Different Modes of Transactivation of Bacteriophage Mu Late Promoters by Transcription Factor C

Transactivator protein C is required for the expression of bacteriophage Mu late genes from lys, I, P and mom promoters during lytic life cycle of the phage. The mechanism of transcription activation of mom gene by C protein is well understood. C activates transcription at Pmom by initial unwinding of the promoter DNA, thereby facilitating RNA polymerase (RNAP) recruitment. Subsequently, C interacts with the ß' subunit of RNAP to enhance promoter clearance. The mechanism by which C activates other late genes of the phage is not known. We carried out promoter-polymerase interaction studies with all the late gene promoters to determine the individual step of C mediated activation. Unlike at Pmom, at the other three promoters, RNAP recruitment and closed complex formation are not C dependent. Instead, the action of C at Plys, PI, and PP is during the isomerization from closed complex to open complex with no apparent effect at other steps of initiation pathway. The mechanism of transcription activation of mom and other late promoters by their common activator is different. This distinction in the mode of activation (promoter recruitment and escape versus isomerization) by the same activator at different promoters appears to be important for optimized expression of each of the late genes.


Introduction
Bacteriophage Mu utilizes the host RNA polymerase for its transcription. Expression of the genes in phage Mu is regulated in a temporal fashion. Like in other phages, the Mu genome is divided into transcription units designated as early, middle and late genes based on their timing of expression during the phage life cycle. Mu late genes that are expressed during the last phase of the lytic cycle, subsequent to the initiation of the phage DNA replication are dependent on C protein for their expression (Fig 1) [1]. Among the four late genes, the lys gene product is required for host cell lysis at the end of the cycle. Mutants defective in lys make normal amounts of functional phage particles but do not release them from the cell [2,3]. I gene has been shown to be required for head synthesis and possibly involved in protein scaffolding during head assembly [3,4], while P gene encodes for one of the enzymes involved in the synthesis of phage tail [3,4]. The mom gene encodes for a unique DNA modification function [5,6] that confers an anti-restriction phenotype to the phage genome making it refractile to the host restriction endonucleases [7]. Although mom is not essential, it confers survival advantage to the genome [7]. Notably, its untimely or inappropriate high level expression leads to cell death owing to its unique DNA modification function [8,9].
The unusual DNA modification by Mom and the associated cytotoxicity due to its expression formed the basis for the detailed analysis of its regulation. A variety of regulatory measures operational at transcriptional and post-transcriptional levels are well elucidated, including, C mediated transactivation [8,[10][11][12][13][14]. C binds at its site overlapping the mom promoter, alters the DNA conformation by unwinding, thereby, facilitating RNAP recruitment at the otherwise out-of phase promoter elements unmalleable for enzyme occupancy [15,16]. In the next step, C interacts with ß' subunit of RNAP, inducing allosteric transitions in the enzyme active site and thus enhancing promoter clearance [17,18]. However, little is known about the mechanism of activator mediated transactivation at lys, I, P promoters which are also expressed only in late phase of lytic cycle in C dependent manner [1,14,19]. Mu late promoter sequences. Late promoter sequences were aligned with respect to their transcription start sites as determined by S1 nuclease mapping (Bolker, et al., 1989, Margolin, et al., 1989, and similarities in -10 and -35 regions. Heavy black line indicates the C-foot print in lys and mom promoters. A 10bp match between lys and I promoters and the 6T stretch common between lys and mom promoters in the spacer region are underlined in green and magenta respectively. Sequences of non-consensus -10 (boxed) and -35 elements (underlined in blue) are depicted. In the present study, we set out to determine the facets of C mediated transactivation at lys, I, P promoters, asking the following questions: Would the mechanism at these promoters involve a multi-step control as seen at P mom ? Is C binding a prerequisite for RNAP occupancy at these promoters? Alternatively, does it follow a single step mechanism typically seen with a majority of promoters subjected to activation? We describe the promoter-polymerase interaction studies in presence of C, aimed to understand C function at lys, I and P promoters. We demonstrate that the mode of action by C at these promoters is distinct to the mechanism seen at Pmom.

Results
Recruitment of RNAP at the lys, I and P promoters does not require C protein In vitro transcription analysis carried out on the templates generated from the late promoter clones revealed that C is essential for transactivation from the lys, I and P promoters similar to that at P mom . No specific transcription was observed at these promoters in the absence of the C protein (Fig 2), confirming their dependence on C described earlier [14,19]. To understand the mechanism of activation by C at these promoters, individual steps of transcription initiation pathway were assessed. To address whether there is any requirement of transactivator C for the recruitment of RNAP at the lys, I and P promoters, Electrophoretic Mobilist Shift Assay (EMSA) was carried out on the end-labeled late promoter DNA fragments. DNA binding (K B ) was estimated by measuring the intensity of DNA-protein complex and plotting the value as a function of RNAP concentration (Fig 3A-3C). Similar experiments were carried out on transactivator independent mom promoter mutant-Ptin7 as a control, where RNAP recruitment occurs in the absence of C protein (Fig 3D). From the data, it is evident that C does not have an effect at the step of closed complex formation, as the K B values in the presence or absence of C were comparable. Thus, unlike at P mom where C facilitates RNAP recruitment by unwinding the DNA [15,16], RNAP recruitment appears to be C independent at lys, I and P promoters. Moreover, the affinity of RNAP to these late promoters is comparable and not altered in presence of C. These results are markedly contrasting to RNAP binding at P mom [17,20,21].

Influence of C protein at the step of isomerization and transcription elongation
Conversion of closed complex to open complex is the next step during initiation and also one of the major steps for transcriptional regulation at a number of promoters. Many transactivator C is essential for transactivation from late promoters. In vitro transcription assays carried out on lys,I,P (lanes 1-6), and at P mom (lanes 9-10) promoters in the absence and presence of transactivator C. At mutant mom promoter-P tin7, RNAP recruitment is C independent. The basal transcription in the absence of C is enhanced in its presence due to second step activation described in the text (lanes 7-8).
doi:10.1371/journal.pone.0129504.g002 proteins exert their effect at this stage [22][23][24][25][26]. Previous studies with P tin7 , a mutant mom promoter in which C is not required for RNAP binding [20], revealed that C does not enhance the formation of open complex at this promoter [17]. Similar experiments were carried out to study the effect of C on the open complex formation at lys, I and P promoters as described in Materials and Methods. In all the three promoters, C facilitated open complex formation. Although very faint heparin resistant complexes were observed in the absence of C (see later section), they are likely to be non-functional open complexes (see later section/paragraph). As shown earlier, extent of formation of open complex was unaltered in the presence and absence of C at P tin7 compared to that observed at lys, I and P promoters (Fig 4A-4D).
At P mom (and also at P tin7 ), C enhances promoter escape by overcoming abortive initiation [17]. To address the effect of C during promoter clearance step of transcription initiation at lys, I, and P promoters, in vitro transcription assays were carried out. The abortive initiation profiles at each of these promoters were compared in the presence or absence of C in the reaction. C is required for the RNAP to enter a productive elongation phase at lys, I, P promoters (Fig 5). The presence of lower levels of abortive transcripts at these promoters (Fig 5 lanes1, 3, 5) could be attributed to the weak complexes formed in the absence of C protein (Fig 4A-4C). These open complexes appear to synthesize only the abortive transcripts that fail to enter productive elongation phase in the absence of transactivator C (Fig 5 lanes 1, 3, 5). These results suggest the formation of a moribund complex/dead-end complex [27] by RNAP at lys, I, P promoters in the absence of transactivator C. Thus, at these three promoters, C is required for productive open complex formation leading to efficient promoter escape and transcription elongation.

Effect of G524DrpoC RNAP on transactivation from lys, I, P promoters
The interaction between DNA bound transactivator C and the β 0 subunit of RNAP, and the concomitant allosteric changes in the enzyme have been shown to be important for RNAP to enter into productive elongation phase at P mom [18]. Positive control (pc) mutants of C or specific mutations in ß' subunit of RNAP affected this second step transactivation at P mom [17,18]. P mom specific transactivation deficient mutant of rpoC (ß' subunit) G524DrpoC RNAP [18] isolated in our previous studies was used to understand the effect of the mutation on transcription at the late gene promoters. Single round transcriptions were carried out from lys, I, P, mom and tin7 promoters with wild type (WT) and G524D RNAP. At P mom and P tin7 where C exerts a dual step activation mechanism, G524D RNAP exhibited reduced transcription in the presence of C (Fig 6A, lanes 7-10) [18]. The same effect was observed with the mutant RNAP at lys, I and P promoters, although at these promoters effect of C mediated activation is only at isomerization step during promoter polymerase interaction (Fig 6A, lanes 1-6). Notably, WT and mutant RNAP transcribe with equal efficiency from T7A1 promoter, which is not subjected to C control (Fig 6A, lanes 11, 12). Transcription efficiencies of WT and G524D RNAP at late promoters is quantitated and represented as bar diagram (Fig 6B). RNAP ß' subunit interaction with C, thus seems to be critical for transcription initiation from all the phage Mu late genes.

Discussion
The mechanism of C-mediated activation from P lys , P I and P P appears to be significantly different from P mom . C is not required for RNAP recruitment per se at these promoters unlike at P mom , where RNAP cannot bind the promoter in the absence of C [17,20]. At these promoters, C exerts an effect at the step of open complex formation, not seen with P mom . Faint open complexes observed with these promoters in the absence of C (Fig 4A-4C), seem to be transcriptionally inactive as they are not converted into productive transcripts (Fig 2 and Fig 5). The abortive transcription seen at these promoters in the absence of C may be a consequence of formation of the ternary complexes termed moribund complexes described earlier at other promoters, which do not proceed into a productive elongation phase [27,28].
How these results compare to the activation mechanisms studied earlier? Some of the well studied activators exert different roles at different promoters. CAP, a global regulator of gene expression in prokaryotes, activates transcription from a number of promoters exerting different modes of activation based on the location of its binding site on the DNA and its interaction surface on RNAP [26]. By interacting with α-CTD of RNAP, CAP activates transcription from the lac promoter by enhancing the closed complex formation. However, at the galP1 promoter, it binds with α-NTD and enhances the isomerization of the closed complex to a transcriptionally competent open complex [24,26]. λcI activates its own expression from the λP RM promoter by enhancing the rate of isomerization, K f [29], by contacting the σ-CTD of RNAP [23,26,30]. Simultaneously, it represses transcription from the divergently positioned λP R promoter. λcI also stimulates closed complex formation (K B ) at the P RM promoter with a mutant polymerase Eσ 70 -R596H RNAP, but not in presence of WT RNAP [31]. The GalR protein that regulates the gal operon in E. coli is also shown to exhibit differential effect at the promoters P1 and P2. Its binding to the DNA on the same face of P1 represses P1 transcription, while stimulating transcription from P2 located on the opposite face of DNA [32]. The p4 protein of bacteriophage ϕ29 represses transcription from the early promoter A2c, but activates late gene Transactivator C facilitates productive transcription at lys, I, P promoters. Abortive transcription profile of lys, I, P promoters. Low amounts of abortive transcripts are seen at lys, I, P in the absence of C protein but no productive transcription is observed. In the presence of C, though the abortive transcription is seen, RNAP enters into productive elongation phase and transcripts are synthesized.
promoter A3 [33,34], by interacting with the α subunit of RNAP [33][34][35]. The activator facilitates RNAP recruitment at the A2c promoter but prevents the elongation step by inhibiting promoter clearance [34]. Activation of transcription from A3 promoter by the protein is achieved through stabilization of closed complex [36]. In a majority of these examples, the regulatory protein binding site is at varied distance with respect to promoter location. However, at phage Mu late promoters, the activator binding site is positioned at upstream of the -35 site (Fig 1B). Normally, activators bound next to -35 elements influence the initiation process by contacting α-NTD or σ subunits. However, C does not interact with either of these subunits [18]. In our previous studies, we demonstrated that C interacts with ß' subunit to induce allosteric transitions that facilitated the promoter escape at P mom [18]. The data presented with the mutant RNAP in Fig 6 suggest the importance interaction of C with the ß' subunit at the other three late promoters. Thus, although there are differences in the mechanism of C mediated transactivation between P mom and other late promoters, similarities are seen in RNAP interaction. Studies with G524D RNAP showed defective transcription with all the four late promoters, indicating the importance of functional interaction between ß' subunit and C to facilitate transactivation. On one hand, the functional interaction is needed for promoter escape at G524DrpoCRNAP exhibits defective transcription from other late promoters similar to that observed at P mom . In vitro transcription assays were carried out on phage Mu late promoter constructs (lys,I,P, mom and mutant mom promoter-tin7) and E. coli 70 promoter T7A1. A) Samples were analyzed on 8% denaturing PAGE to assess the productive transcripts and quantified using Muti gauge software. Transcription at each individual promoter with WT RNAP was taken as 100% and the values were normalized. B) Bar diagram depicts quantitative representation of transcription efficiency. G524D RNAP showed reduced productive transcription at all C-dependent promoters compared to WT RNAP but was competent at T7A1 promoter. The results are an average of three independent experiments. doi:10.1371/journal.pone.0129504.g006 second step of activation of P mom while with other promoters, this interaction appears to be required for isomerization step (Fig 7).
Why a given activator would function in a mechanistically different fashion at different promoters? More specifically, how to envision C acting differently at lys, I and P promoters compared to its action at P mom ? Analysis of the late promoter sequences indicates a variation in the architecture of the promoter elements (-10, -35 elements and spacer region) and the differences in C-binding site (Fig 1B). These features might lend to differences in the C interaction at these promoters and also differences in RNAP binding. Further, the effects of an activator at different kinetic steps could depend on when during the initiation process the activator-RNAP interaction occurs [37].
Finally, what is the biological relevance of such a difference in the regulation of the four late promoters? After the switch from middle to late gene transition, the phage has to ensure that the proteins required for morphogenesis are present in sufficient quantities. In contrast, the Mom protein is required in small quantities for a limited DNA modification and its pre-mature or over-expression causes toxicity [8,9]. Thus, elaborate negative regulatory measures and dual step activation of Pmom is a necessity that is not required for other late gene promoters engaged in production of phage structural proteins.

Chemicals and Reagents
NTPs and dNTPs were purchased from Promega. [γ 32 P] ATP and [α 32 P] UTP were purchased from PerkinElmer life sciences. All column materials used for protein purifications were from Differential activation by C at late promoters. Cartoonic representation of the mechanism of C mediated transactivation at late promoters. C activates transcription by a multi-step mechanism at P mom. C acts initially to recruit RNAP to the mom promoter and then facilitates promoter escape. At promoters lys, I and P-C acts at a single step and facilitates open complex formation which proceeds into productive elongation mode. In the absence of activator C, the open complex formed remains transcriptionally inactive. RNAP holoenzyme-magenta, C dimer-green oval, out of phase promoter elements at P mom is represented on DNA as blue and red rectangles respectively. CBS (C-binding site) over lapping -35 element is shown as green rectangle. CBS on P lys ,P I ,P P promoters is depicted as red rectangle. RNAP open complex at lys, I, P promoters is shown as open circle.
doi:10.1371/journal.pone.0129504.g007 GE Health care. Restriction enzymes were from New England Biolabs. The oligonucleotides and other chemicals used were from Sigma-Aldrich. Sequence of various oligonucleotides used in the study is listed in Table 1. 92 base long top and bottom strands comprising the lys, I and P promoter elements with 5 0 -EcoRI and 3 0 -BamHI over-hangs respectively, were designed in such a way that 20nt were present downstream of the transcription start site. The oligonucleotides were either end labeled with [γ 32 P] ATP, annealed and used for electrophoretic mobility shift assays or were annealed and used for cloning between EcoRI-BamHI sites in pUC19.

Plasmids, promoter templates and protein purification
The annealed synthetic oligonucleotides with EcoRI-BamHI over-hangs were cloned into pUC19. Transcription templates were generated by PCR amplification of the late promoter plasmids using pUC forward and reverse primers, followed by gel purification. C protein was purified by following the procedure described earlier [38]. WT and mutant RNAPs were purified following Kashlev et al [39], using Ni-NTA sepharose and heparin-sepharose affinity columns.

RNAP-promoter interaction assays
92-base pair (bp) 5 0 P 32 labeled lys, I, P and 78-bp 5 0 P 32 labeled tin7 promoter fragments (Table 1) were used in assays for both closed and open complex formation. The assays were carried out either in the presence or absence of C protein, using E. coli RNAP. The closed complex formation experiments were carried out as described earlier [17,18,40]. Briefly, the promoter fragments were incubated with increasing concentrations of RNAP on ice for 10 min, electrophoresed on a 3.5% native PAGE at 4°C and visualized by phosphorimager.

In vitro transcription assays
Transcription reactions were carried out on the linear DNA templates of P lys , P I , P P , P mom and P tin7 with WT and mutant RNAP in transcription buffer [40 mM Tris-Cl pH 8.0, 5 mM (CH3COO) 2 Mg, 0.1mM EDTA, 0.1 mM DTT, 100 mM KCl, 100μg/ml BSA]. Reactions were initiated by incubating 40 nM DNA, 80 nM RNAP in transcription buffer on ice for 10 min to allow the formation of closed complex. 300 nM C protein was used wherever required. The reactions were shifted to 37°C for 10 min to allow the formation of open complex. For single round transcriptions, 50μg/ml heparin was added and incubated at 37°C for 1 min. With the addition of 0.3 mM NTP mix and 3μCi [α 32 P] UTP (6000 Ci/m Mol), the reactions were initiated and after 30 min at 37°C, terminated by the addition of urea loading dye (8M Urea, 0.05% bromophenol blue and 0.05% xylene cyanol), heat inactivated at 65°C for 3 min and quenched on ice. The samples were applied either to a 10% denaturing PAGE or 22% denaturing PAGE for analyzing run-off transcripts and abortive initiation respectively.