The Tungsten-Containing Aldehyde Oxidoreductase from Clostridium thermoaceticum and its Complex with a Viologen-Accepting NADPH Oxidoreductase

Biol. Chem. Hoppe-SeylerVol. 373, pp. 123-132, March 1992

The Tungsten-Containing Aldehyde Oxidoreductase fromClostridium thermoaceticum and its Complex with a Mologen-AcceptingNADPH Oxidoreductase

Gerhard STROBL, Richard FEICHT, Hiltrud WHITE, Friedrich LOTTSPEICH and Helmut SIMON

Lehrstuhl f r Organische Chemie und Biochemie der Technischen Universit t M nchen

(Received 26 November 1991)

Summary: Purification of aldehyde Oxidoreductasefrom C. thermoaceticum, the first detected enzymeable to reduce reversibly non-activated carboxylicacids to the corresponding aldehydes (White, H.,Strobl, G., Feicht, R. & Simon, H. (1989) Eur. J.Biochem. 184, 89-96), results in the generation ofmultiple forms of the enzyme. The specific activitiesfor the viologen-mediated dehydrogenation ofbutyraldehyde for the two main forms of the purifica-tion procedure are 530 and 450 U/mg. Two forms ofthe enzyme composed of α,/3- and α,β,γ-subunits,can be differentiated. The latter binds to red-Sepharose and can be eluted very specifically withNADPH. In contrast to the α,β-types the trimericforms also catalyse the reversible reduction ofoxidised viologen with NADPH (VAPOR activity).

The dimer α,β can oligomerize and the a,/3,y-trimercan easily form various oligomers or split off the y-

subunit. The apparent molecular masses of the sub-units α,β and γ are 64, 14 and 43 kDa. The α,/3-formreveals an apparent molecular mass of 86 kDa con-taining about 29 iron, 25 acid-labile sulphur, 0.8tungsten and forms about l mol pterine-6-carboxylicacid by permanganate oxidation. The correspondingvalues of the trimer showing a mass of 300 kDa, areabout 82 Fe, 54 S, 3.4 W and 2.5 pterine-6-carboxylicacid. In addition, 1.7 mol of FAD could be foundwhich seems to be a component of the y-subunit.Thealdehyde Oxidoreductase from C. thermoaceticumand that from C. formicoaceticum (White, H., Feicht,R., Huber, C., Lottspeich, F. & Simon, H. (1991)Biol. Chem. Hoppe-Seyler 372,999-1005) show qual-itative similarities as far as the Fe, S,Wand pterin con-tent and the broad substrate specificity are con-cerned. However, there are also surprisingly markeddifferences with respect to composition and amino-acid sequence.

Die Wolfram-haltige Aldehyd-Oxidoreduktase aus Clostridium thermoaceticum und ihre Komplexbildung miteiner Viologen-bindenden NADPH-OxidoreduktaseZusammenfassung: Die Reinigung von Aldehyd-Oxidoreduktase von C. thermoaceticum, dem erstenEnzym das in der Lage ist, reversibel nichtaktivierteCarbons uren zu den entsprechenden Aldehyden zureduzieren (White, H., Strobl, G., Feicht, R. & Si-mon, H. (1989) Eur. J. Biochem. 184, 89-96), f hrtzu verschiedenen Formen des Enzyms. Mit einem

Viologen als k nstlichem Mediator werden f r diezwei Hauptformen der Reinigungsprozedur spezifi-sche Aktivit ten f r die Dehydrierung von Butyralde-hyd von 530 bzw. 450 U/mg erreicht.Es k nnen a, - und α,β,γ-Formen unterschiedenwerden. Enzym, das die γ-Komponente enth lt, bin-det an rote Sepharose und kann sehr spezifisch mit

Enzymes:The aldehyde dehydrogenase from C. thermoaceticum is not yet registered in the IUB Enzyme Nomenclature.Abbreviations:DTE, dithioerythritol; FPLC, Fast-Protein Liquid Chromatography; NH2CO-MV, Ι,Ι'-carbamoylmethylviologen; SDS, sodiumdodecyl sulphate;Tris,Tris-(hydroxymethyl)aminomethane;V2®, oxidised; V®", reduced viologen;VAPOR, viologen accepting py-ridine nucleotide Oxidoreductase.

Copyright © by Walter de Gruyter & Co · Berlin · New YorkBrought to you by | University of California - San Diego

Authenticated | 132.239.1.231Download Date | 9/24/13 10:15 PM

124 G. Strobl, R. Feicht, H. White, F. Lottspeich and H. Simon Vol. 373 (1992)

NADPH eluiert werden. Im Gegensatz zu dem Di-mer a, katalysieren die trimeren Formen auch diereversible Reduktion von oxidierten Viologenen mitNADPH (VAPOR-Aktivität).Das Dimere kann oligomerisieren und die , , -For-men bilden ebenfalls verschiedene Oligomere, vondenen die -Untereinheit auch abgespalten werdenkann. Die apparenten molekularen Massen der Un-tereinheiten a, und sind 64, 14 und 43 kDa. Die

, -Form zeigt eine apparente molekulare Massevon 86 kDa und enthält pro Molekül ca. 29 Eisen-atome, 25 Säure-labile Schwefelatome, 0.8 AtomeWolfram und bildet nach Permanganatoxidation ca.l Molekül Pterin-6-carbonsäure. Die entsprechen-

den Werte einer trimeren Form mit einer molekularenMasse von 300 kDa sind ca. 82 Fe, 54 S, 3.4Wund 2.5Pterin-6-carbonsäure. Diese Form enthält zusätzlich1.7 mol FAD, die in der -Untereinheit lokalisiertsein sollten. Die Aldehyd-Oxidoreduktase von C.thermoaceticum und die von C. formicoaceticum(White, H., Feicht, R., Huber, C., Lottspeich, F. &Simon, H. (1991) Biol. Chem. Hoppe-Seyler 372,999-1005) zeigen qualitativ Ähnlichkeiten im Hin-blick auf Eisen, Schwefel, Wolfram und Pteringehaltsowie eine breite Substratspezifität. Jedoch zeigendie Enzyme auch überraschend deutliche Unter-schiede im Hinblick auf die Zusammensetzung unddie Aminosäuresequenz.

Key terms: Aldehyde oxidoreductase; carboxylic-acid reductase, tungsten containing; Clostridium thermoaceticum.

After observing the reduction of non-activated car-boxylic acids to the corresponding alcohols with Clos-tridium thermoaceticum and C. formicoaceticum atthe expense of carbon monoxide or formate[1'2] westudied the enzymology of this surprising reactioni3].It turned out, that in the above-mentioned acetogenicclostridia the reduction of the non-activated car-boxylic acids to the alcohols is catalysed by two en-zymes. One reduces the acid to the aldehyde and asecond the aldehyde to the alcohol.The reactionRCOOH + 2V® + 2H® ̂ RCHO + 2V2® + H2Ois reversible, therefore we call this carboxylic-acid re-ductase an aldehyde oxidoreductase. Pyridine nu-cleotides are not involved in this reaction. Whereasthe natural electron carriers are not known yet, theredox reaction can be conducted with various violo-gens. As already reported the enzyme containstungsten[3].The formate dehydrogenase from C. ther-moaceticum was the first enzyme for which the con-tent of tungsten was reported^. However, growth ofC. thermoaceticum on molybdate instead of tungstatedoes not drastically diminish the formate dehydro-genase activity^. In contrast molybdenum does notseem to be able to form a catalytically active aldehydeoxidoreductase in C. thermoaceticum^. Further-more, the aldehyde-oxidoreductase activity found inthe crude extract of C. thermoaceticum depends heav-ily on the sulphur source applied to the growthmedium[3i. In the meantime a third tungsten-contain-ing enzyme, an aldehyde ferredoxin oxidoreductase,occurring in the archaebacterium Pyrococcusfuriosus has been described. For this enzyme it wasnot shown whether it also acts as carboxylic-acid re-ductase^.Here we report on the subunit structure and prosthe-tic groups of the aldehyde oxidoreductase from C.

thermoaceticum with respect to content of tungsten,pterin, iron and sulphur. It turned out that two typesof complexes exist. One only contains aldehydeoxidoreductase, the other aldehyde oxidoreductaseand a flavin-containing viologen-accepting NADPHoxidoreductase (VAPOR). According to our studiesthe structural properties of the aldehyde oxidoreduc-tase described from C. thermoaceticum and that fromC. formicoaceticum^ are quite different.We have published a purification procedure for al-dehyde oxidoreductase of C. thermoaceticum startingfrom 6 g wet packed cells grown in the presence of185tungstatelJJ. The main goal of the former work wasto see whether enzyme activity and radioactivity, i.e.the content of tungsten, coincided in the various stepsof purification and after gel electrophoresis.This wasdefinitely the case. Now the purification has been im-proved, scaled-up and the various forms of the en-zyme further characterised.

Materials and Methods

Cell growthC. thermoaceticum DSM 521 was grown on a medium reported byLjungdahl and Andreesen^ with the following change and addi-tional supplements: 10 g NaHCO3, 0.1 g Na2S2O4 and 0.134 mgNiCla x 6 H2O were added per / growth medium.

Enzyme assays and other analytical proceduresThe terms aldehyde dehydrogenase and carboxylic-acid reductasedescribe the same enzyme, i.e. the aldehyde oxidoreductase. Its ac-tivity was tested in both ways as described131. The NH2CO-MV2®used for some of the assays was synthesized as described'91. TheVAPOR activity was measured in a total volume of 1.0 ml contain-ing in O.lMTris/HCl buffer pH 9.0, ImM NH2CO-MV2®, 1-10 jagprotein and 0.6mM NADPH.The test was started by the addition ofthe NADPH.

Brought to you by | University of California - San DiegoAuthenticated | 132.239.1.231

Download Date | 9/24/13 10:15 PM

Vol. 373 (1992) Tungsten-Containing Aldehyde Oxidoreductase 125

The determination of tungsten in the aldehyde Oxidoreductasewhich does not bind to the red-Sepharose was carried out accordingto Cardenas and Mortenson'10'. To 2.5 ml of buffer solution contain-ing 1.7 mg protein (19.8 nmol, based on a molecular mass of 86kDa) cone, sulphuric acid (0.714 m/) was added to a final concentra-tion of 4M. This solution was refluxed for 30 min. After adding 1 mlof 30% hydrogen peroxide and refluxing for another 10 min a col-ourless, clear solution resulted. Again cone, sulphuric acid wasadded to make the final concentration 4M, taking into account thevolume of the solution of 3 ml 0.2% toluene-3,4-dithiol in water,which was added later.The tungsten content of the enzyme form binding to red-Sepharosewas determined from 3.8 nmol based on a molecular mass of 300kDa. Other components were determined according to the indi-cated references: acid labile sulphur'11', iron'12', pterine-6-car-boxylic acid after oxidation with alkaline permanganate'13·14' andflavin'15'.Protein was determined according to Bradford using the modifica-tion of Read and Northcote'16'.

Enzvme purificationIf not mentioned otherwise all steps including Chromatographieseparations were carried out in an anaerobic chamber with an at-mosphere of 95:5 = N2:H2 in the presence of 2 pans of catalyst,which reduces traces of oxygen with the hydrogen gas present.Using a French press a crude extract was prepared from 40 g wetpacked cells of C. thermoaceticum resuspended in 80 ml basisbuffer containing 50mM Tris/HCl pH 7.0, 0.02% sodium azide,ImM DTE, 0.2mM NH2CO-MV2@/NH2CO-MV® and O.lmM orless dithionite. Aliquot fractions of 30 ml were heated 2 x 7 min at67 °C in centrifugation tubes. After the first 7-min period the mate-rial was briefly shaken. After 40 min centrifugation at 42000 χ gand 4 °C the combined supernatants were loaded on an octyl-Se-pharose column (45 x 120 mm) which was equilibrated with 10 vol-umes of basis buffer. After adding the protein 2 volumes of basisbuffer were passed through the column followed by 30% glycerol inbasis buffer. Using a flow rate of 11 m//h the eluate was directlyloaded on a hydroxyapatite Biogel HTP column (15 x 150 mm).The glycerol was washed out with basis buffer and the proteineluted with 50mM phosphate buffer pH 7.5 in basis buffer and di-rectly loaded onto a red-Sepharose column (15 x 95 mm) equilib-rated again with basis buffer. Continued elution with basis bufferled, without retardation, to the elution of 75-85% of the aldehydedehydrogenase activity. Afraction of 15-25% enzyme activity can-not be eluted by this procedure. It was eluted from the column in apeak of 15 ml if ImM NADPH in basis buffer was applied.Both fractions were pooled and called aldehyde Oxidoreductaseand aldehyde oxidoreductase/VAPOR complex or, based on theirlater determined composition as α,β- and α,β,γ-complex, respec-tively. These pools were chromatographed separately on variouscolumns.Both enzyme forms were chromatographed on a Mono Q HR 5/5column which was equilibrated with basis buffer containing neitherNH2CO-MV, dithionite nor DTE. The enzyme was desorbed by alinear KC1 gradient (0-0.5M) with a flow rate of 120 m//h.Besides various forms of electrophoresis, FPLC Superose 6 HR10/30 (10 x 300 mm) was used for checking the purity and for deter-mining the relative molecular mass of the preparations. For concen-trating the samples to about 4 mg protein/m/ Ultrafiltration wasapplied, using AmiconYM 10 or Millipore filters both with an ex-clusion limit of 10000. For the elimination of small molecular masscomponents also Econopac DG 10 columns (BIORAD) wereapplied.

Molecular mass determination(i) Gel filtration: The molecular masses of purified enzyme formswere determined by Superose 6 (HR 10/30) FPLC gel filtration.

The column was equilibrated with basis buffer and calibrated with10 proteins ranging in mol masses from 940 kDa (enoate reductase)to 12.4 kDa (cytochrome c). In this range the correlation of logmass versus retention volume was linear.(ii) Gel electrophoresis including Ferguson plots was conducted asdescribed below.

Electrophoretic proceduresAll electrophoretic procedures were performed with PharmaciaPhastSystem gels according to the instructions of the manufacturer.Electrophoresis with native gels and SDS gels was achieved withPhastSystem gradient gels 8-25% or 10-15% as well as with the20% homogeneous gel (Pharmacia) as described by the manufac-turer. For SDS-gel electrophoresis the same gels were used withSDS buffer strips. Molecular mass markers used were phos-phorylase (97000 Da), bovine serum albumin (67000 Da), ovalbu-min (43000 Da), carbonate dehydratase (30000 Da), soybean tryp-sin inhibitor (20100 Da) and a-lactalbumin (14400 Da). The gelswere usually stained with silver and in some cases with Serva BlueG. The aldehyde dehydrogenase and VAPOR activity were vis-ualised after running the electrophoresis under a permanentstream of argon through the electrophoresis chamber and transfer-ring the gel immediately afterwards into the anaerobic chamber.For activity staining of the aldehyde dehydrogenase activity theprocedure described in ref.13' was applied. The VAPOR activity onthe gels was demonstrated by incubating the gel as described for theenzymic assay. After development of the blue colour due to re-duced NH2CO-MVthe image was stabilised by transferring the gelto ImM neotetrazolium chloride. The gel was then washed in H2Oand air-dried. Isoelectrofocusing was performed with the Phast gelsIEF4.0-6.5 or 3-9 following the protocols recommended by themanufacturer.

Immunological techniques and immunoblottingThe immunisation scheme and other techniques were carried out aswill be described*.The rabbit was immunised with purified α,β,γ-complex.The electroblots were transferred with the Pharmacia Phast-Sys-tem Transfer Unit in a semi-dry blot onto nitrocellulose foil accord-ing to the instruction of the manufacturer.The immunological staining was performed as will be describedwith the microtiter plates* except that alkaline phosphatase con-jugated anti-rabbit IgG from goat diluted 1:7 500 was used as thesecond antibody.Staining was achieved by incubation of the blot in O.lM Tris/HCl pH9.5, 0.02% sodium azide, O.lM NaCl, 5mM MgCl2, 0.033% nitro-blue-tetrazolium chloride (dissolved previously 50 mg/m/ in water/dimethylformamide, 1:1) and 0.0165% 5-bromo-4-chloroindolyl-phosphate.The process was stopped with a solution of 5mM EDTAin 20mM Tris/HCl pH S.O.The blots were then air dried in the dark.

Results

When judging the following results one has to keep inmind that the aldehyde Oxidoreductase is extremelysensitive against oxygen. Furthermore, the here de-scribed enzyme complexes aggregate and decomposerather easily. The latter is especially true for the 300-kDa complex.From a practical point of view the test for aldehydedehydrogenase is simpler and more sensitive.Tested

* White, H. & Simon, H. (1991) to be published.



126 G. StrobI, R. Feicht, H. White, F. Lottspeich and H. Simon Vol. 373 (1992)

as described[3] the carboxylic-acid reductase activity isabout 5% of the aldehyde-dehydrogenase activity. Ifnot mentioned otherwise the activities are given forthe aldehyde-dehydrogenase reaction using butyral-dehyde. The purified enzyme has a specific car-boxylic-acid reductase activity of about 25 U/mg pro-tein with propionic acid as the substrate. With manyother carboxylic acids a specific activity of the sameorder of magnitude is observed (publication in prepa-ration) .

Separation and characterisation of two carboxylicacid reductase-containing complexes; one containing aflavoenzymeTwo forms of aldehyde oxidoreductase have been ob-served previously. One form binds to red-Sepharoseand showed a molecular mass 3-4 times higher thanthe form which does not bind to red-Sepharose[3].Thepurification procedure shown in Table 1 started with acrude extract derived from 40 g wet packed cells, issimpler than the one published beforei3], led to higherspecific activities and permitted reproducible separa-tion of the two enzyme forms. The combination ofchromatography on octyl-Sepharose from which thealdehyde oxidoreductase activity was eluted with30% glycerol, of hydroxyapatite for the eliminationof the glycerol and red-Sepharose led to a 88-fold en-richment of that oxidoreductase form which does notbind to the red-Sepharose. Further purification of thisso called, a, /3-complex on Mono-Q resulted in aproduct of 532 U/mg protein instead of 221 U/mgfound previously^. Repeated chromatography onMono Q did not increase the specific activity.The form of the enzyme activity which bound to red-Sepharose (called α,β,γ-complex) could be specifi-cally eluted as a sharp peak with NADPH (Fig. 1).The specific activity was 450 U/mg protein. Carefullyconducted chromatography on Superose 6 did not

lead to higher specific activities. In contrast to theα,/3-complex this material has not only aldehydeoxidoreductase activity but, surprisingly also cata-lyses the reaction:NADP® + H® + 2V®· ?± NADPH + 2V2®The specific activity for this reaction tested withNADPH and NH2CO-MV2® was 7.1 U/mg protein.NAD, NADH, 5'-AMP and 2'- or 3'-AMP were noteffective as eluents nor did NAD® react according tothe above equation. According to our studies in re-cent years there are many enzyme activities in

0 20 60Vol. [m/]

100

Fig. 1. Elution pattern of protein and enzyme activity fromred-Sepharose on which the fractions containing enzyme activ-ity from the hydroxyapatite column were loaded.The total added volume of 20 m/ 50mM phosphate buffer con-tained 25-30 mg protein. After eluting the aldehyde oxidore-ductase which does not bind, ImM NADPH was applied leadingto the elution of the aldehyde oxidoreductase/VAPOR-com-plex. Residual proteins were eluted with IM KC1.

Table 1. Isolation and separation of aldehyde oxidoreductase and aldehyde oxidoreductase/VAPOR-complex from 40 g wet packed cellsof C. ihermoaceticum.The enzyme was always tested as aldehyde dehydrogenase.

Step

Crude extract

Heat treatment

Octyl-Sepharose, hydroxyapatite andred-Sepharose-unbound fraction

Mono-Q of unbound fraction

Red-Sepharose-bound fraction,eluted with NADPH

Protein[mg]

3132

1514

21.4

6.1

2.25

Enzyme activity[U]

6578

7300

3929

3246

1020

Specific activity[U/mg]

2.1

4.8

184

532

453

Enrichment(«-fold)

-

2.3

88

253

216

Yield[%]

100

111

60

49

15




aerobes and especially in anaerobic organisms, whichcatalyse the reversible reduction of NAD or NADP atthe expense of reduced viologens1171. We call theseenzymes VAPORs for viologen accepting pyridinenucleotide oxidoreductases. We have observed highVAPOR activities in crude extract of C. ther-moaceticum, for NADP® (MV®') 5.5 U/mg protein,for NAD® (MV®') 4.6 U/mg protein[18] and forNADPH (NH2CO-MV2®) 4.0 U/mg protein. TheVAPOR activity purified with the α,β,γ-complexamounts to less than 1% of the activity originallyfound in crude extracts (unpublished results).

Characterisation of the two carboxylic acid reductaseformsThe various forms of the α,/3- and α,β,γ-complexes asfound by Chromatographie and electrophoretic proce-dures are summarised in Table 2.

0.20

0.15

0.05

I I I I8 12 16 20

Vol. [m/|

Fig. 2. Elution pattern of pure a, jo-form from a Superose 6 HR10/30 column loaded with 0.67 mg pure α,/3-form in 0.2 ml50mM phosphate.Flow 0.4 m//min of a buffer containing 50mMTris/HCl pH 7.0,200mM KC1,0.02% sodium azide. Peak 1 corresponds to about1600 kDa and peak 2 to about 90 kDa.

The a, -complexThe cx,j8-complex, after passage through the red-Sepharose, had an apparent molecular mass of about86 kDa and exhibited multimers up to 1600 kDa onthe gel column Superose 6 (Fig. 2). The original pro-tein from Mono Q and that from the peaks shown inFig. 2 showed the same pattern after electrophoresison a native PhastSystem gradient gel. A band ap-peared in the range of 420 kDa, a trace at 330 kDa, abroader band which sometimes seemed to be adouble band in the region 60-90 kDa and a band at 14kDa (Fig. 3).The latter could be visualised with silverstaining but not with Serva blue G (not shown).

Table 2. Properties of aldehyde oxidoreductase.The contents of Fe, S and prosthetic groups are given in mol/as-sumed complex composition.

Molecular mass (kDa)Subunits (kDa)Assumed compositionFeSWPterineFADΛ42ο^Μ~' x cm"1)Λ42ο after reduction with

butyraldehydeArjg,4max(X = 265nm)Isoelectric point

Not binding Binding

to red-Sepharose

8664; 14

<*,β25-3321-29

0.80.8-1.1< 0.002

8838

258-

4.6-4.7

30064; 14; 43

«3/3.3775-9050-58

3.42.51.714894

620763

4.6-4.7

With the exception of the 14-kDa band, all theseforms were enzymically active as aldehyde dehydro-genases after electrophoresis (not shown). This indi-cates that the α,/3-complex is able to aggregate tooligomers but also that a small subunit of an apparentmass of 14 kDa can easily be split off.Determination of the molecular masses using the Fer-guson plot techniques revealed a molecular mass of420 and 85-110 kDa, respectively. This is in reason-able agreement with the determinations obtained bygel electrophoresis using the different migration dis-tances. The point of intersection of the straight linesof the Ferguson plots for the 100-kDa and 420-kDaproteins are clearly indicative for an oligomer re-lationship of these two proteins[19]. This is further-more in agreement with SDS gel electrophoresis ofthe proteins of peaks 1 and 2 of Fig. 2. Both are com-posed of two subunits of 64 and 14 kDa, respectively.Fig. 4 shows the subunits of 64 and 14 kDa after SDSgel electrophoresis and silver or immunostaining,which confirmed the observation from the native gel.From these observations it can be concluded that theenzyme activity which does not bind to red-Sepharoseforms oligomers. Under the condition of chromatog-raphy on Superose 6 HR these oligomers may consistof more than 15 α,/8-dimers with an apparent molecu-lar mass around 80 kDa each and under conditionsapplied in native gel electrophoresis 4-5 such dimersappear to be aggregated forming a complex of about330 and 420 kDa.Whether one or two of the 14-kDa subunits are pre-sent in the protein showing a molecular mass in the re-gion from 60-90 kDa cannot be said with certainty.




kDa420

330

160

60-90 I

FPLC on a Mono Q column and electrophoresis ofthe fractions. Fig. 5 shows the surprising but typicalelution pattern of the complex from a Mono Q col-umn after applying a gradient of KC1. The peaks ap-peared at 105, 160, 190 and 290mM KC1. The firstthree of these peaks showed specific aldehyde-dehyd-rogenase activities from 150 to 365 U/mg protein.That of the peak eluting at 290mM KC1 was higherwith 440 U/mg protein. The purified complex elutingat 105mM KC1 from the Mono Q column showed in anative PhastSystem gel a single band of molecularmass of about 160 kDa (Fig. 3, lane B:). The formseluting from Mono Q at 160mM and 190mM KC1showed the same band in native gels (not shown).Theform of the enzyme complex eluting at high KC1 con-centrations after chromatography on Mono Qshowed three bands (Fig. 3, lane C:). In addition to

B

Fig. 3. PhastSystem native 8-25% gradient gel of purified al-dehyde oxidoreductase.Lane A: silver stain of the aldehyde oxidoreductase spread overan area from 60-90 kDa and around 14 kDa. Traces of the 330-kDa and 420-kDa forms are also visible. Lane B: aldehyde oxi-doreductase/VAPOR complex eluting from the Mono Q col-umn at 105mM KC1, with a single band after silver staining at160 kDa; lane C: aldehyde oxidoreductase/VAPOR complexafter the same purification procedure eluting at high KC1 con-centration with three bands at 160 kDa, 330 kDa and 420 kDaafter silver staining.

As already mentioned sometimes a double band in na-tive gel electrophoresis can be seen. The two bandscould be an α,β- and an α,/32-ίοπη.

The a, ,y-complex showing aldehyde oxidoreductaseand VAPOR activityThis complex bound to red-Sepharose and could beobtained in pure form by elution with NADPH, sinceby further Chromatographie steps the specific enzymeactivity could not be increased. The complex showed,according to gel chromatography on Superose 6 HR,a labile species with an apparent molecular mass of300 kDa probably composed of a&y. The presence oftwo γ-subunits cannot be excluded. This complexdoes disintegrate into sub-complexes as shown by

B

Fig. 4. Phastsystem SDS gel of purified aldehyde oxidoreduc-tase.Lane A: silver stain of purified aldehyde oxidoreductase withthe 64-kDa and 14-kDa subunits; lane B: immuno stain; lane C:aldehyde oxidoreductase/VAPOR complex after elution fromMono Q (representative of all fractions), silver-stained with 64-kDa, 43-kDa and 14-kDa subunits; lane D: immunostain of thesame protein.




0.32 -

0.16 -

Fig. 5. Elution pattern of pure a,/3,y-form from a Mono Qcolumn.The enzyme elutes at four different KC1 concentrations: 105,160,190 and 290mM KCl.The specific aldehyde dehydrogenaseactivities were 150, 258, 365 and 442 U/mg protein, respec-tively. Elution was carried out in 50mM Tris/HCl pH 7.0 by alinear KC1 gradient (0-0.SM) with a flow rate of 120 m//h.

the 160-kDa species there was another at 330 kDa anda third at 420 kDa. The latter two were identical withthe bands observed upon native gel electrophoresis ofthe ίχ,/3-complex and should therefore be dissociationproducts of the a,/3,y-complex, i.e. a form of the en-zyme where the γ-subunit was lost.This rather compli-cated behaviour was simpler when the fractions fromthe Mono Q column were analysed by SDS gel elec-trophoresis. Silver staining showed, for materialapplied directly after NADPH elution from red-Sepharose or eluted from the Mono Q column at KC1concentrations of 105, 190 and 290mM KC1, bandscorresponding to 64, 43 and 14 kDa, respectively,(Fig. 4).The behaviour of the a,/3/y-complex on the Mono Qseems to depend on its redox status and on the time ofstorage under anaerobic conditions. If 12mM dithion-ite was added to an enzyme solution obtained fromred-Sepharose the enzyme hardly bound to Mono Q.This was also observed when the solution whichpassed through the Mono Q column was diluted 5-fold and re-chromatographed on the Mono Q col-umn. After anaerobic storage for 8 days at 4 °C therewas an overall decrease of 44% in the enzyme activity.The share of the enzyme eluting from Mono Q at190mM KC1 decreased from 70% to 50%. The shareof the enzyme eluting at 105 and 160mM increasedfrom 3 to 5 % and 7 to 17 %, respectively (not shown).

UVIVis-spectmThe spectra of the α,β- and a,/3,y-complex are shownin Fig. 6a and 6b. The α,/3-complex showed a

maximum near 400 nm which decreased upon the ad-dition of butyraldehyde. Also the addition of dithion-ite caused a decrease in absorption. The same wastrue for the addition of oxygen (not shown). Additionof 180mM propionate resulted in no change of the ab-

100

75

50

25

300 400 500λ[ηιτι]

400 500X[nm]

600

Fig. 6. Spectra of the α,β- and the α,β,γ-complex without andwith aldehyde addition at 22 °C.a) Pathlength 0.2 cm, 0.8 mg protein of the α,β-complex (spe-cific activity 425 U/mg) in 0.4 m/290mM KC1 containing 50mMTris/HCl pH 7.0 andO.02% azide. —0; 0.1; 0.2 and

0.8mM butyraldehyde. b) Pathlength 1.0 cm, 0.48 mg pro-tein of the a,/3,y-complex in 0.4 ml 50mMTris/HCl pH 7.0 con-taining 0.02% azide.—0; 0.06 and 0.3mM butyralde-hyde.




sorbance. The a,/3,y-complex showed an increase ofabsorption upon addition of butyraldehyde at around355 nm. Based on a mM extinction coefficient (ε) of10-12 for a flavin around 372 and 450 nm the con-tribution of the flavin cannot be higher than 0.01 ab-sorption units.

Amino-acid composition and the amino-acidsequence of the Ν-terminus of the a-subunit of theOL, β- and the a, ,y-complexIn the routine procedure for amino-acid analysispeaks appeared which could not be unambiguouslyidentified. There are peaks which may correspond tocysteic acid and methionine 5-oxide.This is surprisingbecause the enzyme samples were never treated withoxidative agents and oxygen was excluded duringsample preparation. Only clearly identified aminoacids are reported. Table 3 shows the amino-acid com-position of the α,/3-complex which does not bind tored-Sepharose and the protein from the α,β,γ-com-plex which elutes from Mono Q with 290mM KC1.From gel electrophoresis it was assumed that this isidentical with the α,/3-complex split off from theα,β,γ-complex.The close similarity of the amino-acidcomposition is in agreement with this assumption, i.e.this enzyme species does not contain the γ-subunit.The partial amino-acid sequence of the 64-kDa sub-unit of the α,β- and the o:,/3/x-complex was identicalfor the first 15 positions starting at the N-terminal.This is further proof that the two proteins are just dif-ferent complexes containing the α-subunit. The par-tial amino-acid sequence of the 43-kDa subunit of theα,β,γ-complex was completely different. The /3-sub-unit has not been sequenced yet.64-kDa (= a-) subunit of the α,/3-complex:Met-Tyr-?Gly-Trp-Thr-Gly-Gln-Leu-Leu-Arg-Val-?Asn-Leu--?Ser-?Asn;64-kDa subunit of the c^ sy-complex:Met-Tyr-Gly-Trp-Thr-Gly-Gln-Leu-Leu-Arg-Val-Asn-Leu--Ser-?Asn;

43-kDa (= γ-) subunit of the o^y-complex:Met-Arg-Tyr-Leu-Ile-Ile-Gly-Asn-Ser-Ala-Ala-Gly-Val-Ala.

Table 3. Amino-acid composition of the α,/3-complex after finalpurification on Mono Q (A) and of the protein species of the α,β,γ-complex eluting from Mono Q at 290mM KC1, which does not con-tain the γ-subunit (B).

Amino acid

AsxThrSerGlxProGlyAlaValHeLeuTyrPheHisLysArg

Residu(A)

10560489065

1181357448

1043134515764

es/mol(B)

9859448564

1261367043

1013033425661

for the α,β- and α,β,γ-complexes, respectively.Molybdenum could not be detected.No flavin could be found in the α,/3-complex whereasthe 300-kDa complex probably contained twomolecules of FAD which should be part of the γ-sub-unit.The higher ε at 420 nm of the 300-kDa complex com-pared with that of 86 kDa is in agreement with thecontent of flavin in the former.During the isolation procedure both forms of the en-zyme were always kept under reducing conditions.Therefore, it was surprising that both showed drasti-cally reduced ε values at 420 nm after adding butyral-dehyde which is dehydrogenated and delivers elec-trons to the enzymes. This decrease must probably bedue to a change of absorption in the a- and/or /3-sub-unit since the decrease of 6420 was almost the same forboth enzyme complexes.Both enzyme forms have the same isoelectric point(Table 2).

Prosthetic groups and other properties of the α, β- andα, β, y-complexWhereas the content of iron in the α,/3-complex isonly slightly higher than the content of acid-labile sul-phur there is about 1.5 times more iron than sulphurin the «s&y-complex (Table 2). In the α,/3-complex0.8 and in the α,β,γ-complex 3.4 mol of tungstenwere found.The amounts of pterine-6-carboxylic acidobtained after alkaline permanganate oxidation ofthe two complexes turned out to be 0.8-1.1 and 2.5

ImmunoblottingThe antibodies used for all these experiments wereproduced by injecting a rabbit with purified a, ,y~complex.Purified aldehyde oxidoreductase was subjected tonative and SDS-gel electrophoresis and then silver-stained, electro-blotted and immuno-stained. As canbe seen in Fig. 4 there was a 64- and a 43-kDa compo-nent in the purified aldehyde oxidoreductase/VAPOR complex of 300 kDa, which could be im-




muno-stained after SDS-gel electrophoresis. The 14-kDa subunit did not stain. Silver and immuno stainsof purified α,β-complex after SDS-gel electro-phoresis are also shown in Fig. 4. Comparing lanes Aand B, the silver and immuno stain respectively,shows again that only the large subunit of 64 kDa wasimmunologically active, the small subunit of 14 kDadid not stain in the immunoblot.

Discussion

The aldehyde oxidoreductase of C. thermoaceticumhas important properties in common with the corre-sponding enzyme from C. formicoaceticum describedin refJ7'. Both enzymes contain tungsten, iron andacid-labile sulphur. Alkaline permanganate oxidationleads to pterine-6-carboxylic acid. The reduction of acarboxylic acid at the expense of a reduced viologenproceeds to the aldehyde only. There are some simi-larities with respect to substrate specificity and Kmvalues. But there are surprising differences in thestructure of both enzymes.The aldehyde oxidoreduc-tase from C. formicoaceticum is a dimer of identicalsubunits.The enzyme from C. thermoaceticum occurs in formof a dimer α,β which is able to polymerise to com-plexes up to 1600 kDa.The α-subunit seems to have amolecular mass of 64 kDa and the -subunit of 14kDa. This subunit could be a ferredoxin. The al-dehyde dehydrogenase isolated from Pyrococcusfuriosus uses a ferredoxin as a cofactor^. FerredoxinI from C. thermoaceticum shows an apparent molec-ular mass of 7.5 kDa in gel filtration conducted in thepresence of O.lM NaCl. In the absence of NaCl or inSDS-gel electrophoresis it shows an abnormally highapparent molecular mass of 12.5-17.5 kDa[20].It is not clear yet how the tungsten, iron and sulphurare distributed in the α,/3-complex. In the α,β,γ-com-plex occurs a subunit containing FAD. Obviously thisis responsible for the capability of the complex to bindspecifically to red-Sepharose and to interact withN AD PH. The complex occurs in various forms. Oneeluted from the Mono Q column at a KC1 concentra-tion of 190mM according to gel filtration reveals an ap-parent molecular mass of 300 kDa. According toSDS-gel electrophoresis it contains subunits of 64,43and 14 kDa. As can be seen from the amino-acid com-position (Table 3) and the amino-acid sequence of thefirst 15 amino acids the 64-kDa subunit of the trimerseems to be identical to that of the α,/3-complex.Therefore, the 43-kDa subunit should contain the fla-vin. The composition of the complex could be 0:3/337.However, other ratios of the three subunits are possi-ble, too. All the different forms of the o:,/3,7-complex

which can be seen by chromatography on Mono Q re-sult after SDS-gel electrophoresis in the same pat-tern. In native gels the 160 kDa species shows theVAPOR reaction (not shown). It is also the majorform occurring, according to these gels. Two otherforms, 330 kDa and 420 kDa appear in native gel elec-trophoresis of the α,β,γ-complex, as well as of theα,/3-complex.The question of the physiological role of the enzymeremains unsolved. Recently Mukund and Adams^reported on a tungsten-containing iron-sulphur al-dehyde ferredoxin oxidoreductase which seems toplay a role in a new pyroglycolytic pathway by oxidiz-ing glyceraldehyde to glycerate. We tested glyceral-dehyde and glycerate with purified aldehyde oxido-reductase from C. thermoaceticum and found lessthan 1% of the activity with butanal and propionateas the substrates, respectively (data not shown). Ittherefore seems unlikely that the C. thermoaceticumenzyme oxidises glyceraldehyde in vivo. Interestinglyamong the first six amino acids of the α-subunit of theC. thermoaceticum enzyme four, including the N-ter-minal methionine, are identical with those reportedfor the enzyme from P. furiosus^.The unusually highε values of the enzyme can at least partially beexplained by the also unusually high iron-sulphur con-tent to which the absorption of the pterine system hasto be added (Table 2). The absorption of a tungstencontaining pterine cofactor is not known yet butshould be similar to the molybdenum cofactor. Forthe latter a surprising 3-fold greater absorption thanexpected has been observed121 \ The aldehyde ferredo-xin oxidoreductase reported by Mukund and Adams[6]

contains 6 iron and 4 acid-labile sulphur atoms andshows a molar absorption coefficient ε425 = Γ7600Μ'1cm'1 in the oxidised form'22·'. The aldehyde oxidore-ductase from C. thermoaceticum contains 4-5 timesthe amount of iron and sulphur with about the samemolecular mass.

The natural electron mediator of the aldehydeoxidoreductase remains another still unsolved prob-lem. Mukund and Adams suggest that ferredoxin ac-cepts the electrons from the tungsten-iron-sulphurprotein of Pyrococcus furiosus^. There are two fer-redoxins reported for C. thermoaceticum^20'23^ Weproceeded to purify both as described and testedthem for enzymic activity with the aldehydeoxidoreductase. None was found to be active (Heinz,unpublished). As already mentioned, ferredoxin Ishows an apparent molecular mass of 12500-17500 inthe absence of NaCl in gel filtration or in SDS-gelelectrophoresis[20]. This fact taken with the high iron-sulphur content of the o:,/3-complex and the unusualstaining behaviour of the 14-kDa subunit makes it




seem possible, that this subunit could be a type of fer-redoxin.This will be a point of interest in future work.

We are especially grateful to Prof. Dr. D. Oesterhelt for the op-portunity to prepare the polyclonal antiserum used in this studyas well as Dr. H. Wieser for amino-acid analysis and discussion.We thank C. Huber for the W and Mo determinations andU. Heinz for the preparation of the ferredoxins.This work wasfinancially supported by Deutsche Forschungsgemeinschaft(SFB 145) and Fonds der Chemischen Industrie.

References

1 Simon, H.,White, H.,Lebertz,H.&Thanos,J.( 1987) -gew. Chem. 99, 7S5-7S7;Angew. Chem. Int. Ed. Engl. 26,785-787.

2 Fraisse, L. & Simon, H. (1988),4rc&. Microbiol. 150,381-386.

3 White, H., Strobl, G., Feicht, R. & Simon, H. (1989) Eur.J. Biochem. 184,89-96.

4 Yamamoto, I., Saiki,T., Liu, S.-M. & Ljungdahl, L.G.(1983)7. Biol. Chem. 258,1826-1832.

5 Andreesen, J.R. & Ljungdahl, L.G. (1973) J. Bacteriol.116, 867-873.

6 Mukund, S. & Adams, M.W.W. (1991) J. Biol. Chem. 266,14208-14216.

7 White, H., Feicht, R., Huber, C., Lottspeich, F. & Simon,H. (1991) Biol. Chem. Hoppe-Seyler3T2,999-1005.

8 Ljungdahl, L.G. & Andreesen, J.R. (1978) Methods Enzy-mol. 53, 360-372.

9 Günther, H., Neumann, S. & Simon, H. (1987) /. Biotech-nol. 5, 53-65.

10 Cardenas, J. & Mortenson, L.E. (1974) Anal. Biochem.60, 372-381.

11 Chen, J.-S. & Mortenson, L.E. (1971) Anal. Biochem. 79,157-165.

12 Crumby, P.E. & Massey, V. (1967) Methods Enzymol. 10,463-474. (Estabrook, R.W. & Pullmann, M.E., eds.)Acad. Press, New York, London.

13 Forrest, H.S. & Mitchell, H.K. (1954) J. Am. Chem. Soc.76, 5658-5662.

14 Johnson, J.L. & Rajagopalan, K.V. (1982) Proc. Natl.Acad. Sei. US 79, 6856-6860.

15 Spencer, R., Fisher, J. & Walsh, C. (1976) Biochemistry 15,1043-1053.

16 Read, S.M. & Northcote, D.H. (1981) Anal. Biochem.116,53-64.

17 Bader, J., Günther, H., Nagata, S., Schütz, HJ., Link,M.L. & Simon, H. (1984)7. Biotechnol. 1, 95-100.

18 Lebertz, H., White, H., Beer, A., Günther, H. & Simon,H. (1988) DECHEMA Biotechnol. Conf. 1,369-374.

19 Thorun,W. (1971) Z. Klin. Chem. Klin. Biochem. 9,3-13.20 Yang, S.-S., Ljungdahl, L.G. & LeGall, J. (1977) /. Bacte-

riol. 130,1084-1090.21 Gardlik, S. & Rajagopalan, K.V. (1991) /. Biol. Chem.

266,4889-4895.22 Mukund, S. & Adams, M.W.W. (1990) /. Biol. Chem. 265,

2311508-11516.Elliott, J.I. & Ljung328-333.

iahl, L.G. (1982) /. Bacteriol. 151,

Dr. Gerhard Strobl, -RAD Laboratory GmbH,Heidemannstr. 164, W-8000 München 45;Richard Feicht, Dr. Hiltrud White, Prof. Dr. Helmut Simon, Lehrstuhl für Organische Chemie und Biochemie,Technische Universität München.Lichtenbergstr. 4, W-8046 Garching;Prof. Dr. Friedrich Lottspeich, Max-Planck-Institut für Biochemie,Am Klopferspitz, W-8033 Martinsried.



Documents

The Tungsten-Containing Aldehyde Oxidoreductase from Clostridium thermoaceticum and its Complex with a Viologen-Accepting NADPH Oxidoreductase