Molecular Genetics of Plant Disease Resistance Brian J. Staskawicz,* Frederick M. Ausubel, Barbara J. Baker, Jeffrey G. Ellis, Jonathan D. G. Jones Plant breeders have used disease resistance genes (R genes) to control plant disease since the turn of the century. Molecular cloning of R genes that enable plants to resist a diverse range of pathogens has revealed that the proteins encoded by these genes have several features in common. These findings suggest that plants may have evolved com- mon signal transduction mechanisms for the expression of resistance to a wide range of unrelated pathogens. Characterization of the molecular signals involved in pathogen recognition and of the molecular events that specify the expression of resistance may lead to novel strategies for plant disease control. Plants, like animals, are continually ex- posed to pathogen attack. Because plants lack a circulatory system and antibodies, they have evolved a defense mechanism that is distinct from the vertebrate immune system (1). In contrast to animal cells, each plant cell is capable of defending itself by means of a combination of constitutive and induced defenses (2). Knowledge about the genetic and biochemical basis of plant dis- ease resistance has accumulated since the turn of the century, when plant breeders first recognized that resistance was often con- trolled by Mendelian genes (3). The dem- onstration that plants have geographical centers of origin (4) and have co-evolved with their pathogens was a pivotal discovery for plant breeders and has led to the use of interspecific hybrids between crops and their wild relatives as sources of resistant germ plasm (5). Until 1992, however, no plant R gene had been cloned and charac- terized at the molecular level (6). Since then, R genes from several plant species Avirulent pathogen Pathogenests Virulent pathogen W.... have been cloned; this constitutes a major advance for molecular plant biology and may lead to the development of novel meth- ods for disease control. Plant Responses to Pathogen Attack The range of phytopathogenic organisms that attack plants is diverse and includes viruses, mycoplasma, bacteria, fungi, nema- todes, protozoa, and parasites (7). Each has a unique mode of pathogenicity. Despite the vast array of potential phytopathogens, resis- tance (lack of susceptibility) is the rule and susceptibility is the exception. Why one pathogen can cause disease in one plant but not in other plants a phenomenon often termed nonhost resistance-remains an im- portant unsolved problem in plant pathology. Resistance to a pathogen is manifested in a variety of ways and is often correlated with a hypersensitive response (HR), local- ized induced cell death in the host plant at Virulent pathogen Plant host cell DISEASE a Fig. 1. Gene-for-gene interactions specify plant disease resistance. Resistance is only expressed when a plant that contains a specific R gene recognizes a pathogen that has the corresponding avirulence gene (upper left panel). All other combinations lead to lack of recognition by the host, and the result is disease. Green represents hypersensitive response; yellow represents susceptibility to disease. SCIENCE * VOL. 268 * 5 MAY 1995 the site of infection (8). Although the mo- lecular mechanism is obscure, HR is thought to be responsible for the limitation of pathogen growth. Resistance does not always invoLve visible HR, which may re- flect either HR limited to individual plant cells or other uncharacterized defense mechanisms. Alternatively, the pathogen could lack a specific pathogenicity function required to cause disease in the host, or the host could lack a specific "susceptibility" factor. Although this review concerns the molecular basis of HR-mediated resistance, the elucidation of the mechanisms involved in nonhost resistance without HR may be an important component of future attempts to control plant disease. The genetic basis of HR-mediated dis- ease resistance was first clarified by Flor, who demonstrated that the resistance of flax to the fungal pathogen Melampsora lini was a consequence of the interaction of paired cognate genes in the host and the pathogen (9). His work provided the theoretical basis for the gene-for-gene hypothesis of plant- pathogen interactions and for the molecular cloning of pathogen avirulence (avr) genes and their corresponding plant R genes. An avr gene gives the pathogen an avirulent phenotype on a host plant that carries the corresponding R gene (Fig. 1) (10). In gene- for-gene interactions, the induction of the plant defense response that leads to HR is initiated by the plant's recognition of spe- cific signal molecules (elicitors) produced by the pathogen; these elicitors are encoded directly or indirectly by avirulence genes, and R genes are thought to encode receptors for these elicitors. Elicitor recognition acti- vates a cascade of host genes that leads to HR and inhibition of pathogen growth (1 1). Gene-for-gene systems involving HR have been described for pathosystems in- volving intracellular obligate pathogens (vi- B. J. Staskawicz is in the Department of Plant Biology, University of California, Berkeley, CA 94720, USA. F. M. Ausubel is in the Department of Genetics, Harvard Medical School, Boston, MA 02115, USA, and Department of Mo- lecular Biology, Massachusetts General Hospital, Boston, MA 02114, USA. B. J. Baker is at the Plant Gene Expres- sion Center, Agricultural Research Service, U.S. Depart- ment of Agriculture, Albany, CA 94710, USA, and Depart- ment of Plant Biology, University of California, Berkeley, CA 94720, USA. J. G. Ellis is at the CSIRO Division of Plant Industry and Cooperative Research Centre for Plant Sci- ence, Post Office Box 1600, Canberra, A.C.T., Australia. J. D. G. Jones is at the Sainsbury Laboratory, John Innes Centre, Norwich Research Park, Colney Lane, Norwich NR4 7UH, UK. *To whom correspondence should be addressed. E-mail: stask~garnet.berkeley.edu 661 wilaoworAmm~ m 1 Plant hnst cell1 am~ a awe, on June 26, 2007 www.sciencemag.org Downloaded from
Transcript
Molecular Genetics of Plant Disease ResistanceBrian J. Staskawicz,* Frederick M. Ausubel, Barbara J. Baker,
Jeffrey G. Ellis, Jonathan D. G. Jones
Plant breeders have used disease resistance genes (R genes) to control plant diseasesince the turn of the century. Molecular cloning of R genes that enable plants to resist adiverse range of pathogens has revealed that the proteins encoded by these genes haveseveral features in common. These findings suggest that plants may have evolved com-mon signal transduction mechanisms for the expression of resistance to a wide range ofunrelated pathogens. Characterization of the molecular signals involved in pathogenrecognition and of the molecular events that specify the expression of resistance may leadto novel strategies for plant disease control.
Plants, like animals, are continually ex-posed to pathogen attack. Because plantslack a circulatory system and antibodies,they have evolved a defense mechanismthat is distinct from the vertebrate immunesystem (1). In contrast to animal cells, eachplant cell is capable of defending itself bymeans of a combination of constitutive andinduced defenses (2). Knowledge about thegenetic and biochemical basis of plant dis-ease resistance has accumulated since theturn of the century, when plant breeders firstrecognized that resistance was often con-trolled by Mendelian genes (3). The dem-onstration that plants have geographicalcenters of origin (4) and have co-evolvedwith their pathogens was a pivotal discoveryfor plant breeders and has led to the use ofinterspecific hybrids between crops andtheir wild relatives as sources of resistantgerm plasm (5). Until 1992, however, noplant R gene had been cloned and charac-terized at the molecular level (6). Sincethen, R genes from several plant species
Avirulent pathogen
Pathogenests
Virulent pathogen
W....
have been cloned; this constitutes a majoradvance for molecular plant biology andmay lead to the development of novel meth-ods for disease control.
Plant Responses toPathogen Attack
The range of phytopathogenic organismsthat attack plants is diverse and includesviruses, mycoplasma, bacteria, fungi, nema-todes, protozoa, and parasites (7). Each has aunique mode of pathogenicity. Despite thevast array of potential phytopathogens, resis-tance (lack of susceptibility) is the rule andsusceptibility is the exception. Why onepathogen can cause disease in one plant butnot in other plants a phenomenon oftentermed nonhost resistance-remains an im-portant unsolved problem in plant pathology.
Resistance to a pathogen is manifestedin a variety of ways and is often correlatedwith a hypersensitive response (HR), local-ized induced cell death in the host plant at
Virulent pathogenPlant host cell
DISEASE
a
Fig. 1. Gene-for-gene interactions specify plant disease resistance. Resistance is only expressed whena plant that contains a specific R gene recognizes a pathogen that has the corresponding avirulence gene(upper left panel). All other combinations lead to lack of recognition by the host, and the result is disease.Green represents hypersensitive response; yellow represents susceptibility to disease.
SCIENCE * VOL. 268 * 5 MAY 1995
the site of infection (8). Although the mo-lecular mechanism is obscure, HR isthought to be responsible for the limitationof pathogen growth. Resistance does notalways invoLve visible HR, which may re-flect either HR limited to individual plantcells or other uncharacterized defensemechanisms. Alternatively, the pathogencould lack a specific pathogenicity functionrequired to cause disease in the host, or thehost could lack a specific "susceptibility"factor. Although this review concerns themolecular basis of HR-mediated resistance,the elucidation of the mechanisms involvedin nonhost resistance without HR may bean important component of future attemptsto control plant disease.
The genetic basis of HR-mediated dis-ease resistance was first clarified by Flor,who demonstrated that the resistance of flaxto the fungal pathogen Melampsora lini was aconsequence of the interaction of pairedcognate genes in the host and the pathogen(9). His work provided the theoretical basisfor the gene-for-gene hypothesis of plant-pathogen interactions and for the molecularcloning of pathogen avirulence (avr) genesand their corresponding plant R genes. Anavr gene gives the pathogen an avirulentphenotype on a host plant that carries thecorresponding R gene (Fig. 1) (10). In gene-for-gene interactions, the induction of theplant defense response that leads to HR isinitiated by the plant's recognition of spe-cific signal molecules (elicitors) produced bythe pathogen; these elicitors are encodeddirectly or indirectly by avirulence genes,and R genes are thought to encode receptorsfor these elicitors. Elicitor recognition acti-vates a cascade of host genes that leads toHR and inhibition of pathogen growth (1 1).
Gene-for-gene systems involving HRhave been described for pathosystems in-volving intracellular obligate pathogens (vi-
B. J. Staskawicz is in the Department of Plant Biology,University of California, Berkeley, CA 94720, USA. F. M.Ausubel is in the Department of Genetics, Harvard MedicalSchool, Boston, MA 02115, USA, and Department of Mo-lecular Biology, Massachusetts General Hospital, Boston,MA 02114, USA. B. J. Baker is at the Plant Gene Expres-sion Center, Agricultural Research Service, U.S. Depart-ment of Agriculture, Albany, CA 94710, USA, and Depart-ment of Plant Biology, University of California, Berkeley, CA94720, USA. J. G. Ellis is at the CSIRO Division of PlantIndustry and Cooperative Research Centre for Plant Sci-ence, Post Office Box 1600, Canberra, A.C.T., Australia. J.D. G. Jones is at the Sainsbury Laboratory, John InnesCentre, Norwich Research Park, Colney Lane, NorwichNR4 7UH, UK.
*To whom correspondence should be addressed.E-mail: stask~garnet.berkeley.edu
ruses and mycoplasmas) as well as for in-tercellular facultative and obligate patho-gens (bacteria, fungi, and nematodes).These findings suggest that common orsimilar recognition and signal transduc-tion mechanisms are involved in differentgene-for-gene signaling pathways (2, 11).Physiological features of HR common to aplant's response to different pathogens in-clude a rapid oxidative burst, ion fluxescharacterized by K'-HW exchange, cellu-lar decompartmentalization, cross-linkingand strengthening of plant cell wall, pro-duction of antimicrobial compounds (phy-toalexins), and induction of pathogenesis-related (PR) proteins such as chitinases
Fig. 2. Disease-resistantand -susceptible pheno- Atypes of TMV, M. lini, andP. syringae inoculated ontheir respective hosts.TMV was inoculated on aresistant NN tobaccoplant (A) and on a sus-ceptible nn tobacco plant(B). Melampsora lini wasinoculated on a resistantL6 L6 flax plant (C) and ona susceptible /6 /6 flaxplant (D). Pseudomonassynngae was inoculatedon a resistant RPS2RPS2 Arabidopsis plant(E) and on a susceptiblerps2 rps2 Arabidopsisplant (F).
and glucanases (2). These events charac-terize a plant's defense response irrespec-tive of the pathogen, although any partic-ular defense response can vary with re-spect to timing, cell autonomy, or inten-sity. The mechanism by which theseevents limit the growth of specific patho-gens remains unknown.HR (12) and other necrotic reactions are
hypothesized to trigger a subsequent re-sponse, referred to as systemic acquired resis-tance (SAR), that acts nonspecificallythroughout the plant: SAR reduces the se-verity of disease caused by all classes ofpathogens, including normally virulentpathogens (13). Experimental evidence sug-
gests that HR induces an unidentified diffus-ible signal; salicylic acid is known to beinvolved in both HR and SAR, but may notparticipate in the systemic signaling pathwaythat induces SAR (14-16). SAR may beinvolved in general resistance in field situa-tions where plants undergo HR. Manipula-tion of SAR by chemical inducers or bygenetic engineering may aid disease control.
Plants that are susceptible to a givenpathogen still attempt to defend them-selves. Indeed, many of the defense respons-es observed in resistant plants, with theexception of HR, are also observed in sus-ceptible plants, although usually later afterthe infection. For example, even diseaselesions in susceptible plants frequently havea defined and delimited shape, suggestingthat the host is limiting the growth of thepathogen. By contrast, Arabidopsis thalianamutants with decreased ability to synthesizephytoalexins (17) or to induce SAR (18),and Arabidopsis and tobacco plants engi-neered to degrade salicylic acid (15), devel-op larger lesions, are susceptible to verysmall amounts of pathogen infiltration, andallow the pathogen to grow to high titers.
Role of Pathogen AvirulenceGenes in Triggering thePlant Defense Response
Genetic, biochemical, and physiologicaltechniques have been used to study respons-es to pathogen attack in heterozygous pop-ulations of crop plants such as maize, soy-bean, bean, parsley, tomato, potato, and bar-ley. Because plant defense responses are sim-ilar irrespective of the pathogen, it has beendifficult to provide compelling evidence forthe significance of particular responses inconferring specific resistance. Thus, tomatoor the crucifer A. dtaliana have been used asmodel hosts to study plant responses topathogen attack (19). The receptor-ligandmodel postulates that pathogen avr genesspecify elicitor molecules that induce diseaseresistance in host plants that contain a cog-nate R gene. This has been confirmed instudies of the interaction between tomatoand its leaf mold pathogen Cladosporiumfulvum (20, 21). Small peptides extractedfrom tomato leaves infected with a pathogenrace carrying an avr gene had the featuresexpected of an avirulence gene-encodedelicitor; the purified peptides elicited anHR-like response on tomato cultivars thatwere resistant to the C. fulvum race used toobtain the peptide (20). In the best-docu-mented cases, two C. fulvum avirulencegenes, avr9 and avr4, were shown to encodeprecursors of elicitor peptides that specifical-ly elicited HR in tomato plants that har-bored the corresponding R genes Cf-9 andCf-4, respectively (22).
ASwhich the avr-generated signal has beendefinitively identified is the Pseudomonassyringae pv. glycinea avrD locus. This locusencodes enzymes involved in the synthesisof exported syringolides that elicit HR insoybean cultivars carrying the R gene Rpg4(23). In the case of tobacco mosaic virus(TMV), the viral-encoded coat protein ap-pears to function as a specific intracellularelicitor that activates HR in Nicotiana syl-vestris cultivars that carry the R gene N'(24). Uncovering the molecular basis ofhow these genes and other cloned aviru-lence genes elicit a plant defense responsewill ultimately be necessary for a completemolecular understanding of host-pathogenspecificity.Why do pathogens contain avirulence
genes? The avr genes may encode pathoge-nicity factors that confer a selective advan-tage for the pathogen, as has been shown forseveral bacterial avirulence genes that conferenhanced virulence on susceptible hosts,that is, on hosts that do not carry a cognateR gene (10). In the case of fungal pathogens,no role in pathogenicity for avr-encodedelicitor peptides has yet been established.
Cloning and Characterizationof Plant R Genes
The cloning of several R genes since 1992reflects, in part, the simultaneous devel-opment of the infrastructures required forinsertional mutagenesis and positionalcloning in several plant species. For manyyears, transposons have been exploited asinsertional mutagens for efficient identifi-cation and isolation of genes (a processtermed transposon tagging) in a widerange of organisms, including plants. TheTam elements of snapdragon and the Ac!Ds, Spm, and Mu elements of maize havebeen used for the isolation of a variety ofgenes. Fortunately, members of the maizeAc and Spm transposon families functionwhen transferred into heterologous plantspecies; this attribute permits the engi-neering of efficient gene tagging systemsin a variety of plant species (25). Trans-poson-based gene tagging systems havebeen used to clone R genes from maize,tobacco, tomato, and flax (26-29).
Map-based positional cloning of toma-to and Arabidopsis genes has become fea-sible with the development of high-densi-ty physical-genetic maps for these two spe-cies (30). The small genome size of Ara-bidopsis (- 150 Mb) and the relativelysmall genome size of tomato (-950 Mb),and the relatively small number of repeat-ed sequences in these species, have facili-tated the successful positional cloning oftwo R genes in these species (31, 32).
The first plant R gene to be cloned wasthe maize Hml gene (26). This gene, which
controls resistance to race 1 isolates ofCochliobolus carbonum, was identified bytransposon tagging with the maize (Mu)transposon. Hml encodes a reduced nico-tinamide adenine dinucleotide phosphate(NADPH)-dependent HC-toxin reduc-tase, HTRC. HTRC inactivates HC-toxin, apathogenicity factor produced by the fungusC. carbonum Nelson race 1 that permits thefungus to infect certain genotypes of maize(26, 33). The genetics of the interactionbetween maize and C. carbonum differ fromthose of gene-for-gene systems because tox-in-deficient C. carbonum strains lose theirability to cause disease in maize cultivars thatdo not carry Hml.
The first plant R gene to be cloned thatconforms to a classic gene-for-gene relationwas the tomato PTO gene (32). The PTOlocus confers resistance to strains of P. sy-ringae pv. tomato (Pst) carrying the aviru-lence gene avrPto (34). A yeast artificialchromosome (YAC) clone that spannedthe PTO region was identified with the useof a map-based cloning strategy and a re-striction fragment length polymorphism(RFLP) marker tightly linked to PTO. ThisYAC clone was then used to isolate com-plementary DNAs (cDNAs) correspondingto the PTO region, and subsequent geneticcomplementation tests identified a cDNAclone corresponding to PTO. The transla-tion product of PTO predicts that it en-codes a serine-threonine protein kinase;hence, this product may play a role in signaltransduction.
Interestingly, PTO appears to be part ofa complex locus that consists of a cluster offive to seven genes, all homologous to PTO.One of these PTO homologs, FEN, conferssensitivity to the organophosphorous insec-ticide fenthion. The evidence that FEN is aseparate gene comes from mutational anal-yses (35) and from the demonstration that acDNA clone with -80% homology to PTOconfers sensitivity to the insecticide (36). Athird gene, involved in both PTO resistanceand fenthion sensitivity, was identified bythe isolation of tomato mutants that weresimultaneously altered in both bacterial re-sistance and fenthion sensitivity. Theseplants carry mutations in a new locus, des-ignated PRF (to indicate Pseudomonas resis-tance and fenthion sensitivity), which istightly linked to PTO (35). Apparently,PRF is part of the signal transduction path-way that includes PTO and FEN. Becausemany tools are available to dissect the PTOsignal transduction pathway, this system is apromising area for future research.
Four additional plant R genes that con-form to classical gene-for-gene relationshave been cloned. The Arabidopsis RPS2gene, which confers resistance to the bac-terial pathogen P. syringae pvs. tomato andmaculicola expressing the avirulence gene
SCIENCE * VOL. 268 * 5 MAY 1995
avrRpt2, was identified by isolation of Ara-rbidopsis mutants that did not exhibit HR in1response to P. syringae strains carrying avr- 'Rpt2. The RPS2 gene was then cloned bymeans of a map-based strategy similar inconcept to the method used to identify thetomato PTO gene (31).
The tobacco N gene, which confers re-sistance to TMV, was isolated by transpo-son tagging with the autonomous maizetransposon Ac (28). At elevated tempera-tures, N does not mediate HR after TMVinfection, but if the temperature is loweredafter TMV infection of seedlings carryingthe N gene, the seedlings become necroticand die. Some survivors contained Ac-tagged mutations at the N locus.
The tomato Cf-9 gene, which confersresistance to the fungal pathogen C. fulvumexpressing the avirulence gene avr9, wastagged by a maize Ds transposable element(27). A tomato line lacking Cf-9 was engi-neered that expressed the C. fulvum avr9gene under the control of a plant genepromoter (37). When this line was crossedwith a line containing both Cf-9 and a Dselement, most of the progeny died becausethe interaction of the avr9 gene productwith the Cf-9 gene product resulted in theelicitation of systemic HR. However, mu-tants carrying a Ds-inactivated tagged Cf-9gene survived.
The flax L6 gene, which confers resis-tance to the fungal pathogen M. lini, wasalso identified by transposon tagging withthe maize transposon Ac. However, becauseno selection for L6 mutations was available,mutants were identified by visual inspectionof thousands of flax plants containing puta-tive transpositions of Ac into the L6 gene(38). The phenotypes of TMV, M. lini, andP. syringae inoculated on both resistant andsusceptible hosts are shown in Fig. 2.
Although the RPS2, N, Cf-9, and L6genes confer resistance to bacterial, viral,and fungal pathogens, DNA sequence anal-ysis revealed that all four genes encodeproteins that contain leucine-rich repeats(LRRs). LRR motifs are found in manyplant and animal proteins and are usuallyinvolved in protein-protein interactions(39). Moreover, all four of these genes arefundamentally different from both themaize Hml and the tomato PTO R genes. Acomparison of the sequences of the RPS2,N, CF-9, and L6 proteins reveals that RPS2,N, and L6 share significant homology,whereas CF-9 appears to belong to a sepa-rate class (Fig. 3). RPS2, N, and L6 allcontain a conserved nucleotide binding site(NBS) in addition to the LRRs. Furtherinspection of these three proteins revealsthat N and L6 are more closely related toeach other than to RPS2. The NH2-termi-nus of RPS2 contains a leucine zipper,which might be involved in protein dimer-
studies and sequence analysis indicate thatthe RPS2 and N proteins, which lack aleader peptide, are most likely cytoplasmicand probably recognize intracellular ligands,
1 10 20 30 40 M 0N m S 1L6 MSY LR EV AT AV AL L LPF IL LN K FWR PN SK DS IV N DDD DS TS EV D A 18D SrN P SFSRPS2 MDF 8 L1I VOGClA 0
IQ 70 80 901071010N RWW SYDVFL F HGE THTL EV ND KGIKTFQ R EAT PF-16 PSVEYE FL8F GP EQ FL OS RYK HTFR DE LKGKE iPNLLRPS2 VICE L ES N 1R D -£1 Ii R
120 130 140 160 160N ATL E F E S I iM ICKT I QTVI P rF D P S HVUNL6 RAIIDOEKI YVPI 8SG ADSK CtIM EvIVRE£DPR R I-P IF YMV D PS DV HRPS2 EISCRS WIRAVR E -L £REQI RiI R
170 180 190 210 220N O0K ElWIFP K U F E TRYKDDVIEQR SA CI DSN AAD IADCIIIRl vDL6 aTGICY|K KFR HTANN FWD QTI QKNWKDIAKKIVO WHI GKNDKQA I AID K IS ARPS2 YL SlG CCDYIL CCKV S A IKS i E RIE R S E AV KITDQIOSlIQVITR
230 RA0 25020 0NISSIKIICKIS SSY LO0NI V Gl tIOIII KillELUEE GIINiGVRE AL6 ID S IH11 SKENLI LETDELVGNlI DIHITA LEK USLD SIN VTMVL ARPS2 i PAI S, VG EITVI L E F S E E B
290 310 320 330N RAI IF$iTLLQRMDsiQF FLKDIIK N KRMiiS N A L. E R E KRAN YL6 INI VK 1 C F ID N T Q E K DL V VL K KI V8VI L RI D SOSVOFRPS2 AN E L TKHL. Bin F EC.T A&A L SWE K
340 211M3 370 380 9N W P1E D GKHQaMIASRIL?1 KVV DN D I LA D LS WF&KMLISVR I i TL6 NNDSGG RKTIKERVISR F KE DIML S Fl a8I IRPS2 RENAKPL VU KVPm
400 41Q 470 430N HI IE KND II VI AiID E1 9KIAlIP 1,iEN~l~LEVNYAK16 RIVLOTLNENQCKLIESIIKPRIBL S6KtAPNT 1V lTNDUTA
N Q N A A I-INI KL.6 _ T I L.L IK 0 E A VE D[T~il- QILI.R - TLN IL lEnO YI 14 UPERPS2 _ JUNS E!- 6N10 ~EM-g520 530 *,, 5E0 5Q
L*RGE LSKIIY1Q1
Cr'AEY .lRIIL"D'KSLOIlIL6 AIKEIDIE CFI GO INIKIEIEP YYIMWITDC NIFIYPASN FLICMlm oDIDDEFKMRPS2 _ _ IEIOI Y_1_1TSSHICNTYi 0J2L K570 §~~~~5059 10NHDIIIK) K Y VNFOQD KDIPIEKIRSR8 IL LUUEVNT AME
L6 HD OILIRIDM RE IREMVPKIRSR S E I D L L L N K K K V KA IS 1RPS2 AC ILIF.DK CIKNEVLALML
620 630 640 660 _70N RF SIN IV KINMiKRLIR VIFIN MORSST H Y A DYP N N RCIF V CITNP ESIF PI S T F EILL6 E F KISEIF INLISEILIR YLIHAREAML T GDI NNL P N KLEPF KIH-GIEDDPIPL T N Y T I
RPS2AILM DNR KM-L_01 N690 700 710 720N MILIV HILIQILIRmN S L R HL WT E T K H L SI DI LSWS KR L T R T P D FTOIP
L6 ENIL1II IIIILIEUSHITADDWGGWRHMMKMAERLKVIL ,ASINNYSLYGRRVRLSDCWR PRPS2 h P VIIE VLDl IL S
7.#n2 750 7610 Z77 7-N NLEYVNLYQCfI5N EVH SLj V1OnIL1CSGR70 PVVEL INNL t Y V N L r aclg ttIE E V H H L-lG C CS VllG - -NDbC K SL K R F P C WVN `E L E Y L GTL6 KSI EVLSMIT IEMDDEVDI IGQELK LTVK K F CPI O K I S G G T F GOMILKGLRELCL ERPS2 Ft S. ITEIPE LI 1KYIL -
790 ~~am' S10 92 0 Ad0N L RSI SlLELI KEl 0 1H MIII GI`Rl SS -FQYIOHVKIIV-IML6 F ND STINLIREVVAD OLISSL KVILIKITTIGIAIKEVEI NIP L GLKELSITISSRI PNILIS LILRPS2 _ E RKLKHIIDL QRT IT PRADSW 860 8~~~~~70 ago 890N NfV LP-S KS7 SILIVSL WZVIRCL ES ERVIFDASDT LI LIRPPS
L6 DEK V Y PDDDMGDMPPAPSEDESSV WK VS KK IE K T R I N V N VVDDARPS2 K EE N L Y VII A G IX L Ot IFIE D F A E E; LSF A DILI Y-V N~F Tr~L( Lt SLVF0 TILIT Lnr-ocN
L6
L6
L6
L6RPS2
-~~~~~~~~~~~MrLirVL r
S SjoI~HLPR Y LLPTrsLTYLK I Y QCTE STWL POG EN NJTS1LE IN DjI jF QT GGDILDOG
0 9970 960 9 9_ 1000ISS SKK1DL S RNNF E HIL PISI r A OL GA LQS 1 D L KL D L RLQORSS IRI RKVNOGLARIKL KDLLCSSTCKLRKfYI TECPDLI E LPEELGVND: ME S VLTL HSLHNNL TRIBVIWG N S VSI Q LRN L
-aLF EL PI ELNEL HVD FHM1ALKF1rIY rTRKK LEHRVKLDDAHNDrMYNLFAYTMFaITVVIVYP S MA E LTr R DICIPIR LIE VOGPIMI RSIL P9F P MIL KKIL DIL A VA NI TKEE DILDA i GSILRICI Ni 61 H.C N E L FLVRFQ K E1070 1080 __ _ ___ 10980 0 1110 1120QON I S-S MRH Dl -S-A-DSISITT L r V F G
EELVS LELELDDTSSI ER VSSSKKTLST VV VIP S1 IR EI IEIOIIEIEEKIQDLYLLIE Vll EILIFYD1CIRIEII EIEII EHGE
1130 1140 116 1160 1170N LPENWYI PDKFLGFAVCYSRISIII OTTIA HL-VCDDKMSIUMT OKLAS EL6 EGCTSLGRLPLEKLKELDI GGCPDILTIELVQITIV VIA VIRGLT PDCRREVGPMMI QRPS2 SSP S V E QPTFaL KT L R T LCI S IPLp
1240 1250 17310 1260Q120G2EE KIMIY~cSR-KG PE V N A LIL-QIMREINI SINIEIPTIN HOTTIVEVPSLREI EGIEL rKIS L GCTSLEILMLP001I KNINVLIDQ K L T T L
ERE ISO EG S E WREvPD aQL G S L KINLNVLD
ER.RTIQMN UUIT CI E KIWIKALE KIDQINE
1290 1300 1310 1320STIGIC R FRTPYNIN RP RT S F YELING1OQGCK SL S~VDHL SA L KT T L P PRA R ITWP DQPYRII CIY L P. RF V P.N
Fig. 3. Amino acid alignment of the proteins expressed by the R genes N, L6, and RPS2. The sequencesfor N (28), L6 (29), and RPS2 (31) were aligned with the use of the programs AMAS (62) and Clustal W (63).Identical matches are shown in blue, and similarities are shown in green.
whereas the L6 protein may attach to thecell membrane by means of a signal anchor.
The CF-9 protein appears to consist pri-marily of extracytoplasmic LRRs, with aCOOH-terminal membrane anchor. Thisstructure suggests that the CF-9 protein is areceptor for the extracellular ligand provid-ed by the Avr9 elicitor peptide. Whether adirect interaction occurs between CF-9 andthe Avr9 peptide is unknown. Membraneprotein preparations from leaves of plantsthat express Cf-0 and Cf-9 bound the Avr9peptide with almost equal affinity, whereasthe intact leaves of the Cf-0-expressingplants did not respond to the Avr9 peptide.Thus, although CF-9 protein binds theAvr9 peptide, other plant proteins (perhapsexpressed by other members of the Cf-9multigene family) also bind the Avr9 pep-tide (40).
Additional R genes will likely be isolatedby positional cloning and transposon tag-ging. Transposon tagging may be a moregeneral method for R gene isolation from awide variety of plant species. Because fourof the six R genes cloned to date have beenfound to encode products with strikinglysimilar sequences and structural features, itis likely that other R genes will be isolatedon the basis of homology to the known Rgenes. Indeed, homologs of RPS2, N, L6,and Cf-9 exist in a variety of species. The L6gene hybridizes to RFLPs linked to the un-linked rust resistance genes at the flax Mresistance locus (38).
Signal Transduction Events andExpression of Disease Resistance
The mechanisms underlying gene-for-generesistance probably involve specific recog-nition of a pathogen-generated ligand (pro-duced by an avr gene) by a plant receptorencoded by an R gene (Fig. 4). The eventsthat occur after recognition are a matter ofspeculation, but the domains in R geneproteins provide clues. For example, if CF-9is a transmembrane receptor and its LRRregion binds the Avr9 peptide directly, thecytoplasmic domain of CF-9 might directlyactivate a kinase such as that encoded bythe tomato R gene PTO. This event wouldbe analogous to the mechanism by whichCD4, a membrane-anchored receptor on Tcells, activates the tyrosine protein kinasep56Lck (41). Alternatively, CF-9 might in-teract with other proteins, including trans-membrane protein kinases that also carryextracellular LRRs (42) or secreted LRR-carrying proteins such as polygalacturonase-inhibiting proteins (PGIPs) (43). A geneticapproach holds promise for the identifica-tion of genes required for Cf-9-dependentresistance (44).
The plant cellular defense responses ac-tivated by the N protein may be analogous
SCIENCE * VOL. 268 * 5 MAY 1995
ization, whereas the NH2-terminus of Nshares homology with the Toll protein ofDrosophila and the mammalian interleukin-lreceptor (IL-iR). Preliminary mutagenesis
to "natural" or innate immunity in verte-brates and insects. In mammals, perceptionof signals produced by pathogens results intranslocation of the Rel-related transcriptionfactor NF-KB (45) from the cytoplasm to thenucleus, where it activates transcription ofdefense-related genes (46); in Drosophila thesame process occurs with the Rel-relatedtranscription factor Dif (47, 48). In themammalian immune system, the cytoplasmicdomain of IL-iR is involved in the transduc-tion of the signal required for the transloca-tion ofNF-KB; this domain has sequence andfunctional similarity to the cytoplasmic do-main of the Drosophila Toll protein (49). InDrosophila development, the perception ofan extracellular signal by Toll results in thetranslocation of Dorsal, a homolog of NF-KB(50). The presence of a domain in the NH2-terminus of the N protein that is similar tothe cytoplasmic domains of Toll and IL-iRsuggests that this domain may trigger anintracellular signal transduction cascade inplants, analogous to the Toll and IL-iRpathways in animals. The N protein, andpossibly other R gene-encoded proteins,may serve as receptors that activate a Rel-related transcription factor that induces theexpression of genes responsible for HR. Un-like that of the N protein, the NH2-terminaldomain of the Arabidopsis RPS2 protein isnot similar to those of Toll or IL-iR. How-ever, RPS2 does contain a leucine zippermotif at the NH2-terminus that may be in-volved in the formation of a heterodimerwith a Toll-like protein.
The chain of events between pathogeninfection in a plant and the onset of HR isnot well defined. However, it has beenwidely observed that HR is preceded by arapid outburst of the reactive oxygen inter-mediates (ROIs) 02- H202, and OH (51).A plasma membrane multisubunit NADPHoxidase complex, similar to the one foundin mammalian phagocytes, might be in-volved in the release of ROIs in plants. If arapid oxidative burst is crucial to HR, theactivation of a protein kinase could lead tothe activation of an NADPH oxidase ratherthan to transcriptional activation. In themammalian innate immune response, ROIshave been shown to induce the expressionof acute phase response genes by activatingthe transcription factors NF-KB (52) andAP-1 (53). In plants, R gene-mediated in-duction of intracellular ROIs suggests that aredox-regulated transcription factor mayalso be involved in the activation of HR.
Evolution of Plant DiseaseResistance
The following scenario for the evolution ofplant disease resistance has been proposed(54): The evolutionary ground state is con-sidered to be a compatible interaction in
which a pathogen has evolved to be viru-lent on a particular host plant. Selectionfavors the evolution and spread of hostindividuals that specifically recognize thepathogen and resist infection. For example,a receptor that evolved to activate defenseresponses to pathogens in general may bemodified so that it specifically recognizes aparticular pathogen product (an avirulencegene product). The pathogen responds bylosing the avirulence gene by mutation.This phenomenon is absolutely essential forthe survival of obligate parasites. The hostis now susceptible, and again selection isbrought to bear on new host R gene speci-ficities. Consequently, the evolution ofgene-for-gene interactions can be seen as acontinuing step-by-step or move-counter-move process, whose consequence in plantpopulations is a diversity of R genes indifferent individuals of a host species and acorresponding diversity of avirulence genesin different pathogen races.
The existing diversity of R genes is theproduct of an evolutionary process that ap-pears to have proceeded along two majorbranches. On one branch, exemplified bythe M rust resistance locus in flax, tandemarrays of related R genes with different spec-ificities are found in the plant genome (38).The other evolutionary branch is exempli-fied by the flax L rust resistance locus; thespecificities at this locus behave geneticallyas alleles of a single gene, and differentspecificities existing in heterozygotes can-not be recombined (38). The cloning of theL6 allele of this locus supports the classicalgenetic interpretation of a simple L locusbut has also provided some surprises (29).The genes at the genetically complex M
Fig. 4. Receptor-ligandmodel for the recognitionand expression of plantdisease resistance. Inthis hypothetical model,the R gene is thought toencode for either an ex-tracellular receptor, suchas the protein product ofthe Cf-9 gene of tomato,or an intracellular recep-tor, such as the productof the N gene of tobac-co. The ligand in thismodel may represent thedirect or indirect productof the pathogen's aviru-lence gene. The specificrecognition event trig-gers a signal transduc-tion cascade that mayinvolve protein kinasesand may lead to the ex-pression of plant diseaseresistance.
locus are related in sequence to the un-slinked L gene, with 70 to 90% nucleotidetidentity. The M locus appears to haveevolved by local duplication and divergencefrom an L-like R gene progenitor, whereasthe L locus appears to have evolved as amultiple allelic series, with only a single Lspecificity capable of existing in a homozy-gote. The contrasting evolution of two suchclosely related genes may be the result of arare duplication event that occurred only atthe M locus and then provided the oppor-tunity for rapid amplification by unequalcrossing over. Similarly, molecular analysesof the TMV resistance locus N in tobaccoand of the Cf-9 locus in tomato have re-vealed a clustered gene family (27, 28).
Molecular Basis and Evolutionof R Gene Specificity
R genes specifically distinguish isolates of asingle pathogen species. The multiple resis-tance specificities encoded by the 13 allelesof the cloned L gene of flax provide anopportunity to study the molecular basis ofthis specificity. Three alleles, LU, L6, andL10, have been cloned and partially charac-terized. Although L6 and L10 are similar, L2has additional numbers of an LRR motifthat occurs in the COOH-terminal regionof the gene product. This region may deter-mine ligand specificity.
Pathogen propagules increase to vastnumbers in comparison with their hosts,and consequently there is a greater oppor-tunity for virulent pathogen races to arisethan for corresponding new host resistancespecificities. Coupled with the differencesin population size, pathogen evolution from
avirulence to virulence usually results fromloss-of-function mutation. The correspond-ing gain of function (that is, resistance inthe host) is unlikely to occur by simplemutation. Do plants have a mechanism togenerate new R gene specificities and keeppace with the evolutionary progress ofpathogens? Pryor, Hulbert, Bennetzen, andtheir colleagues have observed high-fre-quency loss of rust resistance in corn asso-ciated with unequal crossing over at theRpl locus, and they have proposed that thisshuffling of preexisting coding informationmay provide a means for plants to generatenew specificities (55). The LRR domains,which are important in receptor selectivity,may be involved in defining the recogni-tional specificity of the pathogen (56).
Bioengineering for Novel andStable Plant Disease Resistance
The isolation of plant R genes providesopportunities for producing crop plant va-rieties with increased disease resistance. Po-tential approaches can be subdivided intothose that augment classical breeding tech-niques and those that involve direct engi-neering of crop plants.
The identification of a variety of R geneson the basis of amino acid sequence conser-vation will enable plant breeders to monitorR gene segregation using appropriate DNAprobes instead of testing progeny for diseaseresistance and susceptibility. The same ap-proach may greatly facilitate the identifica-tion and introgression of new resistancesfrom wild species that either interbreedpoorly with crop species or do not cross atall. For example, additional nematode resis-tances exist in wild accessions of Lycopersi-con peruvianum, but the amount of worknecessary for the recurrent introgression ofmany such new resistances is daunting.Candidate nematode R genes could be iden-tified by homology, isolated by molecularcloning, and transformed into crop varietiesto evaluate their effectiveness.
It is frequently the case that after pro-tracted breeding efforts of 10 or more years'duration, a new resistant plant variety isproduced only to be overcome by a newpathogen race within a few years of deploy-ment. This is immensely wasteful of timeand effort. Population genetic theory pre-dicts that the breakdown of resistance willhappen more slowly in varietal mixturescarrying an array of different R genes (57).However, such mixtures have not beenadopted in practice because they exhibitvariation in multiple characteristics, includ-ing time to harvest. With several differentcloned R genes responsive to the samepathogen, plant varieties could be producedconsisting of mixtures of lines that differonly in the R gene allele they carry. Such an
666
environmentally and ecologically sound ap-proach to disease control would find accep-tance among consumers and growers.
One exception to the lack of durability ofR gene-mediated resistance is the BS2 geneof pepper, which confers resistance to strainsof Xanthomonas campestris pv. vesicatoria thatcontain the avrBs2 avirulence gene and isextremely effective in controlling the bacte-rial spot disease of pepper. The success ofthis gene seems to be related to the fact thatall strains of the pathogen examined to datecontain an active copy of the avrBs2 gene,and if the gene is lost, the pathogen suffers asevere fitness penalty (58).
The approaches described above to facil-itate the introduction of disease-resistantvarieties all take advantage of cloned Rgenes. However, R genes appear to functionat or near the beginning of a complex signaltransduction cascade that leads to HR andultimately to SAR. It would be desirable todirectly manipulate HR and SAR by engi-neering the signal transduction pathwaysthat lead to their activation. Genetic dis-section of HR and SAR and their regula-tion is beginning in Arabidopsis.
The engineering of HR and SAR cannottake place unless certain problems are ad-dressed, such as the lethality of a constitu-tively activated defense response (37). Tocircumvent this problem, alleles of R genes,or of genes that encode the products withwhich the products of R genes interact,could be found that would partially activatethe defense response. The result would be aphenotype analogous to SAR that conferssome degree of resistance but does not killthe plant. Mutations of this sort are likely tobe selected against in natural populationsbecause they would likely partially cripplethe host in the absence of severe pathogenattack. In agricultural settings, however,they could be advantageous, even thoughthey might be associated with yield penal-ties. Dominant mutations at R gene loci andrecessive mutations at some other locimight be expected to result in partial con-stitutive expression of the defense response.Some necrotic or disease lesion mimic mu-tations may arise in this manner (59). In-deed, the phenotype of the recessive barleymutant mlo, which has been widely used inbarley breeding, can be phenocopied byapplication of 2,6-dichloroisonicotinic acid,a chemical that elicits a SAR-like response(60). Thus, in some circumstances, usefulmutations can be identified in which thedefense response has been "primed."
It has been suggested that if a suitablepathogen-inducible promoter, such as theprpl-1 promoter of potato (61), could befound, it would be possible to induce theexpression of a race-specific elicitor suchasthAv9ptieolinclshaarthe Avr9 peptide only in cells that arebeing challenged by a compatible pathogen
SCIENCE * VOL. 268 * 5 MAY 1995
(20). If this engineered plant also containeda functional Cf-9 gene, then a previouslycompatible pathogen would now elicit HR.Because potato and tomato are so closelyrelated, Cf-9 seems likely to function inpotato, and this system offers real potentialfor increasing resistance to potato lateblight caused by Phytophthora infestans.
In the course of transposon tagging ofthe tomato Cf-9 gene, alleles have beengenerated in which the Ds transposon so-matically excises from Cf-9 and thus re-stores function. In the presence of Avr9,this excision results in the formation oflocalized necrotic sectors in which bothAvr9 and CF-9 are active (27). Preliminaryexperiments indicate that plants with thisphenotype show some characteristics ofSAR, including enhanced resistance topathogens that would otherwise be compat-ible. This phenomenon has been designatedgenetic acquired resistance (GAR) to indi-cate that it is a genetically imposed SAR.Fine tuning of the system will undoubtedlybe required to achieve the optimum balancebetween activation of the defense responseand crop yield.
SummaryThis is an extremely exciting time for thefield of plant pathology. The cloning andcharacterization of several plant R genesconstitutes a major breakthrough in theelucidation of the molecular basis of diseaseresistance to a wide range of phytopatho-gens. As a result, we are finally in a positionto determine the molecular basis of plant-pathogen specificity and expression of dis-ease resistance. Future research challengesinclude the determination of the mecha-nisms by which R gene products recognizepathogen elicitors and the plant defenseresponse blocks pathogen growth. The basicknowledge obtained from this research willundoubtedly help to produce novel forms ofdurable disease resistance and will lead to adecline in the use of environmentally dam-aging pesticides.
REFERENCES AND NOTES
1. V. Walbot, Trends Genet. 1, 165 (1985).2. C. J. Lamb, M. A. Lawton, M. Dron, R. A. Dixon, Cell
56, 215 (1989).3. R. H. Biffen, J. Agric. Sci. 1, 4 (1905).4. J. R. Harlan, Science 174, 468 (1971).5. E. E. Leppik, Annu. Rev. Phytopathol. 8, 323 (1970).6. C. J. Lamb, Cell 76, 419 (1994).7. G. Agrios, Plant Pathology (Academic Press, San
Diego, CA, 1988).8. N. T. Keen, A. Bent, B. J. Staskawicz, in Biotechnol-
ogyin PlantDisease Control, I. Chet, Ed. (Wiley, NewYork, 1993), pp. 65-88.
9. H. H. Flor, Annu. Rev. Phytopathol. 9, 275 (1971).10. J. L. Dangi, in Bacterial Pathogenesis of Plants and
Animals: Molecular and Cellular Mechanisms, J. L.Dangl, Ed. (Springer-Verlag, Heidelberg, Germany,1994), vol. 192, pp. 99-118.
1 1. N. T. Keen, Annu. Rev. Genet. 24,447 (1990); D. W.
.nhiGabriel and B. Rolfe, Annu. Rev. Phytopathol. 28,365 (1990).
12. Z. Klement, in Phytopathogenic Procaryotes, M. S.Mount and G. H. Lacy, Eds. (Academic Press, NewYork, 1982), vol. 2, pp. 149-177.
13. J. Ryals, S. Uknes, E. Ward, Plant Physiol. 104, 1 109(1994).
14. T. Gaffney et al., Science 261, 754 (1993).15. T. P. Delaney et al, ibid. 266, 1247 (1994).16. B. Vernooij et al., Plant Cell 6, 959 (1994); Z. Chen, J.
W. Ricigliano, D. F. Klessig, Proc. Nati. Acad. Sci.U.S.A. 90, 9533 (1993).
17. J. Glazebrook and F. Ausubel, Proc. NatI. Acad. Sci.U.S.A. 91, 8955 (1994).
18. H. Cao, S. A. Bowling, A. S. Gordon, X. Dong, PlantCell 6, 1583 (1994).
19. J. L. Dangi, Adv. Plant Pathol. 10, 127 (1993).20. P. J. G. M. de Wit, Annu. Rev. Phytopathol. 30, 391
(1992).21. Adv. Plant Pathol., in press.22. M. H. A. J. Joosten, T. J. Cozijnsen, P. J. G. M. de
Wit, Nature 367, 384 (1994); G. F. J. M. van derAckerveken, J. A. L. van Kan, P. J. G. M. de Wit,Plant J. 2, 359 (1992).
23. S. L. Midland et al., J. Org. Chem. 58, 2940 (1993);N. T. Keen and R. I. Buzzell, Theor. Appl. Genet. 81,133 (1991).
24. J. N. Culver and W. 0. Dawson, Mol. Plant-MicrobeInteract. 4, 458 (1991).
25. B. Baker, J. Schell, H. Lorz, N. Fedoroff, Proc. Nati.Acad. Sci. U.S.A. 83,4844 (1986); J. D. G. Jones, F.M. Carland, P. Maliga, H. K. Dooner, Science 244,204 (1989).
26. G. S. Johal and S. P. Briggs, Science 258, 985(1992).
27. D. A. Jones, C. M. Thomas, K. E. Hammond-Ko-sack, P. J. Balint-Kurti, J. D. G. Jones, ibid. 266, 789(1994).
28. S. Whitham et al., Cell 78,1101 (1994).29. G. J. Lawrence, E. J. Finnegan, M. A. Ayliffe, J. G.
Ellis, Plant Cell, in press.30. E. M. Meyerowitz, Cell 56, 263 (1989).31. A. F. Bent et al., Science 265,1856 (1994); M. Min-
drinos, F. Katagiri, G.-L. Yu, F. M. Ausubel, Cell 78,1089 (1994).
32. G. B. Martin et al., Science 262, 1432 (1993).33. R. B. Meeley, G. S. Johal, S. P. Briggs, J. D. Walton,
Plant Cell 4, 71 (1992).34. P. C. Ronald, J. M. Salmeron, F. M. Carland, B. J.
Staskawicz, J. Bacterial. 174, 1604 (1992).35. J. M. Salmeron, S. J. Barker, F. M. Carland, A. Y.
Mehta, B. J. Staskawicz, Plant Cell 6, 511 (1994).36. G. Martin et al, ibid., p. 1543; C. Rommens, J. Sal-
meron, D. Baulcombe, B. Staskawicz, ibid., in press.37. K. E. Hammond-Kosack, K. Harrison, J. D. G. Jones,
Proc. Nati. Acad. Sci. U.S.A. 91, 10445 (1994).38. J. G. Ellis, G. J. Lawrence, E. J. Finnegan, P. A.
Anderson, ibid., in press.39. B. Kobe and J. Deisenhofer, Trends Biochem. Sci.
19, 415 (1994).40. G. Honee et al., in Advances in Molecular Genetics of
Plant-Microbe Interactions, M. Daniels, A. Downie,A. Osbourn, Eds. (Kluwer Academic, Dordrecht,Netherlands, 1994), vol. 3, pp. 199-206.
41. B. Tourvieille, S. D. Gorman, E. H. Field, T. Hunka-piller, J. R. Parnes, Science 234, 610 (1986); S. L.Weinstein, C. H. June, A. L. DeFranco, J. Immunol.151, 3829 (1993).
42. C. Chang, E. Schaller, S. E. Patterson, S. F. Kwok, E.Meyerowitz, Plant Cell 4,1263 (1992); J. C. Walker,Plant J. 3, 451 (1993).
43. G. De Lorenzo et al., Biochem. Soc. Trans. 22, 394(1994).
44. K. Hammond-Kosack, D. A. Jones, J. Jones, PlantCell 6, 361 (1994).
45. S. Ghosh and D. Baltimore, Nature 344, 678 (1990).46. P. A. Baeuerle, Biochim. Biophys. Acta 1072, 63
(1991).47. Y. T. lpetal.,Cell75, 753(1993).48. D. Hultmark, Trends Genet. 9, 178 (1993).49. C. Hashimoto, K. Hudson, K. V. Anderson, Cell 52,
269 (1988).50. D. S. Schneider, K. L. Hudson, T. Y. Un, K. V. Ander-
son, Genes Dev. 5, 797 (1991); D. St. Johnston andC. Nusslein-Volhard, Cell 68, 201 (1992); A. Heguy,C. T. Baldar, G. Macchia, J. L. Telford, M. Melli, J.
Biol. Chem. 267, 2605 (1992); S. Leung and N. G.Miyamoto, Nucleic Acids Res. 17, 1177 (1989).
51. M. C. Mehdy, Plant Physiol. 105, 467 (1994); A.Levine, R. Tenhaken, R. Dixon, C. J. Lamb, Cell 79,583 (1994).
52. R. Schreck and P. A. Baeuerie, Trends Cell Biol. 1,39 (1991).
53. M. Meyer, R. Schreck, P. A. Baeuerie, EMBO J. 12,2005 (1993); S. Xanthoudakis and T. Curran, ibid.11, 653 (1992).
54. A. Pryor, Trends Genet. 3, 157 (1987).55. J. L. Bennetzen and S. H. Hulbert, Plant Mol. Biol.
20, 575 (1992); T. P. Pryor and J. Ellis, Adv. PlantPathol. 10, 281 (1993).
56. T. Braun, P. R. Schofield, R. Sprengel, EMBO J. 10,1885 (1991).
57. M. Wolfe, Annu. Rev. Phytopathol. 23, 251 (1985).58. B. Keamey and B. J. Staskawicz, Nature 346, 385
(1 990).59. J. T. Greenberg, A. Guo, D. F. Klessig, F. M. Ausubel,
Cell 77, 551 (1994); R. A. Dietrich etal., ibid., p. 565;
G. S. Johal, S. H. Hulbert, S. P. Briggs, Bioessays, inpress.
60. M. Wolter, K. Hollricher, F. Salamini, P. Schulze-Lefert, Mol. Gen. Genet. 239,122 (1993).
61. N. Martini, M. Egen, I. Runtz, G. Strittmatter, ibid.236,179(1993).
62. C. D. Livingstone and G. J. Barton, Comput. Appl.Biosci. 9, 745 (1993).
63. J. D. Thompson, D. G. Higgins, T. J. Gibson, NucleicAcids Res. 22, 4673 (1994).
64. We thank S. P. Dinesh-Kumar, S. Whitham, M.Whalen, J. Glazebrook, F. Katagiri, and K. Ham-mond-Kosack for their helpful comments and criticalreading of the manuscript, and B. Osborne for help inthe preparation of the figure with the amino acidalignments. Supported by grants from NSF and theU.S. Department of Energy (B.J.S.), NIH (F.M.A.),USDA (B.J.B.), CSIRO (J.G.E.), and the GatsbyCharitable Foundation, UK Biotechnology and Bio-logical Sciences Research Council (BBSRC), andEEC Bridge Program (J.D.G.J.).
The Ethylene SignalTransduction Pathway in Plants
Joseph R. Ecker
Ethylene (C2H4), the chemically simplest plant hormone, is among the best-characterizedplant growth regulators. It participates in a variety of stress responses and developmentalprocesses. Genetic studies in Arabidopsis have defined a number of genes in the ethylenesignal transduction pathway. Isolation of two of these genes has revealed that plantssense this gas through a combination of proteins that resemble both prokaryotic andeukaryotic signaling proteins. Ethylene signaling components are likely conserved forresponses as diverse as cell elongation, cell fate patterning in the root epidermis, and fruitripening. Genetic manipulation of these genes will provide agriculture with new tools toprevent or modify ethylene responses in a variety of plants.
The simple gas ethylene is an endogenousregulator of developmental adaptations inhigher plants (1). Exposure to ethylene canproduce a myriad of effects on plant growth,development, and physiology, most notablythe ripening of fruits, inhibition of stem androot elongation, promotion of seed germi-nation and flowering, senescence of leavesand flowers, and sex determination. Howthis simple olefin evokes such a diversearray of physiological processes has been acentral question in ethylene research.
The biosynthesis of ethylene is stimulat-ed prior to several developmentally pro-grammed senescence processes and in re-sponse to environmental insults such as me-chanical trauma and pathogen infection (2,3). As a result of biochemical analysis, theroute of ethylene synthesis (the Yang Cy-cle) is now largely understood (4, 5). Therate-limiting step is the conversion of S-adenosyl-L-methionine (SAM) to 1-amino-cyclopropane-1-carboxylic acid (ACC),which is catalyzed by ACC synthase. Theenzyme ACC oxidase converts ACC to
The author is in the Plant Science Institute, Department ofBiology, University of Pennsylvania, Philadelphia, PA19104-6018, USA.
ethylene, carbon dioxide, and cyanide.ACC oxidase is constitutively present inmost tissues, but its synthesis is increasedduring fruit ripening in tomato. The genesthat encode ACC synthase and ACC oxi-dase have been cloned and characterizedfrom many plant species (5, 6). ACC syn-thase is encoded by multigene families in allspecies examined, and individual gene fam-ily members are transcriptionally activatedby a variety of inducers. Environmentalstresses (physical, chemical, and biological)and hormonal signals, such as auxin, cyto-kininin, and even ethylene itself, stimulatesynthesis of the ACC synthase enzyme,thereby providing a means for autoregula-tion of its production. Although tremen-dous progress has been made since 1989,questions still remain regarding the com-plex regulation of ethylene biosyntheticgenes. However, it is clear that geneticmanipulation of the ACC synthase andACC oxidase genes by expression of anti-sense RNA (7) will provide a simple meansto control the ripening of fruits in a varietyof plants (4, 8).
By contrast, biochemical approaches to-ward dissection of the mechanisms by
