Fig. 1A)resulted in a pattern of single bands (not shown), which sug-gests the presence of a single copy of the PbMDJ1 gene in thegenome, as seen before with the adjacent PbLON gene (2).This is supported by the previous observations that PbMDJ1and PbLON were mapped to a unique chromosomal band in 12individual P. brasiliensis isolates (13). In other fungi analyzed herefor which the genome is completely sequenced (Aspergillus andCandida), both LON and MDJ1 are present in single copies,reinforcing those observations. We have not found any evi-dence for the occurrence of alternative splicing in PbMDJ1: wevisualized only one PbMDJ1 RNA band in Northern blots.Besides, RT-PCRs using primers designed to elongate thefull-length ORF consistently resulted in a single band of thesame size when we tested different RNA preparations of yeastcells either growing at 36°C or heat shocked at 42°C (notshown). Therefore, the components recognized by anti-rPbMDJ1both in mitochondria and in the cell wall apparently derivefrom the same gene and are not a product of differentiallysorted molecules bearing particular targeting sequences dueto alternative splicing. The gene encoding alanine-glyoxylateaminotransferase from amphibian livers has two alternativetranscription start sites at either side of the 24-nt-long mito-chondrial targeting signal, so that the protein encoded by thelong transcript localizes to the mitochondria, while the productof the shorter transcript is cytosolic (16). We have not beenable to determine the transcription initiation site of PbMDJ1;however, computer analysis does not predict any internaltranslation start site that would result in an ORF lacking themitochondrial matrix-targeting signal. Moreover, the similarSDS-PAGE gel migration of the protein labeled in mitochon-dria and cell wall extracts with anti-rPbMdj1 suggests that themolecule was first sorted to that organelle, where it was prob-ably processed by a matrix-processing peptidase into a mature55-kDa form and then migrated to the cell wall. A matrix-processing peptidase homologue has been found in the P.brasiliensis database.

Recent years have seen a growing number of examples ofproteins that are sorted to the mitochondrial matrix and then

apparently exported to an additional compartment(s) of thecell (extensively reviewed in reference 45). The extramitochon-drial destination varies with the molecules, which include an-tigens, enzymes, receptors, hormones, and chaperones, andwith their specific roles in the destination sites. The mecha-nism(s) involved in mitochondrial export and its control are sofar speculative, but the endosymbiotic origin of mitochondriasuggests that some ancient bacterial secretory pathways musthave been retained and modified by the organelle (45). It isnoticeable that in a recent proteomics study of plasma mem-brane lipid rafts, 24% of the 196 identified proteins weremitochondrial; however, rafts have not been found in thatcompartment (1). In our electron microscopic images, mito-chondria formed a ring near the cell surface. This pattern hasbeen mentioned previously for P. brasiliensis yeast cells (19),and they can be seen in electron microscopic pictures of re-cently transformed yeasts (7), although in hyphae that patternwas not evident (6). The image shown in Fig. 5C suggests thatsome gold particles, in small clusters, are leaving the mitochon-drion and moving toward the cell surface, but it is not clearwhether these clusters are inside vesicles. Thus, the peripherallocalization of mitochondria might not be fortuitous but rathermay suggest that active trafficking of molecules to the cellsurface might be taking place.

We have previously characterized the P. brasiliensis LONhomologue (2). Here we show that PbLON is a heat shockgene that encodes a 110-kDa mitochondrial heat shock pro-tein. In our confocal analyses, PbLon localized to the mito-chondria, showing that the presence of PbMdj1 in the cell walldoes not include helping protein degradation by PbLon. Weobserved that the deduced Lon and Mdj1 sequences are re-markably homologous among dimorphic fungi, especially P.brasiliensis, H. capsulatum, and B. dermatitidis. It will be inter-esting to find out if Mdj1 localizes to the cell wall of thesemicroorganisms. We found that the LON/MDJ1 locus organi-zation is comparable among Eurotiomycetes species. The entirehomologous locus probably also includes Bro1, CreA, a littlefarther from BroA, and a hypothetical protein adjacent to LONso far found in A. nidulans, A. fumigatus, and H. capsulatum. Interms of comparative genomics, this is a relevant finding thatmight have evolutionary significance. This is the first conservedchromosomal organization reported for Ascomycetes, for whichgenome sequencing of several species will soon be finished. Weare presently studying the regulatory elements contained in the5� untranslated region shared by LON and MDJ1, and we hopeto find conserved mechanisms of regulation among fungal heatshock proteins.

ACKNOWLEDGMENTS

We are thankful to Caroline Z. Romera, Elizabeth N. Kanashiro,Marcia F. A. Tanakai, Andre H. Aguillera, and Ronni R. Novaes eBrito for generous and competent technical assistance. We are gratefulto Soraya S. Smaili for the use of confocal microscopy facilities. We arealso grateful to Marcio L. Rodrigues and Gustavo H. Goldman forcritical reading of the manuscript and to Luiz R. Travassos for alwaysreminding us that the cell wall is a fascinating dynamic compartmentwhere one can find any kind of molecule.

This work was supported by FAPESP, CNPq, and Pronex/CNPq.

REFERENCES

1. Bae, T. J., M. S. Kim, J. W. Kim, B. W. Kim, H. J. Choo, J. W. Lee, K. B. Kim,C. S. Lee, J. H. Kim, S. Y. Chang, C. Y. Kang, S. W. Lee, and Y. G. Ko. 2004.

388 BATISTA ET AL. EUKARYOT. CELL

Lipid raft proteome reveals ATP synthase complex in the cell surface. Pro-teomics 4:3536–3548.

2. Barros, T. F., and R. Puccia. 2001. Cloning and characterization of a LONgene homologue from the human pathogen Paracoccidioides brasiliensis.Yeast 18:981–988.

3. Bialek, R., A. Ibricevic, A. Fothergill, and D. Begerow. 2000. Small subunitribosomal DNA sequence shows Paracoccidioides brasiliensis closely relatedto Blastomyces dermatitidis. J. Clin. Microbiol. 38:3190–3193.

4. Bota, D. A., J. K. Ngo, and K. J. Davies. 2005. Downregulation of the humanLon protease impairs mitochondrial structure and function and causes celldeath. Free Radic. Biol. Med. 38:665–677.

5. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation ofmicrogram quantities of protein utilizing the principle of protein-dye bind-ing. Anal. Biochem. 72:248–254.

6. Carbonell, L. M., and J. Rodriguez. 1968. Mycelial phase of Paracoccidioidesbrasiliensis and Blastomyces dermatitidis: an electron microscope study. J.Bacteriol. 96:533–543.

7. Carbonell, L. M. 1969. Ultrastructure of dimorphic transformation in Para-coccidioides brasiliensis. J. Bacteriol. 100:1076–1082.

8. Carvalho, K., L. Ganiko, W. L. Batista, F. V. Morais, E. Marques, G. H.Goldman, M. F. Franco, and R. Puccia. 2005. Virulence of Paracoccidioidesbrasiliensis and gp43 expression in isolates bearing known PbGP43 geno-types. Microbes Infect. 7:55–65.

9. Casanova, M., and W. L. Chaffin. 1991. Cell wall glycoproteins of Candidaalbicans as released by different methods. J. Gen. Microbiol. 137:1045–1051.

10. Chaffin, W. L., J. L. Lopez-Ribot, M. Casanova, D. Gozalbo, and J. P.Martinez. 1998. Cell wall and secreted proteins of Candida albicans: identi-fication, function, and expression. Microbiol. Mol. Biol. Rev. 62:130–180.

11. Cisalpino, P. S., R. Puccia, L. M. Yamauchi, M. I. Cano, J. F. da Silveira,and L. R. Travassos. 1996. Cloning, characterization, and epitope expressionof the major diagnostic antigen of Paracoccidioides brasiliensis. J. Biol. Chem.271:4553–4560.

12. Fan, C. Y., S. Lee, and D. M. Cyr. 2003. Mechanisms for regulation of Hsp70function by Hsp40. Cell Stress Chaperones 8:309–316.

13. Feitosa, L. S., P. S. Cisalpino, M. R. dos Santos, R. A. Mortara, T. F.Barros, F. V. Morais, R. Puccia, J. F. da Silveira, and Z. P. de Camargo.2003. Chromosomal polymorphism, syntenic relationships, and ploidy inthe pathogenic fungus Paracoccidioides brasiliensis. Fungal Genet. Biol.39:60–69.

14. Felipe, M. S., R. V. Andrade, F. B. Arraes, A. M. Nicola, A. Q. Maranhao,F. A. Torres, I. Silva-Pereira, M. J. Pocas-Fonseca, E. G. Campos, L. M.Moraes, P. A. Andrade, A. H. Tavares, S. S. Silva, C. M. Kyaw, D. P. Souza,P. Network, M. Pereira, R. S. Jesuino, E. V. Andrade, J. A. Parente, G. S.Oliveira, M. S. Barbosa, N. F. Martins, A. L. Fachin, R. S. Cardoso, G. A.Passos, N. F. Almeida, M. E. Walter, C. M. Soares, M. J. Carvalho, andM. M. Brigido. 2005. Transcriptional profiles of the human pathogenicfungus Paracoccidioides brasiliensis in mycelium and yeast cells. J. Biol.Chem. 280:24706–24714.

15. Goldman, G. H., M. E. dos Reis, R. Duarte, L. A. Souza Bernardes, A. C.Quiapin, P. M. Vitorelli, M. Savoldi, C. P. Semighini, R. C. de Oliveira, L. R.Nunes, L. R. Travassos, R. Puccia, W. L. Batista, L. E. Ferreira, J. C.Moreira, A. P. Bogossian, F. Tekaia, M. P. Nobrega, F. G. Nobrega, andM. H. Goldman. 2003. Expressed sequence tag analysis of the human patho-gen Paracoccidioides brasiliensis yeast phase: identification of putative ho-mologues of Candida albicans virulence and pathogenicity genes. Eukaryot.Cell 2:34–48.

16. Holbrook, J. D., and C. J. Danpure. 2002. Molecular basis for the dualmitochondrial and cytosolic localization of alanine:glyoxylate aminotransferasein amphibian liver cells. J. Biol. Chem. 277:2336–2344.

17. Huang, H. C., M. Y. Sherman, O. Kandror, and A. L. Goldberg. 2001. Themolecular chaperone DnaJ is required for the degradation of a solubleabnormal protein in Escherichia coli. J. Biol. Chem. 276:3920–3928.

18. Izacc, S. M., F. J. Gomez, R. S. Jesuino, C. A. Fonseca, M. S. Felipe, G. S.Deepe, and C. M. Soares. 2001. Molecular cloning, characterization andexpression of the heat shock protein 60 gene from the human pathogenicfungus Paracoccidioides brasiliensis. Med. Mycol. 39:445–455.

19. Kashino, S. S., V. L. Calich, L. M. Singer-Vermes, P. A. Abrahamsohn, andE. Burger. 1987. Growth curves, morphology and ultrastructure of ten Para-coccidioides brasiliensis isolates. Mycopathologia 99:119–128.

20. Kelley, W. L. 1998. The J-domain family and the recruitment of chaperonepower. Trends Biochem. Sci. 23:222–227.

21. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly ofthe head of bacteriophage T4. Nature 227:680–685.

22. Langer, T., C. Lu, H. Echols, J. Flanagan, M. K. Hayer, and F. U. Hartl.1992. Successive action of DnaK, DnaJ and GroEL along the pathway ofchaperone-mediated protein folding. Nature 356:683–689.

23. Leal, C. V., B. A. Montes, A. C. Mesa, A. L. Rua, M. Corredor, A. Restrepo,and J. G. McEwen. 2004. Agrobacterium tumefaciens-mediated transforma-tion of Paracoccidioides brasiliensis. Med. Mycol. 42:391–395.

24. Long, K. H., F. J. Gomez, R. E. Morris, and S. L. Newman. 2003. Identifi-cation of heat shock protein 60 as the ligand on Histoplasma capsulatum that

mediates binding to CD18 receptors on human macrophages. J. Immunol.170:487–494.

25. Lopez-Ribot, J. L., H. M. Alloush, B. J. Masten, and W. L. Chaffin. 1996.Evidence for presence in the cell wall of Candida albicans of a proteinrelated to the hsp70 family. Infect. Immun. 64:3333–3340.

26. Lu, B., T. Liu, J. A. Crosby, J. Thomas-Wohlever, I. Lee, and C. K. Suzuki.2003. The ATP-dependent Lon protease of Mus musculus is a DNA-bindingprotein that is functionally conserved between yeast and mammals. Gene306:45–55.

27. Marques, E. R., M. E. Ferreira, R. D. Drummond, J. M. Felix, M. Menossi,M. Savoldi, L. R. Travassos, R. Puccia, W. L. Batista, K. C. Carvalho, M. H.Goldman, and G. H. Goldman. 2004. Identification of genes preferentiallyexpressed in the pathogenic yeast phase of Paracoccidioides brasiliensis, usingsuppression subtraction hybridization and differential macroarray analysis.Mol. Genet. Genomics 271:667–677.

28. McEwen, J. G., V. Bedoya, M. M. Patino, M. E. Salazar, and A. Restrepo.1987. Experimental murine paracoccidioidomycosis induced by the inhala-tion of conidia. J. Med. Vet. Mycol. 25:165–175.

29. Morais, F. V., T. F. Barros, M. K. Fukada, P. S. Cisalpino, and R. Puccia.2000. Polymorphism in the gene coding for the immunodominant antigengp43 from the pathogenic fungus Paracoccidioides brasiliensis. J. Clin. Mi-crobiol. 38:3960–3966.

30. Nielsen, H., J. Engelbrecht, S. Brunak, and G. von Heijne. 1997. Identifica-tion of prokaryotic and eukaryotic signal peptides and prediction of theircleavage sites. Protein Eng. 10:1–6.

31. Nimrichter, L., M. L. Rodrigues, E. G. Rodrigues, and L. R. Travassos. 2005.The multitude of targets for the immune system and drug therapy in thefungal cell wall. Microbes Infect. 7:789–798.

32. Nimrichter, L., M. D. Cerqueira, E. A. Leitao, M. Kildare, E. S. Nakayasu,S. R. Almeida, I. C. Almeida, C. S. Alviano, E. Barreto-Bergter, and M. L.Rodrigues. 2005. Structure, cellular distribution, antigenicity, and biologicalfunctions of Fonsecaea pedrosoi ceramide monohexosides. Infect. Immun.73:7860–7868.

33. Nunes, L. R., R. O. Oliveira, D. B. Leite, V. S. Silva, E. R. Marques, M. E. S.Ferreira, D. C. D. Ribeiro, L. A. S. Bernardes, M. H. S. Goldman, R. Puccia,L. R. Travassos, W. L. Batista, M. P. Nobrega, F. G. Nobrega, D. Y. Yang,and G. H. Goldman. 2005. Transcription profiling of Paracoccidioides bra-siliensis cells undergoing the mycelium-to-yeast transition. Eukaryot. Cell4:2115–2128.

34. Odorizzi, G., D. J. Katzmann, M. Babst, A. Audhya, and S. D. Emr. 2003.Bro1 is an endosome-associated protein that functions in the MVB pathwayin Saccharomyces cerevisiae. J. Cell Sci. 116:1893–1903.

35. Pellecchia, M., T. Szyperski, D. Wall, C. Georgopoulos, and K. Wuthrich.1996. NMR structure of the J-domain and the Gly/Phe-rich region of theEscherichia coli DnaJ chaperone. J. Mol. Biol. 260:236–250.

36. Peterson, S. W., and L. Sigler. 1998. Molecular genetic variation in Emmonsiacrescens and Emmonsia parva, etiologic agents of adiaspiromycosis, and theirphylogenetic relationship to Blastomyces dermatitidis (Ajellomyces dermatitidis)and other systemic fungal pathogens. J. Clin. Microbiol. 36:2918–2925.

37. Rodrigues, M. L., L. R. Travassos, K. R. Miranda, A. J. Franzen, S. Rozental,W. de Souza, C. S. Alviano, and E. Barreto-Bergter. 2000. Human antibodiesagainst a purified glucosylceramide from Cryptococcus neoformans inhibitcell budding and fungal growth. Infect. Immun. 68:7049–7060.

38. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: alaboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, ColdSpring Harbor, N.Y.

39. Savel’ev, A. S., L. A. Novikova, I. E. Kovaleva, V. N. Luzikov, W. Neupert,and T. Langer. 1998. ATP-dependent proteolysis in mitochondria. m-AAAprotease and PIM1 protease exert overlapping substrate specificities andcooperate with the mtHsp70 system. J. Biol. Chem. 273:20596–20602.

40. Sheffield, P., S. Garrard, and Z. Derewenda. 1999. Overcoming expressionand purification problems of RhoGDI using a family of “parallel” expressionvectors. Protein Expr. Purif. 15:34–39.

41. Silva, S. P., M. I. Borges-Walmsley, I. S. Pereira, C. M. Soares, A. R.Walmsley, and M. S. Felipe. 1999. Differential expression of an hsp70 geneduring transition from the mycelial to the infective yeast form of the humanpathogenic fungus Paracoccidioides brasiliensis. Mol. Microbiol. 31:1039–1050.

42. Singh, B., B. J. Soltys, Z. C. Wu, H. V. Patel, K. B. Freeman, and R. S. Gupta.1997. Cloning and some novel characteristics of mitochondrial Hsp70 fromChinese hamster cells. Exp. Cell Res. 234:205–216.

43. Soares, R. M., Costa e Silva-Filho, S. Rozental, J. Angluster, W. de Souza,C. S. Alviano, and L. R. Travassos. 1998. Anionogenic groups and surfacesialoglycoconjugate structures of yeast forms of the human pathogen Para-coccidioides brasiliensis. Microbiology 144:309–314.

44. Soltys, B. J., and R. S. Gupta. 1996. Immunoelectron microscopic localiza-tion of the 60-kDa heat shock chaperonin protein (Hsp60) in mammaliancells. Exp. Cell Res. 222:16–27.

45. Soltys, B. J., and R. S. Gupta. 2000. Mitochondrial proteins at unexpectedcellular locations: export of proteins from mitochondria from an evolution-ary perspective. Int. Rev. Cytol. 194:133–196.

VOL. 5, 2006 PbMdj1 LOCALIZES TO THE MITOCHONDRIA AND CELL WALL 389

46. Stahlberg, H., E. Kutejova, K. Suda, B. Wolpensinger, A. Lustig, G. Schatz,A. Engel, and C. K. Suzuki. 1999. Mitochondrial Lon of Saccharomycescerevisiae is a ring-shaped protease with seven flexible subunits. Proc. Natl.Acad. Sci. USA 96:6787–6790.

47. Straus, A. H., E. Freymuller, L. R. Travassos, and H. K. Takahashi. 1996.Immunochemical and subcellular localization of the 43 kDa glycoproteinantigen of Paracoccidioides brasiliensis with monoclonal antibodies. J. Med.Vet. Mycol. 34:181–186.

48. Suzuki, C. K., E. Kutejova, and K. Suda. 1995. Analysis and purification ofATP-dependent mitochondrial lon protease of Saccharomyces cerevisiae.Methods Enzymol. 260:486–494.

49. Szabo, A., R. Korszun, F. U. Hartl, and J. Flanagan. 1996. A zinc finger-likedomain of the molecular chaperone DnaJ is involved in binding to denaturedprotein substrates. EMBO J. 15:408–417.

50. Tang, W., and C. C. Wang. 2001. Zinc fingers and thiol-disulfide oxidoreduc-tase activities of chaperone DnaJ. Biochemistry 40:14985–14994.

51. van Dyck, L., and T. Langer. 1999. ATP-dependent proteases controllingmitochondrial function in the yeast Saccharomyces cerevisiae. Cell. Mol. LifeSci. 56:825–842.

52. Voos, W., and K. Rottgers. 2002. Molecular chaperones as essential media-tors of mitochondrial biogenesis. Biochim. Biophys. Acta 1592:51–62.

53. Wagner, I., H. Arlt, L. van Dyck, T. Langer, and W. Neupert. 1994. Molec-ular chaperones cooperate with PIM1 protease in the degradation of mis-folded proteins in mitochondria. EMBO J. 13:5135–5145.

54. Wagner, I., L. van Dyck, A. S. Savel’ev, W. Neupert, and T. Langer. 1997.Autocatalytic processing of the ATP-dependent PIM1 protease: crucial func-tion of a pro-region for sorting to mitochondria. EMBO J. 16:7317–7325.

55. Walsh, P., D. Bursac, Y. C. Law, D. Cyr, and T. Lithgow. 2004. The J-proteinfamily: modulating protein assembly, disassembly and translocation. EMBOJ. 5:567–571.

56. Wright, R., M. Basson, L. D’Ari, and J. Rine. 1988. Increased amounts ofHMG-CoA reductase induce “karmellae”: a proliferation of stacked mem-brane pairs surrounding the yeast nucleus. J. Cell Biol. 107:101–114.

390 BATISTA ET AL. EUKARYOT. CELL

39

Proteinases de P. brasiliensis: conclusões

A serino-tiol proteinase extracelular de P. brasiliensis (PbST) é a enzima do

fungo melhor caracterizada até o momento. O desenvolvimento desta tese ampliou o

conhecimento sobre essa PbST com os seguintes dados:

1. A atividade da PbST sobre o peptídeo Abz-MKALTLQ-EDDnp sofre uma intensa

modulação por açúcares neutros do tipo dextrana, manana de levedura e galactomanana

de P. brasiliensis. A manana de levedura inibiu a interação da proteinase com o substrato,

enquanto a dextrana e a galactomanana de P. brasiliensis aumentaram a afinidade da

enzima pelo substrato. Adicionalmente, a galactomanana fúngica foi capaz de estabilizar

a atividade hidrolítica em temperaturas elevadas e diminuir a Vmáx, o que provavelmente

previne sua autólise antes de atingir o substrato in vivo (artigo 1).

2. Foram desenvolvidos compostos com intenso poder inibitório da PbST sobre Abz-

MKALTLQ-EDDnp e em ensaios de inibição da aderência celular de LLC-MK2. O

inibidores foram gerados a partir de um substrato peptídico da enzima (MKRLTL)

acoplado com o composto (C)Npys, o qual é capaz de reagir com grupos tiois livre. Bzl-

C(Npys)KRLTL-NH2 apresentou um Ki de 16 nM, Bzl-MKRLTLC(Npys)-NH2 alcançou

160 nM, enquanto (C)Npys inibiu na ordem de 1,7 µM. Os inibidores não foram tóxicos

em ensaios com culturas de células, o que oferece possibilidade de utilização in vivo. A

inibição foi completamente revertida com ditiotreitol, mostrando ser devida à interação

do (C)Npys com um grupo tiol livre posicionado próximo ao sítio ativo, como sugerido

pelo aumento da eficiência de inibição com substrato peptídico (artigo 2).

40

3. Apesar do conhecimento acumulado sobre as características enzimáticas da PbST, sua

identidade continua um mistério. Sua purificação total a partir de sobrenadante de cultura

parece inviável. Utilizamos várias estratégias com o intuito de enriquecer a proteinase e

realizar análises de seqüência por espectrometria de massas ou sequenciamento de N-

terminal. Todavia, nenhuma informação conclusiva foi obtida (artigo 2 e anexo 1).

Talvez os inúmeros peptídeos seqüenciados possam ser identificados em futuro breve,

visto que a análise do genoma completo do Paracoccidioides brasiliensis está em

andamento.

O gene da serino proteinase mitocondrial ATP-dependente Lon de P. brasiliensis

(PbST) foi clonado fortuitamente durante as tentativas de clonar a PbST usando

oligonucleotídeos degenerados correspondentes aos sítios ativos de subtilisinas. Sua

caracterização levou à identificação, no mesmo lócus, do gene MDJ1, codante de uma

chaperone mitocondrial que potencialmente participa na seleção de substratos da Lon. O

desenvolvimento desta tese ampliou o conhecimento sobre a PbLon com os seguintes

dados:

1. A PbLon localiza-se somente na matriz da mitocôndria de leveduras do P.

brasiliensis, enquanto a PbMdj1 foi detectada em grandes quantidades também na parede

celular do fungo. A localização em microscopia confocal foi possível a partir de

anticorpos policlonais de coelho anti-PbLon recombinante truncada. A marcação da

PbLon na mitocôndria foi intensa por microscopia confocal (artigo 3) e discreta nos

cortes de microscopia eletrônica de transmissão (anexo 2).

2. O objetivo de expressar a PbLon ativa em sistema heterólogo e com ela selecionar

substratos peptídicos específicos para a enzima foi abandonado. As inúmeras tentativas

41

de expressão da proteinase recombinante ativa foram em vão (anexo 2), possivelmente

por ser tóxica à célula quando superexpressada.

42

Capítulo 2

Moléculas secretadas de Paracoccidioides

brasiliensis: proteínas associadas à parede

celular diferencialmente expressas e

resultados preliminares sobre vesículas

extracelulares

43

Introdução

A produção intracelular de pequenas vesículas contendo proteínas da membrana

plasmática foi descrita há mais de 25 anos na maturação de reticulócitos de carneiro (Pan

et al., 1983, 1985). Sabe-se que durante a maturação de reticulócito para eritrócito há o

desaparecimento de diversas proteínas da membrana plasmática e o principal mecanismo

parece estar relacionado à formação de vesículas lipidoproteicas no endossomo (Pan et

al., 1985; Harding et al., 1984), seguido de fusão com a membrana plasmática e liberação

dos exossomos na corrente sanguínea (Johnstone et al., 1987). Porém ainda não há

descrição sobre a função fisiológica desses exosomos (revisão em Johnstone, 2006). As

proteínas eliminadas pelos exossomos irão depender do tipo celular, porém o mesmo tipo

celular pode apresentar diferenças, como eritrócitos de carneiros que irão reter a maioria

dos seus transportadores de glucose, enquanto em porcos ficam retidos transportadores de

nucleosídeos. Em células de mamíferos, já foi descrita a liberação de exossomos em

células do sistema imune, epitélio intestinal, epidídimo e túbulos do rim (revisão em

Johnstone, 2006). Atualmente os exossomos são definidos como pequenas vesículas

membranosas lipido-proteicas originadas intracelularmente pelo brotamento de

membranas que formam pequenas vesículas (vesículas intraluminais), as quais se

acumulam em corpos multivesiculados. Sua liberação como exossomos ocorre a partir da

fusão da membrana contendo os corpos multivesiculados com a membrana plasmática e

seu tamanho varia entre 20 e 150 nm (revisões em Février e Raposo, 2004; Johnstone,

2006; Li et al., 2006).

Os exossomos contém citosol e expõem o domínio intracelular de receptores.

Classicamente, podem ser obtidos através de filtrações e centrifugações diferenciais de

44

sobrenadantes de cultura e são precipitados por uma centrifugação final a 100.000 x g por

1 hora. Em gradientes de sacarose, os exossomos alojam-se na densidade entre 1,13 e

1,21 g/mL e esse passo pode ser introduzido para eliminar vesículas associadas não

especificamente a proteínas e/ou lipídeos extracelulares. Sua visualização somente pode

ser feita através de microscopia eletrônica (revisão em Johnstone, 2006). Recentemente,

foi proposto que os exossomos constituem um novo mecanismo de secreção para o

exporte de Hsp70, a qual é desprovida de peptídeo sinal e tem, como outras Hsp,

propriedades imunorregulatórias (Lancaster e Febbraio, 2005).

Além de eliminar proteínas do citosol, os exossomos desempenham importantes

funções imunológicas. Exossomos de células dendríticas maduras são 2 ordens de

grandeza mais efetivos que exossomos de células dentríticas imaturas em indução de

células-T antígeno específicas (Segura et al., 2005). Verificou-se também que exossomos

de células dendríticas estimuladas com antígenos de tumor possuem uma resposta

antitumoral in vivo (Quah et al., 2005), porém esse efeito não foi específico para o tipo

de tumor. Sprent (2005) mostrou que exossomos podem ser imunogênicos para células T

CD8+ mesmo na ausência de células apresentadoras de antígeno. Recentemente

exossomos têm sido usados no tratamento de pacientes com melanoma, pois se sabe que

eles possuem um complexo funcional MHC classe II – peptídeo capaz de promover

resposta imune e rejeição de tumor (revisão em Chaput et al., 2005).

Em bactérias gram-negativas também foi observada a liberação de vesículas

extracelulares de 50 a 250 nm formadas diretamente a partir da membrana externa. A

secreção de produtos em bactérias patogênicas gram-negativas é o principal meio de

comunicação e intoxicação das células do hospedeiro. As vesículas são produzidas em

45

tecidos infectados e participam na colonização, carregando fatores de virulência e

modulando a defesa e resposta imune do hospedeiro. Também já foram detectadas

vesículas em fluidos corpóreos demonstrando sua capacidade de disseminação para

diversos pontos. Algumas bactérias utilizam vesículas para destruir outras bactérias em

situação de competição por nutrientes ou até mesmo transferir material benéfico, como

enzimas de resistência a antibióticos. Por participar de tantas funções é considerado um

potencial fator de virulência em bactérias gram-negativas (revisão em Kuehn e Kesty,

2005).

Em nosso laboratório foi especulada a possibilidade da serino-tiol proteinase

(PbST) ou algum modulador positivo da enzima ser veiculado ao meio extracelular por

vesículas secretadas pelo fungo. Em algumas oportunidades foi observado aumento na

atividade enzimática após congelamentos e degelos sucessivos de sobrenadante, o que

poderia levar à lise de estruturas lipídicas. Estruturas semelhantes aos exossomos

observadas em Trypanosoma cruzi foram capazes de provocar uma intensa resposta pró-

inflamatória de macrófagos murinos via receptor Toll-like 2 (Torrecilhas et al., em

revisão). Em fungos, vesículas extracelulares foram recentemente detectadas em

Cryptococcus neoformans (Yoneda e Doering, 2006; Rodrigues et al., 2007). Além da

parede celular, esse fungo possui uma cápsula composta polissacarídica que é essencial

para a virulência (Chang e Kwon-Chung, 1994), porém o mecanismo de tranporte dessas

macromoléculas nunca foi esclarecido. Recentemente Rodrigues e colaboradores (2007)

evidenciaram e caracterizaram vesículas extracelulares com 50 a 400 nm de diâmetro

contendo alta quantidade de glucoronoxilomanana (GXM), o principal polissacarídeo da

cápsula de C. neoformans. Essas vesículas também foram purificadas a partir de

46

macrófagos infectados com o fungo, demonstrando que sua produção pode ocorrer “in

vivo”. É importante ressaltar que em ensaios com células mortas não houve liberação de

vesículas para o meio extracelular, descartando a possibilidade dessas estruturas serem

meros artefatos de células lisadas. Em fungos, as vesículas possivelmente estão

relacionadas com a secreção de proteínas secretadas que não possuem um sistema

clássico de secreção. Sabe-se que há diversas proteínas citosólicas localizadas na parede

celular de fungos que não possuem peptídeo sinal (revisão em Chaffin et al., 1998).

A parede celular de fungos é uma estrutura dinâmica formada principalmente por

polissacarídeos de quitina e glucana, mas contem também proteínas, glicoproteínas e

lipídeos. Diversas funções são atribuídas à parede celular, entre elas manter o balanço

osmótico, o formato celular, promover proteção mecânica e morfogênese (revisão em de

Groot et al., 2005). O estudo detalhado desta estrutura é importante por 2 motivos: é o

primeiro compartimento em contato com estruturas e com o sistema imune do hospedeiro

e está ausente de células de mamíferos, tornando-se um ótimo alvo farmacológico. Em C.

albicans os dois assuntos mais abordados na literatura em relação à parede celular são a

presença de adesinas, pois a adesão é um pré-requisito para o estabelecimento da infecção

e seu potencial de modular a resposta imune (revisão em Chaffin, 1998).

Em Ascomycetos a parede celular é formada por uma bicamada constituída

principalmente por polímeros de carboidratos e em menor concentração por proteínas,

glicoproteínas e lipídeos. Em Neurospora crassa, a camada externa é alcali-solúvel

formada de fibrilas e glicoproteínas, enquanto a interna é formada por polímeros de β-

1,3-glucana e quitina (Mahadevan e Tatum, 1967). Em C. albicans, a camada

polissacarídica interna é formada pela ligação covalente de β-1,3-glucana à β-1,6-glucana

47

e quitina. Classicamente as proteínas de parede celular estão covalentemente ligadas às

pequenas cadeias laterais de β-1,6-glucana via âncora de glisosilfosfatidilinositol (GPI)

ou diretamente às cadeias de β-1,3-glucana, como é o caso das proteínas Pir (“protein

with internal repeats”). As proteínas da família Pir, que são O-glicosiladas, podem ser

extraídas com NaOH em condições brandas, assim como as proteínas ALS (“alkali

linkage sensitive”). Manoproteínas ligadas covalentemente por pontes dissulfeto podem

ser extraídas por tratamento com agente redutor. Outras proteínas, que são glicosiladas,

possuem peptídeo sinal e não são covalentemente ligadas, podem ser extraídas com água

quente ou detergentes iônicos como SDS. Nos últimos anos diversas proteínas

citoplasmáticas resistentes a esses tratamentos têm sido detectadas na parede celular, e

diversas não possuem peptídeo sinal. Especula-se que são exportadas por uma via não

convencional (revisões em Chaffin et al., 1998; de Groot et al., 2005; Ruiz-Herrera et al.,

2006, Nombela et al., 2006).

As proteínas de parede celular estão relacionadas com diversas funções como

porosidade, retenção de água, proteção contra estresse, manutenção, adesão, formação de

biofilme, antigenicidade, retenção de ions de ferro, hidrofobicidade e virulência. Yun e

colaboradores (1997) observaram que a superexpressão das proteínas Pir em S. cerevisiae

produz um efeito protetor contra a osmotina, uma proteína de defesa em plantas. O

proteoma de parede celular de C. albicans também revelou 3 proteínas GPI-ancoradas

que possivelmente contribuem com a virulência: uma delas é uma proteína ligadora de

grupo heme que poderia facilitar a sobrevivência em condições limitantes de íons ferro e

duas superóxido dismutases que agiriam contra ataques oxidantes de macrófagos (de

Groot et al., 2004; Fradin et al., 2005). Além das proteínas covalentemente ligadas, há

48

importantes proteínas que estão associadas à parede celular como a lacase presente na

camada externa da parede celular de C. neoformans, que é capaz de transformar

substratos fenólicos em melanina (Zhu et al., 2001).

Em P. brasiliensis, poucas proteínas associadas à parede cellular foram descritas,

incluindo duas adesinas (Gonzalez et al., 2005), uma gliceraldeído-3-fosfato

desidrogenase (Barbosa et al., 2006) e gp43 (Vicentini et al., 1994), todas com

capacidade de ligação com componentes da matriz extracelular, além de uma chaperone

mitocondrial PbDnaJ1 (Batista et al., 2006), e melanina (Silva et al., 2006).

49

Proteomic analysis of differentially expressed cell wall-associated

proteins from the human pathogen Paracoccidioides brasiliensis

Alisson L. Matsuoa, Ernesto S. Nakayasub, Luciane Ganikob , Kátia C. Carvalhoa,

Igor C. Almeidab, Rosana Pucciaa,*

aDepartment of Microbiology, Immunology and Parasitology, Cell Biology Division, Federal

University of São Paulo, 04023-62, São Paulo, Brazil

bDepartment of Biological Sciences, University of Texas at El Paso, El Paso, TX 79968, USA

50

Racional

O estudo de proteínas ligadas covalentemente ou associadas à parede de P.

brasiliensis é pobre. Até o momento foram descritos apenas um homólogo da chaperone

Mdj1 (Batista et al., 2006), duas adesinas, uma de 32 kDa (Gonzales et al., 2005) e a

gliceraldeído-3-fosfato desidrogenase (Barbosa et al., 2006) com capacidade de ligar

proteínas da matriz extracelular, além de melanina (Gomez et al., 2001). Algumas das

proteínas encontradas na parede poderiam estar transientemente localizadas nesse

compartimento inseridas em vesículas com destino extracelular. O estudo de vesículas

extracelulares no fungo é inexistente. Neste capítulo apresentamos um manuscrito

contendo as análises de proteoma de frações da parede celular de dois isolados de P.

brasiliensis e o estudo preliminar de vesículas extracelulares do fungo.

51

Abstract

This work aimed at identifying differentially expressed cell wall-associated

proteins (CWAPs) in the dimorphic human pathogen P. brasiliensis. We compared

isolates Pb18 and Pb3, yeast phase, cultivated in F12/glu supplemented or not fetal

bovine serum (FBS). These isolates belong to distinct phylogenetic groups and evoke

different types of immune response and disease progression in the B.10A mouse model.

To obtain cell wall preparations, debris from mechanically disrupted cells were

extensively washed with salt to remove contaminants and a fraction of proteins attached

by disulfides bonds was removed with dithiotreitol. CWAPs were chemically extracted

with NaOH or HF-pyridine, and digested with trypsin. The resulting peptides were

identified by mass spectrometry (LC-MS/MS), validated by Phenyx and Mascot

softwares, and quantified using an exponentially modified protein abundance index. Only

6 serum proteins were detected. We analyzed sixty fungal CWAPs that were identified by

a minimum of 4 peptides in at least one sample. The most abundant were translation

elongation factor 1, hsp70 and histone H2B. The amount of ribosomal proteins, hsp70,

hsp90, catalase, calmodulin, actin and several metabolic proteins increased when both

isolates were cultivated in FBS. Among the CWAPs prevalent in Pb18, thiol-specific

oxidant and glyceraldehyde-3-phosphate dehydrogenase might help explain Pb18 higher

virulence. Prevalence of gp43 in Pb3 cell wall, on the other hand, might contribute to its

faster elimination and decreased virulence. When compared to Pb18, the Pb3 change of

protein profile in serum suggested that it might be less adapted to host conditions.

52

Introduction

Paracoccidioides brasiliensis is a thermodimorphic fungus responsible for

paracoccidioidomycosis (PCM). PCM is a systemic granulomatous mycosis prevalent in

Latin America, where at least 10 million individuals might be infected by inhalation of

conidia (San-Blas et al., 2002). Within the pulmonary alveolar epithelium, the infectious

particles can establish infection once they transform into the parasitic yeast form. The

dimorphic transition appears to be a general requirement for pathogenicity of dimorphic

fungi (Nemecek et al., 2006). In P. brasiliensis, the transition from mycelium to yeast is

accompanied by several changes in the cell wall, such as an increase in chitin content and

a shift in the anomeric structure of the β-1,3- (alkali insoluble) to α-1,3-glucan (alkali

soluble), as reviewed by San-Blas et al. (1993), besides reorganization of membrane

glycosphingolipids (Levery et al., 1998).

The gp43 is a dominant immunodiagnostic antigen from P. brasiliensis found

predominantly in the culture supernatant (Puccia and Travassos, 1991), but also partly

associated to the cell wall (Straus et al., 1996). This glycoprotein is capable of binding to

extracellular matrix-associated components, such as murine laminin (Vicentini et al.,

1994; Gesztesi et al., 1996) and fibronectin (Mendes-Giannini et al., 2006), being

therefore a possible virulence factor. On the other hand, it bears protective T-cell epitopes

that are vaccine candidates (Taborda et al., 1998) and might be used in immunotherapy

associated with chemotherapy (Marques et al., 2006). The PbGP43 gene has up to 6

distinct genotypes (Morais et al., 2000), where the most polymorphic are characteristic of

53

a small group of isolates recently classified as phylogenetic criptic species PS2,

according to a multilocus study conducted by Matute et al. (2006). So far this group

includes Brazilian clinical isolates Pb3, Pb4, BT84, Uberlândia, Venezuelan Pb2 and

armadillo T10B1 (Matute et al., 2006). In a B10.A mouse model, isolates Pb2, Pb3 and

Pb4 evoked regressive PCM, while isolates from the main species S1, where Pb18 is

included, were consistently more virulent (Carvalho et al., 2005). Isolates from the

Colombian phylogenetic group PS3 have not yet been compared to the others in terms of

virulence.

The cell wall of most fungal pathogens is the outermost layer of contact with the

immune system and other host structures; hence its contents are fundamental to determine

the faith of the infection. The cell wall is a dynamic compartment composed mainly of

structural polysaccharides (chitin and glucans), with fewer percentages of proteins,

glycoproteins and lipids. It is responsible for osmotic balance, cell shape, mechanical

protection and morphogenesis. In Candida albicans, an internal layer of β-1,3-glucans is

covalently linked to β-1,6-glucan and chitin. Cell wall proteins (CWPs) are responsible

for remodeling, adhesion, biofilm formation and antigenicity. Classical CWPs are

covalently linked either to β-1,6-glucan short side chains via a

glycosilphosphatidylinusitol (GPI) anchor or directly to the long β-1,3 glucan chain. This

is the case of a family of proteins with internal repeats (Pir) that are highly O-

glycosylated. Pir proteins can be extracted with weak NaOH treatment, which also

extracts other alkali-sensitive proteins (ASL). GPI-anchored proteins are sensitive to HF-

pyridine, while mannoproteins, covalently linked by disulfide bonds, are extracted with

reducing agents. Other proteins, which are glycosylated and bear a signal peptide, are

54

non-covalently linked to the cell wall and are sensitive to extraction with hot water,

chaotropic agents or ionic detergents. A growing number of proteins have been reported

that are covalently linked, and hence resistant to these treatments, do not possess known

signal peptides or sugar moieties and are originally cytoplasmic proteins. The type of

association to or function at the cell wall is not clear, and it has been speculated that they

are transported through a non-conventional export pathway (reviews in Ruiz-Herrera et

al., 2006; Nombela et al., 2006).

In P. brasiliensis, few cell wall-associated proteins have been identified,

including two adhesins (Gonzalez et al., 2005), a glyceraldehyde-3-phosphate

dehydrogenase (Barbosa et al., 2006), and gp43 (Vicentini et al., 1994), all capable of

binding extracellular matrix-associated proteins. Another important protein associated

with the cell wall is the enzyme laccase, responsible for melanin synthesis (Silva et al.,

2006). In addition, our group has recently identified a mitochondrial member of the J

family co-localized to the cell wall (Batista et al., 2006).

In this study, we focused on the proteome of cell wall-associated proteins

(CWPAs). We evaluated by mass spectrometry total CWPAs from fractions extracted

with mild alkali and HF-pyridine of the isolates Pb18 and Pb3 grown in the presence and

in the absence of bovine fetal serum.

55

Materials and methods

Fungal isolates and culture conditions. We analyzed P. brasiliensis isolates 18 (Pb18)

and Pb3 (as in Morais et al., 2000; original name 608; B26 in Matute et al., 2006). Fungal

yeast cells were recovered from the lungs of mouse infected intratracheally, as described

(Carvalho et al., 2005), and maintained short-term at 36oC in solid modified YPD (0.5%

bacto-yeast extract, 0.5% casein peptone and 1.5% dextrose, pH 6.5) before use. For cell

wall extraction, yeast cells were recovered from liquid cultures (900 mL) grown for 7

days (late log phase) in F12 defined medium (Difco)/1.5% glucose (F12/glu)

supplemented or not with 5% fetal bovine serum (FBS), at 36oC, with agitation at 120

rpm in a rotary shaker. For adaptation to the culture medium, fungal cells from solid

medium were first grown in the same conditions (100 mL, 6 days) as pre-inocula before

being transferred to a larger volume of fresh medium.

Cell wall purification. Yeast cells grown in liquid F12/glu supplemented or not with 5%

FBS were microscopically analyzed for purity and viability (> 95%) with Tripan blue.

Fungal cells (10 ml of wet cells) were collected by centrifugation (1,300 x g for 5 min)

and washed 5 times with an excess amount of ice-cold buffer A (PBS, 0.1 M EDTA, 1

mM PMSF, pH 7.5). Washed cells were suspended in the same buffer and mechanically

disrupted with glass beads (425-600 µm, Sigma), at a proportion of 1:2:1 (v:v:v, wet

cells: buffer: glass beads), using a vortex (10 times for 30 sec, alternating with 30-sec

intervals in ice). Cells debris containing the wall fraction were separated from

cytoplasmatic contents (supernatant) by centrifugation at 1,300 x g for 5 min. Cell walls

56

were then washed with ice-cold buffer A another 47 times (1 x 1,300 x g for 5 min; 1 x

1,300 x g for 10 min; 15 x 16,000 x g for 10 min and 30 x 16,000 x g for 15 min),

according to a standard protocol for fungal cell wall isolation (Previato et al., 1979).

Washed cell walls were incubated for 1 hour at 37oC with buffer A containing 50 mM

DTT (dithio-threitol) to release proteins associated or aggregated by S-S bonds. Purified

cell walls were washed 3 extra times in ice-cold milliQ water (16,000 x g, 15 min) and

then lyophilized.

Cell wall protein extraction and fractionation. Cell wall-associated proteins (CWAPs)

sensitive to mild alkali extraction, including covalently linked Pir proteins, were released

by incubating purified cell walls (100 mg) with 30 mM NaOH (5 mL) at 4oC for 16 h.

The reaction was interrupted by adding acetic acid (8 µl), then the suspension was

centrifuged (16,000 x g) and filtrated through a sterile 0.22-micron filter. Samples were

dried in a speed vac apparatus and stored at -20oC prior to trypsin digestion. CWAPs

sensitive to acid treatment were released by HF-pyridine extraction. Purified cell walls

(100 mg) were mixed in a vortex with 500 µL of pure HF-pyridine (70% HF in 30%

pyridine, Fluka) in ice, and incubated for 3h in ice. The reactions were diluted with 9,5

mL of milliQ water and immediately desalted in a POROS R1 cartridges. Briefly, the

cartridges were prepared with 1 mL 40mg/mL POROS 50 R1 resin (Applied Biosystems,

Foster City, CA) in a solid-phase extraction support, washed with methanol and n-

propanol and equilibrated with 0.05%. Then the sample was loaded and washed with

0.05% TFA, before elution with 80% acetonitrile (ACN) containing 0.05% TFA. Proteins

were dried in a speed vac apparatus and stored at -20oC until trypsin digestion.

57

Sample preparation for mass spectrometry (MS) analysis. Protein digestion was

carried out as described by Stone & Williams (1996). Extracted cell wall-associated

proteins were dissolved in 40 µL 400 mM NH4HCO3, pH 8.0, containing 8 M urea and

the disulfide bonds were reduced by adding 10 µL 50 mM DTT and incubating for 15

min at 50oC. Then the cysteine residues were alkylated using 10 µL 100 mM

iodoacetoamide (IA) for 30 min, at room temperature, in the dark. Then the reaction was

diluted with HPLC-grade water to obtain a final concentration of 1 M urea. The digestion

was performed with 4 µg trypsin, for 24 h at 37oC. The reaction was terminated by

adding 1 µL pure formic acid (FA) and the peptides were desalted in a POROS R2

ziptips. The ziptips were prepared using 50 µL of a 40 mg/mL POROS 50 R2 resin

(Applied Biosystems, Foster City, CA) suspension in 2-propanol into 200-µL

micropipette tips. After equilibrating the ziptips with 0.05% TFA, peptide samples were

loaded and washed with 0.05% TFA before eluting with 80% ACN/0.05% TFA.

Peptides were 2-fold diluted in strong cation-exchange (SCX) buffer (25%

ACN/0.5% FA) and fractionated using a POROS 50 HS ion exchange ziptip. The ziptips

were prepared as the POROS R2, except that 20 µL of POROS 50 HS resin (50%

suspension in 20% ethanol) was added instead. The ziptips were equilibrated with SCX

buffer prior to loading the samples. Then the samples were washed with the same buffer

and eluted with increasing NaCl concentrations dissolved in SCX buffer (25, 50, 100, 200

and 500 mM). All fractions were dried in speed vac, desalted in POROS R2 ziptips, dried

in speed vac again and stored at -20oC until analysis.

58

Mass spectrometry analysis and peptide validation. Tryptic peptides were

dissolved in 30 µL of 0.1% F, and subjected (1 µL) to liquid chromatography-mass

spectrometry (LC-MS) analysis. LC was performed in a PepMap reversed-phase column

(15 cm x 75 µm, 3-µm C18, LC Packings, Dionex) coupled to an Ultimate (LC Packings,

Dionex) nanoHPLC. Bound peptides were eluted with 5-35% ACN/0.1% FA over 100

min and directly analyzed in a Q-tof 1 (Micromass, Waters) mass spectrometer. Spectra

in positive-ion mode were collected in the 400–1800 m/z range, and each peptide was

fragmented (MS/MS) for 3 seconds in the 50–2050 m/z range. MS/MS data were

converted into peak lists (PKL format) using ProteinLynx 2.0 (Waters). MS peaks were

smoothed twice with 5 channels using Savitzky-Golay method, and centered with 4

channels in top 80%. For the MS/MS peaks a threshold of 5% was set. After smoothing

twice with 3 channels in Savitzky-Golay method, the peaks were centered with 4

channels in top 80% and converted into monoisotopic ions. Converted peak lists were

submitted to database search analysis using Phenyx (GeneBio) and Mascot (Matrix

Science) softwares against the NCBInr and NCBInr fungi databases. The following

parameters were used: (i) enzyme: trypsin, (ii) one missed cleavage allowed, (iii) fixed

modification: carbamidomethylation of cysteines, (iv) variable modification: oxidation of

methionine, (v) peptide tolerance: 1000 ppm, (vi) MS/MS tolerance: 0.5 Da. The

validated Mascot peptides had a minimum ion score of 48 when using the entire NCBInr

database or 37 and above using the NCBInr fungi database. Peptides validated by Phenyx

had a minimum ion score of 7.0.

59

Relative quantification of proteins. In order to estimate absolute protein

contents in a complex mixture we used the exponentially modified protein abundance

index (emPAI) as described by Ishihama et al. (2005). PAI is an index defined as the

number of observed peptides for an individual protein divided by the number of

observable (theoretical) peptides for that protein upon digestion with a proteinase of

known cleavage site. The number of observable peptides upon trypsin digestion was

estimated with the PepitideMass software (http://ca.expasy.org/tools/peptide-mass.html),

setting as parameters the same conditions used in our experiments. For absolute

abundance, PAI was converted to exponentially modified PAI (emPAI) following the

equation: 10PAI – 1.

Results

Identification of cell wall-associated proteins (CWAPs). We isolated, fractionated and

identified trypsin-cleaved peptides from total P. brasiliensis CWAPs that were sensitive

to individual extractions with NaOH or HF-pyridine. We compared isolates Pb18 and

Pb3 grown in defined F-12/glucose medium supplemented or not with 5% FBS.

Considering that P. brasiliensis cell wall preparations have not been boiled in SDS before

fractionation, our peptide analysis identified not only integral cell wall proteins, but also

those associated with other proteins or carbohydrate moieties, thus justifying the term

“cell wall-associated proteins”. The peptide sequences revealed by mass spectrometry

were scored and validated using Phenyx or Mascot softwares, which generated a list of

proteins containing identical peptides. Many peptides were detected in both NaOH and

60

HF-pyridine fractions. Hence, we merged peptide data from these fractions in order

simplify the analysis of the most abundant CWAPs in each isolate and growth condition.

In addition, we only took into consideration proteins identified by four or more peptides

in at least one of the four samples analyzed (Pb18 or Pb3 in F12/glu; Pb18 or Pb3 in

FBS). The number of peptides identifying each protein can be seen in Table 1 (in

parenthesis). Most of the proteins have identical peptides to fungal homologues, and 17

derive from the P. brasiliensis databases.

Under those stringent parameters, we have identified a total of sixty CWAPs

(Table 1), from which thirty-five were detected in both Pb18 and Pb3 irrespective of

growth condition (Fig. 1). Eight CWAPs were found exclusively in cells cultivated in the

presence of serum. Among them, 3 were common to both isolates (cytochrome c oxidase

subunit VIa, initiation factor 4A, mitochondrial carrier protein), 3 were detected only in

Pb18 (fungal transcription factor, polyketide synthase-related protein, hypothetical

protein) and 2 only in Pb3 (avidin, hypothetical protein). No exclusive protein was found

when the isolates grew only in F12/glu (Figure 1).

Quantitative analysis of CWAPs. The results of emPAI abundance, calculated as

detailed in Materials and Methods, can be seen in Table 1. According to these

calculations, about 75% of total abundance (from the four samples) scored 0.5 or lower,

24% were between 0.5 and 1.0 and only 1% were higher than 1.0.

The most abundant CWAP was by far a homologue of translation elongation

factor 1, followed by heat shock 70 and histone H2B. Two hsp70 family members were

CWAPs in P. brasiliensis, including a mitochondrial one (gi3124921), which was not the

61

most abundant. Other mitochondrial proteins have been identified, such as, putative

cytochrome c oxidase (subunits IV and VIa), carrier protein Ggc1, hsp60, probable

mitochondrial F1 ATPase (subunit alpha) and cpn10. Mitochondrial and ribosomal

proteins have previously been shown to be cross-linked to cell wall components

(Bowman et al., 2006).

It is relevant to point out that we detected only 6 FBS proteins in our samples,

specifically, gamma globin, hemoglobin alpha-1 subunit, alpha-2-HS-glycoprotein,

vitronectin, inter-alpha-trypsin inhibitor and apolipoprotein A-I (Table 2). The small

number and differential distribution between isolates suggests that they might be

specifically bound to cell wall components, but can be displaced under relatively mild

conditions.

Effect of FBS on in vitro expression of CWAP. We used the emPAI values of

abundance (Table 1) to calculate relative values that reflected how growth in 5% FBS

changed the pattern of CWAP expression comparatively in Pb18 and Pb3. The results

were then used to build a color-coded graph (FBS/F12), as explained in the legend of Fig.

2, where each protein of Table 1 is represented. Among the 60 CWAPs analyzed, 22 were

more abundant (green pallet) in both isolates cultivated in FBS, including a series of

ribosomal proteins, hsp60, hsp70 and hsp90, catalase, calmodulin, actin and several

metabolic proteins. The summation of emPAI values from all CWAPs of P. brasiliensis

(both isolates) grown in FBS was 65.48, while the values calculated for cells grown in

F12/glu totalized 38.36, suggesting that serum components evoke general overexpression

of proteins. Total emPAI was considerably higher for Pb18 than PB3 (28.69 x 17.69)

62

when both isolates were cultivated in the absence of serum. When growing FBS,

although Pb18 and PB3 contribute equally well to the total emPAI of 65.48, the number

of individual proteins whose abundance increased was higher in Pb3 (42) than in Pb18

(36). Additionally, more proteins had decrease abundance (red pallet) in Pb18 (17) than

in Pb3 (4), suggesting that these isolates have distinct regulatory mechanisms at the

transcription and/or translation/post translation levels, which ultimately might account for

a distinct host/parasite relationship.

Differentially expressed CWAPs in Pb8 and Pb3. In order to identify more clearly the

CWAPs differentially expressed in Pb18 and Pb3 in terms of abundance, we analyzed

Table 1 data comparing the values according to the isolate (Pb18/Pb3, in FBS or F12), as

seen in Fig.2. When both culture conditions are considered, 20 proteins were more

abundant in Pb18 (green pallet), including some ribosomal proteins, Hsp90, actin,

glyceraldehyde-3-phosphate dehydrogenase, thiol-specific antioxidant and catalase. Only

4 CWAPs were more abundant in Pb3 (red pallet) in both culture conditions, specifically,

Hsp70, triosephosphate isomerase, aconitase and mannitol-1-phosphate dehydrogenase.

When yeast cells cultured in FBS are analyzed, 21 CWAPs were relatively more

abundant in Pb3 than in Pb18 (e.g. gp43, enolase, calmodulin, polyubiquitin, plasma

membrane ATPase), while 33 CWAPs were prevalent in Pb18. Some of these proteins

are discussed below.

63

Discussion

The present work aimed at identifying cell wall-associated proteins (CWAPs)

differentially expressed in two P. brasiliensis isolates (Pb18 and Pb3) grown in the

presence or in the absence of fetal bovine serum (FBS). Pb18 and Pb3 have distinct

phylogenetic background (Matute et al., 2006) and prompt opposite types of disease

progression in a B.10A mouse model (Carvalho et al., 2005). To remove contaminants,

isolated cell walls were extensively washed with salt (Previato et al., 1979), and a

fraction of proteins attached by disulfide bonds was removed with dithiotreitol. We

analyzed CWAPs that have been removed under mild NaOH and HF-pyridine extraction

and blended the results because many of them were found in both fractions. Since

previous hot water and/or detergent removal (de Groot et al., 2004) have not been

undertaken, we assume that the detected proteins are bonded either covalently or not.

That could also justify why our list of proteins does not contain many previously reported

covalently-linked CPWs (de Groot et al., 2004), whose accessibility might be hampered

under our experimental conditions. Alternatively, P. brasiliensis yeast cells do not

necessarily contain the same C. albicans proteins on the cell wall. On the other hand, the

lack of complete genome information for P. brasiliensis does not presently allow for a

complete analysis of the data.

Among the CWAPs prevalent in Pb18, thiol-specific oxidant (TSA) and

glyceraldehyde-3-phosphate dehydrogenase (GADPH) were 2-fold more abundant in

Pb18 than in Pb3 grown in FBS, and that might be relevant for the Pb18 increased

virulence. P. brasiliensis GADPH has previously been detected at the cell wall by

64

immunoelectron microscopy, and found to be involved in the initial colonization and

fungal dissemination due to its capacity to bind fibronectin, type I collagen and laminin

(Barbosa et al., 2005). Assays using P. brasiliensis previously incubated with polyclonal

anti-GADPH or recombinant GADPH promoted an inhibition of adherence and

internalization by pneumocytes in vitro. In C. albicans, TSA was differentially detected

on the cell surface of the pathogenic mycelial form (Urban et al., 2003), and it seems to

be indispensable for the yeast-mycelial transition under oxidative stress (Shin et al.,

2005). In P. brasiliensis, the TSA1 gene was preferentially expressed in the pathogenic

yeast phase (Marques et al., 2004) and presently found associated with the cell wall. TSA

might play a key role in P. brasiliensis survival under an unfavorable oxidative

environment within phagocytes, especially for Pb18, which would then evoke a more

successful infection.

We interestingly detected a polyketide synthase-related protein only in Pb18

grown in serum. This enzyme is involved in the dihydroxynaphtalene (DHN)-melanin

pathway, whose precursor is acetate (Bell e Wheeler, 1986). Melanin is an important

fungal virulence factor (reviewed by Nosanchuk and Casadevall, 2006), and was also

found in P. brasiliensis, where melanized cells were poorly phagocytosed by

macrophages and were less sensitive to antifungal drugs (Silva et al., 2006). Although a

homologue for the laccase gene has not yet been found in P. brasiliensis EST databases,

yeast cells are able to synthesize a melanin-like pigment in the presence of a L-3,4-

dihydroxyphenilalanine (L-DOPA) precursor. Conidia, on the other hand, can produce it

also in the absence of L-DOPA, suggesting the existence of an alternative pathway that

could take acetate as precursor. Two members of a putative kinase regulator family, 14-

65

3-3-like protein and protein 2 were 2-fold more abundant in Pb18 than in Pb3 cultivated

in FSB. The relevance of this difference will be further investigated.

Cytoskeleton proteins have seldom been described associated with the cell wall.

We presently detected β-tubulina only in Pb18 (alkaline fraction), and actin abundantly

distributed in all samples. Both had increased amounts when the isolates were cultivated

in FSB, especially β-tubulina (3.5-fold). Actin is a structural protein intimately related

with S. pombe cell wall synthesis. Reverting protoplasts revealed that new cell wall

formation co-localizes with spots of actin, and treatment with cytochalasin D depleted

cell wall formation (Kobori et al., 1989). A structure called filasome was found to be

composed by a vesicle of 35-70 nm that is surrounded by fine filaments containing actin

(Takagi et al., 2003). The filasome follows an actin filament pathway and fuses with the

plasma membrane releasing the vesicle that probably carries cell wall material. That

could explain the finding of actin in our cell wall preparations, and increase of actin and

tubulin on the cell wall of P. brasiliensis cultivated in serum could reflect increased

protein transportation. Tubulins are the main constituents of microtubules, which are

responsible for many functions including segregation of genetic material, maintenance of

cell shape and intracellular transport of proteins, lipids and organelles (review in Palmer

et al., 2005).

In addition to the surface Pb18 proteins that could aid this isolate to establish a

more successful infection than Pb3, two FBS proteins were differentially bound more

abundantly to Pb18 surface, specifically, vitronectin and apolipoprotein. Vitronectin,

which is involved in the regulation of blood coagulation (Dahlbäck et al., 1987), has also

been found on cell wall extracts of C. albicans (Jakab et al., 1993), possibly bound

66

through a specific cell wall receptor (revision in Chaffin et al., 1998). Considering that in

C. albicans interaction with vitronectin increased binding to and phagocytosis by

macrophages, it might help the microorganism to establish infection (Limper et al.,

1994). The main function of lipoproteins is to transport lipids and lipid-soluble

molecules. There is no previous report on cell wall-bound apolipoprotein A-I, however

bovine apolipoprotein A-II presents an antimicrobial effect against E. coli and S.

cerevisiae (Motizuki et al., 1998). The serum apolipoprotein was detected only in Pb18

cell wall fraction extracted with HF-pyridine, but its relevance is obscure.

Two serum proteinase inhibitors, inter-alpha-trypsin (a serpin) and alpha-2-HS-

glycoprotein (cystatin), were found associated to the P. brasiliensis cell wall. These

inhibitors could play a role in neutralization of fungal proteinases potentially important

for fungal infection, dissemination or nutrition. Presently, we have not found any cell

wall-associated proteinases in our experimental conditions; however inter-alpha-trypsin

was detected associated with only Pb3 cell wall, possibly inhibiting a proteinase not

secreted by Pb18.

Gp43 has been detected in small amounts in the acid-extracted cell wall fraction.

In Pb18, it has been detected only in the sample grown in F12/gluc, while its contents

increase 4-fold in Pb3 cultivated in FSB. We can speculate that its presence in Pb3 cell

wall in vivo might contribute to its more effective clearance by the immune system,

resulting in regressive PCM in B10.A, as opposed to Pb18 (Carvalho et al., 2005).

Polyubiquitin was detected in comparable amounts in the alkali-extracted cell wall

fraction of both Pb18 and Pb3 grown in the absence of serum, but only Pb3 had

detectable amounts when cultivated in serum-supplemented medium, at a 2-fold increase.

67

In the cytoplasm, polyubiquitin classically tags proteins for proteolysis by proteasomes,

but it is now understood that polyubiquitin is involved in many other cellular processes,

which include tagging of membrane proteins before internalization (reviewed in Aguilar

and Wendland, 2003). Its finding on the cell wall remains to be elucidated.

Calmodulin was 3-fold more abundant in Pb3-FBS than in Pb18-FSB.

Calmodulin is a small calcium-binding protein that plays an important role as modulator

of intracellular calcium signaling, and participates in the regulation of gene expression in

response to external stimuli, including stress (Kraus and Heitman, 2003). In P.

brasiliensis, calmodulin seems to be essential for the mycelial-to-yeast-transformation

(Carvalho et al., 2003). Overall, Pb3 seemed to expose at the cell wall fewer proteins than

Pb18 when growing in defined medium and then respond more fiercely to the presence of

serum. That could explain why a potent mediator such as calmodulin prevailed in Pb3.

We have recently shown that, when compared to Pb18, Pb3 has a slower transcriptional

response of some HSP genes to heat shock (Batista et al., 2007). It is possible that a

slower transcriptional response results in a crippled adaptation of Pb3 to the host, which

will in turn more readily eliminate it.

We have detected in P. brasiliensis many cell wall-associated proteins lacking

known secretion signal, which have previously been reported in C. albicans and S.

cerevisiae (reviewed by Nombela et al., 2006). They include glycolysis proteins like

enolase, GADPH, fructose 1,6-bisphosphate aldolase, stress proteins like Hsp70, Hsp60,

Hsp90, and others like catalase, TSA, cytochrome c, 14-3-3-like protein, translation

elongation factors. We presently detected proteins from different intracellular

compartments. Histones, typically nuclear, have previously been microscopically

68

localized to H. capsulatum cell wall and anti-histone monoclonal antibodies protected

against experimental histoplasmosis (Nosanchuk et al., 2003). We found several

mitochondrial proteins, including Hsp 60. In H. capsulatum, cell wall hsp60 was shown

to bind to macrophage complement receptor type 3 (Long et al., 2003) and it protects

vaccinated BALB/C and C57BL/6 (Deepe et al., 1996; Deepe and Gibbons, 2001). Anti-

recombinant Hsp60 have been found in the sera from several paracoccidioidomycosis

patients (Izacc et al., 2001).

Two hsp70 members were abundantly found associated with the P. brasiliensis

cell wall from both Pb3 and Pb18. Recently, Batista et al. (2006) detected an Mdj1

homologue both in cell wall and mitochondria, where it is originally localized. It was the

first time that a protein from the J family was found on the cell wall but its role in this

compartment is speculative. Since it is responsible for mitochondrial hsp70 regulation in

yeast, finding of mitochondrial hsp70 on the cell wall suggests that the chaperone

Mdj1/hsp70 complex could be active also at the cell wall. Presently we have not found

Mdj1 in alkali or acid cell wall fractions, probably because this molecule was washed out

during DTT extraction, as seen to occur with whole cells (Batista et al., 2006).

In conclusion, by using a proteomics approach we identified cell wall-associated

proteins that are differentially expressed in Pb18 and Pb3 growing in the presence of

serum that might help explain their distinct degree of virulence in the mouse model.

69

Legends

Figure 1: Graphic representation of the 60 CWAPs analyzed in this work.

Figure 2. Color representation of the relative abundance of each protein shown in Table

1, calculated as follows: a) FBS/F12 - influence of serum in the change of abundance for

Pb18 and Pb3. Values for growth in FBS were divided by the correspondent control

values in the absence of serum (F12); b) Pb18/Pb3 - comparison between Pb18 and Pb3

cultivated both in the presence or in the absence (F12) of FBS. Red to dark red (values

between 0.5 and 0.99) indicate quantitative decrease in FBS or lower abundance in Pb18,

while dark green to green (values between 1.0 and 2.0) indicate serum-related increase in

the amount of CWAPs or higher abundance in Pb18. Unaltered abundance (1.0) is

represented in black. When peptides for a particular protein were absent (emPAI = 0) the

correspondent protein was considered 2 times repressed or abundant (0.5 or 2.0,

respectively).

70

Figure 1.

Matsuo et al., 2007

71

Figure 2

Matsuo et al., 2007

72

emPAI

Protein AC Pb 18 Pb18

Serum Pb 3 Pb3

Serum similar to eukaryotic translation elongation factor 1 alpha 1 gi|82995146 8.33 (32) 7.11 (30) 4.72 (25) 5.58 (27)

heat shock protein 70 [Paracoccidioides brasiliensis] gi|14538021 1.12 (28) 2.16 (43) 2.71 (49) 3.25 (54)

histone H2B [Rosellinia necatrix] gi|27531291 2.16 (12) 1.87 (11) 0.96 (7) 2.48 (13)

HHF2p [Candida glabrata]/histone 4 gi|21668067 1.93 (7) 1.15 (5) 1.15 (5) 1.15 (5)

actin gi|66804921 0.85 (12) 1.64 (19) 0.76 (11) 1.27 (16)

ATP synthase F1, beta subunit gi|77685559 0.80 (13) 1.25 (18) 0.80 (13) 0.80 (13)

Grp1p [Exophiala dermatitidis] gi|82571189 0.75 (10) 1.08 (13) 0.06 (1) 0.85 (11)

calmodulin gi|168777 0.13 (1) 0.44 (3) 0.13 (1) 1.98 (9)

bipA [Aspergillus awamori] gi|2582633 0.69 (19) 0.56 (16) 0.40 (12) 0.95 (24)ribosomal protein S15, putative [Aspergillus fumigatus Af293] gi|66852980 0.62 (4) 0.62 (4) 0.27 (2) 0.83 (5) hypothetical protein MG00221.4 [Magnaporthe grisea 70-15]/ 40S RIBOSOMAL PROTEIN S7 gi|39975265 0.65 (7) 0.78 (8) 0.15 (2) 0.54 (6) histone H3 [Fusarium proliferatum] gi|5918753 1.51 (6) 0.00 (0) 0.00 (0) 0.58 (3) 40S ribosomal subunit protein S9 [Cryptosporidium hominis] gi|54656881 0.12 (1) 0.78 (5) 0.41 (3) 0.58 (4) putative cytochrome c oxidase subunit VIa [Paracoccidioides brasiliensis] gi|50428882 0.00 (0) 1.40 (8) 0.00 (0) 0.39 (3) 70 kDa heat shock protein [Paracoccidioides brasiliensis] gi|31324921 0.51 (19) 0.44 (17) 0.21 (9) 0.51 (19)

thiol-specific antioxidant [Ajellomyces capsulatus] gi|14161441 0.41 (4) 0.67 (6) 0.29 (3) 0.29 (3)

malate dehydrogenase [Paracoccidioides brasiliensis] gi|47119068 0.54 (10) 0.36 (7) 0.14 (3) 0.54 (10)

14-3-3-like protein 2 [Paracoccidioides brasiliensis] gi|38569374 0.18 (3) 0.73 (10) 0.32 (5) 0.32 (5) putative cytochrome c oxidase subunit IV [Paracoccidioides brasiliensis] gi|50428890 0.37 (3) 0.23 (2) 0.37 (3) 0.52 (4) heat shock protein 60 [Paracoccidioides brasiliensis] gi|3088571 0.39 (14) 0.46 (16) 0.12 (5) 0.46 (16)hypothetical protein [Neurospora crassa OR74A]/Chaperonin 10 Kd subunit gi|85079266 0.58 (4) 0.41 (3) 0.00 (0) 0.41 (3) hypothetical protein AN5800.2 [Aspergillus nidulans /eukaryotic ribosomal protein L18 gi|67539260 0.45(5) 0.45 (5) 0.00 (0) 0.45 (5)

polyubiquitin [Cicer arietinum] gi|24817262 0.31 (2) 0.00 (0) 0.31 (2) 0.72 (4) glyceraldehyde-3-phosphate dehydrogenase [Paracoccidioides brasiliensis] gi|30995493 0.37 (7) 0.50 (9) 0.25 (5) 0.20 (4)

heat shock protein 90 [Paracoccidioides brasiliensis] gi|60656557 0.24 (11) 0.49 (20) 0.10 (5) 0.43 (18)

14-3-3-like protein [Paracoccidioides brasiliensis] gi|30351156 0.51 (7) 0.43 (6) 0.13 (2) 0.19 (3)

thioredoxin [Paracoccidioides brasiliensis] gi|34980254 0.15 (1) 0.15 (1) 0.78 (4) 0.15 (1) ADT_NEUCR ADP,ATP CARRIER PROTEIN (ADP/ATP TRANSLOCASE) gi|42551288 0.48 (7) 0.32 (5) 0.12 (2) 0.25 (4)

73

SconC [Paracoccidioides brasiliensis]

gi|61608602 0.29 (3) 0.29 (3) 0.09 (1) 0.41 (4) peptidyl-prolyl cis/trans isomerase [Paracoccidioides brasiliensis] gi|34979129 0.28 (3) 0.51 (5) 0.28 (3) 0.00 (0)

hypothetical protein MG04489.4 [Magnaporthe grisea gi|39945014 0.00 (0) 0.00 (0) 0.00 (0) 1.05 (5)

ribosomal L10 protein [Paracoccidioides brasiliensis] gi|28797723 0.17 (3) 0.37 (6) 0.23 (4) 0.17 (3) unnamed protein product [Aspergillus oryzae]/mitochondrial F1 ATPase subunit alpha gi|83765516 0.21 (7) 0.35 (11) 0.15 (5) 0.18 (6)

catR [Aspergillus niger] Catalase gi|840716 0.13 (3) 0.39 (8) 0.00 (0) 0.33 (7) Conserved ribosomal protein P0 similar to rat P0, human P0, and E. coli L10e; shown to be phosphory gi|6323371 0.17 (2) 0.37 (4) 0.08 (1) 0.17 (2) unnamed protein product [Aspergillus oryzae]/Ribosomal_L7 gi|83774731 0.30 (4) 0.22 (3) 0.14 (2) 0.14 (2) fructose 1,6-biphosphate aldolase 1 [Paracoccidioides brasiliensis] gi|29826036 0.13 (2) 0.19 (3) 0.06 (1) 0.34 (5)

Plasma membrane ATPase (Proton pump) gi|728908 0.23 (7) 0.13 (4) 0.06 (2) 0.27 (8) Triosephosphate isomerase [Clostridium acetobutylicum ATCC 824] gi|15893999 0.00 (0) 0.13 (2) 0.19 (3) 0.34 (5)

avidin gi|451889 0.00 (0) 0.00 (0) 0.00 (0) 0.65 (5) 40S ribosomal protein S3 [Chaetomium globosum CBS 148.51] gi|88183746 0.00 (0) 0.28 (5) 0.16 (3) 0.16 (3)

enolase [Cryphonectria parasitica] gi|40806812 0.25 (5) 0.05 (1) 0.14 (3) 0.14 (3)

beta-tubulin [Neotyphodium coenophialum] gi|2293 0.12 (2) 0.47 (7) 0.00 (0) 0.00 (0) cytosolic small ribosomal subunit S4 [Aspergillus fumigatus Af293] gi|71000467 0.29 (5) 0.29 (5) 0.00 (0) 0.00 (0)

40S ribosomal protein S18 [Coccidioides immitis RS] gi|90307829 0.21 (3) 0.29 (4) 0.00 (0) 0.07 (1)

hypothetical protein [Yarrowia lipolytica] gi|50556244 0.12 (2) 0.25 (4) 0.00 (0) 0.18 (3) ribosomal protein S0.e.B, cytosolic [Aspergillus fumigatus Af293] gi|70998726 0.05 (1) 0.23 (4) 0.05 (1) 0.17 (3)

elongation factor 2 [Neurospora crassa] gi|13925370 0.04 (2) 0.17 (8) 0.06 (3) 0.15 (7) COG0515: Serine/threonine protein kinase [Nostoc punctiforme PCC 73102] gi|23124801 0.05 (1) 0.21 (4) 0.05 (1) 0.10 (2)

glycoprotein gp43 gi|1588394 0.15 (3) 0.00 (0) 0.05 (1) 0.20 (4) electron transfer flavoprotein, subunit beta imported [Aspergillus fumigatus Af293] gi|70998364 0.06 (1) 0.27 (4) 0.00 (0) 0.06 (1)

aconitase [Aspergillus terreus] gi|3661614 0.00 (0) 0.05 (2) 0.15 (6) 0.15 (6) ATP synthase subunit A [Roseovarius nubinhibens ISM] gi|83952823 0.07 (2) 0.04 (1) 0.04 (1) 0.19 (5) mannitol-1-phosphate dehydrogenase [Paracoccidioides brasiliensis] gi|28797565 0.00 (0) 0.00 (0) 0.04 (1) 0.27 (6) unnamed protein product [Aspergillus oryzae]/60S ribosomal protein L8 gi|83767463 0.20 (4) 0.09 (2) 0.00 (0) 0.00 (0) EUKARYOTIC INITIATION FACTOR 4A (EIF-4A) (EIF4A) [Neurospora crassa] gi|32414453 0.00 (0) 0.17 (4) 0.00 (0) 0.08 (2)

74

putative mitochondrial carrier protein Ggc1 [Candida albicans SC5314] gi|68484497 0.00 (0) 0.20 (4) 0.00 (0) 0.05 (1) fungal specific transcription factor [Aspergillus fumigatus Af293] gi|70981502 0.00 (0) 0.11 (5) 0.00 (0) 0.00 (0)

hypothetical protein [Neurospora crassa N150] gi|85082883 0.00 (0) 0.08 (4) 0.00 (0) 0.00 (0) COG3321: Polyketide synthase modules and related proteins [Burkholderia mallei GB8 horse 4] gi|67640326 0.00 (0) 0.08 (6) 0.00 (0) 0.00 (0)

Table 1. Associated CWPs identified by LC-MS/MS of isolates Pb18 and Pb3 grown in

F12/glu supplemented or not with FCS. The relative abundance of proteins was recorded

as emPAI based on the number of peptides (in parenthesis) detected for each protein. The

listed proteins are organized from the most to the least abundant protein.

75

Serum Proteins

AC Pb 18 NaOH

Pb 18 HF

Pb 3 NaOH

Pb 3 HF

gamma globin gi|393

8 11 4 10

Hemoglobin alpha-1 subunit

gi|122272

0 4 0 5

Vitronectin

gi|73587073

5 3 3 0

inter-alpha-trypsin inhibitor

gi|27806743

0 0 4 0

apolipoprotein A-I, apoA-1

gi|245563

0 8 0 0

alpha-2-HS-glycoprotein

gi|27806751

6 2 5 2

Table 2. FBS proteins bound to cell wall of isolates Pb18 and Pb3. The number of

peptides detected by LC-MS/MS is displayed for each protein.

76

References Aguilar,RC, B Wendland, 2003, Ubiquitin: not just for proteasomes anymore: Curr.Opin.Cell Biol., v. 15, p. 184-190.

Barbosa,MS, S N Bao, P F Andreotti, F P de Faria, M S S Felipe, L D Feitosa, M J S Mendes-Giannini, C M D Soares, 2006, Glyceraldehyde-3-phosphate dehydrogenase of Paracoccidioides brasiliensis is a cell surface protein involved in fungal adhesion to extracellular matrix proteins and interaction with cells: Infection and Immunity, v. 74, p. 382-389.

Batista,WL, A L Matsuo, L Ganiko, T F Barros, T R Veiga, E Freymuller, R Puccia, 2006, The PbMDJ1 gene belongs to a conserved MDJ1/LON locus in thermodimorphic pathogenic fungi and encodes a heat shock protein that localizes to both the mitochondria and cell wall of Paracoccidioides brasiliensis: Eukaryotic Cell, v. 5, p. 379-390.

Batista,WL, T F Barros, G H Goldman, F V Morais, R Puccia, 2007, Identification of transcription elements in the 5' intergenic region shared by LON and MDJ1 heat shock genes from the human pathogen Paracoccidioides brasiliensis. Evaluation of gene expression: Fungal Genet Biol., v. 44, p. 347-356.

Bell,AA, M H Wheeler, 1986, Biosynthesis and functions of fungal melanins.: Annu.Rev.Phytopatol., v. 24, p. 411-451.

Bowman,SM, A Piwowar, M Al Dabbous, J Vierula, S J Free, 2006, Mutational analysis of the glycosylphosphatidylinositol (GPI) anchor pathway demonstrates that GPI-anchored proteins are required for cell wall biogenesis and normal hyphal growth in Neurospora crassa: Eukaryotic Cell, v. 5, p. 587-600.

Carvalho,KC, L Ganiko, W L Batista, F V Morais, E R Marques, G H Goldman, M E Franco, R Puccia, 2005, Virulence of Paracoccidioides brasiliensis and gp43 expression in isolates bearing known PbGP43 genotype: Microbes and Infection, v. 7, p. 55-65.

Chaffin,WL, J L Lopez-Ribot, M Casanova, D Gozalbo, J P Martinez, 1998, Cell wall and secreted proteins of Candida albicans: Identification, function, and expression: Microbiology and Molecular Biology Reviews, v. 62, p. 130-+.

da Silva,MB, A F Marques, J D Nosanchuk, A Casadevall, L R Travassos, C P Taborda, 2006, Melanin in the dimorphic fungal pathogen Paracoccidioides brasiliensis: effects on phagocytosis, intracellular resistance and drug susceptibility: Microbes Infect., v. 8, p. 197-205.

Dahlback,K, H Lofberg, B Dahlback, 1987, Immunohistochemical Demonstration of Vitronectin in Association with Elastin and Amyloid Deposits in Human-Kidney: Histochemistry, v. 87, p. 511-515.

de Carvalho,MJ, R S Amorim Jesuino, B S Daher, I Silva-Pereira, S M de Freitas, C M Soares, M S Felipe, 2003, Functional and genetic characterization of calmodulin from the dimorphic and pathogenic fungus Paracoccidioides brasiliensis: Fungal Genet Biol., v. 39, p. 204-210.

De Groot,PWJ, A D de Boer, J Cunningham, H L Dekker, L de Jong, K J Hellingwerf, C de Koster, F M Klis, 2004, Proteomic analysis of Candida albicans cell walls reveals covalently bound carbohydrate-active enzymes and adhesins: Eukaryotic Cell, v. 3, p. 955-965.

Deepe,GS, Jr., R Gibbons, G D Brunner, F J Gomez, 1996, A protective domain of heat-shock protein 60 from Histoplasma capsulatum: J.Infect.Dis., v. 174, p. 828-834.

Deepe,GS, Jr., R Gibbons, 2001, V beta 6+ T cells are obligatory for vaccine-induced immunity to Histoplasma capsulatum: J.Immunol., v. 167, p. 2219-2226.

77

Gesztesi,JL, R Puccia, L R Travassos, A P Vicentini, J Z de Moraes, M F Franco, J D Lopes, 1996, Monoclonal antibodies against the 43,000 Da glycoprotein from Paracoccidioides brasiliensis modulate laminin-mediated fungal adhesion to epithelial cells and pathogenesis: Hybridoma, v. 15, p. 415-422.

Gonzalez,A, B L Gomez, A Restrepo, A J Hamilton, L E Cano, 2005, Recognition of extracellular matrix proteins by Paracoccidioides brasiliensis yeast cells: Med Mycol., v. 43, p. 637-645.

Goshe,MB, R D Smith, 2003, Stable isotope-coded proteomic mass spectrometry: Current Opinion in Biotechnology, v. 14, p. 101-109.

Ishihama,Y, Y Oda, T Tabata, T Sato, T Nagasu, J Rappsilber, M Mann, 2005, Exponentially modified protein abundance index (emPAI) for estimation of absolute protein amount in proteomics by the number of sequenced peptides per protein: Molecular & Cellular Proteomics, v. 4, p. 1265-1272.

Izacc,SMS, F J Gomez, R S A Jesuino, C A Fonseca, M S S Felipe, G S Deepe, C M A Soares, 2001, Molecular cloning, characterization and expression of the heat shock protein 60 gene from the human pathogenic fungus Paracoccidioides brasiliensis: Medical Mycology, v. 39, p. 445-455.

Jakab,E, M Paulsson, F Ascencio, A Ljungh, 1993, Expression of Vitronectin and Fibronectin-Binding by Candida-Albicans Yeast-Cells: Apmis, v. 101, p. 187-193.

Kobori,H, N Yamada, A Taki, M Osumi, 1989, Actin Is Associated with the Formation of the Cell-Wall in Reverting Protoplasts of the Fission Yeast Schizosaccharomyces-Pombe: Journal of Cell Science, v. 94, p. 635-646.

Kraus,PR, J Heitman, 2003, Coping with stress: calmodulin and calcineurin in model and pathogenic fungi: Biochem.Biophys.Res.Commun., v. 311, p. 1151-1157.

Levery,SB, M S Toledo, A H Straus, H K Takahashi, 1998, Structure elucidation of sphingolipids from the mycopathogen Paracoccidioides brasiliensis: An immunodominant beta-galactofuranose residue is carried by a novel glycosylinositol phosphorylceramide antigen: Biochemistry, v. 37, p. 8764-8775.

Limper,AH, D M Oriordan, J E Standing, E C Crouch, 1994, Surfactant Protein-D Mediates Adherence of Pneumocystis-Carinii to Alveolar Macrophages: Clinical Research, v. 42, p. A427.

Long,KH, F J Gomez, R E Morris, S L Newman, 2003, Identification of heat shock protein 60 as the ligand on Histoplasma capsulatum that mediates binding to CD18 receptors on human macrophages: J.Immunol., v. 170, p. 487-494.

Marques,AF, M B da Silva, M A Juliano, L R Travassos, C P Taborda, 2006, Peptide immunization as an adjuvant to chemotherapy in mice challenged intratracheally with virulent yeast cells of Paracoccidioides brasiliensis: Antimicrob.Agents Chemother., v. 50, p. 2814-2819.

Marques,ER, M E S Ferreira, R D Drummond, J M Felix, M Menossi, M Savoldi, L R Travassos, R Puccia, W L Batista, K C Carvalho, M H S Goldman, G H Goldman, 2004, Identification of genes preferentially expressed in the pathogenic yeast phase of Paracoccidioides brasiliensis, using suppression subtraction hybridization and differential macroarray analysis: Molecular Genetics and Genomics, v. 271, p. 667-677.

Matute,DR, J G Mcewen, B A Montes, G San Blas, E Bagagli, J T Rauscher, A Restrepo, F Morais, G Nino-Vega, J W Taylor, 2006, Cryptic speciation and recombination in the fungus Paracoccidioides brasiliensis as revealed by gene genealogies: Molecular Biology and Evolution, v. 23, p. 65-73.

Mendes-Giannini,MJS, P F Andreotti, L R Vincenzi, J L M da Silva, H L Lenzi, G Benard, R Zancope-Oliveira, H L D Guedes, C P Soares, 2006, Binding of extracellular matrix proteins to Paracoccidioides brasiliensis: Microbes and Infection, v. 8, p. 1550-1559.

78

Morais,FV, T F Barros, M K Fukada, P S Cisalpino, R Puccia, 2000, Polymorphism in the gene coding for the immunodominant antigen gp43 from the pathogenic fungus Paracoccidioides brasiliensis: Journal of Clinical Microbiology, v. 38, p. 3960-3966.

Motizuki,M, T Itoh, M Yamada, S Shimamura, K Tsurugi, 1998, Purification, primary structure, and antimicrobial activities of bovine apolipoprotein A-II: Journal of Biochemistry, v. 123, p. 675-679.

Nemecek,JC, M Wuthrich, B S Klein, 2006, Global control of dimorphism and virulence in fungi: Science, v. 312, p. 583-588.

Nombela,C, C Gil, W L Chaffin, 2006, Non-conventional protein secretion in yeast: Trends Microbiol., v. 14, p. 15-21.

Nosanchuk,JD, J N Steenbergen, L Shi, G S Deepe, Jr., A Casadevall, 2003, Antibodies to a cell surface histone-like protein protect against Histoplasma capsulatum: J.Clin.Invest, v. 112, p. 1164-1175.

Nosanchuk,JD, A Casadevall, 2006, Impact of melanin on microbial virulence and clinical resistance to antimicrobial compounds: Antimicrob.Agents Chemother., v. 50, p. 3519-3528.

Palmer,KJ, P Watson, D J Stephens, 2005, The role of microtubules in transport between the endoplasmic reticulum and Golgi apparatus in mammalian cells: Biochem.Soc.Symp., p. 1-13.

Previato,JO, P A Gorin, R H Haskins, L R Travassos, 1979, Soluble and insoluble glucans from different cell types of the human pathogen Sporothrix schenckii: Experimental Mycology, v. 3, p. 92-105.

Puccia,R, L R Travassos, 1991, 43-Kilodalton Glycoprotein from Paracoccidioides-Brasiliensis - Immunochemical Reactions with Sera from Patients with Paracoccidioidomycosis, Histoplasmosis, Or Jorge Lobos Disease: Journal of Clinical Microbiology, v. 29, p. 1610-1615.

Ruiz-Herrera,J, M V Elorza, E Valentin, R Sentandreu, 2006, Molecular organization of the cell wall of Candida albicans and its relation to pathogenicity: FEMS Yeast Res., v. 6, p. 14-29.

Russell,WK, Z Y Park, D H Russell, 2001, Proteolysis in mixed organic-aqueous solvent systems: Applications for peptide mass mapping using mass spectrometry: Analytical Chemistry, v. 73, p. 2682-2685.

San Blas,G, 1993, Biochemical and physiological aspects in the dimorphism of Paracoccidioides brasiliensis: Arch.Med Res., v. 24, p. 267-268.

San Blas,G, G Nino-Vega, T Iturriaga, 2002, Paracoccidioides brasiliensis and paracoccidioidomycosis: Molecular approaches to morphogenesis, diagnosis, epidemiology, taxonomy and genetics: Medical Mycology, v. 40, p. 225-242.

Shin,DH, S Jung, S J Park, Y J Kim, J M Ahn, W Kim, W Choi, 2005, Characterization of thiol-specific antioxidant 1 (TSA1) of Candida albicans: Yeast, v. 22, p. 907-918.

Straus,AH, E Freymuller, L R Travassos, H K Takahashi, 1996, Immunochemical and subcellular localization of the 43 kDa glycoprotein antigen of Paracoccidioides brasiliensis with monoclonal antibodies: J.Med Vet.Mycol., v. 34, p. 181-186.

Taborda,CP, M A Juliano, R Puccia, M Franco, L R Travassos, 1998, Mapping of the T-cell epitope in the major 43-kilodalton glycoprotein of Paracoccidioides brasiliensis which induces a Th-1 response protective against fungal infection in BALB/c mice: Infection and Immunity, v. 66, p. 786-793.

79

Takagi,T, S A Ishijima, H Ochi, M Osumi, 2003, Ultrastructure and behavior of actin cytoskeleton during cell wall formation in the fission yeast Schizosaccharomyces pombe: Journal of Electron Microscopy, v. 52, p. 161-174.

Urban,C, K Sohn, F Lottspeich, H Brunner, S Rupp, 2003, Identification of cell surface determinants in Candida albicans reveals Tsa1p, a protein differentially localized in the cell: Febs Letters, v. 544, p. 228-235.

Vicentini,AP, J L Gesztesi, M F Franco, W Desouza, J Z Demoraes, L R Travassos, J D Lopes, 1994, Binding of Paracoccidioides-Brasiliensis to Laminin Through Surface Glycoprotein Gp43 Leads to Enhancement of Fungal Pathogenesis: Infection and Immunity, v. 62, p. 1465-1469.

80

Moléculas secretadas de Paracoccidioides

brasiliensis: resultados preliminares sobre

vesículas extracelulares

81

Materiais e métodos

Gel filtração em coluna com Sepharose CL-4B

Uma coluna de gel filtração (16 mm de diâmetro por 40 cm de altura da

Amersham) contendo 70 mL de Sepharose CL-4B (separação entre 20.000 kDa a 60 kDa

- Amersham) foi previamente equilibrada a 4oC com 5 volumes de tampão acetato de

amônio 100 mM, pH 7,3, em fluxo constante de 4 mL/hora, utilizando uma bomba

peristáltica. Amostras de 2 mL de sobrenadantes de cultura de P. brasiliensis concentrado

eram aplicadas na coluna e frações de 1 mL recolhidas com um coletador automático para

posterior análise.

ELISA

A detecção de componentes de alto peso molecular no sobrenadante de cultura de

P. brasiliensis foi realizada através de ELISA, em placas de 96 cavidades (Nunc,

MaxiSorpe surface). Os poços eram sensibilizados com as vesículas adicionados de 50

µL de tampão carbonato/bicarbonato 200 mM pH 9,6 e incubados a 4°C por uma noite

(o/n). Em alguns experimentos as vesículas foram tratadas com metaperiodato de sódio

(reagente que oxida epitopos de carboidratos) em 3 concentrações diferentes (5, 10 e 20

mM). As placas eram lavadas 5 vezes com PBS acrescido de 0,1% de Tween-20 (PBS-T)

e incubadas com PBS-Molico 5% (100 µl/poço) por 3 h à temperatura ambiente. Após 3

lavagens com PBS-T, os antígenos eram incubados por 2 horas a 37oC com 100 µl de

PBS-Molico 5% contendo soro de paciente com PCM (1:1000). Após 5 lavagens com

PBS-T, eram adicionados 100 µl de PBS-Molico 5% contendo anticorpo de cabra anti-

82

IgG de humano (1:1000) ou anticorpo de macaco anti-IgG de coelho (1:1000), ambos

conjugados com HRP (Amersham), e incubados por 1h a 37°C. Depois de 5 lavagens

com PBS-T, era feita uma única lavagem com 200 µL de tampão bicarbonato 50 mM, pH

9,6. A revelação era feita por quimioluminescência, com o uso de luminol (ECL reagent,

Pierce) diluído 1:20 em tampão bicarbonato 50 mM, pH 9,6, preparado de acordo com

instruções do fabricante. A leitura era feita em luminômetro de microplacas (Cambridge

Technology, Watertown, MA, Modelo 7710) e expressa como unidades de relativas de

luminescência (URL).

Purificação de vesículas do sobrenadante de P. brasiliensis

P. brasiliensis isolados Pb3 (nome original: 608), Pb12 (nome original:

Argentina), Pb18 (nome original: 18) ou Pb339 (nome original: B-339) eram crescidos a

36º C, sob agitação, por 7 a 10 dias em meio definido F12 (Invitrogen) acrescido de 1,5%

de glucose (F12glu), e em alguns experimentos complementado com 5% de soro fetal

bovino (SFB), previamente ultracentrifugado a 200.000 g por 4 horas para retirada das

vesículas do soro. Os sobrenadantes de cultura eram filtrados em membrana estéril de

0,22 µm com a ajuda de uma bomba à vácuo e concentrados em sistema AMICON

(Millipore) utilizando as membranas YM10 com limite de exclusão de 10.000 Da ou

BIOMAX (PBMK 076 10 – Millipore) com limite de exclusão de 300.000 Da. A amostra

era lavada com 20 mL de tampão contendo 100 mM de acetato de amônio pH 7,3,

concentrada até um volume final de 2 mL e aplicada na coluna de gel filtração de CL-4B.

Alternativamente, o concentrado era equilibrado em 200 mL de PBS, concentrado a um

volume final de 10 mL e submetido a uma ultracentrifugação por 4 horas a 100.000 x g, a

83

4º C. O precipitado contendo vesículas era liofilizado (para análise de proteínas) ou

estocado em geladeira por período curto (para análises funcionais). Toda a manipulação

era feita a frio. A viabilidade das culturas, recolhidas em fase logarítmica ou logarítmica

tardia de crescimento, era estimada em >90% pelo método de exclusão com Tripan blue

em microscópio ótico, com o qual adicionalmente se determinou a pureza das

preparações.

Gradiente de sacarose para purificação de vesículas de P. brasiliensis

Um precipitado de 100.000 x g conservado a 4oC foi dissolvido em 5 mL de uma

solução 2,6 M de sacarose em 20 mM de Tris pH 7,2 e colocado em um tubo de

ultracentrífuga (12 mL) específico para o rotor SW41 (Sorvall). Acima da amostra foi

adicionado um gradiente linear de 2,0 a 0,25 M de sacarose em 20 mM de Tris pH 7,2, e

a separação foi realizada através de uma ultracentrifugação de 270.000 x g por 16 horas.

As amostras foram recolhidas a partir do fundo do tubo em frações de 1 mL, adaptado-se

uma bomba peristáltica com uma pipeta Pasteur de vidro em um fluxo de 1 mL/min. As

frações recolhidas foram estocadas em geladeira.

Precipitação com ácido tricloroacético (TCA)

Para precipitação de proteínas com TCA, as amostras eram adicionadas de 10%

de TCA e conservadas por 20 min no gelo. Após centrifugar por 20 min a 4º C em

microcentrífuga (14.500 rpm), o sobrenadante era descartado e o precipitado lavado com

1 mL de acetona gelada (estocada a – 20º C). A amostra era novamente centrifugada e o

precipitado seco em speed vac.

84

Digestão de proteínas em solução aquosa/orgânica para análise em espectrometria

de massas (LC-MS/MS)

A digestão de proteínas em solução foi realizada conforme Russel (2001) e Goshe

(2003), com modificações. Proteínas em quantidade de até 100 µg eram dissolvidas em

100 µL de solução contendo metanol e 50 mM de bicarbonato de amônio pH 8,0 (v/v,

60/40) em tubo de rosca. A amostra era incubada por 5 min a 95oC e deixada resfriar à

temperatura ambiente. A redução das proteínas era realizada com 5 mM de dithiotreitol

(DTT) a 37oC por 30 min, seguida de alquilação utilizando 10 mM de iodoacetamida por

90 min no escuro. A amostra era diluída 5x no mesmo tampão e as proteínas digeridas

com tripsina em uma razão de 1/50 (wt/wt) de enzima/substrato por 16 h a 37oC. A

reação era interrompida com uma concentração final de 0,05% de ácido trifluoroacético

(TFA), a amostra seca em “speed vac” e dessalinizada em “ziptip” de fase reversa.

Dessalinização em “ziptip” de fase reversa para análise em LC-MS/MS

Um pequena quantidade de fibra de vidro era adicionada a uma ponteira de 200

µL para evitar vazamento da resina de fase reversa POROS 50 R2 (40 mg/mL em

isopropanol). A coluna era ativada com 5 volumes de metanol, equilibrada com 5

volumes de TFA (0,046%) e a amostra equilibrada na mesma solução era aplicada. A

coluna era lavada com 5 volumes de TFA 0,046%, e eluída com 5 volumes de

acetonitrila 80% contendo 0,046% de TFA. As amostras eram secas em speed vac e

analisadas em LC-MS/MS.

85

Fracionamento de peptídeos em ziptip de troca-iônica

Fibra de vidro era adicionada a uma ponteira de 200 µL para evitar vazamento da

resina POROS HR50 (interação catiônica forte). Após o empacotamento, a resina era

equilibrada com 5 volumes de tampão (EB) contendo 0,5% de ácido fórmico e 25% de

acetonitrila e a amosta em EB era aplicada. Os peptídeos não ligados eram lavados com 5

volumes de EB e a eluição realizada com as concentrações de 25, 50, 100, 200 e 500 mM

de NaCl em EB. A dessalinização era realizada em “ziptip” de fase reversa.

Análise de peptídeos por LC-MS/MS

As amostras previamente secas em speed vac eram ressuspendidas em 30 µL de

ácido fórmico (AF) 0,1%, dos quais 1 µL era aplicado para análise em LC-MS. A

cromatografia dos peptídeos era realizada na coluna de fase reversa PepMap (15 cm x 75

µm, 3µm, LC Packings) em um nanoHPLC Ultimate (LC Packings), com um

autosampler Famos (LC Packings). O gradiente para eluição dos peptídeos era de 0 - 40%

de solvente B (80% acetonitrila/AF 0,1%) em 100 min, ou de 5 - 50% de solvente B em

30 min (para amostras provenientes de digestão em gel).

Os peptídeos eluídos da coluna eram analisados diretamente no espectrômetro de

massas (Q-tof 1, Micromass, Waters). Os espectros no modo MS eram coletados a cada 2

segundos, em uma faixa de 400-1800 m/z e cada peptídeo era submetido a uma

fragmentação (MS/MS) por 3 s, em uma faixa de 50-2050 m/z. Os espectros MS/MS

coletados eram convertidos em peak lists (formato PKL) e analisados com os softwares

Mascot (Matrix Science) and Phenyx (GeneBio) nos bancos NCBInr, P. brasiliensis,

Neurospora e S. cerevisiae.

86

Microscopia eletrônica

Os procedimentos de inclusão, ultramicrotomia e reações de imunocitoquímica

em leveduras de P. brasiliensis foram realizados no Centro de Microscopia Eletrônica

(UNIFESP), sob a coordenação da Dra. Edna F. Haapalainen e em colaboração com a

pós-doutoranda Luciane Ganiko do nosso laboratório na época.

Preparações de vesículas ou leveduras do isolado Pb339 crescido em meio F12glu

em fase logarítmica de crescimento eram lavadas em tampão cacodilato de sódio 0.1 M,

pH 7.2 (TC) e fixadas em solução de Karnowisk modificado (glutaraldeído 2.5%, v/v,

paraformaldeído 2%, m/v, em TC) durante 3½ h à temperatura ambiente, sob agitação

constante. As preparações permaneceram no fixador a 4oC até o momento do

processamento. O precipitado era lavado em TC e incluído em ágar 2.5% (m/v) para

obtenção de fragmentos de ~ 2 mm de espessura. A etapa de pós-fixação era feita com

solução de tetróxido de ósmio 1% (v/v) em TC durante 1 h à temperatura ambiente, sob

agitação constante. Alternativamente, as preparações eram pós-fixadas na presença de

ferricianeto de potássio 1% (v/v) para melhor preservação e resolução dos sistemas de

membranas (Forbes et al., 1977). Parte das amostras era lavada em água ultra-pura (2 x

10 min) para contrastação in bloc com acetato de uranila 0.5% (m/v) durante 30 min,

protegida da luz. As amostras não contrastadas com acetato de uranila eram lavadas em

TC (2 x 10 min).

Posteriormente, as amostras eram desidratadas em concentrações crescentes de

etanol (Merck) 70%, 90% e 100% durante 30 min cada, transferidas para óxido de

propileno (EMS), com 2 trocas de 30 min. Os fragmentos de precipitado celular eram

87

infiltrados com resina Epon (EMS) ou Spurr (EMS), com aumento progressivo da razão

de resina para óxido de propileno (1:1 durante 2 h, 2:1 durante 16 h). Para infiltração com

resina pura, eram realizadas 2 trocas de 2 - 3 h e, posteriormente, os fragmentos eram

transferidos para cápsulas contendo resina. As cápsulas permaneceram na estufa a 65oC

durante 48 h para polimerização da resina. As amostras infiltradas com Epon eram pós-

fixadas em tetróxido de ósmio sem ferricianeto de potássio.

Seções ultra-finas (70 – 120 nm) obtidas no ultramicrótomo Leica eram coletadas

em grades de níquel resvestidas com formvar e carbono, lavadas, fixadas com

glutaraldeído 2.5% (v/v) e enxaguadas em água ultra-pura (MilliQ). As análises eram

feitas em microscópio eletrônico Jeol JEM 1200EX model II.

88

Resultados e Discussão

• Análise da preparação de vesículas por “Western blot” e microscopia eletrônica

Amostras de sobrenadante de cultura do P. brasiliensis Pb339 cultivado em meio

F12/glu complementado com SFB depletado de vesículas foram concentradas em

Amicon e ultracentrifugadas a 100.000 x g por 4 horas a 4oC. O precipitado foi

ressuspendido em tampão de amostra para SDS-PAGE, fervido e analisado por “Western

blot”. Na figura 1 mostramos o padrão de bandas reconhecido por soro de paciente com

PCM (1:1000), como revelado por quimioluminescência. Os componentes do precipitado

100.000 x g reconhecidos pelo soro de paciente concentraram-se na faixa de peso

molecular entre 75 e 45 kDa, aproximadamente e foram específicos, como sugerido pela

comparação com a reação negativa obtida com soro de indivíduo normal.

89

Figura 1. Análise por “Western blot” da antigenicidade das moléculas do precipitado a 100,000 x g com

soro de paciente com PCM.

Uma fração semelhante de amostra foi analisada através de microscopia eletrônica

de transmissão. Na Figura 2A podemos observar a presença neste precipitado de diversas

estruturas delimitadas por membranas, aparentemente vesículas, com tamanhos entre 20 e

100 nm, coerentes com o tamanho dos exossomos descritos na literatura. Revendo

algumas fotos de leveduras de P. brasiliensis analisadas por microscopia eletrônica,

pudemos observar também algumas células aparentemente liberando vesículas para o

meio extracelular (Figura 2B - setas). Coincidentemente estas vesículas têm tamanho de

20 a 100 nm de diâmetro, sugerindo que o fungo realmente libera vesículas para o meio

extracelular.

90

Figura 2. Detecção por microscopia eletrônica de transmissão de vesículas liberadas pelo P. brasiliensis no

meio extracelular. A) Sobrenadante de cultura concentrado e ultracentrifugado a 100.000 x g. B) Célula de

P. brasiliensis aparentemente liberando vesículas para o meio extracelular (setas).

• Cromatografia da preparação de vesículas extracelulares de P. brasiliensis

91

Outra estratégia utilizada para o reconhecimento de estruturas vesiculares

extracelulares foi o fracionamento de sobrenadantes concentrados de cultura de fungo em

uma coluna de Sepharose CL-4B. Previamente, a coluna foi calibrada com 2 marcadores

de peso molecular, a saber, tireoglobulina (~700 kDa) e BSA (dímero com ~135 kDa),

nas mesmas condições descritas em Materiais e Métodos, e as frações analisadas em

espectrofotômetro (Abs280). Na figura 3 podemos observar que o pico de tireoglobulina

está na fração 43, enquanto o dímero de BSA ficou predominantemente na fração 60,

coerente com o resultado esperado.

Figura 3. Calibração de uma coluna de Sepharose CL-4B. Perfil de eluição (A280) de 200 µg de

tireoglobulina (~700 kDa) ou BSA (dímero ~135 kDa) em frações de 1 mL (fluxo de 4 mL/hora).

Cerca de 50 mL de sobrenadante de P. brasiliensis B-339 crescido por 10 dias em

meio F12/glu contendo 5% de SFB foi concentrado em sistema Amicon com membrana

YM10, lavado com 8 mL de tampão contendo 100 mM de acetato de amônio pH 7,3 e

concentrado até um volume final de 2 mL. A amostra concentrada foi fracionada em

92

coluna de Sepharose CL-4B ou ultracentrifugada a 100.000 x g por 4 horas a 4oC, e

estocada.

O fracionamento de sobrenadante concentrado de P. brasiliensis em coluna de

Sepharose CL-4B calibrada pode ser visto na Figura 3. Alíquotas das frações (200 µL)

foram analisadas por ELISA quanto à reatividade com soro de paciente (1:1000) com

PCM. O perfil de reatividade pode ser visto na Figura 4, onde o controle da reação foi

feito na ausência de fungo, ou seja, somente com meio F12/glu contendo ou não SFB,

concentrado e fracionado na mesma coluna. Observa-se que houve uma alta reatividade

do soro com as frações 26 a 33 eluídas a partir de sobrenadante de cultura e, portanto,

específicas para componentes fúngicos. Para uma coluna de 70 mL, o Vo estaria em

torno de 24 mL (fração 24), sugerindo que as frações entre 26 e 33 estão incluídas na

coluna. Essas frações apresentam uma massa molecular acima de 700 kDa (fração 43 –

tireoglobulina), compatíveis com a massa de vesículas (entre 1.000 e 20.000 kDa). Note

que houve uma reatividade de imunoglobulinas de paciente com componentes do soro

(perfil F12+SFB). Sobrenadantes de cultura do fungo cultivado em meio sem soro

apresentaram reatividade específica com componentes do fungo eluídos a partir da

fração 40, os quais podem representar agregados com componentes antigênicos.

93

Figura 4. CL-ELISA das frações eluídas da coluna de Sepharose CL-4B. Os gráficos mostram o perfil da

reatividade, com soro de paciente com PCM, das frações eluídas a partir do sobrenadante de cultura de P.

brasiliensis B-339.

• Análise de vesículas purificadas em gradiente de sacarose

Um dos nossos objetivos em relação às vesículas extracelulares de P. brasiliensis

é verificar se elas carregam moléculas imunomodulatórias, como já evidenciado em

organismos patogênicos como o T. cruzi (Torrecilhas et al., 2007, em revisão). Com o

94

objetivo de testar frações purificadas do pellet de 100.000 x g de sobrenadante de cultura

do fungo, após ultracentrifugação as vesículas foram fracionadas em gradiente linear

entre 2,6 e 0,25 M de sacarose em Tris pH 7,2 (Wubbolts et al., 2003). Estávamos

particularmente interessados em separar a malha que apareceu entremeada nas vesículas,

como visto na Figura 2A. As frações ímpares foram enviadas para análise no laboratório

do Prof. Carlos P. Taborda, USP. As frações foram usadas para estimular macrófagos de

cultura J774 e no sobrenadante foram dosadas as citocinas IL-10, IL-12 e óxido nítrico

(NO). Observou-se que as frações 7 e 9 foram as mais ativas na indução de NO (Figura

5), enquanto IL-10 e IL-12 não foram detectadas em nenhuma fração (resultado não

mostrado). A constatação preliminar de que vesículas extracelulares do P. brasiliensis

podem ter uma importância imunológica é altamente relevante ao nosso estudo.

95

Figura 5. Detecção da liberação de NO por macrófagos após 24 horas de incubação com as frações

vesiculares provenientes da separação por gradiente linear de sacarose.

• Proteoma de vesículas de P. brasiliensis

As informações de que o P. brasiliensis aparentemente libera vesículas, as quais

contêm moléculas antigênicas e potencialmente imunorregulatórias, direcionou nosso

interesse imediato em realizar o proteoma desse material. Paralelamente, foi realizada a

análise do proteoma da parede do fungo (manuscrito anexo), não somente pela

importância de conhecer as proteínas associadas a esse compartimento, com também para

estabelecer uma base de comparação com as proteínas de vesículas que necessariamente

devem transitar por esse compartimento antes de alcançarem o meio externo.

Os isolados a serem estudados foram selecionados segundo sua capacidade

secretora e pelo seu grau de virulência. A Pb339 tem sido usada há décadas como fonte

de antígenos extracelulares e purificação de gp43. Geralmente apresenta crescimento

abundante e secreção de inúmeros componentes moleculares. O Pb18 tem sido

igualmente utilizado há décadas em experimentos de infecção de camundongos devido à

sua alta virulência e representa a maior espécie filogenética (S1), segundo Matute et al.

(2006). O Pb12 foi o isolado mais agressivo em experimentos de PCM murina em

camundongos B10.A, para os quais IFNγ estava ausente durante a infecção (Carvalho et

al., 2005). O Pb3 representa uma espécie filogenética distinta (PS2) e tem um padrão

regressivo de PCM em camundongos B10.A (Carvalho et al., 2005; Matute et al., 2006).

O perfil em SDS-PAGE das proteínas contidas no precipitado 100,000 x g de

vesículas pode ser observado na figura 6. As preparações aparecem com vários e

abundantes componentes. Somente as preparações provenientes de Pb3 apresentaram

96

baixa quantidade de proteínas. Isso já era esperado, já que os precipitados de 100.000 x g

eram consideravelmente menores para esse isolado. Apesar das preparações conterem

uma quantidade razoável de proteínas, as digestões com tripsina mostraram poucos “hits”

significativos com o banco de dados. Para validar o proteoma de vesículas serão

necessárias preparações contendo maior quantidade de proteínas. O conhecimento do

genoma completo do P. brasiliensis também facilitará enormemente sua identificação.

Figura 6. Padrão de banda das proteínas do precipitado 100.000 x g secretadas pelo P. brasiliensis Pb18,

Pb12, Pb3 e Pb339 em gel de SDS-PAGE 10% corado pelo método da prata.

• Marcação da preparação de vesículas com anticorpos anti-alfa galactosil

Um dos nossos objetivos foi detectar um marcador para as vesículas

extracelulares de P. brasiliensis de forma a otimizar sua purificação. Para tal, já foram

testados anticorpos anti-PbLon, anti-gp43 e anti-PbMdj1, porém todos produziram

reações negativas.

97

Em T. cruzi, parte da população de vesículas extracelulares é reconhecida por

anticorpos anti-epitopo alfa-galactosil (anti-Gal), os quais podem ser isolados em grandes

quantidades a partir de soros de pacientes chagásicos (Almeida et. al., 1994). Esses

epitopos estão presentes em cadeias longas de oligossacarídeos O-ligados, principalmente

em resíduos de treonina de mucinas-like do complexo antigênico F2/3 de T. cruzi. O

complexo F2 é composto por glicoproteínas de 74 e 95 kDa e F3 por glicoproteínas de

120 a 200 kDa.

Apesar de epitopos contendo resíduos de α-galactosil não terem sido descritos em

fungos até o momento, foi feita uma reação de “immunoblot” de precipitados 100.000 x g

de sobrenadantes de P. brasiliensis Pb339 com anticorpos anti-Gal. Observou-se uma

reação intensa com um componente de cerca de 70 kDa (Fig. 7) de migração difusa no

gel de SDS-PAGE.

Figura 7. “Immunoblot” com revelação por quimioluminescência mostrando a reatividade de anticorpos

anti-alfa galactosil isolados de soro chagásico com precipitados de sobrenadantes de cultura (100.000 x g)

de P. brasiliensis Pb339: 1, crescido em YPD ou 3, crescido em meio F12/glu. Em 2, componentes do

sobrenadante de cultura < 300.000 Da (não retidos na concentração em Amicon).

98

Ensaios de ELISA utilizando concentrações crescentes do anticorpo (10, 20 e 40

µg/mL) mostraram que a reação foi dose-dependente, como ocorre com o controle

positivo, a fração F2/3 (Fig. 8).

Figura 8. Reatividade em Elisa revelada por quimioluminescência de diferentes concentrações de

anticorpos anti-alfa-galactosil isolados de soro chagásico com precipitados de sobrenadantes de cultura

(100.000 x g) de P. brasiliensis Pb339 (A) ou fração F2/3 (B) de T. cruzi.

Para verificar se os anticorpos reconhecem epitopos de caboidratos, as vesículas

foram tratadas com metaperiodato de sódio em 3 concentrações diferentes (5, 10 e 20

99

mM). Pode-se observar na Fig. 9 que 5 mM do reagente praticamente aboliram a ligação

do anticorpo com o precipitados a 100.000 x g, sugerindo que a reatividade era devida a

epitopos de galactose.

Figura 9. Reatividade em Elisa, revelada por quimioluminescência, de anticorpos anti-alfa galactosil (20

mg/mL) isolados de soro chagásico com precipitado de sobrenadantes de cultura (100.000 x g) de P.

brasiliensis tratado com diferentes concentrações de metaperiodato de sódio (NaIO4).

Um ensaio de inibição com diversos açúcares em concentração de 0,3 M

incubados com o anticorpo primário mostrou inibição principalmente com d(+)-galactose

(45%), seguida de methyl-β-D-xilopiranosídeo (38%) e methyl-α-D-galactopiranosídeo

(35%) (figura 10).

100

Figura 10. Reatividade em Elisa, revelada por quimioluminescência, de anticorpos anti-alfa-galactosil (20

mg/mL) isolados de soro chagásico com precipitado de sobrenadantes de cultura (100.000 x g) de P.

brasiliensis Pb339 na presença de 0,3 M dos açúcares indicados. O controle foi incubado na ausência de

competidor.

Ensaios para confirmar a presença do epitopo anti-alfa galactosil foram realizados

pela Dra. Luciane Ganiko, no laboratório do Dr. Igor C. Almeida da University of Texas

at El Paso (UTEP). Nesses ensaios, o substrato foi tratado com a enzima alfa-

galactosidase, a qual hidroliza resíduos de alfa-galactosidase de cadeias lineares. Na Fig.

11 observamos que o tratamento das vesículas de T. cruzi com alfa-galactosidase impediu

substancialmente o reconhecimento de uma lectina específica para resíduos alfa-

galactosil, conhecida como MOA (Marasmium oreades agglutinin). O reconhecimento da

lectina foi drasticamente reduzido, indicando que esse epitopo também está presente em

preparações de vesículas de P. brasiliensis. Tanto o anticorpo quanto a lectina poderão

101

ser utlizados na purificação das vesículas de P. brasiliensis em coluna de afinidade ou

como marcador.

Figura 11. Reatividade em Elisa, revelada por quimioluminescência, de MOA com precipitado de

sobrenadantes de cultura (100.000 x g) de P. brasiliensis (PbVs) ou vesículas de T. cruzi (TcaGalVes)

tratados (barras negras) ou não (barras brancas) com alfa-galactosidase. Controle: na ausência de antígeno.

102

Capítulo 2: conclusões

Em nosso laboratório foi especulada a possibilidade da serino-tiol proteinase

(PbST) ou algum modulador positivo da enzima ser veiculado ao meio extracelular por

vesículas secretadas pelo fungo. Os resultados preliminares de caracterização de

vesículas extracelulares de P. brasiliensis apontam que:

1. Foi observada por microscopia eletrônica a presença de estruturas semelhantes a

vesículas delimitadas por membrana em precipitados a 100.000 x g de

sobrenadantes de P. brasiliensis. Frações do precipitado correspondentes à

densidade de vesículas em gradiente de sacarose foram capazes de estimular a

produção de NO por macrófagos, porém não de IL-10 ou Il-12 nas condições

testadas. Preparações de vesículas foram reconhecidas especificamente por

anticorpos de pacientes portadores de PCM, anticorpos policlonais

monoespecíficos anti-alfa galactosil e também pela lectina MOA, que reconhece

epitopos contendo α galactose. A atividade da PbST não foi detectada nas

preparações de vesículas até o momento, possivelmente porque sua secreção não é

estimulada no meio F12/glucose. Esses resultados preliminares abriram a

perspectiva de um amplo projeto do laboratório para o estudo de vesículas

extracelulares de P. brasiliensis.

103

A parede celular de fungos é uma estrutura coerente e dinâmica formada

principalmente por polissacarídeos de quitina e glucana, mas contem também proteínas,

glicoproteínas e lipídeos. A parede celular tem fundamental importância na biologia do

fungo e na interação parasita-hospedeiro, além de ser um ótimo alvo farmacológico.

Nossos estudos sobre o proteoma diferencial de frações da parede celular de dois isolados

do P. brasiliensis, que diferem geneticamente e quanto à virulência em camundongos,

podem ser assim resumidos:

1. Foram identificadas proteínas associadas à parede cellular (CWAPs)

diferencialmente expressas no fungo dimórfico P. brasiliensis. Os isolados Pb3 e Pb18

foram cultivados em F12 glucose na presença ou não de 5% de soro fetal bovino (SFB).

Esses isolados pertencem a grupos filogenéticos distintos e provocam tipos diferentes de

resposta immune e progressão da PCM em modelo murino. Preparações de parede

celular, lavadas extensivamente com sal e tratadas com ditiotreitol, foram extraídas com

NaOH e HF, e os peptídeos tripticos foram identificados por espectrometria de massas

(LC-MS/MS) utilizando os softwares Mascot e Phenyx. 60 proteínas foram identificadas

com 4 ou mais peptídeos e as mais abundantes foram o fator de elongação 1, Hsp 70 e

histona H2B. Proteínas ribosomais, hsp70, hsp90, catalase, calmodulina e várias

proteínas metabólicas foram mais abundantes nas culturas em FSB. Entre as CWAPs

prevalentes em Pb18, oxidante tiol específica e gliceraldeído-3-fosfato desidrogenase

poderiam estar relacionadas com sua maior virulência. A prevalência de gp43 na parede

de Pb3, por outro lado, poderia contribuir com a sua eliminação mais eficaz pelo sistema

imune.

104

Referências Bibliográficas Almeida,AJ, M Martins, J A Carmona, L E Cano, A Restrepo, C Leao, F Rodrigues, 2006, New insights into the cell cycle profile of Paracoccidioides brasiliensis: Fungal Genetics and Biology, v. 43, p. 401-409.

Almeida,AJ, J A Carmona, C Cunha, A Carvalho, C A Rappleye, W E Goldman, P J Hooykaas, C Leao, P Ludovico, F Rodrigues, 2007, Towards a molecular genetic system for the pathogenic fungus Paracoccidioides brasiliensis: Fungal Genet Biol..

Almeida,IC, M A Ferguson, S Schenkman, L R Travassos, 1994, GPI-anchored glycoconjugates from Trypanosoma cruzi trypomastigotes are recognized by lytic anti-alpha-galactosyl antibodies isolated from patients with chronic Chagas' disease: Braz.J.Med Biol.Res., v. 27, p. 443-447.

Amerik,AY, V K Antonov, A E Gorbalenya, S A Kotova, T V Rotanova, E V Shimbarevich, 1991, Site-Directed Mutagenesis of la Protease - A Catalytically Active Serine Residue: Febs Letters, v. 287, p. 211-214.

Andreotti,PF, J L Monteiro da Silva, A M Bailao, C M Soares, G Benard, C P Soares, M J Mendes-Giannini, 2005, Isolation and partial characterization of a 30 kDa adhesin from Paracoccidioides brasiliensis: Microbes Infect., v. 7, p. 875-881.

Bagagli,E, A Sano, K I Coelho, S Alquati, M Miyaji, Z P De Camargo, G M Gomes, M Franco, M R Montenegro, 1998, Isolation of coccidioides brasiliensis from armadillos (Dasypus noveminctus) captured in an endemic area of paracoccidioidomycosis: American Journal of Tropical Medicine and Hygiene, v. 58, p. 505-512.

Bagagli,E, M Franco, S D M Bosco, F Hebeler-Barbosa, L A Trinca, M R Montenegro, 2003, High frequency of Paracoccidioides brasiliensis infection in armadillos (Dasypus novemcinctus): an ecological study: Medical Mycology, v. 41, p. 217-223.

Bailao,AM, A Schrank, C L Borges, V Dutra, Walquiria Ines Molinari-Madlum EE, M S Soares Felipe, M J Soares Mendes-Giannini, W S Martins, M Pereira, S C Maria de Almeida, 2006, Differential gene expression by Paracoccidioides brasiliensis in host interaction conditions: representational difference analysis identifies candidate genes associated with fungal pathogenesis: Microbes Infect., v. 8, p. 2686-2697.

Barbosa,MS, S N Bao, P F Andreotti, F P de Faria, M S S Felipe, L D Feitosa, M J S Mendes-Giannini, C M D Soares, 2006, Glyceraldehyde-3-phosphate dehydrogenase of Paracoccidioides brasiliensis is a cell surface protein involved in fungal adhesion to extracellular matrix proteins and interaction with cells: Infection and Immunity, v. 74, p. 382-389.

Barrett,AJ, N D Rawlings, 1991, Types and families of endopeptidases: Biochem.Soc.Trans., v. 19, p. 707-715.

Barros,TF, R Puccia, 2001, Cloning and characterization of a LON gene homologue from the human pathogen Paracoccidioides brasiliensis: Yeast, v. 18, p. 981-988.

Batista,WL, A L Matsuo, L Ganiko, T F Barros, T R Veiga, E Freymuller, R Puccia, 2006, The PbMDJ1 gene belongs to a conserved MDJ1/LON locus in thermodimorphic pathogenic fungi and encodes a heat shock protein that localizes to both the mitochondria and cell wall of Paracoccidioides brasiliensis: Eukaryotic Cell, v. 5, p. 379-390.

105

Batista,WL, T F Barros, G H Goldman, F V Morais, R Puccia, 2007, Identification of transcription elements in the 5' intergenic region shared by LON and MDJ1 heat shock genes from the human pathogen Paracoccidioides brasiliensis. Evaluation of gene expression: Fungal Genet Biol., v. 44, p. 347-356.

Bernard,G, N M Orii, H H S Marques, M Mendonça, M Z Aquino, A E Campeas, G B Delnegro, A Durandy, A J S Duarte, 1994, Severe acuteparacoccidioidomycosis in children: Pediatric Infectious Disease Journal, p. 510-515.

Besche,H, P Zwickl, 2004, The Thermoplasma acidophilum Lon protease has a Ser-Lys dyad active site: European Journal of Biochemistry, v. 271, p. 4361-4365.

Biagioni,L, M J Souza, L G Chamma, R P Mendes, S A Marques, N G Mota, M Franco, 1984, Serology of paracoccidioidomycosis. II. Correlation between class-specific antibodies and clinical forms of the disease: Trans.R.Soc.Trop Med Hyg., v. 78, p. 617-621.

Boddicker,JD, B D Jones, 2004, Lon protease activity causes down-regulation of Salmonella pathogenicity island 1 invasion gene expression after infection of epithelial cells: Infect.Immun., v. 72, p. 2002-2013.

Bota,DA, K J A Davies, 2001, Protein degradation in mitochondria: implications for oxidative stress, aging and disease: a novel etiological classification of mitochondrial proteolytic disorders: Mitochondrion, v. 1, p. 33-49.

Bota,DA, K J A Davies, 2002, Lon protease preferentially degrades oxidized mitochondrial aconitase by an ATP-stimulated mechanism: Nature Cell Biology, v. 4, p. 674-680.

Bota,DA, J K Ngo, K J A Davies, 2005, Downregulation of the human Lon protease impairs mitochondrial structure and function and causes cell death: Free Radical Biology and Medicine, v. 38, p. 665-677.

Botos,I, E E Melnikov, S Cherry, J E Tropea, A G Khalatova, F Rasulova, Z Dauter, M R Maurizi, T V Rotanova, A Wlodawer, A Gustchina, 2004, The catalytic domain of Escherichia coli Lon protease has a unique fold and a Ser-Lys dyad in the active site: Journal of Biological Chemistry, v. 279, p. 8140-8148.

Brown,GD, S Gordon, 2003, Fungal beta-glucans and mammalian immunity: Immunity, v. 19, p. 311-315.

Brummer,E, L H Hanson, A Restrepo, D A Stevens, 1989, Intracellular Multiplication of Paracoccidioides-Brasiliensis in Macrophages - Killing and Restriction of Multiplication by Activated Macrophages: Infection and Immunity, v. 57, p. 2289-2294.

Brummer,E, E Castaneda, A Restrepo, 1993, Paracoccidioidomycosis - An Update: Clinical Microbiology Reviews, v. 6, p. 89-117.

Calich,VLG, L M Singervermes, A M Siqueira, E Burger, 1985, Susceptibility and Resistance of Inbred Mice to Paracoccidioides-Brasiliensis: British Journal of Experimental Pathology, v. 66, p. 585-594.

Calich,VLG, L M Singer-Vermes, M Russo, C A C Vaz, E Burger, 1994, Immunogenetics in paracoccidioidomycosis, in M Franco, CS Lacaz, A Restrepo-Moreno, and G Del Negro (eds), Paracoccidioidomycosis: London, Boca Raton, p. 151-173.

Calich,VLG, S S Kashino, 1998, Cytokines produced by susceptible and resistant mice in the course of Paracoccidioides brasiliensis infection: Brazilian Journal of Medical and Biological Research, v. 31, p. 615-623.

Calich,VLG, C A C Vaz, E Burger, 1998, Immunity to Paracoccidioides brasiliensis infection: Research in Immunology, v. 149, p. 407-417.

106

Cano,LE, L M Singervermes, C A C Vaz, M Russo, V L G Calich, 1995, Pulmonary Paracoccidioidomycosis in Resistant and Susceptible Mice - Relationship Among Progression of Infection, Bronchoalveolar Cell Activation, Cellular Immune-Response, and Specific Isotype Patterns: Infection and Immunity, v. 63, p. 1777-1783.

Carmo,JPM, L A Dias-Melicio, S A Calvi, M T S Peracoli, A M V C Soares, 2006, TNF-alpha activates human monocytes for Paracoccidioides brasiliensis killing by an H2O2-dependent mechanism: Medical Mycology, v. 44, p. 363-368.

Carmona,AK, R Puccia, M C F Oliveira, E G Rodrigues, L Juliano, L R Travassos, 1995, Characterization of An Exocellular Serine-Thiol Proteinase Activity in Paracoccidioides-Brasiliensis: Biochemical Journal, v. 309, p. 209-214.

Carvalhaes,MS, W D Dasilva, E G Birman, O A Santanna, P Abrahamsohn, T L Kipnis, 1986, Experimental Paracoccidioidomycosis in High and Low Antibody-Producer Mice .1. Evolution of the Disease, Its Correlation with the Humoral Immune-Response and the Patterns of Tissue Lesions: Annales de l Institut Pasteur-Immunology, v. C137, p. 127-141.

Carvalho,KC, L Ganiko, W L Batista, F V Morais, E R Marques, G H Goldman, M E Franco, R Puccia, 2005, Virulence of Paracoccidioides brasiliensis and gp43 expression in isolates bearing known PbGP43 genotype: Microbes and Infection, v. 7, p. 55-65.

Chaffin,WL, J L Lopez-Ribot, M Casanova, D Gozalbo, J P Martinez, 1998, Cell wall and secreted proteins of Candida albicans: Identification, function, and expression: Microbiology and Molecular Biology Reviews, v. 62, p. 130-+.

Chagas,JR, I Y Hirata, M A Juliano, W Xiong, C Wang, J Chao, L Juliano, E S Prado, 1992, Substrate specificities of tissue kallikrein and T-kininogenase: their possible role in kininogen processing: Biochemistry, v. 31, p. 4969-4974.

Chang,YC, K J Kwon-Chung, 1994, Complementation of a capsule-deficient mutation of Cryptococcus neoformans restores its virulence: Mol Cell Biol., v. 14, p. 4912-4919.

Chaput,N, J Taieb, F Andre, L Zitvogel, 2005, The potential of exosomes in immunotherapy: Expert.Opin.Biol.Ther., v. 5, p. 737-747.

Chequer-Bou-Habib,D, C Daniel-Ribeiro, D M Banic, A C do Valle, B Galvao-Castro, 1989, Polyclonal B cell activation in paracoccidioidomycosis. Polyclonal activation in paracoccidioidomycosis: Mycopathologia, v. 108, p. 89-93.

Clemons,KV, D Feldman, D A Stevens, 1989, Influence of Estradiol on Protein Expression and Methionine Utilization During Morphogenesis of Paracoccidioides-Brasiliensis: Journal of General Microbiology, v. 135, p. 1607-1617.

Corredor,GG, L A Peralta, J H Castano, J S Zuluaga, B Henao, M Arango, A M Tabares, D R Matute, J G Mcewen, A Restrepo, 2005, The naked-tailed armadillo Cabassous centralis (Miller 1899): a new host to Paracoccidioides brasiliensis. Molecular identification of the isolate: Medical Mycology, v. 43, p. 275-280.

Coutinho,ZF, D Silva, M Lazera, V Petri, R M Oliveira, P C Sabroza, B Wanke, 2002, Paracoccidioidomycosis mortality in Brazil (1980-1995): Cad.Saude Publica, v. 18, p. 1441-1454.

Cunha,AF, M V Sousa, S P Silva, R S Jesuino, C M Soares, M S Felipe, 1999, Identification, N-terminal region sequencing and similarity analysis of differentially expressed proteins in Paracoccidioides brasiliensis: Med Mycol., v. 37, p. 115-121.

107

da Silva,MB, A F Marques, J D Nosanchuk, A Casadevall, L R Travassos, C P Taborda, 2006, Melanin in the dimorphic fungal pathogen Paracoccidioides brasiliensis: effects on phagocytosis, intracellular resistance and drug susceptibility: Microbes Infect., v. 8, p. 197-205.

da Silva,SP, M I Borges-Walmsley, I S Pereira, C M Soares, A R Walmsley, M S Felipe, 1999, Differential expression of an hsp70 gene during transition from the mycelial to the infective yeast form of the human pathogenic fungus Paracoccidioides brasiliensis: Mol Microbiol., v. 31, p. 1039-1050.

De Groot,PWJ, A D de Boer, J Cunningham, H L Dekker, L de Jong, K J Hellingwerf, C de Koster, F M Klis, 2004, Proteomic analysis of Candida albicans cell walls reveals covalently bound carbohydrate-active enzymes and adhesins: Eukaryotic Cell, v. 3, p. 955-965.

De Groot,PWJ, A F Ram, F M Klis, 2005, Features and functions of covalently linked proteins in fungal cell walls: Fungal Genetics and Biology, v. 42, p. 657-675.

de Mattos,GD, S R de Almeida, M Mariano, J D Lopes, 2003, Characterization of gp70 and anti-gp70 monoclonal antibodies in Paracoccidioides brasiliensis pathogenesis: Infect.Immun., v. 71, p. 6534-6542.

de Restrepo,FM, M Restrepo, A Restrepo, 1983, Blood groups and HLA antigens in paracoccidioidomycosis: Sabouraudia, v. 21, p. 35-39.

Diez,S, B L Gomez, A Restrepo, R J Hay, A J Hamilton, 2002, Paracoccidioides brasiliensis 87-kilodalton antigen, a heat shock protein useful in diagnosis: Characterization, purification, and detection in biopsy material via immunohistochemistry: Journal of Clinical Microbiology, v. 40, p. 359-365.

Diniz,SN, R Nomizo, P S Cisalpino, M M Teixeira, G D Brown, A Mantovani, S Gordon, L F Reis, A A Dias, 2004, PTX3 function as an opsonin for the dectin-1-dependent internalization of zymosan by macrophages: J.Leukoc.Biol., v. 75, p. 649-656.

Feitosa,LD, P S Cisalpino, M R M dos Santos, R A Mortara, T F Barros, F V Morais, R Puccia, J F da Silveira, Z P De Camargo, 2003, Chromosomal polymorphism, syntenic relationships, and ploidy in the pathogenic fungus Paracoccidioides brasiliensis: Fungal Genetics and Biology, v. 39, p. 60-69.

Felipe,MSS, R V Andrade, S S Petrofeza, A Q Maranhao, F A G Torres, P Albuquerque, F B M Arraes, M Arruda, M O Azevedo, A J Baptista, L A M Bataus, C L Borges, E G Campos, M R Cruz, B S Daher, A Dantas, M A S V Ferreira, G V Ghil, R S A Jesuino, C M Kyaw, L Leitao, C R Martins, L M P Moraes, E O Neves, A M Nicola, E S Alves, J A Parente, M Pereira, M J Pocas-Fonseca, R Resende, B M Ribeiro, R R Saldanha, S C Santos, I Silva-Pereira, M A S Silva, E Silveira, I C Simoes, R B A Soares, D P Souza, M T De Souza, E V Andrade, M A S Xavier, H P Veiga, E J Venancio, M J A Carvalho, A G Oliveira, M K Inoue, N F Almeida, M E M T Walter, C M A Soares, M M Brigido, 2003, Transcriptome characterization of the dimorphic and pathogenic fungus Paracoccidioides brasiliensis by EST analysis: Yeast, v. 20, p. 263-271.

Felipe,MSS, R V Andrade, F B M Arraes, A M Nicola, A Q Maranhao, F A G Torres, I Silva-Pereira, M J Poc as-Fonseca, E G Campos, L M P Moraes, P A Andrade, A H F P Tavares, S S Silva, C M Kyaw, D P Souza, P Network, M Pereira, R S A Jesuino, E V Andrade, J A Parente, G S Oliveira, M S Barbosa, N F Martins, A L Fachin, R S Cardoso, G A S Passos, N F Almeida, M E M T Walter, C M A Soares, M J A Carvalho, M M Brigido, 2005, Transcriptional profiles of the human pathogenic fungus Paracoccidioides brasiliensis in mycelium and yeast cells: Journal of Biological Chemistry, v. 280, p. 24706-24714.

Ferreira,MED, E D Marques, I Malavazi, I Torres, A Restrepo, L R Nunes, R C de Oliveira, M H S Goldman, G H Goldman, 2006, Transcriptome analysis and molecular studies on sulfur metabolism in the human pathogenic fungus Paracoccidioides brasiliensis: Molecular Genetics and Genomics, v. 276, p. 450-463.

108

Fevrier,B, G Raposo, 2004, Exosomes: endosomal-derived vesicles shipping extracellular messages: Curr.Opin.Cell Biol., v. 16, p. 415-421.

Figueiredo,F, L M C Alves, C L Silva, 1993, Tumor-Necrosis-Factor Production In-Vivo and In-Vitro in Response to Paracoccidioides-Brasiliensis and the Cell-Wall Fractions Thereof: Clinical and Experimental Immunology, v. 93, p. 189-194.

Finlay,BB, F Heffron, S Falkow, 1989, Epithelial-Cell Surfaces Induce Salmonella Proteins Required for Bacterial Adherence and Invasion: Science, v. 243, p. 940-943.

Fischer,H, R Glockshuber, 1994, A point mutation within the ATP-binding site inactivates both catalytic functions of the ATP-dependent protease La (Lon) from Escherichia coli: FEBS Lett., v. 356, p. 101-103.

Fliss,AE, J Rao, M W Melville, M E Cheetham, A J Caplan, 1999, Domain requirements of DnaJ-like (Hsp40) molecular chaperones in the activation of a steroid hormone receptor: Journal of Biological Chemistry, v. 274, p. 34045-34052.

Forbes,MS, M L Rennels, E Nelson, 1977, Ultrastructure of pericytes in mouse heart: Am.J.Anat., v. 149, p. 47-70.

Fradin,C, P De Groot, D MacCallum, M Schaller, F Klis, F C Odds, B Hube, 2005, Granulocytes govern the transcriptional response, morphology and proliferation of Candida albicans in human blood: Mol Microbiol., v. 56, p. 397-415.

Franco,M, M R Montenegro, R P Mendes, S A Marques, N L Dillon, N G Mota, 1987, Paracoccidioidomycosis: a recently proposed classification of its clinical forms: Rev Soc.Bras.Med Trop, v. 20, p. 129-132.

Fu,GK, M J Smith, D M Markovitz, 1997, Bacterial protease Lon is a site-specific DNA-binding protein: J.Biol.Chem., v. 272, p. 534-538.

Galvao-Castro,B, J A Sa Ferreira, K F Marzochi, M C Marzochi, S G Coutinho, P H Lambert, 1984, Polyclonal B cell activation, circulating immune complexes and autoimmunity in human american visceral leishmaniasis: Clin.Exp.Immunol., v. 56, p. 58-66.

Ganiko,L, R Puccia, V S Mariano, O A Sant'Anna, E Freymuller, M C Roque-Barreira, L R Travassos, 2007, Paracoccin, an N-acetyl-glucosamine-binding lectin of Paracoccidioides brasiliensis, is involved in fungal growth: Microbes Infect., v. 9, p. 695-703.

Garcia,NM, G Del Negro, E M Heins-Vaccari, N T de Melo, C M de Assis, C S Lacaz, 1993, [Paracoccidioides brasiliensis, a new sample isolated from feces of a penguin (Pygoscelis adeliae)]: Rev Inst Med Trop Sao Paulo, v. 35, p. 227-235.

Gesztesi,JL, R Puccia, L R Travassos, A P Vicentini, J Z de Moraes, M F Franco, J D Lopes, 1996, Monoclonal antibodies against the 43,000 Da glycoprotein from Paracoccidioides brasiliensis modulate laminin-mediated fungal adhesion to epithelial cells and pathogenesis: Hybridoma, v. 15, p. 415-422.

Goldberg,AL, K H Swamy, C H Chung, F S Larimore, 1981, Proteases in Escherichia coli: Methods Enzymol., v. 80 Pt C, p. 680-702.

Goldman,GH, E D Marques, D C D Ribeiro, L A D Bernardes, A C Quiapin, P M Vitorelli, M Savoldi, C P Semighini, R C de Oliveira, L R Nunes, L R Travassos, R Puccia, W L Batista, L E Ferreira, J C Moreira, A P Bogossian, F Tekaia, M P Nobrega, F G Nobrega, M H S Goldman, 2003, Expressed sequence tag analysis of the human pathogen Paracoccidioides brasiliensis yeast phase: Identification of putative homologues of Candida albicans virulence and pathogenicity genes: Eukaryotic Cell, v. 2, p. 34-48.

109

Gomez,BL, J D Nosanchuk, S Diez, S Youngchim, P Aisen, L E Cano, A Restrepo, A Casadevall, A J Hamilton, 2001, Detection of melanin-like pigments in the dimorphic fungal pathogen Paracoccidioides brasiliensis in vitro and during infection: Infection and Immunity, v. 69, p. 5760-5767.

Gonzalez,A, B L Gomez, A Restrepo, A J Hamilton, L E Cano, 2005, Recognition of extracellular matrix proteins by Paracoccidioides brasiliensis yeast cells: Med Mycol., v. 43, p. 637-645.

Goshe,MB, R D Smith, 2003, Stable isotope-coded proteomic mass spectrometry: Current Opinion in Biotechnology, v. 14, p. 101-109.

Grant,CM, L P Collinson, J H Roe, I W Dawes, 1996, Yeast glutathione reductase is required for protection against oxidative stress and is a target gene for yAP-1 transcriptional regulation: Mol Microbiol., v. 21, p. 171-179.

Griffith,KL, I M Shah, R E Wolf, 2004, Proteolytic degradation of Escherichia coli transcription activators SoxS and MarA as the mechanism for reversing the induction of the superoxide (SoxRS) and multiple antibiotic resistance (Mar) regulons: Molecular Microbiology, v. 51, p. 1801-1816.

Grose,E, J R Tamsitt, 1965, Paracoccidioides brasiliensis recovered from the intestinal tract of three bats (Artibeus lituratus) in Colombia, S.A: Sabouraudia, v. 4, p. 124-125.

Harding,C, J Heuser, P Stahl, 1984, Endocytosis and intracellular processing of transferrin and colloidal gold-transferrin in rat reticulocytes: demonstration of a pathway for receptor shedding: Eur.J.Cell Biol., v. 35, p. 256-263.

Harris,TO, D W Shelver, J F Bohnsack, C E Rubens, 2003, A novel streptococcal surface protease promotes virulence, resistance to opsonophagocytosis, and cleavage of human fibrinogen: J.Clin.Invest, v. 111, p. 61-70.

Hebeler-Barbosa,F, M R Montenegro, E Bagagli, 2003, Virulence profiles of ten Paracoccidioides brasiliensis isolates obtained from armadillos (Dasypus novemcinctus): Medical Mycology, v. 41, p. 89-96.

Hedstrom,L, 2002, Serine protease mechanism and specificity: Chem.Rev, v. 102, p. 4501-4524.

Ishii,Y, F Amano, 2001, Regulation of SulA cleavage by Lon protease by the C-terminal amino acid of SulA, histidine: Biochemical Journal, v. 358, p. 473-480.

Izacc,SMS, F J Gomez, R S A Jesuino, C A Fonseca, M S S Felipe, G S Deepe, C M A Soares, 2001, Molecular cloning, characterization and expression of the heat shock protein 60 gene from the human pathogenic fungus Paracoccidioides brasiliensis: Medical Mycology, v. 39, p. 445-455.

Jesuino,RSA, M O Azevedo, M S S Felipe, M Pereira, C M D Soares, 2002, Characterization of a chaperone ClpB homologue of Paracoccidioides brasiliensis: Yeast, v. 19, p. 963-972.

Johnson,WD, C M Lang, 1977, Paracoccidioidomycosis (South American blastomycosis) in a squirrel monkey (Saimiri sciureus): Vet.Pathol., v. 14, p. 368-371.

Johnstone,RM, M Adam, J R Hammond, L Orr, C Turbide, 1987, Vesicle formation during reticulocyte maturation. Association of plasma membrane activities with released vesicles (exosomes): J.Biol.Chem., v. 262, p. 9412-9420.

Johnstone,RM, 2006, Exosomes biological significance: A concise review: Blood Cells Mol Dis., v. 36, p. 315-321.

110

Jubete,Y, M R Maurizi, S Gottesman, 1996, Role of the heat shock protein DnaJ in the Lon-dependent degradation of naturally unstable proteins: Journal of Biological Chemistry, v. 271, p. 30798-30803.

Kanetsuna,F, L M Carbonel, R E Moreno, J Rodriguez, 1969, Cell Wall Composition of Yeast and Mycelial Forms of Paracoccidioides Brasiliensis: Journal of Bacteriology, v. 97, p. 1036-&.

Kashino,SS, V L G Calich, L M Singervermes, P A Abrahamsohn, E Burger, 1987, Growth-Curves, Morphology and Ultrastructure of 10 Paracoccidioides-Brasiliensis Isolates: Mycopathologia, v. 99, p. 119-128.

Kindler,V, A P Sappino, G E Grau, P F Piguet, P Vassalli, 1989, The Inducing Role of Tumor Necrosis Factor in the Development of Bactericidal Granulomas During Bcg Infection: Cell, v. 56, p. 731-740.

Kominami,E, H Hoffschulte, H Holzer, 1981, Purification and Properties of Proteinase-B from Yeast: Biochimica et Biophysica Acta, v. 661, p. 124-135.

Kuehn,MJ, N C Kesty, 2005, Bacterial outer membrane vesicles and the host-pathogen interaction: Genes Dev., v. 19, p. 2645-2655.

Kurita,N, M Oarada, M Miyaji, E Ito, 2000, Effect of cytokines on antifungal activity of human polymorphonuclear leucocytes against yeast cells of Paracoccidioides brasiliensis: Medical Mycology, v. 38, p. 177-182.

Kwon-Chung,KJ, I Polacheck, T J Popkin, 1982, Melanin-lacking mutants of Cryptococcus neoformans and their virulence for mice: J.Bacteriol., v. 150, p. 1414-1421.

Lacaz,CS, 1994, Mycological diagnosis, in M Franco, CS Lacaz, A Restrepo-Moreno, and G Del Negro (eds), Paracoccidioidomycosis: London, Boca Raton, p. 339-344.

Lacaz,CS, 1994, Historical evolution of the knowledge on paracoccidioidomycosis and its etiologic agent, Paracoccidioides brasiliensis, in M Franco, CS Lacaz, A Restrepo-Moreno, and G Del Negro (eds), Paracoccidioidomycosis: London, Boca Raton, p. 1-11.

Lacaz,CS, 1994, Paracoccidioides brasiliensis: morphology; evolutionary cycle; maintenance during saprophytic life; biology, virulence, taxonomy., in M Franco, CS Lacaz, A Restrepo-Moreno, and G Del Negro (eds), Paracoccidioidomycosis: London, Boca Raton, p. 13-25.

Lacerda,GB, B Arcegomez, F Q Telles, 1988, Increased Frequency of Hla-B40 in Patients with Paracoccidioidomycosis: Journal of Medical and Veterinary Mycology, v. 26, p. 253-256.

Lancaster,GI, M A Febbraio, 2005, Exosome-dependent trafficking of HSP70: a novel secretory pathway for cellular stress proteins: J.Biol.Chem., v. 280, p. 23349-23355.

Langer,T, W Neupert, 1996, Regulated protein degradation in mitochondria: Experientia, v. 52, p. 1069-1076.

Levery,SB, M S Toledo, A H Straus, H K Takahashi, 1998, Structure elucidation of sphingolipids from the mycopathogen Paracoccidioides brasiliensis: An immunodominant beta-galactofuranose residue is carried by a novel glycosylinositol phosphorylceramide antigen: Biochemistry, v. 37, p. 8764-8775.

Li,XB, Z R Zhang, H J Schluesener, S Q Xu, 2006, Role of exosomes in immune regulation: J.Cell Mol Med, v. 10, p. 364-375.

111

Lopes,JD, M C R Mouracampos, A P Vicentini, J L Gesztesi, W Desouza, Z P Camargo, 1994, Characterization of Glycoprotein Gp43, the Major Laminin-Binding Protein of Paracoccidioides-Brasiliensis: Brazilian Journal of Medical and Biological Research, v. 27, p. 2309-2313.

Mahadevan,PR, E L Tatum, 1967, Localization of structural polymers in the cell wall of Neurospora crassa: J.Cell Biol., v. 35, p. 295-302.

Marques,AF, M B da Silva, M A Juliano, L R Travassos, C P Taborda, 2006, Peptide immunization as an adjuvant to chemotherapy in mice challenged intratracheally with virulent yeast cells of Paracoccidioides brasiliensis: Antimicrob.Agents Chemother., v. 50, p. 2814-2819.

Marques,ER, M E S Ferreira, R D Drummond, J M Felix, M Menossi, M Savoldi, L R Travassos, R Puccia, W L Batista, K C Carvalho, M H S Goldman, G H Goldman, 2004, Identification of genes preferentially expressed in the pathogenic yeast phase of Paracoccidioides brasiliensis, using suppression subtraction hybridization and differential macroarray analysis: Molecular Genetics and Genomics, v. 271, p. 667-677.

Matsuo,AL, A K Carmona, L S Silva, C E Cunha, E S Nakayasu, I C Almeida, M A Juliano, R Puccia, 2007, C-Npys (S-3-nitro-2-pyridinesulfenyl) and peptide derivatives can inhibit a serine-thiol proteinase activity from Paracoccidioides brasiliensis: Biochem.Biophys.Res.Commun., v. 355, p. 1000-1005.

Matute,DR, J G Mcewen, B A Montes, G San Blas, E Bagagli, J T Rauscher, A Restrepo, F Morais, G Nino-Vega, J W Taylor, 2006, Cryptic speciation and recombination in the fungus Paracoccidioides brasiliensis as revealed by gene genealogies: Molecular Biology and Evolution, v. 23, p. 65-73.

Mcewen,JG, V Bedoya, M M Patino, M E Salazar, M A Restrepo, 1987, Experimental Murine Paracoccidiodomycosis Induced by the Inhalation of Conidia: Journal of Medical and Veterinary Mycology, v. 25, p. 165-175.

Mckerrow,JH, E Sun, P J Rosenthal, J Bouvier, 1993, The Proteases and Pathogenicity of Parasitic Protozoa: Annual Review of Microbiology, v. 47, p. 821-853.

Mendes-Giannini,MJS, P F Andreotti, L R Vincenzi, J L M da Silva, H L Lenzi, G Benard, R Zancope-Oliveira, H L D Guedes, C P Soares, 2006, Binding of extracellular matrix proteins to Paracoccidioides brasiliensis: Microbes and Infection, v. 8, p. 1550-1559.

Mendes,RP, R Negroni, A Arechavala, 1994, Treatment and control of cure, in M Franco, CS Lacaz, A Restrepo-Moreno, and G Del Negro (eds), Paracoccidioidomycosis: London, Boca Raton, p. 373-387.

Morais,FV, T F Barros, M K Fukada, P S Cisalpino, R Puccia, 2000, Polymorphism in the gene coding for the immunodominant antigen gp43 from the pathogenic fungus Paracoccidioides brasiliensis: Journal of Clinical Microbiology, v. 38, p. 3960-3966.

Moscardi-Bacchi,M, E Brummer, D A Stevens, 1994, Support of Paracoccidioides brasiliensis multiplication by human monocytes or macrophages: inhibition by activated phagocytes: J.Med Microbiol., v. 40, p. 159-164.

Mota,NG, M T Rezkallah-Iwasso, M T Peracoli, R C Audi, R P Mendes, J Marcondes, S A Marques, N L Dillon, M F Franco, 1985, Correlation between cell-mediated immunity and clinical forms of paracoccidioidomycosis: Trans.R.Soc.Trop Med Hyg., v. 79, p. 765-772.

Muller,A, W Saenger, 1993, Studies on the Inhibitory-Action of Mercury Upon Proteinase-K: Journal of Biological Chemistry, v. 268, p. 26150-26154.

112

Munro,CA, K Winter, A Buchan, K Henry, J M Becker, A J P Brown, C E Bulawa, N A R Gow, 2001, Chs1 of Candida albicans is an essential chitin synthase required for synthesis of the septum and for cell integrity: Molecular Microbiology, v. 39, p. 1414-1426.

Munro,CA, S Selvaggini, I de Bruijn, L Walker, M D Lenardon, B Gerssen, S Milne, A J P Brown, N A R Gow, 2007, The PKC, HOG and Ca2+ signalling pathways co-ordinately regulate chitin synthesis in Candida albicans: Molecular Microbiology, v. 63, p. 1399-1413.

Naglik,JR, S J Challacombe, B Hube, 2003, Candida albicans secreted aspartyl proteinases in virulence and pathogenesis: Microbiology and Molecular Biology Reviews, v. 67, p. 400-+.

Nascimento,FRF, V L G Calich, D Rodriguez, M Russo, 2002, Dual role for nitric oxide in paracoccidioidomycosis: Essential for resistance, but overproduction associated with susceptibility: Journal of Immunology, v. 168, p. 4593-4600.

Nemecek,JC, M Wuthrich, B S Klein, 2006, Global control of dimorphism and virulence in fungi: Science, v. 312, p. 583-588.

Nino-Vega,GA, C A Munro, G San Blas, G W Gooday, N A R Gow, 2000, Differential expression of chitin synthase genes during temperature-induced dimorphic transitions in Paracoccidioides brasiliensis: Medical Mycology, v. 38, p. 31-39.

Nombela,C, C Gil, W L Chaffin, 2006, Non-conventional protein secretion in yeast: Trends Microbiol., v. 14, p. 15-21.

Nunes,LR, d O Costa, D B Leite, V S da Silva, M E dos Reis, M E Silva Ferreira, D C Ribeiro, L A Souza Bernardes, M H Goldman, R Puccia, L R Travassos, W L Batista, M P Nobrega, F G Nobrega, D Y Yang, C A Braganca Pereira, G H Goldman, 2005, Transcriptome analysis of Paracoccidioides brasiliensis cells undergoing mycelium-to-yeast transition: Eukaryot.Cell, v. 4, p. 2115-2128.

Ogrydziak,DM, 1993, Yeast Extracellular Proteases: Critical Reviews in Biotechnology, v. 13, p. 1-55.

Ortiz-Ortiz,L, D E Parks, M Rodriguez, W O Weigle, 1980, Polyclonal B lymphocyte activation during Trypanosoma cruzi infection: J.Immunol., v. 124, p. 121-126.

Pan,BT, R Blostein, R M Johnstone, 1983, Loss of the transferrin receptor during the maturation of sheep reticulocytes in vitro. An immunological approach: Biochem.J., v. 210, p. 37-47.

Pan,BT, K Teng, C Wu, M Adam, R M Johnstone, 1985, Electron microscopic evidence for externalization of the transferrin receptor in vesicular form in sheep reticulocytes: J.Cell Biol., v. 101, p. 942-948.

Pinto,AR, R Puccia, S N Diniz, M F Franco, L R Travassos, 2000, DNA-based vaccination against murine paracoccidioidomycosis using the gp43 gene from Paracoccidioides brasiliensis: Vaccine, v. 18, p. 3050-3058.

Posadas,A, 1892, Un nuevo caso de micosis fungoidea con psorospermias: Ann Circ Med Argent, v. 15, p. 585-597.

Puccia,R, S Schenkman, P A J Gorin, L R Travassos, 1986, Exocellular Components of Paracoccidioides-Brasiliensis - Identification of A Specific Antigen: Infection and Immunity, v. 53, p. 199-206.

Puccia,R, L R Travassos, 1991, 43-Kilodalton Glycoprotein from Paracoccidioides-Brasiliensis - Immunochemical Reactions with Sera from Patients with Paracoccidioidomycosis, Histoplasmosis, Or Jorge Lobos Disease: Journal of Clinical Microbiology, v. 29, p. 1610-1615.

113

Puccia,R, A K Carmona, J L Gesztesi, L Juliano, L R Travassos, 1998, Exocellular proteolytic activity of Paracoccidioides brasiliensis: cleavage of components associated with the basement membrane: Medical Mycology, v. 36, p. 345-348.

Puccia,R, M A Juliano, L Juliano, L R Travassos, A K Carmona, 1999, Detection of the basement membrane-degrading proteolytic activity of Paracoccidioides brasiliensis after SDS-PAGE using agarose overlays containing Abz-MKALTLQ-EDDnp: Brazilian Journal of Medical and Biological Research, v. 32, p. 645-649.

Quah,BJ, H C O'Neill, 2005, The immunogenicity of dendritic cell-derived exosomes: Blood Cells Mol Dis., v. 35, p. 94-110.

Rappleye,CA, J T Engle, W E Goldman, 2004, RNA interference in Histoplasma capsulatum demonstrates a role for alpha-(1,3)-glucan in virulence: Mol Microbiol., v. 53, p. 153-165.

Rappleye,CA, L G Eissenberg, W E Goldman, 2007, Histoplasma capsulatum alpha-(1,3)-glucan blocks innate immune recognition by the beta-glucan receptor: Proc.Natl.Acad.Sci.U.S.A, v. 104, p. 1366-1370.

Rawlings,ND, F R Morton, A J Barrett, 2006, MEROPS: the peptidase database: Nucleic Acids Res., v. 34, p. D270-D272.

Restrepo-Moreno,A, 1994, Ecology of Paracoccidioides brasiliensis, in M Franco, CS Lacaz, A Restrepo-Moreno, and G Del Negro (eds), Paracoccidioidomycosis: London, Boca Raton, p. 121-130.

Restrepo,A, I Gomez, M A Robledo, 1982, [Efficacy of ketoconazole in patients with recurrent paracoccidioidomycosis]: Rev Inst Med Trop Sao Paulo, v. 24, p. 173-179.

Restrepo,A, 2000, Morphological aspects of Paracoccidioides brasiliensis in lymph nodes: implications for the prolonged latency of paracoccidioidomycosis?: Medical Mycology, v. 38, p. 317-322.

Restrepo,A, J G Mcewen, E Castaneda, 2001, The habitat of Paracoccidioides brasiliensis: how far from solving the riddle?: Medical Mycology, v. 39, p. 233-241.

Ricci,G, F T Mota, A Wakamatsu, R C Serafim, R C Borra, M Franco, 2004, Canine paracoccidioidomycosis: Med Mycol., v. 42, p. 379-383.

Robertson,GT, M E Kovach, C A Allen, T A Ficht, R M Roop, 2000, The Brucella abortus Lon functions as a generalized stress response protease and is required for wild-type virulence in BALB/c mice: Molecular Microbiology, v. 35, p. 577-588.

Rodrigues,EG, L R Travassos, 1994, Nature of the reactive epitopes in Paracoccidioides brasiliensis polysaccharide antigen: J.Med Vet.Mycol., v. 32, p. 77-81.

Rodrigues,ML, L Nimrichter, D L Oliveira, S Frases, K Miranda, O Zaragoza, M Alvarez, A Nakouzi, M Feldmesser, A Casadevall, 2007, Vesicular polysaccharide export in Cryptococcus neoformans is a eukaryotic solution to the problem of fungal trans-cell wall transport: Eukaryot.Cell, v. 6, p. 48-59.

Rowley,N, C Pripbuus, B Westermann, C Brown, E Schwarz, B Barrell, W Neupert, 1994, Mdj1P, A Novel Chaperone of the Dnaj Family, Is Involved in Mitochondrial Biogenesis and Protein-Folding: Cell, v. 77, p. 249-259.

Ruiz-Herrera,J, M V Elorza, E Valentin, R Sentandreu, 2006, Molecular organization of the cell wall of Candida albicans and its relation to pathogenicity: FEMS Yeast Res., v. 6, p. 14-29.

114

Russell,WK, Z Y Park, D H Russell, 2001, Proteolysis in mixed organic-aqueous solvent systems: Applications for peptide mass mapping using mass spectrometry: Analytical Chemistry, v. 73, p. 2682-2685.

Salem-Izacc,SM, R S Jesuino, W A Brito, M Pereira, M S Felipe, C M Soares, 1997, Protein synthesis patterns of Paracoccidiodes brasiliensis isolates in stage-specific forms and during cellular differentiation: J.Med Vet.Mycol., v. 35, p. 205-211.

Salyers,A, D Witt, 1994, Virulence factors that promote colonization, in A Salyers and D Witt (eds), Bacterial pathogenesis: a molecular approach: Washington, D.C., ASM press, p. 30-46.

San Blas,G, 1993, Biochemical and physiological aspects in the dimorphism of Paracoccidioides brasiliensis: Arch.Med Res., v. 24, p. 267-268.

San Blas,G, F Sorais, G Nino-Vega, C Mendez, F San Blas, 1998, Cytosolic neutral proteinases of Paracoccidioides brasiliensis: Current Microbiology, v. 37, p. 141-143.

San Blas,G, G Nino-Vega, T Iturriaga, 2002, Paracoccidioides brasiliensis and paracoccidioidomycosis: Molecular approaches to morphogenesis, diagnosis, epidemiology, taxonomy and genetics: Medical Mycology, v. 40, p. 225-242.

Sano,A, R Tanaka, K Yokoyama, M Franco, E Bagagli, M R Montenegro, Y Mikami, M Miyaji, K Nishimura, 1998, Comparison between human and armadillo Paracoccidioides brasiliensis isolates by random amplified polymorphic DNA analysis: Mycopathologia, v. 143, p. 165-169.

Sano,A, J Defaveri, R Tanaka, K Yokoyama, N Kurita, M Franco, K I R Coelho, E Bagagli, M R Montenegro, M Miyaji, K Nishimura, 1998, Pathogenicities and GP43kDa gene of three Paracoccidioides brasiliensis isolates originated from a nine-banded armadillo (Dasypus novemcinctus): Mycopathologia, v. 144, p. 61-65.

Saraiva,EC, A Altemani, M F Franco, C S Unterkircher, Z P Camargo, 1996, Paracoccidioides brasiliensis-gp43 used as paracoccidioidin: J.Med Vet.Mycol., v. 34, p. 155-161.

Savel'ev,AS, L A Novikova, I E Kovaleva, V N Luzikov, W Neupert, T Langer, 1998, ATP-dependent proteolysis in mitochondria. m-AAA protease and PIM1 protease exert overlapping substrate specificities and cooperate with the mtHsp70 system: J.Biol.Chem., v. 273, p. 20596-20602.

Schechter,I, A Berger, 1967, On the size of the active site in proteases. I. Papain: Biochem.Biophys.Res.Commun., v. 27, p. 157-162.

Segura,E, S Amigorena, C Thery, 2005, Mature dendritic cells secrete exosomes with strong ability to induce antigen-specific effector immune responses: Blood Cells Mol Dis., v. 35, p. 89-93.

Shikanai-Yasuda,MA, F Q Telles Filho, R P Mendes, A L Colombo, M L Moretti, 2006, Paracoccidioidomicose: Revista da Sociedade Brasileira de Medicina Tropical, v. 39, p. 297-310.

Silva,CL, R A Fazioli, 1985, A Paracoccidioides brasiliensis polysaccharide having granuloma-inducing, toxic and macrophage-stimulating activity: J.Gen.Microbiol., v. 131, p. 1497-1501.

Smith,CK, T A Baker, R T Sauer, 1999, Lon and Clp family proteases and chaperones share homologous substrate-recognition domains: Proc.Natl.Acad.Sci.U.S.A, v. 96, p. 6678-6682.

Smith,EL, F S Markland, C B Kasper, R J DeLange, M Landon, W H Evans, 1966, The complete amino acid sequence of two types of subtilisin, BPN' and Carlsberg: J.Biol.Chem., v. 241, p. 5974-5976.

115

Smithson,SL, I C Paterson, A M Bailey, S E Screen, B A Hunt, B D Cobb, R M Cooper, A K Charnley, J M Clarkson, 1995, Cloning and Characterization of A Gene Encoding A Cuticle-Degrading Protease from the Insect Pathogenic Fungus Metarhizium-Anisopliae: Gene, v. 166, p. 161-165.

Sprent,J, 2005, Direct stimulation of naive T cells by antigen-presenting cell vesicles: Blood Cells Mol Dis., v. 35, p. 17-20.

Stahlberg,H, E Kutejova, K Suda, B Wolpensinger, A Lustig, G Schatz, A Engel, C K Suzuki, 1999, Mitochondrial Lon of Saccharomyces cerevisiae is a ring-shaped protease with seven flexible subunits: Proceedings of the National Academy of Sciences of the United States of America, v. 96, p. 6787-6790.

Straus,AH, E Freymuller, L R Travassos, H K Takahashi, 1996, Immunochemical and subcellular localization of the 43 kDa glycoprotein antigen of Paracoccidioides brasiliensis with monoclonal antibodies: J.Med Vet.Mycol., v. 34, p. 181-186.

Szabo,A, R Korszun, F U Hartl, J Flanagan, 1996, A zinc finger-like domain of the molecular chaperone DnaJ is involved in binding to denatured protein substrates: EMBO J., v. 15, p. 408-417.

Taborda,CP, M A Juliano, R Puccia, M Franco, L R Travassos, 1998, Mapping of the T-cell epitope in the major 43-kilodalton glycoprotein of Paracoccidioides brasiliensis which induces a Th-1 response protective against fungal infection in BALB/c mice: Infection and Immunity, v. 66, p. 786-793.

Taborda,CP, J Rivera, O Zaragoza, A Casadevall, 2003, More is not necessarily better: prozone-like effects in passive immunization with IgG: J.Immunol., v. 170, p. 3621-3630.

Tavares,AH, S S Silva, A Dantas, E G Campos, R V Andrade, A Q Maranhao, M M Brigido, D G Passos-Silva, A L Fachin, S M Teixeira, G A Passos, C M Soares, A L Bocca, M J Carvalho, I Silva-Pereira, M S Felipe, 2007, Early transcriptional response of Paracoccidioides brasiliensis upon internalization by murine macrophages: Microbes Infect., v. 9, p. 583-590.

Tsilibaris,V, G Maenhaut-Michel, L Van Melderen, 2006, Biological roles of the Lon ATP-dependent protease: Research in Microbiology, v. 157, p. 701-713.

Van Dyck,L, D A Pearce, F Sherman, 1994, PIM1 encodes a mitochondrial ATP-dependent protease that is required for mitochondrial function in the yeast Saccharomyces cerevisiae: J.Biol.Chem., v. 269, p. 238-242.

Van Dyck,L, T Langer, 1999, ATP-dependent proteases controlling mitochondrial function in the yeast Saccharomyces cerevisiae: Cell Mol Life Sci., v. 56, p. 825-842.

Van Melderen,L, M H Thi, P Lecchi, S Gottesman, M Couturier, M R Maurizi, 1996, ATP-dependent degradation of CcdA by Lon protease. Effects of secondary structure and heterologous subunit interactions: J.Biol.Chem., v. 271, p. 27730-27738.

Venancio,EJ, C M Kyaw, C V Mello, S P Silva, C M Soares, M S Felipe, I Silva-Pereira, 2002, Identification of differentially expressed transcripts in the human pathogenic fungus Paracoccidioides brasiliensis by differential display: Med Mycol., v. 40, p. 45-51.

Vicentini,AP, J L Gesztesi, M F Franco, W Desouza, J Z Demoraes, L R Travassos, J D Lopes, 1994, Binding of Paracoccidioides-Brasiliensis to Laminin Through Surface Glycoprotein Gp43 Leads to Enhancement of Fungal Pathogenesis: Infection and Immunity, v. 62, p. 1465-1469.

Voos,W, K Rottgers, 2002, Molecular chaperones as essential mediators of mitochondrial biogenesis: Biochimica et Biophysica Acta-Molecular Cell Research, v. 1592, p. 51-62.

116

Wagner,I, H Arlt, L Van Dyck, T Langer, W Neupert, 1994, Molecular chaperones cooperate with PIM1 protease in the degradation of misfolded proteins in mitochondria: EMBO J., v. 13, p. 5135-5145.

Walsh,P, D Bursac, Y C Law, D Cyr, T Lithgow, 2004, The J-protein family: modulating protein assembly, disassembly and translocation: Embo Reports, v. 5, p. 567-571.

Walter,S, J Buchner, 2002, Molecular chaperones--cellular machines for protein folding: Angew.Chem.Int.Ed Engl., v. 41, p. 1098-1113.

Wanke,B, A T Londero, 1994, Epidemiology and Paracoccidioidomycosis infection, in M Franco, CS Lacaz, A Restrepo-Moreno, and G Del Negro (eds), Paracoccidioidomycosis: London, Boca Raton, p. 109-120.

Wright,CS, R A Alden, J Kraut, 1969, Structure of subtilisin BPN' at 2.5 angstrom resolution: Nature, v. 221, p. 235-242.

Wu,AL, W S Moye-Rowley, 1994, GSH1, which encodes gamma-glutamylcysteine synthetase, is a target gene for yAP-1 transcriptional regulation: Mol Cell Biol., v. 14, p. 5832-5839.

Yoneda,A, T L Doering, 2006, A eukaryotic capsular polysaccharide is synthesized intracellularly and secreted via exocytosis: Mol Biol.Cell, v. 17, p. 5131-5140.

Yun,DJ, Y Zhao, J M Pardo, M L Narasimhan, B Damsz, H Lee, L R Abad, M P Durzo, P M Hasegawa, R A Bressan, 1997, Stress proteins on the yeast cell surface determine resistance to osmotin, a plant antifungal protein: Proceedings of the National Academy of Sciences of the United States of America, v. 94, p. 7082-7087.

Zhu,X, J Gibbons, J Garcia-Rivera, A Casadevall, P R Williamson, 2001, Laccase of Cryptococcus neoformans is a cell wall-associated virulence factor: Infect.Immun., v. 69, p. 5589-5596.

117

Anexo 1: Resultados complementares sobre a PbST

1. Purificação da PbST

A purificação da serino-tiol ativa tem sido impedida por ocorrer em baixas

concentrações proteicas, ter tendência à autólise e agregação com componentes

polissacarídeos de alto peso molecular. Anteriormente ao desenvolvimento desta tese,

várias estratégias aplicadas com esse fim tiveram sucesso parcial. Desta forma, um dos

objetivos deste trabalho foi testar outras técnicas no sentido de conseguir preparações

suficientemente enriquecidas para a proteinase que pudessem ser analisadas quanto à

composição de aminoácidos. Nos trabalhos apresentados anteriormente, utilizamos dois

tipos de preparações enzimáticas, a saber, fração 3 eluída de uma coluna de interação

hidrofóbica de Phenyl Superose (Matsuo et al., 2006) e a fração 4 resultante de

cromatografia de afinidade em uma coluna de p-aminometilbenzamidina (pABA),

utilizada apenas recentemente no laboratório (Matsuo et al., 2007). Outras estratégias

usadas na purificação da PbST estão comentadas abaixo e foram extensivamente

detalhadas em relatórios científicos apresentados à FAPESP.

a) Cromatografia de fase reversa em coluna de Sephasil Protein C4: Frações

eluídas da coluna de Phenyl-Superose correspondentes ao pico de atividade proteolítica

(Matsuo et al., 2006) eram fracionados por f.p.l.c, em ÄKTA, em uma coluna de fase

reversa Sephasil Protein C4 (Amersham Biociences) que continha 1,66 ml de resina

equilibrada com 0,1% de TFA (ácido triclorofluoroacético). As frações eram eluídas

através de um gradiente de 0 a 90% acetonitrila em 0,1% de TFA, rapidamente

118

congeladas em gelo-seco, secas em speed-vac, e estocadas a -20ºC. Antes de testadas, as

frações eram ressuspendidas em 100µL de Tris 50 mM, pH 8,0, e então analisadas quanto

ao padrão de bandas e atividade frente ao substrato peptídico específico Abz-

MKALTLQ-EDDnp. Tínhamos consciência de que uma cromatografia em condições

drásticas com solventes orgânicos poderia resultar na morte da atividade enzimática. No

entanto, foi encontrada uma intensa atividade nas frações 5 (pico fino e homogêneo) e 12.

Infelizmente, no entanto, esse resultado não se repetiu em qualquer das 8 vezes

subsequentes e a estratégia foi abandonada.

b) Cromatografia em colunas de troca-iônica: Frações eluídas da coluna de

Phenyl-Superose correspondentes ao pico de atividade hidrolítica sobre Abz-

MKALTLQ-EDDnp (Matsuo et al., 2006) foram fracionadas em colunas trocadoras

catiônicas de Sepharose SP (forte) e Sepharose CM (fraca), e em resinas trocadoras

aniônicas Sepharose Q (forte) e Sepharose DEAE (fraca), da Amersham Pharmacia

Biotech, de acordo com os procedimentos sugeridos pelo fabricante. Os ensaios de

hidrólise de substrato peptídico de fluorescência apagada Abz-MKALTLQ-EDDnp

mostraram que a recuperação da atividade foi melhor com as resinas catiônicas Sepharose

DEAE (28,4%) e Sepharose Q (26,5%), porém o rendimento foi baixo e tornou este tipo

de cromatografia inviável.

c) Precipitação de sobrenadante total com etanol: para enriquecer a atividade

da PbST, sobrenadantes de cultura ativos de P. brasiliensis foram precipitados em série

com etanol nas proporções de 20%, 40%, 60%, 80% ou 95%. As proteínas do

sobrenadante concentraram-se nas frações 20% a 40% e 40% a 60% de etanol. Quando

testadas quanto à clivagem de fibronectina (Fn) por 30 min a 37oC, verificou-se que a

119

fração 20% a 40%, não apresentou atividade proteolítica, enquanto as frações 40% a 60%

e 60% a 80% clivaram a Fn totalmente. Coincidentemente, estas mesmas frações

continham um componente de alta massa molecular, aparentemente a galactomanana do

fungo com efeito modulatório sobre a PbST (Matsuo et al., 2006). A atividade do

precipitado com 40% a 60% de etanol foi inibida por PMSF e pHMB. Não foi dada

sequência aos experimentos de precipitação com etanol porque a coluna de pABA foi

adquirida nesse período e mostrou boa resolução (Matsuo et al., 2007). No entanto, uma

preparação ativa e com reduzido conteúdo proteico resultante da precipitação com 50 a

80% de etanol foi analisada quanto à composição peptídica, com resultados

desanimadores.

2. Resultados de experimentos sugeridos pela banca de

qualificação

Por ocasião do exame de qualificação, dois experimentos foram sugeridos pela

banca, os quais foram realizados como descrito a seguir:

1. Obter anticorpos policlonais reativos com alguma subtilina-tiol e verificar reação

cruzada com a PbST. Para tal, inoculamos uma coelha com 10 µg de proteinase K

(Sigma) purificada do fungo Tritirachium album via intradérmica com adjuvante de

Freund (50 µL). Após duas imunizações, o soro foi recolhido e testado em ELISA,

“Western blot” ou “Dot blot” contra proteinase K, sobrenadante total de P. brasiliensis

ativo para PbST, fração eluída ou não-ligada em coluna de pABA. Porém os resultados

120

foram positivos apenas para a proteinase K, mostrando que o anticorpo não foi capaz de

reagir cruzadamente com a serino-tiol nas condições testadas.

2. Verificar se há anticorpos no soro de paciente com PCM capazes de inibir a atividade

proteolítica, uma vez que a proteinase é secretada. Para tal, imunoglobulinas de um

paciente com PCM foram purificadas em coluna de afinidade de proteína A e testadas em

diversas concentrações (0,5, 0.25 ou 0.1 µg/ µL) quanto à capacidade de inibir a clivagem

de fibronectina pela PbST eluída de coluna pABA. Em nenhuma concentração testada as

imunoglobulinas foram capazes de inibir a ação da proteinase, sugerindo que o soro de

paciente não continha anticorpos anti-PbST ou nas condições testadas os anticorpos não

foram capazes de inibir ou interagir com os epitopos da molécula ao ponto de interferir

com o sítio ativo.

121

Anexo 2: Resultados complementares sobre a PbLon

1. Expressão da PbLon em sistema heterólogo

Um dos objetivos do projeto de doutorado foi a expressão da enzima PbLon

recombinante ativa para estudos de especificidade utilizando peptídeos de fluorescência

apagada do tipo O-aminobenzoil-peptidil-etilenodiaminobenzidina (Abz-EDDnp), os

quais poderiam levar à detecção de inibidores específicos da enzima. Para tal, a

construção deveria conter os fragmentos de cDNA correspondentes aos sítios ativos da

enzima. No trabalho apresentado anteriormente (Batista et al., 2006) obteve-se a

expressão em bactéria de uma porção N-terminal da PbLon, na forma de corpos de

inclusão, a qual foi essencial para obtenção de anticorpos policlonais de coelho usados

em experimentos de localização celular. Foram realizadas, entretanto, inúmeras tentativas

frustadas de expressão da molécula solúvel e ativa, cujas construções e sistemas estão

esquematizamos na Figura 1.

122

Figura 1. Construções contendo o gene da PbLON para expressão em sistema procarioto (E. coli) ou

sistema eucarioto (Pichia pastoris e S. cerevisiae).

Expressão de PbLon recombinante truncada no N-terminal

A primeira tentativa de expressão da PbLon foi realizada pela Dra. Tânia Fraga

Barros utilizando o cDNA completo da proteinase inserido no vetor pHIS1, porém

nenhuma banda correspondente à proteína foi observada. A segunda construção privou a

molécula dos primeiros 178 aminoácidos do N-terminal, o que não comprometeu os sítios

catalíticos. Para tal, usou-se um plasmídeo pHIS1Lon do laboratório (Barros, 2001), onde

o cDNA do gene PbLON truncado no nt 151 foi inserido no sítio EcoRI de pHIS1. Esse

plasmídeo foi presentemente clivado com BamHI (nt 537) e re-ligado no sítio BamHI do

vetor. Essa construção foi usada na transformação das bactérias de expressão

BL21(DE3), BL21(DE3) pLysS resistente ao clorofenicol, BL21(DE3) trxB resistente à

kanamicina e BL21(DE3) Rosetta.

123

As bactérias induzidas com IPTG por 3 horas foram lisadas e as proteínas

submetidas à cromatografia em coluna de Ni-NTA (Qiagen). As proteínas eluídas foram

analisadas em “Western blot” com anti-rPbLon onde foram vizualizadas duas bandas

específicas, uma de 32 kDa e outra de 80 kDa (figura 2), quando o peso molecular

esperado era de 95 a 100 kDa.

Figura 2. Imunoblot das proteínas eluídas de uma coluna de níquel a partir do sobrenadante das bactérias

recombinantes induzidas por 3 horas com IPTG. As massas moleculares estão indicadas em kDa.

Para verificar se essa proteína truncada teria atividade, foram realizados ensaios

de proteólise de azocaseína com extrato total de bactéria, porém não foi observada

nenhuma clivagem ATP-dependente ou diferente do controle. O produto de eluição de

colunas de níquel foi igualmente negativo quanto à atividade proteolítica de azocaseína

dependente de ATP.

124

Expressão de PbLon recombinante usando o vetor pET44a+: Este vetor tem como

características principais a expressão da proteína recombinante com uma cauda da

proteína Nus (495 aminoácidos), que confere solubilidade às proteínas expressas, uma

cauda de histidina na região N-terminal para facilitar da purificação e o gene de

resistência à ampicilina. O fragmento de cDNA de 3,1 kb do gene PbLON foi subclonado

no sítio EcoRI do vetor pET44a+ e transformado em E. coli BL21 pLysS.

Análises em “Western blot” utilizando anti-PbLon não revelaram nenhuma banda

na altura esperada de 160 kDa (~45 kDa da cauda de Nus mais ~117 kDa da proteinase),

porém detectou uma banda com cerca de 120 kDa marcada especificamente (figura 3).

Quando sobrenadantes de bactérias recombinantes foram testados quanto à clivagem de

azocaseína, não foi observada nenhuma clivagem ATP-dependente ou diferente do

controle (bactéria sem plasmídeo de expressão).

125

Figura 3. A) Padrão de bandas do extrato total de bactérias recombinantes pET44a+PbLon após 3 horas de

indução com 1 mM de IPTG. Análise em gel de SDS-PAGE 10% corado por Comassie Blue. Controle,

extrato de bactéria transformada com o vetor expressando uma proteína não relacionada. B) Imunoblot

revelado por quimioluminescência e anti-PbLon (1:500). Foram testados extratos de 3 clones

recombinantes e o mesmo controle de A.

Expressão de PbLon em P. pastoris: ainda com objetivo de expressar PbLon completa e

ativa testamos o sistema de expressão na levedura P. pastoris, o qual foi utilizado com

sucesso em nosso laboratório na secreção de gp43 recombinante solúvel (Carvalho,

2004). Foram usados os vetores pPIC9 e pHIL-S1, que possuem os marcadores de

resistência à ampicilina para seleção em bactéria, e o gene HIS4 (da histidinol

desidrogenase) para seleção em leveduras. Nesse sistema, o fragmento de cDNA

heterólogo é sub-clonado entre as sequências promotora e terminadora do gene da

oxidase alternativa (AOX1), para indução da proteína recombinante em grandes

126

quantidades com metanol. Além disso, o fragmento é clonado em fase com uma uma

sequência sinal na região 5’ para promover a secreção da proteína expressa.

A P. pastoris é uma levedura metilotrópica capaz de metabolizar metanol como

única fonte de carbono, a partir de sua oxidação em formaldeído pela álcool oxidase. Em

nosso experimento foi utilizada a cepa GS115 de P. pastoris, que apresenta o genotipo

HIS4. Essa linhagem sofre inserção do plasmídio linearizado no sítio AOX1 e a levedura

passa a apresentar o fenótipo His+ Mut+ / His+ Muts (com alta ou baixa utilização de

metanol, respectivamente), dependendo da enzima utilizada para linearizar o plasmídeo.

O vetor foi clivado com a enzima SacI, que resulta no fenótipo His+ Mut+. Esse evento

pode ocorrer com plasmídeos fechados ou religados, porém em freqüência baixa.

Foi subclonado no sítio EcoRI dos vetores pPIC9 e pHIL-S1um fragmento de 3,1

kb correspondente ao cDNA da PbLON partir do nucleotídeo 45. As construções foram

selecionadas em bactérias através de resistência à ampicilina e as leveduras selecionadas

em meio mínimo desprovido de histidina, após crescimento por 2 dias a 30ºC. Foram

selecionados 10 clones para análise da expressão de PbLon. Para tal, as leveduras foram

induzidas com metanol por 2, 3, 4, 5 e 6 dias e uma alíquota de 5 µL do sobrenadante das

culturas foi testado. Nenhum deles demonstrou atividade ATP-dependente em ensaios de

azocaseína. Não detectamos a proteína recombinante em gel de SDS-PAGE, “Western

blot” ou “Dot blot”.

Expressão de PbLon em S. cerevisiae: testamos o sistema de expressão na levedura S.

cerevisiae, o qual foi utilizado com sucesso na expressão de genes LON de outros

127

organismos (van Dijl et al., 1998). Presentemente, foram usados os vetores pYES2/NT B

e pYES2/NT C (Invitrogen), que possuem os marcadores de resistência à ampicilina para

seleção em bactéria, e o gene URA3 para seleção das leveduras em meio seletivo URA-.

Nesse sistema, o fragmento de cDNA heterólogo é subclonado em fase com a sequência

promotora GAL1, a qual induz fortemente a expressão de proteínas recombinantes na

presença de galactose e é reprimido na presença de glucose.

No vetor pYES2/NT C, foi subclonado um fragmento de 3,1 kb no sítio EcoRI,

correspondente ao cDNA da PbLON partir do nucleotídeo ~100. No vetor pYES2/NT B,

foi subclonado um fragmento de 2,6 kb nos sítios BamHI e EcoRI, com a região N-

terminal truncada em 178 aminoácidos.

As bactérias foram transformadas com as construções e selecionadas através de

resistência à ampicilina. Os produtos de ligação foram utilizados para transformar as

leveduras, de acordo com os protocolos sugeridos pelo fabricante. As leveduras foram

induzidas, coletadas e analisadas em gel de SDS-PAGE, “Western blot” ou “Dot blot”

porém não foi detectado nenhum indício da proteína recombinante. Ressaltamos que este

kit da Invitrogen possui um vetor controle contendo o gene da β-galactosidade que

funcinou perfeitamente.

2. Localização da PbLon por microscopia eletrônica

Em Batista et al. (2006) foram registrados os resultados de localização

mitocondrial da PbLon usando microscopia confocal. Presentemente mostramos que em

microscopia eletrônica as células marcadas com o anticorpo policlonal monoespecífico de

coelho anti-rPbLon mostraram uma discreta marcação na mitocôndria (setas), como visto

128

na figura 4A representativa das demais. A marcação discreta, que contrasta com aquela

vista em microscopia confocal (Batista et al., 2006), pode estar relacionada ao plano do

corte ultrafino analisado. Com o anticorpo pré-imune, houve uma ligeira marcação

inespecífica na parede celular (figura 4B).

Figura 4. A) Microscopia eletrônica de P brasiliensis marcado com anticorpos policlonais anti-rPbLon

monoespecíficos e B) soro pré-imune. As setas indicam marcações intramitocondriais.