Functional and structural analysis of a catabolite control protein C that responds to citrate

Wei Liu Dalian Minzu University Jinli Chen Dalian University of Technology Liming Jin Dalian Minzu University Zi-Yong Liu Qingdao Institute of Bioenergy and Bioprocess Technology Ming Lu Qingdao Institute of Bioenergy and Bioprocess Technology Ge Jiang Dalian University Qing Yang Dalian University of Technology Chunshan Quan Dalian Minzu University Ki Hyun Nam Pohang University of Science and Technology Yongbin Xu (  yongbinxu@dlnu.edu.cn ) Dalian Minzu University


Introduction
The tricarboxylic acid (TCA) cycle, also known as the Krebs cycle or the citric acid cycle (CAC), is a central metabolic pathway in the cell 1 . The TCA cycle provides organisms with reducing potential, energy, and three of the 13 biosynthetic intermediates 2 . In Bacillus, TCA activity is controlled by several important regulatory proteins, including global regulators CcpA and CodY and the speci c regulator CcpC, which are coordinated by fructose-1,6-bisphosphate (FBP) and glucose-6-phosphate, guanosine-5'-triphosphate (GTP) and branched-chain amino acids (BCAAs), and citrate, respectively 3 . CcpA and CodY are metabolite-responsive global regulators of carbon metabolism pathways 4 . These global regulators coordinate the expression of numerous metabolic, biosynthetic and virulence genes that respond to three metabolites 5 . CcpA is a member of the LacI/GalR family of transcriptional repressors, which exert both direct (through citrate synthase, citZ, and ccpC) and indirect effects on TCA branch enzyme expression 6 . CodY is also a repressor of the citB gene belonging to a unique family of regulators in B. subtilis and other homologues of gram-positive bacteria 7 . Recent studies showed that CcpC and CcpE exclusively regulate the TCA branch enzymes of the TCA cycle (citB, aconitase; citC, isocitrate dehydrogenase; and citZ, citrate synthase) by responding to a pathway-speci c metabolite for both Bacillus subtilis and Staphylococcus aureus, respectively 8 9 .
CcpC widely exists in prokaryotes and is classi ed in the LysR-type transcriptional regulator (LTTR) family 10 . Typical LTTR family proteins comprise approximately 330 amino acids that form structures highly similar to those of N-terminal DNA-binding domains (DBDs), which are directly involved in DNA interactions, and poorly conserved C-terminal inducer-binding domains (IBDs) and are known to adopt different oligomeric states 10 . The DBD is highly conserved and directly involved in DNA interactions, similar to the helix-loop-helix, zinc nger, and β-sheet-anti-parallel domains 11 12 13 . The role of IBD transcription factors in the regulation of bacterial virulence has been investigated in many pathogenic organisms. The coinducer citrate is important for the function of CcpC and appears to function as a key catabolite for coordinating the B. subtilis metabolic state by binding to and activating CcpC 9 . Thus, the CcpC complex with citrate is a signal that morphs CcpC into a conformation that is competent for binding DNA and gene transcription. Several biological functional properties of CcpC are well characterized; however, the structure-based molecular function has been elusive 9 .
In this study, we characterized the full-length Bacillus amyloliquefaciens CcpC and determined the crystal structure of the C-terminal IBD of B. amyloliquefaciens CcpC (BaCcpC-IBD) at 2.3 Å resolution. The citB binding properties and the oligomeric state of BaCcpC were analysed. The crystal structure of BaCcpC-IBD was compared with structures of the LTTR family members. Taken together, our ndings provide insight into the citrate-responsive mechanism of CcpC.

Results And Discussion
Biochemical study of BaCcpC In B. subtilis, CcpC (BsCcpC) negatively regulates citB gene expression, which is responsible for the interconversion of citrate and isocitrate 9 . The BsCcpC binding region forms two dyad symmetry elements centred at positions − 66 and − 27 (Fig. 1A) 9 . These BsCcpC bind to the DNA-binding boxes "ATAA", "TTAT", and "TATT" in the citB promoter region 9 . In the B. amyloliquefaciens genome, the potential promoter region of citB (named citB-P) was found from position − 73 to -20, and it shows high similarity with the same DNA-binding boxes "ATAA", "TTAT", and "TATT" (Fig. 1A). The DNA sequence of the BsCcpC binding box was identical to that of the BaCcpC binding box in citB promoter region I (named citB-PI, -73 to -54), but the nonbinding sequences did not match, whereas the DNA sequences of citB promoter region II (named citB-PII, -40 to -20) matched a consensus sequence (Fig. 1A). Meanwhile, the spacer sequence between citB protomer regions I and II was identical to those of B. subtilis and B. amyloliquefaciens, as was the DNA length.
To verify whether CcpC regulates the predicted citB promoter region in B. amyloliquefaciens, we performed an electrophoretic mobility shift assay (EMSA) using a 56 bp citB promoter DNA fragment (from nucleotides − 73 to -20 relative to the start codon of citB) as a probe (Fig. 1A). A complete shift of the free probe was observed as the concentration of BaCcpC increased (Fig. 1B). This result indicates that BaCcpC has a high binding a nity for citB promoter DNA of B. amyloliquefaciens. Next, the binding of BaCcpC to each of the citB promoter regions (I and II) for each CcpC was assessed. BaCcpC did not bind to either promoter region ( Fig. 1C-D). These results indicate that both dyad symmetry promoter regions are required for BaCcpC to bind to the promoter of citB. To determine whether citrate affects the binding of CcpC to the citB promoter, an EMSA of CcpC for citB was performed with citrate. The results showed that the presence of citrate slightly suppressed the binding of CcpC to citB-P (Fig. 1E).

Oligomerization of BaCcpC
LTTRs are usually functionally active as tetramers and dependent upon a coinducer 14 . To verify the oligomeric state of BaCcpC (MW ~ 30 kDa) in solution, we performed size exclusion chromatography. The chromatogram showed multiple peaks ( Fig. 2A). The calculated molecular weights of BaCcpC for the rst and second peaks were > 500 kDa and 30 kDa, respectively. This result indicated that BaCcpC exists as a large oligomer (or aggregated form) or a monomer in solution. Meanwhile, in equilibrium buffer with 10 mM citrate, the abundance of the large oligomer, i.e., the rst peak intensity, was reduced, whereas the monomer population, i.e., the second peak, was increased ( Fig. 2A). These results indicated that citrate induced the oligomers of BaCcpC proteins to form monomers but not completely. In the crystal structure of BaCcpC-IBD, citrate molecules interact directly with two serine residues (Ser189 and Ser191) by forming hydrogen bonds (see below). To understand whether these two serine residues in uence the oligomeric state of BaCcpC mutants (Ser189Ala and Ser191Ala, named BaCcpC-S189A and BaCcpC-S191A), we performed size exclusion chromatography. Interestingly, when we assessed the oligomeric states of BaCcpC-S189A and BaCcpC-S191A, only one peak appeared (MW ~ 30 kDa) (Fig. 2B).
Overall structure of the IBD of BaCcpC To better understand the molecular function of BaCcpC, we performed a crystallographic study on fulllength BaCcpC; however, it was not successful. Furthermore, crystallographic studies for the DBD and IBD of BaCcpC were separately performed. Finally, we obtained crystals for the IBD of BaCcpC and determined the crystal structure of BaCcpC-IBD in complex with citrate at 2.3 Å resolution using singlewavelength anomalous diffraction (SAD) phasing. The BaCcpC-IBD crystal belonged to space group C2 and had unit-cell parameters of a = 140.96, b = 90.90, c = 105.53 Å and β = 106.18°. The R work and R free of the nal model were 20.7% and 26.6%, respectively. The BaCcpC-IBD molecule is composed of two distinct regulatory domains: IBD-I (His90-Arg155 and Gly266-Gln289) and IBD-II (Asp168-Gly259) (Fig. 3A). IBD-I has three β-sheets, which are surrounded by three α-helices and 3 10 helices (Fig. 3B). IBD-II has four β-sheets, which are surrounded by two α-helices and two 3 10 helices (Fig. 3B). Both IBD-I and IBD-II subdomains adopt the typical α/β fold, which is connected by two crossover regions that form a hinge at central regions of two antiparallel β-strands (β4 and β9) (Fig. 3A). Five BaCcpC-IBD molecules are in the asymmetric unit, and each molecule has an r.m.s.d. of 0.201-0.275 Å for the 144-180 Cα atoms, which emphasizes the similarity of their conformations. BaCcpC-IBD forms dimers with a head-totail arrangement in the asymmetric unit of its crystal structure, and both molecules have essentially the same overall structure (Fig. 3C). Superposition of two dimeric molecules in the asymmetric unit gives an r.m.s.d. of 0.327 Å for 321 Cα. The dimeric interface is stabilized by the main chain interactions Val122-Thr212* (2.82 Å, * denoting the partner molecule), Val122-Leu214* (2.92 Å), Leu124-Asp216* (2.80 Å), and Thr126-Asp216* (3.33 Å) between the β2 strand and β6 strand (Fig. 3D).
The citrate binding site of BaCcpC-IBD To observe the citrate-bound state of BaCcpC-IBD, we added sodium citrate to the puri cation buffer during all protein puri cation steps. The electron density corresponding to citrate molecules is found at the positively charged interface between IBD-I and IBD-II of BaCcpC-IBD (Fig. 4A). In the electron density map, the positions of each carboxyl and hydroxyl group of citrate are clearly distinguished (Fig. 4A). Citrate is a small organic acid that includes three carboxyl groups, one hydroxyl group, and one prochiral centre. To distinguish between the terminal carboxyl groups, they were named pro-R and pro-S. The pro-R carboxyl group accepts a strong hydrogen bond from the backbone nitrogen atom of Ile100 (average distance for ve molecules in the asymmetric unit: 3.12 Å). In addition, the pro-S carboxyl group also has hydrogen bonds from the side-chain NE atoms of Arg147 (2.93 Å) and Arg260 (2.80 Å). The central carboxyl group of citrate accepts hydrogen bonds from the backbone nitrogen atom of Ser129 (2.70 Å), the side-chain hydroxyl group, and Ser189 (2.59 Å) and Ser191 (2.89 Å). The hydroxyl group of citrate interacts with the side-chain hydroxyl group of Ser129 (2.76 Å). The atoms of these residues that contact the citrate molecule are ~ 3.0 Å away from the latter's oxygen atoms, demonstrating that the citrate was coordinated by extensive strong hydrogen-bonding interactions (Fig. 4A).
Inducers are important for the function of LTTRs and often participate in the feedback loop of a speci c metabolic/synthesis pathway 15 . However, citrate molecules are inducers for BaCcpC, BsCcpC, and SaCcpE 8 9 . Sequence alignment shows that the Arg147 and Arg260 residues of BaCcpC are highly conserved in both BsCcpC and SaCcpE, whereas Ile100, Ser129, Ser189, and Ser191 of BaCcpC are not conserved in SaCcpE, being conserved in only BsCcpC (Fig. 4B). Moreover, the citrate-binding Arg147 and Arg260 residues of BaCcpC-IBD analogous to the citrate-binding Arg145 and Arg256 residues of CcpE, which are required for CcpE to evoke an appropriate response in the presence of citrate 16 17 (Fig. 4C). Therefore, we consider these two arginine residues to also play important roles in the citrate binding and functional assembly of BaCcpC.
SaCcpE-IBD, KpYneJ-CTD, and AtOccR-LBD also show dimers with the same tail-to-tail alignments as BaCcpC-IBD, but the dimeric interfaces and the rotations between these molecules exhibit signi cant differences. The dimeric interface of SaCcpE-IBD is formed by hydrogen bonds as well as some salt bridges between α1 and α5*, α1 and loop*, and loop and loop* (an asterisk indicates the partner molecule) (Fig. 5B). The dimer interface of KpYneJ-CTD consists of two α-helices that interact with one β-sheet, namely, α1-α5*, β2-α5*, α5-β2* and α5-α1*, which differ from the dimer interface of AtOccR-LBD in α1-α5*, loop-loop* and β2-β7* (Fig. 5B and Supplementary Fig. 1). These ndings indicated that IBD dimers are relatively stable, even after poor conservation in IBDs. Meanwhile, the rotation angles of dimers are distinct. In addition, CcpC-IBD is functionally distinct from CcpE-IBD but also recognizes citrate molecules. The monomers of citrate-bound BaCcpC-IBD and SaCcpE-IBD dimers are rotated at angles of approximately 33° and 35°, respectively (Fig. 5C). As a result, the rotation angle of the dimer interface of citrate-bound BaCcpC-IBD is very similar to that of citrate-bound SaCcpE-IBD. On the other hand, in the citrate-free state of SaCcpE-IBD, the angle of the dimer interface is approximately 80°, indicating that there is a change in dimer formation depending on citrate binding. Accordingly, the dimeric interface of in BaCcpC-IBD differs when citrate is absent or bound.

Conclusion
BaCcpC required both dyad symmetry regions I and II for recognizing the citB promoter, and the presence of citrate reduced citB binding. Citrate binds the interface between IBD-I and IBD-II of the IBD of BaCcpC. The IBD of BaCcpC shares low sequence similarity with other IBDs of the LTTR family but is similar in terms of the overall structure and dimer formation. Our results provide the framework for a functional analysis of CcpC as well as the diversity and similarity of IBDs of the LTTR family.

Construction, expression, and puri cation
The full-length (residues 1-293; named BaCcpC-FL) and C-terminal region of the IBD (residues 88-293; named BaCcpC-IBD) of CcpC were obtained from genomic DNA of B. amyloliquefaciens by PCR. The gene was cloned into the NcoI and XhoI sites of the pPROEX-HTA vector (Invitrogen, USA), which contains a hexahistidine tag (MSYYHHHHHH), a spacer region (DYDIPTT) and a tobacco etch virus (TEV) protease cleavage site (ENLYFQ) at the N-terminus. The construct was transformed into E. coli BL21 (DE3) competent cells to obtain the target proteins. Protein expression and puri cation procedures were the same for BaCcpC-FL and BaCcpC-IBD. Cells were grown in 2 L of Luria-Bertani (LB) medium containing 0.5 µg ml − 1 ampicillin at 310 K. When the OD 600 of the culture reached 0.8, 0.5 mM isopropyl-β-dthiogalactoside (IPTG) was added, and the culture was incubated at 303 K for 8 h. The bacterial cells were centrifuged for harvesting and resuspended in lysis buffer containing 20 mM Tris (pH 8.0), 10 mM sodium citrate, 150 mM NaCl, and 2 mM β-mercaptoethanol. Then, the cells were disrupted by sonication, and the lysate was centrifuged at 13000 rpm for 30 min at 277 K. The lipid fractions were mixed with a nickel-nitrilotriacetic acid (Ni-NTA) a nity resin (GE Healthcare) that had been preincubated with lysis buffer and stirred for 30 min at 277 K. The resin was washed and eluted with lysis buffer containing 20 mM imidazole and 300 mM imidazole. The fractions containing BaCcpC were pooled, and βmercaptoethanol was added to 10 mM ( nal concentration). To remove the hexahistidine tag, the mixture was incubated with a recombinant TEV protease at 298 K overnight. For further puri cation, the mixture was diluted 4-fold using 20 mM Tris (pH 8.0) buffer and loaded onto a Q anion-exchange column (HiTrap-Q; GE Healthcare, USA). The fractions containing BaCcpC were puri ed using a HiLoad Superdex 200 gel ltration column (GE Healthcare, USA) pre-equilibrated with buffer containing 20 mM Tris (pH 8.0), 10 mM sodium citrate, 150 mM NaCl and 2 mM β-mercaptoethanol. To express selenomethionine (Se-Met)substituted CcpC-IBD protein, the bacterial cells were cultured in 1 L of M9 medium supplemented with an amino acid mixture containing L-(+)-Se-Met at 310 K. When the OD 600 was between 0.6 and 0.8, the cells were induced with 0.5 mM IPTG for 8 h. The Se-Met-substituted protein was puri ed under the same conditions as the native protein. To obtain crystal structures of BaCcpC-IBD bound to citrate, we added 10 mM sodium citrate throughout the whole puri cation process. During puri cation, the presence of the proteins was detected by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) in a 15% gel with Coomassie blue R-250 for staining. Electrophoretic mobility shift assay (EMSA) experiments A chemiluminescent EMSA kit was purchased from Beyotime Biotechnology (Nanjing, China), and a biotin-labelled B. amyloliquefaciens citB promoter was synthesized by Generation (Wuhan, China). Supplementary Table S1 provides the list of oligonucleotide sequences used for EMSA analysis. EMSA experiments between BaCcpC and citB promoter DNA were performed at room temperature. For EMSA between BaCcpC and the citB protomer, various concentrations of puri ed BaCcpC (2.5-40 µM) protein were incubated with citB promoter (800 nM) for 30 min. For EMSA of BaCcpC with regions I and II of the citB promoter, puri ed BaCcpC (40 µM) protein was incubated with each citB promoter (800 nM). To determine the effect of citrate on binding between BaCcpC and the citB promoter, BaCcpC (40 µM) protein was incubated with citB promoter (800 nM) with various citrate concentrations (0-70 µM) for 30 min. After incubation, the reaction mixture was placed in a 6% acrylamide gel on ice using 0.5×Tris/Borate/EDTA (TBE) buffer. The product was analysed by chemiluminescence detection on the Tanon 4600 Chemiluminescent Imaging system (Tanon, China).

Size exclusion chromatography
The oligomer states of BaCcpC-FL and BaCcpC-IBD were analysed using gel ltration chromatography. Five hundred microlitres of the BaCcpC-FL or BaCcpC-IBD protein that had been incubated with 10 mM sodium citrate or not was loaded into a Superdex 200 10/300 GL column (GE Healthcare) at 295 K with a ow rate of 0.5 ml min − 1 ; this column had been was pre-equilibrated with 20 mM Tris buffer (pH 8.0) containing 150 mM NaCl and 2 mM 2-mercaptoethanol. Crystallization, data collection, structure determination, and re nement BaCcpC-IBD was concentrated to 20 mg ml − 1 using a Vivaspin centrifugal concentrator (Cut-off: 10 kDa, Millipore, USA). The initial crystallization was performed using the sitting-drop vapour-diffusion method at 295 K using a Crystal Screen HT high-throughput reagent kit (Hampton Research, USA). Crystals of BaCcpC-IBD were grown in 10% polyethylene glycol 6000, 5% 2-Methyl-2,4-pentanediol (MPD), and 0.1 M HEPES (pH 7.5) using a 1:1 ratio of protein to mother liquor at 287 K. Finally, crystals of BaCcpC-IBD were obtained in sitting drops over 8% (w/v) polyethylene glycol 6000, 6% MPD, and 0.1 M HEPES (pH 7.5) using a 1:1 ratio of protein to mother liquor at 287 K. Immediately after the single crystals were taken from their drop, they were soaked for 5 s in cryoprotectant solution consisting of the mother liquor solution containing 25% (v/v) glycerol and subsequently ash-cooled in liquid nitrogen.
The dataset was collected at 100 K using an ADSC Q310 CCD detector at Beamline 7A, Pohang Accelerator Laboratory (Pohang, Republic of Korea). The peak wavelength of Se-Met in BaCcpC was determined to be 0.9826 Å by a uorescence scan. Data were collected using the inverse beam method with an oscillation range of 1° per frame over a 360° rotation, and the exposure time was 5 sec per frame. The crystal of the BaCcpC-IBD protein diffracted to 2.3 Å resolution. Diffraction data were processed, merged, and scaled using the HKL-2000 program 20 . Initial phases were obtained using AUTOSOL in the software package PHENIX 21 . Re nement was performed using the Crystallographic Object-Oriented Toolkit (COOT) and phenix.re ne 22 23 . The data collection and re nement statistics are given in Supplementary Table S2. Structural images were generated by PyMol 24 .

Declarations
Data availability. The atomic coordinates and structure factors for BaCcpC-IBD (PDB ID 7DMW) have benn deposited in the RCSB Protein Data Bank , www.pdb.org.