Phytophthora infestans Dihydroorotate Dehydrogenase Is a Potential Target for Chemical Control – A Comparison With the Enzyme From Solanum tuberosum

The oomycete Phytophthora infestans is the causal agent of tomato and potato late blight, a disease that causes tremendous economic losses in the production of solanaceous crops. The similarities between oomycetes and the apicomplexa led us to hypothesize that dihydroorotate dehydrogenase (DHODH), the enzyme catalyzing the fourth step in pyrimidine biosynthetic pathway, and a validated drug target in treatment of malaria, could be a potential target for controlling P. infestans growth. In eukaryotes, class 2 DHODHs are mitochondrially associated ubiquinone-linked enzymes that catalyze the fourth, and only redox step of de novo pyrimidine biosynthesis. We characterized the enzymes from both the pathogen and a host, Solanum tuberosum. Plant DHODHs are known to be class 2 enzymes. Sequence analysis suggested that the pathogen enzyme (PiDHODHs) also belongs to this class. We confirmed the mitochondrial localization of GFP-PiDHODH showing colocalization with mCherry-labeled ATPase in a transgenic pathogen. N-terminally truncated versions of the two DHODHs were overproduced in E. coli, purified, and kinetically characterized. StDHODH exhibited a apparent specific activity of 41 ± 1 μmol min-1 mg-1, a kcatapp of 30 ± 1 s-1, and a Kmapp of 20 ± 1 μM for L-dihydroorotate, and a Kmapp= 30 ± 3 μM for decylubiquinone (Qd). PiDHODH exhibited an apparent specific activity of 104 ± 1 μmol min-1 mg-1, a kcatapp of 75 ± 1 s-1, and a Kmapp of 57 ± 3 μM for L-dihydroorotate, and a Kmapp of 15 ± 1 μM for Qd. The two enzymes exhibited different activities with different quinones and napthoquinone derivatives, and different sensitivities to compounds known to cause inhibition of DHODHs from other organisms. The IC50 for A77 1726, a nanomolar inhibitor of human DHODH, was 2.9 ± 0.6 mM for StDHODH, and 79 ± 1 μM for PiDHODH. In vivo, 0.5 mM A77 1726 decreased mycelial growth by approximately 50%, after 92 h. Collectively, our findings suggest that the PiDHODH could be a target for selective inhibitors and we provide a biochemical background for the development of compounds that could be helpful for the control of the pathogen, opening the way to protein crystallization.

The oomycete Phytophthora infestans is the causal agent of tomato and potato late blight, a disease that causes tremendous economic losses in the production of solanaceous crops. The similarities between oomycetes and the apicomplexa led us to hypothesize that dihydroorotate dehydrogenase (DHODH), the enzyme catalyzing the fourth step in pyrimidine biosynthetic pathway, and a validated drug target in treatment of malaria, could be a potential target for controlling P. infestans growth. In eukaryotes, class 2 DHODHs are mitochondrially associated ubiquinone-linked enzymes that catalyze the fourth, and only redox step of de novo pyrimidine biosynthesis. We characterized the enzymes from both the pathogen and a host, Solanum tuberosum. Plant DHODHs are known to be class 2 enzymes. Sequence analysis suggested that the pathogen enzyme (PiDHODHs) also belongs to this class. We confirmed the mitochondrial localization of GFP-PiDHODH showing colocalization with mCherrylabeled ATPase in a transgenic pathogen. N-terminally truncated versions of the two DHODHs were overproduced in E. coli, purified, and kinetically characterized. StDHODH exhibited a apparent specific activity of 41 ± 1 µmol min −1 mg −1 , a k cat app of 30 ± 1 s −1 , and a K m app of 20 ± 1 µM for L-dihydroorotate, and a K m app = 30 ± 3 µM for decylubiquinone (Qd). PiDHODH exhibited an apparent specific activity of 104 ± 1 µmol min −1 mg −1 , a k cat app of 75 ± 1 s −1 , and a K m app of 57 ± 3 µM for L-dihydroorotate, and a K m app of 15 ± 1 µM for Qd. The two enzymes exhibited different activities with different quinones and napthoquinone derivatives, and different sensitivities to compounds known to cause inhibition of DHODHs from other organisms. The IC 50 for A77 1726, a nanomolar inhibitor of human DHODH, was 2.9 ± 0.6 mM for StDHODH, and 79 ± 1 µM for PiDHODH. In vivo, 0.5 mM A77 1726 decreased

INTRODUCTION
Within the Phylum Oomycota, the species of the genus Phytophthora are all considered devastating pathogens of crops and landscape plants, and are responsible annually for huge economic losses worldwide (Lamour et al., 2007;Attard et al., 2008). Phytophthora infestans causes late blight disease in potato, tomato, and other solanaceous crops. In these hosts, the entire plant is destroyed within a few days after the first lesions are observed (Fry, 2008). Despite their economic importance, Phytophthora species remain poorly characterized at the biochemical level. The enzyme dihydroorotate dehydrogenase (DHODH, E.C. 1.3.5.2), catalyzing the fourth step in de novo pyrimidine biosynthesis has been attractive for drug development for decades, with over 100 compounds that inhibit its activity in diverse organisms (Munier-Lehmann et al., 2013). In most eukaryotes, including plants (Witz et al., 2012) and some plant pathogenic fungi (Zameitat et al., 2007), class 2 DHODHs are associated with the outer surface of the inner mitochondrial membrane (Rawls et al., 2000) and transfer electrons from dihydroorotate oxidation to ubiquinone in the respiratory chain. Class 2 DHODHs are also found in gramnegative bacteria (Björnberg et al., 1999), anchored to the inner side of the periplasmic side of the inner cytoplasmic membrane, similarly transferring electrons to the respiratory chain.
Human DHODH, has been intensively studied, and is the target of A77 1726, also known as teriflunomide, the active metabolite of leflunomide, used in the treatment of rheumatoid arthritis (Fragoso and Brooks, 2015), and multiple sclerosis (Faissner and Gold, 2018). Recent work suggests that inhibitors of this enzyme show promise in the treatment of different cancers (Sykes et al., 2016). Structural differences between the human enzyme and pathogen enzymes have been exploited to develop species-specific inhibitors. For example, DSM265, a triazolopyrimidine-based compound targeting DHODH from the apicomplexan parasite Plasmodium falciparum, is currently undergoing phase 2 clinical trials (Ashley, 2017;McCarthy et al., 2017). Similarly, the antifungal properties of F901318, an inhibitor of DHODHs from pathogenic Aspergillus sp., is being tested in phase 1 trials (Oliver et al., 2016). Oomycete DHODH has been evaluated as an enzymatic target in biochemical screenings, leading to the identification of compounds that display in vitro inhibition of Pythium aphanidermatum DHODH, and in vivo whole plant control of Plasmopara viticola (Parker et al., 2002). Interestingly, gene deletion simulations in P. infestans by Rodenburg et al. (2018) identified the DHODH as one of 72 genes essential for growth, thus warranting further study.
Collectively, these observations suggest that selective inhibitors could be developed for P. infestans DHODH that would exert little or no effect on the host, or the human consumer. In the present study we produced recombinant pathogen (PiDHODH) and Solanum tuberosum (StDHODH) enzymes, measured their catalytic properties, and demonstrated differences in their activities with electron acceptors and in their sensitivities to inhibitors. We showed that PiDHODH has a micromolar IC 50 for A77 1726, that is ≈37-fold lower than that of the host enzyme, and demonstrated that this compound inhibits growth of Phytophthora in vivo. Our work highlights the potential of PiDHODH as a possible target for developing novel combination strategies for pathogen control.
Frontiers in Microbiology | www.frontiersin.org number PRJNA17665, and S. tuberosum, EnsemblPlants 1 . Additional sequences were retrieved from National Center for Biotechnology Information (NCBI). Multiple sequence alignments were performed according to sequence and 3Dstructure using the STRAP program (default parameters) 2 . In Figure 1, the amino acid sequences of the secondary structure elements in three proteins with crystallographic structures are identified as specified in the relevant publications as follows: E. coli DHODH (Nørager et al., 2002), P. falciparum DHODH (Deng et al., 2009), human DHODH (Liu et al., 2000). Different designations for the alpha helices and beta sheets are used in these three publications; in Figure 1 we use the designations of Liu and coworkers. Mitochondrial targeting and signal peptide sequences were predicted with Mitoprot II 1.101 (default parameters) (Claros and Vincens, 1996), and TargetP 1.1 (plant network) (Nielsen et al., 1997;Emanuelsson et al., 2000), and MitoFates (plant) (Fukasawa et al., 2015). The N-terminal transmembrane domains were predicted by HMMTOP (default parameters) (Tusnady and Simon, 2001).

Expression Constructs
Solanum tuberosum cDNA was prepared from commercially available potato plants (Sabanera variety). Total RNA was extracted from frozen mycelia or from 300 mg of plant leaves and cDNA was synthesized (Garcia-Bayona et al., 2014). Subsequently, a fraction of the cDNA reactions (2 µL) was used for PCR amplification. cDNA prepared from P. infestans strain 1043, a Colombian isolate from potato with A1 mating type, and also belonging to the EC-1 clonal lineage, was used for the cloning (Vargas et al., 2009). The isolates were cultured routinely on rye agar medium and incubated at 19 • C for 9 days in the dark (Goodwin et al., 1998). Mycelia were collected from two plates and frozen until use.

Expression and Purification
Transformed E. coli BL21-CodonPlus(DE3)-RP electrocompetent cells were precultured overnight in Luria-Bretani broth medium (LB) with 100 µg/mL ampicillin as the selection marker at 37 • C with 200 rpm agitation. The preculture was diluted to 5% in LB containing 100 µg/mL ampicillin and grown at 37 • C until OD 600nm = 0.5-0.6. Induction was performed with 1 mM isopropyl β-D-thiogalactopyranoside (IPTG), supplemented with 0.1 mM flavin mononucleotide (FMN) and growth was continued overnight at room temperature (Baldwin et al., 2002). Induced cultures were harvested by centrifugation at 3,500 × g for 15 min at 4 • C, supernatant discarded and cell pastes were frozen at −80 • C until use.
For the purification of the recombinant proteins, a pellet from 500 mL of bacterial cell culture was resuspended in Buffer A [50 mM Tris-HCl pH 8, 300 mM NaCl and 10% glycerol, 5 mM imidazole, 1 mM phenylmethanesulfonyl fluoride (PMSF) and 1 mM benzamidine] containing 2% Triton X-100. The cell suspension was incubated with lysozyme (1 mg/mL) on ice for 2 h. Cells were disrupted by sonication on ice (30 cycles of 20 s each, output control setting of 8 and 100% duty cycle) using a 250 Analog Sonifier (Branson). After sonication, insoluble and soluble fractions were separated by centrifugation (8,500 × g, 1 h, 4 • C). The recombinant proteins were purified from the clarified cell lysates using Co 2+ affinity columns (Thermo Scientific) equilibrated with Buffer A containing 0.5% Triton X-100 following the manufacturer's recommendations, and eluted with buffer A containing 250 mM imidazole. The yields of purified protein were 2.0 mg/L of cell culture for StDHODH, and 1.3 mg/L of cell culture for PiDHODH. The expected sizes for the truncated recombinant proteins were: PiDHODH 43.1 kDa and StDHODH 43.7 kDa.

SDS-PAGE and Electrotransfer
Protein samples were fractionated by SDS-PAGE on 12% running gels, with 5% stacking gels. Electrophoresis was performed in a BioRad Mini-Protean II electrophoresis cell for 1 h, at 200 volts, constant voltage. Gels were visualized by staining with Coomassie Blue G-250 dye.

Enzymatic Assays and Kinetic Analysis
In our standard activity assay we monitored the reduction of 2,6-dichlorophenol-indophenol (DCIP) at 600 nm (ε = 18,800 M −1 cm −1 ) at 30 • C in a reaction buffer containing 50 mM Tris-HCl pH 8.0, 150 mM KCl, 0.1% Triton X-100, 10% glycerol, 1 mM L-dihydroorotate, 0.1 mM DCIP, 0.1 mM decylubiquinone (Qd) (Zameitat et al., 2007), with concentrations of 19 nM (StDHODH) or 6.6 nM (PiDHODH). The apparent kinetic constants of the substrates were determined by varying L-dihydroorotate concentration (2.5-1,250 µM) while keeping Qd constant at 100 µM, or by varying Qd (1.25-200 µM) at a fixed dihydroorotate concentration of 1 mM. The kinetic data were evaluated by fitting the data to the Michaelis-Menten equation v = Vmax * [S]/ (Km + [S]) using GraphPad Prism v7 software. The apparent kcat was calculated from kcat = Vmax/[ET], where [ET] is total enzyme concentration, based on one active site monomer. Background oxidase activities due to direct reduction of DCIP by the enzyme in the absence of Q d were subtracted from the activities measured in the presence of Q d . For StDHODH the background oxidase activity was 9 ± 2 µmol min −1 mg −1 , and for PiDHODH it was 6 ± 1 µmol min −1 mg −1 .
To test alternative electron acceptors, stock solutions of 10-50 mM quinones, napthoquinones, and benzoquinones were prepared in absolute ethanol. Enzyme activities were measured with the standard assay described above, using 0.1 mM of the acceptor in place of Qd.
Stock solutions of all inhibitors were prepared in reaction buffer, or in dimethyl sulfoxide (DMSO). Activities were measured with the standard DCIP assay with saturating concentrations of Qd (0.1 mM) and dihydroorotate (1 mM), in the presence of 0.5 mM inhibitor. The final concentrations of 5% DMSO used in the assay solutions were found not to affect the DHODH activity of either recombinant enzyme. DHODH activities were measured by the standard assay.
Protein concentration was measured using the bicinchoninic assay (Pierce) with bovine serum albumin as the standard.
Absorbance was measured in microplates with a Thermo Scientific Multiskan GO UV/visible spectrophotometer.

P. infestans Localization Constructs, Transformation, Selection, and Imaging
To determine the subcellular localization of PiDHODH, we made a construct fused on the 3 -ends to eGFP. The desired sequences were amplified by PCR using primer pairs (Supplementary Table 1) which introduced a PacI site at the 5 end and an NheI site at the 3 end. The fragment was ligated to the pGEM-T Easy vector (Promega) and transformed in E. coli DH5-α chemiocompetent cells. The pGEM-T Easy plasmid was subjected to double restriction PacI-NheI, the fragment was subcloned in pGFPH (Ah-Fong and Judelson, 2011), and confirmed by sequencing. The resulting fusion construct was expressed under the control of the promoter and terminator from the Ham34 gene of Bremia lactucae. A control plasmid, pATPase-mCherryN, was used to visualize mitochondria (Ah-Fong and Judelson, 2011). Plasmid DNA for transformation was obtained using the Plasmid DNA Purification Kit NucleoBond PC 500 (Macherey-Nagel, Germany). Stable transformants were generated with isolate 1306 (A1 mating type, United States isolated from tomato) (Judelson et al., 1995), using the modifications proposed by Ah-Fong and Judelson (2011) to the original protoplast method (Judelson et al., 1991). For co-transformation, a total of 40 µg of plasmid DNA were used per experiment; 20 µg of the localization plasmid (PiDHODH-eGFP) and 20 µg of the control plasmid which expresses the b-subunit of the mitochondrial ATPase N-terminally fused to mCherry (pATPase-mCherryN). Transformed strains were grown with 30 µg mL −1 hygromycin and 10 µg mL −1 G418.
GFP-and mCherry-expressing strains were preselected by a visual inspection of the transformants by confocal laser scanning microscope using a Leica TCO-SP2. Transformants were grown in Petri dishes of rye-sucrose media for 6-8 days, mycelia from the external edge of the growth radius were mounted on drops of 10 µL of water in glass slides and sealed with transparent nail polish. GFP expression was visualized using 40X/0.8 or 63X/0.9 objectives with excitation/emission settings 488 nm/500-550 nm for GFP 543 nm/575-700 nm for mCherry. Fluorescent sporangia were obtained after 7-10 days.

Bioassays
The isolate Z3-2 of P. infestans, with EC-1 clonal lineage, and A1 mating type, was grown for 8 to 10 days at 18 • C on ryesucrose agar plates plus 50 µg/mL ampicillin. Plates containing sporulated mycelia were flooded with ultra pure water and rubbed with a glass rod to liberate the sporangia. The sporangia were purified from the suspension by filtration through a 50 µm nylon mesh, and counted with a hemocytometer. Then 4 × 10 6 sporangia/mL were incubated for 2-4 h at 10 • C to induce the liberation of free-swimming zoospores. The suspension containing zoospores and sporangia (125 µL) was dispensed into five replicate wells of a 96-well flat bottom tissue culture treated plates, and combined with 75 µL of Henniger media (Henniger, 2007) and 25 µL of each compound tested. A77 1726 was resuspended in DHODH reaction buffer, and curzate was resuspended in water. Microtiter plates were covered and placed in the dark at 18 • C. Growth was evaluated by measuring the OD 610nm at 20-24 h intervals for 100 h with a Multiskan TM GO Microplate Spectrophotometer (Thermo Scientific). Data analysis and graphs were performed using GraphPad Prism Software v7 for Mac 3 .

StDHODH and PiDHODH Predicted Amino Acid Sequences Are Most Similar to Class 2 Enzymes
The gene encoding the StDHODH was identified as a single copy protein (Gene ID PGSC0003DMG401016396) in the S. tuberosum genome assembly, EnsemblPlants 1 . It had an open reading frame (1275 bp) comprised of 11 exons, which encoded a 48.9 kDa protein (460 residues) with an isoelectric point of 9.35 consistent with the high pI predicted for other class 2 DHODHs. The derived amino acid sequence contained typical features of the class 2 DHODHs, including an N-terminal extension, a transmembrane domain, a conserved serine (S277) that aligned with the catalytic serines of E. coli (S175), human (S215), and P. falciparum (S345) DHODHs, and conserved residues for binding the electron acceptor, FMN, and orotate (Figure 1). The identity with other plant DHODHs was high (Zea mays 73%, A. thaliana 77%), and revealed that the enzyme was more similar to the class 2 DHODHs (Homo sapiens 49%, P. falciparum 30%, P. infestans 48%, E. coli 41%) (Supplementary Table 2) than to class 1 DHODHs (T. cruzi 21%, S. cerevisiae 21%). The N-terminal extension found in plant DHODHs was longer than the extensions found in the human or the oomycete enzymes, but shorter than that found in apicomplexan enzymes. A cleavage site at residue 28 was predicted by both TargetP 1.1 (Nielsen et al., 1997;Emanuelsson et al., 2000) and MitoFates (Fukasawa et al., 2015). The StDHODH displayed a small insertion of 13 residues between positions 225 and 236 that was also present in other plant DHODHs, but not in other eukaryotic class 2 enzymes. Four positions in our cloned sequence of StDHODH (P120T, I273V, I299V, and P359Q) differed from the published sequence shown in Figure 1. Probably these differences are not significant, since all of these residues are found in other sequences shown in the alignment, for example, threonine in position 120 is observed in PiDHODH, valine in position 273 is observed in HsDHODH, valine in position 299 appears in all the eukaryotic DHODHs in Figure 1, and glutamine in position 359 is found in HsDHODH. The last of these differences is in a poorly conserved region of the sequence. Furthermore, it is relevant that cultivated S. tuberosum is a highly heterogenous autotetraploid (Manrique-Carpintero et al., 2018), and thus possesses four alleles of DHODH, which may exhibit sequence differences.

PiDHODH Is Targeted to Mitochondria
The PiDHODH contains a predicted cationic mitochondrial targeting sequence (residues 1-23) at the N-terminus, followed by a predicted hydrophobic transmembrane helix sequence (residues 30-49). A proteolytic cleavage site was predicted at position 23 by MitoprotII v1.101. To confirm the mitochondrial localization we constructed an eGFP PiDHODH fusion, and cotransformed P. infestans with an mCherry tagged ATP synthase construct. PiDHOD was found to colocalize with the ATP synthase in sporangia (Figure 2).

N69StDHODH and N54PiDHODH Recombinant Proteins Are Produced in Soluble Form in E. coli
Class 2 DHODHs from several organisms have been produced as recombinant proteins, and in most cases these expression constructs eliminate the N-terminal sequence, which contains the hydrophobic transmembrane anchor. The elimination of these N-terminal extensions appear to exert relatively small effects on enzyme activity (Baldwin et al., 2002;Zameitat et al., 2006).
We expressed both full-length (data not shown) and N-terminally truncated recombinant proteins of StDHODH, with N-terminal polyhistidine-tags to facilitate purification. While the full-length StDHODH was found in the insoluble fraction, the truncated version, N69StDHODH, was expressed in soluble form, and was purified (Supplementary Figure 2). Interestingly, both full-length and truncated versions of A. thaliana DHODH were soluble and active, although the polyhistidine-tags in these recombinant proteins were located at the C-termini (Ullrich et al., 2002). PiDHODH was expressed using a similar strategy, and only the truncated version N54PiDHODH was soluble and could be purified (Supplementary Figure 2).

Apparent Kinetic Parameters of N69StDHODH and N54PiDHODH
Preliminary kinetic data were collected for the N-terminally truncated DHODHs (Figure 3), and the apparent kinetic parameters were compared to those of other DHODHs ( Table 1) (Ullrich et al., 2001;Hortua Triana et al., 2012). N69StDHODH exhibited an apparent specific activity of 41 ± 1 µmol min −1 mg −1 , a k cat app of 30 ± 1 s −1 , a K m app of 20 ± 1 µM for L-dihydroorotate, and a K m app of 30 ± 3 µM for decylubiquinone (Qd). N54PiDHODH exhibited an apparent specific activity of 104 ± 1 µmol min −1 mg −1 , a k cat app of 75 ± 1 s −1 , a K m app of 57 ± 3 µM for L-dihydroorotate, and a K m app of 15 ± 1 µM for Qd. We evaluated the enzyme activities with a variety of natural and artificial electron acceptors. Activities are expressed as a percentage, taking the activity measured with Qd as 100%.
N69StDHODH, exhibited the highest activities for coenzyme Qs having isoprenyl side chain lengths ranging from Q1 to Q9. N54PiDHODH showed a preference for a slightly shorter range of isoprenyl side chain lengths, from Q1 to Q6. The two enzymes exhibited different activities for six derivatives of napthoquinone ( Table 2).

StDHODH and PiDHODH Show Differential Sensitivities to Inhibitors
We studied the susceptibilities of the S. tuberosum and P. infestans recombinant enzymes to various compounds, which have been demonstrated to be inhibitors of DHODHs in other species Baldwin et al., 2002;Leban et al., 2005;Heikkilä et al., 2006;Davies et al., 2009;Munier-Lehmann et al., 2013). We also tested atovaquone (Munier-Lehmann et al., 2013) and amectoctradin (Dreinert et al., 2018), two inhibitors that bind to ubiquinone sites in cytochrome bc 1 . The first two compounds in Table 3 have similarity to the substrate (dihydroorotate), or to the product (orotate), and the remaining compounds are expected to bind in the electron acceptor (ubiquinone) site. Nine of the compounds decreased PiDHODH activity, and six decreased StDHODH activity to below 50% of the activity observed in absence of inhibitor ( Table 3). Differential inhibition for StDHODH and PiDHODH was observed for compounds A77 1726, toltrazuril, MD108, MD241, and NSC61890. The greatest difference in inhibition was observed for A77 1726, where PiDHODH activity was decreased to 14%, while StDHODH activity was unaffected. This difference was confirmed by the IC 50 values (Figure 4) measured for A77 1726, 2.9 ± 0.6 mM for StDHODH, and 79 ± 8 µM for PiDHODH.

A77 1726 Inhibits the Growth of P. infestans
We measured the growth of P. infestans sporangia and zoospores in liquid culture in the presence of PiDHODH inhibitor A77 1726 ( Figure 5). For comparison, growth was measured in the presence of Curzate R M8, an oomyceticide containing 64% w/w mancozeb [manganese zinc ethylenebis(dithiocarbamate)] and 8% w/w cymoxanil, with the former targeted to enzymes dependent on active sulfhydryl groups (Gullino et al., 2010), and the latter implicated in amino acid synthesis (Tellier et al., 2008). Of the two treatments, Curzate R M8, is more effective, requiring only 1 to 2 µg/mL compared to 136 µg/mL (0.5 mM) required for A77 1726 to reduce the growth of the oomycete by approximately 50% at 92 h. Thus, although PiDHODH is approximately 40-fold more sensitive to A77 1726 compared to StDHODH, the compound is still a relatively poor inhibitor of the oomycete enzyme, and modifications would be needed to achieve a nanomolar IC 50 .

DISCUSSION
Oomycetes had been traditionally misclassified as fungi, since they have fungus-like growth morphology, most sharing filamentous growth habits, producing spores and occupying similar environments (Latijnhouwers et al., 2003). However, morphological features, biochemical analyses and molecular taxonomy studies indicate that oomycetes reside more closely to algae and diatoms in the Stramenopile group, and are a sister group to apicomplexa, ciliates, dinoflagellates, and amoeboids in the protists' supergroup SAR (Adl et al., 2012). The close relation between oomycetes and apicomplexans (Haldar et al., 2006) led us to hypothesize that strategies that have demonstrated effectiveness for control of the human parasites T. gondii and P. falciparum could also be used to control the growth of P. infestans. Indeed, the reverse strategy, the use of oomyceticidal agrochemicals to control apicomplexan parasites, appears to be effective for strobilurins, such as azoxystrobin, trifloxystrobin, and dimoxystrobin, which inhibit P. falciparum with nanomolar IC 50 s (Witschel et al., 2012). De novo pyrimidine biosynthesis is known to be indispensable for Toxoplasma gondii Bzik, 2002, 2010) and for more than a decade, the fourth enzyme in the pathway, DHODH, has been studied as a drug target to control malaria (Baldwin et al., 2002;Ashley, 2017). Thus, we decided to characterize these enzymes in the plant pathogen and its host.
As is the case for many organisms, P. infestans also possesses a salvage pathway for recycling nucleosides and nucleobases, that together with de novo biosynthesis, allow the pathogen to maintain the pyrimidine requirements needed for survival (Garcia-Bayona et al., 2014). Early biotrophic growth is fueled by the de novo pathway (Garcia-Bayona et al., 2014), and is also supported by the uptake of pyrimidines from the host environment, as demonstrated by apparent increases in expression of the pathogen's nucleoside and nucleobase transporters during plant infection (Abrahamian et al., 2016). The fungal plant pathogen Magnaporthe oryzae exhibits a similar requirement for de novo synthesis for its early biotrophic growth in the case of purine nucleotides, which apparently are not sufficiently available from the plant host (Fernandez et al., 2013). In the solanaceous hosts of Phytophthora, such as potato, the de novo pathway is important in growing and developing tissues (Giermann et al., 2002). However, studies of knockdowns of pyrimidine biosynthetic enzymes in transgenic plants demonstrate that the salvage pathway can maintain the required pyrimidine pool for basic metabolism (Schroder et al., 2005) and both pathways are tuned in mutually compensatory fluxes (Zrenner et al., 2006). An additional catabolic pathway, not present in the pathogen, allows for fine-tuning of the plant pyrimidine levels (Zrenner et al., 2009). Dihydroorotate dehydrogenase catalyze the oxidation of dihydroorotate to orotate, and are classified as follows. Class 1 enzymes are soluble, using either fumarate (class 1A) or NAD + (class 1B) as their electron acceptors, and class 2 enzymes are membrane-associated, using ubiquinones in the electron transport chain (ETC) as electron acceptors (Reis et al., 2017). The DHODHs from both S. tuberosum and P. infestans are class 2 enzymes. The mitochondrial localization of plant DHODHs was demonstrated many years ago. In tomato, DHODH co-sediments with the mitochondrial proteins cytochrome oxidase, succinate dehydrogenase, and citrates synthase in sucrose gradients (Miersch et al., 1986). In pea, DHODH is associated with purified mitochondria, and has been shown to reduce cytochrome c in mitochondrial extracts when cytochrome oxidase is inhibited by cyanide. The N-terminus of a class 2 enzyme contains a mitochondrial targeting sequences followed by hydrophobic transmembrane domain that anchors the enzyme to the inner mitochondrial membrane (Chen and Jones, 1976;Zameitat et al., 2007) The structures available for enzymes from this class show that they are comprised of a large α/β-barrel domain containing the orotate/dihydroorotate and FMN binding sites, connected to a small domain containing two alpha helices, αA and αB, forming   a hydrophobic tunnel where ubiquinone is thought to bind (Reis et al., 2017). Differences were observed in the apparent kinetic parameters of the plant and pathogen DHODHs. The K m app for dihydroorotate was threefold higher for N54PiDHODH compared to that of N69StDHODH. In contrast, the K m app for the electron acceptor Qd was twofold higher for N69StDHODH compared to that of N54PiDHODH ( Table 1). The k cat app of the pathogen enzyme was approximately double that of the plant enzyme (Table 1). We tested the activities of the recombinant enzymes with ubiquinones having different isoprenoid tail lengths. While humans use Q10, a ubiquinone with 10 isoprenoid units, both Q9 and Q10 have been found in plants. Ubiquinone content for a few solanaceous plants are available. Tobacco contains Q10 at ≈8 mg/kg (Ohara et al., 2004). Q10 is also found in potato and tomato, but at approximately 10-fold lower concentrations, and levels of Q9 are below the detection limit (Mattila and Kumpulainen, 2001). Although the nature of the quinone in P. infestans is not known, the ubiquinone pool in Phytophthora cactorum appears to be comprised of 80% Q9, and 20% Q8 (Richards and Hemming, 1972), in agreement with a later study showing that Q9 is the major quinone in other oomycetes (Nakamura et al., 1995). Surprisingly, Q9 was not the best substrate for N54PiDHODH, which exhibited approximately double the activity with ubiquinones having shorter isoprenoid tails. Similarly, N69StDHODH appeared to be more active with ubiquinones having isoprenoid tails shorter than Q10 ( Table 2). We also tested the activity of the recombinant enzymes with several 1,4-naphthoquinones ( Table 2). These compounds, which affect diverse cellular targets because of their redox properties (Klotz et al., 2014), are alternate electron acceptors for DHODHs . For N69StDHODH the best acceptor was 5,8-dihydroxy-1,4-naphthoquinone, which exhibited an activity similar to Qd. In contrast, this naphthoquinone was a poor acceptor for N54PiDHODH, showing sixfold less activity compared to Qd. Differences in activities for the two enzymes were also observed for lawsone, juglone, and plumbagin.
We tested compounds known to inhibit DHODHs from other organisms (Table 3). Alloxan and 5 -fluoroorotic acid, analogs of dihydroorotate and orotate, respectively, are weak inhibitors of other class 2 DHODHs (Zameitat et al., 2006;Munier-Lehmann et al., 2013), and were also poor inhibitors of N69StDHODH and N54PiDHODH. The remaining compounds in Table 3 have been shown to interact, or are predicted to interact, with the ubiquinone binding site, which is the target for the most effective inhibitors known to date of class 2 DHODHs (Munier-Lehmann et al., 2013). The variability in this site in class 2 DHODHs has permitted the development of species-specific inhibitors, as illustrated by DSM-265, a triazolopyrimidine inhibitor of P. falciparum DHODH (Ashley, 2017;McCarthy et al., 2017). We found that the related compound DSM190 (Gujjar et al., 2011), had similar and moderate effects on the activities of both recombinant plant and pathogen DHODHs, as did two nanomolar inhibitors of the human enzyme, redoxal  and dichloroallyl lawsone . Brequinar, another nanomolar inhibitor of the human enzyme (Munier-Lehmann et al., 2013), caused moderate inhibition of N69StDHODH, but did not decrease the activity of N54PiDHODH. Greater inhibition of N69StDHODH compared to N54PiDHODH was also observed for toltrazuril, a weak inhibitor of Eimeria pyrimidine biosynthesis (Zameitat et al., 2006;Munier-Lehmann et al., 2013), NSC 61890, and NSC 71097. Several compounds caused greater inhibition of the Phytophthora enzyme than the plant enzyme. For example, the active metabolite of leflunomide, A77 1726, a nanomolar HsDHODH inhibitor used in the treatment rheumatoid arthritis, inhibited N54PiDHODH, but was a poor inhibitor of N69StDHODH. The differential effects of modifying a lead compound scaffold on the inhibition of an enzyme target is nicely illustrated by the MD compounds, a series of inhibitors that were designed in attempt to optimize the binding of A77 1726 to the active sites of PfDHODH or HsDHODH (Leban et al., 2005;Davies et al., 2009). While MD 129 did not affect either enzyme's activity, MD 209 strongly inhibited both enzymes, and MD108 and MD241 preferentially inhibited the pathogen enzyme.
Since we had limited quantities of MD108 and MD241, we selected A77 1726 for further investigation. Measurement of IC 50 s revealed that the pathogen enzyme, with an IC 50 of 79 µM, was 37-fold more sensitive than the plant enzyme. We found that a concentration of 0.5 mM A77 1726 reduced the growth of the oomycete by approximately 50%, at 92 h. Despite the number of DHODH inhibitor sets available (Munier-Lehmann et al., 2013), only two lipophilic inhibitors have been described and suggested as lead compounds for development of oomyceticides that target this enzyme (Parker et al., 2002). These compounds arrested Py. aphinadermatum radial growth in vitro and the addition of uridine was able to reverse this growth inhibition, suggesting that the compounds block pyrimidine metabolism. Nevertheless, these compounds were not stable enough under field conditions to be further developed as oomyceticides.
While the role played by the DHODH in de novo pyrimidine synthesis is well-defined, it is less clear what role it plays in the respiratory chain of P. infestans. In T. gondii, the mitochondrially-associated DHODH has an unknown, but essential, pyrimidine-independent function (Hortua Triana et al., 2016). In mammalian cells, physical associations between DHODH and complexes II and III have been demonstrated, and DHODH knockdown partially inhibits complex III, decreases membrane potential, and increases production of reactive oxygen species (Fang et al., 2013). Relatively little is known about the respiratory chains of plant pathogens. In P. infestans, inhibition of oxygen consumption is observed in the presence of cyanide and in the presence of antimycin A, when succinate is used as the electron donor (Scheepens and Fehrmann, 1977). Electron transport is reported to be insensitive to rotenone (Scheepens and Fehrmann, 1977). P. infestans possesses sequences with similarity to type II NADH dehydrogenases (Feng et al., 2012), plant alternate NADH dehydrogenases (Schertl and Braun, 2014), electron transfer flavoprotein ubiquinone oxidoreductase, and several so-called alternative oxidases (AOX) (Supplementary Table 3). Thus, P. infestans mitochondrial respiration appears to exhibit a complexity due to the increased number of branches, as is also observed in plants (Schertl and Braun, 2014). Oomycete respiratory chains are the target of several commercial oomyceticides, ametoctradin (Initium R ), amisulbrom, and cyazofamid, bind to ubiquinone sites of respiratory complex III, and are effective against late blight disease (Mitani et al., 2001;Dreinert et al., 2018). Since complex III and DHODH both bind ubiquinone, we tested ametoctradin on the recombinant DHODs, and found it had a negligible effect on the plant enzyme, and showed a weak inhibition of the pathogen enzyme. Atovaquone, an antimicrobial ubiquinone analog that inhibits the complexes III of Plasmodium, T. gondii, and Pneumocystis carinii (Meshnick et al., 2001;Kessl et al., 2003;Mather et al., 2005) likewise showed little effect on either recombinant enzyme.
Taken together, our results highlight differences in binding of DHODH inhibitors by the plant and oomycete enzymes, and suggest that these differences could be further exploited to develop species-specific inhibitors that preferentially affect the pathogen enzyme. The inactivation of this enzyme by inhibitors would have a profound effect on pyrimidine biosynthesis, and might also have effects on the mitochondrial respiration of this pathogen. The availability of recombinant N54PiDHODH should expedite the discovery of more potent agents for growth control strategies against P. infestans, and permit the screening of a large number of compounds, the examination of structureactivity relationships of inhibitors, and determination of the 3D structure of enzyme-inhibitor complexes.

CONCLUSION
To our knowledge this is first preliminary characterization of a purified recombinant oomycete DHODH and the first comparison with the corresponding plant DHODH. Whether the differences observed between the two enzymes could be further exploited to develop species-specific compounds for crop management is an intriguing question that remains to be answered. The availability and characterization of the recombinant DHODH from P. infestans in this work permitted the first preliminary screening of potential enzyme inhibitors, with the rationale of interfering with pyrimidine metabolism, and opens the way to protein crystallization, which is a prerequisite for the development of species-specific inhibitors.

FUNDING
This work was supported by the Colciencias grant # 120-4521-28532 (to SR) and by a Colciencias doctoral fellowship from the Becas Doctorado Nacional (to MG). This work was also supported by funding from the Facultad de Ciencias and the Vicerrectoria de Investigaciones (Universidad de los Andes, Colombia).