Vol. 175, No. 3JOURNAL OF BACTERIOLOGY, Feb. 1993, p. 750-7570021-9193/93/030750-08$02.00/0Copyright © 1993, American Society for Microbiology

Polysaccharide Synthesis in Relation to Nodulation Behaviorof Rhizobium leguminosarum

MICHAEL W. BREEDVELD,lt* HAYO C. J. CANTER CREMERS,2t MICHAEL BATLEY,3MAARTEN A. POSTHUMUS,4 LUDOVICUS P. T. M. ZEVENHUIZEN,1

CAREL A. WIJFFELMAN,2 AND ALEXANDER J. B. ZEHNDER1§Department ofMicrobiology, Agricultural University, 6703 CT Wageningen, The Netherlands1; Institute ofPlant

Molecular Biology, State University of Leiden, 2311 VJ Leiden, The Netherlands2; School of Chemistry,Macquarie University, Sydney, New South Wales 2109, Australia3; and Department o/ Organic

Chemistry, Agricultural University, 6703 HB Wageningen, The Netherlands

Received 25 March 1992/Accepted 20 November 1992

In this study, we characterized four TnS mutants derived from Rhizobium leguminosarum RBL5515 withrespect to synthesis and secretion of cellulose fibrils, extracellular polysaccharides (EPS), capsular polysac-charides, and cyclic P-(1,2)-glucans. One mutant, strain RBL5515 exo-344::TnS, synthesizes residual amountsof EPS, the repeating unit of which lacks the terminal galactose molecule and the substituents attached to it.On basis of the polysaccharide production pattern of strain RBL5515 exo-344::Tn5, the structural features ofthe polysaccharides synthesized, and the results of an analysis of the enzyme activities involved, we hypothesizethat this strain is affected in a galactose transferase involved in the synthesis of EPS only. All four mutantsfailed to nodulate plants belonging to the pea cross-inoculation group; on Vicia sativa they induced root hairdeformation and rare abortive infection threads. All of the mutants appeared to be pleiotropic, since in additionto defects in the synthesis of EPS, lipopolysaccharide, and/or capsular polysaccharides significant increases inthe synthesis and secretion of cyclic P-(1,2)-glucans were observed. We concluded that it is impossible tocorrelate a defect in the synthesis of a particular polysaccharide with nodulation characteristics.

Rhizobium leguminosarum is capable of synthesizing avariety of polysaccharides, including lipopolysaccharides(LPS), cellulose fibrils, neutral capsular polysaccharides(CPS), cyclic 13-(1,2)-glucans, and high-molecular-weight an-ionic extracellular polysaccharides (EPS) (5, 16, 36, 41).These polysaccharides can be found on the outer surfaces ofthe cells or secreted in the medium. The structures of theCPS (40) and EPS (31) of many R leguminosanrm strainswere found to be similar (Fig. 1).

Determining the role of Rhizobium meliloti (26-28) and R.leguminosarum (3, 4, 15) polysaccharides in the nodulationprocess has been the goal of many studies. Mutants of R.meliloti unable to synthesize succinoglycan, the typical EPSof this species, formed ineffective pseudonodules on one ofthe R. meliloti host plants, alfalfa (Medicago sativa) (26-28).Furthermore, nodulation of mutants of R. meliloti producinga second EPS (EPS b) was partially restored in that themutants formed effective nodules on alfalfa but failed to formeffective nodules on all other host plants (22). EPS mutantsof R. leguminosarum bv. phaseoli were able to nodulatebeans (4, 17), but R. leguminosarum bv. trifolii and R.leguminosarum bv. viciae mutants with comparable muta-tions formed defective nodules on clover and peas, respec-tively (4, 17). More recently, Canter Cremers et al. describedexoB mutants of R. leguminosarum bv. viciae RBL5523 that

* Corresponding author.t Present address: Department of Food Science, 101 Borland

Laboratory, Pennsylvania State University, University Park, PA16802.

t Present address: AHA! Consultants, 2912 BK Nieuwerkerk a/dIJssel, The Netherlands.

§ Present address: Eidgenossische Anstalt fiir Wasserversorgung,Abwasserreinigung und Gewasserschutz/Eidgenossische Technis-che Hochschulen, CH-8600 Dubendorf, Switzerland.

were defective in EPS, CPS, and LPS synthesis and wereunable to nodulate peas, their host plant (10). It was shownthat the mutants were affected in the enzyme glucose 4'-epimerase, which is necessary for incorporation of galactoseinto the carbohydrate polymers. Unfortunately, most studiesperformed to determine the role of EPS in nodulation havefailed to demonstrate that the mutants used were affected inEPS production alone. Therefore, the symbiotic phenotypemay be due to changes in the synthesis, secretion, orstructure of one of the other polysaccharides.

Previously, we described (9, 13) the isolation of transpo-son mutants of R. leguminosarum bv. viciae that are affectedin the synthesis and/or secretion of (i) EPS (mutantsRBL5515 exo-4::TnS and RBL5515 exo-344::TnS) (ii) LPS(mutant RBL5515 lps-336::Tn5), or (iii) EPS, CPS, and LPS(mutant RBL5515 exoB8: :TnS). In this study, we determinedthe quantities of cellulose fibrils, (1,2)-3-glucans, CPS, andEPS synthesized and compared these amounts with theamounts synthesized by wild-type strains.

MATERIALS AND METHODS

Organisms and cultivation. A list of the organisms used isshown in Table 1. All of the strains were obtained from theDepartment of Plant Molecular Biology, State University,Leiden, The Netherlands. Precultures were prepared byinoculating the organisms into standard medium containing5 g of mannitol per liter, 1 g of glutamic acid per liter, andvitamins and mineral salts (5). Most batch culture experi-ments were done in a production medium, which contained10 g of mannitol per liter. Because of the higher ratio ofcarbon to nitrogen in this medium, higher concentrations ofpolysaccharides were obtained. Cells from the late logarith-mic phase were inoculated (1% vol/vol) into 50-ml portions

750

on August 24, 2015 by guest

http://jb.asm.org/

Dow

nloaded from

R. LEGUMINOSARUM POLYSACCHARIDE SYNTHESIS 751

A a b c d

backbone [-4-GlcA3 4 l-GcAL 4-O-Gc3 4-Glc]a-

chain 6t

sidechain

B

C

h f f3l

Oal3 3-GXc3-G 4-Glc-l 4-GLc e

4116 4116Pyr Pyr

Gat

21

[-4-Glc a. 3-Man aI 3-3Gaat]61n

Gat

Gat

2-Glc Xf2-OGc 2-GtcDP 17-25

FIG. 1. Proposed structures of EPS (A), CPS (B), and cyclic3-(1,2)-glucans (C) of the R. leguminosarum strains used in this

study. (A) Structure of EPS of strains RBL5515, RBL5523,RBL5515 exo-4::TnS, and RBL5515 lps-336::TnS. The EPS ofmutants RBL5515 exoB8::Tn5 and RBL551S exo-344::Tn5 lackterminal galactose h and the hydroxybutanoate and carboxyethyli-dene substituents attached to it. Nonstochiometric substituents are

not shown. DP, degree of polymerization.

of medium in 300-ml Erlenmeyer flasks. The flasks were

incubated on a rotary shaker at 200 rpm and 25°C.Measurement of growth and polysaccharide production.

Cells and culture supernatants were separated by centrifu-gation at 35,000 x g for 30 min. Each resulting cell pellet wasresuspended fivefold concentrated in distilled water andbriefly sonicated to obtain a homogeneous suspension.

TABLE 1. Description of the organisms used in this study

Strain Name

RBL5515..... R. leguminosarum bv. trifolii LPR5 cured of

its Sym plasmidaRBL5523..... Strain RBL5515 harboring R. leguminosarum

bv. viciae pRLlJI::Tnl831 Sym plasmid, str

jitRBL5851..... R. leguminosarum RBL5515 exo-4::TnS cured

of its Sym plasmidcRBL5850..... R. leguminosarum RBL5515 exoB8::Tn5 cured

of its Sym plasmidcRBL5824..... R. leguminosarum RBL5515 lps-336::Tn5

cured of its Sym plasmidcRBL5852..... R. leguminosarum RBL5515 exo-344::TnS

cured of its Sym plasmidc

a See references 24 and 34.b str, streptomycin resistance gene; nif, rifampin resistance gene. See

reference 34.c See reference 13.

(i) Supernatant. The concentrations of excreted glucan andEPS in the supernatant were determined by the anthrone-sulfuric acid method for hexoses (39) and the 3-hydroxy-diphenyl-sulfuric acid method for hexuronic acids (2). Thedetermination of each EPS concentration was based on theuronic acid content of the whole supernatant and the ratio ofhexose to hexuronic acid for a particular EPS (see below);EPS concentrations were expressed as sums of the glucoseand glucuronic acid equivalents (in milligrams per liter ofmedium). The ,-(1,2)-glucan content was derived from theamount of glucose which was not accounted for in thecalculation of the EPS concentration. The mannitol concen-tration was determined spectrophotometrically as describedby Burton (8). The relative viscosity of each culture super-natant was measured with an Ubbelohde viscometer (Tam-son, Zoetermeer, The Netherlands) at 25°C.

(ii) Pellet. The amount of total cellular protein was deter-mined as described by Lowry et al. (30), using bovine serumalbumin as the standard. The CPS was extracted from eachpellet with 1 N NaOH, precipitated with 1 volume of alcohol,and quantified by the anthrone-sulfuric acid method, asdescribed previously (5). The carbohydrate fraction of the50% alcohol supernatant, which contained cyclic 3-(1,2)-glucans(-), was also quantified by the anthrone-sulfuric acidmethod. The alkali-extracted cells were treated with 1 N HClfor 2 h, and the hexose content in the residual material wasused as a measure of the cellulose content of the cells (42).

Isolation and characterization of oligo- and polysaccharides.High-molecular-weight EPS was precipitated from the su-pernatant by adding 3 volumes of ethanol, as describedpreviously (4), and the ratio of hexose to hexuronic acid wasdetermined. The low-molecular-weight cyclic glucans and/orrepeating units remaining in the 75% alcohol supernatantwere separated by using DEAE-Sepharose anion-exchangecolumn chromatography (25 by 2.5 cm) as described previ-ously (6). The hexose and hexuronic acid content of eachfraction was determined. CPS was isolated by alkalineextraction followed by ethanol precipitation, as describedabove. The precipitate was washed with 50% ethanol severaltimes until the pH was 7.0 (40). The neutral sugar composi-tions of EPS and CPS were determined by hydrolysis with2 M trifluoroacetic acid for 8 h at 100°C, followed byconversion of the liberated sugars to alditolacetates andseparation by gas-liquid chromatography (1). Glycosidiclinkages were determined by a methylation analysis (23) andwere identified by mass spectrometry as described previ-ously (12). Pyruvate concentrations were determined spec-trophotometrically as described by Katsuki et al. (25), andhigh-performance liquid chromatography of cyclic glucanswas carried out as described by Breedveld et al. (5).

Depolymerization of EPS. Purified EPS was depolymerizedby using purified phage RL38, fractionated with ethanol, andthen purified further by using columns of Sephadex-DEAEA-25, Biogel P2, and Dowex 5OW-X2 as described previ-ously (10).NMR. For nuclear magnetic resonance (NMR), purified

depolymerized EPS was dissolved in D20, and 1 C spectrawere recorded on a Varian model XL-200 spectrometer byusing methanol (b = 50.04 ppm) as the reference (12).

Analysis of LPS. Isolation of LPS and visualization of LPSprofiles were performed as described previously (10).Enzyme assays. Cell extracts were prepared in 50 mM

N-2-hydroxyethylpiperazine-N'-ethanesulfonic acid (HEPES)(pH 7.6) containing 1 mM dithioerythritol (10). UDP-glucosepyrophosphorylase activity was measured as follows. To amixture containing 0.1 ml of 1 M glycylglycine (pH 8.9), 0.1

VOL. 175, 1993

on August 24, 2015 by guest

http://jb.asm.org/

Dow

nloaded from

752 BREEDVELD ET AL.

TABLE 2. Polysaccharide production by wild-type strains and mutants of R. leguminosaruma

Supernatant PelletStrain Concn of EPS Concn of cyclic glucans Viscosity Concn of CPS Concn of cellulose Concn of cyclic Concn of protein

(mg/liter)' (mg/liter)c (cP) (mg/liter)' (mg/liter)e glucans (mg/gf (mg/liter)

RBL5515 2,010 60 15 990 45 135 540RBL5523 2,105 55 19 750 55 115 525RBLSS1S exo-4 90 380 1 1,590 45 150 560RBL5515 exoB8 280 790 2 70 25 160 510RBL5515 lps-336 1,920 1,050 11 90 15 155 530RBL5515 e-xo-344 350 400 2 1,190 40 150 490

a Bacteria were grown for 14 days at 25'C in the production medium. Experiments were performed three times; in the three experiments values obtained differedby at most 10%. The results of one representative experiment are shown.

b EPS concentrations are expressed as milligrams of glucose and glucuronic acid equivalents per liter of culture.Extracellular cyclic glucan concentrations are expressed as milligrams of glucose equivalents per liter of culture.

d CPS concentrations are expressed as milligrams of galactose equivalents per liter of culture.eCellulose concentrations are expressed as milligrams of glucose equivalents per liter of culture.f The concentrations of low-molecular-weight cellular carbohydrates [mainly cyclic ,3-(1,2)-glucans] are expressed as milligrams of glucose equivalents per gram

of cellular protein.

ml of 0.6 mM glucose 1-phosphate, 0.003 U of UDP-glucosedehydrogenase, 0.1 ml of 10 mM NAD, 0.1 ml of 10 mMUTP, and 0.1 ml of cell extract enough water was added tobring the volume to 1 ml. The conversion of glucose 1-phos-phate into UDP-glucose was measured spectrophotometri-cally at 340 nm.

UDP-glucose 4'-epimerase, glucose dehydrogenase, glu-cokinase, glucose-6-phosphate dehydrogenase, galactose de-hydrogenase, galactokinase, and galactose-1-phosphate-uridyl transferase activities were measured as describedpreviously (10). Enzyme activities were expressed as nano-moles of substrate used per minute per milligram of protein.

Nodulation tests. To test nodulation abilities on differenthost plants, R leguminosarum bv. viciae Sym plasmidpRLlJI marked with TnJ831 was introduced into the bacte-rial strains (10). Nodulation tests were performed as de-scribed previously (11).

RESULTS

Cellulose. We observed that R. leguminosarum RBL5515,RBL5523, and RBL5515 lps-336::Tn5 grew as dispersedhomogeneous liquid cultures, while the cells of mutantsRBL5515 exoB8::TnS, RBL5515 exo-4::TnS, and RBL5515exo-344::TnS aggregated strongly in liquid cultures. Aggre-gation might be the result of an increased cellulose synthesis.Congo red binds preferentially to cellulose in the cell cap-sule, and this phenomenon can be used to measure thecellulose content of a cell pellet (42). On a solidified mediumcontaining Congo red (yeast extract-mannitol-Congo redagar), the colonies of the mutants that were affected in EPSsynthesis were deep red, while the colonies of the otherstrains were pale pink, which suggests that there were

differences in cellulose content. Therefore, we checked theamounts of cellulose synthesized quantitatively (Table 2).No correlation between the cellulose contents of the strainsand their flocculation behavior in liquid cultures or the colorsof the colonies on Congo red agar could be found. Maybe thelack of large amounts of EPS allowed better uptake of Congored by the EPS-deficient mutants.CPS. Sutherland (38) has proposed that the synthesis of

LPS and the synthesis of other exocellular polysaccharidesare correlated since both types of synthesis require the samelipid intermediate. In addition to LPS, R. leguminosarumsynthesizes at least two different exocellular polysaccha-rides, CPS and EPS. To avoid confusion, the term CPS is

used below for the gelling, neutral, high-molecular-weightpolysaccharide produced by stationary-phase cultures of R.leguminosarum; this polysaccharide is firmly attached to thecells and cannot be extracted in cold water (40). The CPS ofa number of R. leguminosarum strains have the same sugarcomposition (glucose-galactose-mannose ratio, 1:4:1) (Fig.1) (40). Therefore, CPS is an entity distinct from both LPSand EPS, and we determined whether our mutants wereaffected in CPS synthesis (Table 2). Strains RBL5515exo-4::TnS and RBL5515 exo-344::TnS synthesized moreCPS than the strain from which they were derived. Acomponent analysis of the CPS revealed a sugar compositionvery similar to that of the wild-type strain.

Table 2 shows that the mutants affected in the synthesis of0-antigen containing LPS I (strains RBL5515 exoB8::TnSand RBL5515 lps-336::TnS) (Fig. 2) formed small amounts ofCPS. However, successive attempts to isolate this capsularmaterial failed, which suggests that these two mutants do notsynthesize CPS at all. We concluded that in our mutantsdefects in the synthesis of 0-antigen containing LPS arecorrelated with defects in the synthesis of CPS.

Cyclic (1,2)-+-glucans. It has been hypothesized thatperiplasmic (1,2)-p-glucans of Rhizobium spp. play a role inosmoregulation (19, 32). Furthermore, it has been found thatcyclic glucans are not directly essential for nodulation, sincesuppressor mutations that restored normal nodulation didnot restore glucan synthesis (20). We analyzed the quantitiesof (1,2)-13-glucans synthesized and secreted by the variousisolates.The pellets of all of the strains contained about the same

quantity of (1,2)-,B-glucans (Table 2). The structures of the(1,2)-p-glucans isolated were studied by performing a meth-ylation analysis. The glucan fractions of both wild-type andmutant strains contained 3,4,6-tri-O-methyl-D-glucose as theonly methylated sugar, confirming the presence of a 1,2-glycosidic bond, as well as the unbranched, cyclic nature ofthe polysaccharide (Fig. 1). The cyclic glucans were neutral,as determined by thin-layer chromatography, 13C NMRspectroscopy, and high-performance liquid chromatography(data not shown), and did not contain any nonsugar substit-uents. On the basis of the elution pattern from the high-performance liquid chromatography preparation, the de-grees of polymerization ranged from 17 to 25, as has beenshown previously for the cyclic glucans of another R.leguminosarum strain (5). In conclusion, the structures and

J. BACTERIOL.

on August 24, 2015 by guest

http://jb.asm.org/

Dow

nloaded from

R. LEGUMINOSARUM POLYSACCHARIDE SYNTHESIS 753

IpsI

Ipsi_

a b cFIG. 2. LPS profiles of R. leguminosarum bv. viciae strains.

Lane a, strain RBL5515 (the LPS profile of strain RBL5523 isidentical); lane b, strain RBL5515 e-xo-344::Tn5 (the LPS profile ofstrain RBL5515 e-xo-4::Tn5 is identical); lane c, strain RBL5515e-xoB8::Tn5 (the LPS profile of strain RBL5515 lps-336::Tn5 isindistinguishable). The spot labeled lpsII represents the lipid A-con-taining core of the LPS, while the spot labeled lpsI represents thelong-chain LPS consisting of 0 antigen attached to the core. Thedifferent bands indicate different degrees of polymerization.

ring size distributions of the cellular (1,2)-,-glucan fractionsisolated from all of the strains were identical.

Large differences were found in the quantities of (1,2)-1-glucans secreted into the medium (Table 2). Significantlymore (1,2)-13-glucans were isolated from the supernatants ofcultures of mutant strains RBL5515 exo-4::Tn5 andRBL5515 exo-344::Tn5 than from the supernatants of a

culture of strain RBL5515 (Table 2). However, the largestamounts were isolated from the culture supernatants ofmutants RBL5515 exoB8::Tn5 and RBL5515 lps-336::Tn5.The structures of the secreted (1,2)-,3-glucans were indistin-guishable from the structures of the (1,2)-,-glucans isolatedfrom the cell pellets. In conclusion, all four mutants showgreatly increased secretion of cyclic 0-(1,2)-glucan.

Synthesis of EPS. Leigh et al. (27) hypothesized that therepeating unit fractions of the EPS synthesized by R. melilotiare important for nodulation. Therefore, we determined theamounts of high- and low-molecular-weight EPS present inthe culture supernatants of the strains used in this study.

Strain RBL5515 and its derivatives all produced about 2 g

of EPS per liter of culture (Table 2). Among the mutants,strain RBL5515 lps-336::Tn5 synthesized amounts of EPSsimilar to the amounts synthesized by the wild-type strain,while the other mutants produced significantly less EPS(Table 2). Therefore, defects in the synthesis of LPS or CPSare not necessarily correlated with defects in the productionof EPS.When culture supernatants of strain RBL5515 and all of its

derivatives and mutants were treated with 3 volumes ofethanol to precipitate high-molecular-weight EPS, 95% ofthe hexuronic acid-containing material was precipitated. Theremaining 5% of the hexuronic acid material was present inthe low-molecular-weight fraction and could be precipitatedwith 10 volumes of ethanol. This fraction was separated into

a neutral component and an acidic component by eluting itfrom a Sephadex-DEAE A-25 anion-exchange column. Af-ter further purification, the neutral components of strainRBL5515 and mutant RBL5515 exo-4::Tn5 were analyzed by13C NMR spectroscopy, and they appeared to consist ofcyclic (1,2)-3-glucan. In the acidic component, repeatingunits of EPS were found; the structural analysis of theseunits is described below.

Structural analysis of EPS. The structures of the EPSsynthesized by strains RBL5515 e-xo-4::Tn5, RBL5515lps-336::TnS, and RBL5515 exo-344::Tn5 were determinedby performing (i) a component analysis, (ii) a methylationanalysis, and (iii) 1H and 13C NMR spectroscopy afterdepolymerization. The structural features of the EPS syn-thesized by strains RBL5515, RBL5523, and RBL5515exoB8::TnS have been described previously (10).Our results showed that the structures of the EPS synthe-

sized by strains RBL5515 exo-4::TnS and RBL5515lps-336::TnS were indistinguishable from the structure of theEPS synthesized by RBL5515 (Fig. 1). A component analy-sis showed that the EPS synthesized by mutant RBL5515exo-344::Tn5 lacked galactose. The spectrum obtained by13C NMR spectroscopy confirmed that strain RBL5515exo-344: :Tn5 synthesized repeating units comparable tothose produced by strain RBL5515 exoB8::Tn5 (10) in thatthe repeating units lacked terminal galactose residue h (Fig.1) and the substituents attached to it.One obvious difference between the spectra of strains

RBL5515 exo-344::TnS (Fig. 3B), RBL5515 (Fig. 3A), andRBL5515 exoB8::TnS exists. In the spectrum of strainRBL5515 exo-344::Tn5 the resonances indicating a smalllevel of O-acetylation of residue d were absent.The repeating units of the EPS isolated from the superna-

tants of the various strains were indistinguishable from thoseobtained by enzymic depolymerization of the long-chainEPS.

In summary, mutant RBL5515 exo-344::Tn5 synthesizessmall amounts of high-molecular-weight EPS which lacksgalactose and the substituents attached to it, as do the exoBmutants (10). Compared with the EPS synthesized by paren-tal strain RBL5515 and the e-xoB mutants, the EPS syn-thesized by strain RBL5515 exo-344: :Tn5 has a slightlydifferent acetylation pattern. Unlike the e-xoB mutants, strainRBL5515 exo-344::TnS synthesizes normal LPS and CPS.

Activities of enzymes involved in galactose metabolism.Previously, we reported that R. leguminosarum bv. viciaee-xoB mutants are deficient in the synthesis of UDP-glucose4'-epierase (10). This enzyme is involved in the synthesis ofUDP-galactose, which is generally used by bacteria to incor-porate galactose in polysaccharides. Like the EPS of e-xoBmutants, the EPS of strain RBL5515 exo-344: :Tn5 lacksgalactose. Therefore, we analyzed strain RBL5515 exo-344::Tn5 and the other mutants not tested previously for thepresence of UDP-glucose 4'-epimerase and other enzymesinvolved in the synthesis of UDP-galactose.

In cell extracts of R. leguminosarum bv. viciae RBL5515e-xo-4::Tn5 pRLlJI, RBL5515 e-xo-344::TnS pRLlJI, andRBL5515 Ips-336::TnS the levels of activity of the followingenzymes were comparable to those of parental strainsRBL5515 and RBL5515 RBL5523: UDP-glucose 4'-epime-rase (90 nmol/min/mg of protein), UDP-glucose pyrophos-phorylase (120 nmol/min/mg of protein), glucose dehydroge-nase (80 nmol/min/mg of protein), glucokinase (90 nmol/min/mg of protein), glucose-6-phosphate dehydrogenase (90nmol/min/mg of protein), galactose dehydrogenase (450nmol/min/g of protein), galactokinase (no activity), and

VOL. 175, 1993

on August 24, 2015 by guest

http://jb.asm.org/

Dow

nloaded from

754 BREEDVELD ET AL.

1

1'3

7

6!1

55' 5"1/

1'

110

45

100 90 80

A

8 9

1"

B

8

70 60

11

ILI

30 20

Chemical Shift (ppm)

FIG. 3. '3C NMR spectra of repeating units isolated from the depolymerized EPS of R. leguminosarum bv. viciae RBL5515 pRLlJI (A)and RBL5515 exo-344::TnS pRLlJI (B). Peak 1, C-1 of glucuronic acid residue b; peak 1', C-1 of 2-O-acetyl glucuronic acid residue b; peak2, C-1 of the carboxyethylidene residue of 4,6-O-(l-carboxyethylidene)glucose residue g; peak 3, C-1 of the carboxyethylidene residue of4,6-O-(l-carboxyethylidene)galactose residue h; peak 4, C-1 of p-glucose residue d; peak 4', C-1 of 3-O-acetyl-13-glucose residue d; peak 4',C-1 of 2-O-acetyl-13-glucose residue d; peak 5, C-1 of a-glucose residue d; peak 5', C-1 of 3-O-acetyl-a-glucose residue d; peak 5", C-1 of2-O-acetyl-a-glucose residue d; peak 6, C-4 of glucuronic acid residue b; peak 6', C-4 of 3-O-acetyl glucuronic acid residue b; peak 7, C-3 of4,6-O-(l-carboxyethylidene)glucose residue g; peak 8, methyl of 4,6-O-(1-carboxyethylidene)glucose residue g; peak 9, methyl of 4,6-0-(1-carboxyethylidene)galactose residue h; peak 10, hydroxybutanoate methyl group; peak 11, acetate methyl group. In panel B resonance 7shifted about 7 ppm upfield because of the absence of galactose residue h and was hidden in the large peak at 74 ppm.

galactose-1-phosphate-uridyl transferase (100 nmol/min/mgof protein).

In a previous paper (10) we confirmed that Rhizobium sp.strain RBL5515 uses the De Ley-Douderoff pathway for theconversion of galactose. Therefore, UDP-galactose can beformed only by the epimerization of UDP-glucose catalyzedby UDP-glucose 4'-epimerase. On the basis of the enzymeactivities which we found and the fact that mutant RBL5515exo-344::TnS can still synthesize EPS, LPS, and CPS, weconcluded that this mutant is able to synthesize UDP-galactose.

Nodulation. Rhizobium mutants affected in the synthesisof EPS are also often affected in nodulation (18, 28), and wetherefore analyzed the nodulation abilities of the mutants.For this purpose, we introduced R. leguminosarum bv.viciae Sym plasmid pRLlJI into mutants RBL5515 exo-4::Tn5, RBL5515 exo-344::TnS, and RBL5515 lps-336::TnS. Onboth peas and Vicia sativa, all three strains failed to inducethe formation of nodules, while on the roots of Vicia hirsuta

several non-nitrogen-fixing, white bumps were present after3 weeks. In comparison, wild-type strain RBL5515 pRLlJIinduced the formation of nitrogen-fixing nodules on all plantswithin 6 to 12 days.On V. sativa, all three mutants induced only root hair

deformation and rare, abortive infection threads. Therefore,the defects in the synthesis of polysaccharides by thesemutants affect nodulation of plants belonging to the peacross-inoculation group at the level of infection threadformation.

DISCUSSION

Cyclic (1,2)-1-glucans. In comparison with the wild-typestrain, all of the mutants described here synthesized morethan 10 time more cyclic glucans. Cyclic (1,2)-3-glucans aregenerally concentrated in the periplasmic space of rhizobia,where they play a role in osmotic adaptation (19, 32).EPS, CPS, and LPS form a shield around the outer

J. BACTERIOL.

on August 24, 2015 by guest

http://jb.asm.org/

Dow

nloaded from

R. LEGUMINOSARUM POLYSACCHARIDE SYNTHESIS 755

membrane of a rhizobial cell, and this shield protects theperiplasmic space from direct contact with the environment.Absence of part of this shield because of a defect in thesynthesis of any of these polysaccharides may facilitateexchange between the periplasmic space and the environ-ment. Therefore, it is possible that the high levels of (1,2)-P-glucans present in the culture supernatants of the exo andIps mutants are the result of increased leakage from theperiplasmic space (7). Since the cell pellets of mutant andwild-type strains contain similar amounts of glucans, themutants react to the enhanced leakage by increasing thesynthesis of (1,2)-p-glucans. This could be a reaction of theorganism in order to osmotically adjust to the environment.On basis of the results of our experiments, we can only pointto a strong correlation between enhanced secretion of cyclicglucans and the inability of mutants to synthesize EPS, LPS,or CPS.

Proposed role for the affected gene(s) of mutant RBL5515exo-4::Tn5. Mutant RBL5515 exo-4::Tn5 synthesizes all ofthe polysaccharides, but only residual amounts of EPS. It istherefore difficult to determine the nature of the affectedgene(s). Since the residual amounts of EPS synthesized stillhave the wild-type structure, the mutant can still synthesizeall of the sugar nucleotides and sugar transferases requiredfor EPS assembly (37). It could be that the mutant is affectedin export or regulation of EPS synthesis.

Proposed role for the affected gene(s) of mutant RBL5515lps-336::TnS. Strain RBL5515 lps-336::Tn5 is a pleiotropicmutant that is affected in the synthesis of both LPS and CPS.The LPS of R. leguminosarum bv. viciae contains glucose,galactose, glucuronic acid, and mannose (14). In addition,Zevenhuizen (40) found that the CPS of R. leguminosarumconsists of galactose, glucose, and mannose. Since mutantRBL5515 lps-336::Tn5 synthesizes EPS at wild-type levels,it is capable of synthesizing UDP-galactose, UDP-glucose,and UDP-glucuronic acid (37). It is possible that this mutantis affected in the synthesis of the other common precursorCPS and LPS, UDP-mannose.

Several R. leguminosarum bv. viciae mutants affected inthe synthesis of LPS but not in the synthesis of EPSdescribed previously by De Maagd et al. (16) and Priefer(33). In contrast to strain RBL5515 lps-336::Tn5, thesemutants induced non-nitrogen-fixing nodules on plants of thepea cross-inoculation group. Unfortunately, these authorsdid not study the synthesis of other polysaccharides, andtherefore, a comparison of the various mutants at this timeseems pointless.

Proposed role for the affected gene(s) of mutant RBL5515exo-344::TnS. The repeating unit of the EPS synthesized inresidual amounts by mutant RBL5515 exo-344::TnS lacks theterminal galactose residue of the side chain and, as aconsequence, the hydroxybutanoyl and carboxyethylidenesubstitutions attached to it. Previously, we described amutant, strain RBL5515 e-xoB8: :Tn5, which synthesizes EPSthat is similar in structure to the EPS synthesized by mutantRBL5515 exo-344: :TnS (10). However, strain RBL5515exo-344::Tn5 (i) synthesizes other polysaccharides whichrequire the presence of UDP-galactose and (ii) harbors theenzymes required for the synthesis of UDP-galactose.Therefore, the defect in the structure of its EPS is not due toan inability to synthesize UDP-galactose. In addition, themutant could not be complemented by plasmids harboringthe e-xoB gene (data not shown).These findings suggest an alternative explanation. When

UDP-galactose is present, the addition of galactose duringsynthesis of EPS requires only the presence of a specific

sugar transferase (37, 38) and thus CPS or LPS is notsynthesis is not affected. It is likely that this transferase isabsent in mutant RBL5515 exo-344::TnS. Therefore, wepropose that mutant RBL5515 exo-344::Tn5 is affected eitherin a structural gene coding for the transferase or in a genewhich regulates the synthesis of this specific galactosetransferase.The relationship between the proposed defect and the

amount of EPS actually synthesized is not yet clear. StrainRBL5515 exo-344::Tn5 produced considerably lower levelsof EPS than the wild-type strain. It could be that theenergetics of synthesis of the truncated EPS synthesized bythe mutant is less favorable than the energetics of synthesisof the fully acetylated and pyruvylated EPS synthesized bythe wild-type strain. An extensive discussion of these kindsof considerations has been published previously for Xanth-omonas campestris mutants that synthesize a truncated EPSwith a reduced capacity (29).

Nodulation ability. The relationship between exopolysac-charide production by Rhizobium spp. and the ability tonodulate host plants has been the subject of many studies,most of which were done with transposon mutants affectedin the synthesis of EPS (17, 26, 28, 35). Unfortunately, moststudies performed with R. leguminosarum strains are ham-pered by the fact that the mutants are defined for only one ortwo polysaccharides. Previously, we have shown that de-fects in the synthesis of EPS can be the result of mutationsin genes coding for basal metabolic enzymes (10). Similarconsequences of mutations in metabolic genes have alsobeen described for other bacterial species, including Salmo-nella spp. (21). Clearly, multiple defects (detectable andnondetectable) can result from a mutation in one of thesegenes. Therefore, firm conclusions concerning the involve-ment of polysaccharides in nodulation which are based onexperiments performed with this kind of mutant should berejected.

Furthermore, authors sometimes use nonisogenic back-grounds to study nodulation. For instance, several mutantsof R. leguminosarum bv. phaseoli CFN42 which fail tosynthesize LPS and EPS were described by Diebold andNoel (17). To study the effect of the mutation on nodulationof plants of the pea cross-inoculation group, the authorstransferred the mutation by transduction into R. leguminosa-rum bv. viciae 128C569. In this background the mutation oftwo strains resulted only in a defect in the synthesis of EPS.This strongly suggests that mutations in homologous genesin different chromosomal backgrounds can have differenteffects. Therefore, we feel that studies to determine relation-ships between nodulation on plants of different cross-inocu-lation groups and defects in polysaccharide synthesis shouldbe performed only when isogenic backgrounds are available.Of the mutants described in this paper, strains RBL5515

e-xoB8::TnS and RBL5515 lps-336::Tn5 are clearly pleiotro-pic. Unfortunately, the other two mutants (strains RBL5515exo-4::Tn5 and RBL5515 exo-344::Tn5) are in fact alsopleiotropic, if only because they both exhibit decreasedsynthesis of EPS and enhanced (1,2)-p-glucan synthesis. Atthis moment, we cannot exclude the possibility that thenodulation defects of all four mutants are caused by en-hanced (1,2)-pi-glucan synthesis and secretion rather than bya defect in the synthesis of a particular polysaccharide.Further research will have to be performed to clarify thesecorrelations.

VOL. 175, 1993

on August 24, 2015 by guest

http://jb.asm.org/

Dow

nloaded from

756 BREEDVELD ET AL.

ACKNOWLEDGMENTS

This investigation was carried out with the support of the DutchNational Innovation Oriented Program on Carbohydrates.H.C.J.C.C. was supported by a grant from the Netherlands Scien-tific Organization.

REFERENCES

1. Blakeney, A. B., P. J. Harris, R. J. Henry, and B. A. Stone. 1983.A simple and rapid preparation of alditolacetates for monosac-charide analysis. Carbohydr. Res. 113:291-299.

2. Blumenkrantz, M., and 0. Asboe-Hansen. 1973. New method forthe quantitative determination of uronic acids. Anal. Biochem.54:484-489.

3. Borthakur, D., C. E. Barber, J. W. Lamb, M. Daniels, J. A.Downie, and A. W. B. Johnston. 1986. A mutation that blocksexopolysaccharide synthesis prevents nodulation of peas byRhizobium leguminosarum but not of beans byR phaseoli andis corrected by cloned DNA from Rhizobium or the phytopatho-gen Xanthomonas. Mol. Gen. Genet. 203:320-323.

4. Borthakur, D., R. F. Barber, J. W. Latchford, L. Rossen, andA. W. B. Johnston. 1988. Analysis of pss genes of Rhizobiumleguminosarum required for exopolysaccharide synthesis andnodulation of peas: their primary structure and their interactionwithpsi and other nodulation genes. Mol. Gen. Genet. 213:155-162.

5. Breedveld, M. W., L. P. T. M. Zevenhuizen, and A. J. B.Zehnder. 1990. Excessive excretion of cyclic 3-(1,2)-glucan byRhizobium tnifolii TA-1. Appl. Environ. Microbiol. 56:2080-2086.

6. Breedveld, M. W., L. P. T. M. Zevenhuizen, and A. J. B.Zehnder. 1990. Osmotically-induced oligo- and polysaccharidesynthesis by Rhizobium meliloti SU-47. J. Gen. Microbiol.136:2511-2519.

7. Breedveld, M. W., L. P. T. M. Zevenhuizen, and A. J. B.Zehnder. 1992. Synthesis of ,B-(1,2)-glucans by Rhizobium legu-minosarum bv. trifolii TA-1: factors influencing excretion. J.Bacteriol. 174:6336-6342.

8. Burton, R. M. 1957. The determination of glycerol and dihy-droxyacetone. Methods Enzymol. 3:246-249.

9. Canter Cremers, H. C. J. 1990. Role of exopolysaccharide innodulation of Rhizobium leguminosarum bv. viciae. Ph.D.thesis. State University of Leiden, Leiden, The Netherlands.

10. Canter Cremers, H. C. J., M. Batley, J. W. Redmond, L.Eydems, M. W. Breedveld, L. P. T. M. Zevenhuizen, E. Pees,C. A. Wiifelman, and B. J. J. Lugtenberg. 1990. Rhizobiumleguminosarum exoB mutants are deficient in the synthesis ofUDP-glucose 4'-epimerase. J. Biol. Chem. 265:21122-21127.

11. Canter Cremers, H. C. J., H. P. Spaink, A. H. M. WUfjes, E.Pees, C. A. WUffelman, R. J. H. Okker, and B. J. J. Lugtenberg.1989. Additional nodulation genes on the Sym plasmid ofRhizobium leguminosarum biovar viciae. Plant Mol. Biol. 13:163-174.

12. Canter Cremers, H. C. J., K. Stevens, B. J. J. Lugtenberg, C. A.Wi3ffelman, M. Batley, J. W. Redmond, M. W. Breedveld, andL. P. T. M. Zevenhuizen. 1991. Unusual structure of theexopolysaccharide of Rhizobium leguminosarum biovar viciaestrain 248. Carbohydr. Res. 218:185-200.

13. Canter Cremers, H. C. J., C. A. Woffelman, E. Pees, M. Engels,F. Hoogerbruggen, K. Stevens, M. van DUk, and B. J. J.Lugtenberg. 1988. Chromosomal nodulation genes of R. legumi-nosarum biovar viciae, p. 484. In H. Bothe, F. J. de Bruijn, andW. J. Newton (ed.), Nitrogen fixation: hundred years after.Gustav Fisher, Stuttgart, Germany.

14. Carlson, R. W., S. Kalembosa, D. T. Thurowski, and K. D. Noel.1987. Characterization of the lipopolysaccharide from a Rhizo-bium phaseoli mutant that is defective in infection threaddevelopment. J. Bacteriol. 169:4923-4928.

15. Chakravorty, A. K., W. Zurkowski, J. Shine, and B. J. Rolfe.1982. Symbiotic nitrogen fixation: molecular cloning of Rhizo-bium genes involved in exopolysaccharide synthesis and effec-tive nodulation. J. Mol. Appl. Genet. 1:585-595.

16. De Maagd, R. A., C. van Rossum, and B. J. J. Lugtenberg. 1989.

Recognition of individual strains of fast-growing rhizobia byusing profiles of membrane proteins and lipopolysaccharides. J.Bacteriol. 170:3782-3785.

17. Diebold, R., and K. B. Noel. 1989. Rhizobium leguminosarumexopolysaccharide mutants: biochemical and genetic analysesand symbiotic behavior on three hosts. J. Bacteriol. 171:4821-4830.

18. Djordjevic, S. P., H. Chen, M. Batley, J. W. Redmond, and B. G.Rolfe. 1987. Nitrogen fixation ability of exopolysaccharide syn-thesis mutants of Rhizobium sp. NGR234 and Rhizobium tnifoliiis restored by the addition of homologous exopolysaccharides.J. Bacteriol. 169:53-60.

19. Dylan, T., D. R. Helinski, and G. S. Ditta. 1990. Hypoosmoticadaptation in Rhizobium meliloti requires 3-(1,2)-glucan. J.Bacteriol. 172:1400-1408.

20. Dylan, T., P. Nagpal, D. R. Helinski, and G. S. Ditta. 1990.Symbiotic pseudorevertants of Rhizobium meliloti ndv mutants.J. Bacteriol. 172:1409-1417.

21. Germanier, R. 1970. Immunity in experimental salmonellosis. I.Protection induced by rough mutants of Salmonella typhimu-rium. Infect. Immun. 2:309-315.

22. Glazebrook, J., and G. C. Walker. 1989. A novel exopolysac-charide can function in place of the calcofluor-binding ex-opolysaccharide in nodulation of alfalfa by Rhizobium meliloti.Cell 56:661-672.

23. Harris, P. J., R. J. Henry, A. B. Blakeney, and B. A. Stone.1984. An improved procedure for the methylation analysis ofoligosaccharides and polysaccharides. Carbohydr. Res. 127:59-73.

24. Hooikaas, P. J. J., A. A. N. Van Brussel, H. Den Dulk-Ras,G. M. S. Van Slogteren, and R. A. Schilperoort. 1981. Symplasmid of Rhizobium tnifolii expressed in different rhizobialspecies and Agrobacterium tumefaciens. Nature (London) 291:351-353.

25. Katsuki, H., T. Yoshida, C. Tanegashima, and S. Tanaka. 1971.Improved method for determination of the keto acids by 2,4-dinitrophenylhydrazine. Anal. Biochem. 43:349-356.

26. Leigh, J. A., and C. C. Lee. 1988. Characterization of polysac-charides of Rhizobium meliloti exo mutants that form ineffectivenodules. J. Bacteriol. 170:3327-3332.

27. Leigh, J. A., J. W. Reed, J. F. Hanks, A. M. Hirsch, and G. C.Walker. 1987. Rhizobium meliloti mutants that fail to succinilatetheir Calcofluor-binding exopolysaccharide are defective in nod-ule invasion. Cell 51:579-587.

28. Leigh, J. A., E. R. Signer, and G. C. Walker. 1985. Exopolysac-charide-deficient mutants of Rhizobium meliloti that form inef-fective nodules. Proc. Natl. Acad. Sci. USA 82:6231-6235.

29. Linton, J. D. 1990. The relationship between metabolite produc-tion and the growth efficiency of the producing organism. FEMSMicrobiol. Rev. 75:1-18.

30. Lowry, 0. H., A. L. Rosebrough, R. J. Farr, and R J. Randall.1951. Protein measurement with the Folin phenol reagent. J.Biol. Chem. 193:265-275.

31. McNeil, M., J. Darvili, A. Darvill, P. Albersheim, A. Van Veen,P. Hooykaas, and R. H. Schilperoort. 1986. The discernable,structural features of the acidic exopolysaccharides secreted bydifferent Rhizobium species are the same. Carbohydr. Res.146:307-326.

32. Miller, K. J., E. P. Kennedy, and V. N. Reinhold. 1986. Osmoticadaptation by Gram-negative bacteria: possible role forperiplasmic oligosaccharides. Science 231:48-51.

33. Priefer, U. B. 1989. Genes involved in lipopolysaccharide pro-duction and symbiosis are clustered on the chromosome ofRhizobium leguminosarum biovar viciae VF39. J. Bacteriol.171:6161-6168.

34. Priem, W. J. E., and C. A. Wiffelman. 1984. Selection of strainscured of the Rhizobium leguminosarum Sym-plasmid pRLlJIby using small bacteriocin. FEMS Microbiol. Lett. 25:247-251.

35. Sanders, R., E. Raleigh, and E. Signer. 1981. Lack of correlationbetween extracellular polysaccharide and nodulation ability inRhizobium. Nature (London) 292:148-149.

36. Smit, G., J. W. Kijne, and B. J. J. Lugtenberg. 1987. Involve-

J. BACTERIOL.

on August 24, 2015 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 175, 1993 R. LEGUMINOSARUM POLYSACCHARIDE SYNTHESIS 757

ment of both cellulose fibrils and a Ca2l-dependent adhesin inthe attachment of Rhizobium leguminosarum to pea root hairtips. J. Bacteriol. 169:4294-4301.

37. Stoddart, R. W. 1984. The biosynthesis of polysaccharides.Croom Helm Ltd., Beckenham, Kent, England.

38. Sutherland, I. W. 1979. Microbial exopolysaccharides: controlof synthesis and acylation, p. 1-34. In R. C. W. Berkeley, G. W.Gooday, and D. C. Ellwood (ed.), Microbial polysaccharidesand polysaccharidases. Academic Press, London.

39. Trevelyan, W. E., and J. S. Harrison. 1952. Studies on yeastmetabolism. I. Fractionation and microdetermination of cell

carbohydrates. Biochem. J. 50:298-310.40. Zevenhuizen, L. P. T. M. 1984. Gel-forming capsular polysac-

charide of fast-growing bacteria: occurrence and rheologicalproperties. Appl. Microbiol. Biotechnol. 20:393-399.

41. Zevenhuizen, L. P. T. M. 1990. Recent developments in Rhizo-bium polysaccharides, p. 387-402. In E. A. Dawes (ed.), Novelbiodegradable microbial polymers. Kluwer Publishers, Dord-recht, The Netherlands.

42. Zevenhuizen, L. P. T. M., C. Bertocchi, and A. R. W. VanNeerven. 1986. Congo-red absorption and cellulose synthesis byRhizobiaceae. Antonie van Leeuwenhoek 52:381-386.

on August 24, 2015 by guest

http://jb.asm.org/

Dow

nloaded from