Recombination, gene conversion, and identity-by-descent at three loci

We investigate the probabilities of identity-by-descent at three loci in order to find a signature which differentiates between the two types of crossing over events: recombination and gene conversion. We use a Markov chain to model coalescence, recombination, gene conversion and mutation in a sample of size two. Using numerical analysis, we calculate the total probability of identity-by-descent at the three loci, and partition these probabilities based on a partial ordering of coalescent events at the three loci. We use these results to compute the probabilities of four different patterns of conditional identity and non-identity at the three loci under recombination and gene conversion. Although recombination and gene conversion do make different predictions, the differences are not likely to be useful in distinguishing between them using three locus patterns between pairs of DNA sequences. This implies that measures of genetic identity in larger samples will be needed to distinguish between gene conversion and recombination.     

Analytical framework for identifying and differentiating recent hitchhiking and severe bottleneck effects from multi-locus DNA sequence data

Hitchhiking and severe bottleneck effects have impact on the dynamics of genetic diversity of a population by inducing homogenization at a single locus and at the genome-wide scale, respectively. As a result, identification and differentiation of the signatures of such events from DNA sequence data at a single locus is challenging. This paper develops an analytical framework for identifying and differentiating recent homogenization events at multiple neutral loci in low recombination regions. The dynamics of genetic diversity at a locus after a recent homogenization event is modeled according to the infinite-sites mutation model and the Wright-Fisher model of reproduction with constant population size. In this setting, I derive analytical expressions for the distribution, mean, and variance of the number of polymorphic sites in a random sample of DNA sequences from a locus affected by a recent homogenization event. Based on this framework, three likelihood-ratio based tests are presented for identifying and differentiating recent homogenization events at multiple loci. Lastly, I apply the framework to two data sets. First, I consider human DNA sequences from four non-coding loci on different chromosomes for inferring evolutionary history of modern human populations. The results suggest, in particular, that recent homogenization events at the loci are identifiable when the effective human population size is 50,000 or greater in contrast to 10,000, and the estimates of the recent homogenization events are agree with the "Out of Africa" hypothesis. Second, I use HIV DNA sequences from HIV-1-infected patients to infer the times of HIV seroconversions. The estimates are contrasted with other estimates derived as the mid-time point between the last HIV-negative and first HIV-positive screening tests. The results show that significant discrepancies can exist between the estimates.     

Statistical methods of DNA sequence analysis: detection of intragenic recombination or gene conversion

Simple but exact statistical tests for detecting a cluster of associated nucleotide changes in DNA are presented. The tests are based on the linear distribution of a set of s sites among a total of n sites, where the s sites may be the variable sites, sites of insertion/deletion, or categorized in some other way. These tests are especially useful for detecting gene conversion and intragenic recombination in a sample of DNA sequences. In this case, the sites of interest are those that correspond to particular ways of splitting the sequences into two groups (e.g., sequences A and D vs. sequences B, C, and E-J). Each such split is termed a phylogenetic partition. Application of these methods to a well-documented case of gene conversion in human gamma-globin genes shows that sites corresponding to two of the three observed partitions are significantly clustered, whereas application to hominoid mitochondrial DNA sequences--among which no recombination is expected to occur--shows no evidence of such clustering. This indicates that clustering of partition-specific sites is largely due to intragenic recombination or gene conversion. Alternative hypotheses explaining the observed clustering of sites, such as biased selection or mutation, are discussed.      

Meiotic Recombination between Duplicated Genetic Elements in SACCHAROMYCES CEREVISIAE

We have studied the meiotic recombination behavior of strains carrying two types of duplications of an 18.6-kilobase HIS4 Bam HI fragment. The first type is a direct duplication of the HIS4 Bam HI fragment in which the repeated sequences are separated by Escherichia coli plasmid sequences. The second type, a tandem duplication, has no sequences intervening between the repeated yeast DNA. The HIS4 genes in each region were marked genetically so that recombination events between the duplicated segments could be identified. Meiotic progeny of the strains carrying the duplication were analyzed genetically and biochemically to determine the types of recombination events that had occurred. Analysis of the direct vs. tandem duplication suggests that the E. coli plasmid sequences are recombinogenic in yeast when homozygous. In both types of duplications recombination between the duplicated HIS4 regions occurs at high frequency and involves predominantly interchromosomal reciprocal exchanges (equal and unequal crossovers). The striking observation is that intrachromosomal reciprocal recombination is very rare in comparison with interchromosomal reciprocal recombination. However, intrachromosomal gene conversion occurs at about the same frequency as interchromosomal gene conversion. Reciprocal recombination events between regions on the same chromatid are the most infrequent exchanges. These data suggest that intrachromosomal reciprocal exchanges are suppressed.     

Effect of repeat copy number on variable-number tandem repeat mutations in Escherichia coli O157:H7

Variable-number tandem repeat (VNTR) loci have shown a remarkable ability to discriminate among isolates of the recently emerged clonal pathogen Escherichia coli O157:H7, making them a very useful molecular epidemiological tool. However, little is known about the rates at which these sequences mutate, the factors that affect mutation rates, or the mechanisms by which mutations occur at these loci. Here, we measure mutation rates for 28 VNTR loci and investigate the effects of repeat copy number and mismatch repair on mutation rate using in vitro-generated populations for 10 E. coli O157:H7 strains. We find single-locus rates as high as 7.0 x 10(-4) mutations/generation and a combined 28-locus rate of 6.4 x 10(-4) mutations/generation. We observed single- and multirepeat mutations that were consistent with a slipped-strand mispairing mutation model, as well as a smaller number of large repeat copy number mutations that were consistent with recombination-mediated events. Repeat copy number within an array was strongly correlated with mutation rate both at the most mutable locus, O157-10 (r2= 0.565, P = 0.0196), and across all mutating loci. The combined locus model was significant whether locus O157-10 was included (r2= 0.833, P < 0.0001) or excluded (r2= 0.452, P < 0.0001) from the analysis. Deficient mismatch repair did not affect mutation rate at any of the 28 VNTRs with repeat unit sizes of >5 bp, although a poly(G) homomeric tract was destabilized in the mutS strain. Finally, we describe a general model for VNTR mutations that encompasses insertions and deletions, single- and multiple-repeat mutations, and their relative frequencies based upon our empirical mutation rate data.     

Effect of limited homology on gene conversion in a Saccharomyces cerevisiae plasmid recombination system.

Plasmids containing heteroallelic copies of the Saccharomyces cerevisiae HIS3 gene undergo intramolecular gene conversion in mitotically dividing S. cerevisiae cells. We have used this plasmid system to determine the minimum amount of homology required for gene conversion, to examine how conversion tract lengths are affected by limited homology, and to analyze the role of flanking DNA sequences on the pattern of exchange. Plasmids with homologous sequences greater than 2 kilobases have mitotic exchange rates as high as 2 x 10(-3) events per cell per generation. As the homology is reduced, the exchange rate decreases dramatically. A plasmid with 26 base pairs (bp) of homology undergoes gene conversion at a rate of approximately 1 x 10(-10) events per cell per generation. These studies have also shown that an 8-bp insertion mutation 13 bp from a border between homologous and nonhomologous sequences undergoes conversion, but that a similar 8-bp insertion 5 bp from a border does not. Examination of independent conversion events which occurred in plasmids with heteroallelic copies of the HIS3 gene shows that markers within 280 bp of a border between homologous and nonhomologous sequences undergo conversion less frequently than the same markers within a more extensive homologous sequence. Thus, proximity to a border between homologous and nonhomologous sequences shortens the conversion tract length.     

High-frequency gene conversion between repeated C mu sequences integrated at the chromosomal immunoglobulin mu locus in mouse hybridoma cells.

The occurrence of mitotic recombination between repeated immunoglobulin mu gene constant (C mu) region sequences stably integrated at the haploid chromosomal immunoglobulin mu locus in murine hybridoma cells was investigated. Recombination events are detected as changes in hapten-specific immunoglobulin M production. Recombination occurs with high frequency (0.5 to 0.8%) by a mechanism consistent with gene conversion. A double-strand break repair-like mechanism is suggested by the finding that repair of a 2-bp deletion mutation and a 2-bp insertion mutation occurs with parity in a donor-directed manner. The results also suggest that the gene conversion process is directional in that the 5' C mu region sequence is preferentially converted.     

Frequency and Directionality of Gene Conversion Events Involving the CYC7-H3 Mutation in SACCHAROMYCES CEREVISIAE

The CYC7-H3 mutation is a 5-kb deletion that causes overproduction of iso-2 cytochrome c. Unlike most mutations in yeast, the CYC7-H3 mutation is preferentially lost when it is involved in a gene conversion event. We have shown that cloned copies of CYC7-H3 DNA that are inserted into the yeast genome are associated with a high frequency of recombination and aberrant segregation events. Since parity in conversion frequency was observed when the extensive insertion/deletion heterozygosity at this locus was eliminated, we conclude that the CYC7-H3 sequences are inherently capable of acting as donors or recipients in gene conversion events, although they are unlikely to act as donors when they are located opposite a large heterology. DNA sequence comparisons revealed similarities between the CYC7-H3 junction region and the 2-micron circle DNA region that is involved in site-specific recombination.     

Direct analysis by small-pool PCR of MS205 minisatellite mutation rates in sperm after mutagenic therapies.

We have used small-pool PCR to analyse mutation in samples of sperm taken from men after mutagenic therapy. Small-pool PCR uses direct analysis of germline DNA at a highly unstable tandem-repeated "minisatellite" locus to measure rates of length-change mutation in individual sperm samples. The advantages of this approach are that the normal mutation rate is extremely high (about 0.4% per gamete at the locus analysed here), so that relatively small increases in mutation rate can be detectable in individual samples. It is known from work on sperm from untreated individuals that different alleles at this locus have different mutation rates. For this reason, we have analysed the germline mutation rates in sperm samples from two men, in each case comparing a post-treatment sample with a pre-treatment sample from the same individual. We find no evidence for altered mutation in the post-treatment sample, suggesting that the repopulation of the germ-cell compartment after treatment may be subject to stringent mechanisms for the detection and elimination of germ-cell damage.     

Rates of mutation and host transmission for an Escherichia coli clone over 3 years

Although over 50 complete Escherichia coli/Shigella genome sequences are available, it is only for closely related strains, for example the O55:H7 and O157:H7 clones of E. coli, that we can assign differences to individual evolutionary events along specific lineages. Here we sequence the genomes of 14 isolates of a uropathogenic E. coli clone that persisted for 3 years within a household, including a dog, causing a urinary tract infection (UTI) in the dog after 2 years. The 20 mutations observed fit a single tree that allows us to estimate the mutation rate to be about 1.1 per genome per year, with minimal evidence for adaptive change, including in relation to the UTI episode. The host data also imply at least 6 host transfer events over the 3 years, with 2 lineages present over much of that period. To our knowledge, these are the first direct measurements for a clone in a well-defined host community that includes rates of mutation and host transmission. There is a concentration of non-synonymous mutations associated with 2 transfers to the dog, suggesting some selection pressure from the change of host. However, there are no changes to which we can attribute the UTI event in the dog, which suggests that this occurrence after 2 years of the clone being in the household may have been due to chance, or some unknown change in the host or environment. The ability of a UTI strain to persist for 2 years and also to transfer readily within a household has implications for epidemiology, diagnosis, and clinical intervention.     

Contrasting Roles of Interallelic Recombination at the Hla-a and Hla-B Loci

A statistical study of DNA sequences of alleles at the highly polymorphic class I MHC loci of humans, HLA-A and HLA-B, showed evidence of both large-scale recombination events (involving recombination of exons 1-2 of one allele with exons 3-8 of another) and small-scale recombination events (involving apparent exchange of short DNA segments). The latter events occurred disproportionately in the region of the gene encoding the antigen recognition site (ARS) of the class I molecule. Furthermore, they involved the ARS codons which are under the strongest selection favoring allelic diversity at the amino acid level. Thus, the frequency of recombinant alleles appears to have been increased by some form of balancing selection (such as overdominant selection) favoring heterozygosity in the ARS. These analyses also revealed a striking difference between the A and B loci. Recombination events appear to have occurred about twice as frequently at the B locus, and recombinants at the B locus were significantly more likely to affect polymorphic sites in the ARS. At the A locus, there are well-defined allelic lineages that have persisted since prior to the human-chimpanzee divergence; but at the B locus, there is no evidence for such long-lasting allelic lineages. Thus, relatively frequent interallelic recombination has apparently been a feature of the long-term evolution of the B locus but not of the A locus.     

Escherichia coli frameshift mutation rate depends on the chromosomal context but not on the GATC content near the mutation site

Different studies have suggested that mutation rate varies at different positions in the genome. In this work we analyzed if the chromosomal context and/or the presence of GATC sites can affect the frameshift mutation rate in the Escherichia coli genome. We show that in a mismatch repair deficient background, a condition where the mutation rate reflects the fidelity of the DNA polymerization process, the frameshift mutation rate could vary up to four times among different chromosomal contexts. Furthermore, the mismatch repair efficiency could vary up to eight times when compared at different chromosomal locations, indicating that detection and/or repair of frameshift events also depends on the chromosomal context. Also, GATC sequences have been proved to be essential for the correct functioning of the E. coli mismatch repair system. Using bacteriophage heteroduplexes molecules it has been shown that GATC influence the mismatch repair efficiency in a distance- and number-dependent manner, being almost nonfunctional when GATC sequences are located at 1 kb or more from the mutation site. Interestingly, we found that in E. coli genomic DNA the mismatch repair system can efficiently function even if the nearest GATC sequence is located more than 2 kb away from the mutation site. The results presented in this work show that even though frameshift mutations can be efficiently generated and/or repaired anywhere in the genome, these processes can be modulated by the chromosomal context that surrounds the mutation site.     

Tetrad analysis shows that gene conversion is the major mechanism involved in mutation at the human minisatellite MS1 integrated in Saccharomyces cerevisiae

Minisatellites are arrays of tandemly repeated DNA sequences which occur at thousands of locations in the human genome. They are frequently hypervariable with respect to allele length as a result of high rates of complex and incompletely understood recombination-based germline mutation events that alter the repeat copy number. MS1 is one of the most variable minisatellites so far isolated from the human genome. We have integrated MS1, flanked by synthetic markers, in the vicinity of a hot spot for meiotic double-strand breaks upstream of the LEU2 locus in chromosome III of Saccharomyces cerevisiae. Here we present the first tetrad analysis of mutations at a human minisatellite locus. The data showed that mutant alleles occur as single mutants in one of the spores in a tetrad, also when the mutant structure was the result of a combination of intra- and inter-allelic rearrangements. The conversional transfer of repeat units from one allele to the other was associated with flanking marker conversion which always involved the same flank of the minisatellite. The results demonstrate that conversion is the predominant mechanism by which minisatellite alleles mutate to new lengths, and also support the assumption that cis-acting elements are involved in the regulation of the mutational process in humans.     

Identification of two new genetically active regions associated with the osmZ locus of Escherichia coli: role in regulation of proU expression and mutagenic effect of cya, the structural gene for adenylate cyclase.

The Escherichia coli K-12 gene coding for the nucleoid-associated protein HNS was cloned together with 5.6 kb of downstream DNA in the vector pACYC184. The cloned DNA complemented a mutation in the osmZ locus of E. coli, which codes for HNS. However, the multicopy plasmid harboring the cloned sequence was found to be mutagenic and to produce at high frequency mutations that mapped to the E. coli cya gene, which codes for adenylate cyclase. Acquisition of the cya mutations was independent of RecA. These mutations were phenotypically suppressed by providing the cells with exogenous cyclic AMP and were complemented in trans by a plasmid carrying an active copy of the cya gene. A deletion analysis of the cloned sequences showed that DNA downstream of the gene coding for HNS was also required for the mutagenic effect of cya and had a role in regulating the expression of the osmZ-dependent proU locus. These sequences appear to contain at least two genetically active regions.     

The effect of recurrent mutation on the frequency spectrum of a segregating site and the age of an allele

The sample frequency spectrum of a segregating site is the probability distribution of a sample of alleles from a genetic locus, conditional on observing the sample to be polymorphic. This distribution is widely used in population genetic inferences, including statistical tests of neutrality in which a skew in the observed frequency spectrum across independent sites is taken as a signature of departure from neutral evolution. Theoretical aspects of the frequency spectrum have been well studied and several interesting results are available, but they are usually under the assumption that a site has undergone at most one mutation event in the history of the sample. Here, we extend previous theoretical results by allowing for at most two mutation events per site, under a general finite allele model in which the mutation rate is independent of current allelic state but the transition matrix is otherwise completely arbitrary. Our results apply to both nested and nonnested mutations. Only the former has been addressed previously, whereas here we show it is the latter that is more likely to be observed except for very small sample sizes. Further, for any mutation transition matrix, we obtain the joint sample frequency spectrum of the two mutant alleles at a triallelic site, and derive a closed-form formula for the expected age of the younger of the two mutations given their frequencies in the population. Several large-scale resequencing projects for various species are presently under way and the resulting data will include some triallelic polymorphisms. The theoretical results described in this paper should prove useful in population genomic analyses of such data.     

Identification and characterization of a gene responsible for inhibiting propagation of methylated DNA sequences in mcrA mcrB1 Escherichia coli strains.

Identifying and eliminating endogenous bacterial enzyme systems can significantly increase the efficiency of propagation of eukaryotic DNA in Escherichia coli. We have recently examined one such system which inhibits the propagation of lambda DNA rescued from transgenic mouse tissues. This rescue procedure utilizes lambda packaging extracts for excision of the lambda DNA from the transgenic mouse genome, as well as E. coli cells for subsequent infection and propagation. This assay, in combination with conjugal mating, P1 transduction, and gene cloning, was used to identify and characterize the E. coli locus responsible for this difference in efficiency. It was determined that the E. coli K-12 mcrB gene when expressed on a high-copy-number plasmid can cause a decrease in rescue efficiency despite the presence of the mcrB1 mutation, which inactivates the classic McrB restriction activity. (This mutation was verified by sequence analysis.) However, this McrB1 activity is not observed when the cloned mcrB1 gene is inserted into the E. coli genome at one copy per chromosome. A second locus was identified which causes a decrease in rescue efficiency both when expressed on a high-copy-number plasmid and when inserted into the genome. The data presented here suggest that this locus is mrr and that the mrr gene product can recognize and restrict cytosine-methylated sequences. Removal of this DNA region including the mrr gene from E. coli K-12 strains allows high rescue efficiencies equal to those of E. coli C strains. These modified E. coli K-12 plating strains and lambda packaging extract strains should also allow a significant improvement in the efficiency and representation of eukaryotic genomic and cDNA libraries.     

Testing for recombination in a short nuclear DNA sequence of the European meadow grasshopper, Chorthippus parallelus

Single-copy nuclear DNA sequences have high potential as a source of genetic markers for population analyses. However, the difficulties that arise when haplotypes that are the product of recombinational rearrangements are present require additional consideration. Two statistical methods for identifying potential recombinants by detecting anomalies in the distribution of variable sites along sequences were used to screen sequences from a single-copy nuclear DNA fragment, cpnl-1, of the European meadow grasshopper (Chorthippus parallelus). Five of the 71 haplotypes in the cpnl-1 data set showed nonrandom distribution of polymorphic sites using both methods. The second method pinpointed an additional four haplotypes. Estimates of the rate of recombination in the entire data set were obtained using standard methods. It is concluded that cpnl-1 haplotypes have been involved in recombination or gene conversion events at a rate more than twice the mutation rate. This confirms that recombination and gene conversion are significant factors in the generation of haplotype variation in nuclear gene sequences. The cpnl-1 haplotypes identified by the tests were present only in populations that have had recent contact; the Balkan and Turkish refugial populations and their post-glacial colonies to the north. This is discussed in relation to the phylogenetic inferences drawn from the same data in a previous report.     

A single base change in the acceptor stem of tRNA(3Leu) confers resistance upon Escherichia coli to the calmodulin inhibitor, 48/80

We have isolated several classes of spontaneous mutants resistant to the calmodulin inhibitor 48/80 which inhibits cell division in Escherichia coli K12. Several mutants were also temperature sensitive for growth and this property was exploited to clone a DNA fragment from an E. coli gene library restoring growth at 42 degrees C and drug sensitivity at 30 degrees C in one such mutant. Physical and genetic mapping confirmed that both the mutation and the cloned DNA were located at 15.5 min on the E. coli chromosome at a locus designated feeB. By subcloning, complementation analysis and sequencing, the feeB locus was identified as identical to the tRNA(CUALEU) gene. When the mutant locus was isolated and sequenced, the mutation was confirmed as a single base change, C to A, at position 77 in the acceptor stem of this rare Leu tRNA. In other studies we obtained evidence that this mutant tRNA, recognizing the rare Leu codon, CUA, was defective in translation at both permissive and non-permissive temperatures. The feeB1 mutant is defective in division and shows a reduced growth rate at non-permissive temperature. We discuss the possibility that the mutant tRNA(3Leu) is limiting for the synthesis of a polypeptide(s), requiring several CUA codons for translation which in turn regulates in some way the level or activity of the drug target, a putative cell cycle protein.     

Recent gene conversions between duplicated glutamate decarboxylase genes (gadA and gadB) in pathogenic Escherichia coli

Escherichia coli have evolved adaptive systems to resist strongly acidic habitats in part through the production of 2 biochemically identical isoforms of glutamate decarboxylase (GAD), encoded by the gadA and gadB genes. These genes occur in E. coli and other members of the genospecies (e.g., Shigella spp.) and originated as part of a genomic fitness island acquired early in Escherichia evolution. The present duplicated gad loci are widely spaced on the E. coli chromosome, and the 2 genes are 97% similar in sequence. Comparison of the nucleotide sequences of the gadA and gadB in 16 strains of pathogenic E. coli revealed 3.8% and 5.0% polymorphism in the 2 genes, respectively. Alignment of the homologous genes identified a total of 120 variable sites, including 21 fixed nucleotide differences between the loci within the first 82 codons of the genes. Twenty-three phylogenetically informative sites were polymorphic for the same nucleotides in both genes suggesting recent gene conversions or intergenic recombination. Phylogenetic analysis based on the synonymous substitutions per synonymous site indicated 2 cases in which specific gadA and gadB alleles were more closely related to one another than to other alleles at the corresponding locus. The results indicate that at least 3 gene conversion events have occurred after the gad gene duplication in the evolution of E. coli. Despite multiple gene conversion events, the upstream regulatory regions and the 5' end of each gene remains distinct, suggesting that maintaining functionally different gad genes is important in this acid-resistance mechanism in pathogenic E. coli.     

Enzyme polymorphism and genetic population structure in Escherichia coli and Shigella

Electrophoretically demonstrable variation in 12 enzymes was studied in more than 1 600 isolates of Escherichia coli from human and animal sources and in 123 strains of the four species of Shigella. All 12 enzymes were polymorphic; and the number of allozymes (mobility variants), which were equated with alleles, averaged 9.3 per locus in E. coli. For Shigella species, the mean number of alleles was 2.9 per locus. Some 77% of the allozymes recorded in Shigella were shared with E. coli. A total of 302 unique genotypic combinations of alleles over the 12 loci (electrophoretic types, ETs) was distinguished, of which 279 represented E. coli and 23 were Shigella. Among electrophoretic types, mean allelic diversity per locus was 0.52 for E. coli and 0.29 for Shigella. It was estimated that there are, on the average, about 0.3 detectable codon differences per locus between pairs of strains of E. coli and Shigella, which is roughly equivalent to 1.2 amino acid differences per enzyme. Evidence that the enzyme loci studied are a random sample of the genome is provided by a significant positive correlation between estimates of genetic divergence between pairs of strains obtained by DNA reassociation tests and estimates of genetic distance between the same strains based on electrophoresis. A principal components analysis of allozyme profiles revealed that the 302 ETs fall into three overlapping clusters, reflecting strong non-random associations of alleles, largely at four loci. Each of the four ETs of E. coli that have been most frequently recovered from natural populations has an allozyme profile that is very similar to, or identical with, the hypothetical modal ET of one of the groups. ETs of Shigella fall into two of the groups. No biological significance can at present bbe attributed to the genetic structure revealed by Multilocus electrophoretic techniques. The electrophoretic data are fully compatible with other molecular and more conventional evidence of a close affinity between E. coli and Shigella, and they raise questions regarding the present assignments of certain strains to species. In support of evidence from DNA reassociation tests and serotyping, the present study suggests that S. sonnei is homogeneous in chromosomal genotype.