rRNA transcription initiation is decreased by inhibitors of the yeast cell cycle control step "start".

Inhibitors of the "start" regulatory step in the cell cycle of the yeast Saccharomyces cerevisiae are known by indirect studies to perturb RNA metabolism. We have investigated these effects further and show here by a pulse-labeling and quantitative hybridization procedure that pre-rRNA transcription was substantially decreased by five inhibitors of start but was transiently stimulated by the mating pheromone alpha-factor. Thus in contrast to the effects of the other start inhibitors, the inhibition of start by alpha-factor is unrelated to this aspect of biosynthetic activity. Mating factor treatment also stimulated the synthesis rate of poly(A)+ RNA. The start inhibitors o-phenanthroline and L-ethionine inhibited pre-rRNA transcription with little effect on poly(A)+ RNA synthesis rates. Northern analysis showed that all inhibitors of start also inhibited pre-rRNA transcript cleavage, a process that has been dissociated from the inhibition of start. Most inhibitors also affected ATP pool size. One inhibitor, o-phenanthroline, markedly induced the general control response.

Inhibitors of the "start" regulatory step in the cell cycle of the yeast Saccharomyces cerevisiae are known by indirect studies to perturb RNA metabolism. We have investigated these effects further and show here by a pulse-labeling and quantitative hybridization procedure that pre-rRNA transcription was substantially decreased by five inhibitors of start but was transiently stimulated by the mating pheromone a-factor. Thus in contrast to the effects of the other start inhibitors, the inhibition of start by a-factor is unrelated to this aspect of biosynthetic activity. Mating factor treatment also stimulated the synthesis rate of poly(A)+ RNA. The start inhibitors o-phenanthroline and L-ethionine inhibited pre-rRNA transcription with little effect on poly(A)+ RNA synthesis rates.
Northern analysis showed that all inhibitors of start also inhibited pre-rRNA transcript cleavage, a process that has been dissociated from the inhibition of start. Most inhibitors also affected ATP pool size. One inhibitor, o-phenanthroline, markedly induced the general control response.
For the budding yeast Saccharomyces cerevisiae the control of cell proliferation is exerted at the regulatory step termed "start" (Hartwell, 1974;Hartwell et al., 1974). Under most nutritional conditions the performance of start takes place when a cell has sufficient biosynthetic capacity to produce a viable daughter cell (Hartwell, 1974). Biosynthetic capacity in this situation is a function of cell mass, because the performance of start occurs only in cells that have acquired a threshold cell size or mass (Johnston et al., 1977). How cell mass is monitored for start is not clear.
Certain inhibitors have been shown to block cell proliferation by disabling the performance of start. For the mating pheromone a-factor this inhibition of start  is part of a response mediated by G-proteins (Dietzel and Kurjan, 1987;Miyajima et al., 1987;Jahng et al., 1988) that prepares the cell for conjugation (Hartwell, 1973). The inhibition of start by a-factor supersedes any biosynthetic control over start and takes place with little effect on global biosynthetic activity (Throm and Duntze, 1970). In contrast, other start inhibitors, within minutes after addition to cultures of growing cells, bring about significant biosynthetic inhibition. In most cases inhibition is significant for RNA * This work was supported by the Medical Research Council of Canada. The costs of publication of this 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.
11 To whom correspondence should be addressed. synthesis with only modest effects on protein synthesis Singer et al., 1978;Singer and Johnston, 1979;Bedard et al., 1980). Indirect tests have indicated that in some of these start arrest situations the inhibition of rRNA or pre-rRNA production may account for most of the effects on RNA synthesis Singer et al., 1978;Singer and Johnston, 1979;Bedard et al., 1980).
In the study described here a quantitative hybridization analysis of pulse-labeled RNA (Veinot-Drebot et al., 1988) was used to measure, for yeast rRNA genes, the transcriptional perturbations caused by six different start inhibitors, including a-factor. With the exception of a-factor, each inhibitor caused rapid and substantial inhibition of pre-rRNA transcription. Northern analysis indicated that these inhibitors also slowed cleavage of the pre-rRNA primary transcript , a perturbation known to be without influence on the control of yeast cell proliferation . Only one inhibitor markedly induced the general control response (Hinnebusch, 1988). A uniform correlation was therefore found between inhibition of pre-rRNA transcription and inhibition of start.

MATERIALS AND METHODS
Strains and Media-Cells of S. cerevisiae strain GR2 (MATa his6 ural;  were grown with gyratory shaking at 22 "C in YNB minimal medium (Johnston et al., 1977) supplemented either with histidine and uracil (40 pg/ml) for transcription rate studies or with histidine, uracil, adenine, arginine, leucine, lysine, tyrosine, and tryptophan for Northern analysis. Cell growth was monitored with an electronic particle counter (Coulter Electronics, Hialeah, FL;Hartwell, 1970 A. Hinnebusch, NIH). Plasmids were prepared as described (Messing, 1983). RNA Labeling, Extraction, and Quuntitation-Procedures for RNA labeling were as described (Veinot-Drebot et al., 1988). Cells were grown for several generations in medium containing ["C]uracil (0.01 pCi/ml; Du Pont-New England Nuclear) to label stable RNA species and thus facilitate quantitation of RNA yields. For pulse labeling, a 2-ml sample of the "C-labeled culture, at 3-5 X lo6 cells/ml, was incubated with [3H]adenine (100 WCi/ml; 21.6 Ci/mmol; Du Pont-New England Nuclear) for a 2-min period and then diluted with 3 volumes of ice-cold water and subjected immediately to RNA extraction by the following modification of published procedures Li et al., 1985). Cells were rapidly harvested by centrifugation, suspended in 0.5 ml of NETS buffer , and vortexed along with 1 ml of acid-washed baked ' The abbreviations used are: ETS, external transcribed spacer; ITS, internal transcribed spacer. glass beads (BT5, 0.45-mm diameter; Flex-0-Lite Division, General Steel Industries) for 45 s to disrupt the cells. The resulting suspension was then supplemented with 4.5 ml of NETS buffer and extracted once with 4.5 ml of pheno1:cresol:hydroxyquinoline (50070:0.5) and once with 5.5 ml of the same organic mixture plus 0.5 ml of 5 M Licl.
The aqueous phase was subjected to centrifugation to remove particulate material and precipitated with ethanol at -20 "C. Precipitate was suspended in 0.5 ml of NETS buffer containing 0.5 M LEI, reextracted with an equal volume of the same organic mixture, precipitated twice with ethanol, and used for hybridizations.
For quantitation of RNA, a sample of "C-labeled total RNA was prepared for each time point from cells treated identically but not pulse-labeled. RNA in this sample was quantified from AZM) and A280 values to determine "C specific activity ("C cpmlpg total RNA); from this value and the levels of incorporated "C and 3H in the doubly labeled parallel sample, 3H specific activity (3H cpm/pg total RNA) was calculated.
Fractions enriched in poly(A)+ RNA were prepared from total labeled RNA by affinity chromatography (Kim and Warner, 1983).
Quantitative Hybridization-The standard procedure for hybridization (Zitomer et al., 1979;Kim and Warner, 1983) was used. To immobilize DNA, 40 pg of plasmid DNA was made to 0.2 mg/ml with 0.1 N NaOH, heated to 100 "C, diluted with 20 volumes of 1 M ammonium acetate and made to pH 7.0 with 0.1 N NaOH, and then loaded under low vacuum onto a 25-mm nitrocellulose filter disc (0.22-mm pore size; Millipore GS) that had been soaked in 6 X SSC (1 X SSC is 0.15 M NaCl, 0.015 M sodium citrate; Maniatis et al., 1982). Each loaded disc was washed twice with 6 X SSC, divided into four equal pieces, and baked at 80 "C under vacuum for 2 h.
Hybridization of [3H]RNA to immobilized DNA was carried out in capped plastic scintillation vials at 40 "C for 69 h. Each vial contained 0.5 ml of hybridization buffer (Kim and Warner, 1983), one filter piece containingpLD11 DNA, one filter piece containingpUC8 DNA, and 5 pg of total [3H]RNA. Following hybridization, filters were washed at 40 "C first with 2 X SSC containing 40% deionized formamide and 2 pg of RNase A (preboiled; Sigma)/ml, then with 2 X SSC containing 40% deionized formamide, and then 10 times with 2 X SSC containing 0.2% sodium dodecyl sulfate. Filters were then washed at 22 "C once with 2 X SSC and finally twice with 95% ethanol. Bound radioactivity was measured by liquid scintillation in a toluene-based mixture. The extent of hybridization determined in this way increased linearly up to at least 75 pg of total RNA per reaction. The routine use of 5 pg of total RNA per reaction ensured DNA excess conditions. ATP Pool Specific Activity Determination-Cells were grown for several generations in medium containing 32P04 (2 pCi/ml; Du Pont-New England Nuclear) to uniformly label cellular ATP. From cultures in the range of 2-5 X lo6 cells/ml, 1-ml samples were incubated for 2-min periods with [3H]adenine (5 pCi/ml). Cells were then quickly harvested by centrifugation and immediately extracted at 4 "C overnight with 50 p1 of 1 M formic acid. The contents of clarified extracts were resolved, along with nucleoside triphosphate standards, by thin layer chromatography (Cashel et al., 1969) on impregnated 0.1-mm cellulose MN300 polyethyleneimine plates (Brinkmann Instruments).
The 3H/32P ratio in ATP was determined by liquid scintillation counting (in a mixture of 1.5 ml of water and 10 ml of universal liquid scintillation mixture) of material from the center of the ATP spot. GTP was not routinely quantitated; preliminary work showed that insignificant [3H]GTP was formed under these labeling conditions.
Northern Analysis-Equal amounts of total RNA, extracted (Penn et al., 1984) and quantified spectrophotometrically, were resolved by formaldehyde-agarose gel electrophoresis and transferred to Gene-Screen hybridization membrane as described by Maniatis et al. (1982) and by the manufacturer (Du Pont-New England Nuclear). After transfer, blots were hybridized with plasmid DNA radiolabeled by use of a nick translation kit (Bethesda Research Laboratories), washed, and exposed to Kodak XAR-5 x-ray film.

RESULTS
Proliferating cells were treated with start inhibitors at concentrations that led in each case to first-cycle arrest of cell proliferation, with over 80% of each arrested cell population in the unbudded phase of the cell cycle (data not shown). For each inhibitor the unbudded cells in an arrested population have been shown to be at start (see the Introduction). The high proportions of unbudded cells verified that in each situation studied here the inhibitor-treated cells promptly became blocked at start.
Rates of transcription of the rRNA genes were then quantified by DNA excess hybridization. In this procedure, which has been used to measure transcription rates for proteincoding genes (Zitomer et al., 1979;Osley and Hereford, 1981;Kim and Warner, 1983), labeled RNA is hybridized to immobilized DNA sequences available in excess. When the rate of incorporation of labeled precursor into RNA is proportional to the transcription rate, the relative specific activity of the labeled RNA can be used to quantify transcription rates for the sequence of interest.
To determine transcription rates for the yeast rRNA genes, a cloned portion (Veinot-Drebot et al., 1988) of the yeast rDNA repeat was used that encodes the pre-rRNA transcription initiation site and most of the ETS of the rRNA primary transcript (Klemenz and Geiduschek, 1980;Bayev et al., 1980). The ETS RNA is the first portion of each 35 S pre-rRNA molecule synthesized. Therefore the rate of synthesis of ETS RNA is proportional to the rate of initiation of pre-rRNA transcripts.
Technical Considerations-The half-life of transcripts that contain ETS sequences is exceedingly short (Veinot-Drebot et al., 1988). Therefore the pulse-labeling time was optimized to quantify rRNA synthesis without interference by ETS sequence degradation by applying the approach used to select an appropriate labeling time to measure rates of synthesis of cytochrome c mRNA (Zitomer et al., 1979). In this approach the specific activity of the transcript of interest is compared with the specific activity of total RNA for different labeling times.
Growing cells were pulse labeled for periods ranging from 0.5 to 10 min, and total RNA was extracted and hybridized to ETS DNA. During these short labeling periods the bulk of the newly synthesized RNA is not expected to be degraded; much of the total RNA synthesized at any one time comprises sequences destined to become mature stable rRNA and tRNA molecules, and most yeast mRNA transcripts have half-lives of 16-22 min (Peterson et al., 1976;Chia and McLaughlin, 1979). As shown in Fig. 1, the specific activity of hybridizable ETS sequences decreased for labeling times over 2 min, reflecting the short half-life of ETS sequences in yeast (Veinot-Drebot et al., 1988). A labeling time of 2 min was therefore chosen for subsequent experiments. This optimal labeling time still underestimates the rate of synthesis because of the short half-life of ETS RNA but allows quantitation of labeled transcripts with minimal interference from ETS degradation. No matter what the rate of RNA synthesis, conditions that alter the size of the ribonucleoside triphosphate pools will as a consequence affect the rate of incorporation of exogenous labeled precursor molecules into RNA and thus affect the apparent rate of transcription. Therefore pool size was taken into consideration; each time RNA was pulse-labeled with [3H]adenine for hybridization, the relative specific activity of the total cellular ATP pool was also determined for a parallel culture that had been labeled continuously with 32P (see "Materials and Methods"). The changes in relative specific activity (defined as the ratio of [3H]ATP from pulse labeling to [3ZP]ATP from continuous long term labeling) were used to correct for pool effects on label incorporation. (Parallel cultures were necessary because the procedure used to pulse label ATP was incompatible with that used to label RNA to high specific activity for hybridization; see "Materials and Methods.") At each labeling time the relative rate of label incorporation into total RNA was also determined for each parallel culture to verify similar responses of the parallel cultures to the same treatment.
Transcriptional Effects of a-Factor Treatment-Cells treated with the mating pheromone a-factor continue to enlarge as mass accumulation continues in the absence of bud formation (Throm and Duntze, 1970). Accordingly, it was found by pulse labeling that total RNA synthesis rates actually increased in a-factor-treated cells (data not shown). Part of this increase was due to increased rates of synthesis of mRNA, represented in the poly(A)+ RNA fraction (Fig.  2 4 ) . Mating factor treatment induces the expression of certain protein-coding genes (Hagen and Sprague, 1984;Hartig et al., 1986;Jahng et al., 1988;Kronstad et al., 1987;McCaffrey et al., 1987;Nakayama et al., 1985;Orlean, 1987;Trueheart et al., 1987), which could account for at least part of this increased rate of poly(A)+ RNA synthesis. In these a-factortreated cells there also was significantly increased transcription initiation for the rRNA genes, as indicated by increased hybridization of pulse-labeled RNA to immobilized ETS DNA ( Fig. 2 A ) . The rRNA genes are therefore included in the set of genes that are modulated by mating factor treatment.
Transcriptional Effects of Other Start Inhibitors-In con- trast to the continued transcription seen upon treatment with a-factor, there was significant inhibition of transcription caused by every other inhibitor of start. More specifically, each inhibitor produced substantial inhibition of pre-rRNA transcription.
The effects on pre-rRNA transcription caused by these inhibitors are exemplified by the effects of the start inhibitor L-ethionine . Most of the transcriptional effects following L-ethionine treatment were limited to inhibition of transcription of the rRNA genes (Fig. 2B). There were only minor effects on the transcription of poly(A)+ RNA (Fig. ZB), consistent with the continued protein synthetic activity of cells treated with L-ethionine . Therefore L-ethionine is an inhibitor of rRNA gene transcription.
AS shown in Fig. 2C, even more substantial inhibition of pre-rRNA transcription was found for the start inhibitor ophenanthroline . Total RNA synthesis was also inhibited, but little of the o-phenanthroline-mediate decrease in the total RNA synthesis rate could be ascribed to inhibition of mRNA synthesis; in fact, the rate of synthesis of RNA in the poly(A)+ fraction was actually increased 1.7-fold at 15 min after inhibitor addition (data not shown), and protein synthesis has been shown to continue in o-phenanthroline-treated cells . Furthermore, the inhibition of total RNA synthesis, of which about half is pre-rRNA (Shulman et al., 1977), was only half that of pre-rRNA transcription (data not shown), indicating that virtually all of the decrease in total RNA synthesis rate was accounted for by the inhibition of pre-rRNA transcription.
The inhibitory effects produced by the start inhibitor 8hydroxyquinoline  were less specific. Although there was rapid and remarkable inhibition of the rate of transcription of the rRNA genes ( Fig. 2 0 ) , there was also similar inhibition of the synthesis rate for total RNA (data not shown), implying that mRNA synthesis was similarly inhibited.
Two other inhibitors of start also produced more general inhibition of transcription, similar to the effects of 8-hydroxyquinoline. Nalidixic acid (Singer and Johnston, 1979) inhibited rRNA gene transcription efficiently ( Fig. 2E) but also inhibited the synthesis of poly(A)+ RNA to a significant extent (Fig. 2E). These effects are consistent with the degree of inhibition of 35 S pre-rRNA accumulation and of amino acid incorporation for nalidixic acid-treated cells (Singer and Johnston, 1979). Sinefungin  had general inhibitory effects on transcription that were less severe than those of nalidixic acid. Not only was rRNA gene transcription inhibited by sinefungin (Fig. 2F), but poly(A)+ RNA synthesis was inhibited (Fig. 2F) to a degree consistent with the reported inhibition of amino acid incorporation .
Inhibition of Pre-rRNA Cleauage-The assay described above indicates the rate of transcription of the rRNA genes but not the fate of the pre-rRNA transcript after synthesis. Therefore the ETS sequences used to measure transcription rates were also used in Northern analysis to measure levels of pre-rRNA transcript in inhibitor-treated cells. From the levels of 35 S pre-rRNA and the extent of inhibition of pre-rRNA transcription quantified above, the effectiveness of the initial cleavage step in the conversion of 35 S pre-rRNA to mature rRNA molecules was deduced.
In cells treated with a-factor the amounts of 35 S pre-rRNA present in cells increased somewhat (Fig. 3 A ) . This increase is consistent with the increased rate of pre-rRNA transcription in a-factor-treated cells (Fig. 2 A ) . Analogous increases in the levels of 35 S pre-rRNA were also seen in cells treated with the start inhibitors o-phenanthroline and 8-hydroxyquinoline ( Fig. 3A) but for a different reason. In these situations the increased levels of 35 S pre-rRNA, seen in parallel with significant inhibition of pre-rRNA transcription (Fig. 2, C and D), do not reflect increased rates of synthesis but instead must result from significantly inhibited rates of initial cleavage of pre-rRNA by these start inhibitors. These findings show that the effects on pre-rRNA metabolism of both ophenanthroline and 8-hydroxyquinoline are more extensive than just the inhibition of pre-rRNA transcription.
The inhibition of initial cleavage of pre-rRNA brought about by these two inhibitors must approximate the degree of inhibition of pre-rRNA transcription initiation. This conclusion stems from kinetic considerations. The intact 35 S pre-rRNA molecule is produced in great quantity but has a very short half-life (approximately 0.5 min under these conditions; Veinot-Drebot et al., 1988), so that for these molecules even a slight perturbation in the relationship between the rate of synthesis and the rate of loss (initial cleavage) would be readily evident as a large change in the amount of intact pre-rRNA. The modest increases in the levels of 35 S pre-rRNA brought about by treatment with o-phenanthroline or 8-hydroxyquinoline and the relative constancy of this increase over time (Fig. 3A) suggest that under these start arrest conditions the degree of inhibition of initial cleavage approximates the inhibition of transcription initiation seen in Fig.  2.
In contrast to the comparable effects on pre-rRNA transcription and initial cleavage caused by o-phenanthroline and 8-hydroxyquinoline, the effects on pre-rRNA cleavage brought about by the other start inhibitors were more severe. For example, Northern analysis indicated that L-ethionine treatment markedly impeded the cleavage of pre-rRNA at many steps. Fig. 3A shows that 35 S pre-rRNA molecules accumulated significantly in cells treated with L-ethionine, indicating that the initial cleavage step that removes ETS sequences from the pre-rRNA transcript was inhibited to an even greater degree than the already substantial inhibition of transcription of the RNA (Fig. 2B). Furthermore, analysis of the same RNA blot with another probe specific for sequences within the first ITS in the pre-rRNA transcript indicated that ITS-containing intermediates of pre-rRNA cleavage also accumulated in L-ethionine-treated cells (Fig. 3B). A similar pattern of accumulation of both 35 S pre-rRNA and ITScontaining intermediates was seen for cells treated with the start inhibitor sinefungin (Fig. 3, A and B ) . An accumulation of 35 S pre-rRNA was also noted for cells treated with the start inhibitor nalidixic acid (Fig. 3A).
Therefore each inhibitor of the start regulatory step for cell proliferation that also inhibits the transcription of pre-rRNA also inhibits at least the initial cleavage step in the conversion of this transcript into mature rRNA. This inhibition of pre-rRNA cleavage, by stabilizing ETS RNA, also leads to overestimation in the assay described here of the rate of ETS RNA synthesis.
Induction of the General Control Response--In addition to effects on pre-rRNA transcription, other transcriptional effects have been associated with some of these start inhibitors. L-Ethionine has been shown (Wolfner et al., 1975;Penn et al., 1983) to be an amino acid analog that can induce the general control response, a regulatory system governing the expression of genes involved in the biosynthesis of amino acids (Hinnebusch, 1988). Certain conditional mutations that affect general control also inhibit start (Wolfner et al., 1975) or bring about unbudded cell arrest (Harashima et al., 1987)) suggesting the possibility that the general control response may be involved in the start inhibition produced by some of these inhibitors. For this reason the general control response was monitored in cells treated with inhibitors at concentrations that caused start arrest.
One characteristic feature of the general control response is induction of the HIS4 gene, as indicated by increased levels of HIS4 mRNA (Donahue et al., 1983; Penn et al., 1984). Therefore HIS4 mRNA levels were determined after the addition of inhibitors and compared to levels of the URA3 transcript; the URA3 gene is not subject to general control -HIS4+ ( . 4. HIS4 and URA3 transcript levels. Equal amounts of total RNA extracted from cells at the indicated times after treatment with start inhibitors were resolved by electrophoresis and probed for HIS4 and URA3 transcripts. Inhibitor designations are as in Fig. 3. " _ 7" (Silverman et al., 1982;Penn et al., 1983). Transfer of the histidine auxotrophic strain used here from supplemented to histidine-free medium caused the expected increased in HIS4 transcript levels (data not shown), For cells of the same strain in the same supplemented medium, the inhibitors at concentrations that cause start arrest produced differing degrees of activation of the general control response. The inhibitor ophenanthroline caused significant induction of HIS4 transcript levels with little effect on URA3 transcript levels (Fig.  4), while 8-hydroxyquinoline had no effect on HIS4 transcript levels (Fig. 4). Other inhibitors, including L-ethionine and nalidixic acid, caused more modest and delayed increases in the relative levels of HIS4 mRNA (Fig. 4), effects similar to but more pronounced than those produced by a-factor treatment (Fig. 4).
These observations sustain the possibility that the general control response could be involved in the start arrest produced by o-phenanthroline treatment. In contrast, for L-ethionine and the other inhibitors at concentrations that bring about start arrest this involvement may not be significant.

DISCUSSION
The mechanisms that make the performance of start and thus all the activities of yeast cell proliferation responsive to the biosynthetic status of the cell are subject to investigation through the use of inhibitors of start. The "natural" inhibitors of start, the mating pheromones, apparently uncouple start from the biosynthetic status of the cell, leaving these activities generally unaffected (Throm and Duntze, 1970; Fig. 2 A ) . Thus other start inhibitors must be used to delineate biosynthetic activities that influence the performance of start. Some of these inhibitors have been studied here.
Many start inhibitors, identified solely by effects on cell proliferation, also affect RNA metabolism (see the Introduction). Here we document that, in fact, each start inhibitor affects expression of the rRNA genes. The mating pheromone a-factor transiently increased transcription of pre-rRNA. In contrast, every other inhibitor of start inhibited the transcription and the cleavage of pre-rRNA. It is notable in this context that heat shock, a treatment that causes transient start arrest , also causes transient inhibition of both transcription (Veinot-Drebot et al., 1989) and cleavage (Shulman and Warner, 1978) of pre-rRNA in yeast. For each inhibitor the kinetics of inhibition of pre-rRNA transcription and cleavage paralleled the RNA inhibition kinetics determined earlier by indirect means. It was also shown that in addition to the effects on pre-rRNA, most of these start inhibitors also had other effects on RNA metabolism; ophenanthroline induced the general control response while 8hydroxyquinoline, nalidixic acid, and to some extent sinefungin inhibited poly(A)+ RNA synthesis generally. Furthermore, a previous study showed that sinefungin retards the methylation of rRNA, leading to a differential loss of undermethylated 18 S rRNA (Li et al., 1985). Each of these inhibitory effects reflected in perturbed RNA metabolism precedes or coincides with the inhibition of start and thus could in principle participate in start regulation.
Despite the correlations just described, some of these inhibitory effects can be eliminated from consideration as possible influences on the performance of start. For example, other studies show that inhibition of the cleavage of pre-rRNA is not itself responsible for start arrest. This conclusion is evident from the effects of an ma2 mutation in yeast . Temperature-sensitive mutations in the RNA2 gene block intron removal from primary transcripts (Rosbash et al., 1981;Bromley et al., 1982;Teem et al., 1983) that code for ribosomal proteins (Rosbash et al., 1981;Fried et al., 1981). The resultant rm2-mediated inhibition of the synthesis of ribosomal proteins also inhibits the synthesis of proteins necessary for proper cleavage of the pre-rRNA primary transcript. As a consequence there is significant inhibition of pre-rRNA cleavage (Warner and , but this inhibition does not affect the performance of start; ma2 mutant cells initiate new cell cycles at the nonpermissive temperature . This ability of cells blocked in pre-rRNA cleavage to perform start suggests that the inhibition of pre-rRNA cleavage caused by the start inhibitors studied here (Fig. 3) does not affect start.
The induction of the general control response by o-phenanthroline is also probably not involved in either the inhibition of pre-rRNA transcription or the start arrest caused by this inhibitor. In other situations that induce the general control response such as growth of prototrophic cells in certain supplemented minimal media (summarized by Hinnebusch, 1988), treatment of cells with amino acid analogs (Wolfner et al., 1975;Hinnebusch and Fink, 1983;Penn et al., 1983;Lucchini et al., 1984), or the presence of gcd mutations under permissive conditions (Wolfner et al., 1975;Harashima and Hinnebusch, 1986), there is ongoing cell proliferation, indicating that rRNA production and the performance of start continue. Conversely, inhibition of pre-rRNA transcription and cleavage does not necessarily induce the general control response, as shown by the effects of 8-hydroxyquinoline (Fig.  4). In this particular situation the absence of HIS4 induction was not simply a result of the general inhibition of poly(A)+ mRNA synthesis caused by 8-hydroxyquinoline treatment (see above). These inhibited cells retain significant biological potential, as indicated by the efficient conjugation of haploid yeast cells treated with 8-hydroxyquinoline (or with o-phenanthroline) (Bedard et al., 1984), and efficient conjugation in turn requires gene induction (Trueheart et al., 1987). Therefore induction of the general control response is unrelated to the inhibition of pre-rRNA transcription or cleavage. These findings also show that during conjugation there is little need for continued rRNA production.
In addition to the RNA effects, most of the inhibitors tested also caused ATP pool perturbations (Fig. 2) that were similar to the effects on pyrimidine triphosphate pool specific activity noted earlier for some of the inhibitors (Singer and Johnston, 1979;. The decrease in ATP pool specific activity cannot easily be ascribed to expansion of ATP pools from turnover of stable RNA ; therefore some of these inhibitors may also release sequestered nucleotides or inhibit uptake of exogenous precursors. Whether these postulated effects influence transcription in the uracil auxotrophic cells used in these experiments remains to be determined. However, ATP pool perturbations do not always accompany start arrest (see Fig. 2H), and start inhibition is unrelated to purine or pyrimidine auxotrophf; both of these observations suggest that nucleoside triphosphate pool perturbations are not involved in the start arrest caused by these inhibitors.
The biosynthetic effects of start inhibitors were assessed here at concentrations that produce start inhibition rather than at concentrations optimized for biosynthetic effects. It remains to be seen if other inhibitor concentrations cause more selective biosynthetic inhibition. It is striking, therefore, that of the parameters assayed for these inhibitors of start only the inhibition of pre-rRNA transcription could be implicated in start inhibition. Even though most of the inhibitors studied here inhibited both pre-rRNA transcription and cleavage, it is unlikely that the inhibitors act through similar mechanisms. This conclusion is illustrated by a comparison of the effects of two of the inhibitors studied here, L-ethionine and sinefungin, with the similar effects of heat shock. Upon heat shock, the inhibition of cleavage leading to accumulation of 35 S pre-rRNA (Veinot-Drebot et al., 1988) accompanies the decreased production of ribosomal proteins (Kim and Warner, 1983). The absence of some of these ribosomal proteins probably inhibits the initial cleavage reaction. In contrast, the two start inhibitors affect the RNA substrate for cleavage, which is methylated pre-rRNA. Both L-ethionine and sinefungin cause undermethylation of rRNA Li et al., 1985), indicating that the methylation of pre-rRNA Klootwijk and Planta, 1973) is decreased by these inhibitors. The accumulation of 35 S pre-rRNA molecules caused by L-ethionine and sinefungin (Fig. 3A) suggests that these undermethylated pre-rRNA molecules may be inefficient substrates for cleavage even in the presence of ribosomal proteins. Therefore these observations suggest that at least with respect to the initial cleavage of pre-rRNA, the start inhibitors L-ethionine and sinefungin work through a different mechanism than heat shock to yield the same general phenotype.
The quantitative features of the dual inhibitions of pre-rRNA transcription and of cleavage also suggest that the inhibitors studied here affect rRNA synthesis through more than one mechanism. For example, both o-phenanthroline and 8-hydroxyquinoline inhibit transcription and cleavage in a relatively balanced way, as indicated by the only modest accumulation of 35 S pre-rRNA in inhibitor-treated cells (Fig.  3A). Therefore each of these inhibitors decreases pre-rRNA transcription to roughly the same extent as cleavage and in comparison with the effects of L-ethionine and sinefungin must affect rRNA production differently.
The inhibition of start that is produced by o-phenanthroline and by 8-hydroxyquinoline is abrogated by zinc ions (Johnston and , in keeping with the known actions of these inhibitors as zinc chelators. Although it has been shown that certain metal chelates of o-phenanthroline can cause single-strand nicks in DNA (Marshall et al., 1981), this DNAnicking activity is not a consideration in this in vivo situation because DNA damage in yeast brings about arrest in G2 (Brunborg and Williamson, 1978;Jordan and Laskowski, 1987) rather than the start arrest seen here. Furthermore, the ability of ferrous ions as well as zinc ions to abrogate inhibition by o-phenanthroline and 8-hydroxyquinoline (Singer and Johnston, 1982) suggests that in chelated configuration these inhibitors are inactive. All these observations suggest that the inhibition caused by o-phenanthroline and 8-hydroxyquinoline may be due to chelation of zinc ions, perhaps the zinc ions found in yeast RNA polymerase I (Auld et a[., 1976;Memet et al., 1988) or in transcription factors (Evans and Hollenberg, 1988).