The PilZ domain of MrkH represents a novel DNA binding motif

MrkH is the first characterized c-di-GMP related transcriptional regulator which affects type 3 fimbrial expression in response to cellular c-di-GMP level, and thus plays an important role in the biofilm formation of Klebsiella pneumoniae (Murphy and Clegg, 2012; Wilksch et al., 2011). However, how MrkH recognizes c-di-GMP and the target DNA sequence remains obscure. Here, we determine the crystal structure of MrkH/c-diGMP complex at 2.3 Å resolution. MrkH adopts a tandem two-domain structure with a canonical YcgR-N/PilZ proteins-fold, it contains two α-helixes (α1, α2) and 18 β-sheets which formed β-barrel 1 and β-barrel 2 (Fig. 1A and 1B). Our size-exclusion chromatography results demonstrate that MrkH forms a stable monomer either in the presence or absence of c-di-GMP (Fig. 1C). A DALI search for globally similar proteins revealed that MrkH/c-di-GMP has notable structural homology with PilZ domain proteins (Benach et al., 2007; Holm and Rosenstrom, 2010; Ko et al., 2010). The binding of c-di-GMP to MrkH is very similar to Pp4397 both MrkH and Pp4397 hold tightly two mutually intercalated c-di-GMP molecules, while VCA0042 combines with only one c-di-GMP (Fig. 1D). Through structure analysis we find that MrkH mainly forms H-bond with two c-di-GMP molecules (Fig. 1D and 1E). The side chains of R107 and R111 also contribute a cation–π interaction with the guanine group of C2E2 and C2E1 respectively. Multiple-sequence alignment revealed these residues are well conserved (Fig. S1). Besides the interactions between protein and ligands, two c-di-GMP molecules are also stabilized by strong base stacking interaction between mutually intercalated guanine groups. We further analyze the interactions between MrkH and c-di-GMP using ITC (Fig. S2A–G). The affinity of MrkH for c-di-GMP is high with Kd of approximately 0.24 μmol/L using the one site specific binding model(Whitney et al., 2015). MrkH also binds to c-di-GMP efficiently. However, MrkH loses the c-di-GMP binding affinity, which indicates that the connecting loop between two β-barrels is crucial for c-diGMP binding. R107 forms both H-bond and cation–π interaction with two c-di-GMPs (Fig. 1D), thus R107A mutant has greatly decreased binding affinity for c-di-GMP with a dissociation constant fifteen times higher than that of the wild-type MrkH (3.55 μmol/L vs 0.24 μmol/L). R111A mutant almost completely loses the c-di-GMP binding affinity, indicating that R111 is the most important residue for c-di-GMP binding. MrkH variants bind to c-di-GMP with an approximately stoichiometry of 1:2 (Table S1). Previous studies have demonstrated that MrkH could bind directly to the promoter of mrkHI or mrkA, and c-diGMP molecule promotes the binding of MrkH to the promoter (Tan et al., 2015; Wilksch et al., 2011; Yang et al., 2013), but it is unclear how MrkH recognizes its targets DNA. To address this problem, we incubated MrkH with various DNA fragments for EMSA. We purified recombinant MBP-MrkH, MBP-MrkH-YcgR-N domain (residues 1-104) and MBP-MrkH-PilZ domain (residues 105-end) which were incubated with the unlabeled mrkHI and mrkA promoter fragments respectively. The evident DNA-protein complex migrations were observed in lanes of MBP-MrkH and MBPMrkH-PilZ domain (Fig. 2A). It suggests that MrkH binds directly to the mrkHI and mrkA promoter sequence mainly through its PilZ domain. The EMSA also shows the migration velocity of DNA fragments constantly slow down as protein concentration increases (Fig. 2B) and MrkH-PilZ without MBP tag gives the same result (Fig. 2C). Size-exclusion chromatography analysis of PilZ domain indicates that the oligomeric state of PilZ domain is not affected by protein concentration (Fig. S3A). These observations implied that a long DNA fragment may recruit multiple PilZ domains and resulting nonspecific binding. In order to locate the DNA binding region in MrkH-PilZ domain, the vacuum electrostatics of MrkH/c-di-GMP structure is carefully analyzed. A highly positive charged hump consisting of six basic residues (R125, K127, K154, K163, K207 and R209) is identified (Fig. 1F). Since DNA is a negatively charged, this positively charged region may be critical for DNA binding. To test this hypothesis, we constructed several cognate mutants of MrkH-PilZ domain and performed EMSA with the mrkA regulatory fragment. The EMSA result shows that all these mutants lose DNA binding abilities

Unbound materials were washed out with one column volume of wash buffer B (25 mM Tris-HCl, pH 8.0, 500 mM NaCl, 10 mM imidazole). The target proteins were eluted with 10 ml buffer C (25 mM Tris-HCl pH 8.0, 500 mM NaCl, 250 mM imidazole). The elutes were then concentrated and loaded onto a Superdex200 size-exclusion column equilibrated with 10 mM Tris-HCl pH 8.0, 500 mM NaCl. The final concentration of MrkH 1-228 -pET-28a was 6 mg•ml -1 and stored at -80 °C for crystallization use. For the MBP fusion MrkH and mutants, the salt concentration of all buffers was 100 mM instead.

Crystallization, data collection and structure determination
MrkH was incubated with c-di-GMP with a molar ratio of 1:2.5 and crystallization screen was performed with kits from Hampton Research using the sitting-drop vapour diffusion method at 4°C. The crystal of MrkH/c-di-GMP complex was observed in a solution containing 0.5 M Ammonium sulfate, 0.1 M Sodium citrate tribasic dihydrate and 1.0 M Lithium sulfate monohydrate after 14 d at 4°C. To solve phase problem, the heavy-atom reagent I3C (Hampton Research) was soaked into the native crystals by addition of 0.5 µl 0.5 M I3C solution directly to the crystallization drop. After 15 min, crystals with pale yellow color were chosen for the next step. All Crystals were flash frozen in liquid nitrogen, with addition of 25% (v/v) ethylene glycol serving as a cryoprotectant.
X-ray diffraction data were collected on beamline BL17U1 at the Shanghai Synchrotron Radiation Facility (SSRF). All data were processed using HKL-2000 (Leslie et al., 2002). The heavy atom soaked crystal was determined by the single-wavelength anomalous dispersion (SAD) phasing. Initial phases were calculated using AutoSol implemented in PHENIX (Adams et al., 2010), and seventeen I3C sites were identified.
AutoBuild in PHENIX was used to automatically trace the chain of MrkH, and the molecular replacement was performed to solve the native structure using I3C structure as the search model by running MOLREP in CCP4 (Murshudov et al., 2011). After several rounds of positional refinement alternated with manual model revision using Coot (Emsley et al., 2010) and the refinement programs REFMAC (Murshudov et al., 2011) and Phenix.Refine (Adams et al., 2010), the quality of final model was evaluated using the PROCHECK program (Laskowski et al., 1993). Details of the data-collection and refinement statistics are shown in Table S2. All of the structure figures were rendered by PyMOL ( www.pymol.org.) (DeLano, 2002).

Electrophoretic mobility shift assay (EMSA)
DNA-binding activities of different MrkH fragments and mutants were analyzed in electrophoretic mobility shift assay (EMSA). The MBP fusion proteins were used for EMSA. Each unlabeled DNA fragment was incubated with varying amounts of purified MBP-MrkH and MrkH-PilZ domain in 50 mM Tris pH 8.0, 100 mM NaCl, 5 mM MgCl2 in a total volume of 20 μl for 100 min on ice. Proteins were added at least at a 103:1 molar ratio to DNA. Reaction samples were then mixed with the loading buffer and separated in 5% native polyacrylamide gels (37.5:1) in 0.5×TBE buffer for 120 min at 80 V and 4°C. Protein-DNA complexes were stained with Ethidium bromide and visualized by UV imaging system. The DNA fragments containing regulatory region and mrkHI regulatory region were generated by PCR using synthetic plasmid containing those regulatory sequences as templates. The used primers were listed in Table S3.

Gel-filtration assay
The MrkH and MrkH/c-di-GMP complex were subjected to gel-filtration analysis (Superdex 200 10/300 GL column; GE Healthcare; 10 mM Tris-HCl pH 8.0; 500 mM NaCl) with or without 50 μM c-di-GMP, the assay was performed at a flow rate of 0.4 ml•min-1 and 0.1 ml of MrkH (about 2.0 mg•ml-1) was injected at 4 °C. The elution volume of the MrkH/c-di-GMP was 15.71 ml under the conditions of the assay, while the elution volume of MrkH monomer was 15.98 ml. MrkH-PilZ proteins were also analyzed with the above method (except for NaCl concentration of 100 mM).

Supplementary Figures
Supplementary Figure S1. Multiple sequence alignment of MrkH and homologues.