Molecular Microbiology (2002) 43(1), 147–157
The Fap1 fimbrial adhesin is a glycoprotein: antibodies specific for the glycan moiety block the adhesion of Streptococcus parasanguis in an in vitro tooth model Aimee E. Stephenson,1 Hui Wu,2 Jan Novak,3 Milan Tomana,4 Keith Mintz1 and Paula Fives-Taylor1* Departments of 1Microbiology and Molecular Genetics, and 2Medicine, University of Vermont, Burlington, VT, USA. Departments of 3Microbiology and 4Medicine, Division of Clinical Immunology and Rheumatology, University of Alabama at Birmingham, AL, USA. Summary Streptococcus parasanguis is a primary colonizer of the tooth surface and plays a pivotal role in the formation of dental plaque. The fimbriae of S. parasanguis are important in mediating adhesion to saliva-coated hydroxylapatite (SHA), an in vitro tooth adhesion model. The Fap1 adhesin has been identified as the major fimbrial subunit, and recent studies suggest that Fap1 is a glycoprotein. Monosaccharide analysis of Fap1 purified from the culture supernatant of S. parasanguis indicated the presence of rhamnose, glucose, galactose, N-acetylglucosamine and N-acetylgalactosamine. A glycopeptide moiety was isolated from a pronase digest of Fap1 and purified by immunoaffinity chromatography. The monosaccharide composition of the purified glycopeptide was similar to that of the intact molecule. The functionality of the glycan moiety was determined using monoclonal antibodies (MAbs) specific for the intact Fap1 glycoprotein. These antibodies were grouped into two categories based on their ability to block adhesion of S. parasanguis to SHA and their corresponding specificity for either protein or glycan epitopes of the Fap1 protein. ‘Non-blocking’ MAb epitopes were mapped to unique protein sequences in the Nterminus of the Fap1 protein using non-glycosylated recombinant Fap1 proteins (rFap1 and drFap1) expressed in Escherichia coli. In contrast, the ‘blocking’ antibodies did not bind to the recombinant Fap1 proteins, and were effectively competed by the binding to the purified glycopeptide. These data suggest that
Accepted 27 September, 2001. *For correspondence. E-mail
[email protected]; Tel. (+1) 802 656 1121; Fax (+1) 802 656 8749.
© 2002 Blackwell Science Ltd
the ‘blocking’ antibodies are specific for the glycan moiety and that the adhesion of S. parasanguis is mediated by sugar residues associated with Fap1.
Introduction The sanguis streptococci are primary colonizers of the tooth surface in humans (Carlsson et al., 1970). They rapidly accrete on oral surfaces and provide a binding substrate for other oral microorganisms. If left undisturbed, this accumulation on the tooth surface leads to the formation of dental plaque, a biofilm composed of bacteria mixed with salivary and serum molecules (Marsh and Bradshaw, 1995). Dental plaque can lead to undesirable disease states in the oral cavity, most notably dental caries and periodontal disease. When oral bacteria are introduced into the bloodstream by way of dental manipulations, other parts of the body are at risk for colonization as well. Multiple adhesins are utilized by the sanguis streptococci to secure binding on oral surfaces (Hasty et al., 1992; Whittaker et al., 1996; Cisar et al., 1997; Jenkinson and Lamont, 1997). In S. parasanguis FW213, one of these adhesins is the long, peritrichous fimbriae that mediates adhesion to saliva-coated hydroxylapatite (SHA), an in vitro tooth model (Fives-Taylor, 1982). Chemically derived mutants of FW213 that lack these fimbriae demonstrate a reduced adhesion to SHA compared with the wild type (Fives-Taylor and Thompson, 1985). Fimbriae-specific antiserum, produced by serial agglutination with a non-fimbriated mutant, inhibits the adhesion of FW213 to SHA by more than 90% (Fachon-Kalweit et al., 1985). In addition, five monoclonal antibodies (MAb) and two monospecific Fabs (fragment antigen binding) are reactive with the adhesive, fimbriated FW213 but fail to react with non-adhesive, non-fimbriated mutants (Fachon-Kalweit et al., 1985; Stephenson et al., 1998). One antibody in particular, MAb F51, is capable of blocking adhesion of S. parasanguis to SHA. MAb F51 is specific for Fap1 and has been used to purify the adhesin from the culture medium of FW213 (Wu et al., 1998). Nterminal sequencing of the purified Fap1 protein enabled cloning and characterization of the fap1 gene, and mutants created by insertional inactivation of fap1 lack long fimbriae and demonstrate reduced adherence to
148 A. E. Stephenson et al. Table 1. Antibodies used in this study. Antibody
Type
Adhesion
Reference or source
B20 D10 F51 A6 E42 Ph5 Ph12 rFap 1-specific Fimbrial-specific
MAb MAb MAb MAb MAb MAb MAb PAb PAb
Blocking Blocking Blocking Non-blocking Non-blocking Non-blocking Non-blocking NA NA
Elder and Fives-Taylor (1986) Elder and Fives-Taylor (1986) Elder and Fives-Taylor (1986) Elder and Fives-Taylor (1986) Elder and Fives-Taylor (1986) Stephenson et al. (1998) Stephenson et al. (1998) This study Fachon-Kalweit et al. (1985)
NA, not applicable. MAb designates monoclonal antibodies and PAb designates polyclonal antibodies. Fap1specific MAbs are categorized by their ability to block adhesion to SHA or not.
SHA (Wu et al., 1998). Together, these data support the idea that the Fap1 adhesin is the major fimbrial subunit. The fap1 open reading frame (ORF) is 7659 nucleotides in length and is predicted to encode a 264 kDa protein. The protein contains an unusually long signal sequence (50 amino acid residues), a cell wall sorting signal (LP(X)TG) and two repeat regions. Repeat regions I and II have a similar dipeptide composition (E/V/I)S and comprise 28 and 1000 repeats respectively. Combined, the two repeat regions account for 80% of the total Fap1 protein. The high degree of conservation of the repeat unit throughout the repeat region suggests that repetitive blocks are structurally important (Wu and Fives-Taylor, 1999). The association of a glycan moiety with the Fap1 protein is suggested by the inability to stain this protein in polyacrylamide gels. Fap1 stains poorly with Coomassie blue and is not detectable by silver staining (Wu et al., 1998). Pretreatment of gels with 1% periodic acid improves silver staining ability and produces intense staining of Fap1 with periodic acid–Schiff (PAS) reagent. Oxidation of the Fap1 protein with sodium meta-periodate results in reduced reactivity with MAb F51 in Western immunoblots. These results suggest that the reactivity of MAb F51 is specific for glycan epitopes of the Fap1 protein. Little is known about the antigenic specificity of the other MAbs. This study was undertaken to examine the role of glycan and protein epitopes of the Fap1 protein in mediating adhesion to SHA using monospecific antibodies. The glycan epitopes of Fap1 were isolated and their linkage to the polypeptide backbone of Fap1 was investigated. The data presented provide additional evidence that Fap1 is a glycoprotein.
Results Antibody characterization The antibodies used in this study are summarized in Table 1. Three monospecific antibodies (F51, Ph5 and
Ph12) had been characterized previously with specificity for the Fap1 protein that runs at an apparent molecular weight of 200 kDa on SDS–PAGE (Stephenson et al., 1998; Wu et al., 1998). Western blot analysis revealed that the four remaining MAbs (D10, B20, E42 and A6) also reacted with the 200 kDa protein, thus demonstrating specificity for the Fap1 protein (data not shown). Electron and fluorescence microscopy of fimbriae Electron micrographs of whole S. parasanguis cells demonstrate the peritrichous fimbriae that coat these bacteria. Free fimbriae isolated from S. parasanguis cells were also visualized by electron microscopy (Fig. 1). The Fap1-specific MAbs were used to visualize the surface of S. parasanguis cells using fluorescence microscopy. Reactivity of MAb F51 and MAb E42 with whole S. parasanguis cells was detected with an Oregon Green® fluorescent secondary antibody. The resulting images demonstrate an even coating of antibody around the cell surface, presumably corresponding to the peritrichous fimbriae that Fap1 comprises (Fig. 2). VT1393, a fap1 insertion mutant, was used as a negative control. Characterization of the rFap1 and drFap1 recombinant proteins expressed in Escherichia coli Expression of an N-terminal region of fap1 in E. coli resulted in a 36 kDa protein designated rFap1. Further modification of the rFap1 construct was undertaken to delete repeat element I, resulting in a clone that expressed only unique protein sequence. The resulting 30 kDa protein was designated drFap1. Previous studies suggest Fap1 is a glycoprotein (Fachon-Kalweit, 1985; Wu et al., 1998). However, the rFap1 and drFap1 proteins migrated at their predicted molecular weights in SDS–PAGE and, therefore, did not appear to be glycosylated in E. coli. To confirm that the rFap1 protein was representative of the intact 200 kDa Fap1 found in wild-type FW213 cells, © 2002 Blackwell Science Ltd, Molecular Microbiology, 43, 147–157
The Fap1 fimbrial adhesin is a glycoprotein
149
B20) failed to react with both rFap1 and drFap1 (Fig. 4). This result may suggest that MAbs F51, D10 and B20 are specific for glycan epitopes of the Fap1 protein. Alternatively, they may be specific for protein epitopes not represented in rFap1 and drFap1. Adhesion blocking ability of antibodies MAb F51 is capable of blocking adhesion of S. parasanguis FW213 to SHA (Fachon-Kalweit et al., 1985). The other purified MAbs were tested for their ability to block adhesion of FW213 to determine if their respective epitopes were important in mediating adhesion to SHA. Adhesion blocking was tested at three concentrations (25, 50 and 100 mg ml–1) for each MAb examined (D10, B20 E42, A6, Ph5 and Ph12). All three concentrations of MAb D10 blocked adhesion equally, whereas adhesion blocking with MAb B20 was dose dependent. The antibodies specific for unique protein epitopes of Fap1, MAbs E42, A6, Ph5 and Ph12 failed to block adhesion at all concentrations (Fig. 5). Treatment of purified Fap1 with N- and O-glycanase Fig. 1. Electron micrographs of fimbriae of S. parasanguis FW213. A. Intact fimbriae on the cell surface of S. parasanguis. B. Purified fimbriae of S. parasanguis FW213. Bars = 0.1 mm.
a rFap1-specific antibody was affinity purified from antiserum made against whole FW213 cells. In Western blots, this antibody reacted strongly with rFap1 as expected, but it also bound to the 200 kDa Fap1 (Fig. 3, lanes 1 and 2). Conversely, a fimbria-specific antibody that reacts with the 200 kDa Fap1 was used to probe rFap1 in Western blots. This antibody recognized the 200 kDa Fap1 as well as rFap1 (Fig. 3, lanes 3 and 4). These data demonstrated that rFap1 contains epitopes found in the intact 200 kDa Fap1 protein. Reactivity of Fap1-specific MAbs for rFap1 and drFap1 Western blot analysis of rFap1 and drFap1 (recombinant proteins expressed in E. coli) was used to determine whether the Fap1-specific MAbs recognized epitopes in the unique or repetitive regions of the Fap1 protein. Of the seven MAbs examined, A6, E42, Ph5 and Ph12 reacted with rFap1. This same group of antibodies also reacted with drFap1, indicating they were specific for epitopes in the unique protein sequence of Fap1 (Fig. 4). The two bands represent rFap1 or drFap1, with and without the pelB leader sequence encoded by the pET27b(+) cloning vector. In contrast, three antibodies (MAbs F51, D10 and © 2002 Blackwell Science Ltd, Molecular Microbiology, 43, 147–157
Purified Fap1 was treated with N-glycanase and Oglycanase to determine the type(s) of glycoprotein linkage(s) present. Reactions were carried out according to the manufacturer’s instructions and the resulting products were analysed by polyacrylamide gel electrophoresis. N-glycanase treatment of Fap1 had no affect on the electrophoretic mobility of the protein. In contrast, the enzyme was effective in removing N-linked glycans from the positive control protein-ovalbumin, as demonstrated by an increase in the electrophoretic mobility (data not shown). These data suggest that Fap1 glycan was not attached through a traditional N-linkage. Treatment with neuraminidase alone, or with a combination of neuraminidase and O-glycanase, did not change the gel mobility of the Fap1 protein. Failure of neuraminidase to affect Fap1 suggested that this protein does not contain terminal sialic acid residues that need to be removed before recognition by O-glycanase. The positive control, fetuin, demonstrated gel mobility shifts after treatment with both neuraminidase and O-glycanase, indicating the enzymes were active. Oglycanase has a narrow specificity, releasing only galactose b1–3-N-acetylgalactosamine linked to either serine or threonine. Unlike N-glycans, O-glycans have fewer structural rules and, as a result, demonstrate much more variability in both eukaryotic and prokaryotic organisms (Schachter and Brockhausen, 1992; Messner, 1997). Therefore, the treatment with O-glycanase was inconclusive.
150 A. E. Stephenson et al.
A
Fig. 2. Fluorescent and phase images of S. parasanguis FW213. A. FW213 probed with MAb F51. B. FW213 probed with MAb E42. C. VT1393 probed with MAb F51 and E42.
B
C
Monosaccharide composition analysis of the Fap1 protein and the glycopeptide of Fap1 purified by F51 affinity chromatography The monosaccharide composition of purified Fap1 was determined by gas–liquid chromatography. This analysis revealed the presence of several different isomers of glucose, galactose, N-acetylglucosamine, N-acetylgalactosamine, and rhamnose at a ratio of 29:5:39:1:1. Mannose and sialic acid, monosaccharides commonly found in mammalian glycoproteins, were not detected. This result supports the hypothesis that Fap1 is a glycoprotein. Purified Fap1 was treated with pronase to cleave the polypeptide backbone and isolate the glycan moiety with minimal peptide attachment (Robertson and Kennedy, 1996). The glycopeptide was further purified on
an F51 immunoaffinity column and the isolated glycopeptide was analysed by gas–liquid chromatography for monosaccharide composition. The isolated glycan was found to have a profile almost identical to that of the purified Fap1 protein. Analysis of amino acid composition of the isolated glycan detected predominantly four amino acids (serine, glutamate, valine and isoleucine; Fig. 6). These amino acids are present in high concentrations in the Fap1 repeat sequences (Wu and Fives-Taylor, 1999). Competitive binding assays with Fap1-specific MAbs and the isolated glycopeptide of Fap1 Competitive binding assays were used as an indirect method to determine if MAbs were specifically binding to © 2002 Blackwell Science Ltd, Molecular Microbiology, 43, 147–157
The Fap1 fimbrial adhesin is a glycoprotein
Anti-rFap1
Fimbriae Specific Antiboby
151
Fig. 5. Adhesion blocking assays with Fap1-specific MAbs. Adhesion blocking assays were carried out with six monoclonal antibodies (MAb) specific for the Fap1 adhesin protein (D10, B20, A6, E42, Ph5 and Ph12). The ability of a MAb to block adhesion to saliva-coated hydroxylapatite (SHA) was tested at three concentrations (25, 50 and 100 mg ml–1) and compared with adhesion without MAb (0 mg ml–1). Percentage adhesion was determined by calculating the counts per minute for cells bound to SHA as a proportion of the total counts per minute (bound and unbound).
Fig. 3. Western blot analysis of Fap1 and rFap1with polyclonal antibodies. The crude ‘fimbrial’ preparation containing Fap1 (lanes 1 and 3) and purified rFap1 (lanes 2 and 4) probed with an antirFap1 antibody or fimbrial-specific antibody.
Fig. 6. Amino acid composition analysis of the isolated glycopeptide of Fap1.The hatched bars are the four amino acids (S, serine; E, glutamate; V, valine; and I, isoleucine) present in the Fap1 repeat sequence.
Fig. 4. Western blot analysis of rFap1 and drFap1. The rFap1 (r) and drFap1 (Dr) proteins were probed with seven Fap1specific MAbs. The two bands in each lane represent proteins with or without the PelB leader sequence that was encoded by the pET27b(+) expression vector.
the isolated glycan. In these assays, FW213 cells with Fap1 present on the surface were immobilized on enzyme-linked immunosorbent assay (ELISA) plates. Isolated glycan or purified Fap1 mixed with a subsaturating concentration of MAb was applied, and the binding of MAb to immobilized Fap1 was measured. Binding of the MAb to the immobilized cells was used to establish 100% binding. A decrease in MAb binding to immobilized Fap1 indicated competition by isolated glycan or purified Fap1. Isolated glycan and purified Fap1 were tested at four concentrations for their ability to compete with MAbs F51, © 2002 Blackwell Science Ltd, Molecular Microbiology, 43, 147–157
D10 and B20 for binding. Increasing concentrations of isolated glycan and purified Fap1 resulted in decreasing MAb D10 and MAb F51 binding to immobilized Fap1. The competitive binding curves for isolated glycan versus purified Fap1 were nearly identical, indicating that these MAbs are specific for epitopes present on both the isolated glycan and intact protein (Fig. 7A and B). Results with MAb B20 were inconclusive as a result of the low affinity of this antibody for Fap1 (data not shown). The binding of an antibody specific for the polypeptide backbone of Fap1 (MAb A6) was not competed by the isolated glycan (Fig. 7C). Discussion Streptococcus parasanguis FW213 binds to SHA, an in
152 A. E. Stephenson et al.
Fig. 7. Competitive binding assays with purified Fap1 and isolated glycopeptide. Purified Fap1 and isolated glycopeptide were tested at four MAb concentrations: 0.18, 0.018, 0.0018 and 0.00018 mM. A, MAb F51; B, MAb D10; and C, MAb A6.
vitro model of the enamel pellicle. The long fimbriae of FW213 mediate adhesion to SHA (Fachon-Kalweit et al., 1985). The Fap1 adhesin protein is essential for fimbrial biogenesis and for adhesion to SHA, possibly as the major fimbrial subunit. Previous studies with periodate
and protein staining suggest that Fap1 is a glycoprotein that consists of both protein and glycan moieties (Wu et al., 1998). MAb F51 is capable of blocking FW213 adhesion to SHA (Fachon-Kalweit et al., 1985). Furthermore, the signals in Western blots with MAb F51 are diminished by periodate treatment, suggesting that F51 may be specific for glycan moieties of Fap1 (Wu et al., 1998). However, periodate treatment may lead to the oxidization of certain amino acids, including serine, threonine and hyroxylysine, resulting in reactive aldehyde groups that can give false positives (Clamp and Hough, 1965). Thus, these studies should be interpreted with caution. The epitopes of Fap1 that are important in adhesion to SHA were studied using a panel of Fap1-specific antibodies. The antibodies used in this study were produced using two techniques: phage display technology (Barbas et al., 1991) and standard hybridoma methodology (Kohler and Milstein, 1975). Two phage display Fabs, Ph5 and Ph12, were derived from mice immunized with purified Fap1 protein (Stephenson et al., 1998). Five hybridoma monoclonal antibodies (A6, B20, D10, E42 and F51) were derived from mice immunized with whole FW213 cells (Elder and Fives-Taylor, 1986). In both cases, reactive antibodies were isolated by screening against adhesive (+ fimbriae) and non-adherent (– fimbriae) strains of S. parasanguis FW213. The specificity of MAbs F51, Ph5, and Ph12 for the Fap1 protein was demonstrated previously (Stephenson et al., 1998; Wu et al., 1998). In this study, the specificity of MAbs D10, A6, B20, and E42 for the Fap1 protein was demonstrated. Electron microscopy was used to visualize the long, peritrichous fimbriae of S. parasanguis. Fluorescent images generated by probing whole S. parasanguis cells with anti-Fap1 MAbs demonstrate an even coating of antibody around the cell surface. We interpret this coat as corresponding to the peritrichous fimbriae. A fap1 insertion mutant that lacks fimbriae failed to produce similar images. Together, these data provide further support for the idea that Fap1 is the major fimbrial subunit. The seven Fap1-specific monoclonal antibodies (MAbs) were used to identify the nature of the Fap1 epitopes (i.e. protein or glycan) and to study their roles in mediating adhesion to SHA. MAbs specific for glycan epitopes of Fap1 were capable of blocking adhesion of FW213 to SHA (‘blocking’ MAbs) whereas antibodies specific for protein backbone were ineffective in blocking adhesion (‘non-blocking’ MAbs; Table 1). The epitopes corresponding to the ‘non-blocking’ MAbs were mapped within unique protein sequences of Fap1, based on their reactivity with rFap1 and drFap1. More interesting was the failure of the ‘blocking’ MAbs to react with rFap1. This could be interpreted to mean that the ‘blocking’ MAbs are
© 2002 Blackwell Science Ltd, Molecular Microbiology, 43, 147–157
The Fap1 fimbrial adhesin is a glycoprotein specific for a unique sequence not present in this protein, or that these MAbs were binding to glycan epitopes of the Fap1 protein. Previous studies with MAb F51 suggested the latter was the more probable scenario. Until recently, prokaryotes were thought to be incapable of synthesizing glycoproteins because they lack the equivalent cellular structure associated with glycosylation in eukaryotic cells. Although reports of glycoproteins in bacteria date back to 1974, their presence in prokaryotes has not been widely accepted until recently (Wieland, 1988; Lechner and Wieland, 1989; Messner, 1997; Moens and Vanderleyden, 1997). Prokaryotic glycoproteins are not entirely unlike their eukaryotic counterparts in that they appear to have classical N- and O-glycan linkages. However, many different structures have been observed in prokaryotes with a much greater overall variability than that found in eukaryotes. Cell-associated and secreted glycoproteins have been identified in numerous bacterial species (archaea and eubacteria), and in several cases, glycosylation of fimbriae (pili) has been reported (Tomoeda and Inuzuka, 1975; Dobson and McCurdy, 1979; Weerkamp et al., 1984; Castric, 1995; Stimson et al., 1995). The linkage of the glycan moiety of Fap1 was characterized using enzymes specific for N- and O-linkages. These enzymes are employed routinely to remove Nlinked and O-linked glycans from eukaryotic glycoproteins and have specific oligosaccharide substrates. All N-linked glycans in eukaryotes share a common core structure that initiates with N-acetylglucosamine linked to asparagine residues in the polypeptide backbone. N-glycanase is used to cleave the N-glycosidic linkage between Nacetylglucosamine and asparagine. Treatment of Fap1 with N-glycanase did not result in a visible shift in mobility on SDS–PAGE, indicating that little or no classically N-linked glycans are associated with Fap1. O-linked glycans have fewer structural rules and do not share a common core structure, but are based on a number of different cores. Fewer enzymes are available to cleave specific O-glycosidic linkages. The most commonly employed is O-glycanase, which has narrow substrate specificity, releasing only galactose b1–3-Nacetylgalactosamine linked to either serine or threonine residues in the polypeptide backbone. The glycoproteins are typically pretreated with neuraminidase to remove sialic acid residues that extend off the core disaccharide that is recognized by the enzyme. Treatment of Fap1 with neuraminidase did not enhance cleavage with Oglycanase; still no release of O-glycans was observed. However, these results cannot rule out the possibility that O-linked glycans are associated with the Fap1 protein. The glycopeptide of the Fap1 protein was isolated by pronase digestion to degrade the polypeptide backbone
© 2002 Blackwell Science Ltd, Molecular Microbiology, 43, 147–157
153
of Fap1. The resulting Fap1 glycopeptide competed equally with intact Fap1 for binding to the ‘blocking’ MAbs, indicating the specificity of these antibodies for the glycan moiety of Fap1. In contrast, an antibody specific for protein epitopes of Fap1 (MAb A6) was exclusively competed by intact protein and was not competed by the Fap1 glycopeptide. This suggests that the polypeptide backbone of Fap1 was absent following degradation with pronase. Taken together with the adhesion blocking data, these results suggest the importance of the Fap1 glycan moiety in mediating adhesion to SHA. However, we cannot rule out the possibility that the protein backbone is involved in the adhesion process as well. Monosaccharide composition profiles for isolated glycopeptide and intact Fap1 were almost identical. The analysis indicated the presence of glucose, galactose, N-acetylglucosamine, N-acetylgalactosamine and rhamnose. In general, glucose and N-acetylglucosamine appeared to be more abundant, followed by galactose, rhamnose, and N-acetylgalactosamine. Monosaccharides commonly found in eukaryotic glycoproteins (mannose, fucose and sialic acid) were not detected, indicating that the sample was free of contaminating proteins such as antibodies. The presence of rhamnose is interesting, but not unique (Moens and Vanderleyden, 1997). The platelet aggregation-associated protein (PAAP) from Streptococcus sanguis has been purified and carbohydrate polymers linked by an N-asparaginyl linkage were found to be rich in rhamnose (Erickson and Herzberg, 1993). Whereas monosaccharides of Fap1 may comprise several different glycan structures, it is more likely that they constitute one type of glycan chain representing the major carbohydrate constituent of the Fap1 protein. Interestingly, the F51immunoaffinity purified glycan was capable of competing with MAb D10, but competition assays with D10 and F51 proved that these MAbs are specific for different epitopes on the glycan chain (unpublished data). MAbs D10 and MAb F51 should prove useful in future studies to determine the specific structure of this glycan chain. Amino acid composition analysis of the isolated glycan detected predominantly the four amino acids present in the Fap1 repeat sequences (E, S, V, I). Preferential detainment of these four amino acids suggests that the glycan moiety of Fap1 is linked within the repeat sequences of this protein. Compared with the unique sequences of Fap1, the repeat regions contain a majority of the putative glycosylation sites and all are putative O-linked sites via serine, the major amino acid present in the dipeptide repeat. Given the importance of this glycan in adhesion to SHA, an abundance of linkage sites within the Fap1 protein backbone is reasonable from a functional and evolutionary perspective. In eukaryotes, the core structure of N-linked glycopro-
154 A. E. Stephenson et al. teins contains N-acetylglucosamine, but lacks N-acetylgalactosamine. In contrast, O-linked glycoproteins demonstrate a greater variability and both N-acetylglucosamine and N-acetylgalactosamine may be present in the common core structures. Both monosaccharides are present in the isolated glycan from Fap1. These results indicate that Fap1 is an O-linked glycoprotein. This result is further supported by the negative result obtained with N-glycanase treatment, indicating that few or no N-linked glycans are present on Fap1, or that N-linked glycans are attached through a different type of linkage not recognized by this enzyme. Whereas O-glycanase treatment was inconclusive, this enzyme has a narrow substrate specificity that contrasts with the substantial variability of O-linkages that has been demonstrated in both eukaryotic and prokaryotic glycoproteins. Previously, it has been noted that the repetition of signature amino acid residues (E/V/I)S is highly conserved throughout the repeat region, suggesting that these regions are structurally important (Wu and Fives-Taylor, 1999). The data presented in this paper suggest that the role of the repetitive regions of Fap1 is to provide sites for glycan attachment, therefore playing an integral role in the ability of Fap1 to mediate adhesion of S. parasanguis FW213 to the tooth surface. Experimental procedures Bacterial strains and growth conditions Frozen stocks of S. parasanguis FW213 preserved in 5% dimethyl sulphoxide were streaked onto trypticase soy broth plates supplemented with 5% sheep’s blood (TSA II, BBL Stacker Plates, Becton Dickinson Microbiology Systems). Liquid cultures were inoculated from TSA II plates into Todd–Hewitt broth. All S. parasanguis plates and cultures were incubated aerobically in 5% CO2 at 37∞C. Desired cell densities were determined based on a growth curve generated using an absorbance of 470 nm. Escherichia coli XL1-Blue containing pCOMBIII plasmids were grown on Superbroth (SB) agar plates or liquid cultures containing selective antibiotics as described previously (Stephenson et al., 1998).
presence of antibody in culture supernatants was verified by ELISA. Mice were injected with the hybridoma cell lines and antibody-rich ascites was produced. MAbs were purified from ascites using standard gel filtration and ion-exchange chromatography techniques (Jenny et al., 1993). On average, 1 mg of MAb was recovered per 1 ml of mouse ascites. The two Fabs used in this study, Ph5 and Ph12, were produced in E. coli using phage display technology and were isolated using a cell-based panning method. A detailed description of this method, and the initial characterization of these Fabs, have been described previously (Stephenson et al., 1998). Soluble Fabs were immunoaffinity purified from culture supernatants using protein G cross-linked to rabbit antimouse IgG F(ab¢)2 (Jackson ImmunoResearch Laboratories). On average, 0.5 mg of Fab was recovered per 100 ml of bacterial culture.
Electron and fluorescence microscopy To prepare intact fimbriae for electron microscopy, middle-log-grown S. parasanguis was washed three times and resuspended in phosphate-buffered saline (PBS) at 4∞C at concentration of 5 ¥ 108 colony-forming units (cfu) ml–1. A drop of suspension was added to a copper grid coated with Formvar. To prepare free fimbriae for electron microscopy, a full loop of S. parasanguis bacteria from a blood agar plate was suspended in one drop of ammonium acetate (0.1 M, pH 7.0). A copper grid coated with Formvar was touched to the drop and quickly removed. The grid was washed by placing it on the surface of ammonium acetate until most of the cells were removed. Dried grids were subjected to platinum–carbon shadow casting and observed under a Philips 300 electron microscope at 60 kv. S. parasanguis cells were incubated for 20 min with saturating concentrations of either MAbF51 or MAbE42. Cells were isolated by centrifugation and washed by three rounds of resuspension in 1 ml of PBS. Cells were incubated for another 20 min with a saturating concentration of Oregon Green® 488 goat anti-mouse IgG (H + l) conjugate (Molecular Probes). Cells were examined on a Nikon Eclipse 400 upright microscope with a 100¥ Plan Fluor objective with a numerical aperture of 1.3. Oregon Green® tagged bacteria were visualized following illumination with a mercury vapour lamp filtered with Omega Optical (Brattleboro) XF100 filter cube that yields excitation and emission optima of 475 and 535 nm respectively. Photographs were taken with a RT Monochrome Spot digital camera (model 2.1.0) and processed with Spot software (version 3.0.4) from Diagnostic Instruments.
Mono-specific antibody production and purification Antibodies used in this study are summarized in Table 1. Monoclonal antibodies (MAbs) from hybridoma cell lines were produced from mice immunized with whole S. parasanguis FW213 cells (Elder and Fives-Taylor, 1986). MAbs F51, E42, A6, D10, and B20 were isolated in a screen for antibodies reactive with wild-type FW213 and non-reactive with non-adherent, non-fimbriated mutants (Elder and FivesTaylor, 1986). Fresh MAb stocks were produced and purified by Green Mountain Antibodies. Hybridoma cell lines for each of the five MAbs were cultured from frozen stocks and the
Cloning and expression of rFap1 (recombinant) and drFap1 (deleted recombinant) The pET system (Novagen) was utilized to clone and express the rFap1 and drFap1 proteins. An 825 bp DNA fragment coding for amino acids 69–342 of the N-terminal region of the mature protein was amplified by polymerase chain reaction (PCR) using FW213 genomic DNA as the template. NcoI and BamHI restriction enzyme sites were introduced via the PCR primers to allow subcloning as a NcoI–BamHI frag© 2002 Blackwell Science Ltd, Molecular Microbiology, 43, 147–157
The Fap1 fimbrial adhesin is a glycoprotein
155
0.2% maltose and 50 mg ml–1 kanamycin. Approximately 1 ¥ 109 plaque-forming units per ml of bacteriophage l CE6 and 10 mM of MgSO4 were added to induce expression in the culture. The infected bacterial cultures were incubated for 0–3.5 h at 30∞C to minimize the formation of inclusion bodies. Cells were harvested by centrifugation at 5000 g for 10 min and resuspended in PBS before boiling in SDS–PAGE sample buffer and loading onto gels.
Polyclonal antibody production
Fig. 8. A schematic diagram of rFap1 and drFap1 clones in E. coli. A. The rFap1 clone contains approximately 1 kb encoding the Nterminus of Fap1. The Repeat element I, the entire first dipeptide repeat, is encompassed within the rFap1 clone (109–166 aa) and is flanked on either side by unique sequence (69–108 aa and 167–342 aa). B. The drFap1 clone was created by deleting the entire repeat element I (109–166 aa) from rFap1 as described in Experimental procedures.
Two polyclonal antibodies were generated in this study to investigate the structure of the fap1 constructs that were cloned and expressed in E. coli: rFap1-specific and fimbriaspecific antibody. Purified rFap1 (100 mg) was used to generate a polyclonal antibody specific for this protein. rFap1, separated on SDS–PAGE, was transferred to a nitrocellulose membrane as determined by staining with Ponceau S (Sigma). The membrane was blocked in 5% non-fat dry milk for 1 h with shaking and incubation at room temperature. The membrane was washed for 3–5 min in 50 mM PBS (pH 7.4) and incubated in 200 ml of polyclonal antiserum prepared from whole FW213 bacteria for 3 h with gentle shaking. Unbound polyclonal antibody was removed and the membrane was washed extensively in PBS. Bound antibody was eluted with 0.2 M glycine/1 mM EGTA pH 2.7 and was neutralized to pH 7.0 with 2 M NaOH.
SDS–PAGE and Western blot analysis ment into the pET27b(+) vector. Ligation of the restricted fap1 PCR product into pET27b(+) resulted in the rFap1 fusion construct, which consisted of a herpes simplex virus (HSV) tag and 6 ¥ His tag at the C-terminus and a pelB leader sequence at the N-terminus in frame with the fap1 insert. Sequence analysis was used to confirm that the rfap1 sequence was identical to the genomic fap1 sequence (Fig. 8A). rFap1 was further modified to derive drFap1. Primers specific for the unique sequences flanking either side of repeat element I were designed and used in separate PCR reactions with the T7 promoter and T7 terminator primers to amplify the unique sequences of rFap1. These fragments were combined in an overlap PCR reaction (no external primers) to generate one contiguous, in-frame sequence consisting only of the unique sequences of rFap1 (Fig. 8B). The T7 promoter and T7 terminator primers were used to amplify large quantities of this fragment before restricting with NcoI and BamHI for reinsertion into pET27b(+). The resulting drFap1 construct was sequenced to confirm that the repeat region was deleted and the unique sequences were ligated in frame. The rFap1 and drFap1 constructs were initially transformed into electrocompetent STBL2 cells to isolate large quantities of the rFap1 and drFap1 plasmids. Use of the restrictive STBL2 host strain was necessary because of the toxicity of these proteins in E. coli. Ultimately, the rFap1 and drFap1 plasmids were electroporated into the expression host strain LE392. The newly transformed colonies were grown to A600 = 0.8 in Luria–Bertani (LB) media, supplemented with © 2002 Blackwell Science Ltd, Molecular Microbiology, 43, 147–157
SDS–PAGE was carried out according to standard protocols (Maniatis et al., 1989). Samples were boiled in sample buffer (0.0625 M Tris, pH 6.8, 2% SDS, 10% glycerol, 0.01% bromophenol blue) for 5–10 min before loading on acrylamide gels. All gels were run at 125 V until the bromophenol blue reached the bottom of the gel. Proteins were transferred from gels onto nitrocellulose according to standard protocols (Maniatis et al., 1989). Non-specific binding sites on nitrocellulose blots were blocked with a solution of PBS plus 0.5% non-fat dried milk (blocking buffer) for 1 h. Blots were then probed with various primary antibodies diluted in blocking buffer. Horseradish peroxidase-conjugated secondary antibody and chemiluminescent substrate (ECL Western Blot detection Reagents) were used to detect bands that reacted specifically with primary antibodies.
Adhesion blocking assay To test the ability of antibodies to block adhesion of S. parasanguis FW213 to SHA, adhesion blocking assays were performed (Fachon-Kalweit et al., 1985). Briefly, broth-grown log-phase FW213 cells were labelled with [3H]-thymidine and sonicated (15 s ¥ 4, at 80–85 W using an ultrasonic cuphorn) to disrupt bacterial chains. Labelled cells were incubated in triplicate with various concentrations of antibody in 50 mM NaPO4 buffer (pH 6.0). The cells were added to 7.5 ml scintillation vials containing SHA, and incubated for 1 h at 37∞C, with gentle shaking. The SHA was allowed to settle; the supernatant fluid was removed and transferred to fresh scintillation vials. The SHA was then washed four times with
156 A. E. Stephenson et al. adhesion buffer and the number of bound bacteria (SHA) and unbound bacteria (supernatant fluid) was determined by liquid scintillation counting. Two independent experiments were performed for each antibody.
Purification of Fap1 Fap1 protein was purified from culture supernatants of stationary phase cells as described previously (Wu et al., 1998). Briefly, NH4SO4 (40%) was used to precipitate proteins from the culture supernatant. After precipitation, Fap1 containing concentrates were dialysed extensively before their application to a F51-protein G immunoaffinity column. The column was prepared by loading F51 ascites onto a protein G column and cross-linking with dimethyl pimelimidate (DMP) (Schneider et al., 1982). On average, 15–30 mg was recovered per litre of culture supernatant.
Treatment of Fap1 with N- and O-glycanase Purified Fap1 protein was treated with N-acetyl-bglucosaminylasparagine amidase (N-glycanase, Oxford Glycosystems,) and endo-a-N-acetylgalactosamidase (Oglycanase, Oxford Glycosystems) according to the manufacturer’s instructions. Both enzymes required boiling of the substrate in a buffer supplemented with detergent before treatment. Additionally, O-glycanase required pretreatment of the substrate with neuraminidase to remove any sialic acid residues (GalNAc), thereby reducing the glycan to the Gal-b(1,3)GalNAc core disaccharide that is recognized by the O-glycanase. Digestion products were visualized on Coomassie-stained gels or by Western blot analysis with MAb F51 and E42.
Isolation of Fap1 glycopeptide Digestion of purified Fap1 with 0.5 mg ml–1 Pronase (Protease from Streptomyces griseus; Calbiochem, Cat#53702) was carried out at pH 7.5 at 40∞C for > 3 h. Reactions were stopped by diluting in PBS supplemented with EDTA to a final concentration of 50 mM. Reaction conditions were determined by Western blot analysis with MAb F51 and E42 to bias towards complete degradation of Fap1 (loss of antibody reactivity). Boiling was found to be essential to ensure complete denaturation of Fap1 before pronase treatment. As a further precaution, digestion products were applied to a Centricon 100 to separate pronase and unattached peptides from the glycopeptide. Isolated glycopeptide was purified from digestion products by immunoaffinity chromatography on a F51 column as described above.
Amino acid and monosaccharide composition analysis Isolated glycopeptide was subjected to 6N HCl hydrolysis at 110∞C for 24 h in sealed vials and applied to an Applied Biosystems 420/H PTC amino acid hydrolyser/analyser system at the University of Texas Medical Branch in the laboratory of Dr Alex Kurosky. The data collected were calculated by Microsoft EXCEL to obtain molarity values for individual amino acids.
Monosaccharide compositions of Fap1 and its glycopeptides were determined as trifluoroacetates of methyl glycosides by gas–liquid chromatography (Tomana et al., 1984; 1997). After liberating the sugars using HCl in methanol (Tomana et al., 1984), excess HCl was evaporated in vacuo, and samples were subjected to methanolysis with 0.5 M HCl in anhydrous methanol, acetylation with pyridine and acetic anhydride and de-O-acetylation (Tomana et al., 1984). Derivatization and gas–liquid chromatography using a Hewlett Packard model 5830 gas chromatograph (Hewlett Packard) equipped with a 25 m fused silica (0.22 mm i.d.) OV-1701 WCOT capillary column electron capture detector (Chrompack) and a Hewlett Packard model 3396 integrator (Hewlett Packard) allowed detection of picomolar quantities of monosaccharides (Tomana et al., 1984).
Competition with isolated glycopeptide and purified Fap1 Varying concentrations of isolated glycopeptide approximating equivalent concentrations of purified Fap1 protein were compared for their ability to compete for antibody binding to immobilized S. parasanguis cells expressing the Fap1 protein. A whole-cell ELISA (BactELISA) was used as the basis for performing these competitive assays. BactELISAs were carried out as described previously with the following modifications (Elder et al., 1982). FW213 cells were dried onto 96-well tissue culture plates and wells were blocked with PBS plus 0.5% gelatin (PBS-gelatin). Antibody stocks were diluted in PBS-gelatin containing either isolated glycopeptide or purified Fap1 protein before addition to wells. Antibody was detected with horseradish peroxidase-conjugated secondary antibody. Plates were washed three times with PBS before each new addition and were incubated for 1 h at 37∞C at each step. After incubating with secondary antibody, plates were washed with PBS plus 0.05% Tween 20. Antibody binding was quantified with an OPD (o-phenylenediamine) substrate that was detectable at 490 nm using a 96-well Microplate Autoreader EL311 (Biotek Instruments). Each concentration of isolated glycopeptide or purified Fap1 protein was tested in triplicate and the percentage of antibody binding in any given test well was calculated in relation to control wells representing 100% antibody binding (no competing glycan or Fap1 protein).
Acknowledgements We thank Robert Ullrich, Susan Wallace, Robert Melamede, Eunice Froeliger and Diane Meyer for their helpful comments and intellectual contribution to this research. This work was supported by Public Health Service grant R37DE11000 from the National Institutes of Health and grant DK57750.
References Barbas, C.F., Kang, A.S., Lerner, R.A., and Benkovic, S.J. (1991) Assembly of combinatorial antibody libraries on phage surfaces: the gene III site. Proc Natl Acad Sci USA 88: 7978–7982. © 2002 Blackwell Science Ltd, Molecular Microbiology, 43, 147–157
The Fap1 fimbrial adhesin is a glycoprotein Carlsson, J., Grahnen, H., Jonsson, G., and Wikner, S. (1970) Establishment of Streptococcus sanguis in the mouths of infants. Arch Oral Biol 15: 1143–1148. Castric, P. (1995) pilO, a gene required for glycosylation of Pseudomonas aeruginosa 1244 pilin. Microbiology 141: 1247–1254. Cisar, J.O., Takahashi, Y., Ruhl, S., Donkersloot, J.A., and Sandberg, A.L. (1997) Specific inhibitors of bacterial adhesion: observations from the study of Gram-positive bacteria that initiate biofilm formation on the tooth surface. Adv Dent Res 11: 168–175. Clamp, J.R., and Hough, L. (1965) The periodate oxidation of amino acids with reference to studies on glycoproteins. Biochem J 94: 17–24. Dobson, W.J., and McCurdy, H.D. (1979) The function of fimbriae in Myxococcus xanthus. I. Purification and properties of M. xanthus fimbriae. Can J Microbiol 25: 1152–1160. Elder, B.L., and Fives-Taylor, P. (1986) Characterization of monoclonal antibodies specific for adhesion: isolation of an adhesin of Streptococcus sanguis FW213. Infect Immun 54: 421–427. Elder, B.L., Boraker, D.K., and Fives-Taylor, P.M. (1982) Whole-bacterial cell enzyme-linked immunosorbent assay for Streptococcus sanguis fimbrial antigens. J Clin Microbiol 16: 141–144. Erickson, P.R., and Herzberg, M.C. (1993) Evidence for the covalent linkage of carbohydrate polymers to a glycoprotein from Streptococcus sanguis. J Biol Chem 268: 23780–23783. Fachon-Kalweit, S. (1985) Physiological characterization of fimbriae and other surface antigens from Streptococcus sanguis FW213. PhD thesis, University of Vermont, Burlington, VT. Fachon-Kalweit, S., Elder, B.L., and Fives-Taylor, P. (1985) Antibodies that bind to fimbriae block adhesion of Streptococcus sanguis to saliva-coated hydroxyapatite. Infect Immun 48: 617–624. Fives-Taylor, P.M. (1982) Isolation and characterization of a Streptococcus sanguis FW213 mutant nonadherent to saliva-coated hyroxyapatite beads. In Microbiology 1982. Schlessinger, D. (ed.). Washington, D.C.: Amer Soc Microbiol, pp. 206–209. Fives-Taylor, P.M., and Thompson, D.W. (1985) Surface properties of Streptococcus sanguis FW213 mutants nonadherent to saliva-coated hydroxyapatite. Infect Immun 47: 752–759. Hasty, D.L., Ofek, I., Courtney, H.S., and Doyle, R.J. (1992) Multiple adhesins of streptococci. Infect Immun 60: 2147– 2152. Jenkinson, H.F., and Lamont, R.J. (1997) Streptococcal adhesion and colonization. Crit Rev Oral Biol Med 8: 175–200. Jenny, R.J., Messier, T.L., Ouellette, L.A., and Church, W.R. (1993) Immunochemical techniques for studying coagulation proteins. Methods Enzymol 222: 400–416. Kohler, G., and Milstein, C. (1975) Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256: 495–497. Lechner, J., and Wieland, F. (1989) Structure and biosynthe-
© 2002 Blackwell Science Ltd, Molecular Microbiology, 43, 147–157
157
sis of prokaryotic glycoproteins. Annu Rev Biochem 58: 173–194. Maniatis, T., Sambrook, J., and Fritsch, E.F. (1989). Molecular Cloning: a Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. Marsh, P.D., and Bradshaw, D.J. (1995) Dental plaque as a biofilm. J Ind Microbiol 15: 169–175. Messner, P. (1997) Bacterial glycoproteins. Glycoconj J 14: 3–11. Moens, S., and Vanderleyden, J. (1997) Glycoproteins in prokaryotes. Arch Microbiol 168: 169–175. Robertson, E.R., and Kennedy, J.F. (1996) Glycoproteins: a consideration of the potential problems and their solutions with respect to purification and characterisation. Bioseparation 6: 1–15. Schachter, H., and Brockhausen, I. (1992) The biosynthesis of serine (threonine)-N-acetylgalactoseamine-linked carbohydrate moieties. In Glycoconjugates: Composition, Structure and Function. Allen, H.J., and Kisailus, E.C. (eds). New York: Marcell Dekker, pp. 263–332. Schneider, C., Newman, R.A., Sutherland, D.R., Asser, U., and Greaves, M.F. (1982) A one-step purification of membrane proteins using a high efficiency immunomatrix. J Biol Chem 257: 10766–10769. Stephenson, A.E., Fives-Taylor, P.M., and Melamede, R.J. (1998) Cell-based panning as a means to isolate phage display Fabs specific for a bacterial surface protein. Methods Cell Science 20: 241–249. Stimson, E., Virji, M., Makepeace, K., Dell, A., Morris, H.R., Payne, G., et al. (1995) Meningococcal pilin: a glycoprotein substituted with digalactosyl 2,4- diacetamido-2,4,6trideoxyhexose. Mol Microbiol 17: 1201–1214. Tomana, M., Prchal, J.T., Garner, L.C., Skalka, H.W., and Barker, S.A. (1984) Gas chromatographic analysis of lens monosaccharides. J Lab Clin Med 103: 137–142. Tomana, M., Matousovic, K., Julian, B.A., Radl, J., Konecny, K., and Mestecky, J. (1997) Galactose-deficient IgA1 in sera of IgA nephropathy patients is present in complexes with IgG. Kidney Int 52: 509–516. Tomoeda, M., and Inuzuka, M. (1975) Bacterial sex pili. Biophys Mol Biol 30: 23–56. Weerkamp, A.H., Van der Mei, H.C., and Liem, R.S.B. (1984) Adhesive cell-wall associated glycsoprotein of Streptococcus salivarius (K+) is a cell surface fibril. FEMS Microbiol Lett 23: 163–166. Whittaker, C.J., Klier, C.M., and Kolenbrander, P.E. (1996) Mechanisms of adhesion by oral bacteria. Annu Rev Microbiol 50: 513–552. Wieland, F. (1988) Structure and biosynthesis of prokaryotic glycoproteins. Biochimie 70: 1493–1504. Wu, H., and Fives-Taylor, P.M. (1999) Identification of dipeptide repeats and a cell wall sorting signal in the fimbriaeassociated adhesin, Fap1, of Streptococcus parasanguis. Mol Microbiol 34: 1070–1081. Wu, H., Mintz, K.P., Ladha, M., and Fives-Taylor, P.M. (1998) Isolation and characterization of Fap1, a fimbriaeassociated adhesin of Streptococcus parasanguis FW213. Mol Microbiol 28: 487–500.
