POLYOL DEHYDROGENASES OF AZOTOBACTER AGILIS LEON MARCUS AND ALLEN G. MARR Department of Bacteriology, University of California, Davis, California Received for publication February 16, 1961

ABSTRACT differ in chain length and configuration are MARCUS, LEON (University of California, available either as natural products or can be Davis), AND ALLEN G. AIARR. Polyol dehydro- obtained by reduction of corresponding sugars; genases of Azotobacter agilis. J. Bacteriol. 82: the availability of such an array of compounds 224-232. 1961.-Two soluble diphosphopyridine- may facilitate a more precise definition of the linked polyol dehydrogenases are formed by structural specificity of induction and of Azotobacter agilis (A. vinelandii). The first, activity. The structures of some of the polyols D- dehydrogenase is induced by D- which are most significant in this investigation mannitol and all of the pentitols except L-. are as follows: is an excellent inducer of mannitol CH20H CH2 OH CH20H dehydrogenase it is not although metabolized, H-C-OH H-C-( nor does the enzyme act upon it. This allows OH HO-C-H study of the gratuitous induction of mannitol H-C-OH HFO-C-lHI HO-C-H dehydrogenase. H-C-OH H-C-(3H H-C-OH Of the polyols tested, mannitol dehydrogenase oxidizes D-mannitol, D-arabitol, D-rhamnitol, and CH20H CH2 OH CH20H perseitol, demonstrating its requirement for Ribitol D-A rabitol substrates bearing the D-manno configuration. (D-lyXitol) The corresponding 2-ketoses, D-, D- xylulose, and presumably D-rhamnulose, and CH 20H CH20H perseulose are reduced. CH20H HO-C--H H-C-OH The second enzyme, L- dehydrogenase is H-C-OH HO-C-H HO-C-H induced only by polyols containing the D-Xy10 configuration, i.e., and xylitol. L-Iditol H-C-OH H-C-OH H-C-OH dehydrogenase oxidizes D-Xy10 polyols seven HO-C-H H-C-iOH H-C-OH times faster than it does D-ribo polyols. Sub- strates oxidized include L-iditol, sorbitol, xylitol, CH20H CH220H CH20H and ribitol. The corresponding 2-ketoses, L- L-Arabitol D-Manni"tol Sorbitol (D- (L-lyXitol) glucitol or sorbose, D-fructose, D-xylulose, and D-ribulose, L-gUlitol) are reduced. The two polyol dehydrogenases have been CH20H CH2OH separated and purified by chromatography on a modified cellulose ion exchanger. H-C-OH H-C-OH HO-C-H HO-C-H This work was begun with the intent of char- HO-C-H H-C-OH acterizing a soluble inducible enzyme suitable H-C-OH HO-C-H for the study of the kinetics of induction of in the azotobacter. Polyol dehydro- CH20H CH20H genases seemed a reasonable choice for two Dulcitol L-lditol reasons. First, Burris, Phelps, and Wilson (1943) () had established that the metabolism of mannitol This paper reports the specificity of induction was inducible in Azotobacter agilis (A. vinelandii and activity of two distinct polyol dehydro- strain 0). Second, a wide variety of polyols which genases produced by A. agilis, D-mannitol and 224 19611 POLYOL DEHYDROGENASES 225 L-iditol dehydrogenases. The purification of assay buffer and held at 30 C. To a 1-cm silica D-mannitol dehydrogenase and the kinetics of absorption cell were added in order: 2.3-x ml of its induction will be discussed elsewhere. 0.05 M tris chloride buffer, pH 8.6; 0.5 ml of 0.5 M polyol; and x ml of enzyme preparation. MATERIALS AND METHODS Ordinarily, sufficient enzyme was added to give a Culture. A. agilis (A. vinelandii strain 0) was change in absorbancy at 340 m,u of 0.1 to 0.35 grown in Burk's nitrogen-free medium (Wilson per min. If the activity was low, a maximum of and Knight, 1952) modified by reducing the 0.5 ml of enzyme preparation was added, and calcium concentration, to contain the following the rate obtained was reported. Addition of 0.2 per liter of distilled water: K2HPO4, 0.8 g; ml of 0.02 M DPN started the reaction. KH2PO4, 0.2 g; MgSO4-7H20, 0.2 g; CaSO4- The reduction of ketose was measured by 2H20, 0.025 g; Na2MoO4*2H20, 0.00025 g; following the oxidation of DPNH. To a 1-cm FeNH4 (S04)2-12H20, 0.0086 g; sucrose, 20.0 g. absorption cell were added in order: 2.4-x ml of The phosphate, sulfate, iron, and molybdate 0.05 M tris chloride buffer, pH 8.6; 0.5 ml of salts were dissolved separately. The solutions 0.5 M ketose; and x ml enzyme. The reaction was were combined to give a medium free of turbidity, started by adding 0.1 ml of a freshly prepared 100 ml of which were dispensed per 250-ml solution of DPNH containing 2 mg DPNH per Erlenmeyer flask. During autoclaving, a slight ml of buffer. precipitate forms which dissolves completely Formation or disappearance of DPNH was on cooling. followed at 340 m, with a spectrophotometer, Inocula were taken from cultures growing the cell compartment of which was maintained exponentially with a specific growth rate of at 30 C. The spectrophotometer has been modified approximately 0.3 per hour. Cultures were for recording. The current of, the phototube was incubated on a rotary shaker at 30 C and har- amplified by a Kiethly electrometer, model 610. vested during the late logarithmic phase by The output of the electrometer, which is pro- centrifugation at 5,000 X g for 5 min. portional to per cent transmission, was converted Cell-free extracts. The enzymes were released to absorbancy by a diode analogue and recorded from the cells by osmotic shock (Robrish and by a Varian strip chart recorder, model G-11. Marr, 19"7). An equal volume of 2 M was Since kinetics for oxidation of polyol are zero added to the centrifuged pellet of cells and order, units of enzymatic activity can be esti- mixed thoroughly in the 50-ml plastic centrifuge mated directly from the slope of the recorded tube. After allowing at least 1 min for glycerol to change in absorbancy. However, the rate of enter the cells; five to eight volumes of cold 0.05 reduction of ketose is not constant. The rates M tris(hydroxymethyl)aminomethane acetate reported are linear estimates of the rates re- (tris acetate) buffer, pH 7.0, were discharged corded during the first minute of reaction. rapidly into the centrifuge tube as the contents The unit of enzymatic activity is defined as were stirred vigorously with a mechanical stirrer. the amount of enzyme which causes a change in The resulting viscous fluid was centrifuged at absorbancy at 340 m,u of 0.001 per min in the 10,000 to 15,000 X g for 10 min to remove standard assay with polyol. Specific activity is residual cells (less than 5% of the initial cells) defined as units of enzyme per mg protein as and the emptied cell envelopes. Centrifugation determined with the Folin-Ciocalteau reagent removes at least 90% of the reduced diphospho- (Lowry et al., 1951) using crystalline serum pyridine (DPNH) oxidase which albumin as the standard. otherxxise would interfere with the dehydrogenase Substrates. Glycerol, , ribitol (ado- assay. The crude extract was stored at 0 to 4 C. nitol), D-arabitol, L-arabitol, xylitol, D-mannitol, Dehydrogenase assay. The assay measures the sorbitol (D-glucitol), galactitol (dulcitol), and rate of formation of DPNH subsequent to the D-rhamnitol were obtained from commercial addition of diphosphopyridine nucleotide (DPN) sources. Perseitol and L-iditol were gifts from to a buffered reaction mixture containing poly- C. E. Ballou, Department of , hydric and cell extract. The assay was University of California, Berkeley, Calif. D- made in the following manner. All solutions Xylulose was obtained from G. Ashwell of the except the enzyme preparation were made in the National Institutes of Health, Bethesda, Md. 226 MARCUS AND MARR [VOL. 82

TABLE 1. Utilization of various compounds as the was equilibrated with 0.05 M tris acetate, pH 7.0, sole carbon source by Azotobacter agilis before use.

Growth No growth RESULTS D-Mannitol Galactitol Carbon source for growth. A. agilis grows well Sorbitol Ribitol in Burk's nitrogen-free basal medium containing D-Arabitol L-Arabitol a variety of polyols as the sole source of carbon. Xylitol Table 1 shows the polyols and a few related Erythritol compounds which support growth. Glycerol Induction of polyol dehydrogenases. The initial experiments on the specificity of induction by D-Fructose D- D-mannitol and sorbitol were misleading because L-Sorbose D- in both of Sucrose Lactose of impurities both polyols. Although D-Glucose these polyols were the best commercial prepara- tions available, with melting points identical One hundred milliliters of Burk's nitrogen-free with the accepted values, each was sufficiently basal medium, containing 0.1 M concentration of contaminated, presumably by the other, to the above compounds, was inoculated with 0.5 ml confuse the results of induction and assay. A. of an exponentially growing culture in Burk's agilis grown on either of the commercial polyols sucrose medium. Cultures were incubated on a as the sole carbon source gave extracts which rotary shaker at 30 C for 24 hr. oxidized both polyols (Table 2). However, the extracts of cells grown on commercial mannitol D-Ribulose was prepared by epimerization of oxidized mannitol more rapidly, and cells grown D-arabinose in dry pyridine (Glatthaar and on commercial sorbitol oxidized sorbitol more Reichstein, 1935). The o-nitrophenylhydrazone rapidly. These results suggested that different of D-ribulose was prepared and recrystallized quantities of two distinct polyol dehydrogenases from absolute (mp 167). The hydrazone were produced by growth on either D-mannitol was decomposed in an aqueous reaction mixture with excess benzaldehyde. Benzaldehydephenyl- TABLE 2. Effect of purification of D-mannitol and hydrazone was removed by filtration, the filtrate sorbitol on the induction of polyol dehydrogenases was extracted with ether, and the aqueous phase was evaporated under reduced pressure yielding Growth substrate D-ribulose as a yellow syrup. Assay substrate Commercial Commercial D-mannitol and sorbitol required purification as explained in the next section. Mannitol Sorbitol D-Mannitol was recrystallized twice from water. Sorbitol was recrystallized twice from pyridine as Recrystallized mannitol .... 100* 27 the pyridine-sorbitol complex (Strain, 1934). The Recrystallized sorbitol 20 100* complex was thermally destroyed and the pyridine was evaporated in a vacuum oven at Recrystallized 50 C giving anhydrous sorbitol which was re- Mannitol Sorbitol crystallized two times from absolute alcohol. Ribitol was recrystallized twice from hot absolute Commercial mannitol ...... 100 10 alcohol leaving behind yellow-brown impurities. Recrystallized mannitol .... 100* 1

The remaining polyols were assumed sufficiently Commercial sorbitol ...... 3.8 100 pure for the purposes of the present investigation. Recrystallized sorbitol... 0.7 100* DEAE-cellulose was purchased from Bio-Rad * The dehydrogenase activities of the extracts Laboratories. After removal of fine particles by are expressed as per cent of the rate with ho- flotation it was converted to the free base with mologous recrystallized polyol as the substrate. NaOH, eliminating a yellow material that Each extract contained approximately 3,000 absorbs ultraviolet light, and then converted to units per ml; 0.05 to 0.4 ml was used in the assay the acetate with sodium acetate. The exchanger depending on the activity obtained. 1961] POLYOL DEHYDROGENASES 227 or sorbitol. After growth on purified polyols TABLE 3. Induction of polyol dehydrogenases by (see Materials and Methods), the ability of various polyols* extracts to oxidize the heterologous commercial Specific activity assayed on: polyol was substantially reduced. The activity of Inducer these extracts on purified polyols was almost D-Man- D-Ara- Sor- Xyli- Ri- completely specific; only the homologous polyol nitol bitol bitol tol bitol was oxidized at a significant rate. A series of polyhydric and pertinent D-Mannitol . 2,)200 1,130 <10 < 10 < 10 D-Arabitol ...... <10 <10 <10 related compounds were tested for their ability 3,250 1,610 Ribitol ...... 1,670 815 <10 <10 <10 to induce polyol dehydrogenases. Since some Sorbitol ...... <10 <10 570 570 85 polyols which do not support growth may induce, Xylitol ...... 940 700 330 265 35 a complete growth medium containing sucrose None ...... <10 <10 <10 <10 <10 was supplemented with the potential inducer. * Prior experiments had demonstrated that sucrose Azotobacter agilis was grown on Burk's did not repress the induction by D-mannitol or sucrose medium supplemented with 0.1 M polyol sorbitol. or other potential inducer. Extracts of the cells The results of this experiment, shown in were tested for the reduction of DPN with the substances listed. Erythritol, glycerol, L-arabitol, Table can presum- 3, be explained most simply by galactitol, D-fructose, L-sorbose, D-mannose, ing that two distinct enzymatic activities can be D-glucose, D-arabinose, D-xylose, and sucrose do induced. D-mannitol, D-arabitol, and ribitol not induce dehydrogenases for the polyols listed induce an activity for D-mannitol and D-arabitol. in the table. Glycerol, erythritol, L-arabitol, Sorbitol induces an activity for sorbitol, D- and galactitol were not effective substrates in xylitol, and ribitol. Xylitol induces both enzy- the standard assay of any of the induced systems matic activities. The first activity mentioned examined. will be designated as D-mannitol dehydrogenase, and the second enzymatic activity, which is D-mannitol dehydrogenase. L-Iditol dehydro- identical in substrate specificity to the dehydro- genase begins eluting at 0.1 M NaCl; the peak is genase isolated from rat liver, has previously sharpened by the next higher NaCl concentration been termed L-iditol dehydrogenase (Blakely, during stepwise elution. D-Mannitol dehydro- 1951; McCorkindale and Edson, 1954). genase elutes with 0.3 M NaCl. Although ribitol is an excellent inducer of The fractions containing L-iditol dehydro- mannitol dehydrogenase, an enzyme which does genase were pooled and tested for the ability to not attack ribitol, it fails to induce L-iditol oxidize polyols and to reduce ketoses (Table 4). dehydrogenase which does oxidize ribitol. Re- Sorbitol, L-iditol, and xylitol, all of which have gardless of whether D-mannitol dehydrogenase is the D-Xy10 configuration, were rapidly oxidized induced by D-mannitol, D-arabitol, or ribitol, the with the reduction of DPN. D-Fructose, L-sorbose, ratio of enzymatic activity against D-mannitol and D-xylulose, the corresponding 2-ketoses of to the activity against D-arabitol is always these polyols, were all reduced with DPNH. In approximately two. Similarly, the relative rates addition to the D-Xy10 substrates, D-ribulose was of oxidation of xylitol, sorbitol, and ribitol by reduced but at a lower rate. Thus, the weak L-iditol dehydrogenase are independent of the activity against ribitol found in crude extracts inducers. induced with sorbitol or xylitol is accounted for To demonstrate conclusively that the activities by the L-iditol dehydrogenase. The relative rates ascribed to D-mannitol dehydrogenase and L- of oxidation of sorbitol and ribitol did not change iditol dehydrogenase were in fact the activities during purification of the enzyme. Repeated of two distinct proteins, we combined an extract crystallization of ribitol from absolute alcohol of Azotobacter grown on D-mannitol with an did not reduce the activity of L-iditol dehydro- extract of cells grown on sorbitol. The combined genase for ribitol. Furthermore, the reduction of extracts were chromatographed on DEAE- D-ribulose confirms the additional activity of cellulose. Figure 1 is the chromatogram obtained L-iditol dehydrogenase for D-ribo substrates. which demonstrates the resolution of L-iditol Table 5 summarizes the activities of D-mannitol dehydrogenase (sorbitol dehydrogenase) from dehydrogenase. D-arabitol, D-mannitol, D-rham- 228 MARCUS AND MARR [VOL. 82

25 c z

cn 20

-I

0 0O 15 c: (\ 0 c) wL rm 10 Z cn

5

0 50 100 150 200 250 FRACTION NUMBER FIG.I. Chromatogram on DEAE cellulose acetate of a mixture of extracts of cells grown on D-mannitol and of cells grown on sorbitol. The column (31 by 1.5 cm) was buffered with 0.05 M tris acetate at pH 7.0. Protein was eluted with the same buffer containing sodium chloride at the concentrations indicated. Each fraction was assayed for L-iditol dehydrogenase with sorbitol as a substrate andfor D-mannitol dehydrogenase with D-mannitol as the substrate. The solid line is the absorbancy per cm of the effluent at 280 mM; the dotted line is units of enzyme per 15 ml fraction. Fractions 118 to 142 and 222 to 246 inclusive were pooled separately. The specific activity of each enzyme was increased 20-fold.

TABLE 4. Substrate specificity of purified L-iditol TABLE 5. Substrate specificity of purified D-mannitol dehydrogenase* dehydrogenase* Polyols oxidized Activity Ketoses reduced Activity Polyols oxidized Activity Ketose reduced Ac- tivity D-Mannitol loot D-Fructose 98 Sorbitol loot D-Fructose 20 L-Iditol 102 L-Sorbose 11 D-ArabitolD-Arabitol 64 {D-xylulose~D-Rihulose 2320 Xylitol 121 D-Xylulose 92 D-Rhamnitol 112 Ribitol 9 D-Ribulose 39 Perseitol 130 * The activity of pturified L-iditol dehydro- * The activity of purified D-mannitol de- genase was measured in the standard assay (see hydrogenase was measured in the standard assay Materials and Methods) except that the one (see Materials and Methods) except that one drop of D-xylulose syrup was added in the D- drop of D-xylulose syrup was added in the D- xylulose assay. xylulose assay. t The rate of oxidation of polyols and reduction t The rates of activity of the enzyme on the of ketoses is expressed as per cent of the rate of various substrates are expressed as per cent of oxidation of sorbitol. Each assay was made with the rate of oxidation of D-mannitol; each assay 150 units of enzyme. was made with 105 units of D-mannitol dehy- drogenase. nitol, and perseitol, all of which have the D- in the configuration of carbon 5, the purified manno configuration, were oxidized with the L reduction of DPN. D-Fructose, the 2-ketose -mannitol dehydrogenase oxidizes sorbitol at less than 1 % the rate of D-mannitol and does not corresponding to D-mannitol, was reduced. D- reduce the corresponding 2-ketose, L-sorbose. Xylulose and D-ribulose are alternative 2-ketoses of D-arabitol; only D-xylulose was reduced by DISCUSSION D-mannitol dehydrogenase. Although sorbitol, Structural specificity for D-mannitol dehydro- viewed as L-gulitol, differs from D-mannitol only genase induction. A chain length of at least five 1961] POLYOL DEHYDROGENASES 229 carbons is presumably a minimal requirement for D-arabitol (R = H), D-mannitol (R = CH20H), the induction of D-mannitol dehydrogenase since D-rhamnitol (R = CH3), and perseitol (R = glycerol and erythritol fail to induce; however, CHOH-CH20H) were the only polyols oxidized as shown in Table 3, all of the pentitols except by D-mannitol dehydrogenase. The ketoses L-arabitol induce the enzyme. Ribitol is a rare reduced by the enzyme are consistent with this example of a compound which induces an enzyme rule of specificity. D-Fructose is reduced, pre- for which it is not itself a substrate, yet fails to sumably to D-mannitol; and D-xylulose, but not induce a related enzyme for which it is a sub- D-ribulose, is reduced, presumably to D-arabitol. strate. Since ribitol is not metabolized, it lends Thus, the D-mannitol dehydrogenase of A. agilis, itself to the study of gratuitous induction; that is, which requires a specified configuration of 5 the kinetics of induction are not complicated by asymmetric carbons, has the most stringent utilization of the inducer. structural requirements for substrates of any of ,Of the three hexitols tested, only D-mannitol the previously described polyol dehydrogenases induced D-mannitol dehydrogenase; sorbitol and (Edson, 1953; Shaw, 1956). galactitol were ineffective. Thus, structural Structural requirements for induction of L-iditol requirements for induction by hexitols are far dehydrogenase. Of the polyols tested only xylitol more stringent than for pentitols. The sole and sorbitol induce this enzymatic activity. dissimilarity between sorbitol, viewed as L- D-Mannitol, ribitol, D- and L-arabitol, as well as gulitol, and D-mannitol is in the configuration of erythritol, glycerol, and galactitol, are ineffectual. C-5, yet mannitol is an inducer and sorbitol is not. The configuration common to the inducers is the D-Xylo configuration. L-Jditol was not available CH20H CH2 OH in quantities sufficient for induction trials but it HO-C-H HO-C-H presumably would act as an inducer. HO-C-H HO-C-H CH20H CH20H CH20H H-C-OH H-C-OH H-C-OH H-C-OH H b-OH HO-C-H H-C-OH HO-C-H HO-C-H HO-C-H I CH20H CH2OH H-C-OH H-C-OH H-C-OH Sorbitol D-Mannitol HO-C-H (L-gUlit0ol) CH2OH H-C-OH Xylitol CH20H CH20H Substrate specificity of D-mannitol dehydro- Sorbitol L-lditol genase. The D-manno configuration is an obliga- (D-glucitol) tory requirement for a substrate of D-mannitol dehydrogenase which catalyzes the reaction: Substrates for L-iditol dehydrogenase. Both the D-xylo and D-ribo configuration are oxidized at CH2 OH carbon 2, although the D-XylO configuration HO-C-H apparently is preferred. The reactions catalyzed by L-iditol dehydrogenase are: HO-C-H H-C-OH + CH2 OH Hl-C-OH H-C-OH + DPN+ 2. HO-b-H R H-C-OH CH20 C=O CH2OH (2a) DPN+ t HO-C-H + DPNH + H C=O + D)PNH + H+ H-C-OH HO-C-H H-C-OH H-C-OH R R 230 MARCUS AND MARR [VOL. 82

Polyol oxidase Glycerol and glycol Galactitol dehydrogenases D-Iditol dehydrogenase (Bertrand-Hudson enzyme) dehydrogenases

CH20H R CH20H CH20H HO-C-H CHOH H-C-OH HO-C-H HO-C-H CHOH HO-C-H H-C-OH R R R R

Ribitol dehydrogenase L-Iditol dehydrogenase D-Mannitol dehydrogenase

CH20H CH20H CH20H CH20H H-C-OH H-C-OH H-C-OH HO-C-H H-C-OH HO-C-H or H-C-OH HO-C-H H-C-OH H-C-OH H-C-OH H-C-OH R R R H-C-OH R

FIG. 2. Substrate speci.ficities for various bacterial polyol dehydrogenases CH20H identical in substrate specificity with the L-iditol H-C-OH + DPN+ t dehydrogenase isolated from rat liver, studied by Blakely (1951) and McCorkindale and Edson, H-C-OH (1954). H-C-OH It is of some interest to compare the substrate specificities of the polyol dehydrogenases produced R CH20H by bacteria. Figure 2 shows the structural spe- cificities written for oxidation at carbon 2 of several + DPNH + H+ bacterial enzymes. For some of these enzymes the H-C-OH specificity has not been definitely established but is postulated on the basis of polyols oxidized H-C-OH by crude extracts. The familiar Bertrand-Hudson R enzyme(s) of Acetobacter is an insoluble, cyto- chrome-linked oxidase with an acid pH optimum (Arcus and Edson, 1956) located in the cell Thus, L-iditol dehydrogenase oxidizes xylitol envelope (Stouthamer, 1959). Configurational (R = CH20H), sorbitol (R = H-C-OH) and requirements for the substrate are simple; the enzyme requires the L-erythro configuration and a CH20H at carbon 1. In many cases, L-iditol (R = HO-C-H) and reduces the however, a methyl group may be substituted for the primary alcohol (Richtmeyer, Stewart, and CH20H Hudson, 1950). The remaining polyol dehydro- corresponding 2-ketoses, D-xylulose, D-fructose, genases are soluble, DPN-requiring enzymes. and L-sorbose, by reaction 2a. The oxidation of A specific DPN-linked glycerol dehydrogenase ribitol (R = CH20H) and reduction of D- from Escherichia coli oxidizes glycerol to dihy- ribulose suggest a second general reaction, 2b. droxyacetone but does not oxidize sorbitol, D- L-Iditol dehydrogenase from A. agilis is thus mannitol, erythritol, or L-2,3-butanediol (Asnis 19611 POLYOL DEHYDROGENASES 231 and Brodie, 1953). A glycerol dehydrogenase ACKNOWLEDGMENT similar in substrate specificity and heat resistance has been isolated from the halophile Pseudomonas This work was supported by grants-in-aid salinaria (Baxter and Gibbons, 1954). A second 4261 and 9860 from the National Science Founda- class of enzymes oxidize a variety of glycols such tion. as 1,2-propane diol and L-2,3-butanediol in addition to glycerol; glycol dehydrogenases have LITERATURE CITED been described in Aerobacter aerogenes (Burton ARcus, A. C., AND N. L. EDSON. 1956. Polyol and Kaplan, 1953; Lamborg and Kaplan, 1960), dehydrogenases 2. The polyol dehydrogenases E. coli (Strecker and Harary, 1954), Vibrio of Acetobacter suboxydans and Candida utilis. costicolus (Baxter and Gibbons, 1954), and Biochem. J. 64:385-394. Acetobacter suboxydans (Goldschmidt and Kram- ASNIS, R. E., AND A. F. BRODIE. 1953. A glycerol pitz, 1954). dehydrogenase from Escherichia coli. J. Biol. The polyols oxidized by extracts of Pseudo- Chem. 203:153-159. monas suggest at least two novel polyol dehydro- BAXTER, R. M., AND N. E. GIBBONS. 1954. The glycerol dehydrogenases of Pseudomonas genases which are distinguished by a difference salinara, Vibrio costicolus, and Escherichia in stability. One, galactitol dehydrogenase, coli in relation to bacterial halophilism. requires the L-threo configuration and, thus, is Can. J. Biochem. Physiol. 32:206-217. responsible for the oxidation of galactitol, L-iditol, BLAKLEY, R. L. 1951. The metabolism and anti- L-arabitol, and xylitol. Another, D-iditol dehydro- ketogenic effects of sorbitol. Sorbitol de- genase, which requires the D-threo configuration, hydrogenase. Biochem. J. 49:257-271. is active against D-iditol, D-gulitol, D-talitol, and BURRIS, R. H., A. S. PHELPS, AND J. B. WILSON. xylitol (Shaw, 1956). 1943. Adaptations of Rhizobium and Azoto- Ribitol dehydrogenase (Wood and Tai, 1958; bacter. Soil Sci. Soc. Am., Proc., 7:272-275. BURTON, R. M., AND N. 0. KAPLAN. 1953. A DPN Fromm, 1958) produced by A. aerogenes differs specific glycerol dehydrogenase from Aero- from L-iditol dehydrogenase in that ribitol bacter aerogenes. J. Am. Chem. Soc. 75: dehydrogenase does not oxidize D-XYIO substrates 1005-1006. and differs from glycol dehydrogenase in that it EDSON, N. L. 1953. The metabolism of the sugar does not oxidize glycerol or erythritol. The alcohols. Rept. Australian New Zealand precise substrate specificity of ribitol dehydro- Assoc. Advance. Sci., 29th Meeting, 281-299. genase has not been reported. FROMM, H. J. 1958. Ribitol dehydrogenase. I. L-Iditol dehydrogenase, which we have found Purification and properties of the enzyme. in A. agilis, was first isolated from rat liver J. Biol. Chem., 233:1049-1052. (Blakely, 1951) and has previously been reported GLATTHAAR, C., AND T. REICHSTEIN. 1935. to be formed by Acetobacter (Arcus and Edson, d-Adonose (d-erythro 2-keto pentose). Helv. 1954). Both the D-xylo and D-ribo polyols are Chim. Acta 18:80-81. substrates. GOLDSCHMIDT, E. P., AND L. 0. KRAMPTIZ. 1954. D-Mannitol dehydrogenase has been definitely Diphosphorpyridine nucleotide-linked propyl- established only in A. agilis. Other genera in- ene glycol dehydrogenase from Acetobacter cluding Acetobacter (MIcCorkindale and Edson, suboxydans. Bacteriol. Proc. 1954:96. 1954) and Pseudomonas (Shaw, 1956) yield LAMBORG, M., AND N. 0. KAPLAN. 1960. A com- extracts that oxidize D-mannitol with the pro- parison of some vic-glycol dehydrogenase duction of DPNH; however, the responsible en- systems found in Aerobacter aerogenes. Bio- zyme may not be identical in substrate specificity chim. et Biophys. Acta 38:272-283. with the D-mannitol dehydrogenase of A. agilis. LOWRY, 0. H., N. J. ROSENBROUGH, A. L. FARR, AND R. J. RANDALL. 1951. Protein measure- for the metabolism A possible alternative route ment with the Folin phenol reagent. J. Biol. of D-mannitol is by phosphorylation followed by Chem. 193:265-275. oxidation of D-mannitol phosphate to D-fructose MCCORKINDALE, J., AND N. L. EDSON. 1954. 6-phosphate. Mannitol 1-phosphate dehydro- Polyol dehydrogenase. I. The specificity of genase has been found in E. coli (Wolff and rat-liver polyol dehydrogenases. Biochem. J. Kaplan, 1956) but has not been detected by us 57:518-523. in A. agilis. RICHTMEYER, N. K., L. C. STEWART, AND C. S. 232 MARCUS AND MARR [VOL. 82

HUDSON. 1950. L-Fuco-4-ketose, a new sugar method of isolation, properties and deriva- produced by the action of Acetobacter sub- tives. J. Am. Chem. Soc. 56:1756-1759. oxydans on L-. J. Am. Chem. Soc. STRECKER, H. J., AND J. HARARY. 1954. Bacterial 72:4934-4937. butylene glycol dehydrogenase and diacetyl ROBRISH, S. A., AND A. G. MARR. 1957. Osmotic reductase. J. Biol. Chem. 211:263-270. disruption of azotobacter. Bacteriol. Proc. WILSON, P. W., AND S. G. KNIGHT. 1952. Experi- 1957:130. ments in bacterial physiology. Burgess Pub- SHAW, D. R. D. 1956. Polyol dehydrogenases. lishing Co., Minneapolis. 61 p. 3. Galactitol dehydrogenase and D-iditol WOLFF, J. B., AND N. D. KAPLAN. 1956. D-Mannitol dehydrogenase. Biochem. J. 64:394-405. 1-phosphate dehydrogenase from Escherichia STOUTHAMER, A. H. 1959. Carbohydrate me- coli. J. Biol. Chem. 218:849-869. tabolism of the acetic acid bacteria. Doctoral WOOD, W. A., AND J. J. TAI. 1958. Pentitol me- thesis, Utrecht, The Netherlands. tabolism by Aerobacter aerogenes. Bacteriol. STRAIN, H. H. 1934. D-Sorbitol: A new source, Proc. 1958:99.