Molecular characterisation of ampeloviruses associated with mealybug wilt of pineapple disease in Ghana

Background: Mealybug wilt of pineapple (MWP) is the most destructive viral disease of pineapple worldwide. The disease is caused by pineapple mealybug wilt-associated virus (PMWaV), a member of the family Closteroviridae and the genus Ampelovirus , and transmitted by mealybugs. Methods: In order to understand the association between closteroviruses and MWP in Ghana, 24 pineapple plant samples showing typical symptoms of MWP were collected during a survey of the Central Region in 2019. Three quantitative reverse transcription-polymerase chain reaction (qRT-PCR) assays with PMWaV species specific primers were performed to amplify the heatshock protein 70 gene (HSP70) of PMWaV-1,-2 and -3 in the pineapple samples. Purified qRT-PCR products of thirteen isolates which consist of PMWaV-1 (7 isolates), PMWaV-2 (4 isolates) and PMWaV-3 (2 isolates) were sequenced in both directions. Sequence and phylogenetic analyses were then conducted. Results : Three different species of Ampelovirus namely PMWaV-1, PMWaV-2, and PMWaV-3, were detected from the plant samples, with abundance of mixed infections. Sixteen out of the 24 samples (66.7%) were infected with at least one of the three species of Ampelovirus identified. PMWaV-2 had the highest infection rate of 62.5% across the districts; this was followed by PMWaV-1 and PMWaV-3 with infection rates of 33.3% and 8.3% respectively. Purified qRT-PCR products of thirteen isolates which consist of PMWaV-1 (7 isolates), PMWaV-2 (4 isolates) and PMWaV-3 (2 isolates) were sequenced in both directions. Sequence comparison using BlastN showed that all the seven sequences of the Ghanaian PMWaV-1 isolates (GenBank Accession Nos. MN427634 - MN427639 and MN399973) shared 95.2% to 99.7% nucleotide identity with each other and 95.2-100% with

nucleotide identity of 98.9-100% with each other and 98.2-100% nucleotide identity with sequences of isolates previously published in GenBank. Also, the two sequences of the Ghanaian PMWaV-3 isolates (GenBank Accession Nos. MN427640 and MN427641) shared 98.3% nucleotide identity to each other and 97.5-99.3% nucleotide identity with sequences of isolates previously published in GenBank. Phylogenetic analyses of the nucleotide sequences of HSP70 gene of the 13 Ghanaian isolates and 24 sequences previously published in GenBank, clustered the PMWaV-1, PMWaV-2 and PMWaV-3 isolates into three distinct genetic groups with > 95% bootstrap support.
Conclusion: The present study shows for the first time the occurrence of PMWaV-1, PMWaV-2 and PMWaV-3 in Ghana pineapple fields as well as in Africa.

Background
Pineapple (Ananas comosus L. Merrill), a Bromeliaceae, is the third most important fresh fruit crop after citrus and banana worldwide [1]. Brazil, Philippines, and Thailand are the leading producers worldwide whilst in Africa, Côte d'Ivoire, Nigeria, Ghana and Kenya are the main pineapple producing countries [2].
In Ghana, the pineapple sector is the most developed horticultural sector [3,4] cultivated mainly in the areas of Central, Greater Accra, Eastern and Volta regions of Ghana, in small and medium scale. Pineapple production is a source of income for thousands of people ranging from farmers to market women and small-holder farmers. The crop provides raw material to feed industries, leading to establishment of cottage industries. Pineapple is a non-traditional export crop in Ghana and hence a source of foreign exchange. It contributed more than USD 283,000,000 in foreign exchange to the economy of Ghana between 1990 and 2013 [5].
Mealybug wilt of pineapple (MWP) is a destructive viral disease that affects pineapple production in many growing regions worldwide including Ghana [6,7]. In Ghana, fruit yield loss due to MWP attack has been estimated at about US$ 248.00 per hectare [7].
Elsewhere in Hawaii, MWP has been reported to cause reduction in fruit yield by 30-55%, depending on the age of the plant at the onset of the disease [8].
Mealybug wilt of pineapple is caused by pineapple mealybug wilt-associated virus (PMWaV), a member of the genus Ampelovirus and family Closteroviridae. Pineapple mealybug wilt-associated virus-1 (PMWaV-1), PMWaV-2, PMWaV-3, PMWaV-4, and PMWaV-5 are the five distinct species identified in Hawaii, Australia and Cuba from diseased pineapple fields [9,10,11]. These viruses are transmitted by two species of mealybugs namely the gray pineapple mealybug (Dysmicoccus neobrevipes, (Beardsley), and the pink pineapple mealybug (Dysmicoccus brevipes (Cockerell) [12], and also by man through inadvertently planting of infected planting materials (suckers, slips or crowns). These mealybugs have a symbiotic association with the ants. The ants help the mealybugs in the foundation of mealybug settlements and consume the honeydew created by the mealybugs and can suppressively affect the mealybugs' natural enemies [13][14][15][16].
Mealybug wilt of pineapple disease symptoms are displayed by serious tip dieback, downward curling, reddening, and wilting of the leaves which can prompt a complete breakdown of the plant [17,18] (Figure 1). Even though PMWaVs have been associated with MWP disease worldwide, there is no record of such an association in Ghana. Knowledge of ampelovirus species associated with MWP in Ghana is very important in devising strategies to manage the disease. Dey et al. [6] reported that PMWaV-2 species alone without the others could make the wilt symptom of pineapple to develop. Also, In Hawaii, MWP symptoms are strongly associated with infection by PMWaV-2 [8]. On the other hand, in Australia, the disease is strongly associated with infections by PMWaV-3 alone or co-infection by PMWaV-1 and -3 [11].
The aim of this study was to identify and characterize ampeloviruses associated with MWP disease that affects pineapple production in Ghana.

Quantitative reverse transcription-polymerase chain reaction (qRT-PCR)
Luna Universal One-Step qRT-PCR Kit (BioLabs Inc.) was used for the qRT-PCR amplification of headstock protein 70 (HSP70) gene of PMWaVs, according to the manufacturer's instructions. Briefly, an initial reaction volume of 12.6 µL containing 10 µL of 2× Luna Universal One-Step Reaction Mix, 1 µL of 20x Luna WarmStart RT Enzyme Mix, 0.8 µL of 10 µM reverse primer, 0.8 µL of 10 µM forward primer, was prepared and placed in qPCR tubes. Total RNA template (< 1 µg) was added to the mixture in the qPCR tubes and nuclease-free water was added to make up a final reaction volume of 20 µL. The qPCR tubes were then spun in a centrifuge for 1 min at 2, 500 rpm to remove the bubbles. The tubes were then incubated in a pre-warmed thermocycler (Applied Biosystems StepOnePlus) according to the programme reaction conditions indicated in Table 1, and SYBR scan mode setting on the real-time instrument (thermocycler). The primer sequences are shown in Table 2.

Gel electrophoresis
The amplification products were assessed by electrophoresis in 1.5% agarose gel in TBE buffer (89 mM Tris-borate and 2 mM EDTA, pH 8.3) and stained with ethidium bromide using a 2 kb ladder. The gel was then visualized in UV light in a gel documentation system and the gel photograph was then taken for further analysis.
Cleaning and sequencing of PMWaV-1, PMWaV-2 and PMWaV-3 Purified qRT-PCR products of thirteen isolates which consist of PMWaV-1, PMWaV-2 and PMWaV-3 were sequenced according to Sether et al. [20] and Gambley et al. [11] in order to assess variation within a virus isolate and to ensure consistent and reliable sequence data. The DNA bands were purified and sequenced in both directions using the BrilliantDye™ Terminator Cycle Sequencing Kit V3.1 (NimaGen BV, The Netherlands).
Sequence data were edited and assembled using BioEdit version 7.0.5 [21]. The quality of each nucleotide in the sequence was examined in order to detect and evaluate changes in nucleotides and for each amplicon construct consensus sequences. Both primer and noncoding sequences from the alignments were also removed. Additional published sequences obtained from GenBank were verified and added to the data sets (Table 3). After editing, the final sequences analysed were the partial HSP70 gene of 420 nt of PMWaV-1, 591 nt of PMWaV-2 and 486 nt of PMWaV-3. The deduced amino acid sequences analysed included 140 aa for PMWaV-1, 197 aa for PMWaV-2 and 162 aa for PMWaV-3. The sequences obtained in this work were analysed together with those retrieved from the GenBank (Table 3).
Multiple sequence alignments were made using the ClustalW programme implemented in MEGA version 7.0 [22]. Alignments were also manually altered to guarantee right reading frames. The analyses included a total of thirteen isolates of PMWaV-1, PMWaV-2 and PMWaV-5 nucleotide sequences of HSP70 genes from pineapple fields in Ghana, and 24 corresponding sequences of isolates previously published in GenBank.

Sequence comparisons and phylogenetic analyses
For HSP70 homologous genes of PMWaV-1, PMWaV-2 and PMWaV-3, the nucleotide and the deduced amino acid sequence identities were determined using BioEdit v7.0.5 [21] and BlastN. For HSP 70 homologous genes between PMWaV-1, PMWaV-2 and PMWaV-3 sequence alignments, MODELTEST [23] implemented in MEGA version 7 program [22] was conducted to select the most suitable nucleotide substitution model using the Akaike Information Criterion, the Bayesian Information Criterion [24] and the hierarchical probability ratio test. The best fit nucleotide substitution model was then used for phylogenetic analyses using the maximum likelihood method with 1000 replicates of bootstrapping using the MEGA 7 software [22]. The neighbour-joining method also implemented in MEGA 7 was used for comparison.

Genetic diversity
The following genetic diversity indices for all samples of the HSP70 homologous gene each of the PMWaV-1, PMWaV-2 and PMWaV-3 were determined using the DnaSP V.5.0 programme [25]: haplotype diversity (h), nucleotide diversity (π), number of segregating sites (S) and total number of mutations (Eta).
Determination of genetic distance and selection pressure For each of PMWaV-1, PMWaV-2 and PMWaV-3 sequence dataset, the overall genetic distance (the number of base substitutions per site from averaging across all sequence pairs in a population) within HSP70 homologous nucleotide sequence data sets were estimated using the Maximum likelihood model [26]. Bootstrap method (1000 replicates) was used to obtain standard error estimates. The analyses were conducted in MEGA 7.
The HyPhy package Maximum Likelihood analysis of the natural codon-by-codon selection technique [27] implemented in MEGA 7 [22] was used to predict the number of synonymous substitutions inferred per synonymous site (dS) and the number of nonsynonymous substitutions per non-synonymous site (dN). These estimates were produced

Results
Detection of the viral species responsible for MWP disease Three different species of Ampelovirus namely PMWaV-1, PMWaV-2, and PMWaV-3 were detected by qRT-PCR from the plant samples during the study (Table 4). Sixteen out of the 24 samples (66.7%) were infected with at least one of the three species of Ampelovirus identified. PMWaV-2 had the highest infection rate of 62.5% across the districts, indicating that it is the most prevalent virus species in the region; this was followed by PMWaV-1 and PMWaV-3 with infection rates of 33.3% and 8.3% respectively (Table 4; Figure 1). All the three viral species were detected from pineapple samples from AAK district, whilst only two (PMWaV-1 and PMWaV-2), were detected from Ekumfi and KEEA districts, implying that PMWaV-3 was found only in AAK district.
Performance of the PMWaVs primers on samples across the growing area The deduced amino acid sequences of the Ghanaian isolates also ranged from 86.5 to 99.2% for the PMWaV-1, 97.1 to 100% for PMWaV-2 and 95.3% for the PMWaV-3 (Table 6) (Table 6).

Phylogenetic analyses
The maximum likelihood tree for the partial HSP70 gene nucleotide sequence dataset  (Table 8).
This provided evidence of heterogenous selection pressures among codon sites in HSP70 genes for PMWaV-1, PMWaV-2 and PMWaV-3 datasets. There was also comparison for the overall selection intensity in the HSP70 genes. The results showed that the selection intensity (mean pairwise d N / d S ) for this gene was 0.2587 for PMWaV-1, 0.2696 for PMWaV-2 and 0.1545 for PMWaV-3) (Table 8). Thus, overall, the values of the d N/ d S were low, i.e. d N /d S < 1, implying that the HSP70 gene of PMWaV-1, PMWaV-2 and PMWaV-3 was under negative selection.

Neutrality tests
The results for the various neutrality tests are summarised in Table 9. Tajima's D test and Fu and Li's F* test for PMWav-1 were significant in terms of neutrality deviation (P < 0.05), but the rest of the tests (Fu and Li's D* and Fu and Li's F* tests) did not detect significant neutrality deviation (P > 0.05) for the PMWaV-1, PMWaV-2, and PMWaV-3 populations.

Discussion
Over the years, symptoms alone have not been effective in the detection of the plant viral disease [36]. The detection of PMWaVs by molecular means has however been shown to be reliable and efficient [11,20]. To identify ampeloviruses associated with MWP disease, al. [6] and Gambley et al. [11] which describe PMWaV-1, PMWaV-2 and PMWaV-3 isolates as three distinct closterovirus species. It was however interesting to note that the HSP70 ORF amino acid phygenetic tree revealed that PMWaV-1 isolates which clustered at clade 1 had two significant sub-clusters supported by 92% bootstrap support with our field isolate with accession number MN427634 clustring with isolate from Genbank with accession number HG940514 (Figure 4). On the other hand, PPMWaV-2 and PMWaV-3 isolates did not form sub-clusters. This suggests that HSP70 gene of PMWaV-1 is more diverse than that of PPMWaV-2 and PMWaV-3. This is supported by the relatively higher nucleotide and haplotype diversities recorded for PMWaV-1 than that of PMWaV-2 and PMWaV-3 (Table 7). This could be due to the greater number of mutations and recombination in the gene of and PMWaV-1 compared to that PMWaV-2 and PMWaV-3, which is consistent with the finding of Melzer et al. [18]. According to report by Roossinck, [40] Table 8).
The high rate of mutation in RNA viruses could not be due to an evolutionary strategy but to the need for replication of their chemically unstable RNA genome [40]. However, high mutation rates for RNA viruses have been revealed to represent an evolutionary strategy [42].
An indication of population substructuring was the important neutrality deviation observed from the neutrality trials. All the tests for neutrality showed negative values (see Table   24), indicating that all PMWaV-1, PMWaV-2 and PMWaV-3 populations were in active evolution.

Conclusions
Three different ampeloviruses namely PMWaV-1, PMWaV-2 and PMWaV-3, were detected from the plant samples following qRT-PCR assays with PMWaVs species specific primers and sequence and confirmed phylogenetic analyses of nucleotide and amino acid sequences of HSP70 gene. The HSP70 gene of the PMWaV-1, PMWaV-2 and PMWaV-3 had low nucleotide diversity and was under negative selection. Mixed viral infections by 2 or all three viral species were detected in the pineapple samples from two out of the three districts in the Central region. This is the first report of PMWaV-1, PMWaV-2 and PMWaV-3 in Ghanaian pineapple and in Africa.

Acknowledgement
We are grateful to Regional University Forum for Capacity Building in Agriculture