THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 250, No. 7, Issue of April 10, pp. 2574-2680, 1976

Printed in U.S.A.

Purification and Properties of a Periplasmic Glutamate- Aspartate Binding Protein from Escherichkz coli K12 Strain W3092*

(Received for publication, July 8, 1974)

RANDALL C. WILLISI AND CLEBIENT E. FURLONGS

From the Department of Biochemistry, University of California, Riverside, California 92502

SUMMARY

A protein which binds both glutamate (KD = 0.8 PM) and aspartate (KD = 1.2 PM) has been purilied to homogeneity (290-fold) from the periplasmic fraction released from Escherichia coli W3092 by the cold osmotic shock procedure. The apparent molecular weight of the glutamate-aspartate binding protein is approximately 31,000 as judged by gel electrophoresis, gel filtration, and sedimentation equilibrium centrifugation; and the protein has a p1 of 9.69. This protein contains 2 half-cystine residues and is dependent on a dithiothreitol-sensitive component for renaturation to an active conformation following urea or guanidine treatment. Of the natural amino acids only the L isomers of glutamate, aspartate, glutamine, asparagine, and alanine were inhibitors of either [14C]glutamate or [“Claspartate binding and the inhibitions were competitive. Only one binding site is indi- cated per molecule of protein.

Antibody prepared against the glutamate-aspartate binding protein does not cross-react with purified glutamine binding protein or any other component of osmotic shock fluid. The antibody does cross-react with osmotic shock fluids obtained from E. coli strains B and W and Salmonella typhimurium LT2. The glutamate-aspartate binding protein-antibody complex does not bind either glutamate or aspartate. The protein may be similar to the glutamate binding activity de- tected in the periplasmic fraction released from E. co2i strain B (MINER, K. M., AND FRANK, L. (1974) J. Bacterial. 117, 1093-1098) and strzK12 CS (BARASH, H., AND HALPERN, Y. S. (1971) Biochem. Biophys. Res. Commun. 45, 681-688). This protein appears to function in the transport of glutamate by E. coli strain W cultured in minimal medium with suc- cinate as the carbon source (WILLIS, R. C., AND FURLONG, C. E. (1975) J. Biol. Chem. 250,2581-2586).

* This work was supported by United States Public Health Service Research Grant GM18311 from the National Institute of Medical Sciences. A preliminarv account of this work has been presented (1). -

1 Present address, Department of Medicine, University of California, La Jolla. California 92037.

fi To whom correspondence should be addressed.

Evidence accumulated over the past several years suggests that a class of osmotic shock releasable, periplasmic proteins termed binding proteins may serve as recognition components for a number of active transport systems (2, 3) and, in at least one case, as receptors for various chemotactic systems (4) of the gram-negative bacteria. A binding activity for Gglutamate has been detected in and partially purified from the supernatant solutions of cold, osmotically shocked Escherichiu coli U cells (5) and plasmolyzed lysozyme-EDTA treated E. coli K12 cells (6). The association of the glutamate binding protein with the active transport of glutamate is not clear, however. Studies with E. coli K12 suggest that glutamate transport may be dependent on the presence of the glutamate binding activity (6), whereas studies with E. coli I3 suggest that only a portion of glutamate transport may be dependent on the binding activity (5). Studies with membrane vesicles from E. coli ML308-225 which are presumably without periplasmic proteins suggest no dependence of glutamate transport on the glutamate binding protein in isolated membranes (7).

This report describes the purification and properties of an osmotic shock releasable glutamate-aspartate binding protein from E. coli K12 strain W3092.

EXPERIMENTAL PROCEDURE

Materials

L-[U-W]Glutamic acid (262 mCi/mmol), and L-[U-Wlaspartic acid (218 mCi/mmol) were obtained from New England Nuclear. The isotopes were diluted to a specific activity of 100 mCi/mmol and adjusted to pH 7.0 with KOH for use. Unlabeled L-amino acids for isotope dilution and competition experiments were obtained from Sigma Chemical Co. or Mann Research Labora- tories. The following compounds were purchased from the Sigma Chemical Co.: nn-cr-methylglutamic acid, cu-ketoglutaric acid, pyruvic acid, L-glutamine, n-asparagine, L-glutamic acid-y- monohydroxamate, ma-aminobutyric acid, nn-aspartic acid-b- monohydroxamate, and dithiothreitol. The following com- pounds were purchased from Nutritional Biochemicals Corp. : L-&aspartyl hydraside, n-4-azaleucine dihydrochloride, glutamic acid-y-ethyl ester, glutamic acid-y-methyl ester, r-L-glutamyl hydraside, n-aspartic acid-&methyl ester, and fi-L-aspartyl hydrazide. Calbiochem was the source for the following chemi- cals: n-y-aminobutyric acid, n-glutamic acid, n-serine, L-methi- onine-n, L-sulfoxamide, n-norleucine, cycloleucine, n-aspartic acid, and n-asparagine. N-Bromosuccinimide was obtained from Eastman Chemical and recrystallized from deionized water. All other chemicals were of at least reagent grade and obtained from the usual commercial sources. Nitrocellulose membrane filters

2574

by guest on Decem

ber 19, 2020http://w

ww

.jbc.org/D

ownloaded from

2575

(16~mm diameter) were punched from sheets (35 X 55 cm) of Schleicher and Schuell Co., Selectron B6, 0.45-pm membrane.

Methods

Erruilibrium Dialysis Membranes-Dialvsis membranes were prepared as described by Englund et al. (8).

Growth and Processina of Bacterial Cells-Escherichia coli K12 W3092 and E. coli strain B-were obtained from Dr. L. A. Heppel, Cornell University, Ithaca, N. Y. E. coli W strain DzW was obtained from Dr. L. Leive, National Institutes of Health, Be- thesda, Md. Salmonella typhimurium LT2 was obtained from Dr. G. F. Ames, University of California, Berkeley. Small cultures were grown from 2yo inocula in 50 mi of Medium B (9). Large cultures were grown in 100 liters of Medium ASA with 2% DO- tassium succinate as the initial carbon source in a 100-liter-New Brunswick Fermacell (10). Cells from large cultures were har- vested, osmotically shocked, and the shock fluid prepared for CM-cellulose chromatography as described previously (9).

Binding Assays-Equilibrium dialysis was carried out in a small Plexiglas multichambered apparatus described previously (9, 11). For quantitation of binding activity during purification, Side A initially contained 0.05 M potassium phosphate, pH 6.9, protein, and 10 PM isotope in a total volume of 0.05 ml. Side B contained the same ingredients without protein.

The nitrocellulose filter binding assay (12, 13) was used for determination of binding activity in column fractions and for characterization of the purified protein. Filter binding assays contained the following (in order of addition and mixing): 20 ~1 of appropriately buffered protein solution, 5 ~1 of radioactive licrand. 5 ~1 of indicated comnetitor or salts. and 5 LIP of 1.2 M magnesium chloride. Total iadioactivity oi the reaction mix- ture was determined by scintillation counting of 5 ~1 of the assay mixture in 5 ml of toluene-Triton scintillation solution (14). Unbound radioactivity was determined as the difference between total radioactivity calculated to be present in 5 ~1 of assay mix- ture and the radioactivity present on the filter after appropriate corrections for background and counting efficiency.

In experiments with the purified glutamate-aspartate binding protein, the protein, as stored in distilled water, required neu- tralization. A neutralized protein solution was prepared from 30 ~1 of a 3.5 mg/ml glutamate-aspartate binding protein solution by addition of 140 ~1 of distilled water and 30 ~1 of 0.2 N HCl. Buffered nrotein solutions were nrenared bv addition of 75 rrl of distilled aster to 10 ~1 of neutr&ed protkin solution followed by 15 ~1 of 1 M buffer. A direct determination of pH was made on each buffered protein solution. Protein used in the kinetic and specificity determinations was prepared by a 6.7-fold dilution of neutralized protein solution with 67.5 mM potassium phos- phate, pH 6.9.

Protein Determinations-Protein concentrations were deter- mined by the method of Lowry et al. (15) using ribonuclease A (Ey = 0.695) as a standard (16).

Immunological Methods-Antiserum was prepared against glutamate-aspartate binding protein by four weekly injections of 0.5 ml of 50yo Freund’s complete adjuvant containing 0.1 mg of pure binding protein into the popliteal lymph nodes’ of one fe- male New Zealand rabbit. Blood (50 ml) was collected from the ear vein at 7- to g-day intervals following 1 month of injections. Booster injections (0.1 ml of 10% Freund’s complete adjuvant containing 0.05 mg of pure binding protein) were administered twice monthly for 2 months.

-y-Globulin was prepared from control sera and antisera by ammonium sulfate precipitation and DEAE-cellulose chroma- tography (17). The r-globulin from each bleed was stored at 4” as a lo-mplml protein solution in 7% ammonium sulfate and 50 mM potas&m phosphate, pH 6.8. ‘-

Polyacrylamide Gel Electrophoresis-Sodium dodecyl sulfate slab gel electrophoresis was performed in an apparatus described by Studier (18.19) with a 10% acrylamide and 0.2% bisacrylamide &solving gel with buffers and stiining procedures described by Fairbanks et al. (20). extent that buffalo black renlaced Coomassie brilliant blue in’thgstaiding procedure. Acidic-, pH 4.5, slab gel electrophoresis was performed with a 7.5yo acryiamide ind O.syo bisacrylamide resolving gel with 2yo acrylamide and 0.25% bis- acrylamide stacking gel and buffer systems as described by Reis-

l G. A. Granger, personal communication.

feld et al. (21). Studies of variation in protein mobility as a function of per cent gel concentration (22) were performed with tube (0.5 cm (inside diameter) X 15 cm) gels with acidic, pH 4.5 (21), and pH 3.5 (23) buffer systems. Apparent molecular weights were determined in 5, 7, 9, and 11% acrylamide gels with ribo- nuclease A (MW 13,700) (24), glutamine binding protein (MW 25,000) (25), and ribose binding protein (MW 28,500) (26) as standards.

Preparative Isoelectric Focusing-Electrofocusing was per- formed as described by Vesterberg (27). The dense solution contained (per 48 ml) 1.9 ml of 20yo Ampholine solution, pH 9 to 11, and 24.0 g of sorbitol. The light solution contained (per 50 ml) 0.7 ml of 40yo Ampholine solution, pH 8 to 10, and 0.3 ml of 40yo Ampholine solution, pH 6 to 8. The cathode solution (per 20 ml) contained 2.0 ml of 1 N sodium hydroxide and 12.0 g of sorbitol. The anode solution was 0.01 N acetic acid. The column was maintained at 4’ and prefocused at 650 volts for 24 hours. The sample was loaded as described under “Results.” The pH values of the fractions (0.7 ml) were read at 4’ with a Radiometer pH meter PHM26 coupled to a Beckman 39505-B3 electrode which had been equilibrated in 4’ buffer overnight.

Sephades G-200 Chromatography-A Sephadex G-200 column was prepared according to the manufacturer’s directions with 50 mM ammonium bicarbonate as the running buffer. The column was calibrated (28) with bovine serum albumin (67,000 MW), E. coli 5’-nucleotidase (52,000 MW (29)), ovalbumin (44,000 MW), E. coli galactose binding.protein (35,060 MW (30)), E.. coli glu- tamine binding protein (25,000 MW (25)), myoglobin (18,000 MW), and cytochrome c (11,700 MW).

Amino Acid Analysis-Aliquots (200 rg) of the purified protein were lvonhilized and either hvdrolvzed directlv at 105” for 22 hours “in- evacuated, nitrogen “flushkd, sealed glass tubes con- taining 6 N HCl or hydrolyzed after performic acid oxidation (31) and lyophilization. The hydrolysates were lyophilized and diluted to 400 ~1 with sample buffer. The hydrolysis and chroma- tography were performed by Eldex Laboratories, Inc. (Menlo Park, Calif.). The tryptophan content was determined by 288:280 ratio with a Cary 118 spectrophotometer as described by Edelhoch (32) and after p-toluene sulfonic acid hydrolysis and chroma- tography with a Beckman model 120 C automatic amino acid analyzer as described by Liu and Chang (33).

Sedimentation Equilibrium Centrifugation-Long column menis- cus depletion sedimentation equilibrium runs were performed as described by Chervenka (34).

Fluorescence Spectroscopy-Fluorescence studies were per- formed with a Farrand Mark I spectrofluorometer with a 0.4-ml cell (3 X 3 mm) and 5-mm entrance and exit slits. The tempera- ture controlled cuvette holder was maintained at 23”.

RESULTS

A protein which binds glutamate and aspartate is found in the crude osmotic shock fluid prepared from succinate-asparagine minimal media cultured cells as described by Willis et al. (9). The glutamate-aspartate binding protein was purified to homo- geneity by the following procedure. Nineteen liters of clarified shock fluid saturated with nitrogen were adjusted to pH 4.5 with acetic acid and loaded onto a IOOO-ml (5.7 x 40 cm) CM- cellulose column equilibrated with 10 mM sodium acetate, pH 4.5. Approximately 7 g of protein were adsorbed onto the resin. The column was then washed with 5 liters of column equilibration buffer and eluted with a 20.liter linear gradient from 0 to 0.25 M

sodium chloride in 10 mM sodium acetate from pH 4.5 to 4.9. Fractions of 400 ml were collected. Glutamate binding activity was found to elute between 0.15 and 0.22 M sodium chloride. Fractions 36 to 41 were pooled and dialyzed against 10 liters of distilled water with five changes at 2- to 4-hour intervals. The dialyzed solution was then adjusted to a final concentration of 10 mM Tris-HCl, pH 7.8. The dialyzed material from CM-cellulose chromatography was loaded onto a 142 ml (1.8 X 56 cm) DEAE- cellulose column equilibrated with 10 mM Tris-HCl, pH 7.8. The glutamate binding activity was not adsorbed and was found in the pass-through fraction. The pass-through fraction and a

by guest on Decem

ber 19, 2020http://w

ww

.jbc.org/D

ownloaded from

2576

2.0

1.5

8

2 I.c

0.5

r

'L t

I-t-

t-

L 0 IO 20 30 40 50 60 70 SO 90 100 IO 20 30

FRACTION NUMBER

FIG. 1. Isoelectric focusing elution profile of glutamate-as- partate binding protein. The preparations and operation of the focusing column are described under “Methods.” Sample prep- aration and application are described in the text. -, relative absorbance at 280 nm; 0, pH of fractions; n , glutamate binding.

600-ml column wash solution were pooled and adjusted to pH 5.2 with acetic acid and loaded onto a 242-ml (5.7 X 9.5 cm) SP- Sephadex column equilibrated with 10 mM Tris-acetate, pH 5.0. The column was washed with 1200 ml of column equilibration buffer and eluted with a 1750-ml linear gradient from 0 to 0.15 M sodium chloride in column equilibration buffer. Fractions of 22 ml were collected. Glutamate binding activity was found to elute between 0.08 and 0.12 M sodium chloride. Fractions 68 to 86 were pooled and concentrated by first dialyzing against 40 volumes of 10 mM sodium acetate buffer, pH 4.5, then adsorbing the dialyzed material onto a CM-cellulose column equilibrated with the dialysis buffer. The protein was eluted with 0.25 M sodium chloride in 10 mM sodium acetate, pH 4.5. A 2-ml aliquot of concentrated glutamate binding activity in 50 mM ammonium bicarbonate-lO$& glycerol was applied to a 437-ml (2.2 X 115 cm) calibrated Sephadex G-200 column equilibrated with 50 InM ammonium bicarbonate. Fractions of 88 drops (approximately 4 ml) were collected. Dextran blue chromato- graphed at Fraction 36, and the glutamate-aspartate binding protein activity eluted with an apparent molecular weight of 29,000. Fractions 72 to 93 were pooled for glutamate binding activity, lyophilized to dryness, and redissolved in a 2-ml aliquot of Ampholyte-sucrose removed from the pH 8 region of a pre- focused electrofocusing column (pH 6 to 11). The sample con- taining the glutamate binding activity was replaced in the column at the position of withdrawal of the Ampholyte-sucrose aliquot and focusing was continued for 24 hours. The voltage was gradually increased to 650 volts during the first 4 hours. The glutamate-aspartate binding protein focuses with an isoelectric point of 9.69 (Fig. 1). Fractions 64 to 75 were pooled and passed through a Sephadex G-50 column equilibrated with 50 mM ammonium bicarbonate to remove the Ampholyte (27). The glutamate binding fractions were lyophilized, resuspended in deionized water, and stored at 4’.

The concentration of glutamate binding activity in the crude shock fluid as well as any other fractions by ultrafiltration with

TABLE I

PuriJZcation of glutamate-asparlale binding protein from Escherichia coli WSOSR

step

Osmotic shock supernatant

CM-cellulose. DEAE-cellu-

lose.. SP-Sephadex. Sephadex

G-200 Isoelectric

focusing*.

Volume

- I- Activity”

Units/ml --

ml

19,ooo 0.05, 2,400 0.27

3,000 0.21 2.2 256

4.0

3.2

128

71

--

5

-

Proteir jpecific Purifi. Lctivity catior

Units

Re- WXY

%

1045 0.49 0.11 0 100 648 0.24 1.1 10 62

630 0.03 563 12.2

60 54

512 4.8

230 2.3

7 64 21 191

27 245

31 281

49

44

y Nanomoles of glutamate bound. * Only 50yo of the material from the previous step was used

A B C D E F

FIG. 2. Slab polyacrylamide gel electrophoretic characteriza- tion (pH 4.5) of the purity of the glutamate-aspartate binding protein at each step of purification. A, crude shock fluid (50 pg); B, CM-cellulose fraction (11 pg); C, DEAE-cellulose pass through fraction (3 fig); D, SP-Yephadex fraction (18 pg); E, Sephadex G-200 fraction (7 pg); F, isoelectric focusing fraction (5 id.

either Amicon UM or I’M membranes was avoided since the glutamate-aspartate binding protein was found to bind tena- ciously to these membranes. Interestingly, the bound protein retains binding activity.

The purification of the glutamate-aspartate binding protein is summarized in Table I, and the purity of each fraction is indi- cated by polyacrylamide slab gel electrophoresis at pH 4.5 (Fig. 2). Antibody prepared against purified glut’amate-aspartate binding protein does not appear to cross-react with any other binding proteins in crude shock fluid (Fig. 3), including the glutamine binding protein. The sera did react with a deter- minant in shock fluid from Escherichia coli strains W and 13 and Salmonella lyphimurium LT2. Sera prepared against the follow- ing binding proteins did not cross-react with glutamate-aspartate binding protein: leucine-isoleucine-valine, leucine-specific, gluta- mine, and galactose. The glutamate-aspartate binding protein- antibody complex does not bind glutamate.

by guest on Decem

ber 19, 2020http://w

ww

.jbc.org/D

ownloaded from

Crude Shock Fluid -7- -

Gin Binding Protein

I Glu/Asp Alndlng Protein

FIG. 3. Ouchterlony double diffusion precipitin test. The central well of an 0.8% agar plate contained 50 pl of undiluted antisera prepared against isoelectric focused glutamate-aspartate binding protein. Other wells contained (as indicated) 10 pg of glutamate-aspartate binding protein, 20 fig of glutamine binding protein, or 1670 pg of crude shock fluid.

I 6 15- P’

do d

f

O50 5 I

51 0

r2 ,cm2 51 5

FIG. 4. Long column meniscus depletion sedimentation equi- librium molecular weight determination. Glutamate-aspartate binding protein in 6 M guanidine-HCl and 20 mM sodium borate, pH !1.2, was centrifuged at 33,600 rpm at 15” until equilibrium was reached. The In of the relative fringe displacement is plotted versus the square of the radial distance from the center of the rotor. Attainment of equilibrium was assumed when no change in interference pattern was observed over a 6-hour period. The total length of the run was 28 hours.

The pure glutamate-aspartate binding protein represented a 290-fold purification from the crude shock fluid with a recovery of 45c/, of the initial units (Table 1). The purified protein appeared to be homogeneous by electrophoresis at pH 3.5 (23) and pH 4.5 (21) in 7, 9, and 11% polyacrylamide gels, as well as by sodium dodecyl sulfate polyacrylamide gel electrophoresis in a 10% gel (24).

The apparent molecular weights of the pure glutamate-nspar- tntc binding protein determined by a Hedrick and Smith (22) treatment of the data obtained from the different percentage of gels for the pH 3.5 and pH 4.5 buffer systems and as determined by a Webcr and Osborn treatment (24) of the sodium dodecyl sulfate data \vere 29,800,31,000, and 31,500, respectively. These values compare well with the value of 29,500 obtained from Sephadex G-200 chromatography. Apparent homogeneit,y and a molecular weight of 30,400 were also established by sedimentation

2577

TABLE II Amino acid analysis of glutamate-aspartate binding protein

Amino acid

Glycine. .................... Alanine ..................... Valine ...................... Leucine .................... Isoleucine .................. Serine ...................... Threonine ................. Aspartic acid. .............. Glutamic acid .............. Lysine ...................... Histidine ................... Arginine ................... Phenylalanine .............. Tyrosine ................... Tryptophan" ............... Half-cystined. .............. Methionine” ................ Proline. .................... Ammonia ................... Molecular weight (from sum.

mation) .................

Mean residues/ histidine?

8.1 16 11.5 23

8.0 16 11.1 22

5.6 11 7.8 16 7.0 14

23.2 46 15.2 30 15.4 31

1.0 2 4.3 9 6.5 13 2.1 4 2.0 4 1.0 2 3.8 8 7.0 14

11.9 24

15,995

Residue/molecule*

n Average of three determinations on asingle sample hydrolyzed 22 hours in 6 N HCl.

* Calculated on the basis of a molecular weight of 32,000. c Determined by measurement of 288:280 ratio in 6 M guanidine-

HCl according to the procedure of Edelhoch (32). d The half-cystine and methionine values were determined

from cysteic acid and methionine sulfone residues, respectively, after 22 hours of hydrolysis in 6 N HCl of performic acid oxidized (31) protein.

equilibrium in 6 hf guanidine HCl and 20 rnhf sodium borate buffer, pH 9.2, as described by Chervenka (34). The molecular weight determination was based on a slope of 1.47 (Fig. 4) and a partial specific volume of 0.731 calculated from the amino acid composition (35).

The amino acid composition of native and performic acid oxidized glutamate-aspartatp binding protein is shown in Table II. A minimum molecular weight of 32,000 was estimated assuming 2 histidines per molecule. The glutamate-aspartatc binding protein is unique from a number of other binding proteins since 2 cysteic acid residues were found per molecule of binding protein suggesting the possibility of an intramolecular disulfide bond. The large number of lysine and arginine residues accounts for the basic nature of the protein.

The glutamate binding activity of native glutamate-aspartate binding protein was not affected by the presence of 10 mhl dithiothreitol in the assay mixtures (Table III), suggesting the absence of or lack of participation of an exposed cysteine or disulfide in the substrate binding. The glutamate-aspartate binding protein lost the capacity to bind substrate in the presence of 6 hl guanidine-HCl or 6 hl urea. The capacity for substrate binding was restored by dilution or dialysis. Khcn 10 rn>f dithiothreitol was included in the above treatments, the binding protein did not regain substrate binding activity upon dilution or overnight dialysis, suggesting that the protein contains a buried disulfide group which is important in protein conformation and the collformatioll-dcpcildeiit binding activity.

Although other binding proteins arc rcportcd to be stable to

by guest on Decem

ber 19, 2020http://w

ww

.jbc.org/D

ownloaded from

2578

TABLE III

Effect of dithiothreitol on glutamate binding

Treatment’=

None 2 mM Dithiothreitol 6 M Guanidine-HCl

6 M Guanidine-HCl + 2 mM dithiothreitol

post-treatment* incubation time

min 1 1 1

10 60

1 10

-

( Xutamate bindingC

nnol/mg

28.6 . 26.4

2.5 15.5 21.3

1.7 1.8

a Samples were incubated at 37” for 30 min at a protein concen- tration of 0.27 mg/ml in 50 mM sodium acetate, pH 5.5, with the indicated additions.

* The treated samples were diluted 1:lO with 50 mM sodium acetate, pH 5.5, and incubated at room temperature for the times indicated.

c The activity of the binding protein was determined with 25 ~1 of the diluted sample at a final glutamate concentration of 14 pM by the membrane filter binding assay.

/

4’ 1 I I I I I 4 5 6 7 0 9 IO

PH

I II

1 I2

FIG. 5. The pH dependence of glutamate binding. The bind- ing of glutamate was measured by the nitrocellulose filter assay described under “Methods.” The assay mixture contained 16 pg of glutamate-aspartate binding protein per ml, 1.66 PM [WI- glutamate, and the final pH values as indicated; A, 0.1 M sodium acetate-O.1 M acetic acid buffer; l ,O.l M sodium dimethyl arsenite- 0.1 M hydrochloric acid; A, 0.1 M dibasic potassium phosphate- 0.1 M monobasic potassium phosphate; 0, 0.1 M Tris-0.1 M hy- drochloric acid; and w, 0.1 M sodium hydroxide-O.1 M sodium bicarbonate.

boiling (2), the glutamate-aspartate binding protein retains only 50% of the binding activity after 2 min of incubation in a boiling water bath.

The glutamate-aspartate binding protein at concentrations of 60 pg/ml was not found to be active in assays for the following activities: glutaminase A and B (EC 3.5.1.2 .) (36)) asparaginase I and II (EC 3.5.1.1.) (37), glutamic dehydrogenase (TPN- linked) (EC 1.4.1.4. ) (38), glutamine synthetase (EC 6.3.1.2) (39), glutamate synthase (EC 1.4.1.X) (36), aspartase (EC 4.3.1.1.) (40), and glutamic-oxalacetic transaminase (EC 2.6.1.1.) (37). Some difficulty was encountered in the equi- librium dialysis assay of glutamate and aspartate binding activity which may have been due to the interaction of these negatively charged amino acids with the dialysis membrane. Therefore, all assays used in the characterization of the binding protein

30 t

GLUTAMATE (pMi

I I I I 10 20

ASPARTATE CPM)

FIG. 6. Concentration dependence of glutamate and aspartate binding. Binding was measured by the nitrocellulose filter assay in 50 mM potassium phosphate, pH 7.0. 0, nanomoles bound; 0, free ‘4C-ligand to bound W-ligand, i.e. s/v transformations,

of binding data. Top, glutamate binding; bottom, aspartate binding.

were, unless stated otherwise, performed by the nitrocellulose filter assay (12, 13). It was fortuitous that the glutamate- aspartate binding protein binds to the nitrocellulose filter with 90 to 100% efficiency over a wide range of pH and ionic strength values. The results of experiments comparing the filter assay to the equilibrium dialysis assay were virtually indistinguishable.

The binding of glutamate by glutamate-aspartate binding protein displays a wide pH optimum over the range of pH 5.5 to 9 (Fig. 5). Sodium carbonate was found to inhibit binding of glutamate; this inhibition was, however, due to the carbonate. Sodium chloride at concentrations as great as 100 mM neither inhibited nor stimulated either aspartate or glutamate binding in assays using either 50 mM Tris-HCl, pH 8.0, or 50 mM potassium phosphate, pH 6.8, buffers.

The glutamate-aspartate binding protein binds L-glutamate and Gaspartate with Ko values of 0.7 and 1.2 pM, respectively (Fig. 6). Scatchard plots (41) of glutamate and aspartate bind- ing indicate that 1 mol of either ligand is bound per mol of pro- tein, (not shown). Both glutamate and aspartate binding was found to be reversible, and competition data (not shown) indicate

by guest on Decem

ber 19, 2020http://w

ww

.jbc.org/D

ownloaded from

2579

TABLE IV

Specificity of glutamate-aspartate binding protein”

W-Ligand

n-Glutamate (Ko = 0.7 pd)

n-Aspartate (KD = 1.2 PM)

Unlabeled competitorb

Glutamate Aspartate Glutamic acid-r ethyl ester Glutamine Glutamic acid-y methyl ester r-Glutamyl hydroxamate

PM 0.8 1.1

72 280 360

2106

Aspartate 1.2 Glutamate 0.4 Asparagine 270

0 Binding was determined by membrane filtration. * L Isomers. 0 Determined from Dixon plots.

that glutamate and aspartate are probably competitive for the same site. There are certainly other explanations; however, this is the simplest. The Ki values determined for unlabeled glu- tamate and unlabeled aspartate competition for labeled glu- tamate binding are 0.8 and 1.1 PM, respectively, and for labeled aspartate binding are 0.4 and 1.2 PM, respectively. Of the glutamate and aspartate analogs tested for inhibition of either glutamate or aspartate binding (Table IV) only n-glutamine and n-asparagine showed some inhibition.2 The K; of L-glutaminefor n-glutamate binding was 280 pM and the Ki of n-asparagine for n-aspartate binding was 270 PM. n-Alanine provided a weak competitive inhibition of glutamate binding and a Ki of 9.3 mM was determined. The K; values for these amino acids were sufficiently high so as to make it difficult to rule out contamina- tion. Glutamate binding is stereospecific as n-glutamate was not an inhibitor. While a-methyl glutamate was not an inhibitor, alterations at the y-carboxyl end of glutamate were somewhat more readily tolerated and the K; values for the following com- pounds were determined: tglutamic acid-y-ethyl ester, 72 PM; n-glutamic acid-y-methyl ester, 360 PM; and n-glutamic acid-y- hydroxamate, 2.1 mM.

The glutamate-aspartate binding protein exhibited native tryptophan fluorescence; an excitation maximum of 284 nm and emission maximum of 335 nm were determined for the protein in sodium acetate buffer at pH 5.8. In the presence of n-glutamate, 10 PM, an excitation maximum of 280 nm and emission maximum of 330 nm were observed; a drop in amplitude (approximately 7ye) was also noted. The emission spectrum shift toward the blue with n-glutamate was not observed with n-glutamate or a-methyl glutamate, or other analogs which do not inhibit glutamate binding. The magnitude of the shift and amplitude drop in the presence of n-glutamate exhibited a pH dependence similar to that observed in binding studies. Oxidation of the 4 tryptophan residues with N-bromosuccinimide in 0.1 M sodium acetate at pH 5.8 completely destroyed the fluorescence at a molar ratio of 3.5 N-bromosuccinimide molecules per tryptophan (Fig. 7). The oxidation was not prevented by tglutamate and

* Other compounds tested and providing no significant inhibi- tion, indicating Ki values greater than observed for L-alanine, were n-a-aminobutyrate, n-proline, L-histidine, n-norleucine, L-leucine, L-4-azaleucine, L-glycine, n-serine, cycloleucine, L- cysteine, n-methionine, n-methionine sulfoximine, n-arginine, succinate, pyruvate, oxaloacetate, a-ketoglutarate, n-fi-hydrox- amyl-aspartate, and r.+aspartyl hydraeide.

- 80

i .!.

70

g 60

iti 0 50 2 g 40 3 $ 30

20

IO

0, 02 .04 .06 .06 N-BROMO SUCCINIMIDE, ,u moles

FIG. 7. Titration of glutamate-aspartate binding protein with N-bromosuccinimide. Glutamate-aspartate binding protein, 0.22 mg/ml, was titrated with increasing amounts of N-bromosuc- cinimide. 0, fluorescence of protein titrated in 0.1 M sodium acetate buffer, pH 5.8; W, fluorescence of protein titrated in 0.1 M sodium acetate-8 M urea, pH 4.0; 0, glutamate binding activity of protein titrated in 0.1 M sodium acetate buffer, pH 5.8. Fluores- cence emission was measured at 335 nm, excitation was at 285 nm. Binding was measured by the nitrocellulose filter assay.

the binding activity was quantitatively destroyed with the loss of tryptophan fluorescence.

DISCUSSION

The glutamate-aspartate binding protein released by E. cd W3092 during osmotic shock is typical of other characterized periplasmic binding proteins (see recent reviews (2, 3)). These proteins have molecular weights ranging from 23,OW for the glutamine binding protein (42) to 42,000 for the phosphate bind- ing protein (43). The isoelectric points of the different proteins vary from a low of 4.5 for the leucine-specific binding protein3 to a high of 9.69 determined in this study for the glutamate-aspartate binding protein.

It has previously been suggested that an unusual feature of periplasmic binding proteins was the absence of cysteine residues. However, recent studies indicated the presence of cysteic acid residues in the hydrolysates of the arginine specific binding pro- tein, the lysine-arginine-ornithine binding protein (44), and the cystine-diaminopimelic acid binding protein (45). The two half-cystines of the glutamate-aspartate binding protein appear to be buried and an intramolecular disulfide bond may be neces- sary for the protein conformation required for substrate binding. Binding conformation is also lost when the glutamate-aspartate binding protein reacts with antibody.

Our preliminary fluorescence data indicate that a tryptophan(s) of the glutamate-aspartate binding protein may undergo an environmental change on binding of a substrate. The flu- orescence emission maximum shift toward the blue suggests the exposure of tryptophans to a more hydrophobic environment.

B C. E. Furlong, unpublished results.

by guest on Decem

ber 19, 2020http://w

ww

.jbc.org/D

ownloaded from

for assistance with the sedimentation equilibrium studies.

REFERENCES 1. WILLIS, R. C., AND FURLONG, C. E. (1974) Fed. Proc. 33.1394 2. HEPPEL. L. A. (1969) J. Gen. Phusiol. 64.955-1095 3. OXENDE’R, D. L: (19i2) Ann. Rev. Biochek 41, 777-814 4. HAZELBAUER, G. L., AND ADLER, J. (1971) Nature New Biol.

230, 101-104 5.

6.

7.

MINER, K. M.. AND FRANK, L. (1974) J. Bacterial. 117. 1093- 1098.

,

BARASH, H., AND HALPERN, Y. S. (1971) Biochem. Biophys. Res. Commun. 3.681-688

LOMBARDI, F. J., AND KABACK, H. R. (1972) J. Biol. Chem. 247, 7844-7857

2580

The shift found in the binding studies is stereospecific for L- glutamate, and the amplitude decrease with substrate binding at different pH values correlates with the pH dependence of bind- ing. Tryptophan(s), or protein conformation dependent on tryptophan(s), appears to be necessary for substrate binding. Similar substrate dependent fluorescence changes have been observed for the galactose binding protein (46) and glutamine binding protein (25). In the studies with glutamine binding protein, oxidation with N-bromosuccinimide was, in contrast to our results with glutamate-aspartate binding protein, prevented by substrate and binding activity was not destroyed.

A great deal of emphasis has been placed on similarities be- tween the K, values and specificity of a transport and the bind- ing constants and specificity of the osmotic shock releasable binding protein associated with the system (3). The association of a binding protein with a particular transport system has been based on (a) apparent coregulation of transport activity with binding protein level and (5) concomitant loss of binding protein or binding protein activity (either genetically or artificially, i.e. osmotic shock) and loss of transport activity. The inferences from these results are that the binding proteins are the recogni- tion and rate determining components in the transport of par- ticular substrates. Recent studies with cells grown and assayed under a number of different conditions as well as with membrane vesicles prepared from these cells suggest that under certain conditions the glutamate-aspartate binding protein may be involved in a rate-limiting step of glutamate and aspartate transport (47).

It is interesting that the attractant receptors for chemotaxis appear to be released by the osmotic shock procedure (4, 48). The fact that so far no chemoreceptor negative mutants have been found among the aspartate taxis negative mutants4 coupled with the fact that membrane vesicles demonstrate an aspartate specific uptake would suggest the possibility that the glutamate- aspartate binding protein could serve as one receptor and a component of the aspartate preferring transport system which is retained in vesicles as a second receptor. Consistent with this possibility, Adler has found that aspartate strongly inhibits glutamate taxis whereas a large excess is required to inhibit aspartate taxis.’ Experiments are underway to explore this possibility.

Acknowledgments-The authors wish to thank Diana Di Iulio (recipient of a Summer Fellowship from the California Founda- tion for Biochemical Research) for assistance with preliminary experiments associated with this work and Dr. G. M. Hathaway

4 J. Adler, personal communication.

8. ENGLUND, P. T., HUBERMAN, J. A., JOVIN, T. M., AND KORN- BERG, A. (1969) J. Biol. Chem. 244, 3038-3044

9. WILLIS, R. C., MORRIS, R. G., CIRAKOGLU, C., SCHELLENBERG, G. D., GERBER, N. H., AND FURLONG, C. E. (1973) Arch. Biochem. Biophys. 161, 64-75

10. OSHIMA, R. G.,.WILLIS, R. C., FURLONG, C. E., AND SCHNEI- DER, J. (1974) J. Biol. Chem. 249.6033-6039

11. FuRL~NG,‘~. E., MORRIS, R. G., KANDRACH, M., AND ROSEN, B. D. (1972) Anal. Biochem. 47, 514-526

12. FURLONG, C. E., AND WEINER, J. H. (1970) Biochem. Biophys. Res. Commun. 38, 1076-1083

13. LEVER, J. (1972) Anal. Biochem. 60, 73-83 14. PATTERSON, M. S., AND GREEN, R. C. (1966) Anal. Chem. 37.

854-857 15. LOWRY, 0. H., ROSEBROUGH, N. J., FARR, A. L., AND RANDALL,

R. J. (1951) J. BioZ. Chem. 193, 265-275 16. SHERWOOD, L. M., AND POTTS. J. T.. JR. (1965) J. BioZ. Chem.

17.

18. 19. 20.

21.

22.

23.

24.

25.

26.

27. 28. 29.

30. 31.

32. 33.

34. 35.

36.

37.

38.

39.

40.

41. 42.

43.

44. 45.

46.

47.

48.

240,3799-3805 , I .

LEVEY, H. B., AND SOBER, H. A. (1960) Proc. Sot. Exp. BioZ. Med. 103, 25&252

STUDIER, F. W. (1972) Science 176.367-376 STUDIER, F. W. (1973) J. Mol. BioZ. 79, 237-248 FAIRBANKS, G., STECK, T. L., AND WALLACH, D. F. H. (1971)

Biochemistry 10, 2606-2617 REISFELD, It. A., LEWIS, 0. J., AND WILLIAMS, D. E. (1962)

Nature, 196, 281-283 HEDRICK, J. L., AND SMITH, A. J. (1968) Arch. Biochem. Bio-

phys. 126, 155-164 PANYIM, S., AND CHALKLEY, R. (1969) Arch. Biochem. Biophys.

130, 337-346 WEBER, K., AND OSBORN, M. (1969) J. BioZ. Chem. 244, 4406-

4412 WEINER, J. H., AND HEPPEL, L. A. (1971) J. BioZ. Chem. 246,

6933-6941 WILLIS, R. C., AND FURLONG, C. E. (1974) J. BioZ. Chem. 249,

6926-6929 VESTERBERG, 0. (1971) Methods Enzymol. 22, 389412 ANDREWS, P. (1964) Biochem. J. 91, 222-233 NEU, H. C., AND HEPPEL, L. A. (1965) J. BioZ. Chem. 240,

3685-3692 ANRAKU, Y. (1968) J. BioZ. Chem. 243, 3123-3127 HATHAWAY, G. M., KIDA, S., .%ND CRAWFORD, I. P. (1969)

Biochemistry 8, 989-997 EDELHOCH, M: (1967) Biochemistry 6,1948-1954 Lru. T.-Y.. AND GANG. Y. H. (1971) J. BioZ. Chem. 246.2842-

2ti8 ’ CHERVENKA, C. H. (1970) Anal. Biochem. 34,24-29 COHN, E. J., AND EDSALL, J. T. (1943) Proteins, Amino Acids,

and Peptides, pp. 370-381, Reinhold, New York PRUSINER, S., MILLER, R. E., AND VALENTINE, R. C. (1972)

Proc. Nat. Acad. Sci. U. S. A. 69, 2922-2926 SCHWARTZ, J. H., REEVES, J. Y., AND BROOME, J. D. (1966)

Proc. Nat. Acad. Sci. U. S. A. 66, 1516-1519 HALPERN, Y. S., AND LUPO, M. (1965) J. Bacterial. 90, 1288-

1295 WOOLFOLK, C. A., SHAPIRO. B.. AND STADTMAN, E. R. (1966)

Arch. B&hem. biophys. i16,‘177-192 MARCUS, M., AND HALPERN, Y. S. (1969) Biochim. Biophys.

Acta 177, 314-320 SCATCHBRD, G. (1949) Ann. N. Y. Acad. Sci. 61,660672 WEINER. J. H.. FURLONG. C. E.. AND HEPPEL. L. A. (1971)

Arch. biocheh. Biophys.‘142, 715-717 ’ ~ ’ MEDVECZKY, N., AND ROSENBERG, H. (1970) Biochim. Biophys.

Acta 211, 158-168 ROSEN, B. P. (1973) J. BioZ. Chem. 246, 1211-1218 BERGER, E. A., AND HEPPEL, L. A. (1972) J. BioZ. Chem. 247,

7684-7694 Boos, W., GORDON, A. S., HALL, R. E., AND PRICE, H. D.

(1972) J. BioZ. Chem. 247,917-924 WILLIS, R. C., AND FURLONG, C. E. (1975) J. BioZ. Chem. 260,

2581-2586 AKSAMIT, R., AND KOSHLAND, D. E. (1972) Biochem. Biophys.

Res. Commun. 48, 1348-1353

by guest on Decem

ber 19, 2020http://w

ww

.jbc.org/D

ownloaded from

R C Willis and C E Furlongfrom Escherichia coli K12 strain W3092.

Purification and properties of a periplasmic glutamate-aspartate binding protein

1975, 250:2574-2580.J. Biol. Chem. 

  http://www.jbc.org/content/250/7/2574Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/250/7/2574.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on Decem

ber 19, 2020http://w

ww

.jbc.org/D

ownloaded from