INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Rolling-circle replication (RCR) of DNA appears particularly well suited for the multiplication of genetic information stored in the form of single-stranded DNA (ssDNA) (29). Genetic entities that multiply their DNA via RCR range from ssDNA plasmids of Archaebacteria (57) and Eubacteria (31, 82), ssDNA phages (e.g., X174) (7), ssDNA viruses of plants (gemini- and nanoviruses) (19, 44, 52) to circoviruses and ssDNA viruses of birds (8, 62) and mammals (34, 60, 76). There is also a linear variant of RCR in the form of a rolling hairpin, a mechanism by which the single-stranded linear genome of parvoviruses is multiplied (11, 21). Common denominators of all these examples are specific replication initiator (Rep) proteins encoded by their own respective plasmid or the virus DNA and their interaction with a target sequence that, in many cases, may form a particular secondary structure, for instance, a hairpin (33, 36, 52, 54, 56).

Unlike the ssDNA plasmids, phages, and circoviruses, the genetic information of some geminiviruses is distributed over two DNAs, and that of the nanoviruses is distributed over at least six different DNAs. This creates the need for the Rep proteins to ensure multiplication of several different DNA molecules. In addition, the apparent redundancy of nanovirus DNAs encoding similar yet distinct Rep proteins raises questions about their respective specific roles in the replication process of the multiple component genome of the nanoviruses.

Nanoviruses (formerly referred to as "plant circoviruses") have only recently been established as a separate genus of plant viruses with a genome consisting of multiple circular ssDNAs, each about 1 kb in size (67). Six to 11 different DNA components have so far been identified in the four assigned nanovirus species, such as the banana bunchy top virus (BBTV), faba bean necrotic yellows virus (FBNYV), milk vetch dwarf virus (MDV), and subterranean clover stunt virus (SCSV) (13, 14, 50, 70). With a single exception, each DNA appears to contain only one gene (9, 10). Each ssDNA is individually encapsidated into small isometric virions, and virus transmission is accomplished only by insects (aphids). The lack of an alternative experimental infection system has so far precluded the identification of the entire set of DNAs that represent the complete genome of a nanovirus.

In contrast, geminiviruses, the first ssDNA plant viruses described (30, 37), are well characterized in biological and molecular terms (12). They differ from the nanoviruses by their twin (geminate) particles, their genome organization, and insect vector (55). Geminivirus replication proteins bind specifically to double-stranded DNA (dsDNA) motifs in the origin region (1, 24, 26) and have a site-specific in vitro cleavage and joining activity on single-stranded oligonucleotides (40, 54) that represent part of the viral replication origin sequences (39, 64, 71). In addition, they possess an ATPase activity essential for virus replication (22, 35). The protein domains and amino acids responsible for the various activities of Rep proteins have been genetically or biochemically identified (18, 22, 42, 46, 53, 65, 66).

However, little is known about the molecular details of nanovirus replication. Minus-strand DNA synthesis of BBTV is primed by virion-associated primers (32), and a Rep protein was inferred to be encoded by one of the six BBTV genome components (36), based on the presence of amino acid motifs conserved among RCR initiator proteins (44). For this protein, a cleavage and joining activity of oligonucleotides corresponding to the inverted repeat sequence of the viral replication origin has been shown in vitro (33). A more complex picture emerged, however, when four further Rep-encoding DNAs (rep components) were identified from a Taiwanese isolate of BBTV (79, 80) and two and four rep components were described for the 7- and 10-component genomes of SCSV (13) and MDV (70), respectively.

For FBNYV, the object of this study, 11 distinct ssDNA components (C1 to C11) have been identified (48-50). Each of them encodes a single protein, such as the virus capsid protein (48), a cell cycle-link protein (4, 5), and four other proteins of as yet unknown function. C1, -2, -7, -9, and -11 encode different but closely related Rep proteins, each of about 33 kDa. The presence in two FBNYV isolates of five DNAs encoding distinct Rep proteins along with (at least) six DNAs encoding other viral proteins immediately raised two questions: whether all rep components are required and which of them are essential for the multiplication of the other genome components.

Here we describe the first biochemical and genetic analyses of FBNYV Rep1 and Rep2 protein activities along with the proof that each of five replication proteins is functional in vivo, i.e., supports autonomous replication of its coding component. Furthermore, we show that only one Rep protein, Rep2, has the capability and is sufficient to initiate in trans the replication of all other genome components of FBNYV that encode no Rep. This demonstrates the existence of a master replication protein in a multipartite ssDNA virus.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Clones of FBNYV DNAs. The DNA sequences of the FBNYV-Sy isolate SV292-88 were previously described (48-50), and their EMBL/GenBank accession numbers are as follows: X80879 (C1), Y11405 to Y11409 (C2 to C6), and AJ005964 to AJ005967 (C7 to C10). C1 and C2 of this isolate are referred to here as C1-Sy and C2-Sy, respectively. Clones of C2 to C10 of the Egyptian isolate EV1-93 were obtained after PCR amplification by Pfu DNA polymerase by using primers designed to create unique restriction sites at the ends of the amplified DNA (see Tables 1 and 2 for details). In this way, all components, except C5 and C7, were cloned in the corresponding sites of pBluescript IISK(+) (Stratagene). To clone C5, two overlapping fragments were amplified by using the combination of the primer pairs P6 and P7 or P18 and P19 (Table 1). Both fragments were then digested with Asp718 and PvuII, ligated, cut by XbaI, and inserted into the XbaI site of pUC19 (81). Since for subsequent replication assays in plant tissue the binary vector pBin19 was used (28), the 1-kb amplification product of C7 had to be circularly permutated prior to insertion into pBin19. Therefore, the DNA was digested by PvuI, self-ligated, cut by HindIII, and inserted into the HindIII site of pUC19. Similarly, a clone of FBNYV C11 (FBNYV C1-Eg in reference 50), available as a PstI insertion in pUC19, was liberated by digestion with PstI, ligated, digested again by XbaI, and inserted into the XbaI site of pUC19.

Rep expression in E. coli and protein purification. Rep1 and Rep2 proteins were expressed in Escherichia coli with an N-terminal hexahistidine tag by using plasmid pQE30 (QIAGEN). DNA of C1-Sy and C2-Sy, cloned in the HindIII site of pGEM-3Zf(+), was released by HindIII and self-ligated, and the respective rep1 and rep2 genes were amplified by PCR with Pfu DNA polymerase and the following primer pairs: rep1 by using oligonucleotides rep1-SphI as the 5'-end primer and rep1-SalI as the 3'-end primer and rep2 by using oligonucleotides rep2-SphI as the 5'-end primer and rep2-SalI as the 3'-end primer (see Table 1 for details). The SphI and SalI restriction sites at the 5' and 3' ends of the amplified rep gene DNAs served for insertion into the corresponding sites of plasmid pQE30. The resulting plasmids, pQE30-rep1 and pQE30-rep2, were introduced into E. coli BL21(DE3)-recA (2, 75) harboring plasmid pRep4 (Qiagen) and expressing a high level of lac repressor to guarantee a tight control of protein synthesis. Bacteria were grown at 28°C in M9 minimal medium supplemented with 0.1% Casamino Acids, 2% glucose, and 0.1% thiamine, to an optical density at 600 nm of about 0.5, and induced by addition of 0.5 mM isopropyl--D-thiogalactopyranoside for 2 h. After centrifugation, the pellets were resuspended in 50 mM phosphate buffer (pH 8.0) containing 300 mM NaCl and 10% glycerol (5 ml of buffer for the bacterial pellet from 250 ml of culture) and frozen. Bacteria were lysed for 30 min with lysozyme (0.5 mg/ml) in the presence of 0.5% Tween 20 and 1 mM phenylmethylsulfonyl fluoride at 0°C, sonicated for 3 min, and further incubated for 30 min on ice with DNase and RNase (5 µg/ml). After clarification of the bacterial lysate by centrifugation at 10,000 × g for 30 min, the supernatant was loaded onto a TALON (Clontech) metal affinity resin column (Co²⁺ as a chelating ion), and the Rep1 and Rep2 proteins were eluted with 250 mM imidazole. The two Rep proteins were sufficiently pure for all subsequent biochemical assays; the yield of Rep2 protein was consistently about 10 times higher than that of Rep1.

Site-directed mutagenesis of the rep1 and rep2 genes. Mutagenesis to change tyrosine residues into phenylalanine (Y78 of Rep1 or Y79 of Rep2) and the lysine residues into alanine (K177 of Rep1 or K187 of Rep2) was performed with plasmids pBSKII:C1-Sy and pBSKII:C2-Sy or directly with expression plasmids pQE30-rep1 and pQE30-rep2 by applying the "Quik Change" site-directed mutagenesis kit (Stratagene). The primers rep1^Y78F(+) and rep1^Y78F() and rep2^Y79F(+) and rep2^Y79F() were used to change the respective tyrosine residues, and rep1^K177A(+) and rep1^K177A() and rep2^K187A(+) and rep2^K187A() were used to change the respective lysine residues.

DNA cleavage and nucleotidyl transfer assay. Approximately 50 to 500 ng of Rep1 or Rep2 protein was incubated with 50 fmol of ³²P-labeled (and nonlabeled) substrate DNA in a total volume of 20 µl for 20 min at 37°C in a reaction buffer containing 25 mM Tris-HCl (pH 7.5), 75 mM NaCl, 2.5 mM MnCl₂, and 2.5 mM dithiothreitol. The reaction was stopped by adding 2 µl of 0.5 M EDTA, lyophilized, resuspended in formamide, and analyzed on 12% sequencing gels (68), which were then dried and autoradiographed.

ATPase assay. About 100 ng of Rep1 or Rep2 protein was incubated at room temperature with 5 nM [-³²P]ATP and various (5 to 50 µM) concentrations of nonlabeled ATP in a 10-µl reaction mixture containing 50 mM PIPES-NaOH (pH 7.0), 50 mM NaCl, 10 mM MgCl₂, and 0.02% NP40. The reaction was stopped by the addition of 1 µl of 0.5 M EDTA. The reaction products were separated on thin-layer chromatography-polyethyleneimine cellulose (TLC-PEI) plates (Schleicher & Schuell) with 0.5 M LiCl-1 M formic acid as running buffer. The amount of ³²P_i liberated was quantified with a PhosphorImager (Molecular Dynamics).

Replication of FBNYV components in N. benthamiana. For replication assays, redundant copies (direct repeats) of the FBNYV C1-Sy, C2-Sy, C3, C4, C5, C6, C9, and C10 were first constructed as dimers in pUC19 or pBluescript IISK(+) (Table 2). The redundant copies were released by using suitable restriction endonucleases (Table 2) and inserted into the binary vector pBin19, which contains a DNA fragment from the T region of Agrobacterium tumefaciens Ti plasmid (T-DNA) (41). Dimers of C2, -7, and -8, as well as of the mutated C1-Sy and C2-Sy (see the section above on site-directed mutagenesis), were directly assembled in the HindIII site of pBin19; a dimer of C11 in the XbaI site of pBin19 was obtained in the same way. Because of a severe toxicity of Rep9 protein in agrobacteria, the redundancy of C9 had to be reduced to 1.1. A 100-bp PvuII-ApoI fragment of C9 was treated with the Klenow fragment of DNA polymerase I to obtain blunt ends prior to cloning in the AccI site of pBluescript IISK(+), which had been similarly filled in. The unique AccI site in the FBNYV DNA served to accept a full-length copy of C9 (liberated from pBSKII:C9 by digestion with PstI and circularly permuted by ligation and subsequent cleavage with AccI). The 1.1-mer of C9 was transferred as a BamHI-KpnI fragment into pBin19.

Dot blot hybridization assays for rep components in FBNYV samples from different countries. The samples of FBNYV-infected legumes from Egypt, Ethiopia, Jordan, Morocco, and Syria, kept as desiccated leaf tissue at 4°C, were those described previously (27). In addition, FBNYV samples from virus surveys in Ethiopia, Pakistan, and Turkey in 1997 (kindly provided by K. M. Makkouk, ICARDA, Aleppo, Syria) and from Algeria (provided by Linda Allala, INA, El-Harrach, Algeria) were also included. Infection by FBNYV was determined by polyclonal and monoclonal antibodies to the virus in double- and triple-antibody sandwich enzyme-linked immunosorbent assays (27). Fresh and dried samples extracted at 1:20 (wt/vol) and 1:200 (wt/vol), respectively, with 20× SSPE (3 M NaCl, 0.2 M Na₂HPO₄, 20 mM EDTA [pH 7.4]) were dotted as 100-µl aliquots onto positively charged nylon membranes (Boehringer Mannheim) by using a 96-well vacuum manifold (Gibco BRL). ³²P-labeled rep component-specific probes (from nucleotide [nt] 598 to nt 929 for C1, 136 to 969 for C2, 67 to 814 for C7, 456 to 972 for C9, and 69 to 436 for C11) were prepared by PCR with primers derived from the respective coding regions and viral DNA of FBNYV-Sy, except for the C11 probe, for which FBNYV-Eg DNA served as template. Random-primed probe labeling with [-³²P]dCTP was carried out with the Megaprime kit (Amersham). Hybridization was conducted essentially as described in reference 68 by incubating the membranes at 42°C and washing the membranes at high stringency (0.1× SSPE and 0.1% sodium dodecyl sulfate at 65°C).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The FBNYV isolate SV292-88 from Syria (FBNYV-Sy), whose DNA sequence had been assembled mostly from partial clones and PCR-amplified DNA, is no longer available as an aphid-transmissible virus isolate. Therefore, full-sized genome components of an FBNYV isolate, EV1-93, from Egypt (FBNYV-Eg) that is serologically indistinguishable from FBNYV-Sy (27) and maintained by continuous insect transmission on fava bean (Vicia faba) were cloned, and their DNA sequences were determined (EMBL/GenBank accession no. AJ132179 to AJ132187 [C2 to C10] and AJ005968 [C11]). Whereas most of the components of both FBNYV isolates were very similar (>94% nucleotide identity), there was a striking difference in both coding (61% amino acid identity) and noncoding regions (68% nucleotide identity) between the rep1 component of FBNYV-Sy and its closest homologue in FBNYV-Eg (50). Therefore, the latter was designated C11. Furthermore, the encoded Rep protein (Rep11) turned out to be functionally distinct from Rep1 (see below). A sequence comparison of the five different Rep proteins encoded by the respective rep components (C1, C2, C7, C9, and C11) and the active site amino acids of Rep1 and Rep2 altered by site-directed mutagenesis are shown in Fig. 1.

View larger version (86K): [in this window] [in a new window]   FIG. 1.   Sequence comparison of FBNYV Rep proteins. Amino acid sequences of Rep1 (FBNYV-Sy) and Rep2, -7, -9, and -11 (FBNYV-Eg) were aligned by using PileUp of the UWGCG sequence analysis software package. Amino acids identical in all five Rep proteins are shown in black. Tyrosine and lysine residues that were altered by site-directed mutagenesis are marked by an asterisk (Y78 and K177 in Rep1 and Y79 and K187 in Rep2).

FBNYV DNAs encoding Rep proteins replicate autonomously. The association with FBNYV of five apparently distinct DNAs encoding putative Rep proteins raised the question of whether all of these Rep proteins were functional. Therefore, the capacity of each rep component to replicate autonomously in leaf discs of N. benthamiana was examined. Leaf discs were inoculated with agrobacteria carrying redundant copies of each of the five rep components (C1, C2, C7, C9, and C11) in the binary T-DNA vector pBin19. One week after agroinoculation, total DNA was isolated from the infected tissue and probed for replicative forms of viral DNA by Southern hybridization analyses (Fig. 2). The results indicated that each of the five rep components replicated autonomously, which means they expressed a functional Rep protein that acts on its cognate component to initiate DNA replication.

View larger version (56K): [in this window] [in a new window]   FIG. 2.   FBNYV DNAs encoding Rep proteins replicate autonomously in N. benthamiana. Southern hybridization of total DNA extracted from N. benthamiana leaf discs inoculated with agrobacteria carrying pBin19 with redundant copies of a respective FBNYV rep component indicated below each blot (rep1, rep2, rep7, and rep11 components were cloned as dimers, and rep9 was cloned as a 1.1-mer). Component-specific probes were used as indicated (rep probe). + S1, DNA was digested with nuclease S1 to digest the ssDNA; + HindIII, DNA was digested with HindIII to linearize rep1 or rep2 DNAs. agro, DNA extracted from agrobacteria used for inoculation. ss, ssDNA; ccc, covalently closed circular DNA; l, linear DNA; oc, open circular DNA. The DNA bands that migrate slower than double-stranded open circular DNA represent dimers and oligomers of replicative forms as they disappeared after digestion by HindIII.

FBNYV Rep1 and Rep2 proteins cleave origin DNA in vitro and have nucleotidyl transfer activity. The noncoding region of all identified FBNYV DNA components contains GC-rich inverted repeats flanking a highly conserved AT-rich sequence of 11 nt (48) (Fig. 3), sequences that potentially form a stem-loop on ssDNA and are supposed to be part of the origin for initiation of RCR. To prove that the FBNYV Rep proteins possess origin DNA cleavage and nucleotidyl transfer activity, the Rep1 and Rep2 proteins were expressed in E. coli, purified, and used for in vitro assays. Oligonucleotides corresponding to the inverted repeat sequences in the ori regions of rep1 and rep2 components of FBNYV-Sy were used as substrates for the cleavage by purified Rep proteins. The precise position of the cleavage was determined by incubating a 5'-end-labeled ori1-specific oligonucleotide, 1a, with Rep1 protein. The results presented in Fig. 4B demonstrate that cleavage occurred between bases 7 and 8 of the consensus nonamer TAGTATTAC. The cleavage product migrated to the same position as a synthetic marker oligonucleotide with a 3'-OH, representing the 23-nt sequence from the 5' end of the substrate to the presumed cleavage position. Both migrated slightly slower than the corresponding Maxam and Gilbert fragment carrying a phosphate at its 3' end. Hence, we infer that the cleavage product has a free 3' hydroxyl group. Figure 4A schematically illustrates the nucleotidyl transfer reaction. When 5'-labeled oligonucleotide 1a and nonlabeled oligonucleotide 1b were incubated with Rep1 or Rep2 protein, two labeled products appeared (Fig. 4C). These are the 5' cleavage product of oligonucleotide 1a and a new recombinant oligonucleotide, corresponding to the joining of the labeled 5' cleavage product of oligonucleotide 1a and the 3' cleavage product of oligonucleotide 1b (transfer product). In addition, Fig. 4C illustrates that both Rep1 and Rep2 proteins cleave the ori1 substrate and that approximately 10-fold more Rep2 than Rep1 was required to obtain comparative amounts of reaction products. At the moment, we cannot be sure whether this difference reflects intrinsic features of the two proteins or whether it results from the strong overproduction of Rep2 and a concomitant partial inactivation of the protein due to aggregation or misfolding. Nevertheless, both Rep proteins showed similar in vitro DNA cleavage and nucleotidyl transfer efficiencies on sets of oligonucleotides corresponding to the replication origins of the rep1 and rep2 components, as well as on substrate oligonucleotides representing the tomato yellow leaf curl geminivirus (TYLCV) origin of replication (data not shown). The cleavage reaction was strictly dependent on the presence of divalent cations (2.5 mM Mg²⁺ or Mn²⁺). The addition of 10 µM ATP neither stimulated nor inhibited the reaction (data not shown).

View larger version (32K): [in this window] [in a new window]   FIG. 3.   Comparison of the replication origin sequences of 11 FBNYV DNAs. The origin DNA sequences of 11 FBNYV components (C1 of FBNYV-Sy and C2 to C11 of FBNYV-Eg) are aligned. Inverted repeat sequences (open horizontal arrows) potentially forming a stem-loop are boxed. The vertical arrow indicates the position of cleavage by Rep protein. Conserved sequences shared by C2, C3, C4, C5, C6, C8, and C10 are indicated by small boxes, and the orientation of iteron-like sequence repeats is indicated by horizontal solid arrows. FBNYV component designations are on the left.

View larger version (51K): [in this window] [in a new window]   FIG. 4.   FBNYV Rep1 and Rep2 proteins possess in vitro cleavage and nucleotidyl transfer activities. (A) Scheme of the cleavage and nucleotidyl transfer reactions. The 5' 32P-labeled oligonucleotide 1a (35 nt [TGGACCAAGGCGGGTATAGTATTACCCCGCCTTGG]) (underlined inverted repeat sequences are indicated in the figure by filled boxes) is cleaved by Rep protein giving rise to a 5' 23-nt labeled and a 3' 12-nt nonlabeled product. Analogously, nonlabeled oligonucleotide 1b (36 nt [GGCGGGTATAGTATTACCCCGCCTTGGAACACCCTC]) is cleaved to yield a 5' product of 15 nt and a 3' product of 21 nt. When both oligonucleotides 1a and 1b are treated simultaneously with Rep protein, their 5' and 3' cleavage products are joined, giving rise to the recombinant molecules of the nucleotidyl transfer reaction, the larger of which is 44 nt and carries the 5' 32P label. (B) Identification of FBNYV origin cleavage position. The 5' 32P-labeled oligonucleotide 1a was subjected to the G+A-specific chemical cleavage according to Maxam and Gilbert (58) or incubated with Rep1 protein. The fragment indicated by an arrow represents the Rep1-specific 5'-terminal cleavage product. M-23 (TGGACCAAGGCGGGTATAGTATT) is a 32P-labeled oligonucleotide identical to the 5' cleavage product. (C) Separation of Rep-mediated DNA cleavage and nucleotidyl transfer products by denaturing polyacrylamide gel electrophoresis. A mixture of labeled oligonucleotide 1a and nonlabeled oligonucleotide 1b was incubated with 70 ng of Rep1 or 600 ng of Rep2 protein. The products of the resulting cleavage and nucleotidyl transfer reactions are indicated. (D) Alteration of the catalytic tyrosine Y78F or Y79F abolishes cleavage and nucleotidyl transfer activity of the Rep1 or Rep2 proteins, respectively. Cleavage and nucleotidyl transfer reactions were performed with labeled oligonucleotide 1a and nonlabeled oligonucleotide 1b by using the following Rep proteins: 1WT and 2WT, wild-type Rep1 and Rep2 proteins; 1Y and 2Y, Rep1Y78F and Rep2Y79F proteins; 1K and 2K, Rep1K177A and Rep2K187A proteins, respectively. Since Rep2K187A protein was available in a limited amount, only 70 ng of wild-type Rep2 or Rep2K187A protein was used in this assay. Hence only faint bands represent cleavage and transfer products. To ensure that Rep2Y79F has no cleavage and nucleotidyl transfer activity, 500 ng of protein was used for this reaction. It is noteworthy that Rep2K187A has the same (low) specific cleavage and joining activities as wild-type Rep2 protein.

View larger version (73K): [in this window] [in a new window]   FIG. 5.   Replication of mutated rep components in N. benthamiana leaf discs. Southern hybridization of total DNA extracted from N. benthamiana leaf discs, inoculated with agrobacteria carrying pBin19 with redundant copies of FBNYV rep components. 1WT and 2WT, wild-type rep1 and rep2 components, respectively (FBNYV-Sy); 1Y and 2Y, rep1 and rep2 components expressing Rep1Y78F and Rep2Y79F proteins, respectively; 1K and 2K, rep1 and rep2 components expressing Rep1K177A and Rep2K187A proteins, respectively. DNA forms are marked the same way as in Fig. 2.

FBNYV Rep1 and Rep2 proteins possess ATPase activity essential for DNA replication. The deduced amino acid sequence of the FBNYV Rep proteins contains putative nucleoside triphosphate (NTP)-binding motifs GxxGxxGKT/S (see Fig. 1) that differ slightly from the canonical P-loop sequence GxxxxGKT/S (77). In order to determine whether the proteins possess ATPase activity, the hydrolysis of [-³²P]ATP by purified Rep1 or Rep2 proteins was assayed. Both proteins were shown to possess ATPase activity with a K_m in the micromolar range similar to that of TYLCV Rep protein (22). K_m and V_max values for ATP hydrolysis are shown in Fig. 6A. Rep2 protein was a more active ATPase than Rep1. The ATPase activity of both proteins was not stimulated by ssDNA (data not shown). When the lysine residues in the P-loop of the NTP binding site of Rep1 protein (K177) or Rep2 protein (K187) were altered by site-directed mutagenesis into alanine, the ATPase activities of both proteins dropped below the level of detection (Fig. 6B). However, the DNA cleavage and nucleotidyl transfer activity of the mutant Rep proteins was not influenced by the alterations in the P-loop (Fig. 4D).

View larger version (28K): [in this window] [in a new window]   FIG. 6.   FBNYV Rep1 and Rep2 proteins possess ATPase activity. About 100 ng of Rep1 or Rep2 protein was incubated at room temperature with 5 nM [-32P]ATP and various (5 to 50 µM) concentrations of nonlabeled ATP and incubated for 5 min (A). Assays of panel B were incubated for 20 min with 10 µM ATP as nonlabeled substrate; the resulting products were separated on TLC-PEI plates. (A) Determination of Km and Vmax for ATP hydrolysis. The amount of 32Pi liberated was calculated upon analysis of chromatograms with a PhosphorImager. The velocity of the reactions is measured as the amount of total ATP hydrolyzed per minute and is displayed in a double-reciprocal Lineweaver-Burk plot against the ATP concentration (1/[ATP]). The Km of Rep1 was 15 ± 5 µM, and the Vmax was 25 ± 10 nmol min1 mg1. The Km of Rep2 was 80 ± 10 µM and the Vmax was 150 ± 30 nmol min1 mg1. (B) Alteration of the conserved lysine in the P-loop abolishes ATPase activity of Rep1 and Rep2 proteins. Autoradiography of a chromatogram of ATP hydrolysis by the following proteins was done: 1WT and 2WT, wild-type Rep1 and Rep2, respectively; 1K and 2K, Rep1K177A and Rep2K187A proteins, respectively; 1Y and 2Y, Rep1Y78F and Rep2Y79F proteins, respectively. (), no protein added.

Rep1 and Rep2 proteins do not functionally cross-complement. Since Rep2 protein in vitro cleaved oligonucleotides corresponding to the rep1 component origin (Fig. 4) and vice versa (data not shown), we determined whether a given Rep protein also initiates in trans the replication of a heterologous rep component. We used a mixture of two Agrobacterium strains, one carrying a mutated rep1 component (Y78F or K177A) and the other carrying a wild-type rep2 component, and vice versa, to inoculate N. benthamiana leaf discs. Total DNA was isolated 7 days after inoculation and probed for replication of the respective rep components by Southern hybridization. As is evident from Fig. 7, Rep1 protein could not replace a mutant Rep2 to initiate replication and, conversely, Rep2 protein could not substitute for a mutant Rep1. Moreover, rep11 also failed to complement a mutated rep1 component, indicating that the two Rep proteins are functionally distinct.

View larger version (69K): [in this window] [in a new window]   FIG. 7.   Complementation assays with combinations of wild-type and mutant FBNYV rep components. Different pairs of agrobacteria carrying pBin19 with redundant copies of wild-type and mutant rep1 and rep2 components (FBNYV-Sy) or rep11 were used to coinoculate N. benthamiana leaf discs. The designations of the respective mutant rep components are as in Fig. 5. The blots were hybridized with probes specific for rep1 and rep2 components as indicated. Because rep11 shares sequence identity with the rep1 probe, total DNA was treated prior to electrophoresis with HindIII or PstI, which specifically linearized either rep1 or rep11 DNA, respectively. The appearance of a linear fragment upon PstI treatment proves that only rep11 DNA replicated when coinoculated with the mutant (Y78F) rep1 component.

Only FBNYV Rep2 protein triggers replication of other viral genome components and only rep2 DNA is always associated with different virus isolates. Sequence comparison of the noncoding regions of 11 FBNYV DNAs revealed conserved sequences of about 70 nt shared by rep2 and all DNAs encoding proteins other than Rep (non-rep components) (Fig. 3). Apart from the inverted repeat sequence, there is only limited similarity in the origin region between the rep2 component and the other rep components. This observation suggested that Rep2 protein might specifically recognize targets in the origin sequences common to all FBNYV DNAs except the rep1, -7, -9, and -11 components. To verify this hypothesis, we tested whether replication of the non-rep components of FBNYV could be initiated in trans by coinoculating each of them in pairwise mixtures with one of the five rep components to N. benthamiana leaf discs. Agroinoculation of the various combinations of DNAs to be tested, DNA extraction, and probing for replication by Southern hybridization were done as described above. Figure 8 shows the summary of a series of replication assays. Only when coinoculated with agrobacteria carrying the rep2 component, was replication of each of the six other FBNYV components observed. Coinoculations with each of the four other rep components did not lead to replication of any non-rep DNA. This proves that only Rep2 protein catalyzes the initiation of DNA replication of all FBNYV genome components that encode viral proteins other than Rep. Therefore, Rep2 is the master Rep protein of FBNYV.

View larger version (47K): [in this window] [in a new window]   FIG. 8.   Rep2 protein mediates replication of other (non-rep) FBNYV components. Shown are Southern blots of DNAs extracted from leaf discs coinoculated with pairwise combinations of agrobacteria carrying the respective FBNYV DNA to be assayed for replication (indicated at the bottom) and agrobacteria carrying the respective rep component (indicated at the top). All DNAs, except C1-Sy, are from FBNYV-Eg. Membranes were hybridized with probes specific for the components indicated below each blot.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

All FBNYV rep components encode functional replication initiator proteins. The recent characterization of the nanoviruses FBNYV and MDV has led to the identification of the largest number of different DNA components among the nanoviruses and, most strikingly, of four (MDV) and five (FBNYV) different DNAs that potentially code for replication initiator proteins (50, 70) (Fig. 1). This prompted us to study whether all five FBNYV rep components encode functional Rep proteins and, if so, which of these Rep proteins supports replication initiation of the other genome components. Here, we have shown that, in fact, all rep components direct the synthesis of functional replication initiator proteins capable of triggering the autonomous replication of their respective DNA component in plant cells (Fig. 2). Moreover, the formation of ssDNA indicates that nanoviruses, here represented by FBNYV, also replicate via a rolling-circle mechanism.

Nanovirus replication is triggered by a master Rep protein. One of the key steps during the initiation of DNA replication is origin recognition (51). Geminivirus Rep proteins specifically recognize the origin of their cognate genome and do not initiate DNA replication of other geminivirus genomes, as has been shown for two different geminiviruses with a single genomic DNA, TYLCV (46) and beet curly top virus (BCTV) (18, 74). Such recognition specificity is particularly relevant for the geminiviruses with a bipartite genome. Their two genomic DNAs (DNA-A and -B) share a common region that contains the replication origin and multiple Rep protein binding sites. The binding sites are recognized by a given Rep protein and have been mapped to a 13-bp element containing 5-bp direct repeats (iterons) for tomato golden mosaic virus (24) and BCTV (17, 74). The iterons have to be identical on a given pair of DNA-A and -B to warrant correct replication initiation and multiplication of both DNAs (25); they differ, however, among different geminiviruses (1, 3, 74).