Kinetics of human pyrroline-5-carboxylate reductase in l-thioproline metabolism

Abstract

l-Thioproline (l-thiazolidine-4-carboxylate, l-T4C) is a cyclic sulfur-containing analog of l-proline found in multiple kingdoms of life. The oxidation of l-T4C leads to l-cysteine formation in bacteria, plants, mammals, and protozoa. The conversion of l-T4C to l-Cys in bacterial cell lysates has been attributed to proline dehydrogenase and l-Δ¹-pyrroline-5-carboxylate (P5C) reductase (PYCR) enzymes but detailed kinetic studies have not been conducted. Here, we characterize the dehydrogenase activity of human PYCR isozymes 1 and 2 with l-T4C using NAD(P)⁺ as the hydride acceptor. Both PYCRs exhibit significant l-T4C dehydrogenase activity; however, PYCR2 displays nearly tenfold higher catalytic efficiency (136 M⁻¹ s⁻¹) than PYCR1 (13.7 M⁻¹ s⁻¹). Interestingly, no activity was observed with either l-Pro or the analog dl-thiazolidine-2-carboxylate, indicating that the sulfur at the 4-position is critical for PYCRs to utilize l-T4C as a substrate. Inhibition kinetics show that l-Pro is a competitive inhibitor of PYCR1 $\left(K_{\mathit{\text{IC}}}^{\mathit{\text{app}}}=15.7\text{ mM}\right)$ with respect to l-T4C, consistent with these ligands occupying the same binding site. We also confirm by mass spectrometry that l-T4C oxidation by PYCRs leads to cysteine product formation. Our results suggest a new enzyme function for human PYCRs in the metabolism of l-T4C.

Introduction

l-Thiazolidine-4-carboxylate (l-T4C, a.k.a. l-thioproline, Fig. 1a) is a metabolite found in protozoa, bacteria, plants, and mammals (Cavallini et al. 1956; Elthon and Stewart 1984; Johnson and Strecker 1962; Mackenzie and Harris 1957; Unger and DeMoss 1966). l-T4C is thought to derive from non-enzymatic condensation of l-cysteine and formaldehyde (Mackenzie and Harris 1957). l-T4C, which is an analog of l-proline, has been reported to have both beneficial and toxic effects in organisms. For example, l-T4C treatment inhibits l-Pro oxidation in mitochondria isolated from etiolated barley (Hordeum vulgare) shoots and leads to the beneficial accumulation of l-Pro, which serves as an important osmolyte and energy source for plants during stress (Elthon and Stewart 1984). In contrast, the toxic effects of l-T4C in Escherichia coli are typically attributed to its mis-incorporation into proteins via the bacterial prolyl-tRNA synthetase, which does not discriminate between l-Pro and l-T4C (Deutch et al. 2001; Papas and Mehler 1970). The toxic effects of l-T4C have been exploited in its use as an anti-infection agent. One study analyzed the effects of l-T4C on parasitic cell viability of the non-infectious epimastigote forms of Trypanosoma cruzi, the causative agent of Chagas disease. l-T4C significantly reduced parasitic cell viability and T. cruzi infection rates in CHO-K1 cells (Magdaleno et al. 2009), suggesting that l-T4C may be a treatment for Chagas disease.

Fig. 1.

Chemical structures and proline metabolic reactions. a Structures of l-Pro, l-thiazolidine-4-caboxylate (T4C), dl--thiazolidine-2-carboxylate (T2C), and l-Cys. The structure of l-T4C also shows the two lone pairs of electrons on the sulfur atom with the arrow indicating the pull of electron density away from the nearby C2 atom, thus potentially generating a partial positive charge at C2. b Reactions catalyzed by proline dehydrogenase (PRODH) and l-Δ¹-pyrroline-5-carboxylate (P5C) reductase (PYCR) enzymes. Molecular structures were prepared using ChemDraw (RRID: SCR_016768)

The toxic effects of l-T4C, however, can be diminished via metabolic conversion to l-Cys, thus mobilizing sulfur for key biological roles such as in oxidative stress defense. The ability of l-T4C to serve as an antioxidant precursor has been suggested to be the central mechanism by which it protects plants and mammals against abiotic stress (Furlan et al. 2020). The potential of l-T4C to serve as a source of l-Cys has been explored in humans for various therapeutic applications. The therapeutic potential of l-T4C was first reported in 1957 when l-thioproline was found to be about five times more potent than l-Cys in preventing massive pleural effusions and death in thiourea-treated rats (Mackenzie and Harris 1957). The higher efficacy of l-T4C compared to l-Cys is likely due to the protection of the sulfur atom in the l-T4C ring, which minimizes unwanted reactions and improves cellular uptake (Mackenzie and Harris 1957). Dispensed as the generic drug Timonacic (brand names Hepalidine and Heparegen), T4C has been used for the treatment of hepatic and biliary diseases, particularly in France; the mechanism of action is unclear and thought to involve remedying cell necrosis by correcting levels of transaminase and alkaline phosphatase enzyme activities (Weber et al. 1982).

Despite the interest in l-T4C as a therapeutic l-Cys source in humans and as an oxidative stress protectant in various organisms such as protozoa and plants, detailed kinetic studies with l-T4C have yet to be conducted. Two enzymes that have been implicated in l-T4C metabolism are l-proline dehydrogenase (PRODH) and l-Δ¹-pyrroline-5-carboxylate (P5C) reductase (PYCR), which catalyze the oxidation of l-Pro and the reduction of l-Δ¹-pyrroline-5-carboxylate (l-P5C), respectively (Fig. 1b). Genetic knock-out studies in E. coli and activity assays with bacterial cell lysates suggests PYCR may have a significant role in converting l-T4C into l-Cys, but the individual contributions of PRODH and PYCR are unclear (Deutch 1992; Deutch et al. 2001). In Entamoeba histolytica, l-T4C is required for the survival of the protozoan which relies on l-Cys as the principal low-molecular-weight thiol due to the lack of glutathione (Jeelani et al. 2014). l-T4C is synthesized in E. histolytica by l-Cys reacting with formaldehyde (Jeelani et al. 2014). Using isotope labeling and mass spectrometry, l-T4C was shown to be a metabolic storage source for l-Cys, with l-Cys being liberated when needed for the growth of the protozoan parasite and protection against oxidative stress (Jeelani et al. 2014). Consistent with the metabolite data, lysates of E. histolytica were shown to convert l-T4C into l-Cys, however, the enzyme responsible for the oxidation of l-T4C was not identified (Jeelani et al. 2014).

Here, we set out to determine the ability of the human PYCR (EC 1.5.1.2) enzymes to catalyze the oxidation of l-T4C. The known function of PYCR is to catalyze the last step in l-Pro biosynthesis by NAD(P)H-dependent reduction of l-P5C into l-Pro. In this study, we characterized the ability of PYCR to catalyze the reverse reaction of reducing NAD(P)⁺ with l-Pro and analogs, l-T4C and dl--thiazolidine-2-carboxylate (dl--T2C) (Fig. 1a) at biologically relevant pH 7.5. We report human PYCR1 and PYCR2 enzymes do not exhibit activity with l-Pro or dl--T2C, however, notable activity is observed with l-T4C. PYCR2 was found to have an apparent catalytic efficiency with l-T4C of nearly tenfold higher than PYCR1. l-Pro was also shown to be a competitive inhibitor of PYCR1 with respect to l-T4C, indicating that l-T4C occupies the l-Pro binding site, consistent with a recent X-ray crystal structure of PYCR1 in complex with l-T4C (Christensen et al. 2020). In summary, the experimental evidence here suggests a novel enzyme function for human PYCRs in l-T4C metabolism.

Materials and methods

Materials, reagents, and enzymes

MilliQ-Ultra purified water was used for the preparation of all buffers and chemicals. l-T4C and dl--T2C were purchased from Millipore-Sigma. Human PYCR1 (UniProt Accession ID: P32322) was expressed and purified following a procedure as described previously (Christensen et al. 2017). The expression and purification conditions for human PYCR2 (UniProt Accession ID: Q96C36) were performed according to recently published procedures (Patel et al. 2020, 2021). The experiments were conducted using two independent enzyme preparations each of PYCR1 and PYCR2. Purified PYCR2 was stored in 50 mM HEPES (pH 7.5) buffer containing 250 mM NaCl, 0.5 mM EDTA, 500 μM THPP, and 10% (v/v) glycerol. Protein concentration was quantified by the 660 nm Pierce Protein Dye Assay (ThermoFisher Scientific-Pierce Biotech) as per the manufacturer’s instructions. All protein samples were flash-frozen in liquid nitrogen in 500 μl aliquots and stored at − 80 °C. The N-terminal (8x)His tag was retained in all the purified PYCR2 proteins.

Kinetic characterization of PYCR reverse activity

Stock solutions of l-T4C, dl--T2C, and l-Pro were adjusted to pH 7.5 using 6 M NaOH prior to performing the assays. Assays were performed at 37 °C in 0.1 M Tris-HCl (pH 7.5) reaction buffer containing 0.01% Brij-35 detergent, 1 mM NAD(P)⁺ (pH 8.0), 1 mM EDTA (pH 7.5), and different test substrate compounds of 10 mM l-proline (pH 7.5), 10 mM l-T4C (pH 7.5), and 10 mM dl--T2C (pH 7.5). Three-spot pH test strips (Ricca Chemical Co.) were used to confirm the final pH ≈ 7.5 of the assay mixture for all reactions. The reaction components were pre-warmed for 5 min in a 1.5 mL microtube at 37 °C and then transferred to a 1 cm path length polystyrene rectangular cuvette (USA Scientific). The assay (600 μl total volume) was initiated by adding PYCR1 or PYCR2 enzyme (0.6 μM final concentration) and reaction progress was monitored at 340 nm using a Varian Cary 50-UV/VIS spectrophotometer at room temperature for 10 min duration.

Steady-state kinetic parameters k_cat and K_M were determined with l-T4C by holding NADP⁺ fixed (1 mM) and varying l-T4C (0–25 mM). Additional steady-state kinetic assays were performed by holding l-T4C fixed (60 mM) and varying either NADP⁺ or NAD⁺ (0–15 mM). In assays varying l-T4C, 1 M Tris-HCl (pH 7.5) buffer was added to each assay to maintain an ionic strength of approximately 450–460 mM Tris⁺Cl⁻. l-Pro inhibition studies were performed under the same conditions with 0.6 μM PYCR1 enzyme (final concentration) with N ADP⁺ fixed (1 mM) and l-T4C varied (0–25 mM) at different concentrations of l-Pro (0–20 mM).

Each cuvette measurement was zeroed against blanks omitting l-T4C, and all assays were performed in triplicate. Initial velocities were calculated from the linear increase in absorbance over 1 min during the initial 2.5 min of the assay. NAD(P)H formation was quantified using ε_{340 nm} = 6.22 mM⁻¹ cm⁻¹ (De Ingeniis et al. 2012; Deutch et al. 2001) according to the modern formulation of the Bouguer-Beer-Lambert law of optical spectroscopy (Mayerhofer et al. 2020). All kinetic assays were performed in triplicate and data (mean ± standard deviation) are plotted as variable substrate concentration (mM) versus initial reaction velocity of NAD(P)H product formation (μM s⁻¹). The data were fit by nonlinear regression analysis to the following Henri-Michaelis-Menten equation (Deichmann et al. 2014):

$$v_0=\left(\frac{V_{\text{lim}}\left[S\right]}{K_M^{\mathit{\text{app}}}+\left[S\right]}\right),$$

where v₀ is the initial reaction velocity, V_lim is the theoretical maximal reaction velocity limit, [S] is the substrate concentration, and K_M^app is the apparent Michaelis–Menten constant defining the substrate concentration that yields half-theoretical maximal reaction velocity (Yadav and Magadum 2017). To determine the mode of inhibition by l-proline, four series of inhibition kinetics data were analyzed by global fitting to different inhibition models using the Enzyme Kinetics Wizard Add-in on SigmaPlot (RRID: SCR_003210) version 12.0. Here, the data were fit by linear regression to the following Hanes-Woolf equation (Yadav and Magadum 2017):

$$\frac{\left[S\right]}{v_0}=\frac{K_M^{\mathit{\text{app}}}}{V_{\text{lim}}}+\left(\frac{1}{V_{\text{lim}}}\times \left[S\right]\right).$$

The data fit best to the model of competitive inhibition (Walsh 2018; Yadav and Magadum 2017; Yarlett et al. 2007):

$$v_0=\frac{V_{\text{lim}}\left[S\right]}{K_M\left(1+\frac{\left[I\right]}{K_I}\right)+\left[S\right]},$$

where [I] is the inhibitor concentration and K_I is the apparent competitive inhibition constant (K_IC^app).

Characterization of l-T4C oxidation reaction compounds by electrospray ionization mass spectrometry

Three samples of the l-T4C reaction mixture with no enzyme (control), PYCR1, and PYCR2 were prepared for product analysis by electrospray ionization mass spectrometry (ESI–MS). The total volume for each sample was 300 μl and the reactions were conducted in 1.5-ml tubes under dark conditions with red LED strips being the only light source in the lab. The reaction mixtures contained after mixing (final concentration) 0.3 mM l-T4C (pH 7.5), 1 mM NADP⁺ (stock in 10 mM HEPES, pH 8), 10 mM HEPES (pH 7.5) and ~ 9 mM potassium phosphate (pH 7.5). The reaction mixtures were prewarmed at 37 °C in a water bath for 5 min and then PYCR enzyme was added (1.8 μM final concentration after mixing) except for the control sample (no enzyme). The reaction mixtures were then incubated for an additional 20 min at 37 °C. Next, 1 mM iodoacetamide (IAM) (final concentration after mixing) was added to each sample and the samples were then incubated for another 10 min at 37 °C to allow for alkylation of thiol species, thus yielding carbamidomethylation (CAM) compounds.

Thereafter, samples were transferred into polyethylene V-vials and sealed with Al-tops. The samples were kept at 5 °C in the autosampler that is part of an LC-1200 HPLC (Agilent, Santa Clara, CA). The liquid chromatography mobile phases consisted of acetonitrile (LC-Grade) (A) and 20 mM ammonium acetate in water (LC-Grade) supplemented with 20 mM ammonium hydroxide (B, pH 9.50). The column was an Amide X-Bridge (2.1 mm × 150 mm, Waters, Milford, MA) running at 350 μl/min at 25 °C with a gradient starting at 95% A (5% B) for 2 min, gradient to 55% A over 5 min, gradient to 5% A over 20 min, then 2 min hold at 5% A, followed by re-equilibration with 95% acetonitrile (A) for 10 min. The conditions of the 4000-QTrap mass spectrometer (Sciex, Framingham, MA) were set to operate in single quadrupole, multiple reaction monitoring (MRM), and MS/MS modes. The general electrospray conditions involve the following conditions: electrospray ionization (ESI) potential = 5000 V, temperature = 500 °C, ion source gas 1 = 60 L/min, ion source gas 2 = 80 L/min, curtain gas = 20 L/min, declustering potential = 60 V, collision entrance potential = 10 V, and collision exit potential = 15 V. The collision energy values for the individual compounds were optimized by injections of authentic standards and were within the range of 15–30 V. The data were collected and analyzed with Analyst (RRID: SCR_015785) version 1.6.1 and exported into SigmaPlot (RRID: SCR_003210) version 14.0 for construction of plots.

Results

PYCR1 and PYCR2 catalyze the NAD(P)⁺-dependent oxidation of l-T4C

The ability of PYCR to catalyze the reduction of NADP⁺ with l-Pro, l-T4C, and dl--T2C was examined at pH 7.5. The progress curves shown in Fig. 2a and c demonstrate that NADPH formation is observed with l-T4C as the substrate, but not with l-Pro or dl--T2C. Similar results were obtained in assays using NAD⁺ (Fig. 2b, d) with activity observed only with l-T4C and not with l-Pro or dl--T2C. The inability of PYCRs to oxidize l-Pro at pH 7.5 is consistent with the physiological role of PYCR1 and PYCR2 strongly favoring the forward reaction toward l-Pro biosynthesis as has been postulated for plant PYCR enzymes (Lebreton et al. 2020).

Fig. 2.

PYCR reverse activity assays with l-Pro and thiazolidine analogs. Final enzyme concentration was 0.6 μM for each assay. Progress of the reaction was monitored for 10 min by following the absorbance at 340 nm. Time-course absorbance profiles of PYCR1 reverse activity with a 1 mM NADP⁺ + 10 mM l-Pro or its analog l-T4C or dl-T2C and with b 1 mM NAD⁺ + 10 mM l-Pro or its analog l-T4C or dl-T2C. Time-course absorbance profiles of PYCR2 reverse activity with c 1 mM NADP⁺ + 10 mM l-Pro or its analog l-T4C or dl-T2C and with d 1 mM NAD⁺ + 10 mM l-Pro or its analog l-T4C or dlT2C. Note the vertical axis scale of absorbance value at 340 nm is not the same in all four plots. Data were plotted using SigmaPlot (RRID: SCR_003210) version 12.0

Steady-state kinetic analysis of PYCR-catalyzed oxidation of l-T4C

The apparent steady-state kinetic parameters k_cat and K_M were then determined for PYCR1 and PYCR2 with varying l-T4C and fixed N ADP⁺ substrate (Fig. 3a, b). Both PYCR1 and PYCR2 exhibited saturation kinetics with l-T4C substrate reaching theoretical maximal reaction velocity limits (V_lim) of 0.16 μM s⁻¹ and 0.55 μM s⁻¹, respectively. PYCR2 exhibited higher catalytic activity than PYCR1 with k_cat^app approaching 1 s⁻¹ (Table 1). The nearly fourfold higher k_cat^app and threefold lower K_m^app values for PYCR2 result in a tenfold higher catalytic efficiency with l-T4C relative to PYCR1 (Table 1). Assays were then conducted by holding l-T4C fixed and varying NADP⁺ and NAD⁺ (Fig. 3c–f). PYCR1 shows a nearly 2.6-fold higher specificity for NAD⁺ over NADP⁺, which is most likely due to a nearly 2.9-fold lower K_M^app observed with NAD⁺ (Table 1). In contrast, PYCR2 demonstrates a nearly 1.6-fold higher cofactor specificity for NADP⁺ over NAD⁺ (Table 1).

Fig. 3.

Steady-state kinetics analysis of PYCR reverse activity with l-T4C, NADP⁺, and NAD⁺. Final enzyme concentration of 0.6 μM (a, c, e) PYCR1 and (b, d, f) PYCR2 were used. Data are plotted as (mean ± SD) of three technical replicates and fit to the Henri-Michaelis-Menten equation by non-linear least-squares regression in Sigma-Plot (RRID: SCR_003210) version 12.0. Note the vertical axis scale of initial reaction rate differs in all six plots

Table 1.

Steady-state kinetics analysis of PYCR reverse activity^a with l-T4C

Enzymeb	Varied substrate	Fixed substrate	Km app (mM)	kcat app (s−1)	(kcat app/KM app) (M−1 s−1)
PYCR1	l-T4C	NADP+c	18.9 ± 2.9	0.26 ± 0.02	13.7 ± 2.3
	NADP+	l-T4Cd	1.4 ± 0.5	0.14 ± 0.01	100 ± 39
	NAD+	l-T4Cd	0.49 ± 0.12	0.13 ± 0.01	259 ± 65
PYCR2	l-T4C	NADP+c	6.8 ± 0.9	0.92 ± 0.04	136 ± 19
	NADP+	l-T4Cd	0.53 ± 0.12	0.46 ± 0.02	857 ± 224
	NAD+	l-T4Cd	0.66 ± 0.11	0.36 ± 0.13	551 ± 223

^a(Value ± Std error) are the apparent best-fit parameters from non-linear least-squares fit of the data to the Henri–Michaelis–Menten equation using SigmaPlot 12.0 (RRID: SCR_003210) with data plotted as the (mean ± SD) of n = 3 technical replicates. Different independent enzyme preparations were used for assays varying L-T4C and NAD(P)⁺ substrates

^bFinal enzyme concentration was 0.6 μM

^cFinal NADP⁺ concentration fixed at 1 mM

^dFinal l-T4C concentration fixed at 60 mM

Inhibition of the PYCR dehydrogenase activity by l-Pro

l-Pro inhibition of l-T4C activity in PYCR1 was examined to test whether l-T4C binds to the same site as l-Pro. Upon holding NADP⁺ constant at 1 mM and varying l-T4C substrate (0–25 mM) at different l-Pro concentrations (0–20 mM), the apparent K_M for l-T4C was observed to trend upward while the apparent k_cat remained relatively constant (Table 2 and Fig. 4). With increasing l-Pro concentration, K_M^app increased while V_lim remained fairly unchanged indicating competitive inhibition with an apparent competitive inhibition constant (K_IC ^app) of 15.7 mM l-Pro (Fig. 4a and Table 2) (Yang et al. 2013). Hanes-Woolf plot analysis of the l-Pro inhibition data shows a parallel linear pattern that is consistent with competitive inhibition (Fig. 4b). In particular, the increasing vertical axis intercept (K_M^app/V_lim) effect shown in Fig. 4b indicates l-Pro is a competitive inhibitor (Yarlett et al. 2007) with respect to the l-T4C substrate.

Table 2.

Steady-state kinetic parameters for l-proline ligand inhibition of PYCR1 with varied l-T4C^a,b,c

l-Proline (mM)	Km app (mM)	kcat app (s−1)	(kcat app/KM app) (M−1 s−1)	KIC app (mM)
0	18.9 ± 2.9	0.26 ± 0.02	13.7 ± 2.3	15.7 ± 1.1d
5	32.0 ± 6.3	0.27 ± 0.35	8.4 ± 11.1
10	35.1 ± 6.9	0.26 ± 0.35	7.5 ± 10.1
20	36.8 ± 7.6	0.23 ± 0.03	6.3 ± 1.6

^a(Value ± Std error) are reported for the apparent best-fit parameters of non-linear least-squares fit of the data to the Henri-Michaelis-Menten equation using SigmaPlot (RRID: SCR_003210) version 12.0 with data plotted as (mean ± SD) of n = 3 technical replicates

^bFinal enzyme concentration was 0.6 μM

^cNADP⁺ concentration was fixed at 1 mM while l-T4C concentration was varied 0–25 mM

^dApparent competitive inhibition constant, K_IC ^app, was estimated by globally fitting the data (0.5–25 mM) to a competitive inhibition (full) model using the Enzyme Kinetics Wizard Add-In on SigmaPlot (RRID: SCR_003210) version 12.0

Fig. 4.

l-Pro ligand inhibition kinetics of PYCR1 with varied l-T4C substrate. a Inhibition of PYCR1 by l-proline (0–20 mM) while varying l-T4C (0–25 mM) and holding NADP⁺ constant (1 mM). The curves are the Michaelis-Menten global fit of the four series of data to the competitive inhibition model. b Hanes-Woolf plot of the l-T4C (2–25 mM) data in panel a. Final enzyme concentration of 0.6 μM PYCR1 was used. Data are plotted as (mean ± SD) of n = 3 technical replicates using Enzyme Kinetics Wizard Add-in on SigmaPlot (RRID: SCR_003210) version 12.0

Characterization of reaction products of PYCR enzyme-mediated l-T4C oxidation

To validate that the PYCR reaction with l-T4C results in l-Cys formation, three l-T4C reaction mixtures were analyzed by HPLC-ESI-MS-MRM. The results shown in Fig. 5 reveal intensity peaks for products in the PYCR1 and PYCR2 enzyme-mediated reactions. The mass spectrometry data not only show evidence for 2,3-thiazoline-4-carboxylate (oxidized l-T4C, m/z 132) (Fig. 5a), but also for the formation of Cys-CAM (m/z 179) (Fig. 5c) and N-formyl-Cys-CAM (m/z 207) (Fig. 5d). The control sample (no enzyme) appears to show a slight peak for oxidized l-T4C but there is no evidence for the presence of N-formyl-Cys-CAM (Fig. 5d) and Cys-CAM (Fig. 5c). In addition, samples with PYCR1 and PYCR2 show a peak for NADPH (m/z 746) (Fig. 5b) whereas no evidence of NADPH product is evident with the no enzyme control sample. Concerning the NADPH profile in Fig. 5b, a second peak was detected in the enzyme reactions after 10 min, which corresponds to NADP⁺ (m/z 744).

Fig. 5.

LC-ESI-MRM-MS characterization of four compounds from samples of PYCR-catalyzed l-T4C oxidation reaction: a oxidized l-T4C (2,3-thiazoline-4-carboxylate), b NADPH, c Cys-CAM, and d N-Formyl-Cys-CAM. All three samples [no enzyme (blue), PYCR1 (red) and PYCR2 (green)] consisted of the following concentrations: 0.3 mM l-T4C (pH 7.5), 1 mM N ADP⁺ (in 10 mM HEPES, pH 8), 10 mM HEPES (pH 7.5) reaction buffer, and the remainder volume is 10 mM potassium phosphate (pH 7.5). The enzyme reactions were initiated by adding PYCR at 1.8 μM and incubated at 37 °C for 20 min. Thereafter, 1 mM IAM was added to derivatize thiols. Collected chromatograms and mass spectra data were analyzed with Analyst (RRID: SCR_015785) v1.6.1 and plotted using SigmaPlot (RRID: SCR_003210) v14.0. The corresponding MRM transitions that were monitored are 132/86, 746/729, 179/162, and 207/144 for T4C_ox, NADPH, Cys-CAM, and N-Formyl-Cys-CAM, respectively. Inserts in panels (c) and (d) correspond to the MS/MS spectra for the carbamoylated derivatives of cysteine (m/z = 179.1) and N-Formylcysteine (m/z = 207.1)

Discussion

We found that both PYCR1 and PYCR2 catalyze the oxidation of l-T4C at a significant rate, with PYCR2 exhibiting a tenfold higher catalytic efficiency than PYCR1. Even so, the catalytic efficiencies of PYCR1 (0.47 × 10⁵ M⁻¹ s⁻¹) (Christensen et al. 2017) and PYCR2 (26 × 10³ M⁻¹ s⁻¹) (Patel et al. 2021) are > 10³-fold higher in the reductive reaction with l-P5C and NADPH than in the oxidative reaction with l-T4C and NADP⁺ (13.7 M⁻¹ s⁻¹ and 136 M⁻¹ s⁻¹ for PYCR1 and PYCR2, respectively) (Table 1). These results are consistent with the main function of PYCRs in l-Pro biosynthesis. In previous studies with PYCR1, a 1.90 Å X-ray crystal structure of the binary complex of PYCR1-l-Pro was solved (PDB code 5UAU) (Christensen et al. 2017). More recently, a 2.30 Å X-ray crystal structure of PYCR1 in complex with l-T4C revealed that l-T4C also binds in the l-Pro site (Christensen et al. 2020). Evidence from the product inhibition kinetics here along with structural information strongly suggests that l-Pro and l-T4C compete for the same binding site in PYCR1.

Our kinetic data draw attention to the chemical importance of sulfur in the structure of l-Pro analogs. As depicted in Fig. 1, the sulfur atom at the C4 position of the pyrrolidine ring enables oxidation of l-T4C whereas sulfur at the C3 position, as in dl-T2C, apparently does not. A reaction mechanism for the oxidation of l-T4C by PYCR is proposed in Fig. 6. The mechanism involves a hydride transfer from the C2 position in l-T4C to NAD(P)⁺ followed by hydrolysis to form products l-cysteine and formate. Due to the adjacent electronegative sulfur atom, C2 is likely to be more acidic in l-T4C relative to l-T2C and l-Pro, which may facilitate hydride transfer to NAD(P)⁺ (Howard-Lock et al. 1986). In addition, the pK_a of the nitrogen atom in l-T4C (pK_a = 6.24) is much lower than in l-Pro (pK_a = 10.6) (Parthasarathy et al. 1976; Ratner and Clarke 1937). Thus, the enzyme bound form of T4C is likely to have the secondary amine deprotonated whereas l--Pro is not. Because deprotonation of the amine is an obligate step in the reaction and there is no residue in the active site of PYCR to facilitate deprotonation (Christensen et al. 2017), oxidation of l--Pro would require pH > 9 (Lebreton et al. 2020). These electronic factors may be sufficient enough to enable l--T4C to serve as a reducing substrate for NAD(P)⁺ in PYCR under the assay conditions used here (pH 7.5). The results from our MS analysis of the reaction products with PYCR1 and PYCR2 support our proposed mechanism by providing evidence for an N-Formyl-Cys intermediate and formation of l-Cys during l-T4C oxidation.

Fig. 6.

Proposed mechanism of l-T4C oxidation to l-Cys and formate catalyzed by PYCR, inspired by previous literature sources (Deutch 1992; Jeelani et al. 2014; Mackenzie and Harris 1957). The pK_a of 6.2 for the secondary amine nitrogen favors the deprotonated form at pH 7.5 which facilitates hydride transfer from l-T4C to NAD(P)⁺. The H⁺ balance of this reaction mechanism may be traced from the initial protonated l-T4C ring structure to the final l-Cys product. The protonated l-T4C ring loses two hydrogens in the steps leading to 2,3-thiazoline-4-carboxylate. Upon nucleophilic attack by water and (C–N) bond lysis of N-Formylcysteine, the free sulfur gains one proton and through proton abstractions, two protons are ultimately added to the amine nitrogen to generate l-Cys. Molecular structures and reaction mechanism were prepared using ChemSketch (RRID: SCR_019272)

l-T4C has long been studied as an anti-aging therapeutic. For example, dietary supplementation with l-T4C (2.0 g/kg) in mice increases median and maximal life span by approximately 10–29% in both genders and, improves neuromuscular coordination and exploratory function, as well as protects brain mitochondrial enzyme activities in aged mice (Navarro et al. 2007). The mechanisms by which l-T4C enhances lifespan are multifaceted and likely revolve around its properties as a protector against protein aggregation (Lyu et al. 2020), and as a scavenger of reactive nitrogen and oxygen species, including N-nitroso compounds. The capacity for l-T4C to trap reactive nitrogen species and N-nitroso compounds has been exploited on the drug market for protection against liver diseases and gastrointestinal disorders. In a 2004 study of pathology with Barrett’s esophagus and development of esophageal cancer, the authors evaluated the nitroxide scavenging efficacy of l-T4C using a surgically made duodenal reflux model in rats (Kumagai et al. 2004). Upon histological examination, the l-T4C-treated rats had milder forms of mucosal changes with lesser degrees of inflammation and zero presence of esophageal adenocarcinoma compared to the control group (Kumagai et al. 2004). These exciting results suggest l-T4C inhibits the production of N-nitroso compounds in the gastroduodenal reflux contents, a property of potential therapeutic benefit in esophageal cancer.

Another important clinical aspect of l-T4C is its impact on biogenic aldehydes and potential to be a source of formaldehyde through oxidative metabolism. A 1980 study assessed the clinical toxicity of l-T4C in patients and concluded that overdose of l-T4C overwhelms the metabolic capacity of the liver, thereby leading to l-T4C distribution throughout the body including the cerebrospinal fluid (CSF). It was proposed that metabolic degradation of l-T4C in the intrameningeal tissue releases toxic formaldehyde, which causes the acute central nervous system (CNS) symptoms of epileptic seizures and convulsions (Garnier et al. 1980). Acute toxic effects of l-T4C in animals have also been suggested to involve inactivation of important biogenic aldehydes by l-T4C, particularly in the brain (Weber et al. 1982).

Regarding the physiological significance of PYCR enzyme activity with l-T4C, it is important to consider the in vivo levels of the relevant metabolites. Various reports have determined the concentrations of l-P5C, l-Pro, and l-T4C in human tissues or fluids. For instance, Fleming et al. determined the l-P5C concentration in normal human plasma as 380 ± 110 nM (Fleming et al. 1984). l-Pro levels in human serum have been reported to range in healthy subjects from 125 to 632 μM (14.4–72.8 μg/ml) (Liang et al. 2015). Information on l-T4C levels is limited but a concentration of 5 μM l-T4C in the plasma of head and neck cancer patients was determined (Lankelma et al. 1981). Previously, PYCR2 was proposed to operate under v₀/K_M^app conditions and be sensitive to feedback regulation by l-Pro (K_I ^app = 145 μM). The data here suggest PYCRs catalyze the oxidation of l-T4C under v₀/K_M^app conditions as well but the reaction would be less susceptible to inhibition by l-Pro (K_I ^app = 15.7 mM).

In summary, we show here that PYCR1 and PYCR2 exhibit activity with l-T4C that is likely to be biologically significant and contribute to the physiological effects reported with l-T4C. Our results indicate PYCRs are capable of P5C(T4C)/NAD(P)H oxidoreductase activity, utilizing l-P5C as a hydride acceptor from NAD(P)H as well as utilizing l-T4C as a hydride donor for NAD(P)⁺. PYCR1 exhibits a higher catalytic efficiency with NAD⁺ relative to NADP⁺, but overall the NAD(P)⁺ cofactor preference is modest for both PYCR1 and PYCR2. A key implication of our findings is that human PYCRs may have a physiological role beyond l-Pro biosynthesis that intersects with l-Cys formation. How PYCR enzymes contribute to the physiological effects reported with L-T4C will require cell-based metabolic studies and assays that mimic cellular conditions of NAD(P)⁺ and L-Cys.

Abbreviations

l-T4C

l-Thiazolidine-4-carboxylate

l-Pro

l-Proline

l-Cys

l-Cysteine

dl--T2C

dl--Thiazolidine-2-carboxylate

l-GSAL

l-Glutamate γ-semialdehyde

GSALDH

l-GSAL dehydrogenase

IAM

Iodoacetamide

NADH

Reduced nicotinamide adenine dinucleotide

NAD⁺

Oxidized nicotinamide adenine dinucleotide

NAD(P)H

Reduced nicotinamide adenine dinucleotide (phosphate)

NAD(P)⁺

Oxidized nicotinamide adenine dinucleotide (phosphate)

l-P5C

l-Δ¹-Pyrroline-5-carboxylate

PYCR

Δ¹-Pyrroline-5-carboxylate reductase

Data availability

The datasets generated and analyzed during the current study for the figures and tables are available from the corresponding author on reasonable request.