Proc. Natl. Acad. Sci. USA Vol. 77, No. 11, pp. 6381-6385, November 1980 Biochemistry

In vitro comparison of initiation properties of X wild-type PR and x3 mutant promoters (RNA polymerase mechanism/abortive initiation) DIANE K. HAWLEY AND WILLIAM R. MCCLURE Department of Biochemistry and Molecular Biology, Conant Laboratory, Harvard University, Cambridge, Massachusetts 02138 Communicated by Mark Ptashne, July 28,1980

ABSTRACT The in vitro initiation properties of the PR Our analysis is based on a simple two-step model of pro- of bacteriophage X and of a PR mutant, x3, were moter-polymerase interaction first proposed by Zillig and his compared. Using the abortive initiation reaction, we measured coworkers (4). According to this model, RNA polymerase first the lags in the approach to a final steady-state rate when dinu- binds to the DNA in a transcriptionally inactive "closed" cleotide synthesis was initiated with RNA polymerase. These unwinds the DNA to form the lags corresponded to the average times required for the forma- complex (RPc) and subsequently tion of transcriptionally active open complexes. By measuring "open" complex (RPO) which then binds the nucleoside tri- the lags at different RNA polymerase concentrations, we could phosphates and initiates (5). Formation of the open separate open complex formation into two steps, based on a complex can be described as follows: simple model in which the initial bimolecular association of free ks k2 promoter and polymerase in a closed complex is followed by an R + Pa SRPco- RPo. [1] isomerization to the open complex. The contribution of each k-1 k-2 step to the overall rate of open complex formation was quanti- tated for both promoters. We found that the x3 mutation, which In this study, we used the abortive initiation assay to monitor is located in the -35 region of PR, resulted in a decrease in the the lags in the formation of open complexes on the two pro- association constant for the initial binding to the closed complex moters when the reaction was initiated with RNA polymerase. to 5% of its wild-type value and a decrease in the rate of the The average time necessary for open complex formation, 'robs, isomerization to 20%. The lifetimes and abortive initiation was measured at different RNA polymerase concentrations. characteristics of the mutant and wild-type promoters were similar. We concluded that the main effect of the x3 mutation These data were analyzed in a way that allowed the equilibrium was to increase the average time of open complex formation and constant for the initial binding (KI = k1/k_1) to be quantitated that the functional properties of the open complexes did not separately from the rate of open complex formation (k2). This differ significantly between the two promoters. latter process is referred to as an isomerization because we do not know whether we have measured the rate of DNA melting Promoters are regions of DNA to which RNA polymerase binds or another rate-determining conformational change. The results to begin RNA synthesis. The structural features of promoters of our study indicate that the main effect of the x3 mutation and the details of the polymerase-DNA interaction are subjects in vitro is a decrease in K, to 5% of the wild-type value and a of great interest because the initiation of transcription is an decrease in k2 to 20%. important control point in the regulation of . Promoters recognized by RNA polymerase MATERIALS AND METHODS contain two regions of sequence homology: (i) the , a six-base-pair sequence centered at -10 with respect to the DNA Templates and Enzyme. The source of X PR was the startpoint of transcription, and (ii) another six-base-pair se- 890-base-pair DNA fragment isolated after Hae III digestion quence located around -35. RNA polymerase interacts with of purified bacteriophage X DNA. The X x3 promoter was bases in the Pribnow box and -35 region, as determined by obtained by Hae III digestion of a pBR313-derived plasmid, chemical probe experiments. The importance of these se- pRB70, constructed by Roger Brent (Harvard University). This quences to in ivo polymerase recognition is demonstrated by plasmid contained an insertion of XcI857x3 DNA spanning the the occurrence of most promoter mutations within these two region from the EcoRI site in the 0 gene to the Bam I site to regions (1, 2). the left of the N gene. The x3 mutation abolishes HindII cutting The frequency of transcription initiation at some promoters within the promoter (3); this characteristic was confirmed for is regulated by accessory , such as and acti- the A Hae III 890 fragment obtained from this plasmid (un- vators. Other promoters appear not to require additional published data). DNA fragments resulting from Hae III di- factors; indeed, mutations that affect transcription from a gestion of phage or plasmid DNA were separated on a 3.5% promoter both in vio and in vitro demonstrate that the DNA polyacrylamide gel and extracted from the gel with phenol (6). sequence alone can directly determine the frequency of initi- The DNA was separated from UV-absorbing impurities by ation on a promoter. One such mutant, x3, is a base-pair change DEAE-Sephadex (A-25) chromatography as described (7) or in the -35 region of PR, the major rightward promoter of by spermine precipitation (unpublished data). DNA concen- bacteriophage X (3). We studied the initiation properties of x3 trations were calculated on the basis of an absorption coefficient and PR to determine what steps in the RNA polymerase- pro- at 260 nm of 6.5 mM-1 cm-1 DNA phosphorus and are ex- moter interaction have been affected by this base pair pressed as molar fragment. E. coli RNA polymerase was isolated by the procedure of Burgess and Jendrisak (8). Holoenzyme was change. separated from core according to Lowe et al. (9). Enzyme = and a molecular The publication costs of this article were defrayed in part by page concentrations are based on E" nm 6.2 charge payment. This article must therefore be hereby marked "ad- weight of 4.9 X 105 (10). The enzyme was 50-60% active in vertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact. Abbreviation: polyfd(A-T)l, poly[d(A-T)]-poly[d(A-T)]. 6381 Downloaded by guest on September 29, 2021 60382 Biochemistry: Hawley and McClure Proc. Natl. Acad. Sci. USA 77 (1980) promoter binding and initiation as judged by titrations of promoter-containing DNA fragments according to Cech and McClure (11). Determination of Tobs. Details of the abortive initiation reaction have been described (7, 12). For each determination of Tobr, a control -initiated reaction was run in parallel with a reaction initiated by addition of RNA polymerase. Both 10 reaction mixtures had a final volume of 250 Ml. For the nucle- x otide-initiated reaction, the DNA in 150 ul of reaction buffer was preincubated at 370C for 3 min before addition of 50 pl of RNA polymerase in reaction buffer. The DNA and RNA 5- polymerase were preincubated together for 14 min when the template was PR or 33.5 min when the template was x3. After this preincubation, AMP (or CpA) and UTP in 50 AI of reaction buffer were added to initiate the reaction. For the polymer- 0 10 20 30 40 ase-initiated reaction, the AMP and UTP were preincubated Time, min with the DNA in 200 Atl of reaction buffer for 5 min before addition of 50 of RNA polymerase at zero time. Aliquots of FIG. 1. Lag in the approach to a final steady-state rate ofabortive Al initiation from the x3 promoter. The radioactivity incorporated into each reaction mixture were spotted onto Whatman 3MM paper pApU is plotted versus reaction time. The standard assay conditions and developed with ascending chromatography in H20/satu- were: 40 mM Tris-HCl, pH 8; 10 mM MgCl2; 120 mM KCl; 1 mM di- rated ammonium sulfate/isopropanol, 18:80:2 (vol/vol), as thiothreitol; 100 jig of bovine serum albumin per ml; 2 mM AMP; 0.05 described (7). mM UTP ([a-32PJUTP was added to a specific activity of 500-600 The rate of the control nucleotide-initiated reaction was cpm/pmol); and 1 nM x3-containing X Hae 890 fragment in 250 ,u determined by a linear least squares analysis- rT was calculated total volume. The RNA polymerase concentration was 133 nM. The steady-state rate of the control nucleotide-initiated reaction (0) was from the polymerase-initiated reaction by extrapolating a least 22 pApU promoter'1 min-'. The curve corresponding to the poly- squares line calculated for data points corresponding to intervals merase-initiated reaction (o) was calculated from a rob. of 10.3 min greater than 4 times the estimated robs to the time axis. These and a final steady state rate of 20 pApU promoter-1 min-1. calculated values for robs and the final steady-state rates were then used as initial estimates of the values determining a single average time required for the free enzyme and promoter to exponential curve. A final value for Tob was obtained from the combine and to isomerize into the open complex. curve best fitting the experimental data. The nucleotide components of the abortive initiation reaction Determination of Open Complex Lifetimes. Because the were varied to determine whether the observed lags were af- lifetimes of the RNA polymerase-promoter complexes were fected by some characteristic of the steady-state reaction. The very long, we chose to measure the initial rate of decay of the robs values determined by monitoring pApU, pppApU, or complex by heparin and poly[d(A-T)]-poly[d(A-T)] (abbre- CpApU synthesis were the same within experimental error, viated herein as poly[d(A-T)]) challenge techniques and to although the steady-state rate of UTP incorporation into the calculate koff from a semilogarithmic plot of fraction of open three products was different in each case. Moreover, when the complex remaining versus time, as described by Cech and AMP concentration was increased 2-fold, the observed lag was McClure (11). When poly[d(A-T)] was used as the challenging unaffected; the final steady-state rate of product synthesis in- reagent, we monitored production of CpApU and limited the creased as expected. The final steady-state rates did not depend concentration of free polymerase by preincubating 1.3 nM PR upon RNA polymerase concentration when the enzyme was promoter and 5 nM RNA polymerase and then diluting with present in 10-fold molar excess over promoter. However, the reaction buffer, with or without poly[d(A-T)], so that the final robs' values determined for the x3 promoter were highly de- PR concentration was 0.9 nM. For x3, 20 nM promoter was pendent upon polymerase concentration (Fig. 2). This obser- preincubated with 55 nM RNA polymerase and was then di- vation and the lack of dependence of the steady-state rates on luted to a final promoter concentration of 1.5 nM. The controls, details of the abortive initiation reaction suggest that the times in which poly[d(A-T)] was added to RNA polymerase alone, being measured are those required to form a stable open com- showed that the synthesis of CpApU from poly[d(A-T)] was plex. negligible under these conditions. The striking dependence of the x3 lag on RNA polymerase concentration is the major observed difference between the RESULTS mutant and wild-type promoters. In Fig. 2 (tau plot), each Preincubation of the PR promoter with RNA polymerase for point corresponds to a separate determination of Tobs similar 10 min before addition of AMP and UTP resulted in steady- to that shown in Fig. 1. Compared to the other steps in RNA state production of pApU which extrapolated to zero product chain initiation and elongation (13, 14) the formation of the at the time when the were added. When dinu- open complex requires a rather long time for the wild-type cleotide synthesis was initiated by the addition of RNA poly- promoter and much longer times for the mutant. Therefore, merase to a solution containing DNA and nucleotides, a lag in these times correspond to the principal rate-limiting steps in the approach to the final steady-state rate was observed. When initiation from these promoters. Extrapolation of the tau plots the same experiment was performed with the x3 mutant pro- to saturating enzyme concentration shows that the rate of open moter, the lag was much longer than with PR under identical complex formation approaches a limiting value. The limiting conditions. This observation was exploited to study and compare value obtained from the ordinate corresponds to the rate of the initiation properties of the wild-type and mutant pro- conversion to RPc to RPO. moters. The slope of the tau plot for x3 is 100 times greater than that A typical lag experiment for the x3 promoter is shown in Fig. for PR. In order to interpret this observation quantitatively, we 1. Extrapolation of the final steady state rate of a lag experiment make the assumptions that the free polymerase and promoter to the time axis, as shown in the figure, gives obs, which is the are in rapid equilibrium with the closed complex and that the Downloaded by guest on September 29, 2021 Biochemistry: Hawley and McClure Proc. Natl. Acad. Sci. USA 77 (1980) 6383 difference in KI values between the mutant and wild-type 750 promoters is not affected by this consideration because the same enzyme preparation was used throughout. 600 The overall binding constant, KO, can be calculated with the information obtained from the tau plot if the lifetime of the a)c; 450 open complex is known. This was determined by the I- poly[d(A-T)] and heparin challenge techniques. Fig. 3 is a .0 0 ,~~~~~~ 0 semilogarithmic plot of fraction RPO remaining versus time for 300[ the PR promoter. The lifetime of the PR-RNA polymerase

- complex was not affected by the concentrations of poly[d(A-T)] 150 F ~~~0 0 I- 0 used. The lifetime of the x3-polymerase complex also was in- El dependent of poly[d(A-T)] concentration (not shown). The koff values obtained for x3 and PR by poly[d(A-T)] challenge are 0 50 100 150 200 listed in Table 1. 1/RNA polymerase, iM-' The lifetimes determined for both promoters by heparin FIG. 2. Tau plots for the X PR and x3 promoters. T0bI, the time challenge were similar to those determined by poly[d(A-T)] required for open complex formation, is plotted versus the reciprocal of RNA polymerase concentration for PR (o) and x3 (0). The stan- challenge; these values are included in the legend to Table 1. dard assay conditions described in Fig. 1 were used. The Trb values The use of both poly[d(A-T)] and heparin in determining the were determined at the RNA polymerase concentrations indicated lifetimes was necessitated by our reluctance to rely solely on by running a control nucleotide-initiated reaction in parallel with a heparin challenge data because there is evidencethat, for some reaction initiated with enzyme and evaluating the observed lag promoters, heparin directly attacks the open or closed com- times. plexes or both (11, 16). We conclude that the x3 and PR open complex dissociation rates were unaffected by heparin at 50 lifetime of the open complex is very long (k-2 << k2). Fur- ,gg/ml. thermore, experimental conditions were chosen so that the Based on the model used to arrive at the equations in this concentration of RNA polymerase was at least in 10-fold excess analysis, k-2 is the rate-limiting step in the decay of the open over promoter concentration. Thus, the free polymerase con- complex. Thus, a measure of koff yields directly a value for k-2. centration remained essentially constant throughout each ex- The apparent equilibrium constant KO can then be calculated periment. Based on these assumptions, the rate equation for as K1-(k2/k-2). These values are also shown in the table. The open complex formation has been solved (15). The rate constant overall equilibrium constant is 100 times higher for PR than for for the reaction (kob) is related to -rob and the RNA polymerase x3. concentration, [R], as follows: The similar lifetimes of the open complexes formed by x3 and PR suggest that the properties of these promoters might be 1 1 + [2] similar or identical after the open complex has been achieved. Tos=kob5 k2 Klk2[R] The similarity of steady-state rates of dinucleotide synthesis A plot of robs versus [R]-' yields values for k2 (reciprocal of the catalyzed by RNA polymerase on both promoters further in- intercept) and for the apparent association constant KI (inter- dicates that any differences in the open complexes are probably cept divided by slope). In this way, the tau plot slope and in- not highly significant. However, because the steady-state rate tercept allows a separate quantitation of the binding and isom- is dependent upon the concentration of abortive initiation erization steps in Eq. 1. substrates and upon the promoter concentration, it is not pos- The kinetic information derived from the tau plot param- sible to conclude from this evidence alone that the details of the eters is summarized in Table 1. The major effect of the x3 abortive initiation reaction are the same for the two promoters. mutation was a 95% decrease in the apparent equilibrium To test this possibility directly, we measured the apparent constant for the initial binding of the free polymerase to the promoter to form the closed complex. In addition, there was an w 80% decrease in the rate of isomerization by the mutant. The I= value for is dependent on the fraction of active, free enzyme. 1. KI 0) The values in Table 1, based on total added RNA polymerase, .0~~~~~o a are lower estimates of this parameter. However, the 20-fold .91- 0 la .0E O.0..8 o Table 1. In vitro comparison of A PR and x3 I-6) PR x3 Qz Ii k,, 6.7 X 106 M-1 sec-1 6.2 X 104 M-1 sec1 0 50 100 150 k2 1 X 10-2 sec-1 2 X 10 sec-1 Time, min koff (k-2) 3.7 X 10-5 sec-' 4.0 X 10-5 sec-1 FIG. 3. Decay of RNA polymerase-PR open complexes. The KI 6.7 x 108 M-1 3.1 x 107 M-1 fraction of RPO remaining is plotted semilogarithmically versus the KII 270 50 time following the addition of poly[d(A-T)]. The standard assay K0 1.8 X 1011 M-1 1.6 x 109 M-l conditions ofFig. 1 were used, except that the concentrations of UTP and CpA were 25 and 250 ,M, respectively. The template was 0.9 nM Values for the kinetic constants k2 and kon were obtained from the PR DNA; RNA polymerase concentration was 3.5 nM. The control intercepts (intercept = 1/k2) and the slopes (1/S = kik2/k-1 = kon) reaction (*), to which poly[d(A-T)] was not added, catalyzed CpApU of Fig. 2. koff was determined by the poly[d(A-T)J challenge tech- synthesis at a constant rate of 32 CpApU promoter-' min-'. The open niques described in Materials and Methods and in Fig. 3. Values for symbols correspond to reactions to which poly[d(A-T)] at the fol- koff determined by heparin challenge (50 ug/ml) are 4.3 X 10-5 sec1 lowing concentrations was added: 45 ,uM (ti), 90,gM (o), and 180,M for PR and 3.9 X 10-5 sec-1 for x3. The equilibrium constants were (0). The line is a least squares fit to all the data points. The initial rate calculated from the ratios KI = kon/k2, KII = k2/koff, Ko = kon/koff corresponds to a turnover of 25 CpApU promoter-' min-'. Downloaded by guest on September 29, 2021 6384 Biochemistry: Hawley and McClure Proc. Natl. Acad. Sci. USA 77 (1980) Michaelis constants for three of the abortive initiation substrates on PR and x3. For two of the three substrates tested, the Km was lower for x3 than for PR (Table 2). The max l velocities were also somewhat lower for x3 than for PR. When CpA was the varied substrate, the maximal velocity was 240 mind1 for x3 and 530 mind for PR. A systematic study of these steady-state ki- netic properties was not pursued further because the differences were not so marked as to suggest major mechanistic differences 03- between the open complexes for the two promoters. X~~~~~~~ The interpretation of the kinetic and equilibrium constants shown in Table 1 has assumed that the closed complex is in rapid 02- equilibrium with the free polymerase and promoter. An al- ternative model is that overall open complex formation is the end result of two essentially irreversible reactions. Solution of the rate equation for this case results in a relationship between 1 ,rob and [RI-' expressed by the equation 0~~~~~~~~ 11I Toib = + [3] 10 20 30 Time, min Here as in Eq. 2 a linear relationship between rT, and [RI-' is expected. However, the interpretation of the tau plot param- FIG. 4. The effect ofheparin addition at various times after ini- tiation of the abortive initiation reaction by addition of RNA poly- eters differs for two models. In both cases the intercept merase. The cpm incorporated into pApU are plotted versus time. The corresponds to 1/k2. In the irreversible model, the slope is re- assay conditions ofFig. 1 were used, except that the RNA polymerase lated to the association rate constant for closed complex for- concentration was 100 nM. The template was 1 nM x3 DNA. The mation. RNA polymerase-initiated reaction (0) consisted of 500 Al of DNA A priori, either model is equally plausible. The key difference and nucleotides prewarmed to 370C in reaction buffer to which 100 is whether >> k2 (rapid equilibrium) or << k2 (irre- Al ofRNA polymerase in reaction buffer was added at zero time. After k-i k-1 0.5 min (&), 2.5 min (0), 4 min (*), and 7 min (v), aliquots of this versible). The irreversible model predicts a two-exponential reaction were added to prewarmed heparin solutions so that the final approach to the steady state that might be apparent at low en- heparin concentration was 48 ,ug/ml. The control reaction (-) was zyme concentrations. However, we sought a more direct test initiated by adding 50 Al of nucleotides in reaction buffer after a of this possibility. 33-min preincubation of DNA and RNA polymerase in 250 Al of re- To examine whether the rapid equilibrium or irreversible action buffer at 370C. An aliquot ofthis reaction mixture was added model is more nearly correct, we initiated pApU synthesis from to heparin after 1 min (03). The steady-state rates ofthe control and RNA polymerase-initiated reactions were 27 and 24 pApU promoter' x3 with RNA polymerase and interrupted the reaction at dif- min1, respectively. A reaction in which polymerase was added to a ferent times during the lag by the addition of heparin. If the mixture of promoter and heparin at 48 jg/ml had a rate of0.1 pApU equilibrium model held, then quenching of free polymerase promoter' min-' (not shown). The steady-state rates of the hepa- by heparin would also quench the closed complexes as a result rin-interrupted reactions were determined by linear least squares of the rapid equilibrium between R and RP,. Thus, a linear rate analysis. of dinucleotide production would be observed, proportional to the fraction of open complexes formed by the time of heparin tent with a rapid equilibrium between enzyme and closed addition. If, instead, the irreversible model obtained, then the complex than with a sequence of two irreversible steps. closed complexes would isomerize to open complexes with a lag equal to 1/k2. DISCUSSION Fig. 4 shows the result of such an experiment. The incorpo- The major in vitro effect of the x3 mutation in the X PR pro- ration of UTP into dinucleotide immediately after addition of moter was to decrease the initial binding constant of enzyme heparin was linear, intersecting the curve at the time of heparin and promoter by 95%. The rate of formation of the open addition. The decrease in rate at longer times probably repre- complex from the closed complex was also decreased 80% by sents the slow dissociation of open complexes, because heparin the mutation. The x3 promoter lacks the wild-type HindII added to the nucleotide-initiated reaction caused the rate of that restriction endonuclease site 33 base pairs upstream from the reaction to decrease at longer times as well. The abrupt inter- startpoint of transcription. This study therefore provides ruption of open complex formation by heparin is more consis- quantitative evidence that the -35 region of this promoter is directly involved in the formation of the closed complex. The possibility that the -35 region is involved in polymerase Table 2. Comparison of apparent Kms for abortive initiation recognition of a promoter was suggested by the observation that substrates on X PR and Xx3 promoter-containing DNA fragments that lacked the -35 re- KM, mM gion as a result of DNase digestion were incapable of rebinding Nucleotide PR x3 the polymerase (18-20). The sequence homology seen in the -35 region when many promoters were compared suggested AMP 4.2 0.4 2.8 0.9 that there is a sequence-dependent requirement for this region CpA 1.6 0.4 2.9 0.3 of the promoter which would not be fulfilled by DNA of ran- UTP 0.26 0.05 0.14k 0.01 dom sequence. The ability of the polymerase to remain stably Standard assay conditions of Fig. 1 were used. When the UTP complexed with DNase-digested promoters and even to initiate concentration was varied, the CpA concentration was held at 0.§ mM; transcription on these fragments led to the proposal that the -35 the UTP concentration was 50 AM for the rest ofthe reactions. Kinetic for initial to constants with the associated uncertainties were calculated according region was required mainly binding prior open to Wilkinson (17) from double reciprocal plots of the initial velocity complex formation (21, 22). data. In light of these observations, the effect of a -35 region Downloaded by guest on September 29, 2021 Biochemistry: Hawley and McClure Proc. Natl. Acad. Sci. USA 77 (1980) 6385 mutation on the rate of isomerization from a closed to open contacts needed to form the closed complex are important for complex might seem rather surprising. The region is well sep- maintenance of the open complex. arated from the Pribnow box and the startpoint of transcription, The rapid equilibrium assumption in the initial binding step where base pairs have been shown to be unwound in the open is important in relating the slopes and intercepts of the tau plots complex (23). However, the unwinding of the DNA may not to KI. The observation that heparin addition prevented further be the rate-determining step in the process that we have termed open complex formation during the lag is most simply inter- isomerization for either of these promoters. preted as deriving from a rapid equilibrium in the binding step. -The kinetic scheme for open complex formation that has This finding would only be consistent with an irreversible for- served as a basis for the interpretation of our observations is the mation of the closed complex if heparin rapidly and quantita- minimal, simplified view. We cannot rule out more compli- tively inactivated this intermediate. However, the observation cated mechanisms involving more than one intermediate either that the open complex lifetimes were similar when determined on the direct pathway or on nonproductive side pathways. In- with heparin and poly[d(A-T)] at several concentrations strongly deed, the effect of the x3 mutation on k2 makes a more com- argues that both of these reagents bind only to the free enzyme. plicated mechanism plausible. For example, it is possible that, It is possible that different reaction conditions could change the before the isomerization step, the RNA polymerase must un- rapid equilibrium mechanism proposed here to one in which dergo a conformational change which is dependent upon the two essentially irreversible steps obtain. Although the experi- -35 region sequence. Then the base pair substitution in x3 ment of Fig. 4 has been performed at 80 mM KCI with com- could be imagined to decrease by a factor of 5 the equilibrium parable results, additional tests on these and other promoters constant for this putative conformational change. Alternatively, will be required to test the generality of this simple mecha- the isomerization of the x3-polymerase complex might involve nism. intermediates not found on the pathway leading to the PR open complex. The complexities considered above are mechanisti- The work was supported by Grant 21052 from the National Institutes cally important but do not affect the basic functional separation of Health, and by a predoctoral fellowship to D.K.H. from the National of the binding and isomerization steps of open complex for- Science Foundation. mation. We have recently found that the major effect of the UV5 1. Rosenberg, M. & Court, D. (1979) Annu. Rev. Gen. 13, 319- mutation in the lac promoter was to increase the isomerization 353. rate about 10-fold compared to the wild-type promoter. In 2. Siebenlist, U., Simpson, R. B. & Gilbert, W. (1980) Cell 20, contrast, KI was nearly the same for both promoters. The UV5 269-281. mutation is a in 3. Maurer, R., Maniatis, T. & Ptashne, M. (1974) Nature (London) 2-base-pair change the Pribnow box. Thus, the 249,221-223. results from the X and lac promoters are consistent with a simple 4. Walter, G., Zillig, W., Palm, P. & Fuchs, E. (1967) Eur. J. Bio- model for open complex formation based on promoter struc- chem. 3, 194-201. ture. If the base pairs in the -35 region are primarily respon- 5. Chamberlin, M. J. (1974) Annu. Rev. Biochem. 43,721-775. sible for the extent of the initial binding reaction and if the in- 6. Rubin, G. M. (1973) J. Biol. Chem. 248,3860-3875. teraction of bound enzyme with DNA in the Pribnow box de- 7. McClure, W. R., Cech, C. L. & Johnston, D. E. (1978) J. Biol. termines the rate of the isomerization process, then we expect Chem. 253, 8941-8948. that quantitative rules for the RNA polymerase-promoter in- 8. Burgess, R. R. & Jendrisak, J. J. (1975) Biochemistry 14, 4634- teraction will follow from studies on additional promoters and 4638. promoter mutants. In fact, the tau plot analysis has now been 9. Lowe, P. A., Hager, D. A. & Burgess, R. R. (1979) Biochemistry to 10 in addition to those considered here. 18, 1344-1352. applied promoters 10. Burgess, R. R. (1976) in RNA Polymerase, eds. Losick, R. & Although the results are largely consistent with the simple Chamberlin, M. (Cold Spring Harbor Laboratory, Cold Spring model, exceptions such as the x3 mutation show that small ef- Harbor, NY), pp. 69-100. fects on the isomerization rate can result from base pair dif- 11. Cech, C. L. & McClure, W. R. (1980) Biochemistry 19,2440- ferences in the -35 region. 2447. The x3 mutation lies in the HindII recognition site (G-T- 12. Johnston, D. E. & McClure, W. R. (1976) in RNA Polymerase, Py-Pu-A-C) which contains the canonical sense-strand sequence eds. Losick, R. & Chamberlin, M. (Cold Spring Harbor Labora- T-T-G-A-C. Recently, Siebenlist et al. (2) have shown that bases tory, Cold Spring Harbor, NY), pp. 413-428. within this sequence were protected from dimethyl sulfate 13. Krakow, J. S., Rhodes, G. & Jovin, T. M. (1976) in RNA Poly- methylation when RNA polymerase was bound in an open merase, eds. Losick, R. & Chamberlin, M. (Cold Spring Harbor on T7 Laboratory, Cold Spring Harbor, NY), pp. 127-157. complex the A3 and lac promoters. Such a finding 14. Nierman, W. C. & Chamberlin, M. J. (1979)J. Biol. Chem. 254, suggests that retention of -35 region contacts may be important 7921-7926. for the stability or function of the open complex. However, the 15. McClure, W. R. (1980) Proc. Nati. Acad. Sci. USA 77, 5634- similar properties of the x3 and PR open complexes suggest that 5638. at least some of the -35 region contacts involved in closed 16. Pfeffer, S. R., Stahl, S. J. & Chamberlin, M. J. (1977) J. Biol. complex formation are not directly involved in maintaining the Chem. 252,5403-5407. open complex. Because the lifetimes of x3 and PR complexes 17. Wilkinson, G. N. (1961) Biochem. J. 80, 324-332. are the same, the -35 region mutation apparently has no effect 18. Schaller, H., Gray, C. & Herrman, K. (1975) Proc. Natl. Acad. on the stability of the open complex. In addition, there are no Sci. USA 72,737-741. differences in the of RNA to catalyze 19. Pribnow, D. (1975) Proc. Natl. Acad. Sci. USA 72,784-788. major ability polymerase 20. Walz, A. & Pirrotta, V. (1975) Nature (London) 254, 118- the abortive initiation reaction on both promoters, indicating 121. that the functional properties of the complexes are not drasti- 21. Pribnow, D. (1975) J. Mol. Biol. 99, 419-443. cally different. It is likely that the properties of the open com- 22. Heyden, B., Nuisslein, C. & Schaller, H. (1975) Eur. J. Biochem. plex are determined by polymerase contacts in the Pribnow box 55, 147-155. region. An additional possibility is that only a subset of the -35 23. Siebenlist, U. (1979) Nature (London) 279,651-652. Downloaded by guest on September 29, 2021