Role of Pyruvate Carboxylase, Phosphoenolpyruvate Carboxykinase, and Malic Enzyme during Growth and Sporulation of Bacillzcs subtilis

Abstract In extracts of Bacillus subtilis, CO2 fixation occurs primarily through the apparently constitutive enzyme pyruvate carboxylase, which is strongly activated by acetyl-CoA. This enzyme is necessary for growth on glucose but is not required for sporulation, as was established with a pyruvate carboxylase mutant. The malic enzyme can use either NAD or, less effectively, NADP as cofactor. The ratio of these activities remains constant through enzyme purification and during enzyme induction by l-malate. Enzyme synthesis is not repressed by glucose. Malic enzyme and pyruvate carboxylase form a "pyruvate shunt" to the citric acid cycle, which apparently is necessary during growth on malate for the production of oxalacetate in substrate amounts; malic dehydrogenase functions mainly to provide energy via the citric acid cycle. A specific and sensitive [γ-32P]ATP assay for P-enolpyruvate carboxykinase has been developed. Using this assay, a purified enzyme preparation gave a Km for oxalacetate of about 25 µm. Enzyme synthesis is repressed by glucose. P-enolpyruvate carboxykinase mutants have established that the enzyme is needed for gluconeogenesis and, under normal growth conditions, for sporulation. Sporulation can be restored by the continuous feeding of gluconate.

which is strongly activated by acetyl-CoA. This enzyme is necessary for growth on glucose but is not required for sporulation, as was established with a pyruvate carboxylase mutant. The malic enzyme can use either NAD or, less effectively, NADP as cofactor.
The ratio of these activities remains constant through enzyme purification and during enzyme induction by L-malate. Enzyme synthesis is not repressed by glucose. Malic enzyme and pyruvate carboxylase form a "pyruvate shunt" to the citric acid cycle, which apparently is necessary during growth on malate for the production of oxalacetate in substrate amounts; malic dehydrogenase functions mainly to provide energy via the citric acid cycle. A specific and sensitive [y-32P]ATP assay for P-enolpyruvate carboxykinase has been developed. Using this assay, a purified enzyme preparation gave a K, for oxalacetate of about 25 PM. Enzyme synthesis is repressed by glucose. P-enolpyruvate carboxykinase mutants have established that the enzyme is needed for gluconeogenesis and, under normal growth conditions, for sporulation. Sporulation can be restored by the continuous feeding of gluconate. l\Iicroorganisms begin to differcnt,iatc into dormant forms when nutritional conditions become growth rate-limiting. This usually occurs w-hen the rapidly metabolizable carbon sources, e.g. carbohydrates, have been exhausted (I, 2). Nevertheless, some carbohydrates are incorporated into polymers during diffcrcntiation, indicating the need for gluc~oneogenesis. WC? use gluc~oncogeIlesis as a general term for the synthesis of any carbohydrate from C3 and cl4 compounds.
To elucidate the control of gluconeogcnesis in differentiation, we have investigated the enzymes connecting t,he citric acid cycle to the Embdcll-llleyrrhof path (Fig. I).
Since its biochemistry and sporulatiou have been well studied, we used Bacillus subtilis.
In this organism, the roles of glycolytic (2, 3) and citric acid cycle enzymes (446) in sporulation have been esamincd.
However, little was howl about t,he enzymes that control CO:! metabolism, some of which arc gluconeogenic. At the branch points of P-enolpyruvate and pyruvate, glycolysis and anaplerosis converge and gluconeogenesis begins. Regulation of this crucial arca determines the direction of carbon flow.
We have determined the mrchanism controlling the synthesis and activation of the cnzymcs and analyzed in mutants the effect of enzyme deficiencies 011 growth and sporulation.
We have found that the malic enzyme of B. subtilis uses either NAD or NADP; the activity rat'io remains constant during purification alltl induction by malate. For P-enolpyruvate carbosykinase we have developed a specific sensitive assay and demonstrated o.lucosc repression. a This enzyme is required for gluconeogenesis and sporulation, as n-as shown with a Penolpyruvate carbosgkinase mutant.
Sporulation of this mutant could be restored by continuous feeding of gluconate.
Illedia--TMW plates, NShlJ', and N have been described (7). n,I medium contained N plus 1 mg per ml of sodium citrate.
13ot.h N and M media always contained 25 pg per ml of L-tryptophan and 10 pg per ml of L-methionine, and a carbon source (50 m&I, unless stated 0therITise).
Bacterial Stmins--All strains wcrc derived from the transformable 168 strain of B. subtilis.
Our strain, 60015, requires L-methionine and L-tryptophan for growth and sporulates well in NSMP.
The two P-enolpyruvate carboxykinase mutants (61101 and 61104) were isolated from G°Co-irradiated spores of strain 60015 as colonies that could not grow on plates with ?j plus L-malate but could grow on N plus glucose. They produced pale colonies on TIM3 plates. We obtained from ,J. Hoch (Scripps Clinic and Research Foundation, La Jolla, C'alif.) the pyruvate carbosylasc mutant 61437 (C50) and the malic dehydrogenase mutant 61461 (JH421); both rcquirc L-tryptophan for growth.
The other malic dehydrogcnase mutant 61421 (lh21) was obtained from R. Hanson (Department of Hacteriology, University of Wiscollsin) ; it requires L-tryptophan and contains an additional mut'ntion that causes the production of large nlesosomes.2 Both malic dehydrogenase mutants produce pale colonies on TBAB or NSMP plates and very few spores. Strains were stored at -60" in X plus 25% glycerol. Growth-Bacteria were incubated overnight on TBAB plates and either used directly or, after streaking once more onto plates containing X plus carbon sources, for induction or repression esperiments.
Media were incubated at an initial AeO,, = 0.1. Cells for P-enolpyruvate carboxykinase isolation were inoculated from N plus L-malate plates into the same medium in Erlenmeyer flasks of 5 times the liquid volume and were shaken at 120 strokes per min at 37". When the Asoo was 1, the cells were inoculated into 12 liters of fresh medium in a fermenter, again grown to Bcao = 1, harvested, and washed in wash buffer (50 mnf Tris-Cl. pH 7.5, 100 mM KCl, 10 mM mercaptoethanol, and 0.1 mM EDTA), and the pellet was stored at -40". The malic enzyme was similarly isolated from the malic dehydrogenase strain, 61461, grown in M plus glucose plus 5 ma< potassium-r-malate.
Preparation of Extracts and Enzyme Assays-All preparation and purification steps were performed at O-5". Cells were suspended in an "extraction buffer" (containing 50 mM Tris-Cl, pH 7.5, 1 m&f MgCl?, and 10 rnM mercaptoethanol) at a concentration of 0.1 g per ml wet weight, ruptured in a French pressure cell, and centrifuged at 37,000 x g for 30 min. The proteiu concentration of the sapernatant extract was 8 to 10 mg per ml.
CO2 fixation was assayed according to Sundarum et al. (8). P-enolpyruvate carboxykinase activity was determined by a 32P transfer assay described in Fig. 5 or, if purified preparations were used, by the coupled spectrophotometric assay of Shrago et al. (9). blalic dehydrogenase was assayed by oxalacetat,e reduction according to Yoshida and Freese (10).
For the malic enzyme assay, cells were extracted in extraction buffer plus 1 mM potassium-n-malate and assayed immediately as described under "Results" (see Table III). Protein was determined by the method of Lowry et al. (11). PuriJication o;F Malic Enzyme-The extract was treated with 20 mg of protamine sulfate per g of protein.
After centrifugation, 313 mg of ammonium sulfate were added per ml of supernatant and the pellet was discarded.
Ammonium sulfate, 214 mg per ml, was added and the pellet was dissolved in >io the original volume of extraction buffer containing 1 mM malate and 50% glycerol and stored at -20".
Two milliliters of the concentrated enzyme solution containing 12 to 15 mg of protein per ml were dialyzed against a buffer containing 0.01 M potassium phosphate, pH 6.5, 1 mM MgC&, 10 m&f mercaptoethanol, 1 mM malate, and 20% glycerol. The sample was applied to a hydroxylapatite column (1 x 17 cm2) which had previously been equilibrated in the dialysis buffer. The column was washed with 20 ml of the buffer and was eluted with the following linear phosphate gradients (in the same buffer) : 0.01 to 0.03 M (40 ml), 0.03 to 0.05 M (80 ml), and 0.05 to 0.1 M (100 ml). All activity eluted as a single symmetric peak in the final gradient step. This activity was concentrated by ultrafiltration (Diaflo, Amicon Corp.). Puri'cation of P-enolpyruvate Carboxykinase-Cell extracts were treated with protamine sulfate as for the malic enzyme. Bmmonium sulfate, 351 mg per ml, was added and the pellet was discarded.
Additional 179 mg of ammonium sulfate per ml were added and the pellet was stored at -20".
The pellet was dissolved in >io the original volume of extract and 277 mg of ammonium sulfate per ml were added.
The new 6063 pellet was dissolved in extraction buffer containing 10 y. glycerol and 1 mM ATP and was dialyzed against the same medium. A DEXE-Sephadex A-50 column (1.3 x 15 cm2) was washed with 10 to 15 void volumes of the above dialysis medium and a 1 ml sample containing 20 mg of protein was applied.
The column was eluted successively with 20 ml each of dialysis medium containing 0.05, 0.1, and 0.2 M Tris-Cl.
A linear gradient prepared by mixing 50 ml each of dialysis buffer containing 0.2 and 0.6 M Tris-Cl was then applied.
The activity eluted as a single peak between 0.2 and 0.4 M Tris-Cl.
It was concentrated by ultrafiltration.
Activities were determined by the V transfer assay prior to the final ammonium sulfate step and by both this assay and the coupled spectrophotometric assay thereafter. Sporulalion-The frequency of heat and octanol resistant spores was measured as described by Freese et al. (12).
ChemicaZs-[y-32P]ATP was prepared according to Glynn and Chappell (13) except that the $Tl-' was further purified on a Dowex 1 (chloride form) column (0.3 x 0.6 cm) as recommended by Dr. R. Lazzarini of our laboratory. Five milliliters of the reaction mixture were applied and the column was washed with water and t.hen with 10 ml of 0.01 M sodium folmate, pH 3.4, and 0.1 M LiCl.
It was precipitated by the addition of 8 ml of absolute alcohol.

Properties of Various Mutants-The
biochemical block of several mutants employed in this paper is shown in Fig. 1, while their doubling times in different media are summarized in Table  I. The two P-enolpyruvate carboxykinase mutants (61101 and 61104) could not grow on malate as sole carbon source but they grew on glycolytic carbon sources. In contrast, the pyruvate carboxylase mutant (61437) grew at a significant rate only on media which supplied citric acid cycle intermediates.
Two malic dehydrogenase mutants (61121 and 61161) grew on all carbon sources but at a very low rate on malate alone.
Apart from the enzymes investigated in this paper, several other enzymes had the normal specific activities in the mutants. These include the inducible P-enolpyruvate transferase, phosphofructokinase, and pyruvate kinase in 61437, and isocitrate dehydrogenase, fumarase, and aconitase in 61101 and 61161.
Whereas the standard strain (60015) and the pyruvate carboxylase mutant sporulated normally, the malic dehydrogenase and P-enolpyruvate carboxykinase mutants were defective in sporulation.
The sporulation deficiency of malic dehydrogenase mutants is typical for mutants of the citric acid cycle (4,5,14), while that of the P-enolpyruvate carboxykinase mutant (61104) is analyzed later in this paper.
CO2 Fixation in B. Xubtilis Extracts and Control of Pyruvate Carboxylase-The fixation of CO2 was assayed in extracts of our standard strain (60015) in the presence of various additions as summarized in Table II  1. B. subtilis reactions involved in COZ metabolism and others discussed in these pages: (1) pyruvate kinase (const,itutive); (8) pyruvate carboxylase (activated by acetyl-CoA) ; (3) malic dehydrogenase; (4) malic enzyme (active with NAD or NADP, induced by malate) ; (5) P-enolpyruvate (PEP) carboxykinase (active with ATP or GTP, repressed by glucose). (Line 2). The removal of either substrate (Lines 3 and 4) or of MgCIZ (Line 5) nearly obliterated the activity. The requirements of ATP, magnesium, and acetyl-CoA and the 90% inhibition by the biotin-complexing protein avidin (Line 6) clearly indicated the presence of pyruvate carboxylase (pyruvate : carbon dioxide ligase (ADP) EC 6.4.1.1) whose activity was not dependent on the growth medium.
The pyruvate carboxylase reaction was linear with time and protein concentration in extracts of the standard strain (60015), while the mutant (61437) produced less than % of this activity (Fig. 2).
Little CO2 fixation was observed with P-enolpyruvate alone, even in the presence of acetyl-CoA or Pi, regardless of the growth condition (Table II (10). Nevertheless, to establish that the observed reaction was caused by malic enzyme WC used two mutants (61121, 61161) that displayed 40 times less malic dehydrogenase activity, than the standard strain (Table III). Their if the reaction mixture contained only Tris, malate, and NAD. In contrast, the sustained malic enzyme reaction depended on the complete reaction mixture (Table III) and was linear, after the 1st min, for 5 min for enzyme activities between 0.5 and 10 nmoles per min. We concluded from these observations that malic enzyme activity can be reliably detected in crude extracts of B. subtilis by using as substrat,es L-malate and NADP or NAD. it decreased much more rapidly when any of the ingredients were omitted. The specific activity of malic enzyme was always 4 to 6 times higher when the bacteria were grown in the presence of malate than in its absence (Table III).
This induction was not prevented by the presence of glucose, whether it was measured in the standard strain or in the malic dehydrogenase mutants (Table IV).
The (sust,ained) malic enzyme activity measured with NAD was always about 5 times higher than that measured with NADP (Tables III and IV).
This constant ratio suggested that the same malic enzyme accepted either cofactor.
To establish this more thoroughly, we have partially purified the malic enzyme activity.
Preliminary experiments had shown that malic dehydrogenase would be difficult to separate from malic enzyme. Therefore, we used extracts of the malic dehydrogenase mutant (61161) which had been grown in N plus glucose and malate.
During the 3to j-hour interval required to purify through the ammonium sulfate steps, more than 50% of the original activity was lost. However, most of this activity was recovered after suspension of the 50 to 80% ammonium sulfate fraction in extraction buffer plus 50% glycerol (Table V). If this suspension was stored at -2O", no loss in activity was noted over a B-week period.
For further purification, the suspension was dialyzed overnight against elution buffer containing 20% glycerol with no loss of activity and was applied to a hydroxylapatite column. After elution more than 90% of the original activity was recovered.
Throughout all purification steps and in all active fractions, the NAD : NADP ratio of activity remained constant (Table V). In the pooled column fractions the NAD-associated activity had a pH optimum of 8.0 whereas the NADP-associated activity had a broad pH profile with a midpoint of about pH 8.0 (Fig. 3). Therefore, we assayed the NADP-associated activity in the foregoing experiments (Tables IV and V) at pH 6.5 to minimize the NAD-associated activity, which in turn was measured at pH 8. When the pooled activity of enzyme was passed through a Sephadex G-200 column, the NAD-and NADP-dependent activities eluted in the same peak distinctly after the void volume. All of our results support the assumption that the NAD and NADP activities are associated with the same enzyme.
PuriJication and Control of P-enolpyruvate Carboxykinase-Initially, we used the COZ fixation assay shown in Table I  For NAD and NADP, the buffers were Tris-Cl, pH 8.0, and potassiummorpholinopropanesulfonate, pH 6.5, respectively. b NG, no growth. mine the linearity of P-enolpyruvate carboxykinase activity wit1 time and protein concentration in crude extracts (Fig. 4, Curve 60015).
Although MnClz (2 mM) in the assay mixture produced twice as much activity as MgC12 (10 mM) we avoided manganese because it often produced a precipitate.
To enable accurate comparisons of the enzyme activities at different times of growth in different media and in mutants a reliable, specific, and sensitive assay was needed. Previously, P-enolpyruvate carboxykinase had been detected by a variety of met,hods. Simplest are the direct spectrophotometric assays coupled either to malic dehydrogenase (9) or to pyruvate kinase and lactic dehydrogenase (I 5). Since NADH and the other substrates can react in other ways, these assays are unreliable in crude extracts.
Other methods, including the one used here, measure the fixation of '4C02 into acid-stable counts (8) ; they are more specific, but less sensitive.
Assays in the direction of P-enolpyruvate formation have the advantage that enzyme turnover is 7 to 10 t.imes higher than in the reverse. But methods employing the Warburg apparatus for determining CO2 liberation or measuring the production of I'-enolpyruvate calorimetrically (16) are also insensitive.
These difficulties can be overcome by measuring the transfer of 32P from ATP to P-enolpyruvate.
'*P can be specifically and quantitatively cleaved from the produced 3"P-labeled P-euolpyruvate by hydrolysis with HgCl, (17). We Cells were grown in NSMP and extracts were assayed as on Line 8 in Table II have verified the quantitative cleavage of P-enolpyruvate to pyruvate and Pi in the concentration range used in the assay; under these conditions ATP was not hydrolyzed.
The unreacted [y-32P]ATP was removed by charcoal absorption and the 32Pi was counted.
At least 957? of "Pi was recovered and less than 1% of the [T-~~P]ATP remained in solution after charcoal treatment.
In crude extracts of cells (60015 grown in N plus malate), the reaction rate increased linearly with time and protein concentration (Fig. 5). The activity observed without oxalacetate was 10 to 15'3, of the activity observed with the substrate.
Using this 32P transfer assay, Tve routinely obtained 5-to IO-fold higher specific activities than with the COz fixation assay. The assay was reproducible and specific activities of extracts from bacteria grown and harvested under identical conditions varied less than 10%.
The specific activity of cells grown in the presence of glucose was always 8 to 10 times lower than in cells grown without glucose, whether it was measured by the 32P transfer or the CO2 fixation assay (Tables II and VI).
This glucose repression was observed in all mutants except the P-enolpyruvate carboxykinase mutants (61101 and 61104) which displayed little activity (Figs. 4 and 5) under any growth conditions (Table VI).
The purification of P-enolpyruvate carboxykinase is summarized in Table VII. Cells (60015) were maximally derepressed by growth in N plus malate.
Another 40% ammonium sulfate cut removed 10yo of the remaining pyruvate kinase activity 25 to 50% protein, and left more than 50%  of the original P-enolpyruvate carboxykinase activity which could now be measured reliably by the coupled spectrophotometric assay. Dialysis against extraction buffer containing substrates and 10 7. glycerol and elution from a DEAE-column produced a single symmetrical activity peak with 85 to 90% recovery.
The activity of the pooled peak, concentrated by ultra-filtration, was unstable, making further purification difficult.
The enzyme eluted (recovery about 50%) from a Sephadex G-100 column with the exclusion volume, indicating a molecular weight above 100,000. Thus tile B. subtilis enzyme appears to resemble the cytoplasmic enzymes of Tetrahymena (18) and yeast (19) rather than the rat liver enzyme (20).
Kinetic measurements of the P-enolpyruvate carboxykinase purified through the DEAE-step showed the greater reliability of the 3zP transfer assay (Fig. 6). Oxalacetate saturated the enzyme at 75 pM concentration with half-saturating concentrations of 20 to 30 pM.
The saturation curve was a normal hyperbola with no suggestion of cooperativity.
The K, value for oxalacetate was an order of magnitude lower than that of other reports (21,22). No activation or inhibition by 1 to 5 mM amounts of fructose-Pz, glucose-B-l', or AMP was detected either in the coupled spectrophotometric or in the 3?? transfer assay. Similarly, 0.1 pM acetyl-CoA was without effect.

Enzyme
Changes during Growth and Sporulation in NSMP-Since some enzyme activities change during sporulation (3,5,12,23), the activities of the above enzymes were measured during growth and sporulation in NSMP. Pyruvate carboxylase and P-enolpyruvate carboxykinase activities remained essentially constant while malic enzyme increased toward the end of exponential growth (Fig. 7). If NSMP was supplemented with an amount of glucose (25 mM) which did not prevent eventual sporulation (2), P-enolpyruvate carboxykinase activity remained repressed to the end of growth and increased subsequently whil the cells entered the sporulation process. Glucose did not ir fluence the activities of pyruvate carboxylase or malic enzyme.  (Table VIII).
The P-enolpyruvate carboxykinase mutant (61104) grew we but lysed after growth had ceased (Fig. 8). Even among th surviving cells very few heat-resistant spores (Table VIII, Fig. 8 or octanol-resistant spores (Fig. 8) were produced. The fel sporulating cells may have salvaged the necessary carbohydrate from the debris of the lysed cells. The lysis could be preventer by the addition of carbohydrates.
The restoration of sporula tion was more difficult, because the onset of spore development i suppressed by carbohydrates (3 cells did not lyse and sporulated reasonably well (Fig. 8). If a carbohydrate was added only once, either at the beginning or during exponential growth, the cells merely grew to a higher A eoO, lysed, and sporulated only slightly more frequently.
To establish that the deficiency of both sporulation and Penolpyruvate carboxykinase activity in strain 61104 were caused by the same point mutation, revertants were isolated by two methods. In the first method lo7 cells were plated on several NSMP plates and 10 sporulating colonies, remaining on the lawn of lysed cells after 3 days at 37", were picked, purified, examined for growth on malate, and assayed for enzyme activity.
In the second method, cells were plated on N plus malate and nine colonies were picked after 7 days, purified, examined for sporulat.ion, and assayed for enzyme activity.
All revertants could grow on malate, sporulated at the normal frequency, and produccd the normal specific activity of P-enolpyruvatc carboxykinase. DISCUSSION The major CO&ixing enzyme of B. suhlilis is pyruvate carboxylase which is strongly activated by acetyl-Coh and apparently constitutive.
Its activit,y is required for growth on carbohydrates, since a pyruvate carboxylase mutant grows in glucose only if the medium also contains a citric acid cycle compound.
The major enzyme coutrolling the reverse path from oxalacetate to Penolpyruvate is P-enolpyruvatc carboxykinase, whose activity is repressible by glucose. Mutants lacking this activity cannot grow on citric acid cycle compounds, although they grow normally on glucose. Similar results have been reported for Escherichia coli (27,28). The repression of P-enolpyruvate carboxykinase by glucose is necessary to prevent a loss of ATP by the cycle of enzymes which convert P-cnolpyruvate to pyruvate, pyruvate to oxalacetate, and osalacetate back to P-enolpyruvate, because the other two enzymes of this cycle are constitutive.
In media (e.g. N plus malate or NSMP) in which P-enolpyruvate carboxykinase is not rcpresscd and the above cycle could therefore function, P-enolpyruvate presumably is used for gluconeogenesis sufficiently fast that little of it reenters the cycle. The cycle is also controlled by the acctyl-CoR activation of pyruvate carboxylase.
The malic enzyme also is involved in CO2 metabolism. In B. subtilis this enzyme reacts with either NAD or NADP, with an activity ratio of about 5: I. In excess malate, this reaction, in contrast to that of malic dehydrogenase, causes a rather complete reduction of the nucleotides.
The major physiological direction of the enzyme apparently is from malatc to pyruvate, since the activity is induced by malate and thepyruvate carboxylase mutant cannot grow on glucose. This enzyme together with pyruvate carboxylasc thus form an inducible "pyruvate shunt" which can be used to supply osalacetate for gluconeogenesis, aspartate production, and metabolism via the citric acid cycle. However, this shunt apparently consumes too much energy to be the only way of converting malatc to oxalacetate; malic dehydrogenase is a necessary alternative, as mutants deficient in its activity grow only extremely slowly on malate.
In the shunt one ATP is used and thus more energy is required than in the direct conversion of malatc to oxalacetate.
Nevertheless, malic dehydrogcnasc alone dots also not produce oxalacetate at a rate necessary for optimal growth, as is shown by the slow growth of a pyruvate carboxylase mutant on malate.
The equilibrium constant of malic dehydrogenasc and the stabilizing effect of NADII (10) favor the production of malate.
We conclude that during growth on malate the inducible pyruvate shunt is used mainly for the production of substrate amounts of oxalacetate, while malic dehydrogenasc functions primarily to provide energy via the citric acid cycle.
At the end of growth in NSMI' all rapidly metabolizable carbon sources have been exhaustctl and the massive development of spores begins. Fnergy tlcmands are then satisfied by the oxidation of direct acetyl-Cob precursors (acetate, acetoin, etc.) via the citric acid cycle (5). In addition, gluconeogenesis is required since mutants lacking Pcnolpyruvate carboxykinase activity cannot sporulatc, uuless they arc slowly and continuously supplied with a carbohytlratc.
In fact glucosamine is required for the synthesis of mucopcptidcs (12), components of the spore cortcs (29,30).