INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Mycoplasma bovis is the most important etiological agent of bovine mycoplasmosis in Europe and North America. It is responsible for outbreaks of therapy-resistant mastitis, mostly in larger dairy herds, and cases of pneumonia and arthritis in calves, as well as infections of the genital tract (16).

The antigen repertoire of this pathogen includes a family of variable surface lipoproteins (Vsps) which represents a set of immunodominant lipoproteins undergoing high-frequency phase and size variations, a phenomenon resulting in a multitude of phenotypes in a cultured mycoplasma population (1). While phase variation involves noncoordinated switching between on and off expression states of individual Vsps and is accompanied by DNA rearrangements (8), size variation leads to a set of differently sized proteins within a given Vsp as a consequence of spontaneous additions or deletions of repeating units within the vsp structural gene.

The biological function of Vsp antigens in M. bovis is not yet understood. Recent data indicated an escape mechanism based on modulation of the expression of certain variable proteins to evade opsonization of specific antibodies (7), which can be regarded as part of the strategy of the pathogen for subverting the host defense system in response to the presence of cognate antibodies. In a more functional aspect, Vsps as a whole or at least some members of the Vsp family are known to be involved in M. bovis cytoadhesion to host cells (6). Variable membrane proteins of other mycoplasma species, such as Vaa of M. hominis (27) and M. synoviae protein A or B (MSPA or MSPB) (12), were also shown to possess adhesive functions. Although considerably longer than those of M. bovis, repeating elements in the genome of M. genitalium are supposed to optimize cellular adhesion and to evade the host immune response (15).

Meanwhile, the vsp genomic locus of M. bovis has been cloned and characterized, and nucleotide sequences of 13 distinct vsp genes are available (8, 9). Examination of deduced amino acid sequences revealed an unusual structural motif. Most of the Vsp molecules are composed of repeating units extending from the N terminus to the C terminus of the protein chain. The majority of repetitive sequences are arranged as tandem domains consisting of units of 6 to 87 amino acids (aa). Since repeated units comprise the major part of most Vsp molecules, they may harbor active sites with certain biological functions, i.e., antigenic determinants, sites for cytoadhesion, or a different, as-yet-unknown function. Detailed characterization of Vsp functional domains appeared to be an essential prerequisite for understanding the molecular interactions between the pathogen and the host cell surface during pathogenesis.

In the present work, the repetitive domains of four selected Vsp antigens of M. bovis, VspA, VspB, VspE, and VspF, were examined for the presence of potential continuous epitopes with respect to immunogenic and/or adhesion sites. Therefore, sera of diseased animals infected with M. bovis were screened by an enzyme-linked immunosorbent assay (ELISA) for antibodies to repeating units. The ability of defined oligopeptides to reduce cytoadhesion was examined with a competitive adherence assay. To characterize the location of functional domains at the amino acid level, mapping of immunodominant epitopes and adherence sites was conducted with overlapping oligopeptides covalently bound to a membrane.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Animal sera. Sera from six dairy cows (cows 1, 4, 7, 14, 22, and 23) with mastitis due to natural infection with M. bovis were investigated. M. bovis in milk samples from all animals was verified by culturing. No other bacterial agent was detected. Serum from an M. bovis-free healthy animal from the same herd was included as a control.

Computer analysis of protein structures. Amino acid sequences of variable surface proteins were deduced from nucleotide sequences of the following genes: vspA (GenBank accession no. L81118), vspB (AF162138), vspE (AF162139), and vspF (AF162140) (8, 9). Hydrophobicity plots, secondary structure analysis, and calculation of total amino acid composition were carried out with the following programs: (i) MacVector version 4.1 (IBI Kodak, New Haven, Conn.) and (ii) Winpep 1.0, developed by Lars Hennig, University of Freiburg, Freiburg, Germany, and available from http://www.biologie.uni-freiburg.de/data/schaefer/winpep1.html.

Synthetic oligopeptides. Oligopeptides were synthesized and purified by reversed-phase (RP) high-performance liquid chromatography (HPLC) at MWG-Biotech (Ebersberg, Germany). Peptides used as capture antigens in ELISAs were linked to an additional terminal cysteine residue to allow covalent coupling to ovalbumin.

Specific antibody ELISA. Eight different oligopeptides based on the repeating units R_A1, R_A2, R_A3, R_A4.1, R_A4.2, R_E1, and R_F2 and a 14-aa peptide of R_F1 (positions 50 to 63) were covalently linked to ovalbumin as a carrier protein with the Imject Activated Immunogen Conjugation Kit (Pierce, Rockford, Ill.).

Competitive adherence assay. The procedure used for the competitive adherence assay was described previously (20). Briefly, embryonic bovine lung (EBL) cells were cultivated in 24-well tissue culture plates to form confluent monolayers. To each well, 10⁸ CFU of ³H-labeled M. bovis PG45 was added together with 0.5 µmol of the oligopeptide. To allow adhesive interactions, the mixture was incubated at 37°C for 30 min with gentle shaking. After three washing steps and solubilization of attached cells, the adherence rates were measured by liquid scintillation counting. Each oligopeptide was tested in at least three separate trials on different plates. Control trials of strain PG45 without the addition of peptides were run in parallel (in quadruplicate) on each plate, and this adherence was taken as 100%. The relative standard deviation of the method is 13.3%.

MAbs. Details on the preparation and characteristics of Vsp-specific monoclonal antibodies (MAbs) 1E5, 4D7, and 2A8 are given elsewhere (2). Immunoglobulin fractions were obtained by affinity chromatography (Sepharose-protein A column; Amersham Pharmacia, Freiburg, Germany) of culture supernatants from a hollow-fiber fermentation system (Tecnomouse; Integra Biosciences, Fernwald, Germany).

Epitope analysis with synthetic peptides spotted on cellulose membranes (Pepscan method). For peptide spot synthesis as described by Frank (5), the Auto Spot Robot ASP 222 (ABIMED Analysentechnik, Langenfeld, Germany) was used. Overlapping oligopeptides (5 to 10 aa) derived from amino acid sequences of repeating units in VspA, VspB, VspE, and VspF were synthesized with 9-fluorenylmethoxy carbonyl-protected amino acids spotted on cellulose membranes derivatized with a polyethylene glycol spacer (300 spots on a membrane measuring 8 by 12 cm).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Structural features of four Vsps of M. bovis. The primary structure of VspA, VspB, VspE, and VspF, as deduced from nucleotide sequences of their genes (8, 9), is schematically depicted in Fig. 1. Examination of hydrophobicity profiles revealed that mature VspA, VspB, and VspE are entirely hydrophilic, whereas the VspF molecule harbors both hydrophilic and hydrophobic regions. As the main structural feature, these antigens possess several distinct domains of reiterated sequences which extend from the N-terminal end to the C-terminal end and create a periodic polypeptide chain. In VspA, for instance, where the repetitive portion represents about 80% of the entire protein chain, four distinct internal regions of contiguous tandemly repeating units were identified and designated R_A1, R_A2, R_A3, and R_A4. The latter can be further divided into subunits R_A4.1 and R_A4.2, which differ in three of eight amino acid residues. While VspB has only two distinct repetitive regions, R_B1 and R_B2, which are highly homologous to R_A3 and R_A4, respectively, the structures of repeated peptide units in VspE and VspF are markedly different. Thus, in VspE there is only one region of short tandem repeats, R_E1, and VspF is unique among these four Vsps for its relatively long repeats, R_F1 (84 aa) and R_F2 (10 aa). (See Fig. 4 and Table 3 for amino acid sequences of repetitive units for R_F1 and all others.) Another structural characteristic is the high content of the amino acid proline, particularly in VspA (10%), VspB (8.2%), and VspE (13.7%). In addition, glycine and glutamine are abundant in VspA and VspB, while glutamic acid and lysine are highly represented in VspE and VspF. A summary of characteristic features of the antigens is given in Table 1. A comparison of both nucleotide and amino acid sequences of Vsps with those contained in GenBank databases failed to identify homologs in other organisms.

View larger version (37K): [in this window] [in a new window]   FIG. 1.   Schematic representation of gene sequences coding for four Vsps of M. bovis: VspA, VspB, VspE, and VspF. Solid blocks at the N terminus depict promoter (PX) or putative signal peptide (black) sequences. Blocks with different types of hatching or shading and designated RAn, RBn, REn, and RFn represent repeating units. Subunits within the RA4 domain (4.1 and 4.2) or within the RB2 domain (2.1 and 2.2) are indicated. Numbers at the right end of each Vsp represent the length of the polypeptide chain.

Detection of antibodies to Vsp repeating units in sera of infected animals. Bovine sera previously found positive for M. bovis by both a polyclonal antibody ELISA and immunoblotting were analyzed together with control sera by means of a specific antibody ELISA designed to capture cognate antibodies. The assay was conducted with microtiter plates coated with eight different ovalbumin-coupled synthetic oligopeptides representing distinct repetitive amino acid sequences of VspA, VspB, VspE, and VspF. Among a group of six naturally infected cows with clinical symptoms of mastitis, at least 50% of the animals tested positive (i.e., more than twice the value for the control) for antibodies to repetitive peptide sequences R_A1, R_A2, R_A4.1, R_E1, and R_F1 (Table 2). The levels of antibodies to peptides R_A3, R_B1, and R_F2 were only slightly increased compared to the levels in the control serum. Because of the high background, the levels of antibodies to peptides R_A4.2 and R_B2.2 could not be defined.

View larger version (76K): [in this window] [in a new window]   FIG. 2.   Examination of sera from a calf experimentally infected with M. bovis by a specific antibody ELISA. Sera were collected on days 0, 6, 15, 21, and 28 p.i. Each serum was examined for antibodies to seven different synthetic oligopeptides representing distinct repetitive amino acid sequences of VspA, VspB, VspE, or VspF. OD, optical density.

Inhibition of M. bovis adhesion by synthetic oligopeptides and MAbs. The possible involvement of Vsp repetitive domains in cytoadhesion was examined through the ability of Vsp oligopeptides to inhibit M. bovis attachment to mammalian host cells. This was done by a competitive adherence assay with tissue culture plates and several synthetic peptides identical to the complete repeat sequences (Table 3). Subsequently, tetrapeptides and pentapeptides representing partial sequences of repeats were used to refine the search for epitopes. A 20% reduction of M. bovis adherence rates due to blocking of host cell receptor sites under standardized conditions was considered a minimal criterion of inhibitory activity. The results presented in Table 3 show that the cytoadhesion of strain PG45 was partially inhibited by 10 oligopeptides derived from repeats R_A2, R_A4.1, R_E1, R_F1, and R_F2. On the other hand, 14 other peptides (tetramers to octamers) derived from repeats R_A1, R_A2, R_A3, R_B1, R_A4.2, R_B2.2, R_E1, R_F1, and R_F2 failed to reduce adherence.

Epitope mapping with membrane-bound overlapping oligopeptides (Pepscan method). The precise location of antigenic determinants on repetitive domains of Vsp molecules was determined with a set of overlapping oligopeptides (trimers to decamers) spanning distinct Vsp repetitive regions and covalently attached to cellulose membranes. These membranes were incubated with the same set of animal sera as those used in the ELISA. Characteristic reactive patterns led to the identification of putative epitopes present within the periodic structures of Vsps (Fig. 3, lanes i).

View larger version (37K): [in this window] [in a new window]   FIG. 3.   Identification of immunogenic epitopes and adherence sites in repeating units of VspA, VspB, VspE, and VspF. Cellulose membranes loaded with a series of covalently attached overlapping oligopeptides were incubated with 11 different sera from infected cattle or 35S-labeled EBL cells. Reactivities of peptide spots with serum from an experimentally infected animal (28 days p.i.) and adherent tissue culture cells are shown in lanes i (immunogenicity) and a (adhesion), respectively. The amino acid sequence corresponding to the respective spot is given in the same line. Presumed immunogenic epitopes are indicated by shaded areas, and adherence sites are boxed. Amino acid sequences of complete repeating units, printed in bold type, are given in each panel for comparison.

View larger version (18K): [in this window] [in a new window]   FIG. 4.   Localization of epitopes of immunogenicity and adhesion sites in repeating unit RF1 of VspF, as determined by epitope mapping. Methodology, serum samples, and adherent cells were the same as those described in the legend to Fig. 3. Potential immunogenic epitopes (shaded areas) are designated i1 to i4, and adherence sites (boxed areas) are designated a1 to a6. (Epitopes i3 and i4 are probably nonspecific, as they were also recognized by preimmune control sera.)

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The ability of microorganisms to alter the antigenic features of their cell surface was suggested to be part of a strategy of adaptive evolution (11), especially if surface components interacting with a changeable environment are involved. Other potential biological functions of variable lipoproteins include substrate binding and participation in protein secretion and signal transduction (18, 22, 28).

For a number of bacterial pathogens, e.g., group B streptococci (10), staphylococci (6), Haemophilus influenzae (23), and Neisseria spp. (14), contiguous repetitive DNA sequences encoding membrane-bound proteins were described, and the corresponding tandemly repeated peptide units were suggested to be involved in eliciting a host immune response and/or to participate in cell attachment and binding processes (4, 10, 24).

While the data of this study provide the first experimental evidence of the involvement of repetitive domains of Vsps in M. bovis interactions with host cells, they also represent further confirmation of the immunogenic and adhesive functions of Vsps.

Detection by an ELISA of specific antibodies to repeating units R_A1, R_A2, R_A4.1, R_B2.1, R_E1, and R_F1 in the sera of at least 50% of the mastitic cows (Table 2) was further substantiated by the identification of immunogenic epitopes in these repeats by the Pepscan method (Fig. 3). For R_F2, for which an active site was also seen on the membrane, ELISA antibody titers appeared to be elevated, but not to the same extent as with the other peptides mentioned. While R_A3 and R_B1 seemed to be nonimmunogenic, the situation with R_A4.2 and R_B2.2 remains unclear from the present data. Very high ELISA readings, as shown in Table 2, as well as an occasional background of immunostained spots on the membrane (data not shown) could be a consequence of nonspecific binding caused by polar residues of R_A4.2 and R_B2.2 (GAGTNSQQ).

The immune response in the experimentally infected pneumonic calf showed a general rise in anti-Vsp titers but appeared to be less intense than that in naturally infected cows (Fig. 2). The M. bovis strain used for infection, 981/84, was previously shown to coexpress in vitro at least two Vsps, VspA and VspB (2).

Considering the locations and sizes of putative adherence epitopes, it is not surprising that the results of epitope mapping could not be verified in all cases in the tissue culture plate adherence assay (Table 3). Obviously, there are essential differences in the spatial arrangement of ligand-receptor reactions between both approaches. The membrane assay involves the specific attachment of tissue culture cells to immobilized but freely accessible peptide chains, so that, in principle, relatively few binding events can generate a positive signal. This characteristic renders the membrane assay more sensitive than the tissue culture plate assay, in which comparatively more recognition events may be necessary to produce measurable adherence inhibition, since a larger number of peptide molecules have to bind to one EBL cell to attain blocking of all receptor sites. Another limitation of the latter includes the difficulties in synthesizing certain tetrapeptides and pentapeptides of sufficient purity.

Important evidence in support of the in vivo function of adherence epitopes identified by the Pepscan method was provided by the demonstration of adherence-inhibiting activity of three MAbs, all of which recognized adherence-related epitopes (Table 4).

Combining adherence data from the present study with previous results from our group, which included (i) selective binding of mammalian host cells to Western-blotted Vsps (21), (ii) enhancement of M. bovis cytoadhesion by purified native Vsps, and (iii) retention of purified Vsp antigens on host cell layers in a cell binding assay (K. Sachse, unpublished results), provides ample evidence demonstrating the involvement of Vsp antigens in cytoadhesion. This conclusion is also supported by reports of proline-rich sequences and proline-rich repeats being involved in a variety of attachment and binding processes (25, 26). Indeed, all epitopes identified in repeats R_A1, R_A2, R_E1, and R_F2 contain one proline residue, and the GTK motif in R_A4.1 is adjacent to a proline.

Concerning positions of epitopes relative to the repeating units, two different patterns were observed. In the case of R_A1, R_A2, and R_E1, epitope sequences covered 5 to 7 aa of two contiguous repeats, whereas those from R_A4.1, R_B2.1, R_F1, and R_F2 (3 to 7 aa) were found to remain within the boundaries of a single unit.

A comparison of positions and lengths of putative antigenic determinants with those of adhesive sites in the short repeating units revealed identity in most cases, with a minor shift only in R_A2 (KPEENK versus NKKP; Fig. 3). It is not certain whether this finding really indicates differences in locations between immunogenic and adherence epitopes or whether it is due to inherent limitations of the Pepscan method. Conversely, mapping data from the considerably longer R_F1 unit indicate that the locations of immunogenic functional sites differed from those of adherence epitopes in three of four instances. Divergent positions between immunodominant and adherence epitopes were also observed for the P1 protein of M. pneumoniae and the MgPa protein of M. genitalium (13, 17). Interestingly, adherence epitopes were found to be distributed over the entire length of the R_F1 repeat, while antigenic determinants appeared to be concentrated in the central part of the C-terminal domain. In addition, the positions of active sites roughly coincided with major or minor peaks in the hydrophobicity plot of this repeating unit (data not shown).

While the present results from epitope mapping will contribute to a better understanding of functional structures of Vsp antigens, it must be emphasized that the methodology used in this study can determine only continuous epitopes. Since the conformation prevailing in the native protein is not retained in linear oligopeptide fragments, it is likely that most of the active sites identified represent only portions of more complex discontinuous epitopes. Consequently, conformational epitopes that are characteristic of the antigens in vivo may either appear as several continuous sites or be left unidentified (3).

Nevertheless, the presence of epitopes in tandemly arranged repetitive peptide units may help to explain the high immunogenicity of several members of the Vsp family, which places VspA, VspB, and VspC among the most prominent antigens of M. bovis (2, 19). In this context, size variation of individual Vsps would result in the generation of more epitopes on the M. bovis surface or a reduction in their numbers, which could be regarded as a way for the pathogen to modulate its immunogenic potential and/or its capability to attach to the host cell surface. The multiplicity of repetitive peptide units may enable the surface protein to bind several ligands, either of the same sort or of a different sort, thus allowing a wide range of avidity and specificity of binding. It is conceivable that favorable or hostile external factors, e.g., from the host immune response or environmental conditions, may cause mycoplasmas to increase or decrease the intensity of their interaction with the host.

Although the findings of the present study on epitope localization should provide important clues for vaccine development, the straightforward approach, i.e., using purified Vsps or synthetic peptides for immunization, may not be effective, as experience with the P1 protein of M. pneumoniae has shown (17). In the case of M. bovis, ways have to be found to deal with its ability to evade the host immune response by modulating the expression of certain Vsps in the presence of cognate antibodies (7). Instead, DNA vaccination with vectors encoding a number of epitopes from distinct variable and nonvariable antigens may be a promising alternative.