The dsRNA Virus Papaya Meleira Virus and an ssRNA Virus Are Associated with Papaya Sticky Disease

Papaya sticky disease, or “meleira”, is one of the major diseases of papaya in Brazil and Mexico, capable of causing complete crop loss. The causal agent of sticky disease was identified as an isometric virus with a double stranded RNA (dsRNA) genome, named papaya meleira virus (PMeV). In the present study, PMeV dsRNA and a second RNA band of approximately 4.5 kb, both isolated from latex of papaya plants with severe symptoms of sticky disease, were deep-sequenced. The nearly complete sequence obtained for PMeV dsRNA is 8,814 nucleotides long and contains two putative ORFs; the predicted ORF1 and ORF2 display similarity to capsid proteins and RdRp's, respectively, from mycoviruses tentatively classified in the family Totiviridae. The sequence obtained for the second RNA is 4,515 nucleotides long and contains two putative ORFs. The predicted ORFs 1 and 2 display 48% and 73% sequence identity, respectively, with the corresponding proteins of papaya virus Q, an umbravirus recently described infecting papaya in Ecuador. Viral purification in a sucrose gradient allowed separation of particles containing each RNA. Mass spectrometry analysis indicated that both PMeV and the second RNA virus (named papaya meleira virus 2, PMeV2) were encapsidated in particles formed by the protein encoded by PMeV ORF1. The presence of both PMeV and PMeV2 was confirmed in field plants showing typical symptoms of sticky disease. Interestingly, PMeV was detected alone in asymptomatic plants. Together, our results indicate that sticky disease is associated with double infection by PMeV and PMeV2.


Introduction
Papaya sticky disease ("meleira" in Portuguese) was first reported in 1980 in the south of Bahia and north of Espírito Santo states of Brazil [1]. The disease has spread and currently occurs

Methods and Materials Ethics Statement
Specific authorization is not required for this experiment since the author and coauthors are Professors and researchers in the Federal University of Espirito Santo (UFES) and Capixaba Institute of Research, Technical Assistance and Rural Extension (Incaper). Incaper and UFES have an official cooperation agreement (Prot. 70557870) for agricultural and biological research. Samples were obtained from plants cultivated at Incaper Experimental Farms in the north and southwest of Espirito Santo state.

Plant material
For deep-sequencing and virus purification, latex samples of papaya plants with typical symptoms of sticky disease were collected at Sooretama, Espírito Santo (ES) state, Brazil. Latex was sampled from each plant using a sterile scalpel blade and mixed (1:1, v/v) with 0.1 M sodium citrate buffer pH 5.0 [13].
Additional samples were collected for detection of sticky disease-associated viruses. Latex (5 ml) and leaf (~2 g) samples of papaya cv. Golden were collected at Baixo Guandu, Iuna, Sooretama and Vitória, ES. Samples were divided into two categories: (1) 55 symptomatic plants for sticky disease and (2) 42 asymptomatic plants. Latex samples were mixed (1:1, v/v) with 0.1 M sodium citrate buffer pH 5.0, cooled in ice and stored at -20°C; leaf samples were cooled in dry ice and stored at -80°C.

Viral genome sequencing and assembly
Viral RNAs were extracted from latex of infected plants as described by Rodrigues et al, [13]. Viral RNA bands were excised from agarose gels and purified using the PureLink Quick Gel Extraction Kit (Invitrogen). The purity was assessed based on the A260/A280 ratio with a Nanodrop ND1000 spectrophotometer (Thermo Scientific). Viral RNAs were sent to Macrogen Inc. (Seoul, South Korea) for deep-sequencing using the 454 GS-FLX Titanium platform. The reads were trimmed and quality-filtered. De novo assembly of the viral genome was carried out using the GS De Novo Assembler (v. 2.6) and SeqMan NGen (DNAStar) with default parameters. The assembled contigs were compared against sequences in GenBank using the BLASTX algorithm.
To confirm the authenticity of the 454-derived sequence, specific primers (S1 Table) were designed for genome amplification and Sanger sequencing. RNA extracted from purified virions as described by Maciel-Zambolim et al, [8] was used as a template for genome amplification. RNA was incubated at 96°C for 3 min or 70°C for 10 min to denature viral dsRNA and viral ssRNA genomes, respectively. First-strand cDNA synthesis was performed using random hexamers and Superscript III Reverse Transcriptase (Invitrogen) following the manufacturer's instructions. All PCR reactions used Platinum High Fidelity DNA Polymerase (Invitrogen) following the manufacturer's instructions in a Mastercycler Thermocycler (Eppendorf). Reactions were performed using the following program: 94°C for 3 min followed by 35 cycles of 94°C for 45 s, Tm Ã for 30 s (S1 Table), 72°C for 2.5 or 1 min (for~2000 bp and~1000 bp, respectively), and a 10-min final extension at 72°C. The amplicons were sent to Macrogen for Sanger sequencing.

Sequence and phylogenetic analyses
Viral consensus sequences of approximately 9 kb for PMeV and 4.5 kb for the second virus (PMeV2) were scanned for the presence of ORFs using Genemark (http://exon.gatech.edu/ GeneMark/eukhmm.cgi). Prediction of pseudoknot structures was performed with DotKnot [14] using as input the 100-nt sequence immediately downstream of the putative heptanucleotide at the ORF1-ORF2 junction of the PMeV coding strand.
Sequence identity/similarity for each ORF was determined using BLAST (http://blast.ncbi. nlm.nih.gov). MUSCLE [15] was used to align RdRp amino acid sequences and the conserved domain RdRP_4 (pfam02123) from PMeV, 14 members of the family Totiviridae and 10 totilike viruses, and to align RdRp amino acid sequences and the conserved domain RdRP_3 (pfam00998) from PMeV2, 16 members of the family Tombusviridae and three umbra-like viruses (S2 Table). Phylogenetic trees were constructed using maximum likelihood and Bayesian inference. Maximum likelihood trees were constructed using MEGA6 (substitution model WAG+G+I) and tested with a bootstrap of 10,000 replicates to ascertain the reliability of each branch pattern. Bayesian inference was performed with MrBayes v. 3.0b4 [16], with the model selected by MrModeltest v. 2.2 (Nylander, 2004) in the Akaike Information Criterion. The analyses were carried out running 20,000,000 generations, excluding the first 2,000,000 generations as burn-in. Trees were visualized using Fig Tree (

Virus purification and characterization of viral structural proteins
Latex collected from diseased papaya plants was centrifuged through a 10-40% (w/v) linear sucrose gradient for virus purification as described by Maciel-Zambolim et al, [8]. Final pellets were resuspended in tris-borate buffer pH 9.0 and the purified virus was stored at 4°C. The virus suspensions were scanned with UV light in the 220-320 nm range to assess their purity. Virus particles from these preparations were negatively stained in 2% potassium phosphotungstate, pH 6.8, for examination in a Zeiss EM-109 transmission electron microscope (TEM).
To identify the structural proteins, purified viral particles collected the from M fraction were digested with trypsin according to Shevchenko et al, [17]. Tryptic peptides were analyzed by MALDI-TOF/TOF (Bruker Daltonics Ultraflex III) and sequenced peptides were compared against the predicted proteins in the viral genome using Prodigal software (downloaded on 01.07.2015, with 12 entries), using the MASCOT application version 2.4.0 (Matrix Science, London, UK). The results obtained by MASCOT were validated by SCAFFOLD version 3.6.4 (Proteome Software Inc., Portland, OR), using the Peptide Prophet [18] and Protein Prophet [19] algorithms with acceptance criteria of 99.9% probability of identification for proteins and 95% for peptides. Putative protease cleavage sites on putative proteins were predicted using NetPicoRNA 1.0 Server [20].

Viruses detection by conventional RT-PCR
Viral RNAs were extracted from latex samples using Trizol reagent (Ambion), according to the manufacturer instructions. The final RNA pellet was dissolved in 20 μL of nuclease-free water and stored at -80°C. The purity (A 260 /A 280 ) was assessed with a Nanodrop ND1000 spectrophotometer. The purified RNA (1 μg) was treated with DNAse I (Invitrogen) and incubated at 96°C for 3 min or 70°C for 10 min to denature dsRNA and ssRNA, respectively. First-strand cDNA synthesis was performed using random hexamers and M-MLV Reverse Transcriptase (Invitrogen) following the manufacturer instructions. All PCR reactions used recombinant Taq DNA Polymerase (Invitrogen) in a Mastercycler Thermocycler (Eppendorf). Reactions were performed using the following programs: (i) PMeV primers: 94°C for 3 min followed by 35 cycles of 94°C for 45 s, 62°C for 30 s, 72°C for 1 min, and a 10-min final extension at 72°C; (ii) PMeV2 primers: 94°C for 3 min followed by 35 cycles of 94°C for 45 s, 62°C for 30 s, 72°C for 1 min, and a 10-min final extension at 72°C; (iii) Luteovirus primers: 94°C for 3 min followed by 35 cycles of 94°C for 45 s, 57°C for 45 s, 72°C for 1 min, and a 10-min final extension at 72°C. PCR amplicons were visualized on 1% agarose gels.

Virus identification and genome sequencing
Viral nucleic acids were extracted from latex samples of infected plants (cv. Golden) and two bands could be visualized in agarose gels, one of approximately 10 kb, likely corresponding to the PMeV dsRNA (henceforth named PMeV-ES), and a second band of approximately 4.5 kb (S1 Fig). Both RNA bands were sliced out, purified from the agarose gel and sent to Macrogen, Inc. for deep-sequencing.
The data obtained from RNA sequencing was comprised of 54,602 reads and 22,706,988 nt, with an average read length of 415 nt. Following de novo assembly, the contigs were submitted to a BLASTX search. The results indicated that it was possible to separate the contigs into two groups, each showing identity to different known viral RNA sequences. One group showed the lowest e-values with the putative structural protein of PMeV-RN (KT013296), and with the RdRp) protein of Phlebiopsis gigantea mycovirus dsRNA-2 (PgV-2; AM111097), a dsRNA virus related to members of family Totiviridae. The other group showed the lowest e-values with the RdRp of papaya virus Q (PpVQ; KP165407), a new papaya virus discovered in Ecuador and which, like PMeV-Mx, is related to members of the genus Umbravirus. Therefore, the 4.5 kb band is not a viral subgenomic RNA as previously suggested by Kitajima et al, [7] and Maciel-Zambolim et al, [8]. Instead, it corresponds to a second virus which is related to umbraviruses. The authenticity of the deep-sequencing data was validated by Sanger sequencing.

Genome organization and characteristics of the two viruses associated with papaya sticky disease
The sequence obtained for PMeV-ES is 8,814 nucleotides long (GenBank accession no. KT921784). Previous studies had estimated the length of the PMeV dsRNA to be 10-12 kb by agarose gel electrophoresis [7,8]. However, as pointed out by Abreu et al, [10], these studies used a dsDNA ladder for the determination of dsRNA length. The electrophoretic mobility of dsRNA is lower than that of dsDNA [21], which may explain the difference in size of the PMeV dsRNA observed in those earlier studies and the genome sequence obtained here and also by Abreu et al, [10].
Totiviruses typically contain two open reading frames (ORFs) that can be translated as a fusion protein by a -1 frameshift. Sequence analysis of PMeV-ES identified a potential slippery heptamer, GGAAAAC, at nt 5,297-5,303, immediately before the ORF1 stop codon (Fig 1a).
The same putative frameshifting sequence is present in PMeV-RN, and similar structures have been identified in the five viruses most closely related to PMeV (PgV-2, FgV-3, PiRV-3, SsNsV-L and BcRV-1). In these viruses, the slippery sequences differ from the general heptanucleotide XXXYYYZ, where X represents any nucleotide, Y represents A or U, and Z represents A, C or U [22,24,25]. Following the slippery sequence there is a stable secondary structure, such as a pseudoknot or hairpin [26]. PMeV-ES has a predicted RNA pseudoknot structure that lies closely downstream of the ORF1 stop codon (Fig 1b).
The partial sequence obtained for the umbravirus-like RNA is 4,515 nucleotides long (Gen-Bank accession no. KT921785), and contains two putative ORFs in different reading frames (Fig 3). The nucleotide sequence has 70% and 71% identity with PpVQ and PMeV-Mx, respectively (the sequences of PpVQ and PMeV-Mx are 79% identical).
The predicted ORF1 (nt 4-816) encodes a putative polypeptide of 270 aminoacids with a predicted molecular mass of approximately 31kDa. This polypeptide has 48% identity to the corresponding proteins of both PMeV-Mx and PpVQ. No significant matches are found for any of these putative proteins in protein databases. Together, these sequence comparison results indicate that the 4.5 kb RNA detected in papaya plants with sticky disease corresponds to an umbravirus which is closely related to both PMeV-Mx and PpVQ. However, considering that the complete sequences of PMeV-Mx and PpVQ have not yet been determined and that the nucleotide sequence identities between the currently available sequences (70-79%) are near the threshold for species demarcation in the genus Umbravirus [27], it is not possible to state at this time whether these are isolates of the same species, or distinct species. Thus, we will refer to the second RNA virus isolated from papaya plants with sticky disease as papaya meleira virus 2 (PMeV2-ES). Determination of the complete sequences of these three isolates will be necessary to clarify their taxonomic status.

Phylogenetic analyses
Phylogenetic trees based on RdRp deduced amino acid sequences were constructed to evaluate the evolutionary relationship between PMeV, 14 members of the family Totiviridae, and 10 toti-like viruses (S2 Table). Unassigned totivirid-like viruses formed a strongly supported clade, distinguishable from all other viruses in the family (Fig 4). Two main groups can be recognized within this cluster, one cluster corresponding to PMeV, PiRV-3, PgV-2, DsRV-1, and  Phylogenetic trees based on RdRp deduced amino acid sequences were also constructed to evaluate the evolutionary relationship between PMeV2-ES, PpVQ, PMeV-Mx and 16 members of the family Tombusviridae (S2 Table). The midpoint-rooted tree showed two main branches: one formed by PMeV2-ES, CYVaV, PpVQ, PMeV-Mx and members of the genus Umbravirus, and the other formed by viruses in the genera Gallantivirus, Alphanecrovirus, Machlomovirus, Carmovirus and Betanecrovirus. Interestingly, PMeV2, PpVQ and PMeV-Mx formed a clade distinct from all the other viruses with a high statistical support (Fig 5). These clusters were also observed in a maximum likelihood tree (data not shown), indicating that they reflect true evolutionary relationships, regardless of the method used for phylogenetic reconstruction.

Virus purification and characterization of viral RNA and structural proteins
Previous electron microscopy studies have indicated the presence of a single type of viral particles (isometric, 42 nm in diameter) in papaya plants with sticky disease. To confirm the association of two distinct genomic RNAs with these particles, virions were purified in a sucrose density gradient. Careful observation of the gradient indicated the presence of three welldefined opalescent zones (Fig 6a). The three fractions were carefully collected and visualized separately by TEM (Fig 6b). Interestingly, the same particle size and morphology was observed for all fractions; however, when purified RNAs from each fraction were subjected to denaturing electrophoreses, a~4,5 kb band was visualized from the middle fraction, and a~9 kb band from bottom fraction (Fig 6c); no RNA bands were visualized from the top fraction. Viral structural proteins from purified virions collected from the M fraction were identified by mass spectrometry. A total of nine peptide fragments matched the PMeV ORF1-encoded polypeptide sequence (Fig 7), accounting for 8% of the amino acid sequence (125/1563 aa) and encompassing the central region of the putative protein (from aa 356 to 785). This result indicated that the PMeV ORF1 translated product could either undergo self-cleavage or be cleaved  Table. doi:10.1371/journal.pone.0155240.g004  Table. doi:10.1371/journal.pone.0155240.g005 by papaya latex proteases, and that the cleaved product(s) comprise(s) the viral structural proteins. NetPicoRNA 1.0 server was used to predict cleavage sites at positions 313 and 1427 of the deduced aa sequence of PMeV ORF1 would result in a mature CP with a calculated molecular mass of approximately 122 kDa. However, viral capsid proteins of non-enveloped, isometric viruses are usually smaller (17-40 kDa) [28]. Additional cleavage sites with atypical consensus sequence were not detected by the algorithms employed.

Papaya sticky disease is associated with the presence of PMeV and PMeV2
To verify whether papaya sticky disease is indeed associated with double infection by PMeV and PMeV2, 55 latex samples of symptomatic plants and 42 latex samples of asymptomatic plants were collected and used for virus detection by RT-PCR. The presence of both PMeV and PMeV2 was confirmed for all symptomatic plants. In the 42 asymptomatic plants, PMeV alone was detected in 33 plants, and 9 plants had the two viruses. Asymptomatic plants infected by PMeV and PMeV2 were monitored for at least 3 months after fructification and, in some cases, eventually developed typical symptoms, while plants infected only by PMeV remained asymptomatic.

Discussion
Initial studies analyzing plants with typical symptoms of sticky disease were based upon the extraction of dsRNA from the plant latex using organic solvents, followed by gel electrophoresis to visualize the resultant dsRNA bands [7,8,13]. Usually two bands could be visualized, one at~10 kb likely corresponding to PMeV dsRNA, and a second band of approximately 4.5 kb. Sequences of the larger RNA indicated a similarity with fungal viruses related to the family Totiviridae [2,9,10]. The smaller band was regarded as a viral subgenomic RNA. However, Two Viruses Associated with Papaya Sticky Disease following recent reports of an umbravirus-related virus associated with sticky disease in Mexico, we decided to sequence and characterize both RNA molecules.
Extended sequence obtained from the 9 kb RNA (PMeV-ES), indicates that it contains two putative ORFs in an arrangement characteristic of members of the family Totiviridae. This family includes viruses of fungi and protozoa that form virions (ie, they encode a coat protein which encapsidates the viral genome) and have a single-component dsRNA genome [29].
The majority of known plant viruses have RNA genomes that generate double-stranded RNA (dsRNA) replicative forms at some point of their life cycle. Thus, dsRNA has been used to assess RNA virus infection in plants. On this regard, biodiversity surveys of plant viruses obtained from dsRNA extraction and sequencing have been done and, surprisingly, members of the Totiviridae family have been identified [30]. Liu et al, [31] performing in silico cloning using known dsRNA viral sequences as queries to search against the NCBI Expressed Sequence Tag (EST) database, obtained plant virus-like sequences that are related to members of the family Totiviridae. In another approach, small RNA (smRNA) sequencing of maize plants from southwest China resulted in the identification of a new dsRNA virus with a sequence and genomic organization resembling that of a totivirus [32]. Interestingly, siRNA sequencing from C. papaya identified totivirus-like sequences [33].
The sequence obtained from the smaller, 4.5 kb RNA, corresponds to a virus related to species in the genus Umbravirus, here named PMeV2-ES. PMeV2-ES has high similarity to PMeV-Mx and to PpVQ, a new umbra-like virus recently described in Ecuador [34]. PMeV-Mx was detected in papaya plants showing typical symptoms of sticky disease in Mexico [12]. PpVQ was detected in papaya plants also infected by Papaya ringspot virus (PRSV). It is noteworthy that symptoms of sticky disease have not been reported in Ecuador. Sequence similarity and phylogenetic analysis including PMeV2, PpVQ and PMeV-Mx suggest that they may be different isolates of the same umbravirus, or closely related umbraviruses. Either way, they are distinct from the dsRNA-containing PMeV.
Umbraviruses are distinguished from most other viruses for their lack of a coat protein (CP) gene and, as a result, umbraviruses do not form conventional virus particles. In nature, they are dependent on the presence of a helper virus-typically, a member of the family Luteoviridae. The CP of the helper virus forms hybrid virus particles encapsidating the umbraviral RNA, which can be transmitted by the helper virus vector to other plants [35]. Since our protein sequence data provided no indication of the presence of a luteovirus in infected papaya plants and no luteovirus was detected in papaya sticky diseased plants using universal luteovirus primers [36], we assumed that PMeV is the helper virus of PMeV2. To identify the capsid protein responsible for the encapsidation of PMeV2 RNA, the structural protein(s) from purified virions were analyzed by mass spectrometry. All sequenced peptides match with the deduced amino acid sequence of ORF1 from PMeV, strongly indicating that PMeV acts as the helper virus for PMeV2 and that the PMeV CP forms hybrid virus particles encapsidating PMeV2 ssRNA. To the best of our knowledge, this is the first known example of a viral coat protein being capable of encapsidating both ssRNA and dsRNA forms, and also the first known case of an umbravirus associated with a non-luteovirid.
Interestingly, all papaya plants analyzed, symptomatic or asymptomatic, were infected by PMeV, suggesting that this virus alone cannot induce sticky disease symptoms. Persistent plant viruses are associated with latent infections and have little or no overt effects on their hosts. They are found in crops, including beans, peppers and rice, where they have been studied more extensively [37]. The origin of persistent plant viruses is unknown, but their similarity to fungal viruses suggests transmission across kingdoms. Plants are almost always colonized by numerous endophytic fungi and, during this interaction, an opportunity for virus exchange between plant and fungal cells exists. Until now, persistent plant and fungal viruses have been recognized as occurring only in the families Partitiviridae and Endornaviridae [38], and phylogenetic analyses of these families support the transmission of these viruses among plant and fungal hosts [37]. A persistent virus discovered in blueberry does not appear to be a member of these families and, interestingly, has a genome organization related to members of the family Totiviridae [39].
Most persistent plant viruses have dsRNA genomes and encode only an RdRp and a CP, lacking the movement protein crucial to systemic infection. Therefore, there is no cell-to-cell movement or transport of the virus within the plant except when cell division takes place [40]. Although the PMeV genome does not encode a movement protein, the virus can infect papaya plants systemically. As PMeV is located only in the laticiferous vessels fo the plant [7], the increased latex fluidity and exudation during the infection suggest that the virus uses the laticiferous vessels to move systemically throughout the plant. However, a different scenario may occur during mixed infection: PMeV could use a movement protein from PMeV2. The ORF3 protein encoded by umbraviruses plays an essential role in mediating long-distance movement [35]. Although PMeV2 does not have a protein with similarity with the ORF3 protein, the putative protein encoded by ORF1 could be responsible for virus movement.
PMeV is an unusual plant virus: it is localized on and linked to the polymers present in the latex, a completely hostile environment for viral replication; it is the first plant virus related to viruses in the family Totiviridae; it acts like a persistent virus in single infections; and its CP encapsidates the ssRNA genome of the umbravirus-like PMeV2.
Our results indicate that papaya sticky disease is associated with a combination of PMeV, a toti-like virus, and PMeV2, an umbra-like virus. Considering the importance of the papaya industry in several tropical countries, a more accurate diagnostic method allowing the detection of both PMeV and PMeV2 is urgently necessary for the early detection of these agents with the purpose of reducing economic losses and preventing the introduction of these two viruses into new areas.