Introduction

5-Fluoro-1H,3H-pyrimidine-2,4-dione (5-fluorouracil, 5-FU) is a cytotoxic chemotherapy medication used to treat cancer and belongs to the antimetabolite and pyrimidine analog families of medications [1,2,3,4]. 5-FU acts as a thymidylate synthase inhibitor and blocks synthesis of the pyrimidine thymidylate, which is a nucleotide required for deoxyribonucleic acid (DNA) replication, resulting in cancerous cell death via thymineless death [4]. Currently, 5-FU is used in combination with other drugs and therapies for the treatment of many types of cancer, including stomach cancer, skin warts, breast cancer, cervical cancer, basal cell carcinoma, and colorectal esophageal cancer [5]. The combined use of 5-FU can maximize its effectiveness against some cancers; however, it may cause strong side effects that need to be addressed. And another point to note is that there is very little difference between the maximum tolerated dose and minimum effective dose of 5-FU because of side effects [6, 7]. Therefore, it is expected to clarify the causes and mechanisms of the side effects to secure diverse use of 5-FU as a combined therapy for cancer treatment.

As many immune checkpoint inhibitors are being developed, further development of immunotherapy is expected in the future, and combination with DNA analog drugs involving 5-FU has the potential to become a highly effective therapy. The immune system of the living body involves numerous proteins and cells such as neutrophils, macrophages, killer T cells, and various cytokines [8, 9]. Immunotherapy acts on any part of the complex immune system to increase its cytotoxic effect on cancer cells, while the terminal of the immune system oxidizes and destroys cancer cells by producing reactive oxygen species (ROS). Therefore, DNA analog drugs such as 5-FU are required to act under a lot of ROS produced in the immune system for the effect of replication inhibitor in the combined use. To maximize its effectiveness as the combined use, it is necessary to not degrade against ROS, but low degradation may increase side effects by increasing concentrations of 5-FU in blood plasma. Conversely, products derived from the oxidative degradation of 5-FU may cause side effects. In addition, if most of the dosed 5-FU is decomposed, the concentration near the cancer cells drops to a level where the desired antimetabolite effect cannot be obtained. Considering the above, it is necessary to demonstrate the chemical reaction between 5-FU and ROS such as superoxide radical anion (O2•−, a precursor of hydroperoxyl radical: HO2•) and hydroxyl radical (OH•). That is, clarifying the chemical mechanism of the oxidative degradation of 5-FU by ROS is a prerequisite to elucidate the related medicinal side effects, since oxidized products may cause several pathologies.

Recently, Xiaofan Li et.al. investigated the oxidation mechanism of 5-FU by ozone (O3) and OH• using density functional theory (DFT) calculations to assess the potential ecotoxicology of 5-FU as a recalcitrant compound to the natural environment [10]. They clarified that 5-FU can be rapidly degraded by O3 and OH•, which subsequently undergoes ring-opening, decomposition, defluorination, and hydroxylation steps, forming fifteen structures of transformation products, including five experimental products [11, 12]. Here, both O3 and OH• are particularly strong oxidants among ROS although are not main oxidants in living organisms. Thus, O2•−, that is the majority in ROS produced by phagocytes (neutrophils, monocytes and macrophages) in the in vivo immune system, may be the oxidant for 5-FU degradation. Since O2•− is not a moderately good electrophile that can accept electrons from 5-FU, HO2• formed after protonation of O2•− as a Brønsted base is a strong oxidant with a short lifetime. Therefore, HO2• is inferred to be the actual oxidants in the living body, where proton transfer (PT) and electron transfer (ET) are closely related to the 5-FU oxidation mechanism. Several oxidation mechanisms of acidic 5-FU are possible, such as single–electron transfer (SET), superoxide-facilitated oxidation [13,14,15], sequential proton-loss electron transfer [16], and hydrogen–atom transfer (HAT) involving proton-coupled electron transfer (PCET) [17,18,19,20]. Currently, it is reasonable that HAT involving PCET is plausible for the oxidation of 5-FU rather than SET. In our previous studies, the PCET reaction between electrogenerated O2•− and acidic antioxidants was analyzed in an aprotic solvent [18, 20,21,22,23,24,25,26,27]. These studies suggested that even a recalcitrant compound such as 5-FU can be oxidized by O2•− or HO2• (protonated form of O2•−) through a PCET mechanism initiated by the initial PT. Therefore, it is essential to reveal the detailed mechanism of the initial reaction between 5-FU and O2•− (HO2•) because complicated metabolic reactions follow it forming the main transformation products in the living body, for evaluating the causes and mechanisms of the side effects.

In this study, the reaction mechanism between 5-FU and O2•− (HO2•) in an aqueous environment was investigated using DFT calculations. Then, DFT results comparatively evaluated the degradability of 5-FU by O2•− (HO2•) with those of two pyrimidine bases (Chart 1), 5-methylpyrimidine-2,4(1H,3H)-dione (thymine), and pyrimidine-2,4(1H,3H)-dione (uracil, a nucleobase in ribonucleic acid: RNA). Accordingly, I present valuable information regarding the oxidative degradation of 5-FU by O2•− (HO2•) in an aqueous environment simulating a living cell, which is assumed to be important for understanding the side effects and expanding the scope of 5-FU application.

Chart 1
scheme 1

Structures of the compounds considered in the present study. a 5-Fluoro-1H,3H-pyrimidine-2,4-dione (5-fluorouracil), b 5-methylpyrimidine-2,4(1H,3H)-dione (thymine), and c pyrimidine-2,4(1H,3H)-dione (uracil)

Full size image

Methods

Solution-phase DFT calculations were performed with two hybrid functionals, the Becke three-parameter Lee–Yang–Parr functional (B3LYP) and the Minnesota 06 functional (the meta exchange-correlated functional, M06-2X) [28], implemented in the Gaussian 16 Program package [29]. These functionals obtain good geometries for the reactants and products in PCET reactions between acidic substrates and free radicals [30]. I obtained the energies of the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) from geometry optimization using frontier orbital theory. In the calculations, I applied the standard split-valence triple ζ basis sets augmented by the polarization 3df,2p and diffusion orbitals 6–311 +  + G(3df,2p). The polarized continuum model (PCM) was employed for the solvent contribution of water to the standard Gibbs free energies under the default settings of Gaussian 16 that is widely employed. The internal energies were converted to standard Gibbs energies at 298.15 K using the zero-point energies, thermal correction, and entropy. The electrons and spins were obtained using the natural bond orbital (NBO) technique in population analysis [31].

Result and discussion

Optimization of stable structure of 5-FU and its deprotonate anion

To elucidate the reaction mechanism between 5-FU and O2•− in aqueous environments such as living organisms, DFT calculations were performed using the B3LYP and M06-2X hybrid functionals employing the PCM method. Here, I focus on the B3LYP results because the results obtained using both functionals are similar. First, the stable structures of the possible reactant species, 5-FU, its deprotonated anion of two conformers (− Ha and − Hb), O2•−, and HO2•, which coexist in aqueous cells, were obtained from energy scanning (Supporting Information, Tables S1–S4). Figure 1 shows the optimized structures with their calculated standard Gibbs free energy changes (ΔG°/kJ mol−1, 298.15 K) along the PT from 5-FU to O2•− (M06-2X results are shown in Supporting Information, Fig. S1). Charges distributed on the amine (NH) protons were obtained by NBO analysis.

Fig. 1
figure 1

Optimized structures of 5-fluorouracil (5-FU) and deprotonated anions (− Ha and − Hb) along a proton transfer (PT) from its amine group at the first and third positions (1NH and 3NH) to O2•− in water, calculated using DFT-B3LYP/PCM/6–311 +  + G(3df,2p) method. The ΔG°s (kJ mol−1, 298.15 K) of PT and charges distributed on the protons of 5-FU obtained in the NBO analysis are also indicated

Full size image

In the living body, O2•− produced in the immune system acts as a Brønsted base though that accepts a proton from the aqueous environment, forming HO2•. Conversely, HO2• is a strong oxidant with a short lifetime that cannot accept any more protons without accepting an electron. Therefore, HO2• immediately reacts with 5-FU only when generated in the vicinity of 5-FU molecule. Comparing the charges on the two NH protons of 5-FU at the first and third positions (1NH: 0.439 and 3NH: 0.431), it was observed that the proton reactivities in the acid–base reaction are slightly higher at 1NH than at 3NH. Similarly, the ΔG° values (kJ mol−1) of the PT from 5-FU to O2•− indicate that deprotonation at 1NH forming − Ha (4.8) is more feasible than that at 3NH forming − Hb (16.7). According to these calculations, PT occurs at the 1NH of 5-FU toward O2•− depending on the acidity of the biological environment.

Change in HOMO–LUMO energies during PCET between 5-FU and O2 •−

The mechanistic analysis of reactivity between O2•− (HO2•) and 5-FU in water was supplemented by frontier molecular orbital analysis based on DFT. Figure 2 shows the HOMO–LUMO (Hartree/a.u.) changes during PCET between 5-FU and O2•− (M06-2X results are shown in Supporting Information, Fig. S2). After the initial PT, the reactant species, 5-FU, − Ha, O2•−, and HO2• coexist in the experimental solution. The HOMO energies of 5-FU (−0.25953) and its anion (−Ha: −0.20788) were much higher than the singly occupied molecular orbital (SOMO) energy of HO2• (−0.31427), indicating that the electron acceptor was HO2• rather than O2•− (−0.16005). Meanwhile, HO2• can be formed after PT to O2•− from aqueous solvent or 5-FU. Therefore, the electron donor was 5-FU or − Ha (the bold red lines in Fig. 2 indicate the downhill energy relationship during the ET). Then, after the subsequent ET, the HOMO–LUMO relationship between the products (i.e., −e/−Ha −e and HO2−) is reversed, as shown by the red dotted lines; thus, the reverse ET may proceed. However, H2O2 has a lower HOMO (−0.27553) than that of hydroperoxyl anion, HO2− (−0.16819), and a SOMO of radical anion,−2H−e (−0.24290), indicating that the second PT from neutral radical−Ha−e forming−2H−e prevents the reverse ET. Notably, the radical intermediates (radical cation: − e and − Ha − e) have very low SOMO (−0.33280 and −0.27958) comparable to that of HO2•, which causes the reverse ET from HO2− or hydrogen peroxide (H2O2). This indicates that 5-FU is difficult to be oxidized and also suggests that intermediate radicals and anions may undergo subsequent reactions between another molecule of O2•− or HO2• (e.g., radical adduct formation) through pathways other than PCET. As a result of the mechanistic analysis using the HOMO–LUMO relationship, 5-FU is difficult to be oxidized by O2•− but HO2• through a PCET.

Fig. 2
figure 2

Change in HOMO − LUMO energies (Hartree/a.u.) during proton-coupled electron transfer between 5-fluorouracil (5-FU) and O2•− in water, calculated using DFT-B3LYP/PCM/6–311 +  + G(3df,2p) method

Full size image

ΔG°s in PCET between 5-FU and O2 •−

For a thermodynamic analysis of the oxidation mechanism of 5-FU by O2•− in water, ΔG°s (kJ mol−1, 298.15 K) along the PCET were obtained using solution-phase vibrational frequency calculations employing the PCM method (Supporting Information, Tables S1–S4). Figure 3 shows the equilibrium schemes and ΔG°s of the six diabatic electronic states in the PCET involving two PTs and one ET between 5-FU and O2•− calculated at the B3LYP/PCM/6–311 +  + G(3df,2p) level (M06-2X results are shown in Table 1). Along the sequential pathway shown in Fig. 3, the main drivers are the ΔG°s of the individual reactions, the redox potentials (ET1–ET3), and the acid–base interactions (PT1–PT4) of the components. Since ET1 (472.8) is strongly endergonic, PT1 (4.8) dominantly forms –Ha and HO2•. In the subsequent pathway (bottom of the panels in Fig. 3), both PT3 (352.5) and ET2 (116.5) are endergonic, so the sequential pathway is unlikely to proceed. However, PT3 and ET2 are promoted by the following exergonic processes, ET3 (− 306.9) and PT4 (− 70.9). As shown in the HOMO–LUMO relationship (Fig. 2), the second PT forming H2O2 is necessary for successful ET along the PCET. Therefore, it is rational that the proton and electron are transferred in one kinetic process corresponding to the diagonal of the rectangle in Fig. 3, labeled concerted PCET (45.6). Thus, the only feasible pathway is PT1 followed by concerted PCET rather than a sequential pathway of PT1–ET2–PT4 (PT1–PT3–ET3), resulting in the oxidative degradation of 5-FU by O2•−. The ΔG° of the concerted PCET (45.6) is not high, although a prereactive step to form a hydrogen-bonded complex between –Ha (at 3NH) and HO2• is required. As a result of the thermodynamic analysis using the ΔG°s, 5-FU is possible to be oxidized by O2•−, where HO2• preformed via the PT (4.8) acts as an actual oxidant through a concerted PCET mechanism which is a weakly endergonic pathway.

Fig. 3
figure 3

Six diabatic electronic states of two proton transfers (PT1–PT4) and one electron transfer (ET1–ET3) between 5-fluorouracil and O2•− in water. The ΔG°s (kJ mol−1, 298.15 K) were calculated using DFT-(U)B3LYP/PCM/6–311 +  + G(3df,2p) method

Full size image
Table 1 ΔGº values (kJ mol−1, 298.15 K) of proton-coupled electron transfer between 5-fluorouracil (5-FU), thymine, uracil, and O2•− in water, calculated using DFT at the B3LYP and M06-2X functionals with the PCM/6–311 +  + G(3df,2p) basis set
Full size table

For a comparative study, the ΔG° values of the PCET pathways of thymine, uracil, and 5-FU with analog structures were calculated (Table 1). Thermodynamically, the energetic driving force of the PCET is the sum of the ΔG°s of the two PTs and one ET, although the sum cannot embody the energetic driving force if the PCET occurs along a pathway involving an infeasible single PT/ET. Along the plausible sequential pathway, PT1–ET2–PT4, the ΔG°s of ET2 for each compound were strongly endergonic when using each of the B3LYP or M06-2X functionals. Thus, the ΔG° results imply that the initial PT (PT1) followed by a concerted PCET which correspond to sums of those for ET2 and PT4 (PT3 and ET3) is the most feasible pathway for the oxidation of each compound by O2•−. Then, the ΔG°s of the concerted PCET (concerted in Table 1) and the net PCET involving two PTs and one ET (total in Table 1) indicate that 5-FU (45.6, 50.5/49.8, 46.3) is more easily oxidized than thymine (47.6, 70.6/54.1, 68.1) and uracil (67.5, 84.8/75.1, 83.0) by O2•−. Thus, this ΔG° results explain that 5-FU is comparatively more degradable by O2•− than uracil and thymine. In addition, 5-FU plays a preventive role in cellular oxidative damage at the pyrimidine base (uracil in RNA, thymine in DNA).

ΔG°s in PCET between 5-FU and HO2 •

Next, a thermodynamic analysis of the oxidation mechanism of 5-FU by HO2• preformed via the protonation of O2•− in aqueous environment was conducted. Figure 4 shows the equilibrium schemes and ΔG°s of the PCET involving one PT and one ET between 5-FU and HO2• calculated at the B3LYP/PCM/6–311 +  + G(3df,2p) level (M06-2X results are shown in Table 2). PT occurs from 1NH (a) or 3NH (b) in 5-FU. In both cases, the concerted PCET (50.2, 105.2, red lines) is a feasible pathway because the initial processes of sequential pathways (ET–PT or PT–ET), ET1 (255.6) and PT1 (266.0, 277.9), are strongly endergonic. A comparison of these two ΔG°s of the concerted PCET clearly suggests that the reaction proceeds at 1NH. Thus, it is rational that the proton and electron are transferred in one kinetic process at 1NH corresponding to the diagonal of the rectangle in Fig. 4a; labeled concerted.

Fig. 4
figure 4

Four diabatic electronic states of one proton transfer (PT1, PT2) and one electron transfer (ET1, ET2) between 5-fluorouracil and HO2• in water. The ΔG°s (kJ mol−1, 298.15 K) were calculated using DFT-(U)B3LYP/PCM/6–311 +  + G(3df,2p) method

Full size image
Table 2 ΔGº values (kJ mol−1, 298.15 K) of proton-coupled electron transfer between 5-fluorouracil (5-FU), thymine, uracil, and HO2• in water, calculated using DFT at the B3LYP and M06-2X functionals with the PCM/6–311 +  + G(3df,2p) basis set
Full size table

In Table 2, the ΔG° values of the PCET between 5-FU, thymine, uracil, and HO2• calculated using the B3LYP or M06-2X functionals are compared. Similar to 5-FU, the ΔG° results indicate that the sequential pathways (ET1–PT2 or PT1–ET2) for thymine and uracil are unfeasible because the initial processes are strongly endergonic. Therefore, only the same concerted PCET pathway at 1NH can occur between HO2• and these pyrimidine bases. The ΔG° values of the concerted PCET indicate that 5-FU (50.2, 54.0) and thymine (48.9, 55.1) may be oxidized because of their weakly endergonic pathway showing similar reactivities toward HO2•. During the reaction in water solution, a few elementary steps are assumed to proceed, generation of HO2• via protonation of O2•−, formation of a prereactive complex (PRC) via a hydrogen bond (HB) from the free reactants (5-FU, thymine, uracil, and HO2•), and the concerted PCET yielding product complex resulting in degradation of the pyrimidine analog. Considering that HO2• is a strong oxidant with a short lifetime, this series of steps must proceed with superior kinetics to oxidize 5-FU. This is because in cells in which a large amount of pyrimidine bases coexists, HO2• reacts with the pyrimidine bases or disappears through a dismutation reaction. Therefore, 5-FU cannot form the PRC in an experimental solution, resulting in not being oxidized by HO2•, although the ΔG° results indicate that 5-FU and thymine can be oxidized through a concerted PCET.

Collectively, these ΔG° analyses indicate that the oxidation of 5-FU by O2•− or HO2• in water can proceed through the concerted PCET. Furthermore, two PCET pathways are feasible for the oxidation of 5-FU, the initial PT followed by a concerted PCET between 5-FU and O2•−, and a concerted PCET between 5-FU and HO2• preformed via the protonation of O2•− by solvent-derived proton. In both pathways, a significant step is the formation of PRC via a HB between HO2• and 5-FU because the actual oxidant is HO2• with a short lifetime.

Analyses of reaction coordinates and potential-energy surfaces of the PCET between 5-FU and HO2 •

To infer the mechanistic insights of the oxidative degradation of 5-FU, I investigated the potential-energy surfaces of the PCET between 5-FU/ − H and HO2• using the (U)B3LYP and M06-2X/PCM/6–311 +  + G(3df,2p) levels of theory combined with NBO calculations. Here, I focus on the B3LYP results because the results obtained using both functionals are similar (M06-2X results are shown in Supporting Information, Fig. S3). During the reaction, the prereactive elementary steps are assumed to proceed, i.e., formation of HO2• and formation of PRC via HBs between 5-FU/ − H and HO2•. Then, PCET occurs, forming a product complex (PC) via TS. First, I scanned some plausible PRCs, the stable intermediate complex, TS, and PC, along with the concerted PCET reaction (Supporting Information, Table S5). As a result, only a 1:1 PRC formed between 5-FU and HO2• (5-FU—HO2•) bonded at the 1NH (and auxiliary at oxygen atom at the second position) can form a PC (− H − e—H2O2) through the concerted PCET involving one ET and one PT along an intrinsic reaction coordinate (IRC) via a TS (Fig. 5). Figure 5a shows the energy profile (ΔG°, kJ mol−1) along the IRC, which forms the PC via the TS with an activation energy (Ea = 91.8 kJ mol−1), where Ea is not so high energetic barrier. Then, the PC (62.1) dissociates yielding − H − e and H2O2 as the free products (FP, 41.0). Here, − H − e is a neutral radical that is expected to undergo various subsequent reactions and be easily decomposed in an aqueous environment.

Fig. 5
figure 5

a Energy profile (kJ mol−1, 298.15 K) along the reaction coordinate of proton-coupled electron transfer between 5-fluorouracil (5-FU) and HO2• in water with the structures of prereactive complex (PRC, 5-FU—HO2•), transition state (TS), and product complex (PC, − H − e—H2O2), forming free products (FP). b Changes in N − H bond distance (1 N − H: black line, nm), and the number of electrons (open circles) in the π-orbitals of 5-FU, corresponding to the labels of the left and right vertical axes. Calculations were performed using the DFT-(U)B3LYP/PCM/6–311 +  + G(3df,2p) method. The spin distributions of the complexes obtained from the NBO analyses are demonstrated

Full size image

Figure 5b shows the dependence of N − H bond distance (1 N − H, black line, nm) on the number of electrons in the π-orbital of the planar 5-FU molecule along the IRC (blue circles). The spin density distributions were localized on the radicals before and after the TS along the concerted PCET, demonstrating that the radical localized on HO2• in the initial PRC was transferred to − H − e (5-FU side) in the resulting PC. The spin changes on the electron-donor side (5-FU) and electron-acceptor side (HO2•) were well correlated with the changes in the π-electrons of 5-FU. The spins also demonstrate that the TS is electronically characterized by the delocalization of the π-electrons over the HB complex of the components. As a result, IRC revealed that an intracomplex PCET between 5-FU (at 1NH) and HO2• occurs in a concerted one-kinetic process without generating intermediates.

Collectively, these findings indicate that the oxidation of 5-FU by O2•− in water is governed by the concerted PCET after the formation of PRC between 5-FU and preformed HO2• via HBs at 1NH, which corresponds to moving along the red diagonal lines in Fig. 4a.

Conclusions

In conclusion, I clarified the initial reaction between 5-FU and O2•− (HO2•) in water solutions using DFT calculations, which is expected to lead to the degradation of 5-FU and complicated metabolic reactions. The main findings are summarized below:

  1. 1.

    Oxidation of 5-FU by O2•− or HO2• in water is feasible through the PCET mechanism.

  2. 2.

    A concerted PCET between 5-FU and HO2• preformed via the protonation of O2•− involving one PT and one ET in one kinetic process is a plausible pathway.

  3. 3.

    The actual oxidant of 5-FU along the PCET mechanism is HO2•.

These findings using DFT calculations revealed the PCET mechanism of 5-FU oxidation by O2•− and HO2•, which cannot be observed in water or aqueous in vivo experiments. I hope that these findings will provide mechanistic insight into the biodegradability of 5-FU through oxidation by O2•−, securing the pharmacological use of 5-FU.