Liganded and unliganded receptors interact with equal affinity with the membrane complex of periplasmic permeases, a subfamily of traffic ATPases.

The histidine-binding protein, HisJ, is the soluble receptor for the periplasmic histidine permease of Salmonella typhimurium. The receptor binds the substrate in the periplasm, interacts with the membrane-bound complex, transmits a transmembrane signal to hydrolyze ATP, and releases the ligand for translocation. HisJ, like other periplasmic receptors, has two lobes that are apart in the unliganded structure (open conformation) and drawn close together in the liganded structure (closed conformation), burying deeply the ligand. Such receptors are postulated to interact with the membrane-bound complex with high affinity in their liganded conformation, and, upon substrate translocation, to undergo a reduction in affinity and therefore be released. Here we show that in contrast to the current postulate, liganded and unliganded receptors have equal affinity for the membrane-bound complex. The affinity is measured both by chemical cross-linking and co-sedimentation procedures. An ATPase activity assay is also used to demonstrate the interaction of unliganded receptor with the membrane-bound complex. These findings support a new model for the transport mechanism, in which the soluble receptor functions independently of the commonly accepted high-low affinity switch.

The histidine-binding protein, HisJ, is the soluble receptor for the periplasmic histidine permease of Salmonella typhimurium. The receptor binds the substrate in the periplasm, interacts with the membrane-bound complex, transmits a transmembrane signal to hydrolyze ATP, and releases the ligand for translocation. HisJ, like other periplasmic receptors, has two lobes that are apart in the unliganded structure (open conformation) and drawn close together in the liganded structure (closed conformation), burying deeply the ligand. Such receptors are postulated to interact with the membranebound complex with high affinity in their liganded conformation, and, upon substrate translocation, to undergo a reduction in affinity and therefore be released. Here we show that in contrast to the current postulate, liganded and unliganded receptors have equal affinity for the membrane-bound complex. The affinity is measured both by chemical cross-linking and co-sedimentation procedures. An ATPase activity assay is also used to demonstrate the interaction of unliganded receptor with the membrane-bound complex. These findings support a new model for the transport mechanism, in which the soluble receptor functions independently of the commonly accepted high-low affinity switch.
Bacterial periplasmic permeases belong to a superfamily of transporters (traffic ATPases (1) or ABC proteins (2)) that comprises many prokaryotic and eukaryotic carriers, including the medically important cystic fibrosis transmembrane conductance regulator (CFTR) and the multidrug resistance Pglycoprotein (MDR) (3,4). Because of the ease of manipulation, prokaryotic permeases are good model systems for understanding the eukaryotic carriers. Periplasmic permeases are composed of a soluble receptor (the periplasmic substrate-binding protein) and a membrane-bound complex, which translocates the substrate with concomitant ATP hydrolysis. Models for their mechanism of action generally depict the liganded receptor as interacting with high affinity with the membrane-bound complex and stimulating ATP hydrolysis by the ATP-binding subunits, which leads to the discharge and translocation of substrate; the unliganded receptor is then released because its affinity for the complex is reduced (reviewed in Ref. 3). The high-low switch in the affinity of the receptor for the membrane complex is central to such models. The histidine permease of Salmonella typhimurium (and Escherichia coli) is a well characterized model system. It consists of the periplasmic receptor, the histidine-binding protein HisJ, and of the four subunits of the membrane-bound complex, HisQ/M/P, (with two integral membrane proteins, HisQ and HisM, and two identical ATPbinding subunits, HisP) 1 (3). HisJ interacts with the membrane-bound complex, as shown indirectly by genetic analysis (5); chemical cross-linking experiments provided direct evidence for a specific interaction between HisJ and HisQ (6).
The structures of many periplasmic binding proteins have been resolved, and despite the fact that these proteins do not usually bear sequence homology, their structures are similar (7). The three-dimensional structure of liganded HisJ and of the homologous lysine-, arginine-, ornithine-binding protein, LAO, shows the existence of two globular domains delimiting a cleft containing the substrate-binding pocket (8,9). The lobes are separate when the protein is unliganded (open conformation) and are positioned much closer to each other in the liganded form (closed conformation) in which the substrate is deeply buried. These receptors undergo a spontaneous conformational change in the absence of ligand, continuously "breathing" and assuming a closed empty form (10 -12), as was demonstrated for HisJ in solution using conformation-specific monoclonal antibodies (10). This finding has raised the possibility that because the receptor in its closed empty form resembles the closed liganded form, it would compete in vivo with the liganded receptor for interaction with the membrane-bound complex. Such a situation would be disadvantageous because transport would be inhibited at low substrate concentrations, when there would be an excess of unliganded over liganded receptor in the periplasm. In addition, in the absence of substrate, the empty closed form interacting with the complex might stimulate a wasteful ATP hydrolysis. Therefore, the question of whether the unliganded receptor indeed interacts with the complex has to be addressed.
We previously observed that unliganded HisJ competes (effectively) with liganded HisJ in histidine transport assays using an in vitro reconstituted system (13). We suggested at the time that the unliganded protein must interact with the same site on the membrane-bound complex as liganded HisJ. Here we have undertaken a detailed study of the interaction between HisJ, both liganded and unliganded, and the membrane complex. We optimized the previously developed cross-linking methods (6) and introduced additional cross-linking reagents. We have also developed a novel approach to studying this interaction that utilizes the signaling properties of the receptor to induce ATP hydrolysis. We show that unliganded receptor indeed interacts with the membrane complex and does so as well as liganded receptor. Therefore, we challenge the conventional model and propose a different one in which the operation of the receptor is independent of a high-low switch in affinity. The model we propose is relevant for understanding the mechanism of action of several important eukaryotic membrane receptors that contain extracellular substrate-binding domains that are homologous to periplasmic receptors (14 -16).

EXPERIMENTAL PROCEDURES
Purification of Receptors and Removal of Bound Ligands-HisJ was purified to over 90% purity (as judged by SDS-PAGE) from strain GA383 (S. typhimurium LT2 carrying plasmid pFA54, which produces the HisJ protein under the control of the Tac promoter). Bound ligand was removed from HisJ, and its absence was verified as described (17). Alternatively, bound ligand was removed as follows: i) the receptor was denatured by exposure to 6 M guanidine HCl and then renatured by extensive dialysis (18); ii) a dialysis bag containing 15 mg of receptor per ml was submerged in 50 mM KP i , pH 7.5 (400 ml, replaced twice) in an electrophoretic tank and electrophoresed for 8 -16 h at room temperature (at 30 mAmp constant current). LAO was purified as described (17); to remove bound ligand from purified LAO, bound substrate was first replaced with histidine by the addition of 10 mM histidine followed by dialysis for 10 h at 4°C against 500 ml of 5 mM histidine in potassium acetate buffer, pH 5.0, with four changes; then histidine was removed as above. The removal of ligand was followed by visualizing either residual histidine or arginine with the respective reagents (19,20); the removal of arginine was also followed by the loss of radioactive arginine added to the protein at the beginning of the treatment. Receptors were considered unliganded if they contained less than 5% ligand.
Cross-linking with Sulfo-SANPAH-Purified de-liganded receptors were derivatized with Sulfo-SANPAH (Pierce) (then liganded with 400 M substrate, when necessary) and cross-linked to right side out membrane vesicles as described (6), with the following modifications. After derivatization, excess Sulfo-SANPAH was quenched by adding Tris/Cl to a final concentration of 50 mM, pH 7.0, incubating for 10 min at room temperature; the derivatized protein was purified on a Sephadex G-50 spin column (100 l of reaction mixture/ml of packed volume) equilibrated with 10 mM, pH 7.0 KP i buffer. Samples were kept in the dark until photoactivation. For photoactivation, samples were scaled down to a final volume of 0.1 ml, and five full-strength flashes were delivered with a Minolta Auto 200X flash held 6 cm from the surface. Samples were then centrifuged for 15 min at 13,000 rpm in a Sigmafuge, the pellets were washed once with 0.1 ml of 50 mM Tris/Cl, pH 7.5, and dissolved in 50 l of Laemmli sample buffer without heating (with the ␤-mercaptoethanol omitted), and 15 l were immediately resolved by SDS-PAGE (11%, pH 8.8). The gels were immunoblotted using Immobilon-P membranes (Millipore, Bedford, MA) and polyclonal anti-HisQ antibody (21) with the following modifications: the blocking reagent contained 2% gelatin in 10 mM Tris/Cl, pH 7.0 buffer, 500 mM NaCl, 0.3% Tween 20. Gelatin 1% was present also in the antibody dilution buffer. Second antibody conjugated with horseradish peroxidase (Bio-Rad) and an Enhanced Chemiluminescence kit (Amersham Corp.) were used to develop the immunoblots; the relative intensities of the responses were quantitated as described (22) or using a scanning densitometer (Molecular Dynamics, Sunnyvale, CA).
Cross-linking with Formaldehyde-For in vivo cross-linking, bacteria were grown at 30°C in LB medium containing ampicillin to an optical density of 0.6 (at 650 nm), when the temperature was raised to 42°C for 60 min. The cells were harvested, washed once with and resuspended in minimal medium E (23) containing glucose, incubated with aeration for 30 min, harvested, and washed once with and resuspended in 0.25 M NaP i buffer, pH 5.0, and the absorbance at 650 nm was adjusted to 1.0. An aliquot (typically 0.2 ml) was placed in an Eppendorf tube and incubated in a 37°C water bath for 3 min, after which the indicated amount of histidine was added. After 3 more min of incubation, 37% formaldehyde (Fisher Scientific) was added to a final concentration of 1%, and incubation continued without shaking for 5 min. The reaction was then quenched by the addition of glycine to a final concentration of 300 mM. Cells were centrifuged, washed once with and resuspended in 10 mM NaP i buffer, pH 6.8, at the same concentration, appropriate amounts were mixed with Laemmli sample buffer (24), incubated at 65°C for 10 min, and resolved by SDS-PAGE. Immunoblotting was performed as described for the Sulfo-SANPAH experiments, with polyclonal anti-HisJ and horseradish peroxidase-conjugated second antibody using the Enhanced Chemiluminescence kit. For in vitro cross-linking 50 l of proteoliposomes (PLS) were mixed at 37°C with 10 l of a solution containing ATP and MgCl 2 to give final concentrations of 4 and 20 mM respectively; then pure unliganded HisJ was added to a final concentration of 20 M together with L-histidine as indicated; 37% formaldehyde (1.6 l) was added to give a final concentration of 1%, and the incubation continued at 37°C for 10 min, when the reaction was quenched by the addition of 3 l of 5 M ethanolamine. After 30 min at room temperature, the PLS were diluted with 10 mM NaP i , pH 6.8, to 1.3 ml, harvested, and then analyzed as for the in vivo experiments.
Affinity Measured by Co-sedimentation-Membranes from heat-induced TA3662 (which carries the hisQ, hisM, and hisP genes under the temperature-sensitive control of the lambda P L promoter) and heatinduced TA3663 (in which the genes are carried in inverted orientation and therefore are not heat inducible) (28) were mixed in an Eppendorf tube with HisJ (5, 2.5, 1.25, or 0.63 M final concentrations) in 50 mM NaP i buffer, pH 7.0, 1 mM L-histidine, and 150 mM NaCl in a total volume of 0.1 ml. After 30 min of incubation at room temperature, the samples were centrifuged at 45,000 rpm for 30 min, the supernatant was removed by aspiration, the pellet was rinsed gently with 1 ml of NaP i buffer, and the wall of the tube was cleaned with a cotton swipe. Pellet proteins were resolved by SDS-PAGE, and HisJ was quantitated by immunoblotting with polyclonal anti-HisJ antibody using pure HisJ as standard (22). Because of the relatively poor affinity of HisJ for the complex, the amount of HisJ bound to membranes in the presence of HisQ/M/P is approximately double the amount bound to membranes lacking HisQ/M/P. Thus, to obtain an accurate estimate of the affinity value, data from three different experiments, each in triplicate, were averaged.
Measurement of ATPase Activity-Various amounts of HisJ, in the presence or the absence of equivalent amounts of L-histidine, were added to reconstituted PLS (in 50 mM MOPS/K ϩ , pH 7.5; final protein concentration: 0.2-0.4 mg/ml) and incubated at 37°C for 5 min. ATP hydrolysis was initiated by the addition of ATP and MgSO 4 (2 and 10 mM final concentration, respectively); at various times 100-l aliquots were mixed with equal volumes of 12% SDS, and the inorganic phosphate released was determined as described (29).

Liganded and Unliganded HisJ Interact Equally Well with
the Membrane Complex-Cross-linking with the light-activated heterobifunctional reagent Sulfo-SANPAH and with formaldehyde were shown previously to yield a cross-linked product between HisJ and HisQ, JϳQ (6). Here we show that this method can be used to determine the affinity of liganded HisJ for the membrane complex by quantitating the amount of JϳQ produced with increasing concentrations of Sulfo-SANPAHderivatized liganded HisJ. Fig. 1A shows that JϳQ increases until it reaches a maximum, as expected for an equilibrium binding process, yielding an apparent K d for the interaction of 15 M. Membranes from a mutant lacking the membrane complex do not yield JϳQ (data not shown and Ref. 6).
The apparent affinity of underivatized liganded HisJ was also determined using a co-sedimentation assay, which gave a value of 19 M. This result indicates that derivatization does not affect the nature of the interaction between HisJ and the complex. It is also consistent with the value of 12-40 M (depending on the calculation method used; see legend to Table  I) calculated for underivatized liganded HisJ in experiments in which it had been used as a competitor for cross-linking (30). The apparent affinity of liganded HisJ was also measured by a transport assay utilizing reconstituted PLS and was determined to be 8 M, 2 which is not significantly different from the above values. In summary, the various values obtained for underivatized HisJ are essentially indistinguishable from that obtained directly for the Sulfo-SANPAH-derivatized protein. It can be concluded that: i) the cross-linked product results from a specific interaction between the derivatized form of liganded HisJ and a binding site(s) on HisQ; ii) the underivatized form of liganded HisJ competes for the same site(s); and iii) derivatization does not greatly perturb the structure of HisJ (6). Therefore, affinities can be determined either by directly measuring the level of cross-linking using the derivatized protein or by competition of cross-linking using underivatized protein.
Sulfo-SANPAH-derivatized unliganded HisJ was then examined for its interaction with HisQ/M/P, and it was also found to be cross-linked. Surprisingly, its apparent affinity is 6 M, a value not significantly different from that obtained for liganded HisJ (Fig. 1B). This finding is consistent with puzzling results obtained previously, which showed that unliganded HisJ competes efficiently with liganded HisJ, thus inhibiting transport, in experiments using an in vitro reconstituted system (13,30). The apparent K i values for unliganded HisJ are calculated to be 22 and 13 M, respectively, which are very similar to those obtained by cross-linking of either liganded or unliganded derivatized HisJ. This also is evidence that cross-linking reflects a physiological interaction.
Because the finding that liganded and unliganded HisJ interact with similar affinities with the membrane complex was unexpected, it was important to determine whether this phenomenon is peculiar to HisJ. Fig. 1 (C and D) shows that also derivatized LAO, unliganded and lysine-liganded, can be crosslinked with HisQ and have comparable apparent affinities (K d values of 12 and 22 M, respectively). Unliganded and arginineliganded underivatized LAO have also been shown to compete with derivatized histidine-liganded HisJ with apparent K i values of 8 and 11 M, respectively (Fig. 1, E and F). Finally, liganded and unliganded LAO compete equally well with HisJ in a transport assay (30). Table I summarizes various combinations tested: all the K d and K i values are essentially the same.
In conclusion, contrary to the assumption made in all current models in which only the liganded form of the protein interacts efficiently with the membrane-bound complex, liganded and unliganded HisJ interact with the same apparent affinity with a site on the membrane, a site that is on the normal pathway leading to transport.
The Mode of Interaction Is Different for Liganded and Unliganded Receptor-Despite the fact that the affinities of the two forms of the receptor for the membrane are very similar, it became clear that the character of the interaction is different. An early indication of this fact was that in any one experiment the level of JϳQ produced in the presence of excess histidine was consistently 2-4-fold higher than in the absence of histidine (using either 15 or 50 M HisJ) (data not shown). Thus,  is the amount prevented from binding divided by the cross-linked product obtained in the absence of inhibitor; [I 0 ] is the initial competitor concentration (i.e., underivatized liganded HisJ); [S 0 ] is the initial concentration of derivatized liganded HisJ; and K m is the known affinity of the substrate (same as the K d in this case) (45) , where S is the fixed concentration of substrate (derivatized liganded HisJ), I is the competitive inhibitor, and v is the fraction of S that has been cross-linked. If we assume that K m ϭ K i , the latter equation can be reduced to: liganded and unliganded HisJ have very similar affinity, but the conformation taken by the liganded protein appears to result in better access of the derivatized reactive group(s) on HisJ to the interacting group(s) on HisQ. The specific nature of the interaction was therefore studied by a completely different approach: the ability of the two conformations to send a signal for ATP hydrolysis. In addition, the effectiveness of a variety of chemical cross-linking reagents to yield JϳQ was also analyzed.
It has been shown that liganded receptors send signals to the ATP-binding site (on the cytoplasmic face of the membranebound complex), activating it to hydrolyze ATP (25,31,32). Because this signal depends on the prior interaction between the receptor and the membrane-bound complex, such an activity could be used as an assay for the interaction. Investigations on the specific nature of the signal had demonstrated that also unliganded HisJ stimulates ATP hydrolysis, although at a lower level (in the case of the maltose receptor) (32). 3 A more detailed analysis shows the relationship between various concentrations of liganded or unliganded HisJ and ATP hydrolysis in PLS reconstituted with HisQ/M/P; both forms stimulate ATP hydrolysis (Fig. 2). The possibility that the stimulation was due to HisJ being liganded because of contaminating histidine was carefully examined, because these proteins retain their ligands with well known tenacity (18). The HisJ preparations (over 90% purity) used were shown to contain less than 0.1 mol of histidine/mol of HisJ by thin layer chromatography (17). As an additional precaution, it was shown that after ligation of pure HisJ with radioactive L-histidine and repurification by HPLC, less than 0.01% of the radioactivity remained, i.e. less than 6 pmol of histidine/mol of HisJ. The possible presence of histidine in reconstituted PLS was considered irrelevant because their preparation involves extensive dilutions steps which, by themselves, would drop the histidine concentration to less than 10 nM in the assay. Thus, histidine was determined to be less than 1 nM in solutions of 160 M HisJ and less than 10 nM in the PLS. In addition, the ATPase activity is dependent on the concentration of HisJ, which excludes the possibility that histidine contaminates the PLS preparation. The stimulation can therefore be safely ascribed to unliganded HisJ. The apparent K m values for ATPase stimulation are essentially the same for liganded and unliganded HisJ (11 and 5 M, respectively). Although unliganded HisJ can take a closed empty conformation mimicking that of the closed liganded conformation, it has been estimated that the open empty form predominates (10). The fact that the V max values in Fig. 2 are different (0.92 and 0.07 mol/min/mg protein for liganded and unliganded HisJ, respectively) excludes the possibility that the form interacting and sending the signal is only a fraction of the added HisJ, which mimics the closed liganded conformation. If the latter were the case, at sufficiently high concentration of unliganded HisJ, the V max should be the same as for liganded HisJ. An important question is why unliganded receptor stimulates ATP hydrolysis. An attractive hypothesis is that unliganded receptor induces slow ATP hydrolysis to convert the membrane complex to a state of readiness for transport; such a state might possibly raise the affinity of the receptor for the ligand. Preliminary results suggest that this may indeed be the case. 4 Because different surface groups would be exposed in the two conformations, cross-linking reagents with a variety of arm lengths also could discriminate between the two types of interaction. The effect of liganding on cross-linking was first tested with formaldehyde. This reagent has been shown to be an effective cross-linking reagent in vivo (6). We first demonstrated that cross-linking does not occur artifactually as a consequence of HisQ/M/P overproduction by the plasmid-borne genes, because it is formed normally in a strain that produces chromosomal levels of both HisJ and HisQ/M/P (it is absent in a strain lacking HisQ/M/P; data not shown). Fig. 3 (top inset,  lanes 1 and 2) show that JϳQ is formed efficiently in vivo only in the presence of histidine. Varying the L-histidine concentration shows that JϳQ is not formed in the absence of histidine and increases in its presence until saturation is reached; a reciprocal plot of these data yields an apparent K d of 56 nM (bottom inset). This value is similar to the K d of HisJ for histidine binding (30 nM; Ref. 33), which indicates that the limiting step in the in vivo cross-linking reaction is the ligation of HisJ. Similar results were obtained when the fusion protein J-LAO (which behaves like LAO (34)), liganded with arginine or histidine, was used as the receptor (data not shown).
To exclude the possibility that the above results reflect a different behavior of the receptor in vivo and in vitro, an adaptation of the formaldehyde cross-linking method was developed to be used in vitro with reconstituted PLS. 2 Fig. 3 (top inset,  lanes 3 and 4) shows that also in reconstituted PLS JϳQ is dependent on the presence of histidine; a K d value of 30 M was obtained by varying the concentration of histidine-liganded HisJ (data not shown). A possible effect of the polymerization state of formaldehyde on the efficiency of cross-linking was excluded because autoclaving a 22% solution of formaldehyde in 0.1 M NaP i , pH 6.8 buffer (35) for 20 min before use had no effect. Thus, the difference between cross-linking with formal- dehyde and Sulfo-SANPAH is not due to the particular assay system used. Rather, because formaldehyde is a small crosslinking molecule (36), it seems reasonable that it would detect a different kind of interaction than a larger molecule, such as Sulfo-SANPAH. On the basis of these results, we tested a number of cross-linking reagents with different arm lengths. Table II shows that unliganded HisJ can be cross-linked by reagents with an arm length of 18 Å, but not by formaldehyde and DFDNB, which have short arm lengths (2 and 3 Å, respec-tively). Reagents with intermediate arm lengths fail to crosslink, whether histidine is present or not. Therefore, the interaction between unliganded HisJ and HisQ can be distinguished from that with liganded HisJ also by the use of appropriate cross-linking reagents.
In conclusion, both cross-linking reagents and induction of ATP hydrolysis provide evidence that liganded and unliganded receptor are in different conformations while interacting with HisQ/M/P. The nature of these conformations will be discussed below. DISCUSSION We have demonstrated that unliganded HisJ interacts with the membrane-bound complex as well as liganded HisJ. We have also shown that liganded and unliganded receptor are in different conformations while interacting with HisQ/M/P. The latter is not surprising, considering the known structures of these two forms (8). However, because unliganded receptor can assume a closed empty conformation spontaneously (10 -12), the question arises as to the nature of its conformation while it interacts with the complex. The data presented show that it cannot be the closed conformation that is normally assumed by liganded receptor. Therefore, the conformation of the interacting unliganded receptor is referred to as "open conformation." Such an open conformation includes the fully open one (that has been resolved for several unliganded structures by x-ray crystallography (7,8,37)) and, presumably, intermediate conformations assumed during breathing because of the dynamic state of the conformations of the receptor.
Here we propose a new model for the mechanism of action of periplasmic permeases (described in Fig. 4). The receptor is associated with the membrane complex even in the absence of ligand, and in this condition it stimulates a low rate of ATP hydrolysis (state I). Upon ligand binding, the receptor assumes the closed form (state II), which induces fast ATP hydrolysis and leads to translocation (state III). The novel feature introduced in this model is that the receptor does not leave the membrane-bound complex after translocation, as required by current "high-low affinity switch" transport models. This model is consistent with past evidence that unliganded HisJ competes effectively with liganded HisJ in histidine transport; the implications of those results were discussed at the time (13). It is also consistent with proposals derived from theoretical calculations, which suggested that both liganded and unliganded receptors interact with the membrane complex (38,39).
What are the biological implications of this model? If the de-liganded receptor does not leave the membrane-bound complex (although this interaction is dynamic), it is not necessary for the complex to recruit another liganded molecule. This is advantageous because the periplasm appears to be a thick gel  (46); incubation after the addition of histidine was decreased to 7 s, and after the addition of formaldehyde the incubation time was 2 min. Before resolution by SDS-PAGE, samples were concentrated to an absorbance at 650 nm of 1.0 by filtration. The top inset shows immunoblots of in vivo (first two lanes) and in vitro (last two lanes) cross-linking experiments. Lanes 1 and 2, strain GA96 (⌬hisF645 his⌬5575 (deleting the hisQ, hisM, and hisP genes) and dhuA1, a promoter-up mutation in the histidine permease operon (47)  The respective arm lengths have been obtained from the following references: formaldehyde (36); 1,5-difluoro-2,4-dinitrobenzene (DFDNB), disulfosuccinimidyl tartarate (SulfoDST), dimethyl adipimidate dihydrochloride (DMA), and dimethyl superimidate dihydrochloride (DMS) (Pierce Catalogue); glutaraldehyde (estimated from formula); Sulfo-SANPAH (30); and sulfosuccinimidyl 2-(p-azidosalicyl-amido) ethyl-1,3Ј-dithiopropionate (SASD) (27).
b The extent of cross-linking (obtained by quantitation as described under "Experimental Procedures") in the absence of histidine is expressed as a percentage of the amount obtained in the presence of 100 M L-histidine. (40), and periplasmic proteins are essentially immobilized (41); therefore, the diffusion of liganded receptor would be inadequate. Instead, a ligand molecule diffuses much faster than its receptor, being considerably smaller. Thus, the receptor is unlikely to physically carry the ligand through the periplasm. In the new model, de-liganded receptor molecules remaining on the complex would be poised to capture repetitively free ligand for sequential cycles of translocation.
Receptors are present in the periplasm in very high concentration (usually 1-10 mM) and are in large excess over the membrane-bound complex (over 30-fold for the histidine permease (42)). If transport can function through a single complexassociated receptor molecule, what would be the function of all the other receptor molecules? In fact, the presence of high concentration of receptor in the periplasm facilitates transport: the combination of high receptor concentration with high affinity for the ligand would effectively convert low external ligand concentrations into high concentrations of bound ligand in the periplasm; the high receptor concentration also increases retention of ligand (as has been indeed demonstrated both theoretically and experimentally (43)), and the retention effect would be an advantage when the external ligand concentration drops, because it maintains the periplasmic ligand concentration high for a long time (44). Thus, the presence of high concentrations of receptor in the periplasm helps recruit and provide ligand for an open unliganded receptor molecule bound to the complex, which immediately translocates it. Several eukaryotic receptors have been shown to possess an extracellular domain that is homologous to periplasmic receptors: the metabotropic glutamate receptor in the signal transduction pathway (14,15) and the calcium receptor of parathyroid cells (16). This domain has been modeled to have the structure of and to function like a periplasmic receptor (the LAO protein, in the case of the glutamate receptor (15)), capturing the substrate and signaling receptor occupancy to its membrane-embedded portion. The validity of this model depends on the ability of the substrate-binding domain to perform its function while being immobilized onto the membranous portion of the molecule. This notion is supported by our results indicating that this is likely the normal mode of action for periplasmic receptors. Indeed, in view of the findings and the model presented here, it is tempting to speculate that the binding and sensing/signaling elements belonging to bacterial permeases receptors have been adopted through evolution by eukaryotic receptors and incorporated into structures unrelated to traffic ATPases (in the above cases, an ion channel and a G protein-coupled receptor, respectively). Because soluble receptors have not yet been identified in eukaryotic traffic ATPases, it is possible that also in these systems a similar substrate-binding and signaling domain may have been incorporated into their membranous portion (at least in some cases). Understanding the mechanism of action of the periplasmic receptors is likely to yield useful information relative to the function of the eukaryotic systems. The hatched box represents the membrane-bound HisQ/M/P complex. HisJ is represented in solid black and is shown to undergo a conformational change upon binding of histidine (a spotted circle) and during translocation; small blank squares and circles represent cross-linking sites for reagents with different arm lengths, respectively. The details of the mechanism are described in the text.