Soil Bid. Biochem. Vol. 28, No. 7, pp. 917-921, 1996 Copyright 0 19% Elsevierscience Ltd Pergamon Printed in Great Britain. Au rights resewed PII: !30038-0717(%)ooos9-4 0038-0717/96$15.00 + 0.00

CELL-BOUND CELLULASE AND POLYGALACTURONASE PRODUCTION BY RHIZOBIUM AND BRAD YRHIZOBIUM SPECIES

JOSE I. JIMENEZ-ZURDO,’ PEDRQ F. MATEOS,‘* FRANK B. DAZZO” and E. MARTINEZ-MOLINA’ ‘Departamento de Microbiologia y Genetica, Facultad de Farmacia, Fkiificio Departamental, Universidad de Salamanca, Salamanca 37007, Spain and 2Department of Microbiology and Center for Microbial Ecology, Michigan State University, East Lansing, MI 48824, U.S.A.

(Accepted 12 February 1996)

Summary-Carboxymethyl cellulase and polygalacturonase activities were evaluated in wild-type strains from different taxonomic groups of rhizobia which nodulate a specific range of legume hosts that do not normally overlap (R. meliloti, R. loti, R. leguminosarumbiovar trifolii, R. leguminosarumbiovar phaseoli, R. leguminosarum biovar viceae, and Bradyrhizobium japonicum). To detect these enzymatic activities we used various methods, a double-layer plate assay, quantitation of reducing sugar products with a bicinchoninate reagent, and viscometric assay. CM-cellulase activity was found in extracts from all the rhizobia tested. In contrast, polygalacturonase activity was only found in R. leguminosarumbiovar trifolii and R. meliloti. CM-cellulase and polygalacturonase enzymes were cell-bound and not detected in the extracellular culture medium. CM-cellulase isozymes were examined using activity gel electrophoresis. The results obtained shown a heterogeneity of CM-cellulase isozymes among the strains tested. These findings are consistent with the hypothesis of an involvement of rhizobial cellulolytic enzymes in the invasion of legumes by Rhizobium. Copyright 0 1996 Elsevier Science Ltd

INTRODUCTION locally induced in the plant by components of the bacteria. The precise mechanism whereby Rhizobium success- The infection of legume roots by Rhizobium is a fully infects temperate legumes remains unknown. delicately balanced process (Dazzo and Hubbell, Nutman (1956) proposed that rhizobia invade root 1982); the infected root hair must remain intact and hairs by an invagination process of the plant cell wall, viable, otherwise infection would abort. An involve- which implies that there is no real penetration of the ment of wall-degrading enzymes would require that bacteria into the root cell. Ljunggren and Fghraeus their activity be regulated to remain very low and (1961) proposed that polygalacturonase production localized at the site of infection in order to avoid by plant root cells, induced by homologous strains destruction of the root hair. Although there is of Rhizobium, would result in an increase in plant indirect evidence for the involvement of hydrolytic cell wall softening at the infection site on the root enzymes in the infection process (Martinez-Molina hair; thus allowing the bacteria to penetrate the cell and Olivares, 1982; Al-Mallah et al., 1987; Chaliflour membrane and initiate an infection thread. The and Benhamou, 1989; Al-Mallah et al., 1989), little is ability of Rhizobium meliloti to induce polygalactur- known about these enzymes in Rhizobium. Some onase production by alfalfa roots is linked to a studies failed to demonstrate the production of plasmid carried by the bacteria (Palomares et al., these enzymes by different species of Rhizobium 1978). (McCoy, 1932; Hunter and Elkan, 1975), but Electron microscopic studies of the infection several studies have detected pectinolytic (Hubbell process have showed a localized degradation of the et al., 1978; Martinez-Molina and Olivares, 1982; root hair wall, suggesting the involvement of cell wall Plazinski and Rolfe, 1985; Angle, 1986; Prasuna and hydrolytic enzymes in the penetration process Ali, 1987; Saleh-Rastin et al., 1991), cellulolytic (Higashi and Abe, 1980; Callaham and Torrey, (Martinez-Molina et al., 1979; Morales et al., 1984; 1981; Ridge and Rolfe, 1985; Turgeon and Bauer, Saleh-Rastin et al., 199 1) and hemicellulolytic 1985). The main point of divergence would be the (Martinez-Molina et al., 1979; Lopez and Signer, possibility that the wall-degrading enzymes involved 1987) enzyme activities from pure cultures of in the process are associated with the bacteria or rhizobia. In general, the activity of these rhizobial enzymes is very low and at the limit of sensitivity of conventional reducing sugars and plate assays; this *Author for correspondence. has hampered biochemical study of these polysaccha-

917 918 Jose I. Jimenez-Zurdo et al. ride-degrading enzymes and research aimed at Mateos et al. (1992). In addition, the BCA reducing determining their role in the establishment of the sugar assay was modified as follows. The reaction symbiotic interaction. mixtures contained 0.4 ml of 1% carboxymethyl Mateos et al. (1992) developed assays with cellulose (CMC) or 0.5% sodium polygalacturonate improved sensitivity and reliability to detect and (NaPp) solution, 0.4 ml of the enzyme sample and 0.8 measure cellulolytic and pectinolytic activities; these ml of 200 mM PCA buffer. After the reagents were were used to demonstrate the production of both mixed, an 0.8 ml aliquot (considered T = 0 h) was enzyme activities by different wild-type strains of the frozen at - 80°C and the remaining sample was kept clover root-nodule symbiont, R. leguminosarum at 40°C for 5 h before quantitation of reducing sugar bv. trifolii. These enzymes were cell-bound rather products. Substrates (CMC medium viscosity and than extracellular, and were commonly produced NaPp; both obtained from Sigma Chemical Co., St by several wild-type strains isolated from different Louis, MO) were dialyzed for 2 days against distilled geographical regions. In strain ANU843, at least water before use. two carboxymethyl (CM)-cellulase isozymes and one Cellulolytic and pectinolytic activities were also polygalacturonase isozyme were found. The total and evaluated in cell extracts of the strain R. meliloti specific activities of these cell-bound enzymes in ATCC 9930 produced by treatment with lysozyme- sonicated cell extracts were unaffected by growth EDTA (Mateos et al., 1992). The presence of in defined B-INOS medium supplemented with cytoplasmic constituents in the intracellular and CM-cellulose or polygalacturonic acid substrates, or extracellular fractions of this strain was examined by by flavone 7,4’-dihydroxyflavone which activates measuring NADP-dependent glucose-6-phosphate expression of its pSym nodulation genes (Mateos dehydrogenase (E.C. 1.1.1.49) (Worthington, 1988) a et al., 1992). cytoplasmic enzyme marker. All enzyme samples These findings have prompted us to investigate were stored at 4C. whether production of these polysaccharide-degrad- CM-cellulase activity was also evaluated by ing enzymes by rhizobia is restricted to the clover measuring the decrease in viscosity of a solution of symbiont R. leguminosarum biovar trifolii, or are CMC (Martinez-Molina et al., 1979) using Cannon- commonly made by various species and biovars of the Fenske viscometers. Reaction mixtures for this assay root-nodule bacteria supporting the hypothesis of an contained 5 ml of a 1% solution of CMC in 100 mM involvement of these rhizobial hydrolytic enzymes in PCA buffer and 1 ml of the enzyme extract. the symbiosis. Sodium dodecyl sulfate-polyacrylamide gel electro- phoresis (SDS-PAGE) was performed (Mateos et al., 1992) in a vertical slab unit (Mini-Protean II; MATERIALS AND METHODS Bio-Rad). The separating gel contained 12% The strains of rhizobia used in this study were R. acrylamide and 0.33% bis. The proteins in SDS gels leguminosarum bv. trifolii ANU843 (B. Rolfe, were renatured by keeping for 2 h with shaking in 10 Australian National University), R. melifoti ATCC mM PCA buffer with periodic changes every 30 min. 9930, R. loti ATCC 33669, R. leguminosarum CMC-agarose overlays were constructed using a bv. phaseoli ATCC 14482, R. Ieguminosarum bv. 0.4-mm layer of 0.2% CMC and 0.5% agarose (w/v) oiceae ATCC 10004 and Bradyrhizobium japonicum in 100 ItIM PCA buffer on electrophoresis film ATCC 10324 (ATCC, American Type Culture (Sigma). The running gel was placed on the substrate Collection). overlay and kept in a moist chamber at 37 C. The Broth cultures were grown in 75 ml of B-INOS CMC-agarose overlay was immersed in aqueous defined medium (Mateos et al., 1992) in 250 ml flasks 0.1% Congo red solution for 20 min and rinsed in IM shaken at 125 rev min - ’ at 28°C. Inocula were NaCl to detect areas of enzymatic activity. prepared by suspending cells from 5-day-old plate cultures into sterile B-INOS medium, centrifuging aseptically at 4000 x g for 15 min and resuspending RESULTS AND DISCUSSION in B-INOS medium to an initial population density of 10’ cells ml - ‘. The results of tests to detect CM-cellulase and After 24 h, at 28’C, cells were pelleted by polygalacturonase activities in sonicated cell extracts centrifugation, resuspended in 100 mM potassium from each of these different rhizobia are summarized phosphate-citric acid buffer (PCA; pH 5.0) sonicated in Table I. All of the methods detected CM-cellulase in five cycles of I-min bursts, and recentrifuged. This and polygalacturonase activities from R. legumi- sonicated, cell-free extract and the original super- nosarum biovar trijblii ANU843 as the positive natant of extracellular culture fluid concentrated control. The double layer plate assay detected 15-fold by ultrafiltration using PM 10 membranes CM-cellulase activity in extracts from each test strain (Amicon Co., Danvers, MA, USA) were assayed for except R. meliloti ATCC 9930, but detected CM-cellulase and polygalacturonase activities by polygalacturonase activity with only the positive the double-layer plate assay, and quantitation of control R. leguminosarum biovar trifblii ANU843. reducing sugar products using methods described by Also, no pectate lyase activity was detected by the Cellulase and polygalacturonases in Rbizobium 919

Table 1. CM-ccllulase and polyplacturonase activities of Rhizobiu* CM-cellulasc Polygalacturonasc Rhizobiwntest strain Double layer plate Viscometric’ Reducing sugarsb Double layer plate Reducing sugar@

R. meliloti ATCC 9930 0 0 0.08 f. 0.01 0 0.05 -I: 0.01 R. hi ATCC 33669 + 0.23 x IO-’ 0.10 f 0.01 0 0 B. japonicumATCC + 0.53 x 10-1 0.36 + 0.01 0 0 10324 R Ieguminosarumbv. viceoeATCC 10004 + 1.16 x IO-’ 0.27 f 0.01 0 0 phaseoliATCC 14482 + 0.50 x to-3 0.29 k 0.02 0 0 trfofolii ANU843 + 0.86 x IO-’ 0.55 * 0.02 + 2.90 k 0.06 ‘One unit of enzyme activity causes a 50% reduction in viscosity of a 1% CMC solution at 40°C and pH 5 mg- ’ protein min- ‘. Wne unit of enzyme activity is the amount releasing 1 nmol reducing sugar at WC and pH 5 mg“ protein min-‘. ‘Values reported are the means of triplicate samples derived from equivalent number of cells f standard errors of the means.

plate assay formulated with 100 mM Tris-HCI (pH among R. leguminosarum bv. trifolii strains. Poly- 8.5) plus 1.5 mM CaCl, (Collmer et al., 1988). galacturonase activity detected by the reducing sugar The quantitative BCA-reducing sugar assay assay in sonicated extracts of R. meliloti ATCC 9930 showed different amounts of CM-cellulase specific and not detected by the qualitative double-layer plate activity among the strains used (Table 1). These assay could also be due to enzymes with exo and not results were consistent with the double layer plate endo action. assay for sonicated extracts from all strains except R. Rhizobium meliloti has always been kept in a meliloti ATCC 9930, where both CM-cellulase and taxonomically stable position within the Family polygalacturonase activities were also detected. None Rhizobiaceae. The data presented in this study show of these assays detected CM-cellulase or polygalac- that the strain R. meliloti ATCC 9930 could produce turonase activity in the concentrated extracellular cellulolytic and pectinolytic enzymes with exo fraction of each strain (data not shown), indicating activity, differing from the other strains tested. As a that when produced, these enzymes are cell-bound complement to the previous determinations, we (Mateos et al., 1992). To evaluate whether ultrafiltra- examined these enzyme activities associated with R. tion might reduce these enzyme activities, we meliloti ATCC 9930 in greater detail. The sensitive concentrated an aliquot of the sonicated fraction; BCA-reducing sugar assay was used to examine the activities were not affected (data not shown). cell-bound vs extracellular distribution of polygalac- CM-cellulase activity in the sonicated extracts of turonase, CM-cellulase, and cellobiase (/I-glycosidase each strain was also evaluated by a viscometric assay, E.C.3.2.1.21) activities associated with R. meliloti specific for measuring endoglucanase activity (Rapp ATCC 9930 grown in the absence or presence of the and Beermann, 1991). The corresponding values of corresponding flavone (luteolin) which activates specific enzymatic units for each strain are shown in expression of its nodulation genes (Peters et al., Table 1. This method gave positive results for 1986). Cellobiase was included since it is cell-bound CM-cellulase activity from the strains of R. loti, B. and often associated with cellulases in both aerobic japonicum and the three biovars of R. leguminosarum and anaerobic bacteria (Rapp and Beermann, 1991). used, indicating the presence of enzymes with Polygalacturonase activity was only detected in the endohydrolytic action, while no activity was detected extracts of sonicated cells, while both CM-cellulase in cell extracts from R. meliloti ATCC 9930. Based on and cellobiase activities were detected in extracts of these findings, it is probable that the CM-cellulase pelleted cells produced by sonication or treatment activity measured in sonicated cell extracts of this with lysozyme-EDTA (Table 2). Although the strain by the BCA-reducing sugar assay is exo- specific activities were lower in sonicated cell extracts hydrolytic. However, direct measurements with due to greater amounts of extraneous protein, total purified enzymes should be carried out to confirm the CM-cellulase and cellobiase activities were higher in presence of exoglucanase activity, since no specific this cellular fraction. Lysozyme-EDTA treatment of substrates and reliable methods are available for the cells released almost 70% of the total activity of determination of this enzyme activity in the presence both enzymes found in the sonicated extract. The of other possible cellulase components (Rapp and cytoplasmic enzyme marker NADP-dependent glu- Beermann, 1991). cose-6-phosphate dehydrogenase (E.C. 1.1.1.49) was Mateos et al. (1992) reported that CM-cellulase present in the supernatant of sonicated cells, but was and polygalacturonase activities are cell-bound and not detected in the fraction of cell-bound proteins are commonly produced by diverse wild-type strains released by lysozyme-EDTA. This verified that there of R. leguminosarum bv. trifolii. Our study corrobo- was no cell lysis during lysozyme-EDTA treatment. rates these results and indicates that cellulolytic The release of CM-cellulase and cellobiase of R. enzymes are also produced commonly by different meliloti ATCC 9930 by lysozyme-EDTA treatment rhizobia species. However, the production of suggests a periplasmic location for these activities, as polygalacturonase activity seems to be common only is commonly found for these types of bacterial 920 Josi I. Jimknez-Zurdo et al.

Table 2. Polygalacturonase. CM-cellulase and cellobiase activities determined by the BCA-reducing sugar assay in different cell fractions of R. melilori ATCC 9930’ Cellobiase activity Polygalacturonase activity (U) CM-cellulase activity (U) -___-. (U) __- Cell fraction or treatment Total Specific Total Specific Total Specific Sonic extract 0.55 +_ 0.05 0.10 * 0.01 1.72 i 0.21 0.34 i 0.04 6.30 + 0.22 1.24 k 0.04 Lysozyme-EDTA 0 0 1.14 f 0.02 21.5 f 0.52 4.40 f 0.15 83.0 f 20.0 Extracellular 0 0 0 0 0 0 “One unit of enzyme activity is the amount releasing 1 nmol reducing sugar at 40 C and pH 5 min- ‘. *Values reported are the means of triplicate samples derived from equivalent number of cells i standard errors of the means enzymes (Rapp and Beermann, 1991). As was mobility as the ones shown in Fig. 1. Only when the previously found for R. leguminosarum bv. trifolii lane was overloaded with a sample containing (Mateos et al., 1992), growth of R. meliloti ATCC the separated cellulase of highest molecular mass, 9930 in the presence of the nod-activating flavone did could the activity of a possible third CM-cellulase not significantly alter its amounts of cell-bound isozyme be observed (data not shown), indicating polygalacturonase, CM-cellulase, and cellobiase ac- that this third isozyme from ANU843 is an aggregate tivities (data not shown), and none of these enzyme of the 44.3 kDa CM-cellulase isozyme. Activity gel activities were detected in the extracellular fraction. electrophoresis did not detect CM-cellulase isozymes, Activity gel electrophoresis (Fig. 1) detected a under the conditions described, in sonicated cell single CM-cellulase isozyme in sonicated cell extracts extracts from R. meliloti ATCC 9930, R. legumi- of R. loti ATCC 33669 and R. leguminosarum bv. nosarum biovar oiciae ATCC 10004, or B. japonicum phaseoli ATCC 14482, with approximate molecular ATCC 10324. masses of 44.3 and 33.3 kDa, respectively, under the In summary, our study indicates that CM-cellulase conditions described, based upon comparison to is not only produced by the clover symbiont, R. protein standards (Bio-Rad Laboratories, Richmond, leguminosarum biovar trifolii, but also by a diversity CA). The same method detected two CM-cellulase of other species and biovars of root-nodule bacteria. isozymes in extracts from R. leguminosarum biovar In contrast, polygalacturonase is less commonly trifolii ANU843, with approximate molecular masses found in rhizobia, and here we document its of 44.3 and 33.3 kDa. However, other studies with production by the alfalfa symbiont, R. meliloti, as ANU843 (Mateos et al., 1992) have indicated that a well as confirm its production by R. leguminosarum third isozyme can be detected when the lane is biovar trijolii ANU843 (Mateos et al., 1992). The overloaded. Ion exchange chromatography of soni- production of cellulolytic enzymes that cleave cated cells extracts of this strain (J. Jimtnez-Zurdo, glycosidic bonds in plant cell wall polymers has been 1993, unpub., Ph.D. University of Salamanca) shown to be a general phenomenon among various separated the two CM-cellulase isozymes. The species of Rhizobium. While the production and low separated enzymes had the same electrophoretic activity of these enzymes strongly suggest a possible role in the root invasion process, purification and further characterization of the enzymes detected should make it feasible to investigate the possible function of these polysaccharide-degrading enzymes kDa in the establishment of the symbiotic interaction.

91.4 Acknowledgrmenrs-This work was supported by Direcci6n General de lnvestigacibn Cientifica y Ttcnica (to E.M.M.) 66.2 and funds provided by the Plan National de Formaci6n de Personal Investigador (fellowship to J.I.J.Z.). and the Center for Microbial Ecology at Michigan State University 45.0 (to F.B.D.).

REFERENCES

31.0 Al-Mallah M. K., Davey M. R. and Cocking E. C. (1987) Enzymatic treatment of clover root hairs removes a Fig. 1. Detection of CM-cellulase isozymes in sonicated cell barrier to Rhkobium-host specificity. Biolrechnology 5, extracts of wild-type strains of different Rhizobium species. 1319-1322. Samples were separated by SDS-PAGE, followed by the Al-Mallah M. K., Davey M. R. and Cocking E. C. (1989) activity gel overlay technique. Lanes (in pg protein per Formation of nodular structures on rice seedlings by sample): I, R. loti ATCC 33669 (78 pg); 2, R. leguminosarum rhizobia. Journal of Experimenral Botan_v 40, 473478. bv. phaseoli ATCC 14482 (62 pg); 3, R. leguminosarum bv. Angle J. S. (1986) Pectic and proteolytic enzymes produced Cceae ATCC 10004 (50 fig); 4, R. leguminosarum bv. trifolii by fast and slow-growing soybean rhizobia. Soil Biolog? ANU843 (66 pg). The molecular masses (in kilodaltons) of & Biochemisrr_v 18, 115-l 16. protein standards are indicated at the right of the panel. Callaham D. A. and Torrey J. G. (1981) The structural basis Cellulase and polygalacturonases in Rhizobium 921

for infection of root hairs of Trifolium repens by the microchemistry of the host’s cell walls. Proceedings of Rhizobium. Canadian Journal of Botany 59, 1647-1664. the Royal Society of London Series B 110, 514-533. Collmer A., Ried J. and Mount M. (1988) Assay methods Morales-V. M., Martinez-Molina E. and Hubbell D. H. for pectic enzymes. Methods in Enzymology 161,329335. (1984) Cellulase production by Rhizobium. Plani and Soil Chahfour F. P. and Benhamou N. (1989) Indirect evidence 80,407415. for cellulase production by Rhizobium in pea root nodules Nutman P. S. (1956) The influence of the legume in root during bacteroid differentiation: cytochemical aspects of nodule symbiosis, A comparative study of host determi- cellulose breakdown in rhizobial droplets. Canadian nants and functions. Biological Reviews (Cambridge) 31, Journal of Microbiology 35, 821-829. 109-151. Dazzo F. B. and Hubbell D. H. (1982) Control of root hair Palomares A., Montoya E. and Olivares J. (1978) Induction infection. In Nitrogen Fixation, Vol. 2. Rhizobium (W. of polygalacturonase production in legume roots as a Broughton, Ed.), pp. 274-310. Clarendon, Oxford. consequence of extrachromosomal DNA carried by Higashi S. and Abe M. (1980) Scanning electron microscopy Rhizobium meliloti. Microbios 21, 33-39. of Rhizobium tr$olii infection sites on root hairs of white Peters N. K., Frost J. W. and Long S. R. (1986) A plant clover. Applied and Environmenral Microbiology 40, flavone, luteolin, induces expression of Rhizobium meliloii 1094-1099. nodulation genes. Science 233, 977-980. Hubbell D. H., Morales V. M. and Umah-Garcia M. (1978) Plazinski J. and Rolfe B. G. (1985) Analysis of the pectolytic Pectolytic enzymes in Rhizobium. Applied and Environ- activity of Rhizobium and Azospirillum strains isolated menral Microbiology 35, 210-213. from Trifolium repens. Journal of Plant Physiology 120, Hunter W. J. and Elkan G. H. (1975) Role of pectic and 181-187. cellulolytic enzymes in the invasion of the soybean by Prasuna P. and Ali S. S. (1987) Detection and characteriz- Rhizobium japonicum. Canadian Journal of Microbiology ation of two thermally reactive pectinases in cultures of 21, 12541258. Rhizobium. Indian Journal of Experimenlal Biology 25, Ljunggren H. and Fahraeus G. (1961) The role of 632-633. polygalacturonase in root-hair invasion by nodule Rapp P. and Beermann A. (1991) Bacterial cellulases. In bacteria. Journal of General Microbiology 26, 521-528. Biosynthesis and Biodegradation of Cellulose (C. H. Lopez M. and Signer E. R. (1987) Degradative enzymes in Haigler and P. J. Weimer, Eds), pp. 535597. Marcel Rhizobium meliloti. In Molecular Genetics of Plant- Dekker, New York. Microbe Interactions (D. Verma and N. Brisson, Eds), Ridge R. W. and Rolfe B. G. (1985) Rhizobium sp. pp. 185-187. Martinus Nijhoff, Dordrecht. degradation of legume root hair ceil wall at the site of Martinez-Molina E., Morales V. M. and Hubbell D. H. infection thread origin. Applied and Environmenral (1979) Hydrolytic enzyme production by Rhizobium. Microbiology 50, 7 17-720. Applied and Environmental Microbiology 38, 118&l 188. Saleh-Rastin N., Petersen M., Coleman S. and Hubbell D. Martinez-Molina E. and Olivares J. (1982) A note on (1991) Rapid plate assay for hydrolytic enzymes of evidence for involvement of pectolytic enzymes in the Rhizobium. In The Rhizosphere and Plant Growth (D. infection process of Medicago saliva by Rhizobium Keister and P. Gregon, Eds), p. 188. Klewvier Academic, meliloti. Journal of Applied Bacieriology 52, 453455. Dordrecht. Mateos P. F., Jimenez-Zurdo J. I., Chen J., Squartini A., Turgeon B. G. and Bauer W. D. (1985) Ultrastructure of Haack S. K., Martinez-Molina E., Hubbell D. H. and infection-thread development during the infection of Dazzo F. B. (1992) Cell-associated pectinolytic and soybean by Rhizobium japonicum. Planta 163, 328-349. cellulolytic enzymes in Rhizobium leguminosarum biovar, Worthington C. C. (1988) Glucose-6-phosphate dehydro- trifolii. Applied and Environmenral Microbiology 58, genase. In Worthington Enzyme Manual(C. C. Worthing- 18161822. ton, Ed.), pp. 159-161. Worthington Biochemical, McCoy E. (1932) Infection by Bact. radicicola in relation to Freehold.