Clathrin heavy and light chain isoforms originated by independent mechanisms of gene duplication during chordate evolution

Abstract

In humans, there are two isoforms each of clathrin heavy chain (CHC17 and CHC22) and light chain (LCa and LCb) subunits, all encoded by separate genes. CHC17 forms the ubiquitous clathrin-coated vesicles that mediate membrane traffic. CHC22 is implicated in specialized membrane organization in skeletal muscle. CHC17 is bound and regulated by LCa and LCb, whereas CHC22 does not functionally interact with either light chain. The imbalanced interactions between clathrin subunit isoforms suggest a distinct evolutionary history for each isoform pair. Phylogenetic and sequence analysis placed both heavy and light chain gene duplications during chordate evolution, 510-600 million years ago. Genes encoding CHC22 orthologues were found in several vertebrate species, with only a pseudogene present in mice. Multiple paralogons surrounding the CHC genes (CLTC and CLTD) were identified, evidence that genomic or large-scale gene duplication produced the two CHC isoforms. In contrast, clathrin light chain genes (CLTA and CLTB) apparently arose by localized duplication, within 1-11 million years of CHC gene duplication. Analysis of sequence divergence patterns suggested that structural features of the CHCs were maintained after gene duplication, but new interactions with regulatory proteins evolved for the CHC22 isoform. Thus, independent mechanisms of gene duplication expanded clathrin functions, concomitant with development of neuromuscular sophistication in chordates.

Fig. 1.

Bayesian phylogenetic tree of CHCs. Animal kingdom phylogenetic tree of CHC proteins, displaying clade probabilities (as percentages of 2,650 replications), with selected outlying species shown. The scale bar shows the number of sequence changes per branch length. D. rerio A, C, D, and E represent sequences found on chromosomes 10, 23, 9, and 11, and sequence B is unassigned in the D. rerio genome.

Fig. 2.

Bayesian phylogenetic tree of clathrin LCs. Animal kingdom phylogenetic tree for LC proteins, displaying clade probabilities (as percentages of 2,650 replications), with selected outlying species shown. The scale bar shows the number of sequence changes per branch length.

Fig. 3.

Paralogous chromosomal regions surrounding CHC genes. Human chromosomal regions 17q12-17q23.2 (Lower Left) and 22q11.21-22q12.1 (Lower Right) contain seven pairs of paralogons for genes encoding ring finger proteins FLJ20315/KIAA1133 (orange), dual-specificity phosphatases MTMR4/MTMR3 (purple), septins PNUTL2/PNUTL1 (blue), partner of pix proteins PPM1E/PPM1F (aqua), yippee-like proteins YPEL2/YPEL1 (yellow), β adaptins AP2B1 and AP1B1 (pink), and CHCs CLTC/CLTD (green). Orthologues of these paralogons (designated by the same colors as the human genes), mapped to one syntenic region in the D. melanogaster (Top), and two syntenic regions in the vertebrate species, are shown. Chromosomes from all species (Upper Left and Right) are on the same scale, with paralogous regions (boxed in red) expanded (Center). M. musculus 11qC is similar to human chromosome 17, whereas human chromosome 22 has orthologues divided between M. musculus 11qA1 and 16qA3. The M. musculus pseudogene similar to human CLTD (ps CHC22) is found on 16qA3, and no orthologue to CLTD is found elsewhere. Asterisks denote genes located by different methodology (see Supporting Text).

Fig. 4.

Site-specific profile for CHC evolutionary rate changes. The posterior probability of functional divergence [P(S₁|X)] was quantitated by using diverge at each amino acid site in an alignment of CHC protein sequences. (A) Posterior probabilities of functional divergence at each sequence position of CHC17 and CHC22. Sixteen residues had posterior probabilities >0.58, which was calculated to indicate significant divergence. The 12 residues circled in red (139, 200, 206, 370, 864, 1382, 1408, 1473, 1494, 1555, 1559, and 1561) are depicted in B, C, and D below. diverge parameters for CHC22/CHC17 are ThetaML = 0.384, AlphaML = 0.283, SE Theta = 0.092, LRT Theta = 17.41. (B) CHC domains are represented on this bar diagram, with the location of residues with significant posterior probabilities of divergence noted in red on the isoform where it is more conserved. The approximate boundaries of each domain are numbered according to residue position in CHC17. Txd, trimerization domain. (C) Residues with predicted functional divergence between CHC17 and CHC22 are mapped onto the crystallographic structure (green) of the CHC proximal leg (27). Residues conserved only in CHC17 are noted in yellow, and those conserved only in CHC22 are noted in orange. (D) Residues with predicted functional divergence between CHC17 and CHC22 are mapped onto the crystallographic structure (green) of the terminal domain with the “clathrin box” peptide bound in its groove noted in purple (31). Residues conserved only in CHC17 are noted in yellow, and those conserved only in CHC22 are noted in orange.

Fig. 5.

Site-specific profile for clathrin LC evolutionary rate changes. The posterior probability of functional divergence [P(S₁|X)] was quantitated by using diverge at each amino acid site in an alignment of LC protein sequences. (A) Posterior probabilities calculated for each LC residue reveal only residue 118 as implicated in functional divergence between LCa and LCb by a calculated posterior probability >0.5. diverge parameters for LCa/LCb are ThetaML = 0.160, AlphaML = 0.471, SE Theta = 0.113, LRT Theta = 1.98. (B) Clathrin LC domains are represented on this bar diagram, with the position of 118 indicated in red on LCb, where it is conserved. The domains are as follows: N, N terminus of LCa; Con, sequence of 100% identity between all mammalian LCa and LCb sequences, which regulates clathrin assembly in vitro; Hsc70bind, binding site for Hsc70 on LCa; Ca, calcium-binding site shared by both LCs; HC bind, CHC-binding region shared by both LCs; Neur, region of neuronal inserts for both LCs; Cam, calmodulin-binding domain shared by both LCs (9). (C) LC-HC interface along the proximal leg (13) with only the interacting portions of each subunit shown. Heavy chain is in green and light chain is in blue, with residue 118 in cyan. Residues 1382 and 1473 are conserved in CHC22 and variable in CHC17 (see Fig. 4).