Characterization of alphasatellites associated with monopartite begomovirus/betasatellite complexes in Yunnan, China

Background Alphasatellites are single-stranded molecules that are associated with monopartite begomovirus/betasatellite complexes. Results Alphasatellites were identified in begomovirus-infected plant samples in Yunnan, China. All samples that contained alphasatellites also contained betasatellites, but only some samples that contained betasatellites contained alphasatellites. Thirty-three alphasatellites were sequenced, and they ranged from 1360 to 1376 nucleotides. All alphasatellites contain 3 conserved features: a single open reading frame (Rep), a conserved hairpin structure, and an adenine-rich (A-rich) region. On the basis of the phylogenetic tree of the complete nucleotide sequences, the alphasatellites were divided into 3 types with one exception. Type 1 was associated with Tomato yellow leaf curl China virus (TYLCCNV)/Tomato yellow leaf curl China betasatellite (TYLCCNB) complex. Type 2 was associated with Tobacco curly shoot virus (TbCSV)/Tobacco curly shoot betasatellite (TbCSB) complex. Type 3 was associated with TbCSV/Ageratum yellow vein betasatellite (AYVB) complex. Within each type, nucleotide sequence identity ranged from 83.4 to 99.7%, while 63.4-81.3% identity was found between types. Mixed infections of alphasatellites associated with begomovirus/betasatellite complexes were documented. Conclusions Our results validate that alphasatellites are only associated with begomovirus/betasatellite complexes. Thirty-three sequenced alphasatellites isolated from Yunnan Province, China were divided into 3 types--each associated with a specific begomovirus/betasatellite complex. Mix-infections of alphasatellite molecules may not be unusual.


Background
Geminiviruses are a group of plant viruses characterized by their geminate shape and the size of their particles, which encapsidate a circular single-stranded DNA genome. Due to their wide host range and high frequency of genome variation, geminiviruses cause substantial yield losses in many crops, including tomato, cassava, and cotton, throughout tropical and sub-tropical regions worldwide [1,2]. The majority of geminiviruses described belong to the genus Begomovirus in the family Geminiviridae, they are transmitted by the whitefly, Bemisia tabaci [3]. Most begomoviruses have 2 components, which are referred to as DNA-A and DNA-B, both are essential for virus proliferation. Many species only have a single genomic component that resembles DNA-A [1,3]. Some monopartite begomoviruses are associated with betasatellites (formerly DNAβ), which affect the replication of their respective helper begomoviruses and alter the symptoms induced in some host plants [4][5][6][7][8][9]. Analysis of betasatellites reveals that they are approximately half the size of the genomic DNA, and except for a conserved hairpin structure and a TAATATTAC loop sequence, they have little sequence similarity to either the DNA-A or DNA-B molecules of begomoviruses. Betasatellites require begomoviruses for replication, encapsidation, insect transmission, and movement in plants [10].
Alphasatellites (formerly DNA1) are circular, singlestranded DNA molecules associated with begomovirus/ betasatellite complexes [11][12][13][14][15]. Alphasatellites are approximately half the size of begomovirus DNA and encode a rolling-circle replication initiator protein similar to nanoviruses. Consequently, alphasatellites are capable of self-replication in host plants, but require helper begomoviruses for movement in plants as well as insect transmission.

Sequence analysis of alphasatellites
The complete nucleotide sequences of the 23 alphasatellites from tobacco, 3 from tomato, 2 from ageratum, and 5 from malvastrum plants (total: 33) were determined to be 1360 to 1376 nucleotides (nts) in length-this is longer than betasatellites, which range from 1333 to 1355 nts in length. The sequences of these 33 alphasatellites have been submitted to GenBank under the accession numbers AJ579345-AJ579361, AJ888445-AJ888455, and FN678899-FN678903 ( Table 2). The alphasatellites are named according to their sample number; thus, Y35A refers to alphasatellites from sample Y35. Nucleotide sequence comparisons show that the 33 alphasatellites can be divided into 3 types (Table 3). Type 1 consists of 9 samples infected by TYLCCNV and 3 samples infected by TYLCCNV + TYLCTHV; overall nucleotide sequence identity is 83.4-99.7%. Type 2 consists of 5 samples infected by TbCSV, 2 samples infected by TbCSV + TbLCYNV, and 3 samples infected by TbCSV + TYLCCNV; the sequences in type 2 share 91.4-98.2% identity. Type 3 consists of samples mixinfected by TbCSV and other begomoviruses, including 4 samples inflected by TbCSV + TbLCYNV, 2 by TbCSV + MYVYNV, and 2 by TbCSV + MYVV; sequences in type 3 share 90.3-99.6% identity. The overall nucleotide sequence identity between types 1 and 2 is 75.9-81.3%, 63.4-72.0% between types 1 and 3 and 69.3-75.5% between types 2 and 3. Y89A is distinct among the 33 alphasatellites and shares only 69.3-79.5% nucleotide sequence identity with alphasatellites of the 3 types. A relatively lower sequence identity (58.9-71.8%) exists between the present 33 and previously reported alphasatellites (data not shown).

Structural features of alphasatellites
All 33 alphasatellites contain 3 conserved features: a conserved hairpin structure, a single open reading frame, and an adenine-rich (A-rich) region ( Figure 2). The highly conserved structure contains a predicted hairpin structure with a loop that includes the nonanucleotide, TAGTATTAC, which is common to nanoviruses and is similar to the TAATATTAC sequence of geminiviruses. For both geminiviruses and nanoviruses, this sequence contains the origin of replication, and is nicked by Rep to initiate virion-strand DNA replication. Alignment analysis indicates that alphasatellite hairpin structures fall into 5 groups. Groups 1 and 2 contain 10 and 4 alphasatellites, respectively; all alphasatellites in groups 1 and 2 belong to type 1 and share the same loop sequences, but in different stems. Group 3 has 10 alphasatellites which belong to type 2. Alphasatellites in groups 1 and 3 share the same stem sequences excluding one different (G/A) nucleotide in the loop. Group 4 contains only 1 alphasatellite (Y89A), which is distinct from the other 32 owing to its unique stem sequence. Group 5 contains 8 alphasatellites belonging to type 3, which share the same loop sequence with groups 1, 2, and 4, but have a distinct stem (Figure 3). A-rich regions are maintained by all alphasatellites immediately downstream of the Rep gene as reported for other alphasatellites. This A-rich region is approximately 153-169 nts long with an A-content of between 52.3-58.4%. The alignment of the sequences of the A-rich region shows that they can be divided into 3 types in accordance with the phylogenetic trees of the complete nucleotide sequences of the alphasatellites (Figure 4).
All alphasatellites encompass a single large virionsense ORF that has the capacity to encode an approximately 36.6 kDa protein consisting of 315 amino acids, which resembles Rep of nanoviruses. Reps encoded by alphasatellites are highly conserved, with 86.3-100.0% amino acid sequence identities among the 33 alphasatellites (Table 3). Therefore, alphasatellite Rep is more conserved than complete alphasatellite sequences. Amino acid sequence comparisons of Reps also show that the 33 alphasatellites can be divided into 3 main types, which correspond to the 3 types of full-length sequence comparison (Figure 1, right).

Mixed infection of alphasatellites
Mixed infections of geminiviruses were readily found. Some samples, including Y70-Y72, Y87, Y115, Y132, Y137, Y143, Y146, Y216, Y249, Y273, Y276-278, and Y290, were infected by 2 different viruses ( Table 2). In order to determine whether each virus associated with its own alphasatellite molecule, more alphasatellites clones from these samples were sequenced. Sequence analysis revealed that mixed infections of alphasatellites occurred in samples Y87 and Y8, but not in any other samples ( Table 2). Y87 was mix-infected by TbCSV and TYLCCNV, 2 alphasatellites (Y87A-2 and Y87A-7) belonging to types 1 and 2, respectively, were identified. Y8 was infected by TYLCCNV, 2 alphasatellites (Y8A-5 and Y8A-6) belonging to type 1 and sharing 85.0% nucleotide acid identity were identified. Because of the obvious divergence, we assumed that the 2 alphasatellites in Y8 were a consequence of a mixed infection by 2 distinct parental alphasatellites belonging to the same type.

Discussion
Two single-stranded DNA components, alpha-and betasatellites, have been found to be associated with monopartite begomoviruses such as AYVV, CLCuMV, and TbCSV [4,8,11,13,23]. Betasatellites are symptom-modulating satellite molecules that depend on a helper virus for proliferation and movement. On the other hand, alphasatellites are apparently dispensable for symptomatic induction and are capable of autonomous replication [12][13][14]. Our results show that the 33 alphasatellites investigated are all associated with begomovirus/betasatellite complexes, which is a similar result to a report by Briddon [11]. However, only some begomovirus/betasatellite complexes were associated with alphasatellites. A better understanding of the relationship between alphasatellites and begomovirus/betasatellite complexes is achievable if future studies concentrate on the identification of alphasatellites from more symptomatic and asymptomatic crop species as well as diverse, agriculturally unimportant plant species from broader areas.
With the exception of Y89A, comparison of alphasatellites shows that they can be divided into 3 types. Type 1 alphasatellites were identified in samples infected by TYLCCNV/TYLCCNB and TYLCCNV/TYLCTHB + TYLCTHV. Since no alphasatellites were found in samples infected by TYLCTHV/TYLCTHB, it is evident that type 1 alphasatellites are associated with TYLCCNV/ TYLCCNB. All type 2 alphasatellites were identified in samples infected by TbCSV/TbCSB, TbCSV/TbCSB + TbLCYNV, TbCSV/TbCSB + TYLCCNV, and TbCSV + TYLCCNV/TYLCCNB. Because no alphasatellites were found in samples infected by TbLCYNV alone, this suggests that type 2 alphasatellites are associated with TbCSV/TbCSB complexes. It is interesting that alphasatellites in sample Y146, which was mix-infected by TbCSV and TYLCCNV/TYLCCNB, were clustered in type 2 but not type 1. Sample Y146 might have been mix-infected by TYLCCNV/TYLCCNB and TbCSV/ TbCSB in addition to alphasatellites associated with TbCSV/TbCSB, TbCSB then disappeared due to competition between TYLCCNB and TbCSB [24]. Most type 3 alphasatellites were found in samples mix-infected by a combination of TbCSV/AYVB and TbLCYNV or MYVYNV/MYVYNB, while 2 type 3 alphasatellites were mix-infected by TbCSV and MYVV/MYVB. Because no alphasatellites were found in samples infected by TbLCYNV, MYVYNV/MYVYNB, or MYVV/MYVB, it is apparent that type 3 alphasatellites are associated with TbCSV/AYVB. Although AYVCNV/AYVB is responsible for ageratum yellow vein disease in Hainan, China [21], AYVCNV was not found in any ageratum yellow vein disease samples in Yunnan. Instead of AYVCNV/AYVB, TbCSV/AYVB is the causal agents of ageratum yellow vein disease (Zhou et al., unpublished). It is probable that TbCSV acquires the heterogenous betasatellite, AYVB, during mixed infections, but we were unable to determine the origin of type 3 alphasatellites in this study. Sample Y89 was infected by TYLCCNV/ TYLCCNB, therefore, its alphasatellite should belong to type 1. However, sequence comparison shows that Y89A shares only 69.3-79.5% nucleotide sequence identity with other alphasatellites of the 3 types. We speculate that Y89A originated from an unidentified begomovirus/ betasatellite complex.
Mix-infections of begomoviruses are common; 16 of 31 isolates in this study were co-infected by 2 begomoviruses (Table 2). For most isolates, each begomovirus is associated with an alpha-and betasatellite. Two type 1 alphasatellites (Y8A-5 and Y8A-6) were identified in sample Y8, while 2 types of alphasatellites (Y87A-7 and Y87A-2) were identified in sample Y87. Our results indicate that mix-infections of alphasatellite molecules may not be unusual.
The origin of alphasatellites is undoubtedly related to nanoviruses. Presently, the function of alphasatellites is not clear, but it is evident that alphasatellites functionally interact with geminivirus/betasatellite complexes resulting in symptom alteration and a reduction in the level of viral DNA and betasatellites [12][13][14]25,26] Available evidence suggests that the ubiquitous association of alphasatellites with begomovirus/betasatellite complexes indicates that alphasatellites may play an important role in the occurrence, diffusion, and epidemiology of begomovirus/betasatellite complexes. More studies are required to elucidate the specific role that alphasatellites play in disease development, virus life cycle, and the evolution of begomoviruses/betasatellite complexes.

Conclusions
Seven viruses, including TbCSV, TbLCYNV, TYLCCNV, TYLCTHV, MYVV, MYVYNV, and SLCYNV, were characterized in Yunnan Province-some of them are associated with betasatellites. Our results show that all samples from Yunnan that contained alphasatellites also had betasatellites. However, only some samples that contained betasatellites had alphasatellites. Thirty-three sequenced alphasatellites were divided into 3 types-each associated with a specific begomovirus/betasatellite complex. Type 1 was associated with TYLCCNV/TYLCCNB; type 2 was associated with TbCSV/TbCSB; and type 3 was associated with TbCSV/AYVB. Alphasatellites have 3 highly conserved structure features: a conserved hairpin structure, a single open reading frame, and an Arich region. The alignment of the sequences of the conserved hairpin structure and the A-rich region shows that the alphasatellites can be further divided into 3 types in accordance with the phylogenetic trees of their complete nucleotide sequences. Reps encoded by the 33 alphasatellites are highly conserved and share more than 86.3% amino acid sequence identity. Alphasatellites may play an important role in the epidemiology of begomovirus/betasatellite complexes.

Virus sources and DNA extraction
Young seedlings were collected from naturally infected tobacco, tomato, Ageratum conyzoides, Malvastrum coromandelianum, and squash plants showing begomovirus-like infection symptoms, from locations separated by 700 km in Yunnan Province, China from 1999 to 2004. Viral DNA from the samples was extracted as previously described [20].

Sequence analysis
Sequence data were assembled and analyzed using DNAStar software version 6.0 (DNAStar Inc., Madison, WI, USA) and MEGA version 4 [30]. Sequence alignments were performed using the CLUSTAL V Multiple Sequence Alignment program in DNAStar, and phylogenetic trees were conducted using the neighbor-joining method using MEGA version 4. Other alphasatellite sequences used for comparisons were alphasatellites of