No, it's not specific to the X chromosome. It's a general trend. CG sequences are underrepresented in mammalian genomes. In fact, the very nucleotides C and G are underrepresented. Instead of the ~25% you'd expect by chance, the real numbers for the human genome are (calculated on the hg19 version of the UCSC human genome):
N: 234350281 = 7.5702% <-- N means unknown
A: 844868045 = 27.2917%
C: 585017944 = 18.8978%
T: 846097277 = 27.3314%
G: 585360436 = 18.9089%
If we break this down by chromosome (chrM is the mitochondrial genome which should probably be ignored here), we get:
chrM A:5113 (30.9%) C:5192 (31.3%) T:4086 (24.7%) G:2180 (13.2%) N:0 (0%)
chr1 A:65570891 (29.1%) C:47024412 (20.9%) T:65668756 (29.1%) G:47016562 (20.9%) N:0 (0%)
chr2 A:71102632 (29.2%) C:47915465 (19.7%) T:71239379 (29.3%) G:47947042 (19.7%) N:4994855 (2.1%)
chr3 A:58713343 (29.6%) C:38653197 (19.5%) T:58760485 (29.7%) G:38670110 (19.5%) N:3225295 (1.6%)
chr4 A:57932980 (30.9%) C:35885806 (19.1%) T:57952068 (30.9%) G:35890822 (19.1%) N:0 (0%)
chr5 A:53672554 (30.2%) C:35089383 (19.7%) T:53804137 (30.3%) G:35129186 (19.8%) N:0 (0%)
chr6 A:50554433 (29.5%) C:33143287 (19.4%) T:50533923 (29.5%) G:33163423 (19.4%) N:3720001 (2.2%)
chr7 A:45997757 (29.6%) C:31671670 (20.4%) T:46047257 (29.6%) G:31636979 (20.4%) N:0 (0%)
chr8 A:42767293 (29.9%) C:28703983 (20.1%) T:42715025 (29.9%) G:28702621 (20.1%) N:0 (0%)
chr9 A:35260078 (29.3%) C:24826212 (20.7%) T:35243882 (29.3%) G:24813259 (20.7%) N:0 (0%)
chr10 A:38330752 (28.3%) C:27308648 (20.1%) T:38376915 (28.3%) G:27298423 (20.1%) N:4220009 (3.1%)
chr11 A:38307244 (29.2%) C:27236798 (20.8%) T:38317436 (29.2%) G:27268038 (20.8%) N:0 (0%)
chr12 A:38604831 (28.8%) C:26634995 (19.9%) T:38624517 (28.9%) G:26617050 (19.9%) N:3370502 (2.5%)
chr13 A:29336945 (30.7%) C:18412698 (19.3%) T:29425459 (30.8%) G:18414776 (19.3%) N:0 (0%)
chr14 A:25992966 (29.4%) C:18027132 (20.4%) T:26197495 (29.7%) G:18071947 (20.5%) N:0 (0%)
chr15 A:23620876 (23%) C:17247582 (16.8%) T:23597921 (23%) G:17228387 (16.8%) N:20836626 (20.3%)
chr16 A:21724083 (27.5%) C:17630040 (22.3%) T:21828642 (27.7%) G:17701988 (22.4%) N:0 (0%)
chr17 A:21159933 (27.2%) C:17727956 (22.8%) T:21206981 (27.3%) G:17700340 (22.8%) N:0 (0%)
chr18 A:22465380 (28.8%) C:14838685 (19%) T:22489493 (28.8%) G:14863671 (19%) N:3420019 (4.4%)
chr19 A:14390632 (25.8%) C:13478255 (24.2%) T:14428951 (25.9%) G:13511145 (24.2%) N:0 (0%)
chr20 A:16523053 (27.8%) C:13107828 (22%) T:16725227 (28.1%) G:13149412 (22.1%) N:0 (0%)
chr21 A:10422924 (21.7%) C:7160212 (14.9%) T:10348785 (21.5%) G:7174721 (14.9%) N:13023253 (27.1%)
chr22 A:9094775 (17.7%) C:8375984 (16.3%) T:9054551 (17.6%) G:8369235 (16.3%) N:16410021 (32%)
chrX A:45648952 (30.2%) C:29813353 (19.7%) T:45772424 (30.3%) G:29865831 (19.8%) N:0 (0%)
chrY A:7667625 (29.9%) C:5099171 (19.9%) T:7733482 (30.1%) G:5153288 (20.1%) N:0 (0%)
As you can see, C and G are consistently less frequent than A and T across all chromosomes. The exception is the mitochondrial genome but that is not the genome of a vertebrate. So, overall, A and T occur almost 10% more often than C and G. But why does this happen?
As far as I know, the reasons are not entirely clear. Some factors that affect this, however, are:
The C of the CG dinucleotides can be methytlated. Methylation is a chemical modification of the nucleotide and plays an important role in the regulation of gene expression. The regulatory sequences that control the way that genes are expressed (promoters), are often rich in CG sequences (these localized overrepresentations of CG are called CpG islands) which enable the control of the associated gene's expression by the mechanism of methylation (which can "turn off" a gene). Not the kind of thing you want to have occurring randomly in your genes.
Methylated Cytosine (C) is often converted to Thymine (T):
This would result in a gradual loss of C in favor of T. This theory had long been considered to be the main reason why CpG is so rare in mammalian genomes, but see the next point.
I found a study from 2004 [1] that shows that, among other things, there is no correlation between CpG and TpA shortages, suggesting that the deamination of Cytosine to Thymine is not enough to explain the phenomenon.
I should also point out that the fact that Alanine represents 7.8% of all aminoacids in proteins does not say anything at all about the frequency with which its codons appear in the genome at large. Only ~2% of the human genome (genes are ~5% and exons, those parts of genes that are translated into proteins, only ~2%) actually codes for proteins so it is only in that 2% where you will see this overrepresentation (if one exists) of Alanine.
In conclusion, yes, the CpG shortage is a general feature of the human genome and not of chromosome X, but the precise reasons for this are not yet entirely clear (as far as I know, anyway).
References
- Jabbari K., and Bernardi G., Gene 333:143-149, 2004
