Singapore’s Anopheles sinensis Form A is susceptible to Plasmodium vivax isolates from the western Thailand–Myanmar border

In December 2013, larvae of the An. sinensis were collected from a grassy pool of a big field at Changi Coast Road, eastern Singapore and they were colonized in the laboratory. Following the first collection, 103 adults and larvae An. sinensis were collected from 13 different locations in Singapore between September 2013 and January 2016 (Fig. 1). Anopheles sinensis larvae were collected by the Environmental Health Officer (EHO) of National Environment Agency (NEA) during the routine malaria surveillance and were submitted to Environmental Health Institute (EHI) for identification. They were reared to adult for this study.

For the An. sinensis adults, some were collected using modified CDC-light traps during spatial distribution study and ad-hoc collection in response to public feedback on high mosquito population. A portion of adult An. sinensis were collected through human landing catch during 2009 malaria outbreak in malaria cluster areas [8] and during surveillance by the Singapore Armed Forces in military training grounds [32]. The remaining adult An. sinensis were collected via Night-Catcher during temporal study and ad-hoc collection was conducted due to high mosquito population. Night-Catcher, an in-house mosquito trap, which was improvised from CDC light trap, enables hourly collection of mosquitoes using incandescent light and dry ice (CO₂) as attractant (Fig. 2). All collections were conducted from 7 p.m. to the 10 a.m. the next morning.

Larvae and adult mosquitoes were identified under compound microscope according to taxonomy keys [17, 33, 34]. Confirmed An. sinensis were reared in EHI’s insectary at 25 °C (± 2 °C) and 70% (± 10%) relative humidity. Upon emergence, the adults were then reconfirmed morphologically to the species level according to taxonomy keys [17, 33, 34]. Due to the absence of morphological trait differences between the two forms of An. sinensis, these mosquitoes were further determined using molecular taxonomic tools to ensure the accurate form determination as well as purity of the colony.

In order to identify the forms of An. sinensis in Singapore, regions of both the COI and rDNA internal transcribed spacer (ITS2) genes were sequenced. All 103 An. sinensis collected from 13 different locations were individually processed. Total DNA were extracted individually using DNeasy blood and tissue kit following the manufacturer’s procedures (Qiagen, Hilden, Germany) and stored at − 20 °C until analysis. Two regions flanking the mitochondrial COI gene and ITS2 gene were amplified by polymerase chain reaction (PCR) as described in previous studies [35‐37]. Amplicons were then visualized on 2% agarose gel stained with GelRed (Biotium Inc., USA), and cleaned using Purelink PCR purification kit (Invitrogen Corp., USA) according to manufacturer’s instructions. Sequencing was carried out by a commercial laboratory using BigDye Terminator Cycle Sequencing kit (Applied Biosystems, USA). For the susceptible study, sixty adult mosquitoes that were used in the susceptible study were transferred individually into separate 2 ml vials and homogenized using a mixer mill (Retsch Mixer Mill MM301). The DNA extraction, PCR and gel visualization protocols were similar to what was described above. The sequences of the remaining 11 samples from nine different locations were extracted from EHI’s previously published data [8, 35].

Contiguous sequences of CO1 and ITS2 genes were created using Lasergene 9.0 software suite (DNASTAR Inc., USA). These sequences were then aligned using Clustal W algorithm [38] executed in BioEdit v7.05 software [39]. Neighbour joining algorithm was adopted during the construction of phylogenetic trees using MEGA 6.06 software suite [40]. Parameters selection included a Kimura-2 parameter substitution model with gamma distributed rates using the nearest neighbour interchange heuristic search method. Robustness of clustering was determined by bootstrap analysis with 1000 replicates. Reference DNA sequences were obtained from the GenBank database. The pairwise distances between each specimens was computed using MEGA 6.06 software package [40].

Colonization was initiated by transferring wild-caught An. sinensis adults into 30 cm × 30 cm × 30 cm large cage made from acrylic plastic sheets. Ten percent sucrose was given as food source by soaking it with cotton wool in a glass bottle. Artificial insemination was conducted in earlier generations to propagate and establish a colony in the insectary. Subsequent generations were based on natural insemination, which was induced by exposing these mosquitoes to stroboscopic blue light from 7 p.m. to 9 p.m. for at least 5 days prior to blood feeding [41, 42]. Six days after emergence, females were transferred into 15 oz. transparent plastic container with plastic net affixed on top. Female mosquitoes were deprived from sugar overnight in these feeding containers with moistened cotton pad on top of the net. Specific Pathogen Free (SPF) mini-pig blood was offered to the mosquitoes via a Hemotek^® feeding system. On the 3rd day post blood feeding, females were transferred into 15 oz. ovipots (transparent plastic containers) lined with moist filter paper for eggs collection. Eggs were collected on filter paper and hatched in Reversed Osmosis water (RO). Larvae food consisting of wheat germ, oats, dry yeast, casein or low fat milk powder, bubble rice, Vitamin B complex and Nestum were mixed and ground into fine powder. Approximately 0.1 g of larvae food was dispensed daily when the larvae grew from 1st to 2nd instars. Larvae food increased to 0.2 and 0.4 g during 3rd instars and 4th instars, respectively. The second generation (F₂) of An. sinensis, collected from a country club in Singapore in July 2015, was also used in this study to compare the differential vector competence with the lab-bred (F₂₂) strain. In order to produce sufficient mosquitoes for the comparison, artificial insemination of F₀ and F₁ An. sinensis were carried out. This strain was colonized following the above described protocol.

A colony of the twenty-second generation of An. sinensis (F₂₂) and another of F₂ were used for the competence study. An approval and an export permit were obtained from Director-General Public Health of National Environment Agency prior to sending the eggs of An. sinensis to SMRU. Eggs produced at the insectary of EHI, were transferred onto a clean piece of filter paper, packed and sealed in a sterile petri dish before transporting to Shoklo Malaria Research Unit (SMRU) laboratory, on the Thai-Myanmar border. Although An. sinensis Form A and B have previously been found in northern Thailand [28], every precaution was taken to ensure that Singapore’s strain An. sinensis used in this study were not released.

Patients seeking consultation at SMRU migrant clinics located along the border (Wang Pha, Mawker Thai) where they were tested by blood smear microscopy, only gametocytes positive patients were selected for the study. After a written informed consent was obtained, five to 10 ml of venous blood were drawn into a heparin tube, and immediately placed in a water bath at 37–38 °C to prevent exflagellation of male microgametes [43]. Within an hour, blood samples were transported from the field clinics to the central SMRU laboratory for processing. Following centrifugation at 1800g for 5 min in an Eppendorf^® centrifuge which was warmed at 38 °C, plasma was replaced with AB+ serum and within 10 min the blood was transported to the insectary.

All experimental mosquito infections were carried out at the SMRU secured insectary in Mae Sot as described by Andolina et al. [44]. The secure insectary that is physically separated from open areas by four sealed and locked doors. Only authorized trained personnel can gain access and conduct infection studies. All infected/engorged mosquitoes were counted and placed in incubators (Sanyo^®, MIR-254) were secured with netting material. Mosquitoes which fed insufficiently were killed in ethanol 70%. Anopheles cracens (An. dirus B), an efficient P. vivax vector [44] was used as a positive control and was fed with the same blood samples, in parallel with An. sinensis.

On seven to 8 days post infection, midguts of both mosquito species were dissected and stained with 1% mercurochrome. Oocyst positive midguts were placed in 100 µl of PBS and stored at − 80 °C until PCR was performed. Dissection of salivary glands for sporozoites detection was carried out 15 days post infection. Salivary glands were placed in an Eppendorf tube filled with 50 µl of Roswell Park Memorial Institute medium (RPMI) and kept on ice until the dissection of all mosquitoes was completed. The sample was spun down for 5 min in a micro centrifuge and salivary glands were pooled and crushed using a 100 µl pipette. 10 µl of salivary glands suspension was placed into a KOVA Glasstic slide with 10 grids. Sporozoites were counted and averaged on four grids, multiplied to the chamber factor and dilution factor in order to calculate the number of sporozoites per µl. Average sporozoites counts in a single mosquito was calculated by dividing the total sporozoites with the number of mosquitoes dissected.

To confirm P. vivax infection, DNA from dissected midguts was extracted in 100 µl PBS using a Qiagen Tissue Kit with minor modifications. Briefly, 180 µl of ATL buffer and 50 µl of proteinase K (Qiagen Tissue Kit) was added to the sample, mixed briefly by vortex and incubated overnight at 56 °C in a shaking incubator. Following digestion, DNA was bound to the silica membrane, washed then eluted in 200 µl water following manufacturer’s instructions. The sample was then concentrated by drying in a vacuum concentrator at 30 °C and re-eluted in 10 µl AE Buffer (Qiagen). Primers and probes described by Perandin et al. [45] were used to amplify and detect species specific regions of the 18S rRNA gene. Real-time PCR was done using QuantiTect Multiplex RT-PCR Kit and an ABI 7500 Fast Cycler.

R-3.1.1 software was used to conduct statistical analysis in this study [46]. Two-way Wilcoxon rank sum test was used when comparison of oocysts development was made between F₂ and F₂₂ An. sinensis Form A.

The study was approved by Oxford Tropical Research Ethics Committee (Reference 28-09).

A total of 103 mosquitoes collected from the various locations were confirmed to be An. sinensis through morphological identification. However, polymorphic wing variation at CuA was noted among the specimens from field collection (Fig. 3). Out of the 103 specimens, only 42 specimens had complete morphological characteristics where wing scales were still intact. Of the 42 specimens, 20 (47.6%) of them had pale CuA fringe spots, while the remaining 22 (52.4%) showed dark fringe spots. The locality and the proportion of the dark and pale CuA fringe spot are listed in Table 1. It was observed that each location could have both An. sinensis with pale and dark CuA fringe spots.

Phylogenetic analysis based on ITS2 gene of 103 Singapore An. sinensis and 42 reference sequences from the NCBI database showed that all Singapore An. sinensis sequences, including those collected during 2009 outbreak, formed a monophyletic clade. Though the Singapore sequences were derived from mosquitoes with CuA pale or dark fringe spots, it is interesting to note the tight clustering despite the differences in wing fringe coloration. The phylogenetic tree also shows that the Singapore An. sinensis clustered together with Thailand’s An. sinensis Form A (bootstrap value = 85) (Fig. 4). On the other hand, An. sinensis Form B from Thailand (AY13047.1), Korea (AY130469.1), China (EU 931614) and Japan (EU 931613) formed a separate clade from our local An. sinensis with strong bootstrap support (bootstrap value = 100). None of the Singapore An. sinensis adults falls into the Form B clade. The data suggests that An. sinensis found in Singapore are Form A.

Similarly, the phylogenetic analysis of CO1 gene sequences using neighbor joining showed that Singapore An. sinensis clustered separately from An. sinensis Form B from Korea (AY444351) with bootstrap value of 73 (Fig. 5). There was no reference sequence of Form A CO1 gene in the NCBI database for comparison with Singapore An. sinensis sequences. Interestingly, unlike the ITS2 sequences, the COI gene sequences of local An. sinensis formed subgroups, though with weak bootstrap support, indicating subtle genetic changes at mitochondrial level.

Previously, it was reported that approximately 75% of An. sinensis in the Malaya region (including Singapore) and Borneo had dark fringe spot [34]. However, in this current study, it was interesting to note that 22 (52.4%) of the examined local An. sinensis has dark fringe spot at CuA (Table 1). When all 42 specimens with pale and dark CuA wing fringe spots were tallied with the CO1 phylogenetic tree, no specific clustering of pale or dark phenotype in subgroups. Both phenotypes randomly occurred within the CO1 phylogenetic tree.

To ensure that the laboratory mosquitoes used for infection is consistent with field caught mosquitoes, the CO1 and ITS2 genes of 30 F₂₂ mosquitoes were also analysed. They were found to be the same as field caught ones (Figs. 4, 5).

The bloods of seven patients with P. vivax were fed to F₂₂ An. sinensis and An. cracens using Hemotek^® feeding system membrane feeding. In total, 50–100% (Fig. 6) of dissected An. sinensis developed one to 92 oocysts (Table 2). Similar results were obtained with An. cracens where the infection rate was between 77.7 and 100%, with each mosquito developing one to 200 oocysts. These findings were further confirmed with real-time PCR.

Out of seven experiments, only the last four yielded enough blood fed mosquitoes for detection of parasite in salivary glands. Salivary glands from each experiment were pooled to ensure minimal loss of sporozoites during manipulation of examining. These four experiments showed that An. sinensis could produce 703–14,538 sporozoites per mosquito (Fig. 7), while An. cracens, produced 2812 sporozoites to 76,764 sporozoites per mosquito (Table 2).

Comparison of infectivity between F₂₂ and F₂ strains of An. sinensis was conducted to determine if they were analogous. Both strains showed similar susceptibility to P. vivax infection (two-way Wilcoxon rank sum test, p = 0.935) with the F₂ strain having 75–100% infection rate and F₂₂ strain having a 50–100% at midguts (Table 3). Second generation of An. sinensis F₂, had on average 25 and 51 oocysts in each midgut, while F₂₂ had 11 and 74 oocysts on average.

Two-way Wilcoxon rank sum test showed no significant difference in the number of oocysts detected between F₂ and F₂₂ strains (p = 0.935). No statistical test was carried out on sporozoites since insufficient data was available.

The average numbers of sporozoites produced by F₂₂ were 4302 and 14,538 sporozoites in each mosquito, while that produced by F₂ strain were 10,928–11,250 sporozoites each. Statistical tests were not carried out on the average number of sporozoites produced due to insufficient data.