Positive selection of the peripheral B cell repertoire in gut-associated lymphoid tissues
http://www.100md.com
《实验药学杂志》
Gut-associated lymphoid tissues (GALTs) interact with intestinal microflora to drive GALT development and diversify the primary antibody repertoire; however, the molecular mechanisms that link these events remain elusive. Alicia rabbits provide an excellent model to investigate the relationship between GALT, intestinal microflora, and modulation of the antibody repertoire. Most B cells in neonatal Alicia rabbits express VHn allotype immunoglobulin (Ig)M. Within weeks, the number of VHn B cells decreases, whereas VHa allotype B cells increase in number and become predominant. We hypothesized that the repertoire shift from VHn to VHa B cells results from interactions between GALT and intestinal microflora. To test this hypothesis, we surgically removed organized GALT from newborn Alicia pups and ligated the appendix to sequester it from intestinal microflora. Flow cytometry and nucleotide sequence analyses revealed that the VHn to VHa repertoire shift did not occur, demonstrating the requirement for interactions between GALT and intestinal microflora in the selective expansion of VHa B cells. By comparing amino acid sequences of VHn and VHa Ig, we identified a putative VH ligand binding site for a bacterial or endogenous B cell superantigen. We propose that interaction of such a superantigen with VHa B cells results in their selective expansion.
Abbreviations used: BCR, B cell receptor; FR, framework region; GALT, gut-associated lymphoid tissue; LigApx, ligated appendix.
K.-J. Rhee and P.J. Jasper contributed equally to this work.
Vertebrates have developed two general strategies for generating a diverse primary B cell repertoire. In humans and mice, the primary B cell repertoire is generated by rearrangement of multiple V, D, and J gene segments in the bone marrow throughout the life of the animal. In other species, such as the rabbit (1–3), chicken (4, 5), and sheep (6, 7), this repertoire initially develops by rearrangement of a limited number of V genes in primary lymphoid tissue and further diversifies in gut-associated lymphoid tissues (GALTs). In rabbits, the D proximal VH gene, VH1, is preferentially rearranged during B cell development in the fetal liver and bone marrow (8). The VDJ genes undergo somatic diversification via somatic hypermutation and gene conversion in GALT in response to intestinal microflora (9). In the absence of appropriate intestinal microflora, GALT develops poorly, and both the number of B cells and the diversification of VH genes are greatly inhibited (9).
Most (80–90%) rabbit serum Ig molecules express VHa allotypic markers that are encoded by the predominantly rearranged gene, VH1 (1). The following three alleles of VH1 are found in laboratory rabbits: VH1-a1, VH1-a2, and VH1-a3; they encode the VHa1, VHa2, and VHa3 allotypes, respectively. These VHa allotypes differ in amino acid residues in framework region (FR)1 and FR3 (10). 10–20% of serum Ig does not react with anti-VHa1, anti-VHa2, or anti-VHa3 allotypic antibodies and is referred to as VHn (VHa-negative) Ig.
Kelus and Weiss (11) identified rabbits with a variant VHa2 allotype-encoding allele, ali, which has a 10-kb deletion of DNA encompassing VH1 (Fig. 1 and reference 1). In contrast with wild-type rabbits, nearly all Ig in young ali/ali rabbits (designated Alicia) is VHn. VHn Ig is encoded predominantly by VHx, VHy, and VHz (12, 13), which reside >50 kb upstream of VH1 (1). In adult Alicia rabbits, high levels of serum Ig with the VHa (a2) allotype are found (11). This increase in VHa Ig is a result of increased numbers of VHa B cells that use VH4, VH7, and VH9, gene segments that encode several of the VHa (a2) allotype-associated amino acids (14, 15). Pospisil et al. (16) found that, in the appendix, more VHa B cells were proliferating and fewer were dying compared with VHn B cells. The molecular basis underlying the repertoire shift from VHn to VHa is unknown and is the subject of the current paper.
A shift in the B cell repertoire could arise from VH gene replacement or from secondary Ig gene rearrangements on the unexpressed Ig allele. Although it is generally believed that these events occur primarily in the bone marrow (17–19), there is evidence that VH gene replacement and secondary Ig gene rearrangement occur in peripheral tissues (20, 21). Another possible explanation for the B cell repertoire shift is that VHa B cells are positively selected in the periphery. Positive selection of B cells in the periphery has been demonstrated in several transgenic mouse models (22–26).
In rabbit, intestinal microflora interacts with GALT to promote development of follicles containing proliferating B cells and to generate the primary B cell repertoire (9). We hypothesized that, in GALT of Alicia rabbits, interactions between GALT and the intestinal microflora also promote the increased proliferation of VHa B cells compared with VHn B cells and lead to the repertoire shift from VHn to VHa B cells. To test this possibility, we surgically disrupted the GALT–bacterial interaction in Alicia rabbits and tested whether the repertoire shift from VHn to VHa B cells was abrogated.
Results
Kinetics of the B cell repertoire shift
The repertoire shift in Alicia rabbits, from the predominant expression of VHn allotype early in life to the predominant expression of VHa allotype later in life, was originally shown by Kelus and Weiss (11), who analyzed Ig allotypes in serum. Pospisil et al. (16) showed that a similar shift occurred in B cells in the appendix. By using antibodies to both VHn and VHa allotypes, we found that, in 9-wk-old rabbits, VHa B cells represented 35–50% of the B cells in spleen, mesenteric lymph nodes, appendix, and PBLs (Fig. 2). We analyzed cells of various tissues from newborn to 2-yr-old Alicia rabbits to follow the appearance and disappearance of VHa and VHn B cells, respectively, throughout life. We found that, although 10–25% of B cells at birth are VHa, at 3 wk of age essentially all B cells (95%) in spleen, appendix, and PBLs were VHn (Fig. 3 and not depicted). Subsequently, the percentage of VHn cells steadily declined, so that by 2 yr of age, <20% of B cells were VHn and >75% were VHa. These data demonstrate that VHa B cells accumulate throughout life, with a sharp increase between 4 and 10 wk of age. The VHa B cells accumulate faster in the appendix than in spleen, suggesting that the B cell repertoire shift from VHn to VHa B cells may occur primarily in GALT.
B cell repertoire shift and GALT
Because GALT development and somatic diversification of Ig genes both require interaction between GALT and intestinal microflora (9, 27), we hypothesized that the repertoire shift in Alicia rabbits also requires this interaction. To investigate this possibility, we generated ligated appendix (LigApx) rabbits by surgically removing the Peyer's patches and the sacculus rotundus and ligating the lumen of the appendix to prevent bacterial colonization (9). If interactions between GALT and intestinal microflora are required for the repertoire shift from VHn to VHa B cells, we expected that the peripheral blood B cells in LigApx rabbits would be predominantly VHn. In each of three 12-wk-old LigApx Alicia rabbits (94S, 353X2, 353X4), we found that the percentage of B cells was approximately eightfold less than in unmanipulated rabbits of that age and that almost all B cells (90–96%) were VHn (Fig. 4). As expected, the percentage of VHn B cells in unmanipulated Alicia rabbits was 50%. We examined one of the LigApx Alicia rabbits (94S) at 8 mo of age and found that >90% of the B cells were still VHn, showing that the B cells remained predominantly VHn for many months. These results indicate that, without interactions between GALT and intestinal flora, the shift from VHn to VHa did not occur.
To confirm that the repertoire shift from VHn to VHa B cells was abrogated in LigApx rabbits, we examined the nucleotide sequences of VDJ genes cloned from peripheral blood of 12-wk-old LigApx Alicia rabbits. We expected that the VH genes used in the VDJ gene rearrangements would be primarily genes that encode VHn molecules rather than VHa molecules. From one LigApx Alicia rabbit (94S), shown in Fig. 4, and from three additional LigApx rabbits (32P2, 144T, 199T1) for which flow cytometry data are not available, we analyzed a total of 80 VDJ gene sequences. As predicted, most (76%) of the VDJ genes used VHn gene segments (Table I), whereas almost none (4%) of the VDJ genes from control (unmanipulated) Alicia rabbits of the same age used VHn gene segments. We think the PCR analysis underestimated the expression of VHn genes because, by FACS analysis, 50% of the peripheral B cells from 12-wk-old control Alicia rabbits were VHn, whereas only 4% of the PCR-amplified VDJ genes were VHn.
To determine whether the low percentage of VHn-encoding genes (Table I) resulted from preferential amplification of VHa cDNA, we conducted two independent experiments in which VDJ genes were PCR amplified from cDNA prepared from a pool of cells containing equivalent numbers of FACS-sorted VHa and VHn B cells from peripheral blood. In the two experiments, 67% (14 out of 21) and 81% (17 out of 21) of PCR-amplified VDJ genes used VHa gene segments. The reduced number of VHn PCR products was also observed with another 5' VH primer, VHldr (5'-GGCTTCTCCTGGTCGCTG-3'), which anneals to a different target site. The preferential amplification of VHa cDNA with independent primers suggests that, even though VHa and VHn B cells appear to express equivalent amounts of surface IgM (Fig. 4), VHa B cells might produce higher levels of IgM mRNA, possibly as a result of their stimulation in GALT (16)
Although we do not understand the molecular basis for the PCR skewing toward VHa-encoding VDJ genes, the data confirm the FACS analysis, which showed that most of the B cells of LigApx Alicia rabbits were VHn instead of VHa. We conclude that the repertoire shift from VHn to VHa B cells required interactions between GALT and the microflora and that expansion of VHa B cells requires such interactions.
Rearrangement status of IgH alleles in VHa B cell
The repertoire shift from VHn to VHa B cells in the periphery could occur by replacement of VHn-using VDJ genes with VHa gene segments (20), by rearrangement of a VHa-encoding VHa gene segment on the second IgH allele (17, 21), or by selective expansion of VHa B cells (16, 24). We think that VH gene replacement is unlikely to explain the repertoire shift because the VHn genes (VHx, VHy, and VHz) used in VDJ gene rearrangements in VHn B cells reside upstream of the VHa genes (VH4, VH7, and VH9) used in VDJ gene rearrangements in peripheral VHa B cells of Alicia rabbits (15). Accordingly, the rearrangement of VHn genes during VDJ gene rearrangements would likely result in deletion of the VHa genes (Fig. 1).
If the repertoire shift from VHn to VHa B cells is caused by gene rearrangements of VHa gene segments on the second IgH allele in VHn B cells, we expected to find VDJ gene rearrangements on both IgH alleles in VHa B cells. To test this possibility, we sorted VHa B cells from an adult Alicia rabbit and assessed the status of VDJ gene rearrangements by single cell PCR. We used PCR primers that would detect rearranged VDJ genes and germline JH genes (Fig. 5 a). Of 26 single cells from which a rearranged VDJ PCR product was obtained, all but one had a product of the expected size for an unrearranged (second) IgH allele (Fig. 5 b). This result showed that essentially all VHa B cells rearranged only one IgH allele, indicating that the B cell repertoire shift from VHn to VHa B cells in Alicia rabbits is not due to secondary IgH gene rearrangements on the other allele. Instead, we propose that the B cell repertoire shift occurs through positive selection due to preferential expansion of VHa B cells.
B cell receptor (BCR) signaling in VHa and VHn B cells
One possible explanation for the preferential expansion of VHa B cells and the concomitant decrease in VHn B cells in Alicia rabbits is that VHa B cells are more responsive to BCR stimulation than VHn B cells. To test this possibility, we assessed the release of intracellular calcium after BCR cross-linking on VHa and VHn B cells from 12-wk-old Alicia rabbits. The Alicia rabbits had the b5 chain allotype; therefore, we incubated PBLs with anti-b5 antibody and measured the release of intracellular calcium, as described in Materials and Methods. We found that the VHn B cells responded to anti-b5 antibody as well as the VHa B cells when the differences in baseline stimulation were taken into account (Fig. 6). Anti-b4 chain allotype antibody served as a negative control. Although we cannot explain the different baseline stimulation of VHa and VHn B cells, we conclude that the inherent signaling capacity of VHa and VHn B cells is similar and, therefore, does not explain the selective expansion of VHn B cells.
Discussion
The intestinal microflora are important in regulating many immune functions, including development of GALT (27), induction of oral tolerance (28), and induction of mucosal immunity (29). In rabbits, intestinal microflora are required not only for GALT to develop but also to generate a diverse primary B cell repertoire (9). Previously, we found that surgical disruption of GALT–bacterial interactions prevented GALT development, B cell expansion, and somatic diversification of the B cell repertoire (9). In the current paper, we found that the repertoire shift from VHn to VHa B cells in Alicia rabbits also depends on GALT–bacterial interactions.
At birth, 10–25% of B cells in peripheral tissues of Alicia rabbits were VHa, and these B cells subsequently declined to nearly undetectable levels by week 3. Although these VHa B cells could represent maternal B cells, we think the percentages are much higher than would be expected for maternal B cells. We also do not think these cells are VHn B cells with maternal VHa Ig bound through Fc receptors because, in this case, we would expect all B cells, rather than a subset, to be VHa. Instead, we think the decline in the percentage of VHa B cells may be due to a dramatic increase in VHn B cells from a second wave of B lymphopoiesis in the bone marrow. We recently identified a burst of both pre–B cells and B cells in bone marrow at 3 wk of age and we suggest that in Alicia rabbits, the newly generated B cells may be primarily VHn (30).
The shift from VHn to VHa B cells after 3 wk of age likely occurs in GALT rather than in the bone marrow because the shift requires GALT–bacterial interactions. Therefore, we favor the idea that this shift is due to selective expansion of VHa B cells as proposed by Pospisil et al. (16), who showed that more VHa B cells proliferate and fewer die than VHn B cells in the appendices of Alicia rabbits. We suggest that VHa B cells are preferentially stimulated by interaction with a bacterial ligand or a bacterially induced GALT-derived ligand. Such preferential stimulation of VHa B cells could be due to differences between VHa and VHn B cells in BCR density (31, 32), in localization of BCR in lipid rafts (33), or in BCR structure leading to differential stimulation and subsequent proliferation. We found no difference in surface IgM levels between VHa and VHn B cells, suggesting that differences in BCR density in VHa and VHn B cells do not contribute to the differential stimulation. Although we have not studied the localization of VHa and VHn BCR in lipid rafts, we suggest that VHa and VHn B cells are differentially stimulated by bacteria because of structural differences between the VH regions of VHa and VHn BCR. Differential stimulation of VHa and VHn B cells by bacteria will be investigated in future studies.
When we compared amino acid sequences encoded by VHa and VHn gene segments, we found many differences in FR1 and FR3. These differences include VHa2 allotype-associated amino acids, which Pospisil et al. (16, 34) proposed may interact with a ligand, leading to expansion of VHa2 B cells. However, because the VHa2 allotype-associated amino acids are not present in allelically encoded VHa1 and VHa3 allotypes (10), and because VHa1 and VHa3 B cells in a1/a1 and a3/a3 rabbits, respectively, also proliferate in GALT, we suggest that the a2 allotype-associated amino acids are not critical for preferential expansion of VHa B cells. Instead, we suggest that the nonallotype-associated amino acids present in VHa molecules, but absent in VHn molecules, are responsible for preferential expansion of VHa B cells.
We examined the amino acid sequences encoded by VHa and VHn gene segments and found six positions in FR1 and FR3 (3, 19, 21, 23, 78, 82A) in which the same amino acids were encoded by all six VH gene segments known to encode VHa molecules, but not by the three VH gene segments known to encode VHn molecules (Fig. 7 a). In addition, we found that, at positions 79 and 82 (FR3), the same amino acids were encoded by five out of six VHa gene segments, but not by the VHn gene segments (Fig. 7 a). If selective expansion of VHa B cells results from interaction of a ligand with VH molecules, the contacting amino acids are likely to be present on the exterior surface of the VH region. By three-dimensional modeling, we found that of these eight amino acids, five (19, 21, 23, 79, 82A) are clustered on the external face of the VH domain with their side chains exposed for potential interaction with a ligand (Fig. 7 b). Two out of the eight amino acids (78 and 82) are nonpolar and, thus, their side chains are not likely to be exposed to solvent. Another conserved amino acid (position 3) is located at a flexible region, making it difficult to predict whether this amino acid will participate in a ligand interaction. We propose that the five amino acids (19, 21, 23, 79, 82A) clustered on the exterior face of the VHa molecules are part of a binding site for a bacterial ligand or a bacterially induced GALT-derived ligand. Closer examination of the putative binding site reveals two additional amino acids (at positions 77 and 81) that may contribute to ligand binding, even though they are present in both VHa and VHn molecules. We propose that a combination of seven VH amino acids at positions 19, 21, 23, 77, 79, 81, and 82A constitutes a ligand binding site and, furthermore, that the ligand interacts more strongly with VHa than with VHn molecules, leading to the differential stimulation and subsequent expansion of VHa B cells.
The putative VH ligand binding site is on the exterior surface of the VH region, similar to the VH binding site of Staphylococcus aureus protein A in human VH Ig molecules (35). Protein A binds to and preferentially stimulates B cells that use VH gene segments of the VH3 family (36). Similarly, we think that a putative bacterial B cell superantigen (37) or a bacterially induced GALT-derived superantigen (38) preferentially binds to and stimulates VHa B cells. If a B cell superantigen promotes positive selection of VHa B cells in GALT, the interaction between such a B cell superantigen and the rabbit VH region would be expected to stimulate the B cells in an antigen-nonspecific, polyclonal manner. Consistent with this idea, Sehgal et al. (39) found that the nature of somatic mutation in VDJ genes in the appendix of young rabbits differed from that which occurs in response to specific antigens in the spleen. Furthermore, Casola et al. (40) demonstrated that anti-HEL transgenic mice had normal-sized Peyer's patches, indicating that B cell expansion in GALT is specific-antigen independent. However, we cannot rule out the possibility that the microflora stimulate B cells in a non-BCR–dependent manner, rather than through interaction with the VH region (40).
Using IgH-transgenic mice, it has been shown that peritoneal B-1 cells undergo antigen-specific B cell–positive selection (23). Evidence for positive selection of conventional B cells (B-2), whether dependent or independent of specific antigen, is more circumstantial (24). Here, we demonstrated in a nontransgenic model that B cells can be positively selected in the GALT during generation of the primary B cell repertoire, likely in an antigen-independent manner (37, 39). Furthermore, this occurs as a result of interactions between GALT and the intestinal microflora. These data demonstrate the potential for commensal intestinal microflora to shape the B cell repertoire. The extent to which commensal microflora play a role in modifying the B cell repertoire in other species remains to be elucidated.
Materials and Methods
Rabbits and antiallotype antibodies
Ali/ali rabbits (designated Alicia; reference 1), which are homozygous for the b5 -chain allotype (b5/b5), were maintained in the Comparative Medicine Facility at Loyola University Chicago, Maywood, IL. All experiments were performed following the guidelines of the Loyola University Chicago Institutional Animal Care and Use Committee. The anti-b4 and anti-b5 anti– chain allotype antisera were as described previously (41).
Anti-VHn antibody directed against VHx and VHy allotypes was produced by immunizing a homozygous a1x–y– (IgH haplotype A/A) rabbit (L76-3) with IgG from a homozygous a2-suppressed a2x32y33 rabbit (42). Ig fractions of the anti-VHn and anti-VHa2 antisera (41) obtained by precipitation with 40% saturated ammonium sulfate were biotinylated for use in immunofluorescence analysis and in Ca2+ mobilization assays. By immunofluorescence, the anti-VHn antibody reacted with <5% of peripheral B cells in adult homozygous a2x32y33 rabbits, as expected (unpublished data).
To confirm that the anti-VHn allotype antibodies reacted with VHx and VHy Ig, we analyzed PCR-amplified VDJ genes from FACS-sorted splenic VHn B cells from Alicia rabbits, using a 5' conserved VH leader primer and a 3' primer specific for JH. Nearly all of the VDJ genes (32 out of 34) encoded amino acids characteristic of the VHn molecules encoded by the VHx and VHy gene segments (references 10, 13 and unpublished data). We also analyzed 12 VDJ genes PCR-amplified from splenic B cells that did not react with anti-VHn antibodies and found that, as expected, all 12 genes encoded amino acids characteristic of those encoded by the VHa gene segments VH4, VH7, and VH9 (10, 12).
Immunofluorescence and flow cytometry
106 PBLs were prepared from buffy coat and stained with biotinylated rabbit anti-VHn or biotinylated rabbit anti-VHa2 allotype antibodies followed by streptavidin-PE as a secondary reagent (Molecular Probes). CD4+ T cells were stained with FITC-conjugated anti-CD4 mAb (clone KEN4; reference 43). B cells were detected using biotinylated affinity-purified goat anti-IgL chain antibodies and streptavidin-PE or FITC-conjugated anti-IgM mAb (clone 367; reference 3). Cells within the side- and forward-scatter lymphocyte gate were analyzed using a FACSCalibur flow cytometer (BD Biosciences) in the FACS core facility at Loyola University Chicago.
PCR analysis to determine rearrangement status of the IgH locus
Single VHa B cells were FACS sorted into 96-well V-bottom plates containing 1x lysis buffer as described previously (30). VDJ genes were PCR amplified using nested primers as follows: the 5' primers were 5'-T[G/C]-GATAT[T/G]AAGGG[T/C]ACACA-3' (sense-outside primer) and 5'-CATAAAAATTCA[T/C]ATGATC-3' (sense-inside primer), taken from conserved sequences 5' of VH promoter regions; the 3' primers were 5'-AGTTGAGTAGGAGAGAGA-3' (antisense-outside primer) and 5'-GAGTTGGCAAGGACTCAC-3' (antisense-inside primer), taken from conserved sequences 3' of JH4 (JH4 is used in 80–90% of VDJ gene rearrangements) and JH2. To determine whether rearrangements in the JH region had occurred, nested PCR amplification was performed by using the 5' primers 5'-TGAGTGCTGTTGGACTGGCT-3' (sense-outside primer) and 5'-CAGAGCTGGAGCTGTGCTAT-3' (sense-inside primer), taken from a region 5' of the JH locus; the antisense primers were the same as those used for VDJ gene rearrangements.
Development of rabbits with a LigApx
The LigApx rabbits were developed as described previously (9). In brief, we removed the sacculus rotundus from newborn rabbits and ligated the lumen of the appendix to prevent bacterial colonization. The vasculature to the appendix was left intact. Peyer's patches were removed at 4 wk of age, when they became macroscopically visible.
Ca2+ mobilization
PBLs isolated with LSMR (ICN Biomedicals) were stained with anti–rabbit T cell mAb (clone KEN5; reference 43) and with biotinylated anti-VHn or anti-VHa allotype antibodies. Secondary reagents were biotinylated Fab goat anti–mouse IgG (Jackson ImmunoResearch Laboratories) and streptavidin-APC (BD Biosciences). The stained cells were suspended in phenol red-free HBSS containing Ca2+ and Mg2+ (GIBCO BRL) and were incubated with rotation for 45 min at room temperature in 10 μM Fura-red, 5 μM Fluo-3 (prepared as 1 mM stocks in 100% DMSO; Molecular Probes), and 2.8 μl 20% pluronic F-127 (Molecular Probes). VHa or VHn B cells were electronically gated as follows: VHn B cells were those cells in the lymphocyte gate that did not react with anti-VHa or anti–T cell antibodies, and the VHa B cells were cells that did not react with anti-VHn or anti–T cell antibodies. The electronically gated VHn and VHa B cells were FACS sorted and, upon reanalysis by FACSCalibur, were shown to be at least 90% pure. The calcium flux of the VHn and VHa B cells in response to anti-b4 and anti-b5 -chain allotype antisera was measured essentially as described previously (47). The fluorescence of Fluo-3 and Fura-red was measured over time, in a linear format. The baseline was determined from data collected 30 s before the addition of antiallotype antibody. The ratio of Fluo-3 to Fura-red and the corresponding mean intracellular calcium ([Ca2+]i) levels were calculated and analyzed using FlowJo software (Tree Star, Inc.).
Three-dimensional modeling of rabbit VH domain
The crystal structure of a Fab fragment of a human IgM antibody-encoding IgM rheumatoid factor (VH3-30/1.9III; reference 35) was retrieved from the Protein Data Bank (http://www.rcsb.org/pdb) and used as a modeling template for the rabbit VH region. Modeling was performed using DeepView/Swiss-PdbViewer v3.7 (http://www.expasy.org/spdbv), and images were rendered using POV-Ray for Windows v3.5 (http://www.povray.org).
Acknowledgments
We thank Dr. A. Edmunson for help with the three-dimensional analysis of rabbit VH regions and identification of a potential ligand binding site.
This work was supported by National Institutes of Health grant no. AI150260.
References
Knight, K.L., and R.S. Becker. 1990. Molecular basis of the allelic inheritance of rabbit immunoglobulin VH allotypes: implications for the generation of antibody diversity. Cell. 60:963–970.
Weinstein, P.D., A.O. Anderson, and R.G. Mage. 1994. Rabbit IgH sequences in appendix germinal centers: VH diversification by gene conversion-like and hypermutation mechanisms. Immunity. 1:647–659.
Vajdy, M., P. Sethupathi, and K.L. Knight. 1998. Dependence of antibody somatic diversification on gut-associated lymphoid tissue in rabbits. J. Immunol. 160:2725–2729.
Reynaud, C.A., V. Anquez, H. Grimal, and J.-C. Weill. 1987. A hyperconversion mechanism generates the chicken light chain preimmune repertoire. Cell. 48:379–388.
Thompson, C.B., and P.E. Neiman. 1987. Somatic diversification of the chicken immunoglobulin light chain gene is limited to the rearranged variable gene segment. Cell. 48:369–378.
Reynaud, C.A., C.R. Mackay, R.G. Muller, and J.-C. Weill. 1991. Somatic generation of diversity in a mammalian primary lymphoid organ: the sheep ileal Peyer's patches. Cell. 64:995–1005.
Jenne, C.N., L.J. Kennedy, P. McCullagh, and J.D. Reynolds. 2003. A new model of sheep Ig diversification: shifting the emphasis toward combinatorial mechanisms and away from hypermutation. J. Immunol. 170:3739–3750.
Tunyaplin, C., and K.L. Knight. 1995. Fetal VDJ gene repertoire in rabbit: evidence for preferential rearrangement of VH1. Eur. J. Immunol. 25:2783–2787
Lanning, D., P. Sethupathi, K.-J. Rhee, S.-K. Zhai, and K.L. Knight. 2000. Intestinal microflora and diversification of the rabbit antibody repertoire. J. Immunol. 165:2012–2019.(Ki-Jong Rhee, Paul J. Jas)
Abbreviations used: BCR, B cell receptor; FR, framework region; GALT, gut-associated lymphoid tissue; LigApx, ligated appendix.
K.-J. Rhee and P.J. Jasper contributed equally to this work.
Vertebrates have developed two general strategies for generating a diverse primary B cell repertoire. In humans and mice, the primary B cell repertoire is generated by rearrangement of multiple V, D, and J gene segments in the bone marrow throughout the life of the animal. In other species, such as the rabbit (1–3), chicken (4, 5), and sheep (6, 7), this repertoire initially develops by rearrangement of a limited number of V genes in primary lymphoid tissue and further diversifies in gut-associated lymphoid tissues (GALTs). In rabbits, the D proximal VH gene, VH1, is preferentially rearranged during B cell development in the fetal liver and bone marrow (8). The VDJ genes undergo somatic diversification via somatic hypermutation and gene conversion in GALT in response to intestinal microflora (9). In the absence of appropriate intestinal microflora, GALT develops poorly, and both the number of B cells and the diversification of VH genes are greatly inhibited (9).
Most (80–90%) rabbit serum Ig molecules express VHa allotypic markers that are encoded by the predominantly rearranged gene, VH1 (1). The following three alleles of VH1 are found in laboratory rabbits: VH1-a1, VH1-a2, and VH1-a3; they encode the VHa1, VHa2, and VHa3 allotypes, respectively. These VHa allotypes differ in amino acid residues in framework region (FR)1 and FR3 (10). 10–20% of serum Ig does not react with anti-VHa1, anti-VHa2, or anti-VHa3 allotypic antibodies and is referred to as VHn (VHa-negative) Ig.
Kelus and Weiss (11) identified rabbits with a variant VHa2 allotype-encoding allele, ali, which has a 10-kb deletion of DNA encompassing VH1 (Fig. 1 and reference 1). In contrast with wild-type rabbits, nearly all Ig in young ali/ali rabbits (designated Alicia) is VHn. VHn Ig is encoded predominantly by VHx, VHy, and VHz (12, 13), which reside >50 kb upstream of VH1 (1). In adult Alicia rabbits, high levels of serum Ig with the VHa (a2) allotype are found (11). This increase in VHa Ig is a result of increased numbers of VHa B cells that use VH4, VH7, and VH9, gene segments that encode several of the VHa (a2) allotype-associated amino acids (14, 15). Pospisil et al. (16) found that, in the appendix, more VHa B cells were proliferating and fewer were dying compared with VHn B cells. The molecular basis underlying the repertoire shift from VHn to VHa is unknown and is the subject of the current paper.
A shift in the B cell repertoire could arise from VH gene replacement or from secondary Ig gene rearrangements on the unexpressed Ig allele. Although it is generally believed that these events occur primarily in the bone marrow (17–19), there is evidence that VH gene replacement and secondary Ig gene rearrangement occur in peripheral tissues (20, 21). Another possible explanation for the B cell repertoire shift is that VHa B cells are positively selected in the periphery. Positive selection of B cells in the periphery has been demonstrated in several transgenic mouse models (22–26).
In rabbit, intestinal microflora interacts with GALT to promote development of follicles containing proliferating B cells and to generate the primary B cell repertoire (9). We hypothesized that, in GALT of Alicia rabbits, interactions between GALT and the intestinal microflora also promote the increased proliferation of VHa B cells compared with VHn B cells and lead to the repertoire shift from VHn to VHa B cells. To test this possibility, we surgically disrupted the GALT–bacterial interaction in Alicia rabbits and tested whether the repertoire shift from VHn to VHa B cells was abrogated.
Results
Kinetics of the B cell repertoire shift
The repertoire shift in Alicia rabbits, from the predominant expression of VHn allotype early in life to the predominant expression of VHa allotype later in life, was originally shown by Kelus and Weiss (11), who analyzed Ig allotypes in serum. Pospisil et al. (16) showed that a similar shift occurred in B cells in the appendix. By using antibodies to both VHn and VHa allotypes, we found that, in 9-wk-old rabbits, VHa B cells represented 35–50% of the B cells in spleen, mesenteric lymph nodes, appendix, and PBLs (Fig. 2). We analyzed cells of various tissues from newborn to 2-yr-old Alicia rabbits to follow the appearance and disappearance of VHa and VHn B cells, respectively, throughout life. We found that, although 10–25% of B cells at birth are VHa, at 3 wk of age essentially all B cells (95%) in spleen, appendix, and PBLs were VHn (Fig. 3 and not depicted). Subsequently, the percentage of VHn cells steadily declined, so that by 2 yr of age, <20% of B cells were VHn and >75% were VHa. These data demonstrate that VHa B cells accumulate throughout life, with a sharp increase between 4 and 10 wk of age. The VHa B cells accumulate faster in the appendix than in spleen, suggesting that the B cell repertoire shift from VHn to VHa B cells may occur primarily in GALT.
B cell repertoire shift and GALT
Because GALT development and somatic diversification of Ig genes both require interaction between GALT and intestinal microflora (9, 27), we hypothesized that the repertoire shift in Alicia rabbits also requires this interaction. To investigate this possibility, we generated ligated appendix (LigApx) rabbits by surgically removing the Peyer's patches and the sacculus rotundus and ligating the lumen of the appendix to prevent bacterial colonization (9). If interactions between GALT and intestinal microflora are required for the repertoire shift from VHn to VHa B cells, we expected that the peripheral blood B cells in LigApx rabbits would be predominantly VHn. In each of three 12-wk-old LigApx Alicia rabbits (94S, 353X2, 353X4), we found that the percentage of B cells was approximately eightfold less than in unmanipulated rabbits of that age and that almost all B cells (90–96%) were VHn (Fig. 4). As expected, the percentage of VHn B cells in unmanipulated Alicia rabbits was 50%. We examined one of the LigApx Alicia rabbits (94S) at 8 mo of age and found that >90% of the B cells were still VHn, showing that the B cells remained predominantly VHn for many months. These results indicate that, without interactions between GALT and intestinal flora, the shift from VHn to VHa did not occur.
To confirm that the repertoire shift from VHn to VHa B cells was abrogated in LigApx rabbits, we examined the nucleotide sequences of VDJ genes cloned from peripheral blood of 12-wk-old LigApx Alicia rabbits. We expected that the VH genes used in the VDJ gene rearrangements would be primarily genes that encode VHn molecules rather than VHa molecules. From one LigApx Alicia rabbit (94S), shown in Fig. 4, and from three additional LigApx rabbits (32P2, 144T, 199T1) for which flow cytometry data are not available, we analyzed a total of 80 VDJ gene sequences. As predicted, most (76%) of the VDJ genes used VHn gene segments (Table I), whereas almost none (4%) of the VDJ genes from control (unmanipulated) Alicia rabbits of the same age used VHn gene segments. We think the PCR analysis underestimated the expression of VHn genes because, by FACS analysis, 50% of the peripheral B cells from 12-wk-old control Alicia rabbits were VHn, whereas only 4% of the PCR-amplified VDJ genes were VHn.
To determine whether the low percentage of VHn-encoding genes (Table I) resulted from preferential amplification of VHa cDNA, we conducted two independent experiments in which VDJ genes were PCR amplified from cDNA prepared from a pool of cells containing equivalent numbers of FACS-sorted VHa and VHn B cells from peripheral blood. In the two experiments, 67% (14 out of 21) and 81% (17 out of 21) of PCR-amplified VDJ genes used VHa gene segments. The reduced number of VHn PCR products was also observed with another 5' VH primer, VHldr (5'-GGCTTCTCCTGGTCGCTG-3'), which anneals to a different target site. The preferential amplification of VHa cDNA with independent primers suggests that, even though VHa and VHn B cells appear to express equivalent amounts of surface IgM (Fig. 4), VHa B cells might produce higher levels of IgM mRNA, possibly as a result of their stimulation in GALT (16)
Although we do not understand the molecular basis for the PCR skewing toward VHa-encoding VDJ genes, the data confirm the FACS analysis, which showed that most of the B cells of LigApx Alicia rabbits were VHn instead of VHa. We conclude that the repertoire shift from VHn to VHa B cells required interactions between GALT and the microflora and that expansion of VHa B cells requires such interactions.
Rearrangement status of IgH alleles in VHa B cell
The repertoire shift from VHn to VHa B cells in the periphery could occur by replacement of VHn-using VDJ genes with VHa gene segments (20), by rearrangement of a VHa-encoding VHa gene segment on the second IgH allele (17, 21), or by selective expansion of VHa B cells (16, 24). We think that VH gene replacement is unlikely to explain the repertoire shift because the VHn genes (VHx, VHy, and VHz) used in VDJ gene rearrangements in VHn B cells reside upstream of the VHa genes (VH4, VH7, and VH9) used in VDJ gene rearrangements in peripheral VHa B cells of Alicia rabbits (15). Accordingly, the rearrangement of VHn genes during VDJ gene rearrangements would likely result in deletion of the VHa genes (Fig. 1).
If the repertoire shift from VHn to VHa B cells is caused by gene rearrangements of VHa gene segments on the second IgH allele in VHn B cells, we expected to find VDJ gene rearrangements on both IgH alleles in VHa B cells. To test this possibility, we sorted VHa B cells from an adult Alicia rabbit and assessed the status of VDJ gene rearrangements by single cell PCR. We used PCR primers that would detect rearranged VDJ genes and germline JH genes (Fig. 5 a). Of 26 single cells from which a rearranged VDJ PCR product was obtained, all but one had a product of the expected size for an unrearranged (second) IgH allele (Fig. 5 b). This result showed that essentially all VHa B cells rearranged only one IgH allele, indicating that the B cell repertoire shift from VHn to VHa B cells in Alicia rabbits is not due to secondary IgH gene rearrangements on the other allele. Instead, we propose that the B cell repertoire shift occurs through positive selection due to preferential expansion of VHa B cells.
B cell receptor (BCR) signaling in VHa and VHn B cells
One possible explanation for the preferential expansion of VHa B cells and the concomitant decrease in VHn B cells in Alicia rabbits is that VHa B cells are more responsive to BCR stimulation than VHn B cells. To test this possibility, we assessed the release of intracellular calcium after BCR cross-linking on VHa and VHn B cells from 12-wk-old Alicia rabbits. The Alicia rabbits had the b5 chain allotype; therefore, we incubated PBLs with anti-b5 antibody and measured the release of intracellular calcium, as described in Materials and Methods. We found that the VHn B cells responded to anti-b5 antibody as well as the VHa B cells when the differences in baseline stimulation were taken into account (Fig. 6). Anti-b4 chain allotype antibody served as a negative control. Although we cannot explain the different baseline stimulation of VHa and VHn B cells, we conclude that the inherent signaling capacity of VHa and VHn B cells is similar and, therefore, does not explain the selective expansion of VHn B cells.
Discussion
The intestinal microflora are important in regulating many immune functions, including development of GALT (27), induction of oral tolerance (28), and induction of mucosal immunity (29). In rabbits, intestinal microflora are required not only for GALT to develop but also to generate a diverse primary B cell repertoire (9). Previously, we found that surgical disruption of GALT–bacterial interactions prevented GALT development, B cell expansion, and somatic diversification of the B cell repertoire (9). In the current paper, we found that the repertoire shift from VHn to VHa B cells in Alicia rabbits also depends on GALT–bacterial interactions.
At birth, 10–25% of B cells in peripheral tissues of Alicia rabbits were VHa, and these B cells subsequently declined to nearly undetectable levels by week 3. Although these VHa B cells could represent maternal B cells, we think the percentages are much higher than would be expected for maternal B cells. We also do not think these cells are VHn B cells with maternal VHa Ig bound through Fc receptors because, in this case, we would expect all B cells, rather than a subset, to be VHa. Instead, we think the decline in the percentage of VHa B cells may be due to a dramatic increase in VHn B cells from a second wave of B lymphopoiesis in the bone marrow. We recently identified a burst of both pre–B cells and B cells in bone marrow at 3 wk of age and we suggest that in Alicia rabbits, the newly generated B cells may be primarily VHn (30).
The shift from VHn to VHa B cells after 3 wk of age likely occurs in GALT rather than in the bone marrow because the shift requires GALT–bacterial interactions. Therefore, we favor the idea that this shift is due to selective expansion of VHa B cells as proposed by Pospisil et al. (16), who showed that more VHa B cells proliferate and fewer die than VHn B cells in the appendices of Alicia rabbits. We suggest that VHa B cells are preferentially stimulated by interaction with a bacterial ligand or a bacterially induced GALT-derived ligand. Such preferential stimulation of VHa B cells could be due to differences between VHa and VHn B cells in BCR density (31, 32), in localization of BCR in lipid rafts (33), or in BCR structure leading to differential stimulation and subsequent proliferation. We found no difference in surface IgM levels between VHa and VHn B cells, suggesting that differences in BCR density in VHa and VHn B cells do not contribute to the differential stimulation. Although we have not studied the localization of VHa and VHn BCR in lipid rafts, we suggest that VHa and VHn B cells are differentially stimulated by bacteria because of structural differences between the VH regions of VHa and VHn BCR. Differential stimulation of VHa and VHn B cells by bacteria will be investigated in future studies.
When we compared amino acid sequences encoded by VHa and VHn gene segments, we found many differences in FR1 and FR3. These differences include VHa2 allotype-associated amino acids, which Pospisil et al. (16, 34) proposed may interact with a ligand, leading to expansion of VHa2 B cells. However, because the VHa2 allotype-associated amino acids are not present in allelically encoded VHa1 and VHa3 allotypes (10), and because VHa1 and VHa3 B cells in a1/a1 and a3/a3 rabbits, respectively, also proliferate in GALT, we suggest that the a2 allotype-associated amino acids are not critical for preferential expansion of VHa B cells. Instead, we suggest that the nonallotype-associated amino acids present in VHa molecules, but absent in VHn molecules, are responsible for preferential expansion of VHa B cells.
We examined the amino acid sequences encoded by VHa and VHn gene segments and found six positions in FR1 and FR3 (3, 19, 21, 23, 78, 82A) in which the same amino acids were encoded by all six VH gene segments known to encode VHa molecules, but not by the three VH gene segments known to encode VHn molecules (Fig. 7 a). In addition, we found that, at positions 79 and 82 (FR3), the same amino acids were encoded by five out of six VHa gene segments, but not by the VHn gene segments (Fig. 7 a). If selective expansion of VHa B cells results from interaction of a ligand with VH molecules, the contacting amino acids are likely to be present on the exterior surface of the VH region. By three-dimensional modeling, we found that of these eight amino acids, five (19, 21, 23, 79, 82A) are clustered on the external face of the VH domain with their side chains exposed for potential interaction with a ligand (Fig. 7 b). Two out of the eight amino acids (78 and 82) are nonpolar and, thus, their side chains are not likely to be exposed to solvent. Another conserved amino acid (position 3) is located at a flexible region, making it difficult to predict whether this amino acid will participate in a ligand interaction. We propose that the five amino acids (19, 21, 23, 79, 82A) clustered on the exterior face of the VHa molecules are part of a binding site for a bacterial ligand or a bacterially induced GALT-derived ligand. Closer examination of the putative binding site reveals two additional amino acids (at positions 77 and 81) that may contribute to ligand binding, even though they are present in both VHa and VHn molecules. We propose that a combination of seven VH amino acids at positions 19, 21, 23, 77, 79, 81, and 82A constitutes a ligand binding site and, furthermore, that the ligand interacts more strongly with VHa than with VHn molecules, leading to the differential stimulation and subsequent expansion of VHa B cells.
The putative VH ligand binding site is on the exterior surface of the VH region, similar to the VH binding site of Staphylococcus aureus protein A in human VH Ig molecules (35). Protein A binds to and preferentially stimulates B cells that use VH gene segments of the VH3 family (36). Similarly, we think that a putative bacterial B cell superantigen (37) or a bacterially induced GALT-derived superantigen (38) preferentially binds to and stimulates VHa B cells. If a B cell superantigen promotes positive selection of VHa B cells in GALT, the interaction between such a B cell superantigen and the rabbit VH region would be expected to stimulate the B cells in an antigen-nonspecific, polyclonal manner. Consistent with this idea, Sehgal et al. (39) found that the nature of somatic mutation in VDJ genes in the appendix of young rabbits differed from that which occurs in response to specific antigens in the spleen. Furthermore, Casola et al. (40) demonstrated that anti-HEL transgenic mice had normal-sized Peyer's patches, indicating that B cell expansion in GALT is specific-antigen independent. However, we cannot rule out the possibility that the microflora stimulate B cells in a non-BCR–dependent manner, rather than through interaction with the VH region (40).
Using IgH-transgenic mice, it has been shown that peritoneal B-1 cells undergo antigen-specific B cell–positive selection (23). Evidence for positive selection of conventional B cells (B-2), whether dependent or independent of specific antigen, is more circumstantial (24). Here, we demonstrated in a nontransgenic model that B cells can be positively selected in the GALT during generation of the primary B cell repertoire, likely in an antigen-independent manner (37, 39). Furthermore, this occurs as a result of interactions between GALT and the intestinal microflora. These data demonstrate the potential for commensal intestinal microflora to shape the B cell repertoire. The extent to which commensal microflora play a role in modifying the B cell repertoire in other species remains to be elucidated.
Materials and Methods
Rabbits and antiallotype antibodies
Ali/ali rabbits (designated Alicia; reference 1), which are homozygous for the b5 -chain allotype (b5/b5), were maintained in the Comparative Medicine Facility at Loyola University Chicago, Maywood, IL. All experiments were performed following the guidelines of the Loyola University Chicago Institutional Animal Care and Use Committee. The anti-b4 and anti-b5 anti– chain allotype antisera were as described previously (41).
Anti-VHn antibody directed against VHx and VHy allotypes was produced by immunizing a homozygous a1x–y– (IgH haplotype A/A) rabbit (L76-3) with IgG from a homozygous a2-suppressed a2x32y33 rabbit (42). Ig fractions of the anti-VHn and anti-VHa2 antisera (41) obtained by precipitation with 40% saturated ammonium sulfate were biotinylated for use in immunofluorescence analysis and in Ca2+ mobilization assays. By immunofluorescence, the anti-VHn antibody reacted with <5% of peripheral B cells in adult homozygous a2x32y33 rabbits, as expected (unpublished data).
To confirm that the anti-VHn allotype antibodies reacted with VHx and VHy Ig, we analyzed PCR-amplified VDJ genes from FACS-sorted splenic VHn B cells from Alicia rabbits, using a 5' conserved VH leader primer and a 3' primer specific for JH. Nearly all of the VDJ genes (32 out of 34) encoded amino acids characteristic of the VHn molecules encoded by the VHx and VHy gene segments (references 10, 13 and unpublished data). We also analyzed 12 VDJ genes PCR-amplified from splenic B cells that did not react with anti-VHn antibodies and found that, as expected, all 12 genes encoded amino acids characteristic of those encoded by the VHa gene segments VH4, VH7, and VH9 (10, 12).
Immunofluorescence and flow cytometry
106 PBLs were prepared from buffy coat and stained with biotinylated rabbit anti-VHn or biotinylated rabbit anti-VHa2 allotype antibodies followed by streptavidin-PE as a secondary reagent (Molecular Probes). CD4+ T cells were stained with FITC-conjugated anti-CD4 mAb (clone KEN4; reference 43). B cells were detected using biotinylated affinity-purified goat anti-IgL chain antibodies and streptavidin-PE or FITC-conjugated anti-IgM mAb (clone 367; reference 3). Cells within the side- and forward-scatter lymphocyte gate were analyzed using a FACSCalibur flow cytometer (BD Biosciences) in the FACS core facility at Loyola University Chicago.
PCR analysis to determine rearrangement status of the IgH locus
Single VHa B cells were FACS sorted into 96-well V-bottom plates containing 1x lysis buffer as described previously (30). VDJ genes were PCR amplified using nested primers as follows: the 5' primers were 5'-T[G/C]-GATAT[T/G]AAGGG[T/C]ACACA-3' (sense-outside primer) and 5'-CATAAAAATTCA[T/C]ATGATC-3' (sense-inside primer), taken from conserved sequences 5' of VH promoter regions; the 3' primers were 5'-AGTTGAGTAGGAGAGAGA-3' (antisense-outside primer) and 5'-GAGTTGGCAAGGACTCAC-3' (antisense-inside primer), taken from conserved sequences 3' of JH4 (JH4 is used in 80–90% of VDJ gene rearrangements) and JH2. To determine whether rearrangements in the JH region had occurred, nested PCR amplification was performed by using the 5' primers 5'-TGAGTGCTGTTGGACTGGCT-3' (sense-outside primer) and 5'-CAGAGCTGGAGCTGTGCTAT-3' (sense-inside primer), taken from a region 5' of the JH locus; the antisense primers were the same as those used for VDJ gene rearrangements.
Development of rabbits with a LigApx
The LigApx rabbits were developed as described previously (9). In brief, we removed the sacculus rotundus from newborn rabbits and ligated the lumen of the appendix to prevent bacterial colonization. The vasculature to the appendix was left intact. Peyer's patches were removed at 4 wk of age, when they became macroscopically visible.
Ca2+ mobilization
PBLs isolated with LSMR (ICN Biomedicals) were stained with anti–rabbit T cell mAb (clone KEN5; reference 43) and with biotinylated anti-VHn or anti-VHa allotype antibodies. Secondary reagents were biotinylated Fab goat anti–mouse IgG (Jackson ImmunoResearch Laboratories) and streptavidin-APC (BD Biosciences). The stained cells were suspended in phenol red-free HBSS containing Ca2+ and Mg2+ (GIBCO BRL) and were incubated with rotation for 45 min at room temperature in 10 μM Fura-red, 5 μM Fluo-3 (prepared as 1 mM stocks in 100% DMSO; Molecular Probes), and 2.8 μl 20% pluronic F-127 (Molecular Probes). VHa or VHn B cells were electronically gated as follows: VHn B cells were those cells in the lymphocyte gate that did not react with anti-VHa or anti–T cell antibodies, and the VHa B cells were cells that did not react with anti-VHn or anti–T cell antibodies. The electronically gated VHn and VHa B cells were FACS sorted and, upon reanalysis by FACSCalibur, were shown to be at least 90% pure. The calcium flux of the VHn and VHa B cells in response to anti-b4 and anti-b5 -chain allotype antisera was measured essentially as described previously (47). The fluorescence of Fluo-3 and Fura-red was measured over time, in a linear format. The baseline was determined from data collected 30 s before the addition of antiallotype antibody. The ratio of Fluo-3 to Fura-red and the corresponding mean intracellular calcium ([Ca2+]i) levels were calculated and analyzed using FlowJo software (Tree Star, Inc.).
Three-dimensional modeling of rabbit VH domain
The crystal structure of a Fab fragment of a human IgM antibody-encoding IgM rheumatoid factor (VH3-30/1.9III; reference 35) was retrieved from the Protein Data Bank (http://www.rcsb.org/pdb) and used as a modeling template for the rabbit VH region. Modeling was performed using DeepView/Swiss-PdbViewer v3.7 (http://www.expasy.org/spdbv), and images were rendered using POV-Ray for Windows v3.5 (http://www.povray.org).
Acknowledgments
We thank Dr. A. Edmunson for help with the three-dimensional analysis of rabbit VH regions and identification of a potential ligand binding site.
This work was supported by National Institutes of Health grant no. AI150260.
References
Knight, K.L., and R.S. Becker. 1990. Molecular basis of the allelic inheritance of rabbit immunoglobulin VH allotypes: implications for the generation of antibody diversity. Cell. 60:963–970.
Weinstein, P.D., A.O. Anderson, and R.G. Mage. 1994. Rabbit IgH sequences in appendix germinal centers: VH diversification by gene conversion-like and hypermutation mechanisms. Immunity. 1:647–659.
Vajdy, M., P. Sethupathi, and K.L. Knight. 1998. Dependence of antibody somatic diversification on gut-associated lymphoid tissue in rabbits. J. Immunol. 160:2725–2729.
Reynaud, C.A., V. Anquez, H. Grimal, and J.-C. Weill. 1987. A hyperconversion mechanism generates the chicken light chain preimmune repertoire. Cell. 48:379–388.
Thompson, C.B., and P.E. Neiman. 1987. Somatic diversification of the chicken immunoglobulin light chain gene is limited to the rearranged variable gene segment. Cell. 48:369–378.
Reynaud, C.A., C.R. Mackay, R.G. Muller, and J.-C. Weill. 1991. Somatic generation of diversity in a mammalian primary lymphoid organ: the sheep ileal Peyer's patches. Cell. 64:995–1005.
Jenne, C.N., L.J. Kennedy, P. McCullagh, and J.D. Reynolds. 2003. A new model of sheep Ig diversification: shifting the emphasis toward combinatorial mechanisms and away from hypermutation. J. Immunol. 170:3739–3750.
Tunyaplin, C., and K.L. Knight. 1995. Fetal VDJ gene repertoire in rabbit: evidence for preferential rearrangement of VH1. Eur. J. Immunol. 25:2783–2787
Lanning, D., P. Sethupathi, K.-J. Rhee, S.-K. Zhai, and K.L. Knight. 2000. Intestinal microflora and diversification of the rabbit antibody repertoire. J. Immunol. 165:2012–2019.(Ki-Jong Rhee, Paul J. Jas)