Gene conversion and evolution of Xq28 duplicons involved in recurring inversions causing severe hemophilia A

Abstract

Inversions breaking the 1041 bp int1h-1 or the 9.5-kb int22h-1 sequence of the F8 gene cause hemophilia A in 1/30,000 males. These inversions are due to homologous recombination between the above sequences and their inverted copies on the same DNA molecule, respectively, int1h-2 and int22h-2 or int22h-3. We find that (1) int1h and int22h duplicated more than 25 million years ago; (2) the identity of the copies (>99%) of these sequences in humans and other primates is due to gene conversion; (3) gene conversion is most frequent in the internal regions of int22h; (4) breakpoints of int22h-related inversions also tend to involve the internal regions of int22h; (5) sequence variations in a sample of human X chromosomes defined eight haplotypes of int22h-1 and 27 of int22h-2 plus int22h-3; (6) the latter two sequences, which lie, respectively, 500 and 600 kb telomeric to int22h-1 are five-fold more identical when in cis than when in trans, thus suggesting that gene conversion may be predominantly intrachromosomal; (7) int1h, int22h, and flanking sequences evolved at a rate of about 0.1% substitutions per million years during the divergence between humans and other primates, except for int1h during the human-chimpanzee divergence, when its rate of evolution was significantly lower. This is reminiscent of the slower evolution of palindrome arms in the male specific regions of the Y chromosome and we propose, as an explanation, that intrachromosomal gene conversion and cosegregation of the duplicated regions favors retention of the ancestral sequence and thus reduces the evolution rate.

Figure 1.

Map of int1h and int22h duplicons. (Top) F8 gene (clear box), with arrow indicating direction of transcription, plus sequence of int1h (black boxes) and int22h (gray boxes) with arrows indicating orientation. Clear and solid arrow heads indicate location and orientation of primers flanking int22h copies and int1h copies, respectively (for primer sequences see Liu et al. 1998 and Bagnall et al. 2002). (Bottom) Int22h and int1h showing location and orientation of primers listed in Supplemental Table 3. These were primers required for completing the sequence of the chimpanzee's int22h sequences (solid arrow heads), or for allele-specific PCR in int22h (clear arrow heads), or for amplification of DNA for FSPCCM analysis (lines terminating with arrow heads). Other primers for sequencing int22h were from Naylor et al. (1995) (not shown). Primers for sequencing int1h (solid arrow heads and underlined names) are from Bagnall et al. (2002).

Figure 2.

Observed and possible pre-inversion int22h haplotypes for patients with int22h-related inversions. For each patient a set of proposed pre-inversion haplotypes (pre-inv) are shown. The haplotype of the proposed pre-inversion proximal and distal int22h sequence are in bold and italics, respectively, and the haplotype number is indicated (e.g., H1#2 is haplotype 2 of int22h-1 and H3#23 is haplotype 23 of distal int22h). The observed haplotypes (obs) show the non-recombined int22h in lower case and the recombined haplotypes. The distribution of bold and italic characters in the latter two haplotypes shows how the pre-inversion haplotypes have rearranged. Standard capital letters indicate noninformative sites. Proposed gene conversions are highlighted gray. h1/3 and h1/2 are the recombined sequences at proximal location while h2/1 and h3/1 are those at distal location. Column “D” indicates number of differences between the patient's non-recombined int22h-2 or int22h-3 sequence and the distal int22h haplotype chosen to represent the pre-inversion distal int22h. The distance between the first and last allele that requires a gene conversion event to account for the observed recombined sequences is the gene conversion tract (last column).