120 Btochimica et Biophvstca A eta 830 (1985l 120-126 Elsevier

BBA 32262

4-Aminobutyrate aminotransferase. Conformational changes induced by reduction of pyridoxal 5-phosphate

Soo-Young Choi and Jorge E. Churchich Department of Biochemistry. Universtty of Tennessee, Knoxville, TN, 37996-0840 (U.S.A.)

(Received March 12th, 1985)

Key words: 4-Aminobutyrate aminotransferase; Conformational change; Pyridoxal phosphate

Conformational changes induced in 4-aminobutyrate aminotransferase (4-aminobutyrate:2-oxoglutarate aminotransferase, EC 2.6.1.19) by conversion of pyridoxal-5-P to pyridoxyl-5-P were examined by two independent methods. The reactivity of the SH groups of the reduced enzyme is increased by chemical modification of the cofactor. 1.8 SH per dimer of modified enzyme react with DTNB, whereas 1.2 SH per dimer of the native enzyme react with the attacking reagent under identical experimental conditions. The modified and native forms of the enzyme bind the fluorescent probe ANS, but the number of binding sites for ANS is increased as result of conversion of P-pyridoxal to P-pyridoxyl. After the eonformationai changes onset by reduction of the cofactor, the modified enzyme binds one molecule of pyridoxal-5-P with a K d of 0.1 /tM to become catalytically competent. The catalytic site of the reduce enzyme was probed with P-pyridoxal analogs. Like resolved 4-aminobutyrate aminotransferase, the reduced species recognize the phosphorothioate analog and regain 40% of the total enzymatic activity. Since the catalytic parameters of reduced and native 4-aminobutyrate aminotransferase are indistinguishable, it is concluded that the additional catalytic site of the reduced enzyme is functionally identical to that of the native enzyme.

Introduction

4-Aminobutyrate aminotransferase (4-amino- butyrate: 2-oxoglutarate aminotransferase, EC 2.6.1.19) from pig brain contains 1 mol of pyri- doxal-5-P firmly bound to the dimeric structure [1].

~ P-NMR and amino acid seqeucne studies have indicated that pyridoxal-5-P forms a Schiff base with an ~-amino group of a lysine residue and that the phosphate group in its dianionic form con- tributes to the correct alignment of the cofactor through strong electrostatic interactions with posi- tively charged amino acids of the catalytic domain [21.

Abbreviations: DTNB, 5,5'-dithiobis(2-nitrobenzoic acid); ANS, 8-anilinonaphthalene 1-sulfonate.

A second catalytic site on the dimeric enzyme becomes functional only after specific chemical modification of the molecule of cofactor tightly bound to the enzyme [1]. Although the enzyme is made up of two identical subunits, the mechanism by which chemical modification of one catalytic site results in activation of a second catalytic site remains to be elucidated.

It has been the purpose of this work to detect conformational changes of the protein induced by conversion of pyridoxal-5-P to pyridoxyl-5-P, and to probe the binding and functional properties of the second catalytic site of reduced 4-amino- butyrate aminotransferase.

Experimental procedures

Purification of enzyme. The purification of 4- aminobutyrate aminotransferase was performed by

0167-4838/85/$03.30 ~:: 1985 Elsevier Science Publishers B.V. (Biomedical Division)

a method developed in our laboratory [1]. The purification steps were conducted at 4°C. Succinic semialdehyde dehydrogenase was purified from pig brain by a method already described [3]. The pyridoxal-5-P content of the purified aminotrans- ferase was determined by the method of Wadda and Snell [4]. Protein concentration was de- termined by the colorimetric method of Lowry et al. [5].

Enzymatic assay. A coupled assay system con- sisting of two purified enzyme, i.e., 4-amino- butyrate aminotransferase and succinic semial- dehyde dehydrogenase, was used to study the cata- lytic conversion of 4-aminobutyrate to succinic semialdehyde. Enzymatic assays were performed in 0.1 M sodium pyrophosphate (pH 8.4) con- taining 5 mM NAD ÷, 30 mM 4-aminobutyrate and 10 mM 2-oxoglutarate. Initial rate measure- ments were carried out by monitoring the changes in absorbance at 340 nm for at least 2 min. A unit of enzyme activity is define as that amount of enzyme which produces 1 ~mol /min of succinic semialdehyde at 25°C.

Polyacrylamide gel electrophoresis. The enzyme preparations were examined by polyacrylamide gel electrophoresis according to the original procedure of Davis [6]. Electrophoresis on gels containing 7.5% polyacrylamide, 1% 2-mercaptoethanol and 0.1% sodium dodecyl sulfate was performed according to the method of Weber and Osborn [7]. Protein bands were detected by staining with Coomassie blue dye.

Spectroscopy. Spectrophotometric measure- ments were carried out in either a Cary Model 15 or an Aminco DW-2 Double Beam Spectropho- tometer. Fluorescence spectra were recorded in a precision spectrofluorimeter equipped with two Bausch and Lomb Monochromators. The slits of the monochromators were set to give a bandwidth of 3 nm.

Titration of thiol groups. The number of reactive thiol groups was determined with DTNB using the procedure of Ellman [8]. A molar extinction coeffi- cient of 13600M 1. cm- l at 412 nm was used for the determination of the concentration of the an- ion 2-nitromercaptobenzoate.

Isolation of mitochondria. Mitochondria were isolated from whole pig brain using a procedure adapted from Schnaitman and Greenawalt [9].

121

Glutamate dehydrogenase was used as mitochon- drial marker enzyme was assayed at 25°C accord- ing to the method of Beaufay et al. [10].

Synthesis of P-pyridoxal aminooxyacetate. The synthesis of P-pyridoxal aminooxyacetate was per- formed according to the procedure of Severin et al. [11]. A methanolic solution of dipotassium pyri- doxal-5-P (1 mmol) was added in small portions to 1 mmol of the potassium salt of aminooxyace- tate (carboxymethoxylamine) in dry methanol (total volume 15 ml) and allowed to react for 1 h at 37°C. The mixture was evaporated and the precipitate was dissolved in water, acidified with HC1 to a pH of 3.5 and applied to a column (2 × 50 cm) of Amberlite XE-G4 (acid form). The column was eluted with water and fractions ab- sorbing at 340 nm were collected and freeze-dried. The dry residue was crystallized; the purity of the compound was tested by thin-layer chromatogra- phy on a silica-gel plate using ethyl acetate/pyri- d ine/water (2 : 1 : 2) as eluting solvent.

Synthesis of the phosphorothioate analog of pyri- doxal 5-phosphate. Pyriodoxai kinase (1 mg) was incubated with pyridoxal (0.1 mM), [3,S]ATP (0.1 mM) and cobalt acetate (0.1 mM) in 10 ml 0.1 M KCI at pH 6.2. The reaction was allowed to pro- ceed at 37°C in the dark for 12 h. The reaction mixture was concentrated to a final volume of 2 ml and applied to a column (15 × 1 cm) of DEAE-cellulose (DE-52, Whatman) equilibrated with 0.1 M ammonium acetate (pH 4).

The column was eluted with the same buffer, and the fractions collected were adjusted to pH 7 with NH4OH prior to spectrophotometric mea- surements. Fractions absorbing at 390 nm were combined and concentrated in a lyophilizer. The purity of the compound was tested by thin-layer chromatography, absorption and 31p-NMR spec- troscopy.

The absorption spectrum of the phosphorothio- ate analog of pyridoxal-5-P shows an intense ab- sorption band centered at 389 nm (~ = 4.6. 103), and it gives a phosphorus resonance signal which is shifted (21.8 ppm) when compared to phor- phoric acid. In marked contrast to pyridoxal-5-P, the phosphorothioate analog is not digested by alkaline phosphatase.

Materials. Pig brains were obtained from Lays Packing Company, Knoxville, TN. Sephadex G-25,

122

DEAE-Sephadex and CM-Sephadex were pur- chased from Pharmacia. DTNB was from Aldrich Chemical Company.

R e s u l t s

4-Aminobutyrate aminotransferase in crude mito- chondrial preparations

Mitochondrial proteins at a concentration of 1 m g / m l in 10 mM potassium phosphate buffer (pH 7.0) were allowed to incubate with various con- centrations of pyridoxal-5-P at 37°C for 1 h. After incubation, the samples of protein were tested for aminotransferase activity using 4-aminobutyrate as substrate.

The results of this experiment, included in Ta- ble I, indicate that the specific activity of the enzyme is not affected by addition of exogenous pyridoxal-5-P. Samples preincubated with excess pyridoxal-5-P were reduced with NaBH 4 (3 m g / m l ) at 4°C and then dialyzed against 10 mM potassium phosphate buffer (pH 7.0)/1 mM 2- mercaptoethanol. After dialysis, all the samples are enzymatically inactive. Catalytic activity is re-

stored to the same level as the holoenzyme when the sample reduced in the absence of exogenous pyridoxal-5-P was dialyzed, mixed with pyridoxal- 5-P and incubated for 1 h at 37°C prior to en- zymatic assays (Table I).

Under the experimental conditions chosen for our studies, binding of pyridoxal-5-P to other mitochondrial proteins takes place, as revealed by the fact that the catalytic activity of glutamate dehydrogenase is impaired following addition of pyridoxal-5-P to the mitochondrial proteins.

The preceding results demonstrate that in crude mitochondrial preparations 4-aminobutyrate ami- notransferase exhibits maximum catalytic activity without the addition of pyridoxal-5-P. At con- centrations of free pyridoxal-5-P higher than 10 /~M, the aminotransferase binds more cofactor, and it becomes irreversibly inactivated by reduc- tion with NaBH 4.

Similar studies conducted with an homogeneous preparation of 4-aminobutyrate aminotransferase have shown that the native enzyme contains one molecule of pyridoxal-5-P per dimer. Reduction with NaBH 4 abrogates catalytic activity, but ad-

TABLE I

SPECIFIC ACTIVITY OF 4-AMINOBUTYRATE AMINOTRANSFERASE IN MITOCHONDRIAL PREPARATIONS

(1). Sample of mitochondria (1 mg) reduced with NaBH4(3 mg) for 10 min at 4°C, then dialyzed against 10 mM potassium phosphate buffer (pH 7)/1 mM 2-mercaptoethanoi at 4°C. (2) Sample of mitochondria (1 mg) preincubated with pyridoxal-5-P (PLP) (0.1 raM) at 37°C for a h, then reduced with NaBH4 (3 mg) for 10 min at 4°C, and dialyzed against 10 mM potassium phosphate buffer (pH 7)/1 mM 2-mercaptoethanol at 4°C.

Treatment Specific activity Enzymatic activity (U/mg) (%)

Mitochondria

Mitochondria (1 mg)+PLP (0.1 mM)

Mitochondria (1 mg) reduced and dialyzed (1)

Mitochondria (1 mg) reduced, dialyzed and reconstituted with PLP (0.1 mM)

Mitochondria (1 mg)+PLP (0.1 mM), reduced and dialyzed (2)

Mitochondria (1 mg)+ PLP (0.1 mM), reduced, dialyzed and incubated with PLP (0.1 raM)

0.2 100

0.2 100

0 0

0.18 90

0 0

0 0

dition of pyridoxal-5-P saturates a second cata- lytic binding site, leading to restoration of full catalytic activity [12].

Since purified 4-aminobutyrate aminotransfer- ase behaves as 4-aminobutyrate aminotransferase in crude preparations, it seems reasonable to con- clude that the presence of one molecule of tightly bound pyridoxal-5-P per dimer is not an artifact introduced by the purification procedures. There- fore, all the experiments designed to detect struct- ural changes in the dimeric structure as result of reduction of the cofactor were performed with purified 4-aminobutyrate aminotransferase.

Conforrnational changes With the aim of detecting structural differences

between native and reduced y-aminobutyrate transaminase, the accessibility of the sulfhydryl groups of both forms of the enzyme was examined at neutral pH.

4-Aminobutyrate aminotransferase is easily in- activated by DTNB, a specific reagent of thiol groups. The reaction is reversible, and addition of 2-mercaptoethanol to the inactive species restores the original catalytic activity. The well-behaved kinetic system can be exploited to probe changes in the reactivity of the sulfhydryl groups due to structural alterations of the enzyme.

When the reduced enzyme was allowed to react with DTNB under pseudo-first-order conditions, the reaction of approx. 1.8 SH per dimer took place with an observed rate constant of 0.06 min- (Fig. 1). Under similar experimental conditions, the reaction of approx. 1.2 SH per dimer of the native enzyme proceeded with an observed rate constant of 0.047 min-1 (Fig. 1).

This increased accessibility of -SH groups of the attacking reagent DTNB reflects local conforma- tional changes of the protein elicited by chemical modification of one molecule of cofactor per di- mer.

In view of the preceding results, it was thought of interest to investigate whether the binding of ligands which do not interfere with catalytic events is also influenced by structural changes induced by chemical modification of the cofactor. The fluo- rescent probe ANS is ideally suited for those stud- ies because it binds to the protein at some site other than the catalytic site [2].

2.0

1.0 :¢

123

10 20 30 40 50 60

Time {rain)

Fig. 1. Time-course of reaction of DTNB (200 pM) with thiol groups of native 4-aminobutyrate aminotransferase (7 tiM) (O), and the reduced 4-aminobutyrate aminotransferase (7 pM) (e) at 25°C in 0.1 M potassium phosphate buffer (pH 7.0). MNB, 2-nitro-5-mercaptobenzoate.

Furthermore, the magnitude of changes in fluo- rescence yield associated with binding of the ligand are easily detected by fluorometric techniques.

The fluorescence enhancement that follows the addition of increasing concentrations of ANS to a fixed concentration of reduced enzyme (1 #M) was used to determine the stoichiometry of binding.

Form plots of ~/[L] vs. ~, it may be inferred that the reduced form of the enzyme possess two non-equivalent binding sites differing by a factor of 7 with respect to their dissociation constants, whereas the native enzyme exhibits only one bind- ing site for ANS (Fig. 2).

Binding of the cofactor The cofactor of the native enzyme exhibits two

absorption bands, centered at 330 nm and 415 nm, respectively.

Upon reduction with NaBH 4, the modified en- zyme displays an intense absorption band at 330 nm which is characteristic of P-pyridoxyl residues [11.

When samples of reduced enzyme at a con- centration of 6.5 #M were allowed to react with pyridoxal-5-P, and the absorbance at 415 nm re- corded after 15 min equilibration, it was found that the maximum increase in absorbance at 415

124

z

1 05 1.0 1.5

Fig. 2. Plots of ~/[L] versus ~ for the binding of ANS to 4-aminobutyrate aminotransferase (0) and reduced 4-amino- butyrate aminotransferase ( 0 ) at protein concentrations of 1 ~M in 0.1 M phosphate buffer (pH 7). From these plots, two binding sites for the reduced enzyme. K1 = 4.3 ~M and K 2 = 0.6 ~M, were obtained. 4-Aminobutyrate aminotransferase yields one dissociation constant K 3 = 6 ~aM.

"~ 0.05

0 .10

/

I I I 2 4 6

• ~, 6 e

I I I I I I I B 1 0 12 14 16 18 2 0

Fig. 3. Increase in absorbance at 415 nm following addition of increasing concentration of pyridoxal-5-P (PLP) to a fixed concentration of reduced ~aminobutyrate aminotransferase (6.5 /~M) at pH 7.4.

nm corresponded to the binding of 1 mol of pyridoxal-5-P per dimer (Fig. 3).

In order to determine the equilibrium dissocia- tion constant of the reduced-enzyme-pyridoxal-5-P complex, it was necessary to perform spectroscopic measurements at protein concentrations of 1 #M or lower.

Since absorption measurements at 415 nm were difficult to perform when the protein concentra- tion approached 1 #M, it was decided to de- termine the equilibrium dissociation constant by resorting to fluorimetric methods.

Quenching of protein fluorescence induced by binding of pyridoxal-5-P is a suitable method for determination of the equilibrium dissociation con- stant because protein fluorescence emitted at 350 nm can be measured with high precision at enzyme concentrations of 0.1/~M.

Fluorescence intensity measurements of sam- ples of reduced enzyme containing a fixed con- centration of protein (0.1 /~M) and varying con- centrations of pyridoxai-5-P were recorded after 15 min incubation at 25°C. The result of the fluorimetric titrations are given in Fig. 4.

The fraction of binding sites occupied (0) by the cofactor was determined by using Eqn. 1:

0 F { } - /-" [ E - PLP] ( 1 }

F{,- F~ [El,

~7

I I I I .__1 1 2 3 4 5

EpLp-] { 1 0 - 7 M )

Fig. 4. Changes in protein fluorescence at 350 nm (/~s0) (excitation 280 nm) induced by binding pyridoxal-5-P (PLP). Titrations conducted at a concentration of reduced 4-amino- butyrate aminotransferase of 0.1 p.M at 25°C in 0.1 M potas- sium phosphate (pH 7.0).

where F 0 is the fluorescence intensity of the en- zyme alone, Fro, the fluorescence emitted by the enzyme saturated with the cofactor, and F the fluorescence observed. [E-PLP] is the concentra- tion of reduced enzyme-pyridoxal-5-P complex and [E] t , the total concentration of enzyme.

By application of the law of mass action, it can be shown that the concentration of free ligand in solution pyridoxal-5-P is related to the fraction of occupied binding sites (0) by Eq. 2:

1 - 0 [PLP] (2)

A dissociation constant K d = 10 -7 M was ob- tained using the fluorescence quenching data of Fig. 4.

Binding of cofactor analogs P-Pyridoxal analogs modified at the 5'-phos-

phate side-chain were chosen as probes of the microenvironment of the catalytic binding site of reduced aminotransferase.

The phosphorothioate analog of pyridoxal-5-P carries a sulfur atom covalently bound to phos- phorus, whereas pyridoxal-5-P O-methyl ester is esterified at the phosphate side-chain. Modifica- tion of the phosphate side-chain by introducing a sulfur atom covalently bound to phosphorus renders a pyridoxal-5-P analog which serves as cofactor of both resolved and reduced forms of v-aminobutyrate transaminase (Table II).

TABLE II

RECONSTITUTION OF AMINOTRANSFERASE ACTIV- ITY BY SEVERAL COFACTOR ANALOGS: EFFECT OF PYRIDOXAL-5-P ANALOGS ON THE RECONSTITU- TION OF REDUCED 4-AMINOBUTYRATE AMINO- TRANSFERASE ACTIVITY

Samples of reduced enzyme (1 ~M) were preincubated with pyridoxal-5-P analogs (10 btM) for 1 h at 37°C in 0.1 M phosphate buffer prior to enzymatic assays.

Cofactor Apoenzyme Reduced activity (%) enzyme

activity (%)

Pyridoxal-5-P 100 100 Phosphorothioate analog 45 40 Pyridoxal-5-P monomethyl ester 0 0 P-Pyridoxal aminooxyacetate 0 0

125

In marked contrast to the phosphorothioate analog, pyridoxal-5-P O-methyl ester, carrying one negative charge on the phosphate side-chain, failed to interact with the reduced enzyme to yield cata- lytically active species.

Since the 5'-phosphate group of the coenzyme plays an important role in binding, as revealed by 31P-NMR spectroscopy [2], it seems reasonable to suggest that the phosphorothioate analog acts as a cofactor of the reduced enzyme, because it pre- serves the two negative charges required for cor- rect alignment of the ligand at the catalytic site.

P-Pyridoxal analogs modified at the 4'C posi- tion of the coenzyme were also tested as inhibitors of reduced aminotransferase. Among these deriva- tives, it was found that P-pyridoxyl aminooxyace- rate acts as a powerful competitive inhibitor with respect to pyridoxai-5-P (Table II). Its inhibitory power ( K , = 2 . 1 0 -s M) is due simply to the ability of this compound to displace pyridoxal-5-P from the catalytic site of the reduced enzyme.

Discussion

The results obtained using chemical and bio- physical methods indicate that conformational changes have taken place in 4-aminobutyrate aminotransferase as a result of specific chemical modification of the molecule of cofactor firmly bound to the dimeric structure of the protein.

Thus, the increased reactivity of thioi groups toward DTNB reflects structural fluctuations in the microenvironment surrounding sulfhydryl groups critically connected with catalytic activity.

On the other hand, changes in the binding parameters of the spectroscopic probe ANS com- plexed to the reduced enzyme can be related to conformational changes of the dimeric structure.

After the conformational changes onset by re- duction of the cofactor, the modified enzyme binds a second molecule of pyridoxal-5-P to become catalytically competent.

The binding of the cofactor to the modified enzyme was analyzed by spectroscopic methods; and the most salient feature of the binding experi- ments is the finding that pyridoxal-5-P interacts with the second binding site of the enzyme with a dissociation constant which is 30-fold lower than the K d value corresponding to the binding of a

126

second molecule of cofactor to the native enzyme [1].

Thus, reduction of the molecule of cofactor of the native enzyme brings about a conformational changes which facilitates the binding of a second molecule of cofactor which becomes catalytically competent.

Several P-pyridoxal analogs recognize the bind- ing site of modified 7-aminobutyrate transaminase and they act as powerful competitive inhibitors of the natural cofactor.

One of the pyridoxal-5-P analogs, the phos- phorothioate analog serves as cofactor of the mod- ified enzyme because introduction of a sulfur atom in the structure of pyridoxal-5-P does not perturb significantly the interaction of the phosphate side- chain of the coenzyme with positively charged amino acid residues of the protein.

Like the native enzyme, modified y-amino- butyrate transaminase catalyzes the transamina- tion of the amino acids fl-alanine and 4-amino- butyrate in the presence of a-ketoglutarate [1]. Since there is no significant difference between the catalytic parameters, gca t and Kin, it seems rea- sonable to conclude that the additional catalytic site of modified y-aminobutyrate transaminase is topographically and functionally identical to the catalytic site of the native enzyme.

Although the results of our studies have demon- strated that an additional functional catalytic site is uncovered upon chemical modification of the cofactor molecule firmly bound to the protein, the nature of subunit interactions in the native enzyme is extremely intriguing and needs further investiga- tion. Obviously, the physiological significance of the oligomeric structure of this enzyme, carrying only one molecule of cofactor, remains to be as- sessed. Whether the second binding site of the native enzyme is used for binding to other uniden- tified proteins or small molecular weight effectors remains to be investigated. In this connection, it is pertinent to note that y-aminobutyrate trans- aminase tends to form stable multienzyme com- plexes with succinic semialdehyde dehydrogenase, an enzyme involved in the catabolism of succinic semialdehyde [13].

An interesting aspect of the present work is the finding that the concentration of the free pyri-

doxal-5-P in brain mitochondria does not reach the saturation level ( K d = 3 /xM) of the second binding site of native aminotransferase. When the concentration of free pyridoxal-5-P is raised to the micromolar level, the catalytic activity of other vitamin B-6-independent enzymes is significantly affected (glutamate dehydrogenase).

The concentration of free pyridoxal-5-P is also very low in the cytosol (1 #M) [14], where most of the cofactor generated by two cytosolic enzymes, pyridoxal kinase and pyridoxine-5-P oxidase, is utilized for reconstitution of vitamin B-6-depen- dent enzymes. It has been shown that pyridoxal-5- P can be transported into isolated mitochondria prior to the binding to proteins located in both the intermembrane and matrix compartments [14]. Since the concentration of cofactor available for transport is very low, it seems reasonable to sug- gest that the catalytic site of 4-aminobutyrate aminotransferase is already saturated with the cofactor prior to translocation of this enzyme into the mitochondria matrix.

References

1 Churchich, J.E. and Moses, U. (1981) J. Biol. Chem. 256. 1101-1104

2 Churchich, J.E., Kim, D.S. and Schnackerz, K.D. (1983) FEBS Lett. 160, 221-225

3 Blaner, W.S. and Churchich, J.E. (1979) J. Biol. Chem. 254. 1794-1798

4 Wadda, H. and Snell, E.E. (1961) J. Biol. Chem. 236, 2089-2095

5 Lowry, O.H., Rosebrough, N.J.. Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275

6 Davis, B.J. (1964) Ann. NY Acad. Sci. 1212, 404-427 7 Weber, K. and Osborn, M. (1969) J. Biol. Chem. 244,

4406-4412 8 Ellman, G.L. (1959) Arch. Biochem. Biophys. 82, 70-76 9 Schnaitman, C. and Greenwalt, J.W. (1968) J. Cell. Biol. 38,

158-175 10 Beaufay, H.. Bendall, D.S., Baudhain, P. and DeDuve, C.

(1959) Biochem. J. 73, 623-630 11 Severin, E.S., Gulyavev, N.N., Khurs. E.N. and Khomotov,

R.M. (1969) Biochem. Biophys. Res. Commun. 35, 318-322 12 Kim, D.S. and Churchich, J.E. (1983) J. Biol. Chem. 258,

11768-11773 13 Hearl, W.G. and Churchich, J.E. (1984) J. Biol. Chem. 259.

11459-14463 14 Lawrence, A.L. and Li, T.K. (1982) J. Biol. Chem. 257,

14903-14906