Lipid modification of prelipoproteins is dispensable for growth in vitro but essential for virulence in Streptococcus pneumoniae

Lipid modification of prelipoproteins is dispensable for growth in vitro but essential for virulence in Streptococcus pneumoniae

FEMS Microbiology Letters 200 (2001) 229^233 www.fems-microbiology.org Lipid modi¢cation of prelipoproteins is dispensable for growth in vitro but e...

316KB Sizes 0 Downloads 45 Views

FEMS Microbiology Letters 200 (2001) 229^233

www.fems-microbiology.org

Lipid modi¢cation of prelipoproteins is dispensable for growth in vitro but essential for virulence in Streptococcus pneumoniae Chantal M. Petit a

a;

*, James R. Brown b , Karen Ingraham a , Alex P. Bryant a , David J. Holmes a

GlaxoSmithKline Pharmaceuticals, Anti-Microbial and Host Defense, 1250 South Collegeville Road, P.O. Box 5089, Collegeville, PA 19426-0989, USA b GlaxoSmithKline Pharmaceuticals, Bio-informatics, 1250 South Collegeville Road, P.O. Box 5089, Collegeville, PA 19426-0989, USA Received 4 December 2000; received in revised form 19 April 2001; accepted 4 May 2001 First published online 5 June 2001

Abstract A vlgt (Lgt, lipoprotein diacylglyceryl transferase) isogenic mutant was obtained which indicates that lgt is not essential for cell growth in vitro, like in the Gram-positive bacterium Bacillus subtilis, but unlike in the proteobacteria Escherichia coli and Salmonella typhimurium. The mutation was transduced to a virulent strain. A 5 log attenuation was observed in a respiratory tract model of infection. Metabolic labeling by [U-14 C]palmitate revealed the presence of eight to ten lipoproteins in the wild-type strain only, with molecular masses between 15 and 80 kDa. Our findings suggest a major difference in the role of lipoproteins in Gram-positive bacteria versus the proteobacteria. ß 2001 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Lipoprotein ; Lgt; Lipoprotein diacylglyceryl transferase; Virulence; Respiratory tract infection ; Streptococcus pneumoniae

1. Introduction Streptococcus pneumoniae is one of the leading causes of septicemia, meningitis, and lower respiratory tract infections in humans. A better understanding of the cellular biochemistry of the pneumococci is essential for the development of novel antimicrobial therapeutics. A key biochemical pathway is the lipid modi¢cation of proteins, which allows their localization to the bacterial cell surface. More than 150 bacterial lipoproteins of diverse structure and function have been identi¢ed to date. All lipoproteins have a characteristic lipobox sequence (commonly -Leu33 -Ser/Ala32 -Ala/Gly31 -Cys‡1 -) [1]. A common biosynthetic pathway for lipoproteins was proposed [2], and subsequently modi¢ed [3]. The ¢rst step is the transfer of the diacylglyceryl moiety from phosphatidylglycerol to the sulfhydryl group of the prospective N-terminal cysteine residue in the unmodi¢ed prolipoprotein by prolipoprotein diacylglyceryl transferase (Lgt) with the concomitant formation of Sn-glycerol-1-phosphate.

* Corresponding author. Tel. : +1 (610) 917 6643; Fax: +1 (610) 917 7901; E-mail : [email protected]

Diacylglyceryl modi¢cation is a prerequisite for the cleavage of the signal peptide by a lipoprotein-speci¢c signal peptidase called signal peptidase II [4,5]. The amino-terminal of the product, apolipoprotein, is further N-acylated by N-acyltransferase which has no speci¢city with respect to phospholipids as acyl donors [6,7]. This pathway appears to be essential in Escherichia coli and Salmonella typhimurium since mutants defective in the activity of any of these three enzymes are temperature sensitive in growth, suggesting that one or more lipoproteins are required for normal growth, division and viability [3,8^11]. Database searches indicate that the key enzyme in this pathway, Lgt, is present in a wide range of bacteria. However the biochemical function of Lgt in Gram-positive bacteria has been little studied. The presence of Lgt in important Gram-positive and Gram-negative human bacterial pathogens, in conjunction with the apparent absence of this protein in eukaryotes, suggests that Lgt might be a potential novel antimicrobial target. In this study we constructed an isogenic vlgt mutant and compared it to the wild-type strain for in vitro growth and for virulence in a respiratory tract model of infection. We show that lgt is not essential for cell growth in vitro but is essential for viability during infection.

0378-1097 / 01 / $20.00 ß 2001 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 0 1 ) 0 0 2 3 3 - 6

FEMSLE 9987 18-6-01

230

C.M. Petit et al. / FEMS Microbiology Letters 200 (2001) 229^233

2. Materials and methods

were taken at time intervals and serial dilutions were plated on TSA II medium.

2.1. Bacterial strains The laboratory strain S. pneumoniae R6 [12], and the virulent strain S. pneumoniae 0100993 (serotype 3) from GlaxoSmithKline's collection, were used in this study. S. pneumoniae strains were routinely cultured in AGCH medium [13] modi¢ed by adding 0.3% (w/v) sucrose and 0.2% (w/v) yeast extract. For enumeration, TSA II (trypticase soy agar) agar plates containing 5% (w/v) sheep blood were used (Becton Dickinson, Franklin Lakes, NJ, USA). Erythromycin was added, when required, at a concentration of 1 ug ml31 . 2.2. Allelic exchange mutagenesis DNA fragments of 522 and 241 bp £anking lgt were fused to an erythromycin resistance determinant (ermAM) [14,15] using crossover PCR [16]. The primers for the upstream £anker were 5P-TGGCAGAACGTACCAGTGTGCACGGTG-3P and 5P-CCGCCATTCTTTGCTGTTTCGTTTGCACCTCATTTCGAGCTATGAG-3P (the sequences in bold are the sequences complimentary to ermAM). The primers for the downstream £anker were 5P-GGAAAGTTACACACGTTACTAAAGGAGCTACCGTCAACCGACTTTCC-3P and 5P-GATCATTATACCGAGACCG-3P. These constructs were used to transform S. pneumoniae R6 by incubating 0.5 ug DNA with 106 bacterial cells at 30³C for 30 min in the presence of the Csp1 competence stimulating heptadecapeptide [17]. Cells were then incubated for a further 90 min at 37³C before plating on AGCH containing 1 ug erythromycin per ml. Transformants were assessed following growth at 37³C for 36 h in 5% (v/v) CO2 . Transformation frequency with control DNA was 103 transformants per ug DNA. Chromosomal DNA was prepared from erythromycin-resistant transformants of S. pneumoniae R6 and used to transform the pathogenic S. pneumoniae strain 0100993 in the presence of Csp1 [17]. Transformants containing the appropriate allelic replacement were con¢rmed by Southern analysis (not shown). In control experiments where S. pneumoniae 0100993 was transformed with chromosomal DNA from wildtype S. pneumoniae R6, no tranformants were obtained. S. pneumoniae 0100993 was also transformed with DNA from an allelic replacement mutant of malC in S. pneumoniae R6 at a frequency of 102 per Wg of DNA. 2.3. In vitro growth The wt and the lgt deletion mutant strains of S. pneumoniae 0100993 were grown from stock solutions in 20 ml modi¢ed AGCH medium. The medium was inoculated with 1U106 CFU ml31 of the respective strains. Aliquots

2.4. Respiratory tract infection The experiment was carried out in duplicate, using ¢ve animals per experiment. Bacteria were prepared from frozen stocks by inoculation of TSA II plates. Following overnight growth at 37³C in 5% (v/v) CO2 , bacteria were recovered from the plates, resuspended in phosphatebu¡erred saline (PBS) and adjusted to an OD600 of 0.90. The male CBA/J mice (14^16 g) were lightly anaesthetized with 3% (v/v) iso£urane. 50 Wl of the inoculum, which corresponds to 1U106 CFU on average was delivered intranasally. Mice were then killed at 48 h post-inoculation by CO2 overdose and lungs were aseptically removed and homogenized in 1 ml of PBS, using a Seward stomacher (Seward Ldt., London, UK). Viable bacteria were enumerated on TSA II plates. 2.5. Detection of lipoproteins For the detection of lipoproteins, S. pneumoniae 0100993 was cultivated in brain heart infusion (BHI) medium to which 2 WCi [U-14 C]palmitate was added. Cells were incubated for 5 h at 37³C. Cells were harvested, washed three times with 150 mM Tris^HCl, pH 8 bu¡er and resuspended in 100 Wl protoplast bu¡er (50 mM Tris^ HCl, pH 8, 15 mM MgCl2 , 66% (w/v) sucrose) containing 2.5 mg ml31 of lyzozyme. The cell suspension was incubated for 1 h at 37³C. The proteins were solubilized by adding 100 Wl of SDS extraction bu¡er (1.1 M sucrose, 500 mM Tris^HCl, 278 mM sodium dodecyl sulfate (SDS), 2 mM EDTA, 0.9 mM Serva Blue G250, pH 8.5) and boiling the samples for 10 min. 20 Wl of each sample was subjected to electrophoresis. Proteins were transferred to a PVDF membrane and detected by autoradiography. 3. Results 3.1. In vitro growth The wt and the lgt deletion mutant strains of S. pneumoniae 0100993 were grown in modi¢ed AGCH medium from stock solutions. 20 ml of fresh modi¢ed AGCH medium was inoculated to reach a cell concentration of approximately 1U106 CFU ml31 and was incubated at 37³C. Growth was recorded by measuring absorbance at 600 nm at time intervals. No di¡erence was observed between the two strains. A doubling time of 45 min for both strains was achieved during logarithmic phase. Furthermore, for both strains a population of 2U108 CFU ml31 was estimated after 6 h of incubation and before both strains entered the cell lysis stage commonly observed when growing S. pneumoniae (data not shown). As further evidence,

FEMSLE 9987 18-6-01

C.M. Petit et al. / FEMS Microbiology Letters 200 (2001) 229^233

Fig. 1. Growth curves of S. pneumoniae 0100993 wt (open symbols) and vlgt mutants (close symbols) on modi¢ed AGCH medium. OD600 , optical density at 600 nm.

another experiment was carried out for which growth was recorded during three passages. For the ¢rst passage, 20 ml of medium was inoculated as described above. When the culture reached an OD600 of 0.3 (mid-logarithmic phase), 200 ul of the culture was inoculated to 20 ml of fresh medium. This was repeated twice. The ¢nal culture was monitored until the end of the logarithmic phase. Again, no di¡erences were recorded (Fig. 1), indicating that the viability of the wt and lgt mutant strains is similar in vitro.

231

Fig. 3. Incorporation of [U-14 C]palmitate into lipoproteins. The samples were as follows : [14 C]protein molecular mass markers from Amersham (M), lipoproteins from S. pneumoniae 0100993 (lanes 1 and 2) and from S. pneumoniae 0100993 vlgt mutant (lanes 3 and 4). [U-14 C]palmitate was added at the time of inoculation (lanes 1 and 3) or when cells reached an OD600 of 0.4, followed by an incubation of 2 h (lanes 2 and 4). A control experiment in which proteins were detected by SDS^ PAGE followed by Coomassie staining was carried out to ensure equal loading (data not shown).

3.2. Respiratory tract infection To further ascertain the potential e¡ect of a deletion in the lgt gene, in vivo studies were carried out. The animals were infected with 1U106 CFU per animal on average. The number of bacteria recovered after infection was 6.3U107 CFU/lungs when animals were infected with the wt strain while it was below the limit of detection for four mice, and below 1U104 CFU/lungs for one mouse when animals were infected with the lgt deleted mutant (Fig. 2). In the second set of experiments, the number was below the limit of detection for the ¢ve mice (data not shown) while 1.0U108 CFU/lungs were recovered from the control experiment. These results indicate that the mutant is largely attenuated during respiratory tract infection. 3.3. Lipoprotein pro¢le SDS^PAGE pro¢le of [U-14 C]palmitate-labeled proteins extracted from S. pneumoniae 0100993 cells revealed the presence of eight to ten lipoproteins, with sizes ranging from approximatively 15 to 80 kDa, and with variable intensity. No radiolabeled proteins were detected in the vlgt mutant strain, thus con¢rming the inactivation of this gene (Fig. 3). 4. Discussion

Fig. 2. Respiratory tract infection of male CBA/J mice caused by S. pneumoniae 0100993 and 0100993 vlgt mutant. The dashed line represents the limit of detection.

Covalent modi¢cation of a cysteine residue with a glyceride thioester group was ¢rst described by Hantke and

FEMSLE 9987 18-6-01

232

C.M. Petit et al. / FEMS Microbiology Letters 200 (2001) 229^233

Braun [18] for the major outer membrane protein (Lpp) of E. coli. Since then, a substantial number of proteins considered to be lipoproteins have been identi¢ed in both Gram-positive and Gram-negative bacteria, directly from biochemical studies or indirectly from sequence analysis. These are less abundant than the Braun lipoprotein, but similar in structure [19]. Sutcli¡e and Russell [20] emphasized the role of lipid modi¢cation of proteins in cell surface localization in Gram-positive bacteria. For example, some L-lactamases in Bacillus licheniformis and Staphylococcus aureus have been shown to be membrane-associated lipoproteins. Also, some lipoproteins could act as substrate-binding molecules involved in transport systems or in chemoreception and sensory mechanisms. Some adhesins in Streptococcus sanguis, S. pneumoniae and Streptococcus gordonii were also reported to be lipoproteins. Considering that the three enzymes involved in this pathway, namely Lgt (prolipoprotein diacylglyceryl transferase), Lsp (prolipoprotein signal peptidase) and Lnt (apolipoprotein N-acyltransferase), appear to be essential in E. coli and S. typhimurium [3], we decided to evaluate the essentiality of the lgt gene/protein in the pathogenic Gram-positive bacterium S. pneumoniae. The lack of homogeneous enzyme preparations has precluded further biochemical characterization of Lgt. Our attempts to overexpress S. pneumoniae lgt were unsuccessful, as were our attempts to assay Lgt activity in vitro as described by Sankaran and Wu [3] and Gan et al. [21]. The results presented here indicate that Lgt is not an essential enzyme in S. pneumoniae, since lack of this enzyme does not result in cell death. In agreement with our results, recent studies indicated that lgt is not essential for growth in Bacillus subtilis [22]. The authors showed that a disruption in lgt completely abolished lipomodi¢cation of pre-

lipoproteins, such as PrsA (foldase involved in protein secretion) and BlaP (penicillinase), but that B. subtilis cells were still fully viable. The lgt mutation in S. pneumoniae R6 was transformed into the virulent strain 0100993, and both the wild-type strain and the transformant were compared for in vivo growth in a respiratory tract model of infection. The full attenuation observed would suggest that lgt is essential for S. pneumoniae to establish and survive infection. A search of GlaxoSmithKline proprietary protein sequence database of S. pneumoniae 0100993 revealed the presence of 17 proteins with a lipobox motif near the extreme N-terminus (Table 1). These included proteins homologous to S. pneumoniae AmiA-AliA-AliB oligopeptide transporter [20], Enterococcus faecalis D-alanyl-D-alanine carboxypeptidase, and proteins of unknown function from Haemophilus in£uenzae and B. subtilis. Interestingly, the search also identi¢ed PsaA, a known lipoprotein in S. pneumoniae involved in adhesion [23] and AdcA. For psaA, Berry and Paton [23] showed that two independent mutants were signi¢cantly less virulent as judged by intranasal or intraperitoneal challenge of mice than the wild-type. The AdcA protein was found to exhibit homology to ATP-binding cassette (ABC) transport operons encoding streptococcal adhesins such as FimA in Streptococcus parasanguis and ScaA in S. gordonii [24]. The PsaA and AdcA prelipoproteins have molecular masses of 34.6 and 56.2 kDa, respectively. The pattern of lipoproteins observed after in vivo labeling by [U-14 C]palmitate indicated the presence of two bands migrating close to the 30 kDa marker and between the 46 and 66 kDa markers (Fig. 3). These bands could correspond to the mature PsaA and AdcA lipoproteins for which we estimated the molecular mass at 33 and 55 kDa, respectively.

Table 1 Putative prelipoproteins identi¢ed by genome analysis of S. pneumoniae 0100993a N-terminal sequence of putative prelipoprotein

MW (kDa)

Nearest homolog

MRKNRVFATAGLVLLASGVLAACSCSKSSDSSAHK VVLSTSAILVACGKTDKEADAPTTFSYVYA MKKSKSKYLTLAGLVLGTGVLLSACGNSST MKKLGTLLVLFLSAIILVACASGKKDTTSG MSSKFMKSTAVLGTVTLASLLLVACGSKTA MKFRKLACTVLAGAAVLGLAACGNSGGSKD VKKISLLLASLCALFLVACSNQKQADGKLN MKKWMLVLVSLMTALFLVACGKNSSETSGD MKNWKKYAFASASVVALAAGLAACGNLTGN MNKKQWLGLGLVAVAAVGLAACGNRSSRNA VIIMKFKKMLTLAAIGLSGFGLVACGNQSA

72.5 25.8 72.6 34.6 45.4 48.3 56.2 31.0 54.5 36.8 30.7

MKKRYLVVTALLALSLAACSQEKAKNEDGT MKKKLLAGAITLLSVATLAACSKGSEGADL MKKWQTCVLGAGSLLCLTACSGKSVTSEHQ MKKTWKVFLTLVTALVAVVLVACGQGTASK MKTSLKLYFTALVASFLLLLGACSTNSSTS MKKKFALSFVALASVXLLAACGEVKSGAVN

26.5 34.0 20.8 37.8 34.9 40.4

S. pneumoniae oligo-binding protein AmiA S. pneumoniae oligopeptide-binding protein AliA S. pneumoniae oligopeptide-binding protein AliB S. pneumoniae adhesion protein PsaA S. pneumoniae maltose/maltodextrine-binding protein MalX S. pneumoniae maltose/maltodextrine-binding protein S. pneumoniae solute-binding protein AdcA B. subtilis probable amino acid ABC transporter-binding protein B. subtilis LplA precursor (function unknown) B. subtilis hypothetical lipoprotein YufN B. subtilis probable ABC transporter-binding protein in sodA-comA intergenic region Enterococcus faecium D-alanyl-Dalanine carboxypeptidase Lactobacillus paracasei protease maturation protein H. in£uenzae hypothetical protein HI1453 H. in£uenzae putative thiamine biosynthesis protein Vibrio anguillarum ferric aguibactin-binding protein P. aeruginosa component of the high a¤nity Leu, Ile, Thr, Val transport system

a

The lipobox consensus sequences are underlined.

FEMSLE 9987 18-6-01

C.M. Petit et al. / FEMS Microbiology Letters 200 (2001) 229^233

From this work, it is clear that the role(s) of lipoproteins in Gram-positive bacteria and the proteobacteria is (are) di¡erent. This was ¢rst demonstrated in the nonpathogenic bacterium B. subtilis and is con¢rmed herein for the human pathogen S. pneumoniae. Nevertheless, our ¢ndings suggest that despite its non-essential nature for growth in vitro, the lgt gene is essential for the bacterium to establish and survive infection. This further supports Lgt as an attractive and broad spectrum antibacterial target. Acknowledgements We thank Marty Rosenberg for helpful discussions and support. References [1] Braun, V. and Wu, H.C. (1993) Lipoproteins: structure, function, biosynthesis and model for protein export. Comp. Biochem. 27, 319^342. [2] Tokunaga, M., Tokunaga, H. and Wu, H.C. (1982) Post-translational modi¢cation and processing of Escherichia coli prolipoprotein in vitro. Proc. Natl Acad. Sci. USA 79, 2255^2259. [3] Sankaran, K. and Wu, H.C. (1994) Lipid modi¢cation of bacterial prolipoprotein - transfer of diacylglyceryl moiety from phosphatidylglycerol. J. Biol. Chem. 269, 19701^19706. [4] Hussain, M., Ichihara, S. and Mizushima, S. (1980) Accumulation of glyceride-containing precursor of the outer membrane lipoprotein in the cytoplasmic membrane of Escherichia coli treated with globomycin. J. Biol. Chem. 255, 3707^3712. [5] Tokunaga, M., Loranger, J.M. and Wu, H.C. (1984) Prolipoprotein modi¢cation and processing enzymes in Escherichia coli. J. Biol. Chem. 259, 3825^3830. [6] Gupta, S.D., Dowhan, W. and Wu, H.C. (1991) Phosphatidylethanolamine is not essential for the N-acylation of apolipoprotein in Escherichia coli. J. Biol. Chem. 266, 9983^9986. [7] Hussain, M., Ichihara, S. and Mizushima, S. (1982) Mechanism of signal peptide cleavage in the biosynthesis of the major lipoprotein of the Escherichia coli outer membrane. J. Biol. Chem. 257, 5177^5182. [8] Gan, K., Gupta, S.D., Sankaran, K. and Wu, H.C. (1993) Isolation and characterization of a temperature-sensitive mutant of Salmonellatyphimurium defective in prolipoprotein modi¢cation. J. Biol. Chem. 268, 16544^16550. [9] Gupta, S.D., Gan, K., Schmid, M.B. and Wu, H.C. (1993) Characterization of a temperature-sensitive mutant of Salmonella-typhimurium defective in apolipoprotein n-acyltransferase. J. Biol. Chem. 268, 16551^16556.

233

[10] Williams, M.G., Fortson, M., Dykstra, C.C., Jensen, P. and Kushner, S.R. (1989) Identi¢cation and genetic mapping of the structural gene for an essential Escherichia coli membrane protein. J. Bacteriol. 171, 565^568. [11] Yamagata, H., Ippolito, C., Inukai, M. and Inouye, M. (1982) Temperature-sensitive processing of outer membrane lipoprotein in an Escherichia coli mutant. J. Bacteriol. 152, 1163^1168. [12] Avery, O.T., MacLeod, C.M. and McCarty, M. (1944) Studies on the chemical nature of the substance inducing transformation of pneumococcal types. J. Exp. Med. 79, 137^158. [13] Lacks, S. (1966) Integration e¤ciency and genetic recombination in pneumococcal transformation. Genetics 53, 207^235. [14] Lunsford, R.D. and London, J. (1996) Natural genetic transformation in Streptococcus gordonii: comX imparts spontaneous competence on strain Wicky. J. Bacteriol. 178, 5831^5835. [15] Martin, B., Alloing, G., Mejean, V. and Claverys, J.P. (1987) Constitutive expression of erythromycin resistance mediated by the ermAM determinant of plasmid pAML1 results from deletion of 5P leader peptide sequences. Plasmid 18, 250^253. [16] Link, A.J., Phillips, D. and Church, G.M. (1997) Methods for generating precise deletions and insertions in the genome of wild-type Escherichia coli: application to open reading frame characterization. J. Bacteriol. 179, 6228^6237. [17] Haverstein, C. and Morrison, D. (1995) An unmodi¢ed heptadecapeptide pheromone induces competence for genetic transformation in Streptococcus pneumoniae. Proc. Natl. Acad. Sci. USA 92, 11140^ 11144. [18] Hantke, K. and Braun, V. (1973) Covalent binding of lipid to protein. Diglyceride and amide-linked fatty acid at the N-terminal end of the murein-lipoprotein of the Escherichia coli outer membrane. Eur. J. Biochem. 34, 284^296. [19] Nielsen, J.B.K. and Lampen, J.O. (1982) Glyceride-cysteine lipoproteins and secretion by Gram-positive bacteria. J. Bacteriol. 152, 315^ 322. [20] Sutcli¡e, I.C. and Russell, R.R.B. (1995) Lipoproteins of Gram-positive bacteria. J. Bacteriol. 177, 1123^1128. [21] Gan, K., Sankaran, K., Williams, M.G., Aldea, M., Rudd, K., Kushner, S.R. and Wu, H.C. (1995) The umpA gene of Escherichia coli encodes phosphatidylglycerol :prolipoprotein diacylglyceryl transferase (lgt) and regulates thymidylate synthase levels through translational coupling. J. Bacteriol. 177, 1879^1882. [22] Leskela, S., Wahlstrom, E., Kontinen, V.P. and Sarvas, M. (1999) Lipid modi¢cation of prelipoproteins is dispensable for growth but essential for e¤cient protein secretion in Bacillus subtilis : characterization of the lgt gene. Mol. Microbiol. 31, 1075^1085. [23] Berry, A.M. and Paton, J.C. (1996) Sequence heterogeneity of PsaA, a 37-kilodalton putative adhesin essential for virulence of Streptococcus pneumoniae. Infect. Immun. 64, 5255^5262. [24] Dintihac, A. and Claverys, J.P. (1997) The adc locus, which a¡ects competence for genetic transformation in Streptococcus pneumoniae, encodes an ABC transporter with a putative lipoprotein homologous to a family of Streptococcal adhesins. Res. Microbiol. 148, 119^131.

FEMSLE 9987 18-6-01