South West and North Central Nigeria: Assessment of cassava mosaic disease and field status of African cassava mosaic virus and East African cassava mosaic virus

Abstract Cassava mosaic disease (CMD), caused by cassava mosaic begomoviruses (CMBs), is a major threat to cassava production in Nigeria. The predominant CMBs in Nigeria are African cassava mosaic virus (ACMV), East African cassava mosaic virus (EACMV) and East African cassava mosaic Cameroon virus (EACMCV), which are transmitted through infected stem cuttings and whitefly vectors. This study was conducted in 2015 and 2017 to assess the epidemiology of CMD and the current distribution of CMBs in cassava farms in South West (SW) and North Central (NC) Nigeria. A survey of cassava farms was undertaken, and samples representative of disease symptoms were collected and assessed using molecular techniques. A total of 184 and 328 cassava farms were sampled in 2015 and 2017, respectively. CMD incidence for both regions surveyed was 43.80 and 12.25% in 2015 and 2017, respectively. Fields in SW recorded a higher incidence rate in 2015 (SW: 45.11%, NC: 42.47%), while the reverse occurred in 2017 (SW: 10.90%, NC: 14.01%). Overall, the CMD incidence in Benue State (NC) was significantly higher than other locations surveyed in both years. CMD symptom severity and mean whitefly population were higher in SW Nigeria in the two survey years. ACMV was widespread across both zones, occurring in 79.1% (453/613) and 54.8% (386/704) of cassava leaf samples analysed in 2015 and 2017, respectively. EACMV was detected in only 6.0% (37/613) and 4.7% (33/704) of all cassava leaf samples analysed in 2015 and 2017, respectively. Overall, a higher proportion of infected samples were found in NC in both 2015 (NC: 85.2%, SW: 75.4%) and 2017 (NC: 73.6%, SW: 45.2%). Detection using strain‐specific primers revealed that 97% of EACMV positive samples were indeed infected by the EACMCV strain of the virus. As previously reported, samples with mixed infections showed a higher symptom severity than samples with single ACMV or EACMV infections. This study provides an update to the distribution of CMBs in SW and NC Nigeria and will be useful in development of monitoring and management strategies for the disease in both regions.

useful in development of monitoring and management strategies for the disease in both regions.

K E Y W O R D S
ACMV, Bemisia tabaci, CMD, EACMV, Manihot esculenta Crantz, Nigeria 1 | INTRODUCTION Cassava, Manihot esculenta Crantz, is an important staple food that serves as an affordable source of carbohydrates for over 800 million people across Africa (FAO, 2013). In Nigeria, cassava provides more than half of the daily calorie requirement for people across various ethnic groups (Akinpelu, Amanigbo, Olojede, & Oyekale, 2011). It is considered an important food security crop, mainly because of the ease of production and processing, as well as its ready-to-eat product referred to as "garri" which can be stored for up to 12 months at room temperature (Akinpelu et al., 2011). Besides its importance for food, cassava is also an industrial raw material and thus a potential source of income to farmers and the country at large. Although Nigeria is the largest producer of cassava in Africa, with about 60 million tonnes produced annually (FAO, 2017), the five to 10 t/ha tuber yield common in Nigeria is much lower than the average tuber yield of 25 t/ha obtained in other cassava growing regions around the world (FAO, 2017). One major biotic constraint to cassava production is its susceptibility to cassava mosaic disease (CMD), a viral disease which causes annual tuber yield losses estimated at USD 1.9 to 2.7 billion (Patil & Fauquet, 2009). CMD is caused by a group of viruses commonly referred to as cassava mosaic begomoviruses (CMBs), belonging to the genus Begomovirus in the family Geminiviridae (Ariyo, Koerbler, Dixon, Atiri, & Winter, 2005;Patil & Fauquet, 2009;Thottappilly, Thresh, Calvert, & Winter, 2003). The CMBs are characterised by their circular, bipartite single stranded DNA genome of about 2.7-2.9 kb (Kathurima, Ateka, Nyende, & Holton, 2016). The bipartite genome consists of two components, DNA-A and DNA-B (Haley, Zhan, Richardson, Head, & Morris, 1992). The CMBs are transmitted through use of infected cassava cuttings as planting materials and by whitefly vectors belonging to the Bemisia tabaci complex (Elfekih et al., 2018;Legg et al., 2015). Characteristic symptoms of CMD vary from mosaic patterns on cassava leaves to leaf distortion, vein clearing and stunted growth (Sseruwagi, Sserubombwe, Legg, Ndunguru, & Thresh, 2004). Nationwide surveys conducted in Nigeria for the assessment of CMD andCMB in 2002 and revealed that ACMV, EACMV and EACMCV are the predominant CMBs responsible for CMD (Ariyo et al., 2005;Ogbe, Dixon, Hughes, Alabi, & Okechukwu, 2006). CMBs Witt, Glycine max (L.) Merr. (Alabi et al., 2007Mgbechi-Ezeri, Alabi, Naidu, & Lava Kumar, 2008;Ogbe et al., 2006;Rossel, Thottappilly, Van Lent, & Huttinga, 1987 , South East and South-South), and cassava has various degrees of importance as food and feed in each zone. Approximately 50% of all cassava production in Nigeria takes place in the two of these six zones; NC and SW zones (FAO, 2004;Philips et al., 2005), making these two zones very important for cassava production in Nigeria. Given that the last CMD surveillance activities in these zones took place over ten years ago (Alabi et al., 2007Ariyo et al., 2005;Ogbe et al., 2006), a comprehensive farm and diagnostic survey was undertaken over two years to assess the current status of CMD in these zones. The expected outcome of this survey is an update on CMD incidence, symptom severity, whitefly abundance as well as the distribution of ACMV, EACMV and EACMCV in these regions of Nigeria and this will also inform the development of future monitoring and management strategies for the disease. A harmonised farm sampling protocol was adopted following a previously described method (Sseruwagi et al., 2004). Survey routes followed a road map which allowed sampling of cassava farms in various local government areas of the States. Surveyed cassava farms were a minimum of 10 km apart as described by Ogbe et al. (2006).

| Survey area
The number of cassava farms between sample locations was recorded as a measure of the relative density of cassava, and the cassava varieties planted in each surveyed farm were also recorded. Geo-location coordinates of farms were recorded using a GPS equipment (Garmin Inc., KS).

| Incidence and symptom severity assessment
In each farm sampled, 30 cassava plants were assessed randomly for the presence or absence of CMD symptoms along two diagonals. The CMD incidence was calculated as the percentage of CMDsymptomatic plants in relation to the number of plants assessed. For each cassava plant, CMD symptom severity was scored on a scale of 1 to 5:1 = asymptomatic plants, 2 = plants with 25% of leaves showing mild chlorotic pattern or mild distortion, 3 = infected plants with 50% exhibiting moderate mosaic pattern, narrowing and distortion at base of the leaves, 4 = infected plants with 75% exhibiting severe mosaic symptom, leaf distortion and general reduction of leaf size, and 5 = infected plants with 100% of plants exhibiting severe mosaic, leaf distortion, reduced leaf size, vein clearing and in most cases stunted growth (Sseruwagi et al., 2004). At each farm, a minimum of one and a maximum of four leaf samples were collected from asymptomatic and symptomatic cassava plants of varying disease severity. Leaf samples from weeds showing mosaic symptoms were also collected for assessment of their role as possible alternative host plants for ACMV or EACMV. All samples were stored in herbarium presses prior to laboratory analysis.

| Source of CMD infection and whitefly assessment
The possible source of the observed CMD infection in each plant was determined based on the location of the leaf symptoms. Cassava plants that showed symptoms either only on the lower leaves or on all leaves were assumed to have been infected through the use of infected cassava cuttings. Plants that showed symptoms only on their upper leaves but not on any lower leaves were assumed to have been infected by the whitefly vector (Sseruwagi et al., 2004). The whitefly population on each assessed plant was estimated by counting the number of whiteflies on the three topmost leaves (Fargette, Fauquet, & Thouvenel, 1985;Samura et al., 2014).

| DNA extraction
Total DNA was extracted from cassava leaf and weed samples following the protocol of Dellaporta, Wood, and Hicks (1983). The DNA

| Polymerase chain reaction
PCR was performed using specific primers (Table 1)

| Analysis of field data
Descriptive statistics were used to describe distributions. Continuous dependent variables, such as CMD incidence, CMD severity, percentage of cutting infection and percentage of whitefly infection were checked for normality using the Shapiro-Wilk and Kolmogorov-Smirnov tests alongside a histogram. Kruskal-Wallis test was performed to assess the difference in the distribution of dependent variables across States and years. Pairwise tests were also performed for post hoc Kruskal-Wallis comparisons. Spearman's correlation was used to examine the relationship between continuous variables. Significance was considered to be p < .05 for all tests. Distribution maps were generated using the GPS data alongside CMD distribution information. All statistical analyses were performed using SPSS v20 for Windows and maps were generated using Tableau 10.5.

| CMD incidence and symptom severity
Visual assessment of symptoms showed an overall CMD incidence of 43.80 and 12.25% in 2015 and 2017, respectively (Table 2). In 2015, CMD incidence was higher in the SW (45.11%) but the NC had a higher incidence in 2017 (14.01%). The differences in CMD incidence between the two Zones in 2015 and 2017 were, however, not  (Table 2). Plateau had the lowest CMD incidence in the NC  Note: Values with alphabetical superscripts across years signify a significant difference (p < .05) in mean cassava mosaic disease (CMD) incidence and mean symptom severity between both years.
States for both years, while Lagos and Osun recorded the lowest incidence among SW States in 2015 and 2017, respectively (Table 2).
Symptom severity was moderate in both years. The mean symptom severity was higher in 2015 (2.73) than in 2017 (2.13). The SW Zone had higher symptom severity scores in both years compared to the NC Zone (Table 2). However, this difference between the two Zones was not statistically significant in either year (2015: p = .735, 2017: p = .634). The CMD symptom severity was positively correlated with CMD incidence in 2015 (r = .231, p = .003) and 2017 (r = .373, p < .0001).  (Tables 4 and 5).

| Origin of infection and adult whitefly distribution
EACMV was detected at significantly lower rates than ACMV in both Zones and in both years. EACMV was detected in only 6.0%  (Table 6).  cassava variety planted in the surveyed fields were documented based on common names provided by the farmers. This data was collected for 92% of the fields surveyed in 2017 but were not recorded where the farmers were unavailable in the farm during the survey and/or where the variety is not one of the common ones. Most farms planted one cassava variety, but in some cases, more than one variety was present in a farm. We recorded only the major cassava variety in each farm. A total of 12 cassava varieties were recorded. Over 75% of the varieties recorded were local varieties with "Akpu" and "Okowayo" as the predominant varieties recorded in both Zones (Table 7). There were two improved varieties as reported by the farmers; the "Agric" and the "TME 419." CMD incidence varied by variety with fields cultivating the Banada variety having the highest CMD incidence and highest proportion of cutting borne infections (Table 8). Symptom severity was generally low in the survey region and as such did not vary significantly by cultivar. Local varieties, however, showed more severe symptoms than the improved varieties. Whitefly abundance varied with fields planting "Agric," "Banada" and "Akpu" having higher whitefly populations than other fields (Table 8). The observed higher whitefly population, however, was not correlated with a high proportion of whitefly borne infections on fields planting these varieties.

| Cassava cultivars
The distribution of ACMV and EACMV varied across the various varieties ( Figure 4). The highest proportion of ACMV and EACMV infected samples were collected from fields cultivating Danwari

| DISCUSSION
This study shows the presence of CMD in surveyed cassava farms with incidence of <50% in both years and symptom severity varying from mild to very severe as was reported in previous CMD surveys in Nigeria (Ogbe et al., 2006). Findings from this study, however, showed no correlation between whitefly population and CMD incidence as  Mwaura, & Obare, 2015) and the Central African Republic (Zinga et al., 2013). This implies that farmers are either unaware of the need for use of virus-free planting materials or have limited access to virusfree planting materials. In the absence of virus-free planting materials, farmers can be trained on how to recognise CMD symptoms and select healthy cuttings for the next planting season (Mallowa, Isutsa, Kamau, Obonyo, & Legg, 2006;Mulenga et al., 2016;Thresh & Cooter, 2005). Studies conducted by Nyirahorana et al. (2017) in Rwanda showed that CMD symptom recognition can be improved by establishing demonstration plots for farmers. Furthermore, the implementation of good agricultural practices such as routine weeding of cassava farms will contribute to successful management of CMD since some of the weeds in farms can be alternative hosts for CMBs and perhaps sources of innoculum for whitefly spread of the virus.
In addition to CMD occurrence in the sampled cassava farms, high incidences of cassava green mite and cassava mealybugs were noted in several States in 2017. Studies have shown that mild mosaic symptoms are masked by leaf discolorations caused by green mites (Zinga et al., 2013), and mealybugs cause the characteristic bunching top symptom which makes CMD leaf distortion more severe (Parsa, Kondo, & Winotai, 2012). These may result in under-or overreporting of CMD incidence and symptom severity.  Ariyo et al., 2005;Ogbe et al., 2006) and to other studies in West Africa (Pita et al., 2001;Torkpo, Offei, Danquah, & Gafni, 2017 Over 97% of the EACMV samples detected were found to be EACMCV which is no surprise given that Ogbe et al. (2003) already confirmed the high similarities between EACMCV and Nigerian EACMV isolates. In this study, the overall proportion of EACMV infections was low (<7%) which further points to the low prevalence of EACMV in Nigeria.
The low occurrence of EACMV in Nigeria is probably because EACMV has not been present in Nigeria as long as ACMV (Ogbe et al., 2006). In this current study, single CMD infections caused by EACMV were observed in both Zones particularly in Ondo, Oyo and Benue States. This is, however, contrary to previous study by Ogbe et al. (2006) where EACMV single infection was only observed in Niger State. This implies that overtime if adequate management efforts are not implemented, EACMV may become even more widespread.
Although the percentage of mixed infections seems to have reduced, the presence of mixed infection poses a risk of genetic recombination and has the potential to compound the problem if a more virulent CMB strain emerges (Berrie, Palmer, Rybicki, & Rey, 1998;Mulenga et al., 2016).
In addition to the potential of genetic recombination that can result from mixed infections, mixed infections also cause increased symptom expression in infected plants (Fondong et al., 2000). In previous studies conducted in Nigeria, plants with mixed CMB infection resulted in more severe symptoms than plants with single infection of either ACMV or EACMV infections (Ogbe et al., 2006). Similarly, in the current study, plants that had mixed ACMV and EACMV infections also had higher symptom severity scores as compared to plants that were infected with either ACMV or EACMV alone. Severe symptoms caused by mixed infections was also observed in Cameroon (Fondong et al., 2000), Cote d'Ivoire (Pita et al., 2001), Zambia (Chikoti et al., 2013), Kenya (Mwatuni et al., 2015), Tanzania and Uganda (Harrison, Zhou, Otim-Nape, Liu, & Robinson, 1997). Increased symptom severity in mixed infection is attributed to the synergistic relationship between strains of CMBs involved in the mixed infection and an increase in plant virus titre (Naseem & Winter, 2016). It is, however, important to note while symptom expression is increased in mixed infection, it is also dependent on the virus strain infecting the plant and the variety of cassava planted (Ogbe et al., 2003).
In this study, approximately 35% of the asymptomatic cassava In addition to the presence of CMBs in asymptomatic samples, symptomatic samples that were unreactive to any of the primer pairs T A B L E 9 Proportion of ACMV, EACMV and mixed (ACMV + EACMV) infections in samples from fields planting each cultivar utilised in this study were also observed. This is similar to previous reports (Aloyce, Tairo, Sseruwagi, Rey, & Ndunguru, 2013;Chikoti et al., 2013;Harimalala et al., 2015;Mulenga et al., 2016;Ogbe et al., 2006;Zinga et al., 2013), where unidentified CMB was reported to have caused the CMD symptoms observed in cassava farms assessed. Further studies are needed for the detection of the virus(es) responsible for the symptoms observed in these plants.

| CONCLUSION
This study provides an updated status of CMD incidence in cassava fields and the distribution of ACMV, EACMV and EACMCV in SW and NC Zones of Nigeria where nearly 50% of cassava production takes place in Nigeria. The study confirms that ACMV is still the predominant CMB strain in the region as EACMV occurred in low percentages in both Zones. The presence of CMB in asymptomatic samples further buttresses the need for a functional clean seed certification system in Nigeria. Furthermore, the presence of symptomatic samples that were unreactive to primer pairs utilised in this study highlights the need for further study to determine the etiology of the CMB involved.