Glycerol Protection and Purification of Bacillus subtilis Glucose Dehydrogenase*

Bacillus subtilis glucose dehydrogenase (EC 1.1.1.47) has been purified from sporulating cell extract to apparent homogeneity (as determined by polyacrylamide gel electrophoresis, sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and isoelectric focusing). The enzyme purified as a single molecular species with no evidence for a multiple form of the enzyme. The B. subtilis glucose dehydrogenase has an apparent isoelectric point of 4.7-4.8 and an apparent Mr = 126,000 and is comprised of four subunits of Mr = 31,500 each. The glucose 2-deoxyglucose and glucosamine substrate specificity of the enzyme is similar to the substrate specificity for B. subtilis spore germination, suggesting that the spore glucose dehydrogenase may play some role in spore germination. The B. subtilis glucose dehydrogenase is extremely dependent on the presence of glycerol or other hydrophobic bond-stabilizing agents (or NAD) for retention of enzymatic activity, and the presence of glycerol (20% w/v) in the extraction and purification buffers was absolutely necessary for the successful purification of this enzyme.

glucosamine substrate specificity of the enzyme is similar to the substrate specificity for B. subtilis spore germination, suggesting that the spore glucose dehydrogenase may play some role in spore germination. The B. subtilis glucose dehydrogenase is extremely dependent on the presence of glycerol or other hydrophobic bond-stabilizing agents (or NAD) for retention of enzymatic activity, and the presence of glycerol (20% w/v) in the extraction and purification buffers was absolutely necessary for the successful purification of this enzyme.
Spore-forming bacteria such as Bacillus subtilis synthesize a number of new proteins during the formation of their heatresistant endospores (1-7). One such protein is the enzyme glucose dehydrogenase (EC 1.1.1.47) (8,9) which presumably functions during glucose-induced spore germination following preincubation of the spores with fructose, asparagine, and potassium (10,11). Although the appearance of glucose dehydrogenase has been repeatedly used as a biochemical marker in the study of bacterial sporulation (12-181, the lability of B. subtilis glucose dehydrogenase (19) has been an obstacle in the purification, characterization, and use of this enzyme in biochemical (20) and genetic (21) studies of bacterial sporulation (22).
The present paper reports the glycerol stabilization of B. subtilis glucose dehydrogenase, thus permitting purification of the enzyme to apparent homogeneity, and reports some of the properties of the purified enzyme. It also reports the similarities between the carbohydrate substrate specificity for the purified enzyme and that for B. subtilis spore germination, suggesting that glucose dehydrogenase may play some role in glucose-initiated spore germination.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

RESULTS
Purification of the B. subtilis Glucose Dehydrogenase-Glucose dehydrogenase can be purified from B. subtilis spores (23). However, rather drastic means (use of glass beads and grinding in a cell mill) are needed to disrupt the spores. The enzyme can also be extracted from sporulating cells of B. subtilis by lysozyme sonication or French cell rupture (23). However, once the forespore has further matured into a completed spore, spore enzymes can no longer be obtained by these milder extraction procedures (13)(14). To obtain sufficient quantities of the glucose dehydrogenase under mild extraction conditions, we have used a B. subtilis mutant (61297) blocked at sporulation stage V (24-26), i.e. after glucose dehydrogenase had been formed (17)  Procedures"). The maximum level of glucose dehydrogenase was obtained from cells harvested about 5 h after the end of truly exponential growth (about 3 h after the culture turbidity was no longer increasing). At this time, about 10-20% of the cells showed phase bright sporangia with dark centers (normal B. subtilis sporangia at this same time would be completely phase bright); the appearance of these sporangia provided the indication of the culture's readiness for harvesting. The sporulating cells were disrupted and the glucose dehydrogenase was purified as described under "Experimental Procedures." The purification procedure employed sequential column chromatography on DEAE-Sephacel, Ultragel AcA-34, hydroxyapatite, phenyl-Sepharose, and finally purification by chromatofocusing. The results of the individual purification steps are summarized in Table I. The presence of 20% (w/v) glycerol in all buffers was absolutely necessary for preservation of enzyme activity. The purified enzyme has been stored at -20 "C in 20% (w/v) glycerol, 0.05 M imidazole buffer (pH 6.5) for more than 4 years without any appreciable loss of activity. The necessity for a serine protease inhibitor phenylmethylsulfonyl fluoride and 2-mercaptoethanol in the  100  64  55  51  45  42  27  19 18 ~~ a A glucose dehydrogenase enzyme unit assayed at room temperature (23 "C) corresponds to 2.5 units at 37 "C. (Thus, 368 pmol/min/ mg of protein at 23 "C is equivalent to 920 pmol/min/mg of protein at 37 "C.) buffers has not been directly demonstrated but these compounds were added as a precautionary measure. Phenylmethylsulfonyl fluoride was always added directly to the buffers in the dry form, or in ethanol, since aqueous solutions of phenylmethylsulfonyl fluoride rapidly lose their activity (28).
Properties of B. subtilis Glucose Dehydrogenase-The purified glucose dehydrogenase migrated as a single protein band in polyacrylamide gel electrophoresis (which co-migrated with glucose dehydrogenase activity as measured by the tetrazolium-linked assay) at pH values of 6.43 ( Fig. S-8), 7.5, and 8.5. 20% (w/v) glycerol was added to the gel to preserve enzymatic activity and to prevent subunit dissociation which gives the appearance of multiple enzyme bands (29-31). Dissociation of glucose dehydrogenase into subunits followed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis showed a single molecular weight species of M, = 31,500 ( with the enzyme having four similar molecular weight subunits. The fact that the value obtained by gel filtration is so close to the four times the value obtained by sodium dodecyl sulfate-polyacrylamide gel electrophoresis subunit indicates that, unlike the B. subtilis inositol dehydrogenase (32), the B. subtilis glucose dehydrogenase is more globular in shape.
(Sucrose density gradient centrifugation of the B. subtilis enzyme also indicated M, = 125,000.*) The isoelectric point of the B. subtilis glucose dehydrogenase is 4.7-4.8 as determined by chromatofocusing ( Fig. S-6) or by column isoelectric focusing ( Fig. S-7). It is slightly more basic than the isoelectric point of B. subtilis inositol dehydrogenase which is 4.4 (32).
Glycerol Protection of B. subtilis Glucose Dehydrogennse-The effect of buffer composition and the necessity of glycerol for the protection of 3. subtilis glucose dehydrogenase activity were examined by dialyzing dilute quantities of both partially purified and purified enzyme against various buffers, pH values, and glycerol concentrations and determining the amount of glucose dehydrogenase activity remaining. Fig. 1 shows the effect of glycerol concentration on the stability of the enzyme in the presence of 0.05 M imidazole (pH 6.5) after dialysis for 48 h at 4 "C followed by incubation for 10 min at 45, 55, 65, and 77 "C. In the absence of glycerol, there was * R. Ramaley, unpublished observations. rapid loss of activity. However, in the presence of glycerol, purified enzyme has been kept at -20 "C for more than 4 years without any loss of activity. Imidazole appears to be the most protective buffer examined to date. HEPES? ACES, MES, and Tris-glycine have almost the same relative protective effect as imidazole (optimum pH for stability is pH 6.0-6.5). However, phosphate is considerably less protective and is no longer used as the buffer for the purification of B. stlbtilis glucose dehydrogenase (23).
The amount of glycerol required to effect protection of the enzyme activity also depends on the buffer; less glycerol was required to effect the same protection when imidazole was used as the buffer (Fig. 1) than when phosphate was used. Relatively large amounts of glycerol (i.e. 20% (w/v) in irnidazole buffer) were required to assure protection during enzyme purification; this is consistent with Gekko and Timasheffs proposal (33, 34) that glycerol's protective effect acts via a "general and nonspecific effect".
Substrate Specificity of Spore Germination and Glucose Dehydrogenase-In the presence of glucose, B. subtilis spores, which have been heat-shocked and incubated in the presence of potassium, fructose, and asparagine, begin spore germination and enter vegetative growth (10, 11). The rate of entry into the initial stage in the spore germination (phase darkening) can be studied spectrophotometrically (determining the decrease in absorbance at 625 nm of resuspended B. subtilis spores), and it is possible from the determination of the maximum rate of phase darkening as a function of glucose concentration to obtain an apparent K,,, of glucose for spore phase darkening (Table 11). Similar K, values can also be obtained for 2-deoxyglucose and glucosamine (35) and can be compared to the in vitro K,,, values observed for these same substrates with glucose dehydrogenase at an assay pH of 6.5 (Table 11). A putative role of glucose dehydrogenase in glucose spore germination is made somewhat more tenable because of the observation that the high K,,, values for glucose previously reported for this enzyme (23) at the conventional assay pH values (pH 7.8-8.0) are substantially reduced when the pH used for the assay (Table 11) is the same as the intracellular pH found in the dormant spores (pH 6.3-6.4 for Bacillus cereus and Bacillus megaterium spores) (36). Fig. 2 shows the effect of pH on the apparent K,,, and V,,, of the B. subtilis  Michaelis constant and maximum velocity for spore darkening and glucose dehydrogenase activity The per cent of V, . . / K, is calculated from the Vmax and K,,, values shown with the value obtained with glucose set at 100%. The Vmax for spore phase darkening is in units of absorbance decrease/min a t 625 nm with an initial spore density of 0.9 absorbance units. Spore phase darkening in the absence of added glucose, 2-deoxyglucose, or glucosamine was less than 0.005 units of absorbance decrease/min. The V,. for glucose dehydrogenase is in units of micromoles of NADH produced per min/mg of protein (at pH 6.5).

Spore phase darkening Substrate
Glucose dehydrogenase glucose dehydrogenase between pH values of 5.75 and 7.8. A similar decrease of apparent K,,, with decreasing pH has also been observed with the B. cereus glucose dehydrogenase by Bach and Sadoff (37).

Glycerol Protection of B. subtilis Glucose Dehydrogeme-
The use of glycerol to protect B. subtilis glucose dehydrogenase has both practical and interpretive values. It has been possible to use the glycerol protection of the enzyme to purify the glucose dehydrogenase of B. subtilis and B. megaterium* to homogeneity. Use of glycerol protection has also clarified some molecular weight discrepancies reported for B. subtilis glucose dehydrogenase. For example, Hachisuka and Tochikubo (38) reported in heated spores of B. subtilis an inactive glucose dehydrogenase (Mr = 40,000) that could be activated by incubation with EDTA and dipicolinic acid, etc to M , = 100,000. More recently Yokota et al. (19) reported in a Bacillus isolate BG 1722 an inactive form of glucose dehydrogenase (with a M , = 45,000) which was activated by NADand NADP-type compounds producing M , = 80,000. R. Ramaley and J. Vary, unpublished observations. However, gel filtration of the glucose dehydrogenase of the BG 1722 isolate of Yokota et al. (19) (in the presence of 20% (w/v) glycerol in 0.05 imidazole (pH 6.5)) gave only a single molecular specific species of approximately 125,0002 (similar to the results observed for the B. subtilis enzyme; Fig. S-11). This and NAD reactivation of B. subtilis glucose dehydrogenase (19, 23) observations are consistent with a hypothesis that, in the absence of glycerol, the B. subtilis glucose dehydrogenase readily dissociates into subunits and that NAD can reverse this dissociation. Incubation of the B. subtilis enzyme at pH 8.5-9.0 at 37 "C, in the absence of NAD or glycerol, results in loss of enzymatic activity with an increasingly longer lag time required for NAD reactivation' (suggesting that a conformational change may be required prior to reassociation to the active tetramer). The dissociation to subunits a t high pH has also been observed with the glucose dehydrogenase from B. megaterium (30), and the dissociation of B. subtilis or B. megaterium glucose dehydrogenase in the absence of NAD or glycerol followed by reassociation under conditions of enzymatic assay (which contains NAD) may explain the additional enzyme forms of glucose dehydrogenase observed following gel filtration (19) or polyacrylamide gel electrophoresis (29, 30) in the absence of glycerol.
Glycerol protection by favoring subunit association via hydrophobic bond stabilization is not unique to sporulation or to Bacillus enzymes but is likely a general phenomenon (33, 34). For example, mammalian glucose-6-phosphate dehydrogenase is stabilized by the presence of NADP (39), but it also can be purified in the absence of NADP, provided glycerol is included in all purification buffers (40,41). We have purified rat liver glucose-6-phosphate dehydrogenase for comparative purposes5 and found that although the rat liver enzyme is quite labile in the absence of glycerol, the B. subtilis glucose dehydrogenase is several orders of magnitude more labile which explains why the B. subtilis glucose dehydrogenase was so difficult to purify unless glycerol was present during purification. Hydrophobic stabilization of Bacillus glucose dehydrogenase had been previously postulated by Bach (42) and Sadoff (43) based on their observation that high ionic strength (which also promotes hydrophobic interaction) stabilized the B. cereus glucose dehydrogenase (44), and high ionic strength was also observed to be an important component in the NAD reactivation of the B. subtilis glucose dehydrogenase reported by Fujita et al. (23).
Purification and Properties of the B. subtilis Glucose Dehydrogenase-There was no evidence obtained during purification of the glucose dehydrogenase that there was more than one molecular species of the enzyme prepared from either the mature spores or sporulating cells of B. subtilis (Figs. S-1-S-6) or from B. megaterium spore^.^ Antibody prepared with the R. Ramaley and K. Barker, unpublished observations. by guest on March 22, 2020 http://www.jbc.org/ Downloaded from B. subtilis glucose dehydrogenase formed a single precipitin line with extracts prepared from mature spore or sporulating cells in double diffusion (micro Ouchterlony) plates (45), and the enzyme in the precipitin line was enzymatically active (provided 20% (w/v) glycerol in 0.05 M imidazole (pH 6.5) was present in the agarose to stabilize the enzyme and unprecipitated enzyme was removed by incubating the Ouchterlony plates in glycerol/imidazole buffer at 4 "C overnight prior to incubation with the tetrazolium reaction mixture). There is also no evidence of any precursor or another immunologically reactive form of the spore enzyme synthesized during vegetative growth, and then modified into the spore enzyme as has been proposed for the B. cerew aldolase (46). Extracts of stage 0 mutants of B. subtilis (i.e. mutants unable to start sporulation) have no antibody cross-reactive material present, which is consistent with the B. subtilis glucose dehydrogenase being a sporulation-specific protein.
The purified B. subtilis glucose dehydrcgenase has four subunits of M, = 31,500 (Fig. S-9) and appears to have a molecular weight just slightly larger than the B. megaterium glucose dehydrogenase (29). The enzymes purified from the two Bacillus species seem to be quite similar in general physical and catalytic properties, and further studies of the B. megaterium glucose dehydrogenase role in spore germination should be quite informative. B. megaterium spores, unlike B.
subtilis spores, will start to germinate (phase darken) upon incubation with glucose without the requirement of any prior incubation with other metabolites (47) although the action of B. megaterium glucose dehydrogenase may not be the trigger event (48)(49)(50) which results in irreversible cortex hydrolysis (51).
Role of Glucose Dehydrogenase in Spore Germination-Although there is similarity between the carbohydrate specificity for spore germination and the carbohydrate specificity for glucose dehydrogenase (10, 1 l), the putative participation of glucose dehydrogenase in B. subtilis spore germination remains based on circumstantial evidence, such as that included in the present paper. Definitive proof of involvement will require glucose dehydrogenase mutants which are not available at present. It seems unlikely that metabolism of glucose by glucose dehydrogenase constitutes the initial trigger step in the germination process in B. subtilis, because we have reconfirmed the absolute requirement (10, 11) for the preincubation of heat-shocked B. subtilis spores with potassium, fructose, and asparagin (30 min at 37 "C) in order to show the phase darkening of the spores, suggesting that some other binding and metabolic events must be occurring during this preincubation period. In addition, although 2-deoxyglucose is a very efficient substrate for glucose dehydrogenase, it was not quite as effective in producing phase darkening of spores (Table II), suggesting that the initial effect of glucose may require some other metabolic or transport event to facilitate spore phase darkening.
In summary, glycerol protection of B. subtilis glucose dehydrogenase has permitted its purification and its subsequent use in obtaining specific polyclonal antibody prepared in rabbits (52) and monoclonal antibody prepared by the mouse hybridoma technique.6 The rabbit antibody has recently been used to isolate the gene that codes for the B. subtilis glucose dehydrogenase (52). The current availability of purified enzyme has also permitted an examination of some of the physical and chemical properties of the enzyme, including the terminal amino acid sequence of the e n~y m e .~ L. Hendricks, M. Heidrick, and R. Ramaley