Novel Acetone Metabolism in a Propane-Utilizing Bacterium, Gordonia sp. Strain TY-5

Abstract

In the propane-utilizing bacterium Gordonia sp. strain TY-5, propane was shown to be oxidized to 2-propanol and then further oxidized to acetone. In this study, the subsequent metabolism of acetone was studied. Acetone-induced proteins were found in extracts of cells induced by acetone, and a gene cluster designated acmAB was cloned on the basis of the N-terminal amino acid sequences of acetone-induced proteins. The acmA and acmB genes encode a Baeyer-Villiger monooxygenase (BVMO) and esterase, respectively. The BVMO encoded by acmA was purified from acetone-induced cells of Gordonia sp. strain TY-5 and characterized. The BVMO exhibited NADPH-dependent oxidation activity for linear ketones (C₃ to C₁₀) and cyclic ketones (C₄ to C₈). Escherichia coli expressing the acmA gene oxidized acetone to methyl acetate, and E. coli expressing the acmB gene hydrolyzed methyl acetate. Northern blot analyses revealed that polycistronic transcription of the acmAB gene cluster was induced by propane, 2-propanol, and acetone. These results indicate that the acmAB gene products play an important role in the metabolism of acetone derived from propane oxidation and clarify the propane metabolism pathway of strain TY-5 (propane → 2-propanol → acetone → methyl acetate → acetic acid + methanol). This paper provides the first evidence for BVMO-dependent acetone metabolism.

Gordonia sp. strain TY-5 is an actinomycete that is capable of aerobic growth on gaseous propane as a carbon and energy source. Our previous study showed that propane is oxidized to 2-propanol by monooxygenase-mediated subterminal oxidation and then 2-propanol is further metabolized to acetone by three distinct NAD⁺-dependent secondary alcohol dehydrogenases (27). Although n-alkanes can be metabolized to the corresponding ketones through subterminal oxidation in some bacteria, microbial metabolism of the downstream ketone is poorly understood (1, 2, 22, 36, 46, 47). A variety of bacteria, including Gordonia sp. strain TY-5, are able to utilize acetone as a source of carbon and energy. Previous studies on bacterial acetone metabolism both in vivo and in vitro suggested that acetone can be metabolized in two ways. In most aerobic bacteria, acetone was hydroxylated by an O₂-dependent reaction producing acetol (hydroxyacetone), although the corresponding enzyme system is not known (13, 29, 42, 45). For most anaerobes (and some aerobes), acetone undergoes CO₂-dependent carboxylation, yielding acetoacetate. Recently, the enzyme responsible for the latter reaction, acetone carboxylase (EC 6.4.1.6), has been purified and characterized (3, 12, 15, 39-41).

This study was conducted to characterize acetone metabolism in propane-utilizing Gordonia sp. strain TY-5 at the enzymatic and gene levels. We first identified two acetone-induced proteins and cloned their corresponding genes, acmA and acmB. Subsequently, we showed that acetone is oxidized to methyl acetate by a novel Baeyer-Villiger monooxygenase (BVMO) (acmA gene product) and that the methyl acetate produced was hydrolyzed to acetate and methanol by an esterase (acmB gene product). This study provides the first evidence for monooxygenase-dependent acetone oxidation and sheds light on the poorly understood microbial pathway of acetone oxidation.

MATERIALS AND METHODS

Bacterial strains, culture conditions, and vectors.

Gordonia sp. strain TY-5, which was previously isolated from a soil sample (27), was used in this work. Strain TY-5 was grown on AY medium containing (per liter) 4 g of K₂HPO₄, 2 g of KH₂PO₄, 2 g of NH₄Cl, 0.2 g of MgSO₄ · 7H₂O, 4 mg of CaCl₂ · 2H₂O, 5 mg of H₃BO₄, 0.4 mg of CuSO₄ · 5 H₂O, 1 mg of KI, 2 mg of FeSO₄ · 7H₂O, 4 mg of MnSO₄ · 4 to 7H₂O, 4 mg of ZnSO₄ · 7H₂O, 1 mg of Na₂MoO₄ · 2H₂O, pH 7.0. When propane was used as a carbon source, a 500-ml shaking flask containing 100 ml of AY medium under a gas mixture (propane-air ratio = 3:7) and sealed with a butyl rubber stopper was shaken at 30°C. Strain TY-5 was also grown in acetone (0.1%, vol/vol), 2-propanol (1.0%, vol/vol), or sodium citrate (1.0%, wt/vol) at 30°C under shaking conditions. Large-scale (10-liter) cultures were performed in a 15-liter jar fermentor (Mitsuwa KMJ-15B; Mitsuwa Co. Ltd., Osaka, Japan) at 30°C and 300 rpm.

Escherichia coli DH5α (TaKaRa, Kyoto, Japan) was used for gene cloning and was usually grown on Luria-Bertani broth (pH 7.0) which contained 1% Bacto Tryptone (Difco Laboratories, Detroit, MI), 0.5% Bacto Yeast Extract (Difco Laboratories), and 0.5% NaCl, in the presence of ampicillin (50 mg/liter) when necessary. pGEM-T Easy (Promega, Madison, WI) was used as a cloning vector. E. coli Rosetta(DE3) (Novagen, Madison, WI) and the T7 expression vector pET-23a(+) (Novagen) were used for gene expression.

Induction of acetone-metabolizing activity.

Gordonia sp. strain TY-5 was grown for 36 h on citrate as described above. Cells were harvested by centrifugation (15,000 × g for 5 min at 4°C) and washed with ice-cold AY medium without a carbon source. The resultant cells were suspended in AY medium containing acetone (0.25%, vol/vol) to 50 U of optical density at 610 nm (OD₆₁₀) and shaken at 30°C. The time-dependent consumption of acetone was monitored by gas chromatography.

Cell extracts.

Citrate- or acetone-incubated cells were harvested and washed with ice-cold 20 mM KH₂PO₄-Na₂HPO₄ buffer (pH 7.5) containing 1 mM dithiothreitol, 1 mM EDTA, and 5% (vol/vol) glycerol (buffer A). The cells were suspended in the same buffer to an OD₆₁₀ of ca. 80 and disrupted by six passages through a High Pressure Laboratory Homogenizer (MINI-LAB type 8.3 H; Rannie a/s, Copenhagen, Denmark) at 80 MPa, followed by centrifugation at 10,000 × g for 15 min at 4°C. Centrifugation of the crude extract at 120,000 × g for 1 h at 4°C yielded a clear supernatant, which was used as the cell extract.

Analytical methods.

Gas chromatography was performed with a Shimadzu GC-14B (Shimadzu Co. Ltd., Kyoto, Japan) equipped with a flame ionization detector and a stainless column (200 cm long by 3.0 mm [inside diameter]) packed with Porapak Q (Shimadzu Co). N₂ was used as the carrier gas (50 ml/min), the injector and detector temperature was 250°C, and the oven temperature was 150°C. Protein was measured with a Bio-Rad protein assay kit (Japan Bio-Rad Laboratories, Tokyo, Japan) with bovine serum albumin as the standard (6). Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) was performed with a 10% polyacrylamide gel by the method of Laemmli (28). Prestained protein markers (low range) for SDS-PAGE (Nacalai Tesque, Inc., Kyoto, Japan) were used as the molecular standards. The relative molecular mass of the native enzyme was determined by gel filtration with a fast protein liquid chromatography system (Amersham Bioscience, Piscataway, N.J.) with a Superose 6 10/300 GL column (1.0 by 30 cm; Amersham Bioscience) preequilibrated with 50 mM KH₂PO₄-Na₂HPO₄ buffer (pH 7.5) containing 1 mM dithiothreitol, 1 mM EDTA, 5% (vol/vol) glycerol, and 0.1 M KCl. The standard proteins used were the products of the Oriental Yeast Co., Ltd. (Tokyo, Japan).

Two-dimensional electrophoresis and N-terminal sequencing.

Polyacrylamide (7.5%) gel electrophoresis without SDS or 2-mercaptoethanol (native PAGE) was performed with a disc gel (60 mm long by 5.0 mm [inside diameter]) as one-dimensional electrophoresis. Cell extract containing 1 mg of total protein was loaded into each lane of the polyacrylamide gel. After the native PAGE, the gel was equilibrated for 30 min with buffer containing 50 mM Tris-HCl (pH 6.8), 2% SDS, and 5% 2-mercaptoethanol. For the second dimension, SDS-PAGE was performed with a 1-mm-thick gel. In order to determine the N-terminal amino acid sequence, protein from the SDS-PAGE gel was electroblotted onto a P^sqPVDF membrane (Millipore Corp., Bedford, Mass) at 15 V for 50 min with a transfer buffer containing 10 mM CAPS-NaOH (pH 11) and 10% (vol/vol) methanol. The membrane was stained with Coomassie brilliant blue R-250 (Nacalai Tesque), desirable protein spots were excised, and the amino acid sequences were determined by Edman's method with a Perkin-Elmer 476A protein sequencer.

Cloning and nucleotide sequencing of a gene cluster encoding proteins involved in acetone metabolism.

The primers used in this work are listed in Table 1. To amplify DNA fragments encoding the BVMO, cassette ligation-mediated PCR (23) with a SalI cassette (TaKaRa) was performed. Nested degenerate primers (primers N1 and N2) designed from the N-terminal amino acid sequence were used as the upstream primers, and cassette primers (TaKaRa) were used as the downstream primers. Chromosomal DNA extracted from Gordonia sp. strain TY-5 with the AquaPure genomic DNA isolation kit (Bio-Rad Laboratories, Hercules, CA) was digested with SalI and then ligated with the SalI cassette. The ligation mixture was used as the template for amplification of a portion of the gene cluster by PCR with Ex Taq polymerase (TaKaRa). The amplified 600-bp fragment was gel purified with MagExtractor (Toyobo, Osaka, Japan) and then ligated with pGEM-T Easy and sequenced. To clone the full-length gene cluster, an inverse PCR (34) and a cassette ligation-mediated PCR were performed. Chromosomal DNA from strain TY-5 was digested with SacII or PstI and then self-ligated. The ligation mixtures were then used as templates for inverse PCR amplification with LA Taq polymerase (TaKaRa) with primers inv-1 and inv-2 (for SacII digestion) or primers inv-3 and inv-4 (for PstI digestion). The amplified 840-bp (primers inv-1 and inv-2) and 2.6-kb (primers inv-3 and inv-4) fragments were ligated with pGEM-T Easy and sequenced. Chromosomal DNA extracted from strain TY-5 was digested with EcoRI and then ligated with an EcoRI cassette (TaKaRa). The ligation mixture was used as the template for cassette ligation-mediated PCR amplification with LA Taq polymerase with primer clm-1, primer clm-2, and cassette primers. The amplified 1.6-kb fragment was ligated with pGEM-T Easy and sequenced. DNA sequencing was performed by the dideoxy-chain termination method (38) with a Thermo Sequenase Primer Cycle Sequencing Kit (Amersham Biosciences) and a DSQ-2000L DNA sequencer (Shimadzu Co. Ltd.).

TABLE 1.

Primers used in this study

Primer	Sequence
N1a	5′-ACSYTSGACGCSGCSGTSATHGGNAC-3′
N2a	5′-GTCGCSGGCYTSTACGAGYTVNANATG-3′
inv-1	5′-AGGTCATCCAGTCGATTGCCGCCGA-3′
inv-2	5′-TAGGCCCGTACCTCGAGTCCCTGCT-3′
inv-3	5′-GTCACTGTCCAACGACGCGGAATTC-3′
inv-4	5′-ATCCGCCGGTGTGCGTGAGCAATTG-3′
clm-1	5′-GTGCCGCTGGAGACCAACTATCTCG-3′
clm-2	5′-AGGTCTACCACCGCGACAACGTGAC-3′
acmA-f	5′-ATGTCAACAACCACACTCGACGCGG-3′
acmA-r	5′-TCAGGACAGTGCGAAGCCCTTGTAG-3′
acmB-f	5′-ATGACCAGCACATTCTCTTCACTCG-3′
acmB-r	5′-TTAGCTCTTCAGGGCGTCGTAGAAG-3′
orf1-f	5′-ATGTCCGCCTTGATGAAGACCTTCA-3′
orf1-r	5′-TCAGAAGGTGATCAGCGGTTTGACG-3′
NdeI-acmA-fb	5′-ggaattcCATATGTCAACAACCACACTCGACGCGGCC-3′
EcoRI-acmA-rb	5′-gGAATTCTCAGGACAGGACAGTGCGAAGCCCTTGT-3′
NdeI-acmB-fb	5′-ggaattcCATATGACCAGCACATTCTCTTCACTCGAT-3′
HindIII-acmB-rb	5′-cccAAGCTTTTAGCTCTTCAGGGCGTCGTAGAA-3′

^aAbbreviations: N, A, G, C, or T; Y, C or T; S, C or G; H, A, C, or T; D, A, G, or T.

^bAdded NdeI, EcoRI, and HindIII sites are underlined, and additional sequences for efficient cleavage are in lowercase.

Enzyme assays.

The enzyme activity of NADPH-dependent BVMO was determined at 30°C in a reaction mixture (1 ml) containing 20 mM HEPES-NaOH buffer (pH 8.5), 0.4 mM NADPH, 20 mM ketone, and an appropriate amount of enzyme. One unit of enzyme activity was defined as the amount of enzyme that catalyzed the oxidation of 1.0 μmol of NADPH per min. For the stoichiometric study, 0.04 U of the enzyme was used in a 1-ml reaction mixture containing 20 mM HEPES-NaOH buffer (pH 8.5) and 0.4 mM NADPH. The reaction at 30°C was started by addition of 13.5 μmol of acetone. For quantitative analysis, the decrease in NADPH was followed at 340 nm on a Shimadzu spectrophotometer (UV-160). O₂ consumption was monitored with a model 10 oxygen electrode (Rank Bros., Bottisham, Cambridge, United Kingdom). After a 5-min reaction, the amount of methyl acetate produced was determined by gas chromatography.

Purification of NADPH-dependent BVMO.

Enzyme purification was performed at 4°C with a fast protein liquid chromatography system. The cell extract obtained, which contained 643 mg of protein, was applied to a DEAE-Toyopearl 650 column (2.2 by 20 cm; Tosoh, Tokyo, Japan) preequilibrated with buffer A and was then eluted with a linear gradient containing increasing KCl concentrations (0 to 1.0 M) in buffer A. Ammonium sulfate was added to the collected active fraction to 1.2 M, and the protein precipitate was removed by centrifugation (30,000 × g for 30 min at 4°C). The resulting supernatant was applied to a HiPrep 16/10 Phenyl FF column (1.6 by 10 cm; Amersham Biosciences) preequilibrated with buffer A containing 1.2 M ammonium sulfate and was then eluted with a linear gradient containing decreasing ammonium sulfate concentrations (1.2 to 0 M) in buffer A. The active fractions were collected, concentrated by ultrafiltration with Centriprep YM-30 (Millipore, Billerica, MA), and then gel filtered as described above. The active fractions were collected, dialyzed against buffer A containing 0.2 M KCl, and then chromatographed on a MonoQ HR 5/5 column (0.5 by 5 cm; Amersham Biosciences) preequilibrated with buffer A with a linear gradient containing increasing concentrations of KCl (0.2 to 0.7 M). The active fractions were collected and stored at 4°C until used for enzyme characterization.

Expression of acm genes in E. coli.

acmA and acmB were amplified by PCR with chromosomal DNA of Gordonia sp. strain TY-5 as the template with primers NdeI-acmA-f and EcoRI-acmA-r for acmA and primers NdeI-acmB-f and HindIII-acmB-r for acmB. PCR was performed with Pyrobest DNA polymerase (TaKaRa). Each amplified DNA fragment was gel purified, digested with the appropriate restriction enzymes, and cloned into linearized pET-23a(+). The resulting plasmids (pEACMA and pEACMB) were introduced into E. coli Rosetta(DE3). E. coli transformants were grown in Luria-Bertani broth containing ampicillin (50 mg/liter) and chloramphenicol (34 mg/liter) at 37°C to which 0.5 mM IPTG (isopropyl-β-d-thiogalactopyranoside) was added at mid-log phase, followed by an additional 3 h of growth and harvesting.

Whole-cell assay.

Gordonia sp. strain TY-5 was grown on 2-propanol as described above. After 4 days, cells were harvested, washed, and suspended in AY medium without a carbon source to an OD₆₁₀ of 50. The reaction was carried out in a 25-ml sealed culture vessel containing 5 ml of the cell suspension, to which 2-propanol (1.96 mM) was added, at 30°C under shaking conditions. A portion of the medium was sampled through a syringe and used for determination of concentrations of 2-propanol and acetone by gas chromatography.

E. coli cells carrying pEACMA or pEACMB were suspended in 20 mM KH₂PO₄-Na₂HPO₄ buffer (pH 7.5) containing ampicillin (50 mg/liter), chloramphenicol (34 mg/liter), and IPTG (0.5 mM) to an OD₆₁₀ of 50. The reaction was carried out in a 25-ml sealed culture vessel containing 5 ml of each cell suspension, to which acetone (13.4 mM) or methyl acetate (62.8 mM) was added, at 30°C under shaking conditions. A portion of the cell suspension was sampled through a syringe and used for determination of concentrations of acetone and methyl acetate by gas chromatography.

Northern blot analysis.

Northern blot analysis was performed as previously described (27). DNA probes specific for individual genes were generated by PCR with primers acmA-f and acmA-r for acmA, acmB-f and acmB-r for acmB, and orf1-f and orf1-r for orf1. DNA probes were labeled with AlkPhos Direct Labeling Reagents (Amersham Biosciences).

Nucleotide sequence accession number.

The sequence of the gene cluster has been submitted to GenBank and assigned accession number AB252677.

RESULTS

Identification of acetone-induced proteins by two-dimensional electrophoretic analysis.

While citrate-grown cells of Gordonia sp. strain TY-5 did not consume acetone in the reaction mixture, acetone-induced cells (12.5 mg [dry cell weight]) consumed acetone at a rate of 0.22 μmol/min (data not shown). This observation suggested that acetone metabolism was induced by acetone in this bacterium. To obtain further insight into the induction of acetone metabolism, two-dimensional electrophoresis was performed with cell extracts from Gordonia sp. strain TY-5 cells with or without acetone induction. Comparative two-dimensional electrophoresis revealed two major acetone-induced proteins: protein A of ca. 63 kDa and protein B of ca. 45 kDa (Fig. 1). The N-terminal amino acid sequences of these proteins were as follows: protein A, XTTTLDAAVIGTXVAGLYELXEQGXXV; protein B, TSTFSSLDVSAFTSAADRILXEAVTGDARVPG (with X denoting residues that could not be reliably identified).

FIG. 1.

Two-dimensional protein analysis of cell extract from Gordonia sp. strain TY-5 cells grown on citrate (left panel) and induced by acetone (right panel). The acetone-induced proteins (A and B) are indicated by arrows.

Gene structures of acetone-induced proteins.

With degenerate primers designed from the N-terminal sequence of protein A, the nucleotide sequence of a 5.4-kb chromosomal DNA fragment around the protein A coding region was determined (Fig. 2). The fragment harbored three putative open reading frames (ORFs) and one incomplete ORF on the same strand. The most upstream ORF (designated acmA) could code for a protein of 533 amino acids. The N-terminal amino acid sequence deduced from the acmA gene product matched the determined amino acid sequence of protein A. The ORF designated acmB, which was located 43 bp downstream of acmA, could encode a protein of 401 amino acids, and its deduced N-terminal sequence coincided with that determined for protein B. Both acmA and acmB had their own putative ribosomal binding sites.

FIG. 2.

The 5,374-bp SacII-PstI DNA fragment of Gordonia sp. strain TY-5 harboring the acmA and acmB genes. The dashed part of orf2 was not cloned in this study.

A BLAST search against the genome databases revealed that the acmA gene product belongs to a flavoprotein monooxygenase family, which consists of multifunctional flavin-containing monooxygenases, the N-hydroxylating monooxygenases, and BVMOs. The acmA gene product also has significant sequence similarity to several cyclohexanone monooxygenases and phenylacetone monooxygenase (Table 2). In the putative amino acid sequence of acmA, two GXGXX(G/A) Rossmann fold motifs are conserved between amino acid residues 12 to 17 (GTGVAG) and amino acid residues 182 to 187 (GTGSSG). The Rossmann fold motif is a common sequence among FAD- and NAD(P)H-dependent oxidoreductases (16, 48). The motif FXGXXXHXXXW(P/D) was found at amino acid residues 156 to 167 (FGGQLVHTARWP), just before the second Rossmann fold motif. This motif is characteristic to BVMOs and discriminates BVMOs from two other flavoprotein monooxygenase families: flavin-containing monooxygenases and N-hydroxylating monooxygenases (18).

TABLE 2.

Sequence similarity between acetone monooxygenase and Baeyer-Villiger monooxygenase proteins

Proteina	Size (aa)b	Identity (%)	Similarity (%)	Organism	Accession no.	Reference
AcmA	533
PAMO	542	42.0	74.1	Thermobifida fusca YX	CP000088	30
SMO	549	41.5	75.2	Rhodococcus rhodochrous	AB010439	33
CHN	541	38.6	73.5	Rhodococcus sp. strain Phi1	AY123974	10
CHN	540	38.6	71.7	Rhodococcus sp. strain Phi2	AY123973	10
CHN	553	38.5	73.7	Brevibacterium sp. strain HCU	AF257214	8
CHN	537	38.5	72.2	Brachymonas petroleovorans	AY437859	9
CHN	591	37.7	73.2	Arthrobacter sp. strain BP2	AY123972	10
CPN	550	37.0	73.4	Comamonas sp. strain NCIMB9872	AB022102	24
CHN	546	36.8	73.2	Xanthobacter flavus	AJ418061	44
CHN	543	35.6	73.0	Acinetobacter sp. strain SE19	AF282240	11

^aAbbreviations: PAMO, phenylacetone monooxygenase; SMO, steroid monooxygenase; CHN, cyclohexanone monooxygenase; CPN, cyclopentanone monooxygenase.

^baa, amino acids.

The amino acid sequence of the acmB gene product had the highest similarity to the 1,4-butanediol diacrylate esterase of Brevibacterium linens (37) (59% identity and 84% similarity). The SXXK motif is found at amino acid residues 71 to 74 (STTK), and the catalytic tyrosine (Tyr-182) is conserved. These data suggested that the acmB gene product catalyzes hydrolysis of methyl acetate formed from acetone oxidation.

The third ORF (orf1) encodes 354 amino acids, and its putative amino acid sequence had high similarity to the NAD⁺-dependent and zinc-containing alcohol dehydrogenase of Phytomonas sp. strain ADU-2003 (32) (45% identity and 77% similarity). The fourth ORF (orf2) is incomplete, and its partial deduced amino acid sequence shows high similarity to the LuxR-like transcriptional regulator of Rhodococcus opacus 1CP (31).

O₂-dependent acetone consumption in Gordonia sp. strain TY-5.

Although Gordonia sp. strain TY-5 was assumed to metabolize acetone in either an O₂-dependent or a CO₂-dependent manner, the structure of the acetone-induced acmA gene product strongly suggested that acetone was metabolized in the former way. Therefore, we attempted to confirm O₂-dependent acetone oxidation with a whole-cell reaction with Gordonia sp. strain TY-5. An aliquot of the cell suspension was removed periodically, and the concentrations of 2-propanol and acetone were determined by gas chromatography.

Under aerobic conditions (atmospheric gas), the reaction mixture consumed 2-propanol completely within 30 min, together with acetone accumulation (Fig. 3A). Then, after 60 min, the accumulated acetone was consumed. In contrast, Gordonia sp. strain TY-5 cells under anaerobic conditions oxidized 2-propanol to acetone with intracellular NAD⁺ as a cofactor but did not consume acetone (Fig. 3B). When CO₂ was depleted from or added to the reaction vessel, the rate of 2-propanol consumption and the rates of accumulation and consumption of acetone were similar to those under aerobic conditions (data not shown). These experiments show that Gordonia sp. strain TY-5 metabolizes acetone by an O₂-dependent mechanism rather than a CO₂-dependent mechanism.

FIG. 3.

Whole-cell reactions with Gordonia sp. strain TY-5 cells grown on 2-propanol under aerobic (A) and anaerobic (B) conditions. 2-Propanol-grown cells were washed to remove 2-propanol and suspended in a reaction mixture containing 2-propanol (1.96 mM). The concentrations (Conc.) of 2-propanol (circles) and acetone (squares) were determined by gas chromatography.

Purification and characterization of NADPH-dependent acetone monooxygenase.

Acetone-dependent oxidation of NADPH was demonstrated in cell extracts from acetone-induced cells of Gordonia sp. strain TY-5. The protein exhibiting this activity was purified by four column chromatography steps (Table 3). Ultimately, this protein was purified to near homogeneity and no other proteins showing this enzyme activity were detected during the purification procedure. After 40.7-fold purification, the preparation exhibited a faint yellow color, suggesting that the purified enzyme is a flavoprotein. The N-terminal amino acid sequence of the protein was determined to be STTTLDA, which corresponds to the N-terminal sequence of the acmA gene product.

TABLE 3.

Purification of acetone monooxygenase from Gordonia sp. strain TY-5

Purification step	Total activity (U)	Total protein (mg)	Sp act (U/mg of protein)	Yield (%)	Purification (fold)
Cell extract	9.77	643	0.015	100	1
DEAE-Toyopearl	6.47	140	0.046	66.2	3.1
HiPrep 16/10 Phenyl FF	1.22	8.1	0.15	12.5	10.0
Superose 6 10/300 GL	1.21	3.0	0.40	12.4	26.7
MonoQ HR 5/5	0.52	0.85	0.61	5.3	40.7

SDS-PAGE analysis revealed that the molecular mass of the purified protein was 63 kDa. By gel filtration, the native molecular mass of the purified protein was estimated to be 230 kDa, suggesting it to be a homotetramer. The maximum activity was found at pH 8 to 8.5 and 35°C.

The stoichiometry of the acetone-oxidizing reaction was determined with the purified enzyme. When 199.4 ± 2.63 nmol of NADPH was consumed, an equivalent amount of methyl acetate (204.5 ± 5.27 nmol) was produced. When 189.0 ± 13.38 nmol of O₂ was consumed, an equivalent amount of methyl acetate (200.2 ± 4.44 nmol) was produced. Without the addition of acetone, neither O₂ consumption nor NADPH oxidation was detectable (controls). The stoichiometric conversion of acetone to methyl acetate (1:1) was determined by using E. coli cells overexpressing acmA (see below, c.f. Fig. 4). These results indicate that the acmA gene product is a novel BVMO with activity toward acetone, i.e., acetone monooxygenase.

FIG. 4.

Stoichiometric production of methyl acetate from acetone with an E. coli transformant overexpressing the acmA gene. The initial concentration (Conc.) of acetone was 13.4 mM. The concentrations of acetone (squares) and methyl acetate (circles) were determined by gas chromatography.

The substrate specificity of acetone monooxygenase was studied with a variety of cyclic and linear ketones under the conditions described in Materials and Methods with the purified acetone monooxygenase (Table 4). The acetone monooxygenase exhibited higher activities toward cyclobutanone, cyclopentanone, cyclohexanone, 2-octanone, 2-nonanone, and 2-decanone than acetone. NADH could not replace NADPH as a cofactor.

TABLE 4.

Substrate specificities of purified acetone monooxygenase

Substrate	Relative activitya (%)
Linear ketones
Acetone	100
2-Butanone	82.4
2-Pentanone	44.8
2-Hexanone	36.8
2-Heptanone	80.7
2-Octanone	128
2-Nonanone	186
2-Decanone	173
Cyclic ketones
Cyclobutanone	131
Cyclopentanone	170
Cyclohexanone	135
Cycloheptanone	10.5
Cyclooctanone	9.6

^aOne hundred percent relative activity was defined as the activity obtained with acetone as the substrate.

Expression of acmA and acmB in E. coli.

For heterologous expression of acmA and acmB in E. coli, pEACMA and pEACMB were constructed and introduced into E. coli Rosetta(DE3). E. coli transformants carrying pEACMA and pEACMB heterologously expressed proteins with molecular masses of 63 and 45 kDa (estimated by SDS-PAGE), which are close to the theoretical molecular masses of the products of acmA (59,676 Da) and acmB (43,454 Da), respectively (data not shown). No enhanced protein bands of these sizes were detectable in the cell extract from E. coli Rosetta(DE3) carrying the pET-23a(+) vector.

Cell extracts from E. coli carrying pEACMA catalyzed acetone-dependent oxidation of NADPH. To confirm the oxidation of acetone, whole-cell assays were carried out with the cells of the transformant. Cell suspensions of E. coli carrying pEACMA consumed acetone and accumulated an equivalent amount of methyl acetate (Fig. 4). Acetone consumption and methyl acetate accumulation ceased after ca. 4.5 h. In contrast, the cell suspension of E. coli Rosetta(DE3) carrying pET-23a(+) showed neither acetone consumption nor methyl acetate accumulation (data not shown).

To examine whether the enzyme encoded by acmB hydrolyzes methyl acetate, a whole-cell assay with the cells of E. coli carrying pEACMB was performed. Methyl acetate that had been added to the assay vessel was consumed rapidly and disappeared completely within 5 min (data not shown). No methyl acetate was consumed within 30 min in the cell suspension by E. coli Rosetta(DE3) carrying pET-23a(+).

Transcript analysis.

To examine expression of the acm gene cluster in Gordonia sp. strain TY-5, Northern blot analysis was conducted with total RNA prepared from cells grown on acetone and with acmA, acmB, and orf1 as probes. The sizes of the hybridizing bands obtained with the acmA and acmB probes were identical, and the longest transcript corresponded to the entire length of the transcriptional message of acmA and acmB (2.9 kb) (Fig. 5A), suggesting that acmA and acmB are induced as a polycistronic transcriptional unit. It is not known whether the shorter transcripts are functional products, higher-order RNA structures, or nonfunctional degradation products. Although orf1 was also induced by acetone, the size of the hybridizing band obtained with the orf1 probe was not identical to the transcription product of acmA and acmB (Fig. 5A).

FIG. 5.

Northern blot analysis of total RNA from Gordonia sp. strain TY-5. (A) Expression of acmAB and orf1 was detected by using the indicated fragments as hybridization probes. Total RNA was prepared from cells grown on sodium citrate (lanes C) and acetone (lanes A), and 10 μg of RNA was loaded into each lane. (B) Expression of acmAB was detected by hybridization with the acmA fragment. Total RNA was prepared from cells grown on sodium citrate (lane C), propane (lane P), 2-propanol (lane 2-P), and acetone (lane A). Ten micrograms of total RNA was loaded into each lane.

To identify the carbon source(s) that induces the acmAB operon, Northern blot analysis with the acmA gene as a probe was carried out with total RNAs prepared from cells grown with citrate, propane, 2-propanol, or acetone as a carbon source (Fig. 5B). Among the compounds tested, the acmAB operon was induced by propane and metabolites of propane oxidation, i.e., 2-propanol and acetone.

DISCUSSION

Gordonia sp. strain TY-5 can utilize propane as a sole carbon and energy source. Our previous study demonstrated that propane is oxidized to 2-propanol via subterminal oxidation by propane monooxygenase and further oxidized to acetone by secondary alcohol dehydrogenases (27). However, the fate of the acetone was not known. We herein describe the pathway of acetone metabolism in this propane-utilizing bacterium.

Several reactions for acetone metabolism have been reported for other microorganisms, and several metabolites were detected downstream of acetone, e.g., acetoacetate via CO₂-dependent carboxylation and acetol through O₂-dependent hydroxylation (13, 29, 42, 45). In mammals, acetone is oxidized to acetol by cytochrome P450 2E1 (5, 25). In contrast, our present study revealed for the first time that acetone can be oxidized to methyl acetate by acetone monooxygenase.

The physiological significance of acetone monooxygenase (acmA gene product) during propane and acetone metabolism is supported by the following results obtained in this study. (i) The acmA gene was transcribed in response not only to acetone but also to propane and 2-propanol. (ii) Only a single protein exhibiting acetone monooxygenase activity could be purified from acetone-induced cells, and the purified protein was identified as the acmA gene product from its N-terminal amino acid sequence. (iii) Gordonia sp. strain TY-5 requires O₂ to metabolize acetone, and this activity was also induced in 2-propanol-grown cells.

The enzymatic and stoichiometric analyses of the purified acmA gene product and the deduced amino acid sequence of acmA have revealed that acetone monooxygenase is a new BVMO. BVMOs have been found in bacteria and fungi and are classified into two distinct groups, i.e., type 1 and 2 enzymes. Type 1 enzymes have FAD and NADPH as cofactors and consist of a single polypeptide chain or two to four identical subunits. On the other hand, type 2 enzymes have FMN and NADH as cofactors and consist of two distinct types of subunits (49). Since acetone monooxygenase exhibited four identical subunits and could not utilize NADH as a cofactor, this enzyme belongs to the type 1 BVMO family.

Different from the acetone metabolism in Gordonia sp. strain TY-5 as revealed in this study, acetol has been considered to be an intermediate in the metabolism of propane, 1,2-propane-diol, and acetone in many bacteria. On the basis of the biochemical properties of acetol monooxygenase from Mycobacterium sp. strain Py1, this enzyme appears to be a BVMO (20). Therefore, it may be noteworthy to point out that two BVMOs are involved in two distinct catabolic pathways for acetone.

Because of their wide substrate spectrum and stereoselectivity, BVMOs are attractive biocatalysts for industrial applications (7, 8, 10, 14, 17, 19, 21, 24, 26, 43, 49). Although some reports describe BVMOs that can oxidize linear ketones to the corresponding esters (7, 10, 17, 19), BVMOs reactive toward acetone have not been described previously. Hence, acetone monooxygenase from Gordonia sp. strain TY-5 is a unique enzyme not only in its physiological significance but also in its catalytic properties. Functional expression of acmA in E. coli will facilitate its use as a biocatalyst in the oxidation of various cyclic and linear aliphatic ketones (Table 4), including acetone.

Analyses of the acmB gene product and its expression strongly suggest that methyl acetate formed by acetone monooxygenase is further metabolized to acetic acid and methanol by methyl acetate hydrolase, the acmB gene product. This is supported by the observations that (i) the acmB gene constitutes an operon with acmA; (ii) the acmAB operon is transcribed and regulated polycistronically; (iii) the acmB product, which exhibits high similarity to esterases, was found in cell extract from acetone-induced cells; and (iv) the cells of E. coli heterologously expressing the acmB gene hydrolyzed methyl acetate.

Both the acmA and acmB gene products were initially identified as major acetone-induced proteins in Gordonia sp. strain TY-5. However, the acmAB operon was inducibly expressed not only by acetone but also by propane and 2-propanol. This strongly suggests that both acmA and acmB are responsible for propane metabolism in Gordonia sp. strain TY-5.

In summary, the present results support our previous study, which showed that propane is oxidized by monooxygenase-mediated subterminal oxidation and that 2-propanol is oxidized to acetone in Gordonia sp. strain TY-5 (27). Various types of 2-propanol metabolism have been implicated in propane-utilizing bacteria (2, 4, 22, 35, 45, 48, 50). However, this research was the first to reveal the subterminal oxidation pathway of propane shown in Fig. 6. This strain can grow on sodium acetate; however, methanol cannot support the growth of this bacterium. Hence, the acetate moiety of methyl acetate may be incorporated into the glyoxylate cycle to be utilized as a carbon and energy source. When Gordonia sp. strain TY-5 was grown on acetone, methanol accumulated temporarily in the medium and finally disappeared (data not shown). The fate of methanol remains to be solved.

FIG. 6.

Propane metabolism in Gordonia sp. strain TY-5. The genes involved in each step of the reaction are indicated in the boxes.

The evidence presented in this report demonstrates that a novel BVMO, designated acetone monooxygenase, is involved in O₂-dependent acetone metabolism. Microbial metabolism of linear ketones has received little attention compared with that of cyclic ketones. This work would provide additional insights into bacterial strategies for metabolism of linear ketones.