Biochemical Society Transactions

Biochemical Society Focused Meeting

Archaeal RNA polymerase: the influence of the protruding stalk in crystal packing and preliminary biophysical analysis of the Rpo13 subunit

Magdalena Wojtas, Bibiana Peralta, Marina Ondiviela, Maria Mogni, Stephen D. Bell, Nicola G.A. Abrescia


We review recent results on the complete structure of the archaeal RNAP (RNA polymerase) enzyme of Sulfolobus shibatae. We compare the three crystal forms in which this RNAP packs (space groups P212121, P21212 and P21) and provide a preliminary biophysical characterization of the newly identified 13-subunit Rpo13. The availability of different crystal forms for this RNAP allows the analysis of the packing degeneracy and the intermolecular interactions that determine this degeneracy. We observe the pivotal role played by the protruding stalk composed of subunits Rpo4 and Rpo7 in the lattice contacts. Aided by MALLS (multi-angle laser light scattering), we have initiated the biophysical characterization of the recombinantly expressed and purified subunit Rpo13, a necessary step towards the understanding of Rpo13's role in archaeal transcription.

  • archaeon
  • RNA polymerase
  • Rpo13
  • transcription
  • X-ray structure


RNAP (RNA polymerase) enzymes are complex multi-subunit machines found in all three domains of life: Eukarya, Bacteria and Archaea. These enzymes have been studied intensively at the structural level for the last two decades. Initial studies of the bacterial RNAP were followed rapidly by elucidation of the structures of the more complex eukaryotic RNAPs [15]. Although an incomplete X-ray structure of the archaeal RNAP (from the genus Sulfolobus) appeared in 2008 [6], it is only recently that the whole enzyme has been visualized [7] (Figure 1A). This latter work has described how the ancestral core enzyme was modulated by incorporation of novel subunits, an evolutionary process that, in eukaryotes, has led to the appearance of three different classes of nuclear RNAPs. In our study of the archaeal transcription machinery, we observed three different crystal forms of the RNAP from Sulfolobus shibatae, one of them (space group P21) is reported for the first time for this species in the present paper. The different crystal forms obtained have prompted the analysis of the lattice contacts and also the comparison with the packing arrangements of the counterpart eukaryotic enzyme, RNA Pol II. This comparison allows rationalization of the contribution of the corresponding domains in crystal lattice formation, most significantly by the protruding heterodimer Rpo4–Rpo7 (and the eukaryotic counterparts Rpb4–Rpb7). Interestingly, this module within the enzyme has been postulated as the docking platform for TFE (transcription factor E) [8] as well as the site of interaction with the nascent RNA [911]. As for the two previous S. shibatae RNAP X-ray structures [7], both subunits Rpo8 and Rpo13 are visible in the crystal form P21 (we will follow the subunit nomenclature analogous to the eukaryotic one; see [7]). Rpo8's gene, present only in the Crenarcheaota phylum, was detected by bioinformatics analyses in 2008 [12]. In contrast, as early as 1994, the existence of an unusual RNAP subunit was noticed in the Sulfolobus genus [13] which now we know to correspond to Rpo13 [7]. The location of Rpo13 in the RNAP complex, overlooking the DNA-binding cleft, its ordered HTH (helix–turn–helix) motif, modelled as α1 and α2 helices, and the bioinformatics secondary-structure prediction of a three-helical protein led us to hypothesize Rpo13's involvement in DNA bubble formation via a putative α3-helix [7] (Figures 1B and 1C).

Figure 1 Complete archaeal RNAP

(A) Surface representation of the current archaeal RNAP model from S. shibatae (PDB code 2WAQ) showing its overall architecture and the location of the different subunits with below the subunit nomenclature. (B) Relative positions of Rpo13's HTH (as cartoon in gold) and DNA (as surface in cyan template strand, light green non-template strand and red RNA) within the RNAP (left as white surface and right as semi-transparent surface). The DNA–RNA location has been obtained by superimposing the Rpb1 (PDB code 1R9T [24]) on to Rpo1N. The yellow semi-transparent sphere schematically approximates the volume that the α3-helix would possibly occupy if its midpoint is at ~15 Å from the last residues of the α2-helix modelled as the closest to the cleft as in Korkhin et al. [7] (assuming α3-helix no folding-back); it is apparent that Rpo13 is on the path of the DNA even in the case of the α2-helix being the helix furthest from the cleft (magenta semi-transparent sphere). (C) Rpo13's primary sequence with the secondary-structure prediction of α1-, α2- and α3-helices and reliability histogram below. The Δ and ● symbols beneath the primary sequence highlight the glutamate and the lysine residues respectively in the N- and C-termini. Reprinted from [7] with permission.

Archaeal and eukaryotic transcription share close evolutionary ancestry [7,14] and this is manifested in the early steps of initiation and the formation of the pre-initiation complex minimally composed in Archaea of RNAP, TFB (transcription factor B) and TBP (TATA-box-binding protein) [5,1517]. In order to extend our results obtained from the archaeal system to the eukaryotic one with the aim of inferring dynamic processes that escape structural analysis [14], we need to address the question of the role that subunit Rpo13 plays in archaeal transcription. With this ultimate goal in mind, we have started the analysis by biophysical methods of purified Rpo13.

Overall archaeal RNAP structure

RNAP from S. shibatae packs in three different space groups: P21212 and P212121, as described previously [7], and P21 reported in the present paper. All of them show the full complement of 13 subunits. Most of the archaeal subunits (Rpo1–Rpo3, Rpo5, Rpo6 and Rpo8–Rpo13) form a globular assembly, with only subunits Rpo4 and Rpo7 protruding asymmetrically originating the so-called ‘stalk’ and conferring on the entire enzyme the largest dimension of ~170 Å (1 Å=0.1 nm) (Figure 1A). The Rpo4–Rpo7 heterodimer is attached to the globular head via subunit Rpo7 with a contact area of ~1100 Å2, a value very similar to that of the eukaryotic counterpart, although with much stronger affinity [14], probably a requirement of the archaeal RNAP to perform efficiently in extreme conditions. Nevertheless, the Rpo4–Rpo7 appendix remains flexible, as also illustrated by the corresponding weak electron density, a feature detected in all three crystal forms. Significantly, Rpo4 and Rpo7 homologues exist in all three nuclear polymerases Pol I (A14 and A43), Pol II (Rpb4 and Rpb7) and Pol III (C17 and C25) [14,1822].

At the heart of the globular head, in the Rpo1N subunit, is located the structurally ordered catalytic core (residues Asp456, Asp458 and Asp460), whereas, at the periphery, fitted between Rpo5 and Rpo1, is located subunit Rpo13, showing an HTH motif visible also in the P21 space group where 38 residues were modelled. As for the two RNAP structures described previously, the terminal ends are mobile, and in none of the structures is there any indication of the location of the third predicted α3 (Figure 1). Despite the current lack of functional data on Rpo13's role in archaeal transcription, it is clear from the crystal structures that Rpo13 is located proximal to the DNA path. This suggests that Rpo13 might interfere with the DNA either through the glutamate-rich N-terminus and/or via the lysine-rich C-terminal end (putative α3-helix; Figures 1B and 1C).

Influence of the Rpo4–Rpo7 heterodimer in S. shibatae RNAP crystal lattices

The packing degeneracy is often a sign of molecular plasticity that is occasionally reflected in the intermolecular contacts within the crystal. Owing to the striking brake of globularity introduced by the protruding stalk in the shape of archaeal and eukaryotic RNAP enzymes and the surface accessibility created, we decided to assess its contribution in crystal-packing contacts as a way of grasping a possible biological inference. We analysed the different packing constraints in the three crystal structures that have been obtained during our studies of the RNAP from S. shibatae: P21212 (PDB code 2WAQ), P212121 (PDB code 2WB1) and P21 (PDB code 2Y0S). For the space group P21212, we have one RNAP molecule in the asymmetric unit contrasting with the two observed in space groups P212121 and P21. This latter space group was also found in the RNAP crystal of Sulfolobus solfataricus [6], albeit the non-crystallographic symmetry axis relating the two molecules is oriented differently. Nevertheless, dimer formation can be brought back to equivalent interactions. Thus, using methods available through PISA (Protein Interfaces, Surfaces and Assemblies) [18], we have analysed the crystal interfaces between the RNAP molecules that form the different crystal lattices. Although the low resolution of all available archaeal RNAP X-ray structures and the inherent flexibility of the stalk limit possible analytical conclusions, it appears that the largest interface between the RNAP molecules in all three space groups is constituted by the interaction of the C-terminal domain of neighbouring Rpo7 subunits with a value of ~840 Å2 for the P21212 crystal (Figure 2A), ~800 Å2 for the P212121 crystal and ~600 Å2 for the P21 crystal (~700 Å2 for S. solfataricus RNAP [6]). The assessment of the ΔiG P-values suggests that this interface does not represent a naturally occurring interaction (P=0.8), but it is functional to the crystal-packing (briefly ΔiG P>0.5 possibly non-specific interface, ΔiG P<0.5 possibly specific interface, for more details see [18]). On the other hand, the globular RNAP head anchors neighbouring molecules at different point locations (we considered only those with areas ≥200 Å2). In the P21212 crystal, these intermolecular contacts contemplate interfaces with areas of ~470 Å2 (Rpo1N–Rpo2, ΔiG P=0.4) and ~240 Å2 (Rpo8–Rpo2, ΔiG P=0.8) and three with values between 200 and 220 Å2 (Rpo1C–Rpo2, ΔiG P=0.5; Rpo1C–Rpo12, ΔiG P=0.6; Rpo3–Rpo1N, ΔiG P=0.6). These interfaces have few significant ΔiG P-values, yet almost all are conserved in the P212121 and in P21 crystals. From these data, it is likely, but not certain, that it is the co-operative effect of this set of smaller interface areas that promotes the lattice network formation rather than the individual interaction between neighbouring Rpo7 stalk domains. Nevertheless, the interface area between neighbouring Rpo7 subunits alone accounts for ~40% of the total interface area. We have compared these observations with the packing arrangement of the eukaryotic enzyme from Saccharomyces cerevisiae, the only example crystallized with and without the stalk domain (PDB codes 1I50 [2] and 1WCM [19] respectively). Significantly, the presence of the Rpb4–Rpo7 stalk induces a change in the packing arrangement and cell dimensions between the two eukaryotic crystal forms. In the case of the complete RNAP II structure, the Rpb7 domain contributes in crystal lattice formation in a comparable manner and with a similar interface area (~800 Å2), but lower ΔiG P-value (0.4) compared with that in the crystal of the archaeal polymerase. The head domain's interfaces are different mainly due to the presence of subunit Rpb9 (Rpb9–Rpb1, ~460 Å2, ΔiG P=0.7; Rpb3–Rpb1, ~380 Å2, ΔiG P=0.5; Rpb4–Rpb1, ~330 Å2, ΔiG P=0.8). From the RNAP structures considered, the stalk, through subunit Rpo7 (Rpb7), appears to act as a point capable of modulating (via a range of hydrogen bonds) the packing arrangement (Figure 2). It is tempting to speculate that the observed stalk–stalk interface might mimic some biological interfaces resulting from contacts present in the landscape of possible RNAP interactions with co-factors required for its recruitment on to the promoter region.

Figure 2 Modes of stalk–stalk interaction

Top, archaeal RNAP (PDB code 2WAQ) colour-coded as Figure 1(A) and a 2-fold symmetry-related molecule at the front in semi-transparent white. The white outline highlights the protruding stalk and in light-yellow are the residues of Rpo7 involved in the interface. Ssh, S. shibatae. Bottom, eukaryotic RNAP (RNA Pol II; PDB code 1WCM) and 2-fold symmetry-related molecule colour-coded as above, in light-pink subunit Rpb9 (absent from the archaeal RNAP). Although the orientation of the two RNAPs is not equivalent within the dimer, the structural conservation of the corresponding subunits and enzyme architecture is apparent.

Biophysical analysis of Rpo13 subunit

Very little is known about Rpo13 [7,13]. In all RNAP structures that we have solved, clear electron density corresponding to an HTH motif is visible at the crevice between Rpo5 and Rpo1 clamp–head domain (Figure 1A). The Rpo13 sequence reveals a strikingly polarized distribution of negative and positive charges (14 glutamate residues in the first 32 residues and 11 lysine residues in the last 22 residues) (Figure 1C). These terminal regions (more precisely residues 1–36 and residues 69–104) are predicted to be disordered and they are not seen in any of the RNAP structures apart from very few residues [7]. Whether these termini are structured in solution or become structured upon ligand binding is not yet known. Purification by SEC (size-exclusion chromatography) of recombinantly expressed Rpo13 (molecular mass 12.1 kDa) yielded a not fully symmetrical elution profile (Figure 3A, left) and suggested an oligomer of ~54 kDa, a large value that has impelled the analysis of Rpo13's mass using the MALLS (multi-angle laser light scattering) technique [23]. Quasi-elastic light scattering and SEC–MALLS analysis have demonstrated that Rpo13 in solution (when injected at ≤0.9 mg/ml) forms low-disperse oligomers with an averaged molar mass of 19.96±0.03 kDa, thus inferring a predominant dimeric oligomeric state (probably a monomeric form is beginning to be eluted at ~15.3 ml; Figure 3A, right). The same sample was cross-linked and separated by SDS/PAGE. On the gel, two major bands are visible corresponding to Rpo13 monomer (running at ~19 kDa) and dimer, although weaker bands appear at a higher molecular mass (Figure 3B, left). Peptide MS fingerprinting and linear MALDI–TOF/TOF (matrix-assisted laser-desorption ionization–tandem time-of-flight) have respectively confirmed the identity of the purified protein, determined its molecular mass (post-translational processing removes the initial methionine residue), but also detected the dimeric form (Figure 3B, right). The probable non-globular shape of the dimer and/or the possible unfolded N- and C-termini could explain the unusual elution time during SEC. In addition, the negatively charged polypeptide chain (see N-terminus in Figure 1C) could explain the observed lower electrophoretic mobility.

Figure 3 Biophysical analysis of Rpo13

(A) Left: Rpo13 elution profile from SEC Superdex 200 16/60; continuous line and broken line correspond respectively to absorbance readings at 280 and 260 nm.Right: the broken line illustrates the molar mass distribution as determined by MALLS along the elution profile (continuous line) represented by the UV reading at 280 nm. mAU, milli-absorbance units. (B) Left: Coomassie Blue-stained SDS/PAGE gel. Lane 1, molecular-mass marker; lane 2, Rpo13 from size-exclusion column; lane 3, same Rpo13 sample cross-linked with 0.01% glutaraldehyde for 10 min; different oligomeric states of Rpo13 are visible. Right: raw MALDI–TOF/TOF ionization MS data obtained from purified Rpo13 (molecular mass, 12 147.5 Da with first methionine residue included; 12 016.3 Da without first methionine residue). The m/z spectrum shows three different mass peaks associated with the different charge states of Rpo13 [1:2, 6001.4; 1:1, 12 008.5 (monomer); 2:1, 23 999.6 (dimer, as shown by the inset)].


We have shown how the stalk heterodomain Rpo4–Rpo7 participates via neighbouring Rpo7 subunits in the network of interactions within the three different RNAP crystal forms of S. shibatae, a recurrence (seen also in the RNAP crystals of S. solfataricus and S. cerevisiae with Rpb4 and Rpo7) that highlights the possible ‘recruitment platform’ role played by this appendix.

Furthermore, the preliminary biophysical analysis of Rpo13 depicts a protein with an unusual electrophoretic mobility, but that purifies as a near-homogenous dimer in solution, an observation reminiscent of the behaviour of the general transcription factor TBP.

Our work represents the starting point for further structural and functional studies that aim to fully elucidate the biological significance of Rpo13 in archaeal transcription and in the evolution of RNAP enzymes.


Our work is supported by the Spanish Ministerio de Cienca y Technología [grant number BFU2009-08123], the Basque Government [grant number PI2010-20], the CICbioGUNE (N.G.A.A.) and the Wellcome Trust and Edward Penley Abraham Trust (S.D.B.). The research leading to these results has received funding from the European Community's Seventh Framework Programme (FP7/2007–2013) [grant number 226716].


We thank Yakov Korkhin for useful discussions, Felix Elortza and Ibon Iloro for biotechnological assistance. We acknowledge the European Synchrotron Radiation Facility (ESRF) and the Swiss Light Synchrotron (SLS) for provision of synchrotron facilities. We also thank the staff at the beamlines ID14-EH1 at ESRF and PXIII at SLS.


  • Molecular Biology of Archaea II: A Biochemical Society Focused Meeting held at Robinson College, Cambridge, U.K., 16–18 August 2010. Organized and Edited by Stephen Bell (Oxford, U.K.) and Finn Werner (University College London, U.K.).

Abbreviations: HTH, helix–turn–helix; MALDI–TOF/TOF, matrix-assisted laser-desorption ionization–tandem time-of-flight; MALLS, multi-angle laser light scattering; RNAP, RNA polymerase; SEC, size-exclusion chromatography; TBP, TATA-box-binding protein; TFB, transcription factor B; TFE, transcription factor E


View Abstract