Interactions among CII protein, RNA polymerase and the λ P_RE promoter: contacts between RNA polymerase and the –35 region of P_RE are identical in the presence and absence of CII protein

Abstract

The DNA recognition sequence for the transcriptional activator, CII protein, which is critical for lysogenization by bacteriophage λ, overlaps the –35 region of the P_RE promoter. Data presented here show that activation by CII does not change the pattern of cleavage of the –35 region of P_RE by iron (S)-1-(p-bromoacetamidobenzyl)-EDTA (Fe-BABE) conjugated to the σ subunit of RNA polymerase (RNAP). Thus, the overall interaction between σ and the –35 region of P_RE is not significantly altered by CII. Therefore, the effects of the activator on RNAP binding to the promoter and formation of open complexes do not reflect a large-scale qualitative change in the nature of the interaction between RNAP and promoter DNA. The ability of CII to stimulate lysogenization is reduced in the presence of plasmid-borne rpoA variants encoding alanine substitutions at several positions in the C-terminal domain of the α subunit. However, it has not been possible to identify residues that directly affect the interaction between the activator and RNA polymerase.

INTRODUCTION

The P_RE promoter of bacteriophage λ is one of three λ promoters that are active only in the presence of the phage-encoded transcriptional activator, CII protein (1). When activated by CII, P_RE (also called P_E) directs the synthesis of CI repressor during the establishment phase of lysogenization. CII binds to directly repeated TTGC sequences (Fig. 1) that flank the –35 region of P_RE (2), which would position CII to activate transcription by contacting the α and/or σ subunits of RNA polymerase (RNAP). Although the sequence in the extended –10 region of P_RE (5′-TGTAAGTAT-3′) agrees at five positions (underlined nucleotides), with the extended consensus sequence TGNTATAAT (3,4) the –35 region of P_RE (5′-GTTTGT-3′) bears little resemblance to the consensus TTGACA (Fig. 1). In vitro, P_RE is activated nearly 700-fold by CII (5), which binds to sequences flanking the –35 consensus region, but on the opposite face of the DNA double helix from RNA polymerase (RNAP) (2). Kinetic analyses have shown that CII protein stimulates formation of closed complexes between the enzyme and P_RE by a factor of 15–20 and the isomerization of closed to open complexes by a factor of 40 (5), presumably by interacting with RNAP (6–9). Kinetic data suggest that the mechanism of activation is similar when CII activates the λ int promoter, P_I, or a third promoter, P_aQ, which is thought to antagonize λ late (lytic) transcription (10,11).

Figure 1.

Nucleotide sequence of P_RE. Sequences cited in the text are on the bottom strand, which should be read from right to left (5′ to 3′). Sequences in the –10 and –35 consensus regions of P_RE are underlined and the corresponding consensus sequences are indicated below the P_RE sequence. The underlined GT near the –10 region is part of the so-called ‘extended –10 sequence’ (4). The TTGC sequences necessary for CII binding are enclosed in boxes. Arrows indicate the transcription start sites for P_RE.

The potential interaction between CII and RNAP raises the question of whether the activator alters the interaction of RNAP with promoter DNA. In this study, we describe experiments that demonstrate that the orientation of RNAP with respect to the –35 region of P_RE is qualitatively similar to its orientation with respect to the –35 regions of strong constitutive promoters, and is essentially the same whether CII protein is present or not. Thus, although the CII-recognition site at P_RE appears to overlap the recognition site for σ⁷⁰ region 4.2 (2,12), the overlap does not substantially alter the interaction between σ⁷⁰ and the –35 consensus region.

In addition, we assayed alanine substitutions in the α-CTD (C-terminal domain) for their effects on the ability of λ to lysogenize (form clear plaques). These assays identified nine residues whose function is very important for clear plaque formation and provide further indication that the α-CTD is required for CII-mediated activation of P_RE. The phenotypes of most of these substitutions are consistent with known functions of the affected residues in binding of the α-CTD to promoter DNA or the interaction between α and σ (13,14). However, as yet, residues that could be contacted directly by CII protein have not been identified.

MATERIALS AND METHODS

Construction of P_RE

Polymerase chain reaction (PCR) was used with plasmid pSB201 (15) as template to produce two overlapping fragments in which the –10 region of P_RE was mutagenized to match the consensus sequence 5′-TATAAT-3′. One fragment extended from +7 to –179 relative to the P_RE transcription start site; it was produced by a λ l strand primer containing three substitutions necessary to produce the complement to the consensus –10 region (5′-ATTATA-3′) and a λ r strand primer whose 5′ end (on the right with respect to the λ genetic and physical map, and upstream from P_RE) encodes a BamHI site. The second fragment extended from +97 to –24 relative to P_RE; it was produced by a λ l strand primer containing an EcoRI restriction site at its 5′ end (on the left with respect to the λ map and downstream from P_RE) and a λ r strand primer containing the –10 consensus sequence 3′-TAATAT-5′. The two isolated fragments were separately denatured, mixed together, annealed and amplified by PCR using the left- and rightmost primers that were used originally to generate the two overlapping fragments. The resulting mutagenized fragment contains phage sequences from –179 to +97 relative to the P_RE transcription start site and terminates with the upstream BamHI and downstream EcoRI sites encoded by the primers. The same primers were used to recover the corresponding wild-type P_RE-containing fragment directly from pSB201 for use in iron (S)-1-(p-bromoacetamidobenzyl)-EDTA (Fe-BABE)-mediated cleavage reactions.

Fe-BABE-mediated DNA cleavage reactions

A derivative of σ⁷⁰ containing a single cysteine at position 581 was purified and conjugated with Fe-BABE as described previously (16–18). Core RNAP was incubated with a 5-fold molar excess of Fe-BABE-conjugated σ for 30 min on ice to form holoenzyme. Open complexes were allowed to form for 30 min at 37°C by incubating reconstituted RNAP and end-labeled DNA (1 nM) in 10 mM HEPES pH 7.9, 50 mM potassium glutamate (KGlu), 10 mM MgCl₂ and 0.1 mM EDTA. For experiments performed in the presence of CII protein, 1 µg of purified native CII (5) or CII containing an N-terminal hexahistidine tag was incubated with template DNA for 5 min at 37°C prior to the addition of RNAP. Final concentrations of RNAP and CII were 25 nM and 10 µg/ml, respectively. Following the incubation of RNAP and DNA, a sample was removed and added to a 20× stock of ascorbate and H₂O₂, to produce final concentrations of 5 mM each. After 10 min at 37°C, reactions were quenched and samples were assayed by gel electrophoresis as described by Marr et al. (18). DNA was isotopically labeled at the 5′ end by phosphorylating either of the two primers used for PCR amplification of the EcoRI–BamHI fragment with γ-[³²P]ATP and T4 polynucleotide kinase.

Transcription in vitro

In parallel with the Fe-BABE-mediated cleavage reactions, in vitro transcription catalyzed by Fe-BABE-conjugated holoenzyme was assayed as follows. After incubation of RNAP and DNA as described above, an aliquot was adjusted to 200 µM ATP, GTP and CTP, 50 µM α-[³²P]UTP (0.5 µCi/µl) and 100 µg/ml of heparin. After 5 min at 37°C, a 20 µl sample was removed and added to 100 µl of transcription stop buffer (0.6 M Tris pH 8.0, 12 mM EDTA, 0.2% SDS, 50 µg/ml tRNA). Samples were prepared for gel electrophoresis and analyzed as outlined previously (18).

Assays of CII-mediated activation of P_RE and CI-mediated activation of P_RM by reconstituted RNAPs containing alanine substitutions were performed under conditions similar to those described previously (7). The DNA templates were a P_RE-containing 663-bp HindIII–EcoRI restriction fragment from pSB201 (15) and a P_RM-containing 1180-bp PvuII–BamHI fragment from pSW301 (19). Activity at P_RE was normalized with respect to activity of the strong constitutive oop promoter (7) and all experiments shown here were performed in excess RNAP.

RNA polymerase, repressor and CII preparations

Native CII protein was purified as described previously (5). N-terminally His-tagged CII was overexpressed from the vector PET-28c (Novagen), and purified on Ni-agarose using methods outlined by the supplier (Qiagen). RNAP holoenzyme was reconstituted from core protein and Fe-BABE-conjugated σ for the experiments described in Figure 3. In experiments involving altered RNAPs, the reconstituted alanine-substituted enzymes were the generous gifts of Drs W. Ross, R. Gourse, T. Santangelo and S. Aiyar. In some experiments, RNAP holoenzyme purchased from Epicentre was used as a control. Purified phage repressor, provided many years ago by R. Sauer, was nearly as active as when it was originally purified.

Figure 3.

Patterns of Fe-BABE induced cleavage at several promoters. Bars represent regions in which Fe-BABE-mediated cleavage was observed. Dots represent nucleotides in promoter DNA. Upper strand is the non-template strand, bottom strand is the template strand. Black bars, strong cleavage; grey bars, moderate cleavage; stippled bars, weak cleavage. Data for P_R′ (λ) and P_R′ (82), lacUV5, and galP1 are taken from Marr et al. (18), Owens et al. (15), and Bown et al. (20), respectively.

RESULTS

Probing σ⁷⁰ contacts with the –35 region of P_RE

Contacts between RNAP holoenzyme and nucleotides in the –35 region of P_RE were identified by analyzing the cleavage of a P_RE-containing DNA fragment by the reagent Fe-BABE conjugated to a cysteine at position 581 of σ⁷⁰ (18). Residue 581 is in σ⁷⁰ region 4.2, which interacts with the –35 consensus region. Upon the addition of H₂O₂ and ascorbate, Fe-BABE creates localized hydroxy-radicals that cleave phosphodiester bonds. Fe-BABE-mediated cleavage indicates sites on promoter DNA with which the reagent comes into contact when σ is bound to the promoter (20). Cleavage of phosphodiester bonds yields single-stranded fragments whose lengths identify sites at which the conjugated probe comes into contact with the DNA.

Contacts between RNAP and two promoters were investigated: P_RE wild-type and P_RE^# (4), which is similar to P_RE, except that the –10 region is mutated to match the consensus TATAAT. As a consequence, P_RE^# is active in the absence of CII protein (4), even though it retains the non-consensus wild-type –35 sequence. The patterns of cleavage following binding of holoenzyme containing Fe-BABE-conjugated σ to ³²P-labeled DNA are illustrated in Figure 2. At wild-type P_RE, there is no significant cleavage of either the non-template (Fig. 2A) or template (Fig. 2C) strand in the absence of CII protein (second lane), but in the presence of native or His-tagged CII protein (lanes labeled ‘+’) there is significant cleavage of both DNA strands. Sites of strong cleavage on the template strand are detected between –38 and –43 and between –28 and –31. On the non-template strand, there is cleavage between –33 and –38 (strong cleavage) and between –48 and –41 (moderate cleavage). Weak cleavage of the non-template strand between –23 and –27 is also observed, but is difficult to see in Figure 2. These cleavage patterns, shown schematically in Figure 3, resemble the patterns observed at λ P_R′ (18), lacUV5 (16) and galP1 (21).

Figure 2.

Fe-BABE-mediated cleavage of P_RE. In each of the four sets of figures, the four lanes represent (in order) the A/G Gilbert-Maxam sequencing pattern, followed by data obtained in the absence of CII, in the presence of native CII and in the presence of His-tagged CII. In each section of the figure, lanes designated ‘RNA’ represent the results of in vitro transcription assays performed in parallel with the Fe-BABE cleavage reactions.

In single round transcription assays, native CII (Figure 2, third lane in each panel) stimulated open complex formation during a 30-min incubation with Fe-BABE-conjugated RNAP 60- or 100-fold when the template was labeled on the template or non-template strand, respectively. Thus, the Fe-BABE-conjugated enzyme responds normally to CII protein.

Similar results were obtained when open complexes at the constitutive promoter P_RE^# were probed with Fe-BABE-conjugated RNAP (Fig. 2B and D). The critical experiments are those performed in the absence of CII protein, which show that RNAP makes qualitatively identical contacts in open complexes formed at two promoters with the same –35 sequences in the presence and absence of CII protein. When CII protein is added to P_RE^#, cleavage is stimulated at least 2-fold (Fig. 2B and D), but the patterns of cleavage of the template and non-template DNA strands are again unchanged. Therefore, CII protein does not detectably alter Fe-BABE-assayable interactions between RNAP and the –35 region of P_RE, regardless of the sequence of the –10 region of the promoter.

Amino acids in the α-CTD that are required for CII-mediated transcriptional activation

The CTD of the RNAP α subunit is known to be required for CII-mediated activation of P_RE, since two α-CTD deletions inhibit activation of P_RE in vitro (7). Furthermore, Wegrzyn et al. (8) showed that rpoA341 (22) affected CII-mediated activation of P_I and P_RE in vivo. Consistent with this result, λ forms clear plaques, but can stably lysogenize the mutant strain rpoA341 at low frequency (S.E.B. and G.N.G, unpublished data), which is the expected phenotype of a mutant defective in CII-mediated activation of P_RE. To identify possible sites of contact for CII in the α-CTD, a bank of 70 plasmids containing alanine substitutions in rpoA (23–25) was screened for effects of the substitutions on the ability of λ to lysogenize. An rpoA341 mutant strain was transformed with each of the rpoA substitution plasmids (obtained from R. Gourse) and assayed for the ability of each plasmid to complement the rpoA341 defect, i.e. to permit λ to form turbid plaques. Since λ can form stable lysogens at low frequency in rpoA341 mutants, clear plaque formation should reflect effects on CII-mediated activation of P_RE, P_I and/or P_aQ (1), but not on repressor-mediated activation of the maintenance promoter, P_RM.

Nine alanine substitutions in the α-CTD strongly affected the ability of λ to form turbid plaques (Table 1, ‘cc’) and four substitutions had weaker effects (Table 1, ‘c’). Of the substitutions that caused very clear plaque-formation, four (N268A, C269A, G296A and L295A) likely affect CTD binding to UP DNA sequences (14,26), because they are part of the ‘265 determinant’ of the α-CTD, which affects DNA binding (13,26–28). Two additional substitutions (V257A and E261A) are likely to be in the ‘261 determinant’ of the α-CTD, which affects the interaction between the α-CTD and σ, although this has not been demonstrated directly for V257A (13,14,26,27). The effect of these substitutions on lysogenization suggests that alterations in the structure of the DNA-binding domain of the α-CTD or in the interaction between α-CTD and σ are critical for CII-mediated activation of P_RE, P_I or P_aQ.

Table 1. Effect of alanine substitutions on α-CTD function and turbid plaque formation.

Amino acid residue/mutation	Activity in UP reporter assaya	Molecular defect	Phenotype in λ plaque assaysb	Activity in PRE reporter assay (%)c
Strong effects on proximal UP element function
D259A	0.51/1.1	α–σ contacta	t	∼75
E261A	0.39/0.49	α–σ contacta	cc	∼70d
T263A	0.41/0.66		t	>100
R265A	0.24/0.15d	UP DNA contacte,f	NT	∼35
N268A	0.42/0.20d	UP DNA contacte,f	cc	∼40g
C269A	0.39/0.20d	DNA bindinge	cc	>100d
L290A	0.62/0.68		t	>100
L295A	0.61/0.39		cc	∼50g
G296A	0.34/0.20	UP DNA contacte,f	cc	∼45g
K298A	0.54/0.22	UP DNA contacte,f	c	∼70
S299A	0.24/0.22	UP DNA contacte,f	t	∼75
E302A	0.44/0.34		t	∼75
Weaker effects on proximal UP element function
D258A	0.73/1.4	α–σ contacta	c	∼75
L260A	0.68/0.92h		t	∼70
L262A	0.73/0.49e,h		t	∼75
V264A	0.63/0.59		t	>100
S266A	0.63/0.68e		t	>100
P293A	0.83/0.48		c	>100d
I303A	0.68/0.41	Core structure	cc	∼60g
V306A	0.68/0.56		t	>100
L314A, L281R	0.88/0.59	CII–α?	cc	∼25g
Marginal or no effect on proximal UP element function
V257A	1.2/1.4	α–σ contacti	cc	∼80d
K271A	1.4/1.4	α–σ contacti	c	∼80d
N320A	0.93/0.76	CII–α?	cc	∼40g

^aCalculated from figure 4 of Ross et al. (13). First datum for each mutant reflects activity of proximal subsite of UP; second datum for each mutant reflects activity of full UP promoter element.

^bcc, very clear plaques, similar to those made on rpoA341 mutant cells containing a control plasmid lacking rpoA; t, turbid plaques, similar to those made on rpoA341 mutant cells containing a control plasmid encoding wild-type rpoA; c, partly turbid plaques, with phenotypes intermediate between cc and t; NT, not tested.

^cCalculated from Figure 2 of Kedzierska et al. (31) and rounded to closest 5%.

^dSubstitutions that affect plaque turbidity but cause <40% reduction in the P_RE reporter assay.

^eDNA-binding assays (25).

^fX-ray data (28).

^gSubstitutions with strong effects in plaque turbidity and P_RE reporter assays.

^hMurakami et al. (46).

ⁱFunction inferred from location in α-CTD and phenotype of E261A.

Since the role of the six residues alluded to above has been well studied in other laboratories and several have been studied in vitro, three remaining substitutions with strong effects on lysogenization (N320A, I303A and the double substitution L281R/L314A) were assayed for their effects on CII-mediated activation of P_RE in vitro. N320A and L314A are surface-exposed in the nuclear magnetic resonance structure of the α-CTD (29). We also assayed K271A, which had a modest effect on plaque turbidity, because it corresponds in position to the rpoA341 mutation, K271E.

Figure 4A shows results of typical single-round in vitro transcription assays of CII-mediated activation of P_RE in the presence of reconstituted alanine-substituted or wild-type RNAP. In these experiments, the final RNAP concentration was 60 nM and open complexes were allowed to form for 15 min in the presence or absence of CII protein. K271A and N320A do not significantly affect P_RE activity under these conditions, but the double substitution L281R/L314A has a very strong effect, and I303A reduces activity by more than a factor of two. Based on additional experiments (not shown) at several RNAP concentrations, we estimate the relative activities of the alanine-substituted enzymes (after normalization to the activity of the constitutive oop promoter) to be: wild type (≈1.0); N320A (0.99); K271A (0.90); I303A (0.43); and L281R/L314A (0.06), with an experimental error approximating 20% of the indicated values. Under conditions in which RNAP was limiting (<10 nM), I303A polymerase was preferentially affected; its activity was reduced to approximately 10–20% of wild type. However, the relative activities of the other RNAPs were not changed by altering the concentration(s) of CII or RNAP, or changing the time allowed for open complex formation (data not shown).

Figure 4.

Activation of P_RE and P_RM by mutant RNAPs. (A) P_RE activity in the presence (+) or absence (–) of CII protein. Final concentrations were: CII, 6 µg/ml; RNAP, 60 nM; DNA, 2 nM. Ratios of activity at P_RE to activity at P_oop in the presence of CII protein: α271 (K271A), 1.2; α303 (I303A), 0.5; α281/314 (L281R/L314A), 0.1; α320 (N320A), 1.4; wild type (WT), 1.2. (B) P_RM activity in the presence (+) or absence (–) of CI protein. Final concentrations were: CI, 10 µg/ml; RNAP, 20 nM; DNA, 2 nM. Ratios of activity at P_RM in the presence of CI protein to activity in the absence of CI are indicated.

We also assayed CI-mediated activation of P_RM (7) in the presence of the alanine-substituted enzymes (Fig. 4B). As expected, none of the altered enzymes was significantly defective in activation of P_RM. The average ratios of activity in the presence and absence of repressor were: wild type (2.2); L281R/N320A (2.6); K271A (2.1); I303A (2.4); and L314A (2.5; one experiment only). Thus, none of the alanine-substituted enzymes can owe the clear-plaque phenotype to a defect in activation of P_RM.

Originally, the doubly substituted variant L281R/L314A was thought to contain only the single substitution at position 314; it was thus thought identify a potential contact point for CII, since L314 is located close to regions contacted by other activators (e.g. CAP) (27). However, the defect in the ability to initiate transcription in vitro appears to be due to L281R, which is likely to affect folding of the α-CTD (13). We have tested the single substituion L314A (the plasmid was provided by Dr G. Christie) (30) and have found that the substitution has very little or no effect on the ability of λ to form clear plaques. Kedzierska et al. (31) also find that the activity of the L314A α subunit in the reporter assay is ∼90% of wild type.

DISCUSSION

Characteristics of P_RE

In spite of the fact that the P_RE –35 sequence does not agree with the consensus TTGACA, early genetic (32) and kinetic (5) analyses demonstrated that –35 region nucleotides are critical for promoter activity in the presence of CII protein. On the other hand, –35 region mutations had no effect on the very low levels of promoter activity detected in vitro in the absence of CII (5). Since these mutations do not affect CII binding per se, it was suggested that even when –35 nucelotides do not match consensus, specific RNAP contacts in the –35 region may be critical for promoter function in the presence of the activator. There are two explanations for this phenomenon: (i) the activator alters the conformation of RNAP so that close, specific contacts are made with non-consensus nucleotides; and (ii) RNAP makes fairly close contacts with the –35 regions of all promoters, regardless of their sequence. These explanations are not mutually exclusive.

The data presented here support the idea that the physical relationship between Fe-BABE conjugated to σ⁷⁰ and the non-consensus –35 region is qualitatively the same in three situations: at P_RE^#, whose –10 region matches consensus, in the presence or absence of CII; and at wild-type P_RE in the presence of CII. The important point is that CII does not alter this relationship, which is very similar—indeed equivalent—to that determined using similar methods at promoters with near-consensus –35 regions (18) and to a CAP-dependent Class II promoter (21). However, subtle changes between σ and the –35 region might not affect Fe-BABE cleavage patterns.

Previous studies have indicated that CII has at most a 2-fold stimulatory effect on transcription in vitro or in vivo from P_RE^# (4). These in vitro results were obtained at limiting RNAP concentration, with no preincubation of RNAP and DNA to allow open complexes to form before the addition of substrates. In the experiments reported here, RNAP was in excess and 30 min were allowed for open complex formation before the addition of substrates for in vitro transcription. These conditions are more favorable for open complex formation and therefore should reduce the stimulatory effect of CII protein. In fact, the degree of stimulation (Figs 2B and D) was not significantly different from 1 (the ratio of activity in the presence and absence of native CII protein on P_RE^# DNAs for which the template or non-template strands were labeled was 1.2 ± 0.5). On the other hand, Fe-BABE-induced cleavage of P_RE^# was stimulated at least 2-fold by both native and His-tagged CII. It is not clear why these two assays (Fe-BABE-induced cleavage and single-round transcription) resulted in different degrees of stimulation by CII.

Effects of α-CTD alanine substitutions on activation

The effects of alanine substitutions on plaque turbidity and their effects in a P_RE reporter assay (31) are in reasonable agreement (Table 1). Seven substitutions decreased P_RE-directed β-galactosidase synthesis in vivo by at least 40% (R265A, N268A, L295A, G296A, I303A, L314A and L320A). Six of these were tested in the plaque assay (R265A has not been tested) and all six caused λ to make very clear plaques. Two substitutions (V257A and E261A) caused λ to make very clear plaques, but only decreased P_RE-directed β-galactosidase synthesis by 20–30% (31), and one substitution (C269A) that caused λ to make very clear plaques had no effect on P_RE-directed β-galactosidase synthesis. Thus, the plaque assay appears to be the more sensitive test for an effect of the alanine substitutions on the ability of CII to activate P_RE, P_I or P_aQ.

In vitro transcription

N320A, which strongly affected turbid plaque formation by λ in the plaque complementation assay and β-galactosidase activity in the reporter assay (31), did not significantly affect activation of either P_RM or P_RE in vitro. N320 is located on the surface of the CTD in a C-terminal loop that is present in Escherichia coli, but not Thermus thermophilus rpoA (33), in a position that would permit contact with an activator. Nearby mutations in this loop, which have been identified in other systems (30,34), have relatively mild effects on transcriptional activation. How can the clear-plaque phenotype of N320A be explained? One possibility is that the substitution has a small (10–20%) effect on activation of several promoters, including P_I and P_aQ, as well as promoters in the bacterial chromosome; these bacterial promoters could direct transcription of many, possibly unidentified genes whose products have some influence on the frequency of lysogenization. On the other hand, the optimal conditions used for transcription in vitro do not reflect conditions in vivo. However, no phenotype was seen for N320A RNAP even when CII or RNAP was limiting or the amount of time allowed for formation of open complexes was varied (data not shown).

Finally, the rpoA341 mutant α may be preferentially assembled into RNAP core enzyme in spite of the fact that the plasmid-borne rpoA gene is thought to be expressed at about twice the level of the chromosomal gene (23). Previous studies of activation defects based on phenotypes of partial diploids containing the chromosomal rpoA341 allele may be subject to the same complication (30).

Does CII contact RNAP?

Although there is, as yet, no direct evidence for a specific interaction between CII and RNAP, the importance of amino acids in the α-CTD for lysogenization, and the proximity of CII to both α (31) and σ (2,12) make it hard to imagine that there is no contact between these two subunits and CII. Furthermore, DNA binding of CII and RNAP are synergistic (6), and CII affects K_B (as well as k_f) in kinetic assays (5,10). However, so far, contact between CII and the α-CTD has not been demonstrated directly. L320A may define a region of contact, based on its location, plaque-morphology phenotype, and effect on β-galactosidase synthesis, but the in vitro transcription data do not support this conclusion. I303A reduces activity in vitro by a factor of approximately two when RNAP is in excess; however, its location suggests that it is not accessible to contact by an activator (29).

An apparent paradox is that rpoA341(K271E) completely prevented CII activation in vivo (9), but deleting the entire α-CTD reduced CII activation of P_RE by only 80–85% in vitro (7). We hypothesize that CII-mediated activation of P_RE depends on contacts not only with α, but also with σ. The paradoxical effect of K271E can be explained by supposing that this mutation completely disrupts normal interaction(s) between α and σ, and thus between σ and CII. According to this hypothesis, α-CTD deletions would not impair contact(s) between CII and σ. It is also possible that deleting the CTD permits CII to make an entirely new activating contact with a previously masked surface on σ or some other part of RNAP.

A set of alanine-substituted rpoD (sigma) variants (35) has been assayed for their ability to support turbid plaque-formation. Four substitutions cause λ to form very clear plaques (S.E.B. and G.N.G., unpublished data), L595A, S602A, R603A and S604A. Substitutions in this region cause defects in the ability of the α-CTD to interact with σ (13,28). In particular, R603 interacts with D259 and E261 of the α-CTD at rrnB P1 (13). The role of L595, S602 and S604 is unclear. They are in a domain of σ that was originally thought to be a site of contact for the λ CI protein (35–38), a class II activator that does not contact the α-CTD (7). Thus, although the effects of these alanine substitutions define residues in σ required for CII-mediated activation, the altered RNAPs have not been tested for their phenotypes in vitro, and there is no evidence that contact actually occurs between CII and σ.

Nevertheless, by analogy with some other promoters (3,39–44), the demonstrated effects of CII on k_f (5) could be mediated by contacts between CII and σ. If that turns out to be the case, then it would be interesting to determine why mutations in the –35 region of P_RE primarily affect K_B (45).

The results presented here, together with those of Kedzierska et al. (31), can be used to draw inferences regarding the structural relationships between the α-CTD 261 determinant and σ region 4, between the α-CTD 265 determinant and base pair –41 of P_RE, and between σ region 4 and the –35 region of P_RE. CII contacts Gs in the template strand at –37 and –27 of P_RE and in the non-template strand of P_RE at –36 and –26 (2). In spite of the fact that σ also appears to contact the template strand at –36/–37 (12), data presented in Figure 2 show that CII protein does not substantially perturb the interaction of σ with P_RE. The positioning of the α-CTD near –41 in the presence of CII (31), which is similar to that found at factor-independent promoters (13,14), is apparently made possible by the fact that CII is bound on the opposite face of the helix from that contacted by σ region 4 or the α-CTD (2,31).