Pre-Steady-State Kinetic Analysis of Enzyme-monitored Turnover during Cystathionine β-Synthase-catalyzed H₂S-Generation

Abstract

Cystathionine β-synthase (CBS) catalyzes the first step in the transsulfuration pathway in mammals, i.e., the condensation of serine and homocysteine to produce cystathionine and water. Recently, we have reported a steady-state kinetic analysis of the three hydrogen sulfide (H₂S)-generating reactions that are catalyzed by human and yeast CBS (Singh et al (2009) J Biol Chem 284: 22457-66). In the current study, we report a pre-steady-state kinetic analysis of intermediates in the H₂S-generating reactions catalyzed by yeast CBS (yCBS). Because yCBS does not have a heme cofactor, in contrast to human CBS, it is easier to observe reaction intermediates with yCBS. The most efficient route for H₂S generation by yCBS is the β-replacement of the cysteine thiol by homocysteine. In this reaction, yCBS first reacts with cysteine to release H₂S and forms an aminoacrylate intermediate (k_obs=1.61 ± 0.04 mM⁻¹ s⁻¹ at low cysteine and 2.8 ± 0.1 mM⁻¹ s⁻¹ at high cysteine concentrations, at 20 °C), which has an absorption maximum at 465 nm. Homocysteine binds to the E•aminoacrylate intermediate with a bimolecular rate constant of 142 mM⁻¹ s⁻¹ and rapidly condenses to form the enzyme-bound external aldimine of cystathionine. The reactions could be partially rate limited by release of the products, cystathionine and H₂S.

Hydrogen sulfide (H₂S)¹, like nitric oxide and carbon monoxide, is a gaseous signaling molecule (1-3) that elicits a variety of physiological effects. In the cardiovascular system, H₂S apparently functions as a vasorelaxant (4) and as a cardioprotective agent (5). A dose-dependent decrease in murine blood pressure by sodium hydrosulfide has been reported (4). Other reported effects of H₂S include protection against ischemia reperfusion injury and anti-inflammatory effects in tissues (5, 6). In lower organisms like yeast, H₂S plays a role in population synchronization during ultradian oscillations (7). There are two known mammalian enzymes that can directly generate H₂S: cystathionase γ-lyase (CGL) and CBS (8). A third enzyme pair, 3-mercaptopyruvate sulfurtransferase together with cysteine aminotransferase, catalyzes the transfer of sulfur to an unknown acceptor and, in the presence of a reductant, can liberate H₂S (3, 9). The role of CGL-dependent H₂S production in the vasculature is controversial with one group reporting development of age-related hypertension in CGL knockout mice (4) and another, a normotensive phenotype (10).

We have recently demonstrated that human CGL catalyzes H₂S production via five alternative reactions that involve chemistry at either the β or the γ carbon of the substrates, cysteine and homocysteine (11). In contrast, CBS generates H₂S via chemistry at the β carbon of the substrate, cysteine (12). The contribution of the mecaptopyruvate sulfurtransferase system to H₂S biogenesis relative to the transsfuluration pathway is not known.

CBS catalyzes the condensation of serine and homocysteine to produce cystathionine in the transsulfuration pathway (Scheme 1, reaction [1]) (13). This reaction provides an avenue for clearing homocysteine in mammals (14, 15). CBS deficiency leads to accumulation of homocysteine and is accompanied by clinical impairments in four major organ systems: the ocular, cardiovascular, central nervous, and skeletal systems (16). CBS also catalyzes the condensation of cysteine and homocysteine (reaction [2]) or of two molecules of cysteine (reaction [3]) to produce cystathionine or lanthionine respectively (12). In both these cysteine-dependent reactions, H₂S is released as a byproduct. Alternatively, CBS can catalyze the β-elimination of H₂S from cysteine and the rehydration of the resulting aminoacrylate to form serine (reaction [4]). The k_cat for H₂S generation by yCBS (55 sec⁻ at 37 °C) via the condensation of cysteine and homocysteine (reaction [2]) is 18-fold faster than that for the β-elimination and rehydration to form serine (3 sec⁻¹). The condensation of two moles of cysteine to form lanthionine occurs at an intermediate rate (8 sec⁻¹); however, the K_M for the second mole of cysteine is much larger (33 mM) than that for homocysteine (0.13 mM) so that formation of lanthionine is not expected to compete well with formation of cystathionine at physiological concentrations of substrates. The catalytic turnover numbers for cystathionine formation via reaction [2] by hCBS and yCBS are comparable (55 and 20 s⁻¹, respectively), but the β-elimination of H₂S from cysteine and the subsequent addition of H₂O to form serine with hCBS occur at only 0.4 sec⁻¹ (i.e., 50-fold slower than for formation of cystathionine). As with yCBS, the K_M for the second mole of cysteine (27 mM) is also high for hCBS compared to the affinity of that site for homocysteine (3 mM).

Scheme 1.

CBS belongs to the fold II family of PLP enzymes (17). Other members of this family include O-acetylserine sulfhydrylase, threonine deaminase, and tryptophan synthase (18). The presence of PLP and the absence of other chromophores in these enzymes have facilitated extensive kinetic analyses of their reaction intermediates (19, 20). However, the UV-visible absorption spectrum of hCBS is dominated by the highly absorbing heme chromophore, which compromises direct observation of PLP-bound intermediates during the catalytic cycle (21); thus, only limited kinetic analysis of an unstable variant lacking the N-terminal heme domain has been reported (22). Hence, the homologous yCBS, which lacks the heme, but catalyzes the same PLP-dependent chemical reactions, has been used as a model of the human enzyme to elucidate the kinetic mechanism of the canonical β-replacement reaction (Scheme 1, reaction [1]) (23, 24).

In this study, we have analyzed the reaction intermediates formed during biogenesis of H₂S by yCBS via the β-replacement of the cysteine thiol by homocysteine (reaction [2]) or by the β-elimination of the cysteine thiol (reaction [3]). Our results demonstrate that the kinetics for the H₂S-generating reactions differ from that for the canonical reaction [1], primarily in the first step, i.e., binding of serine or cysteine. Once the common aminoacrylate intermediate is formed from either substrate, the kinetics of the subsequent addition reactions with homocysteine are similar. Under multiple turnover conditions, the reactions appear to be partially limited by product release.

Experimental Procedures

Materials

L-cysteine, DL-homocysteine, L-cystathionine, L-serine and PLP were purchased from Sigma. Protease tablets were purchased from Roche.

Purification of yCBS

The purification of full-length yCBS was adapted from Jhee et al. (25) with a slight modification. The pSEC plasmid encoding yCBS was provided by Dr. Edith Miles (NIH). The E. coli strain BL21 (DE3) was freshly transformed with the pSEC plasmid. Cells (18 g) obtained from a 1 L culture were suspended in 50 mL of 10 mM potassium phosphate (KPi) buffer, pH 7.8, containing 1 protease tablet, 100 μM PLP, 10 mM β-mercaptoethanol, 1 mM EDTA, 20 mg lysozyme, and Triton (0.1 % v/v). Cells were disrupted by sonication using a power output of 7 for nine cycles of 30 sec pulses and 3 min breaks. The supernatant obtained after centrifugation at 12,000 × g, was loaded on to a 16 × 4 cm Q-sepharose column pre-equilibrated with Buffer A (10 mM KPi, pH 7.8) and washed with Buffer A containing 10 mM NaCl. The fractions were eluted with an 800 mL gradient from 0.01 to 0.5 M NaCl in Buffer A and CBS-containing fractions were pooled, concentrated, and dialyzed overnight against Buffer A. The protein was then loaded onto a 16 × 4 cm hydroxylapatite column pre-equilibrated with Buffer A and washed with the same buffer. Protein was eluted with a 600 mL gradient from 0.01 to 0.5 M KPi, pH 7.8. Fractions of interest were pooled, concentrated and then dialyzed against 100 mM HEPES, pH 7.4 and stored at −80°C. From 1 L of culture, ~300 mg of protein was obtained and was judged to be >95% pure by SDS-PAGE analysis. All the purification steps were performed at 4 °C.

CBS Activity Assays

CBS activity was measured either in the radiolabeled assay (for generation of cystathionine) or a colorimetric assay (for generation of H₂S) as described previously (12).

Rapid Scanning Stopped-flow Spectroscopy

Pre-steady state experiments were performed using an Applied Photophysics stopped-flow spectrophotometer (SX.MV18; Leatherhead, UK) in the photodiode array mode or with a Hi-Tech Scientific, stopped-flow spectrophotometer (Model SF-61DX, Bradford on Avon, UK) in both single wavelength and diode array modes. For diode array assays, a 1.5 ms integration time was used. The temperature was maintained at 20 °C using a circulating water bath. Double mixing experiments were carried out using the Hi-Tech stopped-flow spectrophotometer. All concentrations of reagents listed for the stopped-flow experiments are those prior to mixing. In single-mixing experiments, yCBS (70-145 μM, calculated per 55 kDa monomer) was mixed with various concentrations of substrate in 100 mM HEPES, pH 7.4. For L-cystathionine, the stock solution was made in 100 mM HEPES, pH 7.4, followed by the gradual addition of 5 M NaOH until the solution was clear. Further dilutions of cystathionine were made in 100 mM HEPES, pH 7.4. The substrate and enzyme solutions were filtered through a 0.45 μm filter just prior to their use in the stopped-flow experiment.

In the double-mixing experiments, 120 μM yCBS was initially mixed with 30 mM L-serine or L-cysteine in the first step, and after ~ 15 ms it was mixed with an equal volume of buffer. The reaction was monitored at 465 nm to determine when the formation of the aminoacrylate intermediate was maximal. Based on this information, the delay time was set at 300 msec (with serine) or 500 msec (with cysteine) prior to the second mixing step. After aging the solution for the specified time, the aminoacrylate intermediate was rapidly mixed with various concentrations of DL-homocysteine (0.4–40 mM).

Data from the stopped-flow experiments were fitted using the pro-viewer Software (Applied Photophysics), KinetAsyst (Hi-Tech), Specfit Global Analysis Program (Spectrum Software Associates) or Sigma plot. First-order rate constants (k_obs) for the reactions with serine, cysteine, and cystathionine were determined from fits to single exponential or multiple exponential equations. The k_obs values obtained at various concentrations of substrates were fitted to a linear equation to obtain the apparent second-order rate constants.

The observed rate constants, k_obs, for the reaction of serine and cystathionine with yCBS (described in Scheme 2) to form an aminoacrylate intermediate, shows a hyperbolic dependence on substrate concentration, and the data were fitted using equation 1.

$$k_{\mathit{obsd}}=\frac{k_3\times \left[\mathit{Substrate}\right]}{k_{21}+\left[\mathit{Substrate}\right]}+k_4$$ Equation 1

Here K₂₁ = k₂/k_1, and substrate denotes either serine or cystathionine.

Scheme 2.

Results

Reaction of yCBS with Serine or Cysteine

The UV-visible spectrum of yCBS as isolated exhibits an absorption maximum at 409 nm, which corresponds to the protonated Schiff base (internal aldimine) between Lys53 in the active site and the PLP as described previously (25). Addition of serine results in formation of a PLP adduct of the aminoacrylate species with an absorption maximum at 465 nm and apparent isosbestic points with the initial spectrum of the internal aldimine at 425 and 357 nm (Fig. 1). However, careful inspection of the kinetic traces near 425 nm reveals two kinetic phases suggesting formation of transient species. The dependencies of the rates of formation of the transient and the aminoacrylate species are shown in Fig. 1B. We note that conversion of the internal aldimine to the external aldimine occurs in multiple steps and intermediates are not well resolved either kinetically or spectroscopically in our experiments. Nevertheless, the rate of formation of the aminoacrylate showed a hyperbolic dependence on serine concentration (Fig. 1B). Using equation 1, the maximum rate for aminoacrylate formation and the K_dapp for serine were estimated to be 155 ± 7 s⁻¹ and 14.4 ± 1.7 mM, respectively. At wavelengths near 425 nm two phases could be detected. The linear dependence of the first phase (k_obs) for formation of the apparent transient species on serine is plotted in Fig. 1B, but its interpretation is complex. Hence, the apparent bimolecular rate constant (37.8 ± 3.0 mM⁻¹ s⁻¹) does not simply correspond to the k₁ (i.e. the binding of serine to enzyme). Since there are several kinetically unresolved steps involved in the formation of the aminoacrylate from serine, we cannot unambiguously assign the observed kinetic phases to specific rate constants, i.e. formation of the external aldimine, or obtain a true K_d for serine. If this were a true second order reaction to form bound serine, the plot would have an intercept corresponding to k₋₁ and the ratio of k₋₁/k₁ would give the apparent K_d. However, the intercept is near zero such that the measured ratio would be far less than the apparent K_d. Likewise, the maximum extrapolated rate is not k₂. The maximum rate is limited by the maximum equilibrium concentration of the species (presumably the external aldimine) converting to the aminoacrylate.

Figure 1.

Spectra obtained by rapid-scanning stopped-flow spectrophotometry of the reaction of yCBS with serine in 100 mM HEPES buffer pH 7.4 at 20 °C. (A) yCBS (55 μM) was mixed with 16 mM L-serine, and spectra were recorded for 1.4 sec. The internal aldimine has an absorption maximum at 409 nm, and on addition of serine, it converts to the aminoacrylate intermediate with an absorption maximum at 465 nm. (B) Dependence of the observed first-order rate-constants for the formation of aminoacrylate at 465 nm (open circles, right axis) and appearance of the transient species observed at ~ 425 nm (closed circles, left axis) on the concentration of L-serine. C. Deconvolution of the rapid scan data shown in A. Spectra “a”, “b”, and “c” are assigned to the internal aldimine, transient species with characteristics consistent with it having external aldimine spectral character, and the aminoacrylate respectively. The initially observed spectrum at ~ 1.5 - 2 ms already has a small contribution of the aminoacrylate species, as would be expected in the rapid reactions involving 16 mM serine. At this concentration, Fig. 1B shows that the rate of formation of the aminoacrylate is ~250 s⁻¹.

Global fitting analysis of the spectral data, using the model A → B → C was used as a first-order approximation for the process and yielded the spectra shown in Fig. 1C. We assign spectra “a” and “c” primarily to the internal aldimine and the aminoacrylate, respectively. Spectrum “b” represents the mixture of species that we refer to as the complex transient species. However, spectrum “b” is similar to that previously assigned to the external aldimine for yCBS at pH 8.0 in Tris buffer (23), conditions that better resolved the processes. However, unlike the spectrum obtained for the external aldimine at pH 8.0 (23), the corresponding deconvoluted spectrum “b” in this study is not pure and even shows a small contribution of the 465 nm aminoacrylate species. As noted above, conversion of the internal aldimine (primarily spectrum “a”) to the external aldimine occurs in multiple steps and the intermediate gem-diamine species that is formed during the reaction is not detected. Hence, the determined rate constant in fact is complex and comprises multiple component steps that cannot be deconvoluted in the present study. The spectral changes resulting from the addition of cysteine to yCBS are shown in Fig. 2A. Upon addition of cysteine, the absorption maximum shifts from 409 nm to 465 nm, very similar to the change seen with the addition of serine. The rate of disappearance of the internal aldimine or formation of the aminoacrylate intermediate was not linearly dependent on cysteine concentration (Fig. 2B). At concentrations of cysteine less than 15 mM, a bimolecular rate constant of 1.61 ± 0.04 mM⁻¹ s⁻¹ was estimated and at concentrations of cysteine above 20 mM, a bimolecular rate constant of 2.8 ± 0.1 mM⁻¹ s⁻¹ was estimated. The reaction with cysteine is complex and includes a mixture of reactions (condensation either with a second mole of cysteine or with a water molecule following β-elimination as shown in reactions 3 and 4, respectively). The maximal rate of reaction 3 is significantly greater than that of reaction 4; however, it is limited by the high K_M of CBS for the second mole of cysteine (33 mM). After 1-2 sec of reaction, the absorbance at 465 nm due to the aminoacrylate intermediate decreased, and a new species with an absorbance maximum at 425 nm (Fig. 2C) appeared at 0.24 ± 0.02 s⁻¹, which is too slow to be kinetically relevant for cystathionine formation.

Figure 2.

Spectra obtained by rapid-scanning stopped-flow spectroscopy of the reaction of yCBS with L-cysteine in 100 mM HEPES buffer pH 7.4 at 20 °C. (A) yCBS (70 M) was mixed with 30 mM L-cysteine, and spectra were recorded for 200 msec. The internal aldimine (409 nm), on addition of L-cysteine, converts to an aminoacrylate species, which has an absorption maximum at 465 nm. This spectral change is observed with isosbestic points at 423 and 357 nm. (B) Dependence of the observed first-order rate constant for the formation of aminoacrylate at 465 nm on the concentration of L-cysteine. Inset. The kinetics of absorbance changes at 465 nm at various concentrations of cysteine. (C) yCBS (25 μM) was mixed with 10 mM L-cysteine and spectra were monitored till the 425 nm species was fully formed (solid line). Addition of 10 mM L-homocysteine led to immediate formation of a 410 nm absorbing species (broken line).

Reaction of the aminoacrylate intermediate with homocysteine

Either serine or cysteine can bind to the PLP site and form the aminoacrylate intermediate with the elimination of H₂O or H₂S, respectively. To determine the rate constant for the reaction of the aminoacrylate species with homocysteine, double-mixing experiments were performed. yCBS (120 μM) was mixed with 30 mM serine and aged for 300 ms to pre-form the aminoacrylate intermediate before mixing with 0.5-6 mM L-homocysteine (Fig. 3-C). Immediately upon mixing the aminoacrylate with homocysteine, there was an initial decrease in absorbance at 465 nm (Fig. 3A); this was followed by steady-state turnover, and when homocysteine was completely consumed due to the presence of excess serine, accumulation of the aminoacrylate, as indicated by the final increase in absorbance at 465 nm, was observed (Fig. 3B-C). The duration of the steady-state phase varied with the concentration of homocysteine (Fig. 3B) and corresponded well with the expected duration based on the k_cat (~8 sec⁻¹ and ~18 sec⁻¹ for reactions [1] and [2], respectively) for this reaction. The kinetic course of the reaction in which homocysteine is mixed with the cysteine-derived aminoacrylate is shown in Fig. 3D. At the end of the reaction, i.e., when homocysteine is depleted but cysteine is still present, the absorbance at 465 nm decreases due to formation of the 425 nm absorbing species as shown in Fig. 2C.

Figure 3.

Reaction of serine- or cysteine-derived aminoacrylate with homocysteine. (A) yCBS (120 μM) was premixed with 30 mM of L-serine to form the aminoacrylate, and after 300 ms was mixed with 6 mM L-homocysteine. The spectra are shown for the first 10 msec. (B) Spectral changes for the reaction in (A) monitored for 5 sec. This shows the return of the aminoacrylate due to the presence of excess serine. Inset. Dependence of the observed first-order rate-constant for the decrease in absorbance at 465 nm on the concentration of L-homocysteine in the double mixing experiment. The open and closed circles represent data from experiments in which yCBS was premixed with serine and cysteine respectively. (C) Absorbance changes at 465 nm observed with selected homocysteine concentrations (0.25-3 mM after mixing) added to the aminoacrylate derived from serine (as shown in A). (D) Reaction of the cysteine-derived aminoacrylate with homocysteine. yCBS (120 μM) was premixed with 12 mM L-cysteine to form the aminoacrylate, and after 700 ms of aging time, was mixed with varying concentrations of L-homocysteine (0.25-2 mM). Absorbance changes at 465 nm are shown in the figure.

The disappearance of the aminoacrylate species was monitored at 465 nm (Fig. 3C), and the observed rate constants for this first phase of the reaction were linearly dependent on the concentration of homocysteine (Fig. 3B, inset, open circles). The bimolecular rate constant was estimated to be 183 ± 4 mM⁻¹s⁻¹ and the apparent k_off was 67 ± 4 s⁻¹ (y-intercept), yielding a K_d of 0.37 ± 0.02 mM. A plot of the amplitude of the absorbance decrease at 465 nm during approach to steady state versus the concentration of homocysteine yielded a similar K_dapp value of 0.32 ± 0.02 mM.

When homocysteine was added to the aminoacrylate intermediate generated from cysteine (Fig. 3B, closed circles), a comparable value for the bimolecular rate constant was obtained (142 ± 5 mM⁻¹s⁻¹) and the apparent k_off was estimated to be 21 s⁻¹. Thus, the K_dapp for homocysteine (0.15 ± 0.04 mM) although lower, was similar to the value estimated from the dependence of the change in amplitude at A_465nm during approach to steady-state versus the homocysteine concentration (0.10 ± 0.01 mM).

Reaction of yCBS with cystathionine

yCBS reacts with cystathionine to form an aminoacrylate species with an absorption maximum at 465 nm (Fig. 4). The internal aldimine converts to the aminoacrylate with isosbestic points at 358 and 425 nm, without detectable accumulation of the external aldimine of cystathionine, confirming our previous results (23). The rate of aminoacrylate formation, monitored either by the decrease in absorbance at 409 nm or the increase at 465 nm, was linearly dependent on the concentration of cystathionine. A apparent bimolecular rate constant of 4.1 ± 0.2 mM⁻¹s⁻¹ for the formation of the aminoacrylate from cystathionine was obtained from the linear portion of the curve (Fig. 4B).

Figure 4.

Rapid-scanning stopped-flow spectra of the reaction of yCBS with L-cystathionine (Cth) in 100 mM HEPES buffer pH 7.4 at 20 °C. (A) yCBS (70 μM) was mixed with 20 mM L-cystathionine, and spectra were collected for 100 ms. Upon addition of L-cystathionine, the internal aldimine (409 nm), converts to an aminoacrylate species (465 nm) with isosbestic points at 423 and 357 nm. (B) Plot of the observed first-order rate constant for the formation of the aminoacrylate intermediate on the concentration of L-cystathionine.

Discussion

The mechanisms by which H₂S elicits its biological effects are poorly understood, but even less is known about how its biosynthesis is enabled or disabled. There is significant pharmacological interest in developing inhibitors of H₂S production as well as H₂S-releasing drugs for treating reperfusion injury, inflammation, and circulatory shock (26). Detailed kinetic characterization of the H₂S-generating reactions combined with structural insights into full-length CBS, are needed for guiding rational drug design aimed at inhibiting or activating H₂S production by this enzyme. We have recently reported a steady-state characterization of the H₂S-generating reactions catalyzed by the transsulfuration pathway enzymes, CBS and CGL (11, 12). The promiscuous behavior of both enzymes results in a multitude of routes for H₂S production in reactions that depend on the concentrations of the common amino acids, cysteine and homocysteine. In this study, we present a kinetic characterization of intermediates formed during H₂S biogenesis by yCBS. Since the presence of the heme cofactor in hCBS obscures direct observation of the PLP-bound intermediates, we have employed yCBS, which exhibits robust H₂S producing capacity, as a model for the human enzyme.

The three H₂S-yielding reactions catalyzed by yCBS are described in Scheme 1 and the V/KMCys values for reactions [2], [3] and [4] are 15.3 × 10³, 2.3 × 10³ and 0.9 × 10³ M⁻¹ s⁻¹, respectively. Previously, we reported pre-steady-state kinetic analyses of the canonical yCBS-catalyzed reaction, i.e., the condensation of serine and homocysteine to give cystathionine (23). We note that in the earlier work, reactions were studied in 200 mM Tris buffer, pH 8.0, at 15 °C. Amine buffers can form Schiff bases with PLP and therefore are not ideal for studying PLP enzymes. In the present study, 100 mM HEPES, pH 7.4 was used, and the temperature was set at 20 °C. These changes resulted in some differences in the kinetic parameters, which are noted below.

When serine is rapidly mixed with yCBS, the internal aldimine converts to the transient species with an apparent second-order rate constant of 38 ± 3 mM⁻¹ s⁻¹; this value is ~2.7-fold greater than the value obtained at pH 8.0 and 15 °C (23). The maximal rate for formation of the aminoacrylate intermediate was calculated using equation 1 and yielded a value of 155 ± 6 s⁻¹, which is ~10-fold greater than the value reported previously at pH 8 and 15 °C (15 ± 0.4 s⁻¹) (23) (Scheme 3,A).

Scheme 3.

Kinetic mechanism for reactions catalyzed by yCBS

When cysteine is rapidly mixed with yCBS, the internal aldimine shows an isosbestic conversion to the aminoacrylate intermediate with an apparent second-order rate constants of 1.61 ± 0.04 mM⁻¹ s⁻¹ and 2.8 ±0.1 mM⁻¹ s⁻¹, corresponding to reactions 4 and 3, respectively (Scheme 3B,C). Since an external aldimine did not accumulate at detectable levels in this reaction, we are unable to deconvolute the kinetic parameters for the two steps, i.e., formation of the external aldimine followed by formation of the aminoacrylate. Our data show that the rate of aminoacrylate formation does not depend linearly on the concentration of cysteine (Fig. 2). Rather, it appears to be somewhat cooperative. The slope of the curve at lower concentrations of cysteine is ~1.5–2 fold smaller than that at higher concentrations of cysteine. We speculate that binding of the second mole of cysteine (possibly at the site that binds homocysteine in the canonical reaction [1]), may cause a conformational change or otherwise promote an increase in the rate of aminoacrylate formation. Thus, the biphasic pattern observed during formation of the aminoacrylate species could correspond to a smaller k_obs at low cysteine concentrations when only one mole of cysteine is bound at site 1 (the PLP site) and a larger k_obs at higher cysteine concentrations when site 2 (occupied by homocysteine in the canonical reaction) is also populated with cysteine. This interpretation is consistent with the steady-state kinetics of H₂S formation, which are clearly biphasic as a function of cysteine concentration (12). At low cysteine concentrations, the aminoacrylate intermediate reacts with water to form serine, whereas at high cysteine concentrations, the aminoacrylate intermediate reacts with cysteine bound at site 2.

A few seconds after the reaction with higher concentrations of cysteine (>6 mM) has proceeded, the aminoacrylate converts to a new species with an absorbance maximum at 425 nm (Fig. 2C). The identity of this species is currently not known. However, it is not a dead-end species, because addition of homocysteine to the species with the 425 nm absorbance maximum leads to rapid formation of an aldimine with an absorbance maximum at 409 nm (Fig. 2C). From steady-state experiments in which lanthionine formation was monitored, the K_M for cysteine for the second binding site was estimated to be 33 ± 3 mM (12). We note that hCBS binds cysteine significantly weaker (K_d= 400 ± 27 μM) than serine (K_d= 56 ± 6 μM) (12).

In principle, the aminoacrylate intermediate of yCBS should be the same whether it is formed from serine or cysteine. Hence, the second half reaction, i.e., the nucleophilic attack of homocysteine on the PLP-bound aminoacrylate species, is also expected to be very similar. Indeed, we found the bimolecular rate constants for the addition of homocysteine to the aminoacrylate intermediate derived from serine or cysteine to be 183 ± 4 mM⁻¹s⁻¹ and 142 ± 5 mM⁻¹s⁻¹ respectively (Fig. 3B inset). The small difference in these values could be related to the differential release of H₂O versus H₂S in the two reactions that produce the aminoacrylate or to binding of cysteine to a second and as yet uncharacterized site. Surprisingly, the k_off observed in the case of serine plus homocysteine is ~3-fold larger than for cysteine plus homocysteine (y-intercepts in Fig. 3B inset). The basis for this difference is presently not understood. In the case of the reaction involving serine plus homocysteine, the K_d for homocysteine was determined to be 0.37 ± 0.02 mM, whereas for the reaction involving cysteine plus homocysteine, the K_d for homocysteine was determined to be 0.15 ± 0.03 mM. This value is similar to the K_M (0.13 ± 0.02 mM) for homocysteine determined for the reaction involving cysteine plus homocysteine.

The bimolecular rate constant for the reaction of cystathionine with yCBS to form the aminoacrylate (4.1 ± 0.2 mM⁻¹s⁻¹ at 20 °C) is ~3-fold larger than the value determined previously using Tris buffer at pH 8.0 and 15 °C (1.5 mM⁻¹s⁻¹), and the K_dapp for cystathionine is ~3-fold smaller than that previously determined (1.6 ± 0.3 mM) (18) (Scheme 3D). However, since we are unable to detect an external aldimine, the k_obs represents a composite of the binding and chemical steps. We note that the k_cat values for reactions [1] (serine + homocysteine) and [2] (cysteine + homocysteine) are ~8 s⁻¹ and ~18 s⁻¹ respectively at 20 °C, which match well the estimates obtained from the duration of the steady-state phase in the rapid reaction kinetic data (Figs 3C,D). Since cystathionine is a common product in these two reactions, its release rate could not be completely rate determining, since it would not account for the difference in the k_cat values for the two reactions. Furthermore, differences in the leaving group potential of water from serine versus the H₂S from cysteine are also unlikely to account for the ~2-fold difference in the k_cat between reactions [1] and [2], since formation of the aminoacrylate from serine is rapid (155 s⁻¹), i.e., ~20-fold higher than k_cat. Instead, under steady-state turnover conditions, differences in the rate of release of water (reaction [1]) versus H₂S (reaction [2]) and/or a kinetically invisible conformational change is likely to contribute to the ~2-fold difference in the turnover numbers for these reactions.