Evidence for Proline Catabolic Enzymes in the Metabolism of Thiazolidine Carboxylates

Abstract

Thiazolidine carboxylates such as thiazolidine-4-carboxylate (T4C) and thiazolidine-2-carboxylate (T2C) are naturally occurring sulfur analogues of proline. These compounds have been observed to have both beneficial and toxic effects in cells. Given that proline dehydrogenase has been proposed to be a key enzyme in the oxidative metabolism of thioprolines, we characterized T4C and T2C as substrates of proline catabolic enzymes using proline utilization A (PutA), which is a bifunctional enzyme with proline dehydrogenase (PRODH) and l-glutamate-γ-semialdehyde dehydrogenase (GSALDH) activities. PutA is shown here to catalyze the FAD-dependent PRODH oxidation of both T4C and T2C with catalytic efficiencies significantly higher than with proline. Stopped-flow experiments also demonstrate that l-T4C and l-T2C reduce PutA-bound FAD at rates faster than proline. Unlike proline, however, oxidation of T4C and T2C does not generate a substrate for NAD⁺-dependent GSALDH. Instead, PutA/PRODH oxidation of T4C leads to cysteine formation, whereas oxidation of T2C generates an apparently stable Δ⁴-thiazoline-2-carboxylate species. Our results provide new insights into the metabolism of T2C and T4C.

Graphical Abstract

INTRODUCTION

l-proline is unique among the natural amino acids due to the heterocyclic ring and secondary amine. Besides being essential for protein biosynthesis, has been shown to have critical roles in cellular bioenergetics, redox homeostasis, stress response, osmoprotection, and bacterial pathogenesis.^1–5 Biological compounds rich in proline have also been shown to have promising antimicrobial therapeutic properties.^6–9 One group of naturally occurring sulfur-containing proline analogues that are of important interest includes thiazolidine-4-carboxylate (T4C) and thiazolidine-2-carboxylate (T2C) (Scheme 1). T4C is a condensation product of l-cysteine and formaldehyde^10,11 whereas T2C is generated from glyoxylate and cysteamine.¹² T4C has been reported to serve as a nutrient and cysteine source,^13,14 act as an antioxidant,^4,15 and scavenge free radicals such as nitroxide.^16,17 Suggested therapeutic benefits of T4C include treatment of liver diseases,¹⁶ inhibition of cancer cell growth and reversal of advanced cancer,^18–21 and prevention of aging-related diseases.^22,23 T4C is also an important building block of ß-lactam antibiotics.²⁴

Scheme 1.

Chemical Structures of T2C and T4C

Thioprolines have the potential to be toxic due to misincorporation into proteins or inhibition of protein synthesis.^25–28 In Escherichia coli, it was recently discovered that the Xaa-Pro aminopeptidase, PepP, could rescue thioproline toxicity by removing thioproline misincorporated peptides.¹¹ Intracellular toxic levels of T4C have been proposed to be diminished through oxidative metabolism via the enzymes proline dehydrogenase (PRODH) and Δ-¹ pyrroline-5-carboxylate reductase (PYCR).^11,29–31 PRODH was implicated first in l-T4C metabolism in assays using mitochondrial²⁹ and E. coli membrane fractions.³⁰ Later, E. coli PYCR was also reported to catalyze the oxidation of T4C via a NAD⁺-dependent reaction.³¹ d-T2C was shown previously to be a substrate for d-amino acid oxidase which converts D-T2C to Δ²-thiazoline-2-carboxylate.¹²

Although considered to have a prominent role in oxidative metabolism of thioprolines, to our knowledge, the kinetic properties of a PRODH enzyme with T4C or T2C have not been determined. PRODH is a FAD-dependent enzyme that catalyzes the two-electron oxidation of l-proline to Δ¹-pyrroline-5-carboxylate (P5C, Scheme 2). This reaction is coupled to reduction of ubiquinone in the cytoplasmic membrane and in eukaryotes, the mitochondrial respiratory chain.^1,32–34 P5C subsequently undergoes hydrolysis to form l-glutamate-γ-semialdehyde (GSAL), which is then oxidized by the NAD⁺-dependent aldehyde dehydrogenase superfamily member GSALDH (a.k.a., ALDH4A1 and P5CDH) to form l-glutamate (Scheme 2). The four-electron oxidation of L-proline to L-glutamate by the consecutive reactions of PRODH and GSALDH is conserved among prokaryotes and eukaryotes.^35,36 In Gram-negative bacteria, PRODH and GSALDH are encoded in the same polypeptide chain, known as proline utilization A (PutA).³⁵

Scheme 2.

Reactions Catalyzed by Proline Catabolic Enzymes

To provide insights into the metabolic routes of T4C and T2C-oxidation, we investigated the potential of PutA from Sinorhizobium meliloti (SmPutA) to use these compounds as substrates. SmPutA is structurally well characterized,³⁷ and recently, L-T2C was reported to covalently modify the flavin cofactor of SmPutA;³⁸ however, the modification occurred over a long incubation period (measured in days) and the kinetics of T2C-oxidation was not characterized. Using SmPutA, we show here that both T4C and T2C are facile substrates with PRODH but that the oxidized products do not generate aldehyde substrates for GSALDH. We show that T4C is oxidized to Δ²-thiazoline-4-carboxylic acid (Δ²-T4C), which subsequently hydrolyzes to form cysteine, whereas T2C is oxidized to Δ⁴-thiazoline-2-carboxylate (Δ⁴-T2C) and does not generate additional products.

EXPERIMENTAL PROCEDURES

Materials.

All chemicals and buffers were purchased from Fisher Scientific and Sigma-Aldrich, unless otherwise noted. DL-P5C was synthesized according to the method of Williams and Frank and stored in 1 M HCl at 4 °C.³⁹ P5C concentrations were determined by adding o-aminobenzaldehyde (o-AB) and measuring the absorbance at 443 nm (ε_{443 nm} = 2590 M⁻¹ cm⁻¹).³⁹

Purification of SmPutA.

SmPutA was expressed in E. coli BL21(DE3) pLysS using the pNIC28 expression vector as previously described.^37,38 A 5 mL overnight culture grown in Luria-Bertani (LB) medium was used to inoculate a 4 L LB culture. The cells were grown at 37 °C until an OD₆₀₀ of 0.6, at which point, protein expression was induced with 0.25 mM IPTG for 10 h at 18 °C. The cells were harvested and then resuspended (1:10 ratio of wet cell mass to buffer volume) in 50 mM Tris buffer (pH 7.8) containing 300 mM NaCl, 5 mM imidazole, 5% glycerol, five protease inhibitors (3 mM ε-amino-N-caproic acid, 0.3 mM phenylmethylsulfonyl fluoride, 1.2 μM leupeptin, 48 μM tosyl phenylalanyl chloromethyl ketone, and 78 μM tosyllysine chloromethyl 47 ketone hydrochloride), 20 μM FAD, and 0.2% Triton-100. Resuspended cells were lysed on ice via sonication using a model 550 Sonic dismembrator (Fisher Scientific) with a power setting of 5 for 5 min total using a pulse sequence of 5 s on and 15 s off. The cell lysed material was centrifuged at 20,000 g using a JA 20 rotor for 30 min, and the supernatant was collected. The supernatant was passed through a 0.22 μm filter before loading onto a HisPur Ni-NTA Superflow agarose (Thermo Fisher Scientific) packed in a borosilicate glass column. The flow-through was loaded back onto the column again before washing and eluting with buffers resembling the binding buffer except with increasing imidazole concentrations (20, 40, and 300 mM). The eluted SmPutA was collected in 2 mL fractions, and SDS-PAGE under reducing conditions was used to determine the purity of the fractions. The selected fractions were pooled and dialyzed at 4 °C against the dialysis buffer (50 mM Tris, 150 mM NaCl, 0.5 mM EDTA, 0.5 mM Tris(hydroxypropyl)phosphine, 10% glycerol, pH 7.8). The dialyzed protein was then concentrated via the Amicon stirred cell using a 30 kDa molecular weight cutoff ultrafiltration disc from MilliporeSigma. The concentrated protein was flash-frozen in 500 μL aliquots and stored at -80 °C. The N-terminal hexahistidine tag was retained in purified SmPutA. The concentration of PutA was determined using the bicinchoninic acid method (Pierce) and spectrophotometrically using a molar extinction coefficient of 12,700 M⁻¹ cm⁻¹ at 451 nm.⁴⁰

Steady-State PRODH Kinetics.

The PRODH activity of SmPutA with different substrates was studied by monitoring the reduction of the electron acceptor coenzyme Q₁ (CoQ₁) by absorbance at 278 nm (ε₂₇₈ = 14.3 mM⁻¹ cm⁻¹)⁴¹ using 1 mL cuvettes and a Varian Cary 50 UV-visible spectrophotometer, as previously described.⁴² The assay buffer contained 50 mM potassium phosphate and 25 mM NaCl at pH 7.5. All substrate stock solutions were made in sodium phosphate buffer (pH 7.5). The substrates and CoQ₁ (final concentration of 0.2 mM) were incubated in the assay buffer for 5 min at room temperature before SmPutA (final enzyme concentration of 0.25 μM) was added to initiate the reaction. The substrate concentration was varied as follows: l-proline, 0–200 mM; l-T4C, 0–3 mM; and L-T2C, 0–2.5 mM. We note that T2C is commercially available only as a racemic mixture. All data are reported using the concentration of l-T2C, which was considered to be half of the dl-T2C concentration. The initial velocity (v₀) was determined from the linear portion of the progress curve (0–5 min). Assays were performed in triplicate. Kinetic parameters were estimated by fitting initial rates to the Michaelis-Menten equation using SigmaPlot 12.0.⁴³ The plot for v₀/[E] versus substrate concentration is shown, where V is the maximal reaction velocity of CoQ₁ reduction and [E] is the concentration of SmPutA used in the assays (0.25 μM).

Steady-State PRODH-GSALDH-Coupled Activity Assay.

The PRODH-GSALDH coupled activity of PutA with l-proline, l-T2C, and l-T4C was studied as previously described.⁴⁴ The reactions contained 0.25 μM SmPutA, 0.2 mM CoQ₁ as an electron acceptor, 0.2 mM NAD⁺, and l-1 proline (20 mM), l-T4C(5mM),or l-T2C(5 mM)in the assay buffer (50 mM potassium phosphate,25 mM NaCl, pH 7.5). The coupled activity was measured by monitoring NADH production at 340 nm (ε₃₄₀=6.22 mM⁻¹cm⁻¹).

Single-Turnover PRODH and Coupled Activity Kinetics.

All stopped-flow kinetic measurements were conducted with a Hitech Scientific SF-61DX2 stopped-flow instrument equipped with a photodiode array detector at 23 °C as previously described.³⁸,⁴⁴ Enzyme, buffer, and substrate solutions were degassed using 12 cycles of vacuum and nitrogen. All substrate solutions were made in 50 mM phosphate buffer (25 mM NaCl, pH 7.5). In addition, 100 μM protocatechuic acid (PCA) and 0.05 U/mL protocatechuate 3,4-dioxygenase (PCD) were included in the substrate solutions to scrub residual oxygen.⁴⁵ SmPutA enzyme was rapidly mixed by stopped flow with an equal volume of the following substrates at varying concentrations (all reported concentrations are after mixing): l-T4C (0–100 mM), l-T2C (0–100 mM), and l-proline (0–1000 mM). The reactions were monitored by multiwavelength absorption from 300 to 700 nm. The decrease in absorbance at 450 nm was fit to eq 1 using SigmaPlot 12.0 to obtain the observed rate constants (k_obs1 and k_obs2), where A₀ is the initial absorbance at 450 nm and A₁ and A₂ are the corresponding amplitudes. K_d values were determined by plotting k_obs1 versus substrate concentration. The maximum observed rate constant (k_max) and the apparent K_d were determined by fitting k_obs1 to eq 2 as a function of substrate concentration ([S]).

$$A_{(450)}=A_0+A_1*e^{(-k\mathrm{obs}1*t)}+A_2*e^{(-k\mathrm{obs}21*t)}$$ (1)

$$k_{\mathrm{obs}1}=(k_{\max }*[\mathrm{S}\left]\right)/(K_{\mathrm{d}}+[\mathrm{S}\left]\right)$$ (2)

To test PRODH-GSALDH-coupled activity under single turnover conditions, SmPutA enzyme was rapidly mixed by stopped flow with an equal volume of substrate solution containing 0.2 mM NAD⁺ and 20 mM l-T4C, l-T2C, and l-proline. The reactions were monitored by multiwavelength absorption from 300 to 700 nm.

Prediction of pK_a Values.

The pK_a values of the nitrogen in proline, T2C, T4C, P5C, Δ²-T4C, and Δ⁴-T2C were calculated using the structure-based chemical property prediction tool Chemicalize (https://chemicalize.com).⁴⁶

Characterization of Products by Electrospray Ionization Mass Spectrometry.

Electrospray ionization mass spectrometry (ESI-MS) was used to characterize the products generated during the reaction of SmPutA with l-T2C and l-T4C. The reaction mixtures contained (final concentration) 0.5 mM substrate (l-T4C or l-T2C) (pH 7.5) and 0.5 mM CoQ1 in 50 mM potassium phosphate buffer (50 mM NaCl, pH 7.5) after mixing. The reactions were initiated by adding PutA (2 M final concentration after mixing) except for the control sample (no enzyme). The reaction mixtures were then incubated for 60 min at 23 °C, followed by ESI-MS analysis. To test whether the reaction products also lead to cysteine formation, 1 mM iodoacetamide (final concentration after mixing) was added to each sample and the samples were incubated for an additional 10 min at 37 °C to allow for alkylation of thiol species, which yields stable carbamidomethylation (CAM) product compounds.

The samples were then transferred into polyethylene V-vials and sealed with Al-tops. The samples were kept at 5 °C in the autosampler of the LC-1200 HPLC (Agilent, Santa Clara, CA) prior to column chromatography. The samples were applied onto an Amide X-Bridge (2.1 mm × 150 mm, Waters, Milford, MA) column 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 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 9.5). The 4000-QTrap mass spectrometer (Sciex, Framingham, MA) was set to operate in single quadrupole, multiple reaction monitoring (MRM), and MS/MS modes. The general electrospray conditions were as follows: 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 version 1.6.1 and plotted using SigmaPlot 14.0.

Nuclear Magnetic Resonance Analysis of T2C and T4C Reaction Products.

NMR spectra were acquired on a Bruker AVANCE NEO 600 MHz NMR spectrometer equipped with a 5 mm TCI-H/F cryoprobe. All spectra were acquired under dark lab conditions with overhead lights switched off. Proton spectra were gathered before and after addition of the enzyme to the reaction mixtures. Deuterium oxide was added to the reaction mixture for spectrometer lock with l-T2C, or for l-T4C, the reaction mixture was prepared primarily using deuterium oxide instead of water. For NMR analyses, CoQ₁ was dissolved in DMSO-d₆ and quantified using an extinction coefficient of 14.3 mM⁻¹ cm⁻¹ at 275 nm. In the reaction (250 μL total volume) with l-T2C, 2 μM SmPutA was mixed with 0.5 mM CoQ₁, and 0.5 mM l-T2C in 50 mM phosphate buffer (50 mM NaCl, pH 7.5). A total of 20 μL of deuterium oxide was added to the reaction mixture as noted above. The reactions were initiated by adding SmPutA and quickly transferred to an NMR tube and incubated for 60 min at 25 °C during data acquisition. For the reaction with l-T4C, SmPutA was first flash-frozen and lyophilized and then resuspended in deuterated water. Then, 50 mM potassium phosphate buffer (50 mM NaCl, pH 7.5) and l-T4C were also prepared in deuterated water. The reaction mixture (250 μL total volume) included 0.5 mM CoQ₁, 0.5 mM l-T4C and 2 μM SmPutA. The reactions with l-T4C were initiated by adding SmPutA and incubated in an NMR tube for 60 min at 25 °C during data acquisition.

RESULTS

Steady-State PRODH Kinetics.

The ability of SmPutA to use l-T2C and l-T4C as substrates was first tested using the PRODH activity assay. SmPutA displayed robust PRODH activity with both compounds (Figure 1). The steady-state kinetic parameters for l-T2C, l-T4C, and l-proline are reported in Table 1. The k_cat values range 0.7–1.4 s⁻¹, whereas the K_M values for the thiazolidine carboxylates are about 30-fold lower relative to l-proline. As a result, the catalytic efficiencies with l-T2C and l-T4C are about 20–30 times higher than with l-proline. These results show that the SmPutA PRODH domain effectively catalyzes the oxidation of l-T2C and l-T4C.

Figure 1.

PRODH activity with thiazolidine carboxylates. SmPutA (0.25 μM) was mixed with (A) l-proline (0–200 mM), (B) l-T2C (0–2.5 mM), and (C) l-T4C (0–3 mM) as the varied substrate and 0.2 mM CoQ₁ as the fixed substrate.The data are fit to the Michaelis–Menten equation.

Table 1.

Steady-State Kinetic Parameters of SmPutA PRODH Activity with Thioproline Substrates^a

varied substrate	KM (mM)	kcat (s−1)	kcat/KM (M−1 s−1)
l-T2C	0.22 ± 0.05	0.7 ± 0.1	3000 ± 1600
l-T4C	0.24 ± 0.08	1.2 ± 0.1	4900 ± 2100
l-proline	7.7 ± 1.8	1.4 ± 0.1	149 ± 41

^aassays used 0.25 μM SmPutA and 0.2 mM CoQ₁ and, 0–2.5 mM l-T2C, 0–3 mM l-T4C, or 0–200 mM l-proline.

A hallmark of PutAs is the ability to catalyze the coupled PRODH/GSALDH reaction involving P5C as the intermediate. To test whether the oxidized products of l-T2C and l-T4C from the PRODH reaction can be used as substrates by and l-T4C did not result in NADH formation, indicating that oxidation of these thiazolidine carboxylates by the PRODH domain does not generate a useable substrate for GSALDH.

Single-Turnover Studies.

The rates of l-T2C and l-T4C reduction of SmPutA bound-FAD were determined next by rapidly mixing SmPutA with the different substrates. All substrates resulted in complete reduction of the bound FAD, as shown in Figure 3. Reduction of the FAD occurred in two phases, with k_obs1 corresponding to the major amplitude change in the spectrum. The maximum observed rate constant for FAD reduction was determined by plotting k_obs1 versus substrate concentration. As shown in Figure 4 and reported in Table 2, l-T4C resulted in the highest maximal observed rate constant (k_max), about fourfold higher than that determined with proline. Thus, l-T2C and l-T4C are facile reducing substrates of the FAD in SmPutA.

Figure 3.

UV-visible spectra of PutA reduction by thioproline substrates. Reduction of PutA (32 μM, after mixing) with 20 mM (after mixing) (A) l-proline, (B) l-T2C, and (C) l-T4C. Shown in each inset is the fit of the absorbance change at 450 nm to a double exponential decay (eq 2). The k_obs1 values estimated from the best fits are 7.2 s⁻¹ (l-proline), 32 s⁻¹ (l-T2C), and 53 s⁻¹ (l-T4C).

Figure 4.

Single-turnover kinetics of SmPutA PRODH domain. SmPutA (32 μM, after mixing) was mixed with 0–200 mM l-proline (A), 0–50 mM l-T2C (B), and 0–100 mM l-T4C (C). Shown is the fit of the data to eq 2 from which the parameters k_max and K_d were determined.

Table 2.

Single-Turnover Kinetic Constants of SmPutA PRODH Activity

varied substrate	kmax (s−1)	Kd (mM)
l-proline	43.4 ± 5.7	48 ± 17
l-T2C	91.4 ± 10.8	9.1 ± 3.4
l-T4C	181 ± 19	26 ± 8

The coupled reaction of SmPutA was then monitored by stopped flow to test whether oxidation of l-T2C and l-T4C leads to formation of NADH under single-turnover conditions. Figure 5 shows reduction of FAD and NADH formation measured at 450 and 340 nm, respectively. As expected, mixing SmPutA with proline causes reduction of FAD and generates NADH, consistent with coupled PRODH-GSALDH activity. In contrast, mixing SmPutA with l-T2C and l-T4C results only in reduction of FAD with no NADH formation observed. These results confirm that the oxidized products of l-T2C and l-T4C are not substrates for the GSALDH domain in SmPutA.

Figure 5.

Stopped-flow reaction of coupled PRODH-GSALDH activity in SmPutA. SmPutA (25 μM, after mixing) was rapidly mixed with 0.2 mM NAD⁺ and 20 mM l-proline, l-T2C, or l-T4C. Shown are the changes in concentration for FADox (squares) and NADH (open circles) for the reaction of PutA with each substrate. No NADH formation is observed with l-T2C and l-T4C.

MS and NMR Analysis of T2C and l-T4C Products.

The oxidized products of l-T2C and l-T4C generated by SmPutA were then characterized by mass spectrometry and NMR. Mass spectrometry data of the reaction of SmPutA with l-T4C and l-T2C showed evidence for 2,3-thiazoline-4-carboxylate (or Δ²-T4C, m/z 132) and 3,4-thiazoline-2-carboxylate (or Δ⁴-T2C, m/z 132), respectively (Figure 6). l-T4C and l-T2C, which have the same parent 134 m/z, share the same oxidized m/z of 132. The new 132 m/z species is consistent with a hydride transfer to the bound FAD and formation of a double bond during the reaction with PutA. Because mass spectrometry did not distinguish at which carbon l-T4C and l-T2C are oxidized, NMR was used to further confirm the reaction products. Figure 7A shows the NMR spectrum of the products from the reaction of PutA and l-T2C. The results provide evidence that the C4 atom is oxidized in the reaction (i.e., Δ⁴-T2C is the product). Likewise, oxidation of the C2 atom in l-T4C was indicated in the NMR spectrum of the reaction with l-T4C (Figure 7B). Because water suppression interfered with assignment of the proton at the C4' position of Δ²-T4C at 4.80 ppm, a sample was prepared using only deuterium oxide as the solvent and analyzed by COSY NMR. The COSY NMR spectra (Figure S1) show clear assignment of the C4' proton of Δ²-T4C, further confirming Δ²-T4C as the product. These results suggest that l-T2C and l-T4C are oxidized by SmPutA in a hydride transfer mechanism similar to proline.

Figure 6.

Mass spectrometry of l-T2C and l-T4C reaction products. (A) l-T4C before and (B) after the reaction with SmPutA. (C) dl-T2C before and (D) after the reaction with SmPutA. Oxidized products of T4C and T2C are indicated by the appearance of the 132 m/z peak. Reaction mixtures contained 2 μM SmPutA, 0.5 mM CoQ₁, 0.5 mM l-T2C, or l-T4C in 50 mM potassium phosphate buffer (50 mM NaCl, pH 7.5).

Figure 7.

NMR spectra of T2C and T4C reaction products. (A) Proton NMR spectrum of Δ⁴-T2C after reacting l-T2C and SmPutA for approximately 1 h. (B) Proton NMR spectrum of Δ²-T4C after reacting l-T4C and SmPutA for approximately 1 h. Proton of the C4' position of Δ²-T4C at 4.80 ppm is not visible due to water suppression. NMR spectra of T2C and T4C before addition of SmPutA are provided in Figure S2.

To test whether PutA-mediated oxidation of l-T2C and l-T4C leads to cysteine formation, the reaction products were incubated with iodoacetamide and analyzed by ESI-MS. Figure S3 shows a Cys-CAM species (m/z 179) in the reaction with l-T4C, indicating that cysteine is generated in the reaction. No Cys-CAM was detected in reactions with l-T2C; thus, formation of a free sulfur species is unique to oxidation of l-T4C.

DISCUSSION

PutA was shown here to readily use l-T2C and l-T4C as substrates for the PRODH reaction. Rapid reaction kinetics show that the maximum rate of FAD reduction is faster with the thiazolidine carboxylate substrates than with l-proline and the lower K_d values estimated for l-T2C and l-T4C compared to l-proline indicate that these alternative substrates bind with strong affinity to the PRODH active site. The oxidative step was previously determined to be rate-limiting for k_cat during catalytic turnover for the proline/ubiquinone oxidoreductase activity of PutA.⁴⁷ Thus, although the rate of flavin reduction is higher with T2C and T4C, the overall k_cat does not increase. The catalytic efficiency of PutA with the thiazolidine carboxylates is significantly higher than with L-proline due to the lower K_m values for T2C and T4C. L-T2C was shown recently to be a mechanism-based inactivator of SmPutA; however, the inactivation is very slow with an apparent pseudo-first-order rate constant of 0.44 days^−1.38 Thus, inactivation of SmPutA by T2C is not expected to occur under the conditions used here for the steady-state and rapid reaction experiments.

Covalent modification of FAD by l-T2C was observed in crystals of SmPutA that were derived from cocrystallization of SmPutA and DL-T2C³⁸. A crystal structure of the inactivated enzyme showed that l-T2C forms a covalent bond between the C5 of T2C and N5 of reduced FAD.³⁸ The T2C covalent adduct was observed to occupy the same site as L-tetrahydro-2-furoic acid (L-THFA), a competitive inhibitor that mimics proline binding and ion-pairs with active site residues Arg488, Arg489, and Lys265 (Figure 8A).³⁸ The structural similarity of l-T2C and l-T4C to l-proline suggests that these alternative substrates bind to the PRODH active site like l-proline. Manual docking calculations based on the L-THFA complex structure (PDB 5KF6) suggest that the C4 of l-T2C and the C2 of l-T4C can approach the FAD N5 within 3.2 Å without introducing any steric clashes in the structure. For reference, the analogous distance for the L-THFA complex is 3.3 Å. These results are consistent with l-T2C and l-T4C binding PRODH like L-THFA and proline (see Figure 8A), allowing for hydride transfer from the C4 and C2 atoms of l-T2C and l-T4C, respectively, to the FAD N5 in SmPutA. A reaction mechanism for the reduction of SmPutA-bound FAD by l-T4C is proposed in Figure 8B. The electronegative sulfur atom in the pyrrolidine ring likely increases the acidity of the C4 and C2 atoms in l-T2C and l-T4C, respectively, enabling these compounds to be facile reducing substrates of the bound FAD.⁴⁸ Using the software program Chemicalize (http://www.chemaxon.com), pK_a values for the secondary amine in T4C and T2C were both calculated to be 7.7. The calculated value for T4C is higher than the experimentally determined pK_a value of 6.24 reported previously for T4C.⁴⁹ Thus, L-T2C and l-T4C may already have the secondary amine deprotonated in the enzyme bound form due to the much lower pK_a of the nitrogen atom relative to L-pro (calculated pK_a = 11.3).^49,50 These chemical properties may explain why l-T2C and l-T4C are faster reducing substrates than proline.

Figure 8.

Structure of the substrate binding site and proposed hydride transfer mechanism. (A) Structure L-THFA bound to the PRODH active site in SmPutA (PDB 5KF6).³⁷ Figure adapted from ACS Chem. Biol. 2020.³⁸ (B) Proposed oxidation mechanism of l-T4C. Mechanism was drawn using ChemDraw 20.

Figure 9 shows the steps for the nonenzymatic hydrolysis of P5C that leads to formation of GSAL, the substrate for the GSALDH domain. An interesting observation here is that no formation of NADH is observed with T4C and T2C, indicating that unlike P5C, Δ²-T4C and Δ⁴-T2C are not able to continue along the PRODH/GSALDH coupled reaction pathway. It was also noted in the inactivation study of SmPutA with T2C that NADH was not generated under single-turnover conditions.³⁸ In considering possible explanations for the lack of useable substrates for GSALDH, we sought to determine whether oxidation of these compounds leads to alternative products. We detected cysteine as a reaction product with L-T4C, which as depicted in Figure 9, is formed by hydrolysis of Δ²-T4C to N-formyl-cysteine and subsequently hydrolyzed to release cysteine and formate.^11,51 In contrast to T4C, SmPutA oxidation of T2C did not generate other detectable species besides Δ⁴-T2C (Figure 9).

Figure 9.

Predicted pK_a values for P5C and thiazolidine carboxylates and hydrolysis products. pK_a values for protonation of the nitrogen in P5C, Δ²-T4C, and Δ⁴-T2C. Hydrolysis of the protonated species is shown generating GSAL and N-formyl-Cys from P5C and Δ²-T4C, respectively. Hydrolysis of N-formyl-Cys to generate free cysteine and formaldehyde is not shown. Hydrolysis products from Δ⁴-T2C were not detected. Figure drawn with ChemDraw 20.

Curious as to why Δ⁴-T2C appears less susceptible to hydrolysis under the assay conditions (pH 7.5), we considered that the sulfur atom may have a significant impact on the pK_a of the nitrogen. Hydrolysis and ring opening of P5C to form GSAL is favored when the nitrogen is protonated (Figure 9). The pK_a values of the nitrogen in P5C, Δ²-T4C, and Δ⁴-T2C were thus estimated using the software program Chemicalize from from ChemAxon (http://www.chemaxon.com). The predicted pK_a of the nitrogen atom in P5C is 6.1, which is slightly lower than the experimentally determined value of 6.67.⁵² Surprisingly, the pK_a values for the nitrogen atom in Δ²-T4C (pK_a 1.8) and Δ⁴-T2C (pK_a 1.5) are predicted to be over 4 pH units more acidic than that of P5C (Figure 9). This highly suggests that the nitrogen is predominantly deprotonated in the enzyme-bound forms of Δ²-T4C and Δ⁴-T2C. Thus, hydrolysis may be impeded by the much lower pK_a of the nitrogen atom in Δ²-T4C and Δ⁴-T2C (Figure 9). In the case Δ²-T4C, however, the ring-opening step that leads to N-formyl-cysteine formation may shift the equilibrium in favor of hydrolysis.

In conclusion, steady-state and single-turnover kinetic studies provide strong evidence that l-T2C and l-T4C are facile substrates of the PRODH active site in PutA. The ESI-MS and NMR results support a hydride transfer mechanism, leading to the oxidized products Δ²-T4C and Δ⁴-T2C. The carbon-sulfur bond in thioproline has been shown to increase the perimeter of the pyrrolidine ring and impact the energy levels of different conformers relative to proline.²⁴ This could enhance binding of thioproline to the PRODH active site, which may be reflected in the lower K_m and K_d values determined for l-T2C and l-T4C. Potentially, this would explain why l-T4C has been reported to inhibit proline metabolism at concentrations much less than proline.⁵³ Another chemical effect of sulfur that likely enhances FAD reduction by thioproline is the increased acidity of the nitrogen and adjacent carbon atom. Altogether, PutA exhibits efficient catalytic activity with l-T2C and l-T4C, thus further implicating PRODH as a key enzyme for lowering intracellular levels of thioprolines and, in the case of l-T4C, generating cysteine.

Figure 2.

SmPutA-coupled activity. SmPutA (0.25 μM) was mixed with 0.2 mM NAD⁺, 0.2 mM COQ₁, and proline (20 mM),l-T2C (5 mM), or l-T4C (5 mM). NADH formation with l-proline is indicated by an increase in absorbance at 340 nm.

ABBREVIATIONS

ALDH

aldehyde dehydrogenase

CAM

carbamidomethylation

ESI-MS

electrospray ionization mass spectrometry

GSAL

l-glutamate-γ-semialdehyde

GSALDH

l-glutamate-γ-semialdehyde dehydrogenase

NMR

nuclear magnetic reso-Δ¹-pyrroline-5-carboxylate

PRODH

proline dehydrogenase

PutA

proline utilization A

SmPutA

proline utilization from Sinorhizobium meliloti

Δ⁴-T2C

Δ⁴-thiazoline 2-carboxylate

Δ²-T4C

Δ²-thiazoline-4-carboxylate

T4C

thiazolidine-4-carboxylate

T2C

thiazolidine-2-carboxylate (T2C)

THFA

tetrahydro-2-furoic acid