Non-chemical proton-dependent steps prior to O₂-activation limit Azotobacter vinelandii 3-mercaptopropionic acid dioxygenase (MDO) catalysis

Abstract

3-mercaptopropionate dioxygenase from Azotobacter vinelandii (Av MDO) is a non-heme mononuclear iron enzyme that catalyzes the O₂-dependent oxidation of 3-mercaptopropionate (3mpa) to produce 3-sulfinopropionic acid (3spa). With one exception, the active site residues of MDO are identical to bacterial cysteine dioxygenase (CDO). Specifically, the CDO Arg-residue (R50) is replaced by Gln (Q67) in MDO. Despite this minor active site perturbation, substrate-specificity of Av MDO is more relaxed as compared to CDO. In order to investigate the relative timing of chemical and non-chemical events in Av MDO catalysis, the pH/D-dependence of steady-state kinetic parameters (k_cat and k_cat/K_M) and viscosity effects are measured using two different substrates [3mpa and L-cysteine (cys)]. The pL-dependent activity of Av MDO in these reactions can be rationalized assuming a diprotic enzyme model in which three ionic forms of the enzyme are present [cationic, E^(z+1); neutral, E^z; and anionic, E^(z−1)]. The activities observed for each substrate appear to be dominated by electrostatic interactions within the enzymatic active site. Given the similarity between MDO and the more extensively characterized mammalian CDO, a tentative model for the role of the conserved ‘catalytic triad’ is proposed.

1. Introduction

Thiol dioxygenase enzymes utilize a mononuclear non-heme iron site to catalyze the O₂-dependent oxidation of thiol bearing substrates without the need for an external reductant or cofactor. Among this class of enzymes, the mammalian cysteine dioxygenase (CDO) is the best characterized [1–5]. Imbalances in L-cysteine (cys) metabolism have also been identified in a variety of neurological disease states (motor neuron, Parkinson and Alzheimer) [5–7]. These observations suggest a potential correlation between impaired sulfur metabolism, oxidative stress and neurodegenerative disease [8,9]. For this reason, enzymes involved in sulfur-oxidation and transfer are increasingly being recognized as potential drug targets for development of antimicrobials, therapies for cancer and inflammatory disease [10–13].

Multiple high-resolution crystal structures have been solved highlighting the conserved Fe-coordination sphere among confirmed thiol dioxygenase enzymes. The typical non-heme mononuclear oxidase/oxygenase iron coordination sphere is comprised of two protein-derived neutral His residues and one monoanionic carboxylate ligand, provided by either an Asp or Glu residue. By contrast, the mononuclear iron site in all known thiol dioxygenase enzymes is coordinated by three protein derived histidine residues resulting in a neutral 3-His facial triad.

Structural comparison of annotated bacterial thiol dioxygenases suggest at least two subclasses of enzymes, which differ in a single outer Fe-coordination sphere residue (Arg or Asn) [14]. To illustrate, Fig. 1 shows the active site coordination for the ‘Arg-type’ bacterial cysteine dioxygenase (CDO) isolated form Bacillus subtilis (PDB code 4QM8) as compared to the ‘Gln-type’ enzyme cloned from Pseudomonas aeruginosa (PDB code 4TLF) [15]. Of note, the 3-His facial triad motif in both enzymes is conserved; however, the outer-sphere Arg-residue (R50) involved in electrostatic stabilization of the cys substrate is replaced by Gln (Q62). The spatial orientations of all other conserved residues (S137, H139 and Y141) within the active site ‘catalytic triad’ remain invariant. The primary difference between the eukaryotic and bacterial ‘Arg-type’ CDO enzymes is a post-translational modification adjacent (3.3 Å) to the mononuclear Fe-site in which spatially adjacent Cys93 and Tyr157 residues are covalently cross linked to produce a C93-Y157 pair. This feature is unique to eukaryotic enzymes [16,17].

Fig. 1.

Crystal structure of the substrate-bound ‘Arg-type’ Bacillus subtilis active site (A, PDB code 4QM8) as compared to the annotated ‘Gln-type’ MDO isolated from Pseudomonas aeruginosa (B, PDB code 4TLF) [14,15]. Solvent waters are designated W1-W3. Selected distances A: 1 (2.59 Å); 2 (3.30 Å); 3 (2.14 Å); 4 (2.65 Å); B: 1′ (2.56 Å); 2′ (2.95 Å).

Previously, both ‘Arg-’ and ‘Gln-type’ enzymes were annotated as bacterial CDO enzymes [14]. However, recent spectroscopic and kinetic investigations performed independently demonstrated that the ‘Gln-type’ class of thiol dioxygenase enzymes are more accurately designated as a 3-mercaptopropionic acid dioxygenase (MDO) [15,18]. Remarkably, these ‘Gln-type’ enzymes have vastly relaxed substrate-specificity as compared to CDO [16]. Indeed, at saturating substrate concentration, MDO isolated from the soil bacteria Azotobacter vinelandii (Av MDO) demonstrated comparable activity (k_cat) toward three different thiol-bearing substrates [3-mercaptopropionic acid (3mpa), L-cysteine (cys) and cysteamine (ca)] [18]. Despite similar maximal velocities, the ‘Gln-type’ Av MDO exhibits a catalytic efficiency (k_cat/K_M) for 3mpa nearly two orders of magnitude greater than cys and ca. Thus, 3mpa appears to be the preferred substrate for this enzyme. Similar behavior was reported for the ‘Gln-type’ enzyme isolated from Pseudomonas aeruginosa (Pa) [15].

While the mammalian ‘Arg-type’ CDO isolated from Mus musculus (Mm CDO) has been extensively characterized, no information is available regarding the relative timing of chemical and nonchemical steps in ‘Gln-type’ MDO catalyzed reactions. Comparing the kinetic mechanisms for various thiol dioxygenase enzymes is critical to identifying common mechanistic behavior among this class of enzymes. To this end, the influence of pH, solvent isotope and viscosity effects were evaluated for Av MDO steady-state kinetic parameters (k_cat and k_cat/K_M) obtained using two different substrates (3mpa and cys). Proton-inventory experiments were performed to determine the number of exchangeable protons in flight during chemical and non-chemical steps. The relative timing of diffusional steps was investigated by measuring the influence of solvent viscosity on enzyme kinetics. Collectively, these results provide greater insight into the role of conserved outer Fe-coordination sphere residues and the nature of substrate coordination to the mononuclear iron active site.

2. Materials and methods

2.1. Expression and purification of Av MDO

The isopropyl β-D-1-thiogalactopyranoside (IPTG) inducible T7 vector for Av MDO (designated pRP42), was a generous gift from Professor Tim Larson (Virginia Tech, Department of Biochemistry). Sequence verification of the plasmid was performed by Sequetech (Mountain View, CA, http://sequetech.com/). The pRP42 vector was transformed into chemically competent BL21(DE3) E. coli (Novagen Cat. no. 70236-4) by heat-shock [42 °C for 45 s] and grown overnight at 37 °C on a lysogeny broth (LB) [19] agar plate in the presence of 100 mg/L ampicillin (Amp). The following day, a single colony was selected for growth in liquid LB (Amp) media for training on antibiotic prior to inoculation of 10-L BF-110 fermentor (New Brunswick Scientific) at 37 °C. Cell growth was followed by optical density at 600 nm (OD₆₀₀). Induction was initiated by addition of 1.0 g IPTG, 78 mg ferrous ammonium sulfate and 20 g casamino acids at an OD₆₀₀ ~4. At the time of induction, the temperature of the bioreactor was decreased from 37 °C to 25 °C and agitationwas set to maintain an O₂ concentration of 20% relative to air-saturated media. After 4 h, the cells were harvested and pelleted by centrifugation (Beckman-Coulter Avanti J-E, JA 10.5 rotor) at 18,600 × g for 15 min. The resulting cell paste was stored at −80 °C.

In a typical purification, ~20 g frozen cell paste was added to 150 mL extraction buffer (20 mM HEPES, 50 mM NaCl, pH 8.0). Lysosyme, ribonuclease and deoxyribonuclease were added to the slurry for a final concentration of 10 μg/mL each and stirred slowly on ice for 30 min. The resulting suspension was pulse sonicated (Bronson Digital 250/450) for 15 s on/off at 60% amplitude for a total time of 15 min. The insoluble debriswas removed from the cell free extract by centrifugation at 48,000 × g for 1 h at 4 °C. The supernatant was diluted 1:1 with extraction buffer and then loaded onto a DEAE sepharose fast flow anion exchange column [7 cm W × 20 cm L] (GE Life Sciences #17070901) pre-equilibrated with 20 mM HEPES, 50 mM NaCl, pH 8.0. The column was washed with three column volumes of extraction buffer prior to elution in a linear NaCl gradient (50 mM–350 mM). Fractions (~10 mL) were collected overnight and pooled based on enzymatic activity as described elsewhere [18,20]. SDS PAGE was also used to verify the presence of the recombinant protein (~23 kDa) within each fraction. Broad range protein molecular weight markers utilized in SDS PAGE experiments were purchased from Promega (Madison, WI) Cat. No. V8491. The pooled fractions were concentrated to approximately 5–10 mL using an Amicon stir cell equipped with an YM-10 ultrafiltration membrane. Thrombin protease (Biopharma Laboratories) was added to cleave the C-terminal His-tag from Av MDO. In a typical reaction, ~0.3 molar equivalents of thrombin per Av MDO (based on UV–visible absorbance at 280 nm) was added to batches of purified protein for overnight cleavage at 4 °C in HEPES buffer. The remaining thrombin and free (His)₆-tag was removed from Av MDO by size exclusion chromatography using a sephacryl S100 column. For all batches of Av MDO used in these experiments, spectrophotometric determination of ferrous and ferric iron content was measured using 2,4,6-tripyridyl-s-triazine (TPTZ) as described elsewhere [18,20]. For clarity, the concentrations reported in enzymatic assays reflect the concentration of ferrous iron within samples of Av MDO (Fe^II-MDO). Protein content was determined by Bio-Rad protein assay. NOTE: Although the expressed Av MDO has a C-terminal His-tag, as compared to the DEAE AX method described above, use of immobilized metal affinity chromatography (IMAC) results in substantially lower enzymatic activity and yields due to loss of iron and protein denaturation.

2.2. Enzyme assays

3-sulfinopropionic acid (3spa), cysteine sulfinic acid (csa) and hypotaurine (ht) were assayed using the HPLC method described previously [18,21,22]. Instrumental conditions: column, Phenomenex C18 (100 mm × 4.6 mm); mobile phase, 20 mM sodium acetate, 0.6% methanol, 1% (v/v) heptaflurobutyric acid, pH 2.0. Analytes were detected spectrophotometrically at 218 nm. Reactions (1 mL) were prepared in a buffered solution at the desired pH to obtain a final concentration from 0.1 to 10 mM 3mpa (0.1–60mMcys). Each reaction was initiated by addition of Av MDO (typically 0.5–1.0 μM) at 20 ± 2 °C. Sample aliquots (250 μL) were removed from the reaction vial at selected time points and quenched by addition of 10 μL 1 N HCl (final pH 2.0). Prior to HPLC analysis, each sample was spin-filtered through a 0.22 μm cellulose acetate membrane (Corning, Spin-X). Product concentration was determined by comparison to calibration curves as described elsewhere [16,18]. The rate of dioxygen consumption in activity assays was determined polarographically using a standard Clark electrode (Hansatech Instruments, Norfolk, England) in a jacketed 2.5 mL cell. Calibration of O₂-electrode is described in detail elsewhere [22,23]. All reactions were initiated by addition of 1.0 μM Av MDO under identical buffer conditions as described for HPLC assays. Reaction temperatures were maintained at 20 ± 2 °C by circulating water bath (ThermoFlex 900, Thermo Scientific).

2.3. Synthesis of the dianionic 3-sulfinopropionic acid

The salt of 3-sulfinopropionic acid (3spa) was prepared by saponification of the commercially available methyl ester [sodium 1-methyl 3-sulfinopropanoate; Sigma Aldrich 7,78,168]. Briefly, sodium 1-methyl-3-sulfinopropanoate (100 mg, 0.57 mmol) was dissolved in 5 mL of deionized water. To this solution, LiOH (70 mg, 2.9 mmol) was added prior to overnight reflux under constant stirring on an oil bath (~110 °C). The solution was then cooled to ~4 °C, filtered and the filtrate was dried by rotary evaporation. The resulting solid was dissolved in 2 mL of cold ethanol and filtered again to remove inorganic salts. Evaporation of the ethanol filtrate gave the desired compound as a white solid (68 mg, 0.45 mmol, 79%). Both ¹H and ¹³C {¹H} NMR were measured to verify the identity and relative purity of the 3spa product. ¹H NMR (500 MHz, D₂O) δ 2.49 (t, J = 7.8 Hz, 2H), 2.37 (t, J = 7.8 Hz, 2H). ¹³C {¹H} NMR (126 MHz, D₂O) δ 181.2, 57.3, 30.1 ppm. Additional confirmation of 3spa standard was performed by mass spectrometry as described elsewhere [18] with instrumentation from the Shimadzu Center for Advanced Analytical Chemistry (The University of Texas Arlington).

2.4. Solvent kinetic isotope effects and proton inventory

For pH/D-profiles (collectively pL) and solvent isotope studies, the buffer components were prepared directly in D₂O and adjusted by direct addition of NaOD. Each pD value was obtained from the pH-electrode reading using the relationship [pD = pH + 0.4]. The composition of reaction buffers for all pL-profile experiments consisted of 20 mM Good’s buffer and 50 mM NaCl. 2-(N-morpholino) ethanesulfonic acid (MES) was used to buffer reactions over the pL range of 5.5–6.9, 2-[4-(2-hydroxyethyl)piperazin-1-yl] ethanesulfonic acid (HEPES) was used to buffer reactions over the pH/D range of 7.0–8.4 and 2-(cyclohexylamino)-ethanesulfonic acid (CHES) was used to buffer reactions over the pL range of 8.5–10. For proton inventory experiments, the mole fraction of D₂O (n) was calculated based on combining appropriate ratios of buffer prepared in D₂O and H₂O.

2.5. Viscosity studies

For solvent viscosity studies the steady-state kinetic parameters (k_cat and k_cat/K_M) were determined for Av MDO catalysis using both oxygen electrode and HPLC at pH 8.0 (20 °C) as described above. Sucrose was used to increase the buffer viscosity within reaction mixtures. The viscosity (η) of buffers containing sucrose were measured using an Ostwald viscometer relative to 20 mM HEPES, 50 mM NaCl, pH 8.2 (20 °C). The values obtained represent the average of triplicate measurements. In these experiments, addition of sucrose up to 35% (w/v) was used to increase the relative viscosity of the buffer (η_rel) up to ~3-times that of the control buffer.

2.6. Data analysis

Steady-state kinetic parameters were determined by fitting data to the Michaelis-Menten equation using the program SigmaPlot ver. 11.0 (Systat Software Inc., Chicago, IL). From this analysis, both the kinetic parameters (k_cat and K_M) and error associated with each value were obtained by non-linear regression. For reactions where k_cat- or k_cat/K_M-pH data exhibits limiting nonzero plateaus at low (Y_L) and high (Y_H) pH, followed by transition region and subsequent decrease beyond Y_H, the results were fit to equation (1) [24]. Here, Y is defined by either k_cat or k_cat/K_M, and the variables [H], K₁ and K₂ represent the hydrogen ion concentration and the two observable dissociation constants for ionizable groups involved in catalysis, respectively.

$$logY=\mathit{log}\left(\frac{Y_L+Y_H·\frac{K_1}{[H]}}{1+\frac{K_1}{[H]}+\frac{K_1{K}_2}{{[H]}^2}}\right)$$ (1)

For results in which k_cat- or k_cat/K_M-pH profiles decreased only at low pH, the data were fit to equation (2) [25–27]. This expression is scaled by a constant scalar quantity (C) which represents the maximum kinetic rate (k_cat or k_cat/K_M). The error associated with each parameter (k_cat, k_cat/K_M, pK_a1 and pK_a2) determined from fits to equations (1) and (2) is reported in Table 1.

Table 1.

Summary of pL-dependent steady-state kinetic parameters determined for MDO in reactions utilizing 3-mercaptopropionic acid and L-cysteine.

Av MDO kinetic parameter	Substrate
	3mpa		cys
	H2O	D2O	H2O	D2O
log(kcat)-pL
maximum kcat (s−1)	0.45 + 0.06	0.45 + 0.03	0.28 ± 0.02	0.26 ± 0.02
pKa1	7.8 ± 0.2	7.9 ± 0.1	6.3 ± 0.1	7.1 ± 0.1
pKa2	9.2 ± 0.1	9.3 ± 0.1	–	–
SKIE	1.01 ± 0.11		1.08 ± 0.10
log(kcat/KM)-pL
maximum kcat/KM (M−1 s−1)	27,000 ± 3000	10,000 ± 1100	41.5 ± 0.7	40.8 ± 0.5
pKa1	7.4 ± 0.2	7.8 ± 0.2	6.2 ± 0.1	7.2 ± 0.1
pKa2	9.1 ± 0.1	9.2 ± 0.1	–	–
SKIE	2.7 ± 0.4		1.02 ± 0.09
Proton inventory
pL 6.0	y = 1		n/a
pL = 8.2	y = 1		n/a

$$logY=log\left(\frac{C}{1+\frac{[H]}{K_1}}\right)$$ (2)

Proton inventory results were fit to equation (3) where n is the mol fraction of D₂O in the reaction chamber and E₀ and E_n are the kinetic parameters (either k_cat or k_cat/K_M) in H₂O and the mol fraction of D₂O, respectively. The measured solvent kinetic isotope effect is defined as SKIE and the integer value y describes the number of protons that contribute to the isotope effect.

$$\frac{E_n}{E_0}={\left(1-n+\left(\frac{n}{\mathit{SKIE}}\right)\right)}^y$$ (3)

The effect of solvent viscosity on the steady-state parameters were fit using equation (4). In this equation, Y₀ represents each kinetic parameter determined in the absence of viscogen whereas Yη is the value obtained at each specific relative viscosity measured. All viscosities measured are normalized to reaction buffer in the absence of viscogen (η_rel). The slope of the line (m) represents the extent of diffusion limitation.

$$\frac{Y_0}{Y_{\eta }}=1+m·{\eta }_{\mathit{rel}}$$ (4)

3. Results

3.1. Solvent kinetic isotope effects

It has been previously observed that Av MDO catalyzes the O₂-dependent oxidation of thiol-bearing substrates 3-mercaptopropionic acid (3mpa), L-cysteine (cys) and cysteamine (ca) to yield the corresponding sulfinic acid products 3-sulfinopropionic acid (3spa), cysteine sulfinic acid (csa) and hypotaurine (ht), respectively [18]. In order to probe the rate-limiting chemical and non-chemical steps in Av MDO catalysis, the pH/D-dependent solvent kinetic isotope effects (SKIE) for these reactions were interrogated. However, before these experiments can be presented, it is necessary to discuss some fundamental control experiments first.

For dioxygenase reactions, oxygen is a co-substrate and thus it is important to verify that atmospheric O₂-concentration is sufficient to saturate enzyme kinetics for all substrates utilized in solvent isotope experiments. This control was performed previously at pH 7.5 for all substrates (3mpa, cys and ca) [18]. As an addition control for these studies, this was repeated at pH 5.5, 7.5 and 10 for both 3mpa and cys to verify that pH does not influence O₂-saturation. As before, steady-state rates for Av MDO at saturating substrate concentration are independent of oxygen within the range of 25–400 μM. These observations indicate that (regardless of pH) the apparent $K_M^{O2}$ (for both substrates) must be substantially lower (~10×) than the lowest value of oxygen used (25 μM) [21,28]. Therefore, any differences observed in Av MDO reactivity cannot be attributed to incomplete O₂-saturation. This also means that under these conditions, the reaction of the substrate-bound Av MDO with oxygen can be considered irreversible.

An additional factor to consider is the ‘coupling efficiency’ of the enzyme. The efficiency at which an oxygenase enzyme incorporates one mol of O₂ into the product is commonly referred to as ‘coupling’. As k_cat represents the zero-order limit of catalysis, the coupling efficiency for a dioxygenase can be obtained from the ratio of the k_cat as measured from product formation divided by the k_cat obtained from O₂-consumption. It was previously reported that Av MDO reactions using either 3mpa (102 ± 8%) or cys (97 ± 6%) as a substrate were nearly fully coupled over the accessible pH range of the enzyme [6 < pH < 9] [18]. By contrast, reactions utilizing ca as a substrate exhibited significantly lower coupling (40 ± 9%). Regardless of substrate, the coupling efficiency of Av MDO was not influenced by D₂O. Uncoupled reactions can result in the promiscuous release of reactive oxygen species such as H₂O₂, O₂·⁻ and ·OH, which can complicate interpretation of these experiments. For this reason, the pH/D-dependent solvent isotope studies are focused on Av MDO substrates 3mpa and cys, which exhibit stoichiometric coupling.

Insight into ionizable groups involved in catalysis can be obtained by measuring solvent isotope effects on the pH/D-dependence of both k_cat and k_cat/K_M. The influence of pH on Av MDO catalysis was measured over the accessible pH range of the enzyme (6 < pH < 9). Fig. 2 illustrates the log(k_cat)-pH (A) and log(k_cat/K_M)-pH profiles (B) obtained from the initial rate data of Av MDO reactions using 3mpa (H₂O, white circle) as a substrate. For comparison the log(k_cat)-pD and log(k_cat/K_M)-pD profiles obtained for 3mpa reactions are also included (black circles). Each point within these data sets was obtained by fitting the steady-state kinetic results observed at a fixed pL-value to the standard Michaelis- Menten equation. The error in each kinetic parameter obtained (k_cat and k_cat/K_M) from these fits is indicated graphically using error bars.

Fig. 2.

pL-dependence of steady-state kinetic parameters of MDO using 3-mercaptopropionic acid (circle, A/B) and L-cysteine (square, C/D) as substrates. For both substrates, log(V) and log(V/K) pL-profiles were collected in H₂O (white) and D₂O (black) and fit to equations. (1) and/or (2). A summary of the results obtained from these fits are summarized in Table 1.

In reactions using 3mpa as substrate, both pH-profiles [k_cat-pH and k_cat/K_M-pH] exhibit a skewed ‘bell shaped’ curve with two apparent pKa values. Within the acidic branch of this profile [6 < pH < 8], both data sets exhibit sigmoidal behavior bracketed by limiting nonzero plateaus at low and high pH. Following a short transition, k_cat and k_cat/K_M values decrease beyond with a unitary slope. Over the entire pH range utilized in these experiments, both data sets (k_cat and k_cat/K_M) can be reasonably fit to equation (1) (solid line). The limiting values of k_cat spans 0.17 ± 0.01 s⁻¹ (Y_L) to 0.45 + 0.06 s⁻¹ (Y_H) at 20 °C. Likewise, the limiting values for k_cat/K_M within the sigmoidal region lie between 10,000 ± 1100 M⁻¹ s⁻¹ and 27,000 ± 3000 M⁻¹ s⁻¹, respectively. Relative to H₂O reactions, the observed k_cat values are unaffected for Av MDO assays carried out in D₂O (Fig. 2A). By contrast, a significant decrease in k_cat/K_M is observed for 3mpa reactions carried out in D₂O (Fig. 2B). The sigmoidal limits of k_cat/K_M-values in D₂O lie between 3200 ± 200 M⁻¹ s⁻¹ and 9600 ± 1200 M⁻¹ s⁻¹, respectively.

For each pH-profile, the two ionizable groups (pKa₁ and pKa₂) were determined by fitting data to equation (1). From this analysis, the pKa values (pKa₁, 7.8 ± 0.2) and (pKa₂, 9.2 ± 0.1) were obtained from the 3mpa k_cat-pH profile in H₂O. Similarly, two pKa values (pKa₁, 7.4 ± 0.2) and (pKa₂, 9.1 ± 0.1) were obtained from the 3mpa k_cat/K_M–pH profile. As with H₂O reactions, two ionizable groups are also observed in Av MDO D₂O assays using 3mpa as a substrate. In fact, the log(k_cat)-pD profile obtained is essentially super imposable with the results obtained in H₂O. Thus, within experimental error, the presence of D₂O has no influence on the pKa-values obtained from k_cat data. By contrast, the value of pKa₁ obtained from log(k_cat/K_M)-pD fits (7.8 ± 0.2) is shifted more basic by ΔpKa₁ = + 0.4, whereas the value of pKa₂ (9.2 ± 0.1) remains largely unperturbed. The increased in pKa₁ is consistent with ionizable groups involved with catalysis [29,30].

The pL-profile for Av MDO catalyzed reactions utilizing cys as substrate (Fig. 2C and D) exhibits a ‘bell shaped’ curve suggesting two ionizable groups. However, the second pKa-value is predicted to fall outside of the pH range of the experimental data set (>10). Therefore, k_cat- and k_cat/K_M results obtained in either H₂O or D₂O were fit to equation (2) to obtain pKa₁. In this analysis, data collected beyond pH 9 was neglected. In cys reactions carried out in H₂O, a pKa value of 6.3 ± 0.1 was obtained from log(k_cat)-pH fits. As with 3mpa reactions, no significant perturbation is observed in the value of k_cat in H₂O (0.28 ± 0.02 s⁻¹) as compared to that obtained in D₂O (0.26 ± 0.02 s⁻¹). However, the observed pKa-value is shifted up to 7.1 ± 0.1 (ΔpKa₁=+0.8) in D₂O. Similarly, the observed pKa-value (pKa1) obtained from log(k_cat/K_M)-pL fits of cys reactions in H₂O are shifted upward to from 6.2 ± 0.1 to 7.2 ± 0.1 (ΔpKa₁=+1.0) in D₂O. Regardless, no solvent isotope effect is observed on either k_cat or k_cat/K_M for Av MDO catalyzed reactions with cys. A summary of the all pH/D-dependent kinetic results is provided in Table 1.

For each substrate utilized (3mpa and cys), the solvent kinetic isotope effect (SKIE) on each steady-state kinetic parameters (k_cat and k_cat/K_M) was determined within the pH/D-independent region of each pL-profile. For example, the k_cat/K_M-pL-profile for 3mpa (Fig. 2B) exhibits two pH-independent regions (pL ~ 6.0 and 8.2), both of which exhibit a solvent kinetic isotope effect of 2.7 ± 0.4. This value is significantly larger than possible if solely attributed to differences in the relative viscosity of a D₂O solution as compared to H₂O. Nevertheless, additional viscosity experiments are presented below to determine if the attenuation of the k_cat/K_M-values obtained with 3mpa in D₂O has contributions from viscosity effects. Alternatively, as illustrated in Fig. 2C and D, neither k_cat-pL nor k_cat/K_M-pL profiles for Av MDO catalyzed cys reactions exhibit solvent isotope effects beyond experimental error. Similarly, while the full pD-dependence was not determined for Av MDO catalyzed ca reactions; solvent isotope effects were not observed for either k_cat or k_cat/K_M in assays using ca as a substrate at pL 6.0 or 8.2.

3.2. Proton inventories

Proton inventory experiments were performed to provide greater insight into the observed solvent isotope effect on k_cat/K_M for Av MDO catalyzed 3mpa reactions. In these experiments, both steady-state kinetic parameters (k_cat and k_cat/K_M) were determined for various molar ratios of H₂O and D₂O at both pH/D-independent points (pL ~ 6.0 and 8.2). For each mole fraction of D₂O measured, the value of k_cat/K_M observed ([V/K]_n) is normalized for the value obtained in pure H₂O ([V/K]₀). As illustrated in Fig. 3, the relative change in k_cat/K_M ([V/K]₀/[V/K]_n) decreases with increasing D₂O mole fraction. Proton inventory experiments performed at pL 6.0 (black triangles) and pL 8.2 (white triangles) exhibit nearly equivalent behavior indicating that there is no change in the number of rate limiting protonation steps at either the low or high branch of the pL-profile. Assuming a single proton in flight (solid line, y = 1), the best fit the observed proton inventory data to equation (3) was obtained for a solvent isotope effect of 2.63 ± 0.68 (R² = 0.96₆). Within error, this is in good agreement with the value determined from the k_cat/K_M-pL profile (2.7 ± 0.4). This observation suggests a single solvent exchangeable proton is in flight prior to the first irreversible step. Further, the equivalence of proton inventories within each pL-independent limb (6.0 and 8.2) implies that the ionizable group associated with pKa₁ is not the source of the rate-limiting proton.

Fig. 3.

Proton inventory of MDO catalyzed formation at pL 6.0 (black triangle) and 8.2 (white triangle). Plot of normalized V/K ([V/K]_n/[V/K]₀) versus mole fraction D₂O (n). Experimentally determined SKIE values for k_cat/K_M were used to fit data to equation (3). Proton inventory data were fit assuming a single proton in flight (solid line, y = 1). Fitting results: ^D2OV/K = 2.63 ± 0.68, y = 1, R² = 0.96₆.

3.3. Viscosity effects

The possibility that diffusion controlled events such as substrate and/or product movement into and out of the active site limit the rate of catalysis was evaluated by measuring the perturbation of steady-state parameters (k_cat and k_cat/K_M) as a function of solvent viscosity. Measurements were made at pH 8.2 as described in Material and Methods. For each assay, sucrose was added to increase the relative viscosity of the reaction buffer and the results were fit to equation (4). Fig. 4 illustrates the relative perturbation to k_cat (or k_cat/K_M) obtained for reactions in the absence of viscosigen (k₀) normalized for the values obtained at selected relative viscosities (k_n). Assays using 3mpa and cys as a substrate for Av MDO are shown in Fig. 4 panels A and B, respectively. The dashed lines represent the theoretical limits of this experiment. A positive slope of 1.0 is expected if catalysis is fully diffusion-limited by nonchemical steps in catalysis. Alternatively, the horizontal line represents the expected result in the absence of any diffusional limitation [31]. The absence of any appreciable influence of solvent viscosity on k_cat or k_cat/K_M for either substrate (3mpa or cys) clearly demonstrates that Av MDO catalysis is not limited by either substrate binding or product release. Moreover, this result verifies that the increased viscosity of D₂O buffers does not contribute to the observed SKIE for Av MDO k_cat/K_M data obtained in 3mpa reactions.

Fig. 4.

Effect of solvent viscosity on the maximal rate (v₀/[E]) of MDO catalyzed 3spa and csa formation. A. The effect of solvent viscosity on k_cat (k₀/k_n) and k_cat/K_M [(V/K)₀/(V/K)_n] for 3spa formation is designated by circles (white) and squares (gray), respectively. B. For comparison, the effect of solvent viscosity on k_cat [(k₀/k_n), black circle] and [(V/K)₀/(V/K]_n, gray square) for csa formation was also measured as a function of viscosity. The dashed lines represent the theoretical limits for diffusion-limited product release.

4. Discussion

Among thiol dioxygenase enzymes, the mammalian ‘Arg-type’ CDO has been extensively characterized spectroscopically [32–35], crystallographically [36–38] and kinetically [16,22,23]. However, almost no information is available regarding the relative timing of chemical and non-chemical steps in ‘Gln-type’ MDO catalyzed reactions. A significant point of divergence between CDO and MDO is with respect to substrate coordination to the mononuclear Fe-site. EPR spectroscopic studies indicate that substrate coordination to the Av MDO Fe-site is via thiolate-only [18]. By contrast, all known ‘Arg-type’ CDO enzymes bind cys via bidentate coordination of substrate-thiolate and neutral amine. This difference in Fe-coordination likely explains the relaxed substrate specificity reported among ‘Gln-type’ thiol dioxygenase enzymes [15,18]. Therefore, this subset of enzymes offers a unique point of comparison to better understand the significance of the first- and outer Fe-coordination sphere on catalysis and substrate-specificity among thiol dioxygenase enzymes.

For diffusion-controlled reactions in which product release is rate-limiting, k_cat would be attenuated by increased solvent viscosity. Since addition of D₂O to reaction buffers increases the relative viscosity of reaction mixtures, it is possible that viscosity effects contribute to the apparent solvent isotope effect. However, the results shown in Fig. 4 clearly demonstrate that, regardless of substrate, solvent viscosity has no impact on either k_cat or k_cat/K_M. These experiments verify that non-chemical diffusional steps (product release and/or substrate binding) are not rate-limiting for Av MDO catalyzed reactions with 3mpa or cys. By contrast, product release in Mm CDO is partially rate-limiting [21]. However, substrate (and potentially product) coordination for CDO is bidentate, the resulting chelate effect likely stabilizes both the ES- and EP-complex. By contrast, substrate binding to the Av MDO Fe-site is believed to be via thiolate only, and thus product release is expected to be faster by comparison.

Within the experimental pH range of 6 < pH < 8.5, both k_cat- and k_cat/K_M-pH profiles for Av MDO reactions using 3mpa as a substrate exhibit nonzero limiting values. Only at pH values beyond ~9 does the activity of Av MDO decrease linearly toward zero. By contrast, both k_cat-pH profiles for enzymatic reactions using substrates bearing an amino functional group (cys and ca) decrease to zero within the acidic limb of the profile [18]. These observations are key to the development of a MDO kinetic model.

The pH-dependent behavior observed in 3mpa reactions can be rationalized assuming a diprotic model in which three ionic forms of the enzyme are present but only two of them are catalytically active [39,40]. As illustrated in Scheme 1, three ionic forms of the enzyme are produced by either protonation of the neutral enzyme (designated E^z) to form the cationic E^(z+1) enzyme or deprotonation to yield the anionic E^(z−1) form. The pH-dependent profile for 3mpa reactions suggest that both E^z and E^(z+1) are catalytically active with rates reflected by the asymptotic limits of k_cat within the acidic branch of the pH-profile (27,000 and 10,000 M⁻¹ s⁻¹, respectively). However, the doubly deprotonated enzyme is either unable to bind 3mpa or cannot reductively activate oxygen to generate product. Either way, the E^(z−1) ionic form is catalytically inactive in 3mpa assays.

Scheme 1.

Proposed kinetic mechanism for Av MDO catalyzed 3mpa reaction.

In the context of the reaction scheme proposed, the observed pKa values in k_cat- [pKa₁ =7.4 ± 0.2; pKa₂ = 9.1 ± 0.1] and k_cat/K_M-pH profiles [pKa₁ = 7.8 ± 0.2; pKa₂ = 9.2 ± 0.1] are attributed to catalytically essential ionizable amino acids within the free enzyme [27]. Further, the significant solvent isotope effect (2.7 ± 0.4) observed for 3mpa k_cat/K_M data suggests that proton-dependent steps play an important role in the initial reversible steps leading up to O₂ binding.

Unambiguous identification of the specific ionizing amino acid residue associated with each pKa-value is not possible without direct comparison to Av MDO active site variants. While such experiments are currently ongoing, a reasonable mechanistic hypothesis can be formulated based on: (1) the pKa-values observed in k_cat- and k_cat/K_M-data, (2) steady-state results obtained previously for H155A and Y157F Mm CDO variants, (3) and computationally validated EPR and CD/MCD studies [41,42].

The crystal structure for Av MDO is not available; however, structural comparisons can be made to the Pseudomonas aeruginosa ‘Gln-type’ MDO (PDB code 4TLF) which shares 69.5% sequence identity (84.8% similarity) with Av MDO [15]. As noted previously, the active site of Mm CDO is essentially equivalent to MDO with the notable exception of the C93-Y157 pair and R60 (replaced by Q62 in MDO). The pKa-values observed in 3mpa k_cat- and k_cat/K_M-pL profiles (7.4–7.8), along with the ΔpKa (+0.4) observed in D₂O reactions, is consistent with values reported for catalytically essential histidine residues [29,43]. As illustrated in Fig. 1, all known thiol dioxygenase enzymes have four conserved histidine residues within the active site. Three of which compose the Fe-coordination site (H89, H91 and H142 for Pa MDO, Fig. 1B). Protonation of any of these residues would result in loss of iron and inactivation of the enzyme. Since activity is not abolished below pKa₁ in 3mpa reactions, it can be concluded that these residues are not protonated. This leaves only the conserved His-residue (H157) for further consideration.

The hypothesis that pKa₁ corresponds to H157 is attractive since it has been demonstrated previously that the equivalent residue in Mm CDO (H155) is critical for enzymatic activity. Indeed, despite comparable ferrous iron incorporation, the specific activity of the H155A Mm CDO variant is decreased by ~ 2-orders of magnitude relative to the wild-type enzyme [22]. Furthermore, recent spectroscopic (CD/MCD), crystallographic and computational studies of the Mm CDO H155A variant suggests that H155 serves as a hydrogen bond donor for the Y157 phenol group, thereby positioning the Y157 hydroxyl group of the C93-Y157 pair optimally to donate a hydrogen bond to the Fe-bound cys-carboxylate [42]. Assuming a similar function in the bacterial enzyme, the conserved H157 residue in Pa MDO would be expected to donate a hydrogen bond to the O-atom of the Y159 hydroxyl group to promote interaction with the substrate-carboxylate. This hypothesis presents an attractive explanation for the role of the ‘catalytic triad’ of Ser-His- Tyr residues universally conserved among thiol dioxygenases [44]. In this model, the hydroxyl group of S155 donates a hydrogen bond to the N_ε-atom of H157 to align the opposite Nδ-atom for hydrogen bond donation to the O-atom of the Y159 phenol group. Protonation of H159 would disrupt this hydrogen bond network and result in formation of a positive charge within the enzymatic active site pocket resulting in attenuation of enzymatic activity below pKa₁.

By extension, the basic pKa₂-value observed in 3mpa reactions is close to values typically associated with tyrosine residues (pKa ~10). Furthermore, despite stabilizing a ferrous iron resting state, the Y157F (equivalent to Y159 of Pa MDO) variant of Mm CDO is catalytically inactive, thereby establishing that this residue is also essential for native catalysis [22]. Computational studies validated by EPR spectroscopy confirm that the hydroxyl-group of Y157 within the Mm CDO active site directly interacts with the carboxylate-group of the Fe-bound substrate (cys) [22]. It is therefore reasonable to expect that Y159 of MDO is the second ionizable group observed in 3mpa pL-profiles. A schematic representation summarizing the proposed substrate-bound ionic enzymes forms [E^(z+1)S, E^zS and E^(z−1)S] is shown in Scheme 2.

Scheme 2.

Proposed substrate-bound ionic enzymes forms [E^(z+1)S, E^zS and E^(z−1)S].

Since both pKa-values are observed in k_cat- and k_cat/K_M-pH profiles, it can also be argued that a proton-dependent step is involved in the chemical steps [27]. However, this can also be rationalized in the context of the proposed ionic enzyme forms produced as a function of pH.

As with most non-heme oxidase/oxygenase enzymes, binding and reductive activation of molecular oxygen is gated by substrate coordination to the Fe-site [18]. While the precise mechanism remains a matter of considerable debate, it has been proposed that the coordination of substrate to the Fe-site alters the Fe^II/Fe^III redox couple and/or induces key conformational changes that facilitate direct O₂-coordination and activation [45–47]. Thus, charge stabilization and geometry of the substrate-bound Fe-site are critical for enzymatic activity.

In the doubly deprotonated E^(z−1) form, electrostatic repulsion between the Y159-tyrosinate and 3mpa-carboxylate group would significantly inhibit substrate binding to the opposite face of the 3-His facial triad. For perspective, the distance separating Y159 O-atom and the Fe-bound solvent W3 is 2.67 Å [Fig. 1B]. The short distance separating two negative charges would produce a significant columbic force, likely resulting in both geometric and electrostatic distortions at the substrate-bound Fe-site. Whether the anionic E^(z−1) form cannot bind 3mpa or the E^(z−1)S complex is unable to activate oxygen; the result is the same, abolishment of enzymatic activity beyond pKa₂. This hypothesis also provides an explanation for the absence of a basic pKa in k_cat-pH profiles obtained in assays using ca as a substrate [18]. Unlike 3mpa and cys, cysteamine lacks a carboxylate group. Since there is no electrostatic repulsion between the substrate and the anionic E^(z−1) enzyme, neither substrate-binding nor subsequent O₂-activation are attenuated under basic conditions. Furthermore, the retention of E^(z−1) activity in reactions utilizing ca demonstrates that the second protonation step does not irreversibly inactivate MDO.

Similarly, the differential pH-dependent behavior exhibited within the acidic limb of 3mpa reactions as compared to substrates bearing a positively charged quaternary amine (cys and ca) can also be explained by this kinetic mechanism. As noted previously, both k_cat and k_cat/K_M decrease for Av MDO catalyzed 3mpa reactions, but activity is not abolished within the acidic limb of the pH profile. Conversely, steady-state parameters approach zero for cys and ca enzymatic reactions below pKa₁. Protonation of H157 would produce a cationic E^(z+1) enzyme form. The positive charged imidazole ring of H157 is positioned along the substrate-binding face of the 3-His facial triad (3.90 Å distance from the Fe-bound solvent molecule W3). Therefore, cationic substrates such as cys and ca would experience considerable electrostatic repulsion under acidic conditions. Finally, the zwitterionic cys substrate has both amine- and carboxylate-groups, thus only the neutral E^z enzyme form retains activity, whereas the ionic E^(z+1) and E^(z−1) forms are inactive. As expected, this results in an inverted ‘bell shaped’ curve in k_cat-pH profiles.

Abbreviations

CDO

cysteine dioxygenase

MDO

3-mercaptopropionic acid dioxygenase

3mpa

3-mercaptopropionic acid

cys

L-cysteine

ca

cysteamine (2-aminoethanethiol)

3spa

3-sulfinopropionic acid

csa

cysteine sulfinic acid

ht

hypotaurine

Mm

Mus musculus

Av

Azotobacter vinelandii

Pa

Pseudomonas aeruginosa

HPLC

high performance liquid chromatography

SKIE

solvent kinetic isotope effect