**1. Introduction**

Cytochrome P450 2C8 (CYP2C8) is a member of the cytochrome P450 (CYP) epoxygenase family that metabolizes over 60 clinically relevant drugs on the market [1–3]. For example, CYP2C8 is the primary enzyme involved in the metabolism of paclitaxel (PAC), a common chemotherapeutic that works by interfering with microtubule function [4]. CYP2C8 is primarily expressed hepatically, though it is also present in the vasculature and kidneys [5–7], where it metabolizes lipids, such as arachidonic acid (AA), to form biologically active epoxyeicosatrienoic acids (EETs). EETs are known to be anti-inflammatory [8], angiogenic [9], and inhibit vascular smooth muscle cell migration, implicating CYP2C8 in regulating kidney and vascular function [10].

Polymorphic variations in CYPs have been of clinical interest due to individual differences in drug metabolism. For example, CYP2D6 is among the most highly polymorphic CYPs that greatly contributes to the poor, intermediate, extensive, and ultra-rapid metabolizer phenotypes [11,12]. Two common, naturally occurring polymorphic variants of CYP2C8—CYP2C8\*2 and CYP2C8\*3—display altered drug elimination rates and EET production compared to CYP2C8\*1 (wildtype, WT) [13,14]. The CYP2C8\*3 polymorphism (R139K/K399R) is present in 2% of African-American and 13% of Caucasian

populations [13]. CYP2C8\*3 is associated with an increase in peripheral neuropathy in patients treated with PAC, presumably a result of slower PAC elimination by CYP2C8\*3 [15,16]. However, some studies have suggested PAC metabolism is increased with CYP2C8\*3 [17]. In vitro studies corroborate that CYP2C8\*3 has only 30% and 15% of the activity compared to WT for the turnover of AA and PAC, respectively [13]. However, others report that PAC metabolism is not significantly affected by these polymorphisms [14,18], and one study observed greater PAC turnover by \*3 as compared to WT [19]. Therefore, from previous reports, the effects of CYP2C8\*3 on PAC metabolism are convoluted. The CYP2C8\*2 (I269F) polymorphism, present in 18% of African-American populations [13], shows lower, albeit not always statistically significant, in vitro PAC turnover compared to WT [13,14,20].

Importantly, the amino acid residues that are different in the CYP2C8\*2 and CYP2C8\*3 (R139, K399, and I269) are not located within the enzyme active site of CYP2C8 but on the surface of the protein (Figure 1) [2,13,18,19]. This suggests that these mutations disrupt mechanisms of substrate metabolism that are not directly related to substrate binding. Indeed, CYP2C8\*3 showed WT activity for the deethylation of amiodarone, and so it appears that this polymorphism does not affect substrate binding as a whole [21]. The CYP catalytic cycle is a complex series of redox reactions that require assistance from redox partners such as cytochrome P450 reductase (CPR) (Figure 1). The details of this complex cycle can be found in many reviews [22–25]. Therefore, these mutations may be affecting other steps in the CYP catalytic cycle, such as electron transfer from redox partners. Earlier work demonstrated that CYP2C8\*3 has different binding affinities to its redox partners compared to WT. Particularly, CPR interacts with PAC-bound CYP2C8\*3 better than WT as determined indirectly by PAC metabolism assays using varying CPR concentrations (apparent Km = 5.5 ± 1.5 nM for CYP2C8\*3 versus 35 ± 10 nM for WT) [19]. The greater apparent interaction with CPR ought to lead to a faster transfer of electrons and a greater substrate turnover. In the same study, the authors determined that PAC metabolism was increased with CYP2C8\*3 compared to WT [19]. However, the consensus through other studies is that PAC metabolism is either lower or similar to WT [13,14,18]. Taken together, these data suggest that the CYP-CPR interaction may be disrupted in CYP2C8\*3 and affects PAC metabolism.

CPR transfers two electrons to CYPs during the CYP catalytic cycle, with the first electron reducing the ferric heme to a ferrous heme and the second reducing the dioxygen-heme to a peroxy-heme (Figure 1). Many reactive oxidized intermediates are involved in the CYP catalytic cycle en route to substrate oxidation. These intermediates can sometimes decompose to form reactive oxygen species (ROS) instead of progressing towards substrate oxidation, a process known as uncoupling. These ROS, namely hydrogen peroxide (HOOH) and superoxide, are formed in large quantities by CYP2C enzymes [26,27]. ROS can induce mitochondrial dysfunction in cardiomyocytes, induce coronary artery vasoconstriction, and promote carcinogenesis [26,28–30], and ROS specifically generated by endothelial CYP2C8 has been shown to impair functional recovery after ischemia/reperfusion injury [27]. Another potential effect of these polymorphisms may therefore be on the coupling efficiency of PAC metabolism.

Herein, we determined the effects of the CYP2C8\*2 and CYP2C8\*3 polymorphisms in regards to first-electron transfer (FET) kinetics and PAC metabolism uncoupling. We tested CYP2C8\*1, CYP2C8\*2, and CYP2C8\*3, as well as the single mutations of the CYP2C8\*3 polymorphism (R139K and K399R). To study these polymorphisms, we utilized the Nanodisc technology to solubilize these CYP2C8 variants in a membrane mimic [31,32]. We find that CYP2C8\*2 has a lower PAC turnover rate compared to WT. We further determined that CYP2C8\*3 produces significantly more HOOH compared to WT, indicating a greater uncoupling of the catalytic cycle. Using stopped-flow measurements of the FET, we determined that the CYP2C8\*2 and CYP2C8\*3 have profoundly distinct and slower FET kinetics compared to WT. We determined that there is no change in the reduction potential of the polymorphisms compared to WT, which supports that the altered FET kinetics is due to an altered redox interaction with CPR. None of the single-mutant variants of CYP2C8\*3 (R139K and K399R) reproduce the effects of the CYP2C8\*3, indicating that the effects of this polymorphism are a synergism of both mutations. Taken together, these data demonstrate that these polymorphisms have altered FET kinetics, leading to an increase in HOOH production and greater PAC metabolism uncoupling.

**Figure 1.** Schematic of the CYP catalytic cycle. Inset shows the structure of CYP2C8 with residues R139 (blue), I269 (orange), and K399 (light blue) highlighted. Structure was generated in PyMol v1.3r1 using the PDB entry 1PQ2. A schematic of CYP2C8 (grey) and CPR (black) incorporated into nanodiscs is shown in the center of the cycles. *Catalytic cycle*. Substrate (PAC) binds to the CYP active site, which perturbs the H2O coordination to the iron heme. H2O unbinds leaving a pentacoordinated high-spin iron heme. CPR reduces the iron heme using an electron obtained from NADPH. Under anaerobic conditions, CO ligates the heme to terminate the cycle. Under aerobic conditions, O2 ligates the heme, followed by another one-electron reduction by CPR and the addition of a proton to produce a peroxy-heme. The peroxy-heme can decompose forming HOOH or eliminate an H2O molecule to produce the catalytic ferryl iron heme radical (Compound I). Compound I can oxidize substrate (PAC) to product (PAC-OH, red box), followed by the coordination of an H2O molecule to begin the cycle again. Spectroscopically visible species are indicated with their characteristic absorbance wavelength. More details of the cycle can be found in previous reviews [22–25].
