Gene conversion and the evolution of protocadherin gene cluster diversity

Abstract

The synaptic cell adhesion molecules encoded by the protocadherin gene cluster are hypothesized to provide a molecular code involved in the generation of synaptic complexity in the developing brain. Variation in copy number and sequence content of protocadherin cluster genes among vertebrate species could reflect adaptive differences in protocadherin function. We have completed an analysis of zebrafish protocadherin cluster genes. Zebrafish have two unlinked protocadherin clusters, DrPcdh1 and DrPcdh2. Like mammalian protocadherin clusters, DrPcdh1 has both alpha and gamma variable and constant region exons. A consensus protocadherin promoter motif sequence identified in mammals is also conserved in zebrafish. Few orthologous relationships, however, are apparent between zebrafish and mammalian protocadherin proteins. Here we show that protocadherin cluster genes in human, mouse, rat, and zebrafish are subject to striking gene conversion events. These events are restricted to regions of the coding sequence, particularly the coding sequences of ectodomain 6 and the cytoplasmic domain. Diversity among paralogs is restricted to particular ectodomains that are excluded from conversion events. Conversion events are also strongly correlated with an increase in third-position GC content. We propose that the combination of lineage-specific duplication, restricted gene conversion, and adaptive variation in diversified ectodomains drives vertebrate protocadherin cluster evolution.

Figure 1

Organization of protocadherin gene clusters in zebrafish. Subgroups of paralogous variable exons are indicated by color. Constant region exons are in white. Pseudogenes (ψ) are in gray. (A) Zebrafish Pcdh1α and Pcdh1γ. (B) Partial zebrafish Pcdh2. Sequenced BAC clone names and their locations are shown below each gene cluster.

Figure 2

WebLogo plots of consensus mammalian Pcdh (A), DrPcdh1 (B), and DrPcdh2α (C) core promoter motifs identified using MEME. Mammalian Pcdh and zebrafish Pcdh1 promoter motifs are virtually identical, but the zebrafish Pcdh2α motif has a divergent CGCT box.

Figure 3

Maximum-likelihood phylogenies of zebrafish and human protocadherin cluster proteins. Members of paralog subgroups are indicated by color as in Figure 1. For clarity, subtrees of human protocadherin subgroups are shown as single branches in each tree, except for human C-type protocadherins, which are shown individually in purple. Trees are rooted by midpoint. (A) Protein tree of DrPcdh1α, DrPcdh2α, and human Pcdhα. (B) Protein tree of DrPcdh1γ, human Pcdhβ, and Pcdhγ.

Figure 3

Maximum-likelihood phylogenies of zebrafish and human protocadherin cluster proteins. Members of paralog subgroups are indicated by color as in Figure 1. For clarity, subtrees of human protocadherin subgroups are shown as single branches in each tree, except for human C-type protocadherins, which are shown individually in purple. Trees are rooted by midpoint. (A) Protein tree of DrPcdh1α, DrPcdh2α, and human Pcdhα. (B) Protein tree of DrPcdh1γ, human Pcdhβ, and Pcdhγ.

Figure 4

Ectodomain-specific sequence homogenization in protocadherin cluster genes. The human Pcdhβ ectodomain (EC) 3 gene tree (A) reflects the sequence diversity among Pcdhβ paralogs. Pcdhβ ectodomain 6, however, is almost completely homogenized (B), with otherwise diverse Pcdhβ genes (e.g., β3 and β9) having identical EC6 sequences. This phenomenon is pronounced in zebrafish, where DrPcdh1γ genes with very divergent EC3 domains (C) have nearly identical EC6 domains (D).

Figure 5

Sequence homogenization in protocadherin cluster genes is lineage-specific. The gene tree of human, mouse, and rat Pcdhα (A) and Pcdhβ (C) EC3 domains recapitulates the orthologous relationships among the full-length genes. These relationships break down in Pcdhα EC5 (B) and Pcdhβ EC6 (D) domains, where paralogs within each species are more similar to each other than they are to their orthologs in related species.

Figure 6

The correlation of paralogous sequence diversity at synonymous sites and third-position GC content implicates gene conversion in Pcdh homogenization. Neutral paralog sequence diversity vs. third-position GC content (GC3) is shown for ectodomains 1–6 and the cytoplasmic domain from various protocadherin paralog subgroups. (A) Human, mouse, and rat Pcdhα. (B) Human, mouse, and rat Pcdhβ. (C) Human, mouse, and rat PcdhγA. (D) Human, mouse, and rat PcdhγB. (E) Zebrafish Pcdh1γ. (F) Zebrafish Pcdh1α and Pcdh2α.