Introduction

Cassava (Manihot esculenta Crantz, Family Euphorbiaceae) is widely grown in tropical regions of the world and its tuberous roots are used as a source of dietary carbohydrates in many countries. In India, cassava is grown in substantial amounts in the southern states of Andhra Pradesh, Kerala and Tamil Nadu (1, 2).

Geminiviruses (family Geminiviridae) cause significant yield losses in crop plants throughout the tropical and sub-tropical parts of the world [1,2,3,4]. Within the family Geminiviridae, members of the genus Begomovirus are transmitted by whiteflies of the species Bemisia tabaci [5] and consist of one (monopartite) or two (bipartite) circular single-stranded DNA components of about 2.7 kb in size. Cassava mosaic disease (CMD) is widespread in cassava in the African continent, India and Sri Lanka [1, 2, 6, 7]. Two begomoiruses are associated with CMD in the Indian subcontinent; Sri Lankan cassava mosaic virus (SLCMV) and Indian cassava mosaic virus (ICMV) [8,9,10], SLCMV being more common as compared to ICMV [11,12,13]. Following the initial report of an infectious SLCMV clone from Adivaram, Kerala, (SLCMV-Adi) on Nicotiana benthamiana [10] another cloned DNA of SLCMV, from Attur, Tamil Nadu (SLCMV-Attur) has been recently shown to be infectious on N. benthamiana [14].

Apart from spontaneous variation arising due to the error-prone replication mechanism in RNA viruses [15], recombination has also been proposed to be a reason behind the frequent occurrence of variants [16,17,18,19], especially in geminiviruses, where, the intergenic region of the DNA has a role in the generation of variation through intermolecular recombination and pseudorecombination during mixed infection of viruses [20,21,22,23,24]. Earlier studies on the variability of cassava-infecting begomoviruses from India used the methods of PCR and PCR-RFLP [11, 13], both of which require prior information on nucleotide sequences of the viruses to be studied. To detect novel begomoviruses or recombinants within the known and novel begomoviruses, sequence-independent detection methods such as rolling circle amplification (RCA) are very useful [25, 26].

To generate additional information on begomoviruses associated with CMD in southern India, samples from cassava plants showing symptoms of CMD were collected from 80 locations in 9 districts of Tamil Nadu, India and RCA-RFLP was used to analyze the molecular diversity. As the first step, a conserved CP gene was amplified by PCR, using CP-gene specific primer (based on the Accession No AJ579307) followed by cloning and sequencing. Variability of amino acid sequences of CP was checked from 12 samples. In the second step, RCA-RFLP technique using the restriction enzymes EcoRI, was used to determine variability in the amplified begomoviral DNA patterns from all the samples. Some of the RCA-RFLP generated DNA fragments were cloned, partially sequenced and aligned with SLCMV-Attur genome (Acc. No KC424490) to analyse the mutational changes, protein variability or any hot spot of insertion or deletion. This results indicate that both ICMV and SLCMV display sporadic scattered single nucleotide changes in their DNA, which result in the accumulation of a low level substitutions in the viral proteins. This knowledge will help in developing effective anti-viral strategies to combat the disease with the confidence that it will be applicable uniformity across the regions surveyed, due to the genetic uniformity of the viruses.

Materials and methods

Collections of cassava stem cuttings

A total of more than 100 cassava fields (Supplementary Table 1, Fig. 1) were visited in end of September, 2011. Cassava plants aged three to 6 months of age were selected. The leaves of symptomatic and non-symptomatic cassava and stem cuttings were collected and carried to the University of Delhi South Campus, New Delhi, India. Stem cuttings were planted in soil in pots in the glass house maintained at 30 °C with 16 h light and 8 h dark periods. Emerging fresh leaves, sprouted from the above stem cuttings were used as a source for the further experiments.

Fig. 1
figure 1

Map Showing the sites of collection (shown by dark dot) of cassava stem cuttings

Full size image

Isolation of genomic DNA from cassava leaves

Total genomic DNA was extracted from glasshouse-grown symptomatic cassava leaves using hot SDS method [27].

Cloning and sequencing of coat-protein genes

Genomic DNA extracted from symptomatic leaves were used to amplify the complete CP gene by PCR, using synthetic abutting oligonucleotide based on published nucleotide sequence of SLCMV-Adi (Acc. No AJ579307) [10]. The amplified DNA were cloned into pTZ57R/T vector (MBI Fermentas). DNA sequencing was performed through services provided at University of Delhi South Campus, New Delhi, India.

Rolling circle amplification (RCA)

RCA was carried out using the TempliPhi™ Amplification Kit (GE Healthcare, Amersham) as per manufacturer’s recommendation for amplification and cloning purpose. Ten to twenty nanogram of total genomic DNA were dissolved in 5 µl of sample buffer, denatured for 3 min at 95 °C and cooled down to the room temperature. Then 5 μl of reaction buffer and 0.2 μl enzyme ø29 DNA polymerase were added and incubated at 30 °C for 18–20 h.

Cloning and sequencing of RCA fragments

Digestion with EcoRI restriction enzymes (Fermentas) was performed to linearize the concatameric RCA product and the resulting fragment was cloned with and without gel electrophoresis in pTZ57R vector (MBI Fermentas) and partially sequenced using a universal M13 primer [28, 29]. Randomly selected RCA-RFLP fragments, showing unusual banding patterns were sequenced.

Sequence analysis

Nucleotide sequences from the amplified fragments were searched by BLAST at NCBI server (www.ncbi.nlm.nih.gov/) and analysed by the software Gene Runner version 3.05. Nucleotide identities between cloned viral DNA molecule and other selected begomoviruses were analysed by the ClustalW method and NCBI BLAST server [30]. Multiple sequence alignment was performed by mega 6.0 software and ClustalW. Protein sequence variability was estimated by using the Protein Variability Server [30, 31].

Results

Collection of samples and observation of symptoms

Both mild and severe symptoms could be observed on naturally-infected cassava plants in the fields visited. However, most of the samples collected from the districts of Dharmapuri, Erode, Salem and Villupuram showed mild mosaic symptoms. Plants showing symptoms of severe mosaic were few. Symptoms in the samples collected from Cuddalore district were very mild. The types of symptoms displayed are shown in Supplementary Table 1. In most cases, the symptoms displayed at the field were replicated in the leaves sprouted from the stem cuttings grown under glasshouse conditions.

Genetic diversity of cloned CP genes of SLCMV in southern India

Genetic diversity of twelve complete CP genes were randomly cloned from field grown sprouted leaves from stem cuttings collected from the districts of Cuddalore (2), Dharmapuri (2), Salem (4), Theni (1) and Villupuram (3) (Supplementary Table 1, shown in bold). Although attempts were made to clone CP genes from isolates showing severe symptoms, only the above 12 could be cloned, which showed mild symptoms. Upon multiple alignment, out of the 256 amino acid residues, changes detected at only six positions (Supplementary Fig. 1). Out of a total of 12, only five CP sequences showed substitution mutations. One CP sequence (CP7) showed substitutions at three positions. CP10, CP41, CP46 and CP79 showed substitutions in single positions only as compared to SLCMV-attur. In CP7, at three substitution involved a change of a polar amino acid threonine was changed to basic amino acid arginine, non-polar leucine to non-polar isoleucine and N to polar glycine. In CP10, CP41 and CP46 non-polar amino acid alanine were changed to polar serine amino acid, non-polar valine and polar serine respectively. In CP79, charged aspartic amino acid was changed to polar tyrosine (Supplementary Figure 1).

Analysis of cassava infecting begomoviruses by RCA-RFLP

RCA-RFLP was used to investigate the presence of ICMV and/or SLCMV in sprouted leaves of cassava stem cuttings collected from the fields. The patterns obtained after RCA-RFLP were analyzed and compared with that expected from in silico analysis of the nucleotide sequences of ICMV and SLCMV, using the available full length sequences from publicly available nucleotide databases. The expected band patterns based on the sequence of SLCMV-Attur (KC424490) were 1844, 794, 70 and 50 bp for SLCMV DNA-A and 2722 and 16 bp for SLCMV DNA B Acc. No (AJ579308). Three expected bands were 1623, 923 and 120 for ICMV DNA-A (AY730035) and 2406, 221 and 16 bp for ICMV DNA-B (AY730036). Hence, two bands above 2.0 kb indicate the presence of both ICMV and SLCMV. Such patterns were found in eight samples (sample 2, 20, 27, 30, 46, 56, 60 and 78, Fig. 2). Eleven samples produced band patterns, which were not expected (samples 31, 41, 42, 47, 52, 54, 58, 60, 69, 72 and 78), based on the known sequence information. Fourteen samples showed no amplification and the remaining 47 samples indicated the presence of only SLCMV (Fig. 2). There was no sample which indicated the presence of only ICMV.

Fig. 2
figure 2

RCA/EcoRI RFLP patterns of begomoviral DNAs associated with cassava mosaic disease. Sample collection sites for 1–80 are shown in Supplementary Table 1. Red arrow indicated the ICMV. M1 = GeneRuler™ 100 bp DNA Ladder and M2 = GeneRuler™ 1 Kb Ladder are size markers (color figure online)

Full size image

Thirty one RCA-RFLP generated fragments were randomly cloned and sequenced to an average length of 700 bp. When the sequences were searched for maximum identities with database entries using publicly accessible databases, they showed between 97 and 99% identities with entries described as either as ICMV or SLCMV DNA-A or DNA-B (Supplementary Table 2), isolated from the southern Indian states of Kerala or Tamil Nadu. When the sequenced portions were aligned with the known sequences of SLCMV/ICMV, 15 fragments covered the AC4 gene and 3 covered the AC2 gene completely. This enabled the comparison of the derived amino acid sequences of AC2 and AC4 proteins encoded by the cloned DNA fragments. When these amino acid sequences were studied for variability, 5 amino acid residues were found to be variable in AC2 and 11 in AC4 (Fig. 3a, b), compared to the amino acid sequences of the corresponding viral proteins derived from the infectious clone of SLCMV (SLCMV-Attur, KC424490), although there were an average 2–3 single base changes for each 100 bases analyzed in the nucleotide sequences.

Fig. 3
figure 3

Sequence variability plot of the AC2 and AC4 amino acid sequences of cassava infecting geminiviruses represented according to Shannon’s method [31]. A score higher than 0 indicate variable amino acid positions. Sequence variability ranges from 0 (only one amino acid type is present at that position) to 4.322 (all 20 amino acids are equally represented in that position). a The bar and arrowhead (blue triangle) indicate the locations of the predicted motifs in AC2 (29RRRR32 nuclear localization signal), C37, C39, H44, H54, and C47 (promoter activation). Conserved amino residues of the C-terminal activation domains of AC2 in SLCMV-Attur, 121LDD123 and 127S are indicated by bar. b Myristoylation motif in AC4 is shown by bar (color figure online)

Full size image

Overall, it can be said that although some samples showing severe symptoms were analysed along with a large number of samples showing mild symptoms, no specific genetic relationship could be established between the symptoms and the genetic nature of the viruses.

Discussion

This work was undertaken to try and generate additional information on the nature and variability of cassava infecting begomoviruses in southern India. An attempt was also made to detect any other begomoviral DNA or satellites DNAs, which may be associated with CMD-affected cassava in samples, other than what has been reported earlier [11, 13]. RCA is able to generate sequence non-specific amplification products derived from begomoviruses in plants [25] and hence, was considered appropriate for this investigation.

With that in mind, symptomatic cassava samples were analyzed for the presence of new begomoviral DNAs by using RCA-RFLP. The choice of EcoRI as a selecting enzymes was advantageous because it was capable of producing discrete sets of DNA fragments representing both DNA components of ICMV and SLCMV. The results indicated that a vast majority of the samples analyzed (47 out of 67 which produced patterns) contained only the two SLCMV DNA components and a minor fraction contained the components of both ICMV as well as SLCMV (8). This agreed with the earlier reports of surveys performed using PCR and PCR-RFLP [11, 13] and reinforced the view that SLCMV is more widespread than ICMV in India. Interestingly, a study has recently shown evidence of a possible suppression of ICMV accumulation by SLCMV, in case of joint infection of naturally infected cassava plants collected from the state adjoining Tamil Nadu, Kerala; the ICMV emerging only after “recovery” of the plant from SLCMV infection [32]. Our present observation of low incidence of joint infection with both ICMV and SLCMV and previous reports [11, 13] indicated that ICMV can co-exist with SLCMV in many plants, indicating the existence of possibly more interacting relationships between the two viruses, other than that reported [32].

Eleven samples produced RCA-RFLP patterns not expected from the known sequences of ICMV or SLCMV. To determine whether they represent begomoviral sequences hitherto not reported from cassava, nucleotide sequences were generated from 31 such fragments. Which also included some samples showing the above unexpected patterns. These represented various portions of the begomoviruses and showed high sequence identities (97–99%) with known ICMV and SLCMV sequences, indicating them to be variants of known cassava-infecting begomoviruses, not representing any novel begomoviruses. The most likely explanation for the presence of a novel pattern could be random point mutations in the viral DNA bringing about changes in the recognition sites for EcoRI.

The low variability of the SLCMV and ICMV DNA mainly manifested as scattered single nucleotide changes. This was evident both in the RCA-derived as well as PCR-derived sequences analyzed. The analysed sequences indicated no incidence of recombination or hot-spots of mutation in the RCA-derived fragments. The sequence analysis also showed the presence of scattered point mutations evenly distributed in the sequenced fragments (data not shown). To determine whether the above scattered point mutations can result in any major changes at the protein level, the derived amino acid sequences of AC2 and AC4 were aligned with those of SLCMV-Attur. Of the 135 amino acid sequence of AC2, amino acid changes were seen only at 5 positions and of the 100 amino acid AC4, changes were found at 11 positions. This reinforces the conclusion that cassava begomoviruses display low levels of scattered point mutations, unlikely to affect any viral protein, In addition, none of the sequences of the cloned RCA fragments showed any homology to any satellite sequences known to be associated with begomoviruses, making their existence very unlikely in CMD-affected cassava.

The lack of variability in cassava begomoviruses in India is in stark contrast to the evidence presented by other workers [16, 33] showing high levels of genetic diversity in begomoviruses isolated from tomato [16] and various satellites [33]. One possible reason behind the low variability of cassava begomoviruses could be linked to the vegetative propagation of cassava practiced by farmers, resulting in accumulation of begomoviruses representing those which are carried forward along with the infected propagules (stem cuttings) in the locations surveyed. The begomoviruses displaying high variability infect annual plants and are transmitted by whiteflies in the field. This raises interesting questions about the role of insect vectors in the generation of variability in plant viruses, an aspect which needs more in-depth study.