ApPLIED MICROBIOLOGY, Nov. 1975, p. 731-737 Vo!. 30, No.5 Copyright © 1975 American Society for Microbiology Printed in U.SA.

Carbohydrate Catabolism of Selected Strains In the Genus Agrobacterium

LARRY O. ARTHUR,' LAWRENCE K. NAKAMURA, GRANT ST. JULIAN, AND LEE A. BULLA, JR.'. Northern Regional Research Laboratory, Agricultural Research Service, Peoria, Illinois 61604 Received for publication 3 July 1975

Radiorespirometric and enzyme analyses were used to reveal the glucose­ catabolizing mechanisms functioning in single strains of seven presumed Agro­ bacterium species. The Entner-Doudoroffand pentose cycle pathways functioned in A. radiobacter, A. tumefaciens, A. rubi, and A. rhizogenes. Whereas both catabolic pathways were utilized to an almost equal degree in the A. radiobacter and A. tumefaciens strains, use of the Entner-Doudoroff pathway predominated in the A. rubi and A. rhizogenes strains. A. stellulatum catabolized glucose almost solely through the Entner-Doudoroff pathway. In A. pseudotsugae and A. gypsophilae, glucose was metabolized mainly through the Emden-Meyerhof­ Parnas pathway; the pentose phosphate pathway was also utilized.

According to Bergey's Manual ofDetermina­ itate establishing and delineating their position tive Bacteriology, 7th ed. (3), the genus Agro­ in any taxonomic scheme. bacterium is heterogenous and contains five Although the DNA, serological properties, phytopathogenic species (A. tumefaciens, A. and protein similarities of the agrobacteria rubi, A. rhizogenes, A. gypsophilae, and A. have been extensively investigated, only lim­ pseudotsugae) and two nonphytopathogenic ited work has been done on determining the ones (A. radiobacter and A. stellulatum). In carbohydrate-catabolizing mechanisms of these addition to pathogenicity, classification is organisms. The work done indicated that A. based on indole production, congo red absorp­ tumefaciens utilizes D-glucose by strictly aerobic tion, and nitrate reduction. Modified taxonomic mechanisms involving pentose cycling (PC) schemes based on deoxyribonucleic acid (DNA) reactions and the Entner-Doudoroff (ED) path­ composition and homology, flagellation, 3-keto­ way 0, 22). In the present study we identified lactose production, phytopathogenicity, and nu­ and quantitated by radiorespirometry and en­ merical analyses of nutritional requirements zyme analyses the glucose-catabolizing path­ (5, 7, 9, 24) group A. radiobacter, A. tumefa­ ways functionings in selected strains of seven ciens, A. rhizogenes, and A. rubi as closely species in the genusAgrobacterium. (This paper related organisms separate and apart from A. was presented in part at the 72nd Annual Meet­ gyp,sophilae, A. pseudotsugae, and A. stellula­ ing of the American Society for Microbiology, tum. In fact, some taxonomists suggest thatA. Philadelphia, Pa., 23-28 April 1972.) tumefaciens, A. rhizogenes, and A. rubi are pathogenic variants of A. radiobacter. How­ MATERIALS AND METHODS ever, results of other taximetric DNA composi­ Organisms and cultural conditions. A. radiobac­ tional analyses and hybridization tests (5, 9) ter NRRL B-164, A. tumefaciens NRRL B-36,A. rubi place in question this proposed relatedness of NRRL B-4017, A. rhizogenes NRRL B-193, A. pseu­ A. rubi and A. rhizogenes to A. radiobacter. dotsugae NRRL B-4016, and A. gypsophilae NRRL The validity of including A. stellulatum, A. B-4015 were obtained from the Agricultural Re­ gypsophilae, and A. pseudotsugae in the genus search Service culture collection maintained at the Agrobacterium can be challenged directly from Northern Regional Research Laboratory. A. stellula­ tum TS101 (NRRL B-4018) was kindly furnished by results of these latter experiments. Bergey's M. P. Starr, Department of Bacteriology, University Manual, 8th ed. (4), now places these three of California, Davis. species as incertae sedis. Obviously, additional All cultures, except A. stellulatum, were main­ characterization ofthe agrobacteria would facil- tained on yeast-malt agar slants (8). A. stellulatum was transferred biweekly on artificial seawater agar

I Present address: Frederick Cancer Research Center, consisting of 5.0 g of peptone, 1.0 g of yeast extract, Frederick, Md. 2170l. 0.1 g of iron citrate, 19.45 g of NaCI, 5.0 g of MgCI" 2 Present address: U. S. Grain Marketing Research Cen­ 3.74 g of Na,SO" 1.8 g of CaCI" 0.22 g of KCI, 0.16 g ter. Agricultural Research Service, Manhattan, Kan. ofNaHC03, 80.0 mg ofKBr, 34.0 mg ofSrCl2, 2.0 mg 66502. of H3B03 , 4.0 mg of Na silicate, 2.4 mg of NaF, 1.6 731 Purchased by U. S. Dept. of Agriculture for Official Use 732 ARTHUR ET AL. ApPL. MlcRoBIOL.

mg of NH4 N03 , 8.0 mg of Na2 HP04 , 20.0 g of agar, excess lactic dehydrogenase. KDPG was the gener­ and 1.0 liter of water. ous gift of W. A. Wood, Department of Biochemis­ For enzyme analyses and radiorespirometric ex­ try, Michigan State University, East Lansing. periments A. radiobacter, A. tumefaciens, A. pseu­ Glucosephosphate isomerase (D-glucose-6-phos­ dotsugae, and A. gypsophilae strains were cultured phate ketol-isomerase, EC 5.3.1.9) was measured in in a chemically defined medium described by Kane­ the reverse direction by monitoring the reduction of shiro (11). The strains A. rubi and A. rhizogenes NADP in the presence of fructose-6-phosphate and grew actively in the chemically defined medium with excess glucose-6-phosphate dehydrogenase. Phospho­ yeast-malt broth (20:1 [vollvol]), and A. stellulatum fructokinase (ATP:D-fructose-6-phosphat€-1-phos• in the chemically defined medium with seawater photransferase, EC 2.7.1.11) was assayed by cou­ broth (lO:l[vollvol]). All cultures were incubated pling with fructose-I, 6-diphosphate aldolase, triose­ at 25 C on a rotary shaker (250 rpm). phosphate isomerase, and a-glycerophosphate dehy­ Preparation of cell extracts. Midexponential drogenase and following NADH oxidation. Fructose­ phase cells were collected by centrifugation at 4 C. 1,6-diphosphate aldolase (Fructose-1,6-diphosphate: After being washed once in an equal volume of dis­ D-glyceraldehyde-3-phosphate-lyase, EC 4.1. 2.13) tilled water, the cells were suspended in 0.02 M activity was determined by following NADH oxida­ phosphate buffer (pH 7.4) and disrupted in a French tion during formation of triose phosphates from pressure cell. Nearly all the cells were ruptured fructose-l,6-diphosphate in the presence of excess after the suspensions were processed twice. The mix­ glycerophosphate dehydrogenase and triosephos­ tures were centrifuged at 12,000 x g for 30 min at phate isomerase. 4 C to obtain crude extracts; in turn, the extracts Fructose-6-phosphate phosphoketolase (EC were separated into soluble and particulate fractions 4.1.2.22) was measured by two different methods. In by centrifugation at 105,000 x g for 120 min at 4 C. the first method, acetyl phosphate formed from fruc­ Protein in the extracts was determined by the tose-6-phosphate was determined as a derivative of method of Lowry et al. (16). hydroxamic acid (20). The reaction mixture con­ Enzyme assays. Enzymes were assayed spectro­ sisted of 0.5 ml of 6 mM fructose-6-phosphate, 0.1 photometrically with a Beckman Acta III double­ ml of 100 mM MgCI 2, 0.1 ml of 25 mM 1,2-ethanedi­ beam, multiple-sample absorbance recorder by thiol, 0.1 ml of 45 mM Na2HP04 , 0.1 ml of 1 mM standard procedures (25) unless otherwise indicated. thiamine pyrophosphate, and 0.8 ml of 45 mM Specific activities are expressed as enzyme units per tris(hydroxymethyDaminomethane - hydrochloride milligram of protein in the cell-free extracts. One buffer (pH 7.2) plus cell-free extract and distilled enzyme unit is defined as the quantity of enzyme water. Final volume was 2 ml. This reaction mix­ that converted 1 /-Lmol of substrate per min. ture was incubated at 30 C, and at appropriate inter­ Phosphoenolpyruvate-glucose phosphotransfer­ vals fractions were removed and combined with ase was assayed by the procedure of Ghosh and equal volumes of hydroxylamine solution (4 M hy­ Ghosh (6). Glucokinase (adenosine triphosphate droxylamine-hydrochloride and 3.5 M NaOH) and [ATP]:D-glucose 6-phosphotransferase, EC 2.7.1.2) 0.1 M acetate buffer (pH 5.4). The hydroxamic acid was measured by observing nicotinamide adenine derivative of acetyl phosphate formed in this reac­ dinucleotide (NAD) reduction in the presence of ex­ tion mixture was assayed spectrophotometrically at cess glucose 6-phosphate dehydrogenase. Glucose 505 nm after addition to the above mixture of equal 6-phosphate dehydrogenase (D-glucose-6-phos­ volumes of ferric chloride (5% FeCl2 6H 20 in 0.1 N phate:nicotinamide adenine dinucleotide phosphate HCD, 3.86 N HCI, and 12% trichloroacetic acid. The [NADP] oxidoreductase, EC. 1.1.1.49) and 6-phos­ second method involved the determination ofglycer­ phogluconate dehydrogenase (6-phospho-D-gluco­ aldehyde-3-phosphate by following at 340 nm the nate:NADP oxidoreductase [decarboxylatingJ, EC oxidation of NADH in the presence of fructose-6­ 1.1.1.44) were determined by assays ofNADP reduc­ phosphate, transaldolase, excess trios€ isomerase, tion. and a-glycerophosphate dehydrogenase. Pyruvate Transketolase (D -sedoheptulose-7-phosphate:D­ kinase (ATP:pyruvate phosphotransferase, EC glyceraldehyde-3-phosphate glycoaldehydetransfer­ 2.7.1.40) was assayed by following the formation of ase, EC 2.2.1.1) activity was determined by follow­ pyruvate from phosphoenolpyruvate via NADH oxi­ ing reduced NAD (NADH) oxidation in a triose isom­ dation in a coupled enzyme system containing ex­ erase-a-glycerophosphate dehydrogenase-coupled re­ cess lactic dehydrogenase. Lactic dehydrogenase (L­ action that forms glyceraldehyde-3-phosphate from lactate:NAD oxidoreductase, EC 1.1.1.27) was meas­ ribose-5-phosphat€. Transaldolase (sedoheptulose-7­ ured by following NADH oxidation in the presence phosphate:D-glyceraldehyde-3-phosphate dihydroxy­ of pyruvate. acet~metransferase, EC 2.2.1.2) activity was deter­ Radiorespirometric method. Cells for radiorespi­ mined by following NADH oxidation in the presence rometric experiments were harvested by centrifuga­ of fructose-6-phosphate, erythrose-4-phosphate, and tion at 4 C. The cells were washed three times in excess triose phosphate isomerase and a-glycero­ 0.05 M phosphate buffer (pH 7.4) and then sus­ phosphate dehydrogenase. pended without substrate in phosphate buffer at a The assay for 2-keto-3-deoxy-6-phosphogluconate cell concentration of 2 to 3 mg/ml (dry weight). (KDPG) aldolase (6-phospho-2-keto-3-deoxY-D-glu­ Radiorespirometric experiments, each done in conate:D-glyceraldehyde-3-phosphate lyase, EC triplicate, were performed as described earlier (1). 4.12.14) was accomplished by observing NADH-de­ All radiochemicals used in this study were obtained pendent reduction of pyruvate in the presence of from New England Nuclear Corp., Boston, Mass. VOL. 30, 1975 CARBOHYDRATE CATABOLISM IN AGROBACTERIUM 733

Respirometer flasks containing 60 to 100 mg of cells TABLE 1. Isotope recoveries from metabolism of were incubated at 28 C in a modified Gilson differen­ labeled glucose by Agrobacterium srrains" tial respirometer. Specifically labeled substrate (0.25 jLCil was added from a side arm; each flask Isotope recoveries' (%) contained 0.5 to 1 mg of radioactive substrate. Final [HC]_ Identity glulXlse Me- volume was 30 ml. Flasks were shaken while air was co, Cells dium Total passed through at a rate of 60 mVmin. To decrease endogenous metabolism, cultures were incubated A radiobacter Cre 73 18 8 99 for 60 to 90 min before addition ofsubstrate. Respira­ NRRL B-164 C2 49 32 9 90 tory "C0 was trapped in 10 ml of a mixture of 2 C3 28 50 26 104 absolute ethyl alcohol and monoethanolamine (2:1 C4 40 42 5 87 [voVvol]); the trapping solution was removed and a C6 22 44 30 96 fresh supply was provided at hourly intervals. Trapping solutions containing "C02 were ad­ A. tumefadensd C1 73 17 4 97 justed to 15 ml with absolute ethyl alcohol; a 5-ml NRRL B-36 C2 55 40 5 100 portion of each was mixed with 10 ml of toluene C3 34 57 15 106 containing 6 mg of 2,5-diphenyloxazole per ml and C4 50 33 9 92 0.1 mg of 1,4-bis-[21-(5-phenyloxazolyllbenzene per C6 23 50 24 97 ml. The mixtures were placed into scintillation vials, and radioactivity was measured at balance­ A. rubi NRRL C1 93 1 9 103 point conditions with an automatic liquid scintilla­ B-4017 C2 60 3 29 92 tion spectrometer. Standard deviation ofradioactive C3 46 5 42 93 measurements was no greater than 2%. C4 69 1 25 95 At the end of each experiment, the culture flasks C6 37 6 57 100 were chilled quickly and cells were separated by centrifugation at 4 C. Cells were homogenized in A. rhizogenes C1 75 14 14 105 NCS solubilizer (Amersharn/Searle Corp., Des NRRL B-193 C2 33 39 7 79 Plaines, Ill.) and incubated at 37 C for 12 h; the C3 27 42 12 81 quantity of radioactive carbon incorporated was C4 61 18 6 85 then determined by suspending the cells in a scintil­ C6 27 34 23 84 lation cocktail containing (vol/voll: toluene-2,5-di­ phenyloxazole - 1,4 - bis - [21 - (5 - phenyloxazolyl)ben­ A. stellulatum C1 68 7 17 92 zene: Triton X-100:ethyl alcohol (8:4:3). Samples of NRRL B-4018 C2 51 18 12 81 the phosphate buffer also were mixed with this C3 45 26 15 86 scintillation mixture and analyzed. Counting effi­ C4 64 1 13 78 ciency for each sample was determined by appro­ C6 39 26 36 101 priate internal standards. Calculations. Total 1·C0 recoveries for [4­ 2 A. pseudotsugae C1 38 41 7 86 1·C1glucose were calculated by comparing total re­ NRRL B-4016 C2 25 54 4 83 coveries from both [3-"C1glucose and [3,4­ C3 40 41 12 93 "C1glucose. Quantity of labeled C4 recovered was C4 45 23 10 78 calculated from the following relationship: C4 = C6 16 44 25 85 2(C3,4) - (C3). Relative participation of carbohy­ drat.e catabolic pathways involving simultaneous op­ A. gypsophilae C1 58 6 26 90 eration of the Embden-Meyerhof-Parnas (EMP) and NRRL B-4015 C2 62 6 13 81 pentose phosphate (PP) pathways and concurrent C3 65 4 19 88 activity of the ED and PC pathways was estimated C4 65 2 17 84 according to the methods of Wang et al. (23) and of C6 45 4 36 85 Arthur et al. (1), respectively. " Cells were cultured in a chemically defined me­ dium described by Kaneshiro (11) and harvested RESULTS during midexponential growth. Identifying and estimating the percent partic­ b Average of three separate experiments; incuba­ ipation of pathways involved in glucose catabo­ tion was for 2 to 4 h. lism in the various organisms required (i) com­ C Numbers designate specifically labeled carbon atoms. paring cumulative 14C02 generated from specifi­ d Data previously reported in reference 1. cally labeled n-glucose, (ii) analyzing maximal interval rates of 14C02 released from specifi­ cose utilization by the seven agrobacteria. Maxi­ cally labeled n-glucose, and (iii) assaying for mal interval releases of respiratory "C02 by the presence in extracts ofkey glucose-cataboliz­ each species utilizing specifically labeled n-glu­ ing enzymes. Recorded in Table 1 is an account­ cose are plotted in Fig. 1. Estimates of the ing ofthe various quantities oflabeled radioiso­ relative extent ofutilization ofthe various path­ tope recovered in experiments measuring glu- ways by each organism are presented in Table 734 ARTHUR ET AL. A1'PL. MICROBIOL.

BOr------,,------, r------~

70 A. rodiobocler A. rubi A. rhizogenes HRRl 8·164 HRRl 8·4017 c, HRRl 8·193 c, (. 60

~50

60 A. slellulolum A pseudolsugoe A. gypsophiloe HRRl 8·4018 HRRl 8·4016 HRRl 8·4015

50

-;; 40

1 2 3 4

FIG. 1. Aerobic radiorespirometric patterns of glucose utilization by vegetative cells of Agrobacterium strains (see Materials and Methods for specific media). Cl, C2, C3, C4, and C6 designate specifically labeled carbon atoms. One-half milligram of specifically labeled substrate (0.25 p.Ci) was added to each reaction flask. Final volume per flask was 30 ml.

2. A catalog of key enzymes for glucose catabo­ was Cl > C4 > C2 > C3 > C6 (Fig. Ib), suggest­ lism is recorded in Table 3. ing concurrent operation of the ED and PP A. radiobacter and A. tumefaciens. In pathways (28). In contrast to the activity dis­ terms ofrate and extent ofconversion, the pref­ played by A. radiobacter and A. tumefaciens, erential order of oxidation ofthe glucose carbon the greater and more rapid oxidation ofC4 than skeleton was Cl > C2 > C4 > C3 > C6 for A. C2 by A. rubi and A. rhizogenes revealed that radiobacter (Fig. la). A. tumefaciens displayed catabolism proceeded predominantly through so similar a pattern that separate graphs illus­ the ED pathway. Participation of the ED and trating the activities of both A. tumefaciens PC pathways was 77 and 23%, respectively and A. radiobacter are not presented (see refer­ (Table 2); for A. rhizogenes (Fig. lc), the ED ence 1). Active evolution of Cl, C2, and C4 as was 88% and the PC was 12%. Significantly,

CO 2 indicated that the major mechanisms for A. rubi utilized PC reactions extensively as catabolizing glucose were the ED and PC path­ evidenced by its release of a greater quantity of ways (1), utilized to the extent of 56 and 44%, C2 as CO 2 (Fig. Ib) thanA. rhizogenes achieved respectively, in A. radiobacter (Table 2) and 55 (Fig. lc). The possibility that A . radiobacter, A. and 44%, respectively, in A. tumefaciens (1). tumefaciens, A. rubi, and A. rhizogenes could A. rubi and A. rhizogenes. The preferred operate the ED and PC pathways was con­ order of oxidizing glucose carbons by A. rubi firmed by demonstration of the presence in ex- VOL. 30, 1975 CARBOHYDRATE CATABOLISM IN AGROBACTERIUM 735

tracts of the necessary enzymes (glucose-6­ minimal quantities ofphosphofructokinase and phosphate dehydrogenase, 6-phosphogluconate fructose-l,6-diphosphate aldolase is consistent dehydrogenase, transketolase, transaldolase, with the hypothesis based on results from and KDPG aldolase). Detection ofno more than radiorespirometry that the EMP pathway did not operate in the four organisms. TABLE 2. Relative participation ofglucose pathways A. stellulatum. A. stellulatum oxidized Cl in Agrobacterium strains and C4 to CO2 at approximately equal rates whereas C2, C3, and C6 were oxidized less rap­ Participation of glu­ cose pathways (%) idly (Fig. Id). Equivalence of CO2 evolution Identity from Cl and C4 and the presence of KDPG ED PC EMP aldolase (Table 3) strongly suggest that glucose a was metabolized almost exclusively via the ED A. radiobacter NRRL B-164 56 44 pathway. Although PP pathway enzymes were A. tumefadensa NRRL B-36 55 45 A. rubia NRRL B-4017 77 23 detected, they appeared not to function during A. rhizogenesa NRRL B-193 88 12 primary glucose metabolism in nonproliferat­ A. stellulatumb NRRL B-4018 100 ing cells. The apparent inactivity of the PP A. pseudotsugae C NRRL B-4016 22 d 78 pathway and absence of EMP enzymes leave A. gypsophilae" NRRL B-4015 the tricarboxylic acid (TCA) cycle as the most likely mechanism for oxidizing C2, C3, and C6 " Estimated by method of Arthur et al. (1). to CO2 • Further evidence for involvement ofthe b Estimated by inspection of Fig. Id. TCA cycle is the delayed release of C2, C3, and C Estimated by method of Wang et al. (23). C6 relative to Cl and C4. Preliminary unpub­ d Represents PP pathway rather than PC activity. lished radiorespirometric assays of release of e Because three pathways may function, percent participation calculations were not made for this CO2 from specifically labeled TCA cycle inter­ organism. mediates indicate that in the seven species of

TABLE 3. Activity ofenzymes ofglucose catabolism in Agrobacterium strains"

Sp act' A. radio- A. tume- A. rhizo- A. stellu- A.pseudo- A. gypso- Enzyme A. rubi bacter faciens NRRL genes latum tsugae philae NRRL NRRL B-4017 NRRL NRRL NRRL NRRL B-164 B-36c B-193 B-4018 B-4016 B-4015 Phosphoenolpyruvate-glu- <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 cose phosphotransfer- ase Glucokinase 85 84 66 100 10 83 23 Glucose-6-phosphate dehy- 19 28 46 24 5 22 20 drogenase Phosphoglucoisomerase 89 81 35 51 16 24 71 Fructose-6-phosphate <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 phosphoketolase Phosphofructokinase <0.1 <0.1 <0.1 <0.1 5 15 18 Fructose-1,6-diphosphate <0.8 <0.8 <0.8 <0.8 <0.8 23 18 aldolase 6-Phosphogluconate dehy- 19 23 46 24 8 22 20 drogenase KDPG aldolase 195 185 185 205 23 <0.2 55 Pyruvic kinase 56 83 69 23 16 23 25 Lactic dehydrogenase <0.5 <0.5 28 56 23 16 28 Transaldolase 120 240 40 250 63 160 106 Transketolase 123 106 129 110 11 12 32 Enolase 191 153 177 174 35 137 34

a Cells were cultured as described in Materials and Methods and harvested during midexponential growth. I' Values are expressed as 10-3 enzyme unit per mg of protein and are the average of three separate measurements. To calculate net specific activities of enzyme reactions involving either NADH or reduced NADP oxidation, NADH oxidase and reduced NADP oxidase activities were subtracted from the respective total activity values. C Data previously reported in reference 1. J

736 ARTHUR ET AL. AppL. MrcROBIOL. agrobacteria it is utilization of the TCA cycle stellulatum, A. gypsophilae, and A. pseudotsu­ for oxidization of the catabolic intermediates gae utilized different ones. Previous radiorespi­ during and after depletion of glucose that gives rometric studies demonstrated that utilization the observed order and timing of CO2 labeling. of a common set of mechanisms for glucose A. pseudotsugae and A. gypsophilae. That catabolism may be heavily relied upon to char­ the EMP pathway predominated in A. pseudot­ acterize species of well-defined genera 09, 27, sugae and A . gypsophilae was evidenced by the 28l. preferential oxidation of C3 and C4 to CO2 (Fig. Numerical taxonomic analyses (7, 24) and Ie, D. Furthermore, selective decarboxylation DNA composition (5) and homology (9) reveal of Cl over C6 and C4 over C3 indicates opera­ similarities between A. radiobaeter and A. tu­ tion of the PP pathway in these bacteria. The mefaeiens and serve as bases for arguments respective extent of operation of the EMP and favoring the designation ofA. tumefaeiens as a PP pathways in A. pseudotsugae was 78 and pathogenic variant of A. radiobaeter. That 22% (Table 2). Because three pathways may be strains representative of these two presumed operative, the fractional contribution of each by species used similar mechanisms ofglucose me­ A. gypsophilae was not calculated. Recovery of tabolism is consistent with this hypothesis. Al­ C2 and C6 and CO 2 suggests extensive cycling though there are contradicting opinions 02, of carbons via pentoses. The redistribution of 17), most workers have concluded (as we have carbon atoms presumably followed the pattern also on the basis ofthe current radiorespiromet­ described by Beevers (2). Extracts ofA. pseudo­ ric studies) that A. rhizogenes can be distin­ tsugae exhibited no KDPG aldolase activity, guished from A. tumefaeiens and A. radiobae­ whereas those of A. gypsophilae did (Table 2). ter, and from A. rubi as well. In fact, the specific activity for this enzyme was The species status ofA. rubi is not as clear as as high as that of other enzymes assayed in A. that ofA. rhizogenes. Some taxonomists argue gypsophilae. Therefore, PP and ED pathways that A. rubi and A. tumefaeiens are identical possibly provide supplementary mechanisms because: (i) A. rubi is not specifically patho­ for glucose catabolism in A. gypsophilae; in A. genic for the Rubus plant; (ii) in terms of gua­ pseudotsugae, Cl and C2 ofglucose was catabo­ nine-cytosine and homology, the DNA of A. lized only through the PP and EMP pathways. rubi is very similar to that of A. tumefaeiens; and (iii) nitrite and indole production by each is approximately equivalent (5, 9, 14, 17). In con­ DISCUSSION trast, numerical analyses align A. rubi more From the foregoing studies, the following closely to the other agrobacteria but designate groupings appear valid: (i) in the strains of A. it as a separate entity (24). Our study reveals radiobaeter and A. tumefaeiens used, glucose similar mechanisms of glucose catabolism for catabolism was attributable almost equally to A. rubi and A. tumefaeiens but demonstrates the activity of the enzymes of the ED and PC that the metabolic relationship between these pathways, whereas the A. rhizogenes and A. two organisms is not as close as that between rubi strains utilized the ED pathway predomi­ A. radiobaeter and A. tumefaeiens. nantly and the PC pathway to a lesser degree; The following factors heavily weigh against (ii) the A. stellulatum strain catabolized glu­ including A. stellulatum in the genus Agrobae­ cose solely via the ED pathway; and (iii) the A. terium: (i) polar flagellation, (ii) preferential gypsophilae and A. pseudotsugae strains catab­ growth on seawatermedium, (iii) non-pathogen­ olized glucose mainly by the EMP pathway icity for the tomato and Datura plants, (iv) with minor involvement of the PP or ED path­ failure to produce 3-ketolactDse, and (v) compo­ ways or both. sitional characteristics of its DNA (5, 91. In recent taxonomic schemes based on bio­ Marked use of the ED pathway in A. stellula­ chemical and physiological criteria, A. radio­ tum (see Fig. 1) is consistent with De Ley's sug­ baeter, A. tumefaeiens, A. rhizogenes, and A. gestion (5) that this bacterium is more closely rubi are the only four species recognized as related to the marine pseudomonads than tD agrobacteria (5, 24). Furthermore, it has been the agrobacteria. suggested that A. tumefaeiens, A. rhizogenes, The emergence of the EMP pathway as the and A. rubi are merely pathogenic variants of principal means for glucose utilization in A. A. radiobaeter (5, 7, 9, 13). Accordingly, it is pseudotsugae and A. gypsophilae represents a not surprising that representative strains of clear divergence ofthese bacteria from the met­ these four organisms displayed similar mecha­ abolic pattern exhibited by A. radiobaeter, A. nisms for glucose catabolism, although frac­ tumefaeiens, A. rubi, and A. rhizogenes. Evi­ tional concurrent degrees of utilization of the dence based on differences in DNA properties, pathways varied. Nor is it unexpected that A. lack of tumorogenicity, pigment formation, and VOL. 30, 1975 CARBOHYDRATE CATABOLISM IN AGROBACTERIUM 737 numerical analyses opposed the inclusion ofA. ribonucleic acid homology and taxonomy ofAgrabac­ pseudotsugae and A. gypsophilae in the genus terium, Rhizobium, and Chramobacterium. J. Bacte­ riol. 94:116-124. Agrobacterium (5, 9, 24). 10. Hylemon, P. B., and P. V. Phibbs, Jr. 1972. Independ­ Catabolism of glucose by the ED pathway ent regulation of hexose catabolizing enzymes and without EMP pathway involvement (as in the glucose transport activity in Pseudomonas aerugi­ A. radiobacter, A. tumefaciens, A. rubi, A. nasa. Biochem. Biophys. Res. Commun. 48:1041­ 1048. rhizogenes, and A stellulatum strains used 11. Kaneshiro, T. 1968. Methylation of the cellular lipid of here) is characteristic of other aerobic bacteria methionine-requiring Agrabacterium tumefaciens. J. such as Pseudomonas aeruginosa (15, 21), P. Bacteriol. 95:2078-2082. stutzeri (19), and several strains of Neisseria 12. Keane, P. J., A. Kerr, and P. B. New. 1970. Crown gall of stone fruit. II. Identification and nomenclature of gonorrhoeae (18). In the absence of EMP capac­ Agrobacterium isolates. Aust. J. BioI. Sci. 23:585­ ity, carbon cycling via pentoses could be a signif­ 595. icant feature in those bacteria that utilize the 13. Kersters, K., J. De Ley, P. H. A. Sveath, and M. ED pathway as the major respiratory pathway Sackin. 1973. Numerical taxonomic analysis ofAgro­ bacterium. J. Gen. Microbiol. 78:227-239. and as the primary mechanism for glucose as­ 14. Kline, D. T., and R. M. Klein. 1973. Transmittance of similation as well as for production of biosyn­ tumor-inducing ability to avirulent crown-gall and thetic intermediates. Although glycerol was ob­ related bacteria. J. Bacteriol. 66:220-228. served to induce hexose-catabolizing enzymes 15. Lessie, F., and F. C. Neidhardt. 1967. Adenosine tri­ phosphate-linked control ofPseudomonas aeruginosa in P. aeruginosa (10), we observed no such en­ glucose-6-phosphate dehydrogenase. J. Bacteriol. hancement of activity in glycerol-grown cells of 93:1337-1345. A. tumefaciens (1), and furthermore, no affect 16. Lowry, O. H., N. J. Rosebrough, A. L. Farr, and R. J. by 3-ketoglucose-synthesizing system on the Randall. 1951. Protein measurement with the Folin phenol reagent. J. BioI. Chern. 193:265-275. relative participation of ED and PC pathways. 17. McKeen, W. E. 1954. A study ofcave and crown-galls on Vancouver Island and a comparison of causal orga­ ACKNOWLEDGMENTS nisms. Phytopathology 44:651-655. 18. Morse, S. A., S. Stein, and J. Hines. 1974. Glucose We thank W. A. Wood, Michigan State University, for metabolism in Neisseria gonorrhoeae. J. Bacteriol. his helpful advice during this investigation. Special appreci­ 120:702-714. ation is given G. L. Adams for his able technical assistance 19. Spangler, W. J., and C. M. Gilmour. 1966. Biochemis­ in this work. try of nitrate respiration in Pseudomonas stutzeri. I. L. O. A. was the recipient of a postdoctoral resident Aerobic and nitrate respiration routes of carbohy­ research associateship established by the Agricultural Re­ drate catabolism. J. Bacteriol. 91:245-250. search Service, U.S. Department ofAgriculture, in associa­ 20. Schramm, M., V. Klybas, and E. Racker. 1958. Phospho­ tion with the National Academy of Sciences-National Re­ rolytic cleavage of fructose-6-phosphate by fructose-6­ search Council, 1970-1972. phosphate phosphoketolase from Acetobacter xy­ !inum. J. BioI. Chern. 223:1283-1288. LITERATURE CITED 21. Tiwari, N. P., and J. J. R. Campbell. 1969. Enzymatic control in the metabolic activity of Pseudomonas 1. Arthur, L. 0., L. A. Bulla, Jr., G. St. Julian, and L. K. aeruginosa grown in glucose or succinate media. Nakamura. 1973. Carbohydate metabolism in Agro­ Biochim. Biophys. Acta 192:395-401. bacterium tumefaciens. J. Bacteriol. 116:304-313. 22. Vardanis, A., and R. M. Hochster. 1961. On the mecha­ 2. Beevers, H. 1956. Intermediates of the pentose phos­ nism ofglucose metabolism in the plant tumor-induc­ phate pathway as respiratory substrates. Plant Phys­ ing organismAgrobacterium tumefaciens. Can. J. Bio­ iol. 31:339-347. chern. Physiol. 39:1165-1182. 3. Breed, R. S., E. G. D. Murray, and N. R. Smith (ed.l. 23. Wang, C. H., I. Stem, C. M. Gilmour, S. Klungsoyr, D. 1957. Bergey's manual of determinative bacteriol­ J. Reed, J. J. Bialy, B. E. Christensen, and V. H. ogy, 7th ed. Williams and Wilkins, Baltimore, Md. Cheldelin. 1958. Comparative study ofglucose catabo­ 4. Buchanan, R. E., and N. E. Gibbons (ed.l. 1974. lism by the radiorespirometric method. J. Bacteriol. Bergey's manual of determinative bacteriology, 8th 76:207-216. ed. Williams and Wilkins, Baltimore, Md. 24. White, L. O. 1972. The taxonomy of the crown-gall 5. De Ley, J., M. Bemaerts, A. Rassel, and J. Guilmot. organism Agrabacterium tumefaciens and its relation­ 1966. Approach to an improved taxonomy ofthe genus ship to rhizobia and other agrobacteria. J. Gen. Micro­ Agrobacterium. J. Gen. Microbiol. 43:7-17. bioI. 72:565-574. 6. Ghosh, S., and D. Ghosh. 1968. Probable role ofa mem­ 25. Wood, W. A. (ed.). 1966. Methods in enzymology, vol. 9. brane-bound phosphoenolpyruvate-hexose phospho­ Academic Press Inc., New York. transferase system ofEscherichia coli in the permea­ 26. Zagallo, A. C., and C. H. Wang. 1962. Comparative tion of sugars. Ind. J. Biochem. 5:49-52. carbohydrate catabolism in Arthrobacter. J. Gen. Mi­ 7. Graham, P. H. 1964. Application of computer tech­ crobiol. 29:389-401. niques to the taxonomy of root-nodule bacteria. J. 27. Zagallo, A. C., and C. H. Wang. 1967. Comparative Gen. Microbiol. 35:511-517. carbohydrate catabolism in Corynebacteria. J. Gen. 8. Haynes, W. C., L. J. Wickerham, and C. W. Hesseltine. Microbiol. 47:347-357. 1955. Maintenance of cultures of industrially impor­ 28. Zagallo, A. C., and C. H. Wang. 1967. Comparative tant microorganisms. Appl. Microbiol. 3:361-368. glucose catabolism ofXanthomonas species. J. Bacte­ 9. Heberlein, G. T., J. De Ley, and R. Tijtgat. 1967. Deoxy- riol. 93:970-975.