This information is current as
of June 16, 2017.
Characteristics of Sequences Around
Individual Nucleotide Substitutions in IgV H
Genes Suggest Different GC and AT
Mutators
Jo Spencer, Mark Dunn and Deborah K. Dunn-Walters
J Immunol 1999; 162:6596-6601; ;
http://www.jimmunol.org/content/162/11/6596
Subscription
Permissions
Email Alerts
This article cites 26 articles, 11 of which you can access for free at:
http://www.jimmunol.org/content/162/11/6596.full#ref-list-1
Information about subscribing to The Journal of Immunology is online at:
http://jimmunol.org/subscription
Submit copyright permission requests at:
http://www.aai.org/About/Publications/JI/copyright.html
Receive free email-alerts when new articles cite this article. Sign up at:
http://jimmunol.org/alerts
The Journal of Immunology is published twice each month by
The American Association of Immunologists, Inc.,
1451 Rockville Pike, Suite 650, Rockville, MD 20852
Copyright © 1999 by The American Association of
Immunologists All rights reserved.
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
References
Characteristics of Sequences Around Individual Nucleotide
Substitutions in IgVH Genes Suggest Different GC and AT
Mutators1
Jo Spencer, Mark Dunn, and Deborah K. Dunn-Walters2
I
mmunoglobulin genes accumulate mutations during the
course of a T-dependent immune response. B cells carrying
the mutated genes are selected for survival based on the affinity of the encoded surface Ig for the activating Ag (1). Hypermutation is characterized by single base substitutions, although
insertions and deletions can occur (2, 3). The process of hypermutation is tightly regulated, so that in normal circumstances only
activated B cells mutate their genes, and only Ig genes are hypermutated. Experimental systems using Ig gene constructs in cell
lines or in transgenic mice have shown that the process is regulated, at least in part, by promoter and enhancer elements that work
in conjunction with each other (4). Provided that the correctly orientated enhancers are present, the Ig gene promoter or the Ig gene
itself can be replaced without abolishing hypermutation (5, 6). The
position of the promoter defines the stretch of DNA affected by
hypermutation. The importance of promoters and enhancers suggests
the involvement of transcription in hypermutation. A mechanism of
hypermutation that is related to transcription, occurring preferentially
on one strand of DNA, is consistent with strand bias, where the observed number of mutations from A nucleotides is greater than that
from T nucleotides.
Investigation of hypermutation by sequence analysis can be confounded by the effects of selection on the mutated Ig genes, because selection has been suggested as the cause of strand bias and
“hotspots” (7, 8). Selective influences have been avoided in many
Department of Histopathology, Guy’s, King’s, and St. Thomas’ School of Medicine,
London, United Kingdom
Received for publication December 28, 1998. Accepted for publication March 10,
1999.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1
This work was supported by The Special Trustees of St. Thomas’ Hospital.
studies by looking at noncoding flanking sequences (9), passenger
transgenes in mice (10), in vitro systems (11–13), and unused,
out-of-frame, alleles of human Ig genes (7, 14). These studies reveal biases intrinsic to the hypermutation process, which may provide clues as to the mechanism of hypermutation. Hotspots and
“coldspots” of mutation have been identified, and sequence motifs
that target mutation have been proposed (9, 10, 15, 16). Because
these biases were identified by analyzing sequences around all mutations, it was not possible to identify whether there were differences between the sequences around mutated GC and AT bases.
We have shown previously that there are characteristic sequences
around mutated G’s and C’s, presumably motifs that target the
mutation process (14). These are 4-mer motifs with G in the second
position and C in the third position. They are consistent with the
RGYW motif and many of the dinucleotide and trinucleotide preferences (9, 16). The motif for G is the reverse complement of the motif
for C, which is additional evidence for a mechanism that acts on both
strands of DNA. We were unable to determine the relationship, if any,
between the sequence motifs surrounding mutated A’s and T’s, due to
the low numbers of mutated T’s available within the data studied.
In this study, we have extended our previous investigations of
human out-of-frame IgH genes using a larger data set and computational analysis which includes all mutations and distinguishes
between the different nucleotide substitutions at each base. Inherent biases in Ig gene nucleotide composition were compensated for
by calculating the normal base composition flanking A, C, G, and
T nucleotides and comparing these values with those obtained for
the flanking regions of each different nucleotide substitution. In
this way, we have determined sequence motifs surrounding mutated A’s and T’s and have shown that the motifs surrounding a
particular mutated base can differ depending upon whether the
substitutions are transitions or transversions.
Materials and Methods
2
Address correspondence and reprint requests to Dr. Deborah K. Dunn-Walters, Department of Histopathology, Guy’s, King’s, and St. Thomas’ School of Medicine, St.
Thomas’ Campus, Lambeth Palace Road, London SE1 7EH, U.K. E-mail address:
[email protected]
Copyright © 1999 by The American Association of Immunologists
Nonproductive Ig gene rearrangements that carry mutations (17–21) were
allocated into groups according to their IgVH gene usage. Seven groups of
genes were collated, containing a total of 49 sequences with 670 nucleotide
0022-1767/99/$02.00
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
Somatic hypermutation affects Ig genes during T-dependent B cell responses and is characterized by a high frequency of single
base substitutions. Hypermutation is not a completely random process; a study of mutations in different systems has revealed the
presence of sequence motifs that target mutation. In a recent analysis of the sequences surrounding individual mutated bases in
out-of-frame human IgVH genes, we found that the target motifs around mutated G’s and C’s are reverse complements of each
other. This finding suggests that hypermutation acts on both strands of DNA, which contradicts evidence of a strand-dependent
mechanism as suggested by an observed bias in A and T mutations and the involvement of transcriptional machinery. We have
now extended our database of out-of-frame genes and determined the sequence motifs flanking mutated A and T nucleotides. In
addition, we have analyzed the flanking sequences for different types of nucleotide substitutions separately. Our results confirm
the relationship between the motifs for G and C mutations and show that the motifs surrounding mutated A’s and T’s are weaker
and do not have the same relationship. Taken together with our observation of A/T strand bias in out-of-frame genes, this
observation suggests that there is a semitargeted G/C mutator that is strand-independent and a separate A/T mutator that is
strand-dependent and is less reliant on the local target sequence. The Journal of Immunology, 1999, 162: 6596 – 6601.
The Journal of Immunology
6597
Table I. Mutated, out-of-frame IgV genes used in the analysis
VH Gene
Number of
Sequences
Number of
Mutations
DP63 (421)
20
233
VH32
DP73 (251)
DP47
DP54
DP65
DP71
Total
4
11
3
3
3
5
49
76
189
32
15
51
74
670
GenBank Accession Numbers
X87032, Y13169,72 Y16646–9, Z738020–21, Z80389, Z80708,
Z93132,34,38,41,53–4,56,58–9
Z73863, Z93213–4,16
Y13167–8, Z73839,58,60, Z80570,71, Z80760, Z93198,9, Z93204
Z80673, X87013,64
Z80418,90, Z87055
X87075, X97784, Z80753
X87035,82, Z80396, Z803564, Z803719
Results
Targeting motifs around mutations occurring at G and C
Fig. 1 illustrates the differences between the sequence composition
flanking mutations from G and C compared with the control sequence composition. Significant elements are illustrated by solid
bars. These elements are taken to be a component of a motif that
targets mutation. In this way, the targeting motif around a mutated
G (bold underlined) when all mutations are considered is [A not
C/T], G, [C/T not A/G], [T not C/G] (Fig. 1c). Similarly, the motif
around a mutated C (bold underlined) when all mutations are considered is [A not C/G], [A/G not C/T], C, [T not G] (Fig. 1g). This
confirms our previous results, which showed that the targeting motifs for C and G are reverse complements of each other and are
illustrated by the graphs of C and G being reversed and shown
underneath the graphs of G and C, respectively (Fig. 1, c and d, g
and h).
Analyzing the G transitions and G transversions separately
showed that there were no major differences between the motifs
around these two different types of substitution (Fig. 1, a and b).
However, it would appear that the main contribution to the motif
around mutations from C comes from the sequence around C transitions (Fig. 1, e and f), even though the numbers of C transitions
and transversions are similar (n 5 97 and 84, respectively).
around a mutated T (bold underlined), which is T, T, [A not G], C
when all mutations are taken into account (Fig. 2g).
In contrast to the motifs for G and C mutations, these motifs are
not fully reverse complements of each other, as shown by the
reversal of graphs T and A underneath the graphs of A and T,
respectively (Fig. 2, c and d, g and h). In addition, there was a
difference between the motifs around transitions and transversions
when they were calculated separately. In the sequence around A,
the motifs deduced for transitions and transversions were completely different (Fig. 2, a and b), and there was only one significant point in the sequence around T (A at position 11) that was
common to both transitions and transversions (Fig. 2, e and f).
Strand bias
There is no strand bias seen in mutations from G and C, because
there is no significant difference in the total numbers of mutations
from each (203 and 181, respectively). In addition, when the different types of mutations are paired with their complementary mutations (G-A with C-T, etc.), the proportions of mutations within
each complementary pair are equivalent, even though different
pairs have different proportions of mutations (Fig. 3). Fig. 4, a and
b, show that there is no significant difference between complementary G/C transitions and transversions in individual groups of Ig
genes.
We observed an increase of mutations from A when compared
with mutations from T. Fig. 3 shows that when the complementary
A/T pairs are analyzed, the difference between A-C and T-G and
between A-T and T-A is significant ( p 5 1 3 1025 and p 5 1 3
1023, respectively), whereas the difference between A-G and T-C
just fails to reach significance at the 95% level ( p 5 0.058). Analysis of the complementary A/T transitions for individual groups of
Ig genes shows that the number of A-G mutations is significantly
higher than T-C mutations in only one group (DP73). DP65, DP71,
and VH32 show higher numbers of T-C than A-G. (Fig. 5a). In
contrast, the number of A transversions is significantly higher than
T transversions in DP63, VH32, DP47, and DP73, and the number
of A transversions was consistently higher than T transversions in
all groups, even though the total numbers of mutations in some of
the groups were too small to show any statistical significance (Fig.
5b). The overall effect is that the difference between the total number of A transversions and T transversions is highly significant
( p 5 5 3 1028).
Targeting motifs around mutations occurring at A and T
Fig. 2 illustrates the differences between the sequence composition
around mutations from A and T compared with the control sequence composition. In the same way as before, significant elements (solid bars) are taken to be a component of a motif that
targets mutation. Taking all mutations from A into account, the
motif is [T not C], A (Fig. 2c). This is very different from the motif
Discussion
Separate GC and AT mutators
We have shown previously (14) that the sequences that target mutation toward G and C are reverse complements of each other, and
we suggested that this is evidence that hypermutation of G’s and
C’s acts on both strands of DNA. This possibility was recently
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
substitutions (Table I). The mutated Ig genes were aligned with the appropriate germline IgVH region (22), and a text file of the alignment was
imported into a Microsoft Excel spreadsheet (Redmond, WA). Computations of the number of each type of nucleotide substitution (e.g., A-G, T-C,
A-C, etc.) and the composition of the flanking sequences around these
substitutions were performed using macros in Excel (Visual Basic). The
nucleotide distribution in germline Ig genes is not random; for example, it
is rare to have a C in position 21 from a G. To account for this, the
composition of the flanking sequences around the individual bases (A, C,
G, and T) was determined for the germline genes used in this analysis.
The proportion of each base at each position flanking a mutated base
(from 23 to 13) was calculated as a percentage. The percentage composition in the germline genes at each equivalent position flanking A, C, G,
or T (as appropriate) was subtracted from this percentage around a mutated
base to show any differences that were particular to the mutated base. In
addition to these calculations for total mutations from A, C, G, or T, calculations were performed for individual nucleotide substitutions. The composition of the flanking sequences around different types of transitions and
transversions was compared with the control values by x2 analysis to determine whether any differences from the norm were significant with 95%
confidence limits.
6598
EVIDENCE FOR DIFFERENT GC AND AT MUTATORS
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
FIGURE 1. Difference in nucleotide composition around mutated G’s and C’s compared with germline composition. Graphs show the difference in the
percentage composition of individual bases at each position (from 23 to 13) flanking mutated G’s (a– c) and C’s (e– g) when compared with germline
composition. Base composition is shown flanking transversions only (a and e), transitions only (b and f), and all types of substitutions at G and C (c and
g). Solid bars indicate a significant difference from germline composition at a p value of ,0.05 by x2 analysis. d and h are a reversal of g and c, respectively,
illustrating the relationship between the base composition of the sequence flanking G and the reverse complement of that flanking C (and vice versa).
confirmed by a study in which the number of mutations in individual codons was found to be correlated with the number of mutations in the corresponding complementary codon (23). In this
study, we have undertaken a more comprehensive analysis of the
sequences flanking mutated bases in a larger panel of out-of-frame
Ig genes and have obtained the patterns of the flanking sequences
for all four bases, and for each type of substitution individually.
Our previous motifs surrounding mutated G’s and C’s have been
confirmed as being 4-mers that are reverse complements of each
other, as one would expect if the mechanism of mutation was
targeted and acted on both strands. Further support for a strandindependent mechanism is provided by the correlation between the
The Journal of Immunology
6599
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
FIGURE 2. Difference in nucleotide composition around mutated A’s and T’s compared with germline composition. Graphs show the difference in the
percentage composition of individual bases at each position (from 23 to 13) flanking mutated A’s (a– c) and T’s (e– g) when compared with germline
composition. Base composition is shown flanking transversions only (a and e), transitions only (b and f), and all types of substitutions at A and T (c and
g). Solid bars indicate a significant difference from germline composition at a p value of ,0.05 by x2 analysis. d and h are a reversal of g and c, respectively,
illustrating the dissimilarity between the base composition of the sequence flanking A and the reverse complement of that flanking T.
numbers of individual G mutations and the numbers of their complementary C mutations (Fig. 3).
The motifs surrounding mutated A’s and T’s are very different
from the motifs for G’s and C’s and do not bear the same reverse
complement relationship. That we are able to show some signifi-
cant elements of a motif indicates that there is a targeting component in the AT mutator. However, statistical analysis showed that
the elements of the motifs around mutated G’s and C’s reached a
much higher level of significance than for the sequence around
mutated A’s and T’s, indicating that the AT mutator has less need
6600
EVIDENCE FOR DIFFERENT GC AND AT MUTATORS
of a target sequence. It should be remembered that there are many
mutations, even from G or C nucleotides, which do not fall within
targeting motifs. Therefore, it is likely that there are nontargeted
mutators whose effects may be superimposed over any targeted
mutators when observing the results of hypermutation.
The large difference between the motif for T and the motif for A,
and the fact that the numbers of individual A mutations are consistently higher than their complementary T mutations (Fig. 3), is
consistent with a mechanism acting on only one strand of DNA.
The observation of strand bias in these out-of-frame genes would
confirm that this is an intrinsic part of the hypermutation process
and not an effect of selection. There is a small element of reverse
complement in the motifs, at position 21 from A compared with
position 11 from T ([T not C], A is the reverse complement of T,
FIGURE 5. Comparison of complementary A/T transitions and transversions in individual groups of IgVH genes. Numbers of A/T transitions
(a) and A/T transversions (b) are shown for each different group of mutated
IgVH genes studied, and for the combined total. The number of A transitions and T transitions as well as the number of A transversions and T
transversions were compared by x2 in each case. Solid bars indicate significant differences at p , 0.05.
[A not G]), which might mean that some mutations from A and T
occur as the result of a mechanism acting on both strands. However, when the mutations are divided into transitions and transversions, there is no similarity between the resulting motif from one
and the reverse complement of the motif from the other (Fig. 2).
The differences between the sequence motifs surrounding mutated G/C’s and mutated A/T’s and the identification of strand bias
in A/T’s only suggest that the GC mutator is largely a targeted
mechanism that acts on both strands of DNA, whereas the AT
mutator acts mainly on one strand of DNA only and is less dependent upon a target sequence. Other evidence for independent
GC and AT mutators exists; in shark IgM and Xenopus Ig, mutations are biased toward GC (24), and a cell line harboring a mutator that preferentially targets GC base pairs has been reported
(11). IgV genes from mice deficient in the Msh2 mismatch repair
protein show a bias toward GC mutations (25), as do Ig genes from
a Burkitt’s lymphoma cell line (13) and Bcl2 genes translocated to
the Ig locus (26). We have also found an example of a bias in
mutation at AT base pairs in an in vitro mouse Ig system (12)
(Table II).
Table II. Biased mutation from GC bases and AT bases in different
experimental systemsa
Percent Mutations from
Source of Data
FIGURE 4. Comparison of complementary G/C transitions and transversions in individual groups of IgVH genes. Numbers of G/C transitions
(a) and G/C transversions (b) are shown for each different group of mutated
IgVH genes studied, and for the combined total. The number of G transitions and C transitions as well as the number of G transversions and C
transversions were compared by x2 analysis in each case. There were no
significant differences at p , 0.05.
n
A
Human out-of-frame Ig genes
670 27.5
Msh22/2 mice (25)
108 5.6
Burkitt’s lymphoma cell line (13) 152 8.6
Translocated Bcl2 genes (26)
20 0
Mouse in vitro cell line (12)
70 41.4
C
27.0
30.6
36.8
40
12.9
G
T
30.3 15.2
61.1 2.8
46.7 7.9
60
0
20
25.7
a
Numbers in parentheses refer to the references from which the data were
collated.
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
FIGURE 3. Percent distribution of each type of nucleotide substitution.
The percentage of each type of substitution in the database of out-of-frame
human genes (n 5 670) is shown. Individual substitutions are paired together with their complementary mutation (i.e., G-A with C-T, etc.) Solid
bars indicate pairs of substitutions for which the difference between the
proportion of each substitution within a pair is significant at a p value of
,0.05 by x2 analysis.
The Journal of Immunology
GC and AT mutators themselves are composed of more than one
mechanism
References
1. Jacob, J., G. Kelsoe, K. Rajewsky, and U. Weiss. 1991. Intraclonal generation of
antibody mutants in germinal centres. Nature 354:389.
2. Wilson, P. C., O. de Bouteiller, Y. J. Liu, K. N. Potter, J. Banchereau, J. D. Capra,
and V. Pascual. 1998. Somatic hypermutation introduces insertions and deletions
into immunoglobulin V genes. J. Exp. Med. 187:59.
3. Ohlin, M., and C. A. Borrebaeck. 1998. Insertions and deletions in hypervariable
loops of antibody heavy chains contribute to molecular diversity. Mol. Immunol.
35:233.
4. Storb, U., A. Peters, E. Klotz, N. Kim, H. M. Shen, J. Hackett, B. Rogerson, and
T. E. Martin. 1998. Cis-acting sequences that affect somatic hypermutation of Ig
genes. Immunol. Rev. 162:153.
5. Tumas-Brundage, K., and T. Manser. 1997. The transcriptional promoter regulates hypermutation of the antibody heavy chain locus. J. Exp. Med. 185:239.
6. Yelamos, J., N. Klix, B. Goyenechea, F. Lozano, Y. L. Chui,
A. Gonzalez Fernandez, R. Pannell, M. S. Neuberger, and C. Milstein. 1995.
Targeting of non-Ig sequences in place of the V segment by somatic hypermutation. Nature 376:225.
7. Dorner, T., H. P. Brezinschek, R. I. Brezinschek, S. J. Foster, R. Domiati-Saad,
and P. E. Lipsky. 1997. Analysis of the frequency and pattern of somatic mutations within nonproductively rearranged human variable heavy chain genes.
J. Immunol. 158:2779.
8. Betz, A. G., M. S. Neuberger, and C. Milstein. 1993. Discriminating intrinsic and
antigen-selected mutational hotspots in immunoglobulin V genes. Immunol. Today 14:405.
9. Smith, D. S., G. Creadon, P. K. Jena, J. P. Portanova, B. L. Kotzin, and
L. J. Wysocki. 1997. Di- and trinucleotide target preferences of somatic mutagenesis in normal and autoreactive B cells. J. Immunol. 156:2642.
10. Betz, A. G., C. Rada, R. Pannell, C. Milstein, and M. S. Neuberger. 1993. Passenger transgenes reveal intrinsic specificity of the antibody hypermutation mechanism: clustering, polarity, and specific hot spots. Proc. Natl. Acad. Sci. USA
90:2385.
11. Bachl, J., and M. Wabl. 1996. An immunoglobulin mutator that targets G.C base
pairs. Proc. Natl. Acad. Sci. USA 93:851.
12. Kallberg, E., S. Jainandunsing, D. Gray, and T. Leanderson. 1996. Somatic mutation of immunoglobulin V genes in vitro. Science 271:1285.
13. Denepoux, S., D. Razanajaona, D. Blanchard, G. Meffre, J. D. Capra,
J. Banchereau, and S. Lebecque. 1997. Induction of somatic mutation in a human
B cell line in vitro. Immunity 6:35.
14. Dunn-Walters, D. K., A. Dogan, L. Boursier, C. M. MacDonald, and J. Spencer.
1998. Base-specific sequences that bias somatic hypermutation deduced by analysis of out-of-frame human IgVH genes. J. Immunol. 160:2360.
15. Dorner, T., S. J. Foster, H. P. Brezinschek, and P. E. Lipsky. 1998. Analysis of
the targeting of the hypermutational machinery and the impact of subsequent
selection on the distribution of nucleotide changes in human VHDJH rearrangements. Immunol. Rev. 162:161.
16. Rogozin, I. B., and N. A. Kolchanov. 1992. Somatic hypermutagenesis in immunoglobulin genes: influence of neighbouring base sequences on mutagenesis.
Biochim. Biophys. Acta. 1171:11.
17. Dunn-Walters, D. K., L. Boursier, and J. Spencer. 1997. Hypermutation, diversity, and dissemination of human intestinal lamina propria plasma cells. Eur.
J. Immunol. 27:2959.
18. Dunn-Walters, D. K., P. G. Isaacson, and J. Spencer. 1997. Sequence analysis of
human IgVH genes indicates that ileal lamina propria plasma cells are derived
from Peyer’s patches. Eur. J. Immunol. 27:463.
19. Dunn-Walters, D. K., P. G. Isaacson, and J. Spencer. 1996. Sequence analysis of
rearranged IgVH genes from microdissected human Peyer’s patch marginal zone
B cells. Immunology 88:618.
20. Brezinschek, H. P., R. I. Brezinschek, and P. E. Lipsky. 1997. Analysis of the
heavy chain repertoire of human peripheral B cells using single-cell polymerase
chain reaction. J. Immunol. 155:190.
21. Dorner, T., H. P. Brezinschek, S. J. Foster, R. I. Brezinschek, N. L. Farner, and
P. E. Lipsky. 1998. Comparable impact of mutational and selective influences in
shaping the expressed repertoire of peripheral IgM1/CD52 and IgM1/CD51 B
cells. Eur. J. Immunol. 28:657.
22. Tomlinson, I. M., G. Walter, J. D. Marks, M. B. Llewelyn, and G. Winter. 1992.
The repertoire of human germline VH sequences reveals about fifty groups of VH
segments with different hypervariable loops. J. Mol. Biol. 227:776.
23. Milstein, C., M. S. Neuberger, and R. Staden. 1998. Both DNA strands of antibody genes are hypermutation targets. Proc. Natl. Acad. Sci. USA 95:8791.
24. Diaz, M., and M. F. Flajnik. 1998. Evolution of somatic hypermutation and gene
conversion in adaptive immunity. Immunol. Rev. 162:13.
25. Phung, Q. H., D. B. Winter, A. Cranston, R. E. Tarone, V. A. Bohr, R. Fishel, and
P. J. Gearhart. 1998. Increased hypermutation at G and C nucleotides in immunoglobulin variable genes from mice deficient in the MSH2 mismatch repair
protein. J. Exp. Med. 187:1745.
26. Spencer, J., and D. K. Dunn-Walters. 1999. Somatic hypermutation and B cell
malignancies. J. Pathol. 187:158.
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
When the motifs surrounding the different G substitutions were
determined, there was no major difference between the motif surrounding a G transversion and that around a G transition (Fig. 1, a
and b). However, the same could not be said for the motifs surrounding C transversions and transitions, where the main contribution to the targeting motif for C comes from the sequence
around C-T mutations (Fig. 1, e and f). This contrast implies a
difference in the mechanism by which G and C are mutated, albeit
a small one. In this context, it is interesting to note that although
the numbers of G and C transitions are variable and most likely
equal, a consistent (if not significant) increase of G transversions
over C transversions is seen for all groups of Ig genes studied
(Fig. 4).
The motifs around mutated A’s and T’s vary greatly depending
upon whether transversions or transitions are analyzed. Although
the significant elements in the A motif are both in position 21
from the mutated A, they are very different from one another (Fig.
2, a and b). Similarly, of the four significant elements of the motif
around mutated T’s, only the A in position 11 is common to both
transitions and transversions (Fig. 2, e and f). Interestingly, when
only transitions from A and T are considered, strand bias is much
less apparent. The difference between A-G and T-C mutations was
not consistent between the different groups of Ig genes and just
failed to reach significance on the total number of mutations ( p 5
0.058). The difference between the numbers of A transversions and
the numbers of T transversions, however, is highly significant
( p 5 5 3 1028), and is consistent between different groups of Ig
genes (Fig. 5). This difference between transitions and transversions may account for a previous report where strand bias was not
detected in out-of-frame human Ig genes (7).
The fact that the motifs around the mutated bases can be different, depending upon whether transitions or transversions are
used for the analysis, coupled with the observation that strand bias
is observed more for transversions than transitions, implies that
there may be different mechanisms acting within the GC and AT
mutators. Whether the mechanisms that cause these differences are
ones of mutation and/or of repair is not known.
In conclusion, close study of the patterns of hypermutation that
occur in vivo in humans reveal characteristics that suggest that
multiple mutators act on Ig genes. Some (but not all) of these
characteristics are recapitulated to varying degrees in artificial systems. These systems may have inadvertently separated components of the hypermutation mechanism and therefore will be useful
in elucidating the individual mechanisms that contribute to the
whole.
6601