Resting Behavior of Malaria Vectors in Ghana and Its Implication on Vector Control.

Background: In Sub-Saharan Africa, there is widespread use of long-lasting insecticidal nets (LLINs) and Indoor residual spraying (IRS) to help control the density of malaria vectors and decrease the incidence of malaria in communities. An understanding of the interactions between increased insecticide use and resting behaviour patterns of malaria mosquitoes is important for an effective vector control programme. This study was carried out to investigate the resting behavior, host preference and infection with Plasmodium falciparum of malaria vectors in Ghana in the context of increasing insecticide resistance in malaria vectors in sub-saharan Africa. Methods: Indoor and outdoor resting Anopheline mosquitoes were sampled during the dry and rainy seasons in ve sites that were in 3 ecological landscapes [Sahel savannah (Kpalsogou, Pagaza, Libga), Coastal savannah (Anyakpor) and Forest (Konongo) zones] using pyrethrum spray catches (PSC), mechanical aspiration (Prokopack) for indoor collections, pit shelter and Prokopack for outdoor collections. PCR based molecular diagnostics were used to determine mosquito speciation, genotype for knockdown resistance mutations (L1014S and L1014F), G119S Ace-1 mutation, specic host blood meal origins and sporozoite infection in eld collected mosquitoes. Results: Anopheles gambiae s. l. was the predominant species (89.95%, n = 1,718), followed by An. rupes (8.48%, n=162), and An. funestus s. l. (1.57%, n = 30). Sibling species of the Anopheles gambiae revealed An. coluzzii accounted for 63% (95% CI: 57.10 – 68.91), followed by An. gambiae s. s [27% (95% CI: 21.66 – 32.55)], and An. arabiensis [9% (95% CI: 6.22 – 13.57)]. The mean resting density of An. gambiae s. l. was higher outdoors (79.63%; 1,368/1,718) than indoors (20.37%; 350/1,718) (z = -4.815, p< 0.0001). The kdr west L1014F and the Ace-1 mutations were highest in indoor resting An. coluzzii and An. gambiae in the sahel-savannah sites compared to the forest and coastal savannah sites. Overall, the blood meal analyses revealed a large proportion of the malaria vectors preferred feeding on humans (70.2 %) than animals (29.8%) in all sites. The sporozoite rates was only detected in indoor resting An. coluzzii from the sahel savannah (5.0%) and forest (2.5%) zones. Conclusion: The study reports high outdoor resting densities of An. gambiae and An. coluzzii with high kdr west mutation frequencies, and persistence of malaria transmission indoors despite the use of LLINs and IRS. Continuous monitoring of changes in resting behavior of mosquitoes and implementation of complementary malaria control interventions are needed to target outdoor resting Anopheles mosquitoes in Ghana. mosquitoes and malaria parasites (38). The coastal savannah and forest zone have a bimodal rainfall pattern, allowing for two peaks of malaria transmission while the Sahel zone has a unimodal rainfall pattern making malaria transmission seasonal. Information from this study could provide a better understanding of the impact of current malaria control programme effects on malaria vector population and their interaction on resting behavior. were sampled for mosquitoes. Sampling was done over 4 days each during the dry and rainy seasons. The GPS coordinates of each site was determined and recorded. Malaria vectors resting indoors were sampled using pyrethrum spray catches (PSCs) from 05.00 – 07.00 h (39). The Prokopack aspirator (John W Hock, Gainesville, FL, USA) was used to collect mosquitoes resting indoors and outdoors from 05.00 – 07.00h (40). For indoor collections, mosquitoes resting on the walls and under the roofs of houses or ceilings, under beds were systematically aspirated. Outdoor sampling points included kitchens, granaries, animal resting places and evening outdoor human resting points. Additionally, outdoor resting mosquitoes were collected from pit shelters constructed according to Muirhead-Thomson’s method (41) within 20 m of each selected house. Resting mosquitoes in the cavities created in the pit shelter were collected from 06:00 to 07:00 h by using hand-held mouth aspirators. study revealed high densities of An. and An. gambiae resting outdoors with low genotypic resistance compared to indoor resting mosquitoes that may be triggered by current insecticide-based indoor intervention. Such behavioral change can promote outdoor transmission since current vector strategies mainly target indoor resting malaria vectors. There is a need for further screening of other resistance mutations in this population for better resistance management strategies. Therefore, continuous monitoring of vector behavior in surveillance programmes is recommended and complementary malaria control interventions are needed to control outdoor resting mosquitoes.


Introduction
Malaria is a major public health problem in Africa and was responsible for an estimated 229 million episodes and 409 000 deaths in 2019. In Ghana, malaria is responsible for more than 5.5 million infections and 37 deaths for every 1,000 population (1,2) despite tremendous efforts in the scale-up of vector control interventions particularly in the use of long-lasting insecticidal nets (LLINs) and indoor residual spraying (IRS) (3,4). These anti-vector control interventions previously led to a remarkable reduction of vector population (5-8) and malaria transmission (9,10). However, there are reports of resurgence of malaria in many parts of sub-Saharan Africa (11,12). Some of the challenges implicated in the resurgence is the emergence of insecticide resistance, behavioral modi cation (shift in the biting and resting behavior from indoor to outdoor) and changes in host species preferences from humans to animals (6, [13][14][15]. These challenges have arisen as an adaptation by the malaria vectors response to high use of insecticides for vector control. For instance, following the introduction of LLIN there has been reports of shift in biting behavior of Anopheles gambiae and An. funestus in Kenya (6, 16) and An. funestus in Benin and Senegal (17,18). Whilst the long-term use of LLIN has increased the outdoor feeding proportion of An. gambiae and An. melas in Equatorial Guinea (13), in Tanzania, the long-term use of LLINs was reported to be associated with shifts in the outdoor resting rate of An. gambiae, An. arabiensis and An. funestus (14,19). These behavioral changes however are not consistent with some countries reporting high indoor resting densities of An. gambiae and An. funestus despite long-term use of LLIN and IRS (20)(21)(22). Furthermore, the widespread insecticide resistance in malaria vector populations in Africa is a major threat to current malaria control programmes. Studies from Côte d'Ivoire (23), Togo (24,25), Benin (26), Burkina Faso (27,28), Cameroon (29,30), and Kenya (31,32) have reported high metabolic resistance and target site modi cation of insecticides in the malaria vector. In Ghana, the common target site resistance mechanisms in malaria vectors are the Ace-1 mutation in acetylcholine esterase gene G119S that causes resistance to organophosphates and carbamates, and the voltage-gated sodium channel knockdown resistance (Kdr) which play a major role in resistance to pyrethroids are the most important mechanisms (33)(34)(35).
The primary malaria vectors in Ghana are the Anopheles gambiae sensu lato (An. gambiae s. s., An. arabiensis, An. coluzzii and An. melas) and Anopheles funestus. However, in view of the increasing concern about resurgence of malaria transmission in Africa (36, 37), there is a need to enhance control intervention strategies by having a better understanding of vectors resting and feeding behavior in their different speci c settings and varying seasonal patterns. This is very crucial for the success of the current vector control tools and will provide a guide to improve efforts for the control of malaria in endemic regions.
The objective of this study was to investigate the resting behavior, species composition, insecticide resistance status, and Plasmodium falciparum infections of malaria vectors in three ecological settings of Ghana (the coastal savannah in the south, the forest in the middle, and the Sahel savannah in the north). These ecological zones have varying suitable weather and environmental conditions to sustain the propagation of Anopheles mosquitoes and malaria parasites (38). The coastal savannah and forest zone have a bimodal rainfall pattern, allowing for two peaks of malaria transmission while the Sahel zone has a unimodal rainfall pattern making malaria transmission seasonal. Information from this study could provide a better understanding of the impact of current malaria control programme effects on malaria vector population and their interaction on resting behavior.

Study sites
This study was carried out in ve sites in three ecological landscapes of Ghana; Anyakpor in the coastal savannah zone, Dwease in the forest zone, and Kpalsogou, Libga and Pagaza in the Sahel savannah zone ( Figure 1).
Anyakpor (5° 46'51.96 "N 0° 35'12.84 "E) is a village in the coastal savannah zone, about 5 km west of Ada Foah in Southern Ghana. The coastal savannah has a tropical savannah climate, with an average annual precipitation of 787 mm. Dwease (6° 32'3.05 "N 1° 14'42.22 "W) a village near Konongo, in the Asante-Akim Central district in the middle of Ghana was the site located in the forest zone. The forest zone has a tropical rainforest climate, with an average annual precipitation of 1399.5 mm. The climate in both the coastal savannah and forest area generally consists of a bimodal pattern of rainfall, with the long rainy season from March to June, and the short rainy season from October to November with mean annual temperature of 26.5°C.
Sites in the Sahel savannah ecological zone consisted of Kpalsogou (9° 33'45.2 "N 1° 01'54.6 "W), a village in the Kumbungu district of the northern region, Pagaza (9° 22'33.34 "N 0° 42'29.67 "W) in the Tamale metropolitan area and Libga (9° 35'32.26 "N 0° 50'48.8 "W), a village in the Savelugu-Nanton District. They have a unimodal rainfall pattern from May to November with a mean annual temperature of 28°C, which can get to a maximum of 42°C

Study Design
Mosquitoes were sampled during the rainy season in May for the sites in the coastal and forest zones (Anyakpor and Konongo respectively) and September for the sites in the Sahel savannah zones (Pagaza, Libga, and Kpalsogou) and the dry season in February to March at all the study sites in 2019. Sixteen houses were randomly selected in each study site and during each sampling night four houses were sampled for mosquitoes. Sampling was done over 4 days each during the dry and rainy seasons. The GPS coordinates of each site was determined and recorded.
Malaria vectors resting indoors were sampled using pyrethrum spray catches (PSCs) from 05.00 -07.00 h (39). The Prokopack aspirator (John W Hock, Gainesville, FL, USA) was used to collect mosquitoes resting indoors and outdoors from 05.00 -07.00h (40). For indoor collections, mosquitoes resting on the walls and under the roofs of houses or ceilings, under beds were systematically aspirated. Outdoor sampling points included kitchens, granaries, animal resting places and evening outdoor human resting points. Additionally, outdoor resting mosquitoes were collected from pit shelters constructed according to Muirhead-Thomson's method (41) within 20 m of each selected house. Resting mosquitoes in the cavities created in the pit shelter were collected from 06:00 to 07:00 h by using hand-held mouth aspirators.

Morphological Identi cation
All caught mosquitoes were counted and Anopheline mosquitoes were sorted morphologically according to the identi cation keys of Gillies and Coetzee (42). Sampled mosquitoes were further classi ed according to abdominal status as unfed, freshly fed, half-gravid and gravid. Mosquitoes from each collection method were stored in separately labeled vials with 95% ethanol. Samples were stored at the insectary of the Department of Medical Microbiology, University of Ghana Medical School, Accra, Ghana, until required for further processing.

Sibling Species Discrimination
Sibling species of the An. gambiae s. l. complexes were distinguished using the protocols of Scott et al. (43) and Fanello et al. (44). One leg from each mosquito serving as a DNA template was placed directly into the PCR master mix for Ampli cation.

Detection Of Sporozoites
The head and thorax of individual mosquito samples collected were used to detect the presence of P. falciparum sporozoites using sporozoite Polymerase Chain Reaction (PCR) as described by Echeverry et al. (45).

Detection Of Blood Meal Sources
The abdominal sections of blood-fed Anopheles mosquitoes were cut transversely between the thorax and the abdomen. Genomic DNA was extracted from mosquito abdomens using the ZR DNA MicroPrep™ kit (Zymo Research, CA) following the manufacturer's instructions. One universal reverse primer and ve animal-speci c forward primers (human, cow, goat, pig, and dog) were used for ampli cation of cytochrome b gene encoded in the mitochondrial genome to test for speci c host blood meal origins using conventional PCR (46). Positive controls were included for each host during the PCR analyses and laboratory reared unfed An. gambiae was used as negative control.
Genotyping for kdr and Ace-1 mutations To genotype for kdr mutations, DNA was extracted from mosquito legs using the ZR DNA MicroPrep™ kit (Zymo Research, CA) following the manufacturer's instructions. Standard PCR assays for L1014F kdr allele was used to test the presence of the kdr gene using a modi cation of the protocol by Ahadji-Dabla et al. (2019). In addition, the G119S mutation on the Ace-1 gene was assessed using the PCR protocol described by Weill et al. [23].

Data Analysis
Resting densities of Anopheline mosquitoes was calculated as the number of female mosquitoes per trap/night for each trapping method. The Mann-Whitney U test was used to compare malaria vector density between indoor and outdoor locations. Chi-square was used to test the difference in seasonal abundance and malaria vector species composition between resting locations (indoor and outdoor).
Human blood index (HBI) was calculated as the proportion of blood-fed mosquito samples that had fed on humans to the total tested for blood meal origins. The sporozoite infection rate (IR) expressed as the proportion of mosquitoes positive for Plasmodium sporozoite was calculated by dividing the number of sporozoite positive mosquitoes by the total number of mosquitoes assayed.

Kdr resistance mutations between indoor resting and outdoor resting Anopheles gambiae
A total of 538 Anopheles gambiae s.l samples were genotyped for the presence of L1014S, L1014F and G119S Ace-1 mutations. The L1014F kdr allele was identi ed in 100% (538) of the samples, with the majority of mosquitoes being homozygous for the kdr allele (70.6%; 380/538). Overall, the kdr mutation frequency at Sahel savannah sites (Kpalsogou, Pagaza and Libga) was higher in mosquitoes resting indoors (0.90) than outdoors (0.84). In the forest zone (Konongo), a higher kdr frequency was detected in mosquitoes resting outdoors (0.9) than indoors (0.79). In the coastal zone (Anyakpor), a similar kdr frequency was detected in mosquitoes collected indoors (0.88) and outdoors (0.83). No Kdr L1014S mutation was identi ed in this study.

G119S Ace-1 mutations in indoor and outdoor malaria vectors
The G119S Ace-1 mutations was detected in 79.9% (215/538) of the mosquitoes tested (Table 4). All the mosquitoes with the resistant allele were heterozygous for this mutation. Overall, similar Ace-1 mutations frequencies was detected in indoor and outdoor resting mosquitoes in Sahel savannah zone (0.76 vs. 0.78). In the forest zone, Ace-1 mutations was slightly higher in mosquitoes resting indoors (0.87) than outdoors (0.80), but higher in mosquitoes resting outdoor (1) than indoor (0.76) in the coastal zone.

Discussion
Behavioral diversi cation in vector population threatens to counteract vector control strategies in areas with widespread use of LLINs and IRS (6, 7, [13][14][15]49). This study investigated the behavioral response of malaria vectors for resting, feeding choices and infection rates in the context of increasing insecticide resistance in malaria vectors. Overall, this study revealed high outdoor resting densities of malaria vectors but with lower infection rates. There were high numbers of mosquitoes with insecticide resistance mutations resting indoors compared to outdoors at the different sites in the three ecological zones in Ghana.
Anopheles coluzzii and An. gambiae from Kpalsogou, Libga and Pagaza (Sahel savannah zone) showed increased outdoor resting tendency in both the rainy and dry season despite high coverage and usage of LLINS and IRS) (50). This could drive malaria transmission higher outdoors since these main malaria vectors can bite unprotected humans outdoors and also rest outdoors to avoid contact with the insecticides being used indoors (51,52). This behavioral change in malaria vector population is detrimental to vector control [13,14], which mainly target vectors resting indoors. Such variation in the relative frequency and behavior of this two main malaria vectors has been reported in studies from Equatorial Guinea (13), Tanzania (14,19), but, in contrast to studies from Kenya (20,49) which reported high indoor resting densities of An. gambiae and An. coluzzii.
In this study, it was found that the frequency of kdr west mutation L1014F and the Ace-1 mutation were of higher frequency in indoor resting An. coluzzii and An. gambiae in the sahel-savannah sites compared to the forest and coastal savannah sites. This may be because the interventions indoors are not able to kill the mosquitoes with higher kdr mutations and mosquitoes with less mutations have resorted to feed and rest outdoors to avoid contact with insecticides that are indoors (31,(67)(68)(69). This study is in conformity to previous studies in Ghana (33,34) which reported similar occurrence of Kdr L1014F frequencies in An. coluzzii. This study is also in conformity to previous studies in Ghana (33,34) which reported no Kdr-east allele. However, in contrast to this study, Kdr-east allele (1014S) have been reported in Burkina Faso in An. coluzzii, An. gambiae, and An. arabiensis (53), and in both An. coluzzii and An. gambiae in Togo (25).
The blood meal analyses revealed a large proportion of the malaria vectors preferred feeding on humans than animals in almost all sites. This humanhost choice and higher outdoor resting proportions of An. gambiae and An. coluzzii poses a great concern in malaria elimination efforts due their e ciency in transmitting malaria. This study ndings is in agreement with studies by Orsborne et al. (54) in the coastal area of Ghana, which reported that blood-fed mosquitoes caught indoors had higher HBI and lower BBI than those caught outdoors.
Sporozoite infection was only found in indoor resting malaria vectors collected during the wet season from Kpalsogou and dry season from Konongo. This might mean that malaria transmission is higher indoors than outdoors despite the deployment of LLINs in all sites and IRS in Kpalsogou (55). In contrast, studies from Kenya, Burkina Faso have shown higher sporozoite infection in An. gambiae and An. coluzzii resting outdoors (20,49,56), and in An. gambiae sampled from Northern Ghana (7).

Conclusions
The study revealed high densities of An. coluzzii and An. gambiae resting outdoors with low genotypic resistance compared to indoor resting mosquitoes that may be triggered by current insecticide-based indoor intervention. Such behavioral change can promote outdoor transmission since current vector strategies mainly target indoor resting malaria vectors. There is a need for further screening of other resistance mutations in this population for better resistance management strategies. Therefore, continuous monitoring of vector behavior in surveillance programmes is recommended and complementary malaria control interventions are needed to control outdoor resting mosquitoes. Availability of data and materials

List Of Abbreviations
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
This study was supported by a grant from the National Institute of Health (NIH: R01 A1123074 and D43 TW 011513). The funders had no role or in uence on the design of this study, the collection, analysis, and interpretation of the data collected as well as in writing this manuscript.
Authors' contributions AOF designed, performed the eld and laboratory work, analyzed data and drafted the manuscript. YAA conceived and supervised the study, analyzed data and revised the manuscript. SKA and IAH supervised the study and revised the manuscript. IAH, SBD, ARM and IKS performed eld and laboratory experiments.