Glucose Metabolism and Acetate Switch in Archaea: the Enzymes in Haloferax volcanii

In this work, we provide a comprehensive analysis of glucose degradation via the semiphosphorylative Entner-Doudoroff pathway in the haloarchaeal model organism Haloferax volcanii. The study includes transcriptional analyses, growth experiments with deletion mutants. and characterization of all enzymes involved in the conversion of 3-phosphoglycerate to acetyl coenzyme A (acetyl-CoA) and in anaplerosis.

ABSTRACT

The halophilic archaeon Haloferax volcanii has been proposed to degrade glucose via the semiphosphorylative Entner-Doudoroff (spED) pathway. Following our previous studies on key enzymes of this pathway, we now focus on the characterization of enzymes involved in 3-phosphoglycerate conversion to pyruvate, in anaplerosis, and in acetyl coenzyme A (acetyl-CoA) formation from pyruvate. These enzymes include phosphoglycerate mutase, enolase, pyruvate kinase, phosphoenolpyruvate carboxylase, and pyruvate-ferredoxin oxidoreductase. The essential function of these enzymes were shown by transcript analyses and growth experiments with respective deletion mutants. Furthermore, we show that H. volcanii—during aerobic growth on glucose—excreted significant amounts of acetate, which was consumed in the stationary phase (acetate switch). The enzyme catalyzing the conversion of acetyl-CoA to acetate as part of the acetate overflow mechanism, an ADP-forming acetyl-CoA synthetase (ACD), was characterized. The functional involvement of ACD in acetate formation and of AMP-forming acetyl-CoA synthetases (ACSs) in activation of excreted acetate was proven by using respective deletion mutants. Together, the data provide a comprehensive analysis of enzymes of the spED pathway and of anaplerosis and report the first genetic evidence of the functional involvement of enzymes of the acetate switch in archaea.

IMPORTANCE In this work, we provide a comprehensive analysis of glucose degradation via the semiphosphorylative Entner-Doudoroff pathway in the haloarchaeal model organism Haloferax volcanii. The study includes transcriptional analyses, growth experiments with deletion mutants. and characterization of all enzymes involved in the conversion of 3-phosphoglycerate to acetyl coenzyme A (acetyl-CoA) and in anaplerosis. Phylogenetic analyses of several enzymes indicate various lateral gene transfer events from bacteria to haloarchaea. Furthermore, we analyzed the key players involved in the acetate switch, i.e., in the formation (overflow) and subsequent consumption of acetate during aerobic growth on glucose. Together, the data provide novel aspects of glucose degradation, anaplerosis, and acetate switch in H. volcanii and thus expand our understanding of the unusual sugar metabolism in archaea.

INTRODUCTION

Comparative analyses of the central carbon metabolism in archaea, bacteria, and eukarya revealed a variety of unusual pathways and enzymes in archaea that differ from the canonical pathways of bacteria and eukarya. In particular, sugar metabolism has been studied in detail (1, 2). It was found that anaerobic and microaerophilic hyperthermophilic archaea, including Pyrococcus and Thermococcus species and Pyrobaculum aerophilum, degrade glucose and glucose polymers by modified versions of the Embden-Meyerhof pathway, whereas thermoacidophilic archaea, e.g., Picrophilus torridus and Sulfolobus species, utilize modified Entner-Doudoroff (ED) pathways, namely, the nonphosphorylative ED pathway and the branched ED pathway, respectively (3, 4).

In halophilic archaea, sugar metabolism has been studied in particular in Haloferax and Haloarcula species (5–10). Haloferax volcanii, which represents a model organism of haloarchaea (11), exhibits a high metabolic versatility in utilizing various sugars, hexoses, pentoses and deoxysugars, and other carbon compounds, e.g., acetate, as growth substrates (12). Glucose degradation in H. volcanii has been shown to proceed via the semiphosphorylative ED (spED) pathway, and the enzymes of this pathway, involved in the conversion of glucose to 3-phosphoglycerate (3PG), have been characterized (7, 13, 86). They include glucose dehydrogenase, gluconate dehydratase, 2-keto-3-deoxygluconate kinase, and 2-keto-3-deoxy-6-phosphogluconate aldolase. Furthermore, the enzymes involved in the oxidation of glyceraldehyde-3-phosphate (GAP) to 3PG, GAP dehydrogenase, and phosphoglycerate kinase were characterized. However, the enzymes involved in further degradation of 3PG to pyruvate, including phosphoglycerate mutase, enolase, and pyruvate kinase, have not been analyzed. Also, the enzymes that catalyze the conversion of pyruvate to acetyl-CoA, which is further oxidized in the citric acid cycle, and the formation of oxaloacetate as an anaplerotic reaction from glucose have not been characterized in H. volcanii.

In previous studies, we have shown that during exponential growth on glucose, the haloarchaea Haloarcula marismortui and H. volcanii excreted significant amounts of acetate which were consumed in the stationary phase (14, 15). The excretion of acetate from glucose under aerobic conditions (acetate overflow) and subsequent utilization of acetate following complete glucose consumption—together designated acetate switch—have been reported in various bacteria, e.g., Escherichia coli and Bacillus subtilis (16, 17). It has been proposed that acetate overflow in these bacteria takes place when the rate of glucose degradation exceeds that of subsequent reactions in the citric acid cycle and respiratory chain (18–21). The enzymes involved in the acetate switch in these bacteria have been analyzed; the conversion of acetyl-CoA to acetate is catalyzed by two enzymes, namely, phosphotransacetylase (PTA; acetyl-CoA + P_i → acetylphosphate + CoA) and acetate kinase (AK; acetylphosphate + ADP → acetate + ATP). The respective pta and ackA genes encoding PTA and AK are transcriptionally upregulated during growth on glucose. The subsequent reutilization of excreted acetate involves an inducible AMP-forming acetyl-CoA synthetase (ACS; acetate + CoA + ATP → acetyl-CoA + AMP + PP_i). AK and PTA are known as classical enzymes in acetate formation in anaerobic fermentation processes of bacteria, whereas ACS is the predominant enzyme of acetate activation during growth on acetate in bacteria.

Archaea do not use the bacterial enzymes AK and PTA for acetate formation; instead, it has been shown that all acetate-forming archaea, including anaerobic hyperthermophiles and aerobic haloarchaea, utilize an ADP-forming acetyl-CoA synthetase (ACD; acetyl-CoA + ADP + P_i → acetate + ATP + CoA). ACD is a novel prokaryotic enzyme involved in acetate formation and ATP synthesis replacing the canonical bacterial mechanism involving AK and PTA (22–24). In contrast, the activation of acetate in aerobic archaea is catalyzed by AMP-forming acetyl-CoA synthetase (ACS), and thus, ACS is operative in both archaea and bacteria. Cell extracts of H. marismortui and H. volcanii were shown to contain ACD activity in exponentially grown cells, whereas stationary-phase cells contain an inducible ACS activity (14, 15). So far, the genes encoding ACD and ACS activities and their functional involvement in the acetate switch have not been demonstrated in haloarchaea.

Here, we report the characterization of the spED pathway enzymes phosphoglycerate mutase, enolase, and pyruvate kinase. Also, the anaplerotic enzyme phosphoenolpyruvate carboxylase and the acetyl-CoA-forming enzyme pyruvate-ferredoxin oxidoreductase (POR) are described. Furthermore, we analyzed the functional involvement of genes associated with the acetate switch from glucose, i.e., ACD and ACS enzymes in H. volcanii.

RESULTS AND DISCUSSION

Haloferax volcanii has been reported to degrade glucose to pyruvate via the semiphosphorylative Entner-Doudoroff (spED) pathway. In our previous analyses, we have characterized the spED enzymes that catalyze the conversion of glucose to 3-phosphoglycerate (7, 13, 86). Here, we report the characterization of the spED enzymes phosphoglycerate mutase, enolase, and pyruvate kinase. Furthermore, pyruvate-ferredoxin oxidoreductase forming acetyl-CoA from pyruvate was studied. We also analyzed the enzyme phosphoenolpyruvate (PEP) carboxylase involved in anaplerosis to replenish tricarboxylic acid (TCA) cycle intermediates. The functional involvement of the enzymes was studied by transcriptional analyses and by growth experiments with the respective deletion mutants. Furthermore, we show that H. volcanii excretes significant amounts of acetate from glucose. The enzyme involved in acetate formation from acetyl-CoA, an ADP-forming acetyl-CoA synthetase, was characterized.

Phosphoglycerate mutase.

In the genome of H. volcanii, the gene gpmI (HVO_2516) is annotated to encode a putative 2,3-bisphosphoglycerate-independent phosphoglycerate mutase (iPGM). iPGMs differ from cofactor (2,3-bisphosphoglycerate)-dependent PGMs (dPGMs) that require 2,3-bisphosphoglycerate as effectors. The iPGM from H. volcanii has a calculated molecular mass of 57.2 kDa. Transcription of gpmI analyzed by Northern blotting experiments with cells grown on glucose or Casamino Acids revealed a specific signal of about 1,620 nucleotides each, which corresponds to the length of gpmI of 1,581 nucleotides indicating expression under both glycolytic and gluconeogenetic conditions (Fig. 1). The gene gpmI was overexpressed in H. volcanii H1209, and the recombinant enzyme was purified by Ni-nitrilotriacetic acid (NTA) affinity and size exclusion chromatography. The protein showed on SDS-PAGE a single band of 70 kDa (see Fig. S1 in the supplemental material); by gel filtration, a molecular mass of 62 kDa was determined and characterized iPGM as a monomeric protein. iPGM catalyzed the conversion of 3-phosphoglycerate (3PG) to 2-phosphoglycerate (2PG) with a V_max of 85.4 U/mg and a K_m for 3PG of 2.5 mM. The enzyme also catalyzed the reverse reaction, the conversion of 2PG to 3PG, with a V_max of 9.1 U/mg and a K_m for 2PG of 0.6 mM (see Fig. S2 in the supplemental material).

FIG 1.

Transcriptional analysis of gpmI encoding cofactor-independent phosphoglycerate mutase from H. volcanii. The gene gpmI (1,581 nucleotides) is depicted as a gray arrow, with flanking genes as white arrows (A). Northern blotting was performed with RNA from glucose (Glc)- and Casamino Acid (Cas)-grown cells using a probe against gpmI. 16S rRNA served as a loading control (B).

iPGM from H. volcanii showed high sequence identity (63% to 66%) to putative iPGMs from other haloarchaea, e.g., Haloarcula marismortui, Halobacterium salinarum, and Haloquadratum walsbyi, and to characterized iPGMs from the bacteria E. coli (41%) and Geobacillus stearothermophilus (44%) (25, 26). Less sequence identity (11 to 13%) was found with characterized iPGM from the archaea Archaeoglobus fulgidus, Methanocaldococcus jannaschii, Sulfolobus solfataricus, and Pyrococcus furiosus (27–30).

Phylogenetic analysis of iPGMs indicates that the homologs from bacteria, archaea, and eukarya cluster in three distinct, domain-specific clades (Fig. 2). The iPGMs of H. volcanii and haloarchaeal homologs are members of the bacterial iPGM cluster, which is distantly related to iPGMs of the archaeal cluster. The close phylogenetic relationship of haloarchaeal iPGMs to bacteria suggest a lateral gene transfer event from bacteria to haloarchaea. More detailed phylogenetic analyses of iPGMs showing a similar tree topology have been reported previously (27, 29). It should be noted that iPGMs are phylogenetically and structurally unrelated to cofactor (2,3-bisphosphoglycerate)-dependent PGMs (dPGMs). An archaeal dPGM has been characterized from Thermoplasma acidophilum (27).

FIG 2.

Phylogenetic relationship of cofactor-independent phosphoglycerate mutase from H. volcanii and haloarchaeal homologs, from other archaea, and from bacteria and eukarya. The tree is based upon a multiple amino acid sequence alignment that was generated with ClustalX (83). Numbers at the nodes are bootstrapping values according to neighbor joining. Organism names and UniProt accession numbers are as follows: Haloferax volcanii, D4GTU8; Halobacterium salinarum, Q9HNY7; Haloquadratum walsbyi, Q18GK9; Haloarcula marismortui, Q5UXB9; Archaeoglobus fulgidus, O28523; Pyrococcus furiosus, P58814; Methanocaldococcus jannaschii, Q59007; Sulfolobus solfataricus, Q980A0; Thermoplasma acidophilum, Q9HL27; Methanothermobacter thermautotrophius, O27628; Pyrobaculum aerophilum, Q8ZVE4; Geobacillus stearothermophilus, Q9X519; Bacillus subtilis, P39773; E. coli, P37689; Staphylococcus aureus, PDB entry: 4MY4; Arabidopsis thaliana, Pgm1, O04499; A. thaliana, Pgm2, Q9M9K1; Zea mays, P30792; Trypanosoma brucei, PDB entry: 3NVL.

Enolase.

In the H. volcanii genome, the gene HVO_2774 is annotated as eno encoding a putative enolase of 42 kDa. The gene eno was overexpressed in H. volcanii H1209, and the recombinant protein was purified. The His-tagged enzyme showed on SDS-PAGE a molecular mass of 45 kDa (Fig. S1) and—as analyzed by gel filtration—a molecular mass of 358 kDa characterizing the enolase as a homooctameric protein. The enzyme catalyzed the conversion of 2PG to PEP with a V_max of 121 U/mg and a K_m for 2PG of 0.63 mM. The specific activity in the reverse direction, the conversion of PEP to 2PG, was 41.17 U/mg with a K_m for PEP of 0.92 mM (Fig. S2). Activity was dependent on MgCl₂ with a K_m of 2.7 mM.

Enolases are widely distributed in archaea, bacteria, and eukarya and are highly conserved enzymes composed of about 45-kDa subunits. H. volcanii enolase showed a sequence identity of 32% to 42% to archaeal homologs from Picrophilus torridus, P. furiosus, and S. solfataricus and of 32 to 43% to the bacteria E. coli and Bacillus subtilis and to human. The high sequence identity of the enolases from the three domains of life is in accordance with a conserved function in metabolism. However, the oligomeric composition of enolases differs; in eukarya, dimeric enzymes were reported, whereas the enolases from several bacteria, e.g., Staphylococcus aureus and Bacillus subtilis, are homooctameric proteins (31, 32). In archaea, enolases have been purified from the hyperthermophiles P. furiosus and Methanocaldococcus jannaschii (33, 34). The enolase from M. jannaschii is a homooctamer as shown here for the enolase from H. volcanii. Enolases belong to the enolase superfamily which includes, e.g., various sugar acid dehydratase families that all form distinct clusters in accordance with their specific catalytic function. A phylogenetic analysis of members of the enolase family, which includes the H. volcanii enolase sequence, and of sugar acid dehydratase families within the enolase superfamily has recently been reported (10).

Pyruvate kinase.

Pyruvate kinase from H. volcanii, encoded by pykA (HVO_0806), has recently been characterized as a homotetrameric protein of 200 kDa. The enzyme showed sigmoidal saturation kinetics with the substrates PEP and ADP and exhibited a moderate allosteric activation by AMP. Detailed phylogenetic analyses of pyruvate kinases from archaea, bacteria, and eukarya indicate that pyruvate kinases from haloarchaea originated from bacterial pyruvate kinases that are activated by AMP (35).

Here, we report the transcriptional regulation of pykA and analyzed its functional involvement in glycolysis by growth studies with a deletion mutant. Transcription of pykA was analyzed by Northern blotting experiments in cells grown on glucose or pyruvate. A specific signal at 2,000 nucleotides was detected in glucose-grown cells which corresponds to the length of pykA of 1,758 nucleotides (Fig. 3). No signal was detected in pyruvate-grown cells. A pykA deletion mutant was generated that showed an impaired growth on glucose; wild-type growth was restored by in trans complementation with pykA. Growth of the mutant on pyruvate was not affected (Fig. 3). These results indicate a specific function of pyruvate kinase from H. volcanii in glycolysis. The observed residual growth of the pykA mutant suggests that alternative enzyme(s) might functionally replace pykA in PEP conversion to pyruvate. A potential candidate might be PEP synthetase for which in the thermophilic archaeon Thermococcus kodakarensis a glycolytic function of PEP conversion to pyruvate has been proposed (36). However, in H. volcanii, a significant contribution of PEP synthetase in glucose degradation seems unlikely since a PEP synthetase deletion mutant was previously shown to grow on glucose as the wild type (12).

FIG 3.

Transcriptional and deletion mutant analyses of pykA encoding pyruvate kinase from H. volcanii. The gene pykA (1,758 nucleotides) is depicted as a gray arrow, with flanking genes as white arrows (A). Northern blotting was performed with RNA from glucose (Glc)- and pyruvate (Pyr)-grown cells using a probe against HVO_0806. 16S rRNA served as loading control (B). Growth of the pykA deletion mutant from H. volcanii on 25 mM glucose (C) and on 40 mM pyruvate (D). Square, wild type; circle, ΔpykA mutant; triangle, deletion mutant complemented with pykA.

Pyruvate-ferredoxin oxidoreductase.

The oxidative decarboxylation of pyruvate to acetyl-CoA in haloarchaea has been reported to be catalyzed by pyruvate-ferredoxin oxidoreductase (POR) (pyruvate + CoA + 2 ferredoxin_ox → acetyl-CoA + 2 ferredoxin_red + CO₂). The enzyme from Halobacterium halobium has been characterized (37, 38). Here, we report the characterization of POR from H. volcanii. The functional involvement of POR in glucose degradation, converting pyruvate to acetyl-CoA and CO₂, was analyzed by transcriptional analyses and growth studies with a conditional lethal mutant. In H. volcanii, the genes porA (HVO_1305; 1,896 nucleotides) and porB (HVO_1304, 939 nucleotides) are annotated to encode the α- and β-subunit of POR (Fig. 4). Transcription of porA and porB was analyzed by Northern blotting with cells grown on glucose or acetate. Using specific probes either against porA or porB, a transcript signal each of about 2,650 nucleotides was detected in glucose-grown cells (Fig. 4), which corresponds to the sum of the lengths of porA and porB indicating that both genes are cotranscribed as an operon. No signal was detected in acetate-grown cells, demonstrating a glucose-specific upregulation of POR.

FIG 4.

Transcriptional and conditional lethal mutant analyses of the porAB operon from H. volcanii. Genomic organization of porB (HVO_1304; 939 nucleotides) and porA (HVO_1305; 1,896 nucleotides) encoding the α- and β-subunit of the pyruvate-ferredoxin oxidoreductase are given as gray arrows; flanking genes HVO_1306 (1,158 nucleotides) and HVO_1303 (429 nucleotides) are given as white arrows. The intergenic regions upstream of porA and downstream of porB are 199 and 117 nucleotides, respectively; three nucleotides are between porA and porB. (A). Northern blotting of the porAB operon was performed with RNA from glucose (Glc)- and acetate (Ac)-grown cells using probes against porA (left) and porB (right). 16S rRNA served as loading control (B). Growth of the conditional lethal mutant of porAB on 15 mM glucose (C) and 40 mM acetate (D). Square, wild type; circle, porAB mutant without tryptophan; triangle, the mutant after addition of tryptophan.

The porAB operon was cloned into the vector pTA963 and was homologously overexpressed. The recombinant enzyme was purified and showed on SDS-PAGE one band at 82 kDa (α-subunit) and a second at 44 kDa (β-subunit) (Fig. S1). The molecular masses on SDS-PAGE were higher than the calculated molecular masses of the α-subunit of 68.4 kDa and β-subunit of 34.5 kDa. Higher apparent molecular masses on SDS-PAGE have been reported for several halophilic proteins due to their high content of acidic amino acids (39, 40). By gel filtration, a molecular mass of about 190 kDa for POR was determined indicating a heterotetrameric α₂β₂ structure. POR activity was measured under anaerobic conditions by following the pyruvate- and CoA-dependent reduction of methyl viologen as an artificial electron acceptor. Recombinant POR showed an apparent V_max of 2.07 U/mg and apparent K_m values for pyruvate and CoA of 4.79 mM and 0.06 mM, respectively (Fig. S2).

To analyze the functional involvement of POR in pyruvate conversion to acetyl-CoA, a porAB conditional lethal mutant was generated. In this mutant, the porAB operon was under the control of a tryptophan inducible promoter showing expression of porAB only in the presence of tryptophan. The porAB mutant did not grow on glucose (Fig. 4) or on pyruvate (not shown) in the absence of added tryptophan, whereas the mutant grew on both substrates upon addition of tryptophan. In contrast, growth of the porAB mutant on acetate was not affected, excluding a possible role in acetyl-CoA conversion to pyruvate in the anabolism (Fig. 4). This finding is in accordance with that of previous studies showing that in H. volcanii pyruvate is formed from acetyl-CoA involving the glyoxylic acid cycle and two malic enzymes (12, 41).

The finding that the porAB mutant did not grow on glucose or pyruvate indicates that POR in H. volcanii cannot be replaced by alternative enzymes, such as a pyruvate dehydrogenase (PDH) complex, the typical enzyme of pyruvate conversion to acetyl-CoA in aerobic bacteria and eukarya (pyruvate + CoA + NAD⁺ → acetyl-CoA + CO₂ + NADH). In H. volcanii, genes encoding three 2-oxoacid dehydrogenase complexes have been identified; however, catalytic activities of these complexes, including pyruvate dehydrogenase, have not been shown (42–44). The reason why in haloarchaea and other aerobic archaea PORs rather than PDH complexes catalyze the pyruvate oxidation to acetyl-CoA is still a matter of debate (45).

POR of H. volcanii shows high sequence identity (34% for α-subunit; 38% for β-subunit) to the well-characterized POR from Halobacterium halobium (37, 38). As the H. volcanii POR, the Halobacterium enzyme is a 200-kDa α₂β₂ heterotetramer. In contrast, characterized PORs from the aerobic archaea S. solfataricus and Aeropyrum pernix constitute 100-kDa αβ dimeric enzymes showing sequence homology to the α- and β-subunits of POR from H. volcanii. POR from anaerobic archaea, e.g., P. furiosus and Methanosarcina barkeri, and from the bacterium Thermotoga maritima constitute heterotetrameric proteins composed of four different subunits with molecular masses of 47, 30, 24, and 15 kDa (46–48). However, PORs from many anaerobic bacteria, including Klebsiella pneumonia and Clostridium acetobutylicum, are 240-kDa homodimeric enzymes composed of 120-kDa subunits (49, 50).

PEP carboxylase—the anaplerotic enzyme during growth on glucose.

Acetyl-CoA formed by POR is further oxidized to two molecules of CO₂ via reactions of the citric acid cycle (51). Anaplerotic reactions, i.e., PEP carboxylase or pyruvate carboxylase, during growth on glucose to replenish the cycle have not been characterized in H. volcanii. In the genome of H. volcanii, the gene HVO_2621 is annotated as ppc to encode a putative PEP carboxylase (PPC) with a calculated molecular mass of 102.3 kDa. Transcription of ppc was analyzed by Northern blotting with cells grown on glucose or acetate. A specific transcript of about 2,650 nucleotides which corresponds to the size of the ppc gene (2,694 nucleotides) was detected in glucose-grown cells. This signal was absent in acetate-grown cells indicating glucose-specific induction of ppc expression (Fig. 5).

FIG 5.

Transcriptional analysis of ppc encoding phosphoenolpyruvate carboxylase from H. volcanii. The gene ppc (2,694 nucleotides) is depicted as a gray arrow, with flanking genes as white arrows (A). Northern blotting was performed with RNA from glucose (Glc)- and acetate (Ac)-grown cells using a probe against ppc. 16S rRNA served as loading control (B).

The gene ppc was overexpressed, and the recombinant enzyme was purified by Ni-NTA affinity and size exclusion chromatography. PPC showed on SDS-PAGE a single band of 115 kDa (Fig. S1); by gel filtration, a molecular mass of 140 kDa was determined, indicating PPC to be a monomeric protein. The recombinant enzyme catalyzed the bicarbonate-dependent conversion of PEP to oxaloacetate and phosphate. The activity requires the presence of acetyl-CoA, which is known to be an allosteric activator of PPC from bacteria. The apparent K_m for acetyl-CoA was 0.195 mM (Fig. 6). PPC activity in dependence of PEP concentration, in the presence of 1 mM acetyl-CoA, followed Michaelis-Menten kinetics with an apparent V_max of 2.4 U/mg and a K_m for PEP of 0.5 mM. At 0.025 mM acetyl-CoA, significantly lower PPC activity was detected (Fig. 6). The addition of fructose-1,6-bisphosphate and glucose-6-phosphate (up to 50 mM), which have been shown to act as allosteric effectors of PPC from E. coli and maize, did not affect PPC activity of H. volcanii. PPC was strictly dependent on divalent cations, with Mn²⁺ being most effective (2.2 U/mg at 5 mM MnCl₂). The activity with MgCl₂ (5 mM) was less than 5% compared with MnCl₂. PPC activity was inhibited by aspartate, an allosteric inhibitor of PPC from bacteria, archaea, and plants. Inhibition was dependent on the concentrations of acetyl-CoA and PEP. e.g., at 0.1 mM acetyl-CoA and 1 mM PEP the addition of 1 mM aspartate caused an inhibition of activity by 85%.

FIG 6.

Dependence of phosphoenolpyruvate carboxylase activity from H. volcanii on acetyl-CoA concentration in the presence of 2.5 mM PEP (A) and on PEP concentration (B) in the presence of 1 mM acetyl-CoA (circles) or 0.025 mM acetyl-CoA (squares). Data are fitted to the Michaelis-Menten equation.

To demonstrate an essential anaplerotic role of PPC, we generated a deletion mutant of ppc. The mutant did not grow on glucose; wild-type growth was recovered by complementation with ppc. Growth of the ppc mutant on acetate was not affected (Fig. 7). These data indicate that PPC is the essential anaplerotic reaction during growth on glucose but not on acetate. It has been shown previously that during growth on acetate, H. volcanii uses the glyoxylic acid cycle as anaplerotic sequence (12). Furthermore, the ppc mutant did not grow on pyruvate (Fig. 7), indicating that PPC is the essential anaplerotic reaction also on pyruvate as growth substrate, which is converted to PEP, the substrate of PPC, via PEP synthetase (12).

FIG 7.

Growth of deletion mutant of ppc encoding phosphoenolpyruvate carboxylase from H. volcanii. Growth was performed on 15 mM glucose (A), on 40 mM pyruvate (B), and on 40 mM acetate (C). Square, wild type; circle, deletion mutant; triangle, deletion mutant complemented with ppc.

PEP carboxylases have been characterized from various bacteria, eukarya, and from a few archaea. Bacterial and eukaryal PPCs are usually homotetramers of about 100-kDa subunits and are allosterically activated in bacteria by acetyl-CoA (52, 53). The PPC from H. volcanii is composed of a 100-kDa subunit and was activated by acetyl-CoA, i.e., showed features typical for bacterial PPCs, but it differs from bacterial PPCs, being a monomeric protein. In the archaeal domain PPCs have been characterized from S. solfataricus, Sulfolobus acidocaldarius, Methanothermobacter thermautotrophicus, and Methanothermus sociabilis. These archaeal PPCs are homotetramers composed of small subunits (55 to 60 kDa) and are not activated by acetyl-CoA as an allosteric effector. However, these enzymes showed allosteric inhibition by aspartate (54–57).

PPC from H. volcanii showed the highest sequence identity (64% to 71%) to putative homologs from the haloarchaea Haloarcula marismortui, Halorubrum lacusprofundi, and Halorhabdus utahensis. All these putative haloarchaeal PPC homologs are composed of large, 100-kDa subunits. High sequence identity was also found with characterized bacterial PPCs, e.g., from E. coli (30%) and Rhodopseudomonas palustris (32%), and PPCs from eukarya, e.g., Zea mays (26%) (58–60). A significantly lower sequence identity (<10%) of H. volcanii PPC was found with characterized and putative PPCs from archaea. A sequence identity of less than 10% was also found with PPCs from a few Gram-positive bacteria, e.g., Clostridium perfringens (61), and two putative PPCs from haloarchaea, i.e., Halobacterium salinarum and Haloarcula vallismortis.

An amino acid sequence alignment of PPC from H. volcanii and from E. coli is shown in Fig. 8. The sequences contain the two consensus motifs typical for PPCs. The predicted secondary structure of PPC from H. volcanii matches well to the secondary elements of the PPC structure from E. coli (Fig. 8) (52). Also, the amino acids shown to be involved in the binding of Mn²⁺, PEP (62), and of the allosteric inhibitor aspartate (52) are completely conserved in the PPC form H. volcanii.

FIG 8.

Amino acid sequence alignment of phosphoenolpyruvate carboxylase from H. volcanii and from E. coli. The alignment was calculated with ClustalX (83). Shading in red indicates degree of sequence conservation. Structural-based secondary structure elements of PPC from E. coli and the predicted secondary structure (84) of PPC from H. volcanii are displayed using ESPript 3.0 (85). The residues involved in aspartate binding are marked by asterisks, residues involved in PEP and Mn²⁺ binding are marked by green circles and black triangles, respectively. The signatures I and II of PPC are underlined in red (Prosite entry: PDOC00330).

Phylogenetic analysis of characterized and putative PPC sequences from archaea, bacteria, and eukarya indicates that PPCs form two distinct clusters (Fig. 9), a bacterial type and an archaeal type, as has been previously reported (54, 55). One cluster comprises PPCs from bacteria and eukarya and also contains the PPCs from H. volcanii and related homologs, which are all composed of about 100-kDa subunits. The second cluster comprises small subunit PPCs from archaea and also includes three PPCs from Gram-positive bacteria and the two PPCs from haloarchaea, i.e., H. salinarum and H. vallismortis. The attribution of haloarchaeal PPCs to both the bacterial/eukaryal and the archaeal cluster suggests a different phylogenetic evolutionary history of haloarchaeal PPCs. The close relationship of H. volcanii PPC and homologs with bacterial PPCs indicates a lateral gene transfer event from bacteria to these haloarchaea. Multiple horizontal gene transfers in evolution within the PPC family have previously reported (63).

FIG 9.

Phylogenetic relationship of phosphoenolpyruvate carboxylases from H. volcanii and haloarchaeal homologs, from bacteria, eukarya, and archaea. The tree is based upon a multiple amino acid sequence alignment that was generated with ClustalX (83). Numbers at the nodes are bootstrapping values according to neighbor joining. Organism names and their associated UniProt accession numbers or PDB numbers are as follows: Bacteria/Eukarya: Corynebacterium glutamicum, P12880; Ralstonia solanacearum, Q8XWW2; Synechococcus sp. PCC7002, B1XMB9; Halobacillus halophilus, I0JJ14; E. coli, 1QB4; Rhodopseudomonas palustris, O32483; Plasmodium falciparum, Q8ILJ7; Zea mays, 1JQO; Arabidopsis thaliana, Q9MAH0; Flaveria trinervia, Q9FV66; Haloferax volcanii, D4GUG0; Halorubrum lacusprofundi, B9LS13; Haloarcula marismortui, Q5V4H5; Halorhabdus utahensis, C7NNW9 (haloarchaea are highlighted by dark gray background). Archaea: Halobacterium salinarum, B0R7F9; Haloarcula vallismortis, M0JCI7 (haloarchaea are highlighted by light gray background); Clostridium perfringens, Q8XLE8; Leuconostoc mesenteroides, Q03VI7; Oenococcus oeni, Q04D35; Methanopyrus kandleri, Q8TYV1; Sulfolobus acidocaldarius, Q4JCJ1; Sulfolobus solfataricus, Q97WG4; Methanothermobacter thermautotrophicus, O27026; Methanothermus fervidus, E3GXT0.

ACD—the key enzyme of acetate overflow from glucose in H. volcanii.

During growth on 25 mM glucose and 0.5% yeast extract, H. volcanii excretes—in the exponential-growth phase—about 6 mM acetate which is completely consumed in the stationary phase (Fig. 10). The enzyme functionally involved in acetate formation from acetyl-CoA was characterized as ADP-forming acetyl-CoA synthetase (ACD), encoded by the gene acd (HVO_1000). Transcriptional analyses, growth studies with a deletion mutant, and biochemical characterization of the recombinant ACD are reported.

FIG 10.

Acetate switch during growth on glucose in the wild type, acd mutant, and quadruple acs mutant from H. volcanii. (A) Growth of the wild type on 25 mM glucose (gray circles), consumption of glucose (open circles), and acetate formation (black circles); acetate formation in the Δacd mutant (triangles) and in the Δacd mutant that was complemented with acd in trans (squares). (B) Growth of the quadruple acs mutant on 25 mM glucose (gray triangles); acetate formed was not consumed in this acs mutant (black triangles) compared with acetate formed that was consumed in the stationary phase in the wild type (circles).

The gene ptaN (HVO_0999) is located adjacent to the acd gene and encodes a putative N-terminal domain of phosphotransacetylase. Northern blot analysis with a probe against the acd gene revealed a transcript signal at about 3,200 nucleotides in glucose-grown cells, which corresponds to the sum of the lengths of acd and ptnA, indicating cotranscription of both genes (Fig. 11). In addition to the signal at 3,200 nucleotides, a signal at 1,200 nucleotides was detected with a probe against ptaN, suggesting that ptaN is part of the operon and is also transcribed as a single transcript (Fig. 11). The additional transcription of ptaN as a single transcript is in accordance with the identification of potential basal promoter elements, i.e., a TATA box that centered at −27 and a “WW” motif at −10/−11 (W stands for A or T) (64), in the intergenic region (of 26 nucleotides) of both genes. Significantly weaker transcript signals were detected with RNA from cells grown on acetate, indicating glucose-specific upregulation of these two genes.

FIG 11.

Transcriptional analyses of the acd operon from H. volcanii. Genomic organization of acd (HVO_1000; 2,094 nucleotides) and ptaN (HVO_0999; 1,137 nucleotides) encoding ADP-forming acetyl-CoA synthetase and a putative N-terminal domain of phosphotransacetylase, respectively. The genes acd and ptaN are given as gray arrows; the flanking genes HVO_1001 (261 nucleotides) and HVO_0998 (189 nucleotides) are given as white arrows. (A). Northern blotting of the operon was performed with RNA from glucose (Glc)- and acetate (Ac)-grown cells using probes against acd (B) and ptaN (C). 16S rRNA served as loading control.

Characterization of ACD.

ACD was overexpressed in H. volcanii H1209, and the recombinant protein was purified by Ni-NTA affinity and size exclusion chromatography. SDS-PAGE revealed one single band at 79 kDa (calculated 74.5 kDa) (Fig. S1). The molecular mass of the native protein was 163 kDa, indicating a homodimeric structure. ACD catalyzed the reversible conversion of acetyl-CoA to acetate (acetyl-CoA + ADP + P_i ⇌ acetate + ATP + CoA) with apparent V_max values of 59 U/mg in the direction of acetate formation and of 44.9 U/mg in the direction of acetyl-CoA formation. The apparent K_m values for all substrates measured in both directions are given in Table 1 and Fig. S2. The highest enzyme activity was observed at a pH of 7.5, at 0.5 M KCl, and at 5 mM MgCl₂. In addition to acetyl-CoA (100%; 35 U/mg at 0.1 mM) ACD also utilized, each at 0.1 mM, isobutyryl-CoA (48%) and phenylacetyl-CoA (58%) as the substrates. In the reverse direction, ACD catalyzed—in addition to acetate (100%; 41 U/mg, at 10 mM)—the activation of the short-chain acids propionic acid (113%) and butyric acid (116%), of the branched-chain acids isobutyric acid (64%) and isovaleric acid (82%) and of the arylic acid phenyl acetate (59%). Indolacetate and succinate were used at rates less than 1.5%.

TABLE 1.

Molecular and catalytic properties of ADP-forming acetyl-CoA synthetase from H. volcanii

Parameter	Substrate	Data
Apparent molecular mass of enzyme (kDa)
Native		163
Subunit		79
Calculated		74.5
Oligomeric structure		α2
Apparent Vmax value (U/mg) (direction of acetate formation)		59.1
Apparent Km value (mM)	Acetyl-CoA	0.051
	ADP	0.12
	Pi	2.0
Apparent Vmax value (U/mg) (direction of acetyl-CoA formation)		44.9
Apparent Km value (mM)	Acetate	2.0
	ATP	0.77
	CoA	0.3
pH optimum		7.5
KCl optimum (M)		0.5
MgCl2 optimum (mM)		5

To analyze the functional involvement of ACD in acetate overflow from glucose in H. volcanii, an acd deletion mutant was generated. Growth of the Δacd mutant on glucose was not affected (not shown); however, the amount of acetate formed in the Δacd mutant in the exponential-growth phase was reduced by about 70% compared with the wild type (Fig. 10). Wild-type levels of acetate formation were recovered by complementation of the mutant with acd. These results indicate that ACD is functionally involved in acetate formation during overflow metabolism of H. volcanii on glucose.

Molecular and kinetic properties of H. volcanii ACD compared with those of ACDI isoenzymes.

The ACD from H. volcanii showed high sequence identity (66%) to the characterized ACD from H. marismortui (65). Both haloarchaeal ACDs are homodimeric proteins of 75 kDa subunits which show high sequence similarity to ACDs from hyperthermophilic archaea, e.g., Pyrococcus furiosus and Pyrobaculum aerophilum, that are 140-kDa heterotetrameric, α₂β₂ enzymes (65–67). Thus, the haloarchaeal ACDs are fusions of the homologous α-subunits (47 kDa) and β-subunits (25 kDa) of hyperthermophilic archaea. The hyperthermophilic archaeon Archaeoglobus fulgidus contains both, ACD homologs composed of α₂β₂ subunits or dimers of fused α- and β-subunits (68). Furthermore, in contrast to the haloarchaeal ACDs, which each contain one ACD enzyme, the hyperthermophilic archaea P. furiosus, Thermococcus kodakarensis, and A. fulgidus contain several ACD isoenzymes of different metabolic function, e.g., in the degradation of sugars and amino acids (68–70). The ACD isoenzymes I and II of P. furiosus differ with respect to their substrate specificity as follows: the ACDI isoenzyme preferentially utilizes acetyl-CoA over aryl-CoA esters, such as phenylacetyl-CoA, and is primarily involved in sugar fermentation. In contrast, the ACDII isoenzyme preferentially catalyzes the conversion of aryl-CoA esters and is primarily implicated in the degradation of aromatic amino acids (66, 67).

The ACDs of H. volcanii and of H. marismortui predominantly use acetyl-CoA, which classifies the haloarchaeal ACDs as ACDI isoenzymes involved in acetate formation from glucose. The H. volcanii ACD—in contrast to H. marismortui ACD—showed significant activity with phenylacetyl-CoA, suggesting an additional role in aromatic amino acid degradation.

ACS enzymes are involved in activation of excreted acetate.

In a previous publication, we studied the growth of H. volcanii on acetate and identified four paralogous AMP-forming acetyl-CoA synthetases (ACSs) that are involved in the activation of acetate to acetyl-CoA. This was concluded by showing that growth of a quadruple mutant of ACS1, ACS2, ACS7, and ACS9 on acetate was completely prevented (12). Here, we tested the growth of this quadruple ACS mutant on glucose to analyze its effect on the consumption of the excreted acetate in the stationary phase. This mutant did not affect the growth kinetics and the formation of acetate in the exponential-growth phase; however, the consumption of acetate in the stationary-growth phase was completely prevented (Fig. 10).

Together, the data show that the excretion of acetate is catalyzed by ACD, whereas the reutilization involves ACS enzymes. These data are in accordance with enzyme measurements reported for H. marismortui (15).

Conclusions.

In the present communication, we report a comprehensive description of enzymes of the semiphosphorylative Entner-Doudoroff (spED) pathway in the model haloarchaeon Haloferax volcanii. Following previous analyses, we analyzed here the enzymes of the lower part of the spED pathway, including phosphoglycerate mutase (PGM), enolase, and pyruvate kinase. Furthermore, PEP carboxylase (PPC) as an anaplerotic enzyme and pyruvate-ferredoxin oxidoreductase catalyzing the acetyl-CoA formation from pyruvate were characterized (Fig. 12). Phylogenetic analyses of several enzymes of the spED pathway, including iPGM and PPC, and as shown in our previous reports, 2-keto-3-deoxy-6-phosphogluconate aldolase, glycerinaldehyde-3-phosphate dehydrogenase, and pyruvate kinase, indicate that H. volcanii acquired these enzymes via lateral gene transfer events from bacteria (7, 13, 35). These findings are in accordance with a previous bioinformatic study by Martin and coworkers, proposing that the evolution of the heterotrophic lifestyle of haloarchaea, including sugar degradation pathways, was the result of massive lateral transfer of bacterial genes (71). Furthermore, we present genetic and enzymatic evidence of the key players of acetate switch during aerobic growth on glucose in archaea. The functional involvement of the enzymes in acetate formation and consumption was demonstrated by analyses with knockout mutants (Fig. 12). The enzyme catalyzing acetate overflow in H. volcanii was characterized as an ADP-forming acetyl-CoA synthetase, an enzyme reported so far to be specific for acetate formation in the course of fermentation of sugars in anaerobic archaea. By demonstrating a function of ACD in acetate formation in aerobic glucose metabolism of haloarchaea, we extend the metabolic role of ACD in archaeal metabolism. Furthermore, acetate consumption as part of the acetate switch mechanism involves AMP-forming acetyl-CoA synthetases. Together, the data expand our understanding of glucose degradation and of the acetate switch in the domain of archaea.

FIG 12.

Glucose degradation and acetate switch in H. volcanii. Enzymes shown to be functionally involved in glucose degradation via the semiphosphorylative Entner-Doudoroff pathway, anaplerosis, acetyl-CoA formation from pyruvate, and in acetate formation and consumption are highlighted by gray background. Enzymes that were characterized in previous studies are indicated by a frame (7, 13, 39). Locus tags are given in parentheses. KDG, 2-keto-3-deoxygluconate; KDPG, 2-keto-3-deoxy-6-phosphogluconate; GAP, glyceraldehyde-3-phosphate; 1,3BPG, 1,3-bisphosphoglycerate; 3PG, 3-phosphoglycerate; 2PG, 2-phosphoglycerate; PEP, phosphoenolpyruvate; GDH, glucose dehydrogenase (HVO_1083); GAD, gluconate dehydratase (HVO_1488); KDGK, KDG kinase (HVO_0549); KDPGA, KDPG aldolase (HVO_0950); GAPDH, GAP dehydrogenase (HVO_0481); PGK, phosphoglycerate kinase (HVO_0480); iPGM, cofactor-independent phosphoglycerate mutase (HVO_2516); ENO, enolase (HVO_2774); PK, pyruvate kinase (HVO_0806); POR, pyruvate-ferredoxin oxidoreductase (HVO_1305 and HVO_1304); ACS, AMP-forming acetyl-CoA synthetase (ACS1, HVO_0894; ACS2, HVO_0896; ACS7, HVO_A0156; ACS9, HVO_A0551); ACD, ADP-forming acetyl-CoA synthetase (HVO_1000); PPC, PEP carboxylase (HVO_2621); TCA, tricarboxylic acid cycle.

MATERIALS AND METHODS

Growth experiments.

H. volcanii H26 and deletion mutants grew aerobically at 42°C in 100-ml Erlenmeyer flasks containing 20 ml of synthetic medium (7, 72) supplemented with glucose (15 or 25 mM), acetate (40 mM), or pyruvate (40 mM). In growth experiments analyzing the acetate switch, the medium contained 25 mM glucose and 0.5% yeast extract. Due to the chromosomal deletion of the orotate phosphoribosyltransferase gene (pyrE2) in strain H26, the medium was supplemented with uracil (50 μg/ml). For complementation experiments, deletion mutants were transformed with the pyrE2-containing plasmid pTA963 that carried the respective target gene, and growth was performed without the addition of uracil. Expression of target genes was induced by tryptophan (80 to 200 μM). Growth was followed by measuring the optical density at 578 or 600 nm. Each growth experiment was reproduced independently. Glucose and acetate were determined enzymatically as described (73, 74).

Generation of deletion mutants and of a conditional lethal mutant of porAB.

Markerless gene deletion mutants of HVO_0806, HVO_2621, and HVO_1000 were constructed using the pop-in/pop-out strategy (8, 75). In each case, the up- and downstream regions of the target gene (each of about 500 nucleotides) were amplified by two PCRs (see Table S4 in the supplemental material). Then, the two fragments were fused by PCR, generating a deletion product that was ligated into the vector pTA131. These plasmids carrying the pyrE2 gene for selection in H. volcanii H26 were each multiplied in Escherichia coli XL1 Blue MRF’ followed by transformation of H. volcanii H26. Cells that have integrated the respective plasmid into the genome via homologous recombination (pop-in) were selected and cultivated in uracil-free medium with 1% Casamino Acids. Pop-out clones were generated by plating pop-in cells on medium containing both uracil (30 μg/ml) and 5-fluoroorotic acid (50 μg/ml), followed by passaging the pop-out cells in liquid medium. 5-Fluoroorotic acid is toxic for cells containing the pyrE2 gene, and thus, only cells that had experienced a second homologous recombination event (pop-out) were able to grow. Pop-out clones that were either wild type or the deletion variant of the indicated gene were screened for successful deletion by PCR followed by Southern blotting.

For complementation experiments of the mutants ΔHVO_0806, ΔHVO_2621, and ΔHVO_1000, the strains were transformed with the plasmids also used for overexpression (see below) containing the respective wild-type genes under the control of a tryptophanase (ptnaA) promoter. Expression of target genes was induced by the addition of tryptophan up to 200 μM.

A conditional lethal mutant of the porAB operon was generated. In this mutant, chromosomal expression of the porAB operon was controlled by the inducible ptnaA promoter, showing porAB expression only in the presence of the inductor tryptophan (76, 77). Therefore, the ptnaA promoter was fused to the first 195 nucleotides of porA by PCR (Table S4), followed by ligating this fusion product into the suicide vector pTA131 that carries the pyrE2 gene for selection in H. volcanii H26. H. volcanii H26 was transformed with the plasmid that integrated in the genome via homologous recombination at the porA site (homologous region at the first 195 nucleotides), resulting in a strain in which ptnaA controls the genome-encoded porAB operon. After confirmation of the conditional lethal mutation by PCR and Southern hybridization, growth experiments were performed in the presence (80 μM) and absence of tryptophan.

Homologous overexpression and purification of His-tagged enzymes.

Plasmids for homologous overexpression of target genes were constructed as described (78). Genes were amplified from genomic DNA of H. volcanii and were ligated into the plasmid pTA963. The porAB operon was amplified and as one fragment ligated into pTA963 (see Table S2 in the supplemental material). By this cloning strategy, the genes were fused to 6× CAC at the 5′ end encoding 6 histidine residues. Plasmids were multiplied in E. coli XL1-Blue MRF’ and used for transformation of H. volcanii H1209. For overexpression of His-tagged fusion proteins, the transformed H1209 cells grew in complex medium (8, 78) up to an optical density at 600 nm of about 0.5, and expression was induced by the addition of 2 mM tryptophan. After 10 to 15 h of further growth, cells were harvested by centrifugation. For purification of protein, cell pellets were suspended in 50 mM Tris-HCl (pH 8.2) containing 1.5 M KCl and 5 mM imidazole and disrupted using a French pressure cell, followed by centrifugation. The supernatant was applied onto a Ni-NTA affinity column, and specific elution of the protein was performed with 100 to 250 mM imidazole. This protein solution was applied onto a Superdex 200 HiLoad 16/60 column (GE Healthcare), and elution was performed by an isocratic flow in 50 mM Tris-HCl (pH 7.5) containing 1.5 M KCl. Purification of PEP carboxylase was performed by the same procedure in the presence of 5 mM dithiothreitol (DTT).

Characterization of recombinant, His-tagged enzymes.

Enzyme activities were measured at 42°C. Coupling enzymes were purchased from Sigma-Aldrich, Roche, or Megazyme. It was ensured that the coupling enzymes were not rate limiting. Kinetic data were fitted to the Michaelis-Menten equation with Origin2015 software.

Phosphoglycerate mutase (PGM) activity (3-phosphoglycerate ⇌ 2-phosphoglycerate) was measured in both directions. (i) In the direction of 2-phosphoglycerate (2PG) formation, the assay mixture (1 ml) contained 50 mM Tris-HCl (pH 7.5), 1.5 M KCl, 2 mM ADP, 5 mM MgCl₂, 0.3 mM NADH, 8 mM 3-phosphoglycerate (3PG), 0.5 U enolase, 2 U pyruvate kinase, 5.5 U lactate dehydrogenase, and up to 1.9 μg protein. (ii) In the direction of 3PG formation, the assay mixture (1 ml) contained 50 mM Tris-HCl (pH 7.5), 1.5 M KCl, 5 mM MgCl₂, 2.5 mM ATP, 2.5 mM PEP, 0.3 mM NADH, 5 mM 2PG, 3.9 U phosphoglycerate kinase, 2 U pyruvate kinase, 5.5 U lactate dehydrogenase, and up to 1.9 μg protein.

Enolase activity (2PG ⇌ phosphoenolpyruvate) was measured in both directions as either phosphoenolpyruvate (PEP) formation or consumption at 240 nm using a molar extinction coefficient for PEP of 0.96 mM⁻¹cm⁻¹. The molar extinction coefficient was calculated from a standard curve with PEP up to a concentration of 1.75 mM. The assay mixtures (1 ml) contained 50 mM Tris-HCl (pH 7.5), 1.5 M KCl, 5 mM MgCl₂, 5 mM 2PG or 1.6 mM PEP, and up to 3.2 μg protein. The K_m value for MgCl₂ was determined in the direction of PEP formation with 2.5 mM 2PG.

Pyruvate-ferredoxin oxidoreductase activity (pyruvate + CoA + 2 MV_ox → acetyl-CoA + 2 MV_red + CO₂) was measured under anaerobic conditions in stoppered glass cuvettes (volume 1 ml) under a N₂ gasphase by following the pyruvate- and CoA-dependent reduction of methyl viologen (MV) using a molar extinction coefficient for MV at 578 nm of 9.8 mM⁻¹cm⁻¹ (79). The assay mixture contained 0.1 M Tris-HCl (pH 7.0), 2 M KCl, 5 mM MV, 10 mM pyruvate, 0.3 mM CoA, and 4.6 μg protein. To slightly reduce the assay mixture,1 to 2 μl of a 100 mM sodium-dithionite solution was added.

PEP carboxylase activity (PEP + HCO₃⁻ → oxaloacetate + P_i) was measured in the direction of oxaloacetate formation by coupling the reaction to the oxidation of NADH with malate dehydrogenase. The assay mixture (0.3 ml) contained 100 mM Tris-HCl (pH 7.5), 1.5 M KCl, 5 mM MnCl₂, 10 mM NaHCO₃, 0.3 mM NADH and 2.5 mM PEP, 4 U malate dehydrogenase, 0 to 1 mM acetyl-CoA, and up to 12.3 μg protein. Glucose-6-phosphate and fructose-1,6-bisphosphate as allosteric activators were tested at concentrations up to 50 mM. Aspartate as an allosteric inhibitor was tested at 1 mM in the presence of 0.1 mM acetyl-CoA and 1 mM PEP. The reaction was started by the addition of acetyl-CoA.

ADP-forming acetyl-CoA synthetase (ACD) activity (acetyl-CoA + ADP + P_i ⇌ acetate + ATP + CoA) was measured in both directions as described elsewhere (80) using two assay systems in 200 μl. (i) The P_i- and ADP-dependent CoA release from acetyl-CoA was monitored with Ellman’s thiolreagent (5,5′-dithiobis-2-nitrobenzoic acid; DTNB) by measuring the formation of thiophenolate anion at 412 nm (ε = 13.6 mM⁻¹ cm⁻¹) (81). The assay mixture contained 100 mM HEPES-KOH (pH 7.5), 0.5 M KCl, 5 mM MgCl₂, 0.1 mM DTNB, 0.1 mM acetyl-CoA, 1 mM ADP, 10 mM KH₂PO₄, and up to 0.17 μg protein. The pH dependence was measured between pH 5.5 to 8.0 using piperazine (pH 5.5 to 6.0), MES (pH 6.0 to 7.0), and HEPES-KOH (pH 7.0 to 8.0), at 100 mM each. The salt dependence was measured up to 3 M KCl and 25 mM MgCl₂. The substrate specificity of ACD for CoA esters was measured by replacing acetyl-CoA with 0.1 mM phenylacetyl-CoA or isobutyryl-CoA. (ii) The CoA- and acetate-dependent ADP formation from ATP was measured by coupling the reaction with the oxidation of NADH (ε = 6.2 mM⁻¹ cm⁻¹) at 340 nm via pyruvate kinase and lactate dehydrogenase. The assay mixture contained 100 mM HEPES-KOH (pH 7.5), 0.5 M KCl, 5 mM MgCl₂, 10 mM acetate, 2 mM ATP, 1 mM CoA, 2.5 mM PEP, 0.3 mM NADH, 0.8 U lactate dehydrogenase, 1.2 U pyruvate kinase, and up to 0.17 μg protein. The specificity of ACD for acids was measured by replacing acetate with 10 mM each of propionate, butyrate, isobutyrate, isovalerate, phenylacetate, indole-3-acetate, or succinate.

Analytical assays.

Determination of the molecular masses of native enzymes was performed on a Superdex 200 HighLoad column (1.6 by 60 cm) with a flow rate of 1 ml/min using 50 mM Tris-HCl (pH 7.4), containing 1.5 mM KCl. The column was calibrated with the high-molecular-weight (HMW) and low-molecular-weight (LMW) kits (GE Healthcare) as specified by the manufacturer. The purity of proteins was analyzed by SDS-PAGE. Protein concentration was determined by the method of Bradford with bovine serum albumin V as the standard.

Transcriptional analyses.

RNA from exponentially grown cells of H. volcanii (at optical densities of about 1.3) was isolated as described previously (82). Northern blot analyses were performed with 4 to 15 μg of RNA (6). Probes were amplified using the PCR DIG probe synthesis kit (Roche Diagnostics, Mannheim, Germany) (for primer sequences, see Table S3 in the supplemental material). Transcript sizes were determined with the RiboRuler high-range RNA ladder (Thermo Fisher Scientific, Schwerte, Germany).