Autotrophic CO2 Fixation by Chloroflexus aurantiacus: Study of

JOURNAL OF BACTERIOLOGY, July 2001, p. 4305–4316 0021-9193/01/$04.00⫹0 DOI: 10.1128/JB.183.14.4305–4316.2001 Copyright © 2001, American Society for Microbiology. All Rights Reserved.

Vol. 183, No. 14

Autotrophic CO2 Fixation by Chloroflexus aurantiacus: Study of Glyoxylate Formation and Assimilation via the 3-Hydroxypropionate Cycle SYLVIA HERTER,1 JAN FARFSING,1 NASSER GAD’ON,1 CHRISTOPH RIEDER,2 WOLFGANG EISENREICH,2 ADELBERT BACHER,2 AND GEORG FUCHS1* Mikrobiologie, Institut Biologie II, Universita ¨t Freiburg, Freiburg,1 and Organische Chemie und Biochemie, Technische Universita ¨t Mu ¨nchen, Munich,2 Germany Received 12 January 2001/Accepted 20 April 2001

In the facultative autotrophic organism Chloroflexus aurantiacus, a phototrophic green nonsulfur bacterium, the Calvin cycle does not appear to be operative in autotrophic carbon assimilation. An alternative cyclic pathway, the 3-hydroxypropionate cycle, has been proposed. In this pathway, acetyl coenzyme A (acetyl-CoA) is assumed to be converted to malate, and two CO2 molecules are thereby fixed. Malyl-CoA is supposed to be cleaved to acetyl-CoA, the starting molecule, and glyoxylate, the carbon fixation product. Malyl-CoA cleavage is shown here to be catalyzed by malyl-CoA lyase; this enzyme activity is induced severalfold in autotrophically grown cells. Malate is converted to malyl-CoA via an inducible CoA transferase with succinyl-CoA as a CoA donor. Some enzyme activities involved in the conversion of malonyl-CoA via 3-hydroxypropionate to propionyl-CoA are also induced under autotrophic growth conditions. So far, no clue as to the first step in glyoxylate assimilation has been obtained. One possibility for the assimilation of glyoxylate involves the conversion of glyoxylate to glycine and the subsequent assimilation of glycine. However, such a pathway does not occur, as shown by labeling of whole cells with [1,2-13C2]glycine. Glycine carbon was incorporated only into glycine, serine, and compounds that contained C1 units derived therefrom and not into other cell compounds. Chloroflexus aurantiacus is a phototrophic green nonsulfur bacterium that grows facultatively autotrophically. Energy can be obtained by anoxic photosynthesis, respiration with oxygen, or fermentation (24, 31, 32). The classic Calvin-BensonBassham cycle does not appear to operate in autotrophic carbon fixation, although early work reported the presence of ribulose 1,5-bisphosphate carboxylase (37). The key enzymes of this cycle were not detectable by other groups, and the 13 C-12C isotope discrimination was outside the range of values reported for Calvin cycle organisms (14), indicating the possibility of a different pathway. Indeed, evidence has been accumulated for an alternative carbon fixation cycle, which has been termed the 3-hydroxypropionate cycle (12, 14, 15, 16, 39, 40). This proposed pathway also seems to operate in the branch of archaebacteria comprising autotrophic Acidianus, Sulfolobus, and Metallosphaera species (4, 18, 27, 30). The proposed 3-hydroxypropionate cycle (40) is shown in Fig. 1. Each turn of the cycle results in the net fixation of two molecules of bicarbonate into one molecule of glyoxylate. Acetyl coenzyme A (acetyl-CoA) is carboxylated to malonylCoA by conventional ATP-dependent biotin-containing acetylCoA carboxylase. The unprecedented reduction of malonylCoA to propionyl-CoA requires three NADPH molecules and one MgATP molecule and proceeds via free 3-hydroxypropionate as an intermediate. 3-Hydroxypropionate and succinate are even excreted into the medium by autotrophically grown cultures when factors other than the availability of an electron

donor (hydrogen gas) or a carbon source (carbon dioxide) become limiting (15, 16, 39). Propionyl-CoA is carboxylated to methylmalonyl-CoA, followed by the isomerization of methylmalonyl-CoA to succinyl-CoA; these reactions are conventional and are used in many organisms for propionate assimilation. Interestingly, succinyl-CoA appears to be used for malate activation by CoA transfer, forming succinate and malyl-CoA; succinate in turn is oxidized to malate by conventional enzymes. Malyl-CoA is cleaved to acetyl-CoA and glyoxylate. In conclusion, acetyl-CoA is considered the starting molecule of the cycle and is converted to malate. Two CO2 fixation steps, acetyl-CoA carboxylation and propionyl-CoA carboxylation, are involved in this reaction sequence. A different pathway has been proposed in which acetyl-CoA is reductively carboxylated to pyruvate by ferredoxin-dependent pyruvate synthase (19, 21). Pyruvate then can be converted to phosphoenolpyruvate (PEP) and oxaloacetate via pyruvate phosphate dikinase and PEP carboxylase. Oxaloacetate then can be reduced to malate and malate can be converted to acetyl-CoA and glyoxylate, as described for the 3-hydroxypropionate cycle. Based on this proposal, the labeling pattern of cellular building blocks can be predicted, e.g., when cells are fed with 13C-labeled acetate or succinate. However, the results of labeling studies were not consistent with this proposal (12, 39). Furthermore, the postulated CO2-fixing enzyme pyruvate synthase could not be demonstrated unequivocally (42). In addition, the roles of malonyl-CoA reduction to propionyl-CoA, propionyl-CoA carboxylation, and the excretion of 3-hydroxypropionate could not be explained. According to this hypothetical pathway, malate should be formed solely from oxaloacetate. Contrary to this proposal, in the 3-hydroxypropionate cycle

* Corresponding author. Mailing address: Mikrobiologie, Institut Biologie II, Scha¨nzlestrasse 1, D-79104 Freiburg, Germany. Phone: 49-761-2032649. Fax: 49-761-2032626. E-mail: fuchsgeo@uni-freiburg .de. 4305

4306

J. BACTERIOL.

HERTER ET AL.

FIG. 1. Proposed 3-hydroxypropionate cycle of autotrophic CO2 fixation in the phototrophic green nonsulfur bacterium C. aurantiacus (12, 39). Enzyme activities: (1) acetyl-CoA carboxylase; (2) malonylCoA reductase (NADPH); (3) 3-hydroxypropionate dehydrogenase (NADP⫹); (4) 3-hydroxypropionyl-CoA synthetase; (5) 3-hydroxypropionyl-CoA dehydratase; (6) acryloyl-CoA reductase (NADPH); (7) propionyl-CoA carboxylase; (8) methylmalonyl-CoA epimerase; (9) methylmalonyl-CoA mutase; (10) succinyl-CoA:L-malate CoA transferase; (11) succinate dehydrogenase; (12) fumarate hydratase; (13) L-malyl-CoA lyase. Note that different enzyme activities do not necessarily mean different enzymes.

PEP carboxylation is believed to function as an anaplerotic reaction filling up C4 compounds which are withdrawn from the cycle for the biosynthesis of building blocks such as amino acids of the aspartate family. Hence, malate can be formed either from succinyl-CoA or from oxaloacetate. There are three novel metabolic processes in the proposed 3-hydroxypropionate cycle that need to be clarified. The first one concerns the reformation of acetyl-CoA. Previous work (40) could not provide final evidence as to how malate was converted to the starting molecule acetyl-CoA. Succinyl-CoAand malate-dependent formation of acetyl-CoA and glyoxylate was shown in cell extracts at relatively low rates. It was assumed that malyl-CoA was an intermediate; however, malylCoA was not available to demonstrate such a cleavage reaction. The second unsolved problem concerns the fate of glyoxylate. So far, no mechanism is known for Chloroflexus sp. by which glyoxylate is converted into a central biosynthetic intermediate, such as acetyl-CoA, pyruvate, or oxaloacetate. Some proposals were made, but experimental evidence was scarce (12, 19, 20, 21, 39, 40). These proposals included the following: (i) condensation of two molecules of glyoxylate to form tartronate semialdehyde and subsequent reduction to glycerate— this proposal suffers from the fact that CO2 is released, not a useful trait of a CO2 fixation mechanism; (ii) conversion of glycine to a one-carbon unit, again with the release of CO2, and reassimilation of the one-carbon unit by condensation with

glycine, forming serine—the same objection applies to this proposal; (iii) conversion of one molecule of glyoxylate to glycine and condensation of another molecule of glyoxylate with glycine, forming hydroxyaspartate, or any other condensation reaction of glycine; and (iv) glycine reduction to acetylphosphate. The third issue is the reduction of malonyl-CoA to propionyl-CoA. This complex reaction sequence, which formally involves five steps (Fig. 1), appears to be brought about by only two enzymes, which have been purified and studied (M. Hu ¨gler, B. Alber, and G. Fuchs, unpublished results). The present work addresses the first two problems. The first aim was to provide evidence for malyl-CoA cleavage to acetylCoA and glyoxylate by malyl-CoA lyase and for the activation of malate to malyl-CoA by CoA transfer with succinyl-CoA as a CoA donor. The formation of glyoxylate from malate via these two reactions was unequivocally demonstrated. This result prompted us to further study the second problem, the fate of glyoxylate. A possible route of assimilation would be conversion of glyoxylate to glycine and assimilation of glycine following this conversion. This possibility was tested by longterm labeling of autotrophically grown cells with [1,2-13C2] glycine and determining label incorporation into cellular constituents. MATERIALS AND METHODS Bacteria and growth conditions. C. aurantiacus strain OK-70-fl (DSM636) was grown in 5- or 12-liter glass fermentors to an optical density of 3.5 to 4 at 55°C and a pH of ⬃8. Autotrophic growth under anaerobic conditions on a minimal medium supplemented with vitamins and trace elements and gassed with H2CO2 (80:20 [vol/vol]) was described elsewhere (40). Cells were grown under photoheterotrophic anaerobic conditions on modified minimal medium D (5) supplemented with 0.25% Casamino Acids, 0.1% yeast extract, and trace elements. The medium was buffered with 0.05% glycylglycine–Na⫹ buffer. Cell extracts and membrane fractions. Cell extracts were prepared anaerobically. Cells were suspended in an equal volume of 20 mM Tris-HCl buffer (pH 7.8) containing 1 mg of DNase I per 2 ml of cell suspension. Except for determination of the Mg2⫹ dependence of enzyme activity, the buffer also contained 2 mM MgCl2. The cell suspension was passed through a French pressure cell at 137 MPa, followed by centrifugation (100,000 ⫻ g) at 4°C for 1 h. The supernatant (cell extract) was either used immediately or kept frozen at ⫺70°C. The protein content of the extract was determined by the Bradford method (3) and ranged from 15 to 37 mg of protein ml⫺1. Membrane protein fractions were obtained aerobically from 1 g of cells resuspended in 2 ml of 20 mM Tris-HCl buffer (pH 7.8) containing 20% (wt/vol) glycerol. Cells were broken in a French pressure cell, the suspension was centrifuged at 20,000 ⫻ g (4°C, 1 h), and the supernatant was centrifuged at 100,000 ⫻ g (4°C, 1 h). The formed pellet was washed with 2 ml of 20 mM Tris HCl, pH 7.8, containing 20% (wt/vol) glycerol, resuspended in 2 ml of the same buffer containing 0.5% Triton X-100 (vol/vol), and centrifuged at 100,000 ⫻ g (4°C, 1 h). The supernatant obtained contained the solubilized membrane fraction (2.3 to 3.0 mg of protein ml⫺1). Radiochemicals, isotopes, chemicals, and biochemicals. Radiochemicals were obtained from American Radiolabeled Chemicals Inc./Biotrend Chemikalien GmbH (Ko ¨ln, Germany), Amersham (Braunschweig, Germany), Sigma-Aldrich (Deisenhofen, Germany), or NEN Life Science (Bad Homburg, Germany). 14 ⫺1 14 ⫺1 L-[1,4(2,3)- C]Malate (1.87 MBq ␮mol ), [1,2- C2]oxalate (88.7 kBq ␮mol ), [1,2-14C2]acetate (1.98 MBq ␮mol⫺1), [14C]Na2CO3 (2.02 MBq ␮mol⫺1), and [1,2-14C2]glycine (4.34 MBq ␮mol⫺1) were used. [1,2-13C2]Glycine was obtained from Cambridge Isotope Laboratories (Andover, Mass.). All chemicals except those used for growing cells were analytical grade and were obtained from Fluka (Neu-Ulm, Germany), Sigma-Aldrich, or Roth (Karlsruhe, Germany). Biochemicals were obtained from Roche Diagnostics (Mannheim, Germany), Applichem (Darmstadt, Germany), or Gerbu (Craiberg, Germany). Syntheses. Various possible intermediates of the 3-hydroxypropionate cycle were not commercially available, notably, 14C-labeled compounds. (i) Malonyl-CoA. Monothiophenylmalonate was chemically synthesized as described previously (17, 33) and stored under nitrogen gas at ⫺20°C. CoA (61 ␮mol) was incubated under anaerobic conditions at room temperature in 40 ml

VOL. 183, 2001 of 0.1 M NaHCO3 solution (pH 7.5) with 122 ␮mol of monothiophenylmalonate (dissolved in 300 ␮l of dioxan and 700 ␮l of 0.1 M NaHCO3). After 90 min, the pH was adjusted to pH 3 by the addition of 1 M HCl, and the solution was extracted twice with diethyl ether. The aqueous solution was lyophilized, and the powder was stored at ⫺20°C. (ii) L-Malyl-CoA. L-Malyl-CoA was chemically synthesized as described previously (10), with a slight modification. The synthesis intermediate L-malylcaprylcysteamine (S-[␤-hydroxysuccinyl]-N-caprylcysteamine) was synthesized by Richard Krieger (Institut fu ¨r Organische Chemie, Universita¨t Freiburg, Freiburg, Germany) as described previously (10, 28). L-Malyl-CoA was stored as a freezedried powder at ⫺20°C. It contained 72% CoA-ester and 28% CoA, as determined by high-pressure liquid chromatography (HPLC) separation and detection at 260 nm. (iii) Succinyl-CoA, acetyl-CoA, and propionyl-CoA. The CoA-thioesters of succinate, acetate, and propionate were synthesized from their anhydrides by a slightly modified method described previously (36, 38). (iv) 3-Hydroxypropionate. 3-Hydroxypropionate was obtained by heating an alkaline solution (60 ml; 16 g of NaOH) of 3-hydroxypropionitrile (0.13 mol) until ammonia was completely evaporated. The pH of the solution was adjusted to pH 2, and the free acid formed was extracted in a Kutscher-Steudel apparatus with diethyl ether. The product (9.5 ml) contained ⬃4.4 M 3-hydroxypropionate and was stored at ⫺20°C. (v) [1,2-14C2]Acetyl-CoA. Radioactively labeled acetyl-CoA was synthesized enzymatically. The reaction mixture (1 ml) contained 0.3 mM CoA, 0.2 mM [1,2-14C2]acetate (0.4 MBq), 1 mM ATP, 1 U of acetyl-CoA synthetase (Roche, Basel, Switzerland), and 1 mM MgCl2 in 50 mM Tris-HCl buffer (pH 8.4). An ATP-regenerating system consisting of 1 mM PEP, 0.5 mM NADH, and the enzymes myokinase (1 U), pyruvate kinase (1.5 U), and L-lactate dehydrogenase (2.8 U) was included. The reaction was carried out at 30°C, and NADH oxidation was monitored spectrophotometrically at 365 nm. The addition of a fivefold amount of ethanol stopped the reaction, and the precipitated protein was removed by centrifugation. Ethanol was evaporated, and the radioactive acetylCoA was purified by using an ODS-AQ extraction minicolumn (500 mg; 3 ml; YMC, Schermbeck, Germany) and following the protocols of the supplier. [14C] Acetyl-CoA was eluted with 6 ml of 80% methanol, and samples containing 16.7 kBq of [14C]acetyl-CoA each were dried in a Speedvac concentrator and stored at ⫺20°C. (vi) [1,2-14C2]Glyoxylate. [14C]Glyoxylate was synthesized from [1,2-14C2] oxalate by a previously described method (29). It was purified by HPLC and stored at ⫺20°C and pH 7.0. Enzyme assays. Unless otherwise indicated, tests were carried out at 55°C under aerobic conditions with extracts of autotrophically grown cells; insensitivity of enzyme activity toward oxygen was checked in each case. For comparison, enzyme assays were also performed with extracts of heterotrophically grown cells. (i) L-Malyl-CoA lyase. Enzyme activity was monitored spectrophotometrically at 324 nm as described previously (13), with some modifications (E324 of glyoxylate-phenylhydrazone, 17,000 M⫺1 cm⫺1). The assay mixture (0.5 ml) contained 200 mM morpholinopropanesulfonic acid (MOPS)–K⫹ buffer (pH 7.7), 10 mM MgCl2, 3.5 mM phenylhydrazinium chloride, 0.2 mM L-malyl-CoA, and 0.1 to 0.3 mg of protein. L-Malyl-CoA routinely started the reaction. When the apparent Km was determined, the value for L-malyl-CoA varied between 0.5 and 0.005 mM. The stoichiometry of the reaction was investigated with 0.01, 0.02, and 0.05 mM L-malyl-CoA. (ii) L-Malyl-CoA synthetase. Enzyme activity was measured using endogenous L-malyl-CoA lyase activity, which produces glyoxylate; glyoxylate-phenylhydrazone formation was monitored spectrophotometrically at 324 nm. The assay mixture (0.5 ml) contained 200 mM MOPS–K⫹ buffer (pH 7.7), 10 mM MgCl2, 3.5 mM phenylhydrazinium chloride, 0.3 mM CoA, 3 mM ATP or GTP, 5 mM L-malate, and 0.1 to 1.7 mg of protein. L-Malate started the reaction. The specific activity of endogenous L-malyl-CoA lyase was 52 to 220 nmol min⫺1 mg of protein⫺1, depending on the extract. (iii) Succinyl-CoA:L-malate CoA transferase. The succinyl-CoA- and L-malate-dependent formation of glyoxylate due to L-malyl-CoA cleavage by endogenous L-malyl-CoA lyase was monitored spectrophotometrically at 324 nm to determine enzyme activity (40). The assay mixture (0.5 ml) contained 200 mM MOPS–K⫹ buffer (pH 6.5), 10 mM MgCl2, 3.5 mM phenylhydrazinium chloride, 1 mM succinyl-CoA, 5 mM L-malate, and 0.3 to 1.1 mg of protein. Either substrate could be used to start the reaction. The apparent Kms were determined at saturating concentration of the second substrate using 0.1 to 1.3 mM succinylCoA and 0.1 to 5.0 mM L-malate. The pH optima of the reactions were determined with MOPS–K⫹ buffer at various pHs (6.0 to 8.9).

AUTOTROPHIC 3-HYDROXYPROPIONATE CYCLE

4307

(iv) Apparent malate synthase reaction (reverse reaction of malyl-CoA lyase). The formation of free CoA was monitored spectrophotometrically at 412 nm with 5,5⬘-dithiobis(2-nitrobenzoate) (DTNB; Ellman’s reagent) (E412, 13,600 M⫺1 cm⫺1) (8). The assay mixture (0.5 ml) contained 100 mM potassium phosphate buffer (pH 7.0), 0.25 mM DTNB, 10 mM MgCl2, 0.3 mM acetyl-CoA, 5 mM glyoxylate, and 0.1 to 0.3 mg of protein. The addition of either substrate started the reaction. The apparent Kms were determined at saturating concentrations of the second substrate using 0.01 to 0.2 mM acetyl-CoA and 0.1 to 5 mM glyoxylate. (v) Acetyl-CoA and L-malyl-CoA hydrolysis. The formation of free CoA was monitored spectrophotometrically at 412 nm with 5,5⬘-dithiobis(2-nitrobenzoate). The assay mixture (0.5 ml) contained 100 mM potassium phosphate buffer (pH 7.0), 0.25 mM DTNB, 10 mM MgCl2, 0.2 mM L-malyl-CoA or 0.3 mM acetyl-CoA, and 0.1 to 0.3 mg of protein. Either acetyl-CoA or L-malyl-CoA started the reaction. Cell extract was omitted when chemical CoA-ester stability under these conditions was determined. (vi) Glyoxylate reduction. The reduction of glyoxylate was monitored spectrophotometrically by measuring the oxidation of NADH or NADPH at 365 nm (E365 of NADH, 3,390 M⫺1 cm⫺1; E365 of NADPH, 3,490 M⫺1 cm⫺1) (8) and 45°C. The assay mixture (0.5 ml) contained 100 mM potassium phosphate buffer (pH 7.8), 10 mM KCl, 2 mM MgCl2, 1 mM dithioerythritol (DTE), 0.3 mM thiamine diphosphate, 0.3 mM NADH or NADPH, 10 mM glyoxylate, and 0.5 to 1.0 mg of protein. Glyoxylate started the reaction. The pH optima of the reactions were determined with 100 mM potassium phosphate buffer at different pHs (6.0 to 7.8) and 100 mM Tris-HCl buffer (pH 8.5). The apparent Kms were determined at saturating concentrations of the second substrate using 1 to 10 mM glyoxylate and 0.03 to 0.3 mM NADH or NADPH. (vii) Acetyl-CoA and propionyl-CoA carboxylases. Acetyl-CoA and propionylCoA carboxylases were assayed as described previously (40), with slight modifications. The acetyl-CoA- or propionyl-CoA- and MgATP-dependent fixation of 14 C from [14C]bicarbonate into acid-stable labeled products was monitored. The assay mixture (1 ml) contained 100 mM Tris-HCl buffer (pH 7.8), 5 mM MgCl2, 4 mM ATP, 2 mM NADPH, and 0.4 mM acetyl-CoA or propionyl-CoA. In addition, 10 mM [14C]KHCO3 (36.7 kBq; specific radioactivity, 2.5 Bq nmol⫺1), 5 mM DTE, and 3.0 to 4.0 mg of protein were added. Cell extract started the reaction. (viii) PEP carboxylase. The PEP-dependent fixation of 14C from [14C]bicarbonate into acid-stable labeled products was monitored. The assay mixture (1 ml) contained 100 mM Tris-HCl (pH 7.8), 5 mM MgCl2, 2 mM PEP, 2 mM NADH, 10 mM [14C]KHCO3 (36.7 kBq; specific radioactivity, 2.5 Bq nmol⫺1), 5 mM DTE, and 2 mg of protein. Cell extract started the reaction. (ix) Pyruvate carboxylase. The pyruvate- and MgATP-dependent fixation of 14 C from [14C]bicarbonate into acid-stable labeled products was monitored. The assay mixture contained 100 mM Tris-HCl-buffer (pH 7.8), 5 mM MgCl2, 2 mM pyruvate, 5 mM ATP, 2 mM NADH, 10 mM [14C]KHCO3 (36.7 kBq; specific radioactivity, 2.5 Bq nmol⫺1), 5 mM DTE, and 2 mg of protein. Cell extract started the reaction. The indirect fixation of label via pyruvate water dikinase and PEP carboxylase could be distinguished from biotin-dependent pyruvate carboxylation by the addition of 1 nM avidin to the assay mixture, completely inhibiting pyruvate carboxylation. (x) Pyruvate phosphate dikinase. The pyruvate-, MgATP-, and phosphatedependent fixation of 14C from [14C]bicarbonate into acid-stable labeled products was monitored. PEP formed was carboxylated by endogenous PEP carboxylase to oxaloacetate. The assay mixture (1 ml) contained 100 mM Tris-HCl buffer (pH 7.8), 5 mM MgCl2, 2 mM pyruvate, 5 mM ATP, 5 mM K2HPO4, 2 mM NADH, 10 mM [14C]KHCO3 (36.7 kBq; specific radioactivity, 2.5 Bq nmol⫺1), 5 mM DTE, and 2 mg of protein. Cell extract started the reaction. (xi) Malonyl-CoA reduction to 3-hydroxypropionate. The malonyl-CoA-dependent oxidation of NADPH was monitored spectrophotometrically at 365 nm. The assay mixture (0.5 ml) contained 100 mM Tris-HCl buffer (pH 7.8), 0.3 mM NADPH, 1 mM malonyl-CoA, and 0.5 to 1.0 mg of protein. The addition of malonyl-CoA started the reaction. (xii) Reduction of 3-hydroxypropionate to propionyl-CoA. The 3-hydroxypropionate-, CoA-, and MgATP-dependent oxidation of NADPH was monitored spectrophotometrically at 365 nm. The assay mixture contained 100 mM TrisHCl buffer (pH 7.8), 10 mM KCl, 5 mM MgCl2, 3 mM ATP, 0.5 mM CoA, 0.3 mM NADPH, 1 mM 3-hydroxypropionate, and 0.2 to 1.0 mg of protein. The addition of 3-hydroxypropionate started the reaction. (xiii) Succinate dehydrogenase. Succinate dehydrogenase activity was assayed anaerobically by spectrophotometrically measuring the succinate-dependent reduction of oxidized 2,6-dichlorophenolindophenol at 546 nm in membrane fractions (E546, 13,800 M⫺1 cm⫺1; this value was extrapolated from the UV-visible spectrum of 2,6-dichlorophenolindophenol [E600, 22,000 M⫺1 cm⫺1]) (8). The

4308

HERTER ET AL.

assay mixture (0.5 ml) contained 100 mM Tris-HCl buffer (pH 7.8), 2.5 mM 2,6-dichlorophenolindophenol, 1 mM phenazine methosulfate, 0.5 mM succinate, and 11 to 15 ␮g of protein (solubilized membrane fraction). Succinate started the reaction. (xiv) Fumarate hydratase. Fumarate hydratase activity was monitored spectrophotometrically at 240 nm (E240, 2,440 M⫺1 cm⫺1) (2). The assay mixture (0.5 ml) contained 100 mM Tris-HCl buffer (pH 7.8), 0.2 mM fumarate, and 0.2 to 0.5 mg of protein. The addition of fumarate started the reaction. (xv) Malate dehydrogenase. Malate dehydrogenase activity was assayed spectrophotometrically by measuring the oxaloacetate-dependent oxidation of NADPH or NADH at 365 nm. The assay mixture (0.5 ml) contained 100 mM Tris-HCl buffer (pH 7.8), 0.3 mM NADH or NADPH, 1 mM oxaloacetate, and 0.1 to 0.2 mg of protein. Oxaloacetate started the reaction. (xvi) Citrate synthase. Citrate synthase activity was monitored spectrophotometrically at 412 nm with DTNB as the CoA-detecting agent. The assay mixture (0.5 ml) contained 100 mM Tris-HCl buffer (pH 7.8), 0.25 mM DTNB, 0.2 mM acetyl-CoA, 2 mM oxaloacetate, and 0.3 mg of protein. Either acetyl-CoA or oxaloacetate started the reaction. (xvii) Isocitrate dehydrogenase. Isocitrate dehydrogenase activity was monitored by measuring NADP⫹ reduction spectrophotometrically at 365 nm. The assay mixture (0.5 ml) contained 100 mM Tris-HCl buffer (pH 7.8), 5 mM MgCl2, 0.125 mM FeCl2, 1 mM NADP⫹, 5 mM mercaptoethanol, 10 mM mixture of Dand L-isocitrate, and 0.6 to 0.8 mg of protein. The addition of isocitrate started the reaction. (xviii) Aconitase. Aconitase activity was determined by coupling the reaction to endogenous isocitrate dehydrogenase activity, which reduces NADP⫹. The assay mixture (0.5 ml) was identical to the isocitrate dehydrogenase assay mixture, except for 10 mM citrate instead of isocitrate and 1.0 mg of protein. Citrate started the reaction. (xix) 2-Oxoglutarate and pyruvate dehydrogenases. 2-Oxoglutarate and pyruvate dehydrogenase activities were monitored spectrophotometrically by measuring NAD⫹ or NADP⫹ reduction at 365 nm. The assay mixture (0.5 ml) contained 100 mM Tris-HCl buffer (pH 7.8), 5 mM MgCl2, 1 mM NAD⫹ or NADP⫹, 0.5 mM CoA, 10 mM mercaptoethanol, 0.5 mM thiamine diphosphate, 5 mM 2-oxoglutarate or pyruvate, and up to 2.9 mg of protein. The addition of 2-oxoglutarate or pyruvate started the reaction. Formation of acetyl-CoA from L-malyl-CoA. The assay mixture (0.5 ml) contained 200 mM MOPS–K⫹ buffer (pH 7.7), 10 mM MgCl2, 0.9 mM L-malyl-CoA, 0.3 mM CoA (as an impurity of L-malyl-CoA), and 0.3 mg of protein. The addition of cell extract started the reaction. Samples of 100 ␮l were taken after different periods of incubation, and adding 20 ␮l of 1 M HCl stopped the reaction. Protein was removed by centrifugation, and samples were analyzed by HPLC using system 2 (see below). Formation of [14C]malyl-CoA from [1,4(2,3)-14C]malate and ATP or GTP. The assay mixture (0.5 ml) contained 1 mM L-[14C]malate (36.7 kBq), 10 mM MgCl2, 0.3 mM CoA, 3 mM ATP or GTP, and 0.3 mg of protein. The addition of cell extract started the reaction. After incubation for 1, 10, 20, and 50 min, samples of 100 ␮l were taken; the reaction was stopped by adding 20 ␮l of 1 M HCl. Protein was removed by centrifugation, and samples were analyzed by HPLC using system 2 (see below). Formation of [14C]malyl-CoA and [14C]glyoxylate from [1,4(2,3)-14C]malate and succinyl-CoA. The assay mixture (0.5 ml) contained 100 mM potassium phosphate buffer (pH 6.7), 10 mM KCl, 10 mM MgCl2, 0.5 mM L-[14C]malate (36.7 kBq), 1 mM succinyl-CoA, and 0.5 mg of protein. The addition of cell extract started the reaction. After incubation at 45°C for various time periods, samples of 100 ␮l were taken; the reaction was stopped by adding 12 ␮l of concentrated HCl. Protein was removed by centrifugation, and samples were analyzed by HPLC using systems 1 and 2 (see below). A control reaction was carried out in which succinyl-CoA was omitted. For detection of [14C]glyoxylate formed from L-[14C]malate, a 50 ␮l-sample was retrieved after 20 min of incubation at 55°C; the reaction was stopped by adding 5 ␮l of concentrated HCl. Protein was removed by centrifugation, and 30 ␮l of 2,4-dinitrophenylhydrazine (0.1% [wt/vol] in 2 M HCl) was added to the supernatant. After incubation at 30°C for 30 min, the developed 2,4-dinitrophenylhydrazones were extracted twice in ethyl acetate and separated by thin-layer chromatography (TLC) on cellulose F254 plates with solvent system 3 (see below). The supernatant was analyzed by HPLC using system 2 (see below) to identify [14C]malyl-CoA. The radioactive peak eluting at 11 min was collected and incubated at pH 12 and 70°C for 30 min to hydrolyze CoA-esters. The sample was acidified by adding solid NaHSO4, and the free acids were extracted overnight by Kutscher-Steudel extraction with diethyl ether. TLC on cellulose plates with solvent system 1 (see below) was used to analyze part of the sample. The 14C-labeled product and the malate standard cochromatographed with an Rf of 0.32.

J. BACTERIOL. Formation of [14C]malyl-CoA from [1,2-14C2]acetyl-CoA and glyoxylate. The assay mixture (0.5 ml) contained 100 mM potassium phosphate buffer (pH 7.0), 10 mM MgCl2, 5 mM glyoxylate, 0.5 mM [14C]acetyl-CoA (16.7 kBq), and 0.8 mg of protein. The addition of cell extract started the reaction. Samples of 125 ␮l were taken at intervals, and the reaction was stopped by adding 12 ␮l of 1 M HCl. Protein was removed by centrifugation, and samples were analyzed by HPLC using system 2 (see below). In a control experiment, glyoxylate was omitted. Conversion of [1,2-14C2]glyoxylate to products. [14C]Glyoxylate (1.1 mM, 36.7 kBq) was incubated with and without 0.5 mM NADH or NADPH and cell extract (0.5 to 1.0 mg of protein). In addition, the assay mixture (0.2 ml) contained 100 mM potassium phosphate buffer (pH 7.5), 10 mM KCl, and 0.3 mM thiamine diphosphate. Cell extract started the reaction, and the mixture was then incubated at 45°C for 40 min. The addition of 12 ␮l of concentrated HCl stopped the reaction. Protein was removed by centrifugation, and the supernatant was analyzed by HPLC. When 0.19 mM [14C]glyoxylate (1.7 kBq) was incubated with 2 mM glycine or 2 mM acetyl-CoA, the assay mixture (0.1 ml) additionally contained 2 mM glyoxylate, 100 mM NH4HCO3 (pH 7.8), 5 mM MgCl2, and 0.15 mg of protein. After 50 min of incubation, the addition of 0.5 ml of ethanol stopped the reaction. Protein was removed by centrifugation, and samples were analyzed by TLC on cellulose plates with solvent system 2 (see below). Determination of glyoxylate carboligase and tartronate semialdehyde reductase activities. A slight modification of the method of Chang et al. (6) was used for the detection of 14CO2 released from [1,2-14C2]glyoxylate. The reaction mixture (0.5 ml) contained 100 mM potassium phosphate buffer (pH 7.5), 0.3 mM thiamine diphosphate, 10 mM KCl, 1 mM DTE, 2.8 mM [14C]glyoxylate (35.5 kBq), and 1 mg of protein. The reaction was started with cell extract, and the mixture was shaken for 10 to 50 min at 180 rpm and 45°C in closed vials coated with alkaline Whatman paper. The reaction was stopped by the addition of 20 ␮l of concentrated HCl, and the mixture was then shaken again for 15 min at 45°C. The radioactivity trapped by the Whatman paper was measured. An assay mixture without cell extract was used as a control. HPLC. The following systems were used. (i) System 1. For separation of free organic acids, a Polyspher OA HY column (300 by 6.5 mm; Merck, Darmstadt, Germany) was used with 5 mM H2SO4 as an eluent and a flow rate of 0.8 ml min⫺1. Retention times of organic acids detected at 210 nm were 3.7 min (oxalate), 5.8 min (glyoxylate), 6.5 min (glycerate), and 7.2 min (glycolate). (ii) System 2. An RP-C18 column (LiChrospher 100, end capped, 5 ␮m, 125 by 4 mm; Merck) was used for the separation of CoA-thioesters. A gradient of 1 to 8% acetonitrile in 50 mM potassium phosphate buffer (pH 6.7) and a flow rate of 1 ml min⫺1 over 30 min were used. CoA-esters were detected at 260 nm. Retention times were 2 min (malate), 10 min (L-malyl-CoA), 13 min (CoA and succinyl-CoA), and 19 min (acetyl-CoA). Simultaneous detection of standard compounds and 14C-labeled reaction products was possible by using two detectors (UV and radioactivity) in series. TLC. Cellulose or cellulose F254 plates (0.1 mm; Merck) were used for product separation. The solvent systems used were as follows: 1—pentan-1-ol–formate– water (48.8 ml:48.8 ml:2.4 ml); 2—butan-1-ol–formate–water (10:2:15 [vol/vol]); and 3—butan-1-ol–ethanol–NH3 (35%)–water (140:20:1:39 [vol/vol]). 2,4-Dinitrophenylhydrazones were separated on cellulose F254 plates with solvent system 3 as described previously (26, 34). The corresponding Rfs for the glyoxylate-2,4dinitrophenylhydrazones were 0.41 and 0.62 (cis and trans isomers, respectively). The radioactive products cochromatographed with nonradioactive glyoxylate2,4-dinitrophenylhydrazones. The standard was detected by UV light and sprayed with 10% (wt/vol) NaOH solution (8). The plates were analyzed with a Fujix BAS1000 imaging analyzer (Fuji Film, Tokyo, Japan). Other samples cochromatographed with radioactive or nonradioactive standard compounds. Organic acids were detectable by spraying the plates with a 0.05% bromocresol green solution (Fluka Chemie, Buchs, Switzerland) at an acidic pH (8). Determination of acid-stable 14C. The fixation of 14C from [14C]bicarbonate into nonvolatile acid-stable products was measured with samples (200 ␮l) incubated for 1, 2, and 5 min at 55°C (see above). The addition of 50 ␮l of 5 M H2SO4 (to pH 1 to 2) stopped the reaction. Volatile 14CO2 (nonfixed) was removed from the samples by vigorously shaking the samples in scintillation vials for 2 h. The remaining radioactivity in the samples was determined by liquid scintillation counting. The radioactivity in samples from two control experiments, in which the substrate and the extract were omitted, served as blanks and controls. Determination of amount of 14C. The amount of 14C in liquid samples (up to 200 ␮l) and in Whatman paper was determined by liquid scintillation counting using 3 ml of Rotiszint 2200 scintillation cocktail (Roth). The counting efficiency (75 to 85%) was determined via the channel ratio method. Long-term labeling of growing cells with [1,2-13C2]glycine. For the long-term labeling of an autotrophically grown culture (5 liters) with glycine, [1,2-13C2]

VOL. 183, 2001


FIG. 2. Spectrophotometric assay of glyoxylate-phenylhydrazone formation from L-malyl-CoA by cell extracts (0.16 mg of protein) of autotrophically grown C. aurantiacus at 55°C. L-Malyl-CoA (0.2 mM) started the reaction. In control experiments, the addition of 10 mM L-malate and 0.3 mM CoA could not start the reaction. The assay mixture (0.5 ml) additionally contained MOPS–K⫹ buffer (pH 7.7), 10 mM MgCl2, and 3.5 mM phenylhydrazinium chloride (initial rate: 139 nmol min⫺1 mg of protein⫺1). glycine was proffered from a sterile solution (50 ml) containing 0.2 g of [1,2-13C2] glycine and 0.18 MBq of [1,2-14C2]glycine by use of a peristaltic pump. The feeding rate during the 140-h feeding period (five generations) was adjusted to the cell density. Less than 10% of cell carbon should be proffered by glycine. Aliquots were retrieved at intervals, and protein was determined by a modified Lowry method (25). The radioactivity in the cells and in the cell-free supernatant was determined by liquid scintillation counting. NMR analyses. The isolation of amino acids and ribonucleosides from cell mass and the experimental setup for 1H and 13C nuclear magnetic resonance (NMR) experiments were previously described (12, 39).

RESULTS Conversion of L-malyl-CoA to acetyl-CoA and glyoxylate. From previous work, it has been postulated that C. aurantiacus possesses malyl-CoA lyase to release the CO2 acceptor acetylCoA and glyoxylate as the CO2 fixation product. The indirect evidence was based on the spectrophotometric detection of glyoxylate-phenylhydrazone when both succinyl-CoA and Lmalate were present. We synthesized L-malyl-CoA to prove the existence of L-malyl-CoA lyase and to study the enzyme reaction in more detail at the growth temperature of 55°C. Extracts of autotrophically grown cells catalyzed the L-malylCoA-dependent formation of glyoxylate detected as phenylhydrazone (Fig. 2). The reaction rate was linear for less than 1 min when 0.2 mM L-malyl-CoA was added and then leveled off. The specific activity determined at the pH optimum, pH 7.5, was 52 to 220 nmol min⫺1 mg of protein⫺1, depending on the batch of cells. Extracts of heterotrophically grown cells were far less active, 7 to 19 nmol min⫺1 mg of protein⫺1, also

4309

depending on the batch of cells. The reaction rate was linearly dependent on the amount of protein added in the range of 0 to 0.1 mg of protein/0.5-ml assay. Controls with L-malate, CoA, or both (Fig. 2) did not show glyoxylate formation. The reaction was stimulated twofold by 10 mM Mg2⫹. It showed a bell-shaped pH dependence, with an optimum at pH 7.5 and half-maximal rates at pHs 6.6 and 8.3 in MOPS–K⫹ buffer. The initial rates followed Michaelis-Menten kinetics at concentrations ranging up to 0.1 mM L-malyl-CoA; higher concentrations of L-malyl-CoA resulted in slight substrate inhibition. Half-maximal rates were obtained with 0.03 mM Lmalyl-CoA. To determine the stoichiometry of the reaction, we first tested whether L-malyl-CoA remained stable at neutral pH and 55°C. In the absence of cell extract, 1 nmol of CoA was formed per min; the addition of 0.15 mg of protein/0.5-ml assay increased the rate of CoA release to 11 nmol min⫺1 mg of protein⫺1. The stoichiometric ratio of glyoxylate formed to L-malyl-CoA added was 0.77:1. It must be taken into account that some L-malyl-CoA was hydrolyzed nonenzymatically and enzymatically. Also, L-malyl-CoA was added in the Km range of substrate concentrations; hence, the reaction may not have been completed. Therefore, the corrected stoichiometric ratio may be close to 1:1. To demonstrate acetyl-CoA formation, the assay was performed without phenylhydrazine, and samples were taken after different periods of incubation and separated by HPLC. Separation of the substrate L-malyl-CoA, CoA, and the product acetyl-CoA is shown for one sample in Fig. 3A. The time course of the reaction is given in Fig. 3B. L-Malyl-CoA consumption was concomitant with the formation of acetyl-CoA and of some CoA. The formation of acetyl-CoA and CoA together accounted for the observed consumption of L-malyl-CoA. The initial rates were 70 nmol of acetyl-CoA formed min⫺1 mg of protein⫺1, 109 nmol of L-malyl-CoA consumed min⫺1 mg of protein⫺1, and 30 nmol of CoA released min⫺1 mg of protein⫺1. The equilibrium constant of the malyl-CoA lyase reaction is 3 ⫻ 10⫺3 M (13), allowing the measurement of the reverse reaction. [14C]Acetyl-CoA (0.5 mM) and glyoxylate (5 mM) were incubated at 55°C and pH 7.0 with cell extract; in a control experiment, glyoxylate was omitted. 14C-labeled products were analyzed by HPLC and 14C monitoring. Only two main fractions of 14C-labeled products were observed (Fig. 4A): one was [14C]malyl-CoA, and the other contained polar products (non-CoA-thioesters). Only traces of a compound comigrating with succinyl-CoA were formed. Identification of the formed malyl-CoA was based on the exact cochromatography with the malyl-CoA standard and, after mild alkaline hydrolysis of this compound, the formation of a [14C]-labeled acid which on TLC exactly cochromatographed with the malate standard. The time course of [14C]acetyl-CoA consumption and product formation is shown in Fig. 4B. The specific rate of [14C]acetyl-CoA consumption was 208 nmol min⫺1 mg of protein⫺1. In the absence of glyoxylate, little [14C]acetylCoA was consumed (Fig. 4B). The initial rate of malyl-CoA formation was 91 nmol min⫺1 mg of protein⫺1. It is obvious that malyl-CoA was formed and consumed again and that polar products, which might be [14C]malate or fumarate, ac-

4310

HERTER ET AL.

J. BACTERIOL.

FIG. 3. L-Malyl-CoA cleavage to acetyl-CoA and glyoxylate. (A) HPLC detection of acetyl-CoA formation from L-malyl-CoA by extracts of autotrophically grown cells after 10 min of incubation at 55°C. The reaction mixture (0.5 ml) contained 200 mM MOPS–K⫹ buffer (pH 7.7), 10 mM MgCl2, 0.9 mM L-malyl-CoA, and 0.3 mg of protein. (B) Time course of the reaction: circles, L-malyl-CoA; triangles, acetyl-CoA; and squares, CoA. Note that the L-malyl-CoA preparation contained 28% CoA as an impurity. Therefore, besides 0.9 mM L-malyl-CoA, 0.3 mM CoA was also initially present in the assay mixture. The 77% peak area corresponds to 0.9 mM CoA-thioester or free CoA.

cumulated. Malyl-CoA formation clearly reflects malyl-CoA lyase activity. Under the experimental conditions, acetyl-CoA was shown to be reasonably stable in the presence of cell extract (Fig. 4B). This finding was additionally confirmed by spectrophotometrically measuring the release of CoA from acetyl-CoA with DTNB (2 to 5 nmol min⫺1 mg of protein⫺1). The product malyl-CoA was spontaneously hydrolyzed at 55°C and pH 7.0, at a rate of 1 nmol min⫺1(detected with DTNB) (data not shown). The rate of L-malyl-CoA hydrolysis by cell extracts under these conditions was approximately 11 nmol min⫺1 mg of protein⫺1. When the spectrophotometric assay with DTNB was performed with glyoxylate and acetyl-CoA, CoA formation reached 55 nmol min⫺1 mg of protein⫺1 in extracts of autotrophically grown cells. Half-maximal rates of CoA formation were observed with 0.06 mM glyoxylate and 0.1 mM acetyl-CoA. In extracts of heterotrophically grown cells, only a small amount of CoA formation was measured (⬍1 nmol min⫺1 mg of protein⫺1). These results indicate that in addition to malyl-CoA formation from acetyl-CoA and glyoxylate, enzymatic malyl-CoA hydrolysis to malate or even the direct formation of malate occurred. A plausible explanation for this phenomenon is given in the Discussion. Formation of malyl-CoA from malate. The proposed 3-hydroxypropionate cycle assumes that malyl-CoA is cleaved to— rather than formed from—acetyl-CoA and glyoxylate; both directions would be possible from a thermodynamic point of view. The cleavage direction requires malyl-CoA synthesis

from malate. There are two options for the synthesis of malylCoA: either in a nucleoside triphosphate-dependent synthetase reaction or in a CoA transferase reaction. The MgATP- or MgGTP-dependent formation of malylCoA from malate and CoA was measured by two assays. In one assay, the formation of malyl-CoA was coupled to the endogenous malyl-CoA lyase reaction, and glyoxylate-phenylhydrazone formation was monitored spectrophotometrically. In the other assay, the conversion of [14C]malate to [14C]malyl-CoA or [14C]acetyl-CoA (formed from malyl-CoA due to endogenous malyl-CoA lyase activity) was analyzed by HPLC. In neither case was nucleoside triphosphate-dependent formation of [14C]malyl-CoA or [14C]acetyl-CoA observed. The acyl-CoA-dependent formation of malyl-CoA via CoA transfer was assayed by two independent methods. The first method spectrophotometrically monitored the succinyl-CoAand L-malate-dependent formation of glyoxylate-phenylhydrazone, which is due to two enzyme reactions, those of CoA transferase and L-malyl-CoA lyase: succinyl-CoA ⫹ L-malate 3 succinate ⫹ L-malyl-CoA L-malyl-CoA

⫹ phenylhydrazine 3 acetyl-CoA ⫹

glyoxylate-phenylhydrazone Extracts of autotrophically grown cells catalyzed this overall reaction at a specific rate of 23 to 33 nmol min⫺1 mg of protein⫺1, depending on the batch of cells (Fig. 5). Acetyl-CoA and propionyl-CoA were inactive as CoA donors. The reaction

VOL. 183, 2001


4311

FIG. 4. Formation of malyl-CoA from acetyl-CoA and glyoxylate. (A) HPLC separation of 14C-labeled products formed from [14C]acetyl-CoA and glyoxylate after 1 min of incubation at 55°C with extracts of autotrophically grown cells. (B) Time course of the reaction: circles, [14C]acetylCoA; squares, [14C]malyl-CoA; triangles, [14C]succinyl-CoA; diamonds, 14C-polar products; and hexagons, [14C]acetyl-CoA in a control reaction in which glyoxylate was omitted. The reaction mixture (0.5 ml) contained 100 mM potassium phosphate buffer (pH 7.0), 10 mM MgCl2, 5 mM glyoxylate, 0.5 mM [14C]acetyl-CoA (16.7 kBq), and 0.8 mg of protein. The 100% peak area corresponds to a 0.5 mM concentration.

rate was linearly dependent on the amount of protein added in the range of 0 to 1.0 mg of protein. The pH optimum was 6.4; half-maximal rates were observed at pH 7.0, and 20% activity was observed at pH 7.6. Half-maximal rates of glyoxylate formation were observed with 0.5 mM succinyl-CoA. In extracts of heterotrophically grown cells, virtually no glyoxylate formation was observed (ⱕ1 nmol min⫺1 mg of protein⫺1). This strong effect may have been due partly to the lower L-malylCoA lyase activity in these cells (Table 1); this activity is necessary for the coupled assay. Also, at the pH of the assay (pH 6.4), L-malyl-CoA lyase activity was only approximately 40% the maximal activity measured at the pH optimum, pH 7.5. Still, these experiments suggest that CoA transferase activity also is down-regulated in heterotrophically grown cells. The second method was based on HPLC separation and radiodetection of products formed from [14C]malate and succinyl-CoA (45°C, pH 6.7) (Fig. 6). Labeled malate, malyl-CoA, succinyl-CoA, and acetyl-CoA could be clearly separated (Fig. 6A). The time course of the reaction is shown in Fig. 6B. [14C]Malate (0.5 mM) in the presence of succinyl-CoA (1 mM) was rapidly converted to [14C]malyl-CoA, followed by [14C] acetyl-CoA formation. The formation of a small amount of another labeled CoA-ester, possibly [14C]succinyl-CoA, was also observed. No labeled CoA-esters were formed in control experiments without succinyl-CoA. The same samples were analyzed by TLC for the formation of 14C-labeled glyoxylate from [14C]malate. The results were comparable to those of the other two assays (data not shown); [14C]glyoxylate formation was strictly dependent on the presence of succinyl-CoA.

These results show that, in fact, a CoA transferase using succinyl-CoA as a CoA donor forms malyl-CoA, which subsequently is cleaved by malyl-CoA lyase to acetyl-CoA and glyoxylate. This process is effectively regulated (Table 1). Comparative study of enzyme activities related to carbon assimilation pathways. Various enzyme activities involved in autotrophic carbon metabolism were investigated with cells grown autotrophically with H2-CO2 or heterotrophically with Casamino Acids-yeast extract, anaerobically, and exposed to light. The specific enzyme activities measured at the growth temperature of 55°C are given in Table 1. Many enzyme activities did not show a significant difference under the two growth conditions. However, there were some remarkable exceptions. Activities that were higher in autotrophically grown cells were NADPH-dependent reduction of malonyl-CoA; NADPH-, ATP-, and CoA-dependent reductive conversion of 3-hydroxypropionate; succinyl-CoA:L-malate CoA transferase; and Lmalyl-CoA lyase. Down-regulation of the CoA transferase in heterotrophically grown cells is deduced from the fact that L-malyl-CoA lyase could easily be detected, whereas succinylCoA-dependent malate cleavage was virtually undetectable, suggesting very low CoA transferase activity. All of these enzyme activities are postulated to play a role in the autotrophic CO2 fixation cycle. Succinate dehydrogenase activity was higher in heterotrophically grown cells, possibly in keeping with a higher demand under such conditions. Either no or very small amounts of NAD(P)⫹-dependent pyruvate and 2-oxoglutarate dehydrogenase activities were detectable. This finding indicates that the citric acid cycle is incomplete under autotrophic and

4312

HERTER ET AL.

FIG. 5. Spectrophotometric assay of glyoxylate-phenylhydrazone formation dependent on succinyl-CoA (1 mM) and L-malate (5 mM). The assay mixture (0.5 ml) additionally contained 200 mM MOPS–K⫹ buffer (pH 6.5), 10 mM MgCl2, 3.5 mM phenylhydrazinium chloride, and 0.3 mg of protein. In the first experiment, L-malyl-CoA was added after the L-malate-dependent reaction had been recorded. This procedure served as a control to show that L-malyl-CoA lyase was not the rate-limiting step in the coupled assay.

heterotrophic growth conditions and that pyruvate may not be the precursor of acetyl-CoA. All enzymes forming 2-oxoglutarate from oxaloacetate and acetyl-CoA and those oxidizing succinate to oxaloacetate were present. It appears that pyruvate is an intermediate in the carbon fixation pathway that can be converted via PEP into carbohydrates and C4 compounds: extracts contained pyruvate phosphate dikinase and PEP carboxylase activities, whereas pyruvate carboxylase and pyruvate kinase activities were not detectable. These experiments show that the postulated enzymes of the 3-hydroxypropionate cycle were present at specific activities that were high enough to explain the rate of growth of autotrophic cells growing exponentially at a generation time of 26 to 40 h (see Discussion). Furthermore, the regulation of essential enzyme activities was in keeping with the proposed 3-hydroxypropionate cycle. Conversion of glyoxylate to products. [1,2-14C2]Glyoxylate was synthesized and tested in cell extracts to monitor the further conversion of glyoxylate. One possible glyoxylate assimilation reaction is the conversion of two molecules of [14C]glyoxylate to [14C]tartronate semialdehyde under conditions of 14 CO2 release. This reaction is catalyzed by glyoxylate carboligase (EC 4.1.1.47). Incubation of 1.1 mM [1,2-14C2]glyoxylate did not result in the formation of labeled products or the loss of radioactivity due to the formation of volatile 14CO2. Another possible reaction is the conversion of glyoxylate with glycine to 3-hydroxyaspartate, catalyzed by 3-hydroxyaspartate

J. BACTERIOL.

aldolase (EC 4.1.3.14). Incubation of 2.2 mM [1,2-14C2]glyoxylate with 2 mM glycine did not result in product formation. Incubation of [14C]glyoxylate with acetyl-CoA resulted in the formation of [14C]malyl-CoA, followed by the formation of 14 C-labeled malate and fumarate, as described in the experiment with [14C]acetyl-CoA and glyoxylate (see above). This glyoxylate consumption reaction is ascribed to the reverse reaction of fully reversible malyl-CoA lyase. Extracts of autotrophically grown cells catalyzed the glyoxylate-dependent oxidation of NADPH and NADH, with specific activities at the pH optimum, pH 7.6, of 29 and 18 nmol min⫺1 mg of protein⫺1, respectively (45°C). The product was shown by HPLC and radiodetection to be [14C]glycolate. Half-maximal rates were observed with 2.5 mM glyoxylate and 0.03 mM NADPH. Hence, other than glyoxylate reduction to glycolate, no transformation of glyoxylate alone or in combination with glycine could be detected. Long-term labeling of whole cells with [1,2-13C2]glycine. Glyoxylate conversion to glycine and the assimilation of glycine via several hypothetical or existing reactions would be a possibility for the assimilation of glyoxylate into cell carbon. Some possible reactions were mentioned above. The possibility of glycine assimilation was tested by feeding an autotrophically grown culture [1,2-13C2]glycine continuously for several generations. The feeding rate was exponentially increased and chosen so that less than 10% of cell carbon was derived from

TABLE 1. Enzyme activities relevant for autotrophic CO2 fixation and glyoxylate assimilation in C. aurantiacus at 55°Ca Enzyme or enzyme activity

Sp act (nmol min⫺1 mg of protein⫺1) under the following growth conditions: Autotrophic

Acetyl-CoA carboxylase Propionyl-CoA carboxylase Malonyl-CoA reduction (NADPH) 3-Hydroxypropionate reduction (NADPH) Succinate dehydrogenaseb (oxidized dichlorophenolindophenol) Fumarate hydratase Malyl-CoA lyase plus succinyl-CoA:Lmalate CoA transferase L-Malyl-CoA lyase PEP carboxylase Pyruvate phosphate dikinase Citrate synthase Aconitase Isocitrate dehydrogenase (NADP⫹) 2-Oxoglutarate dehydrogenase NAD⫹ NADP⫹ Malate dehydrogenase NADH NADPH Pyruvate dehydrogenase NAD⫹ NADP⫹

Heterotrophic

10–16 19 162 104 126

25 18 79 32 443

404 23–33

422 ⬍1

52–220 247 12 33 29 47

7–19 181 11 178 48 103

⬍1 ⬍1

3.2 3.6

447 336

645 460

⬍1 ⬍1

⬍1 ⱕ2

a Enzyme activities were comparatively studied with extracts of cells anaerobically grown in the light under autotrophic (H2-CO2) and heterotrophic (Casamino Acids-yeast extract) conditions. The following enzyme activities were searched for but were not detectable under either set of growth conditions: pyruvate carboxylase and pyruvate water dikinase. Mean values were obtained from at least two determinations and two cell batches. Standard deviations generally were ⫾20%. b Milligrams of protein in the solubilized membrane protein fraction.

VOL. 183, 2001


4313

FIG. 6. Formation of malyl-CoA from succinyl-CoA and malate. (A) HPLC separation of 14C-labeled products formed from [14C]malate and succinyl-CoA after 4 min of incubation at 45°C with extracts of autotrophically grown cells. (B) Time course of the reaction: circles, [14C]malylCoA; squares, [14C]acetyl-CoA; triangles, [14C]succinyl-CoA; and hexagons, [14C]acetyl-CoA in a control reaction in which succinyl-CoA was omitted. The assay mixture (0.5 ml) contained 100 mM potassium phosphate buffer (pH 6.7), 10 mM KCl, 10 mM MgCl2, 0.5 mM L-[14C]malate (36.7 kBq), 1 mM succinyl-CoA, and 0.5 mg of protein. The lag phase was due to the low temperature of the assay at the beginning of the experiment, when the assay temperature unintentionally had not yet reached 45°C. The 100% peak area corresponds to a 0.5 mM concentration.

proffered glycine and more than 90% was derived from CO2. The low rate of glycine supply ensured that glycine did not turn down autotrophic carbon fixation. This important prerequisite was found to be correct when the 13C content of cell constituents was compared with that of proffered glycine (less than 10%; Table 2). Biosynthetic building blocks, such as amino acids and nucleosides, were isolated from cellular polymers, and the amount and distribution of 13C in the individual carbon atoms were determined by NMR spectroscopic techniques. These techniques also allowed the determination of the degree of coupling of one carbon atom to another; 100% coupling between two carbon atoms meant that the two carbon atoms of glycine were incorporated as a C2 unit without breakage of the COC bond, rather than by rearrangement reactions involving COC bond cleavage or by separate incorporation of C1 units. From the observed labeling pattern, the patterns of labeling of central precursor metabolites was deduced by a retrobiosynthetic approach. This method has been described elsewhere (1, 12). Only a few of the investigated compounds were labeled: glycine, serine, purine nucleosides, and compounds containing C1 units. All other compounds carried virtually no 13C label. The results are summarized in Table 2. The patterns of labeling of primary metabolites gained from feeding [1,2-13C2]glycine are summarized in Table 3. The amino acids glycine and serine showed not only very high 13C abundance but also high 13 C-13C coupling. These data indicate that a large amount of proffered glycine was directly incorporated into cellular glycine

and serine, as well as into compounds derived from these primary building blocks, such as purine nucleosides and C1 units. All the other metabolites were virtually unlabeled. C-2 carbon atoms of glycine and serine were enriched to a higher degree than C-1 carbon atoms. This finding suggests that the carboxyl of glycine had been partially exchanged with unlabeled CO2 and that part of serine was derived from glycine. The labeling data for glycine and serine were in good accordance with the pattern of labeling of C1 units and of the purine moiety of guanosine: C-2 and C-8 of guanosine, C-5 of methionine, and C-3 of serine were derived from C1 units; C-4 and C-5 of guanosine were derived from glycine. These data clearly show that glycine and serine are not intermediates in the autotrophic carbon fixation pathway. Rather, glycine metabolism is restricted to glycine-serine interconversion and to the supply of one-carbon units. DISCUSSION Required specific enzyme activities. A prerequisite for any proposed pathway is that the postulated enzyme activities can be measured in cell extracts. Furthermore, the in vitro rates must be at least as high as the minimal rates that can explain the growth rate, i.e., in this case, under autotrophic growth conditions. The generation times were 26 to 40 h for autotrophically grown cells and 3 to 10 h for heterotrophically grown cells. The calculated CO2 fixation rate for autotrophically grown cells was 26 to 17 nmol of CO2 fixed min⫺1 mg of

4314

HERTER ET AL.

J. BACTERIOL.

TABLE 2. Absolute 13C abundance of building blocks from C. aurantiacus autotrophically grown for five generations in the presence of small amounts of [1,2-13C2]glycinea Metabolite

Carbon atom

%

13

C

Alanine

1 2 3

2.1 3.1 2.2

Aspartate

1 2 3 4

2.1 4.2 3.7 2.1

Glutamate

1 2 3 4 5

1.3 2.7 3.1 2.9 1.8

Glycine

1 2

21.1 30.4

Serine

1 2 3

9.3 15.6 11.7

Methionine

1 2 3 4 5

1.7 3.5 2.7 1.4 22.9

Threonine

1 2 3 4

5.7 9.6 7.4 3.7

Histidine

1 2 3 4 5 6

2.2 2.6 2.6 4.0 3.0 2.8

Tyrosine

1 2 3 4 5/9 6/8 7

1.4 2.8 1.8 2.6 1.9 2.4 1.3

Cytidine

2 4 5 6 1⬘ 2⬘ 3⬘ 4⬘ 5⬘

1.5 4.2 2.1 2.4 2.2 2.5 1.5 2.7 2.0

Guanosine

2 4 5 6 8 1⬘ 2⬘ 3⬘ 4⬘ 5⬘

17.8 16.8 36.8 Not determined 18.9 1.7 2.0 1.2 2.1 1.6

a

See the text for details.

protein⫺1; this value corresponds to 13 to 8.5 nmol of intermediates cycled in this pathway min⫺1 mg of protein⫺1, since two molecules of CO2 are fixed per one turn of the cycle. The CO2 fixation rate for growing cells was estimated from the equation dS/dt ⫽ (␮/y)x, correlating substrate S (moles of inorganic carbon) consumption per time (t) unit (minute) with the specific growth rate ␮, the molar growth yield y, and the dry cell mass x of the culture. A generation time of 26 h corresponds to a ␮ of 0.00044 min⫺1, y is 24 g of dry cell mass formed per mol of CO2 assimilated (assuming 50% of dry cell mass is carbon), and x is 1 g of dry cell mass, corresponding to 500 mg of total cell protein. All activities of the postulated cycle were detectable, and the rates met this requirement. Notably, the observed specific activity of the postulated enzyme L-malyl-CoA lyase, which was higher than the required minimal activity, the low apparent Km for L-malyl-CoA, and the down-regulation of the activity in heterotrophically grown cells support the proposed role of this enzyme in CO2 fixation. Similar arguments hold true for the CoA transferase, which activates L-malate by using succinyl-CoA. Malate-forming reactions. The conversion of glyoxylate with acetyl-CoA to malate or malyl-CoA is not considered a meaningful reaction in autotrophically grown cells, since malyl-CoA cleavage to acetyl-CoA and glyoxylate is considered the physiological direction of L-malyl-CoA lyase (EC 4.1.3.24). Nonetheless, the in vitro formation of malate from acetyl-CoA and glyoxylate was observed, but only in extracts of autotrophically grown cells. Besides malate synthase (EC 4.1.3.2), several known enzyme reactions might have caused this reaction. Reaction (1) catalyzed by L-malyl-CoA lyase could be followed by a hydrolase activity (2): acetyl-CoA ⫹ glyoxylate N L-malyl-CoA

(1)

⫹ H2O 3 L-malate ⫹ CoA

(2)

L-malyl-CoA

Citrate synthase (EC 4.1.3.7), in the absence of its substrates, shows high nonspecific L-malyl-CoA hydrolase activity (11, 41). Malate synthesis from acetyl-CoA and glyoxylate via the combined action of malyl-CoA lyase and citrate synthase mimicking an apparent malate synthase activity has been described previously by Cox and Quayle (7) for Pseudomonas sp. strain AM1 (now Methylobacterium extorquens). Also, an acyl-CoA thioesterase (EC 3.1.2.20) acting on L-malyl-CoA might be present (22). The situation in C. aurantiacus is very much the same as in M. extorquens, i.e., the simultaneous presence of active malyl-CoA lyase and citrate synthase under autotrophic growth conditions. We therefore believe that, in fact, citrate synthase is responsible for apparent malate synthase and enzymatic L-malyl-CoA hydrolysis activities under autotrophic growth conditions. Since L-malyl-CoA lyase activity was strongly down-regulated under heterotrophic growth conditions, only slight apparent malate synthase activity was measurable under these conditions. Regulation of enzyme activities. The levels of various enzyme activities have been comparatively studied with cells grown autotrophically and heterotrophically. Interestingly, some reactions which are essential for the 3-hyproxypropionate cycle seem to be down-regulated in heterotrophically grown cells. These reactions concern the reduction of malonylCoA, the reductive conversion of 3-hydroxypropionate, and

VOL. 183, 2001


TABLE 3. 13C abundance and standard deviation of 13C content of the individual carbon atoms of central metabolites, glycine, serine, and C1 units from C. aurantiacus autotrophically grown in the presence of [1,2-13C2]glycine Metabolite

Carbon atom

%

13

C

SD (na)

Acetyl-CoA

1 2

1.53 2.63

0.30 (4) 0.46 (4)

Pyruvate

1 2 3 1 2 3 4

1.90 3.15 2.31 2.10 3.32 2.62 1.52

0.44 (3) 0.37 (8) 0.30 (7) 0.40 (3) 0.73 (6) 0.29 (5) 0.35 (6)

2-Oxoglutarate

1 2 3 4 5

1.37 2.53 2.73 2.70 1.53

0.31 (3) 0.38 (3) 0.40 (3) 0.53 (3) 0.38 (3)

Ribose-5-phosphate

1 2 3 4 5

1.87 2.31 1.40 2.32 1.90

0.26 (3) 0.30 (3) 0.14 (3) 0.32 (3) 0.26 (3)

Oxaloacetate

Glycineb

1 2

18.9 33.6

Serinec

1 2 3

9.3 15.6 11.7

5 2 8 3

22.9 17.8 18.9 11.7

C1 units Methionine Guanosine Serine

2.1 (2) 3.2 (2)

a

Number of values used for statistical analysis of the SD. The coupling of C-1 with C-2 was 37.4%; that of C-2 with C-1 was 51.8%. The coupling of C-1 with C-2 was 86.6%; that of C-2 with C-1 was 26.8%; that of C-2 with C-3 was 16.7%; that of C-2 with both C-1 and C-3 was 26.8%; and that of C-3 with C-2 was 56.5%. b c

the conversion of L-malate to glyoxylate and acetyl-CoA. This regulation needs to be studied in more detail to elucidate the underlying mechanisms. Acetyl-CoA and propionyl-CoA carboxylase activities remain unchanged. High acetyl-CoA carboxylase activity under heterotrophic growth conditions may be required due to the four- to eightfold higher growth rate of heterotrophically grown cells, which in any case require acetylCoA carboxylase activity for fatty acid synthesis. In contrast, propionyl-CoA carboxylase activity normally is not required in carbon metabolism, and the unchanged activity in heterotrophically grown cells was unexpected. However, propionyl-CoA carboxylase activity and acetyl-CoA carboxylase activity may be due to one carboxylase enzyme. This assumption could explain the lack of down-regulation of propionyl-CoA carboxylase activity under heterotrophic growth conditions. C1 metabolism. Proffered [1,2-13C2]glycine served as a precursor for cellular glycine and to a lesser extent for L-serine. This finding follows from the labeling pattern and 13C content of the amino acids isolated from cell material. Glycine C-2 obviously also served as a precursor for C1 units. The strong

4315

labeling of C1 units was deduced from the labeling of the methyl group of methionine and C-2 and C-8 of purine; even C-3 of serine was labeled. The specific 13C content of these C1 units was approximately 50% lower than that of C-2 of glycine, which served as a C1 unit precursor. This result indicates that approximately half of the C1 units must have been derived from C-2 of glycine and supports glycine degradation by the reversible glycine cleavage system (formerly glycine synthase; EC 1.4.4.2 and EC 2.1.2.10), leading to 5,10-methylenetetrahydrofolate. This enzyme activity has been reported for C. aurantiacus (19, 20). The other half of the C1 units may have been formed from C-3 of serine; C-1 and C-2 of serine carried less 13C than C-1 and C-2 of glycine, indicating that part of serine was formed from unlabeled precursors other than glycine, e.g., 3-phosphoglycerate. It should be stressed, however, that incorporation of label from glycine into L-serine could simply occur via an exchange reaction due to the full reversibility of serine hydroxymethyltransferase (EC 2.1.2.1): 关12C]serine N [12C]glycine ⫹ C1-tetrahydrofolate 关13C]glycine ⫹ C1-tetrahydrofolate N [13C]serine Theoretical considerations also demand another source of C1 units. The conversion of L-serine to glycine alone does not provide enough C1 units for biosynthesis. For the synthesis of 1 g of cell mass, bacteria need approximately 5 mmol of amino acids. Assuming that glycine represents approximately 1/10 to 1/20 of all amino acids, exclusive synthesis of glycine from serine would yield only 0.5 to 0.25 mmol of C1 units. The actual demand, however, is 1.2 mmol (23). If glycine were formed partly from glyoxylate, the amount of C1 units formed from serine would be even lower. Possible glyoxylate assimilation reactions. The lack of labeling in the other amino acids excludes the possibility that glyoxylate is assimilated via glycine. Several possibilities have been discussed (19, 20, 39). (i) The lack of labeling of aspartate excludes the possibility that glyoxylate was assimilated via the conversion of one molecule of glyoxylate to glycine and, following the condensation of glycine with another molecule of glyoxylate, to ␤-hydroxyaspartate. ␤-Hydroxyaspartate could be converted to oxaloacetate by NH3 elimination; the respective two enzymes are 3-hydroxyaspartate aldolase (EC 4.1.3.14) and 3-hydroxyaspartate dehydratase (EC 4.2.1.38). (ii) The minimal labeling of threonine excludes the possible assimilation of glycine by condensation with acetaldehyde to form threonine. Either threonine aldolase (EC 4.1.2.5) or serine hydroxymethyltransferase (EC 2.1.2.1) catalyzes this reaction. (iii) An interesting possibility is glycine reduction to acetylphosphate catalyzed by glycine reductase. This process would result in the labeling of all carbon compounds, notably, the positions derived from acetyl-CoA. This possibility can also be excluded. (iv) Glycine conversion to serine obviously takes place, but the lack of labeling of the other cell compounds shows that serine formation is a dead end. Condensation of two molecules of glyoxylate to tartronate semialdehyde under conditions of CO2 release, as postulated elsewhere (19, 20), could not be observed in cell extracts. This route also makes little sense as part of an autotrophic pathway, since it would be associated with a loss of CO2. The NAD(P)H

4316

HERTER ET AL.

oxidation observed when glyoxylate was added could not be ascribed to tartronate semialdehyde formation coupled to NAD(P)H-dependent reduction to glycerate. Rather, NAD(P)H oxidation was shown to be due to glyoxylate reduction to glycolate. The rate of glyoxylate reduction was relatively low, and the apparent Km for glyoxylate was high, suggesting that this reaction is nonspecifically catalyzed by a dehydrogenase naturally acting on a different substrate. An enzyme known to reduce glyoxylate nonspecifically is L-lactate dehydrogenase (EC 1.1.1.27) (9, 35). To our knowledge, there is no reported way to form one of the central precursor metabolites from glycolate directly. Anaplerotic reactions. The organism seems to use pyruvate as a precursor for PEP and oxaloacetate synthesis, rather than the reverse option, the formation of pyruvate from these precursors. PEP carboxylase was quite active; pyruvate phosphate dikinase was preliminarily identified based on the pyruvate-, MgATP-, and phosphate-dependent fixation of 14C from [14C] bicarbonate. The presence of pyruvate carboxylase activity was excluded based on the lack of inhibition of pyruvate-dependent CO2 fixation by avidin. If acetyl-CoA is the starting compound of the CO2 fixation cycle, then there must be a way to convert acetyl-CoA to pyruvate. Pyruvate synthase activity catalyzing the ferredoxin-dependent reductive carboxylation of acetyl-CoA to pyruvate is rather low or absent in C. aurantiacus strain OK70-fl (42). Alternatively, 3-hydroxypropionate or propionylCoA, intermediates of the postulated cycle, could be converted to pyruvate, possibly by condensation with glyoxylate. This possibility is currently being tested in our laboratory. ACKNOWLEDGMENTS This work was supported by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie. Special thanks are due to Richard Krieger for the synthesis of L-malylcaprylcysteamine and monothiophenylmalonate. REFERENCES 1. Bacher, A., C. Rieder, D. Eichinger, D. Arigoni, G. Fuchs, and W. Eisenreich. 1999. Elucidation of novel biosynthetic pathways and metabolite flux patterns by retrobiosynthetic NMR analysis. FEMS Microbiol. Rev. 22:567–598. 2. Bergmeyer, H. U. 1970. Methoden der enzymatischen Analyse, 2nd ed., vol. 1. Verlag Chemie, Weinheim, Germany. 3. Bradford, M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248–254. 4. Burton, N. P., T. D. Williams, and P. R. Norris. 1999. Carboxylase genes of Sulfolobus metallicus. Arch. Microbiol. 172:349–353. 5. Castenholz, R. W. 1969. Thermophillic blue-green algae and the thermal environment. Bacteriol. Rev. 33:476–504. 6. Chang, Y.-Y., A.-Y. Wang, and J. E. Cronan, Jr. 1993. Molecular cloning, DNA sequencing, and biochemical analyses of Escherichia coli glyoxylate carboligase. J. Bacteriol. 268:3911–3919. 7. Cox, R. B., and J. R. Quayle. 1976. Synthesis and hydrolysis of malylcoenzyme A by Pseudomonas AM1: an apparent malate synthase activity. J. Gen. Microbiol. 95:121–133. 8. Dawson, R. M. C., D. C. Elliot, W. H. Elliot, and K. M. Jones. 1986. Data for biochemical research, 3rd ed. Clarendon Press, Oxford, United Kingdom. 9. Duncan, R. J. S. 1980. The disproportionation of glyoxylate by lactate dehydrogenase. Arch. Biochem. Biophys. 201:128–136. 10. Eggerer, H., and C. H. Gru ¨newa ¨lder. 1964. Zum Mechanismus der biologischen Umwandlung von Citronensaüre, IV. Synthese von Malyl-Coenzym A und seiner Diastereoisomeren. Liebigs Ann. Chem. 677:200–208. 11. Eggerer, H., U. Remberger, and C. Gru ¨newa ¨lder. 1964. Zum Mechanismus der biologischen Umwandlung von Citronensaüre, V. Citrat-synthase, eine Hydrolase fu ¨r Malyl-Coenzym A. Biochem. Z. 339:436–453. 12. Eisenreich, W., G. Strauss, U. Werz, G. Fuchs, and A. Bacher. 1993. Retrobiosynthetic analysis of carbon fixation in the phototrophic eubacterium Chloroflexus aurantiacus. Eur. J. Biochem. 215:619–632. 13. Hersh, L. B. 1973. Malate adenosine triphosphate lyase. Separation of the reaction into a malate thiokinase and malyl coenzyme A lyase. J. Biol. Chem. 248:7295–7303.

J. BACTERIOL. 14. Holo, H., and R. Sirevåg. 1986. Autotrophic growth and CO2 fixation in Chloroflexus aurantiacus. Arch. Microbiol. 145:173–180. 15. Holo, H., and D. Grace. 1987. Polyglucose synthesis in Chloroflexus aurantiacus studied by 13C-NMR. Arch. Microbiol. 148:292–297. 16. Holo, H. 1989. Chloroflexus aurantiacus secretes 3-hydroxypropionate, a possible intermediate in the assimilation of CO2 and acetate. Arch. Microbiol. 151:252–256. 17. Imamoto, T., M. Kodera, and M. Yokoyama. 1982. A convenient method for the preparation of S-esters of thio analogs of malonic acid. Bull. Chem. Soc. Jpn. 55:2303–2304. 18. Ishii, M., T. Miyake, T. Satoh, H. Sugiyama, Y. Oshima, T. Kodama, and Y. Igarashi. 1997. Autotrophic carbon dioxide fixation in Acidianus brierleyi. Arch. Microbiol. 166:368–371. 19. Ivanovsky, R. N., E. N. Krasilnikova, and Y. I. Fal. 1993. A pathway of the autotrophic CO2 fixation in Chloroflexus aurantiacus. Arch. Microbiol. 159: 257–264. 20. Ivanovsky, R. N., and E. N. Krasilnikova. 1995. Glyoxylate metabolism in Chloroflexus aurantiacus. Microbiology 64:257–261. 21. Kondratieva, E. N., and R. N. Ivanovsky. 1992. Carbon metabolism in Chloroflexus aurantiacus. FEMS Microbiol. Lett. 100:269–272. 22. Lee, K. Y., and H. Schulz. 1979. Isolation, properties, and regulation of a mitochondrial acyl coenzyme A thioesterase from pig heart. J. Biol. Chem. 254:4516–4523. 23. Lengler, J. W., G. Drews, and H. G. Schlegel. 1999. Biology of the prokaryotes, 1st ed. Thieme Verlag, Stuttgart, Germany. 24. Madigan, M. T., S. R. Petersen, and T. D. Brock. 1974. Nutritional studies on Chloroflexus, a filamentous photosynthetic, gliding bacterium. Arch. Microbiol. 100:97–103. 25. Markwell, M. A. K., S. M. Haas, L. L. Bieber, and E. Tolbert. 1978. A modification of the Lowry procedure to simplify protein determination in membrane and lipoprotein samples. Anal. Biochem. 87:206–210. 26. Mehlitz, A., K. Gierschner, and T. Minas. 1963. Du ¨nnschichtchromatographische Trennung von 2,4-Dinitrophenylhydrazonen. Chem. Ztg. 87:573–606. 27. Meńendez, C., Z. Bauer, H. Huber, N. Gad’On, K. O. Stetter, and G. Fuchs. 1999. Presence of acetyl-coenzyme A (CoA) carboxylase and propionyl-CoA carboxylase in autotrophic Crenarchaeota and indication for the operation of a 3-hydroxypropionate cycle in autotrophic carbon fixation. J. Bacteriol. 181: 1088–1098. 28. Mohrhauer, H., K. Christiansen, M. Gan, M. Deubig, and R. T. Holman. 1968. Improved method for the preparation of malonyl coenzyme A. J. Lipid Res. 9:398. 29. Murray, A., C. W. Bills, and A. R. Ronzio. 1952. Micro-syntheses with tracer elements. V. The synthesis of formaldehyde-C14 and of glyoxylic-1,2-C214 acid. J. Am. Chem. Soc. 74:2405–2406. 30. Norris, P. R., A. Nixon, and A. Hart. 1989. Acidophilic, mineral-oxidizing bacteria: the utilization of carbon dioxide with particular reference to autotrophy in Sulfolobus, p. 24–43. In M. S. Da Costa, J. C. Duarte, and R. A. D. Williams (ed.), Microbiology of extreme environments and its potential for biotechnology. Elsevier, London, United Kingdom. 31. Pierson, B. K., and R. W. Castenholz. 1974. A phototrophic gliding filamentous bacterium of hot springs. Chloroflexus aurantiacus, gen. and sp. nov. Arch. Microbiol. 100:5–24. 32. Pierson, B. K., and R. W. Castenholz. 1974. Studies of pigments and growth in Chloroflexus aurantiacus, a phototrophic filamentous bacterium. Arch. Microbiol. 100:283–305. 33. Pollmann, W., and G. Schramm. 1964. Reactivity of metaphosphate esters prepared from P4O10 and ethylether. Biochim. Biophys. Acta 80:1–7. 34. Salem, A. R., A. J. Hacking, and J. R. Quayle. 1973. Cleavage of malylcoenzyme A into acetyl-coenzyme A and glyoxylate by Pseudomonas AM1 and other C1-unit-utilizing bacteria. Biochem. J. 136:89–96. 35. Sawaki, S., N. Hattori, N. Morikawa, and K. Yamada. 1967. Oxidation and reduction of glyoxylate by lactate dehydrogenase. J. Vitaminol. (Kyoto) 13:93–97. 36. Simon, E. 1957. S-succinyl coenzyme A. Biochem. Prep. 5:30–32. 37. Sirevåg, R., and R. Castenholz. 1979. Aspects of carbon metabolism in Chloroflexus. Arch. Microbiol. 120:151–153. 38. Stadtman, E. R. 1957. Preparation and assay of acyl coenzyme A and other thiol esters; use of hydroxylamine. Methods Enzymol. 3:931–941. 39. Strauss, G., W. Eisenreich, A. Bacher, and G. Fuchs. 1992. 13C-NMR study of autotrophic CO2 fixation pathways in the sulfur-reducing archaebacterium Thermoproteus neutrophilus and in the phototrophic eubacterium Chloroflexus aurantiacus. Eur. J. Biochem. 205:853–866. 40. Strauss, G., and G. Fuchs. 1993. Enzymes of a novel autotrophic CO2 fixation pathway in the phototrophic bacterium Chloroflexus aurantiacus, the 3-hydroxypropionate cycle. Eur. J. Biochem. 215:633–643. 41. Subhash, C. G., and E. E. Dekker. 1984. Malyl-CoA formation in the NAD⫹-, CoASH-, and ␣-ketoglutarate dehydrogenase-dependent oxidation of 2-keto-4-hydroxyglutarate. Possible coupled role of this reaction with 2-keto-4-hydroxyglutarate aldolase activity in a pyruvate-catalyzed cyclic oxidation of glyoxylate. J. Biol. Chem. 259:10012–10019. 42. Ugol’kova, N. V., and R. N. Ivanovsky. 2000. On the mechanism of autotrophic fixation of CO2 by Chloroflexus aurantiacus. Mikrobiologiya 69:175–179. (In Russian.)