Structural Effects on the Reaction of Singlet Oxygen with Tertiary Amines in Aqueous Solution

Abstract

Photosensitized excitation of molecular oxygen generates singlet oxygen, a reactive oxygen species that has been studied in biological systems, synthetic methods and in aquatic ecosystems. The reaction of singlet oxygen with tertiary amines is important because they are widely used as electron donors in photochemical reactions. Herein we studied the reaction of singlet oxygen with multiple tertiary amines including ethylenediamine tetraacetic acid (EDTA), triethanolamine (TEOA) and triethylamine (TEA). Singlet oxygen was generated using the photosensitizers methylene blue or chlorin e6 and red light with output at 660 nm. TEOA and TEA generated more hydrogen peroxide (H₂O₂), the stable end product, than EDTA at all pH values tested and regardless of the photosensitizer used. Both histidine and imidazole scavenged singlet oxygen and decreased H₂O₂ yield. The extent of histidine scavenging was pH-dependent for the combination of methylene blue and EDTA but not for TEOA or TEA. The combination of chlorin e6 and EDTA generated less H₂O₂ because both contain multiple negative charges that limit their interaction. Multiple tertiary amines that are used as biochemical buffers produced similar quantities of H₂O₂ as EDTA, TEOA and TEA. However, these sulfonic acid-containing tertiary amines did not function as electron donors in a benzoquinone photoreduction assay.

1. Introduction

Singlet oxygen (¹O₂) that is formed by photoexcitation of paramagnetic triplet molecular oxygen (O₂), is diamagnetic, short-lived and reacts readily with a range of organic functional groups [1,2,3]. Photosensitized generation of an excited state of O₂ was verified by Foote and Wexler in 1964 [4]. Although ¹O₂ forms in some light-independent reactions, most ¹O₂ research focuses on its generation using a photosensitizer and light. Some commonly used chemical photosensitizers include methylene blue (MB), Rose Bengal and phenosafranine, as well as the biological molecules riboflavin and chlorin e6, a chlorophyll metabolite. ¹O₂ reactivity has been explored in a range of contexts from synthetic methodologies, environmental reactions in aquatic ecosystems and in biological systems [1,5,6].

Chlorophyll is the light-harvesting pigment of photosynthesis and photoexcitation of it by red light initiates electron transfer to plastoquinone, a para-benzoquinone that is embedded in chloroplast membranes of plant cells [7,8]. The ultimate end products of electron transfer in photosynthesis are glucose and O₂. Although photoreduction is the predominant outcome, when light is excessive, the photoexcited state of chlorophyll initiates ¹O₂ formation [9]. As a result, plants contain multiple antioxidants that scavenge ¹O₂ to limit cellular damage [10].

In another biological context, the combination of photosensitizers, O₂ and light is used to generate ¹O₂ in photodynamic therapy (PDT) [11,12,13]. PDT is used to initiate cell death in dermatology, oncology and ophthalmology [11]. ¹O₂ reacts with a wide range of biological molecules including DNA, protein amino acids and lipids [5,14]. These reactions often result in the formation of additional, more persistent reactive oxygen species (ROS) including superoxide anion (O₂^−•) and hydrogen peroxide (H₂O₂) [15].

In our studies of benzoquinone and naphthoquinone photoreduction by MB and select chlorophyll metabolites, we routinely use tertiary amines as electron donors [16,17,18]. Ethylenediamine tetraacetic acid (EDTA), triethanolamine (TEOA) and triethylamine (TEA) are widely used electron donors in photoreduction processes because they are inexpensive, and their decomposition products are well-characterized and inert [19]. In one key experiment, we showed that the catechol neurotransmitter, dopamine, could be photo-oxidized by ¹O₂ to its ortho-quinone and then photoreduced in situ using a combination of MB or a chlorophyll metabolite and a tertiary amine electron donor [17].

Our interest in the dual nature of photosensitizers as both mediators of photoreduction and photo-oxidation prompted this work. Under ambient O₂ conditions when photoreduction was first initiated, the reaction of ¹O₂ with the tertiary amine electron donors produced some H₂O₂. Because the only reported rate constant for EDTA and ¹O₂ was ~10⁵ M⁻¹ sec⁻¹, we dismissed this reaction as negligible because photoreduction dominated [20]. However, the ratio of quinone photoreduction to ¹O₂-mediated H₂O₂ formation from the tertiary amines appeared to be dependent on the photosensitizer employed [16].

The competing reaction of ¹O₂ with histidine, a well-characterized scavenger, is also of interest because we showed that histidine enhanced rates of photoreduction and decreased H₂O₂ formation [16,18,21]. Histidine was superior to azide, also an ¹O₂ scavenger, because it reacted with ¹O₂ in a 4 + 2 cycloaddition reaction to effectively remove O₂ from solution. This differs from azide, where O₂ is regenerated and additional azide-based radicals form [22,23]. Rate constants on the order of 10⁷ M⁻¹ sec⁻¹ in aqueous solution have been reported for the reaction of histidine and ¹O₂ though there is some pH dependence with faster rates at higher pH [24,25,26,27].

Further, the reaction of ¹O₂ with tertiary amines is worthy of further investigation because many common biochemical buffers are tertiary amines, including HEPES, PIPES, MES and MOPS. Only limited research has examined the reaction of tertiary amines with ¹O₂ and, to our knowledge, not across a range of pH values using different photosensitizers [6,15,20]. Herein, we employed the straightforward detection of H₂O₂, the end product of ¹O₂-mediated oxidation of tertiary amines, to explore this reaction in more detail. Our objectives were three-fold: (1) to examine the effects of pH and photosensitizer on ¹O₂ reactions with select tertiary amines; (2) to assess histidine and imidazole scavenging of ¹O₂ during reactions with tertiary amines; and (3) to predict reaction outcomes of ¹O₂ with tertiary amines and histidine or imidazole using published rate constants as a guide.

2. Materials and Methods

Materials: All chemicals were from Fisher Scientific (Waltham, MA, USA) or Sigma (St. Louis, MO, USA) and were of the highest purity available. Chlorin e6 and all benzoquinone stock solutions were prepared in DMF and used immediately or stored at −20 °C for up to a month. All other solutions were prepared in water or 10 mM PB pH 7.4. Tertiary amine stock solutions in water (100 or 200 mM) were adjusted to pH 7.4 with NaOH. All phosphate buffers (PB) were equilibrated to room temperature (20–22 °C) to ensure no variation in dissolved O₂. Histidine and imidazole solutions in water were prepared fresh daily.

Red light specifications: A 36-watt red light composed of eighteen 2-watt LEDs was used for all photochemical experiments. Samples were placed under the lamp on an aluminum-lined surface in a shallow box (11 cm from light source to surface). The maximum wavelength of emitted light was 660 nm. The intensity was quantitated in lux, and intensity (as a function of the distance from the light source to the samples) was measured regularly to ensure consistent exposure, as described in [17]. Red light irradiation times were optimized to limit photobleaching of MB or chlorin e6 and did not exceed 7.5 min.

Detection of H₂O₂ with TMB and HRP: For reactions containing only photosensitizers and tertiary amines, TMB and HRP were used to detect H₂O₂. Aliquots (20 μL) of photochemical reactions were combined with 0.5 mM TMB and 0.5 μM HRP in 100 μL. After blue color development, 100 μL 1 M HCl was added. Absorbance was read at 450 nm in a 96-well plate. A standard curve from 0 to 40 μM H₂O₂ was used to calculate H₂O₂ concentrations in the photo-oxidation reactions (Supplemental Figure S1).

Detection of histidine and imidazole: Histidine or imidazole were reacted with diazotized sulfanilic acid under alkaline conditions to form an azo dye based on published methods [28,29]. Briefly, 1% sulfanilic acid in 1 M HCl and 5% NaNO₃ were combined 1:1 to generate diazotized sulfanilic acid. Dark and light-exposed samples containing histidine or imidazole (10 μL aliquots) were combined with 30 μL of diazotized sulfanilic acid solution in a 96-well plate. Na₂CO₃ solution (60 μL, 10%) was added and the resulting red-orange coupling product was detected at 490 nm. Standard curves from 0 to 150 μM histidine or imidazole were used to determine concentrations in photochemical reactions. (Supplemental Figures S2 and S3)

Photoreduction of CoQ₀: Solutions (100 μL total) of CoQ₀ (1 mM), 5 μM MB and 5 mM of each tertiary amine in 10 mM PB 7.4 were aliquoted into a 96-well plate. Absorbance was measured at 405 nm prior to light exposure and at time intervals of 1, 2, 5 and 7 min. Controls with only CoQ₀ and MB (no tertiary amine) were also performed.

Photoreduction of CoQ₀, methyl and methoxyBQ rate enhancement by histidine Benzoquinones (0.6 mM) were combined with 2 μM MB and 2 mM EDTA (1 mL) in a semimicro cuvette in 10 mM PB pH 7.4. Absorbance scans from 250 to 750 nm were collected prior to light exposure and at time intervals of 1, 3 and 5 min. Histidine or imidazole (up to 1 mM) were added prior to light exposure. The cuvette was inverted three times to normalize O₂ concentration. HRP was added after irradiation and the resulting increase in oxidized benzoquinone absorbance was used to calculate H₂O₂ concentration. For CoQ₀, absorbance at 405 nm was used to calculate concentration (740 M⁻¹ cm⁻¹ in 10 mM PB pH 7.4) [16]. For methoxyBQ, absorbance at 366 nm was used. Oxidized methylBQ absorbs weakly at 320-330 nm; therefore, absorbance of reduced methylBQ at 286 nm was used to determine concentration. Sodium ascorbate or NaBH₄ were added after dark scans or after the quinones had been photoreduced to test for complete photoreduction (~5 equivalents).

Data analysis: For each experiment, at least three independent experiments were performed in duplicate or triplicate. Mean values were calculated for each independent experiment (mean ± error). For figures showing error bars, mean values were averaged and error calculated. Details for each are stated in Figure Legends.

3. Results

3.1. Generation of Singlet Oxygen

Singlet oxygen is generated by the combination of visible light, a photosensitizer (PS) and O₂. Its initial reaction with tertiary amines (R₃N:) produces superoxide anion and an amine radical cation, as shown in Scheme 1. Two superoxide anions disproportionate to yield H₂O₂ and regenerate O₂. We employ tertiary amines as sacrificial electron donors in our photoreduction studies and know that photo-oxidation processes mediated by ¹O₂ occur concurrently when O₂ is present [16,18,21]. Only one report determined a rate constant in water for the reaction of ¹O₂ with EDTA, a commonly used tertiary amine, and some metal-EDTA complexes [20]. For these reasons, we sought to study the reaction of ¹O₂ with commonly used tertiary amines in more detail.

3.2. Detection of H₂O₂

To detect H₂O₂ generated by the reaction of ¹O₂ with tertiary amines, we used horseradish peroxidase (HRP) and a colorimetric co-substrate, TMB, that generates a stable yellow product under acidic conditions [30,31,32]. If catalase, an enzyme that scavenges H₂O₂ (Equation (1)), was added prior to HRP and TMB, no yellow color was detected.

2H₂O₂ → 2H₂O + O₂

(1)

To react with ¹O₂ as shown in Scheme 1, tertiary amines must be deprotonated. Therefore, we examined the pH dependence of H₂O₂ formation using the tertiary amines EDTA, triethanolamine (TEOA) and triethylamine (TEA) (Scheme 2). We routinely perform photoreduction studies with methylene blue (MB) as the photosensitizer and a red light with maximum output at 660 nm. The combination of MB and 36-watt red light also generates ¹O₂ efficiently; therefore, it was used herein. We use phosphate buffers (PBs) when studying reactions that generate ¹O₂ and other ROS because phosphate is already oxidized and inert.

3.3. Effect of pH on H₂O₂ Yield

Figure 1 shows that for the three tertiary amines tested, the H₂O₂ concentration increased as buffer pH increased. In the absence of red light or MB, no H₂O₂ was detected. Two ¹O₂ are required to generate one H₂O₂ but, because one O₂ is regenerated, the stoichiometry is one O₂ consumed per H₂O₂. The concentration of dissolved O₂ is ~300 μM in 10 mM PB at 20–22 °C. The H₂O₂ concentrations in Figure 1 are all below 300 μM, confirming that dissolved O₂ was not depleted after 5 min of photo-oxidation.

For EDTA, the H₂O₂ concentration increased two-fold as pH increased from 5 to 8. For TEOA and TEA, H₂O₂ concentrations increased four-fold from pH 5-8 and their yields were equivalent at all pH values tested, despite a substantial difference in amine pK_a values. TEOA and TEA have pKa value of 7.78 and 10.75, respectively. EDTA has two amine groups with pKa values of 6.13 and 10.37; however, despite its lower pKa value and two amines/mole, EDTA generated less H₂O₂ than TEOA or TEA at all pH values tested.

3.4. Singlet Oxygen Scavenging by Histidine

One goal of this work was to examine relative rate constants for the reaction of ¹O₂ with tertiary amines and the amino acid histidine and imidazole, the R group of histidine (Scheme 1). Histidine is a well-characterized ¹O₂ scavenger that we have employed in our photoreduction studies [24,26]. We reported that histidine enhanced rates of benzoquinone and naphthoquinone photoreduction by limiting the reaction of newly reduced quinols with ¹O₂ that was generated concurrently under ambient oxygen conditions [16,18,21]. Further histidine scavenges ¹O₂ to form an oxygenated species, oxo-histidine, thereby depleting O₂ from solution.

To ensure that histidine did not interfere with H₂O₂ detection, up to 1.5 mM histidine was combined with HRP, TMB and known amounts of H₂O₂. Absorbance readings for oxidized TMB with histidine or imidazole were identical to those of controls that did not contain either. We also performed control experiments where histidine and MB were irradiated (no tertiary amine) in 10 mM PB pH 7.4 for 5 min. Under these conditions, no H₂O₂ was generated, which is consistent with the mode of ¹O₂ scavenging by histidine.

Figure 2A shows that the addition of histidine decreased H₂O₂ formation by EDTA, TEOA and TEA in 10 mM PB at pH 7.4 in a dose-dependent manner. The concentration of each amine was 5 mM and the highest histidine concentration was 1.5 mM. For TEOA and TEA, H₂O₂ yield decreased by 75%, with only 0.5 mM histidine and at 1.5 mM histidine, the decrease was 90%. By contrast, for EDTA, with 1.5 mM histidine, H₂O₂ decreased by 60% from 83 μM to 34 μM.

To find a more linear range of histidine concentrations required to decrease H₂O₂ yield with TEOA and TEA, we tested 0.1 to 0.5 mM histidine (Figure 2B). This lower range of histidine concentrations produced the desired dose–response. Again, TEOA and TEA generated equivalent H₂O₂ concentrations despite differences in their amine pKa values.

Based on these results, competition assays to determine relative rate constants for the reaction of ¹O₂ with tertiary amines and histidine are not straightforward. Figure 1 indicates that the structure of the tertiary amine influences its reaction of ¹O₂ with much higher H₂O₂ yields for the neutral amines, TEOA and TEA. With its multiple negative charges, EDTA may be less reactive because ¹O₂ cannot approach it as readily. Given that O₂ is hydrophobic, it may not approach positively charged MB that is associated with EDTA. Further, histidine’s imidazole ring has a pKa of ~7 and it too will have a partial position charge at pH 7.4.

To address this, ¹O₂ scavenging by histidine with EDTA as the tertiary amine was examined in 10 mM PB at pH 6.4, 7.4 and 8.0. As in Figure 1, H₂O₂ yield from EDTA (no histidine) was pH dependent, with increased H₂O₂ as pH increased. For all pH values tested, we observed dose-dependent decreases in H₂O₂ with histidine; however, the decrease was greater at pH 8.0 than at pH 7.4 and much lower at pH 6.4. With 5 mM EDTA and 1.5 mM histidine at pH 6.4, H₂O₂ yield decreased by only 20% vs. 60% at pH 7.4 and by 65% at pH 8.0. The striking change between pH 6.4 and 7.4 implicates the protonation of histidine’s imidazole ring for optimal photo-oxygenation by ¹O₂. Because the imidazole ring undergoes a 4 + 2 cycloaddition with ¹O₂, its protonation may limit reactivity. Alternatively, the protonated imidazole ring of histidine at lower pH may associate with the negative charges on EDTA in a manner that limits its exposure to ¹O₂.

For TEOA and TEA, we did not observe pH-dependent differences in histidine scavenging of ¹O₂ as reflected in H₂O₂ concentration endpoints. As in Figure 1, there was pH dependence for H₂O₂ yield by TEOA and TEA but at pH 6.4, 7.4 and 8.0, H₂O₂ yield consistently decreased by 70–75% with only 0.5 mM histidine present. While this argues against a dependence on imidazole protonation, both TEOA and TEA are neutral and would not interact with the positively charged imidazole ring of histidine.

3.5. Singlet Oxygen Scavenging by Imidazole

When imidazole rather than the amino acid histidine was employed as the ¹O₂ scavenger in competition studies, we also observed a decrease in H₂O₂ yield from EDTA (Figure 3A) as imidazole increased. H₂O₂ yield decreased by 36% from 83 μM (no imidazole) to 53 μM with 1.5 mM imidazole. This differs from Figure 2A, where 1.5 mM histidine decreased H₂O₂ yield by 60% to 34 μM.

In the absence of any tertiary amine, the combination of MB, 1.5 mM imidazole and red light produced 18 μM H₂O₂ (Figure 3A). This was not observed for histidine. The addition of catalase prevented a color change in our HRP assay, confirming that H₂O₂ formed. If the H₂O₂ yield from 1.5 mM imidazole alone (no EDTA) were subtracted from the total H₂O₂ yield of 53 μM (from both imidazole and 5 mM EDTA), it would match that of 1.5 mM histidine and 5 mM EDTA in Figure 2A. There is no method to discern if H₂O₂ was produced from the reaction of ¹O₂ with EDTA or with imidazole.

When additional tertiary amines were assayed with 1.5 mM imidazole, we observed decreased H₂O₂ yield (Figure 3B). Because TEOA and TEA results were identical in Figure 1 and Figure 2, only TEOA is shown in Figure 3B. In addition to TEOA and EDTA, we assayed bicine, which is a hybrid of TEOA and EDTA that has only one negative charge but an amine pKa value of 8.3 (vs. 7.8 for TEOA).

H₂O₂ yield from the reaction of bicine with ¹O₂ was greater than EDTA but less than that of TEOA, as predicted from its hybrid structure. The percent decrease in H₂O₂ yield with 1.5 mM imidazole followed the same trend. For bicine and imidazole, H₂O₂ decreased by 57% vs. 70% for TEOA and by 36% for EDTA and imidazole. As in Figure 2A, some portion of the H₂O₂ detected for amine plus imidazole may be attributed to the reaction of imidazole with ¹O₂.

3.6. Chlorin e6 as Photosensitizer for Singlet Oxygen Production

All data presented in Figure 1, Figure 2 and Figure 3 used 5 μM MB as the photosensitizer. Chlorin e6 is a chlorophyll derivative with three carboxyl groups that will be deprotonated at neutral pH (Scheme 3). It is an effective photosensitizer used in photodynamic therapy to target a wide range of cancer cells [11,12,33]. With its negatively charged carboxylates, it is different from MB, which has one positive charge at neutral pH.

In Figure 4A, we tested the effect of the photosensitizer on the reaction of ¹O₂ with EDTA, TEOA and TEA at multiple pH values. Figure 4A shows that H₂O₂ yield from the reaction of ¹O₂ with EDTA did not increase as pH increased from 5 to 8 but rather remained constant (blue bars). This is in stark contrast with Figure 1, where H₂O₂ from ¹O₂ and EDTA increased two-fold in that pH range. Further, the absolute H₂O₂ values are also considerably lower in Figure 4A (15 μM) than in Figure 1 (83 μM) even though we used 15 μM chlorin e6 vs. 5 μM MB in Figure 1, Figure 2 and Figure 3.

For TEOA and TEA, H₂O₂ yield increased 6-fold and 7-fold, respectively, as pH increased from 5 to 8. At pH 8, H₂O₂ yield with 5 mM TEOA, 15 μM chlorin e6 and 5 min of red light was 65 μM vs. 170 μM with 5 μM MB in Figure 1. One report stated that chlorin e6 can form aggregates at or below pH 5 [34].While this may explain the pH-dependent yesincreases for TEOA and TEA above pH 5, it does not explain the lack of pH dependence for EDTA. We attribute the low H₂O₂ yield to the charge repulsion between EDTA with its four negative charges and chlorin e6 with three negative charges at all pHs tested. For this reason, chlorin e6 is unlikely to generate ¹O₂ in close proximity to EDTA.

Figure 4B shows that histidine competed with EDTA, TEOA and TEA for ¹O₂ when chlorin e6 was the photosensitizer, as evidenced by decreased H₂O₂ yield. The % decrease for TEOA and TEA with 0.5 and 1.0 mM histidine was comparable to that seen in Figure 2A, where MB was the photosensitizer. Because the H₂O₂ values for EDTA were so low in Figure 4B, 1.0 mM histidine was sufficient to block nearly all H₂O₂ formation.

By contrast, when imidazole was used as the ¹O₂ scavenger in competition with EDTA, TEOA and TEA, H₂O₂ concentrations did not increase to the same extent as with histidine (Figure 4C). Indeed, for EDTA and increasing imidazole, H₂O₂ yields increased. Figure 4C shows that, in the absence of amine, some H₂O₂ from imidazole was detected, in agreement with Figure 3A, where MB was the photosensitizer.

Clearly, both histidine and imidazole function as ¹O₂ scavengers to reduce H₂O₂ yield in Figure 2 and Figure 3. As a result, they will have been oxidized. To confirm this, we performed a colorimetric assay in which histidine or imidazole reacted with diazotized sulfanilic acid under alkaline conditions to form an azo dye product [28,29]. If the aromaticity of the imidazole ring is disrupted via an oxidation process, no colored product will form. Numerous reports show that both are oxidized to form 2-oxo-histidine and 2-oxo-imidazole.

Samples containing MB and histidine or imidazole with or without EDTA were irradiated and the amount of azo dye product was determined. The yield of colored product from identical dark samples was also measured so that concentrations consumed could be calculated. Supplemental Figures S4 and S5 show that more histidine or imidazole was consumed in the absence of EDTA, as expected. The concentrations of histidine and imidazole that were oxidized are higher than the H₂O₂ concentrations measured under identical conditions. To disrupt imidazole aromaticity, only one ¹O₂ is required per imidazole ring [29]. However, to generate one H₂O₂, two ¹O₂ are required, according to Scheme 1.

3.7. Tertiary Amine Buffers Generate Hydrogen Peroxide

Multiple common biochemical buffers that are tertiary amines including HEPES, PIPES, MOPS and MES were assayed for their reactivity with ¹O₂, using MB as the photosensitizer. Their structures are shown in Supplemental Figure S6. H₂O₂ yields for each in 10 mM PB at pH 7.4 are summarized in Figure 5A and compared with EDTA, bicine and TEOA. All generated H₂O₂ with nearly identical yields for the two piperazines, HEPES and PIPES. H₂O₂ yields for MOPS and MES that contain a morpholine ring were the lowest of those tested. HEPES, PIPES, MOPS and MES all contain sulfonic acids that will be deprotonated at neutral pH.

Time courses for H₂O₂ production for all tertiary amines were also performed to assess any variability (Figure 5B). Because the reaction of ¹O₂ with tertiary amines is dependent on dissolved O₂ concentration, H₂O₂ formation in the first 2.5 min was often greatest, with only modest increases over time for EDTA, bicine and TEOA in Figure 5A. Time courses for HEPES and PIPES are similar, as were those for MES and MOPS.

3.8. Select Tertiary Amines Function as Electron Donors in Photoreduction

Equal concentrations of each tertiary amine in Figure 5A,B were tested as electron donors in a benzoquinone photoreduction assay using MB as the photosensitizer. CoQ₀, a more water-soluble benzoquinone that is an analog of Coenzyme Q, was used because we routinely monitor its reduction at 405 nm [16]. Scheme 4 shows reduction in a para-quinone to the quinol form. The combination of MB and EDTA is very efficient, with 172 turnovers (mol CoQ₀ reduced per mol MB) (

Table 1. Comparison of H₂O₂ yield and photoreduction turnovers for tertiary amines.

Tertiary Amine	Amine pKa	[H2O2] μM	Mol CoQ0 Red/mol MB
EDTA	6.13, 10.37 *	108 ± 8	170 ± 2
Bicine	8.35	172 ± 2	142 ± 6
TEOA	7.78	183 ± 2	63 ± 2
TEA	10.75	187 ± 6	ND
HEPES	7.48	143 ± 1	ND
PIPES	6.76	150 ± 2	ND
MES	6.15	102 ± 1	ND
MOPS	7.15	81 ± 2	ND

* pKa values are at 25 °C. H₂O₂ reactions contained 5 μM MB and 5 mM tertiary amine that were irradiated for 7.5 min. ND = none detected.

). Bicine, and to a lesser extent TEOA, functioned as electron donors under the conditions employed, but the others did not.

In our prior work, we tested multiple benzoquinones in photoreduction assays using the combination of MB as photosensitizer and EDTA as the electron donor [16,21]. Under ambient O₂, photo-oxidation to generate ¹O₂ was a competing reaction that likely produced some H₂O₂ via the reaction of EDTA with ¹O₂ described herein. However, ¹O₂ also reacts with newly reduced quinols to regenerate oxidized benzoquinones in a photochemical redox cycle (Scheme 4) [35]. The reoxidation of the quinol also generated H₂O₂ via hydrogen atom abstraction to form RO• and a peroxyl radical that deprotonates to superoxide anion.

As in Scheme 1, superoxide anion disproportionation produces H₂O₂. When histidine was included, the rate of photoreduction increased, and less H₂O₂ was generated because ¹O₂ was scavenged by histidine. In support of photochemical redox cycling as the predominant source of H₂O₂, we determined that more H₂O₂ formed during photoreduction reactions with MB, EDTA and benzoquinone than in the absence of benzoquinone, where EDTA was the sole source of H₂O₂. We speculated that we could use rate enhancements by histidine due to ¹O₂ scavenging during photoreduction to evaluate which quinols were more susceptible to reoxidation by ¹O₂ [35,36].

3.9. Histidine Enhances Benzoquinone Photoreduction Rates

Given that this work explores competition between tertiary amines and histidine for ¹O₂, we examined photoreduction rate enhancement in more detail. Figure 6 shows rates of photoreduction for three benzoquinones alone and with three concentrations of histidine. For all three, as histidine increased, the photoreduction rate increased. Rate enhancement was modest for CoQ₀ and methoxyBQ, using 1.0 mM histidine with increases of 24% and 14%, respectively. With methylBQ, the rate enhancement was greater than 100%. These data show that of the three benzoquinones, methylBQ is the easiest to photoreduce because its rate in the absence of histidine was greatest. Further, its quinol is most reactive with ¹O₂ because of the dramatically increased rate when histidine limited reoxidation by ¹O₂ [35,36,37].

Further, as histidine increased, the yield of H₂O₂ during photoreduction for all three benzoquinones decreased, which is consistent with ¹O₂ scavenging. In these assays, we add HRP after photoreduction and use the resulting increase in oxidized benzoquinone to calculate H₂O₂ concentrations. Reduced quinols are HRP co-substrates. When imidazole rather than histidine was used as the ¹O₂ scavenger, we observed equivalent increases in photoreduction rates and similar decreases in H₂O₂.

Although only 2 μM MB and 2 mM EDTA were used for photoreduction (5 min max), H₂O₂ concentrations were as high as 250 μM, whereas with 5 μM MB and 5 mM EDTA, H₂O₂ never exceeded 83 μM after 5 min of irradiation (Figure 1). This supports reoxidation of the newly reduced quinol as the major source of H₂O₂ during photochemical redox cycling, not EDTA oxidation by ¹O₂.

4. Discussion

The results herein show that the tertiary amines we employ as electron donors in quinone photoreduction experiments produce H₂O₂ in a pH-dependent manner when they react with ¹O₂. The pH dependence for EDTA, TEOA and TEA in Figure 1 using MB as a photosensitizer is likely dependent on their chemical structure, polarity and to some extent, on their amine pKa values. The absolute amount of H₂O₂ formed is higher for the neutral tertiary amines, TEOA and TEA, vs. EDTA, which has four negatively charged carboxylates at all pHs tested. Although MB has a positive charge that would interact with EDTA, electrostatics do not appear to be advantageous because EDTA produced less H₂O₂ than TEOA or TEA.

H₂O₂ yields for TEOA and TEA were identical at all pH values tested, despite amine pKa values that differ by three orders of magnitude (Table 1). If amine pK_a were the only criterion for reactivity with ¹O₂, one would expect much lower H₂O₂ yields for TEA with its pKa of 10.75. However, even at pH 5, a small fraction of deprotonated TEA would be present and once it reacted, equilibrium would shift to generate more.

These data suggest that the interaction between the tertiary amine and ¹O₂ is also critical to the formation of H₂O₂. Both O₂ and ¹O₂ are nonpolar and therefore would be expected to have greater affinity for nonpolar TEA rather than TEOA with its three hydroxyl groups. The enhanced interaction between TEA and ¹O₂ due to their hydrophobicity may compensate for the difference in TEOA and TEA amine pKa values.

When chlorin e6 was used as the photosensitizer in Figure 4, TEOA and TEA were superior to EDTA at pH 7.4 and 8.0 but not at the lower pHs tested. We attribute this to chlorin e6 aggregation at low pH which limits its ability to generate ¹O₂ regardless of the tertiary amine used [34,38]. ¹O₂ clearly forms with chlorin e6 at pH 7.4 and 8.0, because we observed that H₂O₂ formed with TEOA and TEA in Figure 4.

The lack of pH dependence for EDTA with chlorin e6 in Figure 4 vs. MB in Figure 1 is likely due to charge repulsion between EDTA and chlorin e6, because both contain multiple negative charges at all pHs tested in Figure 4. While an attractive interaction between MB and EDTA was deemed inconsequential in Figure 1, the negative charge repulsion between EDTA and chlorin e6 likely plays a role in Figure 4. ¹O₂ formation requires direct energy transfer from the excited photosensitizer, PS* to O₂, followed by rapid reaction of ¹O₂ with the tertiary amine (Scheme 1).

In Figure 1, we used 5 μM MB vs. 15 μM chlorin e6 in Figure 4 (tertiary amines at 5 mM) because chlorin e6 was less reactive in our photoreduction experiments. We were hesitant to increase chlorin e6 concentrations because chlorophyll-derived photosensitizers can dimerize and require some organic solvent to ensure full solubility and reactivity [16,38].

The ability of histidine to scavenge ¹O₂ and decrease H₂O₂ yields is evident in Figure 2A,B. In these competition assays, up to 1.5 mM histidine competed with 5 mM tertiary amine for ¹O₂. The reported rate constant for histidine and ¹O₂ is ~10⁷ M⁻¹ s⁻¹, which is two orders of magnitude greater than that of ¹O₂ and EDTA [20,24,26,27]. For this reason, we expected as little as 0.015 mM histidine to compete with EDTA for ¹O₂ but it did not. Under our competitive conditions, the relative rate constants must be comparable at pH 7.4 when EDTA is the tertiary amine. With 5 mM TEOA or TEA, only 0.1-0.5 mM histidine competed effectively, suggesting that the rate constants differed by one order of magnitude (Figure 2B).

Our work herein is important, because it shows that the rate constants for ¹O₂ and different tertiary amines will vary and that the identity of the photosensitizer likely plays a role. The EDTA and ¹O₂ rate constant that was reported by Hessler et al. used Rose Bengal as the photosensitizer [20]. In that work, rate constants for EDTA and its metal complexes were determined from the quenching of the ¹O₂ phosphorescence at 1270 nm. Also, in some kinetic studies, reactions are performed in D₂O because it extends the half-life of ¹O₂ [39]. By comparison, we are using the endpoint of H₂O₂ in H₂O to assess competing ¹O₂ reactions.

Figure 2C shows that the ability of histidine to scavenge ¹O₂ was pH-dependent when EDTA was the tertiary amine, but not with TEOA and TEA. Our data showing greater histidine reactivity with ¹O₂ at pH 8.0 vs. 7.4 or 6.4 is consistent with the range of published rate constants for histidine and ¹O₂. Matheson and Lee reported that the rate constant for histidine and ¹O₂ varied with pH and identified a pKa of 6.9 [24]. This may be attributed to the protonation state of the imidazole ring of histidine, which reacts with ¹O₂ via 4 + 2 cycloaddition. Indeed, we saw the greatest increase in histidine scavenging between pH 6.4 and 7.4 (Figure 2C).

Figure 3A showed that imidazole was also effective as a ¹O₂ scavenger; however, in the absence of tertiary amine, low concentrations of H₂O₂ formed, which differed from histidine that was assayed at the same concentrations. We observed the same low H₂O₂ yields with two different commercial sources of imidazole (99% pure); however, the reasons for this remain unclear at this time.

When the contribution of H₂O₂ from the direct reaction of imidazole with ¹O₂ was subtracted, the H₂O₂ decrease with imidazole was comparable to the histidine results in Figure 2A. In Figure 3B, imidazole scavenging was better with bicine and TEOA than with EDTA, which is also consistent with Figure 2B.

As in Figure 3, imidazole results in Figure 4C, where chlorin e6 was used as the photosensitizer, are complicated by the generation of H₂O₂ in the absence of any tertiary amine. Absolute H₂O₂ values in Figure 4C are low due to the detrimental combination of chlorin e6 and EDTA. As a result, imidazole did not appear to scavenge H₂O₂. For this reason, we do not recommend imidazole as a ¹O₂ scavenger when H₂O₂ yield is the measured outcome.

The generation of H₂O₂ by the combination of MB, light and multiple tertiary amines that are frequently used in biochemistry is noteworthy (Figure 5A). Undesired photo-oxidation of HEPEs in cell culture experiments that also contained riboflavin has been reported [40,41]. Sun et al. also reported that HEPEs and several other tertiary amines reacted with ¹O₂ to produce H₂O₂ [15]. Their interest in this reaction focused on photodynamic therapies to target antibiotic-resistant bacteria.

To our knowledge, the reactivities of the additional tertiary amines in Figure 5 toward ¹O₂ using MB as the photosensitizer has not been explored. From a structural perspective, the piperazine-based HEPES and PIPES were more reactive than EDTA and the morpholines, MES and MOPS. None of them functioned as electron donors in a quinone photoreduction assay where EDTA and bicine, and to a lesser extent TEOA, afforded high turnover numbers (moles quinone reduced per mole MB in Table 1). Like EDTA and bicine, they have at least one negative charge at pH 7.4 due to their sulfonic acid groups. Their lack of reactivity in photoreduction is puzzling, because we believe that electrostatic interactions between MB and negatively charged electron donors like EDTA and bicine facilitate electron transfer during photoreduction.

Data in Figure 6 summarizing the effects of histidine on benzoquinone photoreduction provides further insights into the generation of H₂O₂ during photochemical redox cycling. Consistently with our prior work, we observed more photo-generated H₂O₂ in a mixture of MB, EDTA and quinone vs. MB and EDTA alone. More importantly, Figure 6 shows that each benzoquinone photoreduction rate is different, with the more easily reduced methylBQ showing the highest rate (in the absence of histidine). The rate of methylBQ photoreduction was also more sensitive to histidine. According to Scheme 4, the quinol of methylBQ must react with ¹O₂ more readily than the quinols of the other benzoquinones tested. This is consistent with published rate constants for ¹O₂ and other reduced benzoquinones and naphthoquinones [35,37,42].

Lastly, this work suggests a photochemical method to generate H₂O₂. Herein, we generated H₂O₂ using MB, multiple common tertiary amines, mild red light and ambient O₂ in neutral aqueous solution. While dissolved O₂ will become limiting, this can be rectified by bubbling solutions with air or O₂ gas. The current H₂O₂ industrial process relies on a Pd catalyst, H₂ and O₂ gases, anthraquinone as a redox cycler and a mixture of organic solvents [43,44]. These tertiary amines and MB are very water soluble and nontoxic, with no organic solvents required. Thus, our research on the photochemical reactions of chlorophyll metabolites and methylene blue may be of value for development of a green method for H₂O₂ synthesis.