Immunoglobulin junctional diversity is concentrated in the third complementarity-determining
region of the heavy chain (CDR-H3), which often plays a dominant role in antigen binding.
The range of CDR-H3 lengths in mouse is shorter than in human, and thus the murine
repertoire could be presumed to be a subset of the human one. To test this presumption,
we analyzed 4751 human and 2170 murine unique, functional, published CDR-H3 intervals.
Although tyrosine, glycine, and serine were found to predominate in both species,
the human sequences contained fewer tyrosine residues, more proline residues, and
more hydrophobic residues (p<0.001, respectively). While changes in amino acid utilization
as a function of CDR-H3 length followed similar trends in both species, murine and
human CDR-H3 intervals of identical length were found to differ from each other. These
differences reflect both divergence of germline diversity and joining gene sequence
and somatic selection. Together, these factors promote the production of a rather
uniform repertoire in mice of tyrosine-enriched CDR-H3 loops with stabilized hydrogen
bond-ladders versus a much more diverse repertoire in human that contains CDR-H3 loops
sculpted by the presence of intra-chain disulfide bonds due to germline-encoded cysteine
residues as well as the enhanced presence of somatically generated proline residues
that preclude hydrogen bond ladder formation. Thus, despite the presumed need to recognize
a similar range of antigen epitopes, the murine CDR-H3 repertoire is clearly distinct
from its human counterpart in its amino acid composition and its predicted range of
structures. These findings represent a benchmark to which CDR-H3 repertoires can be
compared to better characterize and understand the shaping of the CDR-H3 repertoire
over evolution and during immune responses. This information may also be useful for
the design of species-specific CDR-H3 sequences in synthetic antibody libraries.