Differences in selective pressure on dhps and dhfr drug resistant mutations in western Kenya.

Abstract

Understanding the origin and spread of mutations associated with drug resistance, especially in the context of combination therapy, will help guide strategies to halt and prevent the emergence of resistance. Unfortunately, studies have assessed these complex processes when resistance is already highly prevalent. Even further, information on the evolutionary dynamics leading to multidrug-resistant parasites is scattered and limited to areas with low or seasonal malaria transmission. This study describes the dynamics of strong selection for mutations conferring resistance against sulphadoxine-pyrimethamine (SP), a combination therapy, in western Kenya between 1992 and 1999, just before SP became first-line therapy (1999). Importantly, the study is based on longitudinal data, which allows for a comprehensive analysis that contrasts with previous cross-sectional studies carried out in other endemic regions.
This study used 236 blood samples collected between 1992 and 1999 in the Asembo Bay area of Kenya. Pyrosequencing was used to determine the alleles of dihydrofolate reductase (dhfr) and dihydropterote synthase (dhps) genes. Microsatellite alleles spanning 138 kb around dhfr and dhps, as well as, neutral markers spanning approximately 100 kb on chromosomes 2 and 3 were characterized.
By 1992, the South-Asian dhfr triple mutant was already spreading, albeit in low frequency, in this holoendemic Kenyan population, prior to the use of SP as a first-line therapy. Additionally, dhfr triple mutant alleles that originated independently from the predominant Southeast Asian lineage were present in the sample set. Likewise, dhps double mutants were already present as early as 1992. There is evidence for soft selective sweeps of two dhfr mutant alleles and the possible emergence of a selective sweep of double mutant dhps alleles between 1992 and 1997. The longitudinal structure of the dataset allowed estimation of selection pressures on various dhfr and dhps mutants relative to each other based on a theoretical model tailored to P. falciparum. The data indicate that drug selection acted differently on the resistant alleles of dhfr and dhps, as evidenced by fitness differences. Thus a combination drug therapy such as SP, by itself, does not appear to select for "multidrug"-resistant parasites in areas with high recombination rate.
The complexity of these observations emphasizes the importance of population-based studies to evaluate the effects of strong drug selection on Plasmodium falciparum populations.

Full text

Page 1

RESEARCHOpen Access
Differences in selective pressure on dhps and dhfr
drug resistant mutations in western Kenya
Andrea M McCollum1,2,3†, Kristan A Schneider4†, Sean M Griffing1,2, Zhiyong Zhou2,3, Simon Kariuki5,
Feiko Ter-Kuile6, Ya Ping Shi2, Laurence Slutsker2, Altaf A Lal2, Venkatachalam Udhayakumar2and
Ananias A Escalante7,8*
Abstract
Background: Understanding the origin and spread of mutations associated with drug resistance, especially in the
context of combination therapy, will help guide strategies to halt and prevent the emergence of resistance.
Unfortunately, studies have assessed these complex processes when resistance is already highly prevalent. Even
further, information on the evolutionary dynamics leading to multidrug-resistant parasites is scattered and limited
to areas with low or seasonal malaria transmission. This study describes the dynamics of strong selection for
mutations conferring resistance against sulphadoxine-pyrimethamine (SP), a combination therapy, in western Kenya
between 1992 and 1999, just before SP became first-line therapy (1999). Importantly, the study is based on
longitudinal data, which allows for a comprehensive analysis that contrasts with previous cross-sectional studies
carried out in other endemic regions.
Methods: This study used 236 blood samples collected between 1992 and 1999 in the Asembo Bay area of Kenya.
Pyrosequencing was used to determine the alleles of dihydrofolate reductase (dhfr) and dihydropterote synthase
(dhps) genes. Microsatellite alleles spanning 138 kb around dhfr and dhps, as well as, neutral markers spanning
approximately 100 kb on chromosomes 2 and 3 were characterized.
Results: By 1992, the South-Asian dhfr triple mutant was already spreading, albeit in low frequency, in this
holoendemic Kenyan population, prior to the use of SP as a first-line therapy. Additionally, dhfr triple mutant alleles
that originated independently from the predominant Southeast Asian lineage were present in the sample set.
Likewise, dhps double mutants were already present as early as 1992. There is evidence for soft selective sweeps of
two dhfr mutant alleles and the possible emergence of a selective sweep of double mutant dhps alleles between
1992 and 1997. The longitudinal structure of the dataset allowed estimation of selection pressures on various dhfr
and dhps mutants relative to each other based on a theoretical model tailored to P. falciparum. The data indicate
that drug selection acted differently on the resistant alleles of dhfr and dhps, as evidenced by fitness differences.
Thus a combination drug therapy such as SP, by itself, does not appear to select for “multidrug"-resistant parasites
in areas with high recombination rate.
Conclusions: The complexity of these observations emphasizes the importance of population-based studies to
evaluate the effects of strong drug selection on Plasmodium falciparum populations.
Keywords: Plasmodium, Malaria, Dihydrofolate Reductase, Dihydropterote synthase, Sulphadoxine-pyrimethamine,
Natural selection, Selective sweep, Drug resistance
* Correspondence: Ananias.Escalante@asu.edu
† Contributed equally
7School of Life Sciences, Arizona State University, Tempe, AZ, USA
Full list of author information is available at the end of the article
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
© 2012 McCollum et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.

Page 2

Background
The massive use of drugs for treating Plasmodium falci-
parum malaria has selected for mutations that confer
resistance in endemic areas worldwide, rendering tradi-
tional anti-malarial drugs ineffective in vast regions of
the globe [1,2]. Artemisinin combination therapy (ACT)
is now being used in many endemic areas; however, there
are concerns that mutations conferring resistance against
ACT could also emerge [3,4]. Understanding such com-
plex evolutionary processes, especially in the context of
combination therapies, is a matter of great interest. Valu-
able information about such dynamics can be obtained
by retrospectively investigating the rise of resistance
against sulphadoxine-pyrimethamine (SP), a combination
drug therapy that has been widely used and for which the
molecular basis of resistance is well known.
SP acts as an inhibitor of the P. falciparum folic acid
pathway, and point mutations in two genes, dihydrofolate
reductase (DHFR) and dihydropteroate synthetase
(DHPS), confer resistance to SP [5]. Point mutations at
dhfr codons 50, 51, 59, 108 and 164 act synergistically to
increase resistance to pyrimethamine. Of note, S108N
has a low level of resistance, the double mutants N51I/
S108N and C59R/S108N have moderate levels of resis-
tance, the triple mutant N51I/C59R/S108N has a higher
level, and the quadruple mutant parasite (N51I/C59R/
S108N/I164L) is considered to be resistant to the effects
of pyrimethamine [6,7]. Similarly, mutations at dhps
codons 436, 437, 540, 581 and 613 act synergistically to
increase the level of resistance to sulphadoxine. Simply,
the mutations S436A and A437G alone confer a low level
of resistance, and when in combination with K540E and/
or A581G and/or A613S/T the parasite has an increased
level of resistance to sulphadoxine [1,8].
The evolution of drug resistance is further complicated
by the fact that resistant alleles may have multiple origins
intertwined with migration patterns among P. falciparum
populations; such complex dynamics are still poorly
understood. There is compelling evidence indicating a
common origin for highly resistant pyrimethamine alleles
across Southeast Asia and at a few sites in Africa [9-14];
however, additional, novel low frequency lineages for the
triple mutant (51I/59R/108 N) dhfr allele have been docu-
mented in Cameroon and also in western Kenyan [14,15].
Similarly, recent studies from sites across Africa and Asia
show multiple independent origins of mutations at dhps
[16-18]. However, the patterns for dhps highlight different
evolutionary processes than those for dhfr. Thus, SP-
induced selection on resistance-associated mutations may
differ for the two genes and across different endemic
regions. Hence, reliable estimates of selective parameters
for various dhfr and dhps mutations are highly desirable.
A few studies have addressed the genetic conse-
quences of SP drug selection, yet the temporal dynamics
of mutations are rarely investigated in both loci. Indeed,
patterns consistent with selective sweeps of highly resis-
tant dhfr alleles have been reported in multiple popula-
tions [9,19,20], but there are only a few studies on dhps
[14,17,18,20]. Despite the limited evidence, dhps shows a
clear pattern of reduced diversity in multiple popula-
tions, indicating an increase in mutant alleles conferring
resistance to sulphadoxine. Notably, the patterns of the
selective sweeps in dhps and dhfr appear to be different,
providing evidence that the strength of selection is not
the same on both loci [14,20]. However, all these studies
are based on cross-sectional data and measures of the
strength of drug selection are limited. Attempts to infer
the strength of selection have been made for dhfr [9,21]
but such estimates focused only on the proportion of
clinical failures, an indirect line of evidence that does
not consider the actual frequency of resistant mutations
and may lead to inaccurate predictions. A direct com-
parison of the selective strengths on dhfr and dhps dur-
ing the early stages of the onset of clinical resistance is
still missing. Indeed, estimates of drug selection have
not been obtained from molecular data. Moreover, pat-
tern of selective sweeps studied so far just indicate drug
selection but the importance of linking estimates of
selection parameters with the pattern of the sweep have
been neglected.
Here, a population-based characterization and analysis
of genetic signatures around dhfr and dhps from samples
collected in western Kenya from 1992-1999 was con-
ducted. At the time these samples were collected, SP had
been exerting selective pressure on P. falciparum popula-
tions since the 1980s. SP was introduced in Kenya as a
second-line treatment for uncomplicated malaria in 1983
and as a first-line treatment in 1999 [22,23]. However,
clinical SP resistance was noted as early as 1982 [23].
Thus, this study captures some of the early events in the
dynamics of drug-resistant mutations in the local P. falci-
parum population. Even before SP was chosen as a first-
line treatment, all alleles in the population had dhfr
mutations associated with pyrimethamine resistance. In
contrast, sulphadoxine-sensitive alleles at dhps were still
present while resistant double-mutant alleles were
increasing in frequency. The longitudinal data, allowed
inferences of the selective strengths on various mutations
at dhfr and dhps based on a theoretical model tailored to
P. falciparum. Overall, these investigations highlight the
differences in selective pressures on these two loci, when
the drugs were part of a combination drug therapy.
Methods
Study subjects
Two hundred thirty-six blood samples collected from
the Asembo Bay Cohort Project, from the years 1992-
1999 [24], were analysed. This study was approved by
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
Page 2 of 14

Page 3

the ethical committee of (Institutional Review Board)
CDC and the Kenya National Ethics Review Committee.
The participants provided written informed consent. In
short, this was a longitudinal study conducted between
1992 and 1999 in western Kenya, a holo-endemic area
of intense transmission estimated at approximately 300
infective bites per person per year [25]. Blood samples
were taken from mother-infant pairs and other siblings
less than five years old once per month until the chil-
dren turned five years old. Malaria parasitaemia was
treated with SP.
DNA isolation and genotyping methods
DNA was isolated from whole blood using the
QIAamp®DNA Mini Kit (Qiagen, Valencia, CA, USA).
All samples were genotyped for P. falciparum mutations
at dhfr codons 50, 51, 59, 108, and 164 and dhps codons
436, 437, 540, 581, and 613 by pyrosequencing as pre-
viously described [20,26].
Microsatellite characterization
Microsatellite characterization was conducted on all sam-
ples. Samples were assayed for 18 microsatellite loci that
span 138 kb on chromosome 4 around dhfr [9-11], 18
loci that span 138 kb on chromosome 8 around dhps
[19], five loci on chromosome 2 that span 101 kb, and
four loci on chromosome 3 that span 94 kb [20]. The
microsatellites used around dhfr are at -89, -58, -30, -17,
-10, -7.5, -5.3, -4.5, -4.4, -3.8, -1.2, -0.3, 0.2, 0.52, 1.48,
4.05, 5.87, 30.3, and 50 kb; where negative numbers refer
to positions 5’ to the gene and positive numbers refer to
positions 3’ to the gene. The microsatellites used around
dhps are at -72.7, -34.5, -18.7, -11, -7.4, -2.8, -1.5, -0.132,
0.034, 0.5, 1.4, 6.4, 9, 16.3, 22.8, 36, 49.5, and 66.1 kb.
The loci around dhps have been previously published
[19,20]; however, it was recently brought to the authors’
attention that the orientation of the microsatellite loci
along chromosome 8 around dhps was incorrect in [20]:
loci that have been reported previously as 5’ to dhps are
actually 3’ and vice versa. To avoid any confusion, the
corrections along with previously published positions and
primers are in Additional file 1: Table S1. The correct
positions of dhps loci have been used throughout this
manuscript.
The microsatellites used on chromosome 2 are at 302,
313, 319, 380, and 403 kb. The microsatellites used on
chromosome 3 are at 335, 363, 383, and 429 kb. The PCR
primers for 403 kb chromosome 2 are 5’-AAATA-
TAAATCTTCTTCTTCTTTTTT-3’ (forward) and 5’-
TAGAGAAATAAATATATCCAT-3’ (reverse); and for
363 kb chromosome 3 are 5’-CAAAAATGAAAAAT-
GAAAAGG-3’ (forward) and 5’-TAAAGGGTGCGCA-
TATCAAT-3’ (reverse). All remaining microsatellite PCR
primers are detailed in [20]. Single reaction PCR and ther-
mal cycling conditions are detailed in [9]; and nested PCR
reactions and thermal cycling conditions are detailed in
[10]. PCR products were separated on Applied Biosystems
3100 capillary sequencer and scored using GeneMapper®
software v3.7 (Applied Biosystems, Foster City, CA, USA).
Genetic variation per locus and allele
The genetic variation for each microsatellite locus was
measured by calculating the expected heterozygosity
(He) and number of alleles per locus (L). Hewas calcu-
lated for each locus as He= n/(n − 1)
where n is the number of isolates sampled and
pi is the frequency of the ith allele (i = 1,...,L).
The sampling variance for He was calculated as
?
was calculated using all alleles that occurred in the
respective group including those in isolates that carried
more than one microsatellite allele.
He, was also calculated for microsatellite loci associated
with specific dhfr and dhps mutant alleles. For dhfr
alleles, only samples with single ‘clone’ infections of the
respective mutant allele were used. This guarantees that
the microsatellite variation is linked to the respective
allele. The pattern of variation present, before the occur-
rence of a beneficial mutation, should be reflected by He
among wildtype alleles; however, since sensitive wildtype
alleles were only present at marginal frequencies an esti-
mate of Hecould not be calculated. As a proxy to esti-
mate the initial variation, Hewas calculated among non-
triple mutant dhfr alleles. For this estimation, all mixed
infections that did not contain the 51I/59R/108 N triple
mutant (e.g. an isolate with mixed codon 51I/S108N was
included, but an isolate with mixed codon 51I/59R/
S108N was excluded) were included. At microsatellite
loci around dhps, Hewas calculated separately among
isolates that contained single infections with the 437 G/
540E mutant allele, and isolates that contained single
infections with the sensitive (wildtype) alleles.
?
1 −
?
p2
i
?
,
2(n − 1)/n3
2(n − 2)
??
p3
i−
??
p2
i
?2??
[19,21]. He
Haplotype characterization
Approximately 70% of the samples used in this study were
‘multiple infections’, i.e. multiple parasite lineages or gen-
omes were present in an infection. Based on dhfr and dhps
genotyping alone, 63.0% and 72.2% were multiple infec-
tions, respectively. The neutral microsatellite markers on
chromosomes 2 and 3 collectively showed that 70.0% of
the samples contained multiple infections, and the micro-
satellites around dhfr and dhps showed 73.9% and 64.5%
multiple infections, respectively. A goal with this study is
to present a population-based perspective and analysis of
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
Page 3 of 14

Page 4

the data; thus, data from multiple infections for appropri-
ate analyses was retained. Multiple infections are inap-
propriate for all of the analyses; and it is stated when data
from multiple infections were excluded.
Microsatellite haplotypes are defined as a collection of
sites close to the genes dhfr and dhps that had low varia-
tion and were more likely to be in linkage disequilibrium.
Thus, 11 microsatellite loci spanning 11.5 kb around dhfr
and nine loci spanning 20 kb around dhps were used to
characterize haplotypes relative to dhfr and dhps alleles.
Haplotypes were classified as different if they contained >
1 different allele across loci. Only samples without mixed
infections detected by pyrosequencing were used for hap-
lotype characterization.
Haplotype analysis
eBURST groups haplotypes, based on a simple evolution
model, which assumes that one lineage or founding hap-
lotype reaches high frequency in the population and then
starts to differentiate, producing closely related haplo-
types; this is depicted as a cluster [27]. Data from the 11
microsatellite loci spanning 11.5 kb around dhfr and nine
loci spanning 20 kb around dhps (as for haplotype char-
acterization) were used to depict genetic relationships in
eBURST. Only samples in which multiple infections were
not detected by pyrosequencing of dhfr or dhps were
used for the eBURST analysis. Since eBURST does not
allow for missing data, samples with incomplete haplo-
types were removed; therefore, there were fewer samples
utilized for the eBURST analysis than for haplotype char-
acterization. If multiple alleles were detected at a single
microsatellite locus in a sample, the most frequent allele
was used, i.e. the one that was present at the highest peak
in the electropherogram.
Genetic differentiation between alleles was measured
using Wright’s F-statistics [28]. The statistic FSTmea-
sures genetic differentiation between populations but,
here, FSTwas used as a statistic to compare groups of
alleles. For dhfr the microsatellite loci from -10 kb to
1.47 kb and for dhps the loci from -2.5 kb to 17.5 kb
were used for the FSTanalysis. FSTcalculations were
computed using Arlequin ver 3.01 [29]. The Excel
Microsatellite Toolkit was used to format data for Arle-
quin [30].
Linkage disequilibrium (LD) between loci along the
chromosomes and also between dhfr and dhps point
mutations was assessed by using an exact test of LD
[31]. Samples with multiple alleles at any locus were
removed from the analysis; this was done for dhfr, dhps,
and the neutral markers independently. Similarly, sam-
ples where multiple infections were detected at any site
were removed from the LD analysis, testing pairs of
point mutations in dhfr and dhps; this was done inde-
pendently for dhfr and dhps for a given sample. Only
loci or sites that showed polymorphism among the used
samples were used for the analysis. Associations were
tested between pairs of loci or sites by using 10,000
Monte Carlo steps in Arlequin version 3.01 [29]. To
correct for multiple testing the Bonferroni-Holm correc-
tion was used.
Measuring the strength of selection
The strength of selection on dhfr and dhps was estimated
from the changes in frequency over time of the various
mutant alleles at each gene. The strength of selection, s,
of allele A compared with allele B, was defined as 1 + s
being the average relative reproductive advantage of A
over B [32]. Hence, if ptand pt+Tare the relative frequen-
cies of A at times t and t + T, log 1 + s = 1/T (log pt+T/
(1-pt+T)- log pt/(1-pt)) [32].
Measurements for the frequency of the advantageous
allele A were made at time t, pt, at different equally
spaced time points (tk= k*180 days (k = 0, 1, 2...))
within the six years covered by the samples. The fre-
quency at tkwas calculated from all samples that were
taken between time tkand tk+ 360 days. Hence, the
intervals [tk, tk+ 360] overlapped (sliding window). The
strength of selection was obtained by performing a lin-
ear regression of the explanatory variable log pt/(1-pt),
where only those time points tkas regressors for which
at least three triple and three non-triple mutations
occurred were included. The actual strength of selection
per generation is derived from the slope of the linear
regression divided by the number of malaria generations
per year, Ngen, which was assumed to be Ngen= 17.3
(i.e., one transmission cycle every three weeks, corre-
sponding to infections throughout the whole year).
More precisely, if s is the strength of selection, and a
and b are the constant and linear regression coefficients
respectively, log pt/(1-pt) = tNgenlog(1 + s)- log p0/(1-p0)
= a+b t. Hence, s = exp(b/Ngen)-1.
Two double mutant dhfr alleles were present in the
Kenyan population, and both confer a level of pyri-
methamine resistance. It is not clear a priori whether
selection for both double mutant alleles is equally
strong; therefore, the strength of selection for 51I/108 N
allele with that for 59R/108 N was compared. The
strength of selection of the triple mutant allele (51I/
59R/108 N) was measured over the 51I/108 N and the
59R/108 N double mutants, separately. For these mea-
surements, only samples with single infections at these
alleles (as detected by pyrosequencing) were included.
The purpose of estimating these three strengths of
selection at dhfr is as follows. If s1, s2, and s3denote the
strength of selection of 51I/59R/108 N over 51I/108 N,
51I/108 N over 59R/108 N, and 51I/59R/108 N over
59R/108 N, then the standard haploid selection model
yields 1 + s3= (1 + s2)*(1 + s1).
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
Page 4 of 14

Page 5

Dhps single mutants were at relatively low frequency
and only the 437 G/540E double mutant was found in
single infections. Thus, for dhps, the strength of selec-
tion of double mutant alleles (jointly) was measured
over all other alleles. To derive the frequencies included
were all samples with 437 G/540E single infections, and
all samples that did not contain the mixed codon
A437G/K540E. More precisely, excluded were only
those samples for which it was unclear whether they
contained the 437 G/540E mutant.
The reduction of Heflanking dhfr and dhps was utilized
to evaluate whether the estimates for the strengths of
selection were meaningful. For this purpose, Hewas com-
pared with the analytical prediction Hepred, given by
Hepred= H0(1-p02r(1-F)/s). Here, H0is the initial expected
heterozygosity, r denotes the recombination rate, F the
inbreeding adjustment (F = 1 corresponds to complete
inbreeding, and F = 0 to random mating), and p0is the
initial frequency of the 51I/59R/108 N or 51I/108 N
allele, or of the 437 G/540E allele. As in [9-14]r =
5.88*10-4Morgans/kb and p0= 10-4were used. Also, F =
0.4, which corresponds to 60-70% mixed clone infections
was used. For Hepredamong dhps 437 G/540E alleles, He
among wildtype alleles was used as an estimate for H0,
since it should not be affected by the sweep. For Hepred
among 51I/59R/108 N alleles, Heamong double, single
mutant and wildtype alleles was used as an estimate for
H0. For Hepredamong 51I/108 N double mutants, He
among 59R/108 N double, single mutant and wildtype
alleles was used as an estimate for H0.
Results
Genotyping results
The frequency of dhfr and dhps alleles in the sample set
was calculated with two analyses: a) including only sin-
gle infections as determined by pyrosequencing of the
mutations in dhfr or dhps, and b) including both the
single and multiple infections that exhibited only one
codon with two amino acids. For the latter, the “multi-
ple” allele codon was used to break down the allele into
two separate alleles (e.g. N51/59R/S108N was analysed
as N51/59R/S108 and N51/59R/108 N). The frequency
of mutant alleles is strikingly similar for the two ana-
lyses (Figure 1), but for consistency the frequency of the
“single infection” sample set will be emphasized here.
There were very few sensitive wildtype dhfr alleles in
the population (3%), and the majority of the sample set
is composed of double (50% 51I/108 N, 27% 59R/108
N) or triple mutant (20%) dhfr alleles. The majority of
dhps alleles were sensitive wildtype (34%) and 437
G/540E mutants (57%). For both dhfr and dhps, the
majority of the mutant alleles were double or triple
mutant alleles.
Haplotype analysis
In an attempt to better understand the evolutionary his-
tory of the alleles in the population, microsatellite hap-
lotypes were characterized for the microsatellite loci
immediately surrounding dhfr and dhps (Additional file
2: Figure S1 and Additional file 3: Figure S2, respec-
tively). There were 25 haplotypes for the dhfr double
mutant 59R/108 N (n = 40) and 23 haplotypes for the
double mutant 51I/108 N (n = 72). The 59R/108 N
allele is not dominated by any particular haplotype;
however, the 51I/108 N allele is dominated by one hap-
lotype at high frequency (haplotype 26, 64%).
Six haplotypes for the dhfr triple mutant (51I/59R/108
N) allele (n = 29) were observed. Two haplotypes (15 and
17) were seen in three of the triple mutant allele samples
that are also represented in allele 59R/108 N. The remain-
ing four haplotypes (48 - 51) are not present for either of
the double mutant alleles; they are unique to the triple
mutant allele. Haplotype 48 is the most predominant hap-
lotype for the triple mutant allele (79%) and is identical or
closely related to the previously characterized Southeast
Asian triple mutant dhfr haplotype [9,11]. The triple
mutant dhfr alleles in this population also have three addi-
tional unique haplotypes, 49, 50, and 51; each present in
one sample.
Haplotype analysis for dhps revealed 54 haplotypes for
the wildtype allele (n = 57) and 13 haplotypes for the
double mutant allele 437 G/540E (n = 95). There were
no predominant haplotypes for wildtype alleles. For the
samples containing the 437 G/540E allele, however, hap-
lotype 55 was present at a high frequency (86%).
Genetic differentiation and relationships among
haplotypes and alleles
The FSTvalues for the comparisons between the three
multiple mutant alleles of dhfr (51I/108 N, 59R/108 N,
51I/59R/108 N) were high and significant (p < 0.01).
There was also a significant value for the comparison
between the dhps wildtype and 437 G/540E double
mutant alleles (p < 0.01).
The application eBURST was used to discern relation-
ships among the dhfr and dhps microsatellite haplotypes.
The analysis for dhfr was conducted by combining the
data from this study with dhfr haplotypes previously
reported from the same region of western Kenya in 38
samples collected in 2002-2004, almost 10 years after
the samples used by this study were collected [15]. The
majority of the dhfr double mutant haplotypes from
2002-2004 were present in the earlier set of samples
(Additional file 4: Figure S3). A large 51I/59R/108 N
cluster is comprised entirely of haplotype 48, which was
the most prevalent haplotype found for the triple
mutant dhfr allele, with the addition of rare independent
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
Page 5 of 14

Page 6

Figure 1 Frequency of dhfr (A) and dhps (B) alleles in the sample set. Bars represent the frequency in “single infections” as detected by
genotyping (dhfr n = 145 and dhps n = 166), and the frequency of the alleles in “mixed infections” - single infections plus multiple infection
where only one codon was a mixture of wildtype and mutant codons (dhfr n = 205 and dhps n = 217).
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
Page 6 of 14

Page 7

haplotypes from 2002-2004 [11]. The minor frequency
triple mutant allele haplotypes 17, 49, 50, and 51 are
not closely related to the major frequency haplotype
(haplotype 48).
An analysis of dhps haplotypes exhibits a single cluster
comprised entirely of haplotypes from the double
mutant alleles (Additional file 5: Figure S4). Specifically,
all points are derived from haplotype 55, which is the
haplotype in highest frequency from the 437 G/540E
sample set (Additional file 3: Figure S2).
Estimates of selection
Using linear regression, the 51I/59R/108 N allele has a
selective advantage over both double mutant alleles at
dhfr: s1= 0.013 (± 0.004) over the 51I/108 N allele
(Figure 2A), and s2= 0.036 (± 0.007) over the 59R/108
N allele (Figure 2B). Note that s2is a lower bound for
the selective advantage of the 51I/59R/108 N allele over
the (extinct) wildtype. Furthermore, the 51I/108 N allele
has a selective advantage of s3= 0.021 (± 0.005) over
59R/108 N allele (Figure 2B). These estimates are con-
sistent since 1 + s3≈(1 + s2)*(1 + s1). At dhps, the 437
G/540E allele has a selective advantage s = 0.009 (±
0.002) over wildtype alleles (Figure 2A).
Variation around genes under selection
The pattern of genetic variation linked to dhfr and dhps
was examined. Number of alleles per locus (A) and het-
erozygosity (He) was calculated as a measure of variation
at each microsatellite locus (Additional file 6: Table S2).
The number of alleles found at each microsatellite locus
around dhfr and dhps alleles is shown in Additional file 7:
Figure S5. There is a stronger reduction in Hesurrounding
dhfr 51I/59R/108 N and 51I/108 N alleles, than the 59R/
108 N allele (Figure 3A). A significant reduction in Heis
observed around dhps 437 G/540E mutant allele com-
pared to the wildtype allele (Figure 3B).
The observed Heamong dhfr 51I/59R/108 N alleles and
dhps 437 G/540E alleles was compared to the Hepre-
dicted by a standard selective sweep model [33], using
the estimates for selection coefficients. Similar models
have been used elsewhere [9,21]; however, these assume
that the beneficial allele has reached a frequency of nearly
100%. As an estimate for the strength of selection, s =
0.036 was used, which is a lower estimate for selective
advantage of the triple mutant over the wildtype. The
initial Hefor dhfr could not be estimated because sensi-
tive wildtype alleles were present only at marginal fre-
quencies. Therefore, the initial Hewas set to Heamong
infections without the triple mutant (including infections
with the mixed codons N51I/S108N or C59R/S108N.
The prediction underestimates Hefor parasites carrying
the 51I/59R/108 N triple mutant at dhfr (Additional file
8: Figure S6). This can be due to the assumption of 17.3
transmissions per year, which leads likely to underesti-
mates of selection.
At dhps the pattern of observed Heis in general agree-
ment with the predicted heterozygosity (Additional file 6:
Table S2). He3’ to the gene is much higher than the
actual observation; however, this is typical for selective
sweeps that the valley of reduced heterozygosity is much
more pronounced on one side of the target of selection
[34].
The pattern of linkage disequilibrium (LD) in the chro-
mosomal region surrounding dhfr and dhps is another
measure of strong selection in the population (Additional
file 9: Figure S7). In addition, LD between the codons
involved in drug resistance of dhfr and dhps was exam-
ined. There was significant LD between codons 437 and
540 of dhps (Additional file 9: Figure S7).
Discussion
Allele populations: genetic differentiation and divergence
Western Kenya is an area of intense transmission with an
entomological inoculation rate (EIR) of approximately
300 infective bites per person per year during the early
1990s, before the introduction of bed nets [25]. There-
fore, there will be many multiple ‘strain’ infections (as
many as 70% of the samples collected in this study as
measured by the neutral microsatellite markers) and, as a
consequence, a great amount of meiotic recombination.
The large amount of recombination along with histori-
cally older parasite populations in Africa contributes to a
large amount of variation in term of the number of differ-
ent genetic lineages circulating in the population. Conse-
quently, reconstructing a pattern of descent is difficult to
do with absolute precision. Nevertheless, the data sup-
port an overall pattern of genetic differentiation between
the major mutant alleles of dhfr and the dhps double
mutant allele from any wildtype alleles in the population.
Genetic differentiation between dhfr alleles was main-
tained over time. The data suggest that the 51I/108 N
and 51I/59R/108 N alleles have remained distinct in the
population. This can be explained by stronger selection
for the 51I/59R/108 N, and the fact that the majority of
these alleles were imported from Southeast Asia along a
haplotype, which was not present in Kenya. Similarly,
genetic differentiation between dhfr alleles in a Cameroo-
nian population was previously noted [14].
The most predominant dhfr triple mutant haplotype in
both sample sets (Kenya and Cameroon) was one pre-
viously described in Southeast Asia and has since been
documented in multiple sites across Africa, including
eastern Kenya [9,10,12,13,15]. This Southeast Asian hap-
lotype appears to have diverged over time in the western
Kenyan population; this result is expected for an allele
that has been in a population for many years. The highly
resistant Southeast Asian lineage is the most prevalent
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
Page 7 of 14

Page 8

triple mutant dhfr lineage in both the 2002-2004 and
1992-1999 sample sets, which is consistent with studies
suggesting that the Southeast Asian haplotype is the only
one present in African populations [11-13]. Furthermore,
a study from eastern Kenya showed only the Southeast
Asian lineage for all triple mutant alleles, and this lineage
was present as early as 1988 [13]. Here, this lineage is
documented in western Kenya as early as 1992. The
Figure 2 Estimates for the strength of selection. (A) 437 G/540E vs wildtype at dhps and 51I/59R/108 N vs 51I/108 N at dhfr. (B) 51I/108 N vs
59R/108 N and 51I/59R/108 N vs 59R/108 N at dhfr.
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
Page 8 of 14

Page 9

mechanisms that allowed the minor and major variant
triple mutant alleles to emerge and be maintained over
time cannot be completely characterized at this time
since there is a paucity of serial samples from the time
period prior to the widespread use of SP. Theoretically, it
is possible that the triple mutant was created by mutation
Figure 3 A) Observed Heat ms loci around dhfr alleles 51I/59R/108 N, 51I/108 N, and 59R/108 N. B). Observed Hearound dhps wildtype
and 437 G/540E alleles. Here, Heis calculated from single infections only. Loci are named according to their positions relative to dhfr or dhps (kb
from the gene). The sampling variance is indicated by error bars. Dashed horizontal line indicates the average Hefrom chromosomes 2 and 3.
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
Page 9 of 14

Page 10

from a rare haplotype, which is similar to the Asian hap-
lotype but too infrequent to be detected in the sample,
and spread to high frequency; however, this alternative is
highly unlikely.
The double mutant dhfr and dhps alleles are also illu-
minating. There is a larger number of haplotypes for
these alleles than has been documented at other sites in
Africa [10,13]. There are, however, fewer haplotypes for
the triple mutant dhfr allele than for the double mutants
and fewer haplotypes for the mutant dhps allele than the
wildtype allele: results which are consistent with the
hypotheses that selection is stronger on the alleles with
fewer haplotypes or that these alleles are newer in the
population [14,19]. Moreover, more haplotypes were
found for the 59R/108 N allele than the 51I/108 N
allele, and this is consistent with results that selection
on the latter allele is stronger. Namely, for the 51I/108
N mutant (i) recombination breaks down the initial
association with the ancestral haplotype less efficiently,
and (ii) the time window in which recurrent mutations
can occur and rise to detectible frequency is narrower.
Hitchhiking and linkage disequilibrium
The shape of the curve of variation around genes under
strong selection is affected by a number of factors includ-
ing the strength of selection, time since the initial selec-
tive event, and the amount of recombination [32,35].
This asymmetry is consistent with hitchhiking models
even when the rates of recombination and mutation are
constant [34]. The asymmetry seen in the lack of varia-
tion surrounding dhfr and dhps has been documented
previously for dhfr [19,21]. The peaks in Hearound the
dhfr triple mutant compared with the 51I/108 N double
mutant at some microsatellite loci are consistent with the
evidence that the triple mutant was imported from
Southeast Asia. Namely, imported microsatellite alleles
(which were not existing in Kenya) could have hitchhiked
with the triple mutant, resulting in relatively large He.
If resistant alleles confer a selective advantage, one can
hypothesize that the genetic variation around these alleles
would be substantially reduced as a result of stronger
selection compared to the wildtype or less resistant/toler-
ant alleles in the population. This has been demonstrated
in African P. falciparum populations previously for dhfr
[19] and dhps [17]. The population of dhfr and dhps alleles
in western Kenya shows dramatic levels of LD around
both genes compared to neutral markers. This is expected
under conditions of very strong natural selection, such as
those that might occur in environments that promote the
rise of drug resistance. Theoretical evidence supports the
hypothesis that LD will decay rapidly after a hitchhiking
event [36,37]. Here, there is LD for approximately 20 kb
around both dhfr and dhps. A more advanced theoretical
model is needed to evaluate the patterns seen here in wes-
tern Kenya.
The data here do not demonstrate significant LD
between mutant dhfr and dhps alleles - another line of
evidence suggesting that selection has not occurred on
any two mutant alleles at a given point in time in the
population; i.e. selection on dhfr alleles is likely indepen-
dent of that on dhps alleles. This finding suggests that, in
a holoendemic area with high recombination, a multi-
drug-resistant haplotype is less likely to be maintained in
the absence of drug pressure, as has been the case in
areas with low transmission [20]. Additionally, since SP
was a second-line treatment, joined selection pressure on
dhfr and dhps must have been weaker (in the overall
population) than as if it was the first-line treatment.
Again, this increases the chance of meiotic recombina-
tion between dhfr and dhps.
Selective sweeps
Sulphadoxine-pyrimethamine resistance in the Asembo
Bay area was reported in the literature as early as 1982;
therefore, the high proportion of mutant dhfr and dhps
alleles in the population is not surprising [23]. Since the
data used in this study provides longitudinal information,
it was possible to directly estimate the selective para-
meters from the frequency changes in the various muta-
tions. Notably, a theoretical model [38,39] tailored to
P. falciparum justifies this approach. Thus, the results
presented here are the first estimates of selective para-
meters from molecular data in Africa and the first for
dhps. The results for dhps, in contrast to dhfr, revealed a
greater proportion of wildtype alleles in the population.
Similar data have been noted in previous studies, and
Nzila et al. [40] have suggested that after the triple
mutant dhfr alleles spread sufficiently through a popula-
tion, dhps mutant alleles increase in frequency due to
selection by sulphadoxine [14,40,41]. While dhfr double
mutants were predominantly present, triple mutants
increased in frequency from 1992-1999, in accordance
with positive selection for those mutants. Sulphadoxine-
sensitive wildtype dhps alleles were still present at appre-
ciable frequency, while resistant mutant alleles increased
in frequency during this observed time period.
Further evidence for strong selection for the double
mutant allele is the loss of variation around dhps 437 G/
540E allele as well as high levels of LD around dhps. It
appears as though this study has captured the dhps 437 G/
540E alleles in the middle of a selective sweep, i.e. while its
frequency increases in the population. The data suggest
that this population in western Kenya possibly experienced
dramatically strong selection events leading to a rapid
increase in frequency of resistant dhfr alleles, but such
events took place before drug selection allowed for the rise
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
Page 10 of 14

Page 11

in frequency of dhps resistant alleles. Indeed, this hypoth-
esis is consistent with previous studies [1,17,42]. It would
be interesting to assay this population again based on
more recent samples to observe how shape and depth of
the heterozygosity as well as the frequency spectrum of
resistant mutations have changed as a result of a change
in drug policy. In particular, by employing the regression
approach presented here, it would be possible to estimate
the amount of metabolic costs associated with drug-resis-
tance of various dhfr and dhps mutations in natural
settings.
Soft selective sweeps, as opposed to the traditional
“hard” selective sweep proposed by Maynard Smith and
Haigh [32], are a case where multiple alleles are favoured
by selection and, consequently, multiple genetic back-
grounds hitchhike with the alleles under selection
[43,44]. The result will be an increase in the amount of
variation surrounding the selected allele as compared to
a hard selective sweep. The presence of different predo-
minant haplotypes for dhfr alleles together with reduced
variation along the chromosome for each of these is evi-
dence of selection on each of these alleles - a soft selec-
tive sweep.
Typically, the classical “hard” sweep of the Southeast
Asian triple mutant allele has been suggested for dhfr
alleles. These arguments, however, do not describe the
dynamics of drug resistant mutations for Plasmodium,
especially in Africa. This study has led to the hypothesis
that soft sweeps involving drug resistant alleles should be
more common in Africa since drug pressure is effectively
lower. Indeed, there is a higher proportion of asympto-
matic (untreated) infections due to higher levels of
acquired immunity. It is worth noting that the data from
[10] is suggestive of a soft sweep in South African and
Tanzanian populations. Nevertheless, understanding the
factors leading to soft sweeps for pyrimethamine resis-
tance, and any other form of drug resistance that involves
such a complex pattern of alleles, requires more studies.
Multiple origins and soft sweeps are expected to be
common if mutation rates are high or population sizes are
large; specifically if the population-based mutation para-
meter 2Neμ > 0.01 [43]. The mutation rate for pyrimetha-
mine resistance has been estimated to be 10-9[45] and the
estimated population size for African P. falciparum popu-
lations based on mtDNA sequence variation is about 105
[46,47]. Given these estimates (2Neμ = 2 × 10-4), a first
approximation is that soft sweeps are expected to be rare
[43]. Note that estimates for Nemay have limited mean-
ing, since these estimates reflect the fact that some charac-
teristic is equivalent to a Wright-Fisher model with
sample size Ne. The complexity of the P. falciparum trans-
mission cycle implies that there are processes that exceed
the simplicity of standard population genetic assumptions
[38]. However, the population-genetic analyses performed
in this article are justified by the theoretical results of
Schneider and Kim [38,39]. Summarizing, soft sweeps
might be more common than naively expected.
Strength of selection
In [9], a similar approach as here was used to estimate
the selective strength at dhfr in a Southeast Asian popu-
lation; however, there are crucial differences. The
strength of selection was estimated from genetic data,
whereas the prior study used historic data from clinical
treatment failures and clinical success. The latter is pro-
blematic, because clinical failures do not directly corre-
late with the presence of resistant mutations and there
could be multiple variables affecting a patient’s treatment
outcome. Moreover, the data are acquired from sympto-
matic infections (asymptomatic infections do not need
treatment), which implies an overestimation of the
strength of selection since it masks the disadvantage of
resistance due to metabolic costs that are only apparent
in untreated infections. Notably, the use of longitudinal
data allowed for a comprehensive analysis on the
strength of selection acting on specific mutants rather
than an overall average of selection. Nair et al. [9]
assumed six transmission cycles per year and an inbreed-
ing adjustment factor of 80%; parameters that properly
describe an area with lower transmission and lower fre-
quency of multiple infections but do not properly
describe holo-endemic areas such as western Kenya. In
this case, 17.3 transmission cycles per year were used,
which, assuming an incubation period of three weeks
corresponds to infections over the entire course of a year.
The choice of 21 days was based on the extremely high
transmission rates, the fact that a mosquito becomes
infective approximately 10 days after the blood meal, and
that it will take seven to 11 days until gametocytaemia
peaks in infected patients. Also the inbreeding was
adjusted to 40%, which is in agreement with the percen-
tage of observed multiple infections (63%-70%). Whereas
the absolute values of s depend on these assumptions,
the relative pattern observed does not. Nevertheless, it is
straightforward to re-calculate s assuming different num-
bers of transmission cycles per year (see methods). It is
worth nothing that there is not such a thing as a univer-
sal “s“; thus, conclusions should be made based on gen-
eral patterns and relative differences at a local level.
For the purpose of estimating the selective strengths,
sample sizes were rather small due to multiple infections
and the inability to discern alleles; however, an approach
was pursued that included as many isolates as possible.
Regardless of its limitations, the results clearly indicate
that drug selection due to a combination drug therapy
operates differently at each individual locus, and for indi-
vidual alleles per locus. Overall, the study provides direct
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
Page 11 of 14

Page 12

temporal evidence of differential selection acting on dhfr
and dhps mutations associated with resistance in Africa.
Conclusions
The three signatures of a selective sweep in a population:
altered distribution of polymorphic sites along the chro-
mosome, altered allele frequency spectrum, and an
increase in the amount of linkage disequilibrium, are all
seen in dhfr and dhps allele populations in western Kenya.
The independent origination, genetic differentiation, and
maintenance of alleles allude to the fact that rapid,
dynamic events in the clinical and ecological settings have
given rise to the patterns of resistant mutations we see
today. Regardless of the fact that SP is a combination drug
therapy, the strength of selection on the two loci is differ-
ent and the drug by itself does not appear to select for
“multidrug"-resistant parasites in areas with high recombi-
nation rate. The various estimates for the selective
strengths on various mutant alleles, allow for a more com-
plete understanding of the evolutionary dynamics asso-
ciated with drug-resistance. Thus, the local demographic
history (effective population size and recombination rate)
needs to be taken into account when investigating the rise
of multi-resistant genotypes in Plasmodium populations.
Additional material
Additional file 1: Table S1. PCR primers used for dhps microsatellite
amplification.
Additional file 2: Figure S1. Haplotype frequencies for dhfr alleles: A)
59R/108N (n = 40), B) 51I/108N (n = 72), and C) 51I/59R/108N (n = 26).
Haplotypes are along the X axis and frequency in the sample set is along
the y axis.
Additional file 3: Figure S2. Haplotype frequencies for dhps alleles: A)
wildtype (n = 57) and B) 437G/540E (n = 95). Haplotypes are along the X
axis and frequency in the sample set is along the y axis.
Additional file 4: Figure S3. Relationships among 95 8-locus dhfr
microsatellite haplotypes from populations in Western Kenya as
determined by eBURST analysis. Samples from 1992-1999 (n = 134
samples) and 2002-2004 (n = 37 samples) were used. Each line connects
haplotypes that are identical at 7 of 8 loci. The size of the circles is
proportional to the number of isolates of the given haplotype. The blue
circles represent founders for the clusters and the yellow circle represent
subgroup founders. Black circles without any shading represent
haplotypes only present for the samples collected in 1992-1999, green
shading represents haplotypes present only in the 2002-2004 collection,
and pink shading represents haplotypes present in both collections. 51I/
59R/108N haplotypes circled in red are triple mutants that originated
independently from the SE Asian haplotype. Two genotypes that include
the mutation 164L are noted not being connected to any other
haplotype.
Additional file 5: Figure S4. Relationships among 128 9-locus dhps
microsatellite haplotypes from western Kenya as determined by eBURST
analysis. Each line connects haplotypes that are identical at 8 out of 9
loci. The size of the circles is proportional to the number of isolates of
the given haplotype. Haplotypes shown as single points differ from the
other haplotypes by alleles in at least 2 loci. The central complex
represents haplotypes from the 437G/540E allele. A total of 44 samples
with the wildtype allele and 84 samples with the 437G/540E allele were
utilized for this analysis.
Additional file 6: Table S2. Number of alleles (A) and heterozygosity
(He) per locus and averaged over loci.
Additional file 7: Figure S5. Relationships among 128 9-locus dhps
microsatellite haplotypes from western Kenya as determined by eBURST
analysis. Each line connects haplotypes that are identical at 8 out of 9
loci. The size of the circles is proportional to the number of isolates of
the given haplotype. Haplotypes shown as single points differ from the
other haplotypes by alleles in at least 2 loci. The central complex
represents haplotypes from the 437G/540E allele. A total of 44 samples
with the wildtype allele and 84 samples with the 437G/540E allele were
utilized for this analysis.
Additional file 8: Figure S6. Observed and predicted Heat ms loci
around the (A) dhfr allele 51I/59R/108N and (B) dhps allele 437G/540E.
For the dhfr prediction (A) we used Heamong all samples that did not
include the 51I/59R/108N triple mutant (i.e. infections with the mixed
codons N51I/S108N and C59R/S108N were included). For the dhps
prediction (B) we used Heamong wildtype alleles as an estimate for the
initial heterozygosity. Loci are labelled according to their positions
relative to dhfr or dhps (kb from the gene). Sampling variance is
indicated by error bars.
Additional file 9: Figure S7. Pairwise LD between microsatellite loci on
different chromosomes (A) and between sites in dhfr and dhps (B). Each
cell represents one comparison between polymorphic pairs of loci. Gray
cells represent significance at p value < 0.01. (A) The position of dhfr and
dhps along the chromosome is denoted by the gray bar. The location of
each microsatellite locus is at the top of the matrix (loci are named
according to their positions relative to dhfr or dhps or along
chromosome 2 or 3 according to the 3D7 genome sequence available
from NCBI). (B) Pairwise LD between sites in dhfr (51, 59, 108) and dhps
(436, 437, 540).
Acknowledgements
Financial support from the CDC Antimicrobial Resistance Working Group and
support from the Atlanta Research and Education Foundation (Atlanta, GA)
are appreciated. AE and KS are supported by the grant R01GM084320 from
the US National Institute of Health. SMG was supported by a National
Science Foundation Graduate Research Fellowship. We thank the CDC
Biotechnology Core Facility for the use the PSQ MA96 system for
pyrosequencing. This paper is published with the permission of KEMRI
Director. The findings and conclusions in this article are those of the authors
and do not necessarily represent the views of the Centers for Disease
Control and Prevention.
Author details
1Program in Population Biology, Ecology, and Evolution, Emory University,
Atlanta, GA, USA.2Malaria Branch, Division of Parasitic Diseases and Malaria,
Center for Global Health, Centers for Disease Control and Prevention,
Atlanta, GA, USA.3Atlanta Research and Education Foundation, Atlanta, GA,
USA.4Department of Mathematics, University of Vienna, Vienna, Austria.
5Kenya Medical Research Institute, Centre for Vector Biology and Control
Research, Kisumu, Kenya.6Liverpool School of Tropical Medicine, Liverpool,
UK.7School of Life Sciences, Arizona State University, Tempe, AZ, USA.
8Center for Evolutionary Medicine & Informatics, The Biodesign Institute,
Arizona State University, Tempe, AZ, USA.
Authors’ contributions
AMM, KAS, AAE, and VU designed the study and drafted the manuscript.
AMM, SMG, and ZZ carried out the molecular genetics studies. KAS carried
out the theoretical and statistical analyses. SK, FK, YPS, LS, and AAL
participated in the design and coordination of sample collection. All authors
read and approved the final manuscript.
Competing interests
The authors declare that they have no competing interests.
Received: 28 November 2011 Accepted: 22 March 2012
Published: 22 March 2012
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
Page 12 of 14

Page 13

References
1. Gregson A, Plowe CV: Mechanisms of resistance of malaria parasites to
antifolates. Pharmacol Rev 2005, 57:117-145.
2.Talisuna AO, Bloland P, D’Alessandro U: History, dynamics, and public
health importance of malaria parasite resistance. Clin Microbiol Rev 2004,
17:235-254.
3.Dondorp AM, Nosten F, Yi P, Das D, Phyo AP, Tarning J, Lwin KM, Ariey F,
Hanpithakpong W, Lee SJ, Ringwald P, Silamut K, Imwong M, Chotinavich K,
Lim P, Herdman T, An SS, Yeung S, Singhasivanon P, Day NP, Lindegardh N,
Socheat D, White NJ: Artemisinin resistance in Plasmodium falciparum
malaria. N Engl J Med 2009, 361:455-467.
4.Dondorp AM, Yeung S, White L, Nguon C, Day NP, Socheat D, Von-
Seidlein L: Artemisinin resistance: current status and scenarios for
containment. Nat Rev Microbiol 2010, 8:272-280.
5.Hayton K, Su XZ: Genetic and biochemical aspects of drug resistance in
malaria parasites. Curr Drug Targets Infect Disord 2004, 4:1-10.
6.Cortese JF, Plowe CV: Antifolate resistance due to new and known
Plasmodium falciparu dihydrofolate reductase mutations expressed in
yeast. Mol Biochem Parasitol 1998, 94:205-214.
7.Plowe CV, Kublin JG, Doumbo O: P. falciparum dihydrofolate reductase
and dihydropteroate synthase mutations: epidemiology and role in
clinical resistance to antifolates. Drug Resist Updat 1998, 1:389-396.
8.Triglia T, Wang P, Sims PF, Hyde JE, Cowman AF: Allelic exchange at the
endogenous genomic locus in Plasmodium falciparum proves the role of
dihydropteroate synthase in sulfadoxine-resistant malaria. EMBO J 1998,
17:3807-3815.
9.Nair S, Williams JT, Brockman A, Paiphun L, Mayxay M, Newton PN,
Guthmann JP, Smithuis FM, Hien TT, White NJ, Nosten F, Anderson TJ: A
selective sweep driven by pyrimethamine treatment in Southeast Asian
malaria parasites. Mol Biol Evol 2003, 20:1526-1536.
10. Roper C, Pearce R, Bredenkamp B, Gumede J, Drakeley C, Mosha F,
Chandramohan D, Sharp B: Antifolate antimalarial resistance in southeast
Africa: a population-based analysis. Lancet 2003, 361:1174-1181.
11.Roper C, Pearce R, Nair S, Sharp B, Nosten F, Anderson T: Intercontinental
spread of pyrimethamine-resistant malaria. Science 2004, 305:1124.
12.Maiga O, Djimde AA, Hubert V, Renard E, Aubouy A, Kironde F, Nsimba B,
Koram K, Doumbo OK, Le-Bras J, Clain J: A shared Asian origin of the
triple-mutant dhfr allele in Plasmodium falciparum from sites across
Africa. J Infect Dis 2007, 196:165-172.
13.Certain LK, Briceno M, Kiara SM, Nzila AM, Watkins WM, Sibley CH:
Characteristics of Plasmodium falciparum dhfr haplotypes that confer
pyrimethamine resistance, Kilifi, Kenya, 1987-2006. J Infect Dis 2008,
197:1743-1751.
14.McCollum AM, Basco LK, Tahar R, Udhayakumar V, Escalante AA:
Hitchhiking and selective sweeps of Plasmodium falciparum sulfadoxine
and pyrimethamine resistance alleles in a population from central Africa.
Antimicrob Agents Chemother 2008, 52:4089-4097.
15.McCollum AM, Poe AC, Hamel M, Huber C, Zhou Z, Shi YP, Ouma P,
Vulule J, Bloland P, Slutsker L, Barnwell JW, Udhayakumar V, Escalante AA:
Antifolate resistance in Plasmodium falciparum: multiple origins and
identification of novel dhfr alleles. J Infect Dis 2006, 194:189-197.
16.Alam MT, Vinayak S, Congpuong K, Wongsrichanalai C, Satimai W,
Slutsker L, Escalante AA, Barnwell JW, Udhayakumar V: Tracking origins and
spread of sulfadoxine-resistant Plasmodium falciparum dhps alleles in
Thailand. Antimicrob Agents Chemother 2011, 55:155-164.
17.Pearce RJ, Pota H, Evehe MS, Ba-el H, Mombo-Ngoma G, Malisa AL, Ord R,
Inojosa W, Matondo A, Diallo DA, Mbacham W, Van den-Broek IV,
Swarthout TD, Getachew A, Dejene S, Grobusch MP, Njie F, Dunyo S,
Kweku M, Owusu-Agyei S, Chandramohan D, Bonnet M, Guthmann JP,
Clarke S, Barnes KI, Streat E, Katokele ST, Uusiku P, Agboghoroma CO,
Elegba OY: Multiple origins and regional dispersal of resistant dhps in
African Plasmodium falciparum malaria. PLoS Med 2009, 6:e1000055.
18.Vinayak S, Alam MT, Mixson-Hayden T, McCollum AM, Sem R, Shah NK,
Lim P, Muth S, Rogers WO, Fandeur T, Barnwell JW, Escalante AA,
Wongsrichanalai C, Ariey F, Meshnick SR, Udhayakumar K: Origin and
evolution of sulfadoxine resistant Plasmodium falciparum. PLoS Pathog
2010, 6:e1000830.
19.Pearce R, Malisa A, Kachur SP, Barnes K, Sharp B, Roper C: Reduced
variation around drug-resistant dhf alleles in African Plasmodium
falciparum. Mol Biol Evol 2005, 22:1834-1844.
20.McCollum AM, Mueller K, Villegas L, Udhayakumar V, Escalante AA:
Common origin and fixation of Plasmodium falciparum dhf and dhp
mutations associated with sulfadoxine-pyrimethamine resistance in a
low-transmission area in South America. Antimicrob Agents Chemother
2007, 51:2085-2091.
Nash D, Nair S, Mayxay M, Newton PN, Guthmann JP, Nosten F,
Anderson TJ: Selection strength and hitchhiking around two anti-malarial
resistance genes. Proc Biol Sci 2005, 272:1153-1161.
Terlouw DJ, Nahlen BL, Courval JM, Kariuki SK, Rosenberg OS, Oloo AJ,
Kolczak MS, Hawley WA, Lal AA, Kuile FO: Sulfadoxine-pyrimethamine in
treatment of malaria in Western Kenya: increasing resistance and
underdosing. Antimicrob Agents Chemother 2003, 47:2929-2932.
Nguyen-Dinh P, Spencer HC, Chemangey-Masaba S, Churchill FC:
Susceptibility of Plasmodium falciparum to pyrimethamine and
sulfadoxine/pyrimethamine in Kisumu, Kenya. Lancet 1982, 1:823-825.
Bloland PB, Boriga DA, Ruebush TK, McCormick JB, Roberts JM, Oloo AJ,
Hawley W, Lal A, Nahlen B, Campbell CC: Longitudinal cohort study of the
epidemiology of malaria infections in an area of intense malaria
transmission II. Descriptive epidemiology of malaria infection and
disease among children. Am J Trop Med Hyg 1999, 60:641-648.
Beier JC, Perkins PV, Onyango FK, Gargan TP, Oster CN, Whitmire RE,
Koech DK, Roberts CR: Characterization of malaria transmission by
Anopheles (Diptera: Culicidae) in western Kenya in preparation for
malaria vaccine trials. J Med Entomol 1990, 27:570-577.
Zhou Z, Poe AC, Limor J, Grady KK, Goldman I, McCollum AM, Escalante AA,
Barnwell JW, Udhayakumar V: Pyrosequencing, a high-throughput method
for detecting single nucleotide polymorphisms in the dihydrofolate
reductase and dihydropteroate synthetase genes of Plasmodium
falciparum. J Clin Microbiol 2006, 44:3900-3910.
Feil EJ, Li BC, Aanensen DM, Hanage WP, Spratt BG: eBURST: inferring
patterns of evolutionary descent among clusters of related bacterial
genotypes from multilocus sequence typing data. J Bacteriol 2004,
186:1518-1530.
Wright S: The interpretation of population structure by F-statistics with
special regards to systems of mating. Evolution 1965, 19:395-420.
Excoffier L, Laval G, Schneider S: Arlequin ver. 3.0: an integrated software
package for population genetics data analysis. Evolutionary Bioinformatics
Online 2005, 1:47-50.
Park SDE: Trypanotolerance in West African cattle and the population
genetic effects of selection. PhD thesis, University of Dublin 2001.
Raymond M, Rousset F: An exact test for population differentiation.
Evolution 1995, 49:1280-1283.
Maynard Smith J, Haigh J: The hitch-hiking effect of a favourable gene.
Genet Res 1974, 23:23-35.
Wiehe T: The effect of selective sweeps on the variance of the allele
distribution of a linked multiallele locus: hitchhiking of microsatellites.
Theor Popul Biol 1998, 53:272-283.
Kim Y, Stephan W: Detecting a local signature of genetic hitchhiking
along a recombining chromosome. Genetics 2002, 160:765-777.
Kaplan NL, Hudson RR, Langley CH: The “hitchhiking effect” revisited.
Genetics 1989, 123:887-899.
McVean G: The structure of linkage disequilibrium around a selective
sweep. Genetics 2007, 175:1395-1406.
Stephan W, Song YS, Langley CH: The hitchhiking effect on linkage
disequilibrium between linked neutral loci. Genetics 2006, 172:2647-2663.
Schneider KA, Kim Y: An analytical model for genetic hitchhiking in the
evolution of antimalarial drug resistance. Theor Popul Biol 2010, 78:93-108.
Schneider KA, Kim Y: Approximations for the hitchhiking effect caused by
the evolution of antimalarial-drug resistance. J Math Biol 2011,
62:789-832.
Nzila AM, Mberu EK, Sulo J, Dayo H, Winstanley PA, Sibley CH, Watkins WM:
Towards an understanding of the mechanism of pyrimethamine-
sulfadoxine resistance in Plasmodium falciparum: genotyping of
dihydrofolate reductase and dihydropteroate synthase of Kenyan
parasites. Antimicrob Agents Chemother 2000, 44:991-996.
Hapuarachchi HC, Dayanath MY, Bandara KB, Abeysundara S,
Abeyewickreme W, De-Silva NR, Hunt SY, Sibley CH: Point mutations in the
dihydrofolate reductase and dihydropteroate synthase genes of
Plasmodium falciparum and resistance to sulfadoxine-pyrimethamine in
Sri Lanka. Am J Trop Med Hyg 2006, 74:198-204.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
Page 13 of 14

Page 14

42.Nzila A, Ochong E, Nduati E, Gilbert K, Winstanley P, Ward S, Marsh K: Why
has the dihydrofolate reductase 164 mutation not consistently been
found in Africa yet? Trans R Soc Trop Med Hyg 2005, 99:341-346.
Pennings PS, Hermisson J: Soft sweeps II-molecular population genetics
of adaptation from recurrent mutation or migration. Mol Biol Evol 2006,
23:1076-1084.
Pennings PS, Hermisson J: Soft sweeps III: the signature of positive
selection from recurrent mutation. PLoS Genet 2006, 2:e186.
Paget-McNicol S, Saul A: Mutation rates in the dihydrofolate reductase
gene of Plasmodium falciparum. Parasitology 2001, 122:497-505.
Joy DA, Feng X, Mu J, Furuya T, Chotivanich K, Krettli AU, Ho M, Wang A,
White NJ, Suh E, Beerli P, Su XZ: Early origin and recent expansion of
Plasmodium falciparum. Science 2003, 300:318-321.
White NJ, Pongtavornpinyo W: The de novo selection of drug-resistant
malaria parasites. Proc Biol Sci 2003, 270:545-554.
43.
44.
45.
46.
47.
doi:10.1186/1475-2875-11-77
Cite this article as: McCollum et al.: Differences in selective pressure on
dhps and dhfr drug resistant mutations in western Kenya. Malaria
Journal 2012 11:77.
Submit your next manuscript to BioMed Central
and take full advantage of:
• Convenient online submission
• Thorough peer review
• No space constraints or color figure charges
• Immediate publication on acceptance
• Inclusion in PubMed, CAS, Scopus and Google Scholar
• Research which is freely available for redistribution
Submit your manuscript at
www.biomedcentral.com/submit
McCollum et al. Malaria Journal 2012, 11:77
http://www.malariajournal.com/content/11/1/77
Page 14 of 14