Allosteric activation unveils protein-mass modulation of ATP phosphoribosyltransferase product release

Abstract

Heavy-isotope substitution into enzymes slows down bond vibrations and may alter transition-state barrier crossing probability if this is coupled to fast protein motions. ATP phosphoribosyltransferase from Acinetobacter baumannii is a multi-protein complex where the regulatory protein HisZ allosterically enhances catalysis by the catalytic protein HisG_S. This is accompanied by a shift in rate-limiting step from chemistry to product release. Here we report that isotope-labelling of HisG_S has no effect on the nonactivated reaction, which involves negative activation heat capacity, while HisZ-activated HisG_S catalytic rate decreases in a strictly mass-dependent fashion across five different HisG_S masses, at low temperatures. Surprisingly, the effect is not linked to the chemical step, but to fast motions governing product release in the activated enzyme. Disruption of a specific enzyme-product interaction abolishes the isotope effects. Results highlight how altered protein mass perturbs allosterically modulated thermal motions relevant to the catalytic cycle beyond the chemical step.

Introduction

A sizable body of evidence point to contributions from protein dynamics to enzyme catalysis, spanning both slow, thermally equilibrated motions that redefine the enzyme conformational ensemble towards active populations^1,2,3,4, and non-statistical, femtosecond/picosecond-timescale vibrations coupled to transition-state barrier crossing^5,6,7. In the context of allostery, where an enzyme’s active site responds to perturbation (e.g., ligand binding) of a remote site in the protein, a role is often invoked for thermal motions governing ligand association/dissociation, product release, and cycling time among conformations^3,8, and, more recently, modulating the chemical step^2,9. Yet it is challenging to separate such motions from those stemming from the inherent flexibility of proteins. In addition, experimental probes aiming to alter protein dynamics (e.g., replacement of specific residues) may also perturb the electrostatic potential surface of the system, making it difficult ascribe effects on rates solely to modulation of specific motions^10,11,12. This difficulty arises from the significant rate enhancement achieved by substrate binding to an electrostatically preorganised active site, which minimises the reorganisation energy necessary to stabilise the charge redistribution as the reaction reaches the transition state¹³. Therefore, the role of protein dynamics in enzyme catalysis remains a controversial topic^10,11,14.

A solution to this problem was envisioned by determining heavy-enzyme kinetic isotope effects: the ratio of a reaction rate constant obtained with an unlabelled enzyme to the reaction rate constant obtained with a heavy isotope-labelled version of that enzyme^5,6. The strategy was predicated on extending the Born-Oppenheimer approximation^15,16 to proteins, whereby labelling of enzymes with heavy isotopes would reduce local vibrational frequencies near the bond vibration timescale without interfering with the electrostatic properties of the system^5,6. Applying this approach to selected enzymes where the chemical step could be isolated resulted most commonly in a normal heavy-enzyme kinetic isotope effect. In other words, the reaction with the isotope-labelled enzyme proceeded more slowly through the chemical step^{5,6,7,17,18,19,20}. This was interpreted as evidence for coupling of fast protein dynamics to the chemical step, with a reduction in protein vibrational frequencies decreasing the probability of either crossing the transition-state energy barrier (purine nucleoside phosphorylase, HIV-1 protease, alanine racemase, lactate dehydrogenase)^5,6,18,20 or reaching the tunnelling-ready state (old yellow enzymes)^7,17, or yet, increasing recrossing of the transition-state dividing surface (bacterial dihydrofolate reductase)^21,22. As an exception, alkaline phosphatase showed no evidence for coupling of protein motions to chemistry²³. Curiously, this approach has not yet been reported for enzymes modulated by allosteric effectors, even though protein motions at various timescales in these systems are proposed to mediate communication between allosteric effector binding and the active-site response^2,3,24,25.

ATP phosphoribosyltransferase (ATPPRT) (EC 2.4.2.17), the enzyme responsible for the first and flux-controlling step of histidine biosynthesis^26,27, offers an opportunity to apply this approach to a complex allosteric system. ATPPRT catalyses the nucleophilic attack of N1 of ATP on C1 of 5-phospho-α-D-ribose 1-pyrophosphate (PRPP), displacing pyrophosphate (PP_i) to generate N¹-(5-phospho-β-D-ribosyl)-ATP (PRATP) in the presence of Mg²⁺ (Fig. 1a), and is allosterically inhibited by histidine to shut down the pathway^26,28. ATPPRT activity is also dependent on KCl.^{29,30,31,32,33} ATPPRT is the focus of protein engineering efforts to optimize histidine biocatalytic production³⁴, and a promising target for novel antibiotic discovery against some pathogenic bacteria, including Acinetobacter baumannii and Mycobacterium tuberculosis^30,35,36. Unlike hexameric long-form ATPPRTs, where one polypeptide chain harbours the catalytic and regulatory domains³⁷, hetero-octameric short-form ATPPRTs constitute a more complex allosteric system made up of catalytic (HisG_S) and regulatory (HisZ) proteins where two dimers of HisG_S flank a tetramer of HisZ^{27,29,38,39,40,41}. HisG_S has low catalytic activity on its own and is insensitive to histidine^29,42. HisZ, a histidyl-tRNA synthetase paralogue without any catalytic activity of its own, binds to HisG_S to form the ATPPRT holoenzyme, which allosterically activates catalysis by HisG_S in the absence of histidine^29,30,40,42. HisZ also contains the histidine binding site and allosterically inhibits ATPPRT catalysis in the presence of histidine, playing a dual regulatory role^40,43. Owing to their architectural versatility alongside their biomedical and biotechnological importance, ATPPRTs have been model systems to interrogate allostery, dynamics, and catalysis^{2,8,30,32,38,40,44,45}.

Fig. 1: The effects of isotope-labelling of AbHisG_S at 25 °C.

a The Mg²⁺-dependent reversible reaction catalysed by ATPPRT. b Cartoon and surface representation of AbATPPRT (PDB ID 8OY0) with AbHisZ (not isotope-labelled) in grey, and AbHisG_S colour-coded according to its isotope-labelling pattern. The mass increases are relative to the masses of AbHisG_S and AbATPPRT carrying natural isotope abundance. c Substrate saturation curves for AbHisG_S isotopologues. d Substrate saturation curves for AbATPPRT isotopologues. All data points are shown for two independent measurements. Lines are best fit of the data to Eq. (2). e Dose-response curves for histidine with AbATPPRT isotopologues in the presence of 1.4 mM ATP and 1.0 mM PRPP. All data points are shown. Two independent measurements were carried out, except for [¹⁵N]AbHisG_S with 0–20 μM histidine, where three independent measurements were performed. Solid lines are best fits of the data to Eq. (5).

For A. baumannii ATPPRT, unique among other reported ATPPRTs due to its reaction proceeding via a rapid equilibrium random mechanism³⁰, steady-state and pre-steady-state kinetics studies point to chemistry as the rate-limiting step for nonactivated HisG_S (henceforth referred to as AbHisG_S). Allosteric activation by HisZ (AbHisZ) disproportionately enhances the chemical step, making product release rate-limiting for the hetero-octameric holoenzyme (henceforth referred to as AbATPPRT)⁴⁶. With AbHisG_S, the lack of a burst of product formation suggested no step after chemistry is rate-limiting for the reaction. In agreement, replacement of Mg²⁺ by Mn²⁺, which leads to more efficient charge balance at the transition state of the related Psychrobacter arcticus ATPPRT, increased AbHisG_S steady-state catalytic constant (k_cat), as it does P. arcticus HisG_S k_cat^2,45, suggesting chemistry is the rate-limiting step for the nonactivated enzyme form⁴⁶. On the other hand, with AbATPPRT, a burst of product formation was inferred, although it was too fast at 25 °C to observe directly even with rapid kinetics, suggesting a step after chemistry is rate-limiting. This was corroborated by high solvent viscosity effects, which showed PRATP diffusion from the enzyme to be rate-determining for k_cat⁴⁶. At 5 °C, the burst phase could finally be observed with AbATPPRT; moreover, the single-turnover rate constant (k_STO), which was much higher than k_cat, showed the chemical step is allosterically activated more than 1300-fold in AbATPPRT as compared with AbHisG_S⁴⁶.

We hypothesize fast protein dynamics are involved in the significant allosteric enhancement of the chemical step in AbATPPRT, which would be susceptible to protein-mass modulation. Taking advantage of the fact that AbHisG_S and AbHisZ are purified independently, and the AbATPPRT holoenzyme generated in vitro by mixing the two proteins at defined concentrations³⁰, we employed various isotope-labelling patterns of AbHisG_S accompanied by differential scanning fluorimetry (DSF), alternative-substrate kinetics, site-directed mutagenesis, steady-state and pre-steady-state enzyme kinetics, and temperature-rate profiles to probe the effect of increased protein mass on AbATPPRT catalysis and allostery.

Results

AbHisG_S and AbATPPRT activities are insensitive to protein mass at 25 °C

To assess the effect of increased AbHisG_S mass in catalysis, we purified AbHisG_S from heterologous expression in M9 medium supplemented with different isotopes to produce AbHisG_S carrying natural isotope abundance (unlabelled AbHisG_S), [¹⁵N]AbHisG_S, [¹³C,¹⁵N]AbHisG_S, and [²H,¹³C,¹⁵N]AbHisG_S (the ²H is incorporated in non-exchangeable positions) (Supplementary Fig. 1). Electrospray ionisation/time-of-flight-mass spectrometry (ESI/TOF-MS) demonstrated the molecular masses of [¹⁵N]AbHisG_S, [¹³C,¹⁵N]AbHisG_S, and [²H,¹³C,¹⁵N]AbHisG_S increased by 1.4%, 5.4%, and 11.2%, respectively, from the unlabelled AbHisG_S molecular mass (Supplementary Fig. 2), which would result in increases of, respectively, 0.5%, 2.0%, and 4.2% in AbATPPRT masses, since the AbHisZ mass³⁰ was never altered (Fig. 1b). DSF-based thermal denaturation assays showed AbHisG_S isotopologues display similar thermal unfolding profiles, and data fit to Eq. (1) yielded similar melting temperatures (T_m), except for [²H,¹³C,¹⁵N]AbHisG_S, which was ~4 °C less thermostable than its counterparts (Supplementary Fig. 3). AbATPPRT substrate saturation curves at 25 °C using unlabelled AbHisG_S produced from heterologous expression in LB medium by our published protocol³⁰ and unlabelled AbHisG_S generated here from expression in M9 medium and in M9 with high cell density induction⁴⁷ (the method used henceforth for all AbHisG_S produced in this work) showed negligible effects of different expression protocols and illustrated the expected data spread from one protein preparation to another (Supplementary Fig. 4; Supplementary Table 1).

Substrate saturation curves for AbHisG_S isotopologues at 25 °C (Fig. 1c) were fit to Eq. (2) to yield apparent steady-state kinetic parameters (Supplementary Table 2). While k_cats for [¹³C,¹⁵N]AbHisG_S and [²H,¹³C,¹⁵N]AbHisG_S are marginally lower than for unlabelled AbHisG_S and [¹⁵N]AbHisG_S, the overlapping nature of most data points in Fig. 1c disagrees with distinguishable heavy-enzyme kinetic isotope effects on k_cat (^HEk_cat). As AbHisG_S k_cat is limited by the chemical step, fast dynamics do not directly influence chemistry in this case. This was further probed by pre-steady-state kinetics under multiple-turnover conditions. The AbHisG_S reaction had been shown not to have a burst of product formation⁴⁶, which is reproduced here for unlabelled AbHisG_S, whose traces overlap with those for [²H,¹³C,¹⁵N]AbHisG_S (Supplementary Fig. 5).

Titration of AbHisG_S with AbHisZ (Supplementary Fig. 6) showed activity increased for all isotopologues upon formation of the holoenzyme, and data fit to Eq. (3) resulted in the apparent equilibrium dissociation constants (K_D) for AbHisZ in Supplementary Table 3, with no mass-dependent effects. This allows the concentrations of each AbATPPRT isotopologue to be calculated using Eq. (4). Substrate saturation curves for AbATPPRT isotopologues at 25 °C (Fig. 1d) were fit to Eq. (2), yielding apparent steady-state kinetic parameters (Supplementary Table 4). No ^HEk_cat was observed for AbATPPRT. This is not surprising as very high solvent viscosity effects on AbATPPRT k_cat have shown this rate constant is determined by the diffusion of PRATP from the enzyme⁴⁶, which is not expected to depend on protein mass. Dose-response curves with histidine were best fit to Eq. (5) (Fig. 1e), yielding protein mass-independent half-maximal inhibitory concentrations (IC₅₀) and Hill coefficients (n) (Supplementary Table 5) in range of previously reported values for unlabelled AbATPPRT³⁰.

Allosteric activation triggers mass-dependent product release at 5 °C

Carrying out the AbATPPRT reaction at 5 °C permits observation of the rate of chemistry with rapid kinetics⁴⁶, a strategy we have repeated here to assess the effect of increased protein mass on the chemical step of AbATPPRT. Substrate saturation curves for AbHisG_S isotopologues showed no ^HEk_cat at 5 °C (Fig. 2a; Supplementary Table 6), which was corroborated by overlapping pre-steady-state kinetics traces for all AbHisG_S isotopologues under multiple-turnover conditions (Fig. 2b). These results confirm the lack of fast dynamics coupling to AbHisG_S chemistry at 5 °C, mirroring the results obtained at 25 °C. We have previously shown rapid kinetics of AbATPPRT under single-turnover conditions at 5 °C produces unimolecular single-turnover rate constants (k_STO) when AbATPPRT concentration is higher than 75 μM. In agreement with the presence of a burst of PRATP formation under multiple-turnover conditions, all k_STO were much higher than k_cat, indicating that chemistry is fast in the holoenzyme⁴⁶. We reproduced those results here with unlabelled AbATPPRT from two different batches and with [¹³C,¹⁵N]AbATPPRT and [²H,¹³C,¹⁵N]AbATPPRT, all at 80 μM enzyme (Fig. 2c). The AbHisZ K_D at 5 °C was also determined for all AbHisG_S isotopologues (Supplementary Fig. 7; Supplementary Table 3). In the single-turnover kinetics, as previously reported⁴⁶, a short lag time in PRATP production is observed, and the data were best fit to Eq. (6), describing product formation in two consecutive irreversible steps: probably an isomerisation (e.g. a conformational change) of the AbATPPRT:ATP:PRPP complex followed by on-enzyme formation of PRATP. There was no mass-dependence of the rate constants (Supplementary Table 7). This suggests the chemical step in AbATPPRT catalysis is not linked to fast protein motions.

Fig. 2: The effects of isotope-labelling of AbHisG_S at 5 °C.

a Substrate saturation curves for unlabelled and isotope-labelled AbHisG_S. All data points for two independent measurements are shown. Lines are best fit of the data to Eq. (2). b Rapid kinetics of PRATP formation at 5 °C by unlabelled AbHisG_S and isotope-labelled AbHisG_S under multiple-turnover conditions. Lines are averages of fifteen replicates. c Pre-steady-state kinetics of PRATP formation by AbATPPRT isotopologues under single-turnover conditions. Lines in colour are averages of six replicates. Thin black lines are best fit of the data to Eq. (6). d Substrate saturation curves for AbATPPRT isotopologues. All data points for two independent measurements are shown. Lines are best fit of the data to Eq. (2). e Solvent viscosity effects on unlabelled AbATPPRT k_cat. All data points for two independent measurements at each PRPP concentration are shown as open circles. Closed circles are the mean ± SD of four measurements at all PRPP concentrations. The line is best fit of the data to Eq. (7). f Kinetic sequence encompassed by AbATPPRT k_cat at 5 °C, highlighting the step proposed to depend on protein mass. The asterisk denotes an isomerised complex.

Surprisingly, substrate saturation curves for AbATPPRT isotopologues showed a clear mass-dependence on k_cat (Fig. 2d, Table 1), an effect that manifested itself only upon allosteric activation of the enzyme. As expected due to just a 0.5% increase in AbATPPRT mass, the [¹⁵N]AbATPPRT ^HEk_cat (¹⁵k_cat) is not statistically significant, but the ^13,15k_cat and ^2,13,15k_cat are. These results pose a conundrum, since a step after chemistry remains rate-limiting for AbATPPRT k_cat at low temperature⁴⁶, but PRATP diffusional release, which is rate-limiting at 25 °C, is incompatible with protein-mass dependence. The T_ms are identical for unlabelled AbATPPRT, [¹³C,¹⁵N]AbATPPRT, and [²H,¹³C,¹⁵N]AbATPPRT, and the presence of PRATP led only to marginal and mass-independent increases in T_m (Supplementary Fig. 8; Supplementary Table 8).

Table 1 Apparent steady-state kinetic parameters and ^HEk_cat (mean ± fitting error) at 5 °C for AbATPPRT isotopologues

We considered whether a change in rate-limiting step, still subsequent to chemistry, occurred at low temperature. To test this, AbATPPRT k_cat at saturating concentrations of substrates was determined at 5 °C in the presence of increasing levels of the microviscogen glycerol (Fig. 2e). Determining the K_D for AbHisZ at 12% glycerol (Supplementary Fig. 9) demonstrated that AbHisG_S remained saturated with AbHisZ in the presence of the microviscogen. Furthermore, increasing the concentration of M. tuberculosis pyrophosphatase (MtPPase) did not alter the AbATPPRT rate (Supplementary Fig. 10), indicating the rate remains independent of the coupled enzyme concentration at 12% glycerol. We use MtPPase in ATPPRT assays to drive forward the reaction equilibrium, rendering the reaction essentially irreversible²⁹. A plot of k_cat ratios against relative viscosity produced a slope of 0.03 ± 0.06 upon best fit of the data to Eq. (7), indistinguishable from 0 within experimental error, indicating diffusional steps do not contribute to AbATPPRT k_cat at 5 °C, in sharp contrast to the scenario at 25 °C, where a similar analysis had yielded a slope of 0.96 ± 0.07⁴⁶, within experimental error of the theoretical maximum value of 1, indicating rate-determining diffusion⁴⁸.

A kinetic sequence is proposed to describe the kinetically relevant steps encompassed by AbATPPRT k_cat at 5 °C (Fig. 2f). Reproducing what we previously observed, an isomerisation of the Michaelis complex is followed by the chemical step, here including the fast release of PP_i which is immediately hydrolysed by MtPPase, making the reaction irreversible⁴⁶. These steps are insensitive to protein mass. Absence of solvent viscosity effects suggests the AbATPPRT:PRATP complex undergoes a rate-limiting isomerisation preceding PRATP departure. As this step is mass-dependent, this isomerisation likely involves a protein vibrational motion. At 25 °C, such motion is fast, and PRATP release becomes diffusion-limited, masking the ^HEk_cat.

“Stress testing” the AbATPPRT heavy-enzyme isotope effects

To challenge further the unusual ^HEk_cat reporting on PRATP release from AbATPPRT, an additional AbHisG_S isotopologue was generated, [²H,¹⁵N]AbHisG_S (Fig. 3a; Supplementary Fig. 1e), resulting in increases of 7.0% and 2.6% over the molecular masses of unlabelled AbHisG_S and AbATPPRT, respectively (Supplementary Fig. 2e). The AbHisZ K_Ds at 5 and 25 °C were also determined for [²H,¹⁵N]AbHisG_S (Supplementary Fig. 11; Supplementary Table 3). Substrate saturation curves for [²H,¹⁵N]AbATPPRT at 25 °C again showed no mass-dependent effect (Fig. 3b; Supplementary Table 4), but at 5 °C, the [²H,¹⁵N]AbATPPRT saturation curves fell qualitatively between those of [¹³C,¹⁵N]AbATPPRT and [²H,¹³C,¹⁵N]AbATPPRT (Fig. 3c), although the [²H,¹⁵N]AbATPPRT k_cat was not statistically distinct from either the [¹³C,¹⁵N]AbATPPRT or the [²H,¹³C,¹⁵N]AbATPPRT k_cat (Table 1).

Fig. 3: Additional mass and temperature probes of AbATPPRT ^HEk_cat.

a Cartoon and surface representation of AbATPPRT (PDB ID 8OY0) with AbHisZ (never isotope-labelled) in grey, and [²H,¹⁵N]AbHisG_S in yellow. The mass increases are relative to the masses of AbHisG_S and AbATPPRT carrying natural isotope abundance. b Substrate saturation curves for AbATPPRT isotopologues at 25 °C. All data points for two independent measurements are shown. Lines are best fit of the data to Eq. (2). Open circles denote the same data first depicted in Fig. 1d. c Substrate saturation curves for AbATPPRT isotopologues at 5 °C. All data points for two independent measurements are shown. Lines are best fit of the data to Eq. (2). Open circles denote the same data first depicted in Fig. 2d. d Temperature-dependence of the ^13,15k_cat, ^2,15k_cat, and ^2,13,15k_cat. Data are shown as mean ± SD of three independent measurements, except at 25 °C where they are the mean of two independent measurements. Dashed lines are linear regressions of the data intended only to aid the eye. e Temperature profiles of AbATPPRT (278–308 K) and AbHisG_S (278–318 K) k_cats. All data points of two independent measurements are shown. The line is the best fit to Eq. (8), which produced the activation thermodynamics parameters (mean ± fitting error) shown as inset. The ΔS^‡ and ΔH^‡ are those at 25 °C (298 K). f Solvent viscosity effects on AbATPPRT k_cat at 35 °C. All data points for two independent measurements at each PRPP concentration are shown as open circles, except at 0% glycerol (v/v) where three independent measurements were carried out. Closed circles are the mean ± SD of four measurements at all PRPP concentrations, except at 0% glycerol (v/v), with six measurements at all PRPP concentrations. The line is best fit of the data to Eq. (7). g Substrate saturation curves for unlabelled and [²H,¹³C,¹⁵N]AbHisG_S at 40 °C (318 K). All data points for two independent measurements are shown. Lines are best fit of the data to Eq. (2).

To dissect the influence temperature exerts on the protein mass-dependence of AbATPPRT k_cat, the ^HEk_cat variation with protein mass was evaluated at increasing temperatures from 5 to 25 °C (Supplementary Fig. 12). While the data trended to more pronounced ^HEk_cat at low temperatures, the relationship was not strict (Fig. 3d). The ^HEk_cats increased monotonically with the relative increase in AbATPPRT mass at 5 and 10 °C, but were higher at 10 °C. The ^HEk_cats decreased at 15 °C and again at 20 °C, but their dependence on protein mass became more disperse. Finally, at 25 °C the ^HEk_cat collapsed, probably masked by the AbATPPRT k_cat becoming fully determined by a diffusional step instead of a protein vibration.

The data displayed in Supplementary Fig. 12 indicate that AbATPPRT k_cat is lower at 25 °C than at 20 °C. To confirm and explore this unusual result, we characterised the temperature-rate profile of AbATPPRT and AbHisG_S (Fig. 3e). For the AbATPPRT profile, the K_D for AbHisZ was determined at 35 °C (Supplementary Fig. 13), to ensure the AbHisG_S remained saturated with AbHisZ even at the highest temperature used. Temperature stability tests showed that AbATPPRT is not stable to incubation for 10 min at 40 °C, upon which activity at 30 °C drops precipitously (Supplementary Fig. 14), even though the AbATPPRT T_m is ~55 °C. As ATPPRT catalysis is dependent on KCl^{29,30,31,32,33}, the rate-dependence on KCl concentration was evaluated at 25 and 35 °C (Supplementary Fig. 15), demonstrating that maximum activity is achieved with 100 mM KCl at both temperatures. AbATPPRT activity was determined at saturating concentration of both substrates at every temperature (Supplementary Fig. 16), ensuring the maximum rate was achieved. The AbATPPRT temperature-rate profile determined between 5 and 35 °C (278 and 308 K) indeed confirmed the reduction in k_cat at 25 °C from that at 20 °C, followed again by a rise in k_cat between 25 and 35 °C (Fig. 3e). The most likely explanation for this unusual behaviour is the change in rate-limiting step from a protein vibration-limited product release below 20 °C to a diffusion-limited product release from 25 °C onwards. At 35 °C, k_cat decreases as the glycerol concentration increases, but is unchanged in the presence of the macroviscogen polyethylene glycol-8000 (PEG-8000), demonstrating sizable viscosity effects on PRATP release due to increased microviscogen concentration (Fig. 3f). The K_D for AbHisZ determined at 12% glycerol at 35 °C (Supplementary Fig. 17) indicates AbHisG_S remained saturated with AbHisZ in the presence of glycerol at the 35 °C. The rate was unaltered upon increasing MtPPase concentration (Supplementary Fig. 18), indicating the reaction remains coupled at the highest temperature employed. A plot of k_cat ratios against relative viscosity produced a slope of 1.09 ± 0.05 (Fig. 3f), indicating a diffusional step determines k_cat at 35 °C as it does at 25 °C⁴⁶.

The AbHisG_S temperature-rate profile was determined between 5 and 45 °C (278 and 318 K) (Fig. 3e). The rate-dependence on KCl concentration showed maximum activity is achieved with 100 mM KCl at 25 and 40 °C (Supplementary Fig. 15), and both ATP and PRPP concentrations were saturating at all temperatures (Supplementary Fig. 19). Intriguingly, the Eyring plot was nonlinear (Fig. 3e). Substrate saturation curves at 40 °C showed k_cat is enhanced when Mn²⁺ replaces Mg²⁺ as the divalent metal (Supplementary Fig. 19), as observed at 5 and 25 °C⁴⁶, suggesting the same rate-limiting step (likely chemistry) is operational at high temperature. This raises the possibility AbHisG_S catalysis involves nonzero heat capacity of activation (ΔC_p^‡), and fitting the Eyring plot to Eq. (8) produced thermodynamic parameters of activation shown in Fig. 3e. A negative ΔC_p^‡ on k_cat predicts a reduction in the number of vibrational modes available to absorb energy in the transition state as compared with the ground state (Michaelis complex)⁴⁹. The AbHisG_S k_cat responded less and less to increases in temperature beyond 30 °C (Fig. 3e). We thus conjectured that any coupling of non-statistical motions on chemical barrier crossing might be brought to the fore as the ΔG^‡ term becomes less responsive to temperature. We evaluated the effect of reducing protein vibrational frequencies on the reaction rate at 40 °C (Fig. 3g). No ^HEk_cat was obtained, indicating no enhanced coupling of fast protein vibrations to transition-state barrier crossing in AbHisG_S at high temperature.

Replacement of ATP by ADP abolishes the AbATPPRT heavy-enzyme isotope effects

Crystal structures of P. arcticus ATPPRT in complexes with PRPP:ATP and with PRPP:ADP (Fig. 4a) revealed remarkably similar Michaelis complexes, except for a salt-bridge between ATP γ-PO₄^2- in one HisG_S subunit and Arg73 of the adjacent HisG_S subunit, which is missing in the PRPP:ADP complex³⁸. This interaction is also present in the P. arcticus ATPPRT:PRATP binary complex³⁸, and overlay of this structure with that of unliganded AbATPPRT⁴⁶ (Fig. 4b) illustrates the equivalent Arg70 of AbHisG_S might make a similar interaction in a complex with PRATP. Since product release is rate-limiting for AbATPPRT k_cat, whereas chemistry is rate-limiting for AbHisG_S k_cat, the presumed absence of this interaction when N¹-(5-phospho-β-D-ribosyl)-ADP (PRADP) is the product may contribute to the higher AbATPPRT k_cat, but unaltered AbHisG_S k_cat, obtained when ADP is the substrate⁴⁶. To gain insight into the AbATPPRT ^HEk_cat at the molecular level, we attempted to perturb the ^HEk_cat obtained with [²H,¹⁵N]AbATPPRT and [²H,¹³C,¹⁵N]AbATPPRT by replacing ATP with ADP. No statistically significant mass-dependence of AbATPPRT k_cat was seen at 25 °C (Fig. 4c, Supplementary Table 9). Interestingly, the unlabelled AbATPPRT k_cat with ADP at 5 °C (Supplementary Table 10) is the same as that with ATP (Table 1), but the use of ADP abolished the AbATPPRT ^HEk_cat (Fig. 4d, Supplementary Table 10). The presence of ^HEk_cat with ATP but not with ADP as substrate suggests the PRATP γ-PO₄^2--Arg70 interaction might be involved in the mass-dependence of the AbATPPRT k_cat. However, the similar AbATPPRT k_cat with either nucleotide would point to the same step limiting both reactions. Surprisingly, contrary to what was obtained with ATP (Fig. 2e), high solvent viscosity effects at 5 °C with ADP as substrate were observed, and the plot of k_cat ratios against relative viscosity produced a slope of 1.03 ± 0.02 (Fig. 4e), compatible with PRADP diffusion from the enzyme determining k_cat at 5 °C, despite the unaltered AbATPPRT k_cat compared with that reflecting a protein motion underpinning release of PRATP.

Fig. 4: The effect of ADP as substrate on AbATPPRT ^HEk_cat.

a Ribbon diagram of overlaid active sites of P. arcticus ATPPRT in complexes with PRPP:ATP (PDB ID 6FU2) and PRPP:ADP (PDB ID 6FUA), with carbon atoms in royal blue and cyan, respectively. Only the interaction between ATP and Arg73 is shown (dashed lines). b Ribbon diagram of overlaid active sites of P. arcticus ATPPRT in complex with PRATP (PDB ID 6FU7) and unliganded AbATPPRT (PDB ID 8OY0), with carbon atoms in cyan and maroon, respectively. The only interaction shown as dashed lines is that between PRATP and Arg73 (the equivalent Arg70 in AbATPPRT is shown in maroon). c Substrate saturation curves for AbATPPRT isotopologues at 25 °C with ADP as substrate. All data points for two independent measurements are shown, except: 0.1 mM PRPP/unlabelled AbATPPRT, 0.05 mM PRPP/[²H,¹⁵N]AbATPPRT, 0.4 mM PRPP/[²H,¹³C,¹⁵N]AbATPPRT (varying PRPP), and 0.8 mM PRPP/[²H,¹³C,¹⁵N]AbATPPRT (varying ADP), where three independent measurements were carried out. Lines are best fit of the data to Eq. (2). d Substrate saturation curves for AbATPPRT isotopologues at 5 °C with ADP as substrate. All data points for three independent measurements are shown. Lines are best fit of the data to Eq. (2). e Solvent viscosity effects on AbATPPRT k_cat at 5 °C with ADP as substrate. All data points for three independent measurements at each PRPP concentration are shown as open circles, except at 12% glycerol (v/v) where two independent measurements were carried out. Closed circles are the mean ± SD of six measurements at all PRPP concentrations, except at 12% glycerol (v/v), with four measurements at all PRPP concentrations. The line is best fit of the data to Eq. (7).

AbHisG_S Arg70 is implicated in the AbATPPRT ^HE k _cat

Arg70 is conserved in 67% of the 205 HisG_S amino-acid sequences used in a multiple sequence alignment, a snapshot of which, containing the 20 most similar sequences to AbHisG_S is shown in Supplementary Fig. 21. The other 33% of sequences harbour a Lys residue in the equivalent position. Mapping the sequence conservation onto the structure of the AbHisG_S dimer (Fig. 5a) produced a structural picture of the most and least conserved regions of the protein, based on a conservation index⁵⁰. Arg70 is the most conserved residue in a three-residue loop, flanked by Leu69 (very low conservation) and Ala71 (moderate conservation) (Fig. 5b).

Fig. 5: The effect of R70A mutation on AbATPPRT.

a Amino-acid sequence conservation mapped onto the AbHisG_S dimer structure (PDB ID 8OY0) from highly conserved (score 2.35) to not conserved (score −1.12). Arg70 are shown in stick model. b Close-up view of the amino-acid sequence alignment depicted in Supplementary Fig. 21, flanking the Arg70 position. c Substrate saturation curves for unlabelled R70A-AbATPPRT at 25 °C with either ATP or ADP as substrate. All data points for two independent measurements are shown, except 0.4 mM ADP, where three data points are shown. Lines are best fit of the data to Eq. (2). d Substrate saturation curves for R70A-AbATPPRT isotopologues at 5 °C. All data points for three independent measurements are shown. Lines are best fit of the data to Eq. (2). e Substrate saturation curves for R70A-AbATPPRT isotopologues at 25 °C. All data points for two independent measurements are shown. Lines are best fit of the data to Eq. (2). f Solvent viscosity effects on R70A-AbATPPRT k_cat at 25 °C. All data points for two independent measurements at each PRPP concentration are shown as open circles. Closed circles are the mean ± SD of four measurements at all PRPP concentrations. The line is best fit of the data to Eq. (7). g Pre-steady-state kinetics of PRATP formation by R70A-AbATPPRT isotopologues under single-turnover conditions at 25 °C. Lines in colour are averages of six replicates. Thin grey lines are best fit of the data to Eq. (9).

To check the importance of Arg70 in AbATPPRT catalysis and uncover clues about the protein mass-dependent isotope effects at the molecular level, we produced an R70A-AbHisG_S variant and its isotopologue [¹³C,¹⁵N]R70A-AbHisG_S (Supplementary Fig. 22). Intriguingly, the mutation increased the T_m of both unlabelled R70A-AbHisG_S and [¹³C,¹⁵N]R70A-AbHisG_S by ~5 °C from the corresponding WT-AbHisG_S T_m, but binding to AbHisZ restored the T_ms to their WT-AbATPPRT values (Supplementary Fig. 23). The R70A-AbHisG_S rates at 25 °C at substrate concentrations that would be saturating for WT-AbHisG_S (Supplementary Fig. 24) were significantly reduced ~131-fold compared to WT-AbHisG_S k_cat. Both R70A-AbHisG_S and [¹³C,¹⁵N]R70A-AbHisG_S are allosterically activated by AbHisZ without detriment to AbHisZ affinity both at 5 °C and 25 °C (Supplementary Fig. 25; Supplementary Fig. 26). The R70A mutation caused an ~5-fold decrease in k_cat compared to WT-AbATPPRT, and nullified the higher AbATPPRT k_cat with ADP as substrate observed with the WT variant at 25 °C (Fig. 5c; Supplementary Table 11). These results agree with a role for Arg70 in PRATP release, as predicted above.

Since no statistically significant ^HEk_cat had been observed with ADP as substrate, we hypothesized the R70A mutation would abrogate the AbATPPRT ^HEk_cat even with ATP as substrate because the Arg70 interaction with the PRATP γ-PO₄^2- would be disrupted. No statistically significant R70A-AbATPPRT ^HEk_cat was observed either at 5 °C (Fig. 5d; Supplementary Table 12) or at 25 °C (Fig. 5e; Supplementary Table 13). To assess the effect of the R70A mutation on the mechanism of PRATP release at 25 °C, solvent viscosity effects were measured on R70A-AbATPPRT k_cat (Fig. 5f; Supplementary Fig. 27), and the plot of k_cat ratios against relative viscosity produced a slope of 0.06 ± 0.05, suggesting diffusional steps are kinetically silent, in sharp contrast with the WT-AbATPPRT where PRATP diffusion determines k_cat⁴⁶. This result also reveals that the R70A-AbATPPRT protein mass-dependent isotope effects can be obscured even in the absence of a diffusion-limited step.

The lack of solvent viscosity effects raised the possibility that R70A-AbATPPRT k_cat is limited by chemistry. Unlike the WT enzyme, the R70A-AbATPPRT k_STO could be measured at 25 °C (Supplementary Fig. 28), and data fit to Eq. (9) yielded a value of 36 ± 2 s⁻¹ (mean ± SD) with 80 μM unlabelled R70A-AbATPPRT. The k_STO increased to 44 ± 2 s⁻¹ with 100 μM unlabelled R70A-AbATPPRT (Fig. 5g), whose average trace overlapped with that obtained with 100 μM [¹³C,¹⁵N]R70A-AbATPPRT (k_STO of 42 ± 1 s⁻¹). These observations indicate the measured k_STO are not unimolecular and probably have a contribution from binding, since increasing protein concentration led to an increase in k_STO. Regardless, chemistry is at least 20-fold faster than a subsequent step and therefore does not limit R70A-AbATPPRT k_cat. Finally, the lack of heavy-enzyme kinetic isotope effect on k_STO does not support coupling of fast protein vibrations to the chemical step in R70A-AbATPPRT catalysis.

Discussion

Previous studies have reported isotope-labelling of the protein decreased k_cat in systems where this macroscopic rate constant is limited by product release instead of chemistry (dihydrofolate reductase, formate dehydrogenase, methylthioadenosine nucleosidase) and takes place at slower timescales (μs–ms)^{22,51,52,53,54}. However, these effects differ from the ones observed for AbATPPRT. At least for formate dehydrogenase and a cold-adapted dihydrofolate reductase, where double- and triple-isotope-labelling patterns produced two different molecular masses for the labelled enzymes, the effect on k_cat was not mass-dependent, i.e. isotope-labelled enzymes with distinct masses had the same k_cat reduction^22,53. Mass modulation of such low-frequency motions are not readily intuitive, since the typically small increase in protein mass (e.g. 5–11%) would lead to an imperceptible decrease in vibration frequency due to negligibly low restoring force constants^53,55. The results have instead been attributed to the breakdown of the Born-Oppenheimer approximation and a consequent perturbation of the electrostatic potential surface of the isotope-labelled proteins^53,55.

Protein mass-dependent effects on AbATPPRT are unique because the heavy-enzyme kinetic isotope effects are evident exclusively upon allosteric activation of the enzyme, are strictly protein-mass-dependent, but are only manifested on k_cat, which in AbATPPRT ultimately emerges as millisecond-timescale events. The ^HEk_cat are attributed to a rate-limiting protein motion associated with PRATP release below 25 °C, above which this motion presumably speeds up and PRATP release becomes diffusional and insensitive to protein mass.

Intriguingly, the absence of AbATPPRT ^HEk_STO and AbHisG_S ^HEk_cat speak against a coupling of fast protein vibrations to the chemical step of catalysis in allosterically activated and nonactivated forms of this enzyme, with the important caveat that AbATPPRT chemistry is too fast near physiological temperature to capture even with stopped-flow spectrophotometry⁴⁶, so that no mass-dependent effect could be evaluated. The only other report of lack of heavy-enzyme isotope effects on the chemical step is the metalloenzyme alkaline phosphatase. In that instance, the unaltered mass of the catalytic Zn²⁺ ions, involved in water activation and leaving group departure stabilisation, was invoked as a potential explanation for the decoupling between protein mass and chemical barrier crossing²³. While ATPPRT is not per se a metalloenzyme, the reaction is dependent on Mg²⁺ ions²⁶, and crystal structures, computational chemistry calculations, and enzyme kinetics suggest it is involved directly in chemistry by positioning the substrates for nucleophilic attack and by stabilising the pyrophosphate leaving group^2,38,45. One might hypothesise, by analogy to the alkaline phosphatase proposal, the unaltered mass of Mg²⁺ prevents coupling from fast protein motions to chemistry in AbATPPRT. Expanding these studies to other ATPPRTs and other metalloenzymes may provide a test of this hypothesis.

It is unlikely that increased AbATPPRT mass by isotope-labelling exerts its effect directly on a putative millisecond-timescale motion (AbATPPRT k_cat is ~10 s⁻¹), as large-amplitude motions, common at this timescale, would not respond to very small mass changes^53,55. Two non-mutually exclusive models might be invoked to account for the mass-dependent effects on AbATPPRT. One model relies on picosecond-timescale dynamics of backbone amide groups enabling global, microsecond-to-millisecond-timescale protein motions triggering product release; this is observed in adenylate kinase⁵⁶. These picosecond fluctuations could be directly slowed down by isotope-labelling, and this mechanism cannot be ruled out as the origin of the AbATPPRT heavy-enzyme isotope effects. In adenylate kinase, these fast backbone motions are located at conserved hinge regions that control a large lid-opening motion to release products⁵⁶. While there are hotspots of amino-acid sequence conservation outside the active site in HisG_S (Fig. 5a), their involvement in enabling PRATP release is yet unknown.

Another model is gleaned from the use of the non-physiological substrate ADP, and site-directed mutagenesis of AbHisG_S. Disrupting a salt-bridge between Arg70 and PRATP γ-PO₄²⁻ eliminates the ^HEk_cat (while chemistry remains fast and protein-mass insensitive), suggesting this interaction is important to bring about the ^HEk_cat. Picosecond-timescale side-chain motions are common in proteins^57,58,59, and the long side chain of arginine residues, whose dynamics affects their polar and nonpolar interactions with ligands^60,61, can retain significant residual conformational entropy even while participating in salt-bridges⁵⁹. Therefore, it is possible the Arg70 side-chain fast dynamics and consequently its interaction with PRATP γ-PO₄^2- are directly perturbed by isotope-labelling, which reduces the probability of additional, slower rearrangements in the protein required to release the product. Further studies will be necessary to interrogate the contributions from these two models, but the AbATPPRT results revealed, to the best of our knowledge, a novel facet of the heavy-enzyme isotope effects approach where mass-dependent effects manifest themselves as physical steps regulated by fast dynamics.

Methods

Reagents

All commercially available chemicals were used without further purification. BaseMuncher endonuclease was purchased from AbCam. Ampicillin, dithiothreitol (DTT), isopropyl β-D-1-thiogalactopyranoside (IPTG) and 2-(N-morpholino)ethanesulfonic acid-sodium dodecyl sulfate (MES-SDS) were purchased from Formedium. DH5α chemically competent Escherichia coli, DpnI were purchased from New England Biolabs (NEB). QIAprep Spin Miniprep was from Qiagen. Ethylenediaminetetraacetic acid (EDTA)-free Complete protease inhibitor cocktail was from Roche. ATP, ADP, BL21(DE3) chemically competent E. coli, D₂O, glycerol, histidine, imidazole, lysozyme, PRPP, KCl, MgSO₄, MgCl₂, MnCl₂, CaCl₂, KH₂PO₄, Na₂HPO₄, NaCl, NH₄Cl, [¹⁵N]NH₄Cl, D-glucose, [¹³C₆]D-glucose, [¹³C₆,²H₇]D-glucose, [²H₇]D-glucose, BME vitamins, and tricine were from Merck. Agarose, dNTPSs, kanamycin, 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES), PageRuler Plus Prestained protein ladder, PageRuler^TM Plus Prestained protein ladder, and SYPRO orange protein gel stain were from ThermoFisher Scientific. DNA oligonucleotide primers were synthesised by Integrated DNA Technologies (IDT). MtPPase and tobacco etch virus protease (TEVP) were produced as previously described²⁹. AbHisZ was produced as published³⁰. AbHisG_S was produced from LB culture as published³⁰. PRATP was produced as previously described⁶².

Site-directed mutagenesis of AbHisG_S

Generation of R70A-AbHisG_S by site-directed mutagenesis was carried out with overlapping primers according to the method of Liu and Naismith⁶³. Forward and reverse primer sequences were 5′-GATTCTGGCCGCCTCTGATGTTCCGACGTACGTTGAAAACG-3′ and 5′-GGCGGCCAGAATCAGGATACGCACCTGTTTGTGGG-3′, respectively. The WT-AbHisG_S expression vector³⁰ was used as DNA template. Correct insertion of the mutation was confirmed by DNA sequencing performed by Eurofins Genomics. The unlabelled R70A-AbHisG_S was expressed and purified by the same protocol as the WT-AbHisG_S³⁰, and its concentration was determined spectrophotometrically (NanoDrop) at 280 nm based on the theoretical extinction coefficient (ε₂₈₀) of 10,430 M⁻¹ cm⁻¹ as calculated in the ProtParam tool of Expasy.

AbHisG_S expression in M9 minimum medium

For unlabelled AbHisG_S expression, E. coli BL21(DE3) cells transformed with the AbHisG_S expression vector³⁰ were grown in 1 LB supplemented with 50 µg mL⁻¹ kanamycin at 37 °C until an OD₆₀₀ of 0.6–0.8 was reached. Cells were harvested by centrifugation (6774 × g, 15 min, 4 °C), rinsed with M9 minimum medium, centrifuged (6774 × g, 10 min, 4 °C), before resuspension in 1 L M9 minimum medium⁴⁷ in the presence of 50 µg mL⁻¹ kanamycin. The culture was equilibrated to 16 °C for 1 h before induction with 0.5 mM IPTG and additional 20-h incubation at 16 °C. Cells were harvested by centrifugation (6774 × g, 15 min, 4 °C) and stored at −20 °C. For unlabelled and isotope-labelled AbHisG_S expression by the high-cell-density expression method, the same procedure was followed except an OD₆₀₀ of 3.0–5.0 was reached⁴⁷. The specific isotope labels were introduced to WT- and R70A-AbHisG_S via the high-cell-density expression method by supplementing the M9 minimum media with differently isotope-labelled chemicals as outlined in Supplementary Table 14. Unlabelled and isotope-labelled AbHisG_S variants were purified, and their concentrations determined, as previously described³⁰. WT- and R70A-AbHisG_S isotopologues had their intact mass determined by ESI/TOF-MS analysis performed by the BSRC Mass-Spectrometry and Proteomics Facility at the University of St Andrews.

DSF

DSF measurements (λ_ex = 490 nm, λ_em = 610 nm) for all WT-AbHisG_S isotopologues were performed in 96-well plates on a Stratagene Mx3005p instrument. Reactions (50 µL) contained 100 mM tricine, 100 mM KCl, 15 mM MgCl₂, 4 mM DTT pH 8.5, 6 µM enzyme, with 5X Sypro Orange (Invitrogen) added to each well. Thermal denaturation curves were recorded over a temperature range of 25–93 °C with increments of 1 °C min⁻¹. DSF measurements (λ_ex = 520 nm, λ_em = 558 nm) for all WT-AbATPPRT isotopologues, in the presence and absence of 250 μM PRATP, and for all R70A-AbHisG_S and R70A-AbATPPRT isotopologues were performed in 96-well plates on a QuantStudioÔ1 Real-Time PCR instrument. Reactions (20 µL) contained 100 mM tricine, 100 mM KCl, 15 mM MgCl₂, 4 mM DTT pH 8.5, 8 μM enzyme, with 5X Sypro Orange (Invitrogen) added to each well. Thermal denaturation curves were recorded over a temperature range of 25–93 °C with increments of 0.05 °C s⁻¹. Control curves lacked protein and were subtracted from curves containing protein.

General assay for enzyme activity under steady-state conditions

Initial rates at 10 °C and above were performed in the forward direction in 100 mM tricine pH 8.5, 15 mM MgCl₂, 100 mM KCl, 4 mM DTT, and 10 μM MtPPase. Either PRATP or N¹-(5-phospho-β-D-ribosyl)-ADP (PRADP) formation was monitored by the increase in absorbance at 290 nm (ε₂₉₀ = 3600 M⁻¹ cm⁻¹)⁶⁴ over 60 s with readings every 1 s in 1-cm path-length quartz cuvettes (Hellma) in a Shimadzu UV-2600 spectrophotometer outfitted with a CPS unit for temperature control. Reactions (500 μL) were incubated for 3 mins at the desired temperature prior to being initiated by the addition of PRPP. Initial rates at 5 °C were obtained by monitoring the increase in absorbance at 290 nm due to either PRATP or PRADP formation in an Applied Photophysics SX-20 stopped-flow spectrophotometer outfitted with a 5-μL mixing cell (0.5-cm path length and 0.9 ms dead time) and a circulating water bath for temperature control. One syringe contained all proteins (AbHisG_S, MtPPase, and AbHisZ where applicable) and either ATP or ADP, while the other contained PRPP. Both syringes contained 100 mM tricine pH 8.5, 100 mM KCl, 15 mM MgCl₂, and 4 mM DTT. Reactions were triggered by rapidly mixing 55 μL from each syringe and monitored for 60 s. Control reactions lacked PRPP. Furthermore, controls were conducted to ensure rates were independent of MtPPase concentration. This was ascertained empirically by increasing the MtPPase concentration in the assay and confirming the rates did not change.

General assay for enzyme activity under pre-steady-state conditions

Rapid kinetics under multiple- and single-turnover conditions at 5 and 25 °C were carried out by monitoring the increase in absorbance at 290 nm due to PRATP formation in an Applied Photophysics SX-20 stopped-flow spectrophotometer outfitted with a 5-μL mixing cell (0.5-cm path length and 0.9 ms dead time) and a circulating water bath for temperature control. In every experiment, one syringe contained all proteins (AbHisG_S, MtPPase, and AbHisZ where applicable) and either ATP or PRPP, while the other contained either PRPP or ATP. Both syringes contained 100 mM tricine pH 8.5, 100 mM KCl, 15 mM MgCl₂, and 4 mM DTT. Reactions were triggered by rapidly mixing 55 μL from each syringe. Control reactions lacked PRPP.

Apparent equilibrium dissociation constant (K _D) for AbHisZ

The K_D for AbHisZ was determined at various temperatures and in the presence and absence of 12% glycerol (v/v) by measuring initial rates of WT-AbHisG_S isotopologues (0.04 μM, except for [¹³C,¹⁵N]AbHisG_S at any temperature, and for unlabelled AbHisG_S at 35 °C, whose concentration was 0.02 μM) in the presence of 1.4 mM ATP, 1.0 mM PRPP, and varying concentrations of AbHisZ (0–0.5 μM, except at 35 °C in 12% glycerol, where the range was 0–0.2 μM). For R70A-AbHisG_S isotopologues (0.24 μM at 25 °C, 0.30 at 5 °C, and 0.30 μM at 25 °C in 12% glycerol) in the presence of 1.4 mM ATP, 1.0 mM PRPP (or 1.6 mM ATP and 1.4 mM PRPP in 12% glycerol), and varying concentrations of AbHisZ (0–1 μM at 25 °C, 0–4 μM at 5 °C).

Histidine dose-dependence for AbATPPRT

The histidine dose-dependence on AbATPPRT isotopologues was determined as previously published³⁰.

Substrate saturation curves for AbHisG_S at 5, 25, and 40 °C

Initial rates for AbHisG_S isotopologues were measured at saturating concentrations of one substrate (either 3.2 mM PRPP or 6.4 mM ATP) and varying concentrations of the other, either ATP (0–6.4 mM) or PRPP (0–3.2 mM). At 5 °C, enzyme concentrations were 3 μM for unlabelled AbHisG_S, [¹⁵N]AbHisG_S, and [²H,¹³C,¹⁵N]AbHisG_S, and 1.5 μM for [¹³C,¹⁵N]AbHisG_S. At 25 °C, enzyme concentrations were 1 μM for unlabelled AbHisG_S and [¹⁵N]AbHisG_S, and 0.75 μM for [¹³C,¹⁵N]AbHisG_S and [²H,¹³C,¹⁵N]AbHisG_S. At 40 °C, enzyme concentration was 0.25 μM for unlabelled AbHisG_S and [²H,¹³C,¹⁵N]AbHisG_S. In addition, at 40 °C, AbHisG_S (0.1 μM) substrate saturation curves were also determined in the presence of 15 mM MnCl₂ instead of MgCl₂.

Substrate saturation curves for AbATPPRT at 5 and 25 °C

Initial rates for AbATPPRT isotopologues were measured at saturating concentrations of one substrate (1.6 mM either PRPP, ATP or ADP) and varying concentrations of the other, either the nucleotide (0–1.6 mM) or PRPP (0–1.6 mM). With ATP as substrate at 5 °C, enzyme concentrations were 0.039 μM for unlabelled AbATPPRT, [¹⁵N]AbATPPRT, [²H,¹⁵N]AbATPPRT and [²H,¹³C,¹⁵N]AbATPPRT, and either 0.014 μM (at high substrate concentration) or 0.028 μM (at low substrate concentration) for [¹³C,¹⁵N]AbATPPRT. At 25 °C, enzyme concentrations were 0.080 μM for unlabelled AbATPPRT, [¹⁵N]AbATPPRT, [²H,¹⁵N]AbATPPRT and [²H,¹³C,¹⁵N]AbATPPRT, and 0.040 for [¹³C,¹⁵N]AbATPPRT. With ADP as substrate at both temperatures, 0.040 μM enzyme was used.

R70A-AbHisG_S activity at 25 °C

Initial rates for 10 μM R70A-AbHisG_S were measured for 5 min in the presence of 3.2 mM PRPP, 6.4 mM either ATP or ADP, and 25 μM MtPPase. Controls lacked enzyme.

Substrate saturation curves for R70A-AbATPPRT at 5 and 25 °C

Initial rates for AbATPPRT isotopologues were measured at saturating concentrations of one substrate (either 3.2 mM PRPP or 6.4 mM ATP or ADP) and varying concentrations of the other, either the nucleotide (0–6.4 mM) or PRPP (0–3.2 mM). At 5 °C, enzyme concentration was 0.500 μM. At 25 °C, enzyme concentration was 0.240 μM.

Solvent viscosity effects for AbATPPRT

At 5 °C, initial rates were determined in the presence 10 μM MtPPase, of 1.4 mM and 1.6 mM PRPP, 1.6 mM either ATP or ADP, and 0–12% glycerol (v/v). WT-AbATPPRT concentration was 0.080 μM with ATP as substrate, and 0.040 μM with ADP as substrate. At 25 °C, initial rates were determined in the presence of 3.0 mM and 3.2 mM PRPP, 6.4 mM ATP, and 0–12% glycerol (v/v). R70A-AbATPPRT concentration was 0.230 μM. At 35 °C, initial rates were determined in the presence of 1.6 mM and 2.0 mM PRPP, 1.6 mM ATP, and 0–12% glycerol (v/v). WT-AbATPPRT concentration was 0.020 μM. Initial rates were also measured in the presence of 15 μM MtPPase at the highest PRPP concentration and 12% glycerol to confirm the rates were not dependent on MtPPase concentration. At 5 °C with ADP as substrate, and at 35 °C with ATP as substrate, where a significant solvent viscosity effect was observed, the rates were also measured in the presence of 5% PEG-8000.

Temperature-rate profile for AbATPPRT

The temperature stability of AbATPPRT was assessed by incubating 0.025 μM AbATPPRT, 10 μM MtPPase, 1.6 mM ATP at either 30 °C, 35 °C, or 40 °C for 10 min before incubating at 30 °C for 3 min and measuring initial rates upon addition of PRPP to a final concentration of 1.6 mM. The temperature-rate profile was determined by measuring initial rates of AbATPPRT at temperatures ranging from 5 °C (278 K) to 35 °C (308 K) in 5-°C increments at saturating concentrations of one substrate (1.6 mM either PRPP or ATP) and varying concentrations of the other substrate near saturation at each temperature (0.4–1.6 mM). AbATPPRT concentrations were 0.040 μM for 5 °C and 25 °C, 0.050 μM for 10 °C and 15 °C, and 0.025 μM for the remaining temperatures.

Temperature-rate profile for AbHisG_S

The temperature-rate profile was determined by measuring initial rates of AbHisG_S at temperatures ranging from 5 °C (278 K) to 45 °C (318 K) in 5-°C increments at saturating concentrations of one substrate (either 3.2 mM PRPP or 6.4 mM ATP) and varying concentrations of the other substrate near saturation at each temperature (0.8–3.2 mM PRPP; 1.6–6.4 mM ATP). AbHisG_S concentrations were 3 μM for 5 °C, 1 μM for 10 °C to 25 °C, 0.5 μM for 30 °C, 0.25 μM for the remaining temperatures.

AbHisG_S and AbATPPRT rate-dependence on KCl concentration

AbHisG_S initial rates (at 25 °C and 40 °C) were measured in the presence of 50–150 mM KCl, 3.2 mM PRPP, 6.4 mM ATP, and either 1 μM (25 °C) or 0.250 μM (40 °C) AbHisG_S. AbATPPRT initial rates (at 25 °C and 35 °C) were measured in the presence of 50–150 mM KCl, 1.6 mM PRPP, 1.6 mM ATP, and either 0.040 μM (25 °C) or 0.025 μM (35 °C) AbATPPRT. Between the lowest and highest temperatures, the maximum pH variation of the buffer was 0.2 units, and the assay pH of 8.5 lies on a pH-independent region of the AbATPPRT pH-rate profile⁴⁶.

Temperature dependence of the AbATPPRT ^HE k _cat

Initial rates of AbATPPRT isotopologues were measured from 5 to 25 °C in 5 °C increments in the presence of 1.6 mM of each substrate and 0.040 μM enzyme except for [¹³C,¹⁵N]AbATPPRT, whose concentration was 0.030 μM.

Multiple-turnover pre-steady-state kinetics

Rapid kinetics of PRATP formation by AbHisG_S isotopologues at 5 and 25 °C were carried out as previously described⁴⁶, except that at 25 °C, 2000 data points were collected in 2.5 s.

Single-turnover pre-steady-state kinetics

Rapid kinetics of PRATP formation by AbATPPRT isotopologues under single-turnover conditions at 5 °C was performed as previously reported⁴⁶, except that enzyme concentration used here was 80 μM. Rapid kinetics of PRATP formation by R70A-AbATPPRT under single-turnover conditions at 25 °C was performed with enzyme concentrations of 80 μM and 100 μM for unlabelled AbATPPRT and 100 μM [¹³C,¹⁵N]R70A-AbATPPRT with 4700 data points collected in 0.47 s, with 6 μM PRPP and 6.4 mM ATP.

Multiple sequence alignment of HisG_S

Sequences for the multiple sequence alignment were acquired by first conducting a BLASTp search of the UniProtKB/Swiss-Prot database using the AbHisG_S sequence as the query. Sequences showing 100% sequence identity to AbHisG_S were excluded. For AbHisG_S this returned a total of 256 ATP phosphoribosyltransferase sequences; those containing over 250 residues were excluded, as it is likely that those corresponded to HisG_L sequences, resulting in a total of 205 different HisG_S sequences. The MSA was conducted using Clustal Omega⁶⁵. The degree of conservation was calculated using the AL2CO algorithm⁵⁰ as implemented in ChimeraX⁶⁶ then mapped onto the structure of the AbHisG_S protein (PDB: 8OY0)⁴⁶.

Kinetics and thermal denaturation data analysis

Kinetics and thermal denaturation data were analysed by the nonlinear regression function of SigmaPlot 14.0 (SPSS Inc.). Thermal denaturation data were fitted to Eq. (1). Substrate saturation curves at a fixed concentration of the co-substrate were fitted to Eq. (2). Initial rate data at varying concentrations of HisZ were fitted to Eq. (3). The concentration of ATPPRT at any concentration of AbHisG_S and AbHisZ was calculated according to Eq. (4). Histidine dose-response data were fitted to Eq. (5). Single-turnover data with WT-AbATPPRT isotopologues were fitted to Eq. (6). Plots of k_cat ratios against relative viscosity were fitted to Eq. (7). Temperature-rate profile for AbHisG_S was fitted to Eq. (8). Single-turnover data with R70A-AbATPPRT isotopologues were fitted to Eq. (9). In Eqs. (1)–(9), F_U is fraction unfolded, T is the temperature in °C, T_m is the melting temperature, c is the slope of the transition region, and LL and UL are folded and unfolded baselines, respectively; k_cat is the steady-state turnover number, v is the initial rate, E_T is total enzyme concentration, K_M is the apparent Michaelis constant, S is the concentration of the varying substrate, V_max is the maximal velocity, G is the concentration of AbHisG_S, Z is the concentration of AbHisZ, K_D^app is the apparent equilibrium dissociation constant, ATPPRT is the concentration of AbATPPRT holoenzyme, IC₅₀ is the half-maximal inhibitory concentration of inhibitor, n is the Hill coefficient, I is the concentration of inhibitor, v_i and v₀ are initial rates in the presence and absence of inhibitor, P(t) is product concentration as a function of time t, k₂ and k₃ are rate constants governing sequential steps in a single turnover, ES is the enzyme-substrate complex concentration, k_cat⁰ and k_cat^η represent the k_cat in the absence and presence of glycerol, respectively, η_rel is the relative viscosity of the solution, m is the slope, k_STO is the apparent single-turnover rate constant, A is the signal amplitude, k_B, h, and R are the Boltzmann, Planck, and gas constants, respectively, T is the temperature, T₀ is the reference temperature (298 K here), ${\Delta H}_{{T}_{0}}^{{{\ddagger}} }$ and ${\Delta S}_{{T}_{0}}^{{{\ddagger}} }$ are the activation enthalpy and entropy, respectively, at T₀, and ΔC_P^‡ is the activation heat capacity.

$${F}_{{{{{{\rm{U}}}}}}}={LL}+\frac{{UL}-{LL}}{1+{{{{{{\rm{e}}}}}}}^{({T}_{{{{{{\rm{m}}}}}}}-T)/c}}$$

(1)

$$\frac{v}{{E}_{{{{{{\rm{T}}}}}}}}=\frac{{k}_{{{{{{\rm{cat}}}}}}}S}{{K}_{{{{{{\rm{M}}}}}}}+S}$$

(2)

$$v={V}_{\max }\frac{G+Z+{K}_{{{{{{\rm{D}}}}}}}^{{{{{{\rm{app}}}}}}}-\sqrt{{\left(G+Z+{K}_{{{{{{\rm{D}}}}}}}^{{{{{{\rm{app}}}}}}}\right)}^{2}-4{GZ}}}{2G}$$

(3)

$${ATPPRT}=\frac{G+Z+{K}_{{{{{{\rm{D}}}}}}}^{{{{{{\rm{app}}}}}}}-\sqrt{{\left(G+Z+{K}_{{{{{{\rm{D}}}}}}}^{{{{{{\rm{app}}}}}}}\right)}^{2}-4{GZ}}}{2}$$

(4)

$$\frac{{v}_{{{{{{\rm{i}}}}}}}}{{v}_{0}}=\frac{1}{1+{\left(\frac{I}{{{{{{{\rm{IC}}}}}}}_{50}}\right)}^{n}}$$

(5)

$$P\left(t\right)=\frac{{ES}}{{k}_{2}+{k}_{3}}\left[{k}_{2}\left(1-{{{{{{\rm{e}}}}}}}^{{-k}_{3}t}\right)-{k}_{3}\left(1-{{{{{{\rm{e}}}}}}}^{{-k}_{2}t}\right)\right]$$

(6)

$$\frac{{k}_{{{{{{\rm{cat}}}}}}}^{0}}{{k}_{{{{{{\rm{cat}}}}}}}^{{{{{{\rm{\eta }}}}}}}}=m\left({{{{{{\rm{\eta }}}}}}}_{{{{{{\rm{rel}}}}}}}-1\right)+1$$

(7)

$${{{{\mathrm{ln}}}}}\frac{{k}_{{{{{{\rm{cat}}}}}}}}{T}={{{{\mathrm{ln}}}}}\frac{{k}_{{{{{{\rm{B}}}}}}}}{h}-\frac{{\Delta H}_{{T}_{0}}^{{{\ddagger}} }+{C}_{{{{{{\rm{P}}}}}}}^{{{\ddagger}} }\left(T-{T}_{0}\right)}{{RT}}+\frac{{\Delta S}_{{T}_{0}}^{{{\ddagger}} }+{C}_{{{{{{\rm{P}}}}}}}^{{{\ddagger}} }{{{{\mathrm{ln}}}}}\left(T/{T}_{0}\right)}{R}$$

(8)

$$P\left(t\right)=A({1-{{{{{\rm{e}}}}}}}^{-{k}_{{{{{{\rm{STO}}}}}}}t})$$

(9)

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.