Gene Expression in Vitro from Deoxyribonucleic Acid of Bacteriophage T7*

Abstract The in vitro synthesis of lysozyme and RNA polymerase directed by DNA of bacteriophage T7 starts with different lag periods after addition of the template. The difference in lag period before appearance of enzyme could be attributed to a difference in the time needed for transcription. Assuming that the difference in lag phase is caused by a difference in the distance between the promoter and the corresponding gene, and taking into account the average rates of transcription and translation, the locations of the polymerase gene and the lysozyme gene on the phage genome were estimated. The distance between the promoter and the distal end of the lysozyme gene corresponds to an RNA of the molecular weight 2.2 x 106. The polymerase gene is closer to the promoter. Transcription of T7 DNA by Escherichia coli RNA polymerase in vitro does indeed result in a polycistronic messenger RNA containing the information for synthesis of both RNA polymerase and lysozyme.

The difference in lag period before appearance of enzyme could be attributed to a difference in the time needed for transcription.
Assuming that the difference in lag phase is caused by a difference in the distance between the promoter and the corresponding gene, and taking into account the average rates of transcription and translation, the locations of the polymerase gene and the lysozyme gene on the phage genome were estimated. The distance between the promoter and the distal end of the lysozyme gene corresponds to an RNA of the molecular weight 2.2 X 106. The polymerase gene is closer to the promoter.
Transcription of T7 DNA by Escherichia coli RNA polymerase in vitro does indeed result in a polycistronic messenger RNA containing the information for synthesis of both RNA polymerase and lysozyme.
A bacterial cell infected by a bacteriophage provides a simple system for studying the molecular basis of the regulation of gene expression.
Bacteriophage T7 is particularly well suited for such a study since it is a relatively small phage. Its DNA has a molecular weight of approximately 2.7 x 10'. Its genetics is well established (1) and T7 has a simple mechanism for activating phage functions (2). After infection, the RNA polymerase of the host cell transcribes efficiently only a minor portion of the T7 DNA (3), but this portion apparently contains the information for a new RNA polymerase (specified by gene 1) (4), which in turn efficiently transcribes the rest of the T7 genome.
In the absence of gene 1 function, only the gene 1 protein, a protein of molecular weight close to 40,000, and one or more smaller proteins are made as efficiently as they would be in a wild type infection (2).
Previous experiments from our laboratory had shown that in an * This investigation was supported in part by Grant  zn vitro enzyme synthesis system, dependent on phage T4 DNA, synthesis of P-glucosyltransferase starts after a shorter lag phase than does synthesis of lysozyme (5).
The difference in the lag periods before synthesis of these enzymes was attributed to a process at the transcriptional level after the initiation of RNA synthesis (5). Two possible reasons were discussed for the difference in the lag phases (5). The delay in the appearance of lysozyme relative to P-glucosyltransferase in vitro could result from (a) a greater distance between the lysozyme gene and the start of transcription, or (6) a difference in the rate of transcription of the lysozyme and /3-glucosyltransferase genes. We could not distinguish between these alternatives. As a possible approach to resolve this problem we have used an in vitro T7 phage system. Here we present data which indicate that the appearance of two T7 enzymes, RNA polymerase (6, 7) and lysozyme (8), during in vitro protein synthesis, is a useful model to study the causes of different lag periods in enzyme synthesis. We found that with T7 DNA Escherichia coli polymerase transcribes a polycistronic messenger RNA containing the messages for both T7 RNA polymerase and lysozyme.
The lag phase of enzyme synthesis reflects the location of the corresponding gene relative to the starting point of synthesis of the polycistronic messenger RNA.
Since E. coli polymerase starts near one end of T7 DNA (9), the lag phases of enzyme synthesis give information about the positions of the genes on the genome.

EXPERIMENTAL PROCEDURE
Bacteriophages-Preparation of wild type and amber mutants of T7 was carried out as in Reference 2. T7 amber mutants were gifts of Dr. Studier and Dr. Hausmann.
T7 am 13 was isolated by Dr. Studier.
T3 phage were grown in E. coli B by infecting at a concentration of 7 to 8 x 108 cells per ml in Tryptone broth, multiplicity of infection l:lOO, at 30", and waiting for complete lysis. T7 and T3 phage were purified by polyethylene glycol precipitation (10) and either two consecutive CsCl centrifugations, or sucrose gradient centrifugation followed by one banding in CsCl (1). Phage DNA was extracted by a mild phenol treatment and dialysis (11).
Bacteria-For preparation of the cell-free systems, E. coli 514 (K-12 (X) lac-) was used as the permissive strain, that is, in the majority of experiments, and E. coli K-38 su-(obtained from Dr. G. Streisinger) as the nonpermissive strain.
E. coli 514 was inhibited with trimethoprim for 30 min immediately before harvesting (12). For the cell-free systems cells were grown in rich (Tris, 0.01 M, pH 7.5) after growth to 5 X lo* cells per ml.
The protein-synthesizing system was prepared as described previously (14). With the exceptions mentioned, ribosomes and supernatant enzymes were from a single extract prepared from E. coli 514. Therefore, changes in the capacity to synthesize proteins from one system to another were largely avoided.
Incubations for protein synthesis were identical with those described earlier (14). The After 15 min at 37", RNA synthesis was stopped by the addition of actinomycin D (20 pg per ml). Then translation was allowed to take place by dilution of this reaction mixture to a final volume of 0.05 ml by the addition of 50 pg of stripped tRNA (from E. coli K-12 or R), 500 pg of ribosomes, 120 pg of SlOO protein, and (in final concen trations) 100 mM NH&l, 0.1 mM N5-formyltetrahydrofolic acid, and varying amounts of magnesium acetate.
Incubation was continued for an additional 15 min. E. coli RNA polymerase was prepared either according to the method of Burgess (15) or, for the in vitro synthesis and purification of T7 mRNA, as previously described (16).
In 'Vitro Synthesized T7 Messenger RNA-As in the "uncoupled" system, T7 mRNA was transcribed from T7 DNA with highly purified RNA polymerase from E. co& The conditions for synthesis were as previously published (16) but with T7 DNA (100 pg per ml) and different ionic conditions: 130 mM NH,Cl and 10 mM magnesium acetate.
After 30 min of synthesis the reaction mixture (1.5 ml) was made to 0.2oj, in sodium dodecyl sulfate and the RNA isolated by sedimentation through CsCl (18) and two subsequent ethanol precipitations.
The RNA was redissolved in 0.5 ml of sodium dodecyl sulfate buffer containing 0.1 M NaCl, 1 mM EDTA, 10 mM Tris-acetate, pH 7.5, and (0.50/ w/v, sodium dodecyl sulfate) and centrifuged through a 12.5 ml of sucrose gradient (5 to 30%, w/v, in the same buffer mixture) at 41,000 rpm and 20" for 23 hours in the Spinco SW 41 rotor. The main RNA peak was precipitated two times with ethanol and finally dissolved in 0.5 ml of 10 mM Tris-acetate, 1 mM EDTA, pH 7.9.

RESULTS
Magnesium Ion Optima of Enzyme Synthesis-Like enzyme formation directed by T4 DNA, (19), T7 DNS-directed enzyme synthesis in vitro is critically dependent on t'he magnesium ion concentration. Fig. 1 shows the influence of different magnesium ion concentrations on synthesis in vitro of T7 RNA polymerase and lysozyme.
RNA polymerase was synthesized best at 10 to 11 mM Mg2"f. This optimum was similar to the one for in vitro synthesis of T4 P-glucosyltransferase (19) or dCMP deaminase (20). However, the curve for T7 lysozyme synthesis differed remarkably from the curve for production of the T4 enzyme (19). Synthesis of T7 lysozyme was maximal at approximately 11 and 13 mM Mg2+ and had a distinct shoulder at 15 mM (Fig. l), whereas synthesis of T4 lysozyme is best at a Mg2+ concentration of 15 mM (19). To determine whether one of the T7-lysozyme optima was attributable to optimal translation conditions, RNA was synthesized using purified RNA polpmerase, T7 DNA, nucleoside triphosphates and salts (10 mM magnesium). This in vitro produced RNA was then translated in the cell-free system in the presence of various magnesium ion concentrations.
The Mgzf optimum for lysozyme translation was found to be 13 rnrvr (Fig. 1). In the coupled system, the optimum at 10 to 11 mM was in the same range of magnesium ion concentration, in which DNA-dependent in vitro synthesis of "early" phage enzymes pro- Kinetics of Appearance of Enzyme Activity and Messenger RNA and of Initiation of Messenger RNA Synthesis-The first type of kinetic experiment measures the minimum time for synthesis of complete enzyme molecules.
In this case chloramphenicol is added at various times to block translation, and only those proteins which have been completed (or completed enough to exhibit enzymatic activity) will be measured.
When this experiment was done at 11 mM Mg2+, rifampicin-resistant UMP-incorporating activity (T7 RNA polymerase) first appeared between 2 to 3 min and lysozyme activity approximately 7 min after the start of the reaction.
At 16 mM magnesium, lysozyme appeared after approximately 5 min (Fig. 2). The second type of kinetic measurement gives an indication of the minimum time necessary for the completion of the message for a particular enzyme. In these experiments actinomycin is added at various times to stop the growth of RNA chains, and then the incubation is continued to permit the translation of the messenger RNA that had been synthesized.
Enzymatic activity will appear only if the RNA chain includes the complete message for that enzyme. When this experiment was done at 11 mM magnesium, lysozyme message did not appear until between 4 and 5 min after the start of the reaction.
In order to measure the kinetics of initiation of messenger RNA The exneriment was performed as described in Fig. 2 synthesis rifampicin is added at various times to prevent further initiation of RNA chains and, then, the reaction mixture is further incubated to permit completion of RNA chains that had already been started.
The earliest time of addition of rifampicin that will still permit the appearance of enzymatic activity gives an indication of the time at which the messenger RNA chain carrying the information for that enzyme was initiated. From Fig. 4 it appears that the messenger RNA chains for lysozyme were initiated within the first minutes at either 11 or 15 mM Mg2f (Fig. 4).
These experiments on Mgaf optima and kinetics give slightly variable results depending on the particular preparations used. However, in sets of experiments done at the same time, using the same preparations, the same relative differences in the Mgz+ optima and in the kinetics of appearance of the two enzymatic activities were repeatedly observed. Characterization of in Vitro Synthesized 2Messenger RNA-Translation may be completely separated from transcription by carrying out the two synthetic steps in separate reactions. RNA was first transcribed from T7 DNA in vitro with purified RNA polymerase (21) from E. coli and isolated by treatment with sodium dodecyl sulfate and precipitation by centrifugation through CsCl (see "Experimental Procedure"). When analyzed by polyacrylamide gel electrophoresis, a large portion of this RNA appeared as a distinct peak with molecular weight of 2.2 X lo6 (21). The main component was further purified from the minor lower and higher molecular weight components by sucrose density gradient centrifugation and again analyzed by gel electrophoresis (Fig. 5).
We previously reported that this RNA species is homogeneous with respect to its 3' terminus (terminating in uridine) and that it is an active template for the synthesis of lysozyme ik vitro (21). The RNA was purified as described under "Experimental Procedure." The major RNA species of 2.2 X 10" daltons was isolated by zone centrifugation and subjected to polyacrylamide gel electrophoresis in 2.4% acrylamide (9 X 0.5 cm) with 0.2% sodium dodecyl sulfate, 10 volts per cm per gel, pH 7.2 (21). The gels were sliced into 1.5-mm discs. The abscissa gives the slice number (starting from the origin). This RNA is not only an active template in vitro for the synthesis of lysozyme, but also for T7 RNA polymerase (gene 1 product), as is shown by the appearance of a rifampicin-resistant UMPincorporating activity ( Table I). Inhibition of protein synthesis by chloramphenicol prevented formation of T7 polymerase. In the T7 DNA-directed enzyme synthesis more lysozyme was made in comparison to polymerase than in the RNA-dependent syn- thesis with purified RNA (Table I). This might be due t.o the secondary structure of the 2.2 X lo6 dalton RNA. A hindrance of translation by the secondary structure of mRNA would affect the translation of distal message more than the translation of information at the 5'-end, assuming that RNA in vitro is also translated starting from internal initiation points (Fig. 8 in Reference 21, not so in DNA-dependent protein synthesis in vitro). This explanation is supported by the fact that with natural T7 mRNA as template, the ratio of lysozyme to polymerase is also lower than that observed with T7 DNA (Table I).
That the appearance of T7 lysozyme and polymerase is due to de novo synthesis in vitro is demonstrated by the inhibition of their formation by inhibitors of protein synthesis (Tables I and  II) and by the inability of DNA from RNA polymerase and lysozyme negative mutants of T7 to direct their synthesis (Table II). DISCUSSION Upon infection, T7 genes are expressed in an orderly progression, proceeding generally from left to right on the genetic map, the same direction as transcription.
Among the first proteins to be made is the T7 RNA polymerase specified by gene 1, the leftmost gene identified in T7 so far (2,4). If the gene 1 protein is not functional, only three or four proteins, the first to appear after infection, are made with normal efficiency; the others are produced in niuch lower amounts (2). The evidence available so far indicates that E. coli RNA polymerase can transcribe efficiently the messengers for only the first three or four T7 proteins to be made, and that the T7 RNA polymerase is necessary for efficient transcription of the remaining T7 genes (4). Davis and Hyman (9)  In both cases, DNA is added to a system which is capable of expressing genetic information by synthesizing active enzyme molecules. If the control of gene expression is the same in both cases, one might expect that the three or four proteins made efficiently in vivo by gene 1 mutants would be made most efficiently in vitro, and the other T7 proteins (including lysozyme) would be made efficiently only if active T7 RNA polymerase were synthesized by the system. However, lysozyme seems to be made in vitro equally well whether active T7 RNA polymerase is present or not, as is shown by synthesis of lysozyme in vitro directed by DNA from phage with mutations in the polymerase gene (gene 1) (Table II). Thus, most lysozyme transcription in the in vitro system (which is prepared from uninfected cells) is due to the E. coli RNA polymerase; any T7 RNA polymerase that is made contributes little to the production of lysozyme under these reaction conditions.
Where does E. coli RNA polymerase initiate transcription under in vitro protein synthesis conditions? At 11 mM Mg2+ it must initiate to the left, of gene 1 since active T7 polymerase is made. Under these conditions, T7 RNA polymerase first, appears between 3 and 4 min earlier than does lysozyme (Fig. 2), thus placing the gene for T7 RNA polymerase (approximately 3000 nucleotides long) (2) closer to the initiation point than the gene for lysozyme.
The approximate speeds of transcription and translation in this system are 28 nucleotides per set and 5 amino acids per sec.' From this, along with the data of Fig. 2 Thus, these data fit well with the idea of a single initiation point immediately to the left, of gene 1, possibly corresponding to the single starting point found by Davis and Hyman (9). Furthermore, they agree with the observed in vitro transcription of a single mRNA molecule of 2.2 x lo6 daltons. At 16 mM Mg2+, lysozyme is made almost as efficiently as at 11 mM, but T7 RNA polymerase is synthesized less well (Fig. 1). The earliest appearance of lysozyme mRNA is about 1 min earlier at 16 mM Mg2+ than it is at 11 mM (Fig. 3). This kinetic difference suggests that at 16 mM Mg2+ either (a) transcription is faster or (b) the initiation point is closer to the lysozyme gene by approximately 0.6 x lo6 daltons of messenger RNA.
The Mg2+ concentrations discussed here are only valid for the complete system. We do not know the effective Mg2+ concentration in our system because of binding of magnesium ions, for instance to nucleic acids, nucleotides, etc. Thus, we do not know which Mg2+ concentration of a pure transcription system (DNA plus RNA polymerase) would correspond to our values, and therefore we could not obtain evidence concerning the dependence of RNA size on MgZ+ concentration.
The shape of the curve for dependence of lysozyme synthesis on magnesium ion concentration is of interest.
Whereas all typical early phage enzymes tested are synthesized best at a magnesium concentration of around 11 mM (19), the magnesium optimum for in vitro synthesis of the "not early" proteins is at 15 mM (19). There seemed to be one exception, as we reported earlier (20 Phage SP82 induces two different dCMP deaminases which were separated by gel electrophoresis (22). dCMP deaminase activity is synthesized in the early as well as in the late phase. The shape of the curve of the magnesium ion dependence of in vitro synthesis can be discussed as a superposition of two curves for the two enzymes: one for the "early" dCMP deaminase and another for the "late" enzyme, resulting in a curve with optima at 11 and 15 mM, respectively.
The similarity of the shape of the curve for T7 lysozyme synthesis could suggest that either two lysozymes are induced or that one lysozyme gene is transcribed by polymerase using two different promoters, an "early" and a "not early" one, thereby causing the two optima at 11 and 15 mM. Genetic results indicate that T7 induces only one lysozyme protein. 2 The existence of two magnesium ion optima for T7 lysozyme synthesis suggests that the corresponding gene could be transcribed starting from two different promoters. From the kinetics of synthesis of lysozyme and lysozyme messenger, the distance between the early promoter and the end of the lysozyme gene was found to correspond to an RNA of molecular weight 2.2 x 106, the same size as the polycistronic messenger which is synthesized in vitro by E. coli polymerase on T7 DNA. As we reported earlier, this RNA carries the information for lysozyme synthesis (21). Since the gene for T7 RNA polymerase is located near the promoter, we expected this messenger RNA to code also for synthesis of this enzyme.
This was found to be the case, as shown in Table II. Thus, E. coli polymerase can transcribe from T7 DNA in vitro a polycistronic messenger which contains the information for polymerase and lysozyme.
The lengths of the lag periods prior to synthesis of polymerase (2 to 4 min) and of lysozyme (7 to 9 min) indicate that the polymerase gene is located near the promoter, whereas the lysozyme gene is further away. The location of the lysozyme gene on the T7 genome suggested by in vitro studies was confirmed by genetic mapping.
It was found that the lysozyme gene is indeed located in the expected area between genes 3 and 4.2