The DNA β satellite component associated with ageratum yellow vein disease encodes an essential pathogenicity protein (βC1)

Virology 324 (2004) 37 – 47 www.elsevier.com/locate/yviro

The DNA h satellite component associated with ageratum yellow vein disease encodes an essential pathogenicity protein (hC1) Keith Saunders, Alexandra Norman, Sebastien Gucciardo, and John Stanley * Department of Disease and Stress Biology, John Innes Centre, Colney, Norwich NR4 7UH, UK Received 13 January 2004; returned to author for revision 3 March 2004; accepted 12 March 2004

Abstract Ageratum yellow vein disease (AYVD) is caused by the geminivirus ageratum yellow vein virus (AYVV) and an associated DNA h satellite. We have mapped a DNA h transcript to a highly conserved open reading frame (hC1 ORF). The most abundant transcript 5Vterminus is located 8 bases upstream of the hC1 ORF putative initiation codon while the transcript terminates at multiple sites downstream from the putative termination codon. Disruption of hC1 protein expression by the introduction of an internal nonsense codon prevented infection of the AYVV – satellite complex in ageratum and altered the phenotype in Nicotiana benthamiana to that produced by AYVV alone although the mutant was maintained in systemically infected tissues. Modification of the putative initiation codon to a nonsense codon produced an intermediate phenotype in N. benthamiana and a mild yellow vein phenotype in ageratum, suggesting that hC1 protein expression could be initiated from an alternative site. N. benthamiana plants containing a dimeric DNA h transgene produced severe developmental abnormalities, vein-greening, and cell proliferation in the vascular bundles. Expression of hC1 protein from a potato virus X (PVX) vector also induced abnormal plant growth. Our results demonstrate that the satellite encodes at least one protein that plays a major role in symptom development and is essential for disease progression in ageratum, the natural host of the AYVD complex. D 2004 Elsevier Inc. All rights reserved. Keywords: Geminivirus; Begomovirus; AYVV; Satellite; DNA h; Transcripts; Ageratum

Introduction Members of the genus Begomovirus in the family Geminiviridae have circular single-stranded DNA genomic components. The first begomoviruses to be studied in detail had bipartite genomes (the genomic components being called DNA A and DNA B), typified by African cassava mosaic virus (ACMV, previously Cassava latent virus; Stanley, 1983; Stanley and Gay, 1983) from the Old World and Tomato golden mosaic virus (TGMV) from the New World (Hamilton et al., 1983). DNA A encodes genes required for amplification and encapsidation of the viral DNA, but DNA B is required for intra- and intercellular movement of the virus (reviewed by Hanley-Bowdoin et al., 1999). Subsequently, the monopartite begomoviruses Tomato yellow leaf curl virus (TYLCV; Kheyr-Pour et al., 1991; Navot et al., 1991) and Tomato leaf curl virus (ToLCV; Dry et al., 1993) * Corresponding author. Department of Disease and Stress Biology, John Innes Centre, Norwich Research Park, Colney, Norwich NR4 7UH, UK. Fax: +44-1603-450045. E-mail address: [email protected] (J. Stanley). 0042-6822/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.virol.2004.03.018

were isolated from tomato. Their single genomic components, analogous to DNA A of the Old World bipartite begomoviruses, were shown to be sufficient to produce a symptomatic infection in this host, implying that these viruses have overcome the need for the movement genes encoded by DNA B. However, monopartite begomoviruses such as Ageratum yellow vein virus (AYVV; Tan et al., 1995) and Cotton leaf curl Multan virus (CLCuMV; Briddon et al., 2000) were unable to produce a symptomatic infection in the hosts from which they were isolated, indicating the requirement of an additional factor. From the analysis of small subgenomic-sized components associated with ageratum yellow vein disease (AYVD; Saunders and Stanley, 1999; Saunders et al., 2001; Stanley et al., 1997) and cotton leaf curl disease (CLCuD; Mansoor et al., 1999), novel satellite components (named DNA h) were isolated and characterized from affected plants (Saunders et al., 2000). Infectivity studies using cloned begomovirus and satellite components demonstrated that DNA h plays an essential role in the maintenance of both diseases (Briddon et al., 2001; Saunders et al., 2000). It is remarkable that such disease complexes have remained undiscovered until quite

38

K. Saunders et al. / Virology 324 (2004) 37–47

recently, particularly as they have since been shown to be associated with several important diseases that have been recognized for many years (Briddon et al., 2003; Jose and Usha, 2003; Saunders et al., 2003; Zhou et al., 2003; reviewed by Mansoor et al., 2003). Comparison of the sequences of DNA h components associated with distinct begomoviruses has identified several common features (Mansoor et al., 2003; Fig. 1A). They contain a highly conserved region, called the satellite conserved region (SCR), adjacent to a putative stem-loop structure containing the nonanucleotide TAATATTAC that has been implicated in the initiation of rolling circle replication (Laufs et al., 1995; Stanley, 1995). The satellites also have a conserved complementary-sense open reading frame (ORF), referred to here as ORF hC1. An A-rich region, located between ORF hC1 and the SCR, may have been responsible for the size adaptation of DNA h from a preexisting progenitor component of unknown origin (Saunders et al., 2000). AYVV can produce a symptomatic infection in a highly permissive host such as Nicotiana benthamiana, but causes a sporadic asymptomatic infection in ageratum, in which the begomovirus accumulates to only low levels (Saunders and Stanley, 1999). The begomovirus accumulates to normal

Fig. 1. Identification of a DNA h transcript. (A) Features of DNA h (shown as a linear molecule for clarity) including the complementary-sense hC1 ORF and its putative TATA box, the A-rich region, and satellite conserved region (SCR), adjacent to the stem-loop structure containing the highly conserved nonanucleotide sequence TAATATTAC. Nucleotide numbering starts from the putative nick site within this sequence in accordance with the convention accepted for geminiviruses. The locations of probes I and II are indicated. (B) Northern blot analysis of polyadenylated RNA from healthy N. benthamiana plants (H) and plants infected using cloned AYVV and DNA h components (I), probed using a full-length copy of the satellite DNA. (C) Northern blot analysis of polyadenylated RNA from a nontransformed N. benthamiana plant (N) and a plant (line 2.1) transformed with a dimeric satellite transgene (T), probed using either a full-length copy of the satellite DNA or fragments I and II. The positions of RNA markers (BRL) are indicated. Relative amounts of samples loaded in each lane are indicated by rRNA markers stained with ethidium bromide.

levels in ageratum in the presence of DNA h, suggesting that the satellite functions either by facilitating the replication or movement of the begomovirus or by suppressing a host defense mechanism such as virus-induced gene silencing. Despite its importance to the disease phenotype, there is still no information available concerning even the most fundamental properties of the satellite. Here, we have investigated the AYVD satellite using a combination of transcript mapping, mutagenesis, and gene expression both in transgenic plants and from a potato virus X (PVX) vector. We demonstrate that the satellite encodes at least one protein that plays an important role in the pathogenicity of the begomovirus – satellite disease complex.

Results AYVD DNA b transcript mapping N. benthamiana plants were infected by agroinoculation of cloned AYVV and DNA h components, and RNA was isolated from healthy and infected plants. Polyadenylated RNA purified from the infected plant sample produced one major band when analyzed by Northern blotting using a fulllength probe specific to DNA h (Fig. 1B). The band was invariably accompanied by a diffuse background signal ranging in size up to that of the full-length DNA h component (1347 nucleotides). This was attributed to the presence of satellite single-stranded DNA, co-purifying with the polyadenylated RNA as a result of its A-rich region, that had not been completely degraded even after two successive treatments with DNase I. In an attempt to overcome this problem, N. benthamiana was transformed with a tandem repeat of DNA h (Fig. 2A). As the transgene contains an intact copy of DNA h, it was anticipated that expression from this linear template would be similar to that from the circular satellite DNA. Furthermore, because the satellite is unable to become mobilized from the transgene in the absence of the replication functions provided by the begomovirus, RNA extracted from the transgenic plant will not be contaminated with extrachromosomal copies of satellite singlestranded DNA. Two independently transformed N. benthamiana lines were produced. Line 2.1 was associated with severe developmental abnormalities that caused distortion of the stems, petioles, and leaves (Fig. 2B). Line 2.2 showed a similar but milder phenotype. Following self-fertilization of the original transformant of line 2.1 and seed germination in the presence of 500 Ag ml 1 kanamycin, 7/86 F1 plants became bleached. However, Southern blot analysis of the genomic DNA of selected plants, digested with SpeI, either failed to detect satellite DNA (Fig. 2C, lane 1) or produced a fragment corresponding to full-length DNA h and an additional larger fragment (lanes 2 – 4). As the DNA h transgene has lost one flanking SpeI site (Fig. 2A), digestion with this

K. Saunders et al. / Virology 324 (2004) 37–47

39

Fig. 2. Transformation of N. benthamiana with a DNA h dimer. (A) Structure of the satellite transgene. One of the flanking SpeI sites (bracketed) has been removed by cloning two copies of the satellite as SpeI fragments into a vector digested with SpeI and XbaI. (B) Phenotype associated with the satellite transgene in line 2.1. (C) Southern blot analysis of genomic DNA isolated from F1 plants of line 2.1 (lanes 1 – 4) and line 2.2 (lanes 5 – 8), digested with SpeI. The position of full-length linearized DNA h is shown. The plant corresponding to lane 1 did not contain the transgene and was phenotypically normal. (D) A leaf from plant line 2.1 showing vein greening (left) compared with a non-transformed plant (right). (E) Section through vascular bundles of transformed (left) and non-transformed (right) plants.

restriction endonuclease would be expected to produce two full-length copies of the component, one of which will be associated with adjacent genomic DNA sequences,

for each independently integrated copy. All F2 generation plants in line 2.1/1 and 62/147 plants in line 2.1/3 (from which samples in lanes 1 and 3, respectively, were

40

K. Saunders et al. / Virology 324 (2004) 37–47

derived) were kanamycin sensitive, consistent with a single copy insert in the original transformant. Selected F1 plants in line 2.2 produced several fragments in addition to that corresponding to full-length DNA h (Fig. 2C, lanes 5 – 8), suggesting multiple copy inserts in the original transformant. Consistent with this, all F1 plants in line 2.2 and F2 plants in lines 2.2/5 and 2.2/7 (from which samples in lanes 5 and 7, respectively, were derived) were kanamycin resistant, suggesting that the mild phenotype associated with this line may be due to transgene silencing. The leaves of transformed plants (line 2.1) showed a vein-greening phenotype (Fig. 2D) that is characteristic of AYVV –DNA h infection in this host as well as in other Nicotiana species (Saunders et al., 2001), and thin sections frequently showed abnormal cell division associated with the vascular bundles of both class II and III veins (Roberts et al., 1997) of the transformed plants (Fig. 2E). These proliferating cells may derive from the phloem parenchyma, although this remains to be confirmed.

When analyzed by Northern blotting using a full-length probe specific for DNA h, polyadenylated RNA extracted from transformed N. benthamiana line 2.1 contained a single abundant DNA h-specific transcript of comparable size to that observed in extracts from infected plants (Fig. 1C). The transcript was also detected using a probe spanning the conserved hC1 ORF (probe I), but was not detected with a probe specific for other DNA h sequences (probe II). The transcript mapped to approximately the first 500 nucleotides of DNA h, suggesting that the conserved hC1 ORF may be functional in protein coding. Accordingly, specific primers (5V-RACE/1 and 3V-RACE; Fig. 3B) were designed to PCR-amplify DNA h fragments using a cDNA template synthesized from polyadenylated RNA that had been isolated from N. benthamiana infected by agroinoculation of cloned AYVV and DNA h components. Although the 3V-RACE reaction produced a visible product when analyzed by agarose gel electrophoresis, the product of the 5V-RACE reaction was much less abundant and required re-

Fig. 3. Fine mapping of the DNA h transcript. (A) Agarose gel electrophoresis of PCR fragments produced from 3V-RACE and 5V-RACE reactions. 5V-RACE reactions were produced using primers 5VRACE/1 (lane 1) and 5’RACE/2 (lane 2). The sizes (bp) of marker DNA fragments are indicated (lane M). (B) Complementary-sense nucleotide sequence encompassing the hC1 ORF. The positions of primers used in 3V-RACE and 5V-RACE reactions are shown. The positions and frequency of transcript 5V-termini (.) and 3V-termini (z), consensus TATA box and AATAAA polyadenylation signal (underlined bold font), and cryptic polyadenylation signals (underlined) are indicated.

K. Saunders et al. / Virology 324 (2004) 37–47

amplification using either 5V-RACE/1 or 5V-RACE/2 primers (Fig. 3A). Gel-purified DNA fragments were cloned and the transcript termini were identified by sequence analysis (Fig.

41

3B). The majority of 5V-termini (19/22 clones) mapped to the A residue at position 555, located 8 nucleotides upstream of the putative initiation codon of the hC1 ORF and

Fig. 4. Mutagenesis of the hC1 ORF. (A) The position and codon changes in mutants hC1mut1 and hC1mut2. (B) Top panel, left to right. N. benthamiana plants inoculated with AYVV, either alone or in the presence of hC1mut1, hC1mut2, and wild-type DNA h. Bottom panels, left to right. N. benthamiana plants inoculated with AYVVC4mut, either alone or in the presence of wild-type DNA h, a healthy N. benthamiana plant, and an ageratum plant inoculated with AYVV and hC1mut1. Plants were photographed 3 weeks after inoculation. (C) Southern blot analysis of AYVV DNA (top panel) and DNA h (bottom panel) extracted either from N. benthamiana plants inoculated with AYVV and DNA h (lanes 1 and 2), AYVV and hC1mut1 (lanes 3, 4, 13, and 14), AYVVC4mut and DNA h (lanes 5 and 6), AYVV (lanes 7 and 8), AYVVC4mut (lanes 9 and 10), and AYVV and hC1mut2 (lanes 11 and 12), or from ageratum plants inoculated with AYVV and DNA h (lanes 15 and 16), AYVV and hC1mut1 (lanes 17 and 18), and AYVVC4mut and DNA h (lanes 19 and 20). DNA was extracted from plants either 14 days (lanes 1 – 10 and 15 – 20) or 20 days (lanes 11 – 14) post-inoculation. Approximately equal amounts of DNA were loaded in each lane. Exposure times for lanes 1 – 14 and 15 – 20 were different. The positions of single-stranded forms of AYVV DNA and DNA h are indicated.

42

K. Saunders et al. / Virology 324 (2004) 37–47

34 nucleotides downstream of a consensus TATA box. Transcript 5V-termini at positions 547 and 551 (one and two clones, respectively) probably reflect incomplete cDNA synthesis on the polyadenylated RNA template. In contrast, the 3V-termini mapped to multiple sites between positions 68 and 157, although termination at position 128 was most frequently observed (6/19 clones). All 3V-termini mapped downstream of the hC1 ORF.

Table 1 Infectivity of AYVV and DNA h mutants Host

Inoculum

Infected/ inoculated (number of experiments)

Symptoms

N. benthamiana

AYVV

25/26(4)

AYVV + DNA h

34/36(7)

AYVV + hC1mut1

43/43(5)

AYVV + hC1mut2

12/12(2)

AYVVC4mut

21/26(3)

AYVVC4mut + DNA h AYVV + DNA h

24/24(3)

AYVV + hC1mut1

20/28(3)

AYVV + DNA h

34/45(6)

AYVV + hC1mut1

20/47(5)

AYVV + hC1mut 2 AYVVC4mut + DNA h

0/24(2) 12/17(2)

severe upward leaf roll and vein swelling severe downward leaf curl downward leaf curl followed by upward leaf roll and vein swelling severe upward leaf roll and vein swelling mild downward leaf curl severe downward leaf curl severe downward leaf curl mild downward leaf curl extensive yellow veins sporadic yellow veins no symptoms sporadic yellow veins

Mutagenesis of the bC1 ORF Two DNA h mutants were constructed in which the hC1 ORF was disrupted by the introduction of nonsense codons, one replacing the putative initiation codon (hC1mut1) and the other replacing a serine residue after the first 20 amino acids (hC1mut2) (Fig. 4A). Plants were routinely challenged by agroinoculation using partial repeats of AYVV DNA and DNA h mutants on the same binary vector to maximize infectivity (Saunders et al., 2000). As previously demonstrated (Saunders et al., 2000, 2001), coinfection with AYVV and DNA h produced a severe downward leaf curl and vein-darkening phenotype in N. benthamiana and Nicotiana glutinosa, and an extensive yellow vein phenotype in ageratum, and the majority of inoculated plants became infected (Table 1; Fig. 4B). In contrast, co-inoculation of plants with AYVV and hC1mut1 produced a mild downward leaf curl in N. benthamiana and N. glutinosa that subsequently developed into an upward leaf roll phenotype in N. benthamiana resembling symptoms induced by AYVV alone, and sporadic yellow veins in ageratum (Fig. 4B). This combination was highly infectious in the Nicotiana spp., but fewer ageratum plants became infected compared with AYVV and wild-type DNA h. Southern blot analysis indicated that hC1mut1 accumulated to high levels in systemically infected N. benthamiana tissues (Fig. 4C, lanes 3, 4, 13, and 14) and was maintained in ageratum (lanes 17 and 18), although accumulation was variable in this host owing to the sporadic nature of the phenotype. AYVV and hC1mut2 produced upward leaf roll symptoms in N. benthamiana, but this clone combination was unable to produce a symptomatic infection in ageratum. Because the hC1mut2 phenotype in N. benthamiana was indistinguishable from that produced by AYVV alone, plants were tested for the presence of the mutant by Southern blot analysis. Of 12 plants analyzed, 10 accumulated the mutant to wild-type DNA h levels (e.g., Fig. 4C, lane 11). The presence of the mutations in the hC1mut1 and hC1mut2 progeny was verified by PCRamplification of fragments of DNA h encompassing the mutations and sequence analysis. Mutagenesis of the AYVV C4 ORF Expression of the Beet curly top virus (BCTV) C4 ORF is known to have a profound effect on symptom

N. glutinosa

A. conyzoides

24/24(3)

development in N. benthamiana and sugar beet (Latham et al., 1997; Stanley and Latham, 1992). To examine the functional equivalence of the AYVV counterpart, mutant AYVVC4mut was constructed in which expression of the C4 ORF was disrupted after eight amino acids by the introduction of a nonsense codon. The mutation did not affect the amino acid encoded by the overlapping Rep ORF. The mutation altered the AYVV phenotype from severe upward leaf roll and vein swelling to mild downward leaf curl in N. benthamiana, resembling the BCTV mutant phenotype (Table 1, Fig. 4B). Co-agroinoculation of AYVVC4mut and DNA h produced a severe downward leaf curl phenotype that was indistinguishable from the phenotype associated with wild-type AYVV and DNA h. Co-agroinoculation of ageratum with AYVVC4mut and DNA h produced sporadic yellow veins, resembling the phenotype associated with AYVV and hC1mut1. Southern blot analysis showed that AYVVC4mut could accumulate to high levels in systemically infected N. benthamiana and was able to maintain DNA h to levels similar to those associated with wild-type AYVV (Fig. 4C, lanes 5 and 6). Both AYVVC4mut and DNA h accumulated in systemically infected ageratum tissues (lanes 19 and 20), although, as observed for AYVV and hC1mut1,

K. Saunders et al. / Virology 324 (2004) 37–47

levels were variable owing to the sporadic nature of the phenotype. bC1 expression in N. benthamiana We have shown that expression from an integrated dimer of DNA h resulted in severe developmental abnormalities in transgenic N. benthamiana plants (Fig. 2). Our transcript mapping and mutagenesis studies suggested that the hC1 protein was responsible for this phenotype. To test this idea, a fragment encompassing the hC1 ORF was inserted into a PVX expression vector to produce the recombinant PVX-hC1 (Fig. 5A). N. benthamiana plants systemically infected with PVX exhibit mild veinal chlorosis. In contrast, plants infected with PVX-hC1 produced severely distorted stems and leaves (Fig. 5B), indicating that hC1 protein expression, when removed from the context of the AYVV genome, can have a marked effect on development. Leaves of plants infected with PVX-hC1 exhibited a vein-greening phenotype resembling the symptom associated with the AYVV –DNA h complex in this host.

Fig. 5. Expression of hC1 protein from a PVX vector. (A) PVX-hC1. The hC1 protein is expressed from the upstream copy of the duplicated coat protein (CP) promoter. (B) The phenotype of PVX-hC1 infection in N. benthamiana, 3 weeks after inoculation.

43

Discussion Begomovirus – satellite disease complexes were only recently discovered (Saunders et al., 2000), yet many examples have now been described for diseases of a wide variety of crops, weeds, and ornamental plants (reviewed by Mansoor et al., 2003). Although a wealth of information is available concerning begomovirus gene expression and function (reviewed by Hanley-Bowdoin et al., 1999), nothing is currently known about the satellites and their role in the infection process. Here, we have taken the first step towards understanding satellite function by analyzing gene expression from DNA h associated with yellow vein disease of ageratum (Saunders et al., 2000). We have identified a single major transcript that maps to a region of DNA h encompassing the hC1 ORF, predicted to be functional on theoretical grounds (Saunders et al., 2000) and highly conserved in all satellites characterized to date (Briddon et al., 2003). Fine mapping has identified the transcript start at nucleotide 555, appropriately located 34 nucleotides downstream of a consensus TATA box (Fig. 3). Assuming that initiation of translation occurs at the first in-frame AUG codon (corresponding to nucleotides 545 –547), the transcript will have a short leader sequence of eight nucleotides. Transcript 3V-termini map uncharacteristically to several different locations although each occurs downstream of the hC1 ORF UAA termination codon at nucleotides 191 – 193, implying that each transcript has the capacity to encode full-length hC1 protein. 3V-Termini at nucleotides 68, 72, 77, and 83 are located downstream of an AAUAAA motif (nucleotides 104– 109) that frequently occurs 10 – 30 nucleotides upstream of the polyadenylation site. Cryptic motifs that differ by a single nucleotide substitution, at nucleotides 137 – 142, 140 – 145 (both AAUAAU), and 190 – 195 (UAUAAA), may be responsible for the alternative 3V-termini. Some alteration to the AAUAAA motif in this way may be tolerated, although this is usually associated with less efficient processing (reviewed by Edmonds, 2002). Although the majority of 3V-termini are represented by a single clone, termination at nucleotide 128 is represented by six clones and is flanked by 3V-termini at nucleotides 127 and 130, suggesting that the overlapping duplicated AAUAAU cryptic motif may provide a functional processing signal. Whether such dispersed 3V-termini is a feature of DNA h awaits transcript mapping of other distinct satellites. Although we have demonstrated that a transcript maps across the hC1 ORF, we cannot rule out the possibility that AYVD DNA h produces additional transcripts that were either below the level of detection or masked by the relatively abundant hC1 ORF transcript when screened by Northern blot analysis. The hC1 ORF has the capacity to encode a 13.8-kDa protein comprising 118 amino acids (Fig. 3). Of the satellites that have so far been shown to be fully functional in their respective hosts, those associated with cotton leaf curl disease (Briddon et al., 2001) and eupatorium yellow vein

44

K. Saunders et al. / Virology 324 (2004) 37–47

disease (Saunders et al., 2003) have the capacity to encode similar-sized proteins, while those associated with bhendi yellow vein mosaic disease (Jose and Usha, 2003) and tomato yellow leaf curl disease (Zhou et al., 2003) have the capacity to encode 22 and 8 additional amino-terminal amino acids, respectively, although internal AUG codons corresponding to the proposed AYVD hC1 ORF initiation codon are also present. The hC1 proteins associated with these diverse diseases show between 29% and 50% amino acid identity, suggesting conservation of biological function despite diversification due to host adaptation and geographic isolation (Briddon et al., 2003). Disruption of the hC1 ORF by introducing an internal inframe nonsense codon did not prevent transreplication and systemic movement of the hC1mut2 mutant by AYVV in N. benthamiana. However, it removed any influence of the satellite on symptom development in this host and prevented symptomatic infection of ageratum. Hence, the hC1 protein is an important pathogenicity factor that plays an essential role in the proliferation of the AYVV – DNA h complex in its natural host. Interestingly, modification of the proposed hC1 ORF initiation codon to a nonsense codon did not completely eliminate DNA h activity as judged by the infectivity of AYVV and hC1mut1 in ageratum and the production of mild symptoms in this host. A similar mutation in the hC1 ORF of a satellite associated with Tomato yellow leaf curl China virus (TYLCCNV) also produced attenuated symptoms in N. benthamiana, although the mutant was not tested in either tomato or tobacco, the host from which the satellite was isolated (Zhou et al., 2003). In this case, a downstream in-frame AUG codon could potentially initiate expression of a protein truncated by only seven amino-terminal amino acids. In contrast, all downstream in-frame AUG codons of the AYVD hC1 ORF occur towards the 3V-terminus, the first of which encodes a methionine at position 73, suggesting that a truncated version of the hC1 protein may be expressed from a noncanonical initiation codon. In eukaryotes, the ribosome scanning mechanism usually ensures that the first AUG codon is recognized by the translational initiation complex. Nonetheless, a growing number of examples have been documented for which translation initiates at alternative codons. This includes nuclear and mitochondrial genes in nematode worms (Ko and Chow, 2003; Okimoto et al., 1990), chloroplast genes in green alga (Hirose et al., 1999), plant nuclear genes (Kobayashi et al., 2002; Riechmann et al., 1999), and animal and plant virus genes (Carroll and Derse, 1993; Corcelette et al., 2000; Futterer et al., 1996; Muralidhar et al., 1994; Prats et al., 1989; Shirako, 1998; van Eyll and Michiels, 2002). Expression from non-canonical initiation codons is usually low compared with downstream in-frame AUG codons that are recognized by a leaky scanning mechanism. The mild phenotype associated with hC1mut1 is consistent with a reduction in the level of expression of a truncated hC1 derivative. Non-canonical initiation codons usually differ from the AUG initiation

codon by alteration of a single nucleotide, and several such codons occur immediately downstream of the proposed hC1 ORF initiation codon. However, these will be embedded in a different sequence context to that of the bona fide initiation codon that may not be favorable for their recognition by the translational initiation complex. Also, there is no strong secondary structure downstream of the modified initiation codon to facilitate recognition of a non-canonical initiation codon (Kozak, 1990). Clearly, additional work is necessary to elucidate how a truncated and possibly functionally impaired hC1 protein is expressed from hC1mut1. Our original infectivity studies in N. benthamiana and ageratum demonstrated a role for both AYVV and DNA h in symptom development (Saunders et al., 2000), although it was not known at the time if the satellite had a direct effect on the phenotype or if it modulated the level of a pathogenicity factor encoded by AYVV. The phenotype of N. benthamiana plants transformed with a tandem repeat of the satellite (Fig. 2) clearly demonstrates that DNA h makes a direct contribution to symptom development of the AYVV – DNA h complex. Mutagenesis of the hC1 ORF (Fig. 4) and transient expression from a PVX vector (Fig. 5) indicate that the hC1 protein plays a major role in this process. Expression of the DNA h gene in transgenic N. benthamiana plants and the transient expression of hC1 in PVX-infected plants caused severe developmental abnormalities including distortion of stems, petioles, and leaf veins, and upward curling of the leaf margins. In both cases, leaves exhibited very characteristic dark green veins, a phenotype that is associated with leaf curl diseases of cotton and hibiscus, and hollyhock leaf crumple disease (HLCD), all of which are caused by similar begomovirus – satellite complexes (Briddon et al., 2001, 2003). The HLCD phenotype is caused by vein thickening and the replacement of spongy parenchyma with chloroplast-rich pallisade parenchyma (Bigarre´ et al., 2001). Ectopic cell division observed in phloem parenchyma cells of HLCDaffected hollyhock (Bigarre´ et al., 2001) and reported here in N. benthamiana plants transgenic for a dimer of DNA h resembles the phenotype induced by expression of BCTV C4 protein (Latham et al., 1997). We have also shown that disruption of the AYVV C4 ORF alters the phenotype in N. benthamiana from upward leaf roll and vein swelling to downward leaf curl, previously observed for a BCTV C4 ORF mutant in this host (Stanley and Latham, 1992), suggesting that their C4 proteins have identical functions. That DNA h affects the AYVV phenotype in a similar manner suggests that hC1 and C4 proteins may perform partially redundant functions involving convergent pathways. We have suggested that the hC1 ORF encodes a pathogenicity determinant that suppresses a host defense mechanism (Saunders et al., 2000), and the behavior of ACMV AC4 and TYLCCNV C4 is consistent with such a function (van Wezel et al., 2002). Having identified hC1 as a functional DNA h protein, we are currently investigating its biological function

K. Saunders et al. / Virology 324 (2004) 37–47

by attempting to identify host factors with which it interacts. Its effect on tissue development may provide an insight into cell-cycle regulation and plant developmental processes.

Materials and methods Cloned virus and satellite components and construction of mutants The construction of a partial repeat of AYVV DNA A in pBin19 (pBinAYVVA), a tandem repeat of AYVD DNA h in pBinPlus (pBinAYVVh), and their cloning together on the same binary vector (pBinAYVVAh) have been described by Saunders et al. (2000, 2001). AYVV and DNA h mutants were constructed using a QuikChange site-directed mutagenesis kit (Stratagene) and clones pHNIC419, containing a partial repeat of AYVV DNA A (Tan et al., 1995), and pBSAYVVh/5 (Saunders et al., 2000) as templates, according to the supplier’s instructions. Modification of A to T at position 2419 of AYVV DNA introduced a nonsense codon (TAA) in the C4 ORF after the first eight amino acids, without altering the amino acid encoded by the overlapping Rep ORF, to produce mutant AYVVC4mut. Modification of AT to TA at positions 547/548 of DNA h changed the putative initiation codon of the hC1 ORF to a nonsense codon (TAG) in mutant hC1mut1. Modification of G to T at position 486 introduced a nonsense codon (TAA) in the hC1 ORF after the first 20 amino acids in mutant hC1mut2. The presence of mutations was confirmed by sequencing. Mutant AYVVC4mut was cloned as a HindIII fragment into either HindIII-digested pBinPlus or p B i nAY V V h t o p r o d uc e p B i n AY V V C 4 m u t an d pBinAYVVC4muth, respectively. Mutants hC1mut1 and hC1mut2 were cloned as SpeI dimers into SpeI – XbaIdigested pBluescript SK+ (Stratagene) to produce pBS2hmut1 and pBS2hmut2, respectively. The pBS2hmut1 mut1 insert was excised with BssHII and cloned into the AscI site of pBinAYVVA to produce pBinAYVVAhmut1. The AYVV DNA insert of pHNIC419 was excised with HindIII and cloned into the same site in pBS2hmut2 to produce pBSAYVVAhmut2. The pBSAYVVAhmut2 insert was excised using SacI and SalI and cloned into pBinPlus (van Engelen et al., 1995) at same sites to produce pBinAYVVAhmut2. Infectivity assay Plants were agroinoculated in the stem when 2– 3 weeks old as described by Tan et al. (1995) using Agrobacterium tumefaciens strain GV3850 (Zambryski et al., 1983) transformed with the binary vector constructs described above. Plants were maintained in an insect-proof glasshouse at 25 jC (reduced to 20 jC at night) with supplementary lighting to give a 16-h photoperiod.

45

Expression of hC1 from a PVX vector A fragment encompassing the hC1 ORF was PCR-amplified using the complementary-sense primer TAATCGATG ACTATATCATATAC that introduced a ClaI site (underlined) immediately upstream of the putative initiation codon (bold font) and the virion-sense primer TCGTCGAC TTATACGGTTAC that introduced a SalI site (underlined) immediately downstream of the nonsense codon (complementary to bold font). The fragment was digested with ClaI and SalI, and cloned into the same sites of the PVX vector pP2C2S (Chapman et al., 1992) to produce pP2C2ShC1. RNA transcripts (PVX-hC1) were produced from pP2C2S-hC1 by in vitro transcription after linearization with SacI, and transcripts were mechanically inoculated onto N. benthamiana plants as described (Chapman et al., 1992). Transformation and analysis of N. benthamiana N. benthamiana was transformed with pBinAYVVh, containing a tandem repeat of DNA h, by a leaf disk procedure (Horsch and Klee, 1986; Horsch et al., 1985), and transformants were selected for resistance to 50 Ag ml 1 kanamycin. Following self-fertilization, progenies were tested for antibiotic sensitivity by germinating seeds in the presence of 500 Ag ml 1 kanamycin. The presence of the transgene was verified by Southern blot analysis of genomic DNA isolated using a Nucleon Phytopure plant DNA extraction kit (Amersham Pharmacia). Transgene copy number was estimated from the segregation of kanamycin sensitivity and from restriction analysis of genomic DNA using SpeI. Thin sections through class II and III veins (Roberts et al., 1997) from emerging (2.5 cm length) and fully developed (15 cm length) leaves of a non-transformed N. benthamiana plant and the equivalent veins from transformed plant line 2.1 were prepared, stained, and examined as described by Pinner et al. (1993). Analysis of viral DNAs Nucleic acids were isolated from systemically infected plant material using the procedure of Covey and Hull (1981), fractionated by agarose gel electrophoresis in TNE buffer (40 mM Tris-acetate pH 7.5, 20 mM Na acetate, 2 mM EDTA), and transferred to Hybond-N (Amersham Pharmacia) as described by Saunders and Stanley (1999). Viral DNAs were detected using probes for AYVV DNA A and DNA h (Saunders et al., 2001). Fragments were radiolabelled according to the procedure of Feinberg and Vogelstein (1983). The presence of mutations in the progeny of mutants AYVVC4mut, hC1mut1, and hC1mut2 was verified by PCR-amplification of fragments encompassing nucleotides 2163 – 2557 of DNA A and nucleotides 214 –1166 of DNA h, cloning the fragments into pGEM-T Easy (Promega), and sequence analysis.

46

K. Saunders et al. / Virology 324 (2004) 37–47

Analysis of satellite transcripts RNA was extracted from 5 g newly emerging leaves from healthy and systemically infected N. benthamiana plants using the Concert plant RNA reagent (Invitrogen) according to the manufacturer’s instructions. Contaminating DNA was removed using 3 M Li acetate as described by Covey and Hull (1981). The resulting RNA solution was was treated with 20 u ml 1 DNase I (RNase-free; Amersham Pharmacia) at 37 jC for 15 min in 40 mM Tris – HCl (pH 7.5), 6 mM MgCl2. Polyadenylated RNA was purified using an Oligotex mRNA kit (Qiagen) according to the manufacturer’s instructions. Because of residual viral DNA contamination, the polyadenylated RNA samples were treated for a second time with DNase I and the RNA was purified using an RNeasy plant minikit (Qiagen). Polyadenylated RNA was also purified from healthy non-transformed N. benthamiana plants and from plants transformed with a tandem repeat of DNA h using the above method except that the DNase I treatments were omitted. RNA samples were fractionated by formaldehyde agarose gel electrophoresis (Oligotex Handbook; Qiagen) and blotted onto Hybond-NX membrane (Amersham Pharmacia). Probes for RNA detection were PCR-amplified from a dimer of DNA h in pBSAYVV2h (Saunders et al., 2000). A full-length DNA h probe was amplified using primers corresponding to nucleotides 1167 – 1193 (virion-sense) and 1139 – 66 (complementary-sense), and DNA h fragments were amplified using primers corresponding to nucleotides 44– 85 (virion-sense) and 526– 552 (complementary-sense) (probe I), and nucleotides 556 –581 (virionsense) and 44– 76 (complementary-sense) (probe II). Fragments were radiolabelled according to the procedure of Feinberg and Vogelstein (1983). Transcript mapping by 5V-RACE and 3V-RACE Complementary DNA was synthesized from polyadenylated RNA using a Smart RACE cDNA amplification kit (Clontech) according to the manufacturer’s instructions. 5VRACE and 3V-RACE reactions were carried out using primers corresponding to DNA h nucleotides 381– 409 (virion-sense) and 425– 450 (complementary-sense), respectively. Due to the low yield of 5V-RACE reaction products, fragments were re-amplified using the above virion-sense primer or a primer corresponding to nucleotides 471 –500. Amplified fragments were purified by agarose gel electrophoresis, cloned into pGEM-T Easy (Promega), and sequenced.

Acknowledgments We thank Marion Pinner for her assistance in sectioning and microscopy of leaf material. This work was funded by the Biotechnology and Biological Sciences Research Council. Viruses were maintained and manipulated with

the authorization of the Department for Environment, Food and Rural Affairs under the Plant Health (Great Britain) Order 1993, license PHL 185A/4538(7/2003). References Bigarre´, L., Chazly, M., Salah, M., Ibrahim, M., Padidam, M., Nicole, M., Peterschmitt, M., Fauquet, C., Thouvenel, J.C., 2001. Characterization of a new begomovirus from Egypt infecting hollyhock (Althea rosea). Eur. J. Plant Pathol. 107, 701 – 711. Briddon, R.W., Mansoor, S., Bedford, I.D., Pinner, M.S., Markham, P.G., 2000. Clones of cotton leaf curl geminivirus induce symptoms atypical of cotton leaf curl disease. Virus Genes 20, 19 – 26. Briddon, R.W., Mansoor, S., Bedford, I.D., Pinner, M.S., Saunders, K., Stanley, J., Zafar, Y., Malik, K.A., Markham, P.G., 2001. Identification of DNA components required for induction of cotton leaf curl disease. Virology 285, 234 – 243. Briddon, R.W., Bull, S.E., Amin, I., Idris, A.M., Mansoor, S., Bedford, I.D., Dhawan, P., Rishi, N., Siwatch, S.S., Abdel-Salam, A.M., Brown, J.K., Zafar, Y., Markham, P.G., 2003. Diversity of DNA h, a satellite molecule associated with some monopartite begomoviruses. Virology 312, 106 – 121. Carroll, R., Derse, D., 1993. Translation of equine infectious anemia virus bicistronic tat-rev mRNA requires leaky ribosome scanning of the tat CTG initiation codon. J. Virol. 67, 1433 – 1440. Chapman, S., Kavanagh, T., Baulcombe, D., 1992. Potato virus X as a vector for gene expression in plants. Plant J. 2, 549 – 557. Corcelette, S., Masse´, T., Madjar, J.-J., 2000. Initiation of translation by non-AUG codons in human T-cell lymphotropic virus type I mRNA encoding both Rex and Tax regulatory proteins. Nucleic Acids Res. 28, 1625 – 1634. Covey, S.N., Hull, R., 1981. Transcription of cauliflower mosaic virus DNA. Detection of transcripts, properties and location of the gene encoding the virus inclusion body protein. Virology 111, 463 – 474. Dry, I.B., Rigden, J.E., Krake, L.R., Mullineaux, P.M., Rezaian, M.A., 1993. Nucleotide sequence and genome organization of tomato leaf curl geminivirus. J. Gen. Virol. 74, 147 – 151. Edmonds, M., 2002. A history of poly A sequences: from formation to factors to function. Prog. Nucleic Acid Res. Mol. Biol. 71, 285 – 389. Feinberg, A.P., Vogelstein, B., 1983. A technique for radiolabeling DNA restriction fragments to high specific activity. Anal. Biochem. 132, 6 – 13. Futterer, J., Potrykus, I., Bao, Y., Li, L., Burns, T.M., Hull, R., Hohn, T., 1996. Position-dependent ATT initiation during plant pararetrovirus rice tungro bacilliform virus translation. J. Virol. 70, 2999 – 3010. Hamilton, W.D.O., Bisaro, D.M., Coutts, R.H.A., Buck, K.W., 1983. Demonstration of the bipartite nature of the genome of a single-stranded DNA plant virus by infection with the cloned DNA components. Nucleic Acids Res. 11, 7387 – 7396. Hanley-Bowdoin, L., Settlage, S.B., Orozco, B.M., Nagar, S., Robertson, D., 1999. Geminiviruses: models for plant DNA replication, transcription, and cell cycle regulation. Crit. Rev. Plant Sci. 18, 71 – 106. Hirose, T., Ideue, T., Wakasugi, T., Sugiura, M., 1999. The chloroplast infA gene with a functional UUG initiation codon. FEBS Lett. 445, 169 – 172. Horsch, R.B., Klee, H.J., 1986. Rapid assay of foreign gene expression in leaf discs transformed by Agrobacterium tumefaciens: role of the TDNA borders in the transfer process. Proc. Natl. Acad. Sci. U.S.A. 83, 4428 – 4432. Horsch, R.B., Fry, J.E., Hoffmann, N.L., Wallroth, W., Eichholz, D., Rogers, S.G., Fraley, R.T., 1985. A simple and general method for transferring genes into plants. Science 227, 1229 – 1231. Jose, J., Usha, R., 2003. Bhendi yellow vein mosaic disease in India is caused by association of a DNA h satellite with a begomovirus. Virology 305, 310 – 317.

K. Saunders et al. / Virology 324 (2004) 37–47 Kheyr-Pour, A., Bendahmane, M., Matzeit, V., Accotto, G.-P., Crespi, S., Gronenborn, B., 1991. Tomato yellow leaf curl virus from Sardinia is a whitefly-transmitted monopartite geminivirus. Nucleic Acids Res. 19, 6763 – 6769. Ko, F.C.F., Chow, K.L., 2003. A mutation at the start codon defines the differential requirement of dpy-11 in Caenorhabditis elegans body hypodermis and male tail. Biochem. Biophys. Res. Commun. 309, 201 – 208. Kobayashi, Y., Dokiya, Y., Kumazawa, Y., Sugita, M., 2002. Non-AUG translation initiation of mRNA encoding plastid-targeted phage-type RNA polymerase in Nicotiana sylvestris. Biochem. Biophys. Res. Commun. 299, 57 – 61. Kozak, M., 1990. Downstream secondary structure facilitates recognition of initiator codons by eukaryotic ribosomes. Proc. Natl. Acad. Sci. U.S.A. 87, 8301 – 8305. Latham, J.R., Saunders, K., Pinner, M.S., Stanley, J., 1997. Induction of plant cell division by beet curly top virus gene C4. Plant J. 11, 1273 – 1283. Laufs, J., Traut, W., Heyraud, F., Matzeit, V., Rogers, S.G., Schell, J., Gronenborn, B., 1995. In vitro cleavage and joining at the viral origin of replication by the replication initiator protein of tomato yellow leaf curl virus. Proc. Natl. Acad. Sci. U.S.A. 92, 3879 – 3883. Mansoor, S., Khan, S.H., Bashir, A., Saeed, M., Zafar, Y., Malik, K.A., Briddon, R., Stanley, J., Markham, P.G., 1999. Identification of a novel circular single-stranded DNA associated with cotton leaf curl disease in Pakistan. Virology 259, 190 – 199. Mansoor, S., Briddon, R.W., Zafar, Y., Stanley, J., 2003. Geminivirus disease complexes: an emerging threat. Trends Plant Sci. 8, 128 – 134. Muralidhar, S., Becerra, S.P., Rose, J.A., 1994. Site-directed mutagenesis of adeno-associated virus type 2 structural protein initiation codons: effects on regulation of synthesis and biological activity. J. Virol. 68, 170 – 176. Navot, N., Pichersky, E., Zeidan, M., Zamir, D., Czosnek, H., 1991. Tomato yellow leaf curl virus: a whitefly-transmitted geminivirus with a single genomic component. Virology 185, 151 – 161. Okimoto, R., Macfarlane, J.L., Wolstenholme, D.R., 1990. Evidence for the frequent use of TTG as the translation initiation codon of mitochondrial protein genes in the nematodes, Ascaris suum and Caenorhabditis elegans. Nucleic Acids Res. 18, 6113 – 6118. Pinner, M.S., Medina, V., Plaskitt, K.A., Markham, P.G., 1993. Viral inclusions in monocotyledons infected by maize streak and related geminiviruses. Plant Pathol. 42, 75 – 87. Prats, A.-C., De Billy, G., Wang, P., Darlix, J.L., 1989. CUG initiation codon used for the synthesis of a cell-surface antigen coded by the murine leukemia virus. J. Mol. Biol. 205, 363 – 372. Riechmann, J.L., Ito, T., Meyerowitz, E.M., 1999. Initiation of AGAMOUS mRNA translation in Arabidopsis thaliana. Mol. Cell. Biol. 19, 8505 – 8512. Roberts, A.G., Santa-Cruz, S., Roberts, I.M., Prior, D.A.M., Turgeon, R., Oparka, K.J., 1997. Phloem unloading in sink leaves of Nicotiana benthamiana: comparison of a fluorescent solute with a fluorescent virus. Plant Cell 9, 1381 – 1396.

47

Saunders, K., Stanley, J., 1999. A nanovirus-like DNA component associated with yellow vein disease of Ageratum conyzoides: evidence for interfamilial recombination between plant DNA viruses. Virology 264, 142 – 152. Saunders, K., Bedford, I.D., Briddon, R.W., Markham, P.G., Wong, S.M., Stanley, J., 2000. A unique virus complex causes Ageratum yellow vein disease. Proc. Natl. Acad. Sci. U.S.A. 97, 6890 – 6895. Saunders, K., Bedford, I.D., Stanley, J., 2001. Pathogenicity of a natural recombinant associated with ageratum yellow vein disease: implications for begomovirus evolution and disease aetiology. Virology 282, 38 – 47. Saunders, K., Bedford, I.D., Yahara, T., Stanley, J., 2003. The earliest recorded plant virus disease. Nature 422, 831. Shirako, Y., 1998. Non-AUG translation initiation in a plant RNA virus: a forty-amino-acid extension is added to the N terminus of the soil-borne wheat mosaic virus capsid protein. J. Virol. 72, 1677 – 1682. Stanley, J., 1983. Infectivity of the cloned geminivirus genome requires sequences from both DNAs. Nature 305, 643 – 645. Stanley, J., 1995. Analysis of African cassava mosaic virus recombinants suggest strand nicking occurs within the conserved nonanucleotide motif during the initiation of rolling circle DNA replication. Virology 206, 707 – 712. Stanley, J., Gay, M.R., 1983. Nucleotide sequence of cassava latent virus DNA. Nature 301, 260 – 262. Stanley, J., Latham, J.R., 1992. A symptom variant of beet curly top geminivirus produced by mutation of open reading frame C4. Virology 190, 506 – 509. Stanley, J., Saunders, K., Pinner, M.S., Wong, S.M., 1997. Novel defective interfering DNAs associated with ageratum yellow vein geminivirus infection of Ageratum conyzoides. Virology 239, 87 – 96. Tan, P.H.N., Wong, S.M., Wu, M., Bedford, I.D., Saunders, K., Stanley, J., 1995. Genome organization of ageratum yellow vein virus, a monopartite whitefly-transmitted geminivirus isolated from a common weed. J. Gen. Virol. 76, 2915 – 2922. van Engelen, F.A., Molthoff, J.W., Conner, A.J., Nap, J.-P., Pereira, A., Stiekema, W.J., 1995. pBINPLUS: an improved plant transformation vector based on pBIN19. Transgenic Res. 4, 288 – 290. van Eyll, O., Michiels, T., 2002. Non-AUG-initiated internal translation of the L* protein of Theiler’s virus and importance of this protein for viral persistence. J. Virol. 76, 10665 – 10673. van Wezel, R., Dong, X., Blake, P., Stanley, J., Hong, Y., 2002. Differential roles of geminivirus Rep and AC4 (C4) in the induction of necrosis in Nicotiana benthamiana. Mol. Plant Pathol. 3, 461 – 471. Zambryski, P., Joos, H., Genetello, C., Leemans, J., van Montagu, M., Schell, J., 1983. Ti plasmid vector for the introduction of DNA into plant cells without alteration of their normal regeneration capacity. EMBO J. 2, 2143 – 2150. Zhou, X., Xie, Y., Tao, X., Zhang, Z., Li, Z., Fauquet, C.M., 2003. Characterization of DNA h associated with begomoviruses in China and evidence for co-evolution with their cognate viral DNA-A. J. Gen. Virol. 84, 237 – 247.