Begomoviruses (genus Begomovirus, family Geminiviridae) are a group of plant viruses transmitted by the whitefly Bemisia tabaci (Aleyrodidae) to a large variety of cultivated and uncultivated plant species. They have a circular single-stranded DNA genome encapsidated in twinned icosahedral particles (for review see ref. [9]).

Previous studies have highlighted the diversity of begomoviruses infecting crops in Madagascar. Members of six species of cassava-infecting bipartite begomoviruses (cassava mosaic geminiviruses, CMGs) have been found in Madagascar (Harimalala et al. [5]; De Bruyn et al., submitted), one of them, tentatively called “cassava mosaic Madagascar virus” (CMMGV; [4]), represents a proposed new species. In the past decade, monopartite begomoviruses from tomato crops in Madagascar belonging to at least six established or tentative species have been characterized. These include Tomato leaf curl Madagascar virus [2], Tomato leaf curl Diana virus [7], Tomato leaf curl Namakely virus [7], Tomato leaf curl Toliara virus [7], and the proposed species “Tomato leaf curl Ambilobe virus” (unpublished) and “Tomato leaf curl Antsiranana virus” [7]. Finally, a novel begomovirus infecting bean has been characterized and tentatively named “Bean leaf curl Madagascar virus” [7].

While African begomoviruses infecting crops of New World origin are most likely descendants of indigenous viruses that probably co-evolved with African native plants [6], knowledge of begomovirus diversity is still largely restricted to cultivated host plants. Here, we describe for the first time, the complete nucleotide sequences of DNA-A and DNA-B of a new bipartite begomovirus isolated in Madagascar from the wild plant Asystasia gangetica (Acanthaceae) and tentatively named asystasia mosaic Madagascar virus (AMMGV). Leaf samples from A. gangetica plants showing bright yellow mosaic symptoms (Supplementary Figure 1) were collected in the northern region (sample MG493; Mangabe; 13.7120 S, 48.4256 E; July 2011) and in the south-eastern region (sample MG147; Beretry; 21.7224 S, 48.0006 E; May 2009) of Madagascar and preserved by dehydration using anhydrous calcium chloride. Total DNA was extracted using a DNeasy Plant Mini Kit (QIAGEN, USA). Viral DNA was amplified by rolling-circle amplification using Phi29 DNA polymerase (TempliPhi, GE Healthcare, USA). Amplified products were digested with BamHI or NcoI endonuclease. The monomeric full-length genomes obtained (~2.8 kb) were purified and cloned into the pGEM-3Zf and pGEM-T vectors, respectively, for BamHI and NcoI-digested fragments. Cloned DNA was sequenced at Macrogen Inc. (Europe) by primer walking; contigs were assembled using Geneious software (Biomatters, New Zealand).

DNA-A and -B component sequences recovered from sample MG493 displayed features typical of Old World bipartite begomoviruses, while only DNA-A was successfully isolated from sample MG147. The DNA-A components were found to be 2737/2738 nucleotides in length and shared 95.1 % nucleotide sequence identity. These DNA-A components encoded six open reading frames (ORFs): AV1 (254 aa; 95.3 % aa sequence identity) and AV2 (118 aa; 97.5 % aa sequence identity) in the viral sense and AC1 (358 aa; 98.6 % aa sequence identity), AC2 (135 aa; 95.6 % aa sequence identity), AC3 (134 aa; 95.5 % aa sequence identity) and AC4 (65 aa; 89.2 % aa sequence identity) in the complementary sense. The DNA-B component of MG493 has two ORFs: BV1 (254 aa) in the viral sense and BC1 (323 aa) in the complementary sense. Both DNA components isolated from sample MG493 shared a common region (CR) of approximately 150 and 200 nucleotides in length for DNA-A and DNA-B, respectively, with only 72.8 % nucleotide sequence identity. However, the iteron GGTGG was present in the CR of both components and was repeated three times, and the corresponding iteron-related domain of the Rep protein (IRD; Rep N-terminal domain interacting with ori-associated iterons [1]) was identified as MPRAGRFNIN (Fig. 1).

Fig. 1
figure 1

A. Alignment of the common region (CR) of DNA-A and DNA-B sequences of AMMGV isolates. The three repeats of the GGTGG iteron, the TATA-box and the stem-loop are indicated by a black frame. B. Alignment of the N-terminal portion of the Rep protein. The iteron-related domain (IRD) is indicated by a black frame

DNA-A and -B of AMMGV were only distantly related to known begomoviruses and shared highest nucleotide sequence identity of 72.9 % and 66.9 % with a recently isolated bipartite begomovirus from Asystasia sp. in West Africa (JF694484, asystasia begomovirus 1 DNA-A; JF694485, asystasia begomovirus 1 isolate ABgV1 segment DNA-B). Maximum-likelihood phylogenetic trees based on the complete nucleotide sequences of DNA-A and DNA-B of AMMGV isolates and representatives of the genus Begomovirus were constructed using PHYML v3.0 [3] (Fig. 2). The ML trees showed the grouping of AMMGV with asystasia begomovirus 1, in a distinct clade only distantly related to other begomoviruses. Analyses of recombination using RDP4 [8] revealed no evidence of a recombination event in any of the genomic components.

Fig. 2
figure 2

Maximum-likelihood (ML) trees inferred from alignments of complete DNA-A (A) and DNA-B (B) sequences of AMMGV and representative begomoviruses. Black circles indicates nodes supported by a SH value above 95 %. For begomovirus acronyms, see Supplementary Table 1

Our results show that the bipartite genome of AMMGV is representative of a distinct clade in the phylogeny of African begomoviruses. The grouping of this particular virus with a recently characterized virus infecting the same host in West Africa suggests that these viruses belong to a novel group of bipartite begomoviruses. Except for the cassava mosaic geminiviruses (CMGs) described on cassava, these two viruses are amongst a very few Old World bipartite begomoviruses described in sub-Saharan Africa and SWIO islands.

FormalPara GenBank accession numbers

KP663483, KP663484, KP663485.