RNA Polymerase Structure, Bacterial

RNA Polymerase Structure, Bacterial S Borukhov, University of Medicine and Dentistry of New Jersey, School of Osteopathic Medicine, Stratford, NJ, USA ã 2013 Elsevier Inc. All rights reserved. This article is a revision of the previous edition article by Michael Anikin, Dmitri Temiakov, and William T. McAllister, volume 3, pp. 781–784, ã 2004, Elsevier Inc.

Introduction

Transcription Cycle

Enzymatic Activities of RNA Polymerase

Transcription process can be roughly divided into three major stages: initiation, elongation, and termination. At each of these steps RNAP is targeted by large number of regulatory factors. Initiation is the first and most complex step in transcription. During initiation, the s70-holoenzyme slides along the DNA scanning for the presence of two conserved hexamer elements of the promoter centered at positions –35 and –10 (with consensus sequences –35TTGACA–30 and –12TATAAT–7, respectively, relative to the transcription þ1 start site) and separated by a 17  1-bp spacer. In addition to the conserved hexamers (that may differ from consensus), other elements immediately upstream and downstream of the –10 element (‘–14TG–13 extended –10’, ‘–15’ regions, and ‘discriminator’ (DSR)) and A/T-rich regions upstream of the –35 element (‘UP-element’ at –45; –65), all contribute to specific recognition by the s70 holoenzyme. The holoenzyme specifically binds to these elements to form a closed promoter complex (RPc). After several transitional steps the enzyme unwinds the DNA duplex around the –10 region (between nt –12 and þ2) creating a 12–14-ntlong transcription bubble, and forms an open promoter complex (RPo). In the presence of rNTPs, RPo transforms into an initiating complex (RPinit) which forms the first phosphodiester bond between rNTPs positioned at þ1 and þ2 sites and the RNA synthesis commences. At this stage, RNAP exists in the form of a scrunched complex where the upstream DNA–RNAP contacts remain intact; the downstream DNA (from þ1 to þ15) is pulled into the enzyme and transcribed into the RNA, while the transcription bubble expands extruding the excess of ssDNA strands outside of the complex. The RPinit may proceed into either abortive or productive pathways. In the former, the enzyme enters the cycle of abortive initiation when it repetitively synthesizes and releases short RNAs without leaving the promoter. In the latter, the enzyme synthesizes RNA till it reaches a critical length (typically 11–15 nt), of which 8–9 nt are basepaired with the DNA template strand (DNA–RNA hybrid) and 2–6 nt are threaded through the RNA exit channel, escapes from promoter, and enters the elongation stage of transcription. This transition leads simultaneously to the loss of RNAP–promoter DNA contacts, s dissociation, and formation of a highly stable and processive ternary elongation complex (EC). Throughout elongation stage, the size of transcription bubble in EC remains constant at 12  1 nt, and the size of RNA/DNA hybrid is maintained at 8–9 bp. EC can transcribe DNA over long distances (>10 000 bp) without dissociation and release of RNA product. However, the monotonous forward movement of EC can be interrupted by pauses which occur through at least two pathways. The first (type I) is induced by RNAP interaction with the nascent RNA secondary structures (such as RNA hairpins). This leads to distortion in RNA/DNA hybrid and temporary disengagement of RNA

In enzymological terms, RNA polymerase (RNAP) belongs to a nucleotidyl transferase class of enzymes. During RNA polymerization (forward reaction), it catalyzes (in a DNA-template dependent manner) the transfer of ribonucleoside monophosphate moiety from ribonucleotide triphosphate (rNTP) to the 30 -terminal hydroxyl group of the growing chain of RNA product, producing pyrophosphate (diphosphate, PPi) as by-product. In a reverse reaction, pyrophosphorolysis, RNAP transfers ribonucleoside monophosphate moiety from the 30 -terminal nucleotide of RNA back to pyrophosphate resulting in degradation of RNA and formation of rNTPs. In addition, RNAP catalyzes RNase-like hydrolysis of nascent RNA acting as 30 -exoribonuclease and endoribonuclease. In the course of these reactions, RNAP produces shortened nascent RNA with the free 30 -hydroxyl group and ribonucleoside-50 -monophosphate (rNMP) or 2–19-nt-long oligoribonucleotide fragments with 50 -monophosphate, respectively. RNAP also possesses a DNAhelicase activity: it unwinds the double-stranded DNA (dsDNA), separates the DNA strands (DNA melting), and induces both positive and negative DNA supercoiling. Finally, RNAP is a DNA-translocase which acts as a powerful molecular motor. During transcription, it utilizes the energy released from rNTP binding and incorporation into the growing chain of RNA to generate substantial mechanical force and to move unidirectionally along the DNA template. All enzymatic activities of RNAP require the presence of two Mg2þ ions (one high-affinity and another low-affinity) at the catalytic center.

Subunit Composition In bacteria, such as Escherichia coli (Eco), the catalytically active core enzyme consists of five subunits (a2bb0 o): two a (36 kDa), b (150 kDa), b0 (155 kDa), and o (8.5 kDa), with the total molecular weight of 380 kDa. On its own, the core enzyme is unable to recognize specific promoter DNA sequences, or melt the DNA and initiate transcription. To carry out these functions, it must bind one of several specificity factors, s (or s subunit, 20–70 kDa) to form RNAP holoenzyme (a2bb0 os). Alternative s factors bind competitively to the core, generating multiple forms of holoenzyme that can utilize different classes of promoters under various growth conditions and in response to stress. The number of s factors in different bacterial species varies from one in Mycoplasma genitalie to seven in Eco to 60 in Streptomyces coelicolor. Besides s, other regulatory transcription factors can bind RNAP to modulate its ability to recognize and bind promoters, modify its enzymatic activity, alter the rate of RNA synthesis and its processivity, and control the release of RNA product.

173

174

Molecular Biology | RNA Polymerase Structure, Bacterial

30 -end from the active site. The second (type II) results from various roadblocks to RNAP translocation caused by DNAbinding proteins, misincorporated substrates, DNA lesions, and certain DNA sequences. During type II pausing, RNAP may slide back by 2–20 nt relative to the RNA 30 end (backtrack). This results in extrusion of RNA 30 -terminus through the secondary channel of RNAP (see below) and concomitant repositioning of the catalytic center from the 30 -end to an internal site in RNA. Both types of pausing play important regulatory roles in transcription and in synchronization of transcription and translation. Under certain conditions, pausing could also lead to transcription arrest or dissociation of EC. Transcription termination occurs when EC encounters a termination signal – a 20–35-nt-long G/C-rich RNA sequence of dyad symmetry that forms stem–loop structure immediately followed by a 7–9-nt-long stretch of Us. During termination, RNAP releases the nascent transcript and dissociates from the DNA template, after which it can rebind s factor and start a new round of transcription. Under certain conditions, transcription termination can also be induced by termination factors r and Mfd.

relatively low resolution of 15 A˚. The medium- and highresolution structures of bacterial RNAP have been obtained by X-ray crystallography using thermophilic organisms Thermus aquaticus (Taq) and Thermus thermophilus (Tth). These include: the structures of Taq core (3.3-A˚ resolution); Taq holoenzyme (4 A˚); Tth holoenzyme (2.6 A˚); Taq and Tth RNAP complexes with several small molecule inhibitors and antibiotics (2.6– 3.5 A˚); the binary complex of Taq holoenzyme with promoter DNA fragment (6.5 A˚); and the EC of Tth core with synthetic DNA/RNA scaffold (2.5 A˚). In addition, high-resolution structures (1.8–2.9 A˚) of individual domains of Esc RNAP subunits b0 , a and s, and their complexes with DNA and regulatory transcription factors were also reported. Together with information gained from a wide range of biochemical, biophysical, and genetic studies, these data refine our understanding of bacterial RNAP structure–function and provide a comprehensive view of transcription process and its regulation. The structure of a typical bacterial RNAP, such as the Taq/ Tth core enzyme, represents an asymmetric irregularly shaped ellipsoid with a deep cleft (internal channel) in the middle of the molecule, resembling a crab’s claw. The overall dimension is about 155 A˚ in length (from the back to the tip of the claw), 115 A˚ in height (from the top to the bottom pincer), and 110 A˚

Structure of RNAP Core Enzyme

in width (alongside the channel). The top pincer of the claw is made up mostly of the b subunit, and the bottom pincer mostly of b0 (Figure 1). The pincers are joined at the base of the claw by the N-terminal domains (NTDs) of asymmetrically

Eco RNAP is the best characterized bacterial enzyme; however, its structure – determined by cryo-electron microscopy – has a b C-terminal coiled-coil

Secondary channel (pore)

b

b bridge b fork-1 (F) helix

b catalytic b fork-2 b fork-1 loop b lobe-1

b fork-2 b lobe-2

b trigger (G) loop b lobe-2

aI b rudder

b rudder

b catalytic loop

bclip 2

w

(a)

b flap tip helix

b lid

aI

b flap tip helix

Main channel w 90

b N-terminal Zn-finger

b flap

b jaw b C-terminal Zn-binding domain

b zipper

b lid

b jaw

Upstream b clamp

b rudder

b lobe-2

b trigger (G) loop

aII bclip 1

b lobe-1

Downstream Upstream b clamp b clamp

aII

b N-terminal Zn-finger

b 20,

b zipper

RNA exit channel

90

b

b

w

b-SI1 b-SI1

aI

b-SI2

aI -NCD1

b-SI3

aII

w

b

(b)

b-SI3

aII

w

(c)

b

w 90

Figure 1 Structural overview of RNAP core. (a) Structure of Taq RNAP core shown in three rotational views, using Molsoft ICM Browser Pro program. Left panel, secondary channel view; middle panel, main channel view; right panel, RNA exit channel view. The structure is represented as lightly colored ribbons (aI, light gray; aII, olive; b, light yellow; b0 , light cyan; o, light blue) with structural elements emphasized in full color and indicated. Mg2þ ion is shown as small magenta sphere. The structures of b0 -trigger loop, b0 -rudder and b0 N-terminal Zn-finger domains are modeled using the structure of Tth holoenzyme. The b0 -NCD1 (G164-S449) and the aCTDs are not shown. (b) Secondary channel view of Taq RNAP core structure with b0 -NCD1 colored blue. (c) Two rotational views of the modeled structure of Eco RNAP core (colored as above, except: b, light orange; b0 , cyan; and nonconserved domains bSI1, bSI2, and b0 SI3 colored bright green, red and light yellow, respectively. Left panel, secondary channel view and right panel, main channel view.

Molecular Biology | RNA Polymerase Structure, Bacterial placed a-dimer serving as a platform for RNAP assembly. aI-NTD binds only the b subunit, whereas aII-NTD binds mostly b0 subunit. The o subunit wraps around the b 0 C-terminus near the bottom pincer, serving as a b0 chaperone.

RNAP Channels The large internal space between the pincers is partitioned into three channels: the main channel (20–27 A˚ in diameter) and two minor channels (10–12 A˚ in diameter) that branch off from the primary channel to form the downstream-facing ‘secondary channel’ and the upstream-facing ‘RNA exit channel’. The main (or primary) channel accommodates 13–15 bp of the downstream DNA duplex, 12–14 nt of the melted region of DNA (the transcription bubble), and up to 8–9 nt of DNA/ RNA hybrid. In the holoenzyme, the main channel also accommodates the bulk of s (domains s1–s3). The back wall of the primary channel – the most conserved part of the RNAP structure – contains a combination of b-sheet and a-helical elements of b and b0 that comprise two juxtaposed six-bstranded double-c b-barrel (DPBB) domains forming the active site cleft (see Table 1 for list of b/b0 functionally relevant structural domains and elements). The DPBB domain of b0 carries the catalytic loop with a triad of aspartates holding essential Mg2þ ions, whereas the DPBB domain of b harbors basic residues of the substrate-binding site (see below). The top and bottom walls of the channel are formed by b and b0 clamps, respectively. The b clamp consists of an upstreamfacing lobe-1 and a downstream-facing lobe-2 with two conserved loops, b fork-1 and b fork-2, protruding into the channel toward the active center. The downstream portion of the b0 clamp is made of the C-terminal Zn-binding domain surrounded by groups of conserved a-helical elements that form the cradle for the downstream dsDNA. These include: two dsDNA-binding three-helix bundles, a long a-helical bridge (F-bridge or bridge helix) connecting the b and b0 clamps near the active center and two adjacent a-helices of the flexible trigger loop (G-loop). Together with the fourstrand b-sheet jaw domain, the trigger loop and bridge helix form the wall separating the primary and secondary channels (Table 1, Figures 1(a) and 1(b)). The upstream portion of the b0 clamp consists of an extended N-terminal coiled-coil element (the major docking site for s and elongation factors NusG and RfaH) with three bulging out loops: rudder, lid, and switch-2, all projecting into the main channel. The secondary channel serves as an entry/exit pore for NTP substrates and by-products of RNA synthesis: the pyrophosphate, hydrolyzed 30 -terminal nucleotides (NMPs), and short oligoribonucleotides. It also provides an exit pathway for the backtracked RNA 30 -terminus and for abortive transcripts during initiation. The channel, which is made entirely of b0 , has a funnel-like shape with wider opening on the outside surface of RNAP and a narrow opening near the catalytic center. This shape allows specialized secondary channel transcription factors to anchor at the entrance of the channel and reach through the channel to the RNAP active center to modulate its activity. The secondary channel’s rim is made of the b0 C-terminal coiled-coil element which serves as the major docking site for such factors as GreA, GreB, DksA, TraR, RNK, and Gfh1. The back wall of the channel consists of the extended loop adjacent

Table 1

175

Main structural elements of bacterial RNAP

Domains and structural elements

Residues (T. thermophilus)

b Subunit b DPBB domain b Clamp lobe-1 domain b Clamp lobe-2 domain b Flap domain

668–698; 832–879; 969–1004 18–142; 332–397 143–331 701–832

b Lobe-2 pincer loop b Lobe-1 loop b Fork-1 helix b Fork-2 b Fork-1 b RNA-binding helix b Flap tip helix b Switch-3 b Helix–loop–helix (HLH)

151–182 387–397 397–411 412–441 442–454 554–568 768–779 998–1005 1050–1080

b0 Subunit Downstream b0 clamp domain b0 NCD1 domain b0 DPBB domain Upstream b0 clamp domain

1–131; 455–605; 1444–1468 132–454 621–646; 694–769 1093–1443

b0 Zipper b0 N-terminal Zn-finger domain b0 Downstream clamp N-terminal HTL loop b0 Downstream clamp helix b0 Rudder b0 Lid b0 Upstream clamp N-terminal coiled coil b0 Switch-2 b0 Exit channel rim helix b0 Catalytic loop b0 Secondary channel rim C-terminal coiled-coil b0 Clip-1 b0 Bridge (F) helix b0 C-terminal Zn-binding domain b0 Clip-2 b0 Trigger (G) loop b0 Jaw b0 Downstream clamp C-terminal HTL loop

22–48 49–83 105–113 486–496 581–603 525–538 540–620 606–620 664–680 737–744 958–1015 1020–1035 1066–1103 1105–1116; 1155–1189 1353–1365 1233–1257 1268–1330 1434–1444

to the bridge helix (F-loop), the helix–loop–helix (HLH) element clip-1 and a bundle of short a-helices connected by unstructured loops. The side wall is formed by the HLH element clip-2, the bridge helix, the trigger loop, and the jaw (Table 1, Figure 1). The RNA exit channel has a curved shape which helps to accommodate the C-terminal domains of s (s3–4 linker and s4) in the holoenzyme, and serves as an outlet for RNA product in ECs. The channel is also involved in RNA/DNA hybrid strand separation, and interactions with RNA hairpins in paused ECs and in transcription termination. The exit channel walls are mostly made of the upstream portions of b and b0 pincers. The double-twisted loop switch-3 and the C-terminal HLH elements of b, together with the four-helix bundle just upstream of the main channel b-barrel of b0 , comprise the back wall. The top wall consists of a flexible b flap domain,

176

Molecular Biology | RNA Polymerase Structure, Bacterial

whereas the bottom wall is made of the b0 rudder, the lid, and the zipper elements (Table 1, Figure 1(c)). The outside rim of the channel is formed by the b flap tip helix, the b0 N-terminal Zn-finger domain, and the b0 exit channel rim helix. The C-terminal a-helical domain of aI, aI-CTD (connected to aI NTDs by flexible linker), may also contribute in forming the channel’s rim.

Nonconserved Domains Despite their overall similarity, bacterial RNAPs are structurally distinct, in part due to the presence of different large nonconserved regions residing in their b and b0 subunits. These domains are dispensable for RNAP basic functions; however, they play important regulatory roles in transcription. For instance, the Taq/Tth RNAP contains a 283-residue domain in the nonconserved portion of b0 present in the upstream b0 clamp (Taq/Tth b0 -NCD1). The crystal structure of this domain comprises five repeats of sandwich-barrel hybrid motifs (SBHMs) visible in Tth RNAP structure as an extended part of the lower b0 pincer (Table 1, Figure 1(d)). The main function of b0 -NCD1 is to stabilize core–s70 interactions (see below). Unlike Taq/Tth, the Eco b0 has just one SBH motif which contributes to selective binding of alternative s factors. Instead, Eco b0 harbors a 188-residue insertion in the conserved G-loop (trigger) element (Eco domain b0 -SI3), which is absent in Taq/ Tth RNAP. The crystal structure of this flexible domain reveals two repeats of SBHM (Figure 1(e)) that can be modeled into the cryo-EM map of Eco RNAP–DNA promoter initiation complex (see Figure 3). Eco b0 -SI3 is required for interaction with transcript cleavage factors GreA and GreB; it also strongly affects RNAP’s intrinsic nucleolytic activity, transcription fidelity, and ability to pause, arrest, and terminate RNA synthesis. On the other hand, Eco b contains two domains that are absent in the Taq/Tth b: a 115-residue element (b-SI1) inserted into lobe-2 between conserved regions B and C and a 99-residue segment (b-SI2) inserted into lobe-1 between conserved regions G and H. The high-resolution structures of these domains have been recently solved and modeled into the cryoEM map of Eco RNAP (Figure 1(e)). Eco b-SI1 contributes to efficient promoter escape by RNAP; it is also targeted by the bacteriophage T4 termination factor Alc, which selectively induces premature transcription termination on E. coli DNA. The role of b-SI2 is currently unknown.

Conformational Flexibility The structural organization of RNAP can be viewed as a combination of stationary (or fixed) and mobile modules. The fixed core module comprises two aNTDs, o and the bulk of b and b0 subunits that carry the catalytic site and form the back wall of all three channels. The mobile modules include elements associated with binding nucleic acids: the upstream and the downstream portions of b0 pincer (the N-terminal s-binding clamp with rudder and switch-2, and the C-terminal dsDNA-binding b0 clamp, respectively); the top b pincer (b lobe-1 and lobe-2 carrying b fork-1 and -2); parts of the active center (b0 bridge-helix and trigger-loop); and parts of the RNA exit channel (b flap, b0 N-terminal Zn-finger, b0 zipper, b0 lid). These mobile modules confer considerable conformational

flexibility to RNAP structure: the swinging motion of the b and b0 clamps, inferred from comparing the structures of Taq and Eco core enzymes, results in the opening of the claws by 25 A˚. This opening is required for initial s-subunit binding and to allow the template DNA strand to enter the primary channel and reach the catalytic cleft during transcription initiation. It is also important for s release during transition to the elongation stage of transcription, and for EC dissociation and transcript release during termination. The subsequent closing of the clamp may help RNAP to tightly hold RNA–DNA hybrid in position during processive elongation.

Structure of RNAP Holoenzyme The high-resolution structure of bacterial holoenzyme is currently available only with Tth s70, the major housekeeping s factor. The Tth holoenzyme maintains its overall crab claw-like shape since most of s is bound on the core surface, except for a short central segment of s (residues 313–342), which is buried in the core molecule (Figures 2(a)–2(d)). The rest of s is wedged between the upper (b) and lower (b0 ) pincers at the upstream side of the core enzyme, creating a wall that partially blocks the opening of the primary channel. Transition from core to holoenzyme is accompanied by changes in the positions of several structural domains of core, by 2 to 12 A˚. The most noticeable are in the b flap domain (especially b flap tip helix by 12 A˚); b lobes-1 and -2, aI-NTD and the b0 main channel clamp (4–6 A˚). The RNA exit channel is almost completely blocked by the presence of s.

Structure of s Tth s70 and all s70-like factors share four regions of sequence homology (1–4), which are further divided into subregions (Figure 2(e)). All conserved regions have been implicated in interactions with core and/or DNA. The visible portion of Tth s70 resembles a curly bracket-shaped structure. It consists of three a-helical domains s2, s3, and s4, and the linker domain s3–4. Domain s2 comprises four helix–turn–helix (HTH) motifs, encompassing conserved regions 1.2–3.0 and the large nonconserved segment between regions 1.2 and 2.1. s3, comprising three a-helical bundle, contains regions 3.0–3.1. The linker domain s3–4 is an extended – mostly unfolded – 30residue-long hairpin loop, which includes region 3.2. s4 is formed by four a-helices arranged as two HTH motifs, and includes conserved regions 4.1 and 4.2 (Figures 2(d) and 2(e)).

s–Core Interactions Domains s2–s4 are located on the enzyme’s surface, wedged between the upper (b) and lower (b0 ) pincers at the upstream opening of the primary channel (Figures 2(a)–2(d)). The hairpin loop of s34 threads through the primary channel reaching the catalytic pocket and comes out from the RNA exit channel. In the holoenzyme, the s2, s3, and s4 domains are situated at an optimal distance to contact the –10, extended –10 region, and –35 elements of the promoter DNA in RPc, respectively. The location of domain s1 containing the

Molecular Biology | RNA Polymerase Structure, Bacterial

177

Secondary channel aI

b lobe-1

b lobe-2 aI b-ruder

Upstream b clamp 2+

Mg s3

b-NCD1

aII aII

w

w b-NCD1

b flap s2

s4 b flap tip helix

b Znfinger

(a)

Upstream b clamp

b

Secondary channel

(c) b lobe-1

a aI

b clamp C-terminal coiled-coil

Mg2+ b-NCD1

s3

aII

s3

b-NCD1

s3-4

s2

w

w

b flap

s2

s4

b b Znzipper finger

(b)

b zipper

(d)

bZnfinger

b flap tip helix

s4

s3-4 hairpin loop

N

C

s3.0 HTH

s4.2 HTH

s2

s1 (not resolved) 1

N

20

40

60

80

1.1

100

120

140

160

1.2

200

2.1

220

2.2

240

2.3

260

2.4

280

3.0

s4

s3-4 300

3.1

DNA melting

Core binding

320

340

360

380

4.1

3.2

400

420

4.2

C

Abortive initiation

+1

Promoter DNA

+15

180

Nonconserved region

Autoinhibitory

(e)

s3

ITR

+10

+5

DSR

+1

–5

–10 region –10

Spacer

–35 region

–15 Extended -10 region region –15

–20

–25

–30

–35

UP element –40

–45

–50

Figure 2 Structural overview of Tth RNAP holoenzyme. The structures are shown as molecular surface (a–c) and ribbons (d, e) views using Accelrys DS Visualizer program, with color coding as follows: aI, light gray; aII, slate gray; b, yellow; b0 , cyan, s, magenta, o, dark cyan. Locations of functionally relevant domains and structural elements are indicated. The Zn2þ ion bound in the b0 Zn-finger domain is shown as small blue sphere and catalytic Mg2þ ion is shown as red sphere. The N-terminal domain of s carrying region 1.1 missing from the structure is not shown. (a, b, d) The back (RNA-exit channel) view of the holoenzyme. (c) The front (secondary channel) view of holoenzyme is obtained from (a) by 180 rotation about the y-axis. (b) The view of holoenzyme as in (a) with b removed to reveal the inside of the RNA exit channel and s3–4 (colored light magenta) buried in it. The location of all visible s domains is indicated. (e) The structural and functional organization of s. Top panel is a ribbon view of s70 from Tth holoenzyme structure shown in (c). Colored regions correspond to the evolutionarily conserved domains of s as shown in the functional map of s70 below. Bottom panel is a linear representation of s polypeptide with structural domains and conserved regions shown as numbered and color-coded boxes. Underneath is a diagram of DNA promoter regions and interactions made by DNA-binding domains of s.

178

Molecular Biology | RNA Polymerase Structure, Bacterial coil domain (b0 540–585) of the b0 main channel clamp – the major s docking site (Figure 2(d)). Multiple contacts of b0 NCD1 with s1.2 and the nonconserved region of s2 also contribute to more stable Tth s70 binding. Additional, although weaker, interactions occur between b flap and s4 [53], and between s3 and b lobe-1 (Figure 2(d)). In the presence of specific activators, s4 also interacts with a-CTD. All reported crystal structures of RNAP core and holoenzyme still lack a few elements, the most notable of which is an 80-residue-long aCTD with a 14-residue flexible linker connecting it to the aNTD. However, the structures of Eco aCTD in complex with Eco transcriptional activator CAP and DNA are now available, and its position in RNAP is modeled using cryoelectron microscopy structures of Eco RNAP–open promoter DNA complex (Figure 3). The aI— and aII— CTDs recognize and bind the upstream (UP) promoter element and serve as targets for many transcriptional activators.

unresolved N-terminal portion of s (residues 1–73), with poorly conserved region 1.1, in the structure of Tth holoenzyme is currently unknown. Biochemical and biophysical data suggest that s1.1 possesses autoinhibitory and regulatory functions. In free s, s1.1 interacts with s4 and blocks s factor from binding to DNA. In the holoenzyme, s1.1 facilitates the initial association of s with core and modulates the efficiency of transcription initiation at some promoters. The s–core interaction interface covers an extended surface area with multiple cooperative contacts between discrete domains of s and different parts of the core. This is likely to be the reason for the high stability of the s-core association (KD ¼ 10–9 M). However, most of these contacts are relatively weak and are distributed over a large area. This explains why different s factors successfully compete with s70 for binding to core. The strongest interactions are observed between conserved regions 2.1 and 2.2 of s2 and the N-terminal coiled-

(a) -15 -35 Extended “enhancer” b region -10 region s4.2 –45 zipper s3.0 UP–10 region element

140

Downstream DNA

b zipper

aI-NTD

s3.0

aII-NTD

+1

aI-, aII-CTD R265, N294, K298

w

s2.4

Downstream DNA

aI-CTD

w b

(b)

b trigger b (G) loop jaw

–35 region s4.2

b bridge s3.2 tip (F) helix hairpin

b

–55 UPelement

–45 UPelement

Downstream DNA

(c)

aII-NTD Catalytic Mg2+ -ion

T +1 NT +1

+1

b lid

DSR (NT –6/–4) s1.2DSRrecognition NT –10 region s2.3 –10-melting

Extended –15 –10 region “enhancer”

b lobe-2

s4.2

b lobe-1

Upstream DNA

Mg2+ a I-NTD

b rudder

aII-NTD

s3.2 tip hairpin D338, K240, D341

Downstream DNA b switch 2

aII-CTD

aI-NTD

s2.3 –15 “Enhancer” Extended –10 region –10 region

–55 UPelement

+1

Downstream DNA

DSR (NT –6/–4) s1.2 DSRrecognition (K216, K215 R208) NT –10 region s2.3 –10-melting F248, T252, Y253, W256

UPelements s3.0 basic cluster (R259, K285, R288, R291)

–35 region

Figure 3 RNAP holoenzyme-promoter DNA complexes. Structural models of Taq RPc (a) and RPo (b, c) with Plac promoter DNA (endpoints –65; þ25) and aI- and aII–CTDs are shown as molecular surfaces generated by Accelrys DS. The color coding are the same as in Figure 2, except for s which is colored pink. Structural elements of s, aCTDs and b0 involved in interactions with DNA are indicated by different colors. DNA is shown in stick representation with T and NT strands colored dark green and light yellow, respectively, except for promoter elements colored as indicated. Small magenta balls indicate the position of catalytic Mg2þ ions. The b0 -NCD1 is not shown. (a) Model of Taq RPc. Left panel, the RNA exit channel view of RPc (same as in Figure 2(a)). Right panel, the bottom (rim of the exit channel) view of RPc, obtained by 140 rotation of the left view about the x-axis. (b) Model of Taq RPo is shown with b removed to reveal the inside of the main channel and to show the interactions of s and b0 elements with transcription bubble and downstream DNA. The view is similar to the left view shown in (a) but rotated by 60 about the x-axis. (c) Zoomed-in detailed view of (b) showing s–promoter DNA interactions. The inset panel shows the full view of RPo and location of b lobe-1 and -2 that close the main channel and block the access to transcription bubble. The view is obtained by rotation of (b) by 90 about y-axis and tilted backward by 25 .

Molecular Biology | RNA Polymerase Structure, Bacterial

Structure of Holoenzyme–Promoter DNA Complexes The high-resolution structures of RNAP–DNA binary complexes are yet unavailable. However, the model structures of the RNAP–closed promoter complex (RPc) and the RNAP–open promoter complex (RPo) were reconstructed from low´ resolution (15–20A˚) cryo-electron microscopy studies of Eco holoenzyme–lacUV5 promoter open DNA complexes, and from crystallographic studies of Eco CAP-aCTD–DNA complex, Taq s4 domain in complex with a DNA fragment containing –35 element (–26 to –37), and of Taq holoenzyme binary complex with fork-junction promoter DNA (containing dsDNA from –12 to –45, and ssDNA from –11 to –7) which partially mimics the RPo. The resulting models of RPc and RPo are consistent with most of the biochemical, biophysical, and genetic data accumulated in the last 20 years.

Closed Promoter Complex (RPc) In the modeled structure of Taq RPc, the promoter dsDNA is located on the surface of holoenzyme, outside the RNAP main channel, bound mostly by s (Figure 3(a)). In RPc the dsDNA is shielded by RNAP from position –55 to –6. Downstream of position –5 dsDNA does not interact with RNAP. All major DNA sequence-specific contacts in RPc, conserved –10, extended –10, and –35 elements of the promoter, are mediated by the s2, s3, and s4 (regions 2.2–2.4, 3.0, and 4.2, respectively) through polar and van der Waals contacts (Figures 2(e) and 3(a)). Specific s–DNA contacts are summarized in Table 2. Depending on the length of the spacer and the presence of noncanonical sequence elements, the s3 region 3.0 and b0 zipper (residues R35, T36, and L37) may contribute to recognition of noncanonical –15 enhancer element, whereas aCTDs (residues R265, N294, and K298) contact A/T-rich sequences at positions from –45 to –65 and up to –90. Because dsDNA is not fully encircled by RNAP in RPc, and because DNA contact area is relatively small and most of the protein–DNA interactions are weak, the RPc is intrinsically unstable. It is sensitive to salt (>0.2 M NaCl) and competitors such as heparin, and can easily dissociate. Though weak, these interactions provide initial promoter recognition by RNAP and increase promoter occupancy. They also cause significant distortion in DNA structure, thereby facilitating further steps in transcription initiation: DNA melting, strand separation, and template strand insertion into the active site cleft. In the RPc the RNAP-bound DNA is Table 2

179

bent or kinked at three places: around –25 (by 8 to accommodate variable spacer length), at the –35 (by 36 induced by insertion of s4 HTH motif into the major groove), and at further upstream of –45 induced by aCTD–DNA minor groove interactions. The DNA bending at –35 aids in proper binding of upstream DNA by aCTD and upstream transcription activators.

Open Promoter Complex (RPo) The proposed structure of RPo includes both the dsDNA and the ssDNA covering positions from –60 to þ25. Unlike RPc, in RPo both strands of DNA up to þ20 position are fully enclosed inside the RNAP main channel (Figures 3(a)–3(c)). The RNAP interactions with the upstream portion of ds DNA (from –60 to –17) are similar to that observed for RPc; however, at –16 the DNA trajectory changes due to a sharp bent by 37 toward the RNAP. At position –11 the template (T) and nontemplate (NT) DNA strands separate, and enter different paths for 15 downstream nucleotides until they reanneal at position þ3, thus creating the transcription bubble. Thermal fluctuation and free energy accumulated in the holoenzyme are believed to be sufficient for DNA melting and transcription bubble formation. The initial DNA melting starts at the A/T bp at position –11. Conserved aromatic residues located on the surface of s region 2.3, Tth F248, Y253, W256, W257, and T252 are positioned near the dsDNA–ssDNA junction of the transcription bubble. The nucleotide bases at –8/–9 and –9/–10 interact with F248 and Y253, respectively, whereas W256 and/or W257 stack on the exposed face of the –12 bp, forming the upstream edge of the transcription bubble (see Table 2). Also, W256 is responsible for capturing the exposed, or flipped A base at the key position –11 of the NT-strand, thereby facilitating DNA melting. The DNA NT-strand (from – 2 to þ4) further continues its path in a groove formed between b lobe-1 and lobe-2 modules (Figures 3b and 3(c)). The conserved s1.2 basic residues (K215, K216, and R208) are implicated in interactions with DNA at positions from –6 to –4 of the DSR region. These interactions stabilize the RPo formation and accelerate transcription initiation at some promoters. Another cluster of conserved basic residues (R259, K285, R288, and R291) of s2.4 and s3.0 is proposed to pull the DNA T-strand (from –7 to þ3), through electrostatic interactions, into a groove formed by s3, b0 lid, and the b0 rudder (Figure 3(c)). The DNA is then placed into the main channel between the active site wall and s3–4 hairpin loop. Conserved charged

Summary of s-DNA contacts

DNA position (T, NT)a

Taq s70 residues

s70 Region (domain)

Major groove at 12A, 12T (base specific) 13, 14, 15 (phosphate backbone) 13A, 13T (base specific) Major groove at 14T, 15G (base specific) 17, 16, 15, 14 (phosphate backbone and base specific) 31G, 30T (base specific) 32T (base specific) 33C (base specific) 35T (base specific) 31, 32, 33; 35, 36 (phosphate/ribose backbone)

Q260, N 263 R237, K241 E281 H278, E281 R274, V277, H278, E281 R409 R413 E410 R411, Q414 R413, R387, L398, E399, R379, T408

2.4 (s2) 2.3 (s2) 3.0 (s3) 3.0 (s3) 3.0 (s3) 4.2 (s4) 4.2 (s4) 4.2 (s4) 4.2 (s4) 4.2 (s4)

a

Roman text is DNA position T, and italic text is DNA position NT.

180

Molecular Biology | RNA Polymerase Structure, Bacterial

residues at the tip of the hairpin participate in juxtaposing DNA þ1 position to the catalytic center and assist in correct placement of the initiator rNTP (Figures 3(b) and 3(c), Table 2). Simultaneously, the dsDNA downstream of þ3 to þ12 is brought inside the downstream DNA binding clamp between the b0 jaw, b lobe-2, and the downstream b0 clamp. The b0 switch-2, a small flexible loop residing in the upstream b0 clamp in the middle of the main channel cavity, controls the binding of the downstream part of the unwound DNA T-strand in the active site cleft. Switch-2 functions as a hinge mediating opening and closing of the b0 clamp, and plays a critical role in downstream propagation of transcription bubble during formation of RPo. It serves as a primary target for the action of antibiotic Myxopyronin (Myx) which binds to b0 switch-2 and prevents its refolding. Thereby, Myx interferes with opening of the active center cleft, and blocks entry and unwinding of the downstream DNA. Finally, b lobes-1 and -2 move toward b0 clamps by 6–8 A˚, which leads to substantial reduction of the main channel opening. This closing hampers DNA dissociation and prevents the collapse of the transcription bubble. Thus, a stable, competitor-resistant and initiation-competent RNAP– promoter DNA complex is formed.

Structure of EC Several high-resolution structures of bacterial ECs are currently available: (1) the 2.5-A˚-resolution structure of Tth RNAP core in complex with a synthetic nucleic acid scaffold containing 14-bp-long downstream dsDNA, 9-bp-long RNA/DNA hybrid, and 7-nt-long ssRNA (Tth EC); (2) the 3-A˚-resolution structures of Tth EC co-crystallized with nonhydrolyzable substrate analog, adenosine-50 -[(a,b)-methyleno]-triphosphate (AMPcPP), and with AMPcPP plus antibiotic streptolydigin (Stl); and (3) the 4.1-A˚-resolution structure of Tth EC with a comparable RNA/DNA scaffold but containing an additional 24-nt-long RNA hairpin, co-crystallized with transcription factor Gfh1. These structures collectively offer a detailed view of the bacterial EC architecture, and help elucidate the mechanisms of RNAP catalysis, that is, substrate binding, incorporation, and enzyme translocation. The structures also provide insights on how the interactions between various structural elements of RNAP, the nucleic acids, substrates, and regulatory molecules determine EC’s stability, high processivity, and response to signals encoded in DNA and RNA. Together with the knowledge gained from the structural studies of eukaryotic ECs we have greatly expanded our understanding of the transcription process carried out by multisubunit RNAPs.

RNA/DNA hybrid. The kink is induced by the protein barrier in the form of the b0 bridge helix and b fork-1 and -2 (Figures 4(a) and 4(b)). The b0 rudder loop residing between the dsDNA and DNA/RNA hybrid stabilizes their position and conformation via direct electrostatic contacts (Table 3). The 9-bp RNA/DNA hybrid is placed at the back wall of the main channel, snugly packed between the upstream b0 clamp, b0 rudder, and b lobe-1. The hybrid is surrounded by the most conserved structural elements in RNAP including b0 DPBB domain with the Mg2þ-binding catalytic loop, b0 switch-2, b DPBB domain with substrate-binding loops, b switch-3, b RNA-binding helix, b fork-1 with fork-1 helix, and b fork-2 (Figure 4(c)). The latter four elements of b-subunit bind antibiotics Rifampicin (Rif) and Sorangicin (Sor), which strongly inhibit the early steps in RNA synthesis. They act by blocking the path of the growing transcript beyond the length of 2–3 nt, thereby locking RNAP in abortive initiation complex. There are numerous of mixed-mode polar and van der Waals contacts between the phosphate backbone of RNA/DNA hybrid and the conserved basic and hydrophobic residues of the hybridbinding pocket, respectively (Table 3, Figure 4(c)). These interactions provide an optimal combination of repulsion and attraction forces that ensures a sterically tight placement of the hybrid in the active site cleft, without excessive binding of the nucleic acids. The b fork-2 participates in dsDNA melting by stacking R422 on base pair at iþ2 register and inducing DNA strand separation (Figures 4(b)–4(d)). The nascent RNA is separated from the T-strand DNA by the b0 lid which sterically blocks the expansion of RNA/DNA duplex beyond 9 bp and fixes the upstream edge of transcription bubble. The b switch-3 also participates in RNA separation by trapping the first displaced RNA nucleotide. The 7-nt-long ssRNA is threaded through the RNA exit channel making weak interaction with the surrounding upstream elements: the b0 zipper, b0 N-terminal Zn-finger, b0 rim helix, and b flap elements (see Table 3, Figures 4(c) and 4(d)). Compared to the structures of core or holoenzyme, the b0 clamp modules in EC are shifted toward b lobes by 4–6A˚. The closing motion of the clamps substantially reduces the main channel aperture and blocks the access to the RNAP active center. This implies that the unobstructed secondary channel serves as the main conduit for substrates and byproducts of RNA synthesis. The observed closure of the clamps also explains the remarkable stability of the EC, for example, its resistance to salt (up to 2 M NaCl) or polyanion DNA competitors (150 mg ml–1 heparin).

Structural Rearrangements at the Active Center Structural Organization of EC and Protein–DNA/RNA Interactions The structures of Tth EC show that 13 bp of the B-form dsDNA resides in the downstream DNA-binding cavity, constricted between the downstream b0 clamp, the b0 jaw and b lobe-2 (Figure 4(a)). The downstream dsDNA makes very few direct polar contacts with RNAP. Most of the RNAP–DNA interactions including electrostatic and hydrophobic are weak and nonspecific (Table 3); this endows EC with fast translocation and high processivity. The T-strand DNA makes a sharp 90 kink at the junction between the downstream dsDNA and

The structures of Tth ECs are captured in the post-translocated state: the 30 -terminal RNA nucleotide, base-paired with the i1 template DNA base, resides in the active site coordinated by one of the two catalytic Mg2þ-ions (Figure 4(e)). The unpaired iþ1 template DNA base is stacked between the i1 nucleotide of the template and the b0 bridge helix, and is positioned in the active site for base pairing with the incoming substrate rNTP (Figure 4(e)). Since the iþ2 register is base-paired, the substrate can bind only to a single base-specific site, iþ1. The structure of Tth ECs with the nonhydrolyzable substrate analog AMPcPP is consistent with this view (Figures 4(e) and 4(f)).

Molecular Biology | RNA Polymerase Structure, Bacterial

Secondary channel

b lobe-2

120

b lobe-2 5 RNA

2+ Mg

180

b flap

3 RNA Downstream ds DNA

Mg2+

b jaw

(a)

181

5 RNA

5 RNA Downstream ds DNA binding cavity

β downstream clamp b lobe-2 b fork-2 pincer loops

3 RNA

+2

b RNAbinding helix

b DPBB domain

b fork-1

Secondary channel

RNA/DNA hybrid

b lobe-1 loop b fork-1 helix

Mg

b fork-1

(I)

b b helixswitch-3 loophelix

b fork-2

b exit channel rim helix b flap tip helix

30

b rudder

180

+14

2+

140 b switch-2

b clamp clamp C-teminal b N-terminal HLH-loop downstream HLH-loop clamp helix

b

(b)

–16 5 RNA

Catalytic Mg2+-ion binding loop

(c)

b b lid rudder

b switch-2

b DPBB domain

b switch-3

(d)

Downstream ds DNA

5 RNA b N-terminal Zn-finger

b zipper

+1+2 DNA (i–1) RNA/DNA hybrid

–9 –16

b catalytic loop

5 RNA

3 RNA

(e)

+14

Bridge helix

Mg2+ (I) Mg2+ (II) Substrate rNTP

+2

b D686

I II g

Unfolded (open) trigger loop

b R783

(f)

DPBB domain

b R1029 Clip-1

p

Closure of the active center I

Nucleotidyl transfer

p

p p II p

4. Pretanslocated complex I

Pyrophosphate release, active center opening

b H1242

p II p

5. Pretanslocated complex*

b trigger helices

I

Reversible translocation

II

6. Post-translocated complex

(g)

p

Open trigger loop

II

3. Insertion complex

b F1241

b R1239

b R879

I

I

b M1238 a b

b E685 Folded (closed) trigger helices

1. Post-translocated complex Substrate NTP binding

b T1088 bridge helix

b D741 b D739 b DPBB domain

DNA (i+1)

2. Preinsertion complex

b D743

Downstream ds DNA –1 +1

RNA 3-end b N737 (i–1)

Open trigger loop

I

Closed trigger helices Closed trigger helices Open trigger loop Open trigger loop

Figure 4 Structure of EC and implications for transcriptional mechanisms. (a) Overall structure of Tth EC showing the location and trajectory of nucleic acids. Left panel, the main channel view of EC in ribbon representation. The color code is same as in Figure 1(a). RNA, red; DNA T strand, dark green; DNA NT strand, light yellow. Middle and right panels are the molecular surface views of EC (white) with nascent RNA and RNA/DNA hybrid (shown as colored semitransparent ribbons with sticks) completely buried inside RNAP main channel. Views of the left and middle panels are the same. View in the right panel is obtained by rotation of the middle panel as indicated. (b, c, and d) Close-up view of the RNAP structural elements involved in interaction with the downstream dsDNA (b), RNA/DNA hybrid (c), and nascent RNA (d). Views in (c) and (d) are obtained by rotation of view shown in (b) about vertical and horizontal axis, as indicated. (e) Nucleic acid scaffold and mobile structural elements at the active center. The view is obtained by rotating the view in (b) around y-axis by 180 and tilting forward by 90 . The nucleic acids end points and positions of DNA T strand bases at registers –1, þ1, and þ2 are indicated. Bridge helix and trigger loop are represented as colored ribbons. Trigger loop is shown in two alternative conformations: as folded helices (visible in Tth EC–AMPcPP structure), and as unfolded loop (visible in holoenzyme structure), colored blue and cyan, respectively. Substrate AMPcPP bound at the insertion site, yellow sticks; catalytic Mg2þ ions, magenta spheres; bridge helix, magenta ribbon. (f) The close-up view of the key components of the catalytic center substrate-insertion site. The view is derived by tilting the view shown in (e) forward by 20 . The DNA, RNA, and AMPcPP are represented as cyan, magenta and yellow stick, respectively. Four groups of residues are shown: those that coordinate Mg-I and Mg-II directly (red) or indirectly (orange), those that participate in binding and proper orientation of a-, b-, and g-phosphates of rNTP (blue), and those responsible for recognition of correct rNTP (light green). (g) Diagram showing the sequential steps of nucleotide addition cycle (NAC). DNA and RNA are depicted as bars, colored coded as in (a–e). Trigger loop, cyan circles connected by long dashed line; trigger helix, blue circles connected by short line; bridge helix, purple circle; Mg2þ ions, small magenta circles; rNTP, short orange bar with three small circles (phosphates).

182

Table 3

Molecular Biology | RNA Polymerase Structure, Bacterial

Summary of protein–DNA/RNA contacts in EC RNA or DNA position (T, NT)a

Tth b or b0 residues

RNAP structural element

Nascent ssRNA

−16 Base −16/−15 Phosphate −14/−13/−12 Phosphates −12/−11 Phosphate

b E770, L773 b0 K671 b E764 b Y1013

b Flap tip helix b0 Exit channel rim helix b Flap tip helix b Helix–loop–helix

RNA/DNA hybrid

−10/−9 Phosphate −9 Base, −9 base −9/−8 Phosphate −9/−8 Phosphate, minor groove −8, −9 Minor groove −8/−7 Phosphate, major groove −7/−6/−5 Phosphate, minor groove −5/−4 Phosphate −4/−3 Phosphate −4/−3 Phosphate, minor groove −3/−2 Phosphate −3/−2 Phosphate −3/−2 Phosphate −3/−2 Phosphate −2/−1 Phosphate −2/−1 Phosphate −2/−1 Phosphate −2/−1 Phosphate

b0 R601 b0 A536, V530 b0 R534 b R388 b0 R598 b0 Q611 b Q390, Q393 b N448, R409 b R409, R405 bE1002*, V1001* b Q567, N563 b K846 b0 R628, R622 b Q1030, R1031 b K846 b0 R704 b Q1030, R1031 b0 R621

b0 Rudder b0 Lid b0 Lid b Lobe-1 loop b0 Rudder b0 Switch-2 b Lobe-1 loop b Fork-1 b Fork-1 helix b Switch-3 b Rifampicin-binding helix b DPBB domain b0 DPBB domain b Switch-3 b DPBB domain b0 DPBB domain b Switch-3 b0 DPBB domain

Downstream dsDNA

−1/þ1 Phosphate −1/þ1 Phosphate þ1/þ2 Phosphate, base þ1/þ2 Phosphate þ2, þ2 Base þ2/þ3 Phosphate þ3/þ4 Phosphate þ3/þ4 Phosphate þ4/þ5 Phosphate þ4/þ5 Phosphate þ4/þ5/þ6 Phosphate þ5/þ6 Phosphate þ8/þ9 Phosphate, minor groove þ12/þ13 Phosphate, minor groove þ13/þ14/þ15 Phosphate

b0 R615 b0 R1096 b0 G1092, A1089 b0 K610 b R422 b0 R1096, Y1093 b0 Y1093 b0 Q1441 b0 N1442 b0 R586 b0 K1266 b0 K106 b0 V108* b0 R486, A487* b0 R488

b0 Switch-2 b0 Bridge (F) helix b0 Bridge (F) helix b0 Switch-2 b Fork-2 b0 Bridge (F) helix b0 Bridge (F) helix b0 Clamp C-terminal HLH-loop b0 Clamp C-terminal HLH-loop b0 Rudder b0 Jaw b0 Clamp N-terminal HLH-loop b0 Clamp N-terminal HLH-loop b0 Downstream clamp helix b0 Downstream clamp helix

a

Underlined text is RNA; roman text is DNA position T, and italic text is DNA position NT.

AMPcPP binds to EC at the insertion site of the catalytic center where it stacks on the i1 RNA base and forms a base pair with the iþ1 acceptor template, while its phosphates coordinate two essential Mg2þ ions that play a key role in RNA catalysis (Figures 4(e) and 4(f)). In the Tth EC–AMPcPP complex the substrate binding induces a refolding of the partially disordered trigger loop (an open conformation observed in the holoenzyme structure) into two a- helices (a closed catalytically active intermediate). Together with the adjacent bridge helix present in straight a-helical conformation, this creates a transient three-helical bundle (Figure 4(e)), which constricts the secondary channel and hinders substrate dissociation. This process is common to all multisubunit RNAPs. Other changes in the EC structure include minor repositioning of the b fork-2 toward the active site (by 1 A˚), and a shift of the b0 bridge helix (by 1.5 A˚) to stabilize the b0 trigger helices. It should be noted that in addition to straight, continuously helical conformation, b0 bridge helix can adopt alternative

conformations where its central part has a kink or is flipped out, as observed in the structures of Taq core, Tth holoenzyme, and Tth EC–Gfh1 complex. However, the functional significance of these conformations is unclear. Similarly, in some structures of Tth EC as well as yeast RNAP (Pol II EC), the trigger loop displays several partially folded conformations, which appear to be intermediate between fully folded a-helices and unfolded loop, as depicted in (Figure 4(e)). It was proposed that the intermediate conformations of trigger loop (and bridge helix) play an important role for transcription pausing, termination and for the action of transcription factors binding in the secondary channel (such as GreA/B, DksA, and Gfh1).

Structure of the Catalytic Site At the center of the catalytic site are two essential Mg2þ ions (high-affinity Mg-I and low-affinity Mg-II) that play a key

Molecular Biology | RNA Polymerase Structure, Bacterial role in RNA catalysis. The two Mg2þ ions are surrounded by four highly conserved structural elements of RNAP: the stationary b0 DPBB domain with several conserved adjacent loops and helices including the catalytic loop; the stationary b DPBB domain; and the mobile b0 bridge helix and b0 trigger loop. The high-affinity Mg-I is stably bound in RNAP by a triad of Asp in the b0 catalytic loop (D739, D741, and D743) (Figure 4(f)). It is primarily responsible for activation of RNA 30 -terminal OH group and for coordination of the a-phosphate of AMPcPP in the insertion site. The low-affinity Mg-II is held at the insertion site of Tth EC, proximal to the location of Mg-I (at a distance of 3.9 A˚), mostly by D739 of the b0 catalytic loop via direct interactions. Also, residues E685 and D686 of the b DPBB domain and D741 of the b0 catalytic loop may contribute to stabilization of Mg-II through additional water-mediated contacts. Mg-II coordinates all three rNTP phosphates providing their optimal spatial alignment for nucleophilic attack of activated 30 -OH group on the a-phosphate. Additionally, the b- and g-phosphates of rNTP (or pyrophosphate) are stabilized by contacts with basic residues of the b0 trigger loop (R1239), b0 DPBB domain (R783), b0 clip-1 loop (R1029), and b DPBB domain (R879) via hydrogen bonding and electrostatic interactions (Figure 4(f)). The H1242 of the b0 trigger loop contacts a-phosphate of rNTP (during RNA synthesis) or the phosphate between RNA i1 and iþ1 nucleotides (during pyrophosphorolysis and hydrolysis). This interaction greatly facilitates the nucleophilic attack of the 30 -OH group (or the OH– ion or the water molecule) and is essential for all enzymatic reactions catalyzed by RNAP, that is, RNA synthesis, pyrophosphorolysis, and nucleolytic hydrolysis. To ensure incorporation of correct nucleotides into the RNA, M1238/F1241 residues of the b0 trigger loop and T1088 of the b0 bridge helix make stacking and hydrophobic/van der Waals contacts with rNTP base and DNA base at iþ1 position, respectively, while N737 and R704 of the b0 catalytic loop establish hydrogen bonds with 20 - and 30 -OH groups of the ribose ring of NTP (Figure 4(f)). Tth EC–AMPcPP–Stl complex captures AMPcPP in an intermediate preinsertion site in the active center. In the structure, Stl resides in the pocket formed between the b0 bridge helix/ trigger helix bundle, b fork-2, and the downstream dsDNA (iþ1/iþ4). The Stl binding leads to displacement of the trigger helix from the insertion site, inducing its unfolded (open) conformation. In the absence of the folded trigger helices, the AMPcPP is found in the preinsertion site. It still maintains proper base-pairing with iþ1 DNA base; however, positions of the a-, b-, and g-phosphates together with the bound Mg-II ion are shifted by 2–2.5 A˚ away from Mg-I to a distal location (at a distance of 6.4 A˚), due to a 35 rotation around the C40 –C50 bond of the ribose. These observations suggest that AMPcPP binds to EC first in the preinsertion site in an inactive substrate configuration, and, then, in response to trigger helix folding, it isomerizes into a catalytically competent insertion site. Structural studies revealed that by blocking the folding of the trigger helix, Stl interferes with one of the key steps of nucleotide addition cycle (NAC), thereby inhibiting RNA synthesis. A similar mechanism of action is ascribed to another antibiotic, Tagetitoxin (phytotoxin), which also binds at the base of the secondary channel near the active center and inhibits conformational cycling of trigger loop/trigger helix.

183

NAC and the Mechanism of Translocation The high-resolution structures of substrate-bound Tth EC offer two key findings: 1. The mobile b0 trigger loop/trigger helix is the second essential component of the catalytic center (after the main Mg2þ-binding b0 catalytic loop). It alternates between the substrate-induced catalytically active a-helical conformation (trigger helix) and the inactive unfolded conformation (trigger loop). 2. In the presence of Mg2þ-ion, rNTP can occupy two alternative sites in EC – the catalytically active insertion site stabilized by the trigger helix, and the inactive preinsertion site. These findings led to a model of NAC and RNAP translocation as illustrated in Figure 4(g). At the beginning of NAC, the EC exists in thermodynamic equilibrium, oscillating between the pretranslocated and posttranslocated states. The rNTP– Mg2þ first binds to EC and stabilizes it in a posttranslocated state acting as a ratchet (Figure 4(g), step 1). The binding occurs in a template-dependent manner at the iþ 1 register in the preinsertion site, when the b0 trigger loop is in an open, unfolded conformation. Mg-II, coordinated by the three phosphates of rNTP, is placed at a distal position relative to Mg-I, and the preinsertion inactive intermediate complex is formed (step 2). Next, rNTP induces the folding of the trigger loop into a helix, displacing the b0 bridge helix and b fork-2, and shifting the position of Mg-II/rNTP triphosphate to a site proximal to Mg-I. This leads to isomerization of EC into the catalytically competent, closed insertion state (step 3). After completion of the catalytic nucleotidyl transfer reaction and transcript extension, the EC transforms into closed, pretranslocated state (step 4). Next, the pyrophosphate–Mg-II complex is released from EC causing destabilization and unfolding of the b0 trigger helix. This results in transition of the EC to the pretranslocated state with an open conformation of the trigger loop (step 5). In the last step of NAC, the EC undergoes reversible translocation and returns to the open, posttranslocated ground-state EC capable of binding new rNTP (step 6). Structural, biochemical, and biophysical data indicate that that the energy of NTP hydrolysis is not directly used for EC translocation during elongation. RNAP moves as a Brownian ratchet machine driven forward only by the energy gained from binding the correct substrate. The concerted action of rNTP binding at the catalytic site, the b0 trigger loop folding/unfolding and the step-wise transcript extension makes RNAP movement unidirectional. The correct incoming rNTP works as a stationary pawl in the Brownian ratchet preventing RNAP from slipping backward, whereas the b0 trigger loop together with the bridge helix and b fork-2 act as a reciprocating pawl, pushing RNAP forward relative to the DNA/RNA scaffold. This system of two pawls couples the forward motion of RNAP with efficient incorporation of correct substrate without falling out of register, nucleotide skipping, or RNA slippage.

Future Prospects In the past 5 years a great progress has been made in our understanding of the structure and function of bacterial

184

Molecular Biology | RNA Polymerase Structure, Bacterial

RNAP, especially its complexes with nucleic acids in the presence of substrates and regulatory molecules. However, our knowledge of RNAP mechanics is still fragmented. Many important details concerning the mechanisms of transcription initiation and termination and their regulation remain to be elucidated. The wish list of RNAP structures that would help address many long-standing questions include complete highresolution structures of RPc, RPo and their intermediates, EC in pretranslocated state with resolved nontemplate ssDNA and upstream dsDNA, ECs locked in paused and arrested (backtracked) states, termination and anti-termination complexes and their intermediates, and RNAP complexes together with transcription factors, regulatory RNAs, antibiotics, and other small molecules that modulate RNAP activity.

See also: Molecular Biology: RNA Polymerase II Elongation Control in Eukaryotes; T7 RNA Polymerase.

Further Reading Borukhov S and Nudler E (2008) RNA polymerase: The vehicle of transcription. Trends in Microbiology 16: 126–134. Ho MX, Hudson BP, Das K, et al. (2009) Structures of RNA polymerase–antibiotic complexes. Current Opinion in Structural Biology 19: 715–723.

Hudson BP, Quispe J, Lara-Gonzalez S, et al. (2009) Three-dimensional EM structure of an intact activator-dependent transcription initiation complex. Proceedings of the National Academy of Sciences of the United States of America 106: 19830–19835. Landick R (2006) The regulatory roles and mechanism of transcriptional pausing. Biochemical Society Transactions 34: 1062–1066. Murakami KS and Darst SA (2003) Bacterial RNA polymerases: The whole story. Current Opinion in Structural Biology 13: 31–39. Murakami KS, Masuda S, Campbell EA, et al. (2002) Structural basis of transcription initiation: An RNA polymerase holoenzyme–DNA complex. Science 296: 1285–1290. Opalka N, Brown J, Lane WJ, et al. (2010) Complete structural model of Escherichia coli RNA polymerase from a hybrid approach. PLoS Biology 8: e1000483. Revyakin A, Liu C, Ebright RH, and Strick TR (2006) Abortive initiation and productive initiation by RNA polymerase involve DNA scrunching. Science 314: 1139–1143. Saecker RM, Record MT Jr., and deHaseth PL (2011) Pathway of prokaryotic transcription initiation: Promoter binding, isomerization to initiation-competent open complexes and initiation of RNA synthesis. Journal of Molecular Biology 412: 754–771. Svetlov V and Nudler E (2009) Macromolecular micromovements: How RNA polymerase translocates. Current Opinion in Structural Biology 19: 701–707. Tagami S, Sekine S, Kumerevel T, et al. (2010) Crystal structure of bacterial RNA polymerase bound with a transcription inhibitor protein. Nature 468: 978–984. Vassylyev DG (2009) Elongation by RNA polymerase: A race through roadblocks. Current Opinion in Structural Biology 19: 691–700. Vassylyev DG, Sekine S, Laptenko O, et al. (2002) Crystal structure of a bacterial RNA polymerase holoenzyme at 2.6 A resolution. Nature 417: 712–719. Vassylyev DG, Vassylyeva MN, Perederina A, et al. (2007) Structural basis for transcription elongation by bacterial RNA polymerase. Nature 448: 157–162. Vassylyev DG, Vassylyeva MN, Zhang J, et al. (2007) Structural basis for substrate loading in bacterial RNA polymerase. Nature 448: 163–168. Zhang G, Campbell EA, Minakhin L, et al. (1999) Crystal structure of Thermus aquaticus core RNA polymerase at 3.3 A resolution. Cell 98: 811–824.