The Malonate Decarboxylase Enzyme System of Malonomonas Rubra Identification, Purification and Biochemical Characterization of Components

Research Collection

Doctoral Thesis

The malonate decarboxylase enzyme system of Malonomonas rubra identification, purification and biochemical characterization of components

Author(s): Hilbi, Hubert Franz Pius

Publication Date: 1994

Permanent Link: https://doi.org/10.3929/ethz-a-000971872

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.

ETH Library Diss. ETH No. 10766

The malonate decarboxylase enzyme System of Malonomonas rubra: Identification, purification and biochemical characterization of components

A dissertation submitted to the

SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZÜRICH for the degree of DOCTOR OF NATURAL SCIENCES

presented by HUBERT FRANZ PIUS HILBI Dipl. Natw. ETH bomMay30,1%5 Citizen of Zug (ZG) and Flums (SG)

accepled on the recommendation of Prof. Dr. P. Dimroth, examiner Prof. Dr. T. Leisinger, coexaminer

Zürich 1994 Dank

Ich möchte Herrn Prof. Dr. P. Dimroth für die fachliche Betreuung dieser Doktorarbeit danken und insbesondere für die Ideen, die in kritischen Momenten verhindert haben, dass das Projekt abgestürzt ist.

Meinem "Diplomvater" Herrn Prof. Dr. T. Leisinger danke ich für die wohlwollende Übernahme des Koreferats.

Uschi & Claudia, die jeweils die fürchterlichen Stunden vor dem ersten Kaffee mit mir

teilten, sei an dieser Stelle für ihr Verständnis gedankt, und gleichzeitig entschuldige ich mich beim Feuerlöscher LFV D19 für den unbeherrschten Tritt, der ihm fast seinen Halt gekostet hätte.

Den übrigen D-Stock-Bewohnerlnnen danke ich für hebevoll ausgewählte Geburtstagsgeschenke, Apdros und Fachdiskussionen; für letztere bedanke ich mich

auch auf Institutsebene.

Corinne danke ich für die Tage in den Cinque Terre & das Salz auf unserer Haut,

Christian für die Idee nach Indien zu fahren & ähnlich schräge Einfälle, Karin für lange Jahre, Urs für nächtliche Gespräche, Marianne für den Weg zwischen Pigalle & Züribar und Martina für gemeinsame Kultur.

Meinen Eltern möchte ich herzlich für die grosszügige Unterstützung dieser Ausbildung danken. Table of Contents

ZUSAMMENFASSUNG

SUMMARY

Chafter I: Introduction S

1.1. Physiology of anaerobic, chemotrophic bacteria 5 1.1. Bioenergetic callenges of chemotrophic anaerobes 5 1.2. Complete anaerobic oxidation of organic matter to CO2 9 1.3. The role of sodium in bacterial physiology 11

1.4. Primary sodium pumps 13

1.5. Endergonic Systems depending on ApNa+ 16

I. 2. Conservation of decarboxylation energy 17

2.1. Fermentation of malonate and other saturated

dicarboxylates 17 2.2. Soluble decarboxylases and substrate/product antiport 19

2.3. Decarboxylation-linked primary sodium pumps 21 1.3. Aimsofthework 27

1.4. References 29

Chafter II: Malonate decarboxylase of MaUmomonas rubra,

a novel type of biotin-containing acetyl enzyme 37 Eur. J. Biochem. 207:117-123 n. 1. Summary 38 II. 2. Introduction 38

n. 3. Materials and Methods 40

n. 4. Results 43

II. 5. Discussion 52

U. 6. References 57 Chafter UI: The malonate decarboxylase enzyme System of Malonomonas rubra: evidence for the cytoplasmic location of the biotin-containing component 59 Arch. Microbiol. 160:126-131

DI. 1. Summary 60

in. 2. Introduction 60

HI. 3. Experimental Procedures 62

HI. 4. Results 65

HI. 5. Discussion 71

DI. 6. References 73

Chafter IV: and Purification characterization of a cytoplasmic enzyme component of the Na+-activated malonate decarboxylase System of Malonomonas rubra: acetyl-S-acyl carrier protein: malonate acyl carrier protein-SH transferase 75 Arch. Microbiol. (in press) IV. 1. Summary 76 IV. 2. Introduction 76

IV. 3. Materials and Methods 79

IV. 4. Results 82

IV. 5. Discussion 93

IV. 6. References 97

Appendix I: N-terminal sequence of the transferase 100

Appendix II: Preliminary experiments on the purification of the biotin protein of malonate decarboxylase and other protein components involved in malonate decarboxylation 101 Summary 102

Introduction 103

Materials and Methods 104

Results 108

Discussion 114

References 117 Chafter V: General Discussion 119

V.l. The malonate decarboxylase enzyme System of M. rubra 119 1.1. Activation of malonate 119

1.2. Purification of a protein thiol transferase and relationship of malonate decarboxylase to citrate lyase 122 1.3. Relationship of malonate decarboxylase to the Na+-

transport decarboxylases 125

V. 2. Anaplerotic needs for fermentaüve growth on malonate 127 2.1. Synthesis of Cj-compounds 127 2.2. Fatty acid synthesis 128 2.3. Redox reactions 129

V. 3. Outlook 130

V. 4. References 131

Curriculum Vitae 133

List of Publications 134 1

Zusammenfassung

Malonomonas rubra, ein mikroaerotolerantes, strikt Na+-abhängiges Faulschlamm- Bakterium, wächst anaerob auf Malonat als einziger Kohlenstoff- und Energiequelle. Malonat Decarboxylase, das Schlüsselenzym dieser Fermentation, setzt dabei das

Substrat quantitativ zu Acetat und C02 um. Die Energieausbeute dieser Reaktion beträgt nur -17.4 kj pro Mol, und daher muss die ATP-Synthese über chemi- osmotische Prozesse erfolgen. Die Decarboxylierung von Malonat stellt ein chemisches

Problem dar, weil die C-C-Bindung für die Spaltung aktiviert werden muss. Ziel der vorliegenden Arbeit war die Aufklärung der Malonat-Aktivierung und die biochemische Charakterisierung der Malonat-Decarboxylase.

Zellfreier Extrakt von M. rubra decarboxyliert Malonat mit einer beträchtlichen Aktivität von 2.7 U/mg Protein. Zugabe von ATP und Acetyl-CoA zeigt keine Wirkung in frisch präpariertem Extrakt Da Malonyl-CoA zehnmal langsamer als freies Malonat umgesetzt wird, ist letzteres offensichtlich das Substrat. Indizien für einen Radikal-Mechanismus sind keine gefunden worden. Stattdessen wird Malonat Decarboxylase durch eine Acetylierung aktiviert, ein Mechanismus, der für Citrat

Lyase bekannt war. Aktives Enzym wird vollständig gehemmt durch desacetylierende Reagentien wie Hydroxylamin, Mercaptoethanol oder Thiocyanat. Diese Inhibition ist sowohl enzymatisch (mittels einer spezifischen Ligase und ATP/Acetat) als auch chemisch (mit Acetanhydrid) reversibel. Dithioerythritol erhöht die reaktivierbare

Aktivität, was auf die Beteiligung eines Thiols schliessen Iässt. Im Verlauf der Katalyse wird der Enzym-gebundene Acetyl-Rest durch einen Malonyl-Rest ersetzt. Das aktivierte Substrat der Malonat Decarboxylase ist daher ein Protein-gebundener Thioester der Malonsäure, nämlich Malonyl-Thio-Acyl-Carrier-Protein (Malonyl-S- ACP). Acetyl-S-Acyl Carrier Protein: Malonat Acyl Carrier Protein-SH Transferase katalysiert die eigentliche Aktivierung von Malonat Diese Transferase setzt als nichtphysiologische Substrate auch die entsprechenden Coenzym A (CoA)-Derivate um. Mit Malonyl-CoA und Acetat als Alternativ-Substraten ist eine lösliche CoA Transferase gereinigt worden, deren Menge 4 % des Proteins im Extrakt beträgt und die an der Malonat Decarboxylierung beteiligt ist. Das monomere Enzym besitzt ein apparentes Molekulargewicht von 67'000 und ein pH-Optimum von 5.5. Die Km- Werte für die CoA Substrate sind 1.9 mM (Malonyl-CoA) und 6.9 mM (Acetyl-CoA) und damit etwa zwei Grössenordnungen grösser als diejenigen von physiologischen CoA Transferasen. Der katalytische Mechanismus läuft nicht über ein kovalentes Transferase-CoA Intermediat, das für physiologische CoA Transferasen nachgewiesen 2

ist Der Umsatz von CoA Derivativen ohne Beteiligung eines kovalenten Enzym-CoA Intermediates ist ebenfalls von der Citrate Lyase bekannt Das Malonat Decarboxylase System enthält neben cytoplasmatischen auch Membran-gebundene Komponenten. Ausserdem ist Biotin als Cofaktor beteiligt und die Umsetzung von Malonat wird im Extrakt spezifisch durch Na+ (Km = 0.8 mM) oder Li+ (Km = 3.3 mM) stimuliert Diese Eigenschaften sind typisch für Na+-

Transport Decarboxylasen, die die Decarboxylierungs-Energie zum Aufbau eines elektrochemischen Na+-Gradienten nutzen. Beim Wachstum von M. rubra auf Malonat wird ein einziges Biotin Protein von ungewöhnlicher Grösse (120 kD) exprimiert Aufgrund biochemischer Analysen (Inhibition durch Avidin, Western Blots) als auch mittels Elektronen-Mikroskopie ist dieses Protein im Cytoplasma lokalisiert worden.

Das Malonat Decarboxylase Enzym System von Malonomonas rubra ist bezüglich der Substrat-Aktivierung (Acetylierung, ACP-Thiol Transfer) mit der Citrat Lyase verwandt, und es gleicht den Na+-Transport Decarboxylasen bezüglich der Substrat- Decarboxylierung (Carboxyltransfer auf Biotin, Decarboxylierung des C02-Biotin) und der Energiekonservierung. 3

Summary

Malonomonas rubra, a microaerotolerant, strictly Na+-dependent bacterium isolated from anoxic Sediments grows anaerobically on malonate as sole source of carbon and energy. Malonate decarboxylase, the key enzyme of this fermentation pathway, decarboxylates the Substrate quantitatively to acetate and CO2. This reaction yields only -17.4 kJ per mol and consequently, the ATP synthesis mandatorily involves chemiosmotic processes. The clecarboxylation of malonate imposes a chemical problem, since the C-C-bond has to be activated for cleavage. Aim of the work presented here was the elucidation of malonate activation and the biochemical characterization of malonate decarboxylase. Cell free extracts of M. rubra decarboxylate malonate with the considerable activity of 2.7 U/mg protein. Addition of ATP and acetyl-CoA has no effect on freshly prepared extracts. Since malonyl-CoA is decarboxylated ten times slower than free malonate, the latter is apparently the Substrate. No evidences for a radical mechanism have been found. Instead, malonate decarboxylase is activated by an acetylation reaction, a mechanism which is also known for citrate lyase. Active enzyme is completely inhibited by deacetylating reagents such as hydroxylamine, mercaptoethanol or thiocyanate. This inhibition is enzymatically (involving a specific ligase and ATP/acetate) and chemically (with acetic anhydride) reversible. Dithioerythritol increases the reactivatable acitvity, which implicates the participation of a thiol residue. During catalysis the enzyme-bound acetyl-residue is exchanged for a malonyl-residue.

The activated Substrate of malonate decarboxylase is, therefore, a protein-bound thioester of malonate, i.e. malonyl-thio-acyl carrier protein (malonyl-S-ACP). Acetyl-S-acyl carrier protein: malonate acyl carrier protein-SH transferase catalyzes the activation of malonate. This transferase also accepts the respective coenzyme A

(CoA) derivatives as nonphysiological Substrates. With malonyl-CoA and acetate as alternative Substrates a soluble transferase has been purified, which amounts to 4 % of the total protein in cell free extracts and which participates in the decarboxylation of malonate. The monomeric enzyme has an apparent molecular weight of 67'000 and a pH optimum of 5.5. The Km-values for the CoA Substrates are 1.9 mM (malonyl-CoA) and 6.9 mM (acetyl-CoA), respectively, and thus about two Orders of magnitude higher than those of physiological CoA transferases. The catalytic mechanism does not involve the formation of a covalent transferase-CoA intermediate, which has been demonstrated for the physiological transferases. CoA derivatives as alternative

Substrates and a mechanism of CoA transfer which does not involve the participation of a covalent enzyme-CoA intermediate are also known for citrate lyase. 4

The malonate decarboxylase enzyme System contains cytoplasmic and membrane- bound components. Biotin is involved as cofactor, and in cell free extracts the decarboxylation of malonate is specüically stimulated by Na+- (Km = 0.8 mM) or Li+- ions (Km = 3.3 mM). These features are typical for Na+-transport decarboxylases which use the decarboxylation energy to build up an electrochemical Na+-gradient.

Upon growth of M. rubra on malonate only one biotin protein is expressed, with the unusually high molecular mass of 120 kD. This protein has been shown to be located in

the cytoplasm by means of biochemical techniques (inhibition with avidin, Western

blot) and by means of electron microscopy.

The malonate decarboxylase enzyme System of Malonomonas rubra is related to

citrate lyase with respect to Substrate activation (acetylation, ACP-thiol transfer) and similar to the Na+-transport decarboxylases with regard to the decarboxylation of the

Substrate (carboxyltransfer to biotin, decarboxylation of a C02-biotin) and the energy conservation strategy. 5

CHAPTER I

Introduction

1.1. Physiology of anaerobic, chemotrophic bacteria

1.1.1. Bioenergetic challenges of chemotrophic anaerobes From a bioenergetic point of view the main challenge of anaerobic bacteria is the low energy yield available from only partial degradation of organic Compounds compared to the complete oxidation with dioxygen (Babcock & Wikström, 1992). The most specialized representatives of these organisms can operate close to the thermodynamic equilibrium (Schink, 1991). As a consequence, catabolism appears to be rate limiting in anaerobic growth, whereas anabolism appears to be rate limiting in aerobic growth. The anaerobes, however, cope with this competitive disadvantage by a variety of catabolic strategies and pathways, invented to exploit even minimal amounts of changes in free energy. Accordingly, a number of unique catabolic enzymes has been discovered and characterized (Gottschalk, 1986).

The formation of ATP is generally coupled to redox processes, which formally can be divided in electron-donating partial processes (dehydrogenation reactions) and electron-accepting partial processes (hydrogenation reactions). Under anaerobic conditions the electron flow associated with ATP synthesis proceeds either intermolecular (anaerobic respiration) or, if this is not possible, intramolecular

(fermentation). In electron transport phosphorylation the electrons flow down a

Potential gradient via electron carriers and finally reduce a terminal electron acceptor.

In fermentation processes, however, the excess electrons have to be consumed through the reduction of intermediate products (see Fig. 1). ATP synthesis by Substrate level phosphorylation involves the intermediate formation of "energy-rich" thioester, acidic anhydride or phosphoenol ester Compounds. These Compounds, in which the carbon atom is at the highest possible oxidation level, are formed by oxidation of aldehydes or by lysis of carboxylic acid derivatives (Thauer et al., 1977). Under physiological conditions the Compounds and their hydrolysis products are far from thermodynamic equilibrium and thus, hydrolysis is exergonic and can be coupled to work (Nicholls & Ferguson, 1992).

Anaerobes are known to use a Wide variety of terminal electron acceptors such as nitrite, nitrate, fumarate, sulfate, sulfur and carbon dioxide (Thauer et al., 1977). Furthermore, there is increasing evidence of organic-matter oxidation coupled to the reduction of Fe3+ or Mn4* (Lovley, 1991). Recently, evidence for the chemiosmotic 6

coupling of reductive dechlorination was reported for Desulfomonile tiedjei, growing on 3-chlorobenzoate and formate or H2, (Mohn & Tiedje, 1991). All terminal oxidants mentioned above are reduced organotrophically or hthotrophically, and energy is conserved by generating electrochemical gradients of H+ (Na+) by means of membrane-bound enzyme Systems.

There are a few examples where ATP synthesis is not coupled to redox processes. Some anaerobic organisms metabolize certain highly oxidized Substrates by lysis rather than by dehydrogenation and hydrogenation reactions. Examples are the deaminations of arginine and agmatine in Streptococcus faecalis, the xanthine fermentation in Clostridium cylindrosporum and pyruvate fermentation to acetate and formate in Proteus rettgeri (Thauer et al., 1977). In Streptococcus bovis a nonredox deamination of glutamine to pyroglutamate was shown to be coupled to energy production by a cyclotransferase reaction (Cook & Russell, 1993). These catabolic nonredox processes

Substrate

ATP < (SLP)

Intermediate NADH Product

> ATP (ETP) Y

Terminal Product Oxidant

Figure 1. Linear catabolic pathway (adapted after Thauer et al., 1977). The intermediate product separates an electron-donating partial process from an electron- accepting partial process. The electron-donating partial process always yields ATP by Substrate level phosphorylation (SLP). In the case of respiration the intermediate product is already the endproduct (generally C02 or acetate). Here, oxidation of NADH (or another reduced Compound) by a terminal electron acceptor is coupled to ATP synthesis by electron transport phosphorylation (ETP). If no terminal oxidant can be utilized, the electron-accepting partial process leads to a reduced organic fermentation product, where no ATP synthesis is possible. 7

are associated with Substrate level phosphorylation Chemiosmoüc mechanisms may also contnbute to energy storage by dissimilatory nonredox processes The change m free energy of a number of decarboxylation reactions is exploited to generate an electrochemical gradient of Na+ or H+ (see paragraphs 2 2 and 2 3)

The only organisms denving energy from the lithotrophic or organotrophic reduction of CO2 are the sürictly anaerobic methanogemc archaea and the acetogemc bacteria The reduction of C02 by methanogens involves a number of novel redox camers and coenzymes, such as methanofuran, tetrahydromethanoptenn (a tetrahydrofolate analogue), the thiol coenzymes M and B (2-mercaptoethanesulfonate and N-7-mercaptoheptanoylthreomne phosphate) and the mckel porphinoid factor

F430 Addiüonally, coenzyme F420 (w NAD(P) analogue), molybdenum or tungsten bound to molybdoptenn dinucleotide, a cornnoid (a Vitamine B^ analogue) and lron- sulfur cofactors are involved in this pathway The latter four classes of cofactors are, however, encountered in bactena and eucaryotes as well In spite of the unusual cofactors involved in methanogenesis, the chemical pnnciples of their function are the same as in eubactena and eucaryotes (Weiss & Thauer, 1993)

The methanogemc archaea have to cope with a bioenergetic problem the H2 partial pressure in their natural habitats is very low (1 to 10 Pa), and thus, the free energy change of the reduction of one mol CO2 with four H2 to CR» yields only -30 kJ/mol, compared to the AG°' of-131 kJ/mol Hence, a chemiosmoüc type of energy storage is mandatonly involved (Thauer, 1990)

The low H2 partial pressure in these ecosystems is maintained by the hydrogen- consuming organisms (methanogens, acetogens, sulfate reducers) At a p(H2> of below

50 Pa the reduction of H+ with NADH becomes exergonic and thus the NAD+- dependent oxidauon of reduced organic Compounds (pnmary alcohols, fatty acids, aromaüc Compounds) with protons as oxidants becomes thermodynamically feasible

These oxidations are carned out by the highly speciahzed obligately proton-reducing bactena in a consortium with hydrogen consuming organisms, which consume the reducing equivalents via interspecies hydrogen transfer The low energy yields denved from these reactions have to be shared by all members of the consortium, and therefore, as little as 12-20 kJ/reaction are avaüable for ATP synthesis, which is in the

conceivable e lowest ränge (reviewed by Schink, 1991) Syntrophomonas wolfei, g , grows from ß-oxidation of saturated C4-Cg fatty acids in mutuai Cooperation with a methanogemc or sulfidogenic organism (Mclnerney et al, 1979) Interestingly,

S wolfei can also grow in pure culture, but only on a more oxidized Substrate such as crotonate, which is dismutated to acetate and butyrate (Beaty & Mclnerney, 1987) The saturated Cg-Cjo dicarboxyhc acids were found to be ß-oxidized and decarboxylated stoichiometncally to acetate and mediane by methanogemc ennchment 8

consortia (Matthies & Schink, 1993) Figure 2 gives an overview of the trophic groups of anaerobic bactena cooperaüng in the methanogemc degradation of complex organic matter

Polymers i®

Monomers

> ^D

Fatty acids, succinate, alcohols, lactate

III

Figure 2 Overview of the steps in degradation of complex organic matter ("the anaerobic food Cham") In step I the pnmary fermenüng bactena (©) degrade polymers (proteins, polysachandes, lipids, nucleic acids, etc ) via mono- and ohgomers (pepüdes, ammo acids, sugars, acids, glycerol, punnes and pynmidines) to the classical fermentation products In step III acetate, H2 and one-carbon Compounds are converted to CH4 directly by bactena of the trophic groups ® (hydrogen-oxidizing methanogemc bactena) and G> (acetoclastic methanogemc bactena) In contrast, longer cham fatty acids, pnmary alcohols, lactate and succinate require further oxidative conversion by bactena of group ® (step II). 1 e the secondary fermenüng bactena (obligate syntrophic or obligate proton reducing bactena) Group CD represents the homoacetogenic bactena (taken from Schink, 1991) 9

1.1.2. Complete anaerobic oxidation of organic matter to C02 Anaerobic bacteria are limited in the oxidation of organic matter to CO2 by the demand to balance electron-donating and electron-accepting partial processes. In order to prevent the formation of "bioenergetically useless" reducing equivalents, anaerobically growing Enterobacteriaceae (and other bacteria) circumvent the oxidative decarboxylation of pyruvate and instead this Compound is cleaved non- oxidatively to acetyl-CoA and formate. The responsible enzyme pyruvate formate-lyase is anaerobically induced by the FNR protein (Knappe & Sawers, 1990). An iron- dependent enzyme, which uses reduced flavodoxin and S-adenosylmethionine as cofactor, activates pyruvate formate-lyase posttranslationally by generating a free carbon atom radical on a glycine residue of the protein backbone (Wagner et al., 1992). In anaerobic bacteria which exploit respiration pathways oxidative decarboxylations occur: e.g. the homoacetogenic Acetobacterium malicum employs an NAD-dependent L-malate decarboxylase (malic enzyme) (Strohhäcker & Schink,

1991) and from Pseudomonas sp. a decarboxylating glutaryl-CoA dehydrogenase is known (Härtelera/., 1993).

Thauer et al. (1977) assumed that complete oxidation of carbon Compounds via acetyl-CoA and the citric acid cycle would demand a terminal electron acceptor with a redox potential (E°') more positive than that of the fumarate/succinate couple (+ 33 mV). This requirement is met by nitrite, nitrate, Fe3+ and trithionate, but not by the other terminal oxidants used by anaerobes. Thus, most anaerobic respiration pathways do not involve the complete oxidation of acetyl-CoA, which rather is the most frequently used source of "high energy phosphate bonds" in anaerobes. ATP is

formed raainly via phosphotransacetylase and acetate kinase and, in some archaea,

through an ADP forming acetyl-CoA synthetase (Schäfer et al, 1993). In spite of this

theoretical prediction, anaerobic bacteria were discovered, realizing the oxidation of

acetyl-CoA to CO2 either with sulfur, sulfate or even protons as terminal electron acceptors (reviewed by Thauer, 1988). The hydrogenic acetate oxidation, however,

was observed only in obligate syntrophic association with H2 consumers (see above). The reduction of sulfuric Compounds has been found to proceed via two distinct

pathways: (1) a modified citric acid cycle or (2) the carbon monoxide dehydrogenase (or acetyl-CoA) pathway (reviewed by Thauer, 1988).

(1) Acetate oxidation through the modified citric acid cycle is accomplished by, e.g., the sulfate-reducer Desulfobacter postgatei and the sulfur-reducer Desulfuromonas acetoxidans. In both organisms, acetate is activated by succinyl-CoA: acetate CoA transferase and the endergonic oxidation of succinate with menaquinone is driven by ApH+. However, these two organisms employ considerably different respiratory

pathways. In D. postgatei, the terminal electron acceptor sulfate (E°' = -516 mV) is 10

adenylated to adenosine phosphosulfate (E°' = -60 mV), which significantiy increases the redox potential and makes the oxidation of menaquinone thermodynamically feasible. The ATP consumed for sulfate activation is regained via Substrate level phosphorylation by means of an ATP citrate lyase (see Fig. 3). The reduced coenzymes obtained from oxidation of acetyl-CoA via the modified citric acid cycle have a more negative average redox potential than the aerobically formed products NADH and reduced ubiquinone. Consequenüy, the energy yield upon reduction of sulfuric acid Compounds is increased. D. acetoxidans reduces unactivated sulfur with menaquinone by means of ApH+ driven reversed electron transport. Energy is conserved in the

ADP + P: ATP

Oxaloacetate 2[H, + ~j

Malate 4 NADPH

Fumarate

Ferredoxin Menaquinone \ (reduced) (reduced) "^N, Succinate

Acetate

Figure 3. Pathway of acetate dehydrogenation in Desulfobacter postgatei via an modified citric acid cycle (according to Möller et al., 1987). Acetate is activated by CoA transfer. The ATP citrate lyase reaction yields one ATP per mol acetate oxidized to 2 CO2 by Substrate level phosphorylation. Concomitantly, NADP, ferredoxin, menaquinone and a yet unidentified electron acceptor are reduced. ATP citrate (pro-3 S)-lyase and 2-oxoglutarate: ferredoxin oxidoreductase are fully reversible enzymes, in contrast to citrate synthase and the 2-oxoglutarate dehydrogenase complex, which operate in the aerobic citric acid cycle. 11

reduction of sulfur with ferredoxin. Here, the condensaüon of acetyl-CoA and oxaloacetate is catalyzed by citrate synthase and is Üierefore not coupled to ATP synthesis. (2) Some anaerobes oxidize acetyl-CoA via the carbon monoxide dehydrogenase

(or acetyl-CoA) pathway. In Desulfotomaculum acetoxidans, acetate is activated to acetyl-CoA by acetate kinase and phosphotransacetylase. Acetyl-CoA is formally spüt to CO and CH3OH by the CO dehydrogenase and the ATP consumed for the activation of acetate is regained in the 10-formyltetrahydrofolate synthetase reaction, which liberales formate from 10-formyltetrahydrofolate. The reducing equivalents obtained from the oxidation of "CO" and "CH3OH" to CO2 are used to reduce sulfuric acid Compounds and thereby ATP is synthesized chemiosmotically.

1.1.3. The role of sodium in bacteria! physiology

Sodium ions are indispensable for the growth of a wide number of bacteria, which include organisms living in saline ecosystems, such as halophilic and marine bacteria

(Tokuda, 1993), and rumen bacteria (Strobel & Rüssel, 1991). Furthermore, methanogens and acetogens (Müller et al., 1990), alkaliphilic bacteria (Ivey et al.,

1993) and some soil bacteria (Page, 1991) show an obligate requirement for this alkali ion. The observed physiological sodium-dependency of these organisms is reflected on a molecular level by numerous membrane-bound enzyme Systems which function by the employment of an inwardly directed electrochemical sodium gradient (ApNa+). Sodium extrusion is generally accomplished by the widely distributed Na+/H+ antiporter, which is crucial for osmotic balance and pH homeostasis. This AßH+- powered antiporter keeps the intracellular pH around 7.6-7.8 (neutrophils) or 8.2

(alkaliphiles). E. coli, e.g., encodes two Na+/H+ antiporters (Padan & Schuldiner, 1993). Whereas the ancillary NhaB antiporter is pH insensitive, the electrogenic NhaA antiporter is strongly alkali-induced (Gerchman et al., 1993) and furthermore, the nhaA gene is regulated positively by the Na+-stimulated NhaR protein. In some bacteria Na+-ions are exported from the cytoplasm by a variety of primary transport

Systems energized by ATP hydrolysis, redox potential or decarboxylation. Figure 4 gives a schematic overview of sodium circuits: here, the sodium motive force is built up either by primary pumps or by the Na+/H+ antiporter and is used to drive various endergonic reactions, e.g. solute uptake, flagellar rotation, or ATP synthesis.

Sodium-coupled processes are alternatives to the more widespread proton-coupled membrane reactions. The Na+-Iinked bioenergetic reactions occur predominantely in anaerobic, marine or alcaliphilic bacteria (Dimroth, 1991). Thus, the use of AfiNa+ instead of AjIH+ in energetics may be favorable especially for bacteria which have to 12

cope with the constraints of small changes in free energy and low ApH+. Since phospholipid membranes are 6-10 Orders of magnitude less permeable for Na+ than for

H+ (Gennis, 1989), bacteria encountering increased membrane permeability, e.g. thermophiles, might benefit from the utilization of sodium cycles in energy coupling (Speelmansera/., 1993b).

Figure 4. Schematic overview of sodium cycling Systems in bacteria (taken from Dimroüi, 1987). Exergonic and endergonic membrane reactions are coupled via an electrochemical sodium gradient which is built up either by primary sodium pumps such as decarboxylases (1) and respiratory oxidoreductases (4) or the Na+/H+ antiporter which is driven by an electrochemical proton gradient (2). The thus stored energy in turn drives a number of endergonic Systems such as ATP synthase (3), flagellar motors (5) and solute symports (6). Stoichiometries were not taken into account. 13

1.1.4. Primary sodium pumps

Procaryotes do not express the electrogenic, oubain-sensitive Na+-/K+-ATPase, which is ubiquitous in the cytoplasmic membrane of animal cells (Harold, 1986). Some bacteria employ, however, unique primary sodium pumps. Among these is the sodium translocating NADH: ubiquinione oxidoreductase of Üie facultative aerobe Vibrio alginolyticus and a number of other gram-negative marine bacteria (Unemoto et al.,

1990). This redox pump is not restricted to marine organisms: Klebsiella pneumoniae, anaerobically grown on citrate, contains a Na+-translocating NADH: ubiquinone oxidoreductase as well (Dimroth & Thomer, 1989). Sodium extrusion coupled to a primary pump driven by respiration has also been reported for the nonmarine, obligately aerobic bacterium Vitreoscilla. Here, AjINa+ was attributed to a cytochromeo terminal oxidase (Efiok & Webster, 1992). Another example for a primary Na+-transport System is Mycobacterium gallisepticum. Here, the sodium cycle was reported to play a vital role in the regulation of the cell volume (Schirvan et al., 1989a,b).

The key enzymes in several fermentation pathways of anaerobic bacteria are sodium transport decarboxylases (see paragraph 2.3.). The ApNa+ built up by these primary sodium pumps either plays a supporting role in the energization of the membrane or it can be the sole source of energy. The first bacterium which has been shown to gain its total energy from a sodium transport coupled decarboxylation is Propionigenium modestum, growing from the decarboxylation of succinate to Propionate. Here, a

AjINa+ generated by methylmalonyl-CoA decarboxylase drives ATP synthesis via a FiF0 ATP synthase (Hilpert et al., 1984; Laubinger & Dimroth, 1988). The F0 moiety of this ATPase was assessed to üanslocate Na+ since a hybrid consisting of F! of

E. coli and F0 of P. modestum showed the same cation specifity as the homologous

P. modestum enzyme (Laubinger et al., 1990). Only recenüy, acetogenesis and ATP synthesis in Acetobacterium woodii was shown to be coupled via a transmembrane sodium gradient which was assumed to be established by the methylenetetrahydrofolate reductase and/or the methyltransferase reaction (Heise et al., 1993). A sodium-dependent FjFq ATPase of A. woodii, related to the P. modestum enzyme, has been purified (Reidlinger & Müller, 1993).

Primary Na+-extruding ATPases exist in some stricüy fermentative bacteria as well:

Enterococcus hirae (formarly Streptococcus faecalis) possesses a unique Na+-ATPase which is distinct from other ion-motive ATPases (F-type and P-type) since this sodium pump is resistant to dicyclohexylcaAodiimide (DCCD) as well as to vanadate

(Kakinuma & Igarashi, 1989; Harold, 1986). This enzyme, which is assumed to expel Na+-ions by exchange for K+, resembles V-type ATPases, as judged from its sensitivity to nitrate and N-ethylmaleimide and from Üie sequences of the hydrophilic 14

major subunits A and B (Kakinuma & Igarashi, 1989; Takase et al., 1993). In studies with membrane vesicles from Üie thermophilic Clostridium fervidus, a Na+-pumping

F/V-type ATPase has been observed. The latter organism lacks a Na+/H+ antiporter, and thus, the narrow growth ränge (pH 6.0 to pH 8.0) may be explained. Since additionally all secondary transport Systems looked at are Na+-driven (Speelmans et al., 1993a), it has been claimed that energy transduction in C. fervidus is accomplished entirely by Na+-cycling (Speelmans et al., 1993b). An overview of primary sodium pumps in bacteria and of Na+-linked endergonic Systems is given in Table 1.

In some bacteria, respiration-linked primary electrochemical gradients are built up which consist of more than one ion species. In these respiration pathways sodium ions and protons are expelled by different enzymes of Üie respiratory chain. In vesicles of the halo- and alkalitolerant Bacillus FTU, H+- and Na+-respiratory chains have been assumed. Based on differences in inhibitor and ionophore sensitivities, Kostyrko et al.

(1991) concluded that pumps with either ion specifity occur in the initial step of the respiratory chain and in the terminal segment. Hence, sodium motive respiratory pumps may not be restricted to the NADH: quinone oxidoreductase reaction, but

Bacillus FTU could also employ a Na+-motive terminal oxidase. Vibrio alginolyticus expresses two different NADH quinone reductases, the sodium motive NADH: ubiquinone oxidoreductase (NQR-1) discussed above and NQR-2, which appears to have no energy transducing capacity (Unemoto et al., 1990). In addition, a cytochrome bo-type terminal ubiquinol oxidase of V. alginolyticus has been suggested to function as a proton pump (Miyoshi-Akiyama et al., 1993).

In methanogenic respiration, energy is conserved by means of both an electrochemical proton and a sodium gradient. The coenzyme F420-oxidizing and the cytochrome fc-containing heterodisulfide oxidoreductases from Methanosarcina strain Göl (Deppenmeier et al., 1990) and Methanosarcina barkeri (Heiden et al., 1993), respectively, üanslocate protons. On Üie other hand, N5-methyltetrahydromethanopterin: coenzyme M methyltransferase from Methanosarcina strain Göl

(Becher et al, 1992) and from Methanobacterium thermoautotrophicum (Gärtner et al, 1993) have been shown to pump sodium ions. The thus established Af£Na+ is assumed to drive Üie endergonic first step in methanogenic CO2 fixation, Üie reduction of CO2 with H2 to formyl-methanofuran (Weiss & Thauer, 1993), which is catalyzed by Üie molybdenum- or tungsten- containing formylmethanofuran dehydrogenase

(Schmitz et al, 1993). Exergonicsys tems linked to Endergonic Systems linked to

Oreanism Growth condition Na+ H+ Na+ H+

Klebsiella pneumoniae Fermentation of Oxaloacetate Citrate carner (CitS) citrate decarboxylase NADH ubiqumone oxidoreductase

Propwmgenium Fermentation of Methylmalonyl-CoA F,F0 ATPase modestum succinate decarboxylase

Enterococcus hirae Fermentation of V-type ATPase F]F0 ATPase glucose

Methanosarcina Methanogenic N5-tetrahydro- F420 and H2 Reduction of CO2 F/V-ATPase barkenl strain Göl, reduction of CO2 methanoptenn heterodisulfide with H2 to formyl- Methanobactenum coenzyme M oxidoreductase methanofuran thermoautotrophicum methyltransferase

Acetobacterium woodu Acetogemc methylenetetrahydro- F,F0 ATPase fermentaüon of folate reductase/ fructose methyltransferase

Clostndium fervidus Fermentation of F/V-type ATPase Glu-, Ser-,Arg- amino acids symporter

Vibrio alginolyttcus, Aerobic growth NADH ubiqumone Cytochrome bo-type Symporter F,F0-ATPase V parahaemolyticus oxidoreductase ubiquinoloxidase Flagellarmotor Flagellarmotor

Vitreoscüla Aerobic growth Cytochrome 0

Table 1 Bactena with pnmary sodium pumps (modifiedafterDunroth, 1991) The sodium transportdecarboxylasesare discussed in more detail in chapter2 3 A Na+/H+ antiporteris presentin allthe above bactena with the exception of Clostridium fervidusIn Propwmgenium modestum the presence of a Na+/H+ antiporterwas not determined 16

1.1.5. Endergonic Systems depending on ApNa+

The consumers of the electrochemical sodium potential built up by pnmary pumps or by the acüon of Üie Na+/H+ antiporter are secondary Na+-symport Systems and

Na+-dnven flagellar motors (Figure 4) Such motors occur in alkahphilic Bacillus sp which show stnctiy Na+-dependent moühty up to pH 11 5 (Hirota et al, 1981)

Manne Vibrio sp not only employ a Single polar flagellum, which is dnven by ApNa+, but in addition numerous ApH+-powered lateral flagella which are produced only under viscous condiüons Whereas the polar flagellum is suited for swimmmg m hquid

medium, the lateral flagella allow swarming over viscous surfaces (Atsumi et al,

1992) Vibrio alginolyticus also performs Na+ -coupled solute uptake and may even employ Na+-dependent protein üanslocation across the cytoplasmic membrane (Padan & Schuldiner, 1993)

Bactena which hve in habitats where sodium is abundant or the pH is high mainly rely on sodium symport Systems These include halophilit, manne and rumen bactena (growing at 2 5-5 M, 460 mM, and 90 mM NaCl, respecüvely), and alkahphihc bactena, growing up to pH 11 (Maloy. 1990) As an example, the manne sulfate reducer Desulfovibno salexigens was shown only recenüy to accumulate sulfate by

ApNa+ which is generated by an electrogenic Na+/H+ antiporter (Kreke & Cypionka, 1994)

Distinct Na+-symport Systems are found in all groups of bactena In E coli prohne, glutamate, senne, üireomne, mehbiose and pantoüienate are cotransported with

sodium ions (Padan & Schuldiner, 1993) Klebsiella pneumoniae expresses a Na+-

dependent citrate camer upon anaerobic growth on citrate (Dimroth & Thomer,

1989) This carner, termed CitS, has been cloned in E coli and sequenced (van der

Rest et al, 1992)

Interesüngly, Üie mehbiose symporter uses either H+, Na+ or Li+ as coupling ions,

depending on the sugar (a-, ß-galactosides or monosacchandes) cotransported

(Maloy, 1990) The feature of switching between different couplmg ions is also shared

by some pnmary pumps The FjF0 ATPase from Propiomgenium modestum Switches from Na+ to H+ pumping at a sodium concenlraüon below 1 mM (Laubinger &

Dimroth, 1989) In the ubiquitous eucaryoüc Na+/K+-ATPase and in Üie gastnc

H+/K+-ATPase protons Substitute for sodium and vice versa (Polvani et al, 1989) An attractive hypothesis to explain the alternative transport of protons or sodium ions was

proposed by Boyer (1988) Since crown ethers form very similar complexes with

H-ifJ* or Na+, the transport of both ions across the membrane could mvolve the

formaüon of coordinaüon complexes with appropnately placed oxygen or nitrogen

atoms within üie transmembrane part of the pump In the case of the mehbiose camer,

üie sugar Compound cotransported could be involved in üie formaüon of this 17

coordination complex. As a consequence, Üie sugar-dependent versaülity of cation specifity would be easily explained, since each sugar would contribute a specific coordination sphere (Maloy, 1990).

1.2. Conservation of decarboxylation energy

1.2.1. Fermentation of malonate and other saturated dicarboxylates

Strict fermentative growth (i.e. redox balanced growth) on saturated dicarboxylic acids imposes a severe bioenergetic problem, since there are only a few exergonic dissimilatory reactions conceivable. Non-oxidative decarboxylation reactions provide a means for a redox neutral catabolism of these Compounds. However, Üie overall change in free energy per mol of dicarboxylate decarboxylated is very low (AG0'« -17-

27 kJ/mol) and it amounts to only a fraction needed for the synthesis of one mol of

ATP under physiological conditions (AG' = 50-70 kJ/mol, Thauer. 1977; Schink,

1991). As a consequence, Üie stoichiometric formation of "energy-rich" (mixed) acidic anhydrides or thioesters is thermodynamically not feasible in these catabolic pathways, and Üierefore, chemiosmoüc processes are mandatory.

Under anaerobic conditions one mol ATP synmesized yields about 10 g dry cell mass (Thauer, 1977). Hence, AG°'-values in the ränge of 17-27 kJ/mol are expected to yield about 1-2 g dry cell mass per mol Substrate decarboxylated. This prediction is in accordance with the growth yields reported for the organisms presented below (0.5-

2.5 g dry cell mass/mol Substrate degraded).

Fermentative energy conservation through decarboxylation is known for dicarboxylic acids with 2-5 carbon atoms. Beyond this chain length dicarboxylates (C6- Ciq) have been reported to be ß-oxidized in coculture with methanogens and decarboxylated at the glutaryl-CoA or succinyl-CoA level (Matthies & Schink, 1993; see paragraph 1.1.). A number of organisms has been isolated which grow anaerobically on dicarboxylates, thereby exploiting decarboxylation reactions as Üie sole source of energy:

Oxalobacter sp. (Allison et al., 1985; Dehning & Schink, 1989b) and Clostridium oxalicum (Dehning & Schink, 1989b) decarboxylate Oxalate. Malonomonas rubra (Dehning & Schink, 1989a), Citrobacter diversus, Klebsiella pneumoniae (Janssen, 1991; Janssen & Harfoot, 1992) and Üie homoacetogenic Sporomusa termitida and

S. malonica (Breznak et al., 1988; Dehning et al., 1989) grow on malonate. Succinate is decarboxylated by S. termitida, S. malonica, Propiomgenium modestum (Schink &

Pfennig, 1982) and strain WoGl 3 (Matthies & Schink, 1992a), which also grows by decarboxylation of glutarate. 18

The decarboxylases involved in decarboxylation of dicarboxylates can be divided

into two groups: (1) soluble decarboxylases, which do not contain biotin, and (2)

membrane-bound, biotin-containing decarboxylases (see Table 2 for an overview).

Decarboxylase

Bacterium Fermentation Substrate Cofactor Subunit pathway composition

Oxalobacter TPP Oxalate CO-SCoA «4 formigenes i 1 65 kD Formate coo-

Malonomonas Malonate CO-S-ACP Biotin aß rubra 4. 1 120, 67 kD Acetate CH2 (and others) 1 BCP:a coo- ACP: ß?

Propiomgenium Succinate CO-SCoA Biotin n.p. modestum i 1

• Propionate H^C-CH

coo-

WoG13 Glutarate CO-SCoA Biotin n.p. i 1 Butyrate, CH II isobutyrate, CH (acetate) 1 CH2 1 coo-

Table 2. Anaerobic bacteria growing exclusively by decarboxylation of saturated dicarboxylic acids. Abbreviaüons: TPP, thiamine pyrophosphate; BCP, biotin carrier protein; ACP, acyl carrier protein; n.p. not purified. 19

Oxalyl-CoA decarboxylase of O fornugenes (Baetz & Allison, 1989) and Üie postulated malonyl-CoA decarboxylase of C diversus (Janssen & Harfoot, 1992)

belong to the first group The second group consists of mefhylmalonyl-CoA

decarboxylase of P modestum (Hilpert et al, 1984), glutaconyl-CoA decarboxylase of strain WoGl 3 (Matthies & Schink, 1992b) and malonate decarboxylase of M rubra (see Chapter II-V) Interestingly, stiain WoGl 3 also harbors methylmalonyl-CoA

decarboxylase, which is speciflcally mduced upon growth on succinate (Matthies & Schink, 1992b)

Figure 5 depicts the different strategies of energy conservation by soluble and

membrane-bound decarboxylases The decarboxylation of a subsüate by a soluble

decarboxylase generates a high product gradient which dnves an electrogenic

substrate/product antiporter Additionally, a proton is consumed m Üie course of the

decarboxylation reaction Hence, decarboxylation energy is stored as ApH+ (see

Paragraph 2 2 ) In conüast, the membrane-bound decarboxylases are pnmary sodium

pumps which störe energy as ApNa+ (see paragraph 2 3)

1.2.2. Soluble decarboxylases and substrate/product antiport

Oxalobacter fornugenes grows from the decarboxylation of Oxalate to formate

Decarboxylation energy is conserved by means of a soluble oxalyl-CoA decarboxylase and the Oxalate formate antiporter The effknent oxalyl-CoA decarboxylase generates

a formate gradient, which dnves the electrogenic Oxalate2 formate1 antiporter Smce one scalar H+ is consumed per decarboxylation event in the cytoplasm, a vectonal

ApH+ is built up Hence, in this case, net extrusion of H+ is accomplished with secondary rather than with pnmary acüve transport (Anantharam et al, 1989) The oxalate/formate antiporter (OxlT) has been punfied and reconstituted into proteohposomes It has an unusually high transport turnover number of about 1000/s (Rum etat, 1992)

Recenüy, a number of funcüonally related antiport Systems have been reported in lacüc acid bactena In Lactobacillus sp the fermentation of malate to lactate (malo- lacüc fermentation) is dnven by a cytoplasmically located, avidin-insensitive malate decarboxylase (malolactic enzyme) which is the only energy yielding reaction The decarboxylaüon energy is assumed to be stored through an electrogenic malate/lactate antiport (Poolman, 1990, Kolb et al, 1992) Lactococcus lactis subsp lactis biovar deacetylactis grows on citrate as sole source of energy The punfied oxaloacetate decarboxylase involved in this degradaüon is soluble, avidin-insensitive and Na+ mdependent It has been postulated that energy is stored by an electrogenic citrate acetate/pyruvate exchange (Hugenholtz et al, 1993) 20

A) Soluble decarboxylase (D) and anion antiporter (A)

B) Membrane-bound, sodium-transport decarboxylase

Na+

Figure 5 Ditferent strategies to conserve decarboxylaüon energy O formigenes employs a soluble decarboxylase and a decarboxylaüon product-dnven antiporter (A) Here, energy is stored as electrochemical proton gradient P modestum uses a membrane-bound, biotin-containing Na+-transport decarboxylase (B) which generates an electrochemical sodium polenüal 21

Moreover, a number of biogenic amine and polyamine antiporters are known in both

Enterobacteriaceae and Lactobacteriaceae. In these cases Üie substrate/product couples are connected either with decarboxylation reactions or with deimination reactions (i.e. the formal release of urea), where the latter also yield ATP by Substrate level phosphorylation (Poolman, 1990). Ureaplasma urealyticum, a mycoplasma, employs a cytoplasmic urease which couples Üie hydrolysis (i.e. Üie decarboxylation) of urea to the generation of a membrane potential (Smith et al., 1993).

1.2.3. Decarboxylation-linked primary sodium pumps Oxaloacetate decarboxylases from Klebsiella pneumoniae and Salmonella typhimurium (Wifling & Dimroth, 1989) are induced upon anaerobic fermentation of citrate (Antranikian & Giffhorn, 1987). The identification of oxaloacetate decarboxylase from Klebsiella pneumoniae as a membrane-bound Na+-ttanslocating biotin enzyme (Dimroth, 1980) led to the discovery and extensive characterization of a family of related decarboxylases which are the key enzymes in various fermentations of gram-negative as well as gram-positive bacteria. The enzymes of this group share the following properties: (1) localization in Üie cytoplasmic membrane, (2) biotin as prosthetic group, (3) Stimulation by sodium ions, (4) conversion of decarboxylation energy into ApNa+ and (5) a similar subunit composition and catalytic mechanism (for a review see Dimroth, 1987). These decarboxylases have been purified by monomeric avidin-Sepharose affinity chromatography (Dimroth, 1986) and they were functionally reconstituted by incorporation into phospholipid vesicles (Dimroth, 1986; Buckel & Semmler, 1983; Hilpert & Dimroth, 1984; Wifling & Dimroth, 1989). The oxaloacetate decarboxylase complex has also been reconstituted from Üie isolated subunits (Dimroth & Thomer, 1988). Whereas the biochemical characterization of the Na+-ttansport decarboxylases was performed in the last decade, data from genetic work have come up only recenüy. Oxaloacetate decarboxylase of K. pneumoniae and S. typhimurium is composed of three different subunits (aßy), as shown in Figure 6 (Dimroth, 1990). Subunit a is a peripheral membrane protein, comprising two domains. The C-terminal domain carries the biotin cofactor covalenüy attached to the lysine residue of the consensus sequence

AMKM 35 amino acids upstream from the C-terminus. The N-terminal carboxyltransferase domain harbors Üie oxoacid binding-site. A unique, extended alanine/proline rieh Stretch is assumed to convey Üie flexibility to Üie biotin moiety which is required for its movement between the two catalytic centres on Üie a and ß subunits, respectively (Schwarz et al., 1988; Woehlke et al., 1992a). Subunits ß and 22

Y are integral membrane proteins, which are assumed to traverse the membrane with 9

helices (ß) or one helix (y) (Laussermair et al., 1989; Woehlke et al., 1992a,b).

The a subunit catalyzes Üie Na+-independent but avidin-sensitive carboxyl-transfer

reaction from oxaloacetate to the biotin moiety, yielding carboxybiotin and pyruvate. The carboxybiotin-enzyme intermediate involved in oxaloacetate decarboxylation has been demonstrated biochemically by 14C02 transfer from [4-14C] oxaloacetate (Dimroth, 1982a). The integral membrane proteins ßy (or ß alone) catalyze the sodium

dependent decarboxylation of Üie carboxybiotin, which is coupled to the electrogenic

translocation of two Na+-ions over üie membrane. A detailed kinetic analysis of the

decarboxylation reaction revealed a mechanism analogous to the ü'anscarboxylase of

Propionibacterium shermanii (Dimroth & Thomer, 1986). This kinetic mechanism was

termed hybrid Ping-Pong-Rapid Equilibrium Random and it takes place in a two Site BiBi system (Segel, 1975). According to this model, oxaloacetate decarboxylase

harbors two different and independent catalytic Sites that are functionally connected by the carboxybiotin intermediate. The two Substrates oxaloacetate and (Na+in/H+) and üie two products pyruvate and (Na+out/C02) are bound and released randomly (Dimroth & Thomer, 1986).

Periplasm

Pyruvate

Oxaloacetate

Cytoplasm

Figure 6. Hypothetical model of oxaloacetate decarboxylase subunits linking structure and function. B represents the biotin cofactor. 23

The reversibility of sodium transport decarboxylases has been demonstrated elegantely with a vectorial ü'anscarboxylase System consisting of oxaloacetate decarboxylase of Klebsiella pneumoniae and methylmalonyl-CoA decarboxylase of Veillonella parvula, co-reconstituted into Üie same vesicles (Dimroth & Hilpert, 1984). The Na+-circuit thus established mediates the transcarboxylation from oxaloacetate and acetyl-CoA to pyruvate and malonyl-CoA and vice versa (see Figure 7). The endergonic carboxylation is accomplished via the AjINa+ generated in the decarboxylation reaction. Reversible energetic coupling of two sodium pumps has also been observed with methylmalonyl-CoA decarboxylase and the F1F0 ATPase of Propiomgenium modestum (Hilpert et al., 1984).

Figure 7. Na+-circuit mediaüng the transcarboxylation from oxaloacetate and acetyl-CoA to pyruvate and malonyl-Coa and vice versa (taken from Dimroth & Hilpert, 1984).

Other well characterized members of Üie sodium-translocating decarboxylase enzyme family include Üie following: methylmalonyl-CoA decarboxylase from lactate fermenüng Veillonella parvula (Hilpert & Dimroth, 1983) or from succinate fermenting Propiomgenium modestum (see paragraph 1.3.) and glutaconyl-CoA decarboxylase from Üie gram-positive Acidaminococcus fermentans, growing anaerobically on glutamate (Buckel & Semmler, 1983). Other gram-positive glutamate fermenting bacteria such as Clostridium symbiosum or Peptostreptococcus 24

asaccharolyticus and Üie gram-negative Fusobacterium nucleatum also contain glutaconyl-CoA decarboxylase (Beatrix et al., 1990). Although Üie subunit composition of these decarboxylases varies from 3-5 components (see Table 3 for an overview), the overall molecular mass of these enzyme

Bacterium, Fermentation Activated Cofactor Subunit com¬ Decarboxylase Pathway Intermediate position of the decarboxylase

Klebsiella Citrate coo- Biotin, aßy pneumoniae, i | Zn2+ Zn2+ 63, 45, 9 kD oxaloacetate Acetate, c=o BCP:a decarboxylase formate, 1 C02 CH2 1 coo-

Veillonella Lactate CO-SCoA Biotin aßvoe parvula, 1 1 55,39, 13, 12, methylmalonyl- Propionate, H,C-CH 6kD CoA acetate, BCP:y coo- decarboxylase C02, H2

Acidaminococ- Glutamate CO-SCoA Biotin aßy8 cus fermentans, i 1 64, 33, 24, glutaconyl-CoA Butyrate, CH 14 kD II decarboxylase acetate, BCP:y CH C02, H2, 1 NH4+ CH2 1 coo-

Table 3. Well characterized members of Üie biotin-containing, Na+-üanslocating decarboxylase enzyme family. The molecular masses of Üie subunits are deduced from the corresponding DNA sequences with the excepüon of glutaconyl-CoA decarboxylase, of which only subunit a is sequenced (Bendrat & Buckel, 1993). In the latter case Üie subunit nomenclature given in Beatrix et al. (1990) is used. BCP: biotin carrier protein. 25

complexes seems to be conserved (127 (± 10) kD) and the catalytic mechanism is very similar (Dimroth, 1990). The a subunits (55-64 kD) represent the carboxyltransferases

(Dimroth & Thomer, 1983; Buckel & Liedtke, 1986; Hoffman et al., 1989) and the ß subunits (33-45 kD) catalyze the Na+-dependent decarboxylation, i.e. Üie carboxylyase reactions. The biotin carrier protein exists either as a domain of the a subunit (oxaloacetate decarboxylase) or as an individual Polypeptide of 13-24 kD (Üie other decarboxylases). The biotin carrier of glutaconyl-CoA decarboxylase

(24 kD) could be visualized on SDS gels neither with Coomassie nor with silver stain and has been identified by affinity staining with streptavidin-alkaline Phosphatase (Beatrix et al, 1990). No specific function has yet been assigned to the small hydrophobic subunits of oxaloacetate decarboxylase (y), methylmalonyl-CoA decarboxylase (5e) and glutaconyl-CoA decarboxylase (8).

The above decarboxylases not only üansport Na+, they are also sümulated specifically by this alkali ion. The apparent Km for Na+ is about 1 mM (0.6-1.5 mM) or 3 mM (F. nucleatum). Furthermore, sodium ions specifically protect the membrane- embedded ß subunit from tryptic degradation (Dimroth & Thomer, 1983; Buckel &

Semmler, 1983; Hoffman et al., 1989) and inactivation by n-alkanols (Buckel &

Liedtke, 1986; Dimroth & Thomer, 1992). As a consequence, the ß subunits are believed to carry a Na+-binding site which when occupied leads to a more resistant (compact) conformation of this protein.

Chemically, decarboxylation reactions often involve a heterolytic cleavage by which the electron pair remains at the organic residue and subsequenüy becomes protonated:

O ii

R - C - O" > R + O = C = O

^H+

Elecüon withdrawing substituents on R thus stabilize the intermediate R" and facilitate the decarboxylation reaction. There are many different ways by which such elecüon withdrawing substituents are introduced in biological Systems to make the decarboxylation reaction chemically feasible, some of which apply to the examples relevant to this chapter. 26

Oxaloacetate decarboxylase as ß-ketoacid already carries an electron attracting

carbonyl group. In Üie catalytic center of Üie enzyme, Üie Zn2+ atom acting as a Lewis acid further polarizes the carbonyl group which facilitates the release of CO2. Another

biochemical strategy to activate decarboxylation of dicarboxylic acids is to create an

electton withdrawing thioester group in ß position to Üie carboxyl group. This is

accomplished in the decarboxylation of succinate. Succinate is first converted to

succinyl-CoA and Üie carbon skeleton is then rearranged to bring Üie thioester group

into Üie ß position of the carboxyl group to be liberated. The strategy of glutamate decarboxylation involves the same principle: the amino acid is converted to glutaconyl-

CoA which is a vinylogue of malonyl-CoA. The primary structure around (he biotin attachment sile of biotin proteins is strongly

conserved. Almost all biotin enzymes contain the tetrapeptide Ala-Met-Lys-Met, 35

amino acid residues upsüeam from Üie C-terminus of a biotin carrier protein (Samols

et al, 1988; Schwarz et al., 1988; Woehlke et al, 1992a; Huder & Dimroth, 1993).

The biotin is attached to Üie e-amino group of the lysine residue via amide linkage, yielding biocytin. In Escherichia coli, Üie ligase which carries out this attachment is

termed BirA (a homodimer of twice 34 kD) and the covalent linkage between apo-

BCCP (the only biotin protein) and the biotin cofactor is formed via biotinyl-AMP as

the intermediate. Interestingly, BirA is not only an enzyme but also a transcriptional

repressor. Its corepressor is biotinyl-AMP (the product of Üie first half-reaction of the

ligase reaction). Upon binding of the BirA-biotinyl-AMP complex to the Operator sequence of Üie Wo-operon, biotin biosynthesis is repressed (Cronan, 1989).

All biotin carboxyl carrier proteins (BCCP) contain a conserved proline (and

alanine) residue in a region about 25-30 amino acids upstream of the biocytin (Samols

et al, 1988; Huder & Dimroth, 1993). Additionally, the y subunit of methylmalonyl-

CoA decarboxylase of Veillonella parvula and (he a subunit of oxaloacetate decarboxylase from K. pneumoniae and S. typhymurium contain extended proline/alanine Imkers of about 25 residues (Huder & Dimroth, 1993). Two conserved

prolines in proximity of Üie prostheüc group have also been found in Üie ACP of citrate lyase and citramalate lyase (Dimroth, 1988). Moreover, the amino acid

sequence of BCCP is similar to those of the lipoyl domains of enzymes such as

pyruvate dehydrogenase, a-ketoglutarate dehydrogenase or the glycine cleavage enzyme complex (Toh et al, 1993), where lipoic acid is attached to a specific lysine residue. These sequences may all serve the same funtional need, namely the oscillation of a chemical group bound to a cofactor (C02-biocytin, acyl-ACP, acetyl-lipoamid) between different catalytic Sites. Hence, in a family of cofactors (biotin, 4'- phosphopanteüieine, 5"-phosphoribosyl-2'-dephospho-CoA and lipoic acid) not only 27

extended sidechains convey flexibility, but additionally, (he protein backbone contributes to the overall mobility. Extensive homologies also exist between subunits and domains of biotin proteins that bind the same Substrate and catalyze the same partial reaction. For example, the CoA ester binding sites of the 12 S subunit of transcarboxylase of Propionibacterium shermanii show homology to the a subunit of methylmalonyl-CoA decarboxylase and to Üie ß subunit of propionyl-CoA carboxylase. These proteins catalyze Üie same partial reaction, i.e. Üie carboxyltransfer from C02-biotin to propionyl-CoA (Samols et al, 1988; Huder & Dimroth, 1993). The 5 S subunit of transcarboxylase, which catalyzes the carboxyltransfer from C02-biotin to pyruvate, yielding oxaloacetate, is homologous to the a subunit of oxaloacetate decarboxylase and to pyruvate carboxylase (Samols et al, 1988; Schwarz et al, 1988; Woehlke et al, 1992a). Interestingly, both the 5 S subunit of transcarboxylase and oxaloacetate decarboxylase

(presumeably the y subunit) contain Zn2+ (Dimroth & Thomer, 1992; M. DiBerardino, personal communicaüon). Furthermore, Üie hydrophobic ß subunits of Üie Na+- transport decarboxylases of K. pneumoniae, S. typhimurium and V. parvula are related by long süetches of complete sequence identity, including two conserved aspartate residues within putative membrane-spanning helices (Laussermair et al., 1989; Woehlke et al, 1992a,b; Huder & Dimroth, 1993). At last, also the biotin carboxylase subunits or domains of a variety of carboxylases clearly show a close relationship (Toh et al, 1993).

1.3. Aims of the work

Malonomonas rubra, a microaerotolerant anaerobic gram-negative bacterium has recenüy been isolated from anoxic marine Sediments (Dehning & Schink, 1989). The organism is moüle, non-sporeforming and it shows a highly specialized physiology since only malonate, fumarate and malate are fermented. At least 150 mM sodium is required for growth with either subsüate. Malonate as sole source of energy and carbon is stoichiometrically decarboxylated to acetate with a concomitant growth yield of about 2 g dry cell matter per mol malonate degraded. The small free energy change of the decarboxylation reaction (AG01 = -17 kJ/mol) does not allow net ATP synthesis via subsüate level phosphorylation and the bacteria do not reduce external electron acceptors, albeit high amounts of a periplasmic cytochrome c and a small amounts of membrane-bound cytochrome b have been detected. Thus, upon growth on malonate, decarboxylation phosphorylation is considered as ATP synthesis mechanism. 28

The decarboxylation of malonate, which at physiological pH predominates as dicarboxylate (pKi 2.85, pK2 5.65), imposes a chemical problem since the C-C-bond is not activated for decarboxylation. Possible routes of decarboxylation are either a homolytic cleavage of Üie C-C-bond involving a radical mechanism or a heterolyüc mechanism, which demands the stabilization of the resulting carbanion by an elecüon wiüidrawing substituent

Object of Üie work presented here was to elucidate the süategy of activation of malonate for decarboxylation. Furthermore, the enzymic components of malonate decarboxylase had to be identified and characterized biochemically. Of special interest was (he question whether this enzyme belongs to the membrane-bound Na+- translocating decarboxylase family or whether some other kind of chemiosmotic energy storage is employed in Üie bioenergetics of the "red malonate consumer". 29

1.4. References

Allison, M. J., Dawson, K. A., Maybenry, W. R. & Foss, J. G. (1985) Oxalobacter

formigenes gen. nov., sp. nov.: oxalate-degrading anaerobes that inhabit the gastrointestinal tract. Arch. Microbiol. 141: 1-7.

Anantharam, V., Allison, M. J. & Maloney, P. C. (1989) Oxalate: formate exchange -

The basis for energy coupling in Oxalobacter. J. Biol. Chem 264:7244-7250. Antranikian, G. & Giffhorn, F. (1987) Citiate metaboUsm in anaerobic bacteria. FEMS

Microbiol. Rev. 46: 175-198.

Atsumi, T., McCarter, L. & Imae, Y. (1992) Polar and lateral flagellar motors of

marine Vibrio are driven by different ion-motive forces. Nature 355:182-184. Babcock, G. T. & Wikström, M. (1992) Oxygen activation and the conservation of

energy in cell respiration. Nature 356: 301-309. Baetz. A. L. & Allison, M. J. (1989) Purification and characterization of oxalyl-

coenzyme A decarboxylase from Oxalobacter formigenes. J. Bacteriol. 171: 2605- 2608.

Beatrix, B., Bendrat, K., Rospert, S. & Buckel, W. (1990) The biotin-dependent

sodium ion pump glutaconyl-CoA decarboxylase from Fusobacterium nucleatum (subsp. nucleatum). Arch. Microbiol. 154: 362-369.

Beaty, P. S. & Mclnerney, M. J. (1987) Growth of Syntrophomonas wolfei in pure

culture on crotonate. Arch. Microbiol. 147: 389-393.

Becher, B., Müller, V. & Gottschalk, G. (1992) N5-methyl-tetrahydromethanopterin:

coenzyme M methyltransferase of Methanosarcina strain Göl is an Na+- translocating membrane protein. /. Bacteriol. 174: 7656-7660.

Bendrat, K. & Buckel, W. (1993) Cloning, sequencing and expression of the gene

encoding Üie carboxytransferase subunit of Üie biotin-dependent Na+ pump glutaconyl-CoA decarboxylase from Acidaminococcus fermentans in Escherichia coli. Eur. J. Biochem 211: 697-702.

Boyer, P. D. (1988) Bioenergetic coupling to protonmotive force: should we be

considering hydronium ion coordination and not group protonation? Trends Biochem. Sei. 13: 5-7.

Breznak, J. A., Switzer, J. M. & Seitz, H.-J. (1988) Sporomusa termitida sp. nov., an H2/C02-utilizing acetogen isolated from termites. Arch. Microbiol. 150: 282-288.

Buckel, W. & Liedtke, H. (1986) The sodium pump glutaconyl-CoA decarboxylase

from Acidaminococcus fermentans - Specific cleavage by n-alcohols. Eur. J.

Biochem 156: 251-257. 30

Buckel, W. & Semmler, R. (1983) Purification, characterization and reConstitution of

glutaconyl-CoA decarboxylase, a biotin-dependent sodium pump from anaerobic bacteria. Eur. J. Biochem 136: 427-434.

Cook, G. M. & Russell, J. B. (1993) The glutamine cyclotransferase reaction of

Streptococcus bovis: a novel mechanism of deriving energy from non-oxidative and non-reductive deamination. FEMS Microbiol. Lett. 111: 263-268.

Cronan, J. E. jr. (1989) The E. coli bio Operon: transcriptional repression by an

essential protein modification enzyme. Cell 58:427-429.

Dehning, I. & Schink, B. (1989a) Malonomonas rubra gen. nov. sp. nov., a microaerotolerant anaerobic bacterium growing by decarboxylation of malonate. Arch. Microbiol. 151: 427-433.

Dehning, I. & Schink, B. (1989b) Two new species of anaerobic oxalate-fermenting

bacteria, Oxalobacter vibrioformis sp. nov. and Clostridium oxalicum sp. nov., from sediment samples. Arch. Microbiol. 153: 79-84.

Dehning, L, Stieb, M. & Schink, B. (1989) Sporomusa malonica sp. nov., a

homoacetogenic bacterium growing by decarboxylation of malonate or succinate. Arch. Microbiol. 151: 421-426.

Deppenmeier, U., Blaut, M., Mahlmann, A. & Gottschalk, G. (1990) Reduced

coenzyme F420: heterodisulfide oxidoreductase, a proton-translocating redox System in methanogenic bacteria. Proc. Natl. Acad. Sei. USA 87: 9449-9453.

Dimroth, P. (1980) A new sodium transport System energized by the decarboxylation of oxaloacetate. FEBSLett. 122: 234-236.

Dimroth, P. (1982) The role of biotin and sodium in Üie decarboxylation of oxaloacetate by the membrane-bound oxaloacetate decarboxylase from Klebsiella pneumoniae. Eur. J. Biochem. 121: 435-441. Dimroth, P. (1986) Preparation, characterization, and reconstitution of oxaloacetate

decarboxylase from Klebsiella aerogenes, a sodium pump. Meth. Enzymol. 125: 530-540.

Dimroth, P. (1987) Sodium ion transport decarboxylases and other aspects of sodium ion cycling in bacteria. Microbiol. Rev. 51: 320-340. Dimroth, P. (1988) The role of Vitamins and their carrier proteins in citrate

fermentation, in The Roots of Modern Biochemistry, pp. 191-204, Kleinkauf, von Döhren, Jaenicke (eds.). W. de Gruyter & Co, Berlin, New York.

Dimroth, P. (1990) Mechanisms of sodium transport in bacteria. Phil. Trans. R. Soc.

Lond. B 326: 465-477.

Dimroth, P. (1991) Na+-coupled alternative to H+-coupled primary tiansport Systems in bacteria. BioEssays 13: 463-468. 31

Dimroth, P. & Hüpert, W. (1984) Carboxylation of pyruvate and acetyl-CoA by a

reversal of the Na+ pumps oxaloacetate decarboxylase and methylmalonyl-CoA decarboxylase. Biochemistry 23:5360-5366. Dimroth, P. & Thomer, A. (1983) Subunit composition of oxaloacetate decarboxylase

and characterization of the a chain as carboxyltransferase. Eur. J. Biochem. 137:

107-112.

Dimroth, P. & Thomer, A. (1986) Kinetic analysis of the reaction mechanism of oxaloacetate decarboxylase from Klebsiella pneumoniae. Eur. J. Biochem. 156: 157-162.

Dimroüi, P. & Thomer, A. (1988) Dissociaüon of Üie sodium-ion-translocating oxaloacetate decarboxylase of Klebsiella pneumoniae and reconstitution of the active complex from the isolated subunits. Eur. J. Biochem 175:175-180.

Dimroüi, P. & Thomer, A. (1989) A primary respiratory Na+ pump of an anaerobic bacterium: the Na+-dependent NADH: quinone oxidoreductase of Klebsiella pneumoniae. Arch. Microbiol 151:439-444. Dimroth, P. & Thomer, A. (1990) Solubüizaüon and reconstitution of the Na+- dependent citiate carrier of Klebsiella pneumoniae. J. Biol. Chem. 265:7221-7224. Dimroth, P. & Thomer, A. (1992) The sodium ion pumping oxaloacetate

decarboxylase of Klebsiella pneumoniae - Metall ion content, inhibitors and proteolytic degradation studies. FEBS Lett. 300: 67-70. Efiok, B. J. S. & Webster, D. A. (1992) Sodium-coupled ATP synthesis in Üie bacterium Vitreoscilla. Arch. Biochem. Biophys. 292: 102-106. Gärtner, P., Ecker, A., Fischer, R., Linder, D., Fuchs, G. & Thauer, R. K. (1993) Purification and properties of N5-methyltetrahydromethanopterin: coenzyme M methyltransferase from Methanobacterium thermoautotrophicum. Eur. J. Biochem. 213: 537-545.

Gennis, R. B. (ed.) (1989), in Biomembranes: Molecular Structure and Function,

p. 265. Springer, New York. Gerchman, Y., Olami, Y., Rimon, A, Taglicht, D., Schuldiner, S. & Padan, E. (1993)

Histidine-226 is pari of the pH sensor of NhaA, a Na+/H+ antiporter in Escherichia

coli. Proc. Natl. Acad. Sei. USA 90:1212-1216.

Gottschalk, G. (ed.) (1986) in Bacteriol Metabolism, second edition, pp. 208-282. Springer, New York, Berlin, Heidelberg, Tokyo. Harold, F. M. (ed.) (1986) in The Vital Force: A Study ofBioenergetics, pp. 137-141, 318. W. H. Freemann and Company, New York.

Härtel, U., Eckel, E., Koch, J., Fuchs, G., Linder, D. & Buckel, W. (1993) Purification

of glutaryl-CoA dehydrogenase from Pseudomonas sp., an enzyme involved in the anaerobic degradation of benzoate. Arch. Microbiol 159: 174-181. 32

Heiden, S., Hedderich, R„ Setzke, E. & Thauer, R. K. (1993) Purification of a cytochrome b containing H2: heterodisulfide oxidoreductase complex from membranes of Methanosarcina barkeri. Eur. J. Biochem. 213: 529-535. Heise, R., Müller, V. & Gottschalk, G. (1993) Acetogenesis and ATP synthesis in

Acetobacterium woodii are coupled via a transmembrane primary sodium ion gradient. FEMS Microbiol. Lett. 112: 261-268.

Hilpert, W. & Dimroüi, P. (1983) Purification and characterization of a new sodium-

transport decarboxylase - Methylmalonyl-CoA decarboxylase from Veillonella alcalescens. Eur. J. Biochem. 132: 579-587.

Hilpert, W. & Dimroth, P. (1984) Reconstitution of Na+-üansport from purified methylmalonyl-CoA decarboxylase and phospholipid vesicles. Eur. J. Biochem

138: 579-583.

Hüpert, W., Schink, B. & Dimroth, P. (1984) Life by a new decarboxylaüon-

dependent energy conservation mechanism with Na+ as coupling ion. EMBO J. 3: 1665-1670.

Hirota, N., Kitada, M. & Imae, Y. (1981) Flagellar motors of alkalophilic Bacillus are

powered by an electrochemical potential gradient of Na+. FEBS Lett. 132: 278-280. Hoffmann, A., Hüpert, W. & Dimroth, P. (1989) The carboxylüansferase activity of the sodium-ion-translocating methylmalonyl-CoA decarboxylase of Veillonella

alcalescens. Eur. J. Biochem. 179: 645-650.

Huder, J. B. and Dimroth, P. (1993) Sequence of Üie sodium ion pump methylmalonyl- CoA decarboxylase from Veillonella parvula. J. Biol. Chem 268: 24564-24571.

Hugenholtz, J., Perdon, L. & Abee, T. (1993) Growth and energy generation by Lactococcus lactis subsp. lactis biovar diacetylactis during citiate metabolism. Appl. Environ. Microbiol. 59: 4216-4222. Ivey, D. M., Guffanti, A. A. & Krulwich, T. A. (1993) The Na+ cycle in alcaliphilic

Bacillus species, in Alkali Cation Transport Systems in Prokaryotes, pp. 101-124, Bakker, E. P. (ed.). CRC Press, Boca Raton, Ann Arbor, London, Tokyo.

Janssen, P. H. (1991) Growth of Enterobacteria on malonate under stricüy anaerobic conditions. System. Appl. Microbiol. 14: 93-97. Janssen, P. H. & Harfoot, C. G. (1992) Anaerobic malonate decarboxylation by Citrobacter diversus. Arch. Microbiol. 157: 471-474.

Kakinuma, Y. & Igarashi, K. (1989) Sodium-translocating adenosine triphosphatase in Streptococcus faecalis. J. Bioenerg. Biomembr. 21: 679-692. Knappe, J. & Sawers, G. (1990) A radical-chemical route to acetyl-CoA: the

anaerobically induced pyruvate formate-lyase System of Escherichia coli. FEMS Microbiol. Rev. 75: 383-398. 33

S B & B conservation in malolactic Kolb, , Otte, H, Nagel, Schink, (1992) Energy fermentation by Lactobaallus plantarum and Lactobacillus sake Arch Microbiol 157 457-463

V A A L V I A M L Kostyrko, , Semeykma, , Skulachev, P, Smirnova, , Vaghina, & Verkhovskaya, M L (1991) The H+-motive and Na+-motive respiratory chains

in Bacillus FTU subcellular vesicles Eur J Biochem 198 527-534

Kreke, B & Cypionka, H (1994) Role of sodium ions for sulfate ttansport and energy

metabohsm in Desulfovibno salexigens Arch Microbiol 161 55-61 Laubinger, W & Dimroth, P (1988) Charactenzaüon of the ATP synthase of

Propiomgenium modestum as a pnmary sodium pump Biochemistry 27 7531- 7537

Laubinger, W & Dimroth, P (1989) The sodium ion translocating ATPase of

Propiomgenium modestum pumps protons at low sodium ion concentrations Biochemistry 28 7194-7198 Laubinger, W, Deckers-Hebestreit, G, Altendorf, K & Dimroth, P (1990) A hybnd ATPase composed of Fj of Escherichia coli and F0 of Propiomgenium modestum

is a functional sodium ion pump Biochemistry 29 5458-5463

E E H K & Dimroth, Laussermair, , Schwarz, , Oesterhelt, D, Reinke, , Beyreuther,

P (1989) The sodium ion üanslocating oxaloacetate decarboxylase of Klebsiella

pneumoniae - Sequence of the integral membrane-bound subunits ß and y J Biol

Chem 264 14710-14715

Lovley, D K (1991) Dissimilatory Fe (III) and Mn (IV) reduction Microbiol Rev 55 259-287

Maloy, S R (1990) Sodium-coupled cotransport, in The Bacteria, Vol XII, pp 203- 224, Krulwich, TA (ed ) Academic Press, San Diego, New York, London

Matthies, C & Schink, B (1992a) Fermentative degradation of glutarate via decarboxylaüon by newly isolated stncüy anaerobic bactena Arch Microbiol 157 290-296

Matthies, C & Schink, B (1992b) Energy conservation in fermentative glutarate degradation by the bactenal süain WoGl 3 FEMS Microbiol Lett 100 221-226 Matthies, C & Schink, B (1993) Anaerobic degradaüon of long-chain dicarboxylic acids bymeüianogemcennchmentcultures FEMS Microbiol Lett 111 177-182 Mclnerney, M J, Bryant, M P & Pfennig, N (1979) Anaerobic bactenum that

degrades fatty acids in synüophic associaüon with methanogens Arch Microbiol 122 129-135

Miyoshi-Akiyama, T, Hayashi, M & Unemoto T (1993) Purification and properties of cytochrome fco-type ubiqumol oxidase from a manne bactenum Vibrio alginolyticus Biochem. Biophys Acta 1141 283-287 34

Mohn, W. W. & Tiedje, J. M. (1991) Evidence for chemiosmotic coupling of reductive dechlorination and ATP synthesis in Desulfomonile tiedjei. Arch. Microbiol. 157: 1-6.

Möller, D., Schauder, R., Fuchs, G. & Thauer, R. K. (1987) Acetate oxidation to COz via a citric acid cycle involving an ATP-ciüate lyase: a mechanism for the synthesis

of ATP via subsüate level phosphorylation in Desulfobacter postgatei growing on acetate and sulfate. Arch. Microbiol. 148: 202-207.

Müller, V., Blaut, M., Heise, R., Winner, C. & Gottschalk, G. (1990) Sodium

bioenergetics in methanogens and acetogens. FEMS Microbiol. Rev. 87: 373-376.

Nicholls, D. G. & Ferguson, S. J. (eds.) (1992) in Bioenergetics 2, pp. 39-63. Academic Press, San Diego, New York, London.

Padan, E. & Schuldiner, S. (1993) Na+ transport Systems in prokaryotes, in Alkali

Cation Transport Systems in Prokaryotes, pp. 3-24, Bakker, E. P. (ed.), CRC Press, Boca Raton, Ann Arbor, London, Tokyo. Page, W. J. (1991) Examination of the role of Na+ in the physiology of the Na+- dependent soil bacterium Azotobacter salinestris. J. Gen. Microbiol. 137: 2891- 2899.

Polvani, C, Sachs, G. & Biostein, R. (1989) Sodium ions as Substitutes for protons in the gastric H,K-ATPase. J. Biol. Chem 264: 17854-17859. Poolman, B. (1990) Precursor/product antiport in bacteria. Molec. Microbiol. 4: 1629- 1636.

Reidlinger, J. & Müller, V. (1993) Purification and characterization of the Na+- translocating ATP synthase of Acetobacterium woodii. Bioengineering 1/93: 19. van der Rest, M. E., Sieve, R. M., Abee, T., Schwarz, E., Oesterhelt, D. & Konings,

W. N. (1992) Nucleotide sequence and functional properties of a sodium-dependent citrate transport System from Klebsiella pneumoniae. J. Biol. Chem 267: 8971- 8976.

Ruan, Z.-S., Ananüiaram, V., Crawford, I. T., Ambudkar, S. V., Rhee, S. Y., Allison, M. J. & Maloney, P. C. (1992) Identification, purification, and reconstitution of OxlT, the Oxalate: formate antiport protein of Oxalobacter formigenes. J. Biol. Chem. 267: 10537-10543.

Samols, D., Thornton, C. G., Murtif, V. L., Kumar, G. K„ Haase, F. C. & Wood, H.

G. (1988) Evolutionary conservation among biotin enzymes. J. Biol. Chem 263: 6461-6464.

Schäfer, T., Selig, M. & Schönheit, P. (1993) Acetyl-CoA synthetase (ADP forming)

in archaea, a novel enzyme involved in acetate formation and ATP synthesis. Arch. Microbiol. 159: 72-83. 35

Schink, B. (1991) Syntrophism in procaryotes, in The Procaryotes, second edition,

vol. 1, pp. 276-299, Balows, A, Trüper, H. G., Dworkin, M., Härder, W. & Schleifer, K. H. (eds). Springer, New York, Berlin, Heidelberg.

Schink, B. & Pfennig, N. (1982) Propiomgenium modestum gen. nov. sp. nov. a new stricuy anaerobic, nonsporing bacterium growing on succinate. Arch. Microbiol. 133: 209-216.

Schmitz, R. A., Richter, M., Linder, D. & Thauer, R. K. (1992) A tungsten-containing active formylmethanofuran dehydrogenase in the thermophilic archaeon Methanobacterium wolfei. Eur. J. Biochem. 207: 559-565. Schwarz, E., Oesterhelt, D., Reinke, H., Beyreufher, K. & Dimroth, P. (1988) The

sodium ion translocating oxaloacetate decarboxylase of Klebsiella pneumoniae - Sequence of the biotin-containing a-subunit and relationship to other biotin-

containing enzymes. J. Biol. Chem 263:9640-9645.

Segel, I. H. (1975) in Enzyme Kinetics - Behavior and Analysis of Rapid Equilibrium

and Steady-State Enzyme Systems, pp. 626-634. J. Wiley & Sons, New York, London, Sydney, Toronto. Shirvan, M. H., Schuldiner, S. & Rottem, S. (1989a) Role of sodium cycle in cell volume regulation of Mycoplasma gallisepticum. J. Bacteriol. 171:4410-4416. Shirvan, M. H., Schuldiner, S. & Rottem, S. (1989b) Volume regulation in

Mycoplasma gallisepticum: evidence that Na+ is extruded via a primary Na+ pump.

J. Bacteriol. 171: 4417-4424.

Smith, D. G. E., Rüssel, W. C, Ingledew, W. J. & Thirkell, D. (1993) Hydrolysis of

urea by Ureaplasma urealyticum generates a transmembrane potential with resultant ATP synthesis. J. Bacteriol. 175: 3253-3258. Speelmans, G., Poolman, B. & Konings, W. N. (1993a) Energy transduction in the thermophilic anaerobic bacterium Clostridium fervidus is exclusively coupled to sodium ions. Proc. Natl. Acad. Sei. USA 90: 7975-7979.

Speelmans, G., Poolman, B., Abee, T. & Konings, W. N. (1993b) Amino acid

transport in Üie thermophilic anaerobe Clostridium fervidus is driven by an electrochemical sodium gradient J. Bacteriol. 175:2060-2066. Strobel, H. J. & Rüssel, J. B. (1991) Role of sodium in the growth of a ruminal selenomonad. Appl. Environ. Microbiol. 57:1663-1668. Strohhäcker, J. & Schink, B. (1991) Energetic aspects of malate and lactate fermentation by Acetobacterium malicum. FEMS Microbiol. Lett. 90: 83-88.

Stryer, L. (ed.) (1988) in Biochemistry, third edition, pp. 480-488. W. H. Freeman and Company, New York. 36

Takase, K., Yamato, I. & Kakinuma, Y. (1993) Cloning and sequencing of the genes coding for the A and B subunits of vacuolar-type Na+-ATPase from Enterococcus hirae. J. Biol. Chem 268: 11610-11616.

Thauer, R. K. (1988) Citric-acid cyle, 50 years on - Modifications and an alternative pathway in anaerobic bacteria. Eur. J. Biochem 176:497-508. Thauer, R. K. (1990) Energy metabolism of methanogenic bacteria. Biochem. Biophys. Acta 1018: 256-259.

Thauer, R. K., Jungermann, K. & Decker, K. (1977) Energy conservation in chemotrophic anaerobes. Bacf. Rev. 41:100-180. Toh, H., Kondo, H. & Tanabe, T. (1993) Molecular evolution of biotin-dependent carboxylases. Eur. J. Biochem 215:687-696. Tokuda, H. (1993) The Na+ cycle in Vibrio alginolyticus, in Alkali Cation Transport

Systems in Prokaryotes, pp. 125-138, Bakker, E. P. (ed.). CRC Press, Boca Raton, Ann Arbor, London, Tokyo.

Unemoto, T., Tokuda, H. & Hayashi, M. (1990) Primary sodium pumps and their

significance in bacterial energetics, in The Bacteria, Vol. XII, pp. 33-54, Krulwich, T. A. (ed.). Academic Press, San Diego, New York, London. Wagner, A. F. V., Frey, M., Neugebauer, F. A, Schäfer, W. & Knappe, J. (1992) The

free radical in pyruvate formate-lyase is located on glycine-734. Proc. Natl. Acad. Sei. USA 89: 996-1000.

Weiss, D. S. & Thauer, R. K. (1993) Methanogenesis and the unity of biochemistry. Cell 72: 819-822.

Wifling, K. & Dimroth, P. (1989) Isolation and characterization of oxaloacetate

decarboxylase of Salmonella typhimurium, a sodium pump. Arch. Microbiol 152: 584-588.

Woehlke, G., Wifling, K. & Dimroüi, P. (1992a) Sequence of the sodium ion pump oxaloacetate decarboxylase from Salmonella typhimurium. J. Biol. Chem. 267: 22798-22803.

Woehlke, G., Laussermair, E., Schwarz, E., Oesterhelt, D., Reinke, H., Beyreuther, K. & Dimroüi, P. (1992b) Sequence of the ß-subunit of oxaloacetate decarboxylase

from Klebsiella pneumoniae: a conection of the C-terminal part. J. Biol. Chem 267: 22804-22805. 37

Chapter II

Malonate decarboxylase of Malonomonas rubra, a novel type of biotin-containing acetyl enzyme

Eur.J.Biochem. 207, 117-123 (1992)

Hubert HILBI1, Irmtraut DEHNING2, Bernhard SCHINK2 and Peter DIMROTH1

1 Mikrobiologisches Institut, Eidgenössische Technische Hochschule, ETH-Zentrum,

Zürich, Switzerland

2 Fakultät für Biologie der Universität Konstanz, Federal Republic of Germany

Correspondence to P. Dimroth, Mikrobiologisches Institut, Eidgenössische

Technische Hochschule, ETH-Zentrum, CH-8092 Zürich, Switzerland

Enzymes. Acetyl-CoA carboxylase (EC 6.4.1.2); adenosinetriphosphatase

(EC 3.6.1.3); citfamalate lyase (EC 4.1.3.22); citrate lyase (EC 4.1.3.6); deoxyribonuclease I (EC 3.1.21.1); glutaconyl-CoA decarboxylase (EC 4.1.1.70); malonate decarboxylase (EC 4.1.1.-); malonyl-CoA decarboxylase (EC 4.1.1.9); methylmalonyl-CoA decarboxylase (EC 4.1.1.41); oxaloacetate decarboxylase

(EC 4.1.1.3). 38

H. 1. Summary

Cell suspensions or crude extracts of Malonomonas rubra grown anaerobically on malonate catalyze the decarboxylation of this Substrate at a rate of 1.7 - 2.5 ujnol • min-1 • mg protein-1 which is consistent with the malonate degradation rate during growth. After fractionation of the cell extract by ultracentrifugation, neither Üie soluble nor the particulate fraction alone catalyzed the decarboxylation of malonate, but on recombination of the two fractions 87% of the activity of the unfractionated extract was restored. The decarboxylation pathway did not involve Üie intermediate formation of malonyl-CoA, but decarboxylation proceeded direcüy with free malonate. The catalytic activity of the enzyme was completely abolished on incubation with hydroxylamine or NaSCN. Approximately 50-65% of Üie original decarboxylase activity was restored by incubation of the extract with ATP in Üie presence of acetate, and the extent of reactivation increased after incubation with dithioerythritol. Reactivation of the enzyme was also obtained by chemical acetylation with acetic anhydride. These results indicate modification of the decarboxylase by deacetylation leading to inactivation and by acetylation of Üie inactivated enzyme specimens leading to reactivation. It is suggested that the catalytic mechanism involves exchange of the enzyme-bound acetyl residues by malonyl residues and subsequent decarboxylation releasing CO2 and regenerating the acetyl- enzyme. The decarboxylase was inhibited by avidin but not by an avidin-biotin complex indicating that biotin is involved in catalysis. A Single biotin-containing

120-kDa Polypeptide was present in Üie exüact and is a likely component of malonate decarboxylase.

n. 2. Introduction

In recent years, several anaerobic bacteria have been shown to conserve the free energy of decarboxylation reactions (1,2). Membrane-bound Na+-tanslocating decarboxylases have been found for oxaloacetate in Klebsiella pneumoniae (3,4) and 39

Salmonella typhimurium (5), for methylmalonyl-CoA in Veillonella alcalescens (6)

and Propiomgenium modestum (7) and for glutaconyl-CoA in Acidaminococcus fermentans and other glutamate fermenting bacteria (8). Upon decarboxylation of

these subsüates, an elecüochemical gradient of Na+ is established over Üie

cytoplasmic membrane that drives various endergonic reactions, e.g. sohlte transport

(9), ATP synthesis (10), or NADH synthesis by reversed elecüon transfer (11). All

Na+-üanslocating decarboxylases are oligomeric enzymes with a similar subunit

composition and all contain a prostheüc group, biotin (1,2). The reaction mechanism

is also similar and involves carboxylation of enzyme-bound biotin by carboxyl

transfer from Üie subsüate and subsequent decarboxylation of carboxybiotin which is

coupled to Na+ translocation (1,2).

The free energy of the decarboxylation of a carboxylic acid allows growth of

P. modestum on succinate (7,12), of Oxalobacter formigenes on Oxalate (13) and of

Malonomonas rubra on malonate (14). P. modestum converts Üie decarboxylation

energy by virtue of methylmalonyl-CoA decarboxylase into ApNa+, which serves a

unique Na+-üanslocating FiF0 ATPase as Üie driving force for ATP synthesis (7,10).

An entirely different mode for conserving decarboxylation energy applies for

O. formigenes (13). Here, a ApH+ is generated by an oxalate/formate antiporter which

receives its driving force from the low Oxalate concentration and high formate

concentration in the cytoplasm as a consequence of an efficient and irreversible

decarboxylation of the Substrate inside the bacteria.

To make the decarboxylation of these subsüates chemically feasible, succinate is

converted to methylmalonyl-CoA (7) and Oxalate is converted to oxalyl-CoA (13). By

analogy, one might expect that in M. rubra malonate is converted to malonyl-CoA

which is then decarboxylated to acetyl-CoA. Energy conservation in this organism

could be analogous to that of P. modestum involving a membrane-bound

decarboxylase or could apply to Üie O. formigenes type with an energy-generating

antiporter. As an approach to elucidating the energy-conservation mechanism in 40

M. rubra, we have determined Üie type of enzyme(s) catalyzing Üie decarboxylation of malonate.

n. 3. Materials and Methods

Materials

DNase I was from Boehringer (Mannheim, FRG), avidin-peroxidase-labeled,

3,3'-dimefhoxybenzidine and diisopropyl fluorophosphate were from Sigma (Buchs,

Switzerland); prestained SDS/PAGE Standards (low ränge) were from Bio-Rad. All other Chemicals, as well as avidin, were from Fluka (Buchs, Switzerland).

Growth ofthe organism

M. rubra was grown anaerobically at 30°C with malonate as sole carbon and energy source essentially as described in (14). The growth medium contained 40 mM malonate, 50 mM NaHC03, 340 mM NaCl, 5 mM NH4C1, 1.15 mM KH2P04,

15 mM MgCl2,1 mM CaCl2,6.7 mM KCl, 1 mM cysteine and 0.5 ml/1 each of filter- sterilized seven-vitamin Solution (15), trace dement Solution SL 10 (16) and selenite/tungstate Solution (17). Stock Solutions of neuüalized malonate, bicarbonate and cysteine were autoclaved separately and the pH was adjusted to pH 7.1 - 7.3. The bacteria were transferred from a 50-ml stock culture (transferred into fresh medium once a month) into 0.5 1 medium, and after 2 days this culture was used to inoculate

12 1 medium in botües sealed with gassing-tube-equipped rubber Stoppers. After approximately 40 h, the culture was supplemented with another 50 mM malonate and allowed to grow for another 32 h. Cells were harvested by conünuous centrifugation at 19000 x g (Contifuge 17 RS, rotor 8575; Hereaus, Zürich). The yield was about 1 g wet packed cells/1 medium. As the cells lysed during washing, even in the presence of up to 500 mM NaCl, they were frozen in liquid niüogen without previous washing.

The malonate decarboxylase was stable on storage under these conditions. Cultures were checked for purity with the agar shake culture method (18). The appearence of 41

Single colonies was biconvex with a brownish-red color developing grey fuzzy edges with age.

Preparation ofcell extract and subcellular fractions

5 g frozen wet cells were suspended in 25 ml buffer A (50 mM potassium

Phosphate, pH 7.5, 100 mM NaCl, 5 mM MgCy, containing 5 mM dithioerythritol,

0.2 mM diisopropyl fluorophosphate and 1 mg DNase I.

The following treatments were performed at 4°C. Cells were disrupted by two passages through a French press at 55 MPa. Whole cells and large debris were removed by centrifugation for 20 min at 15000 x g. The supernatant, referred to as extiact, could be stored in liquid niüogen without loss of malonate decarboxylase activity. Fractionation of üie extiact by ultracentrifugation was carried out with a

Beckman L8-70 ulüacentrifuge at 200'000 x g for 30 min (rotor 70.1 Ti). The resulting membrane pellet was resuspended either in buffer A or, as control, in the cytoplasmic fraction using a motor-driven Teflon plunger (type RW 20 DZM, Janke

& Kunkel).

Determination ofmalonate decarboxylase activity

Malonate decarboxylase activity was determined by measuring CO2 formation from malonate. The reactions were carried out in 13-ml serum botües sealed with rubber Stoppers at 30°C. Decarboxylaüon was initiated by adding 3 u.1 1 M sodium malonate, pH 7.5, to 0.1 ml extract (1.3 mg protein) in buffer A. In some experiments, Üie extract containing endogenous acetate from bacterial metabolism, was incubated for 2 min with 5 u.1 100 mM ATP to reacetylate deacetylated enzyme specimens and thus restore catalytic activity. If not indicated otherwise, Üie reaction was terminated after 30 s by adding 10 ui 2 M HCl with a syringe. The amount of

CO2 formed and released into the gas phase was determined by injecting 0.3 ml gas phase onto a Poropak-N column maintained at 140°C. The amount of CO2 analysed was not dependent on Üie incubation time after acidification. The signal was recorded 42

with a Hewlett Packard 5890 series II gas Chromatograph equipped with a thermal- conductance detector. Peak areas were automatically calculated and converted into an amount of CO2 by comparison with Üie peak areas of CO2 Standards. The stock mixture of CO2 Standard was a 1:200 dilution of CO2 in N2. In our measuring ränge of 10-120 nmol CO2, üie areas corresponded linearly to the amount of CO2 analysed.

Determination ofacetate concentration

Acetate concentration was determined by GC using a Perkin Eimer 8700 gas

Chromatograph equipped with a flame-ionization detector. Samples were acidified with 10% H3PO4 and centrifuged for 5 min at 15 000 x g. 1 u.1 supernatant was injected and separated on a Chromosorb-WAW column (10% SP 1200, 1% H3P04;

120°C). Standards were treated similarly.

SDS/PAGE and Western blotting

The exüact was separated by SDS/PAGE using a Midget apparatus

(Pharmacia/LKB). The sample gel was prepared with 4% arylamide and 0.125% bisacrylamide whereas the separating gel was prepared with 10% acrylamide, 0.3% bisacrylamide and either 6 M urea or 13% glycerol according to (19). The

Polypeptide bands were either visualized by staining with Coomassie blue (20) or the gel was blotted onto a niüocellulose membrane as described (21). After washing the blot twice for 5 min with buffer B (10 mM Tris/Cl, pH 7.5 and 150 mM NaCl), it was blocked 1 h with 2% bovine serum albumin, following incubation with approximately

0.2 mg of an avidin-peroxidase conjugate in 20 ml buffer B for another hour.

Subsequenüy, the blot was washed three times for 10 min with buffer B, and biotin- containing Polypeptides were visualized by incubation with 20 ml buffer B, containing 5 mg 3,3'-dimethoxybenzidine (solubilized in 1 ml methanol) and 20 ul

35% H202. 43

Protein determination

Protein was determined according to Bradford (22), using the Bio-Rad protein- assay reagent mixtum. Bovine serum albumin served as a Standard.

n. 4. Results

Demonstration ofa malonate decarboxylase in cell-free extracts ofM. rubra

Exponentially growing cells of M. rubra consume malonate at a considerable rate. The malonate-decarboxylating activity of these cells was calculated from the growth parameters Ys = 1.9 g/mol and u = 0.154 h_1 (14) to be 2.5 umol • mür1 • mg protein-1. A Suspension of cells harvested in the stationary growth phase catalyzed malonate decarboxylation at a rate of 1.9 umol • min-1 • mg protein-1.

Cell-free exttacts prepared with a French pressure cell retained the malonate decarboxylase activity. The kinetics of CO2 formation from malonate under optimized conditions are shown in Fig. 1. Within Üie first 30 s, Üie amount of CO2 formed increased linearly with time yielding a specific activity of Üie enzyme of

2.7 U/mg protein. This activity is Üierefore commensurate with the malonate- decarboxylation rate of whole cells. After 30 s the decarboxylation of malonate slowed down and after 120 s the Substrate was completely decarboxylated.

These results indicate that M. rubra contains an enzyme System that catalyzes Üie cleavage of Üie unmodified Substrate according to Eqn 1:

malonate + H+ -»acetate + CO2 (1)

Acetyl-CoA was without effect on malonate decarboxylation, and ATP did not stimulate Üie activity of a freshly prepared cell exttact. A complex enzyme System catalyzing malonate decarboxylation via malonyl-CoA to acetyl-CoA is thus not indicated by our experiments. 44

The dependence of malonate decarboxylase activity on protein concentration was

not linear but showed an increase of specific activity with increasing protein

concenüation (Fig. 2). The concenüation-dependent increase of decarboxylase

activity could indicate (he presence of an endogenous cofactor or a multicomponent

enzyme System being subject to a concenüation-dependent dissociation/association

Fig. 1. Kinetics of CO2 formation from malonate, catalyzed by malonate decarboxylase. The decarboxylation of malonate by an exüact of M. rubra was determined from the amount of CO? liberated as described in Materials and Methods. Assays were performed with 30 mM malonate (•) or wiüiout subsüate (). 45

equilibrium (23). The requirement for an exchangeable low-molecular-mass cofactor was excluded by demonsüating malonate decarboxylase activity in dialyzed extracts

(see below).

6 8 10

Protein (mg/ml)

Fig. 2. Dependence of malonate decarboxylase activity on protein concentration. The extract of M. rubra with an initial protein concenüation of 13 mg/ml was diluted with buffer A to Üie concenüations indicated. Malonate decarboxylation was determined with 0.1 ml of each diluted extiact as described under Materials and Methods. The reactions were terminated after 30 s, in the linear part of the kinetics of CO2 formation (Fig. 1). The activities are Üierefore based on initial rates. Values are means of duplicate assays. 46

77j« malonate decarboxylase activity depends on the soluble and particulate fractions ofthe cell extract

Evidence for a multicomponent enzyme System for the decarboxylation of malonate was obtained after separating soluble and particulate fractions of the extract by ulüacentrifugation. The results in Table 1 indicate that neither of these separated fractions retained any significant amount of malonate decarboxylase activity. On recombination of both fractions, however, 87% of the activity of Üie non-fractionated exttact was recovered. Thus, Üie enzyme, after disruption of Üie cells, consists of a soluble and a membrane-bound component. While it is unclear whether Üie soluble part of the enzyme has been sheared off during cell rupture, the results clearly indicate that a membrane-bound component is involved. Location of Üie enzyme in

Üie membrane is expected if energy conservation proceeds by a direct malonate decarboxylase-dependent ion-pumping mechanism.

Table 1. Disappearance of malonate decarboxylase activity after Separation of the extract into cytoplasmic and membrane fractions by ultracentrifugation and reconstitution of enzyme activity by combining these fractions. Centtifugation of the crude extract was performed at 200'000 x g for 30 min, as described in Material and Methods.

Fraction Malonate decarboxylase activity

U/mg (%)

Extract 2.83 (100)

Membrane 0.09 (3)

Cytoplasm 0.02 (1)

Membrane/cytoplasm 2.45 (87) 47

Evidence for an acetyl enzyme

The above results, indicating decarboxylation of free malonate were surprizing

from a chemical point of view, and a malonate decarboxylase enzyme has never been

found before. Malonyl-CoA, however, is decarboxylated by a variety of enzymes (24)

including the Na+-ttanslocating methylmalonyl-CoA decarboxylase of V. alcalescens

(6). The chemical problem of decarboxylating free malonate might be overcome by

forming the malonyl thioester group transienüy on Üie enzyme by an exchange of

malonate for an enzyme-bound acetyl thioester residue. Decarboxylation of the

malonyl thioester on the enzyme would regenerate the acetyl enzyme. This

mechanism would imply that malonate decarboxylase is active only in an acetylated

form.

To test this hypothesis üie enzyme was incubated with hydroxylamine to remove

putative acetyl thioester residues by formation of acetyl hydroxamates. The results in

Fig. 3 show that Üie decarboxylase was inactivated on incubation with

hydroxylamine. The activity decreased with increasing hydroxylamine

concenüations, complete inactivation being observed after 5 min incubation with

500 mM hydroxylamine. Inactivation of the decarboxylase was also observed on

incubation with thiocyanate at approximately three times lower concentations than

with hydroxylamine.

Enzyme specimens thus inactivated were dialyzed for 4 h to remove Üie inhibitor,

and malonate decarboxylase activity was subsequenüy determined in Üie presence of

Üie Compounds listed in Table 2. Enzyme completely inactivated by hydroxylamine

regained 45% of the original activity after 2 min incubation with 5 mM ATP. The

activity increased further to 65% if the exttact was incubated with 20 mM

dithioerythritol, indicating Üie importance of thiol groups for this reactivation. With extracts containing 0.18 mM endogenous acetate, the additional presence of 1-5 mM

sodium acetate was without effect on Üie reactivation of decarboxylase activity. With extensively dialyzed exttact, however, the activation with acetate plus ATP was 2.2

times higher than with ATP alone. After these treatments, 11% of the original enzyme 4g

activity was recovered. A conttol exttact, that was not treated with hydroxylamine but kept overnight at 4°C, lost 80% of its activity which could not be restored to any significant extent by incubation with ATP. The enzyme thus lost activity on storage which was not due to a deacetylation event. If this irreversible loss of activity on

100 200 300 400 500

Concenüation of Inhibitor (mM)

Fig. 3. Inhibition of malonate decarboxylase by hydroxylamine and thiocyanate. The bacterial exttact (75 u.1, 1 mg protein) was mixed with 25 u.1 hydroxylamine in buffer A to yield the inhibitor concentration indicated. After 5 min at 25°C the decarboxylase activity was determined (•). Altematively, 90 |xl bacterial exttact was mixed with 10 ul NaSCN in buffer A to yield the final inhibitor concentration indicated. The malonate decarboxylase activity was subsequenüy determined as described above (). Values represent means of duplicate assays. 49

Table 2. Reactivation of inactivated malonate decarboxylase by acetylation. The freshly prepared enzyme of specific activity 2.67 U/mg protein (not increased by incubation with ATP) was completely inactivated by incubation with 100 mM hydroxylamine for 45 min followed by dialysis for 4 h (A), by incubation with 100 mM NaSCN for 15 min followed by dialysis for 3 h (B), or by dialysis for 18 h (C). The dialysis buffer was buffer A in all cases. As a conttol for the last experiment (C), the exttact was kept without dialysis for 18 h at 4°C. For the chemical acetylation of Üie enzyme with acetic anhydride, 1 uJ 100 mM acetic anhydride in H2O (düuted immediately before) was added to 100 u.1 inactivated enzyme and the mixture was incubated for 1 min at 30°C. The residual malonate decarboxylase activity was 0.5 U/mg protein and could not be increased by incubation with ATP. n.d., not determined. For (A) and (B), activities are given as percentage initial activity, while for (C) activity is given as percentage conttol.

Additions Malonate decarboxylase activities

U/mg (%) U/mg (%) U/mg (%)

Inactivated enzyme 0 (0) 0 (0) 0 (0)

5mMATP/5mM 1.20 (45) 1.32 (49) 0.30 (60)

sodium acetate

5 mM ATP/20 mM 1.73 (65) n.d. (n.d.) n.d. (n.d.) dithioerythritol

1 mM acetic anhydride 0.44 (16) 1.04 (39) 0.23 (46)

storage is taken into account, Üie proportion of decarboxylase that was reactivated by

ATP plus acetate was 60%.

These results suggested that an enzyme-catalyzed acetylation of specific SH groups of the decarboxylase with ATP and acetate is responsible for the observed reactivation. Convincing evidence for catalytically competent acetyl residues bound to Üie decarboxylase was obtained by reactivation of the inactivated enzyme with acetic anhydride. This chemical acetylation restored 36-77% of the activity that could 50

be obtained by enzymic acetylation with ATP and acetate. Similar results on reactivation by acetylation were obtained with enzyme samples that were inactivated with thiocyanate, indicating that this treatment likewise leads to a loss of catalytically important acetyl residues.

Results on Substrate specificity of the acetylating enzyme (Table 3) indicate that

Üie decarboxylase is reactivated with ATP, ADP, or GTP with decreasing efficiency in that order, but not with AMP or with acetyl-CoA. As ATP may be formed from

ADP by adenylate Kinase but not from AMP, the results suggest that ATP is the physiological Substrate of the enzyme acetylating deacetylmalonate decarboxylase.

Table 3. Specificity of the malonate decarboxylase-acetylating enzyme for ATP. The enzyme of specific activity 2.67 U/mg protein was completely inactivated by incubation with 100 mM hydroxylamine (45 min) followed by dialysis (4 h). Activities are given as U/mg and percentage of initial activity.

Additions Malonate decarboxylase activity

U/mg (%)

Inactivated enzyme 0 (0)

5mMATP 1.20 (45)

5mMADP 0.84 (31)

5 mM AMP 0 (0)

5 mM GTP 0.72 (27)

5 mM acetyl-CoA 0 (0) 51

Evidence for a biotin enzyme

A possible involvement of biotin in Üie reaction catalyzed by malonate decarboxylase was determined by measuring the effect of avidin on catalytic activity.

The results in Fig. 4 show decreasing malonate decarboxylase activities after

~\ 1 1 1 1 1 r

20 40 60 80 100 120 140 160

Avidin (ug/mg extract protein)

Fig. 4. Inhibition of malonate decarboxylase by avidin. The bacterial extract (100 u.1, 1.3 mg protein) was mixed with 20 ul avidin of the appropriate concentration in buffer A to yield Üie amount shown. After 15 min at 25°C, residual biotin-binding sites of the avidin were blocked by adding 10 u.1 biotin (7.5 mg/ml, final concentration 2.4 mM). After another 15 min the malonate decarboxylase activity was determined (•). In üie conttol () avidin was incubated with biotin for 30 min prior to the addition to Üie cell extract. Values are means of duplicate assays. 52

incubation of Üie cell exttact with increasing amounts of avidin. With 45 u.g avidin/mg exttact protein, the inhibition was 50%, and 150 ng avidin/mg exttact protein destroyed Üie decarboxylase activity almost completely. Incubation wiüi an avidin-biotin complex performed as a conttol, however, was without effect on the enzyme activity. These results sttongly indicate a biotin prosthetic group on malonate decarboxylase. Accordingly, we found a Single biotin-containing protein band after

SDS/PAGE. After blotting the separated Polypeptides onto niüocellulose membranes, labelling with an avidin-peroxidase conjugate and locating the peroxidase with

H202/3,3'-dimethoxybenzidine, only a Single stained band became visible.

Comparison of the mobility of this band with those of marker proteins indicated a molecular mass for the biotin-containing Polypeptide of approximately 120 kDa. The realtive mobility of this band was not affected if the SDS gel contained additions of

6M urea or 13% glycerol.

II. 5. Discussion

On decarboxylation of a carboxylic acid, Üie carboxyl group is replaced by a

proton. An electron-withdrawing subsütuent on Üie carbon at which this replacement

occurs therefore greatiy facilitates these decarboxylation reactions. For catalysis of

decarboxylation under physiological conditions, the electton-atttacting subsütuent is

provided, e.g. by a carbonyl group in 3-oxoacids, a thioester residue, the pyridinium

cation of pyridoxal phosphate, or the thiazolium cation of thiamine pyrophosphate.

M. rubra grows using malonate as the sole carbon and energy source and thereby

decarboxylates the dicarboxylic acid to acetate and CO2 (14). A chemically feasible

mechanism for malonate decarboxylation in these cells (Eqn 1) would be üie

conversion of malonate to malonyl-CoA by a CoA transferase with acetyl-CoA as üie

second Substrate (Eqn 2), and subsequent decarboxylation of malonyl-CoA to acetyl-

CoA (Eqn 3). Malonyl-CoA could alternately be formed under ATP consumption by

malonyl-CoA synthetase. 53

malonate + acetyl-CoA *— malonyl-CoA + acetate (2)

malonyl-CoA + H+-*acetyl-CoA + C02 (3)

malonate + H+-* acetate + C02 (1)

Malonyl-CoA decarboxylase is present in many organisms including plants, animals and bacteria (24). The enzymes isolated from the uropygial gland of geese or from Mycobacterium tuberculosis are not biotin dependent, are located in Üie cytoplasm and do not participate in energy conservation (24). Therefore, the decarboxylase of M. rubra could more likely be related to methylmalonyl-CoA decarboxylase of V. alcalescens which accepts malonyl-CoA as an alternate Substrate

(6). This enzyme is firmly bound to the membrane, contains biotin as prostheüc group and conserves energy by generating an electrochemical gradient of Na+ across Üie membrane (6).

An investigation of Üie enzyme(s) involved in malonate decarboxylation by

M. rubra, however, led to üie discovery that Üie decarboxylase reacts with free malonate and that malonyl-CoA is not involved in this catalysis. Based on diese results, an alternative mechanism was considered in which malonate was activated by its covalent attachment to the enzyme instead of its linkage to CoA. A catalytic sequence analogous to that shown in Eqn (2) and Eqn (3) could result, with acetyl- enzyme and malonyl-enzyme derivatives as shown in Fig. 5. This mechanism consists of an exchange of malonate for enzyme-bound acetyl-residues and the subsequent decarboxylation of üie malonyl-enzyme intermediate to yield CO2, thereby regenerating Üie acetyl-enzyme. Evidence for this type of catalysis is presented in this paper by the demonsttation of an acetyl-enzyme as the active catalyst. The decarboxylase was completely inactivated by hydroxylamine and no activity was restored by subsequent removal of Üie inhibitor. Under Üie same conditions, an acetyl thioester would be converted into acetyl hydroxamate and the free mercaptane (25).

The enzyme was also inactivated by a variety of other conditions, e.g. incubation with 54

NaSCN, prolonged dialysis, or a high pressure during rupture of the cells. In all these cases, the inactivation was mainly due to deacetylation, as these inactive enzyme specimens recovered activity upon chemical or enzymic acetylation with acetic anhydride or ATP plus acetate, respectively. Acetylation of a mercaptane residue of

Üie decarboxylase is indicated by Üie observed increase of restored catalytic activity if the enzyme was incubated with dithioerythritol. As the crude exttact dialyzed for 3-

4 h contained about 0.2 mM acetate and as the enzyme which acetylates the inactive decarboxylase apparenüy has a low Km for acetate, the exttact had to be extensively dialyzed to demonsttate the acetate requirement of this enzymic reactivation.

Malonate s. • Enzyme-S-acetyl ^^. C02

*~ ^ Acetate Enzyme-S-malonyl / \ H+

Fig. 5. Hypothetical reaction mechanism ofmalonate decarboxylase.

These findings on activation and inactivation of malonate decarboxylase by acetylation or deacetylation are reminiscent of similar observations with citrate lyase

(25-27) and citramalate lyase (28). These lyases are confronted with an analogous chemical problem to that of malonate decarboxylase: a C-C bond next to Üie methylene group of a non-activated acetyl-residue must be cleaved through replacement by a proton. This heterologous C-C bond cleavage creates an electron gap at Üie replaced carbon atom which is filled by the conversion of an hydroxyl into a carbonyl bond. The problem of low chemical reactivity for these lyase reactions is solved in all three cases by the formation of thioester bonds between the acetate- yielding carboxyl group of the Substrate and a thiol group of the enzyme. A 55

comparison of these reactions is given (Fig. 6) for the enzyme-bound thioester derivatives.

A 0 = C-CH2-C-SE > C02 + CH3-C-SE M<

S\JOH O 0 0 (v ' ll II II B,C R^C-CHj-C-SE > R2-C +CH3-C-SE

Fig. 6. A comparison of the reactions of malonate decarboxylase, citrate lyase and citramalate lyase involving enzyme-bound thioester derivatives. (A) Malonate decarboxylase. (B) Cittate lyase: Ri, COOH; R2, CH2-COOH. (C) Citramalate lyase: R,,COOH;R2,CH3.

The thioester linkage creates an electron sink at the adjacent methylene carbon

and thereby makes the replacement of its carbon subsütuent by a proton chemically

feasible. Very similar conditions to those reported here for malonate decarboxylase

apply for üie deacetylation associated with inactivation of cittate lyase and

citramalate lyase with hydroxylamine, and for the acetylation associated with

reactivation of these enzymes with acetic anhydride or with acetate and ATP in üie

presence of a specific ligase (25-29). It will be interesting to determine whether this

analogy extends to the structure of the catalytically competent thiol-carrying residue

which in the lyases is a phosphoribosyl-dephospho CoA prostheüc group bound as a

phosphodiester to a specific serine residue of the proteins (30-32).

M. rubra grows by decarboxylation of malonate to acetate and CO2 and must

therefore conserve energy from this reaction (14). Inhibition of the decarboxylase by

avidin and demonstration of a Single biotin-containing 120-kDa Polypeptide by 56

SDS/PAGE of the crude extract indicate that this Polypeptide belongs to Üie

decarboxylase. From these results, one may speculate that malonate decarboxylase

performs Üie same type of energy conservation as Üie biotin-containing Na+-ttansport

decarboxylases, oxaloacetate decarboxylase, methylmalonyl-CoA decaraboxylase and

glutaconyl-CoA decarboxylase (1,2), although with Üie additional requirement for

substtate activation on Üie enzyme (see above). An energy conservation through a

malonate decarboxylase-catalyzed ion translocation is also indicated by the location

of a component of this enzyme within Üie membrane.

The presence of only a Single biotin-containing Polypeptide in M. rubra growing

on malonate is notable. Most bacteria require Üie biotin enzyme acetyl-CoA

carboxylase to synthesize malonyl-CoA for fatty-acid biosynthesis. In M. rubra,

malonyl-CoA could be readily provided by forming üie thioester from malonate, and acetyl-CoA carboxylase may therefore be repressed by malonate. The enzyme should be present, however, if the bacteria are grown on a different substtate, e.g. fumarate

(14). It appears plausible that Üie malonate decarboxylase mechanism with enzyme- bound thioesters is advantageous over Üie alternate route via malonyl-CoA and acetyl-CoA because competition with fatty-acid synthetase for üie substtate malonyl-

CoA is avoided.

I.D. and B.S. are indebted to W. Buckel for fruitful dicussion. This study was in part supported by the Deutsche Forschungsgemeinschaft. 57

n. 6. References

1. Dimroth, P. (1987) Microbiol Rev. 51,320-340.

2. Dimroth, P. (1990) Phil. Trans. R. Soc. Land. B Biol. Sei. 326,465-477.

3. Dimroth, P. Eur. J. Biochem 121,435-441. , (1982)

4. Dimroth, P. (1982) Eur. J. Biochem 121,443-449.

5. Wifling, K. & Dimroth, P. (1989) Arch. Microbiol. 152, 584-588.

6. Hüpert, W. & Dimroth, P. (1983) Eur. J. Biochem 132,579-587.

7. Hilpert, W., Schink, B. & Dimroth, P. (1984) EMBO J. 3,1665-1670.

8. Buckel, W. & Semmler, R. (1983) Eur. J. Biochem 136, A11-A,U.

9. Dimroth, P. & Thomer, A. (1990) J. Biol. Chem 265,7221-7224.

10. Laubinger, W. & Dimroth, P. (1988) Biochemistry 27,7531-7537.

11. Dimroth, P. & Thomer, A. (1989) Arch. Microbiol. 151,439-444.

12. Schink, B. & Pfennig, N. (1982) Arch. Microbiol. 133,209-216.

13. Ananüiaram, V., Allison, M.J. & Maloney, P.C. (1989) J. Biol. Chem 264,7244-

7250.

14. Dehning, I. & Schink, B. (1989) Arch. Microbiol. 151,427-433.

15. Pfennig, N. (1978) Int. J. Syst. Bacteriol. 28,283-288.

16. Widdel, F., Kohring, G.W. & Mayer, F. (1983) Arch. Microbiol. 134,286-294.

17. Tschech, A. & Pfennig, N. (1984) Arch. Microbiol. 137,163-167.

18. Widdel, F. & Pfennig, N. (1984) in Bergey's manual ofsystematic bacteriology

(Krieg, N.R. & Holt, J.G. eds) vol. 1, pp 663-679, Williams & Wilkins,

Baltimore, London.

19. Schaegger, H. & Von Jagow, G. (1987) Anal. Biochem 166,368-379.

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

21. Towbin, H., Staelin, T. & Gordon, J. (1979) Proc. Natl. Acad. Sei. USA 76,4350-

4356.

22. Bradford, M.M. (1976) Anal Biochem 72,248-251.

23. Segel, LH. (1975) Enzyme kinetics, J. Wiley & Sons, New York. 58

24. Kolattokudy, P.E., Poulose, A.J. & Kim, Y.S. (1982) Methods EnzymoL 71,150-

163.

25. Buckel, W., Buschmeier, V. & Eggerer, H. (1971) Hoppe-Seyler's Z. Physiol.

Chem 352, 1195-1205.

26. Dimroth, P. & Eggerer, H. (1975) Eur. J. Biochem 53,227-235.

27. Dimroüi, P. & Eggerer, H. (1975) Proc. Natl. Acad. Sei. USA 72, 3458-3462.

28. Buckel, W. & Bobi, A. (1976) Eur. J. Biochem 64,255-262.

29. Schmellenkamp, H. & Eggerer, H. (1974) Proc. Natl. Acad. Sei. USA 71,1987-

1991.

30. Dimroth, P. (1976) Eur. J. Biochem 64, 269-281.

31. Dimroth, P. & Loyal, R. (1977) FEBS Lett. 76, 280-283.

32. Robinson, J.B., Singh, M. & Srere, P.A. (1976) Proc. Natl. Acad. Sei. USA 73,

1872-1876. 59

Chapter III

The malonate decarboxylase enzyme System of Malonomonas rubra: evidence for the cytoplasmic location of the biotin-containing component

Arch. Microbiol. 160: 126-131 (1993)

Hubert Hilbi, Ren6 Hermann and Peter Dimroth

Mikrobiologisches Institut, Eidgenössische Technische Hochschule, ETH-Zentrum,

Zürich, Switzerland

Proofs to: P. Dimroth, Mikrobiologisches Institut, Eidgenössische Technische

Hochschule, ETH-Zentrum, CH-8092 Zürich, Switzerland

Key words: Malonomonas rubra - Propiomgenium modestum - malonate

- decarboxylase methylmalonyl-CoA decarboxylase - bioün - avidin - electton microscopy - high pressure freezing - immunolabeling 60

m. 1. Summary

Malonate decarboxylase of Malonomonas rubra is a complex enzyme System involving cytoplasmic and membrane-bound components. One of these is a biotin- containing protein of Mr 120'000, the location of which in Üie cytoplasm was deduced from Üie following criteria:

(i) If Üie cytoplasm was incubated with avidin and Üie malonate decarboxylase subsequenüy completed with Üie membrane fraction the decarboxylase activity was abolished. The corresponding incubation of the membrane with avidin, however, was without effect.

(ii) Western blot analysis identified Üie Single biotin-containing Polypeptide of

Mr 120'000 within Üie cytoplasm.

(iii) Transmission electton micrographs of immuno-gold labeled M. rubra cells clearly showed the location of the bioünyl protein wimin the cytoplasm, whereas Üie same procedure with Propiomgenium modestum cells indicated Üie location of the biotin enzyme methylmalonyl-CoA decarboxylase in the cell membrane.

The biotin-containing protein of the M. rubra malonate decarboxylase enzyme

System was not retained by monomeric avidin-Sepharose columns but could be isolated with üiis column in a catalytically inactive form in the presence of detörgents. If Üie high binding affinity of tettameric avidin towards bioün was reduced by desttucting part of Üie tryptophan residues by irradiaüon or oxidation with periodate, Üie inhibition of malonate decarboxylase by the modified avidin was partially reversed with an excess of biotin. Attempts to purify Üie biotin protein in its catalytically active State using modified avidin columns were without success.

III. 2. Introduction

Malonomonas rubra is an anaerobic, Gram-negative bacterium, isolated recently from anoxic marine Sediments, that grows from decarboxylation of malonate to acetate and CO2 (Dehning and Schink, 1989). The small free energy change of this reaction 61

(AG°' = -17.4 kJ/ mol) does not allow ATP synthesis by substtate level phosphorylation and no redox reactions are involved in üie catabolism of malonate

which could drive electron transport phosphorylation. With respect to energy conservation, M. rubra belongs to a group of anaerobic bacteria that grow from the

decarboxylaüon of dicarboxylic acids (for a review see Dimroth, 1987). Examples are

the decarboxylation of Oxalate to formate by Oxalobacterformigenes (Ananüiaram et

al., 1989) and of succinate to Propionate by Propiomgenium modestum (Hilpert et al.,

1984). Two different mechanisms for the conservation of decarboxylation energy have

been recognized: 0. formigenes generates ApH+ by the obligatory 1:1 exchange of

Oxalate2- for formate1" (Ananüiaram et al., 1989; Baetz and Allison, 1989). A

representative of the second group is P. modestum. In this bacterium energy is

conserved in the form of AflNa+ by a membrane-bound methylmalonyl-CoA

decarboxylase (Hilpert et al., 1984). The mechanism of energy conservation in

M. rubra has not yet been elucidated. The involvement of biotin and a membrane

component for the decarboxylation of malonate may be indicative, however, for

membrane energization (formation of AfXNa+ or AflH+) during the decarboxylation

event In a recent publication we reported the properties of Üie malonate

decarboxylating activity in crude extracts (HUbi et al., 1992). Most interestingly, the

enzyme acts on free malonate and not on malonyl-CoA and carries an essential acetyl

residue at its active site. Our results suggested exchange of these acetyl residues by

malonate and the subsequent decarboxylation of the malonyl-enzyme, thereby

regenerating Üie acetyl-enzyme. Of special interest is Üie involvement of a biotin

protein in Üie overall catalysis of malonate decarboxylaüon and Üie participation of

soluble and membrane-bound components. To further unravel the mechanism of

malonate decarboxylation and its coupling to energy conservation it is necessary to

unequivocally determine the component enzymes (proteins) that participate in Üie

decarboxylaüon event and their subcellular localization. Here we report üie

involvement of a cytoplasmically located biotin-containing protein of Mr 120'000 in

Ulis catalysis. 62

m. 3. Experimental procedures

Cultivation ofbacteria

Malonomonas rubra was grown anaerobically with malonate as sole source of

carbon and energy in a 3001 fermenter (Bioengineering; Wald, Switzerland) containing

270 1 medium. Media and conditions were virtually the same as described by Dehning

and Schink (1989). 40 mM malonic acid (neutralized with NaOH) was added to Üie

salt Solution prior to autoclaving. A NaHC03 Solution (1.25 M, 10 1) was filter-

sterilized under ^-pressure with a 142 mm stainless steel filter holder and a 20 1

pressure tank, both delivered from Sartorius. After cooling üie fermenter to Üie

cultivation temperature of 30°C, Üie buffer Solution, üie filter-sterilized additives

(Dehning and Schink, 1989) and 1 mM cysteine/HCl were added. 12 1 of a culture,

grown for two days as described by Hilbi et al. (1992), was used as inoculum. The cultivation was carried out under an atmosphere of 80 % N2/ 20 % CO2 with genüe

stirring (50 - 100 rpm). After 44 hours, the bacteria were fed with another 50 mM neutralized malonic acid, and they were harvested after 4 days by continuous

centrifugation. The bacteria (220 g wet packed cells per 270 1 medium) were stored

under liquid nittogen without loss of malonate decarboxylase activity. The culture was

checked for purity with Üie agar shake culture method (Widdel and Pfennig, 1984).

Inhibition of subcellularfractions with avidin

To determine the inhibitory effect of avidin on the cytoplasmic and membrane fraction, the cell extract (0.15 ml, prepared as described by Hilbi et al. (1992)) was fractionated by ultracentrifugation (20 min, 200'000 x g). The membrane pellet was washed once with 0.5 ml buffer A (50 mM potassium phosphate, pH 7.5, 0.2 M NaCl,

5mM MgCy, ultracentrifuged again and resuspended in 0.1 ml buffer A.

Subsequenüy, Üie washed membranes as well as the cytoplasmic fraction were incubated 15 min either with 30 u.1 avidin (10 mg/ml), followed by Üie addition of excess biotin (40 |il, 7.5 mg/ml) or 30 min with 70 u.1 of a l:l-mixture of avidin and biotin, preincubated 15 min. To measure decarboxylase activity, Üie two subcellular 63

fractions had to be pooled. Malonate decarboxylase activity was quantified as ouüined

(Hilbi et al., 1992), analyzing the amount of released CO2 with a gas Chromatograph.

Immunolabeling for transmission electron microscopy

M. rubra and Propiomgenium modestum were anaerobically grown in 50 ml of a mineral salts medium (Dehning and Schink, 1989), supplemented with 30 mM malonate or 20 mM succinate, respectively. The cells were harvested in the exponential growth phase and subjected to high pressure freezing (Studer et al., 1989) in cellulose capillary tubes (200 um diameter, 2 mm length). Frozen samples were freeze- substituted in eüianol containing 0.5 % uranyl acetate. Samples were kept in the freeze-substimtion medium for 9 h at -90 °C, 6 h at -60 °C, 3 h at -30 °C and 1 h at

4 °C. They were then washed in anhydrous acetone at 4 °C and infiltrated at this temperature with increasing concenttaüons of Epon/Araldite.

Polyclonal rabbit antibodies against biotin (IgG fraction; Enzo Diagnostics, Inc.) were diluted 1:100, applied to the thin-sections and incubated for 2 hours. The bound antibodies were then marked with protein A-gold complexes (10 nm colloidal gold was prepared according to Slot and Geuze, 1985). Previous masking of unspecific protein binding Sites and post-staining was performed according to Schwarz and Humbel

(1989). To conttol unspecific labeling, incubation of the sections wiüi üie biotin antibodies was omitted. The samples were observed in a Hitachi H-600 transmission electton microscope at 100 kV accelerating voltage.

Photochemical bleaching ofavidin or oxidation with sodium periodate

The exttemely high binding affinity of avidin to biotin (Kj> ~ 10"15 M _1) can be reduced by Üie destruction of ttyptophan residues that participate in binding (Green,

1975). In a photochemical procedure, Üie cuvette containing 10 mg avidin/ml buffer B

(50 mM potassium phosphate, pH 7.5, 100 mM NaCl, 5 mM MgCy was placed at

10 °C into Üie intensive excitation beam (280 nm, slit widüi 20 nm) of a specttofluorophotometer (Shimadzu RF-5001 PC). Tryptophan destruction was 64

recorded by the decrease of fluorescence at 338 nm. Samples were withdrawn during a

5.5 h photoreaction time and analyzed for inhibiting malonate decarboxylase activity at a ratio of 150 u.g bleached avidin per mg exttact protein (0.1 ml extract). In separate assays, the reversion of inhibition was invesügated by adding subsequenüy 75 u.g of biotin (10 u.1). The samples designed to be coupled to a cyanogen bromide-activated

Sepharose column contained 30 or 20 mg avidin in 3 ml 10 mM potassium phosphate, pH 7.0 and were bleached for 2.5 or 4 h, respectively.

Altematively, Üie ttyptophan residues of avidin were oxidized with 1 mM Nal.04, to yield formyl kynurenine derivatives. This procedure leads to a product which possesses 70 % of the initial binding capacity and an decreased affinity for biotin (KD -

10-9 M"1, Green, 1975). 5 mg avidin/ml were incubated with 1 mM sodium periodate, pH 7.0 for 60 min and the progress of the reaction was foUowed as described above from Üie decrease of ttyptophan fluorescence at 338 nm as well as from Üie decrease of absorbance at 280 nm. Inhibition was tested by adding 150 u.g oxidized avidin to

0.1 ml exttact and reversion of inhibition was checked in a separate assay by subsequent addition of 75 u,g biotin. The Solution to be coupled to a column comprised

20 mg avidin in 4 ml 10 mM potassium phosphate, pH 7.0 and was incubated with

1 mM NaI04 for 40 min. The reaction was terminated with 10 mM cysteine/HCl, pH 7.5 (40 min).

Coupling ofavidin derivatives to Sepharose columns

Avidin or the chemically modified avidin derivatives (see above) were covalenüy linked to cyanogen bromide-activated Sepharose (coupling efficiency between 89 and

95 %). Monomeric avidin columns were prepared by the subsequent dissociaüon of the avidin tettamers with 6 M guanidinium/HCl (Dimroth, 1986).

Other methods

SDS-PAGE, Western blotting and protein determinations were performed as described (Hilbi et al., 1992). 65

EX 4. Results

Localization ofthe biotin containing protein ofmalonate decarboxylase

Malonate decarboxylase of Malonomonas rubra is a complex enzyme System that

consists of membrane-bound and cytoplasmic components (Hilbi et al., 1992). All

attemps to stabilize a putative membrane-bound malonate decarboxylase, e.g. by

different conditions of cell rupture and in the presence of various additives failed. One

of the components of malonate decarboxylase is a biotin-containing protein, as

evidenced from Üie inhibition of decarboxylase activity in Üie cell exttact by avidin

(Hilbi et al., 1992). The location of this biotin protein was invesügated by incubating

the cytoplasmic and Üie membrane fraction seperately with avidin, followed by

Saturation of excess avidin with bioün and recombination of the two fractions to

restore malonate decarboxylase activity. The results indicate no effect of avidin

treatment of the membrane fraction on malonate decarboxylase activity (0.87 U/mg

protein vs 0.93 U/mg protein for the conttol), white the corresponding incubation with

Üie cytoplasm led to complete inactivation. The conttol incubation of this fraction with

an avidin/bioün complex was without effect (0.87 U/mg protein of malonate decarboxylase activity). These results clearly indicate üie location of üie biotin- containing component of the malonate decarboxylase enzyme System in the cytoplasm.

This conclusion was confirmed by the results of Western blotting shown in Fig. 1.

Affinity staining of Üie blot with peroxidase conjugated avidin revealed a Single stained biotin-containing protein of an apparent molecular weight of 120'000 in Üie cytoplasm.

Traces of this protein seen in the membrane fraction are considered to result from adhering cytoplasm. Inspite the adherence of some biotin protein to the membrane, malonate decarboxylase activity was not detectable in this fraction, probably because other cytoplasmic enzymes required for this activity are missing (c.f. Fig. 4 and

Discussion).

White these results clearly indicate Üie cytoplasmic location of Üie biotin- containing decarboxylase component in the cell extract, they do not unequivocally deterraine its location within intact cells; a loosely membrane-bound biotin protein 66

might be dissociated from the membranes during cell rupture with a French press.

Another approach was Üierefore made to localize the biotin protein within the bacterial cells by electron microscopy. Cells of the exponenüal growth phase were frozen under high pressure, embedded into Epon/Araldite resin and thin-sectioned. These sections were then exposed to a bioün antibody and subsequenüy to protein A-gold and invesügated by electron microscopy. The results of Fig. 2A show a random distribuüon of the gold particles within the M. rubra cell and no specific enrichment at the cell membrane surface. In contrast, investigation of P. modestum cells by Üie same procedure clearly indicated a preference of Üie gold particles at the internal surface of

Fig. 1: Identification of biotin-containing proteins in cell fractions. Cell rupture and ultracentrifugation were performed as described under Experimental Procedures. The membrane fraction was washed twice with 50 mM potassium phosphate, pH 7.5 and was resuspended in üie same buffer. The samples applied to SDS-PAGE are from left to right: lane 1: a mixture of protein Standards with molecular masses of 205, 166.5, 80, 49.5, 32.5, 27.5 and 18.5 kD; lane 2: 20 ug extract; lane 3: 20 ug washed membranes and lane 4: 20 (j.g cytoplasm. After electrophoresis, the proteins were blotted onto nitrocellulose Sheets and biotin-containing Polypeptides were identified by reaction with an avidin-peroxidase conjugate. 67

the cytoplasmic membrane (Fig. 2B). P. modestum is known to contain the membrane- bound biotin enzyme meüiylmalonyl-CoA decarboxylase. The lower labeling density of

P. modestum as compared to M. rubra was found consistenüy. It may not reflect, however, different amounts of the biotin-containing proteins in these organisms but, more likely, could indicate different affmities of the biotin antibody for Üie two different enzymes. Controls, performed with both bacteria in which the incubation with

Üie biotin antibody was omitted, showed no gold labeling of the cells. We conclude

Üierefore that Üie biotin-containing protein participating in malonate decarboxylaüon in

M. rubra is located in the cytoplasm.

Attempts to purify the biotinylated component with modified avidin columns

Affinily chromatography on avidin-Sepharose columns is a widely used method for the purification of biotin proteins. The affinity of tettameric avidin towards biotin

-15 (KD - 10 M_1) is so high that elution of a biotin protein is only possible under denaturing conditions (Green, 1975). Monomeric avidin, however, has an about

8 Orders of magnitude lower biotin-binding affinity and biotin proteins can be eluted from such columns by simple replacement with free biotin (Green, 1975). Thus, all biotm-containing decarboxylases have been purified on monomeric avidin-Sepharose columns (for a review see Dimroth, 1987).

The biotin-containing cytoplasmic component of malonate decarboxylase, however, was not retained on monomeric avidin-Sepharose, as evidenced from

Western blot analyses after staining with avidin-peroxidase as well as from malonate decarboxylase activity determinations following reconstitution with Üie membrane.

About 50 - 100 % of reconstitutable decarboxylase applied was recovered in the

"break through" fraction and none was found after elution with biotin. This failure to bind to the monomeric avidin column was not due to an excess of free biotin in the cytoplasm, because üie same behavior of the biotin protein was observed after its precipitation with ammonium sulfate (40 - 70 % Saturation) and dialysis. 68

- w * # ** t ' * -* " f * >

• * km . * #fe v

* * , * ** * >'

;,-**-- * TV P:-

Fig 2 Transmission electronmicrographs of Malonomonas rubra (A) and Propiomgenium modestum (B) cells after immunolabeling of thin sections with anti biotin IgG and protein A-gold. The bar represents 300 nm. For details see Expenmental Procedures 69

In the presence of detergents (1.5 % Triton X-100 in the sample applied and

0.1 % Brij 58 in the column buffer) the bioün containing protein was retained and eluted with 2.4 mM biotin, albeit without catalytic activity. These results suggest that

Üie biotin prostheüc group of this protein is not readily available from the outside in its native conformation.

On the other hand, incubation with tetrameric avidin caused complete inhibition of malonate decarboxylase activity that could not be reversed by an excess of free biotin.

Therefore, an avidin derivative with a biotin binding affinity intermediate between Üie tetrameric and Üie monomeric form seemed desirable. As ttyptophan residues of avidin are known to participate in its tight binding of biotin (Green, 1975), we attempted to chemically modify part of these ttyptophan residues. Irradiation of avidin with intense light of 280 nm led to Üie destruction of ttyptophan residues, as shown by the continuous decrease of ttyptophan fluorescence at 338 nm and a concomitant red shift during a 5.5 h Irradiation period (data not shown). The results of inhibition of malonate decarboxylase activity by avidin bleached for various times, shown in Fig. 3, indicate a reduced inhibitory power after prolonged irradiation. It is also shown that Üie inhibition can be reversed parüy if an excess of free biotin is added to the avidin-biotin protein complex prior to the activity measurements. However, avidin bleached for 2.5 or 4 h and then coupled to Sepharose completely retained Üie biotin protein in an irreversible manner.

After oxidation of ttyptophan residues of avidin with sodium periodate Üie KD for biotin binding decreases to 10 ~9 M _1 (Green, 1975). Incubation of the exttact with

150 u.g of thus modified avidin per mg protein completely inhibited malonate decarboxylase activity. The inhibition could be reversed about 50 % by an incubation of the inactive enzyme with 2.4 mM biotin. Attempts to use Üiis oxidized avidin for affinity chromatography failed, however, because the bioün protein was not retained to a significant extent 70

In summary, affinity chromatography with such modified avidin columns is

Üierefore not applicable for the purification of the catalytically active biotin-containing

protein of the malonate decarboxylase of M. rubra.

12 3 4 5

Time of Irradiation (h)

Fig. 3: Inhibition of malonate decarboxylase with bleached avidin. Avidin irradiated at 280 nm for Üie times indicated was assayed for inhibition of malonate decarboxylase () as ouüined under Experimental Procedures. Excess biotin was added in separate assays to the avidin-tteated samples to test reversibility (•). The values are means of duplicate assays. 71

HI. 5. Discussion

The malonate decarboxylase System of Malonomonas rubra was shown recenüy

to be inhibited by avidin which indicates üie participation of a biotin-containing protein jn the catalysis (Hilbi et al., 1992). We have now shown that M. rubra contains a unique biotin protein of Mr 120'000 that is located in Üie cytoplasm not only after cell rupture by a French press but also if Üie cell integrity is preserved, as demonsttated by electton microscopy employing the immuno-gold technique. These results therefore exclude Üie existence of a malonate decarboxylase enzyme complex on Üie cytoplasmic membrane. It is interesüng in this context that üie sodium ion transport decarboxylase enzyme family (oxaloacetate decarboxylase, methylmalonyl-CoA decarboxylase and glutaconyl-CoA decarboxylase) consists of firmly membrane-bound biotin-containing complexes (for a review see Dimroth, 1987; Rohde et al., 1988).

Besides (he different location of Üie malonate decarboxylase System, there are other disünct features that discriminate it from the sodium ion transport decarboxylases mentioned above. White chemically activated subsüates containing a keto or a thioester residue in ß-position to the carboxyl group are decarboxylated by the latter enzymes, the malonate decarboxylase System uses free malonate as a substtate. It is interesüng, however, that an activation of malonate on üie enzyme precedes the decarboxylaüon event. The mechanism of malonate decarboxylation thus involves the intermediary formation of protein-bound malonyl Üiioesters and their subsequent decarboxylation to the respective acetyl Üiioesters (Hilbi et al., 1992). The key features of the chemistry of the decarboxylation event, as catalyzed by either of these decarboxylases, may thus be very similar.

Malonate decarboxylase comprises unique features for the activation of Üie substtate, i.e. conversion of malonate into its enzyme-bound thioester derivative and a distinct Organization of Üie individual proteins involved in üiis catalysis. A hypothetical reaction sequence, based on the data of üiis and our previous paper (Hilbi et al., 1992) is shown in Fig. 4. A soluble protein-SH: transferase is supposed to form the malonyl-

S-protein derivative from malonate and acetyl-S-protein (reaction 1). The malonyl-S- 72

protein is then used as the substtate for the decarboxylation by reactions that are very similar to those previously encountered by the biotin containing decarboxylases. We assume as a next step (reaction 2) üie carboxyl ttansfer from the malonyl-S-protein to the biotin protein of Mr 120'000, again taking place within üie cytoplasm. In Üie following (step 3), üie carboxy biotin protein of Mr 120'000 must move to the membrane, where it is decarboxylated by the membrane-bound decarboxylase that presumably couples Üiis step with Üie conversion of Üie free energy into ApNa+ or

AfiH+

Cytoplasm Membrane

" malonate \ s ACP-S-acetyl »-_ _- E-biotin-CO., N / > Au.Na+ (AliH+)

^- "^ \ ** "* / ^ acetate APP-S-malnnvl E-biotin ACP-S-malonyl'

(1) (2) (3)

Fig. 4: Hypothetical mechanism of malonate decarboxylase. El and E2 represent two domains on a Single protein or two distinct Polypeptides. For details see text.

A peculiarity of the biotinylated component of M. rubra is its rather large size

(Mr ~ 120'000) and its incapability to bind to monomeric avidin-Sepharose columns.

The large size of this protein suggests that it not only harbours the biotin carrier

protein domain but in addition that of the carboxylüansferase and possibly additional

domains for binding of malonyl and acetyl Üiioesters. The low binding affinity of Üie 73

biotin prosthetic group of Üiis protein towards monomeric avidin indicates exposure of the biotin more deeply to the inside than e. g. on üie other biotin-containing decarboxylases. Different availability of the biotin cofactor to avidin, however, is not without precedent; pyruvate carboxylase is inhibited by avidin much slower than other

biotin-containing enzymes (Green, 1975). Affinity chromatography on monomeric avidin-Sepharose has in the past much facilitated Üie study of the Na+ ttansport

decarboxylases (for a review see Dimroüi, 1987). Altough considerable effort was

undertaken to prepare an avidin affinity column with an appropriate binding constant

for the M. rubra biotin protein, this approach has unfortunately been not successful.

Attempts to purify üiis protein and the other component enzymes of üie malonate decarboxylase System by conventional purification procedures are now in progress in

our laboratory.

Hl. 6. References

Anantharam V, Allison MJ, Maloney PC (1989) Oxalate: formate exchange. J Biol

Chem 264: 7244-7250

Baetz AL, AUison MJ (1989) Purification and characterization of oxalyl-coenzyme A

decarboxylase from Oxalobacterformigenes. J Bacteriol 171:2605-2608

Dehning I, Schink B (1989) Malonomonas rubra gen. nov. sp. nov., a microaero-

tolerant anaerobic bacterium growing by decarboxylation of malonate. Arch

Microbiol 151: 427-433

Dimroth P (1986) Preparaüon, characterization, and reconstitution of oxaloacetate

decarboxylase from Klebsiella pneumoniae, a sodium pump. Methods Enzymol

125: 530-540

Dimroth P (1987) Sodium ion ttansport decarboxylases and other aspects of sodium

ion cycling in bacteria. Microbiol Rev 51: 320-340

Green M (1975) Avidin. Adv Prot Chem 29: 85-133 74

Hilbi H, Dehning I, Schink B, Dimroth P (1992) Malonate decarboxylase of

Malonomonas rubra, a novel type of biotin-containing acetyl enzyme. Eur J

Biochem 207:117-123

Hilpert W, Schink B, Dimroth P (1984) Life by a new decarboxylation-dependent

energy conservation mechanism with Na+ as coupling ion. EMBO J 3:1665-1670

Rohde M, Mayer F, Dutscho R, Wohlfarth G, Buckel W (1988) Immunocytochemical

localization of two key enzymes of Üie 2-hydroxyglutarate pathway of glutamate

fermentation in Acidaminococcusfermentans. Arch Microbiol 150: 504-508

Schwarz H, Humbel B (1989) Influence of fixatives and embedding media on

immunolabeling of freeze-substitited cells. Scanning Microsc Suppl 3: 57-64

Slot JW, Geuze HJ (1985) A new method of preparing gold probes for multiple-

labeling cytochemistry. Europ J Cell Biol 38:87-93

Studer D Michel M, Müller M (1989) High pressure freezing comes of age. Scanning

Microsc Suppl 3: 253-269

Widdel F, Pfennig N (1984) Dissimilatory sulfate or sulfur reducing bacteria. In: Krieg

NR, Holt JG (eds) Bergey's manual of systematic bacteriology vol. 1. Williams &

Wilkins, Baltimore London, pp 663-679 75

Chapter IV

Purification and characterization of a cytoplasmic enzyme component of the Na+-activated malonate decarboxylase System of Malonomonas rubra: acetyl-S-acyl carrier protein: malonate acyl carrier protein-SH transferase

Arch. Microbiol., in press

Hubert Hilbi and Peter Dimroth

Mikrobiologisches Institut, Eidgenössische Technische Hochschule, ETH-Zentrum,

Zürich, Switzerland

Qffprints requests to P. Dimroth

Abbreviations. DTNB, 5,5'-dithiobis (2-nittobenzoate); MES, 2-(N-Morpholino) ethanesulfonic acid; TAPS, N-[Tris(hydroxymethyl)-meÜiyl]-3-aminopropanesulfonic acid; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis

Enzymes. citramalate lyase (EC 4.1.3.22), cittate lyase (EC 4.1.3.6), malonate decarboxylase (EC 4.1.1.-)

Key words: malonyl-CoA:acetate CoA transferase - Na+ ttansport decarboxylases -

Na+ cycle - cittate lyase - citramalate lyase - CoA-like prostheüc group 76

IV. 1. Summary

Malonate decarboxylation by crude extracts of Malonomonas rubra was specifically activated by Na+ and less efficienüy by Li+ ions. The extracts contained an enzyme catalyzing CoA-ttansfer from malonyl-CoA to acetate yielding acetyl-CoA and malonate. After about a 26-fold purification of the malonyl-CoA:acetate CoA transferase, an almost pure enzyme was obtained, indicating that about 4% of the cellular protein consisted of the CoA transferase. This abundance of Üie transferase is in accord with its proposed role as an enzyme component of the malonate decarboxylase

System, the key enzyme of energy metabolism in this organism. The apparent molecular

weight of the Polypeptide was 67'000 as revealed from SDS-PAGE. A similar

molecular weight was estimated for the native transferase by gel chromatography,

indicating that Üie enzyme exists as a monomer. Kinetic analyses of Üie CoA transferase

yielded the foUowing: pH-optimum at pH 5.5, an apparent K,,, for malonyl-CoA of

1.9 mM, for acetate of 54 mM, for acetyl-CoA of 6.9 mM, and for malonate of

0.5 mM. Malonate or cittate inhibited the enzyme with an apparent Kj of 0.4 mM and

3.0 mM, respectively. The isolated CoA transferase increased the activity of malonate

decarboxylase of a crude enzyme System, in which part of the endogenous CoA

transferase was inactivated by borohydride, about three-fold. These results indicate that

Üie CoA transferase funcüons physiologically as a component of the malonate

decarboxylase System, in which it catalyzes the ttansfer of acyl carrier protein from

acetyl acyl carrier protein and malonate to yield malonyl acyl carrier protein and

acetate. Malonate is thus activated on the enzyme by exchange for the catalytically

important enzyme-bound acetyl thioester residues noted previously. This type of

substtate activation resembles Üie catalytic mechanism of cittate lyase and citramalate

lyase. 77

IV. 2. Introduction

Decarboxylaüon of malonate is regarded as Üie sole site of energy conservation of

Malonomonas rubra, an anaerobic, gram-negative, marine bacterium that grows from the fermentation of malonate to acetate and CO2 (Dehning and Schink 1989). The pivotal enzyme System, malonate decarboxylase, is related functionally and structurally to members of the sodium ion ttansport decarboxylase enzyme family (for a review see

Dimroth 1987). Common features of these enzymes are Üie participation of biotin in catalysis, activation by Na+, binding to Üie membrane and energization of Üie membrane by Converting decarboxylaüon energy into ApNa+. The Substrates are either a ß-keto acid (oxalacetate) or CoA derivatives of dicarboxylic acids (methylmalonyl-

CoA, glutaconyl-CoA).

Malonate decarboxylase also contains a catalytically active biotin, and Üie cytoplasmic membrane is indispensably involved in the decarboxylation event (Hilbi et al. 1992). However, Üiis enzyme uses free malonate (not malonyl-CoA) as the substtate. Interestingly, the catalytically active decarboxylase carries an acetyl thioester residue (Hilbi et al., 1992). In the course of the decarboxylaüon reaction Üie enzyme- bound acetyl residues are believed to be exchanged for malonyl residues, thus activating malonate on Üie enzyme (step 1, Scheme I). The decarboxylation presumably involves

Cytoplasm Membrane

" malonate v w y ACP-S-acetyl ^* E-biotin-C02\ / Y I —¥ > ^Na+(AjlH+)

*"* ' x acetate ACP-S-malonyl E-biotin *--*C02

(1) (2) (3)

Scheme I. Proposed malonate decarboxylation mechanism (Hilbi et al. 1993), ACP: acyl carrier protein 78

two different Steps: carboxyl transfer from the malonyl acyl carrier protein to the prostheüc biotin group (step 2, Scheme I) that regenerates Üie acetyl acyl carrier protein and Üie subsequent decarboxylation of Üie carboxy biotin (step 3, Scheme I) which should be coupled to ApNa+ or ApH+ generation in order to conserve decarboxylation energy. Unlike other members of the biotin-containing membrane- bound decarboxylase family, malonate decarboxylase does not exist as a tight complex on the cytoplasmic membrane, but rather is composed of at least three proteins located in Üie cytoplasm as well as in Üie membrane as shown in Scheme I (Hilbi et al. 1993).

The proposed mechanism of malonate activation on Üie enzyme by an exchange of malonate for an acetyl thioester residue on the enzyme (step 1, Scheme I) is very similar to Üie mechanism of substtate activation by cittate lyase or cittamalate lyase

(Dimroth and Eggerer 1975; Buckel and Bobi 1976). These enzymes carry acetyl thioester residues on a unique phosphoribosyl dephospho-CoA prostheüc group ffMmroth 1976, 1988; Dimroth and Loyal 1977; Robinson et al. 1976) that are exchanged in the first catalytic step by citrate or cittamalate. The transferase activity of these enzymes can be convenienüy determined with Üie prostheüc group analogue CoA derivatives (Dimroth et al. 1977a). We show here that malonyl-CoA and acetyl-CoA can also be used as substtate analogues to determine the acetyl-S-acyl carrier proteimmalonate acyl carrier protein-SH tranferase of M. rubra. A convenient assay thus became available which could be used for the purification of Üie transferase. In complementation studies the purified ttansferase was demonsttated to participate indeed in the decarboxylaüon of malonate. We also show in this communication that malonate decarboxylation is Na+- or Li+-dependent, indicating a close relationship of the malonate decarboxylase System with Üie other members of the Na+ ttansport decarboxylase family (for a review see Dimroüi 1987). 79

IV. 3. Materials and methods

Materials

The resins Fractogel TSK-DEAE and Fractogel TSK-Butyl were supplied from

Merck. Superose 12 (analytical and preparative grade), the Mono Q column, acetyl-

CoA and HS-CoA were from Pharmacia. Cittate synthase and the calibiation proteins for gel chromatography were from Boehringer Mannheim and Üie prestained protein

Standards (broad ränge) were from BioRad. MES was purchased from Serva, lithium

Chloride (Suprapur) was from Merck and all other reagents were from Fluka.

Malonomonas rubra was a gift from B. Schink (Konstanz, Germany). The bacteria were grown anaerobically on malonate in a 300 1 fermenter as described (Hilbi et al.

1993).

Determination ofCoA transferase activity

A) During the purification procedure the fractions were analyzed for malonyl-

CoA: acetate CoA transferase acüvity with a coupled spectrophotometric test in which

CoA-SH liberated from acetyl-CoA by the cittate synthase reaction was determined with DTNB. The cuvette contained in a total of 1 ml at 30°C: 50 mM potassium phosphate (pH 7.5), 100 mM sodium acetate, 1.0 mM oxaloacetate, 1.0 mM DTNB,

2.2 U cittate synthase and 0.1 mM malonyl-CoA. The reaction was initiated with 10 u.1 of the fractions (20 - 200 ug protein, depending on the stage of purification) and increase of absorption at 405 nm was monitored (Ae = 14 mM"' cm"', Buckel et al.

1981).

B) The kinetic parameters of the CoA ttansferase were assayed in a discontinuous assay separating and determining malonyl-CoA and acetyl-CoA by HPLC. Unless otherwise stated, the reactions were performed in a total volume of 60 ul containing

30 mM MES (pH 6.0), 100 mM sodium acetate, 1 mM malonyl-CoA and 3.2 (ig enzyme. The assay was started with the addiüon of enzyme and stopped after 1 min wiüi 5 ul 70% perchloric acid. With this amount of protein the kinetics of acetyl-CoA 80

formation were linear during 1.5 min. Prior to analysis by HPLC the samples were diluted 10 times with buffer A [0.2 M potassium phosphate (pH 5.0)].

The analysis of CoA Compounds (King and Reiss 1985) as described by Hoffmann et al. (1989) was further simplified as follows: A Hypersil ODS column (250 mm x

4 mm, particle size 5u.m, Hewlett-Packard) was equUibrated with buffer A at a flow rate of 1 ml/min and the acyl-CoA Compounds were eluted with a linear gradient (52 -

60% within 9 min) of buffer B [0.2 M potassium phosphate (pH 5.0), containing 20%, by volume, acetonittile], detected at 254 nm and quantified by comparing peak areas with those of known Standards. The areas increased linearly between 1 and 15 nmol.

Retention times were as follows malonyl-CoA 3.1 min, for HS-CoA 7.2 min and for acetyl-CoA 8.4 min.

The pH Optimum of the CoA ttansferase reaction was assayed in a polybuffer

System containing in a volume of 60 ul: 100 mM sodium acetate (pKa = 4.8), 30 mM each of MES (pKa = 6.2), potassium phosphate (pKa2 = 6.8) and TAPS (pKa = 8.4),

adjusted to the desired pH. To exclude pH effects on the stability of Üie enzyme, Üie

following conttol was performed. CoA ttansferase (25 u.g in 10 u.1) was incubated for

2.5 min in 50 u.1 polybuffer at pH 4.5, 5.0, 6.0 or 9.0, and subsequenüy diluted 15-fold

in polybuffer (pH 6.0). In a total volume of 360 u.1, activity was assayed immediately

with Üie discontinuous test as described above. At pH 4.5, 5.0 and 9.0, Üie enzyme lost

36%, 27% and 32% activity, respectively, as compared to pH 6.0.

Purification ofthe soluble malonyl-CoA: acetate CoA transferase

Frozen cells of M. rubra (10 g wet cells) were suspended in 30 ml buffer and

disrupted by a French press as described in Hilbi et al. (1992). The cell free exttact was

ulttacentrifuged (200'000 x g, 1 h) and the cytoplasm was subjected to ammonium

sulfate precipitation. The pellet from 45 - 70 % Saturation (0°C) was dissolved to in

buffer I [50 mM potassium phosphate (pH 7.5), final volume 10 ml], desalted over a

Sephadex G-25 column (1.6 x 20 cm) equilibrated with buffer I, and subsequenüy

pumped onto a Fractogel TSK-DEAE column (1.6 x 12 cm) connected to an FPLC 81

apparatus (Pharmacia). The column was equUibrated wiüi buffer I and the following gradient of buffer II [50 mM potassium phosphate (pH 7.5), 1 M NaCl] was applied at

2.5 ml/min: 0 - 40 ml, 0% buffer II; 40 - 50 ml, 0 - 10% buffer II; 50 - 200 ml, 10 -

25% buffer II. CoA ttansferase activity eluted in a broad peak between 110 and 180 mM NaCl. The pooled fractions were adjusted to 1.5 M ammonium sulfate, centrifuged

(20'000 x g, 10 min) and loaded onto a Fractogel TSK-Butyl column (1 x 15 cm) equUibrated with buffer III [50 mM potassium phosphate (pH 7.5), 1.5 M ammonium sulfate]. Within 72 min, a linear gradient of buffer I (0 - 100%) at 2.5 ml/min was applied and CoA ttansferase eluted between 0.9 and 0.7 M ammonium sulfate. The active fractions were concenttated to about 2 ml by ulttafiltration (PM-10 membrane,

Amicon) and chromatographed in aliquots of 1 ml on a Superose 12 preparatory grade column (1.6 x 50 cm) equUibrated with buffer IV [100 mM potassium phosphate, (pH

7.5)] at 1 ml/min. Fractions containing CoA ttansferase activity were diluted with one volume of buffer V [20 mM Tris/Cl (pH 7.5)] and loaded onto a Mono Q column (0.5 x 5 cm). At a flow rate of 0.5 ml/min, a linear gradient of buffer VI [20 mM Tris/HCl

(pH 7.5), 1 M NaCl] was applied (0 - 35% within 60 min) and CoA ttansferase eluted at around 250 mM NaCl. Aliquots of these fractions (- 1 mg protein/ml) were stable upon storage in liquid nitrogen. The purification protocol is summarized in Table 1.

Native molecular weight

The estimation of the apparent native molecular weight was carried out with an analytical Superose 12 column (10 x 300 mm) equUibrated with 100 mM sodium phosphate (pH 7.1), at a flow rate of 0.25 ml/min. The elution volume of 80 u.g CoA ttansferase was compared with those of Standard proteins (75 ug each); activity of üie enzyme was determined with Üie coupled assay.

Participation ofthe CoA transferase in the decarboxylation of malonate

Washed membranes of M. rubra [0.55 g; prepared from 10 ml cell exttact according to Hübi et al. (1992)] were resuspended in buffer C [50 mM potassium 82

phosphate (pH 7.5), 50 mM NaCl, 1 mM MgCy. The cytoplasm was subjected to

40% ammomum sulfate precipitation and resuspended in buffer C. A portion of Üiis

Suspension (45 u.1; enriched in the biotin protein) was reduced with 2.5 u.11 M sodium borohydride in 1 M NaOH, immediately foUowed by 2.5 ul 1 M HCl in order partiaüy to inactivate CoA ttansferase.

After 15 min, 50 ul each of sodium borohydride-treated 40% ammonium sulfate precipitate (0.24 mg protein), resuspended membranes (0.57 mg protein), and purified

CoA ttansferase (0.07 mg protein) were combined in a 1.3 ml HPLC vial and 3 mM

ATP was added to assure reacetylaüon of the partially deacetylated enzyme samples

(HUbi et al. 1992). In separate conttol assays, either one of the fractions was replaced with 50 ul buffer C, or ATP was omitted. After incubaüng 2 min, Üie reaction was started with 30 mM sodium malonate (pH 7.5) and malonate decarboxylase activity was quantified by gaschromatography as ouüined in HUbi et al. (1992). The detection limit of malonate decarboxylase with this assay was below 4 mU.

Other methods

Sodium dodecyl sulfate-polyacrylamide electrophoresis, Western blotting, and protein determinations were performed as described in Hilbi et al. (1992)

IV. 4. Results

Dependence ofmalonate decarboxylation on Na+ ions

The demonstration of a specific activation by Na+ is a convenient method to indicate the participation of this metal ion in Üie energy Converting mechanism. The dependence of malonate decarboxylase of a crude exttact from Malonomonas rubra on

Na+ or Li+ concenüation is shown in Fig. 1. The activation profiles for both cations clearly follow Saturation kinetics indicating K^ for Na+ of about 0.8 mM and K,,, for

Li+ of about 3.3 mM. Simüar Kn, values for Na+ have been found for the other Na+- translocating decarboxylases (for review see Dimroth 1987). Some activity (about 10% 83

of Vmax) was also present in the absence of Na+ ions. Analogous fmdings have been

made for the Na+ activation profüe of glutaconyl-CoA decarboxylase (Buckel and

Semmler, 1982), whereas oxaloacetate decarboxylase was completely inactive in the

4 6 8 10 25 50 75 100

Salt Concentration (mM)

Fig. 1. Activation of the malonate decarboxylase by Na+ or Li+ ions. Crude ceU-free exttact of Malonomonas rubra (27 mg/ml) was passed over a Sephadex G-25 column (1.6 x 20 cm) equUibrated with sodium-free 50 mM potassium phosphate buffer (pH 7.5) for the removal of Na+ ions (residual sodium concentration 50 u.M). Addition of 1 mM sodium free potassium ATP and 1 mM magnesium acetate to the column eluted exttact ted to a 57 % increase of activity and these reagents were therefore present in aU assays. The Uberation of CO2 from 30 mM potassium malonate in the presence of the Na^ (•) or Li+ () concenttations indicated was measured as described in Hübi et al. (1992). Preparation of the sodium-free buffer and determinations of sodium by atomic adsorption spectroscopy are ouüined in Kluge and Dimroüi (1992). 84

absence of Na+ ions (Dimroth and Thomer 1986). K+ ions were unable to acüvate malonate decarboxylase since these ions were present in the assay mixture in 50 mM concentration and further addition of K+ up to 150 mM had no effect

Purification ofa soluble malonyl-CoA:acetate CoA transferase

The simUarities between malonate decarboxylase and cittate- or citramalate lyase

(see Introduction) prompted us to determine whether CoA was also a functional analogue for Üie transfer reaction catalyzed by Üie malonate decarboxylase System

(Eq 1). This reaction would provide a convenient assay for the transferase with commerciaUy available CoA derivatives rather than üie acyl carrier protein Compounds.

malonyl-CoA + acetate ^ acetyl-CoA + malonate (1)

Acetyl-CoA formation was monitored in a coupled spectrophotometric assay with citrate synthase by measuring the Uberation of CoA-SH with DTNB. The results indicated Üie presence of a malonyl-CoA:acetate CoA transferase in crude ceU-free extracts of M. rubra with a specific acüvity of 14 mU/mg protein under our assay conditions. For practical reasons the malonyl-CoA concentrations applied (0.1 mM)

were far below Saturation (see below). After ultracentrifugation, 70% of (he CoA

transferase acüvity remained in Üie orange-red cytoplasm, indicating that it is a soluble

enzyme. The enzyme was purified two-fold by fractionated precipitation with

ammonium sulfate, and subsequenüy Üie desalted fraction was chromatographed over

four columns (Fractogel TSK-DEAE, Fractogel TSK-Butyl, Superose 12 and

Mono Q), leading to a final purification of 26-fold. Table 1 gives an overview of the

entire purification procedure.

CoA ttansferase acüvity eluted from the DEAE column in a broad peak,

reproducibly comprising two to three subpeaks. Western blot analysis of three

subpeaks obtained after separating cytoplasm on the DEAE column revealed the 85

Table 1. Purification of a soluble malonyl-CoA:acetate CoA-ttansferase. The enzyme was assayed with üie coupled specttophotomettic test described in Materials and Methods.

Step Activity Protein Specific Yield Purifi¬ (U) (mg) activity (%) cation (mU/mg) (-fold)

Crude extract 10.0 714 14 100 1

Cytoplasm 7.0 568 12 70 - Ammonium sulfate (45- 70 %) and desalting 7.2 339 21 72 2 Fractogel TSK-DEAE 19.4 259 75 194 5 Fractogel TSK-Butyl 16.0 106 151 160 11 Superose 12 9.1 40 228 91 16 MonoQ 5.5 15 367 55 26

presence of the 120 kD biotin protein (HUbi et al. 1993) in only one of them (data not

shown). Therefore, broadening of the elution peak could reflect different populations of

CoA ttansferase, i.e., partial segregation of Üie biotin protein from Üie CoA ttansferase.

The total activity after chromatography of the desalted ammonium sulfate precipitate

on the DEAE column increased about two to three times. This phenomenon could be

explained in the same way: Separation of protein(s) from the CoA ttansferase could

lead to a more favorable conformation.

SDS-PAGE of the individual stages of the purification protocol is shown in Fig. 2

and indicates the purification of a Single Polypeptide with an apparent Mr of 67 000.

The sharp band on the top with a Mr of 120 000 (arrow) comprises the biotin protein

as shown by Western blot analysis. This Polypeptide was completely removed after the

Butyl column (step 4). After 26-fold purification almost pure CoA ttansferase was

obtained indicating that Üiis enzyme comprises about 4% of Üie protein in the crude

extract; this abundance is reflected by the strong band wiüi a M, of 67 000 observed by 86

SDS-PAGE of the cytoplasm (Fig. 2, lane 1). Clearly, the transferase is one of the major protein components present in a M. rubra cell which is in accord with a participation of this enzyme in energy metabolism.

12 3 4 5 6 7

Fig. 2. SDS-Gelelectropherogram of CoA transferase samples at different Steps of purification. The lanes of a 10 % Polyacrylamide gel contained (from left to right): 10

ug cytoplasm (1), 10 ug desalted 45 - 70 % ammonium sulfate precipitate (2), 5 ug each of the pooled active fractions after the following columns: Fractogel TSK-DEAE (3), Fractogel TSK-Butyl (4), Superose 12 (5) and Mono Q (6). Lane 7 shows protein Standards with molecular masses of 205, 116.5, 80, 49.5, 32.5, 27.5, 18.5, and 6.5 kD, respectively. The arrow marks the biotin protein (identified by Western blot analysis).

Native molecular weight

Purified CoA transferase was chromatographed over an analytical Superose 12

column and the elution volume was compared with that of Standard proteins in order to 87

determine the native molecular weight The active enzyme eluted in a Single peak with

13 ml elution volume, corresponding to a molecular weight of 59'000 (Fig. 3). It is, therefore, concluded that Üie native CoA transferase is a monomer.

pH profüe ofthe CoA transferase reaction

The pH profüe of the transformation rate of malonyl-CoA and acetate to acetyl-

CoA and malonate showed a Sharp maximum at pH 5.5 (Fig. 4). The activity declined very sharply at lower pH values and more slowly at higher pH values.

lO -

I14i N, Volume N> O

Elution 00 \

- 6 i ' i i '

4.0 4.5 5.0 5.5 6.0

Log Molecular Weight

Fig. 3. Determination of the apparent native molecular weight of CoA ttansferase. The foUowing Standard proteins (75 \ig each) were applied on an analytical Superose 12 column: chymottypsinogen A (Mr 25'000), hen egg albumin (Mr 45'000), bovine serum albumin (Mr68'000), aldolase (Mr 158'000), ferriün (Mr405'000). The arrow indicates Üie elution volume of CoA transferase. For details see Materials and Methods. 88

The stability of Üie enzyme at extreme pH values (Fig. 4) was about 70% of that at pH 6.0 (see Materials and Methods). The Sharp decrease in acüvity between pH 5.5 and 4.5, therefore, only partially reflects denaturation of the enzyme, but mainly appears to be due to protonation of a catalytically important amino acid.

Fig. 4. pH-rate profüe of CoA ttansferase. CoA transferase activity was determined at different pH-values using a polybuffer System as described in Materials and Methods. The data shown are the means of dupUcate assays that contained 1 mM malonyl-CoA and 100 mM sodium acetate. The amount of acetyl-CoA produced within 2 min was analyzed by HPLC. 89

Kinetics ofihe CoA transfer

Initial velocities of acetyl-CoA formation from malonyl-CoA and acetate at varied

malonyl-CoA (Fig. 5A) or acetate concentration (Fig. 5B) with fixed concentrations of

Üie cosubsttate foUowed Saturation kinetics. At 100 mM acetate, üie apparent Vn^

was 63 U/mg protein and the apparent Kn, for malonyl-CoA was 1.9 mM. At 2 mM

malonyl-CoA, the apparent V,^, was 49 U/mg protein and Üie apparent K^ for

acetate was 54 mM.

Saturation kinetics were also obtained for the reverse reaction, the formation of

malonyl-CoA from acetyl-CoA and malonate (data not shown). At a fixed concentration of 0.5 mM malonate, the apparent V,^ was 0.23 U/mg protein and the

0 12 3 0 50 100 150 200

Malooyl-CoA (mM) Acetate (mM)

Fig. 5. Kinetic constants of CoA ttansferase for malonyl-CoA (A) and acetate (B). The kinetic constants of CoA ttansferase for malonyl-CoA and acetate were determined in dupUcate as ouüined in Materials and Methods quanüfying the acyl-CoA Compounds by HPLC. The concentrations of the cosubsttates were kept constant: 100 mM sodium acetate (A) or 2 mM malonyl-CoA (B), respectively. The insets show double-reciprocal plots. 90

apparent K,,, for acetyl-CoA was 6.9 mM. At a fixed concentration of 10 mM acetyl-

CoA, the apparent Vmax was 0.31 U/mg protein and Üie apparent Km for malonate was

0.5 mM.

Double-reciprocal plots of initial velocity versus acetate concentrations at four different malonyl-CoA concenttaüons lead to an apparent Km for acetate of 40 (±

10) mM (Fig. 6). Since these lines are obviously not parallel, the reaction does not foUow ping-pong-type kinetics like the classic CoA transferases (Jencks 1973).

0.25 -

0.20-

"3i E D 5 °-15 '> a // 3

p o Specific Ä. // >

0.05 - S*^^

- 0.00 i 1 1 1

25 50 75 100 [Acetate]1 (M)'1

Fig.6. Kinetics of acetyl-CoA formation. Initial velocity is plotted double-reciprocaUy versus acetate concentrations at 0.5 mM (V), 1 mM (A), 2 mM () and 3 mM (•) malonyl-CoA. The values are means of duplicate assays. 91

The transformation of malonyl-CoA and acetate to acetyl-CoA and malonate was strongly inhibited by the product malonate and to a lower extent by cittate (Fig. 7).

Added to an assay containing 100 mM sodium acetate and 1 mM malonyl-CoA the apparent Kj was 0.4 mM for malonate and 3.0 mM for citrate. The apparent Kj for malonate was about the same as Üie Km for this Compound, suggesting that malonate is

Inhibitor (mM)

Fig. 7. Kinetics of Üie inhibition of CoA transferase wiüi malonate or cittate. Malonate () or cittate (•) was added in üie concenttaüons indicated to disconünuous assays containing in dupUcate 100 mM sodium acetate and 1 mM malonyl-CoA. For details see Materials and Meüiods. In Üie inset, percent of inhibition and Üie concenttations of inhibitor are plotted double-reciprocaUy. 92

not an inhibitor, but rather a substtate preventing formation of acetyl-CoA according to

Eq. 2:

malonyl-CoA + malonate ^ malonate + malonyl-CoA (2)

Among other substtate analogues tested, Oxalate and Propionate had no effect on the CoA ttansferase reaction up to 10 mM. Succinate, methylmalonate or glycolate

(10 mM each) resulted in about 30% inhibition, whereas the same concenüation of malate inhibited acetyl-CoA formation by 45%. Only glyoxylate (1-10 mM) effected an inhibition comparable to citrate (80% inhibition at 10 mM glyoxylate; data not shown).

Participation ofthe isolated CoA transferase in malonate decarboxylation

In order to verify the anücipated involvement of the isolated CoA ttansferase in the physiological reaction of malonate decarboxylation, we studied the effect of adding

CoA ttansferase to a System containing the other enzymes necessary for the catalysis

(40% ammonium sulfate precipitate and membrane fraction). Without further treatment of the ammonium sulfate fraction, the addition of purified CoA transferase did not increase Üie malonate decarboxylase activity, which could indicate that another component protein other than CoA ttansferase was rate-Umiting. However, after destruction of part of the CoA transferase of the ammonium sulfate fraction by incubation wiüi sodium borohydride, the addition of isolated CoA ttansferase increased

Üie malonate decarboxylase activity about three-fold (Table 2). We conclude, Üierefore, that üie isolated CoA ttansferase is a component enzyme of the malonate decarboxylase System and that the physiological reaction is the ttansfer of an acyl carrier protein with its prostheüc thiol group from an acetyl residue to a malonyl residue. Omission of the membrane or Üie ammonium sulfate precipitate resulted in complete loss of malonate decarboxylase activity, indicating that in addition to CoA

ttansferase, at least two more proteins are required to catalyze malonate 93

Table 2. Participation of CoA ttansferase in the decarboxylation of malonate. The complete System contained purified CoA ttansferase (0.07 mg protein), membranes (0.57 mg protein), and 40% ammonium sulfate precipitate (0.24 mg protein) treated for 15 min with 50 mM sodium borohydride for partial inactivation of CoA ttansferase of this fraction. The inactivation procedure is ouüined in Materials and Methods. Malonate decarboxylase was assayed by gaschromatography.

Omission Specific Activity (%) (mU/mg)

None 148 100

CoA ttansferase 57 39

Ammonium sulfate precipitate (40%) 0 0

Membrane 0 0

3 mM ATP 42 28

decarboxylation. The 40% ammonium sulfate precipitate also contained acetylating

activity, as is indicated by the strong effect of ATP (see Hilbi et al. 1992).

IV. 5. Discussion

The specific activation of malonate decarboxylase by Na+ or Ii+ described in üiis paper is in accord with Üie participation of Üiis enzyme in a membrane-bound energy conservation mechanism wiüi Na+ as coupling ion that is common to aU members of

üie Na+ ttansport decarboxylase famüy (for a review see Dimroüi 1987). Dehning and

Schink (1989) observed growth of Malonomonas rubra on malonate also under

microaerobic conditions (5% O2). The large amount of a periplasmic cytochrome c, however, seemed not to be involved in oxidative phosphorylation since in microaerobiosis growth was retarded and acetate was not utilized. Furthermore, no other inorganic electron acceptors tested were reduced. Therefore, ATP synthesis depends entirely on the free energy of malonate decarboxylation and the above results suggest that M. rubra is another anaerobic bacterium in addition to Propiomgenium 94

modestum (Hilpert et al. 1984) that couples ATP synthesis via a Na+ cycle to a decarboxylation reaction.

Here and elsewhere (HUbi et al. 1992, 1993) we demonstrated that the malonate decarboxylase of M. rubra consists of several different protein components parüaUy located in the cytoplasm and partially located in the membrane. One of the cytoplasmic proteins carries an acetyl thioester residue essential for catalysis that is biosyntheücally attached by an enzymic process with ATP and acetate as subsüates (HUbi et al. 1992).

Activation of malonate, which is required to faciUtate its decarboxylation, does not involve malonyl-CoA but protein-bound malonyl thioester residues that are formed from the acetyl-S-acyl carrier protein and malonate by virtue of an acyl carrier protein-

SH ttansferase reaction (Eq. 3) that is described in this communication.

acetyl-S-acyl carrier protein + malonate ^ malonyl-S-acyl carrier protein + acetate (3)

Interestingly, the acyl carrier protein-SH transferase not only accepts the protein- bound acetyl and malonyl thioester residues as subsüates, but is also functional with üie

CoA derivatives. This finding provided the basis to measure the ttansferase with readüy avaüable CoA derivatives rather than the physiological, but unavaUable acyl carrier protein Compounds. The enzyme could thus be purified and some of its properties could be studied. The rationale for the extended specificity of the ttansferase may be the binding of the acyl residues to Üie acyl carrier protein via a CoA-like prostheüc group. Cittate lyase and citramalate lyase, enzyme complexes that are related to malonate decarboxylase in the mechanism of substtate activation, carry an acyl carrier protein subunit with a 5'-phosphoribosyl dephospho-CoA prostheüc group covalenüy attached by phosphodiester bonding to specific serine residues (Dimroth 1976, 1988;

Dimroüi and Loyal 1977; Robinson et al. 1976). In the lyases, Üie prostheüc group becomes acetylated by a specific ligase with acetate and ATP (Schmellenkamp and

Eggerer 1974), and Üie acetyl residues are exchanged for cittate or citramalate residues in the initial step of the reaction that is catalyzed by üie transferase component of the 95

enzyme complex (Dimroüi et al. 1977a,b). These ttansferases likewise not only catalyze

an acyl carrier protein-SH transfer, but also a CoA ttansfer from acetyl-CoA to citrate

or cittamalate. Thus, there is a süiking simUarity in the activation of malonate with that

of cittate and cittamalate by the respective enzyme System. We do not know yet

whether the malonate decarboxylase System includes a small acyl carrier protein (Mr -

lO'OOO) as in citrate lyase or cittamalate lyase (Dimroth and Eggerer 1975; Dimroth et

al. 1977a). The acyl carrier protein could altematively be located on a domain of the

unusuaüy large cytoplasmic biotin protein (Mr 120'000) that is involved in the

decarboxylaüon of malonate (HUbi et al. 1993).

Unlike these simUarities in the first part of the reaction mechanism, the second part

of the malonate decarboxylaüon mechanism is completely distinct from that of the

lyases. Whüe in malonate decarboxylaüon a soluble biotin protein and a membrane-

bound enzyme are involved and Üie reaction seems to be coupled to a vectorial

translocation of Na+, the lyases contain a distinct subunit that catalyzes Üie cleavage of

the citryl or citramalyl thioester to an acetyl thioester and an oxoacid (oxalacetate or

pyruvate) by a Claisen-type mechanism (Dimroüi and Eggerer 1975; Dimroth et al.

1977a).

As described above, Üie acyl carrier protein transferases of malonate decarboxylase

and the lyases also function as CoA ttansferases in unphysiological in vitro assays.

These enzymes, however, have distincüy different properties from the true CoA

ttansferases that are abundant in bacterial and other cells. The K,,, for the acyl-CoA

derivative is usually in üie ränge of about 20-200 uM for the physiological CoA ttansferases and about 1-10 mM for the unphysiological CoA transferases (acyl carrier protein transferases). For some of the physiological CoA ttansferases the intermediate formation of an enzyme-CoA derivative has been demonsttated (Jencks 1973; Sramek and Frerman 1975; Tung and Wood 1975; Buckel et al. 1981), whüe the CoA

ttansferase of cittate lyase performs catalysis without such an intermediate (Dimroth et al. 1977b). The formation of enzyme-CoA intermediates was demonsttated by destruction of enzymic activity in üie presence of borohydride and the acyl-CoA 96

substtate smce the thioester between coenzyme A and an acidic amino acid of the enzyme is reduced under these conditions with unpairment of catalytic acüvity (Jencks

1973, Sramek and Frerman 1975, Tung and Wood 1975, Buckel et al 1981) A malonyl-CoA-dependent Inhibition of the isolated CoA ttansferase from M rubra by

borohydnde was not observed, 20 mM NaBH^ 1 mM malonyl-CoA mcubated for

15 min rnhibited only 20%, whereas 50 mM NaBH4 incubated for 10 mm resulted in 50

- 70% loss of acüvity independent of whether Üie thioester substtate was present or

not Therefore, there is no lndicaüon for the intermediate formaüon of a covalent

enzyme-CoA denvative on the CoA ttansferase of M rubra Double-reciprocal plots of

initial velocity versus acetate concenlraüon at different concenttations of malonyl-CoA

did not show parallel hnes as would be expected for a prng-pong-type kinetic

mechanism that is charactensüc for the true CoA transferases, but not for üie CoA ttansferase components of cittate lyase or cittamalate lyase that use acyl carner protein

Üiioesters as Üie physiological Substrates In addition, these enzymes do not involve covalent CoA-denvatives m catalysis The CoA ttansferase of M rubra is therefore

more closely related to the CoA ttansferases of citrate lyase and cittamalate lyase than

to the physiological CoA ttansferases

Based on the above descnbed drfferences between true CoA ttansferases and acyl carner protein-SH ttansferases, one might discnminate two famihes of CoA

ttansferases Whereas Üie true CoA ttansferases, acüng physiologicaUy on low molecular weight Compounds, prevent diffusion of the thiol intermediate by Üie formaüon of a covalent thioester bond, the high-molecular-weight protein intermediate of the acyl carner protein-SH ttansferases is presumably stabüized through (multiple) non-covalent protein-protein interacuons and the formaüon of a covalent bond is

Üierefore unnecessary

It is of interest that the maximal velocity of CoA ttansfer from acetyl-CoA to

malonate is about 200-ümes lower than that of CoA ttansfer from malonyl-CoA to acetate The formaüon of malonyl-CoA by the CoA ttansferase thus seems to be kinetically inhibited Malonate, on the other hand, mhibited the formation of acetyl- 97

CoA from malonyl-CoA and acetate with a K, of 0.4 mM. If this inhibition would extend to Üie reaction wiüi Üie physiological subsüates and products acetyl-S-acyl carrier protein and malonyl-S-acyl carrier protein, the rate of the back reaction

(formation of acetyl-S-acyl carrier protein) would be significantiy decreased. The high

apparent K„, for acetate (54 mM) and low apparent K,,, for malonate (0.5 mM) should

also contribute to drive the acyl carrier protein ttansfer in the physiological direction,

i.e., formation of malonyl-S-acyl carrier protein from acetyl-S-acyl carrier protein and

malonate.

Acknowledgements. We would like to thank M. Weiss and H. R. Schläfli of the

Biodegradation group at the ETH-Zentrum for providing their weU-mamtained HPLC

device.

IV. 6. References

Buckel W, Bobi A (1976) The enzyme complex cittamalate lyase from Clostridium

tetanomorphum Eur J Biochem 64: 255-262

Buckel W, Dorn U, Semmler R (1981) Glutaconate CoA-transferase from

Acidaminococcus fermentans. Eur J Biochem 118: 315-321

Buckel W, Semmler, R (1982) A biotin-dependent sodium pump: glutaconyl-CoA

decarboxylase from Acidaminococcus fermentans. FEBS Lett 148:35-38

Dehning I, Schink B (1989) Malonomonas rubra gen. nov. sp. nov., a

microaerotolerant anaerobic bacterium growing by decarboxylation of malonate.

Arch Microbiol 151:427-433

Dimroth P (1976) The prostheüc group of citrate-lyase acyl-carrier protein. Eur J

Biochem 64: 269-281

Dimroth P (1987) Sodium ion ttansport decarboxylases and other aspects of sodium

ion cyling in bacteria. Microbiol Rev 51: 320-340 98

Dimroth P (1988) The role of vitamins and their carrier proteins in cittate fermentation.

In: Kleinkauf H, von Döhren H, Jaenicke L (eds) The Roots of Modern

Biochemistry. de Gruyter, Berlin New York, pp 191-204

Dimroth P, Eggerer H (1975) Evaluation of the protein components of cittate lyase

from Klebsiella aerogenes. Eur I Biochem 53:227-235

Dimroüi P, Loyal R (1977) Structure of the prostheüc groups of cittate lyase and

cittamalate lyase. FEBS Lett 76: 280-283

Dimroth P, Buckel W, Loyal R, Eggerer H (1977a) Isolation and function of the

subunits of cittamalate lyase and formation of hybrids with the subunits of cittate

lyase. Eur J Biochem 80: 469-477

Dimroüi P, Loyal R, Eggerer H (1977b) Characterization of the isolated ttansferase

subunit of cittate lyase as a CoA-transferase. Evidence against a covalent enzyme-

substrate intermediate. Eur J Biochem 80:479-488

Dimroth P, Thomer A (1986) Kinetic analysis of Üie reaction mechanism of

oxaloacetate decarboxylase from Klebsiella aerogenes. Eur J Biochem 156: 157-

162

HUbi H, Dehning 1, Schink B, Dimroth P (1992) Malonate decarboxylase of

Malonomonas rubra, a novel type of biotin-containing acetyl enzyme. Eur J

Biochem 207: 117-123

HUbi H, Hermann R, Dimroth P (1993) The malonate decarboxylase enzyme System of

Malonomonas rubra: evidence for Üie cytoplasmic location of the biotin-

containing component Arch Microbiol 160:126-131

Hüpert W, Schink B, Dimroth P (1984) Life by a new decarboxylation-dependent

energy conservation mechanism wiüi Na+ as coupling ion. EMBO J 3: 1665-1670

Hoffmann A., Hilpert W, Dimroth, P (1989) The carboxylttansferase acüvity of Üie

sodium-ion-ttanslocating methylmalonyl-CoA decarboxylase of Veillonella

alcalescens. Eur J Biochem 179: 645-650

Jencks WP (1973) Coenzyme A transferases. In: Boyer PD (ed) The Enzymes, 3rd edn,

vol 9B. Academic Press, New York, pp 483-496 99

King MT, Reiss PD (1985) Separation and measurement of short-chain coenzyme-A

Compounds in rat Uver by reversed-phase high-performance Uquid

chromatography. Anal Biochem 146:173-179

Kluge C, Dimroth P (1992) Studies on Na+ and H+ ttanslocation through the F0 part

of the Na+-translocating FiF0 ATPase from Propiomgenium modestum:

discovery of a membrane potential dependent step. Biochemistry 31: 12665-12672

Robinson JB, Singh M, Srere PA (1976) Structure of the prostheüc group of

Klebsiella aerogenes cittate (pro-3S)-lyase. Proc Naü Acad Sei USA 73: 1872-

1876

SchmeUenkamp H, Eggerer H (1974) Mechanism of enzymic acetylation of des-acetyl

citrate lyase. Proc Naü Acad Sei USA 71:1987-1991

Sramek SJ, Frerman FE (1975) Escherichia coli coenzyme A-transferase: kinetics,

catalytic pathway and structure. Arch Biochem Biophys 171:27-35

Tung KK, Wood WA (1975) Purification, new assay, and properties of coenzyme A

transferase from Peptostreptococcus elsdenii. J Bacteriol 124:1462-1477 100

Appendix I

N-terminal sequence of the CoA transferase

Experimental

9 pg of purified CoA ttansferase (see Chapter IV) was loaded onto a lane of a gradient SDS-polyacrylamide gel (16, 10, 4 % acrylamide) without glycerol and urea.

Tricine SDS-PAGE was done according to Schägger and Jagow (see Appendix II).

The proteins were electtoblotted onto a PVDF membrane (semidry-blotting, discontinous buffer System; Pharmacia), stained with 0.1 % Coomassie brillant blue

G250 and the 67 kD band was cut out. The N-terminal sequence was analyzed by Edman degradation by Dr. Peter James (Protein Chemistty Facüity, ETH Zürich) with an Applied Biosystems Sequencer (model 470 A) and a HPLC phenylthiohydantoin derivative analyzer (model 120 A).

Results

The foUowing sequence of 29 amino acids was obtained by Edman degradation:

(M)QKEK VWDKL STDTE ERMNA ANELF SDRK

The third glutamate residue (position 16) was determined with some uncertainity. 101

Appendix II

Preliminary experiments on the purification of the biotin protein of malonate decarboxylase and other protein components involved in malonate decarboxylation

Abbreviations. ABTS, 2,2'-azino-bis(3-etiiylbenzothiazoline-6-sulfonic acid)

diammonium salt; BSA, bovine serum albumin; DMB, 3,3'-dimethoxybenzidine;

ELISA, enzyme linked immuno sorbent assay; PBS, phosphate-buffered saline; PVDF, poly (vinyUdene difluoride) 102

Summary Incubaüon of the cytoplasmic fraction of Malonomonas rubra with hydroxylamine and dialysis or a 40-70 % ammomum sulfate precipitaüon and dialysis caused a complete loss of malonate decarboxylase activity The acüvity was restored only partially and only enzymaücaUy by adding ATP/Mg2+/acetate Fractionation of the cytoplasm with either an anion exchange column (Fractogel TSK-DEAE) or a preparaüve gel filtraüon column (Superose 12) resulted in the loss of C02-hberaüng activity The acüvity was not restored upon acetylating the fractions containing the bioün protem

The flowthrough of an overloaded DEAE column containing the bioün protein catalyzed the decarboxylaüon of malonate after recombination with washed membranes of M rubra In contrast, Üie fracüon containing biotin protein, which was eluted from üie DEAE column by a NaCl gradient, was macüve after recombination with washed membranes In an analogous recombination assay, the fractions of a gel filttation column containing the bioün protein were also macüve Hence, diluüon and/or Separation of a further protein (disünct from the bioün protem) is responsible for the loss of malonate decarboxylase activity after these columns

In recombination assays containing punfied acetyl-S-acyl camer protein malonate acyl carner protein-SH transferase and washed membranes, the bioün protein was added as 40 % ammomum sulfate precipitate A pool of DEAE fractions of cytoplasm

(without the fracüons containing the bioün protein) was concenttated by ulttafilttaüon and added to the former components This pool caused a 6 fold increase of malonate decarboxylase activity, which was dependent on ATP/Mg2+/acetate Hence, this pool is assumed to compnse either a putative acyl carner protein disünct from the bioün protein and/or deacetyl malonate decarboxylase acetate ligase

CatalyticaUy inactive bioün protein was assayed with Western blots using a slot blot apparatus With this test the 120 kD bioün protein of the malonate decarboxylase enzyme System was partially punfied from cytoplasm with three consecutive columns

(Fractogel TSK-DEAE Superose 12 and Fractogel-TSK-Butyl) The protein was blotted onto PVDF membranes and subjected to Edman degradation Whereas the

BSA conttol yielded a readable sequence, the N-terminus of Üie bioün protein was blocked

Washed membranes of M rubra did not catalyze Üie decarboxylaüon of CO2- biotm in absence of Na+ 103

Introduction

Malonate decarboxylase of M. rubra is a complex enzyme System, composed of membrane-bound and several cytoplasmic components including an unusuaUy large,

120 kD biotin protein (HUbi et al, 1992). In addition, the enzyme was shown to be activated specificaUy by Na+- (or Li+-) ions (HUbi and Dimroüi, 1994) and may üius be a member of the Na+-ttanslocating decarboxylase enzyme famUy (Dimroth, 1987). The labUity of malonate decarboxylase has severely hampered purification and characterization of the component enzymes/proteins in the past. Furthermore, a monomeric avidin column, suitable for the purification of other membrane-bound Na+- ttanslocating decarboxylases did not retain the cytoplasmically located biotin protein of malonate decarboxylase in a catalyticaUy active form (HUbi et al, 1993).

Malonate decarboxylase acts on free malonate and not on malonyl-CoA OHUbi et al, 1992). The indispensable activation of the substtate takes place directly on the enzyme by forming a protein-bound malonyl thioester in exchange for an acetyl residue. The enzyme catalyzing this activation (acetyl-S-acyl carrier protein: malonate acyl carrier protein-SH ttansferase) has been highly purified and characterized by using the low molecular weight substtate analogue malonyl-CoA and acetate as a cosubsttate (HUbi and Dimroüi, 1994). On Edman degradation of Üiis purified CoA ttansferase 29 residues of the N-terminal sequence were determined (see Appendix of Chapter IV).

An activity assay using a low molecular weight substtate analogue would also be advantagous for determining the membrane-bound component(s) of malonate decarboxylase. Bendrat and Buckel (1993) have shown that the 65 kD carboxylttansferase subunit (GCDA) of the glutaconyl-CoA decarboxylase of

Acidaminococcus fermentans not only accepts the 24 kD bioün carrier protein as a substtate for carboxyl transfer, but also free biotin, although with a high Km of

51 mM. Purified GCDA, therefore, should provide a convenient means for the in situ synthesis of the chemicaUy labüe C02-biotin Compound. Possibly, üiis could serve as an alternative substtate for the membrane-bound decarboxylase subunit(s). Malonate decarboxylase is activated postttanslationaUy by acetylation, which is catalyzed by a specific ligase (HUbi et al, 1992). In ceU free extracts deacetylated malonate decarboxylase was reactivated either enzymaticaUy with ATP/Mg2+/acetate or chemicaUy with acetic anhydride. The experiments described below indicate that the acetyl residue of Üie malonate decarboxylase enzyme System is hydrolyzed very easily and the reacetylation procedure was successful so far only enzymaticaUy. AvaUabUity of the deacetyl malonate decarboxylase: acetate Ugase would, therefore, facUitate Üie purification of the active cytoplasmic components of malonate decarboxylase. 104

Materials and Methods

Materials

A cartridge of Fractogel EMD-SO3"-650 (S) was supplied from Merck. PVDF membranes were purchased from MiUipore and nittoceUulose membranes were from Amersham. Polyethylene imine was from Aldrich and Polyethylene glycol 6000 from

Fluka. The sources of aU other materials and chemicals are the same as in HUbi et al.

(1992) and Hilbi and Dimroth (1994), respectively.

Deacetylation and precipitations of the cytoplasmic components of malonate decarboxylase

Cytoplasmic components of the malonate decarboxylase were completely deacetylated by incubaüng cytoplasm wiüi 0.2 M neutralized hydroxylamine for 2.5 h at 4 °C. Excess hydroxylamine was removed by passing the sample over a Sephadex

G-25 column (1.6 x 20 cm) equUibrated with 50 mM potassium phosphate, pH 7.5, 50 mM NaCl, 1 mM MgCl2.

Fractionation with ammonium sulfate (40 - 70 %) was performed at 4 °C (15 min incubation, 15 min centrifugation at 15'000 x g). The precipitations with polyeüiylene glycol and polyeüiylene imine were done with concenttated stock soluüons (pH 8.0) of (50 %, w/v) and (10 %, w/v), respectively.

Activity assay ofthe membrane-bound decarboxylase compound(s)

In an attempt to assay the membrane-bound component(s) of malonate decarboxylase with Üie analogue C02-biotin, Üiis Compound was generated with enzymes of Acidaminococcus fermentans (Buckel, 1986). Briefly, the carboxyl¬ üansferase subunit GCDA of glutaconyl-CoA decarboxylase (GCD) catalyzes the ttansfer of the CO2 moiety of glutaconyl-CoA to free biotin. Purified GCDA, which has been expressed in a heterologous System (Bendrat and Buckel, 1993), and a pool of five auxüiary enzymes (glutaconate CoA ttansferase, enoyl-CoA hydratase (crotonase), (3S)-3-hydroxybutyryl-CoA dehydrogenase, acetyl-CoA acetyl- ttansferase (thiolase) and phosphate transacetylase) were kindly provided by Prof. Dr. W. Buckel, Marburg. A cuvette contained in 1 ml: 50 mM potassium phosphate, pH 7.0, 1 mM NAD+, 0.1 mM acetyl-CoA, 20 mM d-biotin (neutralized stock Solution of 0.5 M), 10 ul auxüiary enzymes and 10 ul GCDA. The reaction was started with 10 mM glutaconate and Üie increase of absorption at 340 nm was foUowed. For the GCDA a biotin- dependent activity of 0.8 U/ml was determined.

The assays designed to test C02-biotin decarboxylating activity of membranes of

M. rubra were performed in a rubber sealed 1.3 ml HPLC via! and contained 150 ul of 105

the continous assay mixture (preincubated 20 min) described above and either 50 ul

washed membranes or, as a conttol, 50 ul 50 mM potassium phosphate, pH 7.5. CO2-

release was analyzed gaschromatographicaUy as described (Hilbi et al, 1992).

Activity assay ofthe biotin protein

Participation of Üie biotin protein in the decarboxylation of malonate was assayed

in a recombination assay containing 50 ul each of the foUowing components: (1) washed membranes, (2) partially purified biotin protein, (3) purified acetyl-S-acyl

carrier protein: malonate acyl carrier protein-SH ttansferase and (4) a pool (free of biotin protein) containing the putative deacetyl malonate decarboxylase: acetate Ugase

and/or a putative acyl carrier protein distinct from Üie biotin protein.

0.3 g membranes, frozen in Uquid nitrogen, were washed and resuspended in 2 ml

50 mM potassium phosphate, pH 7.5, 50 mM NaCl. Source of the biotin protein was

either a 40 % ammonium sulfate precipitate OHUbi and Dimroüi, 1994), frozen in Uquid

nitrogen, or the fractions containing biotin protein separated with the columns described below. The acetyl-S-acyl carrier protein: malonate acyl carrier protein-SH

ttansferase was purified as described (HUbi and Dimroth, 1994). As a source of the

putative deacetyl malonate decarboxylase: acetate ügase, aU the fractions after a

Fractogel TSK-DEAE column except fractions 16-18 containing the biotin protein

were concenttated with a PM 10 membrane from 240 ml to 9 ml (pool D). In the

assay, 50 ul each of washed membranes, biotin protein fraction, concenttated DEAE- fractions without the biotin protein (pool D) and 10 ul of purified acetyl-S-acyl carrier

protein: malonate acyl carrier protein-SH ttansferase were combined. Omitted

fractions were replaced by the respective volume of 50 mM potassium phosphate, pH

7.5, 50 mM NaCl. If not indicated otherwise, 3 mM ATP/Mg2+/acetate was present in

all assays.

Assays ofthe inactive biotin protein

A) Avidin-peroxidase assay on polystyrene microtiter plates

Bioün protein in the cytoplasm was assayed in analogy to the antibody capture

assay described by Harlowe and Lane (1988). The wells of a polystyrene microtiter

plate were incubated for 2 hours with 50 ul cytoplasm in different dilutions (1 ug - 1 mg/well), washed with phosphate-buffered saline Solution (PBS) and blocked with

3 % BSA in PBS. 50 ul avidin-peroxidase Solution (5 ng - 5 ug/weU), containing 1 %

BSA in PBS, were then incubated for 1 hour at room temperature. The wells were washed extensively and Üie chromogenic Compound ABTS (5 mg/ml in 50 mM cittate-

phoshate buffer pH 4.0) and 0.003 % H202 were added. The reaction was stopped

with 10 ul 37 mM KCN and quantified at 415 nm with an ELISA reader. 106

B) Western blot analysis on nitrocellulose using a slot blot apparatus

Cytoplasm at different diluüons or fractions eluted from columns on which the biotin protein was separated were assayed for the bioün protein with Western blot analysis using a slot blot apparatus (72 slots, Schleicher & SchueU) The dry nittoceUulose membrane was mounted on üie slot blot apparatus and each slot was washed with 0 4 ml PBS After the samples (50 ul each) were loaded applying a vacuum, the nittoceUulose membrane was taken out and the blot was further treated according to HUbi et al (1992) The blot was incubated with avidin-peroxidase

(1 ug/ml in 1 % BSA) to detect bioün protein bound to the membrane With 3,3'- dimethoxybenzidine as chromogemc Compound the bioün protein present in 1 to

100 ug cytoplasm/slot was detected as discrete, strong band, which increased in intensity with increasing amounts of protein over 4 Orders of magnitude 1 mg cytoplasm/slot and above led to an apparent overloading of the nittoceUulose sheet and the stain blurred Assuming a bioün protein content of not more than 1 % of the cytoplasmic proteins of M rubra, bioün protein from approximately 1 ug/slot down to

1-10 ng/slot was detected

Partial purification ofthe biotin protein

A) Monomeric avidin column

The native bioün protein of Malonomonas rubra is not retained by monomenc avidin columns, probably because the bioün moiety is not accessible (HUbi et al,

1993) Attemps were undertaken to bind üie bioün protein m a partially denaturated conformation in presence of detergents or chaottopic agents 0 95 ml cytoplasm were incubated with either 1 % Tnton X-100, 1 M urea or 0 2 M sodium thiocyanate and subsequenüy chromatographed over a monomenc avidin column (Dimroth, 1986) The column was equihbrated with 50 mM potassium phosphate pH 7 5, 0 1 M NaCI (buffer

A) containing either 0 1 % Bnj 58, 0 1 M urea or 20 mM sodium thiocyanate, corresponding to the denatunng agent used in the assay After washmg, reversibly bound bioün protein was eluted with buffer A containing 1 mM bioün and Üie column was regenerated as descnbed (Dimroüi, 1986) ß) Conventional columns

10 ml of cytoplasm, prepared according to HUbi et al (1992) was pumped on a

Fractogel TSK DEAE column (16 x 12 cm), equUibrated with buffer I (50 mM potassium phosphate, pH 7 5) The foUowing program was run at 2 5 ml/mm 0-60 ml, 0 % buffer II (= buffer I plus 1 M NaCl), 60-230 ml, 0-35 % buffer II The bioün protein eluted in a Sharp peak between 70 and 120 mM NaCl ( data not shown) The fractions 16 18 (24 ml), containing bioün protein were concenttated to 1 ml by ulttafilttation and chromatographed on a preparative Superose 12 column (1 6 x 107

50 cm), which was equUibrated with buffer III (= buffer I plus 50 mM NaCl) at

1 ml/min. Fractions 5-8, which contained most of the biotin protein, were concenttated to 1.3 ml. Ammomum sulfate was added to the final concenüation of 0.78 M, incubated for 75 min at 4 °C and centtifuged for 30 min (20'000 x g). The supematant was loaded on a Fractogel TSK-Butyl column (1x15 cm) equUibrated with buffer IV

(= buffer I plus 0.78 M ammomum phosphate). A linear gradient of buffer I (0-100 % within 70 ml) was appüed at 1 ml/min and the biotin protein eluted only at the very end of the run at 100 % buffer I. Fig. 1 shows a slot blot of the fractions obtained after each of the three columns described above and on Fig. 2 a SDS gel of the pooled fractions containing the biotin protein is depicted. Furthermore, the fractions after the Superose 12 column, which contained the biotin protein, were separated on a cation exchange column or on a Mono Q column.

The Fractogel EMD-SO3--650 (S) column (1 x 15 cm) was equUibrated with buffer V

(20 mM potassium phosphate, pH 6.0 or pH 6.5) at 1 ml/min and a gradient (0-100 % within 50 ml) was run with buffer VI (= buffer V plus 1 M NaCl). A Mono Q column

(0.5 x 5 cm) was equUibrated with buffer VII (50 mM Tris/ Cl, pH 7.5) at 0.5 ml/min and a gradient (0-35 %) was applied with buffer VIII (= buffer VII plus IM NaCl) within 35 ml.

N-terminal sequencing

Partially purified biotin protein was separated by Tricine SDS-PAGE according to

Schägger and Jagow (1987), using a gradient gel (16, 10, 4 % acrylamide) without glycerol and urea. The proteins were electtoblotted onto a PVDF membrane (semidry- blotting, discontinous buffer System; Pharmacia), stained with 0.1 % Coomassie

Brillant Blue G250 and the 120 kD band was cut out. As a conttol Üie biotin protein was blotted on nittoceUulose and affinity stained with avidin-peroxidase. The N- terminal sequence was analyzed by Edman degradation by Dr. Peter James (Protein

Chemistty Facüity, ETH Zürich) with an Applied Biosystems Sequencer (model

470 A) and a HPLC phenylthiohydantoin derivative analyzer (model 120 A).

Other methods

Hemoprotein staining in Polyacrylamide gels with 3,3'-dimethoxybenzidine was performed according to Francis and Becker (1984). 108

Results

Stability of the membrane-bound and cytoplasmic components of the malonate decarboxylase enzyme System Whereas ceU free extracts of M. rubra lost 80 % activity wiüiin 18 hours at 4 °C

(HUbi et al, 1992), the membrane component(s) were stable for one week at 4 °C in 50 mM potassium phosphate, pH 7.5, 50 mM NaCl. Activity of washed membranes in recombination assays was not affected by freezing in Uquid nitrogen without previous washing. In conttast, the cytoplasmic components lost 40 to 60 % acüvity within 24 h at 4 °C, which could not be restored by enzymatic reacetylation. No activity of the cytoplasmic components could be measured after storage at -20 °C for two weeks.

Activity assay ofthe membrane-bound decarboxylase compoundfs)

Subunit a of glutaconyl-CoA decarboxylase catalyzed the formaüon of approximately 0.17 mM C02-biotin within 20 min, calculated from the increase of

Aj^ in the continous assay. 150 ul of Üie incubation mixture were added to 50 ul

washed membranes of M. rubra or, as a conttol, to Üie same amount of potassium

phosphate buffer. Whereas the A340 of the parallel running continous assay increased

wiüiin 90 min to 2.49, corresponding to a C02-biotin concentration of 0.4 mM, no

C02-release by the samples was detected within this period. Upon acidification, however, 70 % of the anticipated amount of C02 was detected in the gas phase (11 nmol/ 0.3 ml head space). In order to increase the amount of C02-biotin formed, the concentration of free biotin was doubled and a continous assay containig 40 mM

d-biotin was performed. In this assay an amount of CO2 corresponding to 0.6 mM was released (without acidification) wiüiin 35 min, irrespective, whether washed membranes were present or not. However, no Na+ was added to these assays and since decarboxylation of malonate is specificaUy stimulated by Na+- and Li+-ions (HUbi and Dimroth, 1994), it would be expected that the membrane component(s) of malonate decarboxylase catalyze(s) C02-release from C02-biotin only in presence of these alkali ions.

Partial purification of catalytically inactive biotin protein Polystyrene microtiter plates turned out to be unsatisfactory to quantitate the

biotin protein, since in the ränge of three Orders of magnitude the colour development

of the "ELISA" depended on the concentration of avidin-peroxidase and not on the concentration of the bioün protein applied to the microtiter plate wells. The binding capacity of polystyrene was apparenüy too low to bind sufficient amounts of biotin

protein. On the other hand, nittoceUulose wül bind about 1000-fold more protein per 109

unit surface area (Harlow and Lane, 1988), and therefore, this mattix was used to detect the biotin protein after Separation on different columns.

A) Retention ofthe biotin protein on a monomeric avidin column

It was reported previously that the bioün protein could not be purified in a catalyücaUy active form with a monomeric avidin column or with photochemically or oxidatively modified tetrameric avidin columns. On incubation with 1.5 % Triton

X-100 and addiüon of 0.1 % Brij 58 to Üie elution buffer, a smaU amount of reversibly bound biotin protein was eluted from the monomeric avidin column with 1 mM biotin as shown with Western blots of a SDS gel (Hilbi et al, 1993).

In order to investigate more systematically the effect of denaturing agents on the reversibility of biotin protein binding to a monomeric avidin column, cytoplasm was incubated with 1 % Triton X-100, 1 M urea or 0.2 M thiocyanate, respectively. The flowthrough fractions of the column were subsequenüy analyzed by the slot blot technique. No bioün protein was detected in the assay containing 1 % Triton X-100 and thus, the bioün protein was quantitatively bound. 1 M urea led to an almost quantitative binding of Üie biotin protein. 0.2 M thiocyanate also caused bindmg of most of the bioün protein as judged from a conttol performed with a biotin saturated column which retains no biotin protein. However, in no case biotin protein was eluted after applying 1 mM biotin in presence of Triton X-100/ Brij 58 and thus, binding of the biotin protein to the monomeric avidin column was irreversible in the above assays. The previously reported elution of the biotin protein from the monomeric column

under these conditions OHübi et al, 1993) may be explained with the increased sensitivity of a Western blot compared to Üie slot blot technique applied here, since in the SDS PAGE the sample is concenttated prior to blotting. B) Separation ofthe biotin protein with conventional columns

Cytoplasm of M. rubra was fractionated with three consecutive columns and Üie fractions were analyzed for the biotin protein with the slot blot technique (see Materials and Methods, Fig. 1). Whereas the first purification step, üie anionic exchange column, was very efficient the resoluüon of the foUowing gelfilttation column was not at its Optimum in the ränge of interest (120 kD). Neiüier a reduction of the flow rate from 1 ml/min to 0.5 ml/min, nor alternative gel filttation materials (Sephacryl S-200, Fractogel TSK HW-65) could increase the resolution. The biotin protein eluted from a hydrophobic interaction column only after the gradient had been finished and the column had been washed wiüi one volume of low salt buffer. The cytoplasmically located bioün protein, üierefore, seems to be raüier hydrophobic.

The bioün protein-containing fractions eluted from the Superose 12 column were chromatographed also on other columns. However, they barely improved the purification of the bioün protein. A cation exchange column (Fractogel EMD-SO^") HO

was run at pH 6.0 and pH 6.5 with little effect on the purity of the biotin protein.

Furthermore, whereas pH 6.0 caused a large portion ofthe biotin protein to precipiate,

Fig. 1. Slot blot of the fractions after the foUowing columns: Fractogel TSK-DEAE (A), Superose 12 (B) and Fractogel TSK-Butyl (C) 50 ul each of the fractions indicated on the left of the lanes were applied onto a nitrocellulose membrane as outlined in Materials and Methods. P20 refers to the resuspended pellet of a 20 % (0.78 M) ammonium sulfate precipitate of fractions 5-8 (lane B). The supematant of this precipitation was applied to the Fractogel TSK-Butyl column. 111

at pH 6 5 the binding of the biotin protein to the column was not quantitative anymore Analysis of the pooled fractions containing the biotm protein after a Mono

Q column by SDS PAGE did not reveal a higher purity of the biotm protein (data not shown) Chromatography of the biotin protem on the three colums shown m Figure 1 led to a partial purification of this protein (Fig 2)

Fig 2 SDS gelelectropherogram of fractions ennched in the biotin protein The lanes of a SDS gradient gel (16, 10, 4 % acrylamide) contained 40 ug cytoplasm (1) and the pooled fractions after the foUowing columns Fractogel TSK DEAE, 40 ug (2), Superose 12, 20 ug (3), and Fractogel TSK Butyl, 10 ug (4) respectively Lane 5 shows protein Standards with molecular masses of 205, 116 5, 80, 49 5, 32 5, 27 5 18 5, and 6 5 kD, respectively The arrow at 120 kD marks the biotin protein

The periplasmic cytochrome c described by Dehning and Schink (1989), was ennched by Separation of cytoplasm on a Fractogel TSK-DEAE column At 250 mM

NaCl a deep red fraction eluted from the anionic exchange column (data not shown)

The difterence spectrum (dithionite reduced minus ammonium peroxydisulfate 112

oxidized) of this fraction showed the typical c-type cytochrome maxima at 552, 522

and 426 nm, respectively. Herne stain of the cytoplasm of M. rubra separated by SDS- PAGE revealed three dominant bands at 14,29 and approximately 150 kD.

The protein with an apparent molecular mass of approximately 30 kD, which was considerably enriched after the gelfiltration column (Fig. 2, lane 2) was not red. This

protein, present in Üie biotin protein fraction, is eluted from Üie DEAE column at

around 100 mM NaCl and, therefore, does not represent the abundant cytochrome c mentioned above, which is eluted at 250 mM NaCl from the DEAE column. The function of this protein is not known.

Participation ofthe biotin protein in malonate decarboxylation As outlined in HUbi et al (1992), malonate decarboxylase activity in ceU free

extracts that had been inactivated by deacetylation with 0.2 M hydroxylamine could be reactivated partially either enzymaticaUy with 20 mM DTT/5 mM ATP (65 %

reactivation) or chemicaUy with 1 mM acetic anhydride (16 % reactivation). In

conttast, deacetylated cytoplasm recombined with washed membranes, recovered no

activity upon incubation with 10 mM DTE and 0.5 to 5 mM acetic anhydride. Enzymic

reactivation with 10 mM DTE, 3 mM ATP, 2 mM acetate, however, was successful

leading to an activity of 0.78 U (30 % reactivation as compared to the not inactivated cytoplasm). Ammomum sulfate precipitation of the cytoplasm (40-70 %) and subsequent dialysis completely abolished malonate decarboxylase activity. In Üie past Üie activity

could only be recovered by enzymatic acetylation (46 % recovery as compared to untteated cytoplasm). Up to 20 % Polyethylene glycol 6000 added to the cytoplasm did not precipitate activity, whereas 0.5 % Polyethylene imine precipitated 57 %

activity (compared to cytoplasm diluted accordingly). However, the pellet was very

viscous and activity was not enriched.

Any further fractionation of cytoplasm with a Fractogel TSK-DEAE column led to

a complete and irreversible loss of C02-Uberating activity. Activity of fractions containing the biotin protein could be restored neither with 5 mM ATP/Mg2+/acetate,

nor with 1 mM acetic anhydride. In both attempts purified CoA ttansferase was

present. In conttast, the flowthrough of an overloaded DEAE column, containing the

biotin protein, retained C02-releasing activity. Two explanations are conceivable: (1) In the flowthrough, the biotin protein is not separated from another protein involved in

decarboxylation, possibly a discrete acetylated acyl carrier protein, and therefore, this fraction remains catalyücaUy active. (2) The 120 kD biotin protein, carrying also the

acyl carrier domain, is deacetylated upon binding to Üie anion exchange column and

recovers its catalytic activity only upon reacetylation. 113

Fractions enriched in biotin protein did (in presence of ATP/Mg2+/acetate) not

increase malonate decarboxylase activity of ceU free exttacts of M. rubra which were inhibited with avidin followed by Saturation with bioün.

Cytoplasm passed over a preparative Superose 12 column totally lost its malonate decarboxylase activity. The activity of fractions containing the biotin protein could be

restored neither by addition of ATP/Mg2+/acetate or 5 mM DTT/0.5 mM acetic anhydride, nor by the addition of the supematant of heat inactivated cytoplasm

(100 °C, 1 min). Purified CoA transferase was present in the assays. Neither a more alcaline pH of 8.5 nor heterologously expressed deacetyl cittate lyase: acetate Ugase of

Klebsiella pneumoniae (plus DTE/ATP/Mg2+/acetate) had a positive effect on

reactivation. Only if aU the fractions of the gel filttation column were ulttaconcentrated

approximately 40-fold, some low activity (60 mU) was regained upon incubation with

20 mM DTT and 5 mM ATP/Mg2+/acetate. The gel filttation column, therefore, seems to separate protein components required for malonate decarboxylation. In order to make available the putative deacetyl malonate decarboxylase: acetate

Ugase (or other protein(s) involved in malonate decarboxylation) aU the fractions of a DEAE column, except the biotin protein containing (fractions 16-18, confer Fig. 1),

were ulttaconcentrated extensively from 240 ml to 9 ml. This pool, termed "D", indeed

had a pronounced effect on the C02-release of subcytoplasmic fractions. The increase

of malonate decarboxylase activity amounted to 6 fold as is evident comparing line 2 and line 5 of Table 1. The protein component(s) of D remained active at least for three

days at 4 °C. However, depending on dilution, unfractionated cytoplasm was stül 2-3-

times more active (line 1).

Without ATP üie System was almost inactive (line 3), indicating that üie acetyl moiety is easüy lost and (enzymatic) reacetylation is crucial for recovering activity. Purified CoA ttansferase increased Üie C02-release four times (compare line 2 and 4).

Washed membranes recombined either with the 40 % ammonium sulfate precipitate or

fraction D alone were completely inactive (line 6 and line 7, respectively). An involvement of partially purified biotin protein in malonate decarboxylation could not be demonsttated (lines 8 and 9). Whereas the assay containing apparenüy no cytoplasmic bioün protein showed an acüvity of as much as 43 % (Une 8), the addition of biotin protein partially purified over two columns even decreased the C02-release

(line 9). In principle, assay 8 was not expected being acüve in malonate decarboxylation because the indispensable biotin protein is not added exogenously.

Presumably, some biotin protein stiU adheres to once washed membranes in a catalytically active form (see Discussion). 114

Table 1. Recombination assays of subcytoplasmic fractions of the malonate decarboxylase enzyme System. Each assay contained 50 ul washed membranes of M. rubra and 3 mM ATP/Mg2+/acetate. In assay 3 the ATP was omitted. Abbreviations: B, source of the biotin protein; CT, purified CoA ttansferase; D, concenttated DEAE fractions of chromatographed cytoplasm, without Üie fractions containing the biotin protein; C, cytoplasm; AS, resuspended 40 % ammonium sulfate precipitate; S(D), fractions containing the bioün protein after two columns (DEAE and Superose 12, respectively). For details see Materials and Methods.

B CT D Activity (U) (%) Assay

1) C - - 1.70

2) AS + + 0.54 100

3) AS + + 0.03 6

4) AS - + 0.13 24

5) AS + - 0.09 17

6) AS - - 0 0

7) - - + 0 0

8) - + + 0.23 43

9) S(D) + + 0.16 30

N-terminal sequencing Samples of the bioün protein after the gel filttation column and after the hydrophobic interaction column were subjected to Edman degradation. However, no sequence was obtained. Since a BSA conttol blotted on Üie same PVDF membrane yielded a readable sequence, the N-terminus of the bioün protein of M. rubra apparenüy seems to be blocked.

Discussion

White deacetylated malonate decarboxylase of crude exttacts could be reactivated by chemical acetylation wiüi acetic anhydride, the acetylation procedure faUed if

applied to deacetylated cytoplasm. No catalytic activity was restored after combination

with Üie membrane fraction. It is conceivable, üierefore, that a complex consisting of cytoplasmic and membrane-bound components is required for a functional acetylation with acetic anhydride. Deacetyl cittate lyase: acetate Ugase of Klebsiella pneumoniae does not restore malonate decarboxylase activity. This is not unexpected since the Ugases of 115

Rhodocyclus gelatinosus (a phototrophic anaerobe), Streptococcus lactis ssp. diacetilactis and Clostridium sphenoides are known to be relatively specific for their homologous enzyme subsüates. However, Üie enzyme from Enterobacter aerogenes catalyzes the acetylation of cittate lyase from Streptococcus faecalis and Escherichia coli and the reactivation of citramalate lyase from Clostridium tetanomorphum (Anttanikian and Giffhorn, 1987). Hence, in a reversal of the above experiment the reactivation of deacetyl citrate lyase of K. pneumoniae with exttacts of M. rubra might be possible. Attempts to follow purification of components of the malonate decarboxylase enzyme System by an activity assay were dependent on enzymic reacetylation with deacetyl malonate decarboxylase: acetate Ugase, which seemed to be separated from the biotin protein in the initial step of purification. In a recombination assay Üie pooled fractions of cytoplasm after a DEAE column, from which üie bioün protein had been removed, had a remarkable effect on malonate decarboxylase activity only in presence of ATP/Mg2+/acetate. Thus, this pool is assumed to contain the active Ugase. The results discussed below, however, indicate the involvement of an additional component in this pool, tentatively an acyl carrier protein:

(1) The flowthrough of a DEAE column, in which the biotin protein was detected, retained malonate decarboxylase activity, whereas the acüvity was completely and irreversibly abolished upon binding of the biotin protein to the column. Although Üiis Observation could be explained with Üie loss of the acetyl moiety of Üie bioün protein upon binding to the column, it seemed more probable that a discrete acyl carrier protein was separated and eluted either in the flowthrough or at a different salt concentration.

(2) Even the mild Separation procedure by preparative gelfiltration chromatography caused a total loss of malonate decarboxylase activity of the cytoplasm. It is not easüy explained why this treatment should lead to the break of the covalent acetyl-Üiioester bond. Rather, the Separation of a discrete protein must be taken under consideration. Since the combined and extensively ulttaconcentrated fractions of the gel filttation column restored activity only to a minimal extent upon incubation with DTE/ATP/Mg2+/acetate, a smaU acyl carrier protein might have escaped either collection after gelfiltration chromatography or ulttaconcenttation. Acyl carrier proteins of only 9-10 kD molecular mass are known for cittate lyase of

Klebsiella pneumoniae (Dimroth and Eggerer, 1975) and are involved in fatty acid biosynthesis, e.g. in E. coli (Magnuson et al., 1993). (3) The pooled fractions of cytoplasm after a DEAE column, from which the biotin protein was removed (pool D), were ulttaconcentrated with an exclusion Umit of

10 kD. After enzymic acetylation and recombination with Üie membrane this led to a 116

significant increase of C02-liberating activity. On the other hand, heat inactivated cytoplasm was without effect. Therefore, no evidence was found for a low molecular weight Compound involved in malonate decarboxylation.

A recombination assay consisting only of washed membranes and purified CoA ttansferase (i.e. without Üie 40 % ammonium sulfate pellet) was completely macüve in the decarboxylation of malonate (HUbi and Dimroth, 1994). The recombination assay described in üiis chapter (line 8 on Table 1), which contained pool D in addition to membranes and purified CoA transferase, showed a considerable activity after enzymatic acetylation. Since a biotin protein is indispensibly involved in malonate decarboxylation (HUbi et al, 1992), pool D or üie membranes must still contain catalyücaUy active biotin protein.

The addition of a 40 % ammonium sulfate pellet to an recombination assay containing membranes, purified CoA ttansferase and pool D increased malonate decarboxylation twofold (line 2 and 8, Table 1). Partially purified biotin protein, however, had no positive effect on C02-release (line 9). Thus, the biotin protein was either inactivated through the chromatography procedures or the amounts present in

Üie other fractions were not rate-limiting for malonate decarboxylation.

In the reconstitution assays described above, addition of purified CoA ttansferase resulted in a fourfold increase of decarboxylation activity. Therefore, neither AS nor pool D contained significant amounts of active CoA ttansferase, i.e. the enzyme did not seem to survive freezing in Uquid nitrogen in presence of high concenttations of ammonium sulfate or it did not survive the ulttaconcenttation treatment performed within three days at 4 °C. 117

References

Antranikian G and Giffhorn F (1987) Cittate metaboüsm in anaerobic bacteria. FEMS

Microbiol Rev 46:175-198

Bendrat K and Buckel W (1993) Cloning, sequencing and expression of the gene

encoding the carboxytransferase subunit of the biotin-dependent Na+ pump glutaconyl-CoA decarboxylase from Acidaminococcus fermentans in Escherichia coli. Eur J Biochem 211: 679-702

Buckel W (1986) Biotin-dependent decarboxylases as bacterial sodium pumps:

purification and reconstitution • of glutaconyl-CoA decarboxylase from Acidaminococcus fermentans. MeÜi Enzymol 125: 547-558

Dehning I and Schink B (1989) Malonomonas rubra gen. nov. sp. nov., a microaerotolerant anaerobic bacterium growing by decarboxylaüon of malonate. Arch Microbiol 151: 427-433

Dimroth P and Eggerer H (1975) Evaluation of the protein components of cittate lyase

from Klebsiella aerogenes. Eur J Biochem 53: 227-235 Dimroth P (1986) Preparation, characterization, and reconstitution of oxaloacetate

decarboxylase from Klebsiella pneumoniae, a sodium pump. Methods Enzymol 125: 530-540

Dimroüi P (1987) Sodium ion ttansport decarboxylases and other aspects of sodium ion cycling in bacteria. Microbiol Rev 51: 320-340 Francis RT and Becker RR (1984) Specific indication of hemoproteins in

Polyacrylamide gels using a double-staining process. Anal Biochem 136: 509-514

Harlowe E and Lane D (1988) In Antibodies - a laboratory manual: pp. 555 ff. Cold Spring Harbor Laboratory, New York HUbi H, Dehning I, Schink B and Dimroth P (1992) Malonate decarboxylase of

Malonomonas rubra, a novel type of biotin-containing acetyl enzyme. Eur J Biochem 207: 117-123

HUbi H, Hermann R and Dimroth P (1993) The malonate decarboxylase enzyme

System of Malonomonas rubra: evidence for the cytoplasmic location of the biotin-

containing component Arch Microbiol 160: 126-131

HUbi H and Dimroth P (1994) Purification and characterization of a cytoplasmic

enzyme component of Üie Na+-activated malonate decarboxylase System of Malonomonas rubra: acetyl-S-acyl carrier protein: malonate acyl carrier protein-SH transferase. Arch Microbiol (in press) Magnuson K, Jackowski S, Rock CO and Cronan JE jr (1993) Regulation of fatty acid biosynthesis in Escherichia coli. Microbiol Rev 57: 522-542 118

Schägger H and von Jagow G (1987) Tricine-sodium dodecyl suüate-polyamide gel

electtophoresis for the Separation of proteins in the ränge from 1 to 100 kDa. Anal Biochem 166: 386-379 119

Chapter V

General Discussion

V. 1. The malonate decarboxylase enzyme System of M. rubra

V. 1.1. Activation of malonate

The decarboxylation of malonate imposes a chemical problem, since the C-C-bond

is not activated for decarboxylation. Possible routes of activation are either a homolytic

cleavage of the C-C-bond involving a radical mechanism (Figure IB) or a

C-S-CoA

A) Heterolytic C-C- bond cleavage

B) Homolytic C-C- bond cleavage

Figure 1. Possible mechanisms for Üie decarboxylation of malonate. A) Heterolytic C- C-bond cleavage involving an electton withdrawing subsütuent B) Homolytic C-C- bond cleavage via a radical route. 120

heterolytic mechanism, which demands the stabiUzation of the resulting carbanion by an electton withdrawing subsütuent e.g. by forming a thioester derivative (Figure 1A). Radical chemical routes of cleavage reactions involve the formation of a protein radical by an activating enzyme. The enzyme radical then produces a substtate radical, e.g. an oxygen radical of malonate, by the absttaction of an electton (or hydrogen atom). Subsequenüy, the substtate radical undergoes a rearrangement, leading to the release of CO2 and the product radical is formed, which oxidizes the enzyme üius regenerating the enzyme radical. Immediate protonation finaUy teads to the product acetate.

The oxygen-sensitive protein radical of the pyruvate formate lyase (a homodimer of twice 85 kD) is generated by an iron-dependent activase (28 kD) which requires reduced, FMN-containing flavodoxin and S-adenosyl meüiionine (SAM) as cofactor

(Knappe & Sawers, 1990). Extremely sensitive towards oxygen are also the Converter enzymes of lactyl-CoA dehydratase (Kuchta & Abeles, 1985) and 2-hydroxyglutaryl-

CoA dehydratase (Schweiger et al., 1987), which use ATP (and GTP) as cofactors. Lactyl-CoA dehydratase consists of the activase (E I, 27 kD) and Üie FMN- and riboflavin-containing heterodimer E II (48, 41 kD). The 2-hydroxyglutaryl-CoA dehydratase is composed of an tton-sulfur-containing heterodimer (55, 42 kD) and an activase (which contains neither iron nor inorganic sulfur).

In üie initial course of the work presented here a radical mechanism of malonate decarboxylation has been favored. Since ATP and acetyl-CoA had no effect on freshly prepared ceU free exttacts of Malonomonas rubra and malonyl-CoA (Figure 1A) was decarboxylated at a 10-times slower rate than free malonate, the latter was considered to be the substtate (HUbi et al., 1992). ATP (and GTP), but neither ADP nor SAM activated malonate decarboxylase in crude exttacts and the enzyme was not oxygen sensitive. These results were not in support for a radical mechanism.

The malonate decarboxylase was reversibly inactivated by deacetylation procedures. Incubation of the enzyme with hydroxylamine, ß-mercaptoethanol or sodium thiocyanate led to a complete loss of malonate decarboxylase activity. The inactivated enzyme was reactivated either by enzymic acetylation with ATP/acetate or by chemical acetylation with acetic anhydride (Figure 2). Incubation of the inactivated decarboxylase with dithioerythritol (DTE) had a positive effect on the subsequent enzymic acetylation, which indicates that Üie acetylated residue could be a thiol moiety

(HUbi et al, 1992).

The inactivation and reactivation pattern described above is reminiscent of citrate lyase of Klebsiella pneumoniae, which is induced upon anaerobic growth on cittate.

This enzyme complex (Mr 550'000) is postttanslationaUy activated by the acetylation of a 10 kD acyl carrier protein (ACP) and catalyzes the Mg2+-dependent, reversible 121

cleavage of cittate to acetate and oxaloacetate (Dimroth, 1988). The 55 kD ttansferase subunit of cittate lyase catalyzes the exchange of the ACP-bound acetyl thioester

residue for a cittyl thioester residue. This leads to the formation of acetate and makes the cleavage of the not activated C-C-bond of cittate chemicaUy feasible (Figure 3,

upper part). The protein-bound cittyl thioester is subsequenüy cleaved by the 32 kD lyase subunit of the cittate lyase complex to yield the second product oxaloacetate and

simultaneously regenerates the acetyl-S-ACP.

O II Chemical HS-CHj-CHj-OH E-S-C-CH3 interconversion NH2OH

(Ac)20 E-SH (inactive)

Enzymic ATP/ interconversion Cr^COOH O E-S-C-CH, (active)

Figure 2. Inactivation of malonate decarboxylase by deacetylating agents and reactivation by enzymic and chemical acetylation.

The analogous reaction sequence is shown for malonate decarboxylase (Figure 3,

lower part). This enzyme System, similar to cittate lyase, is activated by acetylation of an active site sulfhydryl residue, and the overall reaction involves the cleavage of free malonate (not malonyl-CoA) to acetate and C02 (HUbi et al. 1992). Taken together, the chemical problem of malonate activation is apparenüy solved by Üie formaüon of a malonyl thioester bond (malonyl-S-ACP), which activates the heterologous C-C-bond cleavage of malonate. On decarboxylation of the protein-bound malonyl-thioester the acetyl-ACP is regenerated. The malonyl-S-acyl carrier protein is formed by the 122

subsequent action of deacetyl malonate decarboxylase: acetate ligase and acetyl-S-acyl carrier protein: malonate acyl carrier protein-SH ttansferase (see Figure 4, reaction (1) and (2)).

Cittate Oxaloacetate frir Acetyl-S-ACP X *""* * Acetate Citryl-S-ACP

"OOC-CHjVC;-C-CrL, -C-S-ACP "OOC

i?H ° 0=C-CH,-C-S-ACP

Malonate _ , Acetyl-S-ACP _„CO-

^ ^ Acetate Malonyl-S-ACP H+

Figure 3. Comparison of the activation mechanism of cittate by cittate lyase and of malonate by malonate decarboxylase.

V. 1.2. Purification of a protein thiol transferase and relationship of malonate decarboxylase to citrate lyase

In order to characterize the malonate decarboxylase enzyme System Üie individual protein components had to be purified. This endeavor, however, was complicated by the loss of malonate decarboxylase activity after Separation of cytoplasmic and membrane-bound components by ultracentrifugation OHübi et al, 1992) and/or by the facile deacetylation of the acetyl-ACP moiety (Chapter V). A further compUcation in Üie purification of the protein components of malonate decarboxylase is that aU enzymes involved in malonate decarboxylaüon act on protein subsüates (see Figure 4). 123

Cytoplasm Membrane

ATP + ACP-SH Acetate

Acetate \/ (1)

Malonate H+ ACP-S-acetyl w. ^ E-biotin-CO,x / ,A > AjiNa Acetate "~ ~*ACP-S-malonyK x E-biotin

(2) (3) (4)

Figure 4. Hypoüietical reaction mechanism of the malonate decarboxylase enzyme System. (1) deacetyl malonate decarboxylase: aceüite ügase, (2) acetyl-S-acyl carrier protein: malonate acyl carrier protein-SH ttansferase, (3) carboxy ttansferase, (4) carboxy-bioün decarboxylase.

As described above, malonate decarboxylation is accompUshed via a malonyl-S- acyl carrier protein. This Compound is formed by acetyl-S-acyl carrier protein: malonate acyl carrier protein-SH transferase, which ttansfers the acyl carrier protein

(ACP) from acetate to malonate. Cittate lyase catalyzes a related ACP ttansfer from the ACP-bound acetate to cittate leading to the formation of citryl-S-ACP

(equation 1). This transferase accepts also CoA as altemate substtate and thus catalyzes the formation of cittyl-CoA from acetyl-CoA and cittate (equation 2).

cittate + acetyl-S-ACP ^ acetate + citryl-S-ACP (1)

cittate + acetyl-S-CoA ?=* acetate + cittyl-S-CoA (2)

The structural background for the turnover of this altemate substtate is Üie relationship between CoA and the Üüol carrying cofactor of citrate lyase, which is

5"-phosphoribosyl-2'-dephospho-CoA (Figure 5). This prostheüc group is covalenüy bound to a serine residue of the ACP via a phosphodiester bond. Another enzyme that 124

contains this prostheüc group is the enzyme complex cittamalate lyase of Clostridium tetanomorphum which catalyzes very similar reactions (Buckel & Bobi, 1976). The analogous reactions for the formaüon of malonyl-thioester derivatives by Üie acetyl-S-acyl carrier protein: malonate acyl carrier protein-SH ttansferase are ouüined for the physiological ACP ttansfer (equation 3) and for the altemate CoA ttansfer (equation 4).

malonate + acetyl-S-ACP ^ acetate + malonyl-S-ACP (3)

malonate + acetyl-S-CoA ?=* acetate + malonyl-S-CoA (4)

Exttacts of M. rubra were tested for the CoA transfer activity with malonyl-CoA and acetate as subsüates. The ttansferase thus detected has been purified to near homogeneity and its participation in malonate decarboxylation has been shown. No covalent transferase-CoA intermediate could be detected (HUbi & Dimroüi, 1994).

Such a covalent intermediate is characteristic of the physiological CoA ttansferases (Jencks, 1973) but it is lacking in the catalyüc mechanism of the ttansferase subunit of

O O H.C OH II II 3 i Adenine CH2-0-P-0-P-0-CH2-C-CH-CO-NH-(CH2)2-CO-NH-(CH2)2-SH ö OH OH CH3 Acyl carrier protein

C = 0 CHj-O-P-O CH2-CH OH NH

Figure 5. Structure of 5"-phosphoribosyl-2'-dephospho-CoA, the prostheüc group of cittate lyase of Klebsiella pneumoniae (according to Oppenheimer et al, 1979). The glycosidic linkage was found to be a (1" -> 2') by 'H-NMR analysis. 125

cittate lyase (Dimroth et al, 1977). Cittate lyase and malonate decarboxylase act on

CoA derivatives as altemate Substrates and both enzymes may use a catalytic mechanism not involving covalent enzyme-CoA intermediates. Hence, these two enzyme Systems not only share the essential acetylation (Figure 4, reaction (1)) but also have related enzymes for the activation of cittate or malate, respectively (Figure 4, reaction (2)). Based on these functional analogies, one may hypothesize that Üie catalytic thiol residue of malonate decarboxylase is part of an enzyme-bound CoA analogue as in cittate lyase and cittamalate lyase (Figure 5).

Another enzyme reacting via enzyme-bound Üiioesters may be arylmalonate decarboxylase from Alcaligenes bronchisepticus, which is a soluble protein of 24 kD that does not contain a biotin residue (Miyomoto & Ohta, 1992). The enzyme decarboxylates various substituted arylmalonates with inversion of configuration

(Miyamoto et al, 1992), but not unsubstituted malonate and methylmalonate.

However, the catalytic thiol group involved in this decarboxylaüon has not been further characterized.

V. 1.3. Relationship of malonate decarboxylase to the Na+-transport decarboxylases

The characteristic biochemical features of the Na+-motive decarboxylase enzyme famüy are discussed extensively in Chapter I, paragraph 2.3. Briefly, these enzymes are membrane-bound, contain biotin, are specifically activated by sodium ions and generate an electrochemical sodium gradient upon decarboxylation of their respective dicarboxylate Substrates (Dimroth, 1987). Malonate decarboxylase in cell free extracts shows some typical features of the Na+-ttansport decarboxylases. These properties and the composition of the enzyme System led to the hypothetical reaction sequence shown in Figure 4 (reaction 3 and 4). The CO2 moiety of the activated substtate (malonyl-S- ACP) is, therefore, supposed to be transferred to biotin in a carboxyltransfer reaction, which regenerates the acetyl-S-ACP (reaction 3). Subsequenüy, Üie carboxy biotin is assumed to be decarboxylated by a membrane-bound decarboxylase, which presumably couples this step to Üie ttanslocation of Na+-ions (reaction 4).

Growth of Malonomonas rubra on malonate, fumarate or malate required at least

150 mM NaCl, which could not be replaced by KCl (Dehning & Schink, 1989). This physiological dependence on sodium ions might already indicate bioenergetics, which rely on a Na+-circuit In ceU free extracts depleted of endogenous sodium, malonate decarboxylase is specifically sümulated approximately 14-times by Na+-ions (Km =

0.8 mM) and the enzyme is sümulated also by Li+-ions (Km = 3.3.mM), albeit with a Vmax of only 50 % compared to Na+ (HUbi & Dimroth, 1994). Such a Stimulation by 126

either Na+- or Li+-ions is a weU known feature of pnmary Na+-pumps (Dimroth,

1987, Dimroth & Thomer, 1989, Laubinger & Dimroth, 1988) and may therefore

indicate that malonate decarboxylase, too, is a primary Na+-pump The addiüonal features hnking malonate decarboxylase to the farmly of Na+-

ttansport decarboxylases are Üie locaüon of enzyme components m the membrane and

the participation of protein-bound bioün in the catalysis (HUbi et al, 1992) The

malonate decarboxylase enzyme System, therefore, seems to funcüon as a pnmary

Na+-ttansport decarboxylase and the altemaüve mechamsm for Üie storage of

decarboxylaüon energy via a soluble decarboxylase and a secondary antiporter does not apply (see Chapter I, paragraph 2 2)

Upon growth on malonate M rubra expresses a Single cytoplasmicaüy located

bioün protein of 120 kD, which is part of the malonate decarboxylase System (jTtibi et

al, 1993) The bioün protein of malonate decarboxylase differs from the biotin

proteins of Na+-translocaüng decarboxylases with respect to size and location The

latter farmly of enzymes contains membrane associated bioün proteins of 13, 24 and

63 kD (Chapter I, paragraph 2 3) Wlule the smaUer of these proteins compnse

separate biotin carboxyl camer proteins, the 63 kD bioün protein of oxaloacetate

decarboxylase is composed of the carboxyltransferase domain of 51 kD and a bioün-

binding domain of 10 kD (Dimroth & Thomer, 1983, Schwarz et al, 1988) According

to its size, the 120 kD bioün protein of M rubra is assumed to harbor addiüonal

domains, e g üie acyl camer domain involved m malonate acüvaüon (see above) and/or the carboxylüansferase funcüon

The stenc course of malonate decarboxylation was determmed by Micklefield &

Floss (personal communicaüon) with R- and S- [l-,3Ci, 2-3H] malonate as subsüates

and subsequent 3H-NMR-analysis The resulting acetate samples were converted to acetyl-CoA and then to malate with malate synthase The couplmg of the prochiral 3H at C3 of malate with Üie 13C revealed retenüon of configuraüon This stereochemical course is also foUowed by oxaloacetate decarboxylase of K pneumoniae (Dimroth, 1981), methylmalonyl-CoA decarboxylase of V parvula (Hoffmann & Dimroth, 1987) and all other biotin-containing enzymes invesügated so far (Knowles, 1989)

Malonomonas rubra growing on malonate probably conserves energy only by

means of an electrochemical sodium gradient It is, therefore, kkely that aU endergonic reactions in uns orgamsm, especiaUy ATP synthesis, are dnven by ApNa+ M. rubra has been shown to contain an ATPase crossreacüng with polyclonal antibodies raised agamst üie ß subunit of üie Escherichta coli FjF0-ATPase (U Genke, personal communicaüon) and may üierefore be anoüier example for an orgamsm employmg a Na+-ttanslocaüng F^ ATPase ATP synthesis in M rubra and Propiomgenium modestum (Hilpert et al, 1984, Laubinger & Dimroth, 1988) could thus be very 127

simUar. Furthermore, the uptake of malonate and rotation of the polar flagella may also be powered by AftNa+.

In summary, malonate decarboxylase of M. rubra combines the substtate

activation sttategy of cittate lyase and cittamalate lyase involving a thiol cofactor with

a biotin-dependent membrane-bound decarboxylation reaction, which is related to the

Na+-pumping decarboxylases. The sttategies for substtate activation and energy

conservation from otherwise unrelated bacteria have Üius been combined in M. rubra.

V. 2. Anaplerotic needs for fermentative growth on malonate

V. 2.1. Synthesis of C4-compounds

Most anaerobic bacteria with a citric acid cycle use üie socaUed pyruvate synthase pathway for the synthesis of C4-compounds from aceüite (Thauer, 1988). Here, acetyl- CoA is reductively carboxylated to pyruvate, from which oxaloacetate is formed by the subsequent action of phosphoenolpyruvate synthetase (AMP forming) and phosphoenolpyruvate carboxylase.

Malonomonas rubra, expressing the citric acid cycle enzymes, however, has not been reported to show phosphoenolpyruvate synthetase acüvity. Rather, in ceU free

exttacts the key enzymes of the glyoxylate bypass (Figure 6), i.e. isocitrate lyase and malate synthase, have been demonsttated with sufficient activities (Dehning & Schink, 1989). Thus, acetate is assimüated according to the overall reaction given in equation (5):

2 Acetyl-CoA + NAD+ + 2 H20 -» succinate + 2 HS-CoA + NADH + 2H+ (5)

Oxaloacetate, Üie initial substtate of gluconeogenesis, is formed oxidatively from succinate. Since üie anaplerotic oxidation of succinate wiüi NAD+, which yields

fumarate and NADH, is an endergonic reaction and takes place at the cytoplasmic membrane, it is tempting to speculate that the AüNa+ generated by the malonate decarboxylase could drive this reaction.

In many organisms acetate is activated to acetyl-CoA by the action of acetyl-CoA

synthetase (AMP-forming), which consumes two mol "energy-rich" phosphoric

anydride per mol acetyl-CoA formed. This enzyme activity has hardly been detected in M. rubra. Here, acetyl-CoA is synthesized by succinyl-CoA: acetate CoA ttansferase,

which consumes only one "energy-rich" bond. 128

V. 2.2. Fatty acid synthesis During growth of M. rubra on malonate fatty acids could be synthesized from acetate via acetyl-CoA and malonyl-CoA. The malonyl moiety is then transferred to an acyl carrier protein (ACP) subunit of the fatty acid synthetase complex. In M. rubra growing on malonate, the ATP-consuming carboxylation of acetyl-CoA to malonyl- CoA would be dispensible if malonyl-CoA could be formed directiy from

Gluconeogenesis Acetate + HS-CoA A

A AMP + PP.

A Aspartate > Pyruvate Phosphoenol A ^" pyruvate

NADH +H

NAD"1

Malate Pyruvate <- ' 4

y y Alanine 2-oxoglutarate

V Glutamate

Porphyrins, corrinoids

Acetate

Figure 6. The glyoxylate (Krebs-Kornberg) cycle (outdrawn Unes) and anaplerotic, respectively biosynthetic sequences (dashed Unes) of M. rubra. Owing to cleamess of the scheme, CO2 as product of the decarboxylation reactions of isocittate dehydrogenase (NADP+-dependent), 2-oxoglutarate dehydrogenase (ferredoxin- dependent), malic enzyme and phosphoenolpyruvate carboxykinase is omitted. Furthermore, CO2 as substtate of the pyruvate synthase (= pyruvate: ferredoxin oxidoreductase) reaction is not depicted. 129

malonate. In fact, there is no indication for the presence of acetyl-CoA carboxylase because only a Single biotin protein of 120 kD, which presumably parücipates in malonate decarboxylation, was found in crude exttacts. A biotin protein of smaU size that is typical for a bacterial acetyl-CoA carboxylase (17 kD in E. coli; Magnuson et al, 1993) was not found.

M. rubra is a physiologically specialized bacterium: besides malonate only fumarate and malate are fermented, yielding succinate and CO2 (Dehning & Schink, 1989). On these altemate Substrates, malonyl-CoA for fatty acid synthesis would have to be formed in the classical way wiüi acetyl-CoA carboxylase and one might speculate

Üiat a small bioün carboxyl carrier protein is induced under these conditions such as the 17 kD biotin carboxyl carrier protein of E. coli.

V. 2.3. Redox reactions

Anaerobic growth on highly oxidized subsüates is sometimes limited by the Provision of reducing equivalents for biosynthesis. During growth of M. rubra from the decarboxylation of malonate to acetate, reducing equivalents must probably be provided by the oxidation of acetate to CO2. This seems to be the case because Üie enzymes of the citric acid cycle have been found, providing reduced pyridine and flavine nucleotides for biosynthetic purposes (Dehning & Schink, 1989).

The benefit of a periplasmic cytochrome c (E°' = + 50 mV) and a membrane- bound cytochrome b present in M. rubra is not yet understood. Participation of the cychrome b in fumarate respiration (E°' = + 33 mV) has been suggested (Dehning &

Schink, 1989) and since the redox potential of the cytochrome c is unusuaUy low, it might participate also in fumarate respiration. A funcüon of the cytochromes during growth on malonate is completely obscure. M. rubra is aerotolerant (up to 5 % O2) and both catalase and cytochrome oxidase acüvity have been found. However, growth of M. rubra under these conditions was retarded, supplemented acetate was not utUized and the stoichiometty of malonate degradation was the same as in Üie absence of oxygen. Thus, it was concluded that O2 was not utUized as electton acceptor in an energy yielding respiratory chain (Dehning & Schink, 1989). Recenüy, terminal oxidases with a high affinity for O2 have been identified in several bacteria. The complex of Bradyrhizobium japonicum, termed FixNOQP, contains cytochromes of the b- and c-type (Preisig et al, 1993). This enzyme is induced under microaerobiosis (0.5 % O2). It is possible Üiat appropriate conditions for the inducüon of a putative analogue of üiis enzyme in M. rubra were not given. 130

V. 3. Outlook

As ouüined in Chapter V, the biochemical characterization of the malonate decarboxylase enzyme System has been greatiy complicated by the labüity of the catalytically essential acetyl moiety and the composition of separate components. Some

Problems, e.g. the identification of the component proteins of malonate decarboxylase of M. rubra, might be overcome by switching to an experimental approach on DNA level. A 29 amino acid N-terminal sequence of üie acetyl-S-ACP: malonate ACP-SH ttansferase is already known (see Appendix to Chapter IV) and cyanogen bromide or ttyptic fragments of the N-terminaUy blocked bioün protein (Chapter V) should yield further peptide sequences. One may also take advantage of the conserved biotin binding consensus sequence (AMKM). Completely conserved sttetches within the hydrophobic ß subunits of Na+-ttansporting decarboxylases can be exploited to amplify genes of related proteins by the Polymerase chain reaction technique (Huder and Dimroth, 1993; P. Burda, personal communicaüon). If these homologies would also extend to a membrane-bound component of Üie malonate decarboxylase, the genetic characterization of the malonate decarboxylase enzyme system of M. rubra would be greatiy facilitated. 131

V. 4. References

Buckel, W. & Bobi, A. (1976) The enzyme complex cittamalate lyase from Clostridium tetanomorphum. Eur. J. Biochem 64: 255-262.

Dehning, I. & Schink, B. (1989) Malonomonas rubra gen. nov. sp. nov., a microaerotolerant anaerobic bacterium growing by decarboxylation of malonate. Arcfc. Microbiol. 151: 427-433.

Dimroth, P. (1981) Characterization of a membrane-bound biotin-containing enzyme:

oxaloacetate decarboxylase from Klebsiella aerogenes. Eur. J. Biochem. 115: 353- 358.

Dimroüi, P. (1987) Sodium ion transport decarboxylases and other aspects of sodium ion cyding in bacteria Microbiol. Rev. 51: 320-340. Dimroth, P., Loyal, R. & Eggerer, H. (1977) Characterization of the isolated

ttansferase subunit of cittate lyase as a CoA ttansferase. Evidence against a covalent enzyme-substrate intermediate. Eur. J. Biochem 80:479-488. Dimroth, P. (1988) The role of vitamins and their carrier proteins in cittate

fermentation, in 77ie Roots of Modern Biochemistry, pp. 191-204, Kleinkauf, von Döhren, Jaenicke (eds.). W. de Gruyter & Co., Berlin, New York. Dimroth, P. & Thomer, A. (1983) Subunit composition of oxaloacetate decarboxylase and characterization of the a chain as carboxylüansferase. Eur. J. Biochem. 137: 107-112.

Dimroth, P. & Thomer, A. (1989) A primary respiratory Na+ pump of an anaerobic bacterium: the Na+-dependent NADH: quinone oxidoreductase of Klebsiella pneumoniae. Arch. Microbiol. 151: 439-444. Hilbi, H., Dehning, 1., Schink, B. & Dimroth, P. (1992) Malonate decarboxylase of

Malonomonas rubra, a novel type of biotin-containing acetyl enzyme. Eur. J. Biochem 207: 117-123.

HUbi, H., Hermann, R. & Dimroth, P. (1993) The malonate decarboxylase enzyme system of Malonomonas rubra: evidence for the cytoplasmic location of the biotin- containing component. Arch. Microbiol. 160:126-131.

HUbi, H. & Dimroüi, P. (1994) Purification and characterization of a cytoplasmic

enzyme component of the Na+-activated malonate decarboxylase System of Malonomonas rubra: acetyl-S-acyl carrier protein: malonate acyl carrier protein-SH ttansferase. Arch. Microbiol. (in press).

Hilpert, W., Schink, B. & Dimroth, P. (1984) Life by a new decarboxylation- dependent energy conservation mechanism with Na+ as coupling ion. EMBO J. 3: 1665-1670. 132

Hoffmann, A. & Dimroth, P. (1987) Stereochemistry of the methylmalonyl-CoA decarboxylation reaction. FEBSLett. 220: 121-125.

Huder, J. B. & Dimroth, P. (1993) Sequence of the sodium ion pump methylmalonyl- CoA decarboxylase from Veillonella parvula. J. Biol. Chem. 268: 24564-24571. Jencks, W. P. (1973) The CoA transferases, in The Enzymes (Boyer, P. D., ed.) third

edition, vol. 9B, pp. 483-496. Academic Press, New York.

Knowles, J. R. (1989) The mechanism of biotin dependent enzymes. Ann. Rev. Biochem. 58:195-221.

Laubinger, W. & Dimroth, P. (1988) Characterization of the ATP synüiase of

Propiomgenium modestum as a primary sodium pump, Biochemistry 27: 7531- 7537. Magnuson, K., Jackowski, S., Rock, C. O. & Cronan, J. E., jr. (1993) Regulation of fatty acid biosynüiesis in Escherichia coli. Microbiol. Rev. 57: 522-542.

Miyamoto, K. & Ohta, H. (1992) Purification and properties of a novel arylmalonate decarboxylase from Alcaligenes bronchisepticus KU 1201. Eur. J. Biochem. 210: 475-481.

Miyamoto, K., Tsuchiya, S. & Ohta, H. (1992) Stereochemistry of enzyme-catalyzed decarboxylaüon of a-methyl-a-phenylmalonic acid. J. Am Chem Soc. 114: 6256- 6257.

Oppenheimer, N. J., Singh, M., Sweeley, C. C, Sung, S.-J. & Srere, P. A. (1979) The

configuration and location of the ribosidic linkage in the prostheüc group of cittate lyase {Klebsiella aerogenes). J. Biol. Chem 254:1000-1002.

Preisig, O., Anüiamatten, D. & Hennecke, H. (1993) Genes for a microaerobically

induced oxidase complex in Bradyrhizobium japonicum are essential for a nittogen- fixing endosymbiosis. Proc. Natl. Acad. Sei. USA 90: 3309-3313. Schwarz, E., Oesterhelt, D., Reinke, H., Beyreuther, K. & Dimroth, P. (1989) The

sodium ion ttanslocating oxaloacetate decarboxylase of Klebsiella pneumoniae - Sequence of the biotin-containing a-subunit and relationship to other biotin-

containing enzymes. J. Biol. Chem 263:9640-9645.

Thauer, R. K. (1988) Citric-acid cyle, 50 years on - Modifications and an alternative pathway in anaerobic bacteria. Eur. J. Biochem 176:497-508. 133

CURRICULUM VlTAE

Hubert Franz Pius Hilbi

Born May 30,1965 in Zug, Switzerland

1972-1978 Primary education in Oberwü/Zug (ZG)

1978-1984 Gymnasium, Kantonsschule Zug

Final examination: Matura type B

1985-1990 Studies in Biochemistry at the Swiss Federal Institute of Technology (ETH) in Zürich Diploma in Natural Sciences

1990 Diploma Thesis at the Institute of Microbiology ETH, Zürich

1990-1994 Assistant researcher at the Institute of Microbiology ETH, Zürich

Ph. D. Thesis (November 1990 - June 1994) 134

LIST OF PUBLICATIONS

HUbi, H., Dehning, I., Schink, B. & Dimroth, P. (1992) Malonate decarboxylase of

Malonomonas rubra, a novel type of biotin-containing acetyl enzyme. Eur. J.

Biochem 207:117-123.

HUbi, H. & Dimroth, P. (1993) Malonate decarboxylase of Malonomonas rubra, a

novel decarboxylating enzyme system. Biol Chem Hoppe-Seyler 374: 810-811.

HUbi, H. & Dimroth, P. (1994) Purification and characterization of a cytoplasmic

enzyme component of the Na+-activated malonate decarboxylase system of Malonomonas rubra: acetyl-S-acyl carrier protein: malonate acyl carrier protein-SH transferase. Arch. Microbiol. (in press).

Micklefield, J., HUbi, H., Dimroth, P. & Floss, H. G. (1994) The stereochemical course of the malonate decarboxylaüon reaction. J. Am Chem Soc. (in preparation).