PHYSIOL, PLANT, 66: 57S-582, Copenhagen If8«

Clover root exudate contains a particuiate form of the lectin,trifoliin A, which binds Rhizobium trifolii

Georges L. Ituchet, John E, Sherwood, H, Stuart Pankratz and Frank B. Bazzo

Truchet, G. L., Sherwood, J. E., Pankratz, H. S. and Dazzo, F. B. 1985. Clover rootexudate contains a particuiate form of the lectin, trifoliin A, which binds to Rhi-zobium trifolii. - Physiol. Plant. 66: 575-582.

A particuiate form of the lectin, trifoliin A, has been isolated from the root exudateof axenically grown seedlings of white clover (Trifolium repens L. cv. Louisiana No-lin). The majority of trifoliin A active in binding to R. trifolii 0403 was sedimentedfrom root exudate by centrifugation at 15 000 g for 30 min, indicating that it was par-ticuiate. Immunofluorescence, electron and immunoelectron microscopy using anti-body prepared against trifoliin A from seeds, suggested that trifoliin A was associatedwith the particles in root exudate that bound specifically to the acidic capsuiar poly-saccharides of Rhizobium trifolii 0403. Electroo microscopic examination alsoshowed that trifoliin A-colloidal gold conjugates bound to these same particles, in-dicating that they also have affinity for the lectin. The particles could be dislodgedfrom intact seedlings by vigorous shaking in isotonic plant growth medium. Isolatedparticles fractionated by ultracentrifugation through a metdzamide gradient had amean density cf 1.12 g cm~'. These isolated particles retained the ability to bind to R.trifolii 0403 as shown by immunofluorescence using anli-trifoliin A antibody. Sodiumdodecyi sulfate polyacrylamide gel electrophoresis showed that the isolated particlesare a mixture of proteins including one with an approximate molecular weight of tri-foliin A. The protein which electrophoretically comigrated with trifoliin A also re-acted with aoti-trifoliin A antibody by western blot analysis, confirming that it wastrifoliin A. The protein composition of the isolated particles did not reflect the totalprotein composition of tiie root exudate, which was far more complex. These studiesindicate that the majority of trifoliin A which can bind to R. trifolii in root exudate ofwhite clover seedlings is associated with electron-dense particles, and that a majorearly interaction between R. trifolii and clover roots involves the binding of these par-ticles containing trifoliin A to acidic polymers on the bacterial cell surface.

Additional key words - Aggregate, capsule, root hair, symbiosis, Trifolium repens.

G. L. Truchet, Lab. de Biologie de la Differenciation, Univ. D'Aix Marseille, Facidtedes Sciences de Marseille-Luminy L.A. C.N.R.S. 179, F-132SS, Marseille Cedex 2,France; J. E. Sherwood, H. Stuart Pankratz and F. B. Dazzo (reprint requests), Deptof Microbiology and Public Health, Michigan State Univ., East Lansing, Ml, 48S24,USA.

Introduction binding specifically to R. trtfolii, the clover symbiont(Dazzo et al. 1978). This lectin has been isolated from

The i{/ii2o6i«ra-!egume symbiosis involves the specific seeds, detected on and eluted from the root surfaceinfection and nodulation of the host root by the homolo- (Dazzo et ai. 1978), and found in the root exudate ofgous bacterium. The interaction of certain plant lectins young, axenically grown seedlings (Dazzo and Hrabakwith glycosylated receptors oti the host root and the 1981). Trifoliin A in the root exudate and on the rootbacterial symbiont surface has been implicated in this surface of white clover is synthesized in the seedlingspecificity (Bohlool and Schmidt 1974, Dazzo and Hub- root (Sherwood e t a i 1984). Recent studies using a newbell 1975). A clover lectiti, tdfoliin A, is capable of method to isolate clover root hairs suggest that trifoliin

Received 21 May, 1985; revised 5 December, 1985

38 Phpiol. Plant. 66. 1986 575

A is a major protein synthesized by these cells (Gerhoidet al. 1985). Transposon mutations in certain genes en-coding essential nodulation functions in the symbioticplasmid of R. trifotii result in significant reductions inability of these non-infective mutants to bind trifoliin Ain the clover root environment (Dazzo et al. 1985). Lec-tins capable of interacting with polysaccharide receptorson R. leguminosarum and Bradyrhizobium japonicumhave also been detected in pea and soybean root exu-dates, respectively (Diaz et al. 1984, Halverson and Sta-cey 1984). In the latter case, binding of soybean lectin inroot exudate to a slow-to-nodulate mutant strain of B.japonicum restored the nodulation kinetics of this mu-tant to that of the wild type (Halverson and Stacey1985). These findings support the hypothesis that cer-tain lectins in the root environment are involved in spe-cific legume-fi/iizofjium interactions that affect rootnodulation.

Electron-dense particles, bound either to plant roothairs or to encapsulated bacteria, have been detectedwhen Rhizobium or Azospirillum were incubated irrroot envirortments (Dart 1971, Dazzo and Hubbell1975, Umali-Garcia et al. 1980). Other studies haveshown that a majoi morphological feature of R. trifoliiin the clover root environment of Fahraeus slide cul-tures, or in the presence of concentrated root exudate,is the rapid binding of these electron-dense particles toacidic polymers surrounding the bacterial cells (Dazzoet al. 1982, 1984). These particles are of plant originsince they were found in concentrated root exudate ofaxenically grown seedlings. In this study, we have exam-ined these particles on root hair tips, isolated theseplant-derived particles which bind to R. trifolii 0403 iothe rhizosphere of white clover, and characterized someof their ultrastructural, physical aod biochemical prop-erties. A preliminary report of this work has been pre-sented (Sherwood et al., Abstr. Annu. Meeting Am.Soc. Microbiol. 1982, K79, p. 147).

Abbreviations - FITC, fluorescein isothiocyanate; G/R/U, glu-taraldehyde/ruthenium red/uranyi acetate; IgG, immuooglobu-liii class G; PBS, phosphate buffered saiine: SDS-PAGE, so-dium dodecyl sulfate-polyacrylamide gel eiectrophoresis;TEM, transmission electron microscopy.

Materials and methods

Rhizobium trifolii 0403, obtained from the RothamstedExperimental Station, Harpenden, U.K., was grown ooBill agar plates (Dazzo 1982) for 5 days at 30°C. Encap-sulated cells in the 5-day-old cultures were obtained bycentrifugation in 10 mM PBS (pH 7.2) at 6000 g for 20min. The soft pellet was washed with filter-sterilizedFahraeus nitrogen-free medium (Dazzo 1982) and cen-trifuged under the same conditions.

White clover seedlings {Trifolium repens L. cv. Loui-siana Nolin) were germinated from surface-sterilizedseeds on Fahraeus medium agar plates. Sterile root exu-date was obtained from seedlings grown at 22°C through

agar blocks and stainless-steel wire supports above ni-trogen-free Fahraeus medium (Dazzo and Hrabak1981), and contained less than 1 [ig protein ml^'. Theroot exudate was concentrated by ultrafiltration with anAmicon PM-10 membrane at 4°C to a protein concen-tration of approximately 3D fig protein ml"'.

Transmission electron microscopy was performed ona Philips TEM 300. Seedling roots were grown asen-ically from surface-sterilized seeds into humid air for 24h. The root hair regions were located by stereomicro-scopy and excised by razor blade, then fixed in glu-taraldehyde in the presence of tannic acid and saponin(Maupin and Pollard 1983) and postfixed with osmiumtetroxide. Following dehydration in graded alcohols andsoaking in propylene oxide, the root sections were em-bedded in DER Epoxy resins (Polysciences Inc., War-rington, PA). Ultrathin sections of root hairs werestained with uranyl acetate and lead citrate. To revealthe bacterial capsule in ultrathin section, cells werestained with ruthenium red in the primary fixative be-fore preparing ultrathin sections (Pate and Ordal 1967).To examine the capsule on whole bacteria by TEM, cellswere contrasted by the G/R/U method (Mutaftschiev etal. 1982).

Trifoliin A was detected on heat-fixed encapsulatedcells of R. trifolii 0403 by indirect immunofluorescencemicroscopy using the IgG fraction of rabbit anti-seedtrifoliin- A antiserum as the primary antibody (Dazzoand Hrabak 1981). FITC-labelled goat anti-rabbitgamma globulin (Difco Laboratories, Detroit, MI) wasused as the secondary antibody.

immunoelectron microscopy was performed usingcolloidal gold, prepared according to the method ofFrens (1973) and coupled to the IgG fraction of rabbitanti-trifoliin A antiserum as described by Horisberger etal. (1975). To locate the sites which bound trifoliin A,colloidal gold was coupled to trifoliin A that had beenisolated from clover root exudate by immunoaffinitychromatography (Dazzo and Hrabak 1981). The bacte-ria on the grids were incubated at room temperature for15 min with the antibody-gold or trifoliin A-gold com-plexes and then contrasted with G/R/U.

Concentrated root exudate of axenically-grown whiteclover seedlings (Dazzo and Hrabak 1981) was cen-trifuged at increasing speeds, each time removing thepellet before recentrifugation of the supernatant at thenext higher speed. The pellet (resuspended in PBS) andthe supernatant of each centrifugation were examinedfor the presence of active trifoliin A which could bind toencapsulated cells of R. trifolii 0403 as detected by indi-rect immunofluorescence (Dazzo and Hrabak 1981).Centrifugation was for 30 min at each speed.

Twenty clover seedlings germinated from surface-sterilized seeds into humid air for 2 days were placed ina 5-ml polyethylene beaker containing 2 ml of filter-sterilized Fahraeus medium and shaken vigorously for 1h. This "seedling wash" was cleared of cell debris bycentrifugation at lOOO g for 3 min, and then passed

576 Physiol. Plain. 66, 1986

through a 0.45 |xm filter. One ml of this filtered seedlingwash (ca 20 jig protein ml"') was layered on a lineardensity gradient of 20-60% metrizamide (Nyegaard .&Co. A/S, Oslo) in distilled-deionized water on top of an80% metrizamide cushion. After centrifugation at150 000 g for 17 h, fractions (0.5 ml) were collected fromthe gradient and assayed for trifoliin A by the indirectimmunofluorescence assay (Dazzo and Hrabak 1981).Densities of the fractions were determined by their re-fractive index.

Additional filtered seedling wash was diluted with fil-tered Fahraeus medium and recentrifuged at 30000 gfor 1 h to pellet the particulate material of interest. Thepellet was washed twice in PBS by ceotrifugation, resus-pended in SDS sample buffer (Laemmli 1970), boiledfor 15 min, and analyzed lor protein composition bySDS-PAGE. Proteins separated in 11% gels were fixedin 50% (v/v) methanol and 10% (v/v) acetic acid for 2 h(changing the fixative after 1 h) and stained with silver(Morrissey 1981). Proteins separated in 10% gels weretransferred electrophoretically to nitrocellulose sheetsfor western blot analysis (Sherwood et al. 1984) usinganti-seed trifoliin A antiserum and ['-'I]-protein A (ICNPharmaceuticals, Irvine, CA, USA) as the probe.

Results

Electron-dense particles were detected by TEM withinthe cell wall at the growing tip of clover root hairs andon the outer face of these differentiated epidermal cells(Fig. 1). Although also found along the sides of the roothair and occasionally on other epidermal cells, theseparticles were more abundant at the root hair tip. Par-ticles having similar ultrastructure bound to the pe-

riphery of the capsule of R. trifolii 0403 which had beenincubated for 1 h with concentrated root exudate of ax-enically grown clover seedlings (Fig. 2a). Particles werealso located on the acidic extracellular fibrils of the bac-teria incubated for 2 days in the onconcentrated root ex-udate of clover seedlings in Fahraeus slide cultures (Fig.2b), but not oo the unencapsulated cell surface (Fig. 2b)or flagella in the same preparation (figure not shown).Similarly, particles which stained positively with uranylacetate were bound to the exofibrils of R. trifolii 04Q3that had been shaken with white clover seedlings in Fah-raeus medium (Fig., 2c). The individual particles had amean diameter of 100-200 nm and tended to form largeraggregates of various sizes. These particles were notpresent on untreated bacteria incubated in Fahraeusmedium alone (Fig. 2d).

Immunoelectron microscopy showed that the parti-cles in root exudate which bound to the acidic poly-saccharide of R. trifolii were reactive with an anti-trifo-liin A IgG colloidal-gold complex (Fig. 3), suggestingthat they contained trifoliin A. As anticipated, cross re-actions between anti-trifoliin A igO and bacterial poly-saccharide itself were not found (Fig. 3). The partic-ulate material which adsorbed to the bacterial poly-saccharide that had been incubated with concentratedroot exudate was more aggregated than the particlesfound on the polysaccharide of bacteria growing in slidecultures with white clover seedlings (compare Figs 2band 3). This was probably due to aggregation of individ-ual particles during pressurized ultrafiltratioo. As a con-trol, a colloidal gold complex of rabbit preimmune IgGdid not bind to the particles from root esudate whichadsorbed to the bacterial polysaccharides. Although 30mM 2-deoxy-D-giuoose, a hapten for trifoliin A, preven-

. ' I I

.. . • • ' I l l , ; . • I I I I

.;• I I I . . I:- ; ••,•. V.

. 1 . . , , I ••_ '^ . i ,

, • I . ' .1 I I . <•! i l ' i i !

I . . I k " I I . 1 . . .

I I • h i - I I I . • I ••>

111 . iiii-i . I l l

• • V I 1 1 - • i ,• I ! '

•• •, I •. - I , . : . , I •

i h i L • • . • ! ^ i n u 111. • I '

. i i i ^ t ' '. ,;;\'" ' . • !

I l l

I .

•• . i l l

.1 • n d

38* Physioi. Plant. 66. 1986 577

Fig. 2. InteractioD of electron-dense particles (arrows) from white clover with encapsulated celts of R. trifo ' i W3 in Fahraeus me-dium. (A) Ceils were incubated for 1 h with concentrated root exudate, fixed in the presence of ruthenium i, d, and post-stainedwith uranyl acetate and lead citrate. Bar = 0.5 |im. (B) Cells were incubated for 2 days in the root environment of clover growingin a Fahraeus slide assembly and stained by the G/R/U method. Individual (arrow) and aggregated particles (double arrow) havebound to the bacterial polysaccharides. Bar = 1 [ira. (C) Cells shaken with seedlings for 1 h and stained with 0.2% uranyl acetate.Aggregated particles (arrow) are present. Bar = 1 (jm. (D) Control cells incubated for 2 days in Fahraeas medium without plantsand contrasted by tlie G/R/U method. Bar = 1 im.

ted the binding of trifoliin A-gold complexes to the bac-terial polysaccharide, this sugar was unable to dissociatethe particles from the polysaccharide or the trifoliin A-gold from the particles once they had bound. TEMshowed no evidence for disaggregation of the particlesby 2-deoxy-D-glucose.

Differential centrifugation indicated that most, butnot all, of the jR. rrrfoffi-binding trifoliin A in the con-centrated root exudate was present in a particuiate form(Tab. 1). The majority of this lectin which couid bind toR. trifolii and be detected by immunofluoresceoce waspelleted from concentrated root exudate at 15 000 g for

Fig. 3. Acidic cxopolymers of J?. in/o/iY0403 preincubated withconcentrated root exudate for 1 h, treated with colloidal goldcoupled to the IgG fraction of aiiti-trifoliin A antiserum andstained by the G/R/U method. The immunogold complexbound only to the aggregated particles, and not to the extracel-lular fibrils. Bar = 1 \xm.

578 Physiol, Planl. 66, 1986

Tab. 1. Presence of trifoliin A in fractions of clover root exu-date following sequential differential centrifugation. a. De-tected by its ability to bind to R. trifolii 0403 and react withanti-seed trifoliin A IgG by indrect immunofluoreseence. In-tensity of fluorescence was rated from 0' for the preimmune se-rum control to '** as the most intense reaction. mCentrifugal force, g Presence of trifoliin A"

pellet supertantant

800015000230003000040000

30 min. However, some residual lectin was still detected r- . .. j . , ,. . ,. , Anr,r,n > c i J » n • ^'S- ^- Aggregated particles pelleted by cenlrifugation atin the 40000 g supernatant of root exudate. Preimmune 15000 g, then treated with trifoliin A coupled to colloidal goldserum gave negative results in the immunofluorescence and contrasted by the G/R/U method. Bar = 1 nm.assay for trifoliin A. The particles isolated by centrifu-

Fi .."!. liflect of vigorou"- shaking on clover seedlings. (A) Whole seedling? after shaking. Phase contrast micrographs of (B) rooihail!, after shaking, (C) root cap cells and debris in pellet after slow-speed centrifugation, and (D) root tip with associated mucigcland cap cells before shaking, liars in B, C, iand D are 40 im.

579

gation also reacted with tnfoliin A coupled to colloidalgold (Fig. 4) and morphologically resembled the aggre-gated form of the particles in concentrated root exudatewhich bound to the capsule of R. trifolii. As further evi-dence of a particle-associated form of trifoliin A, theimmunofluorescence assay indicated that active trifoliinA in concentrated root esudate eluted in the voided vol-umes of a Sepharose 4B column and a glycerol-deactiv-ated Bio-Glass 2000 column using HPLC (figure notshown).

Seedlings were vigorously shaken to dislodge the as-sociated particles from the root. This procedure left theseedlings and the root hairs intact (Fig. 5a and h).Phase-contrast microscopic examination of the attachedroot hairs indicated that they still displayed cytoplasmicstreainiog after Ihis Jreatmeot. .Brief slow-speed cen-trifugation of the "seedling wash" produced a pellet ofdebris and non-root hair cells (Fig. 5c) derived from theroot cap (Fig. 5d). A filtrate of the slow-speed super-tanant was fractionated by equilibrium density ultracen-trifugation through a metrizamide gradient to isolatethe particles containing tnfohm A. Analysis of fractionsb} the immunofluorescence assay (Fig 6) indicated thatthe particle-associated lectio which could bind to encap-sulated R Infolii 0403 had a mean density of1 12 ± 0 01 g cm-

The 30000 g pellet of the cleared seedling wash ofclover roots was analyzed by SDS-FAGE for total pro-tein and western blots for trifohin A (Fig. 7) Separationm 11% gels indicated that this pellet contained severalproteins, including a band which comigrated with tnfo-iiio A standard (ca 53 kdalton) Other major proteins m

A B D E

Fig. 6.;Immunolluoi'escenceofR. rn/o/ri0403 withantl-trifoliinA antiserum after treatment of cells with particles having adensity of 1.12 g'cm" and isolated from the eioverroot wash bydensity gradient ultracentrifugation through metrizamide.Cells were previously heat-fixed to the microscope slides. Bar= 1 nm.

--200

m

10%

Fig. 7. Analysis of proteins in particles by SDS-PAGE (11%running gel) and western blot analysis (10% ruBning gel). LaneA) proteins in particles detached from clover roots by shakingand isolated by centrifugation at 30000 j . Lane B) trifoliin Astandard (star). Lane C) proteins in concentrated root exudatefrom wbtte clover seedlings grown under N-free conditions for5 days. Lane D) Western blot of proteins in particles separatedby SDS-PAGE and probed with anti-trifoliin A antiserum and['-"•IJ-protein A. The upper band comigrated with the glysosy-lated standard of trifoliin A. Lane E) western blot of trifoliin Astandard (glycosylated form, star). Positions of molecularweight markers are shown in kdalton.

this pellet had approximate molecular weights of 93 aod66 kdalton. Not all of the proteins in concentrated rootexudate were found in this pellet. Following separationin 10% gels, the proteins were transferred electropho-retically to a nitrocellulose membraoe for western blotanalysis. The protein io the particles which comigratedwith trifoliin A also reacted with antibody preparedagainst purified seed trifoliin A. In addition, this anti-body reacted with a slightly faster migrating proteinhaving an approximate molecular weight of 51 kdalton.This is the approximate molecular weight of the degly-cosylated form of trifoliin A, which contains approxi-mately 7% (w/w) glycosylation (J. E. Sherwood, 1984,Ph. D. thesis, Michigan State Univ., East Lansing, MI,U.S.A.).

Discussion

We have isolated particles from concentrated whiteclover root exudate and from the surface of intact seed-lings of white clover which bind strongly to encapsu-

580 Physiol. Planl. 66. 1986

lated cells of R. trifolii 0403. Immunofluorescence mi-croscopy, immunoelectron microscopy aod western blotanalysis indicate that a white clover lectin, trifoliin A, isa component of these particles. SDS-PAGB analysisshows that the particles contain several proteins in addi-tion to trifoliin A, but the protein composition of theparticles is not identical to the protein composition ofroot exudate, which is more complex. Electron micro-scopic examinations indicate that these particles stainpositively with uranyl acetate and are within the sizerange (100-200 nm diam.) of globular particles foundassociated with bacteria on the rhizoplane and mucigelof subterranean clover (Dart 1971, Baker aod Cooke1974, Foster et al. 1983), strawberry clover (Dazzo aodHubbell 1975, Napoli and Hubbell 1975), barrel medic(Dart and Mercer 1964), white clover (Callaham andTorrey 1981, Robertson et al. 1981, Dazzo et al. 1984),siratro (Ridge aod Roife 1985) and pearl millet (Umali-Garcia et al. 1981).

The affinity of these particles for acidic exopolymersof the capsule of R. trifolii 0403 and for purified trifoliinA from root exudate is sufficiently strong to withstandthe preparations for electron microscopy. There is someselectivity in these affinities, since the particles do notbind to (i) unencapsulated cells or flagella of R. trifolii0403, or (ii) all proteins present in root exudate. Neitherthe particle-polysaccharide nor the particle-trifoliin Acomplex could be dissociated by 2-deoxy-D-glucose (ahapten for trifoliin A-saccharide interactions) oncethese complexes had formed. Similarly, complexes for-med between trifoliin A and polysaccharides from R.trifolii do not freely dissociate with 2-deoxy-D-glucose,despite the fact that this sugar blocks the formation ofthese specific glycoprotein-polysaccharide complexes(J. E. Sherwood, E. M. Hrabak, and F. B. Dazzo, un-published results). However, hapten inhibition studiesmust be interpreted cautiously sioce some lectins un-dergo conformationai changes after binding to saccha-ride ligands which may prevent subsequent dissociationby hapten sugars (Reeke et ai. 1975). Lectins have beenshown to form secondary, hapten-irreversibie interac-tions once the lectin-saccharide complex has formed(Goldstein 1975). Alternatively, other components ofthe particles may be interacting with the bacterial poly-saccharide in a manner which is not inhibited by 2-deoxy-D-glucose.

The mechanism of formation and fate of these par-ticles in root exudate is as yet unknown. However, sev-eral observations bear relevance to this process. Similarelectron-dense particles are found within the root hairwall at the growing tip and are in abundance along theouter face of the root hair cell wall where trifoliin A ac-cumulates. These particles have affinity for several (butoot all) proteins io root exudate and for each otber.These proteins must be interacting strongly with theparticles, requiring extensive boiling in SDS-bufferprior to electrophoresis. Still, smearing of the proteinsoccurs in SDS-PAGE. These proteins are not easily sol-

ubilized from isolated particles hy sonication in ooo-Ionic detergents (Triton X-100), high salt concentrations(1 M NaCI), or 30 mM 2-deoxyglucose. These particlesalso bind tenaciously to insoluble polyvinylpyrrolidooe,which is known to hydrogen-bond to phenolic com-pounds (Dazzo and Hubbell 1974). Although the par-ticles on root hairs can be detected by TEM after con-ventional glutaraldehyde fixation, their preservation issignificantly improved by fixation in the presence of tan-nic acid using the technique of Maupin and Pollard(1983). Root hair tips are also the site of accumulationof peroxidase (Zaar 1979), which has been proposed toparticipate in crosslinking extensin precursors via isodi-tyrosine bridges (Epstein aod Lamport 1984) and in hg-nin biosynthesis (K. Kirk, personal communication).Considered collectively, these clues suggest to us thatthese particles may arise from compiexing of certain ex-creted proteins (including trifoliin A) with phenoliccompounds (ligoin-like?) at the root hair tip.

A similar situation occurs with beans {Phaseolus vul-garis), which contain a large agglutinin complex of pro-tein and carbohydrate that can be readily washed off ofintact roots or shoots and can differentially bind to cer-tain saprophytic Pseudomonas spp. which colonizethese plant surfaces (Anderson 1983).

When introduced into the hydroponic clover root en-vironment, R. trifolii quickly becomes coated with theseparticles containing many proteins of host origin, in-cluding trifoliin A. The involvement of these and similarparticles in plant-microbe interactions appear to be thenormal case in the rhizosphere and therefore should beconsidered io studies (Nutman 1957; A. van Egeraat,1972. Ph.D. thesis, Wageningen Agricultural Univ.,Wageningen, The Netherlands; C. A. Napoli, 1979.Ph.D. Thesis, Univ. Florida, Gainesville, FL, USA;Bhagwat and Thomas 1982, Dazzo et al. 1982, 1984,Halverson and Stacey 1984, 1985, Sherwood et ai. 1984,Bhuvaneswari and Solheim 1985), examining the effectsof lectins and other biologically active molecules in rootexudates on the early stages of the infection process inthe nitrogen-fixing Rhizobium-legume symbiosis. Thepotential significance of these interactions is increasedhy the recent report that binding of soybean lectin to B.japonicum induces de novo synthesis of a class of pro-teins necessary for nodulation capacity (Halversoo etal., Abstr. OR-14-07, First International Congr. PlantMolecular Biol., 1985).

Acknowledgements — This work was supported bv USDACompetitive Grants 82-CSRS-1-0-1040 and 85-CRCR-1-1627.NSF Grant No. PCM 80-21906, and NIB Grant No. GM34331-02 awarded to F.B.D., and grants from the EuropeanMolecular Biology Organization, Elf-Aquitaine, l'EntrepriseMiniere et Chimique, Rhone Poulenc Recherches et C.D.F.Chimie awarded to G.L.T. This is article No. 11503 of theMichigan Agricultural Experiment Station. We thank KennethNadler and John Wang for helpful suggestions, and Esteile M.Hrabai for excellent technical assistance.

Physiai. Planl. 66, 1986 581

References

Anderson, A. J. 1983. Isolation from root and shoot surfacesof agglutinins that show specificity for saprophytic Pseudo-monads. - Can. J. Bot. 61: 3438-3443.

Baker, K. F. & Cooke, R. J. 1974. Biological Control of PlantPathogens. - W. H. Freeman and Co., San Francisco, CA.pp. 229-232. ISBN 0-89054-045-4.

Bhagwat, A. A. & Thomas, J. 1982. hsgume-Rhizobium inter-actions: cowpea root exudate elicits faster noduiation re-sponse by Rhizobium species. - Appl. Environ. Microbiol.43: 80(>-805.

Bhuvaneswari, T. V. & Solheim, B. 1985. Root hair deforma-tion in the white dovsilRhizobium trifolii symbiosis. -Physiol. Plant. 63: 25-34.

Bohlool. B. B. & Schmidt, E. L. 1974. Lectins: a possible basisfor specificity in the Rhizobium-legnme root nodule sym-biosis. - Science 185: 269-271.

Callaham, D. A. & Torrey, J. G. 1981. The structural basis forinfection of root hairs of Trifolium repens by Rhizobium. -Can. L Bot. 59: 1647-1664.

Dart,P.J. 1971. Scanningelectronmicroscopy of plant roots.-J. Exp. Bot. 22: 163-168.

- & Mercer, F. V. 1964. The legume rhizosphere. - Arch.Mikrobiol. 47: 344-378.

Dazzo, F. B. 1981. Bacterial attachment as related to cellularrecognition in the RMzobium-legumt symbiosis. - J. Su-pramolec. Struct., Cell. Biochem. 16: 29-41.

- 1982. Leguminous root nodules. - In Methods in MicrobialEcology (R. G. Bums and J. H. Slater, eds), pp. 431^M6.Blackwell Scientific Publications, Oxford, U.K. ISBN 0-632-00765-6.

- & Hrabak, E. M. 1981. Presence of trifoliin A, a Rhi-zoWum-binding lectin, in clover root exudate. - J. Su-pramolec. Struct. Cell. Biochem. 16: 133-138.

- & Hubbell, D. H. 1974. A quantitative assay of insolublepolyvinylpyrrolidone. - Plant Soil 40: 435-440.

- & Hubbell, D. H. 1975. Cross-reactive antigens and lectinsas determinants of symbiotic specificity in the Rhizobium-ciover association. - Appl. Microbiol. 30: 1017-1033.

- , Napoli, C. A. & Hubbell, D. H. 1976. Adsorption of bac-teria to roots as related to host specificity in the Rhizobium-cloversymbiosis. -Appl. Environ. Microbiol. 32:166-171.

- , Yanke, W. E. & Brill, W. J. 1978. Trifoliin: a Rhizobiumrecognition protein from white clover. - Biochim. Biophys.Acfa 539: 276-286.

- , Truchet, G. L., Sherwood, J. E., Hrabak, E. M. & Gar-diol, A. E. 1982. Alteration of the trifoiiin A-binding cap-sule of Rhizobium trifolii 0403 by enzymes released fromclover roots. - Appl. Environ. Microbiol. 44: 478-490.

- , Truchet, G. L., Sherwood, J. E., Hrabak, E. M., Abe, M.& Pankratz, H. S. 1984. Specific phases of root hair attach-ment in the Rhizobium trifolii-dovsT symbiosis. - Appl.Environ. Microbiol. 48: 1140-1150.

- ,Hollingsworth, R.L., Sherwood, J. E., Abe, M.,Hrabak,E. M., Gardiol, A. E., Pankratz, H. S., Smith, K. B. &Yang, H. 1985. Recognition and infection of clover roothairs by Rhizobium trifolii. - In Nitrogen Fixation ResearchProgress (H. J. Evans, P. J. Bottomley and W. E. Newton,eds), pp. 239-245. Martinus Nijhoff, Dordrecht, TheNetherlands. ISBN 90-247-3255-7.

Diaz, C. L., Lems-van Kan, P., Van Der Sehaal, L. A. M. &Kijne, J. M. 1984. Determination of pea root lectin usingan enzyme-linked immunoassay. - Planta 101: 302-307.

Epstein, L. & Lamport, D. T. 1984. An intramolecular linkageinvolving isodityrosine io extensin. - Phytochemistry 23:

nn-n4(,.Foster, R. C , Rovira, A. D. & Cock, T. W. 1983. Ultra-

structure of the Root-Soil Interface. - American Phyto-

pathological Society, St. Paul, MN. p. 48. ISBN 0-89054-051-9.

Frens, G. 1973. Controlled nucleation for the regulation of theparticle size in monodispersed gold suspensions. - Nature241: 20-22.

Gerhold, D. L., Dazzo, F. B. & Gresshoff, P. M. 1985. Se-lective removal of seedling root hairs for studies of the Rhi-zobium-\G^nmc symbiosis. — J. Microbiol. Methods 4: 95-102.

Goldstein, L. J. 1975. Studies of the combining sites of Conca-navalin A. - In Concanavalin A (T. K. Chowdhury and A.K. Weiss, eds), pp. 35-53. Plenum Press, N.Y. ISBN 0-306-39055-8.

Halverson, L. J. & Stacey, G. 1984. Host recognition in theRhizobium-s,oybest\ symbiosis. Detection of a protein fac-tor in soybean root exudate which is involved in the nodula-tion process. - Plant Physiol. 74: 84-89.

- & Stacey, G. 1985. Host recognition in the Rhizobium-soy-bean symbiosis. Evidence for the requirement of lectin innodulation. - Plant Physiol. 77: 621-625.

Horisberger, M., Rosset, J. & Bauer, H. 1975. Colloidal goldgranules as markers for cell surface receptors in the scan-ning electron microscope. - Experientia 31: 1147-1151.

Laemmli, U. K. 1970. Cleavage of structural proteins duringthe assembly of the head of bacteriopbage T4. - Nature227: 680-685.

Maupin, P. & Pollard, T. D. 1983. Improved preservation andstaining of HeLa cell actin filaments, clathrin-coated mem-branes, and other cytoplasmic structures by tannic acid-glu-taraldehyde-saponin fixation. - J. Cell. Biol. 96: 51-62.

Morrissey, J. H. 1981. Silver stain for proteins in polyacryla-mide gels: a modified procedure with enhanced uniformsensitivity. - Anal. Biochem. 117: 307-310.

Mutaftschiev, S., Vasse, J. & Truchet, G. L. 1982. Exostruc-tures of Rhizobium melitoti. — FEMS Microbiol. Lett. 13:171-175.

Napoli, C A. & Hubbell, D. H. 1975. Ultrastructure ot Rhi-zobium-\ndnctt^ infection threads in clover root hairs. -Appl. Microbiol. 30: 1003-1009.

Nutmao, P. S. 1957. Studies on the physiology of nodule for-mation. V. Further experiments on the stimulation and in-hibitory effects of root exudate. - Ann. Bot. 21: 321-327.

Pate, J. L. & Ordal, E. J. 1967. The fine structure of Chondro-coccus columnaris. III. The surface layers of Chondrococ-cus columnaris. - J. Cell. Biol. 35: 37—51.

Reeke. G. N., Becker, J. W., Cunningham, B. A., Wang, J.L., Yahara, L. & Edelman, G. M. 1975. Structure andfunction of Concanavalin A. - In Concanavalin A (T. K.Chowdhury and A. K. Weiss, eds), pp. 13-33. PlenumPress, N.Y. ISBN 0-306-39055-8.

Ridge, R. W. & Rolfe, B. G. 1985. Rhizobium sp. degradationof legume root hair cell wall at the site of infection threadorigin. - Appl. Environ. Microbiol. 50: lYI-lTii.

Robertson, J. G., Lyttleton, P. & Pankhurst, C, E. 1981. Pre-infection and infection process in the legume-Rhizobiumsymbiosis. - In Current Perspectives in Nitrogen Fixation(A. H. Gibson and W. E. Newton, eds), pp. 280-291. Aus-tralian Academy of Science, Canberra. ISBN 0-85847-085-3.

Sherwood, J. E., Truchet, G. L. & Dazzo, F. B. 1984. Effect ofnitrate supply on the in-vivo synthesis and distribution oftrifoliin A, a Rhizobium trifolii binding lectin, in Trifoliumrepens seedlings. - Planta 162: 540-547.

Umali-Gareia, M., Hubbell, D. H., Gaskins, M. H. & Dazzo,F. B. 1980. Association of Azospirillum with grass roots. -Appl. Environ. Microbiol. 39: 219-226.

Zaar, K. 1979. Peroxidase activity in root hairs of cress (Lepi-dium sativum L.). Cytochemical localization and radioac-tive labelling of wall-bound peroxidase. - Protoplasma 99:263-274.

Edited by C. Larsson

582 Physiol. Plant. 66, 1986