Identification of the Gene Cluster for the Anaerobic Degradation of 3,5-Dihydroxybenzoate (α-Resorcylate) in Thauera aromatica Strain AR-1

Abstract

Thauera aromatica strain AR-1 degrades 3,5-dihydroxybenzoate (3,5-DHB) with nitrate as an electron acceptor. Previous biochemical studies have shown that this strain converts 3,5-DHB to hydroxyhydroquinone (1,2,4-trihydroxybenzene) through water-dependent hydroxylation of the aromatic ring and subsequent decarboxylation, and they suggest a pathway homologous to that described for the anaerobic degradation of 1,3-dihydroxybenzene (resorcinol) by Azoarcus anaerobius. Southern hybridization of a T. aromatica strain AR-1 gene library identified a 25-kb chromosome region based on its homology with A. anaerobius main pathway genes. Sequence analysis defined 20 open reading frames. Knockout mutations of the most relevant genes in the pathway were generated by reverse genetics. Physiological and biochemical analyses identified the genes for the three main steps in the pathway which were homologous to those described in A. anaerobius and suggested the function of several auxiliary genes possibly involved in enzyme maturation and intermediate stabilization. However, T. aromatica strain AR-1 had an additional enzyme to metabolize hydroxyhydroquinone, a putative cytoplasmic quinone oxidoreductase. In addition, a specific tripartite ATP-independent periplasmic (TRAP) transport system was required for efficient growth on 3,5-DHB. Reverse transcription-PCR (RT-PCR) analysis showed that the pathway genes were organized in five 3,5-DHB-inducible operons, three of which have been shown to be under the control of a single LysR-type transcriptional regulator, DbdR. Despite sequence homology, the genetic organizations of the clusters in T. aromatica strain AR-1 and A. anaerobius differed substantially.

INTRODUCTION

Aromatic compounds are ubiquitous substrates for bacteria in natural habitats. Pathways of aerobic microbial degradation of aromatics have been studied in detail, but our knowledge on anaerobic degradation pathways is comparatively scarce, and the diversity of pathways is certainly underestimated. Anaerobic degradation differs substantially from aerobic degradation, as the well-known O₂-dependent oxygenase reactions cannot be applied in the absence of molecular oxygen. In general, anaerobic degradation attacks the aromatic ring by a reduction reaction to overcome its mesomery-enhanced stability (1, 2). The most common central intermediate formed by bacteria in the absence of oxygen is benzoyl coenzyme A (benzoyl-CoA). Numerous compounds, such as toluene, benzoate, phenol, cresols, and phenylacetate, among others, are transformed through different peripheral pathways to this central intermediate (3). Benzoyl-CoA is reductively dearomatized to render 1,5-dienoyl-CoA, which undergoes further reduction and ring opening through a series of reactions similar to β-oxidation (4). Di- and trihydroxylated aromatic compounds such as resorcinol (1,3-dihydroxybenzene), phloroglucinol (1,3,5-trihydroxybenzene), or hydroxyhydroquinone (1,2,4-trihydroxybenzene) can be dearomatized in fermenting and sulfate-reducing bacteria by reductive reactions directly without channeling reactions (1, 2, 5).

In contrast, two denitrifying bacteria use an entirely different degradation strategy for the anaerobic degradation of dihydroxylated aromatics. The obligately denitrifying betaproteobacterium Azoarcus anaerobius attacks resorcinol by an oxidative rather than a reductive reaction and hydroxylates the aromatic ring at position 4 by using a resorcinol hydroxylase to form hydroxyhydroquinone (HHQ) (6) (Fig. 1). Hydroxylation of resorcinol can be detected in vitro in the membrane fraction of A. anaerobius with K₃Fe(CN)₆ as an electron acceptor (6–8). In the second step, hydroxyhydroquinone is oxidized to 2-hydroxy-1,4-benzoquinone (HBQ). Further metabolism of 2-hydroxy-1,4-benzoquinone to malate and acetate proceeds via so-far-unknown reactions (Fig. 1).

FIG 1.

The T. aromatica strain AR-1 3,5-DHB degradation pathway and the A. anaerobius resorcinol degradation pathway converge in the central intermediate hydroxyhydroquinone.

A similar pathway was described for the degradation of 3,5-dihydroxybenzoate (3,5-DHB) (α-resorcylate) in the denitrifying betaproteobacterium Thauera aromatica strain AR-1. This strain also uses a water-dependent oxidative reaction to convert 3,5-DHB to hydroxyhydroquinone in two steps: the aromatic ring is hydroxylated at position 2, yielding 2,3,5-trihydroxybenzoate, which is decarboxylated to hydroxyhydroquinone in a reaction that is stimulated in the presence of an unidentified cytoplasmic factor (9) (Fig. 1). The initial 3,5-DHB-hydroxylating activity is membrane associated, and a membrane-bound hydroxyhydroquinone dehydrogenase activity was also found in cell extracts of 3,5-DHB-induced cultures (10). This suggests that the anaerobic 3,5-DHB degradation pathway of T. aromatica strain AR-1 parallels A. anaerobius resorcinol degradation, with hydroxyhydroquinone as the common central intermediate (Fig. 1). To our knowledge, these two bacteria are so far the only ones known to use this oxidative pathway to degrade an aromatic compound in the absence of oxygen.

This unexplored metabolic pathway was first accessed at the genetic level by heterologous expression of an A. anaerobius cosmid library in T. aromatica strains (7), which allowed the identification of a cosmid encompassing the gene cluster for resorcinol degradation. Mutational analysis identified the genes for the two first reactions in resorcinol degradation, rehLS for the resorcinol hydroxylase and btdLS for the hydroxyhydroquinone dehydrogenase, and the genes coding for an enzyme system initiating the further degradation of the product hydroxybenzoquinone, bqdLMS (7).

It is so far not possible to grow A. anaerobius on agar surfaces, impeding a more detailed genetic analysis of the pathway. In contrast, we have been able to adapt and apply genetic tools to the manipulation of T. aromatica strain AR-1, providing the possibility to use the 3,5-DHB degradation pathway in this strain as model to further characterize this conserved oxidative pathway. On the other hand, the regulation of these pathways is completely unknown at the molecular level. The main enzymes of the pathway are induced by the presence of the substrate, although in T. aromatica, the 3,5-DHB degradation pathway is also subject to catabolite repression by benzoate (10). The second goal of this study was to identify regulators for 3,5-DHB degradation in T. aromatica strain AR-1.

We initially identified the gene cluster for the pathway based on sequence similarity to key genes of the resorcinol degradation cluster in A. anaerobius, and we characterized mutants with insertion mutations in relevant genes. With the present study, we have gained access to elucidating at the genetic level this anaerobic pathway which uses oxidative reactions to overcome the aromatic structure, thus increasing our knowledge of the diversity of the microbial metabolism of aromatic compounds.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

The bacterial strains and plasmids used in this work are summarized in Table 1. Thauera aromatica strain AR-1 and its mutant derivatives and transconjugants were cultured anaerobically at 30°C without shaking in 50- or 100-ml infusion bottles containing nonreduced Widdel mineral medium under nitrogen gas (6). The medium was buffered with 30 mM 3-(N-morpholino)propanesulfonic acid (MOPS) instead of bicarbonate and supplemented with 8 mM potassium nitrate. Carbon sources stored in sterile infusion bottles under nitrogen gas were added to the cultures with syringes to the specified final concentrations. For all genetic manipulations, T. aromatica strain AR-1 and derivatives were grown aerobically in Widdel mineral medium supplemented with 5 mM succinate or glutarate as the carbon source. Solid media were prepared with 1.6% twice-washed Difco agar. Escherichia coli strains HB101, DH5α, and CC118λpir were grown aerobically at 37°C in Luria-Bertani (LB) medium. When appropriate, antibiotics were used at the following concentration: tetracycline, 10 μg ml⁻¹; chloramphenicol, 30 μg ml⁻¹; ampicillin, 100 μg ml⁻¹; kanamycin, 50 μg ml⁻¹, streptomycin, 50 μg ml⁻¹, and gentamicin, 10 μg ml⁻¹ (except for T. aromatica, for which tetracycline and kanamycin were used at 5 μg ml⁻¹ and 25 μg ml⁻¹, respectively).

TABLE 1.

Bacterial strains and plasmids used in this work

Strain, plasmid, or cosmid	Genotype or relevant characteristics	Reference
Strains
Escherichia coli
HB101	supE44 hsdS20 (rB− mB−) recA13 ara-14 proA2 lacY1 galK2 rpsL20 (Smr) xyl-5 mtl-1	35
CC118	Δ(ara-leu) araD ΔlacX74 galE galK phoA20 thi-1 rpsE (Spr) rpoB (Rifr) argE recA1	36
CC118λpir	CC118 lysogenized with λpir	37
DH5α	endA1 hsdR17 supE44 thi-1 recA1 gyrA (Nalr) relA1 Δ(argF-lac)U169 depR (ϕ80dlacΔ(lacZ)M15)	11
Pseudomonas putida KT2440		38
Thauera aromatica
AR-1	Wild-type strain, degrades 3,5-DHB under denitrifying conditions (DSM-11528)	39
AR-1 orf3	Kmr, T. aromatica AR-1 mutant bearing pPMorf3 plasmid integrated in the chromosome	This work
AR-1 korAB	Gmr, T. aromatica AR-1 mutant, korAB genes partially deleted and replaced by a gentamicin resistance cassette	This work
AR-1 dbhL	Kmr, T. aromatica AR-1 mutant bearing pAMdbhL plasmid integrated in the chromosome	This work
AR-1 dbhS	Kmr, T. aromatica AR-1 mutant bearing plasmid pAMdbhS integrated in the chromosome.	This work
AR-1 dctP	Kmr, T. aromatica AR-1 mutant bearing plasmid pAMorf13 integrated in the chromosome.	This work
AR-1 btdL	Gmr, T. aromatica AR-1 mutant, btdL gene disrupted by a gentamicin resistance cassette	This work
AR-1 orf17	Kmr, T. aromatica AR-1 mutant bearing plasmid pPMorf17 integrated in the chromosome	This work
AR-1 ofr18	Kmr, T. aromatica AR-1 mutant bearing plasmid pAMorf18 integrated in the chromosome	This work
AR-1 bqdM	Kmr, T. aromatica AR-1 mutant bearing plasmid pAMbqdM integrated in the chromosome	This work
AR-1 orf20	Kmr, T. aromatica AR-1 mutant bearing plasmid pAMorf20 integrated in the chromosome	This work
AR-1 orf21	Kmr, T. aromatica AR-1 mutant bearing plasmid pAMorf21 integrated in the chromosome	This work
AR-1 bqdL	Kmr, T. aromatica AR-1 mutant bearing plasmid pAMbqdL integrated in the chromosome	This work
AR-1 dbdR	Kmr, T. aromatica AR-1 mutant bearing plasmid pPMdbdR integrated in the chromosome	This work
AR-1 qorA	Gmr, T. aromatica AR-1 mutant, qorA gene disrupted by a gentamicin resistance cassette	This work
AR-1 orf27	Kmr, T. aromatica AR-1 mutant bearing plasmid pPMorf27 integrated in the chromosome	This work
AR-1 btdL qorA	Kmr, T. aromatica AR-1 qorA derivative bearing an internal deletion in the btdL gene	This work
Plasmids
pCHESIΩKm	Apr, Kmr, pUC18 bearing RP4 oriT and the Ω interposon from pHP45ΩKm as HindIII fragment.	12
pKNG101	Smr, Sacs, RK2oriV	40
pRK600	Cmr, ColE1oriV, RK2mob+tra+	41
pJB3Tc19	Broad-host range IncP1 cloning vector, Tcr, Apr	16
pPMorf3	Internal fragment of orf3 (669 bp) cloned between the KpnI and BamHI sites of pCHESIΩKm	This work
pPMkorAB	Gentamicin resistance cassette flanked by the upstream sequence of korA (842 bp) and the end sequence of korB and its downstream region (715 bp) cloned in pKNG101	This work
pPMbtdL	btdL gene and its flanking region (2,089 bp) interrupted by a Kmr cassette, cloned in pKNG101	This work
pAMdbhL	Internal fragment of dbhL (411 bp) cloned between the SacI and EcoRI sites of pCHESIΩKm	This work
pAMdbhS	Internal fragment of dbhS (584 bp) cloned between the SacI and EcoRI sites of pCHESIΩKm	This work
pAMorf13	Internal fragment of orf13 (650 bp) cloned between the SacI and EcoRI sites of pCHESIΩKm	This work
pPMorf17	Internal fragment of orf17 (546 bp) cloned between the EcoRI and XbaI sites of pCHESIΩKm	This work
pAMorf18	Internal fragment of orf18 (819 bp) cloned between the SacI and EcoRI sites of pCHESIΩKm	This work
pAMbqdM	Internal fragment of bqdM (738 bp) cloned between the SacI and EcoRI sites of pCHESIΩKm	This work
pAMorf20	Internal fragment of orf20 (700 bp) cloned between the SacI and EcoRI sites of pCHESIΩKm	This work
pAMorf21	Internal fragment of orf21 (1,010 bp) cloned between the SacI and EcoRI sites of pCHESIΩKm	This work
pAMbqdL	Internal fragment of bqdL (863 bp) cloned between the SacI and EcoRI sites of pCHESIΩKm	This work
pPMdbdR	Internal fragment of dbdR (677 bp) cloned between the SacI and BamHI sites of pCHESIΩKm	This work
pPMqorA	qorA gene and its flanking regions (780 and 970 bp) interrupted by a gentamicin resistance cassette, cloned in pKNG101	This work
pPMorf27	Internal fragment of orf27 (877 bp) cloned in the EcoRI site of pCHESIΩKm	This work
pJB-dbhLS	pJB3Tc19 derivative harboring the dbhLS genes under control of the Plac promoter	This work
pJB-orf20	pJB3Tc19 derivative harboring the orf20 gene under control of the Plac promoter	This work
pJB-bqdS	pJB3Tc19 derivative harboring the bqdS gene under control of the Plac promoter	This work
pJB-bqdL	pJB3Tc19 derivative harboring the bqdL gene under control of the Plac promoter	This work
Cosmids
pLAFR3	Tcr, cos, RK2oriV, RK2oriT	42
pCOS4	Tcr, pLAFR3 containing a 23.6-kb chromosomal fragment from T. aromatica AR-1	This work
pCOS12	Tcr, pLAFR3 containing a chromosomal fragment from T. aromatica AR-1	This work
pCOS19	Tcr, pLAFR3 containing a chromosomal fragment from T. aromatica AR-1	This work
pCOS2B	Tcr, pLAFR3 containing a chromosomal fragment from T. aromatica AR-1	This work
pCOS6B	Tcr, pLAFR3 containing a chromosomal fragment from T. aromatica AR-1	This work

Growth experiments.

To analyze 3,5-DHB degradation and nitrate reduction by T. aromatica strain AR-1 and mutant derivatives, 100-ml serum bottles containing 75 ml of anaerobic Widdel mineral medium supplemented with the required antibiotics and 3,5-DHB (1 mM) and/or succinate (2 mM) were inoculated with 1% of the exponentially growing wild-type and mutant T. aromatica strain AR-1 cultures and incubated at 30°C. Samples were taken anoxically with a sterile syringe flushed with N₂ at the designated time points and used immediately for determination of optical density at 600 nm (OD₆₀₀) in a Shimadzu UV-266 spectrophotometer.

Construction and screening of a genome library of Thauera aromatica strain AR-1.

Preparation of plasmids and chromosomal DNA, digestion with restriction endonucleases, ligation, agarose gel electrophoresis, and Southern blotting were performed using standard methods (11). To prepare a genome library of Thauera aromatica strain AR-1, 11 μg of genomic DNA from the strain was partially digested with PstI, and 20- to 30-kb fragments were ligated to a PstI-digested and dephosphorylated pLAFR3 vector. Gigapack III XL packaging extract (Stratagene) was used to package the recombinant DNA. The phage particles obtained were transfected into E. coli HB101, and colonies were grown on LB agar plates with tetracycline. The colonies harvested and pooled in LB liquid medium constituted the gene library. Screening of approximately 3,400 clones of the library was carried out by colony hybridization with randomly digoxigenin (DIG)-labeled probes obtained by PCR amplification of internal fragments of the A. anaerobius orf14, rehL, and bqdL genes with specific primers using a DIG-DNA labeling kit (Roche). Twenty cosmids hybridizing with at least one of the probes were initially selected, and finally a 23-kb-insert cosmid (pCOSM4) was shotgun sequenced. Probes against the two ends of this insert were generated as described above and used to select four additional cosmids covering its flanking regions. The four cosmids were pooled and pyrosequenced. Two final contigs could be assembled: one of them (13,363 bp) was homologous to the Thauera sp. strain MZ1T chromosome and contained a gene cluster related to nitrate respiration, while the other one (36,128 bp) showed homology with the Azoarcus anaerobius resorcinol degradation cluster. No connecting sequence between the two cosmids could be found.

Cosmid sequencing.

The selected cosmids from the T. aromatica strain AR-1 library were either shotgun sequenced at Macrogen Inc. (Seoul, South Korea) or pyrosequenced using a Roche 454 GSFLX titanium system with 20× coverage at GATC-Biotech (Constance, Germany). Small gaps between the resulting contigs were covered using PCR, and controversial PstI sites were checked by sequencing the corresponding PCR-amplified region from T. aromatica genomic DNA.

Site-specific homologous inactivation of genes in the T. aromatica strain AR-1 3,5-DHB degradation cluster.

Mutant strains with inactivated chromosomal versions of orf3, dbhL, dbhS, orf13, orf17, orf18, bqdM, orf20, orf21, bqdL, and orf25 were generated by single homologous recombination using plasmid pCHESIΩKm, a pUC18-based vector bearing the oriT transfer origin of RP4 and the Ω-Km interposon of pHP45ΩKm (12) (Table 1). To generate the desired mutation, an internal fragment (minimum size, 500 bp) of the target gene was amplified by PCR with the appropriate primers (Table 2) and cloned in pMBLT vector (Dominion-mbl). The resulting plasmids were cut with the appropriate restriction enzymes and cloned in the corresponding sites of pCHESIΩKm polylinker. The resulting plasmids were mobilized into T. aromatica strain AR-1 by triparental mating (see below), and Km^r recombinant clones were selected. The nature of the mutation in selected kanamycin-resistant clones was confirmed by PCR using a pCHESIΩKm-based primer and a second primer located in the region flanking the cloned truncated fragment, as described previously (12). A clone of each mutant was selected, and the correct insertion of pCHESIΩKm was confirmed by Southern blotting. Mutant strains with inactivated chromosomal versions of korAB, btdL, and qorA were generated by double homologous recombination using plasmid pKNG101 as described previously (13). In short, the gene of interest flanked by 0.5- to 1-kb neighboring regions was amplified with suitable primers (Table 2), cloned in pMBLT, and then inactivated by insertion of the gentamicin resistance cassette from pBBR-MCS5 (14) (for korAB and qorA) or the kanamycin resistance gene from pACYC177 (15). The resulting mutated region was cloned in the suicide plasmid pKNG101 and was mobilized into T. aromatica strain AR-1 by conjugation (see below). Transconjugants selected as gentamicin resistant and streptomycin-sensitive clones were confirmed by PCR and Southern blotting. The use of sucrose for the selection of double recombinants was not efficient in T. aromatica strain AR1 and was skipped in our screening protocol. The bqdL qorA double mutant was obtained in the same way, except that the pKNG101 derivative containing the kanamycin-interrupted btdL gene was transferred to the T. aromatica qorA mutant strain.

TABLE 2.

Oligonucleotide primers used in this work

Primer pair	Sequences (5′ to 3′)	Location (producta)	Expected size (bp)
rt-korA_3′/rt-korB_5′	AACAGAGCCACGGCGCCCAG/CGAACACGTTGGTGTAGGCG	orf5/orf6 intergenic region (1)	350
rt-korB_3′/rt-rhL_5′	GAAGGCCAGCGTCGATTCCG/GTGCGCATCTGCACGGTGAG	orf6/dhbL intergenic region (2)	580
FupK/RdK	CATCTGCGCGACAAGCAG/GTCCGGAGCCGGATCGAG	korAB and flanking regions	3,672
rhLF/rhLR	GACGCAGCGCTCAGCGACG/AACGCCATCGGCACGACC	dhbL internal region	411
rhSF/rhSR	AGAACAGCGGATCGACGGC/GGCCCGGGAGGTGGCCGG	dhbS/orf9 intergenic region (4)	584
rt-rhL_3′/rt-rhS_5′	CGTCAATCAGCTCACGCCG/GGCCGAGCTTGGCCATTTC	dhbL/dhbS intergenic region (3)	277
rt-rhS_3′/rt-cup_5′	TTCGCCATGGAAGTGGAAG/GCACGGTCGAGACGAGCTTG	dhbS/orf9 intergenic region (5)	160
rt-cup_3′/rt-hyp_5′	TCGATCCGCTGTTCTACTTC/TGACGTCGTAGTCCTTGCCG	orf9/orf10 intergenic region (6)	380
rt-hyp_3′/rt-deh_5′	TCCCTCAACGGCAAGGACTAC/TGCCCATCGCCAGATGGACG	orf10/btdL intergenic region (7)	448
rt-deh_3′/rt-m24_5′	GATCGCCGACGAGACCTATG/GGATAGTCGATGCCGAGTTC	btdL/orf12 intergenic region (8)	396
orf9F/orf12R	CTGTTCTACTTCGAATTCCAGG/TCAGGCACTGGATCTCCTGC	btdL and flanking regions	2,089
rt-m24_3′/rt-dctP_5′	AAGTGGTGGTGACGAAGGAC/CCGTCCTCGTATAGATGTG	orf12/dbtP intergenic region (10)	329
m24F/m24R	GGAACTCGGCATCGACTATC/TGCTGGCCATCGACCACCTC	orf12 internal fragment (9)	450
rt-dctP_3′/rt-smallper_5′	TGCGCAAGCAGTTCATCGAC/GGCGCCATGAGGACGTAGC	dbtP/dbtQ intergenic region (11)	259
rt-smallper_3′/rt-dctM_5′	CCACCAGGCCGTGCAGTCG/AAGCGGAATGGCGAGGAACC	dbtQ/dbtM intergenic region (12)	236
rt-esr_5′/rt-αβhyd_3′	CTTGATGATCTTCTGCGG/TGGTGCCGATCGAAGACAGC	orf17/orf16 intergenic region (13)	330
rt-mob_3′/rt-bqdhM_5′	ACAACCAGCCGCCCTAC/TCATAGCTGTTGTTGTCG	orf18/bqdM intergenic region (14)	167
rt-bqdhM_3′/rt-suc_5′	AAGCCGAGCTGATCGCCAC/AGGTCGGGATCGGTCAATTC	bqdM /orf20 intergenic region (15)	338
rt-suc_3′/rt-p47k_5′	TTCTACACCTGCGACCTGG/GCAGCACACGCAACCGTTGG	orf20/orf21 intergenic region (16)	515
rt-p47k_3′/rt-bqdhL-5′	TCGACATCGAGCAGGACGAG/TCGAGATGGGCGAGGCACTTG	orf21/bqdL intergenic region (17)	495
rt-bqdhL_3′/rt-hypprot_5′	ACAGCGTGCGCAAGACCAAC/CCGATCACCTCGACGAAG	bqdL /orf23 intergenic region (18)	400
rt-hypprot_3′/rt-pydh_5′	CCAGACCTGGGTCTATTTCC/CAGGATCTTGCCGAGAATGC	orf23/bqdS intergenic region (19)	338
rt-pydh_3′/rt-lysR_5′	TGTCGGTCGATCATCGCGTG/TTTTCGTCACCGACGACTGC	bqdS/dbdR intergenic region (20)	289
lysRR/lysRR2	GCGAAGGTGCTGTTCCTGTC/GAACAACCGCTTCGGCACC	dbdR internal region	677
rt-lysR_3′/rt-Zaldeh_5′	GATGCCGGCGTGGAACATCG/TCGCGTCGCTCGGATTGACC	dbdR/qorA intergenic region (21)	430
FupQ/RdQ	CGGGCAATTCGATCTCGG/TGAGCAACACCTGATTTTCC	qorA and flanking regions	1,814
HindIIIdbhLSF/XbaIdbhLSR	AAGCTTGGGCTGGATAGACTGCCAGGCATCG/TCTAGACGCCACGTCGATCACCAGATTCCACT	dbhL/dbhS coding sequence	3,979
HindIIIbqdMF/XabIbqdMR	AAGCTTACCCTCGTGCCGCCGCCCTCC/TCTAGAGCCGGCGCACAGCAATCAGATGTG	bqdM coding sequence	1,492
HindIIIbqdLF/XbaIbqdLR	AAGCTTGCGCAGTGCCCGGCACC/AAGCTTGCGCAGTGCCCGGCACC	bqdL coding sequence	2,264
HindIIIorf20F/XabIorf20R	AAGCTTCCATTTCCGAAAATCTTCGAGGAC/TCTAGACGGCCCTCGCCTCATGCCGGAATC	orf20 coding sequence	1,552
HindIIIorf21/XbaIorf21R	AAGCTTACGTCAAACGCGGTGCC/TCTAGACGACGGCATCGGTTCGATTC	orf21 coding sequence	1,154

^aThe numbers in parentheses refer to the PCRs in Fig. 2.

Plasmid construction.

For the complementation of the mutant strains, derivatives of broad-host-range plasmid pJB3Tc19 (16) harboring the gene of interest were constructed as follows. The DNA region encompassing the gene(s) of interest was amplified by PCR in a total volume of 50 μl containing 150 ng chromosomal DNA, 2.6 U Expand high-fidelity DNA polymerase (Roche, Mannheim, Germany), a 300 nM concentration of the appropriate oligonucleotides (Table 2), 3% (vol/vol) dimethyl sulfoxide (DMSO), 300 nM deoxynucleoside triphosphates (dNTPs), and 1× Expand high-fidelity buffer with 1.5 mM MgCl₂. The PCR products were purified with the QIAquick gel extraction kit (Qiagen) and cloned in pGEMT Easy (Promega). The resulting plasmids were sequenced to rule out the presence of possible mutations and digested with HindIII/XbaI, and the fragment harboring the amplified gene(s) was ligated with the pJB3Tc19 vector digested with the same enzymes. The resulting constructs bearing the different pathway genes under the control of the P_lac promoter were transferred to T. aromatica strain AR1 through triparental conjugation.

Triparental conjugation.

Mutations were transferred to T. aromatica strain AR-1 chromosome by RP4-mediated mobilization. T. aromatica strain AR-1 was grown to saturation in Widdel minimal medium supplemented with succinate (5 mM). The E. coli donor strain CC118λpir bearing the corresponding mutagenic pCHESIΩKm or pKNG101 derivatives and the E. coli HB101(pRK600) helper strain were grown overnight on LB medium in the presence of kanamycin or chloramphenicol, respectively. All steps were performed aerobically as follows. One milliliter of E. coli cultures and 15 ml of the T. aromatica strain AR-1 culture were harvested by centrifugation (13,000 × g for 10 min at 10°C). Cell pellets were washed once with 1 ml of Widdel minimal medium. The three resulting cell pellets were combined in 100 μl minimal medium, distributed on a sterile 47-mm-diameter, 0.22-μm-pore-size filter (Schleicher and Schuell, Germany) placed on an LB plate, and incubated overnight at 30°C. The filters were transferred into 1 ml of minimal medium, and cells were washed off by vigorous vortexing. T. aromatica strain AR-1 transconjugants where selected by their resistance to kanamycin (pCHESIΩKm mutants) or gentamicin (double homologous recombination mutants obtained with pKNG101 derivatives) either anaerobically in Widdel minimal medium supplemented with succinate (5 mM) or aerobically in Widdel minimal medium supplemented with glutarate (5 mM) to counterselect against E. coli donor and helper strains. The correct insertion of the mutations was confirmed by PCR analysis and Southern blotting. Complementation plasmids were transferred to T. aromatica AR-1 mutant strains by RP4-mediated mobilization as described above, except that T. aromatica AR-1 mutant transconjugants where selected aerobically by their resistance to tetracycline in Widdel minimal medium supplemented with glutarate (5 mM).

RNA preparation.

Thauera aromatica strain AR-1 was grown at 30°C under nitrate-reducing conditions using 2 mM α-resorcylate, 5 mM succinate, or both compounds as carbon sources. Cells (45 ml) were harvested by centrifugation (8,000 × g, 5 min, 4°C) in disposable plastic tubes precooled in liquid nitrogen, and the pellets were kept at −80°C until use. RNA was extracted using the TRI reagent method (Ambion, Austin, TX) with the following modifications: the lysis step was carried out at 60°C, and a final digestion step with RNase-free DNase was added at the end of the process. The RNA concentration was determined with a NanoDrop instrument (Thermo Scientific), and RNA integrity was assessed by agarose gel electrophoresis.

RT-PCR assays.

Reverse transcriptase PCR (RT-PCR) was done with 400 ng RNA in a final volume of 50 μl using the Titan OneTube RT-PCR system according to the manufacturer's instructions (Roche Laboratories). cDNA was synthesized with random primers at 50°C for 45 min, and the PCR cycling conditions were as follows: 94°C for 20 s, 35 cycles at an adequate annealing temperature for 20 s, and extension at 68°C for 20 s. The annealing temperature was calculated for each reaction based on the melting temperatures of the pair of primers used. Positive (T. aromatica strain AR-1 DNA) and negative (total RNA of the T. aromatica wild type and mutant strains without reverse transcriptase) controls were included in all assays. The primers used to test cotranscription are listed in Table 2.

Preparation of cell extracts.

All steps in the preparation of cell extracts were performed under anoxic conditions. Cultures of T. aromatica strain AR-1 and its knockout mutants were grown in 500-ml infusion bottles with a mixture of 1 mM 3,5-DHB and 2 mM succinate as the carbon source. Cells were harvested and washed once with 100 ml of 50 mM potassium phosphate buffer (pH 7.0). Unless used immediately, cell pellets were frozen in liquid nitrogen and stored at −20°C. To prepare cell extracts, cells were suspended in 50 mM potassium phosphate buffer (pH 7) and passed through a French press at 100 MPa. The crude extract was separated from cell debris by centrifugation at 27,000 × g for 20 min at 4°C. Protein content was quantified with the Bradford method (17).

Determination of 3,5-DHB hydroxylase activity.

All measurements of enzyme activities were performed under anoxic conditions at 30°C in 5-ml Hungate tubes or in 1.5-ml cuvettes using anoxic buffers and solutions. Tubes and cuvettes were flushed with nitrogen gas and closed with butyl septa. All additions and samplings were performed with gas-tight Unimetrix microliter syringes (Macherey-Nagel, Düren, Germany). Linear correlations of protein amount and reaction rates were checked for all assays. The 3,5-DHB-oxidizing activity, which catalyzes the hydroxylation of 3,5-DHB to 2,3,5-trihydroxybenzoate and is located in the membrane fraction (9), was measured with K₃Fe(CN)₆ as an electron acceptor in a photometric assay following the reduction of K₃Fe(CN)₆ at 420 nm. The assay mixture contained 50 mM Tris-HCl (pH 8.0), cell extract (1 mg protein), and 1 mM K₃Fe(CN)₆, and the reaction was started by the addition of 1 mM 3,5-DHB. The rate of 3,5-DHB oxidation was calculated from the K₃Fe(CN)₆ reduction rate based on a 2:1 stoichiometry of electron acceptor to electron donor. One unit of 3,5-DHB hydroxylase activity is defined as the amount of enzyme required to convert 1 μmol of 3,5-DHB in 1 min.

Analytical determinations.

To detect the presence of hydroxyhydroquinone in the cultures, culture samples (1 to 4 μl) were dried under a nitrogen flow to evaporate water, and the dried pellet was directly resuspended in 50 μl of the derivatizing agent N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA)–trimethylchlorosilane (TMCS) (99:1). The resulting samples were analyzed by gas chromatography-mass spectrometry (GC-MS) in a Varian GC-240MS gas chromatograph coupled with a Varian 240 MS ion trap mass spectrometer. The system was equipped with electronic flow control, a 1079 universal capillary injector, and a CTC CombiPal autosampler. Analytes were separated on a SLB-5ms column (30 m by 0.25 mm by 0.25 μm) (Supelco, Sigma-Aldrich). The flow rate was 1 ml/min using He as the carrier gas. The column temperature was held at 50°C for 5 min and then increased at 10°C/min to 300°C, which was held for 5 min. The sample (1 μl) was injected at a 250°C injector temperature in splitless mode (splitless time, 5 min). The mass spectrometer was operated in electron impact ionization mode, acquiring in full scan (50 to 800 atomic mass units), and SIM mode (recording the most abundant ions). The trap and transfer line temperatures were 240°C and 250°C, respectively. MS Workstation software version 6.9.1 was used for instrument control. Detection of compounds was performed using both the NIST08 database and pure standards for comparison. The presence of other intermediates in the culture medium was detected by high-pressure liquid chromatography (HPLC) with a C₁₈ reverse-phase column (Grom, Herrenberg, Germany) and a UV detector (Beckman). The mobile phase was a 1:1 mixture of 100 mM ammonium acetate (pH 2.6) and methanol at a flow rate of 1 ml/min.

Sequence analysis.

Nucleotide and amino acid sequences were analyzed using the tools provided by the NCBI (http://blast.ncbi.nlm.nih.gov/Blast.cgi) and the ExPASy molecular biology server (http://www.expasy.org). The functional annotation of the new gene clusters was performed with Blast2GO based on values for either average similarity to a group of proteins or maximum identity to a specific protein (https://www.blast2go.com/blast2go-pro/download-b2g) (18). Protein alignments were performed with ClustalW (http://www.ebi.ac.uk/Tools/msa/clustalw2), and the resulting alignments were displayed with boxshade (http://www.ch.embnet.org/software/BOX_form.html). Computational models of the protein three-dimensional structure were built with the Phyre2 Server (http://www.sbg.bio.ic.ac.uk/phyre2/) (19). The following servers and database were used to analyze some of the protein sequences: DAS transmembrane prediction server (http://www.sbc.su.se/∼miklos/DAS/) (20), MEROPS peptidase database (http://merops.sanger.ac.uk) (21), and ESTHER esterase and α/β hydrolase enzyme database (http://bioweb.ensam.inra.fr/ESTHER/general?what=index) (22).

Correction of A. anaerobius annotation.

The gene nomenclature of the previously described A. anaerobius resorcinol degradation cluster (accession number EF078692) (7) has been modified to comply with standard genetic nomenclature, as follows: rhLS, bqdhLMS, and btdhLS have been changed to rehLS, bqdLMS, and btdLS, respectively. Moreover, the rehL start codon has been relocated 366 nucleotides upstream from the initially proposed nucleotide based on its homology to dbhL, so that the protein is now 122 residues longer (see Fig. S1a in the supplemental material); a new gene, coding for a cupin family protein, was annotated as orf12b between btdS and orf13. The first 8 nucleotides of the orf12b 5′ end overlapped with orf13 3′ end.

Nucleotide sequence accession number.

The T. aromatica strain AR-1 3,5-DHB-degradation cluster sequence identified in this work has been deposited in GenBank under accession number KJ995609.

RESULTS

Identification of the gene cluster responsible for the anaerobic catabolism of 3,5-DHB in Thauera aromatica strain AR-1.

T. aromatica strain AR-1 and A. anaerobius share an oxygen-independent hydroxylation reaction as the initial step for the degradation of 3,5-DHB and resorcinol, respectively (Fig. 1). We expected genes homologous to those of the A. anaerobius resorcinol degradation pathway to be present in the T. aromatica strain AR-1 chromosome. To verify this hypothesis, we performed a Southern blot of T. aromatica strain AR-1 chromosomal DNA cut with different restriction enzymes and hybridized it under nonstringent conditions with probes against A. anaerobius pathway genes rehL, coding for the resorcinol hydroxylase large subunit, and bqdL, coding for the hydroxybenzoquinone dehydrogenase large subunit, which are the first- and the third-step enzymes in the resorcinol degradation pathway, respectively (Fig. 1) (7). Both probes gave hybridization bands with A. anaerobius and T. aromatica strain AR-1 DNAs, while no signal was observed with Pseudomonas putida KT2440 DNA, used as a negative control (see Fig. S1 in the supplemental material). To map the T. aromatica strain AR-1 gene cluster for the 3,5-DHB degradation pathway, we constructed a cosmid library of the T. aromatica strain AR-1 chromosome. Several cosmids were selected after colony hybridization with different A. anaerobius probes and were sequenced to finally assemble a 36,128-bp fragment showing homology with the A. anaerobius resorcinol degradation cluster (see Materials and Methods). Analysis of this sequence defined 28 open reading frames (ORFs), three of which (orf8, orf9, and orf24) used the alternative start codon GUG. In all of them except orf1, orf3, orf7, and orf19, a conserved Shine-Dalgarno sequence (consensus, AGGAGG) could be identified at 4 to 10 nucleotides upstream from the start codon. The ORFs were compared with sequences in the databases. As expected, several ORFs showed a high degree of identity to genes of the A. anaerobius resorcinol degradation pathway (Fig. 2 and Table 3; see Table S1 in the supplemental material).

FIG 2.

Transcriptional organization of the T. aromatica strain AR-1 genomic region coding for anaerobic 3,5-DHB degradation. (A) Genes having a homolog in A. anaerobius resorcinol degradation cluster are shaded in gray. A pseudogene with no stop codon (labeled 4) is depicted as a box with no arrowhead. The truncated orf28 at the end of the cluster is labeled 28′. A triangle above the gene indicates that a knockout mutant with an impaired (black) or wild-type (white) growth phenotype has been obtained. An asterisk indicates a mutant with a deletion encompassing part of korAB genes. The operons identified in this work are shown below the genes as solid (inducible) or dashed (constitutive) arrows. Curved arrows above the genes show the presence of putative promoters, as deduced from our results. The group of genes constituting the degradation cluster is boxed. (B) RT-PCR analysis of the 3,5-DHB degradation cluster transcripts. Agarose gels show the RT-PCR products with template RNA isolated from T. aromatica strain AR-1 cells grown on succinate (s) or 3,5-DHB (α). The converging black arrows and underlined numbers indicate the amplified region and refer to the PCR products shown below them: M, molecular marker; d, positive control with T. aromatica DNA.

TABLE 3.

Properties of the genes present in the T. aromatica AR-1 3,5-DHB degradation cluster and their gene products

Gene	Properties					Related gene product
	Protein accession no.	% GC content	Distance to next gene (bp)	Product size (amino acids/kDa)	Product pI	% similaritya	% identityb	E value	Accession no.	Hypothetical function of closest homolog	Organism
orf1	AIO06081	69	8	410/44.8	9.07	68	74	2e−148	ACK54553.1	Glycosyltransferase	Thauera sp. MZ1T
orf2	AIO06100	65	626	260/29.8	7.38	74	83	5e−110	ACK54554.1	Metalophosphoesterase	Thauera sp. MZ1T
orf3	AIO06082	67	204c	308/33.2	9.59	53	56	1e−63	EED66667.1	LysR-type transcriptional regulator	Comamonas testosteroni
orf5	AIO06101	70	2	576/60.5	6.38	69	100	0.0	CAA12243.2	Oxoglutarate-ferredoxin oxidoreductase, α subunit	Thauera aromatica
orf6	AIO06083	70	376	307/32.6	6.70	79	97	4e−157	CAD27440.1	Oxoglutarate-ferredoxin oxidoreductase, β subunit	Thauera aromatica
dbhL	AIO06084	65		981/110.6	7.02	60	60	0.0	ABK58620.1	Resorcinol hydroxylase, α subunit	Azoarcus anaerobius
dbhS	AIO06085	65	15	288/32.3	7.10	58	53	3e−79	ABK58619.1	Resorcinol hydroxylase, β subunit	Azoarcus anaerobius
orf9	AIO06103	62	20	137/15.9	5.65	50	29	5e−06	EFW04169.1	Mannose-1-phosphate guanylyl transferase	Coprobacillus sp. 29_1
orf10	AIO06086	68	30	114/11.9	6.06	45	28	0.35	EEP26949.1	Unknown	Abiotrophia defectiva
btdL	AIO06087	74	102	293/29.5	6.0	63	50	8e−69	ABM15958.1	6-Phosphogluconate dehydrogenase	Mycobacterium vanbaalenii
orf12	AIO06088	67	98	417/46.5	5.50	68	93	0.0	ABK58632.1	M24 peptidase family	Azoarcus anaerobius
dbtP	AIO06102	68	72	321/35.2	6.99	53	40	2e−56	CAL95041.1	TRAP transporter, periplasmic protein	Azoarcus sp. BH72
dbtQ	AIO06104	67	4	163/17.8	9.73	52	34	8e−07	CAL95042.1	TRAP transporter, small permease	Azoarcus sp. BH72
dbtM	AIO06105	64	123	423/44.4	5.21	64	50	6e−95	CAL95043.1	TRAP transporter, permease	Azoarcus sp. BH72
orf16	AIO06089	66	100	344/37.9	7.77	66	57	2e−108	ACB33601.1	TRAP transporter, extracytoplasmic receptor	Leptothrix cholodnii sp-6
orf17	AIO06090	74	382	237/25.3	7.68	65	60	6e−72	EFI59106.1	αβ-Hydrolase	Comamonas testosteroni
orf18	AIO06091	67	64	716/80.0	6.16	60	53	0.0	CAJ71291.1	Molybdopterin oxidoreductase	“Candidatus Kuenenia stuttgartiensis”
bqdM	AIO06092	69	193	467/49.9	6.26	80	75	0.0	BAI72155.1	Dihydrolipoamide dehydrogenase	Azospirillum sp. B510
orf20	AIO06093	68	89	499/53.0	5.49	83	72	0.0	ACD96487.1	Succinic semialdehyde dehydrogenase	Geobacter lovleyi SZ
orf21	AIO06094	70	71	354/38.5	6.21	59	46	2e−81	ABK58627.1	p47k family protein	Azoarcus anaerobius
bqdL	AIO06095	70	12	735/78.7	5.64	75	74	0.0	ABK58621.1	Benzoquinone dehydrogenase, α subunit	Azoarcus anaerobius
orf23	AIO06096	68	32	182/20.2	5.95	72	73	1e−74	ABK58617.1	Glyoxalase	Azoarcus anaerobius
bqdS	AIO06106	72	52	437/44.5	5.98	65	54	5e−121	CAM75497.1	Pyruvate/α-ketoglutarate dehydrogenase complex, E2 component	Magnetospirillum gryphiswaldense
dbdR	AIO06107	68	134	317/34.5	7.20	52	33	1e−36	AAZ63058.1	LysR-type transcriptional regulator	Ralstonia eutropha
qorA	AIO06097	71	232	335/34.9	8.42	75	68	9e−148	EHP44701.1	Oxidoreductase	Cupriavidus basilensis
orf27	AIO06098	74	233	145/15.9	8.6	84	79	5e−81	ACK53423.1	Thioredoxin	Thauera sp. MZ1T
orf28d	AIO06099	72	92			78	83	0	ACK53440.1	(S)-Mandelate dehydrogenase	Thauera sp. MZ1T

^aAverage percent similarity to the 20 closest relatives in the databases, as determined with Blast2GO.

^bPercent identity with the closest homolog.

^cDistance to a pseudogene with no stop codon.

^dData for the available partial sequence.

Genes of the first pathway enzyme, 3,5-DHB hydroxylase.

The products of orf7 and orf8 showed 60 and 53% identity to the α and β subunits of the A. anaerobius resorcinol hydroxylase, respectively, suggesting that these genes coded for 3,5-DHB hydroxylase. The genes were annotated as dbhL and dbhS. We constructed site-directed insertion mutants of the different genes in the T. aromatica chromosome, using either single or double recombination (see Materials and Methods). dbhL and dbhS knockout mutants were unable to grow with 3,5-DHB as a carbon source, confirming the crucial role of these genes in 3,5-DHB degradation (Fig. 3A). Both mutants could be complemented by a broad-host-range plasmid bearing the two genes under the control of the P_lac promoter, ruling out possible polar effects on downstream genes. However, growth both on succinate (data not shown) and on 3,5-DHB was significantly delayed, which we attributed to inefficient growth in the presence of two antibiotics (Fig. 3B). Resorcinol hydroxylase activity was determined in cell extracts of wild-type, dbhL mutant, and dbhS mutant T. aromatica cells grown with 2 mM succinate in the presence of 1 mM 3,5-DHB to induce the pathway. Figure 4 shows that 3,5-DHB hydroxylase was impaired in both mutants, although the dbhS mutant reached low activity levels after 72 h, suggesting that a different protein could substitute for this electron transfer protein with low efficiency.

FIG 3.

(A) Growth of T. aromatica strain AR-1 and the indicated mutants with 5 mM succinate (gray) or 2 mM 3,5-DHB (black) as a carbon source. Growth was determined as OD₆₆₀ after 72 h, except for the btdL, bqdM, orf18, and orf21mutants (*), for which growth values were determined after 168 h. Average values from at least 3 repetitions (± standard deviation [SD]) are referred to the growth reached by the wild type with each substrate. OD₆₆₀ values for the wild-type strain grown on succinate and 3,5-DHB after 72 h were 0.456 (±0.036) and 0.175 (± 0.045), respectively. (B) Complementation of mutant strains with the corresponding wild-type gene. T. aromatica strain AR-1 (black), its mutant derivatives (gray), and their complemented strains with pJB3Tc bearing the corresponding wild-type gene (white) were cultivated on 3,5-DHB (2 mM) at 30°C. Growth was determined as OD₆₆₀ after 192 h, except for the dbhL mutant and the mutant complemented with pJB:dbhLS, which was determined after 240 h. Average values from at least 3 repetitions (±SD) are shown.

FIG 4.

Specific activity of 3,5-DHB hydroxylase in crude extracts of wild-type and mutant T. aromatica strain AR-1 cells. Extracts from cells grown for 48 h (black) and 72 h (white) on Widdel medium supplemented with 3,5-DHB (1 mM) plus succinate (2 mM) were obtained as indicated in Materials and Methods. Activity was determined as the rate of 3,5-DHB-dependent K₃F(CN)₆ reduction in extracts containing 1 mg of protein.

We could identify a molybdopterin-binding domain in the DbhL sequence, which classified this protein in the MopB superfamily of molybdopterin-binding proteins (see Fig. S2a in the supplemental material). Proteins known to bind a molybdopterin cofactor generally require a dedicated chaperone to bind the molybdenum-charged cofactor and further direct the protein to the membrane (23) (see below).

Two genes are required for efficient hydroxyhydroquinone oxidation.

The product of the 3,5-DHB hydroxylase reaction is the unstable intermediate 2,3,5-trihydroxybenzoate, which is further transformed to hydroxyhydroquinone in the cytoplasm (9) (Fig. 1). The enzyme responsible for this reaction has not been identified, although spontaneous decarboxylation of 2,3,5-trihydroxybenzoate has also been suggested (9). Hydroxyhydroquinone is then oxidized to hydroxybenzoquinone, a reaction mediated by hydroxyhydroquinone dehydrogenase, which has been detected in the membrane fraction of 3,5-DHB-grown T. aromatica strain AR-1 cells (10). The gene product of orf11 in the T. aromatica cluster showed 35% identity to the A. anaerobius large subunit of hydroxyhydroquinone dehydrogenase, encoded by the btdL gene. We annotated orf11 as btdL. Growth of a btdL mutant on 3,5-DHB was significantly affected (Fig. 3A). In addition, an orf26 mutant showed reduced growth on 3,5-DHB (Fig. 3A) and after 24 h accumulated a pink intermediate that was identified as hydroxyhydroquinone by GC-MS (Fig. 5), suggesting that metabolism of this compound was impaired in the orf26 mutant. Hydroxyhydroquinone was shown to be highly toxic to T. aromatica cells even at low concentrations (D. Pacheco, unpublished data), and accumulation of this intermediate in the orf26 mutant probably caused growth arrest. The product of orf26 showed 64% identity to a hypothetical quinone oxidoreductase belonging to the medium-chain reductase/dehydrogenase (MDR) family (24). Oxidoreductases of this type catalyze the conversion of alcohols to aldehydes or ketones, suggesting that the orf26 gene product would carry out the oxidation of hydroxyhydroquinone to hydroxybenzoquinone. We annotated the orf26 gene as qorA. Thus, efficient processing of hydroxyhydroquinone during 3,5-DHB degradation in T. aromatica strain AR-1 required the joint activity of both enzymes. As expected, a btdL and qorA double mutant was unable to grow on 3,5-DHB (Fig. 3A).

FIG 5.

(A and B) GC-MS analysis of the culture supernatant of a T. aromatica strain AR-1 qorA mutant grown for 72 h on Widdel medium supplemented with 3,5-DHB as a carbon source. (C and D). The GC-MS elution profiles of 3,5-DHB (C) and hydroxyhydroquinone (D) standards are shown for comparison. The mass spectrum of the corresponding trimethylsilyl derivative is included for each standard compound.

Genes for hydroxybenzoquinone degradation.

The products of orf19, orf22, and orf24 showed 65, 74, and 53% identity, respectively, with gene products of A. anaerobius; these were likely to constitute the third step in the resorcinol degradation pathway and were annotated as bqdM, bqdL, and bqdS (Fig. 2 and Table 3; see Table S1 in the supplemental material). These genes showed a high degree of similarity with the three components of the multienzyme complex of pyruvate dehydrogenase. The bqdL gene was essential for growth on 3,5-DHB (Fig. 3A). Growth on 3,5-DHB was restored when the pJB-bqdL plasmid, harboring the wild-type bqdL gene under the control of P_lac, was introduced into the mutant strain (Fig. 3B). A mutant with a mutation in bqdM, encoding the predicted medium-sized component of the enzyme, was also impaired in its capacity to grow on 3,5-DHB, but it still grew on this substrate at a lower rate and reached significant growth values after 1 week (Fig. 3A). This mutant accumulated a colored intermediate in the culture medium; HPLC analysis of the culture supernatant showed two unidentified peaks with retention times of 14.48 min and 10.16 min, which increased with time to reach maximum values when the substrate was completely consumed (not shown).

Genes of unknown function common to the A. anaerobius and T. aromatica strain AR-1 dihydroxy aromatic degradation pathways.

A T. aromatica orf20 mutant was unable to grow on 3,5-DHB (Fig. 3A). This gene was homologous to A. anaerobius orf12 in the resorcinol degradation cluster, which was not essential for growth with resorcinol (7) (see Table S1 in the supplemental material). The product of orf20 showed a strong similarity with succinate semialdehyde dehydrogenases, enzymes generally involved in the oxidation of aliphatic and aromatic aldehydes. The presence of the homologous orf12 in A. anaerobius suggests that this enzyme could be involved in the last steps of the degradation from HBQ to acetate and malate, as suggested for this strain (7). The orf20 mutant could be complemented by pJB-orf20, harboring the wild-type gene under the control of P_lac (Fig. 3B).

The product of orf21, which is homologous to orf10 of the A. anaerobius resorcinol degradation cluster (see Table S1 in the supplemental material), belonged to the COG0523 family of “P-loop” GTPases, putative insertases/metallochaperones involved in metallocenter biosynthesis (25), although they were both predicted to function only as insertases, suggesting that they required an accessory metallochaperone. Frequently, insertases are located near metallopeptidases and proteins involved in molybdopterin cofactor recycling. A gene coding for a peptidase was found in both the 3,5-DHB degradation (orf12) (Table 3) and A. anaerobius resorcinol degradation clusters (7). It is thus tempting to suggest that the products of orf21 and orf12 may participate coordinately in the maturation of DbhL. Although A. anaerobius orf10 was not essential for growth on resorcinol (7), the growth of a T. aromatica orf21 mutant was significantly delayed (Fig. 3A), although a polar effect on the downstream essential gene bqdL could not be ruled out.

Genes unique to the T. aromatica strain AR-1 pathway.

The remaining genes in the cluster had no counterpart in the A. anaerobius resorcinol degradation cluster, but some of them could be annotated with a high degree of confidence. A set of tripartite ATP-independent periplasmic (TRAP) transporter components (orf13, orf14, and orf15) and an additional copy of the extracytoplasmic solute receptor (ESR) part of the transporter (orf16) could be identified in the middle of the cluster. In addition, a pseudogene (orf4) with no identified stop codon between orf3 and orf5 showed a high degree of similarity with the TRAP transporter ESR. In the products of orf14 and orf15, 4 and 12 transmembrane helices could be predicted, respectively, as expected for integral membrane proteins. TRAP transporter systems have been related to the uptake of carboxylic acids (26), consistent with a role of this set of proteins in 3,5-DHB transport. A mutant with a knockout mutation in orf13, encoding the periplasmic component of the TRAP transporter, showed slower growth with 3,5-DHB than the wild type, although it finally reached similar cell densities (see Fig. S3 in the supplemental material). This suggested that this transport system was required for efficient 3,5-DHB uptake into the cell. Furthermore, transcription analysis showed that orf13, orf14, and orf15 constituted an operon that was inducible by 3,5-DHB (see below), supporting the role of this transporter in 3,5-DHB utilization. The three genes were annotated as dbtP, dbtQ, and dbtM, respectively. The 3,5-DHB concentrations found in nature are expected to be much lower than those used in this analysis; thus, utilization of this substrate would probably require the contribution of a dedicated transport system.

Two genes coding for proteins with 100% and 97% identity to the α and β subunits of 2-oxoglutarate-ferredoxin oxidoreductase of T. aromatica, which is involved in the regeneration of reduced ferredoxin for benzoyl-CoA reductase (27) (orf5 and orf6, annotated as korAB), were located upstream from dbhL. The two genes were expressed as one operon and were not induced in the presence of 3,5-DHB (see below). Furthermore, a double mutant in which a region encompassing parts of both genes had been replaced with a gentamicin cassette was unaltered in its capacity to grow with 3,5-DHB as a carbon source, indicating that this electron recycling system was not essential in the pathway (Fig. 3A).

An orf17 insertion mutant was able to grow with 3,5-DHB similarly to the wild type (Fig. 3A). No specific function could be attributed to this gene product, which showed a 65% average similarity with C-C bond hydrolases. The gene product of orf18 showed homology with molybdopterin oxidoreductases, and in fact it was 39% identical to the dbhL (orf7) product, although it was 286 residues shorter and was predicted to include a [Fe₄-S₄] iron-sulfur center not found in DbhL. An orf18 mutant was able to grow on 3,5-DHB, although at a lower rate than the wild type (Fig. 3A). However, a polar effect on the downstream bqdM gene could not be ruled out.

Finally, two genes were identified downstream of qorA: the product of orf27, which was transcribed convergently to qorA, showed 79% identity to a Thauera MZ1T thioredoxin-like gene possibly involved in mandelate metabolism, and the 5′ end of orf28 allowed the prediction of a protein sequence with 81% identity to a Thauera MZ1T gene annotated as mandelate dehydrogenase. An orf28 knockout mutant was able to grow on 3,5-DHB like the wild type, suggesting that these two genes with high similarity to the Thauera MZ1T chromosome were unrelated to the 3,5-DHB degradation pathway and constituted the boundary of the cluster in the T. aromatica strain AR-1 chromosome (Fig. 3A).

Transcriptional organization and induction of the 3,5-DHB degradation gene cluster.

The gene direction and the intergenic distances in the cluster suggested a gene arrangement of at least three operons. To determine the organization of the transcriptional units, we analyzed the cotranscription of each pair of neighbor genes using RT-PCR. It was previously shown that 3,5-DHB degradation enzymes were detected only if the cells were grown with 3,5-DHB as a carbon source, suggesting a substrate-controlled expression of the pathway genes (10). We therefore isolated total RNA from T. aromatica strain AR-1 cells growing anaerobically either on succinate or on 3,5-DHB as a carbon source, and we performed RT-PCRs with a series of oligonucleotide pairs covering the different intergenic regions (Fig. 2B). The results defined three transcriptional units that were induced in the presence of the substrate: operon I spanned from dbhL to orf12, operon II included the three TRAP transporter genes, and operon IV spanned from orf20 to bqdS (Fig. 2). The results also indicated low basal expression levels of a possible internal segment of operon I. In addition, operon III, which covered orf18 and bqdM, and operon V, which included orf25 and qorA, showed basal expression levels during growth on succinate, which seemed to increase in the presence of 3,5-DHB. Semiquantitative RT-PCR assays using increasing RNA concentrations to compare expression levels under induced and uninduced conditions showed amplification of the operon III region at lower RNA concentrations when RNA originated from cells grown in the presence of 3,5-DHB, thus confirming induction of the operon by this substrate (not shown). Finally, the korAB genes constituted an independent operon that was expressed independently of the presence of the aromatic substrate, supporting the suggestion that they did not participate in 3,5-DHB degradation. From these results, we could suggest the presence of promoters initiating transcription from dbhL, dbtP, orf18, orf20, and dbdR (Fig. 2B).

Expression of the pathway is controlled by a LysR-type transcriptional regulator (LTTR).

Two genes coding for regulators belonging to the LysR family (orf3 and orf25) were found at both ends of the cluster. Mutations in the two genes were obtained through reverse genetics. Figure 3A shows that an orf3 mutant could grow on 3,5-DHB like the wild type, indicating that the regulator encoded by orf3 was not an inducer of this pathway. In contrast, an orf25 mutant was unable to grow on 3,5-DHB, suggesting that the product of this gene was an essential regulator of the pathway (Fig. 3A). The gene was annotated as dbdR. To confirm the role of dbdR in the pathway regulation, we determined the expression of the main pathway operons in a dbdR mutant background. To this end, we isolated total RNA from the wild-type and dbdR mutant strains grown on glutarate plus 3,5-DHB. Transcription of operons I, III, and V was determined through RT-PCR of the intergenic regions dbhL-dbhS (operon I), dbtQ-dbtM (operon II), and orf21-bqdL (operon IV). Figure 6 shows that in the wild-type strain, expression of the operons was induced in the presence of 3,5-DHB, while no amplification product was obtained in the dbdR mutant strain grown on succinate plus 3,5-DHB. In contrast, expression of the unrelated korAB operon was unaffected by the dbdR mutation. The expression pattern in an orf3 mutant background was similar to that in the wild-type in both the presence and absence of 3,5-DHB, ruling out a possible role of the orf3 product as a repressor of the pathway.

FIG 6.

DbdR controls expression of the T. aromatica strain AR-1 main 3,5-DBH degradation operons. Products of RT-PCR of total RNA from wild-type (wt) and dbdR mutant strains of T. aromatica strain AR-1 cells grown on 2 mM glutarate plus 2 mM 3,5-DHB (+) are shown. For each operon, a negative control (c) with RNA where the reverse transcription step had been omitted and a positive control with DNA (D) from T. aromatica strain AR-1 were included.

DISCUSSION

A. anaerobius and T. aromatica are two betaproteobacteria that share a unique anaerobic degradation pathway for dihydroxylated aromatics that does not proceed through the canonical benzoyl-CoA intermediate; rather, degradation is initiated through oxidation of the aromatic ring, a reaction that is ultimately linked to nitrate respiration. We have shown that the enzymatic similarities between the two strains are reproduced in their genetic complements. Not only were the genes for the three main steps of the pathway conserved, but also homologous auxiliary genes were present in the two clusters (see Table S1 in the supplemental material). A protein pair putatively involved in metalloprotein maturation, conserved in both pathways, could be tentatively annotated: the product of orf12, carried in the same operon as dbhS, showed 93% identity to the product of the A. anaerobius essential gene orf13. These proteins were 68% identical to an M24B family peptidase suggested to be involved in protein maturation. The MEROPS database predicted that the product of orf12 lacked the critical residues required for peptidase activity (21), suggesting that this gene would code for a dedicated chaperone for 3,5-DHB hydroxylase maturation. This was further supported by the presence in the cluster of orf21, which coded for a protein showing all sequence features of insertases involved in metallocenter synthesis and requiring a complementary chaperone (see Fig. S1b in the supplemental material). The presence of the corresponding homologous genes orf13 and orf10 in A. anaerobius indicates similar processing mechanisms of the enzymatic machinery in both bacteria. Interestingly, two small genes coding for proteins predicted to have a cupin-like structure (orf9 and orf10) were identified in the T. aromatica cluster. Cupins are metal-binding proteins involved in a broad diversity of cellular functions (28, 29). We could detect in A. anaerobius a previously unidentified ORF between btdS and orf13, which coded for a small protein having a cupin-like structure. The β-barrel structural scaffold defining the cupin fold was clearly predictable for the three gene products (see Fig. S4 in the supplemental material), although like other members of the cupin superfamily, they did not show homology in their amino acid sequences. The presence of these metal-binding protein genes in the clusters was consistent with the presence of several metalloproteins in the two pathways and supports they are involved in similar auxiliary mechanisms. The specific role of each cupin remains to be elucidated.

Generally, folding of proteins such as DbhL and RehL that bind molybdopterin as cofactors requires dedicated chaperones, which recognize a consensus sequence at the N-terminal ends of these enzymes. In some cases this motif is composed of a pair of arginine residues close to the N-terminal end of the protein (30, 31). Interestingly, two consecutive arginine residues were present in the N-terminal end of DbhL, which were conserved in A. anaerobius RehL and could represent a chaperone recognition motif (see Fig. S2a in the supplemental material). Moreover, the product of orf23 showed 73% identity to the A. anaerobius orf5 gene product. No clear function had been assigned to these gene products, which belong to the “vicinal oxygen chelate” (VOC) metalloenzyme superfamily (32). Members of this family share a mechanistic feature which involves coordination by a substrate (or an intermediate) through adjacent oxygen atoms to a divalent metal center for activation or stabilization (33). Since several intermediates in both the 3,5-DHB and resorcinol degradation pathways harbor vicinal oxygens (hydroxyhydroquinone and hydroxybenzoquinone), we suggest that the primary function of the orf23 gene product would be to operate in conjunction with some enzyme in the pathway to activate a substrate or stabilize a reaction intermediate. The conserved secondary structure and relevant residues involved in metal binding are present in the sequences of both gene products, and structure modeling of the T. aromatica strain AR-1 orf23 gene product predicted binding of one Ni or Zn atom per molecule.

No transmembrane fragment or anchoring signal could be identified in the sequences of DbhL, BtdL, and QorA. The same was true for the homologous proteins RehL and BtdL in A. anaerobius (7). This was unexpected, since both the hydroxylating activity and the hydroxyhydroquinone dehydrogenase activity had been found almost exclusively in the membrane fractions of T. aromatica and A. anaerobius crude extracts (6, 9). This suggests protein-membrane association through hydrophobic/electrostatic interactions.

Despite sequence homology, the overall gene organization differed considerably between the two strains (Fig. 7). In A. anaerobius the genes appeared to be well organized in transcriptional units according to their function. In contrast, the genes in T. aromatica strain AR-1 were interspersed in the cluster and were transcribed in five different operons. The whole degradation cluster was flanked by genes showing high similarity to the Thauera MZ1T chromosome sequence, while the genes present in the cluster had no homologs in the different Thauera chromosomes sequenced to date. Although the GC content of the cluster (68.1%) and those of other Thauera strains are similar (68% for Thauera sp. MZ1T), horizontal gene transfer from a closely related organism cannot be ruled out, since, for example, some Azoarcus strains also have a similar high GC content, ranging between 65.1 and 67.92%. The presence of two alternative pathways for the metabolism of hydroxyhydroquinone reinforces this idea. Soluble quinone oxidoreductases are enzymes generally involved in protection against oxidative stress through reduction of oxidized quinones (34). T. aromatica strain AR-1 seems to have recruited this type of enzyme to carry out the reverse reaction in 3,5-DHB degradation. Two highly homologous regulatory genes encoding NtrC family proteins work coordinately to regulate A. anaerobius pathway expression (D. Pacheco, unpublished data), while a single LysR family regulator gene was responsible for the pathway expression in the T. aromatica strain AR-1 pathway. Interestingly, the outer membrane component of an ABC-type transporter present in the A. anaerobius cluster was shown to be dispensable for resorcinol utilization in the heterologous host where the mutant was tested, while the TRAP transporter present in T. aromatica was required for efficient growth on 3,5-DHB. This is consistent with the difference in polarity between the two substrates. The carboxylic group of 3,5-DHB would be deprotonated at neutral pH and would require a dedicated transport system to enter the cell.

FIG 7.

Sequence organization and conservation of the T. aromatica strain AR-1 and A. anaerobius dihydroxyaromatic anaerobic degradation clusters. Genes homologous between pathways are connected with a colored band. Regulatory genes are drawn in red. Genes of unknown function present in both strains are shaded in gray. The genes filled in black (A. anaerobius orf12b and T. aromatica strain AR-1 orf9 and orf10) do not show DNA sequence homology, but their products are predicted to share a similar cupin-like structure.

Analysis of the T. aromatica 3,5-DHB degradation cluster led us to propose a degradation pathway that can be summarized as follows (see Fig. S5 in the supplemental material). The substrate 3,5-DHB would enter the cell through the TRAP transport system encoded by dbtPQM. The ESR protein encoded by orf16 could be an alternative receptor protein for the substrate. Maturation of the membrane-associated 3,5-DHB hydroxylase would probably require the coordinated action of the chaperone-peptidase pair encoded by orf21 and orf12. Oxidation of hydroxyhydroquinone would be carried out by either BtdL or QorA to produce hydroxybenzoquinone. Ring cleavage of this compound would involve the three-component enzyme BqdLSM, and the product of orf20 would carry out further metabolism of the resulting linear compound.

Our finding of conserved genetic determinants for dihydroxylated aromatic oxidation in these closely related strains suggests that this degradation strategy was selected in this bacterial group for the efficient degradation of a particular set of aromatic compounds. These pathways with hydroxyhydroquinone as the central intermediate have so far been found only in nitrate-reducing bacteria, which could be explained in terms of the energetic requirements of the reactions involved (5). The oxidation reactions involved in the dearomatization of the two substrates would require electron acceptors of a positive redox potential (the standard redox potentials of the hydroxylating activity are −27 mV for 3,5-DHB, −33 mV for resorcinol, and +180 mV for hydroxyhydroquinone dehydrogenation) (6, 9). This does not preclude the presence of additional anaerobic pathways for aromatic degradation channeled through benzoyl-CoA. In fact, a cluster of genes homologous to the bzd cluster for benzoate degradation has been found in A. anaerobius (J. I. Medina-Bellver, unpublished data), and benzoyl-CoA reductase has been detected by immunoblotting in T. aromatica strain AR-1 cells growing on benzoate (10).