Site-directed Mutagenesis of Rabbit Muscle Phosphofructokinase

Vol. 268, No. 33, Issue of November 25, pp. 24599-’24606,1993 Printed in U.S. A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY

0 1993 by The American Society for Biochemistry and Molecular Biology, Inc

Site-directed Mutagenesisof Rabbit Muscle Phosphofructokinase cDNA MUTATIONSATGLUTAMINE

200 AFFECT T H E ALLOSTERICPROPERTIESOFTHEENZYME* (Received for publication, September 28, 1992, and in revised form, June 18,1993)

Jauyi Lis,Xiaomao Zhu,Malcolm Byrnes, JeffreyW. Nelson, and Simon H. Changs From the Department of Biochemistry, Louisiana State University, BatonRouge, Louisiana 70803

Due to its central role inthecontrol of glycolytic flux Full-length cDNA for rabbit muscle phosphofructokinase hasbeen clonedand characterized (Li, J., Chen, and its allosteric properties, fructose-6-phosphate-1 kinase Z., Lu, L., Byrnes, M., and Chang, S. H. (1990) (PFK,l ATP:~-fructose-6-phosphate-l-phosphotransferase, Biochem. Biophys. Res. Commun. 170, 1056-1060). EC 2.7.1.11) has long been a subject of study in terms of both The 2.8-kilobase cDNA was inserted in the plasmid its enzymology and its protein chemistry (reviewed by Uyeda, vector pPL2 and transformed into Escherichia coli 1979; Dunaway, 1983). The catalytic and allostericmechacells deficient in endogenous phosphofructokinase ac- nisms and the structures of Escherichia coli PFK (EcPFK) tivity (DF 1020). The recombinant phosphofructokiand Bacillus stearotherrnophilus PFK (BsPFK) havebeen nase so prepared is nearly identical in kinetic proper- elucidated at the crystallographic level (Schirakiharaand ties and size of subunits to the enzyme isolated from Evans, 1988; Rypniewski and Evans 1989; Schirmer and Evrabbit muscle. On the basis of the sequence homology between the muscle and the bacterial phosphofructo- ans, 1990). In contrast, due to the absence of any primary kinases and the crystallographic structure of the latter, structure, mammalian PFK species had been studied only at the glutamine atposition 200 of the muscle enzyme is the enzymological level until Poorman andco-workers (1984) reportedtheamino acidsequence of rabbit muscle PFK implicated in the allosteric transitions. This residue (RMPFK). Their results showed that the primary sequence was replacedby alanine (Q200A), glutamate (Q200E), of RMPFK (molecular mass 80 kDa) appears essentially to or arginine (Q200R). The purified enzymes were anabe a direct repeat of bacterial PFK (molecular mass 36 kDa). lyzed for quaternary structure, activity, and allosteric alsobeenobservedin many properties. The native and all the altered enzymes are Thisduplicationpatternhas molecules including that of human liver tetramers. At pH 7.0, the wild-type enzymeis sensitive other eukaryotic PFK to inhibition by ATP at concentration above 0.6 mM, (Levanon et al., 1989), mouse liver (Gehnrich et al., 1988), and its activity responds to fructose 6-phosphatecon- and yeast (Heinisch et al., 1989), thus establishing anevolucentration cooperatively at high ATP concentration. tionaryrelationship between prokaryoticand eukaryotic In contrast, the mutated enzyme Q200R is virtually PFK. The sequence information of rabbit muscle PFK also insensitive to ATP inhibition up 7tomM. Thus at high promoted the cloning of the gene (Lee et al., 1987) and the ATP concentration, its activity responds to fructose 6- full-length cDNA (Li et al., 1990) of this enzyme. phosphate concentration in a manner similar to the Like the bacterial PFK, mammalian PFK responds to Fruactivated form of the native enzyme. Under the same 6-Pconcentration cooperatively (Passonneauand Lowry, conditions, mutant Q200E exhibits cooperative behav-1963). However, its activity issubject to much more complex ior only at much higher concentration of fructose 6- regulation by a variety of effectors and metabolites(reviewed phosphate. Mutant Q200A is active at pH 8.0 but inby Dunaway, 1983), including activation by AMP, inorganic active at pH 7.0. The native enzyme and all three phosphate, Fru-6-P, Fru-1,6-Pz, and Fru-2,6-Pz (Van Schafmutants are activated by inorganicphosphateand tingen et al., 1980; Pilkis et al., 1981; Uyeda et al., 1981) and fructose 2,6-bisphosphate and inhibited by citrate. inhibition by ATP and citrate. The structure and enzymology of mammalian PFK are further complicated by the complex tetramers formed by the assembly of three types of isozymic * This work was supported by National Institutes of Health Grant subunits (Dunaway and Kasten1987; Dunaway et al., 1988): DK31676. The AVIV 62DScirculardichroismspectrometer was M for muscle, L for liver, and C for a third type of subunit. obtained with fundsfrom Louisiana Education Quality Support Fund Each of these homo- or heterotetramers not only binds the Grant (1990-91)-ENH-13 and NIH Shared Instrumentation Grant substrates with different affinity, but also responds to the RR06710. The costs of publication of this article were defrayed in effectors with diverse sensitivity (Foe andKemp, 1985; Dunpart by the payment of page charges. This article must therefore be the binding of hereby marked “aduertisement” in accordance with18 U.S.C. Section away et al., 1988). Residues responsible for 1734 solely to indicate this fact. This work is dedicated to DavidW. H. Chang, 1971-1991. T h e nucleotide sequence(s) reportedin this paperhas been submitted totheGenBankTM/EMBLDataBankwith accessionnumber(s) 502702.

Part of this work was adapted from a dissertation in partial fulfillment of the requirementsfor the degree of Doctor of Philosophy. Present address: Dept. of Physiology, Columbia University, P and S Bldg. 11-420, 630 W. 168 St., New York, NY 10032. To whom correspondence shouldbe addressed Dept.of Biochemistry,Rm. 322 ChoppinHall,LouisianaStateUniversity,Baton Rouge, LA 70803. Tel.: 504-388-5147; Fax: 504-388-5321.

The abbreviations used are: PFK, phosphofructokinase; BsPFK, EcPFK, and RMPFK, the PFK of B. stearothermophilus, E. coli, and rabbit muscle, respectively; p f k , gene encoding PFK; QZOOR, Q200E, and Q200A, rabbit muscle PFK mutated at glutamine 200 to arginine, glutamic acid, andalanine, respectively; pPLS/RMPFK, plasmid containing full-length RMPFK cDNA the transcription of which is under the control of the PLpromoter of X phage; kb, kilobase pair; Fru-6-P, fructose-6-phosphate; Fru-1,6-P2, fructose-1,6-bisphosphate;Fru-2,6-P2, fructose-2,6-bisphosphate;DTT,dithiothreitol; TES, N-tris(hydroxymethy)methyl-2-aminoethanesulfonic acid; MWC, Monod-Wyman-Changeux model for allosteric proteins.

24599

24600

Mutagenesis of Rabbit Muscle Phosphofructokinase

substrates and effectors have been postulated based on the gation. The purified plasmid DNA was then cotransformed with a sequence homology among the NH2 and COOHhalves of second plasmid pCI857 (Remaut et al., 1981) encoding the thermolabile CI repressor into E. coli strain DF1020 which is deficient in both rabbit muscle PFK and bacterial PFK, and the crystalloEcPFK genes p f M and pfkB (Kotlarz and Buc, 1981; Daldal, 1983). graphic structure of the latter (Poormanet al., 1984; Hellinga The transformed cells were grown onLBagarplatescontaining and Evans, 1985). Leading the effort in the identification of ampicillin and kanamycin. Isolated colonies were streaked on M63 these residues, Valaitis and co-workers (1987) have located (Ausubel et al.,1989) agar plates using mannitol as the carbonsource. a heptadecapeptide at the This vector-host system and the restricted mediumprovide a visible the ATP inhibition site within carboxyl terminus of rabbit muscle PFK using the method of screening for the bacterial cellswhich express recombinant PFK. partial proteolysis. Other than this, little has been elucidatedClones of cells purified through this process were grown in 2 X yeast (YT) broth containing ampicillin and kanamycin. regarding the role of the amino acid residues in the catalytic extract-Tryptone The overnight culture was used to inoculate (at 1:lOO ratio) a large and regulatory functions of mammalian PFK. scale culture with the same composition. Incubation a t 28 "C was This work presents studies on the structural basis of the carried out until the absorbance a t 600 nm reached 0.9 to 1.0. The allosteric control of RMPFK using DNA technology. We have culture was brought to 42 "C rapidly and incubated at this temperacloned and expressed the full-length cDNAof this enzyme in ture for 3.5 h. Theculture was chilled on ice. Cellspelleted by were suspended in buffer I containing 0.5 mM ATP E. coli cells and performed site-directed mutagenesis on the centrifugation (see below) and eitherlyzed by sonication for enzyme purification, or glutamine residue a t position 200 (Q200) of RMPFK. The stored frozen until use. rationale for selecting GlnZwas the target for mutagenesis is Site-directed Mutagenesis-The I-kb 5' fragment of the cDNA by inference from the highly resolved crystallographic struc- containing the coding region for the NH2-terminal 310 amino acids tures of the R (relaxed)and T (tense)forms of BsPFK of RMPFK was excised from plasmid pPL2/RMPFK by digestion (Schirmer and Evans, 1990), and the similarity between the with XbaI and BamHI endonucleases (Fig. 1) and inserted by DNA ligase into the polylinker of pSELECT-1 phagemid DNA (Promega primary structuresof this bacterial PFK and RMPFK (Poor-Biotec Inc. Madison, WI)predigested by the same pair of restriction man et al., 1984). The glutamate 161of BsPFK, shown to be enzymes. The construct was transformed into J M 109 E. coli cells. involved in the allosteric transition, forms a salt bridge to The single-stranded DNA template was prepared by infecting the arginine 243 in the T state. In the R state, glutamate 161 transformed E. coli cells with helper phage R408 (Dotto and Zinder, rotates away, and a new salt bridge forms between arginine 1983). Two mutagenic oligonucleotides, one repairing theUAG triplet 200 occupies in the (3-lactamase gene of the phagemid and the other converting 252 and the substrate, Fru-6-P. Since glutamine glutamine (Q200) codon CAG into alanine GCG, arginine CGG, the homologous position in the RMPFK sequence, we propose the or glutamate GAG were annealed to the template DNA. The heterothat this residue is similarly important in the allosteric be- duplex was extended by T4 DNA polymerase and sealed by DNA havior. The results of this seriesof experiments are presented ligase. The double-stranded phagemid DNA was subsequently transformed into a repair deficient strain of E. coli (BMH 71-18 mutS). in this paper. The transformed cells were then grown on agar plates containing ampicillin. PlasmidDNA was prepared fromcolonies randomly picked and selected for correct mutations by sequence analysis using at the Cells and Plasmids-The general methods used for DNA analysis, the single dideoxynucleotide corresponding to the replacement construction, and purificationwere those described by Maniatis and target codon. Wild-type PFK template was included for comparison. co-workers (1982). The cloning and characterizationof the full-length The correctlymutagenized RMPFKcDNAfragment was excised from the phagemid DNA by digestion with XbaI and BamHI restricRMPFK cDNA has been described previously (Li et al., 1990). In order to insert this cDNA at the correct site and orientation down- tion enzymes, then ligated into pPLP/RMPFK plasmidDNA predigstream of the PLpromoter of the expression plasmid pPL2 (Remaut ested by the same pair of endonucleases. Procedures for transformaDNA into DF1020 E. coli et al., 1981), linkers with unique restriction siteswere added to both tion of the mutated pPL2/RMPFK plasmid the cDNA and theplasmid. This plasmid construct(pPL2/RMPFK, cells, selection of the clones expressing PFK, and DNA sequencing were the same asdescribed in the previous section. Fig. 1) was transformed in E. coli strain M5219 carrying a defective Purification of Wild-type and Mutated RMPFK-Cell pellets from prophage encodinga thermolabile repressor CIS57 (Greer, 1975). Minipreparations of plasmid DNA from purified clones were verified a 4- to 5-liter cultureof E. coli cells expressing wild-type or mutated as having the RMPFK insertby double-stranded sequencing analysis RMPFK were suspended at 15% (w/v)in bufferI (50 mM Tris phosphate, pH 8.0, 25 mM NaF, 0.1 mM EDTA, 1 mM DTT, and 0.1 (Chen and Seeburg 1985). Plasmids with the correct DNA sequence were prepared in larger quantity and purified through CsCl centrifu- mM phenylmethylsulfonyl fluoride) containing 0.5 mM ATP. Cells were lysed by sonication. The supernatant was brought to 35-55% saturation with ammonium sulfate. The protein precipitate was dissolved in 20 ml of bufferI containing 0.1 mM ATP and dialyzed against the same buffer. The solution was then loaded onto a Cibacron Blue F3GAagarose (Pierce ChemicalCo.)column(2.5 X 30 cm) equilibrium with the same buffer. The column was washed extensively with the same buffer until the absorbance a t 280 nm reached zero. PFK was eluted with buffer I containing 5 mM each of Fru-6-P and ATP in theabsence of magnesium ion. Fractions containing the PFK activity were combined and loaded onto a column (1.5 X 10 mm) of DEAE-cellulose (Whatman DE52) equilibrated with buffer I containing 0.1 mM ATP. The column was eluted stepwise by the samebuffer containing increasing concentrationof Tris phosphate. Themajority of PFK activity was eluted at 0.2 M phosphate. Fractions containing the PFK activitywere combined and concentratedby dialysis against 50% glycerol in buffer I containing 0.1 mM ATP. Aliquots from each step of the purificationwere analyzed for PFK activity (pH 8.0,lmM ATP, and 10 mM Fru-6-P, see below) andproteinconcentration (Bradford 1976). Purity of the enzyme was analyzed by SDS-polyFIG. 1. Plasmid vector, pPLS/RMPFK, for the expression acrylamide gel electrophoresis using PFK purified from rabbitmuscle of rabbit muscle PFK in E. coli cells. The h PLpromoter, Shine- (Sigma) as a standard. Gels were stained using Coomassie Brilliant Dalgarno box (SD), transcriptionterminationsite,and origin of Blue R-250 (Bio-Rad). The enzyme was dialyzed against either the pH 8.0 (50 mM Tris-HCl)orthepH 7.0 (50 mM TES) buffer replication are indicated. The coding regions for (3-lactamase (Amp') containing 1 mM DTT, 0.1 mM EDTA, and 0.1mM ATP before use. and RMPFK are indicated by solid boxes with arrowheads pointing in thedirection of transcription. Unique and relevant restriction sites Assay of PFKActiuity-PFK activity was assayed usingtwo different protocols (Foe and Kemp,1982). For determination of maximum in the RMPFK cDNA are given. EXPERIMENTALPROCEDURES

Mutagenesis of RabbitPhosphofructokinase Muscle PFK activity (V), the enzyme was assayed in 1ml of reaction mixture containing 50 mM Tris-HC1, pH 8.0, 5 mM MgCl,, 1 mM DTT, 0.2 mM NADH, 10 mM Fru-6-P, 1 mM ATP, 1 mM Fru-2,6-Pz, and the following auxiliary enzymes: aldolase, 40 pg/ml; glycerophosphate dehydrogenase, 24 pg/ml; and triose phosphate isomerase, 2 pg/ml. Under these conditions, PFK obeys Michaelis-Menten kinetics, and all enzymes are in the plateau region of the Fru-6-Psaturation curve. For analyzing the regulatory properties of the recombinant RMPFK or its mutants, the enzyme was analyzed in 1 ml of 50 mM NaTES, pH 7.0, and 150 mM KCl, 1 mM DTT, 0.2 mM NADH, and the same quantities of auxiliary enzymes. ATP and Fru-6-P were added as indicated and MgC12concentration was always 3 mM in excess of the ATP concentration. The reaction rate determined from the latter assays ( u ) was used to calculate the relative activity (u/V). The validity of this procedure is demonstrated in Fig. 6, where the values of u/V for all activated enzymes approach between 90 and 100% at high Fru-6-P concentration. Assays were performed in duplicates at each Fru-6-P concentration, and reactions were started by addition of Fru-6-P (Dunaway and Weber, 1974). Rate of oxidation of NADH was followed using a recording spectrophotometer (Gilford Response 11) equipped with a cell holder regulated at 30 "C. One unit of PFK activity is the quantity of enzyme which catalyzes the conversion of 1 pmol of Fru-6-PtoFru-l,6-Pz/min (Kemp, 1975). The entire procedure of PFK activity assay was repeated on two different preparations of purified enzymes. Analysis of Kinetic Data-The kinetic data were fit to threedifferent equations in order to determine the cooperativity of the kinetics toward Fru-6-P. The cooperative behavior of these enzymes under the different conditions was determined by comparing fits to the Hill equation, the Michaelis-Menten kinetics equation, and the MonodWyman-Changeux (MWC) model for an allosteric enzyme. The Hill equation (Equation 1 , a is defined in Equation 4)answers all-or-none binding of Fru-6-P molecules. The Hill constant, nH, is defined by Equation 2 evaluated at half-maximal velocity, u = Vm/2 (Cantor and Schimmel, 1980; Fersht, 1985). A large Hill constant indicates cooperative interactions, whereas a value of 1.0 indicates no cooperativity. u=

v,- (1 Z n . 1

(Eq. 1)

Hill constants determined by non-linear least square fit to Equation 1 (Bevington, 1969), or linear least square fit to Equation 2 in the region between 0.1 V,,, and 0.9 V,, agreed within the error of the fit. Results calculated by Equation 2 are reported, along with the estimated errors (Bevington, 1969). Michaelis-Menten kinetics, Equation 3, describes the behavior of an enzyme which has no cooperative interactions between subunits.

24601

Equation 4). Since Equation 4reduces to Equation 3 when L becomes small, fits to Equation 4 will always result in a lower x'. In order to determine whether the kinetics exhibit allosteric as opposed to Michaelis-Menten kinetics, we used several criteria: L > 50, and 'x for the fit to Equation 4must be 10 times smaller than thefit to Equation 3. Curve fitting of the kinetic data was performed using the program, InPlot version 4.0 (GraphPad Inc., San Diego, CAI, or a non-linear least-squaresroutine (program CURFIT in Bevington, 1969). Reported errors are the estimates provided by these programs. Circular Dichroism Spectra-CD spectra were recorded on an Aviv 62DS circular dichroism spectrometer. Protein concentrations measured by the procedure of Bradford (1976) were 50 pg/ml for wildtype, 68 pg/ml for Q200R, 22 pg/ml for Q200E, and 28 pg/ml for Q200A in 10 mM potassium phosphate buffer and 10 mM NaF atpH 7.0. Samples were run in a 1-mm cuvette thermostated at 25 "C. At least four scans were averaged from 270 to 184 nm, andthe base-linecorrected spectra were scaled according to their concentration to correspond to the wild-type spectrum. Fig. 7 shows the overlayed spectra. Nondenaturing Agarose Gel Electrophoresis of PFK-The method for electrophoresis of proteins in nondenaturing agarose gelwas similar to thatdescribed by Jiang (1992). Gels at different concentrations of agarose were prepared in 30 mM Tris phosphate at pH 7.0, or 20 mM Tris phosphate at pH 8.0. Each well of the gel was loaded with 5-15 p1 of PFK solution containing 0.05 unit of wild-type or altered RMPFK and 15 pg of tracking dye (xylene cyanol). Electrophoresis was carried out at 4 'C and 55 mA for 12 h in one of the two buffers circulated by a peristatic pump. After electrophoresis, the protein bands were blotted onto nitrocellulose membranes by capillary transfer (Towbin et al.,1979)and stained with 0.5% Amido Black followed by destaining in 45% methanol and 10% acetic acid. The ratio of the mobilities of these PFK proteins relative to that of the tracking dye is defined as R,. The logarithm of these R, values were plotted versus the agarose concentration (%) of the gel (Ferguson, 1964; Hedrick and Smith, 1968; Bryan, 1977; Bras and Garel, 1982). Least-square linear regression analysis was performed using Inplot program. The 0.6% gel at pH 7.0 and 0.8% gel at pH 8.0 were also stained directly for PFK activity using the method described by Brock (1969). In brief, the gel was soaked in 30 ml of solution containing 50 mM Tris-HC1, pH 8.2, 10 mM NanAs04,1 mM EDTA, 4 mM MgC12, 1 mM Fru-6-P, 1 mM ATP, 1 mM NAD, 0.074 mg/ml phenazine methosulfate, 0.4 mg/ml nitro blue tetrazolium chloride, 0.12 mg/ml glyceraldehyde-3-phosphate dehydrogenase, 0.1 mg/ml aldolase, and 7 pg/ml triose phosphate isomerase for 1.5 h at 37 "C followedby destaining in 10% acetic acid. Gel run at pH 7.0was soaked in 20 mM Tris phosphate, pH 8.0, prior to activity staining. RESULTS

Site-directed Mutagenesis at Glutamine 200-Results from DNA sequencing analysis showed correct codon replacement in each mutant, Q200R, Q20OE, or Q200A. All the mutated [Fru-6-P] u = v, (Eq. 3) PFK cDNAs were completely sequenced, and itwas confirmed [Fru-g-P] + K, that no unintentional mutations had been introduced into the The MWCmodel describes the kinetics of an allosteric enzyme remaining sequence of each cDNA. assuming two states of an enzyme, the active R form and inactive T Expression i n E. coli andPurification of Wild-typeand form, with an equilibrium constant L describing the equilibrium Mutants of RMPFK-Successful expression of recombinant between R and T forms in the free enzyme (Monod et al., 1965). Assuming Fru-6-P binds only to the R form, with a dissociation RMPFK was demonstrated by the growth of DF1020 cells constant K,, then therate of reaction as a function of Fru-6-P transformed with pPL2/RMPFK and pCI857 plasmids on a concentration is given in Equation 4 (Cantor and Shimmel, 1980): selective medium using mannitol as carbon source. From 4.5 liters of the culture,0.71 mg of pure RMPFK canbe obtained, a(l + 4 3 u = v, a quantity sufficient for studies on the enzymology and regu(1 + 4 4 + L latory propertiesof both thewild-type and mutated RMPFK. where a = [Fru-6-PJ/Kr.As L becomes larger, the enzyme exhibits A previously reported protocol for the purification of PFK progressively more cooperative behavior. from rabbit muscle using heat treatment and isopropanol Several criteria can be used to determine if the kinetics of an precipitation(Ling et al., 1965) was unsuccessful for the enzyme exhibits allosteric behavior: nH will be larger than 1.0, L will be larger than 1.0, and Equation 4 will fit the data much better than purification of recombinant PFK from bacterial lysate. An alternative procedure was therefore established. The recomEquation 3. The quality of the fits can be compared by calculating binant PFK was purified in three steps: ammonium sulfate 'x = Z(u,, - u,.d'/(N - n) (Eq. 5) fractionation(35%-55%saturation),and chromatography where uelp are the experimental values for the rates, uCsleare the rates through columnsof Cibacron F3GA (Kasten et al., 1983), and calculated by the fit, N is the number of data points, and n is the DEAE-cellulose. The final purification was 1385-fold with a number of parameters in the fit ( n = 2 in Equation 3 and n = 3 in recovery of 49%. All threemutants of RMPFK, Q200R,

Mutagenesis of Rabbit Muscle Phosphofructokinase

24602

Q200E, and Q200A, were purified using the same protocol. Table I shows the resultsof activity assaysat p H 8.0 and 7.0. The properties of recombinant RMPFK, including specific activity and the apparent K , values for Fru-6-P and ATP, are virtually the same as those of the PFK purified from rabbit muscle. As judged by the resultsof SDS-polyacrylamide gel electrophoresis, a high degree of purity has been attained for the wild-type and the three mutants of rabbit muscle PFK (Fig. 2). The electrophoretic patterns also indicate that all the recombinant PFK species have the same subunit size as the commercial standard. Effect of Mutations at Glutamine 200 on the Kinetic Properties of Rabbit Muscle PFK-At pH 8.0, the effect of ATP inhibition is not apparent and RMPFK follows Michaelis. P ) for Q200R and Q200A Menten kinetics. The K , ( F ~ . ~values are essentially the same as that for the wild-type enzyme, while the value for Q200E is 14-fold higher than that of the TABLEI Kinetic parameters for phosphofructokinase purified from rabbit muscle and from E. coli cells transformed by a plasmid bearing the P F K cDNA Assays were carried out at pH 8.0 and 7.0 and the apparent K,,, values were determined as described under "Experimental Procedures." Apparent K,,, activity Enzyme Specific F~l~to~e-6-P" ATP unitslrng

pH 8.0 RMPFK (Sigma) RMPFK (cloned)

mM

220

0.12 f 0.04

0.18 180

f 0.04

pH 7.0 170 RMPFK (Sigma) 0.46 140 RMPFK (cloned) 'ATP concentration was1 mM. Fru-6-P concentration was1 mM.

0.40 f 0.04 f 0.07

0.09 f 0.01 0.10 f 0.02 ND ND

native enzyme. Under these conditions, K,,ATp, values for Q200R and Q200E are 5-fold higher than that of the wildtype enzyme. For Q200A, it is 13-fold higher (Table 11). At pH 7.0, where ATP is both a substrate and an allosteric inhibitor of RMPFK, the effects caused by these mutations are striking. Like the PFK purified from rabbit muscle (Valaitis et al., 1987), the activity of the wild-type PFK starts to decline when the concentration of ATP is greater than 0.6 mM (Fig. 3). However, the activityof Q200R keeps increasing until 7 mM ATP, afterwhich it starts to decline. The activity of Q200E at pH 7.0 is not detectable except in the presence of high (50 mM) Fru-6-P concentration. Under this condition, the optimum ATP concentration for this mutant is 10 mM. For Q200A. the activity is too low a t pH 7.0 to allow proper analysis. Based on these results, the conditions for investigating the effect of Fru-6-P concentration on the activityof native RMPFK and its mutants were adjusted to pH 7.0 in the presenceof low (0.6 mM) or high (7 mM) ATP concentration. With the assumption of a concerted mechanism, the native PFK is in its R state at low ATP concentration; thus TABLE I1 Effect of mutations a t glutamine 200 fQ200) on the kinetic parameters of rabbit m u c k phosphofructokinase Determination of K,,,values and calculation of the Hill numbers were described under "Experimentalprocedures." NA, not applicable, ND, not determined. Apparent K, Hill no. PFK Specific activity F~ctow-6-P ATP p H 8.0 Native Q200R Q200E Q200A

1.0

66,200

0.8

21,500 14,400

Pti

nH

0.18 180 110 44 140

zk 0.04 0.10 f 0.02 0.22 f 0.02 0.47 f 0.03 2.50 f 0.53 0.49 f 0.06 1.30 zk 0.17 0.42 f 0.04

NA NA NA NA 3.2 f 0.1 3.0 f 0.3 2.5 f 0.1

1

1

0.6

4:

> iij

0.4

d 0.2

FIG. 2. SDS-polyacrylamide gel electrophoresis of recombinant and mutated rabbit muscle PFK. The gel was stained using Coomassie Brilliant Blue R-250. Lane 1, RMPFK standard; lanes 2-5, lysates of E. coli cells transformed with plasmid pPL2/ RMPFK, ammonium sulfate fraction (35-55%), pooled active fractions after Cibacron Blue F3GA column, and after DEAE-cellulose column of native recombinant RMPFK, respectively; lanes 6-8, mutants Q200A, Q200R, and Q200E,respectively. All three mutants were purified through DEAE-cellulose columns. LaneM contains size markers.

mM

p H 7.0" ND Native 150 1.70 f 0.1 Q200R 85 0.38 f 0.05 ND Q200Eb 16 16.0 f 1 ND Q200Ab ND ND ND a ATP concentration for these assays was 7 mM. Fru-6-P concentrationwas 50 mM.

91,400

31,000

unitlmg

\

1

0.0

[ATP] mM FIG. 3. Effects of ATP concentration on the activityof the Relative native rabbit muscle PFK and the three mutants. activity is the ratioof u/V where u is the PFK activity at pH7.0 and Fru-6-P concentrationof 1 mM for wild-type RMPFK (0)and Q200R (0)or 50 mM for Q200E).( and Q200A (0)and ATP concentration as indicated. Vis the activity at pH8.0 in the presence of saturating concentrations of Fru-6-P and ATP.

MutagenesisPhosphofructokinase Muscle of Rabbit

24603

and 7 mM ATP by calculating the Hill constant as shown in the rate of reaction versus Fru-6-P concentration exhibits essentially a hyperbolic profile (Fig. 4). Since 0.6 mM ATP is Table 11. In the absence of allosteric effectors, the wild-type, far below the apparent SO.S(ATp) for the mutants (Fig. 3), the QZOOR, and Q200E enzymes all exhibit sigmoidalkinetics, Fru-6-P saturation profiles cannot be analyzed. On the other with Hill constants of 3.2, 3.0, and 2.5, respectively. Fitting hand, under the conditions of 7 mM ATP and pH 7.0, the the data to the MWC model confirms this, with L values of initial conformation of the native PFK is presumably in the 1,900 f 800 for wild-type, 1,100 k 300 for Q200R, and 100 f T state, thus the kinetics are cooperative with respect to Fru- 10 for Q200E (data not shown). 6-P concentration witha Hill coefficient of 3.2 (Table I1 and Wild-type RMPFK and its three mutants are activatedby Fig. 5). Under these conditions, the substrate saturation pro- inorganic phosphate and Fru-2,6-Pz (Fig. 6, A-D). In thecase file of Q200R (Fig. 5) is shifted to lower Fru-6-P concentra- of Q200R, the effect of inorganic phosphateis not noticeable tion, with an apparent K,,,(F~6.p) which is 1/5 of that for the and only slight activation by Fru-2,6-P2is observed (Fig. 6B). native enzyme (Table 11). By contrast, Q200E exhibits 10Mutants Q200E and Q200A require higher concentrations of and much reduced activity theseactivatorsthanthe fold higher apparent Km(Fm.B.P) wild-typeenzyme andmutant (Table I1 and Fig. 5, inset). At pH 7.0, the extremely low Q200R. Inhibition by citrate is clearly observed for the wildactivity of Q200A did not allow a significant comparison of type PFK and QZOOR (Fig. 6, A - 0 ) . However, due to the its activity and properties with thewild-type enzyme (Fig. 5, extremely low activities under the assay conditions, effect the inset). of citrate on mutant Q200A is not detectable. The cooperativity of the enzymes was assessed at pH 7.0 Secondary Structure of Cloned R M P F K and Its MutantsIn order to assess the effect of mutations at Q200 on the secondarystructure of the enzyme, we measured the CD I I spectra of the native RMPFK and the three mutants. Fig. 7 1.0 shows that the CD spectra of these enzyme are essentially identical. These results suggest that mutations at Q200 do not affect the secondary structureof the enzyme. Quaternary Structureof Cloned RMPFK and Its MutantsRabbit muscle PFK is a tetramer of identical subunits (molecular mass 85 kDa). is It possible that the changes in kinetic properties of the various mutantsobserved (Table I1 and Figs. 3,5, and 6)could be caused by alterations in their quaternary structures. To investigate this possibility, we analyzed the quaternary structureof the different mutantsby nondenaturing agarose gel electrophoresis at pH 7.0 as well as pH 8.0 followed by staining for PFK activity and for total protein. ”.” As shown in Figs. 8, A and B, themobilities of the wild-type 0 2 4 6 8 10 RMPFK and its three mutants areclose to that of the PFK [Fru 6-PI mM purified from rabbit muscle. Since the minimum enzymatiFIG. 4. Fructose 6-phosphate saturation curve for the native rabbit muscle PFK. The relative activity ( 0 )is the ratio of u / cally active form of mammalian PFK is a tetramer (Uyeda, V where u is the PFK activity at pH 7.0, 0.6 mM ATP and Fru-6-P 1979; Dunaway, 1983), the positive staining of these protein concentration as indicated. V is the same as defined in Fig. 3. The bands for PFK activitysuggests that the three alteredproteins program and equation used for computer-fit of the dataare described are tetramers. To further support this conclusion, nondenaunder “Experimental Procedures.” turing gel electrophoresis was performed in gels containing different concentrationsof agarose. The logarithm of RF (defined under “Experimental Procedures”) of these proteinswas plotted versus the concentration of the gel (Ferguson plot). 1.0 rThese plots gave parallel lines with the same slope for these 7.0 (Fig. 8C). These results indicate that PFK proteins at pH 0.8 at p H 7.0, the wild-type rabbit muscle PFK and its three 1.o mutants are tetramers of the same size although they may 0.8 differfromeach other in net charge (Hedrick and Smith, 1968). 0.6 A slight change inslope of the Ferguson plot was observed 0.4 for Q200A (0.07 versus 0.08 for other PFK proteins) when the 0.2 electrophoresis was conducted at pH 8.0 (Fig. 80) suggesting 0.0 that, in addition to difference the in netcharge, mutant Q200A 0 20 40 M) 80 100[FN &PI mM I may also differ subtly in conformation, hence size from the other PFK proteins at this hydrogen ion concentration. On 0.0 the other hand,Q200A is almost fully active at pH8.0 (Table 0 2 4 6 8 10 111, thus it must have a tetrameric structure. The tetrameric (Fru 6-PI rnM FIG. 5. Fructose 6-phosphate saturation curves for the structures of the wild-type PFK andQ200A were also verified wild-type rabbit muscle PFK and the three mutants. The by sucrose gradient sedimentation (data not shown).

1

c

1

1

-1

1

relative activity is the same as that defined in Fig. 4. The activities ( u ) of the wild-type enzyme (O),Q200R (O), Q200E (W), or Q200A (0)were measured at pH 7.0, 7 mM ATP and the indicated concentrations of Fru-6-P. The inset shows the concentration of Fru-6-P used for Q200E and Q200A. The program and equations used for computer-fit of the data are described under “Experimental Procedures.”

DISCUSSION

This paper presents the first of case the cloning and expression of a mammalian PFK cDNA in E . coli cells and the first attempt at site-directed mutagenesis on a eukaryotic PFK. Although the level of expression was rather low, the vector-

Mutagenesis Phosphofructokinase Muscle of Rabbit

24604

FIG. 6. Effects of activators and inhibitors on the activities of native and mutated rabbit muscle PFK in response to increasing fructose 6phosphate concentration. Relative activity is the same as that defined in Fig. 5. Activity of the wild-type enzyme ( A ) , Q200R ( B ) , Q200E (C), or Q200A (D)was measured at pH 7.0, 7 mM ATP in the absence (0)or presence of Fru2,6-P2 (O), inorganic phosphate (m) or citrate (0).The concentration of FN2,6-P2 is 1 pM (wild-type RMPFK or Q200R) or 25 ~ L M(Q200E and Q200A); inorganic phosphate, 1 mM (&200R), 10 mM (wild-type RMPFK and Q200E), or 20 mM (Q200A); and citrate, 25 p~ (Q200E and Q200A), 75 p~ (wild-type RMPFK) or 125 p~ (Q200R). The program and equations used for computerfit of the data are described under "Experimental Procedures."

0

2

4

6

8

0

2

4

6

8

10

[Fru 6-PI mM

(Schirmer and Evans,1990) showed a 7 rotation between its two pairs of dimers (A:D uersus B:C) during the R to T transition. The binding site for the cooperative substrate Fru6-P is altered by changes involving the 6-F loop. In the R state, the phosphate group of Fru-6-P interacts with Arg16' and Arg243across the subunit interface (Fig. 9). Inthe T state, ArgI6' swings awayfrom the active site. G1uI6' fills the vacated space by interacting with through the relay of a water molecule across the subunit interface, andforms a salt bridge with A r P . This conformational change physically closes the substrate binding cleft. The positive charges of Arg16', Arg'43, and Arg5' are therefore no longer available for binding the negative charges of Fru-6-P. Glutamine 200 of rabbit muscle PFK, which corresponds to G W 1 of BsPFK (Poorman et al., 1984; Hellinga and Evans, 1985), was thus chosen for mutagenesis in the current work. Indeed, our results demonstrate the critical role of GlnZo0in the activity and allosteric properties of rabbit muscle PFK. Mutations at Gln'O" affect its 180 200 220 240 affinities for both substrates, Fru-6-P andATP. wavelength (nm) A positively charged replacement at position 200 (Q200R) FIG. 7. Circular dichroism spectra of wild-type and QZOOR, results in an enzyme with significantly attenuated ATP inQ200E, and Q200A mutants of rabbit muscle PFK. Ellipticities hibition. The mutation might provide an environment for (milla degree) of CD signal are plotted versus wavelength for wild- Fru-6-P binding to the "T state," making it active. AlternaQ200R (0),Q200E (A), and Q200A (0)mutants type RMPFK (-), At least four scans were averaged from 270 to 184 nm, and the base- tively, the arginine replacement could make the T state strucline-corrected spectra were scaled according to their concentration to ture less stable because its positively charged side chain cannotinteract with theintra- andintersubunit arginine correspond to the wild-type spectrum. residues as shown in the structure of the T state BsPFK host system produced sufficient recombinant RMPFK for the (Shirmer and Evans,1990). The fact that at pH7.0 and 7 mM purpose of structure-function studies. The expression of re- ATP the kinetic profile of mutant Q200R is shifted to a lower combinant RMPFK in E. coli cells might be limited by the 18 Fru-6-P concentration region (Fig. 5) argues for both interAGG/AGA arginine codons in the RMPFK gene (Lee et al., pretations. The tolerance of Q200R for high ATP concentra1987). These arginine codons are rarely used by E. coli cells tion could be the consequence of a reduced affinity for ATP (Brinkman et al., 1989; Wada, et al., 1991). In addition, a at the inhibitor binding site or, more likely, the mutated second possible cause for low level expression of RMPFK proteins can no longer effectively transmit the effect of the could reside at the stepof thermoinduction of the pLpromoter conformational changes taking place at the inhibitor site to at 42 "Cfor 3 h. This heat-shock step induces the pLpromoter the catalytic site. A negatively charged residue (Q200E) at this position drastas well asother heat-shock genes, some of whichencode proteases (Buell et al., 1985). Attempts are being made to ically reduces the affinity of the enzyme for Fru-6-P, most better understandthe reasons for low expression of the cloned likely by charge repulsion. This is reflected by the 10-fold higher apparent Km(Fm.6.P) at pH 8.0 (Table 11).At pH 7.0 and PFK cDNA. The major goalof this work has been to study the structural 7 mM ATP and at Fru-6-P concentration 10-fold higher than basis for the allosteric control of rabbit muscle PFK. The that for wild-type enzyme, the activity of this mutant showed crystallographic structures of the R and T states of BsPFK a sigmoidal response. The cooperativity is lower, as shown by I

"

"

"

"

1

Mutagenesis of Rabbit Muscle Phosphofructokinase

B

"

"""

1.8

C

1.3 0 1.9

,

I 0

1 2 3 Ag8ro¶e Concentration ( X )

4

1

I 1 2 3 Agarose Concentration ( X )

4

FIG.8. Analysis of the quaternary structure of rabbit muscle PFK and its mutants. A and B, electrophoresis of wild-type rabbit muscle PFK ( W T ) ,Q200R, Q200E, and Q200A in a nondenaturing agarose gel at pH 7.0 ( A ) and pH 8.0 ( B ) .The two standards were PFK purified from rabbit muscle (RMPFK)and B. stearotharrno p h i h PFK (RsPFK). The conditions of electrophoresis were described under "Experimental Procedures." The gel was stained for PFK activity by the method described in the text. C and D,Ferguson plot obtained by electrophoresis of wild-type RMPFK (O), Q200R (O), Q200E (W), and Q200A (0)in nondenaturing gels of 0.5, 1.5, 2.5, and 3.5% agarose in 30 mM Tris phosphate, pH 7.0 ( C ) and 20 mM The conditions for electrophoresis were Tris phosphate, pH 8.0 (D). described in the text. The protein bands were transferred to nitrocellulose membranes and stained in 0.5% Amido Black. The logarithm of the R,,as defined in the text, of these PFK proteins were plotted

24605

the smaller Hill constant (Table 11). The Q200E mutant is activated by Fru-2,6-P2,butrequires 25 p~ concentration uersus 1 P M forwild-typeand Q200R mutant. A t pH 7.0, Q200E is inherently lessactive, and exhibits much weaker allosteric behavior (Fig. 5, inset). Preliminary results have shown that a Ql6lE mutant of EcPFK (analogous toQ200E of RMPFK) requires a much higher concentration of Fru-6P to make theT to R transition.' Finally, mutant Q200A exhibited a relatively high specific activity (78% of the wild-type enzyme) at pH 8.0 with a 13fold elevation of the apparent K,,, for ATP (Table 11). AIthough this mutant was essentially inactive a t pH 7.0 over a wide range of Fru-6-P and ATP concentrations(Figs. 3 and 5 ) . it is significantly activatedby a high concentration of Fru2,6-P2. Both Q200E and Q200A are similarin the difficulty of Fru-6-P to activate the enzymes, while being fully activated by high concentrations of activators. This mightsuggest that the T state is stabilized in these mutants, perhaps due to a loss of favorableinteractions between theactivesiteand residue Q200 in the R state. The wild-type RMPFK and the mutated enzymes are activated by inorganic phosphate and Fru-2,6-P2. Citrate may inhibit all forms of RMPFK, although the effect is not discernible for Q200A because of its low activity a t pH 7.0. It is interesting to note that the concentrations of the activators and inhibitor for Q200R, wild-type PFK, Q200E, and Q200A are in opposite order (see legend to Fig. 6). Enzymes which are active a t lower Fru-6-P concentration in the absence of these effectors require lower concentrations of activators or higher concentrations of inhibitor toreveal their effects. The changes in kinetic properties of the various mutants could have been caused by alterations in structure at different levels. We investigated this possibility by analyzing the secondary structural contentof the wild-type PFK and its three mutants by CD, and their quaternary structures by nondenaturing gel electrophoresis in gels of different agarose concentration followed by Ferguson plots (Ferguson, 1964; Hedrick andSmith, 1968). The results of theseexperiments showed that there is no significant change in secondary structure, and all three mutants are tetramers (Figs. 7 and 8). Therefore, the changes in kinetic propertiesobserved forthese mutants reveal the role of Q200 in the allosteric behavior of RMPFK rather than any changes in the secondary or quaternary structure. The techniqueof site-directed mutagenesis hasbeen widely applied to studies on the enzymology of an increasingly larger number of proteins. I t is important, when comparing a mutated enzyme to its wild-type counterpart, to know if the altered amino acids would affect the secondary, tertiary, or quaternary structure of the protein. For analysis of the quaternary structures of the altered proteins, methods such as electrophoresis in nondenaturing gels or sedimentation in sucrose gradients would be suitable. However, we have experienceda technical difficultyin thatrabbit muscle PFK cannot penetrate nondenaturing polyacrylamide gel when gel we concentration is higher than 3.7%. On the other hand, found that electrophoresis in nondenaturing agarose gels of different concentrationsof agarose is much simpler toperform for rabbit muscle PFK than electrophoresis in polyacrylamide gels. Although only residue 6200 has been investigated by sitedirected mutagenesis, this work hasdemonstrateddistinct ~~~

.~

-

. .

" .

M. Byrnes and S. H. Chang, unpublished..observation. .~ ~~

~~~~~

versus the gel concentration. The linea are the linear leant quare fits to the data, and all of the slopes agreed within experimental error. Inplot program was used for the computer-tit of the data.

Mutagenesis of

24606

A

\Asp1

Rabbit Muscle Phosphofructokinase 2

subunit 4

NII!

+

x ,NHd Argl62 ,)

N1l'

\'4Glu7611

\

tlll, \

!111

'1 63

rg243

subunit D

FIG. 9. Schematic view showing the comparison of the T state and R state at the substrates site of B. stearotherrnophilus PFK. Wavy line represents subunit interface. A, T state, showing G1u"jl in the Fu-6-P site; B , R state (with active site ligands Fru-6-P and ADP), showing the phosphate of Fru-6-P bound by Arg16* and Are3. Reprinted with permission from Nature (Schirmer, T., and Evans, P. (1990). Nature 343,140-145) Copyright (1990) Macmillan Magazines Limited. Gehnrich, S. C., Gekakis, N., and SUI, H. S. (1988) J. Biol. Chem. 255, 4240differences inkineticpropertiesamongthenativerabbit A9AK muscle PFK and the three mutants. We speculate that muGree;:H. (1975) Virology 66, 589-604 J. L., and Smith,A. J. (1968) Arch. Biochem. Biophys. 126, 155-164 6-F loop (Schirmer Hedrick, tations at other residues along the putative Heinisch, J., Ritzel, G., von Bostel. R. C., Aueilera. A. Rodicio.. R... and and Evans, 1990) of this enzyme may also give useful results Zimmermann, F. K. (1989) Gene (Amst.)78,305-321 Hellinga, H. W., and Evans, P. R. (1985) Eur. J. Biochem. 1 4 9 , 363-373 which could provide further insights into the structural basis Jiang, J. C., Lee, W. R., Chang, S. H., and Silverman, H. (1992) Enu. & Mol. of the allosteric behavior of this enzyme, whether these inMut. 20,260-270 T. P.,Naqui,D.,Kruep, D., and Dunaway, G. A. (1983) Biochem. volve the cooperative binding of Fru-6-P or the many inter- Kasten, Biophys. Res. Commun. 111,248-254 actions between the Fru-6-P site andseveral regulatory sites. Kemp, R. G. (1975) Methods Enzymol. 4 2 , 71-77

Acknowledgments-We are indebted to Drs. George A. Dunaway, Philip R. Evans, Jean-Renand Garel, and Robert G . Kemp for their valuable suggestions during the performance of this work and their helpful discussions and comments during the preparation of this manuscript. REFERENCES Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A,, and Struhl, K. (1989) Current Protocols in Moleculnr Biology, pp. 1.1.1, Wiley Interscience, New York Bevington, P. R. (1969) Data Reduction and Error Analysis for the Physical Science, pp. 204-246, McGraw-Hill, Inc., New York, NY Bradford, M. M. (1976) Anal. Biochem. 7 2 , 248-254 Bras, G. L., and Garel, J-R.(1982) Biochemistry 21,6656-6670 Brinkman, U., Mattes, R. F., and Buckel P. (1989) Gene (Amst.) 8 5 , 109-114 Brock, D. J. H. (1969) Biochem. J. 1 1 3 , 235-242 Bryan, J. K. (1977) A d . Biochem. 78,513-519 Buell, G., Schulz, M-F., Selzer, G., Chollet, A., Movva, N. R., Semon, D., Escanez, S., and Kawashima, E. (1985) Nucleic Acids Res. 13,1923-1938 Cantor, C. H., and Shimmel, P. R. (1980) Biophysical Chemistry, pp. 939-978, W. H. Freeman & CO., San Francisco, CA Chen, E. Y., and Seeburg, P. H. (1985) DNA 4,165-170 Daldal, F. (1983) J. Mol. Biol. 168, 285-305 Dotto, G. P., and Zinder, N. D. (1983) Virology 114,463-473 Dunaway, G. A. (1983) Mol. Cell. Biochem. 52, 75-91 Dunaway. G. A,, and Kasten, T. P. (1987) Biochem. J. 242,667-671 Dunaway, G. A., and Weber, G. (1974) Arch. Biochem. Biophys. 162,620-628 Dunaway, G. A,, Kasten, T. P., Sebo, T., and Trapp,R. (1988) BiochemJ. 251, 677-683 Foe, L. G., and Kemp, R. G. (1982) J. Biol. Chem. 257,6368-6372 Foe, L. G., and Kemp, R. G. (1985) J. Biol. Chem. 2 6 0 , 726-730 Ferguson, K. A. (1964) Metabolism 13,985-1002 Fersht, A. (1985) Enzyme Structure and Mechanism, 2nd Ed., pp. 263-292, W. H. Freeman, San Francisco,CA

Kotlarz, D., and Buc, H. (1981) Eur. J. Biochem. 117,569-574 Latshaw, S. P., Bazaes, S., Raudoph, A,, Poorman, R. A,, Heinrikson, R. L., and Kemp, R. G. (1987) J. Biol. Chem. 262,10672-10677 Lee, C. P., Kao M. C., French B. A,, Putney, S. D., and Chang, S. H. (1987) J . Biol. Cheml 262,4195-4199 Levanon: D., Danciger, E., Dafni, N., Bernstein, Y., Elson,A,, Moens, W., Brandies, M., and Groner, Y.(1989) DNA 8 , 733-743 Li, J., Chen, Z., Lu, L., Byrnes, M., and Chang, S. H. (1990) Biochem. Biophys. Res. Commun. 166,637-641 Ling, K. H., Marcus, F., and Lardy, H. A. (1965) J . Biol. Chem. 2 4 0 , 18931899 Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY Monod, J., Wyman, J., and Changeaux, J. P. (1965) J. Mol. Biol. 1 2 , 88-118 Passonneau, J. V., and Lowry, 0.H. (1963) Biochem. Biophys. Res. Commun. 13,372-379 Pilkis. S. J.. ELMaehrabi. M. R.. Pilkis., J.., and Claus., T.(1981) . . J. Biol. Chem. 256,3619-3622 ' Poorman, R. A,, Randolph, A,, Kemp, R. G., and Heinrikson, R. L. (1984) Nature 309,467-469 Reinhart, G. D., and Lardy, H. A. (1980) Biochemistry 1 9 , 1477-1484 Remaut, E., Stanssens, P., and Fiers, W. (1981) Gene (Amst.) 15, 81-93 Rypniewski, W. R., and Evans, P. R. (1989) J. Mol. Biol. 207,805-821 Schirakihara, Y., and Evans, P. R. (1988) J. Mol. Biol. 204,973-994 Schirmer, T., and Evans, P. R. (1990) Nature 343,141-145 Towbin, H., Staehelin, T., and Gordon,J. (1979) Proc. Natl. Acad. Sci. U. S. A. 76,4350-4354 Uyeda, K. (1979) Advances in Enzymology and Related Areas of Molecular Biology (Meister, A,, ed) Vol. 48, pp. 193-244, John Wiley and Sons, New York Uyeda, K., Furuya, E., and Luby, L. J. (1981) J. Biol. Chem. 256,8394-8399 Valaitis, A. P., Foe, L. G., and Kemp, R. G. (1987) J. Biol. Chem. 2 6 2 , 50445048 Van Schaftingen, E., Hue, L., and Hers, H. G. (1980) Biochem. J. 1 9 2 , 887895 Wada, K., Wada, Y., Doi, H., Ishibashi, F.,Gojobori, T., and Ikemura,T.(1991) Nucleic Acids Res. 19, 1981-1986 ~