Characterization of virus species associated with sweetpotato virus diseases in Burkina Faso

Abstract Sweetpotato (Ipomoea batatas) production in sub‐Saharan Africa is severely affected by viral diseases caused by several interacting viruses, including sweet potato feathery mottle virus (SPFMV), sweet potato chlorotic stunt virus (SPCSV), and sweet potato leaf curl virus (SPLCV). However, the aetiology of viral symptoms on sweetpotato is rarely established in most countries in Africa. Here, we aimed to investigate and characterize the incidence of sweetpotato viruses in Burkina Faso. We performed a countrywide survey in 18 districts of Burkina Faso and collected 600 plants, with and without symptoms, from 80 fields. Viral strains were identified using nitrocellulose membrane‐ELISA, PCR, and reverse transcription‐PCR. Three scions from each of 50 selected plants with symptoms were grafted to healthy Ipomoea setosa and then serological and molecular tests were performed on the 150 recorded samples. Three viruses were detected: 24% of samples were positive for SPFMV, 18% for SPLCV, and 2% for SPCSV. Across all diagnostic tests, 40% of all plant samples were virus‐negative. Coinfections were found in 16% of samples. Partial sequences were obtained, including 13 that matched SPFMV, one that matched SPLCV, and one that matched SPCSV. All identified SPFMV isolates belonged to either phylogroup B or A‐II. The SPCSV‐positive isolates had 98% gene sequence homology with SPCSV‐West Africa for the coat protein. Begomovirus‐positive isolates clustered with SPLCV‐United States. This first study of sweetpotato viral diseases in Burkina Faso indicates widespread occurrence and suggests a need for further epidemiological investigations, breeding programmes focused on virus‐resistant varieties, and improved farming practices to control disease spread.

by far the most important (Valverde et al., 2007). More than 30 viruses have been reported to infect sweetpotato, and several of these are recently described DNA viruses belonging to the families Geminiviridae and Caulimoviridae (Clark et al., 2012;Cuellar et al., 2015).
In Burkina Faso, sweetpotato is the most important root crop and ranks third, after cereals and legumes, among all crops (Dabiré and Belem, 2001). Sweetpotato production in Burkina Faso increased from 13,618 t in 1998 to 167,550 t in 2013167,550 t in (FAOSTAT, 2013. In western Burkina Faso, sweetpotato is grown as a cash crop. However, because sweetpotato is propagated vegetatively, it is prone to the accumulation of viruses and other pathogens (Souto et al., 2003;Cuellar et al., 2015). Farmers can lose up to 100% of the crop yield; losses differ depending on the cultivar, the infecting virus, the stage of infection, and whether the crop is infected with one or more viruses (Valverde et al., 2007;Rey et al., 2012;Loebenstein, 2015). Given the susceptibility of sweetpotato to virus infection, coinfections are common and cause a generalized disease known as sweet potato virus disease (SPVD; Valverde et al., 2007).
The SPFMV genome consists of a single positive-sense RNA strand with a poly(A) tail at its 3′ terminus and a genome-linked protein (VPg) at its 5′ terminus (Wylie et al., 2017). The genome contains a large open reading frame (ORF) that encodes a polyprotein and the PIPO ORF (Wylie et al., 2017). Previous studies on the diversity of the coat protein (CP) genomic region revealed four phylogenetic lineages: East African (EA) from Tanzania, Uganda, and Kenya; russet crack (RC) from Australia, Africa, Asia, and North America; ordinary (O) from Japan, China, Korea, Niger, Nigeria, and Argentina; and sweet potato virus C (as renamed by Adams and Carstens, 2012) from the USA, China, Australia, Africa, East Africa, and Argentina (Untiveros et al., 2008;Adams and Carstens, 2012). With increased understanding of SPFMV and the biosafety implication of its nomenclature, a neutral classification system has been proposed that uses Latinized numerals to take account of biological and geographical considerations (Jones and Kehoe, 2016). Consequently, SPFMV EA became A-I, O became A-II, and RC became B (Maina et al., 2018).
SPFMV has been characterized in East Africa (Kenya, Tanzania, and Uganda) and West Africa (Nigeria and Niger), where it is a known threat to sweetpotato production (Karyeija et al., 2001;Untiveros et al., 2008).
Sweet potato chlorotic stunt virus (SPCSV) is a bipartite member of the family Closteroviridae, genus Crinivirus. Like other criniviruses, it is phloem-limited (Cohen et al., 1992;van Regenmortel et al., 2000;Loebenstein, 2015). SPCSV has a worldwide distribution (Karyeija et al., 1998) and is present in the main sweetpotato production areas in Africa, and usually coinfects its host plant with other viruses in the field (Mukasa et al., 2003;Tairo et al., 2004).
Coinfection makes separation and purification of SPCSV difficult, thus determination of the SPCSV genomic sequence has been greatly constrained (Qin et al., 2014). To date, only two SPCSV strains have been identified: SPCSV-East African (SPCSV-EA) and SPCSV-West African (SPCSV-WA). The SPCSV-EA strain is more widespread and more studied than SPCSV-WA. Studies have shown that most sweetpotato-infecting viruses can cause severe synergistic disease complexes when they coinfect with SPCSV, leading to increased yield losses. These synergistic disease complexes may include RNA viruses of genera Potyvirus, Ipomovirus, Carlavirus, and Cucumovirus, as well as DNA viruses of genera Begomovirus, Cavemovirus, and Solendovirus, that are symptomless when they infect singly Mukasa et al., 2006;Untiveros et al., 2007;Cuellar et al., 2011;2015). Thus, SPCSV can increase the susceptibility of sweetpotato to a wide range of viruses (Cuellar et al., 2009;2015).
As Burkina Faso is now actively including orange-fleshed sweetpotato in its food security strategy, there is increased focus on major threats to sweetpotato production, particularly from viruses. Symptoms resembling those induced by SPVD have been observed in sweetpotato fields in the main production areas in Burkina Faso, but the viruses associated with SPVD have never been confirmed.
A study in 2010 by the Pan-African Sweet Potato Virome project (http://bioin fo.bti.corne ll.edu/virome) aimed to understand the diversity, distribution, and evolution of sweetpotato viruses across Africa. However, only a few samples from one site in Burkina Faso were available to that project. Nevertheless, this small sample identified the presence of sweet potato leaf curl virus (SPLCV), SPCSV, and SPFMV.
The aim of our countrywide study in Burkina Faso was to assess the presence of viruses in sweetpotato-growing areas and to establish the aetiology of the symptoms observed in the sweetpotato fields surveyed. The knowledge generated will contribute to the current body of knowledge of sweetpotato viruses and will inform decisions concerning production of this crop in Burkina Faso.  Table 1). The leaf samples were immediately put into paper envelopes, dried at 37 °C, and stored at the Environmental and Agricultural Research Institute laboratory located at Kamboinsé Research Station. Cuttings (10-15 cm length) were grown in plastic pots containing an autoclaved mixture of soil, sand, and peat moss (equal volumes) and maintained in an insect-proof greenhouse.

| Nitrocellulose membrane ELISA for virus diagnosis
Fresh leaf samples from stem cuttings growing in the insect-proof greenhouse were subjected to a nitrocellulose membrane ELISA  Table 1. (b) Locations with samples positive for SPCSV, SPFMV, and SPLCV by molecular diagnostic methods. The smallest, single-colour circles represent one specific virus identified at one farm; medium single-colour circles indicate one specific virus identified at more than one farm; largest segmented circles indicate the detection of more than one virus at more than one farm. Viruses detected in each location are shown in Table 1. (c) Locations with samples positive for mixed infection of SPFMV + SPCSV by molecular diagnostic methods. The smallest circles represent one farm with SPFMV + SPCSV infections; medium circles indicate >1 farm with SPFMV + SPCSV infections. Locations with samples positive for mixed infections of SPFMV + SPCSV are shown in Table 1. (d) Locations with samples positive for mixed infection of SPCSV, SPFMV, and SPLCV by molecular diagnostic methods. Within the areas sampled, mixed infections were recorded. The smallest, single-colour circles represent one specific virus combination identified at one farm; the smallest, segmented circles represent detection of more than one virus combination at one farm; larger single-colour circles indicate one specific virus combination found at more than one farm; larger segmented circles indicate the detection of more than one virus combination at more than one farm. Mixed infections detected in each location are shown in

| SPLCV
The 600 dried young leaves collected from the field were processed for DNA extraction. Leaf samples (30 mg) were ground using the TissueLyser II (Qiagen) and DNA was extracted using the CTAB method (Doyle and Doyle, 1987). Degenerate primers SPG1 5′-CCCCKGTGCGWRAATCCAT-3′ and SPG2 5′-ATCCVAAYWTYCAGGGAGCTAA-3′ developed by Li et al. (2004) were used to amplify the CP gene of sweepoviruses that infect

| Sequencing and sequence analysis
PCR products from RT-PCR and PCR were sequenced by the Sanger method by Genewiz Company. Contigs obtained were cleaned and assembled de novo using Geneious v. 8.1.7 (Biomatters Ltd). All the sequences were subjected to the BLAST search tools in NCBI using Geneious and subsequently to pairwise sequence comparison (Altschup et al., 1990;Bao et al., 2014). The homologous sequences were retrieved for phylogenetic analysis. Consensus sequences were MAFFT-aligned using T-coffee v. 11.00.8 tools (Chang et al., 2014).
Using ClustalW in MEGA v. 7.0.14, the sequences were aligned with sequences from other parts of the world retrieved from GenBank (Kumar et al., 2016). Evolutionary history was inferred using maximum likelihood with the Tamura-Nei model (Kumar et al., 2016). Among a number of models, the Tamura-Nei model provided the best nucleotide substitution fit for our data. Phylogenetic reconstruction was performed with bootstrap support values of 1,000.
The trees were visualized and edited using FigTree v. 1.4.3.

| Virus transmission by grafting
Sweetpotato samples to be used in grafting to I. setosa were selected on the basis of the NCM-ELISA, PCR, and RT-PCR results.

| Symptoms observed in fields
The most frequently observed viral symptoms on field plants were stunting, leaf curling, vein-clearing, and leaf distortion. Of these, the most severe were stunting and leaf curling. Chlorotic spots and purpling were also observed on older leaves. Leaf symptoms among the collected samples were generally mild regardless of geographical

| Detection of viruses
The NCM-ELISA revealed the presence of only two virus species: SPFMV and SPCSV. Of the 600 sweetpotato field samples tested, Using RT-PCR, we found that 48% gave negative results for the two viruses identified through NCM-ELISA, 38% were SPFMV-positive, 14% were SPCSV-positive ( Figure 1b, Table 1), and 7% showed SPFMV + SPCSV coinfection ( Figure 1c, Table 1).  where sweetpotato production is high.

| CP gene analysis
Using de novo assembly in Geneious, 15 partial sequences were obtained from RT-PCR and PCR products. From these partial sequences, 13 CP gene sequences were obtained for SPFMV, one for SPCSV and one for SPLCV (  96% nucleotide identity with SPFMV isolates from East Timor and Japan (MF572056 and AB465608, respectively and HQ393473 (Brazil) (Figure 4).
The CP gene from the partially sequenced SPCSV isolate from Burkina Faso was compared with eight CP genes in SPCSV isolates from around the world. Our phylogenetic analysis showed that isolate BFA1279 (sequence LT993430) from Burkina Faso belonged to strain WA together with Spanish sequences KU511274 and FJ807785 ( Figure 5).

| Grafting on I. setosa
The only samples selected for grafting to I. setosa were those that showed single as well as coinfections as detected through NCM-ELISA, RT-PCR, and PCR.
Grafting allowed us to successfully transmit all three viruses identified by molecular analysis to I. setosa (Table 3) (Figure 6a). The other two samples (BFA1284 from Boura and BFA1473 from Matourkou) showed SPFMV + SPCSV coinfection.
For those sweetpotato samples initially found by PCR and RT-PCR to be singly infected by SPCSV (Figure 6c (Table 3).

F I G U R E 4
Maximum-likelihood phylogenetic trees obtained from alignment of nucleotide sequences of coat protein (CP) genes from sweet potato leaf curl virus (SPLCV). The comparisons made were between 22 CP gene sequences obtained from GenBank and one new CP gene sequence obtained in this study. The tree was created in MEGA v. 7.0.14 using ClustalW with 1,000 replicates. Bootstrap values are percentages with values shown at the nodes. The tree was rooted with tomato curly stunt virus (ToCSV) (AF261885). Sequence labels: from this study (red), from GenBank ( Some samples that tested negative using NCM-ELISA were found to be positive for SPFMV and SPCSV when using RT-PCR, thus highlighting the importance of using multiple assays for detecting viruses. Because of the known synergy between SPCSV and some DNA viruses (Cuellar et al., 2015), we also checked for the presence of DNA viruses belonging to the genus Begomovirus using molecular tools.
To determine which strains of the detected viruses are present in Burkina Faso, we sequenced PCR products for SPFMV, SPCSV, and SPLCV. These results demonstrated that SPFMV A-II and B phylogroups, SPCSV-WA, and SPLCV-US were all present in sweetpotato fields in Burkina Faso. Samples BFA340 from Bagré and BFA383 from Di were negative for viruses in all our diagnostic tests; however, after grafting to I. setosa, these two plants showed positive results for at least one of SPFMV, SPCSV, and SPLCV. Similar results were obtained by Abad and Moyer (1992) and Valverde et al. (2007) who proposed that the phenolics and latex found in some sweetpotato cultivars might compromise nucleic acid extraction.
The most widespread virus on sweetpotato crops around the world is SPFMV (Ateka et al., 2004;Njeru et al., 2008;Maina et al., 2018), and it was also the most prevalent virus detected (24%) in our Burkina Faso samples. Leaf vein-clearing is not specific to SPFMV (Moyer and Salazar, 1989) but can indicate the presence of other viral infections. Because mosaic and streak-like symptoms were also prevalent in our sweetpotato samples from fields, it is likely that other viruses in the family Potyviridae were also present (Moyer and Salazar, 1989;Brunt et al., 1996).

F I G U R E 5
Maximum-likelihood phylogenetic trees obtained from alignment of nucleotide sequences of coat protein (CP) genes from sweet potato chlorotic stunt virus (SPCSV). The comparisons made were between 15 CP gene sequences obtained from GenBank and one new CP gene sequence obtained in this study. The tree was created in MEGA v. 7.0.14 using ClustalW with 1,000 replicates. Bootstrap values are percentages with values shown at the nodes. The tree was rooted with lettuce infectious yellows virus (LIYV) (U15440). Sequence labels: from this study (red), from GenBank (

I. batatas I. setosa
The second most prevalent virus that we detected was SPLCV (18%). As a single infection, SPCSV was also present among samples collected but was not very common in the areas surveyed.
However, we did observe instances of SPFMV + SPCSV coinfection, a cause of SPVD. This viral disease has been reported as having the greatest impact on sweetpotato yield worldwide (Valverde et al., 2007;Rey et al., 2012;Loebenstein, 2015). The low prevalence of SPCSV (2%) in our survey areas might mask its true impact if infections are symptomless and thus may also represent a potential threat to sweetpotato production in Burkina Faso. Molecular studies have shown that coinfection of SPCSV enhances SPFMV RNA viral titres by at least 600-fold, whereas SPCSV titres remain equal or are reduced compared to single infection Mukasa et al., 2006;Untiveros et al., 2008). The presence of purple pigmentation, strapping, stunting, and puckering symptoms often result from SPFMV + SPCSV coinfection (Ndunguru and Kapinga, 2007). Some of their findings were applicable to our work: our samples from stunted sweetpotato plants were found to be coinfected by SPFMV and SPCSV in Kenedougou, Houet, Nahouri, and Sissili provinces. Our mixed infections (16%) involving SPFMV + SPCSV + SPLCV, SPFMV + SPCSV, SPFMV + SPLCV, and SPCSV + SPLCV reflect those reported in East Africa (Kenya, Rwanda, Tanzania, and Uganda), the USA, and Korea (Ateka et al., 2004;Clark and Hoy, 2006;Mukasa et al., 2006;Njeru et al., 2008;Cuellar et al., 2015;Kim et al., 2017). Few reports exist on the synergistic interactions between RNA and DNA viruses. However, previous studies, such as Cuellar et al. (2015), showed that SPCSV can interact synergistically with sweepoviruses.
Grafting sweetpotato scions onto I. setosa, and subsequent use of diagnostic tests, confirmed the presence of the three virus species and of coinfections; the I. setosa also displayed the symptoms associated with each virus (Table 3) The majority of symptoms observed in this study are in accordance with the typical SPFMV symptoms described by Moyer and Salazar (1989) and Gibson et al. (1997). Previous research confirmed that SPFMV on its own causes mild or no symptoms in some sweetpotato cultivars (Gibson et al., 1997;Ateka et al., 2004), suggesting that the SPFMV-positive, symptomless samples in our study may have been singly infected with SPFMV. The most severe symptoms were observed for coinfections, mainly involving SPFMV + SPLCV but sometimes SPCSV + SPLCV. These results corroborate those of previous studies using grafting on I. setosa (Maina et al., 2018). In addition, our study showed that   (Moyer and Salazar, 1989;Gibson et al., 1997).
Although symptom scoring, serology, and grafting are useful for the detection of SPFMV, SPCSV, and SPLCV, they are not suitable for virus strain demarcation (Ryu and Choi, 2002;Prasanth and Hegde, 2008;Qin et al., 2014). Previous studies showed that SPFMV phylogroup B and phylogroup A are unrestricted geographically Untiveros et al., 2008;Tugume et al., 2010). Phylogroup B is also known to infect sweetpotato in Asia (Japan, Korea, and China) and North Africa (Egypt) (Tairo et al., 2005). To avoid geographical and biological confusion, Jones and Kehoe (2016) proposed the current approach in which phylogroup nomenclature is based on letters of the alphabet and Latinized numbers.
Our study shows that all SPFMV isolates from Burkina Faso can be assigned to phylogroups A-II and B. Phylogroup A-I was not detected among our samples, even though some sweetpotato cultivars have been introduced from East Africa (Somé et al., 2015). Another finding from our examination of CP amino acid sequences of all isolates from Burkina Faso is the presence of the DAG motif that, according to other research (Shukla et al., 1994;Ateka et al., 2017), is involved in virus transmission by aphids. The presence of the DAG motif in the CP, together with the presence of aphids in the environments sampled, suggests that these insects may play a role in SPFMV transmission F I G U R E 6 Virus-associated symptoms in Ipomoea setosa plants infected by grafting with I. batatas. (a) Vein-clearing like that seen in SPFMV infection; (b) upward leaf curling due to SPLVC; (c) leaf chlorosis due to SPCSV; (d, e) leaf curling, interveinal chlorosis, and blistering due to SPFMV + SPCSV + SPLCV coinfection; (f) symptomless leaf; (g) downward leaf curling and vein-clearing due to SPFMV + SPLCV coinfection; (h, i) downward leaf curling, chlorosis, and light vein-clearing due to SPCSV + SPLCV; (j-l) chlorosis, vein-clearing, severe leaf distortion, blistering, and stunting due to SPFMV + SPCSV ( on sweetpotato in Burkina Faso. We did not investigate viral transmission by aphids, but suggest that this should be considered in future surveys. Of the samples in this study, 60% were infected with viruses that lead to SPVD, either as single infections or as coinfections, on cultivated sweetpotato in Burkina Faso. These viruses do not seem to be widespread in the surveyed areas; however, their presencetogether with that of other viral diseases-could compromise the health of sweetpotato crops in Burkina Faso. Our work confirms that of Ryu and Choi (2002) on the importance of using multiple methods for virus screening, such as NCM-ELISA, RT-PCR, and PCR for the detection of SPFMV, SPCSV, and SPLCV.
In view of our results, we conclude that management approaches for SPVD should continue to include monitoring sweetpotato crops in Burkina Faso and should be performed more widely in West Africa.
Furthermore, we advocate the setting up of management plans to address the development of resistant varieties and to disseminate information about good farming practices to better control spread of viral diseases. The next step in this project is to obtain the full genomes of SPFMV, SPLCV, and SPCSV. Then we will investigate the mechanisms of mixed infection and test for evidence of synergy to better understand the interactions between DNA and RNA viruses and between viruses and plants. Today programme for support with writing, editing, and the design of the maps.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.