Upstream Interactions at the Lambda pRM Promoter Are Sequence ...

JOURNAL OF BACTERIOLOGY, Dec. 1996, p. 6945–6951 0021-9193/96/$04.0010 Copyright q 1996, American Society for Microbiology

Vol. 178, No. 23

Upstream Interactions at the Lambda pRM Promoter Are Sequence Nonspecific and Activate the Promoter to a Lesser Extent than an Introduced UP Element of an rRNA Promoter YANG TANG,1 KATSUHIKO MURAKAMI,2 AKIRA ISHIHAMA,2

AND

PIETER L.

DEHASETH

1

*

1

Department of Biochemistry, Case Western Reserve University, Cleveland, Ohio 44106-4935, and Department of Molecular Genetics, National Institute of Genetics, Mishima, Shizuoka 411, Japan2 Received 30 May 1996/Accepted 25 September 1996

The rightward regulatory region of bacteriophage lambda contains two promoters, pRM and pR, which direct the synthesis of nonoverlapping divergent transcripts from start sites 82 bp apart. Each of the two promoters has an upstream (A1T)-rich region (ATR) within the sequence from 240 to 260 where in the rrnB P1 promoter a stretch of 20 (A1T) bp greatly stimulates promoter function. Here we present an investigation of the possible functional significance of pRM’s ATR. We determined the effects on RNA polymerase-pRM promoter interaction both of (G1C) substitutions in the ATR and of amino acid substitutions in the a subunit, known to affect the upstream interaction. We find small (two- to threefold) effects of selected mutations in the a subunit on open complex formation at pRM. However, the (presumably upstream) interactions underlying these effects are sequence nonspecific, as they are not affected by (G1C) substitutions in the ATR. Substitution of the 20-bp UP element of the rrnB P1 promoter between positions 240 and 260 at pRM stimulates open complex formation to a considerably greater extent (5- to 10-fold). Results from kinetic studies indicate that on this construct the UP element mainly accelerates a step subsequent to the binding of RNA polymerase, although it may also facilitate the binding event itself. Less extensive studies likewise provide evidence for a two- to threefold activation of pR by upstream interactions. The possible involvement of the a subunit in the previously characterized (e.g., B. C. Mita, Y. Tang, and P. L. deHaseth, J. Biol. Chem. 270:30428–30433, 1995) interference of pR-bound RNA polymerase with open complex formation at pRM is discussed. has been implicated in UP element recognition: deletion of this domain greatly impaired the ability of RNAP to respond to the UP element both in vivo and in vitro (27). In subsequent studies, amino acid substitutions in the a subunit were employed to pinpoint the amino acids involved in the interaction with the UP element to those at positions 262 to 269 and 296 to 299 (8). Recent determinations of the structure of the Cterminal domain of the a subunit show that the location of these amino acids is compatible with their direct interaction with DNA (8, 9). Both pR and pRM have 12-bp upstream ATRs between 240 and 260. In the stretch from 248 to 259, pRM has 10 A-T bp, while pR has 11 A-T bp between 247 and 258. We have probed the functional significance of the upstream (A1T) regions of pR and pRM, with emphasis on the latter. In order to do this without interference from the cross-talk between RNAPs bound to pR and pRM, in many instances we have used templates on which pR was inactivated by mutation. We conclude that upstream interactions do lead to a modest activation of pRM, but that they are independent of the (A1T) content of the region. As a control, we show also that an UP element, derived from that of the rrnB P1 promoter, is able to provide a large (5- to 10-fold) activation of pRM. We have observed that pR is activated to a modest extent by a subunitmediated DNA interactions as well; in contrast to our findings with pRM, here the effect was found to be sequence specific.

The rightward regulatory region of bacteriophage lambda contains two promoters, pR and pRM, which are divergently transcribed from start sites separated by only 82 bp (25). Open complex formation at the pR promoter is much faster than at pRM; therefore, the vast majority of RNA polymerases (RNAPs) forming an open complex at the latter promoter do so on a fragment that already contains a polymerase at pR. It was found by us (14, 15) as well as by Gussin’s group (6, 28) that mutations which inactivate the pR promoter facilitate open complex formation at pRM. These results suggest that pR-bound RNAP interferes with open complex formation at pRM. Surprisingly, a 10-bp deletion between the two promoters appears to abolish the interference (22), as it results in increased activity of pRM whether pR is inactivated by mutation or not. The identity of promoters for Escherichia coli RNAP is mostly determined by two DNA elements, the 210 and 235 regions (21). The ability to recognize these two regions is imparted upon the RNAP holoenzyme by its s subunit (12). The promoters for E. coli rRNA (27), as well as some other strong promoters in E. coli (17–19) and Bacillus subtilis (2, 13), have been found to additionally have (A1T)-rich regions (ATRs) upstream of bp 240. These so-called UP elements (27) are typically 20 bp in length. Transcription in vivo and in vitro (27) as well as the rate of open complex formation at the rrnB P1 promoter (26) was decreased by more than 10-fold because of deletion of this UP element. The C-terminal domain (amino acids 249 to 329 [3, 23]) of the a subunit of RNAP

MATERIALS AND METHODS Materials. E. coli RNAP holoenzyme prepared by the method of Burgess and Jendrisak (4) was further purified as described previously (10). Our preparations were approximately 35% active as determined in promoter binding assays. Mutated a subunits were constructed as previously described (9, 16). RNA core polymerases with the altered a subunits were obtained by reconstitution with purified b and b9 subunits as described previously (9, 16). Subsequently, the mutant core enzymes were incubated on ice for 30 min with purified s subunit

* Corresponding author. Mailing address: Department of Biochemistry, School of Medicine, Case Western Reserve University, 10900 Euclid Ave., Cleveland, OH 44106-4935. Phone: (216) 368-3684. Fax: (216) 368-4544. 6945

6946

TANG ET AL.

J. BACTERIOL.

FIG. 1. The sequences of the variants used in this work. The construct designated Wt has the wt pR and pRM promoters, with an 82-bp separation between the start sites. The 210 and 235 regions of pRM are boxed as shown. The 235 regions of pR are doubly underlined, and the 210 regions are singly underlined. The designation 2 pR refers to DNAs that contain an inactivating C-G substitution at position 27 (the constant T) of pR. The designation pR22 refers to DNAs that contain inactivating substitutions at positions 27, 29, and 211 of pR. The (A1T)-rich region at positions 253 to 259 has been referred to as ATR. ATR2 refers to constructs that contain G-C substitutions for A-T base pairs at positions 254 to 257 of pRM. ATR22 templates have G-C substitutions for A-T base pairs at positions 246, 248, 249, 251, and 253 to 257 of pRM. UP1 refers to the substitution of a sequence derived from the UP element of rrnB P1 in the 240 to 260 region of the pRM promoter. D10 indicates a deletion of 10 bp in the DNA separating the 235 regions of pR and pRM. (0.13 mM core and 0.52 mM s in 30 ml of storage buffer [0.01 M Tris-HCl (pH 7.0), 0.1 M KCl, 0.1 mM EDTA, 0.1 mM dithiothreitol, and 50% glycerol]) to obtain the holoenzymes. The relative activities of the reconstituted holoenzymes were determined by monitoring RNA synthesis from poly(dA-dT) templates as described previously (9). The concentrations of RNAP reported here refer to those of the holoenzyme reconstituted with wild-type (wt) a subunit; all modified RNAPs were added in amounts adjusted to give the same units of activity as the wt enzyme in the poly(dA-dT) assay. Promoter variants were assembled from synthetic deoxyoligomers and then cloned as previously described (1, 22). All experiments were carried out on fragments of 261 bp, obtained by PCR and labeled with 32P at the pRM-proximal 59 end (22). Determination of open complex formation. The reaction mixtures contained radiolabeled DNA (1 nM) in transcription buffer (30 mM Tris-HCl [pH 8.2], 100 mM KCl, 3 mM MgCl2, 0.1 mM EDTA, 10 mM dithiothreitol [reduced to 0.5 mM for permanganate probing experiments], and 45 mg of bovine serum albumin per ml). Binding reactions were initiated by the addition of RNAP to the indicated concentrations and terminated by addition of 0.8 ml of heparin (1.25 mg/ml) to 20-ml samples removed from the reaction mix at indicated times. Analysis of the aliquots for open complex formation by the gel mobility shift or KMnO4 assays was exactly as previously indicated (22). In duplicate determinations, the intensities of the bands resulting from treatment with KMnO4 were found to be within 20% of the average for greater than two-thirds of the values. The values in Fig. 2 to 4 contain a contribution from the background; therefore, fold reductions in activities based on these values are underestimated. Rate constants were obtained by fitting to the single exponential equation y 5 a[1 2 exp(2kobst)] 1 c, where y is the experimentally observed quantity, t is the time after mixing RNAP and promoter DNA, kobs is the pseudo-first-order rate constant, and c is the starting and (a 1 c) is the limiting value of y. The fit was to the optimal values of the latter three parameters. The parameter kobs is the reciprocal of tobs. Data obtained as a function of RNAP concentration are presented as tau plots, where tobs is plotted versus the reciprocal of the RNAP concentration (20). If RNAP, promoter DNA, and their initial (closed) complex are in rapid equilibrium (verified for pRM by Hawley and McClure [11]), and if formation of the final (open complex) is irreversible over the course of the experiment (which we verified for the last three constructs of Fig. 1: after heparin challenge, the fraction of DNA in an open complex remained constant for 60 min [27a]), the slope and intercept of the plot provide additional information about the reaction. The slope then equals (KBkf)21 and the intercept equals 1/kf, where KB is the equilibrium association constant for formation of the initial (closed) complex, and kf is the rate constant for conversion of the closed to the open complex. In Table 1, we also present values of KB and kf obtained by nonlinear least-squares fitting of the data to the rectangular hyperbola kobs 5 (kfKB[RNAP])/(KB[RNAP] 1 1). This method was shown to provide the most reliable estimates of KB and kf (24) in comparison with several others, including the double reciprocal method described above.

RESULTS The 10 promoter constructs used in this work are shown in Fig. 1. For several constructs, the pR promoter has been inac-

tivated to eliminate interference with open complex formation at pRM (6, 14, 15, 22). pR2 bears a substitution in the 210 region which greatly reduces pR’s activity, to 5 to 10% of wt as determined by KMnO4 sensitivity after a 10-min incubation with 100 nM RNAP; pR22 has additional substitutions further reducing the activity to 2 to 5% of wt. For the purposes of our experiments, both pR variants have negligible residual activity. To disrupt sequence-specific interactions with pRM’s ATR, substitutions with 4 (G1C) bp were introduced in this region. During the course of this study, it became evident that these substitutions produced only rather small effects, if any, on open complex formation at pRM. In an attempt to minimize any interaction with (A1T) base pairs over the entire 240 to 260 region, the ATR22 mutant bears additional substitutions, converting all but 3 (A1T) bp in this region to (G1C) base pairs, still with minimal effects on pRM function (see below). The D10 constructs have a 10-bp deletion in the region separating the 235 regions of pR and pRM (22). This allows investigation of the relief of the promoter interference described above as well as an assessment of the effect of positioning the ATR closer to the 235 region of pRM. As a positive control for the experiments on the 12-bp ATRs of the pRM and pR promoters, in the UP1 mutant the 240 to 260 region of pRM has been substituted with a sequence similar to the UP element from rrnB P1 (27). It differs from the latter at positions 254 (T instead of A) and 258 to 260 (ATG instead of GAC), in order to accommodate the ligation scheme used for promoter assembly. Our initial experiments to probe for a possible role for the ATR in determining the activity of the pRM promoter employed polymerases with a deletion of the C-terminal domain of the a subunit. RNAP (20 nM) and radiolabeled DNA (1 nM) were incubated for 60 min, whereupon open complex formation was monitored by determining the degree to which T residues in the vicinity of the start site of pRM became susceptible to oxidation by permanganate. The results are presented in Fig. 2; similar results were obtained for an incubation time of 10 min at both 20 nM and 50 nM RNAP concentrations (data not shown). The results in panel a demonstrate that the polymerase reconstituted with wt a subunit is as active as that isolated from E. coli and confirm that open complex formation at pRM by these two polymerases is stimulated to the same

VOL. 178, 1996

UPSTREAM INTERACTION AT pRM

6947

FIG. 2. The comparison of open complex formation at pR and pRM promoters using RNAP with truncations in the C-terminal domain of the a subunit. The aggregate intensity (expressed as the percentage of the total amount of radioactivity on the gel) of bands resulting from KMnO4 modification at positions between 211 and 11 was used as a measure of complex formation. RNAP (20 nM) and radiolabeled DNA were incubated for 60 min at 378C prior to KMnO4 treatment. RNAP, RNAP purified directly from E. coli cells as holoenzyme; a-wt, reconstituted RNAP with wt a subunit; a-256, reconstituted RNAP with the a subunit truncated at position 256; a-235, reconstituted RNAP with the a subunit truncated at position 235.

FIG. 3. Open complex formation at the pRM and pR promoters by RNAPs with alanine substitutions in the C-terminal domain of the a subunit. RNAP (20 nM) and radiolabeled DNA were incubated for 10 min at 378C prior to KMnO4 treatment. The polymerase variants were as indicated on the x axis. Wt, RNAP reconstituted with the wt a subunit; all the others are variants which have an alanine substitution at the positions indicated. The wt sequence of residues 258 to 275 is DDLELTVRSANCLKAEAI and of residues 297 to 298 is KK. Values represent averages of duplicate determinations; greater than two-thirds of the values are within 20% of the average.

extent by the inactivation of pR, and by the deletion of 10 bp between pR and pRM. In contrast, the RNAPs with a C-terminal deletion in a are unaffected by mutations which inactivate the pR promoter (pR2) or decrease the distance between pR and pRM (D10), suggesting a role for the C-terminal domain of the a subunit in the manifestation of both effects. When the ATR has been disrupted by substitution with 4 (G1C) bp, both the inactivation of pR and the 10-bp deletion (as on the pR2 ATR2 and D10 ATR2 templates, respectively [Fig. 2b]) still lead to enhanced open complex formation at pRM. This demonstrates that sequence-specific interactions involving pRM’s ATR do not play a crucial role in the mechanism whereby an RNAP bound at pR interferes with open complex formation at pRM. On the ATR2 templates also, deletion of the C-terminal domain of the a subunit leads to a two- to threefold reduction in the extent of open complex formation at pRM. Thus, the increase in open complex formation at pRM attributable to the C-terminal domain of the a subunit cannot be due to sequence-specific interactions between this domain

and the promoter’s ATR. The results shown in Fig. 2c demonstrate that the pR promoter is also less well utilized by the RNAPs with the a C-terminal deletions, indicative of a role for the C-terminal domain of a in open complex formation at pR as well. The activities of both pRM and pR are unaffected by a 10-bp displacement in the position of either promoter’s ATR: open complex formation proceeds to a similar extent on the D10 as on the wt templates. The above results are suggestive of a role for sequencenonspecific interactions between the C-terminal domain of the a subunit and upstream DNA sequences of the pRM promoter. To explore in greater detail the involvement of the a subunit in open complex formation at pRM as well as pR, we employed a collection of RNAPs bearing alanine substitutions at various positions of the C-terminal domain. These experiments involved incubation of the DNA with 20 nM RNAP for 10 min prior to treatment with KMnO4. The results are shown in Fig. 3. In panels a and b, it can be seen that the effect of the alanine substitutions is less than a factor of two, whether the pRM

6948

TANG ET AL.

construct bears (G1C) substitutions over almost the entire 240 to 260 region (the ATR22 variants [Fig. 1]) or not. It has been shown before (8, 9) as well as below (Fig. 3c) that several of the above substitutions disrupt specific interaction with a UP element. Therefore, these results are consistent with the notion that no specific contacts are made between the a subunit and DNA at the pRM promoter. The pRM construct with an upstream region similar to that of rrnB P1 shows a generally enhanced extent of open complex formation as assayed by permanganate sensitivity (Fig. 3c). On this template, a three- to fivefold reduction in open complex formation (to levels characteristic of pRM without the rrnB UP element) is seen with polymerases bearing substitutions at positions 260, 262, 265, and 268 of the a subunits, in good agreement with published effects of such substitutions on utilization of the rrnB P1 promoter (8, 9). The effects of the a mutations on open complex formation at the pR promoter (panel d) are less pronounced than those at pRM with the substituted UP element (panel c), although the patterns in the two panels are qualitatively remarkably similar in showing the greatest effects of substitutions at positions 263 and 265. The results shown in Fig. 4 were obtained by scanning the 261 to 269 region of the a subunit with tryptophan substitutions. This large amino acid could potentially have an effect beyond its inability to engage in specific contacts with DNA; it could disrupt local protein structure or sterically interfere with DNA binding by other amino acids. The effects on open complex formation at pRM promoters with and without substitutions in their ATR are more distinct than those seen in Fig. 3. The substitutions at positions 265 and 268 have the largest effect: a two- to threefold reduction in the extent of RNAPinduced sensitivity to KMnO4, regardless of whether the ATR has been disrupted by G-C substitution or not (Fig. 4a and b). These two positions are among those where alanine substitutions have been found to have the most pronounced effect on open complex formation at promoters with a 20-bp UP element (Fig. 3c) (8, 9). The same series of substitutions in a was also used on the D10 constructs to test the effect of moving the ATR closer to the 235 region (Fig. 4c and d). The results are very similar to those obtained on the templates with the wt spacing between the ATR and the 235 region of pRM (compare panels a and b with c and d). Results analogous to those presented in Fig. 3 were also obtained with 50 nM RNAP and with a longer incubation time (60 min) (data not shown). These results strongly suggest that in open complexes at pRM amino acids 265 and 268 are in proximity to the DNA, just as they are believed to be in open complexes of UP element-containing promoters. To better compare the effects of the various upstream substitutions on open complex formation, we determined rates of open complex formation by both the KMnO4 and gel shift assays as a function of RNAP concentration. The results are presented in Fig. 5, in the form of tau plots (plots of the lag time, tobs [the reciprocal of kobs] versus the reciprocal of the RNAP concentration). The datum points for the pR22UP1 promoter reflect very fast formation of open complexes (small values of tau) at pRM, comparable to results obtained for strong promoters. The least-squares fit to the data set obtained by both the gel shift and KMnO4 assays yields a small value for both the intercept and the slope (see Table 1). Three sets of datum points are displayed in the middle of the figure. They were obtained by the gel shift and permanganate assays for pR22wt, the construct with wt pRM and an inactivated pR, and by the permanganate assay for the pR22ATR22 construct, which additionally bears G-C substitutions in the 240 to 260 region. These values for tau are similar to each other at each

J. BACTERIOL.

FIG. 4. Open complex formation at the pRM promoter using the mutant RNAPs with Trp substitutions in C-terminal domain of the a subunit as determined by the permanganate assay. RNAP (20 nM) and radiolabeled DNA were incubated for 10 min at 378C prior to KMnO4 treatment. The polymerase variants were as indicated on the x axis. Wt, the reconstituted RNAP with a single mutation to phenylalanine at position 321 of the a subunit (which does not affect activity [data not shown]); 261W to 269W, the mutants which have an additional mutation to Trp at positions 261 to 269, respectively, besides the mutation to phenylalanine at position 321.

of the RNAP concentrations studied; they are on the average 5- to 10-fold higher than the corresponding tau values for the pR22UP1 promoter. The least-squares lines through the datum points obtained with both techniques for the pR22 construct and by the KMnO4 assay for pR22ATR22 are very similar, in agreement with our other results demonstrating that substitutions in the ATR do not affect promoter function. Values for kf and KB calculated from least-squares fitting of the points shown in Fig. 5, as well as by nonlinear fitting of the rectangular hyperbola (which improves the reliability of the values of the parameters obtained [24]; see Materials and Methods), have been collected in Table 1. The two methods of calculating the parameters are in good agreement, while the values obtained agree quite well with results previously reported for pRM in the context of an inactivated pR promoter (6, 14). The main effect of the presence of the UP element is shown to be a 10-fold increase in kf .

UPSTREAM INTERACTION AT pRM

VOL. 178, 1996

6949

DISCUSSION

FIG. 5. The effect of the sequence of the 240 to 260 region on the kinetics of open complex formation at the pRM promoter as determined by the gel shift and the permanganate assays. Lag times (calculated as the inverse of observed rate constants) for open complex formation were determined by the gel shift and KMnO4 assays as a function of RNAP concentration and plotted as the reciprocal of this concentration (tau plot). The insert identifies the DNA constructs used and the particular methods by which different sets of datum points were obtained. The lines shown are least-squares fits to the data, for which the points in brackets were not used. However, the latter were included in the nonlinear fitting procedure (Table 1). For pR22UP1 and pR22, the fit included data obtained by both the gel shift and the KMnO4 assays; for pR22ATR22, the data obtained by the two techniques have been individually fitted.

The gel shift experiments on pR22ATR22 showed a very slow formation of the shifted complex, yielding the much larger tau values shown in the top of the graph, and resulting in a reduced kf as tabulated in Table 1. It is unclear what the reason is for the marked discrepancy between the gel shift results with this promoter and not only the KMnO4 probing data on the same promoter but also other results presented here (Fig. 2 to 4) indicating that substitutions in the ATR have no effect on promoter activity. Inspection of the autoradiographs does not reveal any apparent differences between the gel shift experiments using this promoter and those using the similar construct with a wt ATR (designated pR22 in Fig. 5) other than more slowly appearing shifted bands. Taken at face value, the data would indicate that open complex stabilization (to the point where it would remain largely intact during electrophoresis) is a late process on the pathway of functional complex formation, occurring after the strand opening step. However, it is also quite possible that these results are due to an as yet unrecognized artifact.

TABLE 1. Summary of kinetic data for open complex formation at pRMa Construct

pR22UP1 (all)e pR22 (all)e pR22ATR22 (KMnO4)f pR22ATR22 (gel shift)g a

kf (s21) (103) b

KB (M21) (1027) c

Linear

Nonlinear

Lineard

Nonlinearc

14 6 8 2.6 6 0.7 3.2 6 1.8 0.5 6 0.1

28 6 18 2.7 6 0.7 2.3 6 0.7 1.0 6 0.3

4.0 6 4 1.3 6 0.4 1.0 6 0.6 5.3 6 2

1.5 6 2.5 1.7 6 1.1 2.2 6 1.7 1.0 6 0.6

Parameters and standard errors resulting from fits to the kinetic data. (Intercept)21 in Fig. 5. From two-parameter fit to the equation kobs 5 (kfKB[RNAP])/(KB[RNAP] 1 1); all datum points represented in Fig. 5 were used. d Intercept/slope ratio in Fig. 5. e Fit to data obtained by both gel shift and KMnO4 assays. f Fit to data obtained by the KMnO4 assay. g Fit to data obtained by the gel shift assay. b c

In the experiments presented here, we have addressed the question of the functional significance of the relatively long stretch of (A1T)-rich DNA upstream of the start site of the pRM promoter; in the course of our studies, we also touched upon the role of pR’s similarly positioned ATR. Towards this goal, we have used promoter variants with substitutions in pRM’s ATR and in the pR promoter to avoid the promoter interference we had previously observed. In addition, we studied the effect of RNAPs with deletions of, or amino acid substitutions in, the region that in other promoters was identified as being involved in contacts with upstream DNA. With the KMnO4 assay for strand opening which monitors increased T reactivity in single-stranded DNA compared with the duplex, we did not observe effects of any combination of mutant and wt DNAs and RNAPs on the reactivities of Ts on pRM’s template strand. Thus these upstream perturbations had little downstream effect on the region of pRM that became strand separated in an open complex. (a) Upstream activation at pRM. Our studies on the rate or extent of open complex formation at pRM using mutant RNAPs bearing alanine substitutions in the C-terminal domain of the a subunit or lacking this domain altogether show that the a subunit is not involved in sequence-specific interactions with the ATR of this promoter. On the other hand, the results obtained with the tryptophan substitutions show that a subset of the amino acids involved in the interaction of the a Cterminal domain with the UP element of several promoters is also important for open complex formation at pRM. The simplest interpretation of these observations is that open complex formation at pRM is facilitated by nonspecific interactions of the C-terminal domain of the a subunit with upstream DNA. As specific interactions between the a subunit and promoter DNA have been localized to the region between 240 and 260, we assume that any nonspecific interactions involving the a subunit would also occur in this upstream region. We have been unable to directly demonstrate the existence of such interactions by footprinting (data not shown), but this could be due to their rather weak nature. Experimental evidence for nonspecific interactions at the lacUV5 promoter has also been obtained (22a). The kinetic results obtained by the gel shift assay on the pR22ATR22 construct showed a much lower rate of open complex formation, apparently resulting from substitutions in the ATR. However, in assays of strand separation by KMnO4 probing, no such difference is evident. Thus if the gel shift results indeed reflect a relevant difference in the behavior of the pR22ATR22 and pR22 constructs (rather than an artifact of unknown nature), they would indicate that an ATRaccelerated step beyond initial open complex formation was needed in order to give rise to an interaction stable to gel electrophoresis. We are not aware of any observations which provide support for the existence of such a step. Experiments from several laboratories suggest that for the pRM promoter the first-order conformational change described by the parameter kf is rate limiting: the activation of this promoter by cI (11) and by inactivation of pR (6, 14, 28) have both been found to predominantly result from as much as a 10-fold increase in kf. On the other hand, upstream activation at the rrnB P1 promoter has been ascribed mostly to an effect on the initial binding, as reflected in the parameter KB (26). Thus it was necessary to address the possibility that, because pRM is not binding limited, the interactions at its ATR might be significantly stronger than the relatively small effects that we observed here would suggest. This scenario seems unlikely: as a control for our experiments on the ATR of pRM, we con-

6950

TANG ET AL.

structed the pR22UP1 promoter and found that in fact open complex formation at pRM can be activated about 10-fold by the presence of an extensive upstream ATR similar to the UP element of rrnB P1. Our kinetic results on the rate of open complex formation as a function of RNAP concentration are consistent with UP element activation of pRM resulting from an increase in kf, in agreement with studies on the other modes of activation at this promoter (see above). Because of the large errors in the values (Table 1), it is not possible to draw unequivocal conclusions as to whether KB is affected as well. It is unclear whether the kf effect is more pronounced just at pRM or whether it is a normal consequence of the presence of a UP element, which may not have been obvious at the rrnB P1 promoter because of the particular mechanism of open complex formation at the latter (26). The fact that we have performed all our experiments on linear DNA, while the published results (26) were obtained on supercoiled templates, may also have contributed to the difference. (b) Upstream activation at pR. Several observations presented in this paper are consistent with a small favorable contribution from an upstream contact at pR. As was observed with pRM, deletion of the C-terminal domain of the a subunit reduces open complex formation at pR by two- to threefold (Fig. 2c). The extent of open complex formation was also seen to be consistently less if the RNAP contained an alanine (Fig. 3d) or tryptophan (data not shown) substitution at position 265 of the a subunit. At this position previously (8, 9) and in the work presented here (Fig. 3c), an arginine residue had been shown to be crucial for the sequence-specific interaction with the rrnB P1 element. The magnitude of the effect at pR is small, however, in comparison with that observed at the promoters containing the rrnB P1 UP element (compare Fig. 3c and d). The existence of upstream interactions at pR is consistent with published footprinting results, which show protection in this region (5). In further support of such interactions at pR, an A-T substitution for a G-C base pair at position 251, resulting in an uninterrupted ATR of 12 bp, has been found to lead to a threefold-increased rate of open complex formation at pR (7). Both the latter observation and our own that pR’s activity, but not that of pRM, is sensitive to alanine substitutions in the a subunit are suggestive of some sequence-specific recognition of the (A1T) base pairs in pR’s ATR. Interestingly, it has recently been reported that open complex formation at a minor promoter (pL2) within the integration host factor binding site of the main pL promoter significantly benefits from an (A1T) region upstream of the 235 region (9). (c) Implications for the nature of the interference between RNAPs at pR and at pRM. As seen in Fig. 1, the 235 region of pR is embedded in the 240 to 260 region of pRM. Footprinting and chemical probing results (e.g., reference 5) show that a pR-bound RNAP contacts DNA over the entire region where a pRM-bound polymerase would engage in its upstream interactions, thus making it plausible that the pR-bound RNAP would interfere with the establishment of such interactions. By the KMnO4 assay to monitor open complex formation at pRM, the enhancement in the rate of open complex formation that results from inactivation of pR is readily detected (22). In Fig. 2 to 4, the effect was determined in a fixed-time assay in which the time of incubation between the various forms of RNAP and the promoter was kept constant at 10 min. In this assay, open complex formation at pRM increased two- to threefold upon inactivation of pR (Fig. 2a and b). By this same assay, both the deletion of the C-terminal domain of the a subunit and the substitution of a tryptophan at amino acid 265 are seen to also have an adverse effect of this magnitude on the extent of open complex formation at pRM (Fig. 2 and 4). While the

J. BACTERIOL.

two effects are in opposite directions, it can readily be visualized how both could involve upstream contacts which would be disrupted by either the changes in the a subunit or the presence of an RNAP at pR. On the D10 template, pRM’s 240 to 260 region and the spacer DNA separating pR’s 210 and 235 regions almost coincide (Fig. 1). As E. coli RNAP is known to make few contacts with the spacer DNA of promoters (1, 21), this stretch is now again available for contacts with an a subunit of the polymerase at pRM. We propose that on the D10 DNA the RNAPs at pR and pRM are situated sufficiently close to each other that an a C-terminal domain of the polymerase at pRM can reach over the polymerase at pR to again make the upstream contacts (22). Inherent in the above model is that contacts between the C-terminal domain of the a subunit and DNA increase the rate of open complex formation at pRM. While our results strongly favor such interactions, we have not conclusively proven their existence. Therefore, we also consider the possibility that some of the effects we observe are partially or completely due to protein-protein interactions involving the a subunit(s). It is difficult to thus explain the inhibitory effect of the wt pR promoter on open complex formation at pRM. The deletion in the C-terminal domain, which abolishes this effect (Fig. 2a and b), would then have been expected to additionally lead to an increase in open complex formation at pRM; this is not observed. In the case of the effect of the D10 deletion, however, conceivably favorable interactions between the a subunits of the RNAPs at pR and pRM could serve to approximately counterbalance other unfavorable interactions between these polymerases. This would be consistent with the observed lack of enhancement of open complex formation at pRM on the D10 templates by RNAP with a deleted a C-terminal domain. Both our previous results (22) and the results reported here suggest that open complex formation may actually be facilitated slightly at pRM on a D10 template that also bears a wt pR. In view of the subtle nature of the latter effect, it would be difficult to employ it to assess the importance of favorable proteinprotein interactions on the D10 template. Studies with a variant RNAP (such as a R265W), for which there is no indication that it interacts with upstream DNA, might be more helpful in this regard. A finding that the D10 deletion would still lead to enhanced open complex formation at pRM would be strong support for such protein-protein interactions. ACKNOWLEDGMENTS This work was supported by grant GM 31808 (to P.L.D.) from the National Institutes of Health and grants-in-aid (to A.I.) from the Ministry of Education, Science, and Culture of Japan. The core facility at Case Western Reserve University (oligonucleotide synthesis) is supported by PHS grant P30CA43703. REFERENCES 1. Auble, D. T., T. L. Allen, and P. L. deHaseth. 1986. Promoter recognition by Escherichia coli RNA polymerase: effects of substitutions in the spacer DNA separating the 210 and 235 regions. J. Biol. Chem. 261:11202–11206. 2. Banner, C. D. B., Jr., C. P. Moran, and R. Losick. 1983. Deletion analysis of a complex promoter for a developmentally regulated gene from Bacillus subtilis. J. Mol. Biol. 168:351–365. 3. Blatter, E. E., W. Ross, H. Tang, R. L. Gourse, and R. H. Ebright. 1994. Domain organization of RNA polymerase a subunit: C-terminal 85 amino acids constitute an independently folded domain capable of dimerization and DNA binding. Cell 78:889–896. 4. Burgess, R. R., and J. J. Jendrisak. 1975. A procedure for the rapid, large scale preparation of Escherichia coli DNA-dependent RNA polymerase involving polymin P precipitation and DNA-cellulose chromatography. Biochemistry 14:4634–4638. 5. Craig, M. L., W.-C. Suh, and M. T. Record, Jr. 1995. HO● and DNase I probing of Es70 RNA polymerase-lPR promoter open complexes: Mg21 binding and its structural consequences at the transcription start site. Biochemistry 34:15624–15632.

VOL. 178, 1996 6. Fong, R. S.-C., S. Woody, and G. N. Gussin. 1993. Modulation of PRM activity by the lambda PR promoter in both the presence and absence of repressor. J. Mol. Biol. 232:792–804. 7. Fong, R. S.-C., S. Woody, and G. N. Gussin. 1994. Direct and indirect effects of mutations in lambda PRM on open complex formation at the divergent PR promoter. J. Mol. Biol. 240:119–126. 8. Gaal, T., W. Ross, E. E. Blatter, H. Tang, X. Jia, V. V. Krishnan, N. Assa-Munt, R. H. Ebright, and R. L. Gourse. 1996. DNA-binding determinants of the a subunit of RNA polymerase: novel DNA-binding domain architecture. Genes Dev. 10:16–26. 9. Giladi, H., K. Murakami, A. Ishihama, and A. B. Oppenheim. 1996. Identification of an UP element within the IHF binding site at the PL1-PL2 tandem promoter of bacteriophage l. J. Mol. Biol. 260:484–491. 10. Gonzales, N., J. Wiggs, and M. J. Chamberlin. 1977. A simple procedure for resolution of Escherichia coli RNA polymerase holoenzyme from core polymerase. Arch. Biochem. Biophys. 182:404–408. 11. Hawley, D., and W. R. McClure. 1982. Mechanism of activation of transcription initiation from the lambda PRM promoter. J. Mol. Biol. 157:493–525. 12. Helmann, J. D. 1994. Bacterial sigma factors, p. 1–17. In R. C. Conaway and J. W. Conaway (ed.), Transcription: mechanisms and regulation. Raven Press, New York. 13. Helmann, J. D. 1995. Compilation and analysis of Bacillus subtilis sAdependent promoter sequences: evidence for extended contact between RNA polymerase and upstream promoter DNA. Nucleic Acids Res. 23: 2351–2360. 14. Hershberger, P. A., and P. L. deHaseth. 1991. RNA polymerase bound to the PR promoter of bacteriophage lambda inhibits open complex formation at the divergently transcribed PRM promoter: implications for an indirect mechanism of transcriptional activation by lambda repressor. J. Mol. Biol. 222:479–494. 15. Hershberger, P. A., B. C. Mita, A. Tripatara, and P. L. deHaseth. 1993. Interference by PR-bound RNA polymerase with PRM function in vitro. Modulation by the bacteriophage lambda cI protein. J. Biol. Chem. 268: 8943–8948. 16. Igarashi, K., and A. Ishihama. 1991. Bipartite functional map of the E. coli RNA polymerase alpha subunit: involvement of the C-terminal region in transcription activation by cAMP-CRP. Cell 65:1015–1022. 17. Jafri, S., M. L. Urbanowski, and G. V. Stauffer. 1995. A mutation in the rpoA gene encoding the a subunit of RNA polymerase that affects metE-metR

UPSTREAM INTERACTION AT pRM

6951

transcription in Escherichia coli. J. Bacteriol. 177:524–529. 18. Landini, P., and M. Volkert. 1995. RNA polymerase a subunit binding site in positively controlled promoters: a new model for RNA polymerase-promoter interaction and transcriptional activation in the Escherichia coli ada and aidB genes. EMBO J. 14:4329–4335. 19. Lawley, B., A. Fujita, A. Ishihama, and A. J. Pittard. 1995. The TyrR protein of Escherichia coli is a class I transcription activator. J. Bacteriol. 177:238– 241. 20. McClure, W. R. 1980. Rate limiting steps in RNA chain initiation. Proc. Natl. Acad. Sci. USA 77:5634–5638. 21. McClure, W. R. 1985. Mechanism and control of transcription initiation in prokaryotes. Annu. Rev. Biochem. 54:171–204. 22. Mita, B. C., Y. Tang, and P. L. deHaseth. 1995. Interference of PR-bound RNA polymerase with open complex formation at PRM is relieved by a 10-base pair deletion between the two promoters. J. Biol. Chem. 270:30428– 30433. 22a.Murakami, K., and A. Ishihama. Unpublished data. 23. Negishi, T., N. Fujita, and A. Ishihama. 1995. Structural map of the alpha subunit of Escherichia coli RNA polymerase: structural domains identified by proteolytic cleavage. J. Mol. Biol. 248:723–728. 24. Prosen, D. E., and C. L. Cech. 1985. Bacteriophage T7 E promoter: identification and measurement of kinetics of association with Escherichia coli RNA polymerase. Biochemistry 24:2219–2227. 25. Ptashne, M., A. Jeffrey, A. D. Johnson, R. Maurer, B. J. Meyer, C. O. Pabo, T. M. Roberts, and R. T. Sauer. 1980. How the lambda repressor and Cro work. Cell 19:1–11. 26. Rao, L., W. Ross, J. A. Appleman, T. Gaal, S. Leirmo, P. J. Schlax, M. T. Record, Jr., and R. L. Gourse. 1994. Factor independent activation of rrnB P1. An extended promoter with an upstream element that dramatically increases promoter strength. J. Mol. Biol. 235:1421–1435. 27. Ross, W., K. K. Gosink, J. Salomon, K. Igarashi, C. Zou, A. Ishihama, K. Severinov, and R. L. Gourse. 1993. A third recognition element in bacterial promoters: DNA binding by the alpha subunit of RNA polymerase. Science 262:1407–1413. 27a.Tang, Y. Unpublished data. 28. Woody, S. T., R. S.-C. Fong, and G. N. Gussin. 1993. Effects of a single base-pair deletion in the bacteriophage lambda PRM promoter. Repression of PRM by repressor bound at OR2 and by RNA polymerase bound at PR. J. Mol. Biol. 229:37–51.