Detection of mutations associated with artemisinin resistance at k13-propeller gene and a near complete return of chloroquine susceptible falciparum malaria in Southeast of Tanzania

In Tanzania, chloroquine was replaced by sulphadoxine- pyrimethamine (SP) as a first-line for treatment of uncomplicated malaria. Due to high resistance in malaria parasites, SP lasted for only 5 years and by the end of 2006 it was replaced with the current artemisinin combination therapy. We therefore, set a study to determine the current genotypic mutations associated with Plasmodium falciparum resistance to artemisinin, partner drugs and chloroquine. Parasites DNA were extracted from dried blood spots collected by finger-prick from Tanzanian malaria infected patients. DNA were sequenced using MiSeq then genotypes were translated into drug resistance haplotypes at Wellcome Sanger Institute, UK. About 422 samples were successful sequenced for K13 gene (marker for artemisinin resistance), the wild type (WT) was found in 391 samples (92.7%) whereby 31 samples (7.3%) had mutations in K13 gene. Of 31 samples with mutations, one sample had R561H, a mutation that has been associated with delayed parasite clearance in Southeast Asia, another sample had A578S, a mutation not associated with artemisinin whilst 29 samples had K13 novel mutations. There were no mutations in PGB, EXO, P23_BP and PfMDR1 at position 86 and 1246 (markers for resistance in artemisinin partner drugs) but 270 samples (60.4%) had mutations at PfMDR1 Y184F. Additionally, genotyped PfCRT at positions 72–76 (major predictors for chroquine resistance), found WT gene in 443 out of 444 samples (99.8%). In conclusion, this study found mutations in K13-propeller gene and high prevalence of chloroquine susceptible P. falciparum in Southeast of Tanzania.

treatment of malaria (chloroquine and sulfadoxine-pyrimethamine) in nearly all endemic countries originated in almost similar fashion 8 .
The discovery of mutations in the propeller domain of the kelch (K13) gene were marked as candidate molecular markers and has remained to be the key predictor of ART resistance 9 . Several mutations at K13 are therefore associated with ART resistance 10 , i.e. K13 M476I mutation was first investigated in Tanzanian F32 parasites that were exposed in vitro to escalating concentrations of ART for more than 5 years 11,12 . In addition to that genomic analysis of Cambodian isolates identified four prevalent K13 mutations (Y493H, R539T, I543T and C580Y) that were associated with ART resistance 12,13 . In this regard, the list of K13 has kept on increasing and regularly updated by World Health Organization (WHO) 3 . Additionally, evidence from research reported parasite genetic background (PGB), is the mutations that allowed the emergence of K13 mutations, these mutations include; V127M and D128Y/H in the PfARPS10 (PF3D7_1460900) protein, D193Y in ferredoxin (PfFD, PF3D7_1318100), N326S and I356T in PfCRT (PF3D7_0709000), and T484I in PfMDR2 (PF3D7_1447900) and they are expressed as concatenated haplotype form (VDDNIT) as a reference allele (wild type) 14 .
Moreover, P. falciparum multi-drug resistance gene 1 (PfMDR1, PF3D7_0523000) and particularly, single nucleotide polymorphisms (SNPs) resulting in an amino acid change in codons 86 (N86Y), 184 (Y184F), and 1246 (D1246Y) have been associated with changes in parasite susceptibility to various drugs, including ACT. Mutations at position 86 and 1246 have been associated with parasite resistance to CQ and amodiaquine 15 while mutations at positions 86, 184, and 1246 increase susceptibility of mefloquine and lumefantrine 16 . Genome-wide association study (GWAS), a single nucleotide polymorphism (SNP) in a putative exonuclease gene (PfEXO, Pf3D7_1362500) was associated with an increased tolerance of piperaquine.
On the other hand, gene amplification of a section of chromosome 14 involving the genes plasmepsin 2 and plasmepsin 3, P2/3 breakpoint (P23_BP) has been associated with an increased resistance to piperaquine 17 . Additionally, CQ resistance transporter (PfCRT) and PfMDR1 both located on the food vacuole of the parasite involved in CQ resistance 18 . However, the CQ transporter PfCRT is a stronger predictor of CQ resistance than PfMDR1 19 . Nevertheless, in areas where usage has been strictly regulated, withdrawal of CQ have resulted in dramatic decreases in the prevalence of CQ-resistant parasites 4,20,21 .
Therefore, it was necessary to conduct a molecular surveillance of gene mutations associated with P. falciparum resistance to artemisinin, partner drugs and chloroquine, one decade since ART-based combination therapy was introduced as a first line treatment and 18 years after CQ withdrawal for treatment of uncomplicated malaria in Tanzania 4 .

kelch13 (K13) propeller polymorphisms.
A total of 489 samples were genotyped but 67 (13.7%) could not be detected/missing genotypes. Of 422 successful sequenced samples, wild type (WT) K13 gene was found in 391 samples (92.7%) whereby 31 samples (7.3%) had mutations in K13 gene. Of those mutations, one sample had R561H, a mutation that has been associated with delayed parasite clearance, another sample had A578S, a mutation not associated with clinical or in vitro resistance to artemisinin whilst 24 nonsynonymous mutations are not yet listed (uncharacterized) in WHO artemisinin resistance report of 2018 (Table 1).
prevalence PGB, EXO/P23_BP and PfMDR1 mutations. Generally, 489 samples were genotyped but the total number per every marker differed based on the number genotypes could be detected. All genotypes for PGB and EXO/P23_BP, markers for ART and piperaquine resistance respectively were WT. There were 270 samples (60%) with mutations at PfMDR1 Y184F; a marker for lumefantrine, amodiaquine and mefloquine drug ( Table 2).  Table 3). The PfMDR1 mutation at position N86Y is the first in the 3 amino-acid haplotype (NYD) which enhances resistance to CQ was not detected in the all samples (100%) ( Table 2).

Discussion
To the best of authors' knowledge this is the first study to report mutations at K13-propeller, associated with ART resistance from the southeast of Tanzania. We report resistance patterns, one decade since ART-based combination therapy was introduced as the first and second line treatment of malaria whilst 18 years after CQ withdrawal for treatment of uncomplicated malaria in Tanzania. Generally, this study found a prevalence of 7.3% for K13 mutations, these mutations contained those which are found in WHO list 3 , and those reported elsewhere 10 and undocumented non-synonymous K13 mutations. Prevalence obtained in this study was higher than the one reported the polymorphisms of P. falciparum K13-propeller gene among migrant workers returning to Henan Province, China from Africa, the study found the frequency of the K13-propeller 6.50% in Central Africa, followed by East Africa (5.26%), West Africa (4.55%) and South Africa (4.55%) 6 and study by Kamau et al., 2014 22 reported that, allele frequencies of K13-propeller polymorphisms in P. falciparum parasites from sub-Saharan Africa ranged between 1% and 3%. In contrary to that, the recent findings from Uganda 23 , Kenya 24 as well as Tanzania 25 reported no evidence of K13 mutations. The differences in study periods between the studies could be the cause of the observed discrepancies in prevalence.
K13 mutations at position R561H and A578S, are two mutations previously described by WHO as validated 3 and not associated 26 with ART resistance, respectively. Surprisingly, both two mutations were detected as recombinant which contained WT gene, i.e. WT R561H and WT A578S WT W565C. These mutations especially, the validated mutation R561H cause a delayed parasite clearance 10,27 . Mutation in K13 gene has also been reported from the study conducted in southern Rwanda 28 . More importantly, this study documented 24 K13 mutations which currently don't form part of WHO ART resistance markers list of 2018 3 . On the other hand, all genotypes for PGB and EXO/P23_BP, markers for ART and piperaquine resistance respectively found no mutation.
High prevalence (60.4%) PfMDR1Y184F mutations was detected from Tanzanian samples. These finding are similar from the study conducted in Equatorial Guinea which found high prevalence of PfMDR1 Y184F mutations in P. falciparum isolates. Nevertheless, there are limited evidences which associated the PfMDR1Y184F mutation with lumefantrine and mefloquine 16 susceptibility. Furthermore, the study reported that PfMDR1 at amino acids 86 and 184, demonstrate resistance to the ACT partner drug amodiaquine and the former first-line agent CQ. In contrast, N86Y increases parasite susceptibility to the partner drugs lumefantrine and mefloquine, and the active artemisinin metabolite dihydroartemisinin. The PfMDR1 N86Y plus Y184F isoform moderately    www.nature.com/scientificreports www.nature.com/scientificreports/ reduces piperaquine potency in strains expressing an Asian/African variant of the CQ resistance transporter PfCRT 16 . On the other hand, the current findings suggest that CQ-susceptible P. falciparum parasites have reemerged and are now predominant in Tanzania (sub-Saharan Africa) where CQ was withdrawn in 2001. These findings were similar from the recent study conducted in Zambia 29 , the neighboring country to southwest. In both two countries CQ were withdrawn as a first line treatment in 2001 4 and 2003 21 in Tanzania and Zambia, respectively. These findings were contrary to the country where high levels of CQ resistance have persisted due to incomplete withdrawal of CQ 30 . The study of 2019, conducted in Nigeria revealed a high prevalence of PfCRT mutant genotypes and haplotypes and low frequency of PfMDR1 mutant genotypes, 11 years after the switch in malaria treatment policy from CQ to artemisinin combination therapy (ACT) in Nnewi, Nigeria. The study suggested that continual circulation and spread of CQ-resistant P. falciparum parasites in the study area due to the continued use of unrecommended CQ 30 .
Moreover, this study found a complete deletion of PfMDR1 mutation in all samples at position N86Y, a mutation which is associated with an enhanced resistance to chloroquine. These findings are in line with those reported in a study conducted in Zambian 29 . Another study conducted in Malawi reported a slower decline in prevalence of mutations in PfMDR1 than PfCRT suggesting that PfMDR1 mutations may be less deleterious to parasite fitness than are PfCRT mutations 31 . However, mutations in PfMDR1 by themselves are insufficient to confer CQ-resistance 32 also the combination of PfMDR1 mutations and PfCRT mutations, provided no added advantage to CQ treatment failure than PfCRT mutations alone 33 . Furthermore, PfMDR1 mutations do not add to the predictive value of PfCRT mutations for CQ treatment failure 34 .
Since the design of the current study was cross sectional, there was a limitation in establishing the association between the molecular resistance markers and clinical/treatment outcomes of patients. However, the relationship between malaria treatment outcomes and resistance markers have been well described elsewhere 3,10 .
In conclusion, K13-propeller mutations associated with artemisinin resistance were found in Tanzanian samples. Mutations included K13 R561H and K13 A578S haplotypes validated to cause artemisinin resistance and not associated with artemisinin resistance, respectively. Twenty four K13 non-synonymous mutations not yet listed by WHO (unclassified) are reported. Further phenotypic studies are warranted to investigate the unclassified K13 mutations. Additionally, the return of chloroquine-susceptible P. falciparum malaria, 18 years after the removal of chloroquine drug pressure in Tanzania is documented. In this regard, chloroquine may be considered for malaria prevention, i.e. sickle cell disease children or the reintroduction in future, ideally in combination with other antimalarial drugs, especially in areas where disappearance of chloroquine resistance is evident while safe and affordable alternatives antimalarials are limited.

Materials and Methods
Study design, area, period and population. Surveillance of molecular markers for ART and partner drugs (piperaquine, lumefantrine, amodiaquine, mefloquine and chloroquine) resistance was conducted between April and August 2019 at Kibiti Health Center (KHC), Kibiti District, Tanzania (Fig. 1). Kibiti District is found along the Indian ocean 35 and has malaria prevalence of 10.2% 36 where P. falciparum is responsible for more than 95% of all malaria cases 37 . Patients attending clinic at Kibiti Health Center (KHC) who presented with symptoms suggestive of malaria infection were recruited in the study. The symptoms such as fever, general body weakness and headache were confirmed by the attending physician 38 . Patients screened for malaria and those who tested positive using CareStart Malaria HRP2/pLDH test (Access Bio, Ethiopia) were requested to participate. Then positive samples by rapid tests were subjected to blood smear (BS) microscopy for confirmation. A total of 489 dried blood samples (DBS) from patients tested positive with BS microscopy were subjected to DNA extraction and genotyping.
DBS preparation. DBS were prepared according to MalariaGEN SpotMalaria, DBS collection protocol 39 . A sterilized patient's finger was pricked to allow blood drops where four blood spots from each patient were prepared, two on each paper card. The blood spots were allowed to air dry and placed in the desiccant sachet for storage.   www.nature.com/scientificreports www.nature.com/scientificreports/ DNA extraction. DNA from the DBS was extracted following QIAamp DNA Investigator Kit for isolation of total DNA from filter papers (Qiagen, Valencia, CA, USA) and as previously described by Oyola et al. 40 .
Genotyping of antimalarial resistance markers. Molecular genotyping of ART, partner drugs and chloroquine (K13, PGB, EXO, P23_BP, MDR1 and CRT) were performed by Wellcome Sanger Institute, UK.
Briefly, targets for genotyping were identified and multiplex PCR primers were designed using a modified version of the mPrimer software 41 and the exact design of the primer sequences will be described elsewhere (Goncalves, manuscript in preparation). Primers were designed to amplify products between 190-250 bp and were combined into 3 pools. A two-step protocol was used to first amplify the target regions of the parasite genome, followed by a second PCR to incorporate sequencing and multiplexing adapters. Batched samples (384) were sequenced in a single MiSeq lane combining all PCR products. Samples were de-plexed using the multiplexing adapters and individual CRAM files were aligned to a modified amplicon reference genome. Genotyping was done using bcftools as well as custom scripts to filter and translate genotypes into drug resistance haplotypes.