How Strong Is the Mutagenicity of Recombination in Mammals?
http://www.100md.com
《分子生物学进展》
* Department of Ecology and Evolution, University of Chicago; Institute of Zoology, National Taiwan University, Taipei, Taiwan; and Department of Biological Sciences, University of South Carolina, Columbia
Correspondence: E-mail: whli@uchicago.edu.
Abstract
It is commonly believed that a high recombination rate such as that in a pseudoautosomal region (PAR) greatly increases the mutation rate because a 170-fold increase was estimated for the mouse PAR region. However, sequencing PAR and non-PAR introns of the Fxy gene in four Mus taxa, we found an increase of only twofold to fivefold. Furthermore, analyses of sequence data from human and orangutan PAR and X-linked regions and from autosomal regions showed a weak effect of recombination on mutation rate (a slope of less than 0.2% per cM/Mb), although a much stronger effect on GC content (1% to 2% per cM/Mb). Because typical recombination rates in mammals are much lower than those in PARs, the mutagenicity of recombination is weak or, at best, moderate, although its effect on GC% is much stronger. In addition, contrary to a previous study, we found no Fxy duplicate in Mus spretus.
Key Words: pseudoautosomal region ? recombination rate ? mutation rate ? nucleotide substitution
Introduction
Recombination has been shown to be mutagenic in yeast (Strathern, Shafer, and McGill 1995; Holbeck and Strathern 1997, 1999; Heidenreich et al. 2003) and mammals (Brown and Jiricny 1987, 1989). In mammals, however, the mutagenicity of recombination has experimentally been little studied but has mostly been inferred from comparisons of substitution rates in regions with different recombination rates (Eyre-Walker 1993; Nachman et al. 1998; Perry and Ashworth 1999; Schiebel et al. 2000; Fullerton, Bernardo Carvalho, and Clark 2001; Nachman 2001; Lercher and Hurst 2002; Filatov and Gerrard 2003; Hardison et al. 2003; Hellmann et al. 2003). In particular, by contrasting the synonymous substitution rates in the pseudoautosomal region (PAR) and the non-PAR introns of the Fxy gene in the Mus genus, which has been known to straddle the pseudoautosomal boundary (PAB) in the laboratory mouse C57BL/6 (Mus musculus domesticus) (Palmer et al. 1997), Perry and Ashworth (1999) estimated that the mutation rate in PAR was elevated 170 times by recombination. Is the mutagenicity of recombination really so strong? Because the estimate was based on limited data, it needs to be re-evaluated with a larger set of data. In fact, no strong mutagenicity of recombination was observed in two recent studies of human and ape PARs (Filatov and Gerrard 2003; Yi et al. 2004), although the weak effect in human could have resulted from a reduction in recombination rate because of a shift of the PAR along the X chromosome (Galtier 2004). At the present time, the best way to obtain an upper bound for the mutagenicity of recombination in mammals is perhaps to contrast the substitution rates in PAR and non-PAR introns of the Fxy gene in the genus Mus, because among the known PARs in mammals, the PAR in Mus is the shortest known in more than one species. The Mus PAR is, in fact, a region of only approximately 700 kb of the X chromosome that is obligated to pair with the Y chromosome (Palmer et al. 1997; Perry et al. 2001). Thus, it should have an unusually high rate of recombination (Broman et al. 2002; Froenicke et al. 2002; Blake et al. 2003), and the large difference in recombination rate between the PAR and non-PAR regions of Mus Fxy makes it an ideal case for addressing the above issue.
It has also been observed that recombination, or more precisely gene conversion, is GC biased, promoting mutations from A and T to G and C (Brown and Jiricny 1988, 1989; Eyre-Walker 1993; Perry and Ashworth 1999; Fullerton, Bernardo Carvalho, and Clark 2001; Montoya-Burgos, Boursot, and Galtier 2003; Marais 2003). Again, the strength of bias has not been well established and needs further investigation. For the same reason as above, the Fxy gene in Mus is suitable for addressing this issue. For the above two purposes, we have sequenced parts of introns 1, 2, and 3 (non-PAR located) and parts of introns 4 through 7 (PAR-located) in M. m. musculus, M. m. domesticus, M. spretus, and M. caroli. In addition, we have also used published data from PAR and non-PAR regions in mammals to address the above two issues.
To pursue the present study, it is useful to know the origin of the partially PAR-located Fxy gene in the Mus genus. Montoya-Burgos, Boursot, and Galtier (2003) proposed that a duplication of Fxy occurred and one copy moved to be partially inside the PAR region in the common ancestor of M. spretus and M. musculus. In this study, we found no evidence for either the proposed duplication or the presence of a partially PAR-located Fxy gene in M. spretus. Thus, the sliding of Fxy into the PAR probably occurred in the ancestor of M. musculus.
Methods
Mus Sequences
The introns and exons of Fxy in Mus musculus strain B6 were obtained by searching the Whole Genome Shotgun database at NCBI (table 3 in Supplementary Material online) (Mouse Genome Sequencing Consortium 2002). PCR (polymerase chain reaction) primers were designed to amplify regions of Fxy in four closely related taxa: two subspecies of Mus musculus (M. m. domesticus from Barcelona, Spain and M. m. musculus from Georgia, Europe) and two other species (Mus spretus and Mus caroli). Five regions (fragments of exons 2, 3, 5, and 6 with parts of their respective introns) could be amplified in all four taxa (figure 2 in Supplementary Material online). Additional portions of introns 1 to 7 were obtained solely for M. m. musculus and M. m. domesticus.
FIG. 2.— Phylogeny of the Fxy genes in Mus. The scale bar is in units of numbers of substitutions per site. (A) The first half of Fxy (intron 1 to beginning of intron 3). (B) The second half of Fxy (intron 4 to intron 6).
To verify the second copy of an Fxy in the PAR of M. spretus proposed by Montoya-Burgos, Boursot, and Galtier (2003), PCR primers were designed for exons 9 and 10 in M. spretus, M. m. musculus, and M. m. domesticus. The primer sequences are
Exon 9: 5'-GGCTCATCGCAAGCTGAAGGTGT-3'
5'-CACTCCGTAGCTCCCCTGAC-3'
Exon 10: 5'-GACTACGACAACGGCTCCATC-3'
5'-CAGACACTTGTTCCACACGGTGA-3'
Evolutionary Analysis
Sequences were aligned using ClustalW (Thompson et al. 1997). The neighbor-joining method (Saitou and Nei 1987) and the Kimura two-parameter model of sequence evolution, as implemented in MEGA version 2 (Kumar et al. 2001), were used to infer the phylogeny of Mus Fxy.
To estimate the frequency of mutations in different lineages (Gojobori, Li, and Graur 1982; Li, Wu, and Luo 1984), the sequence of the ancestor was inferred at the phylogenetic node joining M. musculus, M. spretus, and M. caroli (outgroup). This method reconstructs an ancestral nucleotide sequence by assigning bases (A, G, C, or T) to the ancestral node based on the least number of changes along all lineages of a phylogeny that can account for the present-day nucleotide sequence differences. We use this simple method because the sequences under study are highly similar.
The assembled human genomic sequence (Golden Path alignment version hg16; repeat-masked) and recombination rate data (Kong et al. 2002) for all autosomes were obtained from the UCSC Genome Browser. The human autosomal sequence was divided into nonoverlapping bins of size 1 Mb (including repeats) and recombination rates (cM/Mb) assigned. Repetitive sequences were then removed for computing GC%.
Results and Discussion
No Evidence of Fxy Duplication in M. spretus
There are two lines of evidence against Montoya-Burgos, Boursot, and Galtier (2003) claim of a partially PAR-located Fxy gene in M. spretus. First, in their phylogeny of rodent Fxy genes, the M. spretus and M. musculus Fxy_PAR genes were clustered with a nearly zero divergence (see figure 1 for a simplified phylogeny). This condition is highly unlikely, because the PAR region has a higher rate of evolution than the non-PAR region, so that M. musculus Fxy_PAR and M. spretus Fxy_PAR should have a larger divergence than M. musculus Fxy_non-PAR and M. spretus Fxy_non-PAR. Their phylogeny would imply approximately 80 parallel nucleotide substitutions but only four nonparallel substitutions between M. musculus Fxy_PAR and M. spretus Fxy_PAR in the three exons (8, 9, and 10) studied (fig. 1).
FIG. 1.— Phylogeny of the Fxy genes based on exons 8, 9, and 10 (619 sites). The sequence data for the proposed M. spretus_PAR exons are from Montoya-Burgos, Boursot, and Galtier (2003), and the other sequence data are from GenBank. The scale bar is in units of numbers of substitutions. The arrows point to events proposed by Montoya-Burgos, Boursot, and Galtier (2003) and are numbered by order of occurrence.
Second, no PAR-located region of Fxy was detected in M. spretus by PCR amplification in our study (figure 1 in Supplementary Material online). Our PCR primers were specifically designed to detect exons 9 and 10 (both PAR-located) in the alleged Fxy_PAR in M. spretus. However, neither of the two exons was amplified in any of the three M. spretus individuals included in our study, although, as positive controls, both were PCR-amplified in M. musculus. Our PCR results suggest no Fxy_PAR in M. spretus. It is possible that the proposed Fxy_PAR in M. spretus was a result of PCR contamination. In support of this hypothesis, we note that Montoya-Burgos, Boursot, and Galtier's (2003) claim was based on the sequences of exons 8, 9, and 10, but only 4 nt differences were found between their proposed M. spretus Fxy_PAR exons 8, 9, and 10 and M. musculus Fxy exons 8, 9, and 10 (fig. 1).
With no evidence of Fxy_PAR in M. spretus, the movement of Fxy into the PAR occurred after the divergence of M. spretus and M. musculus—likely before the diversification of the M. musculus lineage (Perry and Ashworth 1999). Also, in the absence of a duplication of Fxy in the Mus lineages, there is no need to invoke a gene loss and a translocation of Fxy to explain Fxy_PAR in M. musculus, as proposed by Montoya-Burgos, Boursot, and Galtier (2003). It is more likely that the PAR has slid along the X chromosome in the lineage leading to M. musculus.
Mutagenicity of Recombination
In Mouse
The introns sequenced were divided into two groups, roughly the first half (intron 1 to first-half of intron 3; 2,307 sites sequenced) and the second half (intron 4 to intron 6; 2,378 sites sequenced). Only the second half in M. m. domesticus (dom) and M. m. musculus (mus) is in the PAR.
To compare the rate of nucleotide substitutions in different regions and taxa, we constructed two neighbor-joining trees for the two halves of the Fxy gene (fig. 2). The total branch length between M. caroli (car) and M. spretus (spr) is 0.0163 + 0.0077 + 0.0043 = 0.0283 for the first half (fig. 2A) and 0.0580 for the second half (fig. 2B). Therefore, the second half has evolved two times faster than the first half. On the other hand, the total branch length between dom and mus is 0.0030 for the first half and 0.0150 (five times higher) for the second half. Therefore, we may infer that because of a very high recombination rate in the PAR, the substitution rate in the Mus PAR introns has increased 5/2 = 2.5 times.
We now consider a second estimate. From figure 2, we computed the average branch length from the common ancestral node of spr, dom, and mus to the tips of dom and mus as 0.0071 + (0.0015 + 0.0015)/2 = 0.0086 for the first half and 0.0821 for the second half. The estimated substitution rate for the second half is now 9.5 times faster than that for the first half. Therefore, we may infer that the high recombination rate in the second half has increased the substitution rate by 9.5/2 = 4.8-fold. As in the above estimate, this estimate assumes that the non-PAR second half has evolved two times faster than the first half. This estimate should be taken with caution for two reasons. First, the ratio 0.0821/0.0086 = 9.5 could have been inflated because the denominator is small, so that a small error can be easily amplified. Second, it assumes that the second half has been in the PAR since the divergence between the spr lineage and the dom-mus lineage, which may not be true.
A third estimate was obtained by comparing the branch length (0.0329) from the common node of spr, dom, and mus with spr and comparing the average branch length (0.0821) from the same common node with dom and mus. This comparison gives a ratio of 2.5 for the increase in substitution rate because of the increase in recombination rate. Again, this estimate assumes that the second half has been in the PAR since the divergence between the spr and the dom-mus lineages. Despite this assumption, it might not be an underestimate because in the first half (fig. 2A) the average branch length (0.0071 + 0.0015 = 0.0086) for the dom-mus lineage is twice as long as that for the spr lineage (0.0043), suggesting the possibility of a faster clock in the dom-mus lineage than in the spr lineage.
From the above three estimates, we suggest that the increase in recombination rate in the PAR has, on average, increased the substitution rate in the PAR introns by twofold to fivefold. A very liberal upper bound is 10, which is obtained from the above second estimate (9.5) without dividing it by 2.
In Human
A weak effect of recombination on mutation rate was found in two recent studies of XG (Filatov and Gerrard 2003; Yi et al. 2004), a gene that straddles the PAB in human and apes (Rappold 1993; Lien et al. 2000). Filatov and Gerrard (2003) also studied two additional genes located in the human and ape PAR that did show twofold and threefold increases in the number of substitutions compared with the genomic average (3%) between human and orangutan. However, their analysis of X-linked and autosomal regions in human and orangutan showed no evidence of an association between mutation rate and recombination rate.
In this study, three data sets (Filatov and Gerrard 2003; Filatov 2004; Yi et al. 2004) were pooled along with the available sex-averaged recombination rates in human (Kong et al. 2002), and the human/orangutan intron divergences were correlated with recombination rates (fig. 3). For the PAR and X-linked regions, the influence of recombination rate on the substitution rate is fairly weak; that is, slope = 0.22% substitutions per cM/Mb (fig. 3A), whereas for the autosomal regions, there is no correlation and a slope of approximately 0.0 (fig. 3A).
FIG. 3.— Divergence (% substitutions per site) versus recombination rate (cM/Mb). (A) PAR and X-linked regions only. Each region consisted of the introns within a gene, except for the XG gene, which was divided into PAR and X-linked halves. (B) Autosomal regions only. Data sets obtained from (Filatov and Gerrard 2003; Filatov 2004; Yi et al. 2004) and from a manual search of the human genome.
Lercher and Hurst (2002) observed a weak correlation between human/rodent divergence and recombination rate using a lesser-resolved genetic map than is currently available. Using the more-resolved genetic map assembled by Kong et al. (2002), Hellmann et al. (2003) observed a stronger correlation (R2 = 0.29) between human/chimp divergence and recombination. Both studies had one region with a recombination rate well over 4 cM/Mb, which did not occur in our human/orangutan data set (fig. 3B). However, the analysis of the PAR (fig. 3A), with its higher rates of recombination, did show a correlation (R2 = 0.43) between divergence and recombination rate, which is consistent with the above studies.
Increase of GC Content with Recombination Rate
We inferred lineage-specific mutations in each segment sequenced and found that PAR introns have a higher rate of change from A or T (0.1081 or 0.1110) than from G or C (0.0370 or 0.0302; [table 1, and see also table 1 in Supplementary Material online]). This trend is not apparent in any of the non-PAR regions. Also, using additional sites for the comparison between M. m. musculus and M. m. domesticus revealed an average GC% of 44.3 in the PAR versus 37.2 in the non-PAR. Both results support a GC bias in the PAR (Filatov and Gerrard 2003; Filatov 2004; Yi et al. 2004).
Table 1 Relative Rates of Substitution in the First and Second Half of Fxy in Different Mus Lineages
Next, we pooled the three data sets (Filatov and Gerrard 2003; Filatov 2004; Yi et al. 2004) to study the correlation between GC% and recombination rate (fig. 4A). The estimated effect of recombination on GC (slope = 0.70 GC% per cM/Mb) was moderate.
FIG. 4.— GC (%) versus recombination rate (cM/Mb). (A) Data set same as that used in figure 2A. (B) Data set from 2669 nonoverlapping windows (of size 1 Mb) along all autosomes in the human genome. Repetitive sequences were excluded before computing the GC content.
Finally, we computed the correlation between recombination rate and the GC% in nonoverlapping windows across the human autosomes, with the exclusion of repetitive elements, and found a slope of 2.32 GC% per cM/Mb (fig. 4B). This slope is approximately three times as steep as that revealed by the human PAR and X-linked regions.
One possibility for the difference in slope between the above two data sets is that the effect of recombination on GC% is not linear. However, we did a logistic transformation of the GC values in the human PAR and X-linked regions and obtained a similar regression line. Another possibility is that the recombination rate changes with time, so that the current rate may be different from the rate in the past. A third possibility is that the effect of recombination varies among chromosomes. In this case, the two slopes (0.7 and 2.3) can represent a lower and an upper bound of the effect of recombination on GC%. A value closer to the upper bound seems more reasonable because the window size was large and not as sensitive to past changes in the pattern of recombination along the chromosomes. Galtier (2004) proposed that past events that shifted part of the PAR into a non-PAR region can lead to a great reduction in recombination rate and a slow rate of change in GC%. Under this situation, the effects of recombination near the PAB could be underestimated. If this is true, the upper bound is closer to the truth because our lower bound was estimated from regions surrounding the PAB.
Conclusions
Our data from the Fxy gene suggest only a twofold to fivefold (or at most 10-fold) increase in mutation rate caused by a dramatic increase in recombination rate in the PAR-linked regions of the Fxy gene. This finding is in stark contrast with the previous estimate of an increase of 170 times in Fxy_PAR (Perry and Ashworth 1999). Because the latter estimate was obtained using synonymous sites of exons in Fxy, it would be less reliable than our estimate, which is based on many more sites. Furthermore, our analysis of the human and orangutan data from the PAR and X-linked regions gave a regression line of y = 0.22%x + 2.61%, where x is in cM/Mb (fig. 3A). Thus, even for a recombination rate as high as 20 cM/Mb, y is 4.40% + 2.61% = 7.01%, which is approximately 2.3 times the average genomic divergence (3%) between human and orangutan. Because the recombination rate in the human genome is usually much lower than 20 cM/Mb (i.e., usually below 3.2 cM/Mb), the effect of recombination on mutation rate should generally be far weaker than that in PAR regions. On the other hand, the effect of recombination (gene conversion) on GC content appears to be much stronger than the effect on mutation rate. Our study suggests a slope between 0.7% and 2.3% per cM/Mb. Thus, an increase of 5 cM/Mb in recombination rate can lead to a 5% to 10% increase in GC content. In this respect, our results are consistent with previous studies (Marais 2003; Montoya-Burgos, Boursot, and Galtier 2003).
Acknowledgements
We thank Dr. Francois Bonhomme and Dr. Michael Nachman for sending us DNA samples and Dr. D. Filatov for making available human and orangutan sequence alignments. This study was supported by NIH grants.
References
Blake, J. A., J. E. Richardson, C. J. Bult, J. A. Kadin, J. T. Eppig,the members of the Mouse Genome Database Group. 2003. MGD: The Mouse Genome Database. Nucleic Acids Res. 31:193–195.
Broman, K. W., L. B. Rowe, G. A. Churchill, and K. Paigen. 2002. Crossover interference in the mouse. Genetics 160:1123–11231.
Brown, T. C., and J. Jiricny. 1987. A specific mismatch repair event protects mammalian cells from loss of 5-methylcytosine. Cell 50:945–950.
———. 1988. Different base/base mispairs are corrected with different efficiencies and specificities in monkey kidney cells. Cell 54:705–711.
———. 1989. Repair of base-base mismatches in simian and human cells. Genome 31:578–583.
Eyre-Walker, A. 1993. Recombination and mammalian genome evolution. Proc. R. Soc. Lond. B Biol. Sci. 252:237–243.
Filatov, D. A. 2004. A gradient of silent substitution rate in the human pseudoautosomal region. Mol. Biol. Evol. 21:410–417.
Filatov, D. A., and D. T. Gerrard. 2003. High mutation rates in human and ape pseudoautosomal genes. Gene 317:67–77.
Froenicke, L., L. K. Anderson, J. Wienberg, and T. Ashley. 2002. Male mouse recombination maps for each autosome identified by chromosome painting. Am. J. Hum. Genet. 71:1353–1368.
Fullerton, S. M., A. Bernardo Carvalho, and A. G. Clark. 2001. Local rates of recombination are positively correlated with GC content in the human genome. Mol. Biol. Evol. 18:1139–1142.
Galtier, N. 2004. Recombination, GC-content and the human pseudoautosomal boundary paradox. Trends Genet. 20:347–349.
Gojobori, T., W-H Li, and D. Graur. 1982. Patterns of nucleotide substitution in pseudogenes and functional genes. J. Mol. Evol. 18:360–369.
Hardison, R. C., K. M. Roskin, S. Yang et al. (15 co-authors). 2003. Covariation in frequencies of substitution, deletion, transposition, and recombination during eutherian evolution. Genome Res. 13:13–26.
Heidenreich, E., R. Novotny, B. Kneidinger, V. Holzmann, and U. Wintersberger. 2003. Non-homologous end joining as an important mutagenic process in cell cycle-arrested cells. EMBO J. 22:2274–2283.
Hellmann, I., I. Ebersberger, S. E. Ptak, S. Paabo, and M. Przeworski. 2003. A neutral explanation for the correlation of diversity with recombination rates in humans. Am. J. Hum. Genet. 72:1527–1535.
Holbeck, S. L., and J. N. Strathern. 1997. A role for REV3 in mutagenesis during double-strand break repair in Saccharomyces cerevisiae. Genetics 147:1017–1024.
———. 1999. EXO1 of Saccharomyces cerevisiae functions in mutagenesis during double-strand break repair. Ann. N. Y. Acad. Sci. 870:375–377.
Kong, A., D. F. Gudbjartsson, J. Sainz et al. (13 co-authors). 2002. A high-resolution recombination map of the human genome. Nat. Genet. 31:241–247.
Kumar, S., K. Tamura, I. B. Jakobsen, and M. Nei. 2001. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 17:1244–1245.
Lercher, M. J., and L. D. Hurst. 2002. Human SNP variability and mutation rate are higher in regions of high recombination. Trends. Genet. 18:337–340.
Li, W-H., C-I Wu, and C-C Luo. 1984. Nonrandomness of point mutation as reflected in nucleotide substitutions in pseudogenes and its evolutionary implications. J. Mol. Evol. 21:58–71.
Lien, S., J. Szyda, B. Schechinger, G. Rappold, and N. Arnheim. 2000. Evidence for heterogeneity in recombination in the human pseudoautosomal region: high resolution analysis by sperm typing and radiation-hybrid mapping. Am. J. Hum. Genet. 66:557–566.
Marais, G. 2003. Biased gene conversion: implications for genome and sex evolution. Trends Genet. 19:330–338.
Montoya-Burgos, J. I., P. Boursot, and N. Galtier. 2003. Recombination explains isochores in mammalian genomes. Trends Genet. 19:128–130.
Mouse Genome Sequencing Consortium. 2002. Initial sequencing and comparative analysis of the mouse genome. Nature 420:520–562.
Nachman, M. W. 2001. Single nucleotide polymorphisms and recombination rate in humans. Trends Genet. 17:481–485.
Nachman, M. W., V. L. Bauer, S. L. Crowell, and C. F. Aquadro. 1998. DNA variability and recombination rates at X-linked loci in humans. Genetics 150:1133–1141.
Palmer, S., J. Perry, D. Kipling, and A. Ashworth. 1997. A gene spans the pseudoautosomal boundary in mice. Proc. Natl. Acad. Sci. USA 94:12030–12035.
Perry, J., and A. Ashworth. 1999. Evolutionary rate of a gene affected by chromosomal position. Curr. Biol. 9:987–989.
Perry, J., S. Palmer, A. Gabriel, and A. Ashworth. 2001. A short pseudoautosomal region in laboratory mice. Genome Res. 11:1826–1832.
Rappold, G. A. 1993. The pseudoautosomal regions of the human sex chromosomes. Hum. Genet. 92:315–324.
Saitou, N., and M. Nei. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406–425.
Schiebel, K., J. Meder, A. Rump, A. Rosenthal, M. Winkelmann, C. Fischer, T. Bonk, A. Humeny, and G. Rappold. 2000. Elevated DNA sequence diversity in the genomic region of the phosphatase PPP2R3L gene in the human pseudoautosomal region. Cytogenet. Cell Genet. 91:224–230.
Strathern, J. N., B. K. Shafer, and C. B. McGill. 1995. DNA synthesis errors associated with double-strand-break repair. Genetics 140:965–972.
Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and D. G. Higgins. 1997. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25:4876–4882.
Yi, S., T. J. Summers, N. M. Pearson, and W-H Li. 2004. Recombination has little effect on the rate of sequence divergence in pseudoautosomal boundary 1 among humans and great apes. Genome Res. 14:37–43.(Shiao-Wei Huang*,, Robert)
Correspondence: E-mail: whli@uchicago.edu.
Abstract
It is commonly believed that a high recombination rate such as that in a pseudoautosomal region (PAR) greatly increases the mutation rate because a 170-fold increase was estimated for the mouse PAR region. However, sequencing PAR and non-PAR introns of the Fxy gene in four Mus taxa, we found an increase of only twofold to fivefold. Furthermore, analyses of sequence data from human and orangutan PAR and X-linked regions and from autosomal regions showed a weak effect of recombination on mutation rate (a slope of less than 0.2% per cM/Mb), although a much stronger effect on GC content (1% to 2% per cM/Mb). Because typical recombination rates in mammals are much lower than those in PARs, the mutagenicity of recombination is weak or, at best, moderate, although its effect on GC% is much stronger. In addition, contrary to a previous study, we found no Fxy duplicate in Mus spretus.
Key Words: pseudoautosomal region ? recombination rate ? mutation rate ? nucleotide substitution
Introduction
Recombination has been shown to be mutagenic in yeast (Strathern, Shafer, and McGill 1995; Holbeck and Strathern 1997, 1999; Heidenreich et al. 2003) and mammals (Brown and Jiricny 1987, 1989). In mammals, however, the mutagenicity of recombination has experimentally been little studied but has mostly been inferred from comparisons of substitution rates in regions with different recombination rates (Eyre-Walker 1993; Nachman et al. 1998; Perry and Ashworth 1999; Schiebel et al. 2000; Fullerton, Bernardo Carvalho, and Clark 2001; Nachman 2001; Lercher and Hurst 2002; Filatov and Gerrard 2003; Hardison et al. 2003; Hellmann et al. 2003). In particular, by contrasting the synonymous substitution rates in the pseudoautosomal region (PAR) and the non-PAR introns of the Fxy gene in the Mus genus, which has been known to straddle the pseudoautosomal boundary (PAB) in the laboratory mouse C57BL/6 (Mus musculus domesticus) (Palmer et al. 1997), Perry and Ashworth (1999) estimated that the mutation rate in PAR was elevated 170 times by recombination. Is the mutagenicity of recombination really so strong? Because the estimate was based on limited data, it needs to be re-evaluated with a larger set of data. In fact, no strong mutagenicity of recombination was observed in two recent studies of human and ape PARs (Filatov and Gerrard 2003; Yi et al. 2004), although the weak effect in human could have resulted from a reduction in recombination rate because of a shift of the PAR along the X chromosome (Galtier 2004). At the present time, the best way to obtain an upper bound for the mutagenicity of recombination in mammals is perhaps to contrast the substitution rates in PAR and non-PAR introns of the Fxy gene in the genus Mus, because among the known PARs in mammals, the PAR in Mus is the shortest known in more than one species. The Mus PAR is, in fact, a region of only approximately 700 kb of the X chromosome that is obligated to pair with the Y chromosome (Palmer et al. 1997; Perry et al. 2001). Thus, it should have an unusually high rate of recombination (Broman et al. 2002; Froenicke et al. 2002; Blake et al. 2003), and the large difference in recombination rate between the PAR and non-PAR regions of Mus Fxy makes it an ideal case for addressing the above issue.
It has also been observed that recombination, or more precisely gene conversion, is GC biased, promoting mutations from A and T to G and C (Brown and Jiricny 1988, 1989; Eyre-Walker 1993; Perry and Ashworth 1999; Fullerton, Bernardo Carvalho, and Clark 2001; Montoya-Burgos, Boursot, and Galtier 2003; Marais 2003). Again, the strength of bias has not been well established and needs further investigation. For the same reason as above, the Fxy gene in Mus is suitable for addressing this issue. For the above two purposes, we have sequenced parts of introns 1, 2, and 3 (non-PAR located) and parts of introns 4 through 7 (PAR-located) in M. m. musculus, M. m. domesticus, M. spretus, and M. caroli. In addition, we have also used published data from PAR and non-PAR regions in mammals to address the above two issues.
To pursue the present study, it is useful to know the origin of the partially PAR-located Fxy gene in the Mus genus. Montoya-Burgos, Boursot, and Galtier (2003) proposed that a duplication of Fxy occurred and one copy moved to be partially inside the PAR region in the common ancestor of M. spretus and M. musculus. In this study, we found no evidence for either the proposed duplication or the presence of a partially PAR-located Fxy gene in M. spretus. Thus, the sliding of Fxy into the PAR probably occurred in the ancestor of M. musculus.
Methods
Mus Sequences
The introns and exons of Fxy in Mus musculus strain B6 were obtained by searching the Whole Genome Shotgun database at NCBI (table 3 in Supplementary Material online) (Mouse Genome Sequencing Consortium 2002). PCR (polymerase chain reaction) primers were designed to amplify regions of Fxy in four closely related taxa: two subspecies of Mus musculus (M. m. domesticus from Barcelona, Spain and M. m. musculus from Georgia, Europe) and two other species (Mus spretus and Mus caroli). Five regions (fragments of exons 2, 3, 5, and 6 with parts of their respective introns) could be amplified in all four taxa (figure 2 in Supplementary Material online). Additional portions of introns 1 to 7 were obtained solely for M. m. musculus and M. m. domesticus.
FIG. 2.— Phylogeny of the Fxy genes in Mus. The scale bar is in units of numbers of substitutions per site. (A) The first half of Fxy (intron 1 to beginning of intron 3). (B) The second half of Fxy (intron 4 to intron 6).
To verify the second copy of an Fxy in the PAR of M. spretus proposed by Montoya-Burgos, Boursot, and Galtier (2003), PCR primers were designed for exons 9 and 10 in M. spretus, M. m. musculus, and M. m. domesticus. The primer sequences are
Exon 9: 5'-GGCTCATCGCAAGCTGAAGGTGT-3'
5'-CACTCCGTAGCTCCCCTGAC-3'
Exon 10: 5'-GACTACGACAACGGCTCCATC-3'
5'-CAGACACTTGTTCCACACGGTGA-3'
Evolutionary Analysis
Sequences were aligned using ClustalW (Thompson et al. 1997). The neighbor-joining method (Saitou and Nei 1987) and the Kimura two-parameter model of sequence evolution, as implemented in MEGA version 2 (Kumar et al. 2001), were used to infer the phylogeny of Mus Fxy.
To estimate the frequency of mutations in different lineages (Gojobori, Li, and Graur 1982; Li, Wu, and Luo 1984), the sequence of the ancestor was inferred at the phylogenetic node joining M. musculus, M. spretus, and M. caroli (outgroup). This method reconstructs an ancestral nucleotide sequence by assigning bases (A, G, C, or T) to the ancestral node based on the least number of changes along all lineages of a phylogeny that can account for the present-day nucleotide sequence differences. We use this simple method because the sequences under study are highly similar.
The assembled human genomic sequence (Golden Path alignment version hg16; repeat-masked) and recombination rate data (Kong et al. 2002) for all autosomes were obtained from the UCSC Genome Browser. The human autosomal sequence was divided into nonoverlapping bins of size 1 Mb (including repeats) and recombination rates (cM/Mb) assigned. Repetitive sequences were then removed for computing GC%.
Results and Discussion
No Evidence of Fxy Duplication in M. spretus
There are two lines of evidence against Montoya-Burgos, Boursot, and Galtier (2003) claim of a partially PAR-located Fxy gene in M. spretus. First, in their phylogeny of rodent Fxy genes, the M. spretus and M. musculus Fxy_PAR genes were clustered with a nearly zero divergence (see figure 1 for a simplified phylogeny). This condition is highly unlikely, because the PAR region has a higher rate of evolution than the non-PAR region, so that M. musculus Fxy_PAR and M. spretus Fxy_PAR should have a larger divergence than M. musculus Fxy_non-PAR and M. spretus Fxy_non-PAR. Their phylogeny would imply approximately 80 parallel nucleotide substitutions but only four nonparallel substitutions between M. musculus Fxy_PAR and M. spretus Fxy_PAR in the three exons (8, 9, and 10) studied (fig. 1).
FIG. 1.— Phylogeny of the Fxy genes based on exons 8, 9, and 10 (619 sites). The sequence data for the proposed M. spretus_PAR exons are from Montoya-Burgos, Boursot, and Galtier (2003), and the other sequence data are from GenBank. The scale bar is in units of numbers of substitutions. The arrows point to events proposed by Montoya-Burgos, Boursot, and Galtier (2003) and are numbered by order of occurrence.
Second, no PAR-located region of Fxy was detected in M. spretus by PCR amplification in our study (figure 1 in Supplementary Material online). Our PCR primers were specifically designed to detect exons 9 and 10 (both PAR-located) in the alleged Fxy_PAR in M. spretus. However, neither of the two exons was amplified in any of the three M. spretus individuals included in our study, although, as positive controls, both were PCR-amplified in M. musculus. Our PCR results suggest no Fxy_PAR in M. spretus. It is possible that the proposed Fxy_PAR in M. spretus was a result of PCR contamination. In support of this hypothesis, we note that Montoya-Burgos, Boursot, and Galtier's (2003) claim was based on the sequences of exons 8, 9, and 10, but only 4 nt differences were found between their proposed M. spretus Fxy_PAR exons 8, 9, and 10 and M. musculus Fxy exons 8, 9, and 10 (fig. 1).
With no evidence of Fxy_PAR in M. spretus, the movement of Fxy into the PAR occurred after the divergence of M. spretus and M. musculus—likely before the diversification of the M. musculus lineage (Perry and Ashworth 1999). Also, in the absence of a duplication of Fxy in the Mus lineages, there is no need to invoke a gene loss and a translocation of Fxy to explain Fxy_PAR in M. musculus, as proposed by Montoya-Burgos, Boursot, and Galtier (2003). It is more likely that the PAR has slid along the X chromosome in the lineage leading to M. musculus.
Mutagenicity of Recombination
In Mouse
The introns sequenced were divided into two groups, roughly the first half (intron 1 to first-half of intron 3; 2,307 sites sequenced) and the second half (intron 4 to intron 6; 2,378 sites sequenced). Only the second half in M. m. domesticus (dom) and M. m. musculus (mus) is in the PAR.
To compare the rate of nucleotide substitutions in different regions and taxa, we constructed two neighbor-joining trees for the two halves of the Fxy gene (fig. 2). The total branch length between M. caroli (car) and M. spretus (spr) is 0.0163 + 0.0077 + 0.0043 = 0.0283 for the first half (fig. 2A) and 0.0580 for the second half (fig. 2B). Therefore, the second half has evolved two times faster than the first half. On the other hand, the total branch length between dom and mus is 0.0030 for the first half and 0.0150 (five times higher) for the second half. Therefore, we may infer that because of a very high recombination rate in the PAR, the substitution rate in the Mus PAR introns has increased 5/2 = 2.5 times.
We now consider a second estimate. From figure 2, we computed the average branch length from the common ancestral node of spr, dom, and mus to the tips of dom and mus as 0.0071 + (0.0015 + 0.0015)/2 = 0.0086 for the first half and 0.0821 for the second half. The estimated substitution rate for the second half is now 9.5 times faster than that for the first half. Therefore, we may infer that the high recombination rate in the second half has increased the substitution rate by 9.5/2 = 4.8-fold. As in the above estimate, this estimate assumes that the non-PAR second half has evolved two times faster than the first half. This estimate should be taken with caution for two reasons. First, the ratio 0.0821/0.0086 = 9.5 could have been inflated because the denominator is small, so that a small error can be easily amplified. Second, it assumes that the second half has been in the PAR since the divergence between the spr lineage and the dom-mus lineage, which may not be true.
A third estimate was obtained by comparing the branch length (0.0329) from the common node of spr, dom, and mus with spr and comparing the average branch length (0.0821) from the same common node with dom and mus. This comparison gives a ratio of 2.5 for the increase in substitution rate because of the increase in recombination rate. Again, this estimate assumes that the second half has been in the PAR since the divergence between the spr and the dom-mus lineages. Despite this assumption, it might not be an underestimate because in the first half (fig. 2A) the average branch length (0.0071 + 0.0015 = 0.0086) for the dom-mus lineage is twice as long as that for the spr lineage (0.0043), suggesting the possibility of a faster clock in the dom-mus lineage than in the spr lineage.
From the above three estimates, we suggest that the increase in recombination rate in the PAR has, on average, increased the substitution rate in the PAR introns by twofold to fivefold. A very liberal upper bound is 10, which is obtained from the above second estimate (9.5) without dividing it by 2.
In Human
A weak effect of recombination on mutation rate was found in two recent studies of XG (Filatov and Gerrard 2003; Yi et al. 2004), a gene that straddles the PAB in human and apes (Rappold 1993; Lien et al. 2000). Filatov and Gerrard (2003) also studied two additional genes located in the human and ape PAR that did show twofold and threefold increases in the number of substitutions compared with the genomic average (3%) between human and orangutan. However, their analysis of X-linked and autosomal regions in human and orangutan showed no evidence of an association between mutation rate and recombination rate.
In this study, three data sets (Filatov and Gerrard 2003; Filatov 2004; Yi et al. 2004) were pooled along with the available sex-averaged recombination rates in human (Kong et al. 2002), and the human/orangutan intron divergences were correlated with recombination rates (fig. 3). For the PAR and X-linked regions, the influence of recombination rate on the substitution rate is fairly weak; that is, slope = 0.22% substitutions per cM/Mb (fig. 3A), whereas for the autosomal regions, there is no correlation and a slope of approximately 0.0 (fig. 3A).
FIG. 3.— Divergence (% substitutions per site) versus recombination rate (cM/Mb). (A) PAR and X-linked regions only. Each region consisted of the introns within a gene, except for the XG gene, which was divided into PAR and X-linked halves. (B) Autosomal regions only. Data sets obtained from (Filatov and Gerrard 2003; Filatov 2004; Yi et al. 2004) and from a manual search of the human genome.
Lercher and Hurst (2002) observed a weak correlation between human/rodent divergence and recombination rate using a lesser-resolved genetic map than is currently available. Using the more-resolved genetic map assembled by Kong et al. (2002), Hellmann et al. (2003) observed a stronger correlation (R2 = 0.29) between human/chimp divergence and recombination. Both studies had one region with a recombination rate well over 4 cM/Mb, which did not occur in our human/orangutan data set (fig. 3B). However, the analysis of the PAR (fig. 3A), with its higher rates of recombination, did show a correlation (R2 = 0.43) between divergence and recombination rate, which is consistent with the above studies.
Increase of GC Content with Recombination Rate
We inferred lineage-specific mutations in each segment sequenced and found that PAR introns have a higher rate of change from A or T (0.1081 or 0.1110) than from G or C (0.0370 or 0.0302; [table 1, and see also table 1 in Supplementary Material online]). This trend is not apparent in any of the non-PAR regions. Also, using additional sites for the comparison between M. m. musculus and M. m. domesticus revealed an average GC% of 44.3 in the PAR versus 37.2 in the non-PAR. Both results support a GC bias in the PAR (Filatov and Gerrard 2003; Filatov 2004; Yi et al. 2004).
Table 1 Relative Rates of Substitution in the First and Second Half of Fxy in Different Mus Lineages
Next, we pooled the three data sets (Filatov and Gerrard 2003; Filatov 2004; Yi et al. 2004) to study the correlation between GC% and recombination rate (fig. 4A). The estimated effect of recombination on GC (slope = 0.70 GC% per cM/Mb) was moderate.
FIG. 4.— GC (%) versus recombination rate (cM/Mb). (A) Data set same as that used in figure 2A. (B) Data set from 2669 nonoverlapping windows (of size 1 Mb) along all autosomes in the human genome. Repetitive sequences were excluded before computing the GC content.
Finally, we computed the correlation between recombination rate and the GC% in nonoverlapping windows across the human autosomes, with the exclusion of repetitive elements, and found a slope of 2.32 GC% per cM/Mb (fig. 4B). This slope is approximately three times as steep as that revealed by the human PAR and X-linked regions.
One possibility for the difference in slope between the above two data sets is that the effect of recombination on GC% is not linear. However, we did a logistic transformation of the GC values in the human PAR and X-linked regions and obtained a similar regression line. Another possibility is that the recombination rate changes with time, so that the current rate may be different from the rate in the past. A third possibility is that the effect of recombination varies among chromosomes. In this case, the two slopes (0.7 and 2.3) can represent a lower and an upper bound of the effect of recombination on GC%. A value closer to the upper bound seems more reasonable because the window size was large and not as sensitive to past changes in the pattern of recombination along the chromosomes. Galtier (2004) proposed that past events that shifted part of the PAR into a non-PAR region can lead to a great reduction in recombination rate and a slow rate of change in GC%. Under this situation, the effects of recombination near the PAB could be underestimated. If this is true, the upper bound is closer to the truth because our lower bound was estimated from regions surrounding the PAB.
Conclusions
Our data from the Fxy gene suggest only a twofold to fivefold (or at most 10-fold) increase in mutation rate caused by a dramatic increase in recombination rate in the PAR-linked regions of the Fxy gene. This finding is in stark contrast with the previous estimate of an increase of 170 times in Fxy_PAR (Perry and Ashworth 1999). Because the latter estimate was obtained using synonymous sites of exons in Fxy, it would be less reliable than our estimate, which is based on many more sites. Furthermore, our analysis of the human and orangutan data from the PAR and X-linked regions gave a regression line of y = 0.22%x + 2.61%, where x is in cM/Mb (fig. 3A). Thus, even for a recombination rate as high as 20 cM/Mb, y is 4.40% + 2.61% = 7.01%, which is approximately 2.3 times the average genomic divergence (3%) between human and orangutan. Because the recombination rate in the human genome is usually much lower than 20 cM/Mb (i.e., usually below 3.2 cM/Mb), the effect of recombination on mutation rate should generally be far weaker than that in PAR regions. On the other hand, the effect of recombination (gene conversion) on GC content appears to be much stronger than the effect on mutation rate. Our study suggests a slope between 0.7% and 2.3% per cM/Mb. Thus, an increase of 5 cM/Mb in recombination rate can lead to a 5% to 10% increase in GC content. In this respect, our results are consistent with previous studies (Marais 2003; Montoya-Burgos, Boursot, and Galtier 2003).
Acknowledgements
We thank Dr. Francois Bonhomme and Dr. Michael Nachman for sending us DNA samples and Dr. D. Filatov for making available human and orangutan sequence alignments. This study was supported by NIH grants.
References
Blake, J. A., J. E. Richardson, C. J. Bult, J. A. Kadin, J. T. Eppig,the members of the Mouse Genome Database Group. 2003. MGD: The Mouse Genome Database. Nucleic Acids Res. 31:193–195.
Broman, K. W., L. B. Rowe, G. A. Churchill, and K. Paigen. 2002. Crossover interference in the mouse. Genetics 160:1123–11231.
Brown, T. C., and J. Jiricny. 1987. A specific mismatch repair event protects mammalian cells from loss of 5-methylcytosine. Cell 50:945–950.
———. 1988. Different base/base mispairs are corrected with different efficiencies and specificities in monkey kidney cells. Cell 54:705–711.
———. 1989. Repair of base-base mismatches in simian and human cells. Genome 31:578–583.
Eyre-Walker, A. 1993. Recombination and mammalian genome evolution. Proc. R. Soc. Lond. B Biol. Sci. 252:237–243.
Filatov, D. A. 2004. A gradient of silent substitution rate in the human pseudoautosomal region. Mol. Biol. Evol. 21:410–417.
Filatov, D. A., and D. T. Gerrard. 2003. High mutation rates in human and ape pseudoautosomal genes. Gene 317:67–77.
Froenicke, L., L. K. Anderson, J. Wienberg, and T. Ashley. 2002. Male mouse recombination maps for each autosome identified by chromosome painting. Am. J. Hum. Genet. 71:1353–1368.
Fullerton, S. M., A. Bernardo Carvalho, and A. G. Clark. 2001. Local rates of recombination are positively correlated with GC content in the human genome. Mol. Biol. Evol. 18:1139–1142.
Galtier, N. 2004. Recombination, GC-content and the human pseudoautosomal boundary paradox. Trends Genet. 20:347–349.
Gojobori, T., W-H Li, and D. Graur. 1982. Patterns of nucleotide substitution in pseudogenes and functional genes. J. Mol. Evol. 18:360–369.
Hardison, R. C., K. M. Roskin, S. Yang et al. (15 co-authors). 2003. Covariation in frequencies of substitution, deletion, transposition, and recombination during eutherian evolution. Genome Res. 13:13–26.
Heidenreich, E., R. Novotny, B. Kneidinger, V. Holzmann, and U. Wintersberger. 2003. Non-homologous end joining as an important mutagenic process in cell cycle-arrested cells. EMBO J. 22:2274–2283.
Hellmann, I., I. Ebersberger, S. E. Ptak, S. Paabo, and M. Przeworski. 2003. A neutral explanation for the correlation of diversity with recombination rates in humans. Am. J. Hum. Genet. 72:1527–1535.
Holbeck, S. L., and J. N. Strathern. 1997. A role for REV3 in mutagenesis during double-strand break repair in Saccharomyces cerevisiae. Genetics 147:1017–1024.
———. 1999. EXO1 of Saccharomyces cerevisiae functions in mutagenesis during double-strand break repair. Ann. N. Y. Acad. Sci. 870:375–377.
Kong, A., D. F. Gudbjartsson, J. Sainz et al. (13 co-authors). 2002. A high-resolution recombination map of the human genome. Nat. Genet. 31:241–247.
Kumar, S., K. Tamura, I. B. Jakobsen, and M. Nei. 2001. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 17:1244–1245.
Lercher, M. J., and L. D. Hurst. 2002. Human SNP variability and mutation rate are higher in regions of high recombination. Trends. Genet. 18:337–340.
Li, W-H., C-I Wu, and C-C Luo. 1984. Nonrandomness of point mutation as reflected in nucleotide substitutions in pseudogenes and its evolutionary implications. J. Mol. Evol. 21:58–71.
Lien, S., J. Szyda, B. Schechinger, G. Rappold, and N. Arnheim. 2000. Evidence for heterogeneity in recombination in the human pseudoautosomal region: high resolution analysis by sperm typing and radiation-hybrid mapping. Am. J. Hum. Genet. 66:557–566.
Marais, G. 2003. Biased gene conversion: implications for genome and sex evolution. Trends Genet. 19:330–338.
Montoya-Burgos, J. I., P. Boursot, and N. Galtier. 2003. Recombination explains isochores in mammalian genomes. Trends Genet. 19:128–130.
Mouse Genome Sequencing Consortium. 2002. Initial sequencing and comparative analysis of the mouse genome. Nature 420:520–562.
Nachman, M. W. 2001. Single nucleotide polymorphisms and recombination rate in humans. Trends Genet. 17:481–485.
Nachman, M. W., V. L. Bauer, S. L. Crowell, and C. F. Aquadro. 1998. DNA variability and recombination rates at X-linked loci in humans. Genetics 150:1133–1141.
Palmer, S., J. Perry, D. Kipling, and A. Ashworth. 1997. A gene spans the pseudoautosomal boundary in mice. Proc. Natl. Acad. Sci. USA 94:12030–12035.
Perry, J., and A. Ashworth. 1999. Evolutionary rate of a gene affected by chromosomal position. Curr. Biol. 9:987–989.
Perry, J., S. Palmer, A. Gabriel, and A. Ashworth. 2001. A short pseudoautosomal region in laboratory mice. Genome Res. 11:1826–1832.
Rappold, G. A. 1993. The pseudoautosomal regions of the human sex chromosomes. Hum. Genet. 92:315–324.
Saitou, N., and M. Nei. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406–425.
Schiebel, K., J. Meder, A. Rump, A. Rosenthal, M. Winkelmann, C. Fischer, T. Bonk, A. Humeny, and G. Rappold. 2000. Elevated DNA sequence diversity in the genomic region of the phosphatase PPP2R3L gene in the human pseudoautosomal region. Cytogenet. Cell Genet. 91:224–230.
Strathern, J. N., B. K. Shafer, and C. B. McGill. 1995. DNA synthesis errors associated with double-strand-break repair. Genetics 140:965–972.
Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and D. G. Higgins. 1997. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25:4876–4882.
Yi, S., T. J. Summers, N. M. Pearson, and W-H Li. 2004. Recombination has little effect on the rate of sequence divergence in pseudoautosomal boundary 1 among humans and great apes. Genome Res. 14:37–43.(Shiao-Wei Huang*,, Robert)