Structural, functional, and genetic analyses of the actinobacterial transcription factor RbpA

Significance

Initiation of transcription in bacteria relies on a multisubunit RNA polymerase in concert with a dissociable σ-subunit that confers promoter recognition and opening to reveal the DNA template strand. RbpA, a transcription activator unique to Actinobacteria and essential in Mycobacterium tuberculosis, associates tightly with σ and is required for efficient initiation, although its mechanism of action is unclear. Here, we solve the crystal structure of an M. tuberculosis σ–RbpA complex and present evidence indicating that RbpA activates transcription through unexpected contacts with promoter DNA. The work sheds light on the mechanism of transcription initiation by M. tuberculosis RNA polymerase, which is a proven antibiotic target.

Abstract

Gene expression is highly regulated at the step of transcription initiation, and transcription activators play a critical role in this process. RbpA, an actinobacterial transcription activator that is essential in Mycobacterium tuberculosis (Mtb), binds selectively to group 1 and certain group 2 σ-factors. To delineate the molecular mechanism of RbpA, we show that the Mtb RbpA σ-interacting domain (SID) and basic linker are sufficient for transcription activation. We also present the crystal structure of the Mtb RbpA-SID in complex with domain 2 of the housekeeping σ-factor, σ^A. The structure explains the basis of σ-selectivity by RbpA, showing that RbpA interacts with conserved regions of σ^A as well as the nonconserved region (NCR), which is present only in housekeeping σ-factors. Thus, the structure is the first, to our knowledge, to show a protein interacting with the NCR of a σ-factor. We confirm the basis of selectivity and the observed interactions using mutagenesis and functional studies. In addition, the structure allows for a model of the RbpA-SID in the context of a transcription initiation complex. Unexpectedly, the structural modeling suggests that RbpA contacts the promoter DNA, and we present in vivo and in vitro studies supporting this finding. Our combined data lead to a better understanding of the mechanism of RbpA function as a transcription activator.

Bacterial RNA polymerase (RNAP) comprises a catalytic core (subunit composition α₂ββ′ω) that is active for transcription elongation but requires an additional dissociable subunit, the σ-factor, for promoter-specific initiation (1, 2). All bacteria contain a single primary-σ that is essential for viability and directs transcription of most genes during vegetative growth. Most bacteria also harbor alternative σ-factors that can reprogram the RNAP to orchestrate adaptive responses to specific signals, such as stress and morphological development (3). Primary-σs can make up to four sequence-specific contacts with promoter DNA through three conserved helical domains (σ₂, σ₃, and σ₄) that are spread over one face of the RNAP (4–8). Within each structural domain are defined regions of sequence similarity (e.g., the structural domain-σ₂ comprises regions 1.2, 2.1, 2.2, 2.3, and 2.4) (9). The key interactions involve the σ₂- and σ₄-domains, which are spaced appropriately to contact the −10 and −35 promoter elements, respectively (6).

The vast majority of biochemical and genetic studies on bacterial transcription initiation have focused on Escherichia coli (Eco) RNAP and its primary-σ, σ⁷⁰. Recently, it has emerged that the regulation of transcription initiation in the Actinobacteria phylum, which includes major pathogens, such as Mycobacterium tuberculosis (Mtb), and antibiotic producers, such as Streptomyces spp., is distinct from the Eco system by the dependence on two initiation factors, CarD and RbpA, neither of which is found in Eco (10–12).

In mycobacteria, the essential protein CarD has been shown to be present at most promoters in vivo and function as a transcription activator in vitro (10, 13). More recently, Mtb CarD has been shown to activate transcription initiation by stabilizing the RNAP open complex with promoters (RPo) by preventing collapse of the transcription bubble (14). CarD makes a direct protein–protein interaction with the RNAP β-subunit β1-lobe, and structural models suggest that it also contacts the upstream edge of the −10 promoter element DNA in RPo (13).

RbpA was originally discovered in Streptomyces coelicolor (Sco), where it is a major component of RNAP holoenzyme (11). RbpA is found in almost all Actinobacteria and like CarD, essential for growth in Mtb (11, 12). Compared with CarD, much less is known about the RbpA structural mechanism.

The structural architecture of isolated RbpA has been defined by solution NMR (15, 16). A central RbpA core domain (RCD) comprises a β-barrel fold and is flanked by an unstructured 26-aa N-terminal tail and a C-terminal segment predicted to harbor two α-helices linked to the RCD by a 15-aa basic linker (BL) (Fig. 1A). In the absence of core RNAP, RbpA can form a stable binary complex with the σ₂-domain of the primary σ-factors of both Sco (σ^HrdB) and Mtb (σ^A) (15, 16). The RbpA–σ₂ interaction is mediated by the C-terminal segment [which we designate here the σ-interaction domain (SID)], and point mutations that disrupt σ-binding also disrupt RbpA function (15). In addition to primary σ-factors, RbpA interacts with certain group 2 σ-factors (σ^B in Mtb and σ^HrdA in Sco) but does not interact with group 3 or 4 σ-factors (15–17). RbpA is present at transcription initiation complexes in vivo and stimulates transcription in vitro from a wide range of Sco σ^HrdB-, Mtb σ^A-, and Mtb σ^B-dependent promoters (15, 18), but the mechanism for RbpA-mediated transcription activation is unknown.

Fig. 1.

Structural and functional analyses of RbpA-SID–σ^A₂. (A) Schematic of RbpA and σ^A domains. Domain 2 of Mtb σ^A is shown in orange, with the NCR in cranberry and the remaining regions colored gray. The SID and BL of RbpA are colored purple, and the remaining regions are colored gray. Regions visible in the crystal structure are flanked by dashed lines. (B) Multiple-round in vitro runoff transcription reactions using the vapB10 promoter template. Reactions contained core Mbo RNAP (50 nM), σ^A (250 nM), and RbpA derivatives (500 nM or 1.25 μM) as indicated. Lanes correspond to full-length RbpA (residues 1–111; lanes 4 and 5), SUMO-RbpA full length (RbpA residues 1–111 fused to SUMO; lanes 6 and 7), SUMO-RbpA-RCD (RbpA residues 1–71 fused to SUMO; lanes 8 and 9), and SUMO-RbpA-SID (RbpA residues 71–111 fused to SUMO; lanes 10 and 11). A graphical representation based on duplicate datasets (SD indicated) is illustrated above normalized to the data obtained with σ^A in the absence of RbpA. (C) Crystal structure of Mtb RbpA-SID in complex with σ^A₂ is shown in ribbon and colored as in A.

Here, we show that the RbpA-BL and SID are sufficient for in vitro transcription activation by RbpA, and we determine the X-ray crystal structure of the Mtb RbpA–σ^A₂ complex, revealing the essential RbpA-SID–σ^A₂ interactions as well as representing the first structure, to our knowledge, of a protein interacting with the nonconserved region (NCR) found exclusively in housekeeping σ-factors. From this result, we use a combination of structural modeling, mutagenesis, and in vitro and in vivo functional studies to probe the mechanistic basis of RbpA function. Our results suggest that, in addition to the RbpA-SID–σ^A₂ protein–protein interaction, RbpA–promoter DNA interactions are crucial in the role of RbpA as a transcription activator.

Results

RbpA-BL and SID Are Sufficient for Transcription Activation.

Although the essential role of the RbpA-SID in RbpA function is clear, the roles of the other RbpA structural elements (N-terminal tail, RCD, and BL) are not. We, therefore, tested truncated derivatives of Mtb RbpA [N-terminally fused to small ubiquitin-like modifier (SUMO) to improve stability (Table S1)] for transcription activation function using a mycobacterial transcription system and an Mtb vapB10 (Rv1398c) promoter template. The vapB10p promoter (19) is used here because it exhibits strong dependence on RbpA in vitro. As expected, an Mtb RbpA deletion mutant (containing only residues 1–71) lacking the BL and SID was unable to activate transcription in vitro (Fig. 1B, lanes 8 and 9). By contrast, the RbpA-SID and much of the BL (residue 72 to C terminus) showed partial in vitro function compared with full-length RbpA with an equivalent N-terminal SUMO fusion (Fig. 1B, lanes 10 and 11). We conclude that the RbpA-BL and SID are necessary and sufficient for at least partial RbpA in vitro transcription activation but that other elements of RbpA (N-terminal tail and/or RCD) are required for full activity.

Table S1.

Bacterial strains, plasmids, and PCR primers used in this study

Strain or plasmid	Relevant genotype/comments	Source
Strains
Sco A3(2)
S129	J1981 ΔrbpA::apr (AprR)	15
Eco
ET12567 (pUZ8002)	dam, dcm, hsdM; pUZ8002 is a nontransmissible derivative of RK2 (CmR and KmR)	36
BL21 (DE3)	Eco B F− ompT hsdS(rB−mB−) dcm gal λ(DE3) endA Hte (TetR)	37
BL21 (DE3) pLysS	BL21(DE3) pLysS (CmR)	37
BTH101	cya-99 (SpcR) used for BTH analysis	38
Rosetta 2	BL21(DE3) expressing rare codon tRNA genes	Novagen
Plasmids
pBluescript II SK+	E. coli cloning vector; ori pUC18 (AmpR)
pKT25	Two-hybrid vector; T25 fragment of Bordetella pertussis CyaA for N-terminal fusions (KmR)	38
pUT18	Two-hybrid vector; T18 fragment of Bordetella pertussis CyaA for C-terminal fusions (AmpR)	38
pKT25-hrdB	hrdB (codons 211–511) fused to T25 in pKT25 (KmR)	15
pKT25-sigA	Mtb sigA (codons 1–528) fused to T25 in pKT25 (KmR)	15
pKT25-sigA(RT)	sigA (E248R/K251T; codons 1–528) fused to T25 in pKT25 (KmR)	This study
pKT25-sigA(VR)	sigA (L257V/Y258R; codons 1–528) fused to T25 in pKT25 (KmR)	This study
pKT25-sigA(RTVR)	sigA (E248R/K251T/L257V/Y258R; codons 1–528) fused to T25 in pKT25 (KmR)	This study
pUT18-rbpA(Mtb)	Mtb rbpA (codons 1–111) fused to T18 in pUT18 (AmpR)	15
pUT18-rbpA(Sco)	Mtb rbpA (codons 1–124) fused to T18 in pUT18 (AmpR)	15
pSETΩ	Integrative cloning vector (SpcR)	15
pSX530	pSETΩ containing at the BamHI site a BglII-fragment including rbpA	15
pSX530(R80A)	As pSX530 but with rbpA-R80A mutation	This study
pSX530(M85A)	As pSX530 but with rbpA-M85A mutation	This study
pET15b	Eco expression vector (His6-tagged; AmpR)	Novagen
pET15b-sigA	Mtb sigA cloned in pET15b as an NdeI-BglII fragment	This study
pET15b-sigA(RT)	Mtb sigA (E248R/K251T) cloned in pET15b as an NdeI-BglII fragment	This study
pET15b-sigA(VR)	Mtb sigA (L257V/Y258R) cloned in pET15b as an NdeI-BglII fragment	This study
pET15b-sigA(RTVR)	Mtb sigA (E248R/K251T/L257V/Y258R) cloned in pET15b as an NdeI-BglII fragment	This study
pET20b	Eco expression vector (native; AmpR)	Novagen
pSX500	pET20b containing Mtb rbpA cloned as NdeI-BamHI fragment	15
pSX500(R79A)	pET20b containing Mtb rbpA-R79A cloned as NdeI-BamHI fragment	This study
pSX500(M84A)	pET20b containing Mtb rbpA-R79A cloned as NdeI-BamHI fragment	This study
pET SUMO	Eco expression vector (N-terminal His6/SUMO fusion; KmR)	Life Technologies
pET28	Eco expression vector (His6-tagged; KmR)	Novagen
pET28-rbpA	Mtb rbpA cloned into pET28	This study
pET28-rbpA(R79A)	As pET28-rbpA but with R79A allele	This study
pSUMO-sigA/rbpA	Mtb sigA (codons 224–364) and rbpA (codons 1–111) chemically synthesized and cloned as a BamHI-HindIII fragment into a pET SUMO, generating an SUMO–sigA fusion	This study
pSUMO-rbpA(1–111)	pET SUMO containing Mtb rbpA (codons 1–111) fused to the C terminus of SUMO as a BamHI-HindIII fragment	This study
pSUMO-rbpA(1–71)	pET SUMO containing Mtb rbpA (codons 1–71) fused to the C terminus of SUMO as a BamHI-HindIII fragment	This study
pSUMO-rbpA(72–111)	pET SUMO containing Mtb rbpA (codons 72–111) fused to the C terminus of SUMO as a BamHI-HindIII fragment	This study
PCR primers
Plasmid constructs
pSUMO-RbpA(1–111)	GGATCCCATATGGCTGATCGTGTCCTGAGG
	AAGCTTTCAGCCG CGCCGACGTGACCGAATG
pSUMO-RbpA(1–71)	GGATCCCATATGGCTGATCGTGTCCTGAGG
	AAGCTTTCACTCGGGCAGGTCGCCCTCGATCAG
pSUMO-RbpA(72–111)	GGATCCCATATGGAGCCGAAGAAGGTTAAGCCG
	AAGCTTTCAGCCG CGCCGACGTGACCGAATG

Primer sequences (5′ to 3′) are listed with restriction sites indicated in bold.

X-Ray Crystal Structure of the Mtb RbpA–σ^A₂ Complex.

To provide insight into RbpA function as a transcription activator, we determined the crystal structure of the Mtb RbpA–σ^A₂ complex to 2.2-Å resolution (Figs. 1C and S1A and Table S2). Although crystallization trials were set up with a purified complex of an His₆-SUMO-σ^A₂ fusion (containing σ^A₂ residues 224–364) and full-length RbpA (1–111), MALDI-TOF analysis of the crystal contents revealed that both proteins were N-terminally proteolytically degraded, resulting in crystals containing multiple σ^A₂-fragments (lacking up to 18 N-terminal residues) and multiple RbpA fragments (all including the BL and SID) (Fig. 1A and SI Materials and Methods). Electron density maps revealed RbpA-SID (77–108) bound to σ^A₂ (242–363) (Fig. 1C). Although the MALDI-TOF analysis indicated that the mixture of RbpA fragments crystallized included intact RbpA molecules, additional electron density was missing, and the rest of RbpA was presumed disordered or absent.

Fig. S1.

Molecular interactions in the Mtb RbpA–σ^A₂ complex. (A) Stereoview of a selected portion of a simulated annealing composite omit electron density map (2mFo-DFc) calculated with RbpA omitted and contoured to 1.0σ. Map was calculated in Phenix (34) with a starting temperature of 5,000 K. Clear electron density for RbpA is shown. RbpA (purple) and σ^A₂ (orange with the NCR colored cranberry) are shown in stick, highlighting the hydrophobic interactions between RbpA residues L94 and L98 with σ^A_NCR-Y258. (B) The schematic highlights polar and ionic interactions between RbpA-SID and σ^A. Interactions between residues are indicated by lines, with ionic interactions colored in red, hydrogen bonds colored in gray, and hydrogen bonds mediated by waters colored in blue. (C) The schematic shows nonpolar (van der Waals) interactions between RbpA-SID and σ^A.

Table S2.

Crystallographic statistics for RbpA–σ^A₂ crystals

Parameter	RbpA-σA2
Data collection
Space group	P3121
Combined datasets	1
Cell dimensions
a, b, c	84.459, 84.459, 73.701
α, β, γ	90, 90, 120
Wavelength (Å)	1.74
Resolution (Å)	36.64–2.202 (2.282–2.202)†
Total reflections	182,175 (9,287)
Unique reflections	15,679 (1,423)
Multiplicity	11.6 (6.2)
Completeness (%)	99 (96)
<I>/σI	22.25 (2.07)
Wilson B factor	42.40
Rmerge‡	0.1189 (0.7155)
Rmeas‡	0.1244 (0.7709)
CC1/2‡	0.997 (0.998)
CC*‡	0.999 (0.999)
Refinement
Rwork/Rfree	0.1951/0.2310 (0.2864/0.2966)
CCwork/CCfree§	0.941/0.919 (0.818/0.841)
No. nonhydrogen atoms
Protein	1,260
Ligand/ion	19
Water	58
Protein residues	154
B factors
Protein	56.41
Ligand/ion	96.27
rmsd
Bond lengths (Å)	0.005
Bond angles (°)	0.7
Clash score	5.88
Ramachandran favored (%)	98
Ramachandran outliers (%)	0

CC, correlation coefficient.

^†Values in parentheses are for the highest-resolution shell.

^‡Ref. 39.

^§Ref. 40.

The structure shows that the RbpA-SID comprises two α-helices (α1 and α2) (Fig. 1C), confirming previous sequence-based structural predictions (15–17). Four residues of 15-residue RbpA-BL connecting the RbpA-SID to the RCD are also visible in the structure. Both α-helices of the RbpA-SID contact σ^A₂, forming an intermolecular interface with a buried surface area of 948 Ų.

The RbpA-SID makes extensive contacts with residues from conserved regions of σ^A₂ (1.2 and 2.3) as well as the NCR (Figs. 1C, 2 A–C, and S1). Interacting with the primary σ-NCR is a property that RbpA shares with the unrelated Chlamydia trachomatis transcription factor GrgA, which binds to the Chlamydia trachomatis σ⁶⁶-NCR (20). Moreover, although structurally distinct from RbpA, the holoenzyme assembly factor Crl from enteric bacteria interacts with the equivalent region of the group 2 σ-factor σ^S (21, 22). The cocrystal structure of RbpA–σ^A₂ represents the first structure, to our knowledge, of an activator (or any protein) interacting with the NCR of a housekeeping-σ.

Fig. 2.

The X-ray crystal structure of the Mtb RbpA–σ^A₂ complex explains the σ-specificity of RbpA. (A) Amino acid sequence alignment of conserved regions 1.2–2.3 in σ₂-domains of Mtb σ^A, σ^B, and σ^F and Sco σ^HrdC. The following amino acids, written in single-letter nomenclature, were considered homologous and grouped as indicated within parenthesis: VLMA, RK, ED, YFW, TS, and QN. Remaining residues (H, C, P, and G) were not grouped. The σ^A-residues that contact RbpA are indicated by red dots, with those that are targeted for mutagenesis outlined in black. (B) Alignment of RbpA BL and SID from diverse representatives of the order Actinobacteria. Residues were considered in the same conservation group as in A. Green dots indicate residues that are suggested by the RbpA/RPo model to come in close contact with DNA, blue dots highlight conserved arginine residues shown to be important for binding σ (15), and white dots indicate residues forming a hydrophobic interface between RbpA and σ and shown to be important for binding σ (this study). Agr, Actinomyces graevenitzii; Cac, Catenulispora acidiphila; Fsp, Frankia sp.; Gar, Glycomyces arizonensis; Jga, Jiangella gansuensis; Krd, Kineococcus radiotolerans; Mau, Micromonospora aurantiaca. (C) Highlighted RbpA-SID residues required for σ^A-binding. RbpA residues R88 and R89 form a cluster of salt bridges with σ^A-residues E254 and D336 (selected interactions indicated by dotted lines). RbpA L94, L97, and L98 make van der Waals interactions with σ^A-residues Y258, A259, I284, and K251. (D) BTH analysis of Mtb RbpA interactions with Mtb σ^A-mutants predicted to be defective in RbpA binding. Selected σ^A-residues that contact RbpA were changed to equivalent residues in Sco σ^HrdC as described in the text. The coding sequences of rbpA were fused to the T18 subunit of Bordetella pertussis adenylate cyclase in pUT18, whereas the σ-factor genes were fused to the T25 subunit in pKT25 (15). β-Gal assays were performed in triplicate (SDs indicated), and results are presented as percentage of Miller units relative to WT Mtb σ^A. (E) RbpA activation of the vapB10 promoter using σ^A-mutants highlighted in A. RbpA activation on mutant-σ σ^A(VR) or σ^A(RTVR) was compared with WT σ^A(WT) using multiple-round in vitro runoff transcription reactions. Reactions contained core Mbo RNAP (50 nM), WT σ^A(WT) (250 nM), σ^A(VR) (250 nM), or σ^A(RTVR) (250 nM) and the presence or absence of RbpA (500 nM) as indicated. A graphical representation based on triplicate datasets, with SD indicated, is illustrated above and normalized to the data obtained with σ^A(WT) in the absence of RbpA.

Previous studies identified two conserved arginine residues (Sco RbpA R89 and R90 corresponding to Mtb R88 and R89) (Fig. 2B) that are critical for σ-binding (15). The structure shows that these two residues form extensive electrostatic interactions with σ^A (Fig. 2C), explaining the mutagenesis results. Mtb RbpA-R88 in α1 of the SID forms a salt bridge with σ^A-residue E254, whereas RbpA-R89, located in the linker between the two SID α-helices, makes salt bridges with both E254 and D336 of σ^A.

A cluster of conserved hydrophobic RbpA residues (L94, L97, and L98) (Fig. 2 B and C), located on RbpA-SID α2, makes extensive van der Waals contacts with residues of σ^A (Figs. 2C and S1). To determine if the interactions observed in the crystal structure are of wider importance, we performed bacterial two-hybrid (BTH) assays with an Sco RbpA mutant, in which the residues corresponding to Mtb RbpA L97 and L98 (Sco RbpA V98 and L99) (Fig. 2B) were changed to alanine. The results reveal that these branched hydrophobic residues are necessary for Sco RbpA–σ^HrdB binding (Fig. S2).

Fig. S2.

BTH analysis of Sco RbpA WT and RbpA-V98A/L99A interactions with Sco σ^HrdB. (A) β-Gal assays were conducted with Eco BTH101 strains containing C1, pUT18 and pKT25; C2, pUT18 and pKT25-hrdB; C3, pUT18-rbpA and pKT25; RbpA (WT), pUT18-rbpA(Sco) and pKT25-hrdB; and RbpA (V98A/L99A), pUT18-rbpA(Sco)-V98A/L99A and pKT25-hrdB. rbpA coding sequences were fused to the T18 subunit of Bordetella pertussis adenylate cyclase in pUT18, whereas the σ-factor genes (domains σ₂, σ₃, and σ₄) were fused to the T25 subunit in pKT25 (15). β-Gal assays were performed in triplicate (SDs indicated), and results are presented as percentage of Miller units relative to results obtained with WT Sco RbpA). (B) Western analysis of total cell extracts of Eco BTH101 strains containing: RbpA (WT), pUT18-rbpA(Sco) and pKT25-hrdB; and RbpA (V98A/L99A), pUT18-rbpA(Sco)-V98A/L99A and pKT25-hrdB. Extracts were prepared at the time of β-gal assay by precipitation with 10% (vol/vol) trichloroacetic acid, and T18-RbpA fusion proteins were detected using an antiadenylate cyclase toxin antibody (3D1; Santa Cruz Biotechnology).

Identification of RbpA Residues Involved in σ-Selectivity.

In addition to binding primary σ-factors, RbpA binds some group 2 σ-factors (Mtb σ^B and Sco σ^HrdA) but not others (Sco σ^HrdC and σ^HrdD) and does not interact with the more diverse group 3 (e.g., Mtb σ^F) or 4 (e.g., Sco σ^R) σ-factors (15, 17). The RbpA-SID–σ^A₂ structure provides a basis to understand the σ-selectivity of RbpA. An alignment of σ₂-amino acid sequences of Mtb σ^A and σ^B with Mtb σ^F and Sco σ^HrdC (Fig. 2A) revealed that, of 21 Mtb σ^A-residues contacting RbpA, 16 were identical in Mtb σ^B, whereas only 4 were identical in Mtb σ^F, explaining why RbpA is specific to groups 1 and 2 σ-factors (17). Although there was extensive conservation between Mtb σ^A and Sco σ^HrdC among most RbpA-interacting residues, there were several positions in Sco σ^HrdC that were substituted with physicochemically dissimilar amino acids that might impede RbpA binding (e.g., σ^A E248/σ^HrdC R59, σ^A K251/σ^HrdC T62, and σ^A Y258/σ^HrdC R69). We tested this idea by mutating Mtb σ^A-residues E248 and K251 in conserved region 1.2 and L257and Y258 in the NCR, changing each position to the equivalent residue in Sco σ^HrdC (Fig. 2A). σ^A(RT) contained the mutations E248R and K251T, σ^A(VR) contained L257V and Y258R mutations, and σ^A(RTVR) had all four positions altered to the equivalent positions in Sco σ^HrdC. BTH assays suggested that all three mutants were defective in RbpA binding (Fig. 2D). To support this, σ^A(VR) and σ^A(RTVR) mutants were overexpressed, purified, and tested by in vitro transcription on the vapB10p promoter template. In the absence of RbpA, σ^A(VR) and σ^A(RTVR) directed basal transcription activity to the same extent as WT σ^A. However, neither was responsive to RbpA, confirming that the interaction between σ and RbpA is crucial for the transcription activation role of RbpA (Fig. 2E).

RbpA Interacts Directly with Promoter DNA.

To gain insight into the role of RbpA in transcription initiation by RNAP holoenzyme, we generated a structural model of the complex between RbpA and RPo by superimposing the conserved regions of σ^A₂ from the Mtb RbpA-SID–σ^A₂ complex (Fig. 1C) onto the corresponding regions of Taq σ^A₂ in an RPo model (6, 23) (0.681-Å rmsd over 93 Cα atoms) (Fig. 3), resulting in an RbpA/RPo model with no steric clashes. In the model, the RbpA-SID–σ^A₂ interaction positions the RbpA-BL (Fig. 1A) on the minor-groove side of the duplex promoter DNA just upstream of the −10 element, with absolutely conserved RbpA-R79 (Fig. 2B) positioned to interact with the nontemplate strand (nt strand) DNA phosphate backbone at the −14 position with respect to the transcription start site at +1 (Fig. 3B). In addition, absolutely conserved RbpA-M84 (Fig. 2B) in α1 is positioned to play a role in DNA binding through van der Waals interactions (Fig. 3B). Also, note that the next three residues N-terminal to the modeled portion of RbpA are all lysines (K73, K74, and K76), which may also play a role in forming electrostatic interactions with the DNA phosphate backbone (Fig. 3B).

Fig. 3.

Structural model of Mtb RbpA-SID on an RPo complex. (A) A model of the RbpA-SID–RPo complex was generated by superimposing conserved regions of Mtb σ^A₂ (from the Mtb RbpA_SID–σ^A₂ structure) with Taq σ^A in an RPo model. Protein and DNA elements are colored as indicated. RNAP is shown as a molecular surface, RbpA-SID–σ^A₂ is shown in ribbon, and DNA is shown as a phosphate backbone worm. (B) A magnified view of the RbpA/RPo model shows that RbpA is positioned to make contacts with the DNA upstream of the −10 promoter element. RbpA amino acids R79, located in the RCD–SID linker, and M84, located in α1, are in position to contact the nt-strand phosphate backbone at the −13/−14 position. RbpA-M84 makes nonpolar contacts with Mtb σ^A-residue K334, which interacts with −12T at the beginning of the transcription bubble. RbpA sequence reveals three conserved, positively charged lysine residues (K76, K74, and K73; represented by purple dots) located in the RCD–SID linker that would be well-positioned to interact with the negatively charged DNA phosphate backbone.

To test the hypothesis of an RbpA–DNA interaction, we used formaldehyde cross-linking, which links atoms 2 Å apart (24). To form stable RNAP–DNA complexes, we used a fork-junction template (6, 25), comprising the vapB10 promoter (VFJ-28) (Figs. 4A and S3A) along with an anticonsensus control (VFJ-28anti) (Fig. S3A). The inclusion of RbpA in cross-linking reactions led to the appearance of a new band (Fig. 4A, lane 3), which was confirmed to be an RbpA-DNA cross-link, because a slower migrating band appeared when the assay was repeated with SUMO-RbpA (Fig. 4A, lane 4). Furthermore, RbpA-DNA cross-linking was detected with SUMO-RbpA(72–111), which includes the RbpA-BL and SID, but not with SUMO-RbpA(1–71), which lacks the RbpA-SID essential for σ^A-binding and activity (Fig. 4A, lanes 5 and 6). This finding is consistent with the ability of SUMO-RbpA(72–111) but not SUMO-RbpA(1–71) to activate transcription (Fig. 1B). No cross-linking was detected with VFJ-28anti or reactions that lacked core RNAP (Fig. S3B), indicating that RbpA–σ^A₂ interactions with DNA occur only in the context of RNAP holoenzyme–promoter complexes.

Fig. 4.

RbpA directly contacts the DNA and increases affinity of holenzyme to promoter DNA. (A) RbpA cross-links to fork-junction promoter DNA. Reactions contained VFJ-28 DNA 5′-labeled on the nt strand, core Mbo RNAP (E) at 200 nM, σ^A (1 μM), and native or SUMO-fused derivatives of RbpA (2 μM) as indicated. Complexes were allowed to form at 37 °C for 15 min before treatment with formaldehyde. Cross-linked species were separated by 4–12% (wt/vol) SDS/PAGE and visualized by phosphorimaging. The vapB10 promoter-based fork-junction DNA template VFJ-28 is illustrated, indicating the −10 element (blue). *nt Strand. (B) RbpA increases RNAP affinity for promoter DNA. Representative fluorescence anisotropy binding curves show RNAP binding to Cy3-labeled vapB10 promoter DNA (Fig. S4A). SE bars and average K_d values are based on triplicate trials. Inset shows average fold activation (normalized to no RbpA). WT RbpA increases affinity of RNAP for promoter DNA, whereas RbpA-R79A has little to no effect on binding. Error bars are based on average fold change from nine experiments. (C) RbpA residues R79 and M85, predicted to bind DNA, are important for in vitro transcription activation. Multiple-round in vitro runoff transcription reactions containing Mbo RNAP (50 nM), σ^A (250 nM), and RbpA (500 nM or 1.25 μM) using the vapB10 promoter as a template. A graphical representation based on triplicate datasets with SD indicated is illustrated above normalized to the data obtained with σ^A in the absence of RbpA.

Fig. S3.

Formaldehyde cross-linking indicates specificity for promoter DNA interactions and a requirement for holoenzyme to form RbpA-DNA cross-links. (A) The vapB10 promoter is indicated with the −10 element highlighted in blue. VFJ-28 is the fork-junction template based on the vapB10 promoter. VFJ-28anti is identical to VFJ-28 apart from changes at three positions in the −10 element (underlined). (B) Formaldehyde cross-linking reactions contained VFJ-28 or VFJ-28anti, core Mbo RNAP (E; 200 nM), σ^A (400 nM), and native or SUMO-fused derivatives of RbpA (at increasing molar concentrations relative to E) as indicated. Cross-linked species were separated by 4–12% SDS/PAGE and visualized by phosphorimaging.

An RbpA–promoter DNA interaction would be predicted to increase the overall affinity of RNAP for promoter DNA. To test this hypothesis, we measured RNAP binding to a Cy3-labeled duplex vapB10 promoter template (Fig. S4A) using a fluorescence anisotropy assay (26). We found that addition of RbpA decreased the dissociation constant, K_d, for binding to the promoter DNA nearly twofold (Fig. 4B), consistent with the modest activation activity of RbpA in abortive initiation (Fig. S4B) and runoff (Fig. 1B) transcription assays. However, the RbpA-R79A mutant had no significant effect on RNAP promoter binding (Fig. 4B), consistent with the hypothesis that RbpA-R79 plays an important role in DNA binding (Fig. 3B). WT RbpA had no significant effect on RNAP binding to Cy3-labeled ssDNA comprising only the −10 and discriminator elements (Table S3), supporting the structure-based hypothesis that the effect of RbpA on RNAP-promoter binding is through the interactions of RbpA with duplex DNA just upstream of the −10 element (Fig. 3).

Fig. S4.

Fluorescence anisotropy assays. (A) The Cy3-labeled vapB10 dsDNA probe (−36 to +1) used for anisotropy experiment with the −10 element highlighted in blue. (B) Multiround abortive initiation transcription assays on the vapB10 promoter showing activation by RbpA. SE bars are based on four trials.

Table S3.

Average K_d values of RNAP binding to Cy3-labeled ssDNA (−12 to +1) with and without RbpA

Promoter	Sequence	No RbpA Kd (μM)	RbpA Kd (μM)
Consensus	*TATAATGGGAAGG	0.38 ± 0.07	0.33 ± 0.04
AP3	*TAGACTGGCAGGG	2.24 ± 0.41	1.88 ± 0.31
vapB	*TATGATATGGTGT	1.62 ± 0.26	1.77 ± 0.15

Average K_d values of RNAP binding to Cy3-labeled ssDNA (−12 to +1) were determined using fluorescence anisotropy. The addition of RbpA to the assay has no significant effect on RNAP binding to ssDNA composed of the −10 and discriminator elements; −10 element is bold. Error is based on triplicate trials.

DNA Interaction Through RbpA-R79 Is Critical for Transcription Activation.

We hypothesized that an interaction between RbpA and DNA upstream of the −10 element might underlie the transcription activation function of RbpA. To test this, we used multiround in vitro transcription assays and found that, although Mtb RbpA-M84A retained partial activity, Mtb RbpA-R79A completely failed to stimulate transcription from a vapB10 promoter template (Fig. 4C, lanes 5 and 6). To investigate the importance of these residues in vivo, we performed experiments in Sco, where RbpA is required for normal growth but not essential for viability. The Sco mutants rbpA-R80A and rbpA-M85A (corresponding to Mtb RbpA R79A and M84A, respectively) were cloned into an integrative vector and used to transform the ∆rbpA Sco mutant S129 (15). The rbpA-R80A and rbpA-M85A alleles retarded growth, leading to smaller colonies on agar plates compared with the control rbpA-WT strain (Fig. S5). These data indicate that the conserved Mtb/Sco RbpA R79/80 and M84/85 residues are critical for normal rbpA function in vivo.

Fig. S5.

RbpA R80 and M85 are required for normal growth of Sco. Sco S129 (∆rbpA) containing the integrative vector pSETΩ and S129 derivatives containing WT (rbpA-WT) or mutant derivatives of rbpA (rbpA-R80A or rbpA-M85A) cloned in pSETΩ were grown on mannitol-soya agar containing spectinomycin (25 μg/mL). The agar plate was photographed after incubation at 30 °C for 3 d.

SI Materials and Methods

Protein Expression and Purification.

For crystallography, σ^A₂ (codons 224–364) and rbpA (codons 1–111) from Mtb were chemically synthesized as a single DNA fragment with Eco codon use and independent T7 promoter and translation initiation signals for rbpA (Genscript), cloned as a BamHI-HindIII fragment into an His₆ pET SUMO expression vector, and transformed into Eco BL21(DE3) cells (Novagen). Transformed cells were grown at 37 °C in the presence of 50 μg/mL kanamycin to an OD₆₀₀ of 0.6, at which point the temperature was lowered to 30 °C, and protein expression was induced with 500 µM isopropyl β-D-1-thiogalactopyranoside (IPTG) for 3 h. Cells were harvested by centrifugation and resuspended in lysis buffer [20 mM Tris⋅HCl, pH 8.0, 0.5 M NaCl, 5 mM imidazole, 5% (vol/vol) glycerol, 0.5 mM β-mercaptoethanol] supplemented with 1 mM PMSF and protease inhibitor mixture. Cells were lysed by French press (Avestin), and lysate was cleared by centrifugation. Clarified lysate was loaded on a Hi-Trap IMAC Ni²⁺-Chelating Column (GE Healthcare) and eluted with lysis buffer containing 250 mM imidazole. The elution was directly loaded on a size-exclusion column (SuperDex-200 16/16; GE Healthcare) equilibrated with 20 mM Tris⋅HCl (pH 8), 0.5 M NaCl, 5% (vol/vol) glycerol, 1 mM DTT, and 1 mM EDTA. The sample was concentrated to 15 mg/mL by centrifugal filtration and stored at −80 °C. For fluorescence anisotropy, Mtb RbpA and Mtb RbpA-R79A were cloned into an His₆ pET28-based vector, and the plasmid was transformed into Eco Rosetta 2 BL21(DE3) cells. The proteins were overexpressed and purified using the same buffers and techniques as the RbpA–σ₂ coexpressed complex. Core RNAP, σ^A, and CarD were overexpressed and purified using the previously described method (14).

For in vitro transcription and cross-linking assays, pSX500, which contains Mtb rbpA cloned in pET20b, as well as mutant derivatives were overexpressed in Eco BL21(DE3) pLysS after a 15-min cold shock of cultures (OD₆₀₀ of 0.4–0.7) on ice, which improved protein solubility, and induction with 1 mM IPTG for 3 h. Protein was purified by ion exchange chromatography (HiTrap Q Sepharose Fast Flow and Mono Q 5/50 GL; GE Healthcare) and size-exclusion chromatography (HiLoad 16/60 Superdex 200; GE Healthcare). To generate N-terminal SUMO fusions of Mtb RbpA (amino acid residues 1–111) and truncated derivatives (amino acid residues 1–71 and 72–111), RbpA derivatives were cloned into pET SUMO (Invitrogen) as BamHI–HindIII fragments, overexpressed in Eco BL21(pLysS) as described above, and purified by sequential Ni-affinity chromatography (Ni²⁺-Charged Iminodiacetic Acid Sepharose; Sigma-Aldrich), ion exchange chromatography (Mono Q 5/50 GL; GE Healthcare), and size-exclusion chromatography (HiLoad 16/60 Superdex 200; GE Healthcare); σ^A and mutant derivatives [σ^A(VR) and σ^A(RTVR)] were cloned into pET15b and purified from Eco BL21(pLysS) after induction with 1 mM IPTG for 3 h at 30 °C. Cell pellets were resuspended in ice-cold lysis buffer, and after cell lysis by sonication, clearing of cell lysates by centrifugation, and binding to an Ni column (see above), protein was eluted by thrombin cleavage (Sigma-Aldrich) and then, further purified by size-exclusion chromatography as outlined above.

Crystallization of RbpA–σ^A₂ Complex.

Crystals of the RbpA–σ^A₂ complex were grown at 22 °C by sitting drop vapor diffusion against a reservoir solution of 0.1 M Tris (pH 8.5) and 0.5 M ammonium sulfate at a protein concentration of 15 mg/mL. Drops were set up with a 1:1 ratio of His₆-SUMO-σ^A₂-RbpA:crystallant. Crystals measuring ∼50 μm took 3 d to grow and were cryoprotected in reservoir solution plus 30% (vol/vol) ethylene glycol.

MALDI-TOF MS of Crystals.

One to two crystals were removed from the mother liquor, briefly washed in cold water, and subsequently, dissolved in matrix solution, which consisted of a saturated solution of α-cyano-4-hydroxycinnamic acid in a 1:3:2 (vol/vol/vol) mixture of formic acid:water:isopropanol. An aliquot of 0.5 μL protein–matrix solution was transferred onto an MALDI plate precoated with an ultrathin layer. The ultrathin layer was prepared and laid on the plate as previously published (30, 31). The sample spots were washed for a few seconds with 2 μL cold 0.1% aqueous trifluoroacetic acid (TFA) solution. MALDI spectra were acquired in linear, delayed extraction mode using a Spiral TOF JMS-S3000 (JEOL) equipped with an Nd:YLF laser delivering 10-Hz pulses at 349 nm. Delayed extraction time was set at 1 ms, and acquisition was performed with a sampling rate of 2 ns. Each MALDI spectrum corresponded to an average of 500 scans. Mass calibration was performed using a technique of pseudointernal calibration, wherein a few shots on a nearby calibrant spot were collected and averaged with the sample shots into a single spectrum. The spectra were processed and analyzed using MoverZ (Proteometrics, LLC). MALDI-TOF analysis of the crystals revealed that both σ^A₂ and RbpA were proteolytically degraded at their N termini. The crystals were composed of mixtures of σ^A missing up to 18 N-terminal residues and two sets of RbpA fragments. Fragments of σ^A detected included S224-R364, A230-Q362, Y231-A363, L233-R364, L240-Q362, and N242-R364. Two groups of N-terminal variants of RbpA were detected. One group contained most of the N-terminal region, including all of the RCD (beginning at M1, R10, S15, E17, and R20). The second group lacked the RCD and included N-terminal variants starting at R57, G59, L64, G67, and K74). All RbpA variants contained the native C-terminal residue G111.

Data Collection and Refinement of the σ^A₂–RbpA Complex.

X-ray diffraction data were collected at the Brookhaven National Synchroton Light Source X29 Beamline, integrated, and scaled using HKL2000 (32). Electron density maps were generated using molecular replacement with an Mtb σ^A₂-homology model (ExPASy SWISS MODEL) based on a previously solved structure of the corresponding domain from Taq σ^A [1KU2 (4)]. The model was built using reiterative cycles of manual building with COOT (33) and refinement with Phenix (34) (Table S2). The final model included residues 242–363 of σ^A and 77–108 of RbpA. The RCD of RbpA, which lacked density, was presumed disordered. The PDBePISA server (www.ebi.ac.uk/pdbe/pisa/) was used to calculate intermolecular buried surface areas (35).

Fluorescence Anisotropy.

Fluorescence measurements were performed using an Infinite M1000pro Plate Reader (Tecan) in a 384-well plate with a final reaction volume of 25 μL. Holoenzyme was formed at 37 °C using a 1:1.5 ratio of core RNAP:σ^A. All proteins were dialyzed and then, serially diluted in anisotropy buffer consisting of 100 mM potassium-glutamate, 10 mM Hepes (pH 8), 10 mM MgCl₂, 0.1 mM DTT, 0.02% Tween 20, and 5% (vol/vol) glycerol. When used, RbpA was added to RNAP in fivefold excess. Protein complexes were serially diluted in anisotropy buffer to obtain RNAP concentrations ranging from 5 nM to 3 μM. For assays with Cy3-labeled ssDNA extending from −12 to +1 (Trilink) (Table S3), DNA was diluted in anisotropy buffer and added to the protein mixture at a final concentration 200 nM. For assays performed with dsvapB10 DNA extending from −36 to +10, Cy3-labeled nt-strand DNA (Trilink) was annealed to unlabeled template-strand DNA (Oligos Etc.), diluted in anisotropy buffer, and added to the protein mixture at a final concentration of 10 nM. Measurements of RNAP binding to dsDNA were noisy without the addition of CarD [a result of the unstable open complex and short half-life of Mycobacterium bovis (Mbo) RNAP (14)], and therefore, these assays included CarD (in threefold excess over RNAP). Data were analyzed using Prism software. Assays with ssDNA were performed in triplicate. For assays with duplex DNA, three experiments composed of side-by-side triplicate trials for each condition were performed.

Formaldehyde Cross-Linking.

Fork-junction DNA was prepared by labeling the 5′ end of the nt-strand oligonucleotide using [γ³²P]ATP and annealing with threefold molar excess of template-strand oligonucleotide followed by purification using an NAP-5 Column (GE Healthcare). The fork-junction template VFJ-28 contained an 18-bp double-stranded region from −28 to −12 followed by a single-stranded extension from −11 to −3 on the nt strand (Fig. S3A). In control experiments, an anticonsensus −10 sequence was used (VFJ-28anti), changing the −10 element from TATGAT to AGTGAC; σ^A was mixed with Mtb RbpA and mutant and truncated derivatives on ice for 10 min before the addition of core Mbo RNAP. Reactions were then set up containing 400 nM σ^A, 200 nM core RNAP, various concentrations of RbpA, and ∼10 nM fork-junction DNA in binding buffer [40 mM Tris⋅HCl, pH 7.9, 10 mM MgCl₂, 0.6 mM EDTA, 20% (vol/vol) glycerol, 0.8 mM potassium phosphate, pH 7.5, 0.75 mM DTT, 0.125 mg/mL BSA] and equilibrated at 37 °C for 15 min. Cross-linking was initiated by the addition of formaldehyde (60 mM final) and stopped by the addition of 2× stop buffer [4% (wt/vol) SDS, 20% (vol/vol) glycerol, 10% (vol/vol) β-mercaptoethanol, 0.004% bromophenol blue, 125 mM Tris⋅HCl, pH 6.8] after 2 min. Samples were loaded on Precast NuPAGE 4–12% (wt/vol) Bis-Tris Gels (Invitrogen) at 100 V for 3 h at 10 °C, dried, and then, analyzed by phosphorimagery.

In Vitro Transcription.

Multiround runoff and abortive initiation in vitro transcription assays were carried out using core Mbo RNAP, purified Mtb σ^A, and Mtb RbpA or mutant and truncated derivatives essentially as described (15). The vapB10 promoter template was amplified by PCR using the oligonucleotides vapB10_F (5′-GCGCTGAAGAGGGCGTTGCAC) and vapB10_R (5′-TTCAGCAGGAGGCGGATCAG). Multiround assays, giving rise to a 109-nt runoff transcription product, were performed at 37 °C and contained 5 nM template in reaction buffer [40 mM Tris⋅HCl, pH 7.9, 10 mM MgCl₂, 0.6 mM EDTA, 20% (vol/vol) glycerol, 0.8 mM potassium phosphate, pH 7.5, 0.75 mM DTT, 0.125 mg/mL BSA]. Reactions were preincubated at 37 °C for 10 min before the addition of the NTP mix [200 μM ATP, GTP, and CTP, 50 μM UTP, including 5 μCi [α-³²P]UTP (>800 Ci/mmol); GE Healthcare] and incubated for 10 min. Reactions were halted by the addition of an equal volume of in vitro loading dye [80% (wt/vol) formamide, 0.01% (wt/vol) xylene cyanol, 0.01% (wt/vol) bromophenol blue], separated on denaturing 8% (wt/vol) polyacrylamide gels, and quantified by phosphorimaging. Abortive initiation assays were performed at 37 °C in a buffer containing 10 mM Tris⋅HCl, pH 8.0, 100 mM potassium-glutamate, 10 mM MgCl₂, 0.1 mM EDTA, 0.1 mM DTT, and 50 μg/mL BSA. Core RNAP (50 nM), σ^A (250 nM), and RbpA (500 nM when used) were preincubated at 37 °C for 5 min. DNA (10 nM) was added, and the reaction was incubated for 15 min to allow RPo to form. Abortive transcription was initiated with an ApU dinucleotide primer (250 μM), [α-³²P]GTP (1.25 μCi), and unlabeled GTP (50 μM). After 10 min, reactions were quenched with 2× stop buffer (8 M urea, 0.5× Tris/borate/EDTA buffer, 0.05% bromophenol blue, 0.05% xylene cyanol) and separated on a 23% (wt/vol) urea-polyacrylamide gel. Abortive products were visualized by phosphorimagery and quantified using Image J.

Discussion

We propose that a key role of the RbpA-SID–σ^A₂ interaction in transcription activation is to position the RbpA-BL near the upstream edge of the −10 promoter element to facilitate its interaction with the DNA phosphate backbone. Adding favorable protein/DNA contacts to the transcription initiation complex would potentially stabilize a transcription initiation intermediate, which is consistent with findings that RbpA activates transcription primarily by stimulating RPo formation (15, 17). The role of RbpA-R79, which structural modeling predicts could contact the nt-strand DNA phosphate backbone at the −14 position (Fig. 3B), seems to be particularly significant. This residue is absolutely conserved among RbpA orthologs (Fig. 2B), stabilizes the binding of RNAP to promoter DNA (Fig. 4B), and is required for RbpA transcription activation in vitro (Fig. 4C) and full RbpA function in vivo (Fig. S5). Strikingly, the RbpA-BL also contains three lysine residues (K73, K74, and K76) (Figs. 2B and 3B), which although not modeled in the structure, might also contribute to interactions with the DNA phosphate backbone around the −14/−15 positions (Fig. 3B).

Although the RbpA-BL and SID together are sufficient for partial transcription activation in vitro, full activation requires full-length RbpA (the N-terminal tail/RCD) (Fig. 1B). We suggest that the N-terminal tail and/or RCD likely form additional interactions with RNAP. Interactions between RbpA and core RNAP have been proposed previously, and putative interactions have been mapped to two widely spaced regions of RNAP (18, 27). However, our RbpA-SID/RPo structural model is completely incompatible with previous claims that RbpA interacts with RNAP either in the active-site channel near the Rif binding site (27, 28) (Fig. S6A) or with a different region of the β-subunit (18) (Fig. S6B). Because of the location and orientation of the N terminus of the RbpA-SID in our structural model, it would be impossible for any part of RbpA to span the distance from the N terminus of our model to either of the putative sites (Fig. S6). The binding site near the Rif pocket was originally inferred based on functional data (28) and then refined on the basis of a single cross-link (27). The other β-subunit determinant was identified using cleavage experiments through hydroxyl-radicals generated from Fe(III) (S)-1-(p-Bromoacetamido-benzyl) ethylene diamine tetraacetic acid (Fe-BABE) attached to the lone Cys residue of the RbpA-RCD. In any case, whether the RbpA-RCD or N-terminal tail binds RNAP at all has not been established, and if it does, our RbpA-SID/RPo model suggests that the likely RNAP binding determinant is on the β′-subunit on the top of the clamp domain (Fig. S6A). More experiments will be required to resolve these inconsistencies. A possible interaction with the RNAP clamp domain has also been proposed for Crl, a σ^S-RNAP holoenzyme assembly factor found in Eco and other γ-proteobacteria. Crl is thought to bind to a surface-exposed region of σ^S₂ that is equivalent to the surface of σ^A₂ bound by RbpA (21).

Fig. S6.

Proposed RbpA binding sites on RNAP core are incompatible with the current structure. Protein and DNA elements are colored as in Fig. 2. (A) The proposed RbpA interacting residue R381 (red) on the RNAP β-subunit was predicted through cross-linking to bind K74 in the BL (27). However, RbpA-P77, just four residues away from K74, is 53 Å away from βR381 (direct Cα–Cα distance). (B) The RbpA-RCD binding regions predicted by Fe-BABE cleavage (green) (18) are located at least 80 Å away from the beginning of the RCD-SID linker (measured as a straight line from the Cα of P77). The clamp region, proposed in this study to be the region with which RbpA would be more likely to interact, is highlighted by the arrow.

Both RbpA and CarD are essential transcription activators in Mtb (10, 12). CarD is associated with essentially all σ^A promoters in Mycobacterium smegmatis in vivo (13), and results suggest that RbpA also likely functions at most, if not all, housekeeping σ-dependent promoters in vivo (11, 15). Therefore, it seems likely that RbpA and CarD may function simultaneously on the same RNAP–promoter complexes. Indeed, structural modeling indicates that RbpA and CarD interactions with RPo are completely compatible, each interacting with DNA upstream of the −10 element but from opposite sides of the DNA. Future studies will address the influence of these two transcription regulators on each other and their possible role in modulating global transcription patterns in the Actinobacteria.

Materials and Methods

Full details of the methods used are presented in SI Materials and Methods.

Strains, Plasmids, and Growth Conditions.

Strains, plasmids, and oligonucleotides are described in Table S1. Sco A3 (2) strains were conjugated with Eco ET12567 (pUZ8002) and cultivated on mannitol-soya agar (29). Bacterial two hybrid analysis was carried out using Eco BTH101 derivatives as described (15).

Protein Expression and Purification.

For crystallography, σ₂ (codons 224–364) and RbpA (codons 1–111) from Mtb were cloned into an His₆ pET SUMO coexpression vector, and proteins were overexpressed in Eco BL21(DE3) cells (Novagen). The complex was purified by Ni-affinity chromatography and size-exclusion chromatography. For fluorescence anisotropy, Mtb RbpA and Mtb RbpA-R79A were cloned into an His₆ pET28-based vector, expressed, and purified as described for the complex. Core RNAP, σ^A, and CarD were overexpressed and purified as described (14). For in vitro transcription and cross-linking assays, Mtb RbpA and SUMO-RbpA fusions and mutant derivatives were cloned into pET20b (Novagen) and pET SUMO (Life Technologies), respectively, overexpressed in Eco BL21(DE3)pLysS, and purified. The σ^A- and mutant derivatives were cloned into pET15b and purified from Eco BL21(pLysS).

X-Ray Structure Determination of the RbpA–σ^A₂ Complex.

The purified complex was concentrated to ∼15 mg/mL by centrifugal filtration, and crystals were grown at 22 °C by sitting drop vapor diffusion. MALDI-TOF MS was performed on the washed, dissolved crystals to determine the protein content using a Spiral TOF JMS-S3000 (JEOL). The structure was solved using an Mtb σ^A₂-homology model based on a structure of Taq σ^A₂ with the nonhomologous NCR removed [1KU2 (4)] (Table S2).

Fluorescence Anisotropy.

Fluorescence measurements were performed using an Infinite M1000pro plate reader (Tecan). Reactions were performed with Cy3-labeled duplex vapB10 DNA (Fig. S4A) (Trilink for labeled nt strand and Oligos Etc. for unlabeled template strand) extending from −36 to +1 or single-stranded Cy3-labeled DNA (Table S3) (Trilink) extending from −12 to +1. RNAP binding to duplex DNA gave poor signal-to-noise data without the addition of CarD. Recent studies have shown that Mtb RNAP makes relatively unstable open complexes and displays short half-lives on duplex promoter DNA without CarD (14). Therefore, these assays were performed in the presence of CarD. Data were analyzed using Prism software.

Formaldehyde Cross-Linking.

The cross-linking assay was modified from that in the work in ref. 24 and used as substrate vapB10 fork-junction DNA (VFJ-28), including an 18-bp double-stranded region extending from −28 to −12 followed by a single-stranded extension from −11 to −3 on the nt strand (Fig. S3A). Cross-linking was performed for 2 min in the presence of 60 mM formaldehyde.