|Home | About | Journals | Submit | Contact Us | Français|
Aromatase (CYP19) is the target of several therapeutics used for breast cancer treatment and catalyzes the three-step conversion of androgens to estrogens, with an unusual C-C cleavage reaction in the third step. To better understand the CYP19 reaction, the oxyferrous complex of CYP19 with androstenedione substrate was cryotrapped, characterized by UV-vis spectroscopy, and cryoreduced to generate the next reaction cycle intermediate. EPR analysis revealed that the initial intermediate observed following cryoreduction is the unprotonated g1 = 2.254 peroxo-ferric intermediate, which is stable up to 180 K. Upon gradual cryoannealing, the low-spin (g1 = 2.39) product complex is formed, with no evidence for accumulation of the g1 = 2.30 hydroperoxo-ferric intermediate. The relative stabilization of the peroxo-ferric heme and the lack of observed hydroperoxo-ferric heme distinguish CYP19 from other P450s, suggesting that the proton delivery pathway is more hindered in CYP19 than in most other P450s.
Aromatase (CYP19) is a membrane-bound microsomal P450 heme protein that converts androgens to estrogens. This protein is of key medical and pharmaceutical importance, as it is targeted by multiple therapeutic drugs currently used for first-line treatment of estrogen-responsive breast cancer . While CYP19 is found predominantly in the ovaries, testes, and adrenal glands, it is also expressed in the brain, where recent studies suggest that it may play a neuroprotective role , further highlighting the importance of understanding CYP19 and the details of its function. For years, biophysical studies have been impeded by the membrane localization of CYP19 and its accompanying tendency to aggregate unless large amounts of detergent were present. To circumvent this problem, we incorporated CYP19 into the native-like membrane environment of Nanodiscs, which stabilizes the protein and allows it to be studied in the absence of detergents . An important advance in our understanding of CYP19 was the recent crystal structure determination , providing an atomistic framework for mechanistic understanding.
As a P450 enzyme, CYP19 is thought to proceed through the typical P450 reaction cycle . The reaction begins with substrate binding, which converts the heme iron from 6-coordinate low-spin to 5-coordinate high-spin by displacing a water molecule. The heme iron is then converted from ferric to ferrous by cytochrome P450 reductase (CPR), which acquires reducing equivalents from NADPH. A molecule of dioxygen next binds to the iron, forming the oxy-ferrous complex. This species is reduced by a second electron from CPR, producing the peroxo-ferric intermediate, a nucleophilic species . Two specific protonations ensue, with the first generating the hydroperoxo-ferric intermediate, while the second protonation leads to scission of the O-O bond and the presumed formation of the highly reactive compound I species (Cpd I). Cpd I then generates hydroxylated substrate, which can directly coordinate the heme iron via its hydroxyl group, and subsequently dissociate to allow the next substrate to bind.
Three rounds through this cycle are required for biosynthesis of estrone from AD by CYP19, with the third round potentially deviating from the typical P450 cycle. In the first step, the C19 methyl of AD is hydroxylated, forming 19-hydroxy AD. A second hydroxylation at C19 produces the gem diol, which is dehydrated to yield 19-aldo AD. In the third step, which is much more enigmatic, the C10-C19 bond is cleaved, releasing estrone and formate. Two main classes of mechanisms have been proposed for the C-C cleavage step. In one, Cpd I is the key reactive species [7, 8]. In the other proposed mechanism, the nucleophilic peroxo-ferric intermediate directly reacts with the substrate's aldehyde functionality, and the resulting peroxo hemiacetal intermediate decomposes to yield the observed products .
Chemical precedence for the nucleophilic peroxo-ferric mechanism is given by the non-enzymatic deformylation of a 19-aldo AD mimic by a peroxo-ferric porphyrin model complex . Additionally, evidence suggests that the peroxo-ferric species may be directly used as the reactive intermediate by CYP2B4, CYP17 and CYP51 for C-C cleaving reactions with aldehyde- or ketone-containing substrates, and by nitric oxide synthase [11-14]. However, analysis of a minimal CYP19 active site model system by density functional theory found an energetic preference for Cpd I-mediated hydrogen abstraction from the C1 of 19-aldo AD leading to estrone formation, rather than product formation following direct reaction of the peroxo-ferric species with 19-aldo AD . Thus, a clear and consistent picture as to how CYP19 catalyzes the final step in estrone formation and of the relative stability of the intermediate iron-oxygen states, is currently lacking.
The primary distinction between the two mechanisms proposed for the third step in the CYP19 reaction is that in the nucleophilic mechanism, protonation of the peroxo-ferric intermediate is expected to be somewhat impeded, enabling direct reaction of the peroxo-ferric oxygen with the 19-aldo AD substrate. This nucleophilic reaction would be in competition with the protonation of the peroxo-ferric species to subsequently generate Cpd I. To gain insight into the mechanism of CYP19 and the unique features that may equip it for this unique reaction, we sought to trap and characterize the reduced oxygenated intermediates of CYP19. Cryoreduction by γ-irradiation at 77 K is a powerful technique that allows reactive intermediates to be generated and characterized [16-20]. The progression of the cryoreduced intermediate is then monitored during the stepwise increase of temperature (cryoannealing), which allows conformational relaxation and progressively larger motions to occur in the enzyme complex. In wild-type CYP101, the hydroperoxo-ferric intermediate is the first species observed by EPR after cryoreduction at liquid nitrogen temperatures (77 K) . The peroxo-ferric intermediate was only observed in CYP101 when the first proton delivery was hindered by mutating Asp251 to asparagine, which allowed us to monitor the conversion of the hydroperoxo-ferric species and finally the product complex as the temperature was increased .
To compare the peroxo-ferric intermediate in CYP19 to other P450s, we utilized low temperature oxygenation and subsequent cryoreduction. Surprisingly, we discovered that, in contrast to the other P450 systems, the unprotonated peroxo-ferric intermediate is the predominant species generated by cryoreduction. These are the first observations of the peroxo species in CYP19. Interestingly, when cryoannealed, no hydroperoxo-ferric intermediate is observed, but rather a low-spin product species is directly formed. This end-point complex displayed the same EPR parameters as an authentic ferric 19-hydroxy AD-CYP19 sample. These findings suggest that the proton delivery pathway of CYP19 may be altered relative to that in other P450s, with the important implication for the detailed mechanism of aromatase.
Modified human CYP19 with a C-terminal His6-tag and codon usage optimized for E. coli (pAr5, based on the pCWori plasmid) was co-expressed in DH5α cells with GroEL/ES chaperones from the commercially available pGro7 plasmid (Takara Bio Inc) [3, 21]. To stabilize CYP19 in a native-like membrane environment, the purified CYP19 was self-assembled into Nanodiscs [3, 22, 23]. Final concentrations in the disc reconstitution mixture were: 5 mM POPC (Avanti Polar Lipids), 20 mM sodium cholate, 77 μM MSP1D1(−) scaffold protein [24, 25], 3.8 μM CYP19, 0.1 M KPi pH 7.4, 50 mM NaCl, and 5 mM β-ME. After the mixture was incubated on ice for one hour, an equal volume of moist Amberlite XAD-2 resin (Sigma) was added to the mixture, which was incubated on ice with shaking for 5 hours to remove cholate, initiating the self-assembly of CYP19-containing Nanodiscs. Ni-NTA chromatography (Qiagen) was used to remove Nanodiscs lacking the His6-tagged CYP19. The CYP19-containing Nanodiscs were further purified with a Superdex 200 size-exclusion column (GE Healthcare) to remove any aggregates and CYP19 not incorporated into Nanodiscs. Purified CYP19-containing Nanodiscs (referred to from this point on simply as CYP19) in 0.1 M KPi pH 7.4 were concentrated. BioUltra grade glycerol (Sigma) was added to a final concentration of 15%, and the protein was flash-frozen in liquid nitrogen and stored at −80° C.
Concentrated CYP19 with 1 mM AD, methyl viologen (MV:CYP19 ratio of 1:50), 0.1 M KPi pH 7.4 and 15% glycerol was deoxygenated and reduced with 2.5 molar equivalents of dithionite in a Coy Labs anaerobic glove box. The dithionite concentration was determined using its molar absorption coefficient of 8.05 mM−1cm−1 at 315 nm . To form the oxy-ferrous complex, which was made in triplicate, the reduced CYP19 + AD sample in glycerol was manually injected into a solution of aerobic glycerol-containing buffer precooled to 243 K, mixed for 25 sec, transferred to a 200 K dry ice/ethanol bath, and stored in liquid nitrogen. The ferrous-oxy EPR samples were maintained at liquid nitrogen temperatures and cryoreduced by γ-irradiation from a 60Co source as previously described  with a dose of 4.5 Mrad .
All UV-vis absorbance spectra were acquired with a Cary 3 double-beam spectrophotometer (Varian Instruments), using a previously described home-built cryostat cooled with liquid nitrogen for low-temperature absorbance spectra [16, 17]. X-Band EPR experiments were performed at the University of Illinois EPR Resource Center on a Varian E-122 X-Band (9.03 GHz) spectrometer, with 0.5 mW microwave power and modulation amplitude of 5 Gauss at 100 kHz. A liquid helium flow system (Air Products) was used to maintain the sample temperature at 15 K.
The addition of AD to CYP19 in Nanodiscs (Methods) induced a spectral shift to high-spin 5-coordinate ferric heme absorbing at 394 nm, indicating fully functional P450 protein [3, 29, 30]. When saturating amounts of the first reaction cycle product, 19-hydroxy AD, was added to CYP19, only a partial spin shift was seen (Figure 1), with a larger fraction of low-spin heme than observed for CYP19 + AD . EPR spectra were acquired for ferric CYP19 in the absence of substrate and with AD or 19-hydroxy AD bound. The g-values obtained for substrate-free and AD-bound CYP19 were similar to those previously reported for CYP19 in detergent , including the observation of two distinct low-spin species for substrate-free CYP19 (Figure 2). CYP19 + AD was partially high-spin at 15 K, and also had two low-spin species, suggesting some conformational heterogeneity in the initial substrate complex. The EPR spectrum of CYP19 + 19-hydroxy AD acquired at 15 K represented a single low-spin species: g = [2.39, 2.25, 1.92]. This is consistent with the higher fraction of low-spin heme observed by UV-vis spectroscopy at 298 K for CP19 + 19-hydroxy AD than for CYP19 + AD (Figure 1). Notably the g1 values of substrate-free and 19-hydroxy AD CYP19 differed slightly, and only CYP19 + 19-hydroxy AD was present as a single species, which should allow substrate-free and product-bound CYP19 to be distinguished by EPR.
Substrate-free CYP19 was rapidly reduced by addition of dithionite, but reduction was much slower in the presence of AD, requiring tens of minutes for complete reduction. Addition of CO-saturated buffer to ferrous CYP19 + AD results in a CO difference spectrum peak at 448 nm, demonstrating that CYP19 + AD remains in its active P450 form.
The oxy-ferrous complex of CYP19 + AD was produced by mixing CYP19 + AD with oxygenated glycerol-containing buffer at 243 K and then transferred to 200 K, and stored at 77 K. The absorbance spectrum of the oxy-ferrous CYP19 + AD complex shows the characteristic split hyperporphyrin Soret and Q-band at 555 nm typical of oxy-ferrous intermediates in P450s (Figure 3) , with the same Soret maximum as we previously observed for CYP19 + AD by stopped flow spectroscopy .
The oxy-ferrous CYP19 + AD sample was cryoreduced at 77 K and the resulting intermediate examined by EPR. The characteristic g1 values of the peroxo-ferric (g1 = 2.25) and hydroperoxo-ferric (g1 = 2.3) species in P450 enzymes should allow these potential CYP19 intermediates to be distinguished by EPR, while any remaining oxy-ferrous complex is EPR silent . For the CYP19 oxycomplex cryoreduced at 77 K, peaks were seen at g1 = 2.254 and g2 = 2.163 (Figure 4), with no high-spin ferric iron and no other low-spin species resulting from possible autoxidation of CYP19 during oxygenation. The CYP19 + AD intermediate with g1 = 2.254 is assigned to the peroxo-ferric species, by analogy with the similar peroxo-ferric signal in the D251N, G248T and G248V mutants of CYP101, wild type CYP101 cryoreduced at 6 K, nitric oxide synthase, peroxidase, hemoglobin and myoglobin [19, 31-35] The EPR spectra of Figure 4 are omitted at 3100 Gauss region, as the organic radicals produced during cryoradiolysis of aqueous glycerol solutions prevent any signals in the neighborhood of g = 2 region from being resolved. The signal indicated with a star in the 77 K EPR spectrum of Figure 4 arises from hydrogen atoms produced during cryoradiolysis , and this signal disappears upon annealing at 143 K as the hydrogen radicals recombine. Most striking is our finding that the peroxo-ferric species is the primary CYP19 intermediate observed at 77 K after cryoreduction, which is quite different from other P450s that have been examined by this approach, where the peroxo-ferric species is already protonated at 77 K [19, 20].
We gradually annealed the peroxo-ferric CYP19 + AD samples at higher temperatures, with EPR characterization at 15 K after each two minute cryoannealing step at the indicated temperatures (Figure 4). The initial peroxo-anion species was stable from 77 to 180 K. Over the range of 190 K to 220 K, the g1 = 2.254 species converted to g1 = 2.39, with no observable accumulation of the g1 = 2.3 hydroperoxo species. The g1 = 2.39 species is designated as the initial product complex of CYP19 + 19-hydroxy AD, based on the similarity to the equilibrium complex of authentic ferric CYP19 + 19-hydroxy AD (g1 = 2.39) (Figure 2). Product complexes have previously been observed during the annealing process of the cryoreduced oxy-ferrous complex of camphor-bound CYP101 . Upon further annealing of the CYP19 sample to 298 K, when diffusion becomes possible, the g1 = 2.39 signal diminished, and the predominant feature was the high-spin iron at g1 = 8.1 (data not shown), with less intense low-spin signals at g1 = 2.25 and g1 = 2.39, due to the expected binding of the large excess of AD substrate.
The observation of the peroxo-ferric heme intermediate as the primary species in cryoreduced CYP19, and its direct transition to the g1 = 2.39 product species during cryoannealing suggests fundamental differences in the proton delivery mechanism between CYP19 and other heme proteins such as CYP101, CYP2B4 and heme oxygenase [19, 20, 34]. An explanation for the difference in proton delivery in CYP19 is that the Asp309 sidechain is oriented quite differently than that of the equivalent Asp251 in CYP101 . This would lead to alterations in the water network, resulting in the active site of CYP19 not being optimized for rapid proton delivery, but rather for stabilizing the peroxo-ferric species so that it could react nucleophilically when an aldehyde-containing substrate is present, such as 19-aldo AD. The stabilization of the peroxo-ferric heme may be a strategy used by multiple heme enzymes, as nitric oxide synthase, which is similarly proposed to use a nucleophilc peroxo-ferric mechanism, and also has a stabilized peroxo-ferric heme species that can be observed at 77 K [12, 33].
In summary, in this communication we report for the first time the isolation and characterization of the peroxo-ferric intermediate of CYP19. We propose that the protonation of the CYP19 peroxo-ferric intermediate is hindered, poising it for nucleophilic attack when a suitable aldehyde-containing substrate is present, such as 19-aldo AD. Together, these results provide new insight into the mechanism of this medically important enzyme.
We thank M. Rentea and A. Luthra for their technical help, Dr. S. Toshkov for proving access to the 60Co source (Nuclear Radiation Lab, University of Illinois at Urbana-Champaign), and Dr. M. Nilges for help with the EPR spectroscopy. This work was supported by the National Institute of Health grants GM31756 to S.G. Sligar.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.