Cytochrome P450—The Wonderful Nanomachine Revealed through

Jan 11, 2019 - ... we shall provide here an overview of the factors that cause all of the ... aspects of two-state reactivity were reviewed previously...
0 downloads 0 Views 10MB Size
Article pubs.acs.org/accounts

Cite This: Acc. Chem. Res. XXXX, XXX, XXX−XXX

Cytochrome P450The Wonderful Nanomachine Revealed through Dynamic Simulations of the Catalytic Cycle Kshatresh Dutta Dubey* and Sason Shaik* Institute of Chemistry, The Hebrew University of Jerusalem, Givat Ram Campus, 91904 Jerusalem, Israel

Downloaded via IOWA STATE UNIV on January 11, 2019 at 18:52:07 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

CONSPECTUS: This Account addresses the catalytic cycle of the enzyme cytochrome P450 (CYP450) as a prototypical biological machine with automatic features. CYP450 is a nanomachine that uses dioxygen and two reducing and two proton equivalents to oxidize a plethora of molecules (so-called substrates) as a means of supplying bio-organisms with essential molecules (e.g., brain neurotransmitters, sex hormones, etc.) and protecting biosystems against poisoning. An enticing property of CYP450s is that entrance of an oxidizable substrate into the active site initiates a series of events that constitute the catalytic cycle, which functions “automatically” in a regulated sequence of events culminating in the production of the oxidized substrates (e.g., hydroxylated, epoxidized, etc.), oftentimes with remarkable stereo- and regioselectivities. It is timely to demonstrate how theory uses molecular dynamics (MD) simulations and quantum-mechanical/molecular-mechanical (QM/MM) calculations to complement experiments and elucidate the choreography by which the protein regulates the catalytic cycle. CYP450 is a heme enzyme that contains a ferric ion (FeIII) coordinated by a porphyrin ligand, a water molecule, and a cysteinate ligand that is provided by a strategic residue of the encapsulating protein. While many of the individual steps are sufficiently well-understood, we shall provide here an overview of the factors that cause all of the steps to be sequentially coordinated. To this end, we use examples from three different CYP450 enzymes: the bacterial ones CYP450BM3 and CYP450CAM and the mammalian enzyme CYP4503A4. The treatment is limited to the catalytic cycle, as aspects of two-state reactivity were reviewed previously (e.g., Shaik, S.; et al. Chem. Rev. 2005, 105, 2279). What are the principles that govern the seeming automatic feature? For example, how do substrate entrance and binding gate the enzyme? How does the reductase attachment to the enzyme affect the next steps? What triggers the attachment of the reductase? How does the electron transfer (ET) that converts FeIII to FeII occur? Is the ET coordinated with the entrance of O2 into the active site? What is the mechanism of the latter step? Since the entrance of the substrate expels the water molecules from the active site, how do water molecules re-enter to form a proton channel, which is necessary for creating the ultimate oxidant Compound I? How do mutations that disrupt the water channel nevertheless create a competent oxidant? By what means does the enzyme produce regio- and stereoselective oxidation products? What triggers the departure of the oxidized product, and how does the exit occur in a manner that generates the resting state ready for the next cycle? This Account shows that the entrance of the substrate triggers all of the ensuing events.

1. INTRODUCTION Cytochrome P450 (CYP450) is a versatile enzyme family that catalyzes a variety of essential oxidation reactions, including hydroxylation, epoxidation, sulfoxidation, C−C bond cleavage, and desaturation reactions.1−3 An enticing property of the CYP450 machinery is the catalytic cycle1 depicted in Figure 1 for alkane (R−H) hydroxylationwherein all of the steps are automatically coordinated. CYP450 is thus a biological nanomachine that functions in a tightly choreographed sequence. The active site of CYP450 is located in the hydrophobic core of the protein and is connected to the surface of the enzyme via access channels4 through which substrates and O2 molecules enter, products exit, and water molecules flow. There are also strategic water gates, which are operated by the propionate side chains of the heme and their salt bridge partners and enable timely water influx.5 As such, the P450 © XXXX American Chemical Society

nanomachine is an open system that communicates with the outside molecular world via channels and gates that intervene in the catalytic cycle (Figure 1). The cycle is initiated when the substrate (R−H) enters the active site and interacts with the resting state (I). This drives off the water molecules from the pocket2c and detaches the aqua ligand of I, thus forming the high-spin FeIII−heme complex (II). II has a more positive redox potential,6 which enables electron transfer (ET) from the reducing partner, cytochrome P450 reductase (CPR), to II, which gets reduced to the ferrous FeII complex (III). III, which is a good O2 binder, rapidly takes up an O2 molecule and is transformed to the oxyferrous complex (IV). The latter, being a good electron acceptor, is reduced again by CPR and is transformed to the Received: September 16, 2018

A

DOI: 10.1021/acs.accounts.8b00467 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 1. CYP450 catalytic cycle during R−H hydroxylation. The heme is shown in the middle. Species entering the catalytic cycle are indicated by green arrows and departing species by red arrows. Questions are posed in red. Reproduced with permission from Prof. Jilai Li.

Figure 2. (a) Active site of CYP450BM3 and selected strategic residues and their roles. The protein is shown as thin strips. (b) The solvated enzyme. Water molecules are drawn in red.

molecule; the enzyme is restored to its resting state and is ready for another turnover. This sequence of events is shared by many CYP450s. While each step in Figure 1 is understood,1−3,6−8 an overview of the factors that cause all of the steps to be sequentially choreographed is needed. Therefore, it is important to understand the principles that govern the seeming automation of these nanomachines, as expressed by the questions in Figure 1. This Account addresses the questions and provides an overview of the governing principles on the basis of recent studies combining molecular dynamics

peroxo complex V (also see the references in the Supporting Information (SI)). Somewhere near this junction, the water molecules that left the active pocket re-enter via a water gate and form a water channel that protonates V to give VI. The latter, known as Compound 0 (Cpd 0), is still debated as a putative oxidant.7 The negatively charged Cpd 0 is a good Lewis base. Hence, it accepts an additional proton, releases a water molecule, and forms Compound I (Cpd I, VII). Cpd I is the ultimate oxidant. It abstracts a hydrogen atom from R−H, forming intermediate VIII, which by R· rebound generates the alcohol (R−OH). The R−OH then exits the pocket and is replaced by a water B

DOI: 10.1021/acs.accounts.8b00467 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research (MD) and quantum-mechanical/molecular-mechanical (QM/ MM) calculations on a few CYP450s.4a,5,9 Strategies

Figure 2a uses CYP450BM3 to exemplify the multifaceted computational problem. The species in the center are the active heme species and the fatty acid substrate. This action core is surrounded by strategic residues: At the top left, we find an acid−alcohol pair (Glu267-Thr268), which organize a water chain that transfers the necessary protons (Figure 1). At the upper right, Gln73 is one of three residues that bind the substrate via its carboxylate head.9a,c Further down from Gln73, the residue Phe87 curls the fatty acid chain, exposing its ω−1 ω−3 positions to oxidation. At the bottom, Prop7− Arg398 interactions (similar to other CYP450s) are likely to be responsible for gating of water into the active site. Finally, near the cysteine ligand (Cys400) there are positively charged residues (e.g., Lys391) to which the reductase can attach and transfer electrons to the heme. How does one handle all of these features? Here is a brief outline of the approach (see the SI for details): Features concerned with substrate binding, O2 entrance, reductase attachment, and gating are handled by MD simulations.9 The starting structures are taken from the Protein Data Bank (PDB) and are treated by preparatory steps,3b including protonation and solvation of the enzymes, as shown in Figure 2b, such that one ends up with 30000−40000 atoms. Water channel formation and reorganization require typically 100 ns. Substrate binding requires 350−400 ns, while protein−protein interactions (e.g., reductase−CYP450) require 1000−1500 ns (1−1.5 μs). The longer simulations are performed using a multitrajectory approach in which the simulation is restarted each 200 ns with a random velocity option. Conclusions are drawn from four different longer-scale simulations.9a To treat chemical events of the cycle, we select from the MD trajectories CYP450 conformations, so-called “snapshots”. We subsequently subject these snapshots to QM/MM calculations3 (also see the SI references), which compute the active species and its chemical events and provide geometries, electronic structures, mechanisms, and energies of the species in their native protein environment. We start with substrate entry, which initiates the catalytic cycle in Figure 1. Most of our examples use CYP450BM3, which is a bacterial enzyme that is linked to the reductase domain via a peptide chain, as shown in Figure 3. A few other CYP450s are used too.4,5,9

Figure 3. Domains of CYP450BM3.

2.1. Substrate-Binding-Induced Choreography

What attracts the NPG to probe the open mouth of the substrate access channel is the electrostatic attraction between the NPG’s carboxylate head and the residues Arg47, Tyr51, and Gln73.9c This initiates substrate entrance. The channel involves the structural units marked in Figure 4a,b: the A-helix, β1-sheet, and F-helix residues. During MD, the SF state (Figure 4a) exhibited preparatory choreography, which moves these structural units >6 Å to the right and further opens the channel, thus providing an unperturbed doorway for the substrate. The follow-up MD simulations in Figure 4b generated the SB state. With the NPG substrate docked into the open mouth, we observed that the protein closes on NPG by an inward movement of the same protein units by 11 Å. Thus, NPG slides into the position of the aqua ligand of the resting state. The MD simulation of the already closed SB structure11a,b does not show any major conformational rearrangement. Taken together, these results provide compelling evidence that the conformational transition f rom the open-channel form of the SF state to the closed-channel form of the SB state is instigated by the substrate. Inspection of the MD trajectories in the closed state (Figure 4c) reveals a triad of interactions between the substrate and channel residues, namely, Pro25 (A-helix), Glu43 (β1-sheet), and Leu188 (F/G loop), which mediate closure of the channel. These interactions bring closer the structural subunits that constitute the substrate access channel and thereby close the cavity. This substrate binding mechanism appears in other CYP450s.11 Thus, the starting enzyme is an open state, and as the substrate enters the cavity, the enzyme closes its mouth and “swallows” the substrate, which after oxidation gets released, much like the Biblical story of the prophet Jonah being swallowed by the whale, wherein he is “transformed” and released to fulfill his duty in the city of Nineveh. In summary, the entrance of the substrate is the switch that triggers the ensuing catalytic machine.

2. WHAT TRANSPIRES FOLLOWING SUBSTRATE ENTRY? For CYP450BM3, the PDB contains substrate-free (SF)10 and substrate-bound (SB)11a,b structures of the heme domain. These structures reveal a major conformational rearrangement due to substrate entry. It is essential to answer the following questions: Are these conformational changes instigated by the substrate? If so, what are the mechanisms by which the changes transpire? Finally, how does this prepare for the next step? To answer these questions, we targeted CYP450BM3 with the substrate N-palmitoglycine (NPG) and performed three sets of long-duration MD simulations involving (a) the SF heme domain, which possesses an open cavity; (b) the open-cavity heme domain with the substrate docked; and (c) the SB heme domain having the closed cavity.9c C

DOI: 10.1021/acs.accounts.8b00467 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 5. The most prominent (green and pink) and less prominent (blue) water channels in the (top) SF and (bottom) SB states of CYP450BM3. In the SB state, the substrate is accommodated within the blue channel, thus relocating the three channels.

3. HOW DOES THE REDUCTASE GET A SIGNAL TO TRANSFER ELECTRONS TO THE HEME? The CPR in CYP450BM3 is linked via a peptide chain to the CYP450 domain (see Figure 3) and is a complex entity made from the three reducing domains, as shown in Figure 6a. The ultimate reducing unit is the flavin mononucleotide (FMN) domain, in which the reductant is the semiquinone state (SQ−), shown in Figure 6b. The above water content analysis demonstrates that substrate entry converts the hexacoordinated low-spin FeIII into a pentacoordinated high-spin state (I and II, respectively, in Figure 1). In the absence of a sixth ligand, the high-spin FeIII complex has a more positive redox potential,6 which favors electron acceptance from SQ−. Thus, while the linked CPR may approach the heme frequently, ET transpires only when the redox thermodynamics allows it and when the FMN−heme distance is rather short. Below we show how the substrate enables these requisite conditions for the ET event.

Figure 4. Substrate entrance in CYP450BM3. (a) Widening of the channel in the SF state. (b) Closure of the channel due to binding of the substrate. (c) Interactions that govern the closure of the channel. Adapted from ref 9c. Copyright 2015 American Chemical Society.

2.2. Following Substrate Binding

3.1. Substrate-Binding-Induced ET

As already mentioned, substrate binding detaches the water ligand and brings about the departure of water molecules from the cavity. This in turn increases the entropy and adds a strong driving force for substrate binding.2c Furthermore, the water loss triggers reduction of the heme and what follows thereafter. The presence/absence of water inside the heme cavity was studied using MD simulations of the SF and SB states of CYP450BM3.9b,c Figure 5 shows the most prominent channels (green and pink), which in the SF state pass through the heme’s distal side, while the less prominent channel (blue) passes through the proximal side. Furthermore, the green/pink channels of the SF state exhibit significant water content, including a permanently present molecule close to FeIII (see p S4 in the SI). In the SB state, all of these channels are relocated as a result of substrate binding. The green and pink channels shift to the proximal side, near the heme’s propionate side chains. The blue channel passes through the FeIII coordination axis and accommodates the substrate, which detaches the water ligand of FeIII and depletes the other water molecules. Indeed, the MD simulation of the SB state shows no water near the iron center, hence preparing the heme for reduction by CPR.

As shown in Figure 7, the ET step faces an initial major difficulty. The crystal structure of the heme domain−FMN domain complex shows that these domains are separated by 17.4−18.4 Å.9b,12 The ET rate at these distances is significantly more sluggish than the observed rate constants under physiologically relevant conditions (see ref. S64),13,14 e.g., kET = 100−300 s−1 for CYP450BM3.14 We therefore used CYP450BM3 and the FMN part of the CPR and performed six distinct MD simulations.9a The MD simulation of the SF state (Figure 8a) showed that the acceptor (heme) and the donor (FMN) maintain a large distance, >18.4 Å, throughout the course of the MD simulation. Thus, in the absence of the substrate, the ET process is inefficient. However, when the substrate was fed into the heme−FMN system, the distance was shortened to 12 Å (Figure 8b); further shortening was achieved when FMN− was used. Subsequent QM/MM optimizations (Figure 8c) brought the distance down to 8.8 Å. The QM/MM calculations9a further revealed a weakly endothermic ET event (4.6 kcal/ mol). To assess the role of these conformational changes, we used the Marcus−Sutin equation15 to calculate the ET rates for the SF state (wherein the donor−acceptor distance is ∼18 Å) and D

DOI: 10.1021/acs.accounts.8b00467 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 6. (a) CPR domains of CYP450BM3. (b) Structure of the SQ− reductant in the FMN domain.

the SB state (wherein the donor−acceptor distance is ∼8.8 Å). Our calculated kET in the SB state was within the range of 107−291 s−1, which fits the experimental ET rates of the wildtype (WT) enzyme with bound substrates.14 The calculated kET in the SF state was lower by 5−6 orders of magnitude and much too low relative to the observed rates.14 The substrate is crucial! 3.2. What Can a Substrate Mastermind?

It is therefore clear that substrate binding brings about large conformational11 and other changes that cause the FMN− to nestle closer to the heme and induce a successful ET process. It is instructive to try and understand first why the FMN− heme distance in the SF state is so long This is determined by the choreography of the three helices in Figure 3: the C-helix (in green), which resides in the heme domain; the α1-helix (orange-brown), which belongs to the FMN domain; and the I-helix (blue), which is in the heme domain. The substrate enters between the heme and the I-helix and enhances the interaction between the C-helix and I-helix, which causes the FMN to move 10 Å toward the heme and deliver the electron to the FeIII center. Some details of this choreography can be understood from Figures 9 and 10. Figure 9 shows the C and α1 orientations

Figure 7. Structure of the heme and FMN domains (in green) in the SF state. The FMN−Fe distance of 18.4 Å is from MD simulations.

Figure 8. (a, b) Donor−acceptor distances in the (a) SF and (b) SB states. (c) QM/MM-optimized geometry of the SB−FMN− complex. E

DOI: 10.1021/acs.accounts.8b00467 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

the ferrous complex III (Figure 1). Initially the FeII bonding site is occupied by the substrate, which raises the following question: how does O2 displace the substrate and create the oxyferrous complex IV? The answer was provided by MD simulations using dioxygen efflux for the ferrous species of CYP450BM3 based on a three-site O2 model (which reproduces its quadruple moment; see p S3 in the SI).16 Thus, as shown in Figure 11a, since the substrate occupies the heme binding site Figure 9. (a) Interactions between the α1-helix of CPR (orange) and the C-helix of the heme domain (green) in the SF state at the initial MD stage. (b) Interaction at the end of the MD simulation. The Hbond switch should be noted.

Figure 11. O2 entrance. (a) Initial snapshot. The proximity of the substrate to the iron center leaves no room for O2 molecules. (b) Final snapshot. The increased O2 density pushes the substrate away from iron center, while the space is occupied by O2-450 and O2-466.

at the onset of the simulation, only a low O2 density is detected near the heme (see the plot of O2 density vs simulation time in Figure S4). However, at longer times, as shown in Figure 11b, the substrate is displaced, providing room for O2 molecules, several of which approach to within 2.5 Å of the iron center. The accumulated density of O2 molecules is correlated with the deviation of the end terminus of the fatty acid from the heme center (Figure S4 and Figure 11b). Stated differently, this correlation indicates that the O2 efflux exerts “pressure” that displaces the substrate. This is reminiscent of the recent finding17 that the O2 concentration affects the selectivity of camphor hydroxylation by P450CAM.

Figure 10. Interaction of the C-helix and the I-helix. The enhanced Lys98−Asp250 interaction upon substrate binding should be noted.

during the initial and final MD stages for the SF state. Initially (Figure 9a), the two helices are mutually perpendicular due to H-bonding between Glu494 (on α1) and His100 (on C). At the end of the simulation, the two helices reorganize to the more stable (by 12.3 kcal/mol) parallel orientation (Figure 9b) as a result of the H-bonding switch whereby Glu494 gets tightly bound to Lys97 (on C). This new interaction pushes the two helices apart and uplifts the far end of the α1 helix, which in turn pulls the FMN away from the heme (to 18.4 Å). However, in the SB state, the C-helix and α1-helix prefer the perpendicular conformation. Figure 10 shows that this happens because the bound substrate causes a strong interaction between the C-helix and the I-helix. Thus, the substrate binds between these helices, and consequently, the I-helix kinks and interacts more strongly with the C-helix via a tight H-bonding interaction (Lys98 on I with Asp250 on C), which locks the Chelix perpendicular to α1. This interaction also causes a ripple of H-bond changes that stretches from the C-helix to residues in the α1-helix and onward to residues near the FMN, on the ET loop that connects to the cysteine ligand of the heme. These interactions, especially Asn489−Gly396 and Asn537− Pro386, pull the cysteine ligand toward the FMN, thus establishing the requisite short ET distance. Thus, substrate binding (Figures 4 and 7−10) induces extensive motions (10−11 Å) of protein pieces; these motions are functional (leading to water loss and enabling ET) and propagate the catalytic cycle. We insisted on the above detailed analysis to make the point that these outstanding outcomes are merely combinations of many small interactions, illustrating the hand of evolution in constructing this nanomachine.

5. REDUCTASE-MEDIATED H+ GATING As indicated in Figure 5, substrate entry displaces water from the cavity. How can protonation of the O2 ligand to yield Cpd 0 transpire in the absence of water?2,3,5 This requires a mechanism whereby water re-enters the heme pocket via one of the aqueducts of the enzyme (Figure 5). This puzzle was addressed by MD simulations of three CYP450 members: CYP4503A45 and CYP450CAM and CYP450BM3.9a 5.1. Water Gating in CYP4503A4

Earlier MD simulations5 showed that the aqueduct passes near the propionate (Prop) side chains of the heme, specifically Prop7, which forms a salt bridge with Arg375. This salt bridge serves as a water gate: it is closed when the substrate enters the heme site and blocks the water transport. Figure 12a shows the closed gate, wherein the Prop7−Arg375 distances are ∼2.7 and 2.6 Å and both salt-bridged partners are further H-bonded by Ser437 (on the C-helix). Thus, how do the water molecules enter and what determines their entrance timing? The answer to this question emerged when the simulation was repeated in the presence of the FMN domain, which attaches to the heme via the oppositely charged interface of the heme domain. Thus, as shown in Figure 12b, upon attachment, residue Met490 (along with Asn489, not shown) of FMN interacts with the gatekeeper Ser437 and pulls it. In turn, Ser437 pulls down Arg375, and “Open Sesame”, the gate

4. O2 ENTRY O2 is a small molecule that can easily diffuse into the active site. However, the O2 entrance becomes functional only after the reduction step and the formation of the good O2 binder, F

DOI: 10.1021/acs.accounts.8b00467 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Arg289 gate (p S4 in the SI), including the formation of a water chain connected to the D251−T252 residue pair. The crucial role of Prop7 was demonstrated by Hayashi et al.,19 who mutated CYP450CAM by removing the propionate side chain and found the mutant to be dysfunctional. These studies provide a unified mechanism for closing and opening of the aqueduct that may typify many other CYP450s.5 It follows that substrate binding not only brings the reductase and heme domains closer but also creates an organized water chain connecting the two domains as well as the crucial acid−alcohol pair, thereby playing a major role in coupling the ET and proton transfer events that generate Cpd I. Clearly, substrate binding dominates the operation of the CYP450 machine by inducing an entropy rise2c and switchable weak interactions, which make the cycle automatically choreographed.

Figure 12. Salt bridge gate between Prop7 and Arg375 in CYP4503A4. (a) S437 maintains a closed gate by holding the salt bridge partners by H-bonding. (b) The open gate in the presence of the FMN domain due to the interaction of M490 with S437, which pulls down S437 and Arg375 and opens the gate. Adapted from ref 5. Copyright 2010 American Chemical Society.

6. MUTATION-INDUCED WATER CHANNEL FORMATION The conversion of Cpd 0 to Cpd I (Figure 1) involves protonation of the distal oxygen and liberation of a water molecule, resulting in Cpd I. The process has been amply studied experimentally and computationally.3 However, our MD study9b shows results with a potential for a paradigm shift. A central mechanistic paradigm1,2,20,21 in CYP450 considers the acid−alcohol pair, which organizes the water channel that supplies protons for the formation of the ultimate oxidant Cpd I. Thus, a mutation of Thr to Ala is thought to disrupt the proton shuttle mechanism. Whenever a T → A mutant performs oxidation, the reactivity is attributed to Cpd 0 as a “second oxidant”. However, Cpd 0 is a negatively charged species, and calculations have shown it to be a poor oxidant.3,22 Since water molecules are mobile and the active site has considerable plasticity, we may wonder whether an alternative water channel can be generated in the mutant. The one we found recently9b for the T268A mutant of CYP450BM3 shows that a water channel is generated even in this paradigmatic mutant. Figure 14 shows the results for oxidation of dimethyl((4-methylsulfanyl)phenyl)amine (Sub1) by WT CYP450BM3 and its T268A mutant.7 The WT enzyme oxidizes the two sites of Sub1 with S/Me regioselectivity of 15/1, whereas the T268A mutant shows a S/Me ratio of 60/1.

remains wide open, and water molecules flow into the heme cavity to form an organized water chain stretching between the acid−alcohol pair (Figure S1), ready to deliver the requisite protons for making Cpd 0 and Cpd I.1−3,5 Thus, the key that opens the gate is the FMN, which causes Ser437 to switch interaction partners f rom Prop7−Arg375 to Met490−Ser437− Arg375. 5.2. Water Gating in CYP450BM3 and CYP450CAM

The above water-gating mechanism5 was by-and-large verified using much longer simulation times (1 μs) for CYP450BM39a and CYP450CAM (p S4 in the SI). Figure 13 illustrates the water flow in CYP450BM3 in the presence of FMN.9a The left panel of Figure 13 reveals a closed aqueduct in the SF state and a diffuse water cloud distant from the heme. In contrast, in the SB state (Figure 13 right), the Prop7−Lys69 gate opens up, generating a dense, organized water aggregate that stretches from the open aqueduct all the way to the FMN domain and is connected to the acid−alcohol pair, which shuttles the protons to the oxyferrous complex. The water-gating mechanism induced by the reductase putidaredoxin in CYP450CAM18 was also studied and revealed similar mechanisms for closing and opening of the Prop7−

Figure 13. Water cloud connecting the P450 domain to FMN in the SF and SB states. Blue patches indicate lower density; the region with the highest density of water is highlighted in red. Adapted from ref 9a. Copyright 2017 American Chemical Society. G

DOI: 10.1021/acs.accounts.8b00467 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

decoupling (due to Fe(H2O2) formation and departure of the H2O2), as observed by experiment.7 The reason for the higher S/Me regioselectivity in the T268A mutant can be deduced from Figure 16a, which shows that while the sulfur atom of Sub1 in the WT enzyme is distant from Cpd I because of the π interactions with F87, in the T268A mutant the water “stream” pushes Sub1 away from F87, causing the sulfur moiety to move closer to Cpd I. Figure 16b shows a detailed view of Sub1 in the T268A mutant, which illustrates that Sub1 is barricaded by water molecules and neighboring residues that together push the S end toward Cpd I of the mutant enzyme. Indeed, QM/MM calculations9b showed that in the T268A mutant the sulfoxidation barrier is lowered while the barrier for Me hydroxylation is raised compared with the WT enzyme. Since Cpd I is the oxidant in both cases, the KIE (H/D) remains unchanged and typical of Cpd I.9b Therefore, the study provides compelling evidence that Cpd I does form in T → A mutants and is the sole oxidant in the CYP450BM3 catalytic cycle. The take-home lesson of this story is that the assumption that a single-site mutation is a delicate surgery that helps decipher the precise mechanism may occasionally be wrong. In truth, the active pocket is f lexible and adaptive and f ull of opportunities for alternative routes of generation of its ultimate oxidant.

Figure 14. Regioselectivity of oxidation7 of Sub1 by WT CYP450BM3 and its T268 mutant and KIE values for Sub1 and Sub2.

At the same time, the kinetic isotope effect (KIE) for Me oxidation remains constant at ∼2, which is identical to the KIE for Sub2. The fact that the regioselectivity changes whereas the KIE remains constant were interpreted7 as evidence for the coexistence of Cpd 0 and Cpd I: Cpd 0 performs sulfoxidation whereas Cpd I hydroxylates the Me group, and therefore, the S/Me ratio reflects the Cpd 0/Cpd I relative abundance. Nevertheless, the authors7 did not rule out the possibility that the results may reflect active-site plasticity effects on the reactivity of Cpd I. Since QM and QM/MM calculations have shown that Cpd 0 is a poor oxidant even for sulfoxidation,22 we performed two sets of 200 ns MD simulations of Cpd 0 (for the WT and the T268A mutant) in the presence of Sub1 and characterized the available water and protonation channels. To our surprise, we found two productive water channels, one for the WT enzyme and one for its mutant. Figure 15 traces the two channels. The

7. REGIO- AND STEREOSELECTIVTY As we saw above, MD simulations supplemented by QM/MM calculations revealed that the “automatic” operation of the catalytic cycle and the regioselectivity of, e.g., Sub1 oxidation7 all originate in a collection of weak interactions. A comparison of the regio- and stereoselectivities of fatty acid oxidation by CYP450SPα, CYP450BSβ, and CYP450BM3 leads to the same conclusions.9c,d 7.1. Regioselectivity

Figure 17 illustrates some of these small differences that exert entirely different consequences. In CYP450BM3 (Figure 17a), the carboxylate head of the fatty acid is bound near the enzyme’s surface by Arg47, Gln73, and Tyr51, whereas in CYP450SPα (Figure 17b), the carboxylate head is salt-bridged by Arg241 deep down near Cpd I. Hence, in CYP450BM3 the fatty acid tail slides toward the heme during the MD simulation, and Phe87 causes the tail to curl and expose its ω−1 ω−3 methylene groups to Cpd I, which hydroxylates only these sites in a 36:30:34 ratio.23 The ω end is completely blocked by Phe87, but the F87A mutation exposes primarily the ω end to Cpd I. Indeed, the F87A mutant exhibits >90% ω regioselectivity.23 In CYP450SPα (Figure 17b), the grip of the carboxylate head by Arg241 and Phe288 enforces oxidation of the α-CH2 group24 while excluding any oxidation of β-CH2. In the related CYP450BSβ enzyme,25 the holder of the carboxylate head is Arg242 (one site further compared with Arg241 in CYP450SPα), which provides access also to the β-CH2 group; however, the lack of Phe288 causes the regioselectivity to be lower than in CYP450SPα (>99%).26

Figure 15. Water pathways in the WT enzyme (blue) and the T268A mutant (red) during MD simulations. Adapted with permission from ref 9b. Copyright 2017 Royal Society of Chemistry.

WT enzyme has a traditional channel (in blue), which is guided by the Glu267−Thr268 pair. The T268A mutation indeed destroyed this channel but created a new one (in red on the right side). This channel is also initiated from Glu267 and connected to the distal oxygen of Cpd 0 through a few water molecules. Subsequent QM/MM calculations9b showed that the two channels gave rise to Cpd I formation with rather low barriers, and the T268A channel also resulted in partial

7.2. Enantioselectivity

In CYP450BM3, the MD trajectory also reveals the origins of the enantioselectivity of fatty acid hydroxylation. As a result of the interaction with Phe87, the pro-R hydrogens in ω−1 and ω−2 are closer to the Fe−O center compared with the pro-S hydrogens (Figure 17a). Phe87 was found to be responsible H

DOI: 10.1021/acs.accounts.8b00467 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 16. (a) Superimposed representations of the water channels in WT CYP450BM3 (bluish) and the T268A mutant (reddish). In the WT enzyme, Phe87 (gray) keeps the yellow S moiety of Sub1 away from Cpd I, whereas in the T268A mutant, S is close to Cpd I. (b) Detailed view of Sub 1 in the T268A mutant. The barricading of Sub1 by water molecules and some resides in the site should be noted.

Nothing magical occurs, but there are cues due to the hand of the tinkerer: evolution.

9. PROSPECTS CYP450s are so similar yet so different. The more we analyzed the MD trajectories in the catalytic cycle of CYP450, the more apparent it became that the most fascinating features of the enzyme are rooted in the chemistry of weak interactions. These are the cues that turn on the automatic cycle. Nature is a great tinkerer “who” manages to make nanomachines operated by weak interactions! The take-home lesson from the MD simulations we carried out is the recognition that very few residues and some water molecules are suf ficient to create f unction. This may account for the huge diversity of the CYP450 superfamily. It also highlights the importance of MD simulations in future enzymatic research and the importance of directed evolution, in which the scientist plays the role of the tinkerer.27,28

Figure 17. Schematic representation of residues controlling the regioselectivities of fatty acid oxidations by (a) CYP450BM3 and (b) CYP450SPα.

also for the R enantioselectivity at ω−3.9c These conclusions were verified by QM/MM calculations of the respective hydroxylation barriers.9c The S enantioselectivity during α-CH2 hydroxylation in CYP450SPα was also revealed by MD simulations,9d which, as illustrated in Figure 17b, showed that Pro242 locks the pro-S α-CH bond closer to Cpd I (2.6 Å) and keeps the pro-R α-CH bond farther away (3.3 Å), thus creating a large difference (∼9 kcal/mol) in the corresponding energy barriers.9d The above discussion shows that the selectivity of a few related CYP450s is a function of active-site organization and key residues therein. Therefore, the emergence of the f unctions for different P450 members is predef ined by the active-site architecture shaped during evolution!



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.accounts.8b00467. Details of MD simulations, calculations of ET rates, and a list of additional references (PDF)



AUTHOR INFORMATION

Corresponding Authors

*Tel.: +972 (0)2 658 5909. Fax: +972 (0)2 658 4033. E-mail: [email protected]. *E-mail: [email protected].

8. PRODUCT EXIT The final step is the departure of the oxidized product. The Fe−O bond between the product (e.g., R−OH in Figure 1) and the ferric ion is weaker and longer (2.1−3.3 Å) than the FeIII−OH2 bond (2.0−2.3 Å).3b Therefore, in most cases the product departure starts spontaneously because of the presence of water molecules that displace R−OH, which leaves through an exit channel. A steered MD study4a of the exit of the hydroxylation products of, e.g., testosterone by CYP4503A4 showed that the choice exit channels are lined up with π residues, which help unleash the Fe−O bond of the product by π stacking and expel the substrate to the surface. Again, we see that the seemingly autonomous event of product release is a story of weak interactions and many available exit channels.

ORCID

Kshatresh Dutta Dubey: 0000-0001-8865-7602 Sason Shaik: 0000-0001-7643-9421 Notes

The authors declare no competing financial interest. Biographies Kshatresh Dutta Dubey was a postdoctoral researcher in Jerusalem and a graduate of DDU Gorakhpur University in India. Among his interests are MD simulations of proteins. I

DOI: 10.1021/acs.accounts.8b00467 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Wang, B.; Vajpai, M.; Shaik, S. MD Simulations and QM/MM Calculations Show that Single-Site Mutations of Cytochrome P450BM3 Alter the Active Site’s Complexity and the Chemoselectivity of Oxidation Without Changing the Active Species. Chem. Sci. 2017, 8, 5335−5344. (c) Dubey, K. D.; Wang, B.; Shaik, S. Molecular Dynamics and QM/MM Calculations Predict the Substrate Induced Gating of Cytochrome P450 BM3 and the Regio- and Stereoselectivity of Fatty Acid Hydroxylation. J. Am. Chem. Soc. 2016, 138, 837−845. (d) Ramanan, R.; Dubey, K. D.; Wang, B.; Mandal, D.; Shaik, S. Emergence of Function in P450-Proteins: A Combined Quantum Mechanical/Molecular Mechanical and Molecular Dynamics Study of the Reactive Species in the H2O2-Dependent Cytochrome P450SPα and Its Regio- and Enantioselective Hydroxylation of Fatty Acids. J. Am. Chem. Soc. 2016, 138, 6786−6797. (10) Ravichandran, K. G.; Boddupalli, S. S.; Hasermann, C. A.; Peterson, J. A.; Deisenhofer, J. Crystal Structure of Hemoprotein Domain of P450BM-3, a Prototype for Microsomal P450’s. Science 1993, 261, 731−736. (11) (a) Li, H.; Poulos, T. L. Conformational dynamics in cytochrome P450-Substrate Interactions. Biochimie 1996, 78, 695− 699. (b) Li, H.; Poulos, T. L. Structure of Cytochrome P450 BM3 Haem Domain Complex with the Fatty Acid Substrate, Palmitoleic Acid. Nat. Struct. Biol. 1997, 4, 140−146. (c) Lee, Y.-T.; Wilson, F.; Rupniewski, I.; Goodin, D. B. P450cam Visits an Open Conformation in the Absence of Substrate. Biochemistry 2010, 49, 3412−3419. (12) Sevrioukova, I. F.; Li, H.; Zhang, H.; Peterson, J. A.; Poulos, T. L. Structure of a Cytochrome P450−Redox Partner Electron-Transfer Complex. Proc. Natl. Acad. Sci. U. S. A. 1999, 96, 1863−1868. (13) Page, C. C.; Moser, C. C.; Chen, X.; Dutton, P. L. Natural Engineering Principles of Electron Tunneling in Biological OxidationReduction. Nature 1999, 402, 47−52. (14) For kET values, see Table 4 in: Whitehouse, C. J. C.; Bell, S. G.; Wong, L.-L. P450BM3 (CYP102A1): Connecting the Dots. Chem. Soc. Rev. 2012, 41, 1218−1260. (15) Marcus, R. A.; Sutin, N. Electron Transfer in Chemistry and Biology. Biochim. Biophys. Acta, Rev. Bioenerg. 1985, 811, 265−322. (16) Watanabe, K.; Austin, N.; Stapleton, M. R. Investigation of the Air Separation Properties of the Zeolites Types A, X and Y by Monte Carlo Simulations. Mol. Simul. 1995, 15, 197−221. (17) Prasad, B.; Mah, D. J.; Lewis, A. R.; Plettner, E. Water Oxidation by Cytochrome P450: Mechanism and the Function of the Reaction. PLoS One 2013, 8, No. e61897. (18) Tripathi, S.; Li, H.; Poulos, T. L. Structural Basis for Effector Control and Redox Partner Recognition in Cytochrome P450. Science 2013, 340, 1227−1230. (19) Hayashi, T.; Harada, K.; Sakurai, K.; Shimada, H.; Hirota, S. A Role of the Heme-7-Propionate Side Chain in Cytochrome P450cam as a Gate for Regulating the Access of Water Molecules to the Substrate-Binding Site. J. Am. Chem. Soc. 2009, 131, 1398−1400. (20) Vaz, A. D. N.; McGinnity, D. F.; Coon, M. J. Epoxidation of Olefins by Cytochrome P450: Evidence from Site-Specific Mutagenesis for Hydroperoxo-Iron as an Electrophilic Oxidant. Proc. Natl. Acad. Sci. U. S. A. 1998, 95, 3555−3560. (21) Jin, S.; Makris, T. M.; Bryson, T. A.; Sligar, S. C.; Dawson, J. H. Epoxidation of Olefins by Hydroperoxo-Ferric Cytochrome P450. J. Am. Chem. Soc. 2003, 125, 3406−3407. (22) Li, C.; Zhang, L.; Zhang, C.; Hirao, H.; Wu, W.; Shaik, S. What Oxidant is Really Responsible for Sulfur Oxidation by Cytochrome P450? Angew. Chem., Int. Ed. 2007, 46, 8168−8170. (23) Miura, Y.; Fulco, A. J. ω-1, ω-2 and ω-3 Hydroxylation of Long-Chain Fatty Acids, Amides and Alcohols by a Soluble Enzyme System from Bacillus Megaterium. Biochim. Biophys. Acta, Lipids Lipid Metab. 1975, 388, 305−317. (24) Matsunaga, I.; Yamada, M.; Kusunose, E.; Nishiuchi, Y.; Yano, I.; Ichihara, K. Fatty Acid-Specific, Regiospecific, and Stereospecific Hydroxylation by Cytochrome P450 (CYP152B1) from Sphingomonas Paucimobilis: Substrate Structure Required for α-Hydroxylation. FEBS Lett. 1996, 386, 252−254.

Sason Shaik is an MD simulations convert.



ACKNOWLEDGMENTS This work was supported by the Israel Science Foundation (Grant 1183/13). J. Li designed the boy in Figure 1. C. Li gave his advice on recoloration of the TOC is acknowledged. H. Schwarz raised constructive questions about the driving force for substrate’s entrance.

■ ■

DEDICATION This Account is dedicated to Prof. Yitzhak Apeloig for his forthcoming 75th birthday. REFERENCES

(1) (a) Cytochrome P450: Structure, Mechanism, and Biochemistry, 2nd ed.; Ortiz de Montellano, P. R., Ed.; Plenum Press: New York, 1995. (b) Sono, M.; Roach, M. P.; Coulter, E. D.; Dawson, J. H. Heme Containing Oxygenases. Chem. Rev. 1996, 96, 2841−2881. (c) Poulos, T. L. Heme Enzyme Structure and Function. Chem. Rev. 2014, 114, 3919−3962. (2) (a) Dawson, J. H.; Sono, M. Cytochrome P-450 and Chloroperoxidase: Thiolate-Ligated Heme Enzymes. Spectroscopic Determination of Their Active Site Structures and Mechanistic Implications of Thiolate Ligation. Chem. Rev. 1987, 87, 1255−1276. (b) Ortiz de Montellano, P. R.; De Voss, J. J. Oxidizing Species in the Mechanism of Cytochrome P450. Nat. Prod. Rep. 2002, 19, 477−493. (c) Hui Bon Hoa, G.; McLean, M. A.; Sligar, S. G. High Pressure, a Tool for Exploring Heme Protein Active Sites. Biochim. Biophys. Acta, Protein Struct. Mol. Enzymol. 2002, 1595, 297−308. (d) Guengerich, F. P.; MacDonald, T. L. Mechanisms of Cytochrome P-450 Catalysis. FASEB J. 1990, 4, 2453−2459. (3) (a) Shaik, S.; Kumar, D.; de Visser, S. P.; Altun, A.; Thiel, W. Theoretical Perspective on the Structure and Mechanism of Cytochrome P450 Enzymes. Chem. Rev. 2005, 105, 2279−2328. (b) Shaik, S.; Cohen, S.; Wang, Y.; Chen, H.; Kumar, D.; Thiel, W. P450 Enzymes: Their Structure, Reactivity, and Selectivity − Modeled by QM/MM Calculations. Chem. Rev. 2010, 110, 949− 1017. (4) (a) Fishelovitch, D.; Shaik, S.; Wolfson, H. J.; Nussinov, R. Theoretical Characterization of Substrate Access/Exit Channels in the Human Cytochrome P450 3A4 Enzyme: Involvement of Phenylalanine Residues in the Gating Mechanism. J. Phys. Chem. B 2009, 113, 13018−13025. (b) Cojocaru, V.; Winn, P. J.; Wade, R. C. The Ins and Outs of Cytochrome P450. Biochim. Biophys. Acta, Gen. Subj. 2007, 1770, 390−401. (5) Fishelovitch, D.; Shaik, S.; Wolfson, H.; Nussinov, R. How Does the Reductase Help Regulating the Catalytic Cycle of Cytochrome P450 3A4 Using the Conserved Water Channel? J. Phys. Chem. B 2010, 114, 5964−5970. (6) Sligar, S. G.; Gunsalus, I. C. A Thermodynamic Model of Regulation: Modulation of Redox Equilibria in Camphor. Proc. Natl. Acad. Sci. U. S. A. 1976, 73, 1078−1082. (b) Govindaraj, S.; Poulos, T. L. he Domain Architecture of Cytochrome P450BM-3. J. Biol. Chem. 1997, 272, 7915−7921. (7) Volz, T. J.; Rock, D. A.; Jones, J. P. Evidence for Two Different Active Oxygen Species in Cytochrome P450 BM3Mediated Sulfoxidation and N-Dealkylation Reactions. J. Am. Chem. Soc. 2002, 124, 9724−9725. (8) (a) Zheng, J. J.; Altun, A.; Thiel, W. Common System Setup for the Entire Catalytic Cycle of Cytochrome P450CAM in Quantum Mechanical/Molecular Mechanical Studies. J. Comput. Chem. 2007, 28, 2147−2158. (b) Schöneboom, J. C.; Thiel, W. The Resting State of P450cam: A QM/MM Study. J. Phys. Chem. B 2004, 108, 7468− 7478. (9) (a) Dubey, K. D.; Shaik, S. Choreography of the Reductase and P450BM3 Domains Toward Electron Transfer Is Instigated by the Substrate. J. Am. Chem. Soc. 2018, 140, 683−690. (b) Dubey, K. D.; J

DOI: 10.1021/acs.accounts.8b00467 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research (25) Lee, D. S.; Yamada, A.; Sugimoto, H.; Matsunaga, I.; Ogura, H.; Ichihara, K.; Adachi, S.; Park, S. Y.; Shiro, Y. Substrate Recognition and Molecular Mechanism of Fatty Acid Hydroxylation by Cytochrome P450 from Bacillus Subtilis. J. Biol. Chem. 2003, 278, 9761−9767. (26) Fujishiro, T.; Shoji, O.; Nagano, S.; Sugimoto, H.; Shiro, Y.; Watanabe, Y. Crystal Structure of H2O2-Dependent Cytochrome P450SPα with Its Bound Fatty Acid Substrate. J. Biol. Chem. 2011, 286, 29941−29950. (27) Arnold, F. H. Design by Directed Evolution. Acc. Chem. Res. 1998, 31, 125−131. (28) Reetz, M. N. Controlling the Enantioselectivity of Enzymes by Directed Evolution: Practical and Theoretical Ramifications. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 5716−5722.

K

DOI: 10.1021/acs.accounts.8b00467 Acc. Chem. Res. XXXX, XXX, XXX−XXX