Are human V-region sequences optimized for somatic hypermutation?

Previous studies (), as well as the above analysis showed a segregation of the more mutable codons in CDRs, and less mutable ones in framework regions. This property characterizes all human V region sequences (Fig. ). This is not to say that human V regions have maximal CDR mutability and minimal FR mutability under somatic hypermutation. I can, in fact, construct the exact nucleotide sequence with this property for any of the V region amino acid sequences. However, as I could not assess the significance of the difference in mutability between the observed and the optimal sequence, I designed instead another approach to look at the degree of optimality of the germline sequence. Namely, I explore the neighborhood of the observed germline sequence in the space of silent mutants. These are the sequences to which evolution had most immediate access. I shall describe the new experiment. I start with a germline sequence, and generate 10⁴ variants of it, each of the variants differing from the germline sequence by n silent mutations. I calculate the predicted CDR and FR replacement mutabilities, and their ratio, for each of the variants. I then determine the rank of the observed, germline, sequence among its variants. Returning to the set of human V_H sequences, among the 2 mutation neighbors, the proportion of sequences with higher CDR/FR mutability ratio than the germline sequence varies between 38 and 52 percent. Among 10 mutation neighbors, this proportion varies more widely, 25-65%. Among the 50 mutation neighbors, for some germline sequence we find as few as 4% variants with higher ratio, whereas for some other sequence, this proportion is 98.5%. Thus, the germline sequences are far from being optimal with respect to the differential mutability CDR/FR. It is a different issue whether the selection pressure to select for a higher ratio is sufficiently high. For the 2 mutation neighbors, the CDR/FR mutability ratio changes by fractions of a percent, and only for 10 mutation neighbors do we reach the level of percentages of the germline ratio. My conclusion is therefore that, although all human V_H sequences are characterized by lower FR than CDR replacement mutability per nucleotide, the mutability of individual sequences is quite far from optimal.

 

Table 3.2: Normalized rank of the ratio between the predicted average CDR and FR mutability of observed germline sequences among their 2-, 10- and 50-mutant neighbors.

Gene	2 mutation neighbors		10 mutation neighbors		50 mutation neighbors
human VH1 genes
IGHV1-18	0.541		0.532		0.565
IGHV1-2	0.535		0.449		0.261
IGHV1-24	0.599		0.71		0.867
IGHV1-30	0.527		0.505		0.404
IGHV1-45	0.519		0.459		0.226
IGHV1-46	0.529		0.505		0.449
IGHV1-58	0.52		0.495		0.434
IGHV1-69	0.561		0.593		0.726
IGHV1-8	0.501		0.348		0.069
IGHV1-f	0.544		0.562		0.608
human VH2 genes
IGHV2-26	0.624		0.752		0.956
IGHV2-50	0.601		0.639		0.796
IGHV2-70	0.641		0.751		0.951
human VH3 genes
IGHV3-11	0.525		0.476		0.404
IGHV3-13	0.483		0.447		0.295
IGHV3-15	0.496		0.477		0.377
IGHV3-16	0.514		0.36		0.095
IGHV3-19	0.487		0.359		0.076
IGHV3-20	0.572		0.59		0.583
IGHV3-21	0.503		0.495		0.471
IGHV3-23	0.507		0.477		0.382
IGHV3-30	0.551		0.558		0.714
IGHV3-30.3	0.565		0.627		0.801
IGHV3-33	0.56		0.589		0.702

 

 

Table 3.2: Normalized rank of the ratio between the predicted average CDR and FR mutability of observed germline sequences among their 2-, 10- and 50-mutant neighbors (continued).

Gene	2 mutation neighbors		10 mutation neighbors		50 mutation neighbors
IGHV3-35	0.506		0.383		0.122
IGHV3-38	0.501		0.472		0.337
IGHV3-43	0.564		0.606		0.687
IGHV3-47	0.566		0.638		0.882
IGHV3-48	0.494		0.478		0.457
IGHV3-49	0.526		0.586		0.659
IGHV3-53	0.552		0.556		0.569
IGHV3-64	0.488		0.438		0.31
IGHV3-66	0.533		0.528		0.505
IGHV3-7	0.547		0.55		0.616
IGHV3-72	0.542		0.573		0.66
IGHV3-73	0.585		0.61		0.686
IGHV3-74	0.52		0.554		0.648
IGHV3-9	0.576		0.629		0.778
IGHV3-d	0.482		0.418		0.226
human VH4 genes
IGHV4-28	0.557		0.495		0.444
IGHV4-301-4-31	0.531		0.455		0.322
IGHV4-302	0.535		0.477		0.359
IGHV4-304	0.573		0.549		0.584
IGHV4-31	0.519		0.438		0.306
IGHV4-34	0.616		0.606		0.696
IGHV4-39	0.547		0.546		0.612
IGHV4-4	0.607		0.616		0.756
IGHV4-59	0.574		0.545		0.562
IGHV4-61	0.597		0.574		0.671
IGHV4-b	0.582		0.594		0.737
human VH5 genes
IGHV5-51	0.583		0.701		0.954
IGHV5-a	0.598		0.705		0.938
human VH6 gene
IGHV6-1	0.55		0.625		0.804
human VH7 genes
IGHV7-41	0.586		0.53		0.419
IGHV7-81	0.446		0.274		0.015