Somatic Cell Genetics, Vol. 3, No. 3, 1977, pp. 313-322

Variant Chinese Hamster Cells Resistant to the Proline Analog L-Azetidine 2-Carboxylic Acid M. L. Hooper, B. Carritt, P. S. G. Goldfarb, and C. Slack Cancer Research Campaign Somatic Cell Genetics Group, Institutes o f Genetics and Virology, Glasgow, Gl l 5JS United Kingdom R e c e i v e d 7 D e c e m b e r 1 9 7 6 - - F i n a l 31 J a n u a r y 1977

Abstract--Variants resistant to the toxic effects of the proline analog Lazetidine 2-carboxylic acid (AZCA) have been isolated from the Chinese hamster tissue culture line G3 by a three-step selection procedure using increasing concentrations of AZCA. Cells surviving each of the three selective steps have been examined for AZCA resistance a n d for proline uptake, biosynthesis, and degradation. The largest increment in AZCA resistance is acquired in the third step and is due to overproduction o f proline as a result o f increased activity of the enzyme system responsible for the conversion of glutamic acid to glutamic y-semialdehyde. It is not accompanied by an increase in the rate of formation o f proline from ornithine or in the rate of proline uptake or degradation.

INTRODUCTION Toxic analogs of amino acids have been used to considerable advantage in microorganisms as selective agents for the isolation of mutants deficient in amino acid uptake or utilization (1, 2). Similar mutants of mammalian tissue culture cells would be of great use, not only for the study of amino acid metabolism and its control, but also for investigations into membrane phenomena and for use as selective markers in genetic experiments. Few reports of the isolation of such mutants have appeared (35). In one case (5), variants of C H L cells resistant to the proline analog Lazetidine 2-carboxylic acid (AZCA) have been isolated in which resistance is due to overproduction of proline. These variants owe their increased level of proline production to a loss of inhibition by AZCA of glutamic ~/-semialdehyde formation from glutamate. Other classes of proline-overproducer mutants may, however, exist. We describe here the 313 ~)1977 Plenum Publishing Corp., 227 West 17th Street, New York, N.Y. 10011. To promote freer access to published material in the spirit of the 1976 Copyright Law, Plenum sells reprint articles from all its journals. This availability underlines the fact that no part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written permission of the publisher. Shipment is prompt; rate per article is $7.50.

314

Hooper et al.

independent isolation and the properties of AZCA-resistant variants of the Chinese hamster cell line G3.

MATERIALS AND METHODS

Cells and Reagents. G3 cells, a clone of the Chinese hamster line B14FAF28 (6), were obtained from the American Type Culture Collection, Rockville, Maryland, 63 subcultures after isolation of the clone (approximately 450 subcultures after isolation from tissue of origin), e-Azetidine 2-carboxylic acid and 3,4-dehydro-DL-proline were from Calbiochem; thioproline (L-thiazolidine 4-carboxylic acid) and 4hydroxyproline from Sigma. oL-[ring-14C]Azetidine-2-carboxylic acid (10 mCi/mmol) was obtained from Schwartz-Mann and L-[5-3H]proline (> 5000 mCi/mmol), L-[U-14C]proline (290 mCi/mmol), L-[U-14C]glutamic acid (265 mCi/mmol), and L-[l-~4C]ornithine monohydrochloride (58 mCi/mmol) from the Radiochemical Centre, Amersham, Bucks, United Kingdom. Cell Culture. Cells were grown in Glasgow-modified Eagle' s medium supplemented with 10% dialyzed fetal calf serum, nonessential amino acids with the exception of proline (glycine, alanine, aspartic acid, glutamic acid, and asparagine, each 0.1 raM; serine, 0.2 raM), nucleosides (adenosine, guanosine, cytidine and uridine, each 30/zM; thymidine, 10 /zM) and 1 mM sodium pyruvate. They were routinely subcultured once a week by dispersal with a trypsin-versene mixture (1 volume 0.25% Difco trypsin in Tris-buffered isotonic saline + 4 volumes disodium EDTA, 0.2 g/liter in phosphate-buffered isotonic saline). The fetal calf serum (GibcoBiocult) was dialyzed against three changes of isotonic saline, each change being left overnight, and then filter-sterilized. Some early experiments in high-density culture were carried out on cells grown in Glasgowmodified Eagle's medium supplemented with 10% dialyzed fetal calf serum and 0.5 mM serine only. With the exception of incorporation experiments (see text), this gave results similar to the complete supplemented medium. Cells were monitored at regular intervals for mycoplasma contamination by the method of F o g h and Fogh (7), and contaminated cultures were discarded. Determination of Analog Resistance at Cloning Density. Cells were seeded at 10 cells/cm ~ into a series of replicate dishes and medium containing analog applied 24 hr later. After 1 week dishes were stained with Leishman's stain and colonies counted. The LDs0 of an analog was determined as that concentration required to reduce the number of colonies to 50% of that observed in an analog-free control.

AZCA-Resistant Chinese Hamster Cells

315

Selection of Variants Resistant to AZCA. For the first two rounds of selection for AZCA resistance, cells were plated out at subconfluent density (3 • 103 cells/cm 2) and medium containing AZCA applied 24 hr later. Cells were maintained in selective medium with medium changes at 3-day intervals, until killing appeared complete (7-19 days). They were then transferred to analog-free medium. Surviving colonies were picked and grown up in nonselective medium for further testing. Subsequent reconstruction experiments (data not shown) eliminated the possibility that AZCA resistant mutants are lost when selection is performed at high density (cf. ref. 8), and so the third round of selection for AZCA resistance was performed at 1.5 • 104 cells/cm z. Initial screening of surviving colonies was carried out by seeding cells at 3 • 103 cells/cm 2 into 3-cm 2 wells (Linbro), applying fresh medium containing analog at 24 hr and again at 4 days, and scoring qualitatively for cell growth at 1 week. Quantitation of Proline Biosynthesis from Glutamic Acid. Cells were seeded into 20-cm 2 dishes at 1.5 • 104 cells/cm 2, and 24 hr later 0.125 /zCi/ml[14C]glutamic acid was added. After a further 24 hr, cells were washed three times with phosphate-buffered saline and scraped into 1 ml H20. Protein was precipitated by the addition of 100 /zl 100% trichloracetic acid (TCA) at 4~ and the supernatant extracted twice with an equal volume of ether to remove TCA. The aqueous phase was lyophilized, dissolved in 100/zl water, applied to a silica-gel thin layer (Merck DC-Alufoilen Kieselgel 60), and subjected to ascending chromatography for 16 hr in phenol/water 3 : 1 (w/v) containing 0.2 g/liter NaCN. Proline and glutamic acid spots were located by autoradiography, identified by comparison with standards run in parallel, and quantitated by scraping the gel into scintillation vials and counting in NE233 toluenebased fluid (Nuclear Enterprises Ltd., Sighthill, Edinburgh). Determination of Time Course of Proline Biosynthesis from Glutamic Acid and Ornithine. Cells were seeded as above and at 24 hr the medium replaced by medium containing 10-4 M cycloheximide and 1.25 /zCi/ml [aH]proline as internal standard. At various times ranging from 1.5 to 7.5 hr later [14C]glutamic acid or [14C]ornithine was added to a final concentration of 0.25/zCi/ml, and at 7.5 hr all plates were harvested and processed as described above. Replicate plates were given three washes with phosphate-buffered saline, taken up in 5 ml Folin solution C (9), and an aliquot used for the determination of protein by the method of Lowry et al (9), using bovine serum albumin as standard. For each sample [14CJproline radioactivity was normalized to the internal standard [3H]proline radioactivity, and the resulting figure was converted to dpm [14C]proline per/xg protein using the mean value of dpm [3H]proline per

Hooper et al.

316

/zg protein averaged over all time points for each cell line. This mean value did not differ significantly between G3 and G3ac526. Karyology. Metaphase spreads were prepared and G-banded as previously described (10). RESULTS

Toxicity of AZCA in G3 Cells. AZCA inhibits the growth of G3 cells at 3 txg/ml under cloning conditions and at 30 tzg/ml in high-density culture (data not shown). Protection against AZCA toxicity is seen in the presence of proline and to a lesser extent with certain other L-amino acids. Glutamic acid gives marked protection, presumably because of its metabolism to proline. Protection by the other amino acids tested could be ranked as follows: A l a > Ser - A s n > Gly > Asp. This sequence correlates well with the abilities of these amino acids to act as inhibitors of proline transport, suggesting that they protect by competing with AZCA for entry into the cells. Selection of AZCA-Resistant Variants. Table 1 summarizes the isolation of AZCA-resistant variants from G3 cells. Initial selection was performed at an analog concentration of 30/xg/ml, the survival frequency varying from 10-4 to 5 x 10-3. Thirteen surviving colonies were picked, grown in the absence of analog, and then tested for resistance to AZCA in mass culture. Twelve grew at 30/zg/ml analog but none at 100/zg/ml. One resistant line, G3ac5, was chosen for further study. Fluctuation analysis (11, 12) demonstrated that resistance to 30/xg/ml AZCA is acquired as a result of a process distributed in time throughout the life of the parent clone and generating a stably inherited phenotype, rather than as a response to the selective agent. The rate of this process (between 8 x 10 -7 and 2 x 10-6 cell -I generation -1) is of the same order as the mutation rate to H G P R T - , an X-linked mutation (13). Because of the partial nature of the analog resistance seen in G3ac5 Table 1. Selection of A Z C A - r e s i s t a n t variants

Parent cell line

Selecting conc. A Z C A (/zg/ml)

Frequency of surviving colonies

No. of colonies retested

G3 G3ac5 G3ac52

30 100 1000

10-4-5 x 10-a 8 x 10 -6 10-6

13 12 19

aSevere growth inhibition, i.e., resistance marginal.

No. of tested colonies resistant to A Z C A conc, of (t~g/ml) 30

100

300

12 12

0 11 19

2~ 19

1000

Clone selected for further study

14

G3ac5 G3ac52 G3ac526

AZCA-Resistant

Chinese Hamster C e i l s

317

and its sibs, further rounds of selection were performed (Table 1). On the basis of an initial screening for resistance in mass culture, one of the most resistant variant clones from each selection was chosen for further study. Properties of AZCA-Resistant Variants. Table 2 shows the LD~0 of AZCA for colony formation in G3 and its derivatives. G3ac5 is significantly more resistant to AZCA at low cell density than is G3 (P < 0.005). However, although initial screening (Table 1) suggested that G3ac52 is more resistant to AZCA than G3ac5, no significant difference in resistance is found between these variants at low cell density. We have not investigated whether this anomaly is due to a reproducible difference in response to the two kinds of test, statistical variation, or loss of resistance of G3ac52 with increasing time in culture. Of the three selective steps, the one accompanied by the greatest increment in AZCAresistance is the final step in which G3ac526 is derived from G3ac52. We have also examined the cross-resistance of these clones to killing by other proline analogs and by [3H]proline. 3,4-Dehydro-oL-proline, like AZCA, is incorporated into protein (14), whereas 4-thioproline is thought not to be incorporated into protein (15), although it competes with proline for uptake in G3 cells (unpublished results). G3ac526 is more resistant than G3 to [3H]proline and 3,4-dehydroproline but not to 4-thioproline, which gives a first indication that AZCA resistance here is not due to a transport defect. While G3ac5 and G3ac52 show no reduction in the level of incorporation of.exogenously supplied radioactive proline or AZCA into protein, G3ac526 shows a marked reduction, the effect being somewhat enhanced in the presence of 0.1 mM glutamic acid. Since the growth rates of G3ac526 and G3 are not substantially different, this cannot reflect a decreased rate of protein synthesis, but must represent a decrease in the specific activity of the intracellular proline pool. This could arise because of either a reduction in the rate of transport of proline and AZCA into the cells or an increase in the rate of synthesis of proline, the latter possibility

Table 2. P r o p e r t i e s o f A Z C A - r e s i s t a n t

variants of G3 a

Cell line G3 LD~0 of AZCA for colony formation (~edml) Pro/Glu ratio after Glu labeling

2.2-~ 0.3

G3ac5 (9)

0.104 ~ 0.017 (6)

6.2 • 0.9

G3ac52 (7)

0.106 _+ 0.047 (4)

8.5 • 2.2

G3ac526 (6)

0.070 • 0.020 (4)

"Results are quoted as mean • SEM. Figures in brackets give number of determinations.

> 2 5 0 (3) 0.874 - 0.120 (6)

318

Hooper e t a I.

being suggested by the effect of glutamic acid on proline incorporation. These two possibilities ate examined in the following sections. Transport o f Proline and AZCA into G3 and G3ac526 Cells. In experiments to characterize proline transport in G3 cells (data not shown) we have found that the kinetics of uptake are consistent with entry via a single, carrier-mediated active transport system plus simple diffusion. A number of proline analogs, including AZCA, as well as a substantial pror portion of the naturally occurring L-a-amino acids, inhibit proline transport, suggesting competition for a system of broad specificity. No differences in the rate of transport of either proline or AZCA were found in any of the AZCA-resistant variants. Biosynthesis of Proline in G3 and Its Derivatives. Table 2 shows that after incubation for 24 hr in [14C]glutamic acid, G3a526 cells exhibit a much higher ratio of radioactivity in proline to radioactivity in glutamic acid than do G3 cells, showing an increased activity of the pathway by which proline is synthesized from glutamic acid (Fig. 1). G3ac5 and G3ac52 show proline/glutamic acid ratios not significantly different from that of G3, which demonstrates that the increased activity in this pathway is acquired in the selective step in which G3ac526 is derived from G3ac52. The rate-limiting step in the conversion of glutamic acid to proline in mammalian cells is thought to be, as in microorganisms (16), the formation of glutamic semialdehyde. The enzymes responsible for this conversion have proved difficult to assay in cell-free extracts. They have only relatively recently been identified even in E. coli and then only with the aid of particular mutants (17, 18). Therefore, in order to localize the increased activity of the pathway to a particular step, we have compared the rates of formation of proline from glutamic acid and from ornithine in intact cells. 1 glu

GSA

~

~

orn

2 3

CH pro / / / ~ protein

PCA ~

4

Fig. 1. Biosynthesis and metabolism of proline in animal cells (16). The enzyme(s) catalyzing the conversion of glutamic acid to glutamic semialdehyde (GSA) have not been characterized in cell-free extracts of any animal cell, but are assumed to be analogous to those described by Baich (17, 18). Enzymes catalyzing other interconversions are (2) ALpyrroline dehydrogenase (EC 1.5.1.12); (3) Al-pyrroline-5-carboxylate reductase (EC 1.5.1.2); (4) proline oxidase; (5) ornithine aminotransferase (EC 2.6.1.13). GSA and PCA (Al-pyrroline-5-carboxylic acid) are in spontaneous equilibrium. The site of action of cycloheximide (CH) is also shown.

AZCA-Resistant Chinese Hamster Cells

319

i

8

b

II

6

6

f

4

2

2

|

I

2

4

I

6 0 hrs

I

I

!

2

4

6

Fig. 2. Time course of formation of [14C]proline from (a) [14C]glutamic acid and (b) [14C]ornithine in G3 (o) and G3ac526 (e).

In order to ensure that none of the proline formed was lost from the TCA-soluble fraction of the cells by incorporation into protein, this comparison was carried out in cells treated with 10-4 M cycloheximide, a concentration which inhibits protein synthesis in these cells by 97%. Loss of proline by degradation via glutamic semialdehyde can be ruled out because when either cell line is labeled for 24 hr with [14C]proline, chromatography of the TCA-soluble fraction of the cell reveals no radioactivity in any position other than that of proline. Figure 2 shows the time course of formation of proline from glutamic acid and from ornithine in G3 and G3ac526. It is clear that although the rate of formation from glutamic acid is elevated in G3ac526 cells, the rate of formation from ornithine is unaltered. This result is incompatible with the hypothesis that the increased activity of the pathway is solely due to an increased activity of Al-pyrroline-5-carboxylic acid reductase, the enzyme catalyzing step 3 of Fig. 1, but would be expected on the basis of an increased activity of the system catalyzing step 1. It should be pointed out, however, that we cannot exclude the possibility that the activity of both steps is increased, since our method will only detect changes in activity of the rate-limiting step of each interconversion. A direct assay of Aa-pyrroline-5-carboxylate reductase in cell-free extracts has been reported (19), use of which would resolve this question.

320

Hooper et al.

Karyotypes of G3 and Its Derivatives. While G3, G3ac5, and G3ac52 each have a narrow distribution of chromosome number about a mode of 22, the modal chromosome number of G3ac526 is shifted to 24 (data not shown). Examination of G-banded karyotypes shows that this is due to the presence in G3ac526 of two marker chromosomes, viz, a telocentric and a symmetrical metacentric chromosome. G3ac526 also appears to possess extra material at the distal end of the long arm of one homolog of each of chromosomes 4, 6, and 8. Chromosome counts were made on three sibs of G3ac526 showing comparable levels of AZCA-resistance. All had a modal chromosome number of 24 or 25. However, a causal relationship between the extra chromosomes and AZCA resistance cannot be inferred from this, since all these sibs may be clonally related.

DISCUSSION This paper describes the isolation of AZCA-resistant variants by a three-step selection procedure. The first step is accompanied by a small but significant increase in resistance, and although its biochemical basis is unknown it has some of the properties of a single-gene mutation. The second step is associated with little or no increase in resistance, while the third step is accompanied by a large increase in resistance to AZCA. This increase in resistance is due to overproduction of proline as a result of an increase in activity of the enzyme system responsible for the conversion of glutamic acid to glutamic y-semialdehyde. A number of possible mechanisms of increase in activity would be consistent with our experimental observations, including gene dosage or a mutation affecting the rate of transcription or translation of the structural gene or the specific activity of the enzyme protein itself. To obtain further information about this, we have constructed fusion hybrids between G3ac526 and G3ac52, its immediate parent. Preliminary investigation of the AZCA resistance of uncloned hybrid populations, shown by karyotype analysis to consist mainly of the products of ls + ls fusions with little or no chromosome loss, indicates that the resistant phenotype exhibits partial dominance (unpublished results). This rules out the possibility that overproduction is due to a defect in the production of a diffusible negative-control element unless the latter is normally present in limiting amounts. It suggests, in fact, that each parental genome expresses in the hybrids the same level of proline production as in the parent cell. A question of some interest is whether we are dealing with a mutation whose effect is confined to a single gene product or with a wider-ranging genetic

AZCA-Resistant Chinese Hamster Cells

321

change. While we cannot at present rule out that the increase in activity of glutamic ,/-semialdehyde production from glutamate is accompanied by an increase in activity of hl-pyrroline-5-carboxylic acid reductase, it is certainly not accompanied by an increase in the rate of formation of proline from ornithine, or in the rate ofproline uptake or degradation. Wasmuth and Caskey (5) have recently described the isolation from CHL cells of AZCA-resistant mutants where resistance is also associated with proline overproduction. The independent isolation of overproducer mutants ir two different cell lines suggests that they may represent the most frequently arising class of AZCA-resistant mutants in animal cell lines. Wasmuth and Caskey demonstrate that in wild-type CHL cells AZCA inhibits glutamic 3~-semialdehyde biosynthesis from glutamate and that proline overproduction in the mutants is due to loss of this inhibition. However, two preliminary lines of evidence suggest that in wild-type G3 cells such an inhibition is not present: (1) In an experiment of the type described in Table 2 we were unable to detect any effect of 30/xg/ml AZCA on the rate of proline formation from glutamate (unpublished results). (2) In G3 cells, unlike in CHL cells, glutamate protects against A Z C A toxicity. Further work, however, is required to determine whether this represents a fundamental difference between G3 and CHL cells. In none of the AZCA-resistant variants here described is resistance due to polyploidy as described by Kao and Puck (20). G3ac526 does, however, show various karyotypic abnormalities, one or more of which may be causally related to A Z C A resistance. The role of the first two selective steps in the isolation of G3ac526 from G3 is unclear at present. It is possible that they are purely incidental and that a variant like G3ac526 can be selected in a single step from the wild type using a higher AZCA concentration than was used in the selection of G3ac5. Indeed, the overproducer variants obtained by Wasmuth and Caskey were isolated by a single-step selection procedure, although they show a much lower level of AZCA resistance than G3ac526 and, as already discussed, may have a different basis. Alternatively, the first two steps may play a functional role by inserting mutations providing a more favorable genetic background for the expression of proline overproduction or by allowing a prolonged exposure to AZCA which may play some role in generating overproducer variants. The latter possibility could be investigated by performing selections on wild-type cells previously exposed to concentrations of AZCA insufficient to kill all nonmutant cells. Further characterization of proline overproduction in G3ac526 will require direct assay of the enzymes involved in converting glutamate to glutamic 7-semialdehyde, and investigation of the control mechanisms

322

Hooper et al.

operating on their synthesis and function. Comparison of our variant with that of Wasmuth and Caskey will determine whether both share a common genetic basis or whether overproduction can be brought about by more than one mechanism. ACKNOWLEDGMENTS We would like to thank Professors J. A. Pateman and J. H. SubakSharpe for their support and encouragement, Dr. R. Elton for help in computer analysis of uptake data, and Dr. D. M. Scott for helpful discussion. We are also grateful to Mrs. J. Grieves and Mrs. R. H. M. Morgan for excellent technical assistance. This work was supported by a grant from the Cancer Research Campaign to Professors J. A. Pateman and J. H. Subak-Sharpe. LITERATURE CITED 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

Shive, W., and Skinner, C. G. (1963). In Metabolic Inhibitors, (eds.) Hochster, R. M., and Quastel, J. H., Vol. I: pp. 1-73, Academic Press, New York. Halpern, Y. S. (1974). Annu. Rev. Genet. 8:103-133. Caboche, M. (1976). J. Cell. Physiol. 87:321-335. Kamely, D., and Littlefield, J, W. (1974). Exp. Cell Res. 89:154-160. Wasmuth, J. J., and Caskey, C. T. (1976). Cell 8:71-78. Yerganian, G., and Leonard, M. (1961). Science 133:1600-1601. Fogh, J., and Fogh, H. (1964). Proc. Soc. Exp. Biol. Med. 117:899-901. van Zeeland, A. A., van Diggelen, H. C. E., and Simons, J. W. I. M. (1972). Mutat. Res. 14:355-363. Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951). J. Biol. Chem. 193:265-275. Slack, C., Morgan, R. H. M., Carritt, B., Goldfarb, P. S. G., and Hooper, M. L. (1976). Exp. Cell Res. 98:1-14. Luria, S. E., and Delbruck, M. (1943). Genetics 28:491-511. Lea, D. E., and Coulson, C. A. (1949). J. Genet. 49:264-285. Chu, E. H. Y., Brimer, P., Jacobson, K. B., and Merriam, E. V. (1969). Genetics 62:359-377. Fowden, L., Neale, S., and Tristam, H. (1963). Nature 199:35-38. Takeuchi, T., and Prockop, D. J. (1969). Biochim. Biophys. Acta 175:142-155. Meister, A. (1%5). Biochemistry o f the Amino Acids, Vol. 2, Academic Press, New York. Baich, A. (1969). Biochim. Biophys. Acta 192:462-467. Baich, A. (1971). Biochim. Biophys. Acta 244:129-134. Strecker, H. J. (1971). In Methods in Enzymology, (eds.) Colowick, S. P. and Kaplan, N. O., pp. 258-261, Academic Press, New York. Kao, F. T., and Puck, T. T. (1967). Genetics 55:513-524.