Sex-Chromosomal Differentiation and Amelogenin Genes in Mammals

Paralogous gene pairs on the mammalian sex chromosomes can provide useful information on the tempo and mode of X-Y differentiation. Recently, Lahn and Page (1999) reported intriguing results on the chromosomal locations and synonymous sequence divergences (ks with multiple-hit correction) of 19 pairs of X-Y-paralogous cDNA sequences in humans or squirrel monkeys. It was found that there are four distinct groups (strata) of the gene pairs in terms of the ks value. Of these, seven belong to stratum 4, and the X paralogs are clustered in a single chromosomal region which is adjacent to the pseudoautosomal boundary in the short arm of chromosome X. In contrast, the seven Y paralogs are scattered on chromosome Y, presumably owing to frequent chromosomal rearrangements. Together with similar observations on strata 1, 2, and 3, it was hypothesized that X-Y-chromosomal differentiation was initiated one stratum at a time and that stratum 1 was the first to embark on the differentiation and stratum 4 was the most recent. Lahn and Page (1999) invoked chromosomal inversions as the most likely cause for formation of these strata by suppressing sex-chromosomal recombination. Bengtsson and Goodfellow (1987) argued that X-Y differentiation had occurred only after X-Y recombination ceased and that the most likely mechanism of suppressing recombination was chromosomal inversions. Navarro et al. (1997) theoretically demonstrated that there should be a strong reduction in recombination in the proximal zone of an inverted region owing to the production of unbalanced gametes. Here we report that the junction between stratum 3 (ks = ∼30%) and stratum 4 (ks = ∼10%) in the human and the chimpanzee resides in the second intron of the amelogenin gene, arguing that this junction cannot be a direct breakpoint of any chromosomal inversion. We also make a similar comparison of amelogenin genes in other mammals and show that formation of stratum 4 is specific to mammalian orders. Eutherian mammals generally possess a pair of amelogenin genes on the X and Y chromosomes (e.g., Lau et al. 1989 ; Nakahori, Takenaka, and Nakagome 1991 ; Salido, Yen, and Koprivnikar 1992 ) which are designated AMGX and AMGY, respectively. In humans, the AMGX locus is oriented from the centromere to the telomere. Among the seven loci examined for stratum 4, the AMGX locus is most proximal and nearest to stratum 3. We retrieved two human BAC clone sequences which contained AMGX (AC002366) and AMGY (AC013412) and also determined the genomic DNA sequences of the chimpanzee homologs (data will be presented elsewhere). Quite unexpectedly, it turns out that the upstream region from intron 2 exhibits 25% sequence differences per silent site (ps without multiple-hit corrections), the same level as the corrected ks value of 30% in stratum 3 (fig. 1 ). In sharp contrast, the region downstream of intron 2 exhibits ps = 10%, as in intron 3 (Huang et al. 1997 ) as well as the six other paralogous gene pairs belonging to stratum 4 (Lahn and Page 1999 ). The same pattern and degree of sequence differences were also found in the comparison between the chimpanzee AMGX and AMGY. These findings convincingly indicate that the junction between strata 3 and 4 arranged side by side on chromosome X occurs in the middle of the human and chimpanzee AMGX locus. The first and second exons of the amelogenin gene almost exclusively encode the 5′ untranslated region so that the cDNA or amino acid sequences represent the feature of stratum 4. Accordingly, the phylogenetic analysis showed the sister relationship of X and Y paralogs within both cattle and pigs (Gibson et al. 1991, 1992 ; Hu et al. 1996 ; Chen et al. 1998 ; Girondot and Sire 1998 ; Toyosawa et al. 1998 ). The synonymous differences are about 6% in cattle, which is not significantly different from that in humans. However, since the promoter region and exon 1 in the cow AMGX and AMGY differ by ps = 30% (Chen et al. 1998 ), there is again an abrupt reduction in the extent of sequence differences, as in humans (fig. 1 ). To gain some insight into the position in the step reduction, we determined the genomic DNA sequences of the cattle paralogs. A preliminary result showed that the reduction occurred at almost the same position in intron 2 as in humans. It is also to be noted that the cDNA sequence differences (ps = 3%) between the pig AMGX and AMGY were significantly lower than those in humans and cattle. More extreme are the horse AMGX and AMGY genes, in which no nucleotide difference is found (Hasegawa et al.2000) . Since the genomic DNA sequences of the pig and horse AMGX and AMGY are unavailable at present, it is not clear to what extent they differ from each other in the upstream region of the gene. Yet, the clocklike accumulation of synonymous differences suggests that differentiation of AMGX and AMGY occurred relatively recently in the pig and has not yet occurred in the horse. Thus, the initiation timing of stratum 4 differentiation differs greatly among mammalian orders, being most ancient in humans (Primates) and most recent in horse (Perissodactyla). Rodents are exceptional in that they may not possess AMGY at all: it is either deleted or altered to an undetectable level (Snead et al. 1985 ; Lau et al. 1989 ; Bonass et al. 1994 ). In any case, these amelogenin sequences suggest that the downstream region of the gene, and probably stratum 4 as a whole, began to differentiate after the mammalian radiation. On the contrary, X-Y differentiation in stratum 3 must have occurred well before the mammalian radiation (Lahn and Page 1999 ). This is supported by the relatively large sequence differences between X and Y paralogs in stratum 3, such as those between ZFX and ZFY (Schneider-Gädicke et al. 1989 ). As argued by Bengtsson and Goodfellow (1987) and by Lahn and Page (1999) , X-Y differentiation may result from suppression of X-Y recombination mediated through chromosomal inversions. Also, this process may be a common, irreversible evolutionary step in mammals (Jegalian and Page 1998 ). However, the deposit of stratum 4 in mammals cannot be easily explained as a direct consequence of a single chromosomal inversion for two reasons. First, if the junction between strata 3 and 4 is a breakpoint of a chromosomal inversion, the amelogenin gene on chromosome X or Y should have been disrupted, suggesting that any such inversion, if present, must have occurred somewhere else. Second, the initiation of stratum 4 differentiation is generally mammalian-order-specific, but the junctions between strata 3 and 4 are almost identical, at least between humans and cattle. This finding can be explained by a single inversion that might occur in the common ancestor of humans and cattle. However, since pigs and horses are more closely related to cattle than to humans (Goodman, Czelusniak, and Beeber 1985 ) and the inversion is likely to be shared by all four mammalian species, it is difficult to account for the observed large differences in the ks or ps value among mammalian AMGX and AMGY paralogs. Nonetheless, it is still conceivable that independent chromosomal inversions, particularly on chromosome Y, were responsible for differential initiations of recombination suppression in mammals. Although such inversions must have occurred in regions other than the amelogenin locus itself, they could happen to suppress recombination around the locus. According to Navarro et al. (1997) , the proximal region of chromosomal inversions is expected to undergo a strong reduction in recombination. Recombination appears to be a complex event involving a number of enzymes and proteins, and the event is initiated by the specific binding of one of these proteins to the DNA molecule of the chromosome (Klein and Takahata 1990 ). If this interaction is highly specific, recombination would be initiated only in limited regions of a chromosome. While the molecular mechanism of recombination is still poorly understood, it is tempting to conclude that the upstream region of the amelogenin locus may contain a sequence or structural signal responsible for initiating or suppressing recombination, thereby repeatedly triggering subsequent X-Y differentiation in different mammals. Stephen Palumbi, Reviewing Editor Keywords: recombination chromosomal inversion nucleotide differences BAC clone evolutionary strata Address for correspondence and reprints: Naoyuki Takahata, Department of Biosystems Science, Graduate University for Advanced Studies, Hayama, Kanagawa 240-0193, Japan. takahata@soken.ac.jp . Fig. 1.—Percentages of sequence difference between the human AMGX and AMGY genes and their 5′ and 3′ flanking regions in the comparison between the two BAC clones (AC002366 and AC013412). Plotted are the nucleotide differences in nonoverlapping windows (100-bp window size). The stepwise regression (thick line) is determined by the least-squares method. The decline in the sequence differences is located in the second intron, and the reduced level extends beyond the downstream region of the gene. The exon (numbered from 1 through 7) and intron structure (in a horizontal line) is depicted at the bottom. Various insertion elements in chromosome X (above) and Y (below) are indicated by open triangles, and they are excluded from the analysis We thank two anonymous reviewers for their constructive criticism on an early version of this paper. This work was supported in part by Monkasho grant 12304046 to N.T.