Mechanistic studies of the agmatine deiminase from Listeria monocytogenes.

Listeria monocytogenes is a Gram-positive food-borne pathogen that is capable of living within extreme environments (i.e. low temperatures and pH). This ability to survive in such conditions may arise, at least in part, from agmatine catabolism via the agmatine deiminase system (AgDS). This catabolic pathway utilizes an agmatine deiminase (AgD) to hydrolyse agmatine into N-carbamoylputrescine (NCP), with concomitant release of ammonia, which increases the pH, thus mitigating the ill effects of the acidic environment. Given the potential significance of this pathway for cell survival, we set out to study the catalytic mechanism of the AgD encoded by L. monocytogenes In the present paper, we describe the catalytic mechanism employed by this enzyme based on pH profiles, pKa measurements of the active site cysteine and solvent isotope effects (SIE). In addition, we report inhibition of this enzyme by two novel AgD inhibitors, i.e. N-(4-aminobutyl)-2-fluoro-ethanimidamide (ABFA) and N-(4-aminobutyl)-2-chloro-ethanimidamide (ABCA). In contrast with other orthologues, L. monocytogenes AgD does not use the reverse protonation or substrate-assisted mechanism, which requires an active site cysteine with a high pKa and has been commonly seen in other members of the guanidinium-modifying enzyme (GME) superfamily. Instead, the L. monocytogenes AgD has a low pKa cysteine in the active site leading to an alternative mechanism of catalysis. This is the first time that this mechanism has been observed in the GME superfamily and is significant because it explains why previously developed mechanism-based inactivators of AgDs are ineffective against this orthologue.

INTRODUCTION

Listeriosis is an invasive infection by the Gram-positive bacterium Listeria monocytogenes that can lead to miscarriages in pregnant women, meningitis in newborns and death in immuno-compromised individuals [1]. This bacterial pathogen, which is most commonly found in dairy products, meats and seafood, is the third leading cause of death among food-borne pathogens [2]. Low levels of L. monocytogenes have also been detected in both eggs and potatoes [1]. The prevalence of food-borne illnesses associated with this organism can be attributed to its high tolerance for extreme environmental conditions, such as low temperature and pH: L. monocytogenes can survive at less than 4°C and will tolerate the acidic environment found in the gastrointestinal tract [1].

L. monocytogenes possess a four-gene operon (AguABCD) that encodes a so-called agmatine deiminase system (AgDS), which is required for agmatine catabolism [3]. Analogous AgDSs are found in other bacterial species, such as Enterococcus faecalis and Streptococcus mutans, where they are believed to increase acid tolerance and confer a competitive advantage [4,5]. Agmatine is generated from arginine by arginine decarboxylase, which is found in both humans and bacteria. Agmatine enters the AgDS where is it first hydrolysed to N-carbamoylputrescine (NCP) with concomitant release of ammonia by agmatine deiminase (AgD). The resulting NCP is further processed by N-carbamoylputrescine aminohydrolase (CPA) to form putrescine and carbamoyl phosphate, which is then converted into ATP and ammonia by putrescine transcarbamoylase (PTC) (Figure 1) [6]. At low pH, transcription of the AgDS is activated leading to an increased agmatine catabolism [6]. This increase yields higher levels of ammonia and additional ATP production, which is likely to provide a competitive advantage to the bacteria in the form of acid neutralization and increased energy production. Notably, deletion of the AgD in L. monocytogenes attenuates pathogen survival within acidic environments, which indicates that the AgDS and, more specifically, AgD is critical for acid tolerance [6].

Figure 1

The AgDS found in L. monocytogenes

Agmatine is hydrolysed to NCP by AgD. PTC produces putrescine and carbamoyl phosphate. Finally, carbamate kinase (CK) with ADP produces ATP, ammonia and carbon dioxide from carbamoyl phosphate.

AgDs belongs to a larger superfamily of guanidinium-modifying enzymes (GMEs) that includes protein arginine deiminases (PADs), dimethylarginine dimethylaminohydrolases (DDAHs), amidinotransferases (ATs) and dihydrolases. Members of the GME family share a conserved catalytic Cys-His dyad that catalyses the hydrolysis of guanidinium groups of arginine, agmatine or methylarginine to ureido-containing products. To date, AgDs from a number of bacterial species, including Helicobacter pylori, Porphyromonas gingivalis and S. mutans have been identified and characterized [7]. Recently, sequence analysis identified a putative AgD from the genome of L. monocytogenes that was subsequently isolated and confirmed as a genuine AgD [6].

Previously studied bacterial AgDs from H. pylori, S. mutans and P. gingivalis all use a conserved active site cysteine residue for nucleophilic attack on the guanidinium group of agmatine, as well as a conserved histidine residue as a general acid/base throughout catalysis [7,8]. Using insights gained from mechanistic studies, mechanism-based inhibitors were developed to target and inactivate these AgD enzymes. The most potent inhibitors to date are N-(4-aminobutyl)-2-fluoro-ethanimidamide (ABFA) and N-(4-aminobutyl)-2-chloro-ethanimidamide (ABCA), with IC50 values ranging from 250 nM to 15 μM for the tested AgDs [7,8].

The catalytic mechanisms for the AgD enzymes expressed by S. mutans, H. pylori and P. gingivalis have been identified as reverse protonation mechanisms, where the substrate preferentially binds to the deprotonated cysteine form of the enzyme [7]. This preference also means that, in a reverse protonation mechanism, the majority of the enzyme is inactive at its pH optimum (∼7.5) due to the high pKa of the cysteine residue. An alternative mechanism that has been reported for other GMEs (e.g. PAD2 and DDAH) is the substrate-assisted mechanism, wherein the proximity of the incoming positively charged substrate depresses the pKa of the nucleophilic cysteine to promote catalysis [9,10].

Mutagenesis studies of the L. monocytogenes AgD demonstrated that the cysteine and histidine residues are conserved as Cys356 and His216; however, studies to fully elucidate the catalytic mechanism have not previously been conducted. Given the therapeutic potential of this enzyme as a target against infection, we set out to elucidate the mechanism and thus gain insights to aid inhibitor development. In the present paper, we report the results of pH rate profiles, pKa measurements of the active site cysteine residue and solvent isotope effects (SIEs). The results indicate that the AgD from L. monocytogenes uses a low pKa active site cysteine in contrast with other members of this enzyme family. Given the increased reactivity of this moiety, we predict that it will be trivial to generate irreversible inhibitors targeting this unique therapeutic target.

EXPERIMENTAL

Materials

Agmatine sulfate, CHES, HEPES and MES were purchased from Sigma–Aldrich. Iodoacetamide was purchased from Chem-Impex International and 2-chloracetamidine was purchased from Oakwood Products. ABCA and ABFA were synthesized as previously reported [8]. L. monocytogenes AgD was synthesized based on the sequence from Cheng et al. [6] (accession number AEO05059) and cloned into a pET30a vector by GenScript.

Purification of L. monocytogenes AgD

The pET30a-AgD expression construct was transformed into Escherichia coli BL21(DE3) cells. Expression and purification of pet30a-AgD from L. monocytogenes was adapted from Cheng et al. [6]. Overnight cultures were grown in LB medium supplemented with 50 μg/ml kanamycin at 37°C. Volumes of 10 ml of overnight culture were used to inoculate 1 litre of LB medium supplemented with 50 μg/ml kanamycin in a baffled flask (37°C, 250 rev./min) until the D600 was 0.6–0.8. IPTG (0.4 mM) was added to the culture and the temperature was reduced to 15°C for 12 h. The cells were harvested by centrifugation (3795  for 10 min) and resuspended in Lysis Buffer (50 mM PBS, pH 7.4, and 10% glycerol) before being lysed by sonication (60% amplitude, 1 s pulse, 3 s pause, 3 min total). The lysate was centrifuged at 12000  for 20 min and the cleared lysate was applied to the nickel column. The column was washed with PBS and PBS with 500 mM NaCl, followed by increasing concentrations of imidazole (0–500 mM). Collected fractions were analysed by SDS/PAGE (12% gel). Fractions with >95% purity were dialysed and stored in long term storage buffer (50 mM PBS, pH 7.4, with 20% glycerol). A single band was observed on the SDS/PAGE gel at the expected molecular mass of 46.8 kDa (Figure 2A). Protein concentration was determined using the Bradford assay.

Figure 2

Purification and Michaelis–Menten kinetics of L. monocytogenes AgD

(A) SDS/PAGE analysis of purified L. monocytogenes AgD. Lane 1 is the molecular mass standards and lane 2 is the purified AgD. (B) Michaelis–Menten plot of L. monocytogenes AgD at the optimum pH of 7.5.

Kinetic assay

The production of ureido-containing compounds as a result of AgD activity was monitored using a previously established discontinuous assay that monitors the production of ureido-containing compounds [11]. This assay is designed to react with ureido-containing compounds, such as NCP, resulting in a chromophore that absorbs at 540 nm. The steady-state kinetic parameters were determined by pre-incubating 50 mM HEPES, pH 7.5, and various concentrations of agmatine (0–10 mM) for 10 min at 25°C prior to adding 1 μM L. monocyto-genes AgD (60 μl total volume). Reactions were allowed to proceed for 30 min at 25°C before being quenched and 200 μl of the COLDER solution (2.25 M H3PO4, 4.5 M H2SO4, 1.5 mM NH4Fe(SO4), 20 mM diacetyl monoxime and 1.5 mM thiosemicarbizide) was subsequently added for colour development. Of note the enzymatic activity was linear with respect to time and enzyme concentration under these conditions. The mixture was vortex-mixed and incubated at 95°C for 30 min. The absorbance was measured at 540 nm and the initial rates were fitted to eqn (1):

using GraFit version 7.0.3 software.

pH studies

pH profiles were generated for L. monocytogenes AgD by measuring steady-state kinetic parameters over a range of pH values (5.0–11.0) in 50 mM MES (5.0–6.5), HEPES (6.5–8.5) and CAPS (8.5–11) in the presence of various concentrations of agmatine. Reaction mixtures consisting of 50 mM buffer and agmatine (0–10 mM) were pre-incubated at 25°C for 10 min before the addition of 1 μM L. monocytogenes AgD (60 μl total volume). Reactions were quenched and product formation was determined as described in the previous section. The kcat/Km and kcat values obtained were plotted as a function of pH and fitted to eqn (2) (kcat/Km) or eqn (3) (kcat)

where Lim1 is the amount of activity observed at low pH, Lim2 is pHopt, Lim3 is the amount of activity observed at high pH and y represents kcat/Km.

Iodoacetamide substrate protection

A reaction mixture (480 μl total volume) consisting of 50 mM HEPES, pH 7.5, and agmatine (2 or 10 mM) was pre-incubated in the presence and absence of 3 mM iodoacetamide for 10 min at 25°C before the addition of 1 μM L. monocytogenes AgD. At various time points (0, 5, 10, 15, 20, 25 and 30 min), 60 μl aliquots were removed, quenched and incubated with the COLDER solution for 30 min at 95°C. The progress curves were fitted to eqn (4)

using GraFit version 7.0.3 software.

Iodoacetamide inactivation

Inactivation of L. monocytogenes AgD by iodoacetamide was measured over a range of pH values (6.5–10.0) using the buffers previously listed. Time-course experiments were conducted at each pH under increasing concentrations of iodoacetamide. Briefly, 10 mM agmatine was incubated with iodoacetamide (0–20 mM) in 50 mM buffer for 10 min at 25°C before the AgD was added to the reaction. Aliquots (60 μl) were removed at various time points (0, 5, 10, 15, 20, 25 and 30 min), quenched and incubated with 200 μl of the COLDER solution for 30 min at 95°C. The NCP produced was quantified as previously described and plotted against time. The data were fitted to eqn (5)

using GraFit version 7.0.3 software, where the quantity of NCP produced is [P], the initial velocity is vi and kobs(app) is the apparent pseudo-first order rate constant for the inactivation of the AgD. However, the kobs values were determined by extrapolating the kobs(app) values to zero substrate concentration using the transformation constant described in eqn (6)

using GraFit version 7.0.3. Following this transformation, the kobs values were plotted against iodoacetamide concentration and fitted to eqn (7) to obtain the second-order rate constant, kinact/KI,

where kinact is the maximum inactivation rate and KI is the concentration that gives half-maximum inactivation. Finally, the second-order rate constants, kinact/KI, were plotted against pH and fit to eqn (8)

using GraFit version 7.0.3. The minimum and maximum rates of inactivation are described as Ymin and Ymax respectively.

IC50 determination

Reaction mixtures containing 50 mM HEPES, pH 7.5, various concentrations of the inhibitor (ABCA, ABFA or 2-chloroacetamidine), and 1 μM L. monocytogenes AgD were pre-incubated at 25°C. After 15 min, 10 mM agmatine was added to the reaction mixture and incubated for an additional 30 min before being quenched. The COLDER solution (200 μl) was added, vortex-mixed and incubated at 95°C for 30 min. The NCP produced was measured as described in the above sections and the rates were subsequently fitted to eqn (9)

using GraFit version 7.0.3 software.

Solvent isotope effects

Buffers and substrate were made in 2H2O for all SIE assays. The pD (p2H) values of the buffers were determined using the formula pD=pH+0.4 [12]. All reactions were prepared in 50 mM buffer with various concentrations of agmatine (0–10 mM) in ≥ 95% 2H2O. The reactions and data analysis were performed exactly the same as described previously in the pH studies section.

RESULTS

Steady-state kinetic parameters of L. monocytogenes AgD at optimum pH

Initial experiments to determine the kinetic parameters for the AgD from L. monocytogenes were conducted at the optimal pH of 7.5 with increasing concentrations of agmatine. The deiminase activity was linear with respect to time and enzyme concentration (Supplementary Figures S1A and S1B) and the kinetic parameters were found to be kcat=0.517±0.027 s−1, Km=330±90.0 μM and kcat/Km=1.57×103 M−1·s−1 (Figure 2B). These values are in agreement with the previously published values of 0.572±0.051 s−1 for kcat, 650±230 μM for Km and a kcat/Km of 8.80×102 M−1·s−1 [6].

The steady-state kinetic parameters of L. monocytogenes AgD were determined over a pH range of 5.5–11 to aid in the elucidation of the catalytic mechanism. The pH profile curve for log kcat/Km fitted well to a model in which only one ionizable group contributes to substrate capture. The pKa value for the ascending limb of the pH profile is 5.0±1.0 (Figure 3A). The plot of kcat against pH is bell-shaped and fits to a model with two apparent pKa values and a non-limiting zero plateau for the upper limit of 0.3±0.1. The ascending limb has a pKa value of 6.2±0.50 and the descending limb has a pKa value of 9.4±0.20 (Figure 3B). Deiminase activity was linear with respect to time over the pH range, indicating that the loss of activity at pH extremes is not due to an effect on enzyme stability. The pKa values identified from the pH profiles presumably correspond to the active site residues, Cys356 and His216. More specifically, the ascending limb of the pKa profile probably corresponds to Cys356 and the descending limb to His216, thus maximizing the amount of active enzyme at the optimum pH (pH of 7.5).

Figure 3

pH profile of L. monocytogenes AgD

(A) Plot of log kcat/Km against pH and (B) plot of log kcat against pH for L. monocytogenes AgD.

pKa measurements of the active site cysteine (Cys356) by iodoacetamide inactivation kinetics

To correctly identify the ionization state of Cys356 in the catalytic mechanism, the rates of AgD inactivation by iodoacetamide were determined as a function of pH. Iodoacetamide is a well-known non-specific affinity label for cysteine residues that has been used in elucidating the mechanisms of other GME family members (PADs, DDAH and other AgDs) [9,10,13,14]. For these experiments, the enzyme was incubated with various concentrations of iodoacetamide for a specific amount of time prior to measurement of residual activity (Figure 4A). From these plots, the pseudo-first-order rate constant of inactivation, kobs, was determined and plotted as a function of iodoacetamide concentration (Figure 4B). Subsequently, the values of kinact/KI were determined from either the slope of the line or from the ratio of kinact and KI. To obtain the pKa value of Cys356, kinact/KI values were plotted as a function of pH and fitted to eqn (8) (Figure 4C). Based on the data acquired from the inactivation of the AgD with iodoacetamide, the pKa of Cys356 was determined to be 7.2±0.2, which is in reasonable agreement with the ascending limb of the log kcat/Km against pH rate profile. This value is much lower than the pKa values obtained for the active site cysteine of other GMEs, such as the AgDs from H. pylori, S. mutans and P. gingivalis [7]. These enzymes deviate from the simplest assumption that the majority of the enzyme is in the active form at optimum pH (i.e. a reverse protonation mechanism). Instead, the active site cysteine residue from the L. monocytogenes AgD displays a much lower pKa (pKa of ∼7 instead of ∼9) and results in the majority of the enzyme in the active form. Thus, this implies that the cysteine exists mostly as the thiolate at the optimum pH. It should also be noted that substrate is able to protect against inactivation, indicating that the iodoacetamide specifically reacts with Cys356, the active site cysteine (Figure 4D).

Figure 4

pKa measurements of the active site cysteine (Cys356)

(A) Inactivation of L. monocytogenes AgD by various concentrations of iodoacetamide at pH 7.5: (■) 0 mM, (□) 10 mM, (●) 20 mM and (○) 30 mM iodoacetamide. (B) Plot of the pseudo-first-order rate constant of inactivation (kobs) against iodoacetamide concentration. (C) Plot of kinact/KI against pH identifying the pKa of Cys356, i.e. the second-order rate constant of inactivation. (D) Substrate protection experiments with L. monocytogenes AgD demonstrates that the substrate can protect against iodoacetamide inactivation.

Since L. monocytogenes catalyses a reaction involving a nucleophilic thiolate, the steady-state kinetic parameters were determined in ≥95% 2H2O to provide a better understanding of the catalytic mechanism. The log kcat/Km in 2H2O was plotted against pD and compared with the plot of log kcat/Km in H2O against pH (Figure 5). Of note the data from the log kcat/Km in H2O were from Figure 3. The results with respect to kcat/Km indicate that the reaction is faster in 2H2O and illustrate an inverse SIE of 0.4±0.2 (kcat/KHm/kcat/KDm). By contrast, a normal SIE on kcat was observed with a kHcat/kcatD of 2.1±0.16.

Figure 5

SIEs for L. monocytogenes AgD

Plot of log kcat/Km against pH/pD in (●) H2O and (○) 2H2O.

An inverse SIE on kcat/Km has been observed in other GME family members containing a nucleophilic thiolate in the active site [7,13-15]. It has been suggested that the inverse SIE is a result of some GME family members utilizing a thiolate/imidazolium ion pair [7,13,14]. The fractionation factor of a thiol in 2H2O is approximately 0.5, and suggests that there is an increased concentration of the thiolate species. The inverse SIE observed on kcat/Km (0.41) in 2H2O is similar to the fractionation factor described above, and these data suggest that the increased reactivity in 2H2O is a result of the increased concentration of thiolate.

IC50 values for 2-chloroacetamidine, ABCA and ABFA

The results from the SIE and pH profiles support a mechanism utilizing a low-pKa cysteine residue for catalysis. Since the mechanism for L. monocytogenes AgD differs from other bacterial AgDs, we sought to determine the potency of known cysteine inactivators and other well-studied AgD inhibitors. The inactivation of L. monocytogenes AgD with 2-chloroacetamidine, a positively charged inactivator that has been used in previous mechanistic reports of GME family members, was measured and the IC50 value was determined [7,9,13,14,16]. In previously published mechanistic studies of GME family members, such as DDAH and PAD2, 2-chloroacetamidine was used to confirm a substrate-assisted mechanism [9,10]. For these studies, a large shift in the pKa of the active site cysteine was observed with 2-chloracetamidine in comparison with iodoacetamide due to the perturbation of the active site cysteine pKa by the positively charged substrate [10]. The IC50 value of 2-chloracetamidine was determined as 221±17.0 mM for L. monocytogenes (Figure 6). This finding was surprising given that the IC50 value for 2-chloroacetamidine against the AgD from S. mutans is 0.4±0.1 mM, a >550-fold difference (Figure 6). It is also noteworthy that a similar difference was observed for iodoacetamide (Supplementary Figures S2A and S2B). Given the high concentration of 2-chloroacetamidine required for enzyme inactivation, it was not feasible to measure the pKa of the active site cysteine using this compound.

Figure 6

IC50 plots of L. monocytogenes AgD and S. mutans AgD for 2-chloroacetamidine

The IC50 values for other well-studied AgD inactivators, ABFA and ABCA, against the L. monocytogenes AgD were determined. ABFA and ABCA inhibit this enzyme with IC50 values of 146±15.0 and 78.5±1.70 μM respectively (Figure 7). ABFA and ABCA are the most potent inactivators described for H. pylori, S. mutans and P. gingivalis to date. For example, the IC50 values for ABFA and ABCA against the S. mutans AgD are 0.27±0.10 and 0.26±0.61 μM respectively [7,8]. The difference in inhibition of S. mutans AgD and L. monocytogenes AgD for ABFA and ABCA are >300-fold for ABCA and >500-fold for ABFA.

Figure 7

IC50 plots of L. monocytogenes AgD for ABFA and ABCA

DISCUSSION

The guanidinium-modifying family of enzymes is composed of the AgDs, DDAHs, PADs and ATs, which are all associated with a number of human diseases (e.g. bacterial infections, cancer, rheumatoid arthritis and colitis) [17,18]. For this reason, many of these enzymes have been identified as potential therapeutic targets and subjects of several studies to evaluate features that may be exploited for inhibitor design. One such feature is the highly conserved active site cysteine residue that acts as a nucleophile to attack the guanidinium of the substrate during catalysis. The actual mechanisms of catalysis, however, have been shown to vary between GME family members. For example, DDAH and PAD2 both utilize a substrate-assisted mechanism in which binding of the positively charged substrate to the thiol form of the active site cysteine perturbs its pKa, thereby increasing the reactivity of the thiolate [9,10]. In contrast, PAD4 and other AgDs (S. mutans, P. gingivalis and H. pylori) employ a reverse protonation mechanism in which the substrate preferentially binds to the thiolate form of the free enzyme [7,13,14]. Based on previous studies, a general mechanism of catalysis has been proposed for the AgD enzymes [10,19,20]. The physiological implication of the diversity observed among members of the GME family is unclear.

Recent mutagenesis studies on the L. monocytogenes AgD have identified Asp94, His216, Asp218 and Cys356 as the four conserved catalytic residues [6]. Asp94 and Asp218 are believed to correctly orient the substrate in the active site allowing Cys356 to attack the substrate guanidinium. This attack leads to formation of a tetrahedral intermediate, and His216 can act as a general acid to donate a proton to the departing amine during collapse of the intermediate or during its formation. His216 can also act as a general base to abstract a proton from an active site water molecule to activate it for nucleophilic attack on the thiouronium intermediate, which ultimately leads to hydrolysis (Figure 8).

Figure 8

Proposed mechanism of catalysis for L. monocytogenes AgD

Given the conserved active site found in the AgD encoded by L. monocytogenes, we set out to determine the catalytic mechanism utilized by this enzyme. Analysis of the pH profile for kcat/Km shows a sigmoidal curve, and the decrease in activity at low pH suggests that the protonation state of only a single ionizable group is critical for substrate capture. However, the pH profiles for kcat are bell-shaped, suggesting that two ionizable groups are important for the rate-limiting step of the reaction. These two ionizable groups, with pKa values of ∼6.2 and ∼9.4, are most likely to be Cys356 and His216 respectively. Correctly assigning the pKa values of Cys356 and His216 as the ascending or descending limb of the pH profile requires additional experimental data. For these assignments, the rates of inactivation by iodoacetamide were measured over a range of pH values to identify the pKa of Cys356. This rate increased with pH and resulted in a pKa value of 7.2, which is in reasonable agreement with the ascending limb of the pH profile, suggesting it as the pKa of Cys356. Based on these observations, we can assign the ascending limb of the pH profile to Cys356 and the descending limb to His216. Therefore, it is likely that the AgD in L. monocytogenes AgD utilizes a catalytic mechanism in which the active site cysteine has a low pKa. Interestingly, the L. monocytogenes utilizes a catalytic mechanism more analogous to papain, a cysteine protease. In papain, the nucleophilic cysteine has a low pKa and a higher than expected pKa for the active site histidine [15]. Originally, it was widely believed that the PADs would utilize a mechanism similar to the cysteine proteases; however, until now this had not been observed [13].

Bacterial AgDs, PADs and even papain have displayed an inverse SIE on kcat/Km, and this inverse SIE has been attributed to an increase in the concentration of thiolate [7,13,14]. In 2H2O, the concentration of the reactive thiolate can be increased approximately 2-fold; the fractionation factor of a thiol in 2H2O is ∼0.5. For L. monocytogenes, we observed an inverse SIE on kcat/Km of 0.4±0.2, which correlates with the fractionation factor of the thiol in 2H2O. Based on these data, the higher rate observed in 2H2O is most likely to be due to an equilibrium proton transfer to solvent leading to a higher concentration of reactive thiolate. A normal SIE is observed for GME family members that utilize a substrate-assisted mechanism because the higher concentration of thiolate in 2H2O has no effect on kcat/Km since the enzyme binds the substrate as either the thiol or thiolate. Of note, a normal isotope effect on kcat is expected and suggests that proton transfer contributes significantly to the rate-limiting step of the reaction.

Based on these data, Cys356 exist as a thiolate and His216 is protonated in the active form of the L. monocytogenes AgD. Thus, at low pH (pH <5), the active site cysteine exists as the thiol and the histidine contains a positively charged imidazolium side chain. As pH increases, the concentration of thiolate increases and substrate binding increases, as evident by the ascending limb of the pH profile. At a pH above the optimum, a high concentration of the thiolate exists, which increases affinity for the substrate. In addition, the concentration of the deprotonated His216 increases, leading to a reduction in positive charge of the active site and promotes binding of the agmatine, which is evident by the lack of a descending limb of the log kcat/Km pH profile. The lack of a descending limb in the pH profile of log kcat/Km may suggests proton transfer from His216 occurs after or concomitantly with collapse of the tetrahedral intermediate.

Recently, two novel haloacetamidine-based AgD inactivators (ABCA and ABFA) have been developed and are the most potent AgD inhibitors described to date [7,8,21,22]. Since L. monocytogenes AgD utilizes a conserved active site cysteine, we evaluated the effectiveness of these mechanism-based inactivators on this enzyme. Interestingly, the IC50 values for ABCA and ABFA against L. monocytogenes AgD are much higher than those for other AgDs. The loss in potency for these inhibitors may be due to the high pKa of His216 (pKa ∼ 9.5). Previously characterized AgDs (S. mutans, H. pylori and P. gingivalis) and PADs (e.g. PAD1, PAD3 and PAD4) utilize a reverse protonation mechanism, where the active site histidine has a much lower pKa (pKa ∼ 6). Furthermore, previous studies of the PAD inhibitors, F- and Cl-amidine, suggests that protonation of the inhibitor by the active site histidine prior to collapse of the tetrahedral intermediate is a critical step for complete inactivation [23]. For example, in previously studied AgDs, the inhibitors ABCA and ABFA would bind to the enzyme, undergo nucleophilic attack by the active site cysteine and be stabilized through protonation of the intermediate by the active site histidine. This step leads to halide displacement forming the three-membered sulfonium ring, and ultimately results in the inactivated enzyme (Figure 9, steps 1b–4b). However, the results from the present study suggest an alternative mechanism of inactivation. The AgD from L. monocytogenes incorporates an active site histidine with a much higher pKa than other AgDs or related GME family members, which means that after nucleophilic attack, proton transfer to the intermediate is too slow and the intermediate collapses before halide displacement, which results in the regeneration of free enzyme (Figure 9, steps 1a–2a). Further work will be needed to clarify these differences for developing more potent inhibitors to this bacterial AgD. In addition, the increased reactivity of the active site cysteine in this orthologue could be exploited through the use of different warheads.

Figure 9

Influence of the active site histidine pKa on AgD inactivation by ABFA or ABCA

Proposed mechanism of AgD inactivation by a high-pKa histidine as seen in L. monocytogenes (steps 1a–2a) or by a low-pKa histidine as seen in H. pylori, S. mutans or P. gingivalis (steps 1b–4b).