Conformational Selection and Substrate Binding Regulate the Monomer/Dimer Equilibrium of the C-terminal domain of Escherichia coli Enzyme I*

Background: Conformational changes in the EIC domain of enzyme I upon ligand binding are thought to regulate the phosphotransfer system by modulating the monomer/dimer equilibrium. Results: Binding of phosphoenolpyruvate shifts a preexisting conformational equilibrium in EIC. Conclusion: Conformational selection provides a direct structural link between ligand binding and dimer affinity. Significance: Isolated EIC is an optimal system for investigating dynamic processes regulating EI. 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 the binding of phosphoenolpyruvate (PEP) to the C-terminal domain (EIC) of enzyme I (EI), a highly conserved protein that is common to all sugar branches of the PTS. EIC exists in a dynamic monomer/dimer equilibrium that is modulated by ligand binding and is thought to regulate the overall PTS. Isolation of EIC has proven challenging, and conformational dynamics within the EIC domain during the catalytic cycle are still largely unknown. Here, we present a robust protocol for expression and purification of recombinant EIC from Escherichia coli and show that isolated EIC is capable of hydrolyzing PEP. NMR analysis and residual dipolar coupling measurements indicate that the isolated EIC domain in solution adopts a stable tertiary fold and quaternary structure that is consistent with previously reported crystallographic data. NMR relaxation dispersion measurements indicate that residues around the PEP binding site and in the β3α3 turn (residues 333–366), which is located at the dimer interface, undergo a rapid transition on the sub-millisecond time scale (with an exchange rate constant of ∼1500 s−1) between major open (∼97%) and minor closed (∼3%) conformations. Upon PEP binding, the β3α3 turn is effectively locked in the closed state by the formation of salt bridges between the phosphate group of PEP and the side chains of Lys340 and Arg358, thereby stabilizing the dimer.

transduction pathway that couples phosphoryl transfer through a series of bimolecular protein-protein complexes to sugar transport across the membrane (1,2). The PTS is also involved in the regulation of diverse cellular processes, including transcription, chemotaxis, and glycolysis (2). There are four sugar-specific branches of the PTS, but all require EI to initiate the phosphoryl transfer cascade. EI is autophosphorylated by phosphoenolpyruvate (PEP) and subsequently donates the phosphoryl group to the histidine phosphocarrier protein HPr (3,4). EI and HPr are common to all branches of the PTS. Thereafter, the phosphoryl group is transferred from HPr to the sugar-specific enzymes II and ultimately onto the incoming sugar. It has been shown recently that, under conditions of nitrogen limitation, EI is inhibited by ␣-ketoglutarate (5), the carbon substrate for ammonia assimilation, thereby providing a regulatory link between central carbon and nitrogen metabolism in bacteria. EI is ubiquitous in bacteria and does not have any eukaryotic counterparts. Thus, given the central role of EI in the control and regulation of bacterial metabolism, it is no surprise that EI has been described as an ideal pharmaceutical target for the identification of novel and highly specific antimicrobials (6).
The isolated EIN domain can transfer a phosphoryl group to HPr but only intact dimeric EI can be autophosphorylated by * This work was supported by the Intramural Program of the National Institutes of Health, NIDDK, and the Intramural AIDS Targeted Antiviral Program of the Office of the Director of the National Institutes of Health (to G. M. C.).
PEP (8, 9, 13, 19 -21). It has therefore been suggested that the monomer/dimer equilibrium for the EIC domain plays a central role in the regulation of the overall PTS (22). Indeed, the interaction of EI with the physiological ligands for the EIC domain, Mg 2ϩ and PEP, decreases the equilibrium dissociation constant for dimerization (K D ) by ϳ30-fold (from 4.8 to 0.15 M (14,22). Given the intracellular EI concentration of ϳ10 M subunits (23), binding of Mg 2ϩ and PEP to EI results in a significant shift in the monomer/dimer equilibrium and a concomitant substantial increase in the cellular population of dimeric EI (from ϳ60 to 90%).
There is now a wealth of structural studies on EI. The isolated EIN domain from Escherichia coli has been solved in its free form by crystallography (10) and NMR (11) as well as in its phosphorylated state (24) and bound to HPr (25) by NMR. In all three cases, the structure of the EIN domain and the relative orientation of the ␣ and ␣/␤ subdomains remain unchanged. There are three crystal structures of intact EI, two of free EI from Staphylococcus carnosus (26) and Staphylococcus aureus (16), and one of a trapped phosphorylated intermediate of E. coli EI bound to the inhibitor oxalate (15). In addition, there are two solution structures of E. coli EI, free and bound to HPr, determined by the combined use of NMR and x-ray scattering (17,18). The EI structures reveal large rigid body conformational transitions involving domain reorientation of EIN relative to EIC, as well as reorientation of the two subdomains of EIN relative to each other. These large-scale conformational changes permit transfer of the phosphoryl group from PEP bound to the EIC domain to His 189 located on EIN in the conformation found in the trapped phosphorylated intermediate (15), and subsequent transfer of the phosphoryl group on His 189 to HPr in the conformation found in the structures of free EI and the EI-HPr complex where the structure of the EIN domain is identical to that of the isolated EIN domain (17). In addition, crystal structures of the isolated EIC domain from Thermoanaerobacter tengcongensis in the free form (27) and in complexes with PEP and pyruvate (28) have been obtained. Although the structure of the EIC domain is the same in intact EI and the isolated EIC domain, spectroscopic and kinetic investigations have suggested that in solution the EIC domain may be present as an ensemble of different conformations that are not apparent in the crystal structures (14).
The recombinant E. coli EIC domain has been reported to be proteolytically unstable and difficult to isolate (14,29). Here, we present a robust protocol for expression and purification of recombinant E. coli EIC. We show that the purified protein adopts a single, stable fold in solution and is able to hydrolyze PEP into pyruvate and inorganic phosphate. Using relaxation dispersion NMR spectroscopy (30,31), we demonstrate the existence of a rapid local conformational transition on the submillisecond time scale between two states, a major open state and a minor closed state, involving residues in the vicinity of the active site. Chemical shift analysis indicates that the conformation of the minor species is similar to that of the PEP-bound state, indicating that PEP binding likely proceeds via conformational selection. Furthermore, in combination with the available crystal structures, the data provide a rationale for understanding dimer stabilization by PEP. The data presented here suggest that EIC represents a good model system for studying the combined effects of substrate binding and conformational dynamics on enzymatic regulation.

EXPERIMENTAL PROCEDURES
Protein Expression and Purification-Intact EI (residues 1-575) and the EIN domain (residues 1-249) were expressed and purified as described previously (17,24). The EIC domain (residues 261-575) was cloned into a pET11a vector (Novagen) without tags. 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 either Luria Bertani (LB) or minimal medium (with 15 NH 4 Cl and 13 C 6 -glucose as the sole nitrogen and carbon sources, respectively) in H 2 O or D 2 O, respectively. At A 600 ϳ 0.4 (for LB cultures) or A 600 ϳ 0.8 (for cultures in minimal medium), the temperature was reduced to 20°C, and expression was induced with 1 mM isopropyl-D-thiogalactopyranoside. Cells were harvested by centrifugation after 16 h of induction, and the pellet was resuspended in 20 ml of 20 mM Tris, pH 8.0, 2 mM dithiothreitol (DTT), 1 mM ethylenediaminetetraacetic acid (EDTA), and 1 mM phenylmethylsulfonyl fluoride (PMSF). The suspension was lysed using a microfluidizer and centrifuged at 70,000 ϫ g for 40 min. The supernatant was filtered and loaded onto a DEAE anion exchange column (20 ml; GE Healthcare), and the protein was eluted with a 400 ml gradient of 1 M NaCl. The fractions containing the protein were confirmed by SDS-PAGE and purified by gel filtration on a Superdex-75 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 a monoQ anion exchange column (GE Healthcare), and the protein was eluted with a 400-ml gradient from 150 mM to 400 mM NaCl.
NMR spectra were recorded at 37°C on Bruker 900 and 600 MHz spectrometers equipped with either a z-shielded gradient triple resonance cryoprobe or, for 31 P NMR, an x,y,z-shielded gradient quadruple resonance probe. Spectra were processed using NMRPipe (32) and analyzed using the program SPARKY. Sequential 1 H/ 15 N/ 13 C backbone assignment of the EIC domain was carried out using transverse relaxation optimized (TROSY) versions (33,34) 15 N chemical shift differences in ppm, respectively, between free and bound states. Backbone amide 1 D NH residual dipolar couplings (RDCs) were measured by taking the difference in 1 J NH scalar couplings in aligned and isotropic media (38). The alignment media employed was phage pf1 (14 mg/ml for EIC and 18 mg/ml for the EIC-PEP complex; ASLA Biotech) (39,40), and 1 J NH couplings were measured using the ARTSY (amide RDCs by TROSY spectroscopy) pulse scheme (41). Singular value decomposition (SVD) analysis of RDCs was carried out using Xplor-NIH (42).
Relaxation dispersion experiments were conducted using a pulse sequence that measures the exchange contribution for the TROSY component of the 15 N magnetization (43). Off-resonance effects and pulse imperfections were minimized using a four-pulse phase scheme (44). Carr-Purcell-Meinboom-Gill (CPMG) experiments were performed at 600 and 900 MHz with a fixed relaxation delay but a changing number of refocusing pulses to achieve different effective CPMG fields (45). The transverse relaxation periods were set to 40 and 20 ms at 600 and 900 MHz, respectively. The effective CPMG fields employed were 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000 Hz at 600 MHz, and 200, 400, 600, 800, and 1000 Hz at 900 MHz. The resulting relaxation dispersion curves were fit to a two-state exchange process using Equation 1 (46,47), where R 2 0 is the transverse 15 N R 2 in the absence of exchange; p B is the relative population of the minor populated state; ⌬ N is the 15 N chemical shift difference between the two states; k ex is the exchange rate constant (given by the sum of the forward and backward rate constants k AB and k BA for the transitions from the major species A to the minor species B and back; and cpmg is the CPMG field defined as 1 ⁄ 2 cp , where cp is the spacing between the centers of successive 180°refocusing pulses in the CPMG pulse sequence (45).
Activity Assay-Enzymatic activity of EI and EIC for the hydrolysis of PEP to inorganic phosphate and pyruvate were assayed spectrophotometrically at 37°C using the EnzChek phosphate assay kit (Invitrogen) and a Nanodrop 2000 spectrophotometer. The reaction mixtures were prepared as follows: 25 l of 20ϫ reaction buffer from the kit (1.0 M Tris-HCl, 20 mM MgCl 2 , pH 7.5, 2 mM sodium azide), 7.9 l of 5 M NaCl, x l PEP, y l enzyme, 362.1 -x -y l doubly distilled H 2 O. The mixtures were preincubated at 37°C for 10 min prior to the addition of either EI or EIC at a final concentration of ϳ5 M. For the following 4 h, 39.5-l aliquots of the reaction mixture were taken at regular intervals and incubated for 10 min with 10 l of 1 mM 2-amino-6-mercapto-7-methylpurine riboside (MESG) and 0.5 l of 10 units/ml purine nucleoside phosphorylase. In the presence of purine nucleoside phosphorylase, MESG quantitatively reacts with inorganic phosphate to generate ribose 1-phosphate and 2-amino-6-mercapto-7-methylpurine. Enzymatic conversion of MESG results in a spectrophotometric shift in maximum absorbance from 330 nm for MESG to 360 nm for 2-amino-6-mercapto-7-methylpurine.
Enzymatic assays for EI and EIC were run in duplicate, and the obtained A 360 values were converted to mol units using a calibration curve generated from the standard solution of inorganic phosphate provided by the kit. Initial velocities were determined from the linear portion of the progress curves and were fit using the Michaelis-Menten equation.

RESULTS AND DISCUSSION
Expression and Isolation of Recombinant E. coli EIC-Expression of recombinant E. coli EIC (residues 261-575), without addition of any affinity or solubility tag, is described in the "Experimental Procedures." Purification of EIC from the cell lysate was carried out using a three-step chromatography protocol: weak anion exchange, size exclusion, and a final purification step with a strong anion exchanger (for details, see "Experimental Procedures"). Although the first anion exchange step allowed separation of EIC from most of the nucleic acid and protein contaminants, size exclusion chromatography was the key for successful purification of a stable EIC sample. EIC elutes from a Superdex 75 size exclusion column as two partially resolved peaks that have retention volumes of ϳ125 ml and ϳ150 ml, referred to as species 1 and 2, respectively, in Fig. 1A.
The final anion exchange step on species 1 reveals the presence of at least four similarly populated EIC isoforms, eluting between 250 -350 mM NaCl (Fig. 1B), and the 1 H-15 N TROSY correlation spectrum obtained on a combined sample of these isoforms shows very poor cross-peak dispersion (Fig. 1D). These data indicate that EIC from species 1 is largely unfolded.
In contrast the anion exchange profile for species 2 is dominated by a major peak eluting at ϳ300 mM NaCl (Fig. 1C), and the 1 H-15 N TROSY spectrum acquired for this species is highly dispersed (Fig. 1E), indicative of a stably folded protein. EIC purified from species 2 (note that a slower salt gradient centered at 300 mM NaCl was needed to completely remove contaminations arising from species 1; see "Experimental Procedures") is Ͼ90% pure (as determined by SDS-PAGE) and is stable for Ͼ3 weeks at 37°C in the presence of 4 mM Mg 2ϩ . Moreover, mutation of Met 469 , which is localized at the dimer interface, to the unnatural amino acid p-acetyl-L-phenylalanine (a construct not used in the present study) results in large chemical shift perturbations in the 1 H N / 15 N cross-peaks of opposing residues in the other subunit (48), and the overlay of the EI and EIC 1 H-15 N correlation spectra shows excellent agreement with only small differences in cross-peak positions (Fig. 1E). These data clearly indicate that, in the presence of 4 mM MgCl 2 and 100 mM NaCl, the isolated EIC domain is a dimer with the same fold as that in the full-length EI.
Solution Tertiary Structure of Free E. coli EIC-Crystallographic data show that E. coli EIC is a ϳ 70-kDa dimer exhibiting an (␣/␤) 8 -barrel fold with the active site Cys 502 located in the ␤7␣7 turn (15). Comparison of the x-ray structures of T. tengcongensis EIC free and bound to PEP indicates that substrate binding causes rearrangements of a few side chains (Arg 296 , Phe 354 , Gln 458 , Arg 465 , Met 466 , and Glu 504 ) in the active site but leaves the overall tertiary fold and quaternary structure unperturbed (28). By way of contrast, sedimentation velocity and thermal denaturation exper-iments have shown that the addition of PEP stabilizes the protein fold and decreases the equilibrium dissociation constant for dimerization (14). PEP binding also results in marked changes in the near UV CD and fluorescence spectra, suggesting that EIC undergoes conformational rearrangements upon PEP binding (14). These spectroscopic techniques, however, only provide averaged information on aromatic side chains, so that one cannot ascertain whether the observed spectral changes reflect global effects or local conformational transitions.
Here, we investigated the effect of PEP on the solution structure E. coli EIC by NMR. 1 H, 15 N, and 13 C backbone resonances of free EIC were assigned for 278 of 316 amino acids. The backbone / torsion angles derived from the backbone chemical shifts using the program TALOSϩ (49) are in excellent agreement with the x-ray structure of the trapped phosphorylated intermediate of E. coli EI ( Fig. 2A) (15). Thus, one can conclude that the tertiary structure of the isolated EIC domain in solution is essentially the same as that in the crystal structure of full-length EI. Relaxation Dispersion on Free EIC-To characterize the dynamics in the region of the PEP binding site, we carried out CPMG 15 N-relaxation dispersion experiments (31) at 600 and 900 MHz. These experiments probe exchange dynamics between species with distinct chemical shifts on a time scale ranging from ϳ50 s to 10 ms. Significant relaxation dispersion was observed for the backbone amides of a number of residues in the vicinity of the PEP binding site: Ile 426 and Gly 427 at the N-terminal end of strand ␤5; Arg 465 and Asn 467 located in the turn/loop (residues 453-477) connecting strand ␤6 to helix ␣6; and Gly 337 , Lys 340 , Glu 341 , Leu 355 , Trp 357 , Arg 358 , and Ala 359 located in the turn/loop (residues 333-366) connecting strand ␤3 to helix ␣3 (Fig. 2B). Three of the residues, Leu 355 , Trp 357 , and Asn 467 , are also located at the dimer interface (Fig. 2C). The 15 N-relaxation dispersion curves for all the above residues at both fields (600 and 900 MHz) were fit simultaneously to a two-state model (Equation 1 in "Experimental Procedures") describing the interconversion of two conformational states, optimizing the values of the exchange rate (k ex ), the fractional population of the minor state (p B ), and the residue-specific 15 N chemical shift differences between the two conformational states (⌬ N ). An example of the fits is provided in Fig. 2E, and a summary of the results is provided in Table 1. The population of the minor species is ϳ3%, and the overall exchange rate (sum of forward and backward rate constants) is ϳ1500 s Ϫ1 . These parameters yield values of the rate constants for the conversion from the major to the minor species and from the minor to the major species of ϳ50 and ϳ1500 s Ϫ1 , respectively. Consistent with the relaxation dispersion results, an overlay of the crystal structures of the EIC domain (15, 16, 26 -28) obtained to date (Fig. 2D) shows that the backbone of the Gly 337 -Tyr 344 segment of the ␤3␣3 turn can adopt one of two conformations: an open state observed in the crystal structures of both free EI (from S. aureus and S. carnosus) and the isolated T. tengcongensis EIC domain (free and bound to PEP or pyruvate) (cyan tubes in Fig. 2D); and a closed state found in the crystal structure of the trapped phosphorylated intermediate of E. coli EI (red tube in Fig. 2D). The closed state is stabilized by hydrogen bonds from the backbone amide groups of Gly 337 and Ala 359 to the backbone carbonyls Lys 340 and Ile 336 , respectively. Both of these hydrogen bonds are absent in the open state found in the crystal structure of S. aureus EI (16), and the Gly 337 3 Lys 340 hydrogen bond is missing in all crystal structures of the open state (16, 26 -28). Of note is that among the residues of the ␤3␣3 turn displaying relaxation dispersion, Gly 337 and Ala 359 show the largest ⌬ N values (Table 1), suggesting that the exchange process detected by the relaxation  (15). The data for loop regions (where conformational differences between solution and crystal are to be expected) are depicted as open circles, whereas the data for secondary structure elements are reported as filled circles. The overall correlation coefficient is 0.90, and the correlation coefficient computed using only the data from secondary structure elements is 0.95. deg., degrees. B, structural model for the E. coli EIC-PEP complex. The model was obtained by rigid-body least-square fitting of the backbone atoms of the x-ray structure of the EIC-PEP complex from T. tengcongensis (28) onto the EIC coordinates of the x-ray structure of phosphorylated EI from E. coli (15). The fitting was carried out using residues 261-333 and 367-570, and the C␣ rms difference is 1.1 Å. Ribbons of the two identical subunits are depicted in yellow and white. The PEP molecule is displayed as green bonds, and the side chains of Lys 340 , Arg 358 , and Arg 465 are shown as red bonds. The nitrogen atoms of amide groups exhibiting significant 15 N relaxation dispersion, characteristic of dynamics on the submillisecond to millisecond time scale, are shown as blue spheres. The inset provides a close-up of the ␤3␣3 turn. The turn is highlighted in cyan, the carbonyl groups of Ile 336 and Lys 340 are shown as spheres (gray for carbon, red for oxygen), and the amide groups of residues in this region displaying relaxation dispersion are shown as spheres, blue for nitrogen and white for the associated amide proton.  Table 1. dispersion experiments corresponds to the open/closed transition. The conformational dynamics detected in the ␤3␣3 turn extend all the way to the dimer interface, also affecting Arg 465 and Asn 467 , which directly face the ␤3␣3 turn in the other subunit (Fig. 2, B and C).
Effect of PEP on Spectrum of EIC-Addition of 10 mM PEP to EIC results in substantial changes in the 1 H-15 N TROSY spectrum (Fig. 3A). However, as we discuss in depth in a subsequent section, we noticed that formation of the EIC-PEP complex also results in degradation of PEP into pyruvate and inorganic phos-

population, and chemical shift parameters for conformational exchange between open and closed states of EIC derived from relaxation dispersion experiments
All of the relaxation dispersion curves at two fields (600 and 900 MHz) were fit simultaneously to a two-site exchange model, optimizing the values of two global parameters, the overall exchange rate (k ex ) and the population of the minor state (p B ), and the residue-specific ͉⌬ N ͉ values, which represent the absolute 15 N chemical shift difference between the major and minor states.  phate, making the lifetime of the EIC-PEP complex too short (Ͻ4 h at 37°C with 400 M EIC and 10 mM PEP) to allow acquisition of the three-dimensional triple resonance experiments necessary for sequential backbone resonance assignment. Thus, only those 1 H-15 N cross-peaks that could be tracked unambiguously in a PEP titration experiment were assigned in the EIC-PEP complex and used for further analysis. The assigned backbone amides for the EIC-PEP complex and the corresponding 1 H/ 15 N chemical shift perturbations (⌬ H/N ) arising from PEP binding are depicted on the EIC structure in Fig. 3B. Although the 1 H-15 N cross-peaks of residues directly facing PEP were completely broadened out after addition of ligand, large ⌬ H/N values were still observed for the flanking residues (Ile 275 , Gly 276 , Ile 426 , and Gly 427 ) suggesting that structural rearrangements are occurring at the active site (Fig. 3, B  and D). Moreover, the 1 H-15 N cross-peaks of residues in the ␤3␣3 turn experience extensive line broadening upon addition of PEP (Fig. 3C), indicating that the binding of the substrate affects the conformational dynamics detected in this area.
The perturbations on the 15 N chemical shifts (⌬␦ N ) of EIC observed upon PEP binding are in excellent agreement with the ͉⌬ N ͉ values obtained from the relaxation dispersion measurements on free EIC (Table 1), strongly suggesting that binding of PEP shifts the conformational equilibrium detected in free EIC from the open to the closed state. This is consistent with the model of the E. coli EIC-PEP complex (Fig. 2B), derived from the structures of E. coli phosphorylated EI (15) and the T. tengcongensis EIC-PEP complex (28), which shows that salt bridges between the phosphate group of PEP and the side chains of Lys 340 and Arg 358 lock the ␤3␣3 turn in the closed conformation, providing further stabilization to the Gly 337 3 Lys 340 and Ala 359 3 Ile 336 backbone hydrogen bonds. These key interactions, together with an additional salt bridge between the phosphoryl group of PEP and the guanadino group of Arg 465 (Fig.  2B), also serve to stabilize the backbone conformation of Leu 355 , Trp 357 , and Asn 467 in the closed state observed in the x-ray structure of phosphorylated EI. Because Leu 355 , Trp 357 , and Asn 467 are involved in several intersubunit contacts (Fig.  2C), the above observations explain the lower dimerization K D measured for EI in the presence of PEP (22), as well as the large ⌬ H/N perturbations induced by PEP for Glu 351 , Asn 352 , Gly 356 , and Trp 357 in the ␤3␣3 turn, and for Arg 465 , Asp 468 , Ile 470 , and Asn 475 at the dimer interface (Fig. 3, B and C). In addition, the ␤3␣3 turn is rich in aromatic residues (Tyr 344 , Phe 347 , Phe 354 , and Trp 357 ), so that conformational rearrangements in this region are fully consistent with the changes observed in the CD and fluorescence spectra of EIC upon PEP binding (14).
Effect of PEP on Quaternary Structure of EIC-To assess whether the local conformational changes in the ␤3␣3 turn observed in EIC upon PEP binding affect the subunit orientation in the EIC dimer, backbone amide ( 1 D NH ) RDCs for well resolved 1 H-15 N cross-peaks were measured on samples of weakly aligned EIC and the EIC-PEP complex in a dilute liquid crystalline medium of phage pf1 (39). To avoid structural noise from flexible regions, only the backbone amides from secondary structure elements were included in the analysis. This necessary precaution, together with the fact that a low protein concentration was used in the EIC-PEP sample to limit the rate of PEP hydrolysis, reduced the number the experimental RDCs available for the EIC-PEP complex to 37 (Table 2). However, because 1 D NH RDCs provide orientational information on NH bond vectors relative to an external alignment tensor (38,50), the subunit structure of EIC is known, and one of the principal components of the alignment tensor must lie along the C 2 symmetry axis of the dimer (17), this small number of RDCs is sufficient to fully describe the relative orientation of the two symmetry-related subunits (51).
Singular value decomposition fitting of the experimental 1 D NH RDCs to the coordinates of a single subunit from the phosphorylated EI x-ray structure (15) yields RDCs R-factors (52) of Ͻ22% for both EIC and the EIC-PEP complex ( Fig. 4 and

TABLE 2 SVD analysis of 1 D NH RDCs for the monomeric and dimeric forms of EIC and EIC-PEP
The coordinates used for SVD analysis are those of EIC in the crystal structure of phosphorylated EI (15). The RDCs for the EIC-PEP complex were measured on samples containing 100 M EIC and 50 mM PEP. The stability of the NMR samples used for measuring the RDCs was confirmed by acquiring 1 H-15 N TROSY spectra immediately before and after the acquisition of the ARTSY experiment. a The first and second numbers listed give the results of the SVD fits to a single subunit and to the dimer. b The alignment tensor is described by five parameters: three Euler angles (, , and ), the magnitude of the alignment tensor D a , and the rhombicity . For a symmetric dimer, one of the axes of the principal components of the alignment tensor coincides with the C 2 symmetry axis of the dimer. The fact that the values of the alignment tensor and RDC R-factor are the same for the SVD fits to a monomer and a dimer indicates that the orientation of the two subunits in the dimer in solution is the same as that in the crystal structure. c The RDC R-factor is given by ͓͗(D obs Ϫ D calc ) 2 ͘/(2͗D obs 2 ͔͘ 1 ⁄ 2 , where D obs and D calc are the observed and calculated RDCs, respectively (52). Table 2), confirming that the tertiary subunit fold of the secondary structure elements is unchanged relative to the crystal structure. The decrease in both the axial component of the alignment tensor, D a , and the rhombicity upon addition of PEP (Table 2) reflects changes in the surface charge distribution arising from the presence of PEP. When the RDCs measured for EIC and the EIC-PEP complex are fitted to coordinates of the dimer, the RDC R-factors increase by Ͻ1% (reflecting the doubling of the number of RDCs), and only minor variations within experimental error are observed in the Euler angles (Table 2), relative to the fits obtain with an individual subunit. These data demonstrate unambiguously, that although PEP binding results in local conformational changes within the active site and ␤3␣3 turn, the relative subunit orientation within the dimer is unaffected by the substrate and is the same as that observed in the various crystal structures.

Number of RDCs
Enzymatic Activity of E. coli EIC-During the course of our NMR studies on the EIC-PEP complex, we noticed that the 1 H-15 N TROSY correlation spectrum for the EIC-PEP complex changes over time and slowly reverts to that of the free protein, indicating that EIC degrades PEP and that the degradation products are no longer able to interact with the protein. This is confirmed by 31 P NMR, which shows that a 100 M EIC sample is able to fully hydrolyze 10 mM PEP into phosphate and pyruvate in Ͻ16 h (Fig. 5A). Similar results are obtained with fulllength EI (Fig. 5A).
An enzymatic assay that detects the amount of inorganic phosphate released as a consequence of PEP degradation (see "Experimental Procedures"), reveals that the hydrolysis of PEP by both EI and EIC follows Michaelis-Menten kinetics with K m values of 327 and 374 M, respectively ( Fig. 5B and Table 3). These values agree well with the previously reported K m (ϳ 400 M) for the EI-PEP interaction (20) and indicate that the isolated EIC domain binds PEP with essentially the same affinity as the full-length protein.
Differences, however, are observed for the EI and EIC kinetics, with full-length EI hydrolyzing PEP with ϳ1.5 times higher efficiency (reported as k cat /K m in Table 3) than the isolated EIC domain. A schematic diagram for the PEP degradation pathways catalyzed by EI and EIC is shown in Fig. 5C. The autophosphorylation reaction (equilibrium d in Fig. 5C), which initiates the PTS and occurs in only the full-length protein, is ϳ6 orders of magnitude faster than the direct hydrolysis of PEP into pyruvate and inorganic phosphate catalyzed by EIC (equilibria a and c in Fig. 5C). Phosphorylated EI, that is rapidly formed after addition of PEP to the full-length protein, can generate inorganic phosphate through two independent pathways that occur with similar velocities (equilibria e and f in Fig. 5C), explaining the higher efficiency of PEP degradation exhibited by the fulllength protein relative to EIC. It is also worth noting that no  Table 3 for EI (b) and EIC (a, c, and f). The velocity for the equilibrium d was calculated based on the second-order rate constant obtained for the EI autophosphorylation reaction at 25°C (21). The velocity for the equilibrium e was calculated based on the half-life (t1 ⁄2 ) for the decay of phosphorylated EI to unphosphorylated EI reported by Suh et al. (24). For all conversions, K m was set to 350 M, and the EI concentration was set to 5 M. EI 328 Ϯ 25 383 (Ϯ6) ϫ 10 Ϫ8 1.9(Ϯ1.5) ϫ 10 Ϫ3 5.9 Ϯ 1.7 EIC (species 2) a 374 Ϯ 30 267 (Ϯ5) ϫ 10 Ϫ8 1.3(Ϯ1.1) ϫ 10 Ϫ3 3.6 Ϯ 1.5 EIC (species 1) b Ͻ3 ϫ 10 Ϫ8 a EIC purified from species 2 exhibits a well dispersed 1 H-15 N TROSY correlation spectrum typical of a folded protein (Fig. 1E). b EIC purified from species 1 exhibits a 1 H-15 N TROSY correlation spectrum typical of an unfolded, partially folded, or incorrectly folded protein with minimal chemical shift dispersion (Fig. 1D).
hydrolysis of PEP by EIC purified from species 1 (see "Expression and Isolation of Recombinant E. coli EIC" and Fig. 1, A and B) could be detected by the enzymatic assay (Table 3), further confirming that species 1 is an inactive EIC isoform.
Concluding Remarks-Binding of PEP to the EIC domain of EI initiates the PTS (9) by inducing a series of intra-and interdomain conformational rearrangements (15,17) that result in phosphorylation of EIN and subsequent transfer of the phosphoryl group to HPr. Although conformational changes in EIC are thought to play an important role in regulation of the overall PTS (14,22), the conformations adopted by EIC during catalysis are still poorly understood.
We have presented a protocol for the successful expression and purification of recombinant EIC from E. coli. NMR structural characterization shows that in solution EIC adopts the same tertiary fold and dimeric structure as that observed in the various crystal structures of EI (15,16,26) and EIC (27,28). NMR 15 N relaxation dispersion data also reveal the existence of conformational exchange on the submillisecond time scale in the region of the PEP binding and at the ␤3␣3 turn. The relaxation dispersion data together with the available crystal structures of EI and EIC suggest that the ␤3␣3 turn in free EIC exists in a dynamic conformational equilibrium between open (major) and closed (minor) states that is shifted toward the closed state upon PEP binding. Conformational selection of the closed state by PEP does not perturb the quaternary structure and relative subunit orientation in the dimer but stabilizes the dimer interface, thereby explaining the lower dimerization K D for EIC in the presence of PEP (14).
We have also shown that the EIC domain, both isolated and in the context of the full-length protein, is able to hydrolyze PEP into pyruvate and inorganic phosphate, further supporting the hypothesis that EIC is not merely a scaffold for PEP binding and a dimerization domain but actually plays an active role in the catalytic function of EI (12). Although this enzymatic activity was not reported in previous investigations on the isolated EIC domain from E. coli (14), our finding is not too surprising since the isolated EIC domain from T. tengcongensis was observed to catalyze a similar reaction at 65°C (28). Our data also indicate that EI is capable of hydrolyzing PEP even when phosphorylated at His 189 . However, the PEP degradation rate is ϳ6 orders of magnitude slower than the autophosphorylation reaction of EI (21), suggesting that this additional processing of PEP is negligible in the context of the full PTS.
We expect that the stable and active EIC domain expressed and purified here will serve as an excellent model system for biochemical and structural investigations on the complex interplay between substrate binding and conformational dynamics that regulates EI and, more generally, the activity of many enzymes (53).