Sequence analysis of Plasmodium falciparum SERCA-type pfATPase6 in field isolates collected along Thai-Cambodian border

Polymorphisms of the proposed artemisinin binding pocket of pfATPase6 were investigated via a sequencing procedure. The forward (F) and reverse (R) primers for the PCR amplification of the pfATPase6 gene were designed against the proposed artemisinin-derivative-binding pocket (M3, M5, and M7 helices) [22, 23] as shown in Figure 1 and the coding region around residue 769 [12] using the PyMOL Molecular Graphics System software (Schr dinger, New York, NY, USA) (available from www.pymol.org) to view the proposed catalytic site of the thapsigargin. The sequences of the forward and reverse primers are listed in Table 1. All of the PCR primers were analyzed using a BLAST search for a possibility of ambiguous amplification of other regions in the mammalian genome.

Figure 1

All 177 clinical samples (164 samples collected on D₀ and 13 samples collected on D_f) from the two sites, Tasanh (Cambodia) and Trat (Thailand) had DNA extracted using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany). Fifty picograms of the extracted DNA was used for each PCR reaction. Reactions were performed in a total volume of 50 μL using 2.5 units of AmpliTaq Gold DNA Polymerase (Thermo Fisher Scientific, Waltham, MA, USA), 200 μM of each of dNTP, 2 mM MgCl₂, and 1× PCR buffer II (Thermo Fisher Scientific). The coding regions around the proposed artemisinin binding pocket and the residue 769 and were amplified using the listed pairs of the primer sets (Table 1).

Cycling conditions were 5 min at 95°C, followed by 37 cycles of 1 min at 94°C, 1.5 min at the annealing temperature (45.5°C for the M3 helix amplification, 48.5°C for the 769 residue region and M7 helix amplifications and 47.5°C for the M5 amplification) and 2 min at 72°C. The reaction was then continued using an elongation step of 10 min at 72°C and chilled at 4°C. The amplified DNA fragments for M3, M5, M7, and the region around residue 769 were 151, 207, 141, and 202 bp, respectively.

The DNA purification and sequencing steps were conducted by Macrogen, Seoul, Korea. Briefly, the DNA was purified using a Millipore-Montage filtering kit (Millipore, Billerica, MA, USA). Sequencing reactions were conducted in an MJ Research PTC-225 Peltier Thermal Cycler (MJ Research, Waltham, MA, USA) using AmpliTaq DNA polymerase (Applied Biosystems, Foster City, CA, USA) and an ABI PRISM BigDye Terminator Cycle Sequencing Kit (Applied Biosystems). The conditions followed standard operating protocols supplied by the manufacturer. The samples were analyzed by the ABI 3730XL analyzer (Applied Biosystems). Sequences were analyzed by Sequencher, version 5.1 and MEGA, version 5.2 software (www.megasoftware.net/).

Treatment outcomes for the patients from whom clinical isolates were collected are described in Table 2. There were 4 clinical treatment failures out of 74 patients (5.4%) in the artesunate monotherapy arm conducted in Tasanh. None of the 37 patients in the quinine–tetracycline arm were clinical failures. Among patients taking an ACT in Trat, 9 of 53 patients failed therapy (17.0%). The follow-up periods in the 2 studies were 28 and 42 days, respectively.

The extracted parasite DNA from both cohorts was PCR amplified and the sequence was analyzed near the region of the M3 helix, M5 helix, M7 helix, and residue Ser769 (Figure 1). Some samples failed to be either amplified or sequenced (2 samples for the M3 helix; 2 samples for the residue 769; 6 samples for the M5 helix and 1 sample for the M7 helix) because of low quantity of the DNA provided. The DNA sequence around the M3 helix (175 sequenced samples), M7 helix (176 sequenced samples) and Ser769 (175 sequenced samples) from these isolates was identical to the same DNA regions from the reference P. falciparum 3D7 laboratory clone (Table 3).

No sequence alterations of Leu263 (which resides in the M3 helix) or Ser769 residues were detected. In addition, 46 samples out of a total of 171 sequenced samples (27%) contained a synonymous change (T→A) in the M5 helix (Table 3). This altered silent mutation (T→A) was found in the clinical specimens

collected from all 3 different treatments: 28/46 cases from artesunate monotherapy, 11/46 cases from quinine–tetracycline, and 7/46 cases from artesunate– mefloquine therapy. For the artesunate monotherapy, 24/28 for the ACPR, 2/28 for the treatment failure cases on D₀ and 2/28 from the same treatment cases but on D_f. The same silent mutations were found in both specimens on D₀ and D_f. No synonymous change (T’!A) in the M5 helix from the samples collected from patients failing the treatment from Trat was observed.

There were no polymorphisms of pfATPase6 detected in 13 samples collected on the dates of treatment failure and no nonsynonymous mutations found in the entire 177 samples investigated. The majority of DNA sequences coding for the proposed artesunate binding pocket (M3, M5, and M7 helices) and the regions around Ser769 were conserved in parasite populations collected from patients in both study sites, regardless of clinical outcome.

We attempted to sequence 177 samples; however, 2 samples for M3 helix, 2 samples for 769 residue area, 6 samples for M5 and 1 sample for M7 are unamplifiable (designated as “Unamp”). “n” represents numbers of the clinical samples. “Identical” indicates numbers of patients who share identical or conserved DNA sequence of specific regions of the pfATPase6 compared to the 3D7 clone. Fractions of altered sequence in the indicated regions of the pfATPase6 compared to the 3D7 clone. Sequence alteration observed in M5 helix is synonymous with a mutation or silent mutation, and labeled as “syn”.

The present study investigated nucleotide sequences of the proposed artemisinin binding pocket on pfATPase6 and the regions around the residue 769 in P. falciparum isolates collected from the entire cohort (177 clinical isolates) of two clinical studies (2005-2007) with three different treatment regimens and diverse clinical outcomes. Twenty-seven percent (46/171) of the amplified samples contained a synonymous change, suggesting no alterations in amino acids in the M5 helix of pfATPase6, whose relationship with artemisinin is unknown. These results suggest that the sequence of this specific region of pfATPase6 is either conserved, or contains synonymous change, regardless of the malaria drug treatments or clinical outcomes. No sequence alterations were observed in samples collected on the dates of treatment failure (D_f). Therefore, there appears to be no association between this specific region of the pfATPase6 and clinical outcome, suggesting that these particular polymorphisms of specific regions of the SERCA-type pfATPase are unlikely a useful molecular markers with which to monitor artesunate resistance in people living in the Thai–Cambodian border region.

These findings from sites along the Thai– Cambodian border are in agreement with other molecular epidemiology studies from Southeast Asia [24, 26-29]. All of these field investigations suggest no demonstrable association of pfATPase6 polymorphism with reduced artemisinin susceptibility. Price et al. found nonsynonymous mutations located between the M1 and M2 helices that were not associated with artesunate susceptibility [27]. Because M1/M2 helices are distantly located from the artemisinin binding site and are not associated with artesunate susceptibility, analysis of the M1/M2 regions were not investigated in the samples from the study described herein. In addition, Jambou et al. analyzed 24 samples of the full-length pfATPase6 and identified no polymorphisms in the complete gene of pfATPase6 in Southeast Asian clinical samples with a broad range of IC₅₀ values [24]. Imwong et al. reported sporadic changes in amino acids and reported no correlation with clinical outcomes from clinical specimens collected from northern Thailand and western Cambodia [28]. Furthermore, large GWAS studies in the western Cambodia borders by Miotto et al. and Takala-Harrison et al. demonstrated no association between parasite clearance time or treatment failure with pfATPase6 [30, 31]. Miao et al. investigated clinical samples from the Greater Mekong Subregion, and other regions of Asia, Africa, South America, and Oceania and found no significant signature of pfATPase6 polymorphism among populations indicating the challenge of using this as a marker for ART resistance [29].

Furthermore, several studies had demonstrated in diverse biological systems that pfATPase6 is not a target of artemisinin described in the review [32]. Cardi et al. [33] reported no effects of artemisinin on mammalian SERCA1a (pfATPase6 equivalent) E255L expressed in yeast and COS cells, in contradiction with the previous reports from the expression studies in Xenopus oocytes. Cui et al. and Valderramos et al. observed no changes in the sensitivity of the L263E and S769N mutations in P. falciparum to ART derivatives [34, 35]. Abiodun et al. showed that the in vitro interactions for artemisinin derivatives and thapsigargin were additive [36]. Arnou et al. was not able to demonstrate the inhibition of artemisinin or any binding activity of the drug to SERCA1a [37].

Thus, our findings are consistent with the results of these studies conducted in Southeast Asia, and from basic research studies, that there are no correlations between the sequence of pfATPase6 and clinical outcome. Even though the present study analyzed samples from patients with a variety of clinical outcomes (including cases classified as being artemisinin resistant) from an area of suspected artemisinin resistance, no differences in pfATPase6 sequences were detected between patients who failed and those who were treated successfully.

Despite the correlation between S769N (24), H243Y [38], and the other 13 haplotypes of pfATPase6 [39] and artesunate or artemether IC₅₀ values found in Africa, additional molecular surveillance studies in other regions of the world (Asia and South America) and a study of rodent malarial parasites [40] have all failed to demonstrate an association between pfATPase6 polymorphisms and artemisinin susceptibility. For example, clinical isolates from Peru were found to contain polymorphisms in several regions of pfATPase6; however, these mutations did not show clear correlation with the range of artesunate IC₅₀ s [41]. In addition, a study in artesunate-resistant rodent P. chabaudi showed no mutations in ATPase6 or other genes, such as mdr1, cg10, or tct [40]. Furthermore, Tanabe et al. [42] observed baseline spontaneous mutations in pfATPase6 from malaria parasites collected from Asia, South America, Oceania, and Africa despite these patients never being treated with artemisinins, suggesting that pfATPase6 polymorphisms likely arise from baseline spontaneous mutational events occurring even in the absence of drug exposure.

The present study of clinical isolates from Thailand and Cambodia detected no polymorphisms in the proposed artemisinin-binding-pocket of pfATPase6, a region where mutations were expected if artesunate interacted with this proposed binding pocket. The results from this study and others suggest that artesunate likely does not physically interact with this region of the protein. Another explanation could be that mutations at the proposed artemisinin binding site are lethal or drastically decrease the fitness of parasites. Therefore, mutations of these regions may not be found in natural settings, and it is likely that alterations in the artesunate binding site, such as L263E [22, 23], are generated only in vitro.

To date, no polymorphisms have been identified in other proteins proposed to be the target of artemisinin derivatives. For instance, the translationally controlled tumor protein homolog (pftctp), one of the proposed proteins targeted by artemisinin derivatives, was isolated biochemically from infected erythrocytes physically bound to tritium-labeled [10-³H] dihydroartemisinin [43]. However, when Jambou et al. sequenced the pftctp gene of clinical isolates from malaria patients in Cambodia, French Guiana, and Senegal, all areas at the time of the study practicing broad application of artemether use, they found that all isolates shared an identical nucleotide sequence with the P. falciparum 3D7 laboratory reference clone [24]. In addition, a study of rodent malaria P. chabaudi [40] identified no mutations of the homologous pftctp gene. Thus, there is no evidence of pftctp polymorphisms, even in parasites exposed to artemisinins in vivo.

The region around the proposed artemisinin binding pocket and the area around the 769 residue of pfATPase6 does not appear to be a useful molecular marker with which to monitor artemisinin resistance in parasites from clinical cases in Southeast Asia presenting in this region of known antimalarial drug resistance. By performing molecular surveillance in Southeast Asia and around the globe, we could prevent the spread of multidrug resistant P. falciparum.

Obtaining a molecular fingerprint linked to failure of artesunate treatment is an essential step to facilitate surveillance for emerging artemisinin resistance, and for monitoring elimination and/or containment strategies. Since conducting this investigation, discovery of other potential markers of artemisinin resistance, such as mutations in the P. falciparum DNA repair pathway associated with delayed parasite clearance in patients from Southeast Asia given artesunate [31], offers hope for identification of molecular markers capable of detecting cases of artemisinin-resistant malaria.