Frequency and genetic characterization of V(DD)J recombinants in the human peripheral blood antibody repertoire

Abstract

Antibody heavy-chain recombination that results in the incorporation of multiple diversity (D) genes, although uncommon, contributes substantially to the diversity of the human antibody repertoire. Such recombination allows the generation of heavy chain complementarity determining region 3 (HCDR3) regions of extreme length and enables junctional regions that, because of the nucleotide bias of N-addition regions, are difficult to produce through normal V(D)J recombination. Although this non-classical recombination process has been observed infrequently, comprehensive analysis of the frequency and genetic characteristics of such events in the human peripheral blood antibody repertoire has not been possible because of the rarity of such recombinants and the limitations of traditional sequencing technologies. Here, through the use of high-throughput sequencing of the normal human peripheral blood antibody repertoire, we analysed the frequency and genetic characteristics of V(DD)J recombinants. We found that these recombinations were present in approximately 1 in 800 circulating B cells, and that the frequency was severely reduced in memory cell subsets. We also found that V(DD)J recombination can occur across the spectrum of diversity genes, indicating that virtually all recombination signal sequences that flank diversity genes are amenable to V(DD)J recombination. Finally, we observed a repertoire bias in the diversity gene repertoire at the upstream (5') position, and discovered that this bias was primarily attributable to the order of diversity genes in the genomic locus.

Figure 1

Stringent filtering of V(DD)J recombinants. A sample V(DD)J recombinant (3FV6DI) is shown, along with the germline diversity gene assignments for both the 5′ D (D2-2) and 3′ D (D2-8) positions. Dashes indicate conservation between germline and 3FV6DI and dots indicate mismatches. The match region is indicated above the 3FV6DI sequence, along with the match score (the match region length minus any mismatches within the match region). The germline segment length is shown below the sequence alignment. Below the germline segment length, the scoring calculation is shown. Sequences that contained scores of > 60% for both diversity gene positions were considered V(DD)J rearrangements.

Figure 2

Putative V(DD)J recombinants contain normal N-addition lengths and diversity genes, with low GC content. (a) N-addition length for each recombination site in the total naive repertoire or in the V(DD)J repertoire. The mean N-addition length for each of four healthy individuals ± SEM is shown for each recombination site. (b) GC content (as a percentage of the region sequence) for each N-addition region or diversity gene segment in the V(DD)J repertoire or the total naive repertoire. Combined N-addition at the VD and DJ junctions (N) or diversity gene region (D) are shown for the total repertoire. Diversity genes at the 5′ D position (D1) and 3′ D position (D2) and N-addition sites at the VD junction (N1), DD junction (N2) and DJ junction (N3) are shown for putative V(DD)J recombinants.

Figure 3

Frequency and diversity gene use of putative V(DD)J recombinants. (a) V(DD)J recombinant frequency (as a percentage of the subset repertoire) of naive, IgM memory or IgG memory B-cell subsets isolated from the peripheral blood of four healthy individuals. The V(DD)J frequency for each donor ± SEM for each subset is shown. Pairwise comparisons of V(DD)J recombinant frequency between different subsets were determined using a one-way analysis of variance with Bonferroni's correction. (b) Histogram of CDR3 length distribution of V(DD)J recombinations (filled bars). The length distribution for the entire repertoire (dashed line) is also shown for comparison. **P < 0·01.

Figure 4

Diversity gene use in putative V(DD)J recombinants differs from that in the total naive repertoire. (a) Diversity gene family use for the total naive repertoire (Naive) or for the 5′ D and 3′ D positions in V(DD)J recombinants. (b) Diversity gene use in the total repertoire, at the D1 position in V(DD)J recombinants, and at the D2 position of V(DD)J recombinants. Mean ± SEM for each donor is shown. Pairwise comparisons were determined using a two-way analysis of variance with Bonferroni's correction. *P < 0·05; **P < 0·01; ***P < 0·001.

Figure 5

Genomic orientation of diversity genes matches the orientation in putative V(DD)J recombinants and explains diversity gene use bias at 5′ D and 3′ D positions. (a) All functional diversity genes are represented in the genomic orientation, with arcs connecting diversity genes found paired in a V(DD)J recombinant. Black boxes indicate diversity genes found only in the 5′ D position, white boxes indicate diversity genes found only in the 3′ D position, and grey boxes indicate diversity genes found in both positions. D4-11, D1-14 and D6-25, the only diversity genes not found in any V(DD)J recombinants, are represented by black circles. The orientation of diversity genes in V(DD)J recombinants matches the genomic orientation in every instance, so the leftmost member of any linked pair of diversity genes shown in this diagram was always in the 5′ D position of the V(DD)J recombination. (b) The frequency of each diversity gene in either the 5′ D or 3′ D positions is shown with the diversity genes ordered and labelled by position from the 5′ end of the genomic locus. (c) The frequency of each diversity gene in the total repertoire (including all B-cell subsets). Shown are the mean frequency ± SEM for each donor. (d) The frequency of each recombination span, defined as the distance between the 5′ D gene and the paired 3′ D gene in V(DD)J recombinants (measured in diversity gene segments and including non-functional genes). Recombination between adjacent diversity genes results in a recombination span of 1.