QUINOPROTEIJM GLUCOSE
DEHYDROGENASE AND
THE b -TYPE GTTOGHROMES
OF ACINETOBACTER
CAZJCOACETICUS
L.M.D. 79.41
Pmjl Dokter
TRdissy
QUINOPROTEIN GLUCOSE DEHYDROGENASE AND THE b-TYPE CYTOCHROMES OF ACINETOBACTER CALCOACETICUS L.M.D. 79.41
TR diss
1585
ACINETOBACTBH CALCOACETICUS L.N.D. 79.41
PROEFSCHRIFT
ter verkrijging van de graad van doctor aan de
Technische Universiteit Delft, op gezag van de Rector Magnificus, prof. dr. J. M. Dirken, in het openbaar te verdedigen ten overstaan
van een commissie aangewezen door het College van Dekanen op maandag 16 november 1987 te 14.00 uur
door
Paul Dokter
Dit proefschrift is goedgekeurd door de promotor
Prof. dr. ir. J.A. Duine
This study was carried out at the Department of Microbiology and Enzymology
of the University of Technology Delft, The Netherlands, and subsidized by
Met dank aan allen die aan de totstandkoming hebben meegewerkt binnen en buiten de Technische Universiteit Delft, in het bijzonder mijn promoter Prof. dr. ir. J. A. Duine en dr. J. E. van Wielink.
Honeywell Nederland b.v. heeft financieel bijgedragen in de drukkosten.
Publications
- Dokter, P., Frank, J.Jzn. and Duine, J.A. (1985) Quinoprotein glucose dehydrogenase from Acinetobacter calcoaceticus. Antonie van Leeuwenhoek J. Microbiol. 51: 444
- ^Dokter, P., Van Kleef, M.A.G., Frank, J.Jzn. and Duine, J.A. (1985) Production of quinoprotein glucose dehydrogenase in the culture medium of Acinetobacter calcoaceticus■ Enzyme Microbiol. Technol. 7: 613-617
2)
- Dokter, P., Frank, J.Jzn. and Duine, J.A. (1986) Purification and characterization of quinoprotein glucose dehydrogenase in Acinetobacter calcoaceticus LMD 79.41. Biochem. J. 239: 163-167
3)
- 'Dokter, P., Pronk, J.T., Van Schie, B.J., Van Dijken, J.P. and Duine, J.A. (1987) Jji vivo and in vitro substrate specificity of the quinoprotein glucose dehydrogenase in Acinetobacter calcoaceticus LMD 79.41. FEMS Microbiol. Lett. 43: 195-200
- 'Dokter, P., Van Wielink, J.E., Van Kleef, M.A.G. and Duine, J.A. (1987) Cytochrome b562 from Acinetobacter calcoaceticus LMD 79.41: Its characteris tics and role as electron acceptor for quinoprotein glucose dehydrogenase. Biochem. J. (submitted for publication)
A.H. and Duine, J.A. (1987) Characterization of the membrane-bound
cytochrome b-containing complexes of Acinetobacter calcoaceticus LMD 79.41. J. Gen Microbiol. (submitted for publication)
- De Bont, J.A.M., Dokter, P., Van Schie, B.J., Van Dijken, J.P., Frank, J.Jzn. and Kuenen, J.G. (1984) Role of quinoprotein glucose dehydrogenase in gluconic acid production by Acinetobacter calcoaceticus■ Antonie van
Leeuwenhoek J. Microbiol. 50: 76-77
- Hommes, R.W.J., Postma, P.W., Neijssel, O.M., Tempest, D.W., Dokter, P and Duine, J.A. (1984) Evidence of quinoprotein glucose dehydrogenase apoen-zyme in several strains of Escherichia coli. FEMS Microbiol. Lett. 24: 329-333
- Van Kleef, M.A.G., Dokter, P., Mulder, A.C. and Duine, J.A. (1987) Detection of the cofactor pyrroloquinoline quinone. Anal. Biochem. 162: 143-149
The publications designated by , , }, ' and have been inserted in
STELLINGEN 1
Bij het vinden van "zeer" lage enzym activiteiten in celvrij extract, dient men zich af te vragen of deze activiteiten nog wel fysiologische betekenis hebben voor het organisme.
-Jirausch, M., Asperger, 0., and Kleber, H.P. (1986) J. Basic Microbiol. 26, 351-357
2
De groei van micro-organismen op bepaalde substraten is sterk afhankelijk van de aangelegde condities.
-Bauman, P.(1987) J. Bact. 96, 39-42.
-Juni, E. (1978) Ann. Rev. Microbiol. 32, 349-371. 3
Het postuleren van het bestaan van PQQ synthese in Eschrichia coll, gesteund door. experimentele resultaten, is
waarschijnlijk toe te schrijven aan contaminatie met PQQ van de gebruikte testmedia.
-Ameyama, M., Shinagawa, E., Matsushita, K., and Ada-chi, 0. (1984) Agric. Biol. Chem. 48, 3099-3107. -Hommes, R.W.J., Postma, P.W., Nijssel, O.M., Tempest,
P.W., Dokter, P., and Duine, J.A. (1984) FEMS Micro biol. Lett. 24, 329-333.
-Van Kleef, M.A.C., Dokter, P., Mulder, A.C., and Dui ne, J.A. (1987) Anal. Biochem. 162, 143-149.
4
De K's voor glucose van Acinetobaccer calcoaceticus lijkt te
hoog om een fysiologisch ecologische betekenis aan glucosede-hydrogenase te geven.
-Dit proefschrift. 5
De bewering dat groei van ffiuconobacter oxidans in een com
plex gistextract-glucose medium, glucose gelimiteerd is, wordt niet ondersteund door experimentele resultaten.
-Olijve, W., and Kok, J.J. (1987) Arch. Microbiol. 121, 291-297.
6
De suggestie, dat membraangebonden glucosedehydrogenase in Pseudomonas fluorescens Xn vivo electronen doneert aan
ubi-chinon-9, in de electronentransportketen, op grond van de in vitro gemeten reductiesnelheid van ubichinon-9, getuigt van
moed.
-Matsushita, K., Ohno, Y., Shinagawa, E., Adachiro, 0., and Ameyama, M. (1982) Agric. Biol. Chem. 46,
cus zijn onlangs door onderzoek in Leiden binnen het kader Biotechnologie Delft Lelden (BDL) bekrachtigd.
-Dokter, P., Frank, J. Jzn., and Duine, J.A. (1986) Biochem, J. 239, 163-167.
-Geiger, 0., and Görisch, H. (1986) Biochemistry 25, 6043-6048.
-Matsushita, K., Shinagawa, E., Inoue, T., Adachl, 0., and Ameyama, M. (1986) FEMS Microbiol. Lett. 37, 141-144.
-Cleton-Jansen, A., Goosen, N., Wenzel, T.J., and Van de Putte, P. (1987) J. Gen. Microbiol. (submitted for publication)
8
Het invoeren van een Willie Wortel premie voor universitaire onderzoekers in het geval van verkoop van octrp.o.lenf die op hun onderzoek gebaseerd zijn, zal bevorderend werken op de productiviteit van deze onderzoekers.
-Van Dijken, J.P. (1987) Delta, 30, 8 9
Soms is de wetenschappelijke inhoud van een artikel zo aan sprekend, dat het wordt geaccepteerd voordat het ingezonden
is. " -Dokter, P., Pronk, J.T., Van Schie, B.J., Van Dijken,
J.P., and Duine, J.A. (1987) FEMS Microbiol. Lett. 43, 195-200. 7
10
Het gebruik van de naam obstetrlcus, gevoerd door vele gynae cologen, kreeg in het laatste conflict tussen de medische specialisten en de overheid van mei 1987 een speciale beteke nis.
11
Het zou promovendi verboden moeten worden om, voor het vol trekken van hun promotie, een andere werkkring te aanvaarden (of welke relatie dan ook aan te gaan).
12
Voor de relatie tussen promovendi en promotoren dient de man tel der liefde ruim gesneden te zijn.
Contents
Chapter page
I General introduction 11
II Production of quinoprotein D-glucose dehydrogenase in 23 the culture medium of Acinetobacter calcoaceticus
III Purification and characterization of the quinoprotein 29 glucose dehydrogenase from Acinetobacter calcoaceticus
L.M.D. 79.41
IV The in vivo and in vitro substrate specificity of 37 quinoprotein glucose dehydrogenase of Acinetobacter
calcoaceticus L.M.D. 79.41
V Cytochrome b_562 from Acinetobacter calcoacet icus 45 L.M.D. 79.41: Its characteristics and role as electron
acceptor for quinoprotein glucose dehydrogenase
VI Characterization of the membrane-bound cytochrome b- 75 containing complexes of Acinetobacter calcoaceticus
L.M.D. 79.41
VII Summary 113 Samenvatting
Chapter I
General introduction
The organism:
Acinetobacter calcoaceticue is a Gram-negative, strictly aerobic bacterium. It is oxidase-negative in the sense that it lacks cytochrome c and an aa„-type oxidase. Generally Acinetobacters are rods, 0.9 - 1.6 jjm in diameter and 1.5 - 2.5 pm in length, which become spherical in the exponen tial growth phase. All strains are able to grow at 20-30 C. For most strains the optimum temperature for growth is 33-37 C. The natural habitat of A. calcoaceticus is soil and water, where it seems to be important in the degradation of organic materials (Baumann et al., 1968). Acinetobacter spp. are in general nonpathogenic, but can cause infections in very weak patients (Hosenthal, 1978; Hoffmann et al:, 1982).
The organism was isolated for the first time by Beijerinck in 1911 (Beijerinck, 1911). It was named Micrococcus calcoaceticus. since the cells accumulated as small sphericals in a medium containing calcium acetate as carbon and energy source. Formerly also the names Bacterium anitratum and Herella vaginicola were used. Nowadays the bacterium is called Acinetobacter calcoaceticus (Acinetobacter from akinetos = unable to move and bactrum = a rod).
Incomplete oxidation
Most aerobic microorganisms oxidize their organic energy-source com pletely to carbon dioxide and water via their respiratory metabolism. For instance A. calcoaceticus utilizes a variety of substrates in this way, e.g. acetate, ethanol and hexadecane (Schlegel, 1986). The conversion of acetate
12
-proceeds via the citric acid cycle in combination with the respiratory
chain. The use of ethanol and hexadecane proceeds essentially in the same
way, since ethanol and hexadecane are decompozed to acetate (Schlegel,
1986).
The oxidative process in which partially oxidized substrates accumulate
in the culture medium is referred to as 'incomplete oxidation'. A typical
example is the conversion of glucose into gluconic acid, manifest in
Gram-negative bacteria like A. calcoaceticus and Pseudomonas aeruginosa under
certain conditions of growth. A. calcoaceticus is exceptional since in spite
of its ability to catalize the conversion of glucose into gluconic acid it
can neither grow on glucose nor on gluconate. Therefor, incomplete oxidation
of glucose in this organism is an attractive model system for studying the
bioenergetic consequences of incomplete oxidations since there are no inter
fering phosphorylative routes of glucose degradation.
Energy transduction in substrate oxidation
The overall reaction catalyzed by the enzymes of the citric acid cycle is
given by: CH-COO- + NAD + FAD •♦ 2C0_ + NADH + FADH„. The overall reaction
for electron transport is: NADH + FADH- + 0„ ■» NAD+ + FAD + 2 ^ 0 . Electron
transport can result in generation of ATP (used to drive endergonic
reactions) or in accumulation of solutes like amino acids, sugars etc..
Electron transport from NADH and FADH_ to oxygen is catalyzed by a system
of redox carriers incorporated in the cytoplaamic membrane, the socalled
electron transport chain. ATP synthesis and active transport of solutes
proceed via a local or a bulk H -gradient over the cytoplasmic membrane,
generated by electron transport through the electron transport chain (see
Nicholls, 1982). For Escherichia coli (which has a similar electron
transport chain as A. calcoaceticus. see below), it has generally been
NADH dehydrogenase, leads to the translocation of 2 protons over the
cytoplasmic membrane. Similarly, the reaction QH„ + l/202 ■» Q + H„0,
catalyzed by the cytochrome o- or cytochrome d-containing oxidase, leads to
translocation of 2 protons. (The reaction FADH„ + Q ■♦ FAD + QH„, mediated by
succinate dehydrogenase does not lead to the translocation of protons.)
Proton translocation may proceed by a loop mechanism of alternating an
[H]-carrier (H + e ) and an e -[H]-carrier. For NADH oxidation the first loop is
localized in the NADH dehydrogenase with a flavin cofactor as [H]-carrier
and an iron-sulphur centre as e -carrier. Ubiquinone is in E. coli put
forward as the [H]-carrier of the second loop (also active in succinate
oxidation), the redox groups of the oxidases (cytochromes) as the e
-carriers (for refs. see Nicholls (1982) and Anraku and Gennis (1987)).
Assuming a H /ATP stoichiometry of 3 (3 "translocated" protons needed for
the synthesis of 1 ATP from ADP and anorganic phosphate, see Kashket
(1982)), the two reduction'equivalents going from NADH to oxygen produce 4/3
molecules of ATP (succinate oxidation 2/3 ATP).
Energy transduction in glucose/gluconolacton conversion
Formally, oxidation of 1 glucose produces 1 gluconolactone and 2 reduc
tion equivalents, which could provide useful energy for the organism.
Assuming the entry of electrons occurs at the level of ubiquinone the charge
separation over the cytoplasmic membrane should result in 2/3 ATP.
Oxidation of glucose by A. calcoaceticua provides indeed energy for the
organism since this step can drive solute transport (Van Schie et al., 1985;
Kitagawa et al., 1986; Van Schie, 1987), the production of ATP (Kitagawa et
al., 1986; Van Schie, 1987) and an increase in the molar growth yield (De
Bont et al., 1984; Muller and Babel, 1986). The increase in yield, however,
is higher than expected on account of the 2/3 ATP synthesized per glucose
14
-results that 1 glucose oxidized gave rise to 2 ATP. It was suggested that
the auxiliary energy source glucose, is able to improve the efficiency of
energy generation from the carbon and energy source, acetate. Another ex
planation could be that the cytochrome o- and cytochrome d-containing
oxidases not only take care of the translocation of the two electrons from
the outside to the inside of the cytoplasmic membrane, but also translocate
protons from the inside to the outside. Indications for a Q- or b-cycle
mechanism, sustaining this idea, can indeed be derived from the complex
behaviour of ubiquinone in electron transport (E. coli: Downie and Cox,
1978; Sanchez Crispin et al., 1979; Proteus mirabilis: Van Wielink et al.,
1983, Van Wiel ink, 1986). The measured overall H /0 stoichiometry of 4 may
also be an underestimation (Cox and Haddock, 1978). An early report made
already mention of a H /2e -ratio of 6 (Meyer and Jones, 1973). Moreover,
with cells of Klebsiella aerogenes, another bacterium with a similar
electron transport chain, a value of nearly 6 has been measured (Drozd et
al.. 1984). Also growth yield experiments may point at H /0 ratio's higher
than 4 (for references, see Van Wielink (1986)).
To shed more light on the bioenergetics of the incomplete oxidation of
glucose in A. calcoaceticus, a study was undertaken on that part of the
electron transport chain participating in this process.
The electron transport chain
Electron transport proceeds via a number of redox carriers situated in
the cytoplasmic membrane. The chain contains NAD(P)-independent
dehydrogenases, lipid soluble p_-quinones, and cytochrome-containing oxidase
complexes. Identification and characterization of the redox carriers has
been performed by recording absorption spectra of the components in situ
and/or by purifying the solubilized components. In the case of A. cal
called d) and ubiquinone-9 has been established (Whittaker, 1971; Nakula et al., 1975). Asperger et al. (1978) gave evidence for the existence of two cytochrome oxidases, the o- and the d-type. They postulated a branched terminal oxidase system (Asperger et al., 1981). Ensley and Finnerty (1980) endorsed the statement of Asperger et al. (1981), that the presence of cytochrome o and d was not dependent on the carbon source used to grow the bacteria. They showed, that by growth under high aeration, cytochrome o-type oxidase was predominantly present, and under conditions of poor aeration, cytochrome oxidase type d and type a,.
The NAD(P)-independent dehydrogenases of A. calcoaceticus have been scarcely investigated in the past, recent examples are the flavoproteins lactate dehydrogenase and mandelate dehydrogenase (Allison et al., 1985; Allison and Fewson, 1986). On the other hand, as will be outlined in the following paragraph, NAD(P)-independent glucose dehydrogenase has been studied for a relatively long time.
Quinoprotein glucose dehydrogenase
Dehydrogenases are enzymes that deprive hydrogen from their substrates. The reduction equivalents taken from the substrate are transferred to an organic compound assisting the enzyme, the socalled cofactor or coenzyme. Based on the nature of this compound, dehydrogenases are sub-divided into three groups:
a) dehydrogenases using NAD(P), occurring in the cytoplasm of the cell. The nicotinamides are called coenzymes since they shuttle between the dehydrogenases and the respiratory chain, mono-oxygenases or enzymes in volved in reductive biosynthetic routes.
b) dehydrogenases containing flavins (FAD, FMN), bound to the cytoplasmic membrane. The flavin is called cofactor or prosthetic group since it remains
16
-bound to the protein after substrate conversion. The reduced flavins are
reoxidized in vivo by means of other components of the respiratory chain.
c) dehydrogenases containing pyrroloquinoline quinone (PQQ) as cofactor
or prosthetic group (Duine and Frank, 1981; Duine et al., 1986), bound at
the periplasmic side of the cytoplasmic membrane.
Nicotinamides and flavins are already known for a long time. The history
of this third cofactor that ultimately led to the elucidation of its chemi
cal structure and its name, "pyrrolo-quinoline-quinone", can be traced back
to Aerobacter aerogenes (nowadays known as Klebsiella aerogenes). The
glucose dehydrogenase activity of a cell-free extract from this organism was 2+
stimulated by Ng , but not by any known cofactor (Dalby and Blackwood,
1955). Later, Hauge (Hauge, 1961ab, 1964; Hauge and Murer, 1964), purified
and partially characterized glucose dehydrogenase from Bacterium anitratum
(nowadays known as Acinetobacter calcoaceticus) and detected that its cofac
tor had unusual properties. The cofactor was not restricted to A.
calcoaceticus since it could be used to reconstitute the app-glucose
dehydrogenase of Rhodopseudomonas spaeroides (Niederpruem and Doudoroff,
1965).
Methanol dehydrogenase of Gram-negative methylotrophic bacteria contains
a cofactor with similar properties (Anthony and Zatman, 1967), although
these authors were unaware of this fact at that moment. The cofactor of
methanol dehydrogenase was considered to be a pteridine (Urushibara et al.,
1971) and later a lumazine derivative (Sperl et al., 1973). KSR
spectros-copy, however, revealed that it was a nitrogen-containing o-quinone
(Westerling et al., 1979). The chemical structure was resolved independently
by Salisbury et al. (1979) and Duine et al. (1980).
"pyrrolo-quinoline-quinone" (PQQ), derived from
2,7,9-tricarboxy-lH-pyrrolo-(2,3f)quinoline-4,5-dione, and enzymes that have this cofactor are referred to as quinoproteins.
Several of the bacterial quinoprotein dehydrogenases are involved in incomplete oxidations leading for instance to bacterial gluconic acid and acetic acid production (Duine and Frank, 1981).
As already mentioned, Hauge (1964) purified and partially characterized the enzyme from A. calcoaceticus. Although he made claer that the enzyme contained an unusual cofactor, the identification occured later after the elucidation of the structure of PQQ (Duine et al., 1979). The enzyme is now classified as quinoprotein glucose dehydrogenase (EC 1.1.99.17).
Quinoprotein glucose dehydrogenases have been shown to occur in several other bacteria: Pseudomonas aeruifinosa (Duine and Frank, 1981),
Gluconobacter suboxydans (Ameyama et al., 1981), Pseudomonas fluorescens (Matsushita and Ameyama, 1982) and Klebsiella aerogenes (Neijssel et al., 1983). Apo-quinoprotein glucose dehydrogenase, which become functional after the addition of PQQ, were found in Acinetobacter lwoffi strains (Van Schie et al., 1984), Kscherichia coli strains (Hommes et al., 1984) and many other bacteria (see Van Schie, 1987).
Outline of the study
The goal of the research described in this thesis, was to obtain informa tion, at the molecular level of the components active in glucose oxidation by A. calcoaceticus. The choice of the enzyme was inspired by the idea to use it as a model system for the study of the enzymology of quinoproteins. Since the assay of glucose dehydrogenase does not require special condi tions, PQQ is not covalently bound, and an apo-enzyme is available that can be reconstituted to an active holo-enzyme with PQQ, glucose dehydrogenase seemed an attractive candidate. A.'calcoaceticus was chosen as the source of enzyme and electron transport chain components since parallel studies on the
18
-physiology of incomplete oxidation of glucose (Van Schie, 1987) and biosyn thesis of PQQ were started on this organism in the laboratories of Delft and Leiden (Goosen et al., 1987).
The suitability of glucose dehydrogenase as model enzyme is not
restricted to that for mechanistic studies on quinoproteins but also applies to that for studies on the bioelectrochemistry of NAD(P)-independent
dehydrogenases. PQQ is chemically and photochemically stable while oxygen is not an electron acceptor for glucose dehydrogenase. Therefore, this enzyme is attractive to study the transfer of electrons from reduced enzyme to an „ electrode. The fact that glucose is an important analyte in several fields
of application forms an additional argument. Several reports, describing modified electrodes suitable to accept electrons from glucose dehydrogenase, have already been devoted to this topic (Turner and Pickup, 1985; Mullen et al., 1985; D'Costa et al., 1986).
It is generally accepted that NAD(P)-independent dehydrogenases transfer their reduction equivalents at the level of ubiquinone. During the past few years, it has become clear, however, that quinoproteins are exceptional, the high redox-potential of their cofactor probably preventing the entry of electrons at this level. For instance, methanol dehydrogenase and
methylamine dehydrogenase are coupled to the electron transport chain at the level of cytochrome c. Assuming that glucose dehydrogenase has also a high redox potential and given the fact that cytochrome c is absent in A. cal-coaceticus. the site of coupling cannot be indicated at forehand.
The nature of the primary electron acceptor of quinoprotein glucose dehydrogenase is unclear. Hauge (1960, 1961) postulated glucose
dehydrogenase to be linked to the electron transport chain via a cytochrome b. He based this idea on the finding of a soluble cytochrome b, present in the early stages of the isolation of glucose dehydrogenase, that could be reduced in vitro by glucose in the presence of glucose dehydrogenase. In
contrast to Hauge, Beardmore-Gray and Anthony (1986) concluded, that
glucose, succinate and NADH are all oxidized by way of the sane b-type
cytochromes and glucose dehydrogenase donates its electrons to ubiquinone-9.
The results presented in Chapter II, show that glucose dehydrogenase is
detached from the cells when A. calcoaceticus is grown in the presence of
detergent. Arguments are put forward to propose a periplasmic localization
of the enzyme (Dokter et al., 1985).
The purification and characteristics of the enzyme are described in
Chapter III (Dokter et al., 1986). Comparison of its properties with those
of other glucose dehydrogenases indicates that there are two quite different
groups of glucose dehydrogenase.
It is shown in Chapter IV (Dokter et al. 1987a) that significant changes
in substrate specificity of glucose dehydrogenase occur by detachment of the
enzyme from the cytoplasmic membrane.
The purification and characterization of soluble cytochrome b_„„ and its
role as electron acceptor for glucose dehydrogenase, are described in
Chapter V (Dokter et al., 1987b).
The thesis is concluded with the characterization of the membrane-bound
cytochrome-containing complexes of A. calcoaceticus (Chapter VI, Dokter et
al., 1987c).
REFERENCES
Allison, N., 0'Donnell, M.J., Hoey, M.E. and Fewson, C.A. (1985) Biochem. J.
227, 753-757.
Allison, N. and Fewson, C.A. (1986) FEMS Microbiol. Lett. 36, 183-186.
Ameyama, M., Shinagawa, E., Matsushita, K. and Adachi, 0. (1981) Agric.
Biol. Chem. 45, 851-861.
20
-Anthony, C. and Zatman, L.J. (1967) Biochem. J. 104, 953-959.
Asperger, 0, Kleber, H.P. and Aurich, H. (1978) Acta biol. med. germ. 37, 191-198.
Asperger, 0., Borneleit, P. and Kleber, H.-P. (1981) Abh. d. Akad. Wiss. DDR 3, 259-270.
Baumann, P., Doudoroff, M. and Stanier, R.Y. (1968) J. Bacteriol. 95, 1520-1541.
Beardmore-Gray, M. and Anthony, C. (1983) J. Gen. Microbiol. 129, 2979-2983. Beardmore-Gray, M. and Anthony, C. (1986) J. Gen. Microbiol. 132, 1257-1268. Beijerinck, M. (1911) Proc. K. Ned. Akad. Wet. 13, 1066-1077.
Cox, J.C. and Haddock, B.A. (1978) Biochem. Biophys. Res. Commun. 82, 46-52. Dalby, A. and Blackwood, A.C. (1955) Can. J. Microbiol. 1, 733-742.
D'Costa, E.J., Higgens, I.J. and Turner, A.P.F. (1986) Biosensors 2, 71-87. De Bont, J.A.M., Dokter, P., Van Schie, B.J., Van Dijken, J.P., Frank,
J.Jzn., Kuenen, J.G. (1984) Antonie van Leeuwenhoek 50, 76-77.
Dokter, P., Van Kleef, M.A.G., Frank, J.Jzn. and Duine, J.A. (1985) Enzyme Microbiol. Technol. 7, 613-617.
Dokter, P., Frank, J.Jzn. and Duine, J.A. (1986) Biochem. J. 239, 163-167. Dokter, P., Pronk, J.T., Van Schie, B.J., Van Dijken, J.P. and Duine, J.A.
(1987a) FEMS Microbiol. Lett. 43, 195-200.
Dokter, P., Van Wielink, J.E., Van Kleef, M.A.G. and Duine, J.A. (1987b) Biochem. J. (submitted).
Dokter, P., Van Wielink, J.E., Geerlof, A., Oltmann, L.F., Stouthamer, A.H. and Duine, J.A. (1987c) J. Gen. Microbiol. (submitted).
Duine, J.A., Frank, J.Jzn. and Van Zeeland, J.K. (1979) FEBS Lett. 108, 443-446.
Duine, J.A., Frank, J.Jzn. and Verwiel, P.E.J. (1980) Eur. J. Biochem. 108, 187-192.
Duine, J.A..Frank, J.Jzn. and Jongejan, J.A. (1986) FEMS Microbiol. Rev. 32, 165-178.
Ensley, B.D. and Finnerty, W.R. (1980) J. Bacteriol. 142, 859-868.
Goosen, N., Vermaas, D.A.M, and Van de Putte, P. (1987) J. Bacteriol. 169, 303-307
Hauge, J.G. (1960) Biochim. Biophys. Acta 45, 250-262. Hauge, J.G. (1961) Arch. Biochem. Biophys. 94, 308-318. Hauge, J.G. (1964) J. Biol. Chem. 239, 3630-3639.
Hauge, J.G. and Murer, E.H. (1964) Biochim. Biophys. Acta 81, 244-250. Hoffmann, S. Mabeck, C E . and Vejlsgaard, R. (1982) J. Clin. Microbiol. 16,
443-451.
Hommes, R.W.J., Postma, P.W., Neijssel, O.M., Tempest, D.W., Dokter, P. and Duine, J.A. (1984) FEMS Microbiol. Lett. 24, 329-333.
Kashket, E.R. (1982) Biochemistry 21, 5534-5538.
Kitagawa, K., Tateishi, A., Nakano, F., Matsumoto, T., Morohoshi, T., Tanino, T. and Usui, T. (1986) Agric. Biol. Chem. 50, 2939-2940.
Makula, R.A., Lockwood, P.J. and Finnerty, W.R. (1975) J. Bacteriol. 121, 250-258.
Matsushita, K. and Ameyama, M. (1982) Methods Enzymol. 89, 149-154. Meyer, D.J. and Jones, C.W. (1973) Eur. J. Biochem. 36, 144-151. Muilen, W.H., Churchhouse, S.J. and Vadgama, P.M. (1985) Analyst 110,
925-928.
Muller, R.W. and Babel, W. (1986) Arch. Microbiol. 144, 62-66.
Neijssel, O.M., Tempest, D.W., Postma, P.W., Duine, J.A. and Frank, J.Jzn. (1983) FEMS Microbiol. Lett. 20, 35-39.
Nicholls, D.G. (1982) In: Bioenergetics. An Introduction to the Chemiosmotic Theory. Academic Press, London New York.
Niederpruem, D.J. and Doudoroff, M. (1965) J. Bacteriol. 89, 697-705. Rosenthal, S.L. (1978) In: Glucose non-fermenting Gram-negative bacteria in
22
-clinical microbiology. (Ed.) G.H. Gilardi, CRC Press, Florida.
Salisbury, S.A., Forrest, H.S., Cruse, W.B.T. and Kennard, 0. (1979) Nature 280, 843-844.
Sanchez Crispin, J.A., Dubourdieu, M. and Chippaux, M. (1979) Biochim. Biophys. Acta 547, 198-210.
Schlegel, H.G. (1986) In: General Microbiology (6th edition), Cambridge University Press, Cambridge London New York New Rochelle Melbourne Sydney. Sperl, G.T., Forrest, H.S. and Gibson, D.T. (1973) J. Bacteriol. 118,
541-550.
Turner, A.P.F. and Pickup, J.C. (1985) Biosensors 1, 85-115. Urushibara, T., Forrest, H.S., Hoare, D.S. and Patel, R.N. (1971)
Biochem. J. 125, 141-146.
Van Schie, B.J., Van Dijken, J.P. and Kuenen, J.G. (1984) FEMS Microbiol. Lett. 24, 329-333.
Van Schie, B.J., Hellingwerf, K.J., Van Dijken, J.P., Elferink, M.G.L., Van Dijl, J.M., Kuenen, J.G. and Konings, W.N. (1985) J. Bacteriol. 163, 493-499.
Van Schie, B.J. (1987) Ph. D. Thesis, University of Technology, Delft. Van Wielink, J.E., Reijnders, W.N.M., Oltmann, L.F. and Stouthamer, A.H.
(1983) Arch. Microbiol. 136, 152-157.
Van Wielink, J.E. (1986) Ph. D. Thesis, Vrije Universiteit, Amsterdam. Westerling, J., Frank, J.Jzn. and Duine, J.A. (1979) Biochem. Biophys. Res.
Commun. 87, 719-724.
-Production of quinoprotein D-glucose
dehydrogenase in the culture medium
of Acinetobacter calcoaceticus
P. Dokter, M. A. G. van Kleef, J. Frank, Jzn and J. A. Duine
Laboratory of Biochemistry, Delft University of Technology, Julianalaan 67, 2628 BC Delft, The Netherlands
(Received 29 March 1985; revised 28 May 1985)
On addition of low concentrations (0.005%) of Triton X-100 to a mineral medium supplemented with 0.5% heptadecane, a marked stimulation of growth rate was observed for Acinetobacter calco
aceticus strains able to grow on alkanes while appreciable amounts of soluble quinoprotein D-glucose
dehydrogenase [ D-glucose: (pyrroloquinoline-quinone) 1-oxidoreductase, EC 1.1.99.17] were found in the culture medium. At higher Triton X-100 concentrations (0.04%), still larger amounts of D-glucose dehydrogenase and also cytoplasmic enzyme activities' appeared in the culture medium. Although combinations of other carbon sources plus non-tonic detergents also produced these enzymes in the medium, the combination of heptadecane and Triton X-100 gave higher levels and had a stabilizing effect on D-glucose dehydrogenase. Therefore, by using this combination and cultur-ing within certain pH limits, a stable enzyme solution, havcultur-ing already a high specific activity, is produced while the cell harvesting and disruption steps can be circumvented. The results indicate that D-glucose dehydrogenase in this organism is a periplasmic enzyme, coupled to a cytochrome b. Keywords: Quinoprotein D-glucose dehydrogenase, EC 1.1.99.17; periplasmic enzyme; Acinetobacter calcoaceticus;
alkanes; Triton X-100
Introduction
A large number of industrial enzymes are hydrolases pro duced extracellularly by microorganisms. Among the bacterial representatives of these production organisms, Gram-positives form the main part. The reason for this is that they excrete these enzymes into the culture medium, a process which is probably facilitated by the absence of an outer membrane. Due to this mode of production, the enzymes can be easily isolated as the laborious steps of cell harvesting and disruption are not necessary. Therefore, in order to explore the possibilities for application of enzymes for which an excretion mechanism does not exist, methods should be developed which allow the large-scale isolation of these enzymes in a similar easy way.
Quinoprotein, that is PQQ-containing, D-glucose de hydrogenase [D-glucose :(pyrroloquinoline-quinone) 1-oxidorcductase, EC 1.1.99.17] is an enzyme which has potential for estimating glucose in body fluids' and patent applications for its use in biosensors have been made.2 It
occurs in many Gram-negative bacteria but, although the enzyme from Pseudomonas aentginosa has been assumed to occur in the periplasm,3 evidence for this is still lacking. It was found that during growth of Acinetobacter
calco-aceticus on a basal salt medium supplemented with
hexa-decane and Triton X-100, substantial amounts of D-glucose dehydrogenase appeared in the culture medium. In this report, the relevant factors for this production method are presented.
Materials* and methods Chemicals
All alkanes were from Baker, except hexadecane and heptadecane, which were from Jenssen and Fluka, respec tively. The non-ionic detergents were from Sigma.
Cultivation
Most of the experiments were performed with A.
calco-aceticus LMD 79.41, originally obtained from Professor J.
Hauge.3 Other strains tested were: ATCC 23055 (type
strain), ATCC 23220 and ATCC 23236 (strain HOI). The organisms were grown at 30°C with shaking (200 rev min"1) for about 7 days in 50 ml of an inorganic medium
containing per litre; KH2P04, 4.6 g; K2HP04, 11.5 g;
(NH4)2S04, 2.5 g; MgS04, 0.2 g and 1 ml spore solution
according to Vishniac.5 Alkaline phosphatase was induced
using a low phosphate medium.6
Enzyme assays
D-Glucose dehydrogenase activities were estimated as previously described,7 measuring the reduction rate of
Wurster's Blue in 0.1 M Tris-HCl buffer, pH 7.0. Malate dehydrogenase (L-malate; NAD* oxidoreductase, EC 1.1.1.37) activities were estimated by measuring the oxida tion rate of NADH (0.2 mM) with oxaloacetic acid (0.5 ITIM) in 0.1 M potassium phosphate buffer, pH 7.5.
26
-Alkaline phosphatase [orthophosphoric-monoester phospho-hydrolase (alkaline optimum), EC 3.1.3.1] activities were estimated by measuring the p-nitrophenol production rate at 420 nm with p-nitrophenyl phosphate (440 JJM) in 0.1 M Tris-HCl buffer, pH 8.0. The presence of detergents in the assay did not affect the reaction rates. All measurements were performed at room temperature. Enzyme production was expressed in units per ml culture medium (U ml"1).
One unit of enzyme activity was defined as the amount of enzyme required to convert one micromole of substrate per minute under the assay conditions.
Methods
Protein determinations were done according to Brad ford.8 Cell-free extracts were prepared as reported else
where.7 Electron microscopy pictures were taken from cells
fixed and stained as previously described.'
Results
Growth on alkanes
Growth was observed on straight chain alkanes having a carbon-number between IS and 18. A dramatic stimulation of growth occurred when Triton X-100 was incorporated in the medium (Figure 1). Cells grown under the latter condi tion (not on ethanol as a carbon source) had a similar appearance to that described by Kennedy et a/.10 for growth on hexadecane, showing numerous fimbriae at the outside of their surface. On raising the Triton X-100 con centration to 0.04%, fimbriae were now absent and half of the cells were empty. Table 1 shows that other non-ionic detergents were also active. In general, these detergents were effective in emulsifying heptadecane while the non-stimulating detergents were not.
D-Glucose dehydrogenase production
The finding of D-glucose dehydrogenase activity in the culture medium after growth on hexadecane plus Triton X-100 prompted us to test other carbon sources. From
Figure 2 it appears that the highest production was obtain
ed with heptadecane, while other carbon sources were less active in stimulating production (Table 2).
Table 1 Effect of detergents (0.005%/0.04%) on growth and D-glucose dehydrogenase production in the medium
50 75 100 Incubation time (h)
Figure 1 Growth on heptadecane without and with 0.04% Triton X-100: o, without Triton X-100 precultured on heptadecane;£, with out Triton X-100, precultured on acetate; • , with 0.04% Triton X-100, precultured on heptadecane; », with 0.04% Triton X-100, precultured on acetate Detergent Triton X-1S Triton X-35 Triton X-45 Triton X-100 Triton X-102 Triton X-165 Triton X-305 Triton N-57 Triton N-101
-Growth* 0.005%/0.04% 4/4 1/1.5 1/1.5 1.5/1 1.5/2 2/4 3/4 1.5/1.5 1.5/1.5 4 D-Glucose dehydrogenase activity (U m r ' l 0.21/0.12 0.47/1.22 0.64/1.26 0.50/1.40 0.31/0.47 0.0/0.0 0.0/0.16 0.60/1.20 0.74/1.14 0.0 Growth completed (days)
Table 2 Effect of different growth conditions on the specific activity and amount of D-glucose dehydrogenase in the medium and cells (measured in the cell-free extract) of a SO ml culture
Growth condition 0.4% sodium acetate 0.4% sodium acetate +0.04% Triton X-100 Sodium acetate* 0.5% heptadecane 0.5% heptadecane +0.005% Triton X-100 0.5% heptadecane +0.04% Triton X-100 Specific activity (U mg" Culture medium 0 0.26 0 0 8.8 4.2 protein) Cell-free extract 0.045 0.088 0.94 1.12 0.70 0.90 Total units Culture medium 0 3.55 0 0 32.7 107.1 Cell-free extract 0.66 0.23 3.1 58.3 29.9 23.0
B Carbon-limited continuous culture ( 0 » 0.15)
In addition, the nature of the detergent (Table ]) and its concentration (Table 2) were important factors. Buffering of the medium was essential since no enzyme production was found outside the pH range 6.3 to 7.5. The enzyme in the culture medium was very stable, as determined after incubation at 30°C during several days. This also applied to the apoenzyme form produced in the culture medium by a PQQ" mutant, in contrast to the high lability observed for the apoenzyme in a cell-free extract (unpublished results).
Another attractive point is the high specific activity of the enzyme in the culture medium compared with that in a cell-free extract (Table 2).
Localization
In order to shed light on the localization of D-glucose dehydrogenase in the cell, the occurrence of marker enzymes in the culture medium was tested. Significant pro duction of alkaline phosphatase (a periplasmic enzyme) in the medium already occurred at 0.005% Triton X-100 (at lower Triton X-100 concentrations, a sharp decrease in production was observed), as was the case for D-glucose dehydrogenase (Figure 3). On the other hand, malate de hydrogenase (a cytoplasmic enzyme) production was insig nificant under this condition but became maximal at 0.04% Triton X-100 (Figure 3). Higher concentrations of Triton X-100 did not increase the production of the enzymes. Clearly, the latter condition induced some lysis, as is also indicated by the protein content of the medium (Table 2) and the fact that half of the cells were found to be empty in the electron microscopy pictures. From Figure 4 it appears that this process is not abrupt, since enzyme pro duction follows the growth curve. The D-glucose
dehydro-e
>.
•j to trt O C•
O» e 0.5 X « 1 o-H 15 16 17 ie o C a r b o n a t o m s
Figure 2 D-Glucose dehydrogenase activities in the medium after growth on different straight chain alkanes (0.5%) plus 0.04% Triton X-100
100 r
Z 50
0L i — r
-A B C -A B C -A B C
Figure 3 Enzyme activities in the culture medium after growth on the low phosphate medium supplemented with 0.5% heptadecane and various concentrations of Triton X-100. Vertical blocks mark the activity of: (a) D-glucose dehydrogenase; (b) alkaline phos-phatase; (c) malate dehydrogenase. Triton X-100 concentrations are indicated as follows: A. 0%; B. 0.005%; C, 0.04%
genase was in a soluble form, as appeared from ultracentri-fugation and gel filtration experiments with the culture medium. These also revealed that the enzyme was bound to a cytochrome which became reduced on addition of glucose and had the spectral characteristics of a cytochrome b (results not shown).
Discussion
A. calcoaceticus is a very versatile organism, using many
compounds as a carbon source. During growth on alkanes, however, distinct features are observed. It has been report ed" that extracellular membrane particles with a composi tion similar to that of the outer membrane occur in the culture liquid. The authors11 postulated that these particles
have an emulsifying action and possibly have a role as a vehicle for the transport of the alkane. It is not known, however, in which respects these vesicular particles are related to the emulsifiers produced by various strains.12
Cell morphology of A. calcoaceticus grown on alkanes is also different from that of cells grown on other carbon sources. Numerous thin fimbriae are seen at the cell surface and it has been suggested that they are a major factor in adherence to hydrocarbon droplets.10 Another striking
property is the occurrence of intracellular inclusions under these growth conditions.13 Similar morphological proper
ties were observed in this study with A. calcoaceticus LMD 79.41 grown on heptadecane, without and with
(0.005%) Triton X-100 addition, but not for growth on acetate or ethanol.
Addition of detergents to the culture medium of
Pseudo-monas aeruginosa stimulates growth on alkanes.14 A.similar
effect was observed for .4. calcoaceticus LMD 79.41 (Figure 7). An obvious explanation is that Triton X-100 emulsifies the heptadecane into small droplets. This is in agreement with the observation that detergents which did not stimu late growth (Table 1) mostly failed to emulsify heptadecane. Detergents that stimulated growth are known to emulsify hydrocarbons.15 Enhancement of extracellular enzyme
production on addition of detergent to the growth medium has been observed in the case of Gram-positive bacteria16
and fungi.17 Most probably, the detergent raises the perme
ability of the cytoplasmic membrane for these proteins. Gram-negative bacteria contain an outer membrane so that an extra barrier exists for the excretion of proteins into the medium. Although an export system has been reported in special cases,1 , 1 9 most enzymes outside the
cytoplasmic membrane occur in the periplasmic space and can only be freed by methods which damage the outer membrane (e.g. by osmotic shock procedures). The results presented here indicate that at low Triton X-100 concentra tions, a periplasmic enzyme like alkaline phosphatase and respiratory-chain coupled D-glucose dehydrogenase appear in the culture medium of the Gram-negative bacterium
A. calcoaceticus. This observation could be explained by
assuming that the detergent has the following effect on the bacterium: (a) lysis of part of the cells; (b) stimulation of an excretion mechanism for these enzymes across the outer membrane; (c) partial breakdown of the permeability barrier of the outer membrane with a concomitant solubi-lizing action on the structures to which the enzymes are attached. For the moment, possibility (c) seems the most likely, since the membranes looked normal and empty cells were not seen in the electron microscopy pictures while intracellular enzymes were not found in the culture medium. In view of this, the higher production of D-glucose dehydrogenase with alkanes as a carbon source could origi nate from special alterations of the outer membrane induced by these growth substrates, as described above.
At higher Triton X-100 concentrations, however, cyto plasmic enzymes appear in the culture medium. The higher concentration of detergent probably now also affects the permeability of the cytoplasmic membrane, since about
F 3 c <0 O
*,
5 2 S ' £t<
•/ ^^ ^r I y\1 /
\
/ / \ ^ •
-/ -/Sit-1 / /
JL>^.
1.5 "7 Ë 1.0 3 ■>. 0.5 " o z a a ■15^
E -10 3>.
>
■" a z o z 25 50 75 100 I n c u b a t i o n t i m e ( h )Figure 4 Enzyme production in the medium during growth on
heptadecane (0.5%) and Triton X-100 10.04%). X, Absorbance at 660 nm; o, o-glucose dehydrogenase (GDH) activity; • , malate dehydrogenase (MDH) activity
28
-half of the cells were empty under this condition of growth and substantial amounts of protein were found in the culture medium.
The example presented here might indicate that it is worthwhile to use detergents in order to decrease the per meability barrier for proteins in Gram-negative bacteria without impairing their synthesis or stability. If a direct concentration step from the cell culture is available, an attractive large-scale production should be possible for such an enzyme.
Acknowledgemen ts
We thank Mrs W. H. Batenburg-van der Vegte for perform ing the electron microscopy experiments and Bart van Schie for providing cells grown in continuous culture. This research was supported in part by The Netherlands Tech nology Foundation (STW).
References
1 li. Limbach, R. Helper, Merck Patent Gambll, Ger. OITen. DE 3211167
2 Geneiics-lnt. Uur. Pat. 78-636: 23.10 (1981)
3 Midgley, M. and Dawes, E. A. Biochem. J. 1973, 132, 141 4 Haugc, J. G. Biochim. Biophys. Ada 1960,45, 263 5 Vishniac, W. and Santer, M. Bacleriol. Rev. 1957, 21, 195 6 Cheng, K. J., Ingram, J. M. and Costerton, J. W. J. Bacleriol.
1970, 104,748
7 Duine, J. A., Frank Jzn, J. and Van Zeeland, J. K. FFBS Lett. 1979, 108,443
8 Bradford, M. Anal. Biochem. 1976, 72, 248
9 Strohl, W. R. and Larkin, J. M. Curr. Microbiol. 1978, 1,151 10 Kennedy, R. S., Finnerty, W. R., Sudarsanan, K. anu Young,
R. A. Arch. Micriobiol. 1975, 102, 75
11 Claus, R., Kappeli, O. and Fiechter, A. J. Gen. Microbiol. 1984, 130,1035
12 Sar, N. and Rosenberg, E. Curr. Microbiol. 1983, 9, 309 13 Scott, C. C. L. and Finnerty, W. R. J. Bacleriol. 1976, 127,
481
14 Nakahara, T., Hisalsuka, K. and Minoda, Y. J. Ferment. Technol. 1981,59,413
15 Schónfeldt, N. Surface Active L'lhytctie Oxide Adducts 1 st ed., Pergamon Press, Oxford, 1969, pp. 592, 801
16 Chandra, A. K., Medda, S. and Bhadra, A. K. J. Ferment. Technol. 1980,58, 1
17 Reese, E. T. and Maguire, A. Appl. Environ. Microbiol. 1969, 17,242
18 Lory, S., Tai, P. C. and Davis, B. D. /. Bacleriol. 1983, 156, 695
19 Poole, K. and Hancock, E. W. FEMS Microbiol. Leu. 1983, 16,25
Purification and characterization of quinoprotein glucose
dehydrogenase from Acinetobacter calcoacetkus L.M.D. 79.41
Paul DOKTER, Johannes FRANK, Jzn. and Johannis A. DUINE
Delft University of Technology Laboratory of Microbiology and Enzymology, JuUanalaan 67, 2628 BC Delft, The Netherlands
Quinoprotein glucose dehydrogenase (EC 1.1.99.17) from Acinetobacter calcoacetkus L.M.D. 79.41 was purified to homogeneity. It is a basic protein with an isoelectric point of 9.5 and an M, of 94000. Denaturation yields two molecules of PQQ/molecule and a protein with an MT of 48000, indicating that the enzyme consists of two subunits, which are probably identical because even numbers of aromatic amino acids were found. The oxidized enzyme form has an absorption maximum at 350 nm, and the reduced form, obtained after the addition of glucose, at 338 nm. Since double-reciprocal plots of initial reaction rates with various concentrations of glucose or electron acceptor show parallel lines, and substrate inhibition is observed for glucose as well as for electron acceptor at high concentrations, a ping-pong kinetic behaviour with the two reactants exists. From the plots, Km values for glucose and Wurster's Blue of 22 msi and
0.78 mM respectively, and a Kmax of 7,730 /«mol of glucose oxidized/min per mg of protein were derived. The enzyme shows a broad substrate specificity for aldose sugars. Cationic electron acceptors are active in the assay, anionic acceptors are not. A pH optimum of 9.0 was found with Wurster's Blue and 6.0 with 2,6-dichlorophenol-indophenol. Two types of quinoprotein glucose dehydrogenases seem to exist: type I enzymes are acidic proteins from which PQQ can be removed by dialysis against EDTA-containing buffers (examples are found in Escherichia coli, Klebsiella aerogenes and Pseudomonas sp.); type II enzymes are basic proteins from which PQQ is not removed by dialysis against EDTA-containing buffers (examples are found in A. calcoacetkus and Gluconobacler oxydans).
INTRODUCTION
Bacteria converting aldose sugar(s) into the corres ponding lactones frequently contain an NAD(P)-inde-pendent dye-linked glucose (aldose) dehydrogenase. The enzyme from Acinetobacter calcoacetkus (formerly known as Bacterium anitratum) has been purified to homo geneity and partially characterized (Hauge et al., 1964). Although at that time from the scarce data no structural assignments could be made, it was already clear that the enzyme contained an unusual cofactor. Despite this important finding, no further work on this aspect was reported. After the discovery of PQQ as the cofactor in methanol dehydrogenase (EC 1.1.99.8), comparative studies revealed that the cofactor of glucose dehydrogen ase from A. calcoacetkus was also PQQ (Duine et al., 1979). This resulted in a reclassification and the enzyme is now indicated as 'quinoprotein (PQQ-containing) glucose dehydrogenase' (EC 1.1.99.17) (Nomenclature Committee of the International Union of Biochemistry, 1979). Since in the meantime several similar enzymes have been reported to occur in other bacteria, e.g.
Pseudomonas aeruginosa (Duine & Frank., 1980), Gluconobacler suboxydans (Ameyama et at., 1981), Pseudomonas fluorescens (Matsushita & Ameyama.,
1982) and Klebsiella aerogenes (Neijssel et al., 1983), it seems safe to conclude that all bacterial dye-linked glucose dehydrogenases are quinoproteins.
Recent work on non-acid-producing Acinetobacter
Iwoffi strains revealed that they contain quinoprotein
glucose dehydrogenase apoenzyme (van Schie et a/., 1984). This phenomenon appears to be widespread, and
even the common Escherichia coli laboratory strains harbour it, the apoenzyme becoming functional only after addition of PQQ (Hommes et al., 1984). From this, it might be concluded that the enzyme has been overlooked in the past, and that in the natural environment non-phosphorylative glucose dissimilation occurs more frequently than previously thought.
In view of the recent discoveries, more information on glucose dehydrogenase is required. Moreover, the enzyme is an attractive model system to study the enzymology of quinoproteins, since, in contrast with methanol dehydrogenase, its activity does not depend on the presence of an activator, and reconstitutable apoenzyme forms do exist, or can be prepared from holoenzyme (Duine et al., 1979).
Unfortunately, the strain originally used by Hauge (1964) has been lost, and even some doubt has been expressed about its identity (Beardmore-Gray & Anthony, 1983). As a consequence, it was necessary to purify the enzyme from a genuine A. calcoacetkus strain, to compare the results with those obtained by Hauge (1964), and to characterize the enzyme more extensively.
EXPERIMENTAL Materials
Wurster's Blue was prepared as described previously (Duine et a/., 1978). The PQQ used was prepared by both biological (Duine & Frank, 1980) and synthetic methods (Corey & Tramontano, 1981). 2-Mercaptoethanol was from Bio-Rad Laboratories. Tris was from Janssen
Abbreviations used: PQQ, pyrroloquinoline quinone (2,7,9-tricarboxy-l//-pyrrolo[2,3/]quino!ine-4,5-dione); Wurster's Blue is the free radical of JVAWW-tetramethyl-^-phenylenediamine.
32
-Chimica. Alcohol dehydrogenase (product no. 102709) was from Boehringer. Fractogel TSK HW-50(S), KH2P04, Na2HP04, D-glucose, L-arabinose, D-xylose,
lactose, 2,6-dichlorophenol-indophenol, cone. HC1, SDS, guanidinium chloride and NaCl were from Merck. CM-Sepharose, Sephadex G-100, Sepharose CL-6B, gradient gels PAA 4/30 and molecular-mass calibration (low-Afr and high-Mr) kits were from Pharmacia.
Polyacrylamide, bisacrylamide, iVAWW-tetramethyl-ethylenediamine, Coomassie Brilliant Blue G-250, bovine serum albumin (product no. 11920), carbonic anhydrase (product no. 15880), chymotrypsin (product no. 17160) and conalbumin (product no. 17465) were from Serva. Cytochrome c (product no. C-7752), ribonuclease (product no. R-4875), ovalbumin (product no. A-7641) and phosphorylase b (product no. P-6635) were from Sigma Chemical Co. N2 was from Hoekloos and the Os
filter (model 7970) was from Chrompack.
Culture conditions
Acetate-limited chemostat cultures of A. calcoaceticus L.M.D. 79.41 (identified with the API 20 NE system) were provided by B. J. van Schie (van Schie el al., 1984). The cells were harvested and stored frozen at — 20 °C. Enzyme purification
Cell-free extract (400 ml) was prepared from cells (200 g wet wt.) as described previously (Duine el al., 1979) and applied to a CM-Sepharose column (4.4 cm
x 5 cm) equilibrated with 20 mM-potassium phosphate buffer, pH 7.0. The column was washed with the same buffer (5 column volumes). The enzyme was eluted with 0.2 M-potassium phosphate buffer, pH 7.0. After concen tration by pressure filtration on a type PTGC 047.10 Pellicon membrane (Millipore), gel-filtration chromato-graphy was performed on a Fractogel TSK HW-50(S) column (2 cm x 60 cm) in 50 mM-potassium phosphate buffer, pH 7.0. The pooled fractions were applied to a CM-Sepharose column (1 cm x 5 cm) and eluted with a linear gradient of 0-0.2 M-NaCl in 20 mM-potassium phosphate buffer, pH 7.0. Finally, h.p.l.c. gel filtration was performed on a Serva Si-300-polyol column (0.5 cm x 9.1 cm) in 0.1 M-sodium phosphate buffer, pH 6.5, at a flow rate of 0.5 ml/min.
Enzyme assay
Activities were determined at room temperature by measuring the rate of discoloration of Wurster's Blue at 610 nm in a mixture containing 80 /tM-Wurster's Blue, 1 mM-KCN, enzyme or cell-free extract, 0.1 M-Tris/HCl buffer, pH 7.0, and 20 mM substrate in a final volume of 2 ml. The reaction was started by adding the substrate. One enzyme unit refers to 1 /«mol of aldose sugar converted/min under these conditions. The calculations were based on a molar absorption coefficient (calculated from absorbance measurements) for Wurster's Blue of 1 2 4 0 0 M- 1 cm"1 at 610 nm. Activities were also deter
mined with 2,6-dichlorophenol-indophenol as electron acceptor, by the assay described by Hauge (1960).
Protein determinations
During purification, protein concentrations were determined by the method of Bradford (1976), with bovine serum albumin as a standard. For the purified enzyme the a value was obtained from absorbance measurements at 205 and 280 nm by using the
chromatographic procedure described by van Iersel et al. (1985). Protein concentrations of pure enzyme solutions were calculated from the absorbance values at 280 nm. Polyacrylamide-gel electrophoresis
Glucose dehydrogenase was electrophoresed in gel slabs of 7.7% polyacrylamide cross-linked with 0.2% bisacrylamide, in 0.02 M-potassium phosphate buffer, pH 7.0, with the anode above the application side of the gel. For enzyme activity staining, the gels were immersed in a solution containing 0.1 M-Tris/HCl buffer, pH 7.0, 20mM-glucose and 600/tM-Wurster's Blue. Enzyme activity was visible as a white band on a dark-blue background. Protein staining was done with Coomassie Brilliant Blue G-250 (Pharmacia, 1984). Electrophoresis under denaturing conditions was performed on poly acrylamide gradient gels (PAA 4/30) in the presence of SDS, with the use of high-A^ and low-Afr electrophoresis
calibration kits as a reference (method described by Pharmacia, 1984).
Kinetic measurements
Substrate specificity. Activity of cell-free extract and of partially purified enzyme towards aldose sugars (20 mM) was determined spectrophotometrically as described above.
Kinetic parameters. Initial reaction rates were deter
mined with various concentrations of glucose and Wurster's Blue at room temperature under the conditions used in the enzyme assay. An enzyme concentration of 1 nM was used. The measured values and concentrations were plotted according to the method of Lineweaver & Burk (1934). Kinetic parameters were determined from secondary plots.
M, determinations
The MT of the native enzyme was determined by the
method of Andrews (1965) by gel filtration on a Sephadex G-100 column (1 cm x 55 cm) in 50 mM-potassium phosphate buffer, pH 7.0, containing 0.1 M-NaCl at a flow rate of 7.8 ml/h. Proteins used for calibration were: horse heart cytochrome c (MT 12500),
ribonuclease (A/r 13700), chymotrypsin (Mr 25000),
carbonic anhydrase (Mr 30000), ovalbumin (A/r 43000),
bovine serum albumin (MT 67000), conalbumin (A/r
86000) and baker's-yeast alcohol dehydrogenase (Mr
141000). MT determinations of the denatured enzymes
were performed by gel filtration on a Sepharose CL-6B column (1 cm x 55 cm) in 6 M-guanidinium chloride with or without the addition of 2-mercaptoethanol (0.3 M) to the sample. Phosphorylase b (Mr 92500), bovine serum
albumin (Mr 67000), carbonic anhydrase (MT 30000),
ribonuclease (A/r 13700) and cytochrome c (MT 12500)
were used as standards for calibration of the column (Ansari & Mage, 1977).
PQQ analysis
A 150/tl sample of the enzyme in 0.1 M-sodium phosphate buffer, pH 6.5, containing 0.2% SDS was kept at 100 °C for 10 min. The solution was subsequently injected on a Serva Si-300 (4.5 mm x 250 mm) h.p.l.c. gel-filtration column eluted with 0.1 M-sodium phosphate buffer, pH6.5, containing 0.1% SDS at a flow rate of 0.3 ml/min. The eluate was monitored at 280, 249 and 330 nm with a photodiode array detector
(Hewlett-Table 1. Purification of glucose dehydrogenase from A. cakoacetkus L.M.D. 79.41 For experimental details see the text.
Step Total protein (mg) Total activity (units) Specific activity (units/mg) Yield (%) Cell-free extract CM-Sepharose (0.2 M eluate) Gel filtration (Fractogel) CM-Sepharose (gradient) H.p.l.c. gel filtration 12500 400 42 8.7 1.5 8800 79O0 4600 2700 960 0.7 20 110 340 640 100 90 50 30 10
Packard model 1040A). The PQQ content of the enzyme sample was calculated from the peak height from the low-Afr factor monitored at 330 nm (identified as PQQ, by comparing the retention time and absorption spectrum with those of authentic PQQ). The system was calibrated by injecting 150 ftl samples of six different PQQ concentrations in the range 2-20 fiM.
Aromatic amino acids
The aromatic amino acid composition of the enzyme was determined by multicomponent analysis of the absorption spectrum of the denatured enzyme in 6 M-guanidinium chloride, measured with a Hewlett-Packard model HP 8450A spectrophotometer. Standards used for the calculations were those described by Levine & Frederici (1982). The absorption spectrum above 300 nm appeared to be identical with that of PQQ under these conditions. This was used to correct the spectrum below 300 nm for the presence of PQQ before perfor mance of the multicomponent analysis. It was also used as a method to measure the amount of PQQ in the enzyme. Isoelectric point
Isoelectric focusing was performed on a horizontal gel slab [2 mm thick, 5% acrylamide, 0.15% bisacrylamide and 2% (v/v) Ampholine buffer, pH 3-10 (LKB-Produkter)] by following the method described by LKB-Produkter(1977).
RESULTS Purification
Results of the purification procedure are presented in Table 1. Glucose dehydrogenase was eluted from the CM-Sepharose column at a concentration of 80 mM-NaCl. It had a retention time of 45 min on the h.p.l.c. column. The final preparation gave one single band on polyacrylamide-gel electrophoresis, by protein staining as well as by enzyme-activity staining. The normalized absorption spectra taken from the enzyme peak eluted from the h.p.l.c. column (upslope, at the top and downslope) were identical.
Substrate specificity
Other susbstrates found to be active with the purified enzyme were D-xylose (20%), L-arabinose (35%), lactose (65%), galactose (30%), D-melibiose (10%), cellobiose (70%) and maltose (90%) (activities compared with D-glucose given in parentheses). Similar ratios were obtained with 2,6-dichlorophenol-indophenol as electron acceptor. The enzyme has a pH optimum of 9.0 (in
0 0.15 0.30 1/[Glucose] (mM"'|
Fig. 1. Initial reaction rates of quinoprotein glucose dehydrogen ase (1 nM) with glucose as the variable substrate The Wurster's Blue concentrations were 63 p.M (A), 82 /tM (A), 155/JM ( ■ ) and 286/tin (□)•
Tris/HCl buffers) with Wurster's Blue and 6.0 (in potassium phosphate buffers) with 2,6-dichlorophenol-indophenol as electron acceptor. K3Fe(CN), was not an electron acceptor in the assay, either at pH 9.0 or at pH 6.0.
Kinetic parameters
As determined experimentally, at certain ratios of glucose to Wurster's Blue concentrations the plots of the kinetic data (Fig. 1) gave parallel straight lines. Substrate inhibition was observed with glucose (Fig. 2) as well as with Wurster's Blue (the latter is not shown, but is clear from the downward curvature in Fig. 2). Assuming ping-pong kinetics, Km values of 22 mM and 0.78 miu
could be derived from secondary plots for glucose and Wurster's Blue respectively. The Kmax was calculated to be 7.730/tmol of glucose oxidized/min permg of protein. The specific reaction rate with the assay described by Hauge (1960) was 490 units/mg of protein. The specific reaction rate at pH 9.0 and 25 °C in a mixture containing 80/tM-Wurster's Blue, 1 mM-K.CN, enzyme, 0.1 M-Tris/HCl buffer and 20 mM-glucose in a final volume of 2 ml was 1100 units/mg of protein.
M, determinations
The Afr of the native enzyme, determined by gel
34
-1.0 1.5 1/[Glucose] |mti*'l
Fig. 2. Substrate inhibition of quinoprotein glucose dehydrogen-ase (1 nM) by glucose
The Wurster's Blue concentrations were 16/tM (A), 102 fiM (■) and 336 /»M (G).
enzyme was 47500 (±5000) as determined by gel nitration and 48000 (±5000) as determined by electro-phoresis on a polyacrylamide gradient gel in SDS. P Q Q content
From the h.p.l.c. gel-filtration chromatogram of enzyme denatured with SDS the PQQ concentration was calculated to be 18.4 /M (2.76 nmol) for an enzyme concentration of 10.2/*M (1.53 nmol) (by using a = 14 litre g ~ ' c m- 1 and MT 94000). From the absorption
spectfum of enzyme denatured with guanidinium chloride it was calculated that 1.88/JM(1.5 nmol) enzyme solution contains 4.04/JM-(3.2 nmol) PQQ.
Aromatic amino acid content
All the common aromatic amino acids were present. The numbers of residues per enzyme molecule were found to be: phenylalanine, 21.8; tryptophan, 10.4; tyrosine, 41.6.
Isoelectric point
The isoelectric point was found to be 9.5, so that the enzyme is a basic protein, a property in accordance with its chromatographic behaviour.
Absorption spectrum
The absorption spectrum of the enzyme as it is isolated (Fig. 3) has a broad maximum at 350 ran, and the spectrum of the enzyme measured under anaerobic conditions in the presence of an excess of glucose has a sharp maximum at 338 nm.
DISCUSSION
Quinoprotein glucose dehydrogenase from A.
calco-aceticus L.M.D. 79.41 was purified to homogeneity.
The M, of the native enzyme (94000), the substrate specificity, the apparent Km for glucose (3 mM) and the
pH optimum (6.0) with 2,6-dichlorophenol-indophenol as electron acceptor are comparable with those of the enzyme from Bacterium anitratum (Hauge, 1964). Similarly, the ping-pong kinetic behaviour, as indicated by the parallel lines in Fig. 1 and the dual-substrate
O:ZÏ> 0.20 0.15 0.10 0.05 0
-1
','
1 /•
■ w\
200 300 400 500 Wavelength (nm)Fig. 3. Absorption spectra of quinoprotein glucose dehydro genase
The absorption spectrum was measured in 0.1 M-sodium phosphate buffer, pH 6.5, before ( ) and after ( ) the addition of glucose to the cuvette under anaerobic conditions (bubbling with N„).
inhibition with glucose and Wurster's Blue, were also observed with glucose and 2,6-dichlorophenol-indophenol by Hauge (1960). Further support for the kinetic similarity is reflected by the comparable ratios of the activities (Hauge, 1960) obtained with different aldose sugars.
Just like the glucose dehydrogenases from A.
calco-aceticus L.M.D. 70.9 and L.M.D. 79.39 (Duine et al.,
1979), the enzyme from A. calcoaceticus L.M.D. 79.41 is a quinoprotein. This, most probably, also applies to the enzyme from B. anitratum in view of the shape of the absorption spectra of the reduced and the oxidized enzyme forms (Hauge, 1964). The enzyme has an Mr of
94000 and consists of two subunits with the same Mr
(48000), the subunits probably being identical as even numbers of aromatic amino acid residues and two molecules of PQQ per enzyme molecule were found. The enzyme from B. anitratum (Hauge, 1964) was reported to have a similar MT (86000) but only one cofactor per
enzyme molecule (determined by titrating the oxidized enzyme form with glucose). The lower cofactor content of the enzyme from B. anitratum is also apparent from the absorption spectra (Hauge, 1964). Although the maxima of the reduced and the oxidized enzyme forms (338 and 350 nm) and the specific absorption coefficients
(a = 14 litre • g""1 ■ cm-1) are identical, different ratios exist
for the absorbancies in the spectra (data in parentheses are those estimated for the enzyme from B. anitratum): the oxidized enzyme form shows a broad maximum around 350 nm and ratios of A2m/A3S0 of 7.2 (11.3) and ^s8o/^2«o °f 1-56 (1.65). The reduced enzyme form has
