Regulation of Lac Transcription in Escherichia coli by Cyclic Adenosine 3’, 5’-Monophosphate STUDIES WITH DEOXYRIBONUCLEIC ACID-RIBONUCLEIC ACID HYBRIDIZATION AND HYBRIDIZATION COMPETITION

Abstract Cyclic adenosine 3',5'-monophosphate (cyclic AMP) is required for the synthesis of lac mRNA and β-galactosidase in Escherichia coli and cell-free extracts. Competition hybridization experiments, described in this report, show that the concentrations of lac mRNA in E. coli are controlled by cyclic AMP, as well as by inducer. Both transient and permanent repression by glucose affect lac mRNA production and are reversed by cyclic AMP. In an adenyl cyclase-deficient mutant strain, lac mRNA is not synthesized until the cells are supplied with cyclic AMP. A mutant strain lacking a recently described "cyclic AMP receptor protein" does not make lac mRNA despite exogenous cyclic AMP. We conclude that cyclic AMP, cyclic AMP receptor protein, an intact promoter locus, and other unidentified factor or factors are necessary for efficient transcription of the lac operon.

coli and cell-free extracts. Competition hybridization experiments, described in this report, show that the concentrations of tat mRNA in E. coli are controlled by cyclic AMP, as well as by inducer. Both transient and permanent repression by glucose affect lac mRNA production and are reversed by cyclic AMP. In an adenyl cyclasedeficient mutant strain, lac mRNA is not synthesized until the cells are supplied with cyclic AMP. A mutant strain lacking a recently described "cyclic AMP receptor protein" does not make lac mRNA despite exogenous cyclic AMP. We conclude that cyclic AMP, cyclic AMP receptor protein, an intact promoter locus, and other unidentified factor or factors are necessary for efficient transcription of the lac operon.
Cyclic adenosine 3') 5'-monophosphate is currently believed to be required for the synthesis of inducible enzymes in bacteria and for the cell-free synthesis of fl-galactosidase (l-3). In recent reports we have described DNA-RNA hybridization experiments demonstrating that cyclic AMP' stimulates the synthesis of lac mRNA, both in growing Escherichiu coli and in cell-free extracts (4,5). In this paper, we present a relatively simple competition hybridization technique for the measurement of Zac mRNA concentrations in unlabeled cells. Results with this assay demonstrate regulation of Zac mRNA levels by cyclic AMP, thus confirming predictions made in our earlier studies. In addition, we examine lac transcription in mutant strains deficient in adenyl cyclase and in a recently described "cyclic AMP receptor protein" (6,7). Results with these strains support the notion that cyclic AMP and a normal cyclic AMP receptor protein as well as an intact Zac promotor gene are 1 The abbreviations used are: cyclic AMP, cyclic adenosine 3',5'-monophosphate; IPTG, isopropyl-B-n-thiogalactoside; SSC, standard saline citrate. required for Zac mRNA synthesis (S-10). We also show that cyclic AMP regulates Zac transcription in wild type strains under conditions of permanent as well as transient glucose repression. METHODS

AND MATERIALS
Bacteria, Media, and Chemicals-The bacterial strains employed in this report are listed in Table I. Growth media and chemical sources have been previously described (4).
Enzyme assay-fl-Galactosidase activity was determined as described by Pardee,Jacob,and Monod (11). One unit of enzyme is the amount which hydrolyzes 1 nmole of o-nitrophenylfl-n-galactoside per min at 28".
Phage DNA-Phage were obtained by heat induction of growing cultures of E. coli RV usually labeled with 14C-thymidine for several hours after induction.
After cell lysis, extracted phage were purified by banding in cesium chloride density gradients and deproteinized by phenol treatments as previously described (4).
Bacterial RNA-Labeled and unlabeled RNA were extracted from growing bacterial cultures by the method of Okamoto,Sugino,and Nomura (12). Preparations were either dialyzed against 4-fold concentrated SSC (SSC, 0.15 M NaCl, 0.015 M sodium citrate) or lyophilized to dryness and resuspended in 4-fold concentrated SSC prior to hybridization.
Hybridization Methods-Ah80 DNA and Xh80dlac DNA were immobilized on Schleicher and Schuell type B6 nitrocellulose filters according to the procedure of Gillespie and Spiegelman (13). Filters 8 mm in diameter, containing 0.03 to 0.3 pg of DNA, were punched out of 27-mm filters loaded with 0.33 to 3.3 pg of DNA.
RNA to be tested was incubated with DNA filters in 100 to 200 ~1 of 4-fold concentrated SSC at 75" for I5 to 20 hours.
In competition hybridization experiments, sufficient amounts of 3H-labeled RNA from an induced culture were used to saturate Zac DNA sites, and unlabeled competitor RNA was added in increasing amounts.
In direct hybridization experiments, small amounts of labeled RNA preparations were incubated with a large excess of Zac DNA.
The filters were then washed with 4-fold concentrated SSC, treated with ribonuclease, rewashed, and the content of radioactivity measured as described earlier (4).
Calculations-For competition hybridization, the radioac- Competition curves were then plotted, or the results were transformed to linear functions using the formula derived by Stubbs and Hall (14). This conversion permits an assessment of the relative concentrations of unlabeled Zac mRNA in the tested preparations on the basis of comparative slopes of the linear functions. Data are plotted according to the formula Q0 -Qc/Qc = f.C/tL, where Q0 represents counts per min of Zac hybridized in the absence of unlabeled competitor RNA, Qc equals counts per min of lac annealed in the presence of competitor RNA at concentration C, tL is the amount of labeled RNA used in each reaction mixture (and is therefore a constant in each experiment) and f is the proportion, by weight, of Zac mRNA in an unlabeled RNA preparation.
When Q0 -Qc/Qo is plotted against the concentration of test RNA (C), the resulting slope (f divided by the constant, tL) is a relative measure of the concentration of lac mRNA in each RNA preparation.
In competition experiments, the filters contained about 10 cpm of %-DNA at the settings employed for double isotope counting. 3H values were corrected for the amount of DNA on each filter. This procedure did not materially affect results, which comprised an average of three or four filters, but did reduce the standard error of the mean. In each experiment, the counts per min hybridizable to Xh80 DN,4 were assumed to represent the non-Zac background interactions and were subtracted from experimental points.
For direct hybridization experiments, 3H-labeled RNA preparations of known specific activity were incubated with control and Zac-containing DNA filters.
The percentage of 3H-labeled RNA annealing with each type of filter was calculated, and the "percentage of difference" was determined, as previously described (4). This value is a measure of accumulation of labeled lac mRNA during the period of labeling, and is principally dependent upon the rate of Zac mRNA synthesis (4).

RESULTS
As the first step in development of a competition hybridization assay, it was necessary to determine the amount of 3H-RNA   6000  I  I  I  1 5000 -

Ah80
-a----pg 3H-RNA FIQ. 1. DNA saturation curve. Filters containing 0.03 pg of Xh%Odlac DNA were incubated with increasing amounts of 3Hlabeled RNA extracted from a culture of E. coli 3000 induced for &galactosidase synthesis with 5 X 10-' M IPTG. The cells were labeled with aH-uridine (1 mCi per 5 X log cells) from 7 to 10 min after the addition of IPTG: snecific activitv of the aH-labeled RNA was 6.18 X lo6 cpm per ;g.----, based bn two experimental points, estimates the amount of "background" annealing to Ah80 DNA filters. required to saturate available Zac DNA sites. Increasing amounts of labeled RNA extracted from cells induced for @-galactosidase synthesis were hybridized with filters containing 0.03 pg of Xh80dZac DNA (Fig. 1). The marked decrease in the slope of the curve in Fig. 1 indicates that saturation is approached between 5 and 10 pg of RNA.
In competition experiments shown here, generally 6 to 8 pg of 3H-labeled RNA are incubated with each filter.
The sensitivity of the competition assay was evaluated by preparing unlabeled RNA from wild type cultures grown in the presence of increasing concentrations of inducer; we then compared the capacity of these RNA preparations to compete with labeled Zac mRNA for Zac DNA sites. As shown in Fig. 2A, increasing concentrations of IPTG induce increasing amounts of unlabeled RNA capable of competing for Zac sites. RNA from an uninduced culture competes minimally; although not shown, RNA from the Zac deletion strain, W4032, demonstrates no significant competing ability for the Zac DNA sites. When the competition curves are converted to linear functions, the slopes, which provide an estimate of Zac mRNA concentration, are similar to rates of /l-galactosidase synthesis (Fig. 2, B and C). The competition assay was employed to measure Zac mRNA concentrations when cyclic AMP levels are transiently lowered by glucose and when glucose-repressed cells are replenished with exogenous cyclic AMP.
As illustrated in Fig. 3, glucose lowers Zac mRNA levels to about 10% of the induced value, and cyclic AM.P restores the Zac mRNA concentration to at least 80% of the induced level. Although not shown in the figure, enzyme production in these cultures again parallels Zac mRNA concentrations.
These results indicate that regulation of enzyme synthesis by cyclic AMP, as well as by inducer, occurs at a transcriptional level. Cultures of E. coli 3000 growing in Medium A, 0.5yo glycerol, and thiamine, 5 pg per ml, were induced with three different concentrations of IPTG for 12.5 min; a fourth culture, to which no inducer was added, served as a control.
At the end of the induction period, the RNA was extracted from each cell culture.
Increasing amounts of each unlabeled RNA species were then incubated with filters containing 0.03 pg of ?&?OcZlac DNA in the presence of 6.2 rg of aH-labeled RNA from an induced culture.
The specific activity of the aH-labeled RNA was 3 X 104 cpm per pg, and 417 cpm (kc specific) were annealed at saturation in the absence of competing RNA. The percentage of Zacspecific counts per min competed by the unlabeled RNA preparations is plotted in Panel A. In Panel B, these results are plotted linearly according to the transforming formula described in the text. p-Galactosidase synthesis in each culture is depicted in IPTG 5 X IO+ M, 0-O. Increasing amounts of unlabeled RNA and 5 pg of 8H-labeled RNA from an induced culture were hydridized with filters containing 0.03 pg of hh80dZac DNA.
(Specific activity of the 8H-labeled RNA was 3.6 X lo6 cpm per pg and 4500 cpm (Zac specific) were annealed in the absence of competitor RNA.) Results are plotted linearly as discussed in the text. Because competition hybridization compares unlabeled RNA preparations, it is not subject to variations in rates of 3H-uridine incorporation observed particularly upon addition of glucose to wild type cells. In our previously reported direct hybridization experiments, we largely circumvented this difficulty by studying a mutant strain (E. coli 1103), unable to metabolize glucose because of deficiency of Enzyme I of the p-enolpyruvate transferase system but still susceptible to transient repression (4, 15). In Table II, we present data from direct hybridization experiments with E. coli 3.000. Although glucose substantially increases over-all 3H-uridine incorporation, the difference between percentage of input counts per min hybridized to Xh80dZac DNA and that hybridized to X/GO DNA correlates well both with enzyme synthesis and with Zac mRN-4 concentrations, as determined by competition hybridization.
In our earlier experiments with E. coli 1103, it was not, of course, possible to measure Zac mRN-4 synthesis during permanent (or catabolite) repression by glucose, since this strain does not metabolize glucose. We have shown elsewhere that high concentrations of cyclic AMP (3 X 10e3 M) can reverse catabolite repression of ,&galactosidase synthesis (8). We have now determined Zac mRNA production in wild type cells grown in glucose for several generations (Fig. 4). Induced cells make 2 to 3 times as much Zac mRNA or fl-galactosidase when supplied with adequate amounts of cyclic AMP.
As expected, very little enzyme or message is detected in the absence of inducer.
Because all three cultures are glucose grown, the extent of labeling by 3H-uridine is similar.
These results indicate that the effect of cyclic AMP in catabolite, as well as transient, repression is exerted upon transcriptional events. We felt that a more stringent test of the hypothesis that cyclic AMP is required for Zac mRNA synthesis could be performed with the adenyl cyclase-deficient str&in, E. coli 5336. This strain contains no detectable adenyl cyclase or cyclic AMP and is unable to metabolize a variety of carbohydrates until supplied with exogenous cyclic AMP.
AS shown in Fig. 5, when this Cultures of E. coli 3000 were grown in 0.5% glucose, 0.2% casamino acids, and thiamine, 5 rg per ml.
Medium A supplemented with 0.5yo glucose and 5 pg per ml of The culture was divided into three fractions; one served as an thiamine. Two cultures were induced for &galactosidase syn-uninduced control, one was induced with 10-a M IPTG, and one thesis with 5 X l(r4 M IPTG. One of these cultures was simul-was induced with IPTG in the presence of 5 X 10-* M cyclic AMP. taneously treated with 5 mM cyclic AMP. A third culture served After 7 min of induction, cells were removed for fl-galactosidase as an uninduced control. After 8 min of induction, samples were assay, and 1 min later the cultures were labeled for 120 set with removed for @-galactosidase assay and 2 min later the cultures 3H-uridine (1 mCi ner 1.6 X lOlo cells). The 8H-labeled RNA was were labeled for 120 set with aH-uridine (1 mCi per 10'0 cells). extracted in the presence of a large excess of unlabeled RNA from Samples (2 pg the remainder is divided equally between material sedimenting beyond 26 S and that sedimenting from 6 S to 14 S. Presumably, strain is induced with IPTG, it produces no more enzyme or the largest material represents intact, polycistronic Zac mRNA, Eat mRNA than is found in the uninduced culture.
When cyclic and the other species represent Zac mRNA in the various stages AMP is added, production of both returns to the range seen in of synthesis and degradation expected in view of its short halfinduced wild type strains.
life (4). For each class of RNA, however, the same low levels of A similar result is obtained with the competition hybridiza-Zac-specific material are found in the uninduced culture and in tion assay (Fig. 6). Unlabeled RNA from a culture induced in the induced culture not receiving cyclic AMP.
(As discussed in the absence of cyclic AMP competes poorly for Zac sites compared our previous report, this small amount of RNA hybridizing with the competing ability of RNA from a cyclic AMP-treated preferentially to Xh80dZac DNA probably consists primarily of culture. Therefore, both direct and competition hybridization the transcription product of the i gene (4, 17). A very small experiments with the mutant strain indicate that cyclic AMP amount of material transcribed from genes adjacent to and is required for the synthesis of Zac mRNA.
cotransduced with Zac and a minor amount of RNA transcribed The availability of an adenyl cyclase-deficient strain also from structural Zac genes may also be present.) This result afforded an opportunity to test the hypothesis that cyclic AMP suggests that the effect of cyclic AMP is not to prevent abortive prevents premature termination of Zac transcription (16). This transcription, but to augment frequency of transcription. proposal implies that small species of RNA homologous to Zac Experiments performed with Zac promoter mutants and with DNA might be found, in cultures of E. coli 5336 induced in the inhibitors of transcription support this suggestion (8-10, 18). absence of cyclic AMP.
Cultures exposed to cyclic AMP as A recent report from this laboratory (7) describes the isolawell as inducer would be expected to contain Zac mRNA rang-tion of a cyclic AMP receptor protein which binds cyclic AMP ing in size up to 30 S polycistronic material.
As described in and is believed to be required for cyclic AMP action both in viva detail in the legend to Fig. 7 Cultures of E. coli 5336, grown in Medium A with 0.5% glucose, 0.2% casamino acids, and thiamine, 5 pg per ml, were given UY3 M IPTG with or without 5 X 1O-3 M cyclic AMP. After 10 min of induction, the RNA was extracted, and increasing amounts from each culture were hybridized with filters containing 0.03 pg of Xh80dZac DNA in the presence of 6.3 pg of SH-labeled RNA from an induced culture of E. coli 3000.
(The specific activity of the aH-labeled RNA was 5.9 X lo4 cpm per pg, and 840 cpm (Zac specific) were annealed in the absence of competitor RNA.) The results are plotted linearly as described in "Methods and Materials." IPTG, O-O; IPTG and cyclic AMP, A-A. but contain high adenyl cyclase activity and high levels of cyclic AMP.
Moreover, they do not respond to additional exogenous cyclic AMP.
We measured Zac mRNA synthesis in one of these strains (E. coli 5333) and found minimal production of lac mRNA (or P-galactosidase) in response to inducer, even when extra cyclic AMP was supplied.
These data are presented in Table III and compared with results of experiments with wild type cells. These results demonstrate that efficient lac transcription cannot occur in strains lacking normal cyclic AMP receptor protein, despite the presence of a normal Zac operon, a gratuitous inducer, and cyclic AMP.
Previously reported experiments from this laboratory and others (8-10) have indicated that an intact &promoter gene is also required for normal regulation of Zuc enzyme synthesis by cyclic AMP and glucose. Lac-promoter mutants are generally inducible, but they synthesize much less enzyme than their parent strains (2 to 5%) and are less sensitive (or insensitive) to the effects of glucose or cyclic AMP upon enzyme synthesis. Contesse,C&pin,and Gros (19) have demonstrated by DNA-RNA hybridization that strains carrying a Zuc promoter mutation make much less Zuc mRNA during induction than do wild type strains.
We have confirmed this finding, using E. CO& X8047.
This strain contains the Ll deletion, extending from the latter part of the i gene through the p locus (10). It therefore synthesizes lac enzymes constitutively, but even in the presence of inducer makes only 2% as much enzyme as its The aH-labeled RNA preparations were then placed on 6 to 30% sucrose gradients and centrifuged for 135 min at 2" in a Spinco SW 65 rotor at 65,000 rpm. After 25 fractions were collected from each gradient, the optical density at 260 nm (O-O) was measured, and the counts per min precipitated by 5% trichloracetic acid were determined for each fraction (0---0). These results were not significantly different among the three gradients, and the averaged values are plotted. Fractions 1 through 7,8 through 13, 14 through 19, and 20 through 25 were pooled from each gradient, and samples containing 5 to 10 X lo5 cpm from each pool were hybridized with filters loaded with 0.15 pg of Xh80 or Xh8Odlac DNA.
Counts per min annealing preferentially to Zac DNA were computed, and the relative number of &-specific counts per min in each pool were determined. parent strain X7700.
Lac mRNA synthesis is also markedly impaired; the percentage of pulse labeled RNA found to hybridize preferentially to luc DNA was found to be 0.036y0, thus differing minimally from the 0.025% of uninduced wild type RNA capable of hybridizing to luc DNA. Cyclic AMP in concentrations up to 30 mM did not affect enzyme synthesis in glucose-grown cells and had no significant effect upon lac mRNA production, although both enzyme and luc mRNA production were augmented 2-to 3-fold by cyclic AMP in the parent strain grown under similar conditions. However, small alterations in the very low rate of luc transcription seen in the mutant strain could have occurred and been undetectable with our assay. Thus we cannot definitely exclude an effect by cyclic AMP upon lac transcription in promoter mutants, although it seems unlikely in view of correlations between luc mRNA and enzyme production seen under other conditions of cyclic AMP control. E. coli 5333 was grown in Medium A containing 0.5% glucose, 0.2% casamino acids, and thiamine, 5 @ per ml. Cultures were induced for/%galactosidase synthesis for 12 min with 1w3 M IPTG, alone or supplemented with 0.3 or 1 X lez M cyclic AMP. An uninduced culture was used as a control.
After 8 min of induction, samples were removed for enzyme determinations, and the cells were labeled with SH-uridine (1 mCi per 1.4 X 10"' cells) for the final 120 set of induction.
Uninduced and induced cultures of E. coli 3000, grown on Medium A, glycerol, and thiamine, were similarly treated.
Samples (2 pg) or the 3H-RNA preparations were then hybridized with 0.15 pg of Ah80 or XhSOdlac DNA in the presence of 40 pg of unlabeled RNA from E. coli W4032 and the results were expressed as "percentage of difference" in annealing to control and Zac-containing DNA filters. Our development of a competition hybridization method for measurement of Zac mRNA was stimulated by the reports by Stubbs and Hall (14) of a competition assay for mRNA specific for the tryptophan operon.
Because 3H-RNA from a constitutive tryptophan strain hybridized almost half as well with control as with tryptophan operon-containing DNA, these authors were obliged to purify labeled tryptophan mRNA for their assay. Initially we employed a parallel technique for the isolation of labeled Zac mRNA.
Although we were able to obtain such material in highly purified form, it proved to be relatively rapidly degraded during storage. Since %RNA from a strain induced for fi-galactosidase hybridizes only to a small extent with control DNA (3 to 10% of the radioactivity annealing to Zac-containing DNA), we found that reproducible, significant, and sensitive competition hybridization could be performed using unfractionated labeled RNA from an induced culture.
As we have shown in this report, the capacity of unlabeled RNA to compete with labeled material for Zac DNA sites depends upon the concentration of inducer and the presence of cyclic AMP in the culture from which the RNA is extracted.
These results therefor confirm previous findings that cyclic AMP as well as inducer regulates Zac transcription.
Because the competition hybridization assay measures concentration of Zac mRNA, rather than the rate of Zac mRNA accumulation detected by direct hybridization, it is possible, as illustrated by Stubbs and Hall (14) for tryptophan mRNA, to estimate the number of lac mRNA molecules per cell in a given culture.
In the saturation experiment shown in Fig. 1, the filters contain 0.03 pg of Xh80dlac DNA.
The number of phage DNA copies can be calculated by multiplying this amount of phage DNA (3 X lo-8 g) by Avogadro's number (approximately 6 X 10%) and dividing the product by the molecular weight of the phage genome (3 X 10' daltons).
The result, 6 X 108 copies, provides the number of Zac operons on each filter. At saturation, when presumably 6 x lo* copies of Zac mRNA are annealed to an equal number of lac DNA copies, about 5000 cpm of Zuc mRNA are hybridized. When 0.5 pg of the same 3H-labeled RNA preparation from an induced culture is hybridized to excess XhSOdluc DNA, about 2000 cpm are annealed. Assuming Zac hybridization to be virtually complete under these conditions, 1 pg of such RNA contains about 5 X lo* copies of Zac mRNA.
With our extraction method, we obtain approximately 20 mg of RNA from 1Ol2 cells, or about 2 X 10-E pg of RNA from each cell. Therefore, we estimate that the average cell in an induced culture contains about 10 copies of Zac mRNA. Most cells in uninduced cultures or in cultures unable to synthesize or utilize cyclic AMP have no copies. When cyclic AMP levels are transiently lowered by glucose, cells contain about 1 copy per cell. Since the actual amount of RNA per cell may be somewhat higher than our estimate and since Zuc hybridization may not be more than two-thirds complete under conditions of DNA excess, it is possible that these calculations account for as little as half the actual number of Zuc mRNA copies. However, the minimum value of 10 copies per cell for an induced culture is comparable to the minimum of 8 copies of tryptophan mRNA per derepressed cell as determined by Stubbs and Hall (14) in a similar fashion.
In our previous report of regulation of Zuc mRNA synthesis by cyclic AMP, we used direct hybridization methods which measure accumulation of labeled Zuc mRNA (4). To show that accumulation was not being affected by changes in rates of Zac mRNA degradation, we measured the decay of labeled Zac mRNA.
The half-life proved to be short (60 to 90 set) and unaffected by cyclic AMP.
Hence we assumed that cyclic AMP augments accumulation of lac mRNA by increasing its synthesis, and we predicted that luc mRNA concentrations would vary in proportion to the rates of synthesis. As shown by the data in Table II, this prediction has now been confirmed by competition hybridization experiments. Our first report of hybridization studies of cyclic AMP control of Zuc transcription demonstrated the effect of cyclic AMP only during the severe transient repression observed in the first several minutes after addition of glucose to a growing culture.
We have shown elsewhere that cyclic AMP also reverses the permanent, milder form of repression (catabolite repression) of inducible enzyme synthesis caused by prolonged growth on rich media (8). In this paper, we report that reversal of catabolite repression by cyclic AMP is manifest by stimulation of Zac mRNA as well as fl-galactosidase synthesis.
Further evidence in support of the hypothesis that cyclic AMP is required for Zuc transcription is provided by hybridization experiments with two recently isolated mutant strains. Induced cultures of E. coli 5336, which lacks deny1 cyclase, appear to synthesize no more Zuc mRNA (or @galactosidase) than uninduced cells unless supplied with exogenous cyclic AMP. Moreover, the contention that cyclic AMP prevents premature termination of Zuc transcription seems improbable in view of the absence of small or large pieces of Zuc mRNA in cultures not supplied with cyclic AMP, although RNA fragments of less than 10 to 15 nucleotides would not be detected by our assay. The second mutant strain, E. coli 5333, lacks a normal cyclic AMP receptor protein and therefore cannot bind cyclic AMP or synthesize the inducible enzymes dependent upon cyclic AMP.
Moreover, cell-free extracts of this strain synthesized P-galactosidase at a low rate unless supplied with cyclic AMP receptor protein from a wild type strain (7). As expected, In attempting to determine whether cyclic AMP acts directly upon transcription, we have previously shown that cyclic AMP augments Zac mRNA synthesis in cells in which enzyme production has been halted by the withdrawal of an essential amino acid (4). Moreover, in recent experiments with several antibiotics which block protein synthesis (including chloramphenicol, streptomycin, spectinomycin, and puromycin), responsiveness to cyclic AMP remains in the absence of detectable enzyme synthesis.3 These experiments, however, also favor the notion that Zac transcription and translation are tightly coupled processes, since all the antibiotics thus far tested, excepting puromytin, severely inhibit Zac mRNA accumulation without decreasing total RNA synthesis.
Nevertheless, ribosomal movement along Zac mRNA may occur under such conditions, even without complete enzyme synthesis, and provide adequate translational machinery to serve as an agency through which cyclic AMP could affect the slowed transcriptional events (24). Recent results from our studies of Zac transcription in a cellfree system, however, indicate that a stimulatory effect of cyclic AMP can be produced in reaction mixtures containing only lac DNA and an RNA polymerase-containing extract from ribosomes (25). These experiments, and other experiments in which stimulation of in vitro lac transcription by cyclic AMP is preserved despite interference with protein synthesis, suggest that the transcriptional effect of cyclic AMP is direct and not merely mediated through some translational event. Experimenk which examine the actions of cyclic AMP upon translation of natural Zac mRNA added to a cell-free P-galactosidase-synthesizing system are needed to resolve the question of translational regulation by cyclic AMP.
The possibility that transcriptional control by cyclic AMP might be exerted through translational events raised the further possibility that stimulation of Zac mRNA production by inducer might result from represser-operator interactions at a translational level. Since Zac represser binds to double stranded but not single stranded Zac DNA, we reasoned that a preliminary test of the hypothesis would be to look for double stranded Zac RNA (26). Pulse-labeled RNA preparations from induced and uninduced cultures were digested wit.h Tl and pancreatic ribonuclease.
The 1 to 1.5% of the original trichloracetic acidprecipitable material remaining was then heat denatured and hybridized.
No Zac-specific material was found in either the induced or uninduced preparation, although the sensitivity of the assay was sufficient to detect sequences as short as 15 nucleotides.
Our results suggest at present that several ingredients are required for control of Zac transcription by cyclic AMP in E. coli: adenyl cyclase, cyclic AMP receptor protein, DNA-dependent RNA polymerase, and Zac operon DNA containing a normal Zac promoter locus. However, we have been unable to demonstrate stimulation of Zac transcription by cyclic AMP in simplified reaction mixtures containing cyclic AMP receptor protein, RNA polymerase, and Xh80dZac DNA.4 Perhaps other as yet unidentified factors will be implicated in the regulation of transcription by cyclic AMP.