Rhizobium meliloti exopolysaccharides: genetic ... - Semantic Scholar

Biochemical Society Transactions

~

Rhizobiurn meliloti exopolysaccharides: genetic analyses and symbiotic importance

636

T. Lynne Reuber, Jason Reed, Jane Glazebrook, M. Alexandra Glucksmann, Dianne Ahmann, Andrea Marra, and Graham C. Walker Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02 I 39, U.S.A.

Summary Genetic experiments have indicated that succinoglycan (EPS I), the acidic Calcofluor-binding exopolysaccharide, of the nitrogen-fixing bacterium Rhizobium meliloti strain Rm1021 is required for nodule invasion and possibly for later events in nodule development on alfalfa and other hosts. Fourteen ex0 loci on the second megaplasmid have been identified that are required for, or affect, the synthesis of EPS I. Mutations in certain of these loci completely abolish the production of EPS I and result in mutants that form empty Fix- nodules. We have identified two loci, exoR and exoS, that are involved in the regulation of EPS I synthesis in the free-living state. Certain exo mutations which completely abolish EPS I production are lethal in an exoR9.5 or exoS96 background. Histochemical analyses of the expression of ex0 genes during nodulation using exo::TnphoA fusions have indicated that the ex0 genes are expressed most strongly in the invasion zone. In addition, we have discovered that R. meliloti has a latent capacity to synthesize a second exopolysaccharide (EPS 11) that can substitute for the role(s) of EPS I in nodulation of alfalfa but not of other hosts. Possible roles for exopolysaccharides in symbiosis are discussed.

Introduction Rhizobia fix nitrogen in symbiotic association with leguminous plants. In the course of this symbiotic interaction, the bacteria induce the formation of nodules on the roots of the plant, enter these nodules through tubes called infection threads, are encoated in a membrane of plant origin and released into the interior of plant cells, differentiate into a morphologically distinct form termed bacteroids, and then begin to fix nitrogen (for reviews see [I-71). Rhizobium meliloti is able to form nitrogen-fixing nodules on alfalfa and certain other legumes. Considerable attention has been devoted to the mechanisms whereby a Rhizobium strain and its host plant might recognize each other, and whereby the bacteria enter the nodules which they induce. The products of the rhizobial nod genes have been shown to be central to this process and Abbreviations used: EPS I, succinoglycan, EPS 11. exopolysaccharide 11.

Volume 19

are under active investigation by numerous laboratories [4, 81. A recent study has shown that the R. meliloti nod genes determine the synthesis of a sulphated, acylated glucosamine oligosaccharide that seems to function as an extracellular signal to the plant during nodulation [9]. In addition, most rhizobia produce a variety of polysaccharides [ 1, 101, and over the years it has been hypothesized that these polysaccharides play important roles in bacterial-plant interactions, for example being responsible for the host specificity of various Rhizobium strains. In the course of our attempts to understand the molecular mechanisms of nodulation of legumes by nitrogen-fixing Rhizobia we discovered that exopolysaccharides excreted by R. meliloti play key roles in invasion and development of nodules on alfalfa and other plant hosts. W e have been carrying out a set of genetic analyses of exopolysaccharide biosynthesis by this organism and have been attempting to understand the biological functions of exopolysaccharides in nodulation. The studies have been facilitated by the availability of a wide variety of genetic techniques for R. meliloti which permit essentially any genetic strategy used in the study of Escherichia coli to also be used in the study of R. meliloti [l l ]. In this paper we discuss our genetic analyses of the regulation and synthesis of exopolysaccharides by R. meliloti, possible roles for these exopolysaccharides in nodulation, and examples in which our genetic studies have allowed us to manipulate the regulation of synthesis, structure or molecular mass of exopolysaccharides.

Succinoglycan synthesis by R. meliloti R. meliloti strain Rm 1021 excretes succinoglycan (EPS I), a high molecular mass exopolysaccharide [ 121 composed of polymerized octasaccharide subunits [13]. Each octasaccharide consists of a backbone of three glucoses and one galactose, a side chain of four glucoses, and 1-carboxyethylidene (pyruvate), acetyl, and succinyl modifications in a ratio of approximately 1: 1 : 1. Succinoglycan is also synthesized by Alcaligenes faecalk var. myxogenes [ 141, Agrobacterium tumefaciens [ 151, Agrobacterium radiobacter [ 161, and Pseudomonas sp. [ 171. The structure of EPS I was determined through studies in a variety of laboratories [18-201, although the

Genetics of Oligosaccharide Metabolism

positions of the acetyl and succinyl modifications have not yet been unambiguously established. On the basis of various studies it appears that the succinyl group is on the side chain and the acetyl group is on the backbone [18, 19, 211. Before polymerization, the subunits are assembled on a lipid carrier and labelling studies in vitro have shown that the sequence of assembly is: first, galactose and /?1, 3-glucose, then the rest of the glucose residues, and finally the other substituents [22, 231. Succinoglycan has certain resemblances to the commercially important exopolysaccharide, xanthan gum, of Xanthomonas campestrk, and is of commercial interest in its own right [24, 251. As discussed below, R. meliloti strain Rm1021 also has the capacity to make a second exopolysaccharide, EPS I1 [26].

EPS I of R. meliloti is required for nodule invasion Our laboratory has recently obtained strong genetic evidence that EPS I [ 13, 201 is required for nodule invasion by R . meliloti strain Rm1021 and quite possibly for later events in nodule development [27-291. W e isolated a set of mutants (exo) of R. meliloti Rm1021 on the basis of their failure to fluoresce under U.V.light on medium containing Calcofluor and showed that these mutants did not synthesize EPS I [28]. Alfalfa seedlings inoculated with these ex0 mutants formed ineffective (nonnitrogen-fixing) nodules that contained few if any bacteria and no bacteroids [28]. More detailed characterizations of nodules elicited by an exoB mutant revealed that, in one genetic background, no infection threads were formed following inoculation with this mutant and that the plant cells in the interior of nodules elicited by this strain did not contain bacteria or bacteroids [27]. Subsequent studies [29] of a particular mo mutant generated by insertion of the transposon Tn5 in R. meliloti strain Rm1021 have shown that root hair curling is significantly delayed. Infection threads can form but abort at a very early stage so that the bacteria are never able to reach the interior of the nodule. Furthermore, these ex0 mutants are only able to elicit the synthesis of two of the nodule-specific plant proteins, termed nodulins, as opposed to the 17 that are elicited by infection with a wild-type R. meliloti [29-321. Thus these R. meliloti ex0 mutants uncouple the ability of the bacteria to signal to the plant to initiate nodule formation from the ability of the bacteria to invade the nodule and establish an effective symbiosis.

Similar findings implicating Rhizobium exopolysaccharides in normal nodule development have been reported for the nodulation of clover by R. tnifolii [33], the nodulation of Leucaena by a Rhizobium sp. strain NGR234 [34], and the nodulation of pea by R. leguminosarum [35].

Genetic analyses of the synthesis of EPS I by R. meliloti W e were able to subdivide our initial set of R. meliloti ex0 mutants into five different genetic classes [28]. Four of these classes were then shown to be located on the second symbiotic megaplasmid, pRmeSU47I3, of R. meliloti and another class (exoC) was shown to be located on the chromosome [36]. A detailed genetic analysis of the cluster of ex0 genes located on the second symbiotic megaplasmid has led to the identification of a number of loci that are required for the synthesis of the R . meliloti EPS I [37]. Mutations in exoA, exoB, exoF, exoM, exoL, exoP, exoQ, and exoT, completely abolish production of the EPS I and result in mutants that form Fix- nodules [28, 37, 381. Analyses of the properties of protein fusions of various ex0 genes to alkaline phosphatase, that were generated using TnphoA [ 391, suggest that the moF, exoP, exoQ and exoA gene products are membrane proteins [38, 401. To determine the roles of the ex0 gene products in the synthesis of EPS I, we have initiated studies in vitro of the synthesis of EPS I by these various mutants (T. I,. Reuber & G. C. Walker, unpublished work). In addition, we have found that mutations in exoG, exoJ and exoN [37], diminish the production of Calcofluor-binding material. Mutants of exoG and exoJ form effective nodules with decreased efficiency, whereas plants inoculated with exoN mutants fix nitrogen normally. The exoG and exoJ mutants are of particular interest since they produce no detectable high molecular mass exopolysaccharide, yet the low molecular mass Calcofluor-binding material they do produce seems to be sufficient to allow some amount of nodule invasion and develcause an opment to proceed. Mutations called WCOX increase in exopolysaccharide production and map in the same region as the moG and exor mutations [41]. Recent analyses [42] have shown that the mutations originally referred to as exoG and exoJ appear to define a region involved in the regulation of EPS I biosynthesis. The DNA sequence of this region revealed that it contained two divergently transcribed open reading frames called exoX and RXOYthat have homologues in other Rhizobium species [41, 43-45]. The exoG insertions fall in the

1991

637


638

intergenic regions and may affect the expression of exoX and exoY. The exor mutation falls in the 3'portion of the exoX open reading frame, and is probably an allele of exoX that results in altered function. Gene regulation studies suggested that ExoX and ExoY comprise a system that modulates exopolysaccharide synthesis at a post-translational level [42]. Interestingly, we found that the deduced amino acid sequence of ExoY shares homology with the deduced amino acid sequence of GumD, a protein required for an early step in xanthan gum biosynthesis, further suggesting that the modulatory system may affect the exopolysaccharide biosynthetic apparatus.

R. meliloti mutants that fail to succinylate their EPS I are defective in nodule invasion A very interesting class of ex0 mutants that synthesize a structural variant of EPS I was originally identified on the basis of the failure of colonies of such mutants to form a fluorescent halo under U.V. light, when grown on medium containing Calcofluor [29]. These mutations defined a locus termed exoH which mapped in the middle of a cluster of ex0 genes on the second symbiotic megaplasmid. Alfalfa seedlings inoculated with exoH mutants form ineffective nodules that do not contain intracellular bacteria or bacteroids. Root hair curling is significantly delayed and infection threads abort in the nodule cortex. In other words, despite the fact that these mutants made Calcofluor-binding material, their behaviour on plants was the same as those ex0 mutants that produced no EPS I. Analyses of exopolysaccharide secreted by exoH mutants have shown that it is identical to the Calcofluor-binding exopolysaccharide secreted by the parental exoH+ strain except that it completely lacks the succinyl modification. Translation, in vitro, of total RNA isolated from nodules induced by an exoH mutant has shown that, as in the case of the exopolysaccharide-deficient ex0 mutants, only two of the plant-encoded nodulins are induced, as compared with the 17 nodulins induced by the wild-type strain. These observations raise the possibility that succinylation of the bacterial exopolysaccharide is important for its role(s) in nodule invasion and possibly nodule development [29]. Our ability to isolate exoH mutants suggests that the biosynthetic pathway for the synthesis of EPS I is suficiently flexible to allow the polymerization of non-succinylated octasaccharides to yield

Volume 19

high molecular mass polysaccharide. The resulting polysaccharide has lower charge density. W e are interested in obtaining mutants that synthesize variants of EPS I lacking the acetyl and carboxyethylidene modifications but have not yet succeeded in obtaining such mutants.

Regulation of EPS I biosynthesis in the free-living state W e initially observed that, as for the synthesis of several other bacterial heteropolysaccharides [461, the synthesis of EPS I by R. meliloti was greatly increased if the cells were limited for nitrogen, phosphorus or sulphur in the presence of a good carbon source. We identified two new unlinked loci, exoR and exoS, the products of which play a role in regulating the synthesis of EPS I by R. meliloti strain Rm1021 [47]. Tn5-generated mutations in these loci are recessive and lead to substantial increases in the amount of exopolysaccharide synthesized, indicating that the exoR and exoS gene products play negative roles in regulating exopolysaccharide synthesis. Introduction of an exoR95::TnS or exoS96::TnS mutation into a strain containing an exo::TnphoA fusion results in a 2-5fold increase in the level of expression of alkaline phosphatase activity, suggesting that they negatively regulate ex0 expression. A fundamental difference between the exoR95::TnS and exoS96::TnS mutants is that the exoR95::TnS mutant synthesizes its EPS I at a high constitutive level regardless of the presence or absence of ammonia in the medium, whereas the exoS96::TnS mutant undergoes a further increase in the rate of synthesis upon nitrogen starvation. It seems that the exoR gene product is either involved directly in sensing the level of nitrogen in the medium, or else that it acts later in a putative regulatory cascade than the element(s) that actually does the sensing. The relationship of exoS action to exoR action is not yet clear. In recent work (J. W. Reed, J. Glazebrook & G. C. Walker, unpublished work), we have shown that an exoR strain contains higher levels of mRNA for other ex0 genes than the wild-type parental strain. ExoR therefore most probably exerts its regulatory effect at the level of transcription. W e have also localized, subcloned and sequenced the exoR gene, but the deduced amino acid sequence of ExoR does not show similarity to any other sequenced gene. In the course of examining the regulation of exo::TnphoA fusions by exoR and exoS, we found


that five of the eight classes of ex0 mutations which produced a complete block of EPS 1 biosynthesis (exoL, M , P,Q and T ) are lethal in exoR or exoS backgrounds [38]. These double mutants are viable, however, when a plasmid complementing the ex0 mutation is present. This implies that blocking EPS I biosynthesis at certain stages results in an accumulation of intermediates that are toxic when overproduced. One possibility is that buildup of lipid-linked intermediates may deplete the available pool of lipid carriers and therefore block lipopolysaccharide and peptidoglycan biosynthesis. Availability of lipid carriers has been proposed to be a factor influencing polymer biosynthesis [461.

Regulation of EPS I synthesis during nodulation W e found that it was possible to stain nodules specifically for the alkaline phosphatase activity present in the inducing bacteria by staining at pH 9 [38]. At this pH, nodules induced by a Pho- strain did not stain. Nodules induced by a strain carrying the exoF369::TnphoA fusion and a plasmid which complements the exoF mutation to allow normal nodulation, showed staining primarily in the early symbiotic or invasion zone of the nodule where the bacteria were invading the plant cells. In the late symbiotic zone, which contained mature bacteroids, no staining was seen. Nodules induced by strains carrying exoP::TnphoA and exoA ::TnphoA fusions showed a lower degree of staining than that in nodules induced by the exoF:TnphoA strain, but faint staining of the invasion zone was seen after long incubations. These results suggest that little or no new EPS I synthesis is needed after nodule invasion. exoS96::Tn5 mutants formed Fix+ nodules on alfalfa [47]. In contrast, we found that on alfalfa, wcoR95::Tn5 mutants formed both empty Fixnodules and also Fix+ nodules that contained widely varying numbers of bacteria and bacteroids [47]. All the bacteria we isolated from the Fix+ nodules induced by the exoR95::Tn5 strain had acquired unlinked suppressors that reduced the amount of exopolysaccharide produced, suggesting that the bacteria need to control either how much EPS I they synthesize, or when they synthesize it, to invade nodules. These suppressors could (i) reduce the activity of a positively acting factor, (ii) reduce the activity of an enzyme, not encoded within the ex0 cluster which is required for the synthesis of the exopolysaccharide, or (iii) result from the increased

expression of a gene such as exox, the product of which exerts a negative effect on EPS I synthesis.

R. rneliloti can produce a second exopolysaccharide that can substitute for the role of EPS I in nodule invasion The symbiotic defects of ex0 mutants can be suppressed by the presence of a mutation, expRIO1, which causes overproduction of a second exopolysaccharide, EPS 11 [26]. Genetic analyses have shown that the products of a cluster of at least six exp genes located on the second symbiotic megaplasmid, as well as the product of the exoB gene, are required for EPS 11 synthesis. The presence of the e@RIOl mutation causes increased transcription of the exp genes, resulting in overproduction of EPS 11. As a consequence of this genetic analysis, we were able to construct strains which produced EPS I or EPS I1 exclusively, or neither. Medicago sutivu plants inoculated with an EPS-11-producing strain formed nitrogen-fixing nodules, indicating that EPS I1 is able to substitute for the symbiotic role of EPS I in the nodulation of alfalfa by R. melilotz The structure of EPS 11 has been determined [26, 481, and consists of a polymer of glucosepl-3galactose disaccharides joined by a 1-3 linkages. The glucose carries an 0-6-acetyl modification and the galactose a-4-6 carboxyethylidene linkage. The independent discovery of this exopolysaccharide has been reported by Zhan et ul [49] and the partial structure they reported is in agreement with that determined by Her et ul [48]. Both EPS I and EPS I1 are acidic, contain glucose and galactose, and have acetyl and pyruvate (1carboxyethylidene) modifications. However, the structures differ in many respects: (i) EPS I1 does not contain any succinate groups, (ii) EPS I1 contains more galactose than EPS I, (iii) EPS I1 is unbranched, (iv) EPS I1 has both a and /3 glycosidic linkages, while EPS I has only /3 glycosidic linkages, and (v) the pyruvate group in EPS I is linked to glucose, while in EPS I1 it is linked to galactose. However, it is interesting to note that each exopolysaccharide has a single Glcp 1-3Gal linkage in its backbone and that these two rather diverse exopolysaccharides may share the common structural motif of 0-6-acetylglucose/3 1-3galactose.

Possible roles for exopolysaccharides in nodulation Exopolysaccharides may have more than one function in nodulation. Since exoG mutants, which make

1991

639


640

low molecular mass EPS I, but not high molecular mass EPS I, form effective nodules at reduced efficiency relative to wild type, but much better than mutants which make no exopolysaccharide, it is possible that both the high and low molecular mass forms of EPS have symbiotic functions. One of the most intriguing possibilities for the role of low molecular mass forms of the exopolysaccharides in nodulation is that they act as signals to the plant during the process of nodule invasion and development. Carbohydrates have previously been shown to function as signal molecules in plants [SO]. Several lines of evidence suggest that oligosaccharide fragments of EPS I may have a symbiotic function. First, exoG mutants, which produce mainly low molecular mass EPS I, can invade nodules, albeit at reduced efficiency [37]. Secondly, both we and John Leigh and his colleagues have obtained preliminary evidence that a low molecular mass fraction of EPS I can partially suppress the symbiotic deficiencies of R. meliloti ex0 mutants (personal communication). A similar finding has been reported previously by Djordjevic et al. [ 5 13, who found that ex0 mutants of Rhizobium sp. NGR234 and R. tnyolii can form effective nodules if either high molecular mass EPS or oligosaccharide subunits of EPS are supplied exogenously. However, there appears to be some difference between the R. meliloti-alfalfa system and those studied by Djordjevic et al. [Sl], since neither we nor others [52] have been able to suppress the symbiotic deficiencies of R. meliloti ex0 mutants by the addition of purified high molecular mass exopolysaccharide isolated from their ex0 parent Rm 1021. In the course of these experiments, we observed that we could also partially suppress the symbiotic deficiencies of exoH mutants, which fail to succinylate their EPS I, by the addition of a low molecular mass fraction of EPS I. This observation is consistent with the report of Leigh & Lee [53] that exoH mutants produce less low molecular mass EPS I than the wild type; perhaps the symbiotic defects of these mutants result from a failure to process the non-succinylated EPS I into an oligosaccharide signal molecule. Furthermore, this suppression of exoH mutants appeared to be somewhat more efficient than that of exoA mutants. This observation is consistent with the possibility that low and high molecular mass forms of the exopolysaccharide may have different symbiotic functions and suggests the additional possibility that there may be different structural requirements for these two types of roles. +

Volume 19

Other possible roles for exopolysaccharides include serving as a carrier for extracellular enzymes or signal molecules, forming part of the infection thread matrix, constraining descendants of attached bacteria to the immediate vicinity of the plant, or helping to evade or suppress plant defence responses [26]. With respect to this latter possibility. it is interesting that p 1,3-glucanases are among the major hydrolytic enzymes induced when plants are exposed to pathogens [54] and that each of the two R. meliloti exopolysaccharides has a single p 1-3 linkage in its backbone. Furthermore, this linkage is known to be modified in the case of EPS I1 and is likely to be modified in the case of EPS I. We thank the other members of the laboratory for their support and encouragement. This work was supported by Public Health Service Grant GM3 1030. J. W. K., J. G., and T. L. R. were supported by N. S. F. Predoctoral Fellowships and D. A. by a Howard Hughes I’redoctoral Fellowship. M. A. G. was supported by an American Cancer Society Postdoctoral Fellowship and A. M. by an N. S. F. Postdoctoral Fellowship in Plant Molecular HiolOgY. 1. Rauer, W. L). (1981) Annu. Rev. Plant Physiol. 32, 407-499 2. Halverson, I,. J. & Stacey, G. (1986) Microbiol. Rev. 50,193-225 3. Long, S. K. (1984) in Plant-Microbe Interactions, (Kosuge, T. & Nester, E., eds.), pp. 265-300, Macmillan, New York 4. Long, S. R. (1989) Annu. Rev. Genet. 23,483-506 5. Rolfe, B. G. & Gresshoff, P. M. (1988) Annu. Rev. Plant Physiol. Plant Mol. Biol. 39,297-3 19 6. Vance, C. P., Hoylan, K. I,. M., Stade, S.& Somers, D. A. (1985) Symbiosis 1,69-84 7. Verma, 1). P. S. & Long, S. (1983) Int. Rev. Cytol. 121,211-245 8. Downie. J. A. & Johnston, A. W. B. (1986) Cell (Cambridge, Mass.) 47, 153-154 9. Lerouge, P., Koche. P., Faucher, C., Maillet, F., Truchet, G., Prome. J. C. & Denarie, J. (1990) Nature (1,ondon) 344,78 1-784 10. Sutherland, I. W. (1979) Microbial Polysaccharides and Polysaccharases, Academic Press, London 11. Glazebrook, J. & Walker, G. C. (1991) Methods in Enzymol., in the press 12. Gravanis, G., Milas, M., Kinaudo, M. & Tinland, H. (1987) Carbohydr. Kes. 160,259-265 13. Aman, P., McNeil, M., Franzen, I,.-& Darvill, A. G. & Albersheim, P. (1981) Carbohydr. Res. 95,263-282 14. Hisamatsu, M., Abe., J., Amemura, A. & Harada, T. (1978) Carbohydr. Kes. 66,289-294 15. Zevenhuizen, L. T. P. M. (1973) Carbohydr. Kes. 26, 409-4 19


10. Hisamatsu, M.. Sano, K.. Amemura, A. & Harada, T. (1978) Carbohydr. Kes. 61,89-96 17. Williams, A. G. & Lawson, C. J. (1979) Hr. Patent 1539064 18. Hjorndal, H., Erbing, C., 1,indberg. H.. Fahraeus, G. & Ljunggren, H. (1971) Acta Chem. Scand. 25, 1281- 1296 10. Hisamatsu. M., Abe, J.. Amemura. A. & Harada. T. ( 1 980) Agric. Hiol. Chem. 44, 1049- 1055 20. Jansson. l’.-E., Kenne, I,.. Lindberg, G., Ljunggren, H.. Lonngren, J., Kuden. U. & Svensson, S. (1977) J. Am. Chem. SOC.99.1 1 - 14 2 1. Zevenhuizen. I,. 1’. T. M. & VanNeervan, A. K. (1983) Carbohydr. Res. 118. 127-134 22. Tolmasky, M. E., Staneloni, R. J., llgalde, K. A. & I,eloir7 I,. F. (1980) Arch. Biochem. Biophys. 203, 358-364 23. Tolmasky, M. E., Staneloni, R. J. & Leloir, I,. F. (1982)J. Hiol. Chem. 257,6751-6757 24. Hetlach. M. K., Capage, M. A.. Doherty. D. H., Hassler, K. A,, Henderson, N. M., Vanderslice, K. W., Marrelli, J. D. & Ward, M. B. (1987) in Industrial I’olysaccharides: Genetic Engineering. Structure/ I’roperty Kelations and Applications (Yalpani, M., ed.), pp. 35-50, Elsevier Science Publishers H. V., Amsterdam 25. Linton, J. D., Evans, M., Jones, 1). S. & Gouldney, 1). N. ( 1987) J. Gen. Microbiol. 133, 296 1-2969 26. Glazebrook, J. & Walker, G. C. (1989) Cell (Cambridge, Mass) 56,661-672 27. Finan, T. M., Hirsch, A. M., Leigh, J. A,, Johanson, E.. Kuldau, G. A,. Deegan, S., Walker. G. C. & Signer, E. K. (1985) Cell (Cambridge, Mass) 40,869-877 28. Leigh, J. A., Signer, E. K. & Walker, G. C. (1985) I’roc. Natl. Acad. Sci. U.S.A. 82,623 1-6235 29. Leigh, J. A., Reed, J. W., Hanks, J. F.,Hirsch, A. M. & Walker, G. C. (1987) Cell (Cambridge, Mass) 51, 579-587 30. 1,ang-Unnasch, N. & Ausubel, F. M. (1985) Plant I’hysiol. 77, 833-839 31. Norris, J. H., Marcol, I,. A. & Hirsch, A. M. (1988) I’lant I’hysiol. 88. 32 1-328 32. Dickstein, K., Bisseling, T., Keinhold, V. N. & Ausubel, F.M. (1988) Genes Dev. 2,677-687 33. Chakravorty, A. K., Zurkowsky, W.. Shine, J. & Kolfe, B. G. (1982) J. Mol. Appl. Genet. 1, 585-596 34. Chen, H., Batley, M., Redmond, J. & Rolfe, B. G. (1985) J. Plant Physiol. 120, 331-349

35. Horthakur, D., Barber, C. E., Lamb, J. W., Daniels, M. J.. Downie, J. A. &Johnston, A. W. H. (1986) Mol. Gen. Genet. 203,320-323 36. Finan, T. M., Kunkel, H., L)e Vos, G. F. & Signer, E. K. (1986) J. Hacteriol. 167,66-72 37. Long. S.,Reed, J. W., Himawan, J. W. & Walker, G. C. (1988)J. Hacteriol. 170,4239-4248 38. Kueber. T. I,.. Long, S. & Walker. G. C. (1991) J. Bacteriol. 173, 426-434 39. Manoil, C. & Beckwith, J. (1985) I’roc. Natl. Acad. Sci. USA. 82,8129-8133 40. Long. S., McCune, S. & Walker, G. C. (1988) J. Hacteriol. 170,4257-4265 41. Zhan, H., Gray, J. X., Levery, S. H., Rolfe, H. G. & Leigh, J. A. (1990) J. Hacteriol. 172, 5245-5253 42. Reed, J. W., Capage, M. & Walker. G. C. (1991) J. Hacteriol. 173,3776-3788 43. Horthakur, I). & Johnston, A. W. H. (1987) Mol. Gen. Genet. 207,149-1 54 44. Horthakur, D., Downie, J. A,. Johnston. A. W. H. & Lamb, J. W. (1985) Mol. Gen. Genet. 200,278-282 45. Zhan, H. & Leigh, J. A. (1990) J. Bacteriol. 172, 5254-5259 46. Sutherland, I. W. ( 1 982) Adv. Microbiol. I’hysiol. 23, 79- 1SO 47. Doherty, L)., Leigh, J. A,, Glazebrook. J. & Walker, G. C. (1988)J. Hacteriol. 170,4249-4256 48. Her, G. K., Glazebrook, J., Walker, G. C. & Heinhold, V. N. (1090) Carbohydr. Hes., 198. 305-312 49. Zhan. H.. Levery, S. €3.. Lee, C. C. & Leigh, J. A. (1989) I’roc. Natl. Acad. Sci. U.S.A. 86, 3055-3059 SO. Darvill. A. G. Albersheim, 1’. A,. Bucheli. I’., Doares, S.,Doubrava, N., Eberhard, S.. Collin, D. J., Hahn. M. G., Marfa-Kiera, V., York, W. S. & Mohnen, L). (1989) in NATO AS1 Series H: Signal Molecules in Plants and Plant-Microbe Interactions (Lutgenberg. H. J. J., ed.), vol. H36, pp. 41-48, Springer-Verlag, Berlin, Heidelberg 51. Djordjevic, S.,Chen, H., Hatley. M., Redmond, J. W. & Rolfe, H. G. (1987)J. Hacteriol. 169, 53-60 52. Muller. I’., Hynes. M., Kapp, L)., Niehaus, K. & I’uhler. A. (1988) Mol. Gen. Genet. 21 1, 17-26 53. Leigh, J. A. & I,ee, C. C. (1988) J. Bacteriol. 170, 3327-3332 54. Kombrink. E., Schroeder, M. & Hahlbrock, K. (1988) I’roc. Natl. Acad. Sci. U.S.A. 85, 782-786 Received 9 April 190 1

1991

64 I