Perception of <i>Rhizobium</i> nodulation factors by tomato cells and inactivation by root chitinases

Article

Reference

Perception of Rhizobium nodulation factors by tomato cells and inactivation by root chitinases

STAEHELIN, Christian, et al.

Abstract

The bacterial genera Rhizobium and Bradyrhizobium, nitrogen-fixing symbionts of legumes, secrete specific lipo-chitooligosaccharides that induce the formation of nodules on their host plants. When preparations of such nodulation-inducing factors (Nod factors) were added to suspension-cultured tomato cells, a rapid and transient alkalinization of the culture medium occurred. Lipo-oligosaccharide preparations from Rhizobium or Bradyrhizobium treated with flavonoids, known inducers of Nod factor synthesis, were up to 100 times more potent in inducing alkalinization than the ones from untreated bacteria. The activity was absent from preparations of the mutant strain Rhizobium sp. NGR234 delta nodABC, unable to produce any Nod factors. Preparations of Nod factors from various bacteria as well as individual, highly purified Nod factors from Rhizobium sp. NGR(pA28) induced alkalinization in the tomato cell cultures at nanomolar concentrations. This demonstrates that Nod factors can be perceived by tomato, a nonhost of rhizobia. Using the alkalinization response as a sensitive bioassay, Nod factors were found to be inactivated by [...]

STAEHELIN, Christian, et al. Perception of Rhizobium nodulation factors by tomato cells and inactivation by root chitinases. Proceedings of the National Academy of Sciences, 1994, vol. 91, no. 6, p. 2196-2200

DOI : 10.1073/pnas.91.6.2196 PMID : 8134372

Available at:

http://archive-ouverte.unige.ch/unige:123321

Disclaimer: layout of this document may differ from the published version.

1 / 1

Vol. 91, pp. 21%-2200, March 1994 Plant Biology

Perception of Rlhizobium nodulation factors by tomato cells and inactivation by root chitinases

(chemoperception/chito-oLgosaccharides/sigaflng/symblosis)

CHRISTIAN STAEHELIN*, JosE GRANADO*, JOACHIM MULLER*, ANDRES WIEMKEN*, ROBERT B. MELLOR*, GEORGFELIXt, MARTIN REGENASSt, WILLIAM J.

BROUGHTONt,

AND THOMAS BOLLER*t

*Botanisches Institut der UniversititBasel, Hebelstrasse 1,CH-4056Basel,Switzerland;tFriedrichMiescher-Institut, P.O.Box2543, CH4002 Basel, Switzerland; andtLaboratoiredeBiologieMol6culairedes PlantesSup6rieures,Universitt deGen've,1chemin del'Impdratrice,CH-1292

Chamb6sy,Switzerland

CommunicatedbyHarold J.Evans, October 12, 1993

ABSTRACT The bacterial genera Rhizobium and

Bradyrhizobium,nitrogen-fixing symbionts ofleumes, secrete specificlipo-chitooligosaccharidesthat induce theformation of nodules on their host jilants. When preparations of such nodulation-inducing factors(Nod factors)were addedto sus- pension-cultured tomatocells, arapidand transient alkalin- izationof the culturemediumoccurred. Lipo-oligosaccharide preparations from RhizobiumorBradyrhizobiumtreated with flavonoids,known inducersof Nodfactor synthesis,^wereupto 100 timesmorepotent in ducingalkalinizationthan the ones from untreated bacteria. The activit wasabsent from prep- aratio ofthemutantstrainRhizobiumsp.NGR234AnodABC, unabletoproduceanyNodfactors.Preparationsof Nod factors fromvariousbacteriaaswellasindividual, highly purifiedNod factors fromRhizobiumsp.NGR(pA28)inducedalkliniation inthetomatocellculturesatnanomolar concentrations. This demonstrates that Nod factors canbeperceivedbytomato,a nonhost of rhizobia. Using the alkalinization response as a sensitivebioassay, Nodfactorswerefound^tobe inactivatedby plant chitinases. Root chitinases purified from different le- gumesdiffered in theirpotentialtoinactivate differentlysub- stituted Nodfactorsproduced byRhizobium sp.NGR(pA28).

ThisIndicates that thespecificityof thebacterium-host plant interaction may bedue,atleastinpart, todifferential inacti- vationof Nodfactorsbyrootchitinases.

Theinfection oflegume rootsbyRhizobiumorBradyrhizo- bium resultsin the formationofroot nodules inwhichthe bacteriamayfixnitrogen (1, 2).Duringtheinfectionprocess, flavonoidssecreted bytheplant induce thetranscription of bacterial genes that are required for modulation, the nod genes (3). The products of these nod genes, in turn, are involved in thesynthesis ofthe so-called nodulationfactors (Nod factors), a family of signal molecules consisting of lipo-chitooligosaccharides-i.e., f1,4-linked oligomers of N-acetylglucosaminewithafattyacidmoietyonthenonre- ducing end (3-6). When appliedto hostplants, these Nod factorsinducerapidmembranedepolarization (7), root-hair deformation (3-5), expressionofearlynodulin genes (5, 6), and mitosis in theroot cortex(4-6)and thus appeartoplay akeyrolein the establishment of thenodule symbiosis.Itis generallythoughtthatagivenhostplant has specificrecep- tors for specific Nod factors, accepting a bacterium as a symbiontonly if itrecognizestheNodfactorssecretedby the latter(3-6).

Recognitionof Nod factorsbyhostplantscanbeseen as onevariationonthe themeofperceptionof microbialsignals by plants. Anothervariation is therecognitionof molecules derivedfrommicrobialpathogens, so-called elicitors, which

cause induction ofa defense response (8, 9). Suspension- culturedplantcells arewidely used tostudy elicitorrecog- nitionand signaltransduction(8, 9). Inthesemodelsystems, alkalinization of the growth medium isaparticularly early reactiontovariouselicitor preparationsand canbe used as an early marker for elicitorperception (9-13). Herewe report that Nod factors elicit a similar alkalinization response in suspension-cultured cells oftomato, anonhostofrhizobia, and that the alkalinization-inducing activity of variously substituted Nod factors is differentially inactivated by le- gumechitinases.

MATERIALS ANDMETHODS

Bacterial Strains and Growth Conditions. Rhizobium sp.

NGR(pA28) was grown as described (14). Rhizobium sp.

NGR234 (14), Rhizobiumsp.NGRAnodABC(14), Rhizobium leguminosarum biovar (bv.) viciae RBL5560 (15), Bradyrhizobiumjaponicum 61-A-101 (16),Rhizobium meli- loti 102F34 (17), and Rhizobium phaseoli RCR3622 (from P. R. Hirsch, Rothamsted Experimental Station, Har- penden, U.K.)weregrownat27°Con arotaryshakerat130 rpminaminimal medium(14) with thefollowing carbon and nitrogensources: R.sp.NGR234,R.sp.NGRAnodABC,and R. meliloti, 0.12 Msodiumsuccinateand 1 mM ammonium chloride;^R.leguminosarumand R.phaseoli,8mMmannitol and 5 mM KN03; B.japonicum, 50 mM glycerol, 1 mM mannitol,and 5 mMKNO3.GrowthmediaofR.leguminosa- rum,R.phaseoli,andB.japonicumcontained, in addition,2 mgofyeastextractperml, pretreatedasfollowstoeliminate inducers ofthealkalinizationresponse:theyeast extract(0.1 g/ml) wasincubated with 0.1 mg ofchitinase perml from Serratia marcescens (Sigma) in 100 mM phosphate (K+) buffer at pH 6.0, autoclaved, centrifuged, and partitioned severaltimesagainst fresh 1-butanol (1:1 ratio).

Preparation of Nod Factors. Lipo-oligosaccharides were extracted 24 h (Rhizobium strains) or48 h (B. japonicum) after theaddition ofappropriateflavonoidstoexponentially growing bacteria, using published procedures (14, 15, 18).

Briefly,cultureswerecentrifugedat27,000 x gfor20min, and the supernatant was extracted with 0.25 volume of 1-butanol by shaking for 2 min. The 1-butanol phase was evaporated, and the residue was taken up in water and extracted withanequalvolume ofethylacetate.The aqueous phase was loaded onto C18 SepPac cartridges (Millipore).

Afterwashing with water, theadsorbed material containing lipo-oligosaccharideswas elutedwith methanol anddriedin aSpeed-vacevaporator.The residuewasdissolved inwater containing 0.01% 3-[(3-cholamidopropyl)dimethylammonio]- propanesulfonate (Fluka), filtered through a 0.22-pum Ul- trafree-MC filter(Millipore),and storedat-20°C. Individual

Abbreviation:chitotetraose,N,N',N",N"'-tetraacetylchitotetraose.

2196 Thepublicationcostsofthis articleweredefrayedinpartbypagecharge payment. This article must therefore be hereby marked "advertisement"

in accordance with 18 U.S.C. §1734solelytoindicate this fact.

Proc.Natl. Acad. Sci. USA 91 (1994) 2197

A

5.3 1

- 5

I75.

00

0 5.1

m0.

5.0 I-

5.3 F sp. NGR234

R.sp. NGR Anod ABC

I0

5.2 p

5.1 1-

5.0 I

-2 0 2 4 6 8 10

Time(min)

C

-2 0 2 4 6 8 10

Time(min)

D

I

^{ApH 0.1}

icontrol

lipo-chitooligo- saccharides

K-252a

L-

0)

0)

4.9-

4.9-

-4 -2 0 2 4 6 8

Time(min)

chitotetraose

+chitotetraose

I I l I l l

0 10 20 30 40

Time(min)

FIG. 1. Alkalinization of the culture medium oftomatocells inresponsetopreparations of Nod factors. (A) Comparison oftheresponse toequivalentamountsoflipo-oligosaccharide preparationsfrom R.sp.NGR234and themutantstrainR.sp.NGRAnodABC.(B) Responseto various concentrations of the Nod factor preparation fromR.sp.NGR234 addedatzerotime. (C) Effect of K-252a. K-252a (1juM)wasadded 4minbeforethe Nod factorpreparationfrom R.sp.NGR234(correspondingto100 nMlipo-chitooligosaccharides). Control,cells treated with the Nodfactor preparation only. (D) Refractory behavior of thetomatocells. The pH of the growth mediumwascontinually registered intwo separate samples. Arrows, addition of 1 nMchitotetraose, of theNod factorpreparationfrom R. sp. NGR234(corresponding to 100 nM lipo-chitooligosaccharides), ^orof 10 jug ofxylanaseperml.

Nod factors from theoverproducing strain R.sp.NGR(pA28)

wereisolated andpurified by HPLCasdescribed(14).

Quantification of Lipo-Chitooligosaccharides. Lipo- chitooligosaccharideswerequantified by measuringN-ace- tylglucosamine after enzymatic hydrolysis,asdescribed for chito-oligosaccharides (19): preparations weretreated with 25 mg ofdialysed snail gut enzyme per ml (cytohelicase fromIBF, Villeneuve-la-Garenne, France)in50mMphos- phate (K+)bufferatpH 6.5, 37TC, for 24 h. TheN-acetyl- glucosamine released was determined calorimetrically.

Known amounts of purified Nod factors from R. sp.

NGR(pA28), digested in the same manner, were used as

standards.

Assay and Purification of Root Chitinases. Chitinase was

assayedwith[3H]chitin assubstrate (20), defining1 nkatas

the amountreleasing1 nmol ofN-acetylglucosamine equiv- alentspersatinfinite dilution(19). Chitinaseswerepurified fromcowpea(Vigna unguiculata L.Walp.)cv.Red Caloona

(Rawling Seeds, Orpington, Kent, U.K.), from bean (Phaseo- lus vulgaris L.) cv. Saxa (Samen Wyss, Zuchwil, Switzer- land), and from pea (Pisum sativum L.) cv. Dunkelgrfine Markerbse(Ditzler AG, Mohlin, Switzerland) after exposing plants (3 weeks old)to 10nlofethylenepermlfor 48 h(20).

Plantrootswereground in liquid nitrogen and extracted with 100 mM phosphate (K+) buffer atpH 7.0 (1 ml/g offresh weight). Theextractwascentrifuged (18,000xg,15min)and precipitated with ammonium sulfate (80% saturation). The redissolved precipitatewasloadedonto acolumn ofregen-

eratedchitin (21). Chitin-binding proteins wereelutedwith 100 mM aceticacid, dialyzed,lyophilized, redissolved in 10 mMphosphate (Na+) buffer atpH 6.0 containing 140 mM NaCl, and chromatographed in thesamebufferon aSuper- osel2 column(Pharmacia). Fractions with chitinaseactivity

were pooled, dialyzed, subjected to electrophoresis on a

NaDodSO4/10% (wt/wt) polyacrylamide gel (22), and stained withCoomassiebrilliant blue R250.

5.3 F

5.2 F

L-

0 coo x

I0.

5.1

F-

5.0

Plant

Biology:

Staehelin etal.

_n

Preparation of [3HChito-Oligosaccharides. Water-soluble

[3H]chito-oligosaccharides were prepared from [3H]chitin

with 0.5 nkat of purifiedbeanchitinase asdescribed(21).

Assay of theAlkalnizationResponse. The tomato cellline Msk8wasgrown as afinesuspensionasdescribed (23). For assays,aliquots of2ml, containing z0.3-0.5gofcells,were incubated inopen20-ml vialson a rotatoryshakerat120cycles permin, and the pH of the culture mediumwascontinually registered(13).The maximalpHincrease(ApHmaJ),occurring 2-5minafter applicationof thestimuli,wastakenasameasure of activity (13). Whereindicated, cellsweretreatedwith 1 nM N,N',N"',NI'-tetraacetylchitotetraose (chitotetraose; Bio- Carb,Lund,Sweden), 10Hgofxylanaseperml(fromTricho- dermaviride, Fluka),or 1 ,uM of K-252a(Fluka).

Incubation of Nod Factors with Chitinases. The reaction mixture(150y1)contained100 nMpurified Nod factors from R. sp. NGR(pA28),1%(vol/vol) dimethyl sulfoxide, 10 mM phosphate(Na+)bufferatpH 6.0,10 mMNaCl,4 mMNaN3, and 1 nkat ofpurified chitinase. After incubation at 37°C, sampleswereloadedontoC18SepPaccartridges (Millipore), eluted with 2 ml of methanol, and dried in a Speed-Vac evaporator. The residue was taken up in 5 ulofdimethyl sulfoxide, diluted with45

A4

water,and used forthe assayof thealkalinizationresponse.

RESULTS

Alkalinization Response of Tomato Ceils Induced by Nod Factors. As shown previously, suspension-cultured tomato cells have a sensitive perception system for chito- oligosaccharides andreact to subnanomolarconcentrations of these moleculesbyarapidandtransientalkalinization of thegrowth medium (13).Totestthereaction oftomatocells to Nod factors, lipo-oligosaccharides were isolated from culturesofthebroadhost-range rhizobiumR. sp. NGR234, treated with theflavonoidapigenintoinduce the nodgenes.

1-Butanol extractionfollowedbybatchchromatographyona

C18 SepPac cartridgeyielded about 1 mg (dryweight) ofa

lipo-oligosaccharide preparation per liter of culture. We addedaliquots of thispreparationto tomatocells andfound thatthey induceanalkalinization ofthegrowth medium(Fig.

1A).Similaramountsofapreparation from flavonoid-treated culturesofthemutantR. sp.NGRAnodABC,astrainunable toproduce Nodfactors, did notinduce alkalinization (Fig.

1A), indicatingthat the response depends on Nod factors.

With the preparation from wild-type R. sp. NGR234, con- centrations aslowas 1 ng/mlelicitedameasurable alkalin- ization (Fig. 1B). The initial lag phase and the ApHma depended on the dose; at higher concentrations, the pH startedtoincrease after about1minand reached itspeak after 5min(Fig. 1B).

The tomato cells used are obviously highly sensitive to Nodfactorpreparations, buttheyareeven moresensitiveto underivatized, hydrophilic chitin fragments (13). Totestfor the presence of such hydrophilic factors, the Nod factor preparationofR.sp.NGR234waspartitioned again between 1-butanoland water andchromatographedover aC18SepPac cartridge. The alkalinization-inducing activity was mostly foundin the butanol phase, was practically completelyad- sorbed bythe C18 SepPac material, andcould be quantita- tivelyelutedfromthecartridge by methanol (Table 1). As a control,apreparationof[3H]chito-oligosaccharideswassub- jectedtothesameprocedures. Inthis case, allradioactivity remained in thewaterphase, and allappearedin the flow- through of the C18 SepPac cartridge (Table 1). Thus, the alkalinization-inducingsubstances in the Nod factorprepa- rations arelipophilic.

Thealkalinizationresponseoftomato cellstochitin

frag-

ments canbeblockedbytheproteinkinaseinhibitor K-252a (13), a substance that inhibits various other responses to

Table 1. Partitioningof thealkalinization-inducingactivityof a Nod factorpreparation fromR. sp. NGR234 andofchito- oligosaccharidesduring butanol extraction and

chromatographyon aCisSepPac cartridge

Alkahnization- [3H]Chito- inducing activity, oligosaccharides, Fraction % initial concentration cpm

Waterphase* 19 2428

1-Butanolphase* 81 15

Flow-throught 1 2556

Methanoleluatet 80 12

*Phasesafter 1-butanolextraction.

tFromaC18SepPac cartridge.

elicitorsaswell(10, 24). K-252aalsopreventedthealkalin- ization response elicited by addition of the Nod factor preparations(Fig. 1C). Thissuggeststhat thepH increase is notduetosimple buffer effectsor tononspecific membrane leakage but is triggered by a process involving protein kinases.

Totestif theresponsestotheNod factorpreparationsand tochitinfragments wererelated,wemadeuseof the refrac- tory behavior ofthe tomato cells to repeated stimulation:

cells stimulatedoncewithchitotetraoseareunresponsiveto asecondstimulationwithchitinfragments forupto6 hbut show a full alkalinization response when exposed toa dif- ferent stimulus, xylanase (13). As shown in Fig. ID, cells treated firstwith chitotetraose did notrespondto asubse- quent treatment with the Nod factorpreparation and vice versa.Inbothcases,thecellswerestillsensitivetoxylanase (Fig.ID).Thisstrongly indicates thattomatocellsrespondto the Nod factor preparation in the same way as to chitin fiagments(13).

Dose-response curves were established for individual, highlypurifiedNod factors fromR.sp.NGR(pA28),usingthe

ApHn

as a measure of the response (Fig. 2A). Differently substituted Nod factors fromR. sp. NGR(pA28)were simi- larly active in^oursystem. Concentrations of<1 nMcaused ameasurable alkalinization, and doses of =10 nMinduced the response half-maximally (Fig. 2A). Similar dose- responsecurves were obtained for mixturesof Nod factors

0

EA E

.0

-000.

Cua

0)C

100

801

601

40

20]

A

_1010aa ₀

B ₁₀ oOG

₀₀

0 A

0

o 0

@0 0

1 101001000 1 10 100 1000 Nodfactors(nM)

FIG. 2. Alkalinization induced by different concentrations of Nod factors, expressed in percent of the maximal response. (A) Dose-response curves for individual, purified Nod factorsfromR.

sp.NGR(pA28), NodNGR-V (MeFuc) (o), NodNGR-V (MeFuc, Ac) (a), andNodNGR-V (MeFuc, S)(e).(B)Dose-response curves for Nod factorpreparationsfrom R. sp. NGR234 (0), R.leguminosarum bv.viciaeRBL5560 (o),R.melioti102F34(A),R.phaseoliRCR3622 (A), and B.japonicum 61-A-101(a).

of

Proc. Natl. Acad.Sci. USA 91 (1994) 2199 Table 2. Yield oflipo-chitooligosaccharidesfrom cultures ofvarious flavonoid-inducedbacterial strains and induction

factor of the alkalinization-inducing activity as compared topreparations from uninduced cultures

Yield oflipo-chito- Alkalinization- Flavonoid Conc., oligosaccharides, inducing activity

Strain added AM nmol/liter (induction factor)

R.sp. NGR234 Apigenin 1 87 10

R.sp. NGRAnodABC Apigenin 1 <1

R.leguminosarumbv. viciae Naringemn 1 15 85

R.meliloti Luteolin 10 6 25

R.phaseoli Naringenin 1 5 6

B.japonicum Daidzein 1 21 100

The induction factor was determined by comparing dose-response curves for equivalent preparations of lipo- chitooligosaccharides from flavonoid-inducedand uninduced bacterial cultures, adjusted toequal volumes perliter of culture medium.

from various Rhizobium and Bradyrhizobium strains (Fig.

2B).

Thealkalinizationresponseoftomato cells can be usedas arapid and sensitive bioassay for Nod factors. We used this assay to testthe impact offlavonoids, the inducers of nod genes, on Nod factorproduction. An alkalinization response was observed with preparations from untreated as well as from flavonoid-treatedbacteria, but, based on dose-response curves, thepreparations from the flavonoid-treated bacteria produced up to 100 times more ofthe stimulatingactivitythan theones from untreated cultures (Table 2).

Nod Factors as SubstratesforPlantChitinases. Chitinases were purified from the roots of ethylene-treated plants, a treatment thatincreased chitinase activity in roots [cowpea, 3-fold; bean, 6-fold;pea, 3-fold(datanotshown)]. Analysis of thepurifiedchitinasesbyNaDodSO4/polyacrylamide gel electrophoresis showed single bands with apparent molecular mass values of =30 kDa (Fig. 3), similar to the size of ethylene-inducedchitinasesfrom bean leaves (20), peapods (21),andazuki bean leaves(25).Differentlysubstituted Nod factors from R. sp. NGR(pA28) were incubated with the purified chitinases, adjustedtoequalcatalytic activitytoward chitin. All chitinases retained at least 50% of the initial activity throughout the incubation period (data notshown).

Afterincubation, samples were chromatographed overaC18 SepPac cartridge and thealkalinization-inducing activity of the lipophilic material, determined by establishing dose- response curves,wasexpressed asapercentage of the initial activity. All chitinases examined were able to reduce the activity of NodNGR-V (MeFuc) in a time-dependent way (Fig. 4A). This strongly indicates that this Nod factor is cleaved by the chitinases. Interestingly, substituted Nod

kDa

97-

66- 45-

31- _

21- 14-

1 2 3

FIG. 3. NaDodSO4/polyacrylamidegelelectrophoresisofchiti- nases(5 nkat)purifiedfromrootsofethylene-treatedcowpea(lane 1), bean (lane 2), and pea (lane 3). Positions of molecular mass markersareindicated.

factors-namely, NodNGR-V (MeFuc, Ac), which has an additionalacetyl group (Fig. 4B), and NodNGR-V (MeFuc, S), substituted with a sulfate group (Fig. 4C)-were resistant toinactivationby bean or pea chitinases but not by cowpea chitinase.Thus, chitinases from legumes may differintheir capacitiestohydrolyze variously substituted Nod factors.

DISCUSSION

Nod factors, when added to tomato cell cultures, stimulate a rapid, transient alkalinization of the medium, comparable to the oneelicited by chitin fragments (13). This can be usedas a simple, sensitive bioassay for the Nod factors of the rhizobia. The low, butdetectablepH-stimulatingactivity of cultures prepared without flavonoids may be due to Nod factors thatareproduced in the absence of flavonoids (26) or due to unknown lipo-chitooligosaccharides. Based on the identical refractory behavior toward a second stimulation (Fig.1D), tomato cells do not appear to discriminate between Nod factors and underivatized chitin fragments and detect both Nod factors and chito-oligosaccharides with the same perception system, specific for oligomers ofN-acetylglu- cosamine. This perception system mayfunction in therec- ognitionofchitin-containing fungiandotherchitin-containing organisms in general (13). In contrast, the responses of legume roots to Nod factors, such as root-hair curling, are

10

80 60 Cou

60

CD C

0 co 20

N~~ ~ ~ ~ ~ ~ --I

c0 15 30 45 0 15 30 45 0 15 30 45 Time(h)

FIG.4. Sensitivity of Nod factors fromR. sp. NGR(pA28) to treatmentwith different chitinases. Alkalinization-inducing activity was measuredinpreparations of NodNGR-V (MeFuc) (A), Nod- NGR-V(MeFuc, Ac) (B),andNodNGR-V(MeFuc,S) (C)treated for different lengths of time without chitinase (o) or with 1 nkat of purifiedrootchitinase ofcowpea(o), bean(o), orpea(A).

Plant

Biology:

Staehelinetal.

induced only by Nod factors but not by chitotetraose(15).

Thus, legumesmighthave evolvedamoreselectivepercep- tion system that also recognizes the substituents of Nod factors, thereby discerning specificallytheirsymbioticpart- ners. It remains an open question whether plants contain endogenous molecules homologous to chitin fragments or Nod factors that are recognized by related perception sys- tems, as recently postulated on the basis of theactivity ofa purifiedNod factor in carrotembryogenesis (27).

Byusing thealkalinization bioassay, wewereable toshow thatcertain Nod factors are inactivatedby plantchitinases.

Chitinases are often associatedwith defensereactionsagainst bacterialorfungal attack (28, 29).Inaddition,chitinasesmay have endogenous functions and may be involved in plant somatic embryo development (30), perhaps by releasing endogenous factors related to Nod factors (5, 6, 27).

With respect to the interaction with rhizobia, we have previously found in soybean plants that chitinase accumu- lates in the cortex of root nodules(31): this chitinaseappears atarelativelylatestage inthesymbiosisandmay protectthe infectedzonefrom external pathogens inthe effectivesym- biosis or prevent spreading ofpathogenic rhizobia in the ineffective symbiosis.

The present data indicate afurtherpossible role for chiti- nases constitutively present in legume roots-namely, to hydrolyze and inactivateNod factors. This activitymight be importantalready at an early stage of the infection process.

It is particularly intriguing that different legume rootchiti- nasesdiffer in theirabilitytocleaveNod factors fromR.sp.

NGR(pA28).Thus, the hostspecificityofrhizobia may reside notonly in the postulated specificity of Nod factor receptors butmay bedue,atleast inpart,toinactivation of Nodfactors by chitinasesorotherhydrolases,addinga newdimension of complexity to host-symbiont recognition. This possibility has recently been tested in more detail by studying the pattern andkinetics ofcleavageof Nod factors involved in hostspecificity:differently substituted Nod factorsfromR.

meliloti,known todifferentiallyinducesymbiotic responses inMedicago and Vicia, respectively (4),were cleaved dif- ferentially bytheenzymes present inrootsofMedicagoand Vicia(32).

Thisworkwassupportedby the Swiss National ScienceFounda- tion, Grants 31-25730.88 (to R.B.M. and T.B.), 31-27923.89 (to A.W.), 31-30950.91 (toW.J.B.), and 31-26492.89 (toT.B.).

1. Long,S.R.(1989) Cell 56,203-214.

2. Caetano-Anollds, G. & Gresshoff, P. M. (1991) Annu. Rev.

Microbiol.45, 345-382.

3. Fisher, R. F. & Long, S.R. (1992) Nature (London) 357, 655-660.

4. D6nari6,J., Debelld,F. &Rosenberg, C. (1992)Annu. Rev.

Microbiol. 46, 497-531.

5. Spaink,H.(1992) Plant Mol. Biol. 20,977-986.

6. Vjn, I., das Neves, L,van Kammen, A., Franssen, H. &

Bisseling,T. (1993)Science 260, 1764-1765.

7. Ehrhardt,D.W., Atkinson,E. M.& Long, S. R.(1992)Sci- ence 256,998-1000.

8. Dixon, R. A. & Lamb, C. J.(1990)Annu. Rev.PlantPhysiol.

Plant Mol. Biol. 41, 339-367.

9. Scheel,.D.&Parker,J. E.(1990)Z.Naturforsch. C:Biosci.45, 569-575.

10. Felix, G., Grosskopf,D.G., Regenass,M.& Boiler, T. (1991) Proc.Natl. Acad. Sci. USA 88, 8831-8834.

11. Mathieu, Y., Kurkdjian, A., Xia, H.,Guern, J., Koller, A., Spiro,M.D.,O'Neill, M., Albersheim,P.&Darvill,A. (1991) Plant J. 1, 333-343.

12. Wei Z.-M., Laby, R.J.,Zumoff, C.H., Bauer, D. W., He, S.Y.,Collmer,A.&Beer, S. V.(1992) Science257,85-88.

13. Felix, G., Regenass,M.&Boller,T.(1993)'PlantJ.4,307-316.

14. Price, N. P. J.,Relic, B., Talmont, F.,Lewin, A.,Prom6,D., Pueppke, S.G., Maillet, F., Ddnari6, J., Prom6, J.-C. &

Broughton, W. J. (1992) Mol. Microbiol.6,3575-3584.

15. Spaink, H. P., Sheeley, D. M., van Brussel, A. A.N., Glushka, J.,York,W.S.,Tak,T.,Geiger, 0.,Kennedy, E.P., Reinhold, V. N. & Lugtenberg, B. J. J. (1991) Nature (Lon- don)354,125-130.

16. Stripf, R. & Werner,D.(1978)Z.Naturforsch. C: Biosci.33, 373-381.

17. Foug~re,F.& LeRudulier,D. (1990)J. Gen.Microbiol. 136, 157-163.

18. Roche, P., Lerouge, P., Ponthus, C. & Prom6, J.-C. (1991)J.

Biol.Chem.266,10933-10940.

19. Boiler,T.& Mauch,F.(1988)Methods Enzymol. 161,430-435.

20. Boller, T.,Gehri, A., Mauch, F.& Vgeli,U.(1983)Planta 157,22-32.

21. Mauch, F.,Hadwiger,L. A. & Boller,T.(1988)Plant Physiol.

87, 325-333.

22. Laemmli,U. K.(1970)Nature(London)227,680-85.

23. Felix,G.,Grosskopf, D.G., Regenass, M., Basse, C.W. &

Boiler,T.(1991) Plant Physiol. 97, 19-25.

24. Grosskopf,D.G., Felix, G.&Boller,T.(1990) FEBS Lett.275, 177-180.

25. Ishige,F., Mori,H.,Yamazaki, K. & Imaseki, H. (1993) Plant CellPhysiol. 34, 103-111.

26. Spaink,H.P., Aarts, A.,Stacey, G., Bloemberg, G. V.,Lug- tenberg,B.J. J.&Kennedy,E. P.(1992) Mol.Plant-Microbe Interact. 5, 72-80.

27. DeJong,A.J., Heidstra, R., Spaink, H.P., Hartog, M. V., Meijer, E.A.,Hendriks, T.,LoSchiavo, F.,Terzi, ., Bis- seling, T.,vanKammen,A.& DeVries, S. C. (1993) Plant Cell 5, 615-620.

28. Meins, F., Jr., Neuhaus, J.-M., Sperisen, C. &Ryals,J. (1992) inGenesInvolvedinPlantDefense, eds. Boiler, T. & Meins, F.,Jr. (Springer,Vienna),pp. 245-282.

29. Collinge, D. B., Kragh, K.M., Mikkelsen, J. D., Nielsen, K.K.,Rasmussen, U.&Vad,K.(1993) Plant J. 3, 31-40.

30. DeJong, A.J.,Cordewener, J., Lo Schiavo, F., Terzi, M., Vandekerckhove,J., vanKammen,A.&De Vries,S.C. (1992) PlantCell4, 425-433.

31. Staehelin, C., Muller, J., Mellor,R.B.,Wiemken,A. &Boiler, T.(1992) Planta 187,295-300.

32. Staehelin, C., Schultze, M., Kondorosi, A., Mellor, R. B., Boller,T.&Kondorosi,A. (1994)Plant J. 5, in press.