Adams et al. Parasites & Vectors  (2018) 11:175 
https://doi.org/10.1186/s13071-018-2762-3
RESEARCH Open Access
Prevalence of Plasmodium falciparum
delayed clearance associated polymorphisms
in adaptor protein complex 2 mu subunit
(pfap2mu) and ubiquitin specific protease 1
(pfubp1) genes in Ghanaian isolates
Tryphena Adams1, Nana Aba A. Ennuson1, Neils B. Quashie2,3, Godfred Futagbi1, Sena Matrevi2,
Oheneba C. K. Hagan4, Benjamin Abuaku2, Kwadwo A. Koram2 and Nancy O. Duah2*
Abstract
Background: Plasmodium falciparum delayed clearance with the use of artemisinin-based combination therapy
(ACTs) has been reported in some African countries. Single nucleotide polymorphisms (SNPs) in two genes, P.
falciparum adaptor protein complex 2 mu subunit (pfap2mu) and ubiquitin specific protease 1 (pfubp1), have been
linked to delayed clearance with ACT use in Kenya and recurrent imported malaria in Britain. With over 12 years of
ACT use in Ghana, this study investigated the prevalence of SNPs in the pfap2mu and pfubp1 in Ghanaian clinical
P. falciparum isolates to provide baseline data for antimalarial drug resistance surveillance in the country.
Methods: Filter paper blood blots collected in 2015–2016 from children aged below 9 years presenting with
uncomplicated malaria at hospitals in three sentinel sites Begoro, Cape Coast and Navrongo were used. Parasite
DNA was extracted from 120 samples followed by nested polymerase chain reaction (nPCR). Sanger sequencing
was performed to detect and identify SNPs in pfap2mu and pfubp1 genes.
Results: In all, 11.1% (9/81) of the isolates carried the wildtype genotypes for both genes. A total of 164 pfap2mu
mutations were detected in 67 isolates whilst 271 pfubp1 mutations were observed in 72 isolates. The majority of
the mutations were non-synonymous (NS): 78% (128/164) for pfap2mu and 92.3% (250/271) for pfubp1. Five unique
samples had a total of 215 pfap2mu SNPs, ranging between 15 and 63 SNPs per sample. Genotypes reportedly
associated with ART resistance detected in this study included pfap2mu S160N (7.4%, 6/81) and pfubp1 E1528D
(7.4%, 6/81) as well as D1525E (4.9%, 4/81). There was no significant difference in the prevalence of the SNPs between
the three ecologically distinct study sites (pfap2mu: χ2 = 6.905, df = 2, P = 0.546; pfubp1: χ2 = 4.883, df = 2, P = 0.769).
Conclusions: The detection of pfap2mu and pfubp1 genotypes associated with ACT delayed parasite clearance is
evidence of gradual nascent emergence of resistance in Ghana. The results will serve as baseline data for surveillance
and the selection of the genotypes with drug pressure over time. The pfap2mu S160N, pfubp1 E1528D and D1525E
must be monitored in Ghanaian isolates in ACT susceptibility studies, especially when cure rates of ACTs, particularly
AL, is less than 100%.
Keywords: Plasmodium falciparum, Antimalarial drug resistance, Artemisinin, ACT, pfubp1, pfap2mu, Mutations, Ghana
* Correspondence: nduah@noguchi.ug.edu.gh
2Department of Epidemiology, Noguchi Memorial Institute for Medical
Research, College of Health Sciences, University of Ghana, Accra, Ghana
Full list of author information is available at the end of the article
© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Adams et al. Parasites & Vectors  (2018) 11:175 Page 2 of 12
Background gene mutations were at codon 1525, a change from aspar-
Malaria is still a debilitating disease, especially in sub- tic acid to glutamic acid (D1525E) and codon 1528 from
Saharan Africa (sSA) where there were 212 million cases glutamic acid to aspartic acid (E1528D) in the Kenyan iso-
and 429,000 malaria-related deaths in 2015 [1]. There has lates. In addition, Sutherland et al. [8] reported four UK
been a 21 and 29% reduction in morbidity and mortality, residents with imported malaria who showed recurrent
respectively, since 2010, probably as a result of the imple- malaria after AL treatment [8]. These mutations, pfap2mu
mentation of integrated control strategies [1, 2]. The S160N and pfubp1 E1525D/Q were observed in the recur-
control efforts employed included the use of insecticide- rent parasites but none of the known K-13 gene mutations
treated mosquito nets (ITNs), indoor residual spraying were observed [8]. Although the role played by both genes
(IRS), chemoprevention in pregnant women and children in artemisinin action is not clearly understood, the ap2mu
as well as chemotherapy with artemisinin-based combin- gene is known to encode the μ- subunit of the adaptor
ation therapy (ACT). As such, the development of protein 2 complex (AP2) involved in clathrin-mediated
Plasmodium falciparum resistance to artemisinin (ART) endocytosis into the parasite vacuole [9]; ubp1 encodes a
derivatives as reported from Southeast Asia (SEA) is quite deubiquitinating (DUB) enzyme that functions by cleaving
worrying [1]. Chemotherapy, which is one of the core con- ubiquitin from any protein or peptide to which it is joined
trol strategies for the disease, has been hindered over the [10]. The polymorphic homologues of these two P. falcip-
years by the emergence and spread of parasites resistant arum genes were first identified in the rodent malaria
to the commonly used antimalarial drugs [1]. Currently, parasite, P. chabaudi (pcubp1 encodes ubiquitin carboxy-
the World Health Organisation (WHO) has initiated the terminal hydrolase 1) and pcap2mu encodes clathrin
containment of drug resistance in the SEA region with the vesicle-associated adaptor 2 mu subunit), as being associ-
deployment of a multi-sector strategy [1]. Although this ated with ART resistance [11]. More studies are therefore
initiative is commendable, the need for country-level needed to validate these polymorphisms and their role in
monitoring of the genome of parasite populations for pos- antimalarial drug resistance.
sible evolution and selection due to drug pressure is also The use of ACTs in Ghana began in 2005 and since then
crucial for the early detection of emerging drug resistance. the cure rate of the drugs in use, artesunate-amodiaquine
For over a decade, molecular markers of antimalarial (AS-AQ) and artemether-lumifantrine (AL), has been 100
drug resistance have been used to monitor the emer- and 97.6%, respectively, as of 2014 [12]. Surveillance stud-
gence and spread of drug resistance in malaria endemic ies using the P. falciparum multidrug resistance gene
areas. These genetic markers are mainly single nucleo- (pfmdr1) SNPs (haplotype N86-F184-D1246) linked to re-
tide polymorphisms in genes encoding drug target pro- duced parasite susceptibility to AL showed an increasing
teins in essential biochemical pathways of the parasite. trend over the years in Ghana from 2005 to 2010 [13]. In
The levels of drug susceptibility in the parasites have addition, increased pfmdr1 gene copy number linked to
been linked to SNPs or haplotypes of the genes, and parasite reduced susceptibility to artesunate (AS), meflo-
these markers are relevant in antimalarial drug efficacy quine (MQ), halofantrine and AL [14–16] were also de-
studies. The recent observation of parasite resistance to tected in Ghanaian isolates [13]. The findings from the
ART in SEA set into motion the need to discover a mo- reported studies above are indicative of a subtle emer-
lecular marker for surveillance of drug susceptibility. gence of parasite resistance to ARTand to ACTs especially
Ariey et al. [3] discovered the SNPs in the kelch propel- AL in Ghana. Therefore, the monitoring of newly discov-
ler domain on chromosome 13 of the P. falciparum gen- ered molecular markers is essential as an early warning
ome known as k13 in drug resistance isolates in vitro. signal to the emergence of resistance in Ghana. This study
Three of the k13 polymorphisms, C580Y, R538T and determined the prevalence of known and novel SNPs in
Y493H were also present in slow clearing clinical isolates the pfubp1 and pfap2mu in Ghanaian isolates collected
with ART use. The presence of the SNPs showed varying from three ecologically distinct areas for monitoring anti-
parasite clearance half-life in patients; however, the malarial drug efficacy in Ghana to serve as baseline data
C580Y mutation was linked to longest parasite clearance for antimalarial drug resistance surveillance in Ghana.
half-life compared to the other SNPs. So far these SNPs
have not been detected in African isolates [4–6] and the Methods
quest for novel markers for ART resistance is ongoing. Study sites
Henriques et al. [7] linked SNPs in two genes, the P. falcip- The Noguchi Memorial Institute for Medical Research
arum adaptor protein complex 2 mu subunit (pfap2mu) (NMIMR) in collaboration with the National Malaria
and ubiquitin-specific protease 1 (pfubp1) to delayed clear- Control Programme (NMCP) have set up ten sentinel
ance of parasites. The pfap2mu gene mutation was at sites in the ten regions of Ghana for monitoring antimal-
codon 160 resulting in the amino acid change from serine arial drug efficacy. These sites lie in the three distinct
to either asparagine or threonine (S160N/T). The pfubp1 ecological zones in the country. Of the ten sites, samples
Adams et al. Parasites & Vectors  (2018) 11:175 Page 3 of 12
from three sites were used for this study. The sites in- Detection of pfubp1 and pfap2mu gene polymorphisms
clude Navrongo (10°53'44.05"N, 1°05'31.56"W) located in Parasite DNA was extracted from 120 pre-treatment
the Kassena Nankana District in the Upper East Region blood blot samples on filter paper (WhatmanTM 3 Little
and lies in the guinea savannah zone; Begoro (6° Chalfont, United Kingdom) using the QIAmp DNA mini
23'29.76"N, 0°22' 46.20"W) located in the Fanteakwa Dis- kit (Qiagen GmBH, Hilden, Germany) as per the manu-
trict of the Eastern Region lies in the forest zone; Cape facturer’s protocol. This was followed by amplification of
Coast (5°06'00”N, 1°15'00"W), the capital town of Cen- the pfap2mu and pfubp1 genes using nested polymerase
tral Region lies in the coastal savannah zone (Fig. 1). chain reaction (nPCR) following a previously published
The forest and coastal savannah zones experience perennial protocol [7] with minor modifications. The PCR was
malarial transmission whilst the guinea savannah zone ex- performed in a total volume of 25 μl with the following
periences a seasonal malaria transmission pattern with al- reaction mixture: 0.2 μM of each primer (Table 1), 4.0
most all cases occurring during rainy months between mM MgCl2, 0.4 μM deoxynucleotides triphosphate
May-June and October-November of each year. (dNTPs), 1 U One-Taq polymerase (New England Biolabs,
Massachusetts, USA), 1× PCR buffer, nuclease free water
Study samples and 2 μl of the extracted parasite DNA. One microlitre of
Filter paper blood blots collected in 2015–2016 from chil- the first round product was used as a template in a 50 μl
dren aged below 9 years with uncomplicated malaria report- inner PCR reaction. A DNA sample extracted from the
ing at designated health care facilities in Navrongo, Begoro 3D7 parasite strain was used as a positive control. The
and Cape Coast were used for the study. In total 120 sam- PCR thermal conditions were the same for both genes but
ples (40 from each site) were used for this investigation. different annealing temperatures as shown in Table 1. The
Fig. 1 The map of Ghana showing the three study sites, Navrongo, Begoro and Cape Coast in three different ecological areas, guinea savannah,
forest and coastal savanna
Adams et al. Parasites & Vectors  (2018) 11:175 Page 4 of 12
Table 1 Primer sequences, sizes of PCR amplicons and annealing temperature of the amplification of pfap2mu and pfubp1 genes
Gene Primers (5'–3') Size of PCR amplicon (bp) Annealing temperature (°C)
pfap2mu Primary amplification 2247 50
Forward: AAGACTGTCAAATGTAAAAGACCC
Reverse: CTCATGTAAAACAAAAAGTGAGG
Secondary amplification 841 52
Forward: GATATCCACAAACATTAGAAGTG
Reverse: CCATCTGGTGGTGTGAAGG
pfubp1 Primary amplification 484 52
Forward: CGCCCGTACTATGAAGAAGATC
Reverse: GGCTTTTACCTGAACTGTTCAGG
Secondary amplification 304 57
Forward: CGTAAACAGAATATTCAGGATTGC
Reverse: CTAGCCCTTTATTATCATTATCG
thermal cycle programme for each 1st amplification was followed by 40 cycles of 94 °C for 30 s, annealing
94 °C for 3 min, and 30 cycles of 94 °C for 30 s, annealing temperature for 30 s and 68 °C for 45 s with a final
temperature for 30 s and 68 °C for 1 min with a final exten- extension of 68 °C for 10 min. The PCR amplicons
sion of 68 °C for 15 min. The second round of PCR for the fragments of the two genes were sequenced
consisted of initial denaturation at 94 °C for 3 min, using Sanger sequencing.
Fig. 2 Distribution of pfap2mu and pfubp1 NS and SYN mutations in isolates from the three sites. a pfap2mu. b pfubp1
Adams et al. Parasites & Vectors  (2018) 11:175 Page 5 of 12
Data analysis assurance analysis. Of the 81 good sequences, 35% (28/
The sequence data of the isolates were analysed using 81), 35% (28/81) and 31% (25/81) were from Cape-
the CLC Genomics Workbench 10.01 software (Qiagen, Coast, Begoro and Navrongo, respectively. The propor-
Aarhus, Denmark) and Benchling.com (California, CA, tion of isolates with no mutations in the pfap2mu gene,
USA). PF3D7_1218300 and PF3D7_0104300 (PlasmoDB) that is wildtype sequence as the 3D7 strain, was 11.1%
were used as reference sequences to detect SNPs in (9/81). The sequence analysis revealed several SNPs and
the pfap2mu and pfubp1 respectively. Poor quality se- the total number observed in 67 samples (of the 72 with
quences of isolates after three sequencing trials were not mutations) was 164 SNPs with ≤ 5 mutations per sam-
analysed. The prevalence of individual SNPs was determined ple. Of the 164 SNPs, 78.0% (128/164) were non-
for each site. Chi-square tests were used to compare the synonymous (NS) and 22.0% (36/164) were synonymous
proportions of mutations occurring in the three sites and to (SYN) mutations. The distribution of pfap2mu NS and
determine any significant differences in the prevalence of SYN mutations in isolates from the three sites is shown
the mutations among the three sites using the GraphPad in Fig. 2a. There was no significant difference in the type
Prism 5 (GraphPad Software Inc, La Jolla, CA, USA). Statis- of mutation (SYN or NS) present between the three sites
tical significance was defined as a P-value ≤ 0.05. (χ2 = 1.960, df = 2, P = 0.360). The distribution of the
single or multiple mutations for each site is shown in
Results Fig. 3a. About 68.6% (46/67) of the isolates with muta-
Polymorphisms in pfap2mu and pfubp1 tions had more than one mutation and 36% (9/25) of the
For the pfap2mu gene, 96 samples were sequenced and isolates from Navrongo had more than three mutations
15 were of low quality as determined by quality per isolate. In addition, there were insertions and
Fig. 3 Proportion of isolates from the three sites with varying number of pfap2mu and pfubp1 mutations. a pfap2mu. b pfubp1
Adams et al. Parasites & Vectors  (2018) 11:175 Page 6 of 12
deletions in some of the isolates from the three sites. In For the pfubp1 gene, 81 quality sequence data were
all 22 common SNPs were detected in either two of the analysed comprising 37% (30/81), 35.8% (29/81) and
three sites or all three sites. These include Q149R, 27.2% (22/81) from Cape Coast, Begoro and Navrongo,
S160N, V161K, V161E, D168E, R188R, D203F, D203Y, respectively. About 11.1% (9/81) of the isolates had no
E206*, T235T, N240Y, N240F, K256*, D263V, V270V, mutations in their pfubp1 gene. A total of 271 SNPs
I272I, G284G, K285E, T302T, N317S, T318T and were observed in the 72 sequences of which 92.3% (250/
T325T. Nine of these SNPs were shared in all three sites 271) were NS and 7.7% (21/271) were SYN mutations.
and the proportion of isolates from the study sites is The proportion of isolates with either pfubp1 NS or
shown in Fig. 4a. Of these SNPs, three NS mutations SYN mutations for each site is shown in Fig. 2b. The
were present in isolates from the three sites: S160N, proportion of isolates with varying number of mutations
D168E and V161K. The nucleotide changes for the SNPs is also shown in Fig. 3b for the three study sites. Overall,
are shown in Table 2. The amino acid sequences alignment 93.1% (67/72) of the isolates with mutations had more than
for 21 isolates are shown in Fig. 5. There were 5 other iso- one pfubp1 SNP. The majority of isolates from Begoro
lates (of the 72 with mutations) with a total of 215 pfap2mu (65.5%, 19/29) had more than 3 mutations per isolate for
SNPs (G020, 63 SNPs; G022, 60 SNPs; G025, 47 SNPs; the pfubp1 gene. There were 15 common SNPs detected in
G029, 30 SNPs; G034, 15 SNPs) and were all from isolates from all three sites. These include D1539D,
Begoro. The amino acid sequences for these 5 isolates E1528D, I1487I, I1552M, K1502*, K1537L, N1542D,
are shown in Fig. 6. N1551G, N1560I, N1560K, P1547P, Q1543H, Y1501F,
Fig. 4 Proportion of isolates from the three sites with shared pfap2mu and pfubp1 mutations. a pfap2mu. b pfubp1
Adams et al. Parasites & Vectors  (2018) 11:175 Page 7 of 12
Table 2 Shared pfap2mu mutations observed in the isolates Y1549* and Y1549L. The proportion of isolates with these
from the three sites. Mutations indicated in bold are known mutations from each of the study sites is shown in Fig. 4b.
delayed clearance genotypes, underlined nucleotides are the There were 36 other shared SNPs detected in isolates from
changed bases two out of the three sites. The nucleotide changes for the
Nucleotide position Nucleotide change Amino acid position commonly shared mutations are also shown in Table 3.
and change The amino acid sequence alignment is shown in Fig. 7.
446 CAG to CGG Q149R Two samples, G005 and C329, had wildtype sequences
479 AGT to AAT S160N for both pfap2mu and pfubp1. Of the 5 isolates with
481 GTG to AAG V161K many pfap2mu SNPs, 2 were wildtype for pfubp1 (G022
482 GTG to GAG V161E and G034). For the other three, G020 had 4 SNPs, G025
504 GAT to GAA D168E had 2 SNPs and G029 had 11 SNPs for the pfubp1.
564 AGA to AGG R188R
Prevalence of pfap2mu and pfubp1 SNPs from the three
607 GAT to TAT D203Y sentinel sites
607, 608 GAT to TTT D203F The pfap2mu SNPs were detected in 92.9% (26/28),
616 GAA to TAA E206a 85.7% (24/28) and 88% (22/25) from Begoro, Cape Coast
705 ACA to ACG T235T and Navrongo respectively. There was no significant dif-
718 AAT to TAT N240Y ference in the prevalence of pfap2mu SNPs in the iso-
a lates from the three sites (χ
2 = 6.905, df = 2, P = 0.546).
767 AAG to TAG K256
The most prevalent mutation was R188R and was ob-
789 GAT to GTT D263V served in 25.9% (21/81) of the isolates. The D168E was
810 GTA to GTT V270V also observed in 22.2% (18/81) of the isolates. S160N
852 GGA to GGG G284G which have been reported to be associated with ACT de-
855 AAG to GAG K285E layed clearance was prevalent in 7.4% (6/81) of the iso-
951 AAC to AGC N317S lates whilst the V161K was seen in 6.2% (5/81). The
S160N genotypes occurred mostly in isolates from
954 ACA to ACC T318T
Cape Coast (24.7%) as compared to the other two
aStop codon sites. In all, 35 indels were identified with 17.1% (6/
35) causing a frame shift in the sequence reading
frame (Fig. 5).
For pfubp1, mutations were detected in 86.7% (26/30)
of the Cape Coast isolates, 86.2% (25/29) for Begoro and
Fig. 5 A sequence alignment of pfap2mu gene showing amino acid changes due to single nucleotide polymorphisms. The alignment was done
using pfap2mu reference sequence of the 3D7 strain (PF3D7_1218300). Mutations present at codons 185–282, nucleotide positions 553–849 of
pfap2mu for 21 samples. Samples C314 and G026 had a frameshift, samples C312. C314, C315, C408, G009, G011, G012, G033, N030 and N102 had
an asparagine (N) insertion at codon 233, as well as a lysine (K) insertion at the same position for sample N055
Adams et al. Parasites & Vectors  (2018) 11:175 Page 8 of 12
Fig. 6 A sequence alignment for 5 samples with multiple mutations. The alignment was done using pfap2mu reference sequence of the 3D7
strain (PF3D7_1218300). About 215 SNPs were observed in these isolates ranging from 15 to 63 SNPs per isolate. These mutations were found
from codons 227–324, nucleotide positions 679–970 and are likely due to multiplicity of infection
95.5% (21/22) for Navrongo. There was no significant associated polymorphisms of the pfap2mu and pfubp1
difference in the prevalence of SNPs in the isolates from genes in Ghanaian isolates from three sites located in
the three sites (χ2 = 4.883, df = 2, P = 0.769). The most three distinct ecological zones. Majority of the isolates
prevalent NS mutation was K1502* which was observed had mutations of both genes and were mostly NS muta-
in 18.5% (15/81) of the isolates. The reported SNPs tions. The known delayed clearance genotypes, pfap2mu
linked to delayed clearance of parasite with ACT use, N160 and pfubp1 D1528 and E1525, were observed in
E1528D and D1525E, were observed in 7.4% (6/81) and 7.4, 7.4 and 4.9% of the isolates, respectively. It is inter-
4.9% (4/81) of the isolates, respectively. The predomin- esting to note that although these were not the predom-
ant SYN mutation I1487I was observed in 11.1% (9/81) inant mutations, it is an indication of the presence of
of the isolates. In all, 20 indels were identified with 25% these drug resistance genotypes in Ghanaian parasite
(5/20) causing a frame shift in the sequence reading populations and in the long term their selection with
frame (Table 4). drug use will enhance the emergence of ART resistance
The results showed a number of SNPs that were being in Ghana.
inherited together on the gene as haplotypes in some of The observation that minority of the isolates were
the isolates. However, the prevalence of the haplotypes wildtype with no mutations like the reference 3D7 strain
was low (Table 5). Twenty-two different haplotypes were (11.1% for both pfap2mu and pfubp1) is indicative of the
observed: 3 for pfap2mu and 19 for pfubp1. The most high rate of spontaneous mutations in the two genes.
prevalent haplotype for the pfap2mu gene was V161K- For pfap2mu, a study by Henriques et al. [7] reported
D168E, which was observed in 8.6% (7/81) of the isolates that 41.5% of Kenyan isolates had wildtype gene se-
with mutations. For the pfubp1 gene, the haplotypes quence whilst another study reported 92.8% and 30.9%
Y1548L-N1560I, Y1549L-I1552M and N1560I-L1563 for Ethiopian and Tanzanian isolates, respectively [17].
were each observed in 3.7% (3/81) of the isolates. Most Comparatively, the Ghanaian isolates had low levels of
of the haplotypes for both genes were observed in iso- wildtype strains portraying a rapid genetic recombin-
lates from Begoro. ation of different parasite clones (multiclonal infections
observed in Ghana) during the sexual stage in the vector
Discussion resulting in gene shuffling [18]. About 64% and 93% of
The search for a molecular marker for the early detec- the isolates, respectively, had more than one mutation
tion of ART resistance by the malaria parasite is on- for the pfap2mu and pfubp1 genes.
going. The use of molecular markers to track and The results revealed a high proportion of NS muta-
identify early development of parasite resistance to drugs tions, 78% and 92% for pfap2mu and pfubp1, respect-
is a powerful tool that should be available in all malari- ively, in the Ghanaian isolates. The NS mutations are of
ous regions of the world. With the implementation of much importance because each amino acid substitution
ACTs in Africa, studies to identify possible markers of may affect protein conformation and function [19]. Most
resistance have not been conducted extensively in the of the NS mutations were single base or double base
continent. The key k13 molecular marker, which has substitutions. For SYN mutations, it was initially as-
been linked to drug resistance in SEA has not been ob- sumed that since the resultant change of the nucleotide
served in African isolates, partly because there is no does not affect the amino acid, the change may go un-
‘true resistance’ to ARTs except delayed clearance of par- detected as the gene function may not necessarily be
asites. New markers that have recently been discovered, affected [20]. However, this perception has since chan-
such as the pfap2mu and pfubp1 gene mutations [7], ged due to the evidence that SYN mutations in the
need further validation for their role in delayed parasite pfmdr1 gene resulted in alterations in the functions of
clearance. This study detected ART delayed clearance the P-glycoprotein (P-gp), a product which affects drug
Adams et al. Parasites & Vectors  (2018) 11:175 Page 9 of 12
Table 3 Shared pfubp1 mutations observed in the isolates from Table 3 Shared pfubp1 mutations observed in the isolates from
the three sites. Mutations indicated in bold are known delayed the three sites. Mutations indicated in bold are known delayed
clearance genotypes clearance genotypes (Continued)
Nucleotide position Nucleotide change Amino acid position Nucleotide position Nucleotide change Amino acid position
and change and change
4383 CCT to ACC P1461T 4656 ATT to ATG I1552M
4386 TAT to TTA Y1462L 4659 AAT to AAC N1553N
4389 CGT to TCG R1463S 4667 TAC to TGC Y1556C
4392 AAA to TAA K1464a 4674 AAT to GAT N1558D
4461 ATA to ATC I1487I 4679 AAT to ATA N1560I
4461 ATA to ACC I1487T 4680 AAT to AAA N1560K
4466 ATG to ACG M1489T 4683 AAA to AAG K1561K
4502 TAT to TTT Y1501F 4692 GAG to GAC E1564D
4504 AAA to TAA K1502a 4695 TTC to GGC F1565G
4509 AAT to ATT N1503I 4694 TTC to CAA F1565Q
4509 AAT to ATG N1503M 4765 AAA to TAA K1589a
4527 GAA to GAC E1509D aStop codon
4554 AAC to TAT N1518Y
4557 GAA to GAC E1519D interactions [21]. A high prevalence of novel ap2mu mu-
4561 TAT to AAT Y1521N tation, D168E (25%), was observed in the Ghanaian iso-
4575 GAC to GAA D1525E lates from all three sites followed by V161K (7%). The
codons for the common mutations found among the iso-
4581 TAT to TTT Y1527F
lates from all the three sites were between codons 160
4581 TAT to AAT Y1527N and 170. The known SNP, pfap2mu S160N, found in
4584 GAA to GAC E1528D Kenyan isolates with delayed clearance [7] was also
4599 TAT to TAC Y1533Y found in all three sites (7.2%). In addition, isolates from
4602 GAT to GAA D1534E a UK patient who arrived from Angola and failed AL
4608 TAC to TCC Y1536S treatment had the S160N thereby strengthening the role
of that mutation in recurrent parasitemia with AL use
4611 AAA to TTT K1537F
[8]. The observed pfubp1 E1528D in Ghanaian samples
4609 AAA to TTA K1537L was also seen in Kenyan and Tanzanian isolates [7, 17].
4611 AAA to TAC K1537Y Borrmann et al. [22] first identified this mutation in
4617 GAT to GAC D1539D Kenyan isolates. The prevalence of E1528D in Ghanaian
4623 AAA to CTT K1541L isolates was 7.4%, which is lower than that of Kenya’s
4623 AAA to AAT K1541N 17.1% but higher than that of Tanzania (4.8%) [7, 17].
However, with the continuing use of ACTs, the mutation
4626 AAT to GAT N1542D
may be selected and the prevalence may increase with
4629 CAA to CAT Q1543H time as observed in Kenya. The SYN mutation N1518
4632 CAT to CTT H1544L found in the Kenyan and Burkinabe isolates was also seen
4632 CAT to CCT H1544P in one of the Ghanaian isolates from Begoro. Pre- and
4632 AAA to AAT K1544N post-treatment samples from Burkina Faso, had the
4641 CCA to CCT P1547P D1525E mutation which was seen in four Ghanaian isolates.
a Henriques et al. [7] compared South East Asian and African4647 TAT to TAG Y1549
phenotypes of the pfubp1 gene and detected significant dif-
4647 TAT to TTT Y1549F ferences in the genetic signatures. The recent implication of
4646 TAT to TTG Y1549L SNPs in these genes, especially pfap2mu S160N and pfubp1
4650 GAT to ATT D1550I E1528D and D1525E in ART resistance, raises a genuine
4650 GAT to AAT D1550N concern due to their presence in Ghanaian isolates.
4651 AAT to GGT N1551G Of the three sites, Begoro (forest) had the most pfap2mu
mutations (93%) followed by Cape-Coast (coastal
4653 AAT to ATT N1551I
savannah) (86%) and Navrongo (guinea savannah) (88%).
4653 AAT to CTT N1551L Despite this observation, there is comparatively higher
Adams et al. Parasites & Vectors  (2018) 11:175 Page 10 of 12
Fig. 7 A sequence alignment of pfubp1 gene showing amino acid changes due to single nucleotide polymorphisms. The alignment was done
using pfubp1 reference sequence of the 3D7 strain (PF3D7_0104300). Mutations from codons 1483–1549 at nucleotide positions 4449–4647 is
shown. The gaps are a result of the insertions of amino acids which resulted in a frameshift. The gap between 1519E and 1520K are therefore a
result of a 6 amino acid insertion in C356, G001 and G025. The known mutations D1525E and E1528D are shown in the isolates N061 and N096
respectively. A frameshift mutation is observed in N084, after codon 1536 as a result of a deletion
diversity in the mutations in Navrongo isolates with 72 the spread of resistance genotypes [18]. Therefore, dis-
different SNPs. Most of the isolates (63%) had more than cussing our observations along the line of different trans-
one mutation. The high diversity of SNPs in Navrongo, mission intensities from the three distinct ecological zones
where there is intense seasonal transmission of malaria, is will be premature.
expected. For pfubp1, Navrongo had the most mutations Insertions/deletions (indels) identified in the Ghanaian
(96%), followed by Cape Coast (87%) and Begoro (86%). It isolates were similar to those observed in Burkina Faso,
is quite interesting to observe diversity of mutations from Kenya and the UK patients from Liberia and Uganda
both genes from the guinea savannah zone. However, it [7, 8]. The most common indels, observed in the isolates
must be emphasised that transmission intensity does not from the Ghanaian and the other African countries re-
affect the evolution of resistance but plays a major role in sulted from an insertion of AAT which resulted in an
Table 4 Genetic insertions in the pfubp1 gene of isolates from the study sites
Sample ID Amino acids Insertion Nucleotide position
C315 KYE AAA TAT GAA 4576 to 4584
C337 KYE AAA TAT GAA 4583 to 4591
C344 KYE AAA TAT GAA 4548 to 4556
C356 KYEKYE AAA TAT GAA AAA TAT GAA 4588 to 4605
C360 KYE AAA TAT GAA 4583 to 4591
C321 EKY GAA AAA TAT 4590 to 4598
G001 DKYDKY GAC AAA TAT GAC AAA TAT 4563 to 4580
G018 EKY GAA AAA TAT 4598 to 4606
G025 KYDKYE AAA TAT GAC AAA TAT GAA 4585 to 4602
G026 EKY GAA AAA TAT 4580 to 4588
G034 YDKYDK TAT GAC AAA TAT GAC AAA 4579 to 4596
N023 EKY GAA AAA TAT 4582 to 4590
N031 EKY GAA AAA TAT 4582 to 4590
N050 KNE AAA AAC GAA 4549 to 4557
N111 YEK TAT GAA AAA 4597 to 4605
Adams et al. Parasites & Vectors  (2018) 11:175 Page 11 of 12
Table 5 Haplotypes of the pfap2mu and pfubp1 mutations in Conclusions
Ghanaian isolates The identification of molecular markers of ART resistance
Gene Haplotype No. of isolates in Ghanaian isolates has implications for the development
pfap2mu V161K-D168E 7 of ACT resistance especially for AL use. The findings from
V161E-R188R 6 this study give first-hand information on potential mo-
lecular markers of ART resistance in Ghana and as such
D168E-R188R 6
highlight the possibility of circulating parasites with re-
pfubp1 I1487I-N1488D 2 duced susceptibility to ACTs in use. Further investigations
I1487I-N1490I 2 are underway to ascertain their contribution to the less
Y1549L-I1552M 3 than 100% cure rates observed, particularly for AL.
N1540H-K1541N 2
N1540H-H1544P 3 Additional file
N1540H-N1542D 2
N1488D-D1525H 2 Additional file 1: Table S1. The list of all observed pfap2mu SNPs
in Ghanaian isolates from the three study sites. Table S2. The list of
I1487I-Y1501F 2 pfap2mu SNPs from five isolates with at least 20 SNPs per sample.
Table S3. The list of all observed pfubp1 SNPs in Ghanaian isolates
N1490I-Y1501F 2
from the three study sites. (XLSX 41 kb)
N1560I-L1563A 3
E1528D-N1560K 2 Abbreviations
ACT: artemisinin-based combination therapy; AL: artemether-lumifantrine;
Y1549L-N1560I 3
ART: artemisinin; AS-AQ: artesunate-amodiaquine; DUB: deubiquitinating
Y1549L-I1552M 2 enzyme; IRB: Institutional Review Board; IRS: indoor residual spraying;
ITNS: insecticide-treated mosquito nets; NMCP: National Malaria Control
Y1501W-Y1549L 2 Programme; nPCR: nested polymerase chain reaction; NS: non-synonymous
N1542D-Q1543H 2 mutations; pfap2mu: P. falciparum adaptor protein complex 2 mu subunit
gene; pfk13: P. falciparum kelch propeller domain on chromosome 13;
N1488D-G1492G 2 pfubp1: P. falciparum ubiquitin specific protease 1 gene; SEA: Southeast Asia;
Q1543H-P1547P 2 SNPs: single nucleotide polymorphisms; SYN: synonymous mutations;
WHO: World Health Organization
I1487I-N1490I-H1544L 2
N1490I-Y1501F-Y1533Y-P1547P-K1554N- 2 Acknowledgements
N1555I-D1557N The research work presented was funded by the Global Fund to fight Aids,
Tuberculosis and Malaria (GFATM)/ National Malaria Control Programme
(NMCP, Ghana). The funding covered the field work for the collection of the
samples and the laboratory analysis. The authors wish to thank the Director of
asparagine and lysine residues at codons 226 and 233, re- NMIMR, Professor Kwabena Bosompem, for his permission to publish this article.
spectively [7, 8]. Indels of one to six amino acids were ob-
served for the pfubp1 gene in the Ghanaian isolates and FundingThis study was funded by the Global Fund to Fight Aids, Tuberculosis and
these were similar to those seen in Kenya and Burkina Malaria (GFATM) and the National Malaria Control Programme (NMCP, Ghana).
Faso [7, 8]. These resulted from an insertion of a KYD, The funders stated played no role in the design of the study, sample collection,
KYE or KNE amino acids at codons ranging from 1516 to analysis and interpretation of results as well as manuscript preparation.
1535. Most of the indels that resulted in a frameshift were Availability of data and materials
caused by a deletion or insertion of either a guanosine or The observed SNPs from the sequence data from which the conclusions of
thymidine nucleotide. this manuscript were drawn are available as Additional file 1: Table S1. The
list of all observed pfap2mu SNPs in Ghanaian isolates from the three study
The presence of the molecular markers implicated in sites. Table S2. The list of pfap2mu SNPs from five isolates with at least 20
drug resistance is of great importance in their role in SNPs per sample. Table S3. The list of all observed pfubp1 SNPs in Ghanaian
modulating drug susceptibility and subsequently the pre- isolates from the three study sites.
diction of the dynamics of resistance [23]. The spread of Authors’ contributions
P. falciparum resistance to ARTs is a real global challenge NOD, NBQ, GF, BA and KAK conceived and designed the study. TA, NAAE
and therefore insights into the mechanisms of drug action and SM did the laboratory analysis and generated molecular data. TA, NAAE
and OCKH conducted data analysis. TA and NAAE drafted the manuscript. All
and resistance are critical for early detection of resistance. authors read and approved the final manuscript.
The detection and characterisation of these mutations in
post-treatment Ghanaian isolates is the way forward for Ethics approval and consent to participate
Ethical approval for the study was given by the NMIMR Institutional Review
further validation of their roles in conferring resistance. Board (IRB). The samples were taken after parents or guardians of the children
This study has, therefore, highlighted the existence of de- gave their consent. This work is part of an ongoing surveillance of antimalarial
layed clearance markers of the pfap2mu and pfubp1 genes drug efficacy studies in Ghana approved by the NMIMR IRB (CPN032/05-06a).
in circulating parasites in Ghana and will serve as baseline Consent for publication
data for future surveillance studies. Not applicable.
Adams et al. Parasites & Vectors  (2018) 11:175 Page 12 of 12
Competing interests 16. Wilson CM, Volkman SK, Thaithong S, Martin RK, Kyle DE, Milhous WK, et al.
The authors declare that they have no competing interests. Amplification of pfmdr 1 associated with mefloquine and halofantrine
resistance in Plasmodium falciparum from Thailand. Mol Biochem Parasitol.
1993;57(1):151–60.
Publisher’s Note 17. Golassa L, Kamugisha E, Ishengoma DS, Baraka V, Shayo A, Baliraine FN, et al.
Springer Nature remains neutral with regard to jurisdictional claims in published Identification of large variation in pfcrt, pfmdr-1 and pfubp-1 markers in
maps and institutional affiliations. Plasmodium falciparum isolates from Ethiopia and Tanzania. Malar J. 2015;14:264.
18. Hastings IM, Watkins WM, White NJ. The evolution of drug-resistant malaria:
Author details
1 the role of drug elimination half-life. Phil Trans R Soc Lond B Biol Sci. 2002;Department of Animal Biology and Conservation Science, School of 357(1420):505–19.
Biological Sciences, College of Basic and Allied Sciences, University of Ghana,
2 19. Ng PC, Henikoff S. Predicting deleterious amino acid substitutions. GenomAccra, Ghana. Department of Epidemiology, Noguchi Memorial Institute for Res. 2001;11(5):863–74.
Medical Research, College of Health Sciences, University of Ghana, Accra,
3 20. Kumar P, Henikoff S, Ng PC. Predicting the effects of coding non-synonymousGhana. Centre for Tropical Clinical Pharmacology and Therapeutics, School variants on protein function using the SIFT algorithm. Nature Protoc. 2009;4(7):
of Medicine and Dentistry, College of Health Sciences, University of Ghana, 1073–81.
Accra, Ghana. 4Department of Biochemistry, Cell and Molecular Biology, 21. Kimchi-Sarfaty C, Oh JM, Kim IW, Sauna ZE, Calcagno AM, Ambudkar SV. A
School of Biological Sciences, College of Basic and Allied Sciences, University "silent" polymorphism in the MDR1 gene changes substrate specificity.
of Ghana, Accra, Ghana. Science. 2007;315(5811):525–8.
22. Borrmann S, Straimer J, Mwai L, Abdi A, Rippert A, Okombo J, et al.
Received: 23 November 2017 Accepted: 1 March 2018 Genome-wide screen identifies new candidate genes associated with artemisinin
susceptibility in Plasmodium falciparum in Kenya. Sci Rep. 2013;3:3318.
23. Kwansa-Bentum B, Ayi I, Suzuki T, Otchere J, Kumagai T, Anyan WK, et al.
References Plasmodium falciparum isolates from southern Ghana exhibit polymorphisms
1. WHO. Malaria Fact Sheet 2017. http://www.who.int/mediacentre/factsheets/ in the SERCA-type PfATPase6 though sensitive to artesunate in vitro. Malar J.
fs094/en/. Accessed 08 June 2017. 2011;10:187.
2. WHO. Malaria Fact Sheet: World Malaria Day 2016. www.who.int/malaria/
media/world-malaria-day-2016/en. Accessed 22 Nov 2016.
3. Ariey F, Witkowski B, Amaratunga C, Beghain J, Langlois AC, Khim N, et al. A
molecular marker of artemisinin-resistant Plasmodium falciparum malaria.
Nature. 2014;505(7481):50–5.
4. MalariaGen: MalariaGen Plasmodium falciparum Community Project. Genomic
epidemiology of artemisinin resistant malaria. eLife. 2016;5:e08714.
5. Ouattara A, Kone A, Adams M, Fofana B, Maiga AW, Hampton S, et al.
Polymorphisms in the K13-propeller gene in artemisinin-susceptible
Plasmodium falciparum parasites from Bougoula-Hameau and Bandiagara,
Mali. Am J Trop Med Hyg. 2015;92(6):1202–6.
6. Kamau E, Campino S, Amenga-Etego L, Drury E, Ishengoma D, Johnson K, et
al. K13-propeller polymorphisms in Plasmodium falciparum parasites from
sub-Saharan Africa. J Infect Dis. 2015;211(8):1352–5.
7. Henriques G, Hallett RL, Beshir KB, Gadalla NB, Johnson RE, Burrow R, et al.
Directional selection at the pfmdr1, pfcrt, pfubp1, and pfap2mu loci of
Plasmodium falciparum in Kenyan children treated with ACT. J Infect Dis.
2014;210(12):2001–8.
8. Sutherland CJ, Lansdell P, Sanders M, Muwanguzi J, van Schalkwyk DA,
Kaur H, et al. Pfk13-independent treatment failure in four imported cases of
Plasmodium falciparum malaria treated with artemether-lumefantrine in the
United Kingdom. Antimicrob Agent Chemother. 2017;61(3):e02382–16.
9. Henriques G, van Schalkwyk DA, Burrow R, Warhurst DC, Thompson E, Baker
DA, et al. The Mu subunit of Plasmodium falciparum clathrin-associated
adaptor protein 2 modulates in vitro parasite response to artemisinin and
quinine. Antimicrob Agent Chemother. 2015;59(5):2540–7.
10. Wilkinson KD. The discovery of ubiquitin-dependent proteolysis. Proc Natl
Acad Sci USA. 2005;102(43):15280–2.
11. Henriques G, Martinelli A, Rodrigues L, Modrzynska K, Fawcett R, Houston
DR, et al. Artemisinin resistance in rodent malaria - mutation in the AP2
adaptor mu-chain suggests involvement of endocytosis and membrane
protein trafficking. Malar J. 2013;12:118.
12. Abuaku B, Duah N, Quaye L, Quashie N, Malm K, Bart-Plange C, et al.
Therapeutic efficacy of artesunate-amodiaquine and artemether- Submit your next manuscript to BioMed Central 
lumefantrine combinations in the treatment of uncomplicated malaria in and we will help you at every step:
two ecological zones in Ghana. Malar J. 2016;15(1):6.
13. Duah NO, Matrevi SA, de Souza DK, Binnah DD, Tamakloe MM, Opoku VS, et •  We accept pre-submission inquiries 
al. Increased pfmdr1 gene copy number and the decline in pfcrt and •  Our selector tool helps you to find the most relevant journal
pfmdr1 resistance alleles in Ghanaian Plasmodium falciparum isolates after •  We provide round the clock customer support 
the change of anti-malarial drug treatment policy. Malar J. 2013;12:377.
14. Price RN, Cassar C, Brockman A, Duraisingh M, van Vugt M, White NJ, et al. •  Convenient online submission
The pfmdr1 gene is associated with a multidrug-resistant phenotype in •  Thorough peer review
Plasmodium falciparum from the western border of Thailand. Antimicrob •  Inclusion in PubMed and all major indexing services 
Agent Chemother. 1999;43(12):2943–9.
15. Price RN, Uhlemann AC, Brockman A, McGready R, Ashley E, Phaipun L, et •  Maximum visibility for your research
al. Mefloquine resistance in Plasmodium falciparum and increased pfmdr1
gene copy number. Lancet. 2004;364(9432):438–47. Submit your manuscript at
www.biomedcentral.com/submit