Genetic Variation and Evolutionary Analysis of Eggplant Mottled Dwarf Virus Isolates from Spain

The genetic variation and population structure of gene N (nucleocapsid) and part of gene L (replicase) from 13 eggplant mottle dwarf virus (EMDV) isolates from Spain were evaluated and compared with sequences of EMDV isolates from other countries retrieved from GenBank. Phylogenetic inference of part of gene L showed three main clades, one containing an EMDV isolate from Australia and the other two containing isolates from Iran and Europe, as well as four subclades. EMDV isolates from Spain were genetically very similar and grouped in a subclade together with one isolate from Germany and one from the UK. No new recombination events were detected in addition to one recombination previously reported, suggesting that recombination is rare for EMDV. The comparison of synonymous and non-synonymous rates showed that negative selection played an important role, and only two codons were under positive selection. Genetic differentiation (Fst test), phylogenetic and nucleotide diversity analyses suggest a unique introduction of EMDV to Spain and low gene flow with other countries. In contrast, Greece and Italy showed diverse populations with high gene flow between both.


Introduction
Understanding the genetic variability of virus populations and the factors involved in their evolution is crucial to developing accurate detection and diagnostic tools, implementing efficient disease control strategies and gaining insight into virus epidemiology [1].The main evolutionary mechanisms shaping the genetic structure and variability of virus populations are mutation, recombination, selection, genetic drift and gene flow or migration [2,3].RNA viruses have great potential for rapid evolution due to their rapid replication and high mutation rates since RNA replicases lack proofreading activity.Recombination is another source of genetic variation and the emergence of new plant viruses [4].The genetic variation produced by mutation and recombination is limited by the interplay of selection, genetic drift and gene flow.Natural selection in plant viruses results from the competition among genetic variants differing in some aspects of the life cycle, such as replication, movement between plant cells and transmission to other plants by vectors, so those variants with more reproductive success (fitness) will pass to the next generation.Genetic drift is the change in the frequency of genetic variants in small populations by random chance.Viruses can undergo population bottlenecks or founder events in different life cycle steps such as cell-to-cell movement and transmission by vectors.Gene flow or migration of genetic variants favors genetic uniformity, whereas restricted migration leads to genetic differentiation between populations via selection and genetic drift.
Eggplant mottled dwarf virus (EMDV) has been assigned to the species Alphanucleorhabdovirus melongenae in the genus Alphanucleorhabdovirus, family Rhabdoviridae, order Mononegavirales [5].The EMDV genome is a single-stranded, negative-sense RNA encapsidated by the nucleocapsid protein (N) and wrapped by a phospholipid membrane forming bacilliform particles.The EMDV genome contains seven open reading frames: N (nucleocapsid), X (unknown function), P (phosphoprotein, a polymerase cofactor), Y (putative movement protein), M (matrix protein, which connects the envelope to the ribonucleocapsid core), G (glycoprotein, which protrudes from the lipid envelope exterior) and L (RNA-dependent RNA polymerase) [5,6].
The genetic variability of EMDV and some evolutionary factors have been studied by sequence analyses of genes N, X, P, Y, M and G from seven isolates from Greece and one from Cyprus; part of gene L of these isolates from Greece and Cyprus, as well as eight from Italy and five from Spain [11,12]; and the complete sequence of one isolate from Iran [13].In this work, the nucleotide sequence of gene N and part of gene L in 13 EMDV isolates from four regions of Spain and six hosts were determined and analyzed together with part of the gene L of another 5 Spanish isolates retrieved from GenBank.This work revealed that the population of EMDV in Spain was very homogeneous and had a low migration rate in contrast to those in Italy and Greece.A more comprehensive picture of the genetic variability and evolutionary mechanisms of EMDV globally was obtained by analyzing the entire genome of nine isolates-three from Italy, two from Greece and one each from Iran, Slovenia, Germany and the UK-as well as the N gene from six isolates from Greece and one from Cyprus; part of the L gene of six isolates from Greece, one from Cyprus, two from France and one from Australia; and a Y gene portion from one isolate from Azerbaijan.

Sequencing EMDV Isolates from Spain
We analyzed the N gene (encoding the nucleocapsid) and a portion of the L gene (encoding the RNA-dependent RNA polymerase) to evaluate the genetic variability in EMDV in Spain and other countries.These genomic regions were chosen for three reasons: (i) they are frequently used to study plant virus genetic variability, which facilitates comparison between EMDV and other viruses; (ii) most EMDV sequences in GenBank either corresponded to or contained the selected gene L region, allowing for comparisons of variability in Spain and other countries; and (iii) because both genes, N and L, are separated in the genome, making them useful for recombination analysis.
The complete gene N and a region of gene L (named L1) of 13 EMDV isolates collected in several regions of Spain were amplified by RT-PCR, sequenced and compared with nucleotide sequences of other EMDV isolates retrieved from GenBank (Table 1).The nucleotide sequences were deposited in GenBank under accession numbers OR631742-OR631767.EMDV isolates PV-1127 from Germany and 02923HTS from the UK showed the highest nucleotide identities with the Spanish isolates, ranging from 97.8 to 99.3% for gene N and from 97.3 to 99.1% for the genomic region L1 (Supplementary Tables S1 and S2).Comparison of the L1 region of the 13 Spanish isolates sequenced here and that of five Spanish isolates retrieved from GenBank showed that isolates S4 and S5 from the Granada province of Spain were similar to isolate 443/17, also collected in Granada (nucleotide identity > 99.0%).Isolates S1, S2 and S3 from the Spanish province of Malaga showed lower nucleotide identity with the other Spanish isolates (ranging from 96.3 to 97.1%).Isolate 203/09 also differed from the other Spanish isolates (nucleotide identity, 96.9-97.4%).The rest of the Spanish isolates showed more similar sequences (nucleotide identity 98.1-99.5%).

Phylogenetic Relationships among Worldwide EMDV Isolates
Phylogenetic analysis of the N gene of 29 EMDV isolates from eight countries showed two main clades, A and B, each divided into two subclades, A1, A2, B1 and B2 (Figure 1).Subclade A1 was composed of all isolates from Spain, one isolate from Germany and another from the UK, forming a homogenous and undifferentiated group (the minimum nucleotide identity between isolate pairs was 97.3%).Subclade A2 was composed of isolates from Iran, Italy and Greece.The minimum nucleotide identity within subclade A2 was 97.1%, whereas nucleotide identity between subclades A1 and A2 varied from 91.7 to 92.9%.Subclade B1 had two isolates, one from Greece and another from Cyprus (nucleotide identity of 97.2%), and B2 was composed of isolates from Italy, Greece and Slovenia.The minimum nucleotide identity within subclade B2 was 94.1%, and nucleotide identities between subclades B1 and B2 ranged from 89.0 to 89.6%.Nucleotide identity between clades A and B ranged from 85.3 to 87.6%.
Phylogenetic trees of the L1 genomic region (876 nt) of 39 isolates from eight countries and the L2 region (296) of 44 isolates from 10 countries (Figure 1) had the same topology with high values of bootstrap for the external nodes, but L1 showed better resolution in the external nodes.The phylogenetic tree of L2 comprised three main clades: A, B and C. Clades A and B and their divisions (A1, A2, B1 and B2) contained isolates in the same clades and subclades of the phylogenetic tree of gene N. Clade C contained the EMDV isolate T14_00910 from Australia.Nucleotide identities were 81.4-86.4% between clades A and B, 84.5-88.9% between A and C and 84.5-87.4% between B and C. Subclade A1 contained 1 isolate from Germany, 1 from the UK, the 13 Spanish isolates sequenced here and another 5 Spanish isolates retrieved from GenBank.Subclade A2 contained isolates from Iran, Greece, Italy and France and the isolate EMDVpit from Cyprus that was in the subclade B1 for gene N. Clades A1 and A2 had minimum nucleotide identities of 96.3% and 95.6%, respectively, and the nucleotide identity between these two clades ranged from 87.5 to 91.2%.Subclade B1 had only isolate, EMDVsl from Greece, and subclade B2 contained isolates from Slovenia, Italy and Greece.The minimum nucleotide identity within clade B2 was 90.5%, and that between B1 and B2 ranged from 85.8 to 88.5%.
The phylogenetic tree of a 588 nt region of gene Y (named Y1) in 17 EMDV isolates from eight countries showed three main clades: A, B and C (Figure 1).Clades A and B and the subdivisions A1, A2, B1 and B2 contained the same isolates as the equivalent clades for the N gene and the L1 and L2 genomic regions.Nucleotide identities within each subclade were higher than 94.0%, 91.3-92.9% between subclades A1 and A2 and 88.6-90.1% between B1 and B2.Clade C contained the isolate AZ15-31 from Azerbaijan, showing high divergence from the other isolates (nucleotide identities between 74.4 and 76.4%, whereas clade A and B had nucleotide identities ranging from 83.3 to 88.8%.Blastn analysis showed that isolate AZ15-31 had nucleotide identities of 82.6% with tomato alphanucleorhabdovirus 1 (TARV1) and 78.0% with physostegia chlorotic mottle virus (PhCMoV), both being distinct species belonging to the genus Alphanucleorhabdovirus.The phylogenic trees showed no correlation between genetic distances and hosts.For example, for gene N, the isolate EMDVcs from cucumber was genetically close to the isolate EG1035 from eggplant and genetically distant to another cucumber isolate, PV1212.Another example is for the genomic region L2, where the isolate EM170361 from tomato was close to the eggplant isolate EG1035 and the cucumber isolate EMDVcs but distant from another tomato isolate, Sol1, which was close to the cucumber isolate PV1212.

Recombination between EMDV Variants
Several sets of nucleotide sequences were analyzed with seven recombination detection methods implemented in the package RDP4 [23].Analysis of the complete genome (13163 nt) or each gene individually (N, X, P, Y, M, G and L) of the nine EMDV isolates with the genome completely sequenced (isolates SH-eg, STR20ST2, PV-1127, 02923HTS, Pit-MAIB, Agapanthus, PV-0031, EG1035 and PV-1212, see Table 1) showed no recombination with any of these methods.Also, no recombination was found after analyzing the N gene of 29 EMDV isolates or two regions of gene L: L1 (876 nt) of 39 EMDV isolates and L2 (296 nt) of 44 EMDV isolates.
Finally, the sequences of N and L1 were concatenated for each isolate separately for the 13 Spanish isolates sequenced here, and isolates EMDVcs, EMDVnt and EMDVpit, along with equivalent sequences which were prepared from the nine isolates with the complete genome sequenced.Thus, we generated sequences composed of gene N and the region L1 from 25 EMDV isolates to use as input.The seven methods performed using the RDP4 package detected one recombination event for isolate EMDVpit involving an ancestor of isolate STR20ST2 as the major parental and an SH-eg ancestor as the minor parental.The phylogenetic trees confirmed this recombination event since EMDVpit was placed in different clades depending on the genomic region analyzed.Thus, isolates EMDVpit and STR20ST2 are in clade B and SH-eg is in clade A in the phylogenetic trees of genes N and Y, while in the phylogenetic tree of gene L, EMDVpit and SH-eg are in clade A and STR20ST2 is in clade B (Figure 1).The nucleotide identity of EMDVpit was higher with SH-eg (98.2%) than with STR20ST2 (85.0%) for gene L, whereas the opposite occurred for genes N and Y (Table 2).This result suggests that the breaking point could exist between the genes Y and L. This recombination even was detected previously in the gene G, but only the minor parental was found [11].

Genetic Variation, Selection and Coevolution in Different EMDV Genes
Analysis of the nine completely sequenced EMDV isolates with the genome showed differences in genetic variability between genes.Nucleotide diversity was similar for genes N, P, M and G and slightly higher for genes X and Y (Table 3).Surprisingly, gene Y had the highest nucleotide diversity and the lowest proportion of segregating sites.The ratio between non-synonymous and synonymous rates (dN/dS) was less than one for every gene, indicating negative selection, ranging from 0.010 for gene Y to 0.123 for gene X.Evaluation of selection at individual codons (amino acids) showed that around 10% of the amino acids were under negative (purifying) selection, and amino acid positions 449 of gene N and 10 of gene X were under positive (diversifying) selection.

Genetic Differentiation of EMDV Populations
Genetic differentiation between EMDV populations, determined by the gene flow (migration) rate, was estimated by calculating the nucleotide diversity within and between geographical regions and using the statistic F st .
The nucleotide diversity of EMDV within Italy or Greece was high (>0.100)and about ten times lower than that within Spain (Table 4).Comparison between countries showed nucleotide diversities in the same order as those within Italy or Greece.F st values suggest high gene flow between Italy and Greece and low gene flow between both countries and Spain.To study the population structure of EMDV in Spain, four regions were considered: (i) Malaga province, (ii) Granada province, (iii) Almeria province and (iv) the rest, composed of Navarra, Pontevedra, Valencia and Zaragoza provinces.Nucleotide diversities within and between regions were lower than 0.020, except between Malaga and the other Spanish regions, with values around 0.040 (Table 5).F st values showed two genetically differentiated populations of EMDV in Spain with very low gene flow between them.One was in Malaga province, and the other corresponded to the other provinces of Spain, which showed high gene flow between them.

Discussion
EMDV genetic variability was relatively high compared to that of most plant viruses [24,25], which should be considered for detection and disease control [1,26].Thus, we had to design primers based on conserved sequence stretches for the gene L region since RT-PCR with other primers [12] failed with some Spanish EMDV isolates.
Phylogenetic analyses of genomic regions N, Y1, L1 and L2 were congruent among themselves and with those from previous studies [11,12,21].These phylogenetic trees show a certain geographic structure since the Australian isolate is in clade C, all Spanish isolates are in subclade A1 and most Italian isolates are in subclade B2.However, Greece contains EMDV isolates in clades A (subclade 2) and B (B1 and B2), and France contains isolates in subclade A2, which are more similar to distant isolates like SH-eg from Iran than isolates located in Spain, Germany and the UK.More isolates from different countries should be analyzed to gain insight into the migration routes of EMDV [27].
Phylogenetic analysis of the Y1 genomic region showed that isolate AZ15-31 from Azerbaijan [22] formed a different clade and was divergent from the other EMDV isolates, with a nucleotide identity of about 75.0%, which is in the same range as EMDV with some other species of the genus Alphanucleorhabdovirus.One of the three criteria to differentiate species within the genus Alphanucleorhabdovirus is that the nucleotide sequence identity of entire genomes must be less than 75.0%[5].Thus, the entire genome of isolate AZ15-31 should be sequenced to assign it to EMDV or another species of the genus Alphanucleorhabdovirus such as Alphanucleorhabdovirus physostegiae (physostegia chlorotic mottle virus, PhCMoV) or to propose a new species in this genus.
Recombination seems to be uncommon for EMDV since only one recombination event was detected [11].Homologous recombination is rare in negative-strand RNA viruses [28,29] in contrast to positive-strand RNA viruses [4,30,31].Recombination events found in negative-strand RNA viruses might be artifacts from mixed infections or laboratory contamination [31].So, the presence of recombinant variants should be confirmed by analyzing within-plant viral populations [32].
Some EMDV isolates collected in distant places and times showed similar nucleotide sequences.For example, EMDV isolates 1009/11 collected in Spain in 2011 and 02923HTS collected in the UK in 2022 showed 99.3% nucleotide identity.This suggests genetic stability by strong negative selection within a narrow adaptative peak [33] for the phylogenetic subclade A1.Genetic stability has been found in other plant RNA viruses [34,35].A comparison of synonymous and non-synonymous substitution rates showed that most EMDV genes were subjected to strong negative selection due to functional restrictions.This is frequent for plant viruses [24,36,37] since proteins can have multiple domains and functions in the virus life cycle [38,39].Only two codons were under diversifying positive selection, which could be result of adaptation to a new host, vector or environment [40].Positive selection has been detected in other plant viruses, for example, in the substitution of one amino acid (codon) in the movement protein of tomato spotted wilt virus (TSWV), which is an adaptation to overcome resistance in tomato [41].
The low genetic diversity and the low genetic flow of EMDV isolates from Spain with respect to those from Greece and Italy suggest a unique introduction of EMDV to Spain.The EMDV population within Spain was genetically very homogeneous with high gene flow, except in Malaga province, which showed slight divergence and low gene flow with the rest of Spain.EMDV might have undergone genetic drift after a founder or bottleneck event or experienced adaptation to new conditions in Malaga.Further analysis of additional Malaga EMDV isolates from various hosts would be necessary to verify whether this slight divergence is present in other isolates.
In contrast, Greece and Italy had diverse EMDV populations with high gene flow between both.Greece showed the highest diversity with EMDV isolates in clades A and B, so the center of diversification and dispersion might be in Greece or nearby.Some plant viruses could have dispersed and diversified from Middle East [30], where farming originated.More EMDV sequences from these countries and from Middle East would be necessary to test this hypothesis.

Virus Isolates
We collected 658 plants from different hosts and provinces of Spain from 2009 to 2020 and analyzed them by DAS-ELISA for EMDV (Loewe, Sauerlach, Germany).Only 13 plants were positive for EMDV (Table 1) and were used for RT-PCR, sequencing and sequence analyses.

Nucleic Acid Purification, RT-PCR and Sequencing
Total RNAs were extracted from 100 mg of leaf tissues using the silica-capture protocol [43] and stored at −20 • C until use.
RT was performed by denaturing 3 µL of total RNA extracts by heating at 95 • C for 5 min, incubating them on ice for 1 min and adding them to a 10 µL reaction mixture containing 5 mM DTT, 1 mM of each dNTP, first-strand buffer, 0.4 µM of primer EMDV-N1F2 (Table 6), 10 units of Ribolock RNAse Inhibitor (Thermo Fisher Scientific, Waltham, MA, USA) and 40 units of SuperScript IV Reverse Transcriptase (Thermo Fisher Scientific, Waltham, MA, USA).The reaction mixture was incubated at 50 • C for 10 min and at 80 • C
Figure 1.Unrooted maximum likelihood phylogenetic trees of four EMDV genomic regions: gene N, Y1 (gene Y, nucleotide position 277-865), L1 (gene L, position 1090-1965) and L2 (gene L, position 1375-1670) of EMDV isolates from different countries (see Table 1).Horizontal bars are proportional to evolutionary distances, and bootstrap values greater than 50% are in the nodes.Clades A, B and C and subclades A1, A2, B1 and B2 are indicated in the corresponding branches.

Table 2 .
Nucleotide identity between the recombinant isolate EMDVpit and the parental isolates SH-eg and STR20ST2.

Table 4 .
Genetic diversity and differentiation of EMDV populations in three Mediterranean countries.Gene = Genomic regions analyzed: gene N (1431 nt) and L1, a segment of gene L (876 nt).N = number of isolates per country and genomic region.N values < 5 are too low to represent a country and should be interpreted with caution.Nucleotide diversity estimated as mean nucleotide distance between sequence pairs using the substitution model T92 + G, where G = 0.27.Nucleotide diversity values within countries are in bold on the diagonal, and diversity values between countries are below the diagonal.F st values are between parentheses.

Table 5 .
Genetic diversity and differentiation of EMDV populations in Spain.Genomic regions analyzed: L1, a segment of gene L (876 nt).N = number of isolates per each region of Spain: Malaga, Granada and Almeria provinces, and Rest corresponds to Navarra, Pontevedra, Valencia and Zaragoza provinces.Nucleotide diversity estimated as mean nucleotide distance between sequence pairs using the substitution model T92 + G, where G = 0.27.Nucleotide diversities within Spanish regions are in bold on the diagonal, and diversities between countries are below the diagonal.F st values are between parentheses.

Table 6 .
Oligonucleotide primers designed for RT-PCR.Nucleotide position with respect to the sequence retrieved from GenBank with the accession number LN680656.