Biosynthesis of 6-thioguanine: characterizing two intermediates involved in its thioamide formation reaction

Abstract

6-Thioguanosine 5’-monophosphate (6-TGMP) is the biologically active nucleotide form of the antimetabolite 6-thioguanine produced by Erwinia amylovorans. The YcfA-YcfC enzymatic pair catalyzes the incorporation of the thioamide functional group in GMP via an oxygen-to-sulfur replacement process. YcfA is a member of the adenine nucleotide alpha hydrolase-like (AANH-like) superfamily and YcfC is a pyridoxal 5′-phosphate (PLP)-dependent enzyme. Two mechanistic models exist, depending on whether YcfC is a C-S lyase or a cysteine desulfurase, to explain the trans-sulfuration reaction. Herein, we investigate the YcfA-YcfC trans-sulfuration reaction using anaerobically purified YfcA and YcfC enzymes in the absence of reductants. Steady state and pre-steady state kinetic studies suggest the mechanism involves two distinct intermediates and differs from the common trans-sulfuration pathway used in, for example, tRNA thionucleotide biosynthesis.

Graphical Abstract

INTRODUCTION

Sulfur is an indispensable element in living organisms and sulfur-containing natural products play a central role in biological functions. Thioamides, amides in which the oxygen is replaced by sulfur, exhibit several unique properties. Thioamides are more reactive, more acidic, and more conformationally restrained than amides, in part due to the larger size and lower electronegativity of sulfur.¹ Nature leverages the unique properties of thioamides and several examples of thioamidation in natural product biosynthesis are known.^1–2 Most characterized natural thioamides are of bacterial origin. Examples include methanobactin, which is a ribosomally synthesized and posttranslationally modified peptide (RiPP),³ the nonribosomal natural product closthioamide,^4–5 thioamide-containing nucleotides in transfer RNA (tRNA),^6–7 and thioglycine.⁸ Another example is Cycasthioamide, a non-protein amino acid found in plant seeds.⁹

6-Thioguanine (6-TG) is a naturally occurring thionucleoside produced by Erwinia amylovorans, which is the causal agent of fire blight, a devastating bacterial plant disease affecting a wide range of host species within Rosaceae. The disease can be devastating to the commercial apple and pear industries.^10–11 As an antimetabolite, 6-TG also has a long history of use in patients for the treatment of psoriasis, leukemia, and inflammatory bowel diseases.^{1, 10, 12} Despite its long-standing clinical relevance, the mechanism underlying sulfur incorporation into the guanine scaffold of 6-TG is only recently unfolding. Identification of the 6-thioguanine monophosphate (6-TGMP, 2, Figure 1A) biosynthetic gene cluster reveals two proteins, YcfA and YcfC, essential for 6-TG biosynthesis (Figure 1A).¹⁰ Initial biochemical characterization established that the combined action of YcfA and YcfC converts GMP to 6-TGMP, but the molecular details of sulfur transfer and nucleotide activation remain unresolved.¹⁰

Figure 1.

Transsulfuration reactions in the biosynthesizing 6-TGMP and related systems (e.g., ThiI catalysis). (A), 6-TGMP biosynthesis by YcfA-YcfC catalysis; (B) 4-thiouridine synthesis in tRNA; (C) A YcfA-YcfC mechanistic model in which YcfC functions as a cysteine desulfurase; (D) A YcfA-YcfC mechanistic model in which YcfC functions as a C-S lyase. Discriminating between the models requires determining whether any reductants are needed for YcfA-YcfC catalysis, whether GMP activation by ATP is catalyzed by YcfA alone or needs the YcfA-YcfC complex, and if GMP activation and coupling with a sulfur nucleophile is a stepwise or a concerted process.

Based on sequence homology, YcfA is annotated as a member of the YcaO superfamily, which includes ATP-dependent amide-activating enzymes involved in tRNA thiolation, such as ThiI, MnmA, and TtcA (Figure 1B).^{6, 12–15} This annotation led to the proposal of a ThiI-type transsulfuration mechanism, in which YcfA activates GMP via ATP-dependent phosphorylation or adenylation, followed by sulfur transfer from a YcfC-generated persulfide intermediate (Figure 1C).^{10, 12} In this model, YcfC functions as a cysteine desulfurase, analogous to canonical tRNA thiolation systems.^{10, 12} The recent isolation of a cysteinyl-GMP adduct provides support for an alternative mechanism in which YcfC functions as a PLP-dependent C-S lyase, eliminating the requirement for a protein-bound persulfide intermediate (Figure 1D).^16–17 Complementary structural studies revealed important context/information regarding the organization of the YcfA-containing complex and the conformational features associated with ATP-dependent substrate activation.¹⁷ Together, these studies advocate for a shift from the classical sulfur relay paradigm to a substrate-level sulfur incorporation mechanistic model for 6-TGMP biosynthesis.

However, despite these advances, several fundamental chemical aspects of YcfA-YcfC catalysis remain unresolved/unanswered. In both of the proposed mechanistic frameworks, ATP-dependent activation of GMP is an obligate first step, and YcfA is the implicated enzyme responsible for this transformation due to its membership in the YcaO superfamily.^{6, 15, 18–20} Recent biochemical and structural studies support an adenylation-based activation mechanism.¹⁷ However, the activated nucleotide intermediate remains incompletely defined. It is unclear whether the proposed adenylated GMP intermediate can accumulate or be isolated under controlled reaction conditions. In addition, while a sequential two-step process has been proposed, the extent to which YcfA- and YcfC-catalyzed steps functionally couple at the kinetic or chemical level is unknown. Whether ATP-dependent GMP activation represents a chemically discrete and isolable step, or instead proceeds only in tight coupling with downstream sulfur incorporation, also remains unresolved. It is unclear whether GMP activation by YcfA proceeds independently once initiated, or whether the presence of cysteine modulates the efficiency, stability, or fate of the activated nucleotide intermediate (e.g., Option I versus Option II in Figure 1D). Finally, prior biochemical studies of the YcfA-YcfC system were conducted in the presence of reductants such as dithiothreitol (DTT),^{12, 16} and, thus, it remains unresolved whether a reducing environment is intrinsically required for catalysis or whether exogenous reductants influence intermediate stability or reaction outcome.

To address these outstanding questions, we purified YcfA and YcfC and characterized their catalytic activities under anaerobic conditions in the absence of supplementary reductants. YcfA-YcfC catalysis does not require external reducing agents, in contrast to the canonical ThiI-dependent thiolation reactions in which enzyme-bound disulfide formation necessitates reductive recycling.^{6, 15, 21–22} Reaction kinetics and intermediate trapping experiments reveal that YcfA-YcfC catalysis proceeds through a discrete, multistep process involving at least two chemically distinct intermediates. The first intermediate is a tightly bound GMP-AMP adduct formed by YcfA during ATP-dependent GMP activation. Reaction of this intermediate with cysteine generates a second intermediate, a GMP-Cys adduct, which serves as the direct substrate for YcfC. Isolation of this species is consistent with recent observations by Ishida et al.¹⁶ and supports a mechanism in which YcfA activates and assembles the sulfur-containing nucleotide precursor, while YcfC functions as a PLP-dependent C-S lyase to complete 6-TGMP biosynthesis (Figure 1D).

Collectively, our findings provide direct chemical evidence for the nature and sequence of nucleotide-bound intermediates in YcfA-YcfC catalyzed thioamide biosynthesis, and clarify how ATP-dependent activation, sulfur incorporation, and C-S bond cleavage are coordinated in this noncanonical trans-sulfuration system.

RESULTS AND DISCUSSION

In vitro reconstitution of 6-TGMP biosynthesis

The YcfA and YcfC enzymes are not stable upon purification, and, thus, it was necessary to identify another approach to isolate them.^{12, 16} To avoid the complications associated with reductants, we purified the enzymes under anaerobic conditions, following a method we recently demonstrated for EanB in ergothioneine biosynthesis, which catalyzes a trans-sulfuration reaction.^23–24 More importantly, we then conducted the EanB reaction under single turnover conditions to investigate redox state changes in EanB during catalysis. The insights gained from the EanB studies provided a methodological framework that we subsequently applied to the characterization of the YcfA-YcfC catalytic system. We overexpressed codon-optimized YcfA and YcfC enzymes in E. coli. The YcfA protein included an N-terminal Strep-tag, while the YcfC protein included an N-terminal His₆-tag. We anaerobically purified the YcfA and YcfC proteins using Strep-Tactin resin or Ni-NTA resin, respectively (Figure S1).^23–26

We examined the formation of 6-TGMP from YcfA-YcfC catalysis under anaerobic conditions using either single turnover or multiple turnover conditions. The reaction contained both the YcfA and YcfC enzymes, and the substrates: [β-¹³C]-cysteine, [u-¹³C-¹⁵N]-GMP, and ATP. We separated the resulting products by HPLC (Figure 2B, trace iii). We collected the compound with a retention time t_R of 13.5 min and identified it to be 6-TGMP using one dimensional (¹H, ¹³C, and ³¹P) and two-dimensional (¹H, ¹³C-HMQC) NMR spectroscopy (Figures 2c, 2D and S2). The ¹³C NMR spectrum reveals that the signals of the aromatic carbons of the product (trace ii, Figure 2C) are clearly shifted relative to the signals of GMP (trace i, Figure 2C), consistent with the formation of a thioamide moiety in this transformation.

Figure 2.

In vitro reconstitution of the thioamidation reaction in 6-TGMP biosynthesis. (A) YcfA-YcfC catalyzed thioamidation reaction; (B) HPLC profiles of the YcfA-YcfC reaction: (i) HPLC profile of AMP, GMP and ADP monitored at 260 nm; (ii - iii) HPLC profiles of the YcfA-YcfC catalyzed 6-TGMP production monitored at 260 nm (ii) and 340 nm (iii). The GMP signal decreases as 6-TGMP is produced; (C) ¹³C-NMR analysis of the substrate [u-¹³C-¹⁵N]-GMP (i) and the purified [u-¹³C-¹⁵N]-6-TGMP product (ii) from the YcfA-YcfC catalyzed thioamidation reaction; (D) ¹³C-NMR analysis of YcfA-YcfC reaction using [β-¹³C]-cysteine: (i) YcfA-YcfC anaerobic reaction containing [β-¹³C]-cysteine, GMP and ATP; (ii) ¹³C-NMR of the [β-¹³C]-cysteine substrate; (iii) ¹³C-NMR of the [β-¹³C]-cystine; (iv) ¹³C-NMR of pyruvate; (v) ¹³C-NMR of alanine.

One of the proposed mechanistic models entails a cysteine residue from the enzyme facilitating product release, either through formation of a cysteine persulfide or a cysteine disulfide intermediate (Figure 1C). If this mechanism is correct, an external reductant is required to regenerate the active enzyme and allow multiple turnovers. However, under anaerobic conditions, the YcfA-YcfC reaction proceeds through approximately 50 turnovers without the addition of external reductants. This observation suggests two possibilities: either cysteine itself serves as the reductant, or the catalytic cycle does not require a reductant. To address this question, we analyzed the cysteine turnover products from the YcfA-YcfC multiple turnover reaction under anaerobic condition. For these experiments we included a limiting amount of [β-¹³C]-cysteine in the reaction with excess GMP, and ATP as substrate. The use of [β-¹³C]-cysteine enables direct monitoring of the reaction by ¹³C NMR. If cysteine reduces the YcfA-YcfC catalytic system, a 1:1 correlation will exist between the production of cystine and the product, 6-TGMP. This correlation is not present between cystine and 6-TGMP formation (Figure 2D). After completion of the YcfA-YcfC anaerobic reaction, there is no detectable amount of cysteine (24.6 ppm) nor cystine (36.8 ppm) (trace i, Figure 2D). Instead, the NMR results suggest that pyruvate is generated from cysteine. The dominant ¹³C signal at 27.9 ppm is consistent with the β-carbon of [β-¹³C]-pyruvate (trace iv, Figure 2D) This result does not support the mechanistic model proposed based on ThiI-catalysis (Figure 1C). Cysteine is not a reductant in YcfA-YcfC catalysis and the YcfA-YcfC catalyzed sulfur mobilizing reaction does not require an extra reductant. The findings from the anaerobic studies of YcfA-YcfC indicate that the trans-sulfuration reaction only requires cysteine, GMP and ATP.

Kinetic analysis of YcfA-YcfC reaction indicating the presence of two intermediates.

During the course of the YcfA-YcfC reaction, we observed an absorption increase at 340 nm, and it belongs to 6-TGMP. Analysis of pure 6-thioguanine (Sigma) reveals that the ε_340nm for 6-thioguanine is 24,800 cm⁻¹M⁻¹. Because the absorption band at 340 nm is well separated from the UV absorption maxima for both GMP and ATP, such a feature enables us to perform the kinetic assays of the YcfA-YcfC reaction. Under anaerobic multiple turnover conditions, the kinetic parameters for YcfA-YcfC catalysis are: k_cat of 1.09 ± 0.06 min⁻¹, K_M of 0.256 ± 0.019 mM for cysteine, K_M of 0.221 ± 0.015 mM for GMP, and K_M of 0.259 ± 0.020 mM for ATP (Figure S3).

The reaction required supplementation with pyridoxal 5-phosphate (PLP). After size-exclusion chromatography in Tris buffer, YcfC was colorless, indicating loss of the PLP cofactor and formation of the apo enzyme (Figure S4). The YcfC-PLP complex could be reconstituted by incubating purified YcfC with PLP in potassium phosphate (KPi) buffer. Upon binding to YcfC, the PLP absorption band shifted from 390 nm to 420 nm, consistent with the formation of an internal aldimine between PLP and an active-site lysine residue (Figure S5). For the kinetic analysis, we added one equivalent of PLP to the YcfC enzyme concentration in KPi buffer.

In our steady state characterization of YcfA-YcfC catalysis, we monitored both the kinetics via absorption changes at 340 nm and the UV-visible spectral changes over time in the 220 nm − 800 nm range (Figure 3A & 3B). At early time points (e.g., 1 - 5 min. in Figure 3A), an increase in absorbance at the 300 - 320 nm region was observed. Overtime, it shifts to ~340 nm, and the ratio between the absorptions at 310 nm and 340 nm changes. At the end of the reaction, the spectrum changes to a state similar to that of pure 6-TGMP 2, which possesses an absorption feature centered at ~340 nm. Similar dual-wavelength spectral features have been observed in the YcfA-YcfC system, indicating an intrinsic property of the catalytic reaction.¹⁷

Figure 3.

Evidence for the presence of two intermediates in YcfA-YcfC catalyzed sulfur mobilization reaction in 6-TGMP biosynthesis. (A) Time-course of YcfA-YcfC catalysis at a YcfA and YcfC ratio of 25:1. On the left panel is the UV-vis absorption spectra of the reaction mixture collected at various time points (0, 1, 3, 5 and 10 min.). On the right panel is the HPLC profile for the samples collected at: trace i, 0 min; trace ii, 5 min; trace iii, 10 min; trace iv, 20 min; trace v, 30 min and trace vi, 60 min., respectively; (B) Time-course of YcfA-YcfC catalysis at a YcfA and YcfC ratio of 5:1. On the left panel is the UV-vis absorption spectra of the reaction mixture collected at various time points (0, 1, 3, 5 and 10 min.). On the right panel is the HPLC profile for the samples collected at the same time point as the left panel; (C) Evidence for YcfA-catalyzed intermediate II formation. On the left panel is the UV-visible spectra of YcfA reaction in the presence of cysteine, ATP and GMP. On the right panel is the HPLC trace for samples collected at the same time points as the UV-visible spectra on the left panel; (D) A combination of pre-steady state and steady state analysis of YcfA-catalysis under various conditions. The left panel shows the kinetic traces of 50 μM YcfA with different combination of substrates and the reaction was monitored at 310 nm from 5 ms up to 10 s. On the right panel, we expanded the pre-steady state region (5 ms to 600 ms region) of YcfA preincubated with ATP/GMP and mixed with Cys at multiple YcfA concentrations.

Building on this observation, we systematically varied reaction conditions, including enzyme stoichiometry and reaction time, to further interrogate the origin of the ~310 nm species and to characterize the relationship between these evolving spectral features and the discrete chemical intermediates. By changing the ratio of YcfA and YcfC, the kinetics of the 310 nm species relative to the one at 340 nm also varies (Figure 3B). Along with kinetic studies by monitoring the UV-visible spectral changes over time, we also quenched the reaction by chloroform and then analyzed the quenched reaction mixtures by HPLC. Consistent with the results in UV-visible spectra, the HPLC profile (right panels in Figures 3A & 3B) reveals that besides the product (t_R of 13.5 min.), there is another compound, which we label as intermediate II (t_R of 21.3 min.). In addition, the UV-visible spectrum of intermediate II exhibits a UV-visible absorption feature centered at 312 nm (Figure S6). Also intermediate II forms before 6-TGMP 2 and along with the disappearance of intermediate II, 6-TGMP 2 starts to accumulate.

As there are two enzymes (YcfA and YcfC) in this reaction (Figures 3A & 3B), we next experimentally delineated the enzymatic origin of intermediate II formation and defined the substrate requirements for this step. Specifically, we asked two questions: (1) whether YcfA, YcfC, or a combination of both enzymes generates intermediate II, and (2) which substrates (cysteine, ATP, and/or GMP) are required for its formation. Thus, we examined a series of reactions containing different combinations of enzymes and substrates (Figure 3C & Figure S7). These experiments demonstrate that YcfA exclusively catalyzes intermediate II formation and, moreover, YcfA requires the simultaneous presence of cysteine, ATP, and GMP.

Having established that intermediate II formation is a YcfA-catalyzed process, we then conducted kinetic analysis using SX20 stopped-flow to characterize its formation under pre-steady state conditions. We monitored the reaction kinetics at 310 nm over a time range from 5 ms to 10 s, and representative traces are shown in the left panel of Figure 3D. The result reveals several key features. First, when any one of the three substrates (cysteine, ATP, or GMP) are omitted, little to no accumulation occurs of the 310 nm species (black and green traces in Figure 3D), indicating that formation of intermediate II requires all three components. Second, upon rapidly mixing a purified YcfA with a solution containing cysteine, ATP, and GMP, the absorbance at 310 nm gradually increases (red trace in Figure 3D), consistent with accumulation of intermediate II.

To probe whether intermediate II formation proceeds through an earlier enzyme-bound species, we performed sequential mixing experiments designed to temporally separate nucleotide activation from sulfur incorporation. In these experiments, we pre-incubated YcfA with GMP and ATP prior to rapid mixing with cysteine. Under these conditions, a markedly different kinetic profile is present (blue trace in Figure 3D). Specifically, a rapid increase in absorbance at 310 nm occurs within the first ~600 ms, followed by a slower phase that closely resembles the kinetics observed upon direct mixing of YcfA with all three substrates. The appearance of this initial rapid phase is characteristic of pre-steady state burst kinetics and is indicative of the formation of a tightly enzyme-bound intermediate preceding intermediate II. We designated this upstream species intermediate I.

To provide additional evidence that supports the presence of another intermediate in YcfA-catalysis, we repeated the studies reported in blue trace in the left panel of Figure 3D under different YcfA concentrations (right panel of Figure 3D) and different YcfA-GMP-ATP pre-incubation times (Figure S8). In these studies, the magnitude of pre-steady state phase increases with higher concentrations of YcfA, but are unaffected by different pre-incubation times, indicating intermediate I binds to YcfA rather than being released into the solution.

In summary, a combination of steady-state and pre-steady state kinetic analyses of YcfA-YcfC catalysis revealed the presence of two discrete intermediates, designated intermediate I and intermediate II. YcfA exclusively catalyzes the formation of both intermediates; however, the two species exhibit distinct kinetic and functional properties. YcfA generates Intermediate I upon incubation with ATP and GMP and displays a pronounced pre-steady state kinetic signature, indicating that it is formed as a tightly enzyme-bound species and which is not released into solution. Introduction of cysteine to the pre-assembled YcfA-GMP-ATP complex rapidly converts the enzyme-bound intermediate into intermediate II. The stopped-flow kinetic behavior further indicates that formation of intermediate I constitutes the rate-limiting step of YcfA catalysis, accounting for the rapid burst phase observed within the first ~600 ms when YcfA is pre-incubated with ATP and GMP. In contrast, intermediate II accumulates during YcfA-YcfC turnover, and its accumulation level is sensitive to the relative stoichiometry of YcfA and YcfC. Together, these observations suggest intermediate II is the YcfA reaction product and serves as the physiological substrate for YcfC, ultimately leading to formation of 6-thioguanine monophosphate (6-TGMP, 2) as the final product of YcfA-YcfC catalysis.

Structural characterization of the two intermediates.

6-TGMP 2 elutes at a retention time, t_R, of 13.5 min on the HPLC trace of the YcfA-YcfC reaction with a TE-Click CYS column (Acchrom). By adjusting the ratio between the YcfA and YcfC concentrations and quenching the reaction at the proper time, we increased the production of intermediate II to a higher level (a retention time t_R of 21.3 min (Figure 3C)). Upon discovering intermediate II (compound 12) is the product of YcfA catalysis, we isolated and purified it. We prepared a sufficient quantity of intermediate II for structural characterization from a 100-mL enzymatic reaction containing 35 μM YcfA, 165 μM GMP, 250 μM cysteine and 250 μM ATP. After a one-hour incubation, we lyophilized the reaction mixture and purified the material on a semipreparative HPLC, after removing YcfA by ultrafiltration. We concentrated purified intermediate II (12) by lyophilization prior to structural characterization by NMR and mass spectrometry. During this extended workup procedure, compound 12 spontaneously converts into another compound (13, Figure 4 & Figure S9), which exhibits an absorption maximum at 283 nm and a retention time, t_R, of 16.6 min (Figure S10). Compound 13 is stable, and we fully characterized it using ¹H NMR, ¹³C NMR, 2D NMR, ³¹P NMR, and HRMS (Figure S11).

Figure 4.

Schematic outline of our approaches to trap and characterize the YcfA-YcfC reaction intermediates (intermediates I and II).

Due to the challenge of isolating undegraded compound 12, we used [u-¹³C-¹⁵N]-GMP as a substrate to monitor the YcfA-catalyzed reaction in real time using ¹³C NMR spectroscopy (Figure S12). A new resonance appears at 140.2 ppm after 30 minutes and gradually decreases over time, while a resonance at 136.9 ppm increases and eventually became predominant. The 136.9 ppm peak corresponds to carbon 5 of compound 13. Based on the time-resolved appearance and interconversion of these two signals, together with the corresponding HPLC profiles (Figure S9), the 140.2 ppm signal associates with compound 12. HRMS analysis from an independent reaction sampled at 30 min, prior to compound 13 appearance, reveals that both compounds 12 and 13 possess the same molecular mass (Figure S11 & S13). The compounds are likely structural isomers associated with different modes of cysteine and GMP conjugation. Based on the ¹³C NMR chemical shift changes at the GMP C₅ position, the ~2 ppm downfield shift in compound 12 is consistent with the strong inductive electron-withdrawing effect of a thiolate group under mildly basic conditions (pH 8), suggesting it is the S-linked conjugate. In contrast, the ~1 ppm upfield shift in compound 13 is consistent with a C-N linkage, where the nitrogen exerts a weaker, possibly electron-donating effect. Taken together with identical molecular masses and the observed time-dependent interconversion, these data collectively support assignment of compound 12 as the S-conjugated isomer and compound 13 as the N-conjugated rearrangement product (Figure 4). While characterizing these intermediates, Ishida et al. reported isolation of compound 13 and demonstrated through chemical synthesis that compound 12 spontaneously rearranges to form compound 13, providing complementary support for this assignment.¹⁶ Furthermore, substitution of 13 for 12 as the YcfC substrate affords no reaction and no 6-TGMP 2 production (Figure S14).

Using cysteine persulfide (15, Figure 4) as the nucleophile, instead of cysteine (Figure 4) provides further evidence for compound 12 (Figure 4) as the proposed intermediate II. In our characterization of anaerobic ergothioneine biosynthesis catalyzed by EanB, we reported a method of producing cysteine persulfide 15 in situ.^23–24 Cystathione β lyase (CBL) produces cysteine persulfide 15 from cystine. Consequently, we used cysteine persulfide 15 as the substrate in a YcfA reaction mixture including YcfA, GMP, and ATP (Figure 4). This modified reaction, however, does not produce the expected intermediate. Only a minor increase in absorbance at 310 nm occurs, suggesting minimal accumulation of conjugated intermediates (Figure S15). Furthermore, the addition of a 50 mM reducing agent (e.g. dithiothreitol, tris(2-carboxyethyl)phosphine, dithionate or 2-mercaptoethanol) fails to promote the formation of the final product, 6-TGMP 2. HPLC analysis confirms only trace amounts of compound 12 are produced, and 6-TGMP 2 formation is detected exclusively under conditions where YcfC is also present (Figure S15). These results indicate that cysteine persulfide produced in situ does not function as a productive nucleophile in the YcfA-catalyzed reaction. Instead, the persulfide species may undergo further sulfur transfer or disproportionation reactions, potentially generating trisulfide or higher polysulfide species. During these processes, a small amount of free cysteine may be generated, and subsequently used by YcfA to form the Cys-GMP adduct (compound 12).

Structural assignment of intermediate I as a GMP-AMP adduct also proved to be challenging. Few details are currently available about nucleotide adducts and their activation intermediates in the literature. Only one crystal structure has captured an adduct intermediate, a uridine-AMP species, which in tRNA 2-thiouridine biosynthesis corresponds to the ATP-dependent activation product.²⁰ Prior work suggests that a substrate phosphorylation product or adenylation product are the activation intermediate,²⁷ however biochemical characterization of these intermediate in solution is challenging and limited information is available in the literature.¹⁹ Stopped flow results suggest that intermediate I binds to YcfA after it is formed. While intermediate II accumulates under multiple turnover conditions, and only one equivalent of intermediate I forms relative to YcfA, even under multiple turnover conditions. This reaction behavior is consistent with an enzyme-bound activation intermediate. Obtaining sufficient quantities of intermediate I, by large scale enzymatic reaction, for structural characterization is difficult as this is a single turnover study. Furthermore, the buffer concentration in the reaction is at least two orders of magnitude higher than that of the target compound (intermediate I) and thus the high salt concentration is a barrier for the subsequent purification of the desired compound after lyophilization. To address this challenge, we modified the enzyme purification and reaction protocol. We purified YcfA in NH₄HCO₃ rather than Tris-HCl or KPi buffer. After YcfA elution from the purification column, we added ATP and GMP to the sample. A 100-mL enzymatic reaction containing 60 μM YcfA, 75 μM GMP, and 75 μM ATP. After a 5 min. reaction time, we treated the resulting product mixture with chloroform to inactivate the YcfA protein. We purified the target compound by HPLC followed by lyophilization. ¹H NMR analysis of the compound reveals a downfield shift of 0.05 ppm for the C₅ proton of intermediate I, relative to unmodified GMP 8.00 ppm to 8.05 ppm). This spectral shift is consistent with a keto-to-enol tautomerization at the C₁ position of GMP, which would exert a deshielding effect (Figure S16). To further support this structural assignment, we left the NMR sample of intermediate I at room temperature for 24 hours and reanalyzed. The resulting ¹H NMR spectrum shows that the diagnostic peaks shift to positions identical to those of a GMP/AMP mixture standard, indicating hydrolysis of the adduct. HRMS analysis further confirms the molecular composition is consistent with a GMP-AMP adduct (calculated for [M+H]⁺ of 715.1003, found 715.1024, Figure S17). Together, these results strongly support assignment of intermediate I as a GMP-adenylation product (10, Figure 4).

YcfC is a PLP-dependent C-S Lyase.

Because intermediate II rearranges to compound 13 (Figure 4) under our reaction conditions, it could not be isolated directly as a pure compound. To functionally evaluate both species as potential YcfC substrates, we adopted a two-step procedure based on the time-course of intermediate II formation (Figure 3C). A 2 mL YcfA reaction containing ATP, GMP, and Cys was quenched after 30 min, when intermediate II reaches its maximal abundance, the small molecules were separated from the YcfA protein by ultrafiltration using a 10 kDa membrane. We then used the filtrate in a YcfC reaction. Concurrent with the consumption of intermediate II, we observed 6-TGMP 2 formation (Figure S18). Additionally, in place of intermediate II, we also tested [β-¹³C]-cysteine as a YcfC substrate using identical reaction conditions. In this reaction, the cysteine desulfurase or C-S lyase activity of YcfC is barely detectable (Figure S19).

Given the 30% sequence similarity to PLP-dependent cysteine desulfurase, including FiSufS from thermophilic Fervidobacterium islandicum AW-1, YcfC was annotated as a sulfur-mobilizing enzyme.²⁸ Other proteins that contain similar sequences include serine-pyruvate aminotransferase from Psychrobacter arcticus 273–4 (30.1% identity)²⁹ and aspartate aminotransferase from Thermus thermophilus (29.5% identity).³⁰ A phylogenetic dendrogram of YcfC (Figure 5A) reveals a clear separation of sulfur-mobilizing proteins between the cysteine desulfurases and C-S lyases. YcfC forms a distinct group with PatB, GliI, and Egt2, which are PLP-dependent C-S lyases participating in the biosynthesis of S-thiocysteine, gliotoxin, and ergothioneine, respectively.^31–33 Other closely related PLP-dependent C-S lyases include alliinase and EgtE.^34–35 These related proteins and our discovery that 6-TGMP 2 biosynthesis does not involve persulfide intermediates suggests that YcfC is a PLP-dependent C-S lyase (Figure 1D), rather than a cysteine desulfurase (Figure 1C). A crucial distinction between PLP-dependent C-S lyases and cysteine desulfurases is the product composition when cysteine is the substrate: a C-S lyase converts cysteine into pyruvate, while a cysteine desulfurase produces alanine. The anaerobic YcfA-YcfC reaction, using [β-¹³C] cysteine as the substrate, converts cysteine into pyruvate, and not alanine (Figure 2D).

Figure 5.

In vitro biochemical assay and bioinformatic analysis of a novel PLP-dependent C–S lyase YcfC. (A) Phylogenetic analyses of YcfC, homologues, and characterized cysteine desulfurases or PLP-dependent C–S lyases;¹⁸ (B) Mutagenesis of potential key cysteine residues in YcfA and YcfC and reconstitution analysis. On the left panel is ¹³C-NMR identification analysis. On the right panel are HPLC traces of YcfA-YcfC reactions with cysteine-free mutants. trace i, YcfA-WT/YcfC-WT; trace ii, YcfA-WT/ YcfC_C-all-A; trace iii, YcfA_C-all-A/YcfC-WT; trace iv, YcfA_C-all-A/ YcfC_C-all-A.; (C) HPLC traces of YcfA/YcfC- or YcfA/YcfC variants-catalyzed reaction system. (i) YcfA-WT/YcfC-WT; (ii) YcfA-WT/YcfC_C-all-A(K147C)-EA treated variant; and (iii) YcfA-WT/ YcfC_C-all-A(K147C).

To provide further evidence that YcfC is a lyase, we conducted an ammonia assay as this reaction would also produce ammonia. We incubated the reaction mixture with the Nessler’s reagent (K₂HgI₄ in KOH solution) and measured it colorimetrically.^35–36 The ratio of the resulting products (pyruvate, ammonia, and 6-TGMP) is roughly 1: 1: 1 (Figures S20 – S22).

Next, we conducted site-directed mutagenesis to determine if any of the cysteine residues in YcfA/YcfC are important in catalysis. YcfA contains three cysteine residues (C25A, C43A, and C113A) and YcfC has six cysteine residues (C69, C115, C221, C231, C232, and C264). The three YcfA single Cys mutants are C25A, C43A, and C113A) and the four YcfC single Cys mutants are C69A, C115A, C221A, and C264A). We also prepared one YcfC double Cys mutant (C231A-C232A), and Cys-free mutants of YcfA (C25A-C43A-C113A, i.e., YcfA_C-all-A) and YcfC (C69A-C115A-C221A-C231A-C232A-C264A, i.e., YcfC_C-all-A). These mutants were produced and purified by following a protocol similar to the one used in the production and purification of wild type (WT) YcfA and YcfC (Figure S23). All single variant mutant proteins are soluble and obtained in a yield comparable to that of the WT enzymes. Using isotopically-labelled [u-¹³C-¹⁵N]-GMP and [β-¹³C]-cysteine as the substrates, these mutant reactions were monitored via ¹³C NMR. All single variant mutants catalyze the production of 6-TGMP 2, albeit with slightly varying conversion rates (Figure S24 and S25). To further confirm the results, we also tested the Cys-free YcfA mutant (YcfA_C-all-A) and the Cys free YcfC-mutant (YcfC_C-all-A) (Figure S26) in combination with WT-YcfC or WT-YcfA. In all reactions, ¹³C-NMR spectra reveal the production of 6-TGMP 2 (trace ii and iii, Figure 5B). The results indicate that none of the YcfA and YcfC cysteine residues are required for YcfA-YcfC catalysis.

In our characterization of YcfC, purified YcfC samples are almost colorless in tris buffer (Figure S4), suggesting that the PLP cofactor is weakly bound, unlike many other PLP-dependent enzymes. Thus, we supplemented YcfC with one equivalent of PLP in all studies. To understand the nature of this YcfC-PLP interaction, we first aligned the sequence of YcfC with homologs (Figure S27) retrieved from the NCBI database, and the analysis suggests that K147 might be required for PLP attachment to YcfC.^37–38 To examine the role of K147 in YcfC-catalysis, we conducted the rescue experiment outlined in Figure 5C. If K147 is important for catalysis, we hypothesized that the YcfC K147C mutant will be inactive. Treatment of the YcfC K147C mutant with 2-bromoethylamie results in aminaoethylation of C147, generating a lysine-like side chain with a terminal amino group, which could potentially restore the catalytic function of K147.³⁹

We introduced the K147C mutation into the cysteine-free YcfC mutant (YcfC_C-all-A), to produce the YcfC_C-all-A(K147C) mutant. This variant also generates soluble mutant proteins (Figure S28). PLP does not bind to this variant either, as indicated by the absence of the 420 nm PLP-absorption feature after PLP-reconstitution (Figure S29).³⁵ 6-TGMP 2 is not detected when the YcfC_C-all-A(K147C) variant and WT YcfA are incubated with GMP, ATP and cysteine (Figure 5C, trace iii). Nevertheless, aminoethylation with 2-bromoethylamie could modify C147 to produce a lysine-like side chain with a terminal amino group at this position (YcfC_C-all-A(K147C)-EA, Figure S29) to rescue K147.³⁹ Aminoethylation and reconstitution with PLP generates the PLP-containing YcfC_C-all-A(K147C) variant (Figure S29). In a reaction containing WT YcfA, YcfC K147 rescue mutant, GMP, ATP and cysteine, the production of 6-TGMP 2 recovers (Figure 5C, trace ii). These results clearly indicate that although the PLP cofactor weakly binds YcfC, an active site lysine (K147 in YcfC) is crucial for its activity, as observed in other PLP dependent enzymes.

CONCLUSIONS

Persulfide-mediated trans-sulfuration pathways are a well-established strategy for sulfur mobilization in biological systems, particularly in tRNA thiolation reactions.^{1–2, 12, 40–42} In the biosynthesis of 4-thiouridine in E. coli tRNA, the PLP-dependent cysteine desulfurase, IscS, catalyzes the desulfuration reaction and passes the substrate cysteine sulfur atom to a protein-based cysteine residue (e.g., Cys456 in ThiI) and persulfide intermediates are key players of this process.^{13, 15, 21, 43–44} The persulfide group on Cys456 of ThiI serves as a nucleophile to attack the activated uridine or used as a source of bisulfide (-SH), which attacks the adenylated uridine residue.^{1, 22} In both scenarios, the participation of a second cysteine residue (Cys344) on ThiI is essential for the release of sulfur from the persulfide, which leads to the formation of a disulfide bond (Cys344-Cys456).^{1, 15, 22} Therefore, reduction (e.g., DTT) of disulfide is a key step to regenerate ThiI for another catalytic cycle.¹⁵ Similar catalytic mechanisms are identified in other tRNA thiolation systems accomplished by MnmA and TtcA, which contribute to the generation of 5-methylaminomethyl-2-thiouridine and 2-thiocytidine, respectively.^45–47 In contrast to these canonical persulfide-mediated thiolation systems, the YcfA-YcfC-catalyzed thioamidation proceeds through a fundamentally different sulfur mobilization strategy. None of the cysteine residues in either YcfA or YcfC are essential for catalysis, and the reaction occurs efficiently in the absence of external reductants, thereby excluding the involvement of protein-bound cysteine persulfide intermediates.

Building on recent mechanistic frameworks for 6-TGMP biosynthesis that integrate biochemical and structural insights,^12,17,17 our work provides direct chemical and kinetic evidence that the YcfA-YcfC system operates through a discrete, stepwise process involving two chemically distinct intermediates. Pre-steady state kinetic analyses reveal the formation of an enzyme-bound activation intermediate, identified as a GMP-AMP adduct (intermediate I, 10), generated by YcfA and retained in the enzyme. This activated nucleotide subsequently converts to a second intermediate, a GMP-Cys conjugate featuring a C-S linkage (intermediate II, 12), which accumulates in solution and serves as the direct substrate for YcfC. These findings provide definitive chemical resolution of the nucleotide activation and sulfur incorporation steps underlying 6-TGMP biosynthesis.

Functional, phylogenetic, and biochemical analyses further establish YcfC as a PLP-dependent C-S lyase rather than a cysteine desulfurase. YcfC clusters phylogenetically with characterized PLP-dependent C-S lyases such as PatB, GliI, Egt2, and EgtE, lmbF, which participate in sulfur-containing natural product biosynthesis.^{31–35, 48} Consistent with this assignment, YcfA-YcfC catalysis yields pyruvate and ammonia as coproducts in stoichiometric proportion to 6-TGMP formation, a hallmark of C-S lyase chemistry. Moreover, comprehensive cysteine-to-alanine mutagenesis experiments show that cysteine residues present in either YcfA or YcfC are not required for catalysis, whereas an active-site lysine residue (K147) is essential for PLP-dependent activity, as confirmed by chemical rescue experiments.

Despite exhibiting low sequence identity (<20%) to previously characterized C-S lyases,^31–32 YcfC is a distinct member of the fold type II PLP-dependent C-S lyase family.³⁸ Together with emerging examples such as PatB, GliI, EgtE, Egt2, alliinase, LmbF, and thiocysteine C-S lyases involved in leinamycin and gliotoxin biosynthesis,^{31–35, 49–50 48} The mechanistic insights and chemical criteria established here provide a framework for identifying and functionally annotating related sulfur-mobilizing enzymes encoded in uncharacterized biosynthetic gene clusters, highlighting the potential for discovery of new sulfur-containing natural products through genome mining efforts.²

MATERIALS AND METHODS

Chemicals and chemical analysis:

All the NMR spectra (¹H-NMR, ¹³C-¹H NMR, ¹³C-NMR) were obtained in D₂O at room temperatures by a Varian 500 MHz spectrometer. The solvent peaks include D₂O: δ 4.65 for ¹H NMR. Data of ¹H NMR are reported with multiplicity (s = singlet, d = doublet and m = multiplet) and coupling constants (J). All the ¹³C-¹H NMR and ASAPHMQC spectra were recorded with complete proton decoupling. High resolution mass spectrometry data was obtained on a Waters Qtof (hybrid quadrupolar/time-of-flight) API US system by electrospray (ESI) in the [positive or negative] mode. Mass correction was done by an external reference using a Waters Lockspray accessory. Mobile phases were water and acetonitrile with 0.1% formic acid. The MS settings were: capillary voltage = 3kV, cone voltage = 35, source temperature = 120 °C and dissolvation temperature = 350 °C. All chemicals and solvents used this study are LC-MS grade. The high-performance liquid chromatography (HPLC) was performed using a Shimadzu HPLC system with a TE-Click CYS column (5 μM, 100 Å, 10 x 250 mm, Acchrom). The samples for HPLC analysis were performed eluting with 100 mM ammonium formate (pH 3.2, solvent A)-acetonitrile (solvent B) (60% B for 20 min, 60%-5% B gradient for 10 min, and 5% B for 10 min) (flow rate: 4 mL/minute, column temperature 30 °C).

General molecular techniques.

Reagents used for molecular biology experiments were purchased from New England Biolabs (Ipswich, MA). Strep-Tactin Sepharose resin was purchased from IBA Lifesciences (Edina, MN). All other reagents or compounds were purchased from Fisher Scientific (Hampton, NH), Sigma-Aldrich (St. Louis, MO) and Bio-Rad (Hercules, CA). Escherichia coli DH5α and BL21(DE3) were used for plasmid maintenance and protein overproduction, respectively. Oligonucleotides were purchased from Thermo Fisher Scientific (Waltham, MA). Gene optimization of YcfA and YcfC coding sequences for heterologous expression in E. coli, as well as the variants including YcfA_C-all-A and YcfC_C-all-A were conducted and synthesized by GenScript (Piscataway, NJ). Site-directed mutagenesis was performed using the KLD-based QuickChange method (New England Biolabs) as per the manufacturer’s instructions using Q5^® High-Fidelity DNA polymerase. Mutations were verified by sequencing using the T7 promoter primer or T7 terminator primer. Plasmid inserts were sequenced at the Genewiz company (South Plainfield, NJ).

Heterologous YcfA protein expression and purification.

The ycfA gene was sub-cloned into pASK-IBA5plus vector and transformed into E. coli strain BL21(DE3) for expression of YcfA. Cells in LB broth (2 L) supplemented with ampicillin (100 mg/L) inoculated with BL21(DE3)/pASK-IBA5plus-YcfA were cultured to an OD600 of 0.6. Protein expression was then induced with 200 μg/L of anhydrotetracycline hydrochloride (AHT), followed by further cultivation with shaking at 180 rpm at 25 °C for additional 12 hours. All of the purification processes were performed in a CoyLab vinyl anaerobic chamber at 4 °C using Strep-Tactin affinity chromatography following standard protocols.⁵¹ The expression and purification of YcfA variants were conducted in an identical fashion to wild type. Protein concentration was determined by Bradford assay, then aliquoted into cryogenic storage vials and stored at −80 °C.

Heterologous YcfC protein expression and purification.

The ycfC gene was sub-cloned into pET28a vector and transformed into E. coli BL21(DE3) for expression YcfC. Cells in LB broth (2L) supplemented with kanamycin (50 mg/L) inoculated with BL21(DE3)/ pET28a-YcfC were grown to an OD₆₀₀ of 0.6. Protein expression was then induced with 0.2 mM of isopropylthio-D-galactoside (IPTG). Expression proceeded with shaking at 180 rpm at 16 °C for 12 h. All of the purification process was performed in a CoyLab vinyl anaerobic chamber at 4 °C using Nickel-NTA affinity chromatography following standard protocols.²⁴ Proteins were desalted using G25 Sephadex size-exclusion column (30 × 600 mm). The fractions with >0.3 mg/mL were combined and concentrated. The expression and purification of YcfC variants was conducted similarly to the wild type. Protein concentration was determined by Bradford assay, then aliquoted into cryogenic storage vials and stored at −80 °C.

In vitro activity assay for YcfA and YcfC.

The in vitro assays for reconstitution of thioamide generation were performed in analogy to the protocol described previously.¹² The one-mL reactions consisted of a reaction buffer [10 mM Tris-HCl (pH 8.0), 50 mM KCl, 12 mM MgCl₂], 2 mM ATP, 100 μM PLP, 2 mM cysteine, 1 mM GMP, 20 μM YcfA, and 10 μM YcfC. The reaction mixtures were incubated at room temperature for 2 h under anerobic conditions. The protein was denatured by adding 300 μL chloroform and the samples were obtained by centrifugation at 13,000 g to remove precipitates. The supernatant was lyophilized and redissolved in deionized water. HPLC analyses and purification were conducted as described in the Chemicals and chemical analysis section.

¹³C-NMR YcfA-YcfC reactions using β-¹³C-Cys as one of the substrates.

The 1 mL reaction system consists of a reaction buffer [10 mM Tris-HCl (pH 8.0), 50 mM KCl, 12 mM MgCl₂], 2 mM ATP, 100 μM PLP, 1 mM β-¹³C-Cys (Cambridge Isotopes Laboratories), 2 mM GMP, 20 μM YcfA, and 10 μM YcfC. All the reactions were stirred at room temperature for 2 hours under anerobic conditions. The protein was denatured by adding 300 μL chloroform and the samples were obtained by centrifugation at 13,000 g to remove precipitates. The supernatant was lyophilized and redissolved in 300 μL D₂O for ¹³C-NMR analysis.

YcfA-YcfC reconstitution reactions with different cysteine variants.

The 1 mL reaction system consists of a reaction buffer [10 mM Tris-HCl (pH 8.0), 50 mM KCl, 12 mM MgCl₂], 2 mM ATP, 100 μM PLP, 2 mM cysteine, 1 mM [u-¹³C-¹⁵N]-GMP, with different combinations of 20 μM YcfA variants or 10 μM YcfC variants. For reaction using cysteine free YcfA and/or cysteine free YcfC, the one-mL reaction system consists of a reaction buffer [10 mM Tris-HCl (pH 8.0), 50 mM KCl, 12 mM MgCl₂], 2 mM ATP, 100 μM PLP, 2 mM cysteine, 1 mM [u-¹³C-¹⁵N]-GMP, 100 μM YcfA variants and 100 μM YcfC variants. All the reactions were stirred at room temperature for 2 hours under anerobic conditions. The protein was denatured by adding 300 μL chloroform and the samples were obtained by centrifugation at 13,000 g to remove precipitates. The supernatant was lyophilized and redissolved in 300 μL D₂O for ¹³C-NMR analysis. HPLC analysis was performed using the same conditions described in the Chemicals and chemical analysis section.

Steady state kinetics of YcfA-YcfC reactions.

L-Cysteine concentration dependence: The steady-state kinetic analysis with variable concentration of L-cysteine (0.05 mM to 4 mM) was monitored at room temperature. The one-mL reaction system consists of a reaction buffer [10 mM Kpi (pH 8.0), 50 mM KCl, 12 mM MgCl₂], 1 mM ATP, variable concentration of L-cysteine, 2 mM GMP, 20 μM PLP, 20 μM YcfA and 20 μM YcfC. The reaction mixture was monitored by Agilent Cary 60 UV-vis at 340 nm.

GMP concentration dependence: The steady-state kinetic analysis with variable concentration of GMP (0.05 mM to 4 mM) was monitored at room temperature. The one-mL reaction system consists of a reaction buffer [10 mM KPi (pH 8.0), 50 mM KCl, 12 mM MgCl₂], 1 mM ATP, 2 mM cysteine, variable concentration of GMP, 20 μM PLP, 20 μM YcfA and 20 μM YcfC. The reaction mixture was monitored by Agilent Cary 60 UV-vis at 340 nm.

ATP concentration dependence: The steady-state kinetic analysis with variable concentration of ATP (0.05 mM to 1 mM) was monitored at room temperature. The one-mL reaction system consists of a reaction buffer [10 mM Kpi (pH 8.0), 50 mM KCl, 12 mM MgCl₂], variable concentration ATP, 2 mM cysteine, 2 mM of GMP, 15 μM PLP, 15 μM YcfA and 15 μM YcfC. The reaction mixture was monitored by Agilent Cary 60 UV-vis at 340 nm.

UV-vis spectral characterization of YcfC and PLP interactions.

To assess YcfC and PLP interactions, 50 μM of purified wild type YcfC in 100 mM Tris-HCl (pH 8.0), 500 mM NaCl, 10% glycerol was used. To examine YcfC-PLP interactions, protein samples (50 μM YcfC-WT, YcfC_C-all-A-K147C protein or the YcfC_C-all-A-K147C-EA) were incubated with 50 μM PLP in 10 mM Kpi (pH 8) and 10% glycerol. The UV-vis spectra for two controls were collected: (i) a PLP only solution (50 μM PLP in 10 mM Kpi (pH 8) with 10% glycerol buffer), and (ii) apo-protein samples to confirm the absence of protein-bound PLP.

Time course of YcfA-YcfC reactions in various YcfA/YcfC ratios.

The 1 mL reaction system consists of a reaction buffer [10 mM Tris-HCl (pH 8.0), 50 mM KCl, 12 mM MgCl₂], 2 mM ATP, 100 μM PLP, 1 mM cysteine, 2 mM GMP, with different protein concentrations: 1) 25 μM YcfA and 1 μM YcfC (YcfA:YcfC=25:1); 2) 25 μM YcfA and 5 μM YcfC (YcfA:YcfC=5:1); 3) 25 μM YcfA and no YcfC (YcfA alone). The reaction mixtures were incubated at room temperature. The UV spectra (300 nm to 400 nm region) of the reaction mixtures were monitored using Agilent Carry 60 UV-Vis spectrometer at 0, 1, 3, 5 and 10 min. of the reaction. An aliquot of the reaction at 0, 5, 10, 20, 30 and 60 min. was withdrawn and quenched by adding 300 μL chloroform, followed by centrifugation at 13,000 g for 5 min. to remove the precipitates. HPLC analyses or purification were conducted as described in the Chemicals and chemical analysis section.

YcfA reaction under various substrate combinations.

The 1 mL reaction system consists of a reaction buffer [10 mM Tris-HCl (pH 8.0), 50 mM KCl, 12 mM MgCl₂], 25 μM YcfA with different substrate combination: 1) 2 mM ATP and 1 mM cysteine; 2) 1 mM cysteine and 2 mM GMP; 3) 2 mM ATP and 2 mM GMP. The reaction mixtures were incubated at the room temperature. The UV spectra (300 nm to 400 nm region) of the reaction mixtures were collected using Agilent Carry 60 UV-Vis spectrometer at 1 and 5 min. of the reaction. Then, the third substrate was added into each of the above reactions: 1) adding 2 mM GMP into the reaction containing 2 mM ATP and 1 mM cysteine; 2) adding 2 mM ATP into the reaction containing 1 mM cysteine and 2 mM GMP; 3) adding 1 mM cysteine into the reaction containing 2 mM ATP and 2 mM GMP. The UV spectra (300 nm to 400 nm region) of the reaction mixtures were collected using Agilent Carry 60 UV-Vis spectrometer at 1 and 5 min. of the reaction.

Stopped-flow UV-vis spectrophotometry assay of YcfA reactions.

Stopped-flow absorbance spectrophotometry (SF-Abs) spectra were collected using an Applied Photophysics Ltd. SX20 spectrophotometer, with a photodiode-array detector. The instrument flow-path was configured in the single-mixing orientation, with a light pathlength of 0.2 cm, wherein UV-vis spectra were collected in the region of 300-700 nm. Solutions from two syringes were mixed in a 1:1 volume ratio. Reaction buffer consisted of 100 mM Tris-HCl (pH 8.0), 500 mM NaCl, 12 mM MgCl₂, 50 mM KCl, and 10 % glycerol. Stock solutions of cysteine, GMP, and ATP were prepared in this buffer system prior to use. These YcfA stopped flow studies include:

Experiment i: Syringe A contained 100 μM YcfA pre-incubated with 2 mM ATP for 5 min., and Syringe B contained 2 mM GMP and 2 mM cysteine.

Experiment ii: Syringe A contained 100 μM YcfA pre-incubated with 2 mM ATP and 2 mM GMP for 5 min., and Syringe B contained 2 mM cysteine.

Experiment iii: Syringe A contained 100 μM YcfA pre-incubated with 2 mM ATP, 2 mM GMP, and 2 mM cysteine for 5 min., and solution B contained the same composition; mixing initiated the reaction without changing the final concentrations.

Experiment iv: Syringe A contained 100 μM YcfA pre-incubated with 2 mM ATP for 5 min., and Syringe B contained 2 mM GMP and 2 mM cysteine.

Experiment v: Syringe A contained 100 μM YcfA pre-incubated with 2 mM ATP and 2 mM GMP for 5 min., and Syringe B contained 2 mM cysteine. For experiment v, because of the instrument loading procedure, the shortest achievable interval between combining YcfA with ATP/GMP and initiating the stopped-flow measurement was ~90 s. Accordingly, Syringe A contained 100 μM YcfA pre-incubated with 2 mM ATP and 2 mM GMP in solution A for 90 s, 5 min, 10 min, or 15 min., respectively. In stopped-flow studies, these Syringe A solutions were then mixed 1:1 with Syringe B solution, which contained 2 mM cysteine in the reaction buffer.

Experiment vi also examined stopped-flow using various YcfA concentration. For example, Syringe A contained 2 mM ATP and 2 mM GMP, along with YcfA at concentrations of 100, 200, 500, or 1000 μM in the reaction buffer, respectively. Syringe B contained only 2 mM cysteine in the reaction buffer.

Synthesis and purification of Intermediate II.

For in vitro synthesis of intermediate II, 35 μM YcfA-WT, 165 μM GMP, and 250 μM cysteine, and 250 μM ATP in 5 mM ammonium bicarbonate buffer supplemented with 2 mM MgCl₂, pH 7.8 were incubated in a total 100 mL reaction. The reactions were incubated at room temperature under anaerobic conditions and monitored using UV-Vis spectrometer. The reaction was quenched by adding 30 mL chloroform to denature the YcfA protein when maximal absorption at wavelength 310 nm was reached. After centrifugation at 13,000 rpm for 5 min., the supernatant was lyophilized, redissolved in 500 μL deionized water. Intermediate II (9.6 mg, tR=21.3 min) was then obtained by HPLC purification by using a Shimadzu HPLC system with a TE-Click CYS column (5 μM, 100 Å, 10 x 250 mm, Acchrom). The samples for HPLC analysis were conducted using the following protocol: 100 mM ammonium formate (pH 3.2, solvent A) and acetonitrile (solvent B) using 60% B for 20 min., 60%-5% B gradient for 10 min., and 5% B for 10 min. at a flow rate of 4 mL/minute and by maintaining the column temperature at 30 °C.

Preparation and purification of compound 13.

For a preparation scale of compound 13, the purified intermediate II (around 7.0 mg) was dissolved in one-mL deionized water and was incubated at room temperature overnight. The reaction progress was monitored by an UV-vis spectrometer. The reaction was terminated when the maximal absorption at 280 nm was reached. Compound 13 (4.6 mg, t_R=16.8 min) was purified by HPLC purification by using a Shimadzu HPLC system with a TE-Click CYS column (5 μM, 100 Å, 10 x 250 mm, Acchrom). The samples for HPLC analysis were conducted using the following protocol: 100 mM ammonium formate (pH 3.2, solvent A) and acetonitrile (solvent B) using 60% B for 20 min., 60%-5% B gradient for 10 min., and 5% B for 10 min. at a flow rate of 4 mL/minute and by maintaining the column temperature at 30 °C.

Time-resolved NMR characterization of the YcfA reaction Intermediates.

The 0.4 mL reaction system consisted of a reaction buffer [100 mM Tris-HCl (pH 8.0), 500 mM NaCl, 10% glycerol, 50 mM KCl, 12 mM MgCl₂] in D₂O, 12 mM ATP, 12 mM cysteine, 6 mM [u-¹³C-¹⁵N]-GMP, with 500 μM YcfA. The reaction solution was transferred into an NMR tube and incubated at room temperature for 30 min., 60 min., and 120 min. before ¹³C-NMR measurements.

YcfC reaction using compound 13 as the substrate.

The 0.2 mL reaction consisted of a reaction buffer [10 mM Tris-HCl (pH 8.0)], 1 mM compound 13, 100 μM PLP and 10 μM YcfC. The reaction mixtures were incubated at room temperature for 1 h under the anerobic conditions. The protein was denatured by adding 300 μl chloroform, and then removed by centrifugation at 13,000 g. HPLC analyses was conducted using a Shimadzu HPLC system with a TE-Click CYS column (5 μM, 100 Å, 10 x 250 mm, Acchrom). The samples for HPLC analysis were conducted using the following protocol: 100 mM ammonium formate (pH 3.2, solvent A) and acetonitrile (solvent B) using 60% B for 20 min., 60%-5% B gradient for 10 min., and 5% B for 10 min. at a flow rate of 4 mL/minute and by maintaining the column temperature at 30 °C.

YcfA reaction using cysteine persulfide generated from CBL/cystine reaction as the substrate.

The 0.5-mL reaction system consisted of a reaction buffer of 10 mM Tris-HCl (pH 8.0), 50 mM KCl, 12 mM MgCl₂, 1 mM ATP, 1 mM cystine (suspension solution), 1 mM GMP, 20 μM YcfA and 10 μM CBL. The reaction mixtures were incubated at room temperature for 5 min. Protein was denatured by adding 300 μL chloroform, followed by centrifugation at 13,000 g for 5 min. The UV spectrum of the reaction supernatant (from 300 nm to 400 nm region) was monitored using Agilent Carry 60 UV-Vis spectrometer. After the UV-Vis spectrum was collected, 5 μM YcfC was spiked into the reaction supernatant and incubated at room temperature for another 5 min. and again, the reaction was monitored using Agilent Carry 60 UV-vis spectrometer.

For examining the activity of cysteine persulfide in YcfA-catalysis, a 1.5-mL reaction mixture consisted of the reaction buffer of 10 mM Tris-HCl (pH 8.0), 50 mM KCl, 12 mM MgCl₂, 1 mM ATP, 1 mM cystine (suspension solution), 1 mM GMP, 20 μM YcfA and 10 μM CBL. The reaction mixtures were incubated at room temperature for 5 min. Protein was denatured by adding 300 μl chloroform, followed by centrifugation at 13,000 g for 5 min. The reaction mixture was aliquoted in 6 tubes having 200 μL supernatant each: 1) without any additional reductants; 2) addition YcfC to a final concentration of 5 μM; 3) addition of DTT to a final concentration of 50 mM; 4) addition TCEP to a final concentration of 50 mM; 5) addition dithionate to a final concentration of 50 mM; 6) addition 2-mercaptoethanol to a final concentration of 50 mM. These mixtures were incubated for 15 min., and then analyzed by HPLC using a Shimadzu HPLC system with a TE-Click CYS column (5 μM, 100 Å, 10 x 250 mm, Acchrom). The samples for HPLC analysis were conducted using the following protocol: 100 mM ammonium formate (pH 3.2, solvent A) and acetonitrile (solvent B) using 60% B for 20 min., 60%-5% B gradient for 10 min., and 5% B for 10 min. at a flow rate of 4 mL/minute and by maintaining the column temperature at 30 °C.

As a control reaction using cysteine as substrate, the YcfA reaction (composition and conditions) is the same as described above. After protein denaturation by adding 300 μL chloroform and centrifugation to remove the precipitated protein, the supernatant was analyzed by HPLC. In parallel, the same reaction supernatant was monitored by UV-vis spectroscopy (300-400 nm) before and after addition of YcfC (5 μM), followed by a final HPLC analysis to verify product formation. HPLC analyses was described earlier.

Intermediate I synthesis and purification.

For in vitro synthesis of intermediate I, 60 μM YcfA-WT, 75 μM GMP, and 75 μM ATP in 5 mM ammonium bicarbonate buffer supplemented with 2 mM MgCl₂, pH 8.0, were incubated in a total volume of 100 mL. The reaction mixture was incubated at room temperature under the anaerobic conditions for 5 min. YcfA protein was then denatured by the treatment with 30 mL chloroform. After centrifugation at 13,000 rpm for 5 min. to remove the precipitated protein, the supernatant was lyophilized, redissolved in 500 μL deionized water. Intermediate I was obtained by HPLC purification by using a Shimadzu HPLC system with a TE-Click CYS column (5 μM, 100 Å, 10 x 250 mm, Acchrom). The samples for HPLC analysis was conducted using 100 mM ammonium formate (pH 3.2, solvent A) and acetonitrile (solvent B) according to the following gradient: 60% B for 20 min., 60%-5% B gradient for 10 min., and 5% B for 10 min. at a flow rate of 4 mL/minute and by maintaining the column temperature at 30 °C. Fractions containing the target compound were collected and lyophilized, and then redissolved in 50 μL of ultrapure water, followed by the addition of 10 μL of 1 M KPi buffer (pH 7.0) to stabilize the pH. To this mixture, 400 μL of D₂O was added to prepare samples for NMR analysis. GMP, AMP, and a GMP/AMP mixture standard for NMR analysis were prepared by following the same protocol. The same amount of KPi was added to the reference standards to ensure identical pH conditions during analysis. All ¹H NMR spectra were acquired on a 500 MHz NMR spectrometer. To monitor intermediate I stability, the intermediate I sample was kept at room temperature for 24 hours and re-analyzed by ¹H NMR to monitor its hydrolysis products (GMP and AMP).

Preparation and use of intermediate II as the substrate in the YcfC reaction.

The 2 mL reaction system consisted of a reaction buffer of 10 mM Tris-HCl (pH 8), 50 mM KCl, 12 mM MgCl₂, 2 mM ATP, 1 mM cysteine, 2 mM GMP, with 25 μM YcfA. The reactions were incubated at room temperature. At 30 min. and 6 h, aliquots were withdrawn and YcfA was removed by centrifugation through a 10-kDa molecular-weight-cutoff ultrafiltration device. The filtrates, which contained only the small-molecule reaction products, were collected and divided into two equal portions.

One portion was analyzed directly by HPLC to monitor the accumulation of intermediate II and compound 13. The second portion was supplemented with 5 μM YcfC and 100 μM PLP and incubated for 30 min. at room temperature. Proteins were then denatured by adding 300 μL chloroform. After removing the precipitated protein by centrifugation, the aqueous phase was analyzed by HPLC. HPLC analysis was conducted using a Shimadzu HPLC system with a TE-Click CYS column (5 μM, 100 Å, 10 x 250 mm, Acchrom). The samples for HPLC analysis were conducted using the following protocol: 100 mM ammonium formate (pH 3.2, solvent A) and acetonitrile (solvent B) using 60% B for 20 min., 60%-5% B gradient for 10 min., and 5% B for 10 min. at a flow rate of 4 mL/minute and by maintaining the column temperature at 30 °C.

¹³C NMR analysis the YcfC reaction using cysteine as the substrate.

The 1 mL reaction system consisted of a reaction buffer [10 mM Tris-HCl (pH 8.0), 50 mM KCl, 12 mM MgCl₂], 100 μM PLP, 2 mM β-¹³C-Cys (Cambridge Isotopes Laboratories), and 50 μM YcfC. All the reactions were incubated at room temperature for 2 hours under anerobic conditions. The YcfC protein was denatured by adding 300 μL chloroform treatment and the denatured protein was removed by centrifugation at 13,000 g. The supernatant was lyophilized and redissolved in 300 μL D₂O for ¹³C NMR analysis

YcfA/YcfC product characterization.

The ¹³C NMR analysis YcfA/YcfC indicated the production of 6-TGMP and pyruvate as the products. In this reaction, there should be another product, NH₄⁺. To quantify the ratio between these three products, NH₄⁺, pyruvate, and 6-TGMP, a 2 mL reaction (containing 5 μM YcfA-WT, 5 μM YcfC-WT, 50 μM PLP, 2 mM GMP, 2 mM ATP, 1 mM cysteine, 12 mM MgCl₂ and 50 mM KCl in Kpi buffer (10 mM, pH 8.0)) was run at room temperature for 60 min. To control for background signals arising from the proteins themselves, a parallel control reaction was prepared in which YcfA and YcfC were first heat-denatured (98 °C, 15 min.) and then added to the reaction substrate mixture under identical conditions. At the end of the 60 min. reaction, the reaction mixture was quenched by adding 300 μl chloroform to denature the proteins. After the precipitated proteins were removed by centrifugation, the supernatant was divided in three aliquots (500 μL for NH₄⁺ quantification, 1 mL for pyruvate quantification and 100 μL for 6-TGMP quantification) for further product analysis, including NH₄⁺, pyruvate, and 6-TGMP quantification.

Quantification of NH₄⁺ produced by YcfA-YcfC reaction system: Ammonia produced in the YcfA-YcfC reaction was quantified as NH₄⁺ as previously described.³⁵ Reaction mixtures were prepared as described above for the coupled YcfA-YcfC reaction. A parallel control reaction containing heat-denatured YcfA and YcfC was prepared under identical conditions to account for background NH₄⁺. At the end of the incubation, 500 μL reaction was transferred to a clean tube and combined with 50 μL Nessler’s reagent and 450 μL deionized water to give a final volume of 1 mL. The samples were mixed and allowed to stand at room temperature for 10 min. Absorbance was then measured at 420 nm. A standard curve was generated in parallel using NH₄⁺ standards treated in the same way (500 μL standard solution, 50 μL Nessler’s reagent and 450 μL water). Sample concentrations were calculated from the standard curve, and values from the denatured-enzyme control were subtracted to correct for background. Based on our calculation, the NH₄⁺ yielded from YcfA-YcfC reaction is 96.59 μM.

Quantification of pyruvate production from YcfA-YcfC reaction: Pyruvate formed in the YcfA-YcfC reaction was quantified using 4-fluorophenylhydrazine derivatization followed by ¹H-NMR analysis.⁵² At the end of the enzymatic reaction, 1 mL of reaction mixture was supplemented with 2 mM 4-fluorophenylhydrazine and incubated at 50 °C for 3 h to allow complete conversion of pyruvate to the corresponding hydrazone derivative. 5 μL of 40 mM Ethyl viologen was added to the 1-mL reaction mixture as an internal standard. The samples were then lyophilized and directly subjected to ¹H-NMR analysis. Pyruvate concentration was determined from the integrated methyl resonance of the hydrazone relative to the internal standard, yielding a value of 124 μM for the YcfA-YcfC reaction.

Quantification of the product 6-TGMP from YcfA-YcfC reaction: The amount of 6-TGMP produced in the coupled YcfA-YcfC reaction was determined by HPLC. HPLC analysis was conducted using a Shimadzu HPLC system with a TE-Click CYS column (5 μM, 100 Å, 10 x 250 mm, Acchrom). The HPLC solvent system and protocols were: 100 mM ammonium formate (pH 3.2) assolvent A and acetonitrile as solvent B, 60% B for the first 20 min., 60%-5% B gradient for the next10 min., followed by 5% B for 10 min. at a flow rate: 4 mL/minute and having the column temperature at 30 °C. A calibration curve was produced by using 100-μL aliquots of 6-TGMP standards with known concentrations (250 μM to 2 mM) and plotting peak area versus concentration. Reaction samples (100 μL) were analyzed under the same HPLC conditions, and the 6-TGMP concentration was calculated from the standard curve. The YcfA-YcfC reaction yielded 114.6 μM 6-TGMP.

YcfC_C-all-A(K147C) aminoethylation.

Aminoethylation of YcfC mutant was accomplished by preparing a 250 μL solution of 2-bromo-ethylamine (BrEA, 135 mM) in 100 mM Tris (pH 8.0), 500 mM NaCl, 10% glycerol buffer, followed by the addition of YcfC_C-all-A(K147C) to a final concentration of 135 μM.⁵³ The reaction was incubated at 25 °C and terminated at 20 h. The mixture was first centrifuges at 13,000 g for 10 min. to remove the precipitated protein, and the supernatant was further purified using G-10 desalting column to remove the excess BrEA. The protein-containing fractions were combined and concentrated using ultrafiltration (10 kDa cut-off) at 4 °C. The final yield was 500 μL of protein at 25 μM.

YcfA-YcfC reactions with YcfC_C-all-A(K147C) variants.

The 1-mL reaction system consisted of a reaction buffer [10 mM Tris-HCl (pH 8.0), 50 mM KCl, 12 mM MgCl₂], 2 mM ATP, 100 μM PLP, 2 mM cysteine, 1 mM GMP, 20 μM YcfA, and either10 μM YcfC_C-all-A(K147C) or YcfC_C-all-A(K147C)-EA. All the reactions were incubated at room temperature overnight under the anerobic conditions. The protein was denatured by adding 300 μL of chloroform and the denatured protein was removed by centrifugation at 13,000 g for 10 min. The supernatant was used for HPLC analysis using a Shimadzu HPLC system with a TE-Click CYS column (5 μM, 100 Å, 10 x 250 mm, Acchrom). The HPLC analysis were done using the following protocol: 100 mM ammonium formate (pH 3.2, solvent A) and acetonitrile (solvent B). The elution protocol was: 60% B for 20 min., 60%-5% B gradient for 10 min., and 5% B for 10 min., with a flow rate of 4 mL/minute and the column temperature was maintained at 30 °C.

Supplementary Material

Supporting material and supporting data are available free of charge via the Internet at http://pubs.acs.org.