Properties of recombinant TgCBS
We identified the Toxoplasma gene likely to encode the CBS enzyme by searching the Toxoplasma genome database (https://toxodb.org/toxo/). The T. gondii (strain ME49) genome contains a single-copy gene for CBS (TGME49_259180) encoding a multimodular protein of 514 amino acids (55.8 kDa) that lacks the N-terminal heme-binding motif preceding the catalytic core domain observed in humans but possesses the Bateman module. Moreover, TgCBS also possesses the oxidoreductase (CysXXCys) motif present in the catalytic core of human CBS. Notably, CBSs from other protozoa, e.g. T. cruzi and L. major, lack the N-terminal heme domain, the C-terminal Bateman module and the oxidoreductase motif. The amino acid sequence similarity of TgCBS with its homologs varies significantly along the different domains of its polypeptide chain. For example, TgCBS maintains high identity with the catalytic core of eukaryotic CBSs (e.g. ~ 62%, 56%, 56%, 49%, 53% and 54% identity with respect to humans, fly, honeybee, yeast, Trypanosoma, and Leishmania, respectively), but shows a low similarity with the regulatory domain of these proteins [e.g. ~ 8%, 10%, and 8% with respect to humans, fly and yeast CBSs, respectively (Trypanosoma and Leishmania lack this domain)] (Fig. 1c). These distinct features determine the affinity for the different substrates and regulatory molecules (i.e. SAM in HsCBS), and thus the preferent catalytic activity and regulatory mechanisms across the organisms.
TgCBS was overexpressed in E. coli and purified as His-tagged protein with purity higher than 95% as judged by SDS-PAGE (Fig. 2a). The recombinant protein was yellow and exhibited a UV-visible absorption spectrum with a major peak at 410 nm characteristic of the ketoenamine tautomer of the internal aldimine (PLP bound to active site Lys56) (Fig. 2b)30. No evidence of a heme group was found as illustrated by the absence of Soret band at 430 nm in the absorption profile. In solution, the protein was predominantly present as a dimer (~ 97 kDa) with some high order oligomers [e.g., tetramer (~ 234 kDa), Fig. 2c], in accordance with a monomer molecular mass of ~ 56 kDa. TgCBS binds ~ 1 mol of PLP/mol of monomer with a Kd value for PLP of 0.13 ± 0.01 µM, as deduced by fluorescence titrations of apo-TgCBS with PLP (Fig. 2d). Moreover, PLP affects the thermal stability of the enzyme. Compared to the sample containing apo-protein, the melting temperature (Tm) of TgCBS with PLP increased from 44 ± 1 °C to 56 ± 1 °C. (Fig. 2e).
Steady-state characterization of TgCBS
We determined the steady-state kinetic parameters for TgCBS in the canonical and H2S-generating alternative reactions described in Fig. 1a.
CBS canonical reactions
The steady-state kinetic parameters of TgCBS in the canonical reactions were determined by applying a highly sensitive continuous assay based on recombinant cystathionine beta-lyase (CBL) from Corynebacterium diphtheriae produced in our laboratory31,32 and on commercial lactate dehydrogenase (LDH) as coupling enzymes, following a method described in Ref.33 (see “Materials and methods”). Nascent l-Cth is captured by CBL and converted to l-Hcys, NH3, and pyruvate, which is then detected by LDH assay (decrease in absorbance at 340 nm, reflecting the oxidation of NADH by LDH) (Supplementary Fig. S1a). The continuous nature of the assay avoids accumulation of the l-Cth product, which can compete with l-Ser for free enzyme, therefore preventing the phenomenon of product inhibition.
To test the usefulness of the coupled-coupled enzyme assay to detect CBS activity, we first investigated the kinetic parameters of our recombinant CBL in catalyzing the β-elimination of l-Cth to pyruvate, l-Hcys and NH3 (kcat = 93 ± 2 s−1, Km = 0.8 ± 0.1 mM, kcat/Km = 116 mM−1 s−1) and of l-Ser to pyruvate and NH3 (kcat = 1.2 ± 0.2 s−1, Km = 7.5 ± 1.4 mM, kcat/Km = 0.16 mM−1 s−1) under the conditions used for the CBS coupled-coupled assay. The obtained values agree with those previously published by our laboratory31, 32. Importantly, the catalytic efficiency of CBL toward l-Ser was < 0.2% compared to l-Cth, and thus this low activity does not interfere with the accuracy of the assay. Next, the assay was optimized for the amount of auxiliary enzymes by measuring the NADH oxidation rate in standard assay mixtures containing different concentrations of CBL or LDH. It was necessary to use 1.5 µM CBL and 2 µM LDH in the coupled-coupled assay because these concentrations are each well into the plateau region of coupling enzymes for the range of TgCBS concentrations assayed (dependence of NADH oxidation rate in the coupled-coupled assay was found to be linear in the 0.2–2 μM TgCBS concentration range) (Supplementary Fig. S1b).
Initially, the CBS assay was performed with constant substrate concentrations (10 mM l-Ser and 0.8 mM l-Hcys) from pH 5.5 to 9.5 and the optimum activity was observed around pH 9 (Supplementary Fig. S2a). Thus, pH 9 was used for further CBS enzymatic characterization. Moreover, we evaluated residual activity of purified TgCBS after thermal stresses of 10 min at temperatures ranging from 30 to 70 °C. We found a T50 (half-inactivation temperature) of 44.6 ± 0.3 (Supplementary Fig. S2b).
Experimental data for the condensation of l-Ser and l-Hcys to l-Cth catalyzed by TgCBS are shown in Fig. 3a,b. The phenomenon of substrate inhibition was negligible for L-Ser and evident for l-Hcys as pointed out by the decrease in initial velocity at high substrate concentrations. Both human CBS34 and yeast CBS4,33 are characterized by substrate inhibition by l-Hcys. We collected a large data set as necessary for a bi-substrate system and the data were fit according to Eq. (1), which includes a Ki value representing the inhibition constant for substrate inhibition by l-Hcys4,33. The kinetic parameters are summarized in Table 1. TgCBS showed a kcat for condensation of l-Ser and l-Hcys of 6.3 ± 0.4 s−1 and Km values for l-Ser and l-Hcys of 0.42 ± 0.04 mM and 0.23 ± 0.03 mM, respectively (Table 1). At high concentration of L-Hcys, the enzyme displayed substrate inhibition with a Ki value of 1.0 ± 0.1 mM (Table 1). Thus, the parasitic enzyme is significantly inhibited by l-Hcys which can either compete directly with the l-Ser substrate for the binding site on the free enzyme (E) or bind to the enzyme-substrate complex (E-l-Ser) before releasing water.
Further analysis of the activity of CBS was performed to evaluate if TgCBS also synthetizes l-Cth via the β-replacement reaction of l-OAS and l-Hcys (reaction 2 in Fig. 1a). Interestingly, TgCBS can also act on l-OAS, even if the catalytic efficiency was ~ threefold lower compared to l-Ser, as it is affected by higher Km values. Substrate inhibition was also observed for l-OAS-dependent CBS activity at high concentrations of l-Hcys (Ki = 1.4 ± 0.2 mM) (Table 1).
The ability of TgCBS to use both l-Ser and l-OAS to produce l-Cth was further supported by analysis of reaction products using reverse phase HPLC in combination with ortho-phthaldialdehyde (OPA) derivatization (Fig. 3c,d). The retention times of l-Ser, l-OAS, and l-Cth commercial standards were 6.4 min, 6.5 min and 23.9 min, respectively (Fig. 3c,d). Importantly, HPLC detected only l-Cth upon incubation of TgCBS with l-Hcys and either l-Ser (Fig. 3c) or l-OAS (Fig. 3d).
Since human CBS is allosterically activated by SAM, we also investigated the enzymatic activity of TgCBS in the presence of SAM. However, no significant response to SAM was observed in the 0–0.5 mM SAM concentration range (Supplementary Table S1).
Alternative CBS ability to generate H
The enzymatic ability of TgCBS to produce H2S in the presence of l-Cys alone or with l-Hcys was also analyzed. The generation of H2S catalyzed by TgCBS was monitored by the lead acetate method7.
Experimental data for H2S production from l-Cys followed a markedly biphasic profile (Fig. 4a) which is consistent with the fact that the production of H2S from l-Cys arises from both unimolecular (β-elimination of l-Cys to generate l-Ser, reaction 3, Fig. 1a) and bimolecular (β-replacement of 2 mol of l-Cys to generate lanthionine, reaction 4 Fig. 1a) reactions. This allowed us to deconvolute the kinetic parameters for reactions 3 and 4, respectively (Table 2) by using Eq. (2) with vL-ser defined in Eq. (3) and vlanth defined in Eq. (4), following the procedure described by Singh et al.8. The active-site of CBS can accommodate two substrates, i.e. l-Ser and l-Hcys in the canonical reaction, in the so-called site-1, where the external aldimine with l-Ser is formed, and site-2, where l-Hcys is docked for reaction with α-aminoacrylate to generate l-Cth. Thus, L-Cys can bind to both sites. TgCBS showed a ~ sevenfold higher Km for binding of the second mol of l-Cys (34 ± 3 mM) with respect for binding of l-Cys to site 1 (4.7 ± 0.9 mM). The kcat for condensation of two molecules of l-Cys (reaction 4, Fig. 1a) is ~ fivefold higher than for the β-elimination of l-Cys (reaction 3, Fig. 1a) (Table 2).
The ability of TgCBS to catalyze both the β-elimination and the condensation reactions starting from l-Cys was further confirmed via reverse phase HPLC (Fig. 4b). The comparison of the HPLC profiles obtained following incubation of TgCBS in the presence of l-Cys with those of l-Ser and lanthionine commercial standards allowed the identification of both reaction products l-Ser and lanthionine. Interestingly, quantitative analysis of l-Ser and lanthionine production in the presence of increasing concentration of l-Cys confirmed that the production of l-Ser prevails at low concentrations of l-Cys while at higher concentrations of l-Cys the predominant product is lanthionine (Fig. 4c).
The formation of H2S significantly increased when TgCBS catalyzed the condensation reaction of l-Cys and l-Hcys (reaction 5, Fig. 1a) and the kcat value for the reaction was ~ 11-fold higher than that for l-Cys alone (Fig. 4d, Table 2). Importantly, no H2S formation from l-Hcys alone was detected by the lead acetate assay. In accordance with kinetic data, HPLC analysis of reaction products obtained upon incubation of TgCBS with l-Hcys and l-Cys resulted in a main fluorescence peak corresponding to l-Cth and one minor peak ascribable to l-Ser (Fig. 4e). Thus, TgCBS produces H2S preferentially via β-replacement of l-Cys with l-Hcys than β-elimination of l-Cys or β-replacement of two molecules of l-Cys.
Of note, the kcat value obtained for the β-replacement of l-Ser and l-Hcys (6.3 ± 0.4 s−1, reaction 1, Fig. 1a) is ~ fivefold lower than that for the condensation of l-Cys and l-Hcys (29 ± 1 s−1, reaction 5, Fig. 1a). Since l-Ser and l-Cys likely coexist in the cell at physiological conditions, we investigated the β-replacement reaction of l-Cys with l-Hcys in the presence of l-Ser as a competing substrate. Increasing concentrations of l-Ser resulted in a decrease in H2S production, indicating that l-Ser inhibits the condensation of l-Cys with l-Hcys (Fig. 4f). The IC50 value, i.e. the concentration of l-Ser at which the H2S production and therefore the activity of TgCBS was half-maximal, was 6.7 ± 1.3 mM.
Spectroscopic analysis in the presence of substrates, products and analogs
The absence of heme in TgCBS (Fig. 1b) offered the opportunity to spectroscopically investigate the intermediates in reactions catalyzed by TgCBS. Addition of l-Ser or l-OAS to the TgCBS solution resulted in the disappearance of the 410 nm-peak and the appearance of a major band centered at 440 nm together with an increase at 330 nm (Fig. 5a). The 440–460 nm band is usually ascribed to the aminoacrylate species30. However, since the attribution of the band at 440 nm to the aminoacrylate species in the UV-visible spectra of TgCBS may not be straightforward, we analyzed the PLP-dependent changes elicited by l-Ser and l-OAS by CD spectroscopy. TgCBS alone exhibited a pronounced positive CD peak at 410 nm and a modest band at 280 nm. The addition of l-Ser or l-OAS resulted in a negative band at 460 nm and in an increased signal of the positive band at 280 nm (Fig. 5b). These spectra allowed us to assign the peak at 440–460 nm to the aminoacrylate intermediate. Based on these data, the second absorption peak at 320–330 nm can be assigned to a different tautomer, i.e., the enolimine tautomer of the aminoacrylate species.
Addition of the analog l-alanine (l-Ala) to TgCBS resulted in a shift from a positive 410 nm CD band to a modest negative band centered at 430 nm, in accordance with the conversion of internal to external aldimine (Fig. 5b). No changes in both the absorption and CD spectra were observed following addition of l-Hcys to the enzyme. Thus, l-Hcys cannot form an external aldimine with TgCBS (data not shown).
We also measured the CD spectra of TgCBS in the presence of the product l-Cth to evaluate the reversibility of the CBS canonical reaction. The reaction with l-Cth caused the appearance of a pronounced negative peak at 460 nm and a broader positive peak at 400 nm. Moreover, an increase in the 280 nm band was observed (Fig. 5c). These changes are ascribable to the formation of an aminoacrylate intermediate (Fig. 5d), thus indicating partial reversibility of the reaction. This partial reversibility was also evident for the β-replacement reaction of two molecules of l-Cys (Fig. 5c). Binding of lanthionine to the enzyme (at the putative site described in Fig. 6) resulted in CD spectra that were comparable to those seen in the presence of l-Cth and led to the detectable accumulation of aminoacrylate intermediate at 460 nm.