| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gene Utilization, Junctional VJ and V
J Diversity, and Somatic Hypermutation1
*
Division of Molecular Immunology, Department of Pathology, Cornell University Medical College, and
The Immunology Program, Cornell University Graduate School of Medical Sciences, New York, NY 10021
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
J and V
J genes to complete the primary
structure of the Ag-binding site of these mAbs. While the
anti-rabies virus mAb selection of V genes (2e.2.2 twice, DPL11,
and DPL23) reflected the representation of the V genes in the human
haploid genome (stochastic utilization), that of V genes (O2/O12
twice, O8/O18, A3/A19, A27, and L2) did not (p =
0.0018) (nonstochastic utilization). Furthermore, the selection of both
V and V genes by the anti-rabies virus mAbs vastly overlapped
with that of 557 assorted VJ rearrangements, that of 253 VJ
rearrangements in -type gammopathies, and that of other Abs to
thymus-dependent Ags, including 23 anti-HIV mAbs and 51 rheumatoid
factors, but differed from that of 43 Abs to Haemophilus
influenzae type b polysaccharide, a prototypic thymus-independent
(TI) Ag. The anti-rabies virus mAb VJ and VJ segments
displayed variable numbers of somatic mutations, which, in mAb58 and
the virus-neutralizing mAb57, entailed a significant concentration of
amino acid replacements in the complementarity-determining regions
(p = 0.0028 and p = 0.0023,
respectively), suggesting a selection by Ag. This Ag-dependent somatic
selection process was superimposed on a somatic diversification process
that occurred at the stage of B cell receptor for Ag rearrangement, and
that entailed V gene 3' truncation and N nucleotide additions to yield
heterogeneous CDR3s. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In spite of its importance, the analysis of the human Ab response to TD Ags is virtually limited to our studies of the H chain VDJ genes of mAbs to rabies virus (5, 7). Our choice of rabies virus has been dictated by the lack of prior exposure to this virus in the vast majority of the population, by the ease with which a previous exposure can be verified in the anamnesis of perspective volunteers to be immunized, and by the fact that the full vaccination schedule calls for three sequential injections, thus allowing the study of Ab affinity maturation. Finally, as we have shown, a subject that has not been exposed to rabies virus does not recognize it because of the lack of cross-reacting epitopes with other human viruses (17).
Using B cells from healthy volunteers vaccinated with human diploid
cell rabies virus vaccine, we have measured the frequency of
anti-rabies virus IgM, IgG, and IgA producing cell precursors, and
we have established 10 cell lines secreting IgM, IgG, or IgA mAbs to
rabies virus Ags before or after immunization with inactivated virus
vaccine (5). We then have analyzed the
VHDJH genes of these mAbs, and found that
VH3 family genes were preferentially utilized, and that
most of these genes overlapped with those found to be predominantly
expressed in the fetal liver and in the putatively unselected adult
human B cell repertoire (7). Furthermore, we have
documented traces of somatic hypermutation in all monoreactive high
affinity mAb VHDJH genes, including the rabies
virus-neutralizing IgG1 mAb57 (7).
In the present study, the VLJL gene segments of
the above 10 mAbs to rabies virus were sequenced to characterize the
composition of this TD human Ab response. The V and V genes
utilized by the mAbs to rabies virus vastly overlapped with the V
and V genes expressed in the human B cell repertoire at large, and
showed traces of somatic mutation. Thus, the specific human Ab response
to proteinic TD Ags reflects the overall expressed B cell repertoire
and is somatically diversified, not only at the H, but also at the L
chain level.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Four healthy volunteers were vaccinated with ß-propiolactone-inactivated human diploid cell vaccine against rabies virus (PM-1503-3 M strain) (Merieux Immunovax Rabies Vaccine; Merieux Institute, Miami, FL), according to the recommended schedule. Peripheral blood B cells from these subjects were obtained before and after vaccination to generate IgM, IgG, and IgA mAb-producing cell lines, as reported (5, 7). mAb50 was generated from donor A; mAb52, mAb53, mAb57, mAb58, and mAb59 were generated from donor B; mAb55, mAb56, and mAb105 were generated from donor C; and, finally, mAb107 was generated from donor D. mAb50 and mAb52 were generated using B cells before vaccination. All other mAbs were generated using B cells obtained after the completion of the vaccination schedule (5, 7).
Cloning and sequencing of the anti-rabies virus mAb
V-J and V-J gene segments
mRNA was extracted from the mAb-producing cells, and first
strand cDNA was synthesized using Moloney murine leukemia virus reverse
transcriptase (7). PCR amplification of the L chain
consisted of five individual PCR reactions, each reaction included a
sense leader V or V primer specific for the members of the L
chain gene families, in combination with an antisense C or C
oligonucleotide primer (18, 19). Each sense primer
consisted of a degenerate sequence encompassing an area of the Ig gene
leader region, as follows: V1
[5'-ATG(GA)CC(TG)GCT(CT)CCCTCTCCTCCT-3']; V2-6
[5'-ATG(AG)C(CT)TGGACCC(CT)(AT)CTC(CT)(TG)(TG)TT-3']; V1-2
[5'-AGCTCCTGGGGCT(GC)CT(AG)(AC)TGCTCT-3']; V3
[5'-TCTCTTCCTCCTGCTACTCTGGCT-3']; and V4
[5'-ATGGTGTTGCAGACCCAGGTCTTC-3']. The antisense oligonucleotide
primer consisted of the reverse complement of a 23 nucleotide C
sequence [5'-AGGAGACTCCTCGAAGTTCGGTT-3'] or of a 23 nucleotide C
sequence [5'-AGAAGGGCGGTAGACTACTCGTC-3']. For PCR amplification of
V genes, 30 cycles consisting of the following steps were used:
denaturing, 94°C (1 min); annealing, 52°C (1 min); and extension,
72°C (2 min, last cycle 15 min). For V gene amplification, an
annealing temperature of 58°C was utilized. PCR products were ligated
into PCR II plasmid vectors (Invitrogen, San Diego, CA). The ligation
mixture was used to transform INV
F' competent cells according to the
manufacturer’s protocol (Invitrogen). Recombinant clones were
sequenced by the dideoxy chain termination method using the
Taq Sequencing Kit (Promega, Madison, WI). Each
VL-JL sequence was derived from the analysis of
at least four independent recombinant clones. Differences in nucleotide
sequences among different clones were observed in few cases (less than
2 x 10-4/base, consistent with the error rate of the
Taq polymerase) and such variants were excluded from the
sequence analysis.
Analysis of Ig V gene sequences and mutations
VL and JL sequences were analyzed using the BLAST algorithm as found in the NCBI World-Wide Web home page, accessed through the Netscape Navigator version 3.0, and the GenBank database. In addition, the MacVector 5.0.1 sequence analysis software (Eastman Kodak, Rochester, NY) was used to analyze the current human Ig gene V-BASE database, as found on the World Wide Web (www.mrc-cpe.cam.ac.uk/imt-doc/public/INTRO.htlm).
The number of expected replacement (R) mutations in the Ig V gene CDRs or FRs was calculated using the formula RCDR or RFR = n x (CDRRf or FRRf) x (CDRrel or FRrel), where n is the number of observed (R + S, silent) mutations, Rf is the replacement frequency inherent to complementarity-determining region (CDR) or framework (FR) sequences, and CDRrel and FRrel are the relative size of the CDRs or FRs (20). A binomial probability model was used to verify whether the excess and the scarcity of R mutations in the CDRs and FRs, respectively, were due to chance alone: p = {n!/[k! (n-k)!]} x qk x (1-q)n-k, where q is the probability that a R mutation will localize to CDRs or FRs (q = CDRrel x CDRRf or FRrel x FRRf), and k is the number of observed R mutations in CDRs or FRs (20).
Statistical analysis
The V and V genes utilized by the anti-rabies mAbs,
the V genes of 557 expressed V-J rearrangements, the V
segments of 253 monoclonal gammopathies, and the V and the V
genes of 23 anti-HIV Abs, 51 rheumatoid factors (RFs) and 43
anti-Hib PS Abs were compared with the complexity and number of
V and V genes in the human haploid genome using the
2 goodness-of-fit test. The V and V genes of the
anti-rabies virus mAbs, the V genes of the expressed V-J
rearrangements, the V segments of the monoclonal gammopathies, and
the V and V genes of the anti-HIV Abs, the RFs, and the
anti-Hib PS Abs were compared pairwise to each other, using a
2 test for proportions. In the pairwise comparison of
these groups a Bonferroni correction with 0.005 probability as the
criterion for significance was used. The correlation between the mAb H
and L chain somatic mutations and CDR3 length was analyzed by
calculation of the correlation coefficient (R).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
J and VJ gene
segments of the mAbs
In the human, one of the functional V gene segments is
juxtaposed by rearrangement with one of the five J gene segments on
chromosome 2p11.2 (21). The total germline repertoire
comprises 76 V genes, of which 38 are potentially functional, based
on the sequence of their coding and regulatory regions
(22, 23, 24, 25, 26, 27, 28, 29). In contrast to the locus, which contains a
single C gene cluster, there are 7 to 10 J-C gene clusters on
human chromosome 22q11.2, only four of which, J C1, J C2,
J C3, and J C7, are functional (30, 31, 32, 33). Of
the total 37 germline V genes possessing an open reading frame,
approximately 30 have been found to be rearranged, and 14 are
pseudogenes (34, 35, 36, 37, 38, 39). Virtually all human germline V
and V genes have been sequenced (21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39), and thorough
analysis of the reported V and V cDNAs and proteins has shown
that these sequences could always be ascribed to those of the known
germline V and V genes (24, 26, 27, 28, 29, 38), and the
extent of variation of the V genes (21, 22, 23, 24, 25, 26, 27, 28, 29, 40) and
V genes (30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 41) between different individuals is
only minimal. Thus, the attribution of the expressed V and J or
V and J gene sequences, including those of the anti-rabies
virus mAbs, to the germline genes available in the current Ig V and
J or V and J gene databases is appropriate.
The H chain isotype, the L chain type, and the Ag-binding properties of
our 10 mAbs to rabies virus have been previously reported (5, 7, 17) and are summarized in Table I
.
Six mAbs bear and four L chains. The 10 mAb V and V gene
segment nucleotide and deduced amino acid sequences are depicted in
Figures 1 and
2, respectively. Those of the J and
J gene segments are depicted in Figure 3. The differences of the rabies virus
mAb V, J, V, and J gene segments, as compared with the
nucleotide sequences of the respective germline, are summarized in
Table I.
|
|
|
J and VJ
(CDR3) sequence diversity in the rabies virus
mAbs
To assess the possible contribution of the L chain V-J junctions
to the specificity and diversity of the human anti-rabies virus Ab
response, the nucleotide sequences encoding the mAb L chain CDR3s were
examined in detail. The primary structure of the anti-rabies virus
mAb L chain CDR3 is depicted in Figure 4,
and is compared with that of the H chain CDR3 of the same
anti-rabies virus mAbs in Table II.
The L chain CDR3 length of the anti-rabies virus mAbs ranged from 7
to 12 amino acid residues. They were diverse in composition as a result
of somatic mutation or 3' truncation of the V gene (mAb50, mAb52,
mAb56), 3' truncations of the V gene and/or somatic mutation of the J
gene (mAb55, mAb58, mAb59), N additions (mAb57), or various
combinations of these mechanisms (mAb105, mAb107) (Fig. 4).
|
|
and V gene segmentsIn the absence of negative or positive selective pressure on a gene product, R and S mutations are randomly distributed throughout the coding DNA sequence. If a DNA segment displays a number of R mutations higher than that expected by chance alone, it is likely that a positive pressure was exerted on the gene product to select for those mutations, as it occurs in the V segment CDRs of affinity mature Abs. Conversely, if a DNA segment displays a number of R mutations lower than that expected by chance alone, it is likely that a negative pressure for R mutations was exerted on the gene product so that the protein structure is preserved as it is in the FRs of functional Abs.
The polyreactive IgM mAb50, IgM mAb52, and IgM mAb55 V gene segments
were virtually unmutated (Fig. 1
). The V segment of the fourth
polyreactive IgM, mAb59, was a mutated DPL11. Four of the five
nucleotide changes within CDRs, but only two of the four nucleotide
changes in the FRs were R mutations, yielding R:S mutation ratios of
4.0 and 1.0, respectively (Table I). The three monoreactive IgG1 mAb53,
IgG1 mAb58, and IgA1 mAb105 V segments contained nucleotide changes
(Fig. 1). A total of 7 of the 10 and 6 of the 8 nucleotide changes in
the mAb53 O2/O12 and the mAb105 O8/O18 V genes, respectively,
yielded amino acid replacements, which in both mAbs were distributed
throughout the V segment CDRs or FRs. Eight of the nine nucleotide
changes in the mAb58 O1/O12 V segment CDRs, but only one of five
nucleotide changes in the FRs were R mutations (yielding R:S mutation
ratios of 8.0 and 0.2, respectively), consistent with a negligible
probability that the excess of R mutations in CDRs and the scarcity
thereof in the FRs arose by chance only (Table I). The monoreactive
IgG1 mAb56, IgG1 mAb57, and IgG1 mAb107 V segments also contained
nucleotide changes (Fig. 1). Four of the nine and six of the eight
nucleotide changes in the mAb56 and mAb107 2e.22 and DPL23 V gene
segments yielded amino acid replacements. In both mAbs, these R
mutations were distributed throughout the CDRs and FRs. In contrast,
six (five actual and one double change) of the eight nucleotide changes
in the mAb57 2e.22 V gene segment were R mutations. These R
mutations were all in the CDRs, yielding a R:S mutation ratio of 6.0,
which was consistent with a negligible probability that they arose by
chance only (Table I). Thus, the precursors of mAb57- and
mAb58-producing cells had been subjected to a significant positive
pressure to mutate the expressed Ig V segment CDRs, while conserving
that of FRs.
Comparison of the V genes utilized by the
anti-rabies virus mAbs to those of random
VJ rearrangements, anti-HIV
mAbs, rheumatoid factor (RF) autoantibodies, and
anti-Hib PS Abs
The V genes utilized by the anti-rabies virus mAbs (Fig. 5C) were compared with: 1) the
functional V genes as represented in the human haploid genome
(expected or stochastic utilization) (Fig. 5A) (25, 28, 29); 2) the V genes of 557 putatively random productive
VJ rearrangements in genetically diverse humans (Fig. 5B) (26, 40); 3) the V genes of 15
anti-HIV mAbs (Fig. 5D) (43, 44, 45, 46, 47, 48); 4) the
V genes of 29 RFs in patients with rheumatoid arthritis (Fig. 5E) (49, 50, 51, 52, 53, 54, 55, 56, 57); and 5) the V genes of 36
anti-Hib PS Abs (Fig. 5F) (9, 10, 11, 12, 13, 14, 58, 59).
|
gene rearrangements, the anti-HIV mAbs, and
the RFs, the anti-rabies mAbs utilized predominantly V1 and
V3 and, to a lesser extent, V2 genes. Such V gene utilization
was restricted and differed from that theoretically expected on the
basis of the V gene representation in the human haploid genome (Fig. 5A) (p = 0.0001), but vastly
overlapped with that of the anti-HIV Abs and the RFs. In the
anti-Hib PS Abs, the pattern of the V gene utilization was also
restricted, and was also at variance with that theoretically expected
on the basis of the V gene representation in the human haploid
genome (p = 0.0001), but differed strikingly
from that of the rabies virus mAbs and the anti-HIV Abs and
the RFs (anti-Hib PS vs anti-rabies virus Abs,
p = 0.0013; vs anti-HIV Abs, p =
0.0001; vs RFs, p = 0.0001). In the anti-Hib PS
Abs, V2 genes were dominant, accounting for 23 of the 36 V
segments, although V1 and V3 genes were also utilized (Fig. 5F).
The 10 V germline genes or indistinguishable gene pairs O2/O12,
O8/O18, L8, L12, A17, A3/A19, A27, L2, L6, and B3 accounted for
approximately 80% of the 557 VJ rearrangements (Fig. 5B) (26, 40). Of these V genes, five
(O2/O12, O8/O18, A3/A19, A27, and L2) accounted for approximately 50%
of all 557 VJ rearrangements, and made up all anti-rabies
virus type mAbs (Fig. 5C). Most of the V genes
utilized by the anti-HIV mAbs and the RFs were also part of the
most frequently rearranged V genes (Fig. 5, D and
E). In contrast, 23 of the 36 anti-Hib PS Abs utilized
V2 genes (Fig. 5F). Of these 23 V2 genes, 21, i.e.,
91%, or 59% of the total V anti-Hib PS Abs, were accounted for
by the single V2 A2 gene. A2 would be expected to contribute less
than 3% of the random V gene rearrangements (Fig. 5A),
and was found only twice in the putatively random 557 VJ
rearrangements (Fig. 5B) (9, 10, 11, 12, 13, 14, 26, 40, 58, 59).
Comparison of the V genes utilized by
anti-rabies virus mAbs to those of monoclonal
gammopathies, anti-HIV mAbs, RF autoantibodies, and
anti-Hib PS Abs
The V genes utilized by the anti-rabies virus mAbs (Fig. 6C) were compared with: 1) the
functional V genes as represented in the human haploid genome
(expected or stochastic utilization) (Fig. 6A)
(38); 2) the V genes of 253 -type monoclonal
gammopathies from patients with myeloma or macroglobulinemia (Fig. 6B) (60); 3) the V genes of eight
anti-HIV mAbs (Fig. 6D) (44, 45, 46, 47, 48); 4) the
V genes of 22 RFs in patients with rheumatoid arthritis (Fig. 6E) (49, 50, 51, 52, 53, 54, 55, 56, 57); and 5) the V genes of seven
anti-Hib PS mAbs (Fig. 6F) (9, 10, 11, 12, 13, 14, 58, 59).
Genes of the V2 and V3 subgroups were utilized by the
anti-rabies virus mAbs, as well as by the anti-HIV mAbs,
although some V1 and V6 genes were also utilized by these mAbs.
The majority of the V segments utilized by the monoclonal Igs of
neoplastic origin were members of the three major subgroups V1,
V2, and V3 (Fig. 6B). The V genes utilized by the
RFs comprised members of the V1, V2, V3, and V4 subgroups,
as well as members of the small V5, V8, and V9 subgroups.
Overall, the pattern of V gene utilization by the anti-rabies
virus mAbs vastly overlapped with that of the anti-HIV mAbs, the
RFs, and the 253 monoclonal Ig, and was consistent with that
theoretically expected on the basis of the V gene complexity and
representation in the human haploid genome (Fig. 6A). It,
however, differed from the selection of V genes utilized by the
anti-Hib PS Abs (p = 0.06), which also
differed, in this respect, from the anti-HIV mAbs
(p = 0.049), the RFs (p
= 0.0002), and the monoclonal gammopathies (p =
0.0001), and was at variance with that expected on the basis of the
V gene representation in the human haploid genome
(p = 0.0001). A single gene, DPL19 of the V7
gene subgroup (37, 38), accounted for four of the seven
-type anti-Hib PS mAbs (Fig. 6F). DPL19 was not used
by any of the rabies virus mAbs, or any of the other Abs or 253
-type monoclonal Ig.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
sequences compiled from the literature
and from the GenBank database suggest a biased V gene expression in
the human B cell repertoire of genetically different humans (26, 40). Of the 38 potentially functional V genes, only 20 genes
or indistinguishable gene pairs have been found to be utilized.
Furthermore, of these, only 10 genes or indistinguishable gene pairs
(O2/O12, O8/O18, L8, L12, A17, A3/A19, A27, L2, L6, and B3), most of
which are proximal to the J locus on human chromosome 2, contribute
about 90% of all the expressed V gene sequences, whereas the
majority of J-distal V genes is underrepresented (26, 40). A similar V gene utilization has been reported in
anti-HIV mAbs and autoimmune RFs (44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57), as well as
in human cord blood B cells (29, 61) and, as shown here,
in all six chain mAbs to rabies virus (O2/O12 twice, O8/O18,
A3/A19, A27, and L2), suggesting that this V gene restriction is
characteristic of certain human TD responses, but not of the prototypic
TI Ab response to Hib PS. The dominant V2 utilization by the
anti-Hib PS Abs accounted for their high : ratio (3.6)
compared with the anti-rabies virus mAbs (1.5), the anti-HIV
mAbs (1.9), and the RFs (1.3), and possibly reflected a strong positive
selection pressure exerted by Hib PS on the V2 gene
product-expressing B clonotype precursors.
Our findings further support the speculation that the human locus
is segregated into two phylogenetically distinct sets of V genes. One
set would have been selected for including genes, the products of which
could bind structurally conserved and phylogenetically primordial Ags,
as exemplified by the A2 V gene product and the TI Hib PS Ag, an
elemental repeated disaccharide. The other set would consist of genes
encoding for more diverse V segments that are permissive for binding
heterogeneous and complex structures, such as those of bacterial or
viral glycoproteins, and would include the 10 V genes and gene pairs
O2/O12, O8/O18, L8, L12, A17, A3/A19, A27, L2, L6, and B3. The dominant
expression of these genes in genetically diverse humans may be due to
the preferential selection and expansion of B clonotypes expressing
these V genes by a variety of commonly occurring microorganisms.
Genes of the V2 or V3 subgroups were utilized by the four
-type anti-rabies virus mAbs. Preferential utilization of V2
and V3 genes has also been observed in monoclonal B cells of
neoplastic origin (60), and in human anti-HIV mAbs
(44, 45, 46, 47, 48), and may merely reflect the fact that V2 and
V3, together with V1, are the largest subgroups among the total
37 functional germline V genes, containing 7, 6, and 10 genes,
respectively (34). The frequent utilization of these V
gene segments may also be due to their position (62), as they make up
for most of the J-proximal "cluster A" of the three distinct
clusters of genes in which the human locus is organized on
chromosome 22 (34, 39), and may represent an old
phenomenon in the evolution of terrestrial vertebrates, as the species
that express predominantly -chains, including the chicken, the cow,
the horse, and the sheep, use V genes that are closely related to
members of the human V1, V2, or V3 subgroups
(38). The individual V genes utilized by the
anti-rabies virus mAbs (2e.2.2, DPL11, and DPL12) are also
frequently utilized to encode for Abs to a variety of other TD Ags
(2, 4, 5, 39, 45, 63), but do not include the V gene
that is dominant in the Abs to the TI Hib PS, i.e., DPL19 of the small
V7 gene subgroup. This different and, perhaps, more diverse V
gene utilization by the rabies virus mAbs may reflect the selection of
gene products that are better able to fit these structurally more
complex proteinic or glycoproteinic TD Ags.
There seems to be no obvious correlation between VL and
VH gene usage at the level of the individual mAb. The
diverse anti-rabies virus mAb V1, V2, V3, and V2 and
V3 genes are paired with different VH genes, although in
seven out of the nine mAbs of the same family (VH3). This
lack of correlation is exemplified by the two rabies virus mAbs that
use the same VL gene segment (mAb56 and mAb57, DPL12; mAb53
and mAb58, DPK9), but pair it with different VH gene
segments (mAb56, V3-53, and mAb57, V1-69; mAb53, V3-23, and mAb58,
V4-59). There appears to be no obvious correlation between the rabies
virus structure recognized and the VL gene segment
utilized. mAb50, mAb52, mAb55, mAb105, mAb59, and mAb56, all six
recognize the virus RNP but utilize six different VL genes,
namely four V (L2, A3/A19, A27, O8/O18) and two V (DPL11, 2e.2.2)
(Table I). Likewise, mAb58, mAb57, and mAb107 all recognize the viral
glycoprotein—a specificity that plays a central role in the protection
against lethal rabies virus infection, by neutralizing extracellular
virus, and by lysing virus-infected cells through complement activation
(64)—but utilize three different VL genes,
namely one V (O2/O12) and two V genes (2e.2.2, DPL23). This lack
of correlation is further emphasized by mAb56 and mAb57, which use the
same VL gene (DPL11), but are specific for the RNP and the
glycoprotein, respectively, and by mAb53, which shares the same
VL gene segment (O2/O12) with the anti-RNP mAb58 but
recognizes the viral M protein (Table I), and is consistent with the
lack of correlation between viral structure recognized and
VH gene usage (7).
Most anti-rabies virus mAb VL gene segments bore a
considerable load of mutations, although, overall, the average number
of somatic point mutations in the VL chains (6.1 ±
4.2, mean value ± SD; range, 0 to 12) was about 3.6-fold smaller
than in the VH chains (22.2 ± 5.7; 14 to 31)
(p < 0.0001). In most mAbs, both the V and
the V segment mutations included a significant number of R
mutations. These R mutations paralleled R mutations in the respective
VH gene segments, ranging from none or few R mutations in
the L and H chains in mAb52, mAb53, and mAb55, to as many as eight and
five R mutations in the L chain and 18 and 17 in the H chain of mAb57
and mAb58. Despite its overall smaller load, the number of R mutations
in the L chains correlated significantly with the number of R mutations
in the respective mAb H chains (R = 0.695,
p = 0.037), possibly reflecting the crucial role of R
mutations in selection by Ag and affinity maturation, as in mAb53,
mAb56, mAb58, mAb105, mAb107, and acquisition of virus-neutralizing
activity, as in mAb58 (Kd = 1.0 x
10-9 to 1.1 x 10-10 g/µl)
(17).
A dominant role of the H chain CDR3 in providing the major structural
correlate for Ag binding has been inferred from the crystallographic
analysis of Ag-Ab complexes, and the in vitro expression of recombined
and mutagenized Ig VDJ genes (65, 66, 67, 68, 69, 70, 71). In addition to the
H chain CDR3, the L chain CDR3 also plays a role in Ag binding. As
expected from an orthodox V-J junction, and further extending the
finding that most rearranged human -chain CDR3s are nine amino acids
long (27, 42), the CDR3 of five -chain rabies virus
mAbs consisted of nine residues. Due to the truncation of the 3' end of
the V segment, the mAb52 -chain CDR3 was seven amino acids in
length. As for the -type mAbs, two CDR3s were orthodox in length,
while two (mAb56 and mAb57) deviated from the expected orthodox
VJ junction length. Overall, there was not a significant
correlation between the length of the CDR3 of the mAb L chain and that
of the H chain (R = 0.042, p = 0.92,
Table II), nor was there a significant correlation between the number
of R mutations in the L and H chain CDR3 (R = 0.135,
p = 0.73) (Figs. 1 and 2) (7).
Nevertheless, mAb52, mAb55, and mAb107 were concordant for scarcity of
somatic mutations in the L and H chain CDR3s, which were unmutated or
virtually unmutated in all three mAbs, and mAb59 and 105 were
concordant for abundance of somatic mutations in the L and H chain
CDR3s, which bore a comparable heavy load of somatic mutations in both
mAbs. In addition to the H and L chain CDR3s, the L chain CDR1 and H
chain CDR2 also make an important contribution to the surface structure
of the Ag-binding site. In more than 80% of Ig, these regions possess
a high degree of variability in length and structure, and their primary
structure can be used, together with that of the CDR3s, to predict the
tridimensional configuration of the Ag-binding site, as outlined by
Vargas-Madrazo et al. (72), based on the canonical Ig V
structure model proposed by Chothia and Lesk (73) and
Chothia et al. (74). Using these parameters, we predicted
that the Ag-binding site of mAb52, mAb55, mAb59, and mAb57 is convex,
that of mAb53, mAb58, mAb105, and mAb107 is possibly flat, and finally,
that of mAb56 is putatively concave (Table II).
|
segments were CCT to CCG or CCC
changes, while the VJ junction of mAb57 contained two additional
codons, 95B and 95C, consisting of CCG and GGG triplets. Consistent
with the undetectable level of TdT by immunofluorescence in pre-B cells
at the time of L chain rearrangement (77), possible N
additions are found at VJ junctions (24, 61, 78, 79, 80), but are only rarely found at VJ junctions
(29). By providing further evidence of N additions in not
only V-J but also V-J junctions, our findings suggest that
the residual TdT activity detected in pre-B cells utilizing
quantitative RT-PCR (81) is sufficient to insert unencoded
nucleotides in VJ junctions, thereby further enhancing the
structural diversity of the expressed Ab repertoire.
In spite of the presence of relatively abundant N additions and R
mutations (23 of the 94 CDR3 codons contained R mutations), the rabies
virus mAb and CDR3 codons 90 (all mAbs but mAb107) and 97 (all
mAbs) were conserved at the amino acid level (Gln and Thr in
-chains, and Ser and Val in -chains). This conservation extended
to a Pro at position 95 in mAb55, mAb105, mAb58, and mAb105; 95A in
mAb59 and mAb107; 95B in mAb57; or both positions 94 and 95 in mAb53.
The Pro 95 of the four -type mAbs was encoded by conserved codons in
which the third nucleotide was substituted from the germline sequence.
In all other positions, the conserved Pro was encoded by somatic
mutations and/or N additions. By revealing a significant conservation
of CDR3 residues 90, 94/95, and 97 in both and chains of the
anti-rabies virus mAbs, our findings suggest a major role of these
residues in maintaining a sound geometry of the Ag-binding site. Taken
collectively, they provide experimental evidence for the conservation
of the human -chain CDR3 canonical structures I and II, as proposed
by Tomlinson et al. (28), in the affinity maturation of a
specific human Ab response, and extend a similar principle to the
-chain CDR3. Finally, they identify the CDR3 in both - and
-chains as a major target of the strong selection pressure applied
to the anti-rabies virus mAb-producing cell precursors in vivo. The
Ag epitopes and the Ag-binding site residues, as well as the cellular
and molecular dynamics that underlie such a selection pressure, remain
to be determined.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Paolo Casali, Cornell University Medical College, 1300 York Avenue (C-320), New York, NY 10021. E-mail address:
3 Abbreviations used in this paper: Hib PS, Haemophilus influenzae type b capsular polysaccharide; BCR, B cell receptor for Ag; CDR, complementarity-determining region; FR, framework region; H chain, heavy chain; L chain, light chain; R, replacement; RF, rheumatoid factor; S, silent; TD, thymus-dependent; TI, thymus-independent.
Received for publication January 14, 1998. Accepted for publication May 11, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
II gene and one of several J genes are joined by an invariant arginine to form the most common L chain V region. J. Immunol. 143:4110.[Abstract]
segment structure of anti-insulin IgG autoantibodies in patients with insulin-dependent diabetes mellitus: evidence for somatic selection. J. Immunol. 152:1430.[Abstract]
J region gene. J. Biol. Chem. 257:1516. locus: characterization of the duplicated O regions. Eur. J. Immunol. 21:1821.[Medline]
locus: characterization of the duplicated A regions. Eur. J. Immunol. 22:1023.[Medline]
genes and their hypermutation. Eur. J. Immunol. 23:3248.[Medline]
genes of the L regions and the repertoire of V gene sequences in the human germ line. Eur. J. Immunol. 23:2868.[Medline]
segments reveals a strong bias in their usage. Eur. J. Immunol. 24:827.[Medline]
genes. Ann. NY Acad. Sci. 746:74.
domain. EMBO J. 14:4628.[Medline]
repertoire. J. Immunol. 158:3761.[Abstract]
light-chain genes in human populations. Nature 304:172.[Medline]
genes. J. Exp. Med. 172:609. and J-C gene segments of the human immunoglobulin chain locus are separated by 14 kb and rearrange by a deletion mechanism. Eur. J. Immunol. 21:1513.[Medline]
VI germline gene. Mol. Immunol. 31:531.[Medline]
gene family-VX. J. Immunol. 152:4969.[Abstract]
gene segments. Eur. J. Immunol. 23:1456.[Medline]
repertoire. J. Mol. Biol. 264:220.[Medline]
repertoire. J. Mol. Biol. 268:69.[Medline]
III genes. Mol. Immunol. 31:657.[Medline]
II light chain gene. J. Immunol. 147:2795.3 gene in rheumatoid factors derived from the synovial cells of rheumatoid arthritis patients. Arthritis Rheum. 37:860.[Medline]
-light-chain variable-region subgroups in multiple myeloma, AL amyloidosis, and Waldenstrom’s macroglobulinemia. Clin. Immunol. Immunopathol. 71:183.[Medline]
III gene usage is nonstochastic in both adult and newborn peripheral B cells and that peripheral CD5+ adult B cells are oligoclonal. J. Clin. Invest. 93:2093.
II gene family. J. Clin. Invest. 94:532.
III-J junctional diversity of polyclonal rheumatoid factors during rheumatoid arthritis frequently reveals N addition. Eur. J. Immunol. 22:1773.[Medline]
genes with no or little somatic mutation throughout life. Eur. J. Immunol. 23:3272.[Medline]
This article has been cited by other articles:
|
|
 |
|
  P. Jirholt, L. Strandberg, B. Jansson, E. Krambovitis, E. Soderlind, C. A.K. Borrebaeck, R. Carlsson, L. Danielsson, and M. Ohlin A central core structure in an antibody variable domain determines antigen specificity Protein Eng. Des. Sel., January 1, 2001; 14(1): 67 - 74. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Zan, A. Cerutti, P. Dramitinos, A. Schaffer, Z. Li, and P. Casali Induction of Ig Somatic Hypermutation and Class Switching in a Human Monoclonal IgM+ IgD+ B Cell Line In Vitro: Definition of the Requirements and Modalities of Hypermutation J. Immunol., March 15, 1999; 162(6): 3437 - 3447. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
