Purification and Characterization of Bacteriophage gh-l-induced Deoxyribonucleic Acid-dependent Ribonucleic Acid Polymerase from Pseudomo.nas putida”

Infection of Pseudomonas putida by the bacteriophage gh-L-induced the synthesis of a novel DNA-dependent RNA polymerase. This gh-L-induced RNA polymerase was purified to near homogeneity. It was shown to be distinct from the host RNA polymerase (alpha-2 beta beta sigma) physically and in respect to many of its catalytic properties. The gh-L-induced RNA polymerase was composed of a single polypeptide of approximately 98,000 molecular weight. The divalent metal ion requirement for in vitro RNA synthesis by the gh-L-polymerase could be satisified with Mg-2+, but not with Mn-2+. Rna synthesis by the gh-L polymerase was highly resistant to inhibition by rifampicin and streptolydigin but could be inhibited by relatively low concentrations of KCl or the rifamycin derivative AF/013. The structural analog of ATP, 3'-deoxyadenosine 5'-triphosphate, inhibited both the gh-L-induced and the host RNA polymerases by competing for a single binding site with ATP. The phage polymerase was extremely sensitive to this inhibitor, exhibiting an apparent K-i value (2 times 10-8 M) approximately 100 times lower than that for the host RNA polymerase. The gh-L polymerase had a highly specific template requirement for DNA from the homologous gh-L phage. It would not efficiently utilize denatured DNA templates and had only low levels of activity with pyrimidine-containing polydeoxyribonucleotide homopolymers.

When a bacterial cell becomes infected with a virulent bacteriophage, a shift in RNA synthesis occurs from entirely host-specific (transcription from the host DNA) to largely phage-specific (transcription from the viral DNA). There are two general types of mechanisms by which this shift in transcription can occur. In one mechanism, the host DNA-dependent RNA polymerase is utilized throughout the infectious cycle for the transcription of all classes of viral genes. Modifications of the host RNA polymerase in the viral infected cell, however, alter the specificity of the enzyme to program changes in transcription during the infectious cycle. This mechanism most likely occurs in T4 and X bacteriophage infections of Escherichia coli (1, 2), SPOl and SP82 infections of Bacillus subtilis (3,4), and 429 bacteriophage infection of Bacillus amyloliquefaciens (5). The exact nature of the modification causing altered specificity of the host RNA polymerase is unknown.
Several chemical alterations of the subunits of the host RNA polymerase have been demonstrated after T4 infection of E. coli (6)(7)(8).
Furthermore, polypeptides, some of which have been shown to be the products of T4 regulatory genes, have been found to be associated with the host RNA polymerase in T4-infected cells (9,10). It has not been demonstrated, however, which, if any, of these modifications confer altered transcriptional specificity to the host RNA polymerase.
The second mechanism to account for the shift in DNA transcription after bacteriophage infection involves the synthesis of a new, viral coded DNA-dependent RNA polymerase. This mechanism has been shown to occur in both T3 and T7 infection of E. coli (11)(12)(13)(14).
The new RNA polymerases synthesized after infection by these coliphages are quite different from the host RNA polymerase in both structure and catalytic properties. These phage-induced RNA polymerases are composed of single polypeptides of approximately 108,000 to 110,000 molecular weight (11,14). The E. coli RNA polymerase is composed of five subunits, a&Ya, with a combined molecular weight of 470,000 (15). The phage-induced RNA polymerases show highly stringent template specificities in vitro for their homologous phage DNA, whereas the host RNA polymerase can utilize DNA from a variety of sources (11,13,14). A comparison of other properties of these two types of RNA polymerases has been presented recently (16).
We have examined the regulation of RNA synthesis after the infection of Pseudomonas pulida by the bacteriophage gh-1. gh-1 is a small, virulent bacteriophage isolated in this laboratory (17) (20). The preparation of RNA polymerase used in these studies was more than 95% pure, as determined by SDS-polyacrylamide gel electrophoresis. RESULTS A Novel RNA Polymerase Activity in Bacteriophage gh-l-infected P. p&da-The first evidence that a novel RNA polymerase is synthesized after gh-1 infection of P. putida was obtained from measurements of the RiXA polymerase activity in extracts of uninfected and gh-l-infected cells. In extracts from uninfected cells, RNA polymerase activity was inhibited 97 y0 by the addition to the reaction mixture of the antibiotics, rifampicin and streptolydigin (Table I). This activity is largely, if not entirely, due to the P. putida RKA polymerase, which is known to be sensitive to these antibiotics (20). In extracts from gh-linfected cells, the specific activity of RNA polymerase was 11 times greater than the specific activity in extracts from uninfected cells. Furthermore, this activity from infected cells was inhibited only 4y0 by the addition to the reaction mixture of the two bacterial RNA polymerase inhibitors.
Addition to the reaction mixture of actinomycin D and nogalamycin, which inhibit RNA synthesis by binding to DNA, almost completely inhibited the activity from extracts of both uninfected and gh-l-infected 1725 cells (32,33). These activities were, therefore, due to DNAdirected processes. In extracts from cells infected with gh-1 in the presence of chloramphenicol, the specific activity of RNA polymerase was essentially the same as that in uninfected cells. This activity also was sensitive to rifampicin and streptolydigin. Thus, protein synthesis was necessary for the appearance of the rifampicin-and streptolydigin-resistant RNA polymerase activity.
Although other interpretations are possible, these results can be explained most readily by the synthesis of a novel DNAdependent RNA polymerase after gh-1 infection of P. putida. This explanation was verified by the purification of the gh-linduced RNA polymerase and by a sbudy of its structure and catalytic properties.
Puri$cation of the gh-l-induced RNA Polymerase-The results of the purification of the gh-l-induced RNA polymerase, performed as described under "Experimental l'rocedure," are shown in Table II. The Rio-Gel fraction which was used for many of the catalytic studies reported below had a specific enzyme activity of 42,000 units per mg. This represents a 280.fold purification from the initial extract fraction.
An accurate determination of the specific enzyme activity of the glycerol gradient fraction could not be made due to the difficulty of determining protein concentration at the relatively low level present in this fraction. An estimate of the protein concentration of the glycerol gradient fraction, however, could be made from the SDS-polyacrylamide gel electrophoresis of this fraction (Fig. IC). By measuring the area under the peaks of the scan at 550 nm of the SDS-polyacrylamide gel and comparing with the area under the peaks of known amounts of the reference proteins, the amount of protein present in the gel could be determined.
This determination is dependent on the demonstration that the amount of stain absorbed by the SDS-polyacrylamide gel is linearly related to the amount of protein present (20). From this estimate of protein concentration, a specific enzyme activity of 86,000 units per mg was calculated for the glycerol gradient fraction.
Analysis of the Rio-Gel fraction for RNase and DNase activities, contaminants of RNA polymerase preparations which can alter the observed RNA polymerase activity, were negative. The Rio-Gel fraction also did not contain any RNase III activity, the enzyme involved in the "sizing" of T7 early mRNA in E.
coli (34,35). The Rio-Gel fraction also is free of any host RNA polymerase activity. Procedure." Electrophoresis was performed at 4 volts per cm of gel length for 6.25 hours at 25". After staining and destaining, the gels were scanned at 550 nm on a Gilford linear transport.
The direction of migration was from left to right. The arrows indicate the peak positions of the reference proteins: phosphorylase a (a), bovine serum albumin (b), and catalase (c). migrated significantly ahead of the 0 and /3' subunits of purified P. putida RNA polymerase.
The phosphocellulose and Bio-Gel fractions could be stored at -20" in buffer containing 50% glycerol for several months with little loss of activity, if the protein concentration was equal to or greater than 0.5 mg per ml. Analysis of the phosphocellulose fraction, the Bio-Gel fraction, and the glycerol gradient fraction was performed by SDS-polyacrylamide gel electrophoresis (Fig. 1). The glycerol gradient fraction contained one major polypeptide which comprised approximately SO%, by weight, of the total protein present (Fig.  1C). No other polypeptide present comprised more than 8% of the total protein.
The molecular weight of the major polypeptide was determined by comparison of its mobility to the mobility of the reference proteins, phosphorylase a (subunit molecular weight of 94,000), catalase (SS,OOO), and bovine serum albumin (60,000).
Using a standard curve of the logarithms of the molecular weights of the reference proteins to the distances of migration, a molecular weight of 97,000 was estimated for the major polypeptide of the glycerol gradient fraction.
This major polypeptide is thought to be the only polypeptide comprising the gh-1 polymerase.
It is the only polypeptide which increased in relative purity in the last two steps of the purification procedure. Its increase in purity parallels the increase in specific enzyme activity of gh-1 polymerase in these last two steps. Finally, the FRACTION  molecular weight of the gh-1 polymerase, as determined by SDSpolyacrylamide gel electrophoresis, is consistent with a determination of 98,000 made by glycerol gradient centrifugation and gel filtration (see below).
The gh-1 polymerase polypeptide was 10 to 157& by weight, of the total protein in the phosphocellulose fraction ( Fig. 1A) and 50 to 55% of the Bio-Gel fraction (Fig.  1B).
Molecular Weight and Structure of the gh-l-induced RNA Polymerase-The molecular weight of the gh-1 polymerase was calculated using experimentally obtained values for its sedimentation coefficient and molecular Stokes radius.
A molecular weight value calculated in this manner is not dependent on assumptions concerning the shape of the macromolecule (27). The sedimentation coefficient of gh-1 polymerase was determined by sedimentation velocity centrifugation in a 10 to 30% glycerol gradient (Fig. 2). The reference proteins (alkaline phosphatase, lactate dehydrogenase, and P. putida RXA polymerase) were centrifuged under identical conditions.
Based on the sedimentation coefficients of the reference proteins, the gh-1 polymerase exhibited a sedimentation coefficient of 6.1 & 0.2 S. The molecular Stokes radius of the gh-l-induced RKA polymerase was obtained by gel filtration on a P&Gel P200 column (Fig. 3) Characterization of RNA Synthesis by gh-1 -induced RNA Polymerase Using gh-i DNA as Template-The general requirements for in vitro RNA synthesis by the purified gh-1 polymerase were examined by varying the components of the standard reaction mixture (Table III).
When the enzyme, the gh-1 DNA, one of the four ribonucleoside triphosphates, or the Mg2+ was removed from the reaction mixture, little or no RNA synthesis occurred. Near maximal enzyme activity was maintained over a broad concentration range of 5 to 20 mM Mg2+ with the optimal activity occurring at approximately 10 mM (data not shown). No detectable RNA synthesis occurred when the Mg2+ was replaced in the standard reaction mixture by the divalent metal ions (Mn*+, Zn2+, or Ca2+) at concentrations between 0.5 and 8 mM (Table III).
In fact, the addition of any of these divalent metal ions at 2 mM to the reaction mixture containing Mg2f inhibited the enzyme activity 93 to 100%. The activity of the gh-1 polymerase was also inhibited quite markedly by relatively low concentrations of monovalent ions. At a concentration of 85 mM KCl, the gh-1 polymerase activity was inhibited 50%, whereas at 200 mM, the reaction was essentially completely inhibited. An almost identical inhibition of enzyme activity was observed with either NaCl or NH&l (data not shown).
Apparent K, values for each of the four ribonucleoside triphosphates which are substrates for RNA synthesis were determined. For these studies, the concentration of three of the ribonucleoside triphosphates was fixed at a high level, greater than 5 times the K, value for any ribonucleoside triphosphate.
The concentration of the fourth ribonucleoside triphosphate was varied and the initial reaction rates were measured at each concentration. To analyze the results, Michaelis-Menten kinetics was assumed applicable to this complex reaction, and the results were plotted in Lineweaver-Burk double reciprocal plots (I/V uertis l/[S]). All data were analyzed by a computer program to determine the highest correlation to a least squares straight line for the equation : v = v,,, -K,"(v/[NTP]") as n was varied in increments of 0.05 unit (37). An n value so determined is equivalent to the Hill coefficient, n, and should equal 1.0 if the double reciprocal plot is linear. For the purine ribonucleoside triphosphate ATP, the double reciprocal plot was linear (Fig. 4A). The apparent K, value for ATP was 3.5 X 1O-5 M. Likewise, the pyrimidine ribonucleoside triphosphates, CTP and UTP, yielded linear double reciprocal plots (data not shown).
The apparent K, value for both of these substrates in the RNA polymerase reaction was 4.0 x 10e5 M. For the purine ribonucleoside triphosphate GTP, however, the double reciprocal plot was curvilinear (Fig. 4A). An n value of 1.2 for GTP was determined by the computer analysis. Thus, the best fit to a straight line was obtained when l/v was plotted versus l/[GTP]'." (Fig. 4B). The kinetics of RNA synthesis at the lowest GTP concentration used in the K, study was linear for at least 5 min and showed no appreciable lag in initiation (data not shown).
Thus, the higher order n value is not due to nonlinear reaction rates at the lower substrate concentrations.
The apparent K, value for GTP, using the higher order value of substrate concentration in the Michaelis-Menten equation, was 8.0 X lop5 M or twice that seen for the other three ribonucleoside triphosphates.   esses. The activity of the gh-1 polymerase is highly resistant to the antibiotics, rifampicin and streptolydigin, present at concentrations which markedly inhibit the host RNA polymerase. The effect of 13 other derivatives of rifamycin on gh-1 polymerase activity was tested. The derivatives examined, using the nomenclature of Gruppo Lepetit, were: AF/AOP, Rifamycin AG, AF/APR, AF/DEI, AF/DA-AMP, Rifamide, 4-Dessosi SV, PR/ 14, AF/013, AF/ABDP-cis, AF/AP, AF/BO, AF/DNFI, and PR/19 (for review of structures see Ref. 40). All 13 of these derivatives were effective inhibitors ( >95%) of RNA synthesis by the host RNA polymerase when present at a concentration of 10 pg per ml. When added to the gh-1 polymerase reaction mixture at 100 pg per ml, seven of the derivatives (AF/O13, AF/DNFI, AF/BO, AF/AOP, AF/ABDP, PR/19, and AF/ DEI) were found to inhibit polymerase activity to a significant degree ( >20y0) (data not shown).
The most effective inhibitors were AF/013 and AF/DNFI, which inhibited RNA synthesis by 50% at concentrations of 35 pg per ml and almost completely at concentrations of 80 pg per ml or more. The relative order of effectiveness of the rifamycin derivatives in inhibiting gh-1 polymerase activity was virtually the same as that observed for T7-induced RNA polymerase (41). Inhibitors of the phage polymerase activities, however, were far more effective. Even those rifamycin derivatives which were more effective against the activity of the host RNA polymerase.
3'-Deoxyadenosine 5'.triphosphate, the triphosphate derivative of the antibiotic cordycepin, has been shown to be an i n vitro inhibitor of RNA synthesis by certain bacterial RNA polymerases (42,43). This ATP analog presumably inhibits RNA synthesis by being enzymatically incorporated into an RNA chain at a position normally occupied by an AMP residue. If incorporated, the 3'-dAMP would act as a chain terminator in RNA synthesis, because it does not contain a 3'.hydroxyl group necessary for the formation of the next phosphodiester bond. As shown in Fig. 6, 3'.dATP, when added to the standard reaction mixture, inhibited RNA synthesis by both the gh-l-induced and P. putida RNA polymerases.
It was a much more effective inhibitor, however, of the gh-1 polymerase.
The 3'-dATP concentration required to produce a given level of inhibition with the host RNA polymerase was approximately 80 times greater than that required to inhibit the gh-1 polymerase to the same extent. Thus, at the concentration of ATP present in the standard reaction mixture, 0.4 mM, 50% inhibition of the host polymerase occurred at an ATP :3'-dATP molar ratio of 20, whereas the same degree of inhibition of the phage enzyme occurred at an ATP :3'-dATP molar ratio of 1600. By selecting the appropriate concentration of 3'-dATP, the gh-1 polymerase activity can be essentially completely inhibited, whereas the host polymerase activity is almost completely unaffected.

Double reciprocal plots of l/v versus I/[ATP]
in the absence and presence of 3'-dATP were experimentally determined to study further this interesting inhibitory effect (Fig. 7).  The apparent Ki values for 3'.dAT1' were, however, quite different: 2 X lop6 M for the host enzyme and 2 X lo-* hr for the phage enzyme. Thus, the difference in sensitivity of the two enzymes toward 3'.dhTI', as seen in Fig. 7, was reflected in the relative difference of the apparent Ki values. These results indicate that 3'.dATP inhibited the polymerase by competing for a common binding site with ATI).
This conclusion was substantiated by the finding that the poly(dC) .poly(dG)-primed polymerization of GTP by the gh-1 polymerase (see below) was not affected by the presence of 3'-dRTP at levels which completely inhibit the gh-1 DNA-primed reaction (data not shown). Another structural analog of AT1 is 3'.0-methyladenosine 5'-triphosphatc.
3'.AmTl' is similar to 3'.dATP in that it differs from ATI' only at the 3' position of the ribose moiety. 3'-AmTl' was an inhibitor of RNA synthesis by both the P.
p&da and the gh-l-induced RNA polymerases (data not shown). The large differential inhibitory effect seen for these two RNA polymerases with 3'.dAT1' was not observed for 3'.AmTP.
The apparent K; value for 3'.AmTl' calculated for the P. putida RNA polymerase was 4.1 X lo-" M, or approximately 20 times higher than the apparent Ki value for 3'.dATP.
For the gh-1 polymerase, the apparent Ki value of 3'.AmTP was 1.3 x low4 M; over 3 orders of magnitude greater than that of 3'-dATP.
Thus, the 3'-0-methylated analog of ATl' is not as efficient as an inhibitor of in vitro RNA synthesis as the 3'-H analog for these RNA polymerases.
Template Specificity of gh-1 -induced RNA Polymerase-One of the most striking charact,eristics of the gh-l-induced RNA polymerase-catalyzed reaction is the stringent template specificity. When DNA from many sources was tested, only the homologous phage gh-1 DNA was found to be an efficient template for in vitro RNA synthesis (Table V). The gh-1 polymerase would not utilize DNA from coliphages T3, T4, or T7, nor would it utilize calf thymus or P. putida DNA.
When the gh-I DNA was denatured, it became an inefficient template for the gh-1 polymerase. Thus, some feature inherent in the double-stranded structure of the gh-1 phage DNA is necessary for its function as an efficient template.
Likewise, denatured DNA from either coliphage T3 or calf thymus supported little or no RNA synthesis. By contrast, the host RNA polymerase can utilize all of the above templates, although at varying efficiencies.
Several synthetic polydeoxyribonucleotides were tested as templates for RNA synthesis by the gh-1 polymerase (Table VI). The alternating copolymer, poly[d(A-T)], which was an efficient template for the host RNA polymerase, was not utilized effectively by the gh-1 polymerase.
The gh-1 polymerase utilized the homopolymer duplex, poly(dC) .poly(dG), to direct the polymerization of GTP at a rate 7 times higher than the polymerization of CTP from this template.
Several single-stranded polydeoxyribonucleotide homopolymers also were tested as templates for gh-1 polymerase.
Either of the pyrimidine-containing polymers, poly(dT) or poly(dC), would support the synthesis of the corresponding ribohomopolymers.
Little or no template activity, however, could be detected with the purine-containing homopolymers, poly(dA) or poly(d1).
Thus, with either singlestranded or double-stranded deoxyribonucleotide homopolymers, the gh-1 polymerase markedly prefers to utilize the pyrimidinecontaining templates as compared to the purine-containing ones. It should be noted that the highest enzyme activity on any template other than native gh-1 DNA, namely that for poly(dT), was less than 5y0 of the enzyme activity on native gh-1 DNA, in terms of total nanomoles of nucleotide incorporated per hour per mg of protein.

DISCUSSION
The infection of P. pulida by the bacteriophage gh-1 induces the synthesis of a novel DNA-dependent RNA polymerase. This gh-l-induced RNA polymerase has been purified to near homogeneity.
It is composed of a single polypeptide chain with a molecular weight of approximately 98,000. The structure of Low concentrations of monovalent ions, which do not appreciably affect the activity of the host polymerase, inhibit the gh-1 polymerase markedly.
The antibiotics, rifampicin and streptolydigin, inhibit the activity of the host enzyme at concentrations which do not affect the activity of the phage enzyme. The host RNA polymerase will utilize as an in vitro template every DNA with which it has been tested. The ability of the host polymerase to utilize a wide range of templates may be due to the diversity of sites it must recognize to perform its role in the transcription of the bacterial chromosome.
On the other hand, the gh-1 polymerase is highly specific in its template requirement for DNA from the homologous gh-1 phage.
The infection of E. coli by the coliphages T3 or T7 has been shown to induce the synthesis of viral specific RNA polymerases (11,13,14). These coliphage-induced RNA polymerases are similar in structure to the gh-1 polymerase; both are single polypeptides of 108,000 to 110,000 molecular weight (11,14). The induction of a novel RNA polymerase activity also has been demonstrated after infection of E. coli by the helper-dependent bacteriophage P4 (44). The 1'4.induced RNA polymerase could synthesize polyriboguanylic acid from the duplex homopolymer, poly(dC) .poly(dG) ; however, no naturally occurring DNA has been found yet to serve as an in vitro template for this enzyme. Its actual function, therefore, is still a rnatter of conjecture. These three phage-induced RNA polymerases of E. coli are the only bacteriophage-specific RNA polymerases which have been described previously.
A comparison of the catalytic properties of the gh-l-induced RNA polymerase with those of the T3 and 17 RNA polymerases shows that these three phage polymerases are quite similar (11,13,14). All three phage polymerases cannot utilize RI& as divalent metal ion in place of Mg2+. The activities of the phage polymerases were highly resistant to inhibition by rifampicin and streptolydigin but could be inhibited by the rifamycin derivative AF/013 at concentrations higher than 10 pg per ml (40,45). Low concentrations of monovalent ions inhibited the activities of the three phage enzymes. Finally, all three phage-induced RNA polymerases showed highly stringent specificities for DNA from the homologous bacteriophage as in vitro templates.
The stringent template specificities of the gh-1, T3, and T7 RNA polymerases are quite interesting.
All three polymerases can utilize pyrimidine-containing homopolyrners, either singlestranded or as part of duplex pairs, as templates, but arc far less efficient with the purine-containing homopolymers (13,46). The ability of the pyrimidine-containing polymers to serve as efficient templates may result from the preferential initiation by these enzymes with purine ribonucleoside triphosphates (47). T7 polymerase can utilize T3 DNA approximately 50 y0 as efficiently as T7 DNA, whereas T3 polymerase is approximately 10% as active on T7 DNA as its homologous T3 DNA (13, 14,46). The gh-1 polymerase, however, will not utilize either T3 or T7 DNA as templates to any detectable degree. Thus, the exact nucleotide sequences of DNA necessary for either binding or initiation of RNA synthesis must be different between the coliphage-induced and the gh-l-induced RNA polymerases. The coliphage-induced RNA polymerases can utilize denatured or single-stranded DNA from many sources as templates for RNA synthesis at rates from 4 to 35% of the rates on native homologous phage DNA (46,48). With the gh-1 polymerase, very little RNA synthesis is detected when any denatured templates are used.
The gh-l-induced RNA polymerase can initiate RNA synthesis on gh-1 DNA with the ribonucleoside triphosphate, GTP. This nucleotide has an apparent K, value approximately twice as high as the other three ribonucleosidc triphosphates.
The Hill coefficient of GTP is 1.2, as opposed to 1.0 for ATP, CTP, and UTP.
The higher apparent K, value for GTP and its curvilinear double reciprocal plot may result from the role of GTI' in the initiation process.
The process of RNA synthesis by bacterial and phage-induced RNA polymerases has been postulated to involve two binding sites for ribonucleoside triphosphates: an initiation site, which binds the 5'-terminal ribonucleoside triphosphate during the initiation process and the 3'.terminal nucleotide of the growing RNA chain during elongation, and an elongation site, which binds the ribonucleoside triphosphate which is to be incorporated into the 3' terminus of the growing RNA chain (49, 50). These two sites may have very different K, values. The apparent K, value of any ribonucleoside triphosphatc involved only in the elongation process will be the K, value of the elongation binding site. For gh-1 polymerase, this value is apparently 35 to 40 PX for the ribonucleoside triphosphates.
The over-all apparent K, value for any ribonucleoside triphosphate involved in both initiation and elongation will depend on the relative K, values for the two individual binding processes. If the relative K, value of one of the two binding sites is substantially higher than that of the other site, the over-all apparent K, value for that ribonucleoside triphosphate will reflect primarily the higher K, binding site. This is evidently the case for E. coli and T3 RNA polymerases, for which the apparent K, values of the ribonucleoside triphosphates involved in initiation are 10 times and 5 times higher, respcctively, than the apparent K, values of nucleoside triphosphates involved only in elongation (49, 50). If t,he K, values for the initiation and elongation binding processes are similar for any ribonucleoside triphosphate, the over-all apparent K,n value should contain contributions from both binding processes. This may be the case with gh-1 polymerase for GTP, where the apparent K, value of the initiating ribonucleoside triphosphate GTP is only twice as high as that seen for the nucleoside triphosphates involved only in elongation.
'lherefore, varying the concentration of GTP may aiYect both the initiation and the elongation binding processes simultaneously, causing a curvilinear double reciprocal plot with respect to GTP.
A second explanation for the curvilinear double reciprocal plot with respect to GTP for gh-1 polymerase is based on the requirernent for 2 substrate molecules for the formation of the first phosphodiester bond, whereas subsequent polymerization only involves the addition of single nucleotidex.
If both of t,he first two ribonucleoside triphosphates used in t,he initiation process are the same nucleotide, then the double reciprocal plot with respect to that nucleosidc triphosphate should be second order (50). For the gh-1 polymerase, it is conceivable that a portion of the RNA chains synthesized using gh-1 DNA are initiated with the dinucleotide pppGpG-, giving rise to a Hill coefficient between 1 .O and 2.0 with respect to GTP.
Although it is not possible to distinguish between the two explanations presented for the curvilinear double reciprocal plots of GTP, it should be noted that both ex-planations involve the role of GTP as the initiating nucleoside triphosphate in RNA synthesis.
The inhibitor 3'-dATP was shown to compete with ATP for a common binding site on the gh-1 polymerase and host RNA polymerase molecules.
The exact mechanism by which 3'.dATP inhibited in vitro RNA synthesis of the host and gh-1 RNA polymerases is unknown.
The inhibition of in vitro RNA synthesis by 3'.dATP could be due to a simple competition with ATP for a single binding site. On the other hand, the inhibition could be due to the enzymatic incorporation of 3'.dATP into the growing RNA chain, thus causing chain termination. Once RNA synthesis has terminated, the RNA polymerase molecule would have to be released from the enzyme-DNA-nascent RNA complex and then bind to a proper initiation sequence in the DNA before it could once again participate in normal RNA synthesis. It is also possible that RNA polymerase molecules terminated by incorporation of 3'.dATP could be released less rapidly than RNA polymerase molecules terminated at natural termination sites. The determination of whether 3'-dATP is incorporated into RNA would be greatly facilitated by the use of radioactively labeled 3'.dATP.
Experiments on the size of RNA transcribed in vitro by the gh-1 polymerase after incubation periods long enough to ensure several rounds of transcription revealed that the RNA synthesized in the presence of 3'.dATP was significantly shorter than that synthesized in its absence (data not shown). These experiments indicated that the 3'-dAT1' can cause premature termination of RNA synthesis by the phage enzyme. Although these experiments do not directly demonstrate that 3'-dATP is incorporated into RNA by the gh-1 polymerase, they are presumptive evidence of this point.
The apparent Ki value for 3'-dATP for the gh-1 polymerase (2 X lOWa M) is strikingly low compared to that for the host P. putida RNA polymerase (2 X 1OV M) or for the eukaryot,ic RNA polymerase I and II isolated from Novikoff hepatoma tissue culture cells (1.4 X 10W5 M and 7 X 1Om6 M, respectively). 2 This higher sensitivity of the gh-1 polymerase to 3'-dATP could indicate that it is not as competent at discriminating between the substrate analog (3'.dATP) and the natural substrate (ATP) for binding to the active site as the other RNA polymerases.
3'. dATP provides a tool for selectively inhibiting gh-1 polymerase activity in the presence of host RNA polymerase activity in in vitro RNA synthesis.
For T7 infection of E. coli, development of the bacteriophage requires both the host and phage-induced RNA polymerases (11,12). The host polymerase transcribes approximately 20% of the length of the T7 DNA, giving rise to the early RNA species (51). One of the products of this transcription is the mRNA for the 1'7 polymerase, which is then responsible for the transcription of the late region of coliphage T7 DNA (12). In gh-1 infection of P. putida, two temporally appearing classes of RNA have been identified.a The gh-1 polymerase transcribes from only one strand of gh-1 DNA, the biologically correct strand, and synthesizes both early and late gh-1 RNA.3 It is likely, therefore, that the gh-1 polymerase acts similarly to the T7 and T3 RNA polymerase to provide a positive control in turning on transcription of viral genes.