Functional Coherence of the Human Y Chromosome

See allHide authors and affiliations

Science  24 Oct 1997:
Vol. 278, Issue 5338, pp. 675-680
DOI: 10.1126/science.278.5338.675


A systematic search of the nonrecombining region of the human Y chromosome (NRY) identified 12 novel genes or families, 10 with full-length complementary DNA sequences. All 12 genes, and six of eight NRY genes or families previously isolated by less systematic means, fell into two classes. Genes in the first group were expressed in many organs; these housekeeping genes have X homologs that escape X inactivation. The second group, consisting of Y-chromosomal gene families expressed specifically in testes, may account for infertility among men with Y deletions. The coherence of the NRY's gene content contrasts with the apparently haphazard content of most eukaryotic chromosomes.

Functional or developmental themes have rarely been ascribed to whole chromosomes in eukaryotes. Instead, individual chromosomes appear to contain motley assortments of genes with extremely heterogeneous patterns of developmentally regulated expression. We speculated that the human Y chromosome might be a functionally coherent exception, at least in its nonrecombining portion (the NRY), which makes up 95% of its length (1). It is known to differ from all other nuclear human chromosomes by the absence of recombination, its presence in males only, its common ancestry and persistent meiotic relationship with the X chromosome, and the tendency of its genes to degenerate during evolution (2).

From the 1950s to the present day, many biologists have assumed that the Y chromosome is a functional wasteland, despite the discovery of several NRY genes during this period. Studies of human pedigrees had identified many traits exhibiting autosomal or X-linked inheritance but no convincing cases of Y-linked inheritance (3). In 1959, reports of XO females and XXY males established the existence of a sex-determining gene on the human Y chromosome (4), but this was perceived as a special case on a generally desolate chromosome. The wasteland model has been revised only during the past decade, when eight NRY transcription units (or families of closely related transcription units) were identified, mostly during regionally focused, positional cloning experiments (5-8). Even in recent years, it has been argued that the NRY's gene content is essentially limited to random disintegrating vestiges of its common ancestry with the X (9). The Y-specific repetitive sequences that are so plentiful in the euchromatic regions (10,11) have often been assumed to be functionally inert (12). Realizing that these wasteland theories were based on limited anecdotal data about the NRY's gene content, we decided to embark on a broad, systematic gene hunt that could uncover previously unrecognized functional patterns.

A complete description of the NRY's gene content cannot be obtained with current research methods, short of sequencing the entire NRY. However, it should be possible to obtain a broad representative sampling of NRY genes that could enable us to make comprehensive generalizations. We searched for this sampling in sequences transcribed in a single complex organ, the testis. To assess the suitability of the testis and of a “cDNA selection” protocol (13) for this project, we first sought to crudely measure what fraction of human genes, regardless of developmental regulation, are detectably transcribed there. We did this by testing whether previously identified pseudoautosomal genes (1), whose diversity in developmentally regulated expression is like that of autosomal genes, could be found among testis transcripts. The nine known pseudoautosomal genes were previously identified using mRNA sources as specialized as liver, pineal gland, and skeletal muscle. The extent to which we recovered the nine known pseudoautosomal genes from sampling of testis cDNA would provide a measure of this tissue's adequacy in representing a broad array of genes.

In fact, we recovered testis cDNAs for all nine known pseudoautosomal genes, which suggested that the testis as a single source would be sufficient to provide nearly comprehensive access to NRY genes. From primary, uncloned testis cDNA, we selected and determined the nucleotide sequence of 2539 fragments that hybridized to Y-chromosomal DNA. We anticipated that these sequence fragments would represent a redundant sampling of a much smaller set of genes. Nucleotide sequence analysis revealed that 579 fragments corresponded to known Y genes, including all nine pseudoautosomal genes previously reported and seven of eight known NRY genes. (The one previously reported NRY gene that we failed to recover was AMELY, which is expressed only in developing tooth buds.) After further analysis, both to eliminate human repetitive sequences and to assemble overlapping fragments into contigs, novel sequences were hybridized to Southern (DNA) blots of human genomic DNAs. Sequences that detected at least one prominent male-specific fragment were judged likely to derive from the NRY, and for each we attempted to isolate cDNA clones from a human testis library (13). Nucleotide sequencing of cDNA clones and rescreening of libraries as necessary yielded full-length cDNA sequences for 10 novel NRY genes or families and partial cDNA sequences for two additional ones (Table1). We localized all 12 novel genes on the Y chromosome (Fig. 1) (14) and assessed their expression in diverse human tissues by Northern (RNA) blotting (Fig. 2). The novel genes encode an assortment of proteins (Table 1) and are dispersed throughout the euchromatic portions of the NRY (Fig. 1).

Figure 1

Gene map of NRY. The Y chromosome consists of a large nonrecombining region (NRY; euchromatin plus heterochromatin) flanked by pseudoautosomal regions (yellow). Pter, short-arm telomere; qter, long-arm telomere. The NRY is shown divided into 43 ordered intervals (1A1A through 7) defined by naturally occurring deletions; deletion intervals previously shown to contain Y-specific repeats are shaded blue (10, 11). Listed immediately above the chromosome are nine NRY genes with functional X homologs (red); novel genes are boxed. Immediately below the chromosome are 11 testis-specific genes or families (blue), some with multiple locations. Within deletion intervals, genes have not been ordered. Some testis-specific families probably have members in additional deletion intervals; indicated locations are representative but not necessarily exhaustive. At bottom are shown NRY regions implicated, by deletion mapping, in sex determination, germ cell tumorigenesis (gonadoblastoma), determination of stature, and spermatogenic failure (7, 8,28, 31). For DFFRY, previously thought to be a pseudogene, these mapping studies confirm published findings (19).

Figure 2

Transcription of 12 novel NRY genes in human tissues. Autoradiograms were produced by hybridization of 32P-labeled cDNA probes to Northern blots of polyadenylated RNAs (2 μg per lane) from human tissues (Clontech). Probes used were cDNA clones, either full-length (most genes) or partial (DBY, nucleotides 1476 through 2319 of GenBank no.AF000985; UTY, nucleotides 861 through 1768 of GenBank no. AF000996; and DFFRY, nucleotides 8604 through 9878 of GenBank no. AF000986). Hybridization was done at 65°C in 0.5 M NaiPO4 (pH 7.5) and 7% SDS; washing was done at 65°C in 1× SSC and 0.1% SDS. DBY, TB4Y,EIF1AY, and DFFRY probes cross-hybridize to transcripts derived from their X homologs. For all five X-homologous genes (DBY, PRY, TB4Y,EIF1AY, and DFFRY), expression was tested and confirmed in three male tissues (brain, prostate, and testis) by RT-PCR using Y-specific primers (not shown in figure). ForDFFRY, previously thought to be a transcribed pseudogene, these expression studies confirm published findings (19).

Table 1

Twelve novel genes and gene families in the NRY.

View this table:

Although our gene hunt was systematic, it is likely that some NRY genes in addition to AMELY escaped detection; this could have resulted from failure to select corresponding cDNAs or from discarding them during subsequent screening steps. Like AMELY, other NRY genes may not have been recovered because they are not transcribed in sufficient amounts in the testis. Our screening criteria may have discriminated against NRY genes located in regions of exceptionally high sequence similarity to the X chromosome. In particular, we may have overlooked genes located in a 4-Mb region of the NRY characterized by 99% sequence identity to the X (15). Nonetheless, we suspect that the gene hunt was sufficiently comprehensive for us to form meaningful generalizations about the NRY's gene content.

The 12 novel genes readily sort into two discrete classes (Table 1). The first group, of five novel NRY genes, has several shared features. Each gene has a homolog on the X chromosome encoding a very similar but nonidentical protein isoform; every gene is expressed in a wide range of human tissues; and each gene appears to exist in a single copy on the NRY. The other seven novel NRY genes constitute the second group and share quite different traits. They appear to be expressed specifically in testes. They also seem to exist in multiple copies on the NRY, as judged by (i) the number and intensity of hybridizing fragments on genomic Southern blots or (ii) multiple map locations on the Y. The two classes of genes suggested by our NRY-wide search also accommodate six of eight NRY genes previously identified by less systematic means (5-8, 16), confirming the validity of this bipartite classification.

Many of the X-homologous genes appear to be involved in cellular housekeeping, as suggested by their ubiquitous expression and by the functions of their encoded proteins, which are well established in three cases. TB4Y encodes a Y isoform of thymosin β4, which functions in actin sequestration (17) and which we found to be encoded by the X chromosome.EIF1AY encodes a Y isoform of eIF-1A, an essential translation initiation factor (18). RPS4Y encodes a Y isoform of an essential ribosomal protein (6).

In contrast with these single-copy, X-homologous housekeeping genes, the multicopy NRY gene families appear to encode proteins with more specialized functions. All appear to be expressed specifically in the testis. Our study identified full-length cDNA clones for five of these gene families, which were all found to encode proteins not previously characterized (Table 1). Several of the testis-specific gene families may encode DNA- or RNA-binding proteins, including two small, unrelated basic proteins, BPY1 and BPY2 (Table 1); two putative RNA-binding proteins, RBM and DAZ (7, 8); and CDY, which contains a “chromodomain” [a chromatin-binding motif (Table 1)].

We postulate that the NRY's evolution was dominated by two strategies. The first strategy favored conservation of particular X-Y gene pairs to maintain comparable expression of certain housekeeping functions in males and females. This strategy is at odds with the general behavior of X-Y gene pairs during mammalian evolution. The mammalian X and Y chromosomes evolved from autosomes; most ancestral gene functions were retained on the nascent X chromosome but deteriorated on the nonrecombining portion of the emerging Y chromosome (2). This resulted in females having two copies but males having only one copy of many genes, an inequality predominantly addressed in mammals by transcriptional silencing, or inactivation, of one X chromosome in females. Our findings on X-homologous NRY genes, together with previous studies, suggest the importance in human evolution of an additional solution: preservation of homologous genes on both NRY and X, with male and female cells expressing two copies of such genes. A critical prediction of this model is that the X homologs should escape X inactivation. This is the case for all widely expressed X-linked genes with known NRY homologs, including the X homologs of the five novel NRY genes reported here (6,19-21). A second prediction is that the X- and Y-encoded proteins should be functionally interchangeable despite considerable divergence of their genes' nucleotide sequences. Indeed, each of the eight known X-NRY gene pairs encodes closely related isoforms, with 85 to 97% amino acid identity throughout their lengths; functional interchangeability has been demonstrated in the one case tested to date (22).

These dosage compensation strategies may be relevant to Turner syndrome, which is classically associated with an XO sex chromosome constitution. The Turner phenotype may be due to inadequate expression of certain X-Y common genes that escape X inactivation (23). Given that several X-NRY genes appear to be involved in cellular housekeeping, we speculate that some features of the XO phenotype (such as poor fetal viability) reflect inadequate expression of particular housekeeping functions. The X-homologous NRY genes (Fig. 1) should be investigated as Turner candidates (24).

In addition to the strategy for conserving certain X-Y gene pairs, a second strategy probably shaped the NRY's evolution. This strategy favored the acquisition of testis-specific families, perhaps through selectively retaining and amplifying genes that enhance male reproductive fitness. Animal genomes may contain genes or alleles that enhance male reproductive fitness but are inconsequential or even detrimental to females, as first appreciated by Fisher (25). Fisher recognized that selective pressures would favor the accumulation of such genes in male-specific regions of genomes. Of course, male reproductive fitness depends critically on sperm production, the task of the adult testis. As the only male-specific portion of the mammalian genome, the NRY should have a unique tendency to accumulate male-benefit genes during evolution. Consider the human NRY'sDAZ gene cluster, de novo deletions of which are associated with severe spermatogenic defects (8). The DAZcluster on the human Y chromosome arose during primate evolution by transposition and amplification of an autosomal gene. Similarly, two other testis-specific NRY gene families—RBM andTSPY—may also be the result of the Y chromosome having acquired and amplified autosomal genes (26). We speculate that the selective advantage conferred by the NRY's retention and amplification of male fertility factors (from throughout the genome) accounts for the multitude of testis-specific gene families there. These activities may have been preeminent in shaping the NRY's gene repertoire, because it appears that most NRY transcription units are members of testis-specific families (27). We suspect that most of the NRY's transcription units do not date from the Y chromosome's common ancestry with the X chromosome but instead are more recent acquisitions.

The importance of the human Y chromosome in fertility has been underscored by recent genetic studies. Many men with spermatogenic failure, although otherwise healthy, lack portions of the NRY (7, 8, 28). These findings have suggested the existence of NRY genes that play critical roles in male germ cell development but are not required elsewhere in the body. Previous deletion-mapping studies have implicated four regions of the NRY in either spermatogenic failure or germ cell tumorigenesis, and in each of the four regions we now report novel candidate genes expressed specifically, or most abundantly (29), in testes (Figs. 1and 2).

Although X-homologous and testis-specific genes are somewhat intermingled within the NRY, clustering is evident (Fig. 1). The geographic distribution of the two classes correlates well with previously identified sequence domains within the euchromatic NRY (10, 11). Ten of the 11 known testis-specific families map to previously identified regions of Y-specific repetitive sequences (30). Indeed, one or more testis-specific gene families are found in nearly all known regions of euchromatic Y repeats (Fig. 1). Ironically, it had been widely assumed, partly on theoretical grounds, that these domains consisted of “junk” DNA (12). To the contrary, our results argue that these Y-specific repetitive regions are gene-rich, containing most of the NRY's transcription units (27). We speculate that these were regions of rampant gene amplification during mammalian evolution. By contrast, none of the eight X-homologous genes map to the Y-repeat domains; they all map to regions previously identified as consisting largely of single-copy (or in some cases X-homologous) sequences. We postulate that, earlier in mammalian evolution, these regions of the NRY shared extensive nucleotide sequence identity with the X chromosome.

Although more genes probably remain to be discovered, the 20 genes and families shown in Fig. 1 may constitute the majority of NRY genes, and full-length cDNA sequences are available for 18 of them. The stage is now set for systematic evolutionary, biochemical, and cell-biological studies of this distinctive segment of the human genome.


Stay Connected to Science

Navigate This Article