Crystal structure of N-acetylmannosamine kinase from Fusobacterium nucleatum

The enzyme N-acetylmannosamine kinase (NanK) catalyzes the second step of the bacterial sialic acid catabolic pathway. Here, the structure of F. nucleatum NanK is presented at 2.23 Å resolution.


Introduction
Sialic acids comprise a large family of acidic sugars that contain a core nine-carbon backbone (Angata & Varki, 2002;Vimr et al., 2004;Varki, 1992). The most prevalent type of sialic acid is N-acetylneuraminic acid (Neu5Ac), which is found at the terminal positions of glycoconjugates in humans and other deuterostomes. Sialylation of cell surfaces is crucial for cell-cell interactions and for a range of biological functions that involve cell-signalling processes and modulation of the immune response (Tanner, 2005;Varki, 2007;Vimr et al., 2004;Kazatchkine et al., 1979;Lanoue et al., 2002). Prompted by evolution to adapt to the sialic acid-rich milieu, many bacteria have developed mechanisms to competitively secure their niche on mucosal surfaces (Almagro-Moreno & Boyd, 2009b). Bacteria acquire sialic acids either by cleaving them from the host's glycoconjugates or by scavenging (Vimr, 2013). Once the sialic acid has been transported into the cytosol by specific transporters, some bacteria can incorporate it as a nonreducing terminal sugar on their cell surface for molecular mimicry, and thereby evade the host's innate and adaptive immune response, or can degrade it for use as a carbon and nitrogen source (Mulligan et al., 2011). The ability to utilize sialic acid as an energy source is chiefly exploited by commensal and pathogenic bacteria and requires a cluster of genes, known as the nan-nag cluster (Almagro- Moreno & Boyd, 2009bHaines-Menges et al., 2015;Fig. 1). The sialic acid nan-nag gene cluster was identified in the genome of a Fusobacterium species which exists as a commensal in the gastrointestinal tract and as a periodontal pathogen (Almagro-Moreno & Boyd, 2009a). Once host sialic acid has been transported across the cytoplasmic membrane (by a twocomponent sialic acid tripartite ATP-independent periplasmic transport system in F. nucleatum), degradation of Neu5Ac to fructose 6-phosphate starts with the conversion of sialic acid by an N-acetylneuraminate lyase (NanA), yielding N-acetylmannosamine (ManNAc) and pyruvate. A phosphoryl group from ATP is then transferred to ManNAc by a kinase (NanK), producing N-acetylmannosamine 6-phosphate (ManNAc-6-P), which in turn is converted to N-acetylglucosamine 6-phosphate (GlcNAc-6-P) by N-acetylmannosamine-6-phosphate 2-epimerase. Finally, GlcNAC-6-P is deacylated by N-acetylglucosamine-6-phosphate deacetylase (NagA) and is subsequently deaminated by glucosamine-6-phosphate deaminase (NagB) to yield fructose 6-phosphate (Vimr & Troy, 1985).
F. nucleatum N-acetylmannosamine kinase (FnNanK) belongs to the repressor, open reading frame, kinase (ROK) superfamily of proteins. This collection of polypeptides is primarily composed of transcriptional repressors, sugar kinases and other unknown gene clusters (Titgemeyer et al., 1994). The salient unifying features of the ROK scaffold are a nucleotide-binding region in the N-terminal region, a strictly conserved catalytic aspartate residue that serves as a Schiff base during phosphoryl transfer and a zinc-binding motif that is implicated in the stability of the active site of the enzyme (Martinez et al., 2012;Conejo et al., 2010). Structural representatives of human N-acetylmannosamine kinase (hMNK) domain of UDP-N-acetylglucosamine-2-epimerase/N-acetylmannosamine kinase (Tong et al., 2009;Martinez et al., 2012) and two putative NanKs from Escherichia coli (EcNanK; PDB entry 2aa4; New York SGX Research Center for Structural Genomics, unpublished work) and Listeria monocytogenes (LmNanK; PDB entry 4htl; Midwest Center for Structural Genomics, unpublished work) have been deposited in the Protein Data Bank. The overall fold is a butterfly-shaped homodimer. Each monomer consists of two domains that are connected by two hinge loops, allowing the kinase to change from an open conformation to a closed conformation upon substrate binding.
Here, we present the structural analysis of N-acetylmannosamine kinase from F. nucleatum. This structure is important for inhibitor design, which may lead to the development of antimicrobial agents for the treatment of periodontal disease.
The cells were disrupted using an EmulsiFlex-C3 (Avestin) at 124 MPa for two cycles. The cell debris was removed by centrifugation at 107 000g for 30 min at 4 C. Macromoleculeproduction information is summarized in Table 1.

Protein purification
FnNanK was purified by affinity chromatography at 4 C using a 5 ml HisTrap FF column (GE Healthcare) preequilibrated with buffer A. The bound protein was washed with buffer A and the protein was then eluted with buffer B [20 mM Tris-HCl pH 8.0, 300 mM NaCl, 500 mM imidazole, 5%(v/v) glycerol]. As a final polishing step, the protein was loaded onto a HiLoad 16/600 Superdex 200 size-exclusion column (GE Healthcare) pre-equilibrated with buffer C [20 mM Tris-HCl pH 8.0, 300 mM NaCl, 5%(v/v) glycerol, 1 mM DTT]. The purity of the eluted protein samples was evaluated using SDS-PAGE. The pure samples corresponding to FnNanK were pooled together and concentrated using Vivaspin concentrators to a final concentration of 14 mg ml À1 . The protein concentration was determined using an ND-1000 spectrophotometer at 280 nm, using an extinction coefficient of 27 005 M À1 cm À1 and a molecular weight of 33.9 kDa.

Crystallization
The initial screening for crystallization conditions for F. nucleatum NanK was performed at 293 K using a Mosquito nanolitre-dispensing robot (TTP Labtech) with Crystal Screen HT (Hampton Research). The sitting-drop vapour-diffusion method was used, mixing 0.2 ml protein solution (14 mg ml À1 ) and 0.2 ml reservoir solution. Within one week, rod-shaped crystals of FnNanK were obtained using a reservoir solution consisting of 0.2 M lithium sulfate monohydrate, 0.1 M Tris-HCl pH 8.5, 30%(w/v) PEG 4000. The crystals were flashcooled in liquid nitrogen prior to the diffraction experiment. Crystallization conditions are summarized in Table 2.

Data collection and processing
The crystals of FnNanK diffracted to 2.23 Å resolution. X-ray diffraction data were collected at 100 K on the I911-3 beamline at the MAX-lab National Research Laboratory for Nuclear Physics and Synchrotron Radiation Research, Lund, Sweden using X-rays at a wavelength of 1.0 Å . Diffraction intensities were processed and integrated using iMosflm (Battye et al., 2011) and were scaled using AIMLESS from the CCP4 program suite (Evans & Murshudov, 2013). Datacollection and processing statistics are shown in Table 3.

Structure solution and refinement
The structure of F. nucleatum NanK was determined by molecular replacement using the coordinates of L. monocytogenes NanK (PDB entry 4htl) as a search model using Phaser (Read, 2001) within the PHENIX software suite (Adams et al., 1999(Adams et al., , 2011. The phenix.autobuild program was used for initial model building and electron-density improvement. Subsequently, phenix.refine was used for rigid-body refinement, maximum-likelihood least-squares refinement, simulated annealing and addition of water molecules to the    structure. Manual inspection and model building were performed using Coot (Emsley et al., 2010). Structure-solution and refinement statistics are summarized in Table 4.

Results and discussion
3.1. Protein production, purification and crystallization F. nucleatum NanK (FnNanK) was successfully expressed and purified using a two-step procedure consisting of affinity and size-exclusion chromatography and was concentrated to a final concentration of 14 mg ml À1 . The preparations were homogenous when analyzed by SDS-PAGE and sizeexclusion chromatography. Using the sitting-drop vapourdiffusion method, rod-shaped crystals formed within one week using a reservoir solution consisting of 0.2 M lithium sulfate monohydrate, 0.1 M Tris-HCl pH 8.5, 30%(w/v) PEG 4000.

Crystal structure of F. nucleatum NanK
The structure of FnNanK has one homodimer in the asymmetric unit, corresponding to a solvent content of 38%. The structure was refined to 2.23 Å resolution with an R cryst of 17.7% and an R free of 22% (Tables 3 and 4). No electron density could be attributed to the residues of the N-terminal tag, which are consequently missing from the final model. The structure has no Ramachandran outliers, with 98% and 2% of the residues in the preferred and allowed regions, respectively.
3.2.1. Overall structure. FnNanK is a butterfly-shaped homodimer, as seen in other members of the ROK family (Fig. 2). The monomer structure has an elongated shape and is composed of two / domains that are connected by two hinge loops (residues 119-125 and 269-271). The putative active site is located in a large cleft between the N-terminal domain, which is made of two fragments (residues 1-118 and 272-291), and the slightly smaller C-terminal dimerization domain (residues 126-268). The N-terminal domain contains a central mixed, twisted five-stranded -sheet (1-4 and 7) surrounded by four -helices (1-3 and 11) and a short -hairpin (5-6) (Fig. 3a). The C-terminal dimerization domain consists of a mixed, twisted four-stranded -sheet (8-11) that is sandwiched between the N-terminal domain and a cluster of -helices and 3 10 -helices of the C-terminal domain (Fig. 3b). Forty residues of the helix cluster and connecting loops of the C-terminal domain create a 1479 Å 2 dimer interface stabilized by direct hydrogen bonds and solventmediated hydrogen bonds.
3.2.2. The putative active site. In this study, we report an apo structure of FnNanK. The N-acetylmannosamine kinase domain (hMNK) of the human bifunctional UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase shares 23% sequence identity with FnNanK and has been characterized both functionally and structurally (Martinez et al., 2012). The structure of hMNK in complex with ManNAc and ADP (PDB entry 2yhy) can be superimposed on NanK with an r.m.s. deviation of 2.5 Å for 280 C atoms, making it possible to model the binding of ManNAc and ADP in the active site of FnNanK (Fig. 4a). Residues that are involved in substrate and ATP binding are located in both the N-terminal and C-terminal domains. The conserved residues in hMNK (Asn516, Asp517, Arg477, Glu566, His569 and Glu588) that are required for the coordination of ManNAc (Martinez et al., 2012) are superimposable with Asn106, Asp107, Gln67, Glu156, His159 and Glu168 in FnNanK.
The bacterial EcNanK (PDB entry 2aa4) and LmNanK structures (PDB entry 4htl) were superimposed with FnNanK. The r.m.s. deviations for the structural alignments of EcNanK (289 C atoms) and LmNanK (280 C atoms) with FnNanK are 2.6 and 1.9 Å , respectively. The putative active-site residues in FnNanK (Asn106, Asp107, Gln67, Glu156, His159 and Overall structure of F. nucleatum apo N-acetylmannosamine kinase. The N-terminal domain is coloured in blue shades, the C-terminal dimerization domain in red shades and the the hinge loops are depicted in yellow. For clarity, based on the human hMNK structure (PDB entry 2yhy), ManNAc (green sticks), ADP (white sticks) and Mg 2+ (blue sphere) have been modelled in the putative active site. Glu168) that are predicted to be involved in substrate binding are superimposable with those in EcNanK (Asn104, Asp105, Ile66, His153, His156 and Glu175) and LmNanK (Asn102, Asp103, Tyr64, Glu152, Tyr155 and Asn172) (Fig. 4b).
3.2.3. FnNanK lacks a zinc-binding site. The common signature motifs of the ROK scaffold are (i) an N-terminal region containing the nucleotide-binding site with a DxGxT sequence motif, (ii) a strictly conserved catalytic aspartate within the active-site loop, (iii) an ExGH motif that interacts with the sugar substrate and (iv) a cysteine-rich zinc-binding motif with sequence CxCGxxGCx(E/D) (Conejo et al., 2010). The first three signature motifs are also conserved in FnNanK.
Although FnNanK retains most of the consensus motifs unique to the ROK family, the lack of a zinc-binding site with sequence xCGxxGCx(E/D) is evident both in the sequence and structure alignments. The zinc-binding motif is implicated in upholding the structural integrity of the active site (Mesak et al., 2004;Martinez et al., 2012). A recent report suggested that mutation of the cysteines in the zinc-binding motif through site-directed mutagenesis renders Bacillus subtilis glucokinase inactive (Mesak et al., 2004). The Zn atom is coordinated by three thiols within the cysteine-rich motif and a fourth coordinating conserved histidine relating the zincmotif region to the substrate-binding site (Schiefner et al., 2005). Superimposition of the residues that are involved both in zinc binding and substrate binding in PDB entries 2aa4 (pink) and 4htl (green) and in FnNanK (blue) highlights the absence of the cysteine-rich region in FnNanK (Fig. 4b).
FnNanK lacks the zinc-binding motif, and sequence analysis of the known N-acetylmannosamine kinases shows that the consensus sequence xCGxxGCx(E/D) which denotes the zincmotif region is not present in FnNanK. In LmNanK (PDB entry 4htl; Fig. 4b) there seems to be no deletion; however, the loop contains no cysteine residues. The lack of zinc-binding sequence also extends to methicillin-resistant Staphylococcus aureus (MRSA) NanK (North et al., 2013). However, FnNanK retains the highly conserved His159. The corresponding residue is His156 in EcNanK, and this histidine has been shown to bind both to the zinc ion and to ManNAc in human NanK (His569; Nocek et al., 2011;Martinez et al., 2012). Mutations of the two cysteine residues associated with zinc binding to serine and alanine in the E. coli Mlc repressor compromised its repressor function.
The three cysteine residues and histidine residue engaged in zinc-ion coordination are considered to be a distinct motif in the ROK family (Schiefner et al., 2005). Although the three cysteines are not present in FnNanK, His159 is noted to be significantly shifted on superimposition with EcNanK. The measured distance between His159 in FnNanK and the substrate ManNAc in hMNK is twice as far compared with the distance between His569 in hMNK and ManNAc. In FnNanK, the glutamate residue Glu166 is markedly visible in place of the cysteine residues (Fig. 4b). The structure of FnNanK in complex with substrate analogues and ATP (or an analogue) is required to predict the change in conformation that is needed to complete the binding of the substrate and ATP. Overall structures of the N-terminal domain (a) and C-terminal dimerization domain (b). The helices and strands are numbered. The residues that span and flank each domain are marked. Domain 1 starts from the N-terminus and ends at residue 118 and then continues from residue 272 to the C-terminus (blue). Residues 126-268 form the dimerization domain.