Effects of Mutations and Truncations on the Kinetic Behavior of IIAGlc, a Phosphocarrier and Regulatory Protein of the Phosphoenolpyruvate Phosphotransferase System of Escherichia coli*

IIAGlc, a component of the glucose-specific phosphoenolpyruvate:phosphotransferase system (PTS) of Escherichia coli, is important in regulating carbohydrate metabolism. In Glc uptake, the phosphotransfer sequence is: phosphoenolpyruvate → Enzyme I → HPr → IIAGlc → IICBGlc → Glc. (HPr is the first phosphocarrier protein of the PTS.) We previously reported two classes of IIAGlc mutations that substantially decrease the P-transfer rate constants to/from IIAGlc. A mutant of His75 which adjoins the active site (His90), (H75Q), was 0.5% as active as wild-type IIAGlc in the reversible P-transfer to HPr. Two possible explanations were offered for this result: (a) the imidazole ring of His75 is required for charge delocalization and (b) H75Q disrupts the hydrogen bond network: Thr73, His75, phospho-His90. The present studies directly test the H-bond network hypothesis. Thr73 was replaced by Ser, Ala, or Val to eliminate the network. Because the rate constants for phosphotransfer to/from HPr were largely unaffected, we conclude that the H-bond network hypothesis is not correct. In the second class of mutants, proteolytic truncation of seven residues of the IIAGlc N terminus caused a 20-fold reduction in phosphotransfer to membrane-bound IICBGlc from Salmonella typhimurium. Here, we report the phosphotransfer rates between two genetically constructed N-terminal truncations of IIAGlc (Δ7 and Δ16) and the proteins IICBGlc and IIBGlc (the soluble cytoplasmic domain of IICBGlc). The truncations did not significantly affect reversible P-transfer to IIBGlc but substantially decreased the rate constants to IICBGlc in E. coli and S. typhimurium membranes. The results support the hypothesis (Wang, G., Peterkofsky, A., and Clore, G. M. (2000) J. Biol. Chem. 275, 39811–39814) that the N-terminal 18-residue domain “docks” IIAGlc to the lipid bilayer of membranes containing IICBGlc.

The phosphoenolpyruvate phosphotransferase system (PTS) 2 is a major pathway for the uptake of carbohydrates in the eubacteria (1,2).
Sugars that are PTS substrates are phosphorylated as they are translocated across the cell membrane, as illustrated schematically in Fig. 1 for the Glc-specific uptake systems of Escherichia coli and Salmonella typhimurium.
The PTS also serves several major regulatory functions, including inhibition of the uptake of several carbohydrates that are not PTS substrates and indirect regulation of carbon metabolism on a wide scale by control of the activity of adenylate cyclase (3). In E. coli, the regulatory function is served largely (1,2,4) by the 18.1-kDa phosphocarrier protein of the glucose-specific PTS, IIA Glc (in the older literature, this protein was called III Glc ). Unphosphorylated IIA Glc represses transcription of the uptake systems of three non-PTS carbon sources (lactose, maltose, and melibiose) by inhibiting the uptake of inducer (5), and, in the case of glycerol, which is taken up by facilitated diffusion, by inhibiting glycerol kinase (2). Extensive evidence suggests that [P]IIA Glc is a potent stimulator of adenylate cyclase (3). Therefore, both the presence or absence of IIA Glc , and its state of phosphorylation, are of importance for the regulation of cell growth. The state of phosphorylation of IIA Glc is determined by the relative flux of phospho-groups between it and HPr or IICB Glc , the membrane-associated glucose transporter.
Structural studies of IIA Glc by NMR and x-ray crystallography (6 -8) have shown that it comprises two domains: 1) an N-terminal domain that is unstructured in solution with the following amino acid sequence: (Met)-Gly-Leu-Phe-Asp-Lys-Leu-Lys-Ser-Leu-Val-Ser-Asp-Asp-Lys-Lys-Asp-Thr-Gly; the N-terminal Met is quantitatively removed post-translationally, and is not numbered, and 2) a compact domain that consists of the remaining 150 residues, including the active site, His 90 (9,10). The present report concerns two aspects of the IIA Glc structure thought to play important roles in the phosphotransfer reactions as follows.
The Catalytic Role of Amino Acids Close to His 90 -Transient-state (rapid quench) kinetic methods were adapted to study the phosphotransfer reactions of the PTS (11). Application of these methods to the kinetics of a site-directed mutation of His 75 (H75Q) (9), which lies very close to His 90 in the tertiary structure, showed that the mutation reduced the rate constants for phosphotransfer to and from HPr by a factor of 200.
Gln is isosteric with His, and structural studies of the (H75Q) IIA Glc mutant (12) indicated virtually no detectable change around the active site. Two hypotheses were offered to explain the 200-fold decrease in the rate constants. (a) His 75 stabilizes the negative charge on P-His 90 by virtue of a hydrogen bond and, consequently, delocalizes the negative charge on the intermediate (Fig. 2) (12). This delocalization of the charge would be substantially decreased in the Gln mutant. (b) The active site contains a possible "proton relay" network consisting of  There are five phosphotransfer reactions, each designated by the Roman numeral used throughout the text. The glucose permease, IICB Glc , is shown separated into its two domains, the phosphorylation domain IIB Glc , which extends into the cytoplasm, and the sugar recognition and binding domain IIC Glc , which is an integral membrane domain. Enzyme I is active only as a homodimer of 64-kDa subunit monomers; HPr is 9.1 kDa; IIA Glc is 18.1 kDa; and IICB Glc is a homodimer of 50.4-kDa subunit monomers (1,2). Enzyme I and HPr are called "general proteins," because they are common to all of the sugar-specific proteins in a bacterial species. HPr is a branch point in the flow of phospho-groups to the sugar-specific proteins. The ratio of IIA Glc /[P]IIA Glc is important for the regulation of carbohydrate metabolism, and this ratio is determined by the relative flux of phospho-groups between IIA Glc and HPr or IICB Glc .
Thr 73 -His 75 -P-His 90 . Whether proton transfer might take place between Thr 73 and His 75 during the phosphotransfer reaction had not been investigated.
In the present studies, the validity of the second hypothesis was tested by eliminating the H-bond network by substituting amino acids (Ser, Ala, or Val) for Thr 73 but maintaining His 75 . The rate constants for phosphotransfer between these mutants and HPr were measured.
The Function of the Unstructured N-terminal Domain-Early kinetic studies of IIA Glc by steady-state methods showed that [P]IIA Glc is a Michaelis-Menten substrate of IICB Glc with a definable V max and K m (13,14). A modified form of IIA Glc was isolated that was much less active kinetically (a few percent of wild type) (13). This protein was a truncated form of IIA Glc , lacking 7 residues at the N terminus. The truncation is catalyzed by what appears to be a specific membrane protease (15). The proteolysis occurs between Lys 7 and Ser 8 (13,15).
A function for the N-terminal domain of IIA Glc is suggested by recent studies of the physical properties of the domain. These suggest that it can form an amphipathic helix that serves as a membrane anchor (16,17). In the present work, we investigate the effects of two different length truncations (6 or 17 residues) of the N-terminal domain of IIA Glc on phosphotransfer rates with the soluble proteins HPr and IIB Glc , and the membrane proteins IICB Glc from S. typhimurium and E. coli.

EXPERIMENTAL PROCEDURES
Materials-All buffer salts and other reagents were of the highest purity commercially available. The pH of all buffers is reported at the temperature and concentration at which they were used. Methyl-␣-[U-14 C]glucoside was synthesized from [U-14 C]glucose (NEC 042, PerkinElmer Life Sciences) by the method of Bollenback (18), and the products were purified by the method of Austin et al. (19). The enzymatic synthesis of [ 32 P]PEP (20) was performed with modifications as described previously (11).
Bacterial Strains, Plasmids, and Growth Media-All strains were grown on Luria-Bertani broth containing the required antibiotics and inducers. E. coli strain BL21(DE3) (F Ϫ ompT hsdSB (rB Ϫ mB Ϫ ) gal dcm (DE3)) (Novagen) was used as the host for the pET21a plasmids containing the mutated genes for IIA Glc .
Construction of Site-directed Mutants of IIA Glc -A mutagenesis method based on a one-step PCR was used to construct plasmids containing deletions of either the first 8 or the first 17 codons of the gene. These codons include the N-terminal Met, which is not assigned an amino acid sequence number because it is removed post-translationally in vivo. Thus, the gene contains 169 codons, but the mature protein contains 168 amino acids, so that the number designating an amino acid residue is smaller than its codon number by one. In the case of the 8 codon deletion, Ser 8 was replaced by Met, so that the encoded protein was identical to the proteolytically truncated protein previously isolated, except that Met now replaced Ser as the first residue. This protein is referred to as (⌬7)IIA Glc . To test the effect of a more extensive truncation of the N-terminal domain, Thr 17 was replaced by Met, creating a protein that is 16 amino acids residues shorter than the mature, native protein and lacks virtually the entire N-terminal domain of IIA Glc (18 amino acids). This protein is referred to as (⌬16)IIA Glc .
The plasmid pDS35 (21) containing the crr gene was used as the template. PCR was performed by using two sets of primers. For S8M, the primers were: SS112 (5Ј-GGCGCCATTTTTCACTGCCAGAATTC-TTACTTCTTGATGCG-3Ј) containing the EcoRI site of restriction and S8M (5Ј-GTTCGATAAACTGCATATGCTGGTTTCCGAC-3Ј) containing an NdeI restriction site. For T17M the primers were: SS112 and T17M (5Ј-GACGACAAGAAGCATATGGGAACTATTGAG-3Ј) containing the NdeI site. After the PCR reaction, PCR products were restricted with NdeI and EcoRI and then cloned into plasmid pET21a (Novagen).
Purification and Characterization of HPr, [ (23) and HPr (24) were isolated as described. The various site-directed mutants of IIA Glc were purified by the same methods used for the wild-type protein.
The T73V mutant of IIA Glc behaved somewhat differently during the purification and required a second passage on the ion exchange column to obtain homogeneous protein. IIA Glc was tested for proteolysis by the use of polyacrylamide gel chromatography (13). IIB Glc -6His was purified by the method of Buhr et al. (25), except that a Superose 12 HR 10/30 column (Amersham Biosciences) was substituted for Sephadex G75. The final preparations were apparently homogeneous as judged by SDS-PAGE. The methods previously described (11) were used to determine the concentrations of HPr, IIA Glc , and IIB Glc -6His. [ 32 P]HPr, [ 32 P]IIA Glc , and [ 32 P]IIB Glc -6His were prepared as described previously, with the same attention to the accuracy of the specific activity of the [ 32 P]PEP (11).
Sugar Phosphorylation Assay for PTS Activity and the Determination of Specificity Constants-The PEP-driven sugar phosphorylation assay was performed as reported (26,27). Membrane suspensions (28) for these assays were prepared from E. coli strain ZSC112⌬G harboring plasmid pCB30 (encoding wild-type IICB Glc ) and S. typhimurium, strain PP1133 (29). To determine V max and K m (IIA Glc ), the assay mixture contained fixed concentrations of the following substances in a volume of 0.1 ml: 50 mM Tris/Cl Ϫ buffer, pH 8.0; 10 mM KF; 5 mM MgCl 2 ; 10 mM PEP; 5 mM methyl-␣-[U-14 C]glucoside (5 Bq/nmol); 8 nM Enzyme I (this is 5 units, where a unit is defined as the quantity of enzyme that produces 1 mol of sugar phosphate in 30 min at 37°C); and 6 M HPr. The assays contained Ͻ0.2 unit of IICB Glc ; the concentration of E. coli IICB Glc was ϳ16 nM and S. typhimurium IICB Glc ϳ20 nM determined as described (28). The concentrations of IIA Glc , either wild type or truncated, were 3, 5, 20, and 60 M. V max and K m (IIA Glc ) were calculated by the methods of Eadie-Hofstee and Hanes (30), and the values were averaged.
The specificity constant of IICB Glc for IIA Glc is defined as k cat /K m (IIA Glc ) (30), where k cat is defined as V max /[IICB Glc ] total . We emphasize that the specificity constant is mathematically equal to the rate constant in a reaction catalyzed by an enzyme with a ping-pong mechanism.
Rapid Quench Assays-The present study employed the same rapid quench apparatus described previously, including all the details for its set-up and use, treatment with bovine serum albumin to eliminate adsorption of protein, etc. (11). The quench solution for experiments with HPr and IIA Glc was 3 M KOH with 9 M urea, used as one volume of quench to two volumes of reaction mixture (11). For experiments with IIA Glc and IIB Glc the quench solution was 0.3 M KOH, 9 M urea, also used as one volume of quench to two volumes of reaction mixture (28). Analysis of the quenched reactions by gel-filtration chromatography using high-performance liquid chromatography grade columns at pH 12.3 was also as described (11). Preparation of the solutions for rapid quench experiments required large dilutions from stock solutions of Enzyme I and HPr, and a change from the frozen state to ambient temperature (ϳ23°C), at which all experiments were performed. The solutions were therefore incubated for an hour at ambient temperature before the experiment was started.
Methods Used to Model the Experimental Data on the Rate of Phospho-group Transfer-The numerical integration program, Kinsim (31), as modified by Anderson et al. (32), was used to manually fit mathematical models to the experimental data. When experimental data met the criteria for non-linear least squares fitting of mathematical models, the Fitsim (33) module of Kinsim was used. The model used for the kinetic analysis of the phosphotransfer reactions is shown as Scheme 1.

H-Bond Network
An earlier study showed that replacement of His 75 with Gln in IIA Glc reduced both of the rate constants for the reversible phosphotransfer reaction with HPr by ϳ200-fold (11). His and Gln are approximately isosteric and can make approximately the same hydrogen bonds. That is, each has two locations for H-bonding. Depending on orientation either can be a donor or acceptor, and the H-bonding atoms can be in approximately the same spatial location in each case. It has been experimentally determined that in the H75Q mutant, Gln 75 is a hydrogen bond donor to the phosphogroup, but the bond from Thr 73 is lost (12), i.e. the hydrogen bond network is lost. This result was the basis of one of two hypotheses to explain the 200-fold decrease in activity of the His 75 3 Gln mutant. The idea was tested by retaining His 75 but destroying the H-bond network; amino acids were substituted for Thr 73 . If such mutants showed a greatly lowered activity, similar to the original (His 75 3 Gln), this would lend credibility to the idea of a catalytic proton derived from the H-bond network. If the mutants behaved like wild-type IIA Glc , this would be strong evidence supporting the second hypothesis, i.e. the importance of the imidazole ring of His 75 . The mutants Thr 73 3 Ser, Ala, and Val were therefore constructed and tested.
The results with each of the mutants are summarized and compared with wild-type IIA Glc in Table 1. A progress curve typical of these experiments is shown in Fig. 3.
The rate constants of the T73A and T73S mutant proteins are reduced by a factor of less than two relative to those of the wild-type protein. The T73V mutant shows a somewhat larger defect, about a factor of three. These reductions obviously do not explain the 200-fold effect observed the H75Q mutation.
Thus, the present results suggest that the hydrogen bond between Thr 73 and His 90 is not catalytic. It therefore appears likely that the 200fold decrease in the H75Q mutation ensues from loss of transition state stabilization by the imidazole ring of His 75 and that this function cannot be effectively replaced by the amido group of glutamine.

Function of the N-terminal Domain
In the PTS reaction sequence (Fig. 1), IIA Glc accepts a phosphoryl group from [P]HPr and is a donor to the IIB (cytoplasmic) domain of the membrane protein, IICB Glc . The reactions studied here with the cloned, soluble IIB Glc domain are referred to as Reaction IVa in Scheme 1.
We previously reported that proteolysis of the seven N-terminal residues of IIA Glc from S. typhimurium reduced its activity in sugar phosphorylation assays to a few percent of the activity of the full-length protein (13). In the present studies, we test the role of the N-terminal domain of IIA Glc on its kinetic properties by constructing and assaying two truncations, (⌬7)IIA Glc and (⌬16)IIA Glc , as described under "Experimental Procedures." The phosphotransfer properties of the truncations were tested with the soluble proteins, HPr and IIB Glc , and with the membrane protein IICB Glc . A differential effect on the rate constants might indicate whether the N-terminal domain requires membrane lipid to exert its effect.
Effects of Truncations on Phosphotransfer Reactions with HPr-Experimental and theoretical progress curves for phosphotransfer Reaction III between (⌬7)IIA Glc and HPr are shown in Fig. 3, and a summary of six experiments is given in Table 1. Truncation has only a small effect on Reaction III, reducing both k III and k ϪIII by a factor of about two, and the length of the truncation did not matter.
Two recent studies of the interacting complexes of HPr and IIA Glc (34) and IIA Glc and IIB Glc (35) have shown that the N-terminal domain of IIA Glc remains unstructured in both interactions. The weak effects of truncation on the kinetics of phosphotransfer are consistent with these observations.
Effects of Truncations on Reactions with IIB Glc 6His-A progress curve of P-transfer between the truncated IIA Glc proteins and IIB Glc is given in Fig. 4, and the rate constants for this reaction are summarized in Table  2. Both truncations resulted in an increased rate constant for P-transfer, SCHEME 1. The balanced equations for the phosphotransfer reactions between [P]IIA Glc and HPr or IIB Glc . The signs of the rate constants are positive for reactions proceeding left to right. The balanced equations are for the scheme of reactions used in the kinetic simulator, Kinsim. The convention for numbering the reactions is adapted from Rowher et al. (38). By this convention, the data reported previously (11) pertained to Reaction III. Reaction IVa applies to IIB Glc , the molecularly cloned, cytoplasmic domain of Enzyme IICB Glc . "IIA Glc " pertains to the wild-type protein as well as to the mutant forms used in this report.

TABLE 1
Rate constants for phosphotransfer between HPr and IIA Glc , wild type, and mutants Second order rate constants obtained from rapid quench experiments by manual fitting of the data using Kinsim. The concentration of the phospho-donor proteins ranged from 20 to 50 nM and that of the acceptor proteins ranged from 25 to 50 nM, similar to those used previously (11).
a The values for k III and k ϪIII , obtained at pH 7.5, are similar to those obtained previously at pH 6.5 (11), which were 61 ϫ 10 6 M Ϫ1 s Ϫ1 and 47 ϫ 10 6 M Ϫ1 s Ϫ1 , respectively. b One of the experiments with this protein was started in the reverse direction, i.e., with HPr as the acceptor and ͓ 32 P͔phospho-IIA Glc as the donor. The rate constants derived from experiments started from opposite directions were in good agreement, indicating that there are no significant concentrations of intermediate complexes between the two reacting proteins prior to the last step, transfer of the phosphoryl group to the acceptor, and separation of the proteins to yield the products.
i.e. a small effect, ϳ1.4-fold, in the forward direction, and ϳ6-fold in the reverse reaction (i.e. Reaction IVa). There is no simple structural interpretation of these differing results: the decreased rate constants with HPr and the increased rate constants of the truncated mutants with IIB Glc . However, the results suggest that the N-terminal domain of IIA Glc can interact weakly IIB Glc , even at the low concentrations of the proteins used for these experiments.
Effects of Truncations on Sugar Phosphorylation Rates in the Complete PTS System; Steady-state Assays-Proteolytic truncation of IIA Glc reduced its activity in sugar phosphorylation assays by ϳ30-fold (11). The earlier experiments employed a partially purified preparation of IICB Glc (27) to assay homogeneous, intact, or truncated IIA Glc , all from S. typhimurium, whereas all the work presented here employed proteins from E. coli. To clarify this issue, we performed sugar phosphorylation assays with the genetically modified (⌬7)IIA Glc (E. coli) and IICB Glc from both E. coli and S. typhimurium. The effects of truncation of E. coli IIA Glc on the specificity constants of IICB Glc from E. coli and S. typhimurium are shown in Table 3.
Truncation of IIA Glc reduces the specificity constant of E. coli IICB Glc by a factor of about 4 relative to intact IIA Glc , whereas the specificity constant of S. typhimurium IICB Glc is reduced by a factor of ϳ12, in reasonable agreement with the previously published results.
The 4-fold change with E. coli membranes may seem small, but may be very physiologically significant. In a previous paper (38) we report the results of transport experiments with whole cell and found that the pivotal protein in Glc transport was IICB Glc at its usual cellular levels. As can be seen in Fig. 1 of that paper (panel d), decreasing the activity of the Glc transporter by a factor of 4 would reduce the Glc uptake rate to close to 0, certainly enough to seriously affect growth.
The differences in the effect of truncation of IIA Glc on its activity with E. coli and S. typhimurium IICB Glc membrane preparations  Table 1. Progress curves for the phosphotransfer reaction between HPr and the IIA Glc Thr 73 mutants are very similar to the one shown here.  Table 2.

TABLE 2 Rate constants for phosphotransfer between IIA Glc mutants and IIB Glc 6His
Second order rate constants obtained from rapid quench experiments by manual fitting of the data using Kinsim. The concentration of the phospho-donor and phospho-acceptor proteins ranged from 30 to 50 nM, similar to those used previously (11,28).

IIA Glc
k IVa (k IV ) a k ؊IVa (k ؊IV ) a n K

coli and S. typhimurium
The specificity constant is the ratio: k cat /K m (IIA Glc ) and is identical to the rate constant, k IV , for an enzyme with a ping-pong mechanism (28,30). The data were obtained with sugar phosphorylation assays as described under "Experimental Procedures." IICB Glc in washed membranes was used because it yields more reproducible results than the purified protein (28). could be an anomaly resulting, for instance, from differences in the physical structures of the two membrane preparations, or, could be more significant, reflecting differences in the phospholipid compositions surrounding the Glc transport proteins, or, of the intrinsic properties of the proteins. 3 Conclusions-The N-terminal domain (first 18 amino acids) of IIA Glc has been shown to be unstructured both in crystallographic and NMR experiments (6 -8), and this domain is not located near the active site His 90 . Thus, there was no obvious explanation for the substantial loss in activity when IIA Glc was truncated (N-terminal 7 amino acids) by a specific membrane protease (15). The results reported here with genetically engineered ⌬7 and ⌬16 amino acid truncations explicitly show that the N terminus has no large effect on phosphotransfer to and from the soluble proteins, HPr and IIB Glc , but there is suggestive evidence for a weak interaction with IIB Glc . This observation is important because IIB Glc is a domain of the membrane protein, IICB Glc , that is affected more strongly by the truncations.

IIA
Our results can be explained and provide support for the contention by Wang et al. (16,17) that the N-terminal domain forms an amphipathic helix in the presence of anionic phospholipids and their proposal that this helix enhances the interaction of [P]IIA Glc and IICB Glc during phosphotransfer. In other words, that the N-terminal domain acts in "docking" of phospho-IIA Glc close to the active site Cys of the IIB Glc domain of the membrane protein IICB Glc . But for the N-terminal domain to specifically direct this attachment in the correct position on a membrane containing innumerable proteins requires somewhat more than a nonspecific amphipathic helix. The kinetic results, although relatively small, suggest in fact that the N-terminal domain does play a role in the interaction between P-IIA Glc and IIB Glc .
Perhaps the large effect observed with the truncated IIA Glc in the PTS sugar phosphorylation assay, especially with the S. typhimurium membranes, reflects loss of a synergistic effect in the IIA Glc truncations. That is, the role of the N-terminal domain is to form an amphipathic helix that binds to the lipid bilayer and to interact with the IIB Glc domain in IICB Glc .