Crystal structure of the multiple antibiotic resistance regulator MarR from Clostridium difficile

The crystal structure of MarR from C. difficile is reported.


Introduction
Clostridium difficile is an anaerobic human pathogen that causes acute healthcare-associated diarrhoea. The morbidity and mortality rates of C. difficile infection (CDI) have increased dramatically in Europe and in North America (Heinlen & Ballard, 2010); the emergence of C. difficile strains that are resistant to multiple antibiotic agents can complicate prevention programs and potential treatments (Hunt & Ballard, 2013). Many studies have demonstrated that various bacterial species employ MarR homologues to sense and exert resistance against many cellular toxins from the environment or host immune system, including multiple antibiotics, oxidative reagents and disinfectants (Cohen et al., 1993;Alekshun & Levy, 1999). The regulator of multiple antibiotic resistance (MarR) in Escherichia coli, a member of the MarR family of regulator proteins, modulates bacterial detoxification in response to diverse antibiotics (Hao et al., 2014). The transcription factors of the MarR family regulate diverse genes involved in multiple antibiotic resistance, the synthesis of virulence determinants and many other important biological processes Alekshun & Levy, 1997;Perera & Grove, 2010). The MarR protein, as a member of the MarR family of multiple antibiotic resistance proteins, is a key global regulator in C. difficile. A link between MarR family proteins and antibiotic resistance has been suggested in previous studies (George & Levy, 1983). However, the function of MarR in C. difficile is still unknown. Here, we aim to study the biological function of MarR from the perspective of its crystal structure. Thus, the major work in this article is to report the crystal structure of MarR from C. difficile (MarR C.difficile ).
In this study, we solved the crystal structure of MarR C.difficile by molecular replacement. Diffraction data were collected to 2.3 Å resolution. The overall structure indicated that MarR C.difficile is a homodimer, with each subunit consisting of six helical regions and three -strands. Like other MarR proteins, the helical regions in each subunit contribute to the protein-protein interface in the dimer. An analysis of electrostatic surface potential shows a putative DNA-binding site, as observed in other MarR family proteins. This is the first reported crystal structure of this protein from C. difficile.

Protein preparation
The gene encoding MarR (annotated in GenBank as CAJ67669.1) was PCR-amplified using C. difficile 630 genomic DNA as template, into which NdeI and EcoRI restriction sites were introduced. The purified PCR product was digested with the corresponding restriction enzymes and ligated with T4 DNA ligase into the pET-28a(+) vector (Novagen). The resulting construct contained a hexahistidine tag at the N-terminus of MarR and a thrombin cleavage site. The constructed plasmid was transformed into E. coli BL21 cells for expression. The E. coli cells were grown in LB medium containing 100 mg ml À1 kanamycin at 37 C to an OD 600 of 0.6 before IPTG was added to a final concentration of 0.4 mM. Protein expression was induced at 20 C for 12 h before harvesting. The cell pellets were collected and resuspended in NTA buffer (20 mM Tris-HCl pH 7.6, 200 mM NaCl, 10% glycerol) containing 1 mM phenylmethanesulfonyl fluoride (PMSF). After sonication on ice for 30 min, the cell lysate was spun at 11 000 rev min À1 for 30 min. The clear lysate was filtrated and loaded onto a HisTrap column (5 ml column, GE Healthcare), which had been pre-equilibrated with 50 ml NTA buffer (Hao et al., 2014), for nickel-affinity chromatography. The MarR C.difficile protein was then eluted with a linear imidazole gradient followed by a further purification step using a HiLoad 16/60 Superdex 200 column (GE Healthcare) equilibrated with buffer A (20 mM Tris-HCl pH 7.6, 200 mM NaCl). The eluted protein was purified to >95% homogeneity as determined by 16% SDS-PAGE analysis (Fig. 1). The protein was collected and concentrated for crystallization screening.
Information relating to the production of recombinant MarR is summarized in Table 1.

Crystallization
Crystals of MarR C.difficile were grown at 16 C by hangingdrop vapour diffusion. 2 ml purified protein (5 mg ml À1 ) in 200 mM NaCl, 20 mM Tris pH 7.6 was mixed with 2 ml reservoir buffer [10%(v/v) 2-propanol, 100 mM Tris pH 7.6]. The droplets were equilibrated against 400 ml reservoir buffer. Crystals were looped-out and soaked in cryoprotectant [crystallization buffer containing 20%(v/v) glycerol] before flash-cooling and storage in liquid nitrogen. Crystallization information is summarized in Table 2.

Data collection and processing
X-ray diffraction data were collected to 2.3 Å resolution on beamline BL17U at the Shanghai Synchrotron Radiation Facility (SSRF). Crystals were flash-cooled in mother liquor at the beamline before data collection. All data were processed Elution profile of MarR C.difficile from a Superdex 200 HiLoad 16/60 gel-filtration column on an Ä KTAexplorer FPLC system. The peak at 82 ml (the flow rate was 1 ml min À1 ) represents the MarR C.difficile dimer. The apparent molecular weight of the eluting species was calculated using standard protein markers (Gel Filtration LMW Calibration Kit, GE Healthcare). The right panel shows SDS-PAGE analysis of the collected fraction. and reduced using HKL-2000 (Otwinowski & Minor, 1997). The space group of the MarR crystals was determined to be P4 3 2 1 2, with one molecule in the asymmetric unit and with unit-cell parameters a = b = 66.57, c = 83.65 Å , = = = 90 for the native protein. Data-collection and processing statistics are summarized in Table 3.

Structure solution and refinement
The structure was solved by molecular replacement using SlyA from Listeria monocytogenes (PDB entry 4mnu; Midwest Center for Structural Genomics, unpublished work) as the starting model. The structure was refined using REFMAC 5.7.0032 Potterton et al., 2003;Murshudov et al., 2011). Finally, the structure was deposited in the Protein Data Bank as PDB entry 5eri. The structuresolution and refinement statistics are summarized in Table 4 .

Results and discussion
3.1. Overall structure of MarR C.difficile The crystal structure of MarR C.difficile was determined by molecular replacement using SlyA (PDB entry 4mnu) as a search model. Like other MarR proteins, the MarR C.difficile protein is composed of six -helices and a three-stranded antiparallel -hairpin (Fig. 2a). The 2, 3 and 4 helices and two antiparallel -strands, 2 and 3, are probably responsible for DNA binding, as indicated by the electrostatic surface potential (

Figure 2
The structure of MarR C.difficile . (a) One MarR C.difficile subunit with labelled secondary structure. (b) Ribbon representation of the crystal structure of the MarR C.difficile dimer viewed with the subunit twofold axis close to vertical. MarR C.difficile protein are strongly electropositive, as are other winged-helix DNA-binding proteins (Gajiwala & Burley, 2000).
The structure of MarR from E. coli (MarR E.coli ) is a homodimer (Alekshun et al., 2001), and it has been verified that MarR E.coli binds the marRAB promoter as a dimer (Martin et al., 1996). A recent study showed that disulfide bonds could be formed between MarR E.coli dimers, resulting in the dissociation of MarR E.coli from its cognate DNA and enhanced bacterial resistance (Zhu et al., 2017). The MarR family member MprA also functions as a dimer (Brooun et al., 1999). PISA analysis suggested that MarR C.difficile is a homodimer. In the crystal structure of MarR C.difficile there is one monomer in the asymmetric unit, with the dimer being composed of two subunits related by a crystallographic twofold rotation. The crystal structure also indicates that -helices in the N-and C-terminal regions of each subunit interdigitate with those of the other subunit to form a hydrophobic core burying a surface area of 3100 Å 2 . Two helical regions, 1 and 6 (residues 9-27 in the N-terminus and residues 123-147 in the C-terminus, respectively), are closely juxtaposed and intertwine with the equivalent regions of the second subunit to form a dimer. The dimer is stabilized by several salt bridges, notably that between Arg16 and Glu74 0 and that between Lys155 and Glu145 0 . In addition, a hydrogen bond between the side-chain carbonyl O atom of Glu66 and the guanidinium NH groups of Arg31 0 enhances the stability of the dimeric structure.
3.2. Comparison of MarR C.difficile with MarR E.coli and MgrA S.aureus MgrA from Staphylococcus aureus (MgrA S.aureus ) is a regulator of antibiotic resistance and is also an important virulence determinant during infection (Ingavale et al., 2005). A previous study indicated that the cysteine residue (Cys12) of this protein could be oxidized by various reactive oxygen species. Cysteine oxidation leads to the dissociation of MgrA from DNA, resulting in the initiation of signalling pathways and further enhancing antibiotic resistance in S. aureus (Chen et al., 2006). The oxidation-sensing mechanism is widely used by bacteria to counter challenges of environmental pressure (Lee & Helmann, 2006). In conclusion, MgrA from S. aureus is an oxidation sensor.
Previous studies reported that the MarR family of proteins are typically conserved transcription factors that modulate bacterial resistance to multiple antibiotics, oxidative reagents and detergents . Various bacterial species such as E. coli can respond to environmental stresses such as toxic chemicals and disinfectants by triggering the dissociation of MarR from the cognate DNA in a copperdependent manner. The detailed mechanism is that copper(II) oxidizes a unique cysteine residue (Cys80) that resides in the DNA-binding domain of MarR E.coli to generate inter-dimer disulfide bonds, thereby inducing tetramer formation and the dissociation of MarR from the marRAB promoter (Hao et al., 2014). Therefore, MarR from E. coli is a copper signal oxidation sensor.
Sequence alignment of MarR C.difficile with MarR E.coli and MgrA S.aureus using MUSCLE shows 26 and 19% sequence identity; these proteins share low sequence similarity (Fig. 4). However, superposition of MarR C.difficile with MgrA S.aureus (PDB entry 2bv6; Chen et al., 2006) and MarR E.coli (PDB entry 1jgs; Alekshun et al., 2001) shows structural similarity (Fig. 5). Electrostatic surface representation of the MarR C.difficile dimer. The putative DNA-binding sites are indicated by an arrow. The MgrA S.aureus dimer is triangular in shape, with two wingedhelix DNA-binding domains; the DNA-binding domain includes two -helices, two -sheets and a wing region (Chen et al., 2006). The structure of MarR E.coli is a crystallographic dimer, with each subunit containing a winged-helix DNAbinding motif, and this DNA-binding motif contains three -helices and two -strands (Alekshun et al., 2001). We found that the crystal structure reveals MarR C.difficile to be a dimer, with each monomer consisting of six -helices and a threestranded antiparallel -hairpin. The putative DNA-binding domain of each subunit includes three -helices and two antiparallel -strands. Correspondingly, MarR E.coli and MgrA S.aureus share a similar oxidation-sensing mechanism in which cysteine oxidation leads to the dissociation of MarR E.coli and MgrA S.aureus from DNA. As a result, they exhibit a similar function. However, the function of MarR in C. difficile remains unknown. Structural analysis of MarR C.difficile indicated that two cysteine residues (Cys45 and Cys117) are located in the hydrophobic core; this may suggest that MarR in C. difficile is probably not an oxidation sensor. Although they share structural similarity, these proteins might have diverse molecular mechanisms.  Primary-sequence alignment of MarR C.difficile with representative members of the MarR family (MarR E.coli and MgrA S.aureus ). The secondary-structural elements of MarR C.difficile are indicated above the sequence alignment: -helices () are illustrated as curly lines and arrows represent -sheets (). Numbering is according to the MarR C.difficile primary sequence. Residues that are identical in all homologues are highlighted in red and highly conserved amino acids are shown in blue boxes. Red arrows indicate the cysteine residues of MarR C.difficile ; the key cysteine residues of MarR E.coli and MgrA S.aureus are shown in purple boxes.

Conclusion
Although the MarR protein has been well studied in many species, the MarR protein from C. difficile remains unknown. The relationship between MarR and antibiotic resistance in C. difficile needs to be investigated. The crystal structure reported in this paper reveals MarR C.difficile to be a crystallographic dimer. It shows structural similarity to other MarR family proteins. Furthermore, the structure of MarR in C. difficile suggests a putative DNA-binding site, revealing that the MarR protein in C. difficile might be also a transcription factor that can bind DNA. Based on the structural analysis of MarR C.difficile , we found that the two cysteine residues could not be oxidized easily as they are located in the hydrophobic core. Therefore, MarR from C. difficile may not be an oxidation sensor. The solved crystal structure of MarR C.difficile will be the first step in further functional studies.