High-Resolution Crystal Structure of RpoS Fragment including a Partial Region 1.2 and Region 2 from the Intracellular Pathogen Legionella pneumophila

Abstract

Legionella pneumophila RpoS (LpRpoS), an alternative sigma factor of RNA polymerase (RNAP), is essential for virulence and stress resistance. To investigate the mechanism of RpoS in the intracellular pathogen L. pneumophila, we determined the high-resolution crystal structure of the LpRpoS ^95–195 containing a partial region 1.2 and region 2. The structure of LpRpoS ^95–195 reveals that the conserved residues are critical for promoter melting, DNA and core RNAP binding. The differences in regulatory factor binding site between Escherichia coli RpoS and LpRpoS suggest that LpRpoS may employ a distinct mechanism to recruit alternative regulatory factors controlling transcription initiation.

1. Introduction

Bacteria are able to rapidly sense and adapt to varying environmental conditions in order to survive. In particular, pathogenic bacteria can often resist various environmental stresses during transmission and infection. Many bacteria require alternative sigma factors to initiate transcription and regulate the expression of a specific set of genes in order to adapt to changing environments. RpoS is an alternative sigma factor of RNA polymerase (RNAP) that is essential for diverse bacterial stress resistances and adaptations. In many Gram-negative pathogens, RpoS plays an important role in response to diverse stresses, including starvation, oxidative stress, UV irradiation, and acidic conditions [1].

Legionella pneumophila, the causative agent of Legionnaires’ disease, is a Gram-negative facultative intracellular pathogen capable of multiplying in a wide spectrum of eukaryotic cells. It adopts a biphasic life cycle consisting of a replicative and transmissive phase in a host cell. L. pneumophila replication is necessary to initiate lung destruction, and this process requires massive virulence factors expression [2]. In order to investigate the intracellular pathogens mechanism of improved survival in host eukaryotic cells, L. pneumophila has been exploited as a model intracellular pathogen for a variety of molecular and biochemical studies. To establish a replicative niche inside host cells and escape from host cell vacuoles, L. pneumophila has been shown to utilize RpoS to initiate the expression distinct groups of virulence factors for infection of host cells and intracellular growth [3]. In contrast to a general stress response regulator RpoS in Escherichia coli (EcRpoS), RpoS from L. pneumophila (LpRpoS) was identified as necessary not only for osmotic stress resistance but also for virulence [4]. According to the previous studies, the characterized RpoS contains four structural motifs (regions 1.2, 2, 3, and 4) connected by flexible linkers. The prokaryotic promoters recognized by RpoS contain −35 elements, extended −10 elements, −10 elements and discriminator elements [5]. Region 1.2 plays an important role in RNAP activity, promoter DNA melting, and stabilization of the open promoter complex [6]. Region 2 is the most highly conserved in RpoS family. Several key residues in RpoS region 2 are critical for binding RNAP, recognizing −10 and discriminator element, and melting dsDNA [7]. Extended −10 and −35 elements are recognized by RpoS regions 3 and 4, respectively [8]. Compared to the housekeeping sigma factors, RpoS has relatively weak affinity to the RNAP core [9]. The RpoS binding to the RNAP core can be modulated by regulatory factors. Crl as one of regulatory factors can directly bind to RpoS, and functions to increase the affinity to the RNAP core, thus promoting transcription [10].

Although LpRpoS plays a critical role in the expression of secreted virulence factors and the L. pneumophila pathogenesis, the mechanism of its regulating expression of virulence factors is still elusive. Despite the presence of different RpoS, there are no structures available from the facultative intracellular pathogens. To gain insight into the structure and function of LpRpoS, we try to determine the crystal structure of full-length LpRpoS. However, full-length LpRpoS was unstable and degraded into a more stable fragment (residues 95–195) containing a partial region 1.2 and region 2. Finally, we present the high-resolution crystal structure of LpRpoS ^95–195. This structure confirms that LpRpoS ^95–195 is more resistant to degradation than the other regions. Structural comparisons of LpRpoS ^95–195 and the corresponding region from EcRpoS reveal LpRpoS have non-conserved residues in a helix α2 adjacent to N-terminal of region 2 and an unusual 3₁₀-helix in subregion 2.3, implying that LpRpoS has different regulatory factor binding site. We speculated that L. pneumophila may employ a distinct mechanism to recruit alternative regulatory factors controlling transcription initiation. This study will lead to further explorations of molecular mechanisms used by RpoS to control gene expression in intracellular pathogens.

2. Materials and Methods

2.1. Cloning, Expression and Purification of LpRpoS

The coding sequence for LpRpoS was amplified by PCR from Legionella pneumophila genomic DNA by using sense (5′ GGAATTCCATATGATGTTAAGAAGTAAAAAACTATTTCAAGG 3′) and antisense (5′ CCGCTCGAGTCAAAACAAATCCTCTTGTGTTAGTC 3′) primers. The amplified fragments were then cloned to a modified pET28a vector (Novagen, Madison, MI, USA) with an additional 6× His coding sequence following. The recombinant LpHGPRT expression plasmid was confirmed by restriction-endonuclease digestion and further verified using DNA sequencing (Sangon Biotech, Shanghai, China). 20 µL of a glycerol stock of the transformed E. coli BL21 (DE3) (Novagen) was inoculated into 4 mL Luria Broth (LB) media with supplement of 0.1 mg mL⁻¹ kanamycin. This culture was grown overnight at 310 K and then transferred into 400 mL LB with supplement of 0.01 mg mL⁻¹ kanamycin. The culture was grown at 310 K to an A_600nm of 0.76 and induced at 290 K with 0.25 mM isopropyl β-d-thiogalactoside (IPTG) for 16 h. After harvesting, cells were resuspended in 50 mL buffer of 200 mM NaCl, 20 mM Tris-HCl, pH 8.0. After three cycles of freeze-thaw followed by 3 min sonication, the lysed cells were centrifuged at 15,000× g for 40 min. The supernatant was loaded onto a Ni²⁺ NTA (GE Healthcare, Chicago, IL, USA) equilibrated with binding buffer (200 mM NaCl, 20 mM Tris-HCl, pH 8.0) and the target protein was eluted with elution buffer (20 mM Tris-HCl pH 8.0, 200 mM NaCl, 250 mM imidazole). The eluate was subsequently loaded onto a Superdex 200 column (Amersham Biosciences, Piscataway, NJ, USA) equilibrated with 200 mM NaCl, 20 mM Tris-HCl, pH 8.0. Protein concentration was determined by the Bradford method (Bio-Rad protein assay, Hercules, CA, USA) using bovine serum albumin as a standard. Overall yield for this protein was 53.9 mg per liter of culture. The purity of the fractions was checked on the SDS-PAGE. The purified protein was concentrated to 20 mg mL⁻¹ using Amicon Ultracel 10 K cut-off concentrator (Merck Millipore, Darmstadt, Germany).

2.2. Crystallization and Data Collection

LpRpoS was crystallized at 289 K using sitting-drop vapor diffusion. The crystals of LpRpoS were grown within 3 weeks in a drop of 20 mg mL⁻¹ protein in 200 mM NaCl, 20 mM Tris-HCl, pH 8.0 with an equal volume of reservoir solution containing 0.2 M NaCl, 0.1 M Hepes pH 7.5, 12% Polyethylene glycol monomethyl ether 8000. The crystals were harvested using cryoloops and immersed briefly in a cryoprotectant solution containing the reservoir solution with 15% glycerol. The crystals were subsequently flash-frozen and stored in liquid nitrogen. The X-ray diffraction data were collected at beamline BL17U [11] at the Shanghai Synchrotron Radiation Facility (SSRF). The data set from crystal of LpRpoS ^95–195 were processed with Mosflm [12].

2.3. Structure Determination and Refinement

The structure of LpRpoS ^95–195 was solved by molecular replacement. The sigma 70 (chain Y 375–454) coordinates from E. coli sigma70 holoenzyme structure (PDB code 4yg2 [13]), which has 65% identity to the target structure, was used as the search model. A single solution obtained using Phaser [14] showed a log-likelihood gain (LLG) of 86 and Z scores for the rotation function (RFZ) of 3.9 and translation function (TFZ) of 12.5. This initial model was subjected to 20 cycles of rigid-body refinement and 10 cycles of restrained refinement using REFMAC5 [15], which resulted in R_factor and R_free values of 45.7% and 49.5%, respectively. The model was completed by iterative manual building in Coot [16] and refined with REFMAC [15] and PHENIX [17]. The final refined model contains two LpRpoS ^95–195 molecules in the asymmetric unit and was refined to an R_factor(R_free) of 13.7% (17.7%). The final crystallographic model was evaluated using MolProbity [18]. The final coordinates and structure factors were deposited in the Protein Data Bank (http://www.rcsb.org/pdb) under the accession code 5H6X for LpRpoS ^95–195. The refinement statistics are summarized in

Table 1. Data collection and refinement statistics.

	LpRpoS 95–195
SSRF beamline	BL17U
Wavelength (Å)	0.97915
Space group	P212121
Molecules/ASU	2
Cell parameters
a/b/c (Å)	37.93/68.84/70.38
á/â/ã (°)	90.00/90.00/90.00
Resolution range (Å)	33.4–1.45 (1.53–1.45)
No. of unique reflections	30,122 (3739)
Corresponding % solvent	39.8
Rpim a (%)	3.8 (24.4)
CC1/2	0.99 (0.86)
Average I/ó(I)	12 (3.1)
Redundancy	13.8 (14)
Completeness (%)	90.7 (78.5)
Refinement statistics
Resolution range (Å)	20–1.45 (1.50–1.45)
R-factor b (%)	13.7 (20.8)
R-free c (%)	17.7 (22.9)
RMSD d bond length (Å)	0.007
RMSD bond angles (°)	1.08
No. of non-H atoms
Protein	1616
Solvent	192
Average of B factors (Å2)
Protein	21.5
Solvent	36.0
Ramachandran plot f (%)
Ramachandran favored	99
Ramachandran Outliers	0
PDB ID code	5H6X

Values in parentheses are for the highest resolution shell. ^a R_p.i.m = Σ_hkl[1/(n_hkl – 1)]^1/2Σ_i·|I_i(hkl) – 〈I(hkl)〉|/Σ_hkl·Σ_iI_i(hkl), where n_hkl is the number of observations of reflection hkl. ^b R-factor = Σh||F_obs| − |F_calc||/Σ|F_obs|, where |F_obs| and |F_calc| are the observed and calculated structure factor amplitudes, respectively. Summation includes all reflections used in the refinement. ^c Free R factor = Σ||F_obs| − |F_calc||/Σ|F_obs|, evaluated for a randomly chosen subset of 5% of the diffraction data not included in the refinement. ^d Root-mean square-deviation from ideal values. ^f Categories were defined by Molprobity.

. Amino acid sequences were aligned by Multalin [19], and the figure of structure-based sequence alignment was generated using ESPript [20]. All illustrations were prepared with PyMOL (http://www.pymol.org).

3. Results

3.1. Crystallization of LpRpoS ^95–195

We have attempted to crystallize full length LpRpoS. However, unit cell parameters of crystals were in obvious discrepancy with the molecular weight of the expressed protein. SDS-PAGE analysis of a dissolved crystal revealed a 10 kDa fragment, much smaller than the predicted size of the cloned protein (41 kDa). N-terminal sequencing identified the residues EIGFS. Coupled with mass spectrometry analysis, we determined that the crystallized fragment contains a partial region 1.2 and a region 2 of LpRpoS ^95–195 (Figure 1a).

3.2. Overall Structure of LpRpoS ^95–195

The crystal of LpRpoS ^95–195 diffracts to 1.45 Å resolution and belongs to the P2₁2₁2₁ space group, with two molecules in an asymmetric unit. Analysis using PISA [21] suggested that a monomer is likely to be the biologically relevant form of the protein. LpRpoS ^95–195 adopts a helix-turn-helix fold consisting of five helices (α1, α2, α3, α4 and α5) (Figure 1b). Structure homology search by the DALI [22] server reveals that the LpRpoS ^95–195 structure is very similar to the corresponding regions from the E. coli RpoS (EcRpoS) structure (PDB code 5ipl [23]). LpRpoS ^95–195 shares 61% sequence identity and 76% sequence similarity to the corresponding regions of EcRpoS. The superimposition of LpRpoS ^95–195 monomer structure and its homologs showed that the architecture of LpRpoS ^95–195 is remarkably conserved (Figure 1c). Core RNAP binding, recognition and binding of the −10 promoter element, and promoter melting are all controlled by the well-conserved residues from RpoS region 2 [7]. RpoS region 2 is divided into four parts, subregions 2.1, 2.2, 2.3 and 2.4.

Structure-based alignment of LpRpoS ^95–195 and the corresponding region from EcRpoS revealed that the conserved residues appear to be critical for promoter melting, DNA and core RNAP binding (Figure 2a). The LpRpoS ^95–195 is similar to the structure of the corresponding region from EcRpoS complexed with nucleic acids (Figure 2b). The Arg129 and Lys133 from subregions 2.2 and 2.3 of EcRpoS related to DNA binding are also conserved in LpRpoS ^95–195 (corresponding residues Lys160 and Lys164) (Figure 2c) [24]. Subregion 2.3 from LpRpoS ^95–195 has highly conserved aromatic residues (Phe171, Tyr176, Trp179, Trp180) implicated in promoter melting (Figure 2d) [24]. In EcRpoS, conserved residues Gln152, Glu155 and Arg156 from subregion 2.4 are involved in recognizing the −10 promoter element (Figure 2e) [24]. Meanwhile, these residues are also found in corresponding sites of LpRpoS ^95–195. Similar to EcRpoS structure, LpRpoS ^95–195 contains the conserved residues from subregions 2.2 involving in core RNAP binding (Figure 2a). These structure comparison results indicated that LpRpoS ^95–195 likely have a similar way to melt promoter, bind core RNAP, recognize and bind the −10 promoter element. Although EcRpoS and LpRpoS ^95–195 have the high similarity in recognition and binding of the −10 promoter element, promoter melting sites and core RNAP binding site, there are also some a significant distinction in biological function between EcRpoS and LpRpoS. Unlike EcRpoS, which is required for a general stress resistance, LpRpoS is only involved in osmotic shock but in no other stress resistance, and is also more important for the production of virulence factors [4].

3.3. LpRpoS Show a Different Regulatory Factor Binding Site from EcRpoS

Sigma factors compete for binding to a limited amount of the core RNAP. In contrast to housekeeping E. coli sigma factor 70 with strong affinity for the RNAP core enzyme, EcRpoS binds relatively weakly to the core RNAP in the absence of nucleic acids [9]. It’s predicted that E. coli sigma factor 70 with an additional non-conserved region (NCR) probably increase the affinity to interact with the core RNAP compare to EcRpoS. Many housekeeping sigma factors contain a NCR located between subregions 1.2 and 2.1. NCR is not present in alternative sigma factors like RpoS. In response to a variety of environmental conditions, different sigma factors binding the core RNAP can be modulated by several regulatory factors [25]. The Crl, as the only known regulatory factor, can directly interact with region 2 of RpoS and enhance its affinity to the RNAP core for promoting transcription [26]. The Crl located the similar position as NCR in sigma factor 70, and possibly functions similarly to NCR to provide extra interactions with core RNAP. Despite of non-conserved and less widespread in bacterial species, Crl homologs share a similar mechanism and function. Previously studies showed that RpoS interacting with Crl involves two regions including α-helix adjacent to N-terminal of RpoS region 2 and loop in RpoS subregion 2.3. It is noteworthy that the two regions from RpoS are exposed to the outside of the RNAP main channel for Crl recognition. Meanwhile, it also found that the key residues of Asp87, Asp135, Pro136, and Glu137 from EcRpoS are important for EcRpoS interacting with Crl [26]. Asp87 locate on α-helix near the N-terminal of EcRpoS region 2. Asp135, Pro136, and Glu137 from EcRpoS region 2 with the DPE motif. In particular, the charge of Glu137 in the DPE motif is crucial for Crl recognition [27].

LpRpoS ^95–195 and the corresponding region of EcRpoS have very similar structure and highly conserved the core RNAP binding site. It deduced that LpRpoS may have low affinity to core RNAP, and probably require regulatory factor to bind core RNAP. Unlike E. coli that utilizes Crl regulating EcRpoS, L. pneumophila contains conserved RpoS but does not have Crl. Comparison of LpRpoS ^95–195 and EcRpoS structures reveals significant differences between Crl-binding site of EcRpoS and the corresponding region from LpRpoS. Although the conserved residue Asp118 found in helix α2 near the N-terminal of LpRpoS ^95–195 region 2, LpRpoS ^95–195 does not have the conserved DPE motif. Structural comparison between LpRpoS ^95–195 and EcRpoS show that Lys168 of LpRpoS ^95–195 replaces the corresponding residue Glu137 from EcRpoS (Figure 3). EcRpoS Glu137Gln mutant did directly affect Crl binding, suggesting that the negative charge of Glu137 is importance for Crl binding. The Lys168 of LpRpoS ^95–195 substituted Glu137 from EcRpoS reveals that LpRpoS ^95–195 might utilize a different mechanism to regulate factor recognition. Meanwhile, the residues Asp166, Pro167 and Lys168 of LpRpoS located on a region between α4 and α5 adopt an unusual 3₁₀-helix architecture. It suggests that the special 3₁₀-helix fold and the oppositely charged residue may together compose a distinct regulatory factor binding site. The region 1.2 adjacent to region 2 was proposed as a recognition site of the regulatory factor Crl. Salmonella enterica serovar Typhimurium RpoS R82L mutant is defective in Crl binding and activation [28]. The fact that LpRpoS ^95–195 uses the corresponding Leu113 as a substitute for Arg82 in StRpoS indicates that LpRpoS ^95–195 may not be suitable for Crl binding. It also found that the helix from region 1.2 adjacent to the N-terminal of LpRpoS region 2 contains non-conserved residues compared with the corresponding region in EcRpoS. Altogether, these findings indicate that LpRpoS may enhance its affinity to core RNAP through binding alternative regulatory factor.

4. Discussion

RpoS play a role in stress response gene expression and production of virulence factors in L. pneumophila. Our analysis showed that LpRpoS displays highly conserved residues and has folds that are similar to the corresponding regions in E. coli RpoS. Most interestingly, Sequence and conformation of the region 2 in EcRpoS, which is involved in regulatory factor Crl binding, is different from that observed in LpRpoS ^95–195. It found that the Crl binding site at loop of EcRpoS (Asp87, Asp135, Pro136, and Glu137) is different from the corresponding site of LpRpoS with the residues Asp166, Pro167 and Lys168, which located on a region between α4 and α5 adopt an unusual 3₁₀-helix architecture. Importantly, we determined that the function of Glu137 in E. coli is substituted by Lys168 in LpRpoS. Meanwhile, there is no Crl homologs in L. pneumophila. These observations suggest that L. pneumophila may apply a distinct mechanism to employ alternative regulatory factors controlling LpRpoS transcription initiation. This mechanism is probably important for L. pneumophila survival and virulence. The structures present here further shed light into understanding the mechanism of the intracellular pathogen pathogenesis.