Conditionally Lethal Mutations in Chinese Hamster Cells. Characterization of a Cell Line with a Possible Defect in the Krebs Cycle L. DEFRANCESCO, D. WERNTZ AND I. E. SCHEFFLER Department of Biology, University of California, San Diego, P. 0. Box 109, La Jolla, Calgornia 92037
ABSTRACT A variant Chinese hamster cell line has been isolated from a mutagenized population that has a markedly reduced ability to oxidize a variety of substrates via the Krebs cycle. The production of 14C02from 14C-labeledcompounds was measured using pyruvate, acetate, p-hydroxybutyrate, palmitate and glutamate, and in all cases it was negligible in the mutant. In contrast to this, significant amounts of I4CO2 were produced from l4C-aspartate and 14Csuccinate which suggests that some reactions of the Krebs cycle can take place and this conclusion is supported by tracer experiments with labeled compounds. The rate of respiration measured with a Clark oxygen electrode in the mutant was compared to several normal Chinese hamster cell lines and was found to be only 8 % . Mitochondria appear to be present in normal numbers and with only minor differences in morphology. The measurement of difference spectra between oxidized and reduced states permits us to conclude that the cytochromes are all present and functional. These results lead us to believe that there may be a defect in the Krebs cycle between a-ketoglutarate and succinate. Alternatively a defect in a structural component of the mitochondria or in the electrontransport chain itself may be causing pleiotropic effects in the Krebs cycle and respiration.
In a previous communication (Scheffler, ’74) we have described a mutant or variant Chinese hamster cell line which is dependent on carbon dioxide for growth and is unable to metabolize glucose and pyruvate completely to carbon dioxide. A t the time, we hypothesized that the mutant was defective in its ability to channel pyruvate into the Krebs cycle, possibly because of an alteration in the pyruvate dehydrogenase enzyme complex. The role of carbon dioxide was not obvious and we had shown that it was not needed €or the biosynthesis of purines and pyrimidines, the major end products of COz fixation in cultured mammalian cells (Chang et al., ’61). In the following we present evidence that in these cells the Krebs cycle is almost totally inactive and the rate of respiration as measured by oxygen consumption is reduced by at least a factor of ten relative to other Chinese hamster cells in culture. The need for exogenous COZtherefore may arise from the inability of the cells to make sufficient quantities of their own (in the Krebs cycle) but we will also show that J. CELL. PHYSIOL..85: 293-306.
this requirement can at least in part be overcome by the addition of asparagine to the medium. MATERIALS A N D METHODS
The cells and their culture have been described previously (Scheffler, ’74). The “mutant” auxCOz was derived from the parental fibroblast line CCL16. The SV51 cells were obtained by transformation with SV40 virus of a temperature-sensitive cell line of Chinese hamster, BF113 (Scheffler and Buttin, ’73). All radioisotopes were obtained from New England Nuclear; specific activities and concentrations used are given in the text and figure legends. Our method of measuring the radioactive COz produced from C-14 labeled precursors was described in the previous paper (Scheffler, ’74). Measurement of oxygen consumption Cells in log phase of growth were fed with fresh medium and trypsinized the following day. After washing with TD bufReceived Mar. 14,’74.Accepted Sept. 23, ’74.
293
294
L. DEFRANCESCO, D. WERNTZ AND I. E. SCHEFFLER
fer they were resuspended in TD buffer or complete medium at a final concentration of 3 x 106 to 3 x lo7 cellslml. Immediately after the dilution of the cells the amount of oxygen remaining dissolved in solution was measured with a Clark oxygen electrode and the measurement, after amplification, was recorded continuously with the help of a stripchart r e corder. All measurements were made at 37”C, and throughout the measurement the cell suspension was stirred with a magnetic stirrer. Cell counts in this and other experiments were performed with the Coulter counter. The instrument was calibrated by using buffer as medium with no cells to set the pen deflection corresponding to 100% O2 (saturation at 37”C), and the zero % deflection was obtained after the addition of an excess of the powerful reducing agent dithionite to the solution. We assume that at 37°C the concentration of oxygen in solution is 210 Fmolar or 2.8 p4ml (Maron and Prutton, ’59).
with a small amount of [“HI aspartate or glutamate and then applied to a Dowex-lacetate column, on which the two amino acids can be separated following the procedure of Berl (‘61). Aliquots of the eluted fractions were counted in Triton-toluene scintillation fluid.
Electron microscopy The cells were fixed in 2% glutaraldehyde, post-fixed in 2% medium tetroxide, and after dehydration in a graded series of ethanol-water mixtures they were embedded in Epon A for sectioning. Finally a uranyl-lead stain was applied. RESULTS
Oxidation of C-14 labeled precursors to 14C02 We had previously established that 12I T ] pyruvate would not be oxidized to 14COz in the “mutant” and our first goal was to demonstrate whether or not there was any activity in the Krebs cycle. Unfortunately, most if not all of the intermediates in the Krebs cycle are not very Chromatographic identification of effectively taken up into mammalian cells, aspartate and glutamate and hence a very direct test of the oxidaBetween 1 and 2 X lo6 cells were grown tion of such intermediates is not possible up in micro Fernbach flasks and labeled with intact cells. We therefore attempted for two hours at 37°C with either [U-I4C] first to oxidize precursors which can enter aspartate or [U-I4C]-glutamate. The spe- the Krebs cycle via the intermediate acetyl cific activities were 208 mCilmmole and CoA: in addition to pyruvate these include 234 mCi/mmole, respectively, and the la- acetate and short chain or long chain beled precursors were diluted to 1 gCi/ml fatty acids. Results obtained with [ 1-14C] in a total of 2 ml of complete medium lack- acetate, [ U-14C]palmitate, and [ 3-14C]p ing glucose, aspartate, asparagine, gluta- hydroxybutyrate are shown in figure 1. mate and glutamine. The medium was We were unable to detect any significant removed at the end of the incubation pe- evolution of 14C02 from these precursors riod, the cells were washed twice with in the “mutant” cells, while the wild type TD buffer and covered with 2 ml of 5% cells produced easily measurable amounts. perchloric acid (PCA). Radioactive carbon- This result was obtained under a variety dioxide was trapped in NaOH as described. of conditions, in particular, in TD buffer, The cells were then removed by scraping as shown, and in complete medium lackand the cell suspension in PCA was cen- ing glucose, made up with dialysed serum. trifuged to remove the precipitate material. Acetate is not generally considered to be The supernatant was neutralized with a carbon source for mammalian cells, but KOH and the KC104 was allowed to pre- it appears from these results that the wild cipitate out. The extract was placed on a type cells have some capacity for the acDowex 1-C1column (1 X 14 cm) and elu- tivation of acetate to acetyl CoA. tion was carried out with an HCl gradient. Another path of entry into the Krebs Under these conditions glutamate and cycle is the deamination of glutamate to aspartate are eluted together, but sepa- a-ketoglutarate, a reversible step which rated from other intermediates (Van Korff, links the Krebs cycle with amino acid bio’69). The fractions containing the gluta- synthesis. We have performed experiments mate and aspartate were combined, mixed with both [1-I4C]glutamate and [5J4C]
295
A MUTATION AFFECTING THE KREBS CYCLE
.I 8 r
2
.24-
N
0
f
.02
0
E
01
E 2
4
2
4
1
2
HOURS HOURS HOURS Fig. 1 Formation of 14COz from various isotopically labeled precursors. The 14C02 was trapped and measured as described under METHODS; 0 parental cells, CCLl6; “mutant” cells, auxC@. A: 0.5 wCi/ml [ l - W ] acetate (55 mCilmmole) in TD buffer. B: 0.1 pCi/ml [U-14C] palmitic acid (720 mCi/mmole) in TD buffer. C: 0.1 pCi/ml [3-14C]P-hydroxybutyrate (5.05 mCj/mmole) in TD buffer. Absolute amounts of 14COz were calculated assuming the specific activity of the C q to be identical to the specific activity of a single carbon of the substrate; for acetate and a-hydroxybutyrate this is the same as the specific activity of the precursor, but for palmitic acid it is 1/16of the specific activity of the whole molecule. The amount of 14C02was then normalized with respect to 106 cells.
glutamate which are shown in figure 2. [l-l4C]glutamate is expected to yield I4CO2 immediately from the decarboxylation of a-ketoglutarate in a reaction catalysed by a-ketoglutarate dehydrogenase, an enzyme complex very similar to pyruvate dehydrogenase. On the other hand, 14C02 from [5-l4C]glutamate would be expected only after several interconversions in the cycle. As can be seen in figure 2 , easily measurable quantities of I4CO2 are formed in the parental cells, CCL16, while the corresponding activity in the “mutant” is lower by two orders of magnitude. We have also measured the formation of I4C0p from [ U - l T ] and [1-14C]-aspartate, from [ 1,4-I4C]succinate (fig. 3) and from [1-I4C]asparagine (not shown). In this case the amount of 14C02 formed by the “mutant” is not negligible: the rate of oxidation of succinate is similar in “mutant” and parent cells but low, possibly
because of a slow uptake; the relative rate of formation of 14C02 from aspartate has been found to be somewhat lower and variable in the mutant, and the corresponding rates measured with asparagine are two to five-fold lower in both wild type and mutant cells. We speculated initially that some of the 14C02 resulted from the decarboxylation of orotidine monophosphate, but the formation of 14C02 could not be inhibited by the addition of orotic acid, deoxycytidine, or adenine (Ishii and Green, ’73). Furthermore, the production of 14C02 from [ 1-I4C]aspartate was not inhibited by N-(phosphonacety1)-L-aspartate(PALA), a transition state analogue and powerful inhibitor of aspartate transcafbamylase (Collins and Stark, ’71). Although we may also be observing the decarboxylation of oxaloacetate to phosphoenol pyruvate (pathway of gluconeogenesis), or pyruvate, we believe that at least some of the I4CO2 from aspartate and succinate is generated
296
L. DEFRANCESCO, D. WERNIZ AND I. E. SCHEFFLER
from reactions in the Krebs cycle (see below) .
HOURS
HOURS
Fig. 2 Formation of 14Cq from labeled glutamate. A: 0.5 pCi/ml [ 1-14C]DL-glutamate (19.2 mCi/mmole). B: 0.5 pCi/ml [5-14C] DL-glutamate (4.17 mCi/mmole). 5 X lo5 cells had been suspended in TD buffer at the beginning of each ex“mutant” periment: 0 parental cells, CCLIG; cells, auxCOz.
Interconversion of aminoacids via the Krebs cycle The Krebs cycle plays a role not only in respiration, but also in the interconversion of certain aminoacids, and we have attempted to follow the fate of [U-14C]aspartate and [U-I4C]glutamate in the acid soluble pool of the intracellular extracts of “mutant” and wild type cells. As shown in figure 4, we can demonstrate in wild type cells the formation of aspartate from glutamate and vice versa, but in the “mutant” while aspartate can be converted to glutamate, glutamate cannot be converted to aspartate. The latter result was expected, since we had already shown that [1J4C]glutamate does not yield 14C02 in the “mutant”; the first result confirms our earlier conclusion that CO2 is formed from aspartate by a series of reactions of the Krebs cycle. It should be emphasized that these results are qualitative, in the 0.3
B
A
/
CCLIG
0.2
Kxco2 0.I
I
0
1 2 HOURS
3
0
1
2
3
HOURS
Fig. 3 Formation of 14C02 from labeled aspartate (A) and from labeled succinate (B). The cells were incubated in TD buffer containing 0.25 fiCi/ml [U-*4C]aspartate (specific activity 208 mCi/mmole) or 1.0 pCilml [ 1,4-14C]succinate (specific activity 20.37 mCi/mmole) and the 14C02 was trapped and counted as described under METHODS.
297
A MUTATION AFFECTING THE KREBS CYCLE 100-
-c
h
2ooot
0%
90-
OoO
I
looo z 80W
0
a
W
n
40-
a
30-
y>
20-
P
B
B
3000
z
-
::
OOO
OOO
1
%o
o~%oo
60-
I . . .* a Z
0
70-
OOno
50-
000
000
%o
0
0 0
10-
2000-
4 1000-
IL W U
sense that they indicate which reactions may occur and which pathways are possible, but we cannot be too specific about their absolute rates.
0
L
FRACTION NUMBER
Fig. 4 Interconversion of labeled glutamate and aspartate via the Krebs cycle. Details of the experiment are described under METHODS. The figure shows the resolution of the two aminoacids on the second column. The arrow indicates the peak position of the tritiated aminoacid added a s a marker. The first peak represents glutamate, the second represents aspartate. A: Formation of glutamate from aspartate in parental cells. B: Formation of aspartate from glutamate in parental cells. C: Formation of glutamate from aspartate in auxCOz cells. D: No aspartate being formed from glutamate in a u x C a cells.
Consumption of oxygen Another and completely independent measurement of the rates of respiration in “mutant” and wild type cells can be made by determining directly the rate of oxygen consumption. This rate should be related to, but not necessarily an exact measure of the activity of the Krebs cycle. Our experimental procedure is described under METHODS. Figure 5 shows a plot of the amount of oxygen remaining in solution as a function of time in the presence of 2.6 X 10“ celldm1 (CCL16). A linear decrease is observed up to the first inflection point, at which time glucose was added (6 mM) causing a decrease in the rate of O2 consumption (Crabtree effect). At the second inflection point, dinitrophenol, an uncoupler of oxidative phosphorylation, was added (0.05 mM) and, as expected, the rate of oxygen consumption was increased significantly. From such a plot an initial rate can be calculated with reasonable precision, but for poorly understood reasons the measured rates vary some-
L. DEFRANCESCO, D. WERNTZ AND 1. E. SCHEFFLER
T
CCLIG CHO
I
SV51 auxCOp
Fig. 6 Comparison of the rates of O2 consump tion of different Chinese hamster cell lines. An average is plotted of a number of independent measurements; CCLIG: nine measurements; CHO: three measurements; SV51: two measurements; auxCG; five measurements.
what from experiment to experiment. The history of the cells and their treatment immediately prior to their being used may be reflected in this variability. No consistent differences could be established during these short term incubations between cells in physiological buffer (TD) and cells in complete medium (with glucose and serum). The results of several experiments with different Chinese hamster cell lines, including CCL16 and our “mutant” auxCOn are summarized in figure 6. The number of measurements for each cell line is indicated in the figure legend, and the values are plotted, together with an experimental uncertainty. It is apparent that there is no great difference between the parental fibroblasts (CCLlG), the epithelial cells of ovarian origin (CHO), and SV40-transformed fibroblasts (SV51). The latter were typical, transformed cells, selected for their ability to grow on agar, and with a median chromosome number of 34. By contrast, the rate of oxygen con-
sumption in the “mutant” cells was consistently and appreciably lower, of the order of less than 10% of the wild type rate. Such a very low rate was not easily measured with high precision and it could not be increased by the addition of aminoacids, e.g. asparagine. Oxygen consumption in wild type cells was almost completely inhibited by antimycin A (0.16 pM), cyanide (0.16 mM), and stimulated by dinitrophenol(O.03 mM; see above). On the other hand, it was difficult to demonstrate reproducibly any effects of these substances on the rate of 0 2 consumption in the “mutant,” and we are presently attempting to increase sensitivity and reduce instrument noise to settle this question.
A requirement for asparagine If we assume that the Krebs cycle is not functioning in the “mutant” cells, another likely consequence, in addition to the effect on the energy metabolism of these cells, is a deficiency in certain nonessential aminoacids which are normally derived from the Krebs cycle: glutamic acid, glutamine, aspartic acid and asparagine. As discussed by Meister (‘68) among others, (also Patterson, ’72) glutamine is not properly considered a non-essential aminoacid and is therefore present in all tissue culture media. We therefore placed the emphasis in our studies on the potential need for aspartic acid and asparagine. The relevant results are summarized in figure 7. This figure shows a comparison of the growth rates of parental and mutant cells under four different conditions: (1) in the absence of HCO2C02 and asparagine, (2) in the presence of asparagine only145 ~glml),(3) in the presence of HCOdC02 only, (4) in the presence of HCO3C02 and asparagine. While the growth rate of CCLl6 cells is relatively invariant under these conditions, the “mutant” cells clearly behave quite differently: in the absence of both HCO;/C02 and asparagine they die within a day; the presence of either HCO;/C02 or asparagine alone not only prevents cell death but permits a slow rate of proliferation; when both HCO&O2 and asparagine are present, cell proliferation approaches wild type rates.
A MUTATION AFFECTING THE KREBS CYCLE
9r
98-
: Q
299
I
7-
3
*
6-
%
5-
J J
w
2 I-
a -I W Q
4-
3-
1
0
2 3 0 I 2 3 DAYS DAYS Fig. 7 The effect of asparagine and C@/HCOY on the growth of CCLl6 and auxC@ cells. 1 0 5 cells were plated in 5 cm dishes in minimum essential Eagle’s medium with nonessential amino acids in a Con incubator. At time zero (one day later) cells were counted in one plate and the medium was changed in the other plates to minimum essential Eagle’s medium with glutamine, 10% fetal calf serum and 10 mM HEPES buffer, pH 7.5, without NaHCO,. The cells were counted with the Coulter counter at the indicated times and daily media changes were made in the other aliquot plates. 0,0:no additions; plates kept in air-incubator. 0, m: added 45 pgfml asparagine-H,O; air incubator. a,0 : NaHCO, added (2.2 glliter); CG-incuasparagine and NaHC03 added; C4-incubator. bator. 0 , I
.:
A similar experiment with aspartic acid substituted for asparagine was not successful, even at still higher concentrations of aspartate, perhaps because aspartate is less readily taken up by the cells (DISCUSSION).
Preliminary electron microscopic investigation Since the “mutant” cells were clearly very much deficient in respiratory activity we wished to examine whether there were any obvious differences in the number or morphology of mitochondria in CCLl6 and auxCOz cells. In sections obtained at random from embedded cells it appears that no gross differences are detectable (fig. 8). Numerous mitochondria are seen in the “mutant” cells, which at first glance seem normal, although on closer examination the organization of the cristae may be somewhat less regular, and the mitochondria appear somewhat deflated. The effect of inhibitors Since the “mutant” cells seem to be able to get along with a much reduced
rate of respiration, it seemed of interest to test whether these cells showed an increased resistance to inhibitors of respiration. High concentrations of cyanide or antimycin A were lethal to both mutant and wild type cells, but a narrow concentration range could be found where the growth rate of CCL16 cells was markedly depressed, while that of auxCOr cells was less affected. Figure 9 shows some results with antimycin A. Daily media changes were made in establishing these growth curves to counter the potential instability of the drug. DISCUSSION
The results presented in this paper greatly extend some previous work on an apparent CO2 auxotroph which was unable to oxidize pyruvate completely to carbon dioxide. We now show that in this “mutant,” in contrast to wild type cells, acetate, p-hydroxybutyrate, palmitate, and glutamate also cannot be oxidized to COr. It is most unlikely that this is the result of a defect in uptake, since such a diverse group of compounds is involved, and evenly internally produced [2J4C]
300
L. DEFRANCESCO, D. W E R N T Z A N D I. E. SCHEFFLER
Fig. 8 Electron micrographs of thin sections of C C L l 6 and auxCa cells. Sections were made as described under METHODS. A: CCL16. B : auxC02.For a discussion see the text.
A MUTATION AFFECTING THE
Figure 8
KREBS CYCLE
301
302
L. DEFRANCESCO. D. WERN’lZ AND I. E. SCHEFFLER
in collaboration with V. Mufioz and W. Butler) indicate that both “mutant” and wild type parental cells have the normal K complement of cytochromes, which can W 15 be oxidized in the presence of oxygen or 3 reduced either naturally by oxygen starz vation or by a strong reducing agent. These J -I experiments tend to indicate that the elecg 10 tron transport chain is intact. The enzyme W cytochrome oxidase is present in normal L lamounts (unpublished observation). Ada ditional evidence for a functional electron i i 5 K transport chain comes from studies with isolated mitochondria: with succinate as J the substrate oxygen consumption is com: I parable in normal and mutant mitochondria (DeFrancesco and Scheffler, ’75). Our results so far strongly suggest that a block exists in the Krebs cycle between a-ketoglutarate and succinate, i.e., that a-ketoglutarate dehydrogenase or succinyl CoA synthetase are defective. We cannot yet distinguish between a defect in either of these enzymes themselves or an effect on their activity due to a structural alteration in some as yet unidentified mitochondrial enzyme or protein. The evidence can be summarized as follows: (1) I4CO2is obtained from labeled aspartate and from labeled succinate even in the presence of inhibitors of de nouo synthesis of pyrimidines, suggesting that the COn is obtained via a series of reactions of the Krebs cycle from succinate to a-ketoglutarate; (2) radioactive aspartate can be converted to radioactive glutamate, confirming the above conclusions; (3) in the “mutant” cells [ 1-I4C] glutamate does not become decarboxylated at an appreciable rate (via a-ketoglutarate); (4) radioactive glutamate is not converted to radioactive aspartate in the “mutant” by the sequence of reaction, glutamate + a-ketoglutarate + succinate + + --+ oxaloacetate + aspartate. Since the reactions following succinate have been shown to occur, the block has to be placed in one of the two reactions between a-ketoglutarate and succinate; (5) the observation that labeled precursors entering the Krebs cycle via acetyl CoA will not yield l4CO2 is in agreement with this conclusion: because of the stereospecific interaction of citrate with aconitase, the acetyl carbons do not appear as COz during the first turn of the cycle, and 20
v
A MUTATION AFFECTING T H E KREBS CYCLE
the randomization of the carbons occurs only after free succinate is formed (in normal cells). A consequence of such a defect would be a requirement by the mutant cells of either aspartate or asparagine, since it had been shown (Eagle, ’59; Levintow et al., ’57) that a large portion of the aspartate and asparagine used by cells in culture is derived from glutamine rather than from glucose. Growth of our mutant cells is indeed markedly stimulated by the presence of asparagine but, surprisingly, not by aspartate. Although a difference in the rate of uptake of these two aminoacids may account for this result, a true explanation may require a complete understanding of the equilibria between the various species involved in amino group transfers: asparagine and glutamine as donors, glutamate and aspartate as potential donors or acceptors, and oxaloacetate and a-ketoglutarate as acceptors. This very interesting result deserves further study but we feel at this point that it is peripheral to our discussion of a respiration deficient mammalian cell line. Like some other tumor cells (for a review see, for example, Cooney and Handschumacher, ’70), our mutant may have lost the capacity to make asparagine from aspartate, but the proposal of a defect in L-asparagine synthetase alone would probably not account for all the properties of this “mutant.” We also do not believe that asparagine promotes growth by being converted efficiently to a plentiful supply of carbon dioxide, although some COZ may indeed be derived from this aminoacid: (1) 14C02 can be formed from [1-14C]asparagine, but at a rate which is not greater than that observed with [U-14C]aspartate; (2) high concentrations of asparagine in physiological buffer have no significant effect on the rates of formation of l4COz from [2-14C] and [1-14C]pyruvate (unpublished observations) and a condensation of oxaloacetate with acetyl CoA would certainly be an obligatory intermediate step; (3) the addition of asparagine has no effect on the rate of oxygen consumption by the mutant (unpublished observation$. The requirement for COz/HCOs of these cells is another puzzling problem on which some light might have been shed. The need
303
for exogenous CO, may arise primarily because of the inability of these cells to produce enough of their own, by means of the Krebs cycle, in contrast to wild type cells which become self-sufficient as long as uridine and hypoxanthine are added to the medium (Scheffler, ’74). What the COz is needed for is a second question which we may be able to answer in part: The C o n is required for the carboxylation reactions in which oxaloacetate or malate are formed from pyruvate (involving pyruvate carboxylase or the malic enzyme, respectively), and oxaloacetate in this case is mainly the precursor for aspartate and asparagine. Thus, in the presence of HC03/C02 the requirement for asparagine is partly overcome. The explanation is probably not complete, since our results show that the “mutant” cells grow apprsciably faster in the presence of both HCOd CO, and asparagine than in the presence of either component alone. Even higher concentrations of asparagine than those reported in figure 7 did not increase the growth rate, and the effect of HCOJC02 may therefore have to be sought at another level. The question remains unanswered whether COZ is required for some carboxylation reactions or whether COZ has some regulatory effect. For example, the intracellular pH may be to some extent different from the external pH and strongly influenced by the concentration O f HC0:JCOn. In most mammalian cells both glycolysis and respiration are means of producing metabolic energy and the two are controlled coordinately. It is known, from the work of Paul (’65) and others, that a high pH favors glycolysis while a lower pH favors respiration. Since the “mutant” cells have lost some of this flexibility, they may be more sensitive to the intracellular pH. If we are indeed dealing with a mutant then this would be the first reported case of a respirationdeficient mammalian cell line whose phenotype is absolutely stable. And if, as the evidence suggests, a defect can be located in the Krebs cycle, the mutant would constitute a very interesting addition to the existing collection of somatic cell mutants. A number of attempts have been made in the past to inhibit respiration in mam-
304
L. DEFRANCESCO, D. WERNTZ AND I. E. SCHEFFLER
malian cells in culture either by means of inhibitors or by imposing anaerobic conditions (for a discussion and review see Gregg, ’72). From his own experiments and those of others, Gregg (‘72) concludes that mammalian cells in culture cannot be grown anaerobically for prolonged periods of time (2-3 days). A different approach was taken by Naum and Pious (‘71) who attempted to obtain respiratory-deficient human cells analogous to the “petite” mutants of yeast by culturing the cells in the presence of sublethal concentrations of ethidium bromide. Although cytochrome oxidase activity was reduced 75 to 85% and cytochromes a a:{ had almost completely disappeared after a few days in the presence of ethidium bromide, the cells regained these activities following continued culture in drug-free medium. From the data presented it was not clear whether cells continued to divide at the normal rate once the level of activity of cytochrome oxidase had been reduced so drastically, or whether under these conditions the cells merely survived. More recently, other reports have appeared on permanent changes induced in the mitochondria of cells in culture by treatment with ethidium bromide, but no studies on the respiratory activity of these cells were reported (Klietmann, Sat0 and Nass, ’73; Klietmann et al., ’73). We also observe an increased rate of glycolysis in our “mutant” (Scheffler, ’74 and DeFrancesco, unpublished observations) and experiments are in progress to determine whether the increase in glycolysis can make up for a tenfold reduction in oxygen consumption. This large increase is in itself interesting. It could be due to: (1) a lack of feedback inhibition of glycolysis by Krebs cycle intermediates (e.g. citrate on phosphofructokinase); or (2) due to an increase in the specific activity of certain rate limiting enzymes in glycolysis. The latter may be a unique situation in this “mutant” and not possible for most ordinary cells in culture. The present problem also may have some bearing on the question of the relationship between glycolysis and respiration in normal and tumor cells (Gregg, ’72; Racker, ’72) and one may in fact ask
+
whether the observed phenotype is the result of a transformation or the result of a mutation. Insofar as both the parental and “mutant” cells are established cell lines, they may be already “transformed.” The “mutant” cells exhibit none of the more typical properties of tumor cells: (1) aneuploidy; (2) growth to high cell densities; (3) growth in agar or suspension; (4) low serum requirement. Moreover, a cell line obtained by transformation of a Chinese hamster fibroblast with SV40 virus behaves more like wild type cells with respect to its capacity for oxygen consumption (fig. 6). Finally, although the rate of respiration is often lower in tumor cells compared with normal tissue, the differences are significantly smaller than the ten-to-twelvefold difference observed in our case between parental and “mutant” cells. We, therefore, tend to favor the idea of a specific mutation. Our studies will now focus on cell-free systems and isolated mitochondria to p i n point the defect. We are also very interested in the behavior of this phenotype following somatic cell hybridizations with other hamster and human cells, which will become most meaningful if we are successful in identifying the defective enzyme. ACKNOWLEDGMENTS
This research was supported by Grant GM-18835, National Institutes of Health, and in part by a Program Project Grant from the National Science Foundation. L. DeFrancesco is a predoctoral fellow supported by Grant 2T01 GM-00702, National Institutes of Health. We would like to thank Mr. T. Cope for helping with the electron microscopy, Dr. W. Butler for allowing us to use his oxygen electrode and associated equipment and Dr. K . Poff for instructions on its use. We are also grateful to Dr. G . Stark for a generous amoLtnt of N-(phosphonacety1)-L-aspartate. LITERATURE CITED Berl, S . , A. Iztjtha and H.Waelsch 1961 Aminoacid and protein metabolism. IV. Cerebral compartments of glutamic acid metabolism. J. of Neurochem., 7: 186-197. Chang, R. S., H. Liepins and M. Margolish 1961 Carbon dioxide requirement and nucleic acid metabolism of HeLa and Conjunctival cells. Proc. SOC. Exp. Biol. M e d . , 1 0 6 : 149-152.
A MUTATION AFFECTING THE KREBS CYCLE
Collins, K. D., and G. R. Stark 1971 Aspartate transcarbamylase. Interaction with the transition state analogue N(phosphonacety1)-L-aspartate. J. Biol. Chem., 246: 6599-6605. Cooncy, D. A., and R. E. Handschumacher 1970 L-asparaginase and L-asparagine metabolism. Ann. Rev. Pharmacol., 10: 421-440. DeFrancesco, L., and I. E. Scheffler 1975 In preparation. Eagle, H. 1959 Aminoacid metabolism in mammalian cells i n culture. Science, 130: 432437. Gregg, C. T. 1972 In: Growth, Nutrition, and Metabolism of Cells in Culture. Vol. 1 . G. H. Rothblat and V . J. Cristofalo, eds. Academic Press, New York, pp. 83-136. Ishii, K., and H. Green 1973 Lethality of adenosine for cultured mammalian cells by interference with pyrimidine biosynthcsis. J. Cell. Sci., 13: 4 2 9 4 3 9 . Klietmann, W . , K. Kato, B. Gabara, H. Koprowski and N. Sato 1973 Stable alterations i n HeLa cell mitochondria following ethidium bromide treatment. Exp. Cell Res., 78: 47-58. Klietmann, W . , N. Sato and M. M. Nass 1973 Establishment and characterization of ethidium bromide resistance i n SV40-transformed hamster cells. J. Cell. Biology, 58: 11-26. Levintow, L . , H. Eagle and E. A. Piez 1957 The role of glutamine in protein biosynthesis in tissue culture. J . Biol. Chem., 227: 929-941. Maron, S. H.. and C. F. Prutton 1958 In. Prin-
305
ciples of Physical Chemistry. The Macmillan Co., New York, p. 145. Meister, A. 1968 On the synthesis and utilization ofglutamine. Harvey Lect., 63: 139-178. Naum, Y., and D. A. Pious 1971 Reversible inhibition of cytochrome oxidase accumulation in human cells by ethidium bromide. Exp. Cell. Res., 65: 335-339. Patterson, M. K., Jr. 1972 I n : Growth, Nutrition, and Metabolism of Cells in Culture. Vol. 1 . G. H. Rothblat and V. J. Cristofalo, eds. Academic Press, New York, pp. 171-209. Paul, J. 1965 Carbohydrate and energy metabolism. In: Cells and Tissues in Culture. Vol. 1 . E. N. Wittmer, ed. Academic Press, New York, pp. 239-276. Racker, E. 1972 Bioenergetics and the problem of tumor growth. American Scientist, 60: 5-3. Scheffler, I. E. 1974 Conditional lethal mutants of Chinese hamster cells. Mutants requiring exogenous carbon dioxide for growth. J. Cell. Physiol., 43: 219-230. Schemer, I. E., and G. Buttin 1973 Conditionally lethal mutations in Chinese hamster cells. I. Isolation of a temperature-sensitive line and its characterization by cell cycle studies. J. Cell. Physiol., 81: 199-216. Von Karff, R. W . 1969 Purity and stability of pyruvate and a-ketoglutarate. In: Methods in Enzymology. Vol. XII. J. M . Lowenstein, ed. Academic Press, New York, pp, 514-523.
ABSTRACT A variant Chinese hamster cell line has been isolated from a mutagenized population that has a markedly reduced ability to oxidize a variety of substrates via the Krebs cycle. The production of 14C02from 14C-labeledcompounds was measured using pyruvate, acetate, p-hydroxybutyrate, palmitate and glutamate, and in all cases it was negligible in the mutant. In contrast to this, significant amounts of I4CO2 were produced from l4C-aspartate and 14Csuccinate which suggests that some reactions of the Krebs cycle can take place and this conclusion is supported by tracer experiments with labeled compounds. The rate of respiration measured with a Clark oxygen electrode in the mutant was compared to several normal Chinese hamster cell lines and was found to be only 8 % . Mitochondria appear to be present in normal numbers and with only minor differences in morphology. The measurement of difference spectra between oxidized and reduced states permits us to conclude that the cytochromes are all present and functional. These results lead us to believe that there may be a defect in the Krebs cycle between a-ketoglutarate and succinate. Alternatively a defect in a structural component of the mitochondria or in the electrontransport chain itself may be causing pleiotropic effects in the Krebs cycle and respiration.
In a previous communication (Scheffler, ’74) we have described a mutant or variant Chinese hamster cell line which is dependent on carbon dioxide for growth and is unable to metabolize glucose and pyruvate completely to carbon dioxide. A t the time, we hypothesized that the mutant was defective in its ability to channel pyruvate into the Krebs cycle, possibly because of an alteration in the pyruvate dehydrogenase enzyme complex. The role of carbon dioxide was not obvious and we had shown that it was not needed €or the biosynthesis of purines and pyrimidines, the major end products of COz fixation in cultured mammalian cells (Chang et al., ’61). In the following we present evidence that in these cells the Krebs cycle is almost totally inactive and the rate of respiration as measured by oxygen consumption is reduced by at least a factor of ten relative to other Chinese hamster cells in culture. The need for exogenous COZtherefore may arise from the inability of the cells to make sufficient quantities of their own (in the Krebs cycle) but we will also show that J. CELL. PHYSIOL..85: 293-306.
this requirement can at least in part be overcome by the addition of asparagine to the medium. MATERIALS A N D METHODS
The cells and their culture have been described previously (Scheffler, ’74). The “mutant” auxCOz was derived from the parental fibroblast line CCL16. The SV51 cells were obtained by transformation with SV40 virus of a temperature-sensitive cell line of Chinese hamster, BF113 (Scheffler and Buttin, ’73). All radioisotopes were obtained from New England Nuclear; specific activities and concentrations used are given in the text and figure legends. Our method of measuring the radioactive COz produced from C-14 labeled precursors was described in the previous paper (Scheffler, ’74). Measurement of oxygen consumption Cells in log phase of growth were fed with fresh medium and trypsinized the following day. After washing with TD bufReceived Mar. 14,’74.Accepted Sept. 23, ’74.
293
294
L. DEFRANCESCO, D. WERNTZ AND I. E. SCHEFFLER
fer they were resuspended in TD buffer or complete medium at a final concentration of 3 x 106 to 3 x lo7 cellslml. Immediately after the dilution of the cells the amount of oxygen remaining dissolved in solution was measured with a Clark oxygen electrode and the measurement, after amplification, was recorded continuously with the help of a stripchart r e corder. All measurements were made at 37”C, and throughout the measurement the cell suspension was stirred with a magnetic stirrer. Cell counts in this and other experiments were performed with the Coulter counter. The instrument was calibrated by using buffer as medium with no cells to set the pen deflection corresponding to 100% O2 (saturation at 37”C), and the zero % deflection was obtained after the addition of an excess of the powerful reducing agent dithionite to the solution. We assume that at 37°C the concentration of oxygen in solution is 210 Fmolar or 2.8 p4ml (Maron and Prutton, ’59).
with a small amount of [“HI aspartate or glutamate and then applied to a Dowex-lacetate column, on which the two amino acids can be separated following the procedure of Berl (‘61). Aliquots of the eluted fractions were counted in Triton-toluene scintillation fluid.
Electron microscopy The cells were fixed in 2% glutaraldehyde, post-fixed in 2% medium tetroxide, and after dehydration in a graded series of ethanol-water mixtures they were embedded in Epon A for sectioning. Finally a uranyl-lead stain was applied. RESULTS
Oxidation of C-14 labeled precursors to 14C02 We had previously established that 12I T ] pyruvate would not be oxidized to 14COz in the “mutant” and our first goal was to demonstrate whether or not there was any activity in the Krebs cycle. Unfortunately, most if not all of the intermediates in the Krebs cycle are not very Chromatographic identification of effectively taken up into mammalian cells, aspartate and glutamate and hence a very direct test of the oxidaBetween 1 and 2 X lo6 cells were grown tion of such intermediates is not possible up in micro Fernbach flasks and labeled with intact cells. We therefore attempted for two hours at 37°C with either [U-I4C] first to oxidize precursors which can enter aspartate or [U-I4C]-glutamate. The spe- the Krebs cycle via the intermediate acetyl cific activities were 208 mCilmmole and CoA: in addition to pyruvate these include 234 mCi/mmole, respectively, and the la- acetate and short chain or long chain beled precursors were diluted to 1 gCi/ml fatty acids. Results obtained with [ 1-14C] in a total of 2 ml of complete medium lack- acetate, [ U-14C]palmitate, and [ 3-14C]p ing glucose, aspartate, asparagine, gluta- hydroxybutyrate are shown in figure 1. mate and glutamine. The medium was We were unable to detect any significant removed at the end of the incubation pe- evolution of 14C02 from these precursors riod, the cells were washed twice with in the “mutant” cells, while the wild type TD buffer and covered with 2 ml of 5% cells produced easily measurable amounts. perchloric acid (PCA). Radioactive carbon- This result was obtained under a variety dioxide was trapped in NaOH as described. of conditions, in particular, in TD buffer, The cells were then removed by scraping as shown, and in complete medium lackand the cell suspension in PCA was cen- ing glucose, made up with dialysed serum. trifuged to remove the precipitate material. Acetate is not generally considered to be The supernatant was neutralized with a carbon source for mammalian cells, but KOH and the KC104 was allowed to pre- it appears from these results that the wild cipitate out. The extract was placed on a type cells have some capacity for the acDowex 1-C1column (1 X 14 cm) and elu- tivation of acetate to acetyl CoA. tion was carried out with an HCl gradient. Another path of entry into the Krebs Under these conditions glutamate and cycle is the deamination of glutamate to aspartate are eluted together, but sepa- a-ketoglutarate, a reversible step which rated from other intermediates (Van Korff, links the Krebs cycle with amino acid bio’69). The fractions containing the gluta- synthesis. We have performed experiments mate and aspartate were combined, mixed with both [1-I4C]glutamate and [5J4C]
295
A MUTATION AFFECTING THE KREBS CYCLE
.I 8 r
2
.24-
N
0
f
.02
0
E
01
E 2
4
2
4
1
2
HOURS HOURS HOURS Fig. 1 Formation of 14COz from various isotopically labeled precursors. The 14C02 was trapped and measured as described under METHODS; 0 parental cells, CCLl6; “mutant” cells, auxC@. A: 0.5 wCi/ml [ l - W ] acetate (55 mCilmmole) in TD buffer. B: 0.1 pCi/ml [U-14C] palmitic acid (720 mCi/mmole) in TD buffer. C: 0.1 pCi/ml [3-14C]P-hydroxybutyrate (5.05 mCj/mmole) in TD buffer. Absolute amounts of 14COz were calculated assuming the specific activity of the C q to be identical to the specific activity of a single carbon of the substrate; for acetate and a-hydroxybutyrate this is the same as the specific activity of the precursor, but for palmitic acid it is 1/16of the specific activity of the whole molecule. The amount of 14C02was then normalized with respect to 106 cells.
glutamate which are shown in figure 2. [l-l4C]glutamate is expected to yield I4CO2 immediately from the decarboxylation of a-ketoglutarate in a reaction catalysed by a-ketoglutarate dehydrogenase, an enzyme complex very similar to pyruvate dehydrogenase. On the other hand, 14C02 from [5-l4C]glutamate would be expected only after several interconversions in the cycle. As can be seen in figure 2 , easily measurable quantities of I4CO2 are formed in the parental cells, CCL16, while the corresponding activity in the “mutant” is lower by two orders of magnitude. We have also measured the formation of I4C0p from [ U - l T ] and [1-14C]-aspartate, from [ 1,4-I4C]succinate (fig. 3) and from [1-I4C]asparagine (not shown). In this case the amount of 14C02 formed by the “mutant” is not negligible: the rate of oxidation of succinate is similar in “mutant” and parent cells but low, possibly
because of a slow uptake; the relative rate of formation of 14C02 from aspartate has been found to be somewhat lower and variable in the mutant, and the corresponding rates measured with asparagine are two to five-fold lower in both wild type and mutant cells. We speculated initially that some of the 14C02 resulted from the decarboxylation of orotidine monophosphate, but the formation of 14C02 could not be inhibited by the addition of orotic acid, deoxycytidine, or adenine (Ishii and Green, ’73). Furthermore, the production of 14C02 from [ 1-I4C]aspartate was not inhibited by N-(phosphonacety1)-L-aspartate(PALA), a transition state analogue and powerful inhibitor of aspartate transcafbamylase (Collins and Stark, ’71). Although we may also be observing the decarboxylation of oxaloacetate to phosphoenol pyruvate (pathway of gluconeogenesis), or pyruvate, we believe that at least some of the I4CO2 from aspartate and succinate is generated
296
L. DEFRANCESCO, D. WERNIZ AND I. E. SCHEFFLER
from reactions in the Krebs cycle (see below) .
HOURS
HOURS
Fig. 2 Formation of 14Cq from labeled glutamate. A: 0.5 pCi/ml [ 1-14C]DL-glutamate (19.2 mCi/mmole). B: 0.5 pCi/ml [5-14C] DL-glutamate (4.17 mCi/mmole). 5 X lo5 cells had been suspended in TD buffer at the beginning of each ex“mutant” periment: 0 parental cells, CCLIG; cells, auxCOz.
Interconversion of aminoacids via the Krebs cycle The Krebs cycle plays a role not only in respiration, but also in the interconversion of certain aminoacids, and we have attempted to follow the fate of [U-14C]aspartate and [U-I4C]glutamate in the acid soluble pool of the intracellular extracts of “mutant” and wild type cells. As shown in figure 4, we can demonstrate in wild type cells the formation of aspartate from glutamate and vice versa, but in the “mutant” while aspartate can be converted to glutamate, glutamate cannot be converted to aspartate. The latter result was expected, since we had already shown that [1J4C]glutamate does not yield 14C02 in the “mutant”; the first result confirms our earlier conclusion that CO2 is formed from aspartate by a series of reactions of the Krebs cycle. It should be emphasized that these results are qualitative, in the 0.3
B
A
/
CCLIG
0.2
Kxco2 0.I
I
0
1 2 HOURS
3
0
1
2
3
HOURS
Fig. 3 Formation of 14C02 from labeled aspartate (A) and from labeled succinate (B). The cells were incubated in TD buffer containing 0.25 fiCi/ml [U-*4C]aspartate (specific activity 208 mCi/mmole) or 1.0 pCilml [ 1,4-14C]succinate (specific activity 20.37 mCi/mmole) and the 14C02 was trapped and counted as described under METHODS.
297
A MUTATION AFFECTING THE KREBS CYCLE 100-
-c
h
2ooot
0%
90-
OoO
I
looo z 80W
0
a
W
n
40-
a
30-
y>
20-
P
B
B
3000
z
-
::
OOO
OOO
1
%o
o~%oo
60-
I . . .* a Z
0
70-
OOno
50-
000
000
%o
0
0 0
10-
2000-
4 1000-
IL W U
sense that they indicate which reactions may occur and which pathways are possible, but we cannot be too specific about their absolute rates.
0
L
FRACTION NUMBER
Fig. 4 Interconversion of labeled glutamate and aspartate via the Krebs cycle. Details of the experiment are described under METHODS. The figure shows the resolution of the two aminoacids on the second column. The arrow indicates the peak position of the tritiated aminoacid added a s a marker. The first peak represents glutamate, the second represents aspartate. A: Formation of glutamate from aspartate in parental cells. B: Formation of aspartate from glutamate in parental cells. C: Formation of glutamate from aspartate in auxCOz cells. D: No aspartate being formed from glutamate in a u x C a cells.
Consumption of oxygen Another and completely independent measurement of the rates of respiration in “mutant” and wild type cells can be made by determining directly the rate of oxygen consumption. This rate should be related to, but not necessarily an exact measure of the activity of the Krebs cycle. Our experimental procedure is described under METHODS. Figure 5 shows a plot of the amount of oxygen remaining in solution as a function of time in the presence of 2.6 X 10“ celldm1 (CCL16). A linear decrease is observed up to the first inflection point, at which time glucose was added (6 mM) causing a decrease in the rate of O2 consumption (Crabtree effect). At the second inflection point, dinitrophenol, an uncoupler of oxidative phosphorylation, was added (0.05 mM) and, as expected, the rate of oxygen consumption was increased significantly. From such a plot an initial rate can be calculated with reasonable precision, but for poorly understood reasons the measured rates vary some-
L. DEFRANCESCO, D. WERNTZ AND 1. E. SCHEFFLER
T
CCLIG CHO
I
SV51 auxCOp
Fig. 6 Comparison of the rates of O2 consump tion of different Chinese hamster cell lines. An average is plotted of a number of independent measurements; CCLIG: nine measurements; CHO: three measurements; SV51: two measurements; auxCG; five measurements.
what from experiment to experiment. The history of the cells and their treatment immediately prior to their being used may be reflected in this variability. No consistent differences could be established during these short term incubations between cells in physiological buffer (TD) and cells in complete medium (with glucose and serum). The results of several experiments with different Chinese hamster cell lines, including CCL16 and our “mutant” auxCOn are summarized in figure 6. The number of measurements for each cell line is indicated in the figure legend, and the values are plotted, together with an experimental uncertainty. It is apparent that there is no great difference between the parental fibroblasts (CCLlG), the epithelial cells of ovarian origin (CHO), and SV40-transformed fibroblasts (SV51). The latter were typical, transformed cells, selected for their ability to grow on agar, and with a median chromosome number of 34. By contrast, the rate of oxygen con-
sumption in the “mutant” cells was consistently and appreciably lower, of the order of less than 10% of the wild type rate. Such a very low rate was not easily measured with high precision and it could not be increased by the addition of aminoacids, e.g. asparagine. Oxygen consumption in wild type cells was almost completely inhibited by antimycin A (0.16 pM), cyanide (0.16 mM), and stimulated by dinitrophenol(O.03 mM; see above). On the other hand, it was difficult to demonstrate reproducibly any effects of these substances on the rate of 0 2 consumption in the “mutant,” and we are presently attempting to increase sensitivity and reduce instrument noise to settle this question.
A requirement for asparagine If we assume that the Krebs cycle is not functioning in the “mutant” cells, another likely consequence, in addition to the effect on the energy metabolism of these cells, is a deficiency in certain nonessential aminoacids which are normally derived from the Krebs cycle: glutamic acid, glutamine, aspartic acid and asparagine. As discussed by Meister (‘68) among others, (also Patterson, ’72) glutamine is not properly considered a non-essential aminoacid and is therefore present in all tissue culture media. We therefore placed the emphasis in our studies on the potential need for aspartic acid and asparagine. The relevant results are summarized in figure 7. This figure shows a comparison of the growth rates of parental and mutant cells under four different conditions: (1) in the absence of HCO2C02 and asparagine, (2) in the presence of asparagine only145 ~glml),(3) in the presence of HCOdC02 only, (4) in the presence of HCO3C02 and asparagine. While the growth rate of CCLl6 cells is relatively invariant under these conditions, the “mutant” cells clearly behave quite differently: in the absence of both HCO;/C02 and asparagine they die within a day; the presence of either HCO;/C02 or asparagine alone not only prevents cell death but permits a slow rate of proliferation; when both HCO&O2 and asparagine are present, cell proliferation approaches wild type rates.
A MUTATION AFFECTING THE KREBS CYCLE
9r
98-
: Q
299
I
7-
3
*
6-
%
5-
J J
w
2 I-
a -I W Q
4-
3-
1
0
2 3 0 I 2 3 DAYS DAYS Fig. 7 The effect of asparagine and C@/HCOY on the growth of CCLl6 and auxC@ cells. 1 0 5 cells were plated in 5 cm dishes in minimum essential Eagle’s medium with nonessential amino acids in a Con incubator. At time zero (one day later) cells were counted in one plate and the medium was changed in the other plates to minimum essential Eagle’s medium with glutamine, 10% fetal calf serum and 10 mM HEPES buffer, pH 7.5, without NaHCO,. The cells were counted with the Coulter counter at the indicated times and daily media changes were made in the other aliquot plates. 0,0:no additions; plates kept in air-incubator. 0, m: added 45 pgfml asparagine-H,O; air incubator. a,0 : NaHCO, added (2.2 glliter); CG-incuasparagine and NaHC03 added; C4-incubator. bator. 0 , I
.:
A similar experiment with aspartic acid substituted for asparagine was not successful, even at still higher concentrations of aspartate, perhaps because aspartate is less readily taken up by the cells (DISCUSSION).
Preliminary electron microscopic investigation Since the “mutant” cells were clearly very much deficient in respiratory activity we wished to examine whether there were any obvious differences in the number or morphology of mitochondria in CCLl6 and auxCOz cells. In sections obtained at random from embedded cells it appears that no gross differences are detectable (fig. 8). Numerous mitochondria are seen in the “mutant” cells, which at first glance seem normal, although on closer examination the organization of the cristae may be somewhat less regular, and the mitochondria appear somewhat deflated. The effect of inhibitors Since the “mutant” cells seem to be able to get along with a much reduced
rate of respiration, it seemed of interest to test whether these cells showed an increased resistance to inhibitors of respiration. High concentrations of cyanide or antimycin A were lethal to both mutant and wild type cells, but a narrow concentration range could be found where the growth rate of CCL16 cells was markedly depressed, while that of auxCOr cells was less affected. Figure 9 shows some results with antimycin A. Daily media changes were made in establishing these growth curves to counter the potential instability of the drug. DISCUSSION
The results presented in this paper greatly extend some previous work on an apparent CO2 auxotroph which was unable to oxidize pyruvate completely to carbon dioxide. We now show that in this “mutant,” in contrast to wild type cells, acetate, p-hydroxybutyrate, palmitate, and glutamate also cannot be oxidized to COr. It is most unlikely that this is the result of a defect in uptake, since such a diverse group of compounds is involved, and evenly internally produced [2J4C]
300
L. DEFRANCESCO, D. W E R N T Z A N D I. E. SCHEFFLER
Fig. 8 Electron micrographs of thin sections of C C L l 6 and auxCa cells. Sections were made as described under METHODS. A: CCL16. B : auxC02.For a discussion see the text.
A MUTATION AFFECTING THE
Figure 8
KREBS CYCLE
301
302
L. DEFRANCESCO. D. WERN’lZ AND I. E. SCHEFFLER
in collaboration with V. Mufioz and W. Butler) indicate that both “mutant” and wild type parental cells have the normal K complement of cytochromes, which can W 15 be oxidized in the presence of oxygen or 3 reduced either naturally by oxygen starz vation or by a strong reducing agent. These J -I experiments tend to indicate that the elecg 10 tron transport chain is intact. The enzyme W cytochrome oxidase is present in normal L lamounts (unpublished observation). Ada ditional evidence for a functional electron i i 5 K transport chain comes from studies with isolated mitochondria: with succinate as J the substrate oxygen consumption is com: I parable in normal and mutant mitochondria (DeFrancesco and Scheffler, ’75). Our results so far strongly suggest that a block exists in the Krebs cycle between a-ketoglutarate and succinate, i.e., that a-ketoglutarate dehydrogenase or succinyl CoA synthetase are defective. We cannot yet distinguish between a defect in either of these enzymes themselves or an effect on their activity due to a structural alteration in some as yet unidentified mitochondrial enzyme or protein. The evidence can be summarized as follows: (1) I4CO2is obtained from labeled aspartate and from labeled succinate even in the presence of inhibitors of de nouo synthesis of pyrimidines, suggesting that the COn is obtained via a series of reactions of the Krebs cycle from succinate to a-ketoglutarate; (2) radioactive aspartate can be converted to radioactive glutamate, confirming the above conclusions; (3) in the “mutant” cells [ 1-I4C] glutamate does not become decarboxylated at an appreciable rate (via a-ketoglutarate); (4) radioactive glutamate is not converted to radioactive aspartate in the “mutant” by the sequence of reaction, glutamate + a-ketoglutarate + succinate + + --+ oxaloacetate + aspartate. Since the reactions following succinate have been shown to occur, the block has to be placed in one of the two reactions between a-ketoglutarate and succinate; (5) the observation that labeled precursors entering the Krebs cycle via acetyl CoA will not yield l4CO2 is in agreement with this conclusion: because of the stereospecific interaction of citrate with aconitase, the acetyl carbons do not appear as COz during the first turn of the cycle, and 20
v
A MUTATION AFFECTING T H E KREBS CYCLE
the randomization of the carbons occurs only after free succinate is formed (in normal cells). A consequence of such a defect would be a requirement by the mutant cells of either aspartate or asparagine, since it had been shown (Eagle, ’59; Levintow et al., ’57) that a large portion of the aspartate and asparagine used by cells in culture is derived from glutamine rather than from glucose. Growth of our mutant cells is indeed markedly stimulated by the presence of asparagine but, surprisingly, not by aspartate. Although a difference in the rate of uptake of these two aminoacids may account for this result, a true explanation may require a complete understanding of the equilibria between the various species involved in amino group transfers: asparagine and glutamine as donors, glutamate and aspartate as potential donors or acceptors, and oxaloacetate and a-ketoglutarate as acceptors. This very interesting result deserves further study but we feel at this point that it is peripheral to our discussion of a respiration deficient mammalian cell line. Like some other tumor cells (for a review see, for example, Cooney and Handschumacher, ’70), our mutant may have lost the capacity to make asparagine from aspartate, but the proposal of a defect in L-asparagine synthetase alone would probably not account for all the properties of this “mutant.” We also do not believe that asparagine promotes growth by being converted efficiently to a plentiful supply of carbon dioxide, although some COZ may indeed be derived from this aminoacid: (1) 14C02 can be formed from [1-14C]asparagine, but at a rate which is not greater than that observed with [U-14C]aspartate; (2) high concentrations of asparagine in physiological buffer have no significant effect on the rates of formation of l4COz from [2-14C] and [1-14C]pyruvate (unpublished observations) and a condensation of oxaloacetate with acetyl CoA would certainly be an obligatory intermediate step; (3) the addition of asparagine has no effect on the rate of oxygen consumption by the mutant (unpublished observation$. The requirement for COz/HCOs of these cells is another puzzling problem on which some light might have been shed. The need
303
for exogenous CO, may arise primarily because of the inability of these cells to produce enough of their own, by means of the Krebs cycle, in contrast to wild type cells which become self-sufficient as long as uridine and hypoxanthine are added to the medium (Scheffler, ’74). What the COz is needed for is a second question which we may be able to answer in part: The C o n is required for the carboxylation reactions in which oxaloacetate or malate are formed from pyruvate (involving pyruvate carboxylase or the malic enzyme, respectively), and oxaloacetate in this case is mainly the precursor for aspartate and asparagine. Thus, in the presence of HC03/C02 the requirement for asparagine is partly overcome. The explanation is probably not complete, since our results show that the “mutant” cells grow apprsciably faster in the presence of both HCOd CO, and asparagine than in the presence of either component alone. Even higher concentrations of asparagine than those reported in figure 7 did not increase the growth rate, and the effect of HCOJC02 may therefore have to be sought at another level. The question remains unanswered whether COZ is required for some carboxylation reactions or whether COZ has some regulatory effect. For example, the intracellular pH may be to some extent different from the external pH and strongly influenced by the concentration O f HC0:JCOn. In most mammalian cells both glycolysis and respiration are means of producing metabolic energy and the two are controlled coordinately. It is known, from the work of Paul (’65) and others, that a high pH favors glycolysis while a lower pH favors respiration. Since the “mutant” cells have lost some of this flexibility, they may be more sensitive to the intracellular pH. If we are indeed dealing with a mutant then this would be the first reported case of a respirationdeficient mammalian cell line whose phenotype is absolutely stable. And if, as the evidence suggests, a defect can be located in the Krebs cycle, the mutant would constitute a very interesting addition to the existing collection of somatic cell mutants. A number of attempts have been made in the past to inhibit respiration in mam-
304
L. DEFRANCESCO, D. WERNTZ AND I. E. SCHEFFLER
malian cells in culture either by means of inhibitors or by imposing anaerobic conditions (for a discussion and review see Gregg, ’72). From his own experiments and those of others, Gregg (‘72) concludes that mammalian cells in culture cannot be grown anaerobically for prolonged periods of time (2-3 days). A different approach was taken by Naum and Pious (‘71) who attempted to obtain respiratory-deficient human cells analogous to the “petite” mutants of yeast by culturing the cells in the presence of sublethal concentrations of ethidium bromide. Although cytochrome oxidase activity was reduced 75 to 85% and cytochromes a a:{ had almost completely disappeared after a few days in the presence of ethidium bromide, the cells regained these activities following continued culture in drug-free medium. From the data presented it was not clear whether cells continued to divide at the normal rate once the level of activity of cytochrome oxidase had been reduced so drastically, or whether under these conditions the cells merely survived. More recently, other reports have appeared on permanent changes induced in the mitochondria of cells in culture by treatment with ethidium bromide, but no studies on the respiratory activity of these cells were reported (Klietmann, Sat0 and Nass, ’73; Klietmann et al., ’73). We also observe an increased rate of glycolysis in our “mutant” (Scheffler, ’74 and DeFrancesco, unpublished observations) and experiments are in progress to determine whether the increase in glycolysis can make up for a tenfold reduction in oxygen consumption. This large increase is in itself interesting. It could be due to: (1) a lack of feedback inhibition of glycolysis by Krebs cycle intermediates (e.g. citrate on phosphofructokinase); or (2) due to an increase in the specific activity of certain rate limiting enzymes in glycolysis. The latter may be a unique situation in this “mutant” and not possible for most ordinary cells in culture. The present problem also may have some bearing on the question of the relationship between glycolysis and respiration in normal and tumor cells (Gregg, ’72; Racker, ’72) and one may in fact ask
+
whether the observed phenotype is the result of a transformation or the result of a mutation. Insofar as both the parental and “mutant” cells are established cell lines, they may be already “transformed.” The “mutant” cells exhibit none of the more typical properties of tumor cells: (1) aneuploidy; (2) growth to high cell densities; (3) growth in agar or suspension; (4) low serum requirement. Moreover, a cell line obtained by transformation of a Chinese hamster fibroblast with SV40 virus behaves more like wild type cells with respect to its capacity for oxygen consumption (fig. 6). Finally, although the rate of respiration is often lower in tumor cells compared with normal tissue, the differences are significantly smaller than the ten-to-twelvefold difference observed in our case between parental and “mutant” cells. We, therefore, tend to favor the idea of a specific mutation. Our studies will now focus on cell-free systems and isolated mitochondria to p i n point the defect. We are also very interested in the behavior of this phenotype following somatic cell hybridizations with other hamster and human cells, which will become most meaningful if we are successful in identifying the defective enzyme. ACKNOWLEDGMENTS
This research was supported by Grant GM-18835, National Institutes of Health, and in part by a Program Project Grant from the National Science Foundation. L. DeFrancesco is a predoctoral fellow supported by Grant 2T01 GM-00702, National Institutes of Health. We would like to thank Mr. T. Cope for helping with the electron microscopy, Dr. W. Butler for allowing us to use his oxygen electrode and associated equipment and Dr. K . Poff for instructions on its use. We are also grateful to Dr. G . Stark for a generous amoLtnt of N-(phosphonacety1)-L-aspartate. LITERATURE CITED Berl, S . , A. Iztjtha and H.Waelsch 1961 Aminoacid and protein metabolism. IV. Cerebral compartments of glutamic acid metabolism. J. of Neurochem., 7: 186-197. Chang, R. S., H. Liepins and M. Margolish 1961 Carbon dioxide requirement and nucleic acid metabolism of HeLa and Conjunctival cells. Proc. SOC. Exp. Biol. M e d . , 1 0 6 : 149-152.
A MUTATION AFFECTING THE KREBS CYCLE
Collins, K. D., and G. R. Stark 1971 Aspartate transcarbamylase. Interaction with the transition state analogue N(phosphonacety1)-L-aspartate. J. Biol. Chem., 246: 6599-6605. Cooncy, D. A., and R. E. Handschumacher 1970 L-asparaginase and L-asparagine metabolism. Ann. Rev. Pharmacol., 10: 421-440. DeFrancesco, L., and I. E. Scheffler 1975 In preparation. Eagle, H. 1959 Aminoacid metabolism in mammalian cells i n culture. Science, 130: 432437. Gregg, C. T. 1972 In: Growth, Nutrition, and Metabolism of Cells in Culture. Vol. 1 . G. H. Rothblat and V . J. Cristofalo, eds. Academic Press, New York, pp. 83-136. Ishii, K., and H. Green 1973 Lethality of adenosine for cultured mammalian cells by interference with pyrimidine biosynthcsis. J. Cell. Sci., 13: 4 2 9 4 3 9 . Klietmann, W . , K. Kato, B. Gabara, H. Koprowski and N. Sato 1973 Stable alterations i n HeLa cell mitochondria following ethidium bromide treatment. Exp. Cell Res., 78: 47-58. Klietmann, W . , N. Sato and M. M. Nass 1973 Establishment and characterization of ethidium bromide resistance i n SV40-transformed hamster cells. J. Cell. Biology, 58: 11-26. Levintow, L . , H. Eagle and E. A. Piez 1957 The role of glutamine in protein biosynthesis in tissue culture. J . Biol. Chem., 227: 929-941. Maron, S. H.. and C. F. Prutton 1958 In. Prin-
305
ciples of Physical Chemistry. The Macmillan Co., New York, p. 145. Meister, A. 1968 On the synthesis and utilization ofglutamine. Harvey Lect., 63: 139-178. Naum, Y., and D. A. Pious 1971 Reversible inhibition of cytochrome oxidase accumulation in human cells by ethidium bromide. Exp. Cell. Res., 65: 335-339. Patterson, M. K., Jr. 1972 I n : Growth, Nutrition, and Metabolism of Cells in Culture. Vol. 1 . G. H. Rothblat and V. J. Cristofalo, eds. Academic Press, New York, pp. 171-209. Paul, J. 1965 Carbohydrate and energy metabolism. In: Cells and Tissues in Culture. Vol. 1 . E. N. Wittmer, ed. Academic Press, New York, pp. 239-276. Racker, E. 1972 Bioenergetics and the problem of tumor growth. American Scientist, 60: 5-3. Scheffler, I. E. 1974 Conditional lethal mutants of Chinese hamster cells. Mutants requiring exogenous carbon dioxide for growth. J. Cell. Physiol., 43: 219-230. Schemer, I. E., and G. Buttin 1973 Conditionally lethal mutations in Chinese hamster cells. I. Isolation of a temperature-sensitive line and its characterization by cell cycle studies. J. Cell. Physiol., 81: 199-216. Von Karff, R. W . 1969 Purity and stability of pyruvate and a-ketoglutarate. In: Methods in Enzymology. Vol. XII. J. M . Lowenstein, ed. Academic Press, New York, pp, 514-523.
