Mechanisms of Very Long Abortive Transcript Release during

Nov 27, 2015 - Biology Department, Bryn Mawr College, Bryn Mawr, Pennsylvania 19010, ... Program in Biochemistry, Mount Holyoke College, South Hadley,...
0 downloads 0 Views 7MB Size
Subscriber access provided by Penn State | University Libraries

Article

Mechanisms of Very Long Abortive Transcript (VLAT) Release during Promoter Escape Monica Chander, Ahri Lee, Tenaya K. Vallery, Mya Thandar, Yunnan Jiang, and Lilian Ming-Te Hsu Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.5b00712 • Publication Date (Web): 27 Nov 2015 Downloaded from http://pubs.acs.org on December 6, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biochemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

MECHANISMS OF VERY LONG ABORTIVE TRANSCRIPT (VLAT) RELEASE DURING PROMOTER ESCAPE

Monica Chander1, Ahri Lee2, Tenaya K. Vallery2, Mya Thandar2, Yunnan Jiang2, and Lilian M. Hsu2, * 1

Biology Department, Bryn Mawr College, Bryn Mawr, PA 19010

2

Program in Biochemistry, Mount Holyoke College, South Hadley, MA 01075

Corresponding Author To whom correspondence should be addressed. Tel: 413-549-5346; Email: [email protected]

Funding Sources This work was supported by the National Science Foundation through the Research for Undergraduate Institution Program [MCB0418316 and MCB0841452 to L.M.H.]

ACS Paragon Plus Environment

1

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 44

ABBREVIATIONS αCTD: alpha subunit C-terminal domain; AP: abortive probability; β-ME: β-mercaptoethanol; FL: full-length transcript; ITC: initial transcribing complex; ITS: initial transcribed sequence; OC, open complex; RIF: relative initiation frequency; RNAP: RNA polymerase; TIC: transcription initiation complex; VLATs: very long abortive transcripts

ACS Paragon Plus Environment

2

Page 3 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

ABSTRACT A phage T5 N25 promoter variant, DG203, undergoes the escape transition at the +16 to +19 positions after transcription initiation. By specifically examining the abortive activity of initial transcribing complex positioned at +19 (ITC19), we observe the production of both GreBsensitive and GreB-resistant VLAT19. This suggests that ITC19, which is perched on the brink of escape, is highly unstable and can achieve stabilization through either backtracking or forward translocation. Of the forward-tracked fraction, only a small percentage escapes normally (followed by stepwise elongation) to produce full-length RNA; the rest presumably hypertranslocates to release GreB-resistant VLATs. VLAT formation is dependent not only on consensus -35/-10 promoters with 17-bp spacing, but also on sequence characteristics of the spacer DNA. Analysis of DG203 promoter variants containing different spacer sequences reveals that AT-rich spacers intrinsically elevate VLAT formation. The AT-rich spacer of DG203 joined to the -10 box presents an UP element sequence capable of interacting with the polymerase α subunit C-terminal domain (αCTD) during the escape transition, which in turn enhances VLAT release. Utilization of the spacer/-10 region UP element by αCTD subunits requires a 10-15 bp hyper translocation. We document the physical occurrence of hyper forward translocation using ExoIII footprinting analysis.

KEYWORDS Promoter escape, GreB-resistant abortive transcripts, spacer sequence, spacer/-10 region UP element, hyper forward translocation, ExoIII footprinting of initial transcribing complexes

ACS Paragon Plus Environment

3

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 44

During transcription initiation at promoters that are rate-limited at escape, RNAP binds the promoter DNA to form an open complex, and after the abortive synthesis of numerous short transcripts, relinquishes promoter DNA contacts and moves downstream to continue with elongation thus accomplishing the process known as promoter escape. The T5 phage N25 promoter is a classic example where abortive cycling results in a ladder of 2-11 nt abortive transcripts until escape is achieved at position +12 (1). To escape from this promoter, RNAP must scrunch 9 bp of initially transcribed DNA (2) to generate sufficient stress energy to disrupt the upstream polymerase-DNA contacts, subsequently allowing for the collapse of the open complex bubble and maintenance of the scrunched initial transcribed region as the first elongation complex bubble (3). Promoter escape, therefore, elicits a saltatory movement of RNA polymerase during which the transcription bubble translocates from the open complex region to the elongation complex positions (2, 4). The extent of abortive cycling and the efficiency of promoter escape are reciprocally dependent in part on how well RNAP is anchored by the promoter. In general, the tighter the association of RNAP with promoter DNA, the lower the escape efficiency and the higher the extent of abortive cycling (5,6). A second factor governing these processes is the initial transcribed sequence (ITS) originally defined as spanning positions +1 to +20 (7). Randomization of the N25 ITS resulted in altered promoter escape properties (3); notably, all variants now released abortive RNAs until +15 and only escaped at position +16. The beststudied example of such an ITS variant is N25anti (1,3,7,8). A subset of the variants, with changes in ITS positions +3 to +10, further produced VLATs of 16-19 nts (3,9). Aside from their exceptional length, the VLATs are unusual in that their levels are not diminished in reactions supplemented with GreB (9), which has been shown to rescue nascent transcripts ≤15 nt on backtracked ITCs (3,8). During backtracking, the RNA 3’ end is extruded into the secondary channel where a bound GreB stimulates the hydrolytic activity of RNA polymerase catalytic center, enabling cleavage and re-elongation of the RNA 5’ piece (10,11). Failure of GreBmediated cleavage-rescue implies that GreB-resistant VLATs are unlikely to have arisen through RNAP backtracking. How, then, are GreB-resistant VLATs formed? We previously proposed that VLATs are produced by ITCs that have not undergone escape by the +15/+16 juncture, and the polymerase

ACS Paragon Plus Environment

4

Page 5 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

must transcribe and scrunch in a few more nucleotides to bring about the initiation-elongation transition (9). The cause of the delayed escape may be related to the observation that VLATs were mainly produced by N25 promoter variants containing a GC-rich ITS from +3 to +10 (9). This region must remain open as the initial elongation complex bubble when escape occurs. It is likely that more energy is required to maintain an initial elongation bubble made up of Gs and Cs than one composed of As and Ts (9). RNAP acquires this additional energy by continued scrunching up to positions +16 to +20; only then can escape occur. During such an escape process, those RNAP molecules that translocate stepwise downstream can go on to produce fulllength RNA. However, if RNAP is propelled forward by >1 bp, this would result in hypertranslocation and transcriptional arrest (12). The resulting transcripts, instead of being lodged in the RNAP secondary channel due to backtracking (10), would be bound further up-field in the RNA exit channel, making them impervious to the effects of GreB. This “hyper forward translocation” model does not rule out VLAT formation via backtracking. In this report, we provide evidence that a large fraction of VLAT19 is in fact GreB-sensitive and arises when RNAP backtracks at the escape transition. The most prolific VLAT-releasing N25 promoter variant is DG203 (9). By mutagenizing individual promoter elements, we had shown that a primary determinant of VLAT synthesis/release is the near consensus arrangement of the –35 and –10 hexamers with 17-bp spacing which enables the formation of a highly stable open complex structure (9). Within this consensus arrangement, the composition or sequence characteristic of the 17-bp spacer had not been defined, prompting us to further probe the polymerase-promoter contacts that enhance VLAT formation. Here, we report that the interactions between RNAP sigma subunit (σ) and the -35/-10 elements can vary significantly with the composition of the spacer element, and that ATrich spacers intrinsically elevate the levels of VLATs. Additionally, we found that the αCTD of RNAP plays a role in VLAT formation, not by its interaction with the canonical proximal UP element (9), but through its binding with an UP-like element (13) embedded in the AT-rich spacer/-10 region of the DG203 promoter. The latter interaction is captured only with a 10-15 bp hyper forward translocation by the polymerase right after promoter escape at +19. Evidences in support of the hyper forward translocation model were derived from EcoRI-mediated roadblock transcription and ExoIII footprinting analyses.

ACS Paragon Plus Environment

5

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 44

MATERIAL AND METHODS Enzymes and proteins. Native RNAP holoenzyme was purified from E. coli strain RL721 (gift from R. Landick) and contained ~60% active molecules at the time of use (14). Wild type and αCTD∆235 RNAP holoenzymes, both reconstituted as previously described without the ω subunit (15,16), were generous gifts from W. Ross and R. L. Gourse and contained ~30% active molecules at the time of use (17). The E. coli GreB protein was isolated from IPTG-induced JM109 cells containing plasmid pGF296 (18). N-terminally His6-tagged E111Q-EcoRI protein was prepared from IPTG-induced T7 Express Competent E. coli cells (New England Biolabs) containing the plasmid pVS9 according to the protocol provided by I. Artsimovitch. Plasmid pVS9 was created by cloning the E111Q-EcoRI gene, PCR-amplified from the genomic clone from Modrich (19), into pET33 vector (Novagen). All other enzymes were obtained from New England BioLabs® Inc. DNA primers and promoter templates. All primers were obtained from IDT, Inc. The N25 and N25anti promoters were prepared by PCR-amplification from plasmids pKK-N25 and pKKN25anti using primers N25-u (XE) and N25-d (PH) and spans -85 to +57 as described (3). All other promoter templates were generated by extending a pair of overlapping primers with Klenow Pol I. The set of spacer variant promoters—DG203, SP fullcon, SP Pcon and SP mar— was constructed from individual upstream primers (spanning -60 to +20) that overlap with a common downstream primer (203T; spanning +57 to +3) at +3 to +20. The upstream primers for generating DG203-U19 and DG203/SPfullcon-U19 were respectively N25-u (SSX1) and N25u/SPfullcon (SSX1) (spanning -75 to -1) and overlap their respective downstream primers, DG203U19-d and DG203/SPfullconU19-d (spanning +57 to -26), at -1 to -26. The sequences of above primers are listed in Table 1. The EQ-series of DG203 promoters, each containing an EcoRI site in the transcribed region, were constructed through primer extension. Each EQ promoter is designated with a number corresponding to the first-base position of the EcoRI sequence (e.g. EQ23, EQ28, EQ32, EQ33, EQ34, EQ35, EQ36, EQ37, EQ38, EQ41, and EQ43). The sequences of EQ primers are listed in Table 2. The VL-primers (Table 3) were used to prepare DNA templates by primer extension for footprinting the OCs and ITCs of N25, N25anti, DG203 and SP fullcon promoters. To obtain high quality single-end-labeled DNA for

ACS Paragon Plus Environment

6

Page 7 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

footprinting experiments, the VL fragments were first cloned into the pSA508 vector (20) so that the labeling procedure can be initiated from the plasmid DNA format. Transcription, PAGE analysis, and quantification. Steady state transcription reactions were performed and analyzed as described (3). Each promoter (30 nM) was transcribed for 10 min at 37 °C with 50 nM RNAP (~60% active molecules) in 200 mM KCl, transcription buffer (50 mM Tris-HCl, pH 8, 10 mM MgCl2, 10 mM β-ME, 10 µg/mL acetylated BSA), and 100 µM NTP containing either [γ-32P]-ATP or [α-32P]-UTP label at a specific activity of ~10 cpm/fmol. When supplemented, GreB was added along with RNAP at a 10:1 (GreB:RNAP) molar ratio. Transcription reactions performed with reconstituted RNAPs (~30% active molecules) required 150 nM of the enzymes and 30 minutes of reaction at 37 oC to achieve the same level of activity as the native RNAP. The reaction conditions used with the EQ-series templates were devised to minimize the residual EcoRI cleavage activity and optimize the binding of E111Q-EcoRI protein. A set of three reactions was conducted for each EQ template: 1. with RNAP only; 2. with RNAP and EcoRI; and 3. RNAP+GreB and EcoRI. The 10-µL reaction contained 30 nM DNA, 200 mM KCl, 1x new transcription buffer (50 mM Tris-HCl, pH 7.2, 5 mM MgCl2, 10 mM β-ME and 10 µg/mL acetylated BSA), RNAP at 60 nM (active RNAP: DNA = 1:1) without or with GreB (GreB:RNAP = 10:1) and without or with E111Q-EcoRI (EcoRI:DNA = 30:1), and 100 µM NTP mixture containing [γ-32P] ATP at ~10 cpm/fmol. All reactions were carried out at 37 oC and set up as follows: first, forming OCs of DNA and RNAP (without or with GreB) with a 3-min incubation; next, binding E111Q-EcoRI protein (or buffer, for minus-EcoRI control) with a 3-min incubation; finally, adding the NTP mixture to initiate transcription and the reaction proceeded for 10 min. After terminating the reaction, the transcripts were recovered and analyzed by 23% (10:1) denaturing PAGE as described (21). Gels were visualized on a Storm 820 phosphorimager and quantitated using the ImageQuant software (21). Each reaction was repeated independently at least three times to ensure reproducibility. Generating single end-labeled fragment for footprinting. Plasmid clones containing the N25, N25anti, DG203 or SP fullcon promoter in the pSA508 vector context were constructed for footprinting analysis. To that end, the N25 and N25anti promoters were cloned with a flanking XhoI site at –97 and EcoRV site at +61; the DG203 and SP fullcon promoters were flanked by an Eco53kI site at –97 and Acc65I site at +61. To prepare single end-labeled fragment, plasmid

ACS Paragon Plus Environment

7

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 44

DNA was linearized with the first restriction enzyme, treated with alkaline phosphatase, phosphorylated with [γ-32P]-ATP and T4 polynucleotide kinase, and finally excised out of the plasmid by digestion with the second restriction enzyme. All single 5’-end-labeled fragments were 158 bp long spanning –97 to +61. To obtain the pure promoter fragment, the restriction digest was subjected to a 6% (19:1) native PAGE fractionation. The gel pieces containing the labeled promoter DNA were excised and eluted with a low-salt buffer (20 mM Tris-HCl, pH 7.4, 0.2 M NaCl, 1 mM EDTA). The supernatant was treated with phenol-chloroform extraction, and the DNA was concentrated through NaAc-EtOH precipitation. A successful round of labeling gave rise to labeled DNA with specific activities of 2000 – 5000 cpm/fmol. Maxam-Gilbert sequencing reaction. For footprinting analysis, a Maxam-Gilbert ladder of purines was prepared for each promoter template by reacting 200,000 cpm of the labeled DNA and 1 µg of sonicated salmon sperm DNA with 4% formic acid for 15 min at 20 oC (22). The modified DNA was cleaved in 10% piperidine by a 30-min incubation at 90 oC and recovered through ethanol precipitation with 0.3 M NaAc. ExoIII Footprinting.

ExoIII treatment allows us to probe the upstream and downstream

boundaries of the stationary OCs and the dynamic ITC ensemble. A typical reaction (10 µL) contains 200,000 cpm of labeled DNA, supplemented with the same unlabeled DNA to a final concentration of 20 nM, 1x transcription buffer (50 mM Tris-HCl, pH 8.0, 10 mM MgCl2, 10 mM β-ME, 10 µg/mL acetylated BSA), 50 mM KCl, 100 nM RNAP, 10 µg/mL heparin, without or with 100 µM NTP, and without or with ExoIII (added to a final concentration of 10,000 units/mL). The large RNAP:DNA ratio ensures that ExoIII footprints are left on polymeraseDNA complexes, not naked DNA. The concentration of ExoIII used was previously found effective in reaching the protein boundaries situated 40-50 bp upstream of the DNA 3’ end by the 1-min point. All reactions were carried out at 37 oC and set up as follows: first, forming the OCs with a 10-min incubation; next, mixing in heparin for 30-sec to select the open complexes; finally, adding ExoIII (without or with NTP). For mapping the boundaries of an OC, the limit ExoIII reaction proceeded for 3 min. For monitoring the changing boundaries of ITCs, reaction aliquots were withdrawn at 0.5, 1, 1.5, and 3 min, and terminated by mixing with a stop solution (10 mM Na2EDTA, 1 mg/mL nuclease-free glycogen, 0.3 M NaAc) for EtOH precipitation.

ACS Paragon Plus Environment

8

Page 9 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Modified DNA was directly fractionated in 6% (19:1) denaturing PAGE and visualized through phosphorimaging.

RESULTS GreB treatment of 3’-labeled VLATs reveals their formation via RNAP backtracking. Our original experiments documenting the existence of GreB-resistant VLATs monitored the fate of 5’-labeled RNAs (3,9). Such a design does not rule out the possibility that VLATs can also arise from RNAP backtracking. To clarify this issue, we modified the 57-nt DG203 transcript sequence to contain only two U residues, at +2 and +19, yielding the DG203-U19 template (Fig. 1A). Transcription of this template in the presence of NTP/[α-32P]-UTP led to (pseudo) 5’-end labeling of all abortive RNAs ≥2 nt at U2 and 3’-end labeling of 19- to 21-nt VLATs at U19 (Fig. 1A). This dual labeling strategy allows us to track, in parallel reactions without and with GreB, the ATP-initiated abortive RNA ladder and the 3’-RNA cleaved from backtracked ITC19-21 (Figs. 1A and 1B). The gel mobility of the triphosphorylated abortive RNA bands is necessary for referencing the emergence of 3’-cleaved RNAs which, being 5’monophosphorylated (and designated as pN where N indicates length in nucleotides), would migrate slower than their 5’-triphosphorylated counterparts of the same length and appear between the rungs of the abortive RNA ladder (3). Fig. 1B shows the results of steady-state transcription reactions of DG203-U19 without or with GreB up to 60 min. As noted previously, the persistence and continual accumulation of 17to 19-nt VLATs in the presence of GreB suggests that they are different from backtracked RNAs (9). However, we also observed a significant amount of 3’-cleaved RNAs that are mostly 2-3 nt in length but can range up to ~14 nt, especially in reactions supplemented with GreB (Fig. 1B). Since labeled cleavage products can only arise from 3’-cleavage of 19- to 21-nt RNA (labeled at U19), we conclude that ITCs19-21 must undergo extensive backtracking.

ACS Paragon Plus Environment

9

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 44

Figure 1. ITC19 undergoes both backtracking and forward translocation to release VLAT19. A. 5’- and 3’-labeling of DG203-U19 transcripts with [α-32P]-UTP and predicted RNA cleavage pattern. DG203 ITS was modified to DG203-U19 by replacing all U-residues

ACS Paragon Plus Environment

10

Page 11 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

beyond U19. The replaced residues are underlined, while U2 and U19 positions are highlighted in red. Pseudo 5’-end labeling at U2 allows the tracking of ATP-initiated abortive ladder of 2-19 nt whose gel mobility in turn references the position of 3’-cleavage RNAs. Labeling the 3’-end at U19 allows direct monitoring of the 3’-termini of VLATs 19-21 nt long. Here, the high level of 2-nt cleaved RNA, p2, reflects the most common backtracked form of 19- and 20-nt RNA (pink arrows). In the presence of GreB, the limit 5’-RNA whose length represents the shortest stable RNA-DNA hybrid is 4-nt long (green arrow). B. In vitro transcription gel profile of DG203-U19 and DG203/SPfullcon-U19. Each promoter was transcribed for 7, 20 or 60 min at 37 °C with [α32

P]-UTP label. GreB, when present (indicated by “+”), was in 10-fold molar excess over RNAP.

The abortive ladder is shown on the left border. FL: 57 nt. The 3’-cleavage products are indicated along the right borders by asterisks. Being 5’-monophosphorylated, they migrate slower in the gel than their 5’-triphosphorylated abortive RNAs of the same size and end up between rungs of the triphosphorylated abortive RNA ladder (3). p2 indicates the 2-nt 3’cleavage products; p14, the presumptive 14-nt cleavage RNA. The numbers between the two gels align the identical abortive RNA ladder from these promoters. We performed a similar analysis with the DG203/SPfullcon-U19 template created from DG203-U19 by substituting the spacer region with SP fullcon (23, 24; see Fig. 2A). This promoter produces little, if any, full-length (FL) RNA, but overproduces (compared to DG203U19) GreB-resistant VLATs and 3’-cleaved RNAs (Fig. 1B). This suggests that all ITCs that reach the 19-21 nt stage on this promoter must either backtrack or hyper forward track and become arrested. The arrested transcripts are then released as VLATs. To estimate the amount of VLATs produced by RNAP backtracking versus forward-tracking during promoter escape, we summed for the 60-min reaction aliquot, the cleavage products from p2 to p14 as the GreB-sensitive fraction accumulated through repeated backtracking and cleavage-elongation (Fig. 1), and the 19- to 21-nt VLATs plus FL RNA as the GreB-resistant forward translocating fraction. Table 4 shows that, for the DG203-U19 promoter, RNAP backtracking contributed 81% of the RNA 19 nt or longer while RNAP forward translocation generated 19%. Of the forward-tracked fraction, only 2% are FL RNA produced by complexes that have undergone normal escape followed by stepwise elongation; the rest release their RNA that accumulates as GreB-resistant VLATs, likely because hyper forward translocation (of >1

ACS Paragon Plus Environment

11

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 44

bp) has occurred. In comparison, DG203/SPfullcon-U19 produces little or no FL RNA. All VLATs (19-21 nt) are derived from either backtracking (45%) or hyper forward translocation (55%). This experiment demonstrates that the extent to which ITC19-21 backtracks or forwardtracks is dependent in part on the spacer sequence in the core promoter. Promoters with AT-rich spacers display enhanced VLAT formation. We previously showed that promoter elements most crucial for VLAT formation are the -35 and -10 consensus hexamers separated by 17 bp of DNA (9). The 17-bp spacer optimally positions the -35 and -10 elements for simultaneous contacts with σ domains 4 and 2, respectively (24). Changes in spacer length compromise these interactions and influence RNAP binding and transcriptional efficiency (6,25-27). In the experiment below, we examined if the sequence composition of the spacer also influences VLAT production. This was of interest due to the unusually high A+T content (12/17) of the DG203 spacer (Fig. 2A). To test if the AT-rich 17-bp spacer impedes promoter escape on DG203, we constructed three variant promoters—SP fullcon, SP Pcon, and SP mar—that differ only in composition of the spacer region (Fig. 2A). The spacer sequence in SP fullcon was chosen from a group of in vitro-selected full-consensus promoters whose spacer region, although lacking a specific consensus, is overall AT-rich—composed of 12 centrally located A and T residues—and shows strong sequence preference at the 5’ and 3’ borders to the -35 and -10 elements, respectively (23). The specific fullcon promoter containing this spacer sequence forms a highly stable upstream promoter open complex (24) and supports relatively weak productive transcription due to inefficient escape (23). SP Pcon contains the spacer sequence from another artificial consensus promoter, Pcon (28); it is moderately AT-rich (10/17) with a T-tract located at the 5’ border. Pcon was also shown to support weak productive transcriptional activity (28). The third variant, SP mar, contains the spacer element from the native E. coli mar promoter (29) and is the most dissimilar among the four promoters with 9 ATs interspersed with 8 GCs (Fig. 2A). The four constructs were transcribed under steady state conditions with RNAP and the RNA products analyzed on a sequencing gel (Fig. 2B). Since all four templates have the same ITS sequence (i.e. that of DG203), the abortive ladder and FL RNAs produced were identical. However, several striking differences were immediately apparent. One was the total level of

ACS Paragon Plus Environment

12

Page 13 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 2. AT-rich spacer elements enhance VLAT production. A. The promoter sequence of DG203 extending from –60 to +57 is shown in the top panel. The UP element which shows a 7of 9-base agreement to the proximal subsite consensus (33) is centered at -42 and underlined. The -35 and -10 core hexamers are indicated in bold, while +1 marks the start site of transcription. The bottom panel compares the spacer element (italicized) of DG203 with those in SP fullcon, SP Pcon and SP mar. The number of A + T residues within each spacer is indicated in parenthesis. B. Gel profile of in vitro transcription of DG203/spacer variants. Transcription reactions were carried out with [γ-32P]-ATP label. The abortive ladder is indicated on the left

ACS Paragon Plus Environment

13

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 44

border with a letter-number combination representing the identity of the 3’ most nucleotide and the length of the abortive RNA. FL RNA is 57 nt. RIF (relative initiation frequency) indicates the total initiation obtained in the 10-min reaction period from each promoter and is normalized to that of DG203. PY (productive yield) indicates the full-length transcript as a percentage of total transcription. C. Abortive probability profiles. The probability of aborting at each ITS position was calculated (30) and plotted for each of the four promoters. The higher the abortive probability, the greater likelihood the nascent RNA will be released by RNAP at that position. Note the high abortive probabilities at VLAT positions shown by SP fullcon compared to the very low abortive probabilities shown by SP mar. transcript initiation during the reaction period. Here, we compared relative initiation frequency (RIF) at the different promoters by setting the level of DG203 to 1. The RIF values associated with SP fullcon, SP Pcon and SP mar were 1.2, 0.63, and 0.42, respectively. Although SP mar was the least active of the four promoter variants in terms of initiation frequency, it was the most escape-facile, producing 11% FL RNA, while the constructs with AT-rich spacers were less efficient giving productive yields between 1% (SP fullcon) and 3% (DG203 and SP Pcon) (Fig. 2B). Thus, the spacer sequence composition does influence promoter activity; while AT-rich spacers enhance overall initiation frequency, they concurrently impede escape. A second striking difference between the four promoters was the level of long abortive RNAs produced by each. By comparing their abortive probability (AP) profile (30), we found that all four promoters abort with similar probabilities at positions +2 to +12 (Fig. 2C). Differences become apparent only at positions +13 and beyond, and are particularly pronounced at VLAT positions +17 to +19. SP mar, the most escape-facile, displayed very low AP values from +9 on, indicating that escape is mostly complete at +8. In contrast, SP fullcon is the most escapeimpaired construct and shows very high AP values particularly at VLAT positions +17 to +19. DG203 and SP Pcon display similar AP profiles and abort with moderately high tendencies at the VLAT positions. This analysis indicates that AT-rich spacers affect promoter escape at positions +13 to +19, with the most dramatic effects observed at +17 to +19. αCTD interacts with an UP element-like sequence in the DG203 spacer region during promoter escape to release VLATs.

ACS Paragon Plus Environment

14

Page 15 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

In the promoter context, AT-rich stretches have been shown to bend/activate DNA intrinsically (31,32), or serve as sequence-specific binding sites for RNAP subunits, e.g., σ2 with the consensus -10 box, or αCTD with the AT-rich UP element (12,16). The DG203 core promoter contains three AT-rich stretches—the -10 box, the spacer region, and the proximal UP-element at -42. We previously showed that the canonically-positioned proximal UP element at -42 has little or no effect on VLAT production (9). This observation, however, does not preclude the possibility that αCTD may interact with parts of the AT-rich spacer region and/or the -10 box sequence and influence VLAT formation as RNAP moves downstream during the escape transition. To determine if VLAT formation involves α subunit interaction with the spacer DNA, we transcribed the four spacer variant templates, along with the N25 and N25anti control promoters, using reconstituted wild type (wt) and αCTD∆235 (∆235) RNAP holoenzyme. As shown in Fig. 3A, the reconstituted enzymes were active on four of the six templates but showed little or no activity on SP Pcon and SP mar. The lack of transcription from the latter promoters highlights the importance of the spacer sequence in forming a stable OC with the reconstituted RNAP. Focusing on templates transcribed by the reconstituted enzymes, the αCTD deletion had a pronounced effect on DG203 transcription, showing reduced abortive yield and abortive probability at positions ≥ +14 compared to wt RNAP. This effect was particularly noticeable at positions +17 to +19 where AP values for each VLAT decreased by > 50% (Fig. 3B). Remarkably, a 10% decrease in total VLAT yield led quantitatively to an 11% increase in productive yield (Table 5). Deletion of αCTD only had a negligible effect on SP fullcon (Fig. 3A, 3B), with