Plasmodium-specific molecular assays produce uninterpretable results and non-Plasmodium spp. sequences in field-collected Anopheles vectors.

The Malaria Research and Reference Reagent Resource-recommended PLF/UNR/VIR polymerase chain reaction (PCR) was used to detect Plasmodium vivax in Anopheles spp. mosquitoes collected in South Korea. Samples that were amplified were sequenced and compared with known Plasmodium spp. by using the PlasmoDB.org Basic Local Alignment Search Tool/n and the National Center for Biotechnology Information Basic Local Alignment Search Tool/n tools. Results show that the primers PLF/UNR/VIR used in this PCR can produce uninterpretable results and non-specific sequences in field-collected mosquitoes. Three additional PCRs (PLU/VIV, specific for 18S small subunit ribosomal DNA; Pvr47, specific for a nuclear repeat; and GDCW/PLAS, specific for the mitochondrial marker, cytB) were then used to find a more accurate and interpretable assay. Samples that were amplified were again sequenced. The PLU/VIV and Pvr47 assays showed cross-reactivity with non-Plasmodium spp. and an arthropod fungus (Zoophthora lanceolata). The GDCW/PLAS assay amplified only Plasmodium spp. but also amplified the non-human specific parasite P. berghei from an Anopheles belenrae mosquito. Detection of P. berghei in South Korea is a new finding.

The Malaria Research and Reference Reagent Resource1 recommends a nested polymerase chain reaction (PCR) specific for the 18S small subunit ribosomal DNA (ssrDNA) gene fragment for the detection of human Plasmodium species in Anopheles spp. mosquito vectors.2 This assay was first designed to screen human blood for Plasmodium spp.,3,4 and was later modified to screen mosquito vectors for Plasmodium spp. DNA.2 The modifications were based on the results of a comparison of extraction techniques designed to mitigate the issue of inhibitors to PCR2,5 in the DNA extraction method.1 The same primers (PLF/UNR/VIR) were used in both studies.2,3

The PLF/UNR/VIR assay was used to test for P. vivax in mosquito vectors (Table 1). Anopheles mosquitoes were collected bi-weekly (3–4-day intervals) by using Mosquito Magnet® traps (Pro-Model; American Biophysics Corp., Greenwich, RI) during August–September 2010 in South Korea. Mosquito collections were conducted in the Demilitarized Zone adjacent to the Military Demarcation Line separating North Korea from South Korea and at Warrior Base and Tongilchon located approximately 3 km from the Demilitarized Zone where malaria transmission was suspected.6 Anopheles spp. females were placed individually in 2-mL cryovials, dried, and shipped to the Walter Reed Biosystematics Unit (Suitland, MD).

Table 1

Primers, targets, and fragment sizes used for detecting Plasmodium species

Primer	Locus	Sequence 5′→3′	Specificity	Size of product, basepairs
GDCW 2	cytB	CGGTCGCGTCCGGTAGCGTCTAATGCCTAGACGTATTCCTGATTATCCAG	Plasmodium sp.	–
GDCW4	cytB	CGCATCACCTCTGGGCCGCGTGTTTGCTTGGGAGCTGTAATCATAATGTG	Plasmodium sp.	–
PLAS1	cytB	GAGAATTATGGAGTGGATGGTG	Plasmodium sp.	815
PLAS2	cytB	TGGTAATTGACATCCAATCC	Plasmodium sp.	815
PLF	18S	AGTGTGTATCAATCGAGTTTC	Plasmodium sp.	821 for P. vivax; 787 for P. falciparum
UNR	18S	GACGGTATCTGATCGTCTTC	Plasmodium sp.	–
VIR	18S	AGGACTTCCAAGCCGAAGC	P. vivax	499
PLU5	18S	CTTGTTGTTGCCTTAAACTTC	Plasmodium sp.	1,200
PLU6	18S	TTAAAATTGTTGCAGTTAAAACG	Plasmodium sp.	–
VIV1	18S	CGCTTCTAGCTTAATCCACATAACTGATAC	P. vivax	120
VIV2	18S	ACTTCCAAGCCGAAGCAAAGAAAGTCCTTA	P. vivax	–
Pvr47-F	Pvr47	CTGATTTTCCGCGTAACAATG	P. vivax	333
Pvr47-R	Pvr47	CAAATGTAGCATAAAAATCYAAG	P. vivax	–

The head and thorax were separated from the abdomen from individual mosquitoes to isolate only sporozoite-infected (salivary glands) mosquitoes. Total genomic DNA was extracted by using phenol-chloroform extraction using the Autogen automated DNA extraction robot (AutoGen Inc., Holliston, MA) and eluted in 50 μL of buffer in a 96-well plate format. Mosquitoes were identified to species by sequencing the internal transcribed spacer region 2 and a sequence comparison to voucher specimens available in the National Center for Biotechnology Information (NCBI) (Bethesda, MD) database. Of the mosquitoes tested, 56% were An. kleini, 27% were An. belenrae, 15% were An. sinensis, and 2% were An. pullus. For P. vivax detection, the PCR master mixture contained 1× buffer, 0.4 μM of each primer, 0.1 mM of each dNTP, 1.5 mM MgCl2, 5% dimethyl sulfoxide, 1 unit of Biolase Taq, and 1 μL of DNA template. The total reaction volume was 20 μL. The same master mixture was used in both rounds of amplification (nested PCR), and 1 μL of PCR template was used in the second reaction. For each PCR, a new master mixture was created to mitigate issues with a single batch and to minimize contamination of the master mixture. The cycling parameters were 94°C for 2 minutes; followed by 35 cycles of 94°C for 30 seconds, 62°C for 30 seconds, and 72°C for 1 minute; and a final extension at 72°C for 7 minutes. The same cycling parameters were used in the second PCR with an increase to 40 cycles.

The PCR amplicon was subjected to electrophoresis on a 1.5% agarose gel stained with ethidium bromide in the same 96-well format as the plate layout of the DNA extraction for quick interpretation. Gels were photographed to record the results. Of the 94 individual mosquitoes tested for P. vivax, 20 produced amplicons after PCR, although results were ambiguous (Figure 1). Only six of the amplified samples produced bands that were approximately 499 basepairs, the expected size for this assay. All amplified samples were cleaned by using ExoSapIT (Affymertix Inc., Santa Clara, CA) and sequenced in both directions. Sequences were trimmed, cleaned, and aligned using Sequencher software (Sequencher V5.1; Gene Codes Co., Ann Arbor, MI). Sequences in which 90% of the base calls were quality calls were run through the Plasmodium full genome database PlasmoDB (http//:PlasmoDB.org). The quality of a base call is defined by peak height and peak separation.7 Samples that did not significantly match any Plasmodium species in the database were then run through a standard Basic Local Alignment Search Tool (BLAST)/n search at the NCBI website.

Figure 1.

Agarose gel electrophoresis showing polymerase chain reaction amplification by A, semi-nested primers PLF/UNR/VIR; B, nested primers PLU5/PLU6/VIV1/VIV2; C, single-round Pvr47 F/R primers; and D, cytB nested primers GDCW2/GDCW4/PLAS1/PLAS2, for field collected Anopheles spp. mosquitoes. Pos. = positive; Neg. = negative.

Only 9 of the 20 samples produced a quality sequence. Of these samples, none of the sequences were of the expected size of 499 basepairs once they were trimmed (Table 2). When sequences were run against the known Plasmodium spp. in the PlasmoDBs BLAST database, none of the nine sequences matched any species of Plasmodium with a percent match > 70%. When the same samples were run through the NCBI BLAST/n tool, no significant matches were identified. However, a positive control of P. vivax extracted from An. dirus8 significantly match P. vivax in the PlasmoDBs BLAST/n tool (100% match), as well as in the NCBI BLAST/n tool (100% match), showing a successful amplification and sequencing reaction. The negative control of water produced no amplicon, suggesting that amplification products were not caused by laboratory contaminants.

Table 2

Sequencing results of amplified bands in field-collected Anopheles spp. mosquitoes*

Primer	Sample	Length, nucleotides	QC, %	Result PlasmoDB	% Match	P	Result NCBI	%	Coverage, %
PLAS 1/2	+ Control	629	99	Plasmodium vivax	100	2.30 × 10–9	P. vivax	100	100
PLAS 1/2	10G	663	100	P. berghei	98	9.20 × 10–5	P. berghei
VIV1/2	+ Control	210	98	P. vivax	99	9.50 × 10–34	P. vivax	99	100
VIV1/2	1E	170	92	P. vivax	64	0.057	Zoophthora lanceolata	99	100
VIV1/2	2C	241	93	P. vivax	86	0.042	Z. lanceolata	97	100
VIV1/2	2D	238	90	P. vivax	72	0.015	Z. lanceolata	98	100
PLF/VIR	+ Control	593	99	P. vivax	100	5.90 × 10–9	P. vivax	100	100
PLF/VIR	3E	522	97	P. vivax	67		None
PLF/VIR	6C	119	95	None			None
PLF/VIR	10B	283	98	None			None
PLF/VIR	16	352	100	P. vivax	67	0.00024	None
PLF/VIR	1B	371	100	None			None
PLF/VIR	1E	240	95	P. chabaudi	63	0.76	None
PLF/VIR	2B	318	98	None			None
PLF/VIR	2C	271	99	None			None
PLF/VIR	2D	246	98	P. knowlesi	67	0.9999	None
Pvr47 F/R	+ Control	116	86	P. vivax	99	1.50 × 10–13	NA
Pvr47 F/R	1C	498	95	None			NA
Pvr47 F/R	1H	299	98	None			NA
Pvr47 F/R	2H	543	97	P. falciparum	54	0.2	NA
Pvr47 F/R	5E	474	96	None			NA
Pvr47 F/R	8H	360	96	P. berghei	62	0.0067	NA
Pvr47 F/R	9G	146	91	P. yoelii	69	0.95	NA

QC = query coverage; NCBI = National Center for Biotechnology Information; NA = not available. Values in bold are statistically significant.

When the primers were run through the NCBI database, the PLF primer had a 100% query coverage (QC) and 100% match identity (MI) to a Hepatocyst sp. (HQ605039.1). The UNR primer aligned with species in multiple phyla including Lophotrochozoa (AB679345), Rhodophyta (JX828192), Alveolata (JQ178269), Cnidaria (AB693054), and uncultured fungus (KC218924) with 100% QC and 100% MI. The VIR primer was also significantly similar to species in many phyla, which included Chordata (XM002609218) and Viridiplanta, with 100% MI with 89% QC; several types of fungus, including Myceliophthora thermophila (XM003662658) and Metarhizium anisopliae (AJ251965) with 100% MI and 89% QC; Ecdysozoa (XM002633807) with 84% QC and 100% MI; and several bacterium, including Cryptococcus neoformans (XM771947) and several uncultured bacterium (JQ818134, JQ818132, and JQ818120) with a 100% MI and 78% QC.

Three additional PCRs were then tested for detection of P. vivax in field-collected female mosquitoes. We tested a second 18S ssrDNA assay (PLU/VIV) that was originally designed to screen human blood for parasites,9 a nested PCR specific for mitochondrial cytB (GDCW/PLAS),10 and a single-step PCR specific for the 14–41 copy nuclear tandem repeat region Pvr47 (Pvr47F/R)11 (Table 1).

The PCR master mixtures and cycling parameters for the PLU/VIV and Pvr47 assays were the same as the protocol used for the PLF/UNR/VIR assay above. The PCR mixtures were made and run independently in each assay. The Pvr47 assay was a single round of amplification. The nested GDCW/PLAS assay consisted of a 10-μL reaction with a master mixture comprised of 0.3 μM of each primer, 2 mM MgCl2, 0.125 mM dNTP, 1× buffer, 5% dimethyl sulfoxide, 1 unit of Biolase Taq, and 1 μL of DNA. The cycling parameters consisted of an initial temperature of 94°C; followed by 35 cycles at 94°C for 30 seconds, 60°C for 30 seconds, and 72°C for 1 minute and 30 seconds; and a final extension at 72°C for 10 minutes. The nested PCR used the same cycling parameters for the second round, but the number of cycles was increased to 40.

DNA from the same 94 samples, which were previously tested by using the PLF/UNR/VIR primers, was tested again with the three assays listed above (Figure 1). A positive control of P. vivax extracted from An. dirus was included, as well as a negative control of nuclease-free water. The resulting amplicons were again sequenced in both directions, and the sequences were run through PlasmoDB BLAST and NCBI BLAST/n tools, with the exception of the Pvr47 amplicons. Pvr47 is a novel locus, whose sequences are not available in the NCBI database; therefore, sequences were only run through PlasmoDB BLAST/n.

Pvr47 produced amplicons in 60 of the 94 samples, but only 10 of these were of the approximate expected size of 333 basepairs. Most amplicons was either of the wrong size, smears, or produced multiple-banding patterns. Of the 60 amplified samples, only six produced clean sequences that were obtained from the Pvr47 assay. None of the sequences matched any of the Plasmodium spp. found in the PlasmoDB.org database > 70%. For the PLU/VIV assay, 3 of the 94 samples amplified, but all were approximately 1,000 basepairs instead of the expected 120 basepairs. All samples had a quality score > 90%. However, none of these samples matched any species of the Plasmodium spp. found in the PlasmoDB.org database > 85%. When run through the NCBI database, all three matched Zoophthora lanceolata strain ARSEF (Accession nos. EF392550.1, EF392550.1, and EF392550.1) > 97%. Zoophthora is an arthropod fungus that is sometimes used as an insecticide.12,13 The PLU5/6 primers were aligned with the 18S sequence of Z. lanceolata. The PLU5 primer matched the sequence with only a 1-basepair difference; the PLU6 primer matched the sequence with a 7-basepair difference. This finding might suggest that running a nested PCR with a total of 75 cycles could increase the probability of non-specific binding and explain the amplification of this fungus although the primers are not an exact match.

The GDCW/PLAS was the most successful of the assays because it did not result in smearing or non-specific binding. Amplification was observed in one An. belenrae, which was the approximate correct size of 815 basepairs. This sample was a 100% match to the rodent malaria parasite P. berghei in the PlasmoDB database, as well as in NCBI database (100% MI and 100% QC; accession no. DQ414645). The GDCW/PLAS assay is used to amplify human and primate malaria pathogens, but exclude Apicomplexa and human blood. To verify that this assay amplifies P. berghei, seven An. stephensi with known P. berghei infections were tested by using this PCR. All seven samples produced amplicons. However, before the testing of An. stephensi, no P. berghei had been used in this laboratory, which excluded on-site contamination. The discovery of P. berghei in South Korea and in An. belerae is novel. Also, the head and thorax was separated from the abdomen, suggesting that the sporozoites were in the salivary glands, and that the infection had not merely been ingested but had also propagated. Because P. berghei can infect multiple rodent vectors, and is a temperate species, it may have a wider range than previously expected.

Over the course of this research genomic DNA was extracted from several thousand mosquitoes for vector identification, and many of these mosquitoes were also screened for P. vivax. The Pvr47, PLF/UNR/VIR, and PLU/VIV assays continually produced uninterpretable results and multiple banding patterns, leading to this sequencing work. Samples that produced amplicons were subjected to electrophoresis on agarose gels for one hour to illustrate these results (Figure 2). The GDCW/PLAS assay did not produce any amplicons, illustrating that these primers are more specific than those used in the other three assays.

Figure 2.

Agarose gel electrophoresis showing multiple banding patterns produced by polymerase chain reactions PLU/VIV (A), PLF/UNR/VIR (B), and Pvr47 (C) used to detect Plasmodium spp. in Anopheles spp. mosquitoes.

Our results showed that the PLF/UNR/VIR, PLU/VIV, and Pvr47 assays produce non-specific sequences and uninterpretable results and have primers that could potentially bind to species in multiple phyla. Also, the PLU/VIV assay detects Z. lanceolata, an arthropod fungus. The GDCW/PLAS assay is more specific, but detects P. berghei, which is a rodent pathogen. Therefore, amplicons must be verified by sequencing. Although there is always potential for DNA degradation during shipping that could also result in smearing and irregular banding patterns, the clean sequences for the mosquito and Plasmodium amplicon is evidence that not only is the DNA viable, it is intact enough to provide quality sequences (Table 2). Our positive controls were infected mosquitoes, showing that our DNA extraction and detection methods are optimized for in vivo detection despite using an extraction protocol different from those used in the original publications. Also, if inhibitors were an issue, amplification of the mosquitoes or positive control would not have occurred. The strongest evidence that there may be issues with this detection method is their BLAST results for primers, which were independent of any laboratory work.

The conclusions of this study cast doubt on previously reported infection rates from South Korea and elsewhere where these assays were used.6 Many samples produced clear amplicons, sequences multiple pea for a single base pair, we suspected was caused by co-amplification of multiple organisms. Although one could argue that a band of the incorrect size is not a positive result per se, an assay by definition should produce a binary band or no band result. With the exception of the GDCW/PLAS primers, these assays do not produce such a result. Although these assays may work well with human blood, confounding factors from field-collected mosquitoes, such as sample degradation, exposure to environmental contaminants, and low levels of mosquito infections, may reduce the utility of these assays, and therefore render them unsuitable for use for the detection of specific Plasmodium spp. The cytB GDCW/PLAS assay is the most promising technique for future studies, despite the fact that it is not limited to human-specific Plasmodium spp. detection. Because the interpretation of results when PCR is used can be ambiguous and somewhat subjective, the use of real-time assays in which a present diagnostic negative/positive result is pre-programmed (based on cycle threshold) would produce more robust results in future studies. However, the primers used in a real-time assay must be thoroughly tested with field-collected mosquitoes to ensure they do not amplify non-Plasmodium species because many also target the 18S ssrDNA region.14