The Homogentisate Pathway: a Central Catabolic Pathway Involved in the Degradation of l-Phenylalanine, l-Tyrosine, and 3-Hydroxyphenylacetate in Pseudomonas putida

Abstract

Pseudomonas putida metabolizes Phe and Tyr through a peripheral pathway involving hydroxylation of Phe to Tyr (PhhAB), conversion of Tyr into 4-hydroxyphenylpyruvate (TyrB), and formation of homogentisate (Hpd) as the central intermediate. Homogentisate is then catabolized by a central catabolic pathway that involves three enzymes, homogentisate dioxygenase (HmgA), fumarylacetoacetate hydrolase (HmgB), and maleylacetoacetate isomerase (HmgC), finally yielding fumarate and acetoacetate. Whereas the phh, tyr, and hpd genes are not linked in the P. putida genome, the hmgABC genes appear to form a single transcriptional unit. Gel retardation assays and lacZ translational fusion experiments have shown that hmgR encodes a specific repressor that controls the inducible expression of the divergently transcribed hmgABC catabolic genes, and homogentisate is the inducer molecule. Footprinting analysis revealed that HmgR protects a region in the Phmg promoter that spans a 17-bp palindromic motif and an external direct repetition from position −16 to position 29 with respect to the transcription start site. The HmgR protein is thus the first IclR-type regulator that acts as a repressor of an aromatic catabolic pathway. We engineered a broad-host-range mobilizable catabolic cassette harboring the hmgABC, hpd, and tyrB genes that allows heterologous bacteria to use Tyr as a unique carbon and energy source. Remarkably, we show here that the catabolism of 3-hydroxyphenylacetate in P. putida U funnels also into the homogentisate central pathway, revealing that the hmg cluster is a key catabolic trait for biodegradation of a small number of aromatic compounds.

Eukaryotic organisms catabolize Phe and Tyr by a common peripheral pathway which leads to homogentisate (2,5-OH-PhAc) as a central intermediate (6, 9, 20, 23, 31). The genetic and biochemical interest in this pathway comes from the fact that many severe human diseases (e.g., phenylketonuria, alcaptonuria, tyrosinemia, tyrosinosis, Richner-Hanhart syndrome, and hawkinsinuria) are associated with enzyme deficiencies in the catabolism of Phe and Tyr (16, 19, 24, 32, 52, 73). First, Phe is transformed into Tyr by a pterin-dependent phenylalanine hydroxylase (PhhA), and later, a tyrosine aminotransferase (TyrB) catalyzes the conversion of Tyr into 4-hydroxyphenylpyruvate (4-OH-PhPyr), which is further transformed into 2,5-OH-PhAc by a 4-OH-PhPyr dioxygenase (Hpd) (Fig. 1B). The homogentisate central pathway involves a homogentisate dioxygenase (HmgA) that opens the aromatic ring of 2,5-OH-PhAc, producing maleylacetoacetate, which is isomerized to fumarylacetoacetate by the HmgC isomerase. Finally, fumarylacetoacetate is hydrolyzed by a specific hydrolase (HmgB) to form fumarate and acetoacetate, which are two compounds of the central metabolism (Fig. 1B). In plants and photosynthetic bacteria the catabolism of Tyr is also crucial because homogentisate is a precursor for the biosynthesis of photosynthetic pigments (66).

FIG. 1.

Pathway for the catabolism of Phe and Tyr. (A) Arrangement of the genes involved in catabolism of Phe and Tyr in P. putida U. The gene clusters encoding the peripheral and central (homogentisate) pathways are indicated. Discontinuous lines between genes indicate unknown distances. The relative positions of the gene clusters in the genome of P. putida U are still unknown. (B) Biochemistry of the Phe/Tyr catabolism. The intermediates of the catabolic pathway are indicated. The homogentisate central pathway is enclosed in a box. The enzymes are PhhA (phenylalanine hydroxylase), PhhB (carbinolamine dehydratase), TyrB (tyrosine aminotransferase), Hpd (4-OH-PhPy dioxygenase), HmgA (homogentisate dioxygenase), HmgB (fumarylacetoacetate hydrolase), HmgC (maleylacetoacetate isomerase), Mha (3-hydroxyphenylacetate monooxygenase), and dihydropteridine reductase (DHPR). In previous work, TyrB, HmgB, and HmgC were called PhhC, Fah and Mai, respectively.

Although the catabolism of Phe and Tyr in eukaryotic organisms has been well established, limited information has been obtained about the degradation of these amino acids in prokaryotes (58, 68, 81). The inability of Escherichia coli, the model prokaryotic organism, to mineralize Phe and Tyr might have contributed to the reduction in interest in this pathway in bacterial systems. However, some studies have shown that the catabolism of Phe and Tyr in bacteria is also carried out by a peripheral pathway similar to that of eukaryotes, with formation of homogentisate as a central intermediate (1, 36, 43, 44, 53, 61, 66, 70). Nevertheless, the genes encoding the catabolic enzymes of the homogentisate central pathway and the regulatory elements that control the expression of such genes have only been partially identified and characterized (27, 43, 44, 69, 76).

This work was aimed at identifying and characterizing for the first time the complete set of genes responsible for degradation of Phe and Tyr in bacteria. To this end, we studied the catabolism of Phe and Tyr in Pseudomonas putida, a γ-proteobacterium with great metabolic versatility that can use these amino acids as carbon and energy sources and that is considered a model system for environmental studies (35, 49, 50, 75). By using two strains of P. putida that show different catabolic abilities with some natural aromatic compounds, P. putida U, a strain that is able to grow with 4-hydroxyphenylacetate (4-OH-PhAc) or 3-hydroxyphenylacetate (3-OH-PhAc) as the sole carbon source (48), and P. putida KT2440, a strain whose complete genome is known (47) and that is unable to use 4-OH-PhAc and 3-OH-PhAc (35), we demonstrated that 3-OH-PhAc degradation funnels into the homogentisate central pathway. A transcriptional analysis of the homogentisate cluster revealed the existence of a protein, HmgR, that is the first IclR-type regulator that acts as a repressor of an aromatic catabolic pathway.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

The bacterial strains and plasmids used in this study are listed in Table 1. Unless otherwise stated, bacteria were grown in Luria-Bertani (LB) medium (60) at 37°C (E. coli) or 30°C (P. putida). Growth in minimal medium (MM) (39) was achieved by using the corresponding necessary nutritional supplements. When required, 5 mM citrate, phenylacetate (PhAc), Tyr, Phe, 3-OH-PhAc, or 4-OH-PhAc was added to MM. Recombinant E. coli DH5α cells were cultured in MM with glycerol as the carbon source supplemented with 0.05% Casamino Acids. The appropriate selection markers, including kanamycin (25 to 50 μg/ml), tetracycline (40 μg/ml), ampicillin (100 μg/ml), chloramphenicol (90 μg/ml), gentamicin (30 μg/ml), and rifampin (20 to 50 μg/ml), were added when needed.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Genotype and/or description	Reference or source
E. coli K-12 strains
DH5α	F−ΔlacU169 φ80 dlacZΔM15 hsdR17 recA1 endA1 gyrA96 thi-1 relA1 supE44	77
JM109	F′ traD36 proA+proB+lacIqlacZΔMI5/recA1 endA1 gyrA96 (Nalr) thi hsdR17 supE44 relA1 Δ(lac-proAB)	79
HB101	F− Δ(gpt-proA) 62 leuB6 supE44 ara-14 galK2 lacY1 Δ(mcrC-mrr) rpsL20 (Smr) xyl-5 mtl-1 recA13	3
CC118 λpir	Δ(ara-leu) araD ΔlacX74 galE galK phoA20 thi-I rps-1 rpoB argE(Amp) recA thi pro hsdRM+ RP4-2-Tc	34
S17-1 λpir	λpir recA thi pro hsdR M+, RP4:2-Tc::Mu::Km Tn7Tpr Smr	67
E. coli W strains
AF141	E. coli WΔpaa lacZ Rif+ derivative	25
BMR	AF141 with chromosomal insertion of mini-Tn5Km2 Phmg-lacZ	This study
P. putida strains
U	Wild type	8
U-7	P. putida U hmgA due to a Tn5 insertion	This study
U-95	P. putida U hmgA due to a Tn5 insertion	This study
U-SG6	P. putida U hmgA due to a Tn5 insertion	This study
U-111	P. putida U hpd due to a Tn5 insertion	This study
U-215	P. putida U phhA due to a Tn5 insertion	This study
U-A2	P. putida U hpaB due to a Tn5 insertion	Unpublished data
U-dhmgA	P. putida U hmgA::pK18mob	This study
U-dhmgB	P. putida U hmgB::pK18mob	This study
U-dhmgC	P. putida U hmgC::pK18mob	This study
U-dhpd	P. putida U hpd::pK18mob	This study
U-Δhmg	P. putida U with hmgABC operon deleted from the genome	This study
U-Δhpd	P. putida U lacking the hpd gene	This study
U-dphhA	P. putida U phhA::pK18mob	This study
U-A2dhpd	P. putida U hpaB due to a Tn5 insertion and disrupted in the hpd gene with the pJQhpd construction	This study
KT2440	hsdMR	2
KT2440Rif+	Spontaneous KT2440 rifampin-resistant mutant	This study
KT2440-dhmgA	KT2440Rif+hmgA::pK18mob	This study
KT2440-dhpd	KT2440Rif+hpd::pK18mob	This study
KT2440-dphhA	KT2440Rif+phhA::pK18mob	This study
KT2440-dhmgR	KT2440Rif+hmgR::pK18mob	This study
Plasmidsa
pUC18	AproriColE1 lacZα+lac promoter	79
pGEM-T easy	AproriColE1 lacZα+ SP6 T7 lac promoters, direct cloning of PCR products	Promega
pK18mob	KmroriColE1 Mob+lacZα+, used for directed insertional disruption	62
pRK600	CmroriColE1 oriV Mob+, helper plasmid in triparental matings	37
pBBR1MCS-2	KmroripBBR1 Mob+lac promoter lacZα+, broad-host-range cloning and expression vector	38
pBBR1MCS-3	TcroripBBR1 Mob+lac promoter lacZα+, broad-host-range cloning and expression vector	38
pJQ200KS	Gmrorip15A Mob+lacZα+sacB, used to generate deletions by double-recombination events	56
pQE32	AproriColE1 T5 promoterllac operator, lambda t0/E. coli rrnB T1 terminators, N-terminal 6-His	QIAGEN
pSJ3	AproriColE1r, lacZ promoter-probe vector	25
pUTminiTn5Km2	Apr KmroriR6K Mob+, mini-Tn5Km2 transposon delivery plasmid	11
pUT-Phmg-lacZ	pUTminiTn5Km2 derivative carrying the Phmg-lacZ translational fusion	This study
pU-HH	pUC18 derivative expressing the hmgABC and hpd genes	This study
pU-HHP	pUC18 derivative containing the Tyr catabolic cassette (hmgABC hpd tyrB1)	This study
pM2-HHP	pBBR1MCS-2 derivative harboring the SacI/HindIII Tyr catabolic cassette from pU-HHP	This study
pGS9	Conjugative plasmid containing the Tn5 transposon	65

^aFor plasmids obtained by cloning of a PCR-generated fragment see Table 2.

DNA manipulations.

DNA manipulations and other molecular biology techniques were performed essentially as described previously (60). Transformation of E. coli cells was carried out by using the RbCl method or by electroporation (Gene Pulser; Bio-Rad) (14). Oligonucleotides were synthesized with an Oligo-1000 M nucleotide synthesizer (Beckman Instruments). Nucleotide sequencing was performed with an ABI Prism 3700 DNA sequencer (Applied Biosystems Inc.). DNA fragments were purified by standard procedures by using Gene Clean (Bio 101, Inc.). The method used for preparation of genomic DNA has been described elsewhere (60). The primers used for PCR amplification are summarized in Table 2.

TABLE 2.

Specific primers designed for PCR amplification and driven cloning strategy

Primer	Sequence (5′-3′)a	DNAb template	Cloning strategy	Recombinant plasmid
O-hmgA5 O-hmgA3	GGGGAGCTCGCTTTGGCCATGGGAGGGGAGGTT CAAGGTACCGGCGGCGAACCGTTCACTTC	KT2440	The 3,396-bp PhmgABC fragment that was PCR generated was double digested with SacI/KpnI and cloned into pUC18	pU-HMG
O-hpd5	TCCGGATCCCACGCAATCAATAATTTTCGC	KT2440	The 1,134-bp hpd gene that was PCR generated was double digested with KpnI/BamHI and cloned into pUC18	pU-hpd
O-hpd3	ACGGTACTCTAGAGGCGATCAGTCTGTGC		The 819-bp hpd internal fragment that was PCR generated was double digested with EcoRI/SalI and cloned into pK18mob	pK-dhpd
O-tyrB5 O-tyrB3	GGCTCTAGACTGTACGCCCTATCACACTC CCCTCTAGAAGTCAACCCACCCCCCAG	KT2440	The 1,278-bp tyrB1 gene that was PCR generated was digested with XbaI and cloned into pUC18	pU-tyrBI
O-hmgR5E O-hmgR3	GATGAATTCGGAACCTCCCCTC TATGTCGACTTCTCTCAGGAACTGC	KT2440	The 866-bp hmgR gene that was PCR generated was double digested with EcoRI/SalI and cloned into pQE32	pQ-hmgR
O-dphhA5 O-dphhA3	AGAGCATGCGGTGTGGAACACCC CCTGGATCCGGTCGAAGGCCTGG	KT2440	Cloning into pK18mob of the PCR-generated fragment digested with SphI/BamHI that contained a 539-bp internal sequence of phhA	pK-dphhA
O-dhmgA5 O-dhmgA3	CTGTCGACGAGCCCGCAGCCGATTCCT ACGGTACCTGTTCTCGGCCACCA	KT2440	The 677-bp hmgA internal fragment that was PCR-generated was double digested with SalI/KpnI and cloned into pK18mob	pK-dhmgA
O-dhmgR5 O-dhmgR3	GGCGTCGACGCATCCACCCCCGGCATCAGC GGGGTACCCGAACAGGATGCGGCCACCACC	KT2440	The 428-bp hmgR internal fragment that was PCR generated was double digested with SalI/KpnI and cloned into pK18mob	pK-dhmgR
O-Phmg5 O-Phmg3	GTGGTACCGACGTCGAGGGTGAGAGT GGGGATCCGGGGCCTTCTGCGGGGAGTTCTGC	KT2440	The 335-bp hmgR-hmgA intergenic region that was PCR generated was double digested with KpnI/BamHI and cloned into pSJ3	pSJ-Phmg-lacZ
HGA(5)2 HGA(3)8	CCGGCCGGCGTGAGCATCTACATCTACTG CTCGGCCACCATCCAGCGCGG	U	The 592-bp hmgA internal fragment that was PCR generated was cloned into pGEM-T easy and then digested with EcoRI and subcloned into pK18mob	pKhmgA
HGA(5)4 HGA(3)3	CCGCCCCGACAACCCGCTGCTAC GCGGCTGCGGGTCACCTTCGGG	U	The 433-bp hmgB internal fragment that was PCR generated was cloned into pGEM-T easy and then digested with EcoRI and subcloned into pK18mob	pKhmgB
HGA(5)6 HGA(3)5	GGCCTTGCGCACCGACGGTGG CCGATCCACTGGTTGACCTGCCCCTC	U	The 233-bp hmgC internal fragment that was PCR generated was cloned into pGEM-T easy and then digested with EcoRI and subcloned into pK18mob	pKhmgC
HPD(5)2	GCCGATGGAGCTGCGCCTGCCG	U	The 378-bp hpd internal fragment that was PCR generated was cloned into pGEM-T easy and then digested with EcoRI and subcloned into pK18mob	pKhpd
HPD(3)2	GAACTGCATCAGGAACTCTTCAATCTGGCC		The 378-bp hpd internal fragment that was PCR generated was cloned into pGEM-T easy and then digested with EcoRI and subcloned into pJQ200KS	pJQhpd
Phh(3)4 Phh(5)4	CTGGTGCTCAGGCTCGTCAGACAGACTGTAGAC CAACTGGGCGAGATCAACAAGGTGCTGGG	U	The 419-bp phhA internal fragment that was PCR generated was cloned into pGEM-T easy and then digested with EcoRI and subcloned into pK18mob	pKphhA
Hpd(5) Hpd(3)4	AATCAATAATTTTCGCTGCAGATGAG AGATCGTCCCCTCGCTGATGCTGGTGG	U	Cloning into pGEM-T easy of the PCR-generated fragment (1,245 bp) that contained the Hpd-coding region behind the lac promoter	pGhpd
HpdBam HpdPstB HpdPstX HpdXba	TCTTCGAGGATCCAATGGGCC TTTCAACCTGCAGGGCGCCCA GAGCGGCTGCAGGGTCACGG GGCGGCGCTATCTAGAGGGTATCAGTCGGTGC	U	The PCR-amplified 5′ end (primers HpdBam and HpdPstB) and 3′ end (primers HpdPstX and HpdXba) of the hpd gene were cloned in pJQ200KS as BamHI/PstI and PstI/XbaI fragments, respectively	pJQhpdBX
HmgBam HmgPstB HmgPstX HmgXba	CTACCTCAGCGGATCCGGCAAC GAAGAACACTCGCTGCAGGGAACGG TTTATCCACAGCCTGCAGTGC GGATGCGCGGGAAGCTCTAGAGG	U	The PCR-amplified 5′ end of hmgA (primers HmgBam and HmgPstB) and 3′ end of hmgC (primers HmgPstX and HmgXba) were digested with BamHI/PstI and PstI/XbaI respectively, and cloned into pJQ200KS to delete the hmg operon in P. putida U	pJQhmgBX
HmgA(5) HGA(3)6	GCCAGCAACTAGTCAGTCAGAGCCCGGAGG GGCAAGCTTCCGGCGGCGAACC	U	The PCR-amplified fragment (3,306 bp) containing the hmgABC genes was cloned into pBBR1MCS-3	pMChmg

^aEngineered restriction sites are underlined.

^bEach PCR was carried out with the two primers indicated and the genomic DNA of P. putida KT2440 or P. putida U as the template.

Generation of P. putida U mutants.

Transposon Tn5 was transferred from E. coli HB101(pGS9) (65) to P. putida U by filter mating (34). Transconjugants were selected on LB medium plates containing rifampin (which selected for the Pseudomonas recipient cells) and kanamycin (which selected for the transposon marker) after incubation at 30°C for 36 to 48 h. Colonies were replica plated in two different media, MM containing fructose (0.5 g liter⁻¹) (26), rifampin, and kanamycin (MMA medium) and MM supplemented with Tyr (10 mM), rifampin, and kanamycin (MMB medium). Selected mutants (Tyr⁻) were able to grow in MMA medium but not in MMB medium.

For gene disruption through single homologous recombination, an internal fragment (usually 300 to 800 bp) of the gene to be disrupted was cloned in the polylinker of pK18mob (a mobilizable plasmid which does not replicate in Pseudomonas) (62), and the resulting construct (Table 2) was introduced into P. putida U by triparental filter mating (34). Exconjugants harboring the disrupted gene were isolated on LB medium containing rifampin and kanamycin after 2 days of incubation at 30°C.

Deletion of the hmg and hpd genes in P. putida UΔhmg and P. putida UΔhpd (Table 1) was accomplished by using plasmids pJQhmgBX and pJQhpdBX (Table 2), respectively, through a double-recombination event selected by expression of a lethal sacB gene (13, 56).

All mutants were analyzed by PCR as previously described (46, 49, 59) to define the insertion position of the disrupting element (Tn5 or pK18mob derivative) or to confirm the extent of the deletion.

Construction of a DNA cassette for the catabolism of Tyr.

For construction of a DNA cassette containing the genes responsible for catabolism of Tyr in P. putida, the hmgABC, hpd, and tyrB1 genes were PCR isolated and cloned into plasmid pUC18 to produce plasmids pU-HMG, pU-hpd, and pU-tyrB1, respectively (Table 2). The hmgABC and hpd genes were combined as a SacI-KpnI cassette, producing plasmid pU-HH (Table 1). The tyrB1 gene was subcloned as an XbaI fragment in plasmid pU-HH, giving rise to plasmid pU-HHP (Table 1), a pUC18 derivative that contained the 5.8-kb SacI-HindIII hmg-hpd-tyrB1 DNA cassette (Tyr cassette) expressed under control of the tandem Plac and Phmg promoters. The SacI-HindIII Tyr cassette was then subcloned into a mobilizable broad-host-range vector, pBBR1MCS-2, producing the recombinant pM2-HHP plasmid (Table 1).

Data analysis.

The nucleotide sequence of the P. putida KT2440 genome (accession number AE015451) was analyzed at http://www.tigr.org. Deduced amino acid sequences were analyzed with the Protein Analysis Tool at the World Wide Web Molecular Biology server of the Geneva University Hospital and the University of Geneva. Protein sequence similarity searches were done with the BLAST program by using the National Center for Biotechnology Information server.

Enzyme assays.

β-Galactosidase activities were measured with permeabilized cells as described by Miller (45). The 2,5-OH-PhAc dioxygenase activity was spectrophotometrically determined by measuring the formation of maleylacetoacetate at 330 nm as described elsewhere (15). Exponentially growing cultures were centrifuged (7,000 × g, 5 min, 4°C), and cells were resuspended and concentrated 100-fold in 100 mM potassium phosphate buffer (pH 7.0) containing 20% glycerol. Cell lysis was performed in the same buffer by sonication. Cell extracts were clarified by centrifugation (10,000 × g, 15 min, 4°C) and used in the enzyme assays. The enzyme assay mixtures (final volume, 0.5 ml) contained 100 mM potassium phosphate buffer (pH 7.0), 2 mM ascorbate, 50 μM FeSO₄, 300 μM 2,5-OH-PhAc (unless indicated otherwise), and 5 μl of extract (50 to 100 μg of protein). The reactions were carried out at 37°C with a Beckman DU 520 spectrophotometer. One milliunit corresponded to transformation of 1 nmol of 2,5-OH-PhAc to maleylacetoacetate per min at 37°C under the conditions described above. The molar extinction coefficient of maleylacetoacetate is 13,500 M⁻¹ cm⁻¹ (64).

Isolation of products that accumulated in the culture broth.

The extracellular products accumulated by the wild type or by the mutant strains affected in the homogentisate pathway were identified by high-performance liquid chromatography (HPLC), nuclear magnetic resonance (NMR), and mass spectrometric analyses (7, 10, 28).

P. putida U and the mutant strains were cultured in MM containing Phe, Tyr, or 3-OH-PhAc (5 mM) as a source of intermediates and 4-OH-PhAc (5 mM) for support of bacterial growth. Moreover, 4-OH-PhAc, which requires a specific catabolic route, is not degraded through the homogentisate pathway (48), and it is not a substrate of 4-OH-PhPyr dioxygenase. When required, the cultures were centrifuged (5,000 × g) and filtered through Millipore filters (pore size, 0.45 μm) to eliminate bacteria. The supernatant (culture broth) was acidified with 6 M HCl to pH 1.35 and extracted with n-butanol. The organic phase was washed twice with Milli-Q water, dried with anhydrous Na₂SO₄, and lyophilized.

HPLC analyses.

Cultures were grown in 500-ml Erlenmeyer flasks containing 100 ml of medium and incubated in a rotary shaker (250 rpm) at 30°C. Samples were taken at different times, centrifuged (16,000 × g, 20 min) to eliminate bacteria, and filtered through a Millipore filter (pore size, 0.45 μm). Aliquots (50 μl) were removed and analyzed by using a high-performance liquid chromatograph (Spectra Physics SP8800) equipped with a variable-wavelength UV/VIS detector (SP8450), a computing integrator (SP4290), and a microparticulated (particle size, 10 μm; pore size, 100 Å) reversed-phase column (Nucleosil C₁₈; length, 250 mm; inside diameter, 4.6 mm; Phenomenex Laboratories, Torrance, Calif.). The mobile phase was 0.05 M K₂HPO₄ (pH 4)—CH₃CN (99:1, vol/vol). The flow rate was 2.5 ml/min, and the eluate was monitored at 254 nm. Under these conditions the retention times for Tyr, 2,5-OH-PhAc, 3,4-dihydroxyphenylacetate (3,4-OH-PhAc), 4-OH-PhAc, 3-OH-PhAc, 2-hydroxyphenylacetate (2-OH-PhAc), and PhAc were 3, 7, 11, 20, 23, 29, and 45 min, respectively.

NMR analyses.

NMR spectral analyses were recorded at 20°C with a Varian 400 Mercury VNMRX spectrometer at 400 MHz (¹H) and 100 MHz (¹³C) by using tetramethylsilane as the internal standard. Spectra were measured in CD₃OD.

Liquid chromatography-MS studies.

Mass spectrum analyses were carried out with a Waters ZMD system by using the following parameters: detection range, 200 to 300 nm; capillary power, 3.5 kV; cone power, 25 to 40 V; scan 1, 70 V; scan 2, interphase ES⁺.

Construction of an E. coli strain harboring chromosomal insertions of the Phmg-lacZ translational fusions.

Plasmid pUT-Phmg-lacZ (Table 1; see Fig. 7A), which contained a mini-Tn5Km2 hybrid transposon expressing the Phmg-lacZ fusion, was transferred from E. coli S17-1λpir into the rifampin-resistant E. coli AF141 strain through biparental filter mating as described previously (11). Transconjugants containing the lacZ translational fusions stably inserted into the chromosome were selected for the transposon marker, kanamycin, on rifampin-containing LB medium. One of these transconjugants, E. coli BMR (Table 1), was used in lacZ gene expression experiments.

FIG. 7.

Scheme for subcloning of the hmg regulatory elements: schematic representation of construction of a Phmg::lacZ translational fusion cassette (A) and of plasmid pQ-hmgR harboring the regulatory hmgR gene (B). DNA fragments were PCR amplified by using primers described in Table 2. The Phmg fragment was cloned into the promoter-probe pSJ3 plasmid. The hmgR fragment was cloned into the pQE32 gene expression vector. ΔhmgA indicates a truncated hmgA gene (the number of amino acid [aa] residues fused to the LacZ protein is shown in parentheses). T7, t_o, and rrnB, transcriptional terminators from the T7 and lambda phages and T1 transcriptional terminator from the E. coli rrnB operon, respectively; I and O, termini of the mini-Tn5 transposons; Ap^r and Km^r, genes that confer ampicillin and kanamycin resistance, respectively; tnp*, gene devoid of NotI sites encoding the Tn5 transposase; PT5/lacO, hybrid promoter-operator region composed of the PT5 promoter of phage T5 and the lacO operator from the E. coli lac cluster; oriTRP4, RP4-mediated mobilization functions. The R6K and ColE1 origin of replication (oriR6K and oriColE1) are indicated. Restriction sites: B, BamHI; E, EcoRI; H, HindIII; K, KpnI; N, NotI; S, SalI.

Production of the HmgR repressor.

The hmgR gene was expressed from the strong phage T5 promoter (under lac operator control) in the high-copy-number pQ-hmgR plasmid (Table 1; see Fig. 7B). To prepare crude extract containing the HmgR protein, E. coli JM109(pQ-hmgR) cells were grown in ampicillin-containing LB medium to an A₅₄₀ of about 0.5, and then 1 volume of LB medium containing 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG) was added and the culture was incubated until an A₅₄₀ of about 1.0 was reached. Cells were then collected by centrifugation (5,000 × g, 10 min, 4°C), washed, and resuspended in 0.02 volume of 20 mM Tris-HCl buffer (pH 7.5) containing 10% glycerol, 2 mM β-mercaptoethanol, and 50 mM KCl prior to disruption by sonication. The cell debris was removed by centrifugation at 26,000 × g for 30 min at 4°C. The clear supernatant fluid was decanted and used as the crude cell extract. The control extract (HmgR⁻), obtained from E. coli JM109(pQE32) cells, was prepared in exactly the same way. The protein concentration was determined by the method of Bradford (4) by using bovine serum albumin as the standard.

Mapping the transcription start site by primer extension analysis.

E. coli AF141(pSJ-Phmg-lacZ) cells were grown in MM containing 0.5% glycerol until the culture reached an A₅₄₀ of about 1.0. Total RNA was isolated by using an RNA/DNA Midi kit (QIAGEN) according to the instructions of the supplier. Primer extension reactions were carried out with the avian myeloblastosis virus reverse transcriptase (Promega) by using primer O-Phmg3 (which hybridized with the coding strand between nucleotides 103 and 124 downstream of the hmgA translational start codon [Table 2]). To determine the length of the primer extension product, sequencing reactions with pSJ-Phmg-lacZ were carried out with the same primer (O-Phmg3) by using a T7 sequencing kit and [α-³²P]dCTP (Amershan Pharmacia Biotech) as indicated by the supplier. Products were analyzed on 6% polyacrylamide-urea gels. The gels were dried onto Whatman 3MM paper and exposed to Hyperfilm MP (Amersham Pharmacia Biotech).

Synthesis of the Phmg probe.

The Phmg fragment (335 bp) utilized as a probe was generated by PCR by using plasmid pSJ-Phmg-lacZ (see Fig. 7A) as the template and oligonucleotides O-Phmg5 (which hybridized with the coding strand between nucleotides 89 and 115 downstream of the hmgR translational start codon) and O-Phmg3 as the primers (Table 2). The O-Phmg3 primer was previously labeled (50 pmol) at its 5′ end with phage T4 polynucleotide kinase and [γ-³²P]ATP (3,000 Ci/mmol; Amersham Pharmacia Biotech). To perform the PCR, 10 ng of DNA template (pSJ-Phmg-lacZ), 5 pmol of labeled primer O-Phmg3, and 7.5 pmol of unlabeled primer O-Phmg5 were used; in this way a singly 5′-end-labeled probe at the noncoding strand with respect to the hmgA gene was obtained. The PCR-labeled product was purified with a High Pure PCR product purification kit from Roche Molecular Biochemicals.

Gel retardation assays.

For gel retardation assays the reaction mixtures (final volume, 20 μl) contained in a glutamate buffer solution (20 mM HEPES [pH 8.0], 5 mM magnesium chloride, 2 mM dithiothreitol, 50 mM potassium glutamate) 0.1 nM DNA probe, 500 μg of bovine serum albumin per ml, 100 μg of salmon sperm DNA (competitor) per ml, and cell extract from JM109(pQ-hmgR) or JM109(pQE32) cells. After incubation for 20 min at 20°C, the mixtures were fractionated by electrophoresis in 4% polyacrylamide gels buffered with 0.5× TBE (45 mM Tris-borate, 1 mM EDTA). The gels were dried onto Whatman 3MM paper and exposed to Hyperfilm MP (Amersham Pharmacia Biotech).

DNase I footprinting assays.

The DNase I footprinting assay was carried out in 25 μl (final volume) of a glutamate buffer solution containing 1 nM labeled Phmg probe, 500 μg of bovine serum albumin per ml, and cell extract. This mixture was incubated for 20 min at 30°C, after which 0.15 U of DNase I (Amersham Pharmacia Biotech) (prepared in a solution containing 10 mM CaCl₂, 50 mM MgCl₂, 125 mM KCl, and 10 mM Tris-HCl [pH 7.5]) was added, and incubation was continued at 37°C for 20 s. The reaction was stopped by addition of 180 μl of a solution containing 0.4 M sodium acetate, 2.5 mM EDTA, 50 μg of tRNA per ml, and 5 μg of salmon DNA per ml. After phenol-chloroform extraction, DNA fragments were precipitated with absolute ethanol, washed with 70% ethanol, dried, and directly resuspended in 5 μl of 90% (vol/vol) formamide-loading gel buffer (10 mM Tris-HCl [pH 8.0], 20 mM EDTA [pH 8.0], 0.05% [wt/vol] bromophenol blue, 0.05% [wt/vol] xylene cyanol). Samples were then denatured at 95°C for 2 min and fractionated in a 6% polyacrylamide-urea gel. A+G Maxam-Gilbert reactions (40) were carried out with the same fragments, and the mixtures were loaded in the gels along with the footprinting samples. The gels were dried onto Whatman 3MM paper and exposed to Hyperfilm MP.

Nucleotide sequence accession numbers.

The nucleotide sequences reported in this paper have been submitted to the GenBank/EBI Data Bank; the accession numbers are AY168852, AY168853, AY168854, and AY168855.

RESULTS AND DISCUSSION

Identification of the genes involved in the central pathway for the catabolism of phenylalanine and tyrosine: the homogentisate cluster.

During the course of a research program designed to characterize the genes involved in the catabolism of aromatic compounds in P. putida U, by using a library of mutants constructed by Tn5 transposon mutagenesis, we identified three strains, designated P. putida U-7, U-95, and U-SG6 (Table 1), that were unable to grow on Phe and Tyr as sole carbon and energy sources. When growing on LB medium, these mutants accumulated a black pigment. The same phenotype was observed when these strains were cultured in MM containing Tyr as the source of colored intermediates and 4-OH-PhAc, PhAc, or citrate as the carbon source for supporting bacterial growth (Fig. 2). HPLC and NMR analyses of the culture broth of these mutants revealed the presence of 2,5-OH-PhAc (see Materials and Methods), suggesting that insertion of the Tn5 transposon into the chromosome of the P. putida U mutant strains had disrupted the 2,5-OH-PhAc dioxygenase activity (see below). The NMR data for the 2,5-OH-PhAc (Fig. 3) are as follows. ¹H-NMR (CD₃OD): δ = 3.60 (2H, bs, CH₂), 6.9 (1H, d, J = 8.8 Hz, H-3), 6.7 (1H, dd, J = 8.8, J =2.6 Hz, H-4) and 6.8 (1H, d, J = 2.6 Hz, H-6). ¹³C-NMR (CD₃OD): δ = 175.6 (COOH), 32.6 (CH₂), 124.4 (C-1), 154.1 (C-2), 110.5 (C-3), 114.5 (C-4), 147.8 (C-5′), and 111.7 (C-6). The 2,5-OH-PhAc that accumulates in culture broth becomes oxidized to a quinoid derivative, which by spontaneous polymerization generates melanic compounds that confer the characteristic black or brown color to the medium (61). The accumulation of 2,5-OH-PhAc in the culture broth of the P. putida mutant strains growing in the presence of Tyr suggests that this amino acid is metabolized in this microorganism via the homogentisate pathway. Moreover, sequence analysis of the chromosomal region flanking the transposon insertion site in the three mutant strains allowed us to identify an open reading frame whose product showed a high level of similarity (56% amino acid sequence identity) with the 2,5-OH-PhAc dioxygenase from Sinorhizobium meliloti (43). Additional genetic engineering approaches facilitated cloning and sequencing of a DNA fragment containing a gene cluster (hmg) made up of four open reading frames (Fig. 1A). Three of these open reading frames, hmgA, hmgB, and hmgC, appear to form a single transcriptional unit, and they are likely to encode the 2,5-OH-PhAc dioxygenase (HmgA), fumarylacetoacetate hydrolase (HmgB), and maleylacetoacetate isomerase (HmgC) that convert 2,5-OH-PhAc into fumarate and acetoacetate (Fig. 1B). A putative regulatory gene, hmgR, is divergently transcribed from the hmgABC catabolic genes and encodes a protein that shows similarity with members of the IclR family of transcriptional regulators (72). A homologous cluster (>98% identity) was identified between positions 5241 and 5245 of the P. putida KT2440 genome (Fig. 4). The HmgA, HmgB, and HmgC proteins showed significant amino acid sequence identity with the homogentisate dioxygenase (51%), fumarylacetoacetate hydrolase (44%), and maleylacetoacetate isomerase (39%) involved in degradation of 2,5-OH-PhAc in Emericella nidulans (22). The G+C content of the hmg genes averaged 64.6%, a value close to the mean G+C content (61%) of the P. putida genome (47), which suggests that the hmg cluster either was imprisoned within the chromosome of this bacterium over a long period of evolution or came from a different bacterium having a similar G+C content. A genomic search in microbial databases revealed the existence of similar hmgRABC clusters in other Pseudomonas species whose genomes are totally or partially known, such as Pseudomonas aeruginosa (71) and Pseudomonas fluorescens; the arrangement of the hmgC gene in Pseudomonas syringae is different (5) (Fig. 4). A putative transport gene (hmgT) is located downstream of hmgC in P. fluorescens and P. aeruginosa. Equivalent hmg clusters outside the genus Pseudomonas have been also detected, and they show gene organizations that are different in different bacteria. Whereas in S. meliloti the hmg genes are organized in a way similar to the way in which they are organized in P. putida, in Ralstonia solanacearum, Bordetella bronchiseptica, Bradyrhizobium japonicum, and Silicibacter pomeroyi, the hmgC gene is located outside the hmgRAB cluster, as has been observed in P. syringae (Fig. 4). In Azotobacter vinelandii and Xanthomonas axonopodis, hmgA is not associated with the hmgBC genes. It is worth noting that the hmgB gene is duplicated in Caulobacter crescentus and Mesorhizobium loti; i.e., one copy (hmgB1) is clustered with the hmgRA genes, and the other (hmgB2) is linked to hmgC (Fig. 4).

FIG. 2.

Pigment production (browning) by wild-type and mutant P. putida U strains: growth of wild-type P. putida U (a) and the P. putida U-95 mutant strain (b) on MM containing 5 mM Tyr and 4-OH-PhAc (plate 1) or 5 mM 3-OH-PhAc and 4-OH-PhAc (plate 2).

FIG. 3.

Structure of 2,5-OH-PhAc.

FIG. 4.

Gene organization of the clusters encoding the homogentisate central pathway and the Phe/Tyr peripheral pathway in P. putida KT2440, and comparisons with equivalent gene clusters from other bacteria. Genes are represented by arrows as follows: black, regulatory genes; stippled, transport genes; vertically striped, genes encoding the homogentisate dioxygenase; hatched, genes encoding the hydrolase and isomerase of the homogentisate pathway; cross-hatched, genes encoding the 4-OH-PhPyr dioxygenase; white, genes encoding the tyrosine aminotransferase; horizontally striped, genes encoding the phenylalanine hydroxylase and carbinolamine dehydratase. The numbers beneath the arrows indicate the levels of amino acid sequence identity (expressed as percentages) between the encoded gene products and the equivalent products from P. putida. Identity values are not shown for the hmgR gene products that do not belong to the IclR family of transcriptional regulators. The genomes of P. fluorescens, A. vinelandii, and S. pomeroyi are not completely assembled.

To confirm that the hmg cluster encoded the central homogentisate pathway for the catabolism of Phe and Tyr, as well as to eliminate the possibility that other undetected mutations were responsible for the growth deficiencies observed in the P. putida U-7, U-95, and U-SG6 mutant strains, we constructed insertion mutants with mutations in the hmgA gene (P. putida U-dhmgA), the hmgB gene (P. putida U-dhmgB), and the hmgC gene (P. putida U-dhmgC), as well as a mutant strain in which the three catabolic genes (hmgABC) were deleted (P. putida U-Δhmg) (Table 1) (see Materials and Methods). All these mutants were unable to grow in MM containing Phe or Tyr as the sole carbon source (Fig. 5), but they accumulated 2,5-OH-PhAc in the culture broth when they were grown in MM containing either of these amino acids and either 4-OH-PhAc, PhAc, or citrate as a carbon and energy source (data not shown). Furthermore, the transformation of P. putida U-Δhmg with plasmid pMChmg, which contains the hmgABC genes (Table 2), restored the ability of the mutant to grow in MM (Fig. 5). The fact that mutations in the hmgB and hmgC genes caused secretion of 2,5-OH-PhAc into the broth (as revealed by HPLC analysis) suggests that the homogentisate pathway might be strictly regulated at the level of homogentisate dioxygenase to prevent accumulation of catabolic intermediates, such as maleylacetoacetate or fumarylacetoacetate, that could have toxic effects on the cell. This is in agreement with the observation that traces of 2,5-OH-PhAc were detected in the broth when P. putida U was cultured in MM in the presence of a high concentration (>10 mM) of Tyr (data not shown).

FIG. 5.

Tyr and 3-OH-PhAc share the same central catabolic pathway: growth of wild-type P. putida U (•) and the mutants P. putida U-dhmgA (▪), U-dhmgB (▴), U-dhmgC (▾), U-Δhmg (○), U-Δhmg(pMChmg) (▵) in MM containing 5 mM Tyr (A) or 10 mM 3-OH-PhAc (B) as the sole carbon and energy source.

Remarkably, we observed that whereas P. putida U can grow efficiently in MM containing 3-OH-PhAc as the sole carbon source, all the mutants with disruptions in the homogentisate pathway were unable to degrade this aromatic compound (Fig. 5). HPLC analysis of the culture broth of these mutants grown in MM containing 3-OH-PhAc (as a source of intermediates) and an additional carbon source, such as 4-OH-PhAc, revealed that all of them accumulated 2,5-OH-PhAc that became oxidized to a colored quinoid derivative (Fig. 2). These results strongly suggest that 3-OH-PhAc is assimilated through the homogentisate pathway after previous hydroxylation to 2,5-OH-PhAc (Fig. 1). Thus, the homogentisate pathway appears to be a central convergent route involved in degradation of Phe, Tyr, 3-OH-PhAc, and all the molecules able to generate some of these compounds (e.g., amides and other ester derivatives) in P. putida.

To confirm that the homologous hmg cluster found in the recently sequenced genome of P. putida KT2440 plays a similar role in the catabolism of Phe and Tyr, we constructed the P. putida KT2440-dhmgA strain (Table 1) (see Materials and Methods). As expected, whereas this mutant grew in citrate-containing MM, it did not use Phe or Tyr as a carbon source, and it produced a black or brown pigment when it was cultured on LB medium plates. Furthermore, whereas high homogentisate dioxygenase activity (68 mU/mg of protein) was detected in cell extracts of the wild-type strain when it was cultured in MM containing citrate and Phe or Tyr, no activity was found in crude extracts of the P. putida KT2440-dhmgA mutant cultured in the same medium and conditions or in extracts of the wild-type strain cultured in citrate-containing MM in the absence of Phe and Tyr. These results strongly suggest that the hmgA gene encodes an enzyme able to cleave the aromatic ring of homogentisate in P. putida.

The homogentisate dioxygenase (HmgA) from P. putida KT2440 displays optimal activity and stability at 37°C in 100 mM potassium phosphate buffer (pH 7.0) in the presence of 2 mM ascorbate and 50 μM FeSO₄. Under these conditions, the enzyme showed a hyperbolic kinetic behavior toward increasing concentrations of 2,5-OH-PhAc. The apparent K_m for 2,5-OH-PhAc (27 μM) is similar to that found for the equivalent enzyme from the rat (10 μM) (64) or from E. nidulans (9 μM) (21). However, these data contrast with the high K_m values reported for the murine homogentisate dioxygenase (180 μM) (63) and for the homogentisate dioxygenase from P. fluorescens (600 μM) (1).

Identification of the peripheral pathway for catabolism of phenylalanine and tyrosine in P. putida.

The results presented above suggest that the homogentisate cluster is the central pathway through which Phe and Tyr (in P. putida U and KT2440) and 3-OH-PhAc (in P. putida U) are catabolized. To identify the genes involved in the transformation of Phe and Tyr into 2,5-OH-PhAc, we isolated by Tn5 transposon mutagenesis different P. putida U mutants (strains U-111 and U-215 [Table 1]) which were unable to catabolize Phe and Tyr but which were able to grow in MM containing 3-OH-PhAc and expressed, therefore, a functional homogentisate pathway. Sequence analysis of the DNA fragment flanking the transposon insertion site revealed that whereas P. putida U-111 contains the Tn5 transposon within an open reading frame that encodes a putative 4-OH-PhPyr dioxygenase (Hpd), P. putida U-215 harbors the Tn5 insertion within a gene cluster encoding a putative pterin-dependent phenylalanine hydroxylase (phhA) that converts Phe into Tyr (81), a putative carbinolamine dehydratase (phhB) involved in regeneration of the pterin cofactor (69), and the putative σ⁵⁴-dependent transcriptional activator (phhR) of the phh operon (68) (Fig. 1). The phhRAB and hpd genes are homologous to the genes previously characterized in P. aeruginosa (69) and P. fluorescens (66), respectively (Fig. 4).

When we searched for the homologous genes in the genome of P. putida KT2440, we identified, between positions 5100 and 5111 kb of the chromosome, the phhRABT cluster (the phhT gene encodes a putative transport protein) close to a gene (aroP2) encoding a general aromatic amino acid permease (Fig. 4). In P. aeruginosa the phhC gene encodes a tyrosine aminotransferase that transforms tyrosine into 4-OH-PhPyr and is essential for the catabolism of both Phe and Tyr (33). Although there is no phhC homolog in the phh cluster of P. putida, two genes at positions 2233 kb (tyrB1) and 4080 kb (tyrB2) of the KT2440 genome could encode this tyrosine aminotransferase function. It is worth noting that while in P. aeruginosa and P. fluorescens the phh genes form a cluster, the tyrB genes are not linked to the phhRAB operon in P. putida and P. syringae (Fig. 4) (35). This organization is similar to that found in other bacteria, like A. vinelandii, X. axonopodis, and R. solanacearum. The hpd gene located at position 3890 kb of the P. putida KT2240 genome may encode the 4-OH-PhPyr dioxygenase that converts 4-OH-PhPyr into 2,5-OH-PhAc (Fig. 1). In P. putida, P. fluorescens, A. vinelandii, R. solanacearum, and S. pomeroyi the hpd gene is not linked to other genes involved in Phe and Tyr degradation, while this gene is associated with the phh cluster in P. aeruginosa and with the hmg genes in P. syringae, X. axonopodis, C. crescentus, B. japonicum, M. loti, and S. meliloti (Fig. 4).

To confirm that the phh and hpd genes were involved in the catabolism of Phe and Tyr in P. putida, we disrupted some of these genes in strain KT2440 and strain U, and then we monitored the growth of the resulting mutants in MM containing Phe or Tyr as a carbon source. Insertional inactivation of the phhA gene in P. putida U and in P. putida KT2440 (see Materials and Methods) generated the P. putida U-dphhA and P. putida KT2440-dphhA strains, respectively (Table 1). These strains were unable to grow in MM containing Phe as the sole carbon source, but they grew on Tyr (both mutants) and 3-OH-PhAc (P. putida U-dphhA), producing normal levels of the HmgA enzyme (65 mU/mg of protein). Therefore, these results suggest that the phhA gene is involved in the transformation of Phe into Tyr in P. putida. Since these mutants were not auxotrophs for Tyr, a pathway other than hydroxylation of Phe should be functional in these bacteria for the biosynthesis of Tyr.

P. putida mutants in which the hpd gene was disrupted (P. putida U-dhpd and P. putida KT2440-dhpd) or deleted (P. putida U-Δhpd) (Table 1) were unable to grow in MM containing either Phe or Tyr as the sole carbon and energy source, although they grew in MM containing citrate (or 3-OH-PhAc for P. putida U derivatives). The hpd genes from P. putida KT2440 and P. putida U were cloned and expressed under control of the Plac promoter in plasmids pU-hpd and pG-hpd, respectively (Table 2). When the recombinant E. coli DH5α cells containing pU-hpd or pG-hpd were grown in LB medium or in MM containing glycerol (10 mM) and Tyr or 4-OH-PhPyr (1 mM), secretion and accumulation of 2,5-OH-PhAc in the culture broth were observed (data not shown). These data confirm that the hpd gene encodes a 4-OH-PhPyr dioxygenase, and they are in agreement with previous observations revealing that E. coli, as well as other bacteria that do not use Tyr as a carbon source, can synthesize at least one transaminase that is able to convert Tyr into 4-OH-PhPyr (29, 41).

It is worth noting that the hpd mutant of P. putida U (but not the P. putida KT2440-dhpd mutant) showed a brown pigmentation (browning) when it was cultured in MM containing Tyr and 4-OH-PhAc as carbon sources. However, browning was not observed when this mutant was cultured in MM containing Tyr (as a source of intermediates) and citrate, octanoate, or PhAc as carbon sources. HPLC analysis of the compounds released into the culture medium before the appearance of the brown pigment revealed the presence of 3,4-dihydroxyphenylpyruvate. Thus, these results suggest that the brown compound is a derivative of the 3,4-dihydroxyphenylpyruvate produced through hydroxylation of 4-OH-PhPyr by some enzymes of the 4-OH-PhAc catabolic pathway. Since the 4-OH-PhAc monooxygenase uses a wide range of substrates and is present in P. putida U (48, 55) but not in P. putida KT2440 (35), it could be the enzyme responsible for the hydroxylation of 4-OH-PhPyr. To confirm this hypothesis, we constructed the P. putida U-A2dhpd mutant strain that harbors a disruption of both the hpd and hpaB genes (hpaB encodes the large subunit of the 4-OH-PhAc monooxygenase) (Table 1). The P. putida U-A2dhpd strain did not show the brown phenotype when it was grown in citrate-containing MM supplemented with Tyr and 4-OH-PhAc, and the mutant did not produce 3,4-dihydroxyphenylpyruvate. These data indicate that the production of 3,4-dihydroxyphenylpyruvate and the browning are specifically related to the presence of Tyr (as a source of 4-OH-PhPyr) and to the existence of an active 4-OH-PhAc monooxygenase whose expression is inducible by 4-OH-PhAc.

Engineering a mobile catabolic cassette for the catabolism of tyrosine.

As described above, we characterized a set of genes involved in the catabolism of Tyr in P. putida. To demonstrate that these genes encoded all the functions necessary to degrade Tyr, we constructed plasmids pU-HHP and pM2-HHP that contained the hmgABC, hpd, and tyrB1 genes from P. putida KT2440 engineered as a 5.8-kb DNA cassette (Tyr cassette) (see Materials and Methods). Whereas the Tyr cassette in plasmid pU-HHP was expressed under control of the tandem Plac and Phmg promoters, the mobilizable and broad-host-range plasmid pM2-HHP expressed the Tyr cassette only under control of the Phmg promoter (Fig. 6A). This catabolic cassette was shown to be functional in heterologous hosts since both plasmids conferred to E. coli the capacity to grow in MM containing Tyr as the sole carbon source (Fig. 6B). To our knowledge, this is the first report of an E. coli strain able to mineralize Tyr. These data confirm, therefore, that the hmg, hpd, and tyr gene products are the only gene products required for complete catabolism of Tyr in bacteria. It is worth mentioning that although the tyrB1 gene in the DNA cassette ensures efficient expression of the tyrosine aminotransferase, E. coli and some other bacteria are equipped with their own aminotransferases able to transform Tyr into 4-OH-PhPyr (see above) (42).

FIG. 6.

Catabolic cassette for the catabolism of Tyr. (A) Schematic representation of the construction and expression of a Tyr catabolic cassette. The primer pairs used for PCR amplification are shown in Table 2. Genes are indicated by arrows. The Plac and Phmg promoters are shown. Ap^r and Km^r, genes that confer ampicillin and kanamycin resistance, respectively. The pBBR1 and pUC origins of replication (oripBBR1 and oripUC) are indicated. oriTRP4, RP4-mediated mobilization (Mob⁺) functions. (B) Growth of recombinant strain E. coli(pU-HHP) (•) and control strain E. coli(pUC18) (▴) in MM supplemented with 0.05% Casamino Acids and 5 mM Tyr as the sole carbon and energy source.

The Tyr cassette, reported for the first time in this work, is a useful tool for metabolic engineering. Thus, this cassette can be used both to expand the catabolic potential of many microbes lacking a Tyr catabolic pathway and to improve the Tyr degradation rate in those bacteria that are already able to mineralize this aromatic compound.

In vivo transcriptional analysis of the hmg catabolic genes.

Taking into account the fact that the intergenic regions between the hmgA and hmgB genes and between the hmgB and hmgC genes span 3 and 12 bp, respectively, the hmgABC genes might constitute a catabolic operon. As shown above, the homogentisate pathway is induced when P. putida cells are grown on Phe or Tyr. To investigate in vivo the role of the putative HgmR regulatory protein in the induction of the hmgABC genes, we constructed a P. putida strain with the divergently transcribed hmgR gene disrupted (Fig. 4). The resulting P. putida KT2440-dhmgR mutant (Table 1) was able to use Phe and Tyr as sole carbon and energy sources, but, in contrast with the wild-type strain, this mutant constitutively produced a normal amount of HmgA enzyme (72 mU/mg of protein). Therefore, these results suggest that the hmgR gene product is a repressor protein which regulates the inducible hmg catabolic operon.

To further study the regulatory elements of the hmgABC operon, a DNA fragment containing the potential promoter (Phmg) was PCR isolated and ligated to the lacZ gene of the promoter-probe vector pSJ3 (Table 1). The resulting translational fusion plasmid, pSJ-Phmg-lacZ (Fig. 7A), conferred to the host strain (E. coli CC118) the ability to produce blue colonies on media containing the β-galactosidase indicator 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal), indicating the presence of a functional promoter in the cloned fragment. To further analyze this regulatory system, we engineered the reporter Phmg-lacZ fusion within a mini-Tn5 vector. The resulting construct, pUT-Phmg-lacZ (Fig. 7A), was used to deliver by transposition the corresponding translational fusion into the chromosome of E. coli AF141 (lacZ), giving rise to the reporter strain E. coli BMR (Table 1). To check the influence of the HmgR protein on the expression of the reporter fusion, hmgR was cloned in plasmid pQE32, producing plasmid pQ-hmgR (Fig. 7B). β-Galactosidase assays of permeabilized E. coli BMR cells harboring the control plasmid pQE32 showed that there was constitutive expression of the reporter fusion (Table 3). However, when the hmgR gene was expressed in trans in E. coli BMR(pQ-hmgR) cells growing in glycerol-containing MM, we observed a drastic decrease (more than 2 orders of magnitude) in the β-galactosidase levels, thus indicating that HmgR behaves as a transcriptional repressor of the Phmg promoter. Moreover, since the repressor effect of HmgR was avoided by growing the cells in the presence of 2,5-OH-PhAc, but not when the cells where grown in the presence of Phe, Tyr, or 4-OH-PhPyr (Table 3), we concluded that 2,5-OH-PhAc is the inducer of the homogentisate operon.

TABLE 3.

Expression from the Phmg promoter is controlled by HmgR^a

E. coli strain	Plasmid	Relevant background	β-Galactosidase activity (Miller units)
			Uninduced	Phe	Tyr	4-OH-PhPyr	2,5-OH-PhAc	2,5-OH-benzoate
AF141	None	lacZ	NDb	ND	ND	ND	ND	ND
BMR	None	Phmg-lacZ	2,520	2,462	2,347	2,269	2,017	2,303
BMR	pQE32	Phmg-lacZ	1,890	1,895	1,912	1,425	1,343	1,593
BMR	pQ-hmgR	Phmg-lacZ hmgR	9	15	14	27	1,047	198

^aE. coli strains were grown in glycerol-containing MM in the absence (uninduced) or in the presence of 5 mM Phe, Tyr, 4-OH-PhPyr, 2,5-OH-PhAc (homogentisate), or 2,5-OH-benzoate (gentisate). β-Galactosidase activities were measured with permeabilized cells as described in Materials and Methods.

^bND, not detected.

To determine the transcription initiation site at the Phmg promoter, primer extension analyses were performed with total RNA isolated from E. coli AF141(pSJ-Phmg-lacZ) cells by using primer O-Phmg3 that hybridized within the hmgA gene (see Materials and Methods). The transcription initiation site of the Phmg promoter mapped 37 nucleotides upstream of the ATG translation initiation codon of the hmgA gene (Fig. 8). Analysis of the Phmg promoter region revealed a typical organization of σ⁷⁰-dependent promoters with a −10 box (TACGTT) located at a consensus distance (17 bp) from a highly conserved −35 box (TTGACG) (nucleotides that match the nucleotides in the consensus sequence are underlined) (Fig. 8).

FIG. 8.

Analysis of the hmgR-hmgA intergenic region. (A) Identification of the transcription start site in Phmg. Primer extension experiments were carried out by using total RNA isolated from E. coli AF141 cells bearing the lacZ translational fusion plasmid pSJ-Phmg-lacZ (lane 2) and the control plasmid pSJ3 (lane 1). The size of the extended product was determined by comparison with the DNA sequencing ladder of the Phmg promoter region (lanes T, C, G, and A). Primer extension and sequencing reactions were performed with primer O-Phmg3 as described in Materials and Methods. The nucleotide sequence surrounding the transcription initiation site (enclosed in a box) in the coding strand is shown. (B) Schematic representation of regulation of the hmg cluster and nucleotide sequence of the hmgR-hmgA intergenic region. The hmgR regulatory gene is indicated by a thick grey arrow. The hmgABC catabolic genes are indicated by a thick open arrow. The minus sign indicates transcriptional repression by the HmgR protein. The plus sign indicates transcriptional activation (induction) promoted by homogentisate. Homogentisate is transformed into fumarate and acetoacetate by the HmgABC proteins. The nucleotide sequence of the Phmg probe (335 bp) is indicated. The translation initiation codon for the hmgA and hmgR genes is indicated by boldface lowercase letters; the bent arrows indicate the direction of transcription. The transcription start site (position +1) and the inferred −10 and −35 boxes of the Phmg promoter are indicated. The HmgR binding region is indicated by brackets. The repeated motifs are indicated by thin grey arrows. RBS, ribosome binding site.

The HmgR protein exhibits amino acid sequence identity with other regulators of aromatic catabolic pathways that belong to the IclR family of transcriptional regulators, including the activators PobR (24%) and PcaU (25%) of the 4-hydroxybenzoate and protocatechuate degradation pathways in Acinetobacter sp. strain ADP1 (12, 30), PcaR of the protocatechuate degradation pathways in P. putida and Agrobacterium tumefaciens (27 and 23%, respectively) (51, 57), CatR (23%) and PcaR (23%) of the catechol and protocatechuate degradation pathways in Rhodococcus opacus 1CP (17, 18), and MhpR (26%) of the 3-hydroxyphenylpropionate degradation pathway in E. coli (74). Despite the fact that many IclR-type transcriptional regulators are repressors, HmgR is the first transcriptional regulator that has been described for the catabolism of aromatic compounds. Interestingly, hmgA expression in S. meliloti was shown to be induced under nitrogen deprivation conditions by an activator (NitR) belonging to the ArsR family of regulators, but so far it is not known whether such activation involves a direct interaction of NitR with the Phmg promoter or, on the contrary, NitR controls the expression of another regulator (for instance, the hmgR gene from the hmg cluster [Fig. 4]) which in turn controls the hmgA gene expression (44). Since no homolog of NitR was found in the genome of P. putida and the HmgR proteins do not exhibit similarity (Fig. 4), the expression of the hmg genes seems to be controlled by different regulatory mechanisms in these two bacteria. To further characterize the HmgR-mediated regulation of the Phmg promoter from P. putida, we performed some in vitro studies.

In vitro analyses of the HmgR-dependent control at the Phmg promoter.

To demonstrate the specific interaction of the HmgR regulatory protein with the Phmg promoter, cell extracts from E. coli JM109(pQ-hmgR) were subjected to gel retardation assays by using as the probe a 335-bp PCR-generated fragment carrying the intergenic hmgR-hmgA region (Phmg probe). Whereas extracts containing HmgR were able to retard the migration of the Phmg probe in a protein concentration-dependent manner, control extracts from E. coli JM109(pQE32) did not do this (Fig. 9A), which demonstrates that there is specific binding of the HmgR protein to the Phmg probe. Gel retardation assays were also carried out in the presence of different concentrations of homogentisate at a concentration of HmgR that retards completely migration of the Phmg probe. As shown in Fig. 9B, increasing concentrations of 2,5-OH-PhAc decreased retardation of the DNA probe, and the interaction of HmgR with Phmg was completely abolished at 50 μM 2,5-OH-PhAc. These data are in agreement with the conclusions provided by lacZ-reporter fusion experiments reported above, and they confirm that homogentisate is the inducer of the hmg catabolic genes.

FIG. 9.

Gel retardation analyses of HmgR binding to the hmgR-hmgA intergenic region. Cell extracts were prepared and gel retardation analyses were performed as described in Materials and Methods. The probe DNA used, Phmg, was PCR amplified from plasmid pSJ-Phmg-lacZ as described in Materials and Methods. (A) Lanes 1 to 7, retardation assay mixtures containing 0, 0.5, 0.7, 1.0, 1.5, 2.0, and 3.0 μg of total protein, respectively, of HmgR⁺ extracts obtained from cells bearing plasmid pQ-hmgR; lane 8, assay mixture containing 3.0 μg of total protein of HmgR⁻ extracts obtained from cells bearing the control plasmid pQE32. (B) Lanes 2 to 8, retardation assay mixtures containing 3.0 μg of total protein of HmgR⁺ extracts in the absence of 2,5-OH-PhAc (lane 2) or in the presence of increasing concentrations of 2,5-OH-PhAc, as follows: 1 μM (lane 3), 2.5 μM (lane 4), 5.0 μM (lane 5), 10.0 μM (lane 6), 25.0 μM (lane 7), and 50.0 μM (lane 8). Lane 1 shows migration of the Phmg probe without protein extract. (C) Gel retardation assays with 3.0 μg of total protein of HmgR⁺ extracts and the following different ligands at a concentration of 1 mM: 2,5-OH-PhAc (lane 3), 2,5-OH-benzoate (2,5-OH-Benz) (lane 4), 2-OH-PhAc (lane 5), 3-OH-PhAc (lane 6), 3,4-OH-PhAc (lane 7), 4-OH-PhPyr (lane 8), PhAc (lane 9), Phe (lane 10), and Tyr (lane 11). Lanes 1 and 2 contained assay mixtures lacking HmgR⁺ extract and ligand, respectively. The positions of DNA probes and the DNA-HmgR complexes are indicated by arrows.

To check the ligand profile for HmgR, gel retardation experiments were performed by using the Phmg probe and different structural analogs of 2,5-OH-PhAc, such as 2,5-OH-benzoate (gentisate), 2-OH-PhAc, 3-OH-PhAc, 3,4-OH-PhAc, PhAc, and related compounds of the Phe/Tyr catabolic pathway (Phe, Tyr, and 4-OH-PhPyr). Interestingly, only 2,5-OH-PhAc was able to efficiently inhibit binding of HmgR to the Phmg promoter (Fig. 9C). Gentisate, a structural analog of 2,5-OH-PhAc with a side chain that is one carbon shorter, was also able to disturb the HmgR-Phmg interaction, but this effect was more than 2 orders of magnitude less efficient than that caused by homogentisate (Fig. 9C). lacZ-reporter fusion experiments confirmed that 2,5-OH-benzoate (gentisate) was able to induce expression from the Phmg promoter in vivo, and the induction achieved was fivefold lower than that achieved with 2,5-OH-PhAc (Table 3). These results show that the presence of two hydroxy groups at the para position in the benzene ring of the aromatic acid are indispensable for a productive interaction of the inducer molecule with the HmgR repressor, and an aromatic acid with a two-carbon side chain (2,5-OH-PhAc) is the best inducer. Such high specificity for the inducer molecule suggests that the regulatory elements were recruited a long time ago and that they evolved together with the catabolic genes during the evolutionary history of the hmg cluster.

To characterize the HmgR binding site within the Phmg promoter, DNase I footprinting experiments were performed by using the Phmg probe as the target DNA (Fig. 10). A protected 45-bp region that spans from position −16 to position 29 with respect to the Phmg transcription start site was observed (Fig. 8B). The 3′ end of the HmgR binding region partially overlaps the ribosome binding site of the hmgA gene (Fig. 8B), and, as already reported for the IclR regulator (78), a site hypersensitive to DNase I was detected at the 5′ end of the region (Fig. 10). Analysis of the HmgR binding region revealed a 17-bp perfect palindromic motif (TCGTAATCTGATTACGA) with its pseudodyad axis through a central T residue (located at position 5 with respect to the transcription initiation site) that defines two 8-bp half-sites (Fig. 8B). Other regulators of aromatic catabolic pathways that belong to the IclR family, such as PobR and PcaU from Acinetobacter sp. strain ADP1, PcaR from P. putida, and MhpR from E. coli, also recognize 17-bp palindromic operator regions with the pseudodyad axis through a central base, but the consensus sequence (TGTTCGATAATCGCACA) (30, 74) does not resemble that of the HmgR binding site. Interestingly, the HmgR binding region contains a third 6-bp motif (ATTACG), which is located 4 bp upstream of the palindromic motif, partially overlaps the −10 box, and is arranged as a direct sequence repetition of the right-half site of the palindrome (Fig. 8B). Analyses of the putative Phmg promoters from other Pseudomonas species, such as P. fluorescens, P. aeruginosa, and P. syringae, revealed a very similar organization of the presumed operator regions with a 17-bp imperfect palindromic motif separated by 4 bp of a highly conserved direct repeat identical to that of P. putida (Fig. 11). Based on this observation and on the high amino acid sequence identity among the equivalent HmgR proteins from the four Pseudomonas species (Fig. 4), a common regulatory mechanism can be suggested for the homogentisate cluster in Pseudomonas.

FIG. 10.

DNase I footprinting analyses of the interaction of HmgR with the Phmg promoter region. The DNase I footprinting experiments were carried out by using the Phmg probe labeled at the 5′ end of the noncoding strand as described in Materials and Methods. Lanes 1 and 3 to 6 contained footprinting assay mixtures containing 0, 0.1, 0.3, 1.0, and 3.0 μg of total protein of HmgR⁺ extracts (pQ-hmgR), respectively. Lane 2 contained a footprinting assay mixture with 3.0 μg of total protein of HmgR⁻ extracts (pQE32). Lane 7 shows the results for the A+G Maxam-Gilbert sequencing reaction (40) that provided the sequence of the Phmg probe. The HmgR protected region is indicated, and an expanded view of the nucleotide sequence is indicated by brackets. The asterisk indicates the transcription initiation site in the Phmg promoter. A DNase I-hypersensitive site is indicated by an arrow.

FIG. 11.

Comparison of the Phmg promoter regions in several Pseudomonas species. The hmgR-hmgA intergenic regions from P. putida, P. fluorescens, P. aeruginosa, and P. syringae were aligned from position −50 to position 40 with respect to the transcription initiation site in P. putida. The asterisks indicate the conserved nucleotides. The −35 and −10 boxes, the +1 transcription initiation site, the ribosome binding site (RBS), and the ATG translation initiation codon from hmgA in P. putida are indicated by a black background. The HmgR binding region is indicated by brackets. Repeated motifs are indicated by arrows.

A similarly structured protein binding region with a palindromic motif and an external direct repetition has been described for promoters controlled by other IclR-type regulators, such as PcaU, PobR, and PcaR activators (54), and for the promoter controlled by the DeoR repressor from Bacillus subtilis (80). However, the architectures of these regulatory regions show relevant differences for each individual regulator, which might reflect major differences both in the mechanism by which the regulators interact with the RNA polymerase and in the fundamental method of transcriptional regulation (i.e., activation or repression) of the cognate promoters. Although unraveling the mechanism leading to repression of Phmg by HmgR and to induction by homogentisate requires further research, the location of the HmgR binding site overlapping both the −10 box and the transcription initiation site strongly suggests that HmgR physically competes with the RNA polymerase in promoter binding and that the presence of the inducer must change the nature of the HmgR-Phmg interaction in a way that allows transcription initiation by the RNA polymerase.