Skip to main content
NCBI home page
ACS Omega logoLink to ACS Omega
. 2021 Oct 1;6(40):26065–26076. doi: 10.1021/acsomega.1c02995

Effects of Distal Mutations on Ligand-Binding Affinity in E. coli Dihydrofolate Reductase

Chen-Hua Huang ‡, Yun-Wen Chen ‡, Tsun-Tsao Huang ‡, Ya-Ting Kao †,‡,§,*
PMCID: PMC8515367  PMID: 34660967

Abstract

graphic file with name ao1c02995_0008.jpg

Mutations far from the center of chemical activity in dihydrofolate reductase (DHFR) can affect several steps in the catalytic cycle. Mutations at highly conserved positions and the distal distance of the catalytic center (Met-42, Thr-113, and Gly-121) were designed, including single-point and double-point mutations. Upon ligand binding, the fluorescence of the intrinsic optical probe, tryptophan, decreases due to either fluorescence quenching or energy transfer. We demonstrated an optical approach in measuring the equilibrium dissociation constant for enzyme–cofactor, enzyme–substrate, and enzyme–product complexes in wildtype ecDHFR and each mutant. We propose that the effects of these distal mutations on ligand-binding affinity stem from the spatial steric hindrance, the disturbance on the hydrogen network, or the modification of the protein flexibility. The modified N-terminus tag in DHFR acts as a cap on the entrance of the substrate-binding cavity, squeezes the adenosine binding subdomain, and influences the binding of NADPH in some mutants. If the mutation positions are away from the N-terminus tag and the adenosine binding subdomain, the additive effects due to the N-terminus tag were not observed. In the double-mutant-cycle analysis, double mutations show nonadditive properties upon either cofactor or substrate binding. Also, in general, the first point mutation strongly affects the ligand binding compared to the second one.

Introduction

The enzyme dihydrofolate reductase (DHFR) catalyzes the reduction of 7,8-dihydrofolate (H2F) to 5,6,7,8-tetrahydrofolic acid (H4F) with nicotinamide adenine dinucleotide phosphate (NADPH) as the cofactor. The catalytic product H4F is essential for amino acid catabolism. Hence, DHFR has been recognized as a target for anticancer drugs and antibacterial agents, such as methotrexate (MTX) and trimethoprim (TMP).1,2Escherichia coli DHFR (ecDHFR) is a monomeric protein of 159 amino acid residues (∼18 kD). There are over 40 structures of ecDHFR in both binary and ternary complexes states.3−5 Its three-dimensional structure is dominated by a central eight-stranded β-sheet (strands βA−βH) with four α-helices (αB, αC, αE, and αF) flanking around. The enzyme comprises the adenosine binding subdomain (residues 38–88) and the major subdomain, also named loop subdomain, encompassing the binding sites for the cofactor and the substrate. The loop subdomain contains three critical loops: the Met20 loop (residues 9–24) closes over the active-site pocket, and the βF−βG (residues 116–132) and βG−βH (residues 142–150) loops stabilize the various conformations of the Met20 loop. The substrate and the cofactor bind in a deep hydrophobic cleft at the juncture of the two subdomains. The nicotinamide ring of NADPH is coplanar with the β-sheet and spans the gap between strands βA and βE. The pteridine ring of H2F fits into the cleft between helices αB and αC.

Previous studies have shown that the mutations far from the center of chemical activity in DHFR can affect several steps in the catalytic cycle.6−13 Among these studies, the distal residues Gly-121 and Met-42 are highly conserved. The double-mutant enzymes in which Gly-121 and Met-42 are both substituted exhibit synergistic reduction in the hydride transfer rate.7−10 Gly-121 is located on the βF−βG loop and next to Asp-122, forming hydrogen bonds with Glu-17 and Gly-15 on the Met20 loop. A mutation on this neighboring residue Gly-121 leads to a decrease in the rate of hydride transfer.6 Met-42 is located on the strand βB and close to the Trp-47 on helix αC. The helix αC is in the adenosine binding subdomain and directly interacts with the Met20 loop associating to the cofactor binding and catalytic reaction. Also, Asp-27 is near the active site and plays an essential role in facilitating the proton transfer and hydride transfer steps. Asp-27 may help to orient the substrates and the water molecule through various electrostatic interactions.14−17 Another conserved residue Thr-113 is located in strand βF, away from the adenosine binding subdomain but forming hydrogen bonding with Asp-27.5,14

We systematically measured the equilibrium thermodynamic dissociation constant for enzyme–cofactor, enzyme–substrate, and enzyme–product complexes in wildtype ecDHFR and each mutant. We selected residues at highly conserved positions but different domains (Met-42, Thr-113, and Gly-121) to carry out single and double mutation. We further analyzed the free energy changes to determine the mutation effect on the ligand affinity and possibly on the reaction activity.

Results and Discussion

Enzyme Activity Assays

The E. coli DHRF sequence with a 19-residue lysine biotinylation signal sequence inserted at the N terminal (MGLNDIFEAQKIEWHGGGT; bioseq)18 was subcloned into the pET28a plasmid. The pET28a-6Histag-bioseq-ecDHFR construct was used to express ecDHFR, according to the previous report with some modifications.18 Three mutation points (Met-42, Thr-113, and Gly-121) are highly conserved and distal to the catalytic center, as shown in Figure 1A. We designed these mutations with possible minor disturbance in an active-site structure but with more effects in the local conformation and the interaction network. The residue Met-42 is on the strand βB and in the proximity of helix αC. It was proposed that Met-42 and the helix αC exhibit the correlated motion in wildtype.9 In previous studies, the forward hydride transfer rates decreased 1.4 times in M42F and displayed negligible changes in M42A and other mutations involving smaller amino acids.8 Moreover, with a more massive replacement, M42W, the forward hydride transfer rate decreased 41 times. Hence, we replaced Met residue with the bulky Phe residue (M42F) to induce hindrance in the correlated motion. The Gly-121 is located at the center of the βF−βG loop and is 19 Å from the catalytic center. The βF−βG loop is quite flexible and adopts several conformations during catalytic processes.4 In previous studies, the forward hydride transfer rates decreased 163 and 62 times in G121V and G121S, respectively.8 Val residue and Ser residue are similar in dimension but have different polarities. We replaced Gly residue with either Val or Ser residues to induce a disturbance in loop motion. Thr-113 is located in βF and forms hydrogen bonding with Asp-27. In previous studies, the forward hydride transfer rates decreased 1.3 times in T113V.19 Thr residue and Val residue are similar in dimension but have different polarities. We replaced Thr residue with Val to induce a disturbance in the hydrogen network at the catalytic center. Site-directed mutagenesis steps were carried out to engineer the following mutations: (1) single-point mutation: M42F-ecDHFR, T113V-ecDHFR, G121S-ecDHFR, and G121V-ecDHFR, (2) double-point mutation: M42F/G121S-ecDHFR, M42F/G121V-ecDHFR, T113V/G121S-ecDHFR, and T113V/G121V-ecDHFR. Some mutations have been studied previously and are used for comparison. Some of the mutations are investigated for the first time here.

Figure 1.

Figure 1

Open in a new tab

(A) Crystal structure of ecDHFR (PDB: 1RX2). Five blue spheres represent the tryptophan amino acid residues (W22/W30/W47/W74/W133). The yellow and green sticks represent the cofactor and substrate. The red labels are the mutation sites (M42, T113V, and G121). (B) Results of enzyme activity assays in wildtype and each mutant. The percentage of the H2F–to–H4F conversion was estimated by eq 1.

The activity of ecDHFR was determined at 25 °C following the decrease of NADPH and H2F by the absorbance measurements at 340 nm. The reaction MTEN buffer20 at pH 7.4 containing the reaction solutions (5 μM ecDHFR, 100 μM NADPH, and 50 μM H2F) were kept at 25 °C, while the absorption measurements were performed for 15 min, and the 15 min endpoint absorbance was recorded. Because ecDHFR and NADP+ exhibit no absorption at 340 nm, we estimated the initial point of reaction with a mixture of 100 μM NADPH and 50 μM H2F and the final point of reaction with a mixture of 50 μM NADPH and 50 μM H4F. The concentrations of compounds in solution were determined by UV absorbance using the molar extinction coefficients from the previous study.21

The 340 nm absorbances of the initial point and the final point are 0.98 and 0.56, respectively. The percentage of the H2F–to–H4F conversion in each mutant was estimated by

graphic file with name ao1c02995_m001.jpg 1

All the mutants show catalytic activity with different levels of the substrate-to-product conversion. M42F-ecDHFR and M42F/G121V-ecDHFR show a similar conversion level as the wildtype, while others show lower conversion levels, especially M42F/G121S-ecDHFR and T113V/G121S-ecDHFR. Because the initial 5 μM concentration of ecDHFR is high, we observed the endpoint results occurring right after mixing in the case of most mutants. However, G121V-ecDHFR, M42F/G121V-ecDHFR, and T113V/G121V-ecDHFR show a relatively slow conversion process. Although M42F/G121S-ecDHFR and T113V/G121S-ecDHFR exhibit the lowest conversion level, both of them reach the endpoint rapidly, indicating that the product-binding affinity might be more substantial than the substrate-binding affinity, and such competition reaches equilibrium rapidly. In contrast, three mutants (G121V-ecDHFR, M42F/G121V-ecDHFR, and T113V/G121V-ecDHFR) show a slow turnover but higher conversion above 60%. Among them, M42F/G121V-ecDHFR is the most intriguing mutant. It fulfills conversion of 98% but reaches to endpoint after 14 min of mixing. The slower turnover indicates that this double mutation hinders the binding of ligands and the sampling time of the preorganization configuration in the catalytic center.

Molecular Dynamics Simulation

Both Met-42 and Thr-113 are not on the protein surface and interact with its nearby α helix in the van der Waals contacts. Hence, M42F and T113V should exhibit a negligible difference with wildtype in protein–surface electrostatics and minor effects on the N-terminus-tag structure. Also, Gly-121 is on the flexible βF−βG loop, and the adenosine binding subdomain and the major subdomain of ecDHFR are flanked by the βF−βG loop and the N-terminal. Hence, G121V and G121S should exhibit minor effects on the N-terminus-tag structure. We monitored the interaction between the N-terminus-tag and the main wildtype ecDHFR protein during the molecular dynamics (MD) simulation.22−30 The root-mean-square deviation of the backbone (bbRMSD) for the 1000 ns MD run concerning the reference conformation at 0 ns and the conformation at 1000 ns (red line) are shown in Figure S1. The root-mean-square deviation of the backbone for the 1000 ns MD run indicates that the N-terminus-tag would be stable in the equilibration portion of the trajectory (>600 ns). As shown in Figure 2, the N-terminus tag folds close to the entrance of the substrate-binding cavity. Because residues L28 and I50 are at the entrance of the substrate-binding site, we monitored their distances to the residue tR33 on the N-terminus tag. Figure 2 displays the distances of residue pairs tR33-L28 (red line) and tR33-I50 (green line) with respect to simulation time, indicating that the N-terminus tag could quickly fold and be close to the L28 and I50 in 50 ns. In the equilibration portion of the simulation (600–1000 ns), the minimal distance of tR33-L28 is about 2.5 Å as well, as the tR33-I50 ranges from 3 to 7 Å, demonstrating that the N-terminus tag could fold close to the entrance of H2F binding site. Such a configuration might prevent H2F from binding to the ecDHFR. The driving force pulling the N-terminus tag to the H2F-binding entrance might be the electrostatic interaction between tD32 on the N-terminus tag and the R52 of the ecDHFR. The tD32-R52 distance constantly narrowed down for the first 400 ns simulation. Then, the tD32-R52 distance was stably around 2 Å in the equilibration portion of the trajectory. Our 1000 ns MD simulation of the tag-DHFR suggests that the N-terminus tag might block the entrance of the H2F binding site and interfere with the binding of H2F. While comparing structures at 0 and 1000 ns, we observed that (1) the adenosine binding subdomain leans closer to αF and the entrance of cofactor binding site becomes smaller; (2) the distance between I50 and the nicotinamide-ribose moiety of NADPH becomes larger; (3) the Met20 loop moves away from the active site. These changes might also influence the binding of NADPH.

Figure 2.

Figure 2

Open in a new tab

MD simulation on the N-terminus tag along its folding process and the minimum distances among residue pairs. The N-terminus tag folds close to the entrance of the substrate-binding cavity. Although the distance between L28 and I50 (black line) is relatively stable, the distances of residue pairs tR33-L28 (red line) and tR33-I50 (green line) decrease along with the MD simulation, indicating that the N-terminus tag could quickly fold and be close to the L28 and I50. Besides, the electrostatic interaction between tD32 (on the N-terminus tag) and the R52 drives them close to around 2 Å (blue line).

Cofactor-Binding Assays

The structure of the ecDHFR complex with the oxidized cofactor and substrate analogue is shown in Figure 1 with five intrinsic tryptophan residues (W22/W30/W47/W74/W133) labeled with blue spheres. When excited by 290 nm irradiation, ecDHFR is fluorescent and exhibits a fluorescence band of 300–460 nm with an emission peak at 340 nm. The emission spectrum of tryptophan and the absorption spectrum of NADPH are overlapped well, and five tryptophan residues are located within 16 Å to NADPH.5 Hence, ecDHFR and NADPH could be a Förster resonance energy transfer (FRET) pair.6,8,9,19 Other ligands, such as H2F, H4F, and methotrexate (MTX), could not form a FRET pair with tryptophan, but they show a quenching effect on the tryptophan fluorescence of ecDHFR.6,8,9,19

Figure 3A shows absorption spectra of 16 μM wildtype-ecDHFR, 20 μM cofactor NADPH and a mixture of 16 μM ecDHFR and 20 μM NADPH, along with their emission spectra upon excitation by either 290 or 340 nm irradiation. Under 290 nm excitation, the 340 nm emission bands in ecDHFR originated from tryptophan fluorescence. Under 340 nm excitation, cofactor NADPH exhibits an emission band with a peak of 460 nm, but the emission peak shifts to 440 nm in the enzyme–cofactor mixture. Such spectral blueshifting indicates the binding of NADPH to ecDHFR resulting from slower solvation dynamics and a higher emitting state.31 In an aqueous solution, the fluorescence of nicotinamide is partially quenched by collisions or by stacking with the adenine moiety.32,33 Upon binding to ecDHFR, a nicotinamide–ribose moiety flips in the active-site pocket5 (no collision and stacking with the adenine moiety), resulting in the increment of fluorescence intensity and blue shifting on the emission peak from 460 to 440 nm, as shown in Figure 3A. Figure 3B shows the emission spectra of 30 μM wildtype-ecDHFR, 20 μM NADPH, and a mixture of 30 μM ecDHFR and 20 μM NADPH upon 290 nm excitation. We observed a decrement at 340 nm fluorescence and an increment at 440 nm fluorescence in the mixture compared with ecDHFR alone. Because 20 μM NADPH shows extremely low fluorescence intensity upon 290 nm excitation, the 440 nm emission in the mixture must result from FRET between tryptophans and NADPH in the complex. FRET between tryptophan residues in enzymes, and cofactor NADPH (or NADH) was observed in many cases.34−36

Figure 3.

Figure 3

Open in a new tab

(A) Absorption spectra of ecDHFR (black line), cofactor NADPH (blue line), and the mixture (red line) of ecDHFR (16 μM) and cofactor NADPH (20 μM). The emission spectrum of ecDHFR [(black-bold line)], upon 290 nm excitation, and the emission spectrum of NADPH either in solution (blue-bold line) or the mixture (red-bold line), upon 340 nm excitation. (B) Emission spectra of ecDHFR (black line), cofactor NADPH (blue line), and the mixture (red line) of ecDHFR (30 μM) and cofactor NADPH (20 μM), under 290 nm irradiation. (C) Fluorescence spectra from the cofactor-binding titration experiments and the fluorescence spectrum of 50 μM NADPH (black-thin line). Inset shows the KD analysis curves of five trials only up to 100 μM NADPH titration. The range of ΔF340nm is from −0.1 to 1.1.

We measured the 340 nm fluorescence intensity of ecDHFR at various enzyme concentrations from 500 nM up to 60 μM. The nonlinearity between fluorescence intensity and enzyme concentration occurred at a protein concentration of 20 μM. Hence, all ligand-binding assays were performed with 5 μM ecDHFR. In Figure 3D, we show one of the NADPH-binding titrations. With the increasing concentration of NADPH, the decrease at 340 nm fluorescence and the increase at 440 nm fluorescence were observed. When the concentration of NADPH was more than 100 μM, we observed no increment but a decrement at 440 nm fluorescence. Hence, the high concentration data were not used for the dissociation constant KD analysis and the data were only fitted up to 100 μM. The initial concentration of free ecDHFR is [ET] and that of NADPH is [LT]. When reaching the equilibrium of the complex formation, the concentration of the ENADPH complex is [EL]eq. The dissociation constant KD of the ENADPH complex could be represented by

graphic file with name ao1c02995_m002.jpg 2

Thus, [EL]eq could be solved as follows

graphic file with name ao1c02995_m003.jpg 3

Due to FRET, the ENADPH complex is not fluorescent at 340 but 440 nm. If the 340 nm fluorescence of the initial ecDHFR is F0 and that of the ecDHFR and NADPH mixture is FEL, the 340 nm fluorescent signal of the mixture is from the free ecDHFR in the mixture ([E]free = [ET] – [EL]eq). Because the concentration and fluorescence intensity of ecDHFR were in the linear relation range, hence

graphic file with name ao1c02995_m004.jpg 4

and

graphic file with name ao1c02995_m005.jpg 5

As shown in the inset of Figure 3D, the data were plotted with ΔF340nm as Yaxis and [LT] as Xaxis and fit by a nonlinear least-squares fitting program with the eq 5, where F0 is the initial fluorescence intensity in the absence of NADPH, FEL is the fluorescence intensity in the presence of NADPH of [LT], ET is the initial ecDHFR concentration, LT is the initial NADPH concentration, and the dissociation constant KD is the unknown parameter that was obtained by fitting results. The enzyme–cofactor dissociation constant KDNADPH of wildtype-ecDHFR to NADPH was estimated as 1.23 ± 0.06 μM by 340 nm fluorescence titration (Figure 3D inset). This value is about 3∼4-fold weaker than the previous reports of 0.34 and 0.44 μM.6,8 Because the ENADPH complex is not fluorescent at 340 nm but 440 nm, the 440 nm fluorescence intensity is proportional to the concentration of the ENADPH complex, hence

graphic file with name ao1c02995_m006.jpg 6

where A is the population fluorescence yield relative coefficient, F0440nm is the initial 440 nm fluorescence intensity in the absence of NADPH, and FEL is the fluorescence intensity in the presence of NADPH of [LT]. The dissociation constants KD analyzed from two approaches, ΔF340nm and F440nm analyses, are pretty consistent. The comparison between the two methods is shown in Figure S2. The cofactor-binding assays were also performed with 500 nM ecDHFR and NADPH-titration concentrations of 25, 50, 100, 200, 250, 400, 500 nM, 1, 2, and 4 μM. The dissociation constant KDNADPH was estimated as 1.09 ± 0.45 μM by 340 nm fluorescence titration. Because the cofactor-binding assays in two different titration ranges (ecDHFR 500 nM and 5 μM) exhibit similar results, we further used 5 μM enzyme concentration for all the mutation experiments.

The NADPH titration with the fluorescence measurements and ΔF340nm analyses of ecDHFR mutants are shown in Figure S3 and inset. Similar to the ΔF340nm analysis in wildtype, the data in each mutant were only fitted up to 100 μM, and the corresponding enzyme–cofactor dissociation constants KDNADPH are listed in Table 1. However, to show the same ΔF340nm range (−0.1–1.1), single mutations up to 100 μM NADPH and double mutations up to 400 μM NADPH are displayed in the inset of the figure. In all single-point mutations, M42F-ecDHFR shows the weakest binding with NADPH, which might arise due to the proximity to the NADPH-binding site and bulky replacement of phenylalanine. The phenylalanine replacement induces the stereo-hindrance and affects the rotation and movement of helix αC, resulting in weakening the enzyme–ligand affinity. However, this result differs from the previous report,8 in which the M42F-ecDHFR did not deviate from the wildtype behavior. In earlier studies on M42 with 14 amino acid replacements,37 the results demonstrated that M42 is essential to stabilize the hydrophobic interactions with the residues surrounding this position. Considering the N-terminus-tag in our system, we found that the N-terminus tag pushes the adenosine binding subdomain closer to the helix αF and enhances the hindrance effect from M42F. Three other single-point mutants are one-order weaker in the binding of NADPH. All of them are located in the major subdomain and far from NADPH binding site and the N-terminus tag. These results are similar to those reported previously, which indicated that G121S-ecDHFR and G121V-ecDHFR weakened NADPH binding to 10-fold and 40-fold, respectively.6,8 While the mutation positions away from the N-terminus tag, the additive effects on the NADPH binding are not clearly observed.

Table 1. Dissociation Constants for Wildtype DHFR and Mutants With Various Ligands.

    KDNADPH (μM) KDH2F (μM)
  wildtype 1.23 ± 0.06 2.84 ± 0.31
single-point mutation M42F 33.64 ± 2.67 18.25 ± 2.47
  T113V 13.73 ± 1.82 43.93 ± 4.33
  G121S 9.73 ± 2.40 6.52 ± 3.05
  G121V 12.24 ± 1.41 69.41 ± 9.63
double-point mutation M42F/G121S 115.61 ± 20.12 79.97 ± 8.85
  M42F/G121V 73.96 ± 9.09 44.70 ± 6.19
  T113V/G121S 77.49 ± 19.23 60.24 ± 7.21
  T113V/G121V 137.30 ± 1.94 75.77 ± 4.06
Open in a new tab

All double-point mutations ecDHFR show that the dissociation constant KDNADPH increases around 2 orders of magnitude compared to wildtype-ecDHFR. The weaker the NADPH binding, the smaller the 440 nm fluorescence increment observed in the titration experiments. As shown in Scheme 1 and Table 2, the double mutant cycle analysis38−40 of double mutations demonstrates nonadditive properties upon substrate binding. In general, the first point mutation affects strongly the NADPH binding compared to the second one (ΔG1 > ΔG′1 and ΔG2 > ΔG′2).

graphic file with name ao1c02995_m007.jpg
graphic file with name ao1c02995_m008.jpg
graphic file with name ao1c02995_m009.jpg
graphic file with name ao1c02995_m010.jpg

Scheme 1. General Scheme for a Double Mutant Cycle.

Scheme 1

Open in a new tab

Two residues A and B are mutated to A′ an B′, respectively. A′B and AB′ represent the single-point mutants and A′B′ is the corresponding double-point mutant

Table 2. Double-Mutant Cycle Analysis With NADPH and H2F.

NADPH A′ B′ ΔG1 (kcal/mole) ΔG2 (kcal/mole) ΔG′1 (kcal/mole) ΔG′2 (kcal/mole) ΔΔG (kcal/mole)
M42F/G121S M42F G121S 8.20 5.12 6.13 3.06 –2.06
M42F/G121V M42F G121V 8.20 5.69 4.45 1.95 –3.75
T113V/G121S T113V G121S 5.98 5.12 5.19 4.33 –0.79
T113V/G121V T113V G121V 5.98 5.69 5.99 5.71 0.02
H2F A′ B′ ΔG1 (kcal/mole) ΔG2 (kcal/mole) ΔG′1 (kcal/mole) ΔG′2 (kcal/mole) ΔΔG (kcal/mole)
M42F/G121S M42F G121S 4.61 2.06 6.21 3.66 1.60
M42F/G121V M42F G121V 4.61 7.92 –1.09 3.22 –5.70
T113V/G121S T113V G121S 6.71 2.06 5.86 1.13 –0.93
T113V/G121V T113V G121V 6.71 7.92 0.22 1.35 –6.57
Open in a new tab

Substrate-Binding Assays

Five tryptophan residues in ecDHFR are located within 19 Å of substrate H2F, where W30 is at a distance of 4.8 Å to H2F. Although H2F could not form a FRET pair with tryptophan, a quenching effect on the tryptophan fluorescence of ecDHFR by H2F was observed, as shown in Figure 4A. In Figure 4B, with the increasing concentration of H2F, the 340 nm fluorescence decreases, and the emission peak gradually shifts from 340 to 350 nm. When carefully examining the fluorescence contribution of H2F, H2F exhibits weak fluorescence, as shown in Figure 4C. The fluorescence intensity increases linearly up to 20 μM of H2F and displays a nonlinear effect at higher H2F concentrations. We first considered the data in Figure 4C and plotted the 340 nm fluorescence of H2F at various concentrations. Moreover, then we analyzed the raw data with polynomial fitting and simulated the fluorescence intensity curve of H2F, as shown in Figure 4D. The autofluorescence from H2F was removed by simple subtraction. Figure 4E shows the H2F titration raw data, the H2F-simulated fluorescence intensity, and the final H2F titration-calculated data. Both the raw data and the calculated data were used to construct ΔF340nm analysis. Because the nonlinear effect becomes more pronounced at high H2F concentration, the data were only fitted up to 60 μM, as shown in the inset of Figure 4C. The enzyme–substrate dissociation constant KDH2F of wildtype-ecDHFR to H2F was estimated as 2.84 ± 0.31 and 2.11 ± 0.34 μM with and without the deduction of the effect from H2F fluorescence, respectively. When the data were only fitted up to 20 μM (analysis results not shown), the dissociation constant KDH2F of wildtype-ecDHFR to H2F was estimated as 2.54 ± 0.45 μM with the deduction of the effect from H2F fluorescence. This enzyme–substrate dissociation constant is about 9–11-fold weaker than those reported previously, 0.33 and 0.25.6,8 From the MD simulation result, the N-terminus tag blocking the entrance of H2F might be the main reason for the weakening H2F binding affinity.

Figure 4.

Figure 4

Open in a new tab

(A) Emission spectra of 20 μM ecDHFR (black line), substrate 60 μM H2F (blue line), and the mixture (red line) of 20 μM ecDHFR and 60 μM substrate H2F. In the presence of a substrate, the 340 nm tryptophan emission decreases, indicating fluorescence quenching. (B) Fluorescence spectra from the substrate-binding titration experiments and the fluorescence spectrum of 52 μM H2F (black-thin line). The inset shows the KD analysis curves of three trials only up to 60 μM NADPH titration based on the deduction of the H2F effect described in (C–E). The range of ΔF340nm is from −0.1 to 1.1. (C) Fluorescence spectra of H2F at various concentrations. (D) Concentration–fluorescent intensity dependence of H2F (black-hollow circle) was analyzed with polynomial fitting (black curve) labeled with the simulation data (red-solid circle). (E) H2F titration raw data (green square), the H2F-simulated fluorescence intensity (red-solid circle), and the final H2F titration-calculated data. (blue square).

The H2F titration with the fluorescence measurement and ΔF340nm analysis of ecDHFR mutants are shown in Figure S4 and inset. Similar to the ΔF340nm analysis in wildtype, the data of each mutant were only fitted up to 60 μM, and the corresponding enzyme–substrate dissociation constants KDH2F are listed in Table 1. However, to show the same ΔF340nm range (−0.1–1.1), the titration analysis on single mutations up to 100 μM H2F and on double mutations up to 200 μM H2F was carried out. While compared to wildtype, mutants at the T113 position (T113V-ecDHFR, T113V/G121S-ecDHFR, and T113V/G121V-ecDHFR) weakened the H2F binding up to a maximum of 26-fold in T113V/G121V-ecDHFR. Thr-113 is close to Asp-27. The hydroxyl group of T113 forms hydrogen bonding with the carbonyl group of Asp-27 and hydrogen network with cavity water (W405), substrate H2F, and Asp-27.41 In T113V mutants, such network changes, and the H2F binding decreases. Our results are similar to those reported previously, which indicated a 20–30 fold decrease in H2F-binding affinity.42

Surprisingly, the substrate dissociation constant of G121V-ecDHFR is 10-fold higher than that of G121S-ecDHFR. Our results differ from those reported previously,8 which stated that G121S-ecDHFR and G121V-ecDHFR behaved similarly with wildtype in the H2F-binding. Asp-122 forms hydrogen bonds with Glu-17 and Gly-15 and affect the Met20 loop conformation. In the previous studies, the correlation motion between Asp-122/Gly121 and Gly15/Glu17 is observed in wildtype. Such correlation motion becomes unclear in the G121S mutant and is even hardly observed in the G121V mutant.9 The Gly-121 mutations might exhibit a similar effect on protein flexibility, as reported for Asp-122 mutations.41 This G121V-mutation effect might be less critical in the flexible enzyme and results in similar dissociation constants in G121S-ecDHFR and G121V-ecDHFR, as shown in previous studies.6,8,9 However, in our enzymes, the N-terminus tag acts as a cap on the entrance of the substrate-binding site. Hence, the G121V-mutation effect reinforces the compact H2F-binding site, leading to much weaker binding to the substrate. Also, based on MD simulation, the N-terminus tag causes the Met20 loop to move away from the active site. The Val residues is a bulky and nonpolar sidechain, and it weakens the interaction between Asp-122 and the Met20 loop and possibly affects the H2F-binding site. However, in G121S-ecDHFR, the hydroxyl group of Ser residue might participate in the hydrogen network, resulting in a minor effect on the interaction between Asp-122 and the Met20 loop and the H2F binding.

As shown in Table 2, the double mutant cycle analysis38−40 of double mutations demonstrates nonadditive properties upon substrate binding. The effect of G121V and T113V mutation overwhelms that of G121S and M42F. Hence, in M42F/G121V-ecDHFR and T113V/G121V-ecDHFR, we observed that ΔG1 ≫ ≫ ΔG′1. In T113V/G121S-ecDHFR and T113V/G121V-ecDHFR, we observed that ΔG2 ≫ ≫ ΔG′2.

Product-Binding and Inhibitor-Binding Assays

The effects of ligand autofluorescence and inner filter effect are more enhanced with the H4F ligand, as shown in Figure 5A. The original H4F titration raw data, the polynomial fitting, the H4F-simulated fluorescence intensity, and the final H4F titration-calculated data are shown in Figure 5B,C. One of the H4F titration experiments is shown in Figure 5D. The 340 nm fluorescence decreases, and the emission peak significantly shifts from 340 to 350 nm. Both the raw data and the calculated data were used to construct ΔF340nm analysis. Because the nonlinear effect becomes more pronounced at high concentration, the data were fitted only up to 40 μM. The enzyme–product dissociation constant KDH4F of wildtype-ecDHFR to H4F was estimated as 6.05 ± 0.41 and 4.57 ± 0.47 μM with and without the deduction of the effect from H4F fluorescence, respectively. The compact substrate/product binding site resulting from the N-terminus tag weakens the product binding about 40-fold compared with the previous results.6,8 The inhibitor MTX exhibits no fluorescence, and one of the MTX titration experiments is shown in Figure 5E. The 340 nm fluorescence decreases dramatically, and the enzyme–inhibitor dissociation constant KDMTX in wildtype-ecDHFR was estimated as 42.6 ± 2.6 nM. Interestingly, this value is about 8-fold stronger than that reported previously, which is 362 nM in solution.19 Also, this value is 4-fold weaker than that reported by the studies of 9.5 nM on single-molecule studies.18 The inhibitor MTX and substrate H2F are similar in dimension but are different in the functional group. We assumed that the N-terminal tag affects the kon association rate of these two ligands similarly. However, upon binding, the influence of the N-terminal tag on the koff dissociation rate of these two ligands should be significantly different. In addition, the effects from the N-terminus tag might change upon binding to coverglass. As shown in Figure S5, the enzyme–product dissociation constants KDH4F of M42F-ecDHFR and T113V-ecDHFR were estimated as 44.01 ± 2.73 and 40.87 ± 2.23 μM, respectively. These two mutants show similar enzyme–product dissociation constants but with the H2F-to-H4F conversion of 98 and 74% (Figure 1), respectively. When the enzymatic reaction progressed, the concentration of H2F decreased, and the concentration of H4F increased. There is an affinity competition between substrate and product. In M42F-ecDHFR (KDH4F = 2.6 × KDH2F) and T113V-ecDHFR (KDH4F = 1.0 × KDH2F), competition on the binding of product and substrate contributes to the final conversion percentage. In M42F-ecDHFR, the enzyme binds to the substrate more tightly than the product, leading to higher productivity. In contrast, in T113V-ecDHFR, the enzyme binds to the substrate and the product with a similar tendency, leading to lower productivity.

Figure 5.

Figure 5

Open in a new tab

(A) Fluorescence spectra of H4F at various concentrations. (B) Concentration-fluorescent intensity dependence of H4F (black-hollow circle) was analyzed with polynomial fitting (black curve) labeled with the simulation data (red-solid circle). (C) H4F titration raw data (green square), the H4F-simulated fluorescence intensity (red-solid circle), and the final H4F titration-calculated data (blue square). (D) Fluorescence spectra from the product-binding titration experiments and the fluorescence spectrum of 100 μM H4F (black-thin line). With the increasing concentration of H4F, the decrease at 340 nm fluorescence was observed. The inset shows the KD analysis curves of two trials based on the deduction of the H4F effect describe in (A–C). (E) Fluorescence spectra from the inhibitor-binding titration experiments and the fluorescence spectrum of 15 μM MTX (black-thin line). The inset shows the KD analysis curves of four trials.

Conclusions

Based on enzyme activity assays, we confirmed that all enzymes could convert H2F to H4F and measured the extent of reaction under nonsaturation conditions thermodynamically. When the enzymatic reaction progressed, the concentration of H2F decreased, and the concentration of H4F increased. There is an affinity competition between the substrate and product in the presence of the cofactor. The dissociation constants KDH2F and KDH4F in wildtype ecDHFR are 2.84 and 6.05 μM, respectively. While [H2F] ≈ 15 μM and [H4F] ≈ 35 μM, ligands and 5 μM wildtype ecDHFR formed the same amount of the EH2F complex and EH4F complex in solution. However, the reaction reached about 96%, which indicated the affinity competition between the substrate and product is different in the presence of a cofactor. In the presence of the cofactor, the dissociation constant KDH4F would become much larger than KDH2F. In the previous studies, the dissociation constant of H4F increases about 110 times in the presence of the cofactor (Inline graphic).6

We selected residues at highly conserved positions Met-42, Thr-113, and Gly-121 and designed several mutations, including single-point and double-point mutations. The optical approaches with the FRET effect and fluorescence quenching upon ligand binding were demonstrated. We systematically measured the equilibrium thermodynamic dissociation constant for enzyme–cofactor, enzyme–substrate, and enzyme–product complex in wildtype ecDHFR and each mutant. In the M42F-ecDHFR mutant, we propose that the stereo-hindrance phenylalanine residue squeezes and affects the rotation and movement of helix αC, resulting in the weakening enzyme–ligand affinity. In the T113V-ecDHFR mutant, because the valine and threonine residues are similar in size, this mutation does not create significant space disturbance. However, valine lacks the hydroxyl group as threonine and affects the hydrogen network in the H2F binding, leading to decreasing H2F binding. Mutations at the Gly-121 position might affect the protein flexibility dependent on the type of replacement. In G12lV-ecDHFR mutation, such an effect leads to much weaker binding to the substrate. In the double-mutant-cycle analysis, double mutations show nonadditive properties upon either cofactor or substrate binding. Also, in general, the first point mutation strongly affects the ligand binding (ΔG1 > ΔG′1 and ΔG2 > ΔG′2) compared to the second one.

As shown in Figure 2, the N-terminus tag is an additional peptide at the N-terminus. It acts as a cap on the substrate entrance, pushes the adenosine binding subdomain to the helix αF, and influences the binding of NADPH in most mutants. If the mutation positions are away from the N-terminus tag and the adenosine binding subdomain, the additive effects due to the N-terminus tag were not observed. The bioseq-ecDHFR protein studied here provides insights into the impact of the N-terminus tag. Using a biotin-tag enzyme in fluorescence correlation spectroscopy studies is the main reason to have ecDHFR attached with the N-terminus tag. This biotinylated tag will be used to attach ecDHFR to the coverglass surface through avidin. The effects from the N-terminus tag might change upon binding to coverglass. As discussed previously, KDMTX (42.6 ± 2.6 nM) measured here is between the none-N-terminus tag studies in solution (362 nM)19 and the N-terminus tag studies on coverglass (9.5 nM).18

Materials and Methods

Plasmid Design and Protein Expression and Purification

The E. coli DHRF sequence (ecDHFR Uniprot: P0ABQ4) with the N-terminal insert containing a 19-residue lysine biotinylation signal sequence (MGLNDIFEAQKIEWHGGGT; bioseq)18 was subcloned into the BamHI and EcoRI sites of pRSETa plasmid. The designed pRESTa-bioseq-ecDHFR plasmid was constructed through DNA synthesis services by Genscript with DNA sequence confirmation. Due to the low expression level, the bioseq-ecDHFR DNA sequence was further subcloned into the BamHI and EcoRI sites of the pET28a plasmid. The pET28a-bioseq-ecDHFR construct was used to express ecDHFR as the previous report with some modifications.18 Briefly, E. coli BL21(DE3)pLysS cells containing the plasmid pET28a-bioseq-ecDHFR were grown at 37 °C in LB broth containing 50 mg/L kanamycin to the absorption of 0.6 at 600 nm and induced at 37 °C for 5–8 h with 0.3 mM isopropyl-β-d-1-thiogalactopyranoside. After harvesting, the 6Histag-bioseq-ecDHFR protein was purified by an ÄKTA prime plus liquid chromatography system with HisTrap HP columns (Ni Sepharose affinity resin) followed by a Hiprep-26/60 desalting column. With a gradient elution from 10 to 500 mM imidazole in 50 mM PBS buffer containing 500 mM NaCl at pH 8.0, the bioseq-ecDHFR protein was eluted between 150 and 255 mM imidazole. After desalting and dialysis, the 6Histag-bioseq-ecDHFR protein was stored at MTEN buffer consisting of 50 mM 2-(N-morpholino) ethane sulfonate (MES), 25 mM Tris, 100 mM NaCl, 0.1 mM EDTA, and 25 mM ethanolamine at 7.4 pH.20 All enzyme assays and ligand-binding assays were carried out within 3 days of storage in 4 °C.

Three highly conserved positions, Met-42 (M42), Thr-113 (T113), and Gly-121(G121), were selected. Site-directed mutagenesis steps were carried out with QuikChange Lightning Site-directed Mutagenesis kit (StrateGene) to engineer the following mutations: (1) single-point mutation: M42F-ecDHFR, T113V-ecDHFR, G121S-ecDHFR, and G121V-ecDHFR; (2) double-point mutation: M42F/G121S-ecDHFR, M42F/G121V-ecDHFR, T113V/G121S-ecDHFR, and T113V/G121V-ecDHFR. The ecDHFR mutants with a double-point mutation formed inclusion bodies, while they were induced at 37 °C. Therefore, all cell cultures were quickly cooled down and induced at 20 °C for 18–24 h with 0.3 mM isopropyl-β-d-1-thiogalactopyranoside.

Enzyme Activity Assays

The activity of ecDHFR was determined at 25 °C following the decrease of NADPH and 7,8-dihydrofolate (H2F) by the absorbance measurements at 340 nm using the Hitachi double beam spectrophotometer (U-3900) in MTEN buffer as described above. ecDHFR and NADPH were mixed to form an ENADPH–complex sample. The ENADPH–complex sample and H2F were then mixed to form a reaction solution with starting concentrations of 5 μM ecDHFR, 100 μM NADPH, and 50 μM H2F for carrying out the enzyme assays. The reaction solutions were kept at 25 °C, while the absorption measurements were performed for 15 min. Enzyme assays for wildtype and each mutant were carried out at least three times. Because ecDHFR and NADP+ exhibit no absorption at 340 nm, the initial point and the final point of the reaction were estimated with a mixture of 100 μM NADPH and 50 μM H2F and a mixture of 50 μM NADPH and 50 μM 5,6,7,8-tetrahydrofolate (H4F), respectively. The concentrations of compounds in solution were determined by UV absorbance using the following molar extinction coefficient: NADPH (ε340 = 6.22 mM–1 cm–1), NADP+(ε260 = 18 mM–1 cm–1), H2F (ε282 = 28 mM–1 cm–1), H4F (ε297 = 28 mM–1 cm–1), and MTX (ε302 = 22.1 mM–1 cm–1).21 All the reagent concentrations refer to the final concentrations in the reaction mixture unless otherwise specified. Ligands, such as NADPH (Sigma N5130), H2F (Sigma D7006), H4F (Sigma T3125), and MTX (Sigma M9929), were purchased from Sigma-Aldrich and used directly without further purification. The stationary absorption and fluorescence spectra were recorded with a spectrophotometer (U-3900, Hitachi High-Tech.) and a fluorimeter (FluoroMax4, HORIBA Jobin Yvon Inc).

Molecular Dynamics Simulation

Because there are additional histag and 19-residue lysine biotinylation signal sequences at the N-terminus of ecDHFR, to understand the effect of the extra N-terminus-tag on the ligand-binding affinity, we tracked the N-terminus tag folding process. We monitored the interaction between the N-terminus tag and the main ecDHFR protein. The starting structure of the ecDHFR was from PDB (PDB ID: 1RX1).5 In this structure, the reduced form NADPH binds to ecDHFR. The N-terminus tag was generated by protein structure prediction server (PS) version 3.0.22 Simulation parameters of the NADPH were generated using the program ACPYPE with Antechamber.23,24 All simulations were performed with the GROMACS25 version 2018, and the AMBER99SB force field was used in this study. The ecDHFR and N-terminus tag ecDHFR structures were in boxes, respectively, filled with 7096 and 35,616 TIP3P water molecules, and the buffer distance of 1.0 nm was used in molecular dynamics simulation. Sodium ions were added to both systems to neutralize the system charge, and PropKa was used to set the protonated state of all residues at pH 7.4. The long-range electrostatic interactions were computed by the smooth particle-mesh Ewald (PME) algorithm, while short-range electrostatic and van der Waals cutoffs were set to 1.4 nm.26,27

For each system, after carrying out the relaxation of the system by the steepest descent minimization, 200 ps simulation with protein restraints was conducted to equilibrate the solvent and ion positions around the proteins at constant temperature 300 K. Random initial velocities according to the Maxwell distribution at 300 K were generated for each run. The temperatures of 298 and 300 K displayed a negligible difference in the Maxwell distribution and no effect on the structure of ecDHFR. Both ecDHFR and N-terminus tag ecDHFR systems were then simulated for 1000 ns using a time step of 2 fs, while constraining all bond lengths with the P-LINCS algorithm.28 The temperature was kept at 300 K using the V-rescale thermostat,29 and the pressure was maintained at 1 bar using the Parrinello–Rahman barostat.30

Ligand-Binding Assays

The ecDHFR contains five intrinsic tryptophan residues (W22/W30/W47/W74/W133). When excited by 290 nm irradiation, ecDHFR is fluorescent and exhibits a fluorescence band of 300–460 nm with an emission peak at 340 nm. The emission spectrum of tryptophan and the absorption spectrum of NADPH are overlapped well, and five tryptophan residues are located within 16 Å to NADPH.5 Hence, ecDHFR and NADPH could be a Förster resonance energy transfer (FRET) pair. Other ligands, such as H2F, H4F, and methotrexate (MTX), could not form a FRET pair with tryptophan, but they show a quenching effect on the tryptophan fluorescence of ecDHFR. The 340 nm fluorescence intensity of ecDHFR at different enzyme concentrations was measured to perform all experiments in the linear relation range. The equilibrium thermodynamic dissociation constant (KD) was determined by titration with the increasing ligand concentration. We prepared an ecDHFR of 5 μM with various concentrations of the ligand range of 500 nM to 400 μM. The fluorescence spectra in each mixing sample were measured from 300 to 570 nm under the 290 nm irradiation. Each data set included at least three independent experiments, and each experiment is an average of at least two measurements under the same conditions.

Acknowledgments

This work was supported by grants from the Ministry of Science and Technology, Taiwan (MOST 101-2113-M-009-024-MY2, MOST 103-2113-M-009-012-MY2, and MOST 108-2113-M-009-022-) and the Center for Intelligent Drug Systems and Smart Bio-devices (IDS2 B) in Taiwan.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.1c02995.

  • Root-mean-square deviation of the backbone, comparison between ΔF340 nm and F 440 nm analyses in wildtype DHFR, cofactor-binding titration experiments on (A∼D) single-point mutation and (E∼H) double-point mutation, substrate-binding titration experiments on (A∼D) single-point mutation and (E∼H) double-point mutation, and KD analysis curves of M42F and T113V (PDF)

Author Present Address

Mailing address: Room 424, Jan-Qi Biomedical Engineering building, No. 75, Po-Ai Street, East District, Hsinchu, Taiwan (ZIP Code 300)

Author Contributions

∥ C.-H.H. and Y.-W.C. contributed equally to this work.

The authors declare no competing financial interest.

Notes

Accession ID for E. coli Dihydrofolate reductase: folA: P0ABQ4 (https://www.uniprot.org/uniprot/P0ABQ4).

Supplementary Material

ao1c02995_si_001.pdf (1.8MB, pdf)

References

  1. Schweitzer B. I.; Dicker A. P.; Bertino J. R. Dihydrofolate reductase as a therapeutic target. FASEB J. 1990, 4, 2441–2452. 10.1096/fasebj.4.8.2185970. [DOI] [PubMed] [Google Scholar]
  2. Raimondi M.; Randazzo O.; La Franca M.; Barone G.; Vignoni E.; Rossi D.; Collina S. DHFR inhibitors: reading the past for discovering novel anticancer agents. Molecules 2019, 24, 1140. 10.3390/molecules24061140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Matthews D.; Alden R.; Bolin J.; Freer S.; Hamlin R.; Xuong N.; Kraut J.; Poe M.; Williams M.; Hoogsteen K. Dihydrofolate reductase: x-ray structure of the binary complex with methotrexate. Science 1977, 197, 452–455. 10.1126/science.17920. [DOI] [PubMed] [Google Scholar]
  4. Bystroff C.; Kraut J. Crystal structure of unliganded Escherichia coli dihydrofolate reductase. Ligand-induced conformational changes and cooerativity in binding. Biochemistry 1991, 30, 2227–2239. 10.1021/bi00222a028. [DOI] [PubMed] [Google Scholar]
  5. Sawaya M. R.; Kraut J. Loop and subdomain movements in the mechanism of Escherichia coli dihydrofolate reductase: crystallographic evidence. Biochemistry 1997, 36, 586–603. 10.1021/bi962337c. [DOI] [PubMed] [Google Scholar]
  6. Cameron C. E.; Benkovic S. J. Evidence for a functional role of the dynamics of glycine-121 of Escherichia coli dihydrofolate reductase obtained from kinetic analysis of a site-directed mutant. Biochemistry 1997, 36, 15792–15800. 10.1021/bi9716231. [DOI] [PubMed] [Google Scholar]
  7. Agarwal P. K.; Billeter S. R.; Rajagopalan P. T. R.; Benkovic S. J.; Hammes-Schiffer S. Network of coupled promoting motions in enzyme catalysis. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 2794–2799. 10.1073/pnas.052005999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Rajagopalan P. T. R.; Lutz S.; Benkovic S. J. Coupling interactions of distal residues enhance dihydrofolate reductase catalysis: mutational effects on hydride transfer rates. Biochemistry 2002, 41, 12618–12628. 10.1021/bi026369d. [DOI] [PubMed] [Google Scholar]
  9. Rod T. H.; Radkiewicz J. L.; Brooks C. L. III Correlated motion and effect of distal mutations in dihydrofolate reductase. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 6980–6985. 10.1073/pnas.1230801100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Wong K. F.; Selzer T.; Benkovic S. J.; Hammes-Schiffer S. Impact of distal mutations on the network of coupled motions correlated to hydride transfer in dihydrofolate reductase. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 6807–6812. 10.1073/pnas.0408343102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Mauldin R. V.; Lee A. L. Nuclear magnetic resonance study of the role of M42 in the solution dynamics of Escherichia coli dihydrofolate reductase. Biochemistry 2010, 49, 1606–1615. 10.1021/bi901798g. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Boehr D. D.; Schnell J. R.; McElheny D.; Bae S.-H.; Duggan B. M.; Benkovic S. J.; Dyson H. J.; Wright P. E. A distal mutation perturbs dynamics amino acid networks in dihydrofolate reductase. Biochemistry 2013, 52, 4605–4619. 10.1021/bi400563c. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Singh P.; Sen A.; Francis K.; Kohen A. Extension and limits of the network of coupled motions correlated to hydride transfer in dihydrofolate reductase. J. Am. Chem. Soc. 2014, 136, 2575–2582. 10.1021/ja411998h. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Villafranca J. E.; Howell E. E.; Voet D. H.; Strobel M. S.; Ogden R. C.; Abelson J. N.; Kraut J. Direct mutagenesis of dihydrofolate reductase. Science 1983, 222, 782–788. 10.1126/science.6356360. [DOI] [PubMed] [Google Scholar]
  15. Howell E. E.; Villafranca J. E.; Warren M. S.; Oatley S. J.; Kraut J. Functional role of aspartic acid-27 in dihydrofolate reductase revealed by mutagenesis. Science 1986, 231, 1123–1128. 10.1126/science.3511529. [DOI] [PubMed] [Google Scholar]
  16. Appleman J. R.; Howell E. E.; Kraut J.; Blakley R. L. Role of aspartate 27 of dihydrofolate reductase from Escherichia coli in interconversion of active and inactive enzyme conformers and binding of NADPH. J. Biol. Chem. 1990, 265, 5579–5584. 10.1016/s0021-9258(19)39400-1. [DOI] [PubMed] [Google Scholar]
  17. Liu C. T.; Francis K.; Layfield J. P.; Huang X.; Hammes-Schiffer S.; Kohen A.; Benkovic S. J. Escherichia coli dihydrofolate reductase catalyzed proton and hydride transfers: temporal order and the roles of Asp27 and Tyr100. Proc. Natl. Acad. Sci. U.S.A. 2014, 111, 18231–18236. 10.1073/pnas.1415940111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Rajagopalan P. T. R.; Zhang Z.; McCourt L.; Dwyer M.; Benkovic S. J.; Hammes G. G. Interaction of dihydrofolate reductase with methotrexate: ensemble and single-molecule kinetics. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 13481–13486. 10.1073/pnas.172501499. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Fierke C. A.; Benkovic S. J. Probing the functional role of threonine-113 of Escherichia coli dihydrofolate reductase for its effect on turnover efficiency, catalysis, and binding. Biochemistry 1989, 28, 478–486. 10.1021/bi00428a011. [DOI] [PubMed] [Google Scholar]
  20. Williams J. W.; Morrison J. F.; Duggleby R. G. Methotrexate, a high-affinity pseudosubstrate of dihydrofolate reductase. Biochemistry 1979, 18, 2567–2573. 10.1021/bi00579a021. [DOI] [PubMed] [Google Scholar]
  21. Miller G. P.; Benkovic S. J. Deletion of a highly motional residue affects formation of Michaelis complex for Escherichia coli dihydrofolate reductase. Biochemistry 1998, 37, 6327–6335. 10.1021/bi972922t. [DOI] [PubMed] [Google Scholar]
  22. Huang T. T.; Hwang J. K.; Chen C. H.; Chu C. S.; Lee C. W.; Chen C. C. (PS)2: protein structure prediction server version 3.0. Nucleic Acids Res. 2015, 43, W338–W342. 10.1093/nar/gkv454. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Sousa da Silva A. W.; Vranken W. F. ACPYPE-AnteChamber PYthon Parser interfacE. BMC Res. Notes 2012, 5, 367. 10.1186/1756-0500-5-367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Wang J.; Wang W.; Kollman P. A.; Case D. A. Automatic atom type and bond type perception in molecular mechanical calculations. J. Mol. Graph. Model. 2006, 25, 247–260. 10.1016/j.jmgm.2005.12.005. [DOI] [PubMed] [Google Scholar]
  25. Abraham M. J.; Murtola T.; Schulz R.; Páll S.; Smith J. C.; Hess B.; Lindahl E. GROMACS: high performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX 2015, 1–2, 19–25. 10.1016/j.softx.2015.06.001. [DOI] [Google Scholar]
  26. Darden T.; York D.; Pedersen L. Particle mesh Ewald : an N·Log(N) method for Ewald sums in large systems. J. Chem. Phys. 1993, 98, 10089–10092. 10.1063/1.464397. [DOI] [Google Scholar]
  27. Essmann U.; Perera L.; Berkowitz M. L.; Darden T.; Lee H.; Pedersen L. G. A smooth particle mesh Ewald method. J. Chem. Phys. 1995, 103, 8577–8593. 10.1063/1.470117. [DOI] [Google Scholar]
  28. Hess B. P-LINCS: A parallel linear constraint solver for molecular simulation. J. Chem. Theory Comput. 2008, 4, 116–122. 10.1021/ct700200b. [DOI] [PubMed] [Google Scholar]
  29. Bussi G.; Donadio D.; Parrinello M. Canonical sampling through velocity rescaling. J. Chem. Phys. 2007, 126, 014101. 10.1063/1.2408420. [DOI] [PubMed] [Google Scholar]
  30. Parrinello M.; Rahman A. Polymorphic transitions in single crystals: a new molecular dynamics method. J. Appl. Phys. 1981, 52, 7182–7190. 10.1063/1.328693. [DOI] [Google Scholar]
  31. Zhang L.; Kao Y.-T.; Qiu W.; Wang L.; Zhong D. Femtosecond studies of tryptophan fluorescence dynamics in proteins: local solvation and electronic quenching. J. Phys. Chem. B 2006, 110, 18097–18103. 10.1021/jp063025e. [DOI] [PubMed] [Google Scholar]
  32. Dunn S. M. J.; King R. W. Kinetics of ternary complex formation between dihydrofolate reductase, coenzyme, and inhibitors. Biochemistry 1980, 19, 766–773. 10.1021/bi00545a024. [DOI] [PubMed] [Google Scholar]
  33. Gafni A.; Brand L. Fluorescence decay studies of reduced nicotinamide adenine dinucleotide in solution and bound to liver alcohol dehydrogenase. Biochemistry 1976, 15, 3165–3171. 10.1021/bi00660a001. [DOI] [PubMed] [Google Scholar]
  34. Townley H. E.; Sessions R. B.; Clarke A. R.; Dafforn T. R.; Griffiths W. T. Protochlorophyllide oxidoreductase: A homology model examined by site-directed mutagenesis. Protein 2001, 44, 329–335. 10.1002/prot.1098. [DOI] [PubMed] [Google Scholar]
  35. Fjeld C. C.; Birdsong W. T.; Goodman R. H. Differential binding of NAD+ and NADH allows the transcriptional corepressor carboxyl-terminal binding protein to serve as a metabolic sensor. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 9202–9207. 10.1073/pnas.1633591100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Chen S.; Fahmi N. E.; Bhattacharya C.; Wang L.; Jin Y.; Benkovic S. J.; Hecht S. M. Fluorescent biphenyl derivatives of phenylalanine suitable for protein modification. Biochemistry 2013, 52, 8580–8589. 10.1021/bi401275v. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Ohmae E.; Fukumizu Y.; Iwakura M.; Gekko K. Effects of mutation at methionine-42 of Escherichia coli dihydrofolate reductase on stability and function: implication of hydrophobic interactions. J. Biochem. 2005, 137, 643–652. 10.1093/jb/mvi079. [DOI] [PubMed] [Google Scholar]
  38. Carter P. J.; Winter G.; Wilkinson A. J.; Fersht A. R. The use of double mutants to detect structural changes in the active site of the tyrosyl-tRNA synthetase (Bacillus stearothermophilus). Cell 1984, 38, 835–840. 10.1016/0092-8674(84)90278-2. [DOI] [PubMed] [Google Scholar]
  39. Fersht A. R.; Wells T. N. C. Linear free energy relationships in enzyme binding interactions studied by protein engineering. Protein Eng. 1991, 4, 229–231. 10.1093/protein/4.3.229. [DOI] [PubMed] [Google Scholar]
  40. Pagano L.; Toto A.; Malagrinò F.; Visconti L.; Jemth P.; Gianni S. Double mutant cycles as a tool to address folding, binding, and allostery. Int. J. Mol. Sci. 2021, 22, 828. 10.3390/ijms22020828. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Mhashal A. R.; Pshetitsky Y.; Eitan R.; Cheatum C. M.; Kohen A.; Major D. T. Effect of Asp122 mutation on the hydride transfer in E. coli DHFR demonstrates the goldilocks of enzyme flexibility. J. Phys. Chem. B 2018, 122, 8006–8017. 10.1021/acs.jpcb.8b05556. [DOI] [PubMed] [Google Scholar]
  42. Chen J.-T.; Mayer R. J.; Fierke C. A.; Benkovic S. J. Site-specific mutagenesis of dihydrofolate reductase from Escherichia coli. J. Cell. Biochem. 1985, 29, 73–82. 10.1002/jcb.240290203. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ao1c02995_si_001.pdf (1.8MB, pdf)

Articles from ACS Omega are provided here courtesy of American Chemical Society

RESOURCES