Catalytic Cycle of the Bifunctional Enzyme Phosphoribosyl-ATP Pyrophosphohydrolase/Phosphoribosyl-AMP Cyclohydrolase

The bifunctional enzyme phosphoribosyl-ATP pyrophosphohydrolase/phosphoribosyl-AMP cyclohydrolase (HisIE) catalyzes the second and third steps of histidine biosynthesis: pyrophosphohydrolysis of N1-(5-phospho-β-D-ribosyl)-ATP (PRATP) to N1-(5-phospho-β-D-ribosyl)-AMP (PRAMP) and pyrophosphate in the C-terminal HisE-like domain, and cyclohydrolysis of PRAMP to N-(5′-phospho-D-ribosylformimino)-5-amino-1-(5″-phospho-D-ribosyl)-4-imidazolecarboxamide (ProFAR) in the N-terminal HisI-like domain. Here we use UV–VIS spectroscopy and LC–MS to show Acinetobacter baumannii putative HisIE produces ProFAR from PRATP. Employing an assay to detect pyrophosphate and another to detect ProFAR, we established the pyrophosphohydrolase reaction rate is higher than the overall reaction rate. We produced a truncated version of the enzyme-containing only the C-terminal (HisE) domain. This truncated HisIE was catalytically active, which allowed the synthesis of PRAMP, the substrate for the cyclohydrolysis reaction. PRAMP was kinetically competent for HisIE-catalyzed ProFAR production, demonstrating PRAMP can bind the HisI-like domain from bulk water, and suggesting that the cyclohydrolase reaction is rate-limiting for the overall bifunctional enzyme. The overall kcat increased with increasing pH, while the solvent deuterium kinetic isotope effect decreased at more basic pH but was still large at pH 7.5. The lack of solvent viscosity effects on kcat and kcat/KM ruled out diffusional steps limiting the rates of substrate binding and product release. Rapid kinetics with excess PRATP demonstrated a lag time followed by a burst in ProFAR formation. These observations are consistent with a rate-limiting unimolecular step involving a proton transfer following adenine ring opening. We synthesized N1-(5-phospho-β-D-ribosyl)-ADP (PRADP), which could not be processed by HisIE. PRADP inhibited HisIE-catalyzed ProFAR formation from PRATP but not from PRAMP, suggesting that it binds to the phosphohydrolase active site while still permitting unobstructed access of PRAMP to the cyclohydrolase active site. The kinetics data are incompatible with a build-up of PRAMP in bulk solvent, indicating HisIE catalysis involves preferential channeling of PRAMP, albeit not via a protein tunnel.

Enzymes involved in the histidine biosynthesis pathway are attractive targets for antibiotic development, as they carry out essential functions during infection and have no homologues in humans. For instance, histidine biosynthesis protects M. tuberculosis from host-imposed starvation. 11 In A. baumannii, high-throughput transposon library analysis demonstrated that six enzymes of the histidine biosynthetic pathway, including HisIE, are required for the bacterium's persistence in the lungs during pneumonia. 2 In an independent study, knockout of the gene encoding another enzyme in the pathway, imidazole glycerol phosphate synthase, significantly increased host survival in a murine model of pneumonia caused by A. baumannii. 12 Histidine is required for zinc acquisition and lung infection by A. baumannii, 13 and it is a precursor in the biosynthesis of acinetobactin, a siderophore essential for A. baumannii virulence. 14 Extracellular histidine concentration is lower than 2 μM in the lungs of mice, regardless of A. baumannii infection, 13 while the histidine inhibition constant for A. baumannii ATPPRT, the enzyme allosterically inhibited by the amino acid in a negative feedback control mechanism, lies between 83 and 282 μM. 15 As this value is expected to reflect the metabolite concentration the cell needs to function, 16 this may explain the reliance of this bacterium on histidine biosynthesis to establish and sustain pneumonia. 15 The need for novel antibiotics effective against carbapenemresistant A. baumannii was classified as a critical priority by the World Health Organisation. 17 Moreover, ventilator-associated pneumonia is one of the most common manifestations of A. baumannii infection, and it is linked to high mortality rates. 18 The identification and characterization of promising novel molecular targets are key steps to enable rational drug design. 19 Thus, the characterization of A. baumannii HisIE (AbHisIE) catalysis may lay the foundation for inhibitor design against this enzyme on the path toward novel antibiotics against A. baumannii.
Here, the gene encoding the putative AbHisIE was cloned and expressed, and the recombinant protein was purified. Liquid chromatography-mass spectrometry (LC−MS), differential scanning fluorimetry (DSF), steady-state and pre-steadystate kinetics, biocatalytic syntheses of PRAMP and N 1 -(5phospho-β-D-ribosyl)-ADP (PRADP), solvent deuterium kinetic isotope effects, and viscosity effects were used to characterize the enzyme and its catalyzed reaction. Furthermore, the HisE-like domain of AbHisIE (heretofore referred to as AbHisE domain ) was cloned and shown to catalyze the reaction normally catalyzed by monofunctional HisE, paving the way to study the pyrophosphohydrolysis reaction independently.
Purification of AbHisIE and AbHisE domain . All purification procedures were performed on ice or at 4°C using an ÄTKA Start FPLC system (GE Healthcare). All SDS-PAGE used a NuPAGE Bis-Tris 4−12% Precast Gel (ThermoFisher Scientific). For AbHisIE purification, cells were resuspended in buffer A (50 mM HEPES pH 7.5) supplemented with 0.2 mg mL −1 lysozyme, 750 U BaseMuncher endonuclease, and half a tablet of EDTA-free Complete protease inhibitor cocktail. Cells were lysed in a cell disruptor (Constant systems) at 30 kpsi, and centrifuged at 48,000 g for 30 min to remove cell debris. AbHisIE was precipitated by dropwise addition of 1.5 M ammonium sulfate in buffer A to the supernatant followed by stirring for 1 h. The sample was centrifuged at 48,000 g for 30 min, the supernatant was discarded, and the pellet was resuspended in buffer A, dialyzed against 3 × 2 L of buffer A, filtered through a 0.45 μm membrane and loaded onto a 10 mL HiTrap Q FF column pre-equilibrated with buffer A. The column was washed with 10 column volumes(CV) of 2.5% buffer B (50 mM HEPES pH 7.5, 2 M NaCl) and the adsorbed proteins were eluted with a 30 CV linear gradient of 2.5−15% buffer B. Fractions were analyzed by SDS-PAGE, and those containing AbHisIE were pooled and dialyzed against 2 × 2 L of buffer C (50 mM HEPES pH 8.0, 250 mM NaCl), filtered through a 0.45 μM membrane and loaded onto a ZnCl 2charged 5 mL HisTrap FF column pre-equilibrated with buffer C. The column was washed with 10 CV of buffer C and the adsorbed proteins were eluted with a 20 CV linear gradient of 0−15% buffer D (50 mM HEPES pH 8.0, 250 mM NaCl, 50 mM imidazole). Fractions were analyzed by SDS-PAGE and those containing AbHisIE were pooled, concentrated using a 10,000-MWCO ultrafiltration membrane, and loaded onto a HiPrep 26/60 Sephacryl S200 HR column equilibrated with buffer A. The column was washed with 1 CV of buffer A. Fractions were analyzed via SDS-PAGE and those containing AbHisIE were pooled, concentrated using a 10,000-MWCO ultrafiltration membrane, aliquoted, and stored at −80°C. The concentration of AbHisIE was determined spectrophotometrically (NanoDrop) using a theoretical extinction coefficient (ε 280 ) of 40,910 M −1 cm −1 (ProtParam tool − Expasy). The identity of the protein was confirmed via tryptic digest and LC−MS/MS analysis of the tryptic peptides performed by the University of St Andrews BSRC Proteomics and Mass Spectrometry facility.
For AbHisE domain purification, cells were resuspended in buffer A (50 mM HEPES pH 8.0, 500 mM NaCl, 10 mM imidazole) supplemented with 0.2 mg mL −1 lysozyme, 750 U BaseMuncher endonuclease, and half a tablet of EDTA-free Complete protease inhibitor cocktail. Cells were lysed in a cell disruptor (Constant systems) at 30 kpsi and centrifuged at 48,000 g for 30 min to remove cell debris. The supernatant was filtered through a 0.45 μm membrane and loaded onto a NiCl 2 -charged 5 mL HisTrap FF column pre-equilibrated with buffer A. The column was washed with 10 CV of buffer A, and adsorbed proteins were eluted with a 20 CV gradient of 0− 100% buffer B (50 mM HEPES pH 8.0, 500 mM NaCl, 500 mM imidazole). Fractions were analyzed by SDS-PAGE and those containing AbHisE domain were pooled, mixed with TEVP (1 mg of TEVP to 15 mg of AbHisE domain ) and dialyzed against 2 × 2 L of buffer C (20 mM HEPES pH 7.5, 150 mM NaCl, 10% glycerol (v/v), 2 mM DTT) and against 1 × 2 L of buffer A. Samples were filtered through a 0.45 μm membrane and loaded onto a 5 mL HisTrap FF pre-equilibrated with buffer A. The column was washed with 10 CV of buffer A, and the eluate was collected, analyzed by SDS-PAGE, concentrated using a 10,000-MWCO ultrafiltration membrane, dialyzed against 2 × 2 L of buffer D (20 mM HEPES pH 8.0), aliquoted, and stored at −80°C. The concentration of AbHisE domain was determined spectrophotometrically (Nano-Drop) using an ε 280 23 PRADP and PRATP were solubilized in water and loaded onto an Atlantis Premier BEH C 18 AX column (2.1 × 100 mm, 1.7 μm) on a Waters ACQUITY UPLC system coupled to a Xevo G2-XS QToF mass spectrometer equipped with an ESI source. The UPLC mobile phase was (A) 10 mM ammonium acetate pH 6, (B) acetonitrile, and(C) 10 mM ammonium acetate pH 10. The following sequence was applied: 0−0.5 min at 90% (A) and 10% (B); 0.5−2.5 min step change from 90% (A) and 10% (B) to 50% (A), 10% (B) and 40% (C); 2.5−7 min re-equilibration to 90% (A) and 10% (B), the flow rate of 0.3 mL min −1 . ESI data were acquired in negative mode with a capillary voltage of 2500 V. The source and desolvation gas temperatures of the mass spectrometer were 100 and 250°C, respectively. The cone gas flow was 50 L h −1 , whilst the gas flow was 600 L h −1 . A scan was performed between 50 and 1200 m/z. A lockspray signal was measured and a mass correction was applied by collecting every 10 s, averaging 3 scans of 0.5 s each, using Leucine Enkephalin as a correction factor for mass accuracy.
PRAMP was solubilized in water and loaded onto a Premier BEH C 18 column (2.1 × 50 mm, 1.7 μm) held at 40°C on a Waters Arc HPLC system coupled to a QDa mass detector equipped with an ESI source. PRAMP was eluted with an isocratic mobile phase of 0.1% formic acid, 1% acetonitrile in

pubs.acs.org/acscatalysis
Research Article water at a flow rate of 0.4 mL min −1 . MS scans were performed followed by the selection of the desired ion for PRAMP. ESI data were acquired in negative mode with a capillary voltage of 800 V both as a mass range (50−1250) scan (cone voltage of 30 V) and single-ion (558) recording (cone voltage ramp of 30−100 V). The source and probe temperatures of the mass spectrometer were 120 and 600°C, respectively. Detection of ProFAR by LC−MS. Reactions (500 μL) were prepared in 100 mM HEPES pH 7.5, 15 mM MgCl 2 , 16 μM AbHisIE, and 135 μM PRATP. Control reactions lacked AbHisIE. The reactions were incubated at room temperature for 1 h and passed through 10,000 MWCO Vivaspin centrifugal concentrators. ProFAR was detected exactly as described above for PRADP and PRATP detection by LC− MS.
DSF-Based Thermal Denaturation of AbHisIE and AbHisE domain . DSF measurements (λ ex = 490 nm, λ em = 610 nm) were performed in 96-well plates on a Stratagene Mx3005p instrument. Reactions (50 μL) contained either 8.3 μM AbHisIE (in the presence or absence of 50 μM PRATP) or 8.7 μM AbHisE domain in 100 mM HEPES pH 7.5 and 15 mM MgCl 2 . Invitrogen Sypro Orange (5×) was added to all wells. Controls lacked protein and were subtracted from the corresponding protein-containing samples. Thermal denaturation curves were recorded over a temperature range of 25−93°C with increments of 1°C min −1 . Three independent measurements were carried out.
AbHisIE Direct and Continuous Assay Detecting ProFAR. Typically, initial rates were measured at 25°C in 100 mM HEPES pH 7.5, 15 mM MgCl 2 , and 4 mM DTT, unless otherwise stated. Reactions (500 μL) were monitored for an increase in absorbance at 300 nm corresponding to the formation of ProFAR (Δε 300 = 6700 M −1 cm −1 at pH 7.5) 9 for 60 s in 1 cm path length quartz cuvettes (Hellma) in a Shimadzu UV-2600 spectrophotometer outfitted with a CPS unit for temperature control. Control reactions lacked AbHisIE. The effect of added Zn 2+ on activity was determined by measuring initial rates in the presence or absence of added 50 μM ZnCl 2 and 13.5 μM PRATP, 12 mM MgCl 2 , and 20 nM AbHisIE. The effect of added Mg 2+ on activity was determined by measuring the initial rates in the presence of 0− 24 mM MgCl 2 , 37 μM PRATP, and 20 nM AbHisIE. The enzyme concentration dependence of the initial rate was determined in 100 mM HEPES pH 7. Two independent measurements were carried out.
Determination of PRATP ε 300nm at Different pH Values. The ε 300 of PRATP at pH 7.0, 7.5, and 8.0 were determined by measuring the absorbance (NanoDrop) at 300 nm of known concentrations of PRATP in 100 mM HEPES (pH 7.0: 0.45, 0.91, and 1.8 mM; pH 7.5: 0.51, 1.0, and 2 mM; pH 8.0: 0.46, 0.92, and 1.8 mM). These known concentrations were determined independently via absorbance at 290 nm of a high-concentration stock solution of PRATP in 10 mM HEPES pH 7.5. This stock solution was in turn diluted into 100 mM HEPES at different pH values. The pH of the final PRATP solutions was measured to ensure that they stayed at pH 7.0, 7.5, and 8.0. Controls were measured by following the same dilution procedure in the absence of PRATP, and their absorbance at 300 nm was subtracted from each value with PRATP. The final values were then subtracted from the ε 300 of ProFAR (8000 M −1 cm −1 ). 9 to generate Δε 300 . Three independent measurements were carried out for each concentration at each pH.
AbHisIE Analysis of Kinetics Data. Kinetics data were analyzed by the non-linear regression function of SigmaPlot 14.0 (SPSS Inc.). Data points and error bars represent mean ± SEM, and kinetic and equilibrium constants are given as mean ± fitting error. Substrate saturation curves were fitted to eq 1, solvent viscosity effects were fitted to eq 2, solvent deuterium kinetic isotope effects were fitted to eq 3, and inhibition data were fitted to eq 4. In eqs 1−4, v is the initial rate, k cat is the steadystate turnover number, K M is the Michaelis constant, E T is total enzyme concentration, S is the concentration of substrate, k 0 and k η are the rate constants in the absence and presence of glycerol, respectively, η rel is the relative viscosity of the solution, m is the slope, F i is the fraction of deuterium label, E kd cat /Kd M and E kd cat are the solvent isotope effect minus 1 on k cat / K M ( D2O (k cat /K M )), and k cat ( D2O k cat ) respectively, v i and v 0 are the initial rates in the presence and absence of inhibitor, respectively, IC 50 is the half-maximal inhibitory concentration, and h is the Hill coefficient. D2O (k cat /K M ) and D2O k cat at pL 7.0 and 7.5 were also calculated as the ratios of the relevant rate constants in H 2 O and D 2 ■ RESULTS AND DISCUSSION Purification, Biophysical, and Biochemical Characterization of AbHisIE and AbHisE domain . These results, including Figures S1−S14, are described and discussed in Supporting Information.

Steady-State Kinetic Parameters for AbHisIE with PRATP.
To uncover which of the two reactions, i.e., either pyrophosphohydrolase-catalyzed or cyclohydrolase-catalyzed, is rate-limiting, the steady-state kinetic parameters were determined based on the detection of ProFAR and PP i . The AbHisIE overall reaction displayed Michaelis−Menten kinetics when PRATP concentration was varied and either ProFAR or PPi formation was measured (Figure 1). When ProFAR formation was measured, fitting the data to eq 1 yielded an apparent steady-state catalytic constant (k cat ProFAR ) and apparent Michaelis constant (K M ) shown in Figure 1 , but in the case of these monofunctional HisI, the substrate was PRAMP. 6,9 When PP i formation was monitored, data fitting to eq 1 yielded apparent k cat PPi and apparent K M as depicted in Figure 1, leading to k cat PPi /K M of (3.1 ± 0.4) × 10 6 M −1 s −1 . Assuming catalytically independent active sites in the HisEand HisI-like domains of AbHisIE, and PRAMP binding to the HisI-like domain in a bimolecular step, a minimum kinetic sequence depicting flux through the bifunctional enzyme reaction from PRATP can be summarized in Scheme 2, with the release of PRAMP and PP i from the HisE-like domain combined in one step (k 5 ) since the order (if any) of product release is not known. Besides the irreversibility of the two hydrolytic steps (k 3 and k 9 ), product release steps (k 5 and k 11 ) are presumed irreversible under initial-rate conditions in the absence of added products. 26 The makeup of the steady-state kinetic parameters shown in Figure 1 will differ depending on which product is being measured, even though only PRATP concentration is varied. When PP i is detected, k cat PPi , at saturating levels of PRATP, and k cat PPi /K M , at PRATP levels approaching zero, are defined by simple expressions as in eqs 5 and 6, respectively. On the other hand, when ProFAR is detected, the corresponding steady-state parameters are more complex. Increasing concentrations of PRATP must eventually lead to saturation of the HisI-like domain active site with PRAMP, as seen from the saturation kinetics obtained when the final product is monitored (Figure 1), and k cat ProFAR is given by eq 7. As PRATP levels approach zero, so do PRAMP levels, and k cat ProFAR /K M is defined by eq 8 (eqs 5−8 derived according to Cleland (Figure  2A), and the reaction followed Michaels−Menten kinetics, with data fitting to eq 1 yielding steady-state kinetic parameters displayed in Figure 2B. The similar k cat ProFAR values when either PRATP or PRAMP was the substrate suggest PRAMP can bind the HisI-like domain of AbHisIE from bulk solvent and the cyclohydrolase-catalyzed reaction limits the overall rate of the bifunctional catalytic cycle. The k cat ProFAR /K M for PRAMP of (4.4 ± 0.4) × 10 5 M −1 s −1 is 3-fold lower than the k cat ProFAR / K M for PRATP, which speaks against diffusion in and out of bulk solvent as the main path for PRAMP transfer from the first active site to the second, favoring the hypothesis that AbHisIE catalysis involves proximity channeling. 27,28 Solvent Viscosity Effects on AbHisIE-Catalyzed ProFAR Formation. In order to evaluate whether or not diffusional steps involving either substate binding or product release limit the rate of AbHisIE-catalyzed ProFAR formation from PRATP, solvent viscosity effects were determined by measuring reaction rates at different concentrations of the microviscogen glycerol ( Figure 3A and Table S1). Plots of k cat ProFAR /K M ratios versus relative viscosity ( Figure 3B) and k cat ProFAR ratios versus relative viscosity ( Figure 3C) produced slopes of 0 and −0.01, respectively. These data indicate neither PRATP and PRAMP binding to nor PRAMP, PP i , and ProFAR release from AbHisIE is rate-limiting in the reaction. 29 This contrasts with the first enzyme of histidine biosynthesis, ATPPRT, where significant solvent viscosity effects on k cat revealed diffusion of the product from the enzyme to be ratelimiting. 30 Solvent Deuterium Isotope Effects on AbHisIE-Catalyzed ProFAR Formation. Solvent deuterium isotope effects, defined as the ratio of rate constants (kinetic isotope effects) or equilibrium constants (equilibrium isotope effects) for reactions taking place in H 2 O and D 2 O, can inform on ratelimiting proton-transfer steps in enzymatic reactions. 31 To uncover potential rate-liming proton-transfer steps in the AbHisIE reaction, a solvent deuterium isotope effect study was undertaken. Because the presence of D 2 O can increase the pK a of kinetically relevant ionizable groups by ∼0.5, solvent isotope effects would ideally be measured in a pH-independent region of a pH-rate profile. 32 In the case of AbHisIE, this is further complicated by the fact that, while ProFAR absorbance is pH independent above pH 5, 33 PRATP and PRAMP absorbance at 300 nm is pH-dependent (PRATP and PRAMP have identical absorbance spectra in this region), 23 which would shift the Δε 300 from its value at pH 7.5. 9 Hence, Δε 300 was determined for the conversion of PRATP to ProFAR under Scheme 2. The Minimum Kinetic Sequence for the AbHisIE-catalyzed Reaction  Lines are best fit to eq 2.

ACS Catalysis pubs.acs.org/acscatalysis
Research Article initial-rate conditions by measuring ε 300 for PRATP at pH 7.0, 7.5, and 8.0 ( Figure S14) and subtracting each value from the ε 300 for ProFAR 9 (Table S2). Importantly, the Δε 300 for ProFAR at pH 7.5 obtained by this method (6690 M −1 cm −1 ) is within 0.15% of the published value, lending confidence to the approach. With PRATP as the substrate, both k cat ProFAR /K M and k cat ProFAR increased as the pL increased from 7.0 to 8.0, although k cat ProFAR /K M seemed to peak at pH 7.5 when the reaction took place in H 2 O (Figure 4), indicating that deprotonation of one or more groups increases the reaction rate.
Solvent deuterium isotope effect measurement at a plateau region of the pH-rate profile could not be accomplished here. Therefore, caution must be wielded to interpret the magnitudes of the isotope effects reported in Table 1, especially for D2O (k cat ProFAR /K M ) where the propagated experimental uncertainties were sizable owing to the uncertainties in the second-order rate constants obtained in D 2 O ( Figure 4C). As a trend, D2O (k cat ProFAR /K M ) and D2O k cat ProFAR decreased as the pL increased, suggesting proton-transfer steps accompany steps contributing less to the observed reaction rate as the pH increases. It should be noted that k cat ProFAR /K M and k cat ProFAR varied by a maximum of 4-and 3.4-fold, respectively, across the pH range. This is particularly relevant for D2O (k cat ProFAR /K M ) and D2O k cat ProFAR at pL 7.5, which are higher than what could be accounted for by pL changes alone.
The reactions taking place in the pyrophosphohydrolase and cyclohydrolase active sites are proposed to involve nucleophilic attack by Mg 2+ -and Zn 2+ -activated water molecules, respectively, 5,7,10 and it is reasonable to assume water coordination by the metal will be in equilibrium in the free AbHisIE. As Mg 2+ -activated 34,35 and Zn 2+ -activated 32 H 2 O/ D 2 O can have inverse fractionation factors (ϕ M-OL ∼ 0.7−0.9), the observed D2O (k cat ProFAR /K M ) will be equal to the product of an inverse equilibrium solvent isotope effect ( D2O K eq < 1) and any subsequent solvent kinetic isotope effects. This means the normal kinetic isotope effect portion of D2O (k cat ProFAR /K M ) has a larger value than what was observed, for instance, at pH 7.0 and 7.5, likely the result of more than one proton in flight during a rate-limiting step for k cat ProFAR /K M . At pH 8.0, D2O (k cat ProFAR /K M ) becomes modestly inverse, probably reflecting an inverse D2O K eq on metal-water coordination once a slow proton-transfer step at lower pH becomes fast at this higher pH. The proposed catalytic mechanism for cyclodrolysis of PRAMP 10 is reminiscent of that proposed for the Zn 2+dependent metalloenzyme AMP deaminase, where an inverse D2O (k cat ProFAR /K M ) of ∼0.7 was also reported, followed by a proton inventory implicating at least two proton transfers in a rapid equilibrium step involving Zn 2+ -water coordination. 36 While D2O k cat ProFAR decreases at pH 8.0, it remains normal and significant, indicating protonation steps reporting on k cat ProFAR / K M and k cat ProFAR are separated by an irreversible step. Importantly, at all pHs tested, at least one proton is in flight during the rate-limiting step for k cat ProFAR , which our results indicate has larger contribution from the cyclohydrolysis reaction. At pH 7.5, when PRAMP was employed as a substrate to bypass the pyrophospholysis reaction, the D2O (k cat ProFAR /K M ) was only 1.4 ± 0.1 ( Figure 4E). This suggests a large portion of the D2O (k cat ProFAR /K M ) observed with PRATP as a substrate reports on PRATP pyrophosphohydrolysis. Assuming, hypothetically, the Zn 2+ -bound water molecule responsible for the cyclohydrolysis of PRAMP would induce a D2O K eq of ∼0.7 (based on common fractionation factors attributed to Zn 2+bound water), 32 the kinetic isotope effect portion of the D2O (k cat ProFAR /K M ) with PRAMP as a substrate would have a magnitude of ∼2. 32 This also suggests a D2O (k cat ProFAR /K M ) of ∼8.5 originating in the PRATP pyrophosphohydrolysis reaction, probably involving more than one proton in flight. D2O k cat ProFAR was 2.1 ± 0.1 ( Figure 4E) with PRAMP as a substrate. This suggests at least one proton is in flight during the rate-limiting step for k cat ProFAR from the AbHisIE:PRAMP  To uncover additional information on rate-limiting steps of AbHisIE-catalyzed ProFAR synthesis from PRATP, the approach to steady state was monitored upon rapid mixing of AbHisIE and PRATP at pH 7.5 ( Figure 5). The curves could not be fitted to an equation because, even though PRATP concentration was in 5-fold excess to enzyme concentration, PRAMP concentration was not, as it was being formed in situ. Qualitatively, the data are interpreted as follows. As PRATP concentrations used were at least 10-fold the K M obtained when PP i formation was assayed, the lag phase, which is shorter at the higher enzyme concentration, reflects mostly the PRAMP formation rate from the nearly saturated HisE active site of AbHisIE, with no appreciable formation of ProFAR. As ProFAR production progresses, eventually the HisI-like active site is nearly saturated by PRAMP, leading to ProFAR formation resembling a burst that precedes the steady-state reaction. Supporting this interpretation, linear regressions of the linear phases yielded apparent steady-state rate constants of 4.2 ± 0.1 and 4.16 ± 0.08 s −1 , in reasonable agreement with k cat ProFAR . Furthermore, the y-axis intercepts of the linear regressions indicating the concentrations of on-enzyme ProFAR formed in the burst phase, 4.8 ± 0.2 and 8.2 ± 0.3 μM, approach the corresponding AbHisIE concentrations. However, the k cat PPi of 8.3 s −1 would allow only ∼0.75 turnovers in ∼0.09 s, the apparent time required to saturate the HisI active site with PRAMP ( Figure 5). Even if all AbHisIE is bound to PRATP, only ∼3.75 and ∼7.5 μM of free PRAMP would be produced from 5 and 10 μM AbHisIE, respectively, in 0.09 s, concentrations which are below the PRAMP K M of 11 μM. This suggests the preferred pathway for the transfer of PRAMP from the pyrophosphohydrolase domain to the cyclohydolase domain avoids significant diffusion into bulk solvent. Too short a lag time in consecutive reactions to allow the intermediate to accumulate enough into bulk solvent before rebinding to the next active site has been invoked as characteristic of substrate channeling. 27 A presteady-state burst was also observed in ATPPRT catalysis, 30,37 and product release was shown to be rate-limiting based on solvent viscosity effects on k cat . 30 PRADP Inhibits AbHisIE-Catalyzed Pyrophosphorolysis. ADP can replace ATP as a substrate of ATPPRT, which generates PRADP. 30 AbHisIE, however, failed to produce ProFAR, PP i , or P i when PRADP replaced PRATP as a substrate. As PRADP is a close structural analogue of both PRAMP and PRATP, we tested whether it might act as an AbHisIE inhibitor. PRADP inhibited AbHisIE-catalyzed ProFAR formation from PRATP in a dose-dependent manner ( Figure 6A), and data fitting to eq 4 yieded an IC 50 of 52 ± 4 μM. However, even 268 μM PRADP could not inhibit AbHisIE-catalyzed ProFAR formation from 20 μM PRAMP ( Figure 6B). This suggests that PRADP binds to the HisE-like domain of AbHisIE and inhibits pyrophosphohydrolysis of PRATP to PRAMP. The β-phosphate group of PRADP might prevent its binding to the HisI-like domain of AbHisIE, allowing ProFAR to form unincumbered from PRAMP directly. This provides further evidence of how independently the two active sites are able to operate and demonstrates the probable channeling of PRAMP does not involve a tunnel through the protein connecting the two active sites. 28 A protein tunnel shielded from bulk solvent was also disfavoured as a connection between the two domains based on crystal structures of HisIE orthologues, 7,8 and cannot be readily envisioned from our AlphaFold-based structural model ( Figure  S4A).
Implications for AbHisIE Catalysis. The presence of a burst preceding the steady state indicates a step following adenine ring-opening limits k cat ProFAR . While this is commonly interpreted as product release being rate-limiting, 30,38 the lack of solvent viscosity effects on k cat ProFAR rules out slow diffusion of ProFAR from AbHisIE. 29 Moreover, the sizable D2O k cat ProFAR means at least one proton transfer is associated with this step. 31,32 In addition, it is clear from steady-state and presteady-state kinetic analyses that k cat PPi is not large enough to allow PRAMP to accumulate into the bulk solvent at the levels required to saturate the cyclohydrolase active site, and some form of proximity channeling 28 must be the preferred means of PRAMP transfer. Based on the inhibition of the pyrophosphohydrolase activity but not the cyclohydrolase activity by PRADP, channeling does not involve tunneling through the protein, and probably still involves bimolecular binding of PRAMP to the cyclohydrolase active site being faster than diffusion into bulk solvent. Hence, the catalytic cycle depicted in Scheme 2 must be expanded to include preferential partition of newly synthesized PRAMP toward binding the HisI-like active site as opposed to diffusion into bulk water, and at least a slow unimolecular step (k 11 ) involving a proton transfer following adenine ring opening but preceding ProFAR release (Scheme 3). This might be, for instance, a proton-transfer-  linked conformational change that triggers ProFAR dissociation.
In this revised catalytic sequence, k cat ProFAR for ProFAR synthesis from PRATP is given by eq 9 (see Supporting Information for details). It should be pointed out that the rate constants in Schemes 2 and 3 do necessarily represent microscopic rate constants governing elementary steps, but potentially macroscopic rate constants 38 defining the minimum catalytic path from PRATP to ProFAR.
Metabolic advantages associated with channeling, such as increased flux through a biosynthetic pathway and protection of intermediates from the action of enzymes external to the pathway, 28 may have favored the evolution of bifunctionality in AbHisIE. It should be pointed out, however, that simple fusion of enzymes is neither sufficient nor required to ensure substrate channeling, as exemplified by the lack of channeling in the multifunctional AROM complex 39 and the presence of channeling in the monofunctional proteins constituting the purinosome. 40 Another potential advantage of gene fusion includes a fixed ratio of gene products for a set of consecutive reactions. 41 Any fitness advantage associated with a bifunctional HisIE must be organism-specific, since other bacteria, such as M. tuberculosis have separate genes encoding monofunctional HisE and HisI. 3,5 Future kinetic characterization of the AbHisE domain will elucidate how much, if any, of the catalytic ability of this domain is compromised by the loss of the HisI-like domain.
The inability of AbHisIE to utilize PRADP as a substrate for either of its reactions is somewhat surprising with regards to its pyrophosphohydrolase activity. Other members of the αhelical NTP pyrophosphohydrolase superfamily, to which the HisE-like domain of AbHisIE belongs, such as protozoan dUTPases, can efficiently hydrolyze both dUTP, releasing PP i , and dUDP, releasing P i , to dUMP. 42 Unlike what its threedimensional fold would predict, the pyrophosphohydrolase specificity of AbHisIE seems reminiscent of trimeric all-β dUTPases, which cannot hydrolyze dUDP. 43 In trimeric dUTPases, dUDP acts as a competitive inhibitor, sitting in the active site in the same orientation non-hydrolysable dUTP analogues do, and crystal structures of these enzymes in complex with dUDP shed light on how catalysis proceeds. 44,45 Given their structural similarities, PRADP presumably acts as a competitive inhibitor against PRATP, and it could prove useful for obtaining a crystal structure of AbHisIE, or other HisE enzymes, with a substrate analogue bound in the active site to furnish insight into the catalytic mechanism. ■ ASSOCIATED CONTENT
Additional results and discussion on the characterization of AbHisIE (PDF)