Begomoviruses (family Geminiviridae) are plant viruses with smallsingle-stranded circular DNA molecules encapsulated in twinnedparticles that are transmitted by the white fly Bemisia tabaci (Genn.] todicotyledonous plants and have either monopartite (DNA A genomeof about 2.7 kb, encoding six ORFs) or bipartite genomes (DNA Aand DNA B genomes of 2.5-2.6 kb each, encoding eight ORFs) [1].The DNA A component of the bipartite begomoviruses is involvedin replication and production of virions, but requires the DNA Bcomponent for nuclear localization, systemic infection, host rangedetermination and symptom expression [2]. Both DNA A and DNAB components have a non-coding Common Region (CR) sequence ofapproximately 200 bp that contains sequence motifs required for the control of gene expression and replication [3].

Jute (Corchorus capsularis L. and C. olitorius) is an eco-friendlynatural fiber-producing crop belonging to the family Tiliaceae, or morerecently to Malvaceae. It is known for its versatility, fiber strength andbiodegradability of lingo-cellulosic fibers and is almost exclusivelycultivated in India and Bangladesh; nearly 98% of the world crop isgrown in these two countries [4,5]. Jute plants are infected with manydiseases [6]. Among them, jute yellow mosaic disease causes drasticreduction in the quality and yield of jute, and has been consideredto be one of the most important limiting factors of jute cultivation.The disease was first reported by Finlow [7] in 1917, but the etiologywas not known for many years, until Ha et al. [8,9] identified twobegomoviruses, namely, Corchorus yellow vein virus (CoYVV) andCorchorus golden mosaic virus (CoGMV), from diseased plants inVietnam. These bipartite viruses possess DNA A, which was foundnot to encode an AV2 open reading frame, and DNA B genomiccomponents. Subsequent identification of CoGMV from India [10,11]has provided further evidence that CoGMV has a wide distribution inAsia. CoGMV infection reduces the levels of chlorophyll, total protein,catalase, peroxidase and esterase and also reduces plant height, so itlowers the quality and yield of jute fiber [12]. The incidence of thedisease has been found to be around 50% on some of the leading C. capsularis cultivars in India. Ghosh et al. [11] also demonstrated that, like other begomoviruses, CoGMV can be transmitted by whitefly and the efficiency of its transmission ranges from 20 to 60% for infestation with 3 and 10 viruliferous whiteflies. Considering the importance of the jute fiber and the economic impact of the viral disease, we designed our research work on molecular characterization and confirmation of the causal virus of jute yellow mosaic in Bangladesh.

Virus source

Naturally infected plants of C. capsularis cultivar CLV-1 showingsymptoms of yellow mosaic disease and healthy plants were collectedfrom Mymensingh, Rangpur and Dinajpur districts of Bangladesh(Figure 1A). The plants exhibiting symptoms of typical mosaic diseaseand healthy plants were then subjected to collect sample leaves. Sampleleaves were dried using silica gel and assayed for virus infection andseed transmission at the plant virology laboratory of the Departmentof Agrobiology, Graduate School of Science and Technology, Niigata University, Japan.

DNA extraction

Nucleic acids were extracted from symptomatic leaves (20 mg) bya modified Cetyl Trimethylammonium Bromide (CTAB) method [13].The total DNA was suspended in 15 μl of sterile distilled water and used in PCR procedures.

PCR amplification of DNA A and DNA B components

Extracted DNAs from four samples were used in Polymerase ChainReaction (PCR) to amplify the complete viral DNA A and DNA Bgenomes of CoGMV using the 3 abutting primer pairs, CoGMVAF/CoGMVAR, CoGMVB1F/CoGMVB1R and CoGMVBF/CoGMVBR,were designed according to the determined sequences (sizes 2320, 2190,260 and 138 bp for Co-V1/R1, Co-V1/R2, Co-V2/R1 and Co-V2/R2primer pairs, respectively). The primers used in this study are listed inSupplementary file Table 1 and all amplifications were performed in areaction mixture of 20 μl using PrimeStar Max enzyme according to themanufacturer’s instructions (Takara Bio Inc., Japan). The PCR cycleswere programmed as 20 s for initial denaturation at 98°C, and 15 s eachfor denaturation at 98°C, annealing at 50-58°C and chain extension at72°C (25 cycles). To confirm the completion of amplification of all thetarget templates, a final extension cycle was carried out at 72°C for 7 min.

Table 1: Nucleotide sequence identity of CoGMV isolates with other begomoviruses and comparison of ORF-wise amino acid sequence identity for DNA A and DNA B at the nucleotide level. Highest values are indicated in bold.

Cloning and sequencing

The PCR products were cloned into T-vector pMD20 and usedto transform Escherichia coli (JM109) following the manufacturer’sinstructions (Takara Bio Inc., Japan). Selected clones were completelysequenced by a commercial company (SolGent Co., Ltd., Korea). Thesequence data were assembled and analyzed using GENETYX Win.Software package (GENETYX, Tokyo, Japan) ver. 12. Multiple sequencealignments of viral genomic sequences and pairwise comparisons werecarried out with the help of ClustalW software [14]. Phylogenetictrees were generated with Molecular Evolutionary Genetics Analysis (MEGA) software, version 6 [15] using the neighbor-joining method.

Construction of agroinfectious clones

For construction of CoGMV-[BD:Din:10] [BD3] DNA A infectious clone, the EcoRV-to-KpnI (about 1.7 kb) fragment of DNAA clone (1.0 mer) was separated and removed by gel electrophoresisand the remaining (0.4 mer) of the clone was purified, treated withT4 polymerase and self-ligated which gives pMD20A0.4mer. Theviral fragment was confirmed, and a full length DNA A componentreleased as NheI fragment was further ligated with the NheI linearizedpMD20A0.4mer to generate pMD20A1.4mer. The orientation of theconstructs was confirmed by restriction digestion with AflII, which wasexpected to release 2.7 kb fragment. The 3.7 kb band (1.4 mer), releasedby digestion with HindIII and EcoRI from the recombinant pMD20clone of DNA A was subsequently cloned at the same restriction sitesof a binary vector pRI201-AN-GUS and the final clone was named pRI1.4A.

For construction of CoGMV-[BD: Din: 10] [BD3] DNA Binfectious clone, a similar strategy was adopted. The BglII-to-KpnI (0.5kb) fragment of DNA B clone (1.0 mer) was separated and removedby gel electrophoresis and the remaining (0.8 mer) of the clone waspurified, and treated with T4 polymerase and self-ligated to givepMD20B0.8mer. The viral fragment was confirmed and a full lengthDNA B component released as a FbaI fragment was further ligatedwith the FbaI linearized pMD20B0.8mer to generate pMd20B1.8mer.The orientation was confirmed by restriction digestion with PvuII,which was expected to release 2.7 kb fragment of DNA B. The 4.9 kbband (1.8 mer) released by digestion with HindIII and EcoRI from therecombinant pMD20 clone of DNA B was subsequently cloned at thesame restriction sites of a binary vector pRI201-AN-GUS and the final clone was named pRI1.8B.

Agroinoculation

Jute (Corchorus capsularis) seedlings of the cultivar CVL-1 weregrown in vermiculite inside a temperature controlled growth chamber maintained at 22±2℃ and a 16/8 h light/dark cycle. Third and fourth leaf stage jute seedlings were used for agroinoculation. CoGMV DNA A was agroinoculated either alone or with cognate DNA B. Agrobacterium tumefaciens strain C58C1 cells containing the infectious constructs pRI1.4A and pRI1.8B were grown for 48 hours on Luria Bertani medium (pH 6.8) containing kanamycin (50 μg/ml) and rifampicin (100 μg/ml). Agrobacterium cells were harvested and resuspended in MES buffer (10 mM 2-[N-morpholino) ethanesulfonic acid (MES), 10 Mm Magnesium chloride (MgCl2) and 100 μM acetosyringone). DNA A and DNA B were mixed at equal concentrations (1.0 OD) and introduced by cutting of the leaf petiole with sharp blade and agroinoculated plants were grown in an insect free growth chamber at 22 ± 2°C with 12/12 hour light/dark cycle and observations were recorded periodically.

Symptoms and detection

Yellow flakes appeared on the lamina of young leaves of CoGMVinfected plants, which gradually increased in size to form green andchlorotic intermingled patches (Figure 1B). The full-length DNA A andB clones of three CoGMV isolates, namely, BD1 (Mymensingh), BD2(Rangpur) and BD3 (Dinajpur), were amplified and fully sequenced,using the primer pairs CoGMVAF and CoGMVAR and CoGMVBFand CoGMVBR. Our attempts to amplify CoYVV and sub-genomiccomponents from the jute samples using primer pairs targeting the CoYVV, alphasatellite [16] or betasatellite [17] were unsuccessful.

Figure 1: A) Locations where jute plants were sampled, B) Yellow mosaic symptoms of jute plants infected with CoGMV.

Characterization of CoGMV DNA A and DNA B

The complete nucleotide sequences of CoGMV DNA A were determined to be of 2676 nt [BD1, Accession no. AB849288] and2687 nt [BD2, Accession no. AB849290 and BD3, Accession no.AB849292] in length. The complete BD1 DNA A sequence sharedthe highest nucleotide sequence identity at 94.2% with CoGMV DNAA from Vietnam (Table 1). The DNA A sequences of BD2 and BD3shared the highest nucleotide sequence identities at 98.8 and 98.5%,respectively; with Indian isolate of CoGMV DNA A (Table 1). TheDNA A component of CoGMV had a typical New World bipartitebegomovirus genome organization; all three isolates lacked the AV2 ORF that is present in Old World begomoviruses.

The complete nucleotide sequences of CoGMV DNA B weredetermined to be of 2668 nt [BD1, Accession no. AB849289], 2658nt [BD2, Accession no. AB849291] and 2667 nt [BD3, Accession no.AB849292] in length and to encode two ORFs, BV1 in virion senseand BC1 in complementary sense orientation. The percent nucleotideidentities of the complete DNA B and the two ORFs with the respectivesequences of different begomoviruses used for analysis are listed inTable 1. The DNA B of BD1 shared the highest sequence identity at72.6% with two Indian isolates of CoGMV. The complete DNA Bsequence of other isolates [BD2 and BD3] shared the highest sequence identity at 97.1% with CoGMV Indian isolate.

The complete nucleotide sequences of CoGMV [BD1-3] werealigned with the corresponding sequences of other begomovirusesshowing similarities in BLAST search (Supplementary file Table 2). Theneighbor-joining phylogenetic analysis using complete DNA A andDNA B sequences of three Bangladeshi CoGMV isolates with the other CoGMV reported from India and Vietnam formed a distinct cluster that was more closely related to New World begomoviruses than to viruses from the Old World (100% bootstrap support) (Figure 2).

Figure 2: The neighbor-joining phylogenetic analysis A) using complete DNA A and DNA B sequences of three Bangladeshi CoGMV isolates with the other B) CoGMV reported from India and Vietnam formed a distinct cluster that was more closely related to New World begomoviruses than to viruses from the Old World [100% bootstrap support].

Agroinoculation

In order to study the infectious nature of CoGMV-[BD:Din:10][BD3], agroinfectious constructs of both genomes were introducedinto the leaf petiole of jute using agroinoculation technique. PlainAgrobacterium tumefaciens cells were mock inoculated without anyof the constructs as a control. Agroinoculation of constructs werecarried out either individually or in combination. Initial small spotson leaf lamina was observed at 15 days post inoculation (dpi), which intensified further as typical yellow mosaic symptoms at 20 days post inoculation (dpi) in jute plants when both the genomic components (DNA A and DNA B) were co-inoculated (Figure 3A). The DNA A genomic component failed to develop disease in jute when inoculated alone (Figure 3B). No noticeable phenotypic changes in jute plants were observed which were inoculated with plain Agrobacterium tumefaciens (control). The disease expressions of are summarized in Table 2.

Table 2: Infectivity and symptom induced by CoGMV-[BD:Din:10] (BD3) and the number of symptomatic plats as confirmed by PCR.

Figure 3: Infectivity of CoGMV-[BD:Din:10] (BD3) in jute plants. A) Seedlings agroinoculated with CoGMV DNA A and DNA B. B) Seedlings agroinoculated with CoGMV DNA A alone. C) Seedlings agroinoculated with empty vector.

Total DNA from systemic and symptomatic leaves of agroinoculated jute plants was extracted as described above and subjectedto PCR using primer pair CoGMVAF/CoGMVAR. The expected 2.7kb band was observed only in plants agro inoculated with both thegenomic components. We could not observe any band when plants were inoculated with Co GMV DNA A alone.

Yellow mosaic disease has immense economic importance,including through its association with jute plants in Bangladesh. Inphylogenetic analyses, DNA A sequences of CoGMV BD1-3 isolateswere clustered into a group comprising other isolates of CoGMVreported from India and Vietnam. On the basis of DNA A sequencesimilarity, BD1 was most closely related to Vietnam isolate (94.2%),whereas BD2 and BD3 were closely related to Indian isolates (98.2-98.8%). The level of nucleotide identity (92.1-92.8%) found betweenDNA A sequences of BD1 and other Indian and BD3 isolates ofCoGMV was slightly lower than the strain demarcation level of 93%[18]. However, the DNA A of Vietnam isolate exhibited intermediatelevels of nucleotide identity of 94.2% and 93.9% with BD1 and BD2,respectively, indicating that BD1, Vietnam and the other CoGMVisolates should be regarded as variants. It is also noteworthy that DNAB of BD1 had only 72.1-72.6% nucleotide identity with other CoGMVisolates. Possible recombination events could not be detected in DNAB of BD1 with other CoGMV and CoYVV isolates, using RDP ver. 3[19] (results not shown). DNA B component, by virtue of encoding no overlapping genes, has a greater capacity for variation [20].

Two begomoviruses, CoYVV and CoGMV, were subsequentlyidentified in Vietnam, the infectivity using cloned DNA has beendemonstrated only for CoYVV by microprojectile bombardmenton tobacco culture cells. The mechanical transmission of the yellowmosaic disease has not been demonstrated, presumably due to thepresence of a large amount of mucilage and phenolics in jute plant[13]. Virus isolated from a plant may or may not be the cause ofdisease unless it satisfies the Koch’s postulates. Therefore, to satisfy theKoch’s postulates, the infectivity of CoGMV-[BD: Din: 10] [BD3] wasestablished by inoculating the agro infectious clones in jute plant. Theagro inoculation in jute plants resulted in the disease symptoms that aresimilar to those occurred in the virus infected jute plant in fields. Thus,we fulfilled the Koch’s postulates and showed for the first time thatCoGMV-[BD: Din: 10] [BD3] is responsible for the newly emergingyellow mosaic disease of jute in Bangladesh. Empty agrobacterium cells(control) or only DNA A of CoGMV-[BD: Din: 10] [BD3] inoculatedplants did not show any symptom by 48 day dpi. This showed that onlyDNA A of CoGMV-[BD: Din: 10] [BD3] was neither able to sustainnor produce systemic infection in the host plants. This phenomenon is observed for most of the geminiviruses with the bipartite genomes [21].

The first author is grateful to Honjo International Scholarship foundation,Tokyo, Japan, for providing financial assistance during the tenure of which this work was carried out.