Escherichia coli was studied under various conditions that affect levels of charged tRNA. Protein breakdown in- creased markedly when the synthesis of ...
Proceedings of the Natml Academy of Sciences Vol. 68, No. 2, p 3 -366, February 1971
A Role of Aminoacyl-tRNA in the Regulation of Protein Breakdown in Escherichia coli ALFRED L. GOLDBERG* Department of Physiology, Harvard Medical School, Boston, Massachusetts 02115
Communicated by Bernard D. Davis, November 30, 1970 The rate of degradation of cell proteins in Escherichia coli was studied under various conditions that affect levels of charged tRNA. Protein breakdown increased markedly when the synthesis of valyl-tRNA was prevented in strains containing temperature-sensitive valyl-tRNA synthetase or when the formation of Nformylmethionyl-tRNA was inhibited with trimethoprim. Conversely, protein breakdown decreased in a valine auxotroph-administered valine or an analog capable of attachment to the valyl-tRNA. It appears that changes in the levels of aminoacyl-tRNA regulate protein breakdown by mechanisms similar to those controlling the synthesis of ribosomal RNA. These experiments also demonstrate that protein synthesis is not essential for protein degradation and suggest that the inhibition of degradation in starving cells by chloramphenicol is a secondary effect of the accumulation of charged tRNA.
degradation requires concomitant protein synthesis and have shown that inhibitors of protein synthesis can block degradation by causing accumulation of aminoacyl-tRNA.
ABSTRACT
MATERIALS AND METHODS
The following strains of Escherichia coli were used in these experiments: 1-9, which containsa temperature-sensitive valyltRNA synthetase, was originally isolated by F. C. Neidhardt. 1OB6, an arginine auxotroph, which also contains a temperature-sensitive valyl-tRNA synthetase, was isolated by Dr. S. Kaplan and was obtained from Dr. D. Ezekiel, Philadelphia, Pa. M-48-62, a valine auxotroph, and A33, which requires arginine and tryptophan, were obtained from Dr. B. Davis. All strains were grown on glucose-minimal medium (12) with aeration. For studies with 10B6, arginine was present in concentrations of 60 Mug/ml, and with M-48-62, valine at 60 jug/ml. Strains 1-9 and 1OB6 were grown at 300C, other strains at 370C. The cell concentration at the start of the experiments ranged between 75 and 125 ,ug of cell protein/ml. Cell growth was estimated with a Klett-Summerfield calorimeter. DL-threo-a-amino-,3-chlorobutyric acid hydrochloride was kindly provided by Dr. Marco Rabinovitz of the National Cancer Institute. DLa-aminobutyrate was obtained from the Sigma Chemical Co. [U-3H]Leucine was purchased from the New England Nuclear Corp. Trimethoprim was a gift from Dr. G. H. Hitchings of the Burroughs Wellcome Co. Experimental design was adapted from that of Mandelstam (11) and Schlessinger and Ben-Hamida (4). Bacteria were initially grown for at least two generations in the presence of [3H]leucine (0.04,MCi/ml) to label cell proteins. The cells were collected by centrifugation at 250C and washed twice with the final incubation medium, which contained 75,Mg/ml of unlabeled L-leucine to chase the radioactive amino acids released from proteins from the cell. (Failure to add sufficient leucine to the medium drastically reduced the release of [8H]leucine from the bacteria.) The washed cells were incubated under conditions specified in the text. At various times, 1-ml aliquots were added to 0.1 ml of 50% trichloroacetic acid (TCA), incubated at room temperature for 30 min, and centrifuged in the cold. The supernatant was decanted, and the TCA was extracted with three volumes of ether. Radioactivity was assayed by liquid scintillation counting with Bray's solution (13). For measurements of the radioactivity in the protein precipitate, the precipitate was washed twice with 5% TCA and once with ethanol-ether 1:1, then solubilized with Nuclear-Chicago
When exponentially growing bacteria are deprived of nitrogen, a carbon source, or a required amino acid, the rate of catabolism of cellular protein increases severalfold (1). This response appears to be an important adaptive mechanism for the cell. When the supply of nutrients in the medium is reduced, protein degradation can provide a source of energy or essential amino acids for the synthesis of enzymes appropriate to the new conditions (1, 2). Several observations suggested to us that control of protein breakdown might be coupled to the regulation of net RNA synthesis. Conditions that increase protein breakdown in bacteria (1-4), such as starvation or "step-down", are also those that drastically reduce net RNA synthesis (5), while conditions that maximize synthesis of ribosomal and transfer RNA (3) (e.g., exponential growth) appear to reduce protein breakdown to minimal levels (1-3). While this work was in progress, these correlations were also pointed out by Sussman and Gilvarg (6). Similar relationships between the rates of RNA synthesis and protein breakdown occur in mammalian skeletal muscle, in which rapid growth involves both increased RNA synthesis and decreased protein breakdown, while atrophy involves both decreased RNA synthesis and an increased rate of protein degradation (7). Since the levels of aminoacyl-tRNA play a crucial role in regulating the synthesis of ribosomal RNA in bacteria (8-10), we have tested the possibility that they also determine rates of protein breakdown (4). These experiments indicate that the increased rate of protein breakdown upon starvation results from a decreased supply of charged tRNA. In addition, we have reinvestigated earlier reports (3, 4, 11) that protein Abbreviations: TCA, trichloroacetic acid; CAP, chloramphenicol. * Fellow of the Medical Foundation, Inc.
362
Vol. 68,1]971
Protein Breakdown in E. coli
Solubilizer. Quenching was estimated by the method of external standards.
363
strains without a temperature-sensitive synthetase, leucine was released from proteins only 2-3 times faster at 42°C than at 300C, in accord with previous measurements
RESULTS Requirements for charged tRNA
(17).
Effect of valine analogs
The increased protein breakdown that occurs during starvation for nitrogen or for a required amino acid could result from a decrease in the intracellular pools of the amino acids or from a reduction in the supply of aminoacyl-tRNA. To decide between these alternatives, protein breakdown was studied in strain 1-9, where valyl-tRNA synthetase is temperature-sensitive (8, 14). As shown in Fig. 1, [3H]leucine was released from protein many times faster at the nonpermissive temperature (40'C) than at 30'C. At the lower temperature, the absorbance of the culture doubled in 90 min, while at the higher temperature, the absorbance increased by only 24% in 6 hr; in addition, at 40'C ['4C]arginine was incorporated into proteins only 4% as fast as at 30'C. A sevenfold increase in protein breakdown upon transfer to the nonpermissive temperature was also observed when this strain was grown on rich broth. Similar changes in protein breakdown were obtained when the cells were initially labeled with [U-14C]arginine or [1-14C]valine. Analogous experiments were carried out in strain 1OB6, which also contains a temperature-sensitive valyl-tRNA synthetase (15, 16) but which is not as leaky as I-9 (15) at the nonpermissive temperature. At 420C no growth could be detected and [3H]leucine was released from proteins 9-12 times faster than at 30'C. In contrast, in six other E. coli
These results demonstrated increased breakdown in the absence of a functioning valyl-tRNA synthetase. They do not indicate, however, whether the crucial step regulating the degradative process is the formation of the aminoacylsynthetase complex, the amino acid adenylate-synthetase complex, or the aminoacyl-tRNA. To distinguish between these possibilities, we have employed the valine analogs D-aaminobutyric acid and DL-threo-ca-amino-p-chlorobutyric acid. While both of these compounds are activated by E. coli valyl-tRNA synthetase (18), Freundlich (19) has shown that only the chlorobutyrate is also transferred to valine-specific transfer RNA. As shown in Fig. 2, administration of aminobutyrate to a valine auxotroph deprived of valine did not reduce protein breakdown rates, although chlorobutyrate, valine, or both caused a threefold reduction. Thus, only conditions which permit attachment of an amino acid to the tRNA reduced protein breakdown in the auxotroph. Effect of trimethoprim To test for an influence of other species of charged tRNA on
protein catabolism, I utilized the antimetabolite, trimethoprim, which inhibits the synthesis of N-formyltetrahydrofolate (20). In this way, trimethoprim blocks the formation of N-formylmethionyl-tRNA (21), which is required for peptide + ABA
A - VAL E ICL
40e
,
U)
2000 . TRI
0.
VAL -i
10
CBA
1500 . a d00
, 30-
2 4 TIME (hr )
6
&
TIME
(hr )
00
2
TIfME
4 (hr )
6
(Left) FIG. 1. Protein breakdown in cells containing a temperature-sensitive valyl-tRNA synthetase. E. coli I-9 was grown at 30'C for two generations in minimal medium containing [3H]leucine, washed, and resuspended in minimal medium containing unlabeled leucine. The rates of conversion of labeled proteins to TCA-soluble material were then compared at the permissive (30'C) and nonpermissive (40'C) temperatures. (Center) FIG. 2. Effects of valine analogs on protein breakdown in a valine auxotroph. Cells of E. coli M 48-62 were grown for two generations in minimal medium containing valine and [3H]leucine, washed, and resuspended in minimal medium containing unlabeled leucine and no valine. The following additions were then made: valine (VAL) at a final concentration of 60 ,g/ml, DL-a-aminobutyric acid (ABA) at 60 sg/ml, or DL-threo-a-amino-,i-chlorobutyric acid (CBA) at 75 ,g/ml. Only the culture to which valine was added grew during the incubation period. (Right) FIG. 3. Effect of trimethoprim on protein breakdown. E. coli A33 was grown for two generations in minimal medium supplemented with arginine, tryptophan, and [3H]ileucine. After washing, the cells were resuspended in minimal medium containing arginine, tryptophan, and unlabeled leucine. The final concentration of trimethoprim (TRI) was 50 ,ug/ml, which prevented further growth.
364
2 1600
-
Proc. Nat. Acad. Sci. USA
A. L. Goldberg
Microbiology:
NH4, 30-
40+ CAP
I
1200
J
800 z
+CAP
-i
400
0
2
4
60 TIME (hr)
2
4
6
FIG. 4. Effects of chloramphenicol on protein breakdown in E. coli 1-9. After growth as in Fig. 1, the cells were washed and incubated further in the presence of unlabeled leucine with or without chloramphenicol (CAP) at a concentration of 100 ,.zg/ml. The culture at 300C was suspended in minimal medium lacking ammonium sulfate; that at 40°C was in minimal medium. No culture grew during the incubation. Cells deprived of a carbon source responded to CAP in a fashion similar to those starved for nitrogen.
chain initiation. Addition of trimethoprim to growing cells of strain A33 prevented further growth, inhibited [14C]arginine incorporation by 86%, and increased protein breakdown strikingly (Fig. 3). Similar, although less marked, effects were obtained with A33 growing on Casamino acids. This finding makes it unlikely that trimethoprim is promoting protein breakdown simply by blocking amino acid biosynthesis. Trimethoprim had similar effects in all strains tested, including 1-9, 1OB6, and M-48-62. Independence of protein synthesis and protein breakdown
In Figs. 1-3 increased protein breakdown was observed under conditions where protein synthesis was blocked. These results thus contradict earlier observations in both microbial (3, 4, 6, 11, 22) and mammalian (23) systems that inhibition of protein synthesis also reduces protein breakdown. In E. coli, for example, chloramphenicol (CAP) has been found to inhibit breakdown induced by deprivation of amino acids or nitrogen (3, 4, 6, 11). However, since CAP blocks peptide chain elongation at the ribosome, it also causes the accumulation of charged tRNAs, even in starved cells (24). A build-up of aminoacyl-tRNAs might act to inhibit protein breakdown, just as decreased levels appear to promote this process. This model would predict that CAP should not affect breakdown under conditions where amino-tRNA cannot accumulate. As shown in Fig. 4, CAP inhibited release of TCA-soluble leucine in strain 1-9 starved for nitrogen at the permissive temperature. However, at 40°C, where valyl-tRNA cannot be synthesized, 500 ,g/ml of CAP had no effect on protein degradation. Similarly, though CAP (100 ug/ml) decreased breakdown by about 40% in strain 1OB6 deprived of arginine or of a carbon source at 300C, it had no effect at the nonpermissive temperature. Even at 1 mg/ml, CAP reduced proteolysis in 1OB6 by only 10% at 420C. These results demonstrate that the CAP effect on breakdown in starving cells requires functioning aminoacyl synthetases. DISCUSSION Several types of evidence have been presented that suggest that the increase in protein breakdown in cells starved for nitrogen or a required amino acid occurs in response to
changes in the cellular level of charged tRNA. In cells growing on a complete medium, protein degradation was found to increase markedly when the formation of valyl-tRNA (Fig. 1) was blocked in strains containing a temperature-sensitive valyl-tRNA synthetase. Treatment of growing cells with trimethoprim, which blocks the conversion of methionyltRNA to N-formylmethionyl-tRNA, also promoted protein breakdown (Fig. 3). Moreover, in auxotrophs deprived of valine, rates of degradation decreased (Fig. 2) upon administration not only of valine but also of chlorobutyrate, a valine analog that is capable of attachment to valyl-tRNA but does not permit growth. Finally, treatment of starving cells with chloramphenicol appears to reduce degradation toward levels seen during growth by permitting the accumulation of charged tRNA. These studies, however, cannot distinguish whether the increased catabolism is a direct effect of the altered levels of aminoacyl-tRNA, or whether it is an indirect effect mediated by some metabolite whose formation requires a full complement of charged tRNAs. This latter possibility indeed appears most likely, since it would explain why starvation for any of several amino acids can promote catabolism, and why inhibition of valyl-tRNA synthetase and inhibition of the formylation of methionyl-tRNA yields similar results. Since the only common site at which the various species of charged tRNA interact is the ribosome, it appears likely that some aspect of ribosomal function is involved in the regulation of protein breakdown. To explain the present findings, the critical step in protein synthesis would have to occur after the binding of the aminoacyl-tRNA but prior to the peptidyl transferase reaction, which is blocked by CAP. The present results emphasize the similarity between the physiological regulation of protein breakdown and of net RNA (predominantly ribosomal RNA) synthesis. This latter process also appears to be regulated by the levels of charged tRNA (8-10), although the exact mechanisms of this control are unclear. Recently guanosine-tetraphosphate has been found to accumulate in cells deprived of aminoacyl-tRNA and may be implicated in the control of ribosome synthesis (25); similar mechanisms may well be involved in the regulation of protein breakdown. The various conditions found here to promote protein breakdown have all been previously reported to block ribosome synthesis. For example, transfer of strains 1-9 (8) or 1OB6 (16) to the nonpermissive temperature and the administration of trimethoprim (21) markedly reduce net RNA synthesis. These earlier observations have been confirmed in this laboratory under the present conditions (Goldberg, unpublished). Furthermore, the various conditions that were found here to inhibit degradation have been reported to permit RNA synthesis in E. coli starved for required amino acids. Thus, administration of chlorobutyrate to an auxotroph deprived of valine (9), or treatment with chloramphenicol (5) under conditions which permit accumulation of charged tRNA (14, 16), lead to net RNA build-up in the absence of growth. Finally, in related studies, I have found that "step-up" and "step-down" conditions lead to simultaneous but opposite changes in rates of protein catabolism and RNA synthesis. Similarities between the control of ribosomal RNA synthesis and protein breakdown have also been pointed out by Sussman and Gilvarg (6). These workers observed that in two "relaxed" (rel-) mutants (5) of E. coli, nitrogen starvation
Vol. 68, 1971
not only failed to block ribosomal RNA synthesis but also did not augment protein degradation as markedly as in controls (rel+). We have made similar observations, but have found certain rel- strains that increase breakdown in a normal fashion upon amino acid starvation (Goldberg, unpublished observations). Although the correlation is thus not perfect, it is extensive enough to suggest that the systems regulating protein degradation and net RNA synthesis share common elements. Coordinate control of these two processes would appear of selective advantage to the cell, because of their complementary physiological effects. Thus, maximal ribosome and tRNA synthesisand minimal proteolysis both serve to promote growth in rich medium, while reduced RNA synthesis and enhanced protein breakdown would act synergistically to augment intracellular pools of amino acids, nitrogenous bases, and other intermediary metabolites during "hard times". Analogous changes also occur in mammalian skeletal muscle and kidney, where the rates of net RNA and protein synthesis and of protein breakdown also appear to vary in a complementary manner during rapid hypertrophy and atrophy (7). The present experiments also indicate similarities between the control of protein breakdown and of the biosynthesis of enzymes for amino acid synthesis. Derepression of the biosynthetic pathways for several amino acids (12, 19, 26) have been shown to be regulated by the levels of corresponding charged tRNAs. The conditions that promote protein catabolism in Figs. 1 and 2 also derepress the enzymes for valine and isoleucine biosynthesis (14, 19). Protein catabolism and synthesis of the biosynthetic enzymes also have complementary physiological roles, since both processes serve to augment the supply of amino acids in poor media. These experiments also demonstrate that protein synthesis is not essential for protein breakdown. In fact, maximal degradation was observed under conditions that prevented further protein synthesis (Figs. 1-3). However, inhibition of synthesis with CAP was found to reduce breakdown in cells starved for nitrogen, an amino acid, or a carbon source, as reported previously (3, 4, 6, 22). CAP blocks the transfer of amino acids from charged tRNA to the growing polypeptide chain. Thus, on one hand, protein breakdown increased when formation of aminoacyl-tRNA was prevented; on the other hand, it decreased when unloading of the charged tRNA was prevented. These studies suggest that chloramphenicol reduces protein breakdown toward levels found in growing cells by causing the build-up of charged tRNA in the starved cells (24). This conclusion is supported by the finding that the effect of CAP was not observed when accumulation of valyl-tRNA could not occur (Fig. 4). It is of interest that administration of CAP to growing cultures of E. coli does not reduce protein breakdown further (Goldberg, in preparation). Presumably, under these conditions, in contrast to conditions of starvation, aminoacyltRNAs are not limiting. Since breakdown in starved bacteria is inhibited by CAP, proflavine (4), and rifampicin (6), it has been suggested (4, 6) that this process involves the de novo synthesis of proteolytic enzymes. However, the present findings (Figs. 1-3) make it likely that the degradative enzymes active during starvation also exist in exponentially growing cells and that these proteases do not themselves turn over during starvation. This conclusion is in accord with recent experiments (3, 27) demonstrating significant degradation in exponentially growing cells,
Protein Breakdown in E. coli
365
as was also observed in the present studies (see controls in Figs. 1-3). Presumably, the supply of charged tRNA influences either the activity of preexisting proteases or the sensitivity of certain cell proteins to these enzymes. The demonstration of an influence of aminoacyl-tRNA on breakdown may also explain other observations on the effects of inhibitors of RNA and protein synthesis on protein degradation. For example, inhibitors of messenger-RNA synthesis reduce protein breakdown in bacteria (4, 6, 28) as well as in liver (23) and hepatoma cells (29). Such agents should cause the accumulation of charged tRNA in a similar fashion to CAP and therefore would also be expected to decrease protein breakdown (Goldberg, in preparation). Finally, although the results presented here can explain the effects of inhibitors of protein and RNA synthesis, they cannot account for the observation that inhibitors of energy metabolism block catabolism (1, 4, 22, 28, 30, 31). In related studies, we have confirmed such observations and have found that low concentrations of dinitrophenyl or carbonyl cyanide block degradation under every condition studied here. These findings, however, may still reflect some secondary effect of energy metabolism (31), rather than a direct involvement of an energy source in the proteolytic process. I thank Miss Dawn Tsien for her expert technical assistance and Dr. Bernard Davis for his advice in the preparation of this manuscript. These studies were supported by a grant from the Milton Fund of Harvard University and from the Air Force Office of Scientific Research. 1. Mandelstam, J., Bacteriol. Rev., 24, 289 (1960); Ann. N.Y. Acad. Sci., 102, 621 (1963). 2. Goldberg, A. L., Nature, submitted for publication (1970). 3. Willetts, N. S., Biochem. J., 103, 453 (1967); Biochem. Biophys. Res. Commun., 20, 692 (1965). 4. Schlessinger, D., and F. Ben-Hamida, Biochim. Biophys. Acta, 119, 171 (1966). 5. Maaloe, O., and N. 0. Kjeldgaard, Control of Macromolecular Synthesis (W. A. Benjamin, New York, 1966); Stent, G. S., and S. Brenner, Proc. Nat. Acad. Sci. USA, 47, 2005 (1961); Edlin, G., and P. Broda, Bacteriol. Rev., 32, 206 (1968). 6. Sussman, A. J., and C. Gilvarg, J. Biol. Chem., 244, 6304 (1969). 7. Goldberg, A. L., J. Biol. Chem., 244, 3217, 3223 (1969). 8. Neidhardt, F. C., Bacteriol. Rev., 30, 701 (tO"). to(&e( 9. Williams, L. S., and M. Freundlich, Biochim. Biophys. Acta, 179, 515 (1969). 10. Ezekiel, D. H., Biochim. Biophys. Acta, 95, 48 (1965). 11. Mandelstam, J., Biochem. J., 69, 110 (1958). 12. Davis, B. D., and E. S. Mingioli, J. Bacteriol., 60, 17 (1950). 13. Bray, G. A., Anal. Biochem., 1, 279 (1960). 14. Eidlic, L., and F. C. Neidhardt, J. Bacteriol., 89, 706 (1965). 15. Kaplan, S., and D. Anderson, J. Bacteriol., 95, 991 (1968). 16. Ezekiel, D. H., and B. N. Elkins, Biochim. Biophys. Acta, 166, 466 (1968). 17. Pine, M. J., J. Bacteriol., 93, 1527 (1967); Epstein, I., and N. Grossowicz, J. Bacteriol., 99, 418 (1969). 18. Bergmann, F. B., P. Berg, and M. Dieckmann, J. Biol. Chem., 236, 1735 (1961). 19. Freundlich, M., Science, 157, 823 (1967). 20. Burchall, J. J., and G. H. Hitchings, J. Mol. Pharmacol., 1, 126 (1966). 21. Shih, A. Y., J. Eisenstadt, and P. Lengyel, Proc. Nat. Acad. Sci. USA, 56, 1599 (1966). 22. Halvorson, H. O., Amino Acid Pools, ed. J. T. Holden (American Elsevier Publishing Co., New York, 1964), p. 646. 23. Steinberg, M., and M. Vaughn, Arch. Biochem. Biophys., 65, 93 (1956); Grossman, A., and C. Mavrides, J. Biol. Chem., 242, 1398 (1967); Schimke, R. T., Nat. Cancer Inst. Monogr., 27,
366
Microbiology: A. L. Goldberg
301 (1967); Walter, H., Nature, 188, 643 (1960); Kenney, F., Science, 156, 525 (1967); Levitan, I. B., and T. E. Webb, J. Biol. Chem., 244, 4685 (1969); Feldman, M., and G. Yagil, Biochem. Biophys. Res. Commun., 37, 198 (1969). 24. Morris, D. W., and J. A. Demoss, J. Bacteriol., 90, 1624 (1965); Ezekiel, D. H., and B. Valulis, Biochim. Biophys. Acta, 129, 123 (1966); Aronson, A. I., and S. Spiegelman, Biochim. Biophys. Acta, 53, 70 (1961). 25. Cashel, M., and J. Gallant, Nature, 221, 838 (1969); Cashel, M., J. Biol. Chem., 244, 3133 (1969); Cashel, M., and B. Kalbacher, J. Biol. Chem., 245, 2309 (1970). 26. Schlesinger, S., and B. Magasanik, J. Mol. Biol., 9, 670 (1969); Eidlic, L., and F. C. Neidhardt, Proc. Nat. Acad. Sci. USA, 53, 539 (1965); Roth, J. N., and B. N. Ames, J. Mol. Biol., 22, 325 (1966); Szentermai, A. M., M. Szentermai, and H.
Proc. Nat. Acad. Sci. USA Umbarger, J. Bacteriol., 95, 1672 (1968); Ito, K., S. Hiraga, and T. Yura, Genetics, 61, 521 (1969). 27. Pine, M. J., J. Bacteriol., 103, 207 (1970); Nath, K., and A. L. Koch, J. Biol. Chem., 245, 2889 (1970). 28. Spudich, J. A., and A. Kornberg, J. Biol. Chem., 243,4588 (1968). 29. Peterkofsky, B., and G. M. Tomkins, J. Mol. Biol., 30, 49 (1967); Auricchio, F., D. Martin, and G. Tomkins, Nature, 224, 806 (1969). 30. Simpson, M. V., J. Biol. Chem., 201, 143 (1953); Korner, A., and H. Tarver, J. Gen. Physiol., 41, 219 (1957); Schimke, R. T., E. W. Sweeney, and C. M. Berlin, J. Biol. Chem., 246, 4609 (1965). 31. Brostom, C. O., and H. Jeffay, J. Biol. Chem., 245, 4001 (1970).
