| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Chains in Individual Naive and Memory B Cells1
Department of Internal Medicine and Harold C. Simmons Arthritis Research Center, University of Texas Southwestern Medical Center at Dallas, Dallas TX 75235
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
chains from 144 individual human CD19+/IgM+ B
cells. Pairing of specific VH or V families was not
observed, nor was the length or the amino acid composition of the CDR3s
of VH and V chains in individual B cells similar.
Comparison of VH and V genes in B cells in which one or
both contained evidence of somatic hypermutation with those with no
mutations revealed a significant decrease in the mean length of the
VH CDR3. Moreover, there was a significant correlation
between the frequencies of mutations in VH and V gene
pairs in individual B cells. These results indicate that Ag-mediated
selection as opposed to VHDJH recombination or
subsequent Ig chain pairing tended to approximate the CDR3 lengths and
the frequency of mutations of VH and V in individual B
cells. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
One issue that has not been resolved is whether pairing of
VH and V molecules in the expressed repertoire is random
and unrestricted. It remains possible that Ig expression is favored
when specific VH and V pair. Such restrictions could
limit the array of Ig molecules expressed by mature B cells. This issue
is especially meaningful as preferential pairing of VH and
VL has been suggested to occur in certain pathogenic
autoantibodies (4).
To address this issue, we used a single cell PCR technique (5) to
analyze both productive VH and V rearrangements from
genomic DNA of 144 individual IgM+ B cells obtained from
the peripheral blood of two healthy donors. The antigenic reactivity of
the B cells was not used in their selection, because the goal was to
analyze the influences of possible pairing requirements on the shape of
the Ig repertoire in general. The data indicate that there are no
identifiable limitations on the expressed Ig repertoire imposed by
pairing requirements of VH and VL chains.
However, Ag-mediated selection seems to approximate the CDR3 length and
frequency of mutations of VH and V chains in individual
B cells.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Peripheral blood was obtained from two healthy male donors (donor 1 was a 26-yr-old Hispanic and donor 2 was a 45-yr-old Caucasian) as described (6, 7). B cells were enriched using a commercially available kit (CEPRATE LC Kit, Cellpro, Bothell, WA) that permits selection of CD19+ cells. Cells were stained thereafter with PE-labeled anti-human CD19 Ab (Sigma Chemical, St. Louis, MO), FITC-labeled anti-human IgM Ab (PharMingen, San Diego, CA), a biotinylated anti-human CD5 Ab (Becton Dickinson, San Jose, CA), and RED613-labeled streptavidin (Life Technologies, Gaithersburg, MD). An individual CD19+/IgM+/CD5+ or CD5- B cell was sorted into each well of 96-well PCR plates (Robbins Scientific, Sunnyvale, CA) assembled on a microAmp base (Perkin-Elmer, Norwalk, CT) using a FACStarPlus flow cytometer outfitted with an automatic cell deposition unit (Becton Dickinson). Each well contained 5 µl of an alkaline lysing solution (200 mM KOH/50 mM DTT).
Single cell PCR
Ig genes from individual peripheral blood IgM+
B cells were obtained as published (6, 7). Briefly, single
CD19+IgM+ CD5+ or CD5-
B cells were sorted into lysis solution, and genomic DNA was amplified
nonspecifically. Thereafter, aliquots were subjected to nested PCR
using primers specific for rearranged VH or V genes.
Positive bands were isolated and directly sequenced.
From a total of 736 sorted B cells (184 B cells from donor 1 and 552 B
cells from donor 2) 350 productive VHDJH
rearrangements and 321 productive VJ rearrangements were
obtained. Both productively rearranged VH and V chains
were identified from 144 B cells, 30 from donor 1 and 114 from donor 2.
Within this population, 7 B cells also contained a nonproductively
rearranged VH chain, whereas 32 contained a nonproductively
rearranged V chain and 3 had both a nonproductively rearranged
VH and V chain.
Sequence analysis
Sequences were analyzed using the V Base Sequence directory (8)
to identify the respective germline gene and determine mutations. For
this analysis, nucleotide changes in the V genes as a result of
insertion or deletion of nucleotides were not counted. These events
were found only in the VH gene segment and were
located in the CDR14 (one
VH4 gene with insertion) or CDR2 (two VH3 genes
with insertion and one VH3 gene with deletion). In every
case, three consecutive nucleotides were inserted or deleted. The
single cell PCR introduces <0.5 and 0.3 base pair changes per
VH and V, respectively (6, 7). All sequences used in
this study can be found in the GeneBank, EMBL and DDBJ Nucleotide
Sequence Database under accession numbers Z80363–Z80770 and
Z85397–Z85948.
Statistical analysis
Overall statistical significance between observed and expected
frequencies was calculated with the 
2
goodness-of-fit statistic. The 2 test was used to
analyze differences between the unmutated and mutated population. If
significant, each of the single-df 2 values was examined
for significant contribution to the total. The p value for
the significance of these single-df 2 values was then
adjusted for the accumulation of errors related to multiple testing
according to the Bonferroni method. To determine significant
differences in the distribution of mutations in VH and V
genes, the 2 test was used. p values <0.05
were assumed to be significant. The correlation coefficient,
r, was calculated as described (9), and the t
test was used to evaluate its significance.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
families
To evaluate whether certain VH and V families
paired preferentially, the frequency of occurrence of pairs of
VH and V families within individual B cells was analyzed
and compared with the expected frequency, calculated from the
distribution of each family in the productive repertoire. Within the
heavy and the kappa genes, the distribution of variable gene families
was similar to that reported previously (5, 6, 7). Since there were no
significant differences between the two donors, the data were combined
in the analyses. Overall, no preferential pairing was observed as
determined by the 2 goodness-of-fit test, with the two
largest families (VH3 and V1, respectively) found most
often, followed by VH3/V3 and VH3/V2
(Table I
). A large number of the
IgM+ B cells analyzed appeared to be memory cells in that
their VH and/or V genes contained evidence of somatic
mutation. Because the hypermutation machinery is triggered by antigenic
stimulation (10), these IgM+ B cells presumably had been
stimulated by Ag previously and differentiated into circulating memory
B cells. Previous analysis of the VH repertoire (6) had
shown that the frequency of specific VH families in the
mutated memory population (
98% homology to the respective germline
gene or 
4 mutations per VH gene segment) differed from
that in the naive unmutated B cell population (99% homology).
Therefore, the distribution of VH and V pairs in naive
(99% homology of VH and V gene segments) and memory
(98% homology of VH or V gene segments) B cell
subsets was analyzed separately (Table I). No evidence of preferential
pairing of VH and V families was noted in either naive
or memory B cells. Similar to the distribution previously found in the
entire VH repertoire (6), the frequency of the
VH1 family in the memory B cell subset was diminished,
whereas the VH3 family was increased but neither change was
statistically significant. No evidence of selective pairing of
JH and J elements was found, with the exception of
JH3/J4 that was found infrequently (2 of 144), but more
often than expected (p 0.001).
|
chains
of individual B cells
Not only was VH/V pairing random, but
there was no correlation between the CDR3 length of V and the CDR3
length of VH chains of individual B cells (Fig. 1A). However, there was
a significant decrease in the mean length of the CDR3 of the
VH genes in memory B cells (naive 42.8 ± 12.8 base
pairs, and memory 35.7 ± 10.3 base pairs; p
0.0003). Despite this shortening of the VH CDR3 length to
12 amino acids (aa), there was no correlation between the CDR3s of
VH and V in memory B cells. Finally, there was no
correlation in the frequency of aa with hydrophobic or hydrophilic side
chains when the aa composition of the hypervariable loop 3 of the
VH (H3) and the V chains (L3) was analyzed (aa 96–101
for the VH and aa 91–96 for the V chain; Fig. 1B). A similar lack of correlation was noted when the
entire VH and V CDR3 was compared. Furthermore, this
lack of correlation was similar when the VH and V pairs
from naive or memory B cells were analyzed separately (data not
shown).
|
genes of individual B cells
Of the 144 IgM+ B cells that contained both
productive VH and V rearrangements, 33 (22.9%)
contained no mutations in either Ig chain, whereas 111 (77.1%)
contained nucleotide substitutions in either the VH and/or
the V gene (Table II). Of those B
cells, 61 (42.4%) contained VH or V genes that were
98% homologous to the respective germline gene; i.e., they had 4 or
more mutations per sequence. Within this population 75.4% of B cells
exhibited more extensive mutation of the VH gene, and
26.2% exhibited mutations of the VH gene only. In
aggregate, the VH genes contained 795 mutated nucleotides
within a total of 36,285 base pairs of DNA, whereas the V genes
contained only 389 nucleotide substitutions within 34,497 base pairs.
The overall mutational frequency of VH genes was twice as
great as for V genes (2.2 x 10-2/base pair vs
1.1 x 10-2/base pair, respectively). When the
VH and the V genes that contained mutations were
analyzed separately, the mutational frequencies within mutated genes
were 3.4 x 10-2/base pair for VH genes
and 2.4 x 10-2/base pair for V genes (795
mutations within 23,458 VH base pairs and 389 mutations
within 15,996 V base pairs, respectively). By each of these
analyses, the mutational frequency of VH genes was
significantly greater than that of V genes
(p 0.0001 by 2 test).
|
segment was plotted
as a function of the number of nucleotide substitutions in the
VH gene of the same B cell, a statistically significant
correlation was noted, whether all 144 IgM+ B cells were
analyzed (Fig. 2A) or
when only those memory B cells that contained mutations in both Ig
genes were assessed (Fig. 2B). A much greater
correlation between the number of mutations in VH compared
with V was noted when the CDRs as opposed to the FRs were analyzed.
In the FRs, the correlation was insignificant (Fig. 2D), whereas a significant correlation between the
number of mutations of V CDRs and VH CDRs was noted
(Fig. 2C). Furthermore, there was a significant
correlation between replacement (R) mutations of VH and
V in the CDRs, but not in the FRs (Fig. 3A). However, there was
no correlation between the number of silent (S) mutations of
VH and V (Fig. 3B).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
chains in individual B cells differently.
Although the possibility that specific VH and V gene
products may pair preferentially cannot be excluded, within the group
of IgM+ B cells analyzed, pairing of VH and
V families occurred in a random fashion. Additionally, when only
IgM+ memory B cells that had undergone somatic mutation,
presumably as a result of antigenic stimulation, were analyzed, no
preferential pairing of VH and V families was observed.
Thus, the diversity of expressed Ig molecules that potentially could be
generated from random pairing of VH and V chains does
not appear to be restricted by requirements for specific favored
interactions of particular VH and V families.
No indication was found that the CDR3 lengths of the VH and
the V chains were correlated. Rather, the length of the CDR3 of the
chain was quite restricted to 9.0 ± 0.7 aa (mean ± SD)
as has previously been shown (7), whereas the length of the CDR3 of the
VH was much more variable and tended to be longer than that
of the V partner. The mean CDR3 length of the VH chain
was 13.3 ± 4.1 aa. In only 14 of 144 B cells (9.7%) was the CDR3
length of the VH less than that of the V. Whether the
preference for longer VH CDR3s is related to the inclusion
of the D segment in VH rearrangements or results from
structural constraints cannot be answered yet, but a recent paper has
demonstrated that the VH-VL pairing can
dramatically alter the conformation of VH CDR3 (11).
Exchanging one VL domain for another resulted in main-chain
and side-chain deviations similar to the largest of the conformational
changes induced by Ag binding. Thus, VH chains with longer
CDR3 might have the advantage of being able to pair with several
VL chains since they are more flexible.
There was a significant increase in the number of B cells that had a
shorter VH CDR3 compared with the V CDR3 in the memory B
cell population compared with the naive population (10 of 61 or 16.4%
vs 4 of 83 or 4.8%, p 0.02). Moreover, the mean
CDR3 length of the VH genes of memory B cells was shorter
than that of the VH genes of naive B cells (11.9 ±
3.4 aa vs 14.3 ± 4.3 aa), although the mean CDR3 length of the
V genes was the same in both populations of B cells. Since memory B
cells with mutated Ig chains have had previous antigenic stimulation
(10), this finding suggests that the presence of shorter VH
CDR3s may facilitate more effective Ag binding. No obvious difference
in the memory and naive B cell populations was noted in the aa
composition of VH and the V chains, suggesting that the
length of the CDR3 per se, rather than composition, influences binding
of Ag. Longer CDR3s may interfere with Ag binding by either protruding
out of the Ag binding site or collapsing onto the framework, depending
on the hydrophobicity of the aa (12). The data, therefore, suggest that
Ig molecules with VH and V CDR3s that are more
comparable in length may be more effective at Ag binding. The increased
frequency of B cells expressing heavy and chains with more similar
CDR3s appeared to be accounted for by selective influences solely on
VH. The increased frequency of B cells with shorter
VH CDR3s in the Ag-driven population not only may reflect
selection from the preimmune repertoire but also implies that there may
be Ag-mediated selective pressure on B cells favoring expression of
shorter VH CDR3s.
It has recently been postulated that VH CDR3 with
hydrophobic aa may be deleted from the repertoire (12). The current
data do not support this conclusion, however, as aa with hydrophobic
side chains were more frequent in the H3 than aa with hydrophilic
residues. Moreover, in individual B cells, the H3 as well as the entire
CDR3 of the VH chain was much more hydrophobic than the L3
or the entire CDR3 of the V chain. This was comparable in the memory
and naive populations. The explanation for the increased hydrophobicity
of the VH CDR3/H3 is unclear, although it can be
hypothesized that this might be related to differences in the
contribution of VH or V to Ag binding. In addition,
differences in polarity might limit strong interactions between the H3
and the L3 of the Ab molecule, thereby facilitating availability of the
CDR3s for Ag binding.
Although there were few correlations between features of V and
VH chains expressed by individual B cells, it was striking
that the number of mutational events correlated in VH and
V genes. This was surprising, since molecular events and selection
appeared to affect VH and V differently (13, 14, 15). Thus,
the mutator appeared to target the VH gene preferentially,
whereas selection increased the number of mutations found in expressed
V rearrangements but decreased the number in expressed
VH rearrangements (7, 16). The net result of these
processes, however, was an increase in the frequency of mutations in
VH compared with V in the expressed Ig genes of
individual B cells, as shown here. The limited number of nonproductive
VH (n = 10) and V genes
(n = 35) in the B cells analyzed made it
impossible to assess directly the immediate impact of the mutational
machinery, as was previously conducted with larger panels of
nonproductive rearrangements (6, 7).
The reason for the preferential targeting of the VH gene by
the mutational machinery is not known, but it is interesting that the
VH gene has two matrix attachment regions (MAR) flanking
the intronic enhancer (17). It has been shown that enhancer-mediated
transcription of the Ig µ gene depends on the MAR (18) that
facilitates enhancer-induced local alteration in chromatin structure.
In contrast, in transgenes, the MAR and the intronic enhancer
do not appear to be absolutely essential for transcriptional activation
of rearranged genes (19), but critical for mutation (20). Recently,
it has been suggested that somatic hypermutation of Ig genes is linked
to transcription (21), and it has been proposed that the level of
mutation might be proportional to the rate of transcription (22). It is
unlikely that different transcriptional rates of H and L chains within
the same B cell explain the apparent differential impact of the
mutational machinery. Rather, if mutation is linked to transcription it
is more likely that the nature of transcription initiation, rather than
the transcriptional rate, explains the differences in mutational
frequencies of VH and V genes within the same B cell, as
recently proposed (23).
One reason for the correlation in the frequency of mutations found in
VH and V chains might relate to the observation that
both VH and V chains are limited in the nature of the
mutations they can tolerate and still assemble into an intact Ig
molecule and, ultimately, be secreted. Thus, somatically mutated
VH chains appear to be restricted in their ability to pair
with VL chains (24), whereas certain mutations in L chains
and the H chain abrogate secretion (25, 26, 27, 28, 29, 30). These results imply that
the need to express a functional Ig may impose constraints on the
mutations of VH and V genes permitted within an
individual B cell. Because both VH and V exhibit similar
constraints on tolerated mutations, however, it does not appear to be
likely that this mechanism would normalize the impact of the biased
mutational machinery. Rather, the overriding influence of positive
selection by Ag appears to be more important and sufficiently powerful
to impose a correlation between the number of mutations in
VH and V of productively rearranged and presumably
expressed Ig genes.
Ag-mediated positive selection of V seems to play the major role in
establishing the correlation between VH and V mutations
as evidenced by the markedly increased frequency of mutations in
productive as opposed to nonproductive VJ rearrangements (7). The
effect was apparent even though unselected B cells and not Ag-specific
clones were examined. This conclusion was further supported by the
analysis of a clonal population of memory B cells found in one donor
(data not shown). Analysis of the three members of this clone indicated
that the frequency of mutations in VH and V genes
correlated, primarily because of the correlation of the number of R
mutations in the CDRs, as would be expected from Ag-mediated
selection.
Another possibility for the different frequencies of mutations in the
Ig genes could be receptor editing of mutated V genes that might
result in the combination of a VH chain that experienced
somatic hypermutation with an unmutated replacement gene (31).
Although this possibility needs consideration, it is unlikely to
explain the apparent differences in mutational frequencies of the
entire population, since analysis of the V repertoire revealed
evidence for additional V chain rearrangements in only a small
number of individual B cells (7). Moreover, the analysis of the clonal
population of B cells supported the conclusion that VH was
mutated more frequently than V in individual B cells.
In summary, these data indicate that pairing of VH and V
chains occurs in a random fashion and that subsequent selective forces
do not alter the distribution within the entire B cell population
significantly. However, Ag-mediated selection leads to a variety of
specific changes in the expressed Ig repertoire. Disparate influences
on the VH and V chains, respectively, tend to normalize
the CDR3 lengths and the frequency of mutations in the components of
the Ig molecule of individual B cells. These results imply that
effective Ag binding is favored on a population basis when B cells
express Ig molecules with CDR3s of more similar lengths and containing
a more uniform number of mutations. Thus, despite the nearly limitless
potential of the V(D)J recombination process to generate diversity,
normal antigenic experience tends to limit the repertoire of Ig
molecules expressed by circulating memory B cells to those with more
similar VH and V CDR3s and frequencies of mutations.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 These authors contributed equally to this work.
3 Address correspondence and reprint requests to Dr. Peter E. Lipsky, Department of Internal Medicine and Harold C. Simmons Arthritis Research Center, University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Boulevard, Dallas, TX 75235-8884. E-mail address:
4 Abbreviations used in this paper: CDR, complementarity-determining region; FR, framework region; H3, heavy chain hypervariable loop 3; L3, light chain hypervariable loop 3; aa, amino acid; MAR, matrix attachment regions.
Received for publication October 10, 1997. Accepted for publication January 20, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
gene repertoire of IgM+ B cells. J. Clin. Invest. 99:1614.[Medline]
gene expression after stable integration. II: Role of the intronic MAR and enhancer in transgenic mice. J. Biol. Chem. 264:21190.
2 transgene under the control of the lambda enhancer or the heavy chain intron enhancer. J. Immunol. 157:4458.[Abstract]
This article has been cited by other articles:
|
|
 |
|
  G. F. Widhopf II, C. J. Goldberg, T. L. Toy, L. Z. Rassenti, W. G. Wierda, J. C. Byrd, M. J. Keating, J. G. Gribben, K. R. Rai, and T. J. Kipps Nonstochastic pairing of immunoglobulin heavy and light chains expressed by chronic lymphocytic leukemia B cells is predicated on the heavy chain CDR3 Blood, March 15, 2008; 111(6): 3137 - 3144. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Ohm-Laursen and T. Barington Analysis of 6912 Unselected Somatic Hypermutations in Human VDJ Rearrangements Reveals Lack of Strand Specificity and Correlation between Phase II Substitution Rates and Distance to the Nearest 3' Activation-Induced Cytidine Deaminase Target J. Immunol., April 1, 2007; 178(7): 4322 - 4334. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Stamatopoulos, C. Belessi, C. Moreno, M. Boudjograh, G. Guida, T. Smilevska, L. Belhoul, S. Stella, N. Stavroyianni, M. Crespo, et al. Over 20% of patients with chronic lymphocytic leukemia carry stereotyped receptors: pathogenetic implications and clinical correlations Blood, January 1, 2007; 109(1): 259 - 270. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Tsuiji, S. Yurasov, K. Velinzon, S. Thomas, M. C. Nussenzweig, and H. Wardemann A checkpoint for autoreactivity in human IgM+ memory B cell development J. Exp. Med., February 21, 2006; 203(2): 393 - 400. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Stamatopoulos, C. Belessi, A. Hadzidimitriou, T. Smilevska, E. Kalagiakou, K. Hatzi, N. Stavroyianni, A. Athanasiadou, A. Tsompanakou, T. Papadaki, et al. Immunoglobulin light chain repertoire in chronic lymphocytic leukemia Blood, November 15, 2005; 106(10): 3575 - 3583. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H.-P. Brezinschek, T. Dorner, N. L. Monson, R. I. Brezinschek, and P. E. Lipsky The influence of CD40-CD154 interactions on the expressed human VH repertoire: analysis of VH genes expressed by individual B cells of a patient with X-linked hyper-IgM syndrome Int. Immunol., June 1, 2000; 12(6): 767 - 775. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Rader, G. Ritter, S. Nathan, M. Elia, I. Gout, A. A. Jungbluth, L. S. Cohen, S. Welt, L. J. Old, and C. F. Barbas III The Rabbit Antibody Repertoire as a Novel Source for the Generation of Therapeutic Human Antibodies J. Biol. Chem., April 28, 2000; 275(18): 13668 - 13676. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. C. Nicholson, X. Zou, A. V. Popov, G. P. Cook, E. M. Corps, S. Humphries, C. Ayling, B. Goyenechea, J. Xian, M. J. Taussig, et al. Antibody Repertoires of Four- and Five-Feature Translocus Mice Carrying Human Immunoglobulin Heavy Chain and {kappa} and {lambda} Light Chain Yeast Artificial Chromosomes J. Immunol., December 15, 1999; 163(12): 6898 - 6906. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Schiaffella, D. Sehgal, A. O. Anderson, and R. G. Mage Gene Conversion and Hypermutation During Diversification of VH Sequences in Developing Splenic Germinal Centers of Immunized Rabbits J. Immunol., April 1, 1999; 162(7): 3984 - 3995. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
U. Klein, K. Rajewsky, and R. Kuppers Human Immunoglobulin (Ig)M+IgD+ Peripheral Blood B Cells Expressing the CD27 Cell Surface Antigen Carry Somatically Mutated Variable Region Genes: CD27 as a General Marker for Somatically Mutated (Memory) B Cells J. Exp. Med., November 2, 1998; 188(9): 1679 - 1689. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
