- Email: [email protected]
Electrophoretic Characterization of Exposed Outer Membrane Proteins in Environmental and Human Bacteroides fragilis Strains
Zent.bl. Bakteriol. 287,331-341 (1998) © Gustav Fischer Verlag
Zentralblatt fUr
Bakterioloaie
Electrophoretic Characterization of Exposed Outer Membrane Proteins in Environmental and Human Bacteroides fragilis Strains Regina Maria Cavalcanti Pi lotto Domingues", Karia Eliane Santos Avelar, Wilenes das Gra~as Silva e Souza, Saulo Roni Moraes, Eduardo Nazareno Franco Antunes, and Maria Candida de Souza Ferreira Instituto de Microbiologia Prof. Paulo de Goes, Centro de Ciencias da Saude, Uni versidade Federal do Rio de Janeiro, Ilha do Fundao, 21941-590, Rio de Janeiro, RJ, Brasil
Received March 24, 1997· Revision received September 9, 1997· Accepted October 9, 1997
Summary Bacteroides fragi/is isolated from aquatic environment, from infectious processs and from human feces were compared as to their outer membrane protein electrophoretic profiles after staining with Coomassie blue and reacting with antibodies prepared against whole-cell antigens of a reference strain from a clinical source. A marked ho mogeneity was found among the strains with these methodologies. The profiles of all strains obtained after radio-iodination of the intact cell showed qualitative similarity when compared with the profiles obtained by the other methods. Thus, these data al low us to suggest the designation of the peptides observed in the autoradiograms as surface-exposed proteins. Differences observed in the autoradiograms in the expres sion of bands mainly detected at a molecular weight of 28 in the commensal strain 118310 defined previously as avirulent, in addition to a distinction in the titres of ag glutination with the sera tested and lower reactivity in the immunoblotting assays, sug gest a relationship of the B. fragi/is surface architecture with the virulence potential as well as with the origin of the strain.
* To whom all correspondence and reprint requests should be sent.
332
R. M. C. P. Domingues et al.
Introduction The outer membrane (OM) of Gram-negative bacteria commonly constitutes the outermost cell envelope layer and their polypeptide components have been associated with distinct stages of pathogenesis for some bacteria (9). Bacte roides fragilis is a strictly anaerobic Gram-negative bacterium with a great clinical significance in human endogenous infections (8). Several investiga tions have been done in an attempt to understand not only its significance as an opportunist pathogen but also a possible involvement of this microorgan ism in exogenous infections (3). In this context, in addition to the proposal of hydrolytic enzymes (10, 15), toxins (4,18) and capsular polymers (23,28) as virulence factors, some surface proteins have been well characterized (22, 29). Studies on surface proteins with the purpose of diagnostic identification have already been performed and a significant homogeneity was detected in the electrophoretic profiles of these molecules among different clinical isolates of the species (24). Major proteins, defined as porins, were partially character ized in recent investigations related to the limited OM permeability of the species for a number of antibiotics (13). Some authors have also studied the expression in the OM of protein complexes related to the uptake of essential bioelements such as iron, believing that the bacterial virulence should be di rectly related to a better adaptability of the microorganism to hostile environ ments (20,21). Some studies refer to the specific identification of the OM complexes of ex posed proteins in the bacterial surface. The special interest in these constitu ents is based on the fact that these molecules theoretically represent the first contact of the microorganisms with the environment that surrounds them, and hence, playa crucial role in the steps of the cellular interactions, such as adherence, invasion, evasion from the host defences or antimicrobial agents or in the induction of immunological response (9). Several methodologies have been used to label the surface-exposed proteins including those which use lectins, antibodies or radioisotopes such as 1251 and 131 1 (16). Among the radio-iodination methods, the use of a water-insoluble reagent, the 1,3,4,6tetrachloro-3-a,6-a,diphenylglycoluril (Iodogen), as a catalyst of surface specific iodination of membrane proteins, proved to be a reliable labelling method in studies with a number of pathogens (1, 7, 25). The specificity of this method is based on the reaction of the iodo with tyrosine and histidine residues exposed in the bacterial surface OMP. Although the OM complex of B. fragi/is has already been studied in detail, attempts to define the surface-exposed proteins have not yet been made. The study intends not only to achieve a better characterization of the surface of B. fragi/is but also to establish a possible biological relationship among the strains studied which had been isolated from distinct sources and which were already defined with regard to their virulence potential in an animal model (3).
Bacteroides fragilis exposed proteins
333
Materials and Methods Bacterial strains and cultural conditions
Environmental B. fragilis strains (AA1, AA3, AA7, AA10) were isolated from water (polluted and non-polluted), a commensal strain (118310) from human feces and a clinical strain (1031) from a non-intestinal infection. The isolation and identification procedures have been described previously (3). A strain from a human intestinal infec tion (diarrhea - 283-2-1), defined as non-enterotoxigenic, was kindly supplied by Dr. Lyle Myers from the Montana State University, Montana, USA. A reference strain (ATCC-25285) was also included in this study. These strains were selected based on previous results of studies of their virulence potential (3). All the strains were main tained in our culture collection and they were activated in pre-reduced anaerobic ster ilized brain-heart infusion broth (BHI-PRAS) (12) at the moment of use. The strains were cultured in BHI for 18 h at 37°C in an anaerobic chamber in which an atmosphere of 80% nitrogen, 10% hydrogen and 10% CO 2 was maintained. Extraction of outer membrane and whole cell proteins
The outer membrane proteins (OMP) were obtained based on a published method (2). Briefly, after cultivation the bacterial cells were pelleted by centrifugation, washed and resuspended in 10 mM Tris-HCI buffer pH 8.0, with 1 mM EDTA and 1 mM 2-~ mercaptoethanol. The cells were then disrupted in an ultrasonic disintegrator at 4°C. Remaining intact cells were removed by centrifugation. The crude envelope fraction was collected from the supernatant by centrifugation (100000 x g) for 30 min at 4°C. The pellet was then treated twice with 0.3% (w/v) Sarkosyl (Sigma) solution with 1 mM 2-~-mercaptoethanol to selectively solubilize the inner membrane pan. The in soluble OM fraction was recovered as pelleted by centrifugation at 13 000 x g for 10 min at 4°C. Whole-cell proteins (WP) were obtained as described by Taylor et al. (26). After cultivation, the cells were washed twice with 0.01 M phosphate-buffered saline (PBS) pH 7.2 and 1 ml of the pellet was resuspended in 100 III of a buffer con taining 0.25 M, Tris-HCl 0.192 M glycine 0.1 % SDS pH 8.5. The OMP and WP ex tracts were stored at a -20°C until use. Radio-iodination of intact cells
The B. fragilis cells were radio-iodinated as previously described (7). Cells were washed twice in PBS and resuspended to a final concentration of 109 cells/ml. The cell suspen sions (0.2 ml) were allowed to react with 300 IlCi of carrier-free 1251 (10 mCilml in re action vials containing 45 Ilg of lodogen (Pierce Chemical Co., Rockford, USA) for 30 s at room temperature. Radio-iodination was interrupted by transferring the sample to another tube containing 3.0 ml of 10 mM Nal, followed by washing (three times) with the same solution and another washing with 0.5 M Tris-HCl (pH6.8). Labelled cells were solubilised in electrophoresis buffer and subjected to electrophoresis and autora diography. Before and after the radio-iodination assays, the bacterial populations were analysed in relation to their quantitative viability as proposed by Miles and Misra (17). SDS-PAGE analysis
SDS-PAGE was carried out according to Laemmli's method (14) with a discontinuous Tris-glycine buffer system. The stacking and resolving gels were 4% and 10% acryl-
334
R. M. C. P. Domingues et al.
amide, respectively. Protein samples were mixed with sample buffer containing 0.5 M Tris-HCI (pH 6.8), 10% SDS, 10% glycerol, bromophenol blue (0.05%) and 2% ~ mercaptoethanol. Aliquots were heated at 100°C for 4 min and applied to the gel ap paratus at 20 rnA. The gels obtained after the electrophoresis of OMP and WP extracts were stained for 2 h with Coomassie brilliant blue G-250. Dry gels were scanned with a scanning densitometer. The average similarity between two protein profiles was as sessed by the Dice coefficient (5). Afterwards the electrophoresis resolution of the pro tein extracts labelled with 1251, the radioactivity detection was carried out with frozen gels (-80°C for 24 h) in Kodak X-Omat film. Immunoblotting assays After SDS-PAGE, proteins were electrophoretic ally transferred from the gel onto a ni trocellulose membrane as described by Towbin et al. (27). The transfer was performed for 1 hat 100V in a transfer buffer (25 mM Tris, 193 mM glycine, 20% methanol) in a cooled reservoir. Next, the membrane was washed three times with PBS-O.3% Tween 20 and then immersed in a blocking buffer (5% skim milk, 0.3% Tween 20 and PBS) for 1 h and in the immune serum diluted 1 :100 in blocking buffer for 1 h at room tem perature. The serum was prepared in rabbits against whole cell antigens of a reference strain (ATCC 25285) as described by Elhag and Talbaqchali (6). The membrane was washed three times for 5 min with PBS-O.3 % Tween 20 and probed with a 1: 500 di lution in blocking buffer of an anti-rabbit immunoglobulin G horseradish peroxidase conjugate (Sigma Chemical Co.) for 1 h at room temperature. Following three 5 min washes with PBS, the membrane was placed in freshly prepared 3,3' -diaminobenzidine substrate 0.5 mg/ml in 0.05 M Tris-HCI buffer pH 7.6 with 3 f!l of 30% H 2 0 zlmi of buffer added immediately prior to use. All the above steps were performed under gentle rocking. Agglutination tests Fifty microlitres of the above mentioned rabbit anti-serum were serially (2-fold) dilut ed in PBS, on a microtiter tray and 50 f!l of the bacterial suspension (10 9 CFU/ml in PBS) were added to each well. The plate was gently shaken for 1 min and read after 2 h of incubation at 37°C and after an extra overnight incubation at 4°C. Negative controls were performed with a pre-immune serum and with bacterial suspension with out anti-serum. Agglutination tests were expressed as the reciprocal of the highest dilution that showed a positive reaction. The tests were performed in duplicate.
Results A marked homogeneity in the banding patterns was observed among the strains. Figure 1 shows these profiles as revealed by Coomassie staining. A total of eight common bands could be clearly distinguished in the profiles of all the extracts tested (80, 58, 44, 38, 36, 34, 28 and 22 kDa) and only minor differences were observed. The visual inspections of OMP banding patterns were confirmed by densitometer analysis, which determined similarity coeffi cients (%Dice) above 95% when comparing different profile pairs among the strains. The densitometer tracings are presented in Fig. 2.
Bacteroides fragi/is exposed proteins
335
Fig. 1. SDS-PAGE profiles of OMP extracts of Bacteroides fragi/is strains. Lanes A molecular weight standards (Gibco); B - ATCC 25285; C -118310; D - 283-2-1; E1031; F - molecular weight standards (Gibco); G - AA1; H - AA3; 1- AA7; J - AA10. Common majority bands detected at molecular weights of 80, 58, 44, 38, 36, 34, 28 and 22 (.-). Some variability was observed in the bands of high molecular weight (*).
A
B
c
o
E
F
G
H
Fig. 2. Densitometer tracings showing the similarity of OMP electrophoretic profiles among strains of Bacteroides fragi/is from different sources shown in Fig. 1. A coeffi cient of similarity (% Dice) larger than 90% was estimated among the strains. A - ATCC 25285; B -118310; C -283-2-1; D -1031; E- AA1; F - AA3; G -AA7; H-AA10.
336
R. M. C. P. Domingues etal.
In the autoradiograms, 9 protein bands were regularly detected in B. fragi 1251. These profiles are shown in Fig. 3. The most prominent common bands detected had molecular weights of 62, 58, 54, 44, 36, 34, 28, 25 and 22. The strains ATCC 25285, AA1, AA3, 283-2-1 and 1031 (lanes A, C, D, E and F) presented the same banding pat tern. The strains AA7 and AA10 (lanes G and H) were located in a distinct profile in which the reaction with the bands detected at molecular weights of 36, 34 and 25 were quantitatively altered. Of particular interest was the re duction in the reaction of the protein band detected at 28 kDa for strain 118310 (lane B). Before and after the radio-iodination assays, the bacterial populations (strains ATCC25285, AA1 and 118310) were studied in relation to their viability as a control of the cell integrity and they did not show a sig nificant decrease in the number of viable cells. These data are represented in Fig. 4. The results found in the immunoblotting of OMP preparations against rab bit serum raised to whole cell antigens of a reference B. fragilis strain (ATCC 25285) are shown in Fig. 5. As observed in the autoradiograms, the immuno blotting profiles showed a great homogeneity but also some distinctions. The anti-serum reacted with homologous antigens (lane a) and with heterologous
lis whole cells labelled with Iodogen and
Fig. 3. Autoradiogram of Iodogen-catalysed iodination of Bacteroides fragilis strains. LANES A - 25285; B - 118310; C - AAl; D - AA3; E - 283-2-1; F -1031; G - AA7; H - AAI0. Quantitative differences detected in the bands at molecular weights of 28 (lane B) and 36, 34 and 25 (lanes G and H) when comparing the strains. Molecular weight standards defined by comparison with the electrophoretic pattern defined by the Coomassie blue-stained gel.
Bacteroides fragi/is exposed proteins
337
100 .-. ~
-
~
60
o BEFORE • AFTER
40
25285
AA1
118310
Strains Fig.4. Evaluation of quantitative viability of the strains ATCC 25285, AA1 and 118310, as a control of cell integrity before (D) and after (II ) the radio-iodination as says. No significant decrease was observed in the number of viable cells in any ana lysed population.
Fig. 5. Immunoblotted SDS-PAGE of OMP extracts reacted with rabbit serum anti whole cell antigens of a Bacteroides fragi/is reference strain (ATCC 25285) are re vealed through a horseradish peroxidase conjugate. Lanes a - molecular weight stand ards prestained (Gibco), b - ATCC 25285, c - 118310, d - AA1, e - AA7, f - AA3, g - AA10, h - 1031, i - 283-2-1. Note band of 28 kDa has a greater intensity in lanes b, d, f, g, h, and i (~) and poor expression in lane c ('.). Lane e - strain AA 7 with some distinction in the reactivity of the band detected at 37 kDa (*) and 27 kDa (+).
338
R. M. C. P. Domingues et al.
OMP extracts, revealing a strong cross-reactivity. Regions of broad variabil ity were delineated, from 25 to 40 kDa, and the strain 118310 (lane c) showed a lower reactivity in the band detected at a molecular weight of 28. The strain AA7 (lane e) presented a strong reactivity in the band of 35 kDa. Comparative analyses were done with the strains in relation to the patterns defined by those distinct methods and such analyses raised specific patterns in which ATCC 25285 and 118310 strains were selected as the most represen tative strains (Fig. 6). A greater number of bands was detected in the WP pro files (lanes A and B) than in the OMP profiles (lanes C and D), both of them defined after staining with Coomassie blue. Intact-cell protein extracts radio labelled with 1251 (lanes E and F) presented qualitative profiles quite similar to OMP profiles in Coomassie blue-stained gels. Through the analysis of those profiles, the band of 28 kDa could be detected with greater intensity in the autoradiograms (Fig. 3) and in the immunoblotting membranes (Fig. 5) than in the OMP Coomassie blue-stained gels (Fig. 1) in all strains except 118310. In the agglutination assays, in which the bacterial cells were tested with the same antibodies used in the immunoblotting tests, the titres ranged from 32 in heterologous reactions with the commensal strains (118310) to 512, both in homologous (ATCC 25285 antigens) as well as in heterologous reactions, with strains from clinical (1031, 283-2-1) or from environmental sources (AA1, AA3, AA10) (Table 1).
Fig. 6. Comparative analysis of electrophoretic patterns with WP extracts (lanes A and B) and OMP extracts (lanes C and D), stained with Coomassie blue; intact-cell extracts radiolabelled with 1251 (lanes E and F) and OMP extracts reacted with antiserum pre pared against whole cell antigens of the strain ATCC 25285. Profiles of strain 118310: lanes A, C, E, G. Profiles of strain ATCC 25285: lanes B, D, F, H. Note the qualitative similarity among OMP, autoradiographic and immunoblotted profiles and distinction in the expression of the 28 kDa band in the autoradiogram (>-).
Bacteroides fragi/is exposed proteins
339
Table 1. Reactivity of Bacteroides fragi/is strains isolated from clinical processes, hu man faeces and aquatic environment with rabbit serum prepared against whole cell antigens of a reference strain (ATCC 25285) Strains, whole-cell antigens
Agglutination titres S
ATCC 252851, 1031 1,283-2-1 2, AA1 3, AA31, AA10 3 AA7' 1183104
512
1
2 3 4
5
128
32
non-intestinal infection; intestinal infection; aquatic environment; human faeces; agglutination tests in microtiter tray.
Discussion Although the OM complex of B. fragilis has been widely studied, until now no investigation has been performed with the purpose of identifying surface exposed proteins in this species. This kind of identification has provided inter esting information about cellular structures, since it reveals the real architec ture of the bacterial envelope in a certain moment (16). Methods based on the use of Iodogen and Iodo as a catalyst and a marker, respectively, of a specific radiolabelling of surface-exposed protein, were already applied to other Gram-negative bacteria (1, 7,25). In our study, we could confirm the method as being simple, rapid and reproducible (16). In this work, we could detect a total of 9 major bands in the autoradiograms. Therefore the results of comparative analyses among OMP and WP profiles after staining with Coomassie blue and the immunoblotting membranes after reaction with antibodies prepared against intact cells, from clinical sources, allow us to suggest that the proteins detected in the autoradiograms are indeed surface-located and the Iodogen-catalysed iodination technique is a simple tool for labelling surface-displayed proteins in B. fragilis cells. It is possible that the quantitative differences detected when comparing the OMP with the protein profiles after radio-iodination could be explained by the topographic distribution of the bacterial envelope components. Hence, the proteins defined as major ones in SDS-PAGE gels after staining with Coomas sie blue should have few surface-exposed regions. In relation to the comparative analysis of the protein profiles of the strains from distinct sources, previously categorized as virulent or avirulent, interest ing data have also been found. The distinction in the surface-exposed protein profiles detected, specifically in the band with an apparent molecular ratio of 28 kDa, in the commensal strain, in addition to the observation of the lowest agglutination titres in this strain with antibodies prepared against whole-cell
340
R. M. C. P. Domingues et al.
antigens of clinical strains seems to be relevant. It is tempting to speculate about a particular surface organization in this strain of commensal origin, in relation to strains from other sources. The phenotypical variability was al ready the subject of several studies with B. fragilis (3, 11, 19) and attempts to distinguish commensal strains from clinical isolates have been made in rela tion to various bacterial characters. Admitting the multiplicity of B. fragilis relations in nature, in balanced stages or not, our results allow us to propose a possible relation of the bacte rial surface composition and organization to the origin of the isolates as well as the virulence potential. Acknowledgements. We thank Joaquim Santos Filho for technical assistance. This work was supported by grants of the following institutions: CNPq, PRONEX, FUJB, FINEP and CEPG.
References 1. Abath, F. G. c., A. M. P. Almeida, and 1. C. S. Ferreira: Identification of surface exposed Yersinia pestis proteins by radio-iodination and biotinylation. J. Med. Microbiol. 37 (1992) 420-424 2. Bolin, I., 1. Norlander, and H. Wolf- Watz: Temperature-inducible outer membrane protein of Yersinia pseudotuberculosis and Yersinia enterocolitica is associated with the virulence plasmid. Infect. Immun. 17 (1982) 506-512 3. Domingues, R. M. C. P., M. Oliveira, R. HirataJr., A. F. B.Andrade, andM. C. S. Fer reira: Bacteroides fragilis: an exogenous pathogen? Zbl. Bakt. 282 (1995) 296-302 4. Donelli, G. , A. Fabbri, and C. Fiorentini: Bacteroides fragilis enterotoxin induces cytoskeletal changes and surface blebbing in HT-29 cells. Infect. Imrnun. 64 (1996) 113-119 5. Dice, 1. R.: Measurements of amounts of ecologic association between species. Ecology 26 (1945) 297-302 6. Elhag, K. M. and S. Tabaqchali: A study of the surface and somatic antigens of Bac teroides fragilis. J. Hyg. 180 (1978) 439-449 7. Ferreira, 1. C. S. and D. I. Almeida: Iodogen catalysed iodination for identification of surface-exposed outer membrane proteins of Escherichia coli K12. Braz. J. Ge netics 10 (1987) 625-634 8. Finegold, S. M. and D. M. Citron: Gram negative nonspore forming anaerobe bacilli. In Lenette, E. H., Ballows, A., Hausler Jr. , w.j. , and Truant, j. P. (eds.), Manual of Clinical Microbiology (3rd ed.): (1980) pp.187-192. American Society for Microbiology, Washington DC 9. Finlay, B. B. and S. Falkow: Common themes in microbial pathogenicity. Micro bioI. Rev. 53 (1989) 210-230 10. Godoy, V. G., M. M. Dallas, T. A. Russo, and M. M. Malany: A role for Bacteroides fragilis neuraminidase in bacterial growth in two model systems. Infect. Immun. 61 (1993) 4415-4426 11. Hofstad, T.: Virulence factors in anaerobic bateria . Eur. J. Clin. Microbiol. Infect. Dis. 11 (1992) 1044-1049 12. Holdeman, 1. v., E. P. Cato, and W. E. C. Moore: Anaerobe Laboratory Manual, 4th ed. (1977) Virginia Polytechnic Institute and State University, Blacksburg
Bacteroides fragilis exposed proteins
341
13. Kanazawa, K., Y. Kobayashi, M. Nakano, M. Sakurai, N. Gotoh, and T. Nishino: Identification of three porines in the outer membrane of Bacteroides fragilis. FEMS Microbiol. Lett. 127 (1995) 181-186 14. Laemmli, U. K.: Cleavage of structural proteins during the assembly of the head of bacteriophage T 4 • Nature 227 (1970) 680-685 15. Macfarlane, G. T. and S. Macfarlane: Formation of glycoprotein degrading en zymes by Bacteroides fragilis. FEMS Microbiol. Lett. 77 (1991) 289-294 16. Markwell, M. A. K. and C. F. Fox: Surface-specific iodination of membrane pro teins of viruses and eukaryotic cells using 1,3,4,6-tetrachloro-3-a,6-a,diphenylgly coluril. Biochem. 17 (1978) 4807-4817 17. Miles, A. A. and S. S. Misra: The estimation of the bactericidal power of the blood. J.Hyg.38 (1938) 732-749 18. Myers, 1. 1., D. S. Shoop, 1. 1. Stackhouse, F. S. Newman, R.]. Flaherty, G. W. Letson, and R. B. Sack: Isolation of enterotoxigenic Bacteroides fragilis from hu man with diarrhea. J. Clin. Microbial. 25 (1987) 2330-2333 19. Namavar,F., M.A.].]. Verweij-van Vught, and D.M. MacLaren: A study of the candidate virulence factors of Bacteroides fragilis. J. Gen. Microbiol. 137 (1991) 1431-1435 20. Otto, B. R., M. Sparrius, A. M.].f. Verweij-van Vught, and D. M. MacLaren: Iron regulated outer membrane protein of Bacteroides fragilis involved in heme uptake. Infect. Immun. 58 (1990) 3954-3958 21. Otto, B. R., W. R. Verweij, M. Sparrius, A. M. f.]. Verweij-van Vught, C. E. Nord, and D. M. MacLaren: Human immune response to an iron-repressible outer mem brane protein of Bacteroides fragilis. Infect. Immun. 59 (1991) 2999-3003 22. Patrick, S. and D. A. Lutton: Outer membrane proteins of Bacteroides fragilis grown in vivo. FEMS Microbial. Lett. 71 (1990) 1-4 23. Patrick, S., D. A. Lutton, and A. D. Crockard: Immune reactions of Bacteroides fragilis populations with three different types of capsule in a model of infection. Microbiol. 141 (1995) 1969-1976 24. Poxton, I. R. and R. Brown: Sodium dodecyl sulphate-polyacrylamide gel electro phoresis of cell surface proteins as an aid to the identification of the Bacteroides fragilis group. J. Gen. Microbial. 112 (1979) 211-217 25. Swason,f.: Surface-exposed protein antigens of the gonococcal outer membrane. Infect. Immun. 34 (1981) 804-812 26. Taylor, A. f., C. A. Dawson, and R.]. Owen: The identification of Bacteroides ureo lyticus from patients with non-gonococcal urethritis by conventional biochemical tests and by DNA and protein analyses. J. Med. Microbiol. 21 (1986) 109-116 27. Towbin, H., T. Stachelen, and]. Gordon: Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA 76 (1979) 4350-4354 28. Tzianabos, A. 0., A. B. Onderdonk, B. Rosner, R.1. Cisneros, and D. 1. Kasper: Structural features of polysaccharides that induce intra-abdominal abscesses. Sci ence 262 (1993) 416-419 29. van Doorn,]., B. Oudega, and D. M. MacLaren: Characterization and detection of a 40 kDa fimbrial subunit of Bacteroides fragilis BEL Microbial Pathogen. 13 (1992) 75-79
Corresponding author: Instituto de Microbiologia Prof. Paulo de Goes, Centro de Ciencias da Saude, Universidade Federal do Rio de Janeiro, Ilha do Fundao, 21941590, Rio de Janeiro, RJ, Brasil, Fax: 0055 (021) 560-8028, E-mail: immmcan@micro bio.ufrj.br 23
Zent.bl. Bakteriol. 287/4
Zentralblatt fUr
Bakterioloaie
Electrophoretic Characterization of Exposed Outer Membrane Proteins in Environmental and Human Bacteroides fragilis Strains Regina Maria Cavalcanti Pi lotto Domingues", Karia Eliane Santos Avelar, Wilenes das Gra~as Silva e Souza, Saulo Roni Moraes, Eduardo Nazareno Franco Antunes, and Maria Candida de Souza Ferreira Instituto de Microbiologia Prof. Paulo de Goes, Centro de Ciencias da Saude, Uni versidade Federal do Rio de Janeiro, Ilha do Fundao, 21941-590, Rio de Janeiro, RJ, Brasil
Received March 24, 1997· Revision received September 9, 1997· Accepted October 9, 1997
Summary Bacteroides fragi/is isolated from aquatic environment, from infectious processs and from human feces were compared as to their outer membrane protein electrophoretic profiles after staining with Coomassie blue and reacting with antibodies prepared against whole-cell antigens of a reference strain from a clinical source. A marked ho mogeneity was found among the strains with these methodologies. The profiles of all strains obtained after radio-iodination of the intact cell showed qualitative similarity when compared with the profiles obtained by the other methods. Thus, these data al low us to suggest the designation of the peptides observed in the autoradiograms as surface-exposed proteins. Differences observed in the autoradiograms in the expres sion of bands mainly detected at a molecular weight of 28 in the commensal strain 118310 defined previously as avirulent, in addition to a distinction in the titres of ag glutination with the sera tested and lower reactivity in the immunoblotting assays, sug gest a relationship of the B. fragi/is surface architecture with the virulence potential as well as with the origin of the strain.
* To whom all correspondence and reprint requests should be sent.
332
R. M. C. P. Domingues et al.
Introduction The outer membrane (OM) of Gram-negative bacteria commonly constitutes the outermost cell envelope layer and their polypeptide components have been associated with distinct stages of pathogenesis for some bacteria (9). Bacte roides fragilis is a strictly anaerobic Gram-negative bacterium with a great clinical significance in human endogenous infections (8). Several investiga tions have been done in an attempt to understand not only its significance as an opportunist pathogen but also a possible involvement of this microorgan ism in exogenous infections (3). In this context, in addition to the proposal of hydrolytic enzymes (10, 15), toxins (4,18) and capsular polymers (23,28) as virulence factors, some surface proteins have been well characterized (22, 29). Studies on surface proteins with the purpose of diagnostic identification have already been performed and a significant homogeneity was detected in the electrophoretic profiles of these molecules among different clinical isolates of the species (24). Major proteins, defined as porins, were partially character ized in recent investigations related to the limited OM permeability of the species for a number of antibiotics (13). Some authors have also studied the expression in the OM of protein complexes related to the uptake of essential bioelements such as iron, believing that the bacterial virulence should be di rectly related to a better adaptability of the microorganism to hostile environ ments (20,21). Some studies refer to the specific identification of the OM complexes of ex posed proteins in the bacterial surface. The special interest in these constitu ents is based on the fact that these molecules theoretically represent the first contact of the microorganisms with the environment that surrounds them, and hence, playa crucial role in the steps of the cellular interactions, such as adherence, invasion, evasion from the host defences or antimicrobial agents or in the induction of immunological response (9). Several methodologies have been used to label the surface-exposed proteins including those which use lectins, antibodies or radioisotopes such as 1251 and 131 1 (16). Among the radio-iodination methods, the use of a water-insoluble reagent, the 1,3,4,6tetrachloro-3-a,6-a,diphenylglycoluril (Iodogen), as a catalyst of surface specific iodination of membrane proteins, proved to be a reliable labelling method in studies with a number of pathogens (1, 7, 25). The specificity of this method is based on the reaction of the iodo with tyrosine and histidine residues exposed in the bacterial surface OMP. Although the OM complex of B. fragi/is has already been studied in detail, attempts to define the surface-exposed proteins have not yet been made. The study intends not only to achieve a better characterization of the surface of B. fragi/is but also to establish a possible biological relationship among the strains studied which had been isolated from distinct sources and which were already defined with regard to their virulence potential in an animal model (3).
Bacteroides fragilis exposed proteins
333
Materials and Methods Bacterial strains and cultural conditions
Environmental B. fragilis strains (AA1, AA3, AA7, AA10) were isolated from water (polluted and non-polluted), a commensal strain (118310) from human feces and a clinical strain (1031) from a non-intestinal infection. The isolation and identification procedures have been described previously (3). A strain from a human intestinal infec tion (diarrhea - 283-2-1), defined as non-enterotoxigenic, was kindly supplied by Dr. Lyle Myers from the Montana State University, Montana, USA. A reference strain (ATCC-25285) was also included in this study. These strains were selected based on previous results of studies of their virulence potential (3). All the strains were main tained in our culture collection and they were activated in pre-reduced anaerobic ster ilized brain-heart infusion broth (BHI-PRAS) (12) at the moment of use. The strains were cultured in BHI for 18 h at 37°C in an anaerobic chamber in which an atmosphere of 80% nitrogen, 10% hydrogen and 10% CO 2 was maintained. Extraction of outer membrane and whole cell proteins
The outer membrane proteins (OMP) were obtained based on a published method (2). Briefly, after cultivation the bacterial cells were pelleted by centrifugation, washed and resuspended in 10 mM Tris-HCI buffer pH 8.0, with 1 mM EDTA and 1 mM 2-~ mercaptoethanol. The cells were then disrupted in an ultrasonic disintegrator at 4°C. Remaining intact cells were removed by centrifugation. The crude envelope fraction was collected from the supernatant by centrifugation (100000 x g) for 30 min at 4°C. The pellet was then treated twice with 0.3% (w/v) Sarkosyl (Sigma) solution with 1 mM 2-~-mercaptoethanol to selectively solubilize the inner membrane pan. The in soluble OM fraction was recovered as pelleted by centrifugation at 13 000 x g for 10 min at 4°C. Whole-cell proteins (WP) were obtained as described by Taylor et al. (26). After cultivation, the cells were washed twice with 0.01 M phosphate-buffered saline (PBS) pH 7.2 and 1 ml of the pellet was resuspended in 100 III of a buffer con taining 0.25 M, Tris-HCl 0.192 M glycine 0.1 % SDS pH 8.5. The OMP and WP ex tracts were stored at a -20°C until use. Radio-iodination of intact cells
The B. fragilis cells were radio-iodinated as previously described (7). Cells were washed twice in PBS and resuspended to a final concentration of 109 cells/ml. The cell suspen sions (0.2 ml) were allowed to react with 300 IlCi of carrier-free 1251 (10 mCilml in re action vials containing 45 Ilg of lodogen (Pierce Chemical Co., Rockford, USA) for 30 s at room temperature. Radio-iodination was interrupted by transferring the sample to another tube containing 3.0 ml of 10 mM Nal, followed by washing (three times) with the same solution and another washing with 0.5 M Tris-HCl (pH6.8). Labelled cells were solubilised in electrophoresis buffer and subjected to electrophoresis and autora diography. Before and after the radio-iodination assays, the bacterial populations were analysed in relation to their quantitative viability as proposed by Miles and Misra (17). SDS-PAGE analysis
SDS-PAGE was carried out according to Laemmli's method (14) with a discontinuous Tris-glycine buffer system. The stacking and resolving gels were 4% and 10% acryl-
334
R. M. C. P. Domingues et al.
amide, respectively. Protein samples were mixed with sample buffer containing 0.5 M Tris-HCI (pH 6.8), 10% SDS, 10% glycerol, bromophenol blue (0.05%) and 2% ~ mercaptoethanol. Aliquots were heated at 100°C for 4 min and applied to the gel ap paratus at 20 rnA. The gels obtained after the electrophoresis of OMP and WP extracts were stained for 2 h with Coomassie brilliant blue G-250. Dry gels were scanned with a scanning densitometer. The average similarity between two protein profiles was as sessed by the Dice coefficient (5). Afterwards the electrophoresis resolution of the pro tein extracts labelled with 1251, the radioactivity detection was carried out with frozen gels (-80°C for 24 h) in Kodak X-Omat film. Immunoblotting assays After SDS-PAGE, proteins were electrophoretic ally transferred from the gel onto a ni trocellulose membrane as described by Towbin et al. (27). The transfer was performed for 1 hat 100V in a transfer buffer (25 mM Tris, 193 mM glycine, 20% methanol) in a cooled reservoir. Next, the membrane was washed three times with PBS-O.3% Tween 20 and then immersed in a blocking buffer (5% skim milk, 0.3% Tween 20 and PBS) for 1 h and in the immune serum diluted 1 :100 in blocking buffer for 1 h at room tem perature. The serum was prepared in rabbits against whole cell antigens of a reference strain (ATCC 25285) as described by Elhag and Talbaqchali (6). The membrane was washed three times for 5 min with PBS-O.3 % Tween 20 and probed with a 1: 500 di lution in blocking buffer of an anti-rabbit immunoglobulin G horseradish peroxidase conjugate (Sigma Chemical Co.) for 1 h at room temperature. Following three 5 min washes with PBS, the membrane was placed in freshly prepared 3,3' -diaminobenzidine substrate 0.5 mg/ml in 0.05 M Tris-HCI buffer pH 7.6 with 3 f!l of 30% H 2 0 zlmi of buffer added immediately prior to use. All the above steps were performed under gentle rocking. Agglutination tests Fifty microlitres of the above mentioned rabbit anti-serum were serially (2-fold) dilut ed in PBS, on a microtiter tray and 50 f!l of the bacterial suspension (10 9 CFU/ml in PBS) were added to each well. The plate was gently shaken for 1 min and read after 2 h of incubation at 37°C and after an extra overnight incubation at 4°C. Negative controls were performed with a pre-immune serum and with bacterial suspension with out anti-serum. Agglutination tests were expressed as the reciprocal of the highest dilution that showed a positive reaction. The tests were performed in duplicate.
Results A marked homogeneity in the banding patterns was observed among the strains. Figure 1 shows these profiles as revealed by Coomassie staining. A total of eight common bands could be clearly distinguished in the profiles of all the extracts tested (80, 58, 44, 38, 36, 34, 28 and 22 kDa) and only minor differences were observed. The visual inspections of OMP banding patterns were confirmed by densitometer analysis, which determined similarity coeffi cients (%Dice) above 95% when comparing different profile pairs among the strains. The densitometer tracings are presented in Fig. 2.
Bacteroides fragi/is exposed proteins
335
Fig. 1. SDS-PAGE profiles of OMP extracts of Bacteroides fragi/is strains. Lanes A molecular weight standards (Gibco); B - ATCC 25285; C -118310; D - 283-2-1; E1031; F - molecular weight standards (Gibco); G - AA1; H - AA3; 1- AA7; J - AA10. Common majority bands detected at molecular weights of 80, 58, 44, 38, 36, 34, 28 and 22 (.-). Some variability was observed in the bands of high molecular weight (*).
A
B
c
o
E
F
G
H
Fig. 2. Densitometer tracings showing the similarity of OMP electrophoretic profiles among strains of Bacteroides fragi/is from different sources shown in Fig. 1. A coeffi cient of similarity (% Dice) larger than 90% was estimated among the strains. A - ATCC 25285; B -118310; C -283-2-1; D -1031; E- AA1; F - AA3; G -AA7; H-AA10.
336
R. M. C. P. Domingues etal.
In the autoradiograms, 9 protein bands were regularly detected in B. fragi 1251. These profiles are shown in Fig. 3. The most prominent common bands detected had molecular weights of 62, 58, 54, 44, 36, 34, 28, 25 and 22. The strains ATCC 25285, AA1, AA3, 283-2-1 and 1031 (lanes A, C, D, E and F) presented the same banding pat tern. The strains AA7 and AA10 (lanes G and H) were located in a distinct profile in which the reaction with the bands detected at molecular weights of 36, 34 and 25 were quantitatively altered. Of particular interest was the re duction in the reaction of the protein band detected at 28 kDa for strain 118310 (lane B). Before and after the radio-iodination assays, the bacterial populations (strains ATCC25285, AA1 and 118310) were studied in relation to their viability as a control of the cell integrity and they did not show a sig nificant decrease in the number of viable cells. These data are represented in Fig. 4. The results found in the immunoblotting of OMP preparations against rab bit serum raised to whole cell antigens of a reference B. fragilis strain (ATCC 25285) are shown in Fig. 5. As observed in the autoradiograms, the immuno blotting profiles showed a great homogeneity but also some distinctions. The anti-serum reacted with homologous antigens (lane a) and with heterologous
lis whole cells labelled with Iodogen and
Fig. 3. Autoradiogram of Iodogen-catalysed iodination of Bacteroides fragilis strains. LANES A - 25285; B - 118310; C - AAl; D - AA3; E - 283-2-1; F -1031; G - AA7; H - AAI0. Quantitative differences detected in the bands at molecular weights of 28 (lane B) and 36, 34 and 25 (lanes G and H) when comparing the strains. Molecular weight standards defined by comparison with the electrophoretic pattern defined by the Coomassie blue-stained gel.
Bacteroides fragi/is exposed proteins
337
100 .-. ~
-
~
60
o BEFORE • AFTER
40
25285
AA1
118310
Strains Fig.4. Evaluation of quantitative viability of the strains ATCC 25285, AA1 and 118310, as a control of cell integrity before (D) and after (II ) the radio-iodination as says. No significant decrease was observed in the number of viable cells in any ana lysed population.
Fig. 5. Immunoblotted SDS-PAGE of OMP extracts reacted with rabbit serum anti whole cell antigens of a Bacteroides fragi/is reference strain (ATCC 25285) are re vealed through a horseradish peroxidase conjugate. Lanes a - molecular weight stand ards prestained (Gibco), b - ATCC 25285, c - 118310, d - AA1, e - AA7, f - AA3, g - AA10, h - 1031, i - 283-2-1. Note band of 28 kDa has a greater intensity in lanes b, d, f, g, h, and i (~) and poor expression in lane c ('.). Lane e - strain AA 7 with some distinction in the reactivity of the band detected at 37 kDa (*) and 27 kDa (+).
338
R. M. C. P. Domingues et al.
OMP extracts, revealing a strong cross-reactivity. Regions of broad variabil ity were delineated, from 25 to 40 kDa, and the strain 118310 (lane c) showed a lower reactivity in the band detected at a molecular weight of 28. The strain AA7 (lane e) presented a strong reactivity in the band of 35 kDa. Comparative analyses were done with the strains in relation to the patterns defined by those distinct methods and such analyses raised specific patterns in which ATCC 25285 and 118310 strains were selected as the most represen tative strains (Fig. 6). A greater number of bands was detected in the WP pro files (lanes A and B) than in the OMP profiles (lanes C and D), both of them defined after staining with Coomassie blue. Intact-cell protein extracts radio labelled with 1251 (lanes E and F) presented qualitative profiles quite similar to OMP profiles in Coomassie blue-stained gels. Through the analysis of those profiles, the band of 28 kDa could be detected with greater intensity in the autoradiograms (Fig. 3) and in the immunoblotting membranes (Fig. 5) than in the OMP Coomassie blue-stained gels (Fig. 1) in all strains except 118310. In the agglutination assays, in which the bacterial cells were tested with the same antibodies used in the immunoblotting tests, the titres ranged from 32 in heterologous reactions with the commensal strains (118310) to 512, both in homologous (ATCC 25285 antigens) as well as in heterologous reactions, with strains from clinical (1031, 283-2-1) or from environmental sources (AA1, AA3, AA10) (Table 1).
Fig. 6. Comparative analysis of electrophoretic patterns with WP extracts (lanes A and B) and OMP extracts (lanes C and D), stained with Coomassie blue; intact-cell extracts radiolabelled with 1251 (lanes E and F) and OMP extracts reacted with antiserum pre pared against whole cell antigens of the strain ATCC 25285. Profiles of strain 118310: lanes A, C, E, G. Profiles of strain ATCC 25285: lanes B, D, F, H. Note the qualitative similarity among OMP, autoradiographic and immunoblotted profiles and distinction in the expression of the 28 kDa band in the autoradiogram (>-).
Bacteroides fragi/is exposed proteins
339
Table 1. Reactivity of Bacteroides fragi/is strains isolated from clinical processes, hu man faeces and aquatic environment with rabbit serum prepared against whole cell antigens of a reference strain (ATCC 25285) Strains, whole-cell antigens
Agglutination titres S
ATCC 252851, 1031 1,283-2-1 2, AA1 3, AA31, AA10 3 AA7' 1183104
512
1
2 3 4
5
128
32
non-intestinal infection; intestinal infection; aquatic environment; human faeces; agglutination tests in microtiter tray.
Discussion Although the OM complex of B. fragilis has been widely studied, until now no investigation has been performed with the purpose of identifying surface exposed proteins in this species. This kind of identification has provided inter esting information about cellular structures, since it reveals the real architec ture of the bacterial envelope in a certain moment (16). Methods based on the use of Iodogen and Iodo as a catalyst and a marker, respectively, of a specific radiolabelling of surface-exposed protein, were already applied to other Gram-negative bacteria (1, 7,25). In our study, we could confirm the method as being simple, rapid and reproducible (16). In this work, we could detect a total of 9 major bands in the autoradiograms. Therefore the results of comparative analyses among OMP and WP profiles after staining with Coomassie blue and the immunoblotting membranes after reaction with antibodies prepared against intact cells, from clinical sources, allow us to suggest that the proteins detected in the autoradiograms are indeed surface-located and the Iodogen-catalysed iodination technique is a simple tool for labelling surface-displayed proteins in B. fragilis cells. It is possible that the quantitative differences detected when comparing the OMP with the protein profiles after radio-iodination could be explained by the topographic distribution of the bacterial envelope components. Hence, the proteins defined as major ones in SDS-PAGE gels after staining with Coomas sie blue should have few surface-exposed regions. In relation to the comparative analysis of the protein profiles of the strains from distinct sources, previously categorized as virulent or avirulent, interest ing data have also been found. The distinction in the surface-exposed protein profiles detected, specifically in the band with an apparent molecular ratio of 28 kDa, in the commensal strain, in addition to the observation of the lowest agglutination titres in this strain with antibodies prepared against whole-cell
340
R. M. C. P. Domingues et al.
antigens of clinical strains seems to be relevant. It is tempting to speculate about a particular surface organization in this strain of commensal origin, in relation to strains from other sources. The phenotypical variability was al ready the subject of several studies with B. fragilis (3, 11, 19) and attempts to distinguish commensal strains from clinical isolates have been made in rela tion to various bacterial characters. Admitting the multiplicity of B. fragilis relations in nature, in balanced stages or not, our results allow us to propose a possible relation of the bacte rial surface composition and organization to the origin of the isolates as well as the virulence potential. Acknowledgements. We thank Joaquim Santos Filho for technical assistance. This work was supported by grants of the following institutions: CNPq, PRONEX, FUJB, FINEP and CEPG.
References 1. Abath, F. G. c., A. M. P. Almeida, and 1. C. S. Ferreira: Identification of surface exposed Yersinia pestis proteins by radio-iodination and biotinylation. J. Med. Microbiol. 37 (1992) 420-424 2. Bolin, I., 1. Norlander, and H. Wolf- Watz: Temperature-inducible outer membrane protein of Yersinia pseudotuberculosis and Yersinia enterocolitica is associated with the virulence plasmid. Infect. Immun. 17 (1982) 506-512 3. Domingues, R. M. C. P., M. Oliveira, R. HirataJr., A. F. B.Andrade, andM. C. S. Fer reira: Bacteroides fragilis: an exogenous pathogen? Zbl. Bakt. 282 (1995) 296-302 4. Donelli, G. , A. Fabbri, and C. Fiorentini: Bacteroides fragilis enterotoxin induces cytoskeletal changes and surface blebbing in HT-29 cells. Infect. Imrnun. 64 (1996) 113-119 5. Dice, 1. R.: Measurements of amounts of ecologic association between species. Ecology 26 (1945) 297-302 6. Elhag, K. M. and S. Tabaqchali: A study of the surface and somatic antigens of Bac teroides fragilis. J. Hyg. 180 (1978) 439-449 7. Ferreira, 1. C. S. and D. I. Almeida: Iodogen catalysed iodination for identification of surface-exposed outer membrane proteins of Escherichia coli K12. Braz. J. Ge netics 10 (1987) 625-634 8. Finegold, S. M. and D. M. Citron: Gram negative nonspore forming anaerobe bacilli. In Lenette, E. H., Ballows, A., Hausler Jr. , w.j. , and Truant, j. P. (eds.), Manual of Clinical Microbiology (3rd ed.): (1980) pp.187-192. American Society for Microbiology, Washington DC 9. Finlay, B. B. and S. Falkow: Common themes in microbial pathogenicity. Micro bioI. Rev. 53 (1989) 210-230 10. Godoy, V. G., M. M. Dallas, T. A. Russo, and M. M. Malany: A role for Bacteroides fragilis neuraminidase in bacterial growth in two model systems. Infect. Immun. 61 (1993) 4415-4426 11. Hofstad, T.: Virulence factors in anaerobic bateria . Eur. J. Clin. Microbiol. Infect. Dis. 11 (1992) 1044-1049 12. Holdeman, 1. v., E. P. Cato, and W. E. C. Moore: Anaerobe Laboratory Manual, 4th ed. (1977) Virginia Polytechnic Institute and State University, Blacksburg
Bacteroides fragilis exposed proteins
341
13. Kanazawa, K., Y. Kobayashi, M. Nakano, M. Sakurai, N. Gotoh, and T. Nishino: Identification of three porines in the outer membrane of Bacteroides fragilis. FEMS Microbiol. Lett. 127 (1995) 181-186 14. Laemmli, U. K.: Cleavage of structural proteins during the assembly of the head of bacteriophage T 4 • Nature 227 (1970) 680-685 15. Macfarlane, G. T. and S. Macfarlane: Formation of glycoprotein degrading en zymes by Bacteroides fragilis. FEMS Microbiol. Lett. 77 (1991) 289-294 16. Markwell, M. A. K. and C. F. Fox: Surface-specific iodination of membrane pro teins of viruses and eukaryotic cells using 1,3,4,6-tetrachloro-3-a,6-a,diphenylgly coluril. Biochem. 17 (1978) 4807-4817 17. Miles, A. A. and S. S. Misra: The estimation of the bactericidal power of the blood. J.Hyg.38 (1938) 732-749 18. Myers, 1. 1., D. S. Shoop, 1. 1. Stackhouse, F. S. Newman, R.]. Flaherty, G. W. Letson, and R. B. Sack: Isolation of enterotoxigenic Bacteroides fragilis from hu man with diarrhea. J. Clin. Microbial. 25 (1987) 2330-2333 19. Namavar,F., M.A.].]. Verweij-van Vught, and D.M. MacLaren: A study of the candidate virulence factors of Bacteroides fragilis. J. Gen. Microbiol. 137 (1991) 1431-1435 20. Otto, B. R., M. Sparrius, A. M.].f. Verweij-van Vught, and D. M. MacLaren: Iron regulated outer membrane protein of Bacteroides fragilis involved in heme uptake. Infect. Immun. 58 (1990) 3954-3958 21. Otto, B. R., W. R. Verweij, M. Sparrius, A. M. f.]. Verweij-van Vught, C. E. Nord, and D. M. MacLaren: Human immune response to an iron-repressible outer mem brane protein of Bacteroides fragilis. Infect. Immun. 59 (1991) 2999-3003 22. Patrick, S. and D. A. Lutton: Outer membrane proteins of Bacteroides fragilis grown in vivo. FEMS Microbial. Lett. 71 (1990) 1-4 23. Patrick, S., D. A. Lutton, and A. D. Crockard: Immune reactions of Bacteroides fragilis populations with three different types of capsule in a model of infection. Microbiol. 141 (1995) 1969-1976 24. Poxton, I. R. and R. Brown: Sodium dodecyl sulphate-polyacrylamide gel electro phoresis of cell surface proteins as an aid to the identification of the Bacteroides fragilis group. J. Gen. Microbial. 112 (1979) 211-217 25. Swason,f.: Surface-exposed protein antigens of the gonococcal outer membrane. Infect. Immun. 34 (1981) 804-812 26. Taylor, A. f., C. A. Dawson, and R.]. Owen: The identification of Bacteroides ureo lyticus from patients with non-gonococcal urethritis by conventional biochemical tests and by DNA and protein analyses. J. Med. Microbiol. 21 (1986) 109-116 27. Towbin, H., T. Stachelen, and]. Gordon: Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA 76 (1979) 4350-4354 28. Tzianabos, A. 0., A. B. Onderdonk, B. Rosner, R.1. Cisneros, and D. 1. Kasper: Structural features of polysaccharides that induce intra-abdominal abscesses. Sci ence 262 (1993) 416-419 29. van Doorn,]., B. Oudega, and D. M. MacLaren: Characterization and detection of a 40 kDa fimbrial subunit of Bacteroides fragilis BEL Microbial Pathogen. 13 (1992) 75-79
Corresponding author: Instituto de Microbiologia Prof. Paulo de Goes, Centro de Ciencias da Saude, Universidade Federal do Rio de Janeiro, Ilha do Fundao, 21941590, Rio de Janeiro, RJ, Brasil, Fax: 0055 (021) 560-8028, E-mail: immmcan@micro bio.ufrj.br 23
Zent.bl. Bakteriol. 287/4