Evolutionary origin and functional divergence of totipotent cell homeobox genes in eutherian mammals

Abstract

Background: A central goal of evolutionary biology is to link genomic change to phenotypic evolution. The origin of new transcription factors is a special case of genomic evolution since it brings opportunities for novel regulatory interactions and potentially the emergence of new biological properties.

Results: We demonstrate that a group of four homeobox gene families (Argfx, Leutx, Dprx, Tprx), plus a gene newly described here (Pargfx), arose by tandem gene duplication from the retinal-expressed Crx gene, followed by asymmetric sequence evolution. We show these genes arose as part of repeated gene gain and loss events on a dynamic chromosomal region in the stem lineage of placental mammals, on the forerunner of human chromosome 19. The human orthologues of these genes are expressed specifically in early embryo totipotent cells, peaking from 8-cell to morula, prior to cell fate restrictions; cow orthologues have similar expression. To examine biological roles, we used ectopic gene expression in cultured human cells followed by high-throughput RNA-seq and uncovered extensive transcriptional remodelling driven by three of the genes. Comparison to transcriptional profiles of early human embryos suggest roles in activating and repressing a set of developmentally-important genes that spike at 8-cell to morula, rather than a general role in genome activation.

Conclusions: We conclude that a dynamic chromosome region spawned a set of evolutionarily new homeobox genes, the ETCHbox genes, specifically in eutherian mammals. After these genes diverged from the parental Crx gene, we argue they were recruited for roles in the preimplantation embryo including activation of genes at the 8-cell stage and repression after morula. We propose these new homeobox gene roles permitted fine-tuning of cell fate decisions necessary for specification and function of embryonic and extra-embryonic tissues utilised in mammalian development and pregnancy.

Keywords: Asymmetric evolution; Embryo; Homeodomain; PRD; Placental; Tandem duplication.

Fig. 1

Presence, absence and duplication of Crx and Crx-derived homeobox genes in placental mammals. The order of relevant genes on human chromosomes 19 and 3 is compared to orthologues in eight other mammals; grey boxes with dashed outline indicate probable pseudogenes, X indicates gene deletion, question marks indicate missing data, dotted lines show unconnected scaffolds. The asterisk in mouse indicates that, in this species, Tprx1 and Tprx2 orthologues are especially divergent and named Crxos and Obox genes, respectively. Many intervening non-homeobox genes are not shown. Human, mouse and guinea pig belong to Euarchontoglires, cow, bat, horse and dog to Laurasiatheria; together, these clades form Boreoeutheria. Elephant and tenrec are Afrotheria within the Atlantogenata

Fig. 2

Additional duplication of Tprx and Leutx. a Maximum likelihood phylogenetic tree of duplicate horse and bat Tprx loci rooted with ferret and dog Tprx genes. b Maximum likelihood phylogenetic tree of duplicate guinea pig, elephant and hyrax Leutx genes with a range of other mammalian Leutx genes. c Genomic organisation of duplicated horse Tprx loci, the majority of which are pseudogenes. d Genomic organisation of duplicated guinea pig Leutx loci

Fig. 3

Evidence for origin of Argfx, Pargfx, Leutx and Tprx genes from Crx. a Maximum likelihood phylogenetic tree using homeodomains plus C-terminal Otx-specific domain. b Position of conserved non-coding elements (CNEs) depicted as red arrowheads, located 3’ to human CRX, TPRX1 and TPRX2, plus two ghost loci (duplicates C and D) inferred to have lost coding sequences. Sequence identity between the CNEs (boxed) is shown by the VISTA plot, which also reveals a single location of the same ancient CNE next to Crx in turtle and coelacanth. VISTA peaks: blue, coding; turquoise, untranslated region; pink, noncoding. Masked repetitive sequences are indicated by khaki segments above the VISTA plot

Fig. 4

Evidence for function from sequence and expression. a Heatmaps showing placental mammal phastCons conservation probability scores for the coding sequences of human CRX and ETCHbox genes; from 0 (blue) to 1 (red). The most consistent conserved (red) stretch codes for homeodomain helix 3. b Heatmaps showing expression profiles in developmental stages and adult tissues of human ETCHbox genes compared to paralogues OTX2 and CRX, and stem cell markers, according to fragments per kilobase per million reads (FPKM) on a log2 scale (red, high expression; blue, low expression). c Heatmaps showing expression profiles in developmental stages and adult tissues of cow ETCHbox genes, according to FPKM on a log2 scale (red, high expression; blue, low expression)

Fig. 5

Nuclear localisation and transcriptional activity. a Confocal images of fibroblasts transfected with V5-tagged constructs or empty vector control, showing nuclei (blue DAPI), ectopic protein (red) and, in the merged images, actin cytoskeleton (green phalloidin). b Numbers of genes up- and down-regulated by ectopic expression and overlaps between responsive gene sets. The clearest overlap is between TPRX1 and LEUTX down-regulated genes. For stringency criteria, see Methods

Fig. 6

Association with human embryo gene expression and antagonism between genes. a Profile 27, out of 106 temporal profiles of human embryonic gene expression, is shown; grey lines are expression plots of 313 individual genes in the profile, red marks the median. Oo, oocyte; Zy, zygote; 2C, 2-cell; 4C, 4-cell; 8C, 8-cell; Mo, morula; Bl, blastocyst. b Genes up-regulated by ectopic ARGFX, or down-regulated by LEUTX or TPRX1, are enriched in profile 27. c Venn diagram showing strong overlap between LEUTX and TPRX1-down regulated genes and ARGFX up-regulated genes, demonstrating the antagonistic effect extends to the full transcriptional response induced by ectopic expression not just the subset in profile 27