Controlled oxidation of aliphatic CH bonds in metallo-monooxygenases: mechanistic insights derived from studies on deuterated and fluorinated hydrocarbons.

Journal of Inorganic Biochemistry 134 (2014) 118–133

Contents lists available at ScienceDirect

Journal of Inorganic Biochemistry journal homepage: www.elsevier.com/locate/jinorgbio

Focused review

Controlled oxidation of aliphatic C\H bonds in metallo-monooxygenases: Mechanistic insights derived from studies on deuterated and fluorinated hydrocarbons Yao-Sheng Chen a,b,c, Wen-I Luo a, Chung-Ling Yang a,d, Yi-Jung Tu a, Chun-Wei Chang a, Chih-Hsiang Chiang a,e, Chi-Yao Chang b,c, Sunney I. Chan a, Steve S.-F. Yu a,e,⁎ a

Institute of Chemistry, Academia Sinica, Taipei 115, Taiwan Institute of Fisheries Science, National Taiwan University, Taipei 106, Taiwan c Institute of Cellular and Organismic Biology, Academia Sinica, Taipei 115, Taiwan d Graduate Institute of Applied Science and Technology, National Taiwan University of Science and Technology, Taipei 106, Taiwan e Department of Chemistry, National Cheng Kung University, Tainan 701, Taiwan b

a r t i c l e

i n f o

Article history: Received 27 October 2013 Received in revised form 6 January 2014 Accepted 11 February 2014 Available online 21 February 2014 Keywords: C\H activation Enzyme catalysis Fluorinated substituents Isotopomers Metalloprotein Monooxygenase

a b s t r a c t The control over the regio- and/or stereo-selective aliphatic C\H oxidation by metalloenzymes is of great interest to scientists. Typically, these enzymes invoke host–guest chemistry to sequester the substrates within the protein pockets, exploiting sizes, shapes and specific interactions such as hydrogen-bonding, electrostatic forces and/or van der Waals interactions to control the substrate specificity, regio-specificity and stereo-selectivity. Over the years, we have developed a series of deuterated and fluorinated variants of these hydrocarbon substrates as probes to gain insights into the controlled C\H oxidations of hydrocarbons facilitated by these enzymes. In this review, we illustrate the application of these designed probes in the study of three monooxygenases: (i) the particulate methane monooxygenase (pMMO) from Methylococcus capsulatus (Bath), which oxidizes straight-chain C1–C5 alkanes and alkenes to form their corresponding 2-alcohols and epoxides, respectively; (ii) the recombinant alkane hydroxylase (AlkB) from Pseudomonas putida GPo1, which oxidizes the primary C\H bonds of C5–C12 linear alkanes; and (iii) the recombinant cytochrome P450 from Bacillus megaterium, which oxidizes C12–C20 fatty acids at the ω-1, ω-2 or ω-3 \CH positions. © 2014 Elsevier Inc. All rights reserved.

1. Introduction With the recent impetus toward greener chemistry, chemoenzymatic synthesis has undertaken increasing significance and attention [1–6]. We can learn much from nature to develop new catalysts in the laboratory, even based on new chemistry, as has been shown in biological oxidations [7–11]. Enzymes definitely have much to offer on how to control chemical synthesis. Aside from regio- and stereoselectivity, there are other advantages provided by enzymes. First, some reactions can be carried out at the interface between polar and non-polar environments. This feature of interfacial chemistry allows the protein molecules to come in contact with the substrates more effectively, bringing substrates into the hydrophobic pocket of the enzyme where the active site is confined but without the interference from competing reactions. The effective concentration of the substrates is relatively higher due to their hydrophobic nature and the enzyme can readily reach a higher velocity (Vmax) for the catalysis at lower substrate ⁎ Corresponding author at: Institute of Chemistry, Academia Sinica, 128, Academia Road, Sec. 2, Nankang, Taipei 11529, Taiwan. Tel.: +886 2 2789 8650; fax: +886 2 2783 1237. E-mail address: [email protected] (S.S.-F. Yu).

http://dx.doi.org/10.1016/j.jinorgbio.2014.02.005 0162-0134/© 2014 Elsevier Inc. All rights reserved.

concentrations based on the Michaelis–Menten kinetics. Second, enzyme-based chemical conversions can be easily scaled up through the fermentation process, and the reactions can be carried out on whole-cell platforms without the interference from other side products [12,13]. The overall conversion process is robust and efficient. Once the C\H activation occurs at a given carbon atom of the substrate molecule, the synthetic strategy for successive steps can be straightforward. Monooxygenases for the aliphatic C\H oxidation are enzyme systems that harness molecular oxygen (O2) at the catalytic center to mediate oxidation of hydrocarbon substrates [14–22]. Typically, one of the two oxygen atoms in the O2 is transferred to the substrate either to activate a C\H bond in the aliphatics, or the π-system in the vinylic or aromatic compounds. The other oxygen atom is reduced to a molecule of water with the assistance of two protons. (Eq. 1) þ

−

RH þ O2 þ 2H þ 2e →ROH þ H2 O

ð1Þ

In our laboratory, three monooxygenases are currently under in-depth study to elucidate the molecular basis for their functions: (i) the particulate methane monooxygenase (pMMO) from Methylococcus capsulatus (Bath) [23–27]; (ii) alkane hydroxylase (AlkB) from

Y.-S. Chen et al. / Journal of Inorganic Biochemistry 134 (2014) 118–133

Pseudomonas putida GPo1 [28–31]; and (iii) cytochrome P450 BM3 (CYP102A1) from Bacillus megaterium [32–34]. These three enzyme systems exhibit different functions and possess different metal active sites to mediate the chemistry. pMMO is a membrane-bound protein, whose active site is a copper based cluster [7,25–27,35,36]. In addition to methane, its natural substrate, pMMO also oxidizes the secondary C\H bonds particularly at the C-2 positions of C2–C5 linear alkanes to their corresponding alcohols, and oxidizes related alkenes to oxiranes [37,38]. The stereo-selectivity for n-butane and n-pentane is in preference for the formation of (2R)alcohols with enantiomeric excesses (ee) of 70(8)% and 80(2)%, respectively [37,39]. However, when propene and 1-butene are used as epoxidation substrates for pMMO, the ee of the enzymatic products is only 18% and 37%, respectively, in favor of S-configuration. Usually, radical clock compounds containing cyclopropane moiety as mechanistic probes [40,41] are employed to ascertain whether the C\H bond oxidation is involved with the hydrogen abstraction/oxygen rebound chemistry [40,42–44], or occurs by the “oxenoid” or “oxene” insertion reaction mechanism [23,26,43,45]. The direct insertion mechanism involves a low-spin state of the activated metal active site to form C\O bond without a significant barrier according to the two-state hypothesis [44,46–49]. However, since only linear alkanes and alkenes can be oxidized by pMMO, it is not possible to examine the reaction using radical clock compounds. In the case of pMMO, cryptically chiral (R)- or (S)-[1-2H1,1-3H1]ethanes have been used to examine the degrees of inversion or retention of the C\H bond oxidation that occurred at the sp3 carbon of the chiral methyl end [50] and to determine the lifetime for the sp3 or sp2 radical species formed after hydrogen atom abstraction. The corresponding experimental observations indicate that the hydroxylation mediated by the catalytic site in pMMO takes place with full retention of configuration, suggesting that the C\H bond activation in this system is mediated by a concerted or direct “oxene” insertion reaction mechanism [23,26]. It has been proposed that the active site in pMMO consists of a CuICuICuI tricopper-cluster complex that can be activated by dioxygen to produce a mixed-valence [CuIICuII(μ–O)2CuIII] intermediate that harnesses a highly reactive “singlet oxene” for facile concerted insertion into one of the C\H bonds of methane as well as C2–C5 small alkanes [7,35,51]. AlkB is also a membrane-associated protein that belongs to the superfamily of particulate alkane hydroxylase (pAHs), but its active site is thought to comprise eight histidine ligands coordinated to a diiron center [16,18,20,21]. This structural feature of the catalytic site is similar to the case of xylene monooxygenase, fatty acid desaturases, fatty acid monooxygenases, steroid oxygenases and aldehyde decarbonylases [52,53] but different from the case of ribonucleotide reductase R2, soluble methane monooxygenase (sMMO), toluene monooxygenase (TMO) and other monooxygenases, in which the diiron clusters are bridged by carboxylate residues with the duplicated motif composed of (D/EXXH) [16,54]. It has been shown that the active reactive intermediate of AlkB that participates in the hydrocarbon oxidation is also a nonheme diiron-oxo complex [29]. The substrates of AlkB, medium chain-length alkanes (C3–C12), are specifically activated at the primary carbon [28,55–58]. Besides linear chain alkane, AlkB also hydroxylates branched chain alkanes, cyclic alkanes and simple aromatics with linear alkane substituents [59]. AlkB carries out the epoxidation of olefins, sulfoxidation of methyl sulfides, O-demethylation of methyl ethers and other similar oxidations [60–62]. Recent results further support that AlkB can facilitate the desaturation of norcarane suggesting that the hydroxylation and desaturation mediated by the histidine-rich diiron system are highly related [30]. With respect to the mechanism of C\H bond activation mediated by the whole-cell overexpressing AlkB or the purified AlkB, there are significant amounts of rearranged products observed in experiments with bicyclo[3,1,0]hexane, norcarane, and the ultra-fast probe, trans-2phenyl-1-methylcyclopropane [30,63]. Generally speaking, the reaction mediated by the non-heme diiron center of AlkB is carried out by radical

119

rebound reaction mechanism, where the radical species after hydrogen atom abstraction can sustain for an unusually long time (1–300 ns) [30,64,65]. Cytochrome P450 BM3 is an iron heme enzyme [32,66]. The iron heme center of P450 BM3 consists of a 6-coordinate ferric iron heme with an axial thiolate and an axial water molecule as ligands (Fig. 1). When the hydrophobic substrate interacts with the P450 BM3 to expel water molecules within the hydrophobic pocket, the active site is converted to a 5-coordinate ferric iron heme (pathway a in Fig. 1) [66–69]. The resulting high-spin ferric intermediate has a high redox potential and is readily reduced to ferrous heme by taking the electron(s) from NADPH mediated by the FAD/FMN cofactors (pathway b in Fig. 1). The iron heme center is then poised for binding dioxygen, and upon receiving additional electrons, the high valent ferryl oxo species, namely compound I (Cpd I), is formed (pathway c in Fig. 1) [32,44,70–73]. The formation of Cpd I plays the central role in the hydrogen atom abstraction/radical rebound reaction or O-atom insertion mechanism for the controlled C\H activation in the catalysis (pathway d in Fig. 1)[9,43–45,70,74]. Cytochrome P450 BM3 from B. megaterium comprises a watersoluble monooxygenase (BMP domain) fused with a FAD/FMN reductase (BMR domain) [75] that accepts reducing equivalents from NADPH to activate O2 at the heme iron and mediates the hydroxylation of saturated and unsaturated fatty acids with a chain length of 12 to 20 carbons at their sub-terminal ends [76]. Because of its high yield of heterologous over-expression in Escherichia coli, profound turnover efficiency, and catalytic promiscuity, P450 BM3 has been extensively engineered via site-directed or random mutagenesis for the activation of small alkanes, fine chemical conversions, as well as pharmaceutical lead optimization [4,32,77–82]. To study the hydroxylation reaction mechanism mediated by cytochrome P450 BM3, a series of cyclopropyl fatty acids that are similar to the natural substrates of cytochrome P450 BM3, i.e., C12–C16 fatty acids, have been developed as the mechanistic probes [83–85]. It is shown that the oxidation of C13, C15 and C17 cyclopropyl fatty acids mediated by P450 BM3 results in less than 3% rearranged products that come from the radical intermediates. In addition, no product that come from the cation intermediate is observed [83–85]. The estimated rebound rate constants (kt) are in the range of 2–3 × 1010 s−1. Moreover, it is also shown that fatty acid derivatives with the monosubstituted cyclopropane even yield no rearranged products. The outcomes of these experiments either indicate that the mechanistic probes are not sensitive enough to produce significant or detectable amounts of rearranged products, or implicate the involvement of low-spin state of Cpd I with its low kinetic barrier for the C\O bond formation in the hydroxylation of these fatty acid substrates by cytochrome P450 BM3 according to the two-state model [44,46,48], i.e., the hydroxylation is mediated preferably by the direct “oxenoid” or “oxene” insertion mechanism [43,45]. A portion of the products obtained from C13 cyclopropyl fatty acids, the shortest fatty acid (lauric acid) to serve as a good substrate of P450 BM3, have further been identified to emerge from the cation intermediate during the hydroxylation of these substrates [83]. The most obvious similarity among these enzyme systems is that they all bear high-valent metal active sites as well as hydrophobic substrate-binding pockets to carry out the controlled C\H bond activation of hydrocarbons [11,26,32,38,66,86,87]. Regardless, how each system refuels the electrons consumed to perform functional catalytic cycles is the issue of major significance. To validate mechanisms accounting for the substrate specificity as well as regio- and stereo-selectivity of these enzymes, many molecular manipulation techniques can be deployed to create structural and functional variations to elucidate the molecular machinery used in the aliphatic C\H oxidation. For instance, to manipulate the enzyme itself, one extensively used technique is site-directed or random mutagenesis. The substrates can be manipulated by deuteration and fluorination to

120


Fig. 1. The catalytic cycle of hydroxylation in cytochrome P450 BM3 (CYP102A1) [100].

generate isotopomers and bio-isosteres, respectively, another powerful approach in mechanistic studies. In this review, we briefly summarize current understanding of the mechanism(s) for aliphatic C\H oxidation mediated by the three enzymes highlighted above, with particular focus on the control of the regio-specificity and stereo-selectivity driven by the interactions between the substrate and the hydrophobic pocket of the enzyme. The effects resulting from the introduction of deuterated (isotopomers) and fluorinated (bio-isosteres) substituents on the regio-specificity and stereo-selectivity for the C\H bond activation are discussed. Herein, we demonstrate a compelling alternative strategy to elucidate the chemo-selective controls and their contributions in clarifying our understanding of the mechanistic behaviors of monooxygenases. 2. Regio-specific and stereo-selective oxidation of aliphatics by pMMO The pMMO is a tantalizing enzyme that converts methane to methanol in methanotrophs with high efficiency under ambient temperature and pressure. At a turnover frequency approaching 1 s−1, it is the most efficient methane oxidizer discovered to date [7,88]. It is a complex membrane protein consisting of three subunits (PmoA, PmoB, and PmoC) and multiple copper cofactors [23,24,26,27]. Certain strains of methanotrophic bacteria also utilize a second methane monooxygenase, the soluble methane monooxygenase (sMMO), to accomplish the same controlled oxidation of methane [89,90]. The sMMO, a non-heme diiron enzyme, is usually expressed under low copper-to-biomass conditions. However, at high copper concentrations (N 2 μM), these methanotrophs would switch their carbon assimilation from sMMO to pMMO [91]. According to the studies from the Chan laboratory, pMMO is a copper protein with ~ 15 copper ions associated with the active form of the enzyme. These copper ions could be divided into two groups: the electron-transfer clusters (E-clusters) [92–94] and the catalytic clusters (C-clusters) [23,25,26,35,94,95]. The E-clusters contain about 9 reduced copper ions. It has been suggested that these copper ions provide a reservoir of reducing equivalents to re-reduce the C-clusters after the latter cofactors have completed the oxidative phase of the turnover cycle. The C-clusters comprise 6 copper ions: one type 2 copper ion; one dinuclear

copper center; and one tricopper cluster. The type 2 copper ion and the dinuclear copper center have been identified at the A or C site and the B site in the X-ray crystal structure, respectively [24–26]. The tricopper cluster has been shown recently to reside at the D site in the transmembrane domain of the enzyme and its ligand structure has been characterized [7,25,26,51]. A putative substrate-binding pocket has been also located within 6 Ǻ of the tricopper cluster in the vicinity of the D site [26,38,51]. The enantiomeric distribution of the products has been studied in some detail in the case of n-butane. It is shown that the distribution ratio of the product (R)- and (S)-butan-2-ols reflects the presentation of the C\HR and C\HS faces of the prochiral center in the butane molecule to the active “oxygen atom” species during the oxidation of the secondary carbon. Indeed, the nature of the substrate pocket in pMMO reveals the possibility of a binding equilibrium between these two facial orientations of the hydrocarbon chain toward the activated tricopper cluster harnessing the “singlet” oxene (Fig. 2(a) and (b)). Since the conformational reorientation of the hydrocarbon within the binding pocket (108–1012 s−1) is expected to be much faster than the oxidation of the C\H bond by the activated tricopper cluster (k ~ 104 s−1) [35], according to the Curtin–Hammet principle [96], the ratio of (R)- versus (S)butan-2-ol should be given by Keq × (kHR / kHS), where Keq is the equilibrium constant for orienting the hydrocarbon with the C–HR and C\HS bonds of the secondary carbons directly facing toward the harnessed oxene at the catalytic site; kHR and kHS are the kinetic rate constants for oxene insertion into the C\HR and C\HS bonds, respectively. Similar studies on the designed D,L form chiral deuterated butanes, (2R,3R)-[22 H1,3-2H1] and (2S,3S)-[2-2H1,3-2H1] butanes have enabled us to obtain independent and direct determination of the extent of configurational inversion of the carbon center during the hydroxylation process. For substrates (2R,3R)-[2-2H1,3-2H1] (Fig. 2(b)) and (2S,3S)-[2-2H1,3-2H1] butanes (Fig. 2(a)), the product distributions for (R)- and (S)-butan-2SS ol are given by KRR eq × (kDR / kHS) and Keq × (kHR / kDS), respectively. From these experiments, the values of Keq are in the range of 0.17– 0.25 in favor of the C\HR or C\DR facing toward the tricopper site producing the harnessed oxene. The kinetic isotope effects (kH/kD) for oxidation of the secondary C\H bonds in butane are in the range of 5.2– 5.5. These kinetic isotope effects are consistent with those obtained for cryptically chiral [1-2H1, 1-3H1]ethane experiments [50], where the


121

Fig. 2. Energy diagram for the activation of the C\H or C\D bond of secondary carbon centers in (a) (2S,3S)-[2-2H1,3-2H1]butane and (b) (2R,3R)-[2-2H1,3-2H1]butane by the active “oxygen atom” species for concerted “oxene” transfer. In (a), ΔG0HR and ΔG0DS represent the standard free energy differences of the substrate, (2S,3S)-[2-2H1,3-2H1]butane, for the C\HR and C\DS bonds facing toward the active site in protein bound form versus protein free form, respectively; whereas, in (b), ΔG0DR and ΔG0HS are the free energy differences for C\DR and C\HS bonds of the substrate, (2R,3R)-[2-2H1,3-2H1]butane, facing directly toward the active site. ΔG≠ROH or ΔG≠ROD is the activation energy for the oxidation at the secondary C\H bond or C\D bond in D,L-[2-2H1,3-2H1]butanes. In general, ΔΔG≠ = –RTln (kH / kD).

“O-atom” has been shown to take place with total retention of configuration at the C-atom center oxidized. Taken together, the experiments on n-butane and D,L form chiral deuterated butanes indicate that the hydroxylation of butane by pMMO takes place with full retention of configuration at the C-2 position. Thus, the hydroxylation of the C\H bonds in alkanes can be described by an oxo-transfer process based on side-on singlet “oxene” insertion across the “C\H” bond similar to the previously noted carbene insertion [97]. For a butane molecule occupying the binding site within the hydrophobic pocket of a protein, the replacement of the C\HR or C\HS bond at the prochiral center by a C\D bond is by no means an “innocent” substitution. It is well known that the van der Waals radius of the C\D bond is about 0.01 Å shorter than that of C\H bond. The vibrational frequency of the C\D bond is also significantly lower than that of the C\H bond so that London dispersive interactions between the C\D bond and the amino acid residues in direct contact are expected to be stronger than that for the corresponding C\H bond. This is certainly the case in pMMO, where the binding pocket is an “aromatic box” lined by the aromatic residues Trp48, Phe50, Trp51, and Trp54 of PmoA [26,38]. Presumably, these residues can affect the binding of the hydrocarbon substrate, and also influence the orientation equilibrium

governing the presentation of the C\HR and C\HS bonds to the catalytic site. The replacement of the C\H bond by a C\D bond, thus, may potentially change the regio-selectivity and ee of the products. Typically, the effects highlighted above are presumed to be negligible when a normal substrate is replaced by its isotopomers in enzyme studies. However, the “surface roughness” of the substrate can be significantly modified with the incorporation of two or more deuterium atoms, as when the CH2 or CH3 group of the secondary or primary carbons is replaced by CD2 or CD3 group, respectively. To illustrate this, the activity of the pMMO enzyme toward [2,2-2H2]butane has been studied [39]. This butane substrate displays an average product ee of 97(3)% indicating that only the (2R)-alcohol is formed (Scheme 1). The ratio of Kinsertion for the fractional exposure of the “CH3CD2\” (fCH3CD2) and “CH3CH2\” (fCH3CH2) ends to the active site of pMMO for the hydroxylation is 2.8 (0.9). The insertion of the CH3CD2\ end is more favorable than the CH3CH2\ end due to the tightness of the hydrophobic pocket! To further probe the contact surface of the hydrophobic pocket in pMMO, similar experiments on [1,1,1-2H3]butane and [1,1,1,4,4,4-2H6] butane, where the terminal methyl has been replaced by CD3 at one or both ends of the hydrocarbon molecule, have carried out. The ee of the

122


Scheme 1. The regio- and stereo-selective oxidation of substrates, [2,2-2H2]-, [1,1,1-2H3]-, and [1,1,1,4,4,4-2H6]butanes, mediated by pMMO from M. capsulatus (Bath).

oxidation products for [1,1,1-2H3]butane and [1,1,1,4,4,4-2H6]butane are found to be 94% and 83%, respectively, in favor of the formation of the (2R)-alcohol [98] (Scheme 1). In the case of [1,1,1-2H3]butane, the observed change in the stereo-selectivity may also reflect the different degrees of occupancy resulting from different orientations of the terminal CH3\ versus CD3\ end vis-à-vis the edge of the hydrophobic pocket, presumably biasing the activation of the two different sets of C\H bonds in the two orientations. From the GC–MS analysis of [1,1,1-2H3]butan-2-ol and [1,1,1-2H3]butan-3-ol products, a value of 1.8 (0.1) has been obtained for the fraction fCD3CH2/fCH3CH2, indicating that the insertion is in full agreement with the results derived from the [2,2-2H2]butane experiments [39]. In other words, the C\H oxidation of the [1,1,1-2H3]butane is in favor of the CD3CH2\ end over the CH3CH2\ end because of the shorter C\D bond as well as the stronger C\D/π-interaction with the aromatic residues lining the binding pocket. The results are even more dramatic with the replacement of the terminal methyl by a trifluoromethyl group [38]. With 1,1,1trifluoropropane as the substrate, hydroxylation at the C-2 position exhibits an inverse chiral selectivity (33% in favor of (S)-2-ol) opposed to that seen with normal butane, if we consider the size of the CF3 group in the fluorinated propane to be comparable to one of the ethyl groups in butane (Scheme 2). This observation may be consistent with the insertion of the 1,1,1-trifluoropropane toward the close end of the aromatic box. It is possible that the trifluoromethyl substrate is inserted into

the “aromatic box” with the trifluoromethyl group located at the open edge. The size of CF3 group in the fluorinated propane might somehow cause the steric hindrance that crams the opening of the box and biases the orientation of 1,1,1-trifluoropropane, directing the C2\HS toward the tricopper cluster at the active site for hydroxylation of the secondary carbon. Since the C\F bond is considerably more polar than the C\H bond, C\F/π-interactions with “the walls of the aromatic box” may direct the binding and orienting of the hydrocarbon substrate molecule within the hydrophobic pocket of the protein. Finally, when propene and 1-butene are used as epoxidation substrates for pMMO from M. capsulatus (Bath), the ee of the enzymatic products are only 18% and 37%, respectively. It turns out that this relatively poor stereo-selectivity in the enzymatic epoxidation arises from the low stereo-chemical differentiation between the re and si faces in the hydrophobic pocket of the active site. When the 2-butenes are employed as the substrates, trans-2-butene gives only the D,L-2,3dimethyloxiranes products, and cis-2-butene yields only the meso product [38], indicating that the enzymatic epoxidation indeed proceeds via electrophilic syn addition. To further support that van der Waals interactions influence the orientation of various substrates in the hydrophobic cavity of the active site in the enzyme, when 3,3,3-trifluoro-1propene is used as the substrate, the olefin epoxidation is achieved with excellent facial selectivity yielding 90% of (S)-2-(trifluoromethyl) oxirane (Scheme 2). Clearly, all these observations on olefin epoxidation are indicative of the same oxo-transfer process that the Chan laboratory has proposed for C\H activation based on side-on singlet “oxene” insertion across the “C\H” bond. 3. The development of fluorinated probes to study the controlled oxidation of aliphatics by metalloenzymes

Scheme 2. The synthesis of 1,1,1-trifluoropropane, the conversion of 1,1,1trifluoropentane to (S)-1,1,1-trifluoropropan-2-ol (33% ee) and 3,3,3-trifluoro-1-propene to (S)-2-(trifluoromethyl)oxirane mediated by pMMO [38].

One of the obligatory features of an enzyme designed for oxidation of aliphatics is the hydrophobic pocket for the binding of the hydrocarbon substrate. However, the hydrophobic interactions that provide the driving force for the sequestration of the aliphatic substrate are rather complex. The overall free energy stabilizing the enzyme–substrate complex consists of favorable entropic and enthalpic components arising from the desolvation as the substrate enters a protein pocket as well as the reorganization of the water hydrogen-bonding network within the pocket to accommodate the substrate. The dehydration of the ligand accounts for a favorable entropy for the binding of the hydrocarbon to the hydrophobic pocket, whereas re-ordering of the hydrogenbonding network among the water molecules following the displacement of the water molecules by the substrate leads to the reduction in the enthalpy. This scenario provides the driving force that allows the substrate to exhibit closer contact with the cavity or the wall of the hydrophobic pocket after encapsulation. Thus, the interactions of the substrate with the binding pocket can be tuned by introducing bioisosteric substituents into the hydrocarbon. In cytochrome P450 from B. megaterium, an enzyme that oxidizes C12–C20 fatty acids at ω-1, ω-2 or ω-3 positions, the positioning of the fatty acid substrate is initially stabilized by electrostatic interactions


between the carboxylate group and a hydrated hydrophilic domain sequestered within the hydrophobic pocket. If the aliphatic hydrocarbon contains a fluorinated substituent, there are electrostatic interactions between the polar C\F bond(s) and acidic residues within the binding pocket as well as van der Waals interactions between C\F and polarizable groups, e.g., C\F/π-interactions with aromatic side-chain residue(s) and/or C\F/H\C interactions with aliphatic side-chain residue(s), that might contribute to the positioning of the substrate [99–102]. Interestingly, aliphatics with fluorine substituents are often considered bio-isosteric to their parent molecules with C\H bonds in medicinal chemistry. In fact, 20% of pharmaceuticals and 30% of agrochemicals are fluorinated [103]. However, it is still controversial whether C\F bonds exhibit affinity toward polar or non-polar region [32,99,103–107]. The C\F bond is definitely polar; however, with the poor polarizability of the C\F group, fluorinated aliphatics tend to be hydrophobic [105], with the hydrophobic contacts stabilized further by van der Waals interaction between the C\F bond(s) and a polarizable group such as an aromatic ring or aliphatic side chain. Moreover, since a fluorine atom is relatively fatter than a hydrogen atom, fluorinated substituents are more efficient in expelling waters from the hydrocarbon pocket than their non-fluorinated counterparts, and fluorinated hydrocarbons are predicted to be more hydrophobic [104,105]. Earlier, we demonstrate that fluorinated substrates, like its parent hydrocarbons, could also be oxidized by pMMO and the stereoselectively could be dramatically altered [38]. Accordingly, it would be interesting to see whether or not similar controls can also be exerted by fluorinated substituents in other metallo-monooxygenases, such as alkane hydroxylase (AlkB) and cytochrome P450 BM3. Toward this end, we have developed and synthesized fluorinated hydrocarbons for studies with these two monooxygenases. For the preparations of mono-fluorinated and di-fluorinated alkanes, they can be obtained from their precursor primary and secondary alcohols, aldehydes and ketones by employing the fluorination reagent, Deoxo-Fluor™ [Bis(2-methoxyethyl)aminosulfur trifluoride] in CH2Cl2 [108] (Scheme 3). Terminally trifluorinated alkanes can be prepared either by a modified sulfinato-dehalogenation system [109] (Scheme 4) or a Togni reagent II [100,110,111] (Scheme 5). The fluorinated substrates discussed in this review article are listed in Scheme 6. 4. The oxidation of the primary carbons of n-octanes by alkane hydroxylase from P. putida GPo1 Alkane monooxygenase, an integral membrane-bound diiron ω-hydroxylase, can introduce molecular oxygen regio-selectively into the unactivated terminal methyl group of n-octane to yield 1-octanol and

123

water, as shown in Eq. (2) [56]. Using this protein, the Gram-negative bacterium P. putida GPo1 (formerly known as Pseudomonas oleovorans) is able to utilize linear medium chain-length alkanes (C3–C12) as the sole source of carbon and energy [28,55–58]. þ

þ

C8 H18 þ O2 þ NADH þ H →C8 H17 OH þ H2 O þ NAD

ð2Þ

As purified, the functional AlkB complex comprises three components: (i) a soluble rubredoxin-2; (ii) a soluble NADH-dependent rubredoxin reductase; and (iii) an integral membrane non-heme diiron alkane hydroxylase [29,55,56,60,112–114]. These enzymes are encoded by two operons located on the OCT plasmid [115]. The operon that corresponds to the alkBFGHJKL gene cluster, encodes alkane hydroxylase (AlkB) [116] and rubredoxin-2 (AlkG). The other operon, alkST, encodes the third component of the AlkB complex, the flavoprotein (AlkT) [117]. AlkT mediates the transfer of reducing equivalents from NADH to rubredoxin-2. (1R)-[1-2H1,1-3H1]octane and (1S)-[1-2H1,1-3H1]octane have been employed as substrates of P. putida GPo1 and the oxidations are operated by AlkB [14,118]. The major oxidation products (70–80%) have been reported with the retention in configuration at the carbon center oxidized, with a normal H/D kinetic isotope effect (Scheme 7). There are ca. 20–30% of the products obtained with the inversion of configuration, suggesting a mechanism involving hydrogen abstraction/radical rebound chemistry occurring within an enzymatic transition state and/ or within the solvent “cage” [14,69]. Alternatively, the mechanism involves concerted “singlet oxene” insertion across the C\H bond for the bulk of the hydrocarbon, as in pMMO, but for the remaining molecules, there is “spin-crossover” of the “singlet oxene” to give the triplet oxygen atom in the transition state for a two-step radical process culminating in the inversion of the carbon center [39]. Any racemization and epimerization of the primary chiral sp3 carbon in cryptically chiral n-octane might be impeded by a rather low barrier associated with inversion or rotation after the formation of sp2 radical or cation intermediate [119] leading to partial inversion of configuration occurring with the rotational rate constant of 5 × 1012 s− 1 at ambient temperature [120]. However, this conclusion would seem to conflict with the later radical clock experiments, where the lifetimes of the radical species after hydrogen abstraction have been inferred to be much longer (10 −7 –10−9 s) [30,63–65] (vide supra). In any case, partial racemization of the chiral ethanols is observed in the hydroxylation of (S)- or (R)[1-2H1,1-3H1]-ethane mediated by sMMO from Methylosinus trichosporium OB3b [90] and M. capsulatus (Bath) [121]. These results have been interpreted in terms of the formation of a short-lived alkyl radical

Scheme 3. The synthesis of di-fluorinated octanes from the corresponding octanal 11, octanones 12–14 and 1-octanol 15 [108,109,123].

124


Scheme 4. The synthesis of trifluorinated alkanes by sulfinato-dehalogenation system [109].

[90], or one that is so short-lived that the mechanism is essentially “concerted yet non-synchronous” [121,122]. In the aforementioned AlkB experiments on the cryptically chiral octanes, only one end of the octane is chiral. The oxidation occurs at both terminal methyl ends of the chiral octanes. As expected, the enzyme-mediated oxidation predominantly occurs at the achiral terminus (70–80%). However, after correcting for the statistics (three primary C\H bonds at the achiral terminus versus one at the chiral end) and the kinetic isotope effects among the C\H, C\D and C\T bonds, it appears that there is only a small discrimination between the CHDT and CH3 ends of the isotopomer by the substrate pocket of AlkB. In contrast, with fluorinated octanes, we find that the fluorinated substituents can exert a much stronger control when the C\F groups are embedded within the substrate pocket of AlkB. The terminal hydroxylation of the gem-difluorinated octanes 1 and 2 and trifluorinated octane 6 by recombinant AlkB (Scheme 8) from P. putida GPo1 exhibit 100% selectivity for the formation of fluorinated octan-1-ols at the distal methyl group [123]. Interestingly, even with 3,3- and 4,4difluorooctanes (3 and 4) as substrates, the major products still correspond to C\H activation at the omega position (terminal methyl group) (Scheme 8). However, minor products are also formed corresponding to hydroxylation of the more proximal terminal methyl group. Apparently, the fluorine substituent at the C-3 or C-4 position of n-octane can guide the reaction to occur at the relatively more proximal position, albeit at a lower probability. Since only omega C\H activation is observed with 1,1- and 2,2-difluorooctanes (1 and 2), we surmise that the control must be exerted by van der Waals interactions of the \CF2\group with some unique amino acid residues within the hydrophobic pocket of the alkane hydroxylase, which exert the distal controls toward the diiron active site. Previous studies carried out by van Beilen et al. on alkyl-benzenes [59] have shown that the hydroxylation of these benzene derivatives always occurs at the terminal alkyl position regardless of the extension of the alkyl chain (C1–C4) although the rate and the extent of product formation are varied (Scheme 9). The regio-specificity of hydroxylation

might be attributed to steric factors or hydrophobic interactions between substrates and the substrate binding pocket of AlkB. In the light of these observations on the alkyl-benzenes, together with our results on fluorinated alkanes, we surmise that the planar phenyl group of the alkyl-benzenes, or the \CF2\ group of the fluorinated alkanes, interacts with some amino acid residue(s) within the hydrophobic pocket of AlkB. Possibly, an aromatic side chain can promote π–π interaction with the phenyl group of the alkyl-benzene ring, [124–127] or \CF2\/π interaction with the fluorinated substituent of the fluorinated octane, similar to the interactions that we have reported for pMMO. These interactions, if they are exerted at the proper position within the substrate binding pocket, can steer the long-distance C\H activation at the terminal methyl positions of the substrate. Similar van der Waals interactions might also account for the hydroxylation of 1-methylcyclohexane and 1-ethylcyclohexane that always seem to occur at the trans hydrogen of C-4 position in the cyclohexane ring (Scheme 9). Presumably, for these cyclohexane derivatives, interaction of the CH3\ group with the putative aromatic residue(s) within the substrate binding pocket will provide greater stabilization than a similar interaction of the \CH2\ group in the cyclohexane ring [59]. Since the putative aromatic residue(s) are buried within the hydrophobic pocket distal from the metal active site and the proximal poly-histidine residues, where the non-heme diiron is located [29], it(they) could be exploited to direct the histidine-rich region toward the omega position of the substrate. Similar controls are also observed with the hydroxylation of ethyl-substituted toluene mediated by AlkB. The regio-selectivity is always directed at the methyl group of the ethyl group. In principle, we could vary these van der Waals interactions along the hydrophobic pocket through site-directed mutagenesis study [81,128,129] to uncover the molecular basis of the controls of the regio-specificity of the substrate oxidation within the protein structure. Unfortunately, the crystal structure of membrane-bound AlkB is not available. However, cytochrome P450 BM3 from B. megaterium, a soluble metalloenzyme that oxidizes C12–C20 fatty acids at the ω-1, ω-2 or

Scheme 5. The synthesis of trifluorinated lauric acid (dodecanoic acid; C12 fatty acid) by a Togni reagent II [100,110,111].


125

Scheme 6. The fluorinated substrates.

ω-3 position, has now been well-studied with several high-resolution crystal structures available including RCSB (Research Collaboratory for Structural Bioinformatics) PDB (Protein Data Bank) ID: 1JPZ [66] and 1FAG [130]. At this juncture, this system is probably the most reasonable target to explore these effects in-depth, and we shall summarize some of our efforts in the next Section.

5. Sub-terminal oxidation of alkanes and fatty acids by cytochrome P450 BM3 from B. megaterium Fluorine substitution can usually improve the penetration of aliphatics through membranes, strengthen the protein–ligand binding interaction, tune the values of pKa to render them more biologically compatible and enhance their metabolic stability [103,131,132]. Metabolic stability is especially susceptible to cytochrome P450 proteins [133]. There have been a number of systematic studies of cytochrome P450 CAM, another member in P450 family, and P450 BM3 with fluorinated substrates. From these studies, we now have some insights into

the effects of the fluorinated substituents on the selectivity of the substrate oxidation by these soluble cytochrome P450s [100,109,134–137]. In 1984, Eble and Dawson [134] reported an unusual reactivity of soluble cytochrome P450 CAM with 5,5-difluorocamphor as the substrate (Scheme 10). It was expected that the transient redox intermediate Cpd I can be detected by blocking the C-5 position of camphor for the enzymatic hydroxylation. However, the regio-selectivity for hydroxylation dramatically changes from 5-exo to C-9, a position requiring more energy for activation relative to the methylene group in the ring. Apparently, with the fluorine substituent(s), it is possible to tune the reactivity, regio-specificity, and stereo-selectivity in C\H activation of camphor by P450 CAM. Subsequently, Dawson et al. [136,137] have employed 19F NMR to probe the positions and orientations of 9-fluoro- and 5,5-difluorocamphors when these substrates are bound to the active site of cytochrome P450 CAM with the heme iron either existing in the paramagnetic ferrous form or converted to the diamagnetic ferrous-CO derivative. Changes in the 19F NMR chemical shifts of the order of 1– 2.5 ppm are observed for these fluorinated camphors due to the ring-

Scheme 7. The selective oxidation of substrates, (1R)-[1-2H1,3H1]octane and (1S)-[1-2H1,3H1]octane, mediated by alkane hydroxylase (AlkB) from P. putida GPo1 [14,118].

126


Scheme 8. The regio-selective oxidation of substrates, 3,3-, 4,4-difluoro- and 1,1,1-trifluorooctanes 3, 4 and 6, mediated by E. coli whole-cells with the recombinant alkane hydroxylase system (from genes alkBGT) of P. putida GPo1 [123].

current magnetic anisotropy of the heme porphyrin, a change in the microenvironment of the substrates when they become bound to the hydrophobic pocket of the protein. The 19F resonance of the 9fluorocamphor experiences significant line broadening and is shifted by 3.4 ppm relative to the ferrous-CO state of P450 CAM when the ferrous heme is not inhibited by CO. From these observations, the fluorine atom at the C-9 position is estimated to be 3.8 Å away from the heme iron center [136]. Taken together, these NMR results indicate that the \CF2\ or C\F bond in these fluorinated camphors is essentially in direct contact with the porphyrin π-system (26 π) when these substrates are bound to the cytochrome P450 CAM. In the case of cytochrome P450 BM3, it has been suggested from crystallographic studies that the stereo-chemical controls are exerted on the substrate at a site within the pocket some 6–8 Å above (distal) the iron porphyrin [66,109,130] (Fig. 3). According to crystal structures, PDB ID: 1JPZ [66] and 1FAG [130], this asymmetric pocket can be divided into three sub-pockets designated as “L” (Large), “M” (Medium) and “S” (Small) zones. The substrate is judiciously positioned vis-à-vis these sub-pockets to facilitate stereo-selective oxidation of lauric, myristic and palmitic acids at the ω-1, ω-2 and ω-3 positions with 86–99%, 86–96% and 44–74% ee, respectively, to form the respective secondary alcohols in the R-configuration [138–140]. It has also been demonstrated that introduction of the A74G, F87V, L188Q mutations in cytochrome P450 BM3 yields a mutant protein (henceforth referred to as the 3mt protein) that hydroxylates n-octane to the 2-, 3-, and 4-octanol with reasonable activity [141]. However, the rates of NADPH turnover (1608–1760 min−1) and product formation (150 min−1) by the 3mt protein indicate that the coupling between the electron input for the

formation of ferrous state toward dioxygen binding (pathway b,c in Fig. 1) and the O-insertion from oxy–ferryl species (Cpd I) (pathway d in Fig. 1) is low, with a coupling efficiency (CE, the coupling between the electron input from NADPH and the hydroxylation mediated by Cpd I) of less than 10%. The limited catalytic conversion for alcohol formation indicates that, while the ferric iron heme in P450 BM3 3mt protein is sufficiently facile to take up the electrons from NADPH, the switching of the protein conformation to form Cpd I is too slow to oxidize the octane to form the alcohol product, or the substrate is not found at a suitable position within the binding pocket to facilitate efficient oxidation of the substrate. To illustrate these ideas, three conformations of the protein–substrate complex are shown in Fig. 4. Conformers I and II are two conformations that can facilitate conversion of the octane to the 2-ol and 4-ol, respectively; Conformer III, is a conformation that is not amenable to substrate oxidation. In contrast, when fluorinated n-octanes are used as the substrates for the 3mt protein, the CE between electron input and product formation is dramatically enhanced. For the gem-difluorooctanes 1–4, the NADPH turnover can maintain a rate of 1300–1800 min−1 and the CE ratio can be improved to 72–98% [109] (Table 1). The oxidation of 1fluoro- and 1,1-difluorooctanes (Scheme 11) generates the products 2- to 5-ols in this system. The distribution of products is similar to that from the fatty acid substrates for the wild-type P450 BM3, where the regio-selective controls are facilitated at the ω–x positions (x = 1–4). Interestingly, there is no product corresponding to oxidation of these fluorooctanes at the methyl group, as well as the α– and/or β-positions to the fluorinated carbons. In the case of 4,4-difluorooctane 4, only the C-2 position is oxidized. For 3,3-difluorooctane 3, two products corresponding to oxidation at the C-2 and C-3 positions are observed. If the crystal structures of 1JPZ and 1FAD represent the most stable conformers for the oxidation to occur at the C\HR bond of the ω-3 position in fatty acids, that is, ca. 6–7 Å away from heme iron center, this center may control the direct O-atom insertion or hydroxyl radical rebound reaction toward the C\HR or C\HS bonds right up to the ω-1, ω-2 or ω-3 position (Fig. 3) in Cpd I [100,109]. We would surmise that the “M” pocket can only accommodate up to an n-propyl group. In any case, there is a preference for the system to direct the CHF2\ group of 1,1-difluorooctane 1 to the larger sub-pocket (L zone) (Conformers I′

Scheme 9. AlkB exerts highly regio-selective control for the oxidation of alkyl cyclohexane and benzene derivatives [59].

Scheme 10. 5,5-Difluorocamphor is oxidized at C-9 position by cytochrome P450 CAM [134].


Fig. 3. The reproduction of hydrophobic pocket of cytochrome P450 BM3 for substrates Npalmitoylglycine or palmitoleic acid, respectively, from PDB: (a) 1JPZ [66] (b) 1FAG [130] by Discovery Studio® 2.0 (Accelerys Inc.) [109].

and II′ in Fig. 5). However, the minor 5-ol product observed could be accounted for by placing the fluorinated substituents in the hydrophobic pocket of “M” zone for oxidation at the proximal γ-position of the fluorinated substituent (Conformer III′ in Fig. 5). If this scenario is valid, the controls exerted by the fluorinated alkanes in the 3mt protein may not be quite the same as by the fatty acids, or even the fluorinated fatty acids, in the wild-type P450 BM3 (vide infra). The observed product distributions reflecting the regio-selective hydroxylation of the fluorinated n-octanes do not seem to be controlled by hydrogen-bonding of the C\F bonds to the water networks within the substrate-binding pocket. Irrespective of the location of the gemdifluorination (the C-1, C-2, C-3 or C-4 position), the fluorinated noctanes 1–4 all display comparable specific activities for NADPH turnover and product formation [109]. In any case, we expect only 1,1-difluorooctane to exhibit hydrogen bonding to the water molecules

127

in the water channel (Conformers I′ and II′ in Fig. 5). More likely, the positioning of the gem-difluorinated n-octanes is controlled by some unique van der Waals interactions from the C\F bonds with the amino acid residues around the wall of hydrophobic pocket at the “L” zone rather than with the waters of solvation at the distal site of the heme pocket. In contrast, as we shall describe below, in the wild-type BM3 enzyme, polar residues including Arg47, Tyr51 and Ser72, together with the water network, serve as hydrogen-bonding acceptors or donors to anchor the carboxylate end of fatty acids, and in this manner exert distal controls for the oxidations of C12–C16 fatty acids at ω–x positions (x = 1–3) [11,32,142,143]. Clearly, the hydrogen bonding of the carboxylate end of the lauric acid with the distal water networks together with van der Waals interactions between the fluorinated substituents and the wall at the “M” zone of the hydrophobic pocket of the P450 BM3 works together. These two effects are illustrated through the activation of lauric acid by the P450 BM3 enzyme with the terminal methyl group of the substrate substituted by one, two and three fluorine atoms, respectively [100]. The results indicate that the regio-selectivity is significantly improved, especially for the trifluorinated lauric acid 9 (Scheme 6), which exhibits a single product with the activation occurring at the position of the ω-3 (Fig. 6). Significantly, the CE ratios, defined as the coupling between the electron input from NADPH and the hydroxylation mediated by Cpd I, have increased dramatically from 43% (parent lauric acid) to 85–90% (lauric acids 7–9 in Scheme 6). The NADPH consumption rates (TOFNADPH) are in the range of 2100–3600 min−1 for all the lauric acids (Table 1). The fluorinated substituents also have a significant effect on the substrate binding to the BM3 enzyme. To shed some light on this issue, we have compared the binding affinities of the fluorinated substrates and their parent non-fluorinated fatty acids to the iron heme by examining that the fluorinated substituents affect the displacement of the axial water ligand from 6-coordinated low-spin ferric heme transition to yield the 5-coordinated high-spin ferric heme (pathway a in Fig. 1). The study allows us to derive the differences of standard free energies (ΔΔG0LS) [100] in the initial bindings for fluorinated (lauric acids 7–9) versus their parent lauric acids toward the P450 pocket, which are marginally unfavorable in the order of 0.35–0.66 kcal/mol. This destabilization of the low-spin state implies that the fluorinated substrates within the hydrophobic pocket would stabilize the high spin state. For example, 12,12,12-trifluorododecanoic acid 9 exhibits a much higher stabilization of the high-spin state (ΔΔG0 = –0.32 kcal/mol) compared with other fluorinated fatty acids. From Michaelis–Menten kinetics study, smaller Km values are observed for fluorinated fatty acids suggesting that the system prefers to associate to the high-spin state, namely, a higher high-spin/lowspin ratio [100]. The effect of Km may also reflect that the standard free energies (ΔΔG0) of the conformational intermediate states for the subsequent electron input of NADPH and oxo-transfer of Cpd I are also stabilized. With the fatty acid in the substrate-binding pocket, the conformational transitions that link the electron inputs and the O-atom transfers to the substrate should have already been fairly well optimized as the fatty acid is considered as the natural substrate of the cytochrome P450 BM3 [32]. Therefore, we would not expect dramatic improvement in the performance of enzyme for fluorinated fatty acids. Nevertheless, we still observe significant reductions in the activation free energies (ΔΔG≠) for the substrate 12-fluorododecanoic acid 7. Relative to its parent lauric acid, the differences in the NADPH consumption and Oatom insertion are −0.11 and −0.51 kcal/mol, respectively (Table 1). Finally, n-octane and fluorinated n-octanes 1–6 all exhibit reasonable NADPH turnover frequencies for the electron transfer (1100– 2000 min−1) to the ferric iron heme for dioxygen activation. The values are comparable to the fatty acid substrates (2100–4400 min−1) (Table 1). However, the rates of product formation are substantially different. We have compared the effects of fluorinated substituents on the apparent rate constants in Michaelis–Menten kinetics for NADPH

128


Fig. 4. Cytochrome P450 BM3 3mt variant can efficiently carry out the electron transfer from NADPH to the heme iron, but the oxidation of the n-octane to the 2-, 3- and 4-octanol proceeds with low yields. The CE is less than 10%. Thus, either the switching of the protein conformation to form the Cpd I species for the oxidation of the substrate is slow, or the protein is unable to assume conformations that can facilitate the O-atom transfer to the octane. In conformer III, the n-octane is misplaced and unable to be oxidized after the formation of Cpd I.

consumption (kcat,NADPH,app) and O-transfer from Cpd I (kcat,ROH,app), respectively, under the limiting Vmax conditions [100]. Despite that there is no significant improvement for the rate of NADPH consumption, in comparison with the parent n-octane, there is a substantial enhancement (6.7–12 folds) on the formation of alcoholic compounds for mono- and gem-difluorinated n-octanes (Table 1). The differences in the activation energy differences (ΔΔG≠) relative to n-octane are calculated and they are decreased by − 1.1 to −1.5 kcal/mol in the case of the fluorinated n-octanes. With the better coupling of the fluorinated n-octanes for electron transfer from NADPH and oxo-transfer by Cpd I, there is no question that the fluorinated octanes tighten up heme pocket for better turnover of the alcohol formation (Fig. 5). Presumably, the free energy of the transition state with the fluorinated substituents are significantly stabilized due to the unique van der Waals interactions mostly with the aliphatic residues around the “L” zone near the ironheme center.

6. The role of aromatic residues within the substrate-binding pocket of a metalloenzyme for aliphatics oxidation In general, the active site of a metalloenzyme designed for oxidation of aliphatics consists of many aromatics residues such as Phe, Trp or Tyr instead of simple aliphatic residues such as Leu, Val and Ala [144]. Since these aromatic residues have substantial electric polarizability, they can be exploited for the stabilization of aliphatics via C\H/π-interactions. In addition, these amino acid residues provide contact surface area often times essential for positioning of these substrates within the substrate-binding pocket for enzyme catalysis. In the case of pMMO, a hydrophobic pocket predicted by Global Protein Surface Survey (GPSS) analysis on the GPSS web site: http://gpss.mcsg.anl.gov as well as Dockligand (LigandFit) calculations on Discovery Studio 1.7TM allows us to identify an “aromatic box” capable of binding only the hydrocarbon substrates known to be oxidized by the enzyme [26,38] (Fig. 7).

Table 1 The apparent kcat from the turnover frequency (TOF) of NADPH consumption (kcat,NADPH,app) and product formation (kcat,ROH,app), and the effects of the fluorinated substituents on the activation energy from the ES to the ES≠ states (ΔΔG≠NADPH,app and ΔΔG≠ROH,app) for the substrates lauric acid (LA), pentadecanoic acid (PA) and n-octane (OC), as determined by the conversion of fluorinated fatty acids and octanes 1–10 (shown in Scheme 6) by cytochrome P450 BM3 [100,109]. Substrate

kcat,NADPH, −1 a ) app (min

kcat,ROH, −1 b ) app (min

ΔΔG≠NADPH, −1 c ) app (kcal mol

ΔΔG≠ROH, −1 c ) app (kcal mol

Substrate

kcat,NADPH, −1 a ) app (min

kcat,ROH, −1 b ) app (min

ΔΔG≠NADPH, −1 c ) app (kcal mol

ΔΔG≠ROH, −1 c ) app (kcal mol

Lauric acid 7 8 9 Pentadecanoic acid 10

3000 3600 2600 2100 3700

1300 3100 2200 1900 3500

– −0.11 +0.08 +0.21 –

– −0.51 −0.31 −0.22 –

n-octane 1 2 3 4

1600 1800 1800 1300 1800

150 1800 1600 1000 1300

−0.07 −0.07 +0.12 −0.07

−1.47 −1.40 −1.12 −1.28

4400

1500

−0.10

+0.50

5 6

2000 1100

1800 400

−0.13 0.22

−1.47 −0.58

a b c

kcat,NADPH,app = TOFNADPH. kcat,ROH,app = TOFROH. ΔΔG≠NADPH,app = –RTln (kcat,NADPH,app(fluorinated) / kcat,NADPH,app(non-fluorinated)); ΔΔG≠ROH = –RTln (kcat,ROH,app(fluorinated) / kcat,ROH,app(non-fluorinated)).


129

Scheme 11. Conversion of the fluorinated octanes 1–4 to their corresponding secondary alcohols mediated by the P450 BM3 3mt protein [109].

The location of the pocket is adjacent to site D in the crystal structure of pMMO reported by Rosenzweig and coworkers [24]. The hydrophobic “channel” of PmoA is lined by the aromatic residues Trp48, Phe50, Trp51 and Trp54, and “closed” at one end. Located at the open substrate entrance of the pocket is Gly46. This predicted hydrophobic pocket is sufficiently long to bind C1–C5 linear alkanes and alkenes, but too narrow to accommodate branched hydrocarbons. With the hydrocarbon substrate fully inserted into the pocket, the tricopper cluster that is modeled using the hydrophilic amino acids at site D as ligands (His38, Met42, Asp47, Asp49, Glu100 of PmoA and Glu 154 of PmoC) [25] is directed at the C-2 carbon of the substrate near the depth of the pocket, perfectly poised for O-atom transfer to this secondary carbon when the tricopper cluster is activated by dioxygen (Fig. 8). A dioxygen-activated tricopper–peptide complex of the peptide-

fragment HIHAMLTMGDWD derived from the PmoA subunit of M. capsulatus (Bath) at site D of the crystal structure has proved to be capable of mediating facile epoxidation of propylene as well as methane oxidation [7]. In these experiments, the substrate oxidation is carried out by the CuICuICuI–peptide complex encapsulated into mesoporous carbon materials exhibiting the aromatic functionality similar to the “aromatic box” in the protein structure [145,146]. From topological analysis and growth selection experiments of alkane hydroxylase (AlkB) [28,147], it appears that W55 of AlkB controls the chain length of the hydrocarbon entering the pocket of the AlkB (Fig. 9). This structural model has been extensively deployed as a paradigm in research on the fatty acid desaturases [30,52,87,148], an analogous class of non-heme diiron enzymes that are known to play crucial roles in membrane biology and signalling processes in organisms

Fig. 5. The gem-difluorooctanes can tighten up the heme pocket of cytochrome P450 BM3 3mt variant for efficient electron transfer from the NADPH as well as O-insertion into the C\H bond at the ω-1, 2, 3 or 4 positions.

130


Fig. 6. (a) The distal water network within the hydrophobic pocket of cytochrome P450 BM3 assists to switch the regio-selectivity of the oxidation of lauric acid (LA) to occur at the ω-1, ω2 or ω-3 position. (b) The unique van der Waals interactions between the CF3 group of the fluorinated lauric acid 9 and the wall of hydrophobic pocket in cytochrome P450 BM3 facilitate high regio-selective oxidation at the ω-3 position [100].

ranging from bacteria to human beings. Currently available crystal structures of the soluble fatty acid desaturases, PDB: 1AFR [149], 1ZA0 [150] and 2UW1 [151], all show tight lipophilic pockets lined by aromatic residues to support the desaturation process at regio-specific C\C bonds (Fig. 10). The 3mt variant of cytochrome P450 BM3 with the F87V mutation lacks an aromatic residue around the hydrophobic pocket, although there is a porphyrin ring that is highly polarizable. The smoothness of the landscape provides a limited capability for the porphyrin to confine the substrate for regio-specific oxidation. However, after introducing another mutation A328F, we find substantial regio-specific oxidation of n-octane with the major product being 2-octanol (94% abundance) [109]. Not surprisingly, the major products for the fluorinated substrates 1–4 also occur at sub-terminal carbon atoms, resulting largely in the formation of fluorinated octan-2-ol derivatives. These reactions are actually well-controlled to give the ω-1 alcohol away from the fluorine substituent (96–100% abundance). Again, in both the 3mt enzyme as well as its A328F variant, a dramatic enhancement of the CE ratio between electron input and product formation is observed in the oxidation of fluorooctanes, irrespective of the position of the fluorinated substituent (21% for parent n-octane but 68–82% for fluorinated n-octanes 1–4). To account for this behavior, a specific van der Waals interaction between \CF2\ and the π system has been proposed. In the 3mt A328F variant, the interaction between \CF2\ and the π system at residue 328 must be stronger than that between the \CH2\ and the π system at this position to limit the conformational distribution of the substrates leading to one major binding site. Similar interactions involving the \CF2\ group or C\F bond with the porphyrin π-system (26 π) must exist, to tweak the energetics and dynamics of the P450 in order to tighten up the domain of the protein containing the active site for the O-atom transfer during the substrate oxidation. Thus, through hydrogen bondings, electrostatic or van der Waals interactions around the heme pocket within the “L” and/or “M” zones, it is possible to tune the distribution of accessible enzyme conformations and the associated protein dynamics that activates the iron porphyrin substrate hydroxylation to allow the reactions

mediated by the high valent FeIV = O(Por)+• to become kinetically more commensurate with electron transfer from the BMR domain.

7. Conclusion Due to the bond dissociation energies (BDE) of various C\H bonds in aliphatics, it is more difficult to activate the primary methyl C\H bond than the secondary methylene C\H bond because the bond strength of CH3\ group (101 kcal/mol) is higher than that of \CH2\ group (98 kcal/mol) [152]. It has been a long-standing question whether the oxoferryl species or Cpd I in cytochrome P450 BM3 is capable of oxidizing the stronger C\H bond of primary carbon or even a methane molecule. The latter has been one of the holy grail reactions in organic chemistry. Nonetheless, there are now several cytochrome P450 systems identified that can mediate the oxidation of the long-chain alkanes or fatty acids at the primary carbons [18,22,153,154]. By means of directed evolution and site-directed mutagenesis, the promiscuity of cytochrome P450 BM3 has allowed the system to oxidize the substrates at the primary carbon [32,155]. Recently, it was demonstrated that, by introducing long-chain perfluorocarboxylic acids where fluorine atoms serve as chemical space-fillers or “dummy atoms”, the size of the substrate binding pocket of cytochrome P450 BM3 can be compressed. This strategy has even allowed the activation to occur at the secondary carbon of small alkanes such as n-propane and n-butane [156]. More excitingly, under high pressures, this enzyme is even able to perform methane oxidation and primary carbon oxidation of ethane and npropane [11,86,142]. The high valent active site in metalloproteins, such as the non-heme diiron core in sMMO and AlkB, the tricopper cluster in pMMO, and the oxyferryl of Cpd I in cytochrome P450 BM3, all carry out the oxidation of substrates within a protein scaffold with the active-site pocket comprising hydrophobic amino acid residues to stabilize the activation energy of the transition state for the controlled oxidation. Significantly, isotopomeric or bio-isosteric replacements on the substrates can tune the energetics among the conformational states of metalloproteins as


131

Fig. 8. The hydrophobic pocket is composed of the aromatic residues Trp48, Phe50, Trp51 and Trp54 (PmoA). Gly46 is located at open substrate entrance of the pocket. The residues consisting at site D and served as presumable ligands for the coordination of tricopper cluster are denoted in yellow color [38].

BMR

The subdomain of cytochrome P450 BM3 with the reductase domain for accepting the electrons from NADPH C-clusters The catalytic copper ions of pMMO that participate in dioxygen chemistry and alkane hydroxylation CE ratios Coupling efficiency ratios, defined as the coupling between the electron input from NADPH and hydroxylation mediated by Cpd I Cpd I The transient redox intermediate, compound I of oxyferryl iron species in the heme center of cytochrome P450 BM3 E-clusters The copper ions of pMMO that serve as a buffer of reducing equivalents to re-reduce the C-clusters after the catalytic chemistry ee Enantiomeric excess GC Gas chromatography GPSS Global Protein Surface Survey

Fig. 7. (a) The 122nd CASTp surface calculated from the GPSS program. (b) The amino acid residues in the hydrophobic pocket (aromatic box) are presented as CPK model within the presumed binding site whereas the ones at site D for the accommodation of the copper ions are denoted in yellow color with the stick model [26,38].

well as the protein dynamics that participate both in the electron transfer and/or in C\H bond activation. In summary, we show that both deuterated (isotopomers) and fluorinated (bio-isosteres) substrates can behave like their parent ones and enter the hydrophobic pockets of metallo-monooxygenases including pMMO, AlkB and cytochrome P450 BM3 for their controlled oxidation. Owing to variations of the surface roughness or the onset of unique electrostatic and/or van der Waals interactions introduced by the deuterated or fluorinated substituents, many of these deuterated and fluorinated substrates can exhibit even better chemo-, regio-, and stereo-selectivity and/or enzymatic transformation efficiency, thus serving as better substrates for these metalloenzymes. Abbreviations AlkB Alkane hydroxylase BDE Bond dissociation energy BMP The subdomain of cytochrome P450 BM3 with the heme (porphyrin) center

Fig. 9. The topology model predicted by hydropathy profile [147] suggests that the hydrophobic pocket comprising the Trp 55 can encapsulate n-octanes 1–4 and 6, tighten up the active site and promote the activation at the omega terminal C-1 position of n-octanes by exploiting distal electrostatic forces or van der Waals interactions between \CF2\ or CF3\ group and the aromatic π-system of the tryptophan [28].

132


Fig. 10. The ribbon diagrams of the crystal structures for soluble fatty acid desaturases, 1AFR, 1ZA0 and 2UW1 (Discovery Studio™2.0, Acceryles Inc.). The aromatic residues of Phe, Trp, and Tyr are denoted by CPK models. The metal centers in 1AFR, 1ZA0 and 2UW1 contain two irons, one manganese and two irons, respectively.

RCSB Research Collaboratory for Structural Bioinfomatics PDB Protein Data Bank pMMO Particulate methane monooxygenase PmoA Particulate methane monooxygenase subunit A PmoB Particulate methane monooxygenase subunit B PmoC Particulate methane monooxygenase subunit C sMMO Soluble methane monooxygenase TOF Turnover frequency ω-1, ω-2 or ω-3 hydroxylation The hydroxylation reaction occurring at the α, β or γ position of the distal methyl group from the carboxylate end, respectively.

Acknowledgments This work was supported by the Institute of Chemistry, Academia Sinica, and grants from the National Science Council, R.O.C. (NSC 972113-M-001-006-MY3 and 101-2113-M-001-007-MY3). References [1] O. Pamies, J.E. Backvall, Trends Biotechnol. 22 (2004) 130–135. [2] K. Zhang, B.M. Shafer, M.D. Demars II, H.A. Stern, R. Fasan, J. Am. Chem. Soc. 134 (2012) 18695–18704. [3] J.D. Mortison, D.H. Sherman, J. Org. Chem. 75 (2010) 7041–7051. [4] P.S. Coelho, E.M. Brustad, A. Kannan, F.H. Arnold, Science 339 (2013) 307–310. [5] H.B. Bode, R. Müller, Angew. Chem. Int. Ed. 44 (2005) 6828–6846. [6] J.C. Lewis, S. Bastian, C.S. Bennett, Y. Fu, Y. Mitsuda, M.M. Chen, W.A. Greenberg, C.-H. Wong, F.H. Arnold, Proc. Natl. Acad. Sci. U. S. A. 106 (2009) 16550–16555. [7] S.I. Chan, Y.-J. Lu, P. Nagababu, S. Maji, M.-C. Hung, M.M. Lee, I.-J. Hsu, P.D. Minh, J.C.-H. Lai, K.Y. Ng, S. Ramalingam, S.S.-F. Yu, M.K. Chan, Angew. Chem. Int. Ed. 52 (2013) 3731–3735. [8] P. Nagababu, S. Maji, M.P. Kumar, P.P.-Y. Chen, S.S.-F. Yu, S.I. Chan, Adv. Synth. Catal. 354 (2012) 3275–3282. [9] B. Meunier, S.P. de Visser, S. Shaik, Chem. Rev. 104 (2004) 3947–3980. [10] J.-E. Backvall (Ed.), Modern Oxidation Methods, Wiley-VCH, Weinheim, Germany, 2004. [11] F.E. Zilly, J.P. Acevedo, W. Augustyniak, A. Deege, U.W. Häusig, M.T. Reetz, Angew. Chem. Int. Ed. 50 (2011) 2720–2724. [12] H.L. Holland, H.K. Weber, Curr. Opin. Biotechnol. 11 (2000) 547–553. [13] W.A. Duetz, J.B. van Beilen, B. Witholt, Curr. Opin. Biotechnol. 12 (2001) 419–425. [14] S. Shapiro, E. Caspi, Tetrahedron 54 (1998) 5005–5040. [15] D.E.T. Pazmiño, M. Winkler, A. Glieder, M.W. Fraaije, J. Biotechnol. 146 (2010) 9–24. [16] M. Ayala, E. Torres, Appl. Catal. A 272 (2004) 1–13. [17] R. Ullrich, M. Hofrichter, Cell. Mol. Life Sci. 64 (2007) 271–293. [18] J.B. van Beilen, E.G. Funhoff, Appl. Microbiol. Biotechnol. 74 (2007) 13–21. [19] F. Rojo, Environ. Microbiol. 11 (2009) 2477–2490. [20] Y. Ji, G. Mao, Y. Wang, M. Bartlam, Front. Microbiol. 4 (2013) 58, http://dx.doi.org/10. 3389/fmicb.2013.00058. [21] R.N. Austin, J.T. Groves, Metallomics 3 (2011) 775–787. [22] M. Bordeaux, A. Galarneau, F. Fajula, J. Drone, Angew. Chem. Int. Ed. 50 (2011) 2075–2079.

[23] S.I. Chan, K.H.-C. Chen, S.S.-F. Yu, C.-L. Chen, S.S.-J. Kuo, Biochemistry 43 (2004) 4421–4430. [24] R.L. Lieberman, A.C. Rosenzweig, Nature 434 (2005) 177–182. [25] S.I. Chan, V.C.-C. Wang, J.C.-H. Lai, S.S.-F. Yu, P.P.-Y. Chen, K.H.-C. Chen, C.-L. Chen, M.K. Chan, Angew. Chem. Int. Ed. Engl. 46 (2007) 1992–1994. [26] S.I. Chan, S.S.-F. Yu, Acc. Chem. Res. 41 (2008) 969–979. [27] R. Balasubramanian, S.M. Smith, S. Rawat, L.A. Yatsunyk, T.L. Stemmler, A.C. Rosenzweig, Nature 465 (2010) 115–119. [28] J.B. van Beilen, T.H. Smits, F.F. Roos, T. Brunner, S.B. Balada, M. Rothlisberger, B. Witholt, J. Bacteriol. 187 (2005) 85–91. [29] J. Shanklin, C. Achim, H. Schmidt, B.G. Fox, E. Munck, Proc. Natl. Acad. Sci. U. S. A. 94 (1997) 2981–2986. [30] H.L. Cooper, G. Mishra, X. Huang, M. Pender-Cudlip, R.N. Austin, J. Shanklin, J.T. Groves, J. Am. Chem. Soc. 134 (2012) 20365–20375. [31] H. Alonso, A. Roujeinikova, Appl. Environ. Microbiol. 78 (2012) 7946–7953. [32] C.J.C. Whitehouse, S.G. Bell, L.-L. Wong, Chem. Soc. Rev. 41 (2012) 1218–1260. [33] H. Li, T.L. Poulos, Biochim. Biophys. Acta 1441 (1999) 141–149. [34] S. Eiben, L. Kaysser, S. Maurer, K. Kuhnel, V.B. Urlacher, R.D. Schmid, J. Biotechnol. 124 (2006) 662–669. [35] P.P.-Y. Chen, S.I. Chan, J. Inorg. Biochem. 100 (2006) 801–809. [36] M.A. Culpepper, G.E. Cutsail, B.M. Hoffman, A.C. Rosenzweig, J. Am. Chem. Soc. 134 (2012) 7640–7643. [37] S.J. Elliott, M. Zhu, L. Tso, H.-H.T. Nguyen, J.H.-K. Yip, S.I. Chan, J. Am. Chem. Soc. 119 (1997) 9949–9955. [38] K.Y. Ng, L.-C. Tu, Y.-S. Wang, S.I. Chan, S.S.-F. Yu, ChemBioChem 9 (2008) 1116–1123. [39] S.S.-F. Yu, L.-Y. Wu, K.H.-C. Chen, W.-I. Luo, D.-S. Huang, S.I. Chan, J. Biol. Chem. 278 (2003) 40658–40669. [40] P.R. Ortiz de Montellano, S.D. Nelson, Arch. Biochem. Biophys. 507 (2011) 95–110. [41] M. Newcomb, P.H. Toy, Acc. Chem. Res. 33 (2000) 449–455. [42] J.T. Groves, J. Chem. Educ. 62 (1985) 928–931. [43] C. Jung, Biochim. Biophys. Acta 1814 (2011) 46–57. [44] S. Shaik, S. Cohen, Y. Wang, H. Chen, D. Kumar, W. Thiel, Chem. Rev. 110 (2010) 949–1017. [45] M. Sono, M.P. Roach, E.D. Coulter, J.H. Dawson, Chem. Rev. 96 (1996) 2841–2887. [46] D. Kumar, S.P. de Visser, P.K. Sharma, S. Cohen, S. Shaik, J. Am. Chem. Soc. 126 (2004) 1907–1920. [47] N. Dietl, C. van der Linde, M. Schlangen, M.K. Beyer, H. Schwarz, Angew. Chem. Int. Ed. 50 (2011) 4966–4969. [48] M. Newcomb, P.F. Hollenberg, M.J. Coon, Arch. Biochem. Biophys. 409 (2003) 72–79. [49] H. Schwarz, Angew. Chem. Int. Ed. 50 (2011) 10096–10115. [50] B. Wilkinson, M. Zhu, N.D. Priestley, H.-H.T. Nguyen, H. Morimoto, P.G. Williams, S.I. Chan, H.G. Floss, J. Am. Chem. Soc. 118 (1996) 921–922. [51] P.P.-Y. Chen, P. Nagababu, S.S.-F. Yu, S.I. Chan, ChemCatChem 6 (2014) 429–437. [52] J. Shanklin, E. Whittle, FEBS Lett. 545 (2003) 188–192. [53] J. Shanklin, E. Whittle, B.G. Fox, Biochemistry 33 (1994) 12787–12794. [54] J.G. Leahy, P.J. Batchelor, S.M. Morcomb, FEMS Microbiol. Rev. 27 (2003) 449–479. [55] J.N. Baptist, R.K. Gholson, M.J. Coon, Biochim. Biophys. Acta 69 (1963) 40–47. [56] J.A. Peterson, M. Kusunose, E. Kusunose, M.J. Coon, J. Biol. Chem. 242 (1967) 4334–4340. [57] E.L. Johnson, M.R. Hyman, Appl. Environ. Microbiol. 72 (2006) 950–952. [58] D.J. koch, M.M. Chen, J.B. van Beilen, F.H. Arnold, Appl. Environ. Microbiol. 75 (2009) 337–344. [59] J.B. van Beilen, J. Kingma, B. Witholt, Enzyme Microb. Technol. 16 (1994) 904–911. [60] E.J. McKenna, M.J. Coon, J. Biol. Chem. 245 (1970) 3882–3889. [61] A.G. Katopodis, K. Wimalasena, J. Lee, S.W. May, J. Am. Chem. Soc. 106 (1984) 7928–7935. [62] A.G. Katopodis, J. Homer, A. Smith, S.W. May, J. Am. Chem. Soc. (1988) 897–899.

Y.-S. Chen et al. / Journal of Inorganic Biochemistry 134 (2014) 118–133 [63] H. Fu, M. Newcomb, C.-H. Wong, J. Am. Chem. Soc. 113 (1991) 5878–5880. [64] R.N. Austin, H.K. Chang, G.J. Zylstra, J.T. Groves, J. Am. Chem. Soc. 122 (2000) 11747–11748. [65] R.N. Austin, K. Luddy, K. Erickson, M. Pender-Cudlip, E. Bertrand, D. Deng, R.S. Buzdygon, J.B. van Beilen, J.T. Groves, Angew. Chem. Int. Ed. Engl. 47 (2008) 5232–5234. [66] D.C. Haines, D.R. Tomchick, M. Machius, J.A. Peterson, Biochemistry 40 (2001) 13456–13465. [67] D.F.V. Lewis, M.N. Jacobs, M. Dickins, Drug Discov. Today 9 (2004) 530–537. [68] D. Fishelovitch, S. Shaik, H.J. Wolfson, R. Nussinov, J. Phys. Chem. B 114 (2010) 5964–5970. [69] S. Shaik, D. Kumar, S.P. de Visser, A. Altun, W. Thiel, Chem. Rev. 105 (2005) 2279–2328. [70] P.R. Ortiz de Montellano, Chem. Rev. 110 (2010) 932–948. [71] K.A. Feenstra, E.B. Starikov, V.B. Urlacher, J.N.M. Commandeur, N.P.E. Vermeulen, Protein Sci. 16 (2007) 420–431. [72] J. Rittle, M.T. Green, Science 330 (2010) 933–937. [73] C.M. Krest, E.L. Onderko, T.H. Yosca, J.C. Calixto, R.F. Karp, J. Livada, J. Rittle, M.T. Green, J. Biol. Chem. 288 (2013) 17074–17081. [74] J.T. Groves, K. Shalyaev, J. Lee (Eds.), Academic Press, New York, 2000. [75] L.O. Narhi, A.J. Fulco, J. Biol. Chem. 262 (1987) 6683–6690. [76] V.B. Urlacher, R.D. Schmid, Methods Enzymol. 388 (2004) 208–224. [77] V.B. Urlacher, R.D. Schmid, Curr. Opin. Chem. Biol. 10 (2006) 156–161. [78] R. Fasan, Y.T. Meharenna, C.D. Snow, T.L. Poulos, F.H. Arnold, J. Mol. Biol. 383 (2008) 1069–1080. [79] A.M. Sawayama, M.M.Y. Chen, P. Kulanthaivel, M.S. Kuo, H. Hemmerle, F.H. Arnold, Chem. Eur. J. 15 (2009) 11723–11729. [80] A. Rentmeister, F.H. Arnold, R. Fasan, Nat. Chem. Biol. 5 (2009) 26–28. [81] M.T. Reetz, Angew. Chem. Int. Ed. 50 (2011) 138–174. [82] G.-D. Roiban, M.T. Reetz, Angew. Chem. Int. Ed. 52 (2013) 5439–5440. [83] M.J. Cryle, P.Y. Hayes, J.J. De Voss, Chem. Eur. J. 18 (2012) 15994–15999. [84] M.J. Cryle, J.M.U. Stuthe, P.R. Ortiz de Montellano, J.J. De Voss, Chem. Commun. (2004) 512–513. [85] M.J. Cryle, P.R. Ortiz de Montellano, J.J. De Voss, J. Org. Chem. 70 (2005) 2455–2469. [86] N. Kawakami, O. Shoji, Y. Watanabe, Chem. Sci. 4 (2013) 2344–2348. [87] J. Shanklin, J.E. Guy, G. Mishra, Y. Lindqvist, J. Biol. Chem. 284 (2009) 18559–18563. [88] J.D. Semrau, A.A. DiSpirito, S. Yoon, FEMS Microbiol. Rev. 34 (2010) 496–531. [89] D.A. Kopp, S.J. Lippard, Curr. Opin. Chem. Biol. 6 (5) (2002) 568–576. [90] N.D. Priestley, H.G. Floss, W.A. Froland, J.D. Lipscomb, P.G. Williams, H. Morimoto, J. Am. Chem. Soc. 114 (1992) 7561–7562. [91] W.-C. Kao, Y.-R. Chen, E.-C. Yi, H. Lee, Q. Tian, K.-M. Wu, S.-F. Tsai, S.S.-F. Yu, Y.-J. Chen, R. Aebersold, S.I. Chan, J. Biol. Chem. 279 (2004) 51554–51560. [92] S.S.-F. Yu, C.-Z. Ji, Y.-P. Wu, T.-L. Lee, C.-H. Lai, S.-C. Lin, Z.-L. Yang, V.C.-C. Wang, K.H.-C. Chen, S.I. Chan, Biochemistry 46 (2007) 13762–13774. [93] H.-H.T. Nguyen, S.J. Elliott, J.H.-K. Yip, S.I. Chan, J. Biol. Chem. 273 (1998) 7957–7966. [94] K.H.-C. Chen, C.-L. Chen, C.-F. Tseng, S.S.-F. Yu, S.-C. Ke, J.-F. Lee, H.-H.T. Nguyen, S.J. Elliott, J.O. Alben, S.I. Chan, J. Chin. Chem. Soc. 51 (2004) 1081–1098. [95] H.-H.T. Nguyen, K.H. Nakagawa, B. Hedman, S.J. Elliott, M.E. Lidstrom, K.O. Hodgson, S.I. Chan, J. Am. Chem. Soc. 118 (1996) 12766–12776. [96] N.S. Issaacs, Physical Organic Chemistry, Longman Scientific & Technical, Essex, 1987. [97] W. Kirmse, I.S. Özkir, J. Am. Chem. Soc. 114 (1992) 7590–7591. [98] L.-C. Tu, 2004, Master thesis, Department of Chemistry, National Tsing Hua University. [99] D. O'Hagan, Chem. Soc. Rev. 37 (2008) 308–319. [100] C.-H. Chiang, R. Ramu, Y.-J. Tu, C.-L. Yang, K.Y. Ng, W.-I. Luo, C.H. Chen, Y.-Y. Lu, C.-L. Liu, S.S.-F. Yu, Chem. Eur. J. 19 (2013) 13680–13691. [101] I. Hyla-Kryspin, G. Haufe, S. Grimme, Chem. Eur. J. 10 (2004) 3411–3422. [102] T.V. Rybalova, I.Y. Bagryanskaya, J. Struct. Chem. 50 (2009) 741–753. [103] K. Müller, C. Faeh, F. Diederich, Science 317 (2007) 1881–1886. [104] J. Mecinović, P.W. Snyder, K.A. Mirica, S. Bai, E.T. Mack, R.L. Kwant, D.T. Moustakas, A. Héroux, G.M. Whitesides, J. Am. Chem. Soc. 133 (2011) 14017–14026. [105] V.H. Dalvi, P.J. Rossky, Proc. Natl. Acad. Sci. U. S. A. 107 (2010) 13603–13607. [106] B.E. Smart, J. Fluor. Chem. 109 (2001) 3–11. [107] R. Pongdee, H.-W. Liu, Bioorg. Chem. 32 (2004) 393–437. [108] P.S. Rajendra, J.M. Shreeve, Synthesis 17 (2002) 2561–2578. [109] L.-L. Wu, C.-L. Yang, F.-C. Lo, C.-H. Chiang, C.-W. Chang, K.-Y. Ng, H.-H. Chou, H.-Y. Hung, S.I. Chan, S.S.-F. Yu, Chem. Eur. J. 17 (2011) 4774–4787. [110] X. Wang, Y. Ye, S. Zhang, J. Feng, Y. Xu, Y. Zhang, J. Wang, J. Am. Chem. Soc. 133 (2011) 16410–16413.

[111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121]

[122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143]

[144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156]

133

A.T. Parsons, S.L. Buchwald, Angew. Chem. Int. Ed. 50 (2011) 9120–9123. J.A. Peterson, D. Basu, M.J. Coon, J. Biol. Chem. 241 (1966) 5162–5164. J.A. Peterson, M.J. Coon, J. Biol. Chem. 243 (1968) 329–334. T. Ueda, E.T. Lode, M.J. Coon, J. Biol. Chem. 247 (1972) 2100–2116. A.M. Chakrabarty, G. Chou, I.C. Gunsalus, Proc. Natl. Acad. Sci. U. S. A. 70 (1973) 1137–1140. M. Kok, R. Oldenhuis, M.P. van der Linden, P. Raatjes, J. Kingma, P.H. van Lelyveld, B. Witholt, J. Biol. Chem. 264 (1989) 5435–5441. G. Eggink, H. Engel, G. Vriend, P. Terpstra, B. Witholt, J. Mol. Biol. 212 (1990) 135–142. E. Caspi, S. Shapiro, J.U. Piper, Tetrahedron 37 (1981) 3535–3543. M.H. Baik, M. Newcomb, R.A. Friesner, S.J. Lippard, Chem. Rev. 103 (2003) 2385–2419. T.J. Sears, P.M. Johnson, P. Jin, S. Oatis, J. Chem. Phys. 104 (1996) 781–792. A.M. Valentine, B. Wilkinson, K.E. Liu, S. Komar-Panicucci, N.D. Priestley, P.G. Williams, H. Morimoto, H.G. Floss, S.J. Lippard, J. Am. Chem. Soc. 119 (1997) 1818–1827. S.-Y. Choi, P.E. Eaton, P.F. Hollenber, K.E. Liu, S.J. Lippard, M. Newcomb, D.A. Putt, S.P. Upadhyaya, Y. Xiong, J. Am. Chem. Soc. 118 (1996) 6547–6555. R. Ramu, C.-W. Chang, H.-H. Chou, L.-L. Wu, C.-H. Chiang, S.S.-F. Yu, Tetrahedron Lett. 52 (2011) 2950–2953. G.B. McGaughey, M. Gagné, A.K. Rappé, Proteins 273 (1998) 15458–15463. K.-I. Cho, D. Kim, D. Lee, Nucleic Acids Res. 37 (2009) 2672–2687. T. Steiner, G. Koellner, J. Mol. Biol. 305 (2001) 535–557. V. Soskić, J. Joksimović, Curr. Med. Chem. 5 (1998) 493–512. M.T. Reetz, L.-W. Wang, M. Bocola, Angew. Chem. Int. Ed. 45 (2006) 1236–1241. M.T. Reetz, J.D. Carballeira, Nat. Protoc. 2 (2007) 891–903. H. Li, T.L. Poulos, Nat. Struct. Mol. Biol. 4 (1997) 140–146. S. Purser, P.R. Moore, S. Swallow, V. Gouverneur, Chem. Soc. Rev. 37 (2008) 320–330. H.J. Böhm, D. Banner, S. Bendels, M. Kansy, B. Kuhn, K. Müller, U. Obst-Sander, M. Stahl, ChemBioChem 5 (2004) 637–643. B.K. Park, N.R. Kitteringham, P.M. O'Neill, Annu. Rev. Pharmacol. Toxicol. 41 (2001) 443–470. K.S. Eble, J.H. Dawson, J. Biol. Chem. 259 (1984) 14389–14393. S. Kadkhodayan, E.D. Coulter, D.M. Maryniak, T.A. Bryson, J.H. Dawson, J. Biol. Chem. 270 (1995) 28042–28048. G.B. Crull, J.W. Kennington, A.R. Garber, P.D. Ellis, J.H. Dawson, J. Biol. Chem. 264 (1989) 2649–2655. G.B. Crull, J.V. Nardo, J.H. Dawson, FEBS J. 254 (1989) 39–42. M.J. Cryle, N.J. Matovic, J.J. De Voss, Tetrahedron Lett. 48 (2007) 133–136. M.J. Cryle, J.J. De Voss, Tetrahedron Asymmetry 18 (2007) 547–551. G. Truan, M.R. Komandla, J.R. Falck, J.A. Peterson, Arch. Biochem. Biophys. 366 (1999) 192–198. Q.S. Li, U. Schwaneberg, P. Fischer, R.D. Schmid, Chem. Eur. J. 6 (2000) 1531–1536. C. Li, S. Shaik, RSC Adv. 3 (2013) 2995–3005. A. Hegde, D.C. Haines, M. Bondlela, B. Chen, N. Schaffer, D.R. Tomchick, M. Machius, H. Nguyen, P.K. Chowdhary, L. Stewart, C. Lopez, J.A. Peterson, Biochemistry 46 (2007) 14010–14017. E.V. Anslyn, D.A. Dougherty, Modern Physical Organic Chemistry, University Science Books, Sausalito, 2006. V. Budarin, J.H. Clark, J.J.E. Hardy, R. Luque, K. Milkowski, S.J. Tavener, A.J. Wilson, Angew. Chem. Int. Ed. 45 (2006) 3782–3786. S. Doi, J.H. Clark, D.J. Macquarrie, K. Milkowski, Chem. Commun. (2002) 2632–2633. J.B. van Beilen, D. Penninga, B. Witholt, J. Biol. Chem. 267 (1992) 9194–9201. P.S. Aguilar, D. De Mendoza, Mol. Microbiol. 62 (2006) 1507–1514. D.H. Dyer, K.S. Lyle, I. Rayment, B.G. Fox, Protein Sci. 14 (2005) 1508–1517. Y. Lindqvist, W. Huang, G. Schneider, J. Shanklin, EMBO J. 15 (1996) 4081–4092. J.E. Guy, E. Whittle, D. Kumaran, Y. Lindqvist, J. Shanklin, J. Biol. Chem. 282 (2007) 19863–19871. David R. Lide, CRC Handbook of Chemistry and PhysicsCRC Press/Taylor and Francis, Boca Raton, FL, 2010. S.H. Malca, D. Scheps, L. Kuhnel, E. Venegas-Venegas, A. Seifert, B.M. Nestl, B. Hauer, Chem. Commun. 48 (2012) 5115–5117. I.N.A. Van Bogaert, S. Groeneboer, K. Saerens, W. Soetaert, FEBS J. 278 (2011) 206–221. P. Meinhold, M.W. Peters, A. Hartwick, A.R. Hernandez, F.H. Arnold, Adv. Synth. Catal. 348 (2006) 763–772. N. Kawakami, O. Shoji, Y. Watanabe, Angew. Chem. Int. Ed. 50 (2011) 5315–5318.

Kinetic and mechanistic studies on the reaction of melilotate hydroxylase with deuterated melilotate.

Aliphatic and aromatic hydrocarbons in indoor air.

Highly Enantioselective Oxidation of Nonactivated Aliphatic C-H Bonds with Hydrogen Peroxide Catalyzed by Manganese Complexes.

CH-π hydrogen bonds in biological macromolecules.

Mechanistic insights into xanthine oxidoreductase from development studies of candidate drugs to treat hyperuricemia and gout.

Transpositional recombination: mechanistic insights from studies of mu and other elements.

Mechanistic and kinetic studies on the OH-initiated atmospheric oxidation of fluoranthene.

Oxidation of aliphatic hydroxylamines in aqueous solutions.

Structural and mechanistic insights on nitrate reductases.

Structure and Mechanism of Isopropylmalate Dehydrogenase from Arabidopsis thaliana: INSIGHTS ON LEUCINE AND ALIPHATIC GLUCOSINOLATE BIOSYNTHESIS.

Bidentate cycloimidate palladium complexes with aliphatic and aromatic anagostic bonds.

Weak acidity of vinyl CH bonds enhanced by halogen substitution.

Carcinogenicity studies on halogenated hydrocarbons.

Chemical and enzyme-mediated oxidation of the serotonergic neurotoxin 5,7-dihydroxytryptamine: mechanistic insights.

Chlorinated aliphatic hydrocarbons promote lipid peroxidation in vascular cells.

Mechanistic Studies on RNA Strand Scission from a C2'-Radical.

Mechanistic investigations of the rhodium catalyzed propargylic CH activation.

Mechanistic and therapeutic insights gained from studying rare skeletal diseases.

Neutron diffraction studies on selectively deuterated phospholipid bilayers.

Mechanistic insights from substrate preference in unsaturated glucuronyl hydrolase.

Peroxisomal beta-oxidation: insights from comparative biochemistry.

The association of systemic oxidative stress with insulin resistance: mechanistic insights from studies in Bartter's and Gitelman's syndromes.

A well-defined NHC-Ir(iii) catalyst for the silylation of aromatic C-H bonds: substrate survey and mechanistic insights.

Distributions and sources of petroleum, aliphatic hydrocarbons and polycyclic aromatic hydrocarbons (PAHs) in surface sediments from Bohai Bay and its adjacent river, China.