Crystal structure of di­hydroorotate de­hydrogenase from Helicobacter pylori with bound flavin mononucleotide

This article is part of a focused issue on empowering education through structural genomics.

Helicobacter pylori is the primary causative agent of peptic ulcer disease, among other gastrointestinal ailments, and currently affects over half of the global population. Although some treatments exist, growing resistance to these drugs has prompted efforts to develop novel approaches to fighting this pathogen. To generate many of the nucleotides essential to biochemical processes, H. pylori relies exclusively on the de novo biosynthesis of these molecules. Recent drug-discovery efforts have targeted the first committed step of this pathway, catalysed by a class 2 dihydroorotate dehydrogenase (DHODH). However, these initiatives have been limited by the lack of a crystal structure. Here, we detail the crystal structure of H. pylori DHODH (HpDHODH) at 2.25 Å resolution (PDB entry 6b8s). We performed a large-scale bioinformatics search to find evolutionary homologs. Our results indicate that HpDHODH shows high conservation of both sequence and structure in its active site. We identified key polar interactions between the HpDHODH protein and its requisite flavin mononucleotide (FMN) cofactor, identifying amino-acid residues that are critical to its function. Most notably, we found that HpDHODH maintains several structural features that allow it to associate with the inner membrane and utilize ubiquinone to achieve catalytic turnover. We discovered a hydrophobic channel that runs from the putative membrane interface on the N-terminal microdomain to the core of the protein. We predict that this channel establishes a connection between the ubiquinone pool in the membrane and the FMN in the active site. These findings provide a structural explanation for the competitive inhibition of ubiquinone by pyrazole-based compounds that was determined biochemically in other studies. Understanding this mechanism may facilitate the development of new drugs targeting this enzyme and push the effort to find a resistance-free treatment for H. pylori.

1. Introduction

Helicobacter pylori is a Gram-negative, microaerophilic helix-shaped bacterium that colonizes the gastric mucosa of humans (Dunn et al., 1997). H. pylori is the causative agent of various gastrointestinal diseases, such as chronic gastritis, peptic ulcers, gastric cancer and mucosa-associated lymphoid tissue lymphoma. H. pylori infection is a global problem, affecting more than 50% of the world's population, including around 80% of middle-aged adults in developing countries (Hooi et al., 2017; Suerbaum & Michetti, 2002). Unfortunately, approaches to treating H. pylori in the clinic are becoming increasingly limited due to rampant antibacterial resistance. In 2017, the World Health Organization designated clarithromycin-resistant H. pylori as a high priority for focused antibiotic research and development (Tacconelli et al., 2018). In addition to clarithromycin, resistance to tetracyclines, fluoroquinolones and other antibiotics in H. pylori has increased worldwide (Suzuki et al., 2019; Tshibangu-Kabamba & Yamaoka, 2021). Horizontal gene transfer between H. pylori and other prokaryotes has also been observed, indicating the ease by which this organism can acquire and distribute resistance to new therapeutics (Suerbaum et al., 1998). To combat this resistance, it is paramount to develop novel strategies to fight infection.

As the first bacterium to have two genomes sequenced, comparative studies of H. pylori revealed the gene clusters responsible for the biosynthesis of pyrimidines, a class of molecules that are essential for the generation of biochemically important nucleotides (Liechti & Goldberg, 2012; Alm & Trust, 1999; Doig et al., 1999). While the de novo pyrimidine-biosynthesis pathway is the predominant method for pyrimidine generation, an alternate `pyrimidine salvage' pathway exists that uses exogenous pyrimidines extracted from the environment to generate nucleotides. Interestingly, the biosynthetic gene cluster for the salvage pathway in H. pylori lacks several key enzymes, making this pathogen entirely reliant on de novo pyrimidine biosynthesis for growth and survival (Copeland et al., 2000). It follows that inhibiting the enzymes in the de novo pathway would entirely cut off the production of pyrimidines.

Dihydroorotate dehydrogenase (DHODH) catalyses the first committed step in pyrimidine biosynthesis and has been extensively studied as a druggable target in H. pylori, among other organisms, due to its critical role at the bottleneck of the pathway (Barnes et al., 1993; Okesli et al., 2017). DHODH enzymes are classified into two classes: class 1 and 2 (Nørager et al., 2002; Björnberg et al., 1997). Class 1 DHODHs exist as multimers, are cytosolic and almost exclusively use an active-site cysteine as a base in catalysis (Björnberg et al., 1997; Sørensen & Dandanell, 2002; Nagy et al., 1992; Nielsen et al., 1996; Kahler et al., 1999). Class 2 DHODHs exist as monomers, use an active-site serine as the base in catalysis and are bound to the inner cell membrane in Gram-negative bacteria (Taylor & Taylor, 1964; Chen & Jones, 1976; Rawls et al., 2000). These enzymes also utilize a ubiquinone (coenzyme Q) cofactor as the terminal electron acceptor. This ubiquinone is essential to regenerate the oxidized FMN and allow catalytic turnover. Encoded by the pyrD gene, H. pylori DHODH (HpDHODH) is a class 2 DHODH and uses coenzyme Q₆ as its terminal electron acceptor (Copeland et al., 2000).

Class 2 DHODHs use a bi-substrate ping-pong mechanism of catalysis, in which the transformation of L-dihydroorotate to orotate must be followed by the reduction of ubiquinone to achieve catalytic turnover (Figs. 1a and 1b; Hines & Johnston, 1989; Neidhardt et al., 1999). Firstly, L-dihydroorotate is oxidized to orotate with the concomitant reduction of FMN to FMNH₂. This process uses a proton-relay network composed of ordered water molecules and an active-site threonine to deprotonate the catalytic serine to the basic alkoxide form, which then deprotonates the C5 position of L-dihydroorotate (Kow et al., 2009; Fagan & Palfey, 2009; Fagan et al., 2006; Alves et al., 2015). In combination with the FMN-dependent oxidation, this process yields the orotate product. The reduced flavin, FMNH₂, is finally oxidized back to the FMN form through the action of the second substrate, ubiquinone (Nonato et al., 2019).

Figure 1 Biochemistry of class 2 dihydroorotate dehydrogenases. (a) Reaction scheme. (b) Important co-substrates.

Although researchers have empirically studied HpDHODH as a drug target, the development of small-molecule inhibitors has been limited by the absence of a crystal structure (Copeland et al., 2000; Marcinkeviciene et al., 2000; Haque et al., 2002; Björnberg et al., 1999; Kawada et al., 2013; Abe et al., 2013). We aimed to investigate how the structure of HpDHODH facilitates catalysis in order to better inform such drug-discovery efforts. Here, we determined the crystal structure of HpDHODH at 2.25 Å resolution. We performed a large-scale bioinformatic search to find evolutionary homology across class 2 DHODH proteins. We then evaluated structural and sequence conservation and structurally characterized the important facets of the enzyme that mediate redox chemistry. In doing so, we identified key interactions in the FMN binding cavity that determine the spatial positioning of FMN in the active site. Most notably, we discovered a channel that we hypothesize facilitates the diffusion of ubiquinone to the active site for regenerative catalysis.

2. Materials and methods

2.1. Macromolecule production

Cloning, expression and purification were conducted as part of the Seattle Structural Genomics Center for Infectious Disease (SSGCID; Myler et al., 2009; Stacy et al., 2011) following standard protocols described previously (Bryan et al., 2011; Choi et al., 2011; Serbzhinskiy et al., 2015).

Genomic DNA from H. pylori G27 was kindly provided by Dr Nina Salama, Fred Hutchinson Cancer Center. The full-length protein (UniProt B5Z6I2) encoding amino acids 1–351 was PCR-amplified from genomic DNA using the primers shown in Table 1. The gene was cloned into the ligation-independent cloning (LIC; Aslanidis & de Jong, 1990) expression vector pBG1861 (Choi et al., 2011) encoding a noncleavable 6×His fusion tag (MAHHHHHHM-ORF). Plasmid DNA was transformed into lambda phosphatase chemically competent E. coli BL21(DE3) R3 Rosetta cells. The cells were expression-tested, and 2 l of culture was grown using auto-induction medium (Studier, 2005) in a LEX Bio­reactor (Epiphyte Three Inc.) as described previously (Choi et al., 2011). The expression clone was assigned the SSGCID target identifier HepyC.00487.a.B1.GE40934 and is available at https://www.ssgcid.org/available-materials/expression-clones/.

Table 1
Macromolecule-production information

Source organism	Helicobacter pylori strain G27
DNA source	Dr Nina Salama, Fred Hutchinson Cancer Center
Forward primer	5′-CTCACCACCACCACCACCATATGCTTTATCCATTAGTAAAAAAATATC-3′
Reverse primer	5′-ATCCTATCTTACTCACTTATCTTAAATCCGCTCCTATAGCCTC-3′
Expression vector	pBG1861
Expression host	E. coli BL21(DE3) R3 Rosetta
Complete amino-acid sequence of the construct produced	MAHHHHHHMLYPLVKKYLFSLDAEDAHEKVCKILRTLSKSSFLCSLIHSQWGYKNPKLENEILGLNFPNPLGLAAGFDKNASMLRALIAFGFGYLEAGTLTNEAQVGNERPRLFRHIEEESLQNAMGFNNYGAVLGARSFNRFAPYKTPIGINLGKNKHIEQAHALEDYKAVLNQCLNIGDYYTFNLSSPNTPNLRDLQNKAFVNELFCMAKEMTHKPLFLKIAPDLEIDDMLEIVNSAIEAGAHGIIATNTTIDKSLVFAPKEMGGLSGKCLTKKSREVFKELAKAFFNKSVLVSVGGISDAKEAYERIKMGASLLQIYSAFIYNGPNLCQNILKDLVKLLQKDGFLSVKEAIGADLR

The protein was purified in a three-step protocol consisting of an Ni²⁺-affinity chromatography (IMAC) step, passage over a second Ni²⁺-affinity column and size-exclusion chromatography (SEC). All chromatography runs were performed on an ÄKTApurifier 10 (GE) using automated IMAC and SEC programs according to previously described procedures (Bryan et al., 2011). The final SEC was performed on a HiLoad 26/60 Superdex 75 column (GE Healthcare) using a mobile phase consisting of 25 mM HEPES pH 7.0, 500 mM NaCl, 5% glycerol, 2 mM DTT, 0.025% azide. Peak fractions were eluted as two smooth peaks that were suspected to be the monomer and the dimer. Peak fractions from the second peak (corresponding to the suspected monomer) were pooled and analysed for the presence of the protein of interest using SDS–PAGE. The peak fractions were concentrated to 18.5 mg ml⁻¹ using an Amicon purification system (Millipore). Aliquots of 200 µl were flash-frozen in liquid nitrogen and stored at −80°C until use for crystallization.

2.2. Crystallization

Purified His-HpDHODH was screened for crystallization in 96-well sitting-drop plates against the JCSG++ HTS (Jena Bioscience), MCSG1 (Molecular Dimensions) and Morpheus (Newman et al., 2005; Gorrec, 2009) crystal screens. Equal volumes (0.4 µl) of protein solution at 18.5 mg ml⁻¹ and precipitant solution were mixed and set up at 290 K against reservoir (80 µl) in sitting-drop vapour-diffusion format. The final crystallization precipitant was MCSG condition D3 (see Table 2). The crystals were harvested and flash-cooled by plunging them into liquid nitrogen.

Table 2
Crystallization

Method	Vapour diffusion, sitting drop
Plate type	96-well Compact 300, Rigaku
Temperature (K)	290
Protein concentration (mg ml^−1)	18.5
Buffer composition of protein solution	20 mM HEPES pH 7.0, 300 mM NaCl, 5% glycerol, 1 mM TCEP
Composition of reservoir solution	MCSG1 condition D3: 0.2 M magnesium chloride, 0.1 M Tris–HCl pH 8.5, 30%(v/v) PEG 400
Volume and ratio of drop	0.4 µl protein solution + 0.4 µl reservoir solution
Volume of reservoir (µl)	80

2.3. Data collection and processing

Data were collected at 100 K on beamline 21-ID-F at the Advanced Photon Source (APS), Argonne National Laboratory (see Table 3). Data sets were reduced with XSCALE (Kabsch, 2010). Raw X-ray diffraction images are available from the Integrated Resource for Reproducibility in Macromolecular Crystallography at https://www.proteindiffraction.org/.

Table 3
Data collection and processing

Values in parentheses are for the outer shell.

Diffraction source	Beamline 21-ID-F, APS
Wavelength (Å)	0.97872
Temperature (K)	100
Detector	MAR Mosaic 300 mm CCD
Crystal-to-detector distance (mm)	300
Rotation range per image (°)	1.0
Total rotation range (°)	180
Exposure time per image (s)	1.0
Space group	P21
a, b, c (Å)	49.44, 69.08, 100.10
α, β, γ (°)	90, 98.06, 90
Resolution range (Å)	48.952–2.250 (2.310–2.250)
Total No. of reflections	120853
No. of unique reflections	31777
Completeness (%)	99.700 (100.000)
Multiplicity	3.803 (3.833)
〈I/σ(I)〉	17.870
Rr.i.m.	0.051 (0.654)
Overall B factor from Wilson plot (Å^2)	47.590

2.4. Structure solution and refinement

The structure was solved by molecular replacement with Phaser (McCoy et al., 2007) from the CCP4 suite of programs (Winn et al., 2011; Agirre et al., 2023) using PDB entry 1f76 (Nørager et al., 2002) as the search model. The structure was refined using iterative cycles of Phenix (Liebschner et al., 2019) followed by manual rebuilding of the structure using Coot (Emsley et al., 2010). The quality of the structure was checked using MolProbity (Chen et al., 2010). All data-reduction and refinement statistics are shown in Table 4. The structure of HpDHODH was refined to 2.25 Å resolution. Structural figures were prepared using PyMOL (DeLano, 2002). Coordinates and structure factors have been deposited in the Protein Data Bank (https://www.rcsb.org/; Berman et al., 2000) with accession code 6b8s.

Table 4
Structure solution and refinement

Values in parentheses are for the outer shell.

Resolution range (Å)	48.952–2.250 (2.306–2.250)
Completeness (%)	99.7
σ Cutoff	F > 1.360σ(F)
No. of reflections, working set	31767 (2121)
No. of reflections, test set	1983 (125)
Final Rcryst	0.182 (0.2690)
Final Rfree	0.239 (0.3451)
No. of non-H atoms
Protein	4956
Ligand	82
Solvent	118
Total	5156
Average B factors (Å^2)
Protein	58.3
Ligand	44.7
Water	50.1
Ramachandran plot
Most favoured (%)	97
Allowed (%)	3

3. Results and discussion

3.1. Overall structure of H. pylori DHODH bound to FMN

We determined the structure of dihydroorotate dehydro­genase from H. pylori (HpDHODH; PDB entry 6b8s) at 2.25 Å resolution. The asymmetric unit of the crystal is composed of two identical chains with a total molecular weight of 81.34 kDa comprising 718 amino-acid residues, 671 of which were modelled. We determined that the biological assembly is monomeric via gel filtration (data not shown) and found that it contains 351 residues with a bound FMN cofactor (Fig. 2a). Several residues throughout the structure are only partially modelled due to a lack of electron density (see Supplementary Table S1). Residues 180–191 in chain A and 180–190 in chain B are unmodelled due to insufficient electron density. This lack of electron density indicates that this region was disordered in the crystal, suggesting that these residues form a flexible loop, in line with previously documented findings (Björnberg et al., 1999). This loop also contains the active-site residues Ser181 and Thr184 that are thought to be responsible for catalysing the deprotonation of L-dihydroorotate (Björnberg et al., 1999).

Figure 2 Crystal structure of dihydroorotate dehydrogenase from H. pylori with bound flavin mononucleotide. (a) A cartoon diagram of the biological assembly of H. pylori dihydroorotate dehydrogenase (PDB entry 6b8s chain A, dark green). Flavin mononucleotide (FMN) is bound in the active site (yellow ball-and-stick representation). (b) The α/β TIM-like barrel domain is highlighted with α-helices in burgundy and β-strands in dark green. The N-terminal microdomain is in blue. The view is rotated 90° from (a) to (b).

The majority of the HpDHODH structure is a triose­phosphate isomerase (TIM)-like α/β-barrel with eight α-helices and eight β-strands, determined by analysis via the CATH database (Fig. 2b; Lewis et al., 2018; Sillitoe et al., 2021). This α/β-barrel is capped on either end by two β-sheets and includes two shorter α-helices toward the C-terminus. Both features are absent in TIM, the domain's namesake (Banner et al., 1976). The arrangement of α-helices around the α/β-barrel is disrupted on one side, forming what Orozco Rodriguez et al. (2022) first described as an N-terminal `microdomain' of α-helices spanning residues 1–44. They report two α-helices in both E. coli DHODH (EcDHODH) and human DHODH (HsDHODH), indicating one helix as bent (Orozco Rodriguez et al., 2022). Here, we identified three α-helices in the N-terminal microdomain of HpDHODH.

3.2. Comparison of DHODH from H. pylori with structural homologs from other organisms

To identify patterns of structural conservation critical for the function of HpDHODH, we performed a structure and sequence-based alignment using the DALI Protein Structure Comparison Server (Holm, 2020) to determine the closest evolutionary homologs of HpDHODH (PDB entry 6b8s chain A). To conduct an adequate evolutionary analysis, we picked class 2 DHODH homologs belonging to E. coli (EcDHODH, PDB entry 1f76 chain A), a Gram-negative, filamentous bacterium commonly found in the intestine of warm-blooded animals (Sondi & Salopek-Sondi, 2004), Mycobacterium tuberculosis (MtDHODH; PDB entry 4xq6 chain B), a weakly Gram-positive bacterium that is the causative agent of tuberculosis in humans (Shiloh & Champion, 2010), Plasmodium falciparum (PfDHODH; PDB entry 5boo chain A), a deadly blood parasite that causes malaria in humans (Lee et al., 2012), and finally the enzyme from Homo sapiens (HsDHODH; PDB entry 2prm chain A; Walse et al., 2008). The DALI structural alignment yielded root-mean-square deviation (r.m.s.d.) and percentage identity values of 1.5 Å and 37% for EcDHODH, 1.5 Å and 38% for MtDHODH, 1.6 Å and 31% for PfDHODH and 1.7 Å and 35% for HsDHODH, respectively. The alignment further corroborates our finding that the structure of class 2 DHODH is highly conserved (Fig. 3a). Notably, the core β-strands of the TIM-like α,β-barrel domain are better conserved than the outer α-helices. This finding is significant because FMN is bound in this core structure (Fig. 3b). Further, we found that the position of FMN is highly conserved across the studied homologs (Supplementary Fig. S1). Together, these data indicate that the FMN binding cavity in class 2 DHODH is optimized to tightly hold the coenzyme in a specific position.

Figure 3 Structural and sequence alignments of class 2 dihydroorotate dehydrogenases show high conservation in the active site. (a) Dihydroorotate dehydro­genase (DHODH) from H. pylori (PDB entry 6b8s chain A) is depicted in light green, that from E. coli (PDB entry 1f76 chain A) is depicted in light pink, that from M. tuberculosis (PDB entry 4xq6 chain B) is depicted in light blue, that from P. falciparum (PDB entry 5boo chain A) is depicted in red and that from H. sapiens (PDB entry 2prm chain A) is depicted in dark blue. Flavin mononucleotide (FMN) is shown (yellow ball-and-stick representation) as an overlay from H. pylori DHODH. (b) Class 2 DHODHs have a well conserved FMN binding region. The view is rotated 90° from (a) to (b). (c) The ConSurf alignment shows strong conservation near the FMN binding site. ConSurf analysis was carried out with default parameters using PDB entry 6b8s chain A. Flavin mononucleotide (FMN) is shown (yellow ball-and-stick representation). The view is rotated 90° along a diagonal from (b) to (c).

Next, we sought to identify whether there was strong sequence conservation in addition to structural conservation. To answer this question, we used the ConSurf server to evaluate such homology across a larger sample of DHODH enzymes. ConSurf mapped the sequence conservation onto the protein structure (Fig. 3c; Celniker et al., 2013). The most conserved region of the entire protein is the active site, where FMN is bound, while the most variable regions occur towards the exterior of the protein, away from the FMN binding pocket. In summary, these data demonstrate very strong structural and sequence-based homology among class 2 DHODH enzymes, indicating that HpDHODH is likely to act by a mechanism that is similar, if not identical, to its nearest homologs.

3.3. Ligand–protein interactions between HpDHODH and flavin mononucleotide in the active site

Understanding the strong conservation of the active site, we were interested in examining the noncovalent interactions that facilitate the precise positioning of FMN in the enzyme. We visualized all of the residues within 3.5 Å of FMN with the capacity to form stabilizing interactions (Fig. 4). FMN is stabilized in a cavity at the centre of the HpDHODH protein by hydrogen-bonding interactions with neighbouring amino acids and three ordered water molecules. The bulk of these hydrogen bonds are to the terminal phosphate and the iso­alloxazine tricycle. The O1 of the phosphate group of FMN accepts hydrogen bonds from two water molecules and the Tyr312 amide N atom. The O3 atom of the phosphate group accepts hydrogen bonds from the amide N atoms of Gly291 and Gly262. The O2 atom of the phosphate group accepts one hydrogen bond from the amide and one hydrogen bond from the β-hydroxyl O atom of Ser313.

Figure 4 Intermolecular noncovalent interactions between dihydroorotate dehydrogenase from H. pylori and flavin mononucleotide (ligand) bound in its active site. Noncovalent interactions between flavin mononucleotide (FMN, yellow) and dihydroorotate dehydrogenase (DHODH) from H. pylori (PDB entry 6b8s chain A, dark green) are depicted as magenta dashed lines. Interacting water molecules are shown as red spheres. Active-site residues are labelled using their one-letter codes.

The O4 carbonyl atom of the isoalloxazine accepts hydrogen bonds from the side-chain amine of Lys71, the amide N atom of Thr91 and the β-hydroxyl O atom of Thr91. N3 of the flavin interacts via hydrogen bonding with Thr91 through the β-hydroxyl O atom and an ordered water molecule. The O2 carbonyl atom of the isoalloxazine accepts hydrogen bonds from the side-chain amide N atom of Asn145 and Lys214. Lys214 also forms hydrogen bonds to O2′ and O3′ of the FMN tail. The carbonyl O atom of Thr242 forms an addition hydrogen bond to O3′ of FMN. The carbonyl O atom of Ala67 forms a second hydrogen bond to O2′ and to N10 on ring II of the flavin. In addition to these hydrogen-bonding interactions, Tyr312 is optimally positioned to interact with ring I of the flavin via staggered π–π stacking. These interactions are pivotal in binding FMN tightly in the cavity of HpDHODH, but do not indicate whether they are important across all class 2 DHODHs.

We then returned to our ConSurf analysis, which demonstrated that Gly291, Gly290, Lys71, Asn153, Lys214 and Ala67 are all highly conserved, reaching the maximum ConSurf score of 9, while Tyr312, Ser313 and Thr91 are also highly conserved, with a score of 8 (Fig. 3c and Supplementary Fig. S3; Celniker et al., 2013). All of the amino acids involved in FMN binding are highly evolutionarily conserved. Together, these data indicate that this manner of stabilizing FMN in the active site is highly important for the function of this enzyme.

3.4. The N-terminal microdomain allows H. pylori DHODH to associate with the membrane and channel ubiquinone to the active site

To achieve turnover, HpDHODH must recycle its FMN coenzyme after every round of catalysis. HpDHODH uses ubiquinone to perform this oxidation, converting FMNH₂ back to FMN (Copeland et al., 2000; Munier-Lehmann et al., 2013; Liu et al., 2000). We wanted to investigate how HpDHODH facilitates the entry of ubiquinone to the active site, given that the primary pool of free ubiquinone in the cell is in its inner membrane (Ragan & Cottingham, 1985; Hauss et al., 2005; Yu & Yu, 1981; Samorì et al., 1992). As such, we reasoned that for H. pylori to access the ubiquinone pool, it must be somehow associated with a ubiquinone-containing membrane. Indeed, all class 2 DHODH enzymes are known to be membrane-bound (Orozco Rodriguez et al., 2022; Munier-Lehmann et al., 2013).

To better understand how this may occur, we compared HpDHODH with EcDHODH, as this homolog is known to be associated with the inner membrane of the Gram-negative E. coli through its N-terminal microdomain (Taylor & Taylor, 1964; Orozco Rodriguez et al., 2022; Nørager et al., 2002). Others have also shown that HsDHODH binds ubiquinone via this N-terminal microdomain (Liu et al., 2000). These results pointed to the N-terminal microdomain as the starting place for our search for a ubiquinone binding site.

Orozco Rodriguez et al. (2022) found that the positively charged residues on the α-helices that make up this N-terminal microdomain are important in associating with the negatively charged phospholipid head groups of the membrane. To investigate whether this may also be the case for HpDHODH, we used the Adaptive Poisson–Boltzmann Solver (APBS) plugin in PyMOL to generate electrostatics maps of both HpDHODH and EcDHODH (Fig. 5a; Jurrus et al., 2018). The resulting surface plot shows that in both enzymes the N-terminal microdomain is characterized by electropositivity and is rich in charged arginine and lysine residues. These data indicate that HpDHODH has the potential to be associated with the membrane via electrostatic interactions between its N-terminal microdomain and the phosphate head groups of the membrane.

Figure 5 The N-terminal microdomain forms the membrane–protein interface and a ubiquinone channel. (a) Electrostatic surface plots of dihydroorotate dehydrogenase (DHODH) from H. pylori (PDB entry 6b8s chain A) and E. coli (PDB entry 1f76 chain A), viewing the N-terminal microdomain. Important residues are specified by arrows and labelled using one-letter codes. The colour blue represents high electropositivity and red represents high electronegativity in units of kBT/ec. (b) Hydrophobic amino acids (orange) line the centre of the N-terminal microdomain (blue), with charged residues (purple) at the solvent-accessible border; all residues are labelled using their one-letter codes. Background DHODH protein is coloured dark green. (c) Ubiquinone channel connecting the membrane to the flavin mononucleotide (FMN) binding site. View down the channel (cream colour) to FMN (yellow sticks). Background DHODH protein is coloured dark green. (d) The left and right panels show the protein interior, highlighting the residues that form the ubiquinone channel (cream colour). Hydrophobic amino acids (orange) and charged amino acids (purple) are labelled using one-letter codes. FMN (yellow sticks) is shown in the FMN binding pocket that is connected to the ubiquinone channel. Background DHODH protein is coloured dark green.

We next asked whether this membrane association was solely electrostatic or whether HpDHODH could insert itself into the membrane. To answer this question, we first used DeepTMHMM to predict potential transmembrane elements, but found that none were predicted by the algorithm (Hallgren et al., 2022). We then used the DREAMM algorithm to predict individual amino acids that may be inserted into the membrane (Chatzigoulas & Cournia, 2022a,b). DREAMM predicted Leu2, Leu5, Lys8, Tyr9, Phe34 and Leu38 to insert into the membrane. We then visualized these residues on the N-terminal microdomain (Fig. 5b). The side chains of these amino acids are, for the most part, facing away from the rest of the protein, towards what would be the membrane. These data suggest that HpDHODH may act as a peripheral membrane protein by inserting hydrophobic amino acids near the N-terminus into the lipid core of the membrane, in addition to electrostatic interactions.

Next, we sought to understand the features of the N-terminal microdomain of HpDHODH that may facilitate ubiquinone binding. We observed a channel from the N-terminal microdomain to the core of the protein (Figs. 5c and 5d). Although this channel bends in multiple places, when viewed down the channel, as in Fig. 5(c), the FMN cofactor is just barely visible. Near the channel entrance on the N-terminal microdomain, we observed that the residues with side chains facing the entrance were almost exclusively hydrophobic (Fig. 5b). Thus, we asked whether the rest of the channel might be lined with hydrophobic amino acids to facilitate the passage of ubiquinone to the active site. We found that this channel is almost entirely built from hydrophobic residues, with polar or charged residues only appearing towards the exit into the FMN binding pocket (Fig. 5d). It is possible that this arrangement of amino acids may match the structure of ubiquinone as it enters the channel. The polar dimethoxy­benzoquinone head group of ubiquinone is the redox-reactive region and would be present in this polar region of the channel, while the amphipathic isoprenoid tail would be present in the hydrophobic portion of the channel and may potentially even extend into the membrane itself. These data suggest that this channel may serve as the passageway through which ubiquinone might access the active site from its pool in the membrane. This finding provides a structural explanation for the competitive inhibition between ubiquinone and the pyrazole-based compounds described by Copeland et al. (2000). These compounds are particularly promising, as they retain a >1000-fold greater potency for HpDHODH over HsDHODH. We propose that they block access of ubiquinone to its dedicated channel in a similar manner to macrolide antibiotics blocking the peptide-exit channel in the ribosome. Taken together, these data provide strong evidence for HpDHODH as a membrane-associated enzyme that provides a direct connection between the ubiquinone pool in the membrane and a reduced flavin (FMNH₂) in need of oxidation to reset the enzyme for another round of catalysis.

4. Conclusion

In summary, we determined the crystal structure of HpDHODH and characterized its structure in terms of homology and function. We found extremely high conservation in the binding of FMN in the active site and discovered a hydrophobic channel that facilitates the entry of ubiquinone to the active site of the enzyme to regenerate the oxidized FMN coenzyme. We hope that these findings will assist in the development of new small molecules to treat H. pylori infections.

Footnotes

‡These authors contributed equally.