AU2001264389A1 - Protein translocation into plant cells - Google Patents

Description

PROTEIN TRANSLOCATION INTO PLANT CELLS

The present invention relates to a method of effect- ing a change in a cell, wherein a transfer system is contacted with the cell to be changed, said transfer system comprising a membrane with in it a protein transport system comprising a pore which comprises a VirB complex and VirD4 protein, wherein the transfer system comprises a fusion protein or is capable of making a fusion protein which is introduced into the cell by means of the protein transport system. Such a method is suggested by Zhou, X-R. et al . (Journal of Bacteriology, 181 (14), p. 4342-4352 (1999)). This publication discloses that the protein VirE2, a single stranded DNA-binding protein of Agrobacterium tumefaciens . can be transferred into a host cell by a tumorigenic Agrobac- terium. Zhou et al . disclose that they were only at amino acid 39 (as from the N-terminus, as is customary in the art) able to introduce a heterologous amino acid sequence which was transferred to the host cell. They conclude that it is not practically feasible to base a protein delivery system on fusions at one of the two ends of VirE2. Small proteins at least are tolerated as an insertion at amino acid 39.

The object of the present invention is to provide a method according to the preamble, which is simple to perform and if desired, is performed without native activity of VirE2 in the cell to be changed.

Accordingly, the method according to the invention is characterized in that a fusion protein BA is introduced into the cell to be changed which fusion protein BA i) comprises as a first part A an oligopeptide comprising 3 adjacent amino acids identical to or corresponding to the amino acids of the C-terminal amino acids 1-40 of VirF, VirD2, VirE2 , VirE3, or MobA and ii) comprises as a second part B a polypeptide capable of exercising a cell-changing activity in the cell to be changed, wherein the C-terminal end of the polypeptide is linked to the N-terminal end of the first part A, under the condition that if the fusion protein comprises a first part A derived from VirE2, the fusion protein does not comprise the last 84 N-terminal amino acids of VirE2.

Surprisingly it has been found that, provided the above-mentioned conditions are met, a protein can be introduced from the outside into a cell to be changed, which protein is formed by coupling the polypeptide (second part B) to an (internal) N-terminal amino acid of VirF, VirD2, VirE2, VirE3 or MobA (first part A) , as a result of which the activity of the introduced second part of B may be expressed in the cell to be changed, as a result of which the cell is changed. This change may be temporary (reversible) or permanent (irreversible) . The amino acid sequences of VirF, VirD2, VirE2, VirE3 en MobA are represented by the respective sequence nos. 1-5. A strain comprising a plasmid (LBA8250, a Agrobacterium tu efaciens strain, containing plasmid pTil5955. pTil5955 is described by Sciaky et al . , Plasmid 1, p. 238-253 (1978)) coding for all vir-proteins, is in the Phabagen collection of the Centraalbureau voor Schimmelcultures (CBS) , Baarn, the Netherlands under accession number PC2692. MobA protein is known from incQ plasmid RSF1010, as present in J . coli strain K12 C600 with accession number PC-V3110 of the

NCCB collection of the CBS, and described by Scholz, P. et al . in Gene 75., p. 271-288 (1989) .

In the present invention, a corresponding amino acid is an amino acid according to the following table: A, G; S, T D, E N, Q R, K I, L, M, V; and F, Y, .

It goes without saying that the first part A may comprise more adjacent or non-adjacent amino acids, identical to or corresponding to the sequences defined before. Also, one or more non-corresponding, non-identical amino acids may be present as a spacer between the 3 adjacent amino acids and the second part B. Also, such a spacer may be present upstream of the 3 adjacent amino acids.

The first part A may, if desired, comprise more amino acids from amino acid 41 (as from the C-terminus) of VirF, or VirE2 of VirE3 or VirD2 or MobA. This under the condition mentioned for VirE2. In the present invention a vector is understood to be any DNA- or RNA-sequence which, directly or indirectly, leads to the formation of the fusion protein in the transfer system.

The transfer system may be a cell, such as very suitably a bacterial cell, such as in particular Agrobacterium tumefaciens, but may also be an artificial system, such as a mini cell or an artificial vesicle-system.

The fusion protein to be transferred can either be formed in the transfer system itself, for example, by expressing a vector containing a functional gene system that may be expressed yielding the fusion protein, or the fusion protein itself may be introduced into the transport system. In artificial transport systems generally the latter will be the case. An important advantage of such an artificial transport system is that it may be introduced into the environ- ment, for example for treating a crop, without risk of spreading genetic material.

Preferably, the oligopeptide of the first part A comprises RXR or a derivative thereof

Here X is an arbitrary amino acid. This motif con- sisting of three amino acids is believed to be essential for protein transport by the transfer system.

According to a preferred embodiment, the oligopeptide of the first part A comprises RPR, RAR, RQR, RSR or RVR or a derivative thereof. In the present application "a derivative thereof" in association with an (oligo)peptide is understood to be an (oligo)peptide in which at least one amino acid is replaced by a corresponding amino acid. Preferably the fusion protein is introduced into the cell without the introduction of a DNA- or RNA-sequence . In case of the use of Agrobacterium tumefaciens as the transfer system this implies the absence of T-DNA. This means that it is possible to change a cell without introducing genetic material into the cell to be changed.

According to an important embodiment, the fusion protein introduced has recombinase-activity.

In such a case it is in fact possible to effect a change in the chromosomal DNA of the cell to be changed, without introducing additional genetic material (a vector coding for the T-DNA or peptide B) into the cell. An important application is, for example, the removal of a marker gene, whose admittance in the environment is undesirable. In particular, this may include the removal of antibiotic-resistance genes present between DNA-sequences in direct repeat to be recognized by the recombinase.

Conveniently a bacterium of the class of Rhizobiaceae may be used as a transfer system. Such bacteria, amongst which Agrobacterium, Rhizobium and Phyllobacterium, are very suitable for modification of plants, yeasts or fungi. Other bacteria belonging to this family, such as Br cella, are known for their interaction with human and animal cells and may be used for modification thereof.

According a preferred embodiment, use is made of Aqrobacterium, which is known to be very suitable for the modification of both prokaryotes (bacteria) , and eukaryotes (plants, yeasts, fungi, and animal cells) . According to a preferred embodiment, a cell chosen from the group consisting of i) a plant cell; ii) a yeast cell; and iii) a fungal cell, is used as the cell to be modified.

The invention also relates to a vector, said vector being characterized in that it codes for a protein transport system comprising a pore which contains a VirB complex and VirD4 protein, as well as for a fusion protein BA that comprises i) as a first part A an oligopeptide comprising 3 adjacent amino acids identical to or corresponding to amino acids of the C-terminal amino acids 1-40 of VirF, VirD2, VirE2, VirE3 or MobA, and ii) as a second part B a polypeptide capable of effecting a cell-changing activity in the cell to be changed, wherein the polypeptide with its C-terminal end is linked to the N-terminal end of the first part A, under the condition that if the fusion protein comprises a first part A derived from VirE2, the fusion protein does not comprise the last 84 N-terminal amino acids of VirE2. Such a vector may be introduced into a transfer system, such as a bacterium. The vector then possesses all the information that needs to be expressed for the transfer of the fusion protein, making it possible for the fusion protein to be transferred to the cell to be changed. Finally, the invention relates to a vector set, which is characterized in that the vector set comprises one or more vectors coding for a protein transport system comprising a pore which contains a VirB complex and VirD4 protein as well as a further vector coding for a fusion peptide BA which comprises i) as a first part A an oligopeptide comprising 3 adjacent amino acids identical to or corresponding to the amino acids of the C-terminal amino acids 1-40 of VirF, VirD2, VirE2, VirE3 or MobA, and ii) as a second part B a polypeptide capable of exercising a cell-modifying activity in the cell to be modified, wherein the polypeptide of the C- terminal end of the polypeptide is linked to the N-terminal end of the first part A, under the condition that if the fusion protein comprises a first part A derived from VirE2, the fusion protein does not comprise the last 84 N-terminal amino acids of VirE2.

The advantage of such a vector set is that the vector or vectors coding for the protein transport system, may be introduced into a transfer system separate from the further vector coding for the fusion protein. This makes it possible to use a transfer system, in particular a bacterium, as a standard vehicle for modifying a cell, wherein the transport system is provided with a further vector which is expressed in the transport system for effectuating the modification. The invention will be illustrated with reference to the following examples and to the drawing in which the only figure schematically shows a test system to demonstrate Cre- activity in planta .

Bacterial strains

Ag-roJbacterium strain LBA1010 (Koekman et al . , Plasmid 7 (1982); 119-132; Centraal Bureau voor Schimmelcultures; Baarn, the Netherlands, accession number: PC2805) possesses the wild type Ti-plasmid pTiB6 in a C58 chromosomal background. LBA 1100 (Beijersbergen et al . Science 256 (1992), 1324-1327; easily obtained by culturing CBS 102794, Centraal Bureau voor Schimmelcultures; Baarn, the Netherlands deposited on 17 May 2000 in the absence of gentamycin for curing of plasmid pSDM 3155 and screening for the absence of the small (ca. 5,5 kb) plasmid) is a non-oncogenic derivative of LBA1010. To this end both the left and right T regions in pTiB6 as well as tra and occ genes are replaced by a spectinomycin-resistance marker, resulting in plasmid pALHOO, wherein the Vir region remained intact. For the present application several vir mutants, derived from LBA1100, resp. LBA 1142-1150 (Beijersbergen et al . Science 256 (1992), 1324- 1327) were used (Table 1) . LBA2561 contains a precise deletion of the virF gene in pALllOO (Schrammeijer et al . , Mol . Plant Micr. Int. 5 (1998), 429-433) . Transformation and growth of bacterial strains were performed as described elsewhere (Vergunst et al . Nucl . Acids Res. 26 (1998) 2729-2734) or according to general techniques known to the person skilled in the art (Sambrook et al. Molecular cloning. A lab manual. (1989) ) .

Plasmid constructions

The coding region of the ere gene was cloned translationally to the virE2 and virF gene of pTil5955, under control of the respective vir promoter regions. Both N- terminal and C-terminal fusions were made. The plasmid constructions are detailed below. The ere control plasmid

The coding region of the ere recombinase gene, present in plasmid pUC19cre (Mozo & Hooykaas, Mol . Gen. Genet 236 (1992), 1-7), was cloned as a Sphl/EcόRI fragment into pUC21(Vieira & Messing, Gene 100 (1991), 189-194) resulting in pSDM3120. In order to remove the ATG starting codon, a PCR amplification was performed at the 5' -end of the ere gene with the primers crel (5 ' -ggcagatctgTCCAATTTACTG) and cre2 (5'- GATAATCGCGAACATCTTCAGG) on pSDM3120. After digestion of the PCR fragment with Bglll and Nrul (underlined) this fragment was exchanged for the corresponding fragment in pSDM3120 (resulting in creΔATG or pSDM3121) . A Sail fragment of pRAL3248 (Melchers et al . , Plant Mol. Biol . 14 (1990), 249- 259) into which virEl and virE2 (the last 30 3 '-terminal base pairs being absent) including the virE promoter region are located, was cloned into the Xhol restriction side of pSDM3121 (pSDM3122) . After full digestion of pSDM3122 with Bglll and partial digestion with BstYI, followed by isolation of the vector fragment carrying the virE promoter, the virEl coding area, the ATG starting codon of virE2, as well as ere (ΔATG) , the ere gene was translationally linked through self closure to the ATG starting codon of virE2 (pSDM3126) . As a result the expression of the ere gene is controlled by the virE promoter. Subsequently this construct was transferred as a Stul/Xbal fragment to the Smal/Xbal digested plasmid pRL662, resulting in pSDM3147, i.e. the ere control plasmid used in these experiments. The non-mobilisable plasmid pRL662 with a broad host range is obtained by replacing the kana ycin resistance gene as well as the mob region of pBBRl MCS2 (Kovach et al . , BioTechn. 16 (1994) , 800-802) with a gentamycin resistance marker (J. Escudero, European patent application 00200726.8).

The cre.-:virE2 fusion

In order to create translational fusions between ere and virE2 , the STOP codon of the ere gene was removed; to this end a mutation in the STOP codon (Italics) was introduced using PCR amplification on pSDM3126 DNA using the primers ere 6 (5 ' -acσcgtcσactATCGCCATCTT CAGCAGGCGC) and ere 7 (5¹- cCATCGATTGATTTACGGCGCTAAGG) . After digestion with Clal and Sail (underlined) , the corresponding Clal-Sail fragment of pSDM3126 was replaced (resulting in creΔSTOP or pSDM3127) . A Xhol/Notl fragment of pBluevirE2 (ΔATG) containing the virE2 coding area, without ATG start codon, was subsequently ligated into the vector pSDM3127 digested with Sail and JVotl (pSDM3128) . pBluevirE2 (ΔATG) was the result of cloning the VirEl ~virE2 region (Xhol-Smal ) of plasmid pRAL3248 (Melchers et al., Plant Mol. Biol . 14 (1990), 249-259) in pBLUEscript (Xhol/BcoRV) (Alting-Mees and Short. Nucleic Acids Res. 17 (1989) : 9494) , subsequent to which virEl and the ATG start codon of virE2 were removed using a Xhol-Stul linker (5¹- tcgaGATCTTTCTGGCAATGAGAAATCCAGG and 5'- CCTGGATTTCTCATTGCCAGAAAGATC. The cre::virE2 fusion was subsequently transferred to pRL662 (Smal/Xbal ) as a Stul/Xbal fragment of pSDM3128, resulting in pSDM3129 (cre::virE2 fusion) .

The virE2::cre fusion A Sail fragment of pRAL3248 comprising virEl and virE2 (the last 30 3' -terminal bases being absent) was cloned into pIC19R (Marsh et al . , Gene 32 (1984): 481-485) (pSDM3123) . In the subsequent cloning steps ere was fused to the 3 '-end of virE2, wherein at the same time the last 30 3'- terminal bases of virE2, absent in pSDM3123, were restored. To this end a Bglll/Nrul fragment of pSDM3122, comprising the 200 5' -terminal bases of ere, without the ATG start codon, was cloned into pIC19R (pSDM 3151) . A Sail /Bglll linker, consisting of the last 30 3' -terminal bases of virE2, was synthesized in order to remove the STOP codon of virE2

(Italics) (5' -TCiSAQCGCGTAGCCAAAGCGTCAACAGCTTTcga and 5'- gatctcgAAAGCTGTTGACGCTTTGGCTACGCGG) and to effect translational fusion with ere. This linker was cloned into pSDM 3151 (pSDM3152) . However, after sequence analysis, the Bglll site between the 3 '-end of virE2ΔSTOP and the 5' -end of creΔATG was shown to be lost due to a single base- substitution. This mutation is not present in the region coding for virE2 or ere, and has no effect on the reading frame of the fusion. The Sail fragment of pSDM3123 was introduced into the Sail site of pSDM3152 (pSDM3157) , cloning the complete virE2 sequence in frame with the 5 '-end of ere . A Nrul fragment of about 600 base pairs from pSDM3157 (the 3'- end of virE2ΔSTOP and the 5' -end of creΔATG) was subsequently used to replace the Nrul fragment in pSDM3148 (pSDM3148 is the result of cloning a Stul/Xbal fragment of pSDM3122 into pRL662 Xbal/Smal ) resulting in a translational fusion between the complete virΞ2 region and ere (pSDM3166, i.e. the virE2::cre fusion) under control of the virE promoter, and joint expression of VirEl .

The virF::ere fusion

A 600 base pair SacI/EcoKV fragment of pRAL7088 (Schrammeijer et al. MPMI 11 (1998) 429-433) containing the 5' flanking side of virF and the ribosomal binding place was cloned into pBluescriptSK^" (Alting-Mees and Short. Nucleic Acids Res. 17 (1989): 9494), resulting in pSDM3183. The nuclear localization signal (NLS) of simian virus 40 (SV40) was synthesized with an EcoRV blunt and a 3' Sail sticky end (5 ' -ATCATGGATAAAGCGGAATTAATTCCCGAGCCTCCAAAAAAGAAGAGAAAGGTCG- AATTGGGTACCGG and the complementary strand) , and cloned into pSDM3183, resulting in pSDM3184. The virF gene without ATG and STOP codon was cloned in two steps in frame after SV40 NLS. The ATG start codon was removed by cloning a BamHl/Nsil linker (5' -GATCCGAAATTCGAGTTTGCGTGATGCA) in the BairiΑl en Nail sites of pRAL7088 (pSDM3192) . Subsequently a BaπiHI/SacI fragment of 1,5 kilobase pairs of pSDM3192 was cloned into pIC19H (pSDM3193) . A Sall/Xhol fragment of 500 base pairs of pSDM3193, comprising virFΔ498-609ΔATG was cloned in frame with the SV40 NLS into the Xhol site of pSDM3184, resulting in pSDM3185. Subsequently the STOP codon of virF was removed using two primers (5'- ATCCCTAACTTGGTCTTCAAC and 5'- cttagatcTAGACCGCGCGTTGATCGAGG) in a PCR reaction on pRAL7088. The PCR fragment was sub-cloned as a 175 bp fragment into vector pGEM T (Promega) . A 16 base pair Stul/Bglll linker (5'- cctcgagcccgggata and 5' -gatctatcccgggctcgagg) was cloned into pSDM3121 digested with Stul/Bglll to introduce a Xhol site (underlined) and to effect further in frame cloning of ere with virF (pSDM3121-L) . Subsequently the 3 ' -end of virF was cloned in frame into the Xhol/Bglll sites of pSDM3121-L as a Xhol/Bglll fragment of about 110 base pairs from the pGEM-T vector to the 5 ' -end of ere (resulting in pSDM3186) . A 1,2 kb Xhol/ Sail fragment of pSDM3186 was ligated into the Xhol site of pSDM3185, resulting in pSDM3187. The virF::cre fusion was subsequently introduced into pUC28 (Benes et al . , Gene 130 (1993) ; 151-152) as a Sacl/Pstl fragment, and subsequently ligated from this plasmid as an BcoRI fragment into the EcoRI site of pRL662, resulting in pSDM3153, i.e. the NLS ::virF: : ere fusion plasmid.

The ere: :virF fusions A 0.66 kb fragment of pSDM3184, having the 5' flanking side of virF and the SV40 NLS sequence, was eventually cloned via a Sad/ Sail cloning step into pIC19H, followed by a Hindlll/Sall cloning into pIC19R, as a Sall/Xhol fragment into the Xhol site of pSDM3121. This resulted in an in frame fusion of the SV40 NLS to the 5' -end of ere

(pSDM3188) . The STOP codon of ere in pSDM3188 was subsequently removed by replacing the Clal/ Sail fragment (with the 3 '-end of ere including STOP) with the Clal /Sail creΔSTOP fragment of pSDM3127 (pSDM3179) . A 690 bp Sall/Eagl fragment of pSDM3193 comprising virFΔATG was cloned into pSDM3179 digested with

Sall/Notl , resulting in pSDM3189. A Hindlll/Xbal fragment of pSDM3189 was cloned into the HindiII en Xbal sites of pRL662, resulting in pSDM3154 (the NLS :: ere: :virF fusion plasmid) .

In addition, a number of deletions was made of virF and cloned translationally to the 3 '-terminus of ere. Using the primers F126 (5 ' -acgcgtcgaCCTGTCGAGTCGGCTGAG, position 127 in virF) and pflF2 (5 ' -GACCAGCACACTTAGATACC, in the DNA sequence adjacent to 3¹ virF) , a PCR reaction was performed on pRAL7088. The PCR fragment of 780 base pairs was cloned into vector pT7pBlue-T (Novagen) (pSDM3194) . A Sall/Eagl fragment (virFΔl-126) of 555 base pairs of pSDM3194 was cloned into pSDM3179, resulting in pSDM3190. Cloning the NLS : :cre : -.virFΔl- 126 fusion into vector pRL662 as a Hindlll/Xbal fragment resulted in pSDM3155 (the NLS: :cre: :virFΔl-126 fusion plasmid) .

The 113 3 '-terminal nucleotides of virF were cloned in frame with ere. To this end pSDM3189 was transferred as a Hindlll/Xbal fragment to pUC18 (Yanish-Perron et al . , Gene 33 (1985); 103-119), resulting in pSDM 3172. Digestion of pSDM3172 with Xhol and Sail, removal of the 5 '-sticky end using Mung bean nuclease followed by self-closure of the vector resulted in an in frame fusion of the last 112 3'- terminal base pairs of virF with ere (pSDM3173) . Transfer of the NLS: :cre: :virFΔl-498 fusion to pRL662 as a Xbal/Hindlll fragment resulted in plasmid pSDM3174 (the NLS : :cre : :virFΔl- 498 fusion plasmid) .

The plasmids were transferred to the bacterial strains shown in Table 1. In addition, the plasmid pSDM3191, on which the genes virD3 and virD4 are located under the control of the virD promoter, was mobilized in LBA1147 and LBA1150. pSDM3191 is the result of cloning a 4.4 kilobase pair Bair&ϊl fragment of pMP3 (Vogel and Das, J. Bacteriology 174 (1992) ; 5161-5164) in the BarriHI restriction site of pLM997

(Melchers, unpublished) , which is the result of replacing the T-region of pBin 19 (Bevan et al . , Nucleic Acids Res. 12(1984), 8711-8721) with a pIC19R/H polylinker sequence (Marsh et al . , Gene 32 (1984): 481-485).

TABLE 1 Strains used

The NLS : : ere : : virF-fusion gene was also coupled to the helper plasmid in LBA2561 by selecting for single crossovers between the homologous DNA regions flanking the virF gene and NLS : : cre : :virF (pSDM3189) , resulting in LBA2561 : : NLS : : ere : : vi rF . As a control for transformation of the plant

LBA1115 (=MOG101 {Hood et al . , Transgenic Res. 2 (1993) 208- 218) was used, which is comparable to LBA1100 but containing tra and oec genes. IncP-vector pSDM3088 {Vergunst & Hooykaas, Plant Mol. Biol. 38 (1998), 393-406), containing the ere gene under control of the mannopine synthase promoter between the left and right adjacent sequences of Agro acterium ("T-DNA border repeats"), was transferred to LBA1115. Plant lines

Plasmid pSDM3043 (see fig. and Vergunst & Hooykaas Plant Mol. Biol. 38 (1998), 393-406) was introduced into Arabidopsis thaliana C24 by means of the Agrobacterium- mediated transformation of root explants (Vergunst et al . , in Methods in Mol. Biol. 82 pp 227-244(1998)) and transformants were selected for resistance to phosphinothricin (30 mg/1) . Plants having the excision locus 3043 are sensitive to kanamycin. A transformed plant, with 3:1 segregation for the transgene (1 locus) , was selected based on the efficiency of the Cre-mediated excision of the bar gene (see fig.), flanked by lox sites present in direct repeat (see arrows in the drawing) . Cre-mediated excision could be accomplished by transforming root explants with the Cre-expression vector pSDM3088. Such and excision simultaneously resulted in repair of a kanamycin resistance-marker.

Plant transformation

The protocol for transformation of Arabidopsis-root explants mediated by Agrobacterium was described earlier

(Vergunst et al . , in Methods in Mol. Biol. 82 pp 227- 244(1998)). Roots of 10-day old seedlings being homozygote for the excision locus 3043 (T3 or T4) were used in transformation assays using the strains disclosed in Table 1 containing the constructs described above. After cocultivation for 2 days, the explants were transferred to a medium for inducing the growth of calli and shoots, which medium contained 50 mg/1 kanamycin and 100 mg/1 timentine. The number of kanamycin- resistant calli was determined 2 and 3 weeks after cocultivation. PCR analysis was performed on a number of shoots regenerated from calli resistant to kanamycin.

PCR-analysis

The method used for isolation of chromosomal DNA for PCR-analysis and the protocol for the PCR-reaction were described earlier (Vergunst et al, Plant Mol. Biol. 38 (1998), 393-406) . As primers (a and b in the drawing) for the analysis of kanamycin-resistant calli and plants use was made of a) (5'-GAACTCGCCGTAAAGACTGGCG-3') annealing in the 35S promoter region (pDE35S in the drawing) ; and b) (5'-GCGCTGACAGCCGGAACACG-3') annealing in the nptll coding sequence (see fig.). Primers used for plasmid construction are mentioned in detail.

Results

To show that Agrobacterium transfers fusion proteins into recipient plant cells, a recombination assay was used that allows the detection of the transport of the Cre recombinase. To this end, the plant cell nucleus contained a DNA-segment, which after specific deletion by Cre results in a selectable kanamycin resistance trait. This substrate for Cre recombination (pSDM3043, fig. 1) was transferred by Agrobacterium-mediated transformation into the genome of Arabidopsis by means of selection for resistance to phosphinothricin. The transgenic plants were sensitive to kanamycin. However, the Cre-mediated recombination between two in tandem oriented lox sites results in excision of the intervening DNA-segment, and in the activation of the nptTJ gene by translational fusion of the 35S promoter region including the ATG start codon and an N-terminal coding part (determined by the lox sequence) with the coding region of nptJI. This fusion leads to kanamycin resistance. Thus, only cells in which a Cre-mediated recombination event has taken place survive on the selective medium. A plant line was isolated, comprising a single locus of the excision substrate and whose excision efficiency (as measured by the number of kanamycin resistant calli due to Cre-activity) was comparable to that of the efficiency of Agrobacterium-mediated transformation. This was tested by cocultivation of the plant line with an Agrobacterium-strain carrying a binary vector with a T-DNA harbouring the cre gene driven by a strong promoter sequence (LBA1115-pSDM3088) . The selected excision line was used in further assays to prove direct protein transport, including the Cre-recombinase, from Agrobacterium to plant cells. To this end, Cre-recombinase was expressed under control of the vir-induction system in Agrobacterium, either alone or as an N-terminal or C-terminal fusion with VirF or VirE2 , respectively. Expression of the fusion proteins in Agrobacterium was tested by immunoblot-analysis (data not shown) . Recombination activity of the fusion proteins was assayed by measuring the ability to induce an excision event in plasmid pSDM3043, which plasmid was introduced into the relevant bacterial strains in order to test this effect (data not shown) .

After cocultivation of the above-mentioned plant cells with the Agrobacterium-strains LBA1010, LBA1100,

LBA1149, or LBA2561, or with a derivative of these strains harbouring a plasmid expressing Cre-recombinase (the cre control plasmid) , at most one single survivor was found on a selective medium. Thus the same number of kanamycin-resistant calli ("the background") was obtained during cocultivation with strains which do not express Cre as during cocultivation with Agrobacterium strains which do express Cre-recombinase. Thus, we conclude that bacterially expressed Cre-recombinase is not transferred to plant cells. The results were, however, totally different when experiments were done with Agrobacterium-eelIs expressing fusion proteins between Cre and VirE2 or VirF. Agrobacterium- strains were used (both wild type as well as mutants for the respective Vir-protein) which express either C-terminal or N- terminal fusions of Cre with VirE2 and VirF, respectively. It was found that the fusion with the N-terminal region of the Vir-protein (Cre::virE2; NLS :: cre ::virF) , but not with the C- terminal end (VirE2::cre; NLS ::virF: : cre) was highly efficient in producing kanamycin-resistant calli. No difference was observed whether the Cre: :virF-fusion was expressed from the helper plasmid (LBA2561 : : NLS : : cre : : virF) or from plasmid pRL662 (LBA2561, pRL662 NLS : : cre : : virF) . Cocultivation of Agrobacterium strains (both wild type and virF mutant) expressing a Cre: :virF-fusion with a deletion of the 42 N- terminal amino acids of virF (NLS: : ere: :virFΔl -126) resulted in kanamycin-resistant calli being obtained at a significantly higher frequency. Cocultivation of a strain expressing a fusion protein between the last 37 C-terminal amino acids of VirF and Cre (NLS: :cre: :virFΔl-4S8) resulted in comparable high numbers of kanamycin-resistant calli.

As controls, strains harbouring the fusion plasmids NLS: : cre: :virFΔl -126 and cre: :virE2 were cocultivated with wild type Arabidopsis C24 root explants, but as expected, these did not result in kanamycin-resistance due to the absence of the excision locus.

PCR-analysis proved that kanamycin-resistance was indeed caused by the predicted Cre-mediated excision event. PCR analysis on DNA isolated from the excision-line 3043 resulted in a fragment of 2.3 kb (see fig.), whereas in DNA- samples isolated from kanamycin-resistant calli which were obtained after cocultivation of strains expressing cre : : virE2 and NLS: : cre: : virF, a 0.7 kb fragment is obtained. To exclude contamination with a T-DNA vector expressing cre, a PCR- analysis was performed with primers annealing in the cre coding region. As expected, no fragment was detected.

PCR-analysis of the few calli obtained from the control experiments (including those from experiments in which cultivation took place with an Agro acterium-strain not harbouring a cre gene) showed that these were partly the result of an excision, possibly due to homologous recombination between the lox sites and partly due to continuing growth of sensitive plants on the selection medium ("escapes") . In the latter case, the 2.3 kb-fragment was detected, and not the 0.7 kb-fragment indicating excision. All these results thus show that Agrobacterium can introduce Cre- recombinase into plant cells, but only when expressed as a fusion protein attached to the N-terminus of VirE2 or VirF. The efficiency of obtaining kanamycin-resistant calli after cocultivation of plant line 3043 with said Agrobacterium strains expressing fusions of Cre with 161 or 37 C-terminal amino acids of VirF shows that there must be a transport domain in the last 37 C-terminal amino acids of VirF. In view of the in planta function of VirE2, virF and VirD2 shown before, the three invariant amino acids (RPR-motif) in the C- terminus of VirF, VirE2 and VirD2 suggest the importance of these amino acids for the transport of the Vir proteins to the plant, using the VirB/VirD4 system (see later) . Mutation of one of the R residues in this motif lead to the loss of the transport function, but when the P residue in a residue chosen from A, Q, V and S was changed, the transport activity remained intact .

In the present experiments transport of the fusion proteins, detected as kanamycin resistance after cocultivation with plant line 3043, occurs independent of cotransfer of T- DNA. This implies an important application of the protein transport system, namely the possibility of effecting a change in cells in the absence of T-DNA. In addition, however, transfer may also occur in the presence of T-DNA (that may or may not be oncogenic) , which broadens the applicability of the system. Possible application may be site-directed integration of T-DNA in the genome of the recipient cell by means of cotransport of a recombinase that is expressed as a fusion protein in the same or in cotransforming bacteria.

To determine which of the virulence functions are essential for protein transport, NLS: : cre : : virFΔl -126 and cre : : virE2 were transferred into a set of vir mutants, listed in Table 1.

Cocultivation of the excision line with the virA, virB, virG and virD4 mutants, which harbour the fusion plas- mids, did not result in Cre-mediated excision, whereas virC, virDl/D2, virF and virE2 mutants resulted in calli with efficiencies comparable to the efficiencies of the wild type. Thus we can conclude that in this model system wherein VirA and VirG are responsible for forming VirB and VirD4, expression of vir genes through the regulators VirA and VirG is necessary for transport . But apart from that only the virB genes and virD4 , determining the type IV secretion channel (B-complex) and the coupling factor (VirD4) , are essential. The occurrence of virB and virD4 homologs in other systems (for example O'Callaghan et al . , Mol. Microbiol . 33 (1999): 1210-1220) makes it possible to construct similar protein transport systems in Agrobacterium and other microorganisms based on these homologous systems. The other virulence genes including those coding for the transported proteins VirD2, VirE2 and VirF, ap- H l->

(Jl o Ul

Sequence no . 1

Amino acid sequence of virF_15955 (1-609) Universal code

Total number of amino acids: 202, MW=22394 Max ORF: 1-606, 202 AA, M =22394

ORIGIN

1 MRNSSLRDAS GSNDAQVPHK

21 TELLNLPDHV LTEVAKRLAT 41 NNPVESAENI ANFSKSHRFT

61 RDAVRTEPLE KFSSRLKILS

81 RNAKLLSHAV RHAATLPDGE

101 QLSEAQLSQM RSEVATRPVL

121 GVAYTHQDGQ PEERLSGNHL 141 DHKINNIPNL VFNVAEPIMF

161 NEISALEVMA EVRPIARSIK

181 TAHDDARAEL MSADRPRSTR

201 GL*

Sequence no . 2

Amino acid sequence of VirD2_pTil5955 (1-1275)

Universal code

Total number of amino acids: 424, M =47476 Max ORF: 1-1272, 424 AA, M =47476

ORIGIN

1 MPDRAQVI IR IVPGGGTKTL

21 QQIINQLEYL SRKGKLELQR 41 SARHLDIPVP PDQIRELAQS

61 WVTEAGIYDE SQSDDDRQQD

81 LTTHI IVSFP AGTDQTAAYE

101 ASRE AAEMF GSGYGGGRYN

121 YLTAYHVDRD HPHLHVWNR 141 RELLGHG LK ISRRHPQLNY

161 DGLRKKMAEI SLRHGIVLDA

181 TSRAERGIAE RPITYAEHRR

201 LERMQAQKIQ FEDTDFDETS 221 PEEDRRDLSQ SFDPFRSDPS

241 TGEPDRATRH DKQPLEQHAR

261 FQESAGSSIK ADARIRVSLE

281 SERSAQPSAS KIPVIGHFGI 301 ETSYVAEASV RKRSGIFGTS

321 RPVTDVAMHT VKRQQRSKRR

341 NDEEAGPSGA NRKGLKAAQV

361 DSEANVGEQD TRDDSNKAAD

381 PVSASIGTEQ PEASPKRPRD 401 RHDGELGGRK RARGNRRDDG

421 RGGT*

Sequence no. 3

Amino acid sequence of virE2_pTil5955 (1-1602)

Universal code

Total number of amino acids: 533, M =60502

Max ORF: 1-1599, 533 AA, MW=60502

ORIGIN

1 MDLSGNEKSR PWKKANVSSS

21 TISDIQMTNG ENLESGSPTR

41 TEVLSPRLDD GSVDSSSSLY 61 SGSEHGNQAE IQKELSALFS

81 NMSLPGNDRR PDEYILVRQT

101 GQDAFTGIAK GNLDHMPTKA

121 EFNACCRLYR DGAGNYYPPP

141 LAFDKISVPA QLEET GMME 161 AKERNKLRFQ YKLDV NHAH

181 ADMGITGTEI FYQTDKNIKL

201 DRNYKLRPED RYVQTERYGR

221 REIQKRYQHE LQAGSLLPDI

241 MIKTPKNDIH FVYRFAGDNY 261 A KQFSEFEH TVKRRYGGET

281 EIKLKSKSGI MHDSKYLES

301 ERGSADIRFA EFVGENRAHN

321 RQFPTATVNM GQQPDGQGGL 341 TRDRHVSVEF LMQSAPNSP

361 AQALKKGELW DRVQLLARDG

381 NRYLSPHRLE YSDPEHFTEL

401 MNRVGLPASM GRQSHAASIK 421 FEKFDAQAAV IVINGPELRD

441 IHDLSPENLQ VSTKDVIVA

461 DRNENGQRTG TYTSVAEYER

481 LQLRLPADAA GVLGEAADKY

501 SRDFVRPEPA SRPISDSRRI 521 YESRPRSQSV NSF*

Sequence no . 4

Amino acid sequence of virE3-pTil5955 (1-2019) Universal code Total number of amino acids: 672, M =75584

Max ORF: 1-2016, 672 AA, MW=75584

ORIGIN

1 MVSTTKKSFA KSLTADMRRS 21 AQRWEQMRK ALITEEEALK

41 RQTRLESPDR KRKYAADMAI

61 VDKLDVGFRG EIGYKILGNK

81 RLRVDNPKEL TREHGRLRKT

101 KTVLKRNPVT QEVYLGLHER 121 KS LSVSSHL YAADGTLRMK

141 HVKYKDGRFE ERWERDENGD

161 LIRTRYANRG RLFQPVSEKM

181 GAPYRSGPDN RLYRDLTRQN

201 GFRRETFERD DQGNLERIGS 221 NHVGFSKISV KAANRQTSQT

241 KIQKLGGAFN KSFRSLLDKE

261 GNELGRDILS HRRLYNKRSA

281 VYDEATGQLK SAKHTFGKIY

301 RSETDYLSAG LKKVSKKILG 321 VTVYRKFAAL SERESEAERL

341 RSFESGAHRQ I QERAATPG

361 SPPSESDDIH FAQQSHLAKA

381 NSDHVEADVM RVTDQHIDVA 401 GQTSSSRQRN LEERLDSQSR

421 YKPANMLLSD PDLRADGPRP

441 YEGLAELTLR RDNESDGHKE

461 NDQRLRHFLQ PEPLVLPHPG 481 SPEITKVFGS RGEPLHPSGT

501 LHTAVGETAC EAPVMSSSLD

521 NHQPAPGQQE LLSLLHNAPA

541 PVSVAIHDEQ ERLAGEAPGG

561 SFRGSSGRTS SMSESIFDED 581 VQSHLVRDYS INPTNGFIDP

601 QSLFGGPDLS RGPKSGPEIP

621 SEDYHLSASE QENLLNQLLS

641 VPLPIPSPKP KSARSMIFEG

661 SRPRERSTSR GF*

Sequence no . 5

Derived amino acid sequence of RSF1010_mobA(l-2130)

Universal code

Total number of amino acids: 709, M =77847 1 MAIYHLTAKT GSRSGGQSAR

21 AKADYIQREG KYARDMDEVL

41 HAESGHMPEF VERPADY DA

61 ADLYERANGR LFKEVEFALP

81 VELTLDQQKA LASEFAQHLT 101 GAERLPYTLA IHAGGGENPH

121 CHLMISERIN DGIERPAAQ

141 FKRYNGKTPE KGGAQKTEAL

161 KPKA LEQTR EA ADHANRA

181 LERAGHDARI DHRTLEAQGI 201 ERLPGVHLGP NWEMEGRGI

221 RTDRADVALN IDTANAQIID

241 LQEYREAIDH ERNRQSEEIQ

261 RHQRVSGADR TAGPEHGDTG

281 RRSPAGHEPD PAGQRGAGGG 301 VAESPAPDRG GMGGAGQRVA

321 GGSRRGEQRR AERPERVAGV

341 ALEAMANRDA GFHDAYGGAA

361 DRIVALARPD ATDΝRGRLDL 381 AALGGPMKND RTLQAIGRQL

401 KAMGCERFDI GVRDATTGQM

421 MNRE SAAEV LQNTPWLKRM

441 NAQGNDVYIR PAEQERHGLV 461 LVDDLSEFDL DDMKAEGREP

481 ALWETSPKN YQAWVKVADA

501 AGGELRGQIA RTLASEYDAD

521 PASADSRHYG RLAGFTNRKD

541 KHTTRAGYQP VLLRESKGK 561 TATAGPALVQ QAGQQIEQAQ

581 RQQEKARRLA SLELPERQLS

601 RHRRTALDEY RSEMAGLVKR

621 FGDDLSKCDF IAAQKLASRG

641 RSAEEIGKAM AEASPALAER 661 KPGHEADYIE RTVSKVMGLP

681 SVQLARAELA RAPAPRQRGM

701 DRGGPDFSM*

Claims (9)

1. Method of effecting a change in a cell, wherein a transfer system is contacted with the cell to be changed, said transfer system comprising a membrane with in it a protein transport system comprising a pore which comprises a VirB complex and VirD4 protein, wherein the transfer system comprises a fusion protein or is capable of making a fusion protein which is introduced into the cell using the protein transport system, characterized in that a fusion protein BA is introduced into the cell to be changed which fusion protein BA i) comprises as a first part A an oligopeptide comprising 3 adjacent amino acids identical to or corresponding to the amino acids of the C-terminal amino acids 1-40 of VirF, VirD2, VirE2, VirE3, or MobA and ii) comprises as a second part B a polypeptide capable of exerting a cell-changing activity in the cell to be changed, wherein the C-terminal end of the polypeptide is linked to the N-terminal end of the first part A, under the condition that if the fusion protein comprises a first part A derived from VirE2, the fusion protein does not comprise the last 84 N-terminal amino acids of VirE2.

2. Method according to claim 1, characterized in that the oligopeptide of the part A comprises RXR, in particular RPR, RAR, RQR, RSR or RVR, or a derivative thereof.

3. Method according claim 1 or 2, characterized in that the fusion protein is formed by expression in the transfer system.

4. Method according one of the preceding claims, characterized in that the fusion protein is introduced into the cell to be changed without introducing a DNA- or RNA- sequence .

5. Method according one of the preceding claims , characterized in that the introduced fusion protein possesses a recombinase activity.

6. Method according one of the preceding claims, characterized in that a bacterium from the class of the Rhizobiaceae is used as the transfer system.

7. Method according one of the preceding claims, characterized in that the cell to be modified is chosen from the group consisting of i) a plant cell; ii) a yeast cell; and iii) a fungal cell.

8. Vector, characterized in that it codes for a protein transport system comprising a pore which contains a VirB complex and VirD4 protein, as well as for a fusion protein BA that comprises i) as a first part A an oligopeptide comprising 3 adjacent amino acids identical to or corresponding to the amino acids of the of the C-terminal amino acids 1-40 of VirF, VirD2, VirE2, VirE3 or MobA, and ii) as a second part B a polypeptide capable of effecting a cell-modifying activity in the cell to be modified, wherein the polypeptide with its C-terminal end is linked to the N- terminal end of the first part A, under the condition that if the fusion protein comprises a first part A derived from VirE2, the fusion protein does not comprise the last 84 N- terminal amino acids of VirE2.

9. Vector set, characterized in that the vector set comprises one or more vectors coding for a protein transport system comprising a pore which contains a VirB complex and VirD4 protein as well as a further vector coding for a fusion peptide BA which comprises i) as a first part A an oligopeptide comprising 3 adjacent amino acids identical to or corre- sponding to the amino acids of the C-terminal amino acids 1- 40 of VirF, VirD2, VirE2, VirE3 or MobA, and ii) as a second part B a polypeptide capable of exercising a cell-modifying activity in the cell to be modified, wherein the polypeptide of the C-terminal end of the polypeptide is linked to the N- terminal end of the first part A, under the condition that if the fusion protein comprises a first part A derived from VirE2, the fusion protein does not comprise the last 84 N- terminal amino acids of VirE2.