FEMS Microbiology Letters 121 (1994) 107-114 © 1994 Federation of European Microbiological Societies 0378-1097/94/$07.00 Published by Elsevier

107

FEMSLE 06103

Glucose fermentation to acetate and alanine in resting cell suspensions of Pyrococcus furiosus: Proposal of a novel glycolytic pathway based on 13C labelling data and enzyme activities

T h o m a s Sch~ifer a, K a r i n a B. X a v i e r b, H e l e n a San to s b, P e t e r S c h 6 n h e i t a,,

a Institut fiir Pflanzenphysiologie und Mikrobiologie, Freie Universitiit Berlin, K6nigin-Luise-Strasse 12-16a; D-14195 Berlin, FRG, and b Instituto de Tecnologia Qulmica e Biol6gical, Apartado 127, 2780 Oeiras, Portugal

(Received 9 May 1994; revision received and accepted 6 June 1994)

Abstract: Suspensions of maltose-grown cells of the hyperthermophilic archaeon Pyrococcus furiosus, when incubated at 90°C with 35 mM [1-13C]glucose or [3-13C]glucose, consumed glucose at a rate of about 10 nmol min -1 (mg protein) -1. Acetate (10 raM), alanine (3 raM), CO/ and H 2 were the fermentation products. The i3C-labelling pattern in alanine and acetate were analyzed. With [1-13C]glucose the methyl group of both alanine and acetate was labelled; with [3-13C]glucose only the carboxyl group of alanine was labelled whereas acetate was unlabelled. Extracts of maltose-grown cells contained glucose isomerase (12.8 U mg 1, 100°C), ketohexokinase (0.23 U mg - l , 100°C), and fructose 1-phosphate aldolase (0.06 U mg-1, 100oc). Enzymes catalyzing the formation of fructose 1,6-bisphosphate from fructose 1-phosphate or fructose 6-phosphate could not be detected. As published previously by our group and other authors P. furiosus also contains enzymes of glyceraldehyde conversion to 2-phosphoglycerate according to a non-phosphorylated Entner-Doudoroff pathway, of dihydroxyacetone phosphate conversion to 2-phosphoglycerate according to the Embden-Meyerhof pathway, and of 2-phosphoglycerate conversion - via pyruvate - to acetate and alanine. Based on the enzyme activities in P. furiosus, the following pathway for glucose degradation to alanine and acetate in cell suspensions is proposed which can explain the [13C]glucose labelling data: glucose ~ fructose ~ fructose 1-phosphate ~ dihydroxyacetone phosphate + glyceraldehyde and further conversion of both trioses to alanine and acetate via pyruvate.

Key words: Pyrococcus furiosus; Hyperthermophilic archaeon; Novel glucose fermentation pathway; Ketohexokinase; Fructose 1-phosphate aldolase; Embden-Meyerhof pathway; Non-pbosphorylated Entner-Doudoroff pathway

Introduction

The hyperthermophil ic archaeon Pyrococcus furiosus can grow on maltose, cellobiose and

* Corresponding author. Tel.: (030) 838 3116; Fax: (030) 838 3118

pyruvate as energy and carbon sources [1-4]. Growth on these substrates has been studied in detail with respect to molar growth yields, fer- mentat ion balances, and enzyme activities in- volved in both catabolism and gluconeogenesis [2-7]. Maltose is fermented by P. furiosus to acetate, CO 2 and H 2 and, depending on the H 2 partial pressure, various amounts of alanine [3,4].

SSDI 0378-1097(94)00256-Q

108

Since in cell extracts all enzymes, except glu- conate dehydratase, of a modified non-phos- phorylated Entner-Doudoroff pathway were de- tected, this pathway has been proposed to be operative in sugar degradation to pyruvate [3]. Furthermore, P. furiosus contains all reversible enzymes of the Embden-Meyerhof pathway, which have been implicated in gluconeogenesis, e.g. during growth on pyruvate [6]. In this com- munication we studied the sugar fermentation pathway in P. furiosus by 13C-NMR technique [8]. Specifically labelled [13C]glucose was used to fol- low the metabolic fate of 13C-label in the fermen- tation products. The experiments were performed with cell suspensions since glucose does not serve as growth substrate of P. furiosus [3]. We found that the ~3C-labelling patterns in the fermenta- tion products acetate and alanine were not in accordance with the operation of a Entner- Doudoroff pathway. Based on enzyme activities detected in cell extracts we propose a novel gly- colytic pathway of glucose fermentation in cell suspensions of P. furiosus, involving reactions of both the Embden-Meyerhof pathway and the non-phosphorylated Entner-Doudoroff pathway, which explains the 13C-labelling patterns.

Materials and Methods

Source of material Chemicals used were reagent grade and ob-

tained from E. Merck (Darmstadt, FRG). En- zymes and coenzymes were from Boehringer (Mannheim, FRG). Pyrococcus furiosus DSM 3638 was from the Deutsche Sammlung von Mikroorganismen (Braunscheig, FRG). [1-13C] glucose and [3-13C] glucose, both with 99% iso- topic enrichment were purchased from Omicron Biochemicals, Inc. (USA).

Cell suspension experiments Pyrococcus furiosus was grown at 90°C with 5

mM maltose as carbon and energy source in an open fermenter system as described earlier [2,3]. Cell suspensions were performed under strictly anaerobic conditions [2,3]. Late log-phase cells were harvested, washed twice with 20 mM Tris .

HC1 (pH 8.0) containing 370 mM NaCI, 5 mM MgCI 2 and 8 mM KC1, and finally resuspended in the same buffer at a protein concentration of about 5 mg ml-1. Fermentation of glucose by cell suspensions was performed at 95°C in closed 50- ml serum bottles (gas phase 100% N z) filled with 4 ml cell suspension (see above) supplemented with 35 mM each of either [1-13C]glucose, [3- ~3C]glucose or unlabelled glucose. Consumption of glucose and formation of acetate, alanine, H e and CO 2 was followed for 2 h in assays with unlabelled glucose. In parallel assays, containing [1-13C]glucose and [3-13C]glucose, the cells were centrifuged after 2 h and the labelling patterns of acetate and alanine in the supernatants were analyzed by ~3C-NMR and ~H-NMR. Some sus- pensions experiments were performed in open reaction vessels at a higher cell concentrations (up to 50 mg m1-1) and the ratio of a lan ine /ace- tate formation [4] was varied by gassing the sus- pensions with argon.

NMR analysis of endproducts Spectra were acquired in a Bruker AMX500

spectrometer operating at 125.77 MHz for car- bon-13 and 500.13 MHz for protons. 13C-NMR spectra were acquired by using a 10-mm quadru- ple nuclei probe head with the following parame- ters: spectral width, 38 kHz; data size, 128K; repetition delay, 62 s; pulse width, 12 /xs (corre- sponding to 70 ° flip angle). Chemical shifts are referenced to external methanol at 49.3 ppm. Proton broadband decoupling was applied for 1.6 s during acquisition time by using the WALTZ sequence. 1H_NMR spectra were acquired with a 5-mm probe head with water presaturation and the following parameters: spectral width, 5 kHz; data size, 64K; repetition delay, 17 s; flip angle, 60°C. These conditions ensured full relaxation of the resonances arising from fermentation prod- ucts in both types of spectra.

Preparation of cell extracts Cell extracts were prepared under strictly

anaerobic conditions as described [3]. Late log cells grown on maltose were disrupted by sonica- tion and after centrifugation (18000 × g for 20 min) the extract was depleted of salts and low

molecular compounds by gel filtration Sephadex G-25.

o n

Determination of enzyme activities Enzyme assays were performed anaerobically

in glass cuvettes filled with 1 ml assay mixture (protein concentration 0.5-3 mg protein) as de- scribed earlier [3].

Fructose 1-phosphate aldolase activity was measured by following the fructose 1-phosphate (F-1-P)-dependent formation of both products, dihydroxyacetone phosphate (DHAP) and glycer- aldehyde. DHAP formation was measured at 50°C by following NADH oxidation via glycerol 3-phos- phate dehydrogenase. Assay mixture: 5 mM F-l-P, 0.3 mM NADH, 8.5 U glycerol 3-phosphate dehy- drogenase. Glyceraldehyde formation was mea- sured at 80°C by following benzylviologen reduc- tion via endogenous glyceraldehyde:benzylviolo- gen oxidoreductase present in cell extracts. Assay mixture: 5 mM F-l-P, 5 mM benzylviologen.

Fructose 6-phosphate aldolase activity was measured by following fructose 6-phosphate (F-6- P)-dependent dihydroxyacetone formation via glycerol dehydrogenase.

Fructose kinase activity was determined at 55°C by coupling fructose dependent ADP formation from ATP to the oxidation of NADH via pyru- vate kinase (PK) and lactate dehydrogenase (LDH). Assay mixture: 5 mM fructose, 2 mM ATP, 5 mM phospho-enolpyruvate, 0.3 mM NADH, 10 mM MgC12, 27 U LDH, 4 U PK. This test does not discriminate between the formation of F-1-P and F-6-P.

Hexokinase, fructose 1-phosphate kinase or fructose 6-phosphate kinase were tested using the same assays except that fructose was replaced by 10 mM glucose, 5 mM F-1-P or 5 mM F-6-P, respectively.

Fructose 1-phosphate kinase activity and fruc- tose 6-phosphate kinase activity either dependent on ATP or pyrophosphate were also tested as described previously [9,10].

Analytical procedures Acetate and alanine were determined enzy-

matically [11,12]. Glucose was determined accord- ing to Kunst et al. [13]. Protein was determined

109

with the Biuret method using bovine serum albu- min as standard.

Results

[13C]glucose fermentation to acetate and alanine in cell suspensions of Pyrococcus furiosus

Cell suspensions were incubated at 95°C in closed serum bottles containing 35 mM glucose, either 1-13C-labelled, 3-t3C-labelled or unla- belled. The cells consumed glucose at a rate of about 10 nmol min-1 (mg protein)-1 over a pe- riod of 2 h and acetate (10.7 mM), alanine (3.5 mM), H 2 and CO 2 (not quantified) were formed as fermentation products. The 13C-labelling pat- tern in alanine and acetate formed after fermen- tation of 13C-labelled glucose was determined from ~3C- and ~H-NMR spectra of the super- natants obtained by centrifugation of the assay mixtures. The ~3C-spectra in Fig. 1 show that after fermentation of [1-~3C]glucose the label ended on the methyl groups of acetate and ala- nine (resonances at 16.5 and 2.7 ppm, respec- tively, in spectrum A), whereas the label was found only on the carboxylate group of alanine (resonance a 176.2 ppm in spectrum B) when [3-~3C]glucose was supplied. In this latter spec- trum the two equal intensity resonances at 2.7 and 181.9 ppm are due to natural abundance ~3C in acetate. These conclusions are further sup- ported by ~H-spectra of the same supernatants, shown in Fig. 2. The resonance patterns due to methyl groups of acetate and alanine are high- lighted. In each pattern, the central resonance (singlet for acetate at 1.9 ppm and doublet for alanine a 1.5 ppm) is due to the fully unlabelled compounds; in spectrum A of Fig. 2, the two lateral resonances with equal intensity and sepa- rated by approx. 0.43 ppm (j13C-1H = 130 Hz) are due to isotopomers of alanine or acetate in which only the methyl group is labelled; finally, the central multiplet in the alanine pattern (spec- trum B in Fig. 2), split by a small coupling con- stant of 4 Hz is due to the isotopomer of alanine with label only on the carboxylate carbon. This multiplet is definitely absent in spectrum A (Fig. 2) which shows that no alanine with label on the

110

carboxylate is produced from [1-13C]glucose un- der the experimental conditions used.

The labelling patterns in various experiments were the same independent of the ace t a t e / alanine ratio, which varied between about 5 :1 and 1 : 1 at different gassing conditions.

Enzyme studies The ~3C-labelling found in acetate and alanine

suggests that glucose was fermented by cell sus- pensions via an Embden-Meyerhof pathway type of glycolysis rather than via the Entner-Doudoroff pathway. The 'classical' Embden-Meyerhof path- way appears not to be operative: we could not

glucose + Tris buffer

a lanine

= J_J . . . . . . . . . . . 11 . . . .

A L - ] . . . . .

I I I I I I

175 150 125 100 75 50 25

ppm

Fig. 1. (A, B) 13C-NMR spectra of liquid supernatants con- taining the end-products of metabolism of [1-I3C]glucose (spectrum A) or [3-13C]glucose (spectrum B) by cell suspen- sions of Pyrococcus furiosus. Glucose was fermented under the conditions described in Materials and Methods, the cell suspensions were centrifuged and the supernatants were lyophilized. The resulting residues were dissolved in 2H20 (final pH, 6.8). Spectra were run at 27°C. The spectral region between 60 and 100 ppm is dominated by resonances arising from non-metabolized labelled glucose. The glucose peaks in spectrum A and spectrum B show D-a-glucose and o-fl-glu-

cose in Tris buffer•

acetate

I [ I

B

13CH3COO-

CH3CO0-

13CH3C00-

alanine

[I 'P'I' I]

A

2 2 2 0 1.8 1.6 1.4

pprn

Fig. 2. (A, B) 1H-NMR spectra of the liquid supernatants whose 13C-spectra are shown in Fig. 1, containing end-prod- ucts from fermentation of [1-13C]glucose (A) or [3-13C]glucose (B). The resonance patterns due to the methyl groups of alanine and acetate are highlighted• The strong intensity resonance due to the methyl group of acetate has been

truncated in this representation of spectrum B.

detect hexokinase activity and enzyme activities catalyzing the formation of fructose 1,6-bi- sphosphate (FBP) from hexoses. Neither a fruc- tose 1-phosphate (F-l-P) kinase nor a fructose 6-phosphate (F-6-P) kinase tested with either ATP or pyrophosphate (PPi) as phosporyl donor could be detected in cell extracts. Cell extracts of Ther- moproteus tenax served as a positive control for PPi-dependent F-6-P kinase (0.05 U mg -~ at 55°C) [8]. These findings suggest that glucose degradation to pyruvate does not proceed via FBP.

111

We found ketohexokinase activity and fructose 1-phosphate aldolase activity, which in combina- tion with previously published enzymes can ex- plain the ~3C-labelling patterns by a glycolytic pathway involving reactions of both the Embden- Meyerhof pathway and the non-phosphorylated Entner-Doudoroff pathway.

Ketohexokinase (fructose kinase) Cell extracts catalyzed the fructose-dependent

ADP formation from ATP at a rate of 9.2 mU mg -1 (at 55°C) ( = 225 mU mg -1 at 100°C, as- suming a Ql0 = 2). The apparent K m values were 0.75 mM for fructose and 0.09 mM for ATP, respectively. F-1-P rather than F-6-P is probably the reaction product since cell extracts contain F-1-P aldolase activity (rather then F-6-P aldolase activity) to be involved in subsequent aldol cleav- age (see below). Thus, the data suggest the pres- ence of a ketohexokinase activity (ATP:fructose 1-phosphotransferase, EC 2.7.1.3) in P. furiosus.

Fructose 1-phosphate aldolase activity Cell extracts contained F-1-P aldolase activity

catalyzing cleavage of F-1-P to dihydroxyacetone phosphate (DHAP) and glyceraldehyde. The rate of DHAP formation was 1.9 mU mg -1 at 50°C

e..

<

8.0

6.0

4.0

2.0

.... ! E / - - oo

. ; . . . . 2 8 I/T[K "1 x 10 "~] /

0.0 I

3O I I t I I I

40 50 60 70 80 90

Temperature [°C]

Fig. 3. Fructose 1-phosphate aldolase activity in cell extracts of Pyrococcus furiosus. Rate dependence of glyceraldehyde formation on the temperature. The insert shows logarithm of the rates versus the reciprocal temperatures (Arrhenius plot).

( = 60 mU mg-1 at 100°C). The apparent K m for F-1-P was 1.5 mM. The rate of glyceraldehyde formation was 9 mU mg -1 at 88°C ( = 20 mU mg -1 at 100°C). KCI and MgC12 (5-25 mM each) did not affect enzyme activity. The temperature dependence of enzyme activity is shown in Fig. 3. From the linear Arrhenius plot between 40°C and 80°C a Ql0 value of 2.0 and an activation energy of 59 kJ mol-1 were calculated.

Discussion

It has recently been shown that growing cul- tures of P. furiosus ferment maltose to acetate, alanine, CO 2 and H 2. Since most enzymes of the non-phosphorylated Entner-Doudoroff pathway were detected in cell extracts [3], including glu- cose:ferredoxin oxidoreductase [3,14] and 2-keto- 3-deoxy-gluconate (KDG) aldolase, it has been proposed that this pathway is involved in maltose degradation via glucose to pyruvate. Molar growth yields of about 40-50 g cell dry mass mol - t of maltose indicated an ATP yield of 4-5 mol mol - 1 maltose assuming YATP to be about 10 g cell dry mass mol -~. Since up to 4 mol acetate were formed it was concluded that almost all ATP formed in this fermentation was synthesized dur- ing acetate formation from acetyl-CoA catalyzed by acetyl-CoA synthetase [3,5]. Accordingly, glu- cose conversion to pyruvate was not coupled with energy conservation, which would support with the operation of the non-phosphorylated Entner- Doudoroff pathway.

The laC-labelling patterns obtained after glu- cose fermentation in resting cell suspensions are not in accordance with the operation of the non- phosphorylated Entner-Doudoroff pathway. In such a case fermentation of [1-13C]glucose should result in labelling of the carboxylate group of alanine, whereas fermentation of [3-13C]glucose should result in labelling of the methyl groups of both alanine and acetate. Both labelling patterns were not observed. Instead it was found that with [1-13C]glucose only the methyl groups of alanine and acetate were labelled, whereas with [3- lac]glucose only the carboxylate group of alanine was labelled and acetate was unlabelled. These

112

labelling pat tern indicate the operation of an Embden-Meyerhof pathway type of glycolysis. Since neither hexokinase nor ATP- or PPi-de- pendent kinases forming fructose 1,6-bisphos- phate could be detected the 'classical' Embden- Meyerhof pathway appears not to be operative in glucose fermentation.

From the enzyme activities found in cell ex- tracts of P. furiosus either described here or in previous publications [2-6,14], we propose the following pathway of glucose fermentat ion in cell suspensions, which can explain the ~3C-labelling patterns (Fig. 4): assuming that free glucose is metabolized within the cells, it can be converted to dihydroxyacetone phosphate and glyceralde- hyde via the activity of glucose isomerase [3], ketohexokinase (ATP : fructose 1-phosphotrans- ferase) and fructose 1-phosphate (F-l-P) al- dolase. So far, ketohexokinase activity has been demonstrated in prokaryotes only in halophilic archaea [15]. F-1-P aldolase activity found in cell extracts catalyzes F-1-P cleavage into D H A P and glyceraldehyde. P. furiosus also containes FBP aldolase activity, which has been implicated in gluconeogenesis [6]. F-1-P aldolase and FBP al- dolase had about the same specific activities in P. furiosus. It remains to be shown whether the two aldolytic activities are catalyzed by one or by two different enzymes. Eukaryotic aldolase B from liver is an example for the former possibility: it catalyzes both F-1-P and FBP cleavage at about the same specific activities [16].

F-1-P cleavage yields D H A P and glyceralde- hyde. D H A P can be converted to 2-phosphog- lycerate by the following enzymes of the Embden- Meyerhof pathway which have been found in P. furiosus [6]: t r iosephosphate isomerase, NADP ÷ reducing glyceraldehyde 3-phosphate dehydroge- nase, 3-phosphoglycerate kinase, phosphoglycer- ate mutase. These enzymes have been implicated in gluconeogenesis during growth on pyruvate [6] but can catalyze, as reversible enzymes, D H A P conversion to 2-phosphoglycerate in glucose catabolism by cell suspensions.

Glyceraldehyde, the second product of F-1-P cleavage, can be converted to 2-phosphoglycerate by glyceraldehyde: ferredoxin oxidoreductase and glycerate kinase (2-phosphoglycerate forming), i.e.

Dihydroxyacetone phosphate

Gl(~yc!aldeh yde 3-phosphate

e ]iS p' 1,3 Bisphosphoglycerate

ATP 3-Phosphoglycerate

2-Phosphoglycerate

Fructose

ADP Fructose- l-phosphate

Glyceraldehyde

Glycerate

ADP

2 Phosphoglycerate

Phosphoenolpyruvate

NADP. N.3 ® ~ A D P ATP

~ . ~ Pyruvale

@® coA --.1.-- Fdox --@ 2.* [ ~ ~ ' - ~ " Fd red ~ ' ~ [ ~

Acetyl CoA ®]U-- A°~+p,

CoA .,,M~J~,,..-..~ ATP

Fig. 4. Proposed pathway of glucose fermentation to acetate, alanine, CO 2 and H 2 in cell suspensions of Pyrococcus furio- sus. The numbers in cycles refer to the enzyme activities involved. The numbers in brackets indicate references of publications where the enzyme activities have been described. 1, glucose isomerase [3]; 2, fructose kinase; 3, fructose 1-phos- phate aldolase (this study); 4, triose-phosphate isomerase; 5, glyceraldehyde 3-phosphate dehydrogenase; 6, phosphoglycer- ate kinase; 7, phosphoglycerate mutase [6]; 8, glyceralde- hyde:ferredo×in (viologen dye) oxidoreductase [14,3]; 9, glyc- erate kinase; 10, enolase; 11, pyruvate kinase; 12, pyru- vate:ferredoxin oxidoreductase; 13, acetyl-CoA synthetase (ADP forming) [2,3,5]; 14, hydrogenase [2,17]; 15, glutamate

dehydrogenase; 16, alanine aminotransferase [4].

enzymes of the non-phosphorylated Entner-Dou- doroff pathway. Further conversion of 2-phos- phoglycerate to acetate and alanine proceeds via enolase, pyruvate kinase, pyruvate:ferredoxin oxi- doreductase, acetyl-CoA synthetase (ADP form-

113

ing) and glutamate dehydrogenase/alanine ami- notransferase (Fig. 4) [2-5].

It remains to be shown whether the sugar fermentation pathway proposed here for cell sus- pensions is also valid for maltose fermentation in growing cultures. The physiological conditions in growing cultures and in cell suspensions are most likely different, e.g. growth of P. furiosus on glucose has not been demonstrated so far; fur- thermore the rates of maltose fermentation in growing cuItures are about 100-fold higher than of glucose conversion in cell suspensions which might reflect different transport kinetics. Thus, if the utilization of either sugar fermentation path- way, the glycolytic pathway proposed here and non-phosphorylated Entner-Doudoroff pathway, is regulated under various physiological condi- tions they might be operative to a different extent in resting and growing cells. The presence of two different catabolic pathways of glucose fermenta- tion, a modified Embden-Meyerhof pathway and the non-phosphorylated Entner-Doudoroff path- way, has recently proposed in the hyperther- mophile Thermoproteus tenax by Siebers and Hensel [9].

From molar growth yield data recently pre- sented by Kengen and Stams ([7]; see also in [3]) the authors speculate that an additional energy coupling site might be present in sugar fermenta- tion besides the acetyl-CoA synthetase (ADP-for- ming) reaction. The glucose fermentation path- way proposed here provides an additional ATP formation site: the 3-phosphoglycerate kinase re- action. Accordingly, up to 3 mol of ATP mol- a of glucose equivalent might be formed, if the glu- cose fermentation pathway proposed here for cell suspension is also operative in growing cells, which has to be demonstrated.

Note added

During preparation of the revised version of this manuscript, Kengen and colleagues (Kengen, S.W.M., de Bok, F.A.M., van Loo, N.-D., Dijke- ma, C., Stams, A.J.M., de Vos, W.M. (1994) Evi- dence for the operation of a novel Embden- Meyerhof pathway that involves ADP dependent

kinases during sugar fermentation by Pyrococcus furiosus: J. Biol. Chem., in press) reported on [13C]glucose labelling experiments in cell suspen- sions of Pyrococcus furiosus leading to similar results as described here. Furthermore, they re- ported on the presence of a ADP-dependent hexokinase and a ADP-dependent phosphofruc- tokinase and proposed that glucose degradation to pyruvate in P. furiosus proceeds via a modified Embden-Meyerhof pathway involving these novel kinases and fructose 1,6-bisphosphate as an inter- mediate.

Acknowledgements

The work was supported by grants from the European Union ("Biotechnology of Extremo- philes"), from the Bundesministerium fOr For- schung und Technologie ("Biologische Wasser- stoffgewinnung") and the Fonds der Chemischen Industrie.

References

1 Fiala, G. and Stetter, K.O. (1986) Pyrococcus furiosus sp. nov. represents a novel genus of marine heterotrophic archaebacteria growing optimally at 100°C. Arch. Micro- biol. 145, 56-61.

2 ScNifer, T. and Sch6nheit, P. (1991) Pyruvate metabolism of the hyperthermophilic archaebacterium Pyrococcus fu- riosus. Acetate formation from acetyl-CoA and ATP syn- thesis are catalyzed by an acetyl-CoA synthetase (ADP forming). Arch. Microbiol. 155, 366-377.

3 Schhfer, T. and Sch6nheit, P. (1992) Maltose fermentation to acetate, CO 2 and H 2 in the anaerobic hyperther- mophilic archaeon Pyrococcus furiosus: evidence for the operation of a novel sugar fermentation pathway. Arch. Microbiol. 158, 188-202.

4 Kengen, S.W.M. and Stams, A.J.M. (1994) Formation of L-alanine as a reduced end product in a carbohydrate fermentation by the hyperthermophilic archaeon Pyrococ- cus furiosus. Arch. Microbiol. 161, 168-175.

5 Schhfer, T., Selig, M. and Sch6nheit, P. (1992) Acetyl-CoA synthetase (ADP-forming) in archaea, a novel enzyme involved in acetate formation and ATP synthesis. Arch. Microbiol. 159, 72-83.

6 Sch~ifer, T. and Sch6nheit, P. (1993) Gluconeogenesis from pyruvate in the hyperthermophilic archaeon Pyrococcus

114

furiosus: involvement of reactions of the Embden-Meyer- hof pathway. Arch. Microbiol. 159, 354-363.

7 Kengen, S.W.M. and Stams A.J.M. (1994) Growth and energy conservation in batch cultures of Pyrococcus furio- sus. FEMS Microbiol. Lett. 117, 305-310.

8 Veiga-Da-Cunha, M., Firme, P., San Romfio, V. and San- tos, H. (1992) Application of 13C nuclear magnetic reso- nance to elucidate the unexpected biosynthesis of erythri- tol by Leuconostoc oenos. Appl. Environ. Microbiol. 58, 2271-2279.

9 Siebers, B. and Hensel, R. (1993) Glucose catabolism of the hyperthermophilic archaeum Thermoproteus tenax. FEMS Microbiol. Lett. 111, 1-8.

10 Storey, K.B. (1982) Phosphofructokinase from oyster ad- ductor muscle. In: Methods in Enzymology, Carbohydrate Metabolism (Wood, W.A., Ed.), Vol 90, pp. 39-44.

11 Dorn, M., Andreesen, J.R. and Gottschalk, G. (1978) Fermentation of fumarate and L-malate by Clostridium formicoaceticum. J. Bacteriol. 133, 26-32.

12 GraBl, M. and Supp, M. (1985) Alanine: determination with alanine aminotransferase and lactate dehydrogenase.

In: Methods of Enzymatic Analysis (Bergmeyer, H.U., Ed.), Vol. 8, pp. 345-349. Verlag Chemie, Weinheim

13 Kunst, A., Draeger, B. and Ziegenhorn, J. (1981) Colori- metric methods with glucose oxidase and peroxidase. In: Methods of Enzymatic Analysis (Bergmeyer, H.U., Ed.), Vol. 6., pp. 178-185. Verlag Chemie, Weinheim.

14 Mukund, S. and Adams, M.W.W. (1991) The novel tung- sten-iron-sulfur protein of the hyperthermophilic archae- bacterium, Pyrococcus furiosus, is an aldehyde ferredoxin oxidoreductase. Evidence for its participation in a uaique glycolytic pathway. J. Biol. Chem. 266, 14208-14216.

15 Altekar, W. and Rangaswamy, V. (1991) Ketohexokinase (ATP: D-fructose 1-phosphotransferase) initiates fructose breakdown via the modified EMP pathway in halophilic archabacteria. FEMS Microbiol. Lett. 83, 241-246.

16 Horecker, B.L., Tsolas and Lai, C.Y. (1971) Aldolases. In: The Enzymes (Boyer, P.D. et al., Eds.), Vol. 7, pp. 213-258. Academic Press, New York, NY.

17 Bryant, F.O. and Adams, M.W.W. (1989) Characterization of hydrogenase from the hyperthermophilic archaebac- terium Pyrococcus furiosus. J. Biol. Chem. 264, 5070-5079.