The pathway of the oxidation of propionate to pyruvate in Escherichia coli involves five enzymes, only two of which, methylcitrate synthase and 2-methylisocitrate lyase, have been thoroughly characterized. Here we report that the isomerization of (2S,3S)-methylcitrate to (2R,3S)-2-methylisocitrate requires a novel enzyme, methylcitrate dehydratase (PrpD), and the well-known enzyme, aconitase (AcnB), of the tricarboxylic acid cycle. AcnB was purified as 2-methylaconitate hydratase from E. coli cells grown on propionate and identified by its N-terminus. The enzyme has an apparent Km of 210 µm for (2R,3S)-2-methylisocitrate but shows no activity with (2S,3S)-methylcitrate. On the other hand, PrpD is specific for (2S,3S)-methylcitrate (Km = 440 µm) and catalyses in addition only the hydration of cis-aconitate at a rate that is five times lower. The product of the dehydration of enzymatically synthesized (2S,3S)-methylcitrate was designated cis-2-methylaconitate because of its ability to form a cyclic anhydride at low pH. Hence, PrpD catalyses an unusual syn elimination, whereas the addition of water to cis-2-methylaconitate occurs in the usual anti manner. The different stereochemistries of the elimination and addition of water may be the reason for the requirement for the novel methylcitrate dehydratase (PrpD), the sequence of which seems not to be related to any other enzyme of known function. Northern-blot experiments showed expression of acnB under all conditions tested, whereas the RNA of enzymes of the prp operon (PrpE, a propionyl-CoA synthetase, and PrpD) was exclusively present during growth on propionate. 2D gel electrophoresis showed the production of all proteins encoded by the prp operon during growth on propionate as sole carbon and energy source, except PrpE, which seems to be replaced by acetyl-CoA synthetase. This is in good agreement with investigations on Salmonella enterica LT2, in which disruption of the prpE gene showed no visible phenotype.
Several bacteria and fungi are able to oxidize propionate via methylcitrate to pyruvate. Initially propionyl-CoA condenses with oxaloacetate to (2S,3S)-methylcitrate, which isomerizes to (2R,3S)-2-methylisocitrate. Cleavage leads to pyruvate and succinate. The consecutive oxidative regeneration of oxaloacetate from succinate completes the methylcitrate cycle. Initially this cycle was discovered by growing a mutant strain of the yeast Candida lipolytica on odd-chain fatty acids. The accumulation of a tricarboxylic acid was observed during growth and identified as methylcitrate . Further investigations revealed other enzymes necessary for a functional methylcitrate cycle. The enzymes, however, were only partially characterized and no genomic sequences were identified [2–6]. More recently it was discovered that propionate oxidation in aerobically growing Gram-negative bacteria, especially Escherichia coli and Salmonella enterica serovar Thyphimurium LT2 , also proceeds via methylcitrate. The purification of one of the key enzymes of the methylcitrate cycle, methylcitrate synthase, led to the identification of an operon necessary for propionate degradation. In E. coli and S. enterica this prp operon is composed of the genes prpB, prpC, prpD and prpE. PrpB and PrpC were identified as 2-methylisocitrate lyase  and methylcitrate synthase , respectively. PrpE was shown to catalyse the activation of propionyl-CoA . It remained unclear, however, by which mechanism the dehydration and rehydration of (2S,3S)-methylcitrate is performed to yield (2R,3S)-2-methylisocitrate. In S. enterica it was reported that the first reaction, the dehydration of methylcitrate, is catalysed by the PrpD protein . However, the product of this reaction was not further analysed. It was suggested that 2-methyl-cis-aconitate was formed. Interestingly, this reaction would involve the unusual syn elimination of water, whereas in all other analysed derivatives of malate this β-elimination occurs in an anti manner; for a review see . Aconitase from bovine heart follows this rule by dehydrating both substrates, citrate and (2R,3S)-isocitrate, in an anti manner. Furthermore the enzyme is able to hydrate 2-methyl-cis-aconitate to threo-2-methylisocitrate in an anti manner, but cannot use methylcitrate as substrate . Surprisingly, investigations of the PrpD protein showed that this enzyme is not able to catalyse the hydration of 2-methyl-cis-aconitate to 2-methylisocitrate. There is genetic evidence that an aconitase-like protein or even one of the aconitases (AcnA or AcnB) from S. enterica catalyse this hydration . Other studies on the PrpD protein of E. coli revealed the existence of an iron-sulfur cluster essential for catalytic activity . However, this is in disagreement with results on the S. enterica PrpD protein, in which such a cluster was not found . Therefore, the biochemical characterization of the E. coli PrpD protein will also focus on the activity of this enzyme in the presence of chelating agents such as EDTA and o-phenanthroline.
In this paper we report the in vitro reconstitution of the oxidation of propionyl-CoA to pyruvate by the use of purified PrpC, PrpD, AcnB and PrpB from E. coli. PrpD and AcnB involved in the conversion of methylcitrate into 2-methylisocitrate were biochemically characterized. Furthermore, expression of the genes involved in propionate metabolism was studied in 2D protein gel electrophoresis and Northern-blot experiments.
Bacteria and culture conditions
For purification of wild-type enzymes and for expression studies, the E. coli K12 derivative W3350 (F– gal r+m+λ sensitive) was used . For overexpression of the genes prpD and prpB, E. coli TOP10 cells (Invitrogen) were used, containing plasmids with the corresponding genes and an N-terminal cloned histidine tag. For purification of wild-type enzymes and expression studies, cells were grown aerobically at 37 °C in minimal medium containing 60 mm K2HPO4, 33 mm KH2PO4, 76 mm (NH4)2SO4, 2 mm trisodium citrate, 0.1% (v/v) trace element solution without chelating agent , 1 mm MgSO4, and 50 mm sodium propionate, sodium acetate or glucose. For overproduction of proteins, cells were grown in Standard I medium (peptone, 15.6 g·L−1; yeast extract, 2.8 g·L−1; 100 mm NaCl; 5 mm glucose; Merck, Darmstadt, Germany) and induced with isopropyl thio-β-d-galactoside. Cells were harvested by centrifugation at 10 000 g and used directly or stored at −80 °C.
Purification of 2-methylisocitrate dehydratase (AcnB) from E. coli W3350
For a standard purification, 18 g (wet weight) propionate-grown cells was used. All purification steps were carried out in an anaerobic chamber (95% N2, 5% H2). Cells were thawed on ice and suspended in 20 mL anaerobic buffer I (20 mm potassium phosphate, pH 7.5, 1 mm trisodium citrate and 1 mm dithiothreitol). Cells were broken by sonication (Branson sonifier; 3 × 5 min at 60% pulse and 80% of full power). Cell debris was removed by ultracentrifugation at 96 000 g for 45 min. This crude extract was filtered (0.45 µm pore size; Sarsted, Nümbrecht, Germany) and loaded on to a hydroxyapatite column (20 mL bed volume) equilibrated with buffer I. Unless otherwise indicated, the FPLC system and columns from Amersham Biosciences were used. The hydroxyapatite column was washed with buffer I. The flow through was concentrated with an Amicon chamber over a PM 30 size-exclusion filter (Millipore) and diluted in buffer II (20 mm Tris/HCl, pH 7.5, with 1 mm trisodium citrate and 1 mm dithiothreitol). The enzyme was loaded on to a Q-Sepharose column (30 mL bed volume), previously equilibrated with buffer II. The enzyme was eluted with buffer III (20 mm Tris/HCl, pH 7.5, with 1 mm trisodium citrate, 1 mm dithiothreitol and 1 m NaCl) with a linear NaCl gradient of 150 −200 mm. Active fractions were pooled, and solid (NH4)2SO4 was added to a final concentration of 1 m, filtered and loaded on to a phenyl-Sepharose column (bed volume 30 mL), previously equilibrated with buffer IV (20 mm Tris/citrate, pH 8.0, with 1 mm dithiothreitol and 1 m (NH4)2SO4). The enzyme was eluted with a linear (NH4)2SO4 gradient of 1.0–0 m in buffer V (20 mm Tris/citrate, pH 8.0, with 1 mm dithiothreitol) between 0.2 and 0 m (NH4)2SO4 and was concentrated as described above by changing to buffer II. The enzyme was loaded on to a UnoQ column (Bio-Rad; bed volume 6 mL) equilibrated with buffer II and eluted with buffer III. The purity of the eluted fractions was checked by electrophoresis on a 15% polyacrylamide gel in the presence of SDS.
Overproduction and purification of PrpB and PrpD with N-terminal histidine tags
The source of PrpB protein, the 2-methylisocitrate lyase, was described elsewhere . The prpD ORF from the prp operon of wild-type E. coli W3350 was amplified with Taq polymerase. Primers were constructed with the complete restriction sites of BamHI (primer: 5′-CGGGATCCTCAGCTCAAATCAACAACATCCGC-3′) and PstI (5′-AACTGCAGTTAAATGACGTACAGGTCGAGATAC-3′), respectively. After restriction of the PCR product with both enzymes, the product was cloned into the previously restricted pQE30 vector (Qiagen) for overexpression with an N-terminal His tag. Chemically competent E. coli cells (TOP10) were transformed with the plasmid. Overproduction of the PrpD protein was performed by growing the cells in Standard I medium until D578 = 0.8 and induction with 1 mm of isopropyl thio-β-d-galactoside followed by incubation overnight. Overproduction of methylcitrate dehydratase in four different clones was confirmed by SDS/PAGE. All clones exhibted an induced protein at 54 kDa (data not shown).
Cells from a 1.2-L culture (D578 ≈ 3) were induced for 10 h and harvested by centrifugation. Cells were washed with 50 mm potassium phosphate, pH 7.0, centrifuged, and suspended in the same buffer. Cells were broken by sonication and centrifuged at 96 000 g. The resulting cell-free extract was loaded on to a gravity flow Ni/nitrilotriacetic acid/agarose column with a bed volume of 5 mL. The column was washed with 20 mL 50 mm potassium phosphate, pH 7.0, containing 20 mm histidine to remove unspecifically bound proteins. PrpD was eluted with 50 mm potassium phosphate buffer, pH 7.0, containing 200 mm histidine. Active fractions were concentrated and desalted over a PM 30 size-exclusion filter. After addition of glycerol to a final concentration of 50% (v/v), the protein could be stored at −20 °C without loss of activity.
Enzymatic synthesis of (2S,3S)-methylcitrate
Methylcitrate was produced with the methylcitrate synthases PrpC from E. coli or McsA from the filamentous fungus Aspergillus nidulans. The reaction was carried out at room temperature for 20 h. Propionyl phosphate was synthesized chemically by a modified synthesis described by Stadtman , in which acetic acid anhydride was replaced by propionic acid anhydride. Propionyl phosphate was converted into propionyl-CoA with the help of phosphotransacetylase from Bacillus stearothermophilus (Sigma, Taufkirchen, Germany). A typical reaction for the synthesis of methylcitrate was carried out in a final volume of 60 mL and contained 50 mm propionyl phosphate, 100 mm oxaloacetic acid (neutralized with KHCO3), 0.2 mm CoASH, 500 U phosphotransacetylase and 50 U methylcitrate synthase. The reaction was buffered at pH 7.5 in 20 mm potassium phosphate. After incubation, the enzymes were denatured by heat treatment for 20 min at 80 °C and centrifuged at 10 000 g for 10 min. The supernatant was concentrated to a final volume of 10 mL in a rotary evaporator. Precipitated salts were removed by centrifugation as described above, and the supernatant was loaded on to a Dowex 1x8 column (Cl– form, bed volume 10 mL). Methylcitrate was eluted with 1 m HCl. The methylcitrate-containing fractions, as tested enzymatically with the PrpD protein, were concentrated by evaporation. The residual brownish oil was checked for purity by 1H-NMR (500 MHz, CDCl3): δ = 1.19 (3H, d, 3J = 6.9 Hz CH3), 2.90 (1H, q, 3J = 6.9 Hz, CH), 2.90 (1H, d, 2J = 16.6 Hz, CHH), 3.17 (1H, d, 2J = 16.6 Hz, CHH). Both, the E. coli and the A. nidulans enzyme produced the same enantiomeric pure (2S,3S)-methylcitrate (99.9%) as checked by enantioselective multidimensional capillary gas chromatography (kindly performed by Professor A. Mosandl, Universität Frankfurt/Main, Germany).
2-Methylisocitrate lyase (PrpB) was assayed with the coupled NADH-dependent assay as described previously . Methylcitrate dehydratase (PrpD) activity was measured at 240 nm with a Kontron, model Uvikon 943 double-beam UV/visible spectrophotometer, and the formation of the double bond during dehydration of methylcitrate was monitored. The absorption coefficient, ε240, was taken as 4.5 mm−1·cm−1. The composition of the assay mixture was 50 mm potassium phosphate, pH 7.5, and 1.3 mm methylcitrate in a final volume of 1 mL.
The racemic mixture of chemically synthesized threo-2-methylisocitrate  was used to follow the dehydration and the formation of the double bond in 2-methyl-cis-aconitate at 240 nm; ε240 = 4.5 mm−1·cm−1. The composition of the assay was 50 mm potassium phosphate, pH 7.5, and 0.3 mmthreo-2-methylisocitrate in a final volume of 1 mL. To measure 2-methylisocitrate dehydratase (AcnB), a coupled assay was performed in the reverse direction. The reaction was followed at 340 nm under anaerobic conditions with ε340 = 6.2 mm−1·cm−1. The composition of the assay mixture was 50 mm potassium phosphate buffer, pH 7.5, 2 mm MgCl2, 0.2 mm NADH, 0.64 mm methylcitrate, 0.2 U PrpD, 0.2 U PrpB, 0.3 mm dithiothreitol, 3 U lactate dehydrogenase from rabbit muscle (Roche) and a sample of purified AcnB in a final volume of 1 mL.
Gel electrophoresis and blotting of proteins
The protein fractions obtained from the purification of 2-methylisocitrate dehydratase were analysed by SDS/PAGE. The apparent molecular mass of the 2-methylisocitrate dehydratase subunit was determined by measuring the mobility by SDS/PAGE (15% acrylamide)  with standard proteins as molecular mass markers. Purified 2-methylisocitrate dehydratase was blotted from the gel (10% acrylamide) on a poly(vinylidene difluoride) membrane (Millipore) with the transblot SD semidry transfer cell (Bio-Rad), as described in the manufacturer's protocol, and was then N-terminally sequenced by Edman degradation (kindly performed by D. Linder, Universität Gießen, Germany).
Re-activation and inactivation of AcnB
AcnB was inactivated by exposure to air and by addition of either EDTA or o-phenanthroline (both 2.5 mm final concentration). For reactivation, 98.3 mg FeSO4 × (NH4)2SO4 × 6H2O (final concentration 5 mm) and 136 mg cysteine hydrochloride (monohydrate) (15 mm) were dissolved under anaerobic conditions in 45 mL water, and the pH was adjusted to 7.5 by dropwise addition of 1 m NaOH. Water was added to a final volume of 50 mL. One part of enzyme solution was mixed with one part of re-activation mixture and incubated for 60 min at room temperature under anaerobic conditions.
Iron–sulfur cluster and metal cofactors
Purified PrpD protein was concentrated to 4 mg protein·mL−1 in 20 mm Hepes buffer, pH 7.5, and the activity was measured with methylcitrate as substrate. An aliquot was diluted and a spectrum was determined in the range 220–900 nm. A second 0.5-mL aliquot was taken and incubated for 60 min at room temperature under anaerobic conditions in re-activation mixture (0.5 mL) as described above for the re-activation of the 2-methylisocitrate dehydratase. PrpD was separated from the re-activation mixture by the use of a Sephadex-NAP column (Pharmacia Biotech) and eluted in 20 mm Hepes buffer, pH 7.5. Activity was tested and a spectrum was determined as described above. A third and fourth aliquot were taken and incubated with a 5 mm final concentration of o-phenanthroline or 10 mm EDTA, respectively, and incubated for 20 min at room temperature. PrpD was desalted, and activity and a spectrum were determined as described above.
Synthesis of digoxygenin-labelled RNA probes
For the detection of mRNAs of the genes acs, acnB, prpD and prpE, specific RNA probes labelled with digoxygenin were produced using the T7 polymerase. Oligonucleotides were designed that contained the sequence of the T7 promoter in the reverse primer (the full sequences of all primers are shown in Table 1). A PCR was performed with Taq polymerase, and genomic DNA of E. coli W3350 was used as a template. PCR products were separated by electrophoresis in a 1% agarose gel and purified by the Geneclean Kit II (BIO 101) as described in the manufacturer's protocol. For in vitro transcription, 0.5–1.0 µg PCR product was mixed with 2 µL NTP labelling mixture containing UTP (Dig RNA Labelling Kit T7; Roche), 2 µL reaction buffer (Ambion), 1 µL RNase inhibitor (Dig RNA Labelling Kit T7; Roche), 2 µL T7 polymerase (20 U·µL−1; Ambion) and diethyl pyrocarbonate-treated water to a final volume of 20 µL. Transcription was carried out at 37 °C for 1 h. RNase-free DNase I was added, and the mixture was incubated at 37 °C for a further 15 min. RNA was precipitated by the addition of 2.5 µL 1 m LiCl and 90 µL 100% ethanol and incubated for 1 h at −80 °C. After centrifugation (12 000 g, 4 °C), RNA pellets were dried and dissolved in 100 µL nuclease-free water. The intensity of the digoxygenin label of the probes was checked by cross-linking the specific probes on a nylon membrane and detection of the label by standard methods.
Table 1. Oligonucleotides used for the generation of RNA probes. The reverse primer contains the promoter region for the T7 polymerase at the 5′ end. An asterisk denotes the end of the promoter region.
2D gel electrophoresis
After harvesting of the bacteria by centrifugation, cells were washed in 10 mm Tris/HCl (pH 7.5)/1 mm EDTA, and the cell pellet was suspended in the same buffer. Cells were disrupted by several passages through a French pressure cell, and debris was removed by centrifugation at 4 °C and 20 000 g for 30 min. The protein concentration of the supernatant fraction was assayed by the method of Bradford . For 2D gel electrophoresis, 400 µg crude protein extract was solubilized in a hydration solution containing 8 m urea, 2 m thiourea, 2% (w/v) 3-[(3-chloramidopropyl)dimethylammonio]propane-1-sulfonate (Chaps), 28 mm dithiothreitol, 1.3% (v/v) Pharmalytes, pH 3–10, and bromophenol blue. After hydration in the protein-containing solution for 24 h under low-viscosity paraffin oil, Immobiline DryStrips (IPG-strips; Amersham Biosciences) covering the pH range 4–7 or 3–10 were subjected to isoelectric focusing. The following voltage/time profile was used: a linear increase from 0 to 500 V for 1000 Vh, 500 V for 2000 Vh, a linear increase from 500 to 3500 V for 10 000 Vh and a final phase of 3500 V for 35 000 Vh (pH 4–7) or for 21 000 Vh (pH 3–10). IPG-strips were consecutively incubated for 15 min each in equilibration solution A and B. Solution A contained 50 mm Tris/HCl, pH 6.8, 6 m urea, 30% glycerol, 4% SDS and dithiothreitol (3.5 mg·mL−1). Solution B contained iodoacetamide (45 mg·mL−1) instead of dithiothreitol. In the second dimension, proteins were separated on SDS/12.5% polyacrylamide gels with the Investigator™ System (Perkin–Elmer Life Sciences, Cambridge, UK) at 2 W per gel. Gels were stained with PhastGel BlueR according to the manufacturer's (Amersham Biosciences) instructions. After scanning, the 2D PAGE images were analysed with the Melanie3® software package (Bio-Rad Laboratories GmbH). Three separate gels of each condition and two independent cultivations were analysed, and only spots displaying the same pattern in all parallels were selected for further characterization.
Protein identification by peptide mass fingerprinting
Protein spots were excised from PhastGel BlueR-stained 2D gels, destained, and digested with trypsin (Promega); peptides were then extracted . Peptide mixtures were purified with C18-tips according to the manufacturer's (Millipore) instructions and directly eluted on to a sample template of a MALDI-TOF mass spectrometer with an eluent containing 50% (v/v) acetonitrile, 0.1% (v/v) trifluoroacetic acid, saturating amounts of α-cyano-3-hydroxycinnamic acid and calibration peptides. Peptide masses were determined in the positive ion reflector mode in a Voyager DE RP mass spectrometer (Applied Biosystems) with internal calibration. Mass accuracy was better than 50 p.p.m. Peptide mass fingerprints were compared with databases using the mascot program (http://www.matrixscience.com/cgi/index.pl?page=../home.html). The searches considered oxidation of methionine, pyroglutamic acid formation at the N-terminal glutamine, and modification of cysteine by carbamidomethylation as well as partial cleavage leaving a maximum of one internal site uncleaved.
RNA isolation and Northern blot
E. coli W3350 cells were grown on propionate, acetate or glucose minimal medium to an D578 of 0.8 under vigorous shaking at 37 °C. The cultures (20 mL) were mixed with 20 mL frozen ‘killing buffer’ (20 mm Tris/HCl, pH 7.5, 5 mm MgCl2, 20 mm NaN3; diethyl pyrocarbonate treated) and centrifuged for 10 min at 4000 g. Cell pellets were suspended in 200 µL killing buffer, and frozen in liquid nitrogen. Cells were broken in a frozen state in a Micro-Dismembrator (Braun Biotech International) at 2600 r.p.m. for 2 min. Cell extracts were mixed with 4 mL lysis solution containing 4 m guanidine thiocyanate, 25 mm sodium acetate, pH 5.2, and 0.5%N-lauroylsarcosine (w/v) at 37 °C. One part of this solution was mixed with one part of acidic phenol/chloroform/3-methylbutan-1-ol (50 : 48 : 2, by vol.), shaken at room temperature for 5 min and centrifuged at 12 500 g for 5 min. The upper layer was mixed with 1 mL acidic phenol/chloroform/3-methylbutan-1-ol, shaken again for 5 min, and spun down as described above. The upper layer was mixed with 800 µL chloroform/3-methylbutan-1-ol (24 : 1, v/v), and centrifuged as described above. The aqueous phase was collected, and 80 µL 3 m sodium acetate (pH 5.2) and 1.1 mL propan-2-ol were added. RNA was precipitated by incubation at −80 °C for 1 h. The RNA was centrifuged (23 000 g, 4 °C, 15 min) and the pellet was washed with 70% ice-cold ethanol. The RNA was dried at room temperature and dissolved in 30 µL diethyl pyrocarbonate-treated water. Quality and quantity of isolated RNA was checked with the Agilent 2100 Bioanalyzer (Agilent Technologies, Böblingen, Germany) as described in the manufacturer's protocol.
RNA (8 µg in 4.5 µL) was mixed with 10.5 µL denaturation solution [90 µL formamide, 18 µL formaldehyde, 18 µL 10 × Mops (200 mm Mops, 50 mm sodium acetate, 10 mm EDTA, dissolved in diethyl pyrocarbonate-treated water and adjusted to pH 7.0)] and loaded on to a 1.4% (w/v) agarose gel containing 1.8 m formaldehyde. RNA was separated at 70 V for 3 h and transferred to a nylon transfer membrane (Schleicher and Schuell) . The RNA blot was saturated with blocking reagent and hybridized with the digoxygenin-labelled antisense RNA probes overnight. After a wash, specific hybridization signals were detected by incubation with alkaline phosphatase-conjugated anti-digoxygenin Ig (Roche) and monitoring the conversion of the ECF-vistra substrate with a STORM 860 fluorimager (Amersham Biosciences).
Determination of 2-methyl-cis-aconitate by anhydride formation
Enzymatically synthesized methylcitrate (0.8 mm) was dissolved in a final volume of 5 mL 20 mm Hepes, pH 7.5, and incubated with 0.5 U PrpD. A 1-mL aliquot was taken, and the reaction was monitored at 240 nm until the equilibrium of the reaction was reached. PrpD was inactivated by heating the whole sample for 15 min at 80 °C. Denatured protein was removed by centrifugation. Of this solution, 900 µL was mixed with 100 µL water, and a UV/visible spectrum in the range 220–400 nm was recorded. A second sample was prepared by using another 900 µL of the solution and addition of 100 µL 8 m HCl. The anhydride formation was followed at 259 nm until no further change in absorbance was observed. A second spectrum in the range 220–400 nm was recorded, and the difference spectrum between the neutral and the acidified sample was calculated using the Microsoft Excel worksheet. As a control, a methylcitrate solution without addition of PrpD was treated as described above. No change in absorbance was detectable during acidification.
Biochemical analysis of 2-methylisocitrate dehydratase
E. coli cells were grown to D578 = 1.2 in the presence of 50 mm propionate in the minimal medium. 2-Methylisocitrate dehydratase was identified in extracts of these cells by monitoring the decrease in A340 with enzymatically prepared (2S,3S)-methylcitrate as substrate and with PrpD, PrpB and lactate dehydrogenase as auxiliary enzymes (Fig. 1). Starting from 18 g wet cells, the protein was purified from a specific activity in crude extracts of 0.16 U·mg−1 to 1.9 U·mg−1 with a yield of 3.6%. Purification was performed by chromatography on hydroxyapatite, Q-Sepharose, phenyl-Sepharose and UnoQ (Table 2). A major band was revealed in the resulting protein fractions by SDS/PAGE (Fig. 2, lanes 6 and 7) with an apparent molecular mass of 94 kDa and a turnover number of 3 s−1. Comparison of the forward reaction (hydration of 2-methyl-cis-aconitate) in the coupled assay with the activity of the back reaction measured by the dehydration of chemically synthesized threo-2-methylisocitrate yielded a ratio of 1 : 0.7. The Km of the purified enzyme with threo-2-methylisocitrate as substrate was determined as 210 µm, which is somewhat higher than Km = 51 µm for (2R,3S)-isocitrate determined with the aconitase, AcnB . The enzyme showed no detectable activity with (2S,3S)-methylcitrate as substrate.
Table 2. Purification protocol for AcnB. A unit is defined as the oxidation of 1 µmol NADH·min−1 in the coupled assay.
Specific activity (U·mg−1)
UnoQ (Fr. 48)
UnoQ (Fr. 49)
N-Terminal sequence determination of the purified protein by Edman degradation revealed the peptide sequence, MLEEYXKXVAEXAAE, where X denotes unclear amino acids. Comparison of this sequence with the databases showed 100% identity with the N-terminal sequence of the E. coli citrate cycle aconitase, AcnB (swissprot P36683), with the sequence, MLEEYRKHVAERAAE. The calculated molecular mass of AcnB from its genomic sequence is 93 498 Da, which is in good agreement with the apparent molecular mass of 94 kDa derived from the SDS/PAGE analysis.
The enzyme was rapidly inactivated by exposure to air, which is already known for aconitases , as well as during the purification procedure, especially during chromatography on the phenyl-Sepharose column. Activity was partially restored by incubation in re-activation mixture under anaerobic conditions as described in Experimental procedures. Addition of EDTA or o-phenanthroline totally inactivated enzymatic activity. This is in good agreement with the requirement for a functional [4Fe−4S] cluster for aconitase activity.
Cloning and characterization of PrpD
The prpD gene was cloned and overexpressed as described in Experimental Procedures. The overproduced protein was purified to a specific activity of 11.4 U·(mg protein)−1 by chromatography on a Ni/nitrilotriacetate/agarose column (Fig. 3). PrpD showed maximum activity with enzymatically produced (2S,3S)-methylcitrate as substrate (Km = 440 µm). Another substrate was cis-aconitate, whereas citrate and (2R,3S)-isocitrate (natural occurring stereoisomer) showed no significant activity. Other related compounds such as trans-aconitate, threo-2-methylisocitrate and erythro-2-methylisocitrate, and (S)-malate and (R)-malate gave no activity at all (Table 3). Unfortunately, authentic 2-methyl-cis-aconitate was not available. The enzyme does not require any metal cofactors for full enzymatic activity. In UV/visible spectra, no extra band beside that at 280 nm could be seen. Neither the spectra nor the activity changed after incubation of PrpD with o-phenanthroline or with the re-activation mixture as described for aconitase.
Table 3. Substrate specificity of PrpD. No activity (< 0.01 U·mg−1) was found with threo-2-methylisocitrate and erythro-2-methylisocitrate, trans-aconitate, d-malate and l-malate, fumarate, maleate, d-tartrate and meso-tartrate, d-citramalate and l-citramalate, mesaconate, citraconate, itaconate, and (R,S)-3-methylitaconate.
The most likely product of the dehydration reaction of (2S,3S)-methylcitrate by PrpD is postulated to be 2-methyl-cis-aconitate. Acidification of the reaction mixture with HCl led to an increased A259 as shown in Fig. 4. This can be explained by the formation of a planar five-membered cyclic anhydride from 2-methyl-cis-aconitate under acidic conditions. This condensation would be less likely with 2-methyl-trans-aconitate, which would lead to a nonplanar six-membered cyclic anhydride. A comparable example of a five-membered cyclic anhydride formation between two carboxylic acid groups orientated in a cis conformation is observed in 2,3-dimethylmaleate, which is formed by a δ-isomerase reaction from (R)-3-methylitaconate (2-methylene-3-methylsuccinate) during the nicotinate fermentation by Eubacterium barkeri. Under acidic conditions, dimethyl maleate spontaneously forms the anhydride with a maximal absorbance at 256 nm . The formation of 2-methyl-cis-aconitate from (2S,3S)-methylcitrate is further supported by the substrate specificity of PrpD. cis-Aconitate is a moderate substrate, whereas no activity is detectable with trans-aconitate.
In vitro reconstitution of the methylcitrate cycle
For in vitro reconstitution, purified enzymes and enzymatically produced (2S,3S)-methylcitrate were used. (2S,3S)-Methylcitrate (1.3 mm) was converted into 2-methyl-cis-aconitate by the PrpD protein. 2-Methyl-cis-aconitate acted as substrate for AcnB and was hydrated to (2R,3S)-2-methylisocitrate. This product was cleaved by PrpB into succinate and pyruvate (Fig. 1). To monitor the reaction and to pull the equilibrium to the side of pyruvate formation, lactate dehydrogenase and NADH as cosubstrate were used. This coupled assay was also used to monitor the purification of the aconitase AcnB as described above. Absence of the aconitase or any other enzyme resulted in a loss of pyruvate formation. This result clearly demonstrates that both proteins, PrpD and AcnB, are essential for the conversion of methylcitrate into 2-methylisocitrate.
Northern-blot analysis of prpD, prpE, acnB and acs transcripts
The four genes were selected for the following reasons. Transcription levels of acs, the gene coding for acetyl-CoA synthetase (Acs), were used for comparison of the specificity of transcription during growth on acetate and propionate, respectively. Furthermore, this gene was of interest because of the ability of the Acs to activate propionate to the corresponding CoA ester. In S. enterica it was shown earlier that a strain carrying a deletion of the acs gene was still able to grow on propionate but not on acetate. A propionyl-CoA synthetase mutant was able to grow on propionate as well as on acetate. A double mutant with deletion of both genes did not grow on acetate or propionate . Therefore we postulated that transcripts of acs may be visible under both growth conditions, whereas the transcripts for propionyl-CoA synthetase, prpE, and methylcitrate dehydratase, prpD, should be exclusively formed during growth on propionate. In contrast, transcription of acnB coding for AcnB is expected to achieve similar levels under all conditions tested. AcnB is an essential enzyme of the citrate cycle, as well as of the glyoxylate cycle. During growth on propionate, pyruvate is formed, which is oxidized to acetyl-CoA, a substrate for the citrate and the glyoxylate cycle . Furthermore, AcnB acts as 2-methylisocitrate dehydratase and therefore comprises a twofold function during growth on propionate.
For the transcription experiments, RNA was purified from E. coli W3350 cells grown on glucose, acetate or propionate as sole carbon and energy source. Cells were harvested in the early exponential growth phase (D578 = 0.7–1.2) and broken as described in Experimental Procedures. Quality and quantity of the RNA used in each experiment was confirmed by the use of the Agilent 2100 Bioanalyzer. For each probe, the same quantity of RNA from cells grown on glucose, acetate or propionate was used (Fig. 5). Arrows denote main transcripts. Those with an additional asterisk denote larger transcripts, which may be formed by a read-through and can be observed from high gene expression or because of an alternative starting point of transcription. The first possibility may be correct for the acs gene (2.0 kb), which is not located in an operon, but may be transcribed together with the consecutive genes yjcH, yjcG and yjcF (5.6 kb). The second possibility may be correct for prpE and prpD, because the prp operon of E. coli, in contrast with that of S. enterica, is interrupted by a so-called repetitive extragenic palindromic element. This element is located between prpB and prpC(Fig. 6) and may be responsible for the two transcript sizes (4.6 and 5.9 kb), because these elements are suspected to be involved in transcriptional regulation . As expected, acnB is expressed under all conditions tested and shows a single transcript (Fig. 5). Transcripts of prpD and prpE are exclusively formed during growth on propionate. It can be concluded that acetate is not able to induce transcription of the specific genes involved in propionate catabolism. In contrast, a strong signal for the transcript of acs was observed on acetate as well as on propionate. This coincides with the investigations in S. enterica described above . The acs gene is able to replace prpE but not vice versa. Furthermore, propionate may be able to induce all genes of a functional glyoxylate cycle, because activity measurements for malate synthase of E. coli grown on propionate medium as compared with acetate showed specific activities of 0.50 U·mg−1 and 0.48 U·mg−1, respectively . The weak acs signal detected on glucose is in agreement with the observation of acetate excretion and consumption during growth on glucose medium .
2D gel electrophoresis
2D gel electrophoresis was carried out to monitor differences in the protein pattern of cells grown on acetate or propionate. E. coli W3350 cells were grown on propionate or acetate minimal medium and crude extracts were prepared from exponentially growing cells as described in Experimental procedures. Figure 7 exemplarily displays the protein profile of E. coli W3350 grown with either acetate or propionate as carbon source. Protein spots, which displayed significantly different intensities under the two growth conditions, were isolated from the gels and identified by peptide mass fingerprinting (Table 4). Proteins induced in the presence of propionate at a higher or lower level than in the presence of acetate are labelled with arrowheads and boxes, respectively (Fig. 7). PrpB, PrpC and PrpD encoded by the prp operon were exclusively produced during growth on propionate. PrpE, the propionyl-CoA synthetase, was detected on neither acetate nor propionate minimal medium. However, Acs seems to be present in high amounts, suggesting that it can also serve as a propionate-activating enzyme. Furthermore, increased levels of malate synthase (AceB) were found to be present during growth on propionate. Therefore, the main anaplerotic source of oxaloacetate appears to be the glyoxylate cycle rather than carboxylation of pyruvate or phosphoenolpyruvate as proposed previously . Six proteins, including phosphoglycerate mutase 1 (GpmA), a propanol-preferring alcohol dehydrogenase (AdhP), and pyruvate kinase (PykF) seemed to be present in reduced amounts in propionate-grown cells compared with cultures grown in the presence of acetate.
Table 4. Summary of propionate-induced proteins identified by peptide mass fingerprint matching (see also Figs 5 and 7). The theoretical isoelectric point and molecular mass were calculated with the compute pI/mw tool of the proteomics tools collection at the ExPASy Molecular Biology Server (http://www.expasy.ch/tools/pi_tool.html).
Molecular mass (kDa)
swissprot acc. no.
Sequence coverage (%)
1. Proteins induced at a higher level as compared with growth on acetate:
2. Proteins induced at a lower level as compared with growth on acetate:
Propanol-preferring alcohol dehydrogenase
Phosphoglycerate mutase 1
AcnB purified from E. coli W3350 cells grown on propionate as sole carbon and energy source was the only enzyme that displayed activity as a 2-methylisocitrate dehydratase. Similar results were obtained from S. enterica. AcnA and AcnB from this organism were overproduced, and enzymatic activity for the dehydration of 2-methylisocitrate was studied . This was in agreement with earlier investigations performed on horse and bovine heart aconitases, which both catalyse the reversible hydration of 2-methyl-cis-aconitate to 2-methylisocitrate, but not to methylcitrate [13,28]. The aconitase from E. coli (AcnB) completes the methylcitrate cycle. AcnB possesses a twofold function; it acts as 2-methylisocitrate dehydratase and a citrate/isocitrate isomerase in the citrate cycle. The latter is also active during growth on propionate, because α-oxidation of propionate via methylcitrate yields pyruvate, which is converted into acetyl-CoA and funnelled into the citrate cycle . The observation that AcnB was purified instead of AcnA is in agreement with the different expression of the two genes. AcnB was identified as the major citrate cycle enzyme, whereas AcnA is an anaerobic stationary-phase enzyme which is specifically induced by iron and redox stress .
Interestingly, two enzymes are involved in the conversion of methylcitrate into 2-methylisocitrate. PrpD is involved in the dehydration of (2S,3S)-methylcitrate to 2-methyl-cis-aconitate. The elimination of water from (2S,3S)-methylcitrate to 2-methyl-cis-aconitate is an unusual reaction, because it displays a syn elimination, which has not previously been found in any other dehydration of a derivative of malate. This may explain why PrpD shows no significant identities with other proteins with known function except deduced proteins from prp operons of many proteobacteria, e.g. S. enterica (Fig. 6). In addition, PrpD shows sequence identities with deduced proteins from the Gram-positive Bacillus subtilis (61%, Mmge, accession no. P45859), the eukaroytes Saccharomyces cerevisiae (57%, Pdh1p, accession no. NP-015326) and Mus musculus (14%, immune responsive protein 1, accession no. XP-127883), as well as the archaeon Sulfolobus tokodaii (23%, long hypothetical Mmge protein, accession no. BAB66901).
The PrpD protein from E. coli possesses high substrate specificity. The best substrate was stereochemically pure (2S,3S)-methylcitrate produced by methylcitrate synthases from E. coli or A. nidulans. Partial activity was also observed with cis-aconitate. As the activity with citrate was very low and that with (2R,3S)-isocitrate was almost absent, it would be of interest to identify the product of the syn hydration of cis-aconitate, perhaps one enantiomer of erythro-isocitrate. No significant activity was detected with many other hydroxy or unsaturated dicarboxylic and tricarboxylic acids such as trans-aconitate, threo-2-methylisocitrate and erythro-2-methylisocitrate, d-malate and l-malate, and (R)-citramalate and (S)-citramalate (Table 3). In E. coli the dehydration of methylcitrate is independent of any metal cofactors, which was also shown for the PrpD protein from S. enterica, but is in disagreement with another investigation , in which the specific activity of the purified PrpD from a genetically amplified source (1.65 U·mg−1 protein) was significantly underestimated. The substrate had been produced with the commercially available citrate synthase from pig heart, which yielded all four possible stereoisomers rather than enantiomeric pure (2S,3S)-methylcitrate as obtained with methylcitrate synthases. Furthermore, the only active stereoisomer is produced in the lowest amount [13,30]. Our own observations on the maximum activity of the PrpD protein with a racemic mixture of all four stereoisomers of chemically synthesized methylcitrate revealed a 10-fold decrease in activity. This may also explain the higher relative activities obtained in the former study with substrates other than methylcitrate.
The necessary syn elimination of water performed by PrpD may be the reason why this reaction cannot be catalysed by aconitase. Furthermore, aconitase eliminates a proton from the R-methylene group of citrate, whereas PrpD removes the proton from the methine group of (2S,3S)-methylcitrate equivalent to the S-methylene group of citrate. There is also a steric conflict of the methyl group of methylcitrate with the catalytically active Asp165 as identified in crystals of mitochondrial aconitase with bound 2-methylisocitrate . It remains unclear, however, whether the citrate cycle aconitase B is always involved in the hydration of 2-methyl-cis-aconitate to 2-methylisocitrate in the bacterial methylcitrate pathway. Some organisms, e.g. Ralstonia eutropha, seem to contain an additional aconitase in their prp operon . The functionality of these proteins and their ability to perform both reactions in the conversion of methylcitrate into 2-methylisocitrate has to be established.
Transcription of the genes of the prp operon underlies a strong regulation. Acetate is not able to induce transcription as studied by Northern-blot experiments and 2D gel electrophoresis. Proteins such as PrpC, PrpB and PrpD were not visible after growth on acetate, even on silver staining (data not shown), whereas a strong signal appeared after growth on propionate (Fig. 7). Probably methylcitrate acts as an inducer, as postulated for S. enterica. Interestingly, we were not able to identify the PrpE protein in the 2D gels, despite the fact that a transcript of prpE was detected in Northern-blot experiments. Therefore, the function of PrpE in wild-type E. coli strains remains unclear. Activation of propionate to propionyl-CoA seems to be performed exclusively by the Acs, which was identified in the 2D gels and Northern-blot experiments of cells grown on acetate as well as on propionate. Probably prpE transcripts are translated when the Acs is mutated, as indirectly shown for S. enterica. In this study an acs mutant strain was still able to grow on propionate .
In conclusion, the prp operon does not harbour all genes necessary for a functional methylcitrate cycle. However, propionate catabolism via methylcitrate (Fig. 1) connects the enzymes of three different pathways to a new functional unit: AcnB, succinate dehydrogenase, fumarase and malate dehydrogenase from the citrate cycle, Acs from the glyoxylate cycle and three special enzymes, which are capable of acting on C7 organic acids (PrpC, PrpD and PrpB).
The authors thank Professor A. Mosandl, Universität Frankfurt/Main, Germany for performing the enantioselective multidimensional capillar gas chromatography with our methylcitrate samples, and Dr D. Linder, Universität Gießen, Germany, for the determination of the N-terminus of aconitase B. The work was supported by grants from Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie.