New insights into phylogenetic diversity of maize yellow mosaic virus revealed by viral sequences from Togo

Highlights

• •First MaYMV genomes from Togo reveal a new sub-lineage in Africa.
• •First intercontinental recombination detected in MaYMV genomes.
• •Lineage-specific residues in P5 may impact aphid transmission.

Abstract

Maize yellow mosaic virus (MaYMV) (Polerovirus MAYMV, Solemoviridae) is an emerging virus with a global distribution in maize and other cereals. Transmitted by aphids in a circulative, non-propagative manner, MaYMV forms icosahedral particles and possesses a positive-sense, single-stranded RNA genome. Three distinct phylogenetic groups of MaYMV have been described in Asia, East Africa and America. In this study, 19 genomes from Togo and Burkina Faso were sequenced using Virion-Associated Nucleic Acid-Based Metagenomics (VANA) combined with classical RT-PCR to constitute the first dataset of West African MaYMV genomic sequences. Phylogenetic analyses of global dataset of 92 genomic sequences revealed two sub-lineages within the African strain. We also identified 11 recombinant MaYMV genomes involving intra- and inter-continental exchanges. A comparative analysis of non-recombinant sequences was conducted and identified 19 molecular signatures in proteins P0, P1, P4 and P5. Notably, residue K/R 253 in P5 was lineage-specific and is located near conserved polero- and luteovirus residues known to be involved in aphid interactions, suggesting a potential role in viral transmission.

1. Introduction

Maize yellow mosaic virus (MaYMV) (Polerovirus MAYMV, Solemoviridae), is an emergent polerovirus first identified in sweet corn in China in 2013 (Chen et al., 2016; Wang et al. 2016). Shortly afterwards and since then, it has been reported in multiple countries across Asia, including India and South Korea (Lim, Yoon, Jang, Bae, Lee, et al. 2018; Nithya et al. 2023) ; Africa, including Burkina Faso, Ethiopia, Kenya, Nigeria, Rwanda, South Africa, Tanzania and Uganda (Palanga et al. 2017; Guadie et al. 2018; Massawe et al. 2018; Wamaitha et al. 2018; Read, Featherston, et al. 2019; Yahaya et al. 2019; Asiimwe et al. 2020; Welgemoed et al. 2020) and America, including Brazil, Ecuador and Mexico (Bernreiter et al. 2017; Gonçalves et al. 2017; Lappe et al. 2022). Evidence suggests that MaYMV likely had a wider geographic distribution (Suppl. Fig. 1) long before its first official detection in China (Chen et al. 2016; Wamaitha et al. 2018; Read, Featherston, et al. 2019; Welgemoed et al. 2020). Beyond maize, the virus naturally infects millet (Panicum miliaceum) (Lim, Yoon, Jang, Bae, Kim, et al. 2018), sorghum (Sorghum bicolor) (Lim, Yoon, Jang, Bae, Kim, et al. 2018; Wamaitha et al. 2018), sugarcane (Saccharum spp.) (Yahaya et al. 2017; Sun et al. 2019) and, to a lesser extent, wheat (Triticum aestivum) (Guo, Yuan, et al. 2022) as well as several weedy grass species (Yahaya et al. 2017; Lim, Yoon, Jang, Bae, Kim, et al. 2018; Lappe et al. 2022; Zhang et al. 2023) revealing a broad host range and high adaptability on monocotyledons. Under controlled conditions, MaYMV can also infect other cereals such as barley, foxtail millet, oat and rye though without inducing symptoms (Ohlson et al. 2024).

Initially described as causing mosaic symptoms (Chen et al. 2016; Palanga et al. 2017), MaYMV infection has since been associated with leaf tip reddening and stunting through experimental inoculations (Stewart et al. 2020; Mlotshwa et al. 2021). The frequent occurrence of maize leaf mosaics in MaYMV-infected plants has been linked to high coinfection rate with sugarcane mosaic virus (SCMV, Potyvirus sacchari, Potyviridae) (Adams et al. 2017; Wamaitha et al. 2018; Asiimwe et al. 2020). Restricted to the phloem, MaYMV is exclusively transmitted by aphids, mainly by the corn leaf aphid (Rhopalosiphum maidis), and to a lesser extent by the bird cherry-oat aphid (R. padi) and certain biotypes of greenbug (Schizaphis graminum) (Gonçalves et al. 2020; Stewart et al. 2020; Ohlson et al. 2024). Like other poleroviruses, MaYMV follows a circulative, non-propagative transmission mode (Gray et al. 2014).

Poleroviruses are characterized by a non-segmented, single-stranded positive-sense RNA genome of 5.6–6.2 kb, comprising short 5′ and 3′ untranslated regions (UTRs) and five partially overlapping open reading frames (ORFs) (Fig. 2A) (Delfosse et al. 2021; Sõmera et al. 2021). While some polerovirus species contain two additional ORFs (ORF6, 187 bp and ORF7, 379 bp) located at the 3′ end of the genome (Hwang et al. 2013), this has not been observed in MaYMV. Genome-wide analyses of 23 poleroviruses have revealed high levels of nucleotide substitutions (Latourrette et al. 2021). A detailed analysis of six poleroviruses showed that ORFs in the central region of the genome (ORF2–4) were less variable than those at the extremities (ORF0 and 5) (Latourrette et al. 2021). These ORFs encoding the RNA silencing suppressor protein P0 (ORF0) and the readthrough domain (RTD, ORF5) also contained the highest frequency of sites under positive selection. Another major source of genetic diversity in polerovirus arises from recombination events, which are particularly frequent in the intergenic region between the RNA-dependent RNA polymerase (rdrp) and the coat protein (cp) genes (Pagán and Holmes 2010; Latourrette et al. 2021; Sõmera et al. 2021). This untranslated region plays a crucial role in poleroviruses both due to its location at the 5′ end of subgenomic RNA1 and its specific folding, which blocks degradation by exonucleases (Steckelberg et al. 2020). To date, no significant recombination events have been detected in MaYMV (Sun et al. 2021). Phylogenetic reconstructions have identified distinct genetic lineages or strains (Lin et al. 2014; Kwak et al. 2018; Congdon et al. 2021; Filardo et al. 2021; Sharman et al. 2021). Characterizing these lineages is the first step towards understanding their biological properties and identifying associated genetic determinants. Like for other poleroviruses, MaYMV exhibits geographic structuring, with distinct lineages in Asia, Africa and America (Welgemoed et al. 2020; Sun et al. 2021). In Africa, most of the surveys have focused on Eastern countries in relation to maize lethal necrosis (MLN) outbreaks. Currently, only one complete sequence from West Africa (Nigeria) and seven partial sequences from Burkina Faso are publicly available (Palanga et al. 2017; Yahaya et al. 2017). Partial sequences were excluded from study, as their inclusion could reduce the resolution of phylogenetic reconstructions and hinder the detection of recombination events.

Fig. 2.

(A) Genome organization of MaYMV. White boxes represent the different ORFs. The black triangle indicates the predicted −1 slippery ribosomal frameshifting site between ORF1(P1) and ORF2(P2), asterisk indicates the leaky stop codon in the ORF3–5 readthrough domain (CP-RTD), arrows and black lines highlight the two resulting polyproteins. P0(ORF0), silencing suppressor protein; P1, Protein 1 containing putative chymotrypsin-like serine protease (Pro) and viral protein genome-linked (VPg); P2, RNA dependent RNA polymerase; P3a(ORF3a), long-distance movement proteins; P3(ORF3), capsid protein; P4(ORF4), movement protein; P5, read-through domain. (B) Recombination event of MA154TG isolate (PX998109) from Togo between the isolate KY940544 from Brazil (major parent) and the isolate KY684356 from Nigeria (minor parent). The confidence intervals were 3216–4121 and 5498–88 respectively for the beginning and ending breakpoints. (C) Nucleotide diversity index (π) along the MaYMV genome calculated using DNAsp for the total dataset (N = 92), and (D) for the African and Asian sequences (N = 64 and 24, orange and blue, respectively). The diversity index, average number of nucleotide differences per site, was calculated along the genome by using a sliding window of 50 nt moved by steps of 10 nt after alignment of the 92 sequences. The value is assigned to the nucleotide at the midpoint of the window.

In this study, we report the first detection of MaYMV in Togo. A total of 19 genomes containing complete coding regions were sequenced, including sixteen from Togo and three from Burkina Faso. Phylogenetic and recombination analyses uncovered previously unknown sub-lineages in West Africa and Asia and identified recombinant isolates between lineages and continents. These findings shed light on the evolutionary and epidemiological dynamics of MaYMV on a global scale. Additionally, molecular signatures specific to the major MaYMV lineages in Asia and Africa were identified and their putative functions were investigated.

2. Materials and methods

2.1. Sampling

During the 2020 rainy season (from june to october) in Togo, 257 leaf samples exhibiting viral symptoms were collected from maize plants at the flowering and young ear stage in farmers’ fields across the five administrative regions (Fig. 1B) and dried over calcium chloride (CaCl₂) using Bos’s method (Bos 1977). Additionally, two maize leaves infected with MaYMV isolates BF9 and BF16 (Palanga et al. 2017), and a symptomatic maize leaf (BF2019) collected in Burkina Faso in september 2016 and october 2019, respectively (Fig. 1A), and stored at −20 °C were included in the Togolese sample set (Table 1).

Fig. 1.

(A) West African countries where MaYMV was reported previously (brown, BF, Burkina Faso, and Ni, Nigeria) and in this study (red, Togo). Red dots represented sequenced isolates in this study and brown dot represented the Nigerian isolate. (B) Spatial distribution of MaYMV in Togo (Sav: Savanes, Kar: Kara, Ctr, Central: Pla: Plateaux and Mar: Maritime).

Table 1.

MaYMV West African sequence dataset.

N°	Isolate	Country	Plant host	Accession number	Reference
1	MA35TG	Togo	Zea mays	PX998096	This study
2	MA41TG	Togo	Zea mays	PX998097	This study
3	MA43TG	Togo	Zea mays	PX998098	This study
4	MA48TG	Togo	Zea mays	PX998099	This study
5	MA51TG	Togo	Zea mays	PX998100	This study
6	MA53TG	Togo	Zea mays	PX998101	This study
7	MA61TG	Togo	Zea mays	PX998102	This study
8	MA103TG	Togo	Zea mays	PX998103	This study
9	MA117TG	Togo	Zea mays	PX998104	This study
10	MA128TG	Togo	Zea mays	PX998105	This study
11	MA130TG	Togo	Zea mays	PX998106	This study
12	MA132TG	Togo	Zea mays	PX998107	This study
13	MA144TG	Togo	Zea mays	PX998108	This study
14	MA154TG	Togo	Zea mays	PX998109	This study
15	MA159TG	Togo	Zea mays	PX998110	This study
16	MA195TG	Togo	Zea mays	PX998111	This study
18	BF9	Burkina Faso	Zea mays	PX998112	This study
19	BF16	Burkina Faso	Zea mays	PX998113	This study
17	BF2019	Burkina Faso	Zea mays	PX998114	This study
20	KDHTS	Nigeria	Rottboellia cochinchinensis	KY684356	(Yahaya et al. 2017)

2.2. Maize yellow mosaic virus identification and sequencing

Samples collected in Togo were processed using the VANA-based metagenomic approach combined with Illumina HiSeq sequencing (2 × 150-bp paired-end reads) (Genewiz, South Plainfield, NJ, USA) to identify viruses, as previously described (François et al. 2018). In summary, 140–270 mg of dried leaf samples were ground in Hanks’ buffered salt solution (HBSS) (1:10) in presence of ceramic beads and then homogenized. Following a serial of low-speed centrifugation (3200 g for 5 min and 8228 g for 3 min), a filtration (0.45 μm) and an ultracentrifugation (148,000 g for 2.5 hrs at 4 °C), viral particles were concentrated from the homogenized leaf extracts. The resulting viral pellets were resuspended overnight at 4 °C in 200 μl of HBSS and digested by adding 15 U of bovine pancreas DNase I (Euromedex) and 1.9 U of bovine pancreas RNase A (Euromedex, France) for 90 min at 37 °C to eliminate unencapsidated nucleic acids. 100 μl aliquot of the resuspended virion were used to extract total nucleic acids using a NucleoSpin 96 Virus Core Kit (Macherey-Nagel, Germany) according to the manufacturer's protocol. Subsequently, cDNAs were constructed using 100 pmol of Dodeca linkers and SuperScript III (5U/µl) (Invitrogen, USA) enzyme, followed by purification using the QiaQuick PCR cleanup kit (Qiagen). Then, using Large (Klenow) Fragment DNA polymerase (Promega) and Dodeca linkers, priming and extension were performed. Finally, PCR amplification was performed using 5 μl of Klenow cDNAs, 2 μM of one of the 96 MID-tag primers (François et al. 2018), 10 μl of HotStarTaq Plus Master Mix Kit (Qiagen) and H₂O qsp 20 µl, and incubated in the following cycling conditions: 95 °C for 5 min, five cycles of 95 °C for 1 min, 50 °C for 1 min, 72 °C for 1.5 min, 35 cycles of 95 °C for 30 sec, 50 °C for 30 sec, 72 °C for 1.5 min +2 sec at each cycle and a final extension step of 72 °C for 10 min. After a visualization in 1% agarose gel, DNAs products were pooled by 96 maize leaf samples, cleaned using the NucleoSpin® Gel and PCR Clean-up (Macherey Nagel) and quantified using Qubit™ dsDNA HS Assay Kit (Invitrogen no Q32854), before sending to Genewiz for Illumina HiSeq sequencing.

Illumina reads obtained were quality-checked, trimmed and assembled (François et al. 2018) prior to BLAST analysis against the NCBI GenBank database. Among the MaYMV positive isolates, those with long contigs (> 5300 bp) were selected for further sequencing to obtain genomes containing complete coding regions, as described below.

Specific primer pairs were designed along the MaYMV genome (Suppl. Table 1 and Suppl. Table 2) based on alignments between VANA-Illumina-based MaYMV contigs and GenBank available sequences. These primers were used to validate and sequence internal and external regions of the MaYMV genomes via overlapping PCRs. Total RNAs were extracted from MaYMV infected samples using GeneJET Plant RNA Purification Kit (Thermo Fisher Scientific, Waltham, MA USA), according to the manufacturer’s instructions. cDNAs were synthesized using the universal MaYMV reverse primer MaYMV_5637R (Suppl. Table 1), which corresponds to the 3′ terminal conserved domain of the virus, and the MMLV-RT enzyme kit (200 U/µl) (Promega, Fitchburg Madison/USA) following the manufacturer’s recommendations. Independent PCR amplifications were performed using appropriate primer pairs (Suppl. Table 1) and the Takara polymerase kit (5 U/µl) (TAKARA BIO INC.) with the following reaction mix: 5 µl of PCR buffer, 4 µl of dNTPs (10 mM), 4 µl of each primer (10 µM), 0.25 µl of polymerase Takara and 2 µl of cDNA, qs 50 µl with RNase-free H₂O. After agarose gel electrophoresis (1%), PCR products were directly sequenced in both directions using GENEWIZ’s Sanger sequencing platform. The presence of MaYMV in the sample BF2019 from Burkina Faso was tested using the same RNA extraction and RT-PCR protocols with the primer pair MaYMV-Contig1768A_3915F/MaYMV-Contig1768A_4194R (Suppl. Table 1). Then, the BF2019 sample and isolates BF9 and BF16 were further sequenced following the RT-PCR protocol described above. Sanger sequences were edited and assembled with Illumina contigs (Togolese isolates) or the initial 1763 nt (Burkina Faso isolates BF9 and BF16) or de novo assembled (Burkina Faso isolate BF2019) using Geneious v5.5.9 (https://www.geneious.com) to reconstruct MaYMV genomes containing complete coding regions (Table 1).

2.3. Genetic diversity, recombination and phylogenetic analyses

The total MaYMV dataset constituted of 92 sequences, including 20 West African sequences (this study, Table 1 and KY684356 from Nigeria (Fig. 1A)), 45 complete genomes and 27 nearly complete genomes retrieved from the NCBI Virus Database (2025–02–07) (Suppl. Table 2). Sequences were aligned using MUSCLE (Edgar 2004) in MEGA 12 (Kumar et al. 2024) with default settings. After manual curation and trimming of non-coding extremities, genetic diversity along the genome was analyzed using MEGA and DNAsp (Rozas et al. 2017). Sequences were realigned using MUSCLE, and diversity indices and the average number of nucleotide differences per site were computed along the alignment by using a 50 nt sliding window with a 10 nt step. The genetic differentiation of MaYMV isolates between Togolese regions and Burkina Faso was estimated using an analysis of molecular variance (AMOVA) as implemented in Arlequin v5.3.1.2 (Excoffier et al. 2005). Recombination analyses were performed using RDP5 (Martin et al. 2020) with the following methods: RDP (R), GENECONV (G), BOOTSCAN (B), MAXIMUM (M), CHI SQUARE (C), CHIMAERA (CH), SISCAN (S) and 3SEQ (3), all with default settings (p < 0.05; multiple comparison correction is “on” with Bonferroni correction). Recombination events were considered valid if detected by at least four methods with a p-value < 0.05, and the results were confirmed twice using the Pairwise Homoplasy Index (PHI) test, implemented in SplitsTree (Huson and Bryant 2006) and RDP5. A Maximum Likelihood phylogenetic tree was inferred using PhyML (Guindon et al. 2010) implemented in MEGA 12 (Kumar et al. 2024) with the best-fitted nucleotide substitution model (GTR+G+I) and 100 bootstrap replicates.

2.4. Protein diversity analyses

Sub-alignments were created for each ORF from the genomic sequence alignment. Amino acids specific to the African and the Asian main phylogenetic lineages (i.e., conserved in one lineage and absent in the other) were identified by comparing protein sequences using MEGA. Each protein of the MA35TG isolate, used as reference, was modeled using ColabFold v1.5.5 (Jumper et al. 2021; Mirdita et al. 2022). Only protein models or domains with plDDT >90% scores were considered. The predicted local distance difference test (plDDT) is a per-residue measure of local confidence, higher scores indicating higher confidence and usually a more accurate prediction (Guo, Perminov, et al. 2022). These models were superimposed onto homologous structures present in the Protein Data Bank (PDB 7ULN) and pairwise root mean square deviations (RMSD) were calculated using PyMol (Schrödinger 2015). Sequence identity between 7ULN and MaYMV RTD was calculated after alignment using Ident and Sim (Stothard, 2000). Lineage-specific SNPs were visualized on the models using PyMol.

3. Results

3.1. Genetic diversity and distribution of maize yellow mosaic virus in Togo

In 2020, 257 maize leaf samples showing viral symptoms were collected across five regions of Togo (Savane, Kara, Centrale, Plateaux, and Maritime) (Fig. 1B). Following sequencing and genome reconstruction, MaYMV contigs were detected for the first time in Togo, present in 62 out of 257 samples (24%) from all five regions (Fig. 1B). Sixteen samples showing long MaYMV contigs (>5300 bp) were selected for further sequencing by RT-PCR. Three MaYMV isolates from Burkina Faso were added, including two partially sequenced isolates (Palanga et al. 2017) (Fig. 1A). In total, 19 genomic sequences from Burkina Faso and Togo were obtained. Sequences were nearly complete, lacking only the non-coding extremities (not considered in this study). The 19 sequences were highly conserved (99% nucleotide identity, NI) and their genetic diversity showed no differentiation based on the sampling location in Togo and Burkina Faso (F_ST=0.0956, p = 1.0000). Including the Nigeria sequence (KY684356), this dataset constituted the first corpus of West African MaYMV genomic sequences (Table 1). Noteworthy, the sequence from Nigeria was more genetically distant from the others (98% NI).

3.2. Worldwide phylogeny update: major lineages of maymv and recombinant isolates

Both the phylogenetic relationships and the presence of recombination events within MaYMV genomes were reanalyzed, incorporating the newly obtained sequences from West Africa. Statistically significant evidence (p < 0.05; multiple comparison correction is “on” with Bonferroni correction) for recombination was detected within 11 MaYMV genomes (Table 2). The most probable recombinant event was detected in the MA154TG genome (PX998109) from Togo. It was detected by the seven methods of RDP with significant p-values (p < 10e-5) (Table 2, Fig. 2B and Suppl. Fig. 2) and confirmed using the Pairwise Homoplasy Index (PHI) test in Splitstree (p < 0.001) (Fig. 3A). As the major and minor putative parents were the isolates KY940544 from Brazil and KY684356 from Nigeria, respectively, this is the first inter-continental recombination event described to date (Suppl. Fig. 2). The recombination breakpoints (RB) were located between the end of the RdRp gene and the end of the rtd gene, i.e. a fragment of ∼2300 nts length (Fig. 2B). All the other recombination events involved minor and major parents sharing either intra-country, intra-continental or at least one undetermined geographical origin (Table 2, Fig. 3A).

Table 2.

Recombination events detected in the MaYMV worldwide dataset.

		Putative parent ID and country		Breakpoint				Detection method (p-value)
N°	Putative Recombinant	Major	Minor	Begin	CI	End	CI	R	G	B	M	C	S	T
1	PX998109 Togo	KY940544 Brazilia	KY684356 Nigeria	3316	3216–4121	5614	5498–88	++++	++++	++++	++++	+++	++++	++++
2	PP461400 Ecuador	Unknown	MW447649 USA	1152	1016–1227	2993	2686–3035	+	+	+	++	++	+++	+
3	MF425862 Kenya	PP228203 Kenya	PP228204 Kenya	2872	2544–2998	253	5454–351	-	+	+	+++	+	+++	++
4	MG664790 Tanzania	Unknown	MG570476 South Africa	3060	692–4302	5578	5442–338	-	+	+	++	+	++	+
5	PP461400 Ecuador	Unknown	KY052793 Ecuador	3031	541–1146	4056	3351–4136	+	+	+	+	+	+	-
6	KY940544 Brazil	PP461400 Ecuador	MW447649 USA	649	5301–809	4110	3395–4239	+	-	-	++	++	+	++
7	MF425874 Kenya	Unknown	MG570476 South Africa	2954	921–4303	5587	5443–339	-	+	-	++	+	+++	+
8	PX998111 Togo	PX998114 Burkina Faso	PX998107 Togo	3220	2933–4285	170	168–265	-	-	+	++	-	+	+
9	PP228205 Kenya	PP228203 Kenya	PP228207 Kenya	2386	1490–3919	5500	5308–293	-	+	-	++	-	++	+
10	MG664791 Tanzania	MF425860 Ethiopia	MF425859 Ethiopia	3344	3023–4193	5562	5190–167	-	-	-	++	+	++	+
11	MF425870 Kenya	PP228208 Kenya	MF425867 Kenya	4141	3361–4371	600	16–667	-	-	-	+	+	+	+

RDP (R), GENECONV (G), BOOTSCAN (B), MAXIMUM (M), CHI SQUARE (C), SISCAN (S) and 3SEQ (3).

CI: confidence intervals; “-” indicates p > 0.05; “+” indicates 0.05> p > 10e-6; “++” indicates 10e-6 > p > 10e-9; “+++” indicates 10e-9 > p > 10e-12 and “++++” indicates p < 10e-12.

Fig. 3.

(A) Phylogenetic network analysis performed with Splitstree. The three major lineages in Asia, Africa and America are circled. The putative recombinant isolates are labelled (the inter-continental recombinant PX998109, isolate MA154TG in red). (B) Maximum likelihood tree reconstructed with MEGA (model GTR+I+G) removing UTR extremities of the 92 MaYMV genomes. Asterisks at the major nodes indicate bootstrap values >75 based on 100 replicates. Sequences from Togo (TG) and Burkina Faso (BF) were clustered but not the sequence from Nigeria (arrow), the red branch locates the recombinant PX998109. All the sequences were collected from maize excepted those from sorghum (diamond), barley (circle), itch grass (arrow), sugarcane (triangle) or wheat (cross). (C) Nucleotidic diversity of the three main MaYMV lineages calculated with MEGA, the number of sequences (N) was indicated. * Without the recombinant (PX998109, MA154TG).

The MaYMV phylogeny was reconstructed using the worldwide dataset (Fig. 3B). A maximum likelihood phylogenetic tree confirmed the structuring of MaYMV diversity into three major lineages corresponding to Africa, Asia and America, as previously described (Sun et al. 2021). Our study revealed that the isolates from Togo and Burkina Faso (West Africa) clustered together, forming a well-supported group (bootstrap 95%) within the African lineage. However, the Nigeria isolate was genetically more distant from these West African isolates than from those originating from East and Southern Africa. Genetic diversity was higher in America (95.8–97.4% NI) than in Africa (98.1–99.1% NI) and Asia (97.6–98.9% NI), although the number of available sequences from the Americas was very limited (Fig. 3C). As previously observed (Sun et al. 2021), MaYMV nucleotide diversity was higher in ORF0, ORF1 and ORF5 compared to ORF2-ORF4 (Fig. 2C and D).

3.3. Identification and 3D localization of lineage-specific residues

To investigate the genetic characteristics of the major MaYMV lineages, we compared the sequences of each gene at the amino acid level and identified residues specific to the African and Asian lineages. The American lineage was excluded from the comparison due to the limited number of available sequences. Nineteen sites distributed across four proteins (P0, P1, P4 and P5) consisted of lineage-specific residues (Table 3). Half of these sites (10/19) were located in the protein P5 (RTD), a quarter (5/19) in P1 and the remaining sites in P0 and P4. No lineage-specific residues were identified in proteins P2, P3a or P3. Two of the 19 lineage-specific residues were strictly conserved in all 64 African and 24 Asian sequences (P5 positions 281 and 595 occupied by methionine and valine in the African lineage or, leucine and alanine in the Asian lineage, respectively). In the other cases, the amino acid conservation was not strict, i.e. a few sequences (<6) of one lineage did not harbor the expected lineage-specific residue. As expected, the recombinant isolate MA154TG showed American lineage-specific residues in P1 (S5 and T196) and African lineage-specific in P5 (K253, K356, M357, E525, E534 and V595).

Table 3.

Proteins, positions and nature of lineage-specific polymorphisms.

protein	position	Africa, n = 64		Asia, n = 24		America, n = 4
P0	6	K	-	N	(1Q/2 K)	K	(1X)
	218	F	-	S	(4F/3 V)	F	-
	261	L	(2F)	F	(2 L)	L	(1F)
P1	5	F	(1S/2Y)	S	-	S	(1X)
	20	C	(3Y)	Y	(1C)	C	(1Y/1X)
	89	S	(2 L)	L	(2S)	S	-
	196	M	(2T)	T	(4 M)	T	-
	216	E	-	G	(2E)	E	(1 G)
P4	132	S	-	L	(1S)	S	-
P5a	253	K	(1R)	R	-	R	(1 K)
	281	M	-	L	-	2M/2L
	356	K	(1R)	R	-	R	(1 K)
	357	M	(1 V)	I	-	V	(1M/1X)
	468	T	-	S	(2T)	T	-
	520	L	-	P	(2 L)	L	(1P)
	525	E	-	D	(2E)	D	-
	534	E	-	D	(2E)	V	-
	555	T	-	A	(1T)	T	-
	595	V	-	A	-	A	-

^aResidues numbering corresponded to P3-P5 polyprotein

Strictly conserved residues of one lineage are bold. Number of exceptions are indicated in brackets. Residues present at these sites in two out of four sequences from America were indicated for information.

To identify the putative functional role of these lineage-specific residues, the three-dimensional (3D) structure of each MaYMV protein was modeled using ColabFold (Jumper et al. 2021; Mirdita et al. 2022), and their 3D position was localized on the models. The predicted 3D model of the protein P0 showed local distance difference test (plDDT) scores below 50%, indicating an unreliable structural model. The 3D model of the polyprotein P1 could not be predicted directly, it had to be separated in several domains due to highly variable plDDT scores along the protein sequence. However, high and reliable plDDT scores (>90%) were obtained for the protease core (amino acids 232–413) and specific α-helices (aa 137–155 and 431–446) of the transmembrane domain (TM) and Viral Protein genome-linked (VPg) domain, respectively. However, none of the lineage-specific residues were located in these modeled domains. The predicted 3D models of RdRp, P3a and CP all showed high plDDT scores (aa 576–1046 in P1-P2, aa 1–44 in P3a and aa 58–197 in P3), but no lineage-specific residues were detected in these proteins. Similarly to P1, the reliable parts of the P4 predicted 3D model (aa 43–97) did not contain the only detected lineage-specific polymorphisms (aa 132).

In contrast, four lineage-specific polymorphisms in the RTD N-terminal (^NRTD) of the protein P5 located within the 3D protein model (aa 235–435), exhibited high and reliable plDDT scores. As expected, ^NRTD was organized in two structural domains, a jelly roll and a cap domain (Fig. 4A). The lineage-specific residue 253 was located in the cap domain. The MaYMV ^NRTD model was superimposed on the 7ULN PDB structure of turnip yellows virus (TuYV) (Polerovirus TUYV) ^NRTD using PyMol (Fig. 4B). Although the amino acid sequence identity was low (<30%), high structural similarity was observed with pairwise RMSD between Cα atoms < 1 Å. Residue 253 was located close to conserved polero- and luteovirus residues known to be involved in aphid interactions (Schiltz et al. 2022), suggesting a role in aphid-mediated transmission (Fig. 4C).

Fig. 4.

(A) Cartoon representation of the MaYMV RTD model with jelly roll domain (pink) and cap domain (purple) labeled. The four lineage-specific residues were indicated as sticks (yellow). (B) Cartoon representation of the MaYMV RTD model (purple) and of the TuYV dimer structure (blue, 7ULN PDB). The four lineage-specific residues were indicated as sticks (yellow), residues 281 and 357 were not visible. Compared to panel A, MaYMV RTD model was rotated 90°. (C) Top view of the TuYV RTD dimer structure (7ULN PDB, chain A marine and chain B light blue) illustrating polero-and luteovirus conserved positions of cap domain involved in aphid interactions (green and pale green, respectively) (Schiltz et al. 2022). The MaYMV lineage-specific residue 253 is indicated in yellow.

4. Discussion

In this study, MaYMV was identified for the first time across all administrative regions in Togo, with the highest frequency in the Savanes region (46% of the collected symptomatic samples) and the lowest in the Plateaux region (2.5%). This pattern might be due to factors such as environmental conditions, host/vector distributions or farming practices. Notably, most of the observed symptoms on the positive samples consisted of mosaics but not reddening.

The three major phylogeographic lineages of MaYMV (African, Asian and American) previously described (Sun et al. 2021) were confirmed in this study. Furthermore, a new well-supported sub-lineage was identified in Africa (Fig. 3B). We showed that isolates from Togo cluster with those from Burkina Faso within this newly identified African sub-lineage. The short geographical distance between these two neighboring countries could facilitate the exchange of the same susceptible maize cultivars, the circulation of the same insect-vectors and the same virus strains. Unexpectedly, our phylogenetic analysis revealed that the sequence from Nigeria is more closely related to East and South African MaYMV isolates than to those from West Africa. This sequence was collected from a different host plant (Rottboellia cochinchinensis) rather than maize. However, MaYMV sequences isolated from sugarcane and maize in China have been reported to be genetically similar (Sun et al. 2021), and our study did not identify any clear phylogenetic structuring by host plant among other available sequences from barley and wheat (Fig. 3B).

We showed that reticulate evolution, mediated by recombination events between genetically distant genomes, may play a significant role in MaYMV evolution. Indeed, we identified eleven recombination events in the global MaYMV dataset. We focused on PX998109 genome from Togo that originated from recombination between African and South American parental viruses. The recombination breakpoint occurred within the intergenic region between P2 and CP, a recombination hotspot in the genus Polerovirus (Pagán and Holmes 2010; Latourrette et al. 2021). The consequences of this recombination event on MaYMV biological traits have not yet been investigated but warrant future study.

To date, biological and pathogenic characterizations of MaYMV isolates from distinct lineages are lacking. Our study identified some of the genetic determinants specific to the African and Asian lineages. The lineage-specific residues revealed by our protein comparative analyses may result from adaptation of MaYMV isolates to their respective geographical region and environment, although a founder effect cannot be ruled out as a contributing factor.. These residues are preferentially located in genes p0 (encoding the RNA silencing suppressor), p1 (TM-VPg-Pro) and p5 (RTD), which are the most diverse genes in poleroviruses (Delfosse et al. 2021; Latourrette et al. 2021).

In this study, we leverage recent advances in deep learning approaches that facilitate protein structural modeling (Jumper et al. 2021), a crucial topic for plant viruses, for which 3D protein structures are rare. However, viral proteins such as P0, VPg and CP-RTD of poleroviruses often contain intrinsically disordered regions that prevent structural modeling (Latourrette et al. 2021). Nevertheless, the core domains of MaYMV Pro, RdRp, P3a, CP and RTD could be reconstructed. Only the RTD domain contained lineage-specific residues in a predicted 3D region. To better evaluate the putative role of these residues, we compared more precisely the MaYMV ^NRTD to the atomic structure of the better characterized TuYV. Although the sequence identity was low, the high structural similarity suggested that the lineage-specific residue 253 may be involved in aphid interactions. This hypothesis requires functional validation in the next future, for instance combining directed mutagenesis of available MaYMV infectious clones (Chen et al. 2016; Mlotshwa et al. 2021) and aphid transmission assays under controlled conditions. The recombinant isolate MA154TG harbors a lysine at position 253 of the RTD and all the other residues specific to the African lineage in this domain. The impact of this recombinant event on the viral traits should be further assessed, particularly to detect a putative fitness benefit of the American lineage to the African environment and/or vector specificity. Overall, by providing genomic sequences from less-studied geographical regions, such as West Africa, our study revealed new insights into the diversity, biology, dispersion and evolution of MaYMV.

Funding

EP’s postDoctoral fellowship was funded by the Ministry for Europe and Foreign Affairs (MEAE), Campus France Program Make Our Planet Great Again (MOPGA), reference: mopga-pdf-0000000244. Sample and Data collection were funded by CGIAR Research Program Grain Legumes and Dryland Cereals (CRP GLDC).

We thank IRD program: JEAI VCLML and South-North mobility (MLD) which supported scientific travels. This work was partially supported by the French National Research Agency (ANR-20-CE35–0008–01).

CRediT authorship contribution statement

Essowè Palanga: Writing – review & editing, Writing – original draft, Visualization, Project administration, Methodology, Funding acquisition, Formal analysis, Conceptualization. Agnès Pinel-Galzi: Writing – review & editing, Resources, Methodology. Denis Filloux: Methodology, Formal analysis, Data curation. Adeline Pinault: Methodology. Kossi Essotina Kpemoua: Project administration, Methodology. Essohouna Ali: Methodology. Charlotte Julian: Resources, Methodology. Damien Richard: Writing – review & editing, Formal analysis. Nils Poulicard: Writing – review & editing, Methodology, Funding acquisition, Formal analysis. Philippe Roumagnac: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Eugénie Hébrard: Writing – review & editing, Visualization, Supervision, Project administration, Funding acquisition, Formal analysis, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Genetic data obtained during this study are publicly available (accession numbers indicated in the text and tables).