Figures
Abstract
Plasmodium knowlesi is a newly described zoonosis that causes malaria in the human population that can be severe and fatal. The study of P. knowlesi parasites from human clinical isolates is relatively new and, in order to obtain maximum information from patient sample collections, we explored the possibility of generating P. knowlesi genome sequences from archived clinical isolates. Our patient sample collection consisted of frozen whole blood samples that contained excessive human DNA contamination and, in that form, were not suitable for parasite genome sequencing. We developed a method to reduce the amount of human DNA in the thawed blood samples in preparation for high throughput parasite genome sequencing using Illumina HiSeq and MiSeq sequencing platforms. Seven of fifteen samples processed had sufficiently pure P. knowlesi DNA for whole genome sequencing. The reads were mapped to the P. knowlesi H strain reference genome and an average mapping of 90% was obtained. Genes with low coverage were removed leaving 4623 genes for subsequent analyses. Previously we identified a DNA sequence dimorphism on a small fragment of the P. knowlesi normocyte binding protein xa gene on chromosome 14. We used the genome data to assemble full-length Pknbpxa sequences and discovered that the dimorphism extended along the gene. An in-house algorithm was developed to detect SNP sites co-associating with the dimorphism. More than half of the P. knowlesi genome was dimorphic, involving genes on all chromosomes and suggesting that two distinct types of P. knowlesi infect the human population in Sarawak, Malaysian Borneo. We use P. knowlesi clinical samples to demonstrate that Plasmodium DNA from archived patient samples can produce high quality genome data. We show that analyses, of even small numbers of difficult clinical malaria isolates, can generate comprehensive genomic information that will improve our understanding of malaria parasite diversity and pathobiology.
Citation: Pinheiro MM, Ahmed MA, Millar SB, Sanderson T, Otto TD, Lu WC, et al. (2015) Plasmodium knowlesi Genome Sequences from Clinical Isolates Reveal Extensive Genomic Dimorphism. PLoS ONE 10(4): e0121303. https://doi.org/10.1371/journal.pone.0121303
Academic Editor: Osamu Kaneko, Institute of Tropical Medicine, Nagasaki University, JAPAN
Received: November 21, 2014; Accepted: January 30, 2015; Published: April 1, 2015
Copyright: © 2015 Pinheiro et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: HiSeq and MiSeq reads from P. knowlesi enriched, human DNA depleted, samples were deposited in the EMBL-EBI European Nucleotide Archive http://www.ebi.ac.uk with the archive reference for HiSeq: ERR274221; ERR274222; ERR274224; ERR274225 and MiSeq: ERR366425 and ERR366426.
Funding: MMP and Bioinformatics support was provided by the Bioinformatics Unit at St Andrews University funded by a Wellcome Trust ISSF grant 097831/Z/11/Z. This research was funded by Medial Research Council (www.mrc.ac.uk, grant G0801971) (to JCS and SK) and the University of St Andrews. TS, TDO and JCR were supported by the Wellcome Trust (www.wellcome.ac.uk, grant number 098051). TDO was supported by the European Union 7th framework (EVIMalaR) (www.evimalar.org). TS was supported by the Medical Research Council (www.mrc.ac.uk, grant number MR/J500355/1). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Plasmodium knowlesi is a malaria parasite of old world macaques that causes zoonotic malaria in humans [1]. P. knowlesi has been widely used as an experimental model leading to seminal discoveries in aspects of malaria biology, including antigenic variation, vaccine development and erythrocyte invasion (for example [2,3,4]). More recently, the discovery of severe cases of P. knowlesi malaria in the human population has re-kindled human-disease focussed research on this important parasite [5]. P. knowlesi lacks unique morphological characteristics and human infections are often mis-diagnosed as P. malariae or other Plasmodium species[6]. Novel P. knowlesi-specific PCR assays now allow accurate identification of P. knowlesi malaria and PCR-confirmed cases are continuously reported across Southeast Asia, including severe and fatal cases in Malaysia [7,8,9,10].
P. knowlesi is a widespread human infectious agent in Southeast Asia, yet we currently know very little about naturally circulating parasite populations that enter the human host or the factors that are associated with severe disease. In Sarawak, Malaysian Borneo, we found that P. knowlesi parasitaemia is associated with disease severity [8,9]. To study the relationship between parasitaemia and variation in the proteins that are involved in invasion of human erythrocytes, short regions of two P. knowlesi invasion genes, P. knowlesi normocyte binding protein (Pknbp) xa and Pknbpxb, were sequenced from more than 100 human infections [11]. Both gene fragments were polymorphic and the Pknbpxa fragment was dimorphic with distinct co-associating polymorphisms that segregated into two clusters. In the study cohort, patients were infected with parasites with either Pknbpxa dimorphic type at almost equal frequency but only alleles found in one dimorphic type associated with markers of disease severity [11]. While this suggests a potential link between invasion phenotypes, parasitaemia and virulence, it is critical to extend the study beyond a candidate gene level and out to the whole genome.
A reference P. knowlesi genome sequence has been generated from the macaque-adapted experimental H strain [12], but P. knowlesi genome sequences from clinically well-characterised isolates are not currently available. The generation of parasite genome sequences from clinical Plasmodium samples requires a leucocyte depletion step to minimise the amount of contaminating human DNA. However, many archived sample collections exist, including our own collection of frozen whole blood samples from patients with P. knowlesi malaria, that have not been leucocyte depleted before freezing. Adapting depletion approaches to these frozen sample sets would unlock a wealth of genomic information.
Here we report a method to deplete human DNA from frozen clinical malaria samples and render them suitable for whole genome sequencing. The method exploits two assumptions; 1) that not all leucocytes are lysed when whole blood goes through one freeze/thaw cycle and 2) the more robust parasites would survive the same treatment either in intact infected red blood cells (IRBCs) or as free parasites released from lysed erythrocytes. We developed a simple filtration method to remove leucocytes and recover parasite-rich pellets for Plasmodium genome sequencing. The method offers the malaria research community a means to interrogate Plasmodium species genome data in important archived sample collections. In this case, we use the approach to generate genome sequence data from six previously frozen P. knowlesi clinical isolates, and show that the Pknbpxa dimorphism may extend across the P. knowlesi genome.
Materials and Methods
Patient samples
Archived frozen whole blood samples were used from adult patients recruited into a non-interventional study with informed signed consent that included use of samples in related studies. Patient consent forms are securely stored in the University of St Andrews. Patient recruitment and consent protocols were approved by the Medical Research and Ethics Committee, Ministry of Health Malaysia and the Ethics Committee Faculty of Medicine and Health Sciences, University Malaysia Sarawak. The use of the samples in the study reported here was further approved by the University of St Andrews Teaching and Research Ethics Committee.
Human DNA depletion using Whatman filter paper
EDTA blood samples from P. knowlesi patients were collected and stored at -40°C. The samples were thawed and the volume measured before gentle mixing in ice-cold PBS at a ratio of 300ul thawed blood per 5ml cold PBS. The mixture was pipetted into a 10mL syringe barrel, the base was lined with 3 layers of Whatman No 3 (6uM pore size) to remove small lymphocytes and 3 layers of Whatman No 1 (11uM pore size) on top to remove larger surviving leucocytes. The filter papers were cut to fit the internal diameter of the syringe and pre-wet with PBS before use. Not more than 10mL of diluted blood was loaded per syringe column. The filtrate was collected into sterile 50mL centrifuge tubes following centrifugation at 125g for 2 minutes at 4°C. The columns were washed through with 10mL volumes of cold PBS and each wash was collected into the filtrate tube by centrifugation as above until the filters were no longer blood-stained. The total combined filtrate, up to 40 mL, was centrifuged at 2000g for 10 minutes at 4°C to pellet any surviving IRBCs and free parasites. Pellets were re-suspended in 1ml cold PBS and transferred to 1.5ml Eppendorf tubes and recovered by centrifugation at 14,000g, for 2 minutes at 4°C. Pellets were re-suspended and washed in 1mL cold PBS and collected by centrifugation as described. This wash step was repeated two more times. The washed IRBC/parasite pellets were suspended in 20ul Proteinase K (QIAGEN) followed by 200ul cold PBS. The mixture was vortexed thoroughly before DNA extraction using QIAamp Blood Mini kit (QIAGEN) with RNase A, as per manufacturers instructions. For samples with more than 100,000 parasites /ul blood the initial blood dilution step was 150ul thawed blood into 5mL cold PBS.
TaqMan qPCR multiplexed for human and P. knowlesi DNA
Plasmodium specific 18ssURNA Plasmo1: 5′ GTTAAGGGAGTGAAGACGA TCAGA and Plasmo 2: 5′ AACCCAAAGACTTTGATTTC TCATAA primers were used [13] with the published P. knowlesi TaqMan probe 5′ CTCTCCGGAGATTAGAACTCTTAGATTGCT labelled with 5'FAM and 3'BHQ1 [14]. Human DNA primers: Plat1-A 5′ CTTACCACATCCGCTCCATC, and Plat1-B 5′ TTCACACTCTCCGTCACATTG with the probe 5′ HEX/CACATCCCC/ZEN/AGTGCCGAGTTAGA/3IABkFQ were used. The qPCR master mix contained 250nM Plasmo1, 250nM Plasmo2, 250nM Plat1-A, 250nM Plat1-B, 125nM Pk probe, 125nM Plat1-Probe, 1 x Roche RT-PCR Master Mix and 1ul DNA template in 20ul final volume. qPCR cycling was 10 minutes at 95°C, followed by 45 cycles of 10 seconds at 95°C, 30 seconds at 57°C, and 1 second at 72°C using the Roche LightCycler 480 II.
Illumina sequencing
DNA was quantified (Qubit Fluorometric Quantitation, Invitrogen, Life Technologies) and sheared into fragments of 400–600 bp. Illumina libraries were generated using a) the PCR free protocol (NoPCR) [15] or b) the standard library preparation using the KAPA enzyme [16] with eight PCR cycles. NoPCR libraries were sequenced on the Illumina HiSeq 2000 platform for 100 paired-end cycles and standard PCR libraries were sequenced on Illumina MiSeq for 150 paired ends cycles using V4 or V5 SBS sequencing kits and proprietary reagents according to manufacturer's recommended protocol (https://icom.illumina.com/). Data were analysed from the Illumina sequencing machines using RTA1.6, RTA1.8 or GA v0.3 analysis pipelines.
Reference genome
The Plasmodium knowlesi H strain reference genome version 11.1 GeneDB (www.genedb.org/Homepage/Pknowlesi) was downloaded from PlasmoDB (www.plasmodb.org) [12,17,18]. The region corresponding to the pknbpxa gene (PKH_146970 and PKH_146980) in chromosome 14 was partially missing and fragmented in the current reference genome and we corrected for this using the published pknbpxa gene sequence (GenBank accession number EU867791.1) [2]. Common non-coding DNA regions upstream and downstream of the pknbpxa gene were located in both the Plasmodium knowlesi strain H reference genome and the published pknbpxa gene. With this information it was possible to replace the pknbpxa gene (PKH_146970) in the reference genome sequence with the published EU867791.1 gene sequence without disrupting subsequent mapping. The pknbpxb gene, which was not annotated correctly, was rectified using the EU867792.1 published gene sequence [2].
Genome sequence mapping
HiSeq and MiSeq reads from P. knowlesi enriched, human DNA depleted, samples are deposited in the EMBL-EBI European Nucleotide Archive (http://www.ebi.ac.uk)[19]. The archive references are for HiSeq: ERR274221; ERR274222; ERR274224; ERR274225 and MiSeq: ERR366425 and ERR366426. Sequences mapping to the human genome, representing patient DNA, were removed from this data in the sequencing pipeline. The reads were mapped to the corrected P. knowlesi H strain reference genome sequence (PlasmoDB-11.1_PknowlesiH_Genome.fasta) using Bowtie-2 [20] followed by Bedtools to summarise the coverage of each genome [21].
Single Nucleotide Polymorphism (SNP) calling
Samtools mpileup with threshold base quality set to 13 was used with BCFtools to generate Variant Call SNP Format (VCF) files for each P. knowlesi genome sequence [22]. A varFilter (BCFtools) was applied and all SNP sites with allele frequency less than 0.9 were removed. Insertions and deletions were not included in any of the analyses or scripts. Only SNP sites with a minimum coverage of 13 were taken into consideration.
Linkage Disequilibrium analysis of full-length pknbpxa sequences extracted from P. knowlesi genome sequence data
We used Artemis [12,17,23,24] and the VCF files to generate full-length pknbpxa gene sequences as fasta files from each of the genome sequences (n = 6). The fasta files were converted to the Haploview compatible PLINK format [25]. Linkage disequilibrium was performed on the full-length coding region of Pknbpxa sequences using Haploview and analysed using default parameters [26,27]. Nucleotide diversity (π) was calculated using a 400bp window length with a step size of 25bp, DnaSP [28].
Identification of polymorphisms genome-wide co-associating with the Pknbpxa fragment dimorphism
An algorithm was developed to identify SNP sites in each genome sequence (n = 6), co-associating with the P. knowlesi Pknbpxa dimorphic pattern already identified in a small fragment of this gene [11] and also visible in Artemis on chromosome 14 at the Pknbpxa locus (Fig. 1). Briefly the script was designed to screen VCF files to identify each SNP and test if the SNP co-associated with SNPs defining the Pknbpxa dimorphism. Co-associating patterns were predefined in the algorithm to describe which kind of symbols (SNP pattern) each genome required to fit within either of the P. knowlesi Pknbpxa dimorphic forms. Every time a SNP fit the pattern the event was signalled (recorded). Finally, an image was created to show the density of all SNP sites and then the co-associating SNPs for each chromosome (S1 Fig.).
The P. knowlesi normocyte binding protein xa locus on chromosome 14 is shown. The screen shot shows segregation of the sequences from patient isolates into two groups, (n = 3 in each group) and the dimorphism is clearly visible.
Testing the distribution of co-associated SNPs defined in the Pknbpxa dimorphism
To identify positions on each P. knowlesi chromosome where the density of co-associating sites was more evident a Chi square test of independence was applied followed by a calculation of adjusted residuals. For this, each chromosome was divided into 30 equal parts and a contingency table was created to reflect the number of SNPs co-associating with the Pknbpxa dimorphism per part per chromosome. Adjusted residuals were calculated in a contingency table and a threshold of > 3.00 for more co-associating SNPs than expected and < -3.00 for less co-associating dimorphic SNPs than expected was applied to the resulting values. By applying these thresholds it was possible to identify, within 99.7% limits of confidence, co-associating SNP sites for each chromosome with higher or lower than expected co-associating SNP density.
Gene Ontology (GO)
P. knowlesi genes were analysed using Blast2GO http://www.blast2go.com version 2.7.2 [29]. All genes with complete coverage (4623) were blasted against nr@ncbi database and an InterProScan 5 analysis was performed [30]. The Gene Ontology classification was done with default parameters [31]. Genes with no (0) co-associating SNP sites, with >0–<10 (1–9) and >9 (≥ 10) SNPs that co-associated with the P. knowlesi genome-wide dimorphism were identified within each resulting GO group.
Testing for enrichment of dimorphic genes in particular GO subgroups
Genes with dimorphic SNP sites were tested for statistically significant enrichment of dimorphic genes in GO subgroups using topGO Enrichment analysis for Gene Ontology. R package version 2.14.0. Adrian Alexa and Jorg Rahnenfuhrer (2010). (http://www.bioconductor.org/packages/release/bioc/html/topGO.html). For this we selected two groups of genes those with one or more dimorphic SNP sites (≥1) and a separate group with ten or more dimorphic SNP sites (≥10) and analysed for enrichment against all genes with at least one SNP whether or not dimorphic.
Results
Human DNA depletion from frozen whole blood samples
Frozen whole blood samples from fifteen P. knowlesi patients, with parasite counts ranging from 10,000–400,000 parasite/ul, were thawed and human DNA depleted using an in-house method. Briefly white blood cells were removed by filtration through Whatman filter paper followed by parasite recovery as described in detail (see Methods section). Total human and parasite DNA was quantified using qPCR (Table 1). Nine of fifteen isolates had the required >100ng of P. knowlesi DNA, and seven of the nine had <80% human DNA contamination, the cut-off for Plasmodium genome sequencing, and were suitable for sequencing (Table 1). Five and two DNA samples were used to generate NoPCR and PCR sequencing libraries and multiplexed in a single lane on Illumina HiSeq and MiSeq platforms respectively (Table 1). The remaining six samples had insufficient P. knowlesi DNA and/or >80% human DNA (hDNA) contamination (Table 1).
DNA obtained from frozen P. knowlesi clinical isolates generated high coverage genome sequence
P. knowlesi sequence data was generated from seven patient isolates, five from HiSeq runs and two from MiSeq runs. The HiSeq runs generated >36 million reads and MiSeq >5 million reads. An average mapping of >90% was obtained for both HiSeq and MiSeq data producing an average coverage of >140x for HiSeq and >30x for MiSeq. The total number of reads mapped and not mapped, percent human DNA contamination and coverage per genome sequence are summarized in Table 2.
P. knowlesi genome analysis
The reads were mapped to the P. knowlesi H strain reference genome following correction of the Pknbpxa locus (see materials and methods section). Two genome sequences (ERR274222 and ERR274223) were generated from a single patient representing pre- and post-treatment samples. Only the pre-treatment sample sequence, ERR274222, was included in subsequent analyses along with sequences from five other patients all collected pre-treatment.
The sequences covered 5228 genes, including genes and gene fragments annotated as genes of un-known function. Data from 605 genes were excluded because coverage was zero at one or more base position leaving 4623 genes in subsequent analyses. This filter excluded all but five of the 195 SICAVar genes and gene fragments and all but three of the 67 KIR genes and gene fragments. Both of these gene families are highly polymorphic, and the gene sequences in these contemporary clinical isolates are likely to be very different to those in the historical monkey-adapted reference genome, so mapping issues and low coverage in these gene sets is to be expected. Of the remaining genes 2180 (47.2%) were annotated as genes with unknown function. The SNP distribution across the genome is shown in S1 Fig.
Dimorphism extends across and beyond Pknbpxa
In a previous study we identified a DNA sequence dimorphism in a fragment (885bp) representing 10% of the P. knowlesi normocyte binding protein (Pknbp)xa gene that codes for a protein involved in red blood cell invasion. To determine the extent of the dimorphism across the gene, full-length Pknbpxa, (PKH_146970) coding sequences were assembled from the six genome sequences obtained from the same patient cohort. Ninety-one (91) Pknbpxa SNPs co-associated with the dimorphic pattern (r2 = 1). This dimorphism effectively divides the Pknbpxa gene sequences into two clusters of sequence types, with Pknbpxa sequences from three genomes falling into cluster 1 and three into cluster 2. Nucleotide diversity (π) was higher across the clusters (π = 0.01441), than it was within each cluster, (π = 0.00518 for cluster 1 (n = 3) and π = 0.00868 for cluster 2; n = 3 each). Cluster 1 was less diverse than cluster 2, consistent with Pknbpxa nucleotide diversity found in the previous study, but the significance of this difference cannot be estimated based on six sequences.
The P. knowlesi genome sequences from clinical isolates were viewed in Artemis, a genome browser and annotation tool and referenced to the P. knowesi H strain genome sequence. Two SNP patterns emerged and the Pknbpxa dimorphism was clearly visible with sequences from three patient isolates clustering into each pattern (Fig. 1). To test whether the dimporphism extended beyond the boundaries of the Pknbpxa gene, SNP association with the Pknbpxa dimorphism was examined first along chromosome 14 and then genome-wide using an in-house script (see Materials and Methods section). The dimorphic SNP pattern was evident at multiple genetic loci on all chromosomes (S1 Fig.). The relative distribution of co-associated SNPs on each chromosome was determined by dividing each chromosome into 30 equal parts and using the Chi squared test of independence to test expected and observed events (S1 Table). Although the dimorphism extends across the full genome, the intensity and distribution is not uniform or clustered in any particular chromosomal region. The position and number of non-synonymous and synonymous SNPs co-associating with the dimorphism per gene per chromosome are represented in Fig. 2.
Non-synonymous polymorphisms (red) are shown above the line and synonymous polymorphisms (blue) are shown below the line. The line is drawn at zero. The chromosomes are drawn to scale and the height of the bars represents the number of SNP sites per gene per region of each chromosome. The scale is given in the boxed area and is the number of SNP sites per gene.
More than half of the P. knowlesi genes in the genome, 2801 of 4623 genes (60.8%), appear to be dimorphic. Within the dimorphic group the number of dimorphic SNP sites per gene varied widely. For example, while Pknbpxa had a total of 326 SNPs, of which 91 (27.9%) co-associated with the dimorphism, a related gene of similar size, Pknbpxb, had a total of 197 SNPs of which only 5 (2.2%) co-associated with the dimorphism (S2 Table). Applying a more conservative cut-off identified 507 genes with ≥10 co-associating SNPs, representing 11% of genes in the genome with adequate coverage. Of these 301 (59.5%) were annotated as genes of unknown function. The chromosome location and annotated function of the remaining 206 genes is listed in S2 Table. Notable genes within this high stringency dimorphic group included 12 of 27 (44%) of genes annotated as transcription factors with AP2 domains in the P. knowlesi genome. Several genes associated with drug resistance in other Plasmodium species, such as putative multi drug resistance-associated protein PkMRP1(PKH_144590) and putative multidrug resistance protein, PkMDR 2 (PKH_125840) S2 Table, were also dimorphic, while the putative chloroquine resistance transporter (CRT) gene (PKH_010710) had 23 SNPs, none were dimorphic and only one SNP conferred an amino acid change.
The enrichment of dimorphic genes among genes encoding transcription factors with AP2 domains was obvious and identified manually. We then used Gene Ontology (GO), (Blast2GO) tools to examine whether other P. knowlesi dimorphic genes were enriched in GO groups that served particular biological functions. Genes were sorted into GO term groups and sub-groups with putative or known molecular function, cellular process activity and biological process activity (Table 3). We then calculated the proportion genes with ≥1 dimorphic SNP in each GO group (Table 3). Most of the GO term groups had, as expected, approximately 60% dimorphic genes but there was variation (Table 3). If dimorphic genes have evolved randomly over time then the proportion of genes with dimorphic SNP sites in the GO groups would not be expected to be different from the distribution of genes with dimorphic SNP sites in the whole genome that is: 39% of genes with no dimorphic SNPs; 50% of genes with 1–9 dimorphic SNPs and 11% of gene with ten or more dimorphic SNPs. Several GO sub-groups had more genes than expected with 1–9 dimorphic SNP sites and ≥10 dimorphic SNP sites for example genes with molecular transducer activity, nucleic acid binding transcription factor activity and membrane association (Fig. 3a, 3b and 3c). There were also sub-groups of genes where dimorphic SNP sites were under-represented, including structural and molecular activity, developmental process and immune system process (Table 3 and Fig. 3). We used topGO to test for statistically significant enrichment of dimorphic genes in GO term groups (Table 4). In the first instance all genes with at least one dimorphic SNP were analysed and there was significant enrichment, particularly in the ion binding function, helicase activity and tRNA metabolic process function GO term groups (Table 4). Genes with ≥10 dimorphic SNPs were significantly enriched in the nucleic acid binding transcription factor activity and kinase activity GO term groups (Table 4).
The percent of genes in each GO sub-group of a) molecular function, b) cellular processes and c) biological processes are shown, n = the total number of mapped annotated genes in each GO sub-group. Percent of genes in GO subgroups with: no dimorphic SNP sites shown in brown, genes with 1–9 dimorphic SNP sites turquoise and genes with ≥10 dimorphic SNP sites purple. The expected percent of genes with 1–9 dimorphic SNPs (50%) is marked with a turquoise hatched line and the expected percent of genes with ≥10 dimorphic SNPs (11%) is marked with a purple hatched line. Gene ontology was assigned using Blast2GO—Software for Biologists, http://www.blast2go.com.
Discussion
Here we describe a method for enriching Plasmodium DNA from frozen whole blood samples collected from patients with malaria. The method required at least 200ul of whole blood at >40,000 parasites/ul to obtain sufficient parasite DNA for genome sequencing platforms. Parasite DNA recovery was inconsistent and human DNA contamination was the main problem. Nonetheless, seven of fifteen patient samples had sufficiently enriched P. knowlesi DNA to produce high quality genome sequences using Illumina sequencing platforms. The success may in part be because P. knowlesi is less AT rich (62%) than other Plasmodium genomes [12] perhaps reducing amplification bias. Combining the frozen sample filtration method described here with methylated DNA digestion and target enriched sequencing approaches described by others [16,32], may yield valuable Plasmodium genome data from many precious pre-existing frozen clinical sample collections.
In a previous study we identified a sequence dimorphism in a fragment (885bp) of the P. knowlesi normocyte binding protein (Pknbp)xa that codes for a protein involved in red blood cell invasion [11]. Pknbpxa dimorphic cluster 2 contained alleles associated with markers of disease severity implying that dimorphic cluster 2 may contain more virulent parasites than cluster 1. Our genome data revealed that the dimorphism extended along the full-length (>8000bp) Pknbpxa coding region, along chromosome 14 and beyond. SNPs co-associating with the Pknbpxa dimorphism were distributed genome-wide across all chromosomes.
Interestingly, even within the limitation that only six samples were sequenced, the dimorphism comprised numerous non-synonymous substitutions, suggesting, for the first time, that there may be at least two distinct types of P. knowlesi circulating in Sarawak, Malaysian Borneo, and that some may be more virulent that others. Dimorphic loci have been described in many Plasmodium species, particularly in merozoite surface antigens and invasion ligands of P. falciparum and P. vivax [33,34,35]. In P. ovale dimorphic characteristics at selected loci prompted the division of P. ovale into two sub-species [35]. Even so the evolution and maintenance of allelic dimorphisms in Plasmodium species is difficult to explain [34]. Here we demonstrate a genome-wide dimorphism, involving more than half of the genes in the P. knowlesi genome, including genes coding for functions that transcend from exposed parasite surfaces to protected internal sites. The sub-division of P. knowlesi into distinct types will require further sequence confirmation, yet the genome-wide nature of the dimorphism is striking.
Although there was significant enrichment of dimorphic genes in several GO functional groups it is not clear what is driving a genome-wide dimorphism in P. knowlesi. Interestingly twelve genes implicated in parasite lifecycle stage-specific transcription, the putative transcription factors with Apicomplexan Apetala2 (AP2) domains [36,37,38,39] were dimorphic. Variation at these loci may mark genetically distinct lifecycle characteristics isolating P. knowlesi into strains or subspecies. In addition, all nine members of the ABC, ABC C transporter protein family of genes, annotated in the P. knowlesi genome, were dimorphic [12,40]. These genes are found in all phyla and represent an ancient gene family that, in eukaryotes, expel a wide range of unwanted substrates [41]. This family of genes include P. knowlesi PkMDR2 and PkMRP1 that were both polymorphic and dimorphic implying selection pressure at these loci. PkMDR2 and PkMRP1 are orthologues of P. falciparum PfMDR 2 and PfMRP1, genes that carry genetic markers of drug resistance, including resistance to mefloquine [40,42]. Tantalizingly, experimental lines of P. knowlesi were found innately resistant to mefloquine in Rhesus monkeys and clinical isolates did not respond well to mefloquine ex vivo [43,44]. Patients with uncomplicated P. knowlesi infections responded to mefloquine but one patient with severe disease exhibited RIII type resistance [45,46,47].
Selection at these promiscuous transporter loci in zoonotic parasites that, unlike P. falciparum, are not under conventional drug selection pressure may at first seem surprising. However, domestic and wild animals eat plants with bio-active properties—they self-medicate [48]. The jungles of Sarawak are considered un-mined treasure-troves of plant species with medicinal properties that are freely available to the animal species living there, including the macaque reservoir of P. knowlesi [49]. Selection at Plasmodium loci, that have evolved to eliminate natural toxins, then assumes biological relevance. Unfortunately these loci also evolve to eliminate antimalarial compounds when used to treat patients with malaria.
P. knowlesi is a relatively 'un-tamed' Plasmodium species, therefore P. knowlesi genomes may retain ancient and diverse genetic signatures, that are presently invisible in heavily drug selected human-host restricted parasite populations such as P. falciparum and P. vivax. High throughput pathogen genome sequencing is a powerful new tool for infectious disease research. Here we use Illumina HiSeq and MiSeq platforms to produce high quality P. knowlesi genome sequences from difficult archived frozen samples. Analysis of the sequences uncovered a P. knowlesi genome-wide dimorphism that suggests there are least two types of P. knowlesi parasites in our patient cohort. We further discovered dimorphic genes among transporter genes that are important in antimalarial drug resistance. Genome-wide pathogen analyses, of even a small number of clinical malaria isolates, instantly added context to our understanding of Plasmodium pathobiology, particularly through between-species comparison.
Supporting Information
S1 Fig. P. knowlesi genome SNP density map.
Six P. knowlesi genome sequences from patient isolates were mapped to the P.knowlesi reference genome. Sites that differ from the reference are shown as blue bars (all SNP sites) or grey bars (SNP sites co-associating with the P. knowlesi genome-wide dimorphism). Each bar is 1 pixel wide and represents DNA fragments 809 bases long. The height of the bars represents the number of SNP sites per 809 base fragment. Gaps correspond to regions with low coverage (see results section) or where the reference genome is incomplete (runs of 'N').
https://doi.org/10.1371/journal.pone.0121303.s001
(PNG)
S1 Table. Distribution of co-associating SNPs by chromosome in six P. knowlesi genome sequences from human isolates.
Each chromosome was divided into 30 equal parts.
https://doi.org/10.1371/journal.pone.0121303.s002
(PDF)
S2 Table. List of annotated genes from six clinical isolates with >9 SNPs that co—associate with the Plasmodium knowlesi genome-wide dimorphism.
https://doi.org/10.1371/journal.pone.0121303.s003
(PDF)
Acknowledgments
We thank Matt Berriman, Mandy Sanders, Dan Alcock Wellcome Trust Sanger Institute for their support with whole genome sequencing. Matt Holden, University of St Andrews for helpful comments on the manuscript. An especial thank you to Dr Wong Ing Tien and the staff and patients in Hospitals Sibu and Sarikei, Sarawak, Malaysian Borneo.
Author Contributions
Conceived and designed the experiments: JCS MAA SK JCR. Performed the experiments: MAA TS TDO. Analyzed the data: MMP SBM JCS JCR. Contributed reagents/materials/analysis tools: MMP TDO. Wrote the paper: JCS JCR. Patient recruitment: WCL.
References
- 1. Singh B, Daneshvar C. Human Infections and Detection of Plasmodium knowlesi. Clin Microbiol Rev. 2013; 26: 165–184. pmid:23554413
- 2. Meyer EV, Semenya AA, Okenu DM, Dluzewski AR, Bannister LH, Barnwell JW et al. The reticulocyte binding-like proteins of P. knowlesi locate to the micronemes of merozoites and define two new members of this invasion ligand family. Mol Biochem Parasitol. 2009; 165: 111–121. pmid:19428658
- 3. Miller LH, Hudson D, Haynes JD. Identification of Plasmodium knowlesi erythrocyte binding proteins. Mol Biochem Parasitol. 1988; 31: 217–222. pmid:3221909
- 4. Miller LH, Mason SJ, Dvorak JA, McGinniss MH, Rothman IK. Erythrocyte receptors for (Plasmodium knowlesi) malaria: Duffy blood group determinants. Science 1975; 189: 561–563. pmid:1145213
- 5. Singh B, Kim Sung L, Matusop A, Radhakrishnan A, Shamsul SS, Cox-Singh J et al. A large focus of naturally acquired Plasmodium knowlesi infections in human beings. Lancet 2004; 363: 1017–1024. pmid:15051281
- 6. Lee KS, Cox-Singh J, Singh B. Morphological features and differential counts of Plasmodium knowlesi parasites in naturally acquired human infections. Malar J. 2009; 8: 73. pmid:19383118
- 7. Cox-Singh J. Zoonotic malaria: Plasmodium knowlesi, an emerging pathogen. Cur Opin Infect Dis. 2012; 25: 530–536. pmid:22710318
- 8. Cox-Singh J, Davis TM, Lee KS, Shamsul SS, Matusop A, Ratnam S, et al. Plasmodium knowlesi malaria in humans is widely distributed and potentially life threatening. Clin Infect Dis. 2008; 46: 165–171. pmid:18171245
- 9. Daneshvar C, Davis TM, Cox-Singh J, Rafa'ee MZ, Zakaria SK, Divis PC, et al. Clinical and laboratory features of human Plasmodium knowlesi infection. Clin Infect Dis. 2009; 49: 852–860. pmid:19635025
- 10. William T, Menon J, Rajahram G, Chan L, Ma G, Donaldson S, et al. Severe Plasmodium knowlesi Malaria in a Tertiary Care Hospital, Sabah, Malaysia. Emerg Infect Dis. 2011;17: 1248–1255. pmid:21762579
- 11. Ahmed MA, Pinheiro MM, Divis PC, Siner A, Zainudin R, Wong IT, et al. Disease Progression in Plasmodium knowlesi Malaria Is Linked to Variation in Invasion Gene Family Members. PLoS Neglect Trop Dis. 2014; 8: e3086. pmid:25121807
- 12. Pain A, Bohme U, Berry AE, Mungall K, Finn RD, Jackson AP, et al. The genome of the simian and human malaria parasite Plasmodium knowlesi. Nature 2008; 455: 799–803. pmid:18843368
- 13. Rougemont M, Van Saanen M, Sahli R, Hinrikson HP, Bille J, Jaton K. Detection of four Plasmodium species in blood from humans by 18S rRNA gene subunit-based and species-specific real-time PCR assays. J Clin Microbiol. 2004; 42: 5636–5643. pmid:15583293
- 14. Divis PC, Shokoples SE, Singh B, Yanow SK. A TaqMan real-time PCR assay for the detection and quantitation of Plasmodium knowlesi. Malar J. 2010; 9: 344. pmid:21114872
- 15. Kozarewa I, Ning Z, Quail MA, Sanders MJ, Berriman M, Turner DJ, et al. Amplification-free Illumina sequencing-library preparation facilitates improved mapping and assembly of (G+C)-biased genomes. Nat Methods 2009;6: 291–295. pmid:19287394
- 16. Oyola SO, Gu Y, Manske M, Otto TD, O'Brien J, Alcock D, et al. Efficient depletion of host DNA contamination in malaria clinical sequencing. J Clin Microbiol. 2013; 51: 745–751. pmid:23224084
- 17. Aurrecoechea C, Brestelli J, Brunk BP, Dommer J, Fischer S, Gajria B, et al. PlasmoDB: a functional genomic database for malaria parasites. Nucleic Acids Res. 2009; 37: D539–543. pmid:18957442
- 18. Hertz-Fowler C, Peacock CS, Wood V, Aslett M, Kerhornou A, Mooney P, et al. GeneDB: a resource for prokaryotic and eukaryotic organisms. Nucleic Acids Res. 2004; 32: D339–343. pmid:14681429
- 19. Leinonen R, Akhtar R, Birney E, Bower L, Cerdeno-Tarraga A, Cheng Y, et al. The European Nucleotide Archive. Nucleic Acids Res 39: D28–31. pmid:20972220
- 20. Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nat Methods 2012; 9: 357–359. pmid:22388286
- 21. Quinlan AR, Hall IM. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 2010; 26: 841–842. pmid:20110278
- 22. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 2009; 25: 2078–2079. pmid:19505943
- 23. Carver T, Harris SR, Berriman M, Parkhill J, McQuillan JA. Artemis: an integrated platform for visualization and analysis of high-throughput sequence-based experimental data. Bioinformatics 2012; 28: 464–469. pmid:22199388
- 24. Rutherford K, Parkhill J, Crook J, Horsnell T, Rice P, Rajandrem MA, et al. Artemis: sequence visualization and annotation. Bioinformatics 2000; 16: 944–945. pmid:11120685
- 25. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007; 81: 559–575. pmid:17701901
- 26. Barrett JC Haploview: Visualization and analysis of SNP genotype data. Cold Spring Harb Protoc. Oct; 2009 (10): pdb ip71.
- 27. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005; 21: 263–265. pmid:15297300
- 28. Librado P, Rozas J. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics 2009; 25: 1451–1452. pmid:19346325
- 29. Conesa A, Gotz S, Garcia-Gomez JM, Terol J, Talon M, Robles M. Blast2GO: a universal tool for annotation, visualization and analysis in functional genomics research. Bioinformatics 2005; 21: 3674–3676. pmid:16081474
- 30. Jones P, Binns D, Chang HY, Fraser M, Li W, McAnulla C, et al. InterProScan 5: genome-scale protein function classification. Bioinformatics 2014; 30: 1236–1240. pmid:24451626
- 31. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, et al. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet. 2000; 25: 25–29. pmid:10802651
- 32. Mamanova L, Coffey AJ, Scott CE, Kozarewa I, Turner EH, Kumar A, et al. Target-enrichment strategies for next-generation sequencing. Nat Methods 2010; 7: 111–118. pmid:20111037
- 33. Rayner JC, Tran TM, Corredor V, Huber CS, Barnwell JW, Galinski MR. Dramatic difference in diversity between Plasmodium falciparum and Plasmodium vivax reticulocyte binding-like genes. Am J Trop Med Hyg. 2005;72: 666–674. pmid:15964948
- 34. Roy SW, Ferreira MU, Hartl DL. Evolution of allelic dimorphism in malarial surface antigens. Heredity 2008; 100: 103–110. pmid:17021615
- 35. Sutherland CJ, Tanomsing N, Nolder D, Oguike M, Jennison C, Pukrittayakamee S, et al. Two nonrecombining sympatric forms of the human malaria parasite Plasmodium ovale occur globally. J Infect Dis. 2010; 201: 1544–1550. pmid:20380562
- 36. Iwanaga S, Kaneko I, Kato T, Yuda M. Identification of an AP2-family protein that is critical for malaria liver stage development. PloS ONE 2012; 7: e47557. pmid:23144823
- 37. Lindner SE, De Silva EK, Keck JL, Llinas M. Structural determinants of DNA binding by a P. falciparum ApiAP2 transcriptional regulator. J Mol Biol. 2010;395: 558–567. pmid:19913037
- 38. Sinha A, Hughes KR, Modrzynska KK, Otto TD, Pfander C, Dickens NJ, et al. A cascade of DNA-binding proteins for sexual commitment and development in Plasmodium. Nature 2014; 507: 253–257. pmid:24572359
- 39. Yuda M, Iwanaga S, Shigenobu S, Mair GR, Janse CJ, Waters AP, et al. Identification of a transcription factor in the mosquito-invasive stage of malaria parasites. Mol Microbiol. 2009; 71: 1402–1414. pmid:19220746
- 40. Koenderink JB, Kavishe RA, Rijpma SR, Russel FG. The ABCs of multidrug resistance in malaria. Trends Parasitol. 2010; 26: 440–446. pmid:20541973
- 41. Jones PM, O'Mara ML, George AM. ABC transporters: a riddle wrapped in a mystery inside an enigma. Trends Biochem Sci. 2009; 34: 520–531. pmid:19748784
- 42.
Ecker A, Lehane A, Fiddock DA. Molecular markers of Plasmodium resistance to antiamalrials. In: Staines HMaK S., editor. Treatment and Prevention of Malaria: Antimalarial Drug Chemistry, Acation and Use, Basel: Springer; 2012. pp. 249–280.
- 43. Fatih FA, Staines HM, Siner A, Ahmed MA, Woon LC, Pasini EM, et al. Susceptibility of human Plasmodium knowlesi infections to anti-malarials. Malar J. 2013;12: 425. pmid:24245918
- 44. Tripathi R, Awasthi A, Dutta GP. Mefloquine resistance reversal action of ketoconazole—a cytochrome P450 inhibitor, against mefloquine-resistant malaria. Parasitology 2005;130: 475–479. pmid:15991489
- 45. Bronner U, Divis PC, Farnert A, Singh B. Swedish traveller with Plasmodium knowlesi malaria after visiting Malaysian Borneo. Malar J. 2009;8: 15. pmid:19146706
- 46. Lau YL, Tan LH, Chin LC, Fong MY, Noraishah MA, Rohela M. Plasmodium knowlesi Reinfection in Human. Emerg Infect Dis. 2011;17: 1314–1315. pmid:21762601
- 47. Tanizaki R, Ujiie M, Kato Y, Iwagami M, Hashimoto A, Kutsuna S, et al. First case of Plasmodium knowlesi infection in a Japanese traveller returning from Malaysia. Malar J. 2013;12: 128. pmid:23587117
- 48. de Roode JC, Lefevre T, Hunter MD. Ecology. Self-medication in animals. Science 2013;340: 150–151. pmid:23580516
- 49. Yeo TC, Naming M, Manurung R. Building a discovery partnership with sarawak biodiversity centre: a gateway to access natural products from the rainforests. Comb Chem High Throughput Screen 2014;17: 192–200. pmid:24409959