Dependence of Maltose Transport and Chemotaxis on the Amount of Maltose-binding Protein*

Maltose-binding protein (MBP) is essential for mal- tose transport and chemotaxis in Escherichia coli. To perform these functions it must interact with two sets of cytoplasmic membrane proteins, the MalFGK transport complex and the chemotactic signal transducer Tar. MBP is present at high concentrations, on the order of 1 mM, in the periplasm of maltose-induced or malP constitutive cells. To determine how the amount of MBP affects transport and taxis, we utilized a series of malE signal-sequence mutations that interfere with export of MBP. The MBP content in shock fluid from cells carrying the various mutations ranged from 4 to 23% of the malE+ level. The apparent K , for maltose transport varied by less than a factor of 2 among malE+ and mutant strains. At a saturating maltose concentra- tion 9% (-90 p ~ ) of the malE+ amount of MBP was required for half-maximal uptake rates. Transport ex- hibited a sigmoidal dependence on the amount of periplasmic MBP, indicating that MBP may be involved in a cooperative interaction at some stage of the transport process. The chemotactic response to a saturating mal- tose stimulus exhibited a first-order dependence on the amount of periplasmic MBP. Thus, interaction of a single substrate-bound MBP with Tar appears suffi-cient to initiate a chemotactic signal from the

Maltose-binding protein (MBP) is essential for maltose transport and chemotaxis in Escherichia coli. To perform these functions it must interact with two sets of cytoplasmic membrane proteins, the MalFGK transport complex and the chemotactic signal transducer Tar. MBP is present at high concentrations, on the order of 1 mM, in the periplasm of maltose-induced or m a l P constitutive cells. To determine how the amount of MBP affects transport and taxis, we utilized a series of malE signal-sequence mutations that interfere with export of MBP. The MBP content in shock fluid from cells carrying the various mutations ranged from 4 to 23% of the malE+ level. The apparent K , for maltose transport varied by less than a factor of 2 among malE+ and mutant strains. At a saturating maltose concentration 9% (-90 p~) of the malE+ amount of MBP was required for half-maximal uptake rates. Transport exhibited a sigmoidal dependence on the amount of periplasmic MBP, indicating that MBP may be involved in a cooperative interaction at some stage of the transport process. The chemotactic response to a saturating maltose stimulus exhibited a first-order dependence on the amount of periplasmic MBP. Thus, interaction of a single substrate-bound MBP with Tar appears sufficient to initiate a chemotactic signal from the transducer. A half-maximal chemotactic response occurred at 25% of the malE+ MBP level, suggesting that in vivo the KO for binding of maltose-loaded MBP to Tar is quite high (-250 p~) .
The proteins needed for maltose uptake in Escherichia coli are encoded by genes of two divergent operons in the malB region, located at 91 min on the chromosome (Fig. 1). The malE gene codes for maltose-binding protein (MBP'), which is localized in the periplasmic space (2). MalF and MalG are integral proteins of the cytoplasmic membrane (3)(4)(5), and MalK is apparently attached to the inner surface of the cytoplasmic membrane via MalG (5). The lamB gene codes for the subunit of maltoporin, which exists as a trimeric pore in the outer membrane (6). Maltoporin specifically facilitates *This work was supported in part by Sonderforschungsbereich SFB 156 from the Deutsche Forschungsgemeinschaft. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ' The abbreviation used is: MBP, maltose-binding protein.
permeation of maltose and longer maltodextrins through the outer membrane (7, 8) and also serves as the receptor for phage X (9).
The malT gene, mapping at 74 min, codes for a positive regulator required for expression of all other mal operons, including those of the malB region (10). Growth on maltose induces expression of MalT-dependent operons; m a l T mutations lead to constitutive expression of the mal regulon (11).
When maltose binds to MBP the protein undergoes a change in conformation (7,12). Maltose transport is thought to be initiated by interaction of substrate-loaded MBP with the inner membrane transport components (13, 14). The apparent K , for maltose transport, about 1 PM in lamB' cells, increases 100-fold in mutants lacking maltoporin (7) because maltose entry into the periplasm becomes rate limiting at low maltose concentrations. The MBP-dependent system is the only significant pathway for maltose uptake in E. coli under normal conditions (13).
In addition to its role in transport, MBP functions as the maltose chemoreceptor (15). A second protein, the chemotactic signal transducer Tar (faxis to aspartate and some repellents), is required to generate the chemotactic response to maltose (16). Tar is located in the cytoplasmic membrane (17) and also serves as the receptor for the potent attractant Laspartate and the repellents Co2+ and Ni". Whereas aspartate binds directly to Tar (18) maltose binds indirectly in the form of maltose-loaded MBP (19, 20). Receptor and signal transducer functions in chemotaxis have been recently reviewed (21).
Maltoporin, with lo5 copies of monomer/cell (22), and MBP, with 2 to 4.5 X lo4 copies/cell (23, 24), are major components of the cell envelope in maltose-induced or m a l T strains. The periplasmic MBP concentration has been estimated to be around 1 mM (23). In contrast, maltose-induced cells contain much lower amounts of malF, G, and K gene products, probably about 1000 copies/cell (5). Tar, which is not part of the mal regulon, also is present in about 1000 copies/cell (20, 25).
T o understand why MBP is present in 20-to 40-fold excess relative to its membrane partners, it is necessary to know how transport and chemotaxis depend on the periplasmic MBP concentration. Maltose-induced cells have higher transport and chemotactic activity than uninduced cells (26, 27). For maltose, galactose, and ribose, the increase in chemotactic activities toward each sugar was roughly proportional to the respective binding activities in the periplasm of uninduced and induced cells (24). These experiments, however, permitted the comparison of only two levels of binding protein and, for 9727 This is an Open Access article under the CC BY license.
FIG. 1. The m d B region of E. coli. Five genes whose products are known to be involved in maltose uptake are arranged into two divergent operons. In one operon malE codes for maltose-binding protein, and mnlF and mnlG for proteins found in the cytoplasmic membrane. In the other operon malK codes for a protein that is associated with the cytoplasmic membrane and lamB codes for the subunit of maltoporin, an outer membrane pore that also serves as receptor for phage X. Two positive regulators, the MalT protein and cyclic AMP-receptor protein (CAP), bind to the promoter regionp to stimulate transcription of both operons (1). transport, induction altered the amount of all components involved, not just binding protein.
The resolution to these difficulties lay with a series of malE mutations that affect secretion of MBP (28, 29). MBP is synthesized as a precursor from which 26 amino acids at the NH2 terminus, the signal peptide, are removed to form mature periplasmic MBP (30, 31). The mutations alter the signal peptide so as to interfere in varying degrees (from 4 to 23% of malE+ periplasmic MBP levels) with secretion of MBP (32). Unexported MBP remains as precursor in the cytoplasm, while the reduced amount of MBP that does reach the periplasm is found in the normal, mature form. We measured the periplasmic MBP level and correlated this value with maltose chemotactic and transport activities in mlT" strains carrying five different malE signal sequence mutations. With these data, we determined in vivo affinities for the interaction of MBP with Tar and the MalFGK complex. The maltose chemotactic response depended on the first power of the MBP concentration. Using 1 mM as the periplasmic MBP concentration in malT" malE+ cells, the KO for MBP binding to Tar was determined to be 250 PM. Maltose transport showed a sigmoidal dependence on the MBP concentration, with half-maximal transport occurring at 90 p M MBP. The apparent K , of maltose transport remained essentially constant over the range of MBP levels tested. The implications of these findings for the mechanisms of maltose transport and chemotaxis are discussed.
Bacterial Strains-The bacterial strains used in this study are described in Table I. The construction of strains carrying malE signalsequence mutations, malF::TnlO and lamB102(Am) is explained in Fig. 2, All genetic manipulations were performed according to Miller (38).
Preparation of Osmotic Shock Fluid-Strains "209 through "215 were grown at 37 "C in 1 liter of H1 minimal medium (39) containing 0.2% glycerol and 0.1% casamino acids. The cells were Tris-HCI, pH 7.3, containing 0.03 M NaC1. Osmotic shock was done harvested at an A578nm = 0.6 and resuspended in 100 ml of 10 mM according to Neu and Heppel (40) with a shock volume of 100 ml. The crude shock fluid was lyophilized, resuspended in 2.5 ml of 10 mM Tris-HC1, pH 7.3, and dialyzed against the same buffer. The slightly turbid solution was cleared by centrifugation at 13,000 X g for 5 min and used for the determination of MBP.
Polyacrylamide Gel Electrophoresis in Urea-Analysis of MBP by polyacrylamide gel electrophoresis in sodium dodecyl sulfate proved unsatisfactory, since another protein in shock fluid from our strains co-migrated with MBP (data not shown). In contrast, MBP formed a unique band during polyacrylamide gel electrophoresis in 8 M urea performed according to Pugsley and Schnaitman (41).
Equilibrium Dialysis-Three times 200-pl aliquots of crude shock fluid were transferred into Visking dialysis tubing and dialyzed at 4 "C against 100 ml of Tris-HC1, pH 7.3, containing 3 X lo-' M ["C] maltose and chloramphenicol (20 pg/ml). With the highest MBP concentration used (about 1 mg/ml), equilibrium was reached after 20 h of dialysis. After 36 h the dialysis bags were drained and the volume of the content determined by weighing. Volume changes of up to &20% occurred during dialysis. To determine the amount of maltose bound to MBP, the radioactivity in the bags was measured and the radioactivity of the corresponding volume of the external solution was subtracted. The total MBP concentration (Pt) was determined by the equation, where PL represents MBP complexed with maltose, L the free maltose concentration, and KD the dissociation constant. This equation is valid at free maltose concentrations far below KO (42). A calibration curve for PL/L was obtained with known amounts of MBP ranging from 0.01 to 1.0 mg/ml. A KD of 3 p~ at 4 "C (43) was used, and one binding site/MBP monomer was assumed. PL/L values obtained from the six different MBP preparations ranged from 4 (wild type) to 0.2 (malEl8-1). The corresponding value for a preparation from a AmalE mutant was 0.01. Triplicate determinations showed, at most, 10% variation. The values obtained were normalized to the number of cells from which the shock fluid was obtained.
Immunodiffusion Assay-Shock fluid was tested with antiserum against MBP on Ouchterlony double-diffusion plates (44) at 37 "C. The central wells contained 10 p1 of antiserum and the six peripheral wells 10 p1 of 2-fold serial dilutions of crude shock fluid.
Immunoprecipitation of MBP from Shock Fluid-To ensure that MBP in the shock fluid was in the mature form and not precursor, samples were precipitated with anti-MBP antiserum followed by adsorption of immunoglobulins to glutaraldehyde-treated Staphylococcus aureus. The cells were extensively washed; MBP and antibodies were removed with sodium dodecyl sulfate and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis as previously described (45).
Maltose Transport-Strains "186 through "191 were grown at 32 "C in minimal medium A (38) containing 0.2% glycerol and 0.1% casamino acids. The cells were harvested at A678-= 0.5 (1 to 2 X lo8 cells/ml), washed three times in minimal medium A, and resuspended to A578-= 1.0. To 2.9 ml of cells 100 pl of ["C]maltose was added. 0.5-ml samples were withdrawn at 15,30,60, and 90 s and filtered through Millipore filters of 0.45-p pore size; the filters were washed with 10 ml of minimal medium A, dried at 120 "C, and counted in a scintillation counter after addition of a toluene-based cocktail. For determining the K,,, of maltose transport, a constant amount of radioactive maltose and varying amounts of unlabeled maltose were mixed in the 100-pl substrate solution added to the cell suspension. Depending on the total substrate concentration, the cell density was adjusted so that the initial rate of uptake was linear over a period of 2 min. For instance, the malE+ strain at 0.1 pM substrate concentration was used at a density of 0.01 (578 nm) and at a density of 1.0 at 0.1 mM maltose. With these precautions, at most 10% of the total radioactivity was accumulated by the cell suspension after 2 min.
Capillary Assay-Cells were grown as for maltose transport assays and harvested at A678nm = 0.5 and then washed three times with 10 ml of chemotaxis buffer (10 mM potassium phosphate, pH 7.0, containing 10" M EDTA and M L-methionine). Cells were resuspended very gently to avoid damaging the flagella. The capillary assay was done according to Adler (39) and run for 40 min at 32 "C with cells resuspended in chemotaxis buffer to A578-= 0.005 (1-2 X 1 0 ' cells/ml). The number of cells entering the capillary was determined by diluting the capillary contents and plating on nutrient agar.
Tethered Cell Assay-Strains used in this assay were made tsr to minimize possible thermotactic responses (46). Cells were grown and harvested as for the capillary assay, except that the cells were washed twice in tethering buffer (10 mM potassium phosphate, pH 7.0, containing 0.1 M NaC1, 10" M EDTA, and low5 M L-methionine).
Flagellar filaments were sheared with a Waring blender; a 25-ml cell suspension in a 50-ml blender cup was exposed to the highest speed for 45 s. After shearing, cells were washed once more and resuspended in tethering buffer to A678nm = 0.05. Cells were tethered with anti-  (37) This study This study

This study
This study serum against flagellar filament as described previously (37), except that acid-cleaned rather than siliconized coverslips were used.
Coverslips were sealed with Apiezon L grease onto a flow chamber (47) mounted on the stage of a Nikon Optiphot reverse-phase-contrast microscope. Cells were viewed at a magnification of 400X. Recordings were made with a video camera (Panasonic WV1850/G) and cassette recorder (Panasonic AG-6200) and analyzed by playback on a video monitor (Panasonic WV-5410). Solutions were drawn through the chamber with vacuum; 99% replacement of the contents occurs within 5 s (47). Flow times were 15 s. Maltose and L-aspartate were added at 1 mM, a concentration that saturates the respective chemoreceptors for both attractants in the strains used.

RESULTS
Determination of Periplasmic MBP Leuels-We determined MBP levels in shock fluid prepared from malT" strains carrying the malE signal-sequence mutations, a malE+ gene, or the nonpolar malE444 deletion. Since binding of MBP to maltoporin, Tar, and the MalFGK complex could hinder the release of MBP by cold osmotic shock, all strains contained an amber mutation in lnmB, a deletion of tar, and a TnlO insertion in malF (see Table I). MBP formed a unique band when shock fluid was analyzed by polyacrylamide gel electrophoresis in the presence of 8 M urea (Fig. 3). The intensity of the band decreased in the order malE+, malEl6-1, malElO-1, malEl4-1, and malEl8-1 and malEl9-I, with the last two being very similar. No MBP was seen with shock fluid from a strain containing the nonpolar deletion malE444, Material precipitated from malE+ and mutant shock fluids with antiserum against MBP showed only one band on sodium dodecyl sulfate-polyacrylamide gels, indicating that periplasmic MBP existed only in the mature form (data not shown).
MBP was quantitated by measuring maltose binding with equilibrium dialysis and by the Ouchterlony immunodiffusion test with antiserum against MBP. The results obtained with the two methods were in good agreement (Table 11). According to the binding data, the MBP levels in the mutants ranged from a high of 23% of the malE+ level in malEl6-1 to 3.6% in malEl9-1. We could not calculate absolute periplasmic MBP concentrations since we did not measure the periplasmic volume. The MBP concentration in maltose-induced E. coli cells has been estimated to be 1 mM (23). In what follows, therefore, we will accept 1 mM as the MBP concentration in the periplasm of m a P malE+ cells, also.
Maltose Transport in Signal-Sequence Mutants-The initial rate of maltose transport was determined at maltose concentrations ranging from 0.1 to 33 p M in maly-1 derivatives of the wild-type and mutant strains. All were malF+G+K+, lamB+, and tar+. The curves in Fig. 4 2. Construction of strains carrying malF::TnlO, lamB102.,b, and malE signal-sequence mutations. In the first step, not illustrated, ma1F::TnIO was transduced with phage Plvir into a strain containing the deletion malB11 (33), which confers resistance to phage AVi, (Xr) by removing the malK-lamB promoter. Tetracyclineresistant (Tc') transductants were screened for A'. A Plvi, lysate grown on one such isolate was used in a second step ( A ) to transduce Tc' into strains carrying the malT'-I allele and the various malE signal-sequence mutations (10-1 etc.). Crossovers in the regions marked I and II gave rise to transductants with ma1F::TnlO and the malE mutations; these were identified by their sensitivity to phage AVi.. In the final step ( B ) Plvir lysates grown on isolates from A were used to transduce Tc' into strains containing the nonpolar deletion malE444 and the lamB102 amber mutation. Tc' transductants, after screening for A', were assayed for MBP with the Ouchterlony immunodiffusion test following chloroform-toluene extraction of periplasmic proteins. Isolates producing no detectable MBP were assumed to have retained the malE deletion (crossovers in the regions designated I and II). Isolates producing MBP did so in amounts reflecting the severity of the malE signal-sequence mutation involved (crossovers in the regions designated I and ZII). were solely the consequence of reduced MBP levels, we increased the amount of MBP by lysogenizing with phage XpK3 (48). This phage carries a malE+ gene but no other mal genes. Maltose transport rates in lysogens of malE+, malE444 deletion, and malE18-1 strains were identical to each other and to the nonlysogenic malE+ strain (data not shown). Thus, in a Equilibrium dialysis was performed as described under "Experimental Procedures" a t a ["C]maltose concentration of 3 X lo-" M. The amount of MBP was calculated using a KD of 3 p~ for binding of maltose to MBP at 4 "C (43), a molecular weight of 40,661 daltons for MBP (31), and our measured value of 2.6 X 10" viable cells/ml of exponential culture a t A57anrn = 1.0.
*Two-fold serial dilutions of shock fluid were made, and each dilution challenged with antiserum against MBP in the Ouchterlony immunodiffusion test. For each shock fluid the dilution factor at which the last visible precipitation band formed was determined. If the last band was very faint the shock fluid was assigned the mean value of the last two dilution factors giving visible bands. The results are expressed as fractions of the malE+ dilution factor. the signal-sequence mutant with the lowest transport rate and presumably in the other mutants as well, the cause of the transport defect was the inadequate supply of MBP.
Maltose Chemotaxis in Signal-Sequence Mutants-All of the mutants formed smaller chemotactic swarms than a malE+ strain in soft agar plates containing 100 pM maltose. The swarm diameter became progressively smaller in the Transport, Taxis, and Binding Protein Content 9731 order malEl6-1, malElO-I, malEl4-1, and malEI8-1 and malEl9-I, the last two hardly swarming a t all. Also, none of the mutants formed sharp chemotactic rings, although malEI6-1 showed a faint ring. However, since establishment of a maltose gradient in the swarm plate requires maltose uptake and metabolism, swarming behavior could not distinguish between defects in maltose transport and chemotaxis.
Chemotactic responses can be analyzed without the complicating factor of transport by utilizing the capillary assay (39), in which chemotactic gradients are created by diffusion. Our assays were performed with mal?" strains containing a TnlO insertion in malF. This TnlO insertion completely blocks maltose transport. The results of the capillary test are presented in Fig. 5. T o correct for possible differences in motility and Tar content, accumulations in maltose-containing capillaries were normalized by dividing by the accumulation in aspartate-containing capillaries for each strain. The malE+ strain and all of the mutants showed a peak accumulation in capillaries containing 100 p M maltose, but the magnitude of the accumulation varied markedly among the mutants. For malEl6-I, malElO-I, malEl4-I, malEl8-1, and malEl9-1 the peak accumulations were 64, 36, 18, 8, and 4% of the malE+ accumulation, respectively.
When E. coli cells are exposed to a large, rapid increase in attractant concentration, they respond by totally suppressing clockwise rotation of their flagella for a time proportional to the change in the fraction of receptor bound with attractant (49, 50). The period of exclusively counterclockwise rotation that ensues, called the transition time (49) or recovery time (50) by different authors, is here referred to as the response time. For a maltose stimulus, the response time should be determined by the degree to which Tar is occupied by maltosebound MBP. Thus, the in vivo binding of MBP to Tar can be investigated by recording response times to maximal maltose stimuli in cells with different periplasmic MBP content.
Response times were measured in a flow chamber (47) with cells tethered to a glass coverslip with antiserum against flagellar filament. The strains were all matF tsr derivatives. To prevent possible competition with Tar for binding of MBP, the strains contained a TnlO insertion in malF and an amber Capillaries contained either chemotaxis buffer alone, 1 mM L-aspartate in chemotaxis buffer, or maltose, at the concentrations indicated, in chemotaxis buffer. Accumulation in capillaries containing buffer alone was subtracted as background. To correct for differences in motility or Tar-mediated chemotaxis among preparations, the maltose accumulation for each strain was normalized by expressing it as a fraction of the aspartate response for that strain. The accumulations in aspartate capillaries ranged from 230,000 to 320,000, roughly 40% of the total number of cells. Symbols are as in Fig. 4. mutation in lamB. The maltose concentration was quickly shifted from zero to 1 mM maltose, a change that elicits the maximal response to maltose even in lamB mutants (37). At 1 mM maltose, MBP should be saturated with substrate.
The response time data are given in Table 111. As in the capillary test, the maltose response times were normalized by dividing by the response time to a 1 mM aspartate stimulus. The mean response times for malE16-1, malElO-1, malEl4-I , malEl8-1, and malEl9-1 cells were 60, 36, 20, 11, and 8% those of isogenic malE+ cells, respectively. The correspondence of these results to those of the capillary test (above) was striking.
Dependence of Transport and Chemotaxis on Periplasmic MBP Concentration-The relationship between periplasmic MBP levels, transport, and taxis was examined by plotting transport V,,, or tactic response against the amount of MBP (Fig. 6). The dependence of transport rate on MBP level did not follow simple Michaelis-Menten kinetics (Fig. 6A). To identify possible cooperativity we constructed a Hill plot from the transport data. The points fell on a straight line with a slope of 2.7 (not shown), suggesting that the rate-limiting step of transport involves MBP in a cooperative interaction. Nine per cent (90 pM) of the malE+ MBP concentration was required for half-maximal transport.
A double reciprocal plot of normalized maltose response time uersus MBP level (not shown) was linear at the four highest MBP concentrations. From the linear region of the plot, we determined that the hypothetical maximum response, at infinite MBP concentration, would be 0.34. Our malE+ cells had a maltose response of 0.27, about 80% of the theoretical maximum response. The amount of MBP giving half of the hypothetical maximum response, equivalent to the KD for the MBP-Tar interaction in uiuo, was 25% (250 p~) of the malE+ level. The curve in Fig. 6B was derived using these values for the maximum response and Ku, assuming a firstorder dependence on the MBP concentration. The fit to the data was good except at the two lowest concentrations of MBP.
Strains with the Iowest MBP levels had briefer responses than expected from first-order dependence on MBP level. This deviation from linearity could be an artifact introduced by subtracting the response of a AmalE strain from the maltose responses of all other strains (Table 111). This correction was made assuming that the brief responses of AmalE were elicited by small amounts of contaminating sugars in our purified maltose. If the responses to these contaminants and maltose were not fully additive (51) we would overcorrect by subtracting the AmalE response. This source of error would be greatest for mutants with low MBP levels.

DISCUSSION
We studied the dependence of maltose transport and chemotaxis on periplasmic MBP concentration using strains carrying various malE signal-sequence mutations. Under these conditions MBP was the only component of the transport or taxis systems whose amount varied. Our results demonstrated that more MBP is required for chemotaxis than for transport. For example, malEl6-I strains, with 23% of the malE+ amount of MBP, transported maltose at 90% of the malE+ rate but had only 60% of the malE+ chemotactic response to maltose. This finding is consistent with the report that 5-or 10-fold overproduction of Tar, the signal transducer for maltose taxis, did not inhibit maltose transport in uninduced cells (52). The data also suggest that the reported proportionality of maltose transport and MBP content in uninduced and induced E. coli (24) is fortuitous, simply re- 'The ratio of the maltose to aspartate response times of individual cells was determined. The value shown is the mean of these ratios, +1 S.D., for the cell ensemble. The number of cells analyzed is given in parentheses.
The mean ratios of the maltose to aspartate response times for each strain after subtraction of the mean ratio for the AmalE strain. This correction was made because the brief response of the AmalE strain to maltose stimuli was probably caused by residual contaminants, such as glucose, in our purified maltose. Without this correction we would tend to overestimate maltose resDonse times. The percentages in parentheses are expressed relative to  Table 11). The sigmoidal curve represents the best fit to the data and reflects the Hill coefficient of 2.7 determined for MBP in its transport function (see text). In B, normalized mean response times to chemotactic stimulation with 1 mM maltose (Table 111)  The rates of transport at saturating maltose concentrations appeared to increase in a sigmoidal fashion with the amount of MBP (Fig. 6A). A Hill plot of this data yielded a straight line with a slope of 2.7. We do not want to emphasize this numerical value, but we do believe our data present evidence for some degree of cooperativity. The half-maximal transport rate occurred at 9% (90 p~) of the mdE+ MBP level, which we have taken as 1 mM (23). Since there are 1000 copies each of MalFGK ( 5 ) and Tar (20,25) per cell, the amount of free MBP could be reduced by 20 to 25% through binding to membrane components. Thus, 90 pht is an upper estimate for the concentration of MBP needed to maintain half-maximal transport rates at saturating maltose concentrations.
The apparent maltose K, for transport varied less than a factor of 2 (from 0.8 to 1.4 p~) over a 25-fold range of MBP concentrations (Fig. 4). A model in which MBP exists only as monomer and only substrate-bound MBP interacts with membrane components would predict that the apparent K,,, would increase in mutants with reduced MBP levels. Since such an increase in K,,, was not observed, this model is inadequate to explain how binding protein functions in transport.
Published reports that binding proteins form dimers (53-55) or larger aggregates (56) may have a bearing on our observation of cooperativity. Richarme (53) reported isolation of dimers of MBP or galactose-binding protein that could be converted to monomers by addition of substrate. According to his scheme, substrate-loaded monomers would interact with membrane transport components, then dimerize, causing release of substrate. Dimers would subsequently dissociate from the membrane components and re-enter a monomer-dimer equilibrium. A transport mechanism of this type, or variations thereof, might well involve cooperative interactions of MBP.
Maltose taxis showed a first-order dependence on the MBP concentration (Fig. 6B). Thus, a single substrate-bound MBP molecule can apparently elicit a chemotactic signal from the Tar transducer with which it interacts. Half the hypothetical maximum response to a saturating maltose stimulus (a jump from 0 to 1 mM) occurred at 25% (250 pht) of the rnalE+ MBP level. Since the response time to an attractant is proportional to the per cent change in occupied chemoreceptor (49, 501,