Origin and Loss of Nested LRRTM<i>/</i>α-Catenin Genes during Vertebrate Evolution

<div><p>Leucine-rich repeat transmembrane neuronal proteins (LRRTMs) form in mammals a family of four postsynaptic adhesion proteins, which have been shown to bind neurexins and heparan sulphate proteoglycan (HSPG) glypican on the presynaptic side. Mutations in the genes encoding LRRTMs and neurexins are implicated in human cognitive disorders such as schizophrenia and autism. Our analysis shows that in most jawed vertebrates, <i>lrrtm1, lrrtm2,</i> and <i>lrrtm3</i> genes are nested on opposite strands of large conserved intron of α-catenin genes <i>ctnna2, ctnna1,</i> and <i>ctnna3</i>, respectively. No <i>lrrtm</i> genes could be found in tunicates or lancelets, while two <i>lrrtm</i> genes are found in the lamprey genome, one of which is adjacent to a single <i>ctnna</i> homolog. Based on similar highly positive net charge of lamprey LRRTMs and the HSPG-binding LRRTM3 and LRRTM4 proteins, we speculate that the ancestral LRRTM might have bound HSPG before acquiring neurexins as binding partners. Our model suggests that <i>lrrtm</i> gene translocated into the large <i>ctnna</i> intron in early vertebrates, and that subsequent duplications resulted in three <i>lrrtm/ctnna</i> gene pairs present in most jawed vertebrates. However, we detected three prominent exceptions: (1) the <i>lrrtm3/ctnna3</i> gene structure is absent in the ray-finned fish genomes, (2) the genomes of clawed frogs contain <i>ctnna1</i> but lack the corresponding nested (<i>lrrtm2</i>) gene, and (3) contain <i>lrrtm3</i> gene in the syntenic position but lack the corresponding host (<i>ctnna3</i>) gene. We identified several other protein-coding nested gene structures of which either the host or the nested gene has presumably been lost in the frog or chicken lineages. Interestingly, majority of these nested genes comprise LRR domains.</p>