An active site mutation induces oxygen reactivity in D-arginine dehydrogenase: A case of superoxide diverting protons

Abstract

Enzymes are potent catalysts that increase biochemical reaction rates by several orders of magnitude. Flavoproteins are a class of enzymes whose classification relies on their ability to react with molecular oxygen (O₂) during catalysis using ionizable active site residues. Pseudomonas aeruginosa D-arginine dehydrogenase (PaDADH) is a flavoprotein that oxidizes D-arginine for P. aeruginosa survival and biofilm formation. The crystal structure of PaDADH reveals the interaction of the glutamate 246 (E²⁴⁶) side chain with the substrate and at least three other active site residues, establishing a hydrogen bond network in the active site. Additionally, E²⁴⁶ likely ionizes to facilitate substrate binding during PaDADH catalysis. This study aimed to investigate how replacing the E²⁴⁶ residue with leucine affects PaDADH catalysis and its ability to react with O₂ using steady-state kinetics coupled with pH profile studies. The data reveal a gain of O₂ reactivity in the E²⁴⁶L variant, resulting in a reduced flavin semiquinone species and superoxide (O₂^•ˉ) during substrate oxidation. The O₂^•ˉ reacts with active site protons, resulting in an observed nonstoichiometric slope of 1.5 in the enzyme’s log (k_cat/K_m) pH profile with D-arginine. Adding superoxide dismutase results in an observed correction of the slope to 1.0. This study demonstrates how O₂^•ˉ can alter the slopes of limbs in the pH profiles of flavin-dependent enzymes and serves as a model for correcting nonstoichiometric slopes in elucidating reaction mechanisms of flavoproteins.

Flavin-dependent enzymes are a class of enzymes that often rely on ionization processes during catalysis (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15). Flavoproteins can be classified as oxidases, monooxygenases, and dehydrogenases, depending on their ability to use molecular oxygen (O₂) as an electron acceptor and the products of their reactions (16, 17). Oxidases use O₂ as an electron acceptor to produce H₂O₂. Monooxygenases insert an oxygen atom from O₂ into the substrate with H₂O as a product. Dehydrogenases do not react with O_2, or if they do, they reduce O₂ to superoxide radicals (O₂^•ˉ) without the production of H₂O₂ or H₂O (18). In flavin-dependent enzymes, the spin-forbidden reaction of the triplet state O₂ with the singlet state reduced flavin is overcome by successive electron transfers in a step-wise process to yield a caged O₂^•ˉ/flavin semiquinone radical pair (16, 17, 19, 20, 21). Studies on flavoproteins show that enzymes that react with O₂, such as oxidases and monooxygenases, overcome the thermodynamically unfavorable generation of the O₂^•ˉ/flavin semiquinone radical pair by stabilizing the transition state of the O₂^•ˉ/flavin semiquinone radical pair through electrostatic catalysis using a positive charge close to the flavin C(4a) atom (16, 17, 21, 22, 23, 24).

From structural and mechanistic analysis, a structural motif comprising nonpolar residues in the active site and a positive charge, either in the protein or from the enzyme’s substrate or product, has been identified as a requirement for flavin reactivity with O₂ to yield the O₂^•ˉ/flavin semiquinone radical pair (24). In principle, a flavoprotein can gain the ability to react with O₂, provided these minimum requirements are met. The highly reactive O₂^•ˉ can then undergo an ionization process to acquire a proton from the enzyme or bulk solvent, yielding a hydroperoxyl radical (HO₂^•) with a pK_a value of 4.8 (25, 26). The key players of ionization processes in several enzymes are solvent water, metal ions, and the side chains of ionizable active site residues (10, 16, 17, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36). For several decades, pH profile and mutagenesis studies have been widely employed to identify and assign pK_a values to the ionizing groups during enzyme catalysis (37, 38, 39, 40, 41, 42, 43). However, the pH profiles used to identify ionizing groups and assign pK_a values could be misinterpreted if an enzyme produces highly reactive species such as O₂^•ˉ that can react with ionized protons during catalysis. Hence, the question remains whether the formation and leakage of O₂^•ˉ in flavoproteins that react poorly with O₂ to yield O₂^•ˉ can affect the pH profiles used in assessing the residues important for enzyme catalysis.

The Pseudomonas aeruginosa D-arginine dehydrogenase (PaDADH) is a flavoprotein found in the opportunistic pathogen P. aeruginosa (44). The enzyme is important for P. aeruginosa’s utilization of D-arginine as a primary source of carbon and nitrogen by oxidizing the CN bond between the Cα atom and the amino group of its substrates (45). In recent years, the increased prevalence of multidrug-resistant strains of P. aeruginosa has been largely associated with the formation of bacterial biofilms coupled with bacterial antibiotic-resistance mechanisms (46, 47, 48, 49, 50, 51, 52, 53, 54). Studies have identified that increased levels of L-arginine in P. aeruginosa reduce bacterial mobility, enhancing the formation of bacterial biofilms required for chronic infections (53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64). In P. aeruginosa, PaDADH exists in a two-enzyme conversion system with L-arginine dehydrogenase, allowing the bacterium to accumulate L-arginine from D-arginine oxidation (65). Thus, PaDADH’s activity influences P. aeruginosa to favor the sessile biofilm-forming lifestyle required for chronic infections (63, 64). More recently, the enzyme has emerged as a paradigm for flavin-containing enzymes that oxidize positively charged D-amino acids (66).

PaDADH has broad substrate specificity. The enzyme oxidizes all D-amino acids, except D-aspartate and D-glutamate, to their corresponding α-imino acids, followed by nonenzymatic hydrolysis of the α-imino acid products to yield α-keto acids and ammonia (Fig. 1) (44, 67, 68, 69). PaDADH is a strict dehydrogenase that does not react with O₂ and follows a ping-pong bi-bi steady-state kinetic mechanism (44, 66). Structural analysis of the enzyme’s active site reveals only polar amino acid residues, including H⁴⁸, Y⁵³, E⁸⁷, R²²², E²⁴⁶, Y²⁴⁹, and R³⁰⁵, and a lack of nonpolar residues (69). The enzyme contains four flexible loops, L1, L2, L3, and L4 (69), with residues Y⁵³ of loop L1 and E²⁴⁶ of loop L2 involved in a gating interaction that secures the substrate in the active site for catalysis (Fig. 2) (67, 69). Residue E²⁴⁶ of loop L2 is located 7.6 Å from the flavin N5 atom, 8.2 Å from the flavin C(4a) atom, and does not participate in flavin reduction (70).

Figure 1.

General reaction scheme of PaDADH.PaDADH, Pseudomonas aeruginosa D-arginine dehydrogenase.

Figure 2.

Gating of PaDADH active site by loop L1 and L2 residues Y⁵³and E²⁴⁶. Y⁵³ and E²⁴⁶ are shown in gray. All N atoms are shown in blue, and all O atoms are in red. The FAD cofactor is represented by its isoalloxazine ring with the C atoms in gold. IAR represents the iminoarginine product and is shown in cyan. The E²⁴⁶ residue’s hydrogen bond interaction with Y⁵³ and distance from the flavin are shown as dashed lines. Loop L2 is shown in coral, and loop L1 is shown in green. The PDB file 3NYE was visualized and analyzed using the UCSF Chimera software (110). PaDADH, Pseudomonas aeruginosa D-arginine dehydrogenase.

In this study, the PaDADH residue E²⁴⁶ has been mutated to leucine to yield the glutamate 246 to leucine mutant (E²⁴⁶L) variant enzyme. The effects of the E²⁴⁶L mutation on the enzyme’s catalysis and ability to react with O₂ have been investigated using rapid-reaction kinetics, steady-state kinetics, and pH profile studies. Additionally, the mutation’s effects on the slopes of the pH profile plots have been explored.

Results

Rapid-reaction kinetics of the PaDADH E²⁴⁶L variant enzyme with D-leucine as a substrate

Since ∼80% of the flavin reduction of the PaDADH E²⁴⁶L variant enzyme with D-arginine occurs in the mixing time (2.2 ms) of the stopped-flow spectrophotometer (42, 71), the flavin reduction of the enzyme was investigated with D-leucine as the reducing substrate. The time-resolved aerobic reduction of the PaDADH E²⁴⁶L variant enzyme by D-leucine was investigated under pseudo-first-order conditions by monitoring the loss of the oxidized flavin’s absorbance at 446 nm. The resulting stopped-flow traces showed three distinct reaction phases at [D-leucine] ≤10 mM. As expected for flavin reduction, the absorbance at 446 nm decreased; however, there was a subsequent transient increase in the 446 nm absorbance (Fig. 3A). The stopped-flow traces were fit with triple exponentials (Equation 1).

Figure 3.

Aerobic reductive-half reaction of the PaDADH E²⁴⁶L variant enzyme.A, stopped-flow traces of the absorbance changes at 446 nm at different D-leucine concentrations (1–25 mM) fit with Equation 1. Each trace is the average of triplicate runs at each substrate concentration. For clarity, one out of every 100 experimental points is shown (vertical lines). Note the log time scale. B, time-resolved UV-visible absorption spectra of the various flavin species generated during the aerobic reduction of the PaDADH E²⁴⁶L variant enzyme with D-leucine. The reaction was monitored over 120 s (s) upon mixing 1 mM D-leucine with the E²⁴⁶L variant enzyme of PaDADH in the presence of atmospheric O₂. The black spectrum represents the oxidized enzyme; the red spectrum represents the reduced flavin semiquinone; the green and the blue spectra represent the fully reduced flavin. C, time map for the various flavin species generated during the aerobic reductive-half reaction of the PaDADH E²⁴⁶L variant enzyme with 1 mM D-leucine. D, the observed rate constant for flavin reduction as a function of D-leucine concentration under aerobic conditions fit with Equation 2. The single point shown at each substrate concentration is the k_obs value obtained from the fit of the average of triplicate runs with Equation 1 yielding an error of ≤5%. The assay was performed in 20 mM NaPP_i, pH 10.0, using an SF-61DX2 Hi-Tech KinetAsyst high-performance stopped-flow spectrophotometer thermostated at 25 ^oC and equipped with a photomultiplier detector under aerobic conditions. The instrumental dead time is 2.2 ms. E246L, glutamate 246 to leucine mutant; PaDADH, Pseudomonas aeruginosa D-arginine dehydrogenase.

Analysis of the time-resolved UV-visible absorption spectra of the various flavin species generated during the aerobic reduction of the PaDADH E²⁴⁶L variant enzyme with D-leucine showed an oxidized flavin spectrum at 0.01 s with a λ_max at 446 nm and a second peak at 367 nm. After 1.2 s, there was an observed quenching of the λ_{446 nm} and λ_{367 nm} peaks to yield a spectrum with a λ_max at 367 nm, consistent with a reduced flavin being present. After 8 s, there was an observed slight increase in the reduced flavin absorbance between 367 nm and 500 nm, although the overall spectral characteristics of the reduced flavin at 1.2 s were maintained. After 120 s, there were three observed peaks in the flavin spectrum at 367 nm, 394 nm, and 480 nm, consistent with the formation of a reduced flavin semiquinone species (Fig. 3B).

The first phase of the time-resolved aerobic reduction of the PaDADH E²⁴⁶L variant with D-leucine, characterized by the bleaching of the oxidized flavin absorption at 446 nm, was assigned to flavin reduction. The next phase, characterized by the transient gain of absorbance at 446 nm, was assigned to the imino acid product release. The last phase, showing a quenching of absorbance at 446 nm, was assigned to the reaction of the reduced flavin with O₂ (Fig. 3C), hinting at the formation of a caged O₂^•ˉ/flavin semiquinone radical pair during the aerobic reduction of the PaDADH E²⁴⁶L variant enzyme. The lack of a transient increase in the 446 nm absorbance at [D-leucine] ≥10 mM can be explained as a likely formation of enzyme–substrate complexes between the reduced enzyme and excess substrate following the imino acid product release, which prevents the reduced flavin from reacting with O₂.

To obtain the observed rate constant (k_obs1) for flavin reduction, the rate constants for the first phase at any given substrate concentration were fit with Equation 2, yielding a zero y-intercept hyperbolic dependence of the k_obs1 parameter on D-leucine concentration (Fig. 3D), allowing for the determination of the limiting rate constant for flavin reduction (k_red) and the apparent equilibrium constant for the dissociation of the substrate from the Michaelis complex (K_d). The resulting kinetic parameters are shown in Table 1. There was no observed dependence of the k_obs2 and k_obs3 parameters on [D-leucine].

Table 1.

Rapid-reaction kinetic parameters of the PaDADH E²⁴⁶L variant enzyme with D-leucine as substrate

Kinetic parameter	aAerobic	bAnaerobic
kred (s−1)	54 ± 3	35 ± 1
Kd (mM)	12 ± 2	12 ± 1
kred/Kd (M−1s−1)	4600 ± 400	2900 ± 75

^aReductive-half reaction kinetics were measured at varying concentrations of D-leucine under aerobic conditions. Assays were performed in 20 mM NaPP_i, pH 10.0, at 25 ^oC. The kinetic parameters’ values were obtained after fitting the kinetic data with Equation 2.

^bPreviously reported data for the PaDADH E²⁴⁶L variant enzyme under anaerobic conditions (70).

O₂ reactivity studies of the PaDADH E²⁴⁶L variant enzyme

The ability of the PaDADH E²⁴⁶L variant enzyme to react with O₂ as an electron acceptor was explored by measuring the initial velocities of the enzymatic reaction at various enzyme concentrations with 25 mM D-leucine as substrate and O₂ as the electron acceptor. The data yielded a positive-sloped linear dependence of the initial velocities of enzyme activity as a function of enzyme concentration. Similar results were obtained when the assay was repeated with D-arginine as substrate (Fig. 4A).

Figure 4.

O₂reactivity studies of the PaDADH E²⁴⁶L variant enzyme.A, plot of the initial velocity of the PaDADH E²⁴⁶L variant enzyme’s reactivity with O₂ as a function of enzyme concentration at fixed 5 mM D-arginine in black and 25 mM D-leucine in purple. B, the dependence of the rate of O₂ reactivity of the PaDADH E²⁴⁶L variant enzyme as a function of D-arginine concentration. C, the effect of superoxide dismutase on the PaDADH E²⁴⁶L variant enzyme’s reaction with O₂ with 5 mM D-arginine as substrate. The black trace represents the O₂ reactivity without superoxide dismutase. The blue trace represents the experiment performed in the presence of superoxide dismutase. The assays were carried out in 20 mM NaPP_i, pH 8.0, using a Clark-type oxygen electrode system, thermostated at 25 ^oC. E246L, glutamate 246 to leucine mutant; PaDADH, Pseudomonas aeruginosa D-arginine dehydrogenase.

A separate experiment to investigate the dependence of the O₂ reactivity on D-arginine concentration yielded a flat line with a rate constant for the oxygen-driven dehydrogenase activity† of 0.23 s⁻¹, irrespective of the substrate concentration tested (Fig. 4B). When superoxide dismutase was introduced into the reaction assay to determine the O₂ species generated during the enzyme’s turnover with O₂ as the electron acceptor, there was an observed decrease in the initial velocity of the enzymatic reaction with D-arginine (Fig. 4C). Similar results were obtained when the above assays were repeated with D-leucine as substrate (data not shown).

The steady-state kinetic studies of the PaDADH E246L variant enzyme reported in a previous study yielded a k_cat value of 265 ± 5 s⁻¹, a k_cat/K_m value of 871,000 ± 35,000 M⁻¹s⁻¹, and a K_m value of 0.30 ± 0.02 mM with D-arginine (70). Hence, to determine the effect of the enzyme-generated O₂^•ˉ on PaDADH E²⁴⁶L turnover with D-arginine, assays of the variant enzyme with O₂ as electron acceptor and 0.5 mM D-arginine as substrate were carried out under steady-state conditions. The PaDADH E²⁴⁶L variant underwent multiple turnover cycles spanning 2 h upon reintroduction of D-arginine into the reaction mixture (Fig. 5). For all enzyme turnover cycles, the initial rate of the oxygen-driven dehydrogenase activity was ∼0.4 s⁻¹. Similar rates for D-arginine oxidation were observed when the experiment was repeated using 0.15 and 0.24 mM D-arginine (data not shown).

Figure 5.

Effect of the PaDADH E²⁴⁶L-generated O₂^•ˉ on the turnover ability of the enzyme. O₂ consumption and regeneration cycles were carried out with fixed 6 μM enzyme and 0.5 mM d-arginine. The steady-state regions are shown in black dotted lines and the initial rates of D-arginine oxidation for each turnover cycle is shown below as v_o/e. The assay was performed in 20 mM NaPP_i, pH 8.0, with O₂ as the electron acceptor, using a Clark-type oxygen electrode system at 25 ^oC. E246L, glutamate 246 to leucine mutant; PaDADH, Pseudomonas aeruginosa D-arginine dehydrogenase.

Additionally, O₂ regeneration was observed in every enzyme turnover cycle before the D-arginine substrate was reintroduced (Fig. 5). The observed oxygen regeneration can be explained by a proton-dependent O₂^•ˉ disproportionation to yield hydroperoxide anion HO₂ˉ, hydroxide HOˉ, and O₂ (Fig. 6) (72), followed by a slow re-equilibration of the solution in the O₂ electrode chamber toward atmospheric oxygen. The observed nonzero O₂ level after enzyme turnover with D-arginine can be explained by the O₂^•ˉ disproportionation reaction leading to O₂ accumulation that gradually overturns the O₂ consumption of the D-arginine oxidation.

Figure 6.

The disproportionationof O2•ˉ in water.

Effects of pH on the k_cat/K_m parameter of PaDADH E²⁴⁶L

Due to the O₂ reactivity and formation of the flavin semiquinone species at [D-leucine] ≤10 mM during the aerobic reduction of the PaDADH E²⁴⁶L variant, the kinetic investigations for the steady-state pH effects focused only on the k_cat/K_m parameter, which reports on the enzyme’s behavior at low substrate concentrations and probes the free enzyme. Thus, to understand the effects of pH on PaDADH E²⁴⁶L’s substrate capture, the steady-state reactions of the enzyme were investigated with D-arginine or D-leucine as the substrate from pH 5.0 to 10.5, with phenazine methosulfate (PMS) as an artificial electron acceptor since the physiological electron acceptor for PaDADH activity is not known. The plots of the log values of the k_cat/K_m parameter showed an increase in the k_cat/K_m parameter with increasing pH, a pH-independent region above pH 9.0, and an observed pK_a value for a basic group between 8.2 and 8.8 (Fig. 7 and Table 2).

Figure 7.

Effects of superoxide dismutase and pH on the k_cat/K_mparameter of the PaDADH E²⁴⁶L variant with D-arginine or D-leucine as substrate.A, pH dependence of the k_cat/K_m parameter with D-arginine. B, pH dependence of the k_cat/K_m parameter with D-leucine. Activity assays were carried out with varying concentrations of D-arginine or D-leucine as a substrate and fixed PMS as an artificial electron acceptor at 1 mM from pH 5.0 to 10.5 in 20 mM NaPP_i. Assays without superoxide dismutase are shown in black and red for D-arginine at 25 ^oC and 12 ^oC, respectively, and purple for D-leucine at 25 ^oC. The D-arginine assay with superoxide dismutase at 25 ^oC is shown in blue. The D-arginine plots were obtained by fitting the kinetic data to Equation 3 for the k_cat/K_m parameter without superoxide dismutase at 12 ^oC and 25 ^oC, and Equation 4 for the k_cat/K_m parameter with superoxide dismutase. The D-leucine plot was obtained by fitting the kinetic data with Equation 4. The observed pK_a values for the various assays are recorded in Table 2 for D-arginine and Table 3 for D-leucine. PaDADH, Pseudomonas aeruginosa D-arginine dehydrogenase; PMS, phenazine methosulfate.

Table 2.

Effects of superoxide dismutase on the pH effects on the steady-state kinetic parameters of the PaDADH E²⁴⁶L variant enzyme with D-arginine as substrate^a

kcat/Km
Condition	CH	pKa	Slope	R2
25 oC	665,000 ± 78,000	8.2 ± 0.1	1.5 ± 0.1	0.997
25 oC + superoxide dismutase	263,000 ± 80,000	8.8 ± 0.2	1.1 ± 0.1	0.983
12 oC	531,000 ± 64,000	8.5 ± 0.1	1.4 ± 0.1	0.997

^aEnzymatic activities were measured at varying concentrations of D-arginine and fixed 1 mM PMS. Reactions were carried out in 20 mM sodium pyrophosphate. The k_cat/K_m parameter values were obtained after fitting the kinetic data with Equation 3.

For the D-arginine substrate, the k_cat/K_m pH profile at 25 ^oC fit with Equation 3 yielded a plot with a nonstoichiometric slope of +1.5 for the increasing limb, with an observed pK_a value for a basic group of ∼8.2. When the assay was repeated at 12 ^oC to investigate the effect of temperature on the slope of the k_cat/K_m pH profile plot, similar results were obtained, suggesting that temperature does not contribute to the observed slope. When the effect of the enzyme-generated O₂^•ˉ on the slope of the k_cat/K_m pH profile plot was investigated by adding 200 to 500 units of superoxide dismutase to each apparent steady-state reaction mixture at 25 ^oC, the k_cat/K_m pH profile plot fit with Equation 3 yielded a stoichiometric value of +1 (Table 2).

The effect of pH on the k_cat/K_m pH profile of PaDADH E²⁴⁶L with D-leucine at 25 ^oC fit with Equation 4 yielded a kinetic plot with a slope of +1 for the increasing limb, with an observed pK_a value for a basic group of ∼8.7 (Table 3) and a pH-independent region above pH 9.0 (Fig. 7).

Table 3.

Effects of pH on the steady-state kinetic parameters of the PaDADH E²⁴⁶L variant enzyme with D-leucine as substrate at 25 ^oC^a

Kinetic parameter	CH	pKa	R2
kcat/Km	2200 ± 500	8.7 ± 0.1	0.987

^aEnzymatic activities were measured at varying concentrations of D-leucine and fixed 1 mM PMS. Reactions were carried out in 20 mM sodium pyrophosphate. The k_cat/K_m parameter values were obtained after fitting the kinetic data with Equation 4.

Discussion

This study aimed to investigate the effects of the E²⁴⁶L mutation on the ability of PaDADH to react with O₂. The study also investigated the effect of the E²⁴⁶L mutation on the catalysis of PaDADH using steady-state kinetics coupled with pH profile studies. The data demonstrate that upon the E²⁴⁶L mutation, PaDADH, which is a strict dehydrogenase that does not react with O₂, gains the ability to react with O₂, although poorly, to yield a reduced flavin semiquinone and O₂^•ˉ (66, 67). Consequently, the PaDADH E²⁴⁶L variant turns over with O₂ as an electron acceptor through an alternative dehydrogenase pathway. Following the acquired O₂ reactivity, the variant enzyme yields a nonstoichiometric slope in the plot of the log (k_cat/K_m) parameter as a function of pH with D-arginine as substrate. Details on the gain of function and the implications on PaDADH catalysis are discussed below.

The PaDADH E²⁴⁶L variant enzyme reacts with O₂ to form a flavin semiquinone during substrate oxidation. Evidence supporting this conclusion comes from the enzyme-monitored aerobic flavin reduction with D-leucine and oxygen (Fig. 3), showing the formation and decay of an enzyme intermediate between 1.2 and 120 s. Analysis of the spectral data of the flavin reduction revealed a flavin spectrum with a peak at 367 nm, characteristic of a flavin semiquinone species at 120 s (Fig. 3) (73, 74, 75). By comparison, this feature is not observed with the PaDADH wildtype enzyme under both aerobic and anaerobic conditions (42, 44, 66, 67, 69, 76, 77). In the same way, the flavin semiquinone species was not observed with the E²⁴⁶L variant enzyme under anaerobic conditions, although similar kinetic parameters as reported for the aerobic flavin reduction in this study were observed: k_red = ∼50 s⁻¹ and K_d = 12 mM (Table 1) (70). The data are consistent with the PaDADH E²⁴⁶L variant reacting with O₂ to yield the flavin semiquinone after flavin reduction and product release, as evidenced by the observed transient increase in the λ_{466 nm} absorbance of the reduced flavin at 8 s (Fig. 3). The presence or absence of the product in the active site alters the flavin oscillator strength, yielding different absorbance intensities of the flavin 446 nm peak upon product release (78). The observation that the flavin semiquinone was not formed at [D-leucine] ≥10 mM can be explained as a likely binding of the reduced enzyme to excess substrate. At [D-leucine] ≥10 mM, the reduced enzyme likely forms a complex with free unreacted substrate in the bulk solvent to yield the E_redS complex. Such a complex renders the reduced flavin unavailable, preventing the free reduced flavin from reacting with O₂ to yield the semiquinone.

The reactivity of the PaDADH E²⁴⁶L variant enzyme with O₂ yields O₂^•ˉ during catalysis. Evidence supporting this conclusion comes from the rapid-reaction kinetics with D-leucine and the steady-state enzymatic assay of the PaDADH E²⁴⁶L variant enzyme with O₂ as the electron acceptor with D-arginine or D-leucine as substrate, with and without superoxide dismutase (Figs. 3 and 4). The plot of the initial velocity of the E²⁴⁶L variant enzyme as a function of time showed a decrease in O₂ consumption in the presence of superoxide dismutase (Fig. 4), consistent with a superoxide dismutase–mediated conversion of the PaDADH E²⁴⁶L-generated O₂^•ˉ to O₂, which overturns O₂ consumption (25, 72, 75, 79, 80, 81, 82, 83, 84, 85, 86, 87). The formation of O₂^•ˉ is concomitant with the observed stabilization of a flavin semiquinone species (vide supra) during the aerobic reduction of PaDADH E²⁴⁶L to yield the O₂^•ˉ/flavin semiquinone radical pair. The O₂^•ˉ/flavin semiquinone radical pair likely arises from a one-electron reduction of O₂ by the reduced flavin required to overcome the spin-forbidden reaction of the triplet state O₂ with the singlet state reduced flavin during PaDADH E²⁴⁶L reactivity with O₂, as reported for flavin-dependent enzymes (16, 17, 19, 20, 21, 75, 88).

The PaDADH E²⁴⁶L variant enzyme undergoes multiple turnover cycles with O₂. This conclusion is supported by the steady-state assays of PaDADH E²⁴⁶L with D-arginine or D-leucine as substrate and O₂ as electron acceptor. The positive-sloped linear dependence of the plot of the enzymatic initial velocities against enzyme concentration (Fig. 4) demonstrates that PaDADH E²⁴⁶L turns over with O₂. Accordingly, we propose that PaDADH E²⁴⁶L follows an O₂-driven dehydrogenase mechanism (Fig. 8). In the O₂-driven dehydrogenase mechanism, a hydride transfer from the amino acid substrate reduces the enzyme-bound oxidized flavin to yield the reduced flavin and the imino acid product. The reduced flavin then reacts with O₂ after product release to form the caged O₂^•ˉ/flavin semiquinone radical pair. The flavin semiquinone subsequently donates a hydrogen atom to O₂^•ˉ, producing HO₂ˉ and the re-oxidized flavin. The observation that the initial rates of PaDADH E²⁴⁶L turnover with D-arginine and O₂ remained unchanged at ∼0.4 s⁻¹ for 2 h (Fig. 5) suggests that the PaDADH E²⁴⁶L-generated O₂^•ˉ does not compromise the enzyme’s ability to turnover. The sustained catalytic integrity of PaDADH E²⁴⁶L can then be explained as a likely prevention of the accumulation of the highly reactive O₂^•ˉ that may damage the enzyme. A likely mechanism preventing O₂^•ˉ accumulation is the disproportionation of O₂^•ˉ, which yields HO₂ˉ, HOˉ, and O₂ (Fig. 6) (vide supra). Moreover, the conversion of O₂^•ˉ to HO₂ˉ following a hydrogen transfer from the flavin semiquinone during the O₂-driven dehydrogenase mechanism cannot be ruled out.

Figure 8.

Proposed reaction scheme of the PaDADH E246L dehydrogenase activity with PMS or O2 as an electron acceptor during turnover. A hydride transfer from the amino acid substrate reduces the enzyme-bound oxidized flavin to yield the reduced flavin and the imino acid product. The reduced flavin is then re-oxidized through a PMS-driven or an O2-driven dehydrogenase activity. For the PMS-driven turnover, the reduced flavin is re-oxidized by PMS to yield PMSH2, restoring the enzyme to its resting state. For the O2-driven turnover, the reactivity of O2 with the reduced flavin yields the highly reactive caged O2•ˉ/flavin semiquinone radical pair. The flavin semiquinone then donates a hydrogen atom to the O2•ˉ, yielding HO2ˉ and the re-oxidized flavin in the resting state of the enzyme. E246L, glutamate 246 to leucine mutant; O2•ˉ, superoxide radicals; PMS, phenazine methosulfate.

The O₂-driven dehydrogenase activity of PaDADH E²⁴⁶L with a rate of 0.23 s⁻¹, irrespective of the substrate and concentration tested, can be explained as a likely saturation of the PaDADH E²⁴⁶L variant with the amino acid substrate during turnover with O₂ as the electron acceptor. The data agree well with a recent report investigating the role of the E²⁴⁶ residue in PaDADH catalysis in which a small O₂-driven dehydrogenase activity of ∼0.2 s⁻¹ was reported for the E²⁴⁶L, E²⁴⁶G, and E²⁴⁶Q variant enzymes (70). In the same study, the kinetic parameters were investigated using PMS as an artificial electron acceptor for PaDADH since the physiological electron acceptor is unknown. The study reported a k_cat value of ∼270 s⁻¹ with PMS and a maximum rate of ∼0.2 s⁻¹ with O₂ for the E²⁴⁶L variant enzyme with D-arginine (70). In this way, the previous study demonstrates that the PaDADH E²⁴⁶L variant enzyme turns over primarily through a PMS-driven dehydrogenase mechanism, not an O₂-driven dehydrogenase mechanism (Fig. 8). Hence, the O₂-driven dehydrogenase activity does not affect the PMS-driven dehydrogenase activity of PaDADH E²⁴⁶L.

The E²⁴⁶L variant-generated O₂^•ˉ diverts protons from the active site during D-arginine oxidation. This conclusion is supported by the steady-state pH profiles of the PaDADH E²⁴⁶L variant with D-leucine and D-arginine in the presence and absence of superoxide dismutase (Fig. 7). The pH profile of the k_cat/K_m parameter with D-leucine yielded a plot with a +1 slope for the increasing limb from low to high pH. Conversely, with D-arginine, the k_cat/K_m pH profile yielded a nonstoichiometric slope of +1.5 that was corrected to +1 after superoxide dismutase addition. Conventionally, the slope of a kinetic parameter’s pH profile has been used as an indicator for the number of ionizable processes required to complete the catalytic processes probed by the kinetic parameter (10, 11, 12, 13, 14, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102). Thus, the observed differences in the PaDADH E²⁴⁶L k_cat/K_m pH profile slope with D-arginine and D-leucine can be explained in light of different ionization processes being essential for the catalysis of the two substrates.

With D-arginine, the slope of +1 for the log (k_cat/K_m) pH profile is assigned to the ionization of the substrate’s α-NH₃⁺ during enzyme catalysis. The observed slope of +1 compared to the slope of +2 previously reported for the wildtype enzyme suggests that the unprotonated E²⁴⁶ residue is important for binding D-arginine. However, in a previous study investigating the role of residue E⁸⁷ in PaDADH, the observation of only a single ionizable group with a slope of +1 in the PaDADH E⁸⁷L variant led to the assignment of residue E⁸⁷ as one of the two unprotonated groups required for D-arginine binding, with the substrate’s α-NH₃⁺ group being the other (77). Altogether, the data portray the requirement for three groups: the substrate’s α-NH₃⁺, E⁸⁷, and E²⁴⁶ for the k_cat/K_m kinetic parameter during D-arginine oxidation. With the substrate being the commonality between all enzyme forms under discussion: wildtype, E⁸⁷ variant, and E²⁴⁶ variant, the first ionizable group can be unequivocally assigned to the substrate’s α-NH₃⁺ group. Thus, given that the wildtype enzyme reports only two ionizable groups for the k_cat/K_m pH profile, the assignment of the second group must be critically assessed. From the PaDADH crystal structure (Fig. 9), both E⁸⁷ and E²⁴⁶ interact with the guanidinium side chain and are important for D-arginine binding and imino arginine release (70, 77). Since the guanidinium charge is delocalized, we propose that the second ionizable group reflects not one but the joint ionization of the E⁸⁷ and E²⁴⁶ residues to ensure maximal binding of D-arginine in the wildtype enzyme. Thus, mutation of either E⁸⁷ or E²⁴⁶ yields similar and nonadditive effects on the log (k_cat/K_m) pH profile as reported for His²⁵⁶ and Asp²⁶⁶ of the carbohydrate-binding domain in rat hepatic lectin-1 (103).

Figure 9.

The active site topology of PaDADH showing substrate interactions and the highly polar active site pocket. The PDB file 3NYE was visualized and analyzed using the UCSF Chimera software (110). IAR, iminoarginine product; PaDADH, Pseudomonas aeruginosa D-arginine dehydrogenase.

The correction of the log (k_cat/K_m) pH profile’s slope from +1.5 to +1 following superoxide dismutase addition suggests that the O₂^•ˉ produced by PaDADH E²⁴⁶L’s reaction with D-arginine and O₂ is responsible for the observed nonstoichiometric slope. As the enzyme turns over with D-arginine, the highly reactive O₂^•ˉ generated during flavin reduction progressively accumulates in the solution due to multiple enzyme turnovers. However, during PaDADH E²⁴⁶L oxidation of D-arginine, the proton released from the α-NH₃⁺ ionization is unavailable for O₂^•ˉ reactivity due to the H⁴⁸-mediated proton relay network to the bulk solvent (Fig. 10), as previously described for p-hydroxybenzoate hydroxylase (66, 77, 104). Thus, the accumulated and highly reactive O₂^•ˉ most likely reacts with the E⁸⁷ proton to yield HO₂^• during PaDADH E²⁴⁶L turnover with low concentrations of D-arginine, thereby favoring E⁸⁷ ionization. The observed steeper and nonstoichiometric slope of +1.5 (Fig. 7) in the log (k_cat/K_m) pH profile, therefore, likely reflects inflated values for the k_cat/K_m parameter, which probes the free forms of enzymes at low substrate concentrations rather than enzyme–substrate complexes. Due to the O₂^•ˉ diversion of the E⁸⁷ ionized protons, there is a partial detection of the E⁸⁷ ionization in the log (k_cat/K_m) pH profile with D-arginine despite the absence of E²⁴⁶ (103).

Figure 10.

The proton-relay network of PaDADH. The PDB file 3NYE was visualized and analyzed using the UCSF Chimera software (110). IAR, iminoarginine product; PaDADH, Pseudomonas aeruginosa D-arginine dehydrogenase.

In contrast, the D-leucine substrate has a short and nonpolar side chain and does not require ionization of the E⁸⁷ residue for binding. Thus, there is no observed effect of O₂^•ˉ on the log (k_cat/K_m) pH profile plot. The observed slope of +1 with an intrinsic pK_a value of ∼8.5 and common to both D-leucine and D-arginine corresponds to the ionization of the substrate’s α-NH₃⁺ group that initiates hydride transfer for amine oxidation (42, 77).

The acquired O₂ reactivity of the E²⁴⁶L variant yielding a caged O₂^•ˉ/flavin semiquinone radical pair can be explained by the modified active site topology of the E²⁴⁶L variant. Studies on flavoproteins demonstrate that for an enzyme to react with O₂, there must be a nonpolar residue and a positive charge close to the flavin cofactor to favor the electrostatics required for O₂ reactivity (18, 20, 21, 81, 83). In PaDADH, the interaction between the R²²²/R³⁰⁵ network and the α-carboxylate group renders the amino acid substrate or imino acid product with a net positive charge in the ligand-bound state (Fig. 9). Nonetheless, in the free form of the enzyme, which is the form that reacts with O₂, the positive charge likely arises from either R²²² or R³⁰⁵. However, the wildtype enzyme contains only polar amino acid residues, negating one requirement for O₂ reactivity (66). Conversely, the presence of the nonpolar L²⁴⁶ in PaDADH E²⁴⁶L provides a suitable active site topology for O₂ reactivity, having a positive charge and a nonpolar residue (24). Thus, the E²⁴⁶L variant enzyme gains the ability to react with O₂. Similarly, PaDADH reactivity with O₂ has been recently reported for the Y²⁴⁹F variant enzyme after replacing the polar tyrosine 249 residue with the nonpolar phenylalanine residue in the active site (105, 106). However, unlike this study, the Y²⁴⁹F variant’s O₂ reactivity yielded either a flavin N⁵ adduct or a green 6-OH-FAD species. Nonetheless, these studies exemplify how substrates and protein residues dictate the versatility of flavin reactivity to favor specific reactions (105, 106, 107, 108).

In conclusion, this study has used steady-state and rapid-reaction kinetic approaches to investigate the effects of the E²⁴⁶L mutation on the ability of PaDADH to react with O₂ and the role of the generated O₂^•ˉ in the catalysis of the PaDADH E²⁴⁶L variant enzyme. The study demonstrates a mutation-induced gain of PaDADH reactivity with O₂. pH studies on the steady-state kinetic parameters of the variant enzyme demonstrate that the O₂^•ˉ generated during PaDADH E²⁴⁶L catalysis diverts protons from the active site during D-arginine binding, leading to an observed nonstoichiometric slope of +1.5 in the log (k_cat/K_m) pH profile plot. Conventionally, pH profile slopes of enzyme variants have been the paradigm for identifying ionizable residues involved in enzyme catalysis (37, 38, 39, 40, 41, 42, 43). The technique, which is widely used across the field of enzymology, relies on the identification of integer slopes in pH profile plots for the assignment of catalytic groups after site-directed mutagenesis of enzyme residues. To date, not much information exists on interpreting data from systems with nonstoichiometric pH profile slopes, thereby limiting the applications of pH profiles when elucidating enzyme catalytic mechanisms in such systems. This study provides useful information on correcting and interpreting nonstoichiometric slopes in the pH plots of enzymes that react with O₂ to form O₂^•ˉ during catalysis (99, 101, 109). Additionally, this study demonstrates that a charge and nonpolar residue near the flavin can induce oxygen reactivity in an enzyme.

Experimental procedures

Materials

Escherichia coli strain Rosetta(DE3)pLysS was purchased from Novagen. The QIAprep Spin Miniprep Kit and the QIAquick Polymerase Chain Reaction Purification Kit were obtained from Qiagen. Pfu DNA polymerase was purchased from Stratagene, and DpnI was obtained from New England BioLabs. Oligonucleotides were purchased from Sigma Genosys for site-directed mutagenesis and sequencing of the variant genes. Bovine liver superoxide dismutase and PMS were from MilliporeSigma. D-Amino acids were obtained from Alfa-Aesar. All other reagents used were obtained at the highest purity commercially available.

Site-directed mutagenesis, protein expression, and purification

The E²⁴⁶L variant gene of PaDADH was engineered by mutagenic polymerase chain reaction (PCR) with the pET20b(+)/PA3863 plasmid harboring the wildtype gene (dauA) as a template. A concentration of 5% dimethyl sulfoxide was added to the PCR reaction mixture to ensure proper separation of the high GC-rich, double-stranded DNA template. Site-directed mutagenesis amplicons were purified using the QIAquick PCR Purification Kit following the manufacturer’s protocol. The purified plasmid was then subjected to endonuclease activity using DpnI at 37 ^oC for 2 h. The resulting plasmid was used to transform the DH5α strain of E. coli cells. The success of the mutation was confirmed by sequencing the gene using the services of Humanizing Genomics Microgen USA Corp in Maryland. The E²⁴⁶L variant enzyme of PaDADH was then expressed in E. coli Rosetta(DE3)pLysS and purified to homogeneity as previously described for the PaDADH wildtype enzyme in the presence of 10% (v/v) glycerol for enzyme stability and to prevent the loss of the bound FAD cofactor (44). The purified enzyme was stored at −20 ^oC in 20 mM Tris-Cl, pH 8.0, and 10% glycerol and was found to be active for at least 6 months.

Rapid-reaction kinetics of the PaDADH E²⁴⁶L variant enzyme

To establish the effect of the E²⁴⁶L mutation on the rapid-reaction kinetic parameters of the PaDADH E²⁴⁶L variant enzyme and whether the mutation resulted in an acquired reactivity of the enzyme with O₂, the reductive half-reaction of the PaDADH E²⁴⁶L variant enzyme was carried out under aerobic conditions. Using an SF-61DX2 Hi-Tech KinetAsyst performance stopped-flow spectrophotometer, the reaction was followed in 20 mM NaPP_i, pH 10.0, and 25 ^oC and compared to the anaerobic reductive half-reactions of the wildtype and E²⁴⁶L variant enzymes that were previously investigated (42, 70). Since ∼80% of flavin reduction occurs in the mixing time (2.2 ms) of the stopped-flow spectrophotometer with D-arginine as a substrate (42, 71), the flavin reduction of the E²⁴⁶L variant enzyme was investigated with D-leucine as the reducing substrate. Substrate solutions (1–40 mM) were loaded into syringes and mounted onto the stopped-flow spectrophotometer. The reaction was followed by observing the spectroscopic decay of the 446 nm flavin peak over time upon reacting ∼10 μM enzyme with substrate solutions under pseudo-first-order conditions in single mixing mode.

Data analysis

The time-resolved flavin reductions were fit to Equation 1, which describes a triple exponential process for flavin reduction. Here, k_obs1, k_obs2, and k_obs3 represent the observed first-order rate constant for reducing the enzyme-bound flavin at any given substrate concentration at 446 nm. A represents the absorbance at 446 nm at any given time, B₁, B₂, and B₃ are the amplitudes of the absorption changes, t is time, and C is the absorbance at an infinite time that accounts for the nonzero absorbance of the fully reduced enzyme-bound flavin.

$$A=B_1\exp \left(-k_{\text{obs}1}t\right)+B_2\exp \left(-k_{\text{obs}2}t\right)+B_3\exp \left(-k_{\text{obs}3}t\right)+C$$ (Eq 1)

The resulting kinetic parameters of the reductive half-reaction were determined after fitting the observed rate constants for flavin reduction at various D-leucine concentrations with Equation 2. The equation defines a hyperbolic saturation of the enzyme with the D-leucine substrate, yielding a y-intercept value of zero. The data were fit with the KaleidaGraph software (Synergy Software). Here, k_obs1 represents the observed first-order rate constant for reducing the enzyme-bound flavin at any substrate concentration (S). k_red is the limiting first-order rate constant for flavin reduction at saturating substrate concentrations. K_d is the apparent equilibrium constant for dissociating the enzyme–substrate complex into the free substrate and enzyme. The same data were obtained when an equation that defines a hyperbolic saturation with a finite y-intercept was used.

$$k_{\text{obs}1}=\frac{k_{\text{red}}S}{K_{\mathrm{d}}+S}$$ (Eq 2)

O₂ reactivity studies of the PaDADH E²⁴⁶L variant enzyme

To investigate the effect of the E²⁴⁶L mutation on the ability of the PaDADH E²⁴⁶L variant enzyme to react with O₂ under steady-state conditions, the initial velocities of the enzyme reaction were measured with D-leucine as a substrate and O₂ as an electron acceptor at pH 8.5. The reduction of O₂ was followed using a Clark-type oxygen electrode in a 1 ml reaction volume containing final enzyme concentrations of 0.48 μM to 9.7 μM and fixed D-leucine concentration at 25 mM in 20 mM NaPP_i at 25 ^oC. Substrate solutions were prepared in the reaction buffer, and the pH was readjusted after the amino acid substrate was dissolved. The experiment was repeated with D-arginine as substrate at pH 8.5, in a 1 ml reaction volume containing final enzyme concentrations of 0.17 μM to 8.5 μM and fixed D-arginine concentration at 5 mM in 20 mM NaPP_i at 25 ^oC.

In a separate experiment, the dependence of the PaDADH E²⁴⁶L variant enzyme’s O₂ reactivity on substrate concentration was investigated at fixed 0.5 μM enzyme and varying concentrations of D-arginine (1 mM – 20 mM) or D-leucine (1 mM–10 mM).

To probe the oxygen species generated by the E²⁴⁶L variant O₂ reaction, the enzymatic assay was carried out with 6 μM E²⁴⁶L variant enzyme and fixed 5 mM D-arginine or 25 mM D-leucine. The reaction was then repeated by adding 200 to 500 units of superoxide dismutase to the reaction mixture.

To investigate the effect of the PaDADH E²⁴⁶L-generated O₂^•ˉ on the enzyme’s ability to turnover with D-arginine, the steady-state kinetic properties of the variant enzyme with O₂ as an electron acceptor were investigated using ∼6 μM enzyme with 0.15 mM, 0.24 mM, or 0.5 mM D-arginine. Upon reaching a plateau in the reaction cycle, D-arginine was reintroduced to the reaction mixture to yield a total of four reaction cycles.

pH effects on the steady-state kinetics of the PaDADH E²⁴⁶L variant enzyme

To determine the effects of pH on the steady-state kinetics of the E²⁴⁶L variant enzyme of PaDADH, the apparent steady-state kinetic parameters of the enzyme with D-arginine or D-leucine as a substrate and PMS as an artificial electron acceptor were obtained by monitoring the initial PMS-driven O₂ consumption rates with a computer-interfaced Oxy-32 oxygen-monitoring system (Hansatech Instruments Ltd) under similar conditions. D-Arginine concentrations were between 0.1 and 80 mM, D-leucine concentrations were between 1.25 and 62.5 mM, the enzyme concentration ranged from 1.15 μM to 5.75 μM, PMS concentration was fixed at 1 mM, and the pH ranged from 5.0 to 10.5 at 25 ^oC. Temperature effects on the steady-state pH profiles for D-arginine were investigated by repeating the assays at 12 ^oC. All assays were carried out to ensure that the K_m value was within the range of the substrate concentrations used at each pH. To ensure that the variant enzyme was fully saturated with PMS, the steady-state kinetic parameters were also determined at 1.5 mM PMS, yielding similar results.

Due to the observation of a nonstoichiometric slope of +1.5 for the log (k_cat/K_m) pH profile with D-arginine at 12 ^oC and 25 ^oC, the assay was repeated with 200 to 500 units of superoxide dismutase in each apparent steady-state reaction mixture to investigate the effect of the enzyme-O₂ reactivity on the D-arginine pH profile. The data analyses and interpretation focused only on the k_cat/K_m pH profiles due to the observation that superoxide dismutase affected only the k_cat/K_m pH profiles with D-arginine.

For the D-arginine substrate, the log values of the k_cat/K_m parameters under varying conditions of temperature and superoxide dismutase were plotted by fitting the log values of the k_cat/K_m parameters with Equation 3. The equation describes a curve that increases with increasing pH with a slope of S and a pH-independent limiting value (C_H) at high pH.

The plot of the k_cat/K_m parameter for the D-leucine substrate was made by fitting the log values with Equation 4, which describes a curve that increases with increasing pH with a slope of +1 and a pH-independent limiting value (C_H) at high pH.

$$\mathit{log}Y=\log \frac{C_H}{{(1+\frac{{10}^{-pH}}{{10}^{{-pK}_b}})}^{\wedge }S}$$ (Eq 3)

$$\mathit{log}Y=\log \frac{C_H}{(1+\frac{{10}^{-pH}}{{10}^{{-pK}_b}})}$$ (Eq 4)

Data availability

All data are contained within the manuscript.

Conflict of interest

The authors declare no conflicts of interest with the contents of this article.

Reviewed by members of the JBC Editorial Board. Edited by Joan B. Broderick