Real-time PCR assay for discrimination of Plasmodium ovale curtisi and Plasmodium ovale wallikeri in the Ivory Coast and in the Comoros Islands

Background Plasmodium ovale is one of the five malaria species infecting humans. Recent data have shown that the name of this neglected species masks two distinct genotypes also called curtisi and wallikeri. Some authors show that these species could be sympatric. These two subspecies are not differentiated by microscopy techniques and malaria rapid diagnostic tests. This diagnostic defect is the result of low parasitaemia, antigenic polymorphism and absence of antibodies performance and requires the use of sequencing techniques. An accurate and easy discrimination detection method is necessary. Methods A new molecular assay was developed to easily identify the two genotypes of P. ovale. This tool allowed the study of 90 blood samples containing P. ovale, confirmed by molecular biology techniques, which were obtained from patients with imported malaria. Results The new marker was validated on well genotyped samples. The genotype of 90 P. ovale samples mainly imported from the Ivory Coast and the Comoros Islands was easily and quickly realized. The distribution of the two subspecies was described with a significant number of samples and showed that the two genotypes were present in the studied countries. Conclusion This work confirms the presence of the two species in the same country for the first time, in the Ivory Coast and the Comoros Islands. A better genotyping of P. ovale types may improve a better characterization of the clinical pathophysiology for each.

Plasmodium ovale diagnosis is difficult for several reasons, particularly due to a frequent low level of parasitaemia [8], and the errors of microscopic diagnosis with other malarial species. The low level of parasitaemia is a disadvantage for the RDT detection. The lack of sensitivity and specificity of rapid diagnostic tests (RDT), as no specific monoclonal antibody yet available leads to poor P. ovale identification. Furthermore, the Pan signal is often ineffective for its detection. The failures of P. ovale detection by RDT are well documented [9,10]. Genetic variations based on P. ovale LDH polymorphism could be involved in RDTs failure [11]. To overcome this difficulty, diagnostic methods using molecular biology techniques have been developed. However, few target genes are known because the total sequence of the P. ovale genome is not yet available.
Species-specific sequences within genes encoding small subunit ribosomal RNA (SSU rRNA) have been used as primers and have allowed discrimination between different malaria parasites through sequencing [12]. Different sets of primers were successively designed to improve the diagnosis, which illustrated the problem caused by the two genotypes for the diagnostic of the P. ovale subspecies [13]. Rougemont et al. developed a real-time PCR (RT-PCR) assay, which can separately distinguish the four major Plasmodium species [14]. Recently, another RT-PCR method was developed to identify the rare species from a mixed infection [15,16].
Complementary molecular investigations allowed the identification of a polymorphism within the P. ovale sequences. Species identification by 18S RNA sequencing split the samples into two groups, classic/variant type respectively named Plasmodium ovale curtisi and Plasmodium ovale wallikeri by Sutherland et al. [4,17,18]. Other investigations have confirmed that by the sequencing of nine other target genes [4,11,[18][19][20][21][22] whose genes encoding the lactate dehydrogenase (LDH) and the ookinete surface protein. Further studies with a conventional PCR or SYBR-Green real-time PCR method confirm the existence of the simultaneous presence of two subspecies in Africa [21].
A molecular marker based on hydrolysis probe technology was developed to quickly distinguish the two genotypes. This technique was used to study samples from countries for which less information were available, the Ivory Coast and the Comoros Islands, and the link between RDTs failure and genotyping group was studied.

Samples
Plasmodium ovale samples were selected among previously characterized blood samples collected from 2005 to 2010 from patients hospitalized in France and infected with imported malaria. The samples, collected in EDTA or ACD tubes, were stored at 4°C and were received within 24 to 48 hours. Thin blood films and RDTs were immediately performed. After centrifugation, sera and the red blood cells pellet were separated and stored at −20°C until analysis. For each sample, identification of the parasite species was performed by microscopy, immunochromatography and molecular biology. A total of 90 samples containing P. ovale parasites were retained for further study.

Species detection
Thin blood films were immediately stained by Giemsa to identify parasite species. They were read at least 20 min before to consider the smears as negative. RDT was performed on all the symptomatic samples received. These tests allowed Plasmodium species to be quickly identified. Normally all tests used should detect P. ovale species in a pan-species result, according to the manufacturer's instruction for use. From 2005 to 2008, Now malaria W (Binax, UK) was used, replaced by Core Malaria W (Core Diagnostics Ltd, UK) or Palutop W (Alldiag, Strasbourg, France) since 2009. These RDTs detect ALDOLASE and LDH antigen respectively in the P. ovale samples. All these tests were performed according to the manufacturers' instructions except for blood pipetting with a micropipette (10 μl) for more reproducible results.

Molecular Plasmodium species detection
Parasite genomic DNA templates were isolated from 200 μl of frozen red blood cells pellet using the DNA isolation KIT (QIA amp DNA mini kit, QIAgen, Hilden), according to the manufacturer's instructions. The Plasmodium species detection of P. ovale was performed with primers Pos25S and Pos25R [GenBank:AB074973] ( Table 1) using SYBR green qPCR as previously described [15]. Since 2010, malaria species samples were determined by real-time PCR using TaqMan probes and a set of primers specific to the same target genes ( Table 1).

Plasmodium ovale subspecies detection
The detection was carried out by sequencing the SSU rRNA and/or Po-ldh. The primers P1F-Up and 18S R ( Table 1) from Win [19] bound the SSU rRNA sequence.  Table 1) [23].
The real-time TaqMan PCR (TqPCR) assays were performed in a final volume of 20 μl containing 5 μl of DNA, 1X of master mix (Light Cycler TaqMan master, Roche, Germany), 0.1 μM of probe and 0.8 μM of each primer (Eurogentec, Belgium). The cycle conditions used were as follows: one step at 95°C for 10 min, 45 cycles at 95°C for 10 sec and 60°C for 30 sec, and a final step at 40°C for 30 sec, with a constant ramp rate at 20°C/sec. Fluorescence was read at the end of each cycle on Light-Cycler W 2.0 device (Roche Applied Science, Mannheim, Germany).
A sequence of the SSU rRNA gene was also used to design a pair of primers named PoF and PoR and two probes (pPOC and pPOW) ( Table 1) for the TqPCR analysis (Table 2). A Clustal [24] alignment was made on sequences from "malmai" [seq X99790] and "poc1" [AB182489] for P. o. curtisi and P. o. wallikeri forms, respectively [19]. Each probe is specific to one of the two P. ovale types and displays a distinct HEX or FAM fluorochromes, with readings at 530 and 560 nm for pPOW and pPOC, respectively. The primers, with a Tm of 60°C, were designed by Primer Express W software v2.0, and the probes, with a Tm of 10 degrees higher according to the TaqMan technology specification, were manually designed. The probes were designed to target a deletion at position 1158 in variant SSU rRNA sequence [19]. The TqPCR amplified fragment is 118 bases. The pPOW probe contains the deletion, present in all sequences of this type and the pPOC probe was designed in the same place in the classic sequence.
As positive control, DNA from well characterized samples of each P. ovale type was used. The efficiency of each reaction was assessed using 10-fold dilutions of the  positive control (from 1/10 to 1/1000) in triplicate and the slope of standard curves generated by plotting graphs of genomic DNA concentrations vs Ct values was determined. The specificity was verified by using the four other human malaria species DNA (P. falciparum, P. vivax, Plasmodium malariae and Plasmodium knowlesi) and human DNA as target template.

Samples
The selection of blood samples from imported cases of malaria patients mainly from the Ivory Coast and the Comoros Islands, has allowed the study of 90 P. ovale DNA samples. Species identification was previously performed by microscopy observations and RDTs. The molecular method was used to confirm the species diagnostic. However, these methods were not able to distinguish the P. ovale subtypes. This identification is possible by sequencing the 18S RNA gene or several coding sequences previously published, but the delay to obtain the result is long [19].

Marker set-up
This RT-PCR was set up on two reference samples, defined as "classic" (Poc) and "variant" (Pov), according to both LDH and SSU rRNA gene sequencing. The SSU rRNA sequence polymorphisms are broadly used references [12]. The P. ovale LDH (PoLDH) gene was selected since the corresponding protein is detected in RDT diagnosis and genetic variation in this protein was suggested as a possible cause for RDT failure. The efficiency of primers (named PoF and PoR) and probes set (named pPOC for P. o. curtisi and pPOW for P. o. wallikeri) has been determined in a specific assay. The mean curve slope and coefficient correlation were respectively −3.18 and 0.997 for the P. o. wallikeri type, with corresponding mean reaction efficiency of 99.71% and for the P. o. curtisi form, the mean curve slope and coefficient of correlation were −3.28 and 0.997, respectively, with a corresponding mean reaction efficiency of 99.77%. The sensitivity was tested with parasitaemia dilutions of 0.0065% to 0.000065% (Ct of 27.16 and 33.35 respectively) for P. o. wallikeri, and 0, 01% to 0, 0001% (Ct of 27, 50 and 34, 00 respectively) for P. o. curtisi.
These new sets of primers and probes are very specific, and they did not cross with other malaria species or human DNA.

Samples genotyping
The validation of these markers was made on 62 P. ovale samples, previously characterized either their SSU rRNA or ldh gene sequence ( Table 2). The sequencing of the central part of the ldh gene distinguished the two subspecies thanks to the substitutions on the codon positions S143P, 1N68K, previously describes, and I204V for P. o. curtisi and P. o. wallikeri: respectively (Additional file 1) [11,20].
Among the 62 samples, the sequencing of the ldh and SSU rRNA genes has been done both for eight samples and the new marker validate these results ( Figure 1A and 1B). The "POCPOW" markers were used to identify the subspecies of remaining P. ovale samples. For all the samples studied, 31 were determined as P. o. wallikeri types and 59 as P. o. curtisi types (Additional file 2).

Co-existence of Plasmodium ovale populations
The determination of the genotype gave the opportunity to study the country repartition of the two P. ovale subspecies. The used samples came from different countries, especially from Africa, with a predominant recruitment from Ivory Coast and the Comoros Islands. In general terms, the P. o. curtisi form is more widespread than the P. o. wallikeri in studied sample. The P. o. curtisi form represented 62.8% of samples and 37.1% belonging to the P. o. wallikeri form ( Table 2 samples) whereas the majority of the samples from the Comoros Islands (11/18 samples) were of P. o. wallikeri form. This distribution is statistically different (Chi2 test; p = 0,031).

Parasitaemia level
Easy identification of the two subtypes allowed investigation of the behaviour of the two genotypes. The parasitaemia was established and validated in 63 samples by an experienced microscopist. It was shown to be low ranging from to 0 to 0.5% (Additional file 2).
The two population statistical analysis with Wilcoxon test showed no parasitaemia level difference (Chi2 test p = 0.3353).

RDT failure
Laboratory practice and literature reported that RDT failure is often observed with P. ovale samples. This RDT failure is currently hypothesized by the role of natural variability present in tested species, as well as by the low parasitaemia often encountered. It was observed a higher false negative rate for the P. o. curtisi than for the P. o. wallikeri (about 60% vs 43% respectively). However, due to the relatively small number of samples available for each RDT and lot used, the difference was not statistically significant. From these results, the three tests used have the same efficiency with ALDOLASE or LDH antigens detection. The false negative results rates for the detection of P. o. curtisi with RDTs using ALDO-LASE or LDH antigens were 60% and 63% respectively. For the detection of P. o. wallikeri, the rate of false negative results was 43% no matter the antigen used by the tests. The positive results rates for the detection of P. o. curtisi with RDTs using ALDOLASE or LDH antigens were 40% and 37% respectively and for P. o. wallikeri detection, the positive results rates was 57% for the 2 antigens used.
The other factor commonly associated with RDT failure is low parasitaemia level. In this study, there is no influence of parasitaemia on RDT failure (Chi2 test p = 0.4108). The result of the RDT based on LDH antigen detection seems to be less affected by the low parasitaemia level than those detecting the aldolase antigen (Chi2 test p = 0.9843 and Chi2 test p = 0.0314 without the outlayer data of 0.5% respectively).

Discussion
The single microscopic diagnostic had for a long time limited the description of human malaria to four species. Microscopic diagnostic is an insufficient method to accurately discriminate between the species [7,25], because of its limited resolution, the morphological similarity of certain Plasmodium species and, especially, the expertise of the microscopist. In the same way, P. o. curtisi and P. o. wallikeri were confused due to their morphology. Until recently, the molecular methods detected P. ovale without distinguishing the subspecies. The consequence is doubt in the epidemiological study due to a lack of tools in the rare species diagnostic, such as P. ovale and P. malariae, which are less known and in fact more difficult to be characterized [7]. A better understanding of the circulating malaria species in the environment could increase the success in the identification of the human infectious species. For that, molecular methods have been developed and now bring corresponding tools to specifically identify each species or subspecies.
In this study, a rapid method has been developed to distinguish the two P. ovale subspecies. A TaqMan technology was chosen because the probes provide greater specificity. This method could be used to identify a coinfection between P. o. curtisi and P. o. wallikeri. So, these TaqMan markers with standard conditions could be added to the set of other markers for the detection of all species in one run. This method was clearly capable of to distinguishing the two P. ovale forms.
Ten different genes have been used across the different studies to characterize the 216 P. ovale samples. [4,11,[18][19][20][21][22]. Two sequences were obtained for each gene and defined two groups of alleles without mixing. The lack of recombination between these two groups of gene supports the notion of separated species. This view is in agreement with the results obtained by the sequencing and the TqPCR.
In 2010, Sutherland described the distribution of the two forms according to country origin [4]. The authors demonstrated that P. ovale was divided into two nonrecombining sympatric species in Africa and Asia for 55 samples. To complete these data, the study of P. o. curtisi and P. o. wallikeri in Africa, especially in the Ivory Coast and in the Comoros Islands, was introduced. It complete and deepen the knowledge of the distribution of the two subspecies worldwide ( Table 3).
The meta-analysis of all the data available indicates that the two subspecies are present in several analysed countries. Nevertheless, more detailed studies must be carried out to specify this data. By extrapolation, two subpopulations are suspected to be present in all countries where P. ovale rages if the data was sufficient. Among all the samples included in the study, no co-infection has been detected according results of previous studies [4,21,22]. Recently only one mixed infection of P. o. curtisi and P. o. wallikeri was described [26].
One important question is the origin of this sympatry. To explain that, Oguike [21] put forward a hypothesis of the parasite's evolution, the blood polymorphism host group, or a different recognition of certain molecules involved in the parasite cycle. Another hypothesis should be that different Anopheles species are present in the same area but at different periods of time. The human malaria vectors being described and present in the Ivory Coast and Comoros Islands are the same mosquitoes species (Anopheles gambiae and Anopheles funestus) [27,28]. The correlation between the different seasons and the emergence of different malaria vectors should be investigated. This point cannot be explained in this study, because the number of samples is not large enough and too disparate for a given country and a given period of time. Furthermore, varied climatic zones exist within the same country. In order to enlighten this hypothesis, it would be necessary to collect the samples from the same village at the same time. That would allow the comparison of the subtype identification from different climatic periods. Conversely, the repartition of the P. ovale subspecies into the various populations of Anopheles has never been investigated. This marker would be a tool of choice for it.
The unique difference between the two subspecies was the parasitaemia level. Indeed in one study, P. o. wallikeri has been associated with higher levels of parasitaemia in humans [17,18]. In this study, the both parasitaemia averages are very low and similar. The low parasitaemia is often evoked to explain the false negative results in RDT. A recent statement on the eradication programme emphasizes the role of sensitive diagnostic and especially the role of RDT and its lack of performance for neglected malaria, which remains unavailable for the moment in the breakdown of the malaria effort. Nevertheless, the small cohorts of P. ovale samples and the different RDT used in the literature make the comparisons difficult between the different studies. The fact that the tests available poorly detect P. ovale infection can be the consequence of the genetic difference between the two populations. The RDT must be improved in the light of the information. In the meantime, the specific Tq RT-PCR developed and described here will strongly improve the determination and the characterization of parasitaemia according to the P. ovale subspecies.
The P. ovale species are considered as neglected. However, severe cases pertaining to acute respiratory symptoms have been reported [6,[29][30][31]. Other severe cases describe spleen rupture [32,33]. These cases have never been the subject of molecular investigation to identify the P. o. curtisi or the P. o. wallikeri subtype. Taking this parameter into account would be helpful for the knowledge of clinical survey differences possibly existing between the two strains, which could further indicate whether one of the two species is more aggressive. To verify this hypothesis it will be necessary to upgrade this genotyping tool onto a true molecular diagnostic to replace the former P. ovale marker.

Conclusion
The use of the specific primers and probes described here could be a very accurate, useful, easy and rapid method of genotyping two different P. ovale subspecies. This technique will allow a better understanding of their characteristics, their biology and their responsibility in the clinical symptomatology. Molecular diagnostics are therefore essential for this purpose.
This study, due to the analysis of a large number of samples, confirms the presence of two subspecies in the same country and increases the knowledge of the two