Phenylalanine meta‐Hydroxylase: A Single Residue Mediates Mechanistic Control of Aromatic Amino Acid Hydroxylation

Abstract The rare nonproteinogenic amino acid, meta‐l‐tyrosine is biosynthetically intriguing. Whilst the biogenesis of tyrosine from phenylalanine is well characterised, the mechanistic basis for meta‐hydroxylation is unknown. Herein, we report the analysis of 3‐hydroxylase (Phe3H) from Streptomyces coeruleorubidus. Insights from kinetic analyses of the wild‐type enzyme and key mutants as well as of the biocatalytic conversion of synthetic isotopically labelled substrates and fluorinated substrate analogues advance understanding of the process by which meta‐hydroxylation is mediated, revealing T202 to play an important role. In the case of the WT enzyme, a deuterium label at the 3‐position is lost, whereas in in the T202A mutant 75 % retention is observed, with loss of stereospecificity. These data suggest that one of two possible mechanisms is at play; direct, enzyme‐catalysed deprotonation following electrophilic aromatic substitution or stereospecific loss of one proton after a 1,2‐hydride shift. Furthermore, our kinetic parameters for Phe3H show efficient regiospecific generation of meta‐l‐tyrosine from phenylalanine and demonstrate the enzyme's ability to regiospecifically hydroxylate unnatural fluorinated substrates.


Heterologous production and purification of PhhA and PhhA single point mutants
The phhA gene (encoding Phe3H) was PCR-amplified from Streptomyces coeruleorubidus AB1183F64 using  using E. coli BL21 (DE3) as the host strain as described for above.

Optimisation of Phe3H assay conditions
Prior to measuring the Michaelis-Menten kinetics of Phe3H, we sought to optimise the assay conditions and identify potential inhibitors of the enzyme. Phe3H was expressed heterologously in E. coli and purified by Niaffinity and size exclusion chromatography ( Figure S1). The resultant enzyme was screened for activity in MES, HEPES and Tris buffers across their pH ranges. In contrast to previously reported data, 1 only baseline levels of activity were detected below pH 7, with the optimal tested conditions being HEPES buffer at pH 8. Swapping to Tris buffer at pH 8 resulted in a 2.5-fold decrease in enzyme activity ( Figure S2). This result has also been reported for rat and bovine PhhA, in which it was proposed that Tris may coordinate directly to the active ferric iron centre, thereby decreasing the rate of Fe(III) reduction. 2,3 The influence of additives such as ascorbate, excess ferrous iron, catalase and superoxide dismutase (SOD) was also investigated. SOD and catalase were found to rescue the activity of Phe3H in the presence of excess Fe(II) and protect the tetrahydropterin cofactor from non-enzymatic oxidation, further improving the efficiency of this reaction ( Figures S3-S5). Ascorbate and dithiothreitol (DTT) were also found to slow down the non-enzymatic oxidation of the pterin cofactor. As DTT can scavenge Fe(II) from the active site of the enzyme, however, it was omitted from the enzymatic assays. 3-5
Phe3H reaction mixtures for kinetic analysis were carried out at least in triplicate and contained 0.5 µM Phe3H, 0.5 mM 6MPH 4 or DMPH 4 , 1 U µL -1 catalase, 0.01 U µL -1 superoxide dismutase and varying amounts of L-Phe Tyr was used as standard, whereas for quantitation of fluorinated mTyr, compounds were purified from enzyme reactions and quantified by qNMR using maleic acid as standard.

Kinetic analyses and inhibitor studies 1
Non-linear regression was performed using Prism software. For steady-state kinetics the data points (4 -6 Eq. 1 where v 0 is the initial rate (µM product min -1 ), V max is the maximal reaction velocity, K M is the Michaelis-Menten constant. The rate constant k cat was calculated as V max / [Phe3H].
The associated error in k cat and K M values were calculated using the following equation 6,7 : For steady state kinetics with substrate inhibition, data were fitted to * * Eq. 3 where K i is the inhibitor binding constant for reversible inhibitors.
For inhibition by DL-3F-Phe, the measured rates were presented as Lineweaver-Burk plots for each inhibitor concentration. Slopes were determined by linear regression and plotted against the inhibitor concentration to give the secondary plot. Linear regression was used to determine the X axis intercept x y=0 , with the IC 50 = -x y=0 .

Non-enzymatic oxidation of pterin co-factor
Reactions were conducted in duplicate at 28 °C in Nunc 96-well plates with a 200 µL reaction volume. The oxidation of DMPH 4 to DMPH 2 was monitored by recording the increase in absorbance at 340 nm using a FLUOstar Omega plate reader (BMG Labtech). All reactions contained 50 mM HEPES, pH 7.5 and were started by the addition of 0.2 mM DMPH 4 . The following additives were tested alone or in combination: 1 mM DTT, 1 mM ascorbate, 10 µM ferrous ammonium sulfate/100 µM DTT or 5 µM ferrous ammonium sulfate/50 µM DTT, 0.05 U µL -1 superoxide dismutase, 5 U µL -1 catalase. DMPH 2 was generated from spontaneous oxidation of DMPH 4 in HEPES buffer, its identity was verified by UV spectroscopy. Results are shown in Figure S4.
The solution was then cooled to 0 °C and sodium nitrite (5.15 g, 74.6 mmol, 1.2 eq., in 25 mL distilled water) was added dropwise over 90 minutes whilst keeping the temperature of the reaction mixture below 4 °C. During the course of the addition, the mixture changed first to a pink/red solution before turning into a yellow-brown

[3,5-d 2 ]-Benzaldehyde
The procedure was adapted from Zhao. 8 A solution of cerium(IV) ammonium nitrate (35 g, 64 mmol, 2 eq.) in 3.5 M nitric acid (130 mL) was added to [3,5-d 2 ]-toluene (3.0 g, 32 mmol, 1 eq.) and the orange solution was heated to 80 °C for 2 hours. The yellow solution was cooled to room temperature, then diethylether (100 mL) was added and the organic phase was washed with distilled water until the aqueous layer was pH 7. The organic layer was dried (MgSO 4 ) and the solvent carefully removed in vacuo. This procedure was repeated and the crude benzaldehyde extracts were combined. Distillation afforded the title compound as a yellow liquid that contained 0.5 eq. starting material.

[3,5-d 2 ]-Phenylpyruvic acid
The procedure was adapted from patent literature. 9,10 In two 1 mL reaction vials was assembled each: hydantoin The suspension from both reaction vials was combined in a 50 mL round-bottom flask, the vials rinsed with MW-water (1 mL) and an aqueous NaOH solution (5 mL of a 5 N solution, 25 mmol NaOH, 7 eq.) was added.
The reaction was heated to reflux under nitrogen atmosphere for 90 minutes The resulting solution was cooled to 20 °C and the pH carefully adjusted to 8.5 with concentrated aqueous HCl solution. NaCl (1.5 g) was added to aid precipitation and the mixture was left standing overnight. The solids were collected in a scintered funnel, washed with ice cold distilled water (3 mL) and methanol (3 mL 168.0983 (-3.0 ppm, 100.0%).