The Effect of Increased Phosphoglucose Isomerase on Glucose Metabolism in Saccharomyces cerevisiae"

Comparison of microbial strains with normal and high content of single enzymes is coming into use for metabolic analysis and in vivo assessment of enzyme function. We present an example for phosphoglucose isomerase and glucose metabolism in the yeast Saccharomyces cerevisiae. We use cell suspensions in conditions of inhibited protein synthesis and respiration, with low assimilation, rapid and linear glucose utilization, fer- mentation almost quantitative, and high enough cell density for direct preparation of extracts for metabolite analysis. The mass action ratio and fitting of fructose- 6-P and glucose-6-P concentrations and kinetic parameters of the enzyme are not inconsistent with near equilibrium of the reaction in the wild-type strain and small if any change in the high level strain. However, this conclusion would require that the V,, values underesti- mate the activity in the cell. On the other hand, the specific activities of glucose-6-P and fructose-1,6-P2 dur- ing metabolism of [2-SHlglucose are quite high which, together with knowledge of tritium exchange and isotope effects for the reaction in vitro, would point to the reaction in the wild-type strain being far from equilibrium; the specific activities are lower in the high level strain, indicating that extra enzyme is functional. One way to reconcile the latter results would be for tritium exchange to be considerably lower in vivo than known in vitro.

[2-3Hlglucose. We employ a system of analysis of metabolism in non-growing cells which may have general use.
In the standard resting cell incubation, cells harvested by centrifugation after growth from small inocula (OD,,, values of 0.02 or less) were washed once in B61 and resuspended in B61 also containing 10 pg/ml cycloheximide and 2 pg/ml antimycin A. After 30-60 min of preincubation at 30 "C with gentle mixing, glucose was added (zero time), and the culture was sampled periodically. Other than experiments only measuring glucose or ethanol (Table I), which used culture supernatants, sampling was as described (12): 0.5-ml portions were added to tubes containing 0.028 ml of 11.7 M perchloric acid, 20 m Na2EDTA, vigorously mixed for 30 s and, after at least 30 min neutralized with a solution of 3 M KHC03, and, within 1 h, centrifuged to remove salts and cell debris.
Glucose utilization in resting cells, uGlu, was essentially linear with time (e.g. Fig. 11, and fairly constant over a considerable range of OD,,, values in the incubation. For example, in a comparison of the same cells incubated at 250, 150,75, and 25 OD,,dml, the specific rates of glucose consumption (pmollOD,,&) were 2.48. 2.26, 2.11, and 1.53, respectively. Comparison of different strains or conditions of incubation or previous growth employed the same cell density.
Experiments with f2-3Hl-or fU-'4CIGlucose-Cells from 500-ml cultures in growth on minimal medium were collected at an OD,,; of 0.5-1.8, washed, and the drained pellets were resuspended in 1-2 ml of B61 and kept on ice. One sample of these cells was diluted to ODsso of 50 in 50 m KH2P04 (pH 7.4),2 m M Na2EDTA, 2 m 2-mercaptoethanol for assay of phosphoglucose isomerase. With another sample the usual preincubation with cycloheximide and antimycin A was set up at ODsso of -310, and at zero time diluted to OD580 of 250 by addition of 2% glucose containing [U-14Clglucose (ICN 11047, 348 mCi/mmol) or [2-3Hlglucose (Amersham Corp. TRK.36, 18 mCi/mmol); final specific activities are shown in Table Iv. At the appropriate timds) determined by assessment of volu with unlabeled glucose, a 0.5-ml sample was acidified and extracts prepared as above and frozen until use.
The thawed extracts were recentrifuged, and the supernatants filtered (Gelman 0.45 micro Acrodisks LC13), diluted with 3 volumes of water, and chromatographed by anion-exchange using Beckman Ultra-si1 AX columns and a 26-ml linear gradient of potassium phosphate ( 12). Fractions containing glucose-6-P were combined and the glucose-6-P converted to 6-P-gluconate, by addition to a 1-ml sample of 0.1 m NADP and 2 m MgC1, and 2 p1 of glucose-6-P dehydrogenase (Boehringer Mannheim 127 035.1 mg/ml). The reaction at 30 "C was followed to completion at A,,,. After a 5-min treatment at 100 "C and refiltration, the 6-phosphogluconate was obtained by the same regime of anionexchange chromatography. The appropriate fractions were assayed by direct addition (final volume 0.6 ml) of 5 m potassium phosphate (pH 6.7), 0.1 m NADP, and 2 pl of 6-phosphogluconate dehydrogenase

4879
(Sigma P 0632, 21 mg/ml), the reaction at 30 "C followed at A340, and radioactivity determined using the entire sample. The fructose-1,6-P, region from the first chromatogram was supplemented with 10 m MgC12, and 5 pl of fructose-1,6-bisphosphatase were added (Sigma F0254, 1 mg/ml), and after 15 min at room temperature and 5 min at 100 "C (assay showed only fructose-6-P) the fructose-6-P obtained by ion-exchange chromatography and its specific activity determined.
The [2-3H]glucose employed in these experiments was confirmed by showing conversion to labeled hexose monophosphate (as separated by paper chromatography) by incubation with ATP and hexokinase, and 98% loss of counts if phosphoglucose isomerase was included.
Assays-Phosphoglucose isomerase was assayed at 30 "C in extracts made with a French press and employed 5 m~ triethanolamine containing 10 m MgCl, (pH 7.4),0.3 m NADP, 1 m~ fructose-6-P, and 2 pl of glucose-6-P dehydrogenase (as above). Determination ofK, for fructose-6-P gave a value of 0.21 m, and the V, , values were corrected accordingly. AK, value for glucose-6-P of 2.05 m~ was obtained using the same buffer, with 0.3 m~ NADH, glucose-6-P, 1  Glucose, glucose-6-P, fructose-6-P, and fructose-1,6-Pz were all assayed spectrophotometrically (as in Ref. 12), as were glycerol, pyruvate, acetate, and ethanol; glycogen and trehalose assay were similar to that described (lo), and pyruvate kinase as in Ref. (15). Enzyme activities are expressed in the same units employed for glucose flux, i.e. micromoles of substrate converted/hour/milliliter OD580 of resting cell suspension. The in uiuo concentrations of metabolites given in Table I11 are calculated as millimolar, referred to a 0.5-ml intracellular water volumdg wet weight of packed pellet (16). The wet weight determination for strain DFY425/YEp13 in the present resting cell conditions was 1.34 mg/OD,,, and for strain DFY42WpPGl it was 1.42 mg/OD,,,; a value of 1.38 was used for both strains, or 690 p1 of cell water/OD6,,,.

RESULTS
Glucose Metabolism in Resting Cells-We have previously described (17) the metabolic characteristics, which are unexceptional, of the wild-type S. cereuisiae strain, DFY1, employed in this laboratory. During growth in the usual enriched medium, its rate of glucose metabolism, UG~,,, the product of the reciprocal of the yield times the first order growth rate constant, p, was -2.5 pmoWml ODsB0 of cells, and metabolism largely fermentative, with a yield of -1.5 ethanol and 0.1 glyceroVglucose consumed. After glucose exhaustion under aerobic conditions, with derepression of respiratory metabolism, growth continues at a lower rate at the expense of the ethanol made earlier.
As long known, sugar metabolism in microbes may be rapid in non-growing cells (18,19). When cells obtained from exponential growth in the enriched medium with glucose (i.e. at relatively low cell density when glucose is largely unconsumed) are resuspended in mineral salts buffer B61 (as used for growth but without supplements), and containing cycloheximide and antimycin A to prevent protein synthesis and respiratory ATP formation, glucose hilization was linear with time, as in Fig. 1, with a U G~" about three-fourths that in growth and a similar ethanol yield (Table I, line 1). In such conditions -90% of glucose used can be accounted for, chiefly by the fermentation products, with assimilation, assessed as trichloroacetic acidinsoluble material from incubations in [6-14C]glucose, being only a few percent (Table 11). We therefore use the uGlu value as an estimate of the net flux through phosphoglucose isomerase, uppi. (uppi will be referred to as u and is the difference between forward and reverse rates, i.e. ufur. ) values were about the same in resting cells incubated without antimycin A (Table I, (3) products regardless of cell density would probably require special aeration. As expected for a n organism which ferments aerobically, the presence of antimycin A in growth was without large effect (lines 3 and 4). On the other hand, UGI= values were significantly lower for cells grown in galactose (lines [5][6][7][8] or Products of resting cell incubations Cells of strain DFYl from exponential growth or stationary phase in R61 glucose medium were incubated as specified with 100 IIIM [6-14Clglucose. Fermentation products are shown per mole of glucose consumed, whereas trichloroacetic acid precipitable radioactivity and increments in glycogen and trehalose are expressed as percent of glucose consumed. C h , cycloheximide; Ant, antimycin A EtOH, ethanol; Gly, glycerol; Ace, acetate; Pyr, pyruvate; Ppt, trichloroacetic acid-precipitable material; Gln, glycogen; "e, trehalose; -, not measured.  With high cell densities, metabolites of interest may be obtained in adequate amounts by direct extraction. After 30 min of preincubation without glucose, glucose-6-P, fructose-6-P, and fructose-1,6-P2 are in low concentration. Upon glucose addition they rise and, after a substantial overshoot, reach a quasisteady state, returning to low values after glucose exhaustion (Fig. 1). The overshoot, often reported (e.g. 20, 21), will not be discussed further; short term oscillations of metabolites after metabolic perturbation are also known (e.g. 22).
The Effect of High Phosphoglucose Isomerase on vGlU and Metabolites-We compared a wild-type strain congenic with DFY1, DFY425, either carrying the multicopy vector y E p l 3 or this vector with the phosphoglucose isomerase gene, i.e. pPGZ (Table 111). Phosphoglucose isomerase was in -11 times normal amount in the latter strain; as shown for pyruvate kinase, the plasmid does not affect amounts of other glycolytic enzymes (11). By assay, the V,,, forward value of 21.5 for the strain with normal amount of enzyme is in -20-fold excess to the actual glucose flux value (vcIu) of 1.15 (same units). One would not, therefore, expect the latter value to change in the strain with high level of enzyme, and indeed it was about the same (1.34). Glucose-6-P and fructose-6-P concentrations were barely changed in the high level strain either (Table 111).
The mass action ratios, fructose-6-P/glucose-6-P, for wildtype and high level strain were both 0.29 (Table 1111, not clearly different from the known Keq values at 30 "C of 0.30-0.32 (13), and show that the reaction is at or near equilibrium even in the wild-type strain. This conclusion is not exceptional (23) and indeed was reached previously for wild-type yeast (24). It does require that there be sufficient enzyme. With a net flux unequal to zero the reaction cannot be exactly at equilibrium, and uJvf = (fructose-6-P/glucose-6-P)(1/Ke,) (25). So, for the actual vfv, of 1.2 and, as an example, a v, of 7, the fmctose-6-P/ glucose-6-P ratio would be 0.26, only slightly different from the 0.29 value found. But a u, of 7 would require (see also below) a much higher V, , than its measured value of 7. Thus, it may be that the in vitro V, , , values underestimate their value in the milieu of the cell.
The same conclusion is reached using the concentrations of the two metabolites, the K,,, values, and the steady state rate equation and analogously for v,. The present data for the wild-type strain would give a vf of 4.33 and a v, of 4.02, a net of 0.31 which, as above, underestimates the vGIu of 1.2; 6-phosphogluconate inhibition, known in vitro (13) might further depress the expected rates. However, we note that, aside from the usual possibility that the enzyme parameters are slightly different in the cell, and the reported range of K, values being large anyway (13), the calculation is very sensitive to small errors. (For example, if glucose-6-P concentration were 2.2 mM instead of 1.8 mM and the other values were unchanged, vfu, would be 1.2.) Thus, as with the mass action ratio, the data on concentration of metabolites together with known kinetic constants are not out of line with the reaction being close to equilibrium in the wild-type strain and having more enzyme making little difference.

Metabolism of [2-3HIGlucose-[2-3H]
Glucose has been employed to assess the phosphoglucose isomerase reaction in vivo (reviewed in Ref. 261, ever since demonstration of significant retention of tritium during in vitro catalysis: equilibration of [2-3H]glucose-6-P or [l-3H]fructose-6-P with enzyme gives complete exchange of tritium into water, while trapping of the product shows intramolecular proton transfer between C1 and C2 and only partial exchange with protons from the medium (27)(28)(29). In the present work we have done the resting cell incubations with [2-3Hlglucose and obtained the specific activities of intracellular glucose-6-P and of fructose-1,6-P2.
As tested with [U-14C]glucose (Table IV, lines 1-3), the isolated compounds had specific activities similar to the glucose employed, a s expected (Fig. 1) for the pools being derived from external glucose, With tritiated glucose, experiments with cells from the steady state showed glucose-6-P-specific activities

Specific activities of metabolites during resting cell incubation with
radioactive glucose Glucose-6-P was obtained from the acid-soluble pool and converted to 6-phosphogluconate and fructose-1,6-P2 was obtained from the acid soluble pool and converted to fructose-6-P. Values in parentheses are percent of input glucose specific activity. See "Materials and Methods."

Experiment
Glucose Glucose-6-P Fructose-l,6-P2 261 (36) phase of the incubations where metabolite concentrations peak before a In experiments 6 and 9, the samples were obtained during the early falling to a steady level (see Fig. 1).
-1.1 times that of the glucose itself, in the normal strain (lines 4 and 51, and about half that value in the strain with a high level of phosphoglucose isomerase (lines 7 and 8). Values for fructose-1,6-P2 were about half those of glucose-6-P. Single experiments done with cells obtained from the early time where metabolites are in transiently higher concentration (Fig. 1) gave even higher labeling (lines 6 and 9).
The expected specific activities should be related to retention of tritium in each single one way reaction and isotope discrimination against tritiated substrate (14,28), and to the actual forward and reverse rates. A formulation analogous to that of Katz and Rognstad (26,301 is shown in Table V. Fig. 2 shows G, the normalized specific activity of glucose-6-P, as a function of v, according to Equation 5, using the four parameters, tritium retention, and isotope discrimination for each direction reported recently for the yeast enzyme (14,32).
Because of the amounts of intermediates (Table III), F, the specific activity of fmctose-6-P, was not obtained and application of Equation 3 (which requires values of both F and G ) employed, instead of F , the specific activity of fmctose-1,6-P2. Fructose-l,6-P2 coming from fructose-6-P should have the same specific activity, but later exchange reactions might reduce the value. Hence, use of Equation 4, which requires only the value of G, is preferred. us obtained from Equations 3 and 4 and the data of Table IV are shown in the lower part of   Table V. Both calculations give u, values for the wild-type strain of only -4.

DISCUSSION
As mentioned, strains with increased activity of single enzymes are coming into use for studies in vivo. For example, Brindle (33) concluded that there was acceptable fit, in uivolin vitro, for increased phosphoglycerate kinase in yeast. For another reversible reaction, aldolase in Escherichia coli the data could also be fitted, but required that in vivo V, , , be higher than assayed (12). For irreversible reactions, with increased phosphofructokinase in yeast a compensatory decrease in fmctose-2,6-P2 was found to explain the absence of large effect on fructose-6-P concentration (34). And for high expression of fmctose-1,6-bisphosphatase in yeast, a lack of futile cycling in glucose metabolism was suggested to reflect its inhibition by fmctose-2,6-P2 (35).
For these cases, interpretations have been conservative, and unusual hypotheses (see below) were not needed to reconcile in 'ose Isomerase in Yeast 488 1

TABLE V Calculation of u from specific actiuities
In the equations uGlu is normalized to 1.0, and u is the rate of the phosphoglucose isomerase back reaction, normalized to uGlu. rf and rr are the tritium retention fractions in the forward and reverse reactions, respectively, and df and dr the discriminations against tritium (rate with W/rate with 'HI, values for the reaction in uitro being 0.53, 0.42, and 0.25, 0.43, respectively (14,31). For calculation, the first two values, reported for 37 "C, were corrected for 30 "C by a factor of 1.2 (27). For steady state metabolism of [2-3Hlglucose, Equations 1 and 2 are the conservation equations for radioactivity in glucose-6-P (G) and fmctose-6-P (F), expressed as specific activity normalized to input glucose. Equation 3 is Equation 1 solved for u, which can thus be obtained from values of G and F, while Equations 4 and 5 come from substitution of Equation 2 into Equation 1, showing u as a function of G and vice versa. d r ) (df) ( r b ) (rf), and b = ( d r ) (dn. Equation 5 is plotted in Fig. 2. u, values, according to Equations 3 and 4, using the averaged specific activity obtained for the steady state (Table IV, experiments 3 , 4 , 7 , and 8, respectively), are also shown, and uf obtained as u, + uGlu. To use Equation 3, F was assumed to equal the specific activity of fmctose-1,6-P2 (see text). vivo and in vitro results. In the present work, there are three findings to consider. First, the mass action ratios, which show the phosphoglucose isomerase reaction to be near equilibrium even in the wild-type strain, and hence high one way fluxes. Second, the retention of label from [2-3H]glucose in glucose-6-P, which points to lower flux values in the wild-type strain (u,s of -4, Table VI. And third, the reduced labeling of glucose-6-P in the strain with high level of enzyme, which calculates (Table V) to a vr of only three times wild-type even though the measured factor of enzyme increase was 11 and substrate and product concentrations were almost unchanged. These results would be partially reconciled if, as well as the measured Vmaxs underestimating the values in the cell, intra-molecular proton retention values for the single one way reactions in vivo were higher than known in vitro. As an example, using the average of the two isotope discrimination values employed for Table V (i.e. df = dr = 0.33) but increasing the proton retention fractions to 0.98 (for both rf and rr; they should be the same for the known enzyme mechanism (28)), then for the wild-type strain G value of 1.15 v , would be 8.4, and for the high level strain G value of 0.54 it would be 83. Thus the factor of extra enzyme would be as measured and the actual one way rates also much higher.
Other, perhaps less likely, models should be mentioned. Retention of tritium in products from [2-3Hlglucose has been often reported (see 261, although direct measurement of glucose-6-P is rare (36,371. Dorrer et al. (38) suggested that in Chlorella the fact of similar specific activities to glucose, per hexose in sucrose, pointed to channeling of hexose-phosphates in a multienzyme complex. And Malaisse and Bodur (37) show calculations, based on data from pancreatic islet cells, interpreted as best fitting with partial channeling between phosphoglucose isomerase and phosphofructokinase. The same type of model could be applied here. Or, one might speculate that extra enzyme is at least partially not functional because of being in the wrong place or differently modified. Or, a regulatory circuit acting to establish hexose monophosphates at their apparent equilibrium ratio might be imagined.
Finally, the present method for assessing metabolism in nongrowing yeast differs in some ways from custom, using inhibition of protein synthesis rather than implicitly relying on no new enzyme being made during an incubation, antimycin D for a non-respiratory condition, linear glucose use and proportionality to amount of cells, and some knowledge of assimilation.
There is a similar protocol for E. coli (12). This type of incubation may have general use for in vivo studies of enzymes changed in amount or kinetic characteristics. It will be also be interesting to apply it to cases where enzyme amount can be set to lower than normal.