The dynamics of vertebrate homeobox gene evolution: gain and loss of genes in mouse and human lineages

Abstract

Background: Homeobox genes are a large and diverse group of genes, many of which play important roles in transcriptional regulation during embryonic development. Comparison of homeobox genes between species may provide insights into the evolution of developmental mechanisms.

Results: Here we report an extensive survey of human and mouse homeobox genes based on their most recent genome assemblies, providing the first comprehensive analysis of mouse homeobox genes and updating an earlier survey of human homeobox genes. In total we recognize 333 human homeobox loci comprising 255 probable genes and 78 probable pseudogenes, and 324 mouse homeobox loci comprising 279 probable genes and 45 probable pseudogenes (accessible at http://homeodb.zoo.ox.ac.uk). Comparison to partial genome sequences from other species allows us to resolve which differences are due to gain of genes and which are due to gene losses.

Conclusions: We find there has been much more homeobox gene loss in the rodent evolutionary lineage than in the primate lineage. While humans have lost only the Msx3 gene, mice have lost Ventx, Argfx, Dprx, Shox, Rax2, LOC647589, Tprx1 and Nanognb. This analysis provides insight into the patterns of homeobox gene evolution in the mammals, and a step towards relating genomic evolution to phenotypic evolution.

Figure 1

Chromosomal distribution of human DUX4-like genes and mouse 'chromosome 10' Dux Genes. (A) Twenty-three human DUX4-like sequences include four loci with predicted introns interrupting the coding sequence (orange arrows) and nineteen intronless sequences (grey arrows), most clustering on chromosomes 4, 10 and Y. One putative intron-containing locus at chromosome 10q26.3 is disrupted by a translocation or is the remnants of two loci. (B) The Dux gene on mouse chromosome 10 has been duplicated in tandem to generate three loci. Orientation of arrows indicates direction of transcription; small arrowheads indicate putative introns.

Figure 2

DUXB and CPHX loci in mouse and human. (A) Mouse DUXB-like (Duxbl, orange arrows) and Cphx (red arrows), together with a non-homeobox gene Plac9 (green arrows), have been duplicated giving three copies each at chromosome 14A3. (B) Human Cphx-related homeobox loci flank a DUXBL pseudogene and the DUXB gene on chromosomes 10 and 16 respectively, although the loci on chromosome 10 are disrupted (ragged boxes). Anxa11 (blue arrows) is not a homeobox genes. (C) Amino acid alignments; dots represent identical amino acids, dashes deletons/insertions and red characters conservative substitutions. Hs, Homo sapiens; Mm, Mus musculus.

Figure 3

Presence or absence of the NANOGNB gene in mammalian genomes. (A) The deduced homeodomain sequence of human NANOGNB showing predicted alpha helical regions, compatible with folding into a homeodomain tertiary structure. (B) NANOGNB is located just 15 kb from the human NANOG gene at chromosome 12p13.31 (orange arrows). Orthologous genes at the syntenic position are present in horse and dog, but not in mouse and rat. Grey arrows indicate non-homeobox genes.

Figure 4

Absence of LOC647589 in the mouse genome. LOC647589 was detected in this study at human chromosome 12q24.33. No annotation is found in genomes outside of human currently, although a homologous sequences is present in cow (chr17:46,768,917-46,783,902) and other mammals with available genome sequences, but not in mouse or rat. The orange arrow and box indicate homeobox loci; grey arrows indicate non-homeobox genes.

Figure 5

Rhox (reproductive homeobox) gene clusters. (A) Three Rhox loci are present in the human genome at chromosome Xq24. (B) Thirty-six Rhox loci are clustered in the mouse genome at chromosome X A3.3. Rhox3-ps (white arrow) has a stop codon in the homeobox.

Figure 6

Origin of the LEUTX gene in the human genome. The LEUTX homeobox gene (orange arrow) plus four unusual non-homeobox genes (light grey arrows) form a linked set of genes in the human genome (dashed box). These genes are not present at the syntenic location in mouse, rat, cow or dog genomes.

Figure 7

The mouse Obox loci and associated genes. There are 36 mouse Obox homeobox loci, comprising 6 intron-containing genes (orange arrows) and 28 intronless loci (unfilled arrows) on chromosome 7, plus intronless probable pseudogenes on chromosome 17. Four other homeobox genes, not clearly part of the Obox family, are linked to the large Obox cluster: Crxos1, Gm5585 and Gm7235 and Crx (red arrows).

Figure 8

The MSX3 gene has been lost in human and dog genomes. The Msx3 gene is located in syntenic regions of mouse, rat and pig genomes (orange arrows), but absent at the equivalent location in human and pig. Non-homeobox genes used as indicators of chromosomal synteny are shown as grey arrows.

Figure 9

The Ventx gene has been lost in rodent genomes. The VENTX gene is located in syntenic regions of human, rat, dog and chicken genomes (orange arrows), but absent at the equivalent location in mouse and rat. Non-homeobox genes used as indicators of chromosomal synteny are shown as grey arrows.

Figure 10

The ARGFX locus has been lost in rodent genomes. The ARGFX gene is located in syntenic regions of human and cow (orange arrows), but absent at the equivalent location in mouse and rat. Non-homeobox genes used as indicators of chromosomal synteny are shown as grey arrows.

Figure 11

The DPRX gene has been lost in rodent genomes. The DPRX gene is located in syntenic regions of human, dog and horse (orange arrows), but absent at the equivalent location in mouse and rat. Non-homeobox genes used as indicators of chromosomal synteny are shown as grey arrows.

Figure 12

The SHOX gene has been lost in rodent genomes. The SHOX gene is located in syntenic regions of human (X and Y chromosomes) and chicken (orange arrows), but absent at the equivalent location in mouse and rat. Non-homeobox genes used as indicators of chromosomal synteny are shown as grey arrows.

Figure 13

The RAX2 gene has been lost in rodent genomes. The RAX2 gene is located in syntenic regions of human and chicken (orange arrows), but absent at the equivalent location in mouse and rat. Non-homeobox genes used as indicators of chromosomal synteny are shown as grey arrows.

Figure 14

Tprx loci are not present in rodent genomes. In the human genome, the Tprx family contains one probable functional gene (TPRX1) and a possibly non-functional tandem duplicate (TPRX2P) at chromosome 19q13, either side of the Otx-family CRX gene. The additional Tprx loci, TPRX1P1, TPRX1P2 and TPRXL, are elsewhere in the genome and are not shown. Examination of the region syntenic to human 19q13 in other mammals reveals clear orthologues of Tprx family loci in cow and dog, but not mouse and rat. In rodents, another homeobox gene Crxos1 is found. Orange arrows indicate homeobox genes; grey arrows are non-homeobox genes.

Figure 15

A summary of homeobox gene dynamics in the mouse and human evolutionary lineages. The majority of homeobox genes are conserved between mouse and human lineages (grey squares), although some have undergone duplication to different extents (cascaded boxes). Humans have lost the Msx3 gene; mice have lost VENTX, ARGFX, DPRX, SHOX, RAX2, LOC647589, NANOGB and TPRX1 (dashed boxes). Three new homeobox loci (Gm7235, Gm5585 and Crxos1) and one new cluster (Obox) arose in the rodent lineage; one new gene Leutx arose in the lineage leading to primates.