Effect of Oxygen on the Regulation of Intermediate Metabolism in Tetrahymena”

SUMMARY Tetrahymena grown in poorly aerated cultures have a greater capacity to utilize oxygen than cells grown in relatively well aerated cultures. Paradoxically, the oxidation of [l-14C]-glucose was inhibited, while oxidation of [Z-Wlpyruvate and [Z-W]glyoxylate was enhanced in cells grown under relatively anaerobic conditions. Total glycogen content measured after 17 hours of growth was increased 30 to 60% in cells grown partially anaerobically. In l-hour incubations at the end of this time, the capacity to incorporate label into glycogen from [l-14C]glucose was unchanged, but label incorporation from [Z-Wlpyruvate and [Z-r4C]glyoxylate was increased several fold. The ratio of adenosine di- and triphosphates was nearly identical in the cells grown under different conditions of oxygenation, indicating that this ratio may not play a major role in regulating these changes. After 17 hours of growth in cultures of different depths, cells were also incubated with a mixture of acetate, pyruvate, and octanoate, with one substrate labeled at a time in such a way that [l-14C]-acetyl-CoA is generated at the initial step in the metabolism of each. These results were interpreted in terms of a

conditions. Total glycogen content measured after 17 hours of growth was increased 30 to 60% in cells grown partially anaerobically.
In l-hour incubations at the end of this time, the capacity to incorporate label into glycogen from [l-14C]glucose was unchanged, but label incorporation from [Z-Wlpyruvate and [Z-r4C]glyoxylate was increased several fold.
The ratio of adenosine di-and triphosphates was nearly identical in the cells grown under different conditions of oxygenation, indicating that this ratio may not play a major role in regulating these changes. After 17 hours of growth in cultures of different depths, cells were also incubated with a mixture of acetate, pyruvate, and octanoate, with one substrate labeled at a time in such a way that [l-14C]acetyl-CoA is generated at the initial step in the metabolism of each. These results were interpreted in terms of a pre: viously developed three-compartment model of acetyl-CoA metabolism.
Glyconeogenesis from peroxisomal and mitochondrial precursors was increased in cells grown in low oxygen tension, with the greater contribution coming from the peroxisomes.
Oxidation of acetate and pyruvate was increased under these conditions, but appearance of [l-14C]acetate label in glutamate was decreased.
Lipogenesis from labeled peroxisomal precursors was also increased in cells grown under relatively low oxygen tension.
After a shift down in O2 tension there is a rapid rise in glyconeogenesis from the peroxisomes which levels off after about 4 hours, whereas the rate of oxidation in the Krebs cycle increases steadily for at least 8 hours following the transition to relatively anaerobic conditions. In response to a shift up in O2 tension there is a decline in peroxisomal glyconeogenesis which continues for 8 hours, whereas the rate of oxidation in the Krebs cycle does not begin decreasing until about 4 hours after the increase in O2 tension.
Thus the flux of [1-r4C] temporal pattern in mitochondria as compared to peroxisomes, and in each compartment the sequence of changes in response to a shift up in O2 tension is not the mirror image of the sequence in response to a shift down.
The ciliate Tetrahymena pyriformis can synthesize over 20% of its dry weight as glycogen from noncarbohydrate precursors (1) and is thus an excellent model cell for study of the regulation of glyconeogenesis.
During the transition from the exponential to the stationary phase of growth, the rate of glyconeogenesis from lipid precursors and glycogzn content increase markedly (2). Levy and Scherbaum (3) found that the characteristic increase in glyconeogenesis of transition phase cultures could be induced merely by changing the conditions of growth from a shaken, well aerated state to a static, partially anaerobic state. The 8-to 13.fold increase in glyconeogenesis from trace levels of [2-14C]acetate was associated with a doubling in 3 hours of the specific activities of isocitrate lyase and malate synthetase which comprise the glyoxylate by-pass of the oxidative decarboxylation steps of the Krebs tricarboxylic acid cycle and thus permit the net synthesis of glycogen from acetate or other sources of acetyl-CoA in this cell. Levy (4) later found that although the increase in isocitrate lyase and malate synthetase activity could be prevented by inhibitors of RNA and protein synthesis, the increase in incorporation of label from [2-14Clacetate was not prevented, indicating that activation of the glyoxylate by-pass could be achieved in the absence of protein synthesis. Hogg (5) has similarly found that agitation of Tetrahymena at low partial pressure of oxygen leads to an increase in glyconeogenesis without a large increase in isocitrate lyase activity.
In the absence of pyruvate carboxylase (6), oxaloacetate formed from the product of the malate synthetase step is converted into phosphoenolpyruvate by phosphoenolpyruvate carboxykinase, which is localized in both the cytosol and the mitochondria of Tetrahymena (7). The specific activity of the cytoplasmic enzyme was 20-fold higher in standing cultures than in shaken cultures, while the activity of the mitochondrial enzyme remained at a high level independent of the degree of aeration.
In this respect Tetrahymena is analogous to guinea pig liver, in that both enzyme forms are present and it is the cytosol phosphoenolpyruvate carboxykinase which increases under all conditions in which there is increased glyconeogenesis.
(10) is also adeauate for the auantitative descrintion of inter- between the logarithmically growing and the stationary phase systems and suggested that the difference in oxygen tension between their conditions (shallow well agitated cultures) and those of Connett and Blum (deep  a concentrated sample of the column eluate to two-dimensional chromatography on microcrystalline cellulose thin layer plates with the use of solvent 1 and solvent 2 as described by Raugi et al. (8).
The same sample was also subjected to two-dimensional thin layer chromatography on Silica Gel G plates with the use of phenol (go%)-water (83:17) followed by l-butanol-acetic acid-water (4:l:l) as described by Brenner (16 All other reagents were of highest purity obtainable.

RESULTS
A variety of methods have been used to obtain partially anaerobic conditions of growth. Perhaps the most obvious method is to bubble gas of known composition through the culture at measured flow rates. Although this would appear to be the most reproducible method, this is probably not the case with Tetrahymena, since not only is it difficult to control the size of the bubbles (which in turn determines the rate of solubilization of oxygen into the medium), but also the intensity of bubbling required to approach equilibration with the gas phase is exceedingly high, leading to frothing and cell damage.
Malecki et al. (18) flowed gas mixtures of known composition over the surface of very shallow shaken cultures and obtained highly reproducible growth curves, but found that at low oxygen tensions growth was markedly inhibited by the process of shaking, thus obviating the use of this method for the present purposes. Levy and Scherbaum (3), who have studied the rate of decline of oxygen tension during the growth of deep cultures, showed that flushing air over the surface of such cultures, even when they were in a shaker bath, did not increase the very low oxygen tension once the cell density was over 100,000 cells per ml. We have, therefore, chosen to use air as the gas phase and to vary only the depth of the liquid in the flasks (which were covered with Morton closures to ensure adequate entry of air into the flasks). Since the flasks were grown in the same shaker bath, 5.0 z!z 0.6 (4) 5.3 f 0.6 (4) 4.4 z!z 0.5 (4) nzn l7g 1.11/106 cells/hr 68 zt 12 (4) 84 (2) 120 (2) 12 It 5 (3) 96 (1) 140 (2)  11 + 4 (4) 112 (2) 176 (2) 02 consumption -Substrate + Substrate the only difference between shallow, medium, and deep cultures was the surface to volume ratio and hence the average amounts of oxygen and of carbon dioxide dissolved per ml. We have found this method to give highly reproducible results, as will be seen below.
In what follows, we shall assume that the metabolic changes which occur are due to changes in bhe average oxygen content of the cultures during the 17.hour incubation period, since this is certainly of greater influence than the changes in COZ content.
It must, however, be borne in mind that COz is fixed by Tetruhymena

Content and Oxygen
Tension, and on Rate of Oxygen Consumption-l'able I shows the results of measurements of cell growth, final oxygen tension, and the rate of oxygen consumption from cultures grown in total volumes of 40, 80, and 120 ml in 500-ml flasks, as described under "Materials and Methods." It can be seen that increasing culture depth from 40 to 80 ml had no effect on total cell growth, but a further increase to 120 ml resulted in a small but reproducible inhibit.ion of growth. It was found that cells grown in deep cultures grew at the same rate as those in shallower cultures but reached stationary phase earlier.
Total glycogen content increased as expected with increasing depth, but was quite variable from experiment to experiment (data not shown), ranging from 160 to 280 pg/106 cells in 40.ml cultures.
Values were 1.1 to 1.3 times higher in 80.ml cultures and 1.3 to 1.6 times higher in 120-ml cultures.
Final poz levels were measured in separate cultures with comparable Nf values to those on which the other measurements were made. It is noteworthy that even with only 40 ml in a shaken 500-ml flask the final poz is considerably less than that of a similar solution in equilibrium with room air.
Increasing the culture volume to 80 ml decreased the final poz to about 12 mm Hg, which was not statistically different than the final ~0% of cultures of 120 ml. Since as will be shown below, there arc marked differences in metabolism between medium and deep cultures, the time course of prior exposure to low oxygen tension must play a major role in determining the metabolic state of the cult.ure at the end of the 17.hour incubation period.
Oxygen consumption experiments carried out in a Warburg apparatus showed that cells grown in deep culture had a reproducibly greater rate of oxygen consumption than those grown in shallow cultures. Since the Qo, measurements were made with a high rate of shaking and a total of 2 ml in 15.ml capacity Warburg flask, these measurements were made essentially under "shallow" conditions. Thus, the difference in Qo, between cells from deep and shallow cultures presumably reflects an adaptation which lasts at least for 1 hour after transfer from deep to shallow conditions, although it is possible that the change in QoZ also reflects a change in composition of the growth medium and the availability of oxidizable substrates during the 17.hour growth period.
That the latter effect is not very important for the present experiments is suggested by the observation that although addition of a mixture of acetate, pyruvate, and octanoate to thewarburg flasks increased the Qo, of all the cells, a clear difference remained in the oxidative capacity of cells grown under shallow, medium, and deep conditions. Thus, growth under conditions of decreasing oxygen tension causes an increase in oxidative capacity which lasts for at least 1 hour after transfer to conditions of high oxygen tension.
Eflect Glyoxylate-Because of the apparent inverse relationship of oxygen tension to Qo, and total glycogen content, it was of interest to determine whether these increases occurred at the level of acetyl-CoA fluxes or by the increased incorporation of glucosyl moieties into glycogen.
According to the metabolic scheme shown in Fig. 1 into glycogen should represent the peroxisomal contribution to glycogen synthesis.
From Table II, it can be seen that increasing culture depth (and therefore decreasing the time average oxygen tension of the culture during the li-hour growth period) increased the rate at which [2-'4C]pyruvate and [2-14C]glyoxylate were oxidized during the l-hour assay, which was performed at high surface to volume ratios and with shaking.
The rate of oxidation of [l-14C]glucose, however, was decreased, while there was little change in the incorporation of glucose into glycogen. The incorporation of [2-%]pyruvate and [2-14C]glyoxylate into glycogen increased almost 2-fold and 4-fold, respectively, in deep as compared to shallow cultures.
These results indicate that cultures grown under low surface to volume ratios develop a greater capacity to oxidize 2-and 3-carbon units and a reduced capacity to oxidize glucose when tested under shallow conditions. It is noteworthy that incorporation of label from glyoxylate into glycogen was increased more than that from pyruvate, indicating that the peroxisomal contribution to glyconeogenesis was more sensitive to changes in average oxygen tension than the mitochondrial contribution, a deduction which is sustained by the quantitative analysis presented below.  These results are consistent with the findings (cf. Table I) that oxygen consumption increased in cells grown under partial anaerobiosis and that oxidation of glyoxylate and pyruvate was increased under these conditions (cj .  Table II).

MITOCHONDRIA
There was a small but significant increase in incorporation of label from [PC]acetate into lipid but no significant changes occurred in the incorporation of label from pyruvate or octanoate into lipid. In our original model (8), it was implicitly assumed that all the label derived from [2-r4C]pyruvate that appears in lipid was in the fatty acid moieties of the lipids. Since more of the label from pyruvate appears in glycogen than in lipids, some of the label in the lipid could be in the glycerol moiety.
It was therefore necessary to measure the fraction of label from [2-14C]pyruvate incorporated into the glycerol and fatty acid moieties of the lipids under the conditions of these experiments. It was found ( Table IV)  are also consistent with the three-compartment model (see Fig. 1) since one would expect [1-r4C]acetyl-CoA derived from precursors such as pyruvate entering the outer mitochondrial pool to appear in the glycerol moiety of the lipids to a greater relative extent than acetate which is oxidized in the inner mitochondrial pool and mostly converted into fatty acids in the peroxisomal pool. Because the total amount of label incorporated into lipids from [2-14C]pyruvate in these three-substrate experiments is very small, we have not deemed it worthwhile to explicitly show in the model that about one-third of the label that enters into the lipids from [2-14C]pyruvate is in the glycerol moiety.
Incorporation of label from [l-14C]acetate and from [2-r4C]pyruvate into glycogen was tripled and doubled, respectively, in deep as compared to shallow cultures, and there was a small but statistically insignificant increase in the incorporation of label from [1-14C]octanoate as well.
As noted earlier (8), appearance of label into glutamate was measurable only when [PC]acetate was the labeled substrate. It was found that there was no change in label incorporation from acetate into glutamate in medium versus shallow cultures, but increasing culture depth from 86 to 129 ml halved the amount of label incorporated into glutamate. Incorporation of label into aspartate was not detected under any of these conditions. Table V shows that there was no significant change in the activity of aspartate transaminase or glutamate dehydrogenase in shallow versus deep cultures.
Thus the halving of label incorporation from [l-r4C]acetate into glutamate cannot be explained by changes of either of these activities as measured in vitro and changes in metabolic controls in the mitochondria must be postulated to account for these observations.
In shallow cultures, no label appeared in alanine from any of the labeled substrates.
With increasing culture depth, label from [2-"Clpyruvate appeared in increasing amounts in alanine (see Fig. 1). Since alanine transaminase is localized exclusively   Table III, computer fits to the data were obtained as described in detail by Raugi et al. (8). In each case, the fits obtained mere essentially as good as those described earlier, and it is therefore unnecessary to present examples of these fits. Fig. 1 presents the average of the individual computer fits obtained for the experiments summarized in Table III. Table VI shows the average values for the dist,ribution parameters in Fig. 1.
It can be seen that the inputs of acetate and pyruvate (V, and VT, respectively) increased with each increment of culture depth, but total octanoate input (V,) was relatively independent of culture depth.
Total 14COp production (VJ increased 56%, but ['Qlutamatc output (V,) was decreased 36yc',, indicating that control mechanisms other than simple pool dilution must be operative in the inner mitochondrial compartment.
Lipogenesis (V,) increased slightly with culture depth, the bulk of the increase being derived from peroxisomal acetyl-CoA (V, -(1 -t)V,) rather than from the direct pathway of incorporation of octanoate into lipids.
As the average oxygen tension decreased, glyconeogenesis increased from both the peroxisomal and the outer mitochondrial compartments, with the contribution from the peroxisomal compartment becoming more dominant in deep than in shallow cultures.
This shift is expressed in the parameter y, which changed from 0.25 to 0.188 to 0.165 w&h increasing culture depth.
In contrast to the marked effects on glyconeogenesis, there was no consistent pattern for the effect of reduced oxygen tension on the p oxidation of octanoate. Although the distribution of acetate between the mitochondrial and peroxisomal cornpart.ments (governed by Q) did not change with culture depth, the amount of acetate entering each compartment increased with decreasing oxygen tension via the increase in Vg. Almost all of the increase in lipogenesis and glycogenesis from the peroxisomes is accounted for by the increase in input of acetate to the peroxisomes. About 237, of the increase in 14C02 production from the inner mitochondrial compartment (VJ, however, was due to the increased flux of labeled acetyl-CoA from the outer mitochondrial compartment (V,,). Virtually all of the increase in VI4 is attributable to an increase in VT, reflecting an increase in the activity of pyruvate dehydrogenase.
No significant changes in any of the other interpool fluxes (V,, VlO, V,,) were observed, but, as discussed earlier (8) tion of culture depth, especially since Rooney and Eiler (13) reported that exposure of Tetrahymena to seven hypoxic shocks caused an increase in respiration rate (as observed here in response to continuous growth at low oxygen tension) and a 50% reduction in ATP content.
Although Rooney and Eiler used a very sensitive procedure which did not require centrifugation of t'he cells, their values for AT1 (about 2 to 4 nmol/106 cells) were considerably below the values we have earlier reported (9, 10) with a fluorimetric assay in which the cells had to be centrifuged prior to assay in order to obtain enough ATP.
Because of the ATP:ADP ratios obtained, it was possible that appreciable conversion of ATP to ADP had occurred during the centrifugation step, and it was decided to reinvestigate the question of ATP:ADP levels with the use of the sensitive procedure essent,ially as described by Rooney and Eiler (13), except that the measurements were made immediately after the perchloric acid was neutralized, since we found that about half of the ATP in standards was lost if the neutralized perchloric acid extract was allowed to stand in ice for 4 hours. With this method, the values we obtain for the ATI' + ADP content of Tetrahymena are comparable to the values reported earlier by us (9, 10) but the ATP:ADP ratios are higher than previously reported and comparable to those usually obtained in mammalian tissues. There was no difference in the ATP:ADP ratio between cultures grown overnight under shallow, medium, or deep conditions (Table VII).
We have repeated the measurement of ATP and ADP levels on cells grown under shallow conditions with either tolbutamide or AMP and find ( have reported that the addition of AMP to the culture medium of cells grown without shaking results in an inhibition of growth and an increase in glycogen content (23), whereas when shallow cultures were grown with shaking and then assayed with radioactive substrates to deduce the intracellular flux patterns, it was found that AMP caused an inhibition of glyconeogenesis (10). It was therefore of interest to perform a set of experiments on deep cultures in which AMP was present during the 17-hour growth period (3.0 mM) and during the l-hour assay. The experimental data are shown in Table III and the intracellular   1   . The increase in y shows that the contribution of the peroxisomal acetyl-Cob pool to glyconeogenesis was more strongly inhibited by AMP than was the contribution of the outer mitochondrial pool. Thus, the peroxisomal contribution to glyconeogenesis is more sensitive to inhibition by AMP and to activation by low oxygen tension than is the mitochondrial contribution. AMP had little if any effect on the interpool fluxes, on alanine production from pyruvate, on label incorporation into glutamate, or on t.otal lipid synthesis.

Time Course of Adaptation to
Step Change in Oxygen Tension-Although it is relatively well established that oxygen tension plays a major role in regulating the metabolism of Tetrahymena (5), relatively little work has been done on the time course of the metabolic changes that occur in response to a change in the level of aeration.
Furthermore we have shown that for at least 1 hour after a sudden change in oxygen tension the cells are metabolically quite stable when tested under highly aerobic conditions, irrespective of their prior growth history.
It was therefore of interest to determine the temporal sequence of adaptation to a step change in oxygen tension and in particular what quantitative contribution to the net effect was made by each metabolic compartment.
It should be noted that in the course of these experiments, the steady state assumption is sometimes violated.
However, because the time required for complete adaptation to occur is considerably longer than the period during which the measurements are made (4 to 8 hours versus 1 hour), it seemed that this type of analysis might nevertheless be suitable.
That this assumption is warranted is partially borne out by the extremely close fit to the observed data obtained by our computer model. Nonetheless, where we have represented a particular metabolic rate as a point in time, we mean the average metabolic flux determined in a l-hour period beginning at that point.
In what follows formal statistical evaluation of the results will not be presented because of the limited number of measurements made at some of the time points.
If more experiments had been done, it might have been possible to draw more conclusions about the minor changes in certain fluxes but little would be added to the over-all picture which emerges. We have ATP + ADP ?%m01/10~ cells therefore confined our statements to those measurements in which appreciable changes were apparent between control and experimental cells. In the legends to Figs. 2 and 4, we have presented information relevant to the reproducibility of these measurements.
Adaptation to Wypoxia-The values obtained for the incorporation of label into COz, glycogcn, glutamate, and lipid in a l-hour incubation and for control cells maintained under shallow conditions arc shown in Fig. 2. Incorporation of label from [l-14C]acetate into CO2 increased slightly within 2 hours after transfer to hypoxic conditions and incrcascd steadily thereafter (Fig. 2G). The increase in acetate conversion to glycogen, however, was already noticeable within 1 hour after transfer to hypoxic conditions and the fully adapted level was achieved within 4 hours (Fig. 2Z)). Incorporation of label from [2-Y]pyruvate into both glycogen (Fig. 2E) and CO2 (Fig. 2H) increased with a similar time course, rapidly at first and then steadily.
The significance of the changes during the first few hours are uncertain, however, because of increases in the control cultures which apparently occur at this stage of growth.
There was no significant change in the ratio of incorporation of label from acetate into lipids until 4 hours after transfer to hypoxic conditions when an increase in lipogenesis relative to the shallow (aerobic) controls became apparent ( Fig. 2A).
The amount of label incorporated into lipid from pyruvate is too small for reliable deductions to be made concerning any changes with time, and this information is not presented in Fig. 2, although it was used to obtain the computer fits which serve as the basis for the metabolic parameters presented in Fig. 3. The rate of label incorporation from octanoate into lipid increased in the control and possibly in the adapting cultures up to about 4 hours, but was the same in both the well aerated and hypoxic cultures at the end of the 8-hour period (Fig. 2C). The rate of appearance of label from acetate into glutamate increased markedly as a funct,ion of culture age, a trend which was either stopped or even slightly reversed during the first 4 hours after transfer to hypoxic conditions. The points shown in Fig. 2 are the average values obtained from two to three experiments except that, only one experiment was done for l-hour adaptation.
1%~ means of these average values, single computer fits were obtained to the metabolic model described earlier (see Fig. 1). Fig. 3  The mean values shown were used to generate the computer fits presented in Fig. 3  3. Changes in metabolic parameters during adaptation to hypoxic growth conditions. The distribution parameters (Y, p, y, and e are unitless.
All other values are in nanomoles per lo6 cells per hour. For further details, see legend to Fig. 2. aging under shallow conditions or of 8 hours of aging plus adapta-4 hours, and then increased at the same rate as it increased in tion to the transfer from shallow to deep culture condit,ions.
control cultures (Fig. 3B). CO2 production from the Krebs cycle in the inner mitochon- The glyconeogenic flux of labeled acetyl-CoA from the outer drial compartment (VI) did not change in control cells, but mitochondrial (yVJ and peroxisomal ((1 -r)V,) pools was increased steadily in the cells adapting to hypoxia, going from a unchanged in control cells (Fig. 3C). The mitochondrial convalue of about 600 nmol/106 cells per hour at the start of the tribution to glyconeogenesis increased from about 3 nmol/106 experiment to about 900 after 8 hours of hypoxia (Fig. 3A). cells per hour to about 8 nmol/106 cells per hour between the Glutamate formation (V,), which also occurs from the inner second and eighth hours after transfer from shallow to deep mitochondrial compartment, was reduced almost immediately conditions. The peroxisomal contribution, however, increased after transfer to hypoxic conditions, remained constant for about from about 14 nmol/l06 cells per hour to about 28 nmol/106 cells per hour within the first 4 hours of adaptation to hypoxia and remained at this level for the next 4 hours. The increase in total rate of glyconeogenesis (VJ was thus largely achieved within the first 4 hours after transfer to deep conditions. Because the rates of glyconeogenesis from both pools increased, there was relatively little change in y, the fraction of the glyconeogenic flux derived from the mitochondria (Fig. 31). The rate of /3 oxidation of octanoate (Fig. 30) in the inner mitochondrial compartment (@VJ declined from about 10 nmol/106 cells per hour to about 4 nmol/106 cells per hour in both the control cultures and in those adapting to the transfer to low oxygen tension, but there was only a slight decline in the rate of p oxidation in the peroxisomal compartment (~(1 -/3)V,). Thus e, the fraction of octanoate P-oxidized, declines more or less steadily through the adaptation period (Fig. 31).
The total rate of acetate utilization by the peroxisomes ((1 -a)VJ was almost constant with time in both the control and adapting cultures, but there was a steady rise in the rate of acetate utilization by the mitochondria (~VG) by the adapting cells (Fig. 3E). As can be seen by comparing Fig. 3, E and H, most of the rise in mitochondrial acetate utilization is due to an increase in Vs rather than a change in the distribution of acetate ((Y) between the mitochondrial and peroxisomal compartments. Pyruvate utilization (VT) and the flux of acetyl-CoA from the outer t,o inner mitochondrial compartment (VI,) increased during the first 4 hours in both the control cultures and those adapting to low 02 but then declined in the controls and continued to increase in the cells adapting to low 02 (Fig. 3G).
Change in Metabolic I%LX Pattern during Deadaptation jrom Hypoxia-The experiments so far described characterize the temporal sequence of changes in the flux rates of labeled acetyl-CoA in response to a sudden transition from relatively high oxygen tension to partial anaerobiosis.
It was of interest also to examine the temporal sequence of metabolic changes in response to an abrupt increase in oxygen tension, a process we shall refer to for brevity as deadapt'ation.
The measured values for incorporation of label into CO*, glycogen, lipids, and glutamate during a l-hour incubation with labeled precursors at various times during the deadaptation process are presented, along with t,he values obtained for control cultures maintained under deep conditions in Fig. 4. The values shown are the average values from t,wo to three experiments except that only one experiment was done at zero hours.
There was little change in the oxidation of [I-Wlacetate during the first 4 hours of deadaptation.
After 4 hours, there was a marked reduction in oxidative capacity (Fig. 4G). The rate of oxidation of [2-Wlpyruvate during the first 4 hours of deadaptation was the same as the control, but it is clear (Fig. 4H) that there was an increase in the capacity of cells maintained under hypoxic conditions to oxidize pyruvate whereas the capacity of the deadapting cells decreased considerably by the end of the 8-hour period.
There was no change during 8 hours of aging in the rate of glyconeogenesis from acetate in control cells (Fig. 4D), but the rate of glyconeogenesis began declining rapidly immediately upon transfer of the cells grown in deep cultures to shallow conditions. Glyconeogenesis from [2-Ylpyruvate increased from about 9 to 13 nmol/106 cells per hour in control cultures aging under hypoxic conditions for 8 hours; deadapting cells paralleled the controls during the first 4 hours after transfer to well aerated conditions, then declined markedly in the subsequent 4 hours.
The incorporation of label from acetate into glutamate in-creased markedly after 2 hours of deadaptation (Fig. 4B), while there was no significant change in label incorporation into glutamate in the control cultures.
The experimental data shown in Fig. 4 were analyzed according to the three-pool model of metabolism as described above and the parameters obtained are displayed in Fig. 5. Only the salient features of these graphs will be described.
There was little change in V1, the rate of COz production, from the Krebs cycle for 4 hours in the control and deadapting cells, but a fall in oxidative capacity was seen between 4 and 8 hours in the latter (Fig. 5A). Glutamate production, VZ, stayed relatively constant in control cells but increased steadily after 2 hours under high oxygen tension (Fig. 5B). There was little change in glycogen synthesis from the outer mitochondrial pool (TV,) in either control or deadapting cells, or in the peroxisomal contribution ((1 -y)V,) to glyconeogenesis in the control cells aging under hypoxic conditions, but a steep steady decline in ((1 -y)Vo) from the beginning of the deadaptation process ( Fig.  5C) was observed.
Total acet,ate utilization (V,) and acetate uptake into the inner mitochondrial compartment ((YV,) followed the same temporal pattern as did VI (cf. Fig. 5E with Fig. 5A), since (Y, the parameter governing distribution of acetate between mitochondria and peroxisomes, did not change (Fig. 5H). The fraction of octanoate P-oxidized in the mitochondria, & increased during aging of the control cultures but decreased during the first 4 hours of deadaptation and then stayed constant (Fig.  5H). The fraction of total glyconeogenic flux contributed by the mitochondria, y, increased steadily with time of deadaptation but did not change with time of aging in the control cells (Fig.  51).

DISCUSSION
It was suggested by Raugi et al. (8) that the quantitative discrepancies between their results and those of Connett and Blum (2) were due to differences in growth conditions, and, specifically, in the difference in oxygen tension between the deep cultures used by Connett and Hum and the shallow ones used by Raugi et al. In the present experiments, cultures were grown under identical conditions of shakicg, temperature, and initial cell density but with varying surface to volume ratios. The flux parameters measured in this study reflect the capacity of the cells to perform some metabolic activity and, as discussed earlier, are not necessarily indicative of the flux of unlabeled metabolites derived from the proteose-peptone medium. If growth for 17 hours at various culture depths left, no persistent changes in the cells, then, since all the assays with the radioactive substrates were performed under highly aerobic conditions, one should have detected only minimal changes between the cells grown in deep versus shallow cultures.
Such changes could result from alterations in the composition of t,he culture medium caused by different metabolic patterns during growth under conditions of varying oxygen tension.
That this is probably not a major factor is indicated by Levy's finding (24) that the changed enzyme content which occurs during growth under static conditions reverts to that typical of cells grown under shaken conditions in the absence of cell division, and by our finding that the metabolic flux rates revert to those of cells grown in shallow cultures within a few hours.
The 30% increase in the capacity of Z'etrahymena to consume oxygen after growth under deep conditions parallels the increase in appearance of label from [2-f4C]pyruvate or [2-14C]glyoxylate in CO2 as well as the increase in WO2 production from labeled acetate and pyruvate in the presence of these two substrates ii-l-u-  Levy and Wasmuth (25) found that the specific activities ,of several enzymes, including isocitrate lyase and succinic dehydrogenase, increased in Tetrahymena grown under static uersus shaken conditions which accords well with our findings of increased oxidative capacity in cells grown under relatively anaerobic conditions. An adaptive increase in oxidative capacity in response to hypoxia is not limited to Tetruhymena. In rats, for example, hypoxic acclimatization caused increases in succinic dehydrogenase activity in mitochondria from heart and kidney of about 41y0 and from liver of about 135%, on a per mg of mitochondrial protein basis (26). Recent evidence (27) confirms the increase in succinic dehydrogenase activity and suggests that it may reflect activation by an effector rather than increased enzyme content.
In contrast to the increase in capacity to oxidize glyoxylate, acetate, and pyruvate, the rate of oxidation of [l-14C]glucose to i4C02 was decreased in cells grown in deep cultures.
The finding by Levy and Wasmuth (25) that pyruvate kinase increased in static cultures renders it unlikely that glucose oxidation was limited by this reaction, and the reduction in glucose oxidation capacity caused by growth under partially anaerobic conditions is probably due to a reduction in the enzyme activity between the hexokinase and pyruvate kinase steps. Since phosphofructokinase, localized largely on the mitochondria of Telrahymena (28), appears to be the rate-limiting enzyme of the glycolytic sequence (22), it is likely that partial anaerobiosis reduces the activity of this enzyme in situ despite the fact that no change was observable when this enzyme was assayed in vifro from cells grown in deep or shallow cultures.
The effect of low oxygen tension during growth in causing a decrease in the capacity to oxidize glucose is contrary to the well known Pasteur effect described in Tetrahymena (29). Under conditions of complete anaerobiosis, Tetrahymena rapidly deplete their glycogen reserves (29), indicating a marked acceleration of the glycolytic pathway, whereas under the present' conditions of partial anaerobiosis there was an increase in total glycogen content as compared to lvell aerated cultures.
This suggests that at lower oxygen tensions than were studied hrre there is a switch from net glyconeogenesis to net glycolysis.
In all systems so far studied the Pasteur effect seems to be mediated via the regulatory effects of adenosine nucleotides on phosphofructokinasc (22). The failure of the ATP:ADP ratio to change is thus not inconsistent with the lack of a Pasteur effect under these conditions of growth.
It should be noted, however, that the measurements of adcnosine di-and triphosphate were done on the original culture in contrast to measurements of label incorporation which were done after transfer to a highly aerobic environment.
The increased capacity to oxidize 2-and 3-carbon atom moieties after growth in deep cultures was reflect,ed in an increase in Vg, the total acetate flux measured after the addition of the three substrates.
Since VS wa s increased in each compartment, it would appear that either both the pcroxisomal and the mitochondrial acetyKoA symhetases (30) increased or there was a greater availability of CoA in both the inner mitochondrial and peroxisomal compartments. Levy and Scherbaum (31), however, showed that there was no difference in total acetyl-Co-4 synthetase in cells exposed to static conditions for 3 hours, so the former explanation is unlikely. Total lipogenesis, V8, also increased slightly in deep as compared to shallow cultures. There was no significant change in the pathway for direct incorporation of octanoate into lipids.
A key effect of growth under reduced oxygen tension was the large increase in flux of acetyl-Cob directed to glyconeogtncsis, both from the outer mitochondrial compartment and from the peroxisomal compartment.
Although both isocitrate lyase and malate synthetase increase upon transfer of cells from shaken to static conditions (31), the increase in glyconeogcnesis can occur even when the increase in activity of these two enzymes is prevented (4, 32). The increase in the cytosol form of phosphoenolpyruvate carboxykinase (6, 33) observed in static versus shaken cultures of Tetruhymenu may not be essential either since the lowest measured activity of this enzyme under either condition is much larger than either the rate of glyconeogcnesis (see Fig. 1) or the activity of either of the glyoxylate cycle enzymes. Voichick et al. (34) recently found that the total 3':5'-monophosphate (cyclic AMP) content of a culture of Tetruhymenu decreases during the induction of the prestationary phase in standing cultures.
It seems highly probable, then, that cyclic AMP* levels are lower in deep as compared to shallow cultures, * The abbreviation used is: cyclic AMP, cyclic adenosine 3':s' monophosphate. and this may play a major role in controlling the flow of metabolites along the glyconeogenic pathway.
It is unlikely, however, that cyclic AMP is involved in all the changes reported here, and further experimentation on the role of cyclic AMP in the regulation of Tetrahymenu metabolism is obviously necessary.
Following a report (23) that 5'.-1X11' increases the glycogen content of cultures of Tetruhymenu grown without shaking, we investigated the effect of 5'.AXP on several aspects of metabolism (10). Because the experiments with t,he labeled substrates were done on shallow cultures, and the assays also performed under relatively aerobic conditions where glyconeogcnesis is small, it was of interest to determine the effect of AMP on cells grown under hypoxic conditions.
The finding that the effect of AMP was essentially the same in deep as in shallow cultures indicates that the inhibition of glyconeogcnesis by AMP does not depend on oxygrrl tension. The observation that treatment with AMP reduces Vi and V. while increasing V8 is important in view of the fact that decreasing oxygen tension increased all three. This means that the mechanism of action of 5'.AMP is different than that for oxygen.
These experiments show that reductions in the average oxygen tension during growth cause changes in the pattern of flow of acetyl-CoA along the pathways of intermediate metabolism which last for at least 1 hour after the cells arc transferred to aerobic conditions.
In particular, peroxisomal function in Tetruhymenu appears to be very sensit,ive to alterations in oxygen tension, and in the range of oxygen tension studied here the changes in glyconcogencsis and in glycolysis observed appear to be unrelated to the Pasteur effect which, presumably, comes into play under even more anaerobic conditions.
Changes of Metabolism with Culfure Age-Cells grown for 17 hours with shaking attain cell densities in excess of 500,000 cells per ml whether grown under "shallo\v" or "deep" conditions. There is, of course, nothing special about 17 hours, a number chosen entirely for our convenicncc.
It is t)o bc cxpccted, therefore, that even in cells maintained for 8 additional hours with no change in culture depth, metabolic changts will occur as the composition of the medium changes and as the cells make the transition from the logarithmic phase of growth to the st,ationary phase. The changes that, occur with age in deep cultures when assayed with a mixture of labeled acetate, glutamate, aspartate, and alanine have been described earlier (2). During the &hour interval studied here, which corrcsponds to the t.imc between the transit,ional cells and early stationary cells studied by Connctt and Illum (a), there was a decline in glycogen, and probably in the oxidation of the octanoate as ~11. The oxidation of pyruvatc and glyconcogenesis from pyruvate, however, increased with age. Thus, in tells maintained under low oxygen tension, octanoate utilization in the peroxisomcs decreased while pyruvate utilization in the outer mitochondrial compartment increased during the interval from 17 to 25 hours after the st.art of growth.
1)uring this same interval (i.e. from 17 to 25 hours after the start of growth) cultures maintained under shallow conditions exhibited little change in pyruvatc oxidation, in pyruvatc conversion to glycogcr~, or iii glyconcogcncsis from octanoate. Oxidation of octanoatc, however, dcclincd.
Thus, in contrast to cells maintained under deep conditions, n-here octanoate utilization apparently declined in the pcroxisomes, this decline occurred in the mitochondrial compartment in cells maintained at relatively high oxygen tension.
In shallow cult.ures, incorporation of label from acetate into glutamate increased Cth time during the &hour interval under discussion.
In deep cultures, however, there was little change in the incorporation of label from acetate into glutamate during this interval.
The out,put of glutamate from rat heart mitochondria depends on the balance between citrate xynthetase, aspartate transaminase, and cr-ketoglutaratc dchydrogcnase, and is controlled in part by the A'l'P:ADP ratio and by the NADH :KAD ratio (351. In Tefrahymena, a-ketoglutarate dehydrogcnasc and glutamic dehydrogenase arc also involved in determining the outflow of glutamate, but, except for the observation that the ATP:ADI' ratio does not change as a function of dcgrcc of aeration, nothing is known concerning the regulation of glutamate output,.

Metabolic
Responses fo Changes in Oxygen Tension-In cells adapting to a sudden decrease in oxygen tension, the rate of oxidation in the Krebs cycle, Vr, increases steadily during the next 8 hours (Fig. 38); a transition from low to high oxygen tension, howcvcr, does not cause a decrease in VI until about 4 hours after the dcadaptation process begins (Fig. 5A). Similarly, acetate utilization in the inner mitochondrial compartment (~VG) increases fairly steadily during adaptation to low oxygen (Fig. 3E), whereas CYVB does not begin decreasing until 4 hours after a change from IOK to high oxygen tension (Fig. 5E). Thus, in the mitochondria at least, the temporal changes during adaptation are not the mirror images of those seen during deadaptat,ion. Indeed, because of the differences (described above) which occur during the interval from 17 to 25 hours after inoculation even in cultures maintained at constant depths, one would not expect the temporal pattern of metabolic changes during adaptation to low oxygen to be the same as that during deadaptation from low t,o high oxygen tension.
A similar observation had been made for lactic dehydrogenasc, localized in the mitochondria of Tefrahymena (15), and for pyruvate kinase, localized in the cytosol (33). These two enzymes increase at the same rate following a reduction in oxygen tension but did not decrease for at least 8 hours after shaking of the static cultures was resumed (24). Although it is known that succinic dehydrogenase increases in stat'ic cultures (25), the time course of this change was studied only during intervals of days. Acetyl-CoA symhetasc does not' appear to change in response to a decrease in oxygenation (31). Thus the mechanism for the obscrvcd changes in flow of lab&d acctyl-Coh in the mitochondria are at present obscure, but it is clear that in this compartment the response to a step change in oxygen trnsion happens at different times for different processes and, furthermore, the rate of response is different for a step increase than for a step decrease in oxygen tension.
In response t,o an abrupt decrease in oxygen tension, there is a rapid rise in glyconeogenesis from the peroxisomes ((1 -r)V,) which levels off about 4 hours aft,cr the drop in oxygen level (Fig. 3C); in response to the reverse transition the decline in peroxisomal glyconeogenesis also begins soon after the change in oxygen tension, but, cont,inucs declining throughout the X-hour period (Fig. 5C). Thus in the prroxisomes there is also not a mirror image relation between the time course of change in flow of labrlcd acetyl-CoA into glyconcogencsis in cells adapting to a decrease in oxygen tension as compared to ~11s responding to an increase in oxygen tension.
A doubling in the activities of isocitrate lyasc and malate synthetasc is achieved within 3 hours of a reduction in oxygen tension (4, 31), and a halving of the isocitrate lyase activity n-as observed about 3 hours after an abrupt increase in oxygenation (24). These changes are within the right time scale to account for the rates of change of peroxisomal glyconeogenesis reported here. It must be empha-459 sized, however, the full increase in glyconeogenesis occurred even when the increase in enzyme content was prevented by inhibitors of protein and RNA synthesis (4).
The experimental results presented here are the first studies which permit one to describe the sequence of changes in the flux of labeled acetyl-CoA from each of the three pools known in this cell during adaptation to changes in oxygen tension.
To within the error of the measurements, it can be seen that changes occur in each compartment according to an individual temporal pattern, which in itself depends upon the direction of the shift in oxygen tension. That these changes occur within a range of oxygen tension which alters the rates of glyconeogenesis but does not activate the Pasteur effect suggests that the cytosol, mitochondria, and peroxisomes of this cell are each sensitive to different levels of oxygenation, with peroxisomal function probably being the most responsive to falling oxygen levels. Much further work will be necessary to unravel the complexities of this system.

Aclinowledgments-We
are indebted to Dr. J. W. Moore for use of his computer facilities and to Dr. J. Salzano for use of his pa-gas analyzer.