A. Promoter motifs recognized by primary bacterial RNAP σ factors. Grey circles represent the DNA nucleotides (top, nt-strand; bottom, t-strand). The extent of the transcription bubble in RP_o is illustrated (separated circles). Promoter motifs recognized by primary σ factors are colored black (-35 element, extended –10 element, discriminator element) or yellow (-10 element or Pribnow box), with the consensus sequences above. The position with respect to the transcription start-site (+1) is denoted below.
B. Sequence logo (Schneider and Stephens, 1990) for the bacterial primary σ factor –10 element (adapted from (Shultzaberger et al., 2006).
C. Taq σ^A domain architecture, crystallization construct, and sequence characteristics. The domain architecture of Taq σ^A is represented schematically (structured domains, thick regions; flexible linker, thin regions; (Campbell et al., 2002). The crystallization construct, comprising σ_2–3, is highlighted. The green bar above indicates that σ₂ was ordered in the crystal structure (solid bar) while σ₃ was disordered (dashed bar). The conserved regions within σ₂ (Lonetto et al., 1992) are labeled and color-coded. Expanded below is a sequence alignment of the σ₂ conserved regions for Taq σ^A, Bacillus subtilis (Bsu) σ^A, and E. coli σ⁷⁰. Sequences are shown in one-letter amino acid code. Numbers at the beginning of each line indicate the amino acid positions. Number scales at the top and bottom indicate amino acid position in Taq σ^A, and E. coli σ⁷⁰, respectively. The sequence blocks are color-coded according to the schematic above; the darker bands of color denote protein side-chains that interact with the DNA. Amino acid identity in >50% of the sequences is indicated by a red background, amino acid similarity by a blue background. Groups of residues considered similar are: ST (h), RK (b), DE (a), NQ (m), FYW (o), and ILVM (f). The histogram at the top represents the level of sequence conservation (using the groups denoted above) at each position in an alignment of 50 primary bacterial RNAP σ factors (Campbell et al., 2002). Sequence conservation of 100% is represented by a tall red bar, less than 20% by a small dark blue bar, intermediate levels are represented by orange, light green, and light blue bars.
D. Crystallization oligonucleotide. Synthetic 11mer ssDNA oligonucleotide used for crystallization (Feklistov et al., 2006). The –10 element is colored yellow and labeled.
E. Structural overview. The σ^A₂ protein is shown as a molecular surface and color-coded as in Figure 1C (conserved regions shown in pale colors, residues with side-chains that interact with the DNA shown in darker colors). Selected residues are labeled (see text). The ssDNA is shown with carbon atoms color-coded as in Figure 1D. Other atoms are colored as follows: nitrogen, blue; oxygen, red; phosphorous, orange.

A. Schematic showing DNA/protein contacts. Arrows indicate direct or water-mediated (blue W) hydrogen bonds from side chain (solid) or main chain (dashed) protein atoms. Van der Waals contacts are shown by thick grey dashed lines, stacking/cation-π interactions are shown by thick green dashed lines. The numbers indicate amino acid positions for Taq σ^A and E. coli σ⁷⁰ (in brackets). Grey areas of the oligo indicate regions without biologically relevant contacts.
B. The σ^A₂ protein is shown as an α-carbon backbone ribbon and with a transparent molecular surface (green). Selected protein side chains are shown and labeled. The ssDNA is shown as in Figure 1E (nucleotides upstream of T_-12 have been omitted for clarity). An ordered water molecule is shown as a small red sphere. Polar interactions (H-bonds and/or salt bridges) are denoted (grey dashed lines).

A. Modeling of the upstream fork junction and recognition of the (T/A)_-12 bp. The upstream dsDNA is modeled as B-form double-helix based on (Murakami et al., 2002). In the modeled portion of the DNA, the nt-strand is colored blue; t-strand, grey. The α-helix formed by σ regions 2.3 (turquoise) and 2.4 (green) is shown as a ribbon. The side-chains of W256 (blocks extension of DNA double-helix downstream past the -12 bp and occupies the space vacated by the flipped A_-11) and Q260 (positioned to interact with the major-groove surface of the t-strand A base-paired to T_-12) are shown. Modeling of the recognition of the mutant (C/G)_-12 base pair by Q260H (Waldburger et al., 1990) or Q260R (Kenney et al., 1989) is shown on the right.
B. Edge view of A_-11 base. The ssDNA is shown with the atoms of the A_-11 base in CPK format. An ordered water molecule is shown as a small red sphere. Polar interactions (H-bonds and/or salt bridges) are denoted (grey dashed lines). The σ^A₂ protein is shown as a worm. Side chains that make up the A_-11 binding pocket are shown with transparent CPK atoms.
C. Side view of A_-11 base. The A_-11 binding pocket has been sliced near the level of the base and is viewed from the side. The molecular surface of σ^A₂ within 4.5 Å of the A_-11 nucleotide is shown as a transparent surface.

The σ^A₂ protein, selected protein side chains, the ssDNA, ordered water molecules, as well as polar interactions (H-bonds and/or salt bridges) are shown as in Figure 2B.

A, B, C. Binding isotherms for dsDNA fragments to RNAP holoenzyme. Shown are representative curves for the binding of σ[211-Cys-Alexa555]-holoenzyme at increasing dsDNA concentration as measured using the RNAP beacon assay (Experimental Procedures; (Mekler et al., 2011). Sequences/modifications are shown above binding curves on the left.
D. Summary of DNA modifications studied. Schematic showing orientation of the (A/T)_-11 and (T/A)_-7 base pairs in relation to the DNA-binding surface of σ₂. Modifications decreasing the binding of the dsDNA promoter fragment (sequence shown at the bottom) are highlighted in red, neutral modifications are green. The -10 element sequence on the nt-stand is shown in red, -10-like element sequence on the opposite strand is shown in bold. Modified base pairs investigated in this study are shown on the left [for (A/T)_-11 base pair] and right [for (T/A)_-7 base pair]. Changes from the canonical A/T base pair are highlighted in bold. Measured K_d values (μM) are shown underneath each modification. For each modification, K_d measurements were independently repeated two or three times, and averages were calculated. The experimental variation among replicate measurements usually did not exceed 20% of the average value.

A. A model of RP_c (adapted from (Murakami et al., 2002). Taq RNAP holoenzyme is shown as a molecular surface (α subunits, ω, grey; β, cyan; β', pink; σ is orange, except σ₂, which is green). The position of the RNAP active site Mg²⁺-ion is illustrated by a yellow sphere (viewed through the β’-subunit). The thick black arrow indicates how the dsDNA downstream of the -10 element must move to enter the RNAP active site channel. The DNA t-strand is colored grey, the nt-strand blue, with the –35 and –10 elements labeled. The boxed region is shown in more detail in Figure 6C.
B. View of the RP_c model down the DNA helix axis (the DNA is shown as a P-backbone worm) with the downstream direction into the page. The DNA double-helix sits in a shallow trough formed by σ₂, σ₃, and β. W256 of σ₂, which protrudes into the DNA helix at the position of the -11 bp, is colored red.
C. Magnified view showing σ₂, the dsDNA of the RP_c model (with A_-11c and T_-7c highlighted), and the bound nt-strand ssDNA representing RP_o (yellow). The σ₂ is partially cut-away to reveal the A_-11o and T_-7o binding pockets. The wedge residue W256 is highlighted in red. Red arrows connect A_-11 and T_-7 of RP_c with the same nucleotides of RP_o.