Full shut-off of Escherichia coli RNA-polymerase by T7 phage requires a small phage-encoded DNA-binding protein

Abstract Infection of Escherichia coli by the T7 phage leads to rapid and selective inhibition of the bacterial RNA polymerase (RNAP) by the 7 kDa T7 protein Gp2. We describe the identification and functional and structural characterisation of a novel 7 kDa T7 protein, Gp5.7, which adopts a winged helix-turn-helix-like structure and specifically represses transcription initiation from host RNAP-dependent promoters on the phage genome via a mechanism that involves interaction with DNA and the bacterial RNAP. Whereas Gp2 is indispensable for T7 growth in E. coli, we show that Gp5.7 is required for optimal infection outcome. Our findings provide novel insights into how phages fine-tune the activity of the host transcription machinery to ensure both successful and efficient phage progeny development.


INTRODUCTION
Many bacteriophages (phages) use diverse mechanisms to repurpose, redirect or inhibit the bacterial transcription machinery, the RNA polymerase (RNAP), to coordinate phage gene expression and developmental needs during infection (1,2). T7 is a widely studied obligate lytic phage of Escherichia coli (Ec). The 56 gene products (Gps) of T7 are categorised as early (class I: Gp0.3-1.3), middle (class II: Gp1.6-6.3) and late (class III: Gp6. 5-19.5) to reflect the timing of their expression during the infection process (3,4). Early and middle genes generally encode proteins required for phage RNA synthesis, DNA replication and host takeover; the late genes specify T7 virion assembly and structural proteins. The Ec RNAP containing the major housekeeping 70 factor initially catalyses the entry of the T7 DNA into the cell by transcribing the early genes from three strong early gene promoters T7 A1, A2 and A3 (5). The host RNAP is then shut off by the coordinated action of the early gene product Gp0.7 and the essential middle gene product Gp2; the viral single-subunit RNAP (T7 RNAP, Gp1, a product of an early gene) transcribes the middle and late viral genes. The shutting down of host RNAP is crucial for the successful completion of the infection cycle: Gp0.7 is a protein kinase that phosphorylates the Ec RNAP, leading to increased termination of transcription at sites located between the early and middle genes on the T7 genome (6); and Gp2 binds in the main DNA binding channel in Ec RNAP and thereby prevents the formation of the transcriptionally proficient open promoter complex (RPo) at the T7 A1-3 promoters (7). Gp0.7 is dispensable for T7 growth under standard laboratory conditions, but becomes essential when Gp2 function is compromised (6). However, Gp2 is indispensable for T7 growth and its absence leads to interference by the host RNAP transcription initiated from T7 A1-3 promoters, with middle and late T7 gene transcription by the T7 RNAP (8).
Although several early and middle Gps of T7 have been assigned biological roles in the T7 infection cycle (9), many remain uncharacterised. One of these is Gp5.7, a 7 kDa protein of unknown function. Gene 5.7 is located downstream of gene 5.5 (encodes the Ec HNS inhibitor; see later) and the two genes overlap by one nucleotide (4), suggesting that a fusion gene 5.5/5.7 product resulting from a frame-shift translation could exist. However, Zhang et al. found that a gene 5.5/5.7 fusion protein is not expressed (10). We demonstrate that Gp5.7 is an Ec RNAP-binding protein, which specifically represses the ability of the host RNAP to initiate transcription from the early T7 A1, A2 and A3 promoters. We have elucidated the solution structure of Gp5.7, which reveals a fold similar to the winged helix-turn-helix DNA binding motif and provides a mechanistic insight into how Gp5.7 specifically alters Ec RNAP performance at T7 A1-A3 promoters. A T7 mutant that lacks Gp5.7 is compromised for growth in Ec, which underscores the involvement of Gp5.7 in the optimal inhibition of host RNAP and thus T7 growth in Ec.

Gp5.7 expression and purification
T7 gene 5.7 was PCR amplified from T7 DNA using primers AS1 and AS2 (Supplementary Table S3) containing restriction sites for NdeI and BamHI, respectively, and cloned into the same sites in plasmid pET33b+ (Novagen) to create pSW33b::gp5.7. This plasmid was then digested with XbaI and HindIII restriction endonucleases and the resulting fragment containing T7 gene 5.7 was cloned into the same sites in pBAD33 (ATCC) to create pBAD18::gp5.7, which expresses Gp5.7 as a fusion protein containing a sixhistidine (6xHis) and heart muscle kinase tags at its N terminus. All Gp5.7 mutants were created by site-directed PCR mutagenesis using pBAD18::gp5.7 and appropriate primers (available upon request) as the template and confirmed by DNA sequencing. For the biochemical experiments, recombinant wild-type (WT) and mutant Gp5.7 was made by nickel affinity purification from Ec strain MC1061. Briefly, the culture of MC1061 cells containing pBAD18::gp5.7 was grown at 37 • C to an OD 600 of ∼0.4 and Gp5.7 expression was induced with 0.04% (w/v) arabinose. The cells were left to continue growing at 37 • C for 4 h before harvesting. The cell pellet was re-suspended in Binding buffer (25 mM NaH 2 PO 4 pH 7.0, 0.5 M NaCl, and 5% (v/v) glycerol) containing cocktail of protease inhibitors and lysed by sonication. The cleared cell lysate was loaded onto a His-Trap HP column [GE Healthcare Life Sciences], which was connected to a fast protein liquid chromatograph machine. The purified protein was eluted over a 40 ml gradient of 0-100% Elution buffer (Binding buffer + 1 M imidazole pH 7.0) according to manufacturer's instructions. The purified protein was dialysed into Storage buffer (10 mM Tris-HCl pH 6.8, 50 mM NaCl, 0.1 mM EDTA, 20% (v/v) glycerol, 1 mM DTT) and stored in aliquots at -80 • C. For NMR studies, Ec strain MC1061 containing pBAD18::gp5.7 was grown in Silantes OD1 superrich medium labeled with 15 N and 13 C and Gp5.7 expression was induced with 0.04% Larabinose when the OD 600 of the culture reached 0.4; the cells were harvested following overnight incubation at 18 • C and Gp5.7L42A was purified as described above for the WT protein. The purified protein was then treated with 50 g/unit of enterokinase protease to remove the 6xHis tag; the cleaved protein was then isolated by re-applying the mixture onto a His-Trap HP column. The eluate was then dialysed into NMR spectroscopy buffer containing 50 mM NaH 2 PO 4 at pH 6, 350 mM NaCl and 0.2 mM TCEP. The samples were subsequently concentrated to ∼5 mg/ml for NMR experiments.

Pull-down assays
The MagneHis™ Protein Purification System (Promega) was used for pull-down assays: Approximately 0.02 mg of recombinant 6xHis tagged Gp5.7 in Binding buffer (20 mM Na 2 PO 4 , 50 mM NaCl, 5% glycerol, pH 7.0) was added to 100 l resin and incubated for 30 min at 4 • C. Cleared Ec whole-cell lysate was obtained from exponentially growing Ec strain MC1061, which was added to resin containing Gp5.7 and incubated for 1 h at 4 • C. For pull-down assays involving recombinant proteins, the core and holoenzyme forms of the Ec RNAP were purchased from NEB; M0550S and M0551S, respectively and ∼0.02 mg was added to resin containing Gp5.7 and incubated for 1 h at 4 • C. The beads were washed three times in 1 ml wash buffer (20 mM Na 2 PO 4 , 500 mM NaCl, 5% (v/v) glycerol, 30 mM imidazole, pH 7.0) for 10 min to remove any non-specific proteinprotein interaction. To elute samples from beads, 50 l of binding buffer and 50 l of Laemmli 2x concentrate SDS Sample Buffer was added to beads and boiled for 15 min. Ten microliters was loaded on a 10-15% SDS-PAGE alongside either 5 l of PageRuler™ Prestained Protein Ladder and stained with Coomassie Brilliant Blue or 5 l of Mag-icMark™ XP Western Protein Standard and proteins were detected by western blotting.

Electrophoretic mobility shift assays
Fifty nanomolar 32 P-labeled recombinant Gp5.7 was incubated with varying amounts of Ec core RNAP (NEB) at 37 • C for 10 min in 40 mM Tris-HCl pH 8, 10 mM MgCl 2 , 1 mM DTT and 100 mM NaCl and in the presence of 0.5 g/l ␣-lactoglobulin. The reaction was separated on a non-denaturing gel with the addition of Commassie G250 (Sigma-Aldrich) and the gel was run for 90 min and 100 V in the presence of a cooling block. The dried gel was then analysed by autoradiography.

Isothermal titration calorimetry (ITC)
ITC experiments were performed on an ITC-200 instrument (Microcal). Gp5.7L42A and double-stranded (ds) DNA probes were prepared in the same NMR spectroscopy buffer as above. Gp5.7L42A and ds DNA probes were present at 0.2 and 3 mM, respectively, and the ds DNA probe was titrated by 20 injections of 2 l every 100 s at 25 • C. The raw data were integrated, normalized for the molar concentration and analysed using Origin7.0 according to a 1:1 binding model.

In vitro promoter-independent and dependent transcription assays
These were conducted and analysed exactly as previously described (7). The sequence of the promoter-less minimal scaffold template is provided in Figure 2A. The sequences of the different promoter sequences used in Figure 2B-D are listed in Supplementary Table S3. For all the in vitro transcription assays, the core RNAP was obtained from NEB and the Ec 70 was purified exactly as described by Nechaev and Severinov (11).

Bacterial growth attenuation assays
These were conducted exactly as previously described by Shadrin et al. (12).

NMR structure determination and chemical shift experiments
NMR spectra were collected at 310K on Bruker DRX600 and DRX800 spectrometers equipped with cryo-probes. Spectral assignments were completed using our in-house, semi-automated assignment algorithms and standard triple-resonance assignment methodology (13) using the three-dimensional (3D) HNCA, HNCO, HN(CO)CA, HNCACB and CBCA(CO)NH. H ␣ and H ␤ assignments were obtained using HBHA (CBCACO)NH and the full side-chain assignments were extended using HCCH-total correlation (TOCSY), (H)CC(CO)NH TOCSY and 1 H-15 N/ 13 C NOESY-HSQC experiments. The two NOESY spectra were also used to provide the distance restraints in the final structure calculation (14). The ARIA protocol was used for completion of the NOE assignment and structure calculation. The frequency window tolerance for assigning NOEs was ±0.025 and ±0.03 ppm for direct and indirect proton dimensions and ±0.6 ppm for both nitrogen and carbon dimensions. The ARIA parameters p, Tv and Nv were set to default values. One hundred eight dihedral angle restraints derived from TALOS+ (15) were also implemented. The 10 lowest energy structures had no NOE violations >0.5Å and dihedral angle violations >5 • . The structural statistics are shown in Supplementary Table S1. For the NMR chemical shift experiments, ds DNA probes (Supplementary Table S2) was added to 15 N-labeled Gp5.7 according to stoichiometric ratio to perform NMR titration. Maximal five-fold DNA was added to Gp5.7L42A in order to broad out the entire spectra. The 10 structures were deposited to BMRB (ID: 34024) and PDB (accession number: 5LGM).

Construction of T7 gp5.7 phage
Plasmid pUC5.7::trxA was constructed to replace gene 5.7 in the T7 genome with the Ec trxA gene, a positive selection marker for T7 grown on hosts lacking trxA. The trxA in the pUC5.7::trxA plasmid was flanked by 50 bp upstream and downstream of the DNA sequence encoding gene 5.7. The plasmid was constructed by PCR amplification of Ec trxA gene by using primers MG5F and MG5R (Supplementary  Table S3), which contained the NdeI and XbaI restriction sites, respectively. The PCR fragment was then digested and ligated to a compatibly digested pUC19 plasmid. T7 encoding trxA instead of the gene 5.7 was constructed by homologous recombination as described previously (16) using pUC5.7::trxA plasmid. The obtained lysate was used to infect Ec lacking trxA. Single emerging plaques were purified and the correct insertion was verified by DNA sequencing.

T7 phage infection and plating assays
Ec strain MG1655 rpoC-FLAG and JE1134 were grown in LB at 30 • C to an OD 600 of ∼0.5. For infection with T7 phage, 1 mM CaCl 2 and 1.9 × 10 8 of T7 WT and T7 gp5.7 mutant phage were added to each culture in triplicate and OD 600 reading taken every 10 min until complete lysis was obtained or until 120 min after T7 infection. For the plating assays, Ec strain MG1655 rpoC-FLAG culture was grown to OD 600 0.7-0.8 at 37 • C and 100 l of the culture was mixed with 10 l of T7 WT or T7 gp5.7 phages (∼ corresponding to 10 plaque forming units) and incubated with 1 mM CaCl 2 for 10 min at 30 • C to allow the phage to adsorb to the bacterial cells. Three milliliters of 0.7% (w/v) LB agar, pre-warmed to 50 • C, was added and the mixture was overlaid on LB plates. After 15-18 h at 30 • C, plaques were counted. Only plaques of diameter ≥0.2 cm were taken into account. Three separate experiments were performed, and average values were calculated for each bacterial colony infected with T7 WT or T7 gp5.7.

Gp5.7 is an Ec RNAP-binding protein
In order to identify the potential interacting partner(s) of Gp5.7 in Ec, we made recombinant amino-terminal 6xHis tagged Gp5.7, immobilised it onto nickel-charged affinity resin and incubated the Gp5.7 containing resin with whole-cell lysates from exponentially growing Ec cells. After washing, Gp5.7-interacting proteins were eluted and separated by denaturing SDS polyacrylamide gel electrophoresis (SDS-PAGE). As shown in Figure 1A, inspection of the coomassie blue stained gel revealed a highly enriched doublet of bands migrating at ∼150 kDa in lysate with Gp5.7 (lane 4) compared to the control reaction (containing resin without Gp5.7; lane 3). We suspected that the doublet could be the ␤, ␤' subunits of the Ec RNAP. To confirm this, we probed the denaturing gel with antibodies against the ␤ and ␣ subunits. Results shown in Figure 1B revealed that the antibodies against the ␤ and ␣ subunits cross-reacted with the bands migrating at ∼150 and ∼36 kDa, respectively, suggesting that Gp5.7 is potentially an Ec RNAP binding protein. To ascertain that Gp5.7 can interact with the core RNAP, we incubated a fixed amount of recombinant 32 Plabeled Gp5.7 and varying amounts of Ec core RNAP and separated the reactions on a non-denaturing gel. Results shown in Figure 1C clearly demonstrates that 32 P-labeled Gp5.7 can form a binary complex with the core RNAP in a dose-dependent manner. Next, we repeated the pull-down assays with purified recombinant Ec RNAP with and without 70 and, following several washes, analysed the eluted samples on a denaturing gel. The results unambiguously established that Gp5.7 is indeed an Ec RNAP binding protein, which interacts with both 70 -free (core) and 70 -bound (holoenzyme) forms of the Ec RNAP ( Figure 1D).

Gp5.7 specifically represses Ec RNAP activity at T7 early gene promoters
Having established that Gp5.7 is an Ec RNAP binding protein, we next determined how Gp5.7 affects the activity of the enzyme. To initially determine whether Gp5.7 affects the catalytic activity of Ec RNAP, we measured the activity of the core enzyme on a promoterless minimal nucleic acid scaffold template (hereafter called the minimal scaffold (MS) probe). The MS probe consists of an 18nucleotide-long DNA duplex and an eight-nucleotide-long RNA-DNA heteroduplex separated by two unpaired DNA bases ( Figure 2A) (17). Thus, the MS probe lacks the consensus promoter DNA sequences recognised by 70 . The addition of ␣ 32 P-UTP to the E-MS probe complex results in the synthesis of a nine-nucleotide-long ␣ 32 P-UTP-labeled RNA product, hereafter called RNA-U. Results in Figure  2A revealed that the catalytic activity of Ec RNAP was unaffected by the presence of Gp5.7 in the reaction. We next conducted in vitro multiple-round transcription initiation assays on well-characterised 70 -dependent promoters (lacUV5, N25, and galP1) to determine if Gp5.7 affects the ability of the Ec RNAP to bind to the promoter, nucleate DNA strand-separation and synthesise a dinucleotideprimed RNA product. All the DNA probes used in the transcription initiation assays were 65 base pairs long with 20 base pairs of sequences downstream of the transcription start site at +1, except T7 A2p and T7 A3p which are 59 base pairs long with 14 base pairs of sequences downstream of the transcription start site at +1 (Supplementary Table   Nucleic Acids Research, 2017, Vol. 45, No. 13 7701 S3). As shown in Figure 2B, when E 70 was pre-incubated with Gp5.7 prior to the addition of the promoter probe, no detectable effect in the amount of the RNA product synthesised by the Ec RNAP was detected irrespective of the promoter used in the assay or the amount of Gp5.7 present in the reaction. Strikingly, however, under identical conditions, Gp5.7 had a substantial specific and dose-dependent adverse effect on the ability of the Ec RNAP to synthesise the RNA product from the T7 early gene promoter A1 (Figure 2C and see later Supplementary Figure S1E). Since T7 early genes are expressed from three Ec RNAP-dependent promoters (T7 A1, A2 and A3), we tested whether Gp5.7 also adversely affects the Ec RNAP on T7 early gene promoters A2 and A3. The results shown in Figure 2C clearly indicated that this is indeed the case and revealed that the ability of Ec RNAP to synthesise a RNA product from the T7 A2 and A3 promoters were equally compromised in the presence of Gp5.7 as from the T7 A1 promoter.
To determine which step during transcription initiation Gp5.7 affects on the T7 A1 promoter, Gp5.7 was added at different points of the reaction (conditions I-III; Figure  2D). Results indicated that Gp5.7 had a detectable adverse effect on the amount of the RNA product synthesised when it was added to the RNAP before holoenzyme formation (condition I) or before the initial engagement of RNAP with the promoter to form the closed promoter complex (RPc; condition II). However, Gp5.7 had only a very moderate effect on RNAP activity when added after the RPo had formed (condition III, Figure 2D), suggesting that it affects an early step during transcription initiation. Overall, the results clearly demonstrate that Gp5.7 represses Ec RNAP activity specifically on T7 early gene promoters during infection; further, it seems that Gp5.7, like Gp2 (7), affects a step en route to the RPo.

Alanine scanning mutagenesis of Gp5.7 reveals functional and structural contributions of each amino acid
The conditional expression of plasmid-borne Gp5.7 in exponentially growing Ec, like with Gp2 (12), leads to attenuation of bacterial growth (Supplementary Figure S1A). This implies that Gp5.7 can also potentially repress Ec RNAP activity at some host promoters that are essential for growth. We exploited this observation to explore the functional contribution of each amino acid (aa) residue in Gp5.7 by screening a library of Gp5.7 mutants containing an alanine substitution at every position, excluding the starting methionine and seven alanine residues found in the WT Gp5.7 sequence. The results summarised in Supplementary Figure S1B, revealed that, compared to WT Gp5.7, an alanine substitution at six conserved residues (F18, Q19, R24, S35, L42 and W52) markedly (by >90%) reduced the ability of the mutant protein to attenuate Ec growth upon induction. Although all the mutant proteins were expressed to at least near WT levels, clear differences in the expression levels were detected between the mutant proteins with respect to the WT protein (Supplementary Figure S1C). Therefore, to investigate whether alanine substitutions at F18, Q19, R24, S35, L42 or W52 adversely affected the structural integrity and stability of the mutant proteins, we performed a comparative analysis of their 1 H 1D NMR spectra together with that for the WT protein. The amide regions from the NMR spectra of F18A and L42A mutants, shown in Supplementary Figure S1D, exhibit excellent dispersion with several amide resonances at chemical shifts > 8.5 ppm. The data indicate that F18A and L42A mutants are fully folded, while R24A, S35A, Q19A and W52A lack these spectral features suggesting that these mutants possess defects in their structural integrity. Therefore, differences in structural integrity and stability does not seems to account for the differences in the expression levels observed for some of the mutant proteins with respect to the WT protein and we are not able to provide a conclusive explanation for the differences in the expression levels of the mutant proteins. However, of relevance to our analysis, in vitro transcription assays revealed that Gp5.7 mutants F18A and L42A failed to detectably repress transcription initiation from the T7 A1 promoter, suggesting that the reduced RNAP activity at the T7 early gene promoters in the presence of Gp5.7 is specific and that residues F18 and L42 are required for transcription inhibition (Supplementary Figure S1E). However, it is possible that some other residues in Gp5.7 that do not alleviate the attenuation of Ec growth may play a role in the specific transcription inhibition of Ec RNAP from the T7 A1, A2 and A3 promoters.

Gp5.7 folds into a winged helix-turn-helix (wHTH) like structure
To further understand the repressive effect of Gp5.7 on Ec RNAP activity we determined its solution structure using standard multidimensional NMR spectroscopy. Initial evaluation of the NMR spectra for WT Gp5.7 revealed that its quality deteriorated rapidly during the timescale of the experiments (data not shown). Such instability rendered the structure elucidation of WT Gp5.7 intractable. As the L42A mutant protein was significantly more stable and produced a highly structured 1 H 1D NMR spectrum that is comparable to WT Gp5.7 (Supplementary Figure S1D), we elucidated the solution structure of the Gp5.7L42A ( Figure 3A; Supplementary Table S1). The Gp5.7L42A protein folds into a compact globular domain comprising two ␤ strands packed against a four ␣ helical bundle in a ␣1-␣2-␣3-␤1-␤2-␣4 arrangement ( Figure 3A). A notable feature of the Gp5.7L42A solution structure is that surface exposed charged residues delineate a large positive patch on the ␣1 and ␣4 helices ( Figure 3A and B). The L(A)42 residue is located on the ␤1 strand and is surface exposed, consistent with the view that an alanine substitution at this position would not affect the overall conformation of the protein, but could interfere with a functional interaction.
A search for protein structure similarities using the DALI server (18) revealed that the overall structure of Gp5.7L42A exhibits a statistically significant structural similarity to the winged helix-turn-helix (wHTH) motif often found in MarR family of bacterial DNA binding transcription factors (19). The wHTH DNA binding motif of the Ec MarR protein (Z = 4.6 and an RMSD of 2.9Å over 61 equivalent C␣ atoms) consists of three ␤ strands and ␣ helices with the arrangement of ␣1-␤1-␣2-␣3-␤2-␤3, in which the side chains from the three ␤ strands of the 'wing' and the three helices enclose the hydrophobic core (Supplemen- tary Figure S2). Similarly, in Gp5.7L42A, the two antiparallel ␤ strands (␤1-␤2) form the 'wing' and the four helices (␣1-␣4) form the hydrophobic core (Supplementary Figure  S2). The loop between ␣1 and ␣2 is not well defined in the ensemble of NMR-derived structures and thus likely to be flexible. Gp5.7L42A clearly adopts an atypical wHTH arrangement (␣1-␣2-␣3-␤1-␤2-␣4) (Supplementary Figure  S2) with only two ␤ strands composed from 68 aa residues in the wing region, which is significantly shorter than that found in typical wHTH domains (80-100 aa residues).

Gp5.7 preferentially interacts with the T7 A1 promoter sequence
To determine if Gp5.7 can bind DNA and thus its ability to specifically repress Ec RNAP activity at the T7 early gene promoters also could involve an interaction between Gp5.7 and the promoter, we conducted an NMR titration experiment with a ds DNA probe spanning T7 A1 promoter sequences between -42 and -12 with respect to the transcription start site (at +1) (Supplementary Table S2; probe 1). This sequence was chosen because the binding sites of bacterial MarR protein family members are often located proximal to and/or include the transcription start-site distal consensus promoter element at -35 (19,20). A control DNA probe (probe 2) consisted of the corresponding sequence of the Gp5.7-insensitive lacUV5 promoter (Supplementary  Table S2; probe 2). We recorded the 2D 1 H-15 N HSQC NMR spectra to monitor the backbone amide chemical shift and line-width perturbations in the Gp5.7L42A spectrum in the presence of either DNA probes (∼2-fold molar excess over Gp5.7L42A). Although Gp5.7L42A showed an interaction with both probes, the NMR spectra exhibited two distinct patterns of spectra perturbations depending on which DNA probe was used. In NMR titrations with probe 2 (lacUV5 sequence) several peaks exhibited a measurable broadening with the extent of broadening correlating with the amount of probe 2 added ( Figure 4A). These residues localised to a positively charged regions located proximal to arginine residues R24, R36, R60 and R64 (Figure 4A). Strikingly, Gp5.7L42A showed a much more pronounced broadening effect at low DNA ratios for probe 1 with a larger number of specific residues affected (T7 A1 sequence): This broadening effect was observed for residues located together on the ␤-sheet wing region and ␣3 ( 4B), suggesting higher specificity and affinity for this probe. The peak broadening indicates significant conformation exchange is occurring on an intermediate NMR timescale, the lack of recovery in line-widths at the higher DNA ratios may indicate that an additional state(s) in the bound complex is present. The arginine side chains (R24, R36, R60 and R64), which exhibited perturbations in the presence of probe 2 (lacUV5 sequence) are less affected in the presence of probe 1 (T7 A1 sequence) suggesting that these arginine residues are involved in non-specific, electrostatic interaction with DNA. To localise the specific region bound by Gp5.7L42A within DNA probe 1, we constructed probe 3 (encompassing the sequence between -23 to -12 of T7 A1 promoter) and probe 4 (encompassing the sequence between -42 to -24 of T7 A1 promoter) (Supplementary Table  S2). As shown in Supplementary Figure S3A, probe 4, but not probe 3, induced a similar pattern of spectral broadening to that for probe 1. We thus truncated probe 4 further and made probe 5 (-40 to -27 of T7 A1 promoter; Supplementary Table S2), which also induced similar broadening to probe 1 (Supplementary Figure S3A). Some small chemical shift changes can be observed for some residues near the termini, which assuming that these are reporting the same physical process would indicate a K d of ∼1 mM for T7 A1 and at least an order of magnitude weaker for lacUV5. To confirm this we made isothermal calorimetry (ITC) measurements for probes 1, 2 and 5. Although we are unable to fully saturate binding in the titration due to limitations in available DNA concentrations, we obtain clear exothermic binding isotherms for T7 A1 probes 1 and 5, with estimates for the dissociation constants of K d of ∼8 and 0.6 ± 0.1 mM, respectively, while binding could not be detected for probe 2 (lacUV5) (Supplementary Figure S3B). These data are consistent with NMR observations and overall we conclude that preferential binding exists for Gp5.7L42A to a sequence of the T7 A1-A3 promoters encompassing the consensus -35 motif (which differs by only one nucleotide: TTGACT in T7 A1 and TTGACA in T7 A2 and A3). To provide additional evidence that the specific repression of Ec RNAP activity by Gp5.7 on T7 A1-A3 involves an interaction with the region encompassing the consensus -35 motif, we tested whether Gp5.7 affected transcription initiation by E 70 from the pR promoter, which has the identical consensus -35 sequence as the T7 A1 promoter. Re-  Supplementary Table S2) recorded at pH 6.0, 300 K (see key for details). Peaks with moderate broadening are highlighted in green and residues displaying major spectral perturbations are indicated (R24, R36, R60 and R64) in the spectra and in the ribbon and surface representations (right panel) of Gp5.7L42A structure. The residues footnoted by small triangles do not show notable signs of either chemical shift perturbation or peak broadening. (B) As in (A) but using probe 1 (see Supplementary Table S2). Red and orange are used to highlight major and moderate, respectively, peak broadening in the 2D 1 H-15 N HSQC spectra of the Gp5.7L42A with and without probe 1. The same residues footnoted by small triangles in (A) show significant peak broadening effect. sults shown in Supplementary Figure S3C clearly indicate that E 70 activity on the pR promoter is compromised by Gp5.7 to a similar extent as on T7 A1 ( Supplementary Figure S3C). Overall, we conclude that the specific repression of Ec RNAP activity from the T7 A1-A3 promoter involves an interaction between Gp5.7 and the nucleoprotein complex at the consensus -35 promoter region of these promoters. Although the determined affinity of Gp5.7 to the DNA is weak, it is likely to be enhanced by the interactions Gp5.7 makes with the RNAP in the biologically relevant situation (also see Discussion)

Gp5.7 is required for optimal T7 growth in Ec
The results indicate that the repression of host RNAP activity by Gp5.7 involves a preferential interaction between the T7 A1-A3 promoters. We next investigated how the absence of Gp5.7 would affect the ability of T7 phage to grow in Ec strain MG1655 rpoC-FLAG. We initially compared the ability of T7 WT and T7 gp5.7 to lyse an exponentially growing batch culture of Ec by measuring cell density (OD 600 ) as a function of time. As shown in Figure 5A, under our conditions, a rapid drop in cell density, indicating cell lysis, was observed after ∼50 min in the Ec culture infected with the WT phage. In clear contrast, in the Ec culture infected with the T7 gp5.7 phage, a relatively slower drop in cell density was only observed after ∼70 min. It seems that Gp5.7 is required for optimal growth of T7 in Ec. Consistent with this view, the comparison of the plaque morphology of T7 WT and T7 gp5.7 on Ec MG1655 rpoC-FLAG revealed that the size of the mutant plaques were smaller (by ∼30%) compared with the T7 WT ( Figure 5B). Ec strain JE1134 harbours a deletion in the rpoC gene that removes aa residues 1149-1190, which comprise the Gp2binding site (21). Thus, Gp2 does not bind to and inhibit the RNAP in the JE1134 strain, and, as a consequence, T7 does not grow normally in JE1134. However, JE1134 is efficiently infected and lysed by T7 WT, although no progeny phage results (8). Therefore, we used the JE1134 strain as an 'experimental' chassis to compare the ability of T7 WT and T7 gp5.7 to lyse JE1134 cells to determine and provide further evidence that the repression of host RNAP activity by both Gp2 and Gp5.7 is required for optimal development of T7 in Ec. Results revealed that, in marked contrast to T7 WT, which lysed the JE1134 cells ∼70 min after infection, the T7 gp5.7 phage failed to detectably lyse JE1134 cells even after 120 min after infection ( Figure 5C). We conclude that an inadequate ability to inhibit the host RNAP significantly compromises the ability of T7 gp5.7 to develop in Ec. Overall, the results indicate that, both Gp5.7 and Gp2, are required for the optimal inhibition of host RNAP and thus growth of T7 in Ec.

DISCUSSION
Despite the wealth of knowledge available on the genetic, molecular and structural basis by which T7 infects Ec, as exemplified here, there is much to discover about this simple bacterial virus. The principal finding of this study is that T7 requires, in addition to Gp2, another RNAP inhibiting small protein to temporally control the activity of the host RNAP during infection. Although Gp5.7, unlike Gp2, is dispensable for T7 growth in Ec, Gp5.7 is clearly required for optimal infection. Thus, inadequate inhibition of the host RNAP in the absence of Gp5.7 becomes clearly detrimental for T7 development in Ec (22). Since phage genomes have evolved to be compact and efficient, why does the T7 phage require three different proteins (Gp0.7, Gp2 and Gp5.7) to achieve one biological outcome? We posit that the genomic location, timing of expression and the dependency of T7 on both host and its own RNAP provides a plausible explanation for the co-existence of Gp0.7, Gp2 and Gp5.7. Infection of Ec by T7 begins with the translocation of the left end of the T7 genome into the host which is coupled to the transcription from T7 early genes from T7 A1-A3 promoters by the Ec RNAP. The early gene product, Gp0.7 is a serine/threonine kinase that phosphorylates the Ec RNAP at a single site (T1068) in the catalytic ␤' subunit (6). Phosphorylation of T1068 residue by Gp0.7 is thought to increase transcription termination at termination sites located between the T7 early (transcribed by the host RNAP) and middle genes (transcribed by the T7 RNAP) (6). Gp2, one of the first middle gene products transcribed by the T7 RNAP and is indispensable for T7 growth in Ec (8) binds to a DNA interacting domain of the Ec RNAP, called the ␤' jaw domain, and inhibits transcription initiation from all three T7 early gene promoters (23). Importantly, however, Gp2 cannot bind to or inhibit actively transcribing RNAP molecules (see later) (7). The beneficial consequences of the action of Gp2 (and Gp0.7) for the T7 phage would be to decrease host RNAP read-through into middle genes in order to avoid the slowly transcribing host RNAP interfering with the fast-transcribing T7 RNAP. During normal T7 DNA replication and packaging, the T7 RNAP recognises the genomic end of each separate unit of the progeny T7 genomes that are organised as concatemers, presumably by pausing at a unique site located immediately after the concatemeric junctions. The pausing T7 RNAP is likely to serve as a signal to recruit DNA packaging factors Gp18 and Gp19. A model proposed by Qimron et al. (22), envisages that, in the absence of Gp2, continuous transcription from the strong host RNAP-dependent early gene promoters (T7 A1-A3) during the infection cycle, could create a 'roadblock' to transcription by the much faster transcribing T7 RNAP leading to undesired pausing of the T7 RNAP. This, in turn, could result in the untimely aberrant recruitment of DNA packaging factors and production of less than unit length phage genomes leading to an unsuccessful infection of Ec by T7 phage. In support of this view, T7 infection in the absence of Gp2 fails not at the stage when the switch from host to T7 RNAP occurs, but much later, during T7 DNA packaging (24).
During the course of the infection, two middle gene products, Gp3 and Gp6, which are exo-and endonucleases, respectively, degrade the host chromosome and thereby provide resources to drive phage gene expression and DNA replication. However, the degradation of host chromosome also results in the release of host chromosome-bound proteins, such as actively transcribing (therefore Gp2 unbound; see above) host RNAP molecules or HNS (25), for which now the progeny T7 genome becomes the only available target in the cell. Aberrant binding of HNS to or host RNAP activity on the progeny T7 genomes would be detrimental. Gp5.5 counteracts the inhibitory effect of Ec HNS on phage DNA gene expression (10,26). Since the strong host RNAPdependent early gene promoters T7 A1-A3 are the only target for the host RNAP following host chromosome degradation by Gp3 and Gp6 (which would further explain the weak affinity of Gp5.7 to DNA as it does not have to compete for DNA sites), our results indicate that Gp5.7 could provide another layer of control to specifically counteract the host RNAP activity at T7 A1-A3 that has escaped inhibition of Gp2 at this point in the infection cycle to ensure optimal infection outcome. Since Gp5.7 interacts with both the RNAP ( Figure 1) and DNA (Figure 4 and Supplementary Figure S3B), we envisage a scenario where the mechanism of inhibition of transcription initiation does not simply involve the occlusion of RNAP from the T7 A1-A3 promoters, but could involve the binding of Gp5.7 to the RNAP first (since Gp5.7 seems to have a higher affinity to the RNAP than to DNA) thereby preventing productive (but not initial) promoter complex formation on the T7 A1-A3 promoters. Although deciphering the precise binding site of Gp5.7 on Ec RNAP and the mechanistic basis by which Gp5.7 represses Ec RNAP activity will be the subject of future research, the results of this study underscore that the bacterial RNAP is not only a nexus for the regulation of bacterial gene expression, but also represents an essential target for the diversity of mechanisms by which different phages acquire and modulate the activity of their host transcription machinery to ensure optimal infection outcome.