B-cell responses to a peptide epitope: mutations in heavy chain alone lead to maturation of antibody responses

Abstract

In the present study, the genetic mechanisms responsible for generation of antibodies recognizing the dominant epitope within a synthetic peptide PS1CT3 were examined. PS1CT3 is a peptide model antigen containing residues 28–42 of the large protein of the surface antigen of hepatitis B virus as B epitope (designated PS1), and the known T-helper-cell epitope derived from the circumsporozoite protein of the malaria parasite Plasmodium falciparum (designated CT3). To characterize the repertoire generated, the immunoglobulin heavy chain variable regions from IgM and IgG monoclonal antibodies against PS1CT3 were sequenced. Although all IgG monoclonal antibodies were directed against the immunodominant epitope, the genetic elements used were diverse. Comparison of the sequence of germ line precursor IgM to a mature IgG revealed that during maturation of the primary IgM response only the heavy chain fragment of the antibody molecule underwent somatic mutation.

INTRODUCTION

The nature of the B-cell response to T-cell-dependent antigens is extremely dynamic: extensive changes take place in both the clonal composition of the responding B-cell population and the structure and function of the antibodies these clones express. During this maturation process, somatic hypermutation of antibody variable (V) region genes is induced.^1,2 Proteins are complex antigens and the entire surface of a protein represents an antigenic continuum.^3–5 It has been reported that only a subset of the total epitope repertoire on a protein antigen is preferentially recognized, and antibodies to similar epitopes can be diverse in their V gene expression.^6–11

The model antigen used for the present study is a 15-residue sequence that is derived from the large protein of the surface antigen of hepatitis B virus. It also contains well-characterized T-cell epitopes.¹² It was observed that although the early primary IgM response to this peptide was heterogeneous with multiple specificities, only a subset of antibodies specifically directed against a subsegment of the peptide was stringently selected for maturation into IgG.¹² To understand the genetic mechanisms responsible for the generation of paratopes recognizing the different epitopes, the variable regions of the heavy chain of the IgM and IgG monoclonal antibodies have been cloned, sequenced and characterized. The germ line precursor to a mature IgG monoclonal antibody belonging to the V_H36–60 was isolated from an antigen-specific IgM expression library. The sequence of the IgG heavy chain variable region showed a number of somatic mutations compared with the precursor isolated from the antigen-specific IgM expression library. On the other hand, no significant mutations were found in the light chain variable region of the mature IgG compared with the precursor.

MATERIALS AND METHODS

Peptide, animals and immunization

The peptide PS1CT3 used in this study was generated by solid-phase synthesis, as described previously.¹² Female BALB/c mice (6–8 weeks old) were obtained from the small animal facility at the National Institute of Nutrition (Hyderabad, India). Immunizations of mice were intraperitoneal, with a dose of 50 μg/mouse as an emulsion in complete Freund's adjuvant (CFA). The volume injected per mouse was 100 μl. Mice were bled from the retro-orbital plexus and sera assayed for various antibodies via an enzyme-linked immunosorbent assay (ELISA).

Generation of IgM and IgG monoclonal antibodies

For early primary IgM monoclonal antibodies (mAb), a group of three female BALB/c mice was immunized with peptide PS1CT3 as described above. After 4 days, mice were bled individually and sera were checked for anti-PS1CT3 antibody titre. The highest responder mouse was chosen for hybridoma production. Hybridomas were generated from the total spleen cells of mice using the hypoxanthine-aminopterin-thymidine-sensitive myeloma-derivative SP2/0 fusion partner, as described previously.¹³ For the mature IgG mAb, mice were primed as above, rested for 1 month, then boosted with a 50-μg dose of soluble peptide in phosphate-buffered saline (PBS) and their spleen cells fused 3 days later. Fusions were screened for hybridomas producing anti-PS1CT3 mAb via ELISA, and cells in positive wells were subcloned by limiting dilution.¹³

RNA extraction, cDNA preparation, PCR amplification and sequencing

Total RNA from about 10⁷ hybridoma cells was isolated using the RNAzol B reagent (Wak-Chemie Medical, GmbH, Germany). For reverse transcription about 10 μg of total RNA was used. The first strand of cDNA was synthesized using 20 U of avian myeloblastosis virus reverse transcriptase (Promega, Madison, WI) and 800 pmol of the primer 5′-AGACGAGGGGGAAGACATTT-3′ for Cμ, 5′-GGCCA GTGGATAGAC-3′ for Cγ and 5′-GCTTCACTGGAT GGTGGGAAGATG-3′ for Cκ in a final volume of 25 μl. After reverse transcription 5 μl of product (cDNA) was used for amplification. The 5′ primers used were 5′-AGGT(C/G)(A/C)A(A/G)CTGCAG(G/C)AGTC(A/T) GG-3′ for V_H and 5′-GA(A/C/T)ATTG TG(A/C)T(G/C)AC(A/C)CA(A/G)(A/T)CTCCA-3′ for V_L. A different set of nested 3′ primers were used for amplification, 5′-CATTTGGGAAGGACTGACTC-3′ for Cμ, 5′-GGCCAGTGGATAGAC (T/C/A)GA-3′ for Cγ and 5′-GAAGATGGATACAGTTGGTGCA-3′ for Cκ. The polymerase chain reaction (PCR) mixture contained 200 mm dNTP, 50 mm KCl, 10 mm Tris–HCl (pH 8·8), 1·5 m MgCl₂, 200 pmol of each primer and 2·5 U of Taq DNA polymerase (Stratagene, La Jolla, CA). PCR was performed in a DNA thermal cycler (Perkin-Elmer Cetus, Norwalk, CT) using the following programme: one cycle at 95° for 5 min, followed by 30 cycles each of 1 min at 94°, 1·5 min at 58° and 1 min at 72°. There was a final extension at 72° for 10 min.

An aliquot of the reaction mixture was analysed on an agarose gel and PCR products of about 400 base pairs were blunt-ended by polishing with a PCR polishing kit (Stratagene), and subsequently cloned in the SrfI site of a PCR-script plasmid (Stratagene). Positive clones were identified by restriction digestion of plasmid DNA and sequenced using the T7 dideoxy sequencing system (Pharmacia, Uppsala, Sweden). Both strands of the cloned DNA were sequenced. After each PCR and cloning, multiple bacterial colonies were used for sequencing in order to rule out reverse transcription or PCR-induced error. The gene sequences were analysed using the PC gene software and the Genbank data library. The sequences reported in this paper have been deposited in the Genbank data base (accession numbers Y17364–Y17375).

Construction of antigen-specific IgM immunoexpression libraries

A group of 10 mice was immunized with 50 μg of PS1CT3 in CFA, killed on day 5 and the spleen cells collected. Red blood cells were removed by incubating with red blood cell (RBC) lysis buffer (Sigma, St Louis, MO). Thereafter, antigen-specific B cells were collected by incubation with Dynal beads coupled to antigen as described in the manufacturer's instructions (Dynal, Oslo, Norway). The antigen-specific B cells thus obtained were amplified further by incubation with lipopolysaccharide (LPS) (50 μg/ml; Sigma) for 48 hr. Cells were harvested and total RNA was isolated by the RNAzol method.¹¹ The ImmunoZap cloning and expression system (Stratagene) was used to construct the cDNA library.

Isolated RNA was converted to cDNA using oligo dT primer as described in the supplier's instructions (Stratagene). PCR amplification of the immunoglobulin heavy (H) and light (L) chain sequences was done separately with 3′ primers specific for IgM heavy chain and kappa light chain. The primers used for heavy chain amplification were a common IgM 3′ primer, 5′-TCTGCACTAGTTGGAATGGGCACA TGCAG-3′, and a number of 5′ primers, V_Ha 5′-AGGTC CAGCTGCTCGAGTCTGG-3′, V_Hb 5′-AGGTCCAGCTG CTCGAGTCAGG-3′, V_Hc 5′-AGGTCCAGCTTCTCGAGT CTGG-3′ and V_Hd 5′-AGGTCCAGCTTCTCGAG TCAGG-3′. The primers used for amplification of the heavy chain represent amino acids 1–8 of the heavy chain, and this region is highly conserved between various gene families. Therefore these primers will amplify the sequences from major V_H families such as J558, Q52, 7183 and 36–60. If only V_H36–60 gene family-specific primers are used, the library would be biased and it would not provide the actual representation of the precursor in the antigen-activated pool. The primers used for light chain amplification were 5′ primer V_Lk 5′-GTGCCAGATGTGAGCTCGTGATGACCCAGTCTCCA-3′ and 3′ primer C_Lk 5′-TCCTTCTAGATTACTAACACTCTCCCCTGTTGAA-3′. The PCR-amplified heavy chain DNA fragments were digested with SpeI and XhoI and ligated into the ImmunoZap H vector. Similarly, the light chain products were digested with SacI and XbaI and ligated into the ImmunoZap L vector. The ligated recombinant phage DNA was incorporated into bacteriophage with in vitro packaging extract.

Co-expression of the H fragment and Light chains was also carried out, so that a Fab-like species would result. To construct a phage for co-expression, the right arm of heavy chain phage DNA was digested with HindIII, preserving the left arm of ImmunoZap H with an H insert, and the left arm of light chain DNA was digested with MluI, yielding a right arm of ImmunoZap L with an L insert. Both products were then digested with EcoRI and ligated to create a combinatorial phage. The DNA was packaged and used to infect Escherichia coli. The resulting library was amplified and used for screening.

Screening of immunoexpression libraries

To screen for expression of H or L fragment, E. coli was infected to yield c. 200 plaques per plate. The phage DNA was transferred to nitrocellulose and treated as described elsewhere.¹⁴ The recombinant library was screened with ³²P 5′-end labelled oligonucleotide probes specific for the CDR3 region of mAb PC283 and PC289.¹⁵ The probes used were: heavy chain, 5′-GCAAGAGGGGGGACGGGATTTGCTTA CTGG-3′; light chain, 5′-CACTGTGGACAGACTTACAGC TATCCGACGTTC-3′. Hybridization was carried out at 42° overnight using the Rapid-hyb buffer (Amersham USA, Arlington Heights, IL). The filters were washed first at room temperature for 20 min in 2 × SSC, 0·1% (w/v) sodium dodecyl sulphate (SDS), and then at 42° for 30 min in 0·1 × SSC, 0·1% (w/v) SDS. Positive plaques were purified further through secondary and tertiary screenings. The plasmid DNA was prepared from positive phages by in vivo excision and sequenced, as described above.

RESULTS AND DISCUSSION

The immunogen used in these studies was a synthetic peptide (PS1CT3) of the sequence HQLDPAFGANSTNPDGGDI EKKIAKMEKASSVFNVVNS. The first 15 residues of the peptide represent a B-cell epitope followed by two spacer residues. The carboxy terminal 21 residues correspond to well-characterized T-cell epitopes.¹² It has been shown earlier that the peptide PS1CT3 elicits a T-cell-dependent response, and the early primary IgM response directed against PS1 displays multiple specificities while the mature primary polyclonal response is restricted and is exclusively directed against a tetrapeptide stretch of PS1 residues 4–7 (DPAF).¹²

Antibody repertoire of the PS1CT3-specific IgM response

To examine multiple specificity of the IgM response in greater detail, monoclonal antibodies were generated against PS1CT3. A total of seven independently derived clones was obtained, of which four reacted with multiple overlapping hexapeptides suggesting polyreactivity, while the remaining three reacted with a limited set of hexapeptides. Of these, only one showed specific binding to the tetrapeptide DPAF.¹²

The genetic repertoire of these monoclonal antibodies was examined by determining the nucleotide sequence of their heavy chain variable regions. The results are shown in Fig. 1a. The IgM response to PS1CT3 was oligoclonal and paratypically diverse as it was encoded by different heavy chain V region gene segments. All the antibodies were found to derive from distinct B cells because they had utilized diverse V_H, D_H and J_H gene combinations. The length of the CDR3 region ranged from three to 16 amino acids (Fig. 1b). An abundance of V_H genes from the J558 family was observed (five out of seven) but each represented a separate member of the family (Table 1). From these data it appears that this peptide is capable of inducing a variety of B cells that have distinct phenotypic and genotypic paratopes.

Figure 1.

(a) Nucleotide sequence of the heavy chain variable region of PS1CT3 IgM monoclonal antibodies. The complementarity-determining regions (CDR) are underlined. The codons are numbered according to Kabat et al.¹⁷ (b) Amino acid sequence of the heavy chain variable region of PS1CT3 IgM monoclonal antibodies. The CDR are in bold and the numbering is according to Kabat et al.¹⁷

Table 1.

Immunoglobulin gene segment usage of early primary IgM monoclonal antibodies

Level of antibody heterogeneity in the monoclonal IgG response to PS1CT3

In contrast to IgM heterogeneity, as shown previously,¹² the earliest detectable anti-PS1CT3 IgG antibody response was predominantly monospecific. The monospecificity of this response has already been studied through the generation of monoclonal antibodies.¹² A total of 11 stable hybridomas was produced, all of which were directed to the tetrapeptide stretch DPAF.¹² In a more recent study we have also examined the clonal relatedness of these hybridomas by comparing the CDR regions.¹⁵ The complete nucleotide sequence of the heavy chain variable region genes from these hybridomas is reported in Fig. 2a. The monospecificity of the PS1CT3 IgG response was not due to the bias for one particular V gene usage¹⁵ but was oligoclonal, as shown in Fig. 2a. A large degree of structural diversity among IgG antibodies that share specificity for the dominant epitope is found in the V_H domains¹⁵ that are encoded by genes belonging to the J558, 36–60 and miscellaneous gene families (Fig. 2a). Furthermore, there are no apparent constraints on the D_H or J_H gene segments to which these V_H genes are joined. There are at least four distinct B-cell precursors for these clones with diverse V_H, D_H and J_H gene usage, and all V genes have different numbers of inserted nucleotides (N regions at the V_H–D_H and/or D_H–J_H junctions) resulting in the variation in sequence as well as length of the CDR3 region, which ranges from three (PC 2811) to 13 (PC 2812) amino acids (Fig. 2b). The average length of CDR3 in these hybridomas was 7·3, which compares to an average CDR3 length of 9·5 for murine antibody genes in general.¹⁶ The only apparent restriction is that J_H1 was not observed among these monoclonal antibodies (Fig. 2a). Possibly the side chains of the Tyr and/or the Gln (amino acid residues differing from J_H1) provide an essential hydrogen bond either for antigen contact or for stabilizing the conformation of the combining site.¹⁷

Figure 2.

(a) V_H gene nucleotide sequence of mature anti-PS1CT3 IgG. The sequence of PC 283, 287 and 289 has been taken from reference 15 for comparison. The CDR are underlined. The V_H and J_H segments used by each antibody are shown after the sequence. Dashes in the nucleotide sequence indicate that the sequence continues and no nucleotides appear in those positions. The codons are numbered according to Kabat et al.¹⁷ (b) Amino acid sequence of the heavy chain variable region of mature anti-PS1CT3 IgG. The CDR are in bold and the numbering is according to Kabat et al.¹⁷

It may be possible that the overall paratope structure is almost similar in all cases, which results in the ideal fit for the immunodominant epitope, as it has been suggested that there are relatively few canonical three-dimensional structures for most CDR despite the large variation in amino acid sequences that are observed.¹⁸

As in this study, it has been demonstrated previously that very different antibodies can bind a decapeptide epitope of the tobacco mosaic virus protein.⁸ Both the epitope and the antibody may be viewed as bearing multiple sites for potential contact, with actual binding involving only a few of these potential contact sites. A comparison of IgG and IgM sequences revealed that our IgM monoclonal antibody panel did not contain sequences that were homologous to any of the IgG monoclonal antibody heavy chain variable regions. Using hybridoma technology, it is possible to immortalize only a small fraction of the specific antibody-forming cells available in an immunized animal. It may be possible that due to low clonal frequency of the precursor, such a sequence is not represented among the hybridomas that we have obtained. Therefore, in order to identify germ line precursors to the antigen-activated B cells an antigen-specific IgM expression library was constructed using the primers specific for major V gene families.

Isolation of heavy and light chain precursors from an antigen-specific library

From the IgG pool of monoclonal antibodies we selected two representatives having the same CDR3 regions (PC 283 and 289) to study the process of clonal selection in detail. The IgM expression library was screened with a probe specific for the CDR3 region of heavy or light chains. About 5000 plaques were screened with the heavy chain-specific probe. This resulted in the isolation of five positive plaques, which were validated by secondary and tertiary screening. After in vivo excision of plasmid from the phage, all of these clones were sequenced and found to be similar. A comparison of the precursor sequence with mature IgG clones is shown in Fig. 3a. A number of mutations, at positions 32 and 34 in CDR1, positions 52, 53, 54, 57, 58, 60 and 65 in CDR2 and positions 99 and 100 in CDR3, in the precursor clones was identified, which resulted in an amino acid change in the mature clones (Fig. 3a and Fig. 4). There were a few mutations in the framework regions as well; the difference in the beginning (up to amino acid 8) was due to a difference in the primer sequence used for amplification. The ratio of replacement to silent mutations was significantly higher in CDR regions (10 replacement and none silent in PC283, 11 replacement and none silent in PC289), compared with the framework regions (two replacement at position 23 from Ser to Thr, 81 from Lys to Gln and nine silent in PC283, and three replacement at position 13 from Lys to Asn, 23 from Ser to Thr 81 from Lys to Gln and nine silent in PC289), as antigen–antibody interactions are mediated mainly by amino acid residues in the CDR regions. Because this bias towards replacement mutations is mainly in CDR, it appears likely that the selection was for an improved ability to bind the antigen.

Figure 3.

Comparison of the sequence between the V_H genes (a) and V_L genes (b) of precursor (isolated from the library) and mature IgG hybridoma. Sequence identity to the precursor is indicated by dots, differences are shown explicitly. The CDR are underlined and numbering of codons is according to Kabat et al.¹⁷

Figure 4.

Possible genealogical relationship of V_H and V_KL DNA sequences. Mutations in various CDR are indicated by the codon numbers (Kabat et al.¹⁷) and amino acids are shown using the single letter code.

About 3000 plaques of the library were also screened with the light chain-specific probe. This resulted in the isolation of 10 positive plaques, which were further confirmed by secondary and tertiary screening. All the 10 clones were sequenced and found to be identical. A comparison of the precursor sequence with the mature light chain IgG sequence is shown in Fig. 3b. In this case the precursor and mature light chain sequences were almost identical. There were no mutations in the CDR regions, except one silent mutation in CDR3. The difference in the beginning in the first framework region was due to the difference in the primer sequence used for amplification. There were a few silent mutations but only two replacement mutations (at position 67 from Ser to Pro and 72 from Ile to Thr) that resulted in amino acid change in the third framework region (Fig. 3b); these changes were probably not significant because framework regions do not play an important role in antigen–antibody interactions. The sequence of precursor and mature V_L gene is also c. 97% homologous to the germ line gene¹⁹ (data not shown). These observations show collectively that, during maturation of the IgM response, no significant mutations took place in the antibody V_L gene (Fig. 3b). The possible relationship between the precursor and the mature IgG clones is shown in Fig. 4. It appears that during maturation of the primary IgM response to peptides, only the heavy chain portion of the antibody molecule is somatically mutated because a disproportionately large part of the antigen-binding pocket is contributed by the heavy chain. It has also been shown previously, in the secondary response to the influenza virus haemagglutinin, that in particular the heavy chain is important in determining the antigen specificity.²⁰

Our results presented here are consistent with the fact that an appreciable amount of the binding energy comes from the heavy chain alone, and many antibodies with quite different specificities use similar light chains.²¹ Thus, the combination of a heavy chain with any light chain that allows formation of the combining site without creating unfavourable steric or electrostatic interactions would preserve a significant portion of the antibody specificity.