The pimFABCDE operon from ... - Semantic Scholar
Microbiology (2005), 151, 727–736
DOI 10.1099/mic.0.27731-0
The pimFABCDE operon from Rhodopseudomonas palustris mediates dicarboxylic acid degradation and participates in anaerobic benzoate degradation Faith H. Harrison and Caroline S. Harwood3 Department of Microbiology, The University of Iowa, 3-450 BSB, Iowa City, IA 52242, USA
Correspondence Caroline S. Harwood [email protected]
Received 27 October 2004 Revised 9 December 2004 Accepted 10 December 2004
Bacteria in anoxic environments typically convert aromatic compounds derived from pollutants or green plants to benzoyl-CoA, and then to the C7 dicarboxylic acid derivative 3-hydroxypimelyl-CoA. Inspection of the recently completed genome sequence of the purple nonsulfur phototroph Rhodopseudomonas palustris revealed one predicted cluster of genes for the b-oxidation of dicarboxylic acids. These genes, annotated as pimFABCDE, are predicted to encode acyl-CoA ligase, enoyl-CoA hydratase, acyl-CoA dehydrogenase and acyl-CoA transferase enzymes, which should allow the conversion of odd-chain dicarboxylic acids to glutaryl-CoA, and even-chain dicarboxylic acids to succinyl-CoA. A mutant strain that was deleted in the pim gene cluster grew at about half the rate of the wild-type parent when benzoate or pimelate was supplied as the sole carbon source. The mutant grew five times more slowly than the wild-type on the C14 dicarboxylic acid tetradecanedioate. The mutant was unimpaired in growth on the C8-fatty acid caprylate. The acyl-CoA ligase predicted to be encoded by the pimA gene was purified, and found to be active with C7–C14 dicarboxylic and fatty acids. The expression of a pimA–lacZ chromosomal gene fusion increased twofold when cells were grown in the presence of straight-chain C7–C14 dicarboxylic and fatty acids. These results suggest that the b-oxidation enzymes encoded by the pim gene cluster are active with medium-chain-length dicarboxylic acids, including pimelate. However, the finding that the pim operon deletion mutant is still able to grow on dicarboxylic acids, albeit at a slower rate, indicates that R. palustris has additional genes that can also specify the degradation of these compounds.
INTRODUCTION Medium-chain-length dicarboxylic acids, including the C7 compound pimelate, are found in the natural environment derived from plant matter such as suberin (Bernards & Razen, 2001) and as derivatives of seed oils. Dicarboxylic acids are also intermediates in the metabolism of cyclic alcohols (Donoghue & Trudgill, 1975; Hasegawa et al., 1982) and alkanes (Kester & Foster, 1963). Depending on the bacterial species, either pimelyl-CoA or 3-hydroxypimelylCoA is formed as the immediate ring-cleavage intermediate in the anaerobic degradation of benzoate (Harwood et al., 1999). These ring-cleavage intermediates are then further degraded to three molecules of acetyl-CoA and one molecule of CO2 (Harwood et al., 1999) (Fig. 1a). Since aromatic hydrocarbons and various plant-derived phenylalkanoates are typically converted to benzoyl-CoA in anoxic environments, 3-hydroxypimelyl-CoA degradation is of central 3Present address: Department of Microbiology, University of Washington, Box 357242, Seattle, WA 98195-7242, USA. Abbreviation: Km, kanamycin.
0002-7731 G 2005 SGM
Printed in Great Britain
importance in anaerobic bioremediation and biomass turnover (Boll et al., 2002; Gibson & Harwood, 2002). A pathway for pimelate degradation has been proposed in which pimelyl-CoA is formed, and then degraded by b-oxidation to glutaryl-CoA and acetyl-CoA (Gallus & Schink, 1994) (Fig. 1b). 3-Hydroxypimelyl-CoA is an intermediate in this pathway. Enzymic activities catalysing the conversion of glutaryl-CoA to two acetyl-CoA molecules and one CO2 molecule have been demonstrated in several bacteria (Elshahed et al., 2001; Gallus & Schink, 1994; Ha¨rtel et al., 1993; Schocke & Schink, 1999), but the proposed b-oxidation reactions that act on pimelyl-CoA or other medium-chain-length dicarboxylic acids have not been directly demonstrated. Recently, b-oxidation genes required for aerobic growth on the C6-dicarboxylic acid adipate, and pimelate and longer-chain-length dicarboxylic acids, were described for the first time, and characterized from Acinetobacter sp. strain ADP1 (Parke et al., 2001). Rhodopseudomonas palustris is a purple nonsulfur phototroph that has served as a model organism for studies of 727
F. H. Harrison and C. S. Harwood
Fig. 1. (a) Pathway for anaerobic degradation of aromatic compounds by R. palustris. Structurally diverse aromatic acids are converted to benzoyl-CoA, which undergoes a series of reactions to generate the ring-cleavage substrate 2ketocyclohexanecarboxyl-CoA. Ring cleavage results in the formation of pimelyl-CoA, which is further degraded to acetylCoA and CO2. A slightly different pathway for benzoyl-CoA degradation, which yields 3-hydroxypimelyl-CoA as the immediate product of ring cleavage, is used by Thauera aromatica and other denitrifying bacteria (Harwood et al., 1999). An abbreviated schematic view of the pathway is shown. (b) Proposed pathway for the b-oxidation of dicarboxylic acids. The specific example of pimelate degradation is shown. With longer-chain-length dicarboxylic acids, the b-oxidation pathway would act in a cyclic fashion, and the reactions after PimA would be repeated following the cleavage of acetyl-CoA from the chain. The Pim enzymes proposed to be involved in the pathway are indicated. Enzymes encoded by genes located outside the pim gene cluster catalyse glutaryl-CoA metabolism.
anaerobic benzoate degradation. It grows on pimelate and other dicarboxylic acids, under both anaerobic and aerobic conditions. Inspection of the recently completed genome sequence of R. palustris strain CGA009 revealed that it has a single cluster of genes predicted to encode all the enzymes that would be required for the b-oxidation of fatty acids or dicarboxylic acids (Larimer et al., 2004; GenBank accession no. BX571963). These genes, annotated as pimFABCDE in the R. palustris genome, are only 30–40 % identical at the amino acid level to the genes reported by Parke et al. (2001) to be required for growth of Acinetobacter spp. on pimelate. Also, some of the predicted enzymic activities 728
differ. Here we report evidence that the R. palustris pim genes are organized as an operon, and are required for maximal anaerobic growth on benzoate and dicarboxylic acids. We also purified the acyl-CoA ligase encoded by pimA, and determined that it was active with a wide range of dicarboxylic acid and fatty acid substrates.
METHODS Carbon source nomenclature. The following nomenclature was
used for the saturated, straight-chain dicarboxylic acids: adipate (C6), pimelate (C7), suberate (C8), azelate (C9), sebacate (C10), Microbiology 151
Dicarboxylic acid b-oxidation
Table 1. Primers used in RT-PCR experiments to determine the transcriptional organization of the pim genes Junction Regulator–pimF Regulator–pimF pimF–pimA pimF–pimA pimA–pimB pimA–pimB pimB–pimC pimB–pimC pimC–pimD pimC–pimD pimD–pimE pimD–pimE
Primer name
Primer sequence (5§–3§)
iclRjxnfor iclRjxnrev Hyd.lig.F Hyd.lig.R Lig.tran.F Lig.tran.R Tran.lrg.F Tran.lrg.R Lrg.sml.F Lrg.sml.R Sml.OH.F Sml.OH.R
GCG TCG ATT GTC AAC GAC GTC AGC GCA CAC AAG TAA GC GCG GCG GCC CGA TGC ACT ACG CC GCG AGA AAC CTC TCG GCC ATC GG AGG ACA TGA TCA TCT CCG GCG GC CCT CGA GCG CCG GAT CGA CGG CG GGC CAC CCC TAC GGC ATG TCG GG GCG AGG TCG GAG CCC GAG CCC G TTC TGA TGG AAG TGA TCG GCC CG GGT GGA GAC GTC GCC GAG ATC CC ACG CGC CGG CGC GAT CGC GGC C TGC CGC GTT CCG CCA TCT GCG GG
undecanedioate (C11), dodecanedioate (C12), tridecanedioate (C13) and tetradecanedioate (C14). The saturated, straight-chain fatty acids used were: valerate (C5), caproate (C6), heptanoate (C7), caprylate (C8), caprate (C10), laurate (C12), myristate (C14), palmitate (C16) and stearate (C18). Bacterial strains and culture conditions. R. palustris strains
were grown at 30 uC in defined basal medium (PM) aerobically with shaking, or anaerobically in sealed tubes in light, as previously described (Kim & Harwood, 1991). Fatty acid and dicarboxylic acid carbon sources were added to liquid PM medium at a final concentration of either 1?5 or 3 mM, as indicated in the table and figure legends. Pimelate and benzoate were added to a final concentration of 3 mM, and succinate to 10 mM. Anaerobic fatty acid or dicarboxylic acid cultures were supplemented with sodium bicarbonate (10 mM), which was added to PM medium as a sterile solution after autoclaving. Escherichia coli strains DH5a and S17-1 were grown at 37 uC in Luria–Bertani broth. Where indicated, R. palustris was grown with 100 mg gentamicin ml21 or 100 mg kanamycin (Km) ml21. E. coli was grown with 100 mg ampicillin ml21, 20 mg Km ml21 or 20 mg gentamicin ml21. Growth rates were determined by measuring turbidity OD660 in a Genesys20 spectrophotometer. DNA manipulations. Standard protocols were used for cloning and
transformations. All restriction endonucleases and DNA modification enzymes were purchased from New England Biolabs. Shrimp alkaline phosphatase was purchased from Roche Diagnostics. PCR was performed with Pfu DNA polymerase (Stratagene). Chromosomal DNA was purified using a Puregene DNA isolation kit (Gentra systems). DNA fragments were excised and purified from agarose gels using the Qiaquick gel extraction kit (Qiagen), and plasmid DNA was purified with the Qiaprep spin miniprep kit. DNA was sequenced at the University of Iowa DNA core facility by standard automated-sequencing technology. RNA isolation. RNA was isolated from R. palustris strain CGA009
using a modification of the RNeasy kit (Qiagen). Cells (30 ml cultures) grown aerobically with 10 mM succinate and 1?5 mM pimelate were harvested at 5000 g at an OD660 of 0?3. The cell pellets were resuspended in 1 ml TE buffer (10 mM Tris/HCl, 1 mM EDTA, pH 8?0), and incubated with Ready-Lyse Lysozyme (Epicentre) for 15 min at room temperature. Buffer RLT (4 ml; provided with the Qiagen kit) was added to the suspension, and 1 ml aliquots were placed in 2 ml screw-capped tubes containing approximately 1 ml 0?1 mm zirconia/silica beads (BioSpec Products). The suspension was then treated in a Mini-BeadBeater-8 (BioSpec Products) for 1 min periods, with cooling on ice after each period, for a http://mic.sgmjournals.org
total of 5 min. The broken cells and beads were sedimented by centrifugation, and the supernatants were pooled. The beads were washed by mixing with buffer [100 ml TE (pH 8?0) plus 350 ml RLT] in the 2 ml tubes. After centrifuging a second time, the supernatants were again combined. Ethanol (2?8 ml) was added to the combined cell lysate, and the entire sample was applied to a RNeasy Midicolumn. A DNase I digestion was performed on the Midi-column with the Qiagen RNase-Free DNase set. The RNA was eluted from the column, and a second DNase I treatment was performed with RQ1 RNase-Free DNase. The sample was then applied to RNeasy mini-columns to purify the RNA. Manufacturer’s instructions were followed for an on-column DNase treatment with the Qiagen DNase set. Three DNase treatments were required to obtain RNA suitable for use in RT-PCR experiments. RT-PCR. RT-PCR was performed by using a two-tube protocol. To synthesize cDNA, 0?25 mg template RNA was heated to 65 uC for 5 min to relieve possible secondary structure. The RNA was then transcribed into cDNA using the Omniscript RT kit (Qiagen) and an appropriate primer specific for the pim intergenic regions. The Omniscript was heat-inactivated, and PCR was performed using 2 ml of the reverse-transcription reaction as the template, and primers for the intergenic regions of the pim cluster (Table 1). Negative control reactions were performed, in which the Omniscript reverse transcriptase was omitted, to ensure that the products from the PCR originated from the cDNA, and not from contaminating genomic DNA. Construction of a pim cluster deletion mutant. A mutant
strain containing a deletion of the entire cluster of pim genes was constructed by overlap extension PCR (Horton et al., 1993). Briefly, 1 kb DNA was amplified from both the upstream region of pimF and the downstream region of pimE (Fig. 2). The amplified fragments shared 66 bp of complementary DNA, and were stitched together by PCR to form a 2 kb product, termed the deletion product. This deletion product was engineered to contain a SacI and a XbaI site at the 59 and 39 ends, respectively, while a BamHI site was engineered in the middle of the deletion product. The SacI/XbaIdigested deletion product was ligated with similarly digested pUC19. Plasmid containing the deletion product was then digested with BamHI, and treated with shrimp alkaline phosphatase. The 1?3 kb BamHI fragment from pUC4K (Amersham Biosciences), containing a Km-resistance GenBlock, was ligated into the deletion product at its BamHI site. The 3?3 kb deletion fragment containing the Kmresistance gene was then excised from the pUC19 backbone with SacI/XbaI, and cloned into pJQ200 mp18 (Quandt & Hynes, 1993) 729
F. H. Harrison and C. S. Harwood
Fig. 2. Organization of the pim genes in the R. palustris chromosome. Arrows indicate the direction of transcription. The predicted functions of the genes are given, as are the R. palustris gene designations. The small black arrows indicate the locations of primers used for RT-PCR. The expected PCR fragments are designated by short continuous lines, and those fragments that were detected by RT-PCR are labelled +. A product was obtained for the regions between the pim genes, indicating they are co-transcribed and constitute an operon.
to generate pFH164. pFH164 was then introduced into wild-type R. palustris strain CGA009 from E. coli S17-1(pFH164) by conjugation. Sucrose-resistant, Km-resistant colonies were selected as described previously (Egland et al., 1995). The resulting mutant strain was verified by Southern analysis as having a deletion of the pimFABCDE genes (Fig. 3), and was named CGA151.
kb
A
B
C
D
7.0 6.0 5.0
Construction of a pimA : : lacZ mutant. A 2?5 kb DNA fragment,
containing pimA plus 1 kb upstream and downstream flanking DNA, was generated by PCR, purified, and cloned into pBBR1MCS5 (Kovach et al., 1994) to generate pFH101. A 5?25 kb fragment containing a lacZ–Km cassette was excised from BamHI-digested pHRP314 (Parales & Harwood, 1993), treated with Klenow fragment to create blunt ends, and ligated into ScaI-digested and shrimpalkaline-phosphatase-treated pFH101 to generate pFH103. The correct orientation of the cassette within the pimA gene was verified by PCR. A 7?5 kb fragment containing pimA : : lacZ was excised from pFH103 by digestion with BamHI and PstI, and cloned into pJQ200 mp18 (Quandt & Hynes, 1993) to generate pFH304. pFH304 was then introduced into CGA009 as described above. Colonies were screened by PCR for the loss of the wild-type pimA gene, and the presence of the pimA–lacZ junction. The resulting mutant strain was named CGA140. Expression cloning of pimA. Primers StartTrc (59-GCGAAGTGGGATCCATGTCCCATCCCGG-39) and TrcStop (59-GGTTGGAATTCCGAATTACTTGGTCTGTG-39) containing restriction sites (underlined) were used to amplify the pimA gene. The 1?7 kb product was cloned into pTrcHis (Invitrogen) after digestion with BamHI and EcoRI, to generate pHisLigase. The plasmid insert was sequenced to ensure that mutations were not introduced, and that the full-length gene was present in the His-tag fusion vector. Purification of PimA. E. coli DH5a(pHisLigase) was grown at 30 uC with aeration to an OD660 of 0?3. The 750 ml culture was induced with IPTG (1 mM) for 4 h. Cells were harvested, washed,
730
4.0
3.0
2.0 1.6
Fig. 3. Southern blot analysis of the pim deletion mutant CGA151. Genomic DNA was digested, and separated on a 0?8 % agarose gel. The DNA probe was designed to hybridize to the 1 kb region immediately upstream of the pim cluster. In the wild-type strain, a 1?9 or 2?1 kb fragment was expected after digestion with NotI or EcoRI, respectively. To make the deletion mutant, 8?8 kb was removed from the chromosome. Due to the loss of NotI and EcoRI restriction sites within the deleted region, and the addition of the Km cassette, the expected fragment sizes in the mutant strain were 6?4 kb for NotI, and 6?7 kb for EcoRI. Lanes: A, WT digested with NotI; B, WT digested with EcoRI; C, CGA151 digested with NotI; D, CGA151 digested with EcoRI. The positions of size markers are shown schematically on the left. Microbiology 151
Dicarboxylic acid b-oxidation and resuspended in 20 ml binding buffer [20 mM triethanolamine (TEA/HCl) buffer (pH 8?0), 0?5 M NaCl, 5 % (v/v) glycerol]. Cells were lysed by sonication. PMSF (0?05 M) was added to the cell paste, and the mixture was centrifuged at 10 000 g for 20 min at 4 uC. The resulting supernatant was then centrifuged at 40 000 g for 90 min at 4 uC, after the addition of DNase I (1 mg ml21), RNase I (1 mg ml21) and PMSF (0?05 M). The supernatant from this highspeed centrifugation was termed the crude cell extract. All subsequent steps were carried out at 4 uC. The crude cell extract (20 ml; 62 mg) was loaded onto a 5 ml HiTrap chelating column (Pharmacia Biotech) that had been charged with 0?1 M NiSO4, and equilibrated with binding buffer. After extensive washing with binding buffer, the column was developed over 120 min using a Bio-Rad Biologic System with a linear gradient of 0–0?5 M imidazole in binding buffer with a flow rate of 1 ml min21. Fractions (1 ml) were collected, and active fractions determined with the isotopic assay were pooled. Enzyme assays. Acyl-CoA ligase (PimA) activity was measured with isotopic and spectrophotometric assays as previously described (Geissler et al., 1988). The two assays gave comparable results with pimelate, but the isotopic assay was more sensitive, and had less variability, than the spectrophotometric assay. For the isotopic assay, the enzymic conversion of [14C]pimelate to [14C]pimelyl-CoA was measured. Pimelate that was not converted to pimelyl-CoA was extracted from the reaction mixture with ethyl acetate at acidic pH, while pimelyl-CoA remained hydrophilic. The amount of unreacted pimelate in the reaction mixture was detected in controls where enzyme was omitted from the reaction, and it was used in calculations of enzymic activity. The reaction mixture contained 2?5 mM MgCl2, 0?5 mM ATP, 0?25 mM reduced CoA (CoASH) and [14C]pimelic acid (200 mM in standard assay for 555 Bq per reaction) in 20 mM TEA/HCl, pH 8?0. The reaction was initiated by the addition of 20 mg enzyme for a final volume of 0?5 ml. For determination of kinetic constants, the above reaction mixture was used with 0?5 mM CoASH. The amount of labelled substrate remained constant at 20 mM for these assays, while the addition of unlabelled substrate was used to achieve the desired final substrate concentration. This assay was suitable for use at all stages of enzyme purification. [1,7-14C]Pimelate (2?22–3?33 GBq mmol21) was obtained from American Radiolabelled Chemicals.
The spectrophotometric assay measures the formation of AMP by acyl-CoA ligase by coupling the reaction via a series of auxiliary enzymes to NADH oxidation, as previously described (Geissler et al., 1988). The reaction mixture contained 20 mM TEA/HCl (pH 8?0), 2?5 mM MgCl2, 0?5 mM ATP, 0?25 mM CoASH, 10 mM KCl, 10 mM phosphoenol pyruvate, 0?175 mM NADH, substrate (0?25 mM in standard assays), 2 U of the auxiliary enzymes pyruvate kinase and lactic acid dehydrogenase, and 4 U of myokinase (adenylate kinase), in a total volume of 1 ml. The assays were performed using a Beckman DU800 spectrophotometer with a quartz cuvette of 1 cm path length. For determination of kinetic constants, the above mixture was used with 0?5 mM CoASH. The spectrophotometric assay can only be used with pure enzymes due to the very high background activity of NADH oxidation in crude cell extracts. Kinetic constant determination. Enzyme activities were measured
at a range of substrate concentrations. At least nine concentrations were tested in duplicate on two separate days, and used to generate a Hanes–Woolf plot. Maximal enzymic velocity (Vmax) and the Michaelis constant (Km) were determined from the plot. These values were then used to calculate Kcat (enzyme turnover rate), which was determined from Vmax/[E], where [E] is the total enzyme concentration, and Kcat/Km (catalytic efficiency). Southern blotting. Approximately 3 mg chromosomal DNA was
digested with NotI or EcoRI, and separated on a 0?8 % agarose gel. http://mic.sgmjournals.org
A probe was generated by PCR that was complementary to 1 kb DNA directly upstream of pimF. The probe was labelled with the Ready-To-Go DNA labelling beads (Amersham Biosciences), according to the manufacturer’s procedures. Southern hybridizations were performed by standard procedures. Other procedures. SDS-PAGE was carried out with 12?5 % acrylamide gels by standard procedures (Ausubel et al., 1990). Separated proteins were visualized by Coomassie blue R-250. Molecular-mass standards were from Gibco-BRL. Protein concentrations of cell extracts and enzyme preparations were determined with the Bio-Rad protein assay reagent. b-Galactosidase activity was measured as previously described (Egland & Harwood, 1999). Sedimentation equilibrium centrifugation. The native molecular
mass of the PimA protein was determined using a Beckman XL-1 analytical ultracentrifuge in absorbance mode. All experiments were performed at 12 uC in buffer containing 20 mM TEA/HCl (pH 8?0), 0?5 M NaCl and 10 % (v/v) glycerol. A six-channel centrepiece was used for acquiring absorbance data at 280 nm for a protein concentration range of 2–8 mM. Data were collected at rotor speeds of 10 000, 12 000 and 14 000 r.p.m. Equilibrium was achieved when three consecutive scans taken 2 h apart were unchanged. Data editing was performed with the WinREEDIT program (version 0.999.0028), and fitting of the data was performed with the WinNONLIN program (version 1.06.0048); both programs were developed at the the National Analytical Ultracentrifugation Facility, University of Connecticut Biotechnology Center; http://www.ucc. uconn.edu/~wwwbiotc/AUFMAIN.HTML).
RESULTS Examination of the R. palustris genome for b-oxidation genes The genome of R. palustris has over 160 genes predicted to encode enzymes that might be involved in the b-oxidation of fatty acids and dicarboxylic acids (Larimer et al., 2004; GenBank accession no. BX571963). However, these genes tend to be scattered around the genome, and we identified only a single cluster of genes predicted to encode all the enzymes necessary for dicarboxylic acid or fatty acid boxidation (Fig. 2). This cluster, annotated as the pim genes, includes an acyl-CoA ligase gene (pimA), an acyl-CoA acetyltransferase gene (pimB), a short-chain alcohol dehydrogenase gene (pimE), an enoyl-CoA hydratase (pimF) gene, and two genes, pimC and pimD, predicted to encode flavin-containing dehydrogenases. The genome sequence indicates that the pimE gene has a frameshift mutation. It is annotated as a pseudogene, and is expected to be nonfunctional in R. palustris. The pim genes are organized as an operon, and are required for optimal growth on benzoate and dicarboxylic acids To examine the transcriptional organization of the genes in the pim cluster, we carried out RT-PCR using primers designed to amplify the intergenic regions (Table 1). Products were obtained for all of the intergenic regions in the cluster, but not between the upstream regulatory genes rpa3718 and pimF (Fig. 2). No product was obtained in controls to which no reverse transcriptase was added. These 731
F. H. Harrison and C. S. Harwood
results indicate that the pimFABCDE genes are cotranscribed, and constitute an operon. To examine the contribution of the pim operon to growth with dicarboxylic acids and fatty acids, we constructed a mutant (CGA151) in which the pimFABCDE genes were deleted from the chromosome. PCR and Southern hybridization experiments (Fig. 3) each verified that the expected deletion had occurred in CGA151. The pimFABCDE deletion mutant grew more slowly than the wild-type parent when pimelate (C7), azelate (C9) or tetradecanedioate (C14) was supplied as the sole carbon source under anaerobic photoheterotrophic growth conditions (Table 2). The pim operon deletion strain was also impaired in growth on benzoate, a compound that R. palustris degrades only under anaerobic conditions. The pim operon deletion strain was impaired in aerobic growth on pimelate and azelate. We were unable to obtain consistent growth of wild-type R. palustris with other dicarboxylic and fatty acids under aerobic conditions (Table 2). The pim operon deletion mutant grew anaerobically at wild-type rates on the fatty acid caprylate (C8). This could mean that the pim operon is not involved in the degradation of this fatty acid. Another interpretation is that R. palustris encodes at least two sets of enzymes that can catalyse caprylate degradation, one of which is encoded by the pim operon. The pimA gene is induced by growth with benzoate, fatty acids or dicarboxylic acids
cultures of strain CGA140 were compared with the activities of cultures grown on succinate plus various dicarboxylic acids and fatty acids (Fig. 4). During aerobic growth, dicarboxylic acids from 7 to 14 carbons in length, and the C8 fatty acid caprylic acid, induced pimA : : lacZ expression about twofold over succinate-grown cells. Anaerobically, pimelate or benzoate induced pimA : : lacZ expression about threefold above the levels seen when CGA140 was grown with succinate alone, and b-galactosidase activities in cultures grown under anaerobic conditions with caprylate were twofold higher than in succinate-grown cultures. Strain CGA140 resembled the pim operon deletion strain in that it grew more slowly than its wild-type parent with pimelate and benzoate. When grown with pimelate as a sole carbon source anaerobically in light, the levels of b-galactosidase activity expressed by the pimA : : lacZ mutant strain were similar to those observed when the strain was grown on pimelate and succinate, indicating that succinate does not repress expression of the pimA gene. Characterization of PimA, a broad-substraterange acyl-CoA ligase Because it is difficult to prepare the appropriate CoA thioester intermediates, we were not able to judge the efficiencies of most of the Pim enzymes with a range of possible substrates. We therefore focused our efforts on PimA, the enzyme that is predicted to activate unmodified
We used a strain carrying a pimA : : lacZ chromosomal fusion (strain CGA140) to examine the influence of various carbon compounds on the expression of the pim gene cluster. The b-galactosidase activities of succinate-grown
Table 2. Growth rates of the pim deletion mutant strain CGA151 Carbon source*
Aerobic Succinate (C4-DA) Azelate (C9-DA) Pimelate (C7-DA) Anaerobic Succinate (C4-DA) Caprylate (C8-FA) Azelate (C9-DA) Benzoate (C7-AA) Pimelate (C7-DA) Tetradecanedioate (C14-DA)
Doubling time (h)D WT (CGA009)
CGA151
12?1±2?5 23?6±2?0 25?0±3?4
12?0±1?3 30?5±2?9 55?5±10?3
8?5±0?3 8?2±1?5 7?7±2?2 8?4±0?6 10?3±2?2 11?0±2?5
8?7±0?4 10?2±0?7 22?1±4?4 15?5±3?5 25?8±4?5 63?7±17?2
*Carbon sources were used at a final concentration of 3 mM, except succinate, which was supplied at 10 mM. See the legend of Fig. 4 for key to abbreviations. DValues are the means±SD of three independent cultures assayed in duplicate. 732
Fig. 4. b-Galactosidase activities expressed by the pimA : : lacZ fusion strain CGA140. Activities are expressed as nmol product formed min”1 (mg protein)”1. Cultures were grown on 10 mM succinate plus 1?5 mM test substrate as indicated (number of carbons followed by DA for dicarboxylic acid, FA for fatty acid, AA for aromatic acid): C4-DA, succinate; C7-DA, pimelate; C9DA, azelate; C10-DA, sebacate; C11-DA, undecanedioate; C14DA, tetradecanedioate; C8-FA, caprylate; C7-AA, benzoate. The activities are the means (±SD) of measurements of b-galactosidase activities in cell extracts prepared from at least six independent cultures. Each culture was assayed in duplicate. Microbiology 151
Dicarboxylic acid b-oxidation
Table 3. Purification of the expressed PimA protein from E. coli cells Purification step Cell extract Ni2+ chelating
Protein (mg)
Total activity (U)*
Specific activity (U mg”1)
Activity recovered (%)
Purification (fold)
66 0?57
1?795 0?143
0?0272 0?25
100 22
9
*Units are measured in mmol pimelyl-CoA formed min21.
dicarboxylic acids with CoA to initiate their b-oxidation. An N-terminal His-tagged version of the predicted acylCoA ligase PimA was expressed in E. coli, and purified in a single chromatographic step (Table 3, Fig. 5). Pimelate was the substrate used to follow activity during purification because it is available in radiolabelled form, and it is necessary to use an isotopic assay to measure activity in crude cell extracts. The molecular mass of the purified PimA protein, as determined by SDS-PAGE, was approximately 60 000 Da. The pimA gene sequence predicts a protein of 552 amino acids, with a molecular mass of 60 029 Da. We used analytical ultracentrifugation to determine the oligomeric state of the enzyme. The native molecular mass as determined by sedimentation equilibrium analysis was 59 100 Da. This indicates that native PimA is a monomer. The purified enzyme catalysed the addition of reduced CoA to pimelate at pH 8?0. The activity was unaffected by the presence of oxygen. The enzyme activity was optimal at pH 8?0–8?5.
good activity. These included azelate, sebacate, undecanedioate, tetradecanedioate and caprylate. Kinetic constants for pimelate were determined using the isotopic assay because of its greater sensitivity (Table 4). Caprylic acid, the C8 fatty acid, was the substrate best utilized by the acylCoA ligase. PimA had the highest affinity and the highest catalytic efficiency with caprylate. PimA had similar catalytic efficiencies and turnover rates with azelate, sebacate, tetradecanedioate and undecanedioate (chain lengths defined in Methods). Although PimA was active with pimelate, this dicarboxylic acid was a very poor substrate when compared with other dicarboxylic acids. Cell extracts of R. palustris strain CGA009 had a pimelate-CoA ligase activity of 10 nmol pimelyl-CoA formed min21 (mg protein)21. PimA had no detectable activity with the
PimA has a broad substrate range, and is active with C7–C14 dicarboxylic acids, and C6–C16 fatty acids (Fig. 6). Kinetic constants were determined with a coupled enzymic assay (see Methods) for a subset of substrates that gave
1
2
3
114 81 60
64 50
37
Fig. 5. SDS-PAGE analysis of active protein fractions obtained during the purification of the PimA acyl-CoA ligase. Lanes: 1, molecular-mass marker; 2, crude extract (15 mg); 3, active pooled fractions from the Ni2+ chelating column (1?5 mg). Numbers to the left of the gel represent molecular mass in kDa. http://mic.sgmjournals.org
Fig. 6. Specific activities of the PimA ligase with various dicarboxylic and fatty acid substrates. Each substrate was present in assay mixtures at a final concentration of 250 mM. Activities were measured using the spectrophotometric assay described in Methods. Pimelate was measured with the isotopic assay. Dicarboxylic acids: C7, pimelate; C8, suberate; C9, azelate; C10, sebacate; C11, undecanedioate; C12, dodecanedioate; C13, tridecanedioate; C14, tetradecanedioate. Fatty acids: C6, caproate; C7, heptanoate; C8, caprylate; C10, caprate; C12, laurate; C14, myristate; C16, palmitate. The enzyme was not active with the following substrates: adipate, valerate, stearate and benzoate. Specific activities are means±SD of at least six independent determinations on three different days. 733
F. H. Harrison and C. S. Harwood
Table 4. Kinetic constants of the purified PimA acyl-CoA ligase Data were obtained using the spectrophotometric assay described in Methods. Substrates are followed by the number of carbons in the compound, and designation as either a dicarboxylic (DA) or a fatty acid (FA). Rates were determined by plotting the data for nine concentrations in duplicate. Although similar values for pimelate were obtained using the isotopic or spectrophotometric assay, results using the isotopic assay are reported because of its greater sensitivity. Substrate
Kcat (s”1)
Km (mM)
Kcat/Km (M”1 s”1)
Pimelate (C7-DA) Caprylate (C8-FA) Azelate (C9-DA) Sebacate (C10-DA) Undecanedioate (C11-DA) Tetradecanedioate (C14-DA)
0?03 3?7 3?6 3?3 3?0 2?6
277 4 230 177 47 5?16104
C6-dicarboxylic acid adipate, and it was not active with the fatty acids valerate or stearate, or with benzoate.
DISCUSSION Mammalian mitochondria and many bacteria, including E. coli, degrade medium- and long-chain-length fatty acids by well-characterized b-oxidation pathways. Although dicarboxylic acid degradation has received much less attention, many bacterial species are known to grow with dicarboxylic acids having a range of carbon chain lengths (Gallus & Schink, 1994; Hoet & Stanier, 1970; Parke et al., 2001). Also, some aromatic-compound-degrading bacteria form pimelyl-CoA, and others form 3-hydroxypimelyl-CoA, as the immediate product of ring cleavage during anaerobic benzoate degradation (Harwood et al., 1999). By analogy with aerobic aromatic compound degradation pathways, the metabolism of the benzoate degradation intermediates pimelyl-CoA or 3-hydroxypimelyl-CoA to three molecules of acetyl-CoA plus one molecule of CO2 can be considered to be the ‘lower’ anaerobic benzoate degradation pathway. Data presented here indicate that in R. palustris the pimFABCDE genes are cotranscribed and organized in an operon that encodes enzymes for the b-oxidation of dicarboxylic acids of between 7 and 14 carbons, including pimelate (Fig. 1b). The slow growth phenotype of a pim operon deletion mutant (Table 2) is consistent with the hypothesis that the Pim enzymes also catalyse the lower part of the benzoate degradation pathway. Benzoate and dicarboxylates with an odd number of carbons are predicted to be converted to glutaryl-CoA. Dicarboxylates with an even number of carbons would be converted to succinyl-CoA, and thus enter directly into the citric acid cycle. We cannot formally exclude that the pim operon somehow functions biosynthetically rather than catabolically. However, this seems unlikely because the growth 734
defect of the pim deletion strain is not a general growth defect, but instead is specific to only a subset of the growth substrates that R. palustris degrades. The R. palustris genome has genes predicted to mediate conversion of glutaryl-CoA to acetyl-CoA and CO2 (Larimer et al., 2004) (Fig. 1b). It has a predicted glutaryl-CoA dehydrogenase/glutaconyl-CoA decarboxylase gene (rpa1094), and also encodes predicted 3-hydroxybutyryl-CoA dehydratase (Rpa4339), 3-hydroxybutyryl-CoA dehydrogenase (Rpa4748) and acetoacetyl-CoA thiolase (Rpa0531) enzymes for the conversion of crotonyl-CoA, the product of glutaconylCoA decarboxylation, to two acetyl-CoA (Fig. 1b). The Pim enzymes are obviously not the only R. palustris enzymes that can catalyse dicarboxylic acid degradation because the pim operon deletion strain is still able to grow, albeit more slowly than the wild-type, on these compounds. The R. palustris genome encodes 45 CoA ligases, 32 flavincontaining dehydrogenases, 26 enoyl-CoA hydratases, many short-chain alcohol dehydrogenases and more than 10 acylCoA acetyltransferases. It is likely that combinations of these other b-oxidation enzymes can also contribute to the degradation of dicarboxylic acids. In E. coli, the genes encoding enzymes of aerobic fatty acid b-oxidation are scattered around the chromosome (Clark & Cronan, 1996). The R. palustris pimE ORF is disrupted by a 21 frameshift at nucleotide 4192036. This mutation is expected to render the gene nonfunctional. Thus an alternate R. palustris protein must catalyse the dehydrogenation reaction that is part of the b-oxidation sequence of reactions. It is possible that PimF has 3-hydroxyacyl-CoA dehydrogenase as well as 3-hydroxyacyl-CoA dehydratase activity. PimF possesses the characteristic amino acid motifs to perform both of these functions. A single gene product, FadB, catalyses these two reactions during fatty acid b-oxidation in E. coli (Yang et al., 1988). Another possibility is that one of the other 54 short-chain alcohol dehydrogenases encoded elsewhere on the R. palustris chromosome has 3-hydroxyacylCoA dehydrogenase activity. The PimA protein appears to have one of the broadest substrate ranges of any acyl-CoA ligase described, as it is active with both dicarboxylic and fatty acids of medium and long carbon chain length. Acyl-CoA ligases involved in the b-oxidation of fatty acids, including FadD from E. coli (Kameda & Nunn, 1981), have been described that are active with a broad range of fatty acids; however, these enzymes were not reported to have been tested for activity with dicarboxylic acids. Pimelyl-CoA is the first committed intermediate in the synthesis of the vitamin biotin, and several pimelate-CoA ligases have been purified and characterized in the course of efforts to develop a biotechnological process for the production of biotin (Binieda et al., 1999; Ploux et al., 1992). PimA shares approximately 30 % amino acid identity with these proteins, but the region of identity is mainly confined to the shared AMPbinding domain common to CoA ligases. The described Microbiology 151
Dicarboxylic acid b-oxidation
pimelate-CoA ligases have a narrow substrate range, and are specific to pimelate. A search of the current databases revealed that the only organism with a set of genes that is very similar to the pim genes is the nitrogen-fixing soybean symbiont Bradyrhizobium japonicum (Kaneko et al., 2002). This set of genes lies at nucleotide positions 8567718–8573337 in the B. japonicum genome. Applying pim designations to these genes, B. japonicum possesses the following: bll7817 (pimD), bll7818 (pimC), bll7819 (pimB), bll7820 (pimA) and bll7821 (pimF). These genes have between 73 and 89 % deduced amino acid identity with the homologous R. palustris genes. Based on this high degree of relatedness, we predict a role for these genes in aerobic dicarboxylic acid degradation by B. japonicum. B. japonicum does not have an anaerobic pathway for benzoate degradation. The pimE gene is absent from the B. japonicum genome pim gene cluster. An IclR-like regulatory gene, and genes for an ABC transport system related to the branched chain amino acid (Ilv) uptake system of E. coli, are divergently transcribed from the pim cluster in R. palustris (Fig. 2). R. palustris encodes 20 Ilv-type transporters, and we have speculated that these may be specific for various sorts of hydrophobic compounds, as well as for dicarboxylic acids (Larimer et al., 2004). We predict that the protein products of genes rpa3719–rpa3725 are responsible for transporting substrates for the Pim enzymes, and that the multiple periplasmic binding proteins encoded by this gene cluster enable the transport of a range of dicarboxylic and fatty acids. In addition, we speculate the IclR regulator could control both transport and degradation of these compounds. The B. japonicum genome has an identical gene cluster at position 8575477, directly upstream of its putative pim genes, consisting of an IclR-like regulator, and a transport system with multiple periplasmic binding proteins. The R. palustris and B. japonicum regulatory and transport genes share between 76 and 88 % amino acid identity, while two of the three binding proteins in B. japonicum share approximately 78–80 % identity with those in R. palustris.
ACKNOWLEDGEMENTS
Binieda, A., Fuhrmann, M., Lehner, B., Rey-Berthod, C., FrutigerHughes, S., Hughes, G. & Shaw, N. M. (1999). Purifica-
tion, characterization, DNA sequence and cloning of a pimeloylCoA synthetase from Pseudomonas mendocina 35. Biochem J 340, 793–801. Boll, M., Fuchs, G. & Heider, J. (2002). Anaerobic oxidation of
aromatic compounds and hydrocarbons. Curr Opin Chem Biol 6, 604–611. Clark, D. P. & Cronan, J. E. (1996). Two-carbon compounds and
fatty acids as carbon sources. In Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, pp. 343–357. Edited by F. C. Neidhardt and others. Washington, DC: American Society for Microbiology. Donoghue, N. A. & Trudgill, P. W. (1975). The metabolism
of cyclohexanol by Acinetobacter NCIB 9871. Eur J Biochem 60, 1–7. Egland, P. G. & Harwood, C. S. (1999). BadR, a new MarR family
member, regulates anaerobic benzoate degradation by Rhodopseudomonas palustris in concert with AadR, an Fnr family member. J Bacteriol 181, 2102–2109. Egland, P. G., Gibson, J. & Harwood, C. S. (1995). Benzoate-
coenzyme A ligase, encoded by badA, is one of three ligases able to catalyze benzoyl-coenzyme A formation during anaerobic growth of Rhodopseudomonas palustris on benzoate. J Bacteriol 177, 6545–6551. Elshahed, M. S., Bhupathiraju, V. K., Wofford, N. Q., Nanny, M. A. & McInerney, M. J. (2001). Metabolism of benzoate, cyclohex-1-ene
carboxylate, and cyclohexane carboxylate by ‘‘Syntrophus aciditrophicus’’ strain SB in syntrophic association with H2-using microorganisms. Appl Environ Microbiol 67, 1728–1738. Gallus, C. & Schink, B. (1994). Anaerobic degradation of pimelate
by newly isolated denitrifying bacteria. Microbiology 140, 409–416. Geissler, J. F., Harwood, C. S. & Gibson, J. (1988). Purification and
properties of benzoate-coenzyme A ligase, a Rhodopseudomonas palustris enzyme involved in the anaerobic degradation of benzoate. J Bacteriol 170, 1709–1714. Gibson, J. & Harwood, C. S. (2002). Metabolic diversity in aromatic
compound utilization by anaerobic microbes. Annu Rev Microbiol 56, 345–369. Ha¨rtel, U., Eckel, E., Koch, J., Fuchs, G., Linder, D. & Buckel, W. (1993). Purification of glutaryl-CoA dehydrogenase from Pseudo-
monas sp., an enzyme involved in the anaerobic degradation of benzoate. Arch Microbiol 159, 712–717. Harwood, C. S., Burchhardt, G., Herrmann, H. & Fuchs, G. (1999).
Anaerobic metabolism of aromatic compounds via the benzoyl-CoA pathway. FEMS Microbiol Rev 22, 439–458. Hasegawa, Y., Hamano, K. & Obata, H. (1982). Microbial degrada-
We thank Gary Heil for his assistance with the analytical ultracentrifugation. The Division of Energy Biosciences, US Department of Energy (grant DE-FG02-95ERZ0184), and the US Army Research Office (grant W911NF-04-1-0123), supported this research. F. H. H. was supported by a National Science Foundation research training grant (DBI9602247).
dicarboxylic acids by Pseudomonas fluorescens. Eur J Biochem 13, 65–70.
REFERENCES
Kameda, K. & Nunn, W. (1981). Purification and characterization of
tion of cycloheptanone. Agric Biol Chem 46, 1139–1143. Hoet, P. P. & Stanier, R. Y. (1970). The dissimilation of higher
Horton, R. M., Ho, S. N., Pullen, J. K., Hunt, H. D., Cai, Z. & Pease, L. R. (1993). Gene splicing by overlap extension. Methods Enzymol
217, 270–279.
Ausubel, F. M., Brent, R., Kingston, R. E., More, D. D., Seidman, J. G., Smith, J. A. & Struhl, K. (1990). Current Protocols in Molecular
Biology. New York: Greene Publishing Associates. Bernards, M. A. & Razen, F. A. (2001). The poly(phenolic) domain
of potato suberin: a non-lignin cell wall bio-polymer. Phytochemistry 57, 1115–1122. http://mic.sgmjournals.org
acyl coenzyme A synthetase from Escherichia coli. J Biol Chem 256, 5702–5707. Kaneko, T., Nakamura, Y., Sato, S. & 14 other authors (2002).
Complete genomic sequence of nitrogen-fixing symbiotic bacterium Bradyrhizobium japonicum USDA110. DNA Res 9, 189–197. Kester, A. S. & Foster, J. W. (1963). Diterminal oxidation of long-
chain alkanes by bacteria. J Bacteriol 85, 859–869. 735
F. H. Harrison and C. S. Harwood Kim, M.-K. & Harwood, C. S. (1991). Regulation of benzoate-CoA
ligase in Rhodopseudomonas palustris. FEMS Microbiol Lett 83, 199–204.
Ploux, O., Soularue, P., Marquet, A., Gloeckler, R. & Lemoine, Y. (1992). Investigation of the first step of biotin biosynthesis in
Kovach, M. E., Phillips, R. W., Elzer, P. H., Roop, R. M., II & Peterson, K. M. (1994). pBBR1MCS: a broad-host-range cloning vector.
Bacillus sphaericus. Purification and characterization of the pimeloylCoA synthase, and uptake of pimelate. Biochem J 287, 685–690.
Biotechniques 16, 800–802.
Quandt, J. & Hynes, M. F. (1993). Versatile suicide vectors which
Larimer, F. W., Chain, P., Hauser, L. & 16 other authors (2004).
allow direct selection for gene replacement in Gram-negative bacteria. Gene 127, 15–21.
Complete genome sequence of the metabolically versatile photosynthetic bacterium Rhodopseudomonas palustris. Nat Biotechnol 22, 55–61. Parales, R. E. & Harwood, C. S. (1993). Construction and use of a
new broad-host-range lacZ transcriptional fusion vector, pHRP309, for Gram-negative bacteria. Gene 133, 23–30. Parke, D., Garcia, M. A. & Ornston, L. N. (2001). Cloning and genetic characterization of dca genes required for b-oxidation of
straight-chain dicarboxylic acids in Acinetobacter sp. strain ADP1. Appl Environ Microbiol 67, 4817–4827.
736
Schocke, L. & Schink, B. (1999). Energetics and biochemistry of
fermentative benzoate degradation by Syntrophus gentianae. Arch Microbiol 171, 331–337. Yang, S. Y., Li, J. M., He, X. Y., Cosloy, S. D. & Schulz, H. (1988).
Evidence that the fadB gene of the fadAB operon of Escherichia coli encodes 3-hydroxyacyl-coenzyme A (CoA) epimerase, delta 3-cisdelta 2-trans-enoyl-CoA isomerase, and enoyl-CoA hydratase in addition to 3-hydroxyacyl-CoA dehydrogenase. J Bacteriol 170, 2543–2548.
Microbiology 151
DOI 10.1099/mic.0.27731-0
The pimFABCDE operon from Rhodopseudomonas palustris mediates dicarboxylic acid degradation and participates in anaerobic benzoate degradation Faith H. Harrison and Caroline S. Harwood3 Department of Microbiology, The University of Iowa, 3-450 BSB, Iowa City, IA 52242, USA
Correspondence Caroline S. Harwood [email protected]
Received 27 October 2004 Revised 9 December 2004 Accepted 10 December 2004
Bacteria in anoxic environments typically convert aromatic compounds derived from pollutants or green plants to benzoyl-CoA, and then to the C7 dicarboxylic acid derivative 3-hydroxypimelyl-CoA. Inspection of the recently completed genome sequence of the purple nonsulfur phototroph Rhodopseudomonas palustris revealed one predicted cluster of genes for the b-oxidation of dicarboxylic acids. These genes, annotated as pimFABCDE, are predicted to encode acyl-CoA ligase, enoyl-CoA hydratase, acyl-CoA dehydrogenase and acyl-CoA transferase enzymes, which should allow the conversion of odd-chain dicarboxylic acids to glutaryl-CoA, and even-chain dicarboxylic acids to succinyl-CoA. A mutant strain that was deleted in the pim gene cluster grew at about half the rate of the wild-type parent when benzoate or pimelate was supplied as the sole carbon source. The mutant grew five times more slowly than the wild-type on the C14 dicarboxylic acid tetradecanedioate. The mutant was unimpaired in growth on the C8-fatty acid caprylate. The acyl-CoA ligase predicted to be encoded by the pimA gene was purified, and found to be active with C7–C14 dicarboxylic and fatty acids. The expression of a pimA–lacZ chromosomal gene fusion increased twofold when cells were grown in the presence of straight-chain C7–C14 dicarboxylic and fatty acids. These results suggest that the b-oxidation enzymes encoded by the pim gene cluster are active with medium-chain-length dicarboxylic acids, including pimelate. However, the finding that the pim operon deletion mutant is still able to grow on dicarboxylic acids, albeit at a slower rate, indicates that R. palustris has additional genes that can also specify the degradation of these compounds.
INTRODUCTION Medium-chain-length dicarboxylic acids, including the C7 compound pimelate, are found in the natural environment derived from plant matter such as suberin (Bernards & Razen, 2001) and as derivatives of seed oils. Dicarboxylic acids are also intermediates in the metabolism of cyclic alcohols (Donoghue & Trudgill, 1975; Hasegawa et al., 1982) and alkanes (Kester & Foster, 1963). Depending on the bacterial species, either pimelyl-CoA or 3-hydroxypimelylCoA is formed as the immediate ring-cleavage intermediate in the anaerobic degradation of benzoate (Harwood et al., 1999). These ring-cleavage intermediates are then further degraded to three molecules of acetyl-CoA and one molecule of CO2 (Harwood et al., 1999) (Fig. 1a). Since aromatic hydrocarbons and various plant-derived phenylalkanoates are typically converted to benzoyl-CoA in anoxic environments, 3-hydroxypimelyl-CoA degradation is of central 3Present address: Department of Microbiology, University of Washington, Box 357242, Seattle, WA 98195-7242, USA. Abbreviation: Km, kanamycin.
0002-7731 G 2005 SGM
Printed in Great Britain
importance in anaerobic bioremediation and biomass turnover (Boll et al., 2002; Gibson & Harwood, 2002). A pathway for pimelate degradation has been proposed in which pimelyl-CoA is formed, and then degraded by b-oxidation to glutaryl-CoA and acetyl-CoA (Gallus & Schink, 1994) (Fig. 1b). 3-Hydroxypimelyl-CoA is an intermediate in this pathway. Enzymic activities catalysing the conversion of glutaryl-CoA to two acetyl-CoA molecules and one CO2 molecule have been demonstrated in several bacteria (Elshahed et al., 2001; Gallus & Schink, 1994; Ha¨rtel et al., 1993; Schocke & Schink, 1999), but the proposed b-oxidation reactions that act on pimelyl-CoA or other medium-chain-length dicarboxylic acids have not been directly demonstrated. Recently, b-oxidation genes required for aerobic growth on the C6-dicarboxylic acid adipate, and pimelate and longer-chain-length dicarboxylic acids, were described for the first time, and characterized from Acinetobacter sp. strain ADP1 (Parke et al., 2001). Rhodopseudomonas palustris is a purple nonsulfur phototroph that has served as a model organism for studies of 727
F. H. Harrison and C. S. Harwood
Fig. 1. (a) Pathway for anaerobic degradation of aromatic compounds by R. palustris. Structurally diverse aromatic acids are converted to benzoyl-CoA, which undergoes a series of reactions to generate the ring-cleavage substrate 2ketocyclohexanecarboxyl-CoA. Ring cleavage results in the formation of pimelyl-CoA, which is further degraded to acetylCoA and CO2. A slightly different pathway for benzoyl-CoA degradation, which yields 3-hydroxypimelyl-CoA as the immediate product of ring cleavage, is used by Thauera aromatica and other denitrifying bacteria (Harwood et al., 1999). An abbreviated schematic view of the pathway is shown. (b) Proposed pathway for the b-oxidation of dicarboxylic acids. The specific example of pimelate degradation is shown. With longer-chain-length dicarboxylic acids, the b-oxidation pathway would act in a cyclic fashion, and the reactions after PimA would be repeated following the cleavage of acetyl-CoA from the chain. The Pim enzymes proposed to be involved in the pathway are indicated. Enzymes encoded by genes located outside the pim gene cluster catalyse glutaryl-CoA metabolism.
anaerobic benzoate degradation. It grows on pimelate and other dicarboxylic acids, under both anaerobic and aerobic conditions. Inspection of the recently completed genome sequence of R. palustris strain CGA009 revealed that it has a single cluster of genes predicted to encode all the enzymes that would be required for the b-oxidation of fatty acids or dicarboxylic acids (Larimer et al., 2004; GenBank accession no. BX571963). These genes, annotated as pimFABCDE in the R. palustris genome, are only 30–40 % identical at the amino acid level to the genes reported by Parke et al. (2001) to be required for growth of Acinetobacter spp. on pimelate. Also, some of the predicted enzymic activities 728
differ. Here we report evidence that the R. palustris pim genes are organized as an operon, and are required for maximal anaerobic growth on benzoate and dicarboxylic acids. We also purified the acyl-CoA ligase encoded by pimA, and determined that it was active with a wide range of dicarboxylic acid and fatty acid substrates.
METHODS Carbon source nomenclature. The following nomenclature was
used for the saturated, straight-chain dicarboxylic acids: adipate (C6), pimelate (C7), suberate (C8), azelate (C9), sebacate (C10), Microbiology 151
Dicarboxylic acid b-oxidation
Table 1. Primers used in RT-PCR experiments to determine the transcriptional organization of the pim genes Junction Regulator–pimF Regulator–pimF pimF–pimA pimF–pimA pimA–pimB pimA–pimB pimB–pimC pimB–pimC pimC–pimD pimC–pimD pimD–pimE pimD–pimE
Primer name
Primer sequence (5§–3§)
iclRjxnfor iclRjxnrev Hyd.lig.F Hyd.lig.R Lig.tran.F Lig.tran.R Tran.lrg.F Tran.lrg.R Lrg.sml.F Lrg.sml.R Sml.OH.F Sml.OH.R
GCG TCG ATT GTC AAC GAC GTC AGC GCA CAC AAG TAA GC GCG GCG GCC CGA TGC ACT ACG CC GCG AGA AAC CTC TCG GCC ATC GG AGG ACA TGA TCA TCT CCG GCG GC CCT CGA GCG CCG GAT CGA CGG CG GGC CAC CCC TAC GGC ATG TCG GG GCG AGG TCG GAG CCC GAG CCC G TTC TGA TGG AAG TGA TCG GCC CG GGT GGA GAC GTC GCC GAG ATC CC ACG CGC CGG CGC GAT CGC GGC C TGC CGC GTT CCG CCA TCT GCG GG
undecanedioate (C11), dodecanedioate (C12), tridecanedioate (C13) and tetradecanedioate (C14). The saturated, straight-chain fatty acids used were: valerate (C5), caproate (C6), heptanoate (C7), caprylate (C8), caprate (C10), laurate (C12), myristate (C14), palmitate (C16) and stearate (C18). Bacterial strains and culture conditions. R. palustris strains
were grown at 30 uC in defined basal medium (PM) aerobically with shaking, or anaerobically in sealed tubes in light, as previously described (Kim & Harwood, 1991). Fatty acid and dicarboxylic acid carbon sources were added to liquid PM medium at a final concentration of either 1?5 or 3 mM, as indicated in the table and figure legends. Pimelate and benzoate were added to a final concentration of 3 mM, and succinate to 10 mM. Anaerobic fatty acid or dicarboxylic acid cultures were supplemented with sodium bicarbonate (10 mM), which was added to PM medium as a sterile solution after autoclaving. Escherichia coli strains DH5a and S17-1 were grown at 37 uC in Luria–Bertani broth. Where indicated, R. palustris was grown with 100 mg gentamicin ml21 or 100 mg kanamycin (Km) ml21. E. coli was grown with 100 mg ampicillin ml21, 20 mg Km ml21 or 20 mg gentamicin ml21. Growth rates were determined by measuring turbidity OD660 in a Genesys20 spectrophotometer. DNA manipulations. Standard protocols were used for cloning and
transformations. All restriction endonucleases and DNA modification enzymes were purchased from New England Biolabs. Shrimp alkaline phosphatase was purchased from Roche Diagnostics. PCR was performed with Pfu DNA polymerase (Stratagene). Chromosomal DNA was purified using a Puregene DNA isolation kit (Gentra systems). DNA fragments were excised and purified from agarose gels using the Qiaquick gel extraction kit (Qiagen), and plasmid DNA was purified with the Qiaprep spin miniprep kit. DNA was sequenced at the University of Iowa DNA core facility by standard automated-sequencing technology. RNA isolation. RNA was isolated from R. palustris strain CGA009
using a modification of the RNeasy kit (Qiagen). Cells (30 ml cultures) grown aerobically with 10 mM succinate and 1?5 mM pimelate were harvested at 5000 g at an OD660 of 0?3. The cell pellets were resuspended in 1 ml TE buffer (10 mM Tris/HCl, 1 mM EDTA, pH 8?0), and incubated with Ready-Lyse Lysozyme (Epicentre) for 15 min at room temperature. Buffer RLT (4 ml; provided with the Qiagen kit) was added to the suspension, and 1 ml aliquots were placed in 2 ml screw-capped tubes containing approximately 1 ml 0?1 mm zirconia/silica beads (BioSpec Products). The suspension was then treated in a Mini-BeadBeater-8 (BioSpec Products) for 1 min periods, with cooling on ice after each period, for a http://mic.sgmjournals.org
total of 5 min. The broken cells and beads were sedimented by centrifugation, and the supernatants were pooled. The beads were washed by mixing with buffer [100 ml TE (pH 8?0) plus 350 ml RLT] in the 2 ml tubes. After centrifuging a second time, the supernatants were again combined. Ethanol (2?8 ml) was added to the combined cell lysate, and the entire sample was applied to a RNeasy Midicolumn. A DNase I digestion was performed on the Midi-column with the Qiagen RNase-Free DNase set. The RNA was eluted from the column, and a second DNase I treatment was performed with RQ1 RNase-Free DNase. The sample was then applied to RNeasy mini-columns to purify the RNA. Manufacturer’s instructions were followed for an on-column DNase treatment with the Qiagen DNase set. Three DNase treatments were required to obtain RNA suitable for use in RT-PCR experiments. RT-PCR. RT-PCR was performed by using a two-tube protocol. To synthesize cDNA, 0?25 mg template RNA was heated to 65 uC for 5 min to relieve possible secondary structure. The RNA was then transcribed into cDNA using the Omniscript RT kit (Qiagen) and an appropriate primer specific for the pim intergenic regions. The Omniscript was heat-inactivated, and PCR was performed using 2 ml of the reverse-transcription reaction as the template, and primers for the intergenic regions of the pim cluster (Table 1). Negative control reactions were performed, in which the Omniscript reverse transcriptase was omitted, to ensure that the products from the PCR originated from the cDNA, and not from contaminating genomic DNA. Construction of a pim cluster deletion mutant. A mutant
strain containing a deletion of the entire cluster of pim genes was constructed by overlap extension PCR (Horton et al., 1993). Briefly, 1 kb DNA was amplified from both the upstream region of pimF and the downstream region of pimE (Fig. 2). The amplified fragments shared 66 bp of complementary DNA, and were stitched together by PCR to form a 2 kb product, termed the deletion product. This deletion product was engineered to contain a SacI and a XbaI site at the 59 and 39 ends, respectively, while a BamHI site was engineered in the middle of the deletion product. The SacI/XbaIdigested deletion product was ligated with similarly digested pUC19. Plasmid containing the deletion product was then digested with BamHI, and treated with shrimp alkaline phosphatase. The 1?3 kb BamHI fragment from pUC4K (Amersham Biosciences), containing a Km-resistance GenBlock, was ligated into the deletion product at its BamHI site. The 3?3 kb deletion fragment containing the Kmresistance gene was then excised from the pUC19 backbone with SacI/XbaI, and cloned into pJQ200 mp18 (Quandt & Hynes, 1993) 729
F. H. Harrison and C. S. Harwood
Fig. 2. Organization of the pim genes in the R. palustris chromosome. Arrows indicate the direction of transcription. The predicted functions of the genes are given, as are the R. palustris gene designations. The small black arrows indicate the locations of primers used for RT-PCR. The expected PCR fragments are designated by short continuous lines, and those fragments that were detected by RT-PCR are labelled +. A product was obtained for the regions between the pim genes, indicating they are co-transcribed and constitute an operon.
to generate pFH164. pFH164 was then introduced into wild-type R. palustris strain CGA009 from E. coli S17-1(pFH164) by conjugation. Sucrose-resistant, Km-resistant colonies were selected as described previously (Egland et al., 1995). The resulting mutant strain was verified by Southern analysis as having a deletion of the pimFABCDE genes (Fig. 3), and was named CGA151.
kb
A
B
C
D
7.0 6.0 5.0
Construction of a pimA : : lacZ mutant. A 2?5 kb DNA fragment,
containing pimA plus 1 kb upstream and downstream flanking DNA, was generated by PCR, purified, and cloned into pBBR1MCS5 (Kovach et al., 1994) to generate pFH101. A 5?25 kb fragment containing a lacZ–Km cassette was excised from BamHI-digested pHRP314 (Parales & Harwood, 1993), treated with Klenow fragment to create blunt ends, and ligated into ScaI-digested and shrimpalkaline-phosphatase-treated pFH101 to generate pFH103. The correct orientation of the cassette within the pimA gene was verified by PCR. A 7?5 kb fragment containing pimA : : lacZ was excised from pFH103 by digestion with BamHI and PstI, and cloned into pJQ200 mp18 (Quandt & Hynes, 1993) to generate pFH304. pFH304 was then introduced into CGA009 as described above. Colonies were screened by PCR for the loss of the wild-type pimA gene, and the presence of the pimA–lacZ junction. The resulting mutant strain was named CGA140. Expression cloning of pimA. Primers StartTrc (59-GCGAAGTGGGATCCATGTCCCATCCCGG-39) and TrcStop (59-GGTTGGAATTCCGAATTACTTGGTCTGTG-39) containing restriction sites (underlined) were used to amplify the pimA gene. The 1?7 kb product was cloned into pTrcHis (Invitrogen) after digestion with BamHI and EcoRI, to generate pHisLigase. The plasmid insert was sequenced to ensure that mutations were not introduced, and that the full-length gene was present in the His-tag fusion vector. Purification of PimA. E. coli DH5a(pHisLigase) was grown at 30 uC with aeration to an OD660 of 0?3. The 750 ml culture was induced with IPTG (1 mM) for 4 h. Cells were harvested, washed,
730
4.0
3.0
2.0 1.6
Fig. 3. Southern blot analysis of the pim deletion mutant CGA151. Genomic DNA was digested, and separated on a 0?8 % agarose gel. The DNA probe was designed to hybridize to the 1 kb region immediately upstream of the pim cluster. In the wild-type strain, a 1?9 or 2?1 kb fragment was expected after digestion with NotI or EcoRI, respectively. To make the deletion mutant, 8?8 kb was removed from the chromosome. Due to the loss of NotI and EcoRI restriction sites within the deleted region, and the addition of the Km cassette, the expected fragment sizes in the mutant strain were 6?4 kb for NotI, and 6?7 kb for EcoRI. Lanes: A, WT digested with NotI; B, WT digested with EcoRI; C, CGA151 digested with NotI; D, CGA151 digested with EcoRI. The positions of size markers are shown schematically on the left. Microbiology 151
Dicarboxylic acid b-oxidation and resuspended in 20 ml binding buffer [20 mM triethanolamine (TEA/HCl) buffer (pH 8?0), 0?5 M NaCl, 5 % (v/v) glycerol]. Cells were lysed by sonication. PMSF (0?05 M) was added to the cell paste, and the mixture was centrifuged at 10 000 g for 20 min at 4 uC. The resulting supernatant was then centrifuged at 40 000 g for 90 min at 4 uC, after the addition of DNase I (1 mg ml21), RNase I (1 mg ml21) and PMSF (0?05 M). The supernatant from this highspeed centrifugation was termed the crude cell extract. All subsequent steps were carried out at 4 uC. The crude cell extract (20 ml; 62 mg) was loaded onto a 5 ml HiTrap chelating column (Pharmacia Biotech) that had been charged with 0?1 M NiSO4, and equilibrated with binding buffer. After extensive washing with binding buffer, the column was developed over 120 min using a Bio-Rad Biologic System with a linear gradient of 0–0?5 M imidazole in binding buffer with a flow rate of 1 ml min21. Fractions (1 ml) were collected, and active fractions determined with the isotopic assay were pooled. Enzyme assays. Acyl-CoA ligase (PimA) activity was measured with isotopic and spectrophotometric assays as previously described (Geissler et al., 1988). The two assays gave comparable results with pimelate, but the isotopic assay was more sensitive, and had less variability, than the spectrophotometric assay. For the isotopic assay, the enzymic conversion of [14C]pimelate to [14C]pimelyl-CoA was measured. Pimelate that was not converted to pimelyl-CoA was extracted from the reaction mixture with ethyl acetate at acidic pH, while pimelyl-CoA remained hydrophilic. The amount of unreacted pimelate in the reaction mixture was detected in controls where enzyme was omitted from the reaction, and it was used in calculations of enzymic activity. The reaction mixture contained 2?5 mM MgCl2, 0?5 mM ATP, 0?25 mM reduced CoA (CoASH) and [14C]pimelic acid (200 mM in standard assay for 555 Bq per reaction) in 20 mM TEA/HCl, pH 8?0. The reaction was initiated by the addition of 20 mg enzyme for a final volume of 0?5 ml. For determination of kinetic constants, the above reaction mixture was used with 0?5 mM CoASH. The amount of labelled substrate remained constant at 20 mM for these assays, while the addition of unlabelled substrate was used to achieve the desired final substrate concentration. This assay was suitable for use at all stages of enzyme purification. [1,7-14C]Pimelate (2?22–3?33 GBq mmol21) was obtained from American Radiolabelled Chemicals.
The spectrophotometric assay measures the formation of AMP by acyl-CoA ligase by coupling the reaction via a series of auxiliary enzymes to NADH oxidation, as previously described (Geissler et al., 1988). The reaction mixture contained 20 mM TEA/HCl (pH 8?0), 2?5 mM MgCl2, 0?5 mM ATP, 0?25 mM CoASH, 10 mM KCl, 10 mM phosphoenol pyruvate, 0?175 mM NADH, substrate (0?25 mM in standard assays), 2 U of the auxiliary enzymes pyruvate kinase and lactic acid dehydrogenase, and 4 U of myokinase (adenylate kinase), in a total volume of 1 ml. The assays were performed using a Beckman DU800 spectrophotometer with a quartz cuvette of 1 cm path length. For determination of kinetic constants, the above mixture was used with 0?5 mM CoASH. The spectrophotometric assay can only be used with pure enzymes due to the very high background activity of NADH oxidation in crude cell extracts. Kinetic constant determination. Enzyme activities were measured
at a range of substrate concentrations. At least nine concentrations were tested in duplicate on two separate days, and used to generate a Hanes–Woolf plot. Maximal enzymic velocity (Vmax) and the Michaelis constant (Km) were determined from the plot. These values were then used to calculate Kcat (enzyme turnover rate), which was determined from Vmax/[E], where [E] is the total enzyme concentration, and Kcat/Km (catalytic efficiency). Southern blotting. Approximately 3 mg chromosomal DNA was
digested with NotI or EcoRI, and separated on a 0?8 % agarose gel. http://mic.sgmjournals.org
A probe was generated by PCR that was complementary to 1 kb DNA directly upstream of pimF. The probe was labelled with the Ready-To-Go DNA labelling beads (Amersham Biosciences), according to the manufacturer’s procedures. Southern hybridizations were performed by standard procedures. Other procedures. SDS-PAGE was carried out with 12?5 % acrylamide gels by standard procedures (Ausubel et al., 1990). Separated proteins were visualized by Coomassie blue R-250. Molecular-mass standards were from Gibco-BRL. Protein concentrations of cell extracts and enzyme preparations were determined with the Bio-Rad protein assay reagent. b-Galactosidase activity was measured as previously described (Egland & Harwood, 1999). Sedimentation equilibrium centrifugation. The native molecular
mass of the PimA protein was determined using a Beckman XL-1 analytical ultracentrifuge in absorbance mode. All experiments were performed at 12 uC in buffer containing 20 mM TEA/HCl (pH 8?0), 0?5 M NaCl and 10 % (v/v) glycerol. A six-channel centrepiece was used for acquiring absorbance data at 280 nm for a protein concentration range of 2–8 mM. Data were collected at rotor speeds of 10 000, 12 000 and 14 000 r.p.m. Equilibrium was achieved when three consecutive scans taken 2 h apart were unchanged. Data editing was performed with the WinREEDIT program (version 0.999.0028), and fitting of the data was performed with the WinNONLIN program (version 1.06.0048); both programs were developed at the the National Analytical Ultracentrifugation Facility, University of Connecticut Biotechnology Center; http://www.ucc. uconn.edu/~wwwbiotc/AUFMAIN.HTML).
RESULTS Examination of the R. palustris genome for b-oxidation genes The genome of R. palustris has over 160 genes predicted to encode enzymes that might be involved in the b-oxidation of fatty acids and dicarboxylic acids (Larimer et al., 2004; GenBank accession no. BX571963). However, these genes tend to be scattered around the genome, and we identified only a single cluster of genes predicted to encode all the enzymes necessary for dicarboxylic acid or fatty acid boxidation (Fig. 2). This cluster, annotated as the pim genes, includes an acyl-CoA ligase gene (pimA), an acyl-CoA acetyltransferase gene (pimB), a short-chain alcohol dehydrogenase gene (pimE), an enoyl-CoA hydratase (pimF) gene, and two genes, pimC and pimD, predicted to encode flavin-containing dehydrogenases. The genome sequence indicates that the pimE gene has a frameshift mutation. It is annotated as a pseudogene, and is expected to be nonfunctional in R. palustris. The pim genes are organized as an operon, and are required for optimal growth on benzoate and dicarboxylic acids To examine the transcriptional organization of the genes in the pim cluster, we carried out RT-PCR using primers designed to amplify the intergenic regions (Table 1). Products were obtained for all of the intergenic regions in the cluster, but not between the upstream regulatory genes rpa3718 and pimF (Fig. 2). No product was obtained in controls to which no reverse transcriptase was added. These 731
F. H. Harrison and C. S. Harwood
results indicate that the pimFABCDE genes are cotranscribed, and constitute an operon. To examine the contribution of the pim operon to growth with dicarboxylic acids and fatty acids, we constructed a mutant (CGA151) in which the pimFABCDE genes were deleted from the chromosome. PCR and Southern hybridization experiments (Fig. 3) each verified that the expected deletion had occurred in CGA151. The pimFABCDE deletion mutant grew more slowly than the wild-type parent when pimelate (C7), azelate (C9) or tetradecanedioate (C14) was supplied as the sole carbon source under anaerobic photoheterotrophic growth conditions (Table 2). The pim operon deletion strain was also impaired in growth on benzoate, a compound that R. palustris degrades only under anaerobic conditions. The pim operon deletion strain was impaired in aerobic growth on pimelate and azelate. We were unable to obtain consistent growth of wild-type R. palustris with other dicarboxylic and fatty acids under aerobic conditions (Table 2). The pim operon deletion mutant grew anaerobically at wild-type rates on the fatty acid caprylate (C8). This could mean that the pim operon is not involved in the degradation of this fatty acid. Another interpretation is that R. palustris encodes at least two sets of enzymes that can catalyse caprylate degradation, one of which is encoded by the pim operon. The pimA gene is induced by growth with benzoate, fatty acids or dicarboxylic acids
cultures of strain CGA140 were compared with the activities of cultures grown on succinate plus various dicarboxylic acids and fatty acids (Fig. 4). During aerobic growth, dicarboxylic acids from 7 to 14 carbons in length, and the C8 fatty acid caprylic acid, induced pimA : : lacZ expression about twofold over succinate-grown cells. Anaerobically, pimelate or benzoate induced pimA : : lacZ expression about threefold above the levels seen when CGA140 was grown with succinate alone, and b-galactosidase activities in cultures grown under anaerobic conditions with caprylate were twofold higher than in succinate-grown cultures. Strain CGA140 resembled the pim operon deletion strain in that it grew more slowly than its wild-type parent with pimelate and benzoate. When grown with pimelate as a sole carbon source anaerobically in light, the levels of b-galactosidase activity expressed by the pimA : : lacZ mutant strain were similar to those observed when the strain was grown on pimelate and succinate, indicating that succinate does not repress expression of the pimA gene. Characterization of PimA, a broad-substraterange acyl-CoA ligase Because it is difficult to prepare the appropriate CoA thioester intermediates, we were not able to judge the efficiencies of most of the Pim enzymes with a range of possible substrates. We therefore focused our efforts on PimA, the enzyme that is predicted to activate unmodified
We used a strain carrying a pimA : : lacZ chromosomal fusion (strain CGA140) to examine the influence of various carbon compounds on the expression of the pim gene cluster. The b-galactosidase activities of succinate-grown
Table 2. Growth rates of the pim deletion mutant strain CGA151 Carbon source*
Aerobic Succinate (C4-DA) Azelate (C9-DA) Pimelate (C7-DA) Anaerobic Succinate (C4-DA) Caprylate (C8-FA) Azelate (C9-DA) Benzoate (C7-AA) Pimelate (C7-DA) Tetradecanedioate (C14-DA)
Doubling time (h)D WT (CGA009)
CGA151
12?1±2?5 23?6±2?0 25?0±3?4
12?0±1?3 30?5±2?9 55?5±10?3
8?5±0?3 8?2±1?5 7?7±2?2 8?4±0?6 10?3±2?2 11?0±2?5
8?7±0?4 10?2±0?7 22?1±4?4 15?5±3?5 25?8±4?5 63?7±17?2
*Carbon sources were used at a final concentration of 3 mM, except succinate, which was supplied at 10 mM. See the legend of Fig. 4 for key to abbreviations. DValues are the means±SD of three independent cultures assayed in duplicate. 732
Fig. 4. b-Galactosidase activities expressed by the pimA : : lacZ fusion strain CGA140. Activities are expressed as nmol product formed min”1 (mg protein)”1. Cultures were grown on 10 mM succinate plus 1?5 mM test substrate as indicated (number of carbons followed by DA for dicarboxylic acid, FA for fatty acid, AA for aromatic acid): C4-DA, succinate; C7-DA, pimelate; C9DA, azelate; C10-DA, sebacate; C11-DA, undecanedioate; C14DA, tetradecanedioate; C8-FA, caprylate; C7-AA, benzoate. The activities are the means (±SD) of measurements of b-galactosidase activities in cell extracts prepared from at least six independent cultures. Each culture was assayed in duplicate. Microbiology 151
Dicarboxylic acid b-oxidation
Table 3. Purification of the expressed PimA protein from E. coli cells Purification step Cell extract Ni2+ chelating
Protein (mg)
Total activity (U)*
Specific activity (U mg”1)
Activity recovered (%)
Purification (fold)
66 0?57
1?795 0?143
0?0272 0?25
100 22
9
*Units are measured in mmol pimelyl-CoA formed min21.
dicarboxylic acids with CoA to initiate their b-oxidation. An N-terminal His-tagged version of the predicted acylCoA ligase PimA was expressed in E. coli, and purified in a single chromatographic step (Table 3, Fig. 5). Pimelate was the substrate used to follow activity during purification because it is available in radiolabelled form, and it is necessary to use an isotopic assay to measure activity in crude cell extracts. The molecular mass of the purified PimA protein, as determined by SDS-PAGE, was approximately 60 000 Da. The pimA gene sequence predicts a protein of 552 amino acids, with a molecular mass of 60 029 Da. We used analytical ultracentrifugation to determine the oligomeric state of the enzyme. The native molecular mass as determined by sedimentation equilibrium analysis was 59 100 Da. This indicates that native PimA is a monomer. The purified enzyme catalysed the addition of reduced CoA to pimelate at pH 8?0. The activity was unaffected by the presence of oxygen. The enzyme activity was optimal at pH 8?0–8?5.
good activity. These included azelate, sebacate, undecanedioate, tetradecanedioate and caprylate. Kinetic constants for pimelate were determined using the isotopic assay because of its greater sensitivity (Table 4). Caprylic acid, the C8 fatty acid, was the substrate best utilized by the acylCoA ligase. PimA had the highest affinity and the highest catalytic efficiency with caprylate. PimA had similar catalytic efficiencies and turnover rates with azelate, sebacate, tetradecanedioate and undecanedioate (chain lengths defined in Methods). Although PimA was active with pimelate, this dicarboxylic acid was a very poor substrate when compared with other dicarboxylic acids. Cell extracts of R. palustris strain CGA009 had a pimelate-CoA ligase activity of 10 nmol pimelyl-CoA formed min21 (mg protein)21. PimA had no detectable activity with the
PimA has a broad substrate range, and is active with C7–C14 dicarboxylic acids, and C6–C16 fatty acids (Fig. 6). Kinetic constants were determined with a coupled enzymic assay (see Methods) for a subset of substrates that gave
1
2
3
114 81 60
64 50
37
Fig. 5. SDS-PAGE analysis of active protein fractions obtained during the purification of the PimA acyl-CoA ligase. Lanes: 1, molecular-mass marker; 2, crude extract (15 mg); 3, active pooled fractions from the Ni2+ chelating column (1?5 mg). Numbers to the left of the gel represent molecular mass in kDa. http://mic.sgmjournals.org
Fig. 6. Specific activities of the PimA ligase with various dicarboxylic and fatty acid substrates. Each substrate was present in assay mixtures at a final concentration of 250 mM. Activities were measured using the spectrophotometric assay described in Methods. Pimelate was measured with the isotopic assay. Dicarboxylic acids: C7, pimelate; C8, suberate; C9, azelate; C10, sebacate; C11, undecanedioate; C12, dodecanedioate; C13, tridecanedioate; C14, tetradecanedioate. Fatty acids: C6, caproate; C7, heptanoate; C8, caprylate; C10, caprate; C12, laurate; C14, myristate; C16, palmitate. The enzyme was not active with the following substrates: adipate, valerate, stearate and benzoate. Specific activities are means±SD of at least six independent determinations on three different days. 733
F. H. Harrison and C. S. Harwood
Table 4. Kinetic constants of the purified PimA acyl-CoA ligase Data were obtained using the spectrophotometric assay described in Methods. Substrates are followed by the number of carbons in the compound, and designation as either a dicarboxylic (DA) or a fatty acid (FA). Rates were determined by plotting the data for nine concentrations in duplicate. Although similar values for pimelate were obtained using the isotopic or spectrophotometric assay, results using the isotopic assay are reported because of its greater sensitivity. Substrate
Kcat (s”1)
Km (mM)
Kcat/Km (M”1 s”1)
Pimelate (C7-DA) Caprylate (C8-FA) Azelate (C9-DA) Sebacate (C10-DA) Undecanedioate (C11-DA) Tetradecanedioate (C14-DA)
0?03 3?7 3?6 3?3 3?0 2?6
277 4 230 177 47 5?16104
C6-dicarboxylic acid adipate, and it was not active with the fatty acids valerate or stearate, or with benzoate.
DISCUSSION Mammalian mitochondria and many bacteria, including E. coli, degrade medium- and long-chain-length fatty acids by well-characterized b-oxidation pathways. Although dicarboxylic acid degradation has received much less attention, many bacterial species are known to grow with dicarboxylic acids having a range of carbon chain lengths (Gallus & Schink, 1994; Hoet & Stanier, 1970; Parke et al., 2001). Also, some aromatic-compound-degrading bacteria form pimelyl-CoA, and others form 3-hydroxypimelyl-CoA, as the immediate product of ring cleavage during anaerobic benzoate degradation (Harwood et al., 1999). By analogy with aerobic aromatic compound degradation pathways, the metabolism of the benzoate degradation intermediates pimelyl-CoA or 3-hydroxypimelyl-CoA to three molecules of acetyl-CoA plus one molecule of CO2 can be considered to be the ‘lower’ anaerobic benzoate degradation pathway. Data presented here indicate that in R. palustris the pimFABCDE genes are cotranscribed and organized in an operon that encodes enzymes for the b-oxidation of dicarboxylic acids of between 7 and 14 carbons, including pimelate (Fig. 1b). The slow growth phenotype of a pim operon deletion mutant (Table 2) is consistent with the hypothesis that the Pim enzymes also catalyse the lower part of the benzoate degradation pathway. Benzoate and dicarboxylates with an odd number of carbons are predicted to be converted to glutaryl-CoA. Dicarboxylates with an even number of carbons would be converted to succinyl-CoA, and thus enter directly into the citric acid cycle. We cannot formally exclude that the pim operon somehow functions biosynthetically rather than catabolically. However, this seems unlikely because the growth 734
defect of the pim deletion strain is not a general growth defect, but instead is specific to only a subset of the growth substrates that R. palustris degrades. The R. palustris genome has genes predicted to mediate conversion of glutaryl-CoA to acetyl-CoA and CO2 (Larimer et al., 2004) (Fig. 1b). It has a predicted glutaryl-CoA dehydrogenase/glutaconyl-CoA decarboxylase gene (rpa1094), and also encodes predicted 3-hydroxybutyryl-CoA dehydratase (Rpa4339), 3-hydroxybutyryl-CoA dehydrogenase (Rpa4748) and acetoacetyl-CoA thiolase (Rpa0531) enzymes for the conversion of crotonyl-CoA, the product of glutaconylCoA decarboxylation, to two acetyl-CoA (Fig. 1b). The Pim enzymes are obviously not the only R. palustris enzymes that can catalyse dicarboxylic acid degradation because the pim operon deletion strain is still able to grow, albeit more slowly than the wild-type, on these compounds. The R. palustris genome encodes 45 CoA ligases, 32 flavincontaining dehydrogenases, 26 enoyl-CoA hydratases, many short-chain alcohol dehydrogenases and more than 10 acylCoA acetyltransferases. It is likely that combinations of these other b-oxidation enzymes can also contribute to the degradation of dicarboxylic acids. In E. coli, the genes encoding enzymes of aerobic fatty acid b-oxidation are scattered around the chromosome (Clark & Cronan, 1996). The R. palustris pimE ORF is disrupted by a 21 frameshift at nucleotide 4192036. This mutation is expected to render the gene nonfunctional. Thus an alternate R. palustris protein must catalyse the dehydrogenation reaction that is part of the b-oxidation sequence of reactions. It is possible that PimF has 3-hydroxyacyl-CoA dehydrogenase as well as 3-hydroxyacyl-CoA dehydratase activity. PimF possesses the characteristic amino acid motifs to perform both of these functions. A single gene product, FadB, catalyses these two reactions during fatty acid b-oxidation in E. coli (Yang et al., 1988). Another possibility is that one of the other 54 short-chain alcohol dehydrogenases encoded elsewhere on the R. palustris chromosome has 3-hydroxyacylCoA dehydrogenase activity. The PimA protein appears to have one of the broadest substrate ranges of any acyl-CoA ligase described, as it is active with both dicarboxylic and fatty acids of medium and long carbon chain length. Acyl-CoA ligases involved in the b-oxidation of fatty acids, including FadD from E. coli (Kameda & Nunn, 1981), have been described that are active with a broad range of fatty acids; however, these enzymes were not reported to have been tested for activity with dicarboxylic acids. Pimelyl-CoA is the first committed intermediate in the synthesis of the vitamin biotin, and several pimelate-CoA ligases have been purified and characterized in the course of efforts to develop a biotechnological process for the production of biotin (Binieda et al., 1999; Ploux et al., 1992). PimA shares approximately 30 % amino acid identity with these proteins, but the region of identity is mainly confined to the shared AMPbinding domain common to CoA ligases. The described Microbiology 151
Dicarboxylic acid b-oxidation
pimelate-CoA ligases have a narrow substrate range, and are specific to pimelate. A search of the current databases revealed that the only organism with a set of genes that is very similar to the pim genes is the nitrogen-fixing soybean symbiont Bradyrhizobium japonicum (Kaneko et al., 2002). This set of genes lies at nucleotide positions 8567718–8573337 in the B. japonicum genome. Applying pim designations to these genes, B. japonicum possesses the following: bll7817 (pimD), bll7818 (pimC), bll7819 (pimB), bll7820 (pimA) and bll7821 (pimF). These genes have between 73 and 89 % deduced amino acid identity with the homologous R. palustris genes. Based on this high degree of relatedness, we predict a role for these genes in aerobic dicarboxylic acid degradation by B. japonicum. B. japonicum does not have an anaerobic pathway for benzoate degradation. The pimE gene is absent from the B. japonicum genome pim gene cluster. An IclR-like regulatory gene, and genes for an ABC transport system related to the branched chain amino acid (Ilv) uptake system of E. coli, are divergently transcribed from the pim cluster in R. palustris (Fig. 2). R. palustris encodes 20 Ilv-type transporters, and we have speculated that these may be specific for various sorts of hydrophobic compounds, as well as for dicarboxylic acids (Larimer et al., 2004). We predict that the protein products of genes rpa3719–rpa3725 are responsible for transporting substrates for the Pim enzymes, and that the multiple periplasmic binding proteins encoded by this gene cluster enable the transport of a range of dicarboxylic and fatty acids. In addition, we speculate the IclR regulator could control both transport and degradation of these compounds. The B. japonicum genome has an identical gene cluster at position 8575477, directly upstream of its putative pim genes, consisting of an IclR-like regulator, and a transport system with multiple periplasmic binding proteins. The R. palustris and B. japonicum regulatory and transport genes share between 76 and 88 % amino acid identity, while two of the three binding proteins in B. japonicum share approximately 78–80 % identity with those in R. palustris.
ACKNOWLEDGEMENTS
Binieda, A., Fuhrmann, M., Lehner, B., Rey-Berthod, C., FrutigerHughes, S., Hughes, G. & Shaw, N. M. (1999). Purifica-
tion, characterization, DNA sequence and cloning of a pimeloylCoA synthetase from Pseudomonas mendocina 35. Biochem J 340, 793–801. Boll, M., Fuchs, G. & Heider, J. (2002). Anaerobic oxidation of
aromatic compounds and hydrocarbons. Curr Opin Chem Biol 6, 604–611. Clark, D. P. & Cronan, J. E. (1996). Two-carbon compounds and
fatty acids as carbon sources. In Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, pp. 343–357. Edited by F. C. Neidhardt and others. Washington, DC: American Society for Microbiology. Donoghue, N. A. & Trudgill, P. W. (1975). The metabolism
of cyclohexanol by Acinetobacter NCIB 9871. Eur J Biochem 60, 1–7. Egland, P. G. & Harwood, C. S. (1999). BadR, a new MarR family
member, regulates anaerobic benzoate degradation by Rhodopseudomonas palustris in concert with AadR, an Fnr family member. J Bacteriol 181, 2102–2109. Egland, P. G., Gibson, J. & Harwood, C. S. (1995). Benzoate-
coenzyme A ligase, encoded by badA, is one of three ligases able to catalyze benzoyl-coenzyme A formation during anaerobic growth of Rhodopseudomonas palustris on benzoate. J Bacteriol 177, 6545–6551. Elshahed, M. S., Bhupathiraju, V. K., Wofford, N. Q., Nanny, M. A. & McInerney, M. J. (2001). Metabolism of benzoate, cyclohex-1-ene
carboxylate, and cyclohexane carboxylate by ‘‘Syntrophus aciditrophicus’’ strain SB in syntrophic association with H2-using microorganisms. Appl Environ Microbiol 67, 1728–1738. Gallus, C. & Schink, B. (1994). Anaerobic degradation of pimelate
by newly isolated denitrifying bacteria. Microbiology 140, 409–416. Geissler, J. F., Harwood, C. S. & Gibson, J. (1988). Purification and
properties of benzoate-coenzyme A ligase, a Rhodopseudomonas palustris enzyme involved in the anaerobic degradation of benzoate. J Bacteriol 170, 1709–1714. Gibson, J. & Harwood, C. S. (2002). Metabolic diversity in aromatic
compound utilization by anaerobic microbes. Annu Rev Microbiol 56, 345–369. Ha¨rtel, U., Eckel, E., Koch, J., Fuchs, G., Linder, D. & Buckel, W. (1993). Purification of glutaryl-CoA dehydrogenase from Pseudo-
monas sp., an enzyme involved in the anaerobic degradation of benzoate. Arch Microbiol 159, 712–717. Harwood, C. S., Burchhardt, G., Herrmann, H. & Fuchs, G. (1999).
Anaerobic metabolism of aromatic compounds via the benzoyl-CoA pathway. FEMS Microbiol Rev 22, 439–458. Hasegawa, Y., Hamano, K. & Obata, H. (1982). Microbial degrada-
We thank Gary Heil for his assistance with the analytical ultracentrifugation. The Division of Energy Biosciences, US Department of Energy (grant DE-FG02-95ERZ0184), and the US Army Research Office (grant W911NF-04-1-0123), supported this research. F. H. H. was supported by a National Science Foundation research training grant (DBI9602247).
dicarboxylic acids by Pseudomonas fluorescens. Eur J Biochem 13, 65–70.
REFERENCES
Kameda, K. & Nunn, W. (1981). Purification and characterization of
tion of cycloheptanone. Agric Biol Chem 46, 1139–1143. Hoet, P. P. & Stanier, R. Y. (1970). The dissimilation of higher
Horton, R. M., Ho, S. N., Pullen, J. K., Hunt, H. D., Cai, Z. & Pease, L. R. (1993). Gene splicing by overlap extension. Methods Enzymol
217, 270–279.
Ausubel, F. M., Brent, R., Kingston, R. E., More, D. D., Seidman, J. G., Smith, J. A. & Struhl, K. (1990). Current Protocols in Molecular
Biology. New York: Greene Publishing Associates. Bernards, M. A. & Razen, F. A. (2001). The poly(phenolic) domain
of potato suberin: a non-lignin cell wall bio-polymer. Phytochemistry 57, 1115–1122. http://mic.sgmjournals.org
acyl coenzyme A synthetase from Escherichia coli. J Biol Chem 256, 5702–5707. Kaneko, T., Nakamura, Y., Sato, S. & 14 other authors (2002).
Complete genomic sequence of nitrogen-fixing symbiotic bacterium Bradyrhizobium japonicum USDA110. DNA Res 9, 189–197. Kester, A. S. & Foster, J. W. (1963). Diterminal oxidation of long-
chain alkanes by bacteria. J Bacteriol 85, 859–869. 735
F. H. Harrison and C. S. Harwood Kim, M.-K. & Harwood, C. S. (1991). Regulation of benzoate-CoA
ligase in Rhodopseudomonas palustris. FEMS Microbiol Lett 83, 199–204.
Ploux, O., Soularue, P., Marquet, A., Gloeckler, R. & Lemoine, Y. (1992). Investigation of the first step of biotin biosynthesis in
Kovach, M. E., Phillips, R. W., Elzer, P. H., Roop, R. M., II & Peterson, K. M. (1994). pBBR1MCS: a broad-host-range cloning vector.
Bacillus sphaericus. Purification and characterization of the pimeloylCoA synthase, and uptake of pimelate. Biochem J 287, 685–690.
Biotechniques 16, 800–802.
Quandt, J. & Hynes, M. F. (1993). Versatile suicide vectors which
Larimer, F. W., Chain, P., Hauser, L. & 16 other authors (2004).
allow direct selection for gene replacement in Gram-negative bacteria. Gene 127, 15–21.
Complete genome sequence of the metabolically versatile photosynthetic bacterium Rhodopseudomonas palustris. Nat Biotechnol 22, 55–61. Parales, R. E. & Harwood, C. S. (1993). Construction and use of a
new broad-host-range lacZ transcriptional fusion vector, pHRP309, for Gram-negative bacteria. Gene 133, 23–30. Parke, D., Garcia, M. A. & Ornston, L. N. (2001). Cloning and genetic characterization of dca genes required for b-oxidation of
straight-chain dicarboxylic acids in Acinetobacter sp. strain ADP1. Appl Environ Microbiol 67, 4817–4827.
736
Schocke, L. & Schink, B. (1999). Energetics and biochemistry of
fermentative benzoate degradation by Syntrophus gentianae. Arch Microbiol 171, 331–337. Yang, S. Y., Li, J. M., He, X. Y., Cosloy, S. D. & Schulz, H. (1988).
Evidence that the fadB gene of the fadAB operon of Escherichia coli encodes 3-hydroxyacyl-coenzyme A (CoA) epimerase, delta 3-cisdelta 2-trans-enoyl-CoA isomerase, and enoyl-CoA hydratase in addition to 3-hydroxyacyl-CoA dehydrogenase. J Bacteriol 170, 2543–2548.
Microbiology 151
