The mammalian Y chromosome: molecular search for the sex-determining gene - summary and perspectives

The molecular dissection of the mammalian Y chromosome has been achieved through the application of a variety of approaches. In particular, both molecular and genetic analysis have progressed through the availability of rearranged Y chromosomes. These include both those significant in sexually dysgenic humans and also in the Sxr and Sxr ′ conditions of mice.

Observations on the obligatory pairing and exchange in the ‘pseudoautosomal’ region have been of considerable interest, not only in context of the proposed necessity for such an event in productive meiosis, but also in the ordering of loci with respect to a gradient of recombination within the region (see Burgoyne; Weissenbach et al., this symposium).

The assignment of a locus for testis determination close to the region of active exchange on the human Y chromosome has been reviewed by several contributors. This location for an essential gene in sex determination could be considered to be unexpected and, in this context, it is significant that abnormal exchange between the X and Y chromosomes resulting in the transfer of TDF to the X chromosome occurs at the relatively high frequency of about 10⁻⁴ (see de la Chapelle, this symposium; Bengtsonn & Goodfellow, 1987). In mice, the testis-determining locus is distant from the pseudoautosomal region and, although transfer of Tdy to the X chromosome is observed, such an event is associated with the presence of a rearranged Y chromosome.

Some XX males and some XY females may arise from a similar type of event; that is, the inheritance of either of the reciprocal products of an aberrant paternal X-Y exchange. The comparatively lower frequency of XY females to XX males may reflect an ascertainment bias. Transfer of Y-chromosomal material to the paternal X chromosome of XX males is a frequent, but not an absolute, observation in the aetiology of these cases. However, where such exchanges do take place, there is evidence that they can result in the transfer of a contiguous terminal segment including TDF to the paternal X chromosome. This can occur without a perceptible alteration to the overall length of the recipient X chromosome. Other cases have been described in which unequal crossover events have taken place leading to duplication of terminal regions (Weissenbach et al., this symposium) and one of the X chromosomes in at least some XX males is significantly larger than the other (Ferguson-Smith et al. this symposium).

A wide selection of Y-chromosome-derived sequences has been obtained by random cloning from Y-chromosome libraries. Many of these detect single- or low-copy repeat, Y-specific or Y/autosomal sequences; in addition, a proportion of the probes that have been localized to Yp also exhibits a high degree of homology with a region at Xq21. Pritchard & Goodfellow (this symposium) have described a more direct approach to obtaining Y-chromosomal sequences from specific regions of interest. This has been achieved through chromosome-mediated transfection using thymidine kinase (TK) or antibiotic resistance (neo) as selectable markers followed by fluorescence-activated cell sorting (FACS) for M1C2 (12E7 antigen). Transfectants retaining small Y-chromosomal fragments and which carry flanking markers to the TDF region have provided an enriched source of useful sequences.

Studies in a number of laboratories have allowed a deletion map for the human Y short arm to be deduced. From this map, notwithstanding the uncertainties introduced by the possibility of more complex rearrangements, or by the existence of inversion polymorphisms in the population, it can be estimated that the sequences involved in male determination are located between the proximal boundary of the pseudoautosomal region and the distal boundary of the X/Y homologous region (see Figs 1.2). Indeed, it can be argued that the male-determining locus represents the proximal limit of the region that is normally exchanged (Weissenbach et al. this symposium; Bengtsonn & Goodfellow, 1987). Exchange between the X and Y chromosomes which include the TDF region results in sterile, sexually dysgenic, individuals and therefore the rearrangements will not persist in the population. This provides an effective barrier to the randomization of sequences on the sex chromosomes in the regions below TDF.

Studies of XX males have enabled the identification of probes detecting low-copy Y-specific sequences that map to the interval embracing the testis-determining gene and which are frequently deleted in XY females. These are being investigated by a number of groups as possible starting points for walking, long-range restriction mapping and for the identification of sequence motifs characteristic of transcribed sequences (see Ferguson-Smith et al., Muller, Page et al. and Pritchard & Goodfellow, this symposium).

The presence or absence of a Y-linked gene (or genes) is not the only factor involved in testis determination. Developmental processes in complex organisms depend on the interaction of many genetic and possibly environmental components. Lack of Y-derived sequences in some human XX males and hermaphrodites, coupled with familial inheritance patterns, has led to the postulation of an autosomal testis-determining gene designated TDFA (de la Chapelle, this symposium), which may be homologous to one of the autosomal loci described for mice (see Erickson et al. this symposium). The appearance of Sertoli cells in the testicular cords is a basic event in testis formation; Burgoyne (this symposium) and Singh (this symposium) report that Sertoli cells in XX/XY chimaeras are exclusively XY and, since XXSxr mice are seen to have such cells, it can be postulated that sequences necessary for their development are located in the Sxr region. There have been, however, several reports at this symposium of testis development in the absence of detectable Y sequences and the gene (or genes) necessary for Sertoli cell development could, like those for testis development, (see above) be autoso-mally located but normally regulated by Y-chromosomal sequences (Burgoyne personal communication).

Wiberg and Scherer (this symposium) have presented evidence indicating the presence of a serologically detected H-Y antigen (also known as SDM or Sxs) in true hermaphrodites who lacked detectable Y-genomic sequences. Sxs is normally confined to the heterogametic sex. There is evidence for the location of gene (or genes) responsible for expression of Sxs on human chromosome 6 (Lau, Chan, Kan & Goldberg, 1986) and it is suggested that this may be under the control of X- and Y-linked loci. It is of interest that both this autosomal locus and a testis-determining autosomal locus in mice (see Erickson et al.. this symposium) should be located on chromosomes carrying genes for major histocompatibility determinants (human chromosome 6 and mouse chromosome 17).

The MIC2 gene (see Goodfellow et al., this symposium) represents the only Y-linked gene that has been characterised at the molecular level. It also represents the only known human pseudoautosomal gene. Recent sib testing of the inheritance of 12E7 antigen phenotype, not only confirmed the existence of a further Y-linked gene (previously referred to as Yg). which regulates the level of 12E7 antigen expression, but also provided evidence that this locus can be exchanged with a homologue on the X chromosome by recombination in the pseudoautosomal region. This regulatory sequence on the X chromosome exerts cis control over Xg^- antigen and 12E7 antigen expression on red cells.

A variety of other genes in addition to TDF and MIC2 has been attributed to the human Y chromosome, including those related to the control of growth and spermatogenesis (see Goodfellow, Darling & Wolfe, 1985). Extension of the H-Y transplantation antigen studies (described under mouse sex reversal below) to humans has been achieved through the use of T cell clones specific for H-Y. It now seems clear that the situation for humans is similar to that observed for mice; the locus for testis determination is distinct from that encoding H-Y. Moreover, in humans the two do not appear to be located close together. Whereas the detection of H-Y antigen by the T cell assay is dependent on a locus which is located between intervals 4B and 7 on the published deletion map (see Simpson et al., this symposium), TDF maps to interval 1. Other data presented at this symposium suggest that the H-Y locus may be positioned on Yq. This observation raises the intriguing possibility that it may be identical to the spermatogenesis gene that has been suggested to reside in this region (Tiepolo & Zuffardi, 1976). Again interesting parallels with the mouse Y-chromosome functions are evident (see below). Page (this symposium) has proposed that a pleiotropic effect of the H-Y/ spermatogenesis gene could explain the development of gonadoblastomas, which may occur in dysgenetic gonads. This effect has been attributed to a locus GBY.

There have been several preliminary reports of testis-specific transcripts that are encoded by Y-specific genomic sequences (e.g. see Arnemann et al.. this symposium); however their significance has yet to be fully evaluated. In this context, however, it is worth noting the remarkable detection of two pseudogene sequences (the original genes for which are autosomally located) which detect testis-specific transcripts (Leroy et al., this symposium). This observation suggest that ‘retroposition’ to the Y chromosome may be more active in the testis. The Y chromosome certainly represents a target area with few essential sequences and it is possibly more accessible to colonization by retroviral transposition in the testis.

Burgoyne (this symposium) has reviewed the wide range of phenotypic characteristics that have been attributable to strain-specific variants at Y-chromosomal loci. A minimum of four loci have been postulated (Stewart. 1983; Jutley & Stewart, 1985; Stewart & Jutley, 1987). Arguably the most important recent contribution to arise from studies on sex reversal in mice (Sxr) is the clear distinction between the locus for testis determination (Tdy) and that for H-Y as determined by cytotoxic and proliferative T cell clones (Simpson, 1985; Simpson et al. 1986 & this symposium). A comparison between XOSxr and XOSxr′ mice convincingly demonstrated that, whereas the Sxr region bears both Tdy and H-Y, the Sxr′ region carries only Tdy. Spermatogenesis studies on such individuals and on the testes of XO/XY mosaic mice (Burgoyne, this symposium) have provided evidence for a spermatogenesis gene (Spy) on the Sxr Y-chromosomal fragment, which is excluded in the Sxr′ variant and is expressed in a cell autonomous fashion. Given the paucity of genes on the Y chromosome and the correlation between the loss of spermatogenesis and the absence of H-Y antigen expression, it is tempting to conclude that H-Y antigen is a product of the spermatogenesis gene. It is also plausible that the two functions are encoded by separate but tightly linked genes.

The functional importance of the Sxr region is beyond doubt; attempts to identify genes within this region (Bishop et al., this symposium) have yielded sequences of potential significance. One of these, pCRY8/B, detects a single transcript of approx. 3-5 kb in testis, but does not hybridize to RNA from liver; its presence/absence in Sxr′ will be of considerable interest.

Other Y-genomic sequences have been obtained that detect testis-specific transcripts (e.g. see Bishop et al., this symposium): PY353/B is one such which, although not obviously of retroviral origin, is represented about 250 times and distributed at various positions along the length of the Y chromosome.

Genetic analysis of steroid-sulphatase activity has indicated that the gene necessary for expression of the enzyme is X linked, but only partially inactivated in humans, with no evidence for an expressed Y-chromosomal copy. In contrast, evidence has been presented for X- and Y-localization in mouse with regular exchange between the alleles on the two chromosomes (Keitges, Rivest, Siniscalco & Gartler, 1985). Further support for an pseudoautosomal localization is suggested by its linkage to the most distal X-chromosome marker cream, Crm, (Cattenach & Crocker, 1986) and by its cosegregation with Sxr in an appropriate backcross (Bishop et al., this symposium). Its pattern of linkage to the Crm locus through female meiosis is of interest, as it suggests that recombination in the pseudoautosomal region of female mice is significantly lower than that in males -a similar situation to that observed for the two sexes in humans (Rouyer et al. 1986; Weissenbach, this symposium).

The availability of a cloned probe for the STS structural gene (Ballabio et al. 1987) has allowed us to localize the coding sequences to Xp22.3 (Fraser et al., this symposium). Two additional, Y-chromosomal-specific restriction fragments have been assigned provisionally to Yqll.2. It therefore seems reasonable to speculate that the distribution of the STS sequences in humans reflects a rearrangement from an ancestral sex-chromosome organization in which both X and Y alleles were functional and located in the pseudoautosomal region, as is apparently the situation in mice. However, as no extensive cross-hybridization between human STS coding sequences and rodent DNA has been detected, the nature of the relationship between human and mouse steroid sulphatase genes and their products remains to be established.

The data now available concerning the organization of the mouse and human sex chromosomes can be used to construct a model for the types of rearrangement that would be necessary to explain the different distribution of sequences in the human and mouse Y chromosomes (see Fig. 2). It should be stressed that current evidence indicates a very rapid evolution of the bulk of the Y chromosome. Viral retroposition and adventitious colonization by members of repetitive sequence families is a significant feature. Alterations which may result from these and other Y-chromosome rearrangements may modify the expression of the critical gene functions in a temporal or quantitative manner and may, therefore, be significant in speciation.

Finally, with the molecular pincer movement closing in on the presumptive location of the testis-determining gene on the human Y chromosome and with the rapid dissection of the Sxr region in mice, it is worth speculating what might be at the centre of all this attention. The elegant genetic analysis of sex determination in Drosophila and in nematodes (Nothinger, this symposium; Hodgkin, this symposium) have emphasized the primary signal recognition by a KEY gene which regulates in turn a cascade of subordinate control genes. It appears inescapable that in some systems, at least, sex determination and inactivation mechanisms are inextricably linked, and it would be surprising if a similar situation did not obtain in mammals. The mutation of an autosomal regulatory gene in C. elegans (tra-l) and its potential to give rise to ZZ male/WZ female type of chromosomal sex-determining mechanism when segregating in an appropriate backcross, is a provocative observation. In the case of mammals, two roles can be proposed for the sex-determining sequence(s) on the Y chromosome. It may either be part of a counting system for sequences or regulatory molecules which, on reaching a threshold value, trip a KEY gene into the ON or OFF position; or it may be a gene analogous to one of those in the regulatory cascade seen in simpler organisms. While the exponents of mammalian sex determination still have to consider the various possible functions for TDF (or Tdy), those working on the worm and fly may have the appropriate sequences cloned and available for comparison before the mammalian chromosome walkers arrive at their destination.

With grateful thanks to Peter Goodfellow, Mark Jobling, Mark Ross and Thomas Meitinger for helpful discussions and other contributions to the preparation of the manuscript.