INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Geminiviruses are a family of plant viruses possessing a small genome of either one or two circular single-stranded DNAs (cssDNAs) of about 2.8 kb packaged in geminate particles (44). They replicate in the nucleus through a double-stranded intermediate using a rolling-circle replication (RCR) mechanism (44, 48). Following infection and uncoating of the viral genome, the complementary minus-strand DNA is synthesized on the cssDNA; this process is believed to be entirely under the control of host-encoded proteins. The new double-stranded circular DNA is also assembled in a minichromosome (41) and acts as a template for bidirectional gene expression directed by divergent promoters present on the intergenic region (IR) of the viral genome (49, 50). The RCR requires a site-specific nick on the plus-strand of the DNA to prime synthesis of the plus-sense cssDNA. In the geminiviral genome the nick () has been mapped to the conserved sequence TAATATTAC in the IR (31, 47) and is introduced by the viral replication-associated protein (Rep), which remains covalently linked to the 5' end of the original plus-strand, while the 3' hydroxyl is used to start plus-strand cssDNA synthesis (30).

Rep, which is encoded by open reading frame (ORF) C1 (also called AL1 or AC1 in geminiviruses with bipartite genomes), is a multifunctional protein involved in viral replication (12, 13, 16, 32), autoregulation of its own gene transcription (9, 51), and activation and recruitment of host-encoded proteins related to host DNA synthesis (35). Rep is the only viral protein that is absolutely required for viral replication (10, 11); genetic and biochemical studies have shed light on the functions of its different domains. The amino-terminal domain mediates (i) virus-specific origin of DNA replication and transcriptional repression (4, 5, 15, 25) and (ii) the nicking and joining activity required for initiation and termination of plus-strand DNA synthesis (19, 39). The central portion of Rep has a role in oligomerization (39), while the carboxy-terminal portion contains a nucleoside triphosphate (NTP)-binding domain required for viral replication (7, 18).

The C1 gene of geminiviruses infecting dicotyledonous plants contains in a different frame the small ORF C4. Mutation of ORF C4 impacts viral movement of monopartite but not bipartite geminiviruses in a host-dependent fashion (24, 42, 43), possibly by altering the ability to induce host replication machinery (29). The C4 protein of bipartite Tomato golden mosaic virus (TGMV) weakly represses the transcription of the C1 gene (8).

We have previously shown that expression in Nicotiana benthamiana and Lycopersicon esculentum plants of a truncated C1 gene of Tomato yellow leaf curl Sardinia virus (TYLCSV), encoding the first 210 amino acids (aa) of the Rep protein (T-Rep) and potentially coexpressing the C4 protein, confers resistance to TYLCSV (2, 36). Using a Nicotiana tabacum transient expression system, we showed that inhibition of TYLCSV replication can be achieved by expressing T-Rep and presented indirect evidence that the internal ORF C4 does not inhibit TYLCSV replication (36). T-Rep contains the domains involved in specific recognition of its own origin of replication (ori) (25) and in nicking and joining (19) but it lacks the nucleoside triphosphate (NTP)-binding domain required for viral replication (7).

To investigate the molecular mechanism of T-Rep-mediated interference and to establish the role of C4 protein, we used N. benthamiana stably or transiently expressing T-Rep and C4. We have analyzed the effects of expressing these proteins on viral transcription and replication. We show that T-Rep is alone responsible for the resistance and acts as a trans-dominant-negative mutant by repressing TYLCSV C1 gene transcription and viral replication and inducing accumulation of a heterogeneous population of partially duplex cssDNA possessing incomplete minus strands.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Virus resistance assays in plants. Transgenic plants were agroinoculated with Agrobacterium tumefaciens strains LBA4404/pBin19/TYLCV (27) and LBA4404/pBin19/SP98 (37), containing the infectious clone of TYLCSV and TYLCSV Murcia strain (TYLCSV-ES[1]), respectively. For the sake of brevity, TYLCSV-ES[1] will hereafter be called ES[1]. After inoculation, plants were observed for disease symptoms and assayed weekly by tissue print assay, essentially as described (36). Membranes were hybridized with a digoxigenin-labeled capsid protein-specific probe.

Plasmid constructs. (i) TYLCSV gene expression vectors. The plant expression vectors pTOM100 and pTOM100NT have been described (36); both can potentially express C4 protein from the overlapping ORF C4, but only pTOM100 can express T-Rep. The putative expression of the C4 protein from the internal ATG codon in both pTOM100 and pTOM100NT will be indicated hereafter as C4?. Plasmid pTOM100C4() was derived from pTOM100 by introducing a stop codon in ORF C4 without altering the amino acid sequence of T-Rep. A premature TGA stop codon, truncating the C4 protein after 9 amino acids, was created by a C-to-G mutation at position 339 of the complementary strand of TYLCSV (27); a C-to-T change at position 337 restored a leucine codon in the overlapping sequence of T-Rep. Site-directed mutagenesis was performed by PCR using Pfu DNA polymerase (Stratagene) (20). The template for the PCRs was pGEM102, obtained by subcloning the EcoRI-BamHI fragment of pTOM100 in pGEM4Z (Promega). Two PCRs were primed with the following oligonucleotide pairs: C4plus (CTCATCTCCATATTTTGATCCAATTCGAAG) and M13/pUC sequencing primer 47; C4minus (CTTCGAATTGGATCAAAATATGGAGATGAG) and M13/pUC reverse sequencing primer 48. Bold letters indicate introduced mutations. The two overlapping PCR fragments obtained were mixed and used as templates for a PCR with the external M13/pUC 47 and 48 primers. This PCR product was digested with EcoRI and BamHI, and the resulting fragment was cloned in pJIT60 (kindly provided by P. Mullineaux), giving pTOM100C4(). Plasmids pTOM120 and pTOM120C4() were constructed as follows: pTOM100 and pTOM100C4() were digested with EcoRI and subsequently partially digested with SacI, recovering the 4,030-bp fragments. The SacI-BglII fragment of TYLCSV from pTOM6 (a dimeric clone of TYLCSV in the SacI site of pBluescript SK) was cloned in SacI-BamHI-digested pBluescript SK (Stratagene), and the SacI-EcoRI fragment of this subclone was ligated into the EcoRI-SacI 4,030-bp fragment from pTOM100 or pTOM100C4(), obtaining pTOM120 and pTOM120C4(), respectively. Plasmids pTOM111 and pTOM110 are two different constructs for expression of C4 protein; they contain the C4 ORF from the first ATG (position 2463 of the complementary strand of TYLCSV) and from the second ATG (position 2457), respectively. C4 coding sequences were obtained by PCR from pGEM102 (see above), using the M13/pUC sequencing primer 47 in association with SarC4leader (AAACAATGGGGAACCTCATCTCCATAT) for pTOM110 or with SarC4leader1 (ACAAAAATGAAAATGGGGAACCTCATCTC) for pTOM111. PCR products were digested with EcoRI, generating fragments with an EcoRI and a blunt-ended extremity, which were cloned in EcoRI-PstI-blunt-ended pJIT60.

(ii) GUS reporter vectors. The transcriptional reporter plasmids pIntS/GUS and pIntS-ES[1]/GUS, containing the GUS ORF under the control of the TYLCSV or ES[1] complementary-sense promoter, respectively, were constructed as follow. The BamHI-SacI fragment of pBI121 (22), encompassing the GUS gene, was cloned in pSP65 (Promega), and the filled-in EcoRI-SmaI fragment of this subclone was introduced into SalI-digested and filled-in pJIT61 (obtained from pJIT60 by removal of the SacI site upstream the E35S promoter), producing pJIT61GUS. A fragment encompassing the TYLCSV complementary-sense promoter was PCR amplified from pTOM6 using primers TTTTGCTGTCGTTCTGAATC (nucleotides [nt] 2615 to 2634) and CACGAATGACGGAGATGAGA (nt 278 to 297). Similarly, a fragment encompassing the ES[1] complementary-sense promoter was PCR amplified from pSP97 (a 1.8-mer of ES[1] in pBluescript SK) (37) using primers TTGGTCAATGGGTACCAATTGAC (nt 2620 to 2642) and TGCAAGCATACAACGGAGAC (nt 192 to 211). The PCR products were digested with BamHI (generating fragments with one BamHI extremity and the other blunt-ended) and cloned in HincII-BamHI-digested pGEM4Z (Promega). The KpnI-PstI fragments of these subclones were then cloned in the corresponding sites of pJIT61GUS (replacing the Cauliflower mosaic virus [CaMV] 35S promoter sequence) to obtain pIntS/GUS and pIntS-ES[1]/GUS. pTOM202 is a translational fusion construct carrying the TYLCSV complementary-sense promoter (nt 2606 to 152) fused to the GUS gene (NcoI-HindIII fragment from pV668 (kindly provided by J. M. Bonneville, Grenoble, France) in pUC118. Its translation product consisted of the first five amino acids of Rep and six additional residues derived from the cloning procedure fused to the GUS protein sequence.

Replication assays in protoplasts. N. benthamiana protoplasts were isolated as described (34), and 5 × 10⁵ protoplasts were used for each transfection, with 2 µg of the viral infectious clone (pTOM6 or pSP97); cotransfections were obtained using 1 µg of either pTOM6 or pSP97 together with 5 µg of one of the TYLCSV gene expression vectors. Transfected protoplasts were grown in the dark at 24°C in K3 medium (38) containing vancomycin and cefotaxime (50 µg/ml each). Transfections were performed in duplicate or triplicate in at least three independent experiments.

Characterization of viral DNA forms. TNAs extracted from transfected N. benthamiana protoplasts were subjected to one of the following treatments and analysed by Southern blotting, as described above unless otherwise stated.

(i) Treatment with proteinase K. From 400 to 800 ng of TNAs was incubated in 10 mM Tris-HCl (pH 8.0)-5 mM EDTA-0.5% sodium dodecyl sulfate with 8.4 µg of proteinase K (Losung-Boehringer Mannheim) in a final volume of 50 µl at 56°C for 1 h. An additional 50 µl of buffer containing 8.4 µg of proteinase K was then added, and incubation was continued for 1 h. TNAs were then extracted with phenol-chloroform and ethanol precipitated.

(ii) Denaturation with alkali. TNAs (1 µg) were incubated in 20 µl of 50 mM NaOH at 37°C for 30 min, neutralized with 2 µl of 0.5 N HCl, and buffered with 2.5 µl of 1 M Tris-HCl (pH 8.0). TNAs were then purified through Microcon 100 (Millipore). Blots were hybridized with the digoxigenin-labeled TYLCSV C1 sense RNA probe (minus probe) and reprobed with a digoxigenin-labeled C1 antisense RNA probe (plus probe).

(iii) T4 DNA polymerase reactions. Samples were appropriately diluted to contain similar amounts of viral DNA, and the amounts of TNAs were equalized by adding TNAs from wt N. benthamiana protoplasts. Reactions were carried out at 37°C for 30 min in a final volume of 20 µl containing 400 ng of TNAs, 100 µM each dNTP, and 2 U of T4 DNA polymerase (Boehringer Mannheim) in the incubation buffer supplied, then the concentration of each dNTP was brought to 200 µM, and incubation was prolonged for another 30 min. Reactions were stopped at 75°C for 15 min.

(iv) Taq DNA polymerase reactions. Samples were diluted as described above for T4 DNA polymerase reactions. Reactions were carried out at 72°C for 5 or 10 min in a final volume of 20 µl containing 400 ng of TNAs, 250 µM each dNTP, and 0.5 U of Taq DNA polymerase (Qiagen) in the incubation buffer supplied. Reactions were stopped by adding gel-loading buffer.

Transcriptional repression assays. Total proteins were extracted from protoplasts 24 h posttransfection, and GUS activity was determined according to Jefferson et al. (22). Protein concentrations were determined using a Bradford protein assay kit (Bio-Rad), and GUS activity was corrected for protein concentration. For each experiment, background GUS activity of untransfected protoplasts was subtracted. Each construct was assayed in triplicate in at least three independent experiments. Mean values obtained in independent experiments and standard errors of the means were calculated. For analysis of gene expression in the transgenic system, protoplasts were transfected with 10 µg of each GUS reporter construct. GUS activities in transgenic and wt protoplasts were normalized through transfection of both with pJIT61GUS. For each construct, GUS activity in transgenic protoplasts was calculated as a percentage of the activity recorded from transfection of wt protoplasts. For the analysis of gene expression in the transient system, wt protoplasts were cotransfected with 10 µg of pTOM202 together with 10 µg of one of the TYLCSV gene expression vectors or 8.5 µg of pGEM-P, a pGEM7Zf(+) vector (Promega) carrying the E35S promoter; in each case, the molar ratio between the two cotransfected plasmids was 1:1. For each construct, GUS activity was expressed as a percentage of the activity recorded from cotransfection of pTOM202 with pGEM-P. A two-tailed Student's t test was used to compare the mean GUS activities obtained with the various constructs.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Resistance mechanism operating in 102.22 plants discriminates between two virus strains, acting at single-cell level. Transgenic N. benthamiana plants from line 102.22 (36), expressing T-Rep, were agroinoculated with TYLCSV or with ES[1], and virus infection was monitored weekly. The amino acid sequence identity between TYLCSV and ES[1] individual genes ranges from 80 to 100%, with Rep showing 90% identity (37). In 8 of 11 transgenic plants agroinoculated with TYLCSV, viral infection was delayed at least a week compared to infection on wt N. benthamiana plants. On the contrary, only 1 transgenic plant of the 10 agroinoculated with ES[1] showed a 1-week delay of infection, suggesting that the resistance mechanism operating in 102.22 plants was discriminating between the two strains.

View larger version (33K): [in this window] [in a new window]   FIG. 1.   Replication of TYLCSV and TYLCSV-ES[1] in wt and transgenic (102.22) N. benthamiana protoplasts. Southern blots of total nucleic acids extracted 72 h after transfection with either pTOM6 (TYLCSV) or pSP97 (TYLCSV-ES[1]) were hybridized with a C1-sense RNA probe. DNAs from mock-inoculated protoplasts and infected plants were included as controls. Slowly migrating DNAs are indicated with an asterisk (*). scDNA, supercoiled DNA. When pSP97 was used, slight inhibition of viral replication was observed in most experiments (panel 1), but in one case (panel 2), consistent inhibition coupled with the presence of the slow-migrating DNAs was detected.

Transient expression of T-Rep but not of C4 protein confers virus inhibition similar to that in 102.22 transgenic protoplasts. Wt N. benthamiana protoplasts were cotransfected with pTOM6 or pSP97 together with the plant expression vector pTOM100 (Table 1), which contains the expression cassette used in plant transformation. Transiently expressed T-Rep inhibited replication of TYLCSV but not ES[1] (Fig. 2); inhibition was coupled with the presence of slowly migrating DNAs (Fig. 2). These data show that transient and transgenic expression leads to similar results.

View larger version (40K): [in this window] [in a new window]   FIG. 2.   Consequence of expressing T-Rep and C4 proteins on viral DNA replication. Wt N. benthamiana protoplasts were cotransfected with either pTOM6 (TYLCSV) or pSP97 (TYLCSV-ES[1]), together with the constructs indicated above, which can transiently express the viral protein indicated below (see Table 1 for description of constructs). The C4 protein may start at the first ATG codon (1st ATG) of the ORF or at another nearby in-frame ATG (2nd ATG) C4?, C4 protein potentially expressed from internal ATG codon. Other details as in Fig. 1.

Slowly migrating DNAs are partial duplexes. The slowly migrating DNAs were investigated further. In a first experiment, TNAs extracted from wt protoplasts cotransfected with pTOM6 and pTOM100 were hybridized with a minus probe, corresponding to the region of the C1 gene, or a plus probe, corresponding to the V1 gene of TYLCSV. As shown in Fig. 3A, only the minus probe detected the slowly migrating DNAs. These results, the migration pattern of the DNAs, the functional domains present on T-Rep, and the replication mechanism of geminiviruses suggested that they could be (i) positive ssDNAs covalently linked to T-Rep or Rep, (ii) positive ssDNAs longer than unit length, or (iii) partial duplexes. Gel migration patterns of TYLCSV DNA in T-Rep-expressing protoplasts did not change after proteinase K treatment (Fig. 3B), excluding the hypothesis of DNA-protein complexes.

View larger version (31K): [in this window] [in a new window]   FIG. 3.   Characterization of slowly migrating DNAs, indicated with an asterisk (*). The positions of scDNA and cssDNA forms of viral DNA as well as input DNA are indicated. (A) The slowly migrating DNAs do not hybridize with a plus-sense probe. Wt protoplasts were cotransfected with pTOM6 together with pTOM100 or pTOM100NT, and TNAs were analyzed 72 h posttransfection. The blot was hybridized with a C1-sense RNA probe (probe minus) and reprobed with a V1-sense RNA probe (probe plus). DNAs from mock-inoculated protoplasts and from infected plants were included as controls. (B) The slowly migrating DNAs are not DNA-protein complexes. TNAs extracted from protoplasts cotransfected as in panel A were incubated at 56°C for 2 h with (+) or without () proteinase K prior to Southern blot analysis. The blot was hybridized with a C1-sense RNA probe. (C) Alkali treatment converts the slowly migrating DNAs to DNA migrating as cssDNA. TNAs extracted from transgenic 102.22 protoplasts at 72 h posttransfection with pTOM6 were analyzed directly (no treatment), after denaturation with NaOH (alkali), or after restriction with DpnI followed by NaOH (DpnI + alkali). The blot was hybridized with a C1-sense RNA probe (probe minus) and reprobed with a C1-antisense RNA probe (probe plus).

View larger version (36K): [in this window] [in a new window]   FIG. 4.   Slowly migrating DNAs are converted into ocDNA by Taq (A) or T4 (B) DNA polymerase treatment. The blots were hybridized with a C1-sense RNA probe. The positions of ocDNA, linear DNA (linDNA), scDNA, and cssDNA forms of viral DNA are indicated. Slowly migrating viral DNAs are indicated with an asterisk (*). (A) TNAs were extracted from wt protoplasts at 72 h posttransfection with pTOM6 alone (the two lanes on the right) or together with pTOM100C4() or pTOM100NT and analyzed directly () or following incubation with Taq DNA polymerase for the time indicated below. (B) TNAs were extracted from wt or transgenic (102.22) protoplasts at 72 h posttransfection with pSP97 (TYLCSV-ES[1]) and analyzed following a 1-h incubation with (+) or without () T4 DNA polymerase. Lane C, TNAs from a TYLCSV-infected tomato plant digested with BglII to show migration of linear DNA.

Viral DNAs accumulating early after transfection of wt protoplasts with pTOM6 resemble those observed in T-Rep-expressing transfected ones. Partially duplex cssDNAs are natural intermediates of RCR. At 72 h after protoplast transfection with pTOM6, only a fraction of the total viral DNA was in partial duplex form, as shown above (Fig. 4A). Time course experiments showed that TYLCSV amplification reached a plateau around 72 h posttransfection (data not shown). However, the ratio between the viral DNA forms would be expected to change during viral replication. In particular, viral replication intermediates should accumulate at lower levels late in viral replication. Thus, we asked if, early in TYLCSV replication, partial duplexes resembling those observed in T-Rep-expressing protoplasts would accumulate. Figure 5A shows a Southern blot of TNAs extracted from wt protoplasts at 16, 20, 24, and 72 h posttransfection with pTOM6. Samples at 72 h were diluted 20-fold to give signals of similar intensity on the autoradiography film. Interestingly, at 20 and 24 h posttransfection, a heterogeneous class of viral DNAs migrating slower than cssDNA was observed. Similar results were obtained in four independent experiments. This pattern clearly resembled that in T-Rep-expressing protoplasts. To verify the nature of these slowly migrating molecules, TNAs extracted from the samples at 24 and 72 h were incubated with Taq DNA polymerase and analyzed by Southern blot (Fig. 5B). The slowly migrating DNAs present at 24 h were converted into ocDNA. These data suggest that accumulation of a discrete population of partial duplexes is an early event in viral replication and does not require the expression of T-Rep.

View larger version (29K): [in this window] [in a new window]   FIG. 5.   Slowly migrating DNAs also accumulate at early stages after transfection of wt protoplasts with TYLCSV. TNAs were extracted from protoplasts transfected with pTOM6. The blots were hybridized with a C1-sense RNA probe. The positions of ocDNA, linear DNA (linDNA), scDNA, and cssDNA forms of viral DNA are indicated. Slow-migrating viral DNAs are indicated with an asterisk (*). HPT, hours posttransfection. (A) Time course analysis of viral DNA forms. Sample at 72 h was diluted 20-fold to give signals of satisfactory intensity on the autoradiographic film. (B) Taq DNA polymerase treatment of TNAs at 24 and 72 h after transfection. Samples were incubated for 10 min at 37°C with (+) or without () the polymerase. Lane C, sample at 72 h digested with BglII to show the linear DNA.

Both transgenic and transient expression of T-Rep strongly represses the homologous C1 promoter. To see if virus resistance is correlated with the ability of T-Rep to inhibit C1 gene transcription, 102.22 and wt protoplasts were transfected with constructs containing the GUS reporter gene coding sequence fused to the C1 promoter of TYLCSV or ES[1]. In pTOM202, a GUS gene cassette was fused in frame to the fifth codon of the TYLCSV C1 gene, whereas in pIntS/GUS and pIntS-ES[1]/GUS, the untranslated C1 leader sequences of TYLCSV and ES[1], respectively, were transcriptionally fused to a GUS gene cassette (Table 1). All the constructs contained the complete viral IR. In wt protoplasts, GUS activity of pIntS/GUS was about 15-fold less than that observed with pTOM202 (Fig. 6A). In transgenic protoplasts, the GUS activity of pIntS/GUS and pTOM202 was repressed 83- and 333-fold, respectively, compared with the activity in wt protoplasts, while only a 1.3-fold reduction of activity was observed with the pIntS-ES[1]/GUS construct (Fig. 6A). These results show that the TYLCSV but not ES[1] C1 promoter was strongly repressed in T-Rep transgenic protoplasts.

View larger version (23K): [in this window] [in a new window]   FIG. 6.   Transcriptional repression assays. Values are the averages of at least three independent experiments assayed in triplicate, and error bars indicate the standard error of the mean. Mean absolute GUS activity measured in transfected wt protoplasts, expressed in picomoles of 4-methylumbelliferone (MU) per minute per milligram of protein is indicated in italics. Background GUS activities associated with untransfected protoplasts were 18 and 29 pmol of MU/min/mg for transgenic and wt protoplasts, respectively. (A) 102.22 transgenic and wt N. benthamiana protoplasts were transfected with the GUS reporter constructs indicated on the left. GUS activity was measured in protein extracts at 24 h posttransfection. For each construct, GUS activity in 102.22 protoplasts (open columns) is shown as a percentage of the activity recorded from transfection of wt protoplasts (shaded columns) (in bold). (B) Wt N. benthamiana protoplasts were cotransfected with the GUS reporter construct pTOM202 together with the plant expression cassettes indicated on the left. GUS activity was measured in protein extracts at 24 h posttransfection. GUS activity obtained in cotransfection of pTOM202 with pGEM-P was set at 100%, and other values were standardized to this level (in bold).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Our data show that TYLCSV-resistant 102.22 plants are susceptible to ES[1] and that a direct correlation exists between the virus resistance specificity in plants and the ability of single transgenic cells to strongly inhibit TYLCSV but not ES[1] replication. We also showed that inhibition of viral replication is characterized by reduced levels or absence of cssDNAs and concomitant appearance of a heterogeneous class of DNAs migrating more slowly than cssDNA. These data strongly suggest that the mechanism of resistance operating in transgenic T-Rep-expressing plants acts at the single-cell level through the repression of viral replication. Moreover, we showed that the C4 protein, directly expressed using an enhanced CaMV 35S promoter, is not able to inhibit TYLCSV replication and that pTOM100C4(), expressing only T-Rep, represses TYLCSV amplification even better than pTOM100, expressing T-Rep and potentially the C4 protein (Fig. 2).

The bona fide ability of a transient expression system to mimic what happens in stable transgenics was evident at two different levels. First, both systems showed the same virus specificity. Second, the inhibition of TYLCSV replication was coupled in both systems with the presence of slowly migrating DNAs. Moreover, using the same plant species (N. benthamiana) for both the transient and stable expression assays, we avoided possible misinterpretations due to host-specific behavior of the C1 and C4 proteins. Indeed, TGMV Rep-mediated C1 transcriptional repression was shown to be 10-fold more effective in N. benthamiana than in N. tabacum protoplasts (9), and TYLCSV ORFC4 mutants were able to infect N. benthamiana but not tomato plants (24).

Altogether, these data demonstrate that T-Rep is responsible for the strong and specific inhibition of viral replication, while the C4 protein, if expressed, has no role in this activity. Supporting this conclusion, there is preliminary evidence that tomato plants transformed with the pTOM100C4() construct are resistant to TYLCSV (M. Tavazza and G. P. Accotto, unpublished data).

As mentioned above, a distinctive aspect of the altered viral DNA pattern found in transgenically and transiently T-Rep-expressing protoplasts was the presence of a slowly migrating and heterogeneous population of DNAs. Three orders of evidence clarified the nature of these molecules. First, proteinase K treatment did not increase their mobility (Fig. 3B), and we did not detect a preferential partition of these molecules in the organic-aqueous interphase, as expected for stable DNA-protein complexes (26) (data not shown). Second, alkaline treatment converted them into a sharp band migrating as the cssDNA, which hybridized only with a minus TYLCSV probe; therefore, the hybridization signal cannot derive from denatured scDNA or ocDNA. Third, the in vitro assays with DNA polymerases (Taq, T4, and Pfu; Fig. 4A and 4B and data not shown) showed that, independent of the T-Rep protoplast system used (transgenic or transient expression), the slowly migrating DNAs were converted into ocDNA. The overall data demonstrate that these DNAs are partial duplexes, with the minus strand incomplete and heterogeneous in length, suggesting that they do not derive from interference with nicking and joining activities during virus replication.

The occurrence of partial duplexes is actually expected in the RCR mechanism (44); however, it should be noted that if synthesis of minus-strand DNA proceeds at a constant rate, we should not expect to see a discrete population but rather a smear of partial duplexes migrating between the cssDNA and the ocDNA. Interestingly, time course experiments of TYLCSV replication in wt protoplasts showed that a discrete population of partial duplexes resembling that found in transgenically and transiently T-Rep-expressing protoplasts is abundant in the initial phase of replication. From this, two conclusions can be drawn. First, the presence of partial duplexes in T-Rep-expressing protoplasts does not result from direct interference of T-Rep with complementary-strand synthesis. Second, there are two phases in complementary-strand synthesis, the first phase or the switch between them being rate limiting. The hypothesis of two phase would be in agreement with the presence of heterogeneous RNA/DNA molecules of complementary polarity during African cassava mosaic virus (ACMV) infection (45) and with the minus-strand replication of some bacteriophages (1, 14).

A key element of virus-specific recognition of the viral plus-strand origin is the binding of Rep, through its amino-terminal domain, to an iterative sequence located between the TATA box and the transcription start site of the C1 gene (3, 12, 15). Binding of Rep to the same sequence also mediates virus-specific repression of its own gene (9, 15). Our results show that transgenic protoplasts were able to tightly control TYLCSV but not ES[1] C1 transcription. This is in accordance with the virus-specific inhibition of viral replication at the single-cell level and with the virus-specific resistance observed in plants. Moreover, using a transient-expression system, we showed that T-Rep alone is responsible for the transcriptional repression of C1, while the C4 protein has no role in this activity. These data indirectly suggest that T-Rep is able to bind to the iterative sequence required for viral plus-strand origin.

Truncated Reps of geminiviruses show different abilities to repress C1 gene transcription. The determinants of TGMV Rep-mediated virus-specific transcriptional repression have been mapped to the first 93 amino acids (15), but deletion of as few as 39 residues from the carboxy terminus abolishes in vivo transcriptional repression (17). On the contrary, a truncated version of ACMV Rep containing only the first 57 amino acids is fully active in repressing its own gene transcription (21). Thus, besides the importance of the Rep amino-terminal domain for binding specificity, different Reps appear to require different additional carboxy-terminal portions for a productive interaction with the cis-acting sequences and perhaps with transcription factors in repressing C1 gene transcription.

The ability of T-Rep to repress C1 transcription fivefold more than Rep suggests that the carboxy-terminal part (151 aa) of wt Rep could have a role in downregulating C1 repression. However, when the same proteins were expressed from plasmids potentially coexpressing the C4 protein (pTOM100 and pTOM120), no difference in C1 transcriptional repression was observed. Differences in GUS activities between pIntS/GUS and pTOM202 can be explained by the observation that in pTOM202 the start codon of the GUS gene is in an optimal translational initiation context derived from the untranslated C1 gene leader sequence (23), whereas in pIntS/GUS the same leader sequence is separated from the start codon of the GUS gene by an intervening sequence derived from the cloning procedure. Similarly, an unfavorable translational initiation context occurs in the pIntS-ES[1]/GUS construct.

Our data suggest a model of TYLCSV resistance conferred by T-Rep. After entry into the cell and uncoating of the viral genome, the complementary-sense strand is synthesized on the plus-strand cssDNA, generating double-stranded circular DNA. At this time T-Rep can recognize and bind to the cognate DNA, inhibiting but not abolishing C1 transcription. The limited amount of newly synthesized Rep will not be able to properly synthesize the viral plus-strand, since it will be in competition with T-Rep for utilization of the required sequence; as a result, viral replication will be inhibited. This double action of T-Rep lowers the rate of TYLCSV replication, mimicking the early phase of wt virus infection, characterized by the prevalent accumulation of a discrete population of partial duplexes. We cannot exclude that T-Rep can bind Rep, forming a dysfunctional multimeric complex. However, since Tomato golden mosaic virus (TGMV) Rep can form heterooligomers with Bean golden mosaic virus (BGMV) Rep, but transient expression of TGMV Rep does not inhibit BGMV replication (46), it is unlikely that a mechanism based on T-Rep/Rep dysfunctional multimeric complexes can substantially contribute to the strain-specific resistance observed. The high level of T-Rep required to confer resistance, together with our inability to detect Rep in wt TYLCSV infection (2) and with the evidence that repression of C1 gene transcription close to background level is not able to abolish TYLCSV replication, suggests that a small amount of Rep can efficiently compete with T-Rep. Viral replication and transcriptional repression could require different states of aggregation of Rep (33, 40), and we do not know if T-Rep and Rep have the same ability to form multimeric complexes. However, the ability of geminiviruses to induce posttranscriptional gene silencing (28) should also be considered as an added mechanism to overcome the tight transcriptional repression observed.