From 2016 to 2018 and during the August to October growing seasons, 135 symptomatic Bambara groundnut leaf samples were collected from farmers’ fields across the three agroclimatic zones of Burkina Faso. Agroclimatic zones included the Sudan zone with annual rainfall of 900 - 1100 mm, the Sudan-Sahel zone (600 - 900 mm) and the Sahel zone (less than 600 mm). In order to cover virus diversity, sampling sites were separated by at least 3 km. One hundred samples were collected in the Sudan zone and the Sudan-Sahel zone (50 samples per zone) whereas 35 samples were from the Sahel zone. Leaf Samples were kept on ice before being transferred to the laboratory where they were stored at −80˚C before molecular analyses.

Total RNA from the 135 samples were extracted individually with Trizol (Invitrogen) according to the protocol described by [8]. Before the extraction, the mortars were first sterilized during 2 hours at 180˚C and then kept at −80˚C until use. Briefly, RNA extraction followed was done according to the following steps: first, crushed leaves were transferred into 2 ml Eppendorf tubes and 1 ml of Trizol was added to release cellular content; secondly, 200 µl of chloroform were added. The tubes were kept on ice for 5 min before centrifugation at 14,000 rpm for 15 min at 4˚C. RNA was precipitated by adding 550 μl of cold isopropanol and keeping the tubes at −20˚C for 30 min. Finally, the tubes were then centrifuged at 12,000 rpm for 10 min and the pellets were washed with 75% ethanol. Total RNA was suspended in 30 µl sterile water. RNA quality was checked and concentration measured by a nanodrop before storage at −20˚C.

Detection of potyviruses was carried out in an RT-PCR assay using the universal potyvirus primers P077-F: ATGGTHTGGTGYATHGARAAYGG and P078-R: CARATGAARGCMGCAGCA [9]. Viral cDNAs were obtained from the extracted RNAs using the reverse oligodT primer and the enzyme M-MLV-RT (200 U) (Promega, USA) according to the supplier’s recommendations. Reverse transcription was done in a total volume of 25 µl containing 1 μl of total RNA, 1 μl of 10 mM oligodT, 5 μl of M-MLV RT 5X Buffer, 1 μl of 10 mM dNTP, 200 U M-MLV-RT and sterile water.

PCR amplification was done using the Solis Biodis PCR kit according to the supplier’s instructions. A reaction volume of 25 μl containing 2 μl of cDNA, 2.5 μl of 10x PCR buffer, 0.5 μl of 10 mM dNTP, 2.5 μl of 25 mM MgCl₂, 0.5 μl of each primer P077-F and P078-R (10 μM) and 0.5 U of the enzyme FIREPol DNA polymerase subjected to the following cycling conditions: 95˚C for 5 min followed by 35 cycles of 94˚C for 1 min, 55˚C for 30 sec and 1 min of elongation at 72˚C and a final elongation step performed at 72˚C for 10 min. The amplified fragments were analyzed by electrophoresis on 1% (w/v) agarose gels containing ethidium bromide and visualized under UV to verify the size of the amplified fragments. Expected size fragments (327 bp) were sequenced and the sequences were blasted against the GenBank (NCBI) database.

Samples positive for potyvirus detection were used for amplification of the complete cp gene. Therefore, independent PCRs were performed using primer combinations P105-F/P078-R and P077-F/P106-R (Table 1) as described by [4]. Expected size fragments were sequenced, assembled and blasted against GenBank NCBI. New primers were designed (Table 1) based on the sequences obtained and taking into account closely related potyvirus sequences in GenBank database. Independent PCR amplifications were performed as described above using the newly designed primers and conditions indicated in Table 1. Amplicons of the expected sizes were directly sequenced using respective primer pairs.

All sequences contigs were edited and assembled using DNAMAN software to

generate complete CP sequences. The cp sequences were compared to GenBank database sequences by Blastn NCBI (BLAST, http://www.ncbi.nlm.nih.gov/blast). Nucleotide and protein sequence identities between among query and retrieved sequences were determined using SDT v1 software [10]. Relative synonymous codon usage (RSCU) was determined using Seqinr [11].

Phylogenetic trees were reconstructed from complete CP sequences using the MEGA 6 software [12]. After Clustal alignment, the Maximum Likelihood (ML) method was applied for the phylogenetic trees reconstruction using the TN93 + G + I and Protein JTT + G + I nucleotide substitution models, with a bootstrap of 1000 replications. The amino acid composition and the nucleotide diversity between the different groups of potyviruses identified by this study were determined with the Mega 6 software.

As illustrated in Figure 1, amplicons of expected size (327 bp) were obtained in RT-PCR tests for potyvirus detection in Bambara groundnut using potyvirus universal primers. The universal potyvirus primers [9] associated to the blast NCBI [13] analyses were effective in detecting potyviruses in 36 isolates out of 135 samples analyzed, indicating a virus prevalence of about 26.7%. Twenty of these isolates (20/50 samples) originated from Sudan zone (40% in prevalence), sixteen (16/50 samples) from the Sudan-Sahel zone (30% in prevalence), while no positive samples were detected in the Sahel zone. This distribution makes the Sudan zone the most infested by the Bambara groundnut potyviruses in the country.

Complete CP sequences were obtained from 24 isolates out of 36 positive samples for potyvirus detection. Full CP sequence could not be obtained from the remaining isolates due the insufficient quality of corresponding sequencing data. The 24 full CP sequences were submitted to the Genbank database (Table 2).

Nucleotide and amino acid analysis of the coat protein sequences showed a high virus diversity. As shown in Table 3, three groups of potyvirus sequences designated group 1 (7 isolates), group 2 (4 isolates) and group 3 (13 isolates) were differentiated. Within group nucleotide identity was 94.6% - 100%, 93.7% - 100% and 94.2% - 100% in group 1, group 2 and group 3, respectively. Higher amino acid identity appeared in each group, reaching 97.8% - 100% in group 1, 97.7% - 100% (group 2) and 97.8% - 100% (group 3). Percentages of nucleotide and amino acid identity were significantly lower between groups. At nucleotide level, isolates in group 1 shared roughly 72% - 74% identity with isolates of group 2 or isolates of group 3. Nucleotide identity between group 2 and group 3 ranged between 75.8% and 77.7%. Further between group comparisons indicated that both nucleotide and at amino acid identities showed similar trends.

Isolates in the current study were also compared to the closest virus isolates from the Genbank database. As presented in Table 3 isolates in group 1 shared high sequence identity only with CABMV isolates at both nucleotide (94.6% - 100%) and amino acid (97.8% - 100%) levels. Percentages of nucleotide or amino acid identity with any closest isolate were less than 77%. Isolates in group 2 showed only 70.6% to 77.7% nt identity and less than 80% aa identity with any closest virus isolate outside the group. The highest nucleotide identity (76.9% -

*CABMV: Cowpea aphid-borne mosaic virus; BCMNV: Bean common mosaic necrosis virus; BCMV: Bean common mosaic virus; UPV: Ugandan passiflora virus; PaChV: Passiflora chlorosis virus. Group 1 (MK987189; MK987190; MK987191; MK987192; MK987193; MK987194; MK987195); Group 2 (MK987196; MK987197; MK987198; MK987199); Group 3 (MK987200; MK987201; MK987202; MK987203; MK987204; MK987205; MK987206; MK987207; MK987208; MK987209; MK987210; MK987211; MK987212).

77.1%) and amino acid identity (78.4% - 79.9%) were found with Bean common mosaic necrosis virus (BCMNV). Moreover, group 2 isolates had 780 bp nucleotide sequence similar to BCMNV (786 bp). Isolates of the two other groups had 828 bp, which made a difference of 48 pb. In group 3 isolates, the highest nucleotide identity (77.3% - 78.3%) and amino acid identity (80.7% - 81.5%) were found with Ugandan Passiflora virus [J896003].

The twenty amino acids were found in all three groups but CP sequences in Group 2 isolates consisted of 259 aa which was 16 less amino acids than in the other two groups (275 aa). All CP amino acid sequences displayed the N-terminal DAG motif which is conserved in most potyviruses.

Table 4 summarizes the codon utilization bias in the three potyvirus groups. Codon utilization bias among the potyvirus groups was apparent in all degenerate amino acids. The three groups of potyvirus shared the same preferred codons in some amino acids. However, in some cases, additional preferred codons discriminated the three groups of potyviruses. For example, in leucine the three groups have different codon preferences, group 1 prefers CUU, UUG for group 2 and CUG for group 3. However, in most amino acids group 2 and 3 have the

aRSCU: relative synonymous codon usage in Bambara groundnut potyvirus group 1 (G1), group2 (G2) and group 3 (G3). The most preferred codons are in boldface.

same preferences. Group 1 of potyvirus also differs from group 2 and group 3, preferably using GGG codons for glycine and UCU for Serine. Another distinction between the potyvirus groups was evident in the preferred codon usage for asparagine (AAC in group 1 versus AAU in groups 2 and 3) and by histidine (CAC in groups 1 and 2 with respect to CAU in group 3). The clearest divergences between the potyvirus groups were observed in groups that did not use specific codons for certain amino acids. This was exemplified by codon UGU (Cys) in group 1, codons UGC (Cys) and CGC (Arg) in group 2 and codons CGA (Arg) and UCC (Ser) in group 3.

Phylogenetic analysis performed on the nucleotide and protein sequences of the 24 isolates along with homologous sequences retrieved from GenBank clearly distinguished clades that corresponded to the three groups of potyviruses described previously (Figure 2). In both nucleotide and protein based phylogenetic

trees, group 1 isolates clustered with CABMV isolates only. Notably, this CABMV group included also two isolates (MF277031 and MF277034) recently characterized from Bambara groundnut in Burkina Faso [4]. Group 2 and group 3 formed two new clusters which did not include any already known virus isolate. Therefore, group 2 and group 3 were referred to as Bambara groundnut potyvirus 1 (BGPV1) and Bambara groundnut potyvirus 2 (BGPV2), respectively.

Figure 3 presents the geographical distribution of the three groups of potyviruses characterized on Bambara groundnut in this study. BGPV1 isolates were found in the Sudan zone only whereas both CABMV and BGPV2 isolates were found in the Sudan and Sudan-Sahel zones. However, discrepancies were observed between the two agroclimatic zones. CABMV isolates were most frequent in the Sudan Sahel zone (25% of the isolates) than in the Sudan zone (8.3%). By contrast, most BGPV2 isolates (29.2%) were found in the Sudan zone while only 4.2% of the isolates originated from the Sudan Sahel zone. Altogether, all three virus groups were found in the Sudan zone which held 70% of virus isolates.