Structural analysis of the recognition of the -35 promoter element by SigW from Bacillus subtilis

Sigma factors are key proteins that mediate the recruitment of RNA polymerase to the promoter regions of genes, for the initiation of bacterial transcription. Multiple sigma factors in a bacterium selectively recognize their cognate promoter sequences, thereby inducing the expression of their own regulons. In this paper, we report the crystal structure of the σ4 domain of Bacillus subtilis SigW bound to the -35 promoter element. Purine-specific hydrogen bonds of the -35 promoter element with the recognition helix α9 of the σ4 domain occurs at three nucleotides of the consensus sequence (G-35, A-34, and G’-31 in G-35A-34A-33A-32C-31C-30T-29). The hydrogen bonds of the backbone with the α7 and α8 of the σ4 domain occurs at G’-30. These results elucidate the structural basis of the selective recognition of the promoter by SigW. In addition, comparison of SigW structures complexed with the -35 promoter element or with anti-sigma RsiW reveals that DNA recognition and anti-sigma factor binding of SigW are mutually exclusive.


Introduction
Transcription in bacteria is initiated by sigma factors, which recruit the core RNA polymerase to a cognate promoter [1,2]. Sigma factors selectively recognize promoter elements, -10 and -35 elements, and additional sequences, including extended -10 element and discriminator, which are present upstream of the transcription start site [3,4]. During transcription initiation, the -10 element is strand-separated to form a transcription bubble [5,6], whereas the -35 element is recognized by a helix-turn-helix (HTH) motif in the sigma factor, without strand separation [7,8].
Sigma factors are categorized into two families, based on the sequence homology with Escherichia coli sigma factors: housekeeping σ 70 required for bacterial homeostasis; and σ 54 activated for nitrogen utilization [9]. The σ 70 family is further sub-divided into five groups [4,10]. Group I sigma factors are composed of σ 1.1 , σ 2 , σ 3 , and σ 4 domains, which are responsible for the recognition of the discriminator, -10, extended -10, and -35 elements, respectively. The primary sigma factors, which regulate the transcription of housekeeping genes, belong to group I. Group II-V are classified depending on the presence or absence of the domains and containing 80 mM imidazole. Proteins bound to the resin were eluted by an imidazole gradient (0.08-1.00 M imidazole). Fractions that contained 6 X His-σ W 4 were pooled and treated with TEV protease overnight at 25˚C to cleave the His tag. After complete cleavage, the protein solution was dialyzed against buffer A for 3 h and passed through Ni-NTA resin to remove the 6 X His tag (Thermo Fisher Scientific, Rockford, IL, USA). σ W 4 was further purified by SEC using Superdex 75 preparatory grade column (GE Healthcare Biosciences) pre-equilibrated with buffer B (20 mM HEPES pH 7.5, 1.0 M NaCl, and 5% (v/v) glycerol).

Crystallization, data collection, and structure determination
The crystal structure of σ W 4 /-35 W was determined by the molecular replacement (MR) method using PHASER [33]. The structure of the E. coli σ E 4 /-35 element (PDB ID: 2H27) was used as a template for MR. MR solution was found from the truncated σ E 4 (residues 127-186)/-35 element. Cycles of refinement and model building were performed at 3.1 Å resolution using PHENIX.refine [34] and COOT [35]. Final refinement resulted in R / R free values of 24.8 / 29.0% without residues in the disallowed region of the Ramachandran plot. The data collection and refinement statistics are summarized in Table 1. The final coordinates and structure factors were deposited in the Protein Data Bank (PDB ID: 6JHE). Structural alignment was performed using the DALI server [36]. Protein-ligand interactions were analyzed with Lig-Plot+ [37] and PDBePISA [38]. The free energy change (ΔG) of the protein caused by ligand binding was analyzed using PDBePISA [38]. The conversion of ΔG to a dissociation constant (Kd) was calculated using the equation Kd = e (ΔG/RT) (R = 1.987 cal/molK; T = 293K). DNA geometry was analyzed using w3DNA [39]. Surface charge distribution was calculated using APBS [40]. The figures were drawn using PyMOL [41] and ALSCRIPT [42].

Accession number
The final coordinates and structure factors were deposited in the Protein Data Bank (PDB ID: 6JHE for σ W 4 /-35 W ).

Results and discussion
Overall structure B. subtilis σ W 4 (residues 125-187) recognizes a cognate -35 promoter element (Fig 1A and  1B). Its consensus sequence is identified as T -36 G -35 A -34 A -33 A -32 C -31 X -30 T -29 T -28 T -27 , based on the promoter sequences of the SigW regulon [43]. σ W 4 was purified under high salt conditions (1 M NaCl) to minimize its instability, and dialyzed in low salt buffer together with the double-stranded DNA of A -38 T -37 T -36 G -35 A -34 A -33 A -32 C -31 C -30 T -29 T -28 T -27 (-35 W ) to allow σ W 4 binding to -35 W . Crystals of σ W 4 /-35 W belonging to a hexagonal space group grew under conditions containing PEG1500 and isopropanol as protein precipitants. Diffraction data were collected at a resolution of 3.1 Å ( Table 1), and the structure was determined by molecular replacement using the structure of a truncated E. coli σ E 4 /-35 E (-35 promoter element for SigE binding) as a template [8].
The crystal structure contains a σ W 4 monomer and a double-stranded -35 W in the asymmetric unit. Residues 134-186 of SigW and 11 nucleotide pairs of -35 W were traced into the electron density (S1 Fig) and the final structure was refined at R / R free values of 24.8 / 29.0% (Table 1). σ W 4 is comprised of four α-helices (α6-α9) in the crystal structure of the SigW/ RsiW complex [23]. However, residues 125-133, which correspond to α6, are disordered in the crystal structure of σ W 4 /-35 W (Fig 1A). The residues on α6 are likely to be flexible because they are not bound to DNA directly. α8-α9 of σ W 4 forms the HTH motif, and α9 is inserted into the major groove of -35 W as a DNA recognition helix (Fig 1C and 1D) [7]. Positivelycharged residues are distributed on the DNA binding surface of α8-α9, whereas hydrophobic patches are distributed on the opposite side that interacts with σ W 2 in the crystal structure of SigW/RsiW (Fig 1E and 1F) [23].

Interactions between σ W 4 and the -35 promoter element
The recognition helix α9 mediates the major interactions between σ W 4 and -35 W through hydrogen bonds and hydrophobic interactions. The bases and backbones of three purine nucleotides (G -35 and A -34 in the non-template strand and G' -31' in the template strand; ' indicates the template strand of DNA) form hydrogen bonds with the residues on the N-terminal half of α9 in σ W 4 (Fig 2A-2C and S2 Fig). The guanine oxygen (O6) and backbone phosphate (OP2) of G -35 form hydrogen bonds with side chain amino groups (NH2) of R175 and R172, respectively. The purine nitrogen (N7) and backbone phosphate (OP2) of A -34 interact with the side chain oxygens (OG1) of T171 and T168. The guanine oxygen (O6) of G' -31 forms a hydrogen bond with the side chain amino group (NZ) of K170. The electron density map for the side chain of K170 is relatively weak; however, the most-preferred rotamer is at hydrogen bond distance to the O6 of G' -31 (S2B Fig). Hydrophobic interactions are observed between K170-G' -31 T' -32 , T171-A -34 , H174-T' -32 T' -33 , and R175-T -36 (S3 Fig).
In addition to α9, α7-α8 contributes to -35 W binding without base specificity. The backbone atoms (OP1, OP2, and OP2) of G' -30 form hydrogen bonds with the side chain amino group (NZ) of K148, the side chain oxygen (OG) of S154, and the backbone nitrogen of L155, respectively (Fig 2B and 2C). Altogether, α9 in σ W 4 specifically recognizes -35 W , and α7-α8 provides additional contacts, leading to a tighter interaction.

Structural comparison of σ W 4 /-35 W and σ E 4 /-35 E
SigE is an E. coli ECF sigma factor activated in response to envelope stress and induces transcription of heat shock proteins [45]. Its σ 4 domain (σ E 4 ) recognizes the -35 element, of which the consensus sequence is G -35 G -34 A -33 A -32 C -31 T -30 T -29 (-35 E ) [46]. The σ E 4 structure is highly similar to σ W 4 [8]. σ W 4 is superimposed on σ E 4 with a root mean square deviation (RMSD) value of 1.8 Å for 53 Cα atoms (S4A Fig). The overall fold is conserved between σ W 4 and σ E 4 and the main differences are observed at the N-and C-termini (S4A and S4B Fig). Overall, the interactions of σ E 4 and σ W 4 with the corresponding -35 elements are conserved. Three nucleotides, G -35 , G -34 , and G' -31 , in -35 E , which correspond to G -35 , A -34 , and G' -31 in -35 W , mediate purine nucleotide-specific interactions with the recognition α-helix (Fig 2A). The backbone and base of G -35 (G -35 in σ W 4 ) form hydrogen bonds with R173 (R172 in σ W 4 ) and R176 (R175 in σ W 4 ). The backbone and base of G -34 (A -34 in -35 W ) form hydrogen bonds with T169 (T168 in σ W 4 ) and S172 (T171 in σ W 4 ). Although the -34 position is not identical between -35 E and -35 W (G -34 in -35 E and A -34 in -35 W ), hydrogen bonds mediated by the backbone phosphate and purine N7 are conserved. The base of G' -31 forms a hydrogen bond with R171 (R170 in σ W 4 ). The backbone phosphate of G' -31 also forms a hydrogen bond with R178, which is not observed in the σ W 4 /-35 W structure. A' -30 in-35 E mediates backbone interactions similarly to G' -30 in -35 W (Fig 2). The backbone phosphate of A' -30 forms hydrogen bonds with the NH1 of R149 (K148 in σ W 4 ) and the backbone nitrogen of Y156 (L155 in σ W 4 ) (Fig  2D and 2E). Interaction between S154 and A' -30 in σ W 4 /-35 W is missing in the σ E 4 /-35 E structure. In summary, purine base-specific hydrogen bonds in the structures of σ W 4 /-35 W and σ E 4 /-35 E are conserved, whereas the hydrogen bonds with the nucleotide backbone are slightly different.

DNA geometry of -35 W
Nucleotides A -33 A -32 in -35 E (G -35 G -34 A -33 A -32 C -31 T -30 T -29 ) do not form hydrogen bonds with σ E 4 , although these nucleotides are conserved among the -35 elements of the SigE regulon, and mutating these nucleotides in the Salmonella enterica serovar Typhimurium SigE has been shown to lead to defective transcription [47]. These nucleotides are involved in characteristic oligo(dA)/oligo(dT)-like DNA geometry that is rigid and straight with a narrow minor groove [8,48]. A previous structural study of the σ E 4 /-35 E complex suggested that the geometry of the narrowed minor groove is critical for σ E 4 recognition [8], and we show that -35 W also displays a narrowed minor groove (Fig 3). Like -35 E , the narrowing of the minor groove of -35 W begins at A -33 A -32 and is stabilized downstream of the -35 element, even though the downstream sequence of -35 W has an insertion of two cytosines (A -33 A -32 C -31 C -30 T -29 ) (Fig 3). In contrast, -G -33 A -32 in -35 A (the -35 element of Thermus aquaticus SigA) has a wider minor groove than normal B-DNA (Fig 3). The crystal structure of σ W 4 /-35 W supports the suggestion that the A -33 A -32 conservation in the -35 element for group IV sigma factors is critical for the formation of the narrow minor groove [8].

Structural comparison of σ W 4 /-35 W and SigW/RsiW
The crystal structure of SigW complexed with anti-sigma RsiW was previously reported [23]. The structures of σ W 4 under the binding of -35 W and RsiW superimpose with an RMSD value of 1.6 Å for 53 Cα atoms (S4B Fig). Although conformational differences of σ W 4 in the structures bound to -35W and RsiW are minor, σ W 4 bound to -35 W exists in a slightly compact conformation. σ W 4 interacts with -35 W through the residues K148, S154, L155, T168, K170, T171, R172, and R175, with a surface area of 621.2 Å 2 buried at the binding interface (Fig 4A and  4B). σ W 4 interacts with RsiW through the residues I150, K170, H174, R177, E178, R181, R185, and L187, these interactions result in an approximately 50% larger burial of surface area (915.2 Å 2 ) (Fig 4C and 4D). The surface area of σ W 4 that binds -35 W and RsiW partially overlaps with residue K170 on σ W 4 (Fig 4A and 4C), indicating that the interactions of -35 W and RsiW with σ W 4 are mutually exclusive. In the crystal structure of the SigW/RsiW complex, the -10 element-binding surface of σ W 2 is buried in the surface of σ W 4 [23], whereas the -35 element-binding surface of σ W 4 is directly blocked by RsiW (Fig 4) Role of conserved residues in σ W 4 B. subtilis contains multiple ECF sigma factors that respond to diverse environmental stresses. The positions L137, L147, and E157 of σ W 4 are highly conserved in B. subtilis ECF sigma factors (Fig 5A and 5B). The highly conserved residues are likely to be involved in the intrinsic folding of SigW, but not in DNA binding. L137 and L147 stabilize σ W 4 as part of the central innermost hydrophobic cluster (S6A and S6B Fig). E157 is associated with σ W 2 binding in the crystal structure of the SigW/RsiW complex (S6C and S6D Fig).
The residues involved in DNA binding are less conserved than those involved in intrinsic folding. Residues K170, T171, and R175, which mediate purine-specific hydrogen bonds in σ W 4 , are aligned to K/R, S/T, and K/R in B. subtilis ECF sigma factors, as well as E. coli SigE (Fig 5A and 5B). The conservation of these residues correlates with the three nucleotides that mediate purine-specific interactions (G -35 , A -34 , and G' -31 in -35 W ). For example, A -34 in the SigV promoter interacts with T144 and T147, as does A -34 in -35 W , and T168/T171 in σ W 4 , whereas C -34 in the SigZ promoter is aligned with G138 and S141. These observations suggest that sequence variation in the DNA-binding interface confers the specificity required to discriminate between -35 elements (Fig 5C). In summary, conserved residues in B. subtilis ECF σ 4 are mainly involved in intramolecular stability and interactions with the -35 element. Slight