The DNA sequence of the human X chromosome

Abstract

The human X chromosome has a unique biology that was shaped by its evolution as the sex chromosome shared by males and females. We have determined 99.3% of the euchromatic sequence of the X chromosome. Our analysis illustrates the autosomal origin of the mammalian sex chromosomes, the stepwise process that led to the progressive loss of recombination between X and Y, and the extent of subsequent degradation of the Y chromosome. LINE1 repeat elements cover one-third of the X chromosome, with a distribution that is consistent with their proposed role as way stations in the process of X-chromosome inactivation. We found 1,098 genes in the sequence, of which 99 encode proteins expressed in testis and in various tumour types. A disproportionately high number of mendelian diseases are documented for the X chromosome. Of this number, 168 have been explained by mutations in 113 X-linked genes, which in many cases were characterized with the aid of the DNA sequence.

Figure 2

Xp and Xq pericentromeric contigs extend into the X-chromosome-specific higher-order alpha satellite, DXZ1. The pericentromeric region of the X chromosome is shown as a truncated ideogram. Self-self alignments of proximal sequences from each arm are illustrated by dotter plots below the ideogram. On each plot, the junction between the arm sequence and the arm-specific satellite region is marked by a red arrow, and the junction between the arm-specific satellite region and the X-chromosome-specific alpha satellite array (DXZ1) is marked with a blue arrow. Approximately 594 kb of sequence were analysed from Xp, including ~21 kb of DXZ1 sequence. The ~454 kb of sequence analysed from Xq included ~44 kb of DXZ1 sequence. In each case, ~100 kb of arm sequence were included. The highly repetitive structure of pericentromeric satellites is in stark contrast to the near absence of repetitive structure in the arm sequences, despite an unusually high density of LINE repeats in these regions. Gaps in the dark satellite regions occur where interspersed elements (LINEs, SINEs and LTRs) interrupt the satellite sequences. In the Array Sequences dotter plot, the most proximal ~21 kb of the Xp sequence is joined to the most proximal ~44 kb of the Xq sequence. The periodic nature of the centromeric, higher-order alpha satellite array is evident. Black horizontal lines on the plot reveal near identity of sequences spaced at ~2 kb intervals. This DXZ1 sample represents ~65 kb of the 3 (±0.4) Mb alpha satellite array. The regions outlined in blue are self-self alignments (‘Xp DXZ1’ and ‘Xq DXZ1’), and the remaining rectangular region of the plot is an alignment of Xp versus Xq DXZ1, which reveals the close relationship between DXZ1 sequences from each arm.

Figure 3

Homologies between the human X chromosome and chicken autosomes. a, Plot of BLASTZ sequence alignments between the X chromosome and chicken chromosomes 1 (red) and 4 (blue). Grey bar centred at approximately 60 Mb shows the position of the X centromere. Only the relevant section of each chicken chromosome is shown (see Mb scale at left for chromosome 1 and at right for chromosome 4). A schematic interpretation of the homologies shows the XAR and XCR as red and blue bars, respectively (see Fig. 1). Homologies at the ends of the XAR are indicated with arrows and are expanded in b. b, (Top) Genes at the ends of the human XAR. Genes from distal Xp (magenta arrow in a) are in magenta and genes from Xp11.3 (black arrow in a) are shown in black. (Bottom) Arrangement of the orthologous genes on chicken chromosome 1. A hypothetical ring chromosome, with the equivalent gene order to that observed in the chicken, is indicated by the curved, dotted red line. Recombination between one end of the established X chromosome and the ring chromosome at the arrowed position could, in a single step, have added the XAR and created the gene order observed on the human X chromosome.

Figure 4

Conservation of the X chromosome in eutherian mammals. Plot of BLASTZ sequence alignments between the human X chromosome and the mouse (red) and rat (blue) X chromosomes. The rodent chromosomes are oriented with their centromeres pointing downwards. Regions indicated with arrows are long, highly similar repeats in the mouse sequence that are absent from the human and rat sequences. These repeats were apparently collapsed in an earlier analysed version of the mouse sequence, which also had a large inversion with respect to the mouse assembly used here (NCBI32). The NCBI32 assembly has a gap from 0-3 Mb, which explains the absence of homology to the human X sequence in this part of the plot. The open horizontal bar shows the terminal section of human Xp, which is not conserved on the rodent X chromosomes.

Figure 5

Limited homology between the human sex chromosomes illustrates the extent of Y chromosome erosion in non-recombining regions. a, BLASTN alignments (length ≥80 bp, sequence identity ≥70%) between the finished sequences of the X and Y chromosomes. The centromere positions are represented by grey bars. The analysed Y chromosome sequence ends at the large, heterochromatic segment on Yq, which is indicated by the black bar on the truncated Y chromosome ideogram. b, Major blocks of homology remaining between the XAR and the YAR. Expansion of the BLASTN plot from 0-12 Mb on the X chromosome and 0-20 Mb on the Y chromosome. On the X chromosome, the major homologies lie in the terminal 8.5 Mb of Xp: PAR1 (magenta line) and numbered blocks 1-10. Lesser homologies 11 and 12 contain the TBL1X/TBL1Y and AMELX/AMELY genes, respectively. c, The XTR region in detail (88-93 Mb on X and 2.8-6.8 Mb on Y). Black arrows show large segments deleted from the Y chromosome copy of the XTR. The magenta arrow indicates the short segment that is separated from the rest of the XTR by a paracentric inversion on the Y chromosome. An independent inversion polymorphism on Yp in human populations encompasses this small segment. The position and orientation of the segment shows that the Y chromosome reference sequence is of the less common, derived Y chromosome.

Figure 6

Schematic representation of major homologies between the human sex chromosomes. The entire X and Y chromosomes are shown using the same scale on the left and right sides of the figure, respectively. The major heterochromatic region on Yq is indicated by the pale grey box proximal to PAR2. Expanded sections of X and Y are shown in the centre of the figure. Homologies coloured in the figure are either part of the XAR (PAR1 and blocks 1-12), or were duplicated from the X chromosome to the Y chromosome since the divergence of human and chimpanzee lineages (XTR and PAR2). The numbering of XAR-YAR blocks follows that in Fig. 5b. Blocks inverted on the Y chromosome relative to the X chromosome are assigned red, negative numbers.

Figure 7

Evidence for a fifth evolutionary stratum on the X chromosome. a, Sequence identity between the X and Y homology blocks 1-12 (see Figs 5b and 6) plotted in 5-kb windows. The scale shows the total amount of sequence aligned, excluding insertions and deletions (see Methods). A 10-kb spacer is placed between each consecutive block of homology. Segments of the plot are coloured according to the system used in Figs 6 and 7b. On the basis of this plot, a new evolutionary stratum S5 is defined, which includes homology blocks 1 and 2. b, A most parsimonious series of inversion events from the arrangement of homology blocks 1-12 on the X chromosome (top) to the Y chromosome (bottom), calculated using GRIMM. The grey boxes show the suggested extents of former pseudoautosomal regions within the distal part of the XAR, and the magenta box (bottom row) shows the position of the current pseudoautosomal region. This inversion sequence provides independent support for the proposed pseudoautosomal boundary movements and evolutionary strata. It was previously suggested that AMELX (in block 12) is in S4 (ref. 46), or possibly at the boundary between S3 and S4 (ref. 67). However, the more distal location of block 11, which contains TBL1X (an S3 gene46), is not consistent with these suggestions. The two regions of increased sequence identity within block 10 contain the VCX2 and VCX3B genes on the X chromosome and the VCY1B and VCY genes on the Y chromosome. This gene family might have arisen de novo in the simian lineage, which could account for the unusual characteristics of this part of the alignment.

Figure 8

Sequence compositional changes in the distal evolutionary strata of the X chromosome. Shown are the positions of SINE and LINE repeats and (G+C) content within PAR1, S5 and the distal half of S4. The percentage of Alu, L1 and (G+C) are shown for each region (including the whole of S4). There is an abrupt increase in Alu repeat levels and (G+C) content from S4 to S5. The five euchromatic gaps in PAR1 are shown as light brown bars. Pale blue bars represent clones for which the sequences were unfinished at the time of the sequence assembly.