Detection and diversity of viruses infecting African yam (Dioscorea rotundata) in a collection and F1 progenies in Côte d'Ivoire shed light to plant‐to‐plant viral transmission

Abstract Yam (Dioscorea spp.) is a major staple food whose production is hampered by viral diseases. However, the prevalence, diversity, transmission, and impact of yam‐infecting viruses remain poorly documented. This study reports on the symptomatology, prevalence, and molecular diversity of eight viruses in 38 D. rotundata accessions from a germplasm collection and 206 F1 hybrid progenies maintained in Côte d'Ivoire. Mean severity scores as assessed from leaf symptoms ranged from 2 to 4 in the germplasm collection and from 1 to 3 in F1 hybrids, respectively. Dioscorea mosaic‐associated virus (DMaV), potexviruses, and yam mosaic virus (YMV) were detected by PCR‐based diagnosis tools in single and mixed infections in both the D. rotundata collection and F1 progenies, whereas badnaviruses were detected only in the germplasm collection. In contrast, cucumber mosaic virus (CMV), yam macluraviruses, yam asymptomatic virus 1 (YaV1), and yam mild mosaic virus (YMMV) could not be detected. No correlation could be established between severity scores and indexing results. Phylogenetic analysis performed on partial viral sequences amplified from infected samples unveiled the presence of two putative novel viral species belonging to genera Badnavirus and Potexvirus and provided evidence for plant‐to‐plant transmission of YMV, DMaV, and yam potexviruses.

country (Touré et al., 2003). However, the production is insufficient, especially because yam yields have decreased from 8 to 5.5 t/ha in Côte d'Ivoire since 2007 (FAOSTAT, 2020) despite an increase in cultivated surface. This decrease is likely to result from the impact of viral diseases, the use of infected and/or too old planting material, soil fertility, and postharvest losses (Bakayoko et al., 2017).
Yams are primarily propagated vegetatively, leading to an accumulation of viruses resulting in multiple infections (Eni et al., 2008). Although the impact of viral diseases on yam production is poorly documented, decreases in yield and reduction of the quality of harvested tubers due to viral infections have been reported by Toualy et al. (2014), sometimes threatening entire produc- genus Badnavirus) (Odu et al., 2004). Using mainly viral metagenomics, a large number of novel viruses have been identified in yam during the last 15 years. Thus far, viruses belonging to genera Ampelovirus, Aureusvirus, Badnavirus, Carlavirus, Cucumovirus, Fabavirus, Macluravirus, Potexvirus, Potyvirus, Sadwavirus, and an unknown genus belonging to the family Betaflexiviridae, have been reported in yams (Umber et al., 2020).
In West Africa, five distinct viruses are usually detected in yam, including YMV, yam mild mosaic virus (YMMV; genus Potyvirus), CMV, DBV, and Dioscorea latent virus (DLV; genus Potexvirus) (Eni et al., 2008;Toualy et al., 2014). In Côte d'Ivoire, yam growers report increasing yield losses and observe virus-like symptoms such as chlorosis, mosaic, deformation on leaves, and dwarfism (Séka et al., 2009;Toualy et al., 2014). As a consequence, farmers increase cultivated areas in order to compensate for yield losses and maintain production levels, resulting in conflicts over land use and to negative impacts on the environment.
The control of viral diseases affecting crops relies primarily on the use of clean seeds and/or the breeding of resistant varieties either by conventional methods or by genetic engineering. Yam sanitation programmes have been successfully implemented, resulting in the production of clean seeds (Umber et al., 2020). Such programmes rely on plant regeneration from meristem culture in vitro, thermotherapy, or cryotherapy (Filloux & Girard, 2006). The National Center for Agricultural Research (CNRA) works on the sustainability of yam production in Côte d'Ivoire by the selection and propagation of improved yam varieties that meet the expectations of the yam sector.
These new hybrids have to combine high yield, good tuber quality, and especially resistance to yam mosaic disease, considering the high susceptibility of D. rotundata to viruses. Viral resistance is therefore an important trait for yam selection. The viral status of D. rotundata F 1 progenies obtained from hybridizations by CNRA in Bouaké and of a diverse panel of D. rotundata accessions has been assessed using the comprehensive diagnostic scheme designed by the Biological Resources Center for Tropical Plants (BRC-TP) located in Guadeloupe (French West Indies) (Umber et al., 2020). Molecular tools used for diagnosis include reverse transcription (RT)-PCR for the detection of RNA viruses and immunocapture (IC)-PCR for that of badnaviruses, which avoids false positives resulting from the presence of endogenous badnavirus sequences (eDBVs) in the genome of some yam species, including D. rotundata (Umber et al., 2014). RNA viruses targeted in this study include CMV, Dioscorea mosaic-associated virus (DMaV; genus Sadwavirus), yam asymptomatic virus 1 (YaV1; genus Ampelovirus), YMV, and YMMV with specific tests. Yam potexviruses and yam macluraviruses are detected using generic tests, which are able to target several yam-infecting viral species (Umber et al., 2020).
This study reports on the evaluation of the symptomatology, prevalence, and diversity of these viruses in a field germplasm collection of D. rotundata accessions and in the D. rotundata F 1 progenies maintained in Côte d'Ivoire. The results suggest that the presence of yam-infecting viruses in D. rotundata is not correlated with observed leaf symptoms and that plant-to-plant transmission occurs for at least some of the viruses. This work also reports on the first molecular detection of DMaV and potexviruses in yams in Africa using RT-PCR-based tests. according to various leaf symptoms from the CNRA germplasm collection, which was planted in March 2018 with a spacing of 1 m between and within rows. Two hundred and six F 1 hybrids and their genitors were collected from a nearby plot. They originated from progenies of two biparental crosses involving D. rotundata genitors, one female (Cnraigr09/00001) and two males (TDr95/18555 and TDr00/00380).
The first hybrid population of 70 individuals originated from the Cnraigr09/00001 × TDr99/18555 cross while the second population of 136 individuals resulted from the Cnraigr09/00001 × TDr00/00380 cross. This plot was planted in June 2018, 4 m away from the D. rotundata collection, and was organized in three completely randomized Fisher blocks with one replicate of each hybrid and each genitor per block. The sampling was carried out on the plant with the most important symptoms among the three replicates. Details of the samples are provided in Tables S1 and S2.

| Assessment of leaf symptoms
Leaf symptoms were observed 3, 4, and 6 months after the planting of D. rotundata accessions and F 1 hybrids. Symptom severity was assessed according to a rating scale established by the International Institute for Tropical Agriculture (IITA, 1998; Table 1). For the two progenies and their genitors, the score retained corresponds to that of the plant presenting the most important symptoms among the three replicates.

| Molecular detection of yam viruses
Total nucleic acids (TNAs) were extracted from leaf samples according to procedure 2 developed by Foissac et al. (2005) and used for the detection by RT-PCR of CMV, DMaV, YaV1, YMV, YMMV, yam macluraviruses, and yam potexviruses as described by Umber et al. (2020).
Details of the primers are provided in Table S3. Badnaviruses were detected by immunocapture-multiplex-PCR (IC-M-PCR), as described by Umber et al. (2017), using BEL antiserum (Ndowora et al., 1999) and atpB1/B2 primers to control genomic DNA contamination (Soltis et al., 1999). All tests were performed twice in order to confirm the results.

| Correlation between symptom severity and viral detection
Severity scores observed for D. rotundata accessions of the CNRA collection and the two F 1 hybrid progenies were compared to indexing results. The corrplot package of the open source software R was used to calculate Pearson correlation coefficients between severity scores of viral symptoms and the results of viral indexing (R Core Team, 2021).

| Analyses of the molecular diversity of yam viruses
PCR products amplified from D. rotundata accessions of the CNRA germplasm collection infected by potexviruses, badnaviruses, and/or DMaV were cloned into the pGEM-T Easy vector (Promega) according to the manufacturer's instructions and sequenced (Genewiz). Nonredundant sequences were used for multiple alignments using the ClustalW component of MEGA X (Kumar et al., 2018). Phylogenetic trees were constructed using the maximum-likelihood method and the robustness of the trees was determined using the bootstrap method with 1,000 replicates.

| Correlation between severity scores of leaf symptoms and detected viruses
No overall correlation could be established between symptom severity scores and the detection of a particular virus ( Figure 3). For instance, YMV was detected in the F 1 hybrids Cnraigr17/00096 and Cnraigr17/00669, which were symptomless. In contrast, no virus was detected in 99 out of the 183 F 1 hybrids (54.1%) displaying a score of 2, nor in 11 out of the 18 (61.1%) of those displaying a score of 3 (11/18).
Overall, the correlogram revealed moderate correlations between viral infections and severity scores, regardless of the virus considered (Table 3). Indeed, the detection of DMaV was positively correlated with severity scores 3 and 4 (r = 0.3; p < 0.0001) and negatively correlated with score 2 (r = −0.309; p < 0.0001), with high significance. Similarly, the presence of badnaviruses was positively associated with severity score 3 (r = 0.377; p < 0.0001) and negatively correlated with severity score 2 (r = −0.351; p < 0.0001), with high significance.  ire (Eni et al., 2008). Thus, Adjata (1991)  Thus, among the five symptomless F 1 hybrids (severity score = 1), the genotypes Cnraigr17/00096 and Cnraigr17/00669 were infected by YMV. In similar studies, YMV was detected in symptomless F 1 hybrids, and these genotypes have been considered tolerant (Mignouna et al., 2001). Tolerance, or partial resistance, could be the basis for resistance in field conditions and protection of plants over long periods (Lecoq et al., 2004). Several D. rotundata accessions and Betaflexiviridae family, which seems to be prevalent in Africa (Silva et al., 2019). In addition, primers used for detection tests could fail to target some divergent isolates or tests could be not sensitive enough to detect low viral titre, and explain the negative results.

| Molecular diversity of badnaviruses, DMaV, and potexviruses identified in Bouaké
In order to develop a new preventive control strategy, adapted to existing conditions, data on the relationship between the severity of viral infections and the type of detected viruses is essential.
Thus, statistical analyses showed positive but moderate correlations between severity scores 3 and 4 and the presence of badnaviruses and DMaV, meaning that plant infection by these two viruses leads to severe leaf symptoms, although this result could be due to coinfections, regarding the high prevalence of DMaV and YMV. As expected, their detection was negatively and significantly correlated with severity scores 2 and 1. In contrast, the correlation between YMV and symptom severity scores was very weak, because this virus infected all the accessions of the CNRA collection, as well as more than 40% of the F 1 hybrid progenies.
Thus, the highest severity scores (3 and 4) could rather result from a synergy of the coinfections by YMV and badnaviruses or DMaV.
However, even though significant, correlation rates were very weak. As Njukeng et al. (2014) argued, prediction of the presence of a type of virus based on severity scores is therefore impossible on yam crop.
As the genetic diversity of YMV, which is the most widespread virus in yam crop in West Africa, has already been assessed worldwide (Bousalem et al., 2000), only that of DMaV, badnaviruses, and potexviruses sequences was analysed. The molecular analyses of these viruses detected from the CNRA D. rotundata collection showed a different structure of diversity depending on targeted viruses, and revealed new viral species.
Regarding the badnaviruses, their sequences belong to three different phylogenetic groups representing three distinct species.
Most of these sequences (26 out 31) fit to Group 8, which is the most widespread species of Dioscorea bacilliform virus (DBV) worldwide (Kenyon et al., 2008). Dioscorea bacilliform AL virus (DBALV), belonging to the DBV8 species, was the first badnavirus genome to be fully characterized from yam (Briddon et al., 1999).
Whilst it was first detected in D. alata, it is now known to infect other yam species, such as D. bulbifera, D. nummularia, D. rotundata, and D. trifida (Kenyon et al., 2008;Sukal et al., 2020;Umber et al., 2017). Three of the badnaviral sequences obtained in this study clustered with Group 5 (DBV5), of which the complete genome of the corresponding species, named Dioscorea bacilliform RT virus 3 (DBRTV3), has been recently sequenced in Nigeria from D. rotundata (Bömer et al., 2018). Endogenous viral sequences from DBV8 and DBV5 species are known to be integrated into the genome of yam species of the D. cayenensis-rotundata complex (Umber et al., 2014). Some viral insertions are supposed to lead to viral resistance through posttranscriptional gene silencing, which prevents further plant infection with corresponding viral genomes (Chabannes et al., 2013). However, detection of DBALV and DBRTV3 in episomal form in D. rotundata shows the opposite, and confirms that these endogenous sequences are not involved in viral resistance in this yam species (Umber et al., 2017). The two remaining sequences, CivcDr227_4 and CivcDr402_2, share 88% to 93% of identity with six partial badnaviral sequences identified in Africa belonging to Group 15 (DBV15), from D. rotundata and D. alata (Bömer et al., 2016;Eni et al., 2008 DMaV has been recently characterized in Brazil by metagenomics (Hayashi et al., 2017) and belongs to the genus Sadwavirus (Sanfaçon et al., 2020). DMaV isolates identified in this study form a single species with isolates originating from the American continent; thus, we provide evidence here that this viral species is present in both continents.  (Odu et al., 2004), which need to be transported from plant to plant by ants. DMaV belongs to the family Secovirideae, whose members are transmitted by insects or soilborne nematodes (Thompson et al., 2017). For instance, grapevine fanleaf virus (GFLV), a nepovirus, is transmitted by the nematode Xiphinema index (Schellenberger et al., 2011). Nematode-mediated viral transmission could result in high infection in the field if the same crop is planted in the same infested plot over the years. In contrast, crop rotation leads to natural sanitation of the infected soil (Bilevai et al., 2009). The plot supporting the F 1 hybrid progenies was previously covered by a pepper crop and that may explain the weak DMaV contamination of that population, assuming pepper plants would not be the host for the yam-feeding nematodes. Finally, while seed transmission has been reported for potyviruses (Johansen et al., 1994), this mode of transmission has never been demonstrated for YMV. However, because the female parental accession Cnraigr09/00001 was infected by YMV (Table S2) Pressat (CIRAD, Guadeloupe) for technical assistance, Pierre-Yves Teycheney (CIRAD, Guadeloupe) for his very great assistance in the writing of the manuscript and Sébastien Guyader for English editing.

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