- Email: [email protected]
Complete nucleotide sequence of a plant tumor-inducing Ti plasmid
Gene 242 (2000) 331–336 www.elsevier.com/locate/gene
Complete nucleotide sequence of a plant tumor-inducing Ti plasmid k Katsunori Suzuki a, Yoshiyuki Hattori a, Misugi Uraji a, Nobuyuki Ohta a, Kumi Iwata a, Kenji Murata a, Akira Kato b, Kazuo Yoshida a, * a Department of Biological Science, Faculty of Science, Hiroshima University, Higashi-Hiroshima 739-8526, Japan b Hokkaido Agricultural Experiment Station, Sapporo 061-8555, Japan Received 14 October 1999; accepted 8 November 1999 Received by H. Uchimiya
Abstract Crown gall tumor disease in dicot plants is caused by Agrobacterium tumefaciens harboring a giant tumor-inducing ( Ti) plasmid. Here, for the first time among agrobacterial plasmids, the nucleotide sequence of a typical nopaline-type Ti plasmid (pTi-SAKURA) was determined completely. In total, 195 open reading frames (ORFs) were estimated in the 206 479 bp long sequence. 20 genes for conjugation, three for replication, 22 for pathogenesis and 37 for genetic colonization of host plants were found within two-thirds of the plasmid. These genes formed seven functional gene clusters with narrow inter-cluster spaces. In the remaining one-third of the plasmid, novel genes including homologs of mutT, Rhizobium nodQ and Sphingomonas ligE genes were found, which are likely to be responsible for the broad host range. Restriction fragment length variation indicates extreme plasticity of the part required for conjugational gene transfer and the above-mentioned one-third of the plasmid, even among closely related Ti plasmids. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Agrobacteria; Agrobacterium tumefaciens; Crown gall tumor disease; Horizontal gene transfer; Pathogenesis; Plant–microbial interaction
1. Introduction Agrobacterium tumefaciens and Agrobacterium rhizogenes are plant pathogens with the broadest known host range (De Cleene and De Lay, 1976). By infection at wound sites of dicot plants, the former induces crown gall tumors and the latter induces hairy root disease. Genes directly involved in the pathogenicity of these bacteria are mainly encoded on giant plasmids called the tumor-inducing plasmid ( Ti or pTi) and the hairyroot inducing plasmid (Ri or pRi). Abbreviations: ATP, adenosine triphosphate; GTP, guanosine triphosphate; _h, (extension) homolog; Int, DNA integrase; Inv, DNA invertase; LB, left border of T-DNA; Mw, molecular weight; oriT, origin of transfer; oriV, origin of vegetative replication; PCR, polymerase chain reaction; pRi, Ri plasmid; pTi, Ti plasmid; Ri, root inducing; RB, right border of T-DNA; T-DNA, transferred DNA portion of pTi and pRi; Ti, tumor inducing; tiorf, ORF found in pTi-SAKURA; Tps, transposase; tRNA, transfer RNA; VAR, large variable region. k The complete sequence of pTi-SAKURA reported in this paper has been deposited in the DDBJ/GenBank/EMBL databases under the accession No. AB016260. * Corresponding author. Tel.: +81-824-24-7455; fax: +81-824-24-0734. E-mail address: [email protected] ( K. Yoshida)
A transfer portion ( T-DNA) of the plasmids is moved from the bacteria into the plants (see, e.g., Hooykaas and Beijersbergen, 1994). In infected plant cells, gene expression of T-DNA integrated in plant chromosomes leads to overproduction of opines and phytohormones (see, e.g., Dessaux et al., 1992). The hormones cause neoplastic plant growth. The opines (unusual amino acids and sugar phosphates specific to each plasmid ) are utilized as nutrients by the bacteria. This phenomenon is called genetic colonization by bacteria in a crown gall (see, e.g., Kado, 1993). T-DNA transfer requires cis terminal repeats [ left (LB)- and right (RB)-border sequences] of T-DNA, virulence genes (vir) located outside of the T-DNA region, and additional chromosomal genes. Plant phenolic substances specifically induce expression of the virulence (vir) genes. This DNA transfer system has been widely used as the most reliable method of constructing recombinant plants (see, e.g., Binns and Thomashow, 1988). Therefore, the Ti and Ri plasmids are important not only for plant pathogenesis, but also for plant engineering. T-DNA transfer is the first confirmed example of gene transfer between prokaryotes and eukaryotes. Recently, it was shown that another inter-kingdom
0378-1119/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S0 3 7 8 -1 1 1 9 ( 9 9 ) 0 0 50 2 - 8
332
K. Suzuki et al. / Gene 242 (2000) 331–336
phenomenon, trans-kingdom conjugation, occurs between Escherichia coli and yeasts (Heinemann and Sprague, 1989; Nishikawa et al., 1990, 1992). Unlike T-DNA transfer, trans-kingdom conjugation generally involves the transfer of an entire plasmid into recipient yeast cells by means of bacterial conjugation machinery [origin of DNA transfer (oriT ) and conjugal transfer genes (tra)] (see, e.g., Heinemann, 1991). A. tumefaciens can also transfer DNA into yeasts by both T-DNA transfer (Bundock et al., 1995; Pies et al., 1996) and trans-kingdom conjugation (Sawasaki et al., 1996). In this respect, Ti and Ri plasmids are considered to be the most highly evolved conjugative plasmids. Unfortunately, in spite of the great biological and agricultural importance of these plasmids, there have been no investigations on their entire nucleotide sequence in any agrobacterial plasmid. In this paper, we report the first complete nucleotide sequencing of a nopaline-type Ti plasmid, and summarize the results of gene analysis.
2. Materials and methods 2.1. DNA library and sequencing The nopaline type Ti plasmid pTi-SAKURA was extracted from A. tumefaciens MAFF301001 obtained from the Japanese National Institute for Agrobiological Resources ( Tsukuba) and was used throughout this study. A linking DNA library of pTi-SAKURA and a physical map were constructed (Suzuki et al., 1998; see Fig. 1A), and each DNA segment between the neighboring ends of inserted Ti DNA fragments in overlapping recombinant phages was amplified by polymerase chain reaction (PCR). The nucleotide sequence of the PCR products was determined by the primer walking method. Recombinant l phage DNAs or PCR products obtained directly using the phage DNAs as PCR templates were used as template DNAs for sequencing. During the sequencing, the standard subcloning process was replaced by PCR subcloning (Hattori et al., 1997), which minimizes errors during subcloning. The resultant sequences were assembled using the software Seqed (Applied Biosystems) and Sequencher (Gene Codes). In total, we determined the sequence of both strands of the entire Ti plasmid with a final accuracy greater than 99.99%. 2.2. Genetic analysis Conventional computer sequence analyses were performed using Dnasis (Hitachi Software), Genetyx (Software Development) and GCG (Genetics Computer Groups). Sequence alignment for phylogenetic analysis was performed using ClustalW ( Thompson et al., 1994).
2.3. Assignment of genes and ORFs A homology search was conducted comparing the DNA sequence data with sequences in the EMBL/ GenBank/DDBJ DNA databases, and deduced amino acid sequences were compared with sequences in the non-redundant protein database using FASTA (Pearlson and Lipman, 1988). Coding sequences were predicted by GeneMark analysis (Borodovsky and McIninch, 1993). Sequences were also taken to be candidates of coding sequences (ORFs) if they were similar to the ones in databases in terms of both length and homology of the deduced amino acid sequence to those in the nonredundant protein database. There were long regions which were found to have no highly homologous counterparts in the protein database and did not give an appropriate signal in GeneMark analysis. In the case of such regions, DNA sequences were taken to be ORFs if they were longer than 80 codons and had promoterand ribosome-binding site-like sequence(s) present upstream of either an ATG or GTG codon. In the T-DNA region, sequences with a TATA box upstream of an ATG and with a polyA signal downstream of a termination codon were taken to be candidates of ORFs, even if they were not positive in the GeneMark analysis.
3. Results and discussion 3.1. Source of tumor-inducing plasmid DNA The nopaline-type Ti plasmid pTi-SAKURA (Suzuki et al., 1998; Hattori et al., 1997) was isolated from Agrobacterium tumefaciens MAFF301001. This strain is a broad-host-range pathogen isolated from a crown gall on a Japanese cherry tree Sakura (see Sawada et al., 1992). Nucleotide sequencing of pTi-SAKURA was carried out using the ordered recombinant phage library ( Fig. 1A) as described in Materials and methods. 3.2. General features of the plasmid pTi-SAKURA was a circular double-stranded DNA molecule of 206 479 bp. This is almost as large as the size of 132 MDa (approx. 200 kbp) of another typical nopaline-type plasmid pTiC58 (Holsters et al., 1980) and the sizes of other common Ti plasmids. We set the plasmid sequence starting coordinate, 1 bp, as the 5∞-protruding end base (A) of the actual EcoRI cleavage site located in the hyuB homolog gene (tiorf195) (see Figs. 1B, C and 2). Fig. 1C shows a physical map of pTi-SAKURA based on the complete nucleotide sequence data. When compared with the physical map of pTiC58, 60% of pTi-SAKURA was very similar to pTiC58. However, the remaining 40%, the region between the coordinates 25 kbp and 110 kbp, clearly
K. Suzuki et al. / Gene 242 (2000) 331–336
333
Fig. 1. Phage library, distribution of G+C content, and physical map of pTi-SAKURA. (A) Ordered horizontal lines indicate DNA segments which were cloned in l phage vector DASHII to construct the linking DNA library. (B) The wavy curve indicates G+C value (%) plotted against the nucleotide position in pTi-SAKURA with a window size of 400 bp. Horizontal thick lines represent major gene clusters. acc, genes for import and catabolism of agrocinopine; noc and nox, genes for import and catabolism of nopaline; trb and tra, genes for conjugation; vir, genes for virulence; rep, genes responsible for vegetative DNA replication; T-DNA, transfer DNA. VAR indicates the region in which high restriction fragment length variation between pTi-SAKURA and pTiC58, is evident. A thin line with arrowheads indicates the large inter-cluster space between rep and tra clusters. Vertical arrowheads indicate ORFs for DNA integrase (Int)/ invertase (Inv)/insertion sequence (IS)/transposase ( Tps). (C ) Restriction endonuclease cleavage sites are indicated by vertical lines, with the names of the enzyme at the end of each row. The top scale of nucleotide position in (A) is applicable to (B) and (C ).
showed fragment length polymorphism. We named this region VAR ( large variable) (see Fig. 1B). VAR was composed mainly of a large inter-gene-cluster space between rep and tra regions. The average G+C content of pTi-SAKURA was 56.0%, which is slightly lower than that of the total DNA of A. tumefaciens (57–63%) ( Kersters and De Ley, 1984). As shown in Fig. 1B, where the G+C content of each segment on pTi-SAKURA was plotted against base position, several regions showed a consistently higher G+C content, while other regions had a consistently lower G+C content. Functionally related genes were generally clustered in a DNA region with an even G+C content. 3.3. General features of coding regions Genes and open reading frames (ORFs) were estimated using the complete nucleotide sequence. 195 genes and ORFs in total were found in pTi-SAKURA, as summarized in Table 1. In this paper, all genes and ORFs from tiorf1 to tiorf195 are numbered clockwise (see all ORFs and their characteristics in the accession No. AB016260 datum in DDBJ/EMBL/GenBank databases). Sometimes, tiorf is replaced by the names of homologs. Despite the rigorous search, we were unable to find tRNA or rRNA genes or any pseudo tRNA or rRNA genes in pTi-SAKURA.
As shown in Table 1, typical pTi genes [ T-DNA, virulence genes (vir), conjugation genes (tra and trb) and vegetative replication genes (rep)] (see, e.g., Binns and Thomashow, 1988) and catabolism genes [agrocinopine utilization genes (acc) and nopaline catabolism genes (nox and noc)] were clearly identified on pTiSAKURA. As far as DNA sequence data of pTiC58 genes are available, these genes, except several tra genes, were highly homologous more than 97% between pTiSAKURA and pTiC58. On the circular gene map ( Fig. 2), each functional gene cluster is indicated by a color-coded symbol. These genes represent 44% of the total ORFs, and regions containing these gene clusters occupy 60% of the entire plasmid. Functionally related genes had similar G+C content and were clustered in particular regions. Conjugation genes were split into two clusters, the tra and traI/trb regions (see Figs. 1A and 2). traI/trb, rep, VAR, tra, acc, vir, T-DNA, nox and noc were ordered clockwise in pTi-SAKURA ( Fig. 2). The order of the gene clusters is the same as that in pTiC58 according to genetic and/or sequence data available in pTiC58. Many ORFs were homologous with those in pNGR234a ( Freiberg et al., 1997). As discussed previously (Suzuki et al., 1998), replication genes and conjugation genes are highly homologous with those of the nodule-inducing plasmid pNGR234a. In addition, there are several homologous ORFs in the large inter-cluster space between rep and tra regions.
334
K. Suzuki et al. / Gene 242 (2000) 331–336
Fig. 2. Circular gene map of pTi-SAKURA. Locations of genes including predicted ORFs, regulatory sequences (tra_BOX and vir_BOX ), origins of DNA replication (oriV, oriT, RB) and replication termination signal (LB) are indicated. ORFs outside of the circle are coded clockwise, while ORFs inside are coded counter-clockwise. Open boxes indicate orphan ORFs. Functionally related genes are shown in the same color as follows: green for T-DNA; pink for virulence, including interaction with plants; red for catabolism of opines; blue for conjugation; light pink for DNA modification and recombination; yellow for vegetative DNA replication; black for miscellaneousness. Grey boxes indicate ORFs of unknown function for which homologous proteins were found in databases. The picture at the center indicates a crown gall on a Kalanchoe plant caused by pTi-SAKURA.
Table 1 Genes and ORFs classified according to function Function
(No. of genes or ORFs)
Vegetative DNA replication Plasmid modification
(3): repA, repB, repC (7): tiorf25(stb_h), tiorf47(methylase_h), invertase/recombinase/transposase (5): tiorf22(inv), tiorf26(inv), tiorf36(int), tiorf128(tps), tiorf131(tps_h) (20): traCDG, traAFB, traM,traR, traItrbBCDEJKLFGHI (38): vir (22): virA, virB1B2B3B4B5B6B7B8B9B10B11, virC1C2, virD1D2D3D4, virE1E2, virG, virH T-DNA genes (16): tms1, tms2, tmr, nos, acs (2), e, 5, 6a, 6b, 21_h, rolB_h, 5_h, IS ORF (2), tiorf174 (13): nocR, nocPTQM, hyuA_h-hyuB_h, noxBtiorf192noxAarc-odh-ocd (11): accR, accABCDEFG,tiorf169(accG), tiorf170(accF ), tiorf117(scp) (7): tzs, pinF1F2, virF, virK, tiorf37(mutT_h), tiorf84(ligE ), tiorf53(nodQ_h) (12): ribose transporter tiorf96-98(rtpABC ), chemotaxis tiorf114(mcpA), respiration tiorf92(qor), tiorf58(ribA-h), tiorf66(ankyrin-h), transporter tiorf55, tiorf112, transcriptional regulator tiorf1, tiorf52, tiorf91, secretion/ infection tiorf56 (84) 195
Conjugation T-DNA transfer Nopaline metabolism Agrocinopine metabolism Plant interaction Miscellaneous gene/ORFs
Unclassified ORF Total
K. Suzuki et al. / Gene 242 (2000) 331–336
335
3.4. General features of inter-gene cluster spaces In general, inter-gene cluster spaces between neighboring gene clusters were very short (see Figs. 1 and 2). An exceptionally large inter-gene cluster space between the rep and tra regions contained 84 ORFs in total. tiorf102(hupT ) was the only ORF homologous to previously identified pTi genes in this region. Most of the ORFs in this region were genes of unknown function. To the contrary, outside of this region, 78% of the 111 ORFs were homologous to previously identified pTi and/or pRi genes. As shown in Fig. 1, there was a highly variable region VAR. A large part of VAR consisted of the large spacer. The spacer region showed an uneven distribution of G+C content (Fig. 1A). This might be due in part to transposition. In this region, we found four ORFs (tiorf22, tiorf26, tiorf36 and tiorf47) which are probably responsible for DNA modification. tiorf22(inv) and tiorf26(inv) were putative DNA invertase genes, while tiorf36(int) was an integrase gene. tiorf47(methylase_h) was the largest ORF in pTi-SAKURA. The putative Tiorf47 protein was composed of 1693 amino acid residues and shows 30% identity with adenine-specific methylases over 200 amino acid residues, though Tiorf47 is four to eight times the length of common DNA methylases. 3.5. Novel Genes in VAR potentially responsible for interaction with plants In the large spacer between tra and rep, we discovered a homolog of nodQ which is a Rhizobium gene responsible for host-range determination in nodulation of legumes. The putative product of tiorf53 (nodQ_h) was entirely homologous with the NodQ protein. A motif search analysis indicated that Tiorf53 contains an ATP/GTP-binding motif A (GLSGSGKS ), like the NodQ protein (Schwedock and Long, 1990). tiorf53 is likely to be a member of a putative operon which consists of tiorf53–59 (see Fig. 3). The G+C content of each ORF in this region was as high as 50%. Among the ORFs, two (tiorf55 and tiorf56) can code proteins which are partially homologous with transporter proteins. Like Tiorf53, two other proteins ( Tiorf56 and Tiorf57) were also found to have an ATP/GTP-binding motif. Tiorf56 was partially homologous with Xanthomonas campestris XpsE protein which is responsible for infection to host plants by mediating the secretion of high molecular weight compounds (Dums et al., 1991). Tiorf56 was also homologous with the EpsE protein of Vibrio cholerae which functions to secrete a toxin (Sandkvist et al., 1993). These results suggest that the above putative operon is involved in infection and/or association with host plants, by mediating the secretion of high molecular weight compounds. In Rhizobium nodule-inducing plasmids, nod_box sequences are present for the regulation of nodulation genes (Rostas
Fig. 3. Characteristics of a putative operon containing a nodQ homolog. A thin line ( left) indicates the location in plasmid pTi-SAKURA. Arrows indicate the size and location of the ORF and the coding direction (downward, clockwise; upward, counter-clockwise; see Fig. 1). Filled arrows, ORFs in the putative operon containing the nodQ homolog; shaded arrows, ORFs which do not belong to the operon; MW, molecular weight of the protein encoded by each ORF; Motif/Homology, motif found or proteins found to be homologous to the ORF product in each protein.
et al., 1986). In pTi-SAKURA, however, there was no nod_box sequence, indicating that the nodulation gene homolog in pTi-SAKURA is under its own regulation, which is different from that in the case of Rhizobium. ligE(tiorf84) resembled Sphingomonas paucimobilis ligE, which codes for b-etherase (Masai et al., 1991). This enzyme produces lower molecular weight phenolic compounds from lignin by hydrolysis of ether linkages common in plant lignin. Phenolic compounds are generally toxic to bacterial cells. However, the Agrobacterium cells can utilize simple phenolics as nutrients and, furthermore, phenolics such as acetosyringone induce the expression of vir genes for infection onto plants. ligE(tiorf84) might be involved in production of lower molecular weight phenolics as inducers for virulence genes and as nutrient carbon sources when Agrobacterium cells reach host plants, which contain large amount of lignin. tiorf37(mutT_h) was a homolog of mutT which serves to maintain the high fidelity of DNA synthesis by converting the mutagenic nucleotide 8-oxo-deoxyguanine triphosphate to the non-mutagenic monophosphate in E.
336
K. Suzuki et al. / Gene 242 (2000) 331–336
coli (Maki and Sekiguchi, 1992). The mutT homolog is likely to protect Agrobacterium from such toxic compounds produced by plants as secondary metabolites. These genetic systems are likely to be helpful for the bacterial cell association with plant cells. These notion is supported by the fact that tartrate utilization genes are present in pTiAB3 (Salomone et al., 1996) which have allowed A. vitis to adopt to growth on a tartraterich plant, the grapevine.
Acknowledgements We are grateful to Dr. Atsuhiro Oka ( Kyoto University), Dr. Takeshi Ohama (JT Biohistory Research Hall, Osaka), Ms. Devika de Costa (Hiroshima University) and Dr. Arthur Katoh (Mercy Hospital, Pittsburgh) for critical comments on the manuscript, and to Dr. Takahiko Tsudzuki (Aichi Gakuin University, Nagoya) for his help in constructing the circular gene map. This study was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Culture, Japan to K.Y.
References Binns, A.N., Thomashow, M.F., 1988. Cell biology of Agrobacterium infection and transformation of plants. Annu. Rev. Microbiol. 42, 575–606. Borodovsky, M., McIninch, J., 1993. GeneMark: parallel gene recognition for both DNA strands. Comp. Chem. 17, 123–133. Bundock, P., den Dulk-Ras, A., Beijersbergen, A., Hooykaas, P.J.J., 1995. Trans-kingdom T-DNA transfer from Agrobacterium tumefaciens to Saccharomyces cerevisiae. EMBO J. 14, 3206–3214. De Cleene, M., De Lay, J., 1976. The host range of crown-gall. Bot. Rev. 42, 389–466. Dessaux, Y., Petit, A., Tempe, J., 1992. Opines in Agrobacterium biology. In: Verma, D.P.S. (Ed.), Molecular Signals in Plant–Microbe Interactions. CRC Press, Boca Raton, FL, pp. 109–136. Dums, F., Dow, J.M., Daniels, M.J., 1991. Structural characterization of protein secretion genes of the bacterial phytopathogen Xanthomonas campestris pathovar campestris. Relatedness to secretion systems of other Gram-negative bacteria. Mol. Gen. Genet. 229, 357–364. Freiberg, C., Fellay, R., Bairoch, A., Broughton, W.J., Rosenthal, A., Perret, X., 1997. Molecular basis of symbiosis between Rhizobium and legumes. Nature 387, 394–401. Hattori, Y., Suzuki, K., Ohta, N., Uraji, M., Katoh, A., Yoshida, K., 1997. Genome structure of pTi-SAKURA (I ): Strategy for DNA sequencing of a Japanese cherry-Ti plasmid. Nucleic. Acids Symp. Ser. 37, 159–160. Heinemann, J.A., 1991. Genetics of gene transfer between species. Trends Genet. 7, 181–185. Heinemann, J.A., Sprague, G.F.J., 1989. Bacterial conjugative plas-
mids mobile DNA transfer between bacteria and yeast. Nature 340, 205–209. Holsters, M., Silva, B., Van Vliet, F., Genetello, C., De Block, M., Dhaese, P., Depicker, A., Inze, D., Engler, G., Villarroel, R., Van Montagu, M., Schell, J., 1980. The functional organization of the nopaline A. tumefaciens plasmid pTiC58. Plasmid 3, 212–230. Hooykaas, P.J.J., Beijersbergen, A.G.M., 1994. The virulence system of Agrobacterium tumefaciens. Annu. Rev. Phytopathol. 32, 157–179. Kado, C.I., 1993. Agrobacterium-mediated transfer and stable incorporation of foreign genes in plants. In: Clewell, D.B. ( Ed.), Bacterial Conjugation. Plenum, New York, pp. 243–254. Kersters, K., De Ley, J., 1984. Genus III Agrobacterium Conn 1942 359. In: Krieg, N.R., Holt, J.G. ( Eds.), Bergey’s Manual of Systematic Bacteriology vol. 1. William and Wilkins, Baltimore, pp. 244–254. Maki, H., Sekiguchi, M., 1992. MutT protein specifically hydrolyses a potent mutagenic substrate for DNA systhesis. Nature 355, 273–275. Masai, E., Katayama, Y., Kawai, S., Nishikawa, S., Yamasaki, M., Morohoshi, N., 1991. Cloning and sequencing of the gene for a Psuedomonas paucimobilis enzyme that cleaves b-aryl ether. J. Bacteriol. 173, 7950–7955. Nishikawa, M., Suzuki, K., Yoshida, K., 1990. Structural and functional stability of IncP plasmids during stepwise transmission by trans-kingdom mating: promiscuous conjugation of Escherichia coli and Saccharomyces cerevisiae. Jpn. J. Genet. 65, 323–334. Nishikawa, M., Suzuki, K., Yoshida, K., 1992. DNA integration into recipient yeast chromosomes by trans-kingdom conjugation between Escherichia coli and Saccharomyces cerevisiae. Curr. Genet. 21, 101–108. Pearlson, W.R., Lipman, D.J., 1988. Improved tools for biological sequence comparison. Proc. Natl. Acad. Sci. USA 85, 2444–2448. Pies, K.L., Heath, J.D., Liang, X., Stephens, K., Nester, E.W., 1996. Agrobacterium tumefaciens-mediated transformation of yeast. Proc. Natl. Acad. Sci. USA 93, 1613–1618. Rostas, K., Kondorosi, E., Horvath, B., Simoncsits, A., Kondorosi, A., 1986. Conservation of extended promoter regions of nodulation genes in Rhizobium. Proc. Natl. Acad. Sci. USA 83, 1757–1761. Salomone, J.-Y., Crouzet, P., De Ruffray, P., Otten, L., 1996. Characterization and distribution of tartrate utilization genes in the grapevine pathogen Agrobacterium vitis. Mol. Plant–Microbe Interact. 9, 401–408. Sandkvist, M., Morales, V., Bagdasarian, M., 1993. A protein required for secretion of cholera toxin through the outer membrane of Vibrio cholerae. Gene 123, 81–86. Sawada, H., Ieki, H., Kobayashi, S., Oiyama, I., 1992. Grouping of tumorigenic Agrobacterium spp. based on Ti plasmid-related phenotypes. Ann. Phytopathol. Soc. Jpn 58, 244–252 (in Japanese). Sawasaki, Y., Inomata, K., Yoshida, K., 1996. Trans-kingdom conjugation between Agrobacterium tumefaciens and Saccharomyces cerevisiae, a bacterium and a yeast. Plant Cell Physiol. 37, 103–106. Schwedock, J., Long, S.R., 1990. ATP sulphurylase activity of the nodP and nodQ gene products of Rhozobium meliloti. Nature 348, 644–647. Suzuki, K., Ohta, N., Hattori, Y., Uraji, M., Katoh, A., Yoshida, K., 1998. Novel structural difference between nopaline- and octopinetype trbJ gene: construction of genetic and physical map and sequencing of trb/traI and rep gene clusters of a new Ti plasmid pTi-SAKURA. Biochim. Biophys. Acta 1396, 1–7. Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673–4680.
Complete nucleotide sequence of a plant tumor-inducing Ti plasmid k Katsunori Suzuki a, Yoshiyuki Hattori a, Misugi Uraji a, Nobuyuki Ohta a, Kumi Iwata a, Kenji Murata a, Akira Kato b, Kazuo Yoshida a, * a Department of Biological Science, Faculty of Science, Hiroshima University, Higashi-Hiroshima 739-8526, Japan b Hokkaido Agricultural Experiment Station, Sapporo 061-8555, Japan Received 14 October 1999; accepted 8 November 1999 Received by H. Uchimiya
Abstract Crown gall tumor disease in dicot plants is caused by Agrobacterium tumefaciens harboring a giant tumor-inducing ( Ti) plasmid. Here, for the first time among agrobacterial plasmids, the nucleotide sequence of a typical nopaline-type Ti plasmid (pTi-SAKURA) was determined completely. In total, 195 open reading frames (ORFs) were estimated in the 206 479 bp long sequence. 20 genes for conjugation, three for replication, 22 for pathogenesis and 37 for genetic colonization of host plants were found within two-thirds of the plasmid. These genes formed seven functional gene clusters with narrow inter-cluster spaces. In the remaining one-third of the plasmid, novel genes including homologs of mutT, Rhizobium nodQ and Sphingomonas ligE genes were found, which are likely to be responsible for the broad host range. Restriction fragment length variation indicates extreme plasticity of the part required for conjugational gene transfer and the above-mentioned one-third of the plasmid, even among closely related Ti plasmids. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Agrobacteria; Agrobacterium tumefaciens; Crown gall tumor disease; Horizontal gene transfer; Pathogenesis; Plant–microbial interaction
1. Introduction Agrobacterium tumefaciens and Agrobacterium rhizogenes are plant pathogens with the broadest known host range (De Cleene and De Lay, 1976). By infection at wound sites of dicot plants, the former induces crown gall tumors and the latter induces hairy root disease. Genes directly involved in the pathogenicity of these bacteria are mainly encoded on giant plasmids called the tumor-inducing plasmid ( Ti or pTi) and the hairyroot inducing plasmid (Ri or pRi). Abbreviations: ATP, adenosine triphosphate; GTP, guanosine triphosphate; _h, (extension) homolog; Int, DNA integrase; Inv, DNA invertase; LB, left border of T-DNA; Mw, molecular weight; oriT, origin of transfer; oriV, origin of vegetative replication; PCR, polymerase chain reaction; pRi, Ri plasmid; pTi, Ti plasmid; Ri, root inducing; RB, right border of T-DNA; T-DNA, transferred DNA portion of pTi and pRi; Ti, tumor inducing; tiorf, ORF found in pTi-SAKURA; Tps, transposase; tRNA, transfer RNA; VAR, large variable region. k The complete sequence of pTi-SAKURA reported in this paper has been deposited in the DDBJ/GenBank/EMBL databases under the accession No. AB016260. * Corresponding author. Tel.: +81-824-24-7455; fax: +81-824-24-0734. E-mail address: [email protected] ( K. Yoshida)
A transfer portion ( T-DNA) of the plasmids is moved from the bacteria into the plants (see, e.g., Hooykaas and Beijersbergen, 1994). In infected plant cells, gene expression of T-DNA integrated in plant chromosomes leads to overproduction of opines and phytohormones (see, e.g., Dessaux et al., 1992). The hormones cause neoplastic plant growth. The opines (unusual amino acids and sugar phosphates specific to each plasmid ) are utilized as nutrients by the bacteria. This phenomenon is called genetic colonization by bacteria in a crown gall (see, e.g., Kado, 1993). T-DNA transfer requires cis terminal repeats [ left (LB)- and right (RB)-border sequences] of T-DNA, virulence genes (vir) located outside of the T-DNA region, and additional chromosomal genes. Plant phenolic substances specifically induce expression of the virulence (vir) genes. This DNA transfer system has been widely used as the most reliable method of constructing recombinant plants (see, e.g., Binns and Thomashow, 1988). Therefore, the Ti and Ri plasmids are important not only for plant pathogenesis, but also for plant engineering. T-DNA transfer is the first confirmed example of gene transfer between prokaryotes and eukaryotes. Recently, it was shown that another inter-kingdom
0378-1119/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S0 3 7 8 -1 1 1 9 ( 9 9 ) 0 0 50 2 - 8
332
K. Suzuki et al. / Gene 242 (2000) 331–336
phenomenon, trans-kingdom conjugation, occurs between Escherichia coli and yeasts (Heinemann and Sprague, 1989; Nishikawa et al., 1990, 1992). Unlike T-DNA transfer, trans-kingdom conjugation generally involves the transfer of an entire plasmid into recipient yeast cells by means of bacterial conjugation machinery [origin of DNA transfer (oriT ) and conjugal transfer genes (tra)] (see, e.g., Heinemann, 1991). A. tumefaciens can also transfer DNA into yeasts by both T-DNA transfer (Bundock et al., 1995; Pies et al., 1996) and trans-kingdom conjugation (Sawasaki et al., 1996). In this respect, Ti and Ri plasmids are considered to be the most highly evolved conjugative plasmids. Unfortunately, in spite of the great biological and agricultural importance of these plasmids, there have been no investigations on their entire nucleotide sequence in any agrobacterial plasmid. In this paper, we report the first complete nucleotide sequencing of a nopaline-type Ti plasmid, and summarize the results of gene analysis.
2. Materials and methods 2.1. DNA library and sequencing The nopaline type Ti plasmid pTi-SAKURA was extracted from A. tumefaciens MAFF301001 obtained from the Japanese National Institute for Agrobiological Resources ( Tsukuba) and was used throughout this study. A linking DNA library of pTi-SAKURA and a physical map were constructed (Suzuki et al., 1998; see Fig. 1A), and each DNA segment between the neighboring ends of inserted Ti DNA fragments in overlapping recombinant phages was amplified by polymerase chain reaction (PCR). The nucleotide sequence of the PCR products was determined by the primer walking method. Recombinant l phage DNAs or PCR products obtained directly using the phage DNAs as PCR templates were used as template DNAs for sequencing. During the sequencing, the standard subcloning process was replaced by PCR subcloning (Hattori et al., 1997), which minimizes errors during subcloning. The resultant sequences were assembled using the software Seqed (Applied Biosystems) and Sequencher (Gene Codes). In total, we determined the sequence of both strands of the entire Ti plasmid with a final accuracy greater than 99.99%. 2.2. Genetic analysis Conventional computer sequence analyses were performed using Dnasis (Hitachi Software), Genetyx (Software Development) and GCG (Genetics Computer Groups). Sequence alignment for phylogenetic analysis was performed using ClustalW ( Thompson et al., 1994).
2.3. Assignment of genes and ORFs A homology search was conducted comparing the DNA sequence data with sequences in the EMBL/ GenBank/DDBJ DNA databases, and deduced amino acid sequences were compared with sequences in the non-redundant protein database using FASTA (Pearlson and Lipman, 1988). Coding sequences were predicted by GeneMark analysis (Borodovsky and McIninch, 1993). Sequences were also taken to be candidates of coding sequences (ORFs) if they were similar to the ones in databases in terms of both length and homology of the deduced amino acid sequence to those in the nonredundant protein database. There were long regions which were found to have no highly homologous counterparts in the protein database and did not give an appropriate signal in GeneMark analysis. In the case of such regions, DNA sequences were taken to be ORFs if they were longer than 80 codons and had promoterand ribosome-binding site-like sequence(s) present upstream of either an ATG or GTG codon. In the T-DNA region, sequences with a TATA box upstream of an ATG and with a polyA signal downstream of a termination codon were taken to be candidates of ORFs, even if they were not positive in the GeneMark analysis.
3. Results and discussion 3.1. Source of tumor-inducing plasmid DNA The nopaline-type Ti plasmid pTi-SAKURA (Suzuki et al., 1998; Hattori et al., 1997) was isolated from Agrobacterium tumefaciens MAFF301001. This strain is a broad-host-range pathogen isolated from a crown gall on a Japanese cherry tree Sakura (see Sawada et al., 1992). Nucleotide sequencing of pTi-SAKURA was carried out using the ordered recombinant phage library ( Fig. 1A) as described in Materials and methods. 3.2. General features of the plasmid pTi-SAKURA was a circular double-stranded DNA molecule of 206 479 bp. This is almost as large as the size of 132 MDa (approx. 200 kbp) of another typical nopaline-type plasmid pTiC58 (Holsters et al., 1980) and the sizes of other common Ti plasmids. We set the plasmid sequence starting coordinate, 1 bp, as the 5∞-protruding end base (A) of the actual EcoRI cleavage site located in the hyuB homolog gene (tiorf195) (see Figs. 1B, C and 2). Fig. 1C shows a physical map of pTi-SAKURA based on the complete nucleotide sequence data. When compared with the physical map of pTiC58, 60% of pTi-SAKURA was very similar to pTiC58. However, the remaining 40%, the region between the coordinates 25 kbp and 110 kbp, clearly
K. Suzuki et al. / Gene 242 (2000) 331–336
333
Fig. 1. Phage library, distribution of G+C content, and physical map of pTi-SAKURA. (A) Ordered horizontal lines indicate DNA segments which were cloned in l phage vector DASHII to construct the linking DNA library. (B) The wavy curve indicates G+C value (%) plotted against the nucleotide position in pTi-SAKURA with a window size of 400 bp. Horizontal thick lines represent major gene clusters. acc, genes for import and catabolism of agrocinopine; noc and nox, genes for import and catabolism of nopaline; trb and tra, genes for conjugation; vir, genes for virulence; rep, genes responsible for vegetative DNA replication; T-DNA, transfer DNA. VAR indicates the region in which high restriction fragment length variation between pTi-SAKURA and pTiC58, is evident. A thin line with arrowheads indicates the large inter-cluster space between rep and tra clusters. Vertical arrowheads indicate ORFs for DNA integrase (Int)/ invertase (Inv)/insertion sequence (IS)/transposase ( Tps). (C ) Restriction endonuclease cleavage sites are indicated by vertical lines, with the names of the enzyme at the end of each row. The top scale of nucleotide position in (A) is applicable to (B) and (C ).
showed fragment length polymorphism. We named this region VAR ( large variable) (see Fig. 1B). VAR was composed mainly of a large inter-gene-cluster space between rep and tra regions. The average G+C content of pTi-SAKURA was 56.0%, which is slightly lower than that of the total DNA of A. tumefaciens (57–63%) ( Kersters and De Ley, 1984). As shown in Fig. 1B, where the G+C content of each segment on pTi-SAKURA was plotted against base position, several regions showed a consistently higher G+C content, while other regions had a consistently lower G+C content. Functionally related genes were generally clustered in a DNA region with an even G+C content. 3.3. General features of coding regions Genes and open reading frames (ORFs) were estimated using the complete nucleotide sequence. 195 genes and ORFs in total were found in pTi-SAKURA, as summarized in Table 1. In this paper, all genes and ORFs from tiorf1 to tiorf195 are numbered clockwise (see all ORFs and their characteristics in the accession No. AB016260 datum in DDBJ/EMBL/GenBank databases). Sometimes, tiorf is replaced by the names of homologs. Despite the rigorous search, we were unable to find tRNA or rRNA genes or any pseudo tRNA or rRNA genes in pTi-SAKURA.
As shown in Table 1, typical pTi genes [ T-DNA, virulence genes (vir), conjugation genes (tra and trb) and vegetative replication genes (rep)] (see, e.g., Binns and Thomashow, 1988) and catabolism genes [agrocinopine utilization genes (acc) and nopaline catabolism genes (nox and noc)] were clearly identified on pTiSAKURA. As far as DNA sequence data of pTiC58 genes are available, these genes, except several tra genes, were highly homologous more than 97% between pTiSAKURA and pTiC58. On the circular gene map ( Fig. 2), each functional gene cluster is indicated by a color-coded symbol. These genes represent 44% of the total ORFs, and regions containing these gene clusters occupy 60% of the entire plasmid. Functionally related genes had similar G+C content and were clustered in particular regions. Conjugation genes were split into two clusters, the tra and traI/trb regions (see Figs. 1A and 2). traI/trb, rep, VAR, tra, acc, vir, T-DNA, nox and noc were ordered clockwise in pTi-SAKURA ( Fig. 2). The order of the gene clusters is the same as that in pTiC58 according to genetic and/or sequence data available in pTiC58. Many ORFs were homologous with those in pNGR234a ( Freiberg et al., 1997). As discussed previously (Suzuki et al., 1998), replication genes and conjugation genes are highly homologous with those of the nodule-inducing plasmid pNGR234a. In addition, there are several homologous ORFs in the large inter-cluster space between rep and tra regions.
334
K. Suzuki et al. / Gene 242 (2000) 331–336
Fig. 2. Circular gene map of pTi-SAKURA. Locations of genes including predicted ORFs, regulatory sequences (tra_BOX and vir_BOX ), origins of DNA replication (oriV, oriT, RB) and replication termination signal (LB) are indicated. ORFs outside of the circle are coded clockwise, while ORFs inside are coded counter-clockwise. Open boxes indicate orphan ORFs. Functionally related genes are shown in the same color as follows: green for T-DNA; pink for virulence, including interaction with plants; red for catabolism of opines; blue for conjugation; light pink for DNA modification and recombination; yellow for vegetative DNA replication; black for miscellaneousness. Grey boxes indicate ORFs of unknown function for which homologous proteins were found in databases. The picture at the center indicates a crown gall on a Kalanchoe plant caused by pTi-SAKURA.
Table 1 Genes and ORFs classified according to function Function
(No. of genes or ORFs)
Vegetative DNA replication Plasmid modification
(3): repA, repB, repC (7): tiorf25(stb_h), tiorf47(methylase_h), invertase/recombinase/transposase (5): tiorf22(inv), tiorf26(inv), tiorf36(int), tiorf128(tps), tiorf131(tps_h) (20): traCDG, traAFB, traM,traR, traItrbBCDEJKLFGHI (38): vir (22): virA, virB1B2B3B4B5B6B7B8B9B10B11, virC1C2, virD1D2D3D4, virE1E2, virG, virH T-DNA genes (16): tms1, tms2, tmr, nos, acs (2), e, 5, 6a, 6b, 21_h, rolB_h, 5_h, IS ORF (2), tiorf174 (13): nocR, nocPTQM, hyuA_h-hyuB_h, noxBtiorf192noxAarc-odh-ocd (11): accR, accABCDEFG,tiorf169(accG), tiorf170(accF ), tiorf117(scp) (7): tzs, pinF1F2, virF, virK, tiorf37(mutT_h), tiorf84(ligE ), tiorf53(nodQ_h) (12): ribose transporter tiorf96-98(rtpABC ), chemotaxis tiorf114(mcpA), respiration tiorf92(qor), tiorf58(ribA-h), tiorf66(ankyrin-h), transporter tiorf55, tiorf112, transcriptional regulator tiorf1, tiorf52, tiorf91, secretion/ infection tiorf56 (84) 195
Conjugation T-DNA transfer Nopaline metabolism Agrocinopine metabolism Plant interaction Miscellaneous gene/ORFs
Unclassified ORF Total
K. Suzuki et al. / Gene 242 (2000) 331–336
335
3.4. General features of inter-gene cluster spaces In general, inter-gene cluster spaces between neighboring gene clusters were very short (see Figs. 1 and 2). An exceptionally large inter-gene cluster space between the rep and tra regions contained 84 ORFs in total. tiorf102(hupT ) was the only ORF homologous to previously identified pTi genes in this region. Most of the ORFs in this region were genes of unknown function. To the contrary, outside of this region, 78% of the 111 ORFs were homologous to previously identified pTi and/or pRi genes. As shown in Fig. 1, there was a highly variable region VAR. A large part of VAR consisted of the large spacer. The spacer region showed an uneven distribution of G+C content (Fig. 1A). This might be due in part to transposition. In this region, we found four ORFs (tiorf22, tiorf26, tiorf36 and tiorf47) which are probably responsible for DNA modification. tiorf22(inv) and tiorf26(inv) were putative DNA invertase genes, while tiorf36(int) was an integrase gene. tiorf47(methylase_h) was the largest ORF in pTi-SAKURA. The putative Tiorf47 protein was composed of 1693 amino acid residues and shows 30% identity with adenine-specific methylases over 200 amino acid residues, though Tiorf47 is four to eight times the length of common DNA methylases. 3.5. Novel Genes in VAR potentially responsible for interaction with plants In the large spacer between tra and rep, we discovered a homolog of nodQ which is a Rhizobium gene responsible for host-range determination in nodulation of legumes. The putative product of tiorf53 (nodQ_h) was entirely homologous with the NodQ protein. A motif search analysis indicated that Tiorf53 contains an ATP/GTP-binding motif A (GLSGSGKS ), like the NodQ protein (Schwedock and Long, 1990). tiorf53 is likely to be a member of a putative operon which consists of tiorf53–59 (see Fig. 3). The G+C content of each ORF in this region was as high as 50%. Among the ORFs, two (tiorf55 and tiorf56) can code proteins which are partially homologous with transporter proteins. Like Tiorf53, two other proteins ( Tiorf56 and Tiorf57) were also found to have an ATP/GTP-binding motif. Tiorf56 was partially homologous with Xanthomonas campestris XpsE protein which is responsible for infection to host plants by mediating the secretion of high molecular weight compounds (Dums et al., 1991). Tiorf56 was also homologous with the EpsE protein of Vibrio cholerae which functions to secrete a toxin (Sandkvist et al., 1993). These results suggest that the above putative operon is involved in infection and/or association with host plants, by mediating the secretion of high molecular weight compounds. In Rhizobium nodule-inducing plasmids, nod_box sequences are present for the regulation of nodulation genes (Rostas
Fig. 3. Characteristics of a putative operon containing a nodQ homolog. A thin line ( left) indicates the location in plasmid pTi-SAKURA. Arrows indicate the size and location of the ORF and the coding direction (downward, clockwise; upward, counter-clockwise; see Fig. 1). Filled arrows, ORFs in the putative operon containing the nodQ homolog; shaded arrows, ORFs which do not belong to the operon; MW, molecular weight of the protein encoded by each ORF; Motif/Homology, motif found or proteins found to be homologous to the ORF product in each protein.
et al., 1986). In pTi-SAKURA, however, there was no nod_box sequence, indicating that the nodulation gene homolog in pTi-SAKURA is under its own regulation, which is different from that in the case of Rhizobium. ligE(tiorf84) resembled Sphingomonas paucimobilis ligE, which codes for b-etherase (Masai et al., 1991). This enzyme produces lower molecular weight phenolic compounds from lignin by hydrolysis of ether linkages common in plant lignin. Phenolic compounds are generally toxic to bacterial cells. However, the Agrobacterium cells can utilize simple phenolics as nutrients and, furthermore, phenolics such as acetosyringone induce the expression of vir genes for infection onto plants. ligE(tiorf84) might be involved in production of lower molecular weight phenolics as inducers for virulence genes and as nutrient carbon sources when Agrobacterium cells reach host plants, which contain large amount of lignin. tiorf37(mutT_h) was a homolog of mutT which serves to maintain the high fidelity of DNA synthesis by converting the mutagenic nucleotide 8-oxo-deoxyguanine triphosphate to the non-mutagenic monophosphate in E.
336
K. Suzuki et al. / Gene 242 (2000) 331–336
coli (Maki and Sekiguchi, 1992). The mutT homolog is likely to protect Agrobacterium from such toxic compounds produced by plants as secondary metabolites. These genetic systems are likely to be helpful for the bacterial cell association with plant cells. These notion is supported by the fact that tartrate utilization genes are present in pTiAB3 (Salomone et al., 1996) which have allowed A. vitis to adopt to growth on a tartraterich plant, the grapevine.
Acknowledgements We are grateful to Dr. Atsuhiro Oka ( Kyoto University), Dr. Takeshi Ohama (JT Biohistory Research Hall, Osaka), Ms. Devika de Costa (Hiroshima University) and Dr. Arthur Katoh (Mercy Hospital, Pittsburgh) for critical comments on the manuscript, and to Dr. Takahiko Tsudzuki (Aichi Gakuin University, Nagoya) for his help in constructing the circular gene map. This study was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Culture, Japan to K.Y.
References Binns, A.N., Thomashow, M.F., 1988. Cell biology of Agrobacterium infection and transformation of plants. Annu. Rev. Microbiol. 42, 575–606. Borodovsky, M., McIninch, J., 1993. GeneMark: parallel gene recognition for both DNA strands. Comp. Chem. 17, 123–133. Bundock, P., den Dulk-Ras, A., Beijersbergen, A., Hooykaas, P.J.J., 1995. Trans-kingdom T-DNA transfer from Agrobacterium tumefaciens to Saccharomyces cerevisiae. EMBO J. 14, 3206–3214. De Cleene, M., De Lay, J., 1976. The host range of crown-gall. Bot. Rev. 42, 389–466. Dessaux, Y., Petit, A., Tempe, J., 1992. Opines in Agrobacterium biology. In: Verma, D.P.S. (Ed.), Molecular Signals in Plant–Microbe Interactions. CRC Press, Boca Raton, FL, pp. 109–136. Dums, F., Dow, J.M., Daniels, M.J., 1991. Structural characterization of protein secretion genes of the bacterial phytopathogen Xanthomonas campestris pathovar campestris. Relatedness to secretion systems of other Gram-negative bacteria. Mol. Gen. Genet. 229, 357–364. Freiberg, C., Fellay, R., Bairoch, A., Broughton, W.J., Rosenthal, A., Perret, X., 1997. Molecular basis of symbiosis between Rhizobium and legumes. Nature 387, 394–401. Hattori, Y., Suzuki, K., Ohta, N., Uraji, M., Katoh, A., Yoshida, K., 1997. Genome structure of pTi-SAKURA (I ): Strategy for DNA sequencing of a Japanese cherry-Ti plasmid. Nucleic. Acids Symp. Ser. 37, 159–160. Heinemann, J.A., 1991. Genetics of gene transfer between species. Trends Genet. 7, 181–185. Heinemann, J.A., Sprague, G.F.J., 1989. Bacterial conjugative plas-
mids mobile DNA transfer between bacteria and yeast. Nature 340, 205–209. Holsters, M., Silva, B., Van Vliet, F., Genetello, C., De Block, M., Dhaese, P., Depicker, A., Inze, D., Engler, G., Villarroel, R., Van Montagu, M., Schell, J., 1980. The functional organization of the nopaline A. tumefaciens plasmid pTiC58. Plasmid 3, 212–230. Hooykaas, P.J.J., Beijersbergen, A.G.M., 1994. The virulence system of Agrobacterium tumefaciens. Annu. Rev. Phytopathol. 32, 157–179. Kado, C.I., 1993. Agrobacterium-mediated transfer and stable incorporation of foreign genes in plants. In: Clewell, D.B. ( Ed.), Bacterial Conjugation. Plenum, New York, pp. 243–254. Kersters, K., De Ley, J., 1984. Genus III Agrobacterium Conn 1942 359. In: Krieg, N.R., Holt, J.G. ( Eds.), Bergey’s Manual of Systematic Bacteriology vol. 1. William and Wilkins, Baltimore, pp. 244–254. Maki, H., Sekiguchi, M., 1992. MutT protein specifically hydrolyses a potent mutagenic substrate for DNA systhesis. Nature 355, 273–275. Masai, E., Katayama, Y., Kawai, S., Nishikawa, S., Yamasaki, M., Morohoshi, N., 1991. Cloning and sequencing of the gene for a Psuedomonas paucimobilis enzyme that cleaves b-aryl ether. J. Bacteriol. 173, 7950–7955. Nishikawa, M., Suzuki, K., Yoshida, K., 1990. Structural and functional stability of IncP plasmids during stepwise transmission by trans-kingdom mating: promiscuous conjugation of Escherichia coli and Saccharomyces cerevisiae. Jpn. J. Genet. 65, 323–334. Nishikawa, M., Suzuki, K., Yoshida, K., 1992. DNA integration into recipient yeast chromosomes by trans-kingdom conjugation between Escherichia coli and Saccharomyces cerevisiae. Curr. Genet. 21, 101–108. Pearlson, W.R., Lipman, D.J., 1988. Improved tools for biological sequence comparison. Proc. Natl. Acad. Sci. USA 85, 2444–2448. Pies, K.L., Heath, J.D., Liang, X., Stephens, K., Nester, E.W., 1996. Agrobacterium tumefaciens-mediated transformation of yeast. Proc. Natl. Acad. Sci. USA 93, 1613–1618. Rostas, K., Kondorosi, E., Horvath, B., Simoncsits, A., Kondorosi, A., 1986. Conservation of extended promoter regions of nodulation genes in Rhizobium. Proc. Natl. Acad. Sci. USA 83, 1757–1761. Salomone, J.-Y., Crouzet, P., De Ruffray, P., Otten, L., 1996. Characterization and distribution of tartrate utilization genes in the grapevine pathogen Agrobacterium vitis. Mol. Plant–Microbe Interact. 9, 401–408. Sandkvist, M., Morales, V., Bagdasarian, M., 1993. A protein required for secretion of cholera toxin through the outer membrane of Vibrio cholerae. Gene 123, 81–86. Sawada, H., Ieki, H., Kobayashi, S., Oiyama, I., 1992. Grouping of tumorigenic Agrobacterium spp. based on Ti plasmid-related phenotypes. Ann. Phytopathol. Soc. Jpn 58, 244–252 (in Japanese). Sawasaki, Y., Inomata, K., Yoshida, K., 1996. Trans-kingdom conjugation between Agrobacterium tumefaciens and Saccharomyces cerevisiae, a bacterium and a yeast. Plant Cell Physiol. 37, 103–106. Schwedock, J., Long, S.R., 1990. ATP sulphurylase activity of the nodP and nodQ gene products of Rhozobium meliloti. Nature 348, 644–647. Suzuki, K., Ohta, N., Hattori, Y., Uraji, M., Katoh, A., Yoshida, K., 1998. Novel structural difference between nopaline- and octopinetype trbJ gene: construction of genetic and physical map and sequencing of trb/traI and rep gene clusters of a new Ti plasmid pTi-SAKURA. Biochim. Biophys. Acta 1396, 1–7. Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673–4680.