A Novel 3-Sulfinopropionyl Coenzyme A (3SP-CoA) Desulfinase from Advenella mimigardefordensis Strain DPN7^T Acting as a Key Enzyme during Catabolism of 3,3′-Dithiodipropionic Acid Is a Member of the Acyl-CoA Dehydrogenase Superfamily

Abstract

3-Sulfinopropionyl coenzyme A (3SP-CoA) desulfinase (Acd_DPN7) is a new desulfinase that catalyzes the sulfur abstraction from 3SP-CoA in the betaproteobacterium Advenella mimigardefordensis strain DPN7^T. During investigation of a Tn5::mob-induced mutant defective in growth on 3,3′-dithiodipropionate (DTDP) and also 3-sulfinopropionate (3SP), the transposon insertion was mapped to an open reading frame with the highest homology to an acyl-CoA dehydrogenase (Acd) from Burkholderia phenoliruptrix strain BR3459a (83% identical and 91% similar amino acids). An A. mimigardefordensis Δacd mutant was generated and verified the observed phenotype of the Tn5::mob-induced mutant. For enzymatic studies, Acd_DPN7 was heterologously expressed in Escherichia coli BL21(DE3)/pLysS by using pET23a::acd_DPN7. The purified protein is yellow and contains a noncovalently bound flavin adenine dinucleotide (FAD) cofactor, as verified by high-performance liquid chromatography–electrospray ionization mass spectrometry (HPLC-ESI-MS) analyses. Size-exclusion chromatography revealed a native molecular mass of about 173 kDa, indicating a homotetrameric structure (theoretically 179 kDa), which is in accordance with other members of the acyl-CoA dehydrogenase superfamily. In vitro assays unequivocally demonstrated that the purified enzyme converted 3SP-CoA into propionyl-CoA and sulfite (SO₃²⁻). Kinetic studies of Acd_DPN7 revealed a V_max of 4.19 μmol min⁻¹ mg⁻¹, an apparent K_m of 0.013 mM, and a k_cat/K_m of 240.8 s⁻¹ mM⁻¹ for 3SP-CoA. However, Acd_DPN7 is unable to perform a dehydrogenation, which is the usual reaction catalyzed by members of the acyl-CoA dehydrogenase superfamily. Comparison to other known desulfinases showed a comparably high catalytic efficiency of Acd_DPN7 and indicated a novel reaction mechanism. Hence, Acd_DPN7 encodes a new desulfinase based on an acyl-CoA dehydrogenase (EC 1.3.8.x) scaffold. Concomitantly, we identified the gene product that is responsible for the final desulfination step during catabolism of 3,3′-dithiodipropionate (DTDP), a sulfur-containing precursor substrate for biosynthesis of polythioesters.

INTRODUCTION

Acyl coenzyme A (CoA) dehydrogenases (Acds) (EC 1.3.8.x) are a superfamily of flavoenzymes that usually catalyze the α,β-dehydrogenation of acyl-CoA thioesters using electron-transferring flavoproteins (ETFs) as their physiological electron acceptors (1) (Fig. 1). Acds are involved mainly in fatty acid oxidation or in branched-amino-acid metabolism (1). Furthermore, members of the Acd superfamily that play a role in less common metabolic pathways, particularly in bacteria, have been identified. Examples are the involvement of benzylsuccinyl-CoA dehydrogenase or naphthyl-2-methylsuccincyl-CoA dehydrogenase in anaerobic degradation of aromatic hydrocarbons (2, 3) and of 2-methylsuccinyl-CoA dehydrogenase (Mcd) in the ethylmalonyl-CoA pathway (4). In Mycobacterium tuberculosis, Acds are involved in cholesterol degradation (5), and in Salmonella enterica serovar Typhimurium, they participate in the stress response (6).

Fig 1.

Overview of acyl-CoA dehydrogenase reactions. (A) General reaction scheme of acyl-CoA dehydrogenases. (B) Reaction of glutaryl-CoA dehydrogenases (GCDs). Nondecarboxylating GCDs introduce a double bond, yielding glutaconyl-CoA (1). Decarboxylating GCDs first introduce a double bond, followed by a decarboxylation reaction (2), yielding crotonyl-CoA and CO₂. (C) Reaction of Acd_DPN7 with 3SP-CoA yielding sulfite and propionyl-CoA.

The general three-dimensional (3D) structure of Acds is based on three highly conserved major domains that are found in all members of the Acd superfamily: an N-terminal domain of α-helices, a central domain of β-sheets, and another domain of α-helices at the C terminus (1). Almost all characterized Acds, except human acyl-CoA dehydrogenase 9 (ACAD9) and the very long-chain acyl-CoA dehydrogenase (VLCAD), exist as soluble homotetramers. Therein, the subunits, with a mass of about 43 kDa, exhibit tetrahedral symmetry that can be considered a dimer of dimers (1).

As a common feature, all Acds contain one noncovalently bound flavin adenine dinucleotide (FAD) per subunit, with the flavin moiety positioned at the active site between the central β-sheet and the C-terminal α-helix domain of one subunit. The adenine moiety of the FAD extends into a pocket between the C-terminal α-helix domain of the neighboring subunit (1). Highly conserved residues are involved in binding of the adenine moiety of FAD and are important for the identification of putative Acds (7).

A glutamate residue, a common catalytic base in all described Acds, initiates the dehydrogenation reaction by abstraction of the pro-R hydrogen of the substrate's α-carbon atom (8). The corresponding pro-R β-hydrogen is eliminated as a hydride equivalent to the N-5 position of the flavin (9–12). Regeneration of the so-formed reduced flavin adenine dinucleotide (FADH₂) occurs via ETFs that serve primarily as soluble electron shuttles between various (mainly soluble) flavoprotein dehydrogenases (13–15). However, although members of the Acd superfamily display a highly similar overall 3D structure, the structure of their binding cavity differs between the various Acds and is determining for their substrate specificity.

3,3′-Dithiodipropionate (DTDP) is an organic disulfide and a precursor for the synthesis of polythioesters (PTEs) in bacteria (16, 17). Advenella mimigardefordensis strain DPN7^T, a betaproteobacterium, was isolated due to its ability to utilize DTDP as a sole source of carbon and energy (18, 19). Elucidation of the degradation pathway of DTDP in this bacterium and identification of the genes involved could provide a reasonable strategy to engineer strains suitable for biotechnological PTE production. Genotypic and phenotypic characterization of Tn5::mob-induced mutants as well as characterization of putatively involved enzymes helped to unravel the first steps of this pathway (Fig. 2) (18, 20–22). Initial cleavage of DTDP by a disulfide reductase (LpdA) yields two molecules of 3-mercaptopropionic acid (3MP) (20). In the next step, a thiol dioxygenase (Mdo) catalyzes the oxidation to 3-sulfinopropionate (3SP), a structural succinate analogue (21). The latter is thereafter activated to the corresponding 3SP-CoA thioester by a succinate-CoA ligase (SucCD) that is able to accept several structurally related substrates (22). So far, the metabolic fate of the sulfur moiety has been unclear.

Fig 2.

Degradation pathway of DTDP. DTDP is initially cleaved by an NADH-dependent disulfide reductase (LpdA) (20), yielding two molecules of 3MP, which are further oxygenated by a dioxygenase (Mdo) (21), yielding 3SP. The latter is activated by succinate-CoA ligase (SucCD) (22) to the corresponding CoA thioester. 3SP-CoA desulfinase (Acd_DPN7) catalyzes the final desulfination step, releasing sulfite with the concomitant formation of propionyl-CoA, which subsequently enters the central metabolism via the methyl citric acid cycle.

In this study, insertion of Tn5::mob into acd_DPN7 completely impaired growth on DTDP or 3SP as the sole carbon source. Thus, an involvement of this Acd homologue from A. mimigardefordensis strain DPN7^T during DTDP catabolism was predicted. A reaction similar to that in glutaryl-CoA dehydrogenase (GCDs) (23) was supposed for Acd_DPN7. In contrast to GCDs, Acd_DPN7 cleaves not a carboxylic acid group from glutaryl-CoA with concomitant release of CO₂ but a sulfino group from 3SP-CoA with simultaneous release of sulfite. In terms of structure, the sulfino group differs from the carboxylic acid group only by the replacement of a sulfur atom by a carbon atom. In contrast to the crotonyl-CoA formed by GCDs, the Acd_DPN7 reaction product propionyl-CoA possesses no double bond, indicating that the reaction catalyzed by Acd_DPN7 represents a hydrolytic desulfination rather than an oxidative decarboxylation. Therefore, we report here for the first time this uncommon reaction in a member of the Acd superfamily. This reaction is the final reaction during DTDP degradation in A. mimigardefordensis DPN7^T before propionyl-CoA enters the central metabolism on the level of the methylcitric acid cycle.

MATERIALS AND METHODS

Bacterial strains and cultivation conditions.

All strains used in this study are listed in Table 1. Cells of A. mimigardefordensis were cultivated at 30°C in mineral salt medium (MSM) (29) containing 20 mM gluconate, 20 mM DTDP, or 20 mM 3SP as the sole source of carbon and energy. Cells of Escherichia coli were cultivated in Luria-Bertani (LB) medium at 30°C or 37°C under aerobic conditions on a rotary shaker with agitation at 110 rpm (30). Carbon sources were added to MSM as filter-sterilized stock solutions as indicated in the text. For maintenance of plasmids, antibiotics were prepared according to the method of Sambrook et al. (30) and added to the media at the following concentrations (in μg/ml): ampicillin (75), gentamicin (20), kanamycin (Km) (50), and tetracycline (12.5). In E. coli, heterologous expression of genes under the control of a lac promoter was achieved by cultivation in ZYP-5052 medium, an autoinductive medium, as described previously by Studier (31).

Table 1.

Strains and plasmids used in this study

Strain or plasmid	Descriptiona	Reference(s) and/or source
Strains
A. mimigardefordensis DPN7T	Wild type, DTDP and 3SP utilizing	19; DSM 17166T, LMG 22922T
A. mimigardefordensis Δacd	No growth on DTDP and 3SP	This study
E. coli One Shot Mach1-T1R	F− ϕ80lacZΔM15 ΔlacX74 hsdR(rK− mK+) ΔrecA1398 endA1 tonA	Invitrogen, Carlsbad, CA
E. coli Top10	F− mcrA Δ(mrr-hsdRMS-mcrBC) rpsL nupG ϕ80lacZΔM15 ΔlacX74 deoR recA1 araD139 Δ(ara-leu)7697 galU galK endA1	Invitrogen, Carlsbad, CA
E. coli BL21(DE3)/pLysS	F ompT hsdSB (rB mB) gal dcm (DE3) pLysS (Cmr)	Novagen, Madison, WI
E. coli S17-1	thi-1 proA hsdR17 (rK− mK+) recA1; tra gene of plasmid RP4 integrated into genome	25
Plasmids
pSUP5011	pBR325 containing the transposon Tn5::mob; Apr Cmr Kmr	25
pGEM-T Easy	Apr lacPOZ′	Promega, Madison, WI
pGEM-T-Easy::ΔacdDPN7	Apr lacPOZ′; upstream and downstream flank of acdDPN7 (1.4 kbp)	This study
pJQ200mp18Tc	sacB oriV oriT traJ Tcr	26, 27
pJQ200mp18Tc::ΔacdDPN7	sacB oriV oriT traJ Tcr; upstream and downstream flank of acdDPN7 (1.4 kbp)	This study
pBBR1MCS-5	Broad-host-range cloning vector; Gmr lacZα	28
pBBR1MCS–5::acdDPN7	pBBR1MCS5 containing the acd gene of A. mimigardefordensis DPN7T and a 14–bp upstream region as a 1,220–bp ApaI–PstI fragment; Gmr
pCR2.1–TOPO	Kmr Apr lacZα	Invitrogen, Carlsbad, CA
pCR2.1–TOPO::acdDPN7	Kmr Apr lacZα acdDPN7	This study
pET23a(+)	Apr; C–terminal His6 tag	Novagen, Madison, WI
pET23a(+)::acdDPN7	Apr acdDPN7; expressing AcdDPN7–His6	This study

^aFor abbreviations used in genotypes of E. coli, see reference 24. Ap^r, ampicillin resistance; Cm^r, chloramphenicol resistance; Gm^r, gentamicin resistance; Km^r, kanamycin resistance; Tc^r, tetracycline resistance.

Chemicals.

Thiochemicals of high-purity grade were purchased from Sigma-Aldrich (Steinheim, Germany). 3SP was synthesized according to methods described previously Jollés-Bergeret (32); the procedure was modified by one repetition of the step for alkaline cleavage of the intermediate bis-(2-carboxyethyl)sulfone (18, 22). Synthesis and purity of the substance were confirmed by gas chromatography (GC) and GC/mass spectrometry (MS). According to GC/MS, the purity of the used 3SP was at least 96.5%.

Analysis of 3SP by GC or GC/MS.

To investigate the consumption of DTDP or 3SP and to monitor the formation of catabolic intermediates, cell-free supernatants were analyzed by GC as described previously (22). Samples were subjected to methylation in the presence of 1 ml of chloroform, 0.850 ml of methanol, and 0.150 ml of sulfuric acid for 2 to 4 h at 100°C. Upon methylation, 2 ml of H₂O was added, and the samples were shaken vigorously for 30 s. After phase separation, the organic layer containing the resulting methyl esters of the organic acids was analyzed in an HP6850 gas chromatograph equipped with a BP21 capillary column (50 m by 0.22 mm with a film thickness of 250 nm; SGE, Darmstadt, Germany) and a flame ionization detector.

Aliquots of synthesized 3SP and supernatants, which showed unknown substances during GC analysis, were analyzed by GC/MS. The samples were subjected to acid-catalyzed esterification in the presence of methanol, as described above. The resulting methyl esters were then characterized by using an HP6890 gas chromatograph equipped with a model 5973 EI MSD mass selective detector (Hewlett-Packard, Waldbronn, Germany). Three microliters of the organic phase was then analyzed after split injection (split ratio, 20:1) by using a BPX 35 capillary column (50 m by 0.22 mm, with a film thickness of 250 nm; SGE). Helium was used as the carrier gas at a flow rate of 0.6 ml/min. The temperatures of the injector and detector were 250°C and 240°C, respectively. The same temperature program as that used during GC analysis was applied. Identification of peaks was performed by using AMDIS software in combination with the NIST database (33).

Synthesis of CoA thioesters.

Propionyl-CoA, butyryl-CoA, isobutyryl-CoA, valeryl-CoA, isovaleryl-CoA, succinyl-CoA, or glutaryl-CoA was prepared by a modified method according to the anhydride method of Simon and Shemin (34). For this, 10 mg of the trilithium salt of coenzyme A (Merck KGaA, Darmstadt, Germany) was dissolved in 0.5 M Tris-HCl (pH 8.0). The solution was stirred on ice, and small portions of the respective anhydride were added to this solution until no free coenzyme A was detectable by Ellman's spot test by applying 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB) (35). The pH value of the solution was then adjusted to 5.0 by addition of concentrated hydrochloric acid. The resulting acyl-CoA thioesters were immediately used for enzyme assays.

Isolation and manipulation of DNA.

Chromosomal DNA of A. mimigardefordensis strain DPN7^T was isolated according to the method of Marmur (36). Plasmid DNA was isolated from E. coli by using peqGOLD plasmid miniprep kit I from Peqlab Biotechnologie GmbH (Erlangen, Germany) according to the manufacturer's manual. DNA was digested with restriction endonucleases (Fermentas GmbH, St. Leon-Rot, Germany) under conditions described by the manufacturer. PCR was carried out with an Omnigene HBTR3CM DNA thermal cycler (Hybaid, Heidelberg, Germany) or a PeqSTAR 2× Gradient thermal cycler (Peqlab Biotechnologie GmbH, Erlangen, Germany), using Platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA). T4 DNA ligase was purchased from Fermentas (Fermentas GmbH, St. Leon-Rot, Germany). Primers were synthesized by MWG-Biotech AG (Ebersberg, Germany) and are listed in Table S1 in the supplemental material.

Transfer of DNA.

Competent cells of E. coli strains were prepared and transformed by the CaCl₂ procedure (30).

DNA sequencing and sequence data analysis.

Sequence analysis was performed by Seqlab (Göttingen, Germany) or by the Institut für Klinische Chemie und Laboratoriumsmedizin at the Universitätsklinikum Münster. At the latter, DNA sequencing was performed according to the method of Sanger et al. (37), by applying the BigDye Terminator v3.1 cycle sequencing kit according to the manufacturer's instructions (Applied Biosystems, Darmstadt, Germany). Samples were submitted to purification of the extension products and sequencing in an ABI Prism 3700 DNA analyzer (Applied Biosystems, Darmstadt, Germany). Sequences were analyzed by using the program BLAST (National Center for Biotechnology Information [http://www.ncbi.nlm.nih.gov/BLAST/]) (38).

Transposon mutagenesis.

For transposon mutagenesis of A. mimigardefordensis strain DPN7^T, a suicide plasmid technique described previously (25) was used. The vector pSUP5011 was transferred from E. coli S17-1 to Km-susceptible A. mimigardefordensis strain DPN7^T by conjugation, using the spot agar mating technique (39). Transconjugants were selected on MSM agar plates containing 50 μg of Km ml⁻¹ (MSM^Km) and 0.2% (wt/vol) sodium propionate or 0.3% (wt/vol) 3SP (master plates). Putative Tn5::mob-induced mutants were transferred in a coordinated pattern onto MSM^Km agar plates containing 0.4% (wt/vol) DTDP and 0.3% (wt/vol) 3SP (selection plates) and onto propionate master plates for further analysis.

Genotypic characterization of Tn5::mob-induced mutants.

Flanking regions of the Tn5::mob insertion loci were sequenced by a PCR-based two-step genome walking method (40). IS50 walking and IS50 sequencing primers were constructed as described previously by Pilhofer et al. (40). Genomic DNA of the Tn5::mob-induced mutant was isolated according to methods described previously by Marmur (36) (primers are listed in Table S1 in the supplemental material).

Construction of an acd precise deletion gene replacement plasmid.

The 578- and 837-bp fragments upstream and downstream of acd_DPN7 were amplified by using fwdXbaAcd_up/revEcoR1Acd_up and fwdEcoR1Acd_down/revXbaAcd_down, respectively. The oligonucleotides used for PCR are listed in Table S1 in the supplemental material. The resulting fragments were EcoRI digested and ligated to yield a 1,415-bp fragment. This fragment was amplified by using fwdXbaAcd_up and revXbaAcd_down, and the resulting PCR product was cloned into the XbaI site of pJQ200mp18Tc (26, 27) to yield pJQ200mp18Tc::Δacd.

Construction of an acd gene replacement strain using the sacB system.

Gene replacement was accomplished by adaptation of standard protocols (26, 27). Plasmid pJQ200mp18Tc::Δacd was used to generate the A. mimigardefordensis Δacd mutant. The plasmid was mobilized from E. coli donor strain S17-1 to the A. mimigardefordensis DPN7^T recipient strain by the spot agar mating technique (39). A successfully generated gene replacement strain was identified and confirmed by PCR analyses and DNA sequencing using oligonucleotides listed in Table S1 in the supplemental material.

Construction of plasmids for complementation experiments.

For complementation studies in the broad-host-range vector pBBR1MCS-5 (28), acd_DPN7 and a 14-bp upstream region were amplified by PCR by applying oligonucleotides upAcd_MCS and downAcd_MCSB (see Table S1 in the supplemental material). The PCR product was digested with restriction endonucleases ApaI and PstI and subsequently purified from an agarose gel by using the NucleoTrap kit (Macherey and Nagel, Düren, Germany).

The purified gene was subsequently ligated into pBBR1MCS-5, which was linearized with the same restriction endonucleases, and the ligation products were transformed into CaCl₂-competent cells of E. coli S17-1 and E. coli Top10. Transformants were selected on LB medium containing gentamicin, isopropyl-β-d-thiogalactopyranoside (IPTG), and 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal). The hybrid plasmid pBBR1MCS-5::acd_DPN7 was then transferred into the A. mimigardefordensis Δacd strain by conjugation (25).

Construction of the expression plasmid pET23a::acd_DPN7.

Acd_DPN7 was amplified from total genomic DNA of A. mimigardefordensis strain DPN7^T by PCR using Biomix containing Taq DNA polymerase (Bioline GmbH, Luckenwalde, Germany) and the following oligonucleotides: Acdfwd_NdeI and Acdrev_oStopNotI (see Table S1 in the supplemental material). PCR products were isolated from agarose gels by using the peqGOLD gel extraction kit (Peqlab Biotechnologie GmbH, Erlangen, Germany) and ligated with pCR2.1-TOPO DNA (Invitrogen, Carlsbad, CA). Ligation products were used for transformation of CaCl₂-competent cells of E. coli OneShot Mach1-T1^R, and transformants were selected on LB agar plates containing IPTG, X-Gal, and ampicillin. For heterologous expression in the T7 promoter/polymerase-based expression vector pET23a(+) (Novagen, Madison, WI), acd_DPN7 was obtained by digestion of hybrid plasmid pCR2.1-TOPO::acd_DPN7 with restriction endonucleases NdeI and NotI and purified from an agarose gel by using the peqGOLD gel extraction lit (Peqlab Biotechnologie GmbH, Erlangen, Germany). After ligation with the expression vector pET23a(+), which was linearized with the same restriction endonucleases, the ligation product, pET23b(+)::acd_DPN7, was used for transformation of CaCl₂-competent cells of E. coli Top10. After selection of transformants on LB medium agar plates containing ampicillin, the hybrid plasmids were isolated, analyzed by sequencing, and used for transformation of CaCl₂-competent cells of E. coli BL21(DE3)/pLysS (Novagen, Madison, WI).

Purification of heterologously expressed Acd_DPN7.

Cells from 50- to 500-ml cultures were harvested by centrifugation (15 to 45 min at 4°C at 3,400 × g), washed twice with sterile saline, and stored at −20°C until use. For purification of histidine-tagged fusion proteins, the buffers were prepared as recommended by the manufacturer of the His Spin Trap or HisTrap HP affinity columns (catalogue number 28-4013-53; GE Healthcare, Uppsala, Sweden). Cells were resuspended in 100 mM Tris-HCl binding buffer (pH 7.4) containing 500 mM sodium chloride and 20 mM imidazole and afterwards disrupted by a 3-fold passage through a cooled French press (100 × 10⁶ Pa). Soluble protein fractions of crude extracts were obtained in the supernatants after 1 h of centrifugation at 100,000 × g at 4°C and were subsequently applied for enzyme purifications. To obtain small amounts of purified histidine-tagged fusion protein for initial tests, HisSpinTrap affinity columns (catalogue number 28-4013-53; GE Healthcare, Uppsala, Sweden) were used according to the manufacturer's instructions. Ni-nitrilotriacetic acid (NTA) columns were equilibrated with 100 mM Tris-HCl buffer (pH 7.4) containing 20 mM imidazole and 500 mM sodium chloride. The same buffer containing 40 mM imidazole was used for the washing step, while the elution buffer contained 500 mM imidazole. For large-scale purification, two HisTrap HP columns (catalogue number 17-5247-01; GE Healthcare, Uppsala, Sweden) with a total volume of 2 ml were connected. The columns were equilibrated with 5 volumes of 100 mM Tris-HCl buffer (pH 7.4) containing 20 mM imidazole and 500 mM sodium chloride. Two washing steps with the same buffer containing 100 and 150 mM imidazole, respectively, were performed before elution was done with 500 mM imidazole. Purified enzyme was stable in the elution buffer containing 50% glycerol at −20°C for several months at a final protein concentration of 2 mg/ml. Protein concentrations were determined as described previously by Bradford (41).

Analytical size-exclusion chromatography.

The molecular mass of Acd_DPN7 was determined by analytical size-exclusion chromatography (SEC) using a Superdex 200 HP 16/600 column. The column was equilibrated with 50 mM sodium phosphate buffer (pH 7.4) containing 150 mM sodium chloride. Calibration was performed by applying ovalbumin (44 kDa), conalbumin (75 kDa), aldolase (158 kDa), and ferritin (440 kDa) from the high-molecular-mass gel filtration calibration kit (catalogue number 28-4038-42; GE Healthcare, Uppsala, Sweden), according to the manufacturer's instructions. Eight hundred micrograms of immobilized-metal affinity chromatography (IMAC)-purified Acd_DPN7 was applied onto the column. The column was operated at a flow rate of 1 ml/min. The elution volume was determined by monitoring the absorbance at 280 nm.

Spectroscopic properties and cofactor analysis of recombinant Acd_DPN7.

Spectra of purified recombinant Acd_DPN7 were recorded with a UV-2600 spectrophotometer (Shimadzu, Duisburg, Germany) with 10 mM HEPES buffer (pH 7.4). Measurements occurred aerobically at 25°C between wavelengths of 210 and 800 nm, applying 1.7 mg recombinant Acd_DPN7 in 1 ml buffer (9.65 μM). All spectra were recorded in quartz cuvettes (117.104-QS, with a path length of 10 mm; Hellma Analytics, Müllheim, Germany) against blank buffer solution.

To identify the cofactor, 4.9 mg Acd_DPN7 in 200 μl HisTag elution buffer (100 mM Tris-HCl [pH 7.4], 500 mM NaCl, 500 mM imidazole) was mixed with 50 μl trichloroacetic acid (TCA) (15%, wt/vol) and incubated for 30 min on ice. Precipitated protein was removed by centrifugation (13,000 × g for 15 min), and the supernatant was subsequently analyzed by high-performance liquid chromatography–electrospray ionization mass spectrometry (HPLC-ESI-MS), as described below.

Enzyme assays.

The putative substrate of Acd_DPN7, 3SP-CoA, was not commercially available. Therefore, in situ formation of 3SP-CoA in the reaction mixture was necessary prior to Acd_DPN7 addition. For this reason, purified succinate-CoA ligase (SucCD) from E. coli BL21(DE3), which has been shown to activate 3SP to 3SP-CoA, was applied (J. Nolte, C.-L. Schepers, M. Schürmann, E. Vogel, J. H. Wübbeler, and A. Steinbüchel, unpublished results). After each step, a sample was subjected to HPLC-ESI-MS analysis. The reaction mixture contained 1 mM ATP, 10 mM MgCl₂, 1 mM CoA, and 5 mM 3SP before addition of SucCD. After the addition of 142 μg/ml purified SucCD and incubation for 10 min at 37°C, the enzyme was heat inactivated (10 min at 95°C). Subsequently, 245 μg/ml Acd_DPN7 was added, and the reaction mixture was incubated at 37°C for 10 min. The reaction was stopped by the addition of 30 μl TCA (15%, wt/vol).

Consequently, for activity measurements, 3SP-CoA was synthesized in a way similar to that described above. Therefore, a solution containing 0.4 mM ATP, 7 mM MgCl₂, 0.25 mM CoA, and 10 mM 3SP in 50 mM Tris-HCl (pH 7.4) was incubated in the presence of 200 μg/ml of purified SucCD for up to 1 h at 37°C. From time to time, 2-μl samples were mixed with 2 μl of a 10 mM DTNB solution (in 50 mM Tris-HCl, pH 7.4) (35) to test for complete transformation of CoA to 3SP-CoA until the absence of yellow color indicated complete transformation. Next, SucCD was removed from the reaction mixture, applying Vivaspin 6 tubes with a molecular mass cutoff of 10 kDa (catalogue number 512-3775; Sartorius Group, Göttingen, Germany) at 3,400 × g for 20 min at 4°C.

For kinetic measurements, a spectrophotometric assay utilizing the reaction of DTNB with sulfite was applied (42). The increase in absorption at 412 nm was monitored (ϵ_{412 nm} = 14,150 M⁻¹ cm⁻¹) (43). A typical enzyme assay mixture contained 0.1 mM DTNB and appropriate dilutions of the 3SP-CoA solution for enzyme assays. Activities for 10 different 3SP-CoA concentrations ranging from 0.0 mM to 0.2 mM were measured in triplicate. The reaction was started by the addition of 1 to 10 μg of purified Acd_DPN7. One unit of activity was defined as the amount of enzyme that converts 1 μmol substrate in 1 min.

In an attempt to assess Acd_DPN7's dehydrogenation ability, several acyl-CoA thioesters (propionyl-CoA, butyryl-CoA, isobutyryl-CoA, valeryl-CoA, isovaleryl-CoA, 3SP-CoA, succinyl-CoA, and glutaryl-CoA) were applied in a spectrophotometric assay at 300 nm. The assay mixture contained 50 mM Tris-HCl (pH 7.4), 0.1 mM the respective acyl-CoA thioester, and 0.2 mM ferrocenium hexafluorophosphate as an artificial electron acceptor (ϵ_{300 nm} = 4.3 mM⁻¹ cm⁻¹ [44]). The assay mixture containing ferrocenium hexafluorophosphate was preincubated at 30°C. Subsequently, 200 μg purified Acd_DPN7 was added. After incubation for another minute, the reaction was started by the addition of 10 μl of the respective CoA thioester (final concentration of 0.1 mM). The reaction was stopped by the addition of 15 μl TCA (15%, wt/vol). The samples were analyzed by HPLC-ESI-MS for the formation of the corresponding dehydrogenated form of the CoA thioester.

Analysis of CoA thioesters by HPLC-ESI-MS.

Separation and detection of different CoA thioesters occurred by HPLC-ESI-MS according to a method described previously (22). HPLC-ESI-MS analyses were carried out by using an UltiMate 3000 HPLC apparatus (Dionex GmbH, Idstein, Germany) connected directly to an LXQ Finnigan (ThermoScientific, Dreieich, Germany) mass spectrometer. An Acclaim 120 C₁₈ reversed-phase liquid chromatography (LC) column (4.6 by 250 mm, 5 μm, 120-Å pores; Dionex GmbH, Idstein, Germany) served to separate the CoA thioesters at 30°C. A gradient system was used, with 50 mM ammonium acetate (pH 5.0) adjusted with acetic acid (eluent A) and 100% (vol/vol) methanol (eluent B) as eluents. Elution occurred at a flow rate of 0.3 ml/min. Ramping was performed as follows: equilibration with 90% eluent A for 2 min before injection and afterwards a change from 90% to 0% eluent A in 40 min, followed by holding for 2 min and then returning to 90% eluent A within 5 min. For all steps, eluent B complements to 100%. Detection of CoA thioesters was performed at 259 nm by using a photodiode array detector. The instrument was tuned by direct infusion of a solution of 0.4 mM CoA at a flow rate of 10 μl/min into the ion source of the mass spectrometer to optimize the ESI-MS system for maximum generation of protonated molecular ions (parents) of CoA derivatives. The following tuning parameters were retained for optimum detection of CoA thioesters: capillary temperature of 300°C, sheath gas flow of 12 liters/h, auxiliary gas flow of 6 liters/h, and sweep gas flow of 1 liter/h. The mass range was set to m/z 50 to 1,000 Da when run in the scan mode. The collision energy in the tandem MS (MS-MS) mode was set to 30 V and delivered fragmentation patterns, which are in good accordance with those found previously (45).

Phylogenetic tree analysis and multiple-sequence alignment.

For phylogenetic tree analysis and multiple-sequence alignment, biochemically characterized members of the Acd superfamily were identified in the BRENDA database (http://www.brenda-enzymes.info/). The full-length amino acid sequences of Acd_DPN7 homologues were retrieved from the GenBank database. GenBank accession numbers are given in the corresponding figures. The phylogenetic tree was constructed by using ClustalX (46) and the NJplot program (47). The BioEdit program (48) was used for multiple-sequence alignments.

Nucleotide sequence accession number.

The DNA sequence and the deduced amino acid sequence of acd_DPN7 were deposited in the GenBank database under accession number JX535522.

RESULTS

Identification of the putative desulfinase gene in A. mimigardefordensis strain DPN7^T.

Tn5::mob mutagenesis was the method of choice to elucidate the catabolic pathway of DTDP in A. mimigardefordensis strain DPN7^T. In one of the Tn5::mob-induced mutants, which was unable to use DTDP or 3SP as the sole carbon source for growth, the transposon was localized in a gene (acd_DPN7) that showed the highest homology to a putative Acd from Burkholderia phenoliruptrix strain BR3459a (83% identical and 91% similar amino acids; GenBank accession number AFT90224.1).

Gene organization.

The regions adjacent to acd_DPN7 are shown in Fig. S1 in the supplemental material. lysR, a gene coding for a transcriptional regulator, and act, a gene coding for a putative acyl-CoA transferase family III member, were identified in the upstream region. mdo in the downstream region codes for a 3-mercaptopropionate (3MP) dioxygenase. This enzyme is involved in DTDP catabolism, and it has been shown to catalyze the oxidation of 3MP to 3SP during DTDP catabolism (18, 22). ahpD, an alkylhydroperoxidase gene, and tctC, a gene coding for a putative Bug (Bordetella uptake gene)-like extracytoplasmic solute-binding receptor, are situated further downstream. Strain DPN7^T itself possesses 32 further genes potentially coding for acyl-CoA dehydrogenases, but they all exhibit less than 32% identical amino acids to the acd_DPN7 gene product (best E value, 5e−65; best coverage, 97% [J. H. Wübbeler, unpublished results]).

Precise deletion mutant of acd_DPN7.

To verify the observed phenotype and to exclude polar effects of the transposon insertion, plasmid pJQ200mp18Tc::Δacd harboring the flanking regions of acd_DPN7 was constructed and transferred into wild-type A. mimigardefordensis DPN7^T by conjugation. Upon homologous recombination initiated by parts of this suicide plasmid, this resulted in the precise deletion of the acd_DPN7 gene (see Fig. S1 in the supplemental material). The A. mimigardefordensis Δacd mutant grew normally with propionate, succinate, and gluconate but showed no growth on DTDP and 3SP as the Tn5::mob-induced mutant (Table 2).

Table 2.

Utilization of carbon sources by cells of the wild type and the Δacd mutant of A. mimigardefordensis strain DPN7^Ta

Carbon source	Utilization of carbon source by:
	Wild type	Δacd mutant
Propionate	+	+
Succinate	+	+
Gluconate	+	+
DTDP	+	−
3SP	+	−

^aFor comparison of growth of A. mimigardefordensis strain DPN7^T and the defined A. mimigardefordensis Δacd mutant, cells were cultivated on solid MSM containing the carbon sources mentioned above at a concentration of 0.2% (wt/vol). This is equivalent to 21 mM sodium propionate, 12 mM disodium succinate, 9 mM sodium gluconate, 10 mM DTDP, and 11 mM 3SP.

Utilization of 3SP in the wild type and in the defined deletion mutant and identification of potential accumulating metabolites.

Cells of the wild type and of the Δacd mutant of A. mimigardefordensis were cultivated in MSM containing 0.2% (wt/vol) propionate (21 mM) plus 0.6% (wt/vol) DTDP (29 mM) or 0.4% (wt/vol) 3SP (22 mM) (Fig. 3). Supernatant analyses by GC/MS revealed that DTDP consumption increased significantly in wild-type A. mimigardefordensis DPN7^T after about 58 h. Propionate is most probably utilized first and was depleted at this time. In contrast to this, no significant consumption of DTDP was observed for the A. mimigardefordensis Δacd mutant. A. mimigardefordensis strain DPN7^T consumed most of the 3SP within 24 h, whereas in the culture medium of the A. mimigardefordensis Δacd mutant, the 3SP concentration decreased only slightly. In samples of the A. mimigardefordensis Δacd mutant, which were taken after 115 h and 164 h, a new compound with a retention time of 17.1 min was detected in the GC/MS chromatogram (Fig. 3B). This compound was also detected during a previous cultivation experiment with the Tn5::mob-induced mutant and in another growth experiment with the A. mimigardefordensis Δacd mutant, applying glycerol or glucose as a carbon source in addition to 3SP (data not shown). The corresponding mass spectrum confirmed the structure of dimethylated 3-sulfino-2,3-dehydropropionate (3SDHP) (Fig. 3B). The structure is shown in Fig. 3C. The methylation is most probably due to sample preparation for GC/MS. Hence, 3SDHP is the physiological metabolite. It is unclear whether the cis- or the trans-isomer was excreted by the A. mimigardefordensis Δacd mutant. Due to the lack of standards, the compound could not yet be quantified. 3SDHP was not detected in samples which were taken during cultivation of the wild type.

Fig 3.

Consumption of DTDP and 3SP by wild-type A. mimigardefordensis DPN7^T and by the A. mimigardefordensis Δacd mutant. Formation of 3-sulfino-2,3-dehydropropionate by the A. mimigardefordensis Δacd mutant. (A) Cells of the wild type (squares) and of the A. mimigardefordensis Δacd mutant (triangles) were cultivated in MSM containing 0.2% (wt/vol) propionate (21 mM) and 0.6% (wt/vol) DTDP (29 mM) (solid lines) or 0.4% (wt/vol) 3SP (22 mM) (dashed lines). (B) A new compound at a retention time of 17.1 min was detected by GC/MS in the supernatant of the A. mimigardefordensis Δacd mutant in samples taken after 115 and 164 h. The mass spectrum (inset) confirmed the structure of dimethylated 3-sulfino-2,3-dehydropropionate (3SDHP). The methylation is due to sample preparation. This compound was not detected in samples taken during cultivation of the wild type. (C) Structures of dimethylated 3SDHP. Numbers correspond to the mass-to-charge ratio (m/z). It is unclear whether the cis- or the trans-isomer was excreted by the A. mimigardefordensis Δacd mutant. Due to the lack of standards, the amount of the compound could not be quantified. This compound was also detected during supernatant analysis with the Tn5::mob-induced mutant (data not shown).

Complementation studies applying pBBR1MCS-5::acd_DPN7.

In an attempt to complement the DTDP- and 3SP-negative phenotype of the A. mimigardefordensis Δacd mutant, hybrid plasmid pBBR1MCS-5::acd_DPN7 was transferred to the deletion mutant by conjugation, and transconjugants were selected on MSM containing 0.5% (wt/vol) gluconate and gentamicin. Unexpectedly, no growth was observed on MSM agar plates containing 20 mM DTDP as the sole source of carbon and energy after 3 days, whereas the plasmid rescued growth on MSM agar plates containing 20 mM 3SP. Next, A. mimigardefordensis wild-type, A. mimigardefordensis Δacd, and A. mimigardefordensis Δacd pBBR1MCS-5::acd_DPN7 strains were cultivated in liquid MSM containing 20 mM 3SP as the sole source of carbon and energy (Fig. 4). While the wild type reached the stationary phase within 24 h, the deletion mutant showed no growth at all. Growth of the complemented mutant was delayed but finally reached a high cell mass similar to that of the wild type.

Fig 4.

Growth on 3-sulfinopropionate (3SP). Cells of wild-type A. mimigardefordensis, the A. mimigardefordensis Δacd mutant, and the A. mimigardefordensis Δacd mutant harboring pBBR1MCS-5::acd_DPN7 were precultivated in liquid MSM containing 20 mM sodium gluconate, supplied with gentamicin in the case of the complementation vector. Cells were harvested and washed twice with sterile saline prior to use for inoculation of the main culture. Main cultivation occurred in liquid MSM containing 20 mM 3SP as the sole source of carbon and energy in flasks without baffles at 30°C and with agitation at 120 rpm. ⧫, A. mimigardefordensis wild type; ■, A. mimigardefordensis Δacd mutant; ▲, A. mimigardefordensis Δacd mutant harboring pBBR1MCS-5::acd_DPN7. OD₆₀₀, optical density at 600 nm.

3SP-CoA desulfinase is a member of the Acd superfamily.

To identify structural similarities to and differences from other Acds, a multiple-sequence alignment of biochemically characterized members of the Acd superfamily was performed (see Fig. S2 in the supplemental material). They cover a broad range of substrates, which is shown in Fig. S3 in the supplemental material. Generally, all proteins shared 24 to 33% identical amino acids with Acd_DPN7. The following conserved amino acids were identified in Acd_DPN7. Amino acids residues K154, W156, I157, and T158 (marked by + in Fig. S2 in the supplemental material) correspond to a conserved K-X-W/F-I-T motif, which is involved in noncovalent binding of the isoalloxazine ring of the FAD cofactor (7). In most Acds, a complete G-G-X-G motif is found, which forms hydrogen bonds to the FAD cofactor of the neighboring subunit (49). Two glycine residues of this motif (G342 and G345) (indicated by filled semicircles in Fig. S2 in the supplemental material) are conserved in Acd_DPN7.

The alignments clearly showed that Acds are divided into two distinct groups referring to the position of the catalytic glutamate residue. (i) In isovaleryl-CoA dehydrogenase (IVD), long-chain acyl-CoA dehydrogenase (LCAD), or citronellyl-CoA dehydrogenase (CD), the catalytic glutamate residue is located in helix H (indicated by ▲ in Fig. S2 in the supplemental material) (8, 50, 51). In human LCAD and IVD, the catalytically active residue is E261 (50) or E254, respectively (51). (ii) In Acds acting on small- and medium-chain acyl-CoA thioesters (small-chain acyl-CoA dehydrogenase [SCAD] and medium-chain acyl-CoA dehydrogenase [MCAD]) or glutaryl-CoA (GCD), the catalytic glutamate residue is found in a C-terminal loop between helices K and I (indicated by ▼ in Fig. S2 in the supplemental material). For instance, in MCAD, one of the best-investigated characterized Acds, it is E376 (1). No such glutamate residue is found in either position in Acd_DPN7. Instead, the C-terminal K-I loop of Acd_DPN7 contains a glycine (G368). Most interestingly, a glutamine instead of a glutamate is found at position 248.

In a second step, a phylogenetic tree analysis was conducted (Fig. 5). The same sequences that were used for the multiple-sequence alignment were applied. Additionally, AidB, an IVD homologue from E. coli involved in DNA repair, served as an outgroup (52). According to their different substrates, a kind of grouping of the sequences could be observed. Such a relationship between amino acid sequence and function was supposed previously (4). Based on the neighbor-joining tree, Acd_DPN7 is clearly separated from GCDs, IVDs, CDs, and a group of Acds that act on substituted succinyl-CoA thioesters. It clusters with SCADs and MCADs.

Fig 5.

Neighbor-joining tree of amino acid sequences of biochemically characterized FAD-dependent acyl-CoA dehydrogenases. The sequence of AidB was used as an outgroup. This IVD homologue from E. coli is involved in repair of alkylated DNA (52). Bootstrap values are given for 100 bootstrap repetitions. The distance bar corresponds to 0.1% differences between sequences. PDB or GenBank accession numbers are given at the end of the branches. Abbreviations used are as follows: GCD, glutaryl-CoA dehydrogenase (* indicates a nondecarboxylating GCD); 2MBD, 2-methylbutyryl-CoA dehydrogenase; IVD, isovaleryl-CoA dehydrogenase; CD, citronellyl-CoA dehydrogenase; LCD, long-chain acyl-CoA dehydrogenase; IBD, isobutyryl-CoA dehydrogenase; MCAD, medium-chain acyl-CoA dehydrogenase; BDH, butyryl-CoA dehydrogenase; SCAD, short-chain acyl-CoA dehydrogenase; 2MSD, 2-methylsuccinyl-CoA dehydrogenase; N2MSD, naphthyl-2-methylsuccinyl-CoA dehydrogenase; BSD, benzylsuccinyl-CoA dehydrogenase; CBD, crotonobetainyl-CoA dehydrogenase.

Cloning of the Acd gene from A. mimigardefordensis strain DPN7^T (acd_DPN7), overexpression in E. coli BL21(DE3)/pLysS, and purification and characterization of the translational product.

Based on nucleotide sequence data, Acd_DPN7 has a calculated molecular mass of 43.531 Da (isotopically average) with a theoretical pI of 6.73, and it consists of 401 amino acid residues. In this study, Acd_DPN7 was heterologously expressed by using the T7 promoter/polymerase-based expression vector pET23a(+) and E. coli BL21(DE3)/pLysS as a host strain. Thereby, the protein was equipped with an additional C-terminal His₆ tag with 5 plasmid-encoded amino acids between the protein and His₆ tag. Consequently, this protein consisted of 412 amino acid residues and exhibited a theoretical molecular mass of 44.809 kDa and a calculated pI of 6.71. The overexpressed protein was purified by immobilized metal chelate affinity chromatography to electrophoretic homogeneity (see Fig. S4 in the supplemental material). Afterwards, Acd_DPN7 was applied to analytical size-exclusion chromatography. It revealed an apparent native molecular mass of 173 kDa. This indicates a homotetrameric structure of the protein which is in good accordance with the theoretical molecular mass of 179.2 kDa.

Spectroscopic properties and cofactor analysis of recombinant Acd_DPN7.

The bright yellow appearance of purified Acd_DPN7 strengthened the assumption that FAD is bound to the protein. More specifically, the absorption spectrum of Acd_DPN7 showed two absorption maxima at 368 nm and 446 nm, suggesting the presence of a flavin cofactor (Fig. 6). The A₂₆₈/A₃₆₉/A₄₄₄ ratio was 4.86:0.74:1.00.

Fig 6.

UV-visible spectrum of purified recombinant Acd_DPN7 in the range of 245 to 800 nm. A total of 1,731 μg Acd_DPN7 was applied in 1 ml 10 mM HEPES (pH 7.4) with a final concentration of 9.65 μM referring to the tetramer (38.6 μM monomer). Measurements were done aerobically at 25°C. The A₂₆₈/A₃₆₉/A₄₄₄ ratio was 4.86:0.74:1.00. AU, arbitrary units.

Purified Acd_DPN7 was considered to be in its oxidized form, as purification occurred under aerobic conditions. Purified Acd_DPN7 was denatured, and the supernatant was analyzed by HPLC-ESI-MS. The detection of a substance at m/z 786, corresponding to positively charged FAD, clearly verified the release of FAD from the enzyme (see Fig. S5 in the supplemental material). Neither flavin mononucleotide (m/z 456) nor riboflavin (m/z 377) was detectable. As FAD was copurified, it seems to be tightly bound to the enzyme.

Identification of reaction products.

It was a main task to verify that 3SP-CoA is the substrate of Acd_DPN7 and to identify the reaction products. As 3SP-CoA was not commercially available, it was synthesized in situ in the reaction mixture by purified succinate-CoA ligase (SucCD), as described in Materials and Methods. Afterwards, purified Acd_DPN7 was added to the reaction mixture. Samples were withdrawn before and after addition of SucCD and after addition of Acd_DPN7 (Fig. 7). 3SP-CoA was detectable in the reaction mixture only after addition of purified SucCD (Fig. 7A and B). Upon addition of purified Acd_DPN7 and incubation for 10 min at 37°C, 3SP-CoA was completely consumed, and a new substance at m/z 824 appeared (Fig. 7C and D). The mass spectrum identified this substance as propionyl-CoA (Fig. 7E). AMP, with a retention time of 13.7 min, was identified as an impurity (m/z 348). The increase in the concentration of AMP (Fig. 7B) is explainable by the decomposition of ATP during heat inactivation of SucCD prior to addition of Acd_DPN7. The stronger formation of AMP from ATP shown in Fig. 7C is most probably due to acidification with trichloroacetic acid.

Fig 7.

Identification of propionyl-CoA as a product of the Acd_DPN7 reaction. In order to detect reaction products of Acd_DPN7, a coupled enzyme assay applying succinate-CoA ligase for in situ formation of 3SP-CoA was performed. Samples were withdrawn and analyzed by HPLC-ESI-MS as described in Materials and Methods. (A) Reaction mixture containing 1 mM ATP, 10 mM MgCl₂, 1 mM CoA, and 5 mM 3SP in 50 mM Tris-HCl (pH 7.4) before addition of succinate-CoA ligase (SucCD). AMP was identified (m/z 348) (data not shown) as an impurity with a retention time of 13.7 min. The substance with a retention time of 18.2 min was identified as a CoA dimer (m/z 767) (data not shown). (B) After addition of SucCD and incubation for 10 min at 37°C, the enzyme was heat inactivated (10 min at 95°C). As expected, the putative substrate of Acd_DPN7, 3SP-CoA (retention time, 15.7 min), was formed by SucCD. (C) Subsequently, Acd_DPN7 was added, and the reaction mixture was incubated at 37°C for 10 min. The reaction was stopped by addition of trichloroacetic acid (TCA). 3SP-CoA was completely consumed within the reaction. Another substance with a retention time of 24.0 min was observed. (D) MS spectrum analysis of data in panel C. All ions at m/z 824 ± 0.5 corresponding to propionyl-CoA were detected. (E) Mass spectrum at a retention time of 24.0 min confirmed the formation of propionyl-CoA. The sodium salt of propionyl-CoA gives m/z 842.2. RT, retention time; NL, normalization level.

It was assumed that the sulfino group of 3SP-CoA is released as sulfite. A method applying DTNB for sulfite determination in aqueous solutions was adapted to identify this reaction product (42). As a reference, 0.2 mM DTNB was mixed with an excess of sodium sulfite (final concentration, 1 mM) to form 2-nitro-5-thio-sulfobenzoic acid (NTSB) and was incubated for 30 min at 25°C. Finally, the solution appeared bright yellow due to the formation of 2-nitro-5-thiobenozoic acid (NTB) (see Fig. S6 in the supplemental material). In advance, it was shown that DTNB reacted with sulfite but showed no activity toward sulfate (data not shown). Subsequently, the NTSB formed was analyzed via HPLC-ESI-MS (Fig. 8). One distinct peak with a retention time of 13.9 min was detected (Fig. 8A2). The corresponding mass spectrum showed a major m/z 277.8, which is well in accordance with the calculated mass-to-charge ratio (Fig. 8B2). In a second step, Acd_DPN7 was assessed for the formation of sulfite from 3SP-CoA. Incubation of 0.2 mM DTNB with 0.2 mM 3SP-CoA at 30°C alone did not result in an increase of the absorption at 412 nm. Only after the addition of 10 μg/ml of purified Acd_DPN7 did the absorption at 412 nm increase. After 10 min of incubation, 50 μl of the reaction mixture was analyzed by HPLC-ESI-MS. The analysis verified the formation of NTSB in the Acd_DPN7 assay (Fig. 8A1 and B1). Formation of NTSB is possible only if sulfite is released by Acd_DPN7.

Fig 8.

Identification of sulfite as a reaction product. (A) UV chromatogram at 412 nm. (1) The reaction mixture contained 0.2 mM 3SP-CoA, 0.2 mM DTNB, and 10 μg/ml of purified Acd_DPN7. After incubation at 30°C for 10 min, 50 μl of the reaction mixture was analyzed by HPLC-ESI-MS. (2) A reaction mixture containing 0.2 mM DTNB and an excess of sodium sulfite (1 mM) was incubated for 30 min at 25°C. Fifty microliters of the reaction mixture was analyzed by HPLC-ESI-MS. The so-formed NTSB (retention time, 13.9 min) served as a reference substance. (B) Corresponding ESI-MS mass spectra. (1) Mass spectrum of the product in the reaction mixture containing 3SP-CoA, DTNB, and purified Acd_DPN7. (2) Mass spectrum of the NTSB reference with a retention time of 13.9 min. ITMS, ion trap mass spectrometer.

Catalytic properties of recombinant 3SP-CoA desulfinase.

The detection of sulfite was employed in a continuous spectrophotometric assay of recombinant Acd_DPN7. The reaction was monitored by the rapid and stoichiometric formation of 1 mol NTB (ϵ_{412 nm} = 14,150 M⁻¹ cm⁻¹) per mol sulfite (42, 43). Acd_DPN7 had a V_max of 4.19 μmol min⁻¹ mg⁻¹ and showed an apparent K_m of 0.013 mM for 3SP-CoA (Table 3).

Table 3.

Kinetic characterization of desulfinating enzymes^a

Enzyme	Molecular mass (subunit) (kDa)	Substrate	Vmax (μmol min−1 mg−1)	Km (mM)	kcat (s−1)	kcat/Km (s−1 mM−1)	Reference
AcdDPN7	44.809	3SP-CoA	4.191	0.013	3.13	240.8	This study
CSDE. colic	43.238	Cysteine sulfinic acid	20.00	0.24	15	63	53
AspATI33Q/Y214Q/R280Yc	43.500	Cysteine sulfinic acid	0.690b	0.1	0.50	5.00	54
CystalysinT. denticolac	46.258d	Cysteine sulfinic acid	115.4b	49	89	1.82b	55
DszBKA2-5-1	40.000	2-Phenylbenzene sulfinate	0.155b	0.062	0.10	1.66b	56
DszBKA2-5-1	40.000	2′-Hydroxybiphenyl-2-sulfinate	0.185b	0.0082	0.12	15.00b	56

^aCSD_{E. coli} is cysteine sulfinate desulfinase from E. coli. AspAT_{I33Q/Y214Q/R280Y} is a cysteine sulfinate desulfinase engineered on the aspartate amino transferase (AspAT) scaffold from E. coli. Cystalysin_{T. denticola} is an aspartate aminotransferase-like enzyme isolated from the oral pathogen Treponema denticola. DszB_KA2-5-1 is a 2′-hydroxybiphenyl-2-sulfinate desulfinase from Rhodococcus erythropolis strain KA2-5-1.

^bKinetic parameter has been calculated based on values available in the literature.

^cPyridoxal 5′-phosphate dependent.

^dMolecular mass was reported in reference 57.

A ferrocenium assay (44) was applied to examine whether Acd_DPN7 shows dehydrogenation activity toward several acyl-CoA thioesters. No decreases in the absorption at 300 nm were observed with either acyl-CoA thioesters like propionyl-, butyryl-, isobutyryl-, valeryl-, or isovaleryl-CoA or with CoA thioesters of dicarboxylic acids like succinyl-CoA or glutaryl-CoA. Moreover, no decrease in absorption at 300 nm was detected with 3SP-CoA. For all assay conditions, samples were subsequently analyzed by HPLC-ESI-MS. No dehydrogenation products could be detected in any of the samples (data not shown).

DISCUSSION

Identification of Acd_DPN7 as the enzyme responsible for desulfination of 3SP-CoA.

Inactivation of acd_DPN7 by transposon insertion or gene deletion as well as complementation experiments with pBBR1MCS-5::acd_DPN7 clearly demonstrated the essential involvement of Acd_DPN7 in DTDP degradation and more specifically in the degradation of the metabolite 3SP in A. mimigardefordensis. In a previous study, we showed that 3SP is activated to 3SP-CoA by a succinate-CoA ligase (22). In this study, experiments with heterologously expressed and purified enzyme unequivocally verified that Acd_DPN7 catalyzes the transformation of 3SP-CoA to propionyl-CoA and sulfite.

Complementation of the A. mimigardefordensis Δacd strain with pBBR1MCS-5::acd_DPN7.

Growth of the A. mimigardefordensis Δacd strain harboring pBBR1MCS-5::acd_DPN7 was rescued on 3SP but not on DTDP. This might be explainable due to the less effective expression of acd_DPN7 under the control of the lac promoter in pBBR1MCS-5::acd_DPN7. Consequently, this could lead to an accumulation of 3MP, the first intermediate during DTDP degradation. This thiol compound has been shown to completely inhibit growth of Ralstonia eutropha H16 at comparably low concentrations of 9 mM in the culture medium (0.1%, wt/vol) (16) and might also suppress growth of the A. mimigardefordensis Δacd mutant on DTDP. In contrast to this, 3SP has been shown to be nontoxic to cells of A. mimigardefordensis DPN7^T at concentrations of up to 100 mM (2%, wt/vol) (22; C. Meinert, unpublished results). Thus, cells cultivated with 3SP have more time in which a sufficient amount of Acd_DPN7 can be expressed to sustain growth.

3SP-CoA desulfinase: differences from and similarities with other members of the Acd superfamily.

In general, Acd_DPN7 shows a strong relationship to the Acd superfamily. Several amino acids strictly conserved throughout the Acd superfamily have been identified in Acd_DPN7 in a multiple-sequence alignment (see Fig. S2 in the supplemental material). Phylogenetic tree analyses indicated a more distant relationship to GCD than expected and implied that short- or medium-chain acyl-CoA thioesters are possible substrates of Acd_DPN7 (see Fig. S2 in the supplemental material). Furthermore, FAD was identified as a cofactor (see Fig. S5 in the supplemental material), and the 3SP-CoA thioester, rather than 3SP itself, is the substrate. In contrast to most other members of this family, no dehydrogenation activity with any of the tested typical substrates of Acds (propionyl-, butyryl-, isobutyryl-, valeryl-, isovaleryl-, or glutaryl-CoA) was observed. Furthermore, no decrease in absorption at 300 nm occurred with succinyl-CoA, a structural analogue of 3SP-CoA, or with 3SP-CoA itself. HPLC-ESI-MS analyses verified that all potential substrates, with the exception of 3SP-CoA, were not transformed. The absence of dehydrogenase activity is most probably caused by the absence of a catalytic glutamate residue in either of the two positions conserved throughout the Acd superfamily. Most interestingly, a glutamine was found in a position corresponding to the active glutamate residue of IVDs or LCADs (1, 50, 51). In previous studies, the exchange of glutamate for glutamine has repeatedly been applied to identify glutamate as the catalytically active amino acid residue (50, 51). For instance, an E367Q mutant of SCAD from Megasphaera elsdenii showed only 0.03% residual enzyme activity (58). Referring to the corresponding nucleotide sequence of acd_DPN7, the presence of glutamine instead of glutamate at position 246 could be the result of a single-nucleotide exchange (change from GAA, coding for glutamate, to CAA, coding for glutamine). Hence, Q246 prompts to the affiliation of Acd_DPN7 with the IVD group (active glutamate in helix H) rather than with the other group, which carries the active glutamate in a loop between helices K and I. In contrast to this, Acd_DPN7 clusters with SCADs and MCADs, which belong to the other group regarding the position of the active glutamate residue, by phylogenetic analysis. Thus, Acd_DPN7 displays features of both groups.

Concerning the substrate and the catalyzed reaction, a relationship to GCD is noticeable. To some extent, GCD represent uncommon members of the Acd superfamily, as they catalyze not only the introduction of a double bond but also the oxidative decarboxylation of the γ-carboxylate of glutaconyl-CoA, yielding crotonyl-CoA and CO₂ (Fig. 1) (23, 59). This reaction proceeds in two steps, as demonstrated by an E370Q variant of human glutaryl-CoA dehydrogenase (GCD_Human) that was unable to catalyze the initial dehydrogenation of glutaryl-CoA to glutaconyl-CoA but was still able to catalyze the decarboxylation of glutaconyl-CoA (23). Recently, a structurally related but nondecarboxylating GCD in the obligately anaerobic bacterium Desulfococcus multivorans (60) was characterized. The results led to the assumption that decarboxylating and nondecarboxylating capabilities are provided by distinct structurally conserved amino acid residues that surround the carboxylate group of the glutaconyl-CoA intermediate (61). In all GCDs, decarboxylating and nondecarboxylating, an invariant arginine residue (R94 in GCD_Human) is involved in binding of the terminal carboxylate and right substrate positioning (60–64). Three residues (E87, S95, and Y369, according to GCD_Human numbering), conserved only in decarboxylating GCDs, weaken the binding between R94 and the substrate carboxylate and enable the decarboxylating reaction. Furthermore, R94 and Y369 might form a transient guanidinium-phenolate-CO₂ complex that would increase the leaving-group potential of CO₂ (64). Like glutaconyl-CoA, 3SP-CoA is a substrate with a terminal, negatively charged moiety. The main difference between the carboxy and sulfino groups is the replacement of a carbon by a sulfur atom. In the multiple-sequence alignment (see Fig. S2 in the supplemental material), R84 in Acd_DPN7 attracts attention, as it is in almost the same position as in GCDs. Probably, this arginine residue binds to the sulfino group in a similar way. The three amino acid residues responsible for decarboxylation are absent in Acd_DPN7. Instead, an alanine residue is found in the position corresponding to E87, an isoleucine is found in the position corresponding to S95, and an alanine is found in the position corresponding to Y369. This might also explain why no decarboxylation was detectable with succinyl-CoA or glutaryl-CoA as a substrate.

In summary, Acd_DPN7 represents a structural hybrid within the Acd superfamily, while the presence of a glutamine instead of the usual active-site glutamate may explain the absence of Acd activity in Acd_DPN7.

We suppose that even the introduction of a double bond would not result in sulfite abstraction from 3SP-CoA in an autocatalytic manner. The detection of 3SDHP, both in the culture supernatant of the Tn5::mob-induced mutant and in the culture supernatant of the Δacd defined deletion mutant, corroborated this assumption. We propose that this compound results from dehydrogenation of 3SP-CoA by other Acds or of 3SP by succinate dehydrogenase after prolonged incubation due to substrate inspecificity. However, introduction of a double bond is therefore unlikely to be the initial mechanistic step of the desulfinase reaction of Acd_DPN7.

Is the reaction independent of an electron acceptor?

As described previously by Westover et al. (65), the ratio of the absorbance in the 270 nm region to the absorbance at 447 nm can serve to estimate the protein/FAD ratio. In that study, a ratio of A₂₆₉ to A₃₆₉ to A₄₄₇ of 5.4:0.62:1.0 for wild-type glutaryl-CoA dehydrogenase and a ratio of A₂₇₆ to A₃₇₆ to A₄₄₇ of 16.8:1.0:1.0 for a mutant enzyme, which was almost completely FAD free, were reported. Hence, we expect all FAD-binding sites (¹⁵⁴KYWIT¹⁵⁸) of Acd_DPN7 with A₂₆₈/A₃₆₉/A₄₄₄ ratios of 4.86:0.74:1.00 to be occupied.

The affinity to and the binding of FAD have an essential influence on proper folding and stability, as shown by site-directed mutagenesis experiments with amino acids involved in FAD binding in MCAD (66). These findings are supported by a previous study that showed a lower thermal stability of MCAD variants affected in FAD-binding residues (67). Recently, the positive effect of an increased concentration of FAD on the proper folding of MCAD, SCAD, and GCD was reported by Lucas et al. (68). Furthermore, FAD is stabilizing the structure of Acds by extensive interactions with amino acid residues of both monomers that form a dimer within the tetrameric structure (1, 7). However, FAD seems not to be involved in the reaction of Acd_DPN7 with 3SP-CoA, as the reaction is independent of an artificial electron acceptor. This indicates that FAD is not converted to FADH₂ during the desulfination reaction and consequently needs not to be regenerated by an (artificial) electron acceptor. In contrast with GCD, where dehydrogenation occurs prior to decarboxylation (23), no such introduction of a double bond was observed for the reaction product propionyl-CoA. Consequently, an involvement of FAD as a hydride acceptor in the reaction is not likely. Nonetheless, the 2′-hydroxy group of the ribityl side chain of the FAD and the main-chain NH group of E376 are within hydrogen bond distance to the carbonyl oxygen of the CoA thioester bond in MCAD, as revealed by the crystal structure (7). This mode of substrate binding seems to be a general feature of all described crystal structures of Acds (1). Engst et al. showed previously that the 2′-OH group not only promotes the initial αC-H deprotonation in MCAD but also supports proper substrate orientation (69). Hence, we propose that the cofactor is stabilizing the quaternary structure of Acd_DPN7 and is necessary for correct positioning of 3SP-CoA. Referring to catalytic efficiency, this might provide an advantage in comparison to other desulfinases, as explained further below. Nonetheless, a transient transfer of electrons between FAD and substrate during the reaction process cannot yet be excluded.

Comparison with other desulfinating enzymes.

Until now, only three enzymes that catalyze desulfination reactions have been known (Table 3). One is the pyridoxal-phosphate (PLP)-dependent cysteine sulfinate desulfinase (CSD) from E. coli, which has selenocysteine lyase and cysteine desulfurase side activities (53). The second desulfinating enzyme known so far is aspartate β-decarboxylase (EC 4.1.1.12), which can accept cysteine sulfinates as well and catalyzes the desulfination as a side reaction (54). This enzyme reaction is also PLP dependent. Acd_DPN7 does not show the typical absorption maximum at ∼410 nm of enzyme-bound PLP (70) and therefore has no such PLP cofactor. As 3SP, in contrast to cysteine sulfinate, has no amino group necessary for a reaction with PLP, the underlying mechanism can be excluded for Acd_DPN7. The third known enzyme, DszB (EC 3.13.1.3), catalyzes the hydrolysis of a sulfinate group from 2′-hydroxybiphenyl-2-sulfinate as well as from 2-phenylbenzene sulfinate (Table 3) (56). Here, no additional cofactor or prosthetic group is required. A cysteine residue is involved in desulfination (71). In the first reaction step, the sulfinate sulfur and the cysteine thiolate ion form a thiolsulfonate-like intermediate. This negatively charged intermediate is stabilized by an arginine residue and by a hydrogen bond to the nitrogen of a glycine residue. Addition of H₂O released sulfite from the intermediate in the second step.

Multiple-sequence alignments indicate that Acd_DPN7 has no structural relationship to DszB-like enzymes (data not shown). Acd_DPN7 possesses only three cysteine residues. Experiments to investigate whether these cysteine residues are involved in the reaction are currently in progress to elucidate the underlying reaction mechanism.

Nevertheless, Acd_DPN7 shows a relatively high k_cat/K_m value in comparison to other desulfinases (Table 3). Here, a comparably high catalytic activity is paired with a reasonably high substrate affinity. The latter is most probably caused by the CoA moiety of 3SP-CoA. As mentioned above, several amino acids, conserved throughout the Acd superfamily, and the FAD cofactor interact with the CoA moiety and ensure a correct positioning of the substrate. This might be a benefit of the preceding activation of 3SP to a CoA thioester (22). Such an activation has not been observed for any substrate of the other desulfinases.

Conclusions.

In previous studies, all preceding metabolic steps of DTDP degradation were identified and were clearly allocated to the corresponding gene products applying heterologously expressed and purified enzymes (18–22). This study confirmed the final step in the previously proposed pathway of DTDP catabolism in A. mimigardefordensis strain DPN7^T (18). Multiple-sequence alignments, phylogenetic tree analyses, and cofactor analysis provide conclusive evidence that Acd_DPN7 is a member of the acyl-CoA dehydrogenase superfamily. As clearly demonstrated in this study, the enzyme is not active as a dehydrogenase. Instead, Acd_DPN7 performs a reaction never observed for the Acd superfamily: desulfination.