Fig 1.
Drug resistance evolution from standing variation or de novo mutation.
Drug resistance alleles spread rapidly and reach fixation with a high probability when drug resistance alleles are already present as standing variation. In contrast, if resistance alleles are absent when treatment is initiated, resistance mutation must arise de novo, so there is a waiting time before the resistance alleles appear, and most resistance alleles are lost by genetic drift, so the probability of establishment and fixation is low.
Table 1.
Summary of the samples.
Table 2.
Summary statistics for coding sequence variation in SmSULT-OR.
Fig 2.
Mapping of the resistance mutations on the gene sequence and structure of Schistosoma mansoni SmSULT-OR sulfotransferase.
Exon 1 and exon 2 are represented in orange and beige, respectively. Single nucleotide polymorphisms and duplication/deletion events are represented in cyan and magenta, respectively. (A) Linear representation of the SmSULT-OR gene showing the relative position of the mutations and their translation in amino acid sequences. (B) Positions of mutations on the SmSULT-OR protein. Oxamniquine is represented in yellow, 3’-phosphoadenosine-5’-phosphosulfate (PAPS) co-factor is represented in green.
Fig 3.
Enzymatic activity of recombinant Schistosoma mansoni SmSULT-OR sulfotransferase expressed from different allelic variants.
This in vitro oxamniquine activation assay quantifies DNA-oxamniquine complexes by scintillation (counts per minute). (A) Bars show the mean of three replicates, while error bars are S.E.M. (B) The triangular matrix shows the p-values from the pairwise comparisons (Tukey’s HSD). The color is proportional to the level of significance, from white (not significant) to red (most significant). Enzyme carrying known loss-of-function mutations, such as p.C35R or p.E142del, as well as a newly identified variant (p.L179P) showed no or low oxamniquine activation, while two newly identified variants (p.S160L and p.P225S) showed intermediate activation. The newly identified p.P106S did not impair oxamniquine activation.
Fig 4.
In silico evaluation of mutations on protein stability.
We computed the difference in free enthalpy (ΔΔG) between the mutated and the wild type proteins. The higher the ΔΔG, the more unstable is the mutated protein. Only single amino acid changes within the resolved crystal structure were examined. For completeness, we included mutations from our current dataset and from previous studies [10,13]. Grey bars correspond to known sensitive alleles. Red bars correspond to validated resistance alleles. Grey labels correspond to mutations identified previously from South America [13].
Table 3.
OXA-R mutations scored in the Schistosoma mansoni SmSULT-OR gene and their frequency in the New and Old World.
Fig 5.
World map showing proportion of resistance and sensitive alleles in South America (Brazil), West Africa (Senegal and Niger), East Africa (Tanzania) and Middle East (Oman).
Number of homozygous (hmz) and heterozygous (htz) parasites for resistance alleles, and total number of parasites sampled (N) are shown below the pie charts. East Africa showed the highest frequency of resistance alleles and resistant parasites. Data from Brazil from a previous study [13] was added for comparison.
Fig 6.
Common origin of the p.E142del mutation in the Old and New World.
(A) Haplotype variation in all the samples bearing p.E142del from Caribbean (HR9), Niger (NE) and Tanzania (TZ) across a 102.5 kb region of chr. 6. Each row represents a chromosome, HR9 was used as reference, blank squares reflect HR9 allele state, and black squares correspond to the alternative allele. Relative bp position to p.E142del (0 bp) is shown on the x-axis. The first and last variants showed on the block correspond to the break of the haplotype block. The Caribbean sample (HR9) and a Nigerien sample (Sm.NE_Di158.1) share an identical haplotype block of 102.5 kb. (B) Minimum spanning network of 410 haplotypes of the 102.5 kb region previously identified. The network was built using 399 bi-allelic variants. Node size is proportional to sample size (smallest node: n = 1; biggest node n = 50). Nodes with samples carrying p.E142del are circled in blue. Caribbean and West African haplotypes carrying p.E142del clustered together indicating a common origin. The p.E142del from East Africa has different flanking haplotypes.
Fig 7.
Impact of starting allele frequency on OXA-R allele change.
(A) Monte Carlo simulation over 1,500 generations using strong selection (selection coefficient (s) = 0.1) on a population size (N) of 65,000 with a starting allele frequency (p(R)) of 0.15 for standing variation or 1/(2N) for new mutation. These simulations underestimate time to fixation for true de novo mutation, because we do not account for the waiting time for resistance to appear which is dependant on the rate of drug resistance mutations and Ne [3]. The starting frequency of 0.15 corresponds to the frequencies of OXA-resistance alleles observed in Kenya. (B) Change in frequency of resistant parasites (i.e. homozygotes for OXA-R alleles) from standing variation under a range of selection coefficients. The dashed lines correspond to the two thresholds (10% and 20%) at which treatment efficacy would be compromised. The numbers at the dashed lines correspond to the parasite generations needed to cross these thresholds. These predictions are deterministic, because we expect minimal stochastic variation when starting resistance allele frequencies are high.