The oligomerization state of bacterial enzyme I (EI) determines EI's allosteric stimulation or competitive inhibition by α-ketoglutarate

The bacterial phosphotransferase system (PTS) is a signal transduction pathway that couples phosphoryl transfer to active sugar transport across the cell membrane. The PTS is initiated by phosphorylation of enzyme I (EI) by phosphoenolpyruvate (PEP). The EI phosphorylation state determines the phosphorylation states of all other PTS components and is thought to play a central role in the regulation of several metabolic pathways and to control the biology of bacterial cells at multiple levels, for example, affecting virulence and biofilm formation. Given the pivotal role of EI in bacterial metabolism, an improved understanding of the mechanisms controlling its activity could inform future strategies for bioengineering and antimicrobial design. Here, we report an enzymatic assay, based on Selective Optimized Flip Angle Short Transient (SOFAST) NMR experiments, to investigate the effect of the small-molecule metabolite α-ketoglutarate (αKG) on the kinetics of the EI-catalyzed phosphoryl transfer reaction. We show that at experimental conditions favoring the monomeric form of EI, αKG promotes dimerization and acts as an allosteric stimulator of the enzyme. However, when the oligomerization state of EI is shifted toward the dimeric species, αKG functions as a competitive inhibitor of EI. We developed a kinetic model that fully accounted for the experimental data and indicated that bacterial cells might use the observed interplay between allosteric stimulation and competitive inhibition of EI by αKG to respond to physiological fluctuations in the intracellular environment. We expect that the mechanism for regulating EI activity revealed here is common to several other oligomeric enzymes.

results in active sugar transport across the cell membrane (1)(2)(3). The PTS is initiated by phosphorylation of EI by the small molecule phosphoenolpyruvate (PEP). Phosphorylated EI transfers the phosphoryl group to the phosphocarrier protein HPr. Thereafter, the phosphoryl group is transferred to a sugar-specific enzyme II and finally to the incoming sugar (Fig.  1a). Recently, the small-molecule metabolite ␣-ketoglutarate (␣KG) was shown to act as a competitive inhibitor of EI (inhibition constant, K I ϭ ϳ2.2 mM) (4,5). The intracellular concentration of ␣KG varies considerably in response to a change in the availability of nitrogen source in the culturing medium (from 0.5 mM, in the presence of 10 mM NH 4 Cl, to 10 mM, in the absence of nitrogen source) (4). Thus, inhibition of EI by ␣KG has been proposed as a biochemical mechanism that links the uptake of sugars to the availability of nitrogen source (4,5). In addition to playing a primary role in coupling carbon and nitrogen metabolism in bacteria, the phosphorylation state of EI strictly controls the phosphorylation state of all other PTS components (6), which in turn regulates a large number of bacterial functions, including catabolic gene expression, virulence, biofilm formation, chemotaxis, potassium transport, and inducer exclusion, via phosphorylation-dependent protein-protein interactions (2). Therefore, EI is a central regulator of bacterial metabolism, and obtaining a comprehensive understanding of the mechanisms tuning its biological activity may suggest new strategies in bioengineering and antimicrobial design and might help elucidating the coupling between metabolic networks that controls the biology of all living cells.
EI is a multidomain protein comprising a N-terminal domain (EIN, residues 1-249) that contains the phosphorylation site (His 189 ) and the binding site for HPr and a C-terminal domain (EIC, residues 261-575) that is responsible for protein dimerization and contains the binding site for PEP and the competitive inhibitor ␣KG. The EIN and EIC domains are connected by a short helical linker (residues 250 -260) (1,7). EI undergoes a series of large-scale conformational rearrangements during its catalytic cycle (Fig. 1b), including: (i) a monomer-dimer transition (8), (ii) an expanded-to-compact conformational change within EIC (9), and (iii) an open-to-close transition describing a reorientation of EIN relative to EIC (10 -12). PEP binding to EIC shifts the conformational equilibria toward the catalytically competent dimer/compact/close form and activates the enzyme for catalysis (Fig. 1b) (11). The mono-This work was supported by funds from the Roy J. Carver Charitable Trust and Iowa State University (to V. V.) and by the Intramural Research Program of the NIDDK, National Institutes of Health (to R. G.). The authors declare that they have no conflicts of interest with the contents of this article. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. 1  cro ARTICLE mer-dimer equilibrium of EI has been often suggested as a major regulatory element for PTS because (i) only dimeric EI can be phosphorylated by PEP (13): (ii) the interaction of the enzyme with its physiological ligands Mg 2ϩ and PEP (Michaelis constant, K m , measured in the presence of 4 mM Mg 2ϩ was ϳ300 M) decreases the equilibrium dissociation constant for dimerization (K D ) by more than 10-fold (from ϳ5 to Ͻ0.1 M) (5,8); and (iii) the intracellular concentrations of EI and PEP were reported to vary substantially depending on the experimental conditions (from ϳ30 to ϳ300 M for PEP and from ϳ1 to ϳ10 M for EI) (14 -16).
Here, we develop a flexible enzymatic assay to investigate the effect of perturbations of the monomer-dimer equilibrium of Escherichia coli EI on the activity of ␣KG against the enzyme. We show that at physiological concentrations of EI and PEP that promote dimerization of EI ([EI] Ͼ K D , [PEP] Ͼ K m ), ␣KG acts as a competitive inhibitor of EI. In contrast, at physiological conditions favoring the monomeric form of the enzyme ([EI] Ͻ K D , [PEP] Ͻ K m ), ␣KG allosterically stimulates EI autophosphorylation. To our knowledge, this is one of the few examples of a small molecule metabolite being reported to both inhibit and stimulate the activity of the same enzyme depending on the experimental conditions (the other known case is ATP that can be a substrate or an allosteric inhibitor of phosphofructokinase) (17). The fact that the intracellular concentrations of EI, PEP, and ␣KG are modulated by the composition of the culturing medium (4, 14 -16) suggests that this interplay between allosteric stimulation and competitive inhibition of EI might be used by bacterial cells to regulate the phosphorylation state of PTS in response to a change in the extracellular environment.

Effect of PEP and ␣KG on the monomer-dimer equilibrium of EI
The effect of the EI ligands, PEP and ␣KG, on the monomerdimer equilibrium of the enzyme was investigated by analytical ultracentrifugation (AUC). The sedimentation velocity data indicate that the monomer-dimer equilibrium of EI is shifted toward the monomeric species at concentrations of the enzyme of Ͻ1 M (Fig. 2a) and that addition of PEP or ␣KG results in a substantial stabilization of the dimeric state ( Fig. 2, b and c). Our results are consistent with the more than 10-fold decrease in dimerization K D reported previously for EI upon addition of PEP or ␣KG (5,8).

Kinetics of the phosphoryl transfer reaction
The addition of 10 mM PEP to a NMR sample containing 1 mM 15 N-labeled E. coli HPr and ϳ0.05 M E. coli EI (unlabeled) results in substantial chemical shift perturbations for the 1 H-15 N transverse relaxation optimized spectroscopy (TROSY) (18) peaks originating from HPr residues located in the vicinity of the phosphorylation site (His 15 ; Fig. 3, a and c). As previously noted, HPr does not interact directly with PEP, nor can it be phosphorylated in the absence of EI (19). Therefore, the observed spectral changes are attributed to HPr phosphorylation via EI. After 24 h of incubation at 37°C, the HPr spectrum relaxes back to the unphosphorylated form ( Fig. 3a), which is consistent with the low thermodynamic stability of phosphorylated histidine residues (20).
Here, we use 1 H-15 N SOFAST-TROSY spectra (21) to monitor the time evolution of the phosphoryl transfer reaction from PEP to HPr via EI. SOFAST NMR experiments are ideally suited for real-time investigations on reaction kinetics, because they allow acquisition of 2D NMR spectra within seconds (21). For this particular case, ϳ0.05 M unlabeled EI and 1 mM 15 Nlabeled HPr are mixed in 500 l of reaction buffer (see "Experimental procedures") and incubated at 37°C for 30 min in a conventional 5-mm NMR tube. Thereafter, the reaction is started by addition of the desired amount of PEP (note that the PEP stock solution is preincubated at 37°C). The sample is mixed in the NMR tube and equilibrated at 37°C for 1 min in the NMR magnet. The reaction is then monitored for 20 min by running a series of 2D 1 H-15 N SOFAST-TROSY spectra (1 min each). The phosphoryl transfer reaction is slow on the chemical-shift time scale, and distinct NMR peaks are observed for the phosphorylated and unphosphorylated species (Fig. 3b). To monitor the evolution of the phosphoryl transfer reaction, we have used the NMR peak intensities of residues Ala 10 , Gly 13 , and Gly 54 because they are characterized by high signal-tonoise ratio and are well resolved throughout the experiment ␣KG binding regulates EI of the bacterial PTS ( Fig. 3b). Because the early time points are more important in determining the initial rate of the reaction, we limited our analysis to the disappearance of the unphosphorylated species, for which NMR peaks with high signal-to-noise ratios are obtained at the beginning of the phosphoryl transfer reaction (note the phosphorylated HPr peaks are not present at time zero; Fig. 3b). Signal intensities are plotted versus time, and the linear portion of the decay is fit to obtain the initial rate of change (Fig. 3d). To convert the reaction rate from change in signal intensity over time to change in concentration of unphosphorylated HPr over time, the NMR signal intensities at time zero for Ala 10 , Gly 13 , and Gly 54 were obtained by extrapolation (Fig. 3d) and considered to correspond to the expected signal intensity for a 1 mM HPr sample. Unphosphorylated HPr concentration at any time point is reported as the average over the three analyzed peaks (Fig. 3e).
To evaluate the effect of an increased concentration of dimeric EI on the activity of the enzyme, enzyme kinetic data were collected at a fixed concentration of EI (ϳ0.05 M), PEP (1 mM), and HPr (1 mM), and with increasing concentration of the

␣KG binding regulates EI of the bacterial PTS
inactive EI mutant H189Q (EI Q ). His 189 is located within the N-terminal domain of the enzyme and does not participate in the dimer interface or in PEP/␣KG binding to EIC. Indeed, EI Q has been recently reported to have the same equilibrium dissociation constant for dimerization and to form an identical EIC dimer as the wildtype EI (11,22). Therefore, EI Q cannot receive the phosphoryl group from PEP but can still interact with the wildtype protein (EI WT ) to form an active EI dimer. As expected, increasing the concentration of EI Q from 0 to 10 M doubles the HPr phosphorylation rate measured by our NMR assay (Fig. 4a). It is worth noticing that EI Q is inactive in the absence of EI WT (Fig. 4a). Therefore, the increased enzymatic activity observed by adding EI Q to a sample with a low concentration of EI WT (ϳ0.05 M) is due to an increased population of dimeric EI (which goes from 8% in the absence of EI Q to 80% in the presence of 10 M EI Q ) and not to the eventual presence of EI WT contaminations in purified EI Q .  Fig. 4b). It is worth noticing that increasing the concentration of PEP beyond 1.3 mM makes the phosphoryl transfer reaction too fast to be monitored by our method at our experimen-tal conditions (37°C and ϳ0.05 M enzyme). Therefore, k phosp cannot be accurately determined by the available data. However, our fitted results (k phosp Ͼ 10,000 s Ϫ1 ) are in good agreement with the fast conversion rates previously reported for the EI autophosphorylation reaction (24).

Effect of ␣KG on the activity of EI
The data reported in the previous sections indicate that dimerization stimulates the phosphoryl transfer activity of EI (Fig. 4a) and that increasing the concentration of ␣KG from 0 to 20 mM shifts the monomer-dimer equilibrium toward the enzymatically active EI dimer (Fig. 2). In this section, we evaluate the effect of ␣KG on the phosphoryl transfer activity of EI at experimental conditions that promote the monomeric or dimeric form of the enzyme.
At low concentration of enzyme (ϽK D,free ) and substrate (ϽK m ), we expect EI to exist predominantly as a monomer. In this case, the addition of small concentrations of ␣KG (ϽK I ) will act synergistically with PEP in saturating the binding sites on EI (Fig. 5a). The increased population of EI-ligand adducts will result in stabilization of the enzymatically active EI dimer and allosteric stimulation of the phosphoryl transfer reaction (Fig. 5a). In contrast, increasing the concentration of ␣KG to values larger than K I will result in oversaturation of the binding sites on EI and consequential competitive inhibition of enzymatic activity (Fig. 5a). Indeed, enzyme kinetic data collected at ϳ0.05 M EI, 200 M PEP, and increasing concentrations of ␣KG (0 -10 mM) show an initial stimulation of enzymatic activity followed by a decrease in the rate of phosphoryl transfer at high concentration of ␣KG (Ͼ 2 mM; Fig. 6a). At concentrations of EI Ͼ K D,free and/or concentrations of PEP Ͼ K m , we expect EI to exists predominantly as a dimer, and ␣KG to act exclusively as an inhibitor of the enzyme (Fig. 5b-d). Experimental data collected at ϳ0.05 M EI and 1000 M PEP (Fig.  6b), at 10 M EI and 200 M PEP (Fig. 6c), and at 10 M EI and 1000 M PEP (Fig. 6d) confirm the expected behavior. Interestingly all kinetic data reported in Fig. 6 can be fit considering that (i) only dimeric EI can catalyze the phosphoryl transfer reaction (13), (ii) saturation of the EI dimer-binding sites with PEP and/or ␣KG (dissociation constants K m and K I , respectively) decreases the K D for EI dimerization, and (iii) binding of PEP or ␣KG to one monomeric subunit affects the K D for EI dimerization to a minor extent. As done in the previous section when

␣KG binding regulates EI of the bacterial PTS
fitting the dependence of the phosphoryl transfer reaction on the concentration of enzyme, the model has been simplified by setting the dissociation constant of the EI dimer occupied by a single ligand molecule to K D,free (see Equations 21-37 under "Experimental procedures"). Fits were performed by keeping K m , K I , and K D,free to their measured values (300, 2200, and 1 M, respectively) (5), and optimizing values for K D,bound and k phosp . In all cases, a K D,bound of Ͻ10 Ϫ7 M was obtained. The kinetic model summarized by Equations 21-37 was used to simulate the effect of physiological fluctuations in the intracellular environment on the activity of ␣KG against EI (Fig. 7). In this simulation, K m and K I were set to the literature values for   (14, 15), and 0 -10 mM (4) range, respectively. K D,free is strongly affected by the presence of divalent cations in the buffer (8). Therefore, K D,free was set to 5 or 1 M (8) to simulate low (0.1 mM) or high (4 mM) intracellular concentration of free Mg 2ϩ , respectively. Our simulation (Fig.  7) suggests that ␣KG binding can provide up to 1.5 times stimulation of EI activity at physiological conditions that promote the monomeric form of the enzyme (low concentrations of EI, PEP, and Mg 2ϩ ) but results in strong inhibition of enzymatic activity at physiological concentrations of EI, PEP, and Mg 2ϩ that stabilize the EI dimer.

Discussion
In this work, we describe a novel method based on fast NMR techniques to assay the activity of EI under a wide range of experimental conditions. Previously reported methods to assay the activity of EI required quantification by mass spectrometry of pyruvate (formed as a by-product of the phosphoryl transfer reaction) (4) or quantification of phosphohistidine containing proteins (either EI or some other PTS component) by radioactive labeling (24,25) or by using a recently developed antibody (26). Compared with these methods, our protocol allows for observation of the phosphoryl transfer reaction in real time, therefore reducing the number of reagents and experimental steps required by the assay. On the other hand, our approach does not allow to monitor multiple reactions (i.e. multiple substrate concentrations) simultaneously and can only be applied if a 10% (or larger) reduction in NMR signal intensity is obtained for unphosphorylated HPr upon phosphorylation. This latter condition implies that phosphoryl transfer kinetics at concentrations of PEP lower than 100 M cannot be characterized accurately by our approach.
Using our NMR-based assay, we show that the small molecule metabolite ␣KG can act either as an allosteric stimulator or as a competitive inhibitor of EI depending on the oligomeric state of the enzyme (Figs. 5 and 6). Indeed, at experimental conditions favoring the dimeric form of EI, ␣KG inhibits the phosphoryl transfer activity of the enzyme (Fig. 6, b-d). In contrast, at experimental conditions favoring monomeric EI, addition of ␣KG results in a shift of the monomer-dimer equilibrium toward the enzymatically active dimeric form and a consequential stimulation of enzymatic activity (Fig. 6a). Interestingly, the intracellular concentration of EI was measured to be close to the equilibrium dissociation constant for protein dimerization (16), and the dimer K D of the free enzyme was shown to be affected substantially by varying the concentration of Mg 2ϩ in the experimental buffer (from 5 to 1 M moving from 0 to 4 mM Mg 2ϩ ) (8). In addition, the intracellular amount of PEP and ␣KG are close to the dissociation constants for PEP and ␣KG binding to the enzyme, respectively (4,14,15). In this scenario, small fluctuations in the intracellular concentrations of EI, Mg 2ϩ , PEP, and ␣KG induced by a change in the extracellular environment would drastically affect the activity of ␣KG on the PTS (Fig. 7). The PTS plays multiple regulatory functions in bacterial metabolism (including sugar uptake, virulence, biofilm formation, and chemotaxis) (1-3). These PTSmediated regulatory mechanisms are based either on direct phosphorylation of the target protein by one of the PTS components or on phosphorylation-dependent interactions (2). Therefore, the interplay between allosteric stimulation and competitive inhibition of EI by ␣KG revealed here may be required to tune the phosphorylation state of PTS in response to a change in the extracellular environment. Although the inhibitory activity of ␣KG on EI has been already proven to regulate the uptake of PTS sugars by bacterial cells in response to the availability of nitrogen source (4), understanding the effect of the weak stimulatory activity of ␣KG at low concentration of PEP on the biology of bacterial cells will require further investigations. Finally, this work shows how the activity of small molecule metabolites against their biological targets can change significantly in response to small changes in experimental conditions and illustrates that the dependence of the oligomeric state of the enzyme on the experimental conditions must be considered with great care when interpreting enzyme kinetic data.

Protein expression and purification
Uniformly 15 N-labeled E. coli HPr was expressed and purified as previously described (27). The H189Q (EI Q ) mutant of E. coli EI was created using the QuikChange site-directed mutagenesis kit (Stratagene). Genes for EI and EI Q were cloned into a pET-15b vector (Novagen) incorporating a N-terminal His tag. The plasmid was introduced into E. coli strain BL21star(DE3) (Invitrogen), and the transformed bacteria were plated onto an LB-agar plate containing ampicillin (100 g/ml) for selection. Cells were grown at 37°C in LB medium. At A 600 of ϳ0.4, the temperature was reduced to 20°C, and expression was induced with 1 mM isopropyl-D-thiogalactopyranoside. The cells were harvested by centrifugation (4,000 ϫ g for 30 min) after 16 h of induction, and the pellet was resuspended in 20 ml of 20 mM Tris, pH 8.0 (buffer A). The suspension was ␣KG binding regulates EI of the bacterial PTS lysed using a microfluidizer and centrifuged at 40,000 ϫ g for 40 min. The supernatant was filtrated through a 0.45-m filter membrane to remove cell debris and applied to a His affinity column (GE Healthcare). After the sample was loaded, the column was washed with buffer B (buffer A containing 20 mM imidazole), and the target protein was eluted with buffer C (buffer A containing 300 mM imidazole). The fractions containing the protein were confirmed by SDS-polyacrylamide gel electrophoresis and farther purified by gel filtration on a Superdex 200 column (GE Healthcare) equilibrated with 20 mM Tris, pH 7.4, 200 mM NaCl, 2 mM DTT, and 1 mM EDTA. Relevant fractions were loaded on an EnrichQ anion exchange column (Bio-Rad), and the protein was eluted with a 400-ml gradient from 150 mM to 400 mM NaCl.

Analytical ultracentrifugation
Sedimentation velocity experiments were carried out on a Beckman Coulter ProteomeLab XL-I analytical ultracentrifuge at 50 kilo-revolutions per minute and 20°C following standard protocols (28). A 2.0 mM stock solution of EI was diluted 50-fold in 100 mM NaCl, 20 mM Tris buffer, pH 7.4, 2 mM DTT, and 1 mM EDTA (buffer A) and used to prepare a series of solutions ranging from ϳ1 to 40 M by serial dilution. Samples were loaded into two-channel epon centerpiece cells (12-or 3-mm path length depending on the concentration). Absorbance (280 nm) and Rayleigh interference (655 nm) scans were collected, time-corrected (29), and analyzed in SEDFIT 15.01c (30) in terms a continuous c(s) distribution covering an s range of 0.0 -10.0 S with a resolution of 200 and a maximum entropy regularization confidence level of 0.68. Good fits were obtained with root mean square deviation values corresponding to typical instrumental noise values. Identical experiments were carried out in buffer A containing 20 mM PEP (buffer B) or 20 mM ␣KG (buffer C). Weighted-average sedimentation coefficients obtained by integration of the c(s) distributions for EI in buffer A were used to create an isotherm that was analyzed in SEPD-PHAT 13.0a in terms of a reversible monomer-dimer equilibrium to obtain a K d of 1 M, which is consistent with previous investigations of the EI monomer-dimer equilibrium (5,8). The solution density () and viscosity () for buffer A were calculated based on the solvent composition using SEDNTERP (31). Solution densities for buffers B and C were measured at 20°C on an Anton-Paar DMA 5000 density meter; solution viscosities were measured at 20°C using an Anton-Paar AMVn rolling ball viscometer. The partial specific volume (v) and absorption extinction coefficient for EI were calculated in SEDNTERP (31) based on the amino acid composition. The corresponding interference signal increment (32) was calculated in SEDFIT15.01c (30).

Enzyme kinetic assay
The ability of EI to transfer the phosphoryl group from PEP to HPr was assayed at 37°C using fast NMR methods (21) as described under "Results." NMR spectra were recorded on a Bruker 700 MHz spectrometer equipped with a z-shielded gradient triple resonance cryoprobe. The spectra were processed using NMRPipe (33) and analyzed using the program SPARKY (http://www.cgl.ucsf.edu/home/sparky). 3 The 1 H-15 N correlation spectrum of unphosphorylated HPr was assigned according to previously reported chemical shift tables (34). Composition of the reaction buffer was as follow: 20 mM Tris, pH 7.4, 100 mM NaCl, 4 mM MgCl 2 , 2 mM DTT, 1 mM EDTA, and 95% H 2 O/5% D 2 O (v/v). Unless stated otherwise, all enzymatic assays were run in a reaction volume of 500 l and at fixed concentrations of wildtype EI (ϳ0.05 M) and HPr (1 mM). The assays were run in triplicate. The initial velocities for the phosphoryl transfer reaction in the presence of different amount of EI Q (see "Results" and "Discussion") were fit in DynaFit 4.0 (23) using the following kinetic model, where E is the wildtype enzyme (EI WT ), Q is the concentration of EI Q , S is the substrate (PEP), ES is the EI WT -PEP complex, QS is the EI Q -PEP complex, EQ is the mixed EI WT E Q dimer, EQS is the mixed dimer with PEP bound to the EI Q subunit, ESQ is the mixed dimer with PEP bound to the EI WT subunit, ESQS is the mixed dimer with two PEP molecules, P is the product, K D,free (1 M) is the dimer dissociation constant for free EI, K D,bound (fitted) is the dimer dissociation constant for EI when saturated with ligands, K m (300 M) is the Michaelis constant for the EI-PEP interaction, k phosp (fitted) is the rate constant for the phosphoryl transfer interaction, [dharrow] indicates a thermodynamic equilibrium, and 3 indicates the unidirectional chemical step. Note that given the small amount of EI WT compared with EI Q , the amount of EI WT EI WT dimer is considered to be negligible in this model.
The initial velocities for the phosphoryl transfer reaction in the presence of different amount of PEP (see "Results" and "Discussion") were fit in DynaFit 4.0 (23) using the following kinetic model, E ϩ E N E 2 K D,free (Eq. 14) 3 Please note that the JBC is not responsible for the long-term archiving and maintenance of this site or any other third party hosted site. where E 2 is the EI WT EI WT dimer, E 2 S is the EI dimer complexed to one molecule of PEP, and E 2 S 2 is the dimer complexed with two molecules of PEP. Enzyme kinetic data measured at different concentration of ␣KG were fit in DynaFit 4.0 (23) using the following kinetic model, where I is the inhibitor (␣KG), EI is the EI-␣KG complex, E 2 I is the EI dimer complexed with one ␣KG molecule, E 2 I 2 is the EI dimer complexed with two ␣KG molecules, E 2 SI is the EI dimer complexed with one ␣KG molecule and one PEP molecule, and K I (2.2 mM) is the dissociation constant for free EI-␣KG interaction. In the fits, the concentration of EI is considered to be the sum of the active (EI WT ) and inactive (EI Q ) species.