Evolution of coding and non-coding genes in HOX clusters of a marsupial

Abstract

Background: The HOX gene clusters are thought to be highly conserved amongst mammals and other vertebrates, but the long non-coding RNAs have only been studied in detail in human and mouse. The sequencing of the kangaroo genome provides an opportunity to use comparative analyses to compare the HOX clusters of a mammal with a distinct body plan to those of other mammals.

Results: Here we report a comparative analysis of HOX gene clusters between an Australian marsupial of the kangaroo family and the eutherians. There was a strikingly high level of conservation of HOX gene sequence and structure and non-protein coding genes including the microRNAs miR-196a, miR-196b, miR-10a and miR-10b and the long non-coding RNAs HOTAIR, HOTAIRM1 and HOXA11AS that play critical roles in regulating gene expression and controlling development. By microRNA deep sequencing and comparative genomic analyses, two conserved microRNAs (miR-10a and miR-10b) were identified and one new candidate microRNA with typical hairpin precursor structure that is expressed in both fibroblasts and testes was found. The prediction of microRNA target analysis showed that several known microRNA targets, such as miR-10, miR-414 and miR-464, were found in the tammar HOX clusters. In addition, several novel and putative miRNAs were identified that originated from elsewhere in the tammar genome and that target the tammar HOXB and HOXD clusters.

Conclusions: This study confirms that the emergence of known long non-coding RNAs in the HOX clusters clearly predate the marsupial-eutherian divergence 160 Ma ago. It also identified a new potentially functional microRNA as well as conserved miRNAs. These non-coding RNAs may participate in the regulation of HOX genes to influence the body plan of this marsupial.

Figure 1

Chromosomal locations of tammarHOXgenes by FluorescenceIn-SituHybridization. Tammar HOX genes were mapped to four different chromosomal loci. BAC DNA was hybridized to metaphase chromosomes from a male donor stained with DAPI (blue). The hybridization signal was indicated with anti-DIG-FITC (bright green). HOXA was on the long arm terminal region of chromosome 3; HOXB was located 2/3 of the distance from the centromere on the long arm of chromosome 2; HOXC was on the middle of long arm at chromosome 3; HOXD was on the middle of long arm at chromosome 5.

Figure 2

Organization ofHOXgene clusters, long non-coding RNAs and microRNAs in human and tammar. The tammar has 39 HOX genes located in 4 separate clusters—HOX-A, -B, -C and -D—which show highly conserved organization. Three conserved long non-coding RNAs (HOXA11AS, HOTAIRM1 and HOTAIR) were also present. Orthologous genes are the same color and introns filled with the grey color. The homologous long non-coding RNAs are the same color in the star while homologous microRNAs are also the same color. Question mark (?) represents the predicted microRNAs by sequence alignment.

Figure 3

TammarHOXgene expression in adults. Tammar HOX gene expression pattern were examined in 23 adult tissues including brain, cerebellum, hypothalamus, pituitary, pancreas, spleen, stomach, intestine, cecum, heart, liver, lung, mammary gland, skeletal muscle, kidney, adrenal, testis, epididymis, ovary, uterus and prostate with 39 HOX genes primers (Additional file 12). A, HOXA; B, HOXB; C, HOXC; D, HOXD; 18S, housekeeping gene and positive control.

Figure 4

Conserved miRNA and long non-coding RNAs analysis in theHOXA cluster. The conserved long non-coding RNAs, HOXA11AS and HOTAIRM1, and microRNA miR-196b were shown by mVISTA with comparison of mouse, tammar and frog against human HOXA cluster genomic sequence. The coding genes HOXA13, HOXA10, HOXA9, HOXA2 and HOXA1 are highly conserved in all species. Expression of tammar long non-coding RNAs in bone marrow and endometrium were confirmed by RT-PCR. The blue stands for coding regions, and the green for non-coding RNA regions whilst the pink represents conserved coding sequences. F, forward primer, R, reverse primer.

Figure 5

Sequence conservation in theHOXC cluster in tammar, human, mouse and frog. mVISTA plot of HOXC genomic sequences from tammar, human (chr12:54332691–54396455), mouse (chr15:102751619–102814560) and frog (scaffold_226:281324–390491). The sequence similiarity (50–100%) (vertical axis) is shown in the coordinates of the genomic sequence (horizontal axis) from human, mouse and frog. Genes and their orientation are indicated by grey arrowed line. Exons of genes are indicated by blue solid boxes. Conserved regions above the level of 70%/100 bp are highlighted under the curve, with red indicating conserved non-coding regions, blue representing conserved coding-protein exons, and turquoise representing microRNAs or long non-coding protein exons. The long non-coding RNA HOTAIR located between HOXC12 and HOXC11 was conserved in all mammals and had a much lower conservation in frog. MicroRNA miR-196a2 is extremely highly conserved in all examined species. RT-PCR performed in the tammar with a single band at 81 bp confirmed the presence of the long non-coding RNA HOTAIR providing further evidence of the conservation. In addition, both microRNA miR-196a2 was expressed in tammar cells, verifying the existence of this microRNAs in tammar

Figure 6

Comparative genomic analysis ofHOTAIRM1orthologues in mammals. The genes flanking HOTAIRM1, HOXA1 and HOXA2, from the human genome (chr7:27,132,617–27,142,393; http://genome.ucsc.edu), are shown along with their conservation score (phylop). HOTAIRM1 gene structure consists of three exons in eutherian mammals, but two exons in the tammar (lower left), based on predicted RNA secondary structure and sequence alignment. Phylogenetic trees showing that exon 1 is highly conserved with short genetic distance between them compared to exon 2 and exon 3 consistent with the concept of rapid evolution of non-coding RNAs (lower right).

Figure 7

Evolutionary relationships ofHOXA11AS orthologues. The genes flanking HOXA11AS, HOXA11 and HOXA13, in the human (chr7:27,220,777–27,239,725; http://genome.ucsc.edu) are shown along with their conservation score (phylop). HOXA11AS gene structure consists of two exons in eutherian mammals except mouse, but one exon in tammar (lower left), based on predicted with RNA secondary structure and sequence alignment. Phylogenetic trees showing exon2 is highly conserved in eutherian mammals whilst exon1 is more divergent with the full predicted exon 1 sequence, consistent with the mammalian consensus sequences in the top panel (lower right).

Figure 8

Evolutionary relationships of HOTAIR orthologues. The genes flanking HOTAIR, HOXC11 and HOXC12, in the human genome (shr12:54,348,714–54,370,201; http://genome.ucsc.edu) are shown along with their conservation score (phylop). HOTAIR gene structure consists of 6 exons in the eutherian mammals, except mouse and rat, which have 5 exons. In contrast, only 3 exons were found in tammar. Phylogenetic trees based on exons 4–6 (lower right).

Figure 9

Newly discovered miRNA meu-miR-6313in tammar. A) Centroid secondary structure with a minimum free energy of −43.66 kcal/mol; the bar from blue to red represents base-pair probabilities from low (0) to high (1); B) the reads, precursor and secondary structure of new miRNA; C) sequence alignment of miRNA and precursors in tammar, human, mouse and frog.