Cytokinin induction of RNA polymerase I transcription in Arabidopsis thaliana.

RNA polymerase I (pol I) transcribes the repeated genes that encode the precursor of 17-18, 5.8, and 25-28 S ribosomal RNA (rRNA). Pol I transcription is up-regulated in growing cells and down-regulated in quiescent cells, presumably reflecting the demand for ribosomes and protein synthesis. However, the signal transduction pathways responsible for pol I regulation are poorly understood. We tested the effects of exogenously applied plant hormones on promoter-dependent rRNA transcription in Arabidopsis thaliana Gibberellic acid, abscisic acid, auxin, and ethylene had no detectable effect on rRNA transcription, but kinetin (a cytokinin) stimulated rRNA transcription within 1 h of treatment. Increased steady-state levels of accurately initiated rRNA transcripts, detected by S1 nuclease protection, were paralleled by increased levels of nascent rRNA transcripts in isolated nuclei. Therefore, the primary effect of cytokinin appears to be at the level of transcription initiation rather than rRNA stability. Pol I accounts for ∼34% of total nuclear transcription in untreated plants and ∼60% following cytokinin treatment. The specific responsiveness of pol I transcription to kinetin suggests that cytokinins may act as general regulators of protein synthetic capacity and growth status in plant cells.

RNA polymerase I (pol I) transcribes the repeated genes that encode the precursor of 17-18, 5.8, and 25-28 S ribosomal RNA (rRNA). Pol I transcription is up-regulated in growing cells and down-regulated in quiescent cells, presumably reflecting the demand for ribosomes and protein synthesis. However, the signal transduction pathways responsible for pol I regulation are poorly understood. We tested the effects of exogenously applied plant hormones on promoter-dependent rRNA transcription in Arabidopsis thaliana. Gibberellic acid, abscisic acid, auxin, and ethylene had no detectable effect on rRNA transcription, but kinetin (a cytokinin) stimulated rRNA transcription within 1 h of treatment. Increased steady-state levels of accurately initiated rRNA transcripts, detected by S1 nuclease protection, were paralleled by increased levels of nascent rRNA transcripts in isolated nuclei. Therefore, the primary effect of cytokinin appears to be at the level of transcription initiation rather than rRNA stability. Pol I accounts for ϳ34% of total nuclear transcription in untreated plants and ϳ60% following cytokinin treatment. The specific responsiveness of pol I transcription to kinetin suggests that cytokinins may act as general regulators of protein synthetic capacity and growth status in plant cells.
Eukaryotes have evolved three nuclear RNA polymerases with specialized functions. RNA polymerase I transcribes the tandemly repeated genes encoding the precursor of 17-18, 5.8, and 25-28 S ribosomal RNA (1-3) (rRNA size varies with species). Pol 1 II transcribes protein-coding genes and most small nuclear RNAs (4 -6), and pol III transcribes 5 S rRNA, tRNAs, and one or more small nuclear RNAs (6 -8). All three polymerases are needed to produce the rRNAs and proteins of a functional ribosome (9,10), although the rRNAs transcribed by pol I are thought to comprise the ribosome's catalytic core (11).
Ribosomal RNA transcription and cell growth rate tend to be positively correlated (12)(13)(14)(15)(16). In Escherichia coli and other prokaryotes, the cellular concentration of guanosine tetraphosphate (ppGpp), reflects the level of rRNA transcription, al-though cause and effect are unclear (17). In eukaryotes, there is no strong evidence for an analogous compound. However, as in prokaryotes, rRNA transcription is down-regulated upon nutritional challenge (amino acid, carbon, or nitrogen limitation; serum starvation), cycloheximide treatment, or upon reaching stationary phase (18 -24; recently reviewed in Ref. 12). Glucocorticoid hormones have also been shown to stimulate or repress pol I transcription in different tissues (25)(26)(27)(28). These physiological responses appear to be brought about by phosphorylation or other modifications of the polymerase or its auxiliary transcription factors (21, 22, 29 -31), but the causative signal transduction pathways remain obscure.
Using the model plant species, Arabidopsis thaliana, we have begun to define the rRNA gene loci and cis-acting DNA sequences essential for accurate pol I transcription initiation (32)(33)(34)(35)(36). Recent progress toward dissecting hormone signaling pathways in A. thaliana prompted us to investigate reports that plant hormones such as auxin and cytokinin affect pol I activity in cell-free extracts, isolated chromatin, or nuclei (37)(38)(39)(40)(41). Consequently, we examined hormonal effects on transcription initiation at the rRNA gene promoters as well as elongation of nascent rRNA transcripts. We show that exogenous cytokinin induces pol I transcription within 1 h of treatment, and the stimulatory effect persists for at least 24 h. The effect of cytokinin on rRNA transcription is consistent with its role in growth regulation and cell division.

EXPERIMENTAL PROCEDURES
Sterile Growth of Arabidopsis Seedlings-A. thaliana Columbia seeds were sterilized in 95% ethanol (five washes, 5 min each), followed by 3% sodium hypochlorite wash (10 min) and three washes in sterile distilled water (5 min each). Seeds were recovered by filtration (Nalgene sterile filter unit) and dried for 2 days in a desiccator containing sterile Dri-Rite. Approximately 100 seeds were sprinkled onto semisolid germination medium (42) in 75-mm diameter Petri dishes. Seeds were germinated and plantlets grown for 14 -21 days in a growth chamber (16 h day length, 22°C, 70% relative humidity) prior to hormone treatment.
Hormone Treatment of Plants-Plant hormones were purchased from Sigma and dissolved as 0.1 M -0.2 M stocks. Kinetin was initially dissolved in 1 N NaOH then diluted with water to 0.1 M. Stocks of abscisic acid, gibberellic acid, and 2,4-dichlorophenoxyacetic acid were prepared in 100% ethanol. All hormone stocks were made just prior to use and filter-sterilized. Plants were sprayed to run-off with a fine mist, essentially according to Guilfoyle (37) and harvested by pulling them from the agar with forceps. Plantlets were frozen in liquid nitrogen for RNA isolation or homogenized immediately for nuclei isolation.
RNA Isolation-For S1 nuclease assays, RNA was isolated according to Chirgwin et al. (43) with minor modifications. RNA was resuspended in 400 l of diethyl pyrocarbonate-treated water and extracted twice with an equal volume of phenol:chloroform:isoamyl alcohol (25:24:1; v/v) followed by extraction with 1 volume of chloroform:isoamyl alcohol (24:1, v/v). Sodium acetate (pH 5.2) was added to 0.25 M and nucleic acids precipitated by addition of 2.5 volumes of 100% ethanol and centrifugation at 14,000 ϫ g for 20 min (4°C). Pellets was resuspended in 300 l of diethyl pyrocarbonate-treated water followed by addition of 600 l of ice-cold 4 M lithium chloride to precipitate RNA. Tubes were vortexed and incubated overnight on ice. Following centrifugation as above, pellets were resuspended in 100 l of diethyl pyrocarbonatetreated water and quantified by absorbance at 260 nm. All RNA samples were also checked by agarose gel electrophoresis and ethidium bromide staining to verify their quality and equivalent concentrations. S1 Nuclease Protection-rRNA transcripts were detected using S1 nuclease protection (44) and 5Ј end-labeled probes as described previously (32,33). Ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) small subunit mRNAs were detected using an oligonucleotide perfectly complementary to the A. thaliana Ats3B clone, but capable of hybridizing to transcripts of all four (known) genes (45). The sequence of the oligonucleotide is 5Ј CCACAGCGGCGGAGGAGAGCATAGAG-GAAGCCATTACTACTTCTTGTTGGCGCGC 3Ј; the underlined nucleotides at the 3Ј end are not homologous to any Rubisco mRNAs and are removed by S1 nuclease, discriminating the longest protected fragments from undigested probe.
Nuclei Isolation-Nuclei were isolated according to Feinbaum and Ausubel (46,47) with modifications. Arabidopsis plants (ϳ800) were harvested from 8 to 10 agar plates, washed in ice-cold sterile water, and kept at 4°C. Plants were submerged in cold diethyl ether for 5 min, then washed twice (5 min each) in ice-cold sterile water. Plants were homogenized in 100 ml (approximately 3 volumes) of nuclei isolation buffer (1 M sucrose, 50 mM Tris-HCl (pH 7.2), 5 mM MgCl 2 , 5 mM KCl 2 , 10 mM 2-mercaptoethanol, 0.2 mM PMSF) with a motorized homogenizer (Fisher Scientific Powergen 700). The homogenate was filtered through eight layers of cheesecloth, one layer of Miracloth (Calbiochem), and one layer of 50-m nylon mesh. The filtrate was centrifuged at 14,000 ϫ g for 15 min at 4°C. The pellet was resuspended in 10 ml of nuclei isolation buffer using a 15-ml Dounce homogenizer (A pestle), and the final volume was measured. One volume of "100%" Percoll solution (34.23 g of sucrose, 5 ml of 1 M Tris-HCl (pH 7.2), 0.5 ml of 1 M MgCl 2 , 0.5 ml 1 M KCl, 34 l of 2-mercaptoethanol, 100 l of 2 mM PMSF, and Percoll to 100 ml) was added to 19 volumes of nuclei. 7 ml of crude nuclei in 5% Percoll solution was then layered onto a four-step discontinuous Percoll gradient with 5-ml layers of 15, 30, 45, and 60% Percoll solutions (diluted from the 100% Percoll solution using nuclei isolation buffer). Gradients were centrifuged at 2000 rpm for 10 min, then 8000 rpm for 20 min in a Beckman SW 28 rotor at 4°C. Nuclei were harvested from the 30%/45% and 45%/60% boundaries and pooled as both were found to be transcriptionally equivalent. Nuclei were diluted with 5 volumes of nuclei isolation buffer, mixed by inversion, and collected by centrifugation at 1500 ϫ g, 10 min, 4°C in a clinical centrifuge (Beckman). Nuclei were resuspended gently in 50 ml (approximately 25 volumes) of nuclei isolation buffer and again collected by centrifugation. Final resuspension was in 1 ml of nuclei storage buffer (50 mM HEPES (pH 7.2), 5 mM MgCl 2 , 5 mM KCl, 2 mM dithiothreitol, 0.2 mM PMSF, 50% glycerol). Aliquots were transferred to prechilled 1.5-ml microcentrifuge tubes, frozen in liquid nitrogen, and stored at Ϫ80°C.
Nuclear Run-on-Frozen nuclei were thawed on ice, stained with 4Ј,6-diamidino-2-phenylindole, and counted in a hemacytometer using fluorescence microscopy. Nuclear run-ons for filter hybridizations involved 1 ϫ 10 6 nuclei; for total incorporation experiments, 1 ϫ 10 5 nuclei were used. All nuclear run-on reactions were for 10 min (unless otherwise stated) at 30°C in a final volume of 50 l (40 l of run-on reaction buffer, 10 l of nuclei). Nuclear run-on reaction buffer was: 50 mM Tris-HCl (pH 7.2), 5 mM MgCl 2 , 5 mM KCl, 75 mM (NH 4 ) 2 SO 4 , 1 mM MnCl 2 , 300 M each ATP, GTP, and UTP, 4.5 M CTP, 0.5 M [␣-32 P]CTP, 5 mM dithiothreitol, 0.2 mM PMSF, 2 units of RNasin RNase inhibitor (Promega), 10% glycerol, and 150 g/ml ␣-amanitin (increasing the amanitin level to 500 g/ml was found to have no additional effect). 5 units of RNase-free DNase (Promega RQ1 DNase) was then added and the incubation continued 10 min. 250 l of 2 mg/ml Pronase, 1% SDS was added, and reactions were incubated for 30 min at 55°C. Reactions were extracted twice with phenol/chloroform and once with chloroform. Ammonium acetate was added to a final concentration of 2 M, and RNA was precipitated with 2.5 volumes of ice-cold 100% ethanol. Following centrifugation, RNA was pelleted and washed twice with 70% ethanol. Dried pellets were resuspended in filter hybridization solution, formamide RNA loading buffer, or TE (10 mM Tris-HCl, 0.1 mM EDTA) (pH 7.2) for hybridization, polyacrylamide gel electrophoresis, or scintillation counting, respectively.
Contribution of Pol I to Total Nuclear Transcription-Nuclear runons using 1 ϫ 10 5 nuclei were performed in the presence of 0 or 150 g/ml ␣-amanitin (Sigma). Purified RNA was resuspended in 10 l of TE (pH 7.2) and spotted onto DE81 paper (Whatman; ϳ1 cm 2 ). Filters were washed five times in 300 ml of 5% Na 2 HPO 4 and twice with 250 ml of sterile Milli-Q water. Washes were at room temperature, 5 min each. Filters were rinsed briefly in 100% ethanol to aid drying. After airdrying, filters were subjected to liquid scintillation counting.
Filter Hybridizations for Nuclear Run-on Experiments-435 g of plasmid DNA (with or without cloned rDNA) was denatured for 10 min in 40 ml of boiling 0.4 M NaOH and loaded onto a 44-mm diameter circle of Zeta-Probe (Bio-Rad) nylon membrane by vacuum filtration in a Millipore filter apparatus. A wash with 40 ml of 0.4 M NaOH was followed by two washes with 200 ml of sterile water. DNA was covalently cross-linked to the membrane using UV light (Bio-Rad Gene Linker GS; program C1). Filters were neutralized with 1.5 M NaCl, 0.5 M Tris-HCl (pH 7.2) and rinsed in water. Circular filters of ϳ7-mm diameter were cut from the large filter using a sterilized hole punch. Each filter contained ϳ15 g of DNA. Filters were coded for later identification and were stored under vacuum.
Filters were prewetted with sterile water and submerged in 0.4 -0.6 ml of hybridization solution (50% deionized formamide, 5 ϫ Denhardt's solution, 5 ϫ SSC, 10 mM EDTA, 0.25% SDS, 50 g/ml yeast tRNA) in 15-ml plastic snap-cap tubes for 1 h at 43°C. Each 15-ml tube contained a plain filter (no DNA), two filters with pBluescript (Stratagene) plasmid DNA, and two filters with cloned rRNA gene sequences (essentially a complete gene cloned in pBluescript). Radioactive RNA from a nuclear run-on reaction was resuspended in 200 l of hybridization solution, denatured at 65°C, 10 min and added to a tube with a set of filters. Hybridization reactions were incubated overnight at 43°C. Filters were rinsed in 2 ϫ SSC, incubated in 20 g/ml RNase A (in 2 ϫ SSC) at 30°C for 30 min to remove unhybridized RNA, then washed at high stringency (0.2 ϫ SSC, 0.1% SDS, 65°C, 1 h). Filters were subjected to autoradiography and quantitation by phosphorimaging (Molecular Dynamics).

RESULTS
The A. thaliana rRNA gene promoter is located within the intergenic spacer approximately 2 kilobases upstream of the 18 S coding sequences (Fig. 1). Noncoding sequences are removed in several processing steps, the first involving cleavage within the external transcribed spacer (ETS) (9,48,49). ETS processing occurs rapidly and is thought to be a co-transcriptional event (50 -55). The promoter-proximal ETS RNA is then rapidly degraded, unlike coding sequences which can have halflives longer than a cell cycle. Because ETS RNA does not accumulate, steady-state levels of transcripts just downstream of the promoter are expected to closely reflect rRNA gene transcription activity. Consequently, we measured transcripts accurately initiated at ϩ1 to initially survey plant hormones for effects on rRNA gene transcription.
Aqueous solutions of gibberellic acid, 2,4-dichlorophenoxyacetic acid, abscisic acid, and kinetin were sprayed onto 3-week-old A. thaliana plantlets until run-off, according to the experimental protocol of Guilfoyle (37). Concentrations tested ranged from 10 Ϫ7 to 10 Ϫ3 M. Ethylene was also tested by gassing plants at 10 -60 ppm in a specially designed chamber FIG. 1. Diagram of an A. thaliana rRNA gene. The gene promoter, spacer promoters, and ETS processing site are highlighted. The repetitive elements between the gene and spacer promoters are thought to be transcriptional enhancers (32). The ETS processing site has been mapped to a position approximately 500 base pair upstream of the 18 S coding sequences (Z. Ye and C. S. Pikaard, unpublished data.).
(with the kind assistance of Dr. Harry Klee and the Monsanto Company, St. Louis, MO). RNA was isolated from control and treated plants at various times after treatment, and rRNA transcripts were detected using S1 nuclease protection as described previously (32,33). Only cytokinin exerted a response in this initial survey (data not shown).
Cytokinin treatment induced nascent rRNA transcripts in a dose-and time-dependent manner (Fig. 2). Six hours after treatment, transcript levels were increased by kinetin at concentrations of 10 Ϫ5 M or higher ( Fig. 2A, lanes 3-6), reaching their maximum with 10 Ϫ3 M exogenous hormone (lane 5). However, if RNA was isolated 12 or 24 h following treatment, an effect was discerned at concentrations as low as 10 Ϫ6 M (lane 2). Note that these are the concentrations sprayed onto the exposed upper surfaces of the plants; internal concentrations are expected to be at least 100-fold lower based on surface area to volume estimations. Maximal induction of accurately initiated rRNA transcripts was approximately 6-fold and was observed 24 h after treatment at the highest kinetin concentrations ( Fig. 2A, lanes 5 and 6).
The experiment of Fig. 2A showed that pre-rRNA transcripts increase within a fixed amount of total RNA in response to cytokinin. However, because ϳ80% of total RNA is ribosomal RNA coding sequences, our estimates of cytokinin responsiveness would be underestimates if the total ribosome pool were increased by cytokinin. At early time points, the ribosome pool is so large that such effects might not be significant. Nonetheless, we compared transcription initiation from the rRNA gene promoters to initiation from promoters of the Rubisco small subunit gene family, among the most highly expressed genes in green leaves. rRNA and ribulose bisphosphate carboxylase gene promoters showed different responsiveness to kinetin treatment (Fig. 2B). Whereas nascent rRNA transcript levels were increased severalfold per unit of total RNA in response to increased exogenous cytokinin concentrations (lanes 2-6), transcripts from the Rubisco gene family were unaffected or reduced at high concentrations. The decrease in Rubisco transcripts could be due to a kinetin-dependent inhibition of transcription, as reported for phytochrome mRNA (56), or a decrease in Rubisco RNA relative to the total RNA pool. Nonetheless, the data suggest that increased steady-state levels of rRNA transcripts is not part of a general positive response to cytokinin.
Although steady-state levels of nascent rRNA transcripts are thought to reflect transcription rates, due to the rapid turnover of ETS sequences, this has not been demonstrated in plants. Therefore, we also performed transcription run-on assays with nuclei of control and hormone treated plants (Fig. 3). Plants were again mock-treated (solution lacking hormone) or sprayed with various concentrations of cytokinin. Twelve hours later, nuclei were isolated and treated in four different ways. [ 32 P]GTP was added to all reactions, but the remaining three nucleotides were withheld from half of the tubes (reactions 1 and 2, Fig. 3A) to control for possible nontranscriptional incorporation of the label. The remaining reactions (reactions 3 and 4) were provided with unlabeled ATP, CTP, and UTP to allow transcript elongation. In half of the reactions, ␣-amanitin was added to 150 g/ml to inhibit pol II and pol III transcription, but not pol I (reactions 2 and 4). After a 10-min period to allow nascent transcripts to be elongated by template-engaged polymerases, RNA was purified and hybridized to an excess of DNA affixed to filters. Duplicate filters (lanes a and b) were used for each reaction. The DNA bound to the filters was either a denatured ribosomal gene clone or denatured pBluescript plasmid DNA, the latter serving as a control for the specificity of the hybridization conditions. After washing at high stringency, filters were exposed to x-ray film and were also quantified by phosphorimaging. As can be seen in Fig. 3, A and B, reactions 1 and 2, no signal was obtained if only the labeled nucleotide was provided to the nuclei. In contrast, adding all four nucleotides allowed the synthesis of radioactive transcripts that hybridized to rRNA gene filters (reactions 3 and 4), but not to pBluescript DNA filters (B). Note that run-on transcription signals were unaffected by ␣-amanitin (compare reactions 3 and 4), as expected for pol I transcription of rRNA. Importantly, the amount of run-on transcription was increased by kinetin treatment prior to nuclei isolation. Nuclear run-on signals were maximal following 10 Ϫ3 M kinetin treatment and were 3.8-fold higher than in mock-treated plants (Fig. 3A). An increase in the kinetin concentration to 10 Ϫ2 M decreased the response, suggesting inhibition at excessive concentrations.
The agreement between the nuclear run-on results and the S1 protection results of Fig. 2A is noteworthy (the 12-h time point of Fig. 2A is the relevant comparison). At 10 Ϫ6 M kinetin, nuclear run-on assays showed a 1.4-fold increase over control nuclei (based on the average signals from the duplicate filters quantified by phosphorimaging), whereas S1 protection detected a 1.5-fold change. Likewise, relative signals at 10 Ϫ5 M kinetin were 2.1-versus 1.9-fold; at 10 Ϫ4 M were 2.7-versus 3.5-fold; and at 10 Ϫ3 M were 3.8-versus 4.9-fold. For unknown reasons, the only large discrepancy was at the highest concentration of kinetin tested, 10 Ϫ2 M, yielding a signal 1.7-fold higher than the control in the nuclear run-on assay, but 3.8fold higher in the S1 protection assay. Nonetheless, we conclude that in Arabidopsis, as in other species, steady-state levels of rRNA transcripts initiated at ϩ1 (detected by S1 Mock-treated plants were sprayed with a solution lacking hormone. 6, 12, or 24 h later, RNA was isolated, and nascent rRNA transcripts were detected by S1 nuclease protection using a 5Ј endlabeled probe spanning the promoter region (Ϫ520 to ϩ92, labeled at ϩ92; see Fig. 1 for the position of the probe). The same amount of total RNA (15 g) was probed for each treatment. The panels shown are from the same autoradiographic exposure of the same gel. S1 signals were also quantified by phosphorimaging; the -fold increase over the mocktreated plants is shown below each lane. The signal in the 24-h, 10 Ϫ5 M kinetin treatment (marked with an asterisk below the lane) is low due to a pinched well in the acrylamide gel and loss of part of the sample during loading; no quantitation is provided for this one lane. B, ribosomal RNA and Rubisco transcript levels respond differently to kinetin treatment. Total RNA isolated from plants 12 h after spraying with various kinetin concentrations was hybridized to the rRNA gene promoter probe or to an oligonucleotide probe designed to hybridize to transcripts of the Rubisco gene family. The multiple S1 nuclease-protected bands observed with the Rubisco probe are as expected based on the positions where mRNA sequences of the different family members diverge from the probe sequence in their 5Ј-untranslated region. protection) closely reflects levels of transcription.
Using the sensitive nuclear run-on assay, we examined the kinetics of the cytokinin response. Plants were mock-treated or sprayed with 10 Ϫ4 M kinetin, and nuclei were isolated 1 or 6 h later (Fig. 3B). One hour after hormone treatment, rRNA transcription was stimulated 1.8-fold relative to control nuclei. At 6 h, the response was 2.4-fold in agreement with other experiments ( Figs. 2A and 3A). For technical reasons, time periods shorter than 1 h were impractical. Nonetheless, these data suggest that induction of rRNA transcription by cytokinin occurs rapidly.
Increased rRNA synthesis in nuclear run-on assays could be due to transcription of more rRNA genes, a higher polymerase density on the same number of genes or more rapid transcript elongation. The latter possibility seems unlikely given that hormone-induced increases in transcripts initiated at ϩ1 (detected by S1 protection) were paralleled by increased transcription throughout the body of the gene (detected by nuclear run-on). These data are most easily explained by kinetin affecting the frequency of polymerase I initiation rather than the rate of elongation. Time courses of nuclear run-on reactions also support the hypothesis that more RNA polymerase I molecules are engaged in transcription in hormone-treated plants.
In the experiment of Fig. 4, nuclear run-on reactions were performed using our standard conditions, but early time points in the reaction were examined. Even at the earliest time point (about 12 s), there is an approximately 2-fold increase in label incorporated in nuclei of treated plants. We interpret this to mean that an increased number of polymerase molecules are stalled on the rDNA of kinetin-treated nuclei. Upon addition of nucleotides, these resume transcription, the higher starting level in hormone-treated nuclei presumably reflecting the increased number of engaged polymerases. In contrast, if the same number of polymerase molecules were template-engaged but altered in their transcription rates, total isotope incorporation would be similar initially but would diverge with time (i.e. the lines connecting the data points would converge if extrapolated to time 0). Note that the zero time point provides no data in this assay, because endogenous transcripts are not labeled until transcripts begin to elongate. For this reason, there are no time 0 data points in Fig. 4.
An experiment done in parallel with the one represented by Fig. 4 was performed to estimate the number of nucleotides incorporated per second in the nuclear run-on assays; these data are worth mentioning, though the data are not shown. Nuclei were treated with DNase-free ribonuclease to digest nascent transcripts protruding from the bound polymerase molecules. Nuclei were then washed and incubated in run-on conditions. Aliquots were taken at various intervals (10 -300 s), and labeled transcripts were separated on a sequencing gel and visualized by phosphorimaging. Despite substantial size heterogeneity, computer imaging allowed us to identify the average transcript size at each time point by identifying the size range where the radioactive signal was highest. Such analyses led to an estimated average elongation rate of ϳ2 nucleotides/s. This For this experiment, radioactive RNA was separated from unincorporated GTP by denaturing gel electrophoresis, and radioactive signals were quantified by phosphorimaging. Background signals were subtracted for all time points. The data were fitted to straight lines using CricketGraph software. As discussed in the text, the higher initial incorporation of GTP in nuclei of hormone-treated plants suggests that kinetin primarily affects the number of template-engaged polymerase I molecules rather than the rate of elongation. is in the same range as elongation rates of ϳ5 nucleotides/s within isolated animal nuclei (57). Importantly, we observed no obvious difference in average transcript size over time in nuclei of hormone-treated and control plants. However, transcription signals were approximately twice as strong in kinetin-treated nuclei. These data are consistent with those of Figs. 2-4, suggesting that kinetin increases the number of polymerase I enzymes engaged in rRNA transcription.
Because pol I accounts for a large proportion of all nuclear transcription, we estimated the effect of kinetin on overall transcription by measuring total and ␣-amanitin-resistant [ 32 P]GTP incorporation in nuclei of control and treated plants (12 h following 10 Ϫ4 M treatment). Radioactive RNA from run-on reactions was purified and bound to positively charged DE81 filters. These were washed extensively to remove free nucleotides and quantified by scintillation counting. In control nuclei of two independent experiments, pol I transcription accounted for 32 or 36% of total RNA polymerase activity, whereas in cytokinin-treated plants, pol I activity accounted for 56 or 63% of the total (Fig. 5). These data are in line with studies in animal cells showing that rRNA transcription can account for 40 -80% of total transcription (12). DISCUSSION To the best of our knowledge, our study is the first to examine plant hormone effects on transcription initiation directly at the rRNA gene promoter. Kinetin-induced increases in steadystate transcript levels were mirrored by nuclear run-on data, suggesting that cytokinin elicits a response at the level of gene transcription rather than rRNA stability. In general, our results support the conclusions of prior studies using isolated nuclei of detached cotyledons (58). Agreement between S1 protection and nuclear run-on data also suggests that transcripts upstream of the ETS processing site are rapidly degraded in plants, such that steady-state measurement of these unstable RNAs can essentially substitute for nuclear run-on assays.
Responsiveness of rRNA genes to kinetin is reasonable in light of the role of cytokinins in cell division (cytokinesis). Dividing cells partition their ribosomes among daughter cells, thus rRNA synthesis might need to be up-regulated to replenish the ribosome pool. However, up-regulation within 1 h of kinetin treatment is unlikely to be a consequence of cell division. Furthermore, DNA content in total nucleic acid extracts, detected by Hoechst dye binding and fluorometry, was not appreciably increased at 1, 6, or 12 h following cytokinin treat-ment and was only slightly increased at 24 h (data not shown). In contrast, if rRNA transcription were up-regulated only as a consequence of DNA replication and cell division, every 2-fold increase in rRNA transcription should be paralleled by a 2-fold increase in DNA content. Other data not shown include our observation that rRNA transcription is similarly induced in roots, floral tissues, and whole plants, suggesting that cytokinin responsiveness is not tissue-specific.
We were surprised that other hormones did not exert an effect on rRNA gene transcription in our initial screen. In particular, we thought auxin (specifically 2,4-dichlorophenoxyacetic acid) might elicit a positive response. Guilfoyle (37) clearly showed that RNA polymerase I activity on heterologous template DNA (e.g. sheared calf thymus DNA) was about 10fold higher in cell-free extracts of soybean hypocotyl treated with 2.5 ϫ 10 Ϫ3 M 2,4-dichlorophenoxyacetic acid. Lack of an auxin effect in our experiments might suggest that pol I levels alone are not limiting for promoter-dependent rRNA gene transcription within the plant cell. Perhaps increased levels or modifications of other transcription factors are also needed to facilitate promoter-dependent rRNA transcription, and these events are affected by cytokinin but not auxin.
Cytokinins are the least understood class of plant hormones in terms of their mechanisms of action and signal transduction pathways (59,60). Few mutants affecting cytokinin metabolism are available. The A. thaliana mutant, amp1 (61), has endogenous cytokinin levels six times higher than normal and displays a rapid growth phenotype. Using the S1 protection assay, we found nascent rRNA transcripts in amp1 plants to be slightly higher than in wild-type plants (data not shown). However, kinetin responsiveness was similar in amp1 and wildtype plants. We also tested several mutants we have isolated that are resistant to cytokinin levels that inhibit seed germination and plantlet growth. Thus far, rRNA transcript levels are also increased in these lines in response to exogenous cytokinin treatment, although slight differences relative to control plants are suggested in some lines. 2 Mutants disrupted in cytokinin signaling would clearly be valuable for future studies.
The specific responsiveness of rRNA gene transcription to kinetin suggests that in plants, cytokinins may be key signaling molecules involved in communicating cellular growth status to the transcription and protein synthetic machinery. It is intriguing that kinetin and ppGpp (in prokaryotes) are both modified purines. However, at present it is not clear if this is coincidence or a hint of a functional similarity. FIG. 5. Polymerase I transcription as a percent of total transcription nearly doubles in response to kinetin. Total incorporation of [ 32 P]GTP into RNA (counts/min) in 10-min nuclear run-on reactions are given for two independent trials in the absence or presence of 150 g/ml ␣-amanitin. The effect of kinetin and the relative proportion of ␣-amanitin-resistant transcription (polymerase I) in nuclei of control and treated plants is calculated. The counts/min shown have been adjusted for background, determined by spotting filters with RNA from reactions containing only the labeled nucleotide, but not the remaining three nucleotides needed for transcript elongation.