Bionomics of Anopheles arabiensis from Mamfene in KwaZulu-Natal (South Africa): an area of high malaria transmission in the province

Background Although great strides have been made in controlling malaria, the disease is of signicant public health importance. Historically, efforts to control the vector has concentrated on adult vector control targeting the female Anopheles mosquitoes. As there is now a focus on eliminating residual malaria from KwaZulu-Natal, new strategies are being investigated to increase the impact of malaria elimination strategies. Greater attention is now being given to larval control, as a complementary measure to indoor residual spraying. However, there is a large gap in knowledge of the bionomics of the larval stages of this mosquito vector of malaria in South Africa. In order to focus on both larval and adult mosquito control methods, larval development and the reproductive stages of the vector were investigated since these variables inuences our ability to impact mosquito populations through larval control. This study was therefore conducted to determine the peak eruption times and the emergent sex ratios, as well as the peak egg oviposition time in order to attack the mosquito when it is at its most vulnerable and when control interventions will have the most impact. Results Oviposition studies showed two peaks corresponding with late evening and again just before dawn. Most eggs were also laid in the rst half of the night (18h00 – midnight). Most mosquitoes erupted just after sunset and the sex ratios showed that twice as many females as males emerged. Females readily took a bloodmeal after oviposition or just after erupting. Hatch rate to viable rst instar larvae was 74.5%. Conclusions The results of this study have provided information as to when interventions would be most effective in controlling mosquito populations and have provided information that highlights the value of larval control as a complementary measure to adult mosquito control. The most vulnerable stages of the female Anopheles arabiensis are when they have just emerged or when they have just oviposited. Vector control strategies should be designed to target these vulnerable stages at the breeding sites in order to have maximum impact.


Abstract
Background Although great strides have been made in controlling malaria, the disease is of signi cant public health importance. Historically, efforts to control the vector has concentrated on adult vector control targeting the female Anopheles mosquitoes. As there is now a focus on eliminating residual malaria from KwaZulu-Natal, new strategies are being investigated to increase the impact of malaria elimination strategies. Greater attention is now being given to larval control, as a complementary measure to indoor residual spraying. However, there is a large gap in knowledge of the bionomics of the larval stages of this mosquito vector of malaria in South Africa. In order to focus on both larval and adult mosquito control methods, larval development and the reproductive stages of the vector were investigated since these variables in uences our ability to impact mosquito populations through larval control. This study was therefore conducted to determine the peak eruption times and the emergent sex ratios, as well as the peak egg oviposition time in order to attack the mosquito when it is at its most vulnerable and when control interventions will have the most impact. Results Oviposition studies showed two peaks corresponding with late evening and again just before dawn. Most eggs were also laid in the rst half of the night (18h00 -midnight). Most mosquitoes erupted just after sunset and the sex ratios showed that twice as many females as males emerged. Females readily took a bloodmeal after oviposition or just after erupting. Hatch rate to viable rst instar larvae was 74.5%. Conclusions The results of this study have provided information as to when interventions would be most effective in controlling mosquito populations and have provided information that highlights the value of larval control as a complementary measure to adult mosquito control. The most vulnerable stages of the female Anopheles arabiensis are when they have just emerged or when they have just oviposited. Vector control strategies should be designed to target these vulnerable stages at the breeding sites in order to have maximum impact.

Background
Although the incidence and mortality due to malaria are decreasing remarkably, malaria remains a disease of major public health importance. The latest available statistics report that there has been 405 000 deaths and 219 million cases of malaria world-wide during 2018 [1]. However, in the southern most parts of the continent, malaria has been quite effectively controlled using IRS or LLINs to reduce contact between the host and vector thus resulting in diminished transmission. Eight countries in southern Africa, including South Africa, are now targeting malaria elimination as a result of having achieved low levels of transmission using coordinated and sustained vector control measures, including larval control [2].
However, the road to elimination is paved with many challenges and these would need to be overcome to achieve malaria elimination certi cation.
One of the main obstacles to achieving elimination is residual malaria transmission. Residual malaria is de ned as the low levels of transmission that continues after full universal coverage with effective long-lasting insecticide impregnated nets and/or indoor residual spraying is achieved [3]. Killeen [3] proposed that there should be a greater emphasis on larval control. In South Africa, the province of KwaZulu-Natal has been in elimination for the past decade [4] but residual transmission is su cient to make malaria elimination unattainable without additional vector control tools and technologies which includes the use of larval control.
Due to insecticide resistance and the advent of outdoor biting of mosquitoes, there is increasing attention being paid to larval source management and larval control. Traditionally, larval control has been limited to the control of active larval breeding sites using chemicals, toxins, bacteria, fungi and other biological control methods. However, the usefulness of these techniques is limited by the exorbitant labour costs and the cost of the chemicals but is recommended for use in areas where the breeding sites are few, xed and ndable [5]. There is thus a need to understand the interface between the larval and adult stages in order to identify any opportunities that can be exploited for control purposes. In South Africa the biology of the immature stages and the females feeding behaviour is poorly understood. Over time mosquitoes have adapted to outdoor feeding due to the continued use of IRS and feeding earlier using LLINs [6].
Available information was collected decades ago [7,8] and has changed over time [9]. Environmental changes have also occurred in terms of rainfall and temperature [10]. Changes in the ecological factors can affect malaria transmission by modifying the microclimate of the immature stages and adult mosquitoes [11]. Changes in biology and physiology can also impact mosquito development [12]. In order to focus on both larval and adult mosquito control methods, there needs to be a greater understanding of larval development and the reproductive stages of the vector since these variables in uence the ability to control mosquito populations through larval source management. This study was therefore conducted to determine the peak emergence times and the emergent sex ratios, as well as the peak egg oviposition time in order to better target the vulnerable stages of a mosquito when control interventions will have the most impact. Furthermore, the results from this study can feed into larval source management strategies.

Methodology Collection of Mosquitoes
Laboratory reared An. arabiensis from Mamfene, in KwaZulu-Natal, South Africa was used for the observations in this study. The mosquitoes used were the F 2 progeny of wild-caught females. Field collected females readily laid eggs under insectary conditions maintained at constant temperature of 27°C and relative humidity of 70% with a photoperiod of 12h light: 12h dark with one hour simulated crepuscular period. During the study, the mosquitoes were maintained at a constant temperature of 25°C (the average summer temperature in KwaZulu-Natal).

Oviposition Behaviour
The variables under investigation were the time at which the eggs were laid, the number of eggs laid, the egg hatch rate and the time it took the female to nd a blood meal. At the start of the trial, adult female mosquitoes were collected from Mamfene in KwaZulu-Natal, South Africa. All F 1 mosquitoes emerging on day 1 were pooled together in cages to give the mosquitoes an opportunity to mate. Three hundred unfed three-day old Anopheles arabiensis female mosquitoes were blood fed two days prior to the trial. Thereafter they were kept in 30x30x30cm BugDorm Insect Rearing Cage (BugDorm, USA) under controlled environment with optimum environmental conditions of 27°C and a humidity of 70% which represents the average summer conditions in KwaZulu-Natal. To observe the number of eggs laid per female, 300 fully gravid blood fed female mosquitoes were individually placed into a 60 ml test tube prepared with a water saturated piece of cotton wool beneath a 3 cm diameter disc of lter paper to serve as an egg laying surface. The test tubes were left in a controlled environment for observation. The autotimer switched lights off at 17h00 daily. The test tubes were observed on an hourly basis from 18h00 to 06h00 and the number of eggs laid by each female was then recorded. The eggs per mosquito were counted and recorded. Counting of the eggs was done by viewing the eggs per mosquito under a stereo microscope and counted using a 6-key economy benchtop counter.

Egg Hatch Rate
Eggs from individual females were collected, counted and placed in individual containers containing distilled water. Hatched rst instar larvae were removed and counted daily, up to four days after the rst egg hatched. This life history trait was measured in percentage as the number of rst instar larvae by the total number of collected eggs.

Refeeding rates
The female mosquitoes that had just laid eggs were then offered a blood meal to determine the duration post laying it took to refeed. Feeding was con rmed by observing blood in the abdomen.

Emergence Times and Emergent Sex ratios
The F 2 pupa that developed overnight were counted and used in this aspect of the study. The emergence trial used a thousand 3-day old An. arabiensis pupae. The pupae were put into 5 litre test containers lled with 2 litres of de-ionized water. The test containers were kept in a controlled environment with optimum environmental conditions of 25-27°C and a humidity of 70%. The environmental conditions are maintained using an industrial humidi er and air control system designed speci cally for the insectary. The insectary has an autotimer that gradually switches the lights on or off to mirror the dusk to dawn effect in nature. The autotimer switched the lights off from 17h00 and switched them on from 05h00 daily. Researchers used head lamps on trial day to view the emergence. The test containers were observed on an hourly basis from 18h00 to 06h00 and the newly erupted mosquitoes were collected hourly using a mouth aspirator. The collected mosquitoes were then placed in bug-dorm cages. The male and females were placed in separate cages and counted. A sample of the newly erupted female mosquitoes were removed from the cages and placed individually into a test container. A guinea pig with its fur shaved on the ventral side, was offered as a blood source. The female was then observed for a probing behaviour indicating its willingness to feed off the blood source and the observations of them feeding was recorded.

Results
An analysis of the 12-hour period over which time mosquitoes were observed to have oviposited showed that there are two peaks that corresponds with late evening and again just before dawn (Figure 1).
Periods when most eggs were laid also coincided to the time when most mosquitoes laid eggs (Table 1). Most eggs were also laid in the rst half of the night (18h00-24h00) compared to the second half of the night (z= 29.9, P-value <0.00001, p < 0.05). 58.34% of eggs were laid in the rst half as compared to 41.66 oviposited in the second half (Table 1). Of the 300 mosquitoes that were blood fed and observed only 196 laid eggs. 104 females did not lay eggs which may be an indication of poor blood feeding or egg retention.

Time to refeeding
All females that laid eggs during the course of the night (196) were given access to a blood source and all mosquitoes thus exposed fed on the blood source within an hour of their oviposition. This demonstrated the capacity of females to take a blood meal immediately after egg oviposition to mature the eggs that would contribute to the population of the next generation.

Egg Hatch Rates
Of the 13 594 eggs laid during this study, the hatch rate to viable rst instar larvae were 74.5% resulting in 9 991 offspring with each female producing 52 ± 21.28 larvae (range 14 -97).

Emergence times and ratios
When calculating the absolute number of mosquitoes erupting each hour for the 12-hour observation period, it was found that the majority mosquitoes erupted just after sunset ( Figure 2). The last mosquito emerged from the pupal case just before dawn and no further emergence were recorded after the sun had risen.
Based on the results, unbalanced sex-ratio was observed in the studied population. These eggs tend to give more females than males (z = 2.81, P-value = 0.00248, p < 0.05). Immediately after dusk, almost 25% more females emerged than males and this trend continued until just after midnight. After this period, the number of males and females emerging was similar.

Blood feeding
It was found that newly emerged females took a blood meal. From Table 2 it can be seen that females readily took a blood meal in the rst hour after emergence with a minimum 70% of the exposed females taking blood. Most feeding was completed two hours after exposure, with an average of 2 hours 15 minutes, but this was not the case in the 19h00-20h00 cohort when a single mosquito only took a blood meal in the fourth hour after emergence.

Discussion
Increasing evidence of a change in biting and resting behaviour of the main malaria vectors has been mounting up in recent years as a result of selective pressure by the widespread and long-term use of LLINs and IRS [13,14,15]. Mosquito behaviour is quite variable, with changes in mosquito behaviour posing great challenges to malaria elimination efforts [16,17]. Residual malaria is also in uencing malaria burden in low transmission areas [3].
Anecdotal evidence from Limpopo Province in South Africa shows that the increase in cases may be attributed to the sub-optimal vector control [18]. Since adult control methods have been proven to be insu cient to control the disease, larval source management and especially larviciding is increasingly being recommended as an additional strategy in integrated vector control programmes especially in elimination settings [19]. Mosquito behaviour is quite diverse with distinct preferences in terms of when egg oviposition and adult emergence from the pupal stage occurs [20,21]. It was shown that there are de nite periods during the night when oviposition is favoured. Adult emergence from the pupal cases also occurs at a preferred time in early evening just after sunset. Female mosquitoes are able to take a blood meal soon after emergence as well as immediately after egg-laying. In virgin females, such a blood meal is essential for the development of metabolic reserves prior to mating [22]. Proportionally 25-30% more females are produced which may be due to differential mortality of males and females since climatic variables in uence the lifespan of the sexes [8]. Masters [23] also recorded that most mosquitoes emerged in the early hours of the evening. Newly emerged males are not equipped to mate until genital reorient is complete [24] and An. arabiensis optimal mating occurs with 5-7-day-old males [25].
In order to target the aquatic stages of the mosquito life cycle, a comprehensive knowledge of the larval bionomics is essential. An understanding of mosquito oviposition behavior is necessary for the development of surveillance and control opportunities directed against speci c disease vectors [27]. In this study, vector oviposition and emergence coincide with the bedtime of people living in rural areas as Pates and Curtis [28] found and that the traditional vector control interventions are ineffective as a result of this human behavior. Most people in rural areas go to bed just after sunset (between 20h00 and 22h00) and get up just before sunrise (between 04h00 and 05h00). With vector females laying eggs during these times and with the emergence of a high number of females, there are a large number of vectors that are able to take a blood meal. According to the study by Milali et al [29] this coincides with the peak biting times of Anopheles arabiensis of 21h00 -22h00 and 03h00-04h00.
Approximately 75% of all eggs laid produce viable rst instar larvae and this is in keeping with the observations of the same mosquito populations made by Maharaj [26]. This study complements the study of adult life table characteristics detailed by Maharaj [8,26]. Impoinvil et al [30] found that mosquito eggs held at 22 and 27°C had the highest overall mean hatching count. The temperature at which the eggs in this study was maintained fell within this range so the egg hatch rate was at its optimal. The sex ratio of the emerging mosquitoes Mamfene population reared under ideal conditions showed a clear female bias. Therefore, the population structure of these mosquitoes will require further study before a sterile Insect technique (SIT) programme can be implemented. It will require a sex distortion which is male biases for SIT to succeed [31].
The results of this study can help in streamlining vector control interventions targeting malaria elimination. When temperatures are high and the development of mosquitoes is rapid [26], larviciding can become a costly and labour intensive exercise since the breeding sites are many during summer and the generation time is short [8,26]. However, the use of larval control measures would prevent the rapid buildup of the populations under optimum environmental conditions. There are many proponents for winter larviciding [7] which is the application of chemicals to the breeding sites found during the dry winter season since vectors are in hibernation as larvae and are vulnerable to larval control measures. It was shown that winter larviciding delays the onset of transmission since the mosquito populations emerging from winter hibernation is low [7,26] and hibernating females have undergone gonotrophic dissociation [26].
This study con rms the usefulness of baited traps near larval breeding sites. Immediately upon emergence, both the males and females need to replenish their energy, the males for swarming and the females for mating and host detection. If emerging females were to feed on toxic baits, the population available for blood feeding and transmission would be lowered [16]. However, as this study has shown, newly emerged females as well as those having just oviposited, can also take a blood meal without the bene t of a sugar supplement. Since the females feed opportunistically, livestock treated with endectocides, and placed strategically close to mosquito breeding sites, could be used as a supplementary measure to control females [16]. The study also supports the recommendation from investigations into the use of the sterile insect technique (SIT) that more sterile males be introduced into an area to outcompete non-sterile males [32]. This study has demonstrated that even under optimum conditions fewer males than females are produced and hence the release of a large number of sterile males would dilute the competition from the smaller number of males produced in the wild [33]. Furthermore, male control can be achieved by baits and swarm space spraying [34] However, according to Maharaj [26], there is a slight variation in sex ratio across seasons requiring that SIT programmes constantly monitor the wild male population to adjust the release rate over time.

Conclusion
The results of this study have provided information on egg laying, hatch rates, emergence times, feeding behavior and sex ratios. The sex ratio is biased to females and may be due to differential mortality rates.
The feeding behaviour suggests that the use of attractive toxic sugar baits close to vector breeding sites would be a great tool to eliminate a large number of males and females seeking sustenance soon after emerging or ovipositing. Also, the high larval eclosion rate suggests that targeting larvae for control would serve as an effective population reduction mechanism. The smaller number of males that emerge can also be controlled by baits and swarm space spraying. Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
The study was funded by the South African Medical Research Council by a grant to the corresponding author.
Author's contribution RM developed the concept and the study design. VL and KM conducted the laboratory studies, RM and VL analysed and interpreted the data. RM, VL and KM contributed signi cantly to the write-up of the results.