Topography of the Surface of the Signal-transducing Protein EIIAGlc That Interacts with the MalK Subunits of the Maltose ATP-binding Cassette Transporter (MalFGK2) of Salmonella typhimurium*

The signal-transducing protein EIIAGlc, a component of the phosphoenolpyruvate-glucose phosphotransferase system, plays a key role in carbon regulation in enteric bacteria, such as Escherichia coli and Salmonella typhimurium. The phosphorylation state of EIIAGlc governs transport and metabolism of a number of carbohydrates. When glucose as preferred carbon source is transported, EIIAGlc becomes predominantly unphosphorylated and allosterically inhibits several permeases, including the maltose ATP-binding cassette transport system (MalFGK2) in a process termed “inducer exclusion.” We have mapped the binding surface of EIIAGlc that interacts with the MalK subunits by using synthetic cellulose-bound peptide arrays like pep scan- and substitutional analyses. Three regions constituting two binding sites were identified encompassing residues 69-79 (I), 87-91 (II), and 118-127 (III). Region III is MalK-specific, whereas residues from regions I and II partly overlap but are not identical to the binding interfaces for interaction with glycerol kinase and lactose permease. These results were fully verified by studying the inhibitory effect of purified EIIAGlc variants carrying mutations at positions representative of each of the three regions on the ATPase activity of the purified maltose transport complex reconstituted into proteoliposomes. Moreover, a synthetic peptide encompassing residues 69-91 was demonstrated to partially inhibit ATPase activity. We also show for the first time that the N-terminal domain of EIIAGlc is essential for inducer exclusion.

Complex bacterial activities, like those involved in the quest of food, require the coordination of entire metabolic networks. In enteric bacteria, such as Escherichia coli and Salmonella typhimurium, the phosphoenolpyruvate carbohydrate phosphotransferase system (PTS) 2 plays a key role in this process as a signal transduction system (1). The PTS comprises a cascade of protein kinases and phosphocarriers that constitute a series of transport systems, which couple transport and phosphorylation of numerous sugars. The pathway requires the sequential transfer of a phosphoryl group from phosphoenolpyruvate via EI, HPr to the sugar-specific EII components. "Catabolite repression" allows for the preferential utilization of so-called Class A sugars mostly transported by the PTS, when cells are cultured in the presence of other PTS or non-PTS sugars (Class B). Consequently, the PTS regulates the uptake of Class B-sugars by both transcriptional and post-transcriptional mechanisms. The phosphorylation state of the glucose-specific EIIA Glc component plays a key role in this process. Phosphorylated EIIA Glc is believed to function as an allosteric activator of adenylate cyclase. In contrast, when glucose as a preferred carbon source is transported EIIA Glc becomes predominantly unphosphorylated and inhibits uptake of glycerol, lactose, melibiose, maltose, and other sugars by direct binding to the permeases or glycerol kinase. Thus, at least 10 different proteins have been shown to bind EIIA Glc (2).
EIIA Glc (169 amino acids, M r 18.2 kDa) is an antiparallel ␤-sandwich (3). Histidines 75 and 90, located off center on one face of the ␤-sandwich are required for the phosphoryl transfer reaction. Covalently modified EIIA Glc has been characterized and shown to carry the phosphoryl group attached to His 90 (4). The histidines are mostly buried within a shallow ring of solvent-exposed hydrophobic residues that have been suggested to provide a binding site for regulatory targets (3,4). Structural analyses of complexes of EIIA Glc with glycerol kinase (5), glucose permease (EIICB) (6,7), and HPr (8) largely confirmed this notion. For lactose permease, a biochemical analysis also gave rise to a model that suggests an overlap of the interacting surface of EIIA Glc with that for interaction with other proteins (9).
Thus far, nothing is known on the binding site by which EIIA Glc interacts with the maltose/maltodextrin ATP-binding cassette transport system. The maltose transport complex (MalFGK 2 ) consists of one copy each of the membrane-integral subunits, MalF and MalG, and two copies of the ATPase subunit, MalK. In addition, the periplasmic maltose-binding protein, MalE, is also required for function (10,11). Inhibition of maltose transport by EIIA Glc in vivo (2) is most likely caused by eliminating substrate-stimulated ATPase activity of the transporter as demonstrated in vitro (12,13), suggesting a direct interaction with the MalK subunits. The localization of most mutations that render the transporter insensitive to inducer exclusion further indicated that the C-terminal domain of MalK might interact with EIIA Glc (12,14,15). This notion was also confirmed by a study using monoclonal antibodies recognizing C-terminal epitopes on MalK (16). The observation that the intrinsic ATPase activity of the purified MalK subunit is largely insensitive to EIIA Glc suggests that binding of EIIA Glc might interfere with signal transduction during the transport cycle (13).
Here, we identify the binding surface of EIIA Glc that interacts with MalK by employing synthetic peptide arrays in combination with functional analysis of mutant variants at the level of purified proteins. Our results suggest that two binding interfaces exist; one partly overlaps but is not identical to that for interaction with glycerol kinase (4,5) and lactose permease (9), whereas the other is unique for MalK.

EXPERIMENTAL PROCEDURES
Plasmids-Plasmid pBB04 was constructed by ligating a PCR fragment encompassing the crr gene (encoding enzyme IIA Glc ) as an NdeI-BamHI fragment with expression vector pET15b (Novagene). Derivatives of pBB04 carrying single point mutations were obtained by site-directed mutagenesis using Stratagene's QuikChange kit. Plasmid pBB06 harboring a crr allele that lacks codons 1-17 was constructed similarly by using a PCR primer introducing an NdeI site (comprising an ATG codon) upstream of codon 17. As a consequence, the translated protein (⌬G1-D16) contains a methionine residue fused to the N terminus of Thr 17 of native EIIA Glc . Plasmid pBB07 carrying a crr allele that encodes an EIIA variant lacking the N-terminal residues Gly 1 -Lys 7 was constructed likewise.
Preparative Procedures-MalK (17), MalFGK 2 (13), and MalE (18) of S. typhimurium were purified as described. MalE/maltose-loaded proteoliposomes were prepared as in Ref. 13. Enzyme IIA Glc (wild type and variants) was purified from the cytosolic fraction of E. coli strain BL21 ⌬(pts43crr::kan R ) (16) harboring plasmid pBB04 or derivatives by Ni-NTA affinity chromatography. Subsequent removal of the His tag was carried out by incubation with thrombin according to the manufacturer's instructions (Novagene). Samples were then passed through PD10 followed by a second passage through Ni-NTA. Tagless enzyme IIA Glc was collected in the flow through.
Peptide Synthesis on Cellulose Membranes (SPOT Synthesis)-Cellulose-bound peptide libraries were prepared by semi-automatic SPOT synthesis using a SPOT robot (INTAVIS AG, Köln, Germany). SPOT synthesis was carried out as described in the SPOT synthesis protocol (20), and arrays were synthesized on modified Whatman 50 cellulose membranes (Whatman, Maidstone, UK). Sequence files and a design of the arrays were generated with the in house software LISA. Peptides derived from S. typhimurium EIIA Glc (accession number AAL 21327) (Gly 2 -Lys 169 ) were used for pep scan analyses. To this end, three peptide arrays consisting of 13-meric, 16-meric, or 31-meric peptides, overlapping by 12, 15, and 30 amino acids, respectively, were synthesized. Complete substitutional and length analyses of the interacting 16-mer peptides were generated using the software LISA and subsequently synthesized as described (21,22).
Screening of Cellulose Membrane-bound Peptide Arrays-Before screening, the dried membranes were washed for 10 min in ethanol, 3 ϫ 10 min in TBS (50 mM Tris-HCl, pH 8, 137 mM NaCl, 27 mM KCl) and subsequently incubated in TBS, supplemented with 5% blocking buffer (Sigma) and 5% sucrose, for 3 h at room temperature. After washing with TBS peptide arrays were incubated with 35 S-labeled MalK (10 5 cpm) in blocking buffer, supplemented with 20% (v/v) glycerol for 6 h at room temperature with gentle shaking. Unbound MalK was removed with TBS, and peptide-bound MalK was visualized and quantified using a phosphoimager and associated software (Fuji, Japan).
Peptide Synthesis-Soluble peptides encompassing ␤-strands 5-7 (T66-IGKIFETNHAFSIESDSGIELFVHFGIDT-V96) or a non-binding region of enzyme IIA Glc (T136-PVVISNMDEIKELIKLSGSVTVGETP-VIR-I166) were prepared by automatic solid phase peptide synthesis on a Tentagel-SRam resin (Rapp Polymere, Tübingen, Germany) using Fmoc (N-(9-fluorenyl)methoxycarbonyl) chemistry. Chemical reactions were performed in plastic syringes at room temperature on a multipeptide synthesis robot (Syro2000, MultiSynTech, Witten, Germany) according to the manufacturer's protocol. The final peptides were deprotected and cleaved off the resin using a mixture of 10 ml of trifluoroacetic acid, 0.75 g of phenol, 0.5 ml of water, 0.5 ml of methylphenyl sulfide, and 0.25 ml of 1,2-ethanedithiol. After incubation for 3 h at room temperature, the cleavage solution was collected, and the crude peptides were precipitated with dry ether at 0°C. Purification of the peptides was achieved by high-pressure liquid chromatography on a RP-18 column using a linear solvent gradient (A, 0.05% trifluoroacetic acid in water; B, 0.05% trifluoroacetic acid in acetonitrile; gradient 5-60% B over 30 min). The identity of the purified peptides was validated by mass spectrometry using matrix-assisted laser desorption ionization timeof-flight (VoyagerLT, Applied Biosystems, Weiterstadt, Germany) or electrospray ionization-mass spectrometry (Q-TOFmicro TM , micromass, Manchester, UK).
Determination of the Phosphorylation State of Enzyme IIA Glc -The phosphorylation state of enzyme IIA Glc was determined as described (23). The assay takes advantage of the observation that phosphorylation causes a shift of the apparent molecular weight of EIIA Glc on SDS gels. A cytosolic fraction of E. coli strain LS20 (⌬crr::kan) lacking EIIA Glc but containing EI and HPr was prepared from cells grown in minimal medium E (19) supplemented with glucose (0.4%), thiamine (2 g/ml), and kanamycin (25 g/ml) at 37°C to an optical density of 1.2. Cells were subsequently harvested and disintegrated by ultrasonication. After centrifugation for 30 min at 200,000 ϫ g the supernatant (cytosol) was stored at Ϫ80°C until use. Assay mixtures (final volume: 30 l) contained 50 mM Tris-HCl, pH 7.5, 20% glycerol (v/v), 0.15 M NaCl, 0.1 mM phenylmethylsulfonyl fluoride, 5.7 mM phosphoenolpyruvate, 67 M MgCl 2 , and purified EIIA Glc variants (1 g) as indicated. The reactions were started by the addition of cytosolic fraction (40 g of protein) and terminated after 0 and 20 s, respectively, by adding 4 l of 10ϫ SDS-PAGE sample buffer. Subsequently, the solutions were loaded onto 14% SDS-polyacrylamide gels. After transfer to nitrocellulose, the blots were incubated with polyclonal anti-EIIA Glc antibody (generous gift of K. Jahreis, Universität Osnabrück, Germany) and a horseradish peroxidase-conjugated secondary antibody. Antigen-antibody interactions were visualized using the Western blot chemiluminescence reagent plus system (PerkinElmer Life Science products).
To determine the time dependence of EIIA Glc phosphorylation, the mixtures (140 l) contained 200 g of cytosolic fraction of LS20 and 5 g of EIIA Glc variants. Aliquots (15 l) were taken at the indicated times and further treated as described above.
Analytical Procedures-ATPase assay, protein determination, SDS-PAGE, and immunoblotting were performed as described elsewhere (16).

RESULTS
Screening of EIIA Glc -derived Peptide Arrays for Binding to MalK-To determine the binding motif within EIIA Glc recognized by MalK, we screened cellulose-bound peptide arrays synthesized by SPOT-synthesis and representing the complete EIIA Glc sequence (Gly 1 -Lys 168 ) for MalK binding. According to the convention adopted for E. coli IIA Glc in the literature (2,3), the N-terminal methionine, which can be hydrolyzed, is numbered as zero throughout the manuscript.
The peptide arrays consisted of 13-mers, 16-mers, or 31-mers that overlapped with adjacent peptides by 12, 15, and 30 amino acids, respectively. The cellulose membranes were incubated with 35 S-labeled MalK and washed to remove excess label, and the retained radioactivity was visualized by phosphoimaging. Signals were consistently obtained with all three arrays in four rows but with different intensities (Fig. 1). Strongest signals were observed in rows 2 and 3, whereas the signal in row 1 was rather weak. The reliability of the data was further confirmed by probing three independently synthesized peptide arrays consisting of 16-mers. Each membrane gave the same signals with similar intensities (data not shown).
Substitutional and Length Analyses-To identify those amino acid residues within each peptide that are indispensable for binding of MalK, substitutional analyses of the peptides from signal rows 1-3 were performed (signal row 4 was suspected to result from an unspecific reaction possibly due to the positive charges at the C-terminal end of the protein and was thus not further studied, but see Fig. 7B). In these experiments every position was substituted one-at-a-time by all other genetically encoded amino acids. Thus, all possible single site substitution analogs were synthesized and screened. The results shown in Fig. 2 represent data obtained for spots 64 (A), 80 (B), and 112 (C) of the initial 16-mers array (see Fig. 1, center panel). Basically the same discrete substitution patterns were identified with 31-mers (not shown). The signal intensities obtained with the wild type sequence of each peptide (Fig. 2, left columns) correlated with those seen in Fig. 1. Thus, at least in case of Fig.  2A, the spots were somewhat difficult to detect on a printout. However, close inspection of the images on the computer screen clearly revealed individual residues on each membrane that could not be replaced by any other amino acid without affecting binding of MalK.
We also determined the minimal length of a peptide required to bind MalK. To this end, 16-mers and 31-mers were shortened sequentially from the N-and C-terminal end down to 10-mers and 16-mers, respectively (not shown).
The combined results identified the peptides comprising amino acids 69 -79 (KIFETNHAFSI) (region 1), 87-91 (LFVHF) (region 2), and  118 -127 (PVIEFDLPLL) (region 3) as being crucial for MalK binding (Fig. 3). None of the hydrophobic residues (underlined) can be replaced by amino acids with charged or polar side chains without loss of binding, suggesting that the binding interface is largely hydrophobic. Regions 1 and 2 basically encompass ␤-strands 5-7 of enzyme IIA Glc , which have also been implicated in recognizing other target proteins (4, 5, 7-9) (Fig.  3). However, the binding sites are not identical and, most importantly, region 3, located on the opposite face of the ␤-sandwich (Fig. 4), appears to be unique for binding MalK. Lys 69 represents the only critically charged residue for interaction with MalK. Interestingly enough, hydrophobic residues substituted for Lys 69 are tolerated, whereas negatively charged and some polar uncharged residues (Gln, Pro, Asn, Met) caused a loss of binding ( Fig. 2A).
Mutational Analysis of Selected Residues of EIIA Glc -The results described above suggested that amino acid residues encompassing part of ␤5, ␤6, and ␤7 as well as of ␤11 and ␣2 of EIIA Glc are crucial for binding of MalK. To verify these predictions in the context of the folded protein we monitored ATP hydrolysis catalyzed by the reconstituted MalFGK 2 complex in the presence of purified EIIA Glc variants. To this end, variants of EIIA Glc carrying substitutions of residues from all three peptide regions involved in binding were constructed by site-directed mutagenesis (see "Experimental Procedures"). In particular, residues representing regions 1 (Lys 69 , Phe 71 ) and 2 (Phe 88 , Phe 91 ) were replaced by glutamine and phenylalanine, predicted by substitutional analyses (Fig. 2) to either eliminate or allow binding of MalK. Similarly, residues representing region 3 were changed to isoleucine and glutamine (Phe 122 ) or threonine (Leu 127 ). Moreover, Arg 165 and Lys 167 from the C-terminal end (signal row 4, Fig. 1) were also included. Each variant could be overproduced in soluble form comparable to wild type and was purified accordingly. Furthermore, after purification the N-terminal His tag was removed from each mutant protein to exclude possible artifacts by unspecific interaction (Fig. 5A).
As shown in Figs. 6 and 7, and in agreement with a previous study (13), wild type EIIA Glc when added in a 20 -25-fold molar excess substantially inhibited the MalE-maltose-stimulated ATPase activity of MalFGK 2 -containing proteoliposomes. In comparison, the degree of inhibition displayed by the variants exactly confirmed the above predictions. Glutamine substituting for residues from regions 1 and 2 (Lys 69 , Phe 71 , Phe 88 , Phe 91 ) caused a loss of inhibition (Fig. 6A), whereas  replacement by tyrosine (where tested) had no or only little effect on the inhibitory potential (Fig. 6B). Similar results were obtained with variants carrying mutations in region 3. Although the F122I mutant protein still caused some inhibition, variants carrying F122Q or L127T mutations completely failed to inhibit the ATPase activity of MalFGK 2 (Fig. 7A). Furthermore, mutations of the C-terminal residues Arg 165 and Lys 167 had no effect on the capability of the variants to inhibit ATPase activity (Fig. 7B).
This result is consistent with the above notion that the C-terminal portion of the protein is not involved in the interaction with MalK.
To exclude that the mutations, in particular those that caused loss of inhibition, might have affected EIIA Glc integrity, we analyzed the ability of the variants to accept a phosphoryl group from HPr in vitro. The assay is based on the observation that phosphorylated EIIA Glc is migrating slower in SDS gels than its unphosphorylated form (23). As shown in Figs. 6 and 7 (C and D) all variants behaved like the wild type, whereas a control carrying a mutation of the catalytic His 90 residue (3 A) failed to become phosphorylated. Thus, we conclude that none of the mutations affected the structural integrity of EIIA Glc . Interestingly, the H90A variant also failed to inhibit the ATPase activity of the maltose transporter (not shown), but here the possibility of a structural alteration of the protein must be taken into account. Together, these data not only nicely confirm the results from substitutional analyses (Fig. 2) but support the newly discovered hydrophobic patch constituted by ␤11 and ␣2 as being involved in interaction with MalK.
A Peptide Encompassing ␤5-␤7 Partially Inhibits ATPase Activity of MalFGK 2 -Next, we wished to examine whether one of the two identified binding sites would be sufficient to allow at least partial inhibition of maltose transporter activity. To this end, a soluble peptide (␤5-7) was synthesized encompassing residues Thr 66 -Val 96 . Monitoring the ATPase activity of MalFGK 2 in proteoliposomes in the presence of the ␤5-7 peptide revealed a moderate but consistently observed inhibition by 20% as compared with the control (Fig. 8C). In contrast, a control peptide from the C-terminal region of the protein (residues Thr 136 -Ile 166 ) that was not identified by the pep scan approach to be involved in MalK binding (see Fig. 1) was uneffective, thereby making an unspecific reaction rather unlikely. Thus, the data suggest that a functional interaction of the ␤5-␤7 peptide with the MalK subunits had occurred, which is in further support of the above conclusions.

The N-terminal Region of EIIA Glc Is Required for Inhibition of Maltose
Transporter Activity-The N-terminal region of EIIA Glc is not required to accept a phosphorous from the donor protein HPr but is essential for effective phosphotransfer to the membrane-bound enzyme IIBC Glc (24,25). Structural studies suggested that the N-terminal tail confers amphitropism to the protein, allowing EIIA Glc to shuttle between the cytoplasm and the membrane (26). Whether the N-terminal peptide is cru-cial for binding to target proteins in the context of inducer exclusion is unknown. Thus, we examined the inhibitory potential of a purified EIIA Glc variant lacking amino acids Gly 1 -Asp 16 (Fig. 5B) on the reconstituted maltose transport complex. As shown in Fig. 8A, the deletion caused a complete loss of inhibition even when the truncated protein was added in a 10-fold higher concentration than wild type. Lack of the N-terminal 17 residues did not otherwise affect the  functional state of the protein because time-dependent phosphorylation of the protein comparable to wild type could be demonstrated (Fig. 8B).
Interestingly, a naturally occurring variant of EIIA Glc that lacks only the N-terminal residues Gly 1 -Lys 7 (24) was previously shown to be less inhibitory on lactose transport in everted membrane vesicles than the native protein (27). Thus, we also monitored ATPase activity of the maltose transporter in the presence of purified recombinant EIIA Glc (⌬1-7). The result clearly revealed that even this variant had completely lost its capability to inhibit ATP hydrolysis catalyzed by MalFGK 2 (data not shown).
Together, we conclude that binding of EIIA Glc to complex-assembled MalK in vivo requires association with the membrane via the N-terminal peptide. This finding might also explain why the ATPase activity of soluble MalK is only poorly inhibited by EIIA Glc (13).
The above observation that the ␤5-␤7 peptide moderately affects ATPase activity in the absence of the N-terminal tail might thus be explained by better access of the small peptide to the target site as compared with the intact protein. Unfortunately, the question whether a fusion of ␤5-␤7 to a peptide encompassing the N-terminal domain would increase the inhibitory potential could not be addressed because attempts to synthesize such a peptide were unsuccessful.

DISCUSSION
The signal-transducing protein EIIA Glc interacts with multiple target proteins within the context of glucose uptake, catabolite repression, and inducer exclusion (1). These interactions are either catalytic involving phosphotransfer to or from EIIA Glc (in the case of HPr and EIIB Glc ) or regulatory, being dependent on the phosphorylation state of EIIA Glc (in the case of glycerol kinase, lactose permease, and the maltose ATP-binding cassette transporter). In the work presented here we provide first evidence for the structural requirements of EIIA Glc to recognize the MalK subunits of the maltose ATP-binding cassette transporter.
In the absence of co-crystals between EIIA Glc and the maltose transporter we used knowledge-based peptide arrays synthesized on coherent membranes to map the potential binding site(s) between EIIA Glc and MalK. Over the past decade peptide arrays have become a powerful tool to study molecular recognition events (28), and the reliability of the data has been shown (29,30). The approach is superior over conventional mutational analyses or chemical modification, as employed in the case of lactose permease (9), because peptide arrays and subsequent substitutional analyses provide a complete data set on the residues in question for interaction with the target. Moreover, as demonstrated in this study, the results can perfectly guide a subsequent mutational analysis of functions in the context of the folded protein. By combining both approaches we have identified two putative binding sites located on opposite sites on the surface of EIIA Glc (Fig. 4). One comprises residues from ␤-strands 5-7, whereas the other involves residues located on ␤11 and ␣2 (Fig. 3). The latter region has not been implicated yet in binding of other target proteins and might thus be unique for interaction with the MalK subunits of the maltose transporter.
Structural analyses of complexes between EIIA Glc and several target proteins have revealed that among others the region encompassing ␤-strands 5-7, in particular, residues Lys 69 , Phe 71 , His 90 , Asp 94 , and Glu 97 are crucial for binding of HPr (8), EIIB Glc (6,7), and glycerol kinase (5). Moreover, in a study that combined chemical modification with mutational analysis it was suggested that ␤5-␤7 is also required for interaction of EIIA Glc with lactose permease (9).
The data presented here (see Figs. 2A and 6) also indicate a key role of Lys 69 , although the positive charge seems not to be required as substitution by hydrophobic residues still allowed binding of MalK. This finding rather excludes the possibility that Lys 69 forms a salt bridge with a negatively charged residue in MalK. Our results are also in agreement with data reported for lactose permease. Although the K69E variant was impaired in binding lactose permease, the K69L mutant protein exhibited binding activity comparable to wild type (9). These observations led the authors to speculate that a negative charge at position 69, which is in close proximity to the catalytic histidine 90 might mimic the phosphorylated and, thus, inactive state of the protein with respect to inducer exclusion (9). Our finding that substitution of Lys 69 by polar but uncharged residues also eliminated binding, and, consequently, as shown for K69Q, inhibition of ATPase activity of MalFGK 2 is not in support of this notion.
Phe 71 is also crucial for binding of MalK in that it can be replaced only by residues with other aromatic side chains. Again, this result is consistent with the observation that variants F71K and F71S had lost their capacity to bind to lactose permease (9). In contrast, other residues from region ␤5-␤7 that when mutated affect interaction with lactose permease like Ala 76 (T), and Ser 78 (F) (31) seem not be required for contacting MalK.
Similarly, no interaction was found with the peptide region around ␣-helix 1, which adds to the binding of EIIA Glc to HPr, EIIB Glc , and glycerol kinase (see Figs. 1 and 3). In particular, Asp 38 and Glu 43 form ion pairs with Arg 479 and Arg 402 , respectively, of glycerol kinase (5). In addition, Asp 38 together with Asp 94 is involved in salt bridges formed with Arg 38 /Arg 40 of EIIB Glc (7). Asp 94 was included in the complete substitutional analyses presented here and could be replaced by any other amino acid without eliminating MalK binding (see Fig. 2B). Also in the case of lactose permease, a D94G variant of EIIA Glc still displayed 60% binding activity compared with wild type.
Together, these observations clearly suggest that the binding sites of EIIA Glc for different target proteins overlap but are not identical. In particular, the structural requirements with respect to the catalytic activity of EIIA Glc in the context of glucose transport are in part distinct from those that promote regulatory interactions.
Can we draw any conclusion from these results on the interacting residues on MalK? Previous analyses have revealed that mutations conferring resistance to EIIA Glc inhibition are located in two regions in the helical domain and the C-terminal domain of MalK, respectively (12,14,15). Although additional evidence for the C-terminal domain being involved in EIIA Glc binding was subsequently obtained by competition experiments employing monoclonal antibodies, the same approach did not confirm the involvement of the two residues (Glu 119 , Ala 124 ) from the helical domain (13). However, Samanta et al. (32) noticed that both clusters lie on the same face of the MalK dimer and might nonetheless form a binding site for EIIA Glc . The authors speculated that EIIA Glc might prevent the two N-terminal domains of MalK to move into closer proximity as a consequence of ATP binding (33), thereby arresting the transport cycle. The identification of two binding sites for MalK on EIIA Glc is consistent with this attractive hypothesis but clearly requires further proof. Thus, experiments addressing this question are underway in this laboratory.