Genomic heterogeneity in the density of noncoding single-nucleotide and microsatellite polymorphisms in Plasmodium falciparum

Abstract

The density and distribution of single-nucleotide polymorphisms (SNPs) across the genome has important implications for linkage disequilibrium mapping and association studies, and the level of simple-sequence microsatellite polymorphisms has important implications for the use of oligonucleotide hybridization methods to genotype SNPs. To assess the density of these types of polymorphisms in P. falciparum, we sampled introns and noncoding DNA upstream and downstream of coding regions among a variety of geographically diverse parasites. Across 36,229 base pairs of noncoding sequence representing 41 genetic loci, a total of 307 polymorphisms including 248 polymorphic microsatellites and 39 SNPs were identified. We found a significant excess of microsatellite polymorphisms having a repeat unit length of one or two, compared to those with longer repeat lengths, as well as a nonrandom distribution of SNP polymorphisms. Almost half of the SNPs localized to only three of the 41 genetic loci sampled. Furthermore, we find significant differences in the frequency of polymorphisms across the two chromosomes (2 and 3) examined most extensively, with an excess of SNPs and a surplus of polymorphic microsatellites on chromosome 3 as compared to chromosome 2 (P=0.0001). Furthermore, at some individual genetic loci we also find a nonrandom distribution of polymorphisms between coding and flanking noncoding sequences, where completely monomorphic regions may flank highly polymorphic genes. These data, combined with our previous findings of nonrandom distribution of SNPs across chromosome 2, suggest that the Plasmodium falciparum genome may be a mosaic with regard to genetic diversity, containing chromosomal regions that are highly polymorphic interspersed with regions that are much less polymorphic.

Figure 1

Microsatellites with a Repeat Unit Length of 1 or 2 are Highly Polymorphic. Microsatellite (MS) repeats of unit length one or two were significantly more polymorphic than microsatellite repeats of unit length three or more (P = 5.2E-31) and there was an excess of these short-repeat microsatellites on chromosome 3 relative to chromosome 2 (P = 0.026).

Figure 2

SNPs are Nonrandomly Distributed Between Coding and Noncoding Regions of a Gene. The pfmdr1 gene (PFE1150w; coding region from position 957,885 – 962,144; shown from position 955,110 – 962,912) contains SNPs both within the coding region (arrows corresponding to Tyr86, Phe184, Cys1034, Asp1042, Tyr1246) and in the 5′ (2775 base pairs (bp)) and 3′ (768 bp) flanking regions. One of these SNPs (dotted arrow) is present in both field and laboratory isolates. Several of these mutations may be functionally relevant due to the proximity to the start of transcription (curved arrow). The msp1 gene (PFI1475w; coding region from position 1,201,802 – 1,206,964; shown from position 1,200,803 – 1,208,021) contains several SNPs within the coding region (represented by arrows) within the highly polymorphic blocks (2, 4, 6, 8, 10, 14, 16) defined by Tanabe and colleagues (Miller et al., 1983) shown as conserved (open), semiconserved (hatched), and variable (solid) regions. Most of this variation is dimorphic and grouped into the allelic groups of K1-type and MAD20-type alleles. Exception to this dimorphism is present in Block 2 with additional degenerate tripeptide repeats that represent the RO33-type allele. Both the 5′ (999 bp) and the 3′ (1057 bp) flanking regions are monomorphic among eight or more genetically diverse parasites and lack SNPs in the dataset analyzed.