In planta protein–protein interactions assessed using a nanovirus-based replication and expression system


*For correspondence (fax 33 1 69 82 36 95; e-mail taniat@isv.cnrs-gif-fr)


The multipartite genome of the nanovirus Faba bean necrotic yellows virus, which consists of one gene on each DNA component, was exploited to construct a series of virus-based episomal vectors designed for transient replication and gene expression in plants. This nanovirus based expression system yields high levels of protein which allows isolation of recombinant protein and protein complexes from plant tissues. As examples, we demonstrated in planta interaction between the nanovirus F-box protein Clink and SKP1, a constituent of the ubiquitin-dependent protein turnover pathway. Thus, replicative nanovirus vectors provide a simple and efficient means for in planta characterization of protein-protein interaction.


The development of eukaryotic expression vectors has provided a direct and versatile way to probe gene functions. Ectopic protein expression can be used to study their subcellular localization, stability, post-translational modifications and protein-protein interaction. However, stable integration of foreign DNA may suffer from undesired effects such as insertional mutagenesis, positional variation in expression, or silencing by de novo methylation (Doerfler et al., 1997; Vance and Vaucheret, 2001). To circumvent these problems, the use of episomal, high copy number eukaryotic vectors offers an attractive alternative.

Plant DNA viruses, like geminiviruses and to a lesser extent caulimoviruses, have been used as extrachromosomal replicons for heterologous protein expression (reviewed in Gronenborn and Matzeit, 1989; Timmermans et al., 1994). However, size limitation for encapsidation as well as complex gene expression are factors that have retarded progress in the field (Viaplana et al., 2001). Geminiviruses are completely dependent on DNA replication enzymes of the host cell. Replication occurs many times within a given cell and is facilitated by the creation of a favourable cellular environment through interactions of viral proteins with a member of the retinoblastoma family, pRB, a key cell cycle regulator (Settlage et al., 2001; Xie et al., 1995, 1996). These multiple interactions may even lead to cell death and explain retrospectively why these viruses could not be maintained in cultured cells (reviewed in Gutiérrez, 2000).

Considerable progress has also been made in modifying plant RNA viruses for the expression of foreign genes in plants (Palese and Roizman, 1996; Pogue et al., 1998; Yusibov et al., 1999). Nevertheless, there have been recurrent problems associated with insert instability, low expression, gene silencing, and complex modes of expression. This requires a case by case solution, and tailored viruses that may differ markedly from the original ones (Shivprasad et al., 1999). Furthermore, because RNA viruses replicate in the cytoplasm they are not suitable for promoter and transcription factor studies or experiments addressing mRNA stability.

Nanoviruses are ssDNA plant viruses whose major hosts are legumes or Musa species (reviewed in Chu et al., 1995; Vetten and Katul, 2001). Their genome consists of multiple circular DNAs of about 1000 nucleotides each. In almost all cases, the individual DNA molecules (replicons) carry a single gene (Beetham et al., 1997; Gronenborn et al., 2002). Nanoviruses replicate in the nucleus through double-stranded DNA intermediates by a rolling circle replication mechanism, similar to geminiviruses and certain bacterial plasmids (reviewed in Hanley-Bowdoin et al., 2000; Laufs et al., 1995). Replication is initiated by a master-Rep (M-Rep) protein that interacts with conserved sequence signals on all genomic DNAs of a given nanovirus (Horser et al., 2001; Timchenko et al., 1999, 2000). Replication of nanovirus DNAs is carried out by cellular enzymes, facilitated and enhanced by the action of Clink, a virus-encoded cell cycle modulator protein. Clink was shown to interact with two different types of cell signalling regulators, pRB family members and SKP1. SKP1 is a constituent of the SKP1/Cullin/F-box protein (SCF) complex that targets proteins for degradation through the ubiquitin-proteasome pathway (Deshaies, 1999). These interactions have been analysed in the yeast two-hybrid system and in vitro by pull-down assays using proteins expressed in E. coli (Aronson et al., 2000).

To prove that these interactions do occur in planta, we developed a nanovirus-based expression system. Vector constructs based on a DNA of Faba bean necrotic yellows virus (FBNYV), a legume nanovirus, were devised to express proteins of choice in cells of leguminous and non-leguminous plants. Agrobacterium-mediated inoculation of leaves and leaf discs with such plasmids along with the replication initiator (M-Rep) encoding DNA leads to efficient replication of nanovirus vector DNA and in planta expression of the gene of choice. Complexes of the encoded proteins can be isolated from plant tissue and analysed. Here, we describe the application of the newly developed nanovirus-based expression system to study in planta protein-protein interaction using SKP1 and Clink as examples.


Design of the nanovirus-based gene expression vectors

Initiation of nanovirus DNA replication depends on the interaction of the viral M-Rep protein encoded by DNA-R with sequence signals present on all individual DNA components of a nanovirus (designated C2 in Timchenko et al., 1999). Initial experiments in leaf discs of Nicotiana benthamiana, a non-host of FBNYV, showed that DNA-C (designated C10 in Aronson et al., 2000), the Clink encoding component of FBNYV, conducted high level expression of that protein, in particular, when the M-Rep triggered replication of DNA-C. Therefore DNA-C was modified to create a versatile vector replicon. DNA-C was cloned in a way not to interrupt the clink gene (Figure 1). For efficient in planta replication a partially redundant copy of DNA-C with a 350-bp direct repeat from the non-coding region of DNA-C was constructed (Figure 1b,c). For easy detection of expressed proteins, triple flu- or myc- sequences coding for an HA or Myc epitope, respectively, were inserted after the ATG codon (Figure 1c). To facilitate insertion of any further coding sequence to be expressed by the replicon, a polylinker with a choice of restriction sites was inserted replacing the clink coding DNA (Figure 1d). For Agrobacterium mediated DNA transfer to plant cells, the expression cassette carrying the modified nanovirus replicon was inserted into the T-DNA of pBin19 (Hoekema et al., 1983).

Figure 1.

Construction of nanovirus-based replicons.

(a) pMA104: DNA-C was modified by site-directed mutagenesis, adding a NcoI site at the start codon and a StuI site after the stop codon of clink.

(b) pMA120: containing 350 bp of the non-coding region of DNA-C.

(c) pMA144: Modified DNA-C was inserted into the SmaI site of pMA120 (b), yielding a partially redundant copy of DNA-C. 5′-triple flu- or myc- tagged versions of clink are also available (pMA121, pMA122).

(d) Clink-less DNA-C versions with a versatile polylinker sequence for expression of non-tagged, HA- or Myc-tagged proteins (pMA225, pMA226, pMA227).

The grey arrow represents the clink coding sequence and the black rectangles the DNA-C redundancy flanking the SmaI sites. Unique restriction sites in the polylinker of the DNA-C replicon are shown in bold. Dotted lines represent the plasmid backbone with restriction sites of the pBKSII polylinker on both sides. Relevant restriction sites are shown.

No protein expression from the nanovirus-based vector in Agrobacterium

A number of eukaryotic promoters function in Agrobacterium, demanding introns to interrupt the coding sequences to be studied in transient gene expression assays (Ferrando et al., 2000). For example, when controlled by the CaMV 35S promoter Clink expression is detected in protein extracts of agrobacteria (Figure 2, lane 4). However, no Clink expression is seen in Agrobacteria when under the control of its cognate promoter from DNA-C, whether DNA-R is present (Figure 2, lane 3) or not (Figure 2, lane 2). By contrast, Clink is readily detected in N. benthamiana leaf discs when replication of DNA-C is triggered by DNA-R (Figure 2, lane 6).

Figure 2.

Expression of Clink in Agrobacteria and plants.

Western blot analysis of proteins from Agrobacteria and N. benthamiana, fractionated by 15% SDS-PAGE. Lane 1–4: Extracts from Agrobacteria containing pBin19 (lane 1), a dimer of FBNYV DNA-C in pBin19 (lane 2), dimers of FBNYV DNA-C and DNA-R in pBin19 (lane 3), and pMA12 expressing Clink under control of the CaMV 35S promoter (lane 4).

Lanes 5, 6: Protein extracts corresponding to 3 mg of N. benthamiana leaf discs inoculated with the same Agrobacteria whose proteins were loaded in lanes 1 and 3, respectively. Blots were probed with a polyclonal anti-H6-Clink antiserum.

The migration of pre-stained marker proteins (kDa) (New England Biolabs/Ozyme, Saint Quentin Yuelines, France) are indicated on the left. The arrow marks the position of Clink.

The nanovirus vector replicates and expresses foreign genes in planta

To test the replication and expression capacity of the FBNYV-based system in plant cells, we used the reporter genes GFP and uidA coding for the green fluorescent protein (GFP) and β-glucuronidase (GUS), respectively (Pang et al., 1996; Vancanneyt et al., 1990). Replicons encoding the respective reporter proteins were used to agroinoculate N. benthamiana leaf discs along with a construct carrying redundant copies of DNA-R and DNA-C as expression of Clink significantly increases replication of FBNYV DNA in planta (Aronson et al., 2000). Results of representative experiments are shown in Figures 3 and 4.

Figure 3.

Detection of GFP expressed in planta using the nanovirus-based expression system. Leaf discs from N. benthamiana (a), and M. truncatula (b) or intact M. truncatula leaves (c) inoculated with Agrobacteria carrying binary vectors with redundant copies of DNA-R, DNA-C along with the modified nanovirus replicon expressing GFP.

GFP was revealed under excitation conditions as described. The respective magnifications are indicated under the photographs.

Figure 4.

β-glucuronidase activity in transfected N. benthamiana leaf discs.

(a) Enzyme activity measured in extracts of leaf discs (four independent experiments) agroinoculated with pBin19 derivatives carrying DNA-R, DNA-C, and the nanovirus replicon encoding β-glucuronidase (uidA replicon) was compared to that expressed from a 35S CaMV promoter (p35S-uidA). Enzyme activity is given in µm mg−1 of protein min−1, determined at 24-h intervals for the days post-infection indicated (dpi). The activity from the control leaf discs carrying DNA-C and DNA-R was below the level of detection.

(b) Extent of nanovirus replication assessed by Southern blot. DNA from the leaf discs inoculated as in (a) was extracted and analysed by Southern blotting using uidA as a probe. The amount of viral replicative forms produced at 24-h intervals was quantified by PhosphorImager (Molecular Dynamics, Sunnyvale, CA, USA) and is shown in arbitrary units.

In N. benthamiana GFP expression was restricted to the periphery of the leaf discs, and was rarely visible inside the disc, except when internal wounding had occurred (Figure 3a). When Agrobacteria were vacuum-infiltrated into Medicago truncatula leaf discs, GFP expression foci were observed all over the disc (Figure 3b). Apparently, this technique yields numerous infection sites for Agrobacteria and leads to a higher overall T-DNA transfer and subsequent gene expression. Syringe infiltration of M. truncatula leaves still attached to the plant also yielded high GFP expression levels (Figure 3c).

GUS expression, monitored by enzyme activity, reached its highest level at around 4–5 days post-inoculation (Figure 4). At this time, a six times higher activity was observed when GUS was produced by the nanovirus-based expression system compared to the level obtained when the uidA gene expression was driven by the CaMV 35S promoter (Figure 4a). With both GFP- and uidA reporter genes the amount of protein or the level of enzyme activity was positively correlated with the viral DNA replication levels (Figure 4b).

These results demonstrated that nanovirus-based vectors can be used to express foreign proteins in two different plant species. Moreover, the M-Rep efficiently initiates replication of recombinant DNA-C replicons larger than the original 1000 nt DNA-C molecules, as the uidA gene used is three times larger than clink.

Efficient isolation of clink-H6 expressed in plants

For proper characterization of proteins, isolation from their natural environment is important. For that purpose, the nanovirus-based vector was used for in planta expression of Clink carrying a hexahistidine (H6) tag at the C-terminus. A 1.3-mer of the respective clink component (DNA-C:H6) was inserted 3′ of a dimer of DNA-R on the same T-DNA. Following N. benthamiana leaf disc transfection, the tagged protein was isolated by immobilized metal affinity chromatography (IMAC) from total protein extracts. Proteins bound to the resin were eluted by competition with imidazole, and Clink was readily detected by Western blot with anti-H6-Clink polyclonal antibodies (Figure 5), or a monoclonal anti-H6-tag antibody (results not shown). Clink-H6 (20.8 kDa) migrated approximately as expected for the tagged version of the protein.

Figure 5.

Clink-H6 isolation from N. benthamiana leaf discs by IMAC.

Western blot of a nitrocellulose membrane with proteins from N. benthamiana leaf discs at 4 days after inoculation expressing either Clink (19 kDa) or Clink-H6 (20.8 kDa), fractionated by SDS-PAGE. Leaf discs were inoculated with Agrobacteria carrying redundant copies of DNA-C and DNA-R or DNA-C:H6 and DNA-R, respectively. Lanes marked Clink or Clink-H6 represent total protein extracts loaded. Lanes 1–5 show an elution profile of the IMA column by 400 mm imidazole. The Western blot was probed with a polyclonal anti-H6-Clink antibody.

Detection of in planta protein-protein interaction

Since the advent of the yeast-two hybrid system, a vast number of potentially interacting proteins from many organisms have been described, but proof of their interaction in the original organism is not always feasible. Using yeast two-hybrid and pull-down assays, we had shown that Clink interacts with a plant partner protein, SKP1 (Aronson et al., 2000), but its significance in planta remains to be demonstrated. In order to prove that the two proteins interact in plant cells, they were co-expressed in N. benthamiana leaf discs taking advantage of the nanovirus expression system. SKP1 was tagged with a Myc epitope to facilitate its serological detection. Agrobacteria with the Myc-SKP1 nanovirus replicon as well as as Agrobacteria containing the Clink- or Clink-H6 replicon along with the M-Rep encoding replicons on the same T-DNA were used to transfect N. benthamiana leaf discs. Proteins were then extracted and isolated by IMAC following the scheme described above. Myc-SKP1 co-eluted with Clink-H6 and was readily detected by Western blotting in the eluted protein fractions when both proteins were co-expressed in leaf discs (Figure 6b, lane 6). Complexes between Clink and Myc-SKP1 were also revealed when a Clink variant with an amino-terminal H6-tag was used for affinity purification (data not shown). To rule out the possibility that Myc-SKP1 by itself binds to the affinity resin or to plant proteins retained non-specifically, the same experiment was carried out using the non-tagged version of Clink. Both Clink and Clink-H6 stimulated replication leading to comparable protein amounts produced by the co-transfected replicons (Figure 6a, lanes 2, 3 and Figure 6b, lanes 2, 3). However, no Myc-Skp1 was retained by the affinity resin when co-expressed with non-tagged Clink (Figure 6b, lane 7). We also used a Myc-tagged version of GFP as a protein that does not form specific complexes with Clink. The lack of retention of GFP-Myc by Clink-H6 on the affinity column (Figure 6b, lane 8) confirms that Clink-H6 has no intrinsic affinity to a Myc-tag.

Figure 6.

Isolation of Clink-SKP1 protein complex from N. benthamiana cells.

Total protein extracts of N. benthamiana leaf discs prepared 4 days after inoculation were analysed by SDS-PAGE prior to isolation (‘extract’) or after IMAC (‘eluted proteins’). Proteins expressed following inoculation by the Agrobacteria carrying the respective constructs are indicated above each lane.

Panel A: Blot probed with a polyclonal anti-H6-Clink antiserum.

Panel B: Blot probed with a monoclonal anti-c-myc antibody. The positions of pre-stained protein size markers are indicated on the left.

Therefore, the nanovirus-based protein expression system allows detection and isolation of specific protein complexes formed in planta.


We present a versatile highly efficient gene expression system for plants based on the replicative capacity of nanovirus DNAs. DNA-C of Faba bean necrotic yellows virus was modified to serve as the basic replicon, as its promoter conducts elevated expression of the encoded Clink protein in plant cells (Figure 2). Also for Banana bunchy top virus, a monocot-infecting nanovirus, the clink promoter was shown to be strong and constitutive, both in embryonic banana cells and in transgenic tobacco plants (Dugdale et al., 1998). Since the multiplication of FBNYV replicons is dependent on the M-Rep protein and is greatly enhanced by the action of Clink, the nanovirus-based expression system comprises DNA-R, DNA-C and a modified nanovirus replicon containing the gene of choice. As nanoviruses replicate through DNA intermediates, the vectors may also be used for studying replication and transcription factors. A potential in planta interaction of the expressed proteins is readily assessed, and the complexes can be isolated and subjected to biochemical analyses. As the nanovirus-based expression system does not provide any viral movement or capsid proteins, movement or packaging constraints do not apply. Therefore, the movement of proteins, RNAs or protein-nucleic acid complexes encoded by nanovirus-based replicons may be studied as well. Furthermore, since only two proteins of the original virus (M-Rep and Clink) are required, any cellular (defence) reaction is limited to a potential one against these proteins. Hence, host responses triggered by nanovirus-based expression system itself and interfering with the processes to be studied may be negligible.

The natural host range of FBNYV includes a wide variety of dicotyledonous plant species (Franz, 1997), including Arabidopsis thaliana (C. Ramirez, ISV, Gif sur Yvette, unpublished results). Agroinoculation or introduction of DNA by biolistic techniques permit the use of non-natural hosts as, for example, N. benthamiana.

Several DNA and RNA plant virus-based expression vectors have been described (Dawson et al., 1989; Matzeit et al., 1991; Shivprasad et al., 1999; Timmermans et al., 1994). Meanwhile, improved knowledge of the host responses to RNA virus-based protein expression explains the initial difficulties associated with the use of certain viruses for gene expression as being due to virus-induced gene silencing and co-suppression (reviewed in, Vance and Vaucheret, 2001; Voinnet, 2001). Ways to circumvent these problems are suggested by the biology of these viruses, as some of them encode gene-silencing suppressors to overcome the host defence (reviewed in Voinnet, 2001). As a result, RNA virus-based protein expression has become very promising (reviewed in, Yusibov et al., 1999). Nevertheless, the ease by which foreign DNA is replicated and expressed by the nanovirus-derived vectors, both in leaf discs or entire leaves, allows rapid characterization of multiple protein interactions.

Here, we present versatile nanovirus-based replicative expression vectors and describe their use to demonstrate the in planta interaction between the host protein SKP1 and Clink, a virus encoded protein. The validation of this interaction in a higher eukaryote further strengthens the role of Clink as a modulator connecting ssDNA virus multiplication with the ubiquitin-dependent protein-processing pathway of the cell.

Experimental procedures

Plasmid constructions

Oligonucleotide sequences used in this study are listed Table 1.

Table 1. Sequences of the oligonucleotides used
pr1ACAGGATCCATGGGTCTGAAATATTTCForwardFor amplification of clink, creates a NcoI site
pr10GTTGTTCTTACAACGACCATGGGTCTGAAAForwardCreates NcoI site at the start codon of DNA-C
pr11TTTCAGACCCATGGTCGTTGTAAGAACAACReverseCreates NcoI site at the start codon of DNA-C
pr12CCCGGTACCATGGCCTACCCATACGACGTTCCForwardFor amplification of triple flu sequence with NcoI ends
pr13GGATATCGCCATGGCTGCATAGTCCGGGACReverseFor amplification of triple flu sequence with NcoI ends
pr14GGATCCTCTACCATGGAACAAAAGTTGATTForwardFor amplification of triple myc sequence with NcoI ends
pr15TCTAGAGGCCATGGTCAAGTCTTCTTCTGAReverseFor amplification of triple myc sequence with NcoI ends
pr16GTTCTTACAACGAAGATGGCCTACCCATACForwardRemoves NcoI site at the beginning of flu-clink
pr17GTATGGGTAGGCCATCTTCGTTGTAAGAACReverseRemoves NcoI site at the beginning of flu-clink
pr18GTTCTTACAACGAAGATGGAACAAAAGTTGForwardRemoves NcoI site at the beginning of myc-clink
pr19CAACTTTTGTTCCATCTTCGTTGTAAGAACReverseRemoves NcoI site at the beginning of myc-clink
pr20GTAATTAGTTGAGGCCTTGTAATTAAATGForwardCreates StuI site at the end of clink
pr21CATTTAATTACAAGGCCTCAACTAATTACReverseCreates StuI site at the end of clink
pr22CAGGAATTCGCCATGGGAACAACAAGAAAGForwardFor amplification of SKP1, creates NcoI site
pr23TAATCCATGGTACGTCCTGTAGAAACCCCForwardFor amplification of uidA, creates NcoI site
pr24ATCCCCCGGGTCATTGTTTGCCTCCCTGCReverseFor amplification of uidA, creates SmaI site
pr25GAAGATCTACTAATAACAATATCCReverseFor amplification of clink, creates BglII site
pr26GAAGGCCTTTAGTGATGGTGATGTGReverseFor amplification of clink-H6, creates StuI site
pr28CCTGAGATCTCTCGAGAAGCTTGATATCGAATTCCReversePolylinker with NcoI and StuI overhangs
pT7GTAATACGACTCACTATAGGGC anneals to the pBKSII backbone

Expression of clink from 35S CaMV promoter

Clink DNA was released from pQE30-Clink (Aronson et al., 2000) by BamHI- and SalI-cleavage and inserted between the BamHI and SalI sites of pBluescript KSII(+) (pBKSII) (Stratagene Europe, Amsterdam, The Netherlands), yielding pMA46. Subsequently, pMA46 was digested by BamHI and KpnI, and the DNA containing clink was inserted between the BamHI and KpnI sites of pBI121Gd, yielding pMA12. The binary vector pBI121Gd is a derivative of pBI121 (Jefferson et al., 1987) whose uidA gene sequence was replaced by a polylinker sequence of pUC18 encompassing the XbaI and SstI sites (J. Brevet, ISV, Gif sur Yvette, unpublished).

DNA-C derived expression vectors

A monomer of FBNYV DNA-C encoding the clink gene was excised by SmaI from a cloned DNA-C dimer (Timchenko et al., 1999) and inserted into the SmaI site of pBKSII. An NcoI site at the clink start codon and a StuI site 3′ of the stop codon were introduced by site-directed mutagenesis (Quikchange, Stratagene) using primers pr10 and pr11, and pr20 and pr21, respectively. The resulting plasmid was designated pMA104 (Figure 1a). Triple flu- and myc- sequences were amplified by PCR with NcoI flanking primers (pr12 and pr13, and pr14 and pr15) using plasmids pUC19flu and pUC19myc, as templates, respectively. After digestion with NcoI, the amplified DNA fragments were introduced into the NcoI site of clink in pMA104. The sequence preceding the ATG start codon in the flu-clink and myc-clink constructs was restored to the original DNA-C sequence by site-directed mutagenesis using primers pr16 and pr17, and pr18 and pr19, respectively, thus destroying the first NcoI site. The plasmids expressing the HA-tagged and the Myc-tagged Clink protein from DNA-C:flu and DNA-C:myc were designated pMA107 and pMA108.

For replication in plant tissue, modified viral DNA coding for Clink, HA-Clink and Myc-Clink containing a 350-bp redundancy were constructed. DNA-C was released from pMA104 by SmaI digestion, self ligated and further digested by DraI and StuI. The released 350 bp fragment containing the non-coding region (sequence co-ordinates 826–177 of DNA-C) was subsequently inserted into the SmaI site of pBKSII, yielding pMA120 (Figure 1b). Plasmids pMA104, pMA107 and pMA108 were digested with SmaI and the inserts, DNA-C, DNA-C:flu, and DNA-C:myc, respectively, were transferred into the unique SmaI site of pMA120, yielding pMA144, pMA121, and pMA122 (Figure 1c).

To facilitate insertion of any further sequences to be expressed, the clink coding sequences present in pMA144, pMA121, and pMA122 were replaced by a polylinker. For this purpose, the polylinker originating from pBKSII in plasmids pMA144, pMA121, and pMA122 was removed by EcoRI- and XhoI-cleavage, followed by end-filling with Klenow fragment of PolI and ligation. The resulting DNAs were then cut by NcoI and StuI, releasing the clink gene, and a polylinker encoded by annealed oligonucleotides pr27 and pr28 was inserted between the NcoI and StuI sites, yielding pMA225, pMA226, and pMA227, respectively (Figure 1d).

To add an H6 tag sequence 3′ of clink, its coding sequence was amplified by PCR (pr1 and pr25) from pMA144. The PCR product was digested by NcoI and BglII and inserted between the NcoI and BglII sites of pQE60 (Qiagen, Courtaboeuf, France), yielding pMA215. The thus tagged Clink (20.8 kDa) has eight additional amino acids (…RS{H}6) at its C-terminus. Clink-H6 DNA was amplified by PCR (pr1 and pr26) from pMA215, the PCR product was digested by NcoI and StuI and inserted between the NcoI and StuI sites of pMA144, yielding pMA216. The thus modified DNA-C from pMA216 as well as a dimer of DNA-C cloned in pBKSII (Timchenko et al., 1999) were transferred as XbaI-KpnI and SacI-KpnI fragments, respectively, into pBin19, already containing a dimer of DNA-R (Timchenko et al., 1999), yielding pMA217 and pMA167, respectively.

Marker gene constructs

DNA encoding GFP (Pang et al., 1996) containing a double myc-tag sequence at the 3′ end was released from pGFP (I. Couchy, P. Ratet, ISV, Gif sur Yvette, unpublished) by NcoI and SmaI and inserted between the NcoI and StuI sites of pMA144, yielding pMA151. Subsequently, the modified nanovirus DNA was released from pMA151 by BamHI and HindIII digestion and transferred into pBin19, yielding pMA157. The uidA gene coding for GUS containing an intron was amplified by PCR from p35S GUS INT (Vancanneyt et al., 1990) using primers pr23 and pr24. The PCR product was digested with NcoI and SmaI and inserted between the NcoI and StuI sites of pMA144, yielding pMA149. Subsequently, the modified nanovirus DNA was transferred from pMA149 as a XbaI-KpnI fragment into pBin19, yielding pMA155.

SKP1 expression vector

The SKP1 coding sequence was amplified by PCR from the plasmid pAD-SKP1 (Aronson et al., 2000) using primers pr22 and pT7. After cleavage by XhoI, end-filling by Klenow fragment of PolI and cleavage by NcoI, the PCR product was inserted between the NcoI and StuI sites of plasmid pMA108, yielding pMA110. A dimer of the thus-modified DNA-C was constructed by linearizing pMA110 with SmaI and directly assembling two SmaI fragments in the SmaI site of pBKSII, yielding pMA117. Subsequently, the dimer was transferred from pMA117 as a XbaI-KpnI fragment into pBin19, yielding pMA116.

Isolation of proteins expressed in planta

The pBin19 derivatives described above were introduced into Agrobacterium tumefaciens strain LBA4404 (Hoekema et al., 1983; Ooms et al., 1982) by electroporation (Mozo and Hooykaas, 1991).

Agrobacteria were used to inoculate leaf discs of N. benthamiana or M. truncatula R108-1c3 as described by Horsch et al. (1985) and Trinh et al. (1998), respectively. Agrobacteria, resuspended in induction medium (Yang et al., 2000) were injected into M. truncatula leaves with a 1-ml needle-less syringe.

Total plant protein extracts were prepared from about 700 mg of leaf discs, harvested 4 days post-inoculation, and ground to a fine powder in liquid nitrogen. Ten ml of ice-cold TN buffer (20 mm Tris HCl, pH 8.0, 100 mm NaCl) containing 1 mm PMSF, 10 mmβ-mercaptoethanol and 0.1 KI units aprotinine (Sigma-Aldrich, Saint Quentin Fallavier, France) were added to the powder, and the slurry was centrifuged for 30 min at 48 000 g at 4°C. The protein concentration of the supernatant was estimated by Bradford assay (Bio-Rad Protein Assay) and adjusted to 1 mg ml−1. Isolation of histidine-tagged proteins was performed by IMAC. Ten mg of total protein were loaded onto a column containing a packed volume of 250 µl Sepharose CL-6B beads with Co2+ as the metal chelator (Talon, Clontech/Ozyme, Saint Quentin Yvelines, France), equilibrated with TN buffer. The column was washed with 15 ml of TN buffer, followed by 3 ml of TN buffer containing 10 mm imidazole. Proteins were eluted with 3 ml of 400 mm imidazole in TN buffer, and fractions of 0.2 ml were collected. Proteins in the fractions (10 µl aliquots) were resolved by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and analysed by Western blotting (Ausubel, 1987) using anti-H6 (Qiagen) and anti-c-myc (Boehringer-Mannheim, Mannheim, Germany) monoclonal antibodies. Antigen-antibody complexes were revealed by peroxidase reaction using the ECL detection kit (Boehringer-Mannheim) or by alkaline phosphatase reaction as described (Ausubel, 1987).

Clink isolation and production of anti-Clink serum

Clink protein with an N-terminal H6 tag was expressed in E. coli as described previously (Aronson et al., 2000). Native H6-Clink was isolated by IMAC on Talon resin (as above) and used to immunize a rabbit. Serum from the fifth bleeding was cross-adsorbed against N. benthamiana protein extracts followed by IgG isolation using affinity chromatography on a protein A-column (Pharmacia Biotech, Orsay, France). The purified antibodies were used at 1 : 500 dilution for detection of Clink in Western blots.

Protein extracts from Agrobacteria

Agrobacterium strain (LBA4404) harbouring the respective pBin19 derivatives were grown to an OD600 = 4 and cell pellets from 1 ml of culture were resuspended in 200 µl of loading buffer (20 mm Tris HCl, pH 8.0, 100 mm NaCl, 1 mm PMSF, 3% SDS, 10% glycerol, 1.2% β-mercaptoethanol, 0.003% bromophenol blue). Cells were lysed by three cycles of freezing and thawing using liquid N2 and subsequent boiling for 7 min. Cell debris were pelleted at 14 900 g for 10 min, and proteins in the supernatant (15-µl aliquots) were fractionated by SDS-PAGE and analysed by Western blotting.

Monitoring marker gene expression

GFP fluorescence was monitored using a Leica MZ FLIII UV lamp (Leica Microsystems S.A., Rueil-Malmaison, France) with the following filters: GFP3 (excitation filter 470/40 nm, emission filter 525/50 nm) and B (excitation filter 470/40 nm, emission filter 515 nm). The latter filter does not eliminate background fluorescence of chlorophyll.

β-glucuronidase activity was essentially assayed as described by Jefferson (1987) using 1.25 mm of 4-methylumbelliferyl β-d-glucuronide (MUG) as substrate. Release of 4-methylumbelliferone (MU) was measured by a spectrofluorimeter (Fluoroskan II, Labsystems S.A., Cergy Pointoise, France) at an excitation wavelength of 365 nm and an emission wavelength of 455 nm every 2 min for a 2-h period. GUS activity was calculated based on the molar fluorescence of MU (Sigma-Aldrich).


We are indebted to C. Breda, F. de Kouchkovsky and A. Müller for excellent technical assistance, L. Troussard for DNA sequencing and N. Mansion for photographs. We thank Drs I. Couchy, and P. Ratet for plasmids and Dr J. Brevet and D. Thomas for fruitful discussions. M.N.A and A.C. acknowledge fellowships from the French Ministry of Research. This work was supported in part by the European Commission (INCO-DC Programme; ERBIC18-CT96-0121) and a fellowship from the Fondation pour la Recherche Médicale to M.N.A.