Regulation of Ribonucleic Acid Accumulation in Viva by Nucleoside Triphosphates*

SUMMARY A continuous and unique variation of stable RNA accumulation with pool levels of nucleoside triphosphates was ob-served when Escherichia coli (K12 Leu-, rel+) cells were entering into phosphate starvation or recovering spontaneously from inhibition by dinitrophenol. Since the rate of RNA accumulation in these experimental systems surpassed all other known systems exhibiting the same pool levels of nucleoside triphosphates, it is suggested that stable RNA synthesis in these systems was subjected to kinetic regulation by substrate concentrations. This kinetic relationship in vivo between RNA synthesis and nucleoside triphosphate substrates was sigmoidal in character and exhibited a higher half-saturation constant than the similar relationship obtained with RNA polymerase in vitro. Since the biosynthesis RX-1 RNA polymcrase requires ribonucleoxitle triphosphat,es as substrates, kinetic regulation by triphosphnte levels expected represent a simple mccha-rrism regulating RNA synthesis which can readily respond to mctnbolic changes in the cell. triphosphates did not mediate the inhibition of RNA synthesis in Escherichia coli during :I of nutrient-and

Since the biosynthesis of RX-1 by RNA polymcrase requires ribonucleoxitle triphosphat,es as substrates, kinetic regulation by triphosphnte levels is expected to represent a simple mccharrism regulating RNA synthesis which can readily respond to mctnbolic changes in the cell. The finding that triphosphates did not mediate the inhibition of RNA synthesis in Escherichia coli during :I series of nutrient-and inhibitor-induced shift downs (1) therefore was a surprising one and prompted a search for experimental systems in which kinetic regulation becomes important. Sucll systems should exhibit a unique and continuous corrclntion between pool levels of nuclcosidc triphosphates and rates of RNh synthesis and accumulation.
Moreover, with the nucleoside triphosphates directly acting as the rate-limiting factor, at any given level of triphosphates such systems also should exhibit a rate of RNA synthesis and accumulation which surpasses all other systems utilizin g other modes of regulation. On the basis of these two crit,eria, a number of experimental systema have been analyzed and found not to be subjected to kinetic regulation by nucleoside triphosphates, and these included purinc and pyrimidine starvations, as well as inhibitions by arsenate and carbonylcyanide-p-trifluoromethoxyphenyl- *  What'man microgranular DUE-cellulose, DE-52 (1.0 meq per g of dry weight), was a product of the II. Reeve Angel Company, Clifton, i\;ew Jersey, and N-mcthyl-N'-iiitro-N-iiitrosoguailitlille ~vas from .\ldrich. Bacteria and Growth Conditions-The I:'. coli IX12 Lcu-cells and most of the espcrimental conditions have been described in the preceding paper (1). Since all growth experiments in this study involved 33P-labeling, the Tris-buffered medium 63 (24 pg per ml of I<I-I~I'O~~) TWS used. When phosphate starvation was to be observed, the IIHZ1'O1 content was reduced further to 15 ,ug per ml. To observe adenosine deprivation, a leaky purine auxotroph of I(12 Leu-(isolated, after mutagenesis by A-mcthyl-N'-nitro-~~;-nitrosogu:L11itlinc, as a small colony on agar plates containing 2.5 pg per ml of adenosine) was grown in the presence of only 7 pg per ml of adenosine.
In all instances, cells were allowed to undergo at least8 three doublings in the experimental medium containing ["3P]orthophosphatc to about 1 to 2 x lo* cells per ml before sampling began.
Cell mass was monitored by absorbance at 450 nm in a Gilford dpcctrophotometer; unit absorbance with a lo-mm light path corresponded to approsimately 0.2 mg of dry weight or 4 x IO8 cells per ml.
Incorporation oj [33P]Orthophosphate into Nucleic Acids and Nucleotides--To measure 33P-incorporution into RKA and DKA, one 2O-,~d cell sample was pipettcd into 3 ml of 5% trichloroacetic acid, a duplicate one into 1.5 ml of 0.5 N NaOH, and the two samples were chemically fractionated according to Gallant and IIarada (2). To determine nuclcoside triphosphates, 100.~1 cell samples were pipetted into 150 ~1 of 1 s acetic acid, extracted by freeze-thawing (3), and 25 ~1 of each extract were spotted on a polycthylenrimine-impregnated cellulose thin layer sheet. Chromatography was carried out two-dimensionally as previously described (3), or one-dimensionally by developing the thin layer chromatographic sheet twice to 20 cm above t'he origin in 1 s LiCl and 1 N acetic acid, with an intervening wash in methanol.
Scintillat.ion counting in a Nuclear-Chicago Mark I counter was employed to determine the 33P content in all cases. Measurement of RSA Polymerase Activity-IWA polymerase was purified from frozen E. coli I~12 cells using the procedure of Burgess (4) to the end of chromatography on a DE-52 cellulose column.
Peak fractions from t)hc column were pooled and stored in 50% glycerol at -20".
In the kinetic experiments, measurements of uninitiated polymerizations were carried out according to Anthony, Wu, and Goldthwait (5). The reaction mixture (0.5 ml) contained ATP:GTP:[W]UTP (0.58 mCi per mmole) : CTP in the ratio of 7:4 :3 : 1, 18 pg of enzyme, and 18 pg of native DNA, which had been isolated from frozen E. coli K12 cells using the method of Marmur (6) but omitting treatment with ribonuclease.
Polymerizations at 37" were terminated by either precipitation and washing on a Millipore filter (GS, 0.22 pm) with 5% trichloroacetic acid or by the direct filOration technique of Sentenac, Simon, and Fromageot (7). The latter required adding 25 pmoles of ethylenediaminetetraacetate and 1 mg of UTP to the reaction mixture, and washing the mixture through the Millipore filter with the incubation maleate buffer. In either instance radioactivity on the dried filter was determined by scintillation counting.

RXA Accumulation and Xucleotide Levels during
Phosphate Starvation--As a culture of E. coli K12 Leu-cells progressively depleted the phosphate in a low phosphate medium, cell growth very gradually slowed down. Fig. 1 shows the incorporation of [33P]orthophosphate into RNA during the slow down, and tangents drawn to the smoothed incorporation curve at different times give the rates of RNA accumulation at these times. Simultaneous sampling for ribonucleoside triphosphates yielded the results of Fig. 2. Over a period of 3 hours, the levels of ATP, GTP, and CTP dropped to 25% of their normal levels, and UTP to less than 15oj,.
RNA Accumulation and Nucleotide Levels during Spontaneous Recovery from Diniirophenol-As recorded earlier (I), RNA accumulation and growth of K12 Leu-cells were severely inhibited when first exposed to 0.5 mnr dinitrophenol, but later exhibited a spontaneous recovery to close to 90% control rates. This pattern of spontaneous recovery persisted with higher dinitrophenol concentrations, but the recovered rate of RK-S accumulation as well as triphosphate levels varied with dinitrophenol concentration (Fig. 3).
RNA Accumulation and 1Yucleotide Levels during Allenosine Starvation--When mutant cells of K12 Leu-which required purines for fast growth were cultured in the presence of limited adenosine, the exhaustion of adenosine was marked by a smooth transition of RXA and DNA synthesis to a slower, steady rate. The nuclcoside triphosphatcs, however, exhibit'ed a complex pattern of changes ( Fig. 4).
Kinetics of RNA Polymerase in Vitro-The substrat'e dependence of RNA polymerase was analyzed in order to compare substrate regulation of RNA synthesis in vivo and in vitro.
In Fig In (i) to (iv), the zero time in each case corresponds to the zero time of Fig. 1. The different symbols also correspond to those of Fig. 1 and represent data from different experimental cultures.
5, ATP, GTP, UTP, and CTP were used in the ratio of 7 :4 : 3 : 1, which spproximated the spread of triphosphate concentrat,ions in their decreases during phosphate starvation.
As described under "Materials and Methods," two different procedures were used to terminate RNA polymerization and to collect the newly synthesized RNA.
The progress curves employing acid precipitation showed a significant lag, but the curves employing direct filtration did not, likely reflecting the failure of the acid to precrpitate very short nascent RNA chains. The rate curves derived from t'he two procedures were also significantly different (Fig. 6) In the experiment in Fig. 1, the culture had not yet reached the point of complete phosphate exhaustion, and the turn off of RNA accumulation between zero time and the 40.min mark was close to an average of 20% per 10 min. Consequently the contribution by degradative processes to the turn off of RNA accumulation might be limited. It is important, nevertheless, to define the consequences of any degradative control, as well as nonkinetic types of transcriptional control, operating in the phosphate-depletion system of Fig. 1 The mechanism actually includes two successive kinetic controls by substrate level, first in the phosphorylation of nucleosides and nucleotides, and secondly in the polymerization of RNA. In this study no enquiry has been made into the quantitative control of phosphorylation, but the preservation of moderately constant rat,ios among ATE', GTP, UTP; and CTP during their decreases in Fig. 2 is consistent with a tightly controlled partition of phosphate by the phosphorylating and kinase systems. Because phosphate and nucleosides are both consumed in RX:1 synthesis, the tn-o-stage mechanism given in Fig. 8 might be expected to operate equally during nucleoside starvations. 7. Zn oivo rate ctuves for rLNA synthesis and accumldation. The relationship between the net RNA synthesis and nucleoside triphosphate levels is derived by correlating the tangents drawn at different times in Fig. 1 to the triphosphatcs measured at t,he same timrs ilr E'ig. 2, (i) IO (iv). CUIW I (A) applies to the average valrles for the four iriphosphatcs. CUWC ZZ (a) applies to the KS,\ synthesis V-X :rc~cornpanic,d o111y by modcrate decreases in AYrI' a11d GTP dO\Vll to 607, of their glucose-gromu levels. Duriug guauine exhaustion of the glucose-grown strains R257 and AT2465, howev~~r, GTP dropped to below 20%. Uracil eshau&on of strain lSD1 similarly lowered UTP to the neigllborhood of 20%.
'I'hc fintlilrgq with 11257, AT2465, aud UDl cells, showing a severe iuhibitioll of RNA synthesis at 207, triphosphates, are tjhercEorc cnt,irely in accord with the kinetic calibration giveu by Curves I and 11 iu Fig. 7. Mexsurcments from the DIM29 and 156.2a systems, however, lie far to the right of CWWS I ant1 I/, and suggest the intervention of rcgulntory mechanisms ot,hcr than straightforward substrate control. This type of di~crcpallt behaviour JXXS not restricted to these cells: and t,he rcaddit,ion of uracil to uracil-starved 15-T cells was found by Lazzarim, Xaknta, and Winslow (9) to stimulate RN.1 :Lccumulation wit,11 only a small rise in UTP.
Cornprtrison oj" in. V'ivo and in Vitro Ra,te Curves-During progressive phosphate exhaustion, the concentration of UTP dropped more severely than the other three triphosphates, and thcrc is at present no basis to clrcide which of Curves I and 11 value for UTP, which underwent greater per cent decreases in in Fig. 7 more closely defines the in viva rntc curve for RNA polymerization. Indeed, the exact form and position of the rate curve might vary somewhat with the prevalent ratios betwcen the four substrates, and the area bouuded by Curves I and 11 in Fig. 7 can be more usefully rtlgarded as a regulatory zone where a simple substrate control of RNA synthesis appears to operate.
Curve III in Fig. 7 provides an in vitro rate curve for comparisou with the in vivo rate dependenct.
The control levels of triphosphates in the K12 Leu-cells TT-ere 6.3 pmoles of Al'l', 3.3 pmolcs of GTP, 2.9 of pmolcs UTP, and 1.7 pmoles of CTP per gram dry weight.
The 1OO70 scale used in Fig. 7 thcrcfore corresponds to 1.6 mhl in terms of dTP, assuming a 4: 1 ratio between wet and dry cell weights.
.\gain using ,YYP concentration as a basis of calculation, the in v&o rate curve for RNA pol?mcrase from Fig. 6, obtained by acid precipitation, therefore assumes the form of Curve II in Fig. 7 lying to the left of the Z-11 region. It would move even further left if the rate curve from Fig. 6 obtained by direct filtration were used. Because of the vast differences between i?l vivo and in vitro conditions, a discrepancy between III and I-11 is not entirely unexpected.
Beside the possible activities of degradatire aud nonkinetic transcriptional value but vary with the length of nascent RNA chains selected for measurement, acid precipitation being less efficient than direct filtration in isolating the very short RNA chains polymerized (7). Secondly, part of the substrate triphosphates inside the cells might be bound to membrane structures and therefore not freely available for RNA synthesis (10). Finally, if oxidative phosphorylation occurs at the cell membrane and RNA synthesis at the cell center, the triphosphates have to travel down a diffusion gradient to reach the site of RNA synthesis. From the calculations recorded in Fig. 9, it is apparent that the concentration gradient ACrr likely will be small in the cytoplasmic region.
The gradient AC, in the DNA region, however, may become quantitatively important on account of high viscosity seriously impeding the diffusion of triphosphates.
Because of the packaging of DNA at a density of 0.4 g per cm3 (la), a relative viscosity of lo3 or lo4 is not excluded for the DNA region, extrapolating from its known variation with lower DNA concentrations (13). If a large AC1 indeed exists, the effective triphosphate concentrations at the site of RNA synthesis will be lower than the average cellular concentrations, and the regulatory I-11 zone in Fig. 7 will have to be moved left. Physical factors which affect the organization and viscosity of the DNA also unavoidably will exert an important influence on the rate of RNA synthesis.
Whether or not they will be utilized by the cells for regulatory purposes remains to be considered. controls as already discussed, and the multitude of metabolites and protein factors which might modulate RNA polymerase kinetics in viva, at least three factors can be recognized as possibly contributing to the discrepancy. First, although a survey (11) indicated that the K, for RNA polymerase in vifro was not extensively altered using denatured compared to native E. coli DNA as template, or using the four triphosphates in the ratio of 1: 1: 1: 1 compared to 7 :4: 3 : 1, the two rate curves recorded in Fig. 6 suggest that the in vitro K, may not be a fixed