A Consequence of the rel Gene during a Glucose to Lactate Downshift in Escherichia coli THE RATES OF RIBONUCLEIC ACID SYNTHESIS

SUMMARY Escherichia coli strains with relaxed (CP79) and stringent (CP78) RNA control were examined in regard to their RNA synthesis and accumulation during balanced growth in two media supporting different growth rates, and during downshift transitions between them. Although the RNA content was greater for CP79 at both growth rates, the total rate of RNA synthesis was found to be proportional to the growth rate in both strains. Both strains traversed a downshift transition in about equal periods of time, but as a consequence of the relaxed mutation the reduction of the rate of RNA synthesis in CP79 was much less; the total rate of RNA synthesis was reduced during the downshift to 5 % of the preshift rate in CP78 but to only 50 % of the preshift rate in CP79. Therefore, cessation of RNA accumulation during the downshift in CP79 implies that considerable RNA degradation was required. Thus, in addition to its function during amino acid starvation, the rel gene in-creases the efficiency of the transition from fast to slow growth rates.

SUMMARY Escherichia coli strains with relaxed (CP79) and stringent (CP78) RNA control were examined in regard to their RNA synthesis and accumulation during balanced growth in two media supporting different growth rates, and during downshift transitions between them. Although the RNA content was greater for CP79 at both growth rates, the total rate of RNA synthesis was found to be proportional to the growth rate in both strains.
Both strains traversed a downshift transition in about equal periods of time, but as a consequence of the relaxed mutation the reduction of the rate of RNA synthesis in CP79 was much less; the total rate of RNA synthesis was reduced during the downshift to 5 % of the preshift rate in CP78 but to only 50 % of the preshift rate in CP79. Therefore, cessation of RNA accumulation during the downshift in CP79 implies that considerable RNA degradation was required.
Thus, in addition to its function during amino acid starvation, the rel gene increases the efficiency of the transition from fast to slow growth rates.
Eschetihia coli is able to maintain a constant relation between the concentration of ribosomes and its expressed protein synthesizing capacity.
Thus, at all but very slow growth rates, the number of ribosomes per cell and amount of mRNA per cell increase linearly with the growth rate. (See Kjeldgaard (1) and Maaloe (2) for recent reviews.) The possibility that the rate of RNA synthesis might be constant and that the rate of RNA accumulation might be adjusted to the growth rate by degradative reaction has been proposed (3,4). Although indirect evidence against this model has been adduced (5-7) it has never been eliminated.
The present data exclude such a model and establish that the rate by RNA synthesis varies with the growth rate.
When faced with the necessity of adjusting to a slow growth * Present address, Department of Medicine, Johns Hopkins Hospital, Baltimore, Maryland 21205. rate from a fast one (downshift), E. co%, as well as other bacteria, is able to attenuate the rate of RNA accumulation for a time, until the RNA content is reduced to that characteristic of cells grown at the slower rate. Certain characteristics of this downshift have led several authors to suggest that the mechanism involved may be related to or the same as that, which operates during amino acid starvation (1, 2,8,9).
The cellular response to conditions of amino acid starvation is greatly influenced by the ret gene (see Reference 10 for a review of the properties of the rel gene). Thus, in mutants carrying the relaxed mutation, RNA synthesis and accumulation continue under conditions of amino acid starvation, whereas in the wild type stringent counterpart, RNA accumulat.ion ceases and RNA synthesis is moderately reduced (11)(12)(13).
The present communication is concerned with the importance of the rel gene in the regulation of RNA synthesis during balanced growth and during downshift transitions.
The results of this study demonstrate that both CP78 (rel+)l and CP79 (rel-) are able to adjust their RNA content and rate of RNA synthesis to the values characteristic for slow growth. However, during a downshift transition the rates of RNA synthesis in the two strains are markedly different.
Thus, it is concluded that the rel gene product (the identity of which is at present unknown), functions not only during arnino acid starvation but also during downshift transitions and greatly enhances the efficiency with with which a downshift in growth rate is executed. All experiments were done at 27" in :I reciprocally shaking water bath. Growth was measured by following the absorbance at 600 rnp with a Zeiss PJCQII spectrophotomcter.
In the downshift experiments, 0.025'( glucose and 0.25%> lactate were present. The break in absorbance (11600) denoting the period of transition from glucose to lactate metabolism occurred at 0.25 to 0.30, corresponding to a cell density of about 2.0 x IO8 cells per ml. Measurement of Cellular RNA-At appropriate times, 5 ml of cult,ure were withdrawn, mixed with 0.5 ml of 50% trichloroacetic acid, and chilled.
These suspensions were collected on glass fiber filters, washed with 5% trichloracetic acid, and placed in test tubes with 2.0 ml of 5Yo trichloroacetic acid.
The cells were hydrolyzed by heating to 90" for 20 min. After sedimenting the debris by centrifugation, 0.5 ml of the supernatant fluid was removed and the RN.\ con tent was assayed by the orcinol method (14).
Rates of R.VA Synthesis-When cultures had reached the appro- priate AGoO, [3H]uracil (final concentration, 20 &i per ml; specific activity, 9.2 mCi per pmole) was added, and 5-ml samples were removed at lo-set intervals and mixed with an equal volume of ice-cold 10% trichloroacetic acid. These suspensions were sedimented by centri?ugation, and the cellos were washed three times with 23"/c trichloroacetic acid containing uracil, 0.1 mg per ml. The cells were hydrolyzed in 1.0 ml of 0.3 M KOH for 16 hours at 37", and the radioactivity of the hydrolysates was measured (15). The method, which makes use of a a2P, 3H double labeling technique as well as the measurement of the specific activity of the intracellular CTP and UTP pools, has been previously described (5). In experiments with CP78, in which the specific activity of the UTP pool was very low, acid-soluble extracts were concentrated with charcoal before chromatographic isolation as decribed previously (11).

Concentration. of ppGpp and Nucleoside Triphosphde
Pools-At least one full doubling before the downshift, HSa2P04 was added to give a final concentration of 20 to 25 @Zi per ml. Aliquots (100 ~1) were removed at intervals and mixed with 50 ~1 of icecold 4 M formic acid. These suspensions were sedimented by centrifugation and 10 ~1 of the supernatant fluids applied to PEI sheets for two-dimensional chromatography.
The chromatographic isolation of the four major nucleoside triphosphates was identical with that previously described (5). In addition, some experiments were done with 1.2 M KHzPOl in the second dimension for the isolation of ppGpp. These compounds were located by autoradiography, and the portions of the PEI sheets containing them were cut out and counted in the toluene-Liquifluor scintillator.
The specific activity of the culture medium was determined by counting an aliquot and measuring the phosphate concentration (16).

RESULTS
Rates of RNA Synthesis during Balanced Growth-The general methods for the measurement of the retes of RNA synthesis during balanced growth by means of uracil labeling and the measured specific activities of the UTP pools have been described by guest on July 10, 2020 http://www.jbc.org/ iu detail elsewhere (5). Fig. 1 illustrates this method, showing the results obtained with CP78 growing exponentially in lactate medium.
The rate of R?u'h synthesis for CP79 under the corresponding conditions was determined according to the same method.
The rate of uracil uljtake is slightly faster for CP79 than for ('1'78, and the UTP ])ool specific activity for ('1'79 is slightly less than for CP78. Thus, the total rate of RX.4 synthesis is higher for CP79 than for ('P78, in spite of their similar growth rates. These data are summarized in Table 1. Table I also &on-s the RX;1 content under various conditions of growth. The rates of RX&4 accumulation were calculated with these values and the composition of total cellular RNA given by Midgley (17). The difference between these rates of RKA accumulation and the measured rates of RNA synthesis must be the rates of synthesis of unstable RKA.
If it is assumed that all rRNA and tRNA synthesized is essentially stable, then the unstable fraction of total RNA is mRNA.
Rate of Xynthcsis of Rh:A during the Downshift- Cells were cultured in a shift medium cont,aining 0.25y0 lactate and 0.0"5"; glucose, so that the glucose would be exhausted in mid-log phase and metabolism of lactate would begin. At intervals the cultures were sampled and their d 600 was measured.
Simultaneously, the RN.4 concentration was assayed according to the orci-no1 rnet,hod described under "Experimental Procedure." strates that for both CP78 and CP79 the period of transition between glucose and lactate is marked by an absolute cessation of RNA accumulation for a period of about 2f hours. During this period, as well as before and after it, the instantaneous rates of RNA synthesis were measured.
In these experiments, cells were cultivated in a downshift medium, which contained 0.25% lactate and 0.025($ glucose, so that at a cell density of 2 X lo8 cells per ml, glucose would be exhausted and lactate would serve as the sole source of carbon and energy. -4t an AeOO value of about 0.1, the culture was split, and H332P04 (20 /Ki per ml) was added to to one of the subcultures.
Xt various times during the course of the downshift, the Iwo cultures were labeled simultaneously with [3H]uracil (final concentration, 20 &'i per ml; specific activity, 9.2 niC'i per pmole).
The 32P culture was used to determine specific activity of the HTTP pool, and the other culture was used to determine the [3H]uracil labeling of total cellular RNA.
The results of the rate measurements are described in Table II and in Fig. 3. These data show that in CI'78 there is a drastic inhibition of the rate of RNA synthesis, 20-fold being the maximum measured.
However, the actual value for the maximal reduction is probably not significant, owing to the rapidly changing rate. In contrast, the rate of synthesis in CP79 appears to fall smoothly to reach the new value of about 507; of the glucose rate, without going through a minimum as is the case with CP78. It is clear that very much more RNA is actually synthesized in CP79 during this transition period than is synthesized in CP78, but RN.4 accumulates in neither organism.
Nucleoside Triphosphate and ppCpp Pool Changes-When relf cells are dellrived of a required amino acid, the intracellular concentration of GTP falls precipitously (18). This decrease is accompanied by the similarly abrupt appearance of an unusua1 guanosine tetraphosphate (19)(20)(21).
Recently, this compound has been observed at very low basal concentration in cultures of both relased and stringent cells aud at clevat.ed levels during growth rate transitions (22). The concentration of this compound as well as the changes in the purine triphosphates were measured during the glucose to lactate shift in the present study. Fig. 4 shows that the intracellular concentration of ATP and GTP fall to reach new values in CPSS. This reduction is abrupt for GTP in CP78, but more gradual in C1'79. For CP79, hTP falls only transiently, then returns to the preshift level.
The basal concentration of ppGpp appears to be slightly higher in the glucose medium for CP78 as compared with CP79. These data show considerable scatter, but the observation agrees with similar measurements by Lazzarini, ('ashel, and Gallant (22). The concentration of ppGpp rises very abruptly in both strains, and appears t,o be concomitant with the cessation of RNA accumulation and t,he break in the growth curve. The peak concen-TMILE II Uete,minalion of amount of UMP incorporation into Ri$A ajler SO-see pul se of [SH]uracil Data are shown from experiments during a glucose to lactate downshift at 40 min before (A), 30 min after (II), GO min after (C), and at 165 min after (D) the break in the AHJO curve shown in Fig. 2 per rlsoo. This level is somewhat less than that reached during Further, the disappearance of the compound is more rapid in threonine star\-ntion, as measured by Cashel (20). In ('P79, the CP78 than in c'P79, suggesting either that it is produced in the peak concentration is even less, about, 100 to 150 pmoles per latter for a longer period, or that is is consumed in the former at a Asoo. How-ever. its appearance at all distinguishes this case from faster rate. The rate of RNA accumulation in E. coli is exponentially related to the growth rate (23). The variation of the rate of RNA synthesis with the growth rate has been assumed (2) but never proven.
Moreover, alternate models have been proposed (3,4), which involve degradation of ribosomal RNA at slow growth rates to achieve reduced rates of RNA accumulation.
Indeed, E. coli has been shown to degrade ribosomal and transfer RNA in certain circumstances (3,4,24).
Recently, methods for the measurement of the instantaneous rates of RNA synthesis have been described (5). These methods have been used in the present experiments to measure the rates of RNA synthesis in two media which permit different growth rates. The results indicate that E. coli is able to adjust its rate of RNA synthesis as well as its rate of RNA accumulation to its growth rate. This is true for cells with both relaxed and stringent RNA control.
The instability of RNA in CP79 during the shift, implied by its faster rate of synthesis relative to that of CP78, suggests either that the RNA which is being made is mRNA and normally unstable, or that rRNA is rendered unstable.
Hybridization-competition experiments (13) have demonstrated that the RNA synthesized is predominantly mRNA, and that there is selective inhibition of rRNA synthesis during the downshift. However, as previously noted (13), the failure of RNA to accumulate during the downshift although a small amount of rRNA is synthesized implies that degradation of rRNB also takes place.
In the calculations summarized in Table I, the rates of tRNA and rRNA synthesis were assumed to be the same as the rate of RNA accumulation by the cell. If that assumption is correct, and tRNA and rRNAA are not overproduced and subsequently degraded, proportionality should exist between the rate of mRNA synthesis measured in this way and the growth rate (micra), and between the rate of rRNA synthesis and the square of the growth rate, squared micra (2). This relation predicts that as micra decreases, the proportion of RNA synthesis devoted to rRNA and tRNA should decrease. This prediction is borne out, at least qualitatively, by the calculations shown in Table I, in which the percentage of stable RNA synthesis is shown to be less in lactate medium for both strains than in glucose medium. Hence, a model in which RNA accumulation at different growth rates is regulated by the degradation of excess RNA (25) seems to be ruled out by the present data.
If the rate of elongation of nascent RNA chains is more or less constant at all growth rates (5)(6)(7)26), these observations allow the prediction that the cell maintains a number of nascent RNA chains at any instant which is adjusted to suit its growth rate.
Thus, it appears that although E. coli has mechanisms for the degradation of ribosomal RNA under conditions of unbalanced growth, the most efficient mechanism for the regulation of the rate of synthesis during balanced growth involves maintenance of a number of growing chains which is proportional to the growth rate. This type of regulat,ion minimizes the wastage of newly synthesized RNA.
The downshift transition, like amino acid starvation, represents a period of unbalanced growth.
During this time, no RNA accumulates, and the cell, by continued protein and DNA accumulation, adjusts its composition to that characteristic of cells gron-ing at the slower rate. If indeed there is a relationship betweenregulationduringamino acid starvation and during a down-shift, then manifestation of the rel gene should be detectable during the downshift, as well as during amino acid starvatiom. Accordingly, the rates of RNA synthesis have been measured in CP78 and CP79 during a downshift from glucose to lactate media. Despite the fact that RNA accumulation ceases for 2$ hours during the downshift (Fig. 2) substantial rates of RNA synthesis are demonstrable in both strains. As during amino acid starvation, the relaxed strain synthesizes considerably more RNA than its stringent counterpart (Fig. 3). These findings indicate that the relaxed strain meets the challenge of the downshift much less efficiently than does the stringent.
These results support the conclusion that the amino acid control and growth rate control are interrelated since a single mutation affects both processes.
The guanosine tetraphosphate has been reported to appear during amino acid starvation of relf but not rel-strains (19)(20)(21), and during downshift transitions in both rel+ and rel-strains, but the actual quantitative reIationships are strain-dependent (22). In the present experiments this compound appeared in both strains, in slightly higher concentration in CP78, and with a more rapid disappearance.
Its appearance at all in the relaxed CP79 served to distinguish the downshift from amino acid starvation (20)~ A model in which ppGpp might be an inhibitor of RNA synthesis has been suggested (20), and is consistent with the present data. However, the appearance of this compound in CP79 mitigates against the possibility that ppGpp is a specific inhibitor of RNA polymerase which appears under conditions of unbalanced growth and the production of which is eliminated by the rel-mutation.
The experiments with CP79 demonstrate that ppGpp correlates not with the inhibition of RNA synthesis, but with the the inhibition of RNA accumulation. That, is, in all circumstances studied so far, ppGpp appears o111y under conditions of restricted RNA accumulation.
Therefore, if ppGpp is involved in regulation of RNA synthesis, it seems likely that it participates in the selective inhibition of synthesis of stable forms of RNA rather than in the promotion of degradation of these species.
Equally consistent with these data is the proposal that ppGpp is the product of a reaction which results in the degradation of forms of RNA which would be stable under conditions of balanced growth.
This cause or effect dilemma remains to be resolved.