Journal list menu

Volume 268, Issue 21 p. 5550-5561
Free Access

A novel two-protein component flavoprotein hydroxylase

p-Hydroxyphenylacetate hydroxylase from Acinetobacter baumannii

Pimchai Chaiyen

Pimchai Chaiyen

Department of Biochemistry, Faculty of Science, Mahidol University, Bangkok, Thailand;

Center for Protein Structure and Function, Faculty of Science, Mahidol University, Bangkok, Thailand

Search for more papers by this author
Chutintorn Suadee

Chutintorn Suadee

Department of Biochemistry, Faculty of Science, Mahidol University, Bangkok, Thailand;

Center for Protein Structure and Function, Faculty of Science, Mahidol University, Bangkok, Thailand

Search for more papers by this author
Prapon Wilairat

Prapon Wilairat

Department of Biochemistry, Faculty of Science, Mahidol University, Bangkok, Thailand;

Search for more papers by this author
First published: 19 August 2003
Citations: 102
P. Chaiyen, Department of Biochemistry, Faculty of Science, Mahidol University, Rama 6 Road, Bangkok, 10400, Thailand. Fax: +66 2 248 0375, Tel.: +66 2 201 5607 or +66 2 201 5596, E-mail: [email protected]


p-Hydroxyphenylacetate (HPA) hydroxylase (HPAH) was purified from Acinetobacter baumannii and shown to be a two-protein component enzyme. The small component (C1) is the reductase enzyme with a subunit molecular mass of 32 kDa. C1 alone catalyses HPA-stimulated NADH oxidation without hydroxylation of HPA. C1 is a flavoprotein with FMN as a native cofactor but can also bind to FAD. The large component (C2) is the hydroxylase component that hydroxylates HPA in the presence of C1. C2 is a tetrameric enzyme with a subunit molecular mass of 50 kDa and apparently contains no redox centre. FMN, FAD, or riboflavin could be used as coenzymes for hydroxylase activity with FMN showing the highest activity. Our data demonstrated that C2 alone was capable of utilizing reduced FMN to form the product 3,4-dihydroxyphenylacetate. Mixing reduced flavin with C2 also resulted in the formation of a flavin intermediate that resembled a C(4a)-substituted flavin species indicating that the reaction mechanism of the enzyme proceeded via C(4a)-substituted flavin intermediates. Based on the available evidence, we conclude that the reaction mechanism of HPAH from A. baumannii is similar to that of bacterial luciferase. The enzyme uses a luciferase-like mechanism and reduced flavin (FMNH2, FADH2, or reduced riboflavin) to catalyse the hydroxylation of aromatic compounds, which are usually catalysed by FAD-associated aromatic hydroxylases.


  • HPA
  • p-hydroxyphenylacetate
  • HPAH
  • p-hydroxyphenylacetate hydroxylase
  • DHPA
  • 3,4-dihydroxyphenylacetate
  • 3,4-dihydroxyphenylacetate dioxygenase
  • CHS
  • 5-carboxymethylmuconate semialdehyde
  • C1
  • the small component of Acinetobacter baumannii HPAH
  • C2
  • the large component of Acinetobacter baumannii HPAH
  • hpaC
  • the small component of Escherishia coli HPAH
  • The oxygenation of phenolic compounds is an important step in the biodegradation of synthethic or natural aromatic compounds. Catabolism of these compounds is often initiated by hydroxylase enzymes that incorporate hydroxyl groups into phenolic substrates resulting in catechol products [1]. Hydroxylations at positions ortho to the phenolic group are usually catalysed by aromatic flavoprotein hydroxylases containing FAD as the prosthetic group. The hydroxylation reactions are mostly carried out on single polypeptide chains such as the reactions of p-hydroxybenzoate hydroxylase, phenol hydroxylase, salicylate hydroxylase, anthranilate hydroxylase, and 2-methyl-3-hydroxypyridine-5-carboxylic acid oxygenase [2,3]. However, in the past decade, flavoenzymes catalysing aromatic hydroxylation reactions were found to consist of two proteins, such as the reaction of p-hydroxyphenylacetate (HPA) hydroxylase (HPAH) of Pseudomonas putida[4] and of Escherichia coli W [5], and pyrrole-2-carboxylate monooxygenase [6]. The gene encoding HPAH in Klebsiella pneumoniae also suggests that HPAH exists as a two-component enzyme in this species [7].

    Studies of P. putida HPAH have shown that FAD is tightly bound to the small component and the large component is a coupling protein that enables hydroxylation to occur [4]. The mechanism of P. putida HPAH is similar to the mechanism of other aromatic flavoprotein hydroxylases except that two proteins are required [8,9]. In contrast, studies of E. coli HPAH have shown that the small component of the enzyme is the flavin reductase generating free reduced FAD for the large component to hydroxylate HPA [5,10].

    In these studies, we isolated HPAH from Acinetobacter baumannii. The strain was isolated from Thai soil by using enrichment techniques to obtain bacteria that can utilize HPA as a sole carbon source. HPAH isolated from A. baumannii also requires two-protein components for complete activity. Studies of biochemical and catalytic properties of the enzyme indicated that the enzyme was quite distinct from P. putida HPAH and in many aspects, was different from E. coli HPAH. Our data suggested that the mechanism of A. baumannii HPAH was similar to that of bacterial luciferase. This report also demonstrated the role of FMN in catalysing the hydroxylation of aromatic compounds, the reaction usually found to be associated with FAD.

    Materials and methods


    HPA, 3,4-dihydroxyphenylacetate (DHPA), FAD, FMN, riboflavin, NADH, phenylmethanesulfonyl fluoride, dithiothreitol, DEAE–Sepharose, phenyl-Sepharose, and a standard solution of BSA were from Sigma Chemical Co. Hydroxyapatite, and Bradford dye reagent were from Bio-Rad. Sephadex G-150, and G-25 were from Pharmacia. The concentrations of the following compounds were determined using known extinction coefficients at pH 7.0: NADH, ε340 = 6.22 mm−1·cm−1[11]; HPA, ε277 = 1.55 mm−1·cm−1; DHPA, ε281 = 2.74 mm−1·cm−1. Protein content was determined by the Bradford method [12] using BSA as protein standard. HPA medium contained the following salts per L: K2HPO4, 0.8 g; KH2PO4, 0.2 g; NaCl, 0.05 g; CaSO4·2H2O, 0.05 g; MgSO4·7H2O, 0.5 g; FeSO4·7H2O, 0.01 g; (NH4)2SO4, 1 g; HPA, 1 g; trace elements, 3 mL. The Pseudomonas strain used for production of DHPA dioxygenase (DHPAO) was a generous gift of V. Massey, Department of Biological Chemistry, University of Michigan, Ann Arbor, MI, USA.

    Isolation and growth of organisms

    The enrichment culture technique using HPA as a sole carbon source (HPA medium) was used to isolate an organism that can produce HPAH enzyme. The organism isolated is a coccobacilli Gram-negative bacterium identified as A. baumanii. Large quantities of cells were obtained by culturing bacteria in 16 L of HPA-containing medium using a sterile carboy aerated with sterile air. Cells were harvested at the end of log phase (D600 ≈1.3) by centrifuging at 7000 g, 4 °C for 8 min. Cell pastes were frozen at −80 °C until used.

    HPLC analysis of HPA, DHPA, and HPA analogues

    HPLC analysis of HPA, DHPA and HPA analogues were performed by reversed-phase chromatography (Waters–Nova Pak C18 3.8 × 150 mm) on an HPLC instrument (Waters). The mobile phase was a gradient of 20–40% methanol in 2 mm sodium phosphate pH 3.3, at a flow rate of 1.2 mL·min−1 for 12 min. Detection for HPA, DHPA, and other analogue compounds was carried out by measuring absorbance at 277 nm, except for p-nitrophenol that was monitored at 310 nm. Compounds were identified by comparing their retention times with those of pure substances.

    Enzyme assays

    HPA-stimulated NADH oxidation

    The small component of HPAH (C1) was assayed by monitoring NADH oxidase activity. Decrease in absorbance at 340 nm was used to monitor NADH oxidation (ε340 = 6.22·mm−1·cm−1). Assays (1 mL) were performed at 25 °C and contained 50 mm sodium phosphate pH 7.5, 15 µm FMN (or FAD), 200 µm NADH and C1 (≈ 4–8 nm). Reactions were started by adding 200 µm HPA. Basal NADH oxidase activity was measured by performing the assay in the absence of HPA. Therefore, the C1-specific NADH oxidation was calculated by substracting the basal NADH oxidase activity from the total NADH oxidation. One unit of C1 activity is defined as the amount of enzyme required to oxidize 1 µmol NADH per min under the assay conditions. It should be noted that none of the free flavins in the reactions were reduced to a detectable extent.

    Enzymatic coupled assay with DHPAO

    A coupled enzyme assay was used to determine the activity of C2. DHPAO prepared according to the procedure of Arunachalam et al.[4] converts the product, DHPA, to 5-carboxymethyl-2-hydroxymuconate semialdehyde (CHS), which is a yellow compound with a λmax at 380 nm (ε380 = 38·mm−1·cm−1) [4]. With a saturating amount of DHPAO, the rate of formation of CHS represents the rate of formation of DHPA due to HPAH activities.

    A typical reaction (1 mL) for assaying C2 during enzyme purification contained 15 µm FMN (or FAD), 200 µm HPA, 200 µm NADH, ≈ 1–2 nm C2, excess C1 (≈ 5–10 nm) and excess DHPAO (at least three different concentrations were used for the assays: 100, 150, and 200 µg) in 50 mm sodium phosphate buffer pH 7.5 at 25 °C. Formation of CHS was followed at 400 nm (ε400 = 21.3 mm−1·cm−1). One unit of C2 activity is defined as the amount of the component required to form 1 µmol CHS per min under the above assay conditions.

    Spectroscopic studies

    UV-visible absorbance spectra were recorded with a Hewlett Packard diode array spectrophotometer (HP 8453A), or a Shimadzu 2501PC double-beam spectrophotometer. Fluorescence measurements were carried out with a Shimadzu RF5301PC spectrofluorometer. All spectrophotometers and spectrofluorometers were equipped with thermostated cell compartments. Graphite furnace atomic absorption analysis was performed with a PerkinElmer model 3100 instrument.

    Oxygen consumption and H2O2 measurement

    Measurement of oxygen consumption was performed in a closed reaction vessel (3 mL total volume) fitted with a Clark-type oxygen electrode and a magnetic stirrer (Yellow Springs Instruments, Model 53 oxygen monitor). The assay reaction contained 200 µm NADH, 7.5 µm FMN or FAD, C1 (fixed concentration at 10.9 nm) and various concentrations of C2 (0–1.39 µm) in 50 mm sodium phosphate buffer pH 7.5. The reaction was started by adding HPA to a final concentration of 200 µm to initiate oxygen consumption. At the end of the reaction (O2 consumed to completion), catalase was added to the reaction mixture to convert 50% of the resultant H2O2 to O2. The percentage of H2O2 produced during the reaction was calculated.

    Enzyme purification

    All of purification procedures were performed at 4 °C.

    Preparation of cell extracts and separation of the protein components

    Frozen cell paste (96 g wet weight) was thawed and suspended in 50 mm sodium phosphate buffer pH 7.5, containing 60 µm phenylmethanesulfonyl fluoride and 5 mm dithiothreitol. Cells were disrupted by a French Press (maintaining pressure at 16 000 p.s.i.) and then centrifuged at 100 000 g for 1 h. The pellet was discarded and the supernatant was defined as the crude extract. Nucleic acid materials were removed from the crude extract by protamine sulfate precipitation (9 mg·g cells−1).

    Hydroxyapatite chromatography was used as the initial column to separate the protein components (C1 and C2). Hydroxyapatite (≈ 300 mL) was packed in a column (4.3 × 23 cm) and equilibrated with 10 mm sodium phosphate buffer pH 6.8, 1 mm dithiothreitol. The crude extract was applied to the column and the unbound protein was washed with equilibration buffer (1 L). A gradient (2 L) of 10–150 mm sodium phosphate buffer pH 6.8 containing 1 mm dithiothreitol was used to elute the enzyme. Fractions containing C1 were detected by using HPA-stimulated oxidation of NADH at 340 nm and fractions containing C2 were detected by using the DHPAO-coupled enzyme assay as described above. The fractions containing each enzyme component were pooled, concentrated, and dialysed against 10 mm sodium phosphate buffer pH 6.8, 1 mm dithiothreitol.

    Further purification of C1

    DEAE-Sepharose column chromatography

    The dialysate of C1 was centrifuged to remove insoluble material and applied to a DEAE–Sepharose column (2.5 × 30 cm) previously equilibrated with 10 mm sodium phosphate buffer pH 6.8, 1 mm dithiothreitol. The column was then washed with 500 mL equilibration buffer. C1 was eluted with a gradient (1 L) of 50–250 mm NaCl in 10 mm sodium phosphate buffer pH 6.8, 1 mm dithiothreitol. Fractions containing C1 activity were pooled and precipitated with 80% ammonium sulfate solution.

    Phenyl-Sepharose chromatography

    A phenyl-Sepharose column (1.5 × 17 cm) was equilibrated with 5% ammonium sulfate in 10 mm sodium phosphate buffer pH 7.5, 1 mm dithiothreitol. The ammonium-sulfate-saturated C1 pellet was dissolved in a minimum volume of the equilibration buffer and centrifuged to remove insoluble materials. C1 was loaded on to the column and washed with equilibration buffer. The enzyme was eluted with a combined gradient (800 mL) of 5–0% ammonium sulfate and 0–40% ethylene glycol in 10 mm sodium phosphate buffer, pH 7.5, 1 mm dithiothreitol. The active fractions were pooled, concentrated and stored on ice.

    Further purification of C2

    DEAE-Sepharose column chromatography

    The dialysate of C2 after the hydroxyapatite column was centrifuged to remove insoluble materials and applied to a DEAE-Sepharose column (2.5 × 30 cm) previously equilibrated with 10 mm sodium phosphate buffer pH 6.8, 1 mm dithiothreitol. The enzyme was eluted with a gradient (1 L) of 0–300 mm NaCl in equilibration buffer. The active fractions were pooled and concentrated.

    G-150 gel permeation chromatography

    G-150 gel matrix was packed into a column (1 × 90 cm) and equilibrated with 150 mm NaCl in 50 mm sodium phosphate buffer pH 7.0, 1 mm dithiothreitol. The partially purified C2 was applied and eluted in fractions with the equilibration buffer. The purified C2 was concentrated and stored on ice.

    Molecular mass determination

    Molecular mass of the native enzyme

    The relative molecular mass of the native enzyme was determined by gel filtration chromatography [Superdex 200 HR 10/30 (Pharmacia), operating on an FPLC equipment (Bio-Rad)]. Elution was carried out at a flow rate of 0.5 mL·min−1 with 50 mm sodium phosphate buffer pH 7.0, containing 150 mm NaCl and 1 mm dithiothreitol. The standards used were cytochrome c (12.4 kDa), carbonic anhydrase (29 kDa), BSA (66 kDa), alcohol dehydrogenase (150 kDa) and amylase (200 kDa). Blue dextran (2000 kDa) was used to determine the void volume of the column (Vo). The molecular mass determination of the protein components was carried out by comparing the ratio of Ve/Vo (where Ve is the elution volume) to those of the protein standards. A standard calibration curve was constructed by plotting the logarithms of the known molecular mass of protein standards vs. their respective Ve/Vo values.

    Molecular mass of enzyme subunits

    The subunit molecular mass was determined by 12% SDS/PAGE (12% polyacrylamide). The protein molecular mass markers used were phosphorylase b (97 kDa), BSA (66 kDa), ovalbumin (45 kDa), carbonic anhydrase (31 kDa), soybean trypsin inhibitor (21.5 kDa), and lysozyme (14.4 kDa). Protein bands were visualized by staining the gel with Coomassie brillant blue.

    Anaerobic and photoreduction techniques

    Experiments with limited amounts of oxygen were performed in a special cuvette equipped with a gas inlet that can be connected to a purified argon gas train or vacuum pump [13]. Solution in the main cuvette body contained 3 mm EDTA, 106 µm HPA and 8 µm flavin in 50 mm sodium phosphate buffer pH 7.0, and the side arm of the cuvette contained 8.6 µm C2. The atmosphere in the cuvette was alternatively evacuated and flushed with argon gas. This procedure was repeated for 10–12 cycles. The solution in the cuvette under this condition contained ≈ 2–5 µm oxygen. The absorption spectrum of the oxidized flavin in the reaction mixture was recorded and the flavin was then photoreduced by exposing the cuvette to the high-intensity visible light from a 250-W tungsten lamp (Elmo HP-A290) for a brief period. The absence of the absorbance peaks around 380 and 450 nm indicated that the flavin was fully reduced. C2 in the side arm was then mixed with the reduced flavin and spectra of the reaction were recorded. The negative control experiments were also carried out as above but no C2 was added to the side arm.


    Enzyme activity in cell extracts

    HPAH activity in cell extracts of A. baumannii was analyzed by HPLC. Incubation of crude extract alone with HPA did not result in the formation of DHPA whereas the incubation of crude extract with HPA and NADH resulted in the decrease of HPA and the formation of DHPA. When the reaction mixture with NADH was supplemented with flavins (FMN, FAD or riboflavin), the amount of DHPA product increased (Fig. 1).

    Details are in the caption following the image

    HPAH activity in crude extract of A. baumannii. HPLC chromatogram showing the formation of DHPA and the consumption of HPA when 50 µL crude extract was incubated with 500 µm HPA, 2 mm NADH in 1 mL 50 mm sodium phosphate buffer pH 7.5 for 15 min (solid line). The increase in DHPA when 30 µm FMN was added into the same reaction is shown by the dotted line. DHPA was not formed during the incubation of the crude extract with HPA in the absence of NADH and FMN (dashed line).

    When the rates of NADH oxidation in these reactions were measured, the rate of NADH oxidation increased in the presence of HPA. With extra free flavins added, the rate of NADH oxidation in the presence of HPA was enhanced (data not shown).

    Enzyme purification: evidence for the two-protein components of HPAH

    It was found that the HPA-stimulated NADH oxidation activity lost its ability to hydroxylate HPA during the purification process. This suggested that HPAH might be composed of two components: C1 which alone has HPA-stimulated NADH oxidase activity, and C2 which can combine with C1 to hydroxylate HPA. C1 activity can be measured by the HPA-stimulated NADH oxidase assay and C2 activity can be monitored by using the coupled assay with DHPAO as described previously [4]. Results from enzyme purification indicated that HPAH indeed existed as a two-protein component enzyme. Combination of both components resulted in the formation of DHPA while neither C1 nor C2 alone can hydroxylate HPA (data not shown).

    Typical purifications of C1 and C2 are shown in Table 1 and Table 2. Both components are purified to homogeneity as judged by SDS/PAGE (Fig. 2). C1 has a subunit molecular mass of 32 kDa and C2 has a larger size with a subunit molecular mass of 50 kDa.

    Table 1. Summary of the purification of HPAH C1 from A. baumannii.
    protein (mg)
    activity (U)
    Specific activity
    Crude extract 1340 1560 1.16 1 100
    Protamine sulfate 1160 1610 1.40 1.2 104
    Hydroxyapatite 140 961 6.87 5.9 62
    DEAE–Sepharose 10.4 650 62.3 53.7 42
    Phenyl-Sepharose 2.8 566 201 173 36
    Table 2. Summary of the purification of HPAH C2 from A. baumannii.
    protein (mg)
    activity (U)
    Specific activity
    Crude extract 1860 385 0.21 1 100
    Protamine sulfate 991 338 0.34 1.7 88
    Hydroxyapatite 88.5 145 1.63 7.9 38
    DEAE–Sepharose 13.1 83.3 6.37 30.8 22
    G-150 4.36 38.7 8.89 43.0 10
    Details are in the caption following the image

    SDS/PAGE (12% acrylamide) of purified HPAH. Lanes A and D, molecular mass markers; lane B, C1 of HPAH; lane C, C2 of HPAH.

    Native molecular masses of the two components

    The native molecular mass of C1 was determined to be 73 kDa by the gel filtration method. Based on a subunit molecular mass of 32 kDa, C1 apparently exists as a dimeric enzyme with two identical subunits. Using the same method, the native molecular mass of C2 was estimated to be 209 kDa. With a subunit molecular mass of 50 kDa, C2 appears to be a homotetrameric enzyme.

    Cofactor of HPAH

    At the hydroxyapatite and DEAE–Sepharose chromatography steps of HPAH purification, C1 fractions had a yellow colour with absorption spectra in the visible region resembling those of flavin spectra (Fig. 3). After passage through a phenyl-Sepharose column in the last purification step, the purified C1 became colourless. It was speculated that C1 had a flavin bound as a prosthetic group but that it dissociated from C1 in the high ionic-strength condition (5% ammonium sulfate saturation) of the phenyl-Sepharose column. Therefore, the yellowish enzyme after DEAE-Sepharose chromatography was concentrated and acid-precipitated by 1% trichloroacetic acid. The yellowish supernatant was separated from the denatured protein pellet and analysed by TLC (silica gel 60 F254 Merck) using n-butanol/glacial acetic acid/water (4 : 5 : 1, v/v). FMN, FAD, and riboflavin were used as standards and migration of the compounds was monitored by the characteristic flavin fluorescence under UV light. TLC of the yellow supernatant gave a single fluorescence spot with an Rf of 0.38 identical to that of authentic FMN, whereas FAD has an Rf of 0.17 under the same conditions. The purified colourless C1 after phenyl-Sepharose chromatography was then reconstituted with FMN (Fig. 3). The FMN-reconstituted C1 shows an absorption spectrum (having λmax at 458 nm and a shoulder around 480 nm characteristic of some enzyme-bound flavin spectra) similar to the absorption spectrum of C1 prior to phenyl-Sepharose chromatography. The difference in absorption spectra around 320–390 nm between FMN-reconstituted C1 and the partially purified C1 after the DEAE–Sepharose column could be due to the chromophore of contaminant proteins in the partially purified C1.

    Details are in the caption following the image

    Absorption spectrum of C1. The spectrum of the partially purified C1 after DEAE–Sepharose chromatography is shown by the solid line. Overlaid is the spectrum of free FMN (dotted line). Purified C1 (after the phenyl-Sepharose column) was mixed with excess FMN and passed through a G-25 desalting column (FMN-reconstituted C1; solid line with circles).

    Purified C2 was colourless and no chromogenic cofactors were found in the UV-visible spectrum. A solution of C2 (32 µm) was analysed by graphite furnace atomic absorption spectroscopy. The results did not show any significant amount of Fe. This indicated that C2 did not contain the redox cofactors such Fe-sulfur centres or haem groups.

    Equilibrium binding of C1 to flavins

    The dissociation constants (Kd) for C1–flavin complexes were determined by using fluorescence spectroscopy. Upon titrating FMN or FAD solutions with the apoenzyme (C1), the flavin fluorescence decreased due to the binding of the free flavin to the enzyme. The resulting fluorescence changes were used to calculate Kd values for C1–FMN and C1–FAD complexes (Fig. 4). A Kd of 0.4 ± 0.1 µm was obtained for C1–FMN whereas FAD was less tightly bound to C1 with a Kd value of 1.3 ± 0.1 µm. This result agreed with the finding that FMN was the native cofactor of C1.

    Details are in the caption following the image

    Equilibrium binding of C1 to flavins. FMN (1 µm) in 50 mm sodium phosphate buffer, pH 7.0 was titrated with different concentrations of C1 apoenzyme. The concentration of apoenzyme used, from top to bottom, was 0, 0.192, 0.382, 0.572, 0.761, 0.949, 1.137, 1.320, 1.509, 1.694, 1.878, 2.060, 2.240, 2.426, 2.607 µm. Apoenzyme concentrations were determined by the Bradford method [12] using BSA as standard and the C1 activity assay. The titration was monitored by fluorescence spectroscopy with an excitation wavelength of 450 nm and emission wavelengths > 460 nm. The spectrum with the highest fluorescence intensity is that of free FMN. Increasing the concentration of C1 resulted in decreased fluorescence. The fluorescence changes (ΔF) were plotted against the added apoenzyme concentration to estimate the value of maximum fluorescence change (ΔFmax). The value of ΔFFmax was used to calculate the amount of C1–flavin complex in each titration. Concentrations of free apoenzyme were obtained by substracting the flavin-bound fraction from the total amount of apoenzyme. A similar titration experiment was also performed with FAD (3 µm) (data not shown). Inset, plot of fluorescence change against concentration of free apoenzyme used to calculate Kd values. Kd for C1-FMN, 0.4 ± 0.1 µm; Kd C1-FAD, 1.3 ± 0.1 µm.

    Dependence of HPAH activity on the molar ratio of C2/C1

    The hydroxylation activity of HPAH requires both C1 and C2. In the reaction of C1 alone with HPA and NADH, the oxidation of NADH resulted in the formation of H2O2 without hydroxylation of the aromatic substrate. Increasing the concentration of C2 in the reaction led to less H2O2 formed and more DHPA produced (Fig. 5). C2 did not influence the rate of NADH oxidation of C1 nor did it oxidize NADH by itself (data not shown). This suggested that C1-bound flavin was reduced by NADH and, in the presence of C2, the re-oxidation of reduced flavin was concomitant with the hydroxylation of HPA.

    Details are in the caption following the image

    Influence of the molar ratio of the two components of HPAH on hydroxylation activity and H2O2 formation. Initial rates in the DHPAO-coupled assay were measured while keeping the concentration of C1 constant at 10.9 nm and varying the concentration of C2 (0–1.39 µm) in reaction mixtures containing 150 µm NADH, 15 µm FAD and 200 µm HPA. Initial rates vs. C2/C1 ratio (○); percentage of H2O2 formed (●) as measured at the end of the reaction by adding catalase and measuring the production of oxygen using an oxygen electrode. The same experiment using FMN instead of FAD yielded similar results.

    Flavin specificity of HPAH

    Although FMN was found to be the native cofactor for C1, other flavins (FAD and riboflavin) could be added to the C1–C2 system and resulted in the hydroxylation of HPA. The rate of the NADH oxidation reaction for each flavin was similar when a saturating concentration of flavin was used. However, when the rates of hydroxylation were compared, FMN and FAD had a similar rate of hydroxylation which was significantly higher than the rate with riboflavin. (Activity measurements performed in reaction mixtures containing 200 µm NADH, 200 µm HPA, 136 µg of DHPAO, 4.6 nm C1, 20 nm C2, and 15 µm FMN, FAD or riboflavin in 50 mm sodium phosphate buffer, pH 7.5, showed that the rate of hydroxylation for FAD and riboflavin was 87% and 16%, respectively, when compared with that of FMN.) The slightly lower hydroxylation rate with FAD than that with FMN may be due to the higher Kd values for C1–FAD compared with C1–FMN. The flavin concentration in these reactions (15 µm) is equivalent to 30 × Kd C1–FMN whereas it is equivalent to ≈ 10 × Kd C1–FAD. If a higher amount of the C1–FAD complex could have been attained under the same conditions, the hydroxylation rate of C1–FAD might have been the same as that of C1–FMN. However, a higher amount of C1–FAD complex could not be obtained by adding more FAD as higher concentrations of FAD inhibited the reaction (see Kinetic parameters, below).

    Kinetic parameters of HPAH

    The apparent Km (inline image) of HPA, and of NADH for the HPAH reaction (with FMN or FAD as coenzymes) were investigated and the results are shown in Table 3. inline image values for HPA were similar in the reactions with FMN and FAD whereas inline image of NADH was about twofold higher with FAD than with FMN. When the concentration of flavin was varied up to 15 µm, the rate of the hydroxylation reaction increased with the concentration of FAD or FMN. When the reaction was investigated at higher concentrations of flavins, both FMN and FAD at > 15 µm showed an inhibitory effect on the reaction (Fig. 6). The absorption at 450 nm of all reactions was constant throughout the course of the assay, indicating that no detectable amount of the free flavin in the assay solution was in the reduced form.

    Table 3. Kinetic parameters of the HPAH C1–C2 system.inline image values for HPA were determined in reaction mixtures containing 200 µm NADH, 15 µm FMN (or FAD), 2.9 nm C1, 288 nm C2, and 136 µg DHPAO. inline image values for NADH were determined under the same conditions except that 200 µm HPA was used.
    Substrate inline imagem)
    with FMN with FAD
    NADH 12 ± 1 28 ± 4
    HPA 19 ± 3 14 ± 3
    Details are in the caption following the image

    Inhibition of the hydroxylation reaction by FAD and FMN at high concentration. Assays contained 2.9 nm C1, 288 nm C2, 136 µg DHPAO, 200 µm HPA, 200 µm NADH, and various concentrations of FAD and FMN in 1 mL 50 mm sodium phosphate buffer pH 7.5. Initial rates of the hydroxylation reaction were plotted against the concentration of FMN (●) and FAD (○). The rate of the reaction decreased when the concentration of flavin was > 15 µm.

    Substrate specificity

    A variety of aromatic compounds had a stimulatory effect on NADH oxidation (Table 4), indicating that these compounds could bind to the active site of C1. Only the compounds with a hydroxyl group in the para-position could be hydroxylated by HPAH. This indicated that the hydroxyl group of the phenolic substrate was essential for the hydroxylation. For the single-component aromatic flavoprotein hydroxylase it is known that that the hydroxyl group of the phenolic substrate is necessary for hydroxylation by an electrophilic substitution mechanism [14].

    Table 4. Substrate specificity of HPAH from A. baumannii. NADH oxidase activity was measured in a reaction mixture containing 200 µm NADH, 15 µm FMN, 3 nm C1, and 200 µm substrate analogue in 50 mm sodium phosphate pH 7.0. A relative activity of 100% corresponded to 201 U·mg−1 when HPA was used. Product conversion is the difference between the initial amount of substrate analogue and the amount of substrate analogue remaining after a 16 h incubation. The result is represented by (–) when no significant amount of substrate was consumed. HPLC was used to estimate the amount of the substrate analogue remaining in the reaction. Reactions for HPLC analysis contained 250 µm NADH, 15 µm FMN, 200 µm substrate or substrate analogue, 15 nm C1 and 640 nm C2. The same reactions without C2 were also carried out as a negative controls and the results showed that none of the substrates were consumed in the absence of C2.
    Substrate or substrate analogue Relative NADH
    oxidase activity (%)
    conversion (%)
    No substrate 1.7
    p-Hydroxyphenylacetate (HPA) 100 100
    m-Hydroxyphenylacetate 97.7
    o-Hydroxyphenylacetate 4.8
    p-Chlorophenylacetate 10.3
    p-Fluorophenylacetate 37.6
    p-Nitrophenylacetate 2.0
    4-Hydroxy-3-methoxyphenylacetate 45.4
    Phenylacetate 96.5
    3-(4-Hydroxyphenyl)propionate 35.9 65
    p-Hydroxybenzoate 2.6 85
    p-Nitrophenol 2.6 11

    Evidence for C2 as the hydroxylating component

    To elucidate the role of C2 in this HPAH system, HPA and reduced FMN were mixed with C2 under oxygen-limited conditions in an anaerobic cuvette. The reaction showed the appearance of an intermediate species with λmax at 370 nm (Fig. 7A and the upper inset of Fig. 7A). The absorption characteristics of this species resembled those of a C(4a)-substituted flavin [i.e. C(4a)-hydroperoxyflavin or C(4a)-hydroxyflavin [15–17]] and were distinctly different from those of oxidized and reduced flavin. The negative control experiment was carried out by mixing reduced FMN prepared by the same method with buffer instead of C2 and resulted in the oxidation of some reduced flavin (indicated by the appearance of absorbance peaks with λmax at 450 and 380 nm; data not shown). When reduced FMN was mixed with C2 in the absence of HPA, some of the intermediate absorption was generated, like that seen in the presence of HPA but with more of the oxidized flavin character (Fig. 7B). This indicated that HPA stabilized the formation of the intermediate. The reaction in Fig. 7A (with HPA) was completely open to the air and was analysed for DHPA formation by HPLC and the DHPAO assay. The results showed the formation of DHPA by both methods of analysis (the lower inset to Fig. 7A) indicating clearly that C2 alone can hydroxylate HPA by using reduced FMN. Similar experiments were performed by using FAD and riboflavin instead of FMN. DHPA and the intermediate spectrum were also produced with reduced FAD and riboflavin but to a slightly lesser extent than when FMN was used (data not shown).

    Details are in the caption following the image

    Reaction of C2 with reduced FMN and a limited amount of oxygen. (A) The UV-visible absorption spectrum of oxidized FMN (8 µm) and HPA (106 µm) in 50 mm sodium phosphate buffer (1 mL) pH 7.0 in an anaerobic cuvette is shown by a dashed line (1). The fully reduced FMN spectrum is shown by a solid line (2). The ‘intermediate’ spectrum obtained after the C2 in the side-arm of the cuvette was tipped into the main body is shown by (3). The small amount of oxygen used to generate this intermediate is the residual amount of oxygen left in the cuvette body after the anaerobic procedure. The spectrum obtained when the cuvette was fully opened to air is shown by (4). Because part of spectrum (3) resulted from oxidized flavin, spectral correction was carried out to obtain the more evident ‘intermediate’ spectrum shown in the upper inset. It was approximated that 17% of oxidized flavin was presented in (3) based on the absorbance at 450 nm. The lower inset shows an HPLC analysis of the reaction, clearly indicating the formation of DHPA. (B) Spectra (1)–(4) are the same as in (A) except that no HPA was added to the reaction. Spectrum (3) in this case shows more oxidized enzyme characteristic when compared with the reaction with HPA present in (A). The inset shows the calculated spectrum of the ‘intermediate’ obtained by substracting the spectral contribution of the oxidized flavin. It was approximated that (3) contained 30% of oxidized flavin.


    We have described here the purification and the catalytic properties of the two-component enzyme p-hydroxyphenylacetate hydroxylase (HPAH) from A. baumannii. A comparison of the biochemical properties of the enzyme with other two-component enzymes hydroxylating aromatic substrates is given in Table 5. The enzymes listed are those that use flavins as cofactors and whose enzymatic properties have been investigated previously [4–6,18–20]. It is shown clearly that our HPAH has properties different from those of the HPAH isolated from E. coli and from P. putida and also is distinct from other two-component aromatic hydroxylases with respect to flavin specificity. C1 of HPAH from A. baumannii and the small component (hpaC) of HPAH isolated from E. coli[5] can use either FMN, FAD, or riboflavin for NADH oxidase activity. However, only C1 from A. baumannii binds to flavins and hpaC from E. coli does not bind to FMN, FAD or riboflavin [5]. The small component of HPAH from P. putida is an FAD-dependent enzyme; neither FMN nor riboflavin can substitute for FAD [4]. The small component of 4-chlorophenol hydroxylase and pyrrole-2-carboxylate monooxygenase use only FAD as the coenzyme for their NADH oxidase activity [6,19].

    Table 5. Biochemical properties of two-component aromatic flavoprotein hydroxylases.
    Enzyme Small component
    (native cofactor)
    Flavin specificity
    of small component
    Pyridine nucleotide
    Large component
    Flavin specificity of
    large component
    (A. baumannii)
    Dimer of 32 kDa
    subunit (FMN)
    FMN, FAD,
    NADH Tetramer of 50 kDa
    subunit (hydroxylase)
    FMN, FAD,
    HPAH Dimer of 30.75 kDa FAD only NADH Monomer of 38.5 kDa FAD is tightly
    (P. putida) [4] subunit (FAD) subunit (coupling protein) associated with
    small component
    HPAH Dimer of 19 kDa FMN, FAD, NADH, NADPH Dimer of 59 kDa FAD only
    (E. coli) [5,18] subunit (none) riboflavin subunit (hydroxylase)
    Chlorophenol-4- Monomer of 22 kDa FAD only NADH Monomer of 58 kDa FAD only
    hydroxylase (FAD) (hydroxylase)
    (B. cepacia) [19] Monomer of 19 kDa
    Pyrrole-2-carboxylate (FAD) FAD only NADH Trimer of 54 kDa subunit FAD only
    monooxygenase (monooxygenase)
    (Rhodococcus sp.) [6]
    p-Nitrophenol (FAD) FAD only NADH, NADPH FAD only
    (B. sphaericus) [20]

    The flavin specificity of hydroxylation by HPAH from A. baumannii is also different from other aromatic flavoprotein hydroxylases. All of the two-component (Table 5) and single-component flavoenzymes that hydroxylate aromatic compounds [2,3] use FAD as coenzymes whereas A. baumannii HPAH uses either FMN, FAD, or riboflavin for hydroxylase activity. Our data indicate that Acinetobacter HPAH preferred FMN over FAD and riboflavin for the hydroxylation reaction implying that the flavin binding site on C2 preferred FMN more than the other cofenzymes. It is possible that the active site of C2 contains a positive charge to stabilize the binding of the negatively charged FMN phosphate group but it cannot bind well to the bulky adenine part of FAD. This result and the finding of FMN as the native cofactor of C1 suggest that FMN is the coenzyme of two-component A. baumannii HPAH in vivo.

    The substrate HPA plays a dual role in the HPAH C1–C2 system from A. baumannii: it is the effector for C1 as well as the substrate for C2. In the reaction of C1 alone, HPA binds to C1 and stimulates NADH oxidase activity. With the presence of both C1 and C2, HPA is used as a real substrate leading to HPA hydroxylation, and wasteful utilization of NADH is avoided. Therefore, the presence or absence of HPA is the means of controlling NADH oxidation activity of C1 and the presence of C2 ensures that hydroxylation occurs. The role of the substrate in the HPAH C1–C2 system is similar to the role of the substrate in single-component aromatic flavoprotein hydroxylases [2,21]. Substrate-stimulated NADH oxidation is found in the reductive half-reaction of aromatic flavoprotein hydroxylases such as p-hydroxybenzoate hydroxylase [22,23], salicylate hydroxylase [24,25], anthranilate hydroxylase [26], melilotate hydroxylase [27], phenol hydroxylase [28], and 2-methyl-3-hydroxypyridine-5-carboxylic acid oxygenase [29].

    The ability of aromatic substrates to stimulate NADH oxidation varies among the two-component hydroxylase enzymes. The small component of P. putida HPAH [4] has HPA-stimulated NADH oxidase activity similar to that of the A. baumannii HPAH in this study, whereas NADH oxidation by hpaC of E. coli HPAH is not influenced by the presence of HPA [5]. HPA-stimulated NADH oxidation implies that the small component of Acinetobacter and Pseudomonas enzymes bind the aromatic substrate HPA whereas the E. coli enzyme apparently does not. The size of both Acinetobacter and Pseudomonas enzymes (≈ 31 kDa) is larger than that of the E. coli enzyme (19 kDa). This difference of ≈ 12 kDa of polypeptide chain could correspond to the binding site of HPA.

    Our results indicated that C2 alone was sufficient for HPA hydroxylation if reduced flavin was provided (Fig. 7). This observation suggests that C1 is the reductase component which supplies reduced flavin to C2, and that C2 is the hydroxylase component which hydroxylates the aromatic substrate by using reduced flavin and oxygen. These catalytic properties of A. baumannii HPAH are significantly different from those of P. putida HPAH [4]. In the Pseudomonas enzyme, hydroxylation occurs by an FAD tightly bound to the small component (equivalent to C1). The role of the large component (equivalent to C2) is as a ‘coupling protein’ to enable hydroxylation rather than the formation of H2O2. However, this catalytic property of Acinetobacter HPAH is similar to that of E. coli HPAH. It was shown that the large component (hpaB) of E. coli HPAH could hydroxylate HPA when reduced FAD was steadily provided to the system by using either hpaC or another flavin reductase enzyme [5,10].

    The formation of the intermediate resembling a C(4a)-substituted flavin provides useful insight into the mechanism of this two-component enzyme. Although the intermediate spectrum observed in our experiments did not allow us to distinguish the type of species, it clearly indicated that the mechanism for the hydroxylation reaction proceeded via the formation of the C(4a)-hydroperoxy flavin and the C(4a)-hydroxy flavin. The existence of the C(4a)-hydroperoxy flavin and the C(4a)-hydroxy flavin is well characterized in flavoprotein monooxygenases, both in the single-component aromatic flavoprotein hydroxylases and in the two-protein bacterial luciferase system [2,3]. In the single-component aromatic flavoprotein hydroxylase reaction, C(4a)-substituted flavin intermediates were detected during the oxidative half-reaction of the enzyme-bound FAD and it was shown that C(4a)-hydroperoxy flavin was the reactive intermediate that hydroxylated aromatic compounds [2,15,16]. In bacterial luciferase, a flavin reductase provides reduced FMN to luciferase, and C(4a)-substituted flavin intermediates were found during the oxidative half-reaction [30,31].

    The role of the Acinetobacter HPAH C1–C2 system as reductase/hydroxylase enzymes suggests that, despite the probable similarity in chemical mechanism, the mode of hydroxylation is quite different from the kinetic mechanisms of single-component aromatic flavoprotein hydroxylases and P. putida HPAH, but is rather similar to the kinetic mechanism of bacterial luciferase. Bacterial luciferase generally catalyses the conversion of long-chain aliphatic aldehydes to fatty acid products with concomitant light production [32,33]. There is no evidence to demonstrate the hydroxylation of aromatic substrates by bacterial luciferase.

    Recently, studies of E. coli HPAH have shown a similarity between the catalytic reaction of E. coli HPAH and bacterial luciferase but the hydroxylation of E. coli HPAH specifically requires FADH2[5,10]. Our studies here have shown that A. baumannii HPAH could use FMNH2, FADH2, or reduced riboflavin with a luciferase-like mechanism to catalyse hydroxylation of aromatic compounds. To our knowledge, this is the first example of a hydroxylase that uses various flavins to catalyse the hydroxylation of aromatic compounds; such reactions are usually catalyzed by FAD-containing aromatic hydroxylases.

    Our investigation on the biochemical and catalytic properties of HPAH C1–C2 suggests that the catalytic cycle can be depicted as in Scheme 1. The reaction starts with the binding of HPA to C1, followed by the reduction of the enzyme-bound flavin by NADH. If no C2 is present, oxygen will oxidize the reduced flavin and generate H2O2 as the final product. In the presence of C2, the reduced flavin generated by C1 is transferred to C2 and the reoxidation of flavin occurs concurrently with the hydroxylation of the substrate. Our data cannot distinguish between transfer of the reduced flavin within a C1–C2 complex or its release into solution by C1 followed by binding to C2.

    Details are in the caption following the image

    Proposed scheme for catalysis of Acinetobacter baumannii HPAH.

    Another interesting finding is that a high concentration of flavin (> 15 µm for both FAD and FMN) significantly inhibited HPA hydroxylation activity of the HPAH C1–C2 system but does not inhibit NADH oxidation activity of C1. Therefore, high flavin concentrations probably inhibit C2. This excess oxidized flavin might compete with the reduced flavin for the flavin binding site on C2, leading to an abortive complex. Inhibition by high flavin concentrations was also found in the HPAH from E. coli W [10] and the luciferase enzyme from Vibrio harveyi[34]. In luciferase, the luminescent reaction of the enzyme was inhibited by FMN concentrations > 2 µm. It was proposed that the large amount of oxidized flavin in the reaction competed with the reduced flavin to bind at the flavin binding site on the luciferase component [34].

    HPAH isolated from A. baumannii is the first reported aromatic flavoprotein hydroxylase that utilizes reduced FMN. To our knowledge, all of the FMNH2-utilizing flavoprotein monooxygenases reported so far oxygenate aliphatic compounds such as aldehydes (luciferase [31,35]), EDTA (an enzyme from bacterial strain DSM 9103 [36]), nitrilotriacetate (nitrilotriacetate monooxygenase [37]), alkanesulfonate (alkanesulfonate monooxygenase [38]) whereas the FAD monooxygenases hydroxylate aromatic compounds such as p-hydroxybenzoate hydroxylase [22], melilotate hydroxylase [27], phenol hydroxylase [28], 2-methyl-3-hydroxypyridine-5-carboxylic acid oxygenase [29], chlorophenol-4-monooxygenase [19], pyrrole-2-carboxylate monooxygenase [6] and HPAH isolated from P. putida[4] or E. coli W [5,18]. Our finding suggests that this A. baumannii HPAH enzyme is a ‘hybrid’ that posseses properties of both classes: it carries out the reaction normally found with FAD-associated aromatic hydroxylases by using mechanistic features similar to those of the two-component luciferase enzyme. However, the enzyme still has some ‘control’ properties, such as substrate-stimulated NADH oxidation, that are similar to those of the aromatic flavoprotein hydroxylases. Why nature has evolved such a system to catalyse the hydroxylation of HPA awaits further investigation.


    This work was supported by The Thailand Research Fund, grant PDF 2541/20 (to P. C.) and by the Faculty of Science, Mahidol Univerisity. P. W. is a Senior Research Scholar of The Thailand Research Fund, and C. S. was a recipient of a scholarship of the Institutional Strengthening Program offered by the Faculty of Science, Mahidol University and National Science and Technology Development Agency, Thailand. We thank B. A. Palfey and M. R. Jisnuson Svasti for their valuable suggestions during the preparation of the manuscript. We are grateful to M. R. Jisnuson Svasti for the use of HPLC equipment and V. Meevoothisom and V. Khampha for the use of the FPLC equipment. We thank J. Shiowatana and O. Kitjabuncha for assistance with the atomic absorption analysis. The assistance of J. Sudcharitkul and K. Thotsaporn during enzyme preparation is highly appreciated.