Somatic hypermutation introduces insertions and deletions into immunoglobulin V genes

Abstract

During a germinal center reaction, random mutations are introduced into immunoglobulin V genes to increase the affinity of antibody molecules and to further diversify the B cell repertoire. Antigen-directed selection of B cell clones that generate high affinity surface Ig results in the affinity maturation of the antibody response. The mutations of Ig genes are typically basepair substitutions, although DNA insertions and deletions have been reported to occur at a low frequency. In this study, we describe five insertion and four deletion events in otherwise somatically mutated VH gene cDNA molecules. Two of these insertions and all four deletions were obtained through the sequencing of 395 cDNA clones (approximately 110,000 nucleotides) from CD38+IgD- germinal center, and CD38-IgD- memory B cell populations from a single human tonsil. No germline genes that could have encoded these six cDNA clones were found after an extensive characterization of the genomic VH4 repertoire of the tonsil donor. These six insertions or deletions and three additional insertion events isolated from other sources occurred as triplets or multiples thereof, leaving the transcripts in frame. Additionally, 8 of 9 of these events occurred in the CDR1 or CDR2, following a pattern consistent with selection, and making it unlikely that these events were artifacts of the experimental system. The lack of similar instances in unmutated IgD+CD38- follicular mantle cDNA clones statistically associates these events to the somatic hypermutation process (P = 0.014). Close scrutiny of the 9 insertion/deletion events reported here, and of 25 additional insertions or deletions collected from the literature, suggest that secondary structural elements in the DNA sequences capable of producing loop intermediates may be a prerequisite in most instances. Furthermore, these events most frequently involve sequence motifs resembling known intrinsic hotspots of somatic hypermutation. These insertion/deletion events are consistent with models of somatic hypermutation involving an unstable polymerase enzyme complex lacking proofreading capabilities, and suggest a downregulation or alteration of DNA repair at the V locus during the hypermutation process.

Figure 5

ELISA assays showing the expression of clone pg86 with a six amino acid insertion at the FW1/CDR1 junction. Clone pg86 (IgG heavy chain) was coexpressed with the κ light chain FS6κ in insect cells using the baculovirus expression system. Expression of pg86 and its ability to pair with the FS6κ light chain was tested using capture ELISAs. Wells were coated with goat anti–human IgG. Supernatants containing recombinant antibodies were added in serial twofold dilutions. Bound antibody was detected with phosphatase-conjugated goat anti-human IgG (A), and goat anti–human Cκ (B).

Figure 1

Predicted amino acid sequences of nine cDNA clones with insertions or deletions. Note that all of these clones are extensively mutated. In all but clone tm121, the insertions or deletions occur in CDR1 or CDR2. (A and B) Two clones with insertions (A) and four clones with deletions (B) from a single tonsil. (C) Three additional cDNA clones with insertions isolated from various sources. Sequence data available from GenBank/DDBJ under accession numbers AF013615 through AF013626.

Figure 2

Comparison of the CDR1s of the human V_H4 germline genes. The primary variability between V_H4 family members is 3–6-bp size variances in the CDR1s which is similar to the short insertions and deletions that we attribute to somatic hypermutation in the selected B cell populations studied in this report.

Figure 3

Polyacrylamide gel assay to identify insertions or deletions into V_H genes. (A) Phosphorimage of a polyacrylamide gel: each lane contains the hot-PCR products (³²P-labeled) of the V_H gene and the CDR3 of an individual clone. (B) A comparison of the distribution of CDR3 sizes of the 485 CDR3s assayed to the distribution of 500 CDR3s observed in sequences from this report indicates that the clones assayed by electrophoresis were a polyclonal population. CDR3 sizes were measured from the most 3′ Tyr residue (common to all V genes analyzed) to the most 5′ Cμ or Cγ residue. CDR3 lengths for those assayed by electrophoresis were extrapolated based on sequencing of 75 out of the 485 clones assayed. The x-axis is the number of amino acids greater than the shortest CDR3 found.

Figure 4

The insertions and deletions are related to the surrounding DNA sequence. (A) The insertions involve repetitions of the immediately adjacent sequence. (B) The deletions are deletions of tandem repeats. (C) The 18-base insertion in clone pg86 is a duplication of the adjacent sequence. Nucleotides that mutated before the duplication/insertion are indicated. Sequence data available from EMBL/GenBank/DDBJ under accession numbers AF013615 through AF013626.

Figure 6

Proposed mechanism causing insertion/deletion events: polymerase slippage. This Figure is based on model a of Streisinger et al. (30) and Ripley (31). The same model can account for both (A) insertions and (B) deletions.