The active site of the thermophilic CYP119 from Sulfolobus solfataricus.

CYP119 from Sulfolobus solfataricus, the first thermophilic cytochrome P450, is stable at up to 85 degrees C. UV-visible and resonance Raman show the enzyme is in the low spin state and only modestly shifts to the high spin state at higher temperatures. Styrene only causes a small spin state shift, but T(1) NMR studies confirm that styrene is bound in the active site. CYP119 catalyzes the H(2)O(2)-dependent epoxidation of styrene, cis-beta-methylstyrene, and cis-stilbene with retention of stereochemistry. This catalytic activity is stable to preincubation at 80 degrees C for 90 min. Site-specific mutagenesis shows that Thr-213 is catalytically important and Thr-214 helps to control the iron spin state. Topological analysis by reaction with aryldiazenes shows that Thr-213 lies above pyrrole rings A and B and is close to the iron atom, whereas Thr-214 is some distance away. CYP119 is very slowly reduced by putidaredoxin and putidaredoxin reductase, but these proteins support catalytic turnover of the Thr-214 mutants. Protein melting curves indicate that the thermal stability of CYP119 does not depend on the iron spin state or the active site architecture defined by the threonine residues. Independence of thermal stability from active site structural factors should facilitate the engineering of novel thermostable catalysts.

The cytochrome P450 superfamily of enzymes catalyzes a wide range of oxidative and reductive reactions that are important in xenobiotic detoxification, carcinogen activation, steroid biosynthesis, fatty acid metabolism, and reactions required for the survival of microorganisms on selected nutrients (1). These reactions include hydrocarbon hydroxylation, olefin epoxidation, and heteroatom oxidation (2). The ability of P450 1 enzymes to catalyze a diversity of transformations, combined with the ability of their active sites to bind a range of unrelated structures, makes these enzymes highly versatile catalysts. CYP119, as the first thermophilic member of the P450 superfamily to be identified and cloned (3), is of great interest as a starting point for the construction of useful thermostable P450 catalysts through protein engineering methods.
The CYP119 gene was cloned out of Sulfolobus solfataricus, a hyperthermophilic, acidophilic archaebacteria found in sulfurous volcanic hot springs (3). Sulfolobus are sulfur autotrophs that can be found attached to sulfur crystals at Pisciarelli Springs in Italy and in mud pools at Yellowstone National Park. Optimal growth conditions for Sulfolobus occur at temperatures between 78 and 86°C and pH values between 3 and 4 (4). The growing interest in thermophilic proteins has led to an ongoing effort to sequence the complete Sulfolobus genome (5)(6)(7).
Several recognizable structural features are evident from the CYP119 primary sequence. First, CYP119 lacks a hydrophobic tail and is therefore expected, as found, to be a soluble protein.
Second, CYP119 has a high identity in the heme binding region with mammalian, fungal, and bacterial P450 enzymes and exhibits a particularly high overall identity of 33% with P450 eryF . Last, CYP119 possesses both the conserved cysteine involved in heme coordination and a highly conserved threonine within the I helix thought to facilitate the coordination of a water ligand to the ferric heme, the binding of oxygen to the ferrous prosthetic group, and cleavage of the bound dioxygen molecule to give the final activated oxidizing species (8,9). In P450 cam , the conserved threonine is Thr-252. This residue is crucial for catalysis, and its mutation to an alanine or valine impairs substrate binding and greatly enhances the uncoupled reduction of oxygen to H 2 O 2 and H 2 O (8). In addition, mutation of the conserved threonine to a large hydrophobic residue can increase the fraction of the protein in the high spin state due to displacement of the water ligand by the amino acid side chain (8 -10).
Conversion of P450 enzymes from the low to the high spin state, which correlates with loss of the distal water ligand, can be indirectly measured by monitoring the shift in the Soret absorbance maximum from 415 to 390 nm. The P450 spin state is temperature-, pH-, and ionic strength-dependent (11)(12)(13). It is worth noting that the conserved threonine (Thr-213) in CYP119 is followed by two additional threonine residues (Thr-214 and Thr-215). Such a sequence of three consecutive threonines is not unique within the P450 family, but is unique among the well studied bacterial P450 enzymes (14).
We report here expression and characterization of CYP119, including the first determination of its catalytic properties. We also describe the construction, expression, and characterization of its Thr-213 and Thr-214 mutants. These studies show that the threonine residues in CYP119 play an important role in controlling the active site structure, spin state, and catalytic activity of the enzyme. They are not, however, critical for its thermal stability. An abstract describing some of our work on the basic properties of CYP119 has been published (15), as has a communication on the same enzyme by another laboratory (16).

EXPERIMENTAL PROCEDURES
Materials-Styrene, cis-stilbene, R-and S-styrene oxide, thiophenol, 4-ethylthiophenol, and other chemicals were purchased from Aldrich. cis-␤-Methylstyrene was from K & K Laboratories (Cleveland, OH), and methylphenyldiazene carboxylate azo ester was from Research Organics (Cleveland, OH). Methyl or ethyl aryldiazene carboxylate azo esters were prepared from the corresponding nitro, amino, and hydrazino precursors according to the method of Huang and Kosower (17).
General Procedures-UV-visible spectra were recorded on a Cary 1E double beam spectrophotometer. Gas chromatography was performed on an HP5890 Series II chromatograph with a flame ionization detector. CD measurements were performed on a Jasco J-710 instrument fitted with a Peltier water bath. 1 H spectra were recorded on a Varian Unity Inova AS400 NMR spectrometer.
Expression and Purification of CYP119 -The CYP119 cDNA was a kind gift from Peter Kennelly (Virginia Tech). For overexpression, CYP119 was subcloned into pCWori using the NdeI and XbaI sites. An overnight culture (3 ml) of a clone picked from a single colony in 2YT medium with ampicillin was used to inoculate 1.5 liters of 2YT medium. Protein expression in Escherichia coli DH5␣ cells was induced using 1 mM isopropyl-1-thio-␤-D-galactopyranoside after the cell density reached an OD of 0.8. The incubation temperature was then lowered to 30°C, and the cells were grown for 40 h. Cells were harvested and lysed using lysozyme in 50 mM Tris buffer, pH 7.0, at 4°C. After sonication and centrifugation, the supernatant was heated to 60°C for 1 h. Precipitated proteins were pelleted (20 min at 15,000 rpm), and the supernatant was loaded onto a Q Sepharose column; washed with 50 mM Tris, pH 7.0, buffer; and eluted using a gradient of 0 -250 mM NaCl. All of the reddish fractions were combined and dialyzed overnight against 50 mM bis-Tris, pH 7.0. The protein was then loaded onto a PBE94 column equilibrated in 50 mM bis-Tris, pH 7.0, buffer and eluted using Polybuffer 74 at pH 5.0. CYP119 eluted at pH 6.6 and was judged to be pure by SDS-polyacrylamide gel electrophoresis. The polybuffer was removed by ammonium sulfate precipitation (80%), and the protein was stored at Ϫ20°C in 50 mM bis-Tris, pH 6.0, buffer containing 5% glycerol. Protein concentration was determined by reduced CO difference spectra using excess dithionite and ⑀ 450 ϭ 100 mM -1 cm Ϫ1 .
Construction of CYP119 Mutants-The CYP119 Thr-213 and Thr-214 mutants were constructed using the Stratagene QuikChange mutagenesis kit. Briefly, the forward and reverse oligonucleotides containing the desired mutation were used to amplify CYP119-pCWori using the polymerase chain reaction. DpnI was then used to digest the parental DNA. Transformation of the remaining DNA resulted in colonies that were screened by unique restriction sites and DNA sequencing. By this method, Thr-213 was mutated to Ala, Val, Ser, Trp, and Phe, and Thr-214 was mutated to Ala and Val. One additional double T213A/ T214A mutant was also constructed. All mutants were expressed and purified in the same manner as wild-type CYP119 in the pCWori vector. The mutagenic oligonucleotides were from the University of California, San Francisco Biomolecular Resource Center and Life Technologies, Inc. For the T213S mutant, a SalI site was introduced (bold face indicates the positions of the mutations, and underlines indicate new restriction sites), 5Ј-CTT CTC ATA GCG GGT AAT GAG TCG ACA ACT AAC TTA ATA TCA AAC-3Ј. An EcoRI site was introduced into 5Ј-CTT CTC ATA GCG GGT AAT GAA TTC ACA ACT AAC TTA ATA TCA AAC-3Ј for the T213F mutant. New NaeI sites were introduced into 5Ј-ATT TTA CTT CTC ATA GCC GGC AAT GAG GTT ACA ACT  AAC TTA ATA TC-3Ј for T213V, 5Ј-ATT TTA CTT CTC ATA GCC GGC  AAT GAG GCT ACA ACT AAC TTA ATA TC-3Ј for T213A, 5Ј-ATT TTA  CTT CTC ATA GCC GGC AAT GAG TGG ACA ACT AAC TTA ATA  TC-3Ј for T213W, 5Ј-ATT TTA CTT CTC ATA GCC GGC AAT GAG ACT  GCA ACT AAC TTA ATA TC-3Ј for T214A, 5Ј-ATT TTA CTT CTC ATA  GCC GGC AAT GAG ACT GTA ACT AAC TTA ATA TC-3Ј for T214V,  and 5Ј-ATT TTA CTT CTC ATA GCC GGC AAT GAG GCT GCA ACT  AAC TTA ATA TC-3Ј for the double T213A/T214A mutants. Styrene Epoxidation by CYP119 -Peroxide-dependent epoxidation assays were carried out with styrene, cis-␤-methylstyrene, and cisstilbene in 50 mM bis-Tris buffer, pH 6.0. Reaction mixtures (total volume 200 l) contained CYP119 (12.5 M), styrene (saturated), and H 2 O 2 (5 mM). Styrene, cis-␤-methylstyrene, or cis-stilbene was added (3 l) as a 0.5 M solution in acetonitrile to give a final acetonitrile concentration of 1.5%. Control incubations were done without CYP119 or without H 2 O 2 . Each reaction was allowed to proceed for 30 min at 30°C. Samples were then extracted with CH 2 Cl 2 (2 ϫ 150 l). The combined CH 2 Cl 2 layers were concentrated under a stream of argon and then analyzed by isothermal gas chromatography at 80°C on a fused silica DB-1 column (30 m ϫ 0.25 mm inner diameter). Under these conditions, the retention times for styrene and styrene oxide were 5.3 and 11.8 min, respectively. Authentic standards of the other epoxides were obtained by adding meta-chloroperbenzoic acid to the olefins in CHCl 3 , allowing the mixture to react for 30 min at room temperature, and then injecting the products directly onto the gas chromatographic column. The retention times for cis-␤-methylstyrene and cis-␤-methylstyrene oxide were 8.4 and 14.6 min, respectively. The gas chromatographic conditions were modified to 130°C for 2 min, followed by 7°C/min rise to 180°C, and finally 180°C for 5 min, for the separation of cis-stilbene and cis-stilbene oxide (retention times of 10.5 and 12.2 min, respectively). The styrene epoxide formation was quantitated using relative peak areas with excess styrene as the internal standard. The initial CYP119 concentration was used for all calculations. The enantiomeric selectivity of styrene epoxidation was analyzed by isothermal gas chromatography at 100°C on a Chiraldex G-TA column (30 m ϫ 0.25 mm inner diameter). Pd-and PdR-dependent styrene epoxidation assays were also carried out using wild-type, T213S, T214A, T214V, and T213A/ T214A CYP119. Reaction mixtures (total volume 200 l) contained CYP119 (12.5 M), Pd (125 M), PdR (25 M), styrene (saturated), catalase (1 mg/ml), and NADH (1.5 mM). Catalysis was allowed to proceed for 5 h at 30°C. Products were analyzed and quantitated as described above. Reactions with wild-type CYP119 were used as the control incubation.
Characterization of CYP119 -For thermal stability studies, 75 M CYP119 in 50 mM bis-Tris buffer, pH 6.0, was incubated in a water bath at temperatures ranging from 20 to 90°C for 1.5 h. An aliquot of each sample was taken and immediately reduced and analyzed at room temperature for CO complex formation. Each CYP119 sample was then used in the styrene epoxidation assay carried out at room temperature. For comparison, P450 cam was incubated in 50 mM Tris buffer, pH 7.0, at 40 and 60°C and assessed for reduced-CO complex formation after 1.5 h.
Styrene binding studies were performed in 50 mM bis-Tris buffer at pH 6.0 saturated with styrene (5 l undiluted in 5 ml of buffer) using 1.3 or 2.5 M CYP119. Under these conditions, the styrene concentration exceeded that of the enzyme by at least 1000-fold. CYP119 was added to the styrene-saturated buffer in a capped cuvette at the temperature of interest, and the mixture was incubated for 3 min before the UV-visible spectrum was recorded. The percentage conversion of CYP119 to the high spin state was calculated using an extinction coefficient of 6.4 mM Ϫ1 cm Ϫ1 at max ϭ 646 nm. The binding of thiophenol and 4-ethylthiophenol to 4 M CYP119 in 50 mM bis-Tris buffer, pH 6.0, saturated with each of the two compounds (3 l undiluted in 1 ml of buffer) was monitored by UV-visible spectroscopy.
The initial rates of CYP119 bleaching by H 2 O 2 were calculated for the reaction from 2 to 6.5 min. A 12.5 M solution of CYP119 in 50 mM bis-Tris buffer, pH 6.0, was incubated for 3 min at 30°C before the H 2 O 2 (final concentration 10 mM) was added, and the decrease in the Soret absorbance was recorded.
Resonance Raman Spectroscopy-RR spectra were obtained using a custom McPherson 2061/207 spectrograph (0.67 m) with a Princeton Instruments (LN-1100PB) liquid N 2 -cooled CCD detector. Rayleigh scattering was attenuated with Kaiser Optical supernotch filters. Excitation sources consisted of a Coherent Innova 302 krypton laser (413 nm) and an Innova 90-6 argon laser (514.5 nm). Spectra were collected in a 90°scattering geometry with collection time of a few min. Frequencies were calibrated relative to indene and CCl 4 standards and are accurate to 1 cm Ϫ1 . CCl 4 was also used to check the polarization conditions. The samples contained in glass capillaries were either inserted in a copper finger immersed in a water bath or directly exposed to a thermostated air flow that maintained the sample at ϳ22 and 70°C (18). Typical enzyme concentrations were ϳ90 M in 50 mM bis-Tris (pH 6.0) and ϳ1 mM in styrene when present. The integrity of the Raman samples, before and after laser illumination, was confirmed by direct monitoring of their UV-visible spectra in the Raman capillaries with a Perkin-Elmer Lambda 9 spectrometer.
NMR Relaxation Rate Data Acquisition-Relaxation rates in three independent titrations of styrene with CYP119 were measured using the standard inversion recovery method at 400 MHz with a spectral width of 4500 Hz and 0.1-Hz digital resolution at 298 K using a 5 mm probe. The 700 l D 2 O samples contained a saturated amount of styrene (2.4 mM) (19), 0.1 M potassium phosphate, 100 mM NaCl, and 0.5-10 M CYP119 at a pH (uncorrected for isotope effects) of 5.5. All solutions were pretreated with Chelex to remove free metal ions. The typical pulse sequence contained a long delay (5-7 ϫ T 1 ) between scans, 30 s of low power irradiation on the residual water peak, a 180°pulse, a variable time delay, a 90°read pulse, and acquisition. When enzyme concentrations exceeded 1 M, the integrity of the protein was verified by formation of the ferrous carbon monoxide complex. All T 1 data sets contained 15 time delay points over a 10 -15-fold range centered around the null point.
Relaxation Rate Data Analysis-The amplitudes of the peak heights were fit to standard relaxation equations using a nonlinear method (20 -22). Under the limits of fast exchange, the observed rate (R 1obs ) is a weighted average of the rate in free solution (R 1, f ) and the rate bound to the paramagnetic enzyme (R 1, p ) (23,24).
When S o Ͼ Ͼ K d , the fraction of enzyme bound at 1.5 mM styrene may be approximated as unity. With p p ϳ 1, Equation 1 may be expressed in terms of total enzyme concentration, total substrate concentration, and the dissociation constant (25).
The relaxation enhancement of a ligand in the presence of a paramagnetic protein contains paramagnetic and diamagnetic components. The paramagnetic contribution may be dissected from the total enhancement by duplicating the relaxation measurement with the heme protein in the diamagnetic ferrous carbon monoxide complex (25).
R 1, p and K d may be calculated from a two parameter curve fit when Equation 2 is varied as a function of enzyme and/or substrate concentration. The K d value may also be approached by measuring the spectroscopic binding constant and evaluating R 1, p as the sole unknown. The Solomon-Bloembergen equation relates the paramagnetic effect to a function of the square of the interaction energy, the frequency of the nuclear and electronic transitions, and a correlation time for the motion that modulates the interaction (26,27). The assumptions and restrictions of this equation have been discussed in detail elsewhere (28,29).
The parameter r is the distance from the heme iron to the ligand proton, s is the spin state of the heme iron, I and s are the nuclear and electronic Larmor frequencies, and c is the correlation time for the motion that modulates the interaction. c Ϫ1 is the sum of three correlation times: the electronic relaxation rate s Ϫ1 , the rotational correlation rate r

Ϫ1
, and the chemical exchange rate m Ϫ1 (29). Due to the uncertainty in the precise ratio of high to low spin protein in the solution, limiting values were calculated by assuming that the protein was all high spin or all low spin.
The electronic relaxation rate for heme proteins is usually 10 10 to 10 11 s Ϫ1 (29). Chemical exchange rates rarely exceed 10 6 s Ϫ1 , and rotational rates rarely exceed 10 8 s Ϫ1 based on the Stokes-Einstein equation. c Ϫ1 is therefore usually dominated by s Ϫ1 and has been estimated by the frequency dependence of R 1, p for several proteins. In most cases, c Ϫ1 lies between 10 Ϫ10 and 10 Ϫ11 (25,30,31). In accord with these previous results, we have assumed that c Ϫ1 is equal to 5 ϫ 10 Ϫ11 . Aryl-Iron Complex Formation-Aryldiazenes were prepared immediately prior to use by adding 1 l of 2 N NaOH to 5 l of a 100 mM stock solution of the desired alkyl aryldiazene carboxylate azo ester in methanol, followed by dilution of the resulting mixture with 44 l of potassium phosphate buffer (50 mM, pH 7). Typically, 5-20 l of the final aryldiazene was added to the protein solution, and the formation of the complex was monitored by UV-visible spectroscopy. The protein solution was either air-saturated or made anaerobic by purging with argon for 10 -20 min. The rate of complex formation or decomposition was obtained by following the absorbance change at 468 -478 nm as a function of time. The precise wavelength monitored depended on the aryldiazene and the absorbance maximum of the resulting aryl-iron complex. The data were fitted to single or double exponential equations.
N-Aryl-PPIX Analysis-The heme products were isolated by pouring the protein mixtures into 8 ml of 5% H 2 SO 4 and extracting with 3 ϫ 1 ml of CHCl 3 . After evaporation of the solvent from the combined organic fractions, the residue was redissolved in 50 l of solvent A (see below) and injected onto the HPLC. HPLC was carried out on an HP1090 system equipped with a diode array detector and a Partisil ODS-3 column (5 m ϫ 4.6 mm ϫ 250 mm) (Alltech, San Jose, CA) eluted with an isocratic mixture of solvent B (methanol/acetic acid, 10:1) into solvent A (methanol/water/acetic acid, 6:4:1). The effluent was monitored at 416 nm. Circular Dichroism-All measurements were performed on a Jasco J-715 instrument fitted with a Peltier water bath. A 5 M enzyme concentration in 50 mM bis-Tris buffer, pH 6.0, containing 100 mM KCl was used. The melting point for wild-type CYP119 and each of its mutants was determined by monitoring the change in molar ellipticity at 221 nm over the temperature range from 40 to 100°C.

Expression, Purification, and Characterization of CYP119 -
CYP119 was heterologously expressed in E. coli behind two tac promoters using the pCWori vector. The overexpressed protein was purified to apparent homogeneity ( Fig. 1) using anion exchange and chromatofocusing chromatography. As determined by SDS-polyacrylamide gel electrophoresis ( Fig. 1) and by quantitation of the reduced-CO difference spectrum, each liter of culture yielded ϳ20 mg of purified CYP119. Analysis by gel filtration fast protein liquid chromatography showed that CYP119 is a 43-kDa protein, as predicted by the cDNA-encoded primary sequence. Purified CYP119 at a 75 M concentration is stable for over 90 min at 80°C in 50 mM bis-Tris buffer at pH 6.0 and can be stored at 4°C for at least 1 year. As purified, CYP119 was completely low spin with a Soret absorbance maximum at 415 nm (Fig. 2). Both the reduced CO difference (Fig.  2) and RR spectra (see below) of CYP119 are indicative of a typical P450 low spin six-coordinate iron. Replacing the water ligand of CYP119 by imidazole shifts the Soret maximum from 415 to 424 nm (not shown). The addition of thioanisole, 4-ethylthiophenol, and ethyl methyl sulfide produces a split Soret band with maxima at 380 and 465 nm similar to that previously observed by Ullrich and co-workers (32) and Dawson and co-workers (33) with other P450 enzymes (Fig. 2, inset). CYP119, which was first identified as a possible P450 enzyme from the S. solfataricus genome sequence, indeed has the general properties characteristic of a cytochrome P450 enzyme. Styrene Epoxidation by CYP119 -Since the endogenous electron transfer partners for CYP119 remain unknown, the ability of alternate proteins to transfer electrons to CYP119 was investigated. Incubation of CYP119 with Pd/PdR, spinach ferredoxin/ferredoxin reductase, and human cytochrome P450 reductase under a CO atmosphere, conditions that should lead to formation of the ferrous-CO complex with a max at ϳ450 nm, showed that none of these proteins functioned effectively as an electron donor partner for CYP119. This conclusion was confirmed by unsuccessful efforts to use these electron donor proteins to support the CYP119-catalyzed oxidation of styrene (see below). However, the catalytic activity of CYP119 could be assayed in the absence of electron donor proteins using H 2 O 2 as the source of oxidizing equivalents.
No information is available on the endogenous substrates for CYP119, although the growth of S. solfataricus in sulfur-rich environments makes sulfur compounds good candidates for this role. We have used styrene and substituted styrenes as test substrates because they are readily available, they give products that are readily analyzed, and there are extensive data on their oxidation by other P450 enzymes (34 -37). H 2 O 2 was used for the assays to minimize the likelihood of homolytic oxygen radical formation, although preliminary experiments suggest that tert-butylhydroperoxide is a more efficient donor of oxidizing equivalents. The sole detectable product detected by gas chromatography in a 60-min incubation of CYP119 with styrene and 10 mM H 2 O 2 was styrene oxide (Fig. 3A). The styrene oxide was identified by direct chromatographic comparison with an authentic standard. Although a longer incubation time increased the yield of epoxide, the reaction time was held at 30 min to minimize the concomitant, H 2 O 2 -dependent bleaching of the CYP119 porphyrin chromophore. Kinetic studies show that with 10 mM H 2 O 2 at 30°C CYP119 converts styrene to styrene oxide with a V max of 0.6 nmol min Ϫ1 nmol Ϫ1 protein. Control experiments establish that the reaction requires both H 2 O 2 and CYP119. This is the first CYP119 catalytic activity to be identified.
Incubation of CYP119 with cis-␤-methylstyrene and cis-stilbene shows that the epoxidation reaction takes place with complete retention of the olefin stereochemistry. The only products identified in these reactions by gas chromatography were cis-␤-methylstyrene oxide (Fig. 3B) and cis-stilbene oxide (data not shown), respectively. As observed with other P450 enzymes (34), the epoxidation of styrene by CYP119 is enanti- Thermal Stability-The effect of temperature on both the CYP119 heme chromophore and styrene epoxidation activity was investigated. After preincubation of CYP119 at temperatures between 20 and 80°C for 90 min, the absorbance maximum of the ferrous CO complex of the enzyme determined at 25°C remained entirely at 450 nm. Abrupt precipitation occurred when the protein was preincubated between 80 and 90°C (Fig. 4). This thermal stability contrasts with that of P450 cam , which at 60°C partially precipitated and gave mixtures of species absorbing at 420 and 450 nm (data not shown). Preincubation of CYP119 for 90 min at temperatures between 20 and 80°C did not measurably alter the styrene epoxidation activity of the enzyme when subsequently measured at 25°C (Fig. 4).
Substrate Binding Studies by UV-visible Spectroscopy-As shown by changes at 646 nm in the difference absorption spectrum of substrate-bound versus substrate-free CYP119, styrene only converts a small fraction of the enzyme to the high spin state at temperatures between 20 and 80°C (Fig. 5). Similar results were obtained using cis-␤-methylstyrene, cis-stilbene, and 4-methoxystyrene (data not shown). The fraction of the high spin species increased slightly with increasing temperature, but it has not been possible to determine whether a temperature-dependent correlation exists between the spin state and the catalytic activity. Preliminary experiments suggest that styrene oxidation increases as the incubation temperature is raised, but substrate and product recovery were not sufficiently quantitative at the higher temperatures to make possible accurate rate comparisons.
Resonance Raman Characterization of Styrene Binding-The minor spin state changes that accompany the addition of styrene to CYP119 suggest that this substrate either very poorly displaces the distal water ligand from the iron or does not actually bind in the active site and only allosterically perturbs the spin state. We have therefore also examined the binding of styrene to CYP119 by resonance Raman and NMR methods. Fig. 6 shows the effect of temperature and excess substrate on the high frequency region of the CYP119 RR spectrum obtained with Soret excitation. The data obtained at room temperature are very similar in the absence and presence of substrate (Fig. 6, A versus B). The heme core marker bands 4 , 3 , 2 , and 10 , at 1370, 1500, 1581, and 1635 cm Ϫ1 , respectively, are characteristic of a hexacoordinate low spin (Fig. 6,  6cLS) heme and are very similar to the literature values for substrate-free P450 enzymes (38). When the temperature is raised to 70°C, a significant amount of the pentacoordinate high spin species (Fig. 6, 5cHS) is observed. This low to high spin shift appears as a 3 component at 1485 cm Ϫ1 and as additional intensity in the 2 region at ϳ1570 cm Ϫ1 , while hexacoordinate low spin contributions are diminished (Fig. 6). Although at 70°C, the addition of excess styrene increases the conversion of LS (Fig. 6D) to HS (Fig. 6C), a large proportion of LS heme is still present. The percentage of LS to HS conversion deduced from the RR data is consistent with the UV-visible absorption data (Table I).
Resonance Raman using excitation wavelengths away from the Soret absorption can result in resonance enhancement of iron-ligand stretching vibrations. In heme proteins with cysteinate-proximal ligands, the Fe-S stretching mode is observed only in pentacoordinate high spin species ferric hemes using excitation wavelengths around 540 or 365 nm (39 -41). The low frequency RR spectra of substrate-bound CYP119 at 22 and 70°C was obtained with 514.5-nm excitation (Fig. 7). A new signal at 352 cm Ϫ1 can be isolated from the 70°C data by subtraction of the room temperature spectrum (Fig. 7C). The 352 cm Ϫ1 band is assigned to the (Fe-S) of the pentacoordinate high spin species substrate-bound CYP119, since it is absent in the purely LS sample, and it is not observed with 413-nm excitation. The relatively small amount of HS species formed in CYP119 even with excess styrene at 70°C is consistent with the low intensity of the 352-cm Ϫ1 band as compared with modes insensitive to spin states such as 16 at 753 cm Ϫ1 . Whereas the intense 16 vibration at 753 cm Ϫ1 is subtracted out in the difference spectra, the positive signal at 352 cm Ϫ1 is accompanied by a negative band of comparable intensity at 390 cm Ϫ1 . 2 The signal at 352 cm Ϫ1 is comparable with the Fe-S vibration observed at 351 cm Ϫ1 in P450 cam (41). The similarity in (Fe-S) frequencies between the two P450s indicates that the interaction within the proximal side of the heme pocket of pentacoordinate high spin species CYP119 does not differ from that of other P450 enzymes.

FIG. 7. Low frequency region of the RR spectra of ferric wildtype CYP119 obtained with 514.5-nm excitation (20 milliwatts) in the presence of excess styrene at 70°C (solid trace A) and 22°C (dotted trace B) and the difference spectrum A minus B (trace C).
The inset is an enlargement of the same spectra in the 300 -500-cm Ϫ1 region.

T 1 NMR Studies of Substrate Binding-Titration of an NMR
sample saturated with styrene with 0.06 -10 M CYP119 caused a large decrease in the relaxation rate of the styrene protons (Fig. 8). After measuring the relaxation rate of the styrene protons in the presence of the ferric CYP119, the heme was converted to the diamagnetic ferrous carbon monoxide complex. The relaxation rates of the styrene protons in the presence of the ferrous carbon monoxide complex of CYP119 were equal to those in the absence of CYP119, indicating that the diamagnetic contribution of micromolar CYP119 was negligible. When concentrations of CYP119 were greater than 1 M, the reduced carbon monoxide complex was observed by UV-visible spectroscopy as P450 with no P420. Analysis of the decrease of the relaxation rate of the styrene protons as a function of CYP119 concentration, assuming that the protein was completely in the high spin state, yields distances from each of the styrene protons to the iron of approximately 6.4 Å (Table II). If the assumption is made that the protein was completely in the low spin state, these distances are calculated to be 4.4 -4.5 Å. These results clearly place styrene within the active site of the enzyme. Furthermore, the equivalence of the distances for all of the protons suggests either that (a) styrene binds in a single position from which the distance of each styrene proton to the heme iron is roughly the same, or (b) styrene is able to enter into the active site in various orientations or, once inside, tumbles sufficiently rapidly to average out the distances of the various protons. Since it is highly unlikely that styrene is the natural substrate for CYP119, it is likely that tumbling rather than a fixed orientation is responsible for the similarity of the distances.
CYP119 Spin State Equilibrium-The inability of ligands such as styrene to displace the CYP119 water ligand from the iron, as indicated by the very small shift from the low to the high spin state, could be due to unusually strong hydrogenbonding interactions between the water ligand and protein residues. Furthermore, the inability to significantly shift the spin state could account for the failure of the surrogate electron donor proteins examined to reduce the iron, since the spin state shift in P450 cam is associated with a change in the redox potential that facilitates reduction of the iron by Pd/PdR (42).
Alignment of the CYP119 sequence with those of the five P450 enzymes for which crystal structures are available suggests that Thr-213 is the highly conserved CYP119 catalytic threonine residue (Fig. 9). To explore the possible role of Thr-213 in controlling the spin state and properties of CYP119, it was mutated to an alanine, serine, valine, phenylalanine, and tryptophan. Each mutant exhibited a normal ferric P450 spectrum, formed a reduced-CO complex with a Soret maximum at approximately 450 nm, and had a CD spectrum very similar to that of the wild-type protein (data not shown). As shown in Table I, the spin states of both the substrate-free and styrenebound mutant proteins are highly temperature-dependent. The percentage of the protein in the high spin state in the absence of substrate varied from between 0 and 17% at 40°C to between 5 and 38% at 70°C. In the presence of styrene, the fraction of the T213W mutant in the high spin state was 31% at 40°C and 51% at 70°C. However, whereas in wild-type CYP119 the spin state effects of increasing the temperature and adding styrene were additive, the changes in the spin states of the Thr-213 mutants appear to be due mostly to the increase in temperature.
CYP119 has an unusual sequence of three threonine residues beginning with the conserved Thr-213 (Fig. 9). To determine whether Thr-214 is also involved in stabilization of the CYP119 water ligand, it was mutated to an alanine and valine. These mutants were found to have 10 -15% high spin character at room temperature in the substrate-free form (Table I). Interestingly, the T214V mutant separated into two distinct bands during chromatofocusing. The first band contained predominantly low spin protein, and the second band contained predominantly high spin protein (Fig. 10). Both species were shown by their ferrous CO spectra, which exhibited maxima at 450 nm, to be entirely in the undenatured P450 state (data not shown). Similar phenomena were observed with the other T213 and T214 mutants, but only in the case of T214V were the low and high spin species almost completely resolved chromatographically. The basis for this separation of the proteins into two fractions differing in spin state is not known at this time. The T213A/T214A double mutant was constructed to see if the spin state effects of the two mutations were additive. Surprisingly, the changes in Thr-214 appear to have a much greater impact on the spin state than the changes in the putative conserved Thr-213 residue (Table I).
Active Site Topology of Threonine Mutants-As indicated above, sequence alignments suggest that Thr-213 and Thr-214 are probably situated within the CYP119 active site. To examine their proximity to the heme moiety and their role in determining the active site topology, the Thr-213 and Thr-214 mutants were allowed to react with aryldiazene probes. Aryldiazenes react with P450 enzymes to give aryl-iron complexes that can be induced to undergo a migration of the aryl group from the iron to the porphyrin nitrogens by incubation with ferricyanide (43). The distribution of the four possible  N-aryl-PPIX isomers isolated from the reaction can then be related to the active site topology and can be used, in particular, to explore changes in the topology caused by mutations. We have found that CYP119 is unusual in that it forms the aryliron complex, but the complex is unstable in the presence of oxygen and directly undergoes a shift of the aryl group to the porphyrin nitrogens (44). The CYP119 studies thus do not require the use of ferricyanide but simply require anaerobic incubation of the enzyme with the aryldiazene to generate the aryl-iron complex followed by aerobic incubation to induce migration of the aryl group. The process is thus described by the following equation, where A and B are the enzyme and the aryldiazene, respectively, C is the aryl-iron complex, and D is the N-aryl heme.
The rate of formation at 60°C of the (4-trifluoromethylphenyl)-iron complex of the T213A mutant was roughly twice that for the wild-type enzyme (Table III), but the rate of complex decomposition was essentially the same for all of the mutants (data not shown). HPLC analysis of the N-aryl-PPIX regioisomers extracted from the reactions revealed that the mutations profoundly altered the active site topology. The aryl group migrates primarily to the nitrogens of pyrrole rings C and D in wild-type CYP119 (Fig. 11), suggesting that the active site is least sterically encumbered over these two pyrrole rings (44). However, in the T213A and T213S mutants, the nitrogens of pyrrole rings A and B become much more accessible, and the aryl group migrates preferentially to these two pyrrole rings (Fig. 11). The effect is most pronounced with 4-bromophenyldiazene, for which the regioselectivity of the iron to nitrogen migration is completely altered. For this probe, the favored migration of the probe to pyrrole rings C and D in the wild-type enzyme is completely suppressed in favor of migration to pyrrole rings A and B (Table III). A second striking example is provided by the (3,5-difluoro-4-nitrophenyl)-iron complex, which with wild-type CYP119 yielded no N-aryl-PPIX adducts, but with the T213S mutant gave rise primarily to the pyrrole ring A and B adducts. In contrast, mutations that replaced Thr-213 with bulkier phenylalanine or tryptophan residues slowed down the rate of complex formation, suggesting that the   side chain at position 213 is located above, or close to, the heme in a manner that interferes with reaction of the aryldiazenes with the iron. The regioisomers obtained with these bulkier mutants were essentially the same as those observed for the wild type protein, although in the case of the T213F mutant access to nitrogen D by the p-trifluoromethylphenyl moiety appears somewhat reduced (Fig. 11). Mutations at the adjacent threonine residue had much less influence on the reactions of the enzyme with aryldiazenes. Thus, the T214A and T214V mutations slightly increased the rate of complex formation but had little effect on the N-alkyl-PPIX isomer ratio. Although quantitation of the products is neither trivial nor reliable, it appears that the amount of product based on total integration of the extracted material was much higher for the T213A and T213S mutants than for the wild type or the T213F or T213W mutants. The higher yields of N-aryl-PPIX adducts when Thr-213 is mutated to smaller residues suggest that the yield of adducts is enhanced by decreasing the steric hindrance in the active site.

Rates of complex formation at 60°C and N-aryl-PPIX regioisomer ratios in the reactions of CYP119 and its Thr-213 and Thr-214 mutants with 4-trifluoromethylphenyldiazene, 4-bromophenyldiazene, and 3,5-difluoro-4-nitrophenyldiazene
Effect of Mutations on CYP119 Thermostability-To investigate whether CYP119 thermostability was influenced by alteration of the active site cavity, the melting points of wild-type CYP119 and its Thr-213 and Thr-214 mutants were determined using circular dichroism. These melting point studies reveal that the Thr-213 mutations have little or no influence on CYP119 thermostability (Table IV). In all cases, the melting temperature varied, on average, by only 2°C, and the CD spectra obtained at all temperatures were similar to those obtained with wild-type CYP119 (Fig. 12). These results demonstrate that the CYP119 catalytic machinery (see below) can be separated from the thermostability of the protein.
Effect of Mutations on CYP119 Catalysis-All but one of the active site threonine mutants retained H 2 O 2 -dependent catalytic activity, but they produced varying amounts of styrene oxide. The rates for the Thr-213 mutants roughly correlate with the relative active site access implied by the aryl-iron shift studies (Table IV). In general, a lower rate of enzyme bleaching paralleled a lower rate of styrene epoxidation. However, the Thr-214 mutants are slightly more resistant to H 2 O 2 bleaching but are able to epoxidize styrene at higher rates than wild-type CYP119 while still giving the same 75:25 S:R epoxide ratio.
Surprisingly, the double alanine mutant had properties similar to the T213A mutant. Given that the T214A mutant undergoes a larger spin state conversion and has a higher activity than the wild-type protein, one might have expected the double mutant to exhibit additive effects due to the two single mutations. This result suggests that Thr-213 plays a catalytic role and/or is involved in substrate binding. Furthermore, the data agree with the findings from the aryldiazene reactions that Thr-214 is further away from the heme than Thr-213 and only indirectly influences the CYP119 spin state.
The ability of the CYP119 mutants to accept electrons from electron transfer proteins was tested by measuring the ability of Pd/PdR to support styrene epoxidation. The reactions were carried out with wild-type CYP119 and its T213S, T214A, T214V, and T213A/T214A mutants. Styrene epoxidation by the T214A and T214V mutants is supported by Pd and PdR, whereas catalysis by the wild-type protein and the Thr-213    (Table V). Epoxide formation by the Thr-214 mutants, with rates of approximately 4 and 7 pmol/min/nmol P450, is 100-fold slower than the H 2 O 2 -supported reactions of the same mutants but is comparable with the H 2 O 2 -dependent reactions of the Thr-213 mutants with larger side chains. DISCUSSION CYP119, the first P450 enzyme from a thermophilic organism, withstands prolonged incubation at high temperatures at a pH of 6.0. This is clearly demonstrated by the finding that the absorbance at 450 nm of the ferrous CO complex remains unchanged after preincubation of the protein at a temperature of up to 80°C for 90 min (Fig. 4) and by the finding that the catalytic oxidation of styrene is unimpaired by the same preincubation (Fig. 4). Site-specific mutagenesis has been used in this study to investigate the role of the spin state, and of two residues that control the spin state, on the thermostability and catalytic activity of the protein.
Clues to the stability of the enzyme are provided by the fact that the primary protein sequence contains a large number of charged residues and is shorter by 46 residues than that of P450 cam (3). A second feature of CYP119 that distinguishes it from mesophilic P450 enzymes is the fact that it is in the low spin state and that its conversion to the high spin state is much more difficult than it is, for example, for P450 cam . As illustrated by both UV-visible (Table I) and RR spectroscopy (Fig. 6), the enzyme is shifted slightly to the high spin state by incubating it with styrene or increasing the temperature, but the net spin state shift is minor even with styrene at 70°C. Nevertheless, 1 H NMR relaxation studies indicate that styrene is bound in the active site of the protein (Table II). Since the spin state shift is normally linked to dissociation of the distal water ligand from the iron atom, the recalcitrance of CYP119 to undergo the spin state shift suggests that the distal water ligand is more tightly bound to the iron than usual. The basis for the unusual stability of the distal water ligand is not known, but it could involve the formation of an unusually strong set of hydrogen bonds with active site residues or water molecules hydrogenbonded to those residues or electrostatic effects on the distal side of the active site (45).
If hydrogen bonding is important for stabilization of the water ligand, two possible candidates for the hydrogen-bonding residues are Thr-213, the residue that aligns with the conserved threonine in the sequences of other P450 enzymes, and the adjacent residue Thr-214 (Fig. 9). Indeed, mutation of these threonine residues strongly perturbs the spin state of the enzyme without affecting either the folding or thermostability of the proteins (Table IV). At all temperatures, the T214A, T214V, T213S, T213F, and T213W mutant proteins exhibit a higher proportion of HS species than the wild-type enzyme (Table I).
In contrast, the T213A and T213V mutants have spin state equilibria very similar to those of the parent enzyme. The same pattern is observed when the proteins are incubated with styrene at either 40 or 70°C, with all of the Thr-214 mutants and the Thr-213 mutants with large amino acid replacements showing a considerably increased high spin component. The high spin state of CYP119 is thus favored by groups larger than a threonine at position 213 and by mutation of Thr-214 to nonpolar (non-hydrogen-bonding) residues.
If Thr-213 is hydrogen-bonded to the distal water ligand, removal of this interaction is not sufficient to destabilize the iron-water bond. Either this hydrogen bond does not contribute significantly to stabilization of the water ligand, or the hydrogen bond to Thr-213 is replaced when this residue is mutated by a hydrogen bond to a new water molecule in the active site. However, when the hydrogen bond is eliminated and the new residue at position 213 is large enough to block the positioning of a new water molecule near the distal water ligand, dissociation of the water ligand becomes more favorable, and an increase in the fraction of the protein in the high spin state is observed. In contrast, simple removal of the hydrogen-bonding group from position 214 is sufficient to perturb the stability of the distal water ligand, either because the water and Thr-214 are directly hydrogen-bonded or because Thr-214 participates in a hydrogen-bonding network that helps to stabilize the distal water ligand.
To examine the active site alterations accompanying the threonine mutations, we investigated the reactions of CYP119 and the CYP119 mutants with aryldiazene probes. Analysis of the reaction rates indicates that mutation of Thr-213 to a smaller residue increases, whereas mutation to a larger residue decreases, the rate of formation of the aryl-iron intermediate. In contrast, the Thr-214 mutations have a relatively minor effect on the rate of complex formation. These results clearly suggest that Thr-213 is located close to the iron atom in a position that sterically interferes with formation of the aryliron complex, whereas Thr-214 is so placed that changes in the volume of the side chain at that position have little effect on the rate of the reaction despite their major effect on the spin state of the protein.
As already reported (44), the N-aryl-PPIX regioisomer patterns obtained in the reactions of wild-type CYP119 with the 4-trifluoromethylphenyldiazenes and 4-bromophenyldiazenes indicate that the wild-type active site is primarily open above pyrrole rings C and D (Table III). In both cases, 92% of the resulting N-aryl-PPIX adducts bear the aryl moiety on the nitrogens of these two pyrrole rings. The central role of Thr-213 in the CYP119 active site is beautifully confirmed here by the changes in the N-aryl-PPIX regioisomer ratio when this residue is mutated. Introduction of larger side chains at this position, as in the T213F, T213W, and T213V mutants, causes only small changes in the regioisomer distribution, suggesting a minimal alteration in the active site steric environment. However, the T213A and T213S mutations, in which the size of the side chain is reduced, clearly alter the regioisomer ratio from one that favors pyrrole rings C and D to one that favors pyrrole rings A and B. This is most clearly seen with the 4-bromophenyl probe, which shifts only to the A and B rings in the latter mutants but almost exclusively to pyrrole rings C and D in the wild-type protein (Table III). The aryl shift data suggesting that reducing the size of the residue at position 213 decreases steric encumbrance over pyrrole rings A and B are clearly consistent with the finding that the same mutations increase the rates of formation of the aryl-iron complexes (Table III). In contrast, mutation of Thr-214 to an alanine or valine has little effect on the N-aryl PPIX regioisomer ratios, in accord with the observation that the rates of aryl-iron complex formation are also not altered. Thr-214 is therefore in a sterically insensitive region of the active site despite the fact that mutation of this residue strongly alters the spin state equilibrium. This finding supports the inference that the effects of Thr-214 on the distal water ligand are mediated by polar effects or through a hydrogen-bonding network rather than by a direct hydrogen-bonding interaction.
The unusually high stability of the low spin state of CYP119 and the architecture of the active site defined by Thr-213 and Thr-214 are not linked to the thermostability of the enzyme. Thus, despite the range of spin state compositions and the degree of active site steric crowding represented by the mutants, the melting temperatures of the proteins varied by no more than 2-3°C (Table IV). Although the stability of the hexacoordinate, low spin state is apparently not structurally important for the thermostability of the protein, it could still be important in preventing oxidative degradation of the protein under the high temperature conditions in which it is normally expressed. The low spin state, as shown below, disfavors electron transfer to the iron and the subsequent activation of molecular oxygen, reactions that are deleterious in the absence of a bound substrate. It is possible that the low to high spin shift is normally facilitated by specific interactions with either the native electron donor or the endogenous substrate, so that catalysis only occurs under optimal conditions. In the absence of the endogenous redox partners for CYP119, we have examined the catalytic activity of the enzyme with styrene as the substrate and H 2 O 2 as a surrogate donor of oxidizing equivalents. Under these conditions, CYP119 catalyzes the enantioselective oxidation of styrene to styrene oxide (Fig. 3A), giving a 25:75 ratio of the R-and S-enantiomers of the epoxide. Furthermore, the enzyme oxidizes cis-␤-methylstyrene exclusively to cis-␤-methylstyrene oxide (Fig. 3B) and cis-stilbene to cis-stilbene oxide (not shown). No trace was found of the trans-isomers expected from a non-P450-like epoxidation mechanism. The fact that the oxidation is enantioselective and proceeds with complete retention of the olefin stereochemistry confirms that it occurs within the active site of the enzyme. This epoxidation of styrene and styrene analogues is the first catalytic activity identified for CYP119, although styrene is unlikely to be the natural substrate for the enzyme.
The catalytic oxidation of styrenes by CYP119 would appear to conflict with the observation that these substrates cause only a small shift in the spin state of the enzyme, particularly at the lower temperatures used for the catalytic studies (Table  I). However, RR studies confirm that the binding of styrene causes a small shift in the spin state (Fig. 6). Furthermore, 1 H NMR relaxation studies demonstrate that styrene is bound in the active site of the enzyme with the styrene protons at a distance of approximately 6.4 Å from the heme iron atom (Table II). Although the binding of the substrate may alter as the enzyme traverses the intermediates required to generate the reactive oxygen species, it is clear that styrene binds in the active site although it does not greatly alter the spin state.
Altering the CYP119 spin state equilibrium and the active site environment by mutating Thr-213 and Thr-214 alters both the stability of the heme toward degradation by H 2 O 2 and the styrene epoxidation activity of the enzyme. In all of the mutants, the heme chromophore is bleached more slowly by H 2 O 2 than it is in the wild-type enzyme. The mutants with large side chains at position 213 are most resistant toward bleaching, but they are also catalytically the least active. Thus, mutation of Thr-213 to a large, non-hydrogen-bonding residue (i.e. valine, tryptophan, or phenylalanine) greatly diminishes the H 2 O 2 -dependent epoxidation activity of the enzyme. Even replacement of Thr-213 with a serine, a mutation that increases the high spin state and decreases the active site steric encumbrance, decreases the rate of styrene epoxidation (Table IV). In contrast, a higher epoxidation activity is observed when Thr-213 is replaced by a non-hydrogen-bonding alanine residue in both the single and double mutants. However, one cannot conclude from this that Thr-213 is not important for activation of the ferric peroxide complex, because the role of Thr-213 could be fulfilled by a water molecule in the mutants with smaller side chains. The increased rates of reaction with the aryldiazene probes and the changes in the resulting aryl shift patterns indicate that the smaller side chains decrease steric encumbrance in the vicinity of the iron atom. This decrease in steric interference provides space for additional water molecules and could also help to compensate for a decreased rate of peroxide activation by increasing access of both the peroxide and the substrate to the active site iron. Accessibility of the active site to H 2 O 2 , rather than the spin state of the enzyme, appears to be the crucial determinant of the H 2 O 2 -dependent activity of the enzyme.
A clearer picture of the catalytic roles of Thr-213 and Thr-214 emerges from studies of the Pd-, PdR-, and NADH-dependent epoxidation of styrene by CYP119 mutants. The native electron donor partner for CYP119 is not known, and efforts to identify alternative electron donor partners among the proteins that support other P450 systems have not been successful. However, even if wild-type CYP119 does not detectably oxidize styrene with Pd and PdR as electron donors, the same is not true of some of the mutants. The T213S mutant, which has an increased high spin component (Table I) and a higher H 2 O 2dependent activity (Table IV), also does not detectably catalyze styrene epoxidation using Pd and PdR. However, both the T214A and T214V mutants are able to epoxidize styrene using Pd and PdR as electron transfer partners (Table V). The ability of these CYP119 mutants to accept electrons from Pd and PdR is consistent with the fact that in the native P450 cam system the change in redox potential that accompanies the low to high spin shift is necessary for efficient electron transfer (46). The T213A/T214A double mutant, in contrast, does not retain the ability to utilize Pd and PdR to epoxidize styrene, confirming that the conserved Thr-213 residue is important for formation of the oxidizing species. The Thr-213 results are in keeping with previous results on P450 cam (47). The results also establish that Thr-214 has a major role in controlling the spin state of the enzyme but is not catalytically essential. These results and the evidence obtained from the aryldiazene reactions strongly suggest that Thr-214 influences the spin state and the catalytic activity indirectly, probably as the result of a hydrogen-bonding network that extends some distance away from the iron atom.
In summary, CYP119 is a highly thermostable P450 enzyme. Thr-213, which aligns with the conserved threonine in other P450 enzymes, is close to the heme iron atom as shown by the changes in the spin state, the catalysis of styrene epoxidation, the rates of reaction with aryldiazenes, and the regioisomer patterns obtained from those reactions when Thr-213 is mutated. Thr-214 strongly influences the CYP119 spin state but is at some distance from the iron. As a result, Thr-214 mutants are able to support styrene epoxidation by Pd and PdR, an activity that is lost when Thr-213 is also mutated. These results suggest that Thr-213 interacts with the distal water ligand and functions as a catalytic residue, whereas Thr-214 indirectly influences the protein spin state through a hydrogenbonding network. Significantly, the data indicate that the thermostability of CYP119 does not depend critically on the spin state or the structural features of the active site represented by Thr-213 and Thr-214. This is important because it implies that site-directed mutagenesis may be employed to change the substrate specificity of CYP119 without altering its thermostability, a critical requirement for the development of novel CYP119-based catalysts through protein engineering.