RNA Polymerase: Structural Similarities Between Bacterial RNA Polymerase and Eukaryotic RNA Polymerase II

RNA Polymerase: Structural Similarities Between Bacterial RNA Polymerase and Eukaryotic RNA Polymerase II

doi:10.1006/jmbi.2000.4309 available online at http://www.idealibrary.com on J. Mol. Biol. (2000) 304, 687±698 REVIEW ARTICLE RNA Polymerase: Struc...

861KB Sizes 1 Downloads 64 Views

doi:10.1006/jmbi.2000.4309 available online at http://www.idealibrary.com on

J. Mol. Biol. (2000) 304, 687±698

REVIEW ARTICLE

RNA Polymerase: Structural Similarities Between Bacterial RNA Polymerase and Eukaryotic RNA Polymerase II Richard H. Ebright* Howard Hughes Medical Institute Waksman Institute and Department of Chemistry Rutgers University Piscataway NJ 08854, USA

Bacterial RNA polymerase and eukaryotic RNA polymerase II exhibit striking structural similarities, including similarities in overall structure, relative positions of subunits, relative positions of functional determinants, and structures and folding topologies of subunits. These structural similarities are paralleled by similarities in mechanisms of interaction with DNA. # 2000 Academic Press

Keywords: RNA polymerase; RNA polymerase II; open complex; elongation complex; transcription

Introduction Transcription is the ®rst step in gene expression and is the step at which most regulation of gene expression occurs. RNA polymerase (RNAP) is the enzyme responsible for transcription and is the target, directly or indirectly, of most regulation of transcription.1 ± 3 RNAP is conserved in all living organisms.4 ± 6 Thus, bacterial RNAP, archaeal RNAP, and eukaryotic RNAP I, RNAP II, and RNAP III, are members of a conserved protein family, termed the `multisubunit RNAP family.' 4 ± 6 Members of this protein family contain a conserved subunit of 160 kDa (b0 in bacterial RNAP; A in archaeal RNAP; RPA1, RPB1, and RPC1 in eukaryotic RNAP I, II, and III), a conserved subunit of 150 kDa (b in bacterial RNAP; B in archaeal RNAP; RPA2, RPB2, and RPC2 in eukaryotic RNAP I, II, and III), a conserved subunit of 35 kDa (aI in bacterial RNAP; D in archaeal RNAP; RPC5 in eukaryotic RNAP I and RNAP III; RPB3 in eukaryotic RNAP II), a conserved subunit of 10-35 kDa (aII in bacterial RNAP; L in archaeal RNAP; RPC9 in eukaryotic RNAP I and RNAP III; RPB11 in eukaryotic RNAP II), and a conserved subunit of 6 kDa (o in bacterial RNAP; K in archaeal RNAP; RPB6 in eukaryotic RNAP I, II, Abbreviations used: RNAP, RNA polymerase; aNTD, a subunit N-terminal domain; aCTD, a subunit C-terminal domain. E-mail address of the corresponding author: [email protected] 0022-2836/00/050687±12 $35.00/0

and III) (Figure 1).4 ± 6 Bacterial RNAP contains only these conserved subunits. Archaeal and eukaryotic RNAP contain these conserved subunits, and also contain additional subunits. Structures of bacterial RNA polymerase and eukaryotic RNA polymerase II A milestone in our understanding of the multisubunit RNAP family was reached in 1999, with the determination of a crystallographic structure of bacterial RNAP: i.e., the structure of Thermus aquaÊ resolution (PDB ticus RNAP core enzyme at 3.3 A accession code 1DDQ).7 Figure 2 shows three views of the structure. Bacterial RNAP has dimenÊ  100 A Ê  100 A Ê and has a sions of 150 A shape reminiscent of a crab claw, with two ``pincers'' de®ning a central cleft (Figure 2(a)). The cleft Ê , suf®cient to accommohas a diameter of 25 A date a double-stranded nucleic acid, and has the active-center Mg2 ‡ at its base (white sphere in Figure 2(b)). The b0 subunit makes up one pincer and part of the base of the cleft. The b subunit makes up the other pincer and part of the base of the cleft. The aI and aII subunits are located distal to the cleft. aI and aII have identical sequences, but their locations within RNAP, and their interactions within RNAP, differ: aI is located closer to the cleft and interacts with b; aII is located farther from the cleft and interacts with b0 . Each a subunit consists of two domains: an N-terminal domain (aNTD) responsible for interaction with b or b0 , and a C-terminal domain (aCTD) responsible for protein# 2000 Academic Press

688

Figure 1. Conserved subunits of bacterial RNAP, archaeal RNAP, eukaryotic RNAP I, eukaryotic RNAP II, and eukaryotic RNAP III.4 ± 6

DNA interactions with upstream promoter DNA and protein-protein interactions with upstream activators and repressors.8 aNTD is connected to aCTD, and thus to the remainder of RNAP, through a long, ¯exible linker.8 In the crystallographic structure, aCTD is disordered (presumably re¯ecting the fact that the linker allows aCTD to occupy different positions relative to the remainder of RNAP in different molecules in the crystal lattice).7 Nevertheless, it is possible to estimate the range of possible positions of aCTD relative to the remainder of RNAP, based on the positions of the C-terminal residues of aNTDI and aNTDII, the length of the linker, and the diameter of aCTD.9 The remaining subunit, o, is located distal to the cleft, at the base of the pincer formed by b0 .6,7 A second milestone in our understanding of the multisubunit RNAP family was reached in 2000, with the determination of a crystallographic structure of eukaryotic RNAP II: i.e., the Ca structure of yeast RNAP II 4/7 (a derivative of yeast RNAP II lacking non-conserved Ê resolution subunits RPB4 and RPB7) at 3.0 A (PDB accession code 1EN0).10 Figure 3 shows three views of the structure. RNAP II has a shape reminiscent of a crab claw, with two pincers de®ning a central cleft, and thus is similar in shape to bacterial RNAP (compare Figures 2 and 3).7,10 The relative positions of the conserved subunits within yeast RNAP match those in bacterial RNAP (compare Figures 2 and 3).6,7,10 Thus, the counterpart of b0 (RPB1) forms one pincer and part of the base of the cleft; the counterpart of b (RPB2) forms the other pincer and part of the base of the cleft; the counterparts of aI (RPB3) and aII (RPB11) are located distal to the cleft; and the counterpart of o (RPB6) is located at the base of the pincer formed by the counterpart of b0 . The non-conserved subunits (RPB5, RPB8, RPB9, RPB10, and RPB12) are located on the periphery of the structure and, with one exception (RPB5; see below), are located distal to the cleft (Figure 3). In addition to the overall structural similarity of bacterial RNAP and yeast RNAP II, there also are equivalences in the locations of important functional determinants:

RNA Polymerase

(i) Each enzyme contains an equivalently positioned, and identically coordinated, activecenter Mg2 ‡ (white spheres in Figures 2(b) and 3(b)). (ii) Each enzyme contains an equivalently Ê long, 15 A Ê wide channel that positioned 25 A connects the ¯oor of the active-center cleft with the surface of RNAP to one side of the active-center cleft, and that mediates exit of the nascent RNA during elongation (``exit channel'' in Figures 2(b) and 3(b)).7,9 ± 11 The walls of this channel are formed by the pincer composed of the largest subunit (b0 in bacterial RNAP, RPB1 in RNAP II) and a ¯ap-like feature of the second-largest subunit (b in bacterial RNAP, RPB2 in RNAP II; ``¯ap'' in Figures 2(a),(b) and 3(a),(b)). (iii) Each enzyme contains an equivalently Ê long, 10 A Ê wide tunnel that positioned 20 A connects the ¯oor of the active-center cleft with the surface of RNAP opposite the mouth of the activecenter cleft, and that mediates entry of NTPs to the active center during elongation, extrusion of the 30 end of RNA during backtracking, and release of the cleaved 30 end of RNA during editing (``secondary channel'' in Figures 2(b),(c) and 3(b),(c)).7,10 ± 12 (iv) Each enzyme contains an equivalently positioned activation target required for response to a subset of transcriptional activators (located within aNTDI in bacterial RNAP, and within RPB3 in yeast RNAP II; ``AT'' in Figures 2(a),(b) and 3(a),(b)).13 ± 15 (v) The largest subunit of each enzyme (b0 in bacterial RNAP, RPB1 in RNAP II) contains an equivalently positioned C-terminal segment involved in transcriptional activation (see correction to structure of RNAP II in ref. 6; ``C-ter'' in Figures 2 and 3).16 ± 18 In each enzyme, this segment threads through an opening in the smallest conserved subunit (o in bacterial RNAP, RPB6 in RNAP II) and extends away from the main mass of RNAP.6 In RNAP II, this segment serves as the point of departure for the ``CTD,'' an extension of the largest subunit involved in transcriptional regulation, RNA processing, and RNA localization.18,19 The similarity between bacterial RNAP and yeast RNAP II extends to the three-dimensional structures and folding topologies of the individual subunits. Indeed, it is at this level that the similarity is most striking. The largest subunit ±b0 in the case of bacterial RNAP, RPB1 in the case of RNAP II ±consists of a central mass and two smaller lobes, numbered 1 and 2, based on their order within the primary structure (Figure 4(a)). The central masses of b0 and RPB1 have essentially identical folds; thus, the central mass of b0 can be superimposed on the central mass of RPB1, with 356 residues aligning and an Ê (an rmsd comparable to the sum of rmsd of 2.3 A error in the coordinate sets; Figure 4(a)). The lobes of b0 and RPB1 also have essentially identical folds; thus, lobe 1 of b0 can be superimposed on lobe 1 of

RNA Polymerase

689

Ê resolution.7 b0 is in orange; b is in green; aNTDI is in cyan; Figure 2. Structure of T. aquaticus RNAP at 3.3 A aNTDII is in blue; and o is in gray. The active-center Mg2‡ is shown as a white sphere (visible in (b)). The active-center-proximal and active-center-distal openings of the secondary channel are shown as black circles (visible in (b) and (c), respectively). The activation target within aNTDII 13,14 is shown in yellow. Atomic coordinates were obtained from the PDB (accession code 1DDQ); stereodiagrams were prepared using INSIGHT II (MSI, San Diego, CA). (a) ``Upstream'' face. (b) ``Top'' face (view into active-center cleft; 90 rotation about x-axis relative to (a)). (c) ``Downstream'' face ( 90 rotation about x-axis relative to (b)).

RPB1, with 43 residues aligning and an rmsd of Ê (Figure 4(b)), and lobe 2 of b0 can be super2.4 A imposed on lobe 2 of RPB1, with 61 residues alignÊ (Figure 4(c)). Lobes 1 and ing and an rmsd of 2.4 A 2 are oriented differently relative to the central mass in the crystallographic structures of bacterial RNAP and yeast RNAP II, with the orientations differing by 10  rotations about points of attach-

ment of lobes to the central mass (imperfectly aligned dashed ovals in Figure 4(a)). The differences in orientation of lobes 1 and 2, together with the small numbers of covalent connections between lobes and the central mass (one covalent connection for lobe 1, two covalent connections for lobe 2), suggests that there may be ¯exibility in points of attachment of lobes to the central mass, such

690

RNA Polymerase

Ê resolution.10 RPB1 is in orange; RPB2 is in green; RPB3 is in Figure 3. Structure of yeast RNAP II 4/7 at 3.0 A cyan; RPB11 is in blue; and RPB6 is in gray. Non-conserved subunits RPB5, RPB8, RPB9, RPB10, and RPB12 are in red, aqua, violet, magenta, and pink, respectively. The active-center Mg2‡ is shown as a white sphere (visible in (b)). The active-center-proximal and active-center-distal openings of the secondary channel are shown as black circles (visible in (b) and (c), respectively). The activation target within RPB315 is shown in yellow. Ca coordinates were obtained from the PDB (accession code 1EN0), and full-backbone and Cb coordinates were generated using MAXSPROUT (www.ebi.ac.uk/dali/maxsprout);33 stereodiagrams were prepared using INSIGHT II (MSI, San Diego, CA). (a) ``Upstream'' face. (b) ``Top'' face (view into active-center cleft; 90 rotation about x-axis relative to (a)). (c) ``Downstream'' face ( 90 rotation about x-axis relative to (b)).

that lobes might occupy different locations in different crystal lattices or different functional complexes (see below). The structural similarity between b0 and RPB1 is remarkably high, higher, for example, than the structural similarity among DNA polymerases.20 The structural similarity of b0 and RPB1 is substan-

tially higher than expected based on sequence comparisons. The structural similarity comprises fully one-half of the structurally characterized residues of b0 (Figure 4 and data not shown); in contrast, sequence similarity is con®ned to small regions and comprises only one-third of the structurally characterized residues of b0 .4,5

RNA Polymerase

691

Figure 4. Structural similarity between bacterial RNAP b0 (blue) and eukaryotic RNAP II RPB1 (red). (a) Superimposition of residues of the central mass of T. aquaticus RNAP b0 (residues 624-651, 699-794, 875-952, 1017-1102, 11941233, 1252-1266, and 1328-1374; residues numbered as in sequence ®le CAB65466, not as in structure ®le 1DDQ) on corresponding residues of yeast RNAP II RPB1. The active-center Mg2‡ in T. aquaticus RNAP is shown as a green sphere; the active-center Mg2‡ in yeast RNAP II is shown as an orange sphere (barely visible, behind green sphere). Positions of lobes 1 and 2 of b0 and RPB1 are shown as dashed ovals. (b) Superimposition of residues of lobe 1 of b0 (residues 535-623) on corresponding residues of RPB1. (c) Superimposition of residues of lobe 2 of b0 (residues 9531016) on corresponding residues of RPB1. DALI (www.ebi.ac.uk/dali)34 alignment statistics are summarized to right in each panel.

An even higher degree of structural similarity is observed for the second-largest subunit: b in the case of bacterial RNAP, RPB2 in the case of RNAP II. The second-largest subunit consists of a large central mass and four smaller lobes, numbered 1, 2, 3, and 4, based on their order within the primary structure (Figure 5(a)). The central mass and lobes 2, 3, and 4 of b can be superimposed on the corresponding structural elements of RPB2, with 413 Ê residues aligning and an rmsd of 2.1 A (Figure 5(a)). In addition, lobe 1 of b can be superimposed on lobe 1 of RPB2, with 128 residues Ê (Figure 5(b)). The aligning and an rmsd of 2.7 A orientation of lobe 1 relative to the remainder of the second-largest subunit differs in the crystallographic structures of bacterial RNAP and yeast

RNAP II (dashed ovals in Figure 5(a)), suggesting that there may be ¯exibility in the point of attachment between lobe 1 and the remainder of the second-largest subunit (see below). As with the largest subunit, the structural similarity for the second-largest subunit is substantially higher than expected based on sequence comparisons. Structural similarity spans fully two-thirds of the length of b (Figure 5 and data not shown); in contrast, sequence similarity is con®ned to small regions and spans only one-quarter to one-third the length of b.4,5 Structural similarity also is observed for the small conserved subunits: aI, aII, and o in bacterial RNAP; and RPB3, RPB11, and RPB6 in RNAP II. Subdomain 1 of aNTDI can be superimposed on

692

RNA Polymerase

Figure 5. Structural similarity between bacterial RNAP b (blue) and eukaryotic RNAP II RPB2 (red). (a) Superimposition of residues of the central mass (residues 136-143, 325-334, 395-413, 430-474, 534-593, 627-637, 659-700, 833-921, and 967-999), lobe 2 (residues 19-135 and 335-394), lobe 3 (residues 701-722, 739-757, and 787-832), and lobe 4 (residues 1031-1056) of T. aquaticus RNAP b on corresponding residues of yeast RNAPII RPB2. The active-center Mg2‡ in T. aquaticus RNAP is shown as a green sphere; the active-center Mg2‡ in yeast RNAP II is shown as an orange sphere. Positions of lobe 1 of b and lobe 1 of RPB2 are shown as dashed ovals. (b) Superimposition of residues of lobe 1 of b (residues 535-623) on corresponding residues of RPB2. DALI34 alignment statistics are summarized at right in each panel.

subdomain 1 of RPB3 (subdomains de®ned as in refs 15 and 21), with 94 residues aligning and an Ê (Figure 6(a) and data not shown), rmsd of 2.8 A and subdomain 2 of aNTDI can be superimposed on subdomain 2 of RPB3, with 90 residues aligning Ê (Figure 6(b)). Subdomain 1 and an rmsd of 2.5 A of aNTDII can be superimposed on RPB11, with 92 Ê (Figure 6(a) residues aligning and an rmsd of 2.8 A and data not shown). o can be superimposed on RPB6, with 47 residues aligning and an rmsd of Ê (Figure 7).6 The observed structural simi2.2 A larities for the small conserved subunits are very substantially higher than expected from sequence comparisons. Structural similarity between aNTDI and RPB3 spans nearly the entire length of aNTDI, structural similarity between aNTDII and RPB11 spans one-half the length of aNTDII, and structural similarity between o and RPB6 spans one-half the length of o (Figures 6 and 7 and data not shown); in contrast, sequence similarities are very limited6,15,21 (so limited in the case of o that sequence similarity remained undetected until recently).6 Structures of transcription complexes of bacterial RNA polymerase and eukaryotic RNA polymerase II RNAP carries out a complex series of reactions during transcription initiation:22 (i) RNAP, together with initiation factors (s in the case of bacterial

RNAP;22 transcription factors IIB, IID (or TBP), IIE, IIF, and IIH, in the case of eukaryotic RNAP II23), binds to promoter DNA, to yield an RNAP-promoter closed complex; (ii) RNAP melts approximately 14 nucleotides of DNA surrounding the transcription start site, rendering accessible the genetic information in the template strand of DNA, to yield an RNAP-promoter open complex; (iii) RNAP begins synthesis of RNA as an RNAP-promoter initial transcribing complex; and (iv) upon synthesis of an RNA product of a critical threshold length of 9-11 nt, RNAP breaks its interactions with promoter DNA, breaks or at least weakens its interactions with initiation factors, and begins to translocate along DNA, processively synthesizing RNA as an RNAP-DNA elongation complex. Site-speci®c protein-DNA photocrosslinking has been used to map protein-DNA interactions in the bacterial RNAP-promoter open complex (Figure 8(a))9 and the bacterial RNAP-DNA elongation complex.11 In conjunction with the structure of bacterial RNAP,7 the results de®ne the orientation of DNA relative to RNAP, de®ne the position and rotational phasing of the doublestranded DNA segment downstream of the melted region (``downstream duplex''), de®ne the positions of the template and non-template strands of the melted region (``transcription bubble''), constrain the position and rotational phasing of the double-stranded DNA segment upstream of the

RNA Polymerase

693

Figure 6. Structural similarity between bacterial RNAP aNTDI and aNTDII (blue and green) and eukaryotic RNAP II RPB3 and RPB11 (red and orange). (a) Superimposition of residues of subdomain 1 of aNTDI (residues 6-50 and 170-231) and subdomain 1 of aNTDII (residues 6-50 and 170-231) of T. aquaticus RNAP on corresponding residues of yeast RNAP II RPB3 and RPB11. Positions of subdomain 2 of aNTDI and subdomain 2 of RPB3 are shown as dashed ovals. (b) Superimposition of residues of subdomain 2 of aNTDI (residues 51-169) on corresponding residues of RPB3. The location of the activation target within bacterial RNAP aNTDI (residues 155-158; ``AT'' in Figure 2(a),(b))13,14 is shown as a green sphere; the location of the activation target within RNAPII RPB3 is shown as two orange spheres (residues 92-95 and 159-162; ``AT'' in Figure 3(a),(b)).15 DALI34 alignment statistics are summarized to right in each panel.

melted region (``upstream duplex''), and permit construction of structural models of the open and elongation complexes (Figures 8(a) and 9).9,11 In both the RNAP-promoter open complex and the RNAP-DNA elongation complex, the downstream duplex binds deep within the active-center cleft (Figure 9(b),(c)).9,11 The downstream end of the transcription bubble also binds deep within the active-center cleft, with position ‡1 of the template strand being located Ê on the ¯oor of the active-center cleft, 20 A from the active-center Mg2 ‡ (Figure 9(b),(c)).9,11 The remainder of the transcription bubble rises from the ¯oor of the active-center cleft, along an axis nearly perpendicular to the downstream duplex, with the upstream end of the transcripÊ from the ¯oor tion bubble being located 60 A of the active-center cleft (Figure 9(b),(c)).9,11 The template strand of the transcription bubble is positioned in a manner that can accommodate formation of a 7-9 bp A-form RNA:DNA hybrid with the RNA product, with the RNA 30 end being adjacent to the active-center Mg2 ‡ and secondary channel, and with the RNA 50 end directed toward the exit channel (see Figures 2(b)

and 9(b) for locations of secondary channel and exit channel).9,11 In the RNAP-promoter open complex (the only complex for which data for positions upstream of position 15 have been reported), positions 30 to 12 of the upstream duplex bind within a shallow channel that is distinct from the active-center cleft and that lies above the exit channel (Figure 9(a),(b); see Figure 2(b) for location of exit channel),9 and positions 93 to 40 bend toward RNAP and interact with aCTD.9 Upon formation of open and elongation complexes, the tips of the two pincers of RNAP are likely to move closer to DNA, narrowing the active-center cleft, and clamping DNA in place. Lobe 1 of b0 , which appears to be ¯exibly connected to the remainder of RNAP (Figure 4(a),(b); see above), forms the tip of one pincer of RNAP and is positioned such that it could move closer to the template strand of the transcription bubble (Figure 9(a),(b)). (Lobe 1 of b0 contains a short loop, termed the ``rudder,`` that has been proposed to make interactions with the upstream separation point of the template-strand:RNA hybrid.7,9,11 It is attractive to speculate that establishment of these

Figure 7. Structural similarity between bacterial RNAP o (blue) and eukaryotic RNAPII RPB6 (red). Superimposition of residues 9-81 of o of T. aquaticus RNAP on corresponding residues of yeast RNAPII RPB6. DALI34 alignment statistics are summarized on right.

694

RNA Polymerase

Figure 8 (legend opposite)

interactions triggers and/or stabilizes re-positioning of lobe 1 of b0 .) Lobe 1 of b, which likewise appears to be ¯exibly connected to the remainder of RNAP (Figure 5(a),(b); see above), forms part of the tip of the other pincer and is positioned such

that it could move closer to the downstream duplex and the non-template strand of the transcription bubble (Figure 9(b),(c)).11,24 (The non-template strand of the transcription bubble binds within a narrow groove formed by the interface

695

RNA Polymerase

between lobe 1 of b and lobe 2 of b (Figure 9(b)).9,11 It is attractive to speculate that binding of the nontemplate strand within this groove triggers and/or stabilizes re-positioning of lobe 1 of b.) The crosslinking results for the bacterial RNAP-promoter open complex also de®ne locations of segments of the initiation factor s relative to promoter DNA (Figure 8(a)).9 s region 4 crosslinks in the 35 element, s region 3 crosslinks between the 35 and 10 elements, and s region 2 crosslinks in the 10 element (Figure 8(a)).9 Incorporation of these results into the structural model of the open complex indicates that s region 3 and region 4 are located outside of the active-center cleft, in proximity to lobe 3 of b, and that s region 2 ``caps'' DNA in the active-center cleft, being located directly above the active-center cleft, in proximity to lobe 1 of b0 and lobe 2 of b (Figure 9(a),(b); ref. 9; see also refs 25, 26). Site-speci®c protein-DNA photocrosslinking also has been used to map protein-DNA interactions in the eukaryotic RNAP II-promoter open complex (Figure 8(b))27 and the eukaryotic RNAP II-DNA elongation complex.28 The results establish that eukaryotic RNAP II interacts with the downstream duplex in the same manner as does bacterial RNAP. The key result is the observation that the ‡10 region of DNA can be crosslinked to RPB5 (Figure 8(b)),27,28 a subunit located on the ¯oor of the active-center cleft of RNAP II (red in Figure 3(b),(c)).10 This result permits the conclusion that the ‡10 region of DNA is located on the ¯oor of the active-center cleft, in proximity to RPB5. Mechanistic considerations permit the conclusion that position ‡1 of DNA is located on the ¯oor of the active-center cleft, in proximity to the activecenter Mg2‡. Taken together, these conclusions de®ne the location and orientation of the downstream duplex relative to RNAP II: i.e., the downstream duplex is located on the ¯oor of the activecenter cleft, with its helix axis oriented along the y-axis in the view in Figure 3(b) (compare location and orientation of downstream duplex relative to bacterial RNAP in Figure 9(b)). Low-resolution

structural analysis of a tailed-template RNAP IIDNA elongation complex yields a completely compatible model.10,12,29 The site-speci®c protein-DNA photocrosslinking results for the eukaryotic RNAP II-promoter open complex also de®ne positions of the initiation factors IIB, IID (TBP), IIE, IIF, and IIH relative to promoter DNA (Figure 8(b)).27 TBP and IIB bind to DNA in a location formally equivalent to that bound by region 4 of s (Figure 8). Transcription factor IIF binds to DNA in a location formally equivalent to that bound by regions 2 and 3 of s (Figure 8). Transcription factors IIE and IIH are involved in ATP-dependent promoter melting,23,27 a process that has no counterpart in bacterial transcription initiation;22 IIE and IIH bind to DNA in the transcription-bubble region and downstream duplex, in locations where there is no counterpart initiation factor in the bacterial RNAP-promoter open complex (Figure 8). Analogous site-speci®c protein-DNA photocrosslinking analyses have been performed with archaeal RNAP30,31 and eukaryotic RNAP III.32 The results establish that these additional members of the multisubunit RNAP family also interact with the downstream duplex in the same manner as does bacterial RNAP.

Prospect The analysis in this review establishes that structural similarity between bacterial RNAP and eukaryotic RNAP II involves not only similarity in overall structural organization (Figures 2 and 3),7,10 but also detailed similarity in folding topologies of subunits (Figures 4-7). The structural similarity between bacterial RNAP and eukaryotic RNAP II is substantially more extensive than has been reported previously and substantially more extensive than had been anticipated based on sequence analysis. The extent of the structural similarity implies that results of structural and mechanistic studies of bacterial RNAP (the smallest and by far best characterized member of the multisubunit

Figure 8. Structures of transcription complexes: site-speci®c protein-DNA photocrosslinking data for RNAP-promoter open complexes. In each panel, results for the promoter non-template strand are shown above the sequence; results for the promoter template strand are shown beneath the sequence. Crosslinks are indicated by bars (strong, consistently reproducible, crosslinks by ®lled bars; weak crosslinks by open bars). Phosphates at which crosslinking agents were incorporated are indicated by asterisks. The transcription start site and promoter sequence elements are indicated by shading, and the transcription bubble (melted region) is indicated by having nucleotides of the non-template strand raised and nucleotides of the template strand lowered. (a) Site-speci®c protein-DNA photocrosslinking data for the bacterial RNAP-promoter open complex (ref 9; N. Naryshkin and R.H.E., unpublished results; positions 40 to ‡25 of the lacUV5(ICAP) promoter). Crosslinks to b0 (1-580), b0 (545-878), b0 (821-1407), b(1-235), b(235-643), b(643-989), b(951-1342), aNTDI (none), aCTDI, aNTDII (none), and aCTDII, of RNAP are in black; crosslinks to s region 1 (none), s region 2, s region 3, and s region 4, are in blue. Experiments were performed with, and residues are numbered as in, Escherichia coli RNAP and E. coli s70. (b) Site-speci®c protein-DNA photocrosslinking data for the eukaryotic RNAP II-promoter open complex (positions 40 to ‡25 of the adenovirus major late promoter).27 Crosslinks to subunits of RNAP II are in black; crosslinks to subunits of transcription factors IID (TBP), IIB, and IIF are in blue; crosslinks to subunits of transcription factor IIE and IIH, which mediate an ATP-dependent promoter-melting step with no counterpart in bacterial transcription,23,27 are in green and red.

696

RNA Polymerase

Figure 9. Structures of transcription complexes: model for the structure of the bacterial RNAP-promoter open complex (positions 40 to ‡20).9 b0 is in orange; b is in green; aNTDI is in cyan; aNTDII is in blue; o is in gray; the DNA template strand is in gray, and the DNA non-template strand is in pink. The active-center Mg2‡ is shown as a white sphere (visible in (b)). The active-center-proximal and active-center-distal openings of the secondary channel are shown as black circles (visible in (b) and (c), respectively). The activation target within aNTDI 13,14 is shown in yellow. In the RNAP-promoter open complex, the transcription initiation factor s would interact with positions 37 to 3 of promoter DNA, with s region 4 interacting with positions 37 to 31, s region 3 interacting with positions 31 to 7, and s region 2 interacting with positions to 17 to 3 (Figure 8(a)).9 In the RNAP-DNA elongation complex, positions ‡1 to 6 of the DNA template strand would be engaged in an A-form RNA:DNA hybrid, with the RNA 30 end being adjacent to the active-center Mg2 ‡ and secondary channel, and with the RNA 50 end directed toward the exit channel.9,11 (a) ``Upstream'' face. (b) ``Top'' face (view into active-center cleft; 90 rotation about xaxis relative to (a)). (c) ``Downstream'' face ( 90 rotation about x-axis relative to (b)).

RNAP family) are likely to be directly relevant to understanding eukaryotic RNAP II. Priorities for future work include: (i) determination of structures of transcription initiation and

elongation complexes; (ii) elucidation of mechanistic and kinetic aspects of transcription initiation and elongation; and (iii) determination of the functions of the subunits of eukaryotic RNAP II that

RNA Polymerase

lack counterparts in bacterial RNAP. Progress in all areas is likely to be rapid.

Acknowledgments This work was supported by NIH grants GM41376 and GM53665 and a Howard Hughes Medical Institute Investigatorship.

References 1. Chamberlin, M. (1976). RNA polymerase ± an overview. In RNA Polymerase (Losick, R. & Chamberlin, M., eds), pp. 17-67, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 2. von Hippel, P. (1998). An integrated model of the transcription complex in elongation, termination, and editing. Science, 281, 660-665. 3. Nudler, E. (1999). Transcription elongation: structural basis and mechanisms. J. Mol. Biol. 288, 1-12. 4. Young, R. (1991). RNA polymerase II. Annu. Rev. Biochem. 60, 689-715. 5. Sentenac, A., Riva, M., Thuriaux, P., Buhler, J.-M., Treich, I., Carles, C., Werner, M., Ruet, A., Huet, J., Mann, C., Chiannilkulchai, N., Stettler, S. & Mariotte, S. (1992). Yeast RNA polymerase subunits and genes. In Transcriptional Regulation (McKnight, S. & Yamamoto, K., eds), vol. 1, pp. 27-54, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 6. Minakhin, L., Bhagat, S., Brunning, A., Campbell, E., Darst, S., Ebright, R. & Severinov, K. (2001). Bacterial RNA polymerase subunit and eukaryotic RNA polymerase subunit RPB6 are sequence, structural, and functional homologs and promote RNA polymerase assembly. Proc. Natl Acad. Sci. USA, In the press. 7. Zhang, G., Campbell, E., Minakhin, L., Richter, C., Severinov, K. & Darst, S. (1999). Crystal structure of Ê Thermus aquaticus core RNA polymerase at 3.3 A resolution. Cell, 98, 811-824. 8. Busby, S. & Ebright, R. (1994). Promoter structure, promoter recognition, and transcription activation in prokaryotes. Cell, 79, 743-746. 9. Naryshkin, N., Revyakin, A., Kim, Y., Mekler, V. & Ebright, R. (2000). Structural organization of the RNA polymerase-promoter open complex. Cell, 101, 601-611. 10. Cramer, P., Bushnell, D., Fu, J., Gnatt, A., MaierDavis, B., Thompson, N., Burgess, R., Edwards, A., David, P. & Kornberg, R. (2000). Architecture of RNA polymerase II and implications for the transcription mechanism. Science, 288, 640-649. 11. Korzheva, N., Mustaev, A., Kozlov, M., Malhotra, A., Nikiforov, V., Goldfarb, A. & Darst, S. (2000). A structural model of transcription elongation. Science, 289, 619-625. 12. Fu, J., Gnatt, A., Bushnell, D., Jensen, G., Thompson, N., Burgess, R., David, P. & Kornberg, R. (1999). Ê resolution. Cell, 98, Yeast RNA polymerase II at 5 A 799-810. 13. Niu, W., Kim, Y., Tau, G., Heyduk, T. & Ebright, R. (1996). Transcription activation at Class II CAPdependent promoters: two interactions between CAP and RNA polymerase. Cell, 87, 1123-1134.

697 14. Busby, S. & Ebright, R. (1999). Transcription activation by catabolite activator protein (CAP). J. Mol. Biol. 293, 199-213. 15. Tan, Q., Linask, K., Ebright, R. & Woychik, N. (2000). Activation mutants in yeast RNA polymerase II subunit RPB3 provide evidence for a conserved activation target in eukaryotes and bacteria. Genes Dev. 14, 339-348. 16. Miller, A. D. W., Ebright, R. & Rothman-Denes, L. (1997). RNA polymerase beta subunit: a target for DNA binding-independent activation. Science, 275, 1655-1657. 17. Xiao, H., Friesen, J. & Lis, J. (1994). A highly conserved domain of RNA polymerase II shares a functional element with acidic activation domains of upstream transcription factors. Mol. Cell. Biol. 14, 7507-7516. 18. Carlson, M. (1997). Genetics of transcriptional regulation in yeast: connections to the RNA polymerase II CTD. Annu. Rev. Cell Dev. Biol. 13, 1-23. 19. Hirose, Y. & Manley, J. (2000). RNA polymerase II and the integration of nuclear events. Genes Dev. 14, 1415-1429. 20. Steitz, T. (1999). DNA polymerases: structural diversity and common mechanisms. J. Biol. Chem. 274, 17395-17398. 21. Zhang, G. & Darst, S. (1998). Structure of the Escherichia coli RNA polymerase a subunit amino-terminal domain. Science, 281, 262-266. 22. Record, M. T., Reznikoff, W., Craig, M., McQuade, K. & Schlax, P. (1996). Escherichia coli RNA polymerase (Es70), promoters, and the kinetics of the steps of transcription initiation. In Escherichia coli and Salmonella (Neidhart, F. C., ed.), pp. 792-820, ASM Press, Washington, D.C. 23. Orphanides, G., Lagrange, T. & Reinberg, D. (1996). The general transcription factors of RNA polymerase II. Genes Dev. 10, 2657-2683. 24. Nechaev, S., Tchlenov, M. & Severinov, K. (2000). Dissection of two hallmarks of the open promoter complex by mutation in an RNA polymerase core subunit. J. Biol. Chem. 275, 25516-25522. 25. Arthur, T., Anthony, L. & Burgess, R. (2000). Mutational analysis of b0 260-309, a s70 binding site located on Escherichia coli core RNA polymerase. J. Biol. Chem. 275, 23113-23119. 26. Datwyler, S. & Meares, C. (2000). Protein-protein interactions mapped by arti®cial proteases: where sigma factors bind to RNA polymerase. Trends Biochem. Sci. 25, 408-414. 27. Kim, T.-K., Ebright, R. & Reinberg, D. (2000). Mechanism of ATP-dependent promoter melting by transcription factor IIH. Science, 288, 1418-1421. 28. Wooddell, C. & Burgess, R. (2000). Topology of yeast RNA polymerase II subunits in transcription elongation complexes studied by photoaf®nity crosslinking. J. Biochem. 39, 13405-13421. 29. Poglitsch, C., Meredith, G., Gnatt, A., Jensen, G., Chang, W.-H., Fu, J. & Kornberg, R. (1999). Electron crystal structure of an RNA polymerase II transcription elongation complex. Cell, 98, 791-798. 30. Bartlett, M., Thomm, M. & Geiduschek, E. (2000). The orientation of DNA in an archaeal transcription initiation complex. Nature Struct. Biol. 7, 782-785. 31. Bell, S. & Jackson, S. (2000). Charting a course through RNA polymerase. Nature Struct. Biol. 7, 703705. 32. Bartholomew, B., Durkovich, D., Kassavetis, G. & Geiduschek, E. (1993). Orientation and topography

698

RNA Polymerase

of RNA polymerase III in transcription complexes. Mol. Cell. Biol. 13, 942-952. 33. Holm, L. & Sander, C. (1991). Database algorithm for generating protein backbone and side-chain coordinates from a Ca trace application to model

building and detection of co-ordinate errors. J. Mol. Biol. 218, 183-194. 34. Holm, L. & Sander, C. (1993). Protein structure comparison by alignment of distance matrices. J. Mol. Biol. 233, 123-138.

Edited by P. Wright (Received 10 November 2000; received in revised form 13 November 2000; accepted 13 November 2000)