Structural Analyses of Helicobacter Pylori FolC Conducting Glutamation in Folate Metabolism

Abstract

FolC plays important roles in the folate metabolism of cells by attaching l-Glu to dihydropteroate (DHP) and folate, which are known activities of dihydrofolate synthetase (DHFS) and folylpolyglutamate synthetase (FPGS), respectively. Here, we determined the crystal structure of Helicobacter pylori FolC (HpFolC) at 1.95 Å resolution using the single-wavelength anomalous diffraction method. HpFolC has globular N- and C-terminal domains connected by a single loop, and a binding site for ATP is located between the two domains. Apo-HpFolC was crystallized in the presence of citrate in a crystallization solution, which was held in the ATP-binding site. Structural motifs such as the P-loop and Ω-loop of HpFolC for binding of ATP and two magnesium ions are well conserved in spite of the low overall sequence similarity to other FolC/FPGSs. The Ω-loop would also recognize a folate molecule, and the DHP-binding loop of HpFolC is expected to exhibit a unique recognition mode on DHP, compared with other FolCs. Because human FolC is known to only exhibit FPGS activity, the DHFS activity of bacterial FolC is an attractive target for the eradication of pathogenic bacteria. Consequently, our structural analyses of HpFolC provide a valuable foundation for a universal antibacterial strategy against H. pylori as well as other pathogenic bacteria.

1. Introduction

Helicobacter pylori is a pathogenic Gram-negative bacterium, which has been classified as a class 1 carcinogen that contributes to chronic inflammation within the human gastric niche and can lead to stomach cancer [1,2,3]. Eradication of H. pylori usually starts with triple therapy, which is a combination of a proton pump inhibitor, amoxicillin, and clarithromycin. Failure of this therapy as a result of antibiotic resistance to clarithromycin has been combated with newly developed approaches, including bismuth and tetracycline, but long-term plans still require novel antibacterial targets for H. pylori [4,5].

In cells, folate plays a central role in the single-carbon transfer system involved in DNA (thymidine and purines) and amino acid (Gly, Ser, and Met) synthesis; hence, it is essential for cell growth and replication [6]. Folate and its derivatives in cells are reported to exist mainly in polyglutamated forms, which were initially speculated to be a reservoir of cellular glutamate [6] and, further, to be substrate forms with elevated affinities to several folate-processing enzymes [7,8,9]. Several folate-derived drugs targeting folate metabolism also exhibit increased efficiencies by polyglutamation [10]. Folate polyglutamation is achieved through folylpolyglutamate synthetase (FPGS, EC 6.3.2.17), which attaches l-Glu to acid moieties of folate and its polyglutamated derivatives in an ATP-dependent manner (Figure 1).

FPGS is known as a ubiquitous enzyme found in all living cells. Interestingly, in some organisms such as Corynebacterium species, Escherichia coli, Neisseria gonorrhoeae, Mycobacterium tuberculosis, and Plasmodium falciparum, FPGS has been reported to additionally perform the activity of dihydrofolate synthetase (DHFS, EC 6.3.2.12) as a bifunctional enzyme (referred to as FolC in this paper) and is involved in the de novo synthesis of folate [11,12,13,14,15] (Figure 1). H. pylori is also predicted to have bifunctional FPGS exhibiting DHFS activity [16], which implies that a selective inhibitor against the DHFS activity of H. pylori FolC (HpFolC) could be an attractive candidate for an antibacterial drug [17].

So far, structures of FPGSs from five different bacterial species have been determined: E. coli FolC (EcFolC, PDB codes: 1W78 and 1W7K) [18], Lactobacillus casei FPGS (LcFPGS, PDB codes: 1FGS, 1JBV, 1JBW, 2GC5, 2GC6, 2GCA, and 2GCB) [19,20,21], M. tuberculosis FolC (MtFolC, PDB codes: 2VOR and 2VOS) [22], Thermotoga maritima FolC (TmFolC, PDB code: 1O5Z), and Yersinia pestis FolC (YpFolC, PDB codes: 3N2A, 3NRS, 3PYZ, and 3QCZ). Here, we determined the crystal structure of HpFolC at 1.95 Å resolution using the single-wavelength anomalous diffraction (SAD) method. We analyzed the structure of HpFolC in comparison with the other FolC/FPGS structures, which should enable a deeper understanding of folate polyglutamation in H. pylori and provide a structural basis for the development of new antibiotics against H. pylori.

2. Materials and Methods

2.1. Cloning, Protein Expression, and Purification

The HpFolC gene (hp_1545, residues 1–394) was cloned into the pET-21a(+) vector (Novagen, Darmstadt, Germany) with a C-terminal hexahistidine tag. The recombinant vector was transformed into Rosetta 2(DE3), an E. coli strain. The transformed cells were selected and cultured with Luria-Bertani media containing 50 μg/mL ampicillin. Cells were shake-incubated at 37 °C until OD₆₀₀ reached 0.5, and 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) was added for induction. After induction, the cells were incubated for an additional 16 h at 20 °C. Cultured cells were harvested at 6,000× g and lysed by sonication in a buffer containing 20 mM Tris-HCl (pH 7.5), 500 mM NaCl, 35 mM imidazole, and 1 mM phenylmethanesulfonyl fluoride. The cell lysate was centrifuged at 35,000× g for 1 h and the supernatant was filtered with 0.45 μm syringe filter (Sartorius, Göttingen, Germany). The filtrate was loaded onto a HiTrap chelating HP column (GE Healthcare, Chicago, IL, USA) and eluted by adding an increasing gradient of a buffer containing 20 mM Tris-HCl (pH 7.5), 500 mM NaCl, and 1 M imidazole. The eluted fractions containing HpFolC protein were desalted using a HiPrep 26/10 desalting column (GE Healthcare, Chicago, IL, USA) with a buffer containing 20 mM Tris-HCl (pH 9.0) and 50 mM NaCl. The desalted sample was loaded onto a HiTrap 5 ml Q HP column (GE Healthcare, Chicago, IL, USA) and eluted by adding an increasing gradient of a buffer containing 20 mM Tris-HCl (pH 9.0) and 1 M NaCl. The eluted protein was loaded onto a HiLoad 16/600 Superdex 200 pg column (GE Healthcare, Chicago, IL, USA) equilibrated with a buffer containing 10 mM HEPES-NaOH (pH 7.0) and 100 mM NaCl. The eluted protein was concentrated for crystallization.

Selenomethionine (SeMet)-derived HpFolC was expressed with B834(DE3), a methionine auxotrophic E. coli strain. The transformed cells were cultured in the media containing M9 minimal salts (Sigma-Aldrich, Darmstadt, Germany) and an amino acid mix with l-selenomethionine. Cells were shake-incubated at 37 °C until OD₆₀₀ reached 0.5, and 0.5 mM IPTG was added for induction. After induction, the cells were incubated for an additional 16 h at 20 °C. The subsequent purification procedure was the same as for native HpFolC.

2.2. Crystallography

Initial crystal of HpFolC was obtained using the sitting-drop vapor diffusion method from a commercial crystallization-screening reagent containing 20% (w/v) polyethylene glycol 3350 and 0.2 M ammonium citrate tribasic (pH 7.0) (The Classic Ⅱ suite; QIAGEN, Hilden, Germany). HpFolC crystals were further optimized using the hanging-drop vapor diffusion method for X-ray diffraction experiments. Before exposure to X-ray beams, crystals were soaked in the crystallization solution supplemented with 25% glycerol for cryoprotection and flash-cooled in a 100 K nitrogen gas stream. Diffraction data were indexed and scaled with HKL2000 [23]. The phase of the SeMet-derived crystal was solved using the SAD method with the Autosol program [24,25], and the obtained initial model was used as the template for MR phasing of native crystal data with the Phaser program [26]. Models were built, refined, and validated with Coot [27], phenix.refine [28], and MolProbity [29], respectively. Statistics for data collection and model refinement are summarized in

Table 1. Statistics for data collection and model refinement.

	HpFolC	SeMet-substituted HpFolC
PDB code	6K8C
Data collection
Beam line	PLS-5C
Space group	C2	P41212
Wavelength (Å)	0.9795	0.9798 (Peak)
Unit cell parameter
a, b, c (Å)	131.45, 61.64, 69.74	117.90, 117.90, 69.18
α, β, γ (°)	90, 115.96, 90	90, 90, 90
Resolution range (Å)	50.00–1.95 (1.98–1.95)1	50.00–2.50 (2.54–2.50)1
No. of unique reflections	36178 (3591)1	17401 (1701)1
I/σ(I)	26.5 (4.2)1	69.2 (15.9)1
Wilson B-factor (Å2)	33.9	28.4
Rmerge (%)	0.059 (0.431)1	0.148 (0.635)1
Redundancy	7.4 (7.5)1	38.1 (39.9)1
Completeness (%)	98.9 (98.9)1	100.0 (100.0)1
Refinement
Resolution	30.00–1.95 (2.02–1.95)1
Rwork/Rfree2	0.184/0.219
R.m.s. deviations
Bonds (Å)	0.004
Angles (°)	0.640
No. of non-H atoms	3362
Protein	3029
Ligand	193
Water	314
Average B-factors (Å2)	41.22
Protein	40.42
Ligand	62.68
Water	47.59
Ramachandran favored/outlier (%)	98.4/0.04
Rotamer outlier (%)	0.00

¹ Values of the highest shell are indicated in parentheses. ² 5% of total reflections were used to calculate the R_free value. ³ One citrate (C₆H₈O₇) and one glycerol (C₃H₈O₃) molecule were included. ⁴ One cis-Pro (Pro82) was excluded. Abbreviations: HpFolC—Helicobacter pylori FolC; SeMet—selenomethionine.

.

2.3. Data Availibility

Atomic coordinates and structure factor files of the HpFolC structure described in this research have been deposited in the Protein Data Bank (www.wwpdb.org) with the accession code 6K8C.

3. Results and Discussions

3.1. HpFolC Shares Structural Features of Other FolC/FPGSs

HpFolC largely contains N- and C-terminal domains separated by a turn (Ile272 and Gly273) (Figure 2a,b). The N-terminal domain (residues 1–272) contains a mixed β-sheet of six parallel strands and one antiparallel strand flanked by 13 helices (10 α-helices and 3 3₁₀-helices), and the C-terminal domain (residues 273–394) contains a mixed β-sheet of five parallel strands and one antiparallel strand flanked by four helices (3 α-helices and 1 α-3₁₀ mixed helix, αη15). The N- and C-terminal domains form an interface with a hollow cavity inside, and a citrate molecule from the crystallization solution seems to be captured (Figure 2c). FolC is classified as a member of the Mur ligase family, in which other member proteins are usually involved in peptidoglycan synthesis as amino acid ligases. The overall structure of HpFolC is quite similar to other FolC/FPGS structures from Dali analyses (Z-score: 31.8–36.1). An outstanding topological difference of HpFolC is its α12 in the N-terminal domain, where the other reported FolC/FPGS structures have an antiparallel β-sheet with three strands instead (Figure 2b).

HpFolC was observed to share characteristic functional components, such as the dihydropteroate (DHP)-binding loop (α1–α2, residues 18–27), P-loop (phosphate-binding loop, β1–α3, residues 54–58), and Ω-loop (β2–β3, residues 80–90) (Figure 2a). Considering two activities of FolC, dihydropteroate, tetrahydrofolate, ATP, and l-Glu could be the substrates of HpFolC. Among the 16 reported FolC/FPGS structures from five species, two structures (PDB codes: 1W78 and 1JBW) were elucidated in complex with phosphorylated dihydropteroate (DHP-P) [18] and 5,10-methylene-6-hydrofolate [20], respectively, and nine structures (PDB codes: 1JBW, 1JBV, 1W7K, 1W78, 3QCZ, 3PYZ, 3N2A, 2VOR, and 2VOS) were elucidated in complex with ATP or its homologues [18,20,22]. The YpFolC structure (PDB code: 3QCZ) was elucidated in complex with the substrate glutamate.

3.2. Binding Site of HpFolC for DHP in the Vicinity of the DHP-Binding Loop

The DHP-binding loop of HpFolC is mainly composed of charged/aromatic residues (¹⁸TKPKEYHKFD²⁷). Previous studies on EcFolC revealed that hydrophobic residues on the DHP-binding loop (Ile28 and Leu30) form a stacking interaction with DHP-P together with Phe124 and Ala155 [18]. The stacking interaction causes the DHP-binding loop of EcFolC to become ordered and to have the “closed” conformation in complex with DHP. Additionally, adjacent polar residues (Thr122, Glu125, and Asp154) of EcFolC interact with hydrogen-bond donors/acceptors of the pterin moiety. Considering the bifunctional properties of FolC, the DHP-binding site would be conserved among FolCs. Our sequence alignment with HpFolC, MtFolC, EcFolC, and LcFPGS showed that the corresponding residues to Thr122, Phe124, Glu125, and Asp154 of EcFolC are well-conserved in the other proteins (Figure 3). These conserved residues are thought to interact with DHP as in EcFolC. However, the sequence of the DHP-binding loop of EcFolC is not conserved. Instead of the hydrophobic stacking of EcFolC with DHP, the charged/aromatic residues of the DHP-binding loop in HpFolC could form probable hydrogen-bond/hydrophobic/stacking interactions with the pterin moiety of DHP.

3.3. P-Loop, a Structural Motif of HpFolC for ATP Binding

HpFolC features the P-loop (residues 54–60, GTNGKGS), which is conserved among other FolC/FPGSs as the glycine-rich sequence “GTNGKx(S/T)” (Figure 3). Previous studies on complex structures of FolC/FPGSs with ATP-homologue molecules revealed that an ATP is sandwiched between the N- and C-terminal domains of FolC/FPGS. Especially, the P-loop provides contacts and interactions with ATP from the N-terminal domain of FolC/FPGS [18,20,22]: the Asn residue forms a hydrogen bond with the ribose moiety of ATP, the “GKx” sequence interacts with the β-phosphate of ATP, and the “x(S/T)” sequence interacts with the α-phosphate with its backbone and the side chain of the “(S/T)” residue. Based on similarities of sequence and structure, the P-loop of HpFolC seems to recognize an ATP in a similar way (Figure 4b).

At the opposite side of the P-loop, binding of the ribose moiety, α-phosphate, and γ-phosphate would be stabilized by hydrogen bonds with strictly conserved Asp285, Arg274, and His288 of HpFolC, respectively, as in LcFPGS (Figure 4c). The side chain of His288 of HpFolC in our structure is out of binding conformation and expected to move toward the ATP-binding site with ATP binding.

3.4. Structural Motifs of HpFolC for ATP Binding via Two Magnesium Ions

In the ATP binding of FolC, two magnesium ions (Mg1 and Mg2) are involved [18,19,20]. One magnesium (Mg1) is located between β- and γ-phosphates of ATP to interact with them and can be coordinated by the Ω-loop of FolC/FPGS, which is named after its Ω-like shape. The Ω-loop is an additional feature of the N-terminal domains of FolC/FPGSs when compared to other structurally similar nucleotide-binding proteins and is also observable in HpFolC [19]. The Ω-loop contains approximately 10 residues, in which “Ser-cis-Pro” and “Glu-Arg” motifs are strictly conserved, including their locations. The residue following the Ser-cis-Pro motif is the conserved His/Phe residue, which is discussed below. The next conserved hydrophobic residue (Val84 in HpFolC, Leu101 in MtFolC, Leu86 in EcFolC, and Ile76 in LcFPGS) makes a hydrophobic contact with the C-terminal domain. The Ω-loop holds Mg1 with the carbonyl group of Ser in the Ser-cis-Pro motif in FolC/FPGS-ATP complex structures [18,19,20]. Together with the carbonyl group of Ser in the Ω-loop, the side chain of the adjacent conserved Glu (Glu174 of MtFolC, mentioned as Glu176 in the original paper but the numbering in our paper follows the MtFolC sequence in the UniProt database (entry: I6Y0R5); Glu146 of EcFolC; and Glu143 of LcFPGS) also coordinates Mg1. The corresponding Ser81 and Glu142 of HpFolC would provide similar functions to other FolC/FPGSs (Figure 4d).

The other magnesium ion (Mg2) coordinates the γ-phosphate of ATP [18,20]. His and carbamylated Lys residues (His173 and Lys188 in EcFolC; His170 and Lys185 in LcFPGS) were observed to hold Mg2 together, and they are also conserved as His169 and Lys184 in HpFolC, respectively (Figure 4d, His169 of HpFolC is disordered). The carbamylated Lys has been observed in the Mur ligase superfamily [30,31]. However, the carbamylated Lys as well as the coordinated Mg2 was not observed in HpFolC and MtFolC-AMPPCP complex structures. It seems that Lys carbamylation is essential for the recruitment of Mg2 and subsequent catalytic reactions. A previous study on MtFolC suggests that crystallization conditions with a pH higher than 7.0 would result in the carbamylation of the conserved Lys, and further, this carbamylation condition would be a factor for the pH-dependent activity of FolC/FPGS [22]. Although HpFolC was crystallized at pH 7.0, Lys184 of HpFolC was not carbamylated.

3.5. Analyses of Folate Binding of HpFolC, Focused on the Ω-Loop

As mentioned above, HpFolC has the conserved His83 residue preceded by the Ser-cis-Pro motif of the Ω-loop, similar to other FolC/FPGSs (His100 in MtFolC, His85 in EcFolC, and Phe75 in LcFPGS). In the LcFPGS-folate complex structure, Phe75 was shown to interact with the pterin moiety of folate by stacking and a hydrogen bond with its side chain and backbone, respectively. The equivalent His of other FolCs and HpFolC are also thought to function similarly (Figure 5b). Tyr414 at the opposite side of Phe75 of LcFPGS sandwiches the pterin moiety together with Phe75, and equivalent residues of other FolC/FPGS (Leu380 of HpFolC, Val473 in MtFolC, and His405 in EcFolC) could also provide supporting surfaces against pterin moieties with their side chains (Figure 5c). In addition, the glutamate moiety of folate was suggested to extend toward the γ-phosphate binding site with dynamic movement in the binding mode of folate to LcFPGS [20].

3.6. l-Glu Binding Site of FolC/FPGSs

In the PDB data, the structure of YpFolC in complex with an l-Glu molecule is deposited (PDB code: 3QCZ). In the structure, the loop between the fifth β-strand and sixth helix (residues Thr167–Tyr179) is open toward the solvent area and the resultant space accommodates the l-Glu molecule. The equivalent loop in HpFolC (residues 163–176) is partially disordered, which implies its flexible movement for ligand binding. However, His169 on this loop (His201 in MtFolC, His173 in EcFolC, and His170 in LcFPGS) was suggested to coordinate Mg2 in the presence of carbamylated Lys, which induces the closed loop conformation. This could provide some clues regarding the mechanism of FolC/FPGS, in which l-Glu binding could occur after dissociation of Mg2 and γ-phosphate or l-Glu binding would induce dissociation of Mg2 and γ-phosphate.

4. Conclusions

Our analyses of the HpFolC structure revealed that its substrate-recognition components are shared among other reported FolC/FPGS structures. Because the DHFS activity of HpFolC is of great interest for the development of new antibiotics against H. pylori, we compared substrate bindings of HpFolC with LcFPGS that would lack DHFS activity as well as human FPGS. However, LcFPGS is structurally similar to bifunctional FolCs, and any outstanding structural difference between FolC and FPGS could not be observed. Thus, further structural and functional investigations of FolC/FPGS are needed for a deeper understanding of the glutamation of folate and DHP in cells. Finally, further structure determination of human FPGS would accelerate the development of selective antibiotics targeting HpFolC.