Frequent expansion of Plasmodium vivax Duffy Binding Protein in Ethiopia and its epidemiological significance

Abstract

Plasmodium vivax invasion of human erythrocytes depends on the Duffy Binding Protein (PvDBP) which interacts with the Duffy antigen. PvDBP copy number has been recently shown to vary between P. vivax isolates in Sub-Saharan Africa. However, the extent of PvDBP copy number variation, the type of PvDBP multiplications, as well as its significance across broad samples are still unclear. We determined the prevalence and type of PvDBP duplications, as well as PvDBP copy number variation among 178 Ethiopian P. vivax isolates using a PCR-based diagnostic method, a novel quantitative real-time PCR assay and whole genome sequencing. For the 145 symptomatic samples, PvDBP duplications were detected in 95 isolates, of which 81 had the Cambodian and 14 Malagasy-type PvDBP duplications. PvDBP varied from 1 to >4 copies. Isolates with multiple PvDBP copies were found to be higher in symptomatic than asymptomatic infections. For the 33 asymptomatic samples, PvDBP was detected with two copies in two of the isolates, and both were the Cambodian-type PvDBP duplication. PvDBP copy number in Duffy-negative heterozygotes was not significantly different from that in Duffy-positives, providing no support for the hypothesis that increased copy number is a specific association with Duffy-negativity, although the number of Duffy-negatives was small and further sampling is required to test this association thoroughly.

Fig 1. Ethiopian P. vivax samples contain both the Malagasy- and Cambodian-type duplications.

AF/AR, BF/BR and AF2/AR2 are control primers where all samples indicated amplifications except the negative control sample. Sample SGH(1)-357produced a band of ~650bp with BF/AR primers, indicative of a Malagasy PvDBP duplication. Sample BBH(1)-125produced a band of ~750bp with BF/AR2 primers, indicative of a Cambodian PvDBP duplication. Sample BBH(1)-132showed no amplification with both BF/AR and BF/AR2 primers, indicative of a single PvDBP region without any duplications. No bands were observed for DNA-negative control (-) in all amplifications.

Fig 2. Comparison of PvDBP copy number estimated using different approaches.

(A) A scatter plot showing PvDBP copy number estimated by quantitative real-time polymerase chain reaction (y-axis) and by PvDBP coverage in whole genome sequencing (x-axis) of 20 symptomatic P. vivax samples. Regression coefficient and P-value were indicated. (B) A dot plot showing PvDBP copy number detected by quantitative real-time PCR of all 145 symptomatic P. vivax samples by three different copy number categories. The geometric median and standard deviation of the data were indicated.

Fig 3. Comparison of PvDBP copy number in symptomatic and asymptomatic P. vivax samples.

PvDBP copy number was estimated by qPCR. The geometric median and stand deviation of the data were shown for each type of infection.

Fig 4. Comparison of PvDBP copy number among symptomatic samples with different Duffy genotypes.

These included Duffy-null homozygotes (FyB^ES/B^ES), Duffy-null heterozygotes (FyA/B^ES or FyB/B^ES), Duffy-positive homozygotes (FyA/A or FyB/B), and Duffy-positive heterozygotes (FyA/B or FyB/X2). The geometric median and stand deviation of the data were shown for each of the genotypes.

Fig 5. Scatter plot based on principle component analyses showing variation in malaria symptoms among the P. vivax patients.

PvDBP copy number of P. vivax observed in each of the patient samples was indicated by different color circles. The first three axes represented 99.5% of the total variation from 12 malaria symptoms. No clear cluster was observed among samples with single or multiple PvDBP copies, suggesting that malaria symptoms of these patients did not relate to PvDBP copy number. Malaria symptom data are presented in S3 Table.