Hydroxyl Radical Generation by the H 2 O 2 /Cu II /Phenanthroline System under Both Neutral and Alkaline Conditions: An EPR/Spin-Trapping Investigation

: The copper–phenanthroline complex Cu I (Phen) 2 was the ﬁrst artiﬁcial nuclease studied in biology. The mechanism responsible for this activity involves Cu II (Phen) 2 and H 2 O 2 . Even if H 2 O 2 /Cu systems have been extensively studied in biology and oxidative chemistry, most of these studies were carried out at physiological pH only, and little information is available on the generation of radicals by the H 2 O 2 /Cu II -Phen system. In the context of paper pulp bleaching to improve the bleaching ability of H 2 O 2 , this system has been investigated, mostly at alkaline pH, and more recently at near-neutral pH in the case of dyed cellulosic ﬁbers. Hence, this paper aims at studying the production of radicals with the H 2 O 2 /Cu II -Phen system at near-neutral and alkaline pHs. Using the EPR/spin-trapping method, HO • formation was monitored to understand the mechanisms involved. DMPO was used as a spin-trap to form DMPO–OH in the presence of HO • , and two HO • scavengers were compared to identify the origin of the observed DMPO–OH adduct, as nucleophilic addition of water onto DMPO leads to the same adduct. H 2 O 2 decomposition was enhanced by the addition of Cu II –Phen (and only slightly by addition of CuSO 4 ), reaching a level similar to the Fenton reagent at near-neutral pH. This evidences the role of Phen, which improves the effect of Cu II by tuning the electronic structure and structural properties of the corresponding Cu II complexes.


Introduction
In 1979, Cu + (phenanthroline) 2 (the molecular structure of phenanthroline is presented in Figure 1) was discovered to display nuclease activity [1], by cleaving the phosphodiester bonds of DNA or RNA. The scission of DNA occurs in the presence of two co-reactants: Cu + (phenanthroline) 2 and H 2 O 2 [2,3]. According to Sigman [2], the Cu + (phenanthroline) 2 complex coordinates to DNA and the oxidation of this coordinate by H 2 O 2 further generates a cupric hydroxyl radical-like coordinate, responsible for the DNA scission. Hence, copper-phenanthroline coordinates (Cu-Phen) represent the first artificial nucleases studied in biology. Generally, they consist of bis(1,10-phenanthroline)copper(I) complexes [1][2][3][4][5], although a mono(1,10-phenanthroline)copper complex was found to However, the observation of DMPO-OH does not necessarily mean that HO • radicals are formed. Indeed, several phenomena lead to the DMPO-OH adduct, including (i) superoxide radical spin-trapping and decomposition of the corresponding DMPO-OOH adduct into DMPO-OH; (ii) the oxidation of DMPO by very strong oxidants (called inverted spin-trapping) [18]; or (iii) the nucleophilic addition of water onto the DMPO spin-trap (Forrester-Hepburn reaction) [19]. Besides, the nucleophilic addition of water is quite usual in aqueous solution and has been proved to be catalyzed by copper ions [13,15,20]. In the case of Cu(II), the metal binds to DMPO and the polarization of the DMPO's double bond promotes the nucleophilic attack of H 2 O (Figure 2).
To confirm that the observed DMPO-OH adduct originates from HO • addition and not from these above-mentioned side reactions, an indirect method using scavenger molecules can be applied. The selected scavengers should be more reactive towards hydroxyl radicals than DMPO (rate constants higher than that of the spin-trap, k(DMPO/HO • )) and could be added at high concentrations. Among scavengers, dimethyl sulfoxide (DMSO) and formate (HCOO − ) are often used, and the resulting adducts are DMPO-CH 3 • with a typical six-line signal (a N = 16.1 G, a H = 23 G) and DMPO-COO •− with another six-line pattern (a N = 15.75 G, a H = 19.09 G), see Figure S1 (Supplementary Materials).
While the spin-trapping technique has been extensively used to study hydroxyl radical formation by H 2 O 2 /Cu(I) [15] and H 2 O 2 /Cu(II) systems [21] especially in biology at physiological pH, less information is available when alkaline solutions are considered. In this context, this paper deals with the investigation of the decomposition of hydrogen peroxide into hydroxyl radicals in the presence of copper(II)-phenanthroline complexes in both neutral and alkaline aqueous media.

Analysis Conditions
The spin-trapping trials were carried out in aqueous solution at room temperature, except for some exceptional assays with heating.
Under near-neutral conditions (no addition of NaOH), the pH was between 7.5 and 9 (7.5-7.6 in the presence of H 2 O 2 , around 8-8.6 in the absence of H 2 O 2 , and pH 9 when adding sodium formate). Under strong alkaline conditions (0.02 M NaOH), the pH was of 12.5. The use of scavengers induced some pH variations. With formate, the pH was around 9 in the absence of NaOH and was equal to 12.5 with NaOH. With DMSO, the pH was around 9 alone, around 8 with Cu-Phen, 7.6 with Cu-Phen + H 2 O 2 , and 14 with Cu-Phen + H 2 O 2 + NaOH.
A "Fenton control" was performed for comparison with FeSO 4 replacing CuSO 4 , at the same molar concentrations, and with the addition of H 2 SO 4 to reach a pH of 3.

EPR Experiments
X-band EPR spectra were recorded with a Bruker EMX Plus spectrometer equipped with a standard ER4102ST Bruker cavity, either at room temperature or after heating. The instrument settings were the following: receiver gain 30 dB, modulation amplitude 1 G, and microwave power 0.02 mW. The tested samples were aqueous solutions of CuSO 4 , Phen and Cu-Phen (called activators) with or without hydrogen peroxide, in the presence of DMPO as a spin-trap and DMSO or sodium formate as hydroxyl radical scavengers. After preparation, the solution was transferred to a bottom-sealed Pasteur pipette and immediately analyzed. For the trials requiring heating, a heat gun was used to heat the pipette for approximately 15 s, after which the pipette was placed in the EPR cavity. All spectra in Section 3 display the same intensity scale. Note that a first measurement was made immediately and was repeated after 5 and/or 10 min to verify that the signal was quite stable. As it was, the presented spectra were considered representative of the system during the first 10 min.

Result Analysis: Integration and Simulations
Most EPR simulations were conducted using the free isotropic simulation program SimEPR, provided by the (US) National Institute of Environmental Health Sciences' Public Electron Paramagnetic Resonance Software Tools [27,28]. It was useful to confirm the assignment of the experimental constants to the most probable radical species.
For the EPR experiments at 100 K, the Easyspin software [29] was used with graphical interface Simultispin [30].
Additional data regarding the simulation process and fitted spectra can be found in the Supplementary Materials.

Results and Discussion
The formation of HO • was evaluated using the DMPO spin-trap, both in the absence of alkali and under strong alkaline conditions. The main results are gathered in Table 1a,b. The entry numbers will be used in the paper to refer to each experiment, e.g., DMPO + CuSO 4 is "exp 1". Table 1: (a) Integration and simulation results at near-neutral pH: total peak area, and distribution of DMPO adducts and degradation products for some experiments, as relative areas of each radical, in percent. (b) Integration and simulation results at alkaline pH (with NaOH): total peak area, and distribution of DMPO adducts and degradation products for some experiments, as relative areas of each radical, in percent.

Results at Near-Neutral pH
Before investigating the decomposition of H 2 O 2 by Cu-Phen, the stability of DMPO in the presence of Cu-Phen (exp 2) was evaluated ( Figure 3). In the presence of DMPO, the EPR spectrum of a solution of Cu-Phen displays features corresponding to three different radical adducts: DMPO-OH, DMPO-R, and a radical showing triplet splitting lines ( Figure 3a). The addition of DMSO (exp 13) does not reduce the DMPO-OH signal, indicating that DMPO-OH does not originate from the trapping of HO • , but from nucleophilic addition of water. This was confirmed using the other radical scavenger, HCOO − (exp 14), as illustrated in Figure S6. Note that the hyperfine splitting constants for DMPO-OH in the presence of DMSO (and water) are different from those in water only, as described in the literature [31,32].
The DMPO-R and the species that shows triplet splitting lines (we will name it "triplet radical" in figure legends for easier reading) are most likely due to the decomposition of DMPO in the presence of copper. Controls have been carried out to understand the origin of these species. While DMPO does not exhibit any signal in the absence or presence of phenanthroline, an EPR signal is observed in the presence of CuSO 4 (exp 1), similar to that of DMPO + Cu-Phen, but less intense (about 50%; see Table 1a entries 1 and 2 and Figure 3b). The higher intensity of the DMPO-R and triplet signals observed with Cu-Phen compared to CuSO 4 demonstrates that DMPO is degraded faster in the presence of Cu-Phen. This is consistent with the redox potential of Cu II (Phen) 2 /Cu I (Phen) 2 that is slightly higher than that of Cu II /Cu I (0.17 and 0.16 V, respectively) [15]. In the present work, the coordination of DMPO to the Cu(II) ion was evidenced by EPR spectra recorded at 100 K with DMPO 60 mM and Cu(II) 0.15 mM (Figure 4). Under such conditions (low temperature and Cu(II) concentration 10 times larger than in the spin-trapping experiments), the EPR signal corresponding to the Cu(II) ion (S = 1/2, d 9 ) is observed with a typical axial spectrum characterized by a quartet in the parallel component arising from hyperfine interaction (I Cu = 3/2). The addition of DMPO notably modifies the EPR spectra of both Cu-Phen and CuSO 4, as demonstrated by changes in the g-value anisotropy. Such changes evidence the difference in coordination geometry around the Cu ion. The present work also highlights the fact that Phen remains bound to Cu when DMPO and Cu-Phen are mixed, since the spectra of Cu-DMPO adducts with or without Phen display different EPR parameters (see Table S24 in the Supplementary Materials). Besides, the absence of a decrease of the signal integration when DMPO is added demonstrates that Cu(II) is not reduced by DMPO into Cu(I) under these experimental conditions.
The H 2 O 2 /Cu-Phen system (exp 3) was tested in the presence of DMPO ( Figure 5), leading to an intense DMPO-OH spectrum, a multiplet assigned to DMPO-R, and a triplet splitting. The simulations confirm that the DMPO-R and the radical that shows triplet splitting lines are the same as without H 2 O 2 . The DMPO-OH signal is 4.8 times more intense than with Cu-Phen alone (38% DMPO-OH for a total area of 1.29 (entry 3 in Table 1) vs. 7% for a total area of 1.46 (entry 2)). This already indicates that hydroxyl radicals are the major species responsible for the formation of DMPO-OH with H 2 O 2 /Cu-Phen. Note that no hydroxyl radicals are present at room temperature with hydrogen peroxide alone.
In the presence of DMSO (exp 4, see Figure 5 and entry 4 in Table 1a), a drop of the DMPO-OH signal is accompanied with the detection of a new carbon-centered DMPO adduct, assigned to DMPO-CH 3 based on its EPR parameters [33,36]. DMPO-CH 3 is highly predominant, representing 73% of the detected radicals. This was confirmed when sodium formate was used as scavenger (exp 5, Figure S9), with 49% of DMPO-COO − among the detected radicals (entry 5 in Table 1a).
These experiments demonstrate that hydroxyl radicals are produced in aqueous solution in the presence of both H 2 O 2 and Cu-Phen. The apparent disappearance of the DMPO-OH signal in the presence of both radical scavengers is consistent with the absence of nucleophilic water addition in this case. This can be explained by the low constant rate of the DMPO/H 2 O reaction, or by the fact that copper activates H 2 O 2 s decomposition rather than induces nucleophilic addition of water onto DMPO. When analyzing the solution after 10 min at room temperature ( Figure S10), the DMPO-COO − signal is more intense than after direct acquisition, meaning that the system is still active after a few minutes.  Table 1a). The DMPO-OH signal is very weak compared to that observed earlier with H 2 O 2 /Cu-Phen. Since DMSO does not seem to scavenge hydroxyl radicals (exp 9), sodium formate was also tested (exp 5, Figure 6). In its presence, the DMPO-OH signal almost disappears with the concomitant appearance of the signal typical of DMPO-COO − . This shows that only a low amount of hydroxyl radicals is generated under such conditions. Consistently, the group of Bhattacharjee [37] also observed DMPO-OH at physiological pH with DMPO, H 2 O 2 , and CuCl 2 . However, they did not identify the origin of the DMPO-OH adduct because they did not use a radical scavenger.
As illustrated in Figure 6, the DMPO-COO − signal with H 2 O 2 + Cu-Phen is more intense than with H 2 O 2 + CuSO 4 , approximately 85% higher (total area of 0.65 including 49% DMPO-COO − vs. 0.21 including 23% DMPO-COO − , see entries 5 and 10 respectively in Table 1a). This unambiguously confirms the stronger H 2 O 2 decomposition in the presence of Cu-Phen. Similarly, in the context of dye decolorization, Nerud et al. [38] observed that H 2 O 2 /Cu II -pyridine was more effective than H 2 O 2 /Cu II at pH 3 to 9. The authors proposed that hydroxyl radicals were probably involved in the decolorization, since HO • scavengers such as superoxide dismutase inhibited the color-stripping. Since H 2 O 2 /Cu-Phen is the most efficient system to produce hydroxyl radicals, its efficiency was compared to a reference system: the Fenton reagent. FeSO 4 was thus used instead of Cu-Phen, under strong acidic conditions, with the same H 2 O 2 and metal concentrations (exp 11, Figure 7).  Table 1a). Yet, it is 25% less intense than for H 2 O 2 /Cu-Phen (total area of 0.43 including 86% DMPO-OH found with H 2 O 2 /Fe, compared to 1.29 including 38% DMPO-OH with H 2 O 2 /Cu-Phen). Assuming that nucleophilic addition of water does not occur neither in the Fenton system nor in the H 2 O 2 /Cu-Phen system (as discussed earlier), the H 2 O 2 /Cu-Phen system is thus a little more active than the Fenton reactant under similar conditions. Still, the HO • generation of H 2 O 2 /Cu-Phen is in the same range as that of the Fenton system.

Results at Alkaline pH
At alkaline pH, DMPO is degraded faster than at near-neutral pH, both alone and with phenanthroline (same intensity, see Figure S11), and the three-line signal attesting the DMPO degradation is observed even in the absence of copper or hydrogen peroxide. This is consistent with easier nucleophilic addition at alkaline pH.
With CuSO 4 (exp 17), the signal exhibits mostly DMPO degradation products, whereas with Cu-Phen (exp 18), the DMPO-OH signal can be clearly observed ( Figure S12). Therefore, the Cu-Phen solution (the main species being CuPhen(OH) 2 in this case) was analyzed in the presence of the sodium formate scavenger (exp 19, Figure S13), leading to the generation of the typical DMPO-COO − signal, and thus confirming that hydroxyl radicals are generated under such conditions. The generation of HO • in the absence of hydrogen peroxide can be rationalized by the catalytic reduction of soluble oxygen by Cu-Phen under alkaline conditions [39][40][41].
As previously examined under near-neutral conditions, the influence of DMPO on the electronic structure of the Cu II ion was tested at alkaline pH. The difference in the 100 K-EPR spectra of Cu-Phen recorded at both pHs (see Figures S19 and S23) is consistent with the binding of a hydroxo ligand at the Cu site [36]. When DMPO is added, a new EPR spectrum with well-resolved hyperfine couplings is observed, distinct from that obtained with CuSO 4 at alkaline pH. This agrees with the fact that DMPO can coordinate copper at both alkaline and near-neutral pHs (Figure 8). When H 2 O 2 is added to the Cu-Phen system (exp 20), a DMPO-OH signal appears (Figure 9). HO • scavenging with DMSO (exp 21) leads to the disappearance of the DMPO-OH signal and the apparition of DMPO-CH 3 , thus confirming that the DMPO-OH signal is only due to the presence of hydroxyl radicals (Figure 9a). As a confirmation, the formation of DMPO-COO − is observed in the presence of sodium formate (exp 22, Figure 9b). However, the generation of HO • is less intense ( Figure 10) and slower ( Figure S15) than at lower pH. Note that as observed at near-neutral pH, the hydroxyl radical is not produced in the alkaline H 2 O 2 /CuSO 4 system ( Figure S16).

Concluding Remarks
Under near-neutral and alkaline conditions, DMPO was shown to be degraded in the presence of copper (CuSO 4 and Cu-Phen), but not with phenanthroline alone. Indeed, nucleophilic addition of water onto DMPO was evidenced, leading to the formation of the DMPO-OH adduct in the absence of hydroxyl radicals. To identify the origin of the observed DMPO-OH, two HO • scavengers (sodium formate and DMSO) have been compared, and their efficiency is pH dependent. While sodium formate was found to be the best scavenger at near neutral pH, under alkaline conditions, both scavengers display similar activity. Besides, we have shown that DMPO coordinates the copper(II) ion under both conditions in the presence or absence of Phen.
The addition of Cu-Phen into a hydrogen peroxide solution at near-neutral pH and alkaline pH was proven to cause its decomposition into hydroxyl radicals. Interestingly, at near-neutral pH, the Cu-Phen system demonstrates an efficiency similar to that of the Fenton system at acidic pH, while CuSO 4 displays only a poor production of hydroxyl radicals. This evidences the role of the Phen ligand.
The presence of Phen influences both the electronic structure of the copper(II) complexes and their structural properties. Accordingly, Burkitt et al. [15] showed that at neutral pH, the redox potential of Cu II (Phen) 2 /Cu I (Phen) 2 is slightly higher than that of Cu II /Cu I (0.17 and 0.16 V, respectively), consistently with higher degradation of DMPO in the presence of Phen.
Author Contributions: This study was supervised by N.M. and G.M. Formal analysis and investigation were performed by E.W., with the help of F.M. who ran the spectroscopic measurements and software. E.W. wrote the original draft with the guidance of C.D., and F.M. and C.D. participated to review and editing. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study is available in the article and supplementary material.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.