We examined 645 archived blood samples that we had collected during 2011–2014 among indigenous populations of various subtribes from 14 villages in 7 states in Malaysia (Appendix Table 1). We first screened the extracted DNA samples at Universiti Malaya (UM) for the presence of Plasmodium with the aid of genus-specific primers (rPLU1 and rPLU5; rPLU3 and rPLU4) (Appendix). Of the 645 indigenous community samples, 102 (15.8%) were positive for Plasmodium. Using species-specific nested PCR assays (Appendix), we identified these infections as monoinfections with P. knowlesi (n = 40), P. vivax (n = 21), P. cynomolgi (n = 9), P. falciparum (n = 6), P. coatneyi (n = 3), P. inui (n = 3), P. malariae (n = 2), and P. ovale curtisi (n = 1) (Table 1). In 17 samples, the species could not be identified despite repeated attempts. Our species-specific primer pairs were designed on the basis of either the asexually (A) or sexually (S) transcribed forms of Plasmodium small subunit (SSU) rRNA genes (7); the genus-specific primer pairs anneal to both asexual and sexual forms of the SSU rRNA genes, and therefore the genus-specific assay is more sensitive.

Figure 1

Figure 1. Neighbor-joining phylogenetic tree of Plasmodium species based on partial sequence of SSU rRNA genes for identification of Plasmodiummalaria species from indigenous community blood samples, Malaysia. Nucleotide...

We further characterized the 55 samples that tested positive for simian malaria parasites by amplifying a longer fragment of the SSU rRNA gene (914 bp–950 bp) for direct sequencing. Phylogenetic analysis using the neighbor-joining method (Figure 1) revealed the presence of P. knowlesi (samples PK1–40), P. coatneyi (UM1–3), P. cynomolgi (UM9, UM11, UM12, UM14, UM15, UM17, UM18), and P. inui (UM5–7). Meanwhile, 2 sequences derived from samples UM10 and UM16 were found to be closely related to P. simiovale.

We then reextracted DNA from 15 blood samples that were positive for P. coatneyi, P. cynomolgi, and P. inui and sent these samples (blinded) together with 5 Plasmodium-negative samples to Universiti Malaysia Sarawak (UNIMAS) to confirm their identities by PCR and sequencing of part of the cytochrome c oxidase subunit 1 (COX1) gene. At UNIMAS, using nested PCR assays based on SSU rRNA genes, we found 1 single and 9 double species infections. We could not identify the species of Plasmodium for sample UM6, 4 of the Plasmodium-positive samples from UM were Plasmodium negative, and all 5 Plasmodium-negative samples from UM (UM4, 8, 13, 19, 20) tested negative (Table 2). Furthermore, because both laboratories at UM and UNIMAS had previously extracted DNA from macaque blood to examine for simian malaria parasites, we tested the samples for macaque DNA to rule out the possibility that the simian malaria parasites detected were the result of contamination with macaque blood. We obtained negative results using nested PCR for detection of macaque DNA for the 20 DNA samples when they were first received at UNIMAS and also when we repeated testing after completing the sequencing of COX1 genes, indicating that these samples were not contaminated with macaque blood upon receipt or during subsequent experiments at UNIMAS.

Figure 2

Figure 2. Neighbor-joining phylogenetic tree of Plasmodium species based on partial sequence of COX1 genes for identification of Plasmodiummalaria species from indigenous community blood samples, Malaysia. Nucleotide sequences...

We then subjected the PCR-positive samples (UM6–7, UM9–12, UM14–18) to amplification and sequencing of partial COX1 genes. Neighbor-joining (Figure 2) phylogenetic inference of these sequences, together with available referral sequences from GenBank, indicated that 32 haplotypes from samples UM9–12 and UM14–18 were genetically indistinguishable from P. cynomolgi. Our phylogenetic analyses also demonstrated that sample UM7 had a single infection with P. inui–like parasites, whereas UM6 had a double infection with P. simiovale and P. inui–like parasites and UM16 had a triple infection with P. cynomolgi, P. simiovale, and P. inui–like parasites.

We generated phylogenetic trees of similar topology by the maximum-likelihood method for the SSU rRNA genes (Appendix Figure 1) and by the Bayesian maximum clade credibility method for the COX1 genes (Appendix Figure 2). There were discrepancies between the nested PCR assay results and the sequencing results between our 2 laboratories; mixed species of Plasmodium were identified only at UNIMAS. A possible explanation is that the DNA samples analyzed at UNIMAS were newly extracted and were different from the ones used in the experiments at UM. There might also be a compromise of the sensitivity in detecting the species with lower parasitemia in mixed infections as a result of competition for nest 1 primers by the species with higher parasite loads. Furthermore, for sequencing of the SSU rRNA genes at UM, primers that were specific for the species identified by nested PCR assays were used, whereas for the COX1 genes, both P. cynomolgi–specific primers and primers that could amplify other species of Plasmodium were used. Therefore, additional species of Plasmodium were identified at UNIMAS in these samples, such as P. simiovale and P. inui–like, for which no species-specific PCR primers exist.

The 40 P. knowlesi infections we detected originated from 6 states in Malaysia, thereby confirming the widespread distribution of human P. knowlesi malaria cases in Malaysia (1). We detected P. cynomolgi infections among indigenous communities in 4 states in Malaysia. Taken together with previous reports of naturally acquired P. cynomolgi infections in humans in the states of Terengganu, Sabah, and Sarawak (3,8,9), our findings indicate that human infections caused by P. cynomolgi are also widely distributed in Malaysia.

Our study highlights the occurrence of naturally acquired human infections with P. inui, P. inui–like, P. coatneyi, and P. simiovale. Natural human P. inui infections have not been described (10), although the parasite is experimentally transmissible to humans (2). For P. coatneyi, attempts to infect humans with blood from an infected rhesus monkey and through infected mosquitoes were unsuccessful (2). P. simiovale is a lesser-studied simian malaria parasite that was previously described only in toque macaques (Macaca sinica) of Sri Lanka (2) until it was recently identified, together with P. inui–like parasites, in long-tailed macaques from Sarawak in Malaysian Borneo (11). All these simian malaria parasites would have been diagnosed by microscopy as human malaria parasites because they share morphological similarities with human malaria parasites. The early blood stages of P. knowlesi resemble those of P. falciparum, and the other forms are similar to P. malariae (2,6). P. cynomolgi is morphologically similar to P. vivax (2), and both P. inui and P. inui–like parasites are morphologically identical to P. malariae (2,11), whereas P. coatneyi bears morphologic similarities to P. falciparum and P. simiovale bears morphologic similarities to P. ovale (2,12). Besides misdiagnosis of simian malaria parasites as human malaria parasites, there are other limitations of microscopy for diagnosis of malaria; thus, utilization of molecular tools is paramount in generating accurate epidemiology data (6). It is envisaged that screening with molecular tools of other communities living at the forest fringes will demonstrate the widespread distribution of zoonotic malaria and uncover more newly emergent malaria parasites.