Insights into the inhibitory mechanisms of the regulatory protein IIAGlc on melibiose permease activity.

The phosphotransfer protein IIA(Glc) of the bacterial phosphoenolpyruvate:carbohydrate phosphotransferase system plays a key role in the regulation of carbohydrate metabolism. Melibiose permease (MelB) is one among several permeases subject to IIA(Glc) regulation. The regulatory mechanisms are poorly understood; in addition, thermodynamic features of IIA(Glc) binding to other proteins are also unknown. Applying isothermal titration calorimetry and amine-specific cross-linking, we show that IIA(Glc) directly binds to MelB of Salmonella typhimurium (MelB(St)) and Escherichia coli MelB (MelB(Ec)) at a stoichiometry of unity in the absence or presence of melibiose. The dissociation constant values are 3-10 μM for MelB(St) and 25 μM for MelB(Ec). All of the binding is solely driven by favorable enthalpy forces. IIA(Glc) binding to MelB(St) in the absence or presence of melibiose yields a large negative heat capacity change; in addition, the conformational entropy is constrained upon the binding. We further found that the IIA(Glc)-bound MelB(St) exhibits a decreased binding affinity for melibiose or nitrophenyl-α-galactoside. It is believed that sugar binding to the permease is involved in an induced fit mechanism, and the transport process requires conformational cycling between different states. Thus, the thermodynamic data are consistent with the interpretation that IIA(Glc) inhibits the induced fit process and restricts the conformational dynamics of MelB(St).

Sugar transport is an important process for all living organisms. To secure the energy supply, bacterial cells usually contain multiple sugar transport systems. Melibiose permease (MelB), 2 which catalyzes electrogenic symport of galactoside with Na ϩ , Li ϩ , or H ϩ (1)(2)(3)(4)(5)(6), is one of the bacterial sugar transporters. MelB is encoded by the mel operon, which requires transcriptional activation induced by melibiose, as well as a global transcriptional activator (the cAMP-CAP (catabolite activator protein) complex) (7,8). In certain bacteria, such as Escherichia and Salmonella, the phosphoenolpyruvate:carbohydrate phosphotransferase system (PTS) carries out both catalytic and regulatory functions, which promotes preferential utilization of glucose (9 -11). The phosphotransfer protein IIA Glc plays a key role in the regulation of carbohydrate metabolism. It regulates the transcription level of various operons encoding non-PTS permeases (12), such as the mel and lac operons, by modulating the level of cAMP. Unphosphorylated IIA Glc regulates the activity of various non-PTS transporters (9,11,(13)(14)(15)(16). In addition, as a central signaling molecule, it also binds to many other proteins, such as adenylate cyclase (17) and glycerol kinase (18). Recently, it has been reported that IIA Glc also binds to a carbon storage regulator and affects Vibrio cholerae biofilm formation (19). It is interesting that a single rigid protein is able to bind to various families of soluble proteins as well as membrane transporters. As demonstrated by x-ray crystallography, IIA Glc binds to the ATPase subunit of maltose permease, preventing structural rearrangements necessary for ATP hydrolysis (16). In the H ϩ -coupled lactose permease (LacY), it has been reported that the binding of IIA Glc requires the presence of sugar (13,20,21), but the reported stoichiometry number is not convincing.
In the Na ϩ -coupled MelB of Salmonella typhimurium (MeB St ), it was shown that mutations affect the PTS regulation (22,23); however, there is, at present, no evidence that IIA Glc directly binds to MelB (11). In addition, there is no information about the energetics of IIA Glc binding to other proteins. Recently, we solved the three-dimensional crystal structure of MelB St and demonstrated that MelB (24,25), a member of the glycoside-pentoside-hexuronide:cation symporter family (26), belongs to a subgroup of the major facilitator superfamily (MFS) permeases (5,(27)(28)(29), like LacY (30,31). The structures were captured in an outward partially occluded and a partially outward-facing conformation, and suggest a single sugar-binding pocket within the central internal cavity (5,(32)(33)(34). The structure provides important mechanistic insights for a major facilitator superfamily permease that catalyzes Na ϩ -coupled symport (26,35,36). However, it was not clear where the IIA Glcbinding site is and how IIA Glc regulates MelB activity.
It is generally believed that these IIA Glc binding partners have little or no sequence or structural homology with one another (37). It is surprising that the C-terminal tail of MelB Ec and MelB St , as well as other MelB orthologues, contains the consensus region 443 IQIHLLDK 450 that has a high sequence similarity to 121 LQLAHLLDR 129 in MalK (16,38). Both stretches form a short helical structure (16,25), and the underlined residues in MalK directly contact with IIA Glc (16). The crystal structure determination of MelB St reveals that this motif occupies two different conformations; one is closer to the membrane domain (see also Fig. 2c). The previously characterized MelB St mutants (D438Y, R441S, or I445N) (38), which are resistant to PTS inhibition, are mapped near or within this motif (see Fig. 2c). Based on this structural information, we agree with the previous postulation (38) that the C-terminal fragment of MelB could be a part of the IIA Glc -binding sites.
In this study, we determined the thermodynamics of the IIA Glc -MelB interaction, sugar binding to MelB St , as well as the effect of IIA Glc on the sugar-MelB interaction using isothermal titration calorimetry (ITC). We observed that IIA Glc binds to MelB St in the absence or presence of melibiose, and inhibits the conformational entropy and sugar affinity of the transporter.
Gene Cloning of IIA Glc -The gene encoding IIA Glc was amplified from the chromosomal DNA of Escherichia coli DW2 strain (5) by PCR (sense primer, 5Ј-TATATGCTCTTCTAGT-ATGGGTTTGTTCGATAAACTAAAATC-3Ј; antisense primer, 5Ј-TATATAGCTCTTCATGCTCATTACTTCTTAATGCGG-ATAACCGGAGT-3Ј), cloned into the T7-based expression vector p7XNH3 with a kanamycin resistance marker by the fragment-exchange cloning method (39). The resultant plasmid contains a 10-His tag sequence at the N terminus with a 9-residue linker (MHHHHHHHHHHLEVLFQGPS), which was verified by DNA sequencing analysis.
Protein Expression and Purification-The overexpression of IIA Glc was performed in the E. coli T7 express strain (New England Biolabs). The cells were grown in LB containing 0.5% glycerol, 0.2% glucose, and 50 mg/liter kanamycin. The overnight cultures were diluted to 2% with the same medium and shaken at 30°C. Isopropyl-1-thio-␤-D-galactopyranoside at 0.4 mM was added at A 600 of 0.8, and the incubation was continued for another 4 h. Cells were harvested and suspended in a buffer containing 50 mM NaP i , pH 7.5, 200 mM NaCl, 5% glycerol, and 0.1% PMSF, and broken by passage through an EmulsiFlex at 10000 p.s.i. The supernatant, after ultracentrifugation at 70.409 ϫ g for 30 min in a Beckman rotor, type 45 Ti at 4°C, was loaded onto a column containing Talon resin (Clontech) preequilibrated with 50 mM NaP i , pH 7.5, 200 mM NaCl, 5% glycerol, 5 mM imidazole for cobalt affinity chromatography. After washing with the same buffer containing 30 mM imidazole, elu-tion was performed using the same buffer containing 200 mM imidazole; the eluate was concentrated to ϳ100 mg/ml using a VIVASPIN 20 (5,000 molecular weight cut-off polyethersulfone, Millipore) and dialyzed against three changes of 1 liter of dialysis buffer containing 20 mM Tris-HCl, pH 7.5, 100 mM NaCl, 10% glycerol. The protein concentration of IIA Glc was measured by the Micro BCA protein assay (Pierce Biotechnology, Inc.). The protein samples were flash-frozen in liquid nitrogen and stored at Ϫ80°C. From a 1-liter culture, about 50 mg of highly pure IIA Glc protein can be obtained routinely. The purified IIA Glc protein was analyzed with both SDS-14% PAGE and Phos-tag SDS-12% PAGE.
Isothermal Titration Calorimetry-ITC measurements were performed in a Nano isothermal titration calorimeter (TA Instruments). Purified MelB (50 M) in MelB buffer without or with 10 mM melibiose was placed into the sample cell with a reaction volume of 163 l, and IIA Glc (455 M) in the MelB buffer without or with 10 mM melibiose, prepared by dilution from a highly concentrated protein sample, was titrated incrementally into the MelB sample. For measuring sugar binding to MelB, melibiose at 10 or 100 mM was dissolved in the MelB buffer; ␣-NPG or ␣-NPGlu at 1 mM was prepared by diluting 200 mM ␣-NPG or ␣-NPGlu in dimethyl sulfoxide into the MelB buffer. In this case, the same concentration of dimethyl sulfoxide was added into the sample placed in the sample cell. When testing the IIA Glc effect, IIA Glc was preincubated with MelB at a 2 to 1 ratio for 1 h.
Titrations were performed by injection of titrant with an interval of 250 or 300 s at a constant stirring rate of 250 rpm. By plotting integrated rates of heat change against the molar ratio of IIA Glc /MelB or sugar/MelB, the binding stoichiometry number (n), the association constant (K a ), and the enthalpy change (⌬H) are directly determined by fitting the data using the onesite independent binding model provided by the instrument. The dissociation constant (K d ) and the entropy change (⌬S) are obtained by calculation using the equation of ⌬G ϭ ϪRT ln K a and T⌬G ϭ ⌬H Ϫ ⌬G, where ⌬G is free energy change, R is the Faraday constant, and T is absolute temperature. Parameterization of ⌬S was calculated as described previously (42). Total ⌬S ϭ ⌬S solv ϩ ⌬S mix ϩ ⌬S conf . Mixing entropy change (⌬S mix ) ϭ R ln (1/55.5) ϭ Ϫ33 J/mol/K. Solvent entropy change (⌬S solv ) ϭ ⌬C p ln (298.15/385.15), where ⌬C p is the heat capacity change. Conformational entropy change ⌬S conf ϭ ⌬S Ϫ ⌬S mix Ϫ ⌬S solv .
Protein Cross-linking-The amine-reactive cross-linking reagent DSP was used for the cross-linking reaction between MelB St and IIA Glc . Briefly, 1 g of MelB St (1.88 M) in the absence or presence of 10 mM melibiose and IIA Glc (4.8 M) in 20 mM HEPES, pH 7.6, 50 mM NaCl, and 0.035% undecyl ␤-Dmaltoside were preincubated for 15 min, and the cross-linking reaction was carried out by incubating with 200 M DSP for 15 min at room temperature and stopped by the addition of 100 mM Tris-HCl. The reaction samples were analyzed with SDS-12% PAGE and visualized by silver staining.

IIA Glc Binding to MelB St in the Absence or Presence of
Melibiose-IIA Glc was purified to homogeneity from the E. coli T7 express strain (Fig. 1), and the Phos-tag SDS-PAGE analysis indicates that the affinity-purified recombinant IIA Glc protein is unphosphorylated (Fig. 1, right). By titrating IIA Glc into a MelB St sample at 25°C, ITC measurements show exothermic binding in the absence or presence of melibiose (Fig. 2a). No detectable changes were observed when injecting IIA Glc into the buffer (Fig. 2a, inset) or buffer to MelB St (data not shown). The data fitting (Fig. 2b) suggests dissociation constant (K d ) values of 3.62 or 10.15 M (Table 1) in the absence or presence of melibiose, respectively. The protein-protein interaction under both conditions is solely driven by favorable enthalpy change (⌬H) and opposed by negative entropy change (T⌬S) (Fig. 2b, inset; Table 1). The results indicate that polar or hydrophilic interactions are the major forces governing IIA Glc binding and that the charged or polar residues in the proposed IIA-Glc -binding site in the C-terminal tail may contribute to the enthalpy forces (Fig. 2c).
When melibiose is preincubated with MelB St (Fig. 2a, Table  1), the ⌬S becomes less unfavorable. The measured stoichiometry number (n) without melibiose is 0.98. In the presence of  melibiose, the n number is about 0.78; however, at 20°C, it is 1.1 ( Table 2). It is noteworthy that asymmetric peaks appear at the beginning of the titration in the absence of sugar (Fig. 2a).
The interaction of IIA Glc with MelB St was further tested by amine-specific cross-linking studies. In the absence or presence of melibiose, a band with M r ϭ ϳ62,000 was obtained only in the presence of cross-linking reagents, which corresponds to the cross-linked product containing one MelB St and one IIA Glc (Fig. 2d). These data are consistent with the results from the ITC measurements and support the conclusion that the stoichiometry of IIA Glc to MelB St is unity in the absence or presence of melibiose.
IIA Glc binding to E. coli MelB (MelB Ec ) was also examined by ITC. The data reveal thermodynamic features similar to that observed when injecting IIA Glc into MelB St , except for the higher K d value of 25 M (Fig. 3, Table 2); furthermore, there is no difference in the absence or presence of melibiose. The following studies focused only on IIA Glc binding to MelB St. Inhibition of Conformational Entropy by IIA Glc -IIA Glc binding to MelB St was further tested in the temperatures ranging from 20 to 35°C. With an increase in temperature ( Table 3, Fig. 4a), ⌬H values become more favorable, and ⌬S exhibits compensation to ⌬H, as indicated by the parallel curves. As a result of the compensation, there is little change  Table 2. Error bars, S.E.     Fig. 4b), which reveals a large increase in the solvent-entropy change (⌬S solv ) and a large unfavorable conformational entropy change (⌬S conf ). Without sugar, the binding yields a more favorable change in ⌬S solv and a more unfavorable change in ⌬S conf .

IIA Glc Restrains Conformational Entropy of MelB
Inhibition of Melibiose Affinity by IIA Glc -We further analyzed the effect of IIA Glc binding on sugar affinity. As we previously showed (25), the binding of melibiose to the Na ϩ -bound MelB St was detected by ITC measurement at 25°C; the binding is exothermic with a K d value of 950 M (Fig. 5a). The binding is driven by both favorable ⌬H and favorable ⌬S (Table 5). Strikingly, when injecting the melibiose solution to the MelB St -IIA Glc complex at 25°C (Fig. 5b) or 35°C (data not shown), the heat changes are smaller than the control. By injecting 10-fold higher melibiose (100 mM), a titration curve after correction for the buffer control shows an endothermic effect (Fig. 5c, inset). As a control, injection of 100 mM sucrose (a non-substrate) does not produce detectable binding signals (Fig. 5d). Although the weak signals were not amenable for fitting, the affinity is apparently inhibited.  Fig. 2 and plotted against temperature. The ⌬C p values were obtained by a linear fit of ⌬H. b, ⌬S solv and ⌬S conf . Both parameters were obtained as described in legend for the Table 4.   Inhibition of ␣-NPG Affinity by IIA Glc -When titrating ␣-NPG into MelB St at 25°C, large exothermic peaks were detected (Fig. 6a, red). As the control for the hydrophobic substrate binding to the detergent-solubilized membrane protein, we tested binding of the ␣-NPGlu. Under the same conditions, injection of ␣-NPGlu yields a flat titration curve with small exothermic peaks (Fig. 6a, black). The results strongly support the notion that the NPG titration curve reflects a specific binding to MelB St , which is consistent with the previous conclusion that MelB recognizes di-and trisaccharides containing the galactosyl moiety or galactose (1,5,25,43). The exothermic binding curve was also fitted using the one-site independent binding model (Fig. 6b). The K d is 15.45 M, and energetically, the binding is solely driven by ⌬H and opposed by ⌬S (Fig. 6b, inset; Table 5), which is different from the melibiose binding, as well as the NPG binding to LacY (44).
When titrating ␣-NPG into the IIA Glc -MelB St complex, similar to that observed with melibiose, an endothermic thermogram with small heat changes was observed (Fig. 6c, red). The K d value increases to 76.13 M, indicating that IIA Glc inhibits the ␣-NPG binding by 5-fold. The binding becomes solely driven by ⌬S with compensation of ⌬H ( Fig. 6d; Table 5).

DISCUSSION
Previous studies reported that the binding of IIA Glc to LacY requires the presence of sugar substrate (13,20,21); in contrast, we show here that IIA Glc binds to MelB Ec or MelB St in the absence or presence of the sugar. The binding to MelB St has higher affinity than that to MelB Ec . Without or with melibiose, the titration curves of IIA Glc to MelB Ec are similar. When injecting IIA Glc into MelB St without sugar, the heat change peaks in the beginning of the thermogram are asymmetric, which indicates that IIA Glc binding is relatively slow. It is likely that MelB St occupies several conformational states; not all the conformers can be recognized by IIA Glc , which causes the slow binding. The results suggest that the conformational transition is slow in the absence of melibiose. It is possible that the lower number of MelB St molecules available for the IIA Glc binding in the absence of melibiose or at a low temperature (Table 3) results in an apparently faster decrease of the peak height in the titration. Since MelB Ec is conformationally more flexible, a faster rate of IIA Glc binding and no sugar effect are observed. Thus, sugar binding facilitates the conformational dynamics associated with IIA Glc binding.
Although the IIA Glc -bound state of MelB and the details of the interaction are not known, two conformations, an outward partially occluded conformation and a partially outward-facing conformation, were determined by crystallography (25); both structures show a close surface at the cytoplasmic side, reflecting its low energy state. Accordingly, it seems reasonable to postulate that the cytosolic IIA Glc binds between the C-terminal tail and a closed face on the cytoplasmic side of MelB (Fig.  2c).
IIA Glc binding to MelB is involved in a large favorable ⌬S solv with compensation of a large unfavorable ⌬S conf . Because IIA Glc is structurally rigid (16,45), it is likely that the large change in ⌬S solv may mainly result from the conformationally flexible MelB St , as well as the binding interface. The inhibition of conformational entropy upon the binding implies that IIA Glc bind-  Table 5. The one-site independent binding model was used for data fitting. Error bars, S.E., number of tests ϭ 2.
ing prevents the conformational changes of MelB, similar to that proposed by the structural approach in maltose permease (16). The less restrained ⌬S conf is observed in the presence of melibiose, which is probably due to the idea that the conformational entropy was restrained to a certain extent by the sugar binding prior to the IIA Glc binding.
IIA Glc effects on MelB St affinity for sugars were examined with the lower affinity melibiose and the higher affinity ligand ␣-NPG. The ITC measurements show that IIA Glc binding affects the sugar binding significantly. It inhibits the sugar binding affinity; on the other hand, it alters the thermodynamic features of sugar binding from an exothermic to an endothermic reaction. Because the affinity of MelB St for melibiose is at the lower sensitivity boundary of the Nano ITC equipment, the decreased sugar binding affinity by IIA Glc is out of the detectable range of this method. We confirmed the IIA Glc inhibition of sugar binding with ␣-NPG. The ITC measurements reveal that ␣-NPG binds to MelB St with Ͼ60-fold higher affinity than does melibiose. Different from the entropy-driven melibiose binding, the ␣-NPG binding is solely driven by ⌬H, suggesting that the increased affinity mainly results from polar or hydrophilic interactions between ␣-NPG and MelB St . Furthermore, ␣-NPG binding to MelB St in the IIA Glc -bound complex becomes solely driven by ⌬S with a 5-fold higher K d value. Because the sugar binding to the permease is likely involved in an induced fit process as demonstrated in LacY (46,47), the results could be explained by the observed unfavorable ⌬S conf of the IIA Glc -MelB St complex. Thus, IIA Glc inhibits the induced fit process for sugar binding by restricting the conformational entropy.
Overall, the observed direct interaction of IIA Glc with MelB St inhibits the sugar affinity and conformational dynamics of the transporter protein. It is likely that by such a mechanism, unphosphorylated IIA Glc blocks entry of melibiose, the inducer of the mel operon, so that the cells utilize glucose via the PTS transport system.