The Effects of Cycloheximide upon Transcription of rRNA, 5 S RNA, and tRNA Genes*

The effects of inhibitors of protein synthesis upon transcription have been re-examined. Cycloheximide (1 rg/ml) inhibits incorporation of uridine into RNA of P1798.520 lymphosarcoma cells. Filter hybridization studies indicate that labeling of pre-rRNA is inhibited 6 0 4 0 % after 1 h and quantitative S1 nuclease mapping reveals a corresponding decrease in the amount of cellular pre-rRNA. Cycloheximide also inhibits labeling of 5 S RNA and tRNA, but incorporation of uridine into poly(A+) RNA is unaffected. Transcription experiments carried out in nuclei from cycloheximide- treated cells indicate that the inhibitor causes a selec-tive decrease in the activity of RNA polymerases I and 111. Cell-free extracts from P1798.520 were used to transcribe the cloned mouse rRNA gene, Syrian ham- ster 5 S RNA gene, and the Drosophila tRNAArg gene. Extracts from cycloheximide-treated cells were inhib- ited in this respect. Transcription of rRNA and 5 S RNA genes was inhibited by 90% after 2 h and 50% inhibition occurred within 20-30 min. Transcription of the tRNA gene was inhibited 75% after 2 h with a half-time of -1 h. Inhibition was due neither to a direct effect of cycloheximide nor to the presence of nucleases or diffusible inhibitors of transcription. Moreover, transcription of rDNA in extracts from cycloheximide- treated cells could be restored by the addition of a partially purified initiation

The Effects of Cycloheximide upon Transcription of rRNA, 5 S RNA, and tRNA Genes* (Received for publication, September 10, 1985) Preeti K. Gokal, Alice H. Cavanaugh, and E. Aubrey Thompson, Jr.$ From the Department of Biology, University of South Carolina, Columbia, South Carolina 29208 The effects of inhibitors of protein synthesis upon transcription have been re-examined. Cycloheximide (1 rg/ml) inhibits incorporation of uridine into RNA of P1798.520 lymphosarcoma cells. Filter hybridization studies indicate that labeling of pre-rRNA is inhibited 6 0 4 0 % after 1 h and quantitative S1 nuclease mapping reveals a corresponding decrease in the amount of cellular pre-rRNA. Cycloheximide also inhibits labeling of 5 S RNA and tRNA, but incorporation of uridine into poly(A+) RNA is unaffected. Transcription experiments carried out in nuclei from cycloheximidetreated cells indicate that the inhibitor causes a selective decrease in the activity of RNA polymerases I and 111.
Cell-free extracts from P1798.520 were used to transcribe the cloned mouse rRNA gene, Syrian hamster 5 S RNA gene, and the Drosophila tRNAArg gene. Extracts from cycloheximide-treated cells were inhibited in this respect. Transcription of rRNA and 5 S RNA genes was inhibited by 90% after 2 h and 50% inhibition occurred within 20-30 min. Transcription of the tRNA gene was inhibited 75% after 2 h with a half-time of -1 h. Inhibition was due neither to a direct effect of cycloheximide nor to the presence of nucleases or diffusible inhibitors of transcription. Moreover, transcription of rDNA in extracts from cycloheximidetreated cells could be restored by the addition of a partially purified initiation factor preparation. The data indicate that inhibition of protein synthesis results in rapid depletion of transcription factors that are required for initiation by RNA polymerases I and 111. Among these is the glucocorticoid-regulated rDNA initiation factor designated TFIC.
Inhibition of protein synthesis in yeast is accompanied by inhibition of pre-rRNA synthesis (Foury and Goffeau, 1973;Gross and Pogo, 1976, a and b). The synthesis of pre-rRNA in mammalian cell culture lines is inhibited by amino acid starvation (Franze-Fernandez and Pogo, 1971;Grummt and Grummt, 1976) as well as inhibition of protein synthesis by cycloheximide and puromycin (Warner et al., 1966;Soeiro et al., 1968;Craig and Perry, 1970;Chesterton et al., 1975;Mishima et al., 1979; see also Warner, 1974;Hadjiolov and Nikolaev, 1976). Injection of cycloheximide into rats has been reported to cause rapid inhibition of rRNA synthesis in liver CA22394 and CA24347 to E. A. T. The costs of publication of this * This work was supported by National Institutes of Health Grants article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
j Recipient of American Cancer Society Faculty Research Award FRA299. To whom correspondence and requests for reprints should be addressed. and liver nucleoli (Yu and Feigelson, 1972), although these observations have been challenged (Stoyanova and Dabeva, 1980). These data have been interpreted to indicate the existence of an rDNA transcription factor that exhibits a very short biological half-life.
In this laboratory, we have begun to study hormonal regulation of transcription of rDNA and have identified a hormone-regulated transcription initiation factor (Cavanaugh et al., 1984). Kinetic studies suggest that this protein turns over rapidly and may be a logical candidate to account for the effects of protein synthesis inhibitors on transcription of rDNA (Yu and Feigelson, 1973). Other studies suggest that the rate of transcription of the 5 S RNA gene may be regulated by partitioning of transcription initiation factor IIIA between the gene and nascent 5 S RNA (Pelham and Brown, 1980;Honda and Roeder, 1980;Hanas et al., 1983). In the absence of ribosome assembly, 5 S RNA may accumulate to the extent that transcription of the gene is inhibited as the factor is diverted to form the TFIIIA-5 S RNA complex. If this mechanism is correct, inhibition of translation should result in rapid inhibition of synthesis of 5 S RNA as TFIIIA is depleted. Similar factors may play a role in transcription of tRNA and rRNA genes and could provide a general mechanism of coordination of synthesis of ribosomal components. To test this hypothesis, we have re-examined the effects of cycloheximide upon RNA synthesis in lymphosarcoma P1798 cells in culture.
The data demonstrate that synthesis of pre-rRNA, 5 S RNA, and tRNA is rapidly inhibited after cessation of protein synthesis and suggest that transcription of class I and I11 genes requires a protein(s) of short biological half-life.

Effect of Cycloheximide upon RNA Synthesis in Cells+
Preliminary experiments were undertaken to determine the optimum concentration of cycloheximide for inhibition of protein synthesis in P1798 cells. Incorporation of [3H]leucine into trichloroacetic acid-precipitable material was inhibited 90-95% in the presence of 1 pg/ml cycloheximide. Higher concentrations produced no additional effect and maximum inhibition was achieved within 1-2 h. The effects of cycloheximide upon incorporation of [3H]uridine are shown in Fig. 1. Total uridine incorporation during a 15-min pulse-labeling period (diamonds in Fig. 1) was inhibited approximately 80% The "Experimental Procedures are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document No. 85M-3048, cite the authors, and include a check or money order for $1.20 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press. The efficiency of hybridization, monitored by hybridization of 3Hlabeled cRNA internal controls, was estimated to be 6-8% under the conditions employed and the control (0 h) filter bound 1.2 X lo4 3H cpm. Nonspecific binding to pBR322 was 50-60 'H cpm. In a parallel series of experiments, unlabeled RNA was extracted from lo7 cells that had been exposed to 1 pg/ml cycloheximide and 2.5 pg of RNA was analyzed for pre-rRNA sequences by quantitative S1 mapping as described under "Experimental Procedures" (closed circles).

Hours
in this experiment and maximum inhibition was observed within 2-4 h after the ad4tion of cycloheximide. RNA was extracted and the amount of labeled prerrRNA was measured by filter hybridization. As shown in Fig. 1, incorporation of [3H]uridine into pre-rRNA (open circles) paralleled inhibition of total [3H]uridine incorporation. Quantitative S1 nuclease mapping was used to estimate the amount of pre-rRNA in cycloheximide-treated cells. The observed decrease in the amount of cellular pre-rRNA ( Fig. 1, closed circles) paralleled the inhibitory effects of cycloheximide upon synthesis of labeled pre-rRNA. These observations are consistent with the very short half-life (5-10 min) of the rRNA primary transcript (Perry, 1976;Gurney, 1985). In a brief pulse-labeling experiment of this sort, one would expect this species of RNA to be disproportionately labeled. The rapid turnover likewise accounts for the rapid disappearance of the primary transcript following inhibition of synthesis.
Cultures of P1798 were treated with cycloheximide for various periods of time, the cells were pulse-labeled with [3H] uridine, and labeled RNA species were resolved as described under "Experimental Procedures." As shown in Fig. 2, pulselabeling of pre-rRNA was inhibited -60%. Labeling of 5 S RNA and tRNA was reduced by 40-50%; there was no detectable decrease in the amount of 5 S RNA or tRNA, as estimated by the intensity of staining with ethidium bromide. Inhibition of labeling of 5 S RNA was also confirmed by filter hybridization to pTH1 (data not shown). Labeling of poly(A+) RNA was unaffected by a 2-h exposure to cycloheximide indicating that the apparent inhibition of synthesis of other RNA species is not a reflection of decreased specific activity of intracellular nucleotide pools.
The data indicate that cycloheximide causes inhibition of transcription by RNA polymerases I and 111. This was confirmed by measuring the activity of these enzymes in nuclei isolated from control and cycloheximide-treated cells, as shown in Table IA. The assays were carried out in the presence of exogenous template and under these conditions one measures primarily elongation of RNA chains that were initiated prior to isolation of nuclei. Nuclei from cycloheximidetreated cells exhibited a significant decrease in the activity attributable to RNA polymerases I and 111. No significant decrease was observed in RNA polymerase I1 activity, consistent with the lack of effect upon labeling of poly(A+) RNA. RNA polymerase activity was also measured in extracts from control and cycloheximide-treated cells (Table IB). Exposure to cycloheximide (1 pg/ml, 2 h) did not affect transcription of calf thymus DNA by RNA polymerases I or 111. The data indicate that neither nonspecific initiation nor nucleotidyltransferase activity was inhibited and suggest that cycloheximide acts at the level of specific initiation upon class I and class I11 genes.

Effect of Cycloheximide or Emetine upon RNA Synthesis in
Vitro-S100 extracts were prepared and assayed for the ability to transcribe cloned mouse rDNA (prMAB/PuuII), the Syrian hamster 5 S RNA gene (pTHl), and the Drosophila tRNAArg gene (pYH48). The results are shown in Fig. 3. Transcription of the PuuII-truncated, cloned mouse rRNA TABLE I RNA polymerase activity A, two 1-liter cultures were prepared and one was treated for 2 h with 1 pg/ml cycloheximide. Nuclei were prepared and polymerase activity was estimated in the presence of appropriate concentrations of a-amanitin, as described under "Experimental Procedures." The numbers in parentheses are the activities relative to RNA polymerase I1 in that preparation. B, SlOO extracts were prepared from control and cycloheximide-treated cells. Activities of RNA polymerases were estimated in the presence of a-amanitin using calf thymus DNA as described under "Experimental Procedures." The activity of RNA polymerase I1 in such extracts was too low to reliably measure.
[   , lane a). The ability to transcribe 5 S RNA and tRNA genes was reduced in SlOO extracts from cycloheximidetreated or emetine-treated cells (lanes b and c, respectively).
The time course of inhibition was determined by preparing SlOO extracts from cells treated with cycloheximide for various intervals. As shown in Fig. 4, transcription of the cloned rRNA and 5 S RNA genes was inhibited -90% within 2 h; 50% inhibition occurred approximately 20 min after addition of cycloheximide. Transcription of the cloned tRNA gene was reduced by -75% under these conditions and 50% inhibition occurred 50-60 min after addition of the inhibitor.
Inhibition of transcription in vitro was not due to a direct effect of cycloheximide upon the enzymes and/or initiation factors. As shown in Fig. 5, addition of cycloheximide to SlOO extracts did not inhibit transcription of the genes encoding rRNA, 5 S RNA, or tRNA. As an additional control, a mixing experiment was carried out as shown in Fig. 6. Mixtures of SlOO extracts from control and cycloheximide-treated cells were able to transcribe the cloned genes, indicating that inhibition of transcription was not due to the presence of nucleases or diffusible inhibitors in extracts from cycloheximide-treated cells.
Reconstitution of Transcription in Vitro-An SlOO extract was fractionated by chromatography on DEAE-cellulose and phosphocellulose as described under "Experimental Procedures." Four fractions, designated A, B, C, and D, were obtained and none of these was capable of specific transcription of mouse rDNA (Fig. 7, lanes a-d). Fraction C from the phosphocellulose column (eluted at 600 mM KCl) was capable of reconstituting transcription of rDNA in extracts from cycloheximide-treated cells (Fig. 7, lane g).
Phosphocellulose fraction C contains at least two components of the rDNA transcription complex: a small but detectable amount of RNA polymerase I and the hormone-regulated initiation factor TFIC (Cavanaugh and Thompson, 1985). Either or both of these factors could account for the effects of cycloheximide. A template exclusion protocol was employed to determine if RNA polymerase I from cycloheximide-treated cells was capable of specific initiation. The rationale for such tracts were prepared from lo9 cells that had been treated with 1 pg/ ml cycloheximide for various periods of time. These were used to transcribe the cloned genes encoding pre-rRNA (prMAB, open circles), 5 S RNA (pTH1, closed circles), or tRNAAg (pYH48, closed diamonds). The products were resolved by electrophoresis on poly-acry1amide:urea gels and specific transcripts were detected by autoradiography, excised from the gels, and counted for 32P by liquid scintillation. a procedure is as follows. Extracts were preincubated with template under conditions that permitted formation of stable, preinitiation complexes. Formation of such complexes was evidenced by the observation that preincubation with a given template precluded transcription of a second template added subsequently. An extract from control cells was heated for 15 min at 45 "C to inactivate RNA polymerase I. Such extracts have been previously shown to be incapable of transcription but may be reconstituted by the addition of partially purified or highly purified RNA polymerase I (Cavanaugh and Thompson, 1984). The heat-inactivated extract was preincubated with prMAB truncated with SmaI (prMABISma1) so as to introduce a double-stranded break at +155 bp. An extract from cycloheximide-treated cells was preincubated with prMAB truncated with PuuII at +292 bp (prMAB/PuuII). After preincubation, the extracts were mixed and prMAT11/ PuuII was added. This third template was truncated with PvuII at +350 bp. Transcription was carried out and the products were characterized as shown in Fig. 8, lanes a- transcriptional activity was reduced in the cycloheximidetreated extract (lane b). When the two reactions were mixed, however, both of the preincubated templates were transcribed to yield RNAs of 292 nt (prMABIPuuI1) and 155 nt (prMAB/ S m I ) . The third template (prMATll/PuuII) was not transcribed. These data suggest that both heated and cycloheximide-treated extracts are capable of forming stable, preinitiation complexes, thereby excluding transcription of templates added subsequent to preincubation. Preinitiation complexes formed in heat-treated extracts cannot support transcription because they do not contain functional RNA polymerase I. However, RNA polymerase I from cycloheximide-treated cells can reconstitute heat-treated extracts indicating that cycloheximide does not cause inactivation of the enzyme. Moreover, preinitiation complexes form in extracts from cycloheximide-treated cells and can support transcription if supplemented with one or more heat-stable factors from control extracts.
Extracts from dexamethasone-treated cells contain functional RNA polymerase I but are depleted of the transcription initiation factor TFIC Thompson, 1983, 1985): Mixing experiments indicate that extracts from hormone-treated cells do not reconstitute extracts from cycloheximide-treated cells (Fig. 8, lanes d, e, and f). Therefore, at least one factor in common is involved in both cycloheximideand glucocorticoid-mediated inhibition of transcription of rDNA. Finally, partially purified TFIC is capable of reconstituting extracts from cycloheximide-treated cells (Fig. 8, lanes  g and h). In this experiment, TFIC was preincubated with prMAB/SmaI and a cycloheximide-treated extract was preincubated with prMAB/PuuII. TFIC does not carry out specific transcription of rDNA (lane g); the high level of nonspecific transcription is due to contamination of this fraction with RNA polymerase I. When the reactions were mixed, only prMAB/PuuII was transcribed (lane h). These data confirm  8. Reconstitution of transcription of rDNA. Lune a, a heat-treated extract was used to transcribe prMAB/SmaI. Lane b, a cycloheximide-treated extract was used to transcribe prMAB/PuuII. Lune c contains the products of a mixing experiment in which a heattreated extract was preincubated with prMAB/SmaI as described under "Experimental Procedures." In parallel, a cycloheximidetreated extract was preincubated with prMAB/PuuII. Thereafter, the reactions were mixed, 0.4 pmol of prMATIIIPvuI1 was added, and transcription was initiated by the addition of nucleoside triphosphates. The arrows indicate the calculated positions of RNA of 155 nt (prMABISmaI), 292 nt (prMAB/PuuII), and 350 nt (prMATII/ PuuII). Lune d, an extract from P1798 cells, treated 24 h with 0.1 ~L M dexamethasone, transcribing prMAB/SmaI. Lune e, an extract from cycloheximide-treated cells transcribing prMABIPuuI1. Lane f, a mixing experiment in which a hormone-treated extract was preincubated with prMABISma1 and mixed with a cycloheximide-treated extract that had been preincubated with prMABIPuuI1. Lune g, 5 pl of TFIC (0.5 pg of protein) transcribing prMABISmaI1. Lune h, a cycloheximide-treated extract was preincubated with prMAB/PuuII and mixed with a similar reaction in which 5 pl of TFIC had been preincubated with prMABISma1. Transcription was initiated by the addition of nucleoside triphosphates and the products were resolved as described. Lune m contains 5' end-labeled fragments derived by digestion of 6x174 with HaeIII. The smallest fragment shown in lune m has a size of 118 nt. the observation that extracts from cycloheximide-treated cells are capable of forming preinitiation complexes but are depleted of TFIC.

DISCUSSION
Cycloheximide inhibits the synthesis of pre-rRNA, 5 S RNA, and tRNA in P1798 cells. The effect prevails in nuclei and in SlOO extracts from cycloheximide-treated cells indicating that inhibition occurs at the level of transcription. Three observations suggest that the effects of cycloheximide are mediated via inhibition of translation rather than by a direct effect upon transcription: (i) inhibition in culture is specific for RNA polymerases I and 111, indicating that cycloheximide is not a nonspecific inhibitor of transcription; (ii) cycloheximide does not inhibit transcription in vitro; (iii) the effects of cycloheximide are identical with those of emetine, a structurally dissimilar inhibitor of translation. Moreover, transcription of rDNA in extracts from cycloheximide-treated cells can be reconstituted by the addition of partially purified transcription factors from control cells. This indicates that the transcriptional defect exhibited by these extracts is not due to general degradation or denaturation.
It has been reported that transcription of rDNA is influenced by intracellular nucleotide pools (Grummt and Grummt, 1976). Cycloheximide may alter pool sizes under certain circumstances (Stoyanova and Dabeva, 1980). However, labeling of poly(A+) RNA was not reduced in cyclohex-S RNA, and tRNA Genes imide-treated cells suggesting that changes in UTP pools do not occur under the experimental conditions employed. Mishima et al. (1979) have observed that cellular ATP and GTP pools are unaffected by treatment with 100 pg/ml cycloheximide for 90 min. Although we have not measured nucleotide pool sizes, we assume that pool sizes remain relatively constant and do not account for inhibition of transcription under the circumstances described in this report.
Our working hypothesis proposes that initiation of transcription by RNA polymerases I and I11 requires the continual synthesis of factors that are rapidly degraded or consumed during the transcription process. Attempts to reconstitute transcription of 5 S RNA and tRNA genes have yielded equivocal results in that more than one phosphocellulose fraction contains stimulatory activity. Our primary interest is in transcription of rDNA and the mechanism of action of cycloheximide upon RNA polymerase I11 has not been pursued.
Inhibition of rRNA synthesis results from a decrease in the amount or activity of a protein that co-purifies with the glucocorticoid-regulated rDNA transcription initiation factor TFIC. The data strongly suggest that TFIC exhibits a short biological half-life and is rapidly depleted after inhibition of translation. This is consistent with an early report that indicated that a protein of short half-life is involved in glucocorticoid regulation of rRNA synthesis in rat liver (Yu andFeigelson, 1972, 1973). Other data suggest that TFIC may be involved in inhibition of transcription of rDNA in serum or amino acid-starved cells (Grummt, 1982). This protein therefore appears to play a central role in regulating the rate of synthesis of the precursor for rRNAs. In order to prove this hypothesis, it will be necessary to purify TFIC.
To date, none of the ancillary RNA polymerase I transcription factors has been purified to homogeneity and the biochemical properties of the rDNA transcription complex are not completely elucidated. Fractionation studies suggest that at least three classes of proteins are involved RNA polymerase I; the stable complex factor(s) that elute from phosphocellulose in 1 M KC1 (TFID); and the initiation factor TFIC (Mishima et al., 1982). Poly(ADP-ribose) polymerase may also be involved (Kurl and Jacob, 1985). Template exclusion experiments indicate that extracts from cycloheximidetreated cells are capable of forming stable preinitiation complexes. It is therefore unlikely that the effects of cycloheximide are attributable to the stable complex factors designated TFID. TFIC is rather strongly associated with the polymerase and, in our experience, it is difficult to completely resolve TFIC from the enzyme. The most highly purified preparations of TFIC contain 4 % of the polymerase I activity that was present in the SlOO extracts and it is difficult to preclude the possibility that TFIC represents a minor subpopulation of functional RNA polymerase I. Reconstitution studies with heat-inactivated extracts argue against this possibility. The polymerase from cycloheximide-treated extracts is functional when mixed with heated extracts from control cells. Conversely, a heat-stable factor from control cells can reconstitute cycloheximide-treated extracts. Since heat-treated extracts contain no detectable RNA polymerase activity, it is difficult to argue that reconstitution of transcription under these circumstances is attributable to the polymerase. The simplest interpretation of these data is that TFIC is distinct from RNA polymerase I. However, one may not preclude the possibility that TFIC is an exchangeable polymerase subunit or a heatstable enzyme that modifies the polymerase in some way so as to facilitate specific initiation.
Elucidation of the mechanism that underlies the effect of inhibitors of translation requires purification and characterization of the factors involved. At the present time, the number of factors involved is unknown. We are intrigued by the possibility that a class of proteins of short biological half-life may be involved in coordinating the rate of translation and the rates of synthesis of rRNAs, 5 S RNAs, and tRNAs. Such a mechanism could have physiological significance in preventing overproduction of these RNA species under conditions in which ribosome assembly is limited by the availability of the ribosomal proteins and may be analogous to a stringent control mechanism in mammalian cells.