Normocyte-binding protein required for human erythrocyte invasion by the zoonotic malaria parasite Plasmodium knowlesi

Significance

Plasmodium knowlesi is a parasite that naturally infects cynomolgus monkeys but is also a major cause of severe zoonotic malaria in humans in South East Asia. Comparing the genomes of parasites restricted to growth in culture with cynomolgus RBCs and those adapted to growth in human RBCs identified a gene specifically required for invasion of human RBCs, a process that is critical for parasite replication. This gene encodes normocyte-binding protein Xa, a protein previously shown to bind human RBCs and implicated in invasion. Disruption of this gene blocks invasion of human but not cynomolgus RBCs, thus confirming a key mediator of human infection and a potential target for inclusion in vaccines to prevent human infection.

Abstract

The dominant cause of malaria in Malaysia is now Plasmodium knowlesi, a zoonotic parasite of cynomolgus macaque monkeys found throughout South East Asia. Comparative genomic analysis of parasites adapted to in vitro growth in either cynomolgus or human RBCs identified a genomic deletion that includes the gene encoding normocyte-binding protein Xa (NBPXa) in parasites growing in cynomolgus RBCs but not in human RBCs. Experimental deletion of the NBPXa gene in parasites adapted to growth in human RBCs (which retain the ability to grow in cynomolgus RBCs) restricted them to cynomolgus RBCs, demonstrating that this gene is selectively required for parasite multiplication and growth in human RBCs. NBPXa-null parasites could bind to human RBCs, but invasion of these cells was severely impaired. Therefore, NBPXa is identified as a key mediator of P. knowlesi human infection and may be a target for vaccine development against this emerging pathogen.

A causative agent of severe zoonotic disease in South East Asia is Plasmodium knowlesi, a malaria parasite of the cynomolgus macaque, Macaca fascicularis (1, 2). In countries such as Malaysia where other malaria parasites are being controlled effectively, P. knowlesi infections are the dominant cause of human malaria (3, 4). In all malaria parasite species, disease is caused by cyclic parasite multiplication within RBCs, releasing invasive merozoites that invade new RBCs. Parasite adhesins transported onto the merozoite surface bind specific host cell receptors to facilitate RBC invasion (reviewed in ref. 5). The first adhesin–RBC interaction was identified in P. knowlesi and Plasmodium vivax, in which it was shown that parasites must bind to the Duffy antigen/chemokine receptor (DARC) to invade human RBCs (6). These adhesins are determinants of parasite virulence (7), host cell tropism (8), and potential vaccine candidates (9), and understanding the role of these adhesins may be critical for identifying the causes of increased risk of human infection and for the development of novel interventions.

Reticulocyte-binding proteins (RBPs) were first described in P. vivax (10) and are the prototypical examples of the reticulocyte binding-like/reticulocyte-binding homolog (RBL/RH) proteins also characterized in other malaria parasite species including Plasmodium cynomolgi (11) and Plasmodium yoelii (12). Plasmodium falciparum normocyte-binding proteins (NBPs) (13, 14) are members of this family that includes four conventional RBL/RH proteins and one unconventional short RBL/RH protein (RH5) in this species (reviewed in ref. 5). In P. knowlesi two proteins in the RBL family, NBPXa and NBPXb, have been identified (15), and recombinant fragments of these proteins bind RBCs (16). The RBL/RH adhesins and the Duffy-binding protein/erythrocyte-binding ligand (DBP/EBL) family of proteins play key roles during merozoite invasion of RBCs (reviewed in refs. 5 and 9). However, ascribing specific roles to these proteins has been hampered by functional redundancy, and only RH5 may be essential for P. falciparum invasion (17). In P. knowlesi, DBPα is the parasite adhesin binding DARC, and two paralogues, DBPβ and DBPγ, also have been identified (18), but thus far only DBPα has been shown to be required for RBC invasion (19).

P. knowlesi can be grown readily in both cynomolgus and rhesus (Macaca mulatta) macaque RBCs in vitro (20–22) but requires extended adaptation for growth in human RBCs in vitro, displaying insufficient invasion efficiency to support continuous culture in human RBCs. In previous work we overcame this impediment by extended adaptation in a mixture of human and cynomolgus RBCs, producing a parasite line that can be maintained in human RBCs alone (21). This previous analysis suggested that improved parasite invasion of human RBCs was essential for the maintenance of parasite growth in vitro.

Here, we undertook a comparative analysis of the genomes of P. knowlesi lines adapted to human or cynomolgus RBCs in culture, in parallel with analysis of the preadapted parasites. The comparison highlighted a crucial role for NBPXa in the invasion of human RBCs. Using the unique capacity to grow this parasite in two host cell types, we were able to disrupt the gene encoding NBPXa and show that it is required for invasion of human but not of cynomolgus macaque RBCs. Human-adapted P. knowlesi parasites provide a robust experimental platform to unravel the mechanisms of host cell invasion in both P. knowlesi and the related pathogen P. vivax.

Results

Culture Adaptation of P. knowlesi Is Associated with Distinct Large Genomic Deletions and Duplications.

To identify the genetic changes occurring during adaptation to in vitro culture, we compared the parasite genomes at two critical points in the process: within 2 mo of introducing the parasite to culture (A1-O) and after full adaptation to growth in human RBCs (A1-H) or in cynomolgus RBCs (A1-C) for a similar period (Fig. 1A). Genome-sequence analysis confirmed that the starting A1-O parasite is closely related to the H strain (23, 24). Use of the Pacific Biosciences (PacBio) sequencing platform allowed the generation of high-quality genome assemblies (Table S1) for clones of the adapted parasite lines (A1-H.1 and A1-C.2) to use as references to identify SNPs and indels among both the uncloned lines and the other derived clones. Few nucleotide differences were identified (Fig. 1A and Dataset S1), with only 22 nonsynonymous variants in protein-coding genes throughout all the lines sequenced. Only three of these differences were found exclusively in the human-adapted parasite, and none was present in all clones, so they were unlikely to be responsible for the adaptation phenotype (Dataset S1). Similarly there were no nonsynonymous substitutions that were present in long-term culture adapted lines but absent from the A1-O parasite. In addition to these SNPs, we identified four larger deletions (3.6–66.3 kb) and one duplication (7.7 kb). In contrast to the SNPs, these appeared to be specific to either the human- or the cynomolgus-adapted lines, resulting in the deletion of 13 genes in A1-C–derived lines or the deletion of two genes and the duplication of four genes in A1-H–derived lines (Table S2). The largest deletion was of ∼66.3 kb at one end of chromosome 14 in the A1-C line (Fig. 1B and Fig. S1). In A1-O and A1-H lines, this region encompasses 12 genes (Table S2), including PKNH_1472300 which encodes NBPXa (15). This observation was of particular interest, because adaptation to human RBCs appeared to result from improved invasion efficiency rather than from improved intracellular growth (21). The genomic comparisons identified no substantial differences in other known invasion gene loci, including DBPα, DBPβ, DBPγ, and NBPXb genes. Similarly, although in A1-H derived lines we detected a duplication which contained four genes, including GTP cyclohydrolase I, we found no evidence of increased copy number for any of the DBP or RBL genes. However, using our data and a set of 48 P. knowlesi genomes from malaria infections in Malaysia (25), a high level of SNP diversity can be shown for the NBPX and DBP genes (Fig. S2 and Dataset S2).

Fig. 1.

Genomic comparisons of adapted parasite lines. (A) Genomic differences across P. knowlesi lines and derived clones used in this study. The starting material was the A1 stabilate derived from the H strain grown in a rhesus macaque. The schematic shows SNPs (green) and small indels (blue) identified between the H strain, the original A1-O line, and the human RBC-adapted A1-H and the cynomolgus RBC-adapted A1-C lines, and respective clones. Venn diagrams show the relationship among the A1-O, A1-H, and A1-C lines to the genome of a clone from each line. The presence of the known invasion genes in the clones (A1-H.1 to H.3, and A1-C.1 to C.2) is shown. (B) Mapping Illumina reads coverage of P. knowlesi A1-O (green), A1-C (red), and A1-H (blue) against chromosome 14 of P. knowlesi H viewed in Artemis. The enlarged area shows the lack of A1-C read coverage, corresponding to a 66.3-kbp deletion encompassing the invasion-related gene NBPXa (PKNH_1472300).

Table S1.

Nuclear genome assemblies of PKA1-C.2 and PKA1-H.1

Feature	PKNA1-C.2	PKNA1-H.1
Sequencing SMRT cells	3	3
No. of reads	361,197	355,195
Average read length (bp)	8,143.76	8,525.33
Total sequence bases	2,941,501,128	3,028,153,055
Theoretical coverage	120.8	126.4
Final assembly size (bp)	24,359,887	23,958,038
Number of contigs	45	37
Percent guanine–cytosine (%)	39	39
Nuclear genome coding DNA sequences	5,208	5,222
Chromosomes*
PKNH_01	869,812 (1)	872,018 (1)
PKNH_02	723,636 (1)	701,237 (1)
PKNH_03	1,015,880 (1)	1,011,696 (2)
PKNH_04	1,117,236 (1)	1,106,313 (1)
PKNH_05	722,893 (2)	718,347 (1)
PKNH_06	1,060,559 (1)	1,063,567 (1)
PKNH_07	1,505,548 (2)	1,504,978 (2)
PKNH_08	1,838,988 (2)	1,821,383 (4)
PKNH_09	2,144,239 (1)	2,113,701 (2)
PKNH_10	1,478,676 (1)	1,479,457 (1)
PKNH_11	2,320,732 (3)	2,321,793 (3)
PKNH_12	3,128,288 (3)	3,082,790 (3)
PKNH_13	2,549,650 (3)	2,549,747 (3)
PKNH_14	2,951,386 (3)	3,186,221 (4)

PacBio genome sequencing, assembly, and annotation information, and chromosomal assignments of contigs are based on unique mapping to publicly available (January 2015 version, Wellcome Trust Sanger Institute File Transfer Protocol server) P. knowlesi H strain chromosome data.

^*Numbers of contigs are shown in parentheses.

Table S2.

Genes affected by genomic deletions or duplications in A1-H and A1-C lines

Parasite line	Chromosome	Deletion/duplication	Size of region, kbp	Gene ID	Protein product
A1-H.1	11	Deletion	8.1	PKNH_1111800	Conserved Plasmodium protein, unknown function
	12	Deletion	7.4	PKNH_1245200	Trailer hitch homolog, putative
	14	Duplication	—	PKNH_1443200	GTP cyclohydrolase I (GCH1)
	14	Duplication	—	PKNH_1443300	Hypothetical protein, conserved in Apicomplexan species
	14	Duplication	—	PKNH_1443400	Conserved Plasmodium protein, unknown function
A1-C.2	4	Deletion	3.6	PKNH_0420600	CCR4-associated factor 16, putative
	14	Deletion	33.8	PKNH_1472300	Reticulocyte binding protein
	14	Deletion	33.8	PKNH_1472400	Tryptophan-rich antigen
	14	Deletion	33.8	PKNH_1472500	Conserved Plasmodium protein, unknown function
	14	Deletion	33.8	PKNH_1472600	Plasmodium exported protein, unknown function
	14	Deletion	33.8	PKNH_1472700	Plasmodium exported protein, unknown function
	14	Deletion	33.8	PKNH_1472800	Plasmodium exported protein, unknown function

In addition to SNPs and indels, several larger genomic deletions/duplications were detected by comparing the A1-H.1 and A1-C.2 genomes with the genome of the H strain (24). The accession numbers and product names of the deleted and duplicated genes are listed.

Fig. S1.

Mapping PCR-free 101-bp paired-end Illumina reads against all 14 P. knowlesi strain H chromosomes: A1-O (green), A1-H (blue), and A1-C (red). The duplication of the GCH1 locus of A1-H on chromosome 14 is indicated by the dashed rectangle. The chromosome 14 deletion containing NBPXa is shown (dotted circle). All chromosomes are to scale.

Fig. S2.

Natural SNP variation in the five invasion-related genes. The scatterplots show the frequency of different SNP variants across 48 naturally occurring clinical isolates and five laboratory lines (17) compared with PKA1-H.1. The x axis represents the nucleotide position of SNPs across the sequences spanning the DBLα (PKNA1_H1_0623500), DBLβ (PKNA1_H1_1400800), DBLγ (PKNA1_H1_1356900), NBPXa (PKNA1_H1_1472300), and NBPXb (PKNA1_H1_0700200) genes. The y axis shows the variant frequency as a fraction of the total number of samples included. The darker shaded areas indicate the positions of the predicted genes; the lighter shaded areas show the 10-kbp flanking regions wherever applicable. The datasets used to generate these plots are presented in Dataset S2.

Adaptation to Human RBCs Results in Increased Invasion Efficiency for both Human and Cynomolgus RBCs.

The genetic differences between cynomolgus and human RBC-adapted parasite lines, including the loss of NBPXa in the A1-C line, suggested that the A1-C and A1-H lines had been subjected to distinct selection processes during culture adaptation. Therefore, the growth of these lines was compared with that of the A1-O line to try to correlate genetic and phenotypic differences (Fig. 2). To compare invasion efficiencies, schizonts purified from each line grown in cynomolgus RBCs were added to fresh human or cynomolgus RBCs, and then new intracellular parasites were quantified 24 h later. The invasion efficiency of the A1-O line in cynomolgus RBCs was around half that of both the A1-H.1 and A1-C.1 lines, as is consistent with selection for improved invasion following prolonged culture (Fig. 2A). Further examination of the cultures showed no evidence of defects in subsequent intracellular growth (Fig. 2 B and C). In human RBCs the invasion efficiency of the A1-O line was approximately half that of the A1-H.1 line but was four times that of the A1-C.1 line. Thus, the A1-H.1 line increased in invasion efficiency for both human and cynomolgus RBCs during more than a year in vitro, but the A1-C.1 line did so only in cynomolgus RBCs, with a reduction in invasion efficiency for human RBCs. Using the 24-h invasion data to calculate a mean host RBC selectivity ratio for the different lines (cynomolgus RBC invasion efficiency/human RBC invasion efficiency) revealed no significant differences between host cell selectivity of the A1-O and A1-H.1 lines (cynomolgus selectivity ± 1 SD = 1.53 ± 0.14 and 1.54 ± 0.49, respectively), clearly demonstrating that no specific increase in human RBC invasion efficiency had occurred in the A1-H.1 line. In contrast, the selectivity of the A1-C.1 line for cynomolgus RBC had increased significantly (cynomolgus selectivity = 16.1 ± 8.1) compared with both the A1-H.1 and A1-O lines (P = 0.0044, one-way ANOVA). The absence of NBPXa in the A1-C.1 line clearly did not reduce its invasion efficiency for cynomolgus cells. We therefore reasoned that its loss may have caused the shift in host cell selectivity through reduced invasion efficiency for human RBCs. The ability of the A1-H.1 line to grow interchangeably in the two host RBCs offered a unique opportunity to test this hypothesis. No other clear relationship between genotype and adaptation to culture was established.

Fig. 2.

Phenotypic characterization of the A1-O parasite line. (A) Growth of the A1-O line and the long-term culture-adapted A1-C.1 and A1-H.1 clones in cynomolgus or human RBCs after 24 h. The data shown are the mean of four independent experiments ± 1 SD. Statistical significance was determined using a two-factor ANOVA with the Sidak post hoc test to provide multiplicity-adjusted P values. Significant (*P < 0.05; **P < 0.005) and nonsignificant (ns) results are denoted. In cynomolgus RBCs both A1-C.1 (P = 0.0040) and A1-H.1 (P = 0.0021) lines grew significantly better than A1-O. In human RBCs only A1-H.1 (P = 0.0285) showed increased growth rate relative to A1-O; in contrast A1-C.1 demonstrated a nonsignificant but consistent reduction in growth relative to A1-O in human RBCs. (B and C) Parasitemia was determined after invasion (at 7 h) and after a cycle of intracellular growth (at 24 h) in cynomolgus (B) or human (C) RBCs, revealing that growth rates were determined by invasion efficiency and not by differences in subsequent intracellular growth. The data shown are the means of triplicate data points from a single experiment ± 1 SD.

The NBPXa Gene Can Be Disrupted in A1-H.1 Parasites Grown in Cynomolgus but Not in Human RBCs.

To analyze the role of NBPXa, A1-H.1 parasites were transfected with a gene-disruption construct (Fig. 3A) and were maintained under drug selection in either human or cynomolgus RBCs until resistant parasites emerged (8–15 d). Analysis of these parasites by PCR revealed successful disruption of NBPXa in parasites from independent transfections maintained in cynomolgus RBCs (n = 5) but not in parasites grown in human RBCs (n = 3), despite successful integration of control plasmid (Fig. 3B). Thus, the NBPXa gene appeared to be refractory to disruption in parasites grown in human but not in cynomolgus RBCs, indicating a host species-specific requirement for this gene. Two independent ΔNBPXa cloned lines were derived from A1-H.1 parasite transfections in cynomolgus RBCs and were used for subsequent phenotypic analysis.

Fig. 3.

Disruption of NBPXa can be achieved when parasites are grown in macaque RBCs but not in human RBCs. (A) Disruption of NBPXa by the introduction of linearized pNBPXa_1XKO into the genome via integration at the 5′ end of the NBPXa locus within a targeting region (hatched box). Primers for diagnostic analysis are indicated by thick black arrows. (B) PCR analysis of independently transfected parasites grown in human (T1 to T3) or cynomolgus (T1 to T5) RBCs and of two independent clones derived from the latter (T1.1 and T2.1) confirmed correct integration only for parasites grown in cynomolgus RBCs. (Upper) Amplification of a product from wild-type NBPXa (primers ol1 and ol2). (Lower) Amplification of a product from a species that can be produced only following integration (primers ol1 and ol3).

NBPXa Is Unnecessary for Invasion of Cynomolgus RBCs but Is Crucial for Invasion of Human RBCs.

We examined the ability of the ΔNBPXa clones to invade and grow within either human or cynomolgus RBCs in the 24-h invasion assay. Disruption of NBPXa had no effect on invasion and growth in cynomolgus RBCs, but invasion of human RBCs was significantly impeded, matching the phenotype of the A1-C.1 line (Fig. 4A and Fig. S3). The defect in human RBC invasion caused by the loss of the NBPXa gene was severe and prevented multiplicative growth, but a low frequency of successful invasions was detected for both the A1-C.1 and ΔNBPXa lines. To rule out the possibility that this result was caused by contamination of purified schizonts with cynomolgus RBCs, the parasites were allowed to invade RBCs prelabeled with a fluorescent dye (26). The results showed that although contaminating cynomolgus RBCs contributed to a slight overestimation of invasion efficiency in human RBCs, only a small fraction of both the ΔNBPXa and A1C.1 parasites was able to infect human RBCs. Parasitemia measurements at 48 h showed that the ΔNBPXa human RBC invasion phenotype was expressed in subsequent cycles, and the parasite could not undergo multiplicative growth in human RBCs (Fig. 4 B and C). Extended culture of ΔNBPXa parasites in human RBCs resulted in a rapid reduction of parasitemia consistent with the severe defect. Although some parasites were detectable after 27 d of culture, these were revertants that had deleted the NBPXa disruptant construct (Fig. S3). These data demonstrate the severity of the ΔNBPXa phenotype, accounting for both the inability to disrupt the NBPXa gene and the severely reduced growth of the A1-C.1 line in human RBCs.

Fig. 4.

Disruption of NBPXa prevents multiplicative growth in human RBCs through impaired invasion. (A) Growth of A1-C.1, A1-H.1, and ΔNBPXa in cynomolgus or human RBCs after 24 h. Data shown are the mean of four independent experiments ± 1 SD. Statistical significance was determined using a two-factor ANOVA with the Sidak post hoc test to provide multiplicity-adjusted P values. Significant (*) and nonsignificant (ns) results are shown. Compared with A1-H.1, ΔNBPXa showed no significant difference in growth in cynomolgus but a significant reduction in human RBCs (*P = 0.0481). The growth of ΔNBPXa and A1-C.1 did not differ significantly in human or cynomolgus RBCs. (B and C) Parasite growth and invasion over two cycles in either cynomolgus (B) or human (C) RBCs revealed a growth defect at or around invasion. Data shown are the means of triplicate data points from a single experiment ± 1 SD. (D and E) Light microscopic images of Giemsa-stained thin blood films from A1-H.1 (D) and ΔNBPXa (E) parasites in human RBCs at 7 h. Black arrowheads indicate newly formed rings; red arrowheads indicate attached but uninvaded merozoites. (Scale bar, 5 µm.)

Fig. S3.

The ΔNBPXa phenotype in human RBCs is displayed by two independent clones. Growth in reticulocyte-enriched human RBCs is improved; there is no effect on the invasion of rhesus RBCs. (A) Attempts to readapt ΔNBPXa parasites to growth in human RBC were unsuccessful. No ΔNBPXa parasites were apparent after 4 d, but parasites reappeared at 28 d; PCR analysis revealed that only revertant to wild-type parasites had survived, and no ΔNBPXa parasites remained. PCR genotyping shows amplification of a product from wild-type NBPXa (Upper, primers ol1 and ol2) in the A1-H.1 line and 28 d ΔNBPXa revertants. A band that can only result from parasites with plasmid integration (Lower, primers ol1 and ol3) was seen only in the control ΔNBPXa starting population (clone T1.1). (B) A1-C.1, A1-H.1, and two independent ΔNBPXa clones grown for 24 h in cynomolgus or human RBCs reveal a significant defect in human RBC invasion by both ΔNBPXa clones. (C and D) The growth of A1-O, A1-C.1, and A1-H.1 (C) and A1-C.1, A1-H.1 and ΔNBPXa (D) in rhesus, cynomolgus, reticulocyte-enriched, and normal human RBCs showed no invasion phenotype for ΔNBPXa rhesus RBCs; a marginal increase in invasion efficiency was observed for all lines in reticulocyte-enriched human RBCs. Data shown are the mean of triplicate data points from a single experiment ± 1 SD.

Further analysis showed that the ΔNBPXa block occurred within the first 7 h of the assay, as is consistent with a defect in invasion rather than in subsequent intracellular growth (Fig. 4 B and C); examination of Giemsa-stained blood smears taken at 7 h showed that although A1-H.1 parasites had successfully developed as ring stages in human RBCs, most of the ΔNBPXa parasites appeared to be extracellular merozoites that had attached to but had not invaded RBCs (Fig. 4 D and E). To see whether this observation was consistent with a phenotype expressed subsequent to RBC binding, we used an immunofluorescence-based merozoite surface protein 1 (MSP1) processing assay (27) to distinguish merozoites that had successfully invaded and formed rings (MSP1₁₉-positive only) from those that had attached but failed to invade, (MSP1₃₃ and MSP1₁₉ double positive) (Fig. 5A). Although 92% of A1-H.1 merozoites successfully formed rings in human RBCs, 76% of ΔNBPXa merozoites attached but failed to invade human RBCs (Fig. 5B). In contrast, about 90% of merozoites from both lines successfully formed rings in cynomolgus RBCs (Fig. 5B). These results show that NBPXa disruption results in a block in invasion following initial merozoite attachment.

Fig. 5.

ΔNBPXa parasites attach to human RBCs but fail to invade. (A) A1-H.1 and ΔNBPXa schizonts were purified, and released merozoites were allowed to invade human or cynomolgus RBCs for 3 h. Panels show blood smears stained with anti-MSP1₃₃ (green), anti-MSP1₁₉ (red), and a merge of both MSP1 antibodies with anti-human RBC (magenta) and DAPI (blue). MSP1 is processed during invasion, with MSP1₃₃ being shed and MSP1₁₉ retained, distinguishing external attached merozoites (both MSP1₃₃- and MSP1₁₉-positive) from those that had successfully invaded to form rings (only MSP1₁₉-positive). Although A1-H.1 merozoites (Upper Row) were able to invade human RBC and form rings (white arrows), most ΔNBPXa merozoites (Lower Row) attached but did not invade human RBC (white arrowheads). (Scale bars, 10 µm.) (B) The parasite stages present in the immunofluorescence assays were quantified. The graph shows the number of free merozoites, RBC-attached merozoites, and rings stages expressed as a percentage of total merozoite-derived forms counted. Data shown are the means of triplicate cultures from a single experiment ± 1 SD. The numbers in parentheses indicate the average ring-stage parasitemia for each of the parasite/blood combinations.

To ensure that up- (or down-) regulation of another gene was not responsible for the observed phenotype, we performed RNA-sequencing (RNA-seq) analysis of schizont stages of A1-H.1 and ΔNBPXa parasites. This analysis did not identify any significantly differentially expressed genes.

Deletion of NBPXa Does Not Result in Restriction to Human Reticulocytes.

Previous work by others had suggested that the ability to invade and grow in human reticulocytes as well as normocytes is a key parasite adaptation to culture in human RBCs (22). Our observation that both the ΔNBPXa and A1-C.1 lines invaded a small percentage of human RBCs raised the possibility that the loss of NBPXa function restricted the parasite to invasion of reticulocytes, which typically constitute 0.5–2% of total RBCs. To investigate this possibility, we used an adapted protocol (28, 29) to enrich reticulocytes to around 10% of human RBCs and used these cells for invasion assays. Reticulocyte enrichment resulted in a small increase in invasion efficiency over normal human RBCs for the A1-O, A1-H.1, A1-C.1, and ΔNBPXa lines, suggesting that all lines favor younger RBCs (Fig. S3). However, these increases were insufficient to explain the ΔNBPXa phenotype as the result of reticulocyte restriction. Furthermore, the few successfully invasive ΔNBPXa parasites were present in both reticulocytes and normocytes.

NBPXa Is Unnecessary for Invasion of Rhesus RBCs.

Disruption of NBPXa had no effect on invasion into rhesus RBCs (Fig. S3), although the invasion efficiency for all lines tested (A1-H.1, A1-C.1, and ΔNBPXa) was significantly higher in cynomolgus RBCs than in rhesus RBCs, suggesting that the parasite has slightly different adhesin specificities even for these closely related hosts. However this difference also may be explained by phenotypic culture adaptation to cynomolgus RBCs.

Discussion

In this work we set out to identify parasite features that facilitate growth in human RBCs in vitro, exploiting P. knowlesi lines that we previously adapted to growth in either human or cynomolgus RBCs. Comparative genome analysis of the culture-adapted lines revealed relatively few nonsynonymous SNPs, and none of these associated universally with either adaptation to human RBCs or long-term culture. In contrast, we detected several large genome deletions and a duplication that were present in either human- or cynomolgus-adapted lines but not in both and that were present in the uncloned lines as well as in clones derived from them, suggesting that selection had led to fixation within the population. Further analysis of these deletions either by individual gene deletion, as we have demonstrated for NBPXa, or by gene complementation may be informative. The duplication in A1-H–derived lines contains the gene coding for GTP cyclohydrolase, a rate-limiting enzyme in folate metabolism, and duplications of this gene are associated with drug resistance in South East Asian strains of P. falciparum (30). Increased activity in this pathway and metabolic adaptation may provide a general selective advantage for growth in culture.

Correct assembly and read mapping for several key gene families, in particular the DBPs, was only possible with the use of PacBio sequencing, which produces longer read lengths. Using these data as a scaffold to map the higher-coverage but shorter Illumina reads facilitated high-quality de novo assemblies of both A1H.1 and A1-C.2 cloned lines, which complement the existing H-strain data and will benefit further work with the culture-adapted P. knowlesi lines.

Our previous comparison of cynomolgus- and human-adapted P. knowlesi parasite lines had suggested that an increase in human RBC invasion efficiency enabled the parasite to adapt to long-term culture in human RBCs (21). However, the additional comparison with the invasion characteristics of the A1-O line, which had been in culture for less than 2 mo, demonstrated that the host cell specificity of the A1-H lines had remained unchanged. Instead the overall invasion efficiency for both human and cynomolgus RBCs had improved, with invasion of human RBCs being above the minimum threshold required to support in vitro multiplication and growth. In contrast, the invasion efficiency of the A1-C line (and its derived clones) increased only for cynomolgus RBCs. This increased selectivity is likely the result of the loss of a section of one end of chromosome 14 containing the gene for NBPXa, which we have demonstrated to be crucial for human RBC invasion. Given that the cynomolgus RBC invasion phenotypes of the A1-C.1 and A1-H.1 lines are indistinguishable, it is quite possible that if A1-C parasites had not lost NBPXa, their invasion efficiency for human RBCs in culture would have risen to levels equivalent to that seen for A1-H.1 without prior exposure to human RBCs. In principle this idea could be investigated by genetic complementation of the lost NBPXa gene within the A1-C.1 line, but so far such efforts have been hampered by its large size (8.5 kb). The reasons for the different behavior of the H-strain parasite [which was isolated from a patient (23)] in vivo and in vitro are unclear, but our results suggest that for successful parasite multiplication and survival in vitro, prolonged physiological adaptation to culture is more important than adaptation to a specific host RBC. Our data identified clear differences in invasion efficiencies into RBCs from rhesus and cynomolgus macaques; this finding is surprising, given that these two primate species are closely related (31) and that experimental infection of rhesus monkeys (not a natural host of P. knowlesi) results in hyperparasitemia (23), in contrast to the chronic low-level parasitemia characteristic of P. knowlesi infection in cynomolgus macaques. Recombinant NBPXb expressed on cells binds cynomolgus RBCs four times more effectively than rhesus RBC (16), offering one possible mechanism for this difference in invasion efficiency. Independently established culture-adapted parasite strains developed using rhesus RBCs (20, 22) may have undergone adaptive changes quite different from those observed for cynomolgus RBCs. It also will be of interest to determine the contribution of metabolic adaptation to growth in different RBCs in culture.

P. knowlesi is phylogenetically closely related to P. vivax (32), a malaria parasite that causes the majority of malaria cases outside of Africa and that currently cannot be cultured in vitro. Although both these parasites use an orthologous DBP to bind to the DARC receptor on host RBCs (6, 18, 33, 34), there are greater differences in the RBL family. P. vivax can invade only reticulocytes, a phenotype attributed to the presence of RBP1 and RBP2, which bind only to reticulocytes (10) and are not conserved in P. knowlesi. Orthologs of the P. knowlesi RBLs are present in P. vivax (35, 36), but, given that our data indicate the P. knowlesi NBPXa protein is functional in invasion of both normocytes and reticulocytes, it is tempting to speculate that the absence of this protein in P. vivax could account for the parasite’s restriction to reticulocytes. Intriguingly another closely related simian malaria parasite, P. cynomolgi, contains orthologs for RBL proteins specific for both P. vivax and P. knowlesi (11, 37). Transferring orthologs from P. vivax or P. cynomolgi into P. knowlesi may offer opportunities to study both the requirements for and the evolution of P. vivax invasion.

The unique advantages of the P. knowlesi culture system have enabled us to demonstrate that NBPXa plays a crucial role in the invasion of human RBCs. Its role appears to occur after initial merozoite attachment, perhaps during establishment of the moving junction between RBC and parasite plasma membranes, which is required for invasion. In other Plasmodium species both the RBL and DBP protein families are highly redundant, and ascribing precise roles has been difficult, although in P. falciparum they have been implicated in triggering signaling and release of secretory organelles (38). The ability to grow P. knowlesi interchangeably in different host cells with distinct adhesin requirements enables the generation of lines with gene knockouts for important parasite adhesins such as NBPXa, providing a powerful and unique opportunity to investigate their precise role in red cell invasion.

The adhesin repertoire is smaller in P. knowlesi than in many malaria parasites (9, 15, 18), and the loss of either DBPa (6, 19) or NBPXa creates a critical block in the invasion of human RBCs, perhaps because the paralogues cannot bind human RBCs. For example, although PkNBPXa and DBPa proteins bind to both human and macaque RBCs, DBPβ, DBPγ, and NBPXb bind only to macaque RBCs (16, 39). This apparent lack of adhesin redundancy for human RBC invasion suggests that P. knowlesi is vulnerable to perturbation of DBPα/NBPXa function. Although this lack of redundancy may encourage the use of these antigens in vaccine development, their diversity in the parasites responsible for human infection (25, 40) provides an important caveat. Vaccine development may be an important strategy for control of P. knowlesi malaria, given the presence of a primate reservoir in Southeast Asia.

NBPXa polymorphisms are important determinants of high parasitemia and disease severity in P. knowlesi infection (7) and may impact the potential for human-to-human transmission. Because we have shown a critical role for this protein in human RBC invasion, it is possible that the increased virulence and multiplication rates of parasites containing particular alleles result from an improved ability to bind and invade human RBCs. However, even the human-adapted P. knowlesi line invades cynomolgus RBCs more efficiently than human RBCs, suggesting that invasion efficiency may remain a bottleneck in human infections. Interestingly, variation in NBPXb, which had been found not to bind to human RBCs, also has been linked to disease severity in humans (7, 16). The role of NBPXb and variants of both P. knowlesi NBPs now can be investigated in vitro. Human-adapted P. knowlesi parasites provide a robust experimental platform to unravel the mechanisms of host cell invasion and to aid in the development of strategies to control malaria caused by P. knowlesi and the closely related pathogen P. vivax.

Methods

Macaque and Human RBCs.

Cynomolgus and rhesus blood was collected by venous puncture. Animal work was reviewed and approved by the local National Institute for Biological Standards and Control Animal Welfare and Ethical Review Body (the Institutional Review Board) and by the United Kingdom Home Office as governed by United Kingdom law under the Animals (Scientific Procedures) Act 1986. Animals were handled in strict accordance with the “Code of Practice Part 1 for the housing and care of animals (21/03/05)” available at https://www.gov.uk/research-and-testing-using-animals. The work also met the National Centre for the Replacement Refinement and Reduction of Animals in Research (NC3Rs) guidelines on primate accommodation, care, and use (https://www.nc3rs.org.uk/non-human-primate-accommodation-care-and-use), which exceed the legal minimum standards required by the United Kingdom Animals (Scientific Procedures) Act 1986, associated Codes of Practice, and the US Institute for Laboratory Animal Research Guide. Human blood (Groups A⁺ and O⁺; Duffy⁺) was obtained from the United Kingdom National Blood Transfusion Service.

Parasite Lines and Maintenance.

P. knowlesi parasite cultures were derived from a frozen in vivo stabilate (A1, dated July 22, 1975) supplied by Louis Miller, National Institutes of Health, Bethesda, MD, and were obtained following passage in vivo in rhesus macaque of the P. knowlesi H strain derived from a clinical infection (23). A1-O is a nonclonal cultured line from the A1 stabilate propagated in cynomolgus RBCs in vitro for less than 2 mo. A1-H is a nonclonal line derived from A1-O following maintenance in 80% human and 20% cynomolgus RBCs for 10 mo and for up to 7 mo in human RBCs alone, with clones A1-H.1, A1-H.2, and A1-H.3 derived at this time. P. knowlesi A1-C is a nonclonal line derived from A1-O after maintenance in 100% cynomolgus RBCs for 15 mo, with clones A1-C.1 and A1-C.2 derived at this time. Parasites were grown and synchronized as described previously (21).

P. knowlesi Genome Sequencing, Analysis, and de Novo Assembly.

DNA for the A1-H.1 and A1-C.2 lines was used for preparation of libraries: HiSeq 2000 (Illumina, Inc.) compatible genomic shotgun libraries (500-bp fragment size); PCR-free libraries sequenced on MiSeq (Illumina); and, for genome assemblies, 20-kbp insert libraries sequenced on a PacBio RS-II instrument (Pacific Biosciences Inc.). Genome assembly was performed using PacBio's FALCON tools and Genome Consensus suite. Annotation was transferred from the P. knowlesi H strain reference genome (24), and gene models were checked in Artemis BAM view and corrected using strand-specific RNA-seq datasets.

Adhesion, Invasion, and Growth Assays.

Purified schizonts were used to set up triplicate cultures with cynomolgus, rhesus, and human or human reticulocyte-enriched RBCs, at a 2% hematocrit and 1% parasitemia. All parasites (including A1-H.1) were grown in cynomolgus RBCs for at least 1 mo before growth assays to control for any effect of prior host cell type on subsequent invasion and growth. Parasitemia was measured with a flow cytometry (FACS)-based assay (21, 26) at before and after incubation at 37 °C in a gassed chamber for 7, 24, or 48 h. Samples were fixed, washed, permeabilized with Triton X-100, and then washed again before RNase treatment, staining with SYBR Green I (Life Technologies), and FACS analysis. To control for invasion of RBCs carried over during schizont purification, target cells were prelabeled with a far red fluorescence dye (26) using 7-hydroxy-9H-(1,3-dichloro-9,9-dimethylacridin-2-one) succinimidyl ester (DDAO-SE) (Life Technologies). Following washing the samples were analyzed on a CyAN ADP Analyzer (Beckman Coulter). Data were acquired and analyzed using CellQuest (BD) or HyperCyt (Beckman Coulter) software.

To investigate invasion/binding phenotype and reticulocyte preference, samples taken at 7 h were incubated with reticulocyte stain solution (Sigma). Then the cells were pelleted, and Giemsa-stained thin blood films were prepared. To attempt to readapt ΔNBPXa parasites to growth in human RBC, a culture at 1% parasitemia was set up and maintained for 1 mo. After parasites were no longer detectable in blood smears (4 d), the culture was maintained by replacement with 30% fresh human RBCs and medium every 2–3 d. The (revertant) parasites that were detectable after 28 d were analyzed by PCR together with A1-H.1 and ΔNBPXa controls.

The difference between adhesion and invasion of merozoites was also measured using an immunofluorescence assay. Synchronous P. knowlesi A1-H.1 or ΔNBPXa schizonts were added to triplicate cultures at 2% parasitemia with cynomolgus or human RBC, and after 3 h thin blood films were prepared for immunofluorescence staining. The number of free merozoites (MSP1₁₉-positive, MSP1₃₃-positive, no contact with RBC), attached merozoites (MSP1₁₉-positive, MSP1₃₃-positive, in contact with RBCs), rings (MSP1₁₉-positive, MSP1₃₃-negative, within RBCs), and RBCs were counted. A minimum of 1,000 RBCs was examined.

DNA Constructs and PCR.

The DNA construct (NBPXa_1XKO) for targeted disruption of NBPXa (GeneDB ID PKNH_1472300) via single–cross-over homologous recombination was adapted from a precursor to the PkconGFP_ep vector (21) and used a 1.2-kb region of the 5′ end of the NBPXa gene to target the NBPXa locus, with linearization for transfection using the BsiWI site situated in the NBPXa fragment.

Transfection, Cloning, and Genotyping.

Parasites were transfected with NBPXa_1XKO or with plasmid PkconGFPp230p (21) as a positive control, and fresh human or cynomolgus RBCs were added every 4 d until parasites were visible. Parasites were cloned by limiting dilution, a process requiring 9 or 14 d in cynomolgus or human RBCs, respectively, and then were expanded for genotypic and phenotypic analysis. Transfected parasites were analyzed by diagnostic PCR with two different primer sets. The first was designed to amplify a 1,478-bp fragment only from the intact wild-type NBPXa locus (ol1 and ol2), and the second was designed to amplify a 1,306-bp fragment after correct integration of NBPXa_1XKO (ol1 and ol3). PCR reactions were carried out using GoTaq Master Mix (Promega) using the following cycle conditions: 3 min at 96 °C, then 30 cycles of 25 s at 96 °C, 25 s at 52 °C, and 2 min at 64 °C, and a final extension of 5 min at 64 °C.

SI Methods

Macaque, Human, and Reticulocyte-Enriched Blood.

Cynomolgus and rhesus blood was collected by venous puncture into K2 EDTA vacutainers (Becton Dickinson). Human blood (Groups A⁺ and O⁺; Duffy⁺) was obtained from the United Kingdom National Blood Transfusion Service and was used within 2 wk of receipt for all growth invasion assays. Reticulocyte enrichment was achieved using a protocol adapted from a previous method (28, 29), which uses a high K⁺ buffer to prevent reticulocyte dehydration and allows density-based enrichment from normal blood. Human blood cells were washed twice in high K⁺ buffer (1× K⁺ buffer; 20 mM K⋅Hepes, 1 mM MgCl₂, 1 mM NaH₂PO₄, 10 mM glucose, 0.5 mM EGTA, 5 mM NaCl, 116 mM KCl). The washed cells were passed through a sterile CF11 cellulose column (Whatman) under gravity to remove white blood cells, and the column was washed with one volume of K⁺ buffer. Reticulocytes were enriched by centrifugation at 900 × g for 12 min on a cushion of 65% Percoll (GE Healthcare) [stock solution (90% Percoll, 10% 10× K⁺ buffer) diluted into 1× K⁺ buffer]. The RBCs at the layer interface were collected and washed once with K⁺ solution. Ten microliters of RBCs were incubated with 50 µL of reticulocyte stain solution (Sigma) for 15 min; then reticulocytemia was assessed by microscopy on thin blood films, counting reticulocytes (which contain two or more blue-stained granules) and total RBCs.

Parasite Lines and Maintenance.

All P. knowlesi parasite cultures were derived from a frozen in vivo stabilate (A1, dated July 22, 1975) supplied by Louis Miller, National Institutes of Health, Bethesda to Guys Hospital Medical School and obtained following passage in vivo in the rhesus macaque of the H strain of P. knowlesi that was derived from a clinical infection (23). A1-O is a nonclonal line adapted to culture from the A1 stabilate, which had been propagated in cynomolgus RBCs in vitro for less than 2 mo. A1-H is a nonclonal line derived from A1-O following maintenance in 80% human and 20% cynomolgus RBCs for 10 mo and for up to 7 mo in human RBCs alone, with clones A1-H.1, A1-H.2, and A1-H.3 derived at this time. P. knowlesi A1-C is a nonclonal line derived from A1-O after maintenance in 100% cynomolgus RBCs for 15 mo, with clones A1-C.1 and A1-C.2 derived at this time. Parasites were grown in complete medium comprising RPMI 1640 (Life Technologies) with the following modifications: 2.3 g/L sodium bicarbonate, 4 g/L dextrose, 5.957 g/L Hepes, 0.05 g/L hypoxanthine, 5 g/L AlbuMAX II (Thermo Fisher Scientific), 0.025 g/L gentamycin sulfate, 0.292 g/L l-glutamine, and 10% serum. For routine parasite culture 10% (vol/vol) equine serum (Life Technologies) was used, but for sensitive procedures, including growth assays and transfections, human AB⁺ serum was used. Parasites were grown and synchronized as described previously (21).

P. knowlesi Genome Sequencing, Analysis, and de Novo Assembly.

After passage through a Plasmodipur filter (EuroProxima) to remove white blood cells, DNA was purified from schizont-stage parasites using the phenol/chloroform extraction method. DNA quality was highest for the A1-H.1 and A1-C.2 lines, so these were selected for the generation of large insert libraries. HiSeq 2000 (Illumina) compatible genomic shotgun libraries (500-bp fragment size) were prepared according to the manufacturer’s protocols (Illumina). Additional PCR-free libraries were prepared and sequenced on MiSeq (Illumina). Poor-quality bases from the raw reads were trimmed using Trimmomatic (www.usadellab.org/cms/?page=trimmomatic). For genome assemblies, 20-kbp insert template libraries were prepared for single-molecule real-time (SMRT) sequencing chemistry at Macrogen, Inc. Samples were sequenced on a PacBio RS-II instrument (Pacific Biosciences). Genome assembly was performed using PacBios’ FALCON tools (https://github.com/PacificBiosciences/FALCON), and consensus scaffolding was achieved through the Genome Consensus suite (https://github.com/PacificBiosciences/GenomicConsensus). PacBio sequencing generated 361,197 and 355,195 reads from three SMRT cells each for PKNA1-C.2 and PKNA1-H.1 genomic DNA, respectively. Reads with a mean length just above 8 kb were generated to obtain a theoretical coverage of 122.6-fold (for PKNA1-C.2) and 124.1-fold (for PKNA1-H.1) of the genomes. Genome assembly resulted in 45 and 37 contigs for PKNA1-C.2 and PKNA1-H.1, respectively, and after PCR-free Illumina read correction the total assembly sizes were 24,393,488 bp and 23,988,125 bp, respectively. Consensus scaffolds were error-corrected using the PCR-free Illumina sequence, and annotation was transferred from the reference genome of the P. knowlesi H strain (24) using the Rapid Annotation Transfer Tool, RATT (41). Gene models were checked in Artemis BAM view where flawed models (containing in-frame stop codons or lacking start or end codons) were corrected using strand-specific RNA-seq datasets from schizont stages of the human RBC-adapted parasite PKA1-H.1. RNA-seq analysis of schizont stages of A1-H.1 and ΔNBPXa parasites was performed to identify differential gene expression (three biological replicates for each, P value of 0.05 corrected for false discovery rate). Significance was determined using cuffdiff and CummeRbund (bioconductor.org/packages/release/bioc/html/cummeRbund.html).

The PacBio sequencing, assembly, and annotation information for both genomes is summarized in Table S1. Genomic deletions were determined by cross-mapping the reads against the reciprocal assembly and against the reference H strain (24) using the Burrows–Wheeler Alignment tool. GATK pipeline (Broad Institute) was used to perform variant calling with the corrected assemblies as references. GATK-based filtering was applied with the following parameters: quality-to-depth ratio of 5 and Fisher strand value of 60 for SNPs and 200 for indels. Variants having an allele frequency of 80% were retained for further analysis. Additional variant filtering and comparisons were performed using VCFtools (https://github.com/vcftools/vcftools). Natural variation (SNP frequency) in the five invasion-related genes NBPXa (PKNA1_H1_1472300), NBPXb (PKNA1_H1_0700200), DBLα (PKNA1_H1_0623500), DBLβ (PKNA1_H1_1400800), and DBLγ (PKNA1_H1_1356900) was determined using the Illumina raw sequence datasets from 48 P. knowlesi infections (25) compared with the reference PKA1-H.1 sequences using the GATK pipeline, and the SNPs were visualized using ggplot2 (ggplot2.org). A collection of three tools was used for computational analysis of copy number variation in addition to manual examination of the Illumina reads mapped onto the genome assembly of P. knowlesi H strain: Control-FREEC (42) was used with a permissive breakpoint threshold of 0.35 and a coefficient of variation of 0.07; CNVnator (43) was used with a window size of 150; and GROM-RD (44) was run with a P value of 0.001, a duplication coverage threshold of 3, and a sampling rate of 15.

Antibody Production.

The mouse P. knowlesi MSP1₃₃-specific antibody has been described (27). A P. knowlesi MSP1₁₉-specific antibody was produced following rabbit immunization with recombinant MSP1₁₉. The DNA for a synthetic codon-optimized MSP1₁₉ sequence including a C-terminal His tag was amplified using primers pGEX-MSP1₁₉ for (5′-AATTATatcctgAACATGGCTTCCGCTCAC-3′) and rev (5′-AATTATgaattcCTACTAGTGGTGGTGATGGTGATGAGAGG-3′) (restriction enzyme sites are in lowercase). The PCR product was cloned into pGEX3X via BamHI/EcoRI cleavage, and pGEX3XPkMSP1₁₉ was transformed into Escherichia coli BL21Gold and was grown in LB broth to an OD at 600 nm of 0.6, before induction with 0.5 mM isopropyl β-d-1-thiogalactopyranoside for 3 h at 37 °C. Cell pellets were solubilized in BugBuster (Novagen), and the clarified bacterial lysate was incubated with 2.5 mL of glutathione-Sepharose 4B (GE Healthcare) at 4 °C overnight. The Sepharose resin was washed with 100 mL cold PBS followed by 3 mL of resuspension buffer [50 mM Tris⋅HCl (pH 8.0), 150 mM NaCl] before protein elution with 15 mM reduced glutathione in resuspension buffer. The eluted protein was dialyzed against PBS, and 5 mg was rebound to 2.5 mL of glutathione Sepharose beads. The beads were washed using cleavage buffer [50 mM Tris⋅HCl (pH 8.0), 100 mM NaCl, 5 mM CaCl₂] before on-bead protein digestion with three units per microliter of Factor Xa (Novagen) overnight at 20 °C. Eluted recombinant MSP1₁₉ lacking GST was concentrated and buffer-exchanged into PBS using a Vivaspin concentrator (Sartorius). Polyclonal rabbit anti-MSP1₁₉ antibodies were produced commercially (Harlan Laboratories) using 200 µg of MSP1₁₉ per inoculation. Rat anti-human RBC ghost antibody was prepared using A⁺ human erythrocytes (leukocyte depleted) lysed in 10-times the cell volume of cold 5-mM sodium phosphate before collection of ghosts by centrifugation at 12,000 × g for 20 min at 4 °C. The ghost pellet was washed three times in sodium phosphate buffer with centrifugation until all visible hemoglobin had been removed. The pellet was suspended in PBS, and 25 µg of total protein at 250 µg/mL was supplied to Harlan Laboratories for antibody production in rats following the company’s 35-d protocol. The resulting antiserum could be used for immunolabeling both human and cynomolgus RBC because of cross-reactivity.

Adhesion, Invasion, and Growth Assays.

A flow cytometry (FACS)-based assay adapted from previous reports (21, 26) was used to determine parasitemia in growth and invasion assays. Schizonts were purified and used to set up triplicate cultures in 12-well plates with cynomolgus, rhesus, human, or human reticulocyte-enriched blood, with 2 mL medium per well at 2% hematocrit and a schizont parasitemia of 1%. All parasites (including A1-H.1) were grown in cynomolgus RBCs for at least 1 mo before growth assays to control for any effect of prior host cell type on subsequent invasion and growth.

Samples of each culture (40 µL) were taken for FACS analysis before and after incubation at 37 °C in a gassed chamber for 7, 24, or 48 h. The culture samples were transferred to a 96-well microtiter plate and were fixed using a solution of 2% (wt/vol) paraformaldehyde (Sigma) and 0.2% glutaraldehyde (Sigma) for 1 h at 4 °C. Samples were washed with PBS, permeabilized with 0.3% (vol/vol) Triton X-100 (Sigma) in PBS for 10 min, and then washed again twice with PBS, all at room temperature. Following RNase treatment for 1 h at 37 °C with 0.5 mg/mL ribonuclease A (MP Biomedicals) in PBS, the samples were washed again, stained for 30 min with 1:5,000 SYBR Green I (Life Technologies) in PBS, diluted into 500 μL PBS, and transferred to FACS tubes. Samples were analyzed on a FACSCalibur (BD), with the exception of the DDAO-SE–labeled experiment, which was analyzed on a CyAN ADP Analyzer (Beckman Coulter). Data were acquired and analyzed using CellQuest (BD) or HyperCyt (Beckman Coulter) software. The graphs detailing 24-h growth rates in Figs. 2A and 4A use data from four independent experiments (carried out on 4 different days, using different blood samples and independent parasite preparations). Statistical significance was determined using a two-factor ANOVA with the Sidak post hoc test to provide multiplicity-adjusted P values. Other graphs show results from triplicate cultures carried out in a single representative experiment. To control for invasion of any RBCs carried over during schizont purification, target cells were prelabeled with a far-red fluorescence dye using a protocol previously described (22). A suspension in RPMI 1640 of either cynomolgus or human RBCs (4% hematocrit) was labeled for 2 h at 37 °C with 10 µM DDAO-SE (Life Technologies). The suspension was washed in complete RPMI medium 1640, then was washed again, and the sample was incubated at 37 °C for 30 min. The samples then were washed two more times with complete medium and were used for the assays within 24 h.

In attempts to readapt ΔNBPXa parasites to growth in human RBC, a 50-mL culture at 1% parasitemia was set up and maintained for 1 mo. After parasites were no longer detectable in blood smears (4 d), the culture was maintained by replacement with 30% fresh human RBCs and medium every 2–3 d. The (revertant) parasites that were detectable after 28 d were analyzed by PCR together with A1-H.1 and ΔNBPXa controls, as described below.

To investigate invasion/binding phenotype and reticulocyte preference, a 100-μL sample was taken from each culture at 7 h and incubated for 20 min at 37 °C with 200 μL reticulocyte stain solution (Sigma). The cells then were pelleted for 1 min at 3, 500 × g, and thin blood films were prepared. Dried films were fixed with methanol for 30 s and were counterstained with Giemsa stain for 15 min.

The difference between adhesion and invasion of merozoites was also measured using an immunofluorescence assay. Synchronous P. knowlesi A1-H.1 or ΔNBPXa schizonts were purified and used in triplicate cultures at 2% parasitemia in six-well plates with cynomolgus or human blood (5 mL medium, 2% hematocrit). After 3 h under standard conditions, thin blood films were prepared for each sample and placed at −20 °C before immunofluorescence staining. Frozen blood films were thawed to room temperature, fixed with 4% (wt/vol) paraformaldehyde (Sigma) in PBS for 30 min, washed with PBS, and then permeabilized with 0.1% (vol/vol) Triton X-100 (Sigma) in PBS for 10 min. Films were washed with PBS and then were incubated with blocking solution [3% (wt/vol) BSA (Sigma) in PBS] overnight at 4 °C.

All primary and secondary antibodies (Thermo Fisher Scientific) were diluted in blocking solution and incubated sequentially for 1 h at room temperature followed by two washes in PBS for 15 min each. Slides were stained in the following order: mouse anti-PkMSP1₃₃ antibody (diluted 1:200), anti-mouse IgG Alexa Fluor-488 (diluted 1:50,000), rabbit anti-PkMSP1₁₉ (diluted 1:5,000), anti-rabbit IgG Alexa Fluor-594 (diluted 1:50,000), rat anti-human RBC (diluted 1:2,000), and anti-rat IgG Alexa Fluor-647 (diluted 1:50,000). After the last antibody incubation, the film was washed in PBS for 30 min and was mounted in ProLong Antifade mountant with DAPI (Thermo Fisher Scientific).

Slides were imaged using a Leica DMi8 epifluorescence microscope and analyzed using the LASX software suite (Leica). The number of free merozoites (MSP1₁₉-positive, MSP1₃₃-positive, no contact with RBC), attached merozoites (MSP1₁₉-positive, MSP1₃₃-positive, in contact with RBCs), rings (MSP1₁₉-positive, MSP1₃₃-negative, within RBCs), and RBCs were counted. A minimum of 1,000 RBCs were examined, and data files were blinded before counting. The counts were used to determine the total percentage of each merozoite-derived form (free, attached, ring), and ring/RBC counts were used to determine the ring-stage parasitemia. Residual schizont parasitemia also was quantified and ranged from 0.34 to 0.65%.

DNA Constructs and PCR.

The DNA construct for targeted disruption of NBPXa (GeneDB ID PKNH_1472300) via single–cross-over homologous recombination was adapted from a precursor to the PkconGFP_ep vector (21), which contains an hdhfr cassette driven by the Pkef1α promoter. To target the plasmid to the NBPXa locus, a 1.2-kb region of the 5′ end of the NBPXa gene, commencing 565 bp downstream of the start codon, was amplified by PCR using oligonucleotide primers ol301: CATGAGcccgggAGTACTAAGGATGCAAATAATAGGCTTAATGG and ol302: CATGAGagatctTTTCTGACTATGATAACTTCTACATTTTCATTCC (incorporated restriction sites are indicated in bold and lowercase letters). This fragment was cloned into the construct using the XmaI and BglII sites to form the NBPXa_1XKO construct. The NBPXa_1XKO construct was linearized for transfection using the BsiWI site situated in the middle of the NBPXa fragment.

Transfection and Dilution Cloning of P. knowlesi.

Tightly synchronized mature schizonts were transfected as described previously (21) using the Amaxa 4D electroporator (Lonza) and the P3 Primary cell 4D Nucleofector X Kit L (Lonza). After 24 h, and at daily intervals, the medium was replaced with fresh medium containing 100 nM pyrimethamine (Sigma). Fresh human or cynomolgus RBCs (50 µL) were added every 4 d until parasites were visible in Giemsa-stained blood smears. For each transfection experiment, a positive control plasmid (PkconGFPp230p) (21) was used and demonstrated successful integration for parasites grown in either blood type. The integration and genotyping strategies for this control plasmid have been detailed previously (21). Parasites were cloned by limiting dilution as described previously (21), a process requiring 9 or 14 d in cynomolgus or human RBCs, respectively, and then were expanded for genotypic and phenotypic analysis.

Transgenic P. knowlesi Genotyping.

Parasites transfected with the NBPXa_1XKO construct were analyzed by diagnostic PCR. DNA was prepared using a DNeasy blood and tissue kit (Qiagen). The diagnostic PCR used two different primer sets. The first was designed to amplify a 1,478-bp fragment only from the intact wild-type NBPXa locus (ol1: CACATTTGGAAGGAAGCCATCACGC and ol2: GCACTGTATTCATTTTGCTATTTACTTGTTCC). The second primer pair was designed to amplify a 1,306-bp fragment after correct integration of the NBPXa_1XKO construct into the NBPXa locus (ol1: CACATTTGGAAGGAAGCCATCACGC and ol3: CGTATTTTTATTAAGGGGCATATGCAAGGG). PCR reactions were carried out using GoTaq Master Mix (Promega) with ∼50 ng of gDNA template and 300 nM of each primer, using the following cycle conditions: 3 min at 96 °C, then 30 cycles of 25 s at 96 °C, 25 s at 52 °C, and 2 min at 64 °C, and a final extension of 5 min at 64 °C.