Leishmania braziliensis causing human disease in Northeast Brazil presents loci with genotypes in long-term equilibrium

Abstract

Background: Leishmaniases are neglected tropical diseases that inflict great burden to poor areas of the globe. Intense research has aimed to identify parasite genetic signatures predictive of infection outcomes. Consistency of diagnostic tools based on these markers would greatly benefit from accurate understanding of Leishmania spp. population genetics. We explored two chromosomal loci to characterize a population of L. braziliensis causing human disease in Northeast Brazil.

Methodology/principal findings: Two temporally distinct samples of L. braziliensis were obtained from patients attending the leishmaniasis clinic at the village of Corte de Pedra: (2008-2011) primary sample, N = 120; (1999-2001) validation sample, N = 35. Parasites were genotyped by Sanger's sequencing of two 600 base pairs loci starting at nucleotide positions 3,074 and 425,451 of chromosomes 24 and 28, respectively. Genotypes based on haplotypes of biallelic positions in each locus were tested for several population genetic parameters as well as for geographic clustering within the region. Ample geographic overlap of genotypes at the two loci was observed as indicated by non-significant Cusick and Edward's comparisons. No linkage disequilibrium was detected among combinations of haplotypes for both parasite samples. Homozygous and heterozygous genotypes displayed Hardy-Weinberg equilibrium (HWE) at both loci in the two samples when straight observed and expected counts were compared by Chi-square (p>0.5). However, Bayesian statistics using one million Monte-Carlo randomizations disclosed a less robust HWE for chromosome 24 genotypes, particularly in the primary sample (p = 0.04). Fixation indices (Fst) were consistently lower than 0.05 among individuals of the two samples at both tested loci, and no intra-populational structuralization could be detected using STRUCTURE software.

Conclusions/significance: These findings suggest that L. braziliensis can maintain stable populations in foci of human leishmaniasis and are capable of robust genetic recombination possibly due to events of sexual reproduction during the parasite's lifecycle.

Fig 1. Time distributions of L. braziliensis isolations for the primary (2008–2011) and validation (1999–2001) parasite samples used in the study.

(A) Primary sample of parasite isolates obtained from ATL patients enrolled during six consecutive semesters from July/2008 to June/2011 in Corte de Pedra, Northeast Brazil. (B) Validation sample obtained from ATL patients enrolled during five consecutive semesters from July/1999 to December/2001 in the same region. Columns depict the number of isolates obtained per semester, according to the left Y axes. Lines consist in percent contribution of each semester to the total number of isolates in the corresponding sample, according to right Y axes.

Fig 2. Haplotypes of nucleotides found in biallelic polymorphic positions at chromosomal loci CHR24/3074 and CHR28/425451 of at least two different isolates in the collection of L. braziliensis used in the study.

For each locus, DNA sequences of four to six clones per L. braziliensis isolate of the study collection (i.e. primary plus validation samples) were aligned using MEGA X. Multiple alignment across all study isolates permitted identify polymorphic nucleotides at biallelic positions 31, 52, 68, 129, 469 and 604 in locus CHR24/3074, and 30, 286 and 545 in CHR28/425451. (A) Haplotypes of nucleotides detected within polymorphic positions at CHR24/3074; (B) Haplotypes of polymorphic nucleotides detected in CHR28/425451. The proportional representation of each haplotype in the sample is within parenthesis at the tips of the dendrograms.

Fig 3. Time distributions of homozygous and heterozygous genotypes in CHR24/3074 and CHR28/425451.

Each L. braziliensis isolate in the collection was genotyped at each locus according to their haplotype contents, considering diploid genomes. (A and C) Proportions of CTTCAG: CTTCAG (green), CTTCAG: TCATGA (red) and TCATGA: TCATGA (blue) in CHR24/3074 for the primary (A; i.e. 2008–2011) and validation (C; i.e. 1999–2001) parasite samples, respectively. (B and D) Proportions of CCT:CCT (green), CCT:TT- (red) and TT-:TT- (blue) in CHR28/425451 for the primary (B) and validation (D) samples, respectively. Overall fluctuations of genotype frequencies within each locus in each sample were not statistically significant (Chi-square p>0.05).

Fig 4. Homozygous and heterozygous genotypes in L. braziliensis loci greatly overlap in Corte de Pedra.

Homozygous and heterozygous genotypes observed in L. braziliensis loci CHR24/3074 and CHR28/425451 of parasites obtained during 1999–2001 (A and B) and 2008–2011 (C and D) sampling periods were mapped and the resulting sets of geographic events were statistically compared. CHR24/3074 (A and C): CTTCAG: CTTCAG (yellow), CTTCAG: TCATGA (red) and TCATGA: TCATGA (blue). CHR28/425451 (B and D): CCT:CCT (yellow), CCT:TT- (red) and TT-:TT- (blue). Cuzick and Edward´s comparisons of combined homozygous versus heterozygous spatial distributions rendered non-significant (p>0.05) for data depicted in all four maps of the panel. Total number of dots plotted in each map may be smaller than the number of corresponding parasite isolates described in text due to overlap of some ATL patients’ geographic coordinates. Link to the United States Geological Survey (USGS) Landsat satellite photograph used in the figure [39]: https://earthexplorer.usgs.gov/scene/metadata/full/5e83d1193824e4fc/LT52160691994219CUB00/.

Fig 5. L. braziliensis individuals contribute similarly to putative subpopulations predicted by Structure, based on their genotypes in CHR24/3074 and CHR28/425451 loci, in Corte de Pedra.

Structuralization within Corte de Pedra’s L. braziliensis population into discrete subpopulations was assessed using the software STRUCTURE [21]. The analyses were performed using combined data of CHR24/3074 and CHR28/425451 loci. Two conditions were tested: (A) structuralization into two subpopulations, i.e. K = 2; and (B) structuralization into 20 subpopulations, i.e. K = 20. Histograms show the proportional contributions of each individual in the overall sample to each theoretical subpopulation ordered by sampling period (i.e. 1999–2001 and 2008–2011) after a burn-in step of 10,000 runs followed by 100,000 MCMC replications of the data.