Peter R. Sibbald

In both immunoglobulins (Ig) and T cell receptors (TCR), the rearrangement of V, D, and J region sequence elements during lymphocyte maturation creates an enormous degree of diversity in an area referred to as the complementarity determining region 3 (CDR3) loop. Variations in the particular V, D, and J elements used, precise points of recombination, and random nucleotide addition all lead to extensive length and sequence heterogeneity. CDR3 loops are often critical for antigen binding in Igs and appear to provide the principal peptide binding residues in TCRs. To better understand the physical and selective constraints on these sequences, we have compiled information on CDR3 size variation for Ig H, L (kappa and lambda) and TCR alpha, beta, gamma, and delta. Ig H and TCR delta CDR3s are the most variable in size and are significantly longer than L and gamma chains, respectively. In contrast, TCR alpha and beta chain distributions are highly constrained, with nearly identical average CDR3 lengths, and their length distributions are not altered by thymic selection. Perhaps most significantly, these CDR3 length profiles suggest that gamma/delta TCRs are more similar to Igs than to alpha/beta TCRs in their putative ligand binding region, and thus gamma/delta and alpha/beta T cells may have fundamentally different recognition properties.

Four methods for weighting aligned biological sequences have recently appeared that differ mathematically, philosophically, and in their results. Thus, while there is consensus about the need to weight sequences, the method to use is contentious. A geometric analysis based on a continuous sequence space is presented that provides a common framework in which to compare the methods. It is concluded that there are two "best" methods. When the sequences are known to be phylogenetically related and a tree can be generated without introducing excessive stress into the data, the method of Altschul et al. [Altschul, S. F., Carroll, R. J. & Lipman, D. J. (1989) J. Mol. Biol. 207, 647-653] is appropriate. When the sequences are not known to be phylogenetically related or a tree cannot be produced without unduly distorting the distances between the sequences, a modification of the method of Sibbald and Argos [Sibbald, P. R. & Argos, P. (1990) J. Mol. Biol. 216, 813-818] is preferable.

Several known trans-splicing RNA structures were used to define a canonical trans-splicing structure which was then used to perform a computer search of the EMBL nucleotide database. In addition to most known trans-splicing structures, many putative new trans-splicing sites were detected. These were found in a broad range of organisms including the vertebrates. Control experiments indicate that the search predicts known false positives at a rate of only 20%. Trans-splicing may therefore be a very wide-spread phenomenon.