Soybean Resistance Genes Specific for Different Pseudomonas Syringue Avirwlence Genes Are Auelic, Or Closely Linked, at the RPG1 Locus

Home , Plant disease resistance, Pseudomonas syringae

Soybean Resistance Genes Specific for Different Pseudomonas syringue Avirwlence Genes are AUelic, or Closely Linked, at the RPG1 Locus

Tom Ashfield,* Noel T. Keen,+ Richard I. Buzzell: and Roger W. Innes* *Department of Biology, Indiana University, Bloomington, Indiana 47405, $Department of Plant Pathology, University of California, Riverside, California 92521, and fAgriculture & Agri-food Canada, Research Station, Harrow, Ontario NOR lG0, Canada Manuscript received July6, 1995 Accepted for publication September 11, 1995

ABSTRACT RPGl and RPMl are disease resistance genes in soybean and Arabidopsis, respectively, that confer resistance to Pseudomonas syringae strains expressing the avirulence gene awB. RPMl has recently been demonstrated to have a second specificity, also conferring resistance to P. syringae strains expressing [email protected] we show that alleles, or closely linked genes, exist at the RPGl locus in soybean that are specific for either awB or awR@ml and thus can distinguish between these two avirulence genes.

ESISTANCE displayed by particular plant cultivars genes specific to avirulence genes of both the soybean R to specific races of a pathogen is often mediated pathogen Psgand the tomato pathogen Pst. The inabil- by single dominant resistance genes (R-genes). Typi- ity of Pstto cause disease in any soybean cultivarcan be cally, these R-genes interact with single dominant “avir- explained, atleast in part, by the presence of a battery of ulence” (aw) genes in the pathogen. Such specific in- resistance genes in soybean that correspond to one or teractions between races of pathogens and cultivars of more avirulence genes present in all Pst strains. host plants are the basis of the “gene-for-gene” model There arenow severalexamples of bacterial avrgenes of disease resistance developed by H. H. FLOR over 50 detected by multiple plant species (WHALENet al. 1991; years ago (FLOR1955). This model states that resistance DANGL et al. 1992; FILLINGHAMet al. 1992; RONALDet of a plant cultivar to a specific pathogen race is con- al. 1992; INNESet al. 1993; SIMONICHand INNES1995). trolled by a single dominant resistance gene, the prod- These studies suggest that R-genes sharing the same uct of which specificallyinteracts (directly or indirectly) specificities are present in different plant species. This with the productof a “corresponding” avirulence gene. has been demonstrated genetically for interactions in- Thus, for each avirulence gene in the pathogen, there volving awB (KEEN and BUZZELL1991; INNESet al. is a corresponding resistance gene in a resistant plant, 1993), avrRpml (VIVIANet al. 1989; DEBENERet al. 1991; and resistance is observed only when both genes are FILLINGHAMet al. 1992) and awPph3 (JENNER et al. 1991; present. Often this resistance is associated with a “hy- SIMONICHand INNES 1995). Itis unclear as to whether persensitive resistance response” (HR) thatis visualized this phenomenon represents theconservation of ances- as rapid localized necrosis of plant tissue at theinfection tral R-genes through speciation or whether convergent site. The HR appears to be an important component evolution is responsible. If functionally analogous R- of the defense response in many plant species (GOOD- genes in different species represent the conservation of MAN and NOVACKY1994). ancestral genes during speciation, it seems paradoxical Recently, the “gene-for-gene” model has been ex- that they should be lost (or change specificity) at a high tended beyond race-cultivar interactions to include in- frequency within a species; however, multiple alleles of teractions between plant pathogens and “nonhosts.” differing specificities is a hallmark of R-geneloci For example, the tomato pathogen Pseudomonas syringue (PRYORand ELLIS1993). The cloning of R-genes shar- pv. tomato (Pst) possesses multiple avirulence genes that, ing common specificities will help address this question. when expressed in P. syringue pv. glycinea (Psg), induce Only recently have the first three R-genes specificfor an HR in various cultivars of soybean (KOBAYASHIet al. bacterial avrgenes been cloned.These are the Pto gene 1989). This interaction was shown to be a true gene- from tomato and RPS2 and RPMl from Arabidopsis for-gene interaction when KEEN and BUZZELL(1991) (-TIN et al. 1993; BENTet al. 1994; MINDRINOSet al. established that the resistance response in specific soy- 1994; GRANTet al. 1995). Pto and RPS2 interact with the bean cultivars was controlled by single dominant resis- P. syringue avirulence genes aurPto and avrRpt2, respec- tance genes corresponding to the individual Pst aviru- tively (RONALDet al. 1992; KUNJSELet al. 1993). RPMl lence genes. Thus, soybeancultivars carry resistance displays a dual specificity, responding to both awRpml and avrBand consequently is also known as RPS3(DEBE- Cmrespondzng author: Roger W. Innes, Department of Biology, Indi- ana University, Jordan Hall 142, Bloomington, IN 47405. NER et al. 1991; INNES et al. 1993; BISGROVEet al. 1994; E-mail: [email protected] GRANTet al. 1995). Sequence analysis has revealed that

Genetics 141: 1597-1604 (December, 1995) 1598 T. Ashfield et al.

TABLE 1 Bacterial strains and plasmids used

Bacterial strain/ Description Reference plasmid Strain PsgR 4 PseudomonasglycineaPsgR4 pv.syringm race 4 LONGet al. (1985) (rifamycin resistant isolate) PsgR4 ( avrB) PsgR4 carrying the plasmid pVBO1 INNESet al. (1993) PsgR4 (aurB:: Q) PsgR4 carrying the plasmid pVl3Ol: :C2 INNESet a,Z. (1993) PsgR4 ( avrRpml) PsgR4 carrying the plasmid This paper pVSPGl/avrRprnl Plasmid pvBo1 awBin cloned the vector pVSP61 INNESet al. (1993) pvBo1::sz aurB disrupted by the insertion of an INNESet al. (1993) R fragment cloned in the vector pVSP61 pVSP6 1/ avrrPml avrRpml cloned in the vector pVSP61BISCROVE et al. (1994)

Pto shows homology to known serine-threonine protein Bacterial strains and plasmids are described in Table 1. kinases, suggestive of a role in signal transduction. In Growth of bacteria and inoculumpreparation: Bacterial lawns were grown on King's medium B (KING et al. 1954) contrast, RPS2 and RPMl display no homology to pro- supplemented with the appropriate antibiotics at 30"over- tein kinases but contain leucine-rich-repeats, a putative night. Rifamycin (Sigma) was includedat 100 pg/mland leucine zipper and a potential nucleotide binding do- kanamycin (Sigma) at 50 pg/ml. Bacterial suspensions were main. These motifs are also present in other recently prepared from the lawns in 10 mM MgC12 and diluted to -1 cloned R-genes corresponding to viral and fungal X 10' cfu/ml (an OD600of 0.1) for the HR tests and -5 X lo5cfu/ml for in9lantagrowth analysis. The suspensions were pathogens,but their role in R-gene function is un- used within 4 hr of preparation. known (reviewed by BRICCS 1995;DANGL 1995; INNES HR hand-inoculation tests: Primary leaves were inoculated 1995; ST~~KAWICZet ul. 1995). Neither is it known 2-3 wk after planting. The undersides of the leaves were whether any of these R-gene products interact directly nicked with a razor blade before the inoculum was forced with pathogen-derived elicitors. into theapoplastwith a 1-ml disposable syringe with no needle We have been analyzing the R-genes RPMl and RPGl fitted. The inoculated panelswere scored 20-24 hr after injection. Incompatible (hypersensitive) responses were observed from Arabidopsis and soybean, respectively. Bothgenes as areas of brown sunken tissue. Typically, no macroscopic confer resistance to P. syringae strains expressing the response was seenin compatible interactions at this time, avirulence gene aurB (MUKHERJEEet al. 1966; KEEN and although occasionally mild chlorosis was obsemed. At least BUZZEL1991; INNESet al. 1993). However, it was not five individuals were scored from each recombinant inbred known whether RPGl, like RPMl, also confers resis- family. Each FZ individual was injected twice with each bacterial strain being tested. tance to Psgstrains expressing avrRpml. Here we show In-planta growth analysis: Znplanta bacterial growth analysis that in most soybean cultivars, RPGl is specific only to was conducted essentially as described by BISGROVEet al. aurB. However an R-gene specific to aurRpml is present (1994). Primary leaves were inoculated when they were fully in some cultivars, and this gene isclosely linked, or expanded (2-3 wk after planting). The plants to be inocu- allelic, to RPGI. We also demonstrate that in a soybean lated were vacuum infiltrated with an inoculum containing 10 mM MgCI2, 0.001% Silwet L77 surfactant (Osi Specialties, cultivar responsive to both uur genes, both resistance Inc.) and 5 X 105cfu/ml bacteria. A cork borer was used to specificities are determined either by an allele of RPGl remove leaf-disc samples from the inoculated leaves 0, 2 and or by RPGl and a second closely linked gene. 4 days after inoculation. The bacterial titer in these samples was determined by homogenizing the leaf discs in 10 mM MgC12 and then plating serial dilutions of the homogenate MATERIALS AND METHODS on trypticase soy agar (Becton Dickinson, Cockeysville, MD) Plant lines and growth. All soybean [Glycine max (L.) Merr.] containing 100 pg/ml rifamycin and 50 pg/ml cyclohexa- seed used in this study was propagated at Harrow, Ontario, mide (Sigma). Colonies were counted after 48 hr. Each data- Canada. The Flambeau X Merit recombinant inbred lines point represents the average of four independent samples, were derived from a cross between these two cultivars followed and error bars represent one standard error. All in$Zanta by inbreeding to the F8 generation by single-seed descent. bacterial growth analyses were performed at least twice. All plants for pathogen tests were grown in clay pots (4 Linkage analysis: Map distances in the RI lines were cal- inch diam)containing a soi1:peat:vermiculite:perlite (2:l: culated using the Haldane and Waddington equation p = 0.5:0.5) mix supplemented with osmocote slow-release fertil- R/(2 - 2R), where p is the frequency of recombinant ga- izer. For the first 2-3 wk after planting, the seedlings were metes in a single meiosis and R is the proportion of recom- grown in a glasshouse. A photoperiod of 216 hr was main- binant individuals. The standard error of p (sp) was calcu- tained with supplementary lighting when required. The day lated using the formula st, = the square root of p(1 - p)/ before inoculation, plants were transferred lo a growth room n, where n is the number of RI lines examined (ALIARD (16-hr photoperiod, 180 microeinsteinsm . s", 22"). 1956). p and 4 were converted to centimorgans using the Bacterial Blight Resistance 1599

TABLE 2 Resistance of soybean cultivars to Pseudomonas syringae pv. &cinea race 4 expressing amB or avrRpml

Cultivar/line Pedigree or origin

H/H = hypersensitive resistant to awB/hypersensitive resistant to awRpml Coles" Hark X (Provar X Disoy X Magna) Hark" Hawkeye X Harosoy Hawkeye Mukden X Richland Mukden Hsiao Chin Huang Tou from China Norchief Hawkeye X Flambeau

H/S = hypersensitive resistant to awB/susceptible to avrRpm1' AK (Harrow) Selected from AK (from China) Blackhawk Mukden X Richland Capital No. 171 X AK (Harrow) Clark Lincoln (2) X Richland Evans Merit X Harosoy Harcor Corsoy (2) X Harosoy 63 Harosoy Mandarin-Ottawa (2) X AK (Harrow) Merit Blackhawk X Capital Provar Harosoy X Clark Richland PI-70.502-2 (from China)

S/H = susceptible to avrB/hypersensitive resistant to awRpml'

Disoy [Mandarin (Ott.) X Kanro] X (Richland X Jogun) Flambeau Wisc. 839-14 Grande Anoka X Magna Jogun Shirobana from Korea Kanro From Korea Magna [Mandarin (Ott.) X Jogun] X [Mandarin(Ott.) X Kanro] Vinton Hark X (Provar X Disoy X Magna)

S/S = susceptible to awB/susceptible to awRpmld Bonminori Japan PI-360.835 from CNS ProbablyChina Nanking from Higan Japan Higan Mame from Peking From China Raiden From Japan OX615 Harcor (2) X Raiden OX735 Coles X OX615 Resistance/susceptibility was determined by flooding small areas of the leaf with P. syn'ngae pv. glycinea race 4 expressing aurB or awRpml. Resistance reactions were scored after 24 hr and disease symptoms monitored after 3 days. H, hypersensitive resistance (HR) response; S, susceptible response. In some cultivars/lines, the HR response to awRpml was consistently weak. *Forty tested cultivars/lines not shown. Fifteen cultivars/lines not shown. Thirty-six cultivars/lines not shown.

Kosambi function as described by KOORNEEF and STAM soybeancultivars expressing RPGl. Onehundred (1992). The standard errors on map distances in the F4 twenty soybean cultivarsand lines were hand inoculated families were calculated using the equation sp = the square with race 4 containing avrB [PsgR4(awB)] or root of (4 - p') /4n, where n is the numberof F2 individuals Psg examined (ALLARD 1956). avrRpml [ PsgR4 (awRpml)3 and scored for HRs (PsgR4 is virulent on all tested soybean cultivars). Only 5 of the RESULTS 55 cultivars responding to awB (and so carrying RPGl ) also responded to awRpml (Table 2). These were Muk- Soybean can distinguish between wrB and mwl: den and four cultivars tracing to it.However, many The observation that theArabidopsis RPMl gene is spe- cultivars lacking a functional RPGl gene did respond cific for both avrB and awRpml prompted us to deter- hypersensitively to awqml. Cultivars were also identi- mine whether awRpml, from the nonhost pathogen fied that did not respond to either awB or awRpml. P. syringae pv. maculicola (Psm), could be detected by Based on pedigree analysis of the cultivars and lines 1600 Ashfield T. et al.

FIGURE 1.-Hypersensitive responses displayed by soybean leaves after interveinal injectionwith Pseudomonas syringaepv. glycinea race 4 expressing awB or awRpml. The primary leaves of 2-Swk-old soybean plants of cultivars Merit, Flambeau, Norchief and Peking were hand inoculated with PsgR4(uwB), PsgR4(awB:R) and PsgR4( awQml). The leaves were photographed after 24 hr.

(Table 2, and others not shown), cultivar reactions were growth is reflected in attenuated disease symptoms in heritable, that is, each cultivar had at least one parent the infiltrated plants (data notshown) and occasionally displaying the same reaction. The cultivars Merit (awB led to Peking being scored as avrRpml responsive dur- responsive), Flambeau (uwRpml responsive); Norchief ingthe initial cultivar screen. Interestingly, the (responsive to aznB and aw8pml) and Peking (respon- PsgR4( UWB)strain displayed a small, but statistically sig- sive to neither) were chosen as representative of the nificant, increase in growth over PsgR4(awB:n) in the four classes and selected for further study. The macro- avrB susceptible cultivars Flambeau (t= 6.23, P = 0.001 scopic responses of these four cultivars to avrB and on day 4) and Peking (t = 3.38, P = 0.01 on day 4). This avrRpml are shown in Figure 1. observation suggests that uwB has a role in virulence in To confirm that the HR tests accurately reflected the compatible interactions. resistance specificitiesof thefour cultivars, in-planta Resistance to awB and aw&ml resides at the RPG1 bacterial growth was monitored (Figure 2). In all four locus: Soybean resistance to Psg race 4 expressing aurB cultivars, the control PsgR4(uvrB::R) strain, which car- has been shown previously to be inherited as a single ries aurB disrupted with a R fragment, multiplied 100- dominant Mendelian trait (KEEN and BUZZELL1991). 1000-fold over 4 days. The growth of PsgR4(awB) in The locus responsible has been designated RPG1 Merit and Norchief and PsgR4(aurRpml) in Flambeau (MUKHERJEEet al. 1966). To determine whether resis- and Norchief was severely restricted, reaching a level tance to avrRpml is alsoinherited in a monogenic fash- 50-100-fold less than thePsgR4(awfk:R) control strain. ion, the resistance specificities of 95 recombinant in- PsgR4 (aw8pml) consistently multiplied to a lower level bred lines (RILs) derived from a cross between the in Peking than the control strain. This difference was cultivars Flambeau and Merit were determined. Rns significant (t-test: t = 3.16 P = 0.02) at day 2 but not were chosen for this study because the R-gene specific at day 4 (t= 1.43, P = 0.20), and was reproducible over to avrRpml was found to be incompletely dominant, three replicates, suggesting a very weak resistance gene and we were unable to reliably distinguish individuals specific for avr8pml in this cultivar. This reduced heterozygous for this R-gene from homozygous suscep Bacterial Blight Resistance 1601 I I I I 1 7 s

d 1 3 4 5 0 1 3 4 5 diys diys

I I I I I I I1 I I Norchibf Peking - 1

0 1 2 3 4 5 0 1 3 4 5 days dgys FIGURE2.-Growth of Psgrace 4 expressing awB or awRpml in the leaves of four differentsoybean cultivars. Psgrace4 strains carrying the indicated aur genes were vacuum infiltrated into theleaves of 2-3-wk-old soybean plants at a concentration of 5 X lo5 cfu/ml. Strain PsgR4(awB::R) carriesaurB disrupted with a R fragment. At the indicated time points, leaf tissuewas removed with a cork borer and the bacterial titer determined. Each data point represents the averageof four independent samples and the error bars equal one standard error. cfu, colony forming units. tibles in F2 populations. RILs, which are homozygous the RI lines for resistance to PsgR4( avrB). As predicted over most of their genomes, avoid this problem. Resis- from the previous study (KEEN and BUZZELL1991), re- tance/susceptibility to PsgR4 (awRpml) in the Flam- sistance/susceptibility to PsgR4( avrB) segregated 1:l beau X Merit RIL population segregated 1:l (x2= 0.17, (Table 3; x2 = 0.38, P > 0.5), confirming the mono- P > 0.5), indicating the involvement of a single locus genic nature of this resistance. None of the RI lines that differs between these two cultivars (Table 3). displayed resistance to both avrgenes andonly a single Because resistance to awB and awRpml is deter- potential recombinant, susceptible to both PsgR4( avrB) mined by a single gene in Arabidopsis (GRANTet al. and PsgR4(avrRpml), was detected (line RI-61). These 1995), we hypothesized that these two resistance speci- datademonstrate that the resistance specificities for ficities might be controlled by alleles at theRPGl locus avrB and awRpml are closely linked (0.56 f 0.77 cM) in soybean. To determine whether genetic linkage ex- in soybean. To confirm that family RI-61 represents a ists between the locus responsible for resistance to true double susceptible, in$lanta bacterial growth was PsgR4(aw~ml)and RPGl, we therefore also scored monitored for both PsgR4( avrB) and PsgR4( avrRpml). 1602 T. Ashfield et al

TABLE 3 Norchief also resides at, or near, theRPG1 locus: Culti- Segregation of resistance to Psg race 4 expressing aurB or var Norchief displays resistance to Psg race 4 expressing avrRpml in soybean recombinant inbred families derived aurB or awRpml. Because this is analogous to the situa- from a cross between the cultivars Merit and Flambeau tion observed for Arabidopsis accessions expressing RPMl, we hypothesized that a similar dual-specificity No. of families allele might be present at the RPGl locus. To address awB resistant/avrRpml awBsusceptible 44 this hypothesis, allelism tests were conducted between avrRpm l resistant/aurBavrRpmlsusceptible 49 the avrB and aurRpml specificities in Norchief and avrB andresistant awRpml 0 those in Merit and Flambeau, respectively. avrB andsusceptible awRpml 1 Two hundred oneF, individuals derived from a cross families still segregating still families 1 between cultivars Merit and Norchief were scored for Resistance phenotype wasdetermined by flooding leafpan- their resistance to PsgR4(avrB) (Table 4). All plants els with bacterial suspensions at a concentration of 1 X 10' were resistant, demonstrating that the Norchief resis- cfu/ml. Hypersensitive responses were scored 20-30 hr after tance specificity is allelic, or tightly linked (0 2 7.0 cM) , injection. to that in Merit. We were unable to reliably score this family for resistance to PsgR4 (awRpml) (a large excess Both of these strains were virulent on this genotype of the susceptible class was observed), which we attri- (Figure 3). Genotyping of this double susceptible line bute to incomplete dominance of the R-gene relative with four different microsatellite markers revealed only to avrRpml recognition. parental alleles (data not shown); thus we found no Two hundred fourteen F2 individuals from the Flam- evidence of a contaminating soybean genotype in this beau X Norchief family were scored for their resistance line. to PsgR4 (avrRpm1) (Table 4). No susceptible plants The above data are consistent with resistance to avrB were identified (Table 4), indicating close linkage (0 and aurRpml in soybean being mediated by two closely ? 6.8 cM) between the avrRpml specificR-genes in linked genes. However, it is possible that alleles at the Norchief and Flambeau. Resistance/susceptibility to RPGl locus are responsible as we have failed to recom- avrB in this population segregated 3:l (x' = 0.16, P > bine both specificities onto a single chromosome. 0.5), confirming that the resistance to avrB displayed The dual resistance specificity displayed by cultivar by Norchief is mediated by a single dominant locus. We did not analyze an F2 family derived from a cross between Norchief X Peking (doublesusceptible), as the incomplete dominance of the awwml-specific R- gene in Norchief rendered such an analysis noninfor- mative without a large amount of progeny testing. How- ever, theincomplete dominance relative to avrRpml strengthens the data obtained from the Flambeau X Norchief cross, as individuals with a single recombinant chromosome that lacked both the Flambeau and Nor- chief R-genes would likely havebeen scored as susceptible to PsgR4 (awRpm1). These data demonstrate that the dualresistance specificity displayed by Norchief is mediated either by an allele of RPGl or by RPGl and a second RPG gene closely linked to it.

DISCUSSION

1 I I I I In this study we confirm that some soybean cultivars 0 2 4 respond to awwml (DANGLet al. 1992). Furthermore, days we demonstrate that, unlike the situation observed in Arabidopsis, soybean cultivars exist that can distinguish FIGURE3.-Growth of Psg race 4 expressingawB or between awB and avrRpml. Cultivars also exist that re- avrRpml in the leaves of recombinant inbred line 61. Psg race 4 strains carrying the indicated aur genes were vacuum spond to both aur genes, or neither. Finally, by follow- infiltrated into the leaves of 2-3-wk-old soybean plants at a ing the segregation of resistance to awB and awRpml concentration of 5 X io5 cfu/ml. Strain PsgR4(awB:f2)car- in recombinant inbred and F2 populations, we demon- ries aurB disrupted with an 0 fragment. At the indicatedtime strate that these resistance specificities map at, or are points, leaf tissue was removed a cork borer and the with tightly linked to, the locus. bacterial titer determined. Each data point represents the RPG1 average of three independent samples and the error bars That soybean cultivars exist that can distinguish be- equal one standard error. cfu, colony forming units. tween awB and awRpml demonstrates that the elicitors Bacterial Blight Resistance 1603

TABLE 4 Segregation of resistance to Psg race 4 expressing awB or awRpml in soybean F4populations

~ ~ ~~ ~ ~ ~ awB avrB awRpml avrBpml population Resistant Susceptible Resistant Susceptible Resistant Susceptible Resistant F2 population

~~ ~~ ~ ~ ~ ~ ~ ~ Flambeau X Norchief 158 56 214 0 Merit X Norchief 201 0 ND" ND" Resistance phenotype._ determined as described in Table 2. a Not determined. produced by these two awgenes must be distinct. This are linked (DEBENERet al. 1991; KUNKEL et al. 1993; was not necessarily to be expected; although these two HINSCHand STASKAWICZ1995; SIMONICHand INNES avr genes appearto have unrelated sequences, they are 1995). The only examples of linked bacterial resistance detected (directly or indirectly) by a single R-gene genes of whichwe are aware are Xu-I O and Xu-4 (YOSHI- (RPMI) in Arabidopsis (BISGROVEet al. 1994; GRANT et MURA et al. 1983),and Ptol and Bo2 (STOCKINGERand al. 1995). It was theoretically possible that awB and WALLING1994). Xu-IO and Xu-4 are rice genes that con- avrRpml directed the production of identical elicitor fer resistance to races of Xanthomonas o9rae and are molecules. RPMl must therefore code for a receptor -27 cM apart. Ptol and Pt02 are tomato genes that able to detect two distinct elicitors or for a component confer resistance to specific races of P. syringae pathovar of a signal transduction pathway used by receptors spe- tomato and are reportedto be within 9 cM of each other, cific for the awB and aurRpml elicitors. but no linkage data have been published. There are no The distinction between the avrB and awRpml elici- reports of complex R-gene loci specificfor bacterial avr tors was also apparent in compatible interactions in genes analogous to the Rpl or L loci. soybean as only aurB appeared to contribute to viru- We identified one potential recombinant family lence. That avrB should act as a virulence factor but among 95 RI lines segregating for resistance to not avrRpml is intriguing because during compatible PsgR4( awB) and PsgR4( avrqml). The simplest expla- interactions between Psm and Arabidopsis, the inverse nation for this observation is that resistance to awB and is true (RITTER and DANGL1995). This is perhaps not aurRpml in soybean is controlled by two tightly linked surprising as avrB originates from Psg (a soybean patho- genes. It is possible, however,that bothresistance speci- gen) and awqml originates from Psm (an Arabidopsis ficities are mediated by alleles and that the double- pathogen). These observations are consistent with pre- susceptible family resulted from intragenic recombina- vious evidence that demonstrated that bacterial viru- tion, unequalcrossing over, or transposon-induced mu- lence factors can be host specificin their action tation. These processes have been proposed to account (SWARUPet al. 1991; DE FEWERet al. 1993; RITTER and for the recovery of double-susceptible progeny from DANGL1995). individuals heterozygous for distinct "alleles" at the L Our results indicate that soybean resistance to Psg locus in flax (ISLAMet al. 1989). The dual specificity strains expressing awB or avrRpml is mediated by alleles displayed by the soybean cultivar Norchief could be of RPGl or by RPGl and a second closely linked gene. mediated by an RPGl allele able to respond to both When a cultivar-race series exists between a crop plant avrB and avrRpml. This is a plausible hypothesis as it and a fungal pathogen, resistance genes Corresponding has been shown that the Arabidopsis RPMI gene re- to different races of the pathogen are often clustered sponds to both these aw genes, demonstrating that R- either as closely linked genes or as alleles (reviewed by genes may have multiple specificities (BISGROVEet al. PRYORand ELLIS 1993). For example, the maize Rpl 1994; GRANTet al. 1995). locus contains numerous tightly linked R-genes corresponding to specific races of the rust pathogen Puccinia RPMl has recently been cloned (GRANTet al. 1995) swghi (HULBERTand BENNETZEN1991). Incontrast, the and shown to contain motifs conserved in other R-genes available evidence suggests that the multiple resistance corresponding to bacterial, fungal and viral pathogens genes corresponding to races of the flax rust Melampsora (reviewed by BRIGGS 1995; DANGL 1995; INNES1995; lini are alleles at theL locus as it has not been possible to STASKAWICZ et al. 1995).We are now focused on cloning recombine two specificities onto the same chromosome RPGl and the allele/linked gene specific for awqml. (ISLAMet al. 1989). Comparison of these soybean alleles/genes to each Clustering of R-genes specific to bacterial pathogens other and to RPMl may provide valuable information appears to be uncommon,however. The fourpreviously on how specificity is conferred to R-genes. identified R-genesin soybean thatare specific to P. We thank membersof our laboratories and two anonymous review- qringae avirulence genes are not closely linked (KEEN ers for critically reading the manuscript. Excellent technical assis- and BUZZELL1991). Likewise, none of the fourR-genes tance was provided byJoHN DANZER,SANDRA SZERSZEN and PATIUCIA in Arabidopsis specific to P. syringae avirulence genes MOWERY.T.A. acknowledges receipt of a NATO postdoctoral fellow- 1604 T. Ashfield et al.

ship. This work was supported by U.S. Department of Agriculture KOBAYMHI, D.Y., S. J. TAMAK~and N. T. KEEN,1989 Cloned aviru- grant no. 93-37303-9136 to R.W.I. lence genes from the tomato pathogenPseudomonas syringae pv. tomato confer cultivar specificity on soybean. Proc. Natl. Acad. Sci. USA 86: 157-161. LITERATURE CITED KOORNEEF,M., and P. STAM,1992 Genetic analysis, pp. 83-99 in ALLARD, R. W. 1956 Formulas and tables to facilitate the calculation Methods in Arabidopsis Research, edited by C. KONCZ,N. CHUAand of recombination values in heredity. Hilgardia 24: 235-278. J. SCHELL.World Scientific, London. BENT, A. F., B. N. KUNKEI., D.DAHLBECK, K. L. BROWN,R. SCHMIDT, KUNKEL,B. N., A. F. BENT,D. DAHLBECK, R. W. INNESand B. J. et al. 1994 RPSP of Arabidopsis thaliana: aleucine-rich repeat STASKAWICZ, 1993 RPS2, an Arabidopsis disease resistance locus class of plant disease resistance genes. Science 265: 1856-1860. specifymg recognition of Pseudomonas syringae strains expressing BISGROVE,S. R.,M. T.SIMONICH, N. M. SMITH,A. SATTLER and the avirulence gene avrRpt2. Plant Cell 5: 865-875. R. W. INNES,1994 A disease resistance gene in Arabidopsis with LONG,M., P. BARTON-WILLIS,B. J. STASKAWICZ,D. DAHLBECKand specificity for two different pathogen avirulence genes. Plant Cell N. T. KEEN, 1985 Further studies on the relationship between 6: 927-933. glyceollin accumulation and the resistance of soybean leaves to BRIGGS,S. P., 1995 Grand unification in sight. Molecular character- Pseudomonas syringae pv. glycinea. Phytopathology 75 235-239. ization of the components of signalling pathways that mediate MARTIN,G. B., S. H. BROMMONSCHENKEL,J. CHUNWONGSE, A. FRARY, disease resistance is at last providing a unified picture of how M. W. GANAL, et al. 1993 Mapbased cloning of a protein kinase plants fight disease. Cum. Biol. 5: 128-131. gene conferringdisease resistance in tomato. Science 262: 1432- DANGL,J. L., 1995 Piece de resistance: novel classes of plant disease 1436. resistance genes. Cell 80: 363-366. MINDRINOS,M., F. KATAGIRI,G. Yu and F.M. AUSUBEL,1994 The DANGL,J. L., C. RITTER, M.J. GIBBON,I,. A. MUR,J. R. WOOD, et al. A. thaliana disease resistance gene RPS2 encodes a protein con- 1992 Functionalhomologs of the Arabidopsis RPMl disease taining a nucleotide-binding site and leucine-rich repeats. Cell resistance gene in bean and pea. Plant Cell 4: 1359-1369. 78: 1089-1099. DEBENER,T., H. LEHNACKERS,M. ARNoI.D and J. L.DANGI., 1991 MUKHERJEE,D., J. W. LAMBERT,R. L. COOPERand B. W. KENNEDY, Identification and molecular mapping of a single Arabidopsis thali- 1966 Inheritance of resistance to bacterial blight (Pseudomonas ana locus determining resistance to a phytopathogenicPseudome glycinea Coerper) in soybeans (Glycine max L.) . Crop Sci. 6: 324- nus syringae isolate. Plant J. 1: 289-302. 326. DE FEWER, R., Y. Y~Gand D. W. GABRIEL,1993 Gene-for-genes PRYOR,T., and J. ELLIS,1993 The geneticcomplexity of fungal resis- interactions between cotton R genes and Xanthomonas campestris tance genes in plants. Adv. Plant Pathol. 10: 281-305. pv. malvacearum aurgenes. Mol. Plant. Microbe Interact. 6: 225- RITTER, C., and J. L. DANGL,1995 The awRpml gene of Pseudomonas 237. syringae pv. maculicola is required for virulence on Arabidopsis. FILLINGHAM, A. J., J. WOOD,J. R. BEVAN,I. R. CRUTE,J. W. MANSFIELD, Mol. Plant. Microbe Interact. 8: 444-453. et al. 1992 Avirulence genes from Pseudomonas syringae patho- RONALD,P. C.,J. M. SALMERON, M. F. CARLAND and B. J. STASKAWICZ, vars phaseolicola and pisi confer specificity towards both host and 1992 The cloned avirulence gene awPto induces disease resis- non-host species. Physiol. Mol. Plant Pathol. 40: 1-15. tance in tomato cultivars containing the Pto resistance gene. J. FLOR,A. H., 1955 Host-parasite interactions in flax rust-its genetics Bacteriol. 174: 1604-1611. and other implications. Phytopathology 45: 680-685. SIMONICH,M. T., and R. W. INNES,1995 A disease resistance gene GOODMAN,R.N., and A. J. NOVACKY,1994 The Hypersensitive Reaction in Arabidopsis with specificiv for theawPph3 gene of Pseudomonas in Plants to Pathogens. APS Press, St. Paul. syringaepv.phaseolicola. Mol. Plant. Microbe Interact. 8: 637-640. GRANT,M. R., L. GODIARD,E. STRAUBE,T. ASHFIELD,J. LEWALD,et STMKAWICZ,B.J.,D.DAHLBECK,N.KEENandC.NAPoLI,1987Molecu- al. 1995 Structure of the Arabidopsis RPMl gene enabling dual lar characterization of cloned avirulence genes from race 0 and specificity disease resistance. Science 269: 843-846. race 1 of Pseudomonas syringaepv. glycinea.J. Bacteriol. 169: 5789- HINSCH,M., and B. STASKAWICL,1995 Identification of a new Arabi- 5794. dopsis disease resistance locus, RPS4, and cloning of the corre- STMKAWICZ,B. J., F.M. AUSUBEL,B. J. BAKER,J. G. ELLIS and sponding avirulence gene, aurRps4, from Pseudomonas syringaepv. J. D. G. JONES,1995 Molecular genetics of plant disease resis- pisi. Mol. Plant. Microbe Interact. (in press). tance. Science 268: 661-667. HULBERT,S. H., and J. L. BENNETZEN,1991 Recombination at the STOCKINGER,E. J., and L. L. WALLING,1994 Pto3 and Pt04 novel Rpl locus of maize. Mol. Gen. Genet. 226: 377-382. genes from Lycopersicon hirsutum var. glabratum that confer resis- INNES,R. W., 1995 Plant-parasite interactions: is the gene-for-gene tance to Pseudomonas syringae pv. tomato. Theor. Appl. Genet. 89: model outdated? Trend Microbiol. (in press). 879-884. INNES,R.W., S. R. BISGROW.,N. M. SMITH,A. F. BENT,B. J. STASKAWICZ SWARUP,S., R. DE FEWER,R. H. BRLANSKYand D. W. GABRIEL, 1991 et al. 1993 Identification of a disease resistance locus in Arabi- A pathogenicity locusfrom Xanthomonas citri enables strains from dopsis that is functionally homologous to the RPG1 locus of soy- several pathovars of X. campestris to elicit canker-like lesions on bean. Plant J. 4: 813-820. citrus. Phytopathology 81: 802-809. ISLAM,M. R., K. W. SHEPHERDand G. M. E. YO, 1989 Recombina- VIVIAN,A,, G. T. ATHERTON,J. R. BEVAN,I. R. CRUTE,L. A. J. MUR tion among genes at the L group in flax conferring resistance et al. 1989 Isolation and characterization of cloned DNA confer- to rust. Theor. Appl. Genet. 77: 540-546. ring specific avirulence in Pseudomonas syringae pv. pisi to pea JENNER, C., E. HITCHIN,J. MANSFIEID,K. WALTER?,,P. BETTERIDGE,et (Pisumsatiuum ) cultivars, which possess the resistance allele, R2. al. 1991 Gene-for-gene interactions between Pseudomonas syrin- Physiol. Mol. Plant Physiol. 34: 335-344. gae pv. phaseolicola and Phaseolus. Mol. Plant. Microbe Interact. WHALEN,M. C., R. W. INNES,A. F. BENT and B. J. STMKAWICZ,1991 4 553-562. Identification of Pseudomonas syringae pathogens of Arabidopsis KEEN,N. T., and R. I. BUZZELL,1991 New disease resistance genes and a bacterial locusdetermining avirulence on bothArabidopsis in soybean against Pseudomonas syringae pv. glycinea: evidence that and soybean. Plant Cell 3: 49-59. one of them interacts with abacterial elicitor. Theor. Appl. YOSHIMURA,A,, T. W. MEW,G. S. KHuSH and T.OMURA, 1983 Inheri- Genet. 81: 133-138. tance of resistance to bacterial blight in rice cultivar Cas 209. KING,E. O., M. K. WARDand D. E. RANEY, 1954 Two simple media Phytopathology 73: 1409-1412. for the demonstration of phycocyanin and fluorescin. J. Lab. Clin. Med. 44: 301-307. Communicating editor: J. CHORY