A DNAβ Associated with Tomato Yellow Leaf Curl China Virus Is Required for Symptom Induction

Abstract

We report here that all 25 isolates of Tomato yellow leaf curl China virus (TYLCCNV) collected from tobacco, tomato, or Siegesbeckia orientalis plants in different regions of Yunnan Province, China, were associated with DNAβ molecules. To investigate the biological role of DNAβ, full-length infectious clones of viral DNA and DNAβ of TYLCCNV isolate Y10 (TYLCCNV-Y10) were agroinoculated into Nicotiana benthamiana, Nicotiana glutinosa, Nicotiana. tabacum Samsun (NN or nn), tomato, and petunia plants. We found that TYLCCNV-Y10 alone could systemically infect these plants, but no symptoms were induced. TYLCCNV-Y10 DNAβ was required, in addition to TYLCCNV-Y10, for induction of leaf curl disease in these hosts. Similar to TYLCCNV-Y10, DNAβ of TYLCCNV isolate Y64 was also found to be required for induction of typical leaf curl diseases in the hosts tested. When the βC1 gene of TYLCCNV-Y10 DNAβ was mutated, the mutants failed to induce leaf curl symptoms in N. benthamiana when coinoculated with TYLCCNV-Y10. However, Southern blot hybridization analyses showed that the mutated DNAβ molecules were replicated. When N. benthamiana and N. tabacum plants were transformed with a construct containing the βC1 gene under the control of the Cauliflower mosaic virus 35S promoter, many transgenic plants developed leaf curl symptoms similar to those caused by a virus, the severity of which paralleled the level of βC1 transcripts, while transgenic plants transformed with the βC1 gene containing a stop codon after the start codon remained symptomless. Thus, expression of a βC1 gene is adequate for induction of symptoms of viral infection in the absence of virus.

Whitefly-transmitted geminiviruses in the genus Begomovirus have genomes typically consisting of two molecules of circular single-stranded DNA, referred to as DNA-A and DNA-B, each of about 2.8 kb. However, a few Begomovirus species have only a single genomic component, which resembles DNA-A of the bipartite begomoviruses. In recent years, a novel satellite DNA, referred to as DNAβ, has been found associated with monopartite begomoviruses, including Ageratum yellow vein virus, Bhendi yellow vein mosaic virus, Cotton leaf curl Multan virus, and Eupatorium yellow vein virus. These DNAβ molecules are required for induction of typical disease symptoms in Ageratum, bhendi, cotton, and Eupatorium plants, respectively (3, 11, 22, 23).

DNAβ is a circular single-stranded DNA of approximately 1,350 nucleotides. Several putative genes have been noted on the virion-sense and complementary-sense strands of DNAβ. Only the βC1 gene, located on the complementary-sense strand of all DNAβ species, is conserved in position and size, but the function of the βC1 gene is unknown (34). DNAβ depends on the helper begomovirus for its replication, encapsidation, insect transmission, and movement in plants; it has little sequence similarity to either DNA-A or DNA-B molecules of begomoviruses except for a conserved hairpin structure that includes a TAATATTAC loop sequence (34).

In China, DNAβ molecules were associated with begomoviruses obtained from tobacco, tomato, and weeds (Malvastrum coromandelianum and Siegesbeckia orientalis). Comparison of these DNAβ sequences suggested that they coevolved with their cognate helper viruses (31, 34). In the present work, we show that the DNAβ associated with Tomato yellow leaf curl China virus (TYLCCNV) is required for induction of leaf curl disease in tobacco, tomato, and petunia and that the βC1 gene of DNAβ is necessary for symptom induction but not DNAβ replication. We also find that transgenic Nicotiana plants expressing the βC1 gene display symptoms of viral infection.

MATERIALS AND METHODS

Virus sources and DNA extraction.

Between 1999 and 2002, 93 naturally infected plants of tobacco, tomato, and Siegesbeckia orientalis showing begomoviral symptoms were collected at farms in Yunnan Province, China, at sites separated by up to 700 km. Viral DNA was extracted from these plants as described previously (31).

Detection of TYLCCNV and associated DNAβ.

Based on the available complete nucleotide sequences of TYLCCNV isolates and their associated DNAβ (30, 34), primer pairs specific for TYLCCNV (TYLCCNV/SF and TYLCCNV/SR) and DNAβ (TYLCCNV/Beta and Beta02) were designed. PCR was done as described previously (33). Primer sequences are shown in Table 1.

TABLE 1.

Sequences of primers used for this study

Purpose	Primer	Sequence (5′-3′)a	Position in TYLCCNV-Y10 or DNA β (nucleotides)
Specific detection of TYLCCNV and DNAβ
	TYLCCNV/SF	CGTCACCAGAAGACAAATGT	398-417
	TYLCCNV/SR	ATGCG/AA/GT/AC/ATTCG/ATCACCCTC	1604-1585
	TYLCCNV/Beta	CGGCATTATTTTGAGGCAGT	345-364
	Beta02	GGTACCTACCCTCCCAGGGGTACAC	1277-1253
Construction of infectious clone of TYLCCNV
	Y10 Full-E/F	GAATTCTTTAAAGTGCTTTAG	1771-1791
	Y10 Full-E/R	GAATTCATGGGGGCCCAAAGG	1776-1756
	Y10 PL/F	CGTCGACACCTGTTTGGGGATATGAGAT	2343-2363
	Y64PL-Sal/F	CCGTCGACTGAATGTTCGGATGGAAATGTG	2319-2340b
	Y64FL-Xba/F	TCTAGATACTCTTTATAGCTGG	1647-1668b
	Y64FL-Xba/R	TCTAGAAGAGGAAAAGAATTCC	1652-1631b
Construction of βC1 mutants and plant expression vectors
	BetaCM-F/F	GTGGATGATAGATGAGTCTGTT	578-599
	BetaCM-F/R	AACAGACTCATCTATCATCCAC	599-578
	BetaCM-S/F	GTTGTATTTGATTCACATGTTTATTTG	546-575
	BetaCM-S/R	CAAATAAACATGTGAATCAAATACAAC	575-546
	BetaCM-B/F	GTGATGTTTATTTGTTGTGGATGATAGAT	562-590
	BetaCM-B/R	ATCTATCATCCACAACAAATAAACATCAC	590-562
	BetaCM-T/F	ACTCCAAACCCTACATGTTGTTGT	530-553
	BetaCM-T/R	ACAACAACATGTAGGGTTTGGAGT	553-530
	Beta01	GGTACCACTACGCTACGCAGCAGCC	1272-1296
	Beta05	GAAACCACTACGCTACGCAGCAGCC	1276-1296
	C1F	ACGGATCCATGTATCATCCACAAC	575-590
	C1R	CCGTCGACTCATACATCTGAATTTG	210-226

^aB is C, T, or G; K is G or T; R is A or G; S is C or G; V is A, C, or G; W is A or T; Y is C or T.

^bPosition in TYLCCNV-Y64 (AJ457823).

Construction of infectious clones of TYLCCNV and its associated DNAβ.

Based on the complete nucleotide sequence of TYLCCNV-Y10 (accession no. AJ319675), primers Y10PL/F (a SalI site was introduced) and Y10 Full-E/R (containing an EcoRI site) were designated to amplify a 2.0-kb fragment, and the PCR product was inserted into the pGEM-T vector (Promega) to produce pGEM-0.7A. pGEM-0.7A was digested with SalI and EcoRI and the 2.0-kb fragment was introduced into the binary vector pBinPLUS (28) to produce pBinPLUS-0.7A. The complete sequence of TYLCCNV-Y10 was amplified by two specific primers, Y10 Full-E/F (containing an EcoRI site) and Y10 Full-E/R, and the 2.7-kb PCR product was cloned into pGEM-T (Promega) to produce pGEM-AFL. Clone pGEM-AFL was digested with EcoRI to obtain a full-length unit of Y10 and inserted into the unique EcoRI site of pBinPLUS-0.7A to produce pBinPLUS-1.7A, containing a 1.7-mer tandem repeat of TYLCCNV-Y10.

The infectious clone of TYLCCNV-Y64 (pBinPLUS-Y64-1.7A) was constructed with a similar strategy except that three different primers, Y64PL-Sal/F (a SalI site was introduced) and Y64FL-Xba/F and Y64FL-Xba/R (each containing an XbaI site at the 5′ end), were used to obtain a 2.0-kb fragment and the complete genome. The infectious clones of TYLCCNV-Y10 DNAβ (pBinPLUS-2β) and TYLCCNV-Y64 DNAβ (pBinPLUS-Y64-2β) were produced as described (34). Primer sequences are shown in Table 1. All the amplified fragments used for construction of infectious clones were sequenced entirely with the automated model 377 DNA sequencing system (Perkin Elmer Inc.), and sequence analysis indicated that no mutation was introduced into the clones by PCR.

Construction of mutants of the βC1 gene of TYLCCNV DNAβ and plant expression vectors.

Mutations were introduced into the βC1 gene of TYLCCNV-Y10 DNAβ to study the function of this open reading frame (ORF). The mutation strategy is detailed in Fig. 1. The βC1 gene of TYLCCNV-Y10 DNAβ has three possible start codons (ATG). Site-directed mutagenesis was performed to alter the first or second ATG, both the first and second ATG, or the third ATG alone, with complementary primer pairs BetaCM-F/F and BetaCM-F/R, BetaCM-S/F and BetaCM-S/R, BetaCM-B/F and BetaCM-B/R, and BetaCM-T/F and BetaCM-T/R, respectively. Two independent PCRs were performed with each of the above primer pairs and with primers Beta05 and Beta02, with the previously obtained clone pGEMβ as the template (34).

FIG. 1.

Mutational strategy for the βC1 gene of TYLCCNV-Y10 DNAβ. The three possible start codons are underlined. Mutation sites are in italic.

The PCR products were fused and amplified with overlap extension PCR as described previously (26), and the overlapping PCR products were inserted into the pGEM-T-Easy vector to produce clones pGEMC1M-Fβ, pGEMC1M-Sβ, pGEMC1M-Bβ, and pGEMC1M-Tβ, respectively. The fidelity of the mutants was confirmed by sequencing. The strategy described previously (34) was then used for construction the tandem repeats of βC1 gene mutants to produce pBinPLUS-C1M-Fβ, pBinPLUS-C1M-Sβ, pBinPLUS-C1M-Bβ, and pBinPLUS-C1M-Tβ, respectively. The clones were mobilized into Agrobacterium tumefaciens strain EHA105 by triparental mating (34).

The βC1 gene (381 nucleotides) was PCR amplified from plasmid pGEMβ containing TYLCCNV-Y10 DNAβ (34) with the complementary-sense primer C1F containing a BamHI site and virion-sense primer C1R containing a SalI site. After digestion with BamHI and SalI, the PCR fragment was cloned between a duplicated Cauliflower mosaic virus 35S promoter and the nopaline synthase terminator (nos) in the expression vector pBin438 (14) to produce pBin-C1. For obtaining a mutated βC1 gene construct, the plasmid pGEMC1M-Tβ was used as the template for PCR with primer pair C1F and C1R. The resulting DNA fragment was cloned into pBin438 as described above to produce pBin-mC1. The fidelity of the wild-type βC1 gene and of its mutated version in the expression vectors was confirmed by sequencing. Primer sequences are shown in Table 1.

Agroinoculation of plants.

A. tumefaciens cultures were grown at 28°C for 48 h (optical density at 550 nm = 1), after which a fine needle was used to inject 0.2 ml of culture into the stems or petioles of plants at the six-leaf stage. Nicotiana benthamiana, Nicotiana glutinosa, Nicotiana tabacum Samsun NN, Nicotiana tabacum Samsun nn, Lycopersicon esculentum, and Petunia hybrida plants were agroinoculated either with pBinPLUS-1.7A alone or together with pBinPLUS-2β. N. benthamiana, N. glutinosa, and L. esculentum plants were also agroinoculated with pBinPLUS-Y64-1.7A either alone or together with pBinPLUS-Y64-2β. Other N. benthamiana plants were agroinoculated with each of the βC1 gene mutants together with pBinPLUS-1.7A. Inoculated plants were grown in an insect-free cabinet with supplementary lighting corresponding to a 16-h day length.

Analysis of viral DNA replication in leaf disks.

The ability of cloned DNA to replicate after agroinoculation was also tested in N. benthamiana leaf disks as described (12). Leaf disks 6 mm in diameter were incubated on precallusing medium plates for 24 h at 25°C under continuous lighting, then dipped into an overnight culture of the appropriate transconjugant, and returned to the plates for a further 48 h. The leaf disks were next transferred to selective medium containing 100 μg of kanamycin per ml and 500 μg of carbenicillin per ml (Sigma), incubated at 25°C for 2, 4, or 6 days, and then used for isolation of viral DNA. DNA molecules were detected by Southern blotting (see below).

Viral DNA detection.

Nucleic acids were isolated from young leaves of tobacco plants as previously described (35), fractionated by 1% agarose gel electrophoresis in TBE buffer (90 mM Tris-borate, 2 mM EDTA, pH 8.3), and then transferred to Hybond-N⁺ membranes (Amersham Pharmacia, Buckinghamshire, England). Following alkali denaturation and neutralization, membranes were hybridized with randomly labeled probes (7). Hybridization signals were detected by phosphorimaging with a Typhoon 9200 imager (Amersham Pharmacia). All the leaves were collected 20 days after inoculation.

Serological tests.

Triple antibody sandwich enzyme-linked immunosorbent assay (ELISA) was done as described (8). Monoclonal antibody SCR18, which was kindly provided by B. D. Harrison, Scottish Crop Research Institute, United Kingdom, was used to detect TYLCCNV-Y10.

Production of transgenic tobacco plants.

The expression vectors were introduced into A. tumefaciens EHA105 by triparental mating. The derived A. tumefaciens strains were then used to transform leaf explants of N. benthamiana and N. tabacum as described (18). Selection for transformation was done on medium containing kanamycin (200 μg/ml). Kanamycin-resistant shootlets were collected, placed on rooting medium, grown to a height of 5 to 6 cm, and transferred to soil.

Analysis of DNA and RNA from transgenic Nicotiana plants.

The presence of the wild-type and mutant βC1 transgenes in N. benthamiana and N. tabacum was determined by PCR and Southern blot analysis. Tobacco genomic DNA was extracted by the cetyltrimethylammonium bromide (CTAB) method (17) and used as the template for PCR analysis with primer pair C1F and C1R. Healthy Nicotiana plants were used as negative controls. PCR products from each sample were fractionated on 1.2% agarose gels, transferred to a Hybond N⁺ membrane, and hybridized with a specific probe. The DNA probe was labeled with denatured PCR products of the βC1 gene with [α-³²P]dCTP by the random priming method according to the manufacturer's instructions (Promega).

For Northern blot analysis, total RNA was extracted from leaf tissues with Trizol reagent following the manufacturer's protocol (Invitrogen, Carlsbad, Calif.). RNA electrophoresis and gel blotting were performed as described (21). Equal amounts of total RNA (40 μg) were subjected to 1.2% agarose gel electrophoresis under denaturing conditions and subsequently transferred to Hybond N⁺ membranes. The blots were hybridized at 42°C with a radiolabeled DNA probe as described above.

RESULTS

Association of DNAβ with TYLCCNV isolates.

Nineteen DNA samples from tobacco, five from tomato, and one from Siegesbeckia orientalis were found to be infected with TYLCCNV when tested with primers specific for TYLCCNV, while the other 59 tobacco and nine tomato samples were found to be infected with other begomoviruses. The complete sequences were determined for seven tobacco, one tomato, and one Siegesbeckia sample of TYLCCNV (accession numbers AJ319674 to AJ319677, AJ420316 to AJ420317, AJ457823, AJ457985, and AJ781302), and sequence analysis showed that they shared more than 89% nucleotide sequence identity, further indicating that they were infected by TYLCCNV. When these samples were tested by PCR with primers specific for TYLCCNV DNAβ, a band of approximately 1.3 kb was consistently amplified from all samples. The results show that all the samples of TYLCCNV collected in Yunnan Province were associated with DNAβ molecules.

Infectivity of TYLCCNV and its associated DNAβ.

To investigate the biological role of DNAβ, the infectious clone pBinPLUS-1.7A containing a 1.7-mer TYLCCNV-Y10 and the clone pBinPLUS-2β containing a dimer of TYLCCNV-Y10 DNAβ were constructed in the binary vector pBinPLUS, and these clones were found to be infectious in N. benthamiana plants, producing leaf curl symptoms. pBinPLUS-1.7A was then agroinoculated either alone or together with pBinPLUS-2β to plants. Triple antibody sandwich ELISA and PCR showed that clone pBinPLUS-A-1.7 was infectious in N. benthamiana, N. glutinosa, N. tabacum Samsun (NN or nn), L. esculentum, and P. hybrida, but no symptoms developed in these plants. When seedlings were coagroinoculated with pBinPLUS-1.7A and pBinPLUS-2β, these plants developed typical symptoms, including downward leaf curling, vein swelling, stunting, and enations (Table 2). Agroinoculation of DNAβ alone did not produce a local or systemic infection (data not shown). Similarly, systemic infection of plants with TYLCCNV-Y64 induced no symptoms, and the phenotype changed from symptomless to leaf curl and vein darkening when the plants were coinoculated with TYLCCNV-Y64 DNAβ (Table 2).

TABLE 2.

Infectivity and symptoms induced by TYLCCNV with and without DNAβ

Isolate	Plant species	No. of plants with symptoms/no. inoculated
		TYLCCNV			TYLCCNV and DNAβ
		Expt 1	Expt 2	Expt 3	Expt 1	Expt 2	Expt 3	Typical symptoms
Y10	N. benthamiana	0/8	0/16	0/8	12/12	24/24	8/8	Severe downward leaf curling, stunting, leaf crumpling, vein darkening, yellow vein, curly shoot, and enations
	N. glutinosa	0/8	0/6	0/10	10/12	16/20	13/16	Severe downward leaf curling, vein darkening, stunting, and enations
	N. tabacum Samsun NN	0/12	0/6	0/8	13/22	10/15	6/10	Downward leaf curling, vein darkening, stunting, and enations
	L. esculentum	0/16	0/16	0/12	2/10	4/18	3/16	Downward leaf curling, some upward leaf curling, vein darkening, stunting, and enations
	P. hybrida	0/6	0/8	0/8	6/6	8/8	8/8	Downward leaf curling, leaf chlorosis
Y64	N. benthamiana	0/7	0/9		5/5	7/7		Downward leaf curling and vein darkening
	N. glutinosa	0/8	0/6		6/6	6/7		Downward leaf curling and vein darkening
	L. esculentum	0/5	0/8		2/8	2/6		Downward leaf curling and vein darkening

Interaction between TYLCCNV and its associated DNAβ.

Southern blot hybridization analysis of N. benthamiana plants coinfected with TYLCCNV-Y10 and DNAβ showed that both viral DNA and DNAβ were present (Fig. 2, lanes 3 and 4). N. benthamiana plants coinfected with TYLCCNV-Y10 and DNAβ accumulated more DNA than those infected with TYLCCNV-Y10 alone (Fig. 2), whereas DNAβ was not detected in inoculated or noninoculated tissues of N. benthamiana plants inoculated with TYLCCNV-Y10 DNAβ alone (data not shown), indicating that TYLCCNV-Y10 is required for replication of DNAβ in plants. Both TYLCCNV-Y10 and DNAβ were detected in leaf curl-affected tobacco plants that had been infected via whiteflies (Bemisia tabaci) that had fed on symptomatic plants agroinoculated with cloned TYLCCNV-Y10 and DNAβ (Fig. 2, lane 5).

FIG. 2.

Detection of viral DNAs in agroinoculated N. benthamiana plants. Total nucleic acids (5 μg) were extracted 20 days postinoculation from individual plants agroinoculated with clones of TYLCCNV-Y10 (lanes 1 and 2) or TYLCCNV-Y10 plus DNAβ (lanes 3 and 4). Lane 5 represents nucleic acids (15 μg) extracted from leaf curl-affected leaves of an N. tabacum Samsun NN plant infected with Bemisia tabaci. Total nucleic acids were fractionated in agarose gels, blots were probed with the CP gene sequence of TYLCCNV-Y10 (top) or the full-length sequence of TYLCCNV-Y10 DNAβ (bottom). The positions of single-stranded (ssDNA) and subgenomic (sgDNA) forms of TYLCCNV-Y10 and DNAβ are indicated.

The replication of TYLCCNV-Y10 and DNAβ was also investigated with the N. benthamiana leaf disk assay. TYLCCNV-Y10 was detected at 2 days postinoculation both in leaf disks coinoculated with TYLCCNV-Y10 and DNAβ and in those inoculated with TYLCCNV-Y10 alone. DNAβ did not replicate autonomously but was replicated in trans by TYLCCNV-Y10. DNAβ was barely detectable in leaf disks inoculated with TYLCCNV-Y10 and DNAβ at 2 days postinoculation but had increased greatly by 6 days postinoculation (data not shown).

The 65-day time course of accumulation in N. benthamiana of TYLCCNV-Y10 with or without DNAβ was also compared by triple antibody sandwich ELISA (Fig. 3). In plants coinfected with TYLCCNV-Y10 and DNAβ, a low level of virus was detected at 6 days postinoculation. The virus concentration reached its highest level at about 9 to 50 days postinoculation and then decreased. In plants infected with TYLCCNV-Y10 alone, the virus became detectable later than in plants coinfected with TYLCCNV-Y10 and DNAβ and reached its highest level at 25 to 35 days postinoculation. However, it accumulated in somewhat smaller amounts, which later decreased more rapidly than in the presence of DNAβ (Fig. 3). It is striking that plants with a symptomless infection have such high concentrations of virus, indicating that essential pathogenicity determinants may not be encoded by viral DNA but by DNAβ.

FIG. 3.

Time course of accumulation of TYLCCNV-Y10 in N. benthamiana coinfected with or lacking DNAβ. The youngest infected leaves (about 200 mg) were collected at intervals from 6 to 65 days postinoculation (p.i.). Triple antibody sandwich ELISA was used to estimate viral content (A represents A₄₀₅). The error bars indicate the standard deviation of each sample. A, TYLCCNV-Y10, A+β, TYLCCNV-Y10 and DNAβ.

Function of the βC1 gene of DNAβ in symptom induction and DNA replication.

TYLCCNV-Y10 DNAβ has three possible start codons (ATG) for expression of the βC1 gene. When the four mutated clones pBinPLUS-C1M-Fβ, pBinPLUS-C1M-Sβ, pBinPLUS-C1M-Bβ, and pBinPLUS-C1M-Tβ, which represent alterations of the first ATG, the second ATG, both the first and second ATG, or the third ATG alone, respectively, were coinoculated with pBinPLUS-1.7A, only pBinPLUS-C1M-Fβ (first ATG mutated) induced leaf curl symptoms in N. benthamiana (Fig. 4A). However, in these plants leaf curling was not accompanied by the usual stunting. All the DNAβ mutants were replicated by TYLCCNV-Y10 (Fig. 4B), and sequence analysis confirmed that each mutation was maintained in systemically infected leaves. These results indicate that the βC1 protein is not required for replication of DNAβ. The results also show that the second ATG, which is conserved in all previously reported species of DNAβ, is the most important start codon for expression of the βC1 gene but suggest that the first start codon may also affect its expression.

FIG. 4.

(A) Symptoms in N. benthamiana plants induced by pBinPLUS-A-1.7 coinoculated with different βC1 mutants. From left to right, plants were coinoculated with pBinPLUS-2β, pBinPLUS-C1M-Fβ, pBinPLUS-C1M-Sβ, pBinPLUS-C1M-Bβ, and pBinPLUS-C1M-Tβ, respectively. The wild-type DNAβ induced severe downward leaf curling, stunting, vein swelling and darkening, and enations. When the first ATG of the βC1 gene was mutated, plants developed leaf curl symptoms but did not become stunted. Plants inoculated with any of the other βC1 mutants remained symptomless. (B) Southern blot analysis of plants inoculated with βC1 mutants of TYLCCNV-Y10 DNAβ. Samples (10 μg) of total nucleic acids were extracted from individual N. benthamiana plants infected by agroinoculation with TYLCCNV-Y10 (lanes 1 and 2), TYLCCNV-Y10 and DNAβ (lanes 3 and 4), TYLCCNV-Y10 and DNAβ-C1MF (lanes 5 and 6), TYLCCNV-Y10 and DNAβ-C1MS (lanes 7 and 8), TYLCCNV-Y10 and DNAβ-C1MB (lanes 9 and 10), and TYLCCNV-Y10 and DNAβ-C1MT (lanes 11 and 12). The blots were probed with the full-length sequence of TYLCCNV-Y10 DNAβ. ssDNA, single-stranded DNA; scDNA, supercoiled DNA.

Phenotypes of R₀ transgenic tobacco plants expressing the βC1 gene.

The above results show that mutation of the βC1 gene abolished disease symptoms in N. benthamiana coinoculated with cloned TYLCCNV. The βC1 gene therefore may play an important role in symptom induction. To test this possibility, N. benthamiana and N. tabacum were transformed with A. tumefaciens containing construct pBin-C1. During the in vitro regeneration of transgenic plants, it was observed that many shoots of N. benthamiana were severely deformed, while some shoots of N. tabacum also developed abnormally. Of the rooted plantlets, about 40% of transgenic N. benthamiana and 50% of transgenic N. tabacum plants were abnormal.

All rooted plantlets were transferred to soil. As summarized in Table 3, 75 separate lines of transgenic N. benthamiana and 39 of N. tabacum were obtained. Most abnormal plantlets of both species continued to display abnormal phenotypes, including leaf distortion, upward leaf curling, and blistering of leaves (Fig. 5). In addition, abnormal phenotypes such as interveinal protuberances or small interveinal tissue outgrowths could be observed on the undersides of some leaves of transgenic N. tabacum (Fig. 5I and 5J). Four phenotypes of transgenic Nicotiana plants were distinguished: severely abnormal, moderately abnormal, “recovered,” and symptomless. Severely abnormal means that leaf distortion or severe curling developed in N. benthamiana (10 of 75 plants) (Fig. 5A and 5D) or leaf distortion, severe curling, or protuberance of leaves developed in N. tabacum (10 of 39 plants) (Fig. 5F and 5G). Moderately abnormal N. benthamiana plants (12 of 75 plants) had curled leaves (Fig. 5B), whereas those of N. tabacum (5 of 39 plants) had curled leaves and numerous minute outgrowths on the leaf undersides (Fig. 5H). Some transgenic plants of both species exhibited abnormal phenotypes such as leaf curling or small outgrowths on the undersides of leaves when they were young but then produced normal systemic leaves after several weeks of growth in soil and were assigned to the recovered phenotype (Fig. 5C). In contrast, many transgenic plants developed normally and remained symptomless (Table 3).

TABLE 3.

Phenotypes of R₀ transgenic plants expressing the βC1 gene of TYLCCNV-Y10 DNAβ

Transgenic species	No. of lines
	Total	Severely abnormal	Moderately abnormal	Recovered	Symptomless
N. benthamiana	75	10	12	5	48
N. tabacum	39	10	5	4	20

FIG. 5.

Phenotypes of transgenic N. benthamiana and N. tabacum expressing the βC1 gene of TYLCCNV-Y10 DNAβ. (A) Example of a severely abnormal plant. (B) Example of a moderately abnormal plant. (C) Example of a plant showing the recovered phenotype. The white arrow marks a curled leaf. (D) Distorted leaf from a severely abnormal plant. (E) Upward curling and blistering of leaves from a moderately abnormal plant. Left, top side; right, underside. (F) Examples of severely abnormal plants. (G) Distorted leaf from a severely abnormal plant. Left, top side; right, underside. (H) Curled leaf from a moderately abnormal plant. Left, top side; right, underside. (I) Interveinal protuberances and small interveinal tissue outgrowths on the underside of a leaf. (J) Enlarged view of a protuberance on the leaf underside.

To guard against the unlikely possibility that the abnormal phenotypes were caused by virus contamination, leaf extracts from the transgenic tobacco plants were inoculated onto nontransgenic healthy tobacco and examined by electron microscopy. No evidence of virus infection was obtained. Transgenic plants were also tested for TYLCCNV infection by triple antibody sandwich ELISA and PCR, but no infection was detected (data not shown).

To preclude the possibility that the transformation procedure had induced the symptoms of viral infection, the vector pBin-mC1, which contains the mutated βC1 gene of TYLCCNV-Y10 DNAβ, was used in parallel with pBin-C1 to produce additional sets of transgenic plants with the same transformation method. The pBin-C1 construct again produced abnormal phenotypes of transgenic N. benthamiana and N. tabacum, whereas no transgenic N. benthamiana and N. tabacum plants transformed with the pBin-mC1 construct developed any abnormality (data not shown).

Analysis of DNA and RNA from transgenic plants.

The presence of the βC1 transgene was verified by PCR amplification from genomic DNA of R₀ transgenic plants of N. benthamiana and N. tabacum. The βC1 transgene sequences were amplified from all transgenic plants but not from nontransgenic healthy plants. Sequence analysis of the PCR products showed that over 15 plants tested contained an unaltered nucleotide sequence of the βC1 gene. Several lines representing all four phenotypes of transgenic N. benthamiana and N. tabacum were further analyzed for the integrity of the βC1 transgene. Southern blot analysis indicated that the βC1 transgene had integrated into the genome of these plant lines (data not shown).

To examine the expression of the βC1 transgene in transgenic N. benthamiana and N. tabacum, total RNA was isolated from representative transgenic lines and hybridized with a radiolabeled βC1 gene probe. As shown in Fig. 6, the probe hybridized with a 380-bp transcript in transgenic N. benthamiana and N. tabacum. This transcript was not present in nontransgenic control plants (data not shown). As hypothesized, the presence of detectable amounts of βC1 transcript in transgenic plants coincided with the presence of symptoms of viral infection, with the phenotypically normal βC1 transgenic lines failing to accumulate detectable amounts of βC1 transcript. Moreover, the amounts of βC1 transcript in severely abnormal transgenic N. benthamiana were over 10 times greater than in moderately abnormal ones. Similarly, severely abnormal N. tabacum plants accumulated more βC1 transcript than moderately abnormal ones (Fig. 6). In all instances, there was a strong positive relationship between the severity of symptoms of virus-like infection and the amount of βC1 transcript in transgenic plants.

FIG. 6.

Northern blot analysis of RNA from transgenic plants containing the βC1 gene of TYLCCNV-Y10 DNAβ. Aliquots (40 μg) of plant total RNA extracted from transgenic leaf tissue of individual plants were loaded onto a 1.2% agarose gel containing formaldehyde. The blot was probed with a radiolabeled βC1 gene fragment. The ethidium bromide-stained gel shown below the blot indicates equal loading of RNA. The source plants were severely abnormal (SA), moderately abnormal (A), or symptomless (N). The arrow marks the location of βC1 mRNA.

Western blots were used in attempts to detect βC1 protein in transgenic N. benthamiana and N. tabacum plants with the antiserum raised against the βC1 fusion protein expressed in Escherichia coli. However, no βC1 protein was detected in transgenic plants exhibiting severely abnormal or symptomless phenotypes or in nontransgenic plants infected with TYLCCNV-Y10 and DNAβ (data not shown). Our failure to detect the βC1 protein suggests that it accumulates at low levels in vivo.

DISCUSSION

Novel begomovirus-DNAβ complexes were recently found to be associated with numerous economically important diseases occurring in Africa and Asia (2, 16). In Yunnan Province in the People's Republic of China, we found these complexes involving three begomovirus species: TYLCCNV, Tobacco curly shoot virus, and Malvastrum yellow vein virus (34). Here we show that all 25 TYLCCNV isolates collected from tobacco, tomato, or Siegesbeckia orientalis plants at widely separated locations in Yunnan Province were associated with DNAβ molecules, strongly suggesting that these DNAβ molecules have an important biological role.

Agroinfection tests showed that TYLCCNV DNAβ is required for the induction of typical disease symptoms in the plants tested, suggesting that the DNAβ molecule associated with TYLCCNV isolates is needed for induction of symptoms and that this complex may be responsible for disease symptoms in tobacco or tomato in Yunnan Province. This role parallels that of DNAβ species associated with Ageratum yellow vein virus, Cotton leaf curl Multan virus, Eupatorium yellow vein virus, and Bhendi yellow vein mosaic virus (3, 11, 22, 23). In contrast, a TYLCCNV isolate from Guangxi Province was previously shown to be a monopartite begomovirus, and its DNA-A-like molecule alone could infect tobacco and tomato systemically, inducing leaf curl symptoms (32). Y10 and Y64 have 90.1 and 95.5% nucleotide sequence identity, respectively, to the Guangxi isolate. Based on examination of the viral sequences, all three TYLCCNV isolates have functional C4 ORFs and their C4 ORFs share 95.2 to 98.3% nucleotide identity and 89.7 to 97.9% amino acid identity. Further work is needed to examine the reasons for this discrepancy with our results.

In our tests, TYLCCNV-Y10 DNAβ could not replicate autonomously. It depended on TYLCCNV-Y10 for replication and was probably encapsidated by the coat protein encoded by TYLCCNV-Y10, as evidenced by transmission of DNAβ by B. tabaci in association with Y10, because encapsidation of DNA is a probable requirement for vector transmission. However, TYLCCNV-Y10 DNAβ does not possess the cis-acting elements (iterons) needed for replication of TYLCCNV-Y10. These iterons are found in the common region of the DNA-A and DNA-B molecules of bipartite-genome begomoviruses (6). TYLCCNV-Y10 DNAβ presumably has one or more cis-acting elements needed for binding of TYLCCNV replication protein (Rep) and for replication, and these elements are most probably located in the 115-nucleotide highly conserved region of DNAβ upstream of its stem-loop structure.

Previous studies of the specificity of Rep binding have shown that a 5-bp core sequence (GGN₁N₂N₃) is a typical constituent of Rep-binding iterons (1). Conserved GG motifs occur upstream of the115-nucleotide highly conserved region of DNAβ. One or more of these GG motifs, combined with the 115-nucleotide highly conserved region, may be responsible for Rep binding to DNAβ. Iteron sequences typically differ between begomovirus species, and Rep binding during the initiation of viral DNA replication is highly specific, which conserves the genetic integrity of bipartite begomoviruses by preventing genome component exchange. However, the Rep-binding activity of the Tomato leaf curl virus satellite from Australia seems much less specific: Tomato leaf curl virus satellite contains an A-rich region but lacks a βC1 gene and is believed to be a defective DNAβ molecule (15). Tomato leaf curl virus satellite is replicated and causes systemic infection in association with distinct begomoviruses (Tomato yellow leaf curl virus and African cassava mosaic virus) and even with a curtovirus (Beet curly top virus) (4). The possibility that DNAβ also has a relaxed specificity in its recognition sequence for Rep is supported by the finding that Ageratum yellow vein virus DNAβ can be trans-replicated by Sri Lankan cassava mosaic virus DNA-A, although Ageratum yellow vein virus and Sri Lankan cassava mosaic virus have different iterons (24). Thus, DNAβ may be able to exchange helper viruses in mixed infections.

TYLCCNV-Y10 DNAβ enhances the accumulation of TYLCCNV-Y10 and is required for symptom induction. The effect of mutation of the conserved βC1 gene of TYLCCNV-Y10 DNAβ indicates that the βC1 protein plays a key role in symptom induction. A possible role of the βC1 protein is to interact directly with Rep protein to enhance begomovirus replication, either by combining with some other factor that affects DNA replication or by facilitating viral DNA movement within the plant. Other possible roles for the βC1 protein are to interact directly with a host factor to induce symptoms or to act as a suppressor of gene silencing, thereby allowing more efficient systemic infection of the plant. The sequences controlling expression of the βC1 protein were not determined, but it is worth noting that bidirectional promoters and TATA box elements typical of DNA-A were not found in the highly conserved region of TYLCCNV DNAβ. However, we note that a putative promoter and TATA box are located upstream of the βC1 gene.

Although the βC1 protein is responsible for symptom induction, the βC1 protein is not essential for the replication of DNAβ. Thus, the βC1 gene of DNAβ could be replaced by a foreign gene and be modified to convert it into an expression vector. The modified DNAβ might be an candidate gene silencing vector to study functional genomics in plants (16).

We have shown that leaf curl symptoms in Nicotiana species can be induced by transgenic expression of the βC1 gene of TYLCCNV-Y10 DNAβ that the severity of the symptoms parallels the level of βC1 transcript in the transgenic plants and their ability to induce symptoms is abolished by mutation of the βC1 gene. In naturally infected plants, DNAβ has been found only in association with begomoviruses that possess a monopartite genome (and so lack a DNA-B) (2). Indeed, evidence is increasing that DNAβ may play some of the same roles as DNA-B. For example, symptoms similar to those caused by viral infection are induced by transgenic expression of the DNA-B-encoded BC1 gene of Squash leaf curl virus in N. benthamiana (18), Tomato mottle virus in N. tabacum (5), and Bean dwarf mosaic virus in tomato (9). BC1 has been identified as a movement protein and shown to be a symptom determinant for bipartite geminiviruses (10, 29). Based on our transgenic expression and mutation analysis results in this study, we think that βC1 protein likely has a similar function in symptom development for some monopartite begomoviruses, it may also be involved in movement of the virus in conjunction with other virally encoded ORFs, which needs to be confirmed experimentally.

The situation with the monopartite genome Tomato leaf curl virus differs from that with either TYLCCNV or the bipartite genome begomoviruses. Tomato leaf curl virus induces symptoms without the aid of its small satellite DNA, the C4 gene in viral DNA was involved in symptom development of Tomato leaf curl virus and that transgenic plants expressing the C4 gene developed a viral-infection-related phenotype (13, 20). In contrast, the C4 gene in TYLCCNV seems not to induce symptoms because plants infected with TYLCCNV but not DNAβ are symptomless. Precisely how the DNAβ βC1 protein controls symptom expression is unclear. Some of the gene products of RNA viruses that induce symptoms in transgenic plants, such as the P0 protein of poleroviruses and P19 protein of tombusviruses, also act as suppressors of posttranscriptional gene silencing (19, 25, 27). Preliminary evidence suggests that the βC1 protein may have a similar function (16). Future work is needed to explore this possibility.