Expression, purification, and inhibition profile of dihydrofolate reductase from the filarial nematode Wuchereria bancrofti

Filariasis is a tropical disease caused by the parasitic nematodes Wuchereria bancrofti and Brugia malayi. Known inhibitors of dihydrofolate reductase (DHFR) have been previously shown to kill Brugia malayi nematodes and to inhibit Brugia malayi DHFR (BmDHFR) at nanomolar concentrations. These data suggest that BmDHFR is a potential target for the treatment of filariasis. Here, protocols for cloning, expression and purification of Wuchereria bancrofti DHFR (WbDHFR) were developed. The Uniprot entry J9F199-1 predicts a 172 amino acid protein for WbDHFR but alignment of this sequence to the previously described BmDHFR shows that this WbDHFR sequence lacks a crucial, conserved 13 amino acid loop. The presence of the loop in WbDHFR is supported by a noncanonical splicing event and the loop sequence was therefore included in the gene design. Subsequently, the KM for dihydrofolate (3.7 ± 2 μM), kcat (7.4 ± 0.6 s-1), and pH dependence of activity were determined. IC50 values of methotrexate, trimethoprim, pyrimethamine, raltitrexed, aminopterin, (-)-epicatechin gallate, (-)-epicatechin, and vitexin were measured for WbDHFR and BmDHFR. Methotrexate and structurally related aminopterin were found to be effective inhibitors of WbDHFR, with an KI of 1.2 ± 0.2 nM and 2.1 ± 0.5 nM, respectively, suggesting that repurposing of known antifolate compound may be an effective strategy to treating filariasis. Most compounds showed similar inhibition profiles toward both enzymes, suggesting that the two enzymes have important similarities in their active site environments and can be targeted with the same compound, once a successful inhibitor is identified.


Introduction
Lymphatic filariasis (elephantiasis) is a disfiguring and incapacitating disease caused by three species of mosquito borne parasitic worms, Wuchereria bancrofti, which is responsible for 90% of the cases, Brugia malayi and Brugia timori. This disease threatens the well-being of 947 million people in 54 countries. Clinical manifestations include lymphedema of the limbs a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 (currently approximately 15 million cases worldwide) and hydrocele (swelling of the scrotum and penis, approximately 25 million cases) [1,2]. Those infected with filariasis further suffer from stigma, disabilities, and the associated economic consequences.
Dihydrofolate reductase (DHFR) is an NADPH dependent enzyme that catalyzes the formation of tetrahydrofolate from dihydrofolate [3,4]. With the exception of some prokaryotes [5], DHFR is a ubiquitous enzyme required for folate metabolism and DNA synthesis. As such, DHFR inhibition by "antifolates" has proven to be a successful strategy in the treatment of cancer, bacterial infections and malaria [6,7]. A recent Brugia malayi DHFR (BmDHFR) 3-D structural modeling and docking analysis predicted several antifolate compounds to be effective inhibitors of the enzyme [8]. These predictions are potentially supported by findings reported in three recent articles that show Brugia malayi nematode mobility decreased in the presence of antifolate agents [9][10][11]. Moreover, folic acid reversal studies have shown that the mobility of microfilariae decreased less when the nematodes were pre-incubated with folic acid before treatment with the antifolate compounds. Hande and coworkers also predicted vitexin, a compound found in passion flower, and the green tea compounds epicatechin and (-)-epicatechin gallate to be inhibitors of BmDHFR [8].
We recently developed methods to clone, express and purify BmDHFR, and have demonstrated its inhibition by well-known antifolates [12]. DHFR from Wuchereria bancrofti (WbDHFR) is 96% identical to BmDHFR in amino acid sequence. We now report the development of methods to clone, express and purify WbDHFR and compare its kinetic parameters and inhibitor profile to those of BmDHFR. Such a comparison allows insights into whether the amino acid differences between the two sequences have impact on kinetic parameters and inhibitor binding.

Wuchereria bancrofti (Wb) gene sequence development
A nucleotide sequence encoding WbDHFR with an N-terminal His-6 tag was designed, synthesized, and codon optimized for expression in E. coli by Genewiz. The resulting DHFR gene sequence was subcloned into pET25b via NdeI and XhoI sites and transformed into the E. coli LOBSTR strain for protein expression.

Expression and purification of WbDHFR
WbDHFR was expressed at 25˚C in LB media with 100 μg/mL ampicillin and induction overnight with IPTG at 0.3 mM. The enzyme was harvested by centrifuging the E. coli mixture at 5,000 rpm for 30 min at 4˚C using a JA-10 rotor in a Beckman Avanti J-26S XP centrifuge. The pellet was collected and supernatant discarded. This pellet was then resuspended using equilibration buffer (10 mM imidazole, 20 mM Na 2 HPO 4 , 300 mM NaCl, 0.1 mM DTT, at pH 7.4) and soluble protein prepared by sonication of the wet cell paste followed by centrifugation of the mixture using a Sorvall ST16R centrifuge at 5,000 rpm for 30 min at 4˚C. The supernatant, rich in soluble WbDHFR, was collected and pellet discarded. His-tagged WbDHFR was purified at pH 7.4 using Ni-NTA resin. The column was washed with 100 mM imidazole wash buffer (100 mM imidazole, 20 mM Na 2 HPO 4 , 300 mM NaCl, 0.1 mM DTT, at pH 7.4) before being eluted with 250 mM imidazole elution buffer (250 mM imidazole, 20 mM Na 2 HPO 4 , 300 mM NaCl, 0.1 mM DTT, at pH 7.4). Protein was concentrated, and the buffer was exchanged to 20 mM Na 2 HPO 4 , 300 mM NaCl, at pH 7.4 and the concentration was determined spectroscopically at 280 nm using the extinction coefficient 25,440 M -1 cm -1 .

Enzymatic activity assays
To characterize DHFR enzymatic activity, we measured absorbance at 340 nm to follow the disappearance of DHF substrate and NADPH cofactor over time [12]. The K M of WbDHFR for DHF was determined over a concentration range of 3.8 to 195 μM DHF, at 25˚C, in MTEN buffer (50 mM 2-morpholinoethane sulphonic acid (MES), 25 mM Tris, 25 mM ethanolamine, 100 mM NaCl, and 1mM DTT) at pH 6.0. Initial velocity was plotted as a function of DHF concentration using KaleidaGraph and the Michaelis-Menten equation was fitted to the data. Catalytic activities of WbDHFR and BmDHFR were determined at various pH values (5.5-9.0) in MTEN buffer. The MTEN buffer used for all the reported assays has essentially a constant ionic strength at 0.15 over the pH range for which pH values were measured [13,14]. Initial velocity was plotted as a function of pH using Excel.

Inhibition studies
Previous computational research predicted some green tea compounds to be inhibitors of BmDHFR [8]. Compounds (-)-epicatechin, (-)-epicatechin gallate, and vitexin were tested as inhibitors of WbDHFR and BmDHFR. The compounds (-)-epicatechin and (-)-epicatechin gallate were synthesized as described previously [15]. Additionally, methotrexate, trimethoprim, pyrimethamine, aminopterin, and raltitrexed were tested as inhibitors of WbDHFR and BmDHFR. Stock solutions of aminopterin and raltitrexed were prepared in water and stock solutions of the other drugs were prepared in DMSO. Control experiments were conducted to confirm that 5% DMSO (final concentration in the experimental wells) did not affect WbDHFR and BmDHFR activity (data not shown). The concentrations of methotrexate and aminopterin were determined spectroscopically in 0.1 M NaOH at 302 nm using an extinction coefficient of 22,700 M -1 cm -1 . Enzyme activity was measured in wells (200 μL) with 12.5 nM WbDHFR or 40 nM BmDHFR and 100 μM NADPH and 50 μM DHF in MTEN buffer at pH 6.0 at 25˚C. Disappearance of DHF and NADPH was observed by measuring absorbance at 340 nm to measure the DHFR activity in a SpectraMax M3 microplate reader. For active inhibitors, IC 50 curves were generated using KaleidaGraph and the IC 50 values were obtained by fitting the data to the Hill equation with Hill coefficient, n H. = 1. All experiments were completed in triplicate.

Design and subcloning of a WbDHFR gene into a bacterial expression vector
The Uniprot entry J9F199-1 predicts a 172 amino acid DHFR protein for Wb. Alignment of this sequence to the previously described BmDHFR [12], however, shows that this WbDHFR sequence lacks a crucial 13 amino acid loop that is conserved across a number of DHFR proteins from different species (data not shown). The presence of the loop in WbDHFR can be supported by a noncanonical splicing event (data not shown) and the loop sequence was therefore included in the gene design.

Expression and purification of WbDHFR
To make in vitro studies of WbDHFR possible, a protocol was developed for expression and purification of WbDHFR using Ni-NTA resin; approximately 0.9 mg protein / 1 L of culture was obtained (Fig 1). Attempting to purify WbDHFR using the protocol previously developed for BmDHFR [12] resulted in protein with larger molecular weight impurities. To obtain WbDHFR of increased purity, the imidazole concentration in the wash buffer was changed from 25 mM to 100 mM. With this modification, we were able to successfully purify WbDHFR (Fig 1).

Kinetic characterization of WbDHFR
Kinetic characterization of WbDHFR revealed a catalytic activity of 7.4 ± 0.6 s -1 (S.E) at pH 6. This k cat is higher than what was found for BmDHFR, 2.2 ± 0.2 s -1 (S.E.), at the same pH value. The K M found for DHF and WbDHFR, 3.7 ± 2.0 μM (S.D., Fig 2), is lower compared to the K M value previously determined for BmDHFR (14.7 ± 3.6 μM); data for individual trials is included in S1 Table [12]. The activity versus pH profile of WbDHFR was found to be similar to that of BmDHFR (Fig 3). The different y-axis values in the two profiles indicate that WbDHFR catalyzes the reaction faster than BmDHFR at optimal pH values.
WbDHFR and BmDHFR have similar but not identical steady-state kinetic characteristics. Comparison of the WbDHFR and BmDHFR amino acid sequences shows eight residues to be different (Fig 4). There are no crystal structures available for either WbDHFR or BmDHFR and we therefore cannot directly examine the location of the residues with different sidechains. Supporting information shows the locations of the corresponding residues superimposed on a mouse DHFR structure (PDB# 1U70) (S1 Fig), which is the DHFR with the highest level of sequence identity to WbDHFR and BmDHFR and an available solved structure.

Inhibition profile of WbDHFR and BmDHFR
We determined IC 50 values for several known antifolate and green tea compounds against BmDHFR and WbDHFR using the Hill Equation in KaleidaGraph (Table 1); data for individual values is shown in S2 Table [16]. The data show that methotrexate, trimethoprim, raltitrexed, pyrimethamine, and aminopterin inhibit WbDHFR. We did not observe inhibition for (-)-epicatechin gallate, (-)-epicatechin, or vitexin against either WbDHFR or BmDHFR. We used Dixon plots to experimentally investigate whether the five compounds that show inhibition based on IC 50 values (Table 1) act as competitive inhibitors for WbDHFR. We plotted the reciprocals of the initial velocity at different substrate concentrations against inhibitor concentrations. A linear equation was fitted to the data at each substrate concentration. The resulting lines for all inhibitors tested crossed in the top left quadrant, indicating a competitive inhibition mechanism (Fig 5 and S3 Fig) [17]. The negative x-axis values of the point of intersection of the lines for all pairs of individual lines were determined and the average of these values was used to obtain the K I values listed in Table 1; each experiment was conducted in triplicate and the standard deviations are shown and values from individual trials are shown in S3 Table. The K I values for BmDHFR in Table 1 were obtained using the Cheng-Prusoff Equation [18]. For the two tight-binding inhibitors aminopterin and methotrexate against BmDHFR, a modification of the Cheng-Prusoff Equation for competitive inhibition for tightly bound inhibitors was needed and we report an upper limit for the K I values [19,20]. Inhibitor structures are shown in supporting information (S2 Fig). Most inhibitors that were tested have similar IC 50 and K I values towards both nematode homologs but pyrimethamine inhibits BmDHFR with a K I value of 3.6 ± 1.5 μM while this drug binds WbDHFR four-times less tightly with a K I of 15 ± 6 μM. These data suggest similarities but also subtle differences in the active sites of the two enzymes that have only eight different amino acids in their sequences.
The IC 50 value determined here for pyrimethamine for BmDHFR is different compared to the previously determined value of the same drug against the same enzyme: 15.6 ± 6.6 μM found now versus 109 ± 34 μM found previously [12]. To verify the drug stock concentration in the current study, the extinction coefficient for pyrimethamine was determined to be 6.7 ± 0.8 mM -1 cm -1 at 268 nm in 1 X MTEN at pH 6.0.

Comparison of current data to previous computational predictions
The data agrees with some of the computational predictions by Hande and coworkers [8]; for example, they authors predicted that trimethoprim would inhibit BmDHFR with a K I of 11 μM and we found the K I of trimethoprim to be 15 μM against BmDHFR. On the other hand, vitexin was predicted to be a 465 nM inhibitor of BmDHFR, but in our assays we did not observe any inhibition for vitexin against BmDHFR (Table 1). Similarly, (-)-epicatechin and (-)-epicatechin gallate were predicted to have K I values of 76 μM and 48 μM against BmDHFR [8] but neither compound showed any inhibitory activity against BmDHFR or WbDHFR, even  https://doi.org/10.1371/journal.pone.0197173.t001 at concentrations greater than 10 mM. We also examined other compounds that are structurally related to (-)-epicatechin and (-)-epicatechin gallate and observed similar results. As the authors state themselves, the computational predictions must be interpreted with caution due to a lack of a crystal structure for any of the filarial parasite DHFRs.

Discussion
We found that the Uniprot entry J9F199-1 for WbDHFR lacks a crucial 13 amino acid loop. WbDHFR, consisting of 185 amino acids (Fig 4), was successfully designed and subcloned into the pET25b expression vector and expressed in LOBSTR E. coli cells using a modified version of a protocol previously developed for BmDHFR. The methods that were developed to purify active WbDHFR for in vitro studies will facilitate the testing of additional antifolate compounds as potential inhibitors in the treatment of filariasis. Well known antifolates, methotrexate and trimethoprim, were found to inhibit WbDHFR with K I values of 1.2 ± 0.2 nM and 6 ± 0.06 μM, respectively. These K I values are significantly different from those of methotrexate and trimethoprim against human DHFR (40 pM and 1.38 μM, respectively) [21], indicating that there are differences in the inhibitor binding of the human DHFR compared to the parasite homologs that will likely enable discovery of selective inhibitors. These data suggest that repurposing of known antifolate compounds can be an effective approach for the treatment of filariasis. The expression, purification and basic kinetic analysis of WbDHFR we publish here make it possible to test other synthetic molecules proven to act on DHFRs from other organisms as inhibitors of WbDHFR. BmDHFR and WbDHFR have similar kinetic and inhibition parameters; 177 of the 185 amino acid residues are conserved (Fig 4, S1 Fig). We are currently working toward obtaining an x-ray crystal structure of WbDHFR with an inhibitor and NADPH bound. Such a structure will further facilitate the development of antifolate compounds in the treatment of filariasis. Most of the antifolates that were tested, including those with lower IC 50 values, inhibit the two homologs similarly, suggesting the possibility that one DHFR inhibitor could be used to treat both filarial parasites. Such an approach would be helpful in resource-poor settings where the infrastructure to determine which parasitic infection is present is not available.  Dixon Plots for methotrexate (A.), raltitrexed (B.), pyrimethamine (C.), and aminopterin (D.) for WbDHFR. All reactions were performed at 25˚C in 1 X MTEN buffer at pH 6.0. The concentration of WbDHFR and NADPH were kept constant at 6 nM and 100 μM, respectively. DHF concentrations of 2, 4, and 8 μM were used. All experiments were performed in triplicate. The plots were generated in Excel. The K I values are shown in S1 Table. Data for trimethoprim is shown in Fig 5. (TIF) S1