Mutations in Plasmodium falciparum dihydrofolate reductase and dihydropteroate synthase genes in Senegal


J. P. Daily (corresponding author) and D. F. Wirth, Department of Immunology and Infectious Diseases, Harvard School of Public Health, Boston, Massachusetts 02115, USA. E-mail:
D. Ndiaye, O. Ndir and O. Gaye, Faculty of Medicine and Pharmacy, Cheikh Anta Diop University, Dakar, Senegal.
O. Sarr and S. Mboup, Laboratory of Bacteriology and Virology, Dantec Hospital, Dakar, Senegal.


Senegal recently (2004) switched to sulfadoxine-pyrimethamine (SP) with amodiaquine as first line therapy for malaria in response to increasing chloroquine resistance. In anticipation of emerging resistance to SP as a result of this change in drug pressure, we set out to define the baseline prevalence of SP-associated mutations in the dhfr and dhps genes in Plasmodium falciparum using geographically diverse and longitudinally collected samples. A total of 153 blood samples were analysed from patients (5 years or older) with mild malaria after informed consent was obtained. Longitudinal samples were collected between 2000 and 2003 in Pikine, a suburb of Dakar. Geographically diverse site sampling was carried out in 2003. The mutation prevalence in DHFR codons 51, 59 and 108 is 65%, 61% and 78% in Pikine, 2003. The overall prevalence of the triple mutation that is associated with high-level pyrimethamine resistance is 61%. The mutation prevalence rate in DHPS codons 436 and 437 is 21% and 40%, respectively. There is significant geographic variation in genotypic resistance, as samples from Pikine in 2003 had higher mutation prevalence in the pfdhfr and pfdhps genes compared to samples from Tambacounda (P < 0.015). In summary, this study demonstrates a high background prevalence of SP resistance mutations already present in P. falciparum in Senegal.


Many African countries have turned to sulfadoxine pyrimethamine (SP) as the first line antimalarial therapy in response to the emergence of chloroquine resistant parasites (White 2004). Specific point mutations in the parasite's dhfr and dhps genes have been identified that correlate with resistance to pyrimethamine and sulfadoxine, respectively. Mutations in dhfr codons 51, 59, 108 and 164 correlate with in vitro pyrimethamine resistance with each additional mutation resulting in higher resistance (Cowman et al. 1988; Peterson et al. 1988). Similarly, field studies have documented mutations within dhps at codons 436, 437, 540, 613 occurring in strains worldwide. There are numerous reports correlating these resistance markers with SP in vitro and in vivo resistance (Plowe et al. 1995; de Pecoulas et al. 1996; Basco et al. 1998; Wongsrichanalai et al. 2002; Dorsey et al. 2004). In this report, we set out to define the baseline prevalence of parasite SP associated resistance mutations in Senegal before the implementation of SP amodiaquine as first line antimalarial therapy.

Patients and methods

Study site and DNA extraction

The samples for this study were collected as part of ongoing malaria treatment studies conducted in outpatient health clinics in areas that vary in transmission intensity: Pikine (EIR = 1), 15 km from Dakar, Thies (EIR 1–20) 70 km from Dakar and Tambacounda (EIR > 100), 400 km Southeast of Dakar (Trape et al. 1992; Faye et al. 1995;Diallo et al. 1998;Thomas et al. 2002). Blood samples were collected after informed consent was obtained and patients were treated with the standard chloroquine treatment regimen. Patients greater than 5 years of age who presented to these clinics with symptoms consistent with mild malaria including fever, chills, headache and a positive thick blood smear with Plasmodium falciparum were offered enrolment into the study. Exclusion criteria included pregnancy, severe disease or history of recent treatment with anti-malarials. The Human Subjects Committee of Harvard School of Public Health in Boston and the Ethics Committee of Cheikh Anta Diop University in Dakar approved the protocol used in this study.

Analysis of mutations

DNA was extracted from whole blood using standard methods or from filter paper using DNA Minikit (Qiagen) following manufacturer's instructions (Thomas et al. 2002). Primers were used to amplify a region that included codons 16, 50, 51, 59, 108, 164 in pfdhfr and codons 436, 437, 540, 581, 613 in pfdhps [dhfr.1. (F, 5′-ATG GAA CAA GTC TGC GAC GTT TTC-3′; R, 5′-ATG ACA TGT ATC TTT GTC ATC ATT -3′); dhfr.2. (F, 5′- ATG GAA CAA GTC TGC GAC GTT TTC-3′; R, 5′-ATT GTT ACT AGT ATA TAC ATC GCT-3′); dhps.1 (F, 5′-CCATTCCTCATGTGTATA CAACAC-3′; R, 5′-CATCTG AAACATCCAATTGTGT GA-3′) and dhps.2 (F, 5′-TATGATTCTTTT TCAGAT GGAGGT-3′; R, 5′-CATCTGAAACATCCAATTGTG TGA-3′)]. 15 picomoles of each primer and 0.5 μl taq polymerase were added to a reaction volume of 50 μl. For secondary amplification, 1 μl of initial amplification product was added. PCR was carried out with 35 cycles of 95°C 30 s denaturing, 56°C 45 s annealing, and 62°C 1 min extension.

PCR amplified products were cloned into TopoTA vector and transfected into E. coli. Cells were grown overnight in terrific broth and DNA was isolated using the SNAP Gel Purification Kit (Invitrogen, Carlsbad, CA). Each clone was sequenced in both the forward and reverse direction using M13F and M13R primers. Sequence alignment was performed using the SeqManTM and MegAlignTM programs (Seqwright, Houston, TX; DNASTAR, Madison WI).

Statistical analysis

Fischer's exact test was used for two-tailed significance at P < 0.05. Pearson chi-square test was used to determine significance of mutation prevalence over time (STATA, 7.0, Stata Corporation).


We first determined the prevalence of mutations associated with SP resistance during the rainy seasons of 2000, 2002 and 2003 in pfdhfr and pfdhps in Pikine. In 2003, the most prevalent mutation is a substitution at codon 108 (pfdhfrS108N) in 14 of 18 (78%) samples followed by mutation in codon 51 (pfdhfrN51I) in 11 of 17 samples (65%) (Table 1) Finally, there is mutation prevalence in codon 59 (pfdhfrC59R) in 61% (11 of 18 samples). The prevalence rates of these mutations in samples collected in nearby Thies were not statistically different; however in comparison to distant Tambacounda, the prevalence of mutations in pfdhfr was significantly less in codons 51 (P = 0.044) and 108 (P = 0.018).

Table 1.  Percent prevalence of SP associated resistance mutations in P. falciparumdhfr and dhps in drug resistance associated codons from isolates collected in Pikine, Thies and Tambacounda in Senegal
YearPrevalence of mutations in DHFR and DHPS (%)
DHFR codonDHPS codon
  1. n is number of isolates that were successfully amplified by PCR.


In contrast, there was a lower prevalence of mutations in pfdhps and only a single sample had more than one mutation from Pikine 2003 samples. The predominant mutation was found in at codon 437 (dhpsA437G) in 6 of 15 (40%) isolates at and there was a 21% mutation rate (3 of 14 samples) at pfdhpsS436A. Although pfdhfr and pfdhps are encoded on different chromosomes, there is an association between the pfdhfr triple mutation and pfdhps mutation at 437 from all samples in 2003 (P = 0.001, Fisher exact test); however there was no association with the pfdhfr triple mutation and pfdhps mutation at 436. There is an increase in prevalence of these mutations in pfdhfr and pfdhps over the years 2000, 2002 and 2003 at Pikine, however these were not statistically different by Pearson's chi-squared statistic.

Pfdhfr haplotypes

Isolates containing a single dhfr mutation were identified at codon 108 exclusively. Pikine had a single isolate containing a double mutation, occurring at codons 51 and 108; the three isolates from Tambacounda with dual mutations occurred at codons 51 and 59. The triple mutation haplotype is found to be most prevalent in Pikine and Thies (Figure 1). This is in contrast to the results from Tambacounda where isolates containing the triple mutation were the least prevalent. The difference between the prevalence of the triple mutation in Tambacounda compared to Pikine is statistically significant (P = 0.015).

Figure 1.

Haplotype frequencies of number of mutations in codons 51, 59, 108 present in dhfr in P. falciparum isolates from Pikine, Thies and Tambacounda in 2003.


We characterized the prevalence of SP associated resistance mutations in the parasites’dhfr and dhps genes in isolates collected over time and in three distinct geographical sites prior to the initiation of SP-amodiaquine therapy for the treatment of P. falciparum in Senegal. Our data shows a high prevalence of mutations at the key positions in pfdhfr, with 61% of isolates containing a triple mutation at codons 51, 59 and 108 in Pikine. The prevalence of strains containing the pfdhfr triple mutation varied by geographic location. In addition, sulfadoxine associated mutations in pfdhps was determined. The overall lower prevalence of drug resistance mutations in pfdhps compared to pfdhfr has been reported at other sites (Basco et al. 2002a). The presence of pfdhps mutations in addition to pfdhfr mutations are thought to increase the clinical failure rate of SP in some reported field studies.

The pfdhfr triple mutation haplotype found in Senegal has been correlated with P. falciparum SP treatment failure in Malawi (Kublin et al. 2002). The high-background rate of this haplotype may exist due to the use of SP as a second line treatment for malaria and the use of antifolates for treatment of bacterial infections.

The pre-dominance of the triple mutation haplotype in Pikine and Thies is striking. Tambacounda is geographically isolated from these other two cities, which may explain the lower prevalence of this haplotype. We will use microsatellite analysis to test whether a selective sweep of an individual strain with multiple dhfr mutations may be responsible as reported in South Africa and Southeast Asia (Roper et al. 2003,2004; Nair et al. 2003).

In summary, this study demonstrates a high background prevalence of SP resistance mutations already present in P. falciparum in a background of genotypic resistance to chloroquine (Thomas et al. 2002). We have the opportunity to study the evolution of these genotypes under SP-amodiaquine drug pressure. Other sites in Africa, with high prevalence of these mutations have demonstrated good outcome using this regimen (Dorsey et al. 2002; Basco et al. 2002b; Gasasira et al. 2003; Rwagacondo et al. 2003; Talisuna et al. 2004). Correlation of the longitudinal monitoring of drug resistance markers with in vivo outcomes will provide data to guide future regimens for treatment of malaria in Senegal.


DN and OS were supported by a Fogarty International D43 Training Grant TW01503 to DFW; JPD was supported by a K23 award A1054518. JPD and DN are equal contributors to this work.