A(maize)ing attraction: gravid Anopheles arabiensis are attracted and oviposit in response to maize pollen odours

Anopheles arabiensis, Nazareth and Dongola strains, both maintained in the laboratory for over 30 years, were used for behavioural and electrophysiological analyses in Ethiopia and Sweden, respectively. The colonies were maintained at 27 ± 2 °C, 75 ± 5% relative humidity and under a 12 h light/12 h dark cycle. Briefly, the aquatic stages of the mosquitoes were reared in distilled water, and fed Faffa food (Ethiopia) or Tetramin^® fish food (Sweden). Pupae were transferred from the rearing trays in 100 ml polypropylene cups (Qingdao Ori-Color Industry and Commerce Co., Ltd., China), containing distilled water, to cages (30 cm × 30 cm × 30 cm; custom-made or Bugdorm, MegaView Science, Taiwan). The emerging adults were supplied with 10% sucrose solution ad libitum. Five days after emergence, female mosquitoes were offered a rabbit (Ethiopia) or sheep blood (Sweden) from an artificial feeder (Hemotek, Discovery Workshops, Accrington, UK) over the course of two days, 3 h per day, to ensure that females took a complete blood meal. For all experiments, gravid females, 3 days post-blood feeding, were selected by visually inspecting the pale white abdomen and used for bioassays.

The headspace of BH-660 and ZM-521 maize pollen was collected in the field under shaded ambient conditions at Melkassa Agricultural Research Center, in East Oromia region of Ethiopia. In addition, headspace was collected from the water of a major malaria mosquito breeding site at the shore of lake Ziway, Ethiopia. The inflorescence of a fully mature male flower (45 replicates per cultivar) was enclosed in a polyacetate bag (Toppits, Cofresco, Germany), and a charcoal-filtered continuous airstream (1 l min⁻¹) was drawn by a Personal Air Sampler (PAS-500, Spectrex, Redwood City, CA, USA) over the tassel, onto an aeration column for 3 h. Alternatively, for collecting the headspace from breeding water (35 replicates), 1 l was poured into a polyacetate bag, after which the headspace was collected for 3 h using a diaphragm vacuum pump (12 V, KNF-Neuberger, Freiburg, Germany), using charcoal-filtered air as described above. The aeration columns were made of Teflon tubing (6 cm × 3 mm i.d.), filled with 35 mg Super Q (80/100 mesh; Alltech, Deerfield, IL, USA) between polypropylene wool plugs and nylon stoppers. The aeration columns were cleaned with 1 ml n-hexane (LabScan, Malmö, Sweden), re-distilled before use. Adsorbed volatiles were eluted with 300 µl re-distilled n-hexane. Headspace volatile extracts were stored in glass vials at −20 °C until used for behavioural and electrophysiological analyses.

A two-port olfactometer [17] was used to test the attraction preference of the mosquitoes for the headspace volatiles collected from the BH-660 and ZM-521 pollen, and from natural breeding water. All assays were conducted between 18:00 and 21:00, which is the peak period of oviposition activity as determined in pilot experiments. For each experiment, 10 gravid females were allowed to acclimatize for 5 min in a custom-made cage (22 cm × 30 cm × 12 cm; L:W:H) constructed of clear vinyl for easy viewing. Thereafter, two dental-wick (4 cm × 1 cm; L:D; DAB Dental AB, Upplands Väsby, Sweden) odour dispensers were simultaneously introduced into the cylindrical vinyl arms (13 cm × 9 cm; L:D) positioned at opposite ends of the cage. Treatment and control wicks were exchanged in between experiments to assure for no positional bias. The ends of the cylindrical arms were covered by mesh. Attraction preference of the mosquitoes to the following treatments were analysed: (a) headspace volatiles of breeding water vs hexane control, (b) headspace volatiles of BH-660 or ZM-521 pollen vs hexane, (c) headspace volatiles of BH-660 or ZM-521 pollen vs headspace volatiles of breeding water, and (d) headspace volatiles of BH-660 vs headspace volatiles ZM-521 pollen. For all treatments, a hexane vs hexane control was performed. The behavioural response to the volatile headspace of the two maize pollen varieties was analysed to increasing amounts of headspace extract from the two pollen varieties. After 5 min, the behavioural responses of the mosquitoes were scored by counting the number of mosquitoes in each arm. Ten replicates per treatment and per release rate were performed. The treatments were tested one release rate at a time, with a solvent only control for each replicate to control for day effects. Between each replicate, the bioassay apparatus was cleaned with 70% ethanol and the position of the treatments changed to avoid bias.

The oviposition preference of gravid mosquitoes was analysed in a two-choice assay [17]. Metal wire framed cages (30 cm × 30 cm × 30 cm) covered with white nylon mosquito netting were used. Two 100 ml polypropylene cups (Qingdao, Ori-Color Industry and Commerce, Co. Ltd, China), placed in opposite corners of the cages and filled to the brim (100 ml) with distilled or field collected breeding water, served as the oviposition substrate. The position of the cups was exchanged between experiments. Treatment cups were conditioned by dosing the oviposition substrate with the same range of release rates of headspace volatile extracts of BH-660 and ZM-521 pollen in hexane, as described above. An equivalent amount of hexane was used as a control. In addition, a hexane vs hexane control was performed for all treatments. Treatment and control cups were exchanged in between experiments to assure for no positional bias. Gravid mosquitoes were transferred from the maintenance cages at dusk (18:00), and on the following day (09:00) the numbers of eggs in the two cups were counted. All experiments were replicated ten times. The treatments were tested one release rate at a time, with a solvent only control for each replicate to control for day effects.

Antennal responses of gravid female An. arabiensis to the preferred headspace extract of BH-660 pollen were analysed using combined gas chromatography and electroantennographic detection (GC-EAD). An Agilent Technologies 6890 GC (Santa Clara, CA, USA) was equipped with a HP-5 column (30 m × 0.25 mm i.d., 0.25 μm film thickness, Agilent Technologies), and hydrogen was used as the mobile phase at an average linear flow rate of 45 cm s⁻¹. Each sample (2 µl) was injected in splitless mode (30 s, injector temperature 225 °C). The GC oven temperature was programmed from 35 °C (3 min hold) at 10 °C min⁻¹ to 290 °C (10 min hold). At the GC effluent, 4 psi of nitrogen was added and split 1:1 in a Gerstel 3D/2 low dead volume four-way cross (Gerstel, Mülheim, Germany) between the flame ionization detector and the EAD. The GC effluent capillary for the EAD passed through a Gerstel olfactory detection port-2 transfer line, that tracked the GC oven temperature, into a glass tube (30 cm × 8 mm), where it was mixed with charcoal-filtered, humidified air (1.5 l min⁻¹). The antenna was placed 0.5 cm from the outlet of this tube. The antennal preparation was made by using the excised head. The distal end of the antenna was inserted into a recording glass electrode filled with Beadle–Ephrussi Ringer, after cutting the distal segment. The recording electrode was connected to a pre-amplifier probe and then to a high impedance DC amplifier interface box (IDAC-2; Syntech, Kirchgarten, Germany). The reference electrode, filled with Beadle–Ephrussi Ringer, was inserted into the head capsule. Each individual animal accounted for a single replicate, and at least five replicates were performed.

The BH-660 pollen headspace extract was injected (2 µl) and analysed on a combined gas chromatograph and mass spectrometer (GC–MS; 6890 GC and 5975 MS; Agilent Technologies), operated in the electron impact ionization mode at 70 eV. The GC was equipped with fused silica capillary columns (30 m × 0.25 mm, 0.25 µm film thickness), DB-wax (J&W Scientific, Folsom, CA, USA) or HP-5MS (Agilent Technologies). Helium was used as the mobile phase at an average linear flow rate of 35 cm s⁻¹. The temperature programmes were the same as for the GC-EAD analysis. Compounds were identified according to retention times (Kovat’s indices) and mass spectra, in comparison with custom made and NIST05 libraries (Agilent), and confirmed by co-injection of authentic standards: (±)-α-pinene (Cas no. 7785-70-8; Aldrich, 98%), (±)-limonene (Cas no. 5989-27-5; Sigma, 97%), nonanal (Cas no. 124-19-6; Aldrich, 95%), benzaldehyde (Cas no. 100-52-7; Aldrich, 99%), p-cymene (Cas no. 99-87-6; Aldrich, 97%). For quantification, 100 ng of heptyl acetate (99.8% chemical purity; Aldrich) was added as an internal standard.

The assays were carried out in the same two-port olfactometer and oviposition bioassay that were used for the natural headspace extract experiments. The synthetic blend mimicked the composition and ratio of compounds in the headspace collected from BH-660 pollen. Synthetic blends were prepared at seven different doses in half orders of magnitude between 1 and 1000 ng of α-pinene in pentane dispensed from dental-wicks and in distilled water for the attraction and oviposition assays, respectively. The ratio among the compounds within the blend was maintained as a constant across all doses. Thereafter, subtractive blends, in which single compounds of the full blend were removed, were tested against the full blend (100 ng).

A preference index was calculated, (T − C)/(T + C), for both attraction preference (AP) and oviposition (OP) preference; where T is the number of mosquitoes or eggs associated with the test odours, and C the number of mosquitoes or eggs associated with the control odours. The behavioural responses of gravid An. arabiensis in the two-port olfactometer and oviposition bioassay were analysed using a nominal logistic fit model, in which choice was the dependent variable, weighted by the number of (1) mosquitoes in the attraction assays and (2) eggs laid in the oviposition assays, with release rate or dose as the independent fixed effect and replicate (day) as a random effect (JMP^® Pro 12.0.1. SAS Institute Inc., Cary, NC, USA). Here, we report the χ² and P value from the likelihood ratio test. The generation of the good fit logistic models were made by omitting the highest release rate or dose, which in most cases was shown to cause either a neutral or avoidance behaviour.

Gravid An. arabiensis were significantly attracted, over a range of release rates, to the headspace collections of pollen collected from the ZM-521 and BH-660 maize varieties in a two-port olfactometer, when compared to the hexane control (ZM-521: χ² = 9.647, P = 0.0019; BH-660: χ² = 8.976, P = 0.0027; Fig. 1a, c; Additional file 1) and the headspace of breeding water (ZM-521: χ² = 10.66, P = 0.0011; BH-660: χ² = 16.77, P < 0.0001; Fig. 1b, d; Additional file 1). Comparison between the two varieties showed a significantly higher attraction to the headspace of BH-660 pollen by gravid An. arabiensis than to that of ZM-521 (χ² = 9.648, P = 0.0019; Fig. 1e; Additional file 1). No significant difference was observed in the attraction to the hexane control and the headspace of breeding water when tested in the same assay (χ² = 0.5968, P = 0.4398).

Gravid females preferred to lay their eggs in both distilled and breeding water conditioned with the headspace of ZM-51 and BH-660 maize pollen, over water with hexane added as a control (ZM-521 vs distilled water: χ² = 4.405, P = 0.0358; BH-660 vs distilled water: χ² = 7.887, P = 0.0050; ZM-521 vs breeding water: χ² = 8.980, P = 0.0027; BH-660 vs breeding water: χ² = 6.812, P = 0.0091; Fig. 2a–d; Additional file 2). Comparison between the headspace of the two varieties demonstrated a significantly higher number of eggs laid by An. arabiensis in response to the headspace of BH-660 pollen than to that of ZM-521, irrespective of the oviposition substrate (distilled water: χ² = 11.71, P = 0.0006; Fig. 2e; breeding water: χ² = 8.492, P = 0.0036; Fig. 2f; Additional file 2). The number of eggs laid by An. arabiensis was not significantly different between the two controls (χ² = 0.1959, P = 0.6581; data not shown).

The GC-EAD and GC–MS analyses identified five bioactive compounds in the headspace of the more attractive BH-660 pollen: benzaldehyde, nonanal, p-cymene, limonene and α-pinene (Fig. 3). The overall volatile release rate was 100 ng min⁻¹, with limonene as the most abundant compound, followed by nonanal and α-pinene.

A synthetic blend containing all five GC-EAD-active compounds, approximating their natural ratio (10:5:5:1:1, limonene:nonanal:α-pinene:benzaldehyde:p-cymene), elicited short-range attraction and stimulated oviposition in gravid An. arabiensis over a range of doses (χ² = 9.581, P = 0.0020; χ² = 8.125, P = 0.0044, respectively) (Fig. 4; Additional file 3). The release rate that elicited the optimal behavioural responses was found to be similar to the natural release rate (Fig. 4). To assess the role of each individual component in the blend, five subtractive blends, from which single compounds were removed, were evaluated against the full blend in both behavioural assays (Fig. 5; Additional file 4). All subtractive blends were found to be less attractive (χ² = 4.02, P < 0.0449) and females laid fewer eggs in the subtractive blends (χ² = 5.13, P < 0.0236) than in the full blend (Fig. 5; Additional file 4).