Biochem. J. (1978) 173, 773-786Printed in Great Britain

Rapid Purification and Properties of Potassium-Activated AldehydeDehydrogenase from Saccharomyces cerevisiae

By KEITH A. BOSTIAN* and GRAHAM F. BETTSDepartment ofPlant Biology and Microbiology, Queen Mary College,

London El 4NS, U.K.

(Received 16 December 1977)

A method for the purification of yeast K+-activated aldehyde dehydrogenase is presentedwhich can be completed in substantially less time than other published procedures. Theenzyme has a different N-terminal amino acid from preparations previously reported,and other small differences in amino acid content. These differences may be the resultof differential proteolytic digestion rather than a different protein in vivo. A purificationstep involves the biospecific adsorption on affinity columns containing immobilizednucleotides in the absence of the substrate aldehyde. Direct binding studies with thecoenzyme in the absence of aldehyde reveal 4 NAD sites per tetrameric molecule, eachwith a dissociation constant of 120pM. These results conflict with properties of prepara-tions previously reported and may conflict with kinetic models that have aldehyde as theleading substrate. Binding to Blue Dextran affinity columns suggests the presence of adinucleotide fold in common with other dehydrogenases and kinases.

Ethanol metabolism involves two oxidative steps,both catalysed by nicotinamide nucleotide-dependentdehydrogenases:

RCH2OH + NAD(P)+ xRCHO + NAD(P)H + H+

RCHO + NAD(P)+ + H20 eRCO2H + NAD(P)H + H+

In the past 30 years extensive work has been done onalcohol dehydrogenase, but comparatively little hasbeen done on aldehyde dehydrogenase. Part of theexplanation for this may be a number of difficultiesassociated with its purification. It was not until 1967that a homogeneous preparation was reported withthe isolation of the K+-activated enzyme from yeast(Steinman & Jakoby, 1967). Difficulties in thepurification of this enzyme are its instability in theabsence of high concentrations of K+ (Sorger &Evans, 1966), its increasing instability with purifica-tion in the absence of high concentrations of poly-hydric alcohol (Bradbury & Jakoby, 1972), and itsmodification by endogenous proteinases (Clark &Jakoby, 1970a,b). For these reasons we havedeveloped a rapid method of purification for thisenzyme by using an affinity-chromatographic step,which results in higher yields of enzyme activity.Differences are revealed between this preparationand those previously published (Sorger & Evans,

Abbreviations used: SDS, sodium dodecyl sulphate;Dns-, 5-dimethylaminonaphthalene-1-sulphonyl-.

* Present address: Rosenstiel Basic Medical ResearchCentre, Brandeis University, Waltham, MA 02154,U.S.A.

Vol. 173

1966; Steinman & Jakoby, 1967; Clark & Jakoby,1970a,b; Bradbury & Jakoby, 1972).

Materials and Methods

Baker's yeast (N.G. & S.F., British FermentationProducts, London E.C.2, U.K.) was purchased froma local baker, stored at 4°C and used within 24h.The yeast is produced commercially from an anaero-bic axenic culture by bulk fermentation (for 12h)with extensive aeration in a molasses wort supple-mented with inorganic salts. In addition, yeast waspurchased from a U.S.A. distributor of Anheuser-Busch, St. Louis, MO, U.S.A., and handled as above.Sepharose 6B CL 2,3-dibromopropanol cross-linked CH-Sepharose 4B (with 6 carbon spacer andfree carboxyl group), 5'-AMP-Sepharose 4B, DEAE-Sephadex A-50, Sephadex G-25 and Blue Dextranwere purchased from Pharmacia (G.B.), LondonW.5, U.K. Whatman DE-23 ion-exchange DEAE-cellulose was purchased from Whatman Bio-chemicals, Maidstone, Kent, U.K. NAD+ (gradeIII), NAD+ (grade AA), 5'-AMP, 2-mercapto-ethanol, a-thioglycerol, phenylmethanesulphonylfluoride, phenazine methosulphate, Nitro BlueTetrazolium, glutamate dehydrogenase, thyro-globulin, catalase (bovine), ovalbumin, cytochrome c

(type 2), bovine serum albumin (A4378) andcarbonic anhydrase were purchased from SigmaChemical Co., Kingston upon Thames, Surrey, U.K.fl-Galactosidase, aldolase and muscle lactate de-hydrogenase were supplied by Boehringer Corp.(London) Ltd., London W.5, U.K. Standard proteins

773

K. A. 1OSTIAN AND G. F. BETTS

for SDS/polyacrylamide-gel electrophoresis wereprepared from a Boehringer protein calibration kit,Combithek-size 1. Reagents for elfctrophoresis werepurchased from Kodak, Kirkby, Liverpool, U.K.Reference Dns-amino acids, pglyamide thin-layersheets and AnalaR grade ethanediol were suppliedby BDH Chemicals, Poole, Dorset, U.K. 3-Methyl-2-benzothiazolone hydrazone was purchased fromHopkin and Williams, Chadwell Heath, Essex, U.K.Nicotinamide-[carbonyl-"Cjadenine dinucleotide(59.2mCi/mmol) and nicotinamide-[U-14C]adeninedinucleotide (271 mCi/mmol) were supplied by TheRadiochemical Centre, Amersham, Bucks., U.K.Insta-Gel was from Packard Nuclear ChemicalSupplies, Caversham, Berks., U.K.. Acetaldehydewas redistilled and stored at 18°C as a 2M solution.All other reagents were of AnalaR grade where com-mercially available.

Enzyme assay

Enzyme activity was measured at 25'C by follow-ing the rate of reduction of NAD+ by observing theincrease in A340 in a Pye-Unicam SP. 1800 spectro-photometer. A standard reaction mixture containedin a final volume of 2.5 ml: 2mM-acetaldehyde,0.5mM-NAD+, 0.1 M-Tris/HCl buffer, pH 8.0, 10mM-2-mercaptoethanol and 0.1 M-KCl. Reaction wasbegun by the addition of 0.1 ml or less of enzyme toan otherwise complete reaction mixture. Under theseconditions a unit of enzyme activity is the amount ofenzyme producing 1 pmol of NADH/min.

Protein determination

Samples containing 20-100AUg of protein wereprecipitated with 4 vol. of 25% (w/v) trichloroaceticacid. The precipitate was resuppended in 1 M-NaOHand protein measured by the method of Lowry et al.(1951), with bovine serum albumin as a standard.

Aldehyde determinations

Aldehydes were measured by the colorimetricmethod ofSawicki et al. (1961). To 0.5 ml ofunknownsample or standard aldehyde solution (5-100,ug/100ml) was added 0.5ml of 0.4% (w/v) 3-methyl-2-benzothiazolone hydrazone in a ground-glassstoppered test tube. The solution was placed in awater bath at 98.5°C for 3min, The mixture wascooled to 25°C in an ice-water bath, then 2.5ml of0.2% (w/v) FeCl3 was added. After 5 min the volumewas adjusted to 10ml by the addition of 6.5ml ofacetone, and the 4635 or A664 measured.Aldehyde was also determined by its enzymic

reactivity with aldehyde dehydrogenase in thepresence of 0.1 M-KCl and 1 mM-NAD+. Fraction Venzyme (see purification schedule below) was

equilibrated by passage through a Sephadex G-25column into 0.1 M-K2HPQ4 buffer, pH 8.0, and 1 mm-EDTA, both showing low 3-methyl-2-benzothia-zolone hydrazone reactivity. The equilibrated enzyme(1 unit) was added to a solution containing 1 mm-NAD+, 0.1 M-K2HPO4, 1 mM-EDTA, with or withoutthe sample, resulting in a final volume of 2.5 ml.A340 was monitored with time. Lactate dehydro-genase and pyruvic acid were then added to oxidizeany NADH formed, to confirm the amount ofrecorded absorbance change relating to NADHIproductiop.

Polyacrylamide-disc-gel electrophoresis

Electrophoresis was performed by the method ofDavis (1964) at 4°C, except that running buffercontained 1 mM-mercaptoethanol, and the stackinggel was omitted. Enzyme samples containing 25%(y/v) glycerol were layered directly on to the poly-tcrylamide gels. Enzyme samples not containingglycerol were first made 20% (w/v) sucrose by theaddition of solid sucrose.

Gels were stained for protein by the method ofChrambach et al. (1967) and recorded by scanningwith a Joyce-Loebl gel scanner. Gels were stainedfor activity by placing them into a standard reactionmixture supplemented to 25mg of phenazine metho-sulphate/ml and 0.5mg of Nitro Blue Tetrazolium/ml, from stocks freshly made up before use. Aftermixing, the gels were left to incubate at 23°C(2-Smin), washed with water and finally stored in10% (w/v) trichloroacetic acid. Alternatively,enzyme activity in gels placed in standard rqactionmixture was followed by monitoring the increase inA340 with a scanning accessory to a Gilford 240spectrophotometer.

SDS/polyacrylamide-gel electrophoresis

Enzyme and known molecular-weight markerproteins were prepared for SDS/polyacrylamide-gelelectrophoresis by dissociation in a boiling-waterbath for 3mi in 0,01 M-Tris/hydrogen phosphatebuffer, pH 7.0, 1 % (w/v) SDS and 5% (v/v) mercapto-ethanol, at a protein concentration of 1.6mg/ml;25-50,ug of protein was used for loading 7°% (w/v)polyacrylamide gels, and electrophoresis performedessentially by the method of Weber & Osborn (1969),except that running buffer consisted of 75mm-Tris/hydrogen phosphate, pH7.0, 0.1% (w/v) SDS and0.1 % (v/v) mercaptoethanol. Subunit molecularweight was estimated by calibrating the gels againstbovine serum albumin dimer and monomer (mol.wt.134000 and 67000 respectively), phosphorylase a(mol.wt. 92500), catalase (mol.wt, 60000), oval-bumin (mol.wt. 45000) and carbonic anhydrase(mol.wt. 30000) standards. Mobilities were repro-

1978

774

PURIFICATION AND PROPERTIES OF YEAST ALDEHYDE DEHYDROGENASE

ducible within 1.5% and were inversely proportionalto logarithm (molecular weight) over the molecular-weight range 134000-30000.

Sepharose 6B CL chromatography

The molecular weight of the undissociated proteinwas estimated on columns (1.75cm x 53cm) ofSepharose 6B CL equilibrated and run at 4°C in0.1 M-KH2PO4 (pH 7.4)/1 mm-mercaptoethanol/l mm-MgSO4/0.125M-NaCl buffer. Flow rate was 0.254ml/min and fractions (1 ml) were collected. Sampleswere added in a volume of 1 ml. Elution profiles ofproteins were monitored with an Isco dual-beamabsorptiometer at A280. In addition, aldehyde de-hydrogenase activity in the column effluent wasmonitored by the standard assay procedure. Columnswere calibrated with Blue Dextran (mol.wt. 2000000),glutamate dehydrogenase (mol.wt. 1020000), thyro-globulin (mol.wt. 670000), ,-galactosidase (mol.wt.515000), catalase (mol.wt. 247000), aldolase (mol.wt.145000), bovine serum albumin (mol.wt. 67000),ovalbumin (mol.wt. 45000), carbonic anhydrase(mol.wt. 30000) and cytochrome c (mol.wt. 12300).Elution volumes were reproducible to ±2.5% andwere inversely proportional to logarithm (molecularweight) over the molecular-weight range 1041_06.

Amino acid analysis

Amino acid analysis was performed by g.l.c. oftrifluoroacetylated amino acid methyl esters. Samplesof freeze-dried fraction VI enzyme (see purificationschedule below) were hydrolysed in triplicate for 24or 48 h in 6M-HCl at 105°C. Samples of hydrolysateswere converted into derivatives by the method ofDarbre & Islam (1968) and quantitative identificationwas achieved on mixed silicone columns in a Pyeseries 104 chromatograph as described by Davy &Morris (1976).

End-group analysis

N-Terminal amino acid analysis was performed bythe dansyl chloride procedure recommended forproteins by Gray (1967), with identification of theDns-amino acids by two-dimensional chromato-graphy on thin-layer polyamide sheets as describedby Hartley (1970). Concentrated fraction VI enzyme(see purification schedule below) was prepared fordansylation by dialysis for 72h against severalchanges of distilled deionized water. After freeze-drying, 1.5mg samples containing approx. 25nmolof N-terminal amino acid were taken for analysis.

Equilibrium dialysis

Equilibrium dialysis binding experiments wereperformed in a semi-micro dialysis apparatus con-

Vol. 173

sisting of two Perspex blocks each with five matchingchambers (15mm diam. x 2mm deep) equipped withsmall screw-capped inlet ports. The blocks were heldtogether by ten wing-nut screws with a Visking 24/32dialysis membrane suspended between the adjoiningchambers. The menmbrane was prepared for use byboiling twice for 10min in 1 mM-EDTA, thoroughlywashing and then boiling for 10min in distilleddeionized water. The membrane was then rinsed andused. In chambers on one side of the dialysis mem-brane were placed 0.3 ml samples of a known con-centration of labelled ligand in 'dialysis buffer',consisting of 0.1 M-Tris/HCl, pH 8.0, 20% (v/v)ethanediol, 1 M-KCl, 1 mM-mercaptoethanol, 2mM-pyruvic acid and supplemented to 0.5 g/ml withlactate dehydrogenase also equilibrated against thedialysis buffer. The additions were made with 1 mldisposable syringes. Equal volumes of enzyme,equilibrated by passing through a column (1.25cm x6.8 cm) of Sephadex G-25 into dialysis buffer andsupplemented to 0.5 mg/ml with lactate dehydro-genase in dialysis buffer, were placed in the chamberson the other side of the membrane, and the wholeapparatus was coupled end-on to a Watson-Marlowpump with rotation at approx. 60rev./min. Fillingeach chamber with 0.3ml of solution left enough ofan air-space to ensure adequate mixing across themembrane surface. All experiments were; performedat 4°C. Lactate dehydrogenase (at 0.1 % aldehydedehydrogenase concentration) and pyruvic acid wereadded to buffer on both sides of the membrane toregenerate any NAD+ reduced by the presence ofendogenous aldehydes.

In the absence of enzyme but under otherwisesimilar conditions the time required for NAD+ toreach equilibrium was determined as 48h. In othertests no detectable enzymne activity was observed inthe ligand chamber, cross contamination was not aproblem, and recovery was consistent with the totalvolume of solution recovered, i.e. no adsorption ofligand to the membrane was observed,

Radioactive [carbonyl-'4C]NAD+ solutions wereprepared by dissolving the freeze-dried radioactivematerial in 1 ml of 5 mM-non-radioactive NAD+(Sigma grade III). This solution was diluted 1: 10 ormore and supplemented with non-radioactive NAD+to make the appropriate starting solutions. Afterdialysis three weighed samples (approx. 85,u1) wereremoved from each chamber and measured forradioactivity with lOml of Insta-Gel supplementedwith 4m1 of water, in a Packard Tri-Carb liquid-scintillation counter. Overall standard deviation indetermining radioactivity in either chamber atequilibrium was 0.15%.

Rate of dialysisRate of dialysis binding studies were performed by

the method of Colowick & Womack (1969) using a

775

K. A. BOSTIAN AND G. F. BETTS

dialysis flow cell previously described (Womack &Colowick, 1973), but containing a Perspex insert inthe lower chamber to decrease its capacity to 2ml.Visking 24/32 dialysis membrane was used asprepared above. Connections were made withminimal lengths of 1.5mm inner diameter Teflontubing constituting a dead-space volume betweencell and fraction collector of 0.15ml. The lowerchamber was filled with the same buffer used forequilibrium dialysis and 1 ml of enzyme equilibratedwith the same buffer by passage through a column(1.25cm x 6.8cm) of Sephadex G-25 was added tothe upper chamber. A small volume of very highspecific radioactivity [U-'4C]NAD+ was added tothe enzyme solution and 3.25 ml fractions werecollected at a flow rate of 9ml/min. Under theseconditions, rate of dialysis reached a steady statewithin 10 fractions, so that after every 10 fractions asample of non-radioactive ligand was added. Thetotal initial volume of radioactive ligand added wasless than one-tenth the initial enzyme volume, andsubsequent additions of non-radioactive ligand didnot amount to more than 1% of the total initialvolume. All experiments were performed at 4°C.That no artifacts were present in the system wasobserved from test experiments minus enzyme.

Radioactive NAD+ effluent solution in fractionswere prepared for counting in the same manner as inequilibrium dialysis, except that weighed samples ofapprox. 3 ml were used, and 2ml of water was addedwith the 10ml of Insta-Gel. Overall standarddeviation in determining radioactivity in fractionsof eluate was 0.3 %.

Preparation of immobilized ligands

e-Aminohexanoyl-NAD+-Sepharose 4B was pre-pared by the method of Mosbach et al. (1972), byusing 45g of CH-Sepharose 4B. Sigma-grade-AANAD+ was used. This gel was used to pack a column(2.5cm x 25 cm) at 4°C, providing a gel volume of125ml. The column was equilibrated with 5-6vol.of 45mM-K2HPO4 buffer, pH 8.0, containing 22%(v/v) glycerol, 1 mM-mercaptoethanol, 0.025% (w/v)phenylmethanesulphonyl fluoride (the 'affinity-chromatography buffer').

Blue Dextran-Sepharose 4B was prepared by theCNBr method of Ryan & Vestling (1974). Columnswere packed with gel in 10mM-Tris/HCl buffer,pH 8.0, and equilibrated before use in an appropriatebuffer. Column dimensions were 1.2cm x 7cm.

Enzyme concentration

Enzyme was concentrated by a combination ofhollow-fibre ultrafiltration (Bio-Fiber 80 beaker;Bio-Rad Laboratories, Bromley, Kent, U.K.) andcentrifuged at 3000g through an Amicon PM 10

membrane supported on a porous plastic disc in a50ml centrifuge tube.

Preparation of aldehyde dehydrogenase type A fromAnheuser-Busch yeast

Attempts at preparing aldehyde dehydrogenase Afrom Anheuser-Busch yeast were made by followingthe outline of Clark & Jacoby (1970a) as closely aspossible. Unfortunately, a Manton-Gaulin homo-genizer was not available for use. Alternatively, cellbreakage was accomplished by the glass-bead methoddescribed in the rapid affinity-chromatographicpurification (see below). Breakage was performed atroom temperature with a yeast/glass-bead suspensionin the buffer used by Clark & Jakoby (1970a), at astarting temperature of 4°C. Time of disruption wasapproximately the same in both procedures. Yeastin 0.75 kg lots was used, and subsequent steps werescaled appropriately. The crude extract was heat-and acid-precipitated (Clark & Jakoby, 1970a) andthe (NH4)2SO4 fraction, after dialysis, was loaded ona column (2.5cm x 30cm) of DEAE-Sephadex A-50and chromatography conditions that allowedduplication of the chromatography times for thelarger columns of Clark & Jakoby (1970a) were used.Chromatography buffer was prepared with Kodak'spectro'-grade glycerol. A major peak of activitywas eluted at the same ionic strength as describedby Clark & Jakoby (1970a); these fractions werepooled, supplemented to 0.01 % phenylmethane-sulphonyl fluoride and stored at-80°C until furtheruse. The hydroxyapatite chromatography describedby Clark & Jakoby (1970a) was not reproducible inour hands.

Rapid affinity-chromatographic enzyme purification

At all stages of the purification phenylmethane-sulphonyl fluoride was added to reagents immediatelybefore use. The temperature was 4°C unless other-wise stated.

Extraction procedure

Method I. Yeast acetone-dried powder (200g)prepared by the method of Stoppani & Milstein(1957) was resuspended in 1200ml of 0.1 M-K2HPO4/1 mM-mercaptoethanol/0.024% (w/v) phenylmethane-sulphonyl fluoride. The pH was adjusted to 8.6 with1 M-KOH and the mixture stirred for 4h. The pH wasadjusted to 5.5 by the dropwise addition of I M-citricacid and this mixture centrifuged for 30min at23000g. The supernatant is fraction I.Method IL Packed yeast cells (0.7 kg) were

suspended in 1.5 vol. of 0.1 M-K2HPO4/1 mM-mercaptoethanol /0.024% (w/v) phenylmethane-sulphonyl fluoride. Acid-washed and oven-baked

1978

776

PURIFICATION AND PROPERTIES OF YEAST ALDEHYDE DEHYDROGENASE

0.45,um glass beads at -18°C were added at a ratioof 3 g/ml of cell suspension. Cells were then dis-rupted by shaking on an orbital shaker at 350rev./min. Breakage was continued until it was estimatedby microscopy to be 90-95% complete, whichrequired about 15min. The resulting slurry waspassed rapidly through a glass-fibre plug and thebeads were washed with extraction buffer. Thecombined filtrate, at a 4-5-fold dilution of theoriginal suspension, was called fraction I.Heat precipitation. Fraction I was adjusted to

pH6.5 with lM-KOH, heated rapidly to 54°C andmaintained at this temperature for 15min withoccasional stirring. It was then cooled to 10°C overliquid N2 and centrifuged for 30min at 23000g.The supernatant is fraction II.Acid and second heat precipitation. Fraction II was

adjusted to pH4.3 by the dropwise addition of 1 M-citric acid and centrifuged at 23000g for 1 h. Thepellet was resuspended in 250ml of extraction buffer.After homogenization the pH was adjusted to 6.9with 1M-HCl, the mixture heated rapidly to 50°C,and maintained at this temperature for 10min withoccasional stirring. The solution was chilled in anice bath to 10°C and centrifuged for 30min at23000g. This is fraction III.(NH4)2SO4 fractionation. Fraction III was diluted

to 300ml with extraction buffer and adjusted topH 7.0 with IM-HCl. (NH4)2SO4 (63 g) was addedgradually, and the mixture was stirred for 30min.The precipitate was removed by centrifugation for20min at 23000g. An additional 54g of (NH4)2SO4was added to the supernatant solution and afterstirring for 30min, the precipitate collected bycentrifugation for 30min at 23000g. The pellet wasresuspended in 80ml of ion-exchange buffer consist-ing of 10mM-KH2PO4, SmM-Tris/HCl, 45mM-KCl,0.45 mM-EDTA, 0.5% (v/v) a-thioglycerol and 25%(v/v) glycerol, pH8.0 at 4°C, and if necessary stored at-18°C. This is fraction IV.

Ion-exchange step. DEAE-Sephadex A-50 wasequilibrated with ion-exchange buffer to within0.01 pI unit, and the slurry supplemented with0.024% (w/v) phenylmethanesulphonyl fluoride.Fraction IV was added in a ratio of 2.5 units/g ofgel that had been allowed to filter to completionwithout suction. Then 1 gel volume of ion-exchangebuffer was added and the slurry stirred for 50min. Asmall portion of liquid was filtered off and assayedfor activity. Additional gel (2-5%) was added ifrequired and stirred for 10min until the activity in afiltered portion of liquid was less than 0.02unit/mi.The gel was collected on a Biuchner funnel. To thegel was added 2 gel volumes of ion-exchange buffersupplemented with 52.5mM-KCI, and the slurry wasstirred for 10min. The gel was collected on aBuchner funnel, and washed with 0.5 gel volume ofion-exchange buffer containing 52.5 mM-KCl and

Vol. 173

the filtrate discarded. Enzyme activity was elutedwith three washes of 1 gel volume of 0.15M-KCI-supplemented ion-exchange buffer followed by0.5 gel volume of 0.21 M-KCl-supplemented ion-exchange buffer. The elution filtrates were pooledand successive portions of 200ml were diluted to1 litre with 5% (v/v) glycerol/I mM-mercaptoethanol,and the enzyme was collected by suction filtrationon a bed of Whatman DE-23 ion-exchange cellulosecontaining 1 g wet wt. of gel for every 70 units ofenzyme activity. The enzyme was eluted by stirringthe gel for 10min in 5-10 gel volumes of 1 M-K2HPO4/1 mM-mercaptoethanol/22% (v/v) glycerol/0.024% (w/v) phenylmethanesulphonyl fluoride, andthe eluate collected by suction filtration. If necessary,the eluate may be stored at -18°C. This is fraction V.

Affinity chromatography. Fraction V enzyme wasequilibrated by passage through a Sephadex G-25column into the affinity-chromatography buffer. Theequilibrated enzyme, containing 500 units in avolume less than 35 ml, was loaded on the NAD+-Sepharose 4B column equilibrated with the samebuffer. Fractions (lOml) were collected at 1.2ml/min.The column was then washed with 1 column volumeof affinity-chromatography buffer, followed byenzyme elution with 200ml of buffer containing10mM-NAD+. Fractions containing activity werepooled, supplemented with solid K2HPO4 to 0.1 M,concentrated by ultrafiltration and stored at -1 8°C.This is fraction VI.

Results

Table 1 summarizes the rapid enzyme-purificationprocedure described above with data for preparationsfrom N.G. & S.F.-supplied yeast. When a sample ofenzyme was measured for activity in the standardassay of Steinman & Jakoby (1967) with benzalde-hyde as substrate, the activity was 29.5% of thatdetermined by the standard assay used for thepresent work. Yields of activity were independent ofthe method of extraction and independent of thesource of yeast.Under the conditions of the described purification

procedure phenylmethanesulphonyl fluoride is hydro-lysed with a half-life of 100min (Pringle, 1976). Whenattempts were made to compensate for this hydrolysisby continuous addition of phenylmethanesulphonylfluoride during extraction of the acetone-driedpowder, and the addition of frequent small portionsduring subsequent procedures, activity yields weredecreased. Fraction III had only 25% of the activityof the standard preparation, and subsequent frac-tions showed progressively poorer activities, althoughthe mobility of the active band on polyacrylamide-gel electrophoresis remained the same.A portion (1 mg) of freeze-dried fraction VI gave a

result of 1.05mg of protein by the method of Lowry

777

K. A. BOSTIAN AND G. F. BETTS

Table 1. Summary of aldehyde dehydrogenase purificationResults are based on the use of 200g of acetone-driedpowder derived from 0.7 kg of packed yeast cells.Fractions were obtained as given in the text. Activitywas measured in standard reaction mixtures andprotein was determined by the method of Lowryet al. (1951).

VolumeFraction (ml)

I 1190II 1140III 278IV 80V 215VI 1200Biofiber 75

concentrate

Totalprotein(mg)2323011172307518201289590

Totalactivity(units)

*

*

3690t3640345032103090

Specificactivity

(units/mg)>0.2t>0.3t1.22.0

27.034.034.0

* Aldehyde dehydrogenase activity masked by otherNAD(H)-dependent oxidoreductase.

t On the basis of total activity demonstrated byfraction III.

$ All alcohol dehydrogenase removed by this step.

et al. (1951) standardized against bovine serumalbumin. The same amount of enzyme in 1 mlof 0.1M-Tris/HCl buffer (pH8.0)/lmM-mercapto-ethanol/0.1 M-KCl/10% (v/v) ethanediol had A260values of 0.659 and A280 values of 1.192, giving avalue of 1.35mg of protein by the method ofWarburg& Christian (1941).Disc polyacrylamide-gel electrophoresis of N.G. &

S.F.-supplied yeast aldehyde dehydrogenase prepara-tions made in the presence or absence of phenyl-methanesulphonyl fluoride is shown in Fig. 1. Thephenylmethanesulphonyl fluoride preparation ex-hibits a single sharp coincident band of protein andactivity with a mobility relative to BromophenolBlue of 0.185. Preparations stored at -18°C or 4°Cin the absence of phenylmethanesulphonyl fluoridedid not change in electrophoretic mobility. Enzymeprepared in the absence of phenylmethanesulphonylfluoride always resulted in discrete bands, althoughthe consequent proteolysis resulted in various electro-phoretic patterns. On several occasions we haveobserved only single bands of activity which whentested on gels pre-run for 1 h before use result againin an electrophoretic mobility of 0.185. The samesamples run on gels not pre-run demonstratedelectrophoretic mobilities of 0.262. More com-monly, fractions before affinity chromatographyprepared in the absence of phenylmethanesulphonylfluoride show multiple bands of activity at relativemobilities of either 0.185 and 0.230 if pre-run or0.262, 0.290 and in addition in some preparations at0.310 if not (Fig. la). When purification is continuedthrough the affinity chromatography there is a partial

Fig. 1. Polyacrylamide-gel electrophoresis of aldehydedehydrogenase

(a) Diagrammatic representation of gel-el8ctro-phoresis patterns obtained from fraction V preparedin the absence (gel A) and in the presence (gel B) of0.024%4 phenylmethanesulphonyl fluoride, bothelectrophoresed on non-pre-run gels, and in thepresence of0.024%/ phenylmethanesulphonyl fluoride,electrophoresed on pre-run gels (gel C). Activeenzyme bands were-stained by enzymically reducedNitro Blue Tetrazolium. (b) and (c), Densitometertraces of gels stained respectively for enzyme activitywith Nitro Blue Tetrazolium and for protein withCoomassie Blue, by using fraction VI prepared in0.024%/ (w/v) phenylmethanesulphonyl fluoride andelectrophoresed on pre-run gels. Arrow indicates themarker dye Bromophenol Blue.

separation of these multiple forms, such that thosewith faster electrophoretic mobility tend to beeluted earlier from the affinity column. This indicatesthat enzyme molecules modified by proteolysis haveless affinity for immobilized ligand. Enzyme purifiedfrom Anheuser-Busch yeast by either the rapid

1978

778

PURIFICATION AND PROPERTIES OF YEAST ALDEHYDE DEHYDROGENASE

(b) (c)

10 bb

8

6~~~~~~

0 6 . -. 0.e. 04 5 0 7 0 9

xf

Relative moby E10fg

h

2-

0 0.2 0.4 0.6 0.8 30 40 50 60 70 80 90

Relative m'obility Elution volu'me (ml)

Fig. 2. SDS/polyacrylamide-gel electrophoresis and Sepharose 6B CL chromatography of aldehyde dehydrogenase andmarker proteins

(a) Densitometer trace of Coomassie Blue-stained polyacrylamide gel run in 0.1% (w/v) SDS. The peaks are:a, bovine serum albumin dimer; b, aldehyde dehydrogenase dimer; c, bovine serum albumin monomer; d, aldehydedehydrogenase monomer; e, aldehyde dehydrogenase A; f, carbonic anhydrase. Arrow (2) indicates marker dyeBromophenol Blue. (b) Molecular-weight calibration curve for SDS/polyacrylamide gels run with the proteinsstandards: a, bovine serum albumin dimer; b, phosphorylase A; c, bovine serum albumin monomer; d, catalase;e, ovalbumin; f, carbonic anhydrase. Arrow indicates relative mobility of aldehyde dehydrogenase. (c) Molecular-weight calibration curve for Sepharose 6B CL chromatography of the standards: a, Blue Dextran; b, glutamatedehydrogenase; c, thyroglobulin; d, f,-galactosidase; e, catalase; f, aldolase; g, bovine serum albumin; h, ovalbumin;i, carbonic anhydrase; j, cytochrome c. Arrow indicates elution volume for aldehyde dehydrogenase. Aldehydedehydrogenase was obtained from fraction VI prepared in the presence of 0.024% phenylmethanesulphonyl fluoride.SDS/polyacrylamide gel-electrophoresis and Sepharose 6B CL chromatography were performed and proteins were

prepared as in the Materials and Methods section.

Vol. 173

(1)

d(a)

b

779

K. A. BOSTIAN AND G. F. BETTS

affinity-chromatographic procedure or by thedehydrogenase A-type procedure had identical gel-electrophoretic behaviour.

SDS/polyacrylamide gel-electrophoresis offractionVI enzyme is shown in Fig. 2(a). Bovine serumalbumin and carbonic anhydrase were used asmarkers and were run in the same gels as the enzymesample. Arrow (1) (Fig. 2a) indicates the mobility ofa yeast aldehyde dehydrogenase designated A byClark & Jakoby (1970a) and was calculated fromtheir data (Clark & Jakoby, 1970b) based on thesame marker proteins. Under the conditionsdescribed, most of oligomeric enzymes dissociateinto subunits of uniform charge and migrate relativeto size on electrophoresis (Weber & Osborn, 1969).Calibration against protein standards of knownmolecular weight (see the Materials and Methodssection) provide a monomeric molecular weight forfraction VI enzyme from N.G. & S.F.-supplied yeastof 62500. Preparations of enzyme from Anheuser-Busch yeast made by either purification procedureprovide an identical monomeric mol.wt. estimate of62500.The molecular weight of intact fraction VI enzyme

from N. G. & S.F.-supplied yeast was estimated bymolecular exclusion chromatography on Sepharose6B CL columns (Porath et al., 1971) calibratedagainst globular proteins of known molecular weight(see the Materials and Methods section; Fig. 2b). Acoincident band of protein and activity was elutedat a volume of 65.5 ml (±2.5 %), corresponding to amol.wt. of 240000.Amino acid analysis of fraction VI enzyme

prepared in phenylmethanesulphonyl fluoride isshown in Table 2. Results are expressed as integralnumbers of residues per 50000g of protein andcompared with the ion-exchange data of Clark &Jakoby (1970a). Histidine forms derivatives poorly,and this method gives irregular results with meth-ionine and cysteine, so no data were collected forthese amino acids. The overall pattern for the twodeterminations is similar. The reliability of such com-parisons is ±17% or 1 residue in 6. Only arginine andtyrosine show differences outside this error.

N-Terminal amino acid analysis was performed onfraction VI enzyme prepared in the presence ofphenylmethanesulphonyl fluoride. For N.G. & S.F.-supplied yeast enzyme other than the reactive sidegroups Dns-e-lysine and Dns-a-tyrosine, the onlyspot identifiable as an N-terminal amino acid wasDns-valine. A spot corresponding to Dns-serine, theN-terminus of aldehyde dehydrogenase A (Clark &Jakoby, 1 970a), was not present in any of thechromatograms produced from samples prepared byseveral different dansylation techniques. No otherspots were present in sufficient intensity to representheterogeneity at the N-terminus. Enzyme prepara-tions from Anheuser-Busch yeast made either by the

Table 2. Amino acid analysis of aldehyde dehydrogenaseComparison of the amino acid composition offraction VI enzyme and aldehyde dehydrogenase Afrom the data of Clark & Jakoby (1970a). G.l.c.analysis of amino acid derivatives of fraction VIprepared in the presence of 0.024°% phenylmethane-sulphonyl fluoride was performed as in the Materialsand Methods section. Threonine, tyrosine and serineare extrapolated to zero time of hydrolysis and theremainder are averages of 24 and 48h hydrolyses.

Amino acidLysineArginineAspartic acidThreonineSerineGlutamic acidProlineGlycineAlanineValineIsoleucineLeucineTyrosinePhenylalanine

Integral no. of residues/50000g

Dehydrogenase A(Clark & Jakoby,

Fraction VI 1970a)32 3212 1544 4420 2323 2338 4116 1939 3940 3629 2931 3526 309 1217 17

dehydrogenase A-type procedure or by the rapidaffinity-chromatographic procedure (both in phenyl-methanesulphonyl fluoride) were micro-hetero-geneous at the N-terminus, in which Dns-serineconstitutes approx. 20% of the a-amine derivedamino acids present. Dns-valine was about equalin intensity with Dns-serine, whereas the majorityof the remaining fluorescence corresponding to thea-amine derived Dns-amino acids is identifiableas Dns-aspartic acid and Dns-glutamic acid.The behaviour of fraction V enzyme from N.G. &

S.F.-supplied yeast on NAD+-Sepharose affinitycolumns in the absence of added aldehyde and freeNAD+ is shown in Fig. 3. With the high-ionic-strength buffer used for these experiments, non-specific effects are minimized, and evidence for thebiospecific retardation of enzyme is clearly seen bycomparison with the elution profile of bovine serumalbumin, run under identical conditions (Fig. 3).Supplementation with 5mM-NAD+ at fraction 20resulted in elution of enzyme with the nucleotidefront (Fig. 3). The sharpness of the elution profilewas directly related to the concentration of nucleotidein the elution buffer. NAD+ or 5'-AMP were equallyeffective eluting agents. Decreasing the ionic strengthof the chromatography buffer results in an increasein affinity. As the potassium phosphate concentra-tion is decreased the front of activity occurs later

1978

780

PURIFICATION AND PROPERTIES OF YEAST ALDEHYDE DEHYDROGENASE

8

°a0 6-

x

10 4-

2

0 30 40

Fraction number

1.5

E1.

1 .0 .4;Y

UCd

EN

0.5 r.

Fig. 3. Affinity chromatography of aldehyde dehydrogenase on NAD+-Sepharose 4BChromatography was performed in 0.1 M-K2HPO4 (pH8.0)/22% (v/v) glycerol/I mm-mercaptoethanol/1.2mM-phenylmethanesulphonyl fluoride buffer. *, Fraction V aldehyde dehydrogenase (26 units); *, bovine serum albumin(1.3 mg); *, 26 units of fraction V aldehyde dehydrogenase in which running buffer was supplemented with 5mM-NAD+ at fraction 20. Samples were added in a volume of 1 ml. ,,*, in units/ml of activity determined in the standardassay; *, A280 units. Column dimensions were 1.25cm x 16cm. The volume of each fraction was 1.2ml. Flow ratewas 0.9ml/min.

and is smaller than under the conditions of Fig. 3.There is a consequent increase in the amount ofenzyme eluted with the front of free nucleotide.Elution of contaminating protein is only marginallyeffected by this decrease in ionic strength, so thiseffect reinforces the biospecific interaction andenables more enzyme to be processed. A 45mM-potassium phosphate buffer, pH 8.0, allows asufficient amount of retardation, but fulfils the K+requirement for stability and was adopted forroutine purification. Attempts at batchwise pro-cedure were unsuccessful.

Blue Dextran binding to several enzymes has beenestablished by molecular-sieve chromatography(Staal et al., 1969, 1971), affinity chromatography(Kopperschlager et al., 1971; Ryan & Vestling, 1974;Thompson et al., 1975) and competitive inhibition(Thompson et al., 1975). The specific binding ofenzyme to immobilized Blue Dextran has beencorrelated with the possession of the dinucleotidefold common to several classes of enzymes (Thomp-son et al., 1975). 5'-AMP is a competitive inhibitorof several dehydrogenases and has proved a usefulimmobilized ligand for affinity chromatography(Brodelius & Mosbach, 1973; Craven et al., 1974).Fig. 4 shows that both 5'-AMP and Blue Dextranare inhibitors of aldehyde dehydrogenase, com-petitive with NAD+. The Ki values for 5'-AMP andfor Blue Dextran are 790#M and I AM respectively.The binding ofenzyme to 5'-AMP affinity columns

is shown in Fig. 5. Binding is similar to that on an

Vol. 173

NAD+ column, but the column used here has approx.80% less binding capacity. 5'-AMP and NAD+ areequally effective eluting agents at equivalent con-centrations. Fig. 5 also shows that inclusion through-out the chromatography of acetaldehyde at con-centrations optimal for activity only marginallyimproves the affinity of the column for enzyme. BothAnheuser-Busch fraction-V enzyme and that pro-duced by the dehydrogenase A-type purificationbind to 5'-AMP columns with a similar or slightlygreater affinity than the N.G. & S.F.-supplied yeastenzyme affinity shown in Fig. 5.The affinity ofN.G. & S.F.-supplied yeast aldehyde

dehydrogenase for Blue Dextran columns in lowionic strength buffer is demonstrated in Fig. 6.Affinity is increased by the presence of acetaldehydein the running buffer. Enzyme is eluted from thecolumn by buffers of high ionic strength. Unlike thebehaviour of other enzymes that bind specifically toBlue Dextran columns, low concentrations of freenucleotide ligand (1-10mM) fail to accelerate elutionin low-ionic-strength buffers. However, at concen-trations (0.1 M) of inorganic salts producing onlymoderate increases in elution, the addition of 5mM-NAD+ or -5'-AMP causes a sharp specific elution.The addition of 20% (v/v) glycerol required forstability decreases enzyme affinity.The ubiquitous contamination of endogenous

aldehydes in a number of general reagents haspreviously been demonstrated by Bradbury & Jakoby(1971) using the colorimetric method of Sawicki et al.

781

K. A. BOSTIAN AND G. F. BETTS

-10-4/[NAD+ I (m- )Fig. 4. Inhibition of aldehyde dehydrogenase by Blue

Dextran or 5'-AMPDouble-reciprocal plots of initial velocity and NAD+concentrations at several fixed concentrations ofinhibitor. In (a) concentrations of Blue Dextranwere: *, 0; A, 0.05pM; U, 1 AM. In (b) concentrationsof 5'-AMP were: 0, 0; A, 1mM; U, 2mM; o, 3mr;A, 4mM. Activity was measured in standard reactionmixtures, except for varied NAD+ concentrationsand the presence of Blue Dextran or 5'-AMP asshown.

(1961). An assessment of contamination in thereagents used here for affinity chromatography wastherefore made both by the colorimetric method andby enzymic reactivity with aldehyde dehydrogenase.Tris/HCl, mercaptoethanol, a-thioglycerol and pos-sibly glycerol interfered with the colorimetricreaction.By the 3-methyl-2-benzothiazolone hydrazone

method, polyhydric alcohol and secondly NAD+ arethe most highly contaminated constituents in theaffinity-chromatography buffer; 100% glycerol gavean aldehyde concentration of 0.3mm and aldehydewas 2.4% of the concentration of NAD+. Con-tamination by enzymically reactive aldehyde in thealdehyde dehydrogenase reaction mixture sup-plemented to 25% (v/v) with glycerol was estimatedat 40M. Chemical analyses using 3-methyl-2-benzothiazolone hydrazone performed on samples

3

2

0 10 20 30 40 50Fraction number

Fig. 5. Affinity chromatography ofaldehyde dehydrogenaseon 5'-AMP-Sepharose 4B

Chromatography was performed in 0.1M-K2HPO4(pH 8.0)/22% (v/v) glycerol/I mM-mercaptoethanol/1.2mM-phenylmethanesulphonyl fluoride buffer, with(M) or without (U) 2mM-acetaldehyde; 17 units offraction V aldehyde dehydrogenase were used ineach case, and buffer was supplemented with 5mM-5'-AMP at fraction 20. All other conditions were asin Fig. 3.

E 0.6

1-

>, 0.44:E

CE0 0.2

N

0 10 20 30Fraction number

40 50

Fig. 6. Affinity chromatography ofaldehyde dehydrogenaseon Blhe Dextran-Sepharose 4B

Chromatography was performed in lOmM-Tris/HCI(pH8.0)/0.4mM-mercaptoethanol buffer, with (0) orwithout (N) 2mM-acetaldehyde; 12 units of fractionV aldehyde dehydrogenase in a volume of 1.5mlwere used in each case. Elution buffer [0.1 M-K2HPO4(pH 8.0)/22% (v.v) glycerol/0.4 mM-mercaptoethanol/5 mM-NAD+] was added at fraction 25 in each case.Column dimensions were 1.2cm x 7cm, and thevolume of each fraction was 1.2 ml. Flow rate was0.8mI/min. Activity was determined in the standardassay mixture.

removed from the reaction mixture at various timeintervals provide an estimated initial aldehyde con-centration of 100tM, with no reduction during thereaction period, in spite of the fact that 40#M-aldehyde is removed by enzymic oxidation. The datasuggest that the 3-methyl-2-benzothiazolone hydra-zone test (as used here) is not a reliable test forcontaminating aldehyde.

1978

I

782

PURIFICATION AND PROPERTIES OF YEAST ALDEHYDE DEHYDROGENASE

Contamination of glycerol by enzymically reactivealdehyde present in the affinity-chromatographybuffer is 26AM. When 10% (v/v) ethanediol was sub-stituted for glycerol in the affinity-chromatographybuffer, a 10-fold reduction in enzymically reactivecontaminating aldehyde (to 3AuM) was achieved with-out altering column capacity and elution profilesof enzyme from NAD+ and 5'-AMP columns.3 - Methyl - 2 - benzothiazolone hydrazone - estimatedaldehyde contamination by 10% (v/v) ethanediolis 10pM based on the standard curve for acetal-dehyde.

In these affinity-chromatographic investigations, ifaldehyde is to be an obligatory leading ligand beforethe binding of immobilized NAD+, 5'-AMP or BlueDextran, then for biospecific separation to beunchanged by the addition of saturating concentra-tions of aldehyde, contaminating aldehyde must beclose to saturation for enzyme binding sites. More-over, since a 10-50-fold decrease in contaminatingaldehyde does not result in an alteration of bio-specific elution, within experimental error, thealdehyde sites would have to be 99% saturated. Thusat a total contaminating aldehyde concentrationmeasured enzymically or colorimetrically of 10puMor less the K, for contaminating aldehyde wouldhave to be less than 0.1 AM. On this basis the actualKm values for good aldehyde substrates would haveto be less than that determined with these omnipresentcontaminating aldehydes by a factor of 1 + I/K,, or100-fold. This requires that true Km values for atleast two aldehyde substrates, isobutyraldehyde andp-nitrobenzaldehyde (Steinman & Jakoby, 1968), areless than 10nM, an unreasonably high affinity forsuch enzyme substrate complexes.

The binding of fraction VI enzyme to NAD+ in theabsence of added aldehyde was determined directlyby equilibrium dialysis and by the rate of dialysismethod (Colowick & Womack, 1969). Protein wasdetermined by the method of Lowry et al. (1951) andmolarity calculated with a mol.wt. of 248000. Forboth experiments enzyme was approx. 5,UM, andNADI varied from 1 to 0.01 mM. Scatchard-typeplots of the binding data for enzymes from N.G. &S.F.-supplied yeast are shown in Fig. 7. A least-squares fit to the equilibrium-dialysis data in Fig.7 (a) provides a dissociation constant (1/slope) of120pM for each of 4 coenzyme-binding sites permolecule. Data for the rate of dialysis experiment inFig. 7(b), performed under similar experimentalconditions, provide a dissoc}dtion constant of 68AuMfor each of 4 coenzyme-binding sites per molecule.NADI binding to fraction-VI enzyme purified fromAnheuser-Busch yeast determined by an identicalequilibrium-dialysis procedure gave a dissociationconstant of 200pM for each of 4 coenzyme-bindingsites per molecule.

Discussion

Evidence is presented above that the enzymepurified from N.G. & S.F.-supplied yeast asdescribed here is different in primary structure fromthat previously reported (Steinman & Jakoby, 1967;Clark & Jakoby, 1970a). The most unequivocalindication of a difference is the N-terminal aminoacid valine for this enzyme preparation, comparedwith serine for the enzyme designated dehydrogenaseA (Clark & Jakoby, 1970a). The total amino acidanalysis indicates minor differences between the

4

3

Z 2

x

0-1.

0 1 2

V

3

U

0

c)

x

I0

0 2 3 4v

Fig. 7. Binding ofNAD+ to aldehyde dehydrogenase(a) Scatchard type plot of variation of v with V/CNAD+ from equilibrium dialysis data; (b) Scatchard type plot ofvariation of vU with iV/CNAD+ from rate of dialysis data. Dialysis was performed in 0.1 M-Tris/HCI (pH 8.0)/20% (v/v)ethanediol/l mM-mercaptoethanol, at 4°C. Aldehyde dehydrogenase was obtained from fraction VI prepared in thepresence of 0.024%4 (w/v) phenylmethanesulphonyl fluroide. Dialysis was performed as in the Materials and Methodssection.

Vol. 173

783

K. A. BOSTIAN AND G. F. BETTS

preparations with only arginine and tyrosine contentbeing significantly different.

SDS/polyacrylamide-gel electrophoresis gives amol.wt. of 62500 for the monomer, compared with56000 reported for dehydrogenase A (Clark &Jakoby, 1970b), determined by the same method.Molecular-sieve chromatography gives a mol.wt. of240000 for the intact enzyme, compared with 214000reported for dehydrogenase A (Clark & Jakoby,1970a). However, in our hands enzyme preparationsfrom yeast supplied by either N.G. & S.F.or Anheuser-Busch and purified by the methodreported here or as previously (Clark & Jakoby,1970a) all co-migrate on SDS/polyacrylamide gels.These data might reflect a minor difference in the

form of the proteins in vivo isolated from twodifferent yeast strains. An alternative hypothesis inview of considerations below is that both proteinshave the same primary structure in vivo and thatdifferences between the isolated proteins are theresult of differential proteolytic digestion duringpurification which results in the removal of a segmenttoo small to modify the migration rate on SDS/poly-acrylamide gels.

Highly specific proteolytic modification of yeastaldehyde dehydrogenase in vitro resulting in theproduction of distinct but active enzyme forms hasbeen demonstrated both in the original work ofClark & Jakoby (1970a) and by the findings presentedhere (Fig. l). By avoiding autolytic procedures and/or by the addition of high concentrations of protein-ase inhibitors di-isopropyl phosphorofluoridate or

phenylmethanesulphonyl fluoride, a single electro-phoretically homogeneous enzyme activity is ob-served. It has been clearly documented, however, inan exhaustive review by Pringle (1976) that no singleprecaution, such as the inclusion of di-isopropylphosphorofluoridate or phenylmethanesulphonylfluoride, is completely adequate for safe-guardingagainst modification. In many cases, proteinasesshow residual activity or are unaffected by thepresence of these agents, and therefore, any pro-

cedure that speeds the separations of proteinasesfrom the required enzyme is likely to result in a

lesser amount of the degraded forms.The slowest steps in published purification pro-

cedures of yeast aldehyde dehydrogenase are ion-exchange chromatography on DEAE-Sephadex A-50and hydroxyapatite (Steinman & Jakoby, 1967;Clark & Jakoby, 1970a). These were used at a timewhen it was recognized that proteolytic activitypersisted in the preparation even in the presence ofphenylmethanesulphonyl fluoride or di-isopropylphosphorofluoridate. Moreover, to allow rapidequilibration and sample binding to the ion-exchangematerial it is necessary to lower glycerol and K+concentrations, both ofwhich are required for enzymestability.

As a possible alternative to these lengthy ion-exchange chromatographic steps, we attempted tointroduce an early rapid affinity-chromatographystep. After the report of a compulsory sequentialmechanism in which aldehyde is the leading substrate(Bradbury & Jakoby, 1971a,b) and the report thatdirect binding studies reveal no nucleotide binding inthe absence of aldehyde (Bradbury & Jakoby, 1971a),we made several unsuccessful attempts to immobilizean aldehyde on a polysaccharide for affinity chroma-tography. However, enzyme does bind to columnscontaining immobilized coenzyme, partial coenzyme(AMP) and probe for dinucleotide fold (BlueDextran). Unfortunately, dispensing with the ion-exchange steps completely and proceeding from(NH4)2SO4 fractionation directly to affinity chroma-tography results in preparations demonstratingmultiple protein bands on disc polyacrylamide gelelectrophoresis of which only one band has aldehydedehydrogenase activity. Presumably an ion-exchangestep removes other NAD+-linked enzymes. With abulk procedure for DEAE-Sephadex A-50 ionexchange rather than column chromatography, thepurification step is marginally less efficient, but onlytakes one-tenth of the time. The enzyme can then bepurified to homogeneity (as judged by polyacryl-amide-gel electrophoresis and N-terminal amino acidanalysis) by a relatively rapid NAD+ affinitychromatography.A slow step (4h) in the present purification is

the initial resuspension of acetone-dried powder.Although it is true that proteinases are present inhigher concentration at this stage than any other,,sotoo are the yeast's own proteinase inhibitors. A pHwas purposely chosen at which proteinase-inhibitorcomplexes are most stable (Pringle, 1976). The highconcentration of other proteins present at this stageoffers some protection. Moreover, it has been shownthat an alternative glass-bead breakage procedurefor extraction yields an enzyme with apparentlyidentical characteristics. This more direct and rapidmethod of extraction is the method of choice at themoment as it was recently made applicable to largeamounts of cells.An attempt was made to ascertain whether

differences in protein characteristics between thepresent preparation and that previously reported(Steinman & Jakoby, 1968; Clark & Jakoby, 1970a;Bradbury & Jakoby, 1971a) are due to differences inisolation methods rather than to source. Thus thepresent more rapid isolation methods were used onAnheuser-Busch yeast (that used by Clark & Jakoby,1970a) to see if an enzyme could be obtained havingcharacteristics of the present preparation. N-Terminal amino acid analysis was ambiguous. Innone of the Anheuser-Busch yeast enzyme prepara-tions was a single a-amino derivative of amino acidsobserved. Aspartic acid and glutamic acid comprised

* 1978

784

PURIFICATION AND PROPERTIES OF YEAST ALDEHYDE DEHYDROGENASE

about 50% of the a-amine derivatives, with the nextmost prevalent spots corresponding to Dns-valineand Dns-serine. These results were observed regard-less of the method of purification. In di-isopropylphosphorofluoridate preparations of dehydrogenaseA only 80% of the a-amine derivatives of aminoacids were serine, and in purification with phenyl-methanesulphonyl fluoride as little as 60% (W.Jakoby, personal communication). The generallyhigher proteinase activity in Anheuser-Busch yeastenzyme preparations indicated by N-terminal aminoacid analysis is also reflected in the lower enzymeactivity recoverable during purification, even inphenylmethanesulphonyl fluoride (Clark & Jakoby,1970a; the present work).To explain these results one must assume that

even when careful attempts have been made tofollow the dehydrogenase A-type purification out-line, some alteration in procedure (e.g. breakage withglass beads rather than Manton-Gaulin homogeniza-tion, or a change in the Anheuser-Busch strain ofyeast, either physiologically or genetically, avoids aproteolysis or is otherwise responsible for modifyingthe enzyme.A significant feature of the above work is that a

purification technique involving binding to NAD+,AMP or Blue Dextran affinity columns has beensuccessfully applied to yeast aldehyde dehydrogenasein spite of the report of Bradbury & Jakoby (1971a)of a compulsory order of substrate addition in whichaldehyde is the leading substrate. Moreover, ourenzyme preparations bind to these affinity columns inbuffers similar to those that failed to demonstrateNAD+ binding to dehydrogenase A (Bradbury &Jakoby, 1971a). Supplementation with aldehyde atconcentrations optimal for activity does little toenhance enzyme affinity toward 5'-AMP columnsand only slightly increases affinity for Blue Dextrancolumns. Conceivably, this might indicate priorbinding by contaminating aldehydes rather thanbinary enzyme-NAD+ complex. Aldehyde deter-minations reveal the presence of contaminatingaldehyde in the chromatography buffer. These areintroduced predominantly by polyhydric alcohol.However, best estimates indicating a 10-50-folddecrease in contaminating aldehyde concentrationby removal of glycerol or its substitution by ethane-diol does not alter the biospecific affinity of aldehydedehydrogenase for these columns.For several reasons (Porath & Kristiansen, 1975)

it is not possible to determine accurately the effectiveligand concentration of the affinity gels used here.However, the total amount of immobilized ligandpresent is 50,umol/g and 150,umol/g of dry gel forimmobilized NAD+ and 5'-AMP respectively.Assuming that total immobilized ligand approxi-mates to effective ligand concentration, and by usingbovine serum albumin as a protein for which bio-

Vol. 173

affinity is zero, behaviour can be compared withpredicted binding and elution for simulated binarysystems ofenzyme and immobilized ligand (Graves &Wu, 1974). The behaviour of aldehyde dehydro-genase is consistent with a dissociation constant forimmobilized NAD+ and 5'-AMP in the range 1-0.10mM. The K1 for 5'-AMP is 0.79mM (Fig. 4),and under the chromatographic conditions describedthe K. for NAD+ has been determined by us as0.12mm by equilibrium dialysis. These data demon-strate that a specific interaction between aldehydedehydrogenase and immobilized NAD+ or 5'-AMPis taking place. The introduction of non-specificinteractions at low ionic strength is advantageouslyused to increase this affinity, and therefore theamount of enzyme bound to either column, withno other obvious alteration in chromatographicbehaviour.The qualitative data on binding to Blue Dextran

columns as well as kinetic inhibition data indicatethat at least some portion of the dinucleotide foldcommon to all dehydrogenases thus far investigated(Thompson et al., 1975) is present in aldehydedehydrogenase.

Binding to columns of immobilized AMP, BlueDextran or NAD+ is not a critical test of binding tofree NAD+. However, the data (Fig. 7) clearly showthat NAD+ interacts with aldehyde dehydrogenasein the absence of added aldehyde. Respective valuesfor coenzyme dissociation constants from the twodialysis studies were 120 and 68AM for each of fourcoenzyme-binding sites per tetrameric molecule. Thisimplies one active site per monomer, a general casefor most enzymes with fewer than six subunits.Known reactive aldehydes are present to a certain

extent in the rate of dialysis experiments, but havebeen completely oxidized during equilibrium dialysis.In either case an NAD+-regenerating system wasincluded so the system was not measuring the bindingof product NADH. The concentration of enzymic-ally unreactive aldehyde is negligible with the buffersystem used for equilibrium dialysis and is unlikelyto exceed 10% saturation of aldehyde sites. Thepresence of additional enzymically reactive aldehydesin the rate of dialysis experiments does not alter thenumber of available binding sites per enzyme mole-cule. Thus unless unreasonable assumptions aremade about contaminating aldehyde concentrations,NAD+ binds to the free enzyme.Once again studies have been made in an attempt

to resolve the discrepancy in NAD+-bindingcharacteristics between this enzyme preparation andthat previously reported (Bradbury & Jakoby, 1971a).This has been done both by using our purificationprocedures on the yeast used previously and byvarying the preparation procedure of enzyme fromN.G. & S.F.-supplied yeast in an attempt to obtainenzyme with identical coenzyme-binding charac-

785

786 K. A. BOSTIAN AND G. F. BETTS

teristics to those previously reported. Enzymeprepared rapidly from Anheuser-Busch yeast wasfound to bind to the affinity columns in a fashionidentical with enzyme rapidly purified from N.G. &S.F.-supplied yeast. NAD+ binding was subsequentlyconfirmed by equilibrium dialysis. Variations in thesource of yeast or the method of purification havealways resulted in enzyme that binds to coenzymeanalogue affinity columns and where tested binds toNAD+ during equilibrium dialysis.The implication from this work is that enzyme-

NAD+ binary complex does form, and that thismight therefore reflect an important binary complexin the kinetic mechanism. This is consistent with theknown mechanisms of all other dehydrogenasessufficiently investigated. Alcohol dehydrogenases,lactate dehydrogenase, malate dehydrogenase, for-mate dehydrogenase and glyceraldehyde 3-phosphatedehydrogenase all have ordered mechanism withcoenzyme as the leading substrate (Dalziel, 1975).The kinetic mechanisms of glutamate dehydrogenaseand isocitrate dehydrogenase are consistent with arandom-order mechanism where free enzyme bindscoenzyme (Dalziel, 1975). Coenzyme has also beenshown to bind with free rat and horse liver aldehydedehydrogenase (Takio et al., 1974; Tottmar et al.,1974; Weiner et al., 1974). Such a distinct differencein the proposed inability of free yeast aldehydedehydrogenase to bind NAD+ in a kineticallysignificant pathway has led us to a complete kineticinvestigation of our enzyme preparations. The dataare presented in the following paper (Bostian &Betts, 1978).

We thank Mr. P. Poole for his ready technical assistanceand Dr. C. Bruton for his generous help in the determina-tion of N-terminal amino acid. We also thank Mr.Kenneth Davy and Professor C. J. 0. R. Morris forperforming the amino acid analysis, and Ms. DianaGorringe for making available a calibrated Sepharose6B CL column. This work was supported, in part, byU.S. Public Health Service grant GM20755 awarded toD. J. Tipper.

ReferencesBostian, K. A. & Betts, G. F. (1978) Biochem. J. 173,

787-798Bradbury, S. L. & Jakoby, W. B. (1971a) J. Biol. Chem.

246, 1834-1840Bradbury, S. L. & Jakoby, W. B. (1971b) J. Biol. Chem.

246, 6929-6932Bradbury, S. L. & Jakoby, W. B. (1972) Proc. Natl.

Acad. Sci. U.S.A. 69, 2373-2376Brodelius, P. & Mosbach, K. (1973) FEBS Lett. 35,

223-226Chrambach, A., Reisfeld, R. A., Wyckoff, M. & Zaccari,

J. (1967) Anal. Biochem. 20, 150-154Clark, J. F. & Jakoby, W. B. (1970a) J. Biol. Chem. 245,6065-6071

Clark, J. F. & Jakoby, W. B. (1970b) J. Biol. Chem. 245,6072-6077

Colowick, S. P. & Womack, F. C. (1969) J. Biol. Chem.244, 774-777

Craven, D. B., Harvey, M. J., Lowe, C. R. & Dean,P. D. G. (1974) Eur. J. Biochem. 41, 329-357

Dalziel, K. (1975) Enzymes 3rd Ed. 11, 2-52Darbre, A. & Islam, A. (1968) Biochem. J. 106, 923-925Davis, B. J. (1964) Ann. N. Y. Acad. Sci. 121, 404-427Davy, K. W. & Morris, C. J. 0. R. (1976) J. Chromatogr.

116, 305-314Graves, D. S. & Wu, Y. T. (1974) Methods Enzymol. 34,

140-162Gray, W. R. (1967) Methods Enzymol. 11, 139-151Hartley, B. S. (1970) Biochem. J. 119, 805-822Kopperschlager, G., Diezel, W., Freyer, R., Liebe, S. &Hoffman, E. (1971) Eur. J. Biochem. 22, 40-45

Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall,R. J. (1951) J. Biol. Chem. 193, 265-275

Mosbach, K., Guilford, H., Ohlsson, R. & Scott, M.(1972) Biochem. J. 127, 625-631

Porath, J. & Kristiansen, T. (1975) in The Proteins(Neurath, H. & Hill, R. L., eds.) vol. 1, pp. 95-178,Academic Press, London and New York

Porath, J., Janson, J. C. & LMAs, T. (1971) J. Chromatogr.60, 167-177

Pringle, J. R. (1975) Methods Cell Biol. 12, 149-184Ryan, L. D. & Vestling, C. S. (1974) Arch. Biochem.

Biophys. 160, 279-284Sawicki, E., Hauser, T. R., Stanley, T. W. & Elbert, W.

(1961) Anal. Chem. 33, 93-96Sorger, G. J. & Evans, H. J. (1966) Biochim. Biophys.Acta 118, 1-8

Staal, G. E. J., Visser, J. & Veeger, C. (1969) Biochim.Biophys. Acta 185, 39-48

Staal, G. E., Koster, J. F., Kamp, H., Van Milligen-Boersma, C. & Veeger, C. (1971) Biochim. Biophys.Acta 227, 86-96

Steinman, C. R. & Jakoby, W. B. (1967) J. Biol. Chem.242, 5019-5023

Steinman, C. R. & Jakoby, W. B. (1968) J. Biol. Chem.243, 730-734

Stoppani, A. 0. M. & Milstein, C. (1957) Biochem. J. 67,406-416

Takio, K., Saho, Y. & Yonetani, T. (1974) in Alcohol andAldehyde Metabolizing Systems (Thurman, R. G.,Yonetani, T., Williamson, J. R. & Chance, B., eds.),pp. 115-123, Academic Press, London and New York

Thompson, S. T., Cass, K. H. & Stellwagen, E. (1975)Proc. Natl. Acad. Sci. U.S.A. 72, 669-672

Tottmar, S. 0. C., Petterson, H. & Kiessling, K. H. (1974)in Alcohol and Aldehyde Metabolizing Systems(Thurman, R. G., Yonetani, T., Williamson, J. R. &Chance, B., eds.), pp. 147-159, Academic Press,London and New York

Warburg, 0. & Christian, W. (1941) Biochem. Z. 310,382-421

Weber, K. & Osborn, M. (1969) J. Biol. Chem. 244, 4406-4412

Weiner, H., King, P., Hu, J. H. J. & Bensch, W. R. (1974)in Alcohol and Aldehyde Metabolizing Systems(Thurman, R. G., Yonetani, T., Williamson, J. R. &Chance, B., eds.), pp. 101-113, Academic Press,London and New York

Womack, F. C. & Colowick, S. P. (1973) MethodsEnzymol. 27, 464 471

1978