Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Molecular Microbioiogy (1993) 10(3), 597-605 Restoration of attachment, virulence and noduiation of Agrobacterium tumefaciens chvB mutants by rhicadhesin Saskia Swart,* Gerrit Smit, Ben J.J. Lugtenberg and Jan W. Kijne institute of Moiecular Plant Sciences, Leiden University, Nonnensteeg 3. 2311 VJ Leiden, The Netheriands. Summary In contrast to wild-type Agrobacterium tumefaciens strains, p-1,2-glucan-deficjent chvB mutants were found to be unable to attach to pea root hair tips. The mutants appeared to produce rhicadhesin, the protein that mediates the first step in attachment of Rhizobiaceae cells to plant root hairs, but the protein was inactive. Both attachment to root hairs and virulence of the ChvB mutants could be restored by treatment of the plants with active rhicadhesin, whereas treatment of plants with |3-1,2-glucan had no effect on attachment or virulence. Moreover, nodulation ability of a chvB mutant carrying a Sym plasmid could be restored by pretreatment of the host plant with rhicadhesjn. Apparently the attachment-minus and avirulence phenotype of chvB mutants is caused by tack of active rhicadhesin, rather than directly being caused by a deficiency in p-1,2-glucan synthesis. The results strongly suggest that rhicadhesin is essential for attachment and virulence of A. tumefaciens cells. They also indicate that the mechanisms of binding of Agrobacterium and Rhizobium bacteria to plant target cells are similar, despite differences between these target cells. Introduction Attachment of Rhizobium and Agrobacterium bacteria, both members of the Rhizobiaceae family, to host plant cells is an early step in the establishment of the plantbacterium associations leading to the formation of nitrogen-fixing root nodules and plant tumours, respectively. Attachment of Agrobacterium and Rhizobium cells appears to be a two-step process in which binding to the plant target cell surface is followed by accumulation of the bacteria on the plant cell (Matthysse, 1983; Kljne e( a/., Received 7 June, 1993; revised and accepted 26 July, 1993. *For correspondence. Tel. (71) 275059; Fax (71) 275088. 1988; Smit et ai, 1989b; 1992), The majority of genes that are responsible for tumour induction by Agrobacterium are located on the Ti (tumour-inducing) plasmid (Van Larebeke etai., 1974; Zaenen etai., 1974), whereas most nodulation genes of Rhizobium are localized on the Sym (symbiosis) plasmid (Johnston etai., 1978; Nuti et ai., 1979; Hirsch etai., 1980), Introduction of a Ti piasmid into Rhizobium can yield tumorigenic Rhizobium strains, whereas Agrobacterium harbouring a Sym plasmid can nodulate certain leguminous plants (Hooykaas et ai., 1977; 1981; 1985). In contrast, genes involved in the attachment of Agrobacterium to cells of host plants are localized on the bacterial chromosome (Douglas et ai., 1982; Smit etai., 1987). Similarly, attachment of Rhizobium to the root hairs of host plants is chromosomally encoded (Smitefa/., 1986; Kijne ef a/., 1988). Cell-to-cell attachment is considered to be a prerequisite for plant tumour formation. Plant molecules involved in the attachment process are poorly characterized. Recently a vitronectin-like protein of plant origin was proposed to play a role in attachment of Agrobacterium (Wagner and Matthysse, 1992). Several bacterial chromosomal loci are reported to be involved in the ability of Agrobacterium tumefaciens to attach to plant cells (Douglas etai., 1982; Cangelosi etai., 1987; Matthysse. 1987; Thomashow etai., 1987). The corresponding mutants are all avirulent. Two of these loci. chvB and exoC. are required for the production of p-1,2-glucan (Puvanesarajah etai., 1985; Cangelosi etai., 1987). cftvSmutants do not produce cyclic p-1.2-glucan, since the chvB gene encodes a 235 kDa cytoplasmic membrane protein, that is required for p-1,2-glucan synthesis (Zorreguieta et al., 1988). A third mutant, chvA, is deficient in the excretion of cyclic p-1,2-glucan from the cell (Cangelosi et ai., 1989; Inon de lannino and Ugalde, 1989; O'Connel and Handelsman. 1989), However, the mutant still exports some p-1,2-glucan (O'Connel and Handelsman, 1989) and has been reported to have a leaky phenotype (Van Veen et ai., 1987), Since p-1.2-glucan mutants studied so far were found to be attachment deficient, ^-1,2-glucan was hypothesized to be necessary for attachment. However, it should be noted that these mutants show a pleiotropic phenotype (Douglas et ai.. 1985; Puvanesarajah et ai, 1985; Cangelosi etai., 1987; Thomashow etai, 1987). Results from our laboratory showed that binding of Rhizobiaceae to the surface of plant root hairs is mediated by 598 S. Swart, G. Smit, B. J. J. Lugtenberg and J. W. Kijne Table 1. Attachment of Agrobacterium and Rhizobium cells to pea root haJrtips^. I attachment in class'' Stfain characteristics t 2 3 A. tumetaciens LBA1010 1251 A1011 A1020 A1045 ME117 Wild type pTi-. pCr-cured ChvB ChvB chvB ChvB 12 9 73 75 90 63 33 25 24 24 9 15 55 66 3 1 1 2 7 23 70 R. leguminosarum bv. viciae 24B Wild type a. Bacteria were incubated with the roots as described previously (Smit et ai., 1986). b. Class 1, no attached bacteria; class 2, tew bacteria, directly attached to the root hair; class 3, many attached bacteria, forming two or more layers of bacteria or a cap-like aggregate at the tip ot the root hair (cap formation). rhicadhesin, a calcium-binding protein located on the bacterial cell surface (Smit et al., 1989a,b). Since chvB mutants are unable to attach to plant cells, and since it was unknown whether these bacteria produce rhicadhesin, we studied the attachment characteristics of these mutants to pea root hair tips in relation to production of rhicadhesin. Study of chvB mutants was preferred over chvA mutants because of the leaky phenotype of chvA mutants. The results show that the attachment-minus and avirulence phenotypes of chvB mutants result from a lack of active rhicadhesin and not from a lack of p-1.2-glucan. Furthermore, evidence is presented that in the case of nodulation by Sym plasmid-carrying Agrobacterium cells attachment is a prerequisite for infection, being dependent on the presence of active rhicadhesin. plasmid-cured strain, and tested for the presence of rhicadhesin, that is, for its ability to inhibit attachment of A. tumefaciens and R. leguminosarum cells to pea root hairs. As shown for A. tumefaciens in Table 2, cell surface components isolated from chvB mutants did not possess rhicadhesin activity, whereas cell surface preparations isolated from A. tumefaciens wild-type and plasmid-cured cells did. The same results were obtained with inhibition of attachment of R leguminosarum 248. This result indicates that no active rhicadhesin is present at the cell surface of chvS mutants. For one c/ii/S mutant strain, Al O i l , the culture medium as well as a sonicated cell extract were tested for adhesin activity, since it could not be excluded that these mutants still produce rhicadhesin, but are somehow affected in transiocation or anchoring of the adhesin. No adhesin activity was found in either fraction (Table 2} indicating that strain A1011 is affected in the production of active rhicadhesin. In order to check whether the chvB mutant produces rhicadhesin in an inactive form, ME117 bacteria were sheared and the rhicadhesin purification method was followed. By SDS-PAGE, a protein band of about 14kDa was observed for rhicadhesin purified from strain ME117 (Fig. 1, lane A) at the same position as expected for wildtype rhicadhesin (Fig. 1, lane B) (Smit etaL, 1991). Since several concentrations of rhicadhesin isolated from chvB mutants were tested for activity and no inhibition of attachment of wild-type cells was observed (data not shown), it was concluded that ME117 produces rhicadhesin in an inactive form. Table 2. Influence of cell surface preparations, cell extract and culture supernatant from various A. tumetaciens stfains on attachment of A. fumeCac/ens wild-type strain LBA1010 to pea root hairtips^. 7o attachment in class" Results Attachment ability of A. tumefaciens chvB mutants The chvB mutants grown in TYC and harvested at various growth stages were found to be unable to attach to pea root hair tips, in contrast to the wild-type strain LBA1010, the Ti plasmid-cured strain LBA1251, and Rhizobium leguminosarum 248 (Table 1). This result is consistent with earlier reports on attachment to tobacco and Zinnia cells (Douglas et al., 1982: 1985). The chvB mutants appeared to be affected in the first step of attachment, since the majority of the root hairs were devoid of adhering bacteria. chvB mutants do not synthesize active rhicadhesin Cell surface preparations were isolated from A. tumefaciens chvB mutants, the wild-type strain and the Ti Addition to standard assay 2 3 9 31 60 Cell surface preparation from LBA1010 1251 (pTi-. pCr-cured) A1011 (ChvB) M020 {chvB\ A1045 (c/1 vB) ME117(c/iv6) 56 43 12 19 10 5 25 34 28 28 26 32 19 23 60 53 64 63 Culture supernatant from A1011{chvBi 11 32 57 14 28 58 None 1 Cell extract from AlOti {ChvBi a. Additions corresponding to 4 ml bacterial culture were incubated with the pea roots. A. tumetaciens LBA1010 cells were harvested from TY medium, suspended in phosphate bufter and incubated with the roots for 2h. b. See Table 1, footnote b. Role of rhicadhesin in virulence and nodulation 599 reported by O'Connel and Handelsman (1989) concerning attachment of chvB mutants to Zinnia cells. Restoration of virulence of A. tumefaciens chvB mutants by wild-type rhicadhesin A B Fig. 1. SDS-PAGE of purified rhicadhesin isolated from +1.8 x lO'^ cells of A. tumefaciens chvB mutant strain ME117 (A) and of A. tumefaciens wild-type strain LBA1010 (B). Restoration of attachment ability of chvB mutants by wildtype rhicadhesin R. leguminosarum 248 and A. tumefaciens LBA1010 cells, grown under Ca^*-limiting conditions, are unable to attach to pea root hair tips (Smit etai., 1987; 1991). However, after pre-incubation of pea roots with purified rhicadhesin such bacteria appeared to attach to pea root hairs. This physiological restoration of attachment ability is most likely to be caused by binding of the bacteria to rootbound rhicadhesin (Smit ef al., 1991). The results presented in Table 3 show that attachment ability of chvB mutant ME117 could be restored after pretreatment of roots with rhicadhesin, isolated from wild-type A. tumefaciens, followed by incubation with mutant cells grown in TY medium, that is, under Ca^*-limiting conditions. Also rhicadhesin isolated from R. leguminosarum bv. viciae was able to restore attachment (data not shown). However, wild-type levels of attachment, such as that found with restoration of attachment of wild-type Rhizobium and Agrobacterium grown under Ca^*-limiting conditions, were not reached. This indicates that binding of rhicadhesin to the cell surface of the chvB mutants is indeed possible, but is less efficient than to wild-type strains of Agrobacterium and Rhizobium. Restoration of attachment ability was also found when chioramphenicol (50)igmr') was present duhng the attachment assay (data not shown). No restoration of attachment ability was found after protease treatment of rhicadhesin (Tabie 3). Cells of strain ME117 grown in TYC medium did not adhere to the rhicadhesin-pretreated roots (Table 3). Pretreatment of roots with p-1,2-glucan did not result in restoration of attachment ability of ME117 cells (Table 3). The presence of |3-1,2-glucan during the incubation with bacteria had no effect on the attachment of ME117 cells; however, this treatment did not inhibit attachment of wild-type strains (data not shown). This result is In agreement with results Virulence of chvB mutants on Kaianchoe daigremontiana and Kaianchoe tubiflora plants was tested after treatment of plants with purified rhicadhesin, isolated from A. tumefaciens or R. leguminosarum 248, The chvB mutants, A1011 and ME117, grown in TY or TYC, appeared to be avirulent in the absence of rhicadhesin (less than 1% of the wound sites developed a small tumour) (Fig.2B). Since attachment ability of chvB mutants could be optimally restored by rhicadhesin after growth of the cells under Ca^*-limiting conditions, bacteria grown under such conditions were used for testing the possible restoration of virulence. Treatment of K. daigremontiana leaf wound sites with rhicadhesin isolated from A. tumefaciens LBA1010, prior to inoculation with the chvB mutants, resulted in tumour formation on an average of 50% of the wound sites (Fig.2E). No tumours were observed when rhicadhesin was added in the absence of bacteria (data not shown). Moreover, no tumours were formed after treatment with rhicadhesin pretreated with protease (Fig. 2C). Addition of p-1,2-glucan to the wound sites prior to inoculation with ME117 did not result in tumour formation (Fig.2D), This result is in agreement with earlier reports (O'Connel and Handelsman. 1989). Addition of Table 3. Influence ot pretreatment ol pea roots with purified rhicadhesin and [i-1,2-glucan isolated from wild-type A. tumefaciens on attachment of cells of ,4. tumefaciens chvB n\u\an\ ME117topea root hair lips". %• attachment in class'' Bacterial growth medium" Root pretreatment TY TYC TY TYC TY None None Rhicadhesin Rhicadhesin Protease-treated rhicadhesin" p-1,2-glucan p-i ,2-glucan 0-1,2-glucan +rhicadhesin TY TYC TY 1 2 3 73 74 14 75 23 23 48 25 4 3 38 0 70 69 65 25 23 27 5 8 8 11 50 39 a. Roots were incubated with rhicadhesin (lOngml ') and/or [i-1,2-glucan (0.5tigmr'). The roots were washed in phosphate buffer betore incubation with the bacteria. Agrobacterium cells were harvested, suspended in phosphate bufter and incubated with the roots for 2 h. See the Experimental procedures section for details. b. TY medium contatns (per litre); tryptone (Difco), 5.0g and yeast extract (Oifco), 3 0 g . TYC medium is TY medium supplemented with 7mM CaClj. c. See Table 1, footnote b. d- Rhicadhesin (100 ng ml"^) was treated with proteinase K (1 mg m r ' ) for 60 mm at 37'-C prior to incubation with the roots. 600 S. Swan. G. Smit, B. J. J. Lugtenberg and J. W. Kijne B leguminosarum bv. phaseoli strain RCC3622 forms nodules on siratro (Fig.3A). Also A. tumefaciens strain LBA2728, grown in TY or TYC medium, forms nodules on siratro (Fig. 3B). However, c^vSstrain LBA2753, grown in either TY or TYC medium, does not form nodules on siratro (Fig. 3C). Pretreatment of the roots with rhicadhesin followed by inoculation with LBA2753 cells, grown in TY medium, resulted in the formation of several nodules per plant {Fig. 3D). Incubation of fresh plants, not pretreated with rhicadhesin, with plant medium of the nodulated siratro plants did not result in nodulation. This indicates that; {i) no nodulating wild-type bacteria were present in the medium of the nodulated plants, and {ii) LBA2753 did not permanently regain nodulation ability. Apparently, pretreatment of roots with rhicadhesin is required for nodulation by LBA2753. Similar results were obtained when bean was used as a test plant. Discussion cf)vB mutants of A. tumefaciens were found to be unable Fig. 2. Restoration of virulence of A. tumefaciens chvB mutant strain ME117on K. daipremonftana by treatmeni of the wound sites with rhicadhesin, isoiated from A tumefaciens wild type. Before mocuiaton with A tumefaciensWM-t^pe strain LBAIOIO (A) or chvSmutant strain ME117 (B-F). wound sites were incubated for 30 mm with either buffer (A, B), rtiicadhesin treated with proteinase K(C), lOiag |3-1,2-glucan (D), lOngrtiicadhesin(E)of botti(J-1,2-glucanandrhicadhesin(F). Seethe Experimental procedures sec\\or\ lor 6e\a\\s. both rhicadhesin and [i-1,2-glucan prior to inoculation gave on average a greater number of, and larger tumours than treatment with rhicadhesin oniy (Fig.2F). Similar results were obtained when K. tubiflora was used as a test plant (data not shown). The use of rhicadhesin purified from R. leguminosarum strain 248 ied to identical results {data not shown). Restoration ofnodulation of A. tumefaciens chvB mutants Van Veen etai (1987) showed that, unlike wiid-type A. tumefaciens (Hooykaas et al., 1985). chvB mutants harbouring a Sym plasmid of R. leguminosarum bv. phaseoli are not capable of nodulating bean. To investigate if this phenomenon was because of the lack of active rhicadhesin, nodulation of siratro (Macroptiiium atropurpureum) by LBA2753 ceils grown in TY was initially tested with and without pretreatment of the roots with purified rhicadhesin isolated from strain LBAIOIO. Siratro was chosen because of its ease as a test system. Wild-type R. Fig. 3. Nodulation of siratro after inoculation with (A) R. leguminosarum bv. phaseoii strain RCC3622, (B) A., tumefaciens sUam LBA2728. (C) A. tumefaciens chvB strain LBA2753, (D) A tumefaciens chvB strain LBA2753 after trealment of the rools with rhicadhesin, isolated from A. fumefeciens strain LBAIOIO. See the Experimental procedures section for details Role of rhicadhesin in virulence and nodulation to attach to pea root hairs. According to the model of Smit etal. {1987; 1989b), attachment appeared to be affected in the first step, that is, direct attachment of the bacteria to the surface of the root hair tip. For Rhizobium, this step is non-host-plant-specific and mediated by rhicadhesin. SDS-PAGE of a rhicadhesin preparation of the chvB mutant ME117 yielded a protein band of about 14kDa running at the same position as wild-type rhicadhesin. However, rhicadhesin preparations isolated from chvB mutants were found to be devoid of adhesin activity. Attachment ability of chvB mutants could be restored by addition of active wild-type rhicadhesin. Presumably chvB mutants produce an inactive form of rhicadhesin. It was concluded that lack of active rhicadhesin is directly responsible for the attachment deficiency of chvB mutants and that other components that are not produced by these mutants do not seem to be directly involved in attachment. Since we were able to isolate rhicadhesin from the cell surface, lack of translocation to the cell surface is not responsible for the attachment deficiency of cfivB mutants. The fact that more than one chvB mutant strain has been used and none displays rhicadhesin activity (Table 2) renders the possibility of a {point) mutation in the gene encoding rhicadhesin unlikely. The defect may be caused by improper processing of the rhicadhesin protein or to impaired Ca^* binding. The role of p-1,2-glucan in activation of rhicadhesin will be discussed later. Restoration of attachment ability of chvB mutants by rhicadhesin did not result in wild-type attachment levels, leaving the possibility for requirement of a second efficiency-increasing factor. It is not very likely that this factor is extracellular p-1,2-glucan since addition of p-1,2-glucan did not result in enhancement of attachment, neither in absence or presence of rhicadhesin. Taken together these results are consistent with the hypothesis that, like in the case of Rhizobium, the first step in the attachment of A. tumefaciens is mediated by rhicadhesin. Douglas et al. (1982) reported that chvB mutants are not affected in the synthesis of cellulose fibrils. Our results are consistent with this obsetvation, since accumulation of mutant cells on the plant cell surface (class 3 attachment), a process mediated by cellulose fibrils (Matthysse, 1983; Smit et al., 1987), is observed after treatment of the roots with active rhicadhesin. Binding of rhicadhesin to Rhizobium and Agrobacterium cells is mediated by Ca^"" ions. Only Rhizobiaceae cells grown under Ca^'^-limiting conditions or grown with calcium and treated with EDTA, were found to be able to bind to rhicadhesin-treated pea roots. Under such growth conditions rhicadhesin bound to bacteria is released into the medium, which probably enables binding of the cells to root-bound adhesin (Smit etai. 1989b; 1991). Accordingly, in order to achieve binding of rhicadhesin to the cell surface of the chvB mutants, the level of Ca^^ ions on the 601 cell surface of the mutants had to be tow. Furthermore, several phenotypic changes in the cell envelope of chvB mutants could be prevented by growth under calcium-limiting conditions {S. Swart et al., submitted), yielding a more wild-type-like surface for the mutant cells, which may be necessary for anchoring on root-bound rhicadhesin. Virulence of chvB mutants on K. daigremontiana and K. tubifiora plants was restored by treatment of the wound sites with rhicadhesin prior to inoculation with chvB mutants. On average, the number of tumours was lower in comparison with that induced by the wild-type strain, although the size of the tumours was similar. This result corresponds with the suboptimal restoration of attachment ability found after pretreatment of roots or bacteria with rhicadhesin. Treatment of wound sites with j3-1,2-glucan prior to inoculation did not result in tumour formation by ChvB mutants, strongly suggesting that restored tumour formation is preceded by restored attachment. Taken together, it can be concluded that lack of active rhicadhesin on the cell surface of the chvB mutants causes the attachment-minus and virulence-minus phenotype of the chvB mutants, rather than lack of p-1,2-glucan, as proposed by Puvanesarajah etal. (1985). However, it is possible that endogenous p-1,2-glucan is necessary for production of active rhicadhesin. The active rhicadhesin used was isolated from a [5-1,2-glucan-producing wildtype strain. Under hypo-osmotic conditions p-1,2-glucan may be required for proper membrane assembling, which may be necessary for correct anchoring of rhicadhesin. Beta-1,2-glucan may also be required for stability of rhicadhesin, such as cyclodextrins, that are used for protein stabilization {Szejtti, 1991). This could explain why a greater number, and larger tumours were found in the case where both rhicadhesin and p-1,2-glucan were added in the restoration of virulence. A. tumefaciens chvA mutants still export some of the wild-type amount of p-1,2glucan {O'Connel and Handelsman, 1989). This may explain why chvA mutants have a leaky phenotype with regard to attachment and rhicadhesin activity (S. Swart, unpublished results) and virulence and nodulation {Van Veen etal., 1987). Dylan etai (1986) reported homology between the chvA and chvB loci of A. tumefaciens and so-called ndvA and ndvB (for nodulation development) loci in Rhizobium meliloti. In reports on R. meliloti ndvB mutants it is proposed that p-1,2-glucan is not absolutely required for nodule development but might play a role in enhancement of nodulation (Dylan etai, 1990). Wild-type A. tumefaciens cells harbouring a Sym plasmid of H. leguminosarum bv. phaseoli are able to nodulate siratro and bean whereas c/ivS mutants harbouring a Sym plasmid cannot {Van Veen e( al., 1987; this study). We demonstrated that, only after treatment of the roots of siratro or bean plants with rhicadhesin, strain LBA2753 602 S. Swart, G. Smit, B. J. J. Lugtenberg and J. W. Kijne Table 4. Sacterial strains. Strain Genotype or phenotype Reference/Source LBA1010 A. tumefaciens C58 chromosome, pTiB6 A. tumefaciens C58 chromosome, pTi- and pCr-cured A. tumefaciens C58 chromosome chvB::Jn5 A. tumefaciens C58 chromosome chv&.:Tn5 A. tumefaciens C58 chromosome chvB:.Tn5 A. tumefaciens C58 chromosome chvS::Tn3::HoHo1 A. (umefec/ensLBAIOlO pSym9 ot RCC3622 A. lumefaciens M E n 7 , pSym9 of RCC3622 R. ieguminosarum bv. phaseoli (wild-type) R. ieguminosarum bv. viciae (wild-type) Koekman era/. (1982) LBA1251 A1011 A1020 A1045 ME117 LBA2728 LBA2753 RCC3622 248 Rosenberg and Huguet (1984) Douglas e( af. (1982) Douglas ef a/. (1982) Douglas efa/. (1982) Douglas sfaA (1985) Van Veen e/a/.(1987) Van Veen efa/. (1987) Hooykaas efa/. (1985) Josey efa/. (1979) encoded (Douglas etai., 1982; Matthysse, 1987; Smit ef al., 1986; 1987). (iii) Bacteria from both genera produce rhicadhesin, and, moreover, rhicadhesin, purified from one genus, can inhibit attachment of cells of the other genus to legume root hair tips (Smit e/a/., 1989b). (tv) For both genera attachment is a two-step process. The second step, mediated by cellulose fibrils, is not essential for the establishment of a successful association (Matthysse, 1983; Smit etaL, 1987). The first step, mediated by rhicadhesin, is necessary for an interaction between plant and bacterium (Smit ef aL, 1989b; this paper), (v) The attachment ability to root hairs as well as virulence of ChvB mutants can be restored by addition of rhicadhesin isolated from either Rhizobium or Agrobacterium (this paper), (vi) The virulence of chvB mutants on Kaianchoe as well as nodulation of bean and siratro by cArvS mutants harbouring a Sym plasmid of R. leguminosarum bv. phaseoli can be restored by pretreatment of the roots with rhicadhesin (this paper). Experimental procedures was able to nodulate these roots, Heterologous complementation of A. tumefaciens chvB mutants with R. meliloti ndvB genes and vice versa demonstrated the functional interchangeability of the chv and ndv loci (Dylan ef al., 1986). Also ndvB mutants of R. meliloti were found to be affected in infection (Geremia et ai, 1987). However ndvB mutants form pseudonodules whereas chvB mutants harbouring a Sym plasmid do not form nodules at all (Van Veen ef al., 1987). Furthermore, ndvB mutants form abortive infection threads and show attachment, although diminished (Dylan et aL, 1990). Indeed, we found rhicadhesin activity in cell surface preparations of a ndvB mutant (S. Swart, unpublished results). Apparently, a mutation in the gene encoding the intermediate protein in the p-1,2-glucan synthesis, chvB in the case of A. tumefaciens or ndvB in the case of R. meliloti, does not result in the same phenotype for these Rhizobiaceae. Rhicadhesin activity might be regulated in a different manner in A. tumefaciens and R. meliloti. Comparision of binding properties of Rhizobium and Agrobacterium The following evidence supports the hypothesis that Rhizobium and Agrobacterium bind to the surface of plant celis through a similar mechanism, (i) The chromosome of bacteria from each genus contains the necessary information for establishing an association with host plants of the other genus, provided that the proper plasmid (Sym or Ti) is present (Hooykaas etai., 1977; 1981; 1985; Kondorosi etaL, 1982; Hirsch et ai., 1985). (ii) The binding properties of cells of both genera are chromosomally Bacterial strains and culture conditions Bacterial strains of A. tumefaciens and R. leguminosarum used are listed in Table 4. Compositions of the media A+ and LC (Smit etai., 1987) have been described previously. TY medium contains (per litre); tryptone (Difco), 5.0g and yeast extract (Difco), 3.0 g. TYC medium is TY medium supplemented with 7 mM CaClg. Rhizobium and Agrobacterium species were maintained on solid A+ and solid LC medium, respectively. For attachment assays, bacteria were cultivated on a rotary shaker (180 r.p.m.) at 28^C in 100 ml Erienmeyer flasks, containing 50 ml TYC medium. For purification ot rhicadhesin, bacteria were grown in either 500 ml or 21 Erienmeyer flasks containing 250 ml or 1.25 I TYC medium, respectively. Plants and plant culture conditions Seeds of pea {Pisum sativum cv. finale; Cebeco) were surface sterilized and cultivated as described previously (Smit et al., 1986). Seeds of siratro {Macroptilium atropurpureum, a generous gift from Dr B. G. Rolfe, Australian National University, Canberra, Australia) were surface sterilized as described for pea seeds and germinated on solid Jensen agar (Vincent, 1970) at 28'^C for about 20 h. Subsequently, plants were grown in liquid Jensen medium at 28''C in a growth chamber. The light intensity at table surface was 20000 iux during a 16h lightperiod. K. daigremontiana and K. tubiflora plants were cultivated as described by Ooms efa/- (1980). Bean seeds (Phaseolus vuigaris cv. prelude) were surface sterilized and germinated between wetted Whatman filter paper at 22°C. After germination for 4d the seedlings were grown in glass tubes on solid Jensen medium slopes at 22°C in a growth chamber with light intensity of 20 000 lux during a 16h lightperiod. Role of rhicadhesin in virulence and nodulation Attachment assay The attachment assay used has been described previously by Smit etal. (1986). Briefly, roots were immersed in a suspension of bacteria, washed 10 times and attachment was quantified by randomly screening at least 100 developing root hairs with a phase-contrast microscope. Attachment was classified in three classes: class 1, no attached bacteria; class 2, few bacteria, directly attached to the root hair; class 3, many attached bacteria, forming two or more layers of bacteria or a cap-like aggregate at the tip of the root hair (cap formation). The percentage of root hairs of each class was calculated. The variability of the test is about 10% and depends largely on the condition of the roots. Purification of rhicadhesin Procedures for the purification of rhizobial rtiicadhesin (Smit ef ai, 1989b), isolation of cell surface preparations from A. tumefaciens strains (Smit etai, 1989a), and testing the presence of rhicadhesin in culture supernatants of cells (Smit et al., 1991) have been described previously. Presence of rhicadhesin inside bacteria was examined after sonication of the cells for 5x 30s in a sonifier cell disrupter (Branson Sonic Power Comp.). After sonication of the cells, the suspension was clarified by ultracentrifugation (100 000 x gfor 2 h) and tested for adhesin activity. SDS-PAGE and protein staining SDS-PAGE was performed as descibed by Schagger and Von Jagow (1987). Proteins were silver-stained according to Oakley etai (1980), The molecular masses of the marker proteins were; bovine serum albumin, 66 kDa; egg albumin, 45 kDa; carbonic anhydrase, 29 kDa; p-lactoglobu!in, 18 kDa; and lysozyme, 14 kDa. Isolation ofp-1,2-glucan Stationary cells of A. tumefaciens LBA1010, grown in TYC. were harvested by centrifugation at 10000 x pfor lOmin. Pellets were extracted with 1 % trichloroacetic acid (TCA) for 30 min at room temperature. TCA extracts were neutralized, concentrated, chromatographed on Sephadex G-50 (Pharmacia), and tested according to the method of Miller e/a/. (1986). Test for rhicadhesin activity Activity of rhicadhesin was determined by measuring the ability to inhibit attachment of LBAIOIO cells to pea root hair tips. To test whether a fraction contained active rhicadhesin, two lateral pea roots were incubated together with the fraction in 1.0 ml of 25 mM potassium phosphate buffer, pH 7-5, for 60 min at room temperature on a rotary table (2 r.p.m.). Controls were incubated for 60 min in buffer only. After this treatment the roots were washed five times in phosphate buffer and incubated with bacteria in an attachment assay as described above (see also Smit era/., 1989b). Restoration of attachment ability of A. tumefaciens chvB mutants was tested by incubating rhicadhesin-treated roots 603 with bacteria grown in TY as described in the attachment assay. Virulence assays In the greenhouse K- daigremontiana and K. tubifiora plants were wounded on leafs and stems, respectively. A solution of rhicadhesin (2-10ng per wound site) and/or p-1,2-glucan (10 ng per wound site) in 25 mM phosphate buffer, pH7,5, was added in a volume of 5-25^1 per wound site. Controls were incubated with phosphate buffer only. Ten to twenty microlitres of A. tumefaciens chvB mutant cells, grown in TY medium, harvested and suspended in phosphate buffer to an Agjo value of 1.0, were added to each wound site 30 min after incubation with rhicadhesin or buffer. As a control wild-type A. tumefaciens was added to wound sites. Tumour formation was quantified three weeks after inoculation. Nodulation assay After germination, sirato seedlings were transferred to Jensen medium (5-6 seedlings per glass tube containing 25 ml medium) and grown overnight at 28°C in the dark. Subsequently, an appropriate number of roots were treated with 0.1 fig ml"' rhicadhesin in Jensen medium for 30 min. Twentyfive microlitres of A. tumefaciens LBA2753 cells, grown in TY medium, harvested and resuspended in Jensen medium to an A620 value of 0.1, were added to each tube. Strains LBA2728 and RCC3622 were used as controls. Nodulation was quantified 3 weeks after inoculation. After germination, bean seedlings were grown on Jensen medium for 3d in the dark before inoculation. One millilitre 25 mM phosphate buffer, pH7.5, or 1.0 ml rhicadhesin (0.5|igmr') in phosphate buffer was added to the roots and after 1 h the roots were inoculated with 3.0 ml of A. tumefaciens strain LBA2753 cells, grown in TY medium, harvested and resuspended in Jensen medium to an Agao value of 0.05. Strains LBA2728 and RCC3622 were used as controls. Nodulation was quantified 5 weeks after inoculation. Acknowledgements This work was supported by the Foundation for Fundamental Biological Research (BION), which is subsidized by the Netherlands Organization of Scientific Research (NWO). The authors thank Trudy Logman, Teun Tak and Bram Wetselaar for supply of test plants, and Ron van Veen for helpful discussions. References Cangelosi, G.A., Hung, L., Puvanesarajah, V., Stacey, G., Ozga, A.D., Leigh, J.A., and Nester, E.W. (1987) Common loci for Agrobacterium tumefaciens and Hhizobium meliloti exopolysaccharide synthesis and their roles in plant interactions, J Bacteriol 169: 2086-2091, Cangelosi, G.A., Martinetti, G., Leigh, J.A., Chang Lee, C , Theines, C , and Nester, E.W. (1989) Role for Agrobacterium tumefaciens ChvA protein in export of p-1,2-glucan. J Bacteriol ni: 1609-1615. 604 S. Swart, G. Smit, B. J. J. Lugtenberg and J. W. Kijne Douglas, C , Halpecin, W., and Nester, E.W. (1982) Agrobacterium tumefaciens affected in attachment to plant cells. J 0acter/o/152: 1265-1275. Douglas, C.J., Staneloni, R.J., Rubin, R.A., and Nester, E.W. (1985) Identification and genetic analysis of an Agrobacterium tumefaciens chromosomal virulence region, J Bacterio/161:850-860. Dylan, T., lelpi, L., Stanfield, S., Kashyap, L., Douglas. C , Yanofsky, H., Nester, E,, Helinski, D.R., and Ditta, G. (1986) Rhizobium meliloti genes genes required for nodule development are related to chromosomal virulence genes in Agrobacterium tumefaciens. Proc NatI Acad Sci USA 83: 4403-4407. Dylan, T., Nagpal, P., Helinski, D.R., and Ditta, G.S. (1990) Symbiotic pseudorevertants of Rhizotjium meiiioti ndv mutants. J Bacteriol 172:1409-1417. Geremia, R.A., Cavaignac, S., Zorreguieta, A., Toro, N., Olivares, J., and Ugalde, R.A, (1987) A Rhizobium meliloti mutant that forms ineffective pseudonodules in alfalfa produces exopolysaccharide but fails to form p-1,2-glucan. J Sacferio/169: 880-684. Hirsch. A.M., Drake, D., Jacobs, T.W,, and Long, S.R. (1985) Nodules are induced on alfalfa roots by Agrobacterium tumefaciens and Rhizobium trifolH containing small segments of the Rhizobium meliloti nodulation region. J Bacteriol ^61: 223-230. Hirsch, P.R., Van Montagu, M., Johnston, A.W.B., Brewin, N.J., and Schell, J. (1980) Physical identification of bacteriocinogenic, nodulation and other piasmids in strains of Rhizobium leguminosarum. J Gen Microbiol MO: 403-412. Hooykaas, P.J.J., Klapwijk, P.M,, Nuti, M.P., Schilperoort, R.A., and Rorsch, A. (1977) Transfer of the Agrobacterium tumefaciens Ti plasmid to avirulent agrobacterla and to Rhizobium ex planta. J Gen Microbio!9B: 477-484. Hooykaas, P.J.J,, Van Brussel, A.A.N., Den Dulk-Ras, H., Van Slogteren, G,H.S., and Schilperoort, R.A. (1981) Sym plasmid of Rhizobium trifolii expressed in different rhizobial species and Agrobacterium tumefaciens. Nature 291: 351-353. Hooykaas, P.J.J., Den Dulk-Ras, H., Regensburg-Tuink, A.J.G., Van Brussel, A,A.N., and Schilperoort, R.A. (1985) Expression of a Rhizcbium phaseoli Sym plasmid in R. tritoiii and Agrobacterium tumefaciens: Incompatibility with a R. trifoliiSym plasmid. PlasmidAA: 47-52. Ifion de lannino, N., and Ugalde, R.A. (1989) Biochemical characterization of avirulent Agrobacterium tumefaciens chvA mutants: synthesis and excretion of p-(1,2)glucan. J Bacter/o/171: 2842-2849Johnston, A.W.B., Beynon, J.L., Buchanan-Wollaston, A.V., Setchell, S.M., Hirsch, P.R., and Beringer, J.E. (1978) High frequency transfer of nodulating ability between strains and species of Rhizobium. Nature 276: 635-636. Josey, D.P., Beynon, J.L, Johnston, A.W.B,, and Beringer, J.E. (1979) Strain identification in Rhizobium using intrinsic antibiotic resistance. J AppI Microbiol A6: 343-350. Koekman, B,P., Hooykaas, P.J.J., and Schilperoort, R.A. (1982) A functional map of the replicator region of the octopine Ti plasmid. Plasmidl: 119-132. Kondorosi, A-, Kondorosi, E., Pankhurst, C.E., Broughton, W.J., and Banfalvi, Z- (1982) Mobilization of a Rhizobium meliloti megaplasmid carrying nodutation and nitrogen fixation genes into other rhizobia and Agrobacterium. Mol Gen Genef 188: 433-439. Kijne, J.W-, Smit, G., Diaz, C,L, and Lugtenberg, B.J,J. (1988) Lectin enhanced accumulation of manganese-limited Rhizobium ieguminosarum cells on pea root hair tips. J Bacteriol 170:2994-3000. Matthysse, A.G. (1983) Role of bacterial cellulose fibrils in Agrobacterium tumefaciens infections. J Bacteriol 154: 906-915. Matthysse, A.G. (1987) Characterization of nonattaching mutants of Agrobacterium tumefaciens. J Bacteriol 169: 313-323. Miller, K.J., Kennedy, E.P., and Reinhold, V,N. (1986) Osmotic adaptation by gram-negative bacteria: possible role for periplasmic oligosaccharides. Science 231: 48-51. Nuti, M.P., Lepidi, A.A., Prakash, R.H., Schilperoort, R.A., and Cannon, F.C. (1979) Evidence for nitrogen fixation {nif) genes on indigous Rhizcbium piasmids. Nature 282: 533-535. Oakley, B.R., Kirch, DR., and Morris, N.R- (1980) A simplified ultrasensitive silver stain for detecting proteins in polyacrylamide gels. Anal Biochem W5: 36-\-363. O'Connel, K.P,, and Handelsman, J. (1989) chvA locus may be involved in export of neutral cyclic p-1,2-linked D-glucan from Agrobacterium tumefaciens. Mol Plant-Microbe Int 2: 11-16. Ooms, G., Hooykaas, P.J.J., Poulis, J.A., and Schilperoort, R.A. (1980) Characterization of Jn904 insertions in octopine Ti plasmid mutants of Agrobacterium tumefaciens. J Bacter/D/144: 82-91. Puvanesarajah, V-, Schell, F.M., Stacey, G., Douglas, C.J., and Nester, E.W. (1985) Role for 2-linked-p-D-glucan in the virulence of Agrcbacterium tumefaciens. J Bacteriol 164: 102-106. Rosenberg, C , and Huguet, T. (1984) The pAtC58 plasmid of Agrobacterium tumefaciens is not essential for tumor induction, Mol Gen Genet 196: 433-436. Schagger H., and Von Jagow, G- (1987) Tricine-sodium dodecyl su Ifate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa- Anal e/oc/iem 166: 368-379. Smit, G., Kijne, J.W., and Lugtenberg, B,J-J. (1986) Correlation between extracellular fibrils and attachment of Rhizobium leguminosarum to pea root hair tips. J Bacteriol 168: 8 2 1 827, Smit, G., Kijne, J,W,, and Lugtenberg, B,J.J, (1987) Both cellulose fibrils and a Ca^*-dependent adhesin are involved in the attachment of Rhizobium leguminosarum to pea root hair tips. JSacterio/169: 4294-^301. Smit, G-, Kijne, J.W., and Lugtenberg, B.J.J. (1989a) Roles of flagella, lipopolysaccharide, and a Ca^'-dependent cell surface protein in attachment of Rhizobium leguminosarum biovar viciae to pea root hair tips. JBacterioHT'l: 569-572. Smit, G., Logman, T.J.J., Boerrigter, M.E.T.I., Kijne, J.W., and Lugtenberg, B.J.J. (1989b) Purification and partial characterization of the Ca^* dependent adhesin from Rhizobium leguminosarum biovar viciae, which mediates the first step in attachment of Rhizobiaceae cells to plant root hair tips. J Bacteriol 171: 4054^062. Smit, G., Tubbing, D.M.J., Kijne, J.W., and Lugtenberg, B.J.J. (1991) Role of Ca^* in the activity fo rhicadhesin from Rhizobium leguminosarum biovar viciae, which mediates the first step in attachment of Rhizobiaceae cells to plant root hair tips. Arch Microbion55: 278-283. Smit, G., Swart, S., Lugtenberg, B.J.J., and Kijne, J.W. (1992) Molecular mechanisms of attachment of Rhizobium bacteria to plant roots. Mol Microbiol S: 2897-2903. Role of rhicadhesin in virulence and nodulation Szejtii, J. (1991) Cyclodextrins in drug formulation: part II, Phartn TechnintZ: 16-24Thomastiow. M F., Karlinsey, J.E., Marks, J.R., and Hulbert. R.E. (1987) Identification of a new locus in Agrobacterium tumefaciens that effects polysaccharide composition and plant cell attachment. J Bacteriol 169: 3209-3216. Van Larebeke. N.. Engler, G., Holsters, M., Van den Elsacker, S.. Zaenen, I , Schilperoort, R.A., and Schell, J, (1974) Large plasmids in Agrobacterium tumefaciens essential for crown gall-inducing ability. Wafure242:171-172. Van Veen, R.J.H,, Den Dulk-Ras. H.. Schilperoort, R.A., and Hooykaas, P.J.J. (1987) Chromosomal nodulation genes: Sym-plasmid containing Agrobacterium strains need chromosomal virulence genes {chvA and chvB) for nodulation. Plant Molec Biol 8:105-108. 605 Vincent, J.M. (1970) A Manual for the Practical Study of Rootnodule Bacteria. I.B.P. Handbook no. 15. Oxford: Btackwell Scientific Publications, Ltd. pp 75-82. Wagner. V.T., and Mattytisse, A.G. (1992) Involvement of a vitronectin-like protein in attachment of Agrobacterium tumefaciens to carrot suspension culture cells. J Bacteriol 174: 5999-6003 Zaenen. i.. Van Larebeke. N., Teuchy, L., Van Montagu. H.. and Schell, J. (1974) Supercoiled circular DNA in crown-gall inducing^groftacferiumstrains. JMolBiol8B: 109-127. Zorreguieta, A., Geremia, R.A,, Cavaignac, S., Cangelosi, G.A., Nester, E.W., and Ugalde, R.A. (1988) Identification of the product of an Agrobacterium tumefaciens chromosomal virulence gene. Mol Plant-Microbe Int^: 121-127.