Infectivity of nanovirus DNAs: induction of disease by cloned genome components of Faba bean necrotic yellows virus

Viruses depend on host cells for multiplication and have developed many ways of redirecting the host cell's biosynthetic machinery to their use (Kornberg & Baker, 1992; Jansen-Dürr, 1996). Gemini- and nanoviruses are the only known single-stranded DNA (ssDNA) viruses of plants. Geminivirus genomes comprise one or two circular ssDNA molecules of 2.53.0 kb, individually encapsidated in characteristic twinned virions, transmitted by either whiteflies or leafhoppers (Harrison, 1985; Rybicki, 1994). Nanovirus genomes consist of multiple circular ssDNAs of about 1 kb each (Vetten et al., 2005). They are encapsidated in small, isometric virions transmitted by aphids (Chu & Helms, 1988; Harding et al., 1991; Katul et al., 1993). Up to 12 different DNAs have been described for the nanoviruses Banana bunchy top virus (BBTV), Faba bean necrotic yellows virus (FBNYV), Milk vetch dwarf virus (MDV) and Subterranean clover stunt virus (reviewed by Vetten et al., 2005). With a single exception, each DNA molecule appears to encode only one protein (Beetham et al., 1997; Vetten et al., 2005). This contrasts with geminiviruses, where both strands of the replicative double-stranded DNA (dsDNA) are transcribed and encode proteins (reviewed by Gutierrez, 2000).

Gemini- and nanoviruses modulate the host's cell cycle and replicate in the nucleus by a rolling-circle mechanism exploiting host DNA polymerases (Timchenko et al., 1999; Aronson et al., 2000; Gutierrez, 2000; Hanley-Bowdoin et al., 2004). Multifunctional virus replication-initiator proteins (Rep proteins) are essential for replication of both gemini- and nanoviruses (Hafner et al., 1997; Timchenko et al., 1999; Hanley-Bowdoin et al., 2004). In contrast to geminiviruses, two to five similar yet clearly distinct Rep-encoding DNAs (rep components) have been described for each nanovirus (Vetten et al., 2005). However, only one of them encodes the essential master Rep (M-Rep) protein capable of initiating replication of the DNAs encoding the other nanovirus proteins (Timchenko et al., 1999, 2000; Horser et al., 2001). The master Rep-encoding DNA (DNA-R) and all non-rep DNAs are considered integral parts of the nanovirus genome, as they are consistently found associated with nanovirus infections and share a common origin of replication (Timchenko et al., 2000). The additional rep components occur erratically in nanovirus infections and encode Rep proteins that only initiate replication of their cognate DNAs (Timchenko et al., 1999). They depend on their respective nanovirus for encapsidation and dissemination (Vetten et al., 2005). The non-rep DNAs, encoding the capsid protein (DNA-S) (Chu et al., 1993; Katul et al., 1997; Wanitchakorn et al., 1997), the cell-cycle link protein Clink (DNA-C) (Aronson et al., 2000), a movement protein (DNA-M) (Wanitchakorn et al., 2000) and a nuclear-shuttle protein (DNA-N) (Wanitchakorn et al., 2000), have been identified from all nanovirus infections. Additional DNAs (DNA-U1 to DNA-U4) encoding proteins of unknown function have been described for some nanoviruses (Vetten et al., 2005). Circumstantial evidence for FBNYV suggests that eight DNAs, DNA-R, -C, -M, -N, -S, -U1, -U2 and -U4 (Fig. 1), are the integral genome components of this virus (Vetten et al., 2005).

(22K): Fig. 1. Genome components of FBNYV. Eight FBNYV genome components are displayed as replicative dsDNAs. Component designations are given below the respective DNA; names and sizes of the encoded proteins are indicated inside each circle. Bold arrows represent the position and size of the corresponding ORFs. ori marks the replication origin containing the conserved inverted-repeat sequence (>.<) shared by these eight DNAs. Asterisks denote presumed TATA boxes and a wavy line (∼) indicates putative polyadenylation signals. A large dash () at the top of some DNAs indicates that the respective component was individually dispensable for symptom development in inoculations of V. faba with only seven other DNAs, upon both agroinoculation and bombardment. Components DNA-C, DNA-N and DNA-U4 (in brackets) were collectively not required for symptom development in agroinoculations using only the five components DNA-R, -S, -M, -U1 and -U2.

Infectivity of cloned genomic DNAs has not previously been shown for any nanovirus. Also, the minimal number of genomic DNAs required for eliciting nanovirus disease was not known. Here, we report the first nanovirus infections resulting from the use of cloned DNA components of the FBNYV genome. Disease symptoms typical of an FBNYV infection were obtained in Vicia faba plants bombarded or agroinoculated with different combinations of cloned full-length FBNYV DNAs. Cloned nanovirus DNAs.
DNA components of the Egyptian isolate (EV1-93) of FBNYV (FBNYV-Eg) including rep11 DNA were cloned as described previously (Timchenko et al., 1999). In the following, the individual DNAs are designated according to current nomenclature (Vetten et al., 2005). Full-length DNA-U4 of FBNYV-Eg (GenBank accession no. AJ749902[GenBank] ) was amplified by PCR using primer pair C12XbaI(+) and C12XbaI() (see Supplementary Table S1, available in JGV Online) and inserted in the XbaI site of pBluescript II KS(+) (Stratagene). Thus, an XbaI site without amino acid change in the encoded protein was introduced at position 499 of DNA-U4. The DNA-S clone of MDV has been described previously (Timchenko et al., 2000).

Tandemly repeated copies of FBNYV-Eg DNA in pBluescript II KS(+) and the binary T-DNA vector pBin19 (Bevan, 1984) were as described previously (Timchenko et al., 1999). DNA-U4 was duplicated in the XbaI site of pBluescript II KS(+) and subsequently transferred as a BamHISalI fragment to pBin19. A tandemly repeated copy of MDV DNA-S was transferred from pBluescript II KS(+) to pBin19 as a BamHISalI fragment.

In addition to the tandemly repeated copies of individual genome components, pairs of two different tandemly repeated FBNYV DNAs were assembled on the same T-DNA of pBin19 (see Supplementary Methods, available in JGV Online).

Inoculation of V. faba using cloned nanovirus DNAs
Plant material.
V. faba (variety Condor) plants were grown in soil in growth chambers with a 16 h photoperiod, 50 % humidity and 25 °C within a restricted-access S3 confinement facility.

Agroinoculation.
Tandemly repeated copies of all FBNYV genome components and pairs of tandemly repeated copies in pBin19 were introduced as described previously (Timchenko et al., 1999) by electroporation into Agrobacterium tumefaciens strain COR308 (Cornell Research Foundation). This strain carries copies of the virG and virE1/E2 genes on the additional plasmid pCH32 (Hamilton et al., 1996). Agrobacteria were grown in YEB medium (0.5 % Difco nutrient broth, 0.5 % peptone, 0.1 % yeast extract, 0.5 % sucrose, 2 mM MgSO₄, pH 7.2) containing 50 µg kanamycin ml¹ and 5 µg tetracycline ml¹. An overnight culture was diluted 10-fold in the same medium, adjusted to pH 6.0 by adding morpholinoethanesulfonic acid (MES) to 10 mM and containing 50 µM acetosyringone (3',5'-dimethoxy-4'-hydroxyacetophenone) (Sigma-Aldrich). Cells were grown overnight to a final OD₆₀₀ of about 2, harvested by centrifugation (3500 g, 20 min) and resuspended in 0.1 vol. 0.5x MS medium (Murashige & Skoog, 1962), containing 10 mM MES and 150 µM acetosyringone. One-week-old V. faba plants (three- to four-leaf stage) were needle-inoculated with a suspension containing equal amounts of bacteria carrying a given FBNYV genome component or component pairs.

Bombardment.
The Helios gene gun system (Bio-Rad) was used for delivery of the tandemly repeated copies of the FBNYV DNAs in pBluescript II KS(+). All procedures of cartridge preparation were performed according to the Helios gene gun system instruction manual (Bio-Rad, 1996). Gold particles (0.6 µm size) were coated with an equimolar mixture of pBluescript II KS(+) derivatives carrying tandemly repeated copies of the respective FBNYV DNAs, each at a concentration of about 1 µg ml¹. For about 60 cm tubing, 25 mg particles and 50 µg total DNA were used. Polyvinyl pyrrolidone (molecular mass 360 000 Da) was used at 0.01, 0.02 and 0.03 mg ml¹. One-week-old V. faba plants (three- to four-leaf stage) were bombarded, using six to eight shots per plant at a helium pressure of 200 p.s.i. (1.38 MPa).

Detection of nanovirus DNAs in infected plants
Southern hybridization.
For rapid detection of FBNYV in V. faba, leaves of symptomatic plants were squashed onto a Hybond-N membrane (Amersham Biosciences) to which viral ssDNA binds without denaturation (Navot et al., 1989). Replicative FBNYV DNA forms present in total DNA preparations from V. faba were resolved by electrophoresis and detected by Southern hybridization as described previously (Timchenko et al., 1999), using gel-purified component-specific probes.

PCR.
Total DNA was extracted from V. faba according to Edwards et al. (1991) with slight modifications. About 50 mg plant tissue was ground in 400 µl extraction buffer [200 mM Tris/HCl (pH 7.5), 250 mM NaCl, 25 mM EDTA, 0.5 % SDS, 0.1 % 2-mercaptoethanol] and the extract was clarified by centrifugation (15 000 g, 5 min). DNA was precipitated by adding an equal volume of 2-propanol and resupended in 100 µl RTE buffer [10 mM Tris/HCl (pH 7.0), 1 mM EDTA, 10 µg RNase ml¹]. One microlitre of DNA was used for PCR in a 20 µl reaction mix as specified for Eurobiotaq II DNA polymerase (EuroBio). PCR conditions were: 95 °C, 5 min; 35 cycles of 95 °C, 30 s; 53 °C, 30 s; 72 °C, 2 min; and 72 °C, 5 min. Individual FBNYV DNAs were identified by using component-specific primer pairs (see Supplementary Table S1, available in JGV Online).

Detection of FBNYV virions in infected plants.
For virion purification, all above-ground parts of symptomatic plants that had served as source for virus acquisition by Acyrthosiphon pisum were harvested, pulverized in liquid nitrogen and stored at 80 °C until use. About 80 g leaf and stem tissue was processed essentially as described by D'Arcy et al. (1989). The frozen tissue was further macerated in a Waring blender using 2.5 vols 0.1 M sodium citrate buffer, pH 6, containing 5 ml 2-mercaptoethanol l¹, 1 mg NaN₃ l¹ and 100 ml Pectinex SP-L l¹ (Novo Nordisk Ferment Ltd) and stirred overnight at room temperature. Following clarification of the extracts with 0.25 vol. chloroform/butanol (1 : 1) and low-speed centrifugation, virions were concentrated and purified further by precipitation in polyethylene glycol 6000 (8 g in 100 ml), differential centrifugation and sucrose density-gradient centrifugation. Negatively stained virion preparations were examined by using a Zeiss EM906 electron microscope at a magnification of x36 000.

Detection of nanovirus proteins in infected plants.
About 100 mg plant tissue was ground in liquid nitrogen and the powder was transferred to an Eppendorf tube with 400 µl extraction buffer containing 20 mM HEPES (pH 8.2), 50 mM EDTA, 150 mM NaCl, 0.5 % NP40, 1 mM PMSF, 10 mM 2-mercaptoethanol and 100 KIU aprotinin ml¹ (Calbiochem) and incubated at 4 °C for 20 min on a shaker. Extracts were clarified by centrifugation at 4 °C (15 000 g, 5 min) and used freshly or stored at 20 °C. Aliquots (20 µl) of the extracts were fractionated by SDS-PAGE (12 or 15 % polyacrylamide for M-Rep or capsid protein, respectively) and transferred to a Hybond-P membrane (Amersham Biosciences) for 40 min at 200 mA at 4 °C by using a Semi-Dry Electroblotter (Ancos). For detection of capsid protein, a 1 : 1000 dilution of the IgG fraction of a rabbit antiserum to FBNYV virions was used (Katul et al., 1993). M-Rep protein was detected by using a 1 : 2000 dilution of a rabbit antiserum to His₆-tagged M-Rep protein (Timchenko et al., 1999).

Insect transmission
Aphis craccivora.
Viruliferous and non-viruliferous aphids were reared as described previously (Vega-Arreguín et al., 2005). For virus acquisition, aphids were allowed to feed for 47 days on individual diseased plants in soil or on diseased plants assembled into a group (kept in water or in soil). For inoculation access, they were transferred to 1-week-old V. faba seedlings and killed after 1 week by fumigation (0.2 % Dedevap; Bayer).

Acyrthosiphon pisum.
All symptomatic plants were transplanted to a large plastic pot (approx. 20 cm diameter). About 50 virus-free Acyrthosiphon aphids (larval stages L2L4) were placed on each plant and given an acquisition access feeding of about 72 h. Daily disturbing forced them to move and feed on different plants. Thereafter, about 25 aphids per plant were transferred to a total of 25 V. faba seedlings for an inoculation access feeding period of 72 h.

Reconstitution of FBNYV symptoms in V. faba plants infected with cloned viral DNAs
To test the hypothesis that the M-Rep-encoding DNA (DNA-R) along with the seven different non-rep components (DNA-S, -M, -N, -C, -U1, -U2 and -U4) may constitute the FBNYV genome (Fig. 1), attempts were made to use cloned genome components to reproduce an FBNYV infection. Mixtures of the eight different FBNYV DNAs, each inserted as a tandemly repeated copy in pBluescript II KS(+), were used for coating gold particles and delivered biolistically to young V. faba plants. In three independent experiments, we obtained one, one and 10 symptomatic plants per 12, 14 and 50 bombarded plants, respectively.

Similarly, diseased faba bean plants were obtained after agroinoculation with derivatives of the Agrobacterium strain COR308, each carrying a tandemly repeated copy of a given FBNYV DNA in the binary T-DNA vector pBin19. Eight individual cultures of agrobacteria were mixed and needle-inoculated to V. faba seedlings. Symptoms such as leaf yellowing, rolling and necrosis, as well as stunted growth (Katul et al., 1993), appeared about 23 weeks post-inoculation (p.i.) and were indistinguishable from those observed on plants following vector transmission of a field isolate of FBNYV (Fig. 2).

(68K): Fig. 2. Disease symptoms on V. faba agroinoculated with eight cloned FBNYV DNAs. Plants were inoculated with agrobacteria carrying pBin19 or with a mixture of eight cultures, each containing bacteria harbouring a pBin19 derivative with a tandemly repeated copy of a different FBNYV DNA. A plant exhibiting FBNYV symptoms following aphid transmission of a field isolate of FBNYV is shown for comparison (insert). Magnifications of symptoms are shown on the right-hand side.

Based on the assumption that the presence of all eight genomic DNAs in the same cell is required to initiate a successful infection, we combined pairs of two different FBNYV DNAs, each tandemly repeated, on the same T-DNA to increase the probability of such an event. Indeed, using four different cultures of agrobacteria containing pairs of tandem repeats of FBNYV DNA-R and DNA-C, DNA-M and DNA-N, DNA-U1 and DNA-U4, DNA-S and DNA-U2, respectively, in pBin19 increased the frequency of infection. Ten to fifty-five per cent of the plants inoculated this way developed symptoms, as opposed to 318 % when using individual tandemly repeated copies of the eight DNAs in agrobacteria. A comparable increase in infection efficacy was described for the combination on the same T-DNA of a redundant copy of a geminivirus genome with a tandemly repeated copy of a DNA β (Tao & Zhou, 2004).

V. faba inoculated with cloned FBNYV DNAs produce virus
To confirm that the disease symptoms resulting from plant inoculation with cloned FBNYV DNAs were due to virus multiplication, we assayed viral DNA replication and protein expression. A rapid first screening was done by analysing leaf squashes of the inoculated plants on nylon membranes (Navot et al., 1989), which were then hybridized with a DNA-S-specific probe (see Supplementary Fig. S1, available in JGV Online). DNA prepared from bombarded or agroinoculated plants that had tested positive in this assay was fractionated by agarose-gel electrophoresis followed by Southern blotting. An FBNYV DNA-S-specific probe used for hybridization revealed the presence of DNA-S, both as circular (relaxed and supercoiled) dsDNA and as ssDNA forms in symptomatic tissue (Fig. 3a). Similarly, a DNA-C-specific probe identified replicative forms of DNA-C (data not shown).

(76K): Fig. 3. Detection of FBNYV in V. faba agroinoculated with eight FBNYV DNAs. (a) Southern hybridization of total DNA extracted from newly emerged leaves that had been sampled from V. faba plants 3 weeks after bombardment or 6 weeks after agroinoculation with eight cloned FBNYV DNAs. Numbers refer to individual bombarded or agroinoculated plants, e.g. 26 and 26b represent a young and an old leaf, respectively, of the same plant. Hybridization was with a DNA-S-specific probe. Positions of the replicative viral DNA forms are indicated. oc, Open circular dsDNA; ccc, covalently closed circular dsDNA; ss,single-stranded DNA. (b) Western blot of proteins extracted from agroinoculated V. faba plants that had tested positive in the leaf hybridization assay. IgG to FBNYV virions was used as primary antibody. An arrow indicates the position of the capsid protein. Numbers above the lanes correspond to the samples in (a), except for lanes 8 and 20, not included in (a). C1, Non-inoculated plant; 8, inoculated plant that had scored negative for FBNYV DNA in the squash blot (see Supplementary Fig. S1, available in JGV Online); 20, a plant that had scored positive for FBNYV DNA (Supplementary Fig. S1); Fi, protein extract from a field-infected V. faba from Egypt. M, Pre-stained marker proteins (New England Biolabs); their respective molecular masses (kDa) are indicated on the right. (c) Detection of FBNYV M-Rep protein. Polyclonal anti-His6-M-Rep antiserum was used for detection in a Western blot of plant proteins fractionated by SDS-PAGE (12 % gel). Lane 1, 30 ng His6-M-Rep protein expressed in and purified from E. coli. Lanes 2 and 3, protein extracts of V. faba agroinoculated with eight cloned FBNYV DNAs. Lanes 4 and 5, extracts of a field-infected (4) and a healthy (5) faba bean plant. Lanes 6 and 7, extracts from Nicotiana benthamiana leaf discs inoculated with either agrobacteria containing pBin19 (6) or pBin19 carrying a tandemly repeated copy of FBNYV DNA-R (7). Arrows denote the respective positions of wild-type M-Rep and histidine-tagged M-Rep. (d) Electron-microscopic visualization of negatively stained nanovirus particles in partially purified preparations typically obtained from symptomatic V. faba following agroinoculation or biolistic bombardment. Bar, 200 nm.

Western blot analysis of plants in which DNA-S was detected also expressed the capsid protein encoded by that DNA (Fig. 3b). Furthermore, the FBNYV M-Rep protein was detected readily by Western blot in symptomatic plants (Fig. 3c), despite the fact that the m-rep promoters of BBTV and MDV have been reported to be weak (Dugdale et al., 1998; Shirasawa-Seo et al., 2005). Finally, virus preparations from symptomatic V. faba agroinoculated with all eight FBNYV DNAs contained numerous nanovirus particles (Fig. 3d) and gave strong ELISA reactions (data not shown). The yield of purified virions from agroinoculated plants was comparable to that of plants infected with a field isolate of FBNYV following aphid transmission. The observation that the sucrose-gradient sedimentation rates of the virions produced after agroinoculation or aphid transmission appeared indistinguishable suggests that the clonally derived virions contained viral DNA.

These results demonstrate that FBNYV multiplication, virion formation and the virus-associated disease can be reproduced by using cloned viral DNAs for particle bombardment or agroinoculation. This is the first demonstration of a nanovirus infection using cloned copies of the viral DNAs.

Clink protein is dispensable for FBNYV disease in V. faba
The Clink protein encoded by DNA-C interacts with the cell-cycle regulator pRB and SKP1, a constituent of the ubiquitin protein-turnover pathway (Aronson et al., 2000). Having obtained FBNYV disease symptoms by using eight cloned genomic DNAs, we further elucidated the role of Clink in viral infection. DNA-C mutated in the sequence encoding the F-box or the LxCxE motif of Clink (Aronson et al., 2000) was agroinoculated along with the other seven FBNYV DNAs. The alterations of Clink changed neither the disease symptoms nor the proportion of symptomatic plants. In a representative experiment, two out of 40 plants inoculated were diseased when the RB-binding motif of Clink was altered; likewise, two out of 40 plants were diseased when SKP1 binding of Clink was impaired. Moreover, when omitting DNA-C completely and using only the other seven FBNYV DNAs, also two out of 40 plants became symptomatic, compared with three out of 40 diseased plants when using eight DNAs including wild-type DNA-C. These results are in line with those obtained for mutants in the pRB-binding site of the Rep proteins of Bean yellow dwarf virus (Liu et al., 1999). In the case of Maize streak virus, a mastrevirus infecting monocotyledonous plants, interaction of Rep with pRB was also not essential for replication or timing or efficiency of infection, whereas symptom severity was reduced (McGivern et al., 2005; Shepherd et al., 2005). Although Clink as a general enhancer of DNA replication (Aronson et al., 2000) was dispensable for FBNYV symptom development in V. faba, there seems to be a positive selection for maintenance of DNA-C in the course of an infection (see Table 1).

Table 1. Presence of FBNYV DNAs in symptomatic V. faba agroinoculated with eight cloned virus DNAs The respective FBNYV DNAs (DNA-R, -M, -S, -N, -C, -U1, -U2 and -U4) were detected by PCR using total DNA of symptomatic plants from at least four independent inoculation experiments.

How many and which viral DNAs are required for manifestation of the FBNYV disease?
Having determined that DNA-C encoding Clink was dispensable for the development of the FBNYV disease, we investigated whether other FBNYV DNAs were also dispensable for infection in V. faba. We used different combinations of seven cloned DNAs for bombardment or agroinoculation: each of the eight genomic DNAs, except DNA-R, which is essential for replication, was omitted one at a time. As a result, DNA-N, -U1, -U2 and -U4 turned out to be dispensable, as the combination of the respective other seven DNAs was still able to elicit FBNYV disease symptoms. This is indicated by a dash () in Fig. 1. In experiments omitting either DNA-U1 or DNA-U2, symptom development appeared to be less pronounced and was delayed, appearing 45 weeks versus 23 weeks post-inoculation. However, the number of infected plants was too low to draw any further conclusions on symptom severity or timing, as only one of 13 bombarded plants and one of 40 agroinoculated plants developed disease when DNA-U2 was omitted. Similarly, one of eight bombarded plants showed FBNYV-like symptoms when DNA-U1 was omitted. DNA-M and DNA-S appeared essential for infection, as their omission from the inoculum never led to any diseased plants.

Moreover, when individual viral DNAs in symptomatic plants agroinoculated with eight DNAs were analysed by PCR, only DNA-R, -M and -S were invariably detected in all symptomatic plants, whereas DNA-C, -N, -U1, -U2 and -U4 sometimes appeared to be missing (Table 1). In some cases, symptomatic plants even lacked two or three different viral DNAs. In contrast, all eight DNAs were consistently detected by PCR in symptomatic plants following aphid transmission of a field isolate of FBNYV.

We then further reduced the number of DNAs used for agroinoculation. Surprisingly, we obtained FBNYV symptoms by using only DNA-R, -M, -S, -U1 and -U2, and disease symptoms and frequency of infection using only these five virus DNAs were similar to those in agroinoculations with eight virus DNAs (Table 2). We did not assay other combinations of only five FBNYV DNAs. Diseased plants were not obtained by using only three (DNA-R, -M and -S) or four (DNA-R, -M, -S and DNA-U1 or -U2) DNAs for inoculation (Table 2).

Table 2. Determination by agroinoculation of the minimal number of different FBNYV DNAs causing disease DNA designations are those used in Table 1. The number of symptomatic plants out of inoculated plants is given; NT, nottested.

In summary, our data suggest that only five FBNYV DNAs are sufficient to induce disease symptoms in V. faba. DNA-U1 and DNA-U2 were required in combination with DNA-R, -S and -M (Table 2), but they could be omitted individually when using all other seven DNAs (see Fig. 1). However, the infection rates in the experiments using component combinations were very low (012 %). Thus, the observed differences are probably not statistically significant. It therefore may also be possible that, given a higher infection rate, both U1 and U2 are collectively dispensable, if the presence of either would have a quantitative rather than a qualitative effect on infectivity.

These findings suggest that there may be a functional redundancy or complementation between different nanovirus proteins, despite the lack of significant overall amino acid sequence similarities. As both Clink and M-Rep protein interact with pRB (Aronson et al., 2000; Supplementary Fig. S2, available in JGV Online), it may well be possible that, in the absence of Clink, M-Rep is able to compensate for this deficiency, at least in the host V. faba. Several other examples of partial functional complementation between proteins with no evident sequence similarity have been described. For instance, Btcd, a mouse protein binding curved DNA, can substitute for Escherichia coli H-NS (Timchenko et al., 1996) or two different microbial pyridoxine 5'-phosphate synthases complement each other (Wetzel et al., 2004).

Aphid transmission of FBNYV from artificially infected V. faba
Having established FBNYV infection of V. faba by using cloned viral DNAs, we tried to transmit the virus to V. faba seedlings using nymphs of Aphis craccivora and Acyrthosiphon pisum, two efficient aphid vector species of FBNYV (Katul et al., 1993; Franz et al., 1999). Given that not all viral DNAs were always present in individual diseased plants following bombardment or agroinoculation, all symptomatic plants of a respective experiment were combined into a group and exposed to >100 Aphis craccivora individuals for an acquisition access feeding period of 47 days. Aphids were forced to move between the plants in order to increase their chances of acquiring all encapsidated viral DNAs.

In a further experiment employing agroinoculation of about 400 V. faba seedlings with eight DNAs, only the symptomatic plants verified by component-specific PCR to contain all eight DNAs were used for transmission. Ten such plants were combined into a group and used for a 7-day acquisition access feeding by Aphis craccivora. Subsequently, the aphids were distributed among young faba bean seedlings and allowed a 1-week inoculation access feeding. Another 10 plants verified to contain all eight FBNYV DNAs were exposed to 3050 virus-free Acyrthosiphon pisum nymphs per plant for an acquisition access feeding of 72 h, followed by a 72 h inoculation access feeding on young V. faba seedlings.

In none of these transmission experiments did the aphid-exposed plants develop any disease symptoms, irrespective of whether biolistic DNA delivery or agroinoculation had caused infection of the plants later used as transmission source. Virus transmission by aphids also failed from symptomatic plants that had been inoculated with eight FBNYV DNAs in combination with the additional rep DNAs. Hence, we suspect that determinants required for vector transmission might be mutated in the cloned DNAs. Insect non-transmissible mutants frequently arise in virus populations (Lee et al., 1993; Andrejeva et al., 1996; Noris et al., 1998; Liu et al., 1999; Kheyr-Pour et al., 2000; Höhnle et al., 2001). In ssDNA viruses, such mutants affect the capsid-protein gene and are complemented by their wild-type counterparts present in the population (Noris et al., 1998; Liu et al., 1999; Kheyr-Pour et al., 2000). Therefore, we tried to complement a potential defect in the capsid protein of FBNYV DNA-S by using for agroinoculation the cloned DNA-S of MDV, the capsid protein of which is 84 % similar in sequence to that of FBNYV (Vetten et al., 2005). FBNYV M-Rep also allows efficient replication of MDV DNA-S (Timchenko et al., 2000). In an agroinoculation experiment using seven DNAs of FBNYV in combination with the DNA-S of MDV, four out of 40 plants became symptomatic, compared with 10 diseased plants out of 40 when using FBNYV DNA-S, confirming that indeed the capsid proteins of the two nanoviruses could substitute for each other.

Symptomatic plants from the MDV DNA-S and FBNYV DNA-S complementation experiments were pooled and used for virus acquisition and transmission by aphids. However, as in the previous transmission attempts, no diseased plants were obtained in these aphid-transmission experiments. Franz et al. (1999) have provided evidence for a virus-encoded helper component required for aphid transmission of FBNYV. However, neither the molecular nature of such a helper component nor its coding DNA(s) is known. Hence, virus transmission by aphids from symptomatic plants obtained by inoculation with cloned FBNYV DNAs remains unsolved.

Role of additional Rep-encoding DNAs in FBNYV infections
Apart from the observation that V. faba infected with some FBNYV isolates for which no additional rep DNAs were identified showed more severe symptoms (L. Katul & H. J. Vetten, unpublished results), no experimental data are available as to whether these additional rep DNAs have any biological significance for the nanovirus that they are associated with. Therefore, we tried to elucidate the role of one of the additional rep DNAs. V. faba plants were agroinoculated with eight FBNYV DNAs or with the same eight FBNYV DNAs in combination with rep11 DNA, the most abundant additional rep DNA (Timchenko et al., 1999). Out of 77 V. faba plants inoculated with eight DNAs, we obtained 42 symptomatic plants (lot 1). By contrast, only 21 of 74 plants agroinoculated with eight DNAs plus rep11 DNA were symptomatic (lot 2). Moreover, 16 of the 42 symptomatic plants from lot 1 showed severe symptoms of yellowing and necrosis at 28 days p.i., whereas only four of the 21 plants from lot 2 showed equally severe symptoms. These results suggest an interference of rep11 DNA with FBNYV multiplication to such an extent that it reduces the number of plants in which the virus can establish disease. This may be due to competition between the additional rep DNAs and the genomic nanovirus DNAs for factors required for replication, systemic movement or encapsidation. Remarkably, Rep1 protein encoded by the rep1 DNA was about 10 times more active in an in vitro origin-cleavage and nucleotidyl-transfer reaction than the FBNYV M-Rep protein (Timchenko et al., 1999). Furthermore, in many cases, the additional rep DNAs were identified prior to the master Rep-encoding DNA-R (Katul et al., 1995; Timchenko et al., 2000), suggesting that they attain higher concentrations than DNA-R in nanovirus-infected plants.

The fact that not all FBNYV DNAs are required for disease in the host V. faba is reminiscent of some geminiviruses (reviewed by Mansoor et al., 2003). For instance, Ageratum yellow vein virus (AYVV) DNA A replicates and induces leaf-curl symptoms in tomato and Nicotiana spp. However, it is unable to induce typical yellow-vein symptoms when reintroduced into Ageratum conyzoides unless in mixed infection with an additional smaller molecule, DNA β (Saunders et al., 2000). Similar findings were described for Cotton leaf curl virus (CLCuV) DNA A, where DNA β is also required for the full manifestation of typical disease in cotton (Briddon et al., 2001). Furthermore, a nanovirus-like rep DNA, DNA 1, which by itself replicates autonomously, but relies on DNA A for encapsidation and systemic movement, was found associated with CLCuV and AYVV (Mansoor et al., 1999; Saunders & Stanley, 1999). Contrary to DNA β, DNA 1 is not required for maintenance of the disease induced by AYVV or CLCuV, yet was proposed to exert a modulating effect on symptom severity (Mansoor et al., 2003).

In this context, it appears conceivable that fewer than all eight identified DNAs are required to elicit FBNYV disease symptoms in V. faba, a cultivated plant. However, in the various wild species that are the natural reservoirs of FBNYV (Franz et al., 1997), very probably all of them are required. Whatever the role of the additional rep DNAs in wild species may be, they appear to have been maintained in nanovirus evolution as paralogues of rep or para rep DNAs. They may attenuate disease symptoms, prolong the life expectancy of FBNYV-infected plants (which typically die prematurely) and increase the chance of the respective virus of being spread. As they probably provide an evolutionary advantage to the nano- and geminiviruses that they associate with, they are commonly encountered in nanovirus infections and in an increasing number of geminivirus infections (Briddon & Stanley, 2006).

The first reconstitution of typical disease symptoms in V. faba using cloned copies of FBNYV DNAs and biolistic DNA delivery or agroinoculation will contribute to elucidating the role of each individual genome component and their respectively encoded proteins, including the gene(s) for the hitherto-elusive aphid-transmission factor. It represents a considerable advancement towards a better understanding of the biological cycle of nanoviruses in general.

We are indebted to F. de Kouchkovsky, D. Clérot and A. Sieg-Müller for technical assistance. We thank Cornell Research Foundation Inc. for A. tumefaciens strain COR308. J. C. V.-A. was supported by a fellowship from CONACYT, México. The work was supported in part by the European Commission (INCO-DC programme ERBIC18-CT96-0121).

Footnotes

^,†, J. C. Vega-Arreguín^{1 ,‡}, B. C. Ramirez¹ †Present address: Ambassade de France, 00180 Helsinki, Finland. ‡Present address: CINVESTAV, Unidad Irapuato, 36500 Irapuato, Guanajuato, Mexico. §Present address: Institut Pasteur, 75724 Paris Cedex 15, France.