Testis-specific transcripts detected by a human Y-DNA-derived probe

The putative sex-determining gene(s) on the human Y(TDF) is localized on Yp (Goodfellow, Darling & Wolfe, 1985), but in the mouse the corresponding gene TDY is localized to the pericentric region of the chromosome (Eicher & Washburn, 1986). In both man and mouse, evidence for the presence of additional male-specific genes has been reported, such as genes involved in spermatogenesis (Tiepolo & Zuffardi, 1976; Levy & Burgoyne, 1986). In an attempt to find such genes on the human Y-chromosome, random DNA sequences obtained from a human Y-specific cosmid library (Bishop et al. 1983) were studied by Northern blot analysis of human RNA. Two probes, derived from the long arm of the Y chromosome, were found to detect mRNAs specifically expressed in the testis: probe 49F previously described (Seboun et al. 1986) and probe 12f3 reported here. These probes are probably derived from pseudogenes on the Y chromosome. Because mRNAs detected by them may be involved in spermatogenesis, they may represent useful tools for studying this poorly defined process at a molecular level. They may also be of use for understanding the evolution of the Y chromosome.

Approximately 50 random Y genomic sequences, isolated from a human Y-chromosome library (Bishop et al. 1983), were used to probe Northern blots of human testis poly(A)⁺ RNA and several other male and female human tissues. In this way, probe 12f3 was shown to detect an mRNA expressed in human testis but not in other tissues tested.

Probe 12f3 is a 1·7 kb Bgl II fragment which with the flanking sequences 12f 1 and 12f2 was subcloned from the genomic probe 12f (Fig. 1A), a 5·1 kb EcoRl fragment derived from cosmid 12 (Fig. IB).

Southern blot analysis with human DNA digested by Eco RI (Fig. 2A) reveals that, under either stringent or nonstringent conditions, 12f3 detects the 5·1 kb cognate fragment which can be assigned to the Y chromosome due to its increased hybridization intensity on the 49X,YYYY line. In addition, the probe detects several homologous sequences which cannot be localized to the Y or the X (as shown by no increased hybridization signal on the 49XXXX.Y line (Fig. 2B)) and are therefore presumed to be autosomal. Using a Y-deletion panel (data not shown) 12f3 has been mapped to the long arm of the human Y chromosome, between Yq 11.22 and Yq 11.23.

In the chimpanzee, the Southern blot profile (using DNA digested by Eco RI) is very similar to the human (Fig. 3). A male-specific band of 5·1 kb can be detected and as this band has the same size as the Y-located band of the human, it is probably also a Y-specific band.

Human Northern blot analysis with 12f3 detects a 1·6 kb mRNA in human testis, but not in other tissues tested, including ovary (Fig. 4). Occasionally, a 3·8 kb transcript which is not testis specific can be weakly detected in several tissues. Using mouse mRNA. the same specificity is also obtained but Southern analysis reveals only one hybridization band of approximately 7·2 kb on both male and female mouse DNA digested by EcoRI.

Due to the fact that 12f3 cross-reacts with the mouse and the ready availability of these tissues, we chose to investigate the function of 12f3 in this species. Approximately 7×10⁴ plaques from an amplified adult mouse testis cDNA library (Bishop & Hatat, 1987) were screened with 12f3 and eight positive plaques were purified. The longest cDNA obtained, designated PL7, is 1·4kb and contains a repetitive sequence at one of its extremities (probably the 3’ end). Further analyses were carried out using the shorter overlapping clone PL5, a 0·8 kb cDNA which is included in the restriction map of PL7 and which does not carry the repeated element (Fig. 5). We have neither determined the relationship between the two transcripts detected by 12f3 nor from which of them these cDNAs have been synthesized.

The restriction maps of PL7 and PL5 are shown in Fig. 5. The repetitive sequence present in PL7 has been localized to the left end of the map. Fig. 6 shows the hybridization pattern observed on a Southern blot of human DNA, hybridized consecutively with the human genomic probe 12f3 and the mouse cDNA PL5. Under low-stringency conditions, 12f3 and PL5 detect the same bands but with different intensities; the very low intensity of the 5·1 kb Y-specific band observed with PL5 suggests that the mRNA is probably expressed from an autosomally located gene.

Probing Eco RI-digested mouse DNA with PL5 (Fig. 7A) detects, under either nonstringent or stringent conditions, only one band of 7·2 kb present on both male and female (the same profile that is obtained using human probe 12f3). No mouse signal is obtained with a hybrid cell line containing only the mouse chromosomes X and 16 on a Chinese hamster background. Thus, we can exclude the possibility of an X–Y common sequence. Southern blot analysis of DNA from several vertebrate species shows that PL5 corresponds to a putative gene which is highly conserved from reptiles to mammals (with the possible exception of the viper) (Fig. 8).

By Northern blot analysis (Fig. 7B). PL5 detects (like 12f3) a 1·6 kb mRNA highly expressed in testis (both mouse and human) but not in numerous other tissues tested, including ovary. In contrast, the 3·8 kb mRNA is present in all tissues tested but apparently at different levels. Partial sequence analysis of PL7 has revealed an open reading frame of at least 500 bp. Searching the Los Alamos data bank with 50bp of this sequence did not detect any significant homologies at the 72 % level.

A second category of cDNA (represented by PL10). which only hybridizes weakly to the probe 12f3, was obtained from the adult mouse testis cDNA library. When this 2·1 kb cDNA is used to probe mouse DNA restricted by Eco RI. a single hybridization band of 6· 8 kb can be seen on both male and female (Fig. 9A).

By Northern blot analysis, PL10 detects an approximately 4 kb mRNA which is highly expressed in mouse testis but not in male or female liver (Fig. 9B). Occasionally, a 4·3 kb mRNA can be detected in testis and liver.

PL10, like PL5, is highly conserved among vertebrates (data not shown) suggesting a role of fundamental importance. At present, we do not know the exact relationship between PL10 and PL5.

Initial results from in situ hybridization to mouse metaphase chromosomes (unpublished observations) indicates that PL5 is located in the central region of chromosome 13 and PL10 in the telomeric region of chromosome 1. Under nonstringent conditions, a secondary hybridization site is detected with PL 10 on the X chromosome.

Preliminary in situ hybridizations to sheep testis tissue sections, using probe 12f3, shows a high expression at late spermatid stage. These preliminary results are consistent with the RNA playing a role during the postmeiotic stages of spermatogenesis.

A random human sequence, 12f3, derived from the long arm of the Y chromosome, detects an abundant testis-specific transcript of 1·6 kb in both man and mouse. Using the mouse cDNA PLS, isolated with 12f3, we conclude that the L2f3 sequence does not represent the active gene, which is probably auto-somally located. This conclusion is drawn from a comparison of human Southern blot patterns obtained with 12f3and PL5. Under stringent conditions, in contrast to the genomic probe 12f3, the intensity of the Y-located band is drastically reduced. We cannot rule out the possibility, however, that PL5 is derived from the non-testis-specific transcript (3·8 kb) and thus that the 1·6 kb transcript could be expressed from the human Y chromosome. But, in mouse, in situ hybridizations to metaphase chromosomes indicate that these transcripts originate from a single region of mouse chromosome 13. In addition, results obtained by Southern blot analysis using mouse DNA suggest that these transcripts are synthesized from a single-copy gene. Possibly the 1·6kb mRNA is obtained by a testis-specific splicing. Moreover, in the different vertebrates species tested (except for primates and man), a single band common to both male and female is detected on Southern blotting. Thus, we propose that, in lower vertebrates, there is one coding gene, autosomally located, which has spread into both autosomes and the Y before the separation of the chimpanzee and human but after rodent and primate divergence. These results are similar to those obtained with the human Y genomic probe 49F, which has been localized to the same region of the Y chromosome as 12f3 (Seboun et al. 1986).

We do not know the mechanism of the genomic dispersion. DNA sequences 12f 1 and 12f2 which flank 12f3 (Fig. 1 A) do not detect transcripts (Seboun et al. 1986). Moreover, in cosmid 12 (the cosmid from which 12f was subcloned and in which 12f is at least 5kb from one end (Fig. IB), PL5 detects no homologous sequences other than 12f3. As the RNA is expressed during spermatogenesis, reintegration by a mechanism of retroposition in the germ line is possible and would give rise to such hereditary events. Several examples of pseudogenes found in the human Y chromosome have been described, such as the housekeeping genes arginosuccinate synthetase and actin (Daiger, Wildin & Su, 1982; Heilig et al. 1984). The essentially nonfunctional nature of the Y chromosome may explain the easy reintegration of sequences in this chromosome.

The two cDNAs are both highly conserved among vertebrates; although, no signal was observed with the viper. We have not yet tested amphibian, fish or invertebrates. The conservation observed suggests an important role for these genes among higher vertebrates. Preliminary histological localization of the transcripts to the germ line using 12f3 suggests a role during the haploid stages of spermatogenesis. Northern blot analysis, using ovarian poly(A)⁺ RNA, failed to detect the 1·6kb transcript but the number of germ cells is much lower than in the testis and perhaps this type of analysis is not sensitive enough. In addition, the stage during which transcripts are detected in the male germ line would correspond to a stage reached by the oocytes after fertilization and thus not possible to detect in an ovary. This ambiguity could possibly be resolved by in situ studies on ovarian sections of fertilized oocytes. Finally, we do not know the exact relationship between PL5 and PL10. Although both were identified using the same genomic probe, PL5 reacts much more strongly with 12f3 than PL10. In addition, PL5 and PL10 are only weakly cross-reactive. Neither the genomic sequence nor the two cDNAs cross-react with the 49F sequence which also detects a germ line transcript (Seboun et al. unpublished result).

We wish to thank Drs Andreas Weith and David Sassoon for their helpful comments in preparing this manuscript.