Distribution and diversity of alternate hosts of Maruca vitrata Fabricius in three West African countries

The evolution of resistance to the Bacillus thuringiensis (Bt) toxins by insect pests is a major threat to Bt technology. However, the rate of resistance can be slowed with appropriate integrated insect resistance management (IRM) strategies. Surveys were conducted to identify alternate host species for Maruca vitrata (commonly called the legume pod borer or Maruca) that could serve as refuges for Pod-Borer Resistant (PBR) cowpea in three West African countries (Ghana, Nigeria, and Burkina Faso). Survey sites included 25 in northern Ghana, 44 in northern Nigeria, and 52 in north-central and southwestern Burkina Faso. Alternate hosts of Maruca identified from plant species belonging to the Fabaceae family that showed signs of Maruca damage on cowpea tissues were collected and dissected. Larvae that were found during these dissections were reared to adult moths in the laboratory then identified to species. The alternate host plants including species of Crotolaria, Sesbania, Tephrosia, and Vigna were the most frequently encountered among sites and locations. Flowering and podding of these plants overlapped with flowering and podding of the nearby (~200 m) cowpea crop. Abundance of these wild hosts and overlapping flowering patterns with the cowpea crop in most locations have the potential to sustain ample numbers of Bt susceptible Maruca that will mate with possible resistant Maruca and deter resistance development. Further quantitative studies, however, are required from each location to determine if actual Maruca production from alternate hosts is sufficient for a PBR IRM strategy. If verified, this approach would be compatible with the high dose/refuge IRM strategy that includes alternate hosts and non-Bt crops as refuges.


Introduction
Cowpea, Vigna unguiculata (L) Walp., is an important food for people and feed for livestock in Sub-Saharan Africa. However, cowpea production has major insect pest constraints especially by Maruca vitrata Fabricius (Lep. Crambidae) (commonly called the legume pod borer or Maruca; latter used hereafter), which can reduce yields 20-80% (Singh et al. 1990). Controlling Maruca is a challenge for smallholder farmers (Jackai and Adalla 1997). Many growers cannot afford chemical insecticides and when some synthetic insecticides, such as cypermethrin, dimethoate, and endosulfan have been used, Maruca populations have evolved resistance (Bottenberg 1995;Ekesi 1999). Also, many growers lack the proper safety equipment and, because of literacy challenges, do not know how to handle these compounds safely (Onstad et al. 2012;Jepson et al. 2014). Moreover, despite years of research, no cowpea varieties with natural resistance to Maruca have been identified and, when chemical insecticides are used, often 5-8 sprays are required in one cropping season (Fatokun 2002). The chemical insecticide-use challenge and the possibility that Maruca could evolve resistance to available insecticides require other options for pest control in cowpea.
One promising solution to control Maruca is to use genetically engineered (GE) cowpea varieties that provide protection against insects. The lepidopteran specific cry1Ab gene expressed in Pod-Borer Resistant (PBR) cowpea, also known as Bacillus thuringiensis (Bt) cowpea, has demonstrated near-complete control of Maruca in field trials in West Africa (Ba et al. 2019;Addae et al. 2020). Bt cowpea has now been approved by regulators for commercial release in Nigeria (NBMA 2019).
Insect resistance to GE plants in other parts of the world, however, suggests that an insect resistance management (IRM) plan is necessary to protect PBR cowpea from selecting for resistance in Maruca populations (Addae et al. 2020). The African stalk borer, Busseola fusca (Fuller), has evolved resistance to Bt maize in South Africa (Van Rensburgh 2007). Also, other moths (Lepidoptera) have evolved resistance to GE crops, for example, the fall armyworm, Spodoptera frugiperda (J.E. Smith), to Bt maize in Puerto Rico, Brazil, and United States (Tabashnik et al. 2009;Storer et al. 2010;Huang et al. 2011). Developing an integrated IRM plan that includes high-dose expression of the trait and use of refuges is recommended as an effective strategy for preventing insect resistance (Carrière et al. 2010;Campagne et al. 2016). The size of the refuge will depend on the distribution of alternate hosts of Maruca in the locations where the PBR cowpea will be grown. Therefore, the objective for this study was to identify the distribution and diversity of alternate hosts in West Africa, specifically Ghana, Nigeria, and Burkina Faso, where cowpea is commonly grown. Determination of alternate hosts will provide essential parameters for developing an IRM strategy that will delay the evolution of resistance to the Bt toxin in Maruca populations.

Materials and methods
Surveys of alternate hosts were conducted in Ghana, Nigeria, and Burkina Faso during October 2014, which is within the cowpea cropping season. There were 25 sites in the Tolon-Kumbungu location in northern Ghana; 25 sites near Bunkure, Kano State and 19 sites near Bomo, Kaduna State in northern Nigeria (44 total sites); 12 sites near Ziniaré, Oubritenga Province in northcentral Burkina Faso and 40 sites near Bobo Dioulasso, Houet Province in southwestern Burkina Faso (52 total sites). The surveys were carried out in three agro-ecological zones, which in some cases overlap: Guinea Savanna (Tolon-Kumbungu and Bomo), Sudan Savanna (Bunkure, Ziniaré, and Bobo Dioulasso), and Sahel Savanna (Ziniaré).
Survey sites were non-systematically selected at each location. Global Positioning System (GPS) coordinates using decimal degrees (DD) were used to identify the location of the alternate hosts. Assessment of alternate hosts within 200 m of cultivated cowpea farms began by using the transect intercept method of sampling vegetation as described by Greig-Smith (1983). At the beginning of each sampling point, a 200 m line, referred to as baseline, was established in a designated direction, south-north in this example (Fig. 1). The baseline (200 m) was divided into ten equal parts (marks) starting from 0 to the 10th mark (the end of 200 m length). Three sets of numbers from the ten designated marks were randomly generated and then allocated to each baseline. Three line transects (transects 1, 2, and 3) of 50 m each were established starting at the marked points, running perpendicular to the baseline and parallel to one another, in this example a west-east direction. Sampling was done along each transect for terminal shoots, flower buds, flowers, or pods of all Fabaceae plant species. These plants were carefully examined for the presence of Maruca larvae or Maruca-like damage. This procedure was repeated at all the survey points, that is, 25 times in Ghana, 44 in Nigeria, and 52 in Burkina Faso. This resulted in an assessment of alternate hosts in areas within 200 m of cultivated cowpea that included grazing lands, forests, and other cultivated crops (Fig. 2).
All species belonging to the Fabaceae intercepting the transects were recorded. Crown cover estimates (proportion of the ground cover occupied by a perpendicular projection of the individual plant species) of each plant intercepting the transects were recorded (Greig-Smith 1983). The intercept lengths for each species were summed and each of these values was divided by the total length of the transect. These values then were converted to coverage percentages for each species.
The formula used for calculating plant cover was: Plant cover% = Σi∕Tl × 100 Σi = sum of individual plant intercept lengths Tl=Total length of the transect line Samples of potential alternate hosts found within 200 m from the border of cultivated cowpea fields were examined for Maruca damage symptoms on terminal shoots, flower buds, flowers, and pods, which included feeding scars, tunnels, and evidence of frass. Infested plants were dissected for recovery of Maruca larvae. Plants from which the larvae were collected were then identified to species. The larvae were identified by the characteristic translucent, shining body with six rows of black spots running from thorax to abdomen and a dark brown head. The adults were identified by three white patches with black margins on the brown forewings (Sharma et al. 1999). In cases where the infested plants could not be identified readily in the field, voucher specimens were collected, wrapped in tissue paper, placed in plastic bags, and then sent to the authors' respective research institutes for identification. Plants were identified to genera and, if possible, species using the following references: Akobundu 1987; Akobundu and Agyakwa 1987. Cowpea field infestations were estimated in Ghana and Nigeria by inspecting 100 randomly distributed cowpea plants for Maruca infestation damage on terminal shoots, flower buds, flowers, and pods. Sampling of Maruca larvae on cowpea farms was timed generally to coincide with flowering. Cowpea plants can compensate for flower removal, so gently removing flowers during sampling reduces plant destruction and limits yield loss (Abudulai and Shepard 2003).

Laboratory rearing and identification of Maruca
The Maruca larvae collected were sent to the laboratory for rearing and identification. Each specimen was placed in a glass vial containing artificial diet made from a modified Ostrinia nubilalis Hübner diet obtained commercially from Bio-Serv Company, Flemington, NJ. USA (Bio-Serv product No. F9478B-M) without corncob grits but supplemented with cowpea flour. The vials were plugged with cotton wool, labeled, and incubated at 25 ± 1 °C and 60-70% relative humidity (RH) until adult (moths) emergence. Maruca moths were identified by entomologists at each institute. If a Maruca adult was confirmed then the host plant was recorded; otherwise, the plant was not recorded.

Results
The results for distribution and diversity of alternate hosts of Maruca in five locations within three countries in West Africa are presented in Tables 1-5. The survey found many Fabaceae plant species in areas where cowpea and    Maruca coexist. The Fabaceae plant species found in the survey areas were classified into three categories based on: 1) presence of Maruca larvae on the plants classified as alternate host, 2) only Maruca-like damage found (potential alternate host), and 3) none, where neither Maruca nor Maruca-like damage, was found. These three categories (larvae, damage, and none) are presented in the Tables 1-5. At the Tolon-Kumbungu location in Ghana, nine taxa of plants-namely, Senna obtusifolia L., Tephrosia linearis (Willd.) Pers., Tephrosia bracteolata Guill. & Perr., Corchorus olitorius L., Stylosanthes spp., Pueraria phaseoloides (Roxb.) Benth., Crotalaria quinquefolia L., Cajanus cajan (L.) Millsp., and Crotalaria retusa L. or combinations of them-were found at 24 of the 25 survey sites ( Table 1). Most of the plant species were at flowering or podding stages except Stylosanthes spp. that was at the vegetative stage. Besides Stylosanthes spp, all of them had either only Maruca damage symptoms or presence of Maruca larvae. No alternate host plants were found at site 4. Numbers and mean cover percentages (n, x̅ ± SE) for the four most common plant species found at this location were S. obtusifolia (20, 0.96 ± 0.16), T. linearis (9, 0.83 ± 0.08), T. bracteolate (5, 0.92 ± 0.06), and Corchorus olitorius (5, 0.34 ± 0.04). The average number of potential host plants (larvae, damage, and none) found along the three transects for each site was 9.3 with 1.5% average plant cover (Table 1) (Table 2). Only two locations had the latter two species. Most of the plant species were at flowering or podding stages except C. senegalensis species that was at vegetative stage. All the plant species found had either presence of Maruca larvae or only damage symptoms. No alternate host plants were found at sites 8, 9 and 16. The most common species found at this location were S. sesban (12, 1.29 ± 0.21), C. senegalensis (10, 1.35 ± 0.19), and T. bracteolata (8, 0.81 ± 0.17). The mean number of host plants found along the three transects for each site was 8.4 with 1.5% average plant cover ( Table 2). The average number of Maruca larvae on 100 cowpea plants growing in cultivated cowpea fields was 7.5.
At the Bomo location in Nigeria, four plant speciesnamely, V. racemosa, C. senegalensis, Centrosema pubescens Benth., and Sesbania sesban, or combinations of them-were found in 11 of the 19 survey sites (    stages except C. pubescens that was at the vegetative stage. Also, all the plant species found had either presence of Maruca larvae or only damage symptoms with the exception of two species, C. senegalensis, and C. pubescens, where neither larvae nor Maruca-like damage were found. No alternate host plants were found at sites 1, 4, 5, 9, 10, 12, and 13. The numbers and mean cover percentages for the most common plant species found at this location were V. racemosa (7, 0.86 ± 0.14) and Crotalaria senegalensis (4, 0.63 ± 0.06). The mean number of potential host plants found along the three transects for each site was 4.5 with 1.3% average plant cover (Table 3). The average number of Maruca larvae found on 100 cowpea plants growing in cultivated cowpea fields was 5.2.
Three species of alternate hosts-namely, C. retusa, S. pachycarpa, and T. bracteolata, all of which had Maruca larvae present-were found at the Ziniaré location in Burkina Faso (Table 4). T. bracteolata was found only in two locations. The other two species either occurred alone or at two locations in combination. Numbers and mean cover percentages for the most common plant species found at this location were C. retusa (8, 5.1 ± 2.20) and S. pachycarpa (4, 3.1 ± 1.77). The mean number of host plants where Maruca was found along the three transects for each site was 5.1 with 7.6% plant coverage. Most of the plant species were at the flowering stage with a few also at the podding stage. Maruca larvae were not counted in cultivated cowpea fields in Burkina Faso.

Discussion
Our study aimed to assess the distribution and diversity of alternate host plants for Maruca in selected, important cowpea-growing countries in West Africa. Generally, the alternate host plants including Crotolaria, Sesbania, Tephrosia, and Vigna, species were the most frequently *V Vegetative, F Flowering, and P Podding a Maruca larvae or Maruca-like damage found, plants classified as alternate or potential alternate hosts, respectively   (Tamò et al. 2002;Arodokoun et al. 2003), corroborating our findings. There were instances where only damage symptoms were observed on the plants and those plants were classified as potential alternate hosts of Maruca. Also, there were plants where neither Maruca larvae nor Maruca-like damage were found. This occurred especially at the Bomo (Nigeria) location. These same plant species, however, except for Centrosema pubescens, had Maruca larvae or damage at other locations. There were sites where no plant species were found because of scarce vegetation or poor accessibility, especially at Bomo. Unlike the higher numbers of Maruca larvae found on the cowpea plants in Ghana, few Maruca larvae were found on cowpea plants in Nigeria because the growers had sprayed their crops with chemical insecticides prior to conducting the survey or the pest pressure was low that year. Another challenge was sorting through the many plant species that were encountered during the surveys. Future studies could address this by focusing on the major known alternate species at a location.
The extensive distribution and diversity of the alternate plant species within the surveyed locations indicate that they could be used as refuges when Bt cowpea is introduced. Many Maruca larvae were found on both cowpea plants on the farms and alternate hosts growing within 200 m of the farms in Tolon-Kumbungu (Ghana), and Bunkure (Nigeria) locations. While the presence of Maruca larvae on cowpea plants was not determined in Burkina Faso, earlier studies by Traore et al. (2014) reported that Maruca populations overlapped on cultivated cowpea and alternate hosts plants during the rainy season in southwestern Burkina Faso. These observations suggest that adult Maruca emerging from cultivated cowpea and alternate host plants in close vicinity are very likely to mix and mate. Small farms embedded in areas with alternate hosts particularly could benefit. For example, a one-hectare farm (10,000 m 2 ) potentially could attract Maruca from a radius of 200 m (or more). In this case, Maruca would be attracted from a minimum area of 125,600 m 2 , which is more than 12 times the size of the farm. Likewise, half, quarter, and one-tenth hectare farms would attract Maruca from proportionally larger areas, that is, more than 25, 50, and 125 times larger, respectively. Consequently, when Bt cowpea is planted in similar environments, resistant moths could mate with susceptible moths, which lowers the frequency of the resistance alleles    (Roush 1997;Gould 1998). This random mating of moths will help delay and perhaps prevent Maruca from developing resistance to the Bt toxin. In this study, many Maruca larvae were found on alternate hosts during the rainy season. Traore et al. (2014) reported that during the dry season, Maruca maintained a permanent population on the wild host plants Mucuna poggei Taub. and Daniella oliveri Rolfe in Burkina Faso. Arodokoun et al. (2003) also found Maruca in the dry season in southern and central Benin. There is evidence that Maruca moths migrate to the north during the rainy season and disperse from local wild hosts (Bottenberg et al. 1997;Ba et al. 2009;Margam et al. 2010).
Utilization of wild hosts plants can be an effective component of IRM strategies for transgenic crops (Jackson et al. 2008). Establishing refuges, however, will be required at some locations where alternate hosts plants are not found or production is low. In fact, Margam et al. (2010) after surveying a total of 67 sites in proximity to cowpea fields in the Sudan Savannah (Kano, Nigeria), the Northern Guinea Savannah (Zaria, Nigeria), and the Sahel Savannah (Maradi, Niger) ecological zones, indicated that alternate host plants for Maruca are scarce or absent during the cowpea-growing season in these areas. Their findings were inconsistent with earlier studies that showed an abundance of alternate hosts in West Africa (Atachi and Djihou 1994;Tamò et al. 2002;Arodokoun et al. 2003). They hypothesized that the lack of alternate hosts in their survey was likely due to the northerly arid region of the survey. They concluded that since Maruca is migratory into those northerly arid regions and becomes locally extinct at the end of each growing season, the lack of alternate hosts would not impact IRM strategies (Margam et al. 2010). While our study had some geographical overlap with the Margam et al. (2010) study, our survey areas were focused more in the semi-arid cowpea zones where more alternate hosts would be expected (Ba et al. 2019). It is noteworthy that our study was qualitative and only identified the potential for Maruca production. Further quantitative studies are needed to determine Maruca population sizes that originate from alternate hosts relative to those that originate from cultivated cowpea in order to model their potential for use in an IRM plan.
Another way to increase the number of Maruca that originate from alternate hosts includes the use of a cultivated crop, e.g., pigeon pea, Cajanus cajan. This crop is an alternate host for Maruca in Ghana and could be planted as a refuge as well as food for the growers. The refuge can be comprised of cultivated non-transgenic crop plants or perhaps any other host plants that can support significant population sizes for the targeted insect pest species (Agunbiade et al. 2014). Vacher et al. (2003) suggested refuge fractions of less than 25% will minimize pest density while efficiently delaying resistance. Further studies should be done to ascertain the appropriate refuge percentage required for Bt cowpea to ensure effective insect resistance management. The high dose/refuge strategy is considered central to managing resistance to Bt toxins (Carrière et al. 2010;Campagne et al. 2016). Pod-borer Resistant cowpea has near complete control of Maruca in field tests in West Africa (Addae et al. 2020). Breeding is currently in progress to introduce a second Bt gene in addition to the cry1Ab gene at research institutes in West Africa (Bett et al. 2017). Two or more genes in a crop has been reported to delay development of resistance to Bt toxins better than a single gene. The second gene, however, does not need to be high dose (Head and Greenplate 2012;Onstad et al. 2012). Other information required for fine-tuning an IRM strategy should include a better understanding of pest biology/ecology including interactions with biocontrol agents, product deployment patterns, local cropping systems, and plans for insect susceptibility monitoring, grower communications on compliance, and remedial action should resistance occur.

Conclusion
There are abundant and diverse alternate hosts for Maruca in the surveyed locations where cowpea is grown in Ghana, Nigeria, and Burkina Faso. The alternate hosts with their flowering and podding patterns may be adequate for most of the sites in each location to serve as a refuge to delay Maruca from developing resistance to the Bt toxin until a Bt cowpea variety with two insecticidal genes is deployed. This qualitative study suggests there is potential for adequate refuge from alternate hosts, but follow-up quantitative studies are necessary to verify this. Refuges should be established where alternate hosts are not available or uncertain. A robust IRM strategy is required for growing Pod-borer Resistant Bt cowpea. This strategy should include abundant alternate hosts for Maruca. However, in areas where alternate hosts are limited or missing, growers' compliance to produce non-Bt cowpea will be needed. Future studies could determine better estimates of the population size of Maruca in the refuge as compared to those of Bt cowpea. Such insights would allow effective modeling to understand the longterm potential for the evolution of resistance in Maruca populations in Bt cowpea.