Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Cell growth wikipedia , lookup
Tissue engineering wikipedia , lookup
Magnesium transporter wikipedia , lookup
Protein moonlighting wikipedia , lookup
Cytokinesis wikipedia , lookup
Signal transduction wikipedia , lookup
Cellular differentiation wikipedia , lookup
Organ-on-a-chip wikipedia , lookup
Cell encapsulation wikipedia , lookup
Vol. 175, No. 21 JO(URNAL OF BACTERIOLOGY, Nov. 1993, P. 7056-7065 0021-9193/93/217056-1 0$02.00/0 Copyright © 1993, American Society for Microbiology Cloning, Expression in Escherichia coli, and Characterization of Cellulolytic Enzymes of Azoarcus sp., a Root-Invading Diazotroph BARBARA REINHOLD-HUREK,' t* THOMAS HUREK,' t MARC CLAEYSSENS,2 AND MARC VAN MONTAGU' Laboratorium voor Genetika' and Laboratorium voor Biochemie,2 Universiteit Gent, B-9000 Ghent, Belgium Received 25 May 1993/Accepted 2 September 1993 We screened members of a new genus of grass-associated diazotrophs (Azoarcus spp.) for the presence of cellulolytic enzymes. Out of five Azoarcus strains representing different species, only in the endorhizosphere isolate BH72, which is also capable of invading grass roots, was significant endoglucanase activity, in addition to 13-glucosidase and cellobiohydrolase activity, present. Reducing sugars were readily released from medium-viscosity carboxymethylcellulose (CMC), but neither CMC, cellulose filter strips, Avicel, cellobiose, nor D-glucose served as the sole carbon source for growth of Azoarcus spp. Clones from a plasmid library of strain BH72 expressed all three enzymes in Escherichia coli, apparently not from their own promoter. According to restriction endonuclease mapping and subclone analysis, ,-glucosidase and cellobiohydrolase activities were localized on a single 2.6-kb fragment not physically linked to a 1.45-kb fragment from which endoglucanase (egl) was expressed. Two isoenzymes of endoglucanase probably resulting from proteolytic cleavage had pl values of 6.4 and 6.1 and an apparent molecular mass of approximately 36 kDa. Cellobiohydrolase and ,I-glucosidase activity were conferred by one enzyme 41 kDa in size with a pl of 5.4, which we classified as an unspecific exoglycanase (exg) according to substrate utilization and specificity mapping; hydrolysis of various oligomeric substrates differentiated it from endoglucanase, which degraded substituted soluble cellulose derivatives but not microcrystalline cellulose. Both enzymes were not excreted but were associated with the surface of Azoarcus cells. Both activities were only slightly influenced by the presence of CMC or D-glucose in the growth medium but were enhanced by ethanol. egl was located on a large transcript 15 kb in size, which was detectable only in cells grown under microaerobic conditions on N2. Surface-bound exo- and endoglucanases with some unusual regulatory features, detected in this study in a strain which is unable to metabolize cellulose or sugars, might assist Azoarcus sp. strain BH72 in infection of grass roots. Cellulose is a constituent of lignocellulosic materials of plant origin and is composed of D-glucopyranosyl units linked by 3-1,4 bonds. For its microbial conversion complex, batteries of several enzymes are required (52). At least three different classes of hydrolytic enzymes are thought to participate: endo1,4-p-glucanase (EC 3.2.1.4), exo- l,4-cellobiohydrolase (EC 3.2.1.91), and 1-glucosidase (EC 3.2.1.21). Cellulose degradation has been reported to occur concomitantly with nitrogen fixation in aerobic, shipworm-associated bacteria (57) and in anaerobic freshwater mud and soil isolates (34). Among plant-associated nitrogen-fixing bacteria, cellulolytic activity is present in the actinomycete Franckia sp. (49) and in Rhizobium sp. (38, 43). In the legume-Rhizobium symbiosis, ultrastructural studies provided evidence that rhizobial cellulases may be involved in various steps of the infection process (8, 9). The grass-associated diazotroph Azospirillulm brasilense, although able to colonize the interior of grass roots (35, 51), is pectinolytic (55), but so far no cellulolytic activity has been reported to be present. Cellulose only supported growth of Azospirillum cells in coculture with cellulolytic bacteria (22). Corresponding author. t Present address: Max-Planck-Institut fur Terrestrische Mikrobiologie, Karl-von-Frisch-Str., D-35043 Marburg, Germany. t Present address: Pfarracker 5, D-35043 Bauerbach, Germany. We were interested in the cellulolytic features of Azoarcus spp., a new genus of gram-negative, nitrogen-fixing bacteria colonizing grass roots. Taxonomic studies located them in the beta subclass of the Proteobacteria, in which they cluster in five groups differing at the species level (47). Two of them were proposed as new named species, Azoarcus indigensT and Azoarclis commlunis (47). Members of the genusAzoarcus were found to be in root-zone-specific association with Kallar grass (Leptochloa fiusca (L.) Kunth), in which they were isolated from the root interior (46). Kallar grass is grown as a pioneer plant on salt-affected, frequently flooded, low-fertility soils in the Punjab of Pakistan (50). Indirect evidence for colonization of the root interior obtained by isolation procedures was first supported by immunofluorescence with antibodies which are Azoarcus genus specific (44). Light and electron microscopic immunogold studies of better resolution revealed that these bacteria colonize the aerenchymatic air spaces but also deeply penetrate roots into stele and stem bases (27, 28). In gnotobiotic culture, strain BH72 was capable of systemic infection of Kallar grass and rice seedlings, invading the cortex inter- and intracellularly and even penetrating into xylem vessels. Anticipating that plant polymer-degrading enzymes of bacterial origin were involved, we wanted to screen members of Azoarcus spp. for occurrence of cellulases. Here, we report cloning and characterization of two classes of cellulolytic enzymes, which share an unusual combination of features indicating they might be involved in the infection process. 7056 V()L. 175, 19)93 CELLULASES OF AZOARCUS SP. 7057 TABLE 1. Bacterial strains and plasmids" Dcscriptioin Strain or plasmid Strains A. indigenis VB32' A. comnhunis SWub3' Azoarcclus sp. strain BH72 Azoarcus sp. strain S5b2 Azoarcus sp. strain 6a3 E. coli DH5st F' Wild type Wild type Wild type Wild typc Wild type F' ecdA1 hsdRJ7 (rK m-iK+) supE44 thi-i recAI gyrA re/lAl 80/a(cZAM15 Sourcc or reference A(/acZYA-argF)1,,,X, 47 47 47 47 47 23 Plasmids pUC I9 pUCl8 Ap' ColEl replicon Apr ColEl replicon pBGV5.2 Ap'-, 5.3-kb KpnI fragment of pBGV5 cloned in pUC18 pBGVI 1.58 Ap', 1.45-kb SstI-KpsnI fragment of pBGV5 cloned in pUCI8 pBGV8.l Apr, 4.8-kb SmaI-XbalI fragment of pBGV8 cloned in pUC18 Plasmitds not shown in this talbIc arc described in Fig. 2. 59 59 This study This study This study " MATERIALS AND METHODS Bacterial strains and plasmids. The strains and plasmids used in this study are listed in Table 1. Azoarcius strains originated from roots of Kallar grass, Leptochloa fusca (L.) Kunth, grown in Pakistan, and represented five groups of Azoarclus spp. differing at the species level (47). Physical maps of plasmids are illustrated in Fig. 2. Media and growth conditions. Escherichia coli cells were grown at 37°C in Luria broth (LB) (2) or on LB agar supplemented with carbenicillin (150 pg/ml) to maintain plasmids and were grown, when indicated, with isopropyl-p3-Dthiogalactopyranoside (IPTG; 24 p.g/ml). Azoarculs spp. were grown on VM medium (47) with rotary shaking at 28°C as an inoculum for cellulase tests and rotary shaking at 37°C for cell mass production. To obtain microaerobic growth, Azoarcus sp. strain BH72 was grown at 33°C in 100 ml of liquid SM medium (46), with fourfold-strength phosphate buffer and 20 mM L-proline added, in a gas-tight sealed 1 liter Erlenmeyer vial with reciprocal shaking (70 strokes per min). Cultures were harvested when the initial 02 concentration of 1.7% in the headspace had decreased to 0.8%; atmospheric 02 was measured with a GC1 1 gas chromatograph with a thermal conductivity detector (Delsi Instruments, Suresnes, France) by using a Molecular Sieve DS column (2 m, 1/8-mm DS, 80 to 100 mesh; Alltech). To measure the expression of cellulase by Azoarcus spp., a modification of a medium proposed by Kim and Wimpenny (KW medium [29]) was used, which contained 0.4% (NH4)2S04, 0.01% NaCl, 0.01% MgSO4, 0.01% CaCl2, 0.05% yeast extract, 33 mg of Fe(III)EDTA per liter, 0.05 M potassium phosphate buffer (pH 7.0), and 0.04% medium-viscosity carboxymethylcellulose (CMC; no. C 4888; Sigma, St. Louis, Mo.). Addition of 0.5% filtersterilized D-glucose or 1.5% agar was optional. When tested for utilization of carbon sources, Azoarcus cells were grown on KW plates for 4 days, scraped off, and washed twice in saline. They were used as an inoculum (at 6 [ig of protein per ml) for test cultures in liquid KW medium without yeast extract (25 ml in 125-ml Erlenmeyer flasks) and were incubated at 37°C with rotary shaking (150 rpm) with three replicates each. Complex carbohydrates were autoclaved, and D-glucose and cellobiose were sterilized by filtration and added at 1% (wt/vol). Nitrogen-fixing growth on the respective carbon sources was tested with nitrogen-free semisolid malate medium (46), with potassium malate being replaced by the respective carbohydrate (1% [wt/vol]). Culture tubes were inoculated with cells washed as described above (at -1 [ig of protein per culture), and nitrogen fixation was determined by the acetylene reduction assay (47). Azoarctus cells for protein extractions were grown in 150 ml of liquid KW medium supplemented with 3 ml of filtersterilized ethanol per liter in screw-capped 500-ml Erlenmeyer flasks at 28°C with rotary shaking (60 rpm); after 48 h, ethanol (3 ml/liter) was again added and the cells were grown for a further 12 h. Screening of bacterial colonies for B-glucosidase, cellobiohydrolase, and endoglucanase activity. Azoarcus strains growing exponentially on VM medium were streaked on KW plates with or without D-glucose, incubatcd for 4 days at 37°C, and then incubated at 37°C with the appropriate overlay to screen for cellulase activities. Overlays consisted of 8 ml of 0.05 M potassium phosphate (pH 7.0)-buffered agarose (0.7%) containing 0.5 mg of either 4-methylumbelliferyl-3-D-glucoside (MUG; Sigma) or 4-methylumbelliferyl-3-cellobioside (MUC; Sigma) per ml for 3-glucosidase or cellobiohydrolase detection or 0.2% CMC for endoglucanase detection. After 4 to 10 h, plates were exposed to 302 nm of UV light on a transilluminator, and the active colonies were identified by the appearance of blue fluorescence. Endoglucanase activity was detected after overnight incubation by staining with Congo red (Merck, Darmstadt, Germany); the appearance of a clear-yellowish halo on a red background indicated endoglucanase activity (53). When necessary, plates were incubated for 5 min with 10 ml of 1 M HCI to enhance contrast (49). To detect 3-glucosidase and cellobiohydrolase activity in E. coli clones, cells were grown overnight on LB plates with IPTG and then were incubated with a MUG or MUC overlay for I to 4 h at 37°C. To visualize expression of endoglucanase, LB plates were overlaid with CMC-agarose supplemented with carbenicillin and IPTG and clones were inoculated deeply into the agar. After incubation at 37°C overnight, plates were stained with Congo red as described above. General cloning and DNA-analysis techniques. Plasmids from E. coli were prepared by the alkaline lysis procedure of Birnboim and Doly (7), as modified by Ausubel and Frederick (2), for both small-scale and large-scale preparations. Restriction endonuclease digestions, ligations, analysis and recovery of DNA fragments on agarose gels, and autoradiography were carried out according to standard protocols (2). DNA fragments were labeled with 32p with a random primed labeling kit (Boehringer Mannheim). T4 DNA ligase and calf intestine alkaline phosphatase were purchased from Bethesda Research Laboratories and Promega, respectively. E. coli competent 7058 REINHOLD-HUREK ET AL. cells were prepared and transformed according to the method of Kushner (30). Construction and screening of a representative genomic library. High-molecular-weight genomic DNA was prepared from Azoarcus sp. strain BH72 by the method of Marmur (37), modified as described previously (45), except that further purification by CsCl gradient was omitted. DNA (240 pLg) was partially digested with Sau3AI and was size fractionated with a sucrose gradient (2). Fragments 5 to 15 kb in size were used for construction of a library in pUC19 which was digested to completion with BamHI and dephosphorylated. Vector and insert DNA in equimolar amounts, as well as vector DNA alone as a control, were ligated and subsequently used for transformation of E. coli. A total of 2,880 clones positive in blue-white selection with 5-bromo-4-chloro-3-indolyl-,3-D-galactopyranoside (20 ,ug/ml of LB agar) were transferred to microtiter plates and replica plated. Bacterial colonies were screened for expression of 3-glucosidase, cellobiohydrolase, and endoglucanase. Positive clones were restreaked and tested again to confirm activity. Preparation of whole-cell extracts and cell fractionation. For enzymatic quantitation, Azoarcus sp. strain BH72 cells were grown on liquid KW medium supplemented with ethanol unless stated otherwise. Cells from 15 to 25 ml of culture were processed at 4°C; they were pelleted, washed once in 10 ml of 10 mM Tris-HCI (pH 8.0), and suspended in I ml of the same buffer to give an optical density (at 578 nm) of -60. To obtain total cellular extracts, cells were disrupted by ultrasonication with a B-10 cell disrupter (Branson Sonic Power Corp.) on ice at a 40-W output for three cycles of 30 s each. Cell debris was pelleted by centrifugation at 13,000 x g, and supernatants were used for enzyme assays. For extraction of proteins with Triton X-100, washed cells were shaken for 10 min at 28°C and 60 rpm in Tris-HCl buffer containing 0.1% Triton X-100 and 10 mM Na2-EDTA. After centrifugation at 9,000 x g for 10 min, the supernatant was used. The first step of cell fractionation was performed by centrifugation at 10,000 x g for 10 min. To obtain the extracellular fraction, the culture supernatant was further centrifuged at 20,000 x g for 25 min and concentrated 30-fold by ultrafiltration through an Amicon YM10 membrane. Pelleted cells were subjected to fractionation by the osmotic shock method as described by Cornelis et al. (14), except that the 25% sucrose solution was made up in 10 mM Tris-HCl (pH 7.0). Residual cells from which the periplasmic fraction had been prepared were ultrasonicated as described above to obtain the cytoplasmic fraction. The total cellular protein of E. coli was prepared by ultrasonication as described above or by the freeze-thaw method. Cells suspended in Tris-HCI buffer were rapidly frozen and thawed (dry ice-cooled ethanol, 30°C). Determination of protein concentrations. The protein content of cell extracts was determined by the Bio-Rad protein assay (Bio-Rad Laboratories Ltd.). Cellular protein was estimated by the micro-Goa method (4). IEF, zymograms, and SDS-PAGE. For isoelectric focusing (IEF) LKB Ampholine PAGplates in the pH range 3.5 to 9.5 (no. 1804-101, LKB) were used with a Multiphor II electrophoresis chamber (LKB). Electrophoresis and Coomassie brilliant blue staining were carried out according to the manufacturer's instructions, except that experiments were run at lower voltage (400 V) for 2.5 h. For zymograms, we followed the outlines given by Coughlan (15). Agarose overlays for IEF gels contained 0.8% agarose in 0.05 M potassium phosphate buffer (pH 7.0) and the respective substrate at 0.01% (MUG or MUC) or 0.1% (CMC). After incubation at 37°C, overlays J. BA(CF[-'RIOL. proceeded as described above, except that CMC overlays were washed for 10 min in phosphate buffer prior to Congo red staining. Proteins from extracts obtained by ultrasonication were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on 12% polyacrylamide gels as described by Laemmli (31). Enzyme assays. ,B-Glucosidase and cellobiohydrolase activities were quantified with the chromophoric substrates 2chloro-4-nitrophenyl-r3-D-glucoside or 2-chloro-4-nitrophenyl3-cellobioside (11) at 2 mM in 200 mM potassium phosphate buffer (pH 6.8) with 20 [tg of cell protein extract added (heat inactivated at 95°C as a negative control); the increase in A405 was monitored spectrophotometrically upon incubation at 37°C. One unit of specific activity is defined as 1 ,umol of chromophore released min'-'' mg protein', with the extinction coefficient at 405 nm equalling 9,000 M-' cm '. Qualitative assays with other substrates with the same chromophore were similarly performed. Enzymic hydrolysis of fluorogenic substrates was qualitatively monitored by UV illumination. Endoglucanase activity was assayed with the chromogenic substrate Remazol brilliant blue R-CMC according to the method of Cleary (13) with modifications. Twenty micrograms of cell protein extract was added to 0.2% Remazol brilliant blue R-CMC dissolved in 50 mM Tris-HCl buffer (pH 6.8), and samples were incubated at 37°C for 10 h. The specific increase in A5910 was calculated, taking into account a negative control (heat-inactivated extract), as AA599 h -' mg of protein . Qualitative tests for degradation of Ostazin brilliant redhydroxyethylcellulose (Sigma) were carried out as described by Biely et al. (6) with the buffer described above. Enzymatic hydrolysis of Remazol brilliant blue R-Avicell (0.2%) was tested with 160 mM Tris-HCI (pH 6.8) and calculated from the increase in A5,0 of the supernatant after centrifugation. Reducing sugars were quantified by a modification of the Somogyi-Nelson method as described by Marais et al. (36). Northern analysis. RNA was isolated from bacterial cells from 30 ml of an exponentially growing culture by the hotphenol method (1). Northern (RNA) blotting and hybridization were carried out according to standard protocols (2). After separation of nucleic acids on a 0.9% agarose gel, they were blotted onto a nylon membrane (Hybond N; Amersham), fixed to the membrane by baking at 80°C, and finally were hybridized at 65°C. RESULTS Occurrence of cellulolytic activities in Azoarcus strains. Representative members of all five Azoarcus spp. were screened for the presence of hydrolytic enzymes known to be involved in cellulose degradation. Screening of colonies was carried out with fluorophoric substrates to detect exo-1,4-,Bglucanases (,-glucosidase, EC 3.2.1.21; exo-1,4-cellobiohydrolase, EC 3.2.1.91), and a CMC-plate-clearing test was used to detect endo-1,4-p-D-glucanase (EC 3.2.1.4). All but one strain, 6a3, were positive or weakly positive for f-glucosidase, whereas cellobiohydrolase activity was detected only in A. communis SWub3T. Endoglucanase activity was only detectable in strain BH72. Whether D-glucose was added to KW medium or not, no significant difference was observed. However, when ethanol was added to glucose-free KW agar to promote colony growth, detection levels were considerably enhanced. Strong 3-glucosidase activity could be found in strains BH72 and SWub3T (Fig. IA). Although cellobiohydrolase activity was generally weaker, it could be detected in all four strains (Fig. IB). In contrast, considerable endoglucanase VOL. 175, 19933 CELLULASES OF AZOARCUS SP. 7059 FIG. 1. Cellulolytic activities of Azoarcus sp. strains in colony assays. Bacteria werc grown on KW agar plates supplemented with ethnilol (6 ml/liter), overlaid with the respective substrate and viewed under UV illumination (A and B) or evaluated by Congo red staining (C). Overlays contained fluorophoric substrates to detect P-glucosidase (MUG [A]) or cellobiohydrolase (MUC [B]) activity or CMC to detect endoglucanasc activity (C) and were incubated for 4 h, 10 h, or overnight, respectively. Numbers on plates: 1, Azowrcus sp. strain BH72; 2, A. indigens VB32"'; 3, A. communis SWub3?, 4, Azoarcius sp. strain S5b2. activity was only seen in strain BH72; A. commutnis SWub3T showed only very weak activity (Fig. IC). Also, with ethanol as a carbon source, none of these enzymic activities were found in strain 6a3. All Azoarculs strains were tested for degradation of cellulose filter strips (9 by 1 cm; Whatman 40) in culture solutions consisting of KW medium without CMC, KW medium supplemented with either D-glucose, potassium malate, or ethanol, or KW medium in the absence of any additional carbon source. Even within 1 month of incubation at 37°C, no structural disintegration of the filter strips was visible, indicating that there was no significant cellulolytic activity on this substrate. However, cells of strain BH72 readily released reducing sugars from soluble cellulose CMC. When cells were grown on liquid KW medium with 0.02% CMC and 6 ml of ethanol per liter for 3 days, culture supernatants contained 12 ± 4 pg of reducing sugar per ml more than controls, corresponding to 29 ± 10 p.g per mg of protein. Controls consisted of the sum of the reducing sugar contents of uninoculated medium and inoculated, CMC-free cultures. All assays were performed in triplicate. Utilization of cellulose and other carbohydrates. Utilization by Azoarculs sp. strain BH72 of the following cellulose types, derivatives, and hydrolysis products was tested: medium-viscosity CMC, cx-cellulose fiber (C 8002; Sigma), microcrystalline Avicel (Merck), cellobiose, and D-glucose. Growth was monitored by measurement of optical density (at 578 nm) or, in the case of insoluble cellulose, by microscopic examination after 2, 7, and 12 days of incubation, and at the end cellular protein was determined. There was no evidence for growth on any of the carbohydrates. The protein contents found with D-glucose (134 ± 6 pg), cellobiose (132 ± 6 kg), and CMC (129 ± 9 pg) coincided with those found in control cultures (131 ± 8 p.g). The same compounds also did not support nitrogen-fixing growth or nitrogen fixation. Within 2 weeks of incubation, there was no formation of a subsurface pellicle, which is typical for nitrogen-fixing growth of Azoarclus spp., and no acetylene reduction could be detected. Cloning of cellulases and expression in E. coli. E. coli clones of a total genomic library from Azoarcuts sp. strain BH72 constructed in pUC19 were screened by the overlay technique for all three cellulase activities found in this strain: P-glucosidase, exo-1,4-cellobiohydrolase, and endo-1,4-3-glucanase. Of 2,880 colonies tested, seven clones showed P-glucosidase activity within 1 to 2 h. The same clones according to their positions in the microtiter plates also showed a positive, albeit weaker, reaction for cellobiohydrolase, which was well detectable within 4 h of incubation. As deduced from different positions in the microtiter plates, five different clones were positive for endoglucanase activity. To allow an initial grouping of the clones, PstI restriction patterns of plasmid DNA from individual clones were compared. Clones pBGV7 and pBGV6, as well as pBGV9 and pBGV8, had almost identical patterns, so that only one of each was further analyzed. P-Glucosidase and endoglucanase clones were subjected to endonuclease restriction mapping (Fig. 2). Restriction maps of clones pBGV10, pBGV12, pBGV6, and pBGV8 overlapped in a central genomic region which was likely to carry the 3-glucosidase (g,Ic) and cellobiohydrolase (cbh) genes (Fig. 2A). To confirm this assumption and to localize the genes more accurately, fragments were subcloned into pUCI9 in the same orientation as in clone pBGV6. A 2.6-kb subclone (pBGV6.6) expressed both activities strongly, whereas smaller subclones had no or only weak 3-glucosidase activity (Fig. 2A), indicating that parts of the gene necessary for function might be located upstream from the HindlIl site. Also the endoglucanase activity (egi) seemed to be located as one gene in an overlapping region of pBGV2, pBGV3, pBGVl 1, pBGVI, and pBGV5. Subcloning could localize the activity on a 1.45-kb SstI-KpnI fragment (Fig. 2B, pBGV11.5). There was no evidence for a second endoglucanase gene located within 5 kb upstream or 12.5 kb downstream of egl in this genomic region (pBGV2.1 and pBGV2.2, respectively). Restriction maps of genomic regions harboring 3-glucosidase or endoglucanase gave no evidence for physical linkage. Because all clones of one type which we had obtained from screening of the library were in the same orientation with respect to the IacZ promoter of pUC 19, we anticipated that cellulase genes were not under control of their own promoter when expressed in E. coli. When the orientation of egl-active fragments (pBGV1 1.5 or pBGV5) was reversed (pBGV1 1.58 or pBGV5.2, respectively), no endoglucanase activity could be detected. Similarly, reversion of a glc- and cbh-active fragment (pBGV8 to pBGV8. 1) resulted in loss of 3-glucosidasc or cellobiohydrolase activity, indicating that the respective Azo- 7060 REINHOLD-HUREK ET AL. J. BA(TFI-RIOL.. ½½~~~~~~~~~-- 1 ½ 2 3 2 3 2 3 4, Phenotype glc cbh glc/cbh 7.0 - pBGV10 pBGV12 6.5 - pBGV6 pBGV8 pBGV6.6 6.0 - pBGV6.6 K pBGV6.6H 1kb II egl pBGV2 pBGV3 pBGVll pBGV1 pBGVS pBGV21 pBGV2.2 pBGV5.1 pBGV11.5 lkb FIG. 2. Physical map of genomic regions carrying putative ,-glucosidase and cellobiohydrolase (glc/cbh) genes corresponding to an exoglycanase (exg) and the endoglucanase gene (egi) of Azoarcus sp. strain BH72 and plasmid inserts in pUCI9 transcribed from the 1acZ promoter (right side of figure). For clones pBGV2 and pBGV3, the position of the second Sall site from the right is one of two possible locations, the second being 0.7 kb to the left. The enzymatic activities of P-glucosidase (glc), cellobiohydrolase (cbh), and endoglucanase (egl) determined by colony assays are given to the right: +, positive; w, weak activity; -, no activity detectable. arcus promoters either were not located on the fragments or were not active in E. coli. In order to compare cloned and wild-type cellulases, enzyme activities were located on IEF gels by zymograms. Sonicated cell extracts of Azoarcus sp. strain BH72 were not separated into distinct patterns in IEF-PAG plates but gave smears, probably due to the presence of lipids. We overcame this problem by extracting samples once with chloroform, a treat- ment compatible with endoglucanase but detrimental to ,B-glucosidase activity. The latter was resistant to Triton X-100 and EDTA, which we used to release cell surface proteins, including ,-glucosidase (see below). ,B-Glucosidase and cellobiohydrolase activities of the E. coli clone were focused at the same pH (pl 5.4); wild-type enzymes behaved the same way (Fig. 3). In contrast, endoglucanase activity obtained from a clone carrying a small 1.45-kb insert exhibited two distinct bands with pIs of 6.4 and 6.0, respectively (Fig. 4). These two isoenzymes also appeared in Azoarcus spp., although with different intensity ratios. The apparent molecular masses of the cellulases expressed by E. coli were estimated by SDS-PAGE (Fig. 5). A new, 41-kDa band appeared in E. coli cells expressing 3-glucosidase and cellobiohydrolase from a 2.7-kb fragment when induced with IPTG, which was not seen in uninduced cells or induced cells carrying insert-free plasmid (Fig. SA). When endoglucanase was induced from a 1.45-kb fragment, newly formed bands were not clearly visible (not shown). However, with a different method of protein extraction (SDS-soluble proteins [45]) a new, 36-kDa band was clearly visible (Fig. SB). Enzymatic properties of the cloned cellulases. Substrate 5.1 - C A B FIG. 3. IEF and zymograms of cloned (lanes 2) and wild-type (lanes 3) exoglucanase. Protein extracts were obtained by the freezethaw method from E. coli cells carrying plasmid pUC19 as a negative control (lane 1) or pBGV6.6 (lanes 2) and from cell surface proteins prepared by Triton X- 100 washes from Azoarcus sp. strain BH72 (lanes 3). Gels were stained with Coomassie brilliant blue (A) and for 13-glucosidase (B) or cellobiohydrolase (C) activity with overlays incubated for 15 or 45 min, respectively. Twenty and 8 ,utg of protein (lanes Al and A2 and lane A3, respectively) were loaded for Coomassie staining; 10 and 12 ,ug of protein (lanes B2 and C2 and lanes B3 and C3, respectively) were loaded for zymograms. Numbers to the left correspond to pls of protein standards. Enzymatically activc bands at pl 5.1 are found in both cloned (lanes 2) and wild-type (lanes 3) preparations. utilization was determined for Azoarcus sp. strain BH72 cells and for individual cloned cellulases expressed in E. coli and tested with cell extracts (Table 2). E. coli cells carrying insert-free pUC19 were negative for all substrates tested (not shown). The substrate specificity of cloned ,B-glucosidase and cellobiohydrolase (exoglucanase) was clearly different from that of endoglucanase. The activity of the cellobiohydrolase and ,B-glucosidase is incompatible with substituted, soluble cellulose derivatives, but a variety of oligomeric sugars are substrates, indicating that the presumed glc/cbh gene codes for an exoglucanase (exg). The activity of the endocellulase is clearly in contrast with this behavior. Microcrystalline cellulose (Avicel) was not attacked by any of the enzymes or the wild-type Azoarcus strain. This clearcut difference in substrate specificity allowed us to differentiate activities in cell extracts containing both enzymes. To elucidate the mode of action and substrate specificity of cellulases, cleavage sites of 3-glucosidase and cellobiohydrolase and endoglucanase were determined with cell extracts from E. coli clones. 1,4-,-Cellooligosaccharides carrying a fluorophoric group (MUGj) were tested as substrates for the respective enzymes (4 h of incubation), and the products were separated by thin-layer chromatography. Inspection of the products obtained (Fig. 6) indicated that endoglucanase preferably attacked oligosaccharides larger than cellobiose and released large oligomers when sufficiently long substrates were used. 3-Glucosidase attacked dimers as expected (Table 2), but it also attacked larger substrates, preferably cleaving the bond at the fluorophoric group (nonreducing end). Apparently, more-highly oligomeric substrates were also cleaved at internal bonds. Temperature and pH profiles of exoglucanase and endoglucanase activities were obtained from cell extracts of Azoarcus sp. strain BH72 (not shown). Profiles for exoglucanase measured either as 3-glucosidase or as cellobiohydrolase activity correlated satisfactorily. The specific activity was 3- to 10-fold VOL. 1 75, 19'93 CELLULASES OF AZOARCUS SP. 1 2 2 3 TABLE 2. Substrate specificity of cloned cellulases and wholc cells of Azoarcus sp. strain BH72 3 8.1 - Substratc specificity of: Exoglu- Endoglu- Whole canaIsc" canase" cells" Substrate 7.06.5- MUG MUC 2-Chloro-4-nitrophenyl-f3-D-glucoside +" - + + - + - + + 2-Chloro-4-nitrophenyl-3-cellobioside + + + + - - + 2-Chloro-4-nitrophenyl-l3-lactoside p-Nitrophenyl-P-D-xyloside 4-Methylumbelliferyl-13-D-xyloside 4-Methylumbelliferyl-l3-xylobioside CMC' Remazol brilliant blue R-CMC Remazol brilliant blue-Avicel Ostazin brilliant red-hydroxyethylcellulose 6.05.1- A b FIG. 4. IEF and zymograms of cloned (lanes 2) and wild-type (lanes 3) endoglucanase. Protein extracts were obtained by the freeze-thaw method from E. coli cells carrying plasmid pUC19 as a negative control (lane 1) or pBGV1 1.5 (lanes 2) and from whole-cell preparations from Azoarculs sp. strain BH72 extracted with chloroform (lanes 3). Gels were stained with Coomassie brilliant blue (A) or for endoglucanase activity (B) with overlays incubated for 3.5 h. Fifteen, 7, and 12 xg of protein were loaded on lanes Al, A2, and A3, respectively, for Coomassie staining; 15 and 24 ,ug of protein were loaded on lanes B2 and B3, respectively, for zymograms. Numbers to the left correspond to pIs of protein standards. Two isoenzymes with pIs of 6.4 and 6.0, respectively, appear from cloned (lanes 2) and wild-type (lanes 3) preparations. higher for 3-glucosidase than for cellobiohydrolase. The pH optima were 6.4 and 6.0 for exoglucanase and endoglucanase, respectively. The temperature optima coincided at 46°C. Cellular location of cellulases in Azoarcus sp. Determination A 1 2 3 4 --- B 97.4 66.2 42.7 31.2 21.5 5 97.4 - 66.2 a 42.7 _ 6 7061 7 - + + + ND" ND ND + + + + - - " Cell extracts of E. coli cells carrying plasmid pBGV6.6. "'Cell extracts of E. coli cells carrying plasmid pBGVI 1.5. 'Azoarcus sp. strain BH72 cells grown on KW medium, with complex substrates added during growth. d Cleavage was estimated by appearance of the released chromophore, seen as fluorescence or color. ' ND, not determined. /'Cleavage was estimated by appearance of reducing sugars. of specific activities after cell fractionation revealed that both cellulases were mainly surface associated (Table 3). Exoglucanase quantified as ,B-glucosidase activity was not detectable in the culture supernatant, and only low specific activities were found in the cytoplasmic fraction. To enhance limits of detection, proteins in the culture supernatant were concentrated by ultrafiltration. Even after 30-fold concentration, no 3-glucosidase activity was detected. To evaluate whether ultrafiltration might reduce endo- and exoglucanase activities, we concentrated an aliquot of the cell surface fraction; specific activities were not affected (data not shown). For endoglucanase activity, we obtained a similar pattern of distribution. Endoglucanase activity could be detected neither in the culture supernatant nor in the cytoplasmic fraction but was detected only in the cell surface and periplasmic fractions. Regulation of cellulases and transcription of endoglucanase. To test the effect of carbon sources on the expression "t" _ "f . MUG, + endoglucanase MUG, + exoglucanase El-31.2 21.5 _ _ El-Ei--- 4 El-El-EI-. - FIG. 5. Apparent molecular weights of exoglucanase (A) and endoglucanase (B) expressed in E. coli. Whole-cell extracts prepared by ultrasonication (A) or SDS extraction (B) from E. coli cells carrying plasmid pUC19 (lanes 2 and 6), pBGV6.6 (lanes 3 and 4), or pBGVI 1.5 (lane 7) were separated by SDS-PAGE and stained with Coomassie brilliant blue. Approximately 40 p.g of protein extracts per lane was loaded. For lane 3, cellulases were repressed by addition of 1% glucose and omission of IPTG. For lanes 4 and 7, cellulases were induced by omission of glucose and addition of IPTG. Lanes 1 and 5 show molecular mass markers (sizes are given in kilodaltons). Additional protein bands appearing in cells expressing cellulases are marked with arrowheads and have apparent molecular masses of 41 kDa (A) and 36 kDa (B). 4 El4El4 ElE lEl- lE4lE4 FIG. 6. Specificity mapping of cloned endoglucanase and exoglucanase with MUGn derived from the cellooligosaccharides (n = 1 to 6). Protein extracts from E. coli cells carrying plasmid pBGBI 1.5 or pBGV6.6 were incubated with MUG,,, and products were separated by thin-layer chromatography (12). Solid circles indicate the positions of the fluorophoric group. Arrows mark cleavage sites estimated from the mobility of fluorescing bands. 70)62 REINHOLD-HUREK ET AL. J. BACT11RIOL. TABLE 3. Cellular location of cellulase activity in Azoarcus sp. strain BH72 kb Endoglucanase of -tity(Exoglucmngasc acprteprotcin ivit (U/mg ± SD) Fraction Friiction Not detectable Not detectable Culture supernatant Concentrated culture supernatant" Cell surface fraction Periplasmic fraction Cytoplasmic fraction Total cells 107 92 5 30 activity (AA590 h mg of protein -- SD)" Not detectable Not detectable 2.15 ± 0.30 2.50 ± 0.04 Not detcctable 0.71 ± 0.05 ± 7 ± 1 ± (0 ± 0 14.5 -* 9.5 7.54.4 - 2.4 - 1.4 - SpCcific activity of -glUcosidase was determined with 2-chloro-4-nitrophenyl as substrate. P-i)-glucoside / Specific increase in absorption was determined with Remazol brilliant blue R-CMC as a substrate. High-molccukar-wcight fraction was concentrated 30-fold by ultratfiltration. " 0.24 of cellulases by Azoarcuis sp. strain BH72, cells were grown for 48 h on the respective carbon sources, and specific enzyme activities in cell extracts were determined. Both exoglucanase and endoglucanase showed similar patterns of activity (Table 4) and were expressed under all conditions tested. A slight induction by the substrate CMC was seen only for exoglucanase and not for endoglucanase. There was no strong repression by the end product D-glucose. Other carbon sources which well support growth of Azoarcus cells acted differently: DL-malate hardly affected or even reduced exo- and endoglucanase activity, whereas ethanol had the strongest inducing effect, increasing both activities 4- and 1.6-fold, respectively. Oxygen also had a pronounced effect on expression of endoglucanase. In Northern analysis with the 1.45-kb SstI-Pstl fragment as a probe, we found a large, -15-kb transcript on which egl was located (Fig. 7). It was only detected in cells grown microaerobically on N2 and was not detected when cells were grown aerobically on combined nitrogen (Fig. 7). As a control for the absence of genomic contamination in the RNA preparation of Fig. 7A, we visualized nucleic acids by ethidium bromide staining after gel separation (not shown). Additionally, we hybridized Northern blots from the same preparation with genomic probes for other Azoarcus genes, namely nifH and nifDK; hybridizing bands were obtained only at lower molecular sizes (-6.2 kb), indicating therc was no genomic contamination at 15 kb (not shown). TABLE 4. Effect of carbon sources on cellulase activities of Azoarcus sp. strain BH72" Cairbon source added' Yeast Yeast Yeast Yeast Yeast extract extract extract extract extract + + + + CMC CMC + glucose CMC + malate CMC + ethanol P-Glucosidase Endoglucanasc activity (AA590 protein + SD) protcin ± SD) 8.56 ± 0.01 10.67 ± 0.t)8 9.55 ± (1.11 12.22 ± 0.88 41.11 ± 1.0() 0.030 ± 0.001 0.028 ± 0.001 0.026 ± 0.002 0.019 ± 0.0(1 0.044 ± 0.000 activity (U/mg ot h mg of " Specific aictivitics of cnzymes in cell extracts were determined after ultriasonication of total cells. ' Carbon sources in mineral salts of KW medium were as follows: yeast extract, 11.5 g/liter; CMC, 11.4 g/liter; 1)-glucose, 1) g/liter; il.-malic acid, 2.5 g/liter (neutralized with KOII); ethalnol, 4 ml/litcr. ++ 02 + FIG. 7. Northern blot analysis of endoglucanase expression. RNA was isolated from Azoar-cus sp. strain BH72 grown on malate microaerobically on N, (A) or aerobically on combined nitrogen (B). Twelve micrograms of RNA in each lane was separated by agarose gel electrophoresis, blotted onto nylon membrane, and hybridized to a 329P-labeled endoglucanase probe consisting of the 1.45-kb insert of pBGVI 1.5. The arrow marks labeled transcript. Symbols: + +, high 02 concentration (aerobic); +, low 0O concentration (microaerobic). DISCUSSION Although cnzymes known to be involved in cellulase breakdown were present in four of five Azoarcus spp., we focused on Azoarcus sp. strain BH72 for further studies because it had the widest spectrum of cellulolytic enzymes. From this root-invading diazotroph, we detected and cloned two types, an exoglucanase and an endoglucanase. Although strain BH72 showed 3-glucosidase as well as cellobiohydrolase activities, we obtained evidence that both are conferred by only one enzyme, as follows. (i) All clones from a representative genomic library expressing one activity also express the other; however, with 2.7 kb, the smallest fully positive subclone was still large enough to harbor two open reading frames. (ii) Wild-type and cloned enzymes of both activities have the same pl (5.4). (iii) Activity profiles of pH and temperature dependence coincide. (iv) From the smallest fully active subclone, only one major protein is expressed (41 kDa). (v) The cloned enzyme not only cleaves cellobiose or releases it but it also hydrolyzes larger oligomeric sugars, releasing, e.g., cellotriose. Furthermore, since aryl-3-D-xylosides are cleaved, the enzymc can best be characterized as an unspecific exoglycanase. According to its degradation pattern of chromophoric cellooligosaccharides (Fig. 6), the enzyme can tentatively be classified as a family F cellulase (12, 24). To this transition group of cellulases and xylanases belong an exoglucanase (Cex) from Cellulomonas fimi (19) and a xylanase (XynZ) from Clostridium thermocellum (20). Although evidence for two endoglucanase isoenzymes was obtained from zymograms, both forms are probably derived from one gene product. The 36-kDa protein would require an open reading frame -1 kb in size, which would not allow expression of a second open reading frame of this size from the 1.45-kb fragment if they were not overlapping. Two bands in Voi.- 1 75? 1 993 IEF presumably result from proteolytic cleavage. Cleavage sites appear to be identical in Azoarcuts sp. strain BH72 and E. coli, because the cloned and wild-type enzymes have the same pIs. The appearance of two active bands 55 and 53 kDa in size was probably due to proteolytic cleavage when cloned C. thermocellum celB was expressed (3). Also, cellulases of C. fimni (CenB [41]) and Clostridium cell/ulolytictum (EGCCA [17]) were reported to be subjected to proteolytic attack. As in the case of Pseudomonias solanacearuim (25, 26), the second product might also represent a mature form after removal of a signal sequence. Specificity mapping (Fig. 6) allows a preliminary classification of the endocellulase in family D (12), although sequence analysis should be decisive in this matter. It is remarkable that Azoarcus sp. strain BH72, although possessing a cellulolytic system, cannot grow on cellulose or on its breakdown products. In addition to filter paper, CMC, cellobiose, and D-glucose, for which results are presented here, Azoarcus spp. grew on none of 49 carbohydrates which we tested for taxonomic purposes (47). There are cellulase-positive eubacteria which are regarded as not truly cellulolytic because they show no or weak growth on celluloses; in Erwinia chrysanit/lemiii (10) or P. solanaceanlini (48), cellulases might assist in releasing nutrients from the host plant; also a cellulase was detected in Franckia sp., which is the symbiont of many actinorhizal plants, but growth on filter paper was not obvious, although growth on sugars occurs (49). Bacteroides ruminicola does not use celluloses for growth, but breakdown products such as cellodextrines and cellobiose are used (39). In contrast, extreme nutritional constraints which do not allow consumption of any of the breakdown products as in Azoarcuts spp. are unique among cellulose-degrading bacteria. This raises the intriguing question of which physiological or ecological function such an enzyme system might have for these bacteria. Bearing in mind that they are capable of infecting grass roots where they can penetrate into the stele and that they were even found in the stem base of Kallar grass grown in situ (27, 28), one might anticipate that cellulases are involved in the infection process, allowing Azoarcus cells to reach a certain microhabitat. Also, an unusual combination of some further features of this cellulolytic system point in this direction. In some gram-positive cellulolytic bacteria, cellulases are surface associated. In C. thermocellum, cellulases are organized in a complex, the cellulosome, which is cell surfaceassociated at some stages of the process of cellulose degradation (32). Also, in Ruminococcus albus, most of the cellulase activity was found to be cell bound (58). In other gram-positive bacteria, cellulases are present in the culture supernatant, e.g., in C. fimi (33), Streptomyces lividans (54), and Franckia sp. (49); many gram-negative bacteria also release cellulases, e.g., Fibrobacter sluccinogenes (40) and Prevotella rluminicola (39). In our context, gram-negative phytopathogens are of special interest; P. solanacearlum (26) and E. chrysanthemi (10) excrete endoglucanases into the culture supernatant. In contrast, exoand endoglucanase of the gram-negative Azoarcuts spp. are cell surface associated and might therefore mediate a more localized digestion of plant polymers in comparison with phytopathogens, causing less damage to the host. Most bacterial cellulases seem to be regulated by inductionrepression mechanisms depending on the carbon source. In accordance with such observations, it has been suggested that, e.g., several cel genes of C. therrnocell/tm are regulated by a mechanism analogous to catabolite repression (42). However, in some rumen bacteria, such as F. succinogenes (40) and P. ruiminicola (39), there are more constitutively expressed cellulases which are not repressed by D-glucose or cellobiose. Both CELLULASES OF AZOARCUS SP. 7063 cellulases of Azoarcus sp. strain BH72 also do not appear to be subject to strong induction by CMC or repression by D-glucose. Although genes are not physically linked, they seem to be under coordinate regulation. Taking into account the fact that this strain is restricted in use of carbohydrates for growth, end product inhibition or catabolite repression would be of no obvious physiological advantage. In contrast to Franckia sp., carbon sources supporting growth well (49) do not necessarily enhance cellulase activity, as exemplified by malic acid. The only substrate we found to significantly enhance activity is ethanol, a substance which might be of importance in the Azoarcuts habitat. When aeration of roots is reduced by flooding, ethanol concentrations in roots cain rise immediately (16), a situation which occurs in rice and Kallar grass culture; in root tips, ethanol can always be detected, even under atmospheric oxygen pressure (5). Induction of endoglucanase by ethanol and microaerobic conditions in strain BH72 might thus support penetration of roots under these conditions. An unusual feature of the Azoarcus endoglucanase is its location on a large transcript -15 kb in size. Evidence for a large transcript also comes from the lack of transcription of endoglucanase from its own promoter in E. coli. In contrast, most bacterial cellulases studied so far are transcribed monocistronically. Also, E. chrysanthemi (21) and P. solanacearlum (48) endoglucanases appear to be transcribed from their own promoters. To our knowledge, the only reported evidence for polycistronic transcription comes from primer extension experiments with P. ruminicola endoglucanase (56). So far, we have no data concerning genes being cotranscribed with eg/. In several bacteria, multiple cellulases are present, such as in C. fimni, C thermocellum, R. albus (e.g., see reference 18), and E. chrysanthemi (10). InAzoarcius sp. strain BH72, no evidence was obtained for such a multiplicity either from the results of library screening or from low-stringency hybridization with the homologous egl probe (data not shown). Therefore, mutational analysis might give clear answers about the putative role of this enzyme in the infection process. Biochemical evidence for an unusual type of regulation of cellulases will be further investigated genetically by transcriptional fusions with reporter genes. ACKNOWLEDGMENTS B.R.-H. and T.H. were supported by postdoctoral training grants awarded by the EC. We thank Marcelle Holsters for hosting us in her laboratory and Jan Desomer for valuable suggestions. We are grateful to Karel Spruyt, Vera Vermaercke, and Claudia Nickel-Reuter (Marburg) for help with the artwork. REFERENCES 1. Aiba, et al. 1981. Evidence for two functional gal/ promoters in intact Eschericlhia coli cells. J. Biol. Chem. 256:11905-1 19 t1). 2. Ausubel, I., and M. Frederick. 1987. Current protocols in molecular biology. John Wiley & Sons, New York. 3. Beguin, P., P. Cornet, and J. Millet. 1983. Identification of the endoglucanase encoded by the celB gene of Clostuidiiui,i tlilermoce/llum. Biochimie 65:495-5(t(. 4. Bergersen, F. J. 198t). Measuremenlt of nitrogen fixation by direct means, p. 65-110. In F. J. Bergersen (ed.), Methods for evaluating biological nitrogen fixation. John Wiley and Sons Ltd., Chichester, United Kingdom. 5. Betz, A. 1957. Zur Atmung wachsender Wurzelspitzen. Planta 50:122-143. 6. Biely, P., D. Mislovicova, and R. Toman. 1985. Soluble chromogenic substrates for the assay of endo-1,4-3-xylanases and endo1,4-p-glucanases. Anal. Biochem. 144:142-146. 7. Birnboim, H. C., and J. Doly. 1979. A rapid alkaline extraction 7064 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. J. BAC1 ECRIOL. REINHOLD-HUREK ET AL. procedure for screening recombinant plasmid DNA. Nucleic Acids Res. 7:1513-1523. Callaham, D., and J. Torrey. 1981. The structural basis for infection of root hairs of Trifolium repens by Rhizobium. Can. J. Microbiol. 59:1647-1664. Chalifour, F., and A. Benhamou. 1989. Indirect evidence for cellulase production by Rhizobium in pea root nodules during bacteroid differentiation: cytochemical aspects of cellulose breakdown in rhizobial droplets. Can. J. Microbiol. 35:821-829. Chippaux, M. 1988. Genetics of cellulases in Erwinia chrysanthemi, p. 219-234. In J.-P. Aubert, P. Beguin, and J. Millet (ed.), Biochemistry and genetics of cellulose degradation. Academic Press, London. Claeyssens, M. 1988. The use of chromophoric substrates and specific assays in the study of structure-activity relationships of cellulolytic enzymes, p. 393-397. In J. P. Aubert, P. Beguin, and J. Millet (ed.), Biochemistry and genetics of cellulose degradation. Academic Press, London. Claeyssens, M., and B. Henrissat. 1992. Specificity mapping of cellulolytic enzymes: classification into families of structurally related proteins confirmed by biochemical analysis. Protein Sci. 1:1293-1297. Cleary, B. V. 1988. Soluble, dye-labeled polysaccharides for the assay of endohydrolases. Methods Enzymol. 160:74-86. Cornelis, P., C. Digneffe, and K. Willemont. 1982. Cloning and expression of a Bacillus coagulans amylase gene in Escherichia coli. Mol. Gen. Genet. 186:507-511. Coughlan, M. P. 1988. Staining techniques for the detection of the individual components of cellulolytic enzyme systems. Methods Enzymol. 160:135-144. Crawford, R. M. M., and M. Baines. 1977. Tolerance of anoxia and ethanol metabolism in tree roots. New Phytol. 79:519-526. Fierobe, H.-P., C. Gaudin, A. Belaich, M. Loutfi, E. Faure, C. Bagnara, D. Baty, and J.-P. Belaich. 1991. Characterization of endoglucanase A from Clostridium cellulolyticum. J. Bacteriol. 173:7956-7962. Gilkes, N. R., B. Henrissat, D. G. Kilburn, R. C. Miller, Jr., and R. A. J. Warren. 1991. Domains in microbial P3-1,4-glycanases: sequence conservation, function, and enzyme families. Microbiol. Rev. 55:303-315. Gilkes, N. R., M. L. Langsford, D. G. Kilburn, R. C. Miller, Jr., and R. A. Warren. 1984. Mode of action and substrate specificities of cellulases from cloned bacterial genes. J. Biol. Chem. 259: 10455-10459. Grepinet, O., M.-C. Chebrou, and P. Beguin. 1988. Nucleotide sequence and deletion analysis of the xylanase gene (xynZ) of Clostridium thermocellum. J. Bacteriol. 170:4582-4588. Guiseppi, A., B. Cami, J.-L. Aymeric, G. Ball, and N. Creuzet. 1988. Homology between endoglucanase Z of Erwinia chrysanthemi and endoglucanase of Bacillus subtilis and alkalopholic Bacillus. Mol. Microbiol. 2:159-164. Halsall, D. M., and A. H. Gibson. 1986. Comparison of two Cellulomonas strains and their interaction with Azospirillum brasilense in degradation of wheat straw and associated nitrogen fixation. Appl. Environ. Microbiol. 51:855-861. Hanahan, D. 1983. Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166:557-580. Henrissat, B., M. Claeyssens, P. Tomme, L. Lemesle, and J.-P. Mornon. 1989. Cellulase families revealed by hydrophobic cluster analysis. Gene 81:83-95. Huang, J., and M. A. Schell. 1992. Role of the two-component leader sequence and mature amino acid sequences in extracellular export of endoglucanase EGL from Pseudomonas solanacearum. J. Bacteriol. 174:1314-1323. Huang, J., M. Sukordhaman, and M. A. Schell. 1989. Excretion of the egl gene product of Pseudomonas solanacearum. J. Bacteriol. 171:3767-3774. Hurek, T., B. Reinhold-Hurek, M. van Montagu, and E. Kellenberger. Submitted for publication. Hurek, T., B. Reinhold-Hurek, M. van Montagu, and E. Kellenberger. 1991. Infection of intact roots of Kallar grass and rice seedlings by "Azoarcus." Dev. Plant Soil Sci. 48:235-242. Kim, B. H., and J. W. T. Wimpenny. 1981. Growth and cellulolytic activity of Cellulomonas flavigena. Can. J. Microbiol. 27:12601266. 30. Kushner, S. R. 1978. An improved method for transformation of Escherichia coli with ColEI derived plasmids, p. 17-23. In H. W. Boyer and S. Nicosia (ed.), Genetic engineering. Elsevier/North Holland, Amsterdam. 31. Laemmli, U. K. 1970. Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature (London) 227:680-685. 32. Lamed, R., and E. A. Bayer. 1988. The cellulosome concept: exocellular/extracellular enzyme reactor centers for efficient binding and cellulolysis, p. 1t)1-1 16. In J. P. Aubert, P. Beguin, and J. Millet (ed.), Biochemistry and genetics of cellulose degradation. Academic Press, London. 33. Langsford, M. L., N. R. Gilkes, N. R. Warkarchuk, W. W. Kilburn, D. G. Miller, Jr., and R. A. J. Warren. 1984. The cellulase system of Cellulomonas fimi. J. Gen. Microbiol. 130:1367-1376. 34. Leschine, S. B., K. Holwell, and E. Canale-Parola. 1988. Nitrogen fixation by anaerobic cellulolytic bacteria. Science 242:1157-1159. 35. Levanony, H., Y. Bashan, B. Romano, and E. Klein. 1989. Ultrastructural localization and identification of Azospirillum brasilense Cd on and within wheat root by immuno-gold labeling. Plant Soil 117:207-218. 36. Marais, J. P., J. L. De Wit, and G. V. Quicke. 1966. A critical examination of the Nelson-Somogyi method for the determination of reducing sugars. Anal. Biochem. 15:373-381. 37. Marmur, J. 1961. A procedure for the isolation of deoxyribonucleic acid from microorganisms. J. Mol. Biol. 3:208-218. 38. Mateos, P. F., J. I. Jiminez-Zurdo, J. Chen, A. S. Squartini, S. K. Haack, E. Martinez-Molina, D. H. Hubbell, and F. B. Dazzo. 1992. Cell-associated pectinolytic and cellulolytic enzymes in Rhizobium leguminosarum biovar Trifolii. Appl. Environ. Microbiol. 58:18161822. 39. Matsushita, O., J. B. Russell, and D. A. Wilson. 1990. Cloning and sequencing of Bacteroides ruminicola B,4 endoglucanase gene. J. Bacteriol. 172:3620-3630. 40. McGavin, M., J. Lam, and C. W. Forsberg. 1990. Regulation and distribution of Fibrobacter succinogenes subsp. succinogenes S85 endoglucanases. AppI. Environ. Microbiol. 56:1235-1244. 41. Meinke, A., N. R. Gilkes, D. G. Kilburn, R. C. Miller, Jr., and R. A. J. Warren. 1991. Multiple domains in endoglucanase B (CenB) from Cellulomonas fimi: functions and relatedness to domains in other polypeptides. J. Bacteriol. 173:7126-7135. 42. Mishra, S., P. Beguin, and J.-P. Aubert. 1991. Transcription of Clostridium thermocellum endoglucanase genes celF and celD. J. Bacteriol. 173:80-85. 43. Morales, V., E. Martinez-Molina, and D. Hubbell. 1984. Cellulase production by Rhizobium. Plant Soil 80:407-415. 44. Reinhold, B., T. Hurek, and I. Fendrik. 1987. Cross-reaction of predominant nitrogen-fixing bacteria with enveloped, round bodies in the root interior of Kallar grass. Appl. Environ. Microbiol. 53:889-891. 45. Reinhold, B., T. Hurek, I. Fendrik, B. Pot, M. Gillis, K. Kersters, S. Thielemans, and J. De Ley. 1987. Azospirillum halopraeferens sp. nov., a nitrogen-fixing organism associated with the roots of Kallar grass (Leptochloa fusca (L.) Kunth). Int. J. Syst. Bacteriol. 37:4351. 46. Reinhold, B., T. Hurek, E.-G. Niemann, and I. Fendrik. 1986. Close association of Azospirillum and diazotrophic rods with different root zones of Kallar grass. Appl. Environ. Microbiol. 52:520-526. 47. Reinhold-Hurek, B., T. Hurek, M. Gillis, B. Hoste, M. Vancanneyt, K. Kersters, and J. De Ley. 1993. Azoarcus gen. nov., nitrogenfixing Proteobacteria associated with roots of Kallar grass (Leptochloa fusca (L.) Kunth), and description of two species, Azoarcus indigens sp. nov. and Azoarcus communis sp. nov. Int. J. Syst. Bacteriol. 43:574-584. 48. Roberts, D. P., T. P. Denny, and M. A. Schell. 1988. Cloning of the egl gene of Pseudomonas solanacearum and analysis of its role in phytopathogenicity. J. Bacteriol. 170:1445-145 1. 49. Safo-Sampah, S., and J. G. Torrey. 1988. Polysaccharide-hydrolyzing enzymes of Franckia (Actinomycetales). Plant Soil 112:8997. VOL. 175, 1993 50. Sandhu, G. R., and K. A. Malik. 1975. Plant succession-a key to the utilization of saline soils. Nucleus 12:35-38. 51. Schank, S. C., R. L. Smith, G. C. Weiser, D. A. Zuberer, J. H. Bouton, K. H. Quesenberry, M. E. Tyler, J. R. Milam, and R. C. Littell. 1979. Fluorescent antibody technique to identify Azospirillim brasilense associated with roots of grasses. Soil Biol. Biochem. 11:287-295. 52. Sharrock, K. R. 1988. Cellulase assay methods: a review. J. Biochem. Biophys. Methods 17:81-106. 53. Teather, R. M., and P. J. Wood. 1982. Use of Congo redpolysaccharide interactions in enumeration and characterization of cellulolytic bacteria from the bovine rumen. Appl. Environ. Microbiol. 43:777-780. 54. Theberge, M., P. Lacaze, F. Shareck, R. Morosoli, and D. Kluepfel. 1992. Purification and characterization of an endoglucanase from Streptomyces lividans 66 and DNA sequence of the gene. Appl. Environ. Microbiol. 58:815-820). CELLULASES OF AZOARCUS SP. 7065 55. Tien, T. M., H. G. Diem, M. H. Gaskins, and D. H. Hubbell. 1981. Polygalacturonic acid transeliminase production by Azospirilllin species. Can. J. Microbiol. 27:426-431. 56. Vercoe, P. E., and K. Gregg. 1992. DNA sequence and transcription of an endoglucanase gene from Prevotella (Bacwroides) ruminicola AR20. Mol. Gen. Genet. 233:284-292. 57. Waterbury, J. B., C. B. Calloway, and R. D. Turner. 1983. A cellulolytic nitrogen-fixing bacterium cultured from the gland of deshayes in shipworms (Bivalvia: Teredinidae). Science 221:14011403. 58. Wood, T. M., C. A. Wilson, and C. S. Stewart. 1982. Preparation of the cellulase from the cellulolytic anaerobic bacterium Ruininococcus albus and its release from the bacterial cell wall. Biochem. J. 205:129-137. 59. Yanisch-Perron, C., J. Vieira, and J. Messing. 1985. Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mpl8 and pUC19 vectors. Gene 33:10)3-119.