Malaria is one of the most important infectious diseases of humanity, and continues to cause significant mortality and morbidity worldwide. Early diagnosis is important for case management and treatment of the disease, and in guiding treatment for non-malaria fevers. Symptom-based clinical diagnosis is inaccurate, and contributes to poor management of febrile illness, over-treatment of malaria, and may promote drug resistance to current anti-malarials 1 .

Rapid diagnostic tests (RDTs) for malaria have the potential to improve case management and thereby reduce morbidity and mortality, especially in remote areas, facilitating the timely delivery of appropriate treatment. Indeed, many RDTs today can achieve excellent sensitivity and specificity for Plasmodium falciparum at a parasitaemia greater than 500 parasites per microlitre (parasites/μL) 2 . At lower parasitaemia, however, variability in sensitivity is more common 3 4 5 6 7 .

Today, over 150 malaria RDTs are commercially available, with most using a P. falciparum detecting component targeting P. falciparum histidine-rich protein 2 (PfHRP2). The gene encoding the protein, pfhrp2, is a single copy subtelomeric gene located on chromosome 7 encoding an amino acid sequence containing 34% histidine, 37% alanine and 10% aspartic acid 8 9 10 11 . PfHRP2 is characterized by multiple contiguous repeats of the sequences AHH and AHHAAD 8 9 . PfHRP2 is a 60-105 kD water-soluble protein specific to P. falciparum, synthesized and present throughout the asexual life cycle, identified as a surface-exposed protein in infected erythrocytes 8 9 10 12 13 14 15 16 17 18 . The protein is also found circulating in the peripheral blood of infected individuals 19 . These features make PfHRP2 a good target for diagnosis of P. falciparum infection.

Pfhrp3 encodes P. falciparum histidine-rich protein 3 (PfHRP3), also known as the small histidine-rich protein (SHARP), located near one end of chromosome 13 17 20 . Pfhrp3 shares many structural similarities with pfhrp2. Both genes have an interrupted structure and contain a signal peptide sequence in exon 1 followed by an intron. The intron is followed by the main coding region, exon 2. Exon 2 in both pfhrp2 and pfhrp3 encodes histidine-rich amino acid repeats beginning 75-90 nucleotides downstream from its start 18 . Although the histidine composition of PfHRP3 is slightly less than that of PfHRP2 (28% compared to 34%), both genes share many histidine and alanine rich repeats 10 . It has been suggested that, due to their similarity, both genes are related, derived from an ancestral duplication and interchromosomal divergence from a common ancestral gene, and may complement each other in function 8 10 17 18 . Antibodies against PfHRP2 cross-react with PfHRP3 8 18 . Thus, PfHRP3 also contributes to the detection of P. falciparum infections in PfHRP2-detecting malaria RDTs.

As part of the World Health Organization (WHO) and Foundation for Innovative and New Diagnostics (FIND) Malaria RDT Quality Assurance Programme, the levels of diversity for antigens targeted by malaria RDTs have been systematically investigated. While parasite aldolase and pLDH appear to be highly conserved 21 22 23 , pfhrp2 was found to be highly variable. In the preliminary analysis of 74 isolates from mostly Southwest Pacific and Asian countries, a significant sequence variation in pfhrp2 and pfhrp3 was observed in isolates within the same country and between different countries 24 . This raised a serious concern that the sequence variation could result in significant variation in the presence and frequency of epitopes recognized by monoclonal antibodies (MABs) and hence impact on the RDT detection sensitivities for different parasites. This concern was strengthened by a regression analysis based on 16 cultured parasite lines tested where the number of type 2 (AHHAHHAAD) and type 7 (AHHAAD) repeats in PfHRP2 were identified to be a contributing factor to the variable sensitivity reported at low level parasitaemia (below 250 parasites/μl) 24 25 . While this established extensive variation and a preliminary link between sequence diversity and RDT sensitivity within this sample set, it did not comprehensively cover diversity from all areas of global malaria transmission or testing sensitivity on a large set of samples with a wide range of RDT products.

The aims of the current study were to extend the diversity investigation to include isolates from African and South American countries, and to better understand the implication of global diversity in pfhrp2 and pfhrp3 on the performance of PfHRP2- detecting RDTs. In this paper, the global diversity of pfhrp2 and pfhrp3 was examined and the distribution of variants mapped. Furthermore, the results of the recently completed WHO product testing of malaria RDTs: Round 1 (2008) 26 were used to examine the effect of PfHRP2 sequence structure on RDT sensitivity.

PfHRP2 is the target antigen for many RDTs detecting P. falciparum. In this study, pfhrp2 from 458 P. falciparum isolates collected from 38 countries, all major malaria endemic areas, were sequenced and evaluated. It was shown that this gene exhibits extensive diversity both within and between countries and regions. A similar protein, PfHRP3, also exhibited extensive diversity between parasite isolates examined. This comprehensive examination and analysis of the extent of the diversity and geographic variation in these genes provides important information for laboratory and field evaluation of malaria RDTs. It also provides an indication of whether the genetic diversity in the antigen contributes to the variability in RDT sensitivity.

Both pfhrp2 and pfhrp3 are located in the subtelomeric regions of chromosomes. In general, genes located in telomeric and subtelomeric regions of Plasmodium have vast genetic diversity, and are highly susceptible to changes during recombination events 28 29 30 31 32 33 . Subtelomeric regions from different malaria species appear to have undergone rapid evolution, with significant sequence variation generated in the complex repeats in these regions 32 33 . Molecular mechanisms contributing to the generation of tandemly repeated regions, changes in the length of repeat blocks and other variation events include slipped strand mispairing followed by DNA replication or repair, unequal reciprocal combination and gene conversion 34 35 36 37 38 39 . The different organization and varying number of repeats observed in both pfhrp2 and pfhrp3 are likely the result of frequent recombination of the chromosomes. Therefore, a correlation between malaria transmission intensity and pfhrp2 diversity should be expected, as malaria infections in high transmission settings often involve co-infection of multiple strains which increase the probability of recombination during sexual reproduction in the mosquito vector. Indeed, this general trend was observed in this study with the ratio of different pfhrp2 sequence types to total sequences being higher in countries with high transmission intensity such as in Africa, and lower in South American and Asian countries. However, this may also reflect differences in the geographical spread of collections within countries, which varied between sites.

The function of PfHRP2 still remains to be determined. Early theories suggested that PfHRP2 may be involved in detoxification of free haem by converting it to inactive haemozoin 40 41 . Other theories suggest that PfHRP2 may be involved in remodeling the infected erythrocyte cytoskeleton 42 and in modulating host immune responses 43 . The extensive diversity in the pfhrp2 sequence provides evidence that the function(s) of the molecule is not affected by the sequence diversity in exon 2 and that parasites with a particular pfhrp2 sequence do not appear to have a significant fitness or survival advantage. The host immune responses may also contribute to the maintenance of diversity by selecting for immunologicly different types.

A mechanism of chromosome breakage and healing is believed to contribute to generation of deletions in various parasite chromosomes 27 . Chromosome deletions at the pfhrp2 locus of chromosome 7 and pfhrp3 locus of chromosome 13 have been observed in laboratory-adapted lines 17 20 27 , and in field isolates from the Amazon region of Peru 44 . This observation of gene deletions in field isolates suggests that PfHRP2-detecting RDTs may not be reliable for detecting P. falciparum infections in this region of South America, and also raised a serious question as to whether parasites lacking PfHRP2 and PfHRP3 may exist in other parts of the world and what are possible fitness or survival advantages. In the 485 isolates, collected in over 30 malaria endemic countries, that we examined, we did not observe any deletions of pfhrp2 or pfhrp3 in any field isolates, suggesting that the parasites with gene deletions may not be widespread outside of South America. However, it should be noted that the sampling did not include some endemic regions (e.g. South and Western Asia) and the number of samples examined for some countries was quite small, not coming from different areas within the country. To ensure the performance of RDTs, efforts should be taken to monitor the existence and spread of parasites with gene deletions, especially when false negative results associated with a high parasitaemia are reported.

Pfhrp2 and pfhrp3 variability appears to occur independent of the other, as no correlation between the lengths of the two genes was evident. However, both proteins share several repeats, such as type 1 and 2 repeats. These shared repeats are probably the basis for observed cross reactivity between PfHRP2 and PfHRP3 by monoclonal antibodies against PfHRP2 8 18 , which may contribute to the detection sensitivity of PfHRP2-detecting RDTs and reduce the effect of PfHRP2 variability on RDT sensitivity, particularly at high parasite densities.

The organization and the number of repeats in PfHRP2 vary extensively between parasite isolates. Theoretically, the presence and absence, as well as the number of repeats could affect the binding affinity of the antibodies used in RDTs to the parasites' antigen. Indeed, this preliminary analysis based on 16 cultured lines and tested on two earlier malaria RDTs, Paracheck Pf (Orchid Biomedical Systems, India) and ICT Malaria (ICT Diagnostics, South Africa), yielded a binary logistic regression model that was able to predict detection sensitivity of these two RDTs based on sequence structure 24 . However, the number of isolates used was limited and cultured parasites were used. In this study, the analysis was repeated using the recently completed WHO product testing of malaria RDTs (Round 1) results that included the testing of 34 PfHRP2-detecting RDTs tested against 79 isolates at 200 parasites/μl. This much larger sample size enabled more stringent analysis. The regression analysis this time did not show a correlation between PfHRP2 structure and the overall RDT detection rates. This result suggests that the performance of this group of RDTs is not greatly affected by the diversity of PfHRP2 at parasite densities of 200 parasite/μL or above and, therefore, should increase the confidence in the performance of the devices in various geographic locations. However, this does not exclude an effect of sequence diversity on RDT detection rates at lower parasite densities, or by RDTs employing other monoclonal antibodies.

PfHRP2, the most common target of malaria RDTs, is highly polymorphic throughout the malaria-endemic regions included in this study. While this may affect RDT sensitivity at very low parasite densities ( < 200 parasites/μL), it appears unlikely to have a major impact in most endemic regions at parasite densities usually seen in clinical malaria (200 parasites/μL and above). The extensive diversity in the pfhrp2 sequence also provides evidence that there has not been a strong evolutionary selection for any particular type of sequence. The better understanding of the structure of PfHRP2 and its variation contributes to the evaluation and testing of malaria RDTs, the RDT quality assurance programmes and helps improve malaria RDTs.

The authors declare that they have no competing interests.

JB1 carried out molecular genetic studies, sequence alignment, statistical analyses, and drafted the manuscript. MFH carried out molecular genetic studies and submission of sequences to GenBank. AP and MG carried out statistical analyses. NC assisted with molecular genetic studies for the Nigerian isolates. SA, FA, JB2, AA, DB, JC, DD, DFE, DG, JH, MPK, JL, CM1, DM, CM2, SN, BO, PO, WO, SQW supervised, carried out and coordinated field collection of isolates from patients. DB, JM and QC conceived of the study, participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.