We have shown that DHH is a highly conserved mammal specific hedgehog paralogue with conserved expression during mammalian gonadogenesis. DHH and its receptors PTCH1 and 2 are highly conserved at the protein level and are expressed in an analogous pattern to that seen in the mouse gonad. However, DHH was expressed in the developing marsupial ovary in contrast to the mouse, in which it is testis-specific during development.
Phylogenetic analysis of the hedgehog gene family across vertebrates shows that non-mammalian DHH genes in fish form a distinct subgroup, distantly related to mammalian DHH genes, indicating they have had an independent evolutionary origin. We have re-named this sub-group fishy hedgehog (FHH) to emphasise their distinction from the DHH genes. This suggests that the evolution of mammalian DHH is a recent event (Figure 2) making it quite unique among the gonadal differentiation genes, all of which have orthologues in the non-mammalian vertebrates with the notable exception of the sex determination switch gene SRY, which is also mammal specific . Despite its recent origin, DHH was extremely highly conserved between marsupials and eutherians, suggesting it quickly adopted an essential function in mammalian reproduction.
The hedgehog receptors PTCH1 and 2 were highly conserved between marsupials, eutherians and non-mammalian vertebrates. Marsupial PTCH2 was the most divergent (especially in the C-terminal region, consistent with findings in other vertebrate species ) but still shared 89% amino acid similarity with eutherian orthologues. The tammar PTCH2 C-terminus contained a 70 amino acid additional exon not found in eutherian PTCH2 proteins. Interestingly, significant homology to the additional tammar exon was identified in the human PTCH2 genomic sequence, in intron 21, and shared 70% identity at the nucleotide level and 79% amino acid similarity with the tammar additional exon (hereafter referred to as exon 21a). The level of conservation of this exon between marsupials and humans was much higher than that of non-functional intronic DNA, suggesting functional conservation of the sequence. Translation of the human sequence revealed a premature stop codon at amino acid 35, so its inclusion in the transcript would lead to a PTCH2 receptor with a severely truncated intracellular signalling domain (Figure 1b). Such an isoform, lacking exon 22, identical to the one predicted from the inclusion of the human putative exon 21a, has been previously identified (the Δ-22 isoform). The human Δ-22 PTCH2 isoform is the only one capable of acting as a strong inhibitor of SHH induction, similar in function to PTCH1 . It appears that the ability to produce such an isoform was derived from a stochastic nonsense mutation in the original exon 21a leading to a truncated protein. The tammar does not have a premature stop (exon 21a is an intact ORF) and so this tammar PTCH2 isoform does not share redundancy with PTCH1 function. The degree of conservation of this region in humans suggests that it has only recently become non-functional in primate evolution. It is intriguing then, that this sequence could not be identified in any other mammalian PTCH2 loci, but only in the tammar, opossum, and human. These findings suggest that the exon was present in the ancestral PTCH2 gene and has been independently lost in different eutherian lineages (Figure 1c). We also identified several PTCH2 isoforms that appear to be dynamically regulated at specific developmental time points. This is also consistent with findings in humans that identified PTCH2 isoforms lacking exons 9 and 10 (PTCH2-Δ9-10). Taken together, these data suggest that PTCH2 has divergent species-specific roles in development, while PTCH1 is likely to maintain a highly conserved function in hedgehog signal transduction. Furthermore, it suggests that the human PTCH2 Δ-22 isoform may have evolved to compensate for a loss of PTCH1 in tissues in which they are co-expressed.
DHH, PTCH1 and PTCH2 mRNA and protein were present throughout gonadal development in both males and females, from early development through to adult stages. The presence of ligand and both receptor proteins throughout gonadal development is consistent with findings in mouse testis, but not ovary  and suggests a conserved role for hedgehog signalling in mammalian gonad formation. These findings are also consistent with the observed disruption to normal gonadal patterning and significant reduction in the expression of the downstream target gene GLI1, in the tammar when hedgehog signalling is ablated in vitro .
In the testis, DHH could be seen within the pre-Sertoli cells of the aggregating cords. Once the testes differentiated, DHH staining was concentrated in the Sertoli cells, especially at the basal lamina of the cord. This protein distribution is similar to that reported in mouse [49, 24, 49] and suggests it is critical for testicular patterning. However, there were some differences in PTCH distribution from predicted mouse patterns. PTCH1 staining was similar to that of DHH, and was distributed mainly within the seminiferous cords (containing germ cells and Sertoli cells). This was in contrast to the interstitial expression seen for Ptch1 in the developing mouse testis  but similar to the expression of Ptch2 . Conversely, PTCH2 staining in the tammar was more reminiscent of Ptch1 distribution in the mouse testis  and was located throughout the gonad but concentrated in the interstitium and Leydig cells. This suggests there may have been a reversal in the roles of these receptors in marsupial testicular development relative to the mouse. Since detailed localisation of the PTCH receptors during gonad development in other mammals and vertebrates is not available we cannot determine which profile is more typical during development.
There was discrete staining of DHH, PTCH1 and PTCH2 proteins in the adult testis. DHH was concentrated in the differentiating germ cells, but restricted to the post pachytene primary spermatocyte stage through to the mature sperm. There was faint PTCH1 and PTCH2 staining throughout the testis but the proteins were concentrated in the Leydig cells and Sertoli cells respectively. This is consistent with in situ results in the adult mouse testis , suggesting a conserved role for these genes in maintaining testicular function and spermatogenesis in all therian mammals.
Unlike in the mouse, in which Dhh is testis-specific in early development , DHH was expressed in the developing tammar ovary throughout development. Activation of hedgehog signalling in the developing mouse ovary leads to Leydig cell development . However, early ovarian development was not affected by the presence of DHH in the tammar, despite the presence of similar mRNA and protein levels of both ligand and receptors as in the developing testis. These findings show that SRY is not needed for DHH activation in the developing gonad. In the juvenile ovary, DHH was abundant in the oocytes consistent with the suggested role for DHH in maintaining the germ line . In the adult ovary, DHH was broadly co-localized with PTCH1 and 2, in follicles and the corpus luteum suggesting it may be important for normal folliculogenesis and steroidogenesis, consistent with recent findings in the mouse . As in the testis, PTCH2 appeared to be the predominant receptor throughout ovarian development and in the adult.