Promoter Escape with Bacterial Two-component σ Factor Suggests Retention of σ Region Two in the Elongation Complex*

Background: The proposed model for promoter escape predicts the destabilization of interactions of σ region 4 with RNA polymerase and DNA. Results: Using a two-component σ factor, we show that YvrI, mimicking the σ region 4, is released, whereas YvrHa, mimicking σ region 2, is retained after promoter escape. Conclusion: This study validates the proposed mechanism for promoter escape. Significance: This study suggests the possibility of certain σ-factors to be retained in elongation complex. The transition from the formation of the RNA polymerase (RNAP)-promoter open complex step to the productive elongation complex step involves “promoter escape” of RNAP. From the structure of RNAP, a promoter escape model has been proposed that suggests that the interactions between σR4 and RNAP and σR4 and DNA are destabilized upon transition to elongation. This accounts for the reduced affinity of σ to RNAP and stochastic release of σ. However, as the loss of interaction of σR4 with RNAP results in the release of intact σ, assessing this interaction remains challenging to be experimentally verified. Here we study the promoter escape model using a two-component σ factor YvrI and YvrHa from Bacillus subtilis that independently contributes to the functions of σR4 and σR2 in a RNAP-promoter complex. Our results show that YvrI, which mimics σR4, is released gradually as transcription elongation proceeds, whereas YvrHa, which mimics σR2 is retained throughout the elongation complexes. Thus our result validates the proposed model for promoter escape and also suggests that promoter escape involves little or no change in the interaction of σR2 with RNAP.

Transcription is the first step in gene regulation that involves initiation, elongation, and termination. For transcription initiation in bacteria, a factor must bind to RNA polymerase (RNAP) 3 rendering the enzyme capable of promoter recognition and nucleation of DNA prior to the open complex formation (RP o ). Upon addition of NTP, RNAP in the open complex begins synthesis of RNA as an initial transcribing complex and releases short RNA transcripts between 2 and 8 nucleotides. Once the length of the nascent RNA becomes greater than 9 -11 nucleotides, RNAP is able to escape from the promoter and start RNA synthesis productively as a transcription elongation complex (EC). The step between the open complex formation and elongation complex formation is known as "promoter escape." Initially, it was proposed that promoter escape is possible only upon the release of from RNAP (1)(2)(3)(4)(5). The conclusion was based on biochemical assays with Escherichia coli 70 , in which 70 was observed to be present in the open complex, but absent in the elongation complex. This observation lead to the generalized proposal that promoter escape in bacteria involves "-cycle" in which binds to RNAP to initiate transcription and is released from RNAP upon transition from transcription initiation to elongation (1)(2)(3)(4)(5). This obligatory release of 70 for promoter escape was subsequently challenged by several observations that suggested that the release of 70 does not occur immediately upon transition from transcription initiation to elongation, but rather occurs slowly during the course of transcription elongation (6 -12). In contrary to the "-release model," it was further reported that a fraction of RNAP of E. coli in stationary phase does not release 70 throughout the elongation step on some genes (13,14). When RNAP covalently tethered with 70 was expressed in E. coli, there was no adverse effect on the cell growth. Because there was no release of 70 in vivo in these cells, this observation argues that the -cycle may not be essential for bacteria.
The principal factors of bacterial species contain four conserved regions: (R1.1, R2, R3, and R4), all of which interact extensively with ␤ and ␤Ј subunits of RNAP (15)(16)(17). In addition, R4 and R2 are responsible for recognition and interaction with promoter elements Ϫ35 and Ϫ10, respectively. R1.1 is located inside the active center cleft in the RNAP holo but is displaced from the cleft once the open complex is formed (15). The structures of RNAP holoenzyme of different bacterial species show that the R3/R4 linker occupies the "RNA exit channel" that mediates the egress of nascent RNA. This linker encounters a steric clash with the nascent RNA once it reaches a threshold length of ϳ9 -11 nucleotides and tries to enter the RNA exit channel (15)(16)(17). Thus, the transition from transcription initiation to elongation involves the displacement of the R3/R4 linker from the RNA exit channel. From the structure of RNAP, Murakami and Darst (18) proposed a model for the promoter escape, which predicts that displacement of the R3/ R4 linker could destabilize the interactions between R4 and ␤, which in turn, further destabilizes the interactions of R4 and the Ϫ35 element, allowing RNAP to escape from the promoter and start productive elongation. Although the interaction of R2 with RNAP possibly remains intact at this stage, this transition into elongation is responsible for the reduced affinity of for RNAP (19) and results in a stochastic release of .
In Bacillus subtilis -like factors YvrI and coregulator YvrHa activate transcription from a small set of conserved promoters (P oxdC , P yvrI , and P yvrJ ) (20). YvrI includes a region 4 domain that interacts with the ␤ subunit of RNAP and is responsible for recognizing the Ϫ35 element of specific promoters. On the other hand, YvrHa functions as R2 and is responsible for recognizing the Ϫ10 element and DNA melting of the promoter (20 -22). Thus, these two proteins together function as a factor in an RNAP-promoter complex and is referred to as a twocomponent factor. This two-component factor could posit an ideal system to study the proposed promoter escape model that predicts that the interaction between RNAP and YvrI (which mimics R4) may break once nascent RNA reaches a length of 9ϳ11 nucleotides, whereas interaction between YvrHa (which mimics R2) and RNAP may remain intact. Here, we observed that YvrI was released as RNAP proceeds through elongation, whereas YvrHa was fully retained in the elongation complex. Thus, our study validates the predictions of the proposed model for promoter escape and establishes that R2 retains its interaction with RNAP in the elongation phase.

Materials and Methods
Cloning Strategies-The genes encoding YvrI and YvrHa were subcloned from plasmids pSM11 and pSM17, respectively (a kind gift from Dr. Shawn Maclellan), into pET32a with NcoI and BamHI, for the incorporation of the C-terminal histidine tag. DNA fragments containing oxdC and yvrI promoters were amplified from B. subtilis genomic DNA (strain Bs168) using primers ( Table 1). The portions of the rpoD gene encoding regions 2/3.1 and 3.2/4 of 70 were amplified from plasmids pGEMD-396Cys and pGEMD-578Cys (7), respectively (a kind gift from Dr. Richard H. Ebright), using primers (Table 1) and cloned into pET32a with NcoI-EcoRI and NcoI-BamHI, respectively.
Preparation of RNAP Core of B. subtilis-For the purification of B. subtilis (Bs) RNAP core, plasmids pNG545 (containing ␣ and ␤) and pNG540 (containing ␤Ј and ) (a kind gift from Dr. Xiao Yang and Peter J. Lewis) were transformed into E. coli B384 (DE3) cells (23), and grown in 4 liters of LB (with 0.1% dextrose, 100 g/ml of ampicillin, and 35 g/ml of chloramphenicol) at 37°C until the OD reached 0.5. Protein expression was induced by adding 0.5 mM isopropyl 1-thio-␤-D-galactopyranoside and further grown at 16°C until the OD reached 1.2. The cells were harvested, and RNAP was purified essentially as described by Mukhopadhyay et al. (7), except that the pellet obtained from the polyminP precipitation was washed with buffer containing 0.4 M NaCl instead of 0.5 M NaCl.
Purification of YvrI and YvrHa Proteins of B. subtilis-pET32a-yvrI and pET32a-yvrHa were transformed into E. coli BL21(DE3) cells and grown in 2 liters of 2ϫ YT (16 g of tryptone, 10 g of yeast extract, and 5 g of NaCl per liter) with 0.1% dextrose and 100 g/ml of ampicillin at 37°C until the OD reached 0.4. Protein expressions were induced with addition of 0.5 mM isopropyl 1-thio-␤-D-galactopyranoside and the cells were grown for another 3 h. The cells were lysed with TGB (50 mM Tris-HCl, 5% glycerol, 2 mM ␤-mercaptoethanol) and 0.2 M NaCl, 0.25% deoxycholate, 1 mM PMSF, followed by sonication and centrifugation at 14000 rpm for 20 min at 4°C. The pellets were washed with TGB, 0.2 M NaCl, and 0.5% Triton X-100 followed by centrifugation at 8000 rpm for 20 min at 4°C. The pellets were solubilized in buffer A (TGB ϩ 8 M urea) and incubated for 1 h, followed by centrifugation at 14000 rpm at room temperature. The supernatant containing the solubilized protein was passed through 5-ml nickel-nitrilotriacetic acid-agarose, pre-equilibrated with buffer A, washed with 25 ml of buffer A, and eluted with 5 ml each of buffer A containing 10, 20, 40, 80, and 160 mM imidazole. Both proteins were eluted at 40 mM or higher concentration of imidazole. The purified YvrI sample was then dialyzed in 4 liters of TGB buffer for 24 h with two changes at 8-h intervals and the renatured protein was stored with 50% glycerol at Ϫ80°C.
The 5 ml of YvrHa protein sample was renatured by serial dilution with an equal volume of TGB buffer to make a final volume of 80 ml. The sample was applied to a MonoQ HR 10/10 column in Akta Purifier (GE Healthcare), preequilibrated with TGB buffer. The column was washed with 16 ml of TGB buffer and eluted in 1-ml fractions with a linear gradient 0 to 1 M NaCl in 160 ml of TGB buffer. The protein was eluted at 0.4 M NaCl.
The purified protein sample was added with an equal volume of 100% glycerol and stored at Ϫ80°C.
Preparation of E. coli RNAP Core-E. coli RNAP core was purified as in Mukhopadhyay et al. (7).
Purification of E. coli 70 -E. coli 70 (a derivative with a single-Cys residue at position 578) was purified as in Mukhopadhyay et al. (7).
Purification of E. coli 70 -R2 and 70 -R4 -Plasmids pET32a-rpoD-R2 and pET32a-rpoD-R4 containing Cys at positions 396 and 578, respectively, were transformed into E. coli C43 cells and grown in 1 liter of 2ϫ YT containing 100 g/ml of ampicillin. After growing the cells to OD of 0.4 at 37°C, the protein induction was initiated at 16°C with the addition of 0.5 mM isopropyl 1-thio-␤-D-galactopyranoside, and the cells were further grown for 12 h. The cells were lysed and the proteins that appeared in soluble forms were purified essentially following the same method as for YvrI except that protein samples were passed through a nickel-nitrilotriacetic acid column using TGB ϩ 0.2 M NaCl buffer, and no denaturation/renaturation steps were involved.
Labeling of YvrI and YvrHa mutants of B. subtilis and 70 , 70 -R2, and 70 -R4 of E. coli with TMR-6-maleimide-For labeling of YvrI and YvrHa, we generated single Cys derivatives of each protein. Because YvrI does not contain any cysteine, this residue was incorporated at amino acid position 25 (originally Ser) using a site-directed mutagenesis. YvrHa contains two cysteines at amino acid positions 24 and 61. First these cysteines residues were mutated to serine to generate a no-Cys protein derivative. Then Cys was incorporated at amino acid position 15 (originally Thr) of YvrHa using a site-directed mutagenesis kit (Table 1). Both the single Cys YvrI and YvrHa derivatives were purified as their wild type counterparts.
Preparation of 5Ј-Biotinylated Promoter DNA Fragments-The promoter DNA templates ( Table 2) were generated by PCR amplification of either genomic DNA (Bs168) or from synthetic oligonucleotides with their respective 5Ј-biotinylated forward and reverse primers ( Table 1).
B. subtilis RNAP holoenzyme was prepared by incubation of 200 nM unlabeled RNAP core and 500 nM each of TMR-labeled YvrI and YvrHa in transcription buffer (18 mM Tris (pH 8.0), 10 mM NaCl, 8 mM ␤-mercaptoethanol,10 mM MgCl 2 ) for 20 min at room temperature. Open complexes were formed by incubating the RNAP holoenzyme sample with 100 nM 5Ј-biotinylated DNA fragments containing either PoxdC or PyvrI promoter at 37°C for 20 min. 10 l RP o samples were immobilized on 20 l of streptavidin beads pre-equilibrated with the respective transcription buffer, by incubating the samples at 25°C for 1 h with regular tapping. The beads were washed 4 times with the respective transcription buffer; each time the samples were centrifuged at 4000 ϫ g for 1 min.
Formation of Stalled Elongation Complexes-For formation of stalled elongation complex ECϩn, we used the DNA derivative that contained the first non-template strand CTP or UTP at position nϩ1, ( Table 2). After formation of the open complex as above, transcription reactions were initiated with subset of NTPs and 0.25 g/l of heparin as following: ECϩ11 on PoxdC, 500 M ApG, GTP, UTP, ATP; ECϩ20 and ECϩ39 on PoxdC, These complexes were washed thoroughly several times with TB to remove the excess unbound proteins, followed by centrifugation at 4000 ϫ g for 1 min each time. A set of reactions containing RP o and EC are run on 10% SDS-PAGE and scanned in the TMR channel (excitation 530 nm, emission 580 nm) of a fluorescence scanner, (Typhoon Trioϩ, GE Healthcare). The amounts of YvrI, YvrHa, and 70 in each of the complexes were quantitated from the fluorescence intensities of the corresponding protein bands on the fluorescence scanned gel, and the amount of RNAP was quantitated from the intensity of the ␤␤Ј band on the same gel stained with Coomassie Blue. Previously a standard curve with intensity versus concentration was generated for each protein sample by running various amounts of each protein on a 10% SDS-PAGE and measuring the intensity of each band on the gel either scanned by fluorescence on the TMR channel or stained by Coomassie Blue. The fractional occupancies of YvrI, YvrHa, 70 , 70 -R2, and 70 -R4 with respect to RNAP were calculated from the ratio of the amount of the proteins to RNAP.
Determination of the Transcription Efficiency of RP o -To quantitate the amount of transcripts in stalled EC, an in vitro transcription assay was performed with each RP o as above but with 32 P-labeled ATP. The samples were run on 12% urea PAGE and scanned by phosphorimaging (Typhoon Trioϩ, GE Healthcare). The amounts of transcripts were determined from the intensity of band for each stalled EC as described by Mukhopadhyay et al. (7). The amounts of open complexes formed were determined from the quantity of DNA bound to the beads by a fluorescence based assay with SyBr Gold as per the manufacturer's protocol (Life Technologies). The subpopulation of RP o competent to form EC (f) was determined from the ratio of the amount of transcripts to RP o .
Fluorescence Anisotropy Assay-20 nM TMR-labeled YvrI and unlabeled YvrHa in 60 l of TB was titrated with an increasing concentration of RNAP at 37°C and the anisotropy values were monitored with excitation at 540 nm and emission at 580 nm using a PTI Fluorescence master QM400 system fitted with automatic polarizer (25,26). The anisotropy values remained unchanged on the addition of unlabeled YvrHa to labeled YvrI, in the absence of RNAP (considered as anisotropy value (A o ) of free YvrI. When saturation was reached with the addition of around 50 nM RNAP, PoxdC promoter DNA was added at a concentration equal to the saturating level of RNAP core, to form the open complex. The addition of DNA did not change the anisotropy value (considered as A 1 of fully bound YvrI).
Stalled elongation complexes (ECϩ11, ECϩ20, and ECϩ39) were formed by adding appropriate NTP and heparin (as indi-cated above) and incubating with RP o and the anisotropy values of the samples were monitored. The fractions of bound YvrI in RP o and ECs with anisotropy value A (with both free and bound molecules) were determined from the equation: where f is the fraction of free molecules of YvrI.
The subpopulation of RP o competent to form EC was estimated using in vitro transcription assay as above but in solution. The fractional occupancies of YvrI and YvrHa in the stalled elongation complexes were corrected for the subpopulation of RP o competent to form EC using the formula mentioned earlier. A similar set of experiments were repeated with labeled YvrHa (and unlabeled YvrI) to determine the fractional occupancies of this protein in RP o and ECs.

Results
Previously the occupancy of relative to RNAP in the context of the open complex and elongation complex had been determined by separating the complexes from the free components either by gel electrophoresis or chromatography and subsequently analyzing their contents by SDS-PAGE and Coomassie staining. In this report, we use a similar technique that involves immobilization of open complex (RP o ) on streptavidin beads using biotin-labeled promoter DNA fragments. First RNAP core and TMR-labeled or two-component factor were incubated to form RNAP holo before further incubation with promoter DNA fragments to form RP o . The beads were washed several times to remove unbound excess RNAP and . A part of the open complexes were used to form the stalled elongation complexes (ECϩn, where n is the length of RNA) and further washed. To form EC, the promoter DNA fragment derivatives first having UTP or CTP residues at the position nϩ1 base on the non-template strand of the transcribed region (sequences listed in Table 2) were used to form the RP o and added with ATP, GTP, and CTP or UTP, so that RNAP synthesizes RNA of length n nucleotide and halts. Both RP o and ECϩn were resolved on SDS-PAGE gel, scanned with a fluorescence imager at the TMR channel, stained with Coomassie Blue, and scanned further. The amounts of RNAP in the complexes were quantified by comparing the intensity of ␤␤Ј band on the Coomassie-stained gel with the intensity of the same band of RNAP of known quantity run in parallel on the same gel. The amount of labeled YvrI and YvrHa (or ) in the complexes were quantified in the same way by estimating the fluorescence intensity of the band on the fluorescently scanned gel. Because the size of YvrI and YvrHa is small (much less as compared with ␤␤Ј), the protein bands from the above complexes were not visible on the Coomassie-stained gel. Therefore these small proteins were labeled with fluorescent dye for better quantification. The labeled YvrI and YvrHa were active in transcription (results not shown). The fractional occupancy of these factors in RP o and EC were determined from the ratio of the amount of protein to RNAP in the complexes. The fractional occupancy was further corrected for the subpopulation of RP o that was unable to form EC. The amount of RP o was determined from the quantity of DNA bound to the beads as the DNA was the limiting factor in the formation of RP o . The amount of EC was determined from the quantity of radiolabeled transcripts formed with RP o using the in vitro transcription assay identical to the formation of EC as above, but with 32 P-labeled NTP.
Study with E. coli 70 -Because it was previously shown that 70 in E. coli is released upon transition from transcription initiation to elongation, we first tested whether the result was reproducible using the above assay. Two different promoter DNA fragments, lacUV5 and T7A1, were used to form RP o . Subsequently ECϩ23 and ECϩ15 were formed with T7A1 and lacUV5 promoter DNA fragments, respectively (Fig. 1A). Six replicates of a set of RP o and EC were formed for each promoter. Fractional occupancy of 70 in these complexes were estimated for each set. The mean value of fractional occupancy of 70 was found to be 0.61 for the T7A1 DNA fragment and 0.57 for the lacUV5 DNA fragment (Fig. 1B). Fractions of RP o that were competent to form elongation complexes were estimated for each set and the typical values were 0.55 of RP o in T7A1 and 0.60 in lacUV5 (Fig. 1C). After correcting the subpopulation that were competent to undergo the transition from transcription initiation to elongation, the fractional occupancy of 70 was estimated to be 0.19 and 0.13 in T7A1 and lacUV5 promoters, respectively (Fig. 1D). Thus, the majority of 70 was released from the elongation complexes on both promoters in this assay, consistent with previous observation that 70 was released during transition from transcription initiation to elongation using a similar assay.
Study with B. subtilis YvrI and YvrHa-We then tested the fate of two-component factors YvrI and YvrHa of B. subtilis during the transition from transcription initiation to elongation. RP o were formed on the PoxdC promoter DNA fragments with RNAP core, YvrI and YvrHa. Three PoxdC promoter DNA derivatives were used in the assay that permitted formation of ECϩ11, ECϩ20, and ECϩ39, respectively, using a subset of NTPs ( Fig. 2A). The assay was repeated six times and the mean value of the fractional occupancies (Fig. 2B) of YvrI with respect to RNAP were estimated to be 0.85, 0.85, and 0.69 in ECϩ11, ECϩ20, and ECϩ39, respectively. The subpopulation of RP o competent to undergo transition to elongation was estimated to be 0.45 for ECϩ11, 0.36 for ECϩ20, and 0.30 for ECϩ39 (Fig.  2C). After correction, the fractional occupancies of YvrI were estimated to be 0.48 in ECϩ11, 0.16 in ECϩ20, and 0.0 in ECϩ39 (Fig. 2D). On the other hand, the fractional occupancy of YvrHa for all three ECs was estimated to be around 0.99 and 0.97, respectively, before and after the correction. Please note that as the fraction of RP o that was competent to undergo transition to elongation was estimated only from the intensity of stalled EC, not taking into account the intermediate paused EC, and therefore there could be a 10 -20% error in the estimation, and these errors could be incorporated in the estimation of fractional occupancies of YvrI and YvrHa. The data indicate a gradual release of YvrI from RNAP as transcription elongation proceeds, whereas there is no release of YvrHa from RNAP for all the ECs.  o were formed by incubating 200 nM Ec RNAP core, 500 nM each of TMR 70 , and 100 nM biotinylated promoter DNA fragment at 37°C for 20 min, and immobilized on streptavidin beads. ECs were formed by adding NTP (indicated under "Materials and Methods"). The beads containing RP o and EC were washed before resolving on 10% SDS-PAGE, followed by fluorescence scanning and Coomassie staining. Left panel, data for T7A1 promoter: ECϩ23 were formed. Right panel, data for lacUV5 promoter: ECϩ15 were formed. B, fractional occupancies of 70 with respect to RNAP core: determined by quantifying the amount of each protein from the gel. Data were average of six replicates. Left panel, data for T7A1 promoter: ECϩ23 were formed. Right panel: data for lacUV5 promoter: ECϩ15 were formed. C, representative data for in vitro transcription assay with RP o immobilized on streptavidin beads. Stalled elongation complexes were generated using 32 P-labeled NTP; chase, all four NTP were added to EC to produce runoff products. The samples were run on 12% urea PAGE and scanned on a phosphorimager. Left panel, data for T7A1 promoter: ECϩ23 were formed. Right panel, data for the lacUV5 promoter: ECϩ15 were formed. D, fractional occupancies of 70 after correction for the subpopulation that are competent to form EC. Data were average of six replicates. Left panel, data for T7A1 promoter: ECϩ23 were formed. Right panel, data for lacUV5 promoter: ECϩ15 were formed.
Identical results were obtained with these factors using another promoter DNA PyvrI. Two elongation complexes, ECϩ10 and ECϩ19, were formed using the PyvrI promoter derivatives (Fig. 3A). The fractional occupancy of YvrI at ECϩ10 and ECϩ19 was estimated to be 0.79 and 0.72 before correction (Fig. 3C); and 0.45 and 0.11 after correcting for the subpopulation not competent to undergo transition to elongation (fraction of RP o efficient in transcription were 0.39 for ECϩ10 and 0.30 for ECϩ19, Fig. 3D). The fractional occupancy of YvrHa for the above ECs was estimated to be around 1 and 1.1, respectively, before and after the correction.
Study with B. subtilis YvrI and YvrHa Using Anisotropy-In addition to the above method that involved a separation step, we used another complimentary approach involving fluores-cence anisotropy that does not involve any separation step for RP o or EC from the free components. As the fluorescence anisotropy value of a free molecule (A o ) is different from the anisotropy value of the molecule bound to another protein (A 1 ), the fraction of free molecules (f) in a mixture of samples with anisotropy value A (both free and bound molecules) can be monitored from the equation: A ϭ fA o ϩ (1 Ϫ f)A 1 . The anisotropy values of TMR-labeled YvrI were monitored in the presence of an equal amount of unlabeled YvrHa (taken as an anisotropy value of free YvrI, Fig. 4A) before and after titration with RNAP core. Once saturation was reached, the PoxdC promoter DNA fragment was added (at a concentration equal to the saturating level of RNAP core) to form the open complex. After addition of DNA, the anisotropy value of the complex did . The beads containing RP o and EC were washed before resolving on 10% SDS-PAGE, followed by fluorescence scanning and Coomassie staining. B, fractional occupancy of YvrI and YvrHa with respect to the RNAP core: determined by quantifying the amount of each protein from the gel. Data were average of six replicates. Gray bar, YvrI; black bar, YvrHa. C, representative data for in vitro transcription assay with RP o immobilized on streptavidin beads. Stalled elongation complexes ECϩ11, ECϩ20, and ECϩ39 were generated using 32 P-labeled NTP; chase, all four NTP were added to EC to produce runoff products. The samples were run on 12% urea PAGE and scanned on a phosphorimager. D, fractional occupancies of YvrI and YvrHa after correction for the subpopulation that are competent to form EC. Data were average of six replicates. Gray bar, YvrI; black bar, YvrHa.  Fig. 2A but with the PyvrI promoter. B, fractional occupancy of YvrI and YvrHa with respect to the RNAP core: determined by quantifying the amount of each protein from gel. Data were the average of six replicates. Gray bar, YvrI; black bar, YvrHa. C, representative data for the in vitro transcription assay with RP o immobilized on streptavidin beads. Stalled elongation complexes ECϩ10 and ECϩ19 were generated using 32 P-labeled NTP; chase, all four NTP were added to EC to produce runoff products. The samples were run on 12% urea PAGE and scanned on a phosphorimager. D, fractional occupancies of YvrI and YvrHa after correction for the subpopulations that are competent to form EC. Data were average of six replicates. Gray bar, YvrI; black bar, YvrHa.
not change and was taken as the anisotropy value of bound YvrI (Fig. 4A). Three sets of RP o and ECs (ECϩ11, ECϩ20, and ECϩ39) were formed, and the anisotropy values were monitored. Each set was repeated six times and the mean values of fractional occupancies of YvrI in ECs were estimated to be 0.75 for ECϩ11, 0.65 for ECϩ20, and 0.52 for ECϩ39 (Fig. 4C). The competent subpopulation of RP o efficient in forming ECs in solution was determined to be 0.42 for ECϩ11, 0.36 for ECϩ20, and 0.30 for ECϩ39 (Fig. 4D), which in turn, determined the corrected fractional occupancy of ECϩ11, ECϩ20, and ECϩ39, respectively, as 0.52, 0.22, and 0 (Fig. 4E). Similar assays were performed with TMR-labeled YvrHa and unlabeled YvrI (Fig.  4B), in which the fractional occupancy of YvrHa remained around 0.98 for all three ECs. Observation of the fluorescence anisotropy assay was consistent with the previous assay.
Study with E. coli 70 -R2 and 70 -R4 -To test whether the promoter escape study with YvRI and YvrHa is valid for any other two-component factor, we prepared two truncated derivatives of 70 , comprising amino acids 130 -500 of the region 2/3.1 ( 70 -R2) and amino acids 501-613 of the region 3.2/4 ( 70 -R4), respectively. Interestingly, these two truncated derivatives of 70 are able to initiate transcription from the 70 -dependent promoter. Each protein derivative contained a single Cys residue and was labeled with TMR. The promoter escape study with these two 70 derivatives was performed essentially as with YvrI and YvrHa of B. subtilis except that the lacUV5 promoter derivatives and E. coli RNAP core were used to form RP o and ECs (Fig. 5A). The mean value of the fractional occupancies of 70 -R4 with respect to RNAP was estimated to be 0.67, 0.56, and 0.49 in ECϩ15, ECϩ25, and ECϩ40, respectively (Fig. 5B). The subpopulations of RP o competent to undergo the transition to elongation were estimated to be 0.60 for ECϩ15, 0.52 for ECϩ25, and 0.45 for ECϩ40 (Fig. 5C). After correction, the fractional occupancies of 70 -R4 were estimated to be 0.45 in ECϩ15, 0.15 in ECϩ25, and 0.0 in ECϩ40 (Fig.  2D). On the other hand, the fractional occupancies of 70 -R2 for all three ECs were estimated to be around 0.99 and 0.98, respectively, before and after the correction.

Discussion
Our results show that the amount of YvrI relative to RNAP is gradually decreased in the elongation complexes as elongation proceeds, whereas the amount of YvrHa relative to RNAP remains constant in the elongation complexes (Fig. 6). Because a certain amount of YvrI is present in the early elongation complexes (fractional occupancy is 0.48 at ECϩ11), our result is consistent with previous observations that shows a significant fraction of 70 remains associated with RNAP in the early elongation complex (6,12). The result (fractional occupancy of YvrI are 0.48, 0.16, and 0, respectively for ECϩ11, ECϩ20, and ECϩ39) is also consistent with the observation that demonstrated the stochastic release of in transcription elongation (11). The RNAP structure-based proposed model for promoter escape predicts that, upon transition from transcription initiation to elongation, the interaction of R4 with RNAP is destabilized upon transition to elongation reducing the overall affinity of to RNAP and this accounts for stochastic release of . However, as the destabilization of interactions of R4 with RNAP in EC results in the release of intact 70 , assessing the overall interaction of this region with RNAP during the promoter escape is not possible using a simple biochemical assay. In principle, assessing the above interaction could be possible using a FRET assay involving fluorescently labeled RNAP and derivatives. However, this assay could be technically challenging because of the difficulty in labeling RNAP at specific sites. Thus validation of the proposed model for promoter escape remains a difficult task. The two-component factor YvrI and YvrHa individually contribute to the functions of R4 and R2 to an RNAP-promoter complex. Using simple biochemical and biophysical assays we show that YvrI (that mimics R4) is released from EC due to the loss of its interaction with RNAP, Stalled elongation complexes are generated by adding 500 M each of ApG, heparin, and other NTPs at ϩ11 (without CTP), ϩ20 (without UTP), and ϩ39 (without UTP). The mean anisotropy values of RP o , ECϩ11, ECϩ20, and ECϩ39 were 0.297, 0.26, 0.24, and 0.22, respectively, for YvrI; 0.104, 0.102, 0.1, and 0.1, respectively, for YvrHa. The fractional occupancies of YvrI and YvrHa were determined by estimating the amount of bound proteins in RP o and EC from the anisotropy values of the protein in the respective complexes. Gray bar, YvrI; black bar, YvrHa. D, representative data for in vitro transcription assay in solution. Stalled elongation complexes ECϩ11, ECϩ20, and ECϩ39 were generated using 32 P-labeled NTP; chase, all four NTP were added to EC to produce runoff products. The samples were run on 12% urea PAGE and scanned on a phosphorimager. E, fractional occupancies of YvrI and YvrHa after correction for the subpopulations that are competent to form EC. Data were average of six replicates. Gray bar, YvrI; black bar, YvrHa.
whereas YvrHa (that mimics R2) is retained in the EC. Similar results were obtained with the truncated 70 derivatives: 70 -R4 is released, whereas 70 -R2 is retained in the EC. Therefore, our results validate the proposed model for promoter escape in bacteria. As YvrHa is not covalently attached to YvrI, the release of YvrI does not alter the interaction of YvrHa with RNAP in EC. This further suggests that there is no or little change in the interaction between R2 with RNAP upon transition to elongation. But this interaction is not strong enough to retain the whole 70 in the elongation complex as the rest of regions, R1.1, R2, and R3/4 linker, lose their interaction with RNAP. However, the observation raises the possibility that the R2-RNAP interaction of a certain factor from other bacterial species could be strong enough to retain the factor in the EC throughout the elongation phase.
Author Contributions-J. M. conceived and designed the experiments; S. S. performed the experiments; S. S. preformed PAGE analysis, in vitro transcription assays, protein labeling, and fluorescence anisotropy assays, R. K. P. purified the B. subtilis RNAP; S. S. analyzed the data; and S. S. and J. M. wrote the paper.