Cell, Vol. 71, 191-199,

October

16, 1992, Copyright

0 1992 by Cell Press

Molecutar Signals in the Interactions between Plants and Microbes Howard R. G. Clarke,’ John A. Leigh,” and Carl J. Douglast *Department of Microbiology University of Washington Seattle, Washington 98195 tDepartment of Botany University of British Columbia Vancouver, British Columbia V8T 124 Canada

Plants coexist with a great number of microorganisms. Plant-microbe interactions can be beneficial, as in symbiotic Rhizobium-legume and mycorrhizal-root associations, but in other cases, microbes represent potential pathogens with the ability to parasitize plants and cause disease. Early in the interactions between plants and microbes, signals are produced that elicit discrete responses in the respective partners, the first steps in the cascade that determines the outcomeof the symbiosisor pathogenesis or the expression of disease resistance. The sixth International Symposium on Molecular Plant-Microbe Interactions (University of Washington, July 11-16, 1992) brought together a diverse group of investigators focusing on the genetics, physiology, and molecular biology of bacteria, fungi, and plants as they relate to plant-microbe interactions. Like other organisms, signal generation, perception, and transduction are important in plant growth and development. The sessile nature of plants requires that they perceive and respond to many environmental signals as well as to endogenous signals. In contrast with some other prokaryotic and eukaryotic systems, however, the molecular basis of signal perception and transduction and the nature of some of the signaling molecules are just beginning to be elucidated (reviewed by Becraft and Freeling, 1992). Genetic and molecular approaches are being used to identify genes involved in response to growth and developmental regulators, as well as to external stimuli such as light, gravity, wounding, touch, water stress, and interactions with microbes. In most cases, however, the receptors and signaling mechanisms mediating these responses are unknown or ill defined. Signal generation, perception, and transduction in many of the microbes that interact with plants are understood in greater detail and parallel findings from other prokaryotic and eukaryotic microbes. Plant-microbe interactions offer opportunities for the elucidation of basic mechanisms governing signal perception and transduction in plants, as well as of mechanisms by which microbes respond to plant signals. This review will highlight recent discoveries regarding the molecular interactions between plants and microorganisms, focusing on signaling events that are crucial in establishing both beneficial and detrimental interactions.

Meeting Review

The Agrobacterium-Plant

Interaction

Agrobacterium tumefaciens, the causative agent of crown gall disease in dicotyledonous plants, transfers a copy of a DNA segment of its own tumor-inducing (Ti) plasmid into plant cells, wherein the DNA moves into the nucleus and integrates into host chromosomal DNA (reviewed by Winans, 1992; Hooykaas and Schilperoort, 1992). The transferred DNA (T-DNA) encodes enzymes that direct the synthesis of the plant growth regulators auxin and cytokinin, leading to the disorganized, neoplastic cell division characteristic of the disease. T-DNA genes also direct the production of opines, such as octopine and nopaline, which can be utilized by Agrobacterium as carbon and nitrogen sources. T-DNA transfer to the plant nucleus requires bacterial virulence (vir) gene induction, T-complex formation, its export out of Agrobacterium and import into plant cells, and targeting to the plant nucleus. These processes are carried out by the products of several Ti plasmid-encoded vir operons (reviewed in Winans, 1992). vir gene expression is induced by plant-derived phenolic compounds, such as acetosyringone, which are released from wounded plant tissues. The induction of vir gene expression requires the products of two vir genes, virA and viffi, which encode members of a large family of prokaryotic twocomponent regulatory systems that respond to a variety of environmental signals (reviewed in Winans, 1991). VirAand VirG-mediated signaling is relatively well understood and offers insights into mechanismsof interspecific signaling (Figure 1). Signal Perception by VirA VirA is an inner membrane protein belonging to a family of histidine protein kinases (Melchers et al., 1989; Winans et al., 1989) and is thought to be the sensor protein that responds to the presence of plant-derived phenolic compounds. The VirA kinase (Figure 1) undergoes autophosphorylation in the cytoplasmic kinase domain (Jin et al., 1990a; Huang et al., 1990) which is essential for both vir gene induction in vitro and tumorigenesis (Jin et al., 199Oa), and transfers a phosphate group to a conserved aspartate residue in the N-terminal receiver domain of the VirG protein (Jin et al., 199Ob). VirG is a member of the response regulator family of proteins that activate gene transcription (Figure 1); activation of vir gene transcription by phosphorylated VirG proteins is thought to occur due to enhanced binding of the protein to regulatory sequences associated with vir genes or through enhanced interaction with RNA polymerase subsequent to binding (see Winans, 1992). How does VirA sense plant phenolic molecules? Deletion of most of the periplasmic domain does not abolish acetosyringone-inducible vir gene expression, arguing against a role for this domain in phenolic signal perception (Melchers et al., 1989). A. Binns (University of Pennsylva-

Cell 192

WOUND-RELEASED PLANT SUGARS AND PHENOLICS e

00

I+@

00 I

p4mdic

ChvE

biting protein’,

VirA

PERIPIASM bansmembrane 2

tmnanembram 1

VirG

inactive

Et&

CYTOPLASM

&&,,.

DNA-binding I t

v A

9 B

G

v c

v

t

D

tE

Virulence genes

Figure Model

1. Structure of Agrobacterium VirA and VirG Proteins for Plant Signal-Induced Expression of vir Genes

and a

The functional domains of VirA and VirG and their cellular locations are indicated. The model depicts the interaction of plant phenolics and sugars with VirA and ChvE, respectively, and the ensuing cascade of events leading to vir gene expression. See text for details,

nia) described the use of acetosyringone analogs (e.g., a-bromoacetosyringone) that are specific, irreversible inhibitors of vir gene induction to address the question of whether acetosyringone interacts directly or indirectly with the VirA receptor (Hess et al., 1991). Such analogs can be used as affinity labels for identifying phenolic-binding proteins. Incubation of intact Agrobacterium cells or cell extracts with the radiolabeled analog a-bromod’-iodoacetovanillone did not appear to label VirA. Instead, specific binding to two proteins 10 kd and 21 kd in size was observed. These data are inconsistent with VirA as a phenolit-binding protein and argue for the involvement of other phenolic-binding proteins in transducing plant signals to the t&AA&G two-component regulatory system. There are precedents in other two-component systems for an indirect interaction of the environmental signal with the sensor protein. Acetosyringone-mediated induction of vir gene expression is markedly enhanced in the presence of monosaccharides, such as glucose and arabinose, as well as noncatabolized substituted sugars such as e-deoxy-oglucose (Shimoda et al., 1990). The chromosomal virulence gene chvE, which shares sequence identity with a glucose/galactose-binding protein from A. radiobacter, is proposed to potentiate sugar-responsive vir gene induction (Cangelosi et al., 1990). Mutants of chvE show an attenuatedvirulence and nosugar enhancement. This mutant phenotype has led to speculation that the ChvE-sugar complex interacts directly with the periplasmic domain of VirA, resulting in increased &induction in the presence of

phenolic inducers. Y. Machida(Nagoya University, Japan) and A. Binns showed a loss of sugar enhancement in virA point mutants mapping to the periplasmic domain near transmembrane domain 2 (Figure l), and Machida further showed that point mutations in chvE suppressed these sugar-nonresponsive phenotypes. These data support the hypothesis that a ChvE-sugar complex interacts directly with VirA. A model depicting these interactions and the ensuing cascade of events is outlined in Figure 1. Agrobacterium does not normally infect monocotyledonous plants such as maize, and the inefficiency of A. tumefaciens-induced tumorigenesis on maize has been proposed to involve the virA locus. D. Raineri (University of Washington) has used agroinfection (Grimsley et al., 1997)-a technique in which maize streak virus DNA is inserted between T-DNA borders to enable detection of DNA transfer from Agrobacterium to plant-to show that VirA plays a role in determining differences in the ability of two closely related strains of A. tumefaciens to transfer T-DNA to maize. Octopineproducing Agrobacterium strains do not normally agroinfect maize, whereas nopaline-producing strains do. Infectivity was conferred upon the noninfective octopine strain by transformation with the nopaline virA locus, suggesting the direct involvement of VirA in potentiating host range limitations, possibly by controlling the ability to perceive plant signals. T-Complex Targeting In Agrobacterium, the T-DNA is tightly associated with VirD2, part of the endonuclease complex that nicks the T-DNA borders. VirEP, a single-stranded DNA-binding protein, is thought to coat the DNA. This T-complex is transported from the bacterium into the plant cell and then moves to the nucleus wherein the T-DNA is inserted into the plant genome. A major question of interest concerns the T-complex-specific signals that mediate transfer from the bacterium to the plant nucleus. The discovery of nuclear localization signals (NLS) in VirD2 (Howard et al., 1992, and references therein), similar to those in animals that are responsible for receptor-mediated targeting to nuclei, has led to intense study of their function in T-DNA targeting. B. Hohn (Friedrich-Miescher Institut, Basel) described work demonstrating that the bipartite NLS of VirD2, which is capable of targeting a heterologous protein to the plant nucleus, is required for T-DNA transfer from bacterium to plant, suggesting the VirD2 NLS is involved in the T-DNA transport process. The VirD2 NLS is also functional in yeast, indicating functional conservation of nuclear targeting mechanisms between yeast and plants. P. Zambryski (University of California, Berkeley) reported on the discovery of a bipartite NLS in VirE2 (Citovsky et al., 1992). VirE2 is proposed to play a significant role in nuclear targeting of the T-complex. Opines as Signal Molecules Opines such as octopine and nopaline have traditionally been viewed as sources of bacterial nutrition in crown gall tumors. S. Winans (Cornell University) presented data demonstrating a new role for octopine as a signal for the transcriptional activation of an octopine catabolic operon

f$yting

Review

on the Ti plasmid. Expression of the operon is induced by octopine and is under the control of the OccR protein, a LysR-type transcriptional regulator. Octopine can directly affect the activity of OccR in vitro: octopine alters the DNA binding specificity of OccR and relaxes DNA bending in the OccR-DNA complex (Wang et al., 1992) events that correlate with octopine-activated transcription. Signaling by this and possibly other opines completes a circle of signal exchange in the Agrobacterium-plant interaction: transfer of the bacterial opine biosynthetic gene to plant cells via the T-DNA is dependent on plant-derived signals; the opine in turn acts as a signal regulating catabolic activity of the parasitizing bacteria.

9-l ,Cltnked N-acetyl glucosamine bearing an N-acyl substitution (often 16 or 16 carbons) at the nonreducing end and variably acetylated at carbon 6 of the nonreducing end (Figure2; Lerougeet al., 1991; Truchet et al., 1991; Spaink et al., 1991). Modifications at carbon 6 of the reducing terminus and in the structure of the acyl substituent produce host-specific variants. Thus, J. Denarie (Laboratoire de Biologie Moleculaire des Relations Plantes-Microorganismes, Centre National de la Recherche Scientifique [CNRS)-lnstitut National de la Recherche Agronomique [INRA], Toulouse) reported that, depending on the species, the reducing terminus was unmodified or modified with sulfate, fucose, or methyl fucose. The length and pattern of unsaturation of the acyl substituent varied similarly. Given that at least some of these variations have been shown to be critical in host-specific induction of nodule primordia (Roche et al., 1991; Spaink et al., 1991) this structural variation in the Nod factors probably accounts for host specificity. Since the Nod factors act at nanomolar concentrations, plant receptors are probably required to amplify the signal, although such receptors have not been directly detected. Genetic complementation experiments between species or biovars of Rhizobium have placed the nod genes into two classes: common nod genes, which function equally in nearly all Rhizobia for nodulation of the host of that organism, and host-specific nod genes, which function only for nodulation of the host of a particular species or biovar (reviewed by Fisher and Long, 1992). These nod gene-dependent plant responses to bacterial inoculation are now being correlated to actual functions of the nod gene products in synthesizing specific Nod factors. For example, S. Long (Stanford University)described how Rhizobium meliloti NodP and NodQ function in activating sulfate for addition to the Nod factor (Schwedock and Long, 1990) and that they require GTP for activity, similarly to Escherichia coli CysD, CysN, and CysC. NodH was shown to add sulfate to the Nod factor in vitro (Long). A novel function for a nod gene product was described by A. Downie (John lnnes Institute, Norwich, England)the Rhizobium leguminosarum Nod0 protein is a secreted protein that can form calcium-regulated ion channels in an artificial membrane, similar to hemolysin-like toxins of certain mammalian pathogens. This observation suggests a model in which localization of a Rhizobium protein in the plant membrane may enhance the response of the plant to Rhizobium signals, perhaps by enhancing ion fluxes (see below). Other issues under analysis concern the effects of the

Rhizobium-Legume Interactions: Developmental Mutualism The Rhizobium-legume interaction stands out from other plant-microbe interactions discussed at the meeting as one in which a true developmental mutualism occurs (reviewed by Fisher and Long, 1992). In the successful interaction, rather than fight off the invader, the plant grows an organ, the root nodule, to house Rhizobium cells and develops a structure, the infection thread, to facilitate invasion. These events are controlled by specific molecular signals. The initial give-and-take occurs when the bacterium responds to the presence of specific plant-secreted substances by secreting Nod factors, the morphogens that induce the formation of nodule primordia in the root cortex as well as root hair deformations at the root epidermis. Nod factor synthesis is specified by the bacterial nod genes, which are induced by plant flavonoids via NodD proteins, members of the LysR family of transcriptional activators. Questions that are of current importance include the following: Do structural variations in the Nod factors account for host specificity, i.e., the fact that each species or biovar of Rhizobium (or the related genera Bradyrhizobium and Azorhizobium) interacts only with a particular set of host plants? How do the Rhizobium nod gene products function in the synthesis of these specific structures? What is the nature of the plant cellular response to these factors? What signals seem to function in subsequent nodule development, nodule colonization, and the onset of nitrogen fixation? Nod Factors as Host-Specific Morphogens The structures of Nod factors from various Rhizobium and Bradyrhizobium strains reported at the meeting were variants on a theme: modified tetramers and pentamers of

w2 3-w NH o-c’ 1n=1.2.cf3 \ R =sunale0,Other \

Figure 2. Generalized Nod Factors

7

Host-specific

Structure

modifications

occur

of Rhizobium

at the two R

Nod factors on the plant, ranging from possible signaling cascades to the ultimate morphological effects. bong reported that the R. meliloti Nod factor induced plant cell membrane depolarization (Ehrhardt et al., 1992). In addition, since nodule primordia form before bacteria penetrate the root, a second messenger may act to transduce the nodule formation signal to the interior of the root. Auxins and cytokinins are obvious candidates, and A. Kondorosi (Institut des Sciences Vegetales, CNRS, Girsur-Yvette, France) reported a correlation between auxin sensitivities of alfalfa cultivars and their tendencies to form nodules. These results corroborate a report (A. Hirsch, University of California, Los Angeles) that auxin transport inhibitors induce nodule-like structures on alfalfa (Hirsch et al., 1989). A picture emerges in which Nod factors induce a signaling cascade that may involve at some level membrane depolarization and possibly local changes in plant hormone levels. Several workers are using molecular indicators of plant gene expression as in situ markers of early nodulation events, including those promoted by Nod factors (Scheres et al., 1990). T. Bisseling (Agricultural University, Wageningen, The Netherlands) showed that some early pea nodulins (nodule-specific plant genes) are expressed in nodule primordia prior to infection thread penetration, while others are expressed only in cells that have been penetrated. D. Barker (CNRS-INRA, Toulouse) described the use of GUS reporter gene fusions to early nodulin genes in alfalfa to detect expression in differentiating epidermal cells (root hairs) and in nodule primordia. But what morphological responses do Nod factors promote? In alfalfa, they elicit the deformation of root hairs and the formation of genuine nodules (Truchet et al., 1991). In vetch, the Nod factor seems to initiate preinfection thread structures as well as nodule primordia: in the outer root cortex, cytoplasmic bridges resembling those preceding infection threads were observed, and at the same time in the inner cortex cell divisions were evident (H. Spaink, Leiden University, The Netherlands; van Brussel et al., 1992). These observations, together with the finding that some lines of alfalfa generate nodules spontaneously (Truchet et al., 1989) suggest that Nod factors trigger events that are preprogrammed in the plant to form nodules and to direct the bacterial infection. Post-Nod Factor Signaling and Evasion of Defense Response One widely posed question is how can the infection occur without succumbing to plant defenses as would often occur with pathogenic microorganisms? G. Truchet (CNRSINRA, Toulouse) showed that, in fact, a portion of initiated infections do abort with symptoms of a protective hypersensitive response (localized necrosis and accumulation of phenolic compounds). What occurs differently in the successful infections remains to be determined. Intercellular signaling between Rhizobium and the legume does not stop after the Nod factors. What signals function subsequently? Specific exopolysaccharides of Rhizobium were reported to function in promoting nodule invasion in alfalfa (J. Leigh; Battisti et al., 1992; Urzainqui

and Walker, 1992). N. Brewin (John lnnes Institute) showed that in pea, lipopolysaccharides may play similar roles, as well as functioning in the avoidance of a plant defense response. Once inside the nodule, Rhizobium cells respond to markedly decreased oxygen concentrations to express the nitrogen fixation @if) genes. J. Batut (CNRS-INRA, Toulouse) and A. Lois (University of California, San Diego) reported the detailed characterization of a regulatory cascade involving the oxygen-sensing heme protein FixL (Gilles-Gonzalez et al., 1991) and the transcriptional response regulator FixJ, members of the twocomponent regulatory system family of proteins (David et al., 1988). It was apparent that the study of the Rhizobium-legume interaction is entering an exciting phase in which the detailed structures and functions of intercellular signals are being defined. In addition, several questions were just being hinted at. What is the nature of the presumed plant receptors or families of receptors that mediate responses to Nod factors? Do plants have endogenous signal molecules that resemble Nod factors? Do plant hormones participate directly in the nodule development process? What other signals function in nodulation? And, finally, what is the relationship between nodulation and plant defense processes? Regulation

of Pathogenic@

in Ustilago

Ustilago maydis incites galls on corn plants and forms diploid spores capable of undergoing meiosis within the infected tissue. Phytopathogenic smut fungi such as U. maydis are dimorphic, having the ability to switch between a yeast-like nonpathogenic haploid phase and a filamentous pathogenic dikaryotic phase (reviewed by Banuett, 1992). The regulation of the dimorphic switch and the genetics of infectious dikaryon formation were the subjects of presentations by I. Herskowitz (University of California, San Francisco), J. Kronstad (University of British Columbia, Vancouver), and M. Balker (Institut fijr Genbiologische Forschung, Berlin). In U. maydis, mating and dikaryon formation is regulated by two mating type genes, a and b. The a locus is proposed to control cell-cell signaling in mating by producing mating pheromones and their receptors (Balker et al., 1992). Maintenance of filamentous growth also appears to require an autocrine response involving a-specific pheromones and their receptors (Biilker; Banuett, 1992). Interestingly, there was some indication that expression of the mfal gene, which encodes an a-specific pheromone, is nutritionally regulated, raising the possibility that plant-fungus signaling is involved in regulating expression of pheromone genes. The b region has at least 25 different alleles and seems to control events required for filamentous growth and pathogenicity; two homeodomain proteins (bE and bW) are encoded at each b allele (Gillissen et al., 1992). The combination of the bE product from one b allele and bW from another is thought to form a novel heterodimeric regulatory protein that triggers filamentous growth and pathogenicity (Gillissen et al., 1992; Kronstad; Herskowitz). Target genes for b regulatory proteins for filamentous growth

Meeting 195

Review

and pathogenicity have not yet been cloned, but candidate target genes have been identified as mutants defective in filamentous growth (Banuett, 1991; Herskowitz; Kronstad). What is the nature of the pathogenicity factors produced by the filamentous form? Since normal plant development is obviously perturbed in gall formation, it was suggested that one of the genes identified genetically (fo may be involved in regulating auxin production in the plant (Herskowitz). Pathogenic

Interactions:

Recognizing

the Bad Guys

Interspecific signaling and intracellular signaling are also important in establishing the ability of pathogenic microbes successfully to parasitize some plants and in the ability of other plants to resist infection. However, the nature of the signals and mechanisms of signal perception are generally less well understood. Many questions remain unanswered. For example, how do pathogenicity genes function, and what regulates their expression? How do single genes, defined by classical genetics in both plants and pathogens, specify plant-pathogen recognition? What pathogen-induced signals are generated within plant cells, and how do they initiate resistance responses? Bacterial Pathogenicity and Hypersensitive Reaction and Pathogenicity (hrp) Genes Bacterial plant pathogens have evolved a number of mechanisms to perturb plant metabolism and cause disease. These mechanisms include the alteration of plant growth regulator balance and the production of toxins, plant cell wall-degrading enzymes, and extracellular polysaccharides. More enigmatic are the products of hypersensitive reaction and pathogenic@ (hrp) genes, first identified in Pseudomonas syringae pv. phaseolicola and subsequently in a large number of other plant pathogenic bacteria (reviewed by Willis et al., 1991). As the name suggests, the products of these genes are required for pathogenicity on (susceptible) host plants and for the induction of a localized resistance, or hypersensitive reaction (HR), on resistant host and nonhost plants. A likely scenario for hrp gene action starts with transcriptional activation by plant-derived signals. Expression of hrp genes can be environmentally regulated by such factors as osmoticum, nutrient source, and pH and can be induced in planta or by plant culture filtrates (U. Bonas, Institute fur Genbiologische Forschung, Berlin; N. Panopoulos, University of California, Berkeley). Some hrp genes are clearly regulatory, responding to plant signals to induce transcription of hrp operons in a manner that may be analogous to the two-component signaling systems. For example, the hrpsgeneof P. syringae pv. phaseolicola is an ntrC-like regulator that controls hrpL expression, which is in turn required for the expression of several other hrp operons (Panopoulos). Expression of all hrpcontrolled operons also requires a functional n&A gene for full induction. The nature of the plant-derived molecules that induce hrp gene transcription are unknown. What do the remaining hrp genes do? An intriguing hypothesis is that hrp genes are involved in the synthesis

and export of pathogenicity factors that also induce an HR on nonhost plants. Support for this idea comes from the results of database searches (C. Boucher, CNRS-INRA, Toulouse; Bonas) that revealed significant similarities of several hrp genes to pathogenicity genes in the mammalian pathogens Yersinia and Shigella. The Yersinia and Shigella genes encode membrane proteins involved in the secretion of Yersinia outer membrane proteins, or yops, important virulence factors (Cornelius et al., 1989). Evidence was presented by Boucher suggesting that Pseudomonas solanacearum hrp genes function in the export and synthesis of an HR-inducing factor. This factor, which can induce an HR on the nonhost tobacco, was detected in bacterial filtrates and is sensitive to proteinase K. Significantly, the factor appeared to be present in cell lysates of most hrp- strains, suggesting a block in secretion. Strains without a functional hrpB gene (of the araC family of transcriptional regulators), however, were unable to produce the factor intracellularly, suggesting that production of the factor is regulated by a component of the cluster. S. Beer (Cornell University) presented evidence for a hrp-encoded proteinaceous HR elicitor in Erwinia amylovora (Wei et al., 1992) but it does not appear to be exported from bacterial cells. Recognition and Plant Resistance Specific recognition between plants and potential pathogens often determines the ultimate result of the interaction (reviewed in Knogge, 1991; Keen, 1998). Genes that are likely to be involved in recognition signal production by pathogens and signal perception by host plants have long been known to plant pathologists. Pioneering genetic analyses of the interaction between flax and the flax rust pathogen Melampsora lini by Flor (1955) demonstrated that recognition functions are encoded by dominant alleles of plant resistance(R) genes that interact genetically with the products of single dominant avirulence (avr) genes encoded by the pathogen. According to this gene-for-gene hypothesis for plant cultivar-specific resistance, each R gene is specific for a corresponding avr gene (Figure 3). A widely

0

Pathogen genotype:

Host genotype:

avrz

+

RI r1

r1 r1

t Recognltlon

Host response: Figure 3. Genetic Plant R Genes

4

t

t Reslstent (incompatlble) Interactions

t No rocognltlon

t Susceptible (compatible) between

Pathogen

%‘l t No mognltlon t Susceptible (compatible) avr

Genes

and

Resistant, or incompatible, reactions occur when a pathogen carrying an aw gene is recognized by a corresponding R gene allele in the plant. When there is no match between an R gene and a pathogen avf gene, the pathogen escapes recognition and acompatible, or susceptible, reaction results.

Cdl 196

accepted model, based solely on the genetics of the interaction, suggests that R gene products act as receptors for the products of avr genes, and, given interaction with the proper avr-specified ligand, trigger intracellular signaling events culminating in a resistance response (Keen, 1990). Resistance reactions can also be triggered by nonhost pathogens(those without the basic ability to cause disease on a given plant species) in a manner not governed by the specific R and avr genes discussed above. These responses are thought to be indicative of the basic resistance that plants have against most pathogens (Knogge, 1991). Signal molecules, or elicitors, which activate defense responses independently of R genes, generally function nonspecifically (i.e., trigger responses in many plants) and can be of pathogen or host plant origin. avr Gene Structure and Function Shotgun cloning strategies have been used to isolate avr genes from several bacterial pathogens (for example, Bonas et al., 1989). Recent work on the avrBs3 gene from Xanthomonas campestris pv. vesicatoria (which causes bacterial spot disease on pepper and tomato and interacts genetically with the pepper Bs3 R gene) was described by U. Bonas. The structure of the gene is remarkable: within the 3.5 kb open reading frame of the cloned gene are 17.5 repeats of an almost identical 34 amino acid motif (Herbers et al., 1992). While the function of the encoded polypeptide is unknown, it is soluble and is constitutively expressed. It was speculated that the product may interact directly with a plant receptor, presumably encoded by the 6~3 gene. Insight into the interaction with R genes was gained by generating avrBs3 deletion derivatives in which the number and position of repeats was altered and testing these derivatives for altered recognition on pepper and tomato plants. Removal of repeats 11-14 abolished the incompatible interaction with Bs3-carrying plants: a susceptible reaction was observed. Surprisingly, however, this deleted avr gene did not lose its function entirely but was apparently converted into an avr gene with a new specificity. Instead of a susceptible interaction with a plant homozygous for the recessive bs3 allele (unable to recognize the avrBs3 gene), a resistance reaction was induced. Thus, by altering the structure of the avrBs3 gene product, a new plant resistance specificity was revealed, although there is no proof yet that the new avf allele interacts with Bs3. These results suggest that there may be many more undiscovered R gene specificities in plants. avrgenes with structures very similar to avrBs3 have been cloned from other Xanthomonas species and pathovars (J. Leach, Kansas State University; Y. Yang and D. Gabriel, University of Florida). The finding that avr genes of conserved structure are recognized by R genes from several different plants raises the possibility of conserved R gene structure and function in these plants. An avr gene, awl), has also been cloned from P. syringae pv. tomato (Keen, 1990). The avrD locus is an operon with at least five open reading frames and appears to encode proteins with enzymatic functions (N. Keen, University of California, Riverside). Cverexpression of avrD in P. syringae pv. glycinea or other gram negative bacteria

resulted in the secretion of two related hydroxylipids that act as elicitors specifically recognized by soybean plants carrying the Rpg4 R gene. In contrast with avrBs3, avrD expression may be plant regulated, suggesting potential exchange of signals between host and pathogen: plant factors could activate avrD expression, leading to the biosynthesis of discrete hydroxylipid elicitors that interact with the Rpg4 gene product and trigger an HR. Is the Rpg4 R gene product a receptor for avrD-specified elicitors? The answer awaits cloning of the R gene or the receptor for the avrD elicitor. Recent progress has also been made in cloning fungal avr genes (van Kan et al., 1991; Van den Ackervecken et al., 1992). P. de Wit (Agricultural University, Wageningen, The Netherlands) provided genetic evidence that the product of the tomato fungal pathogen Cladosporium fulvum avr9 gene is a peptide elicitor that interacts genetically with the tomato Rgene, cf9, to induce an HR. Interestingly, the avr9 gene is activated specifically when the fungus penetrates plant tissue, as was demonstrated by expression of an avr9 promoter-GUS gene fusion in the fungus during infection. Whether the avr9 elicitor is a ligand that interacts directly with the Cf9 gene product, the nature of the plant signals that induce avr9 expression and the function of the avr9 product in the fungus remain open questions. It is apparent, however, that avr9 is solely responsible for incompatibility (resistance) on Cf9 plants. This was proven by disruption of avr9 in avirulent races, which then became virulent, and by transforming virulent races with avr9, which then became avirulent on Cf9 plants. B. Valent (Du Pont) described the cloning of avr genes from the rice blast fungus Magnaporthe grisea. Two such genes, Pwll and Avr2-YAMO, are genetically instable: there is a high rate of reversion to virulence in strains containing these avrgenes. The genes have been cloned, and this should aid in determination of their biochemical functions in recognition and the reasons for their instability.

Progress in Cloning R Genes A major impediment to understanding recognition and signaling in pathogenic plant-microbe interactions is the dearth of information regarding R gene function. Since R genes are defined only by the resistance phenotype they specify, strategies for R gene cloning must be based on genetic analyses. Two major approaches have been the use of transposable element mutagenesis and map-based cloning. Groups using both techniques are tantalizingly close to having Rgenes in hand. Using traditional methods of transposon tagging with the Ac/Ds element system in maize, T. Pryor (Commonwealth Scientific and Industrial Research Organisation, Canberra) reported potential insertional inactivation of the Rp5 R gene for resistance to Puccinia sorghi. The active transposition systems of some lines of Antirrhinium majus have been useful in cloning interesting developmental genes. Introduction of an Antirrhinium R gene specific to Puccinia antirrhini into an active line resulted in the isolation of unstable susceptible lines with a potentially tagged R gene (J. Newbury, University

Meeting 197

Review

of Birmingham). Further progress will depend on identification of the putative transposon that is the mutagen in these lines. The utility of the transposon tagging approach for isolating resistance genes in maize was illustrated by the successful cloning of the maize Hml gene, which confers resistance to the cyclic tetrapeptide HC toxin produced by the maize fungal pathogen Cochliobolus carbonum. A functional Hml gene is correlated with HC toxin reductase (HCTR) activity, which detoxifies the HC toxin (R. Meeley and J. Walton, Michigan State University). The work of G. Johal and S. Briggs (Pioneer Hi-Bred International, Johnston, Iowa) was described in which tagging of the /-/ml gene by Mu and Spm transposable elements enabled the gene to be cloned. No formal proof yet exists that Hml encodes HCTR, but Hm7 is homologous to the maize A 7 gene, which encodes dihydroflavanol reductase. Meeley showed that HCTR activity and Hm7 mRNA accumulation appear to be inducible by the fungal infection, suggesting that pathogen-plant signaling is involved in regulating the defense response. Progress in map-based cloning of the tomato Pro R gene, which specifies resistance to P. syringae pv. tomato, was described by G. Martin (Cornell University). A YAC clone containing RFLP markers flanking the gene was identified, and the cDNA clone of a gene encoded by the YAC is a candidate for the Pro gene: the cDNA is highly polymorphic, yet no recombinants between cDNA-detected polymorphisms and Pro were observed in a population of 1600 F2 progeny. Expression of this and other candidate cDNAs in transgenic tomato lines should reveal in the near future whether an R gene has actually been cloned. Arabidopsis is providing fertile ground for the identification of R genes. This is significant because the small genome size, short generation time, genetics, and transformability of this plant make map-based cloning methods particularly practical. Arabidopsis R genes, which interact in a gene-for-gene manner with P. syringae avr genes, have been identified (Debener et al., 1991; Whalen et al., 1991). J. Dangl (Max Delbrijck Laboratorium, Kbln) described progress towards the map-based cloning of the RPM7 R gene, which interacts specifically with the P. syringae pv. maculicola avrRpm7 gene to give a typical resistance reaction. The RPM7 gene was localized to a six map unit region flanked by two RFLP markers, facilitating the identification of a YAC clone spanning the RPM7 locus. The avrRpm7 gene and the avrPpiA7 gene of P. syringae pv. pisi (a pathogen of pea and bean) appear to be interchangeably recognized by the Arabidopsis RPM7 R gene and by R genes in pea and bean (Dangl). Moreover, the two avr genes are very closely related (greater than 95% homology at the amino acid level) and may be interacting with R genes of conserved structure and function in Arabidopsis, pea, and bean (Dangl et al., 1992).

rapid and localized accumulation of defensive compounds following recognition is correlated with resistance in many cases, the number and nature of plant genes actually required for a resistance response is unknown. Genetic approaches to elucidating steps in R gene-mediated signaling in Arabidopsis were discussed by F. Ausubel (Massachusetts General Hospital, Boston) and B. Kunkel (University of California, Berkeley). Several HR-negative (susceptible) mutants from both labs have been characterized to date. So far, mutations appear to be allelic with the R gene itself (in this case Rp.92, which governs resistance to P. syringae avrRpt2), although several other mutations remain to be mapped and fully characterized. Analogous strategies for isolating tomato mutants defective in mounting a resistance response mediated by the Cladosporium fulvum avr9-Cf9 interaction were discussed by K. Hammond-Kosack (The Sainsbury Laboratory, Norwich, England). So far, three fully susceptible mutants map to the Cf9 locus, while two partially susceptible mutants do not. More efficient methods of screening for resistance mutants were described (Ausubel; Hammond-Kosack; Kunkel); the further identification and characterization of pathogen response mutants in Arabidopsis and tomato will be important in establishing the number of post-R gene signaling steps and in elucidating the mechanism(s) by which R genes work in specifying resistance. Biochemical investigations of intracellular signaling following pathogen recognition have largely depended on the use of cell culture systems in which transcriptional activation of defense-related genes is triggered by nonrace-specific fungal elicitors. Elicitor-induced signaling is likely to be mediated by membrane-bound elicitor receptors. Soybean plasma membranes have binding sites for the Phytophthora megaspermaf. sp. glycineagiucan elicitor (Cosio et al., 1992; Cheong and Hahn, 1991) and several discrete elicitor-binding proteins are candidates for the receptor (J. Ebel, University of Freiburg). In elicitortreated soybean cells and parsley cells (Ebel; D. Scheel, Max Planck Institut, Koln), calcium channels seem to be elicitor activated and calcium channel blockers and calmodulin antagonists inhibited transcriptional activation. In another system, tomato suspension-cultured cells respond to oligo-N-acetylglucosamine (n I 4-10) elicitors (components of fungal cell walls and also the backbone of the Rhizobium Nod factors) with a rapid increase in extracellular pH (G. Felix, Thomas Boiler Lab, FriedrichMiescher Institut, Basel). This rapid alkalinization is blocked in the presenceof a protein kinase inhibitor. These results, which point to intracellular roles for calcium, calmodulin, and protein phosphorylation, provide the outlines for potential signal transduction events in elicitor-activated gene transcription.

Postrecognition Signaling and Host Response The function of R genes in triggering resistance and the nature of signal transduction events downstream of R gene recognition have yet to be elucidated. Although the

The field of plant-microbe interactions has witnessed several recent breakthroughs, such as the molecular details of virgene induction, identification of Nod factors, and the cloning and characterization of avr genes. Other break-

Summary

through% such as the cloning and characterization of R genes, appear imminent. Parallels to mammalian systems are emerging in the world of plant-microbe interactions, for example, ion channels formed by Rhizobium proteins, similarities of hrp genes to pathogenicity genes of mammalian pathogens, and plant signal transduction via calcium and protein phosphorylation. We remain, however, largely ignorant of many facets of signaling in plantmicrobe interactions. We know little about how microbial signals are perceived by plants or how subsequent signal transduction occurs within plant cells and are probably unaware of many of the microbe-generated signals to which plants respond or of plant-generated signals to which bacteria and fungi respond. Contributions from those working on the genetics, molecular biology, and physiology of bacteria, fungi, and plants will be required to address these questions. The many nonpathogenic plant-microbe interactions in addition to the Rhizobium-plant interaction remain relatively unexplored. Genetic and molecular approaches are being initiated to investigate the signaling that is likely to underlie interactions such as those between mycorrhizal fungi and plant roots and between epiphytic bacteria and plant leaf surfaces. The importance of these interactions to plant growth and development makes it likely that they will figure more prominently at future symposia. Finally, it is clear that there will be increased efforts to capitalize on the accumulating knowledge regarding plant-microbe interactions to increase crop plant productivity by encouraging beneficial plant-microbe interactions and enhancing plant resistance to pathogens. Several investigators presented data suggesting that it is possible to generate transgenic plants with increased capacity to resist pathogen infection. A potentially useful strategy is to engineer the expression of defense genes such that they are not dependent on pathogen-generated or strict developmental signals for expression. As more is learned about molecular interactions between plants and microorganisms, it is inevitable that more light will be shed on pathways involved in interspecific signaling and signaling within plants and microorganisms. This should, in turn, lead to technological breakthroughs that can be used to benefit agriculture. Acknowledgments We apologize to those investigators whose work we could not mention owing to the limited scope of this review. We thank Eugene Nester for directing the scientific organization of a stimulating meeting and are grateful to several meeting participants for their suggestions and critical review of the manuscript.

Rtierences Banuett, F. (1991). Identification of genes governing filamentous growth and tumor induction by the plant pathogen Ustilago maydis. Proc. Nati. Acad. Sci. USA 88, 39223926. Banuett, F. (1992). 8, 174-160.

Usfilegomaydis,

thedelightful

blight. TrendsGenet.

Battisti, L., Lara, J. C., and Leigh, J. A. (1992). Specificoligosaccharide form of the Rhizobium meliloti exopolysaccharide promotes nodule invasion in alfalfa. Proc. Natl. Acad. Sci. USA 89, 5625-5629.

Becraft, P. W., and Freeling, M. (1992). Curr. Opin. Genet. Dev. 2, 751-755.

Cell interactions

in plants.

Balker, M., Urban, M., and Kahmann, R. (1992). The a mating type locus of U. maydis specifies cell signaling components. Cell 68,441550. Bonas, U., Stall, R. E., and Staskawicz, B. J. (1969). Genetic and structural characterization of the avirulence gene avr6s3 from Xanthomonas campestris pv. vesicatoria. Mol. Gen. Genet. 278, 127-136. Cangelosi, G. A., Ankenbauer, R. G., and Nester, E. W. (1990). Sugars induce the Agrobacterium virulence genes through a periplasmic binding protein and a transmembrane signal protein. Proc. Natl. Acad. Sci. USA 87,6706-6712. Cornelius, G. R., Biot,T., LambertdeRouvroit, C., Michiels,T., Mulder, B., Sluiters, C., Sory, M-P., Van Bouchaute, M., and Vanooteghem, J.-C. (1969). The Yersinia yop regulon. Mol. Microbial. 3, 1455-1469. Cheong, J.J., and Hahn, M. (1991). site for the hepta-p-glucoside elicitor Plant Cell 3, 137-147.

A specific, high-affinity binding exists in soybean membranes.

Citovsky, V., &pan, J., Warnick, D., andZambryski, P. (1992). Nuclear localization of Agrobacterium VirEP in plant cells. Science 256, 18021605. Cosio, E. G., Frey, T., and Ebel, J. (1992). Identification affinity binding protein for a hepta+glucoside phytoalexin soybean. Eur. J. Biochem. 204, 1115-I 123.

of a highelicitor in

Dangl, J. L., Ritter, C., Humphrey, M. J., Mur, L. A. J., Wood, J. R., Goss, S., Mansfield, J., and Vivian, A. (1992). Functional homologs of the Arabidopsis RPM1 disease resistance gene in bean and pea. Plant Cell 4, in press. David, M., Daveran, M.-L., Batut, J., Dedieu, A., Domergue, O., Ghai, J., Hertig, C., Boistard, P., and Kahn, D. (1968). Cascade regulation of nif gene expression in Rhizobium meliloti. Cell 54, 671-663. Debener, T., Lehnackers, H., Arnold, M., and Dangl, J. L. (1991). Identification and molecular mapping of a single Arabidopsis locus conferring resistance against a phytopathogenic Pseudomonas syringae isolate. Plant J. 7, 289-302. Ehrhardt, D. W., Atkinson, E. M., and Long, S. R. (1992). Depolarization of alfalfa root hair membrane potential by Rhizobium meliloti Nod factors. Science 256, 998-1000. Fisher, change.

R. F., and Long, S. R. (1992). Nature 357, 655-660.

Rhizobium-plant

Flor, A. H. (1955). Host-parasite interactions in flax rust-its and other implications. Phytopatholgy 45, 680-685.

signal

ex-

genetics

Gilles-Gonzalez, M. A., Ditta, G. S., and Helinski, D. R. (1991). haemoprotein with kinase activity encoded by the oxygen sensor Rhizobium meliloti. Nature 350, 170-172.

A of

Gillissen, B., Bergemann, J., Sandmann, C., Schroeer, B., Bblker, M., and Kahmann, R. (1992). A two-component regulatory system for self/ non-self recognition in Ustilago maydis. Cell 88, 647-657. Grimsley, N., Hohn, T., Davies, J. W., and Hohn, B. (1987). Agrobacterium-mediated delivery of infectious maize streak virus into maize plants. Nature 325, 177-l 79. Herbers, specificity repetitive

K., Conrads-Strauch, of plant resistance motifs in a bacterial

Hess, K. M., Dudley, (1991). Mechanisms genes: development 88, 7854-7656.

J., and Bonas, U. (1992). Raceto bacterial spot disease determined by avirulence gene. Nature 358, 172-174.

D. G., Lynn, R. D., Joerger, A. N., and Binns, A. N. of phenolic activation of Agrobacterium virulence of specific inhibitors. Proc. Natl. Acad. Sci. USA

Hirsch, A. M., Bhuvaneswari, T. V.. Torrey, J. G., and Bisseling, T. (1989). Early nodulin genes are induced in alfalfa root outgrowths elicited byauxin transporl inhibitors. Proc. Natl. Acad. Sci. USA 86,12441248. Hooykaas, P. J. J., and Schilperoort, R. A. (1992). Agrobacterium plant genetic engineering. Plant Mol. Biol. 79, 15-36.

and

Howard, E. A., Zupan, J. R., Citovsky, V., and Zambryski, P. C. (1992). The VirD2 protein of A. tumefaciens contains a C-terminal bipartite nuclear localization signal: implications for nuclear uptake of DNA in plant cells. Cell 88, 109-l 18.

Meeting 199

Review

Huang, Y., Morel, P., Powell, B.. and Kado, C. I. (1990). lator of Ti-specified virulence genes, is phosphorylated teriol. 7 72, 1142-I 144.

VirA, acoreguin vitro. J. Bac-

ciens transcriptional activator OccR causes a bend at a target promoter, which is partially relaxed by a plant tumor metabolite. Cell 69, 659-867.

Jin, S., Roitsch. T., Ankenbauer, A. G., Gordon, M. P., and Nester, E. W. (199Oa). The VirA protein of Agrobacterium tumefaciens is autophosphorylated and is essential for vir gene regulation. J. Bacterial. 7 72, 525-530.

Wei, Z.-M., Laby, R. J., Zumoff, C. H., Bauer, D. W., He, S. Y., Collmer, A., and Beer, S. V. (1992). Harpin, elicitor of the hypersensitive response produced by the plant pathogen Efwinia amy/ovora. Science 257,85-88.

Jin, S., Prusti, Ft. K., Roitsch, T., Ankenbauer, R. G., and Nester, E. W. (1990b). Phosphorylation of the VirG protein of Agrobacterium tumefeciens by the autophosphorylated VirA protein: essential role in biological activity of VirG. J. Bacterial. 172, 4945-4950.

Whalen, M. C., Innes, R. W., Bent, A. F., and Staskawicz, B. J. (1991). Identification of Pseudomonas syringae pathogens of Arabidopsis and a bacterial locus determining virulence on both Arabidopsis and soybean. Plant Cell 3, 49-59.

Keen, N. T. (1990). Gene-forgene interactions. Annu. Rev. Genet.

Willis, D. K., Rich, J. J., and Hrabak, E. M. (1991). thogenic bacteria. Mol. Plant Microbe Interact.

Knogge, W. (1991). physiological models 46c, 969-961.

complementarity 24, 447-463.

Plant resistance and identification

in plant-pathogen

genes for fungal pathogensin cereal crops. 2. Naturforsch.

Lerouge, P., Roche, P., Faucher, C., Maillet, F., Truchet, G., Prome, J. C., and Denarie, J. (1990). Symbiotic host-specificity of RhizoMun meliloti is determined by a sulphated and acylated glucosamine oligosaccharide signal. Nature 344, 781-784. Melchers, L. S., Regensburg-Tuink, T. J. G., Bourret, R. B., Sedee, N. J. A., Schilperoort, R. A., and Hooykaas, P. J. J. (1989). Membrane topology and functional analysis of the sensory protein VirA of Agrobacterium tumefaciens. EMBO J. 8, 1919-1925. Roche. P., Debelle, F., Maillet, F., Lerouge, P., Faucher, C., Truchet, G., DenarM, J., and Prome, J.-C. (1991). Molecular basis of symbiotic host specificity in Rhizobium meliloti: nod/+ and nodW genes encode the sulfation of lipo-oligosaccharide signals. Ceil 67, 1131-l 143. Scheres, 8.. Van De Wiel, C., Zalensky, A., Howath, B., Spaink, H., Van Eck, H., Zwartkruis, F., Wolters, A. M., Gloudemans, T., Van Kammen, A., and Bisseling, T. (1990). The ENODIP gene product is involved in the infection process during pea-Rhizobium interaction. Cell 60, 261-294. Schwedock, J., and Long, S. R. (1990). ATP sulphurylase activity of the nodP and nodC gene products of Rhizobium meliloti. Nature 348, 644-647. Shimoda, N., Toyoda-Yamamoto, A., Nagamine, J., Usami, S., Katayama, M., Sakagami, Y., and Machida, Y. (1990). Control of expression of Agrobacterium vir genes by synergistic actions of phenolic signal molecules and monosaccharides. Proc. Natl. Acad. Sci. USA 87, 8684-6688. Spaink, H. P., Sheeley, D. M., van Brussel, A. A. N., Glushka, J., York, W. S., Tak, T., Geiger, O., Kennedy, E. P., Reinhold, V. N., and Lugtenberg, B. J. J. (1991).Anovel highlyunsaturatedfattyacidmoiety of lipo-oligosaccharide signals determines host specificity of Rhizobium. Nature 354, 125-130. Truchet, G., Barker, D. G., Camut, S., de Billy, Huguet, T. (1989). Alfalfa nodulation in the absence Gen. Genet. 279, 65-68.

F., Vasse, J., and of Rhizobium. Mol.

Truchet, G., Roche, P., Lerouge, P., Vasse, J., Camut, S., de Billy, F., Prom& J.-C., and Denarie, J. (1991). Sulphated lipo-oligosaccharide signals of Rhizobium melilotielicit rcot nodule organogenesis in alfalfa. Nature 357, 670-673. Urzainqui, A., and Walker, G. C. (1992). Exogenous suppression of the symbiotic deficiencies of Ffhizobium melilotiexo mutants. J. Bacterial. 7 74, 3403-3406. van Brussel, A. A. N., Bakhuizen, Ft., van Spronsen, P. C., Spaink, H. P., Tak, T., Lugtenberg, B. J. J., and Kijne, J. W. (1992). Induction of pm-infection thread structures in the leguminous host plant by mitogenie lipe-oligosaccharides of Rhizobium. Science 257, 70-72. Van den Ackerveken, G. F. J. M., van Kan, J. A. L., and De Wit, P. J. G. M. (1992). Molecular analysis of the avirulence gene avr9 of the fungal tomato pathogen Cladospodum fulvum fully supports the gene-forgene hypothesis. Plant J. 2, 359-368. van Ken, J. A. L., Van den Ackerveken, G. F. J. M., and De Wit, P. J. (1991 ). Cloning and characterization of cDNA of avirulence gene avr9 of the fungal pathogen Cladosporium fulvum, causal agent of tomato leaf mould. Mol. Plant Microbe Interact. 4, 52-59. Wang,

L., Helmann,

J. D., and Winans,

S. C. (1992).

The A. tumefa-

hrpgenesof 4, 132-138.

phytopa-

Winans, S. C. (1991). An Agrobacterium two-component regulatory system for the detection of chemicals released from plant wounds. Mol. Microbial. 5, 2345-2350. Winans. S. C. (1992). Two-way chemical signaling plant interactions. Microbial. Rev. 56, 12-31.

in Agrobacterium-

Winans, S. C., Kerstetter, R. A., Ward, J. E., and Nester, E. W. (1989). A protein required for transcriptional regulation of Agrobacterium virulence genes spans the cytoplasmic membrane. J. Bacterial. 777, 1616-1622.

Molecular signals in the interactions between plants and microbes.

The field of plant-microbe interactions has witnessed several recent breakthroughs, such as the molecular details of vir gene induction, identificatio...
1MB Sizes 0 Downloads 0 Views