INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Streptococcus pneumoniae is a significant pathogen, accounting for a large fraction of pneumonia, sepsis, and meningitis (12). Antibiotic treatment of pneumococcal infections has become less effective due to a recent dramatic increase in the prevalence of antibiotic-resistant strains of S. pneumoniae in many parts of the world (4). While pneumococcal infections could be prevented with pneumococcal vaccines, the currently available 23-valent pneumococcal vaccine, which contains the capsular polysaccharide (PS) of 23 commonly found serotypes of S. pneumoniae, is not protective for young children (24) and may be effective only for subpopulations of older adults (28), the two populations most susceptible to S. pneumoniae infections. Thus, there is a great need for an effective pneumococcal vaccine; the main approach used for improving the pneumococcal vaccine is to conjugate the PS to a carrier protein as done for Haemophilus influenzae type b vaccines (29, 35).

Although this PS-protein conjugate vaccine may elicit antibody responses in young children, the conjugation process can alter the antigenic epitopes, and the conjugate vaccine may elicit antibodies with altered repertoire or induce undesirable antibodies. It has been suggested that some PS (particularly those that are linear and contain phosphodiester bonds) may be mimotopes of DNA, as some anti-DNA antibodies bind the bacterial PS from Klebsiella species (22) and Neisseria meningitidis group B (16). The currently available 23-valent pneumococcal vaccine has been known to increase the antibody displaying the 8.12 idiotope (14), which is expressed on nephropathic anti-double-stranded-DNA (dsDNA) antibodies (20). The capsular PS from S. pneumoniae serotype 6B is, like DNA, a linear polymer with phosphodiester bonds (8). Although 6B capsular PS is poorly immunogenic, all new vaccines will contain 6B capsular PS (35) because infection by S. pneumoniae serotype 6B is common. Our preliminary study showed that some anti-6B antibodies frequently express  light chains (25), which suggests that anti-6B antibodies may, like anti-dsDNA antibodies, display the 8.12 idiotype that is expressed on  light chain.

Expression of human V-C genes has not been systematically studied so far, although these genes have several unique features. Unlike the case for other constant-region genes, there are seven C genes, each associated with one unique J gene (23). In some individuals, the region between C2 and C3 is amplified and may contain up to 10 C genes (40). Furthermore, unlike human V genes and mouse V genes, human V genes are grouped according to gene family. For instance, V2 gene family members are located together near the J1 gene whereas all members of the V4 gene family are found together away from the J1 gene in the 5' direction (13). Thus, there may be limits in extrapolating our observations on the expression of murine V genes or of human V genes to those of human V genes. Since antibodies to 6B PS are often ⁺ (25), studies of human antibodies to 6B PS would allow studies of V expression in young children and adults. We therefore investigated the V-region structure of antibodies to the capsular PS of serotype 6B at the molecular level, using 23-valent PS vaccine.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Antipneumococcal antisera. Healthy adult volunteers were immunized with the 23-valent PS vaccine from Lederle Laboratories (Pearl River, N.Y.) or from Pasteur Merieux (Lyon, France). Serum samples were collected from the volunteers 1 month after vaccination. A serum pool (89-SF) was obtained from C. Frasch (Food and Drug Administration, Bethesda, Md.) and used as the standard in all assays. The standard contains 24.3 µg of total (27), 17.6 µg of K⁺, and 6.7 µg of ⁺ anti-6B PS antibody per ml (25).

ELISA. The amount of anti-6B PS antibody was determined by sandwich-type enzyme-linked immunosorbent assays (ELISAs). Briefly, the wells of Immulon II plates (Dynatech, Chantilly, Va.) were coated at 37°C with 6B PS (10 µg/ml; American Type Culture Collection, Rockville, Md.) overnight in phosphate-buffered saline, which was prepared fresh, using water from a Milli-Q UF water purification system (Millipore, Bedford, Mass.), to minimize the background signal. The plates were washed and blocked with phosphate-buffered saline containing 1% nonfat milk (Carnation, Los Angeles, Calif.). The human serum pool (89-SF) was used as a standard. Samples were preabsorbed with 3 µg of C-PS (Statens Seruminstitut, Copenhagen, Denmark) per 20 µg of serum in a total volume of 1 ml of diluent for 30 min at room temperature. The serum samples were then added to wells, serially diluted, and incubated for 3 to 5 h at room temperature. Plates were then washed, and alkaline phosphatase-conjugated goat antibody against total human immunoglobulin (Ig), kappa or lambda chain, was added. To measure 8.12⁺ anti-6B antibody, 8.12 monoclonal antibody was added to wells, and goat anti-mouse Ig antibody labeled with alkaline phosphatase was added later. 8.12 recognizes human antibodies expressing the V2 family gene product. The amount of enzyme immobilized to the well was determined with para-nitrophenyl phosphate substrate (Sigma) in diethanolamine buffer. Optical density at 405 nm was read with a microplate reader (Cambridge Technology, Watertown, Mass.). The amount of antibody in the sample was determined by comparing the optical density of the samples to the standard, using piecewise linear interpolation.

Purification of anti-6B antibody from immune serum and determination of its amino acid sequence. Purification was performed as described for antibodies to H. influenzae type b PS (34), with slight modifications as described below. An Ig fraction was separated from the immune serum (50 to 200 ml) by precipitation in 50% saturated ammonium sulfate. The Ig fraction was passed over a Sepharose column conjugated with 6B PS. The 6B PS was conjugated to Sepharose following CNBr treatment. The 6B-Sepharose column saturated with anti-6B antibody was washed first with 100 mM NaCl containing 50 mM Tris (pH 7.4), next with 100 mM borate buffer (pH 8.4) containing 5 mM phosphocholine (PC) and 0.01% Tween 20, and last with 150 mM NaCl. Preliminary studies showed that 5 mM PC eluted most anti-C-PS antibodies binding the haptenic determinant PC. Anti-6B antibody (50 to 200 µg) was eluted from the column with 3.5 M MgCl₂ (pH 3.5), and the elution fraction containing purified protein was neutralized with 1 M Tris and dialyzed against normal saline buffered with 50 mM Tris (pH 7.5). To obtain the IgG fraction, the antibody was passed over Sepharose columns coupled with HB57 and HA1, monoclonal antibodies specific for IgM and IgA, respectively. The  and  fractions of the IgG antibody were obtained by depleting  and  antibodies with the use of anti- (monoclonal antibody HK-2) and anti- (monoclonal antibody HL-1) antibody columns, respectively. The purified antibody was assessed for antigen specificity for 6B; purity and clonality were tested by ELISA and isoelectric focusing methods as described elsewhere (34).

Amino acid sequencing. Purified antibodies were separated into the light and heavy chains in a sodium dodecyl sulfate-polyacrylamide gel, and the separated chains were electrophoretically transferred to a polyvinylidene difluoride (PVDF) membrane. The amino acid sequence was obtained from the peptide-containing membrane with an Applied Biosystems Inc. (Foster City, Calif.) model 470A gas-phase sequencer by Midwest Analytical, Inc. (St. Louis, Mo.). To obtain the sequences in the CDR2 and CDR3 regions, the PVDF membrane containing the peptide chain was treated with CNBr or BNPS-Skatole to cleave the chain at either methionine or tryptophan, respectively, before the amino acid sequence was obtained (33). To obtain the sequence in the CDR2 region of VH, the VH blot was treated with 0.02 mg of o-phthalaldehyde (OPA) per 1 ml of butylchloride at the appropriate sequencing cycle. OPA treatment makes all the N-terminal amino acids except proline resist the peptide degradation chemistry and reduces the sequencing noise, thereby extending the sequencing depth (3).

Nucleotide sequencing of VH and VL region cDNAs from Dob1 hybridoma. Dob1 hybridoma secreting human IgG2 anti-6B antibody was produced by fusing K6H6/B5 cells (5) with human peripheral blood mononuclear cells, which were obtained from an individual immunized with the 23-valent PS vaccine (unpublished data). Dob1 mRNA, extracted by the method of Chomczynski and Sacchi (7), was reverse transcribed with oligo(dT) and a reverse transcriptase. cDNAs of VH and VL regions were obtained from the Dob1 cDNA by 35 cycles of PCR using primers GAGGTGCAGCTG(G/T30)TGGAGTCT and GAC(C/G)GATGGGCCCTTGGTGGA for VH and CAGTCTGCGCTGACTCA(A/G)CCG(G/C)CCTCT and AGAGGA(G/C25)GG(C/T30)GGGAACAGAGTGAC for VL (18). The PCR cycle was 3 min at 72°C for extension, 1.5 min at 95°C for melting, and 2 min for annealing. To minimize spurious PCR products, the annealing temperature of 65°C for the first cycle was decreased by 1°C per cycle for the first 15 cycles and maintained at 50°C for the remaining cycles. After confirmation of the presence of a PCR product of appropriate length by electrophoresis, the PCR product was cloned in Escherichia coli with a TA cloning kit (Invitrogen, San Diego, Calif.) by ligation to the pCR2.1 vector and transformation of E. coli with that vector. Plasmid DNA was isolated from five ampicillin-resistant, galactosidase-deficient E. coli colonies by using a commercial kit (Wizard Plus Miniprep; Promega, Madison, Wis.), and the DNA insert containing the VH (or VL) region was recovered from the plasmid. The forward and reverse sequences of the DNA insert were determined by the Sanger dideoxy method, using fluorescent terminators (DNA sequencing kit from Perkin-Elmer, Foster City, Calif.) and by measuring electrophoretically separated fluorescent bands with a model 377 DNA sequencer from Applied Biosystems. The forward and reverse sequences from several bacterial colonies were assembled into the complete VH and VL sequences. DNA sequences were compared by the clustered method, using a commercial program from DNAstar Inc. (Madison, Wis.).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Idiotope 8.12 is commonly expressed among anti-6B PS antibodies. Pneumococcal vaccination has been shown to increase the level of antibodies displaying the 8.12 idiotype, which is now known to be expressed by V2 family gene products. Previously, we reported frequent expression of the  light chain by anti-6B PS antibodies, especially among the high responders (Fig. 1A) (25). Consequently, we tested whether the ⁺ anti-6B antibodies express the 8.12 idiotope by measuring the concentrations of anti-6B PS antibodies expressing the 8.12 idiotope in the serum samples from 25 adults and by comparing them with the anti-6B antibodies expressing the  light chain (Fig. 1B). Almost all of the ⁺ anti-6B PS antibodies expressed the 8.12 idiotope, with a strong correlation between the two parameters (r = 0.81). In only 3 of 25 samples was the 8.12 idiotope expressed in less than 10% of the ⁺ anti-6B PS antibodies. Thus, VL genes belonging to the V2 family must encode most of ⁺ anti-6B PS antibodies.

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Relationship between the amount of anti-6B antibodies (Ab) expressing the 8.12 idiotype and the amount of anti-6B antibodies expressing  light chain for volunteers vaccinated with a PS vaccine (B), compared with the relationship between total anti-6B and anti-6B antibodies expressing either  or  chain (A). Donor P26, who was found to express antibody derived from a V10 family gene, is indicated with an arrow. Panel A was reproduced with permission of the publisher (J. Infect. Dis. 174:75-82, 1996).

Products of several similar V2 genes and the C1 gene are used for anti-6B PS antibodies. Initial attempts to determine the amino acid sequences of the -chain antibodies revealed a blocked N terminus. Since the  light chain often has methionine at position 49, we treated the PVDF blot (with  chain) with CNBr and determined the sequence of 14 amino acids starting at position 50 (Fig. 2). The amino acid sequences from the P9C and P703 fractions were identical and matched exactly to that of the 2c V gene. The sequence of P704C was identical to that of the 2e gene, which differs from the 2c gene only by one amino acid (E versus D at position 52) in this part of the V region. However, the sequences of these three samples at this region (positions 49 to 64) differed from other V gene sequences by more than three amino acids. We therefore conclude that P704C, P9C, and P703C are derived from either the 2c or 2e V2 gene and not from other V genes. In contrast, the sequence from sample P18C matched exactly that of the 2a2 V gene and differed from other V gene sequences by more than three amino acids. Thus, the P18C V region is likely the product of the 2a2 V gene. The amino acid sequences from positions 38 to 49 obtained following Skatole digestion of the light chain matched exactly those of the 3 V genes listed above and supported our conclusion.

View larger version (20K): [in this window] [in a new window]   FIG. 2.   Amino acid sequences of the V region of the light chain of anti-6B antibodies expressing the  light chain. Dayhoff amino acid notation and the residue numbering system of Kabat et al. (17) are used. X indicates no residue identified; the X at position 22 of P26 is likely an invariant cysteine not identifiable by our sequencing method. The lowercase letter denotes the recovery of a smaller than expected amount of amino acid during the Edman degradation cycle; the solid line denotes identity to the reference sequence at the top. Tryptophan at positions 38, 148, and 186 are inferred by the cleavage method.

The V10 gene product can be used for anti-6B antibodies. Several serum samples (e.g., P26C [indicated with an arrow in Fig. 1B]) had high levels of ⁺ anti-6B antibodies without expressing the 8.12 idiotope in proportion. This finding suggested that these samples may be the product of a V gene other than V2. Indeed, the  chain of anti-6B antibodies isolated from serum sample P26C had an unblocked N terminus and the N-terminal sequence closely matched that of V gene 10a, which belongs to the recently defined V10 family (Fig. 2) (42). Compared with the sequence of V gene 10a, our sequence lacked the first two amino acids at the N terminus and had one potential mismatch (from S to A) at the 24th cycle.

Many V genes are used to form anti-6B antibodies. To obtain a more complete picture of the V region repertoire of anti-6B antibodies, we determined the sequences of anti-6B antibodies expressing the kappa chain which were prepared by affinity chromatography (Fig. 3). In two cases, we found that one antibody clone dominated the kappa response and were able to separate the clone and obtain the unique amino acid sequence of the VL region. In these two cases, the sequences of the N-terminal 25 bases of the two light-chain preparations match perfectly the A27 sequence and differ from sequences of other V genes by at least two amino acids (Fig. 3). In most cases, ⁺ anti-6B antibodies from an individual often have amino acid sequences of the genes of several different V gene families (data not shown), and we could not identify the dominant  clones. While some of the V sequences that we observed may have been those of contaminants, ⁺ anti-6B antibodies in a given individual appear to be derived from several V genes belonging to different families.

View larger version (12K): [in this window] [in a new window]   FIG. 3.   Amino acid sequences of the V region of the light chain of anti-6B antibodies expressing the  light chain. Symbols are as in Fig. 2.

Characterization of the VH region of anti-6B antibodies. The heavy-chain sequences were mostly of the VH3 subgroup (Fig. 4). However, the N-terminus of the heavy chain of the antibody P18C was blocked. The PVDF membrane bearing the P18C heavy chain was digested with Skatole and then treated with OPA at the fifth cycle of Edman degradation to suppress the amino acids arising from the C-region fragments. The sequence of 18 amino acids matched perfectly to that of the FR2-CDR2 region of VH1 gene 1-03 (Fig. 4) but differed from that of other VH1 genes by at least three residues. We confirmed our sequence by cleaving the heavy chain at methionine (position 34) with CNBr and treating it with OPA at the seventh cycle. Taken together, the results suggest that the VH gene 1-03 is the candidate gene for P18C VH.

View larger version (19K): [in this window] [in a new window]   FIG. 4.   Amino acid sequences of the V region of the heavy chain of anti-6B antibodies. Symbols are as in Fig. 2.

The human anti-6B hybridoma Dob1 expresses the gene products of the V2 and VH3 gene families. To confirm the findings obtained with amino acid sequences of anti-6B antibodies, we produced one human-mouse hybridoma (Dob1) secreting anti-6B antibody and determined the DNA sequences of its VH and VL regions. As shown in Fig. 5, the VH and VL regions are clearly derived from genes in the VH3 and V2 families. The VH region nucleotide sequence of Dob1 displayed an 88% match with VH gene 3-15 and 79% match with 3-72, whereas the match with all other VH3 genes ranged from 73 to 68% by the clustered method. The Dob1 JH region sequence best (92%) matched the JH4 sequence, followed by less than 83% match with all other JH sequences. Taken together, the results suggest that the candidate VH and JH genes for Dob1 are 3-15 and JH4. The D-JH splicing occurred 5' to TTGA, which is a common splice site (43). The D-region sequence of Dob1 lacked obvious resemblance to any of the 27 D genes in the human genome (9). Analysis of the VL region of Dob1 showed that it best matched 2a2 (91%) and 2b2 (91%) sequences but was still similar to the 2e gene sequence (88%). Other V gene sequences displayed less than 74% similarity with the Dob1 sequence. At the J region, the Dob1 sequence matched best the J1 sequence. There was no evidence for the extra nucleotides inserted at the VL-JL junction. There are several nucleotide sequence changes consistent with somatic mutations. For instance, position 49 is leucine in Dob1 but methionine in all V2 genes. Although this particular change was not observed with purified antibody proteins, the Dob1 sequence supports our conclusion that the most common form of anti-6B antibodies is formed with a VH3 gene and a V2 gene.

View larger version (73K): [in this window] [in a new window]   FIG. 5.   cDNA sequences of VH (A) and VL (B) regions of Dob1. The first lines denote the VH and VL domains as assigned by Kabat et al. (17). The third lines denote the reference DNA sequences for the heavy and light chains, which were based on the sequences of VH gene 3-15 (DP-38) (GenBank accession no. Z12338), JH4 gene (Z14191), IgG2 gene (J00230), VL gene 2a2 (Z73664), and J1 and C1 genes (X51755). The second lines denote amino acids translated from the reference sequences. The fourth and fifth lines, respectively, denote cDNA sequences of Dob1 VH and VL regions and their amino acid translations. The fifth lines show only the translated amino acids of Dob1 cDNAs that are different from those of the reference sequences.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

A major technical difficulty in studying the human antibody repertoire is that it is very difficult to obtain human hybridomas stably producing antibodies. Consequently, a practical alternative approach is to purify clonal antibodies from immune sera and determine their amino acid sequences (34). This alternative approach is limited by difficulties in obtaining the sequences of the internal part of the V region, which requires the cleavage of antibody molecules, generally at methionine or tryptophan. Purification of a specific peptide fragment from other cleavage peptides is difficult because only a limited amount of antibody protein (about 10 to 200 µg from 1 U of blood) can be obtained from immune sera. On the other hand, sequencing the mixture of cleavage peptides produces many amino acids at each sequencing cycle, and it is difficult to identify the sequence of the desired peptide fragment. Our strategy was to treat the mixture of peptide fragments, at the appropriate cycle of amino acid sequencing, with OPA, which blocks all peptides at the N termini except the one with proline at the N terminus. This is effective because both VH and VL regions of antibody regularly have methionine-proline or tryptophan-proline pairs. For instance, VH invariantly has tryptophan at position 36 and proline at position 41. We found that this simple strategy permits one to obtain the FR2-CDR2 region sequence with very small amounts of purified antibodies.

Our amino acid analysis of anti-6B antibodies shows that the antibodies generally use the products of genes in the VH3 (Table 1) and V2 (Fig. 1) gene families. Because our amino acid sequences are variable and incomplete, it is hard to determine exactly what VH3 family genes are used to form anti-6B antibodies. Nevertheless, we believe that the candidate VH genes are 3-07, 3-23, 3-66, and 3-74. V3-23 gene is expressed in a large number of B cells in the normal peripheral blood (38) as well as among some anti-dsDNA antibodies (6). Examining the light chain of anti-6B antibody, we observed the expression of both V and V genes. Although the product of the A27 V gene, a very commonly expressed gene (10), was found to dominate ⁺ anti-6B antibodies in some cases, the most common observation was that anti-6B antibodies encoded by multiple V genes are expressed in small amounts in all individuals. In contrast to the  chain, the  chain of anti-6B antibodies is encoded by only a few, closely related V2 genes. Among them, 2a2 has been found to be expressed most commonly (27% of all  chains [15]). There are alleles among 2a2 genes, and this allelism may affect the choice of the specific V2 genes. Taken together, the V-region structure of anti-6B antibody is typically derived from VH3 and V2 genes.

This conclusion is further supported by the analysis of a hybridoma (Dob1) secreting anti-6B antibody. Also, detailed analysis of the Dob1 sequence strongly suggests that anti-6B antibodies can be made with commonly expressed V-gene elements. The VL region of Dob1 is derived from one of the two genes belonging to V2 family (2a2 or 2b2), probably the commonly used 2a2 gene (15). Its V gene is directly joined to J1 without any evidence for unique amino acid(s) at the N region of VL, and the VH region is comprised of V3-15 and JH4, which is expressed over half of all antibodies (31). Incidentally, the V3-15 and 2a2 V combination was also observed for an antibody to H. influenzae type b PS, JB21 (1). Although the D region could not be associated with a specific D gene, it is joined with the JH gene at the typical site of the JH gene. Both VH and VL regions of Dob1 display evidence for somatic mutations, which were reported to be present among anti-PS antibodies (1, 30, 32). Taken together, the results indicate that a canonical antibody for 6B PS is formed with the V-gene components that are frequently expressed, and the structure does not suggest why humans have difficulty in producing anti-6B antibodies upon immunization.

Our study of anti-6B antibodies provides several pieces of information about -chain usage in humans. So far, the expression of human  gene components has not been studied even though the genomic organization of human V and C genes is quite distinct from those of other Ig chains. First, we show for the first time an expressed V10 gene product at the protein level. Previously, expression of this gene was noted only in an in vitro variant of a cell line (39). Compared to the translated sequence of the 10a V gene, our amino acid sequence lacked the two N-terminal residues. The cleavage of N- or C-terminal residues has been observed for antibodies with blocked N termini (37a) or enzymes (2). Second, we observed that four anti-6B antibody clones use the C1 gene only; this observation is unlikely due to random occurrence since C1 is expressed in only about 10% of all ⁺ myelomas (21) or serum Igs expressing the  light chain (37).

Our study provides a clear explanation for the observation made previously by Livneh et al. that pneumococcal vaccine elicits antibodies expressing the 8.12 idiotype (19). In fact, we found that there is a striking similarity between anti-6B antibodies and those anti-dsDNA antibodies containing VH3 and V2 gene products (26, 36). Interestingly, P16C, which uses a VL that is very similar to anti-dsDNA antibody, uses a VH1 gene product although VH3 genes are most commonly expressed for anti-6B. Perhaps anti-6B antibodies cross-reactive to dsDNA have been eliminated by apoptosis during the immune response to the pneumococcal vaccines. To test if there is a censoring of the B-cell repertoire, we are currently producing hybridomas by using fusion partners expressing Bcl-2. Despite this structural similarity, we have not yet seen a strong binding of the anti-6B antibodies to dsDNA. Also, immunization with 23-valent PS vaccines does not increase the level of anti-dsDNA antibodies in adults (14). Nevertheless, new pneumococcal conjugate vaccines may need to be tested for eliciting antibodies to dsDNA.