Photochemical Redox Reactions of 2,6-Dichlorophenolindophenol and Its Use to Detect Photoreduced Quinones

Abstract

Photosynthesis in plants and the electron transport chain in mitochondria are examples of life-sustaining electron transfer processes. The benzoquinones plastoquinone and ubiquinone are key components of these pathways that cycle through their oxidized and reduced forms. Previously, we reported direct photoreduction of biologically relevant quinones mediated by photosensitizers, red light and electron donors. Herein we examined direct photoreduction of the quinone imine 2,6-dichlorophenolindophenol (DCPIP) using red light, methylene blue as the photosensitizer and ethylenediaminetetraacetic acid (EDTA) as the electron donor. Photoreduction of DCPIP by methylene blue and EDTA was very pH-dependent, with three-fold enhanced rates at pH 6.9 vs. pH 7.4. Photochemical redox cycling of DCPIP produced hydrogen peroxide via singlet oxygen-dependent reoxidation of reduced DCPIP. Histidine enhanced photoreduction by scavenging singlet oxygen, whereas increased molecular oxygen exposure slowed DCPIP photoreduction. Attempts to photoreduce DCPIP with pheophorbide A, a chlorophyll metabolite, and triethanolamine as the electron donor in 20% dimethylformamide were unsuccessful. Photoreduced benzoquinones including 2,3-dimethoxy-5-methyl-p-benzoquinone (CoQ₀), methoxy-benzoquinone and methyl-benzoquinone were used to examine electron transfer to DCPIP. For photoreduced CoQ₀ and methoxy-benzoquinone, electron transfer to DCPIP was rapid and complete, whereas for reduced methyl benzoquinone, it was incomplete due to differences in reduction potential. Nonetheless, electron transfer from photoreduced quinols to DCPIP is a rapid and sensitive method to investigate quinone photoreduction by chlorophyll metabolites.

1. Introduction

Redox processes with multiple electron transfer steps are integral to all life. In photosynthesis, light-dependent excitation of chlorophyll initiates electron transfer through several protein complexes with plastoquinone, a membrane-bound benzoquinone, as a key intermediary (Scheme 1) [1,2]. In the electron transport chain (ETC) of mitochondria, ubiquinone (also called Coenzyme Q) undergoes cycles of reduction and oxidation within the membrane as electrons flow through protein complexes [3,4]. Electrons are ultimately transferred to molecular oxygen (O₂) in a process coupled to ATP synthesis.

Our interest in light-mediated redox reactions of benzoquinones was prompted by reports that dietary chlorophyll metabolites remain photoexcitable after ingestion and therefore may affect health outcomes. Zhang et al. determined that chlorophyll metabolites accumulated in mouse brain, gut and fat by examining their diagnostic fluorescence emission [5]. In vitro work by Qu et al. showed that the combination of red light and chlorophyll metabolites reduced ubiquinone to its quinol form [6]. Their subsequent work using intact mouse mitochondria showed that ATP yield increased with the combination of chlorophyll metabolites and red light [7].

Previously, we have reported that benzoquinones and both ortho- and para-naphthoquinones can be photoreduced using a combination of red light, photosensitizers and tertiary amine electron donors [8,9,10]. We typically employ methylene blue, a well-characterized chemical photosensitizer, as well as several chlorophyll metabolites including pheophorbide A (pheoA), pyropheophorbide A and chlorin e6. PheoA is the primary chlorophyll metabolite derived from dietary chlorophyll [11,12]. Of particular interest is vitamin K, a group of fat-soluble naphthoquinones, because they must be reduced to perform their essential roles in blood coagulation [13]. Vitamin K is abundant in leafy greens, which are also excellent sources of dietary chlorophyll [14,15]. Further, many natural products contain quinone functionalities that may be potential photochemical substrates with health implications [16,17].

We routinely use UV/Visible absorbance measurements to study photoreduction of quinone substrates. Many of the chlorophyll metabolites absorb strongly ~400 nm, which coincides with the absorption of benzoquinone and naphthoquinone substrates, thereby complicating quantitation of photoreduction events. For this reason, we are interested in additional methods to detect photoreduced quinones.

2,6-dichlorophenolindophenol (DCPIP) is a well-characterized redox indicator that is blue in its oxidized form and colorless when reduced (Scheme 2) [18,19]. Oxidized DCPIP absorbs strongly at 605 nm (ε~19,100 M⁻¹ cm⁻¹ at pH 7.4) and is water-soluble. Its best-known use is as an artificial electron acceptor in the Hill reaction of chloroplasts [20]. When intact chloroplasts are irradiated, electron transfer through membrane-bound photosystems is initiated and electrons flow to DCPIP, rather than to NADP⁺. Hill was the first to report that O₂ was generated during this process [21]. DCPIP is also used routinely to measure ascorbic acid; electron transfer from ascorbic acid to it can be readily quantified at 605 nm [22,23].

DCPIP is a quinone imine, analogous to a quinone except that one of the oxygens is replaced by nitrogen [24]. Quinone imines typically react like quinones as electrophiles; oxidized acetominophen is one example [25]. The published reduction potential for DCPIP is 0.22 V vs. a standard hydrogen electrode, more positive than simple benzoquinones like 1,4-benzoquinone (0.10 V) or methylbenzoquinone (0.023 V) [18,26]. For this reason, we were also interested in studying photoreduction of DCPIP using the same combinations of red light, photosensitizers and tertiary amine electron donors that we employed in prior studies.

Herein, our objectives were two-fold: (1) to explore direct photoreduction of DCPIP by our optimized photoreducing systems and (2) to quantify photoreduced quinones using DCPIP as an electron acceptor.

2. Materials and Methods

All chemicals were from Fisher Scientific or Sigma (St. Louis, MO, USA) and were of the highest purity available. PheoA 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. Solutions of 2,6-dichloroindophenol (DCPIP) sodium salt (5 mM in water) were prepared fresh daily. EDTA and TEOA stock solutions were adjusted to 7.4 or 8.0 with NaOH or HCl. Phosphate buffers were equilibrated to room temperature (20–22 °C) to ensure no variation in dissolved O₂.

2.1. Red Light Specifications

A 36-watt red light composed of eighteen 2-watt LEDs was used for all photochemical experiments. 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. Samples were placed under the lamp on an aluminum-lined surface in a shallow box (11 cm from light source to surface).

2.2. Concentration Measurements

For CoQ₀, absorbance at 405 nm was used to calculate concentration (740 M⁻¹ cm⁻¹ in 10 mM PB pH 7.4) [9]. 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. DCPIP standard curves from 0 to 0.10 mM in 10 mM PB (pH 4.5–8) were linear.

2.3. DCPIP Photoreduction Assays

DCPIP (0.10 mM final) in 10 mM PB of varying pHs (pH 6.0, 6.4, 6.9, 7.4 or 8.0) was combined with 2 μM MB and 4 mM EDTA or 1 μM MB and 2 mM EDTA. DCPIP photoreduction was performed in either a semimicro cuvette (1 mL) or a 96-well plate (100 μL or 200 μL). Prior to absorbance readings, the cuvette was inverted three times to normalize dissolved O₂. Catalase (30 μL, 2 mg/mL) or histidine (0.5 to 1.5 mM) were added prior to red-light irradiation.

Red light irradiation times were optimized to limit photobleaching of MB or pheoA and did not exceed 8 min. Absorbance scans for the cuvette or 605 nm in the 96-well plate were conducted prior to irradiation and at time intervals. Dark samples containing all reactants served as controls.

2.4. Detection of H₂O₂ in Photoreduction Samples

HRP (8 μL, 2 mg/mL) was added directly into the cuvette solution (1 mL) after irradiation and mixed gently to avoid introducing O₂. For 96-well plate reactions, 2 or 4 μL HRP solution was added to 100 or 200 μL reaction volumes, respectively. Absorbance increases indicative of quinol reoxidation were monitored for up to 5 min until no further increase in absorbance occurred.

2.5. Quinone Photoreduction Assays

Quinones were combined with MB and EDTA (1 mL) in a semimicro cuvette. Typical concentrations of quinone, MB and EDTA were 0.4 mM, 2 μM and 2 or 4 mM, respectively, in 10 mM PB pH 7.4. For pheoA/TEOA, 20% DMF was required for optimal quinone photoreduction. PheoA (8 μM) and TEOA (16 mM) were employed to photoreduce 0.40 mM quinone. This maintains a molar ratio of quinone to photosensitizer of 1000:1 or 2000:1 to determine the turnover number. In some cases, sodium ascorbate or NaBH₄ were added after dark scans or after the quinones had been photoreduced to test for complete photoreduction (~5 equivalents).

2.6. Detection of Reduced Quinones with DCPIP

Quinone photoreduction reactions were prepared as described above in a cuvette. Aliquots (100 or 150 μL) were combined with DCPIP in 10 mM PB pH 7.4. Final concentrations were 80 μM DCPIP and 40 or 60 μM quinone (1 mL). UV/Vis scans were collected from 250 to 850 nm. For the 96-well plate assay, 30 μL quinone reaction (60 μM) was combined with 80 μM DCPIP in 200 μL. Solutions were mixing by pipetting gently 3–4 times to limit O₂ exposure. Absorbance readings at 605 nm (plate reader) were collected at 30 s intervals until the DCPIP absorbance was constant (up to 3 min).

2.7. Oxidation of Blue DCPIP by H₂O₂/HRP

DCPIP (80 μM) was combined with 20–80 μM H₂O₂ in 10 mM PB pH 7.4 and ~1 μM HRP (8 μL, 2 mg/mL) in a total volume of 1 mL. The decrease in absorbance at 605 nm was measured in semimicro cuvette. The color change was complete after 1.5 min.

2.8. Detection of H₂O₂ with Tetramethylbenzidine (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 1 mM TMB and 1 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 (based on 200 μL total volume) was used to calculate H₂O₂ concentrations in the photochemical reactions.

2.9. 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. Rate constants were determined by plotting the natural log of the DCPIP concentrations vs. time. Error in our rate constant or half-life calculations was not shown because all errors were less than 5%.

3. Results

We reported that p-benzoquinones, o- and p-naphthoquinones can be photoreduced by the combination of photosensitizers, red light and electron donors [9,10]. In those studies, we observed oxygen uptake and detected H₂O₂ because photoexcitation generated singlet oxygen (¹O₂) concurrently during quinone photoreduction. Newly photoreduced quinols reacted with ¹O₂ resulting in reoxidation to the quinone as depicted in Scheme 3.

Scheme 4 summarizes the photoreduction cycle where the sensitizer (PS) is excited to the triplet state, PS*. Once formed, PS* may react with O₂ to generate ¹O₂ or transfer an electron to the quinone substrate to yield the one-electron reduced semiquinone. The electron-deficient photosensitizer (PS^+•) is restored to its resting state by electron donation from the tertiary amine (R₃N). Two semiquinones (RO•) disproportionate rapidly to yield a quinol and regenerate a quinone [27].

H₂O₂ formation is the result of ¹O₂-mediated hydrogen atom abstraction from the quinol to produce a semiquinone and peroxyl radical (•OOH). Because the pK_a of •OOH is 4.8, it will deprotonate at neutral pH to a superoxide anion [28]. Disproportionation of two superoxide anions yields O₂ and H₂O₂ according to Equation (1).

O₂^−• + O₂^−• + 2H⁺ → H₂O₂ + O₂

(1)

Horseradish peroxidase (HRP), an enzyme that requires H₂O₂ to oxidize numerous organic substrates including catechols and quinols, was used to detect H₂O₂ [29,30]. Because quinols are oxidized by HRP only when H₂O₂ is available, benzoquinone or naphthoquinone substrates were regenerated when HRP was added following irradiation (Scheme 3). Previously, for 1,2-naphthoquinone, we monitored the loss of absorbance at 415 nm for photoreduction and the subsequent increase in A₄₁₅ when HRP was added [10]. If catalase was added prior to HRP, no increase in A₄₁₅ was detected because it consumed H₂O₂ according to Equation (2):

2 H₂O₂ → 2H₂O + O₂

(2)

Direct photoreduction of DCPIP was examined using the methodology described above for benzoquinones and naphthoquinones. Consistent with our earlier work, we employed the combination of methylene blue (MB) as photosensitizer and EDTA as the tertiary amine electron donor. We initiated this work using a 1 mL semimicro cuvette so that it could be irradiated and scanned directly at time intervals to limit mixing and O₂ exposure to only the O₂ that was in the solution.

Figure 1A shows that DCPIP absorbance decreased upon irradiation with a 36-watt red light with maximum output at 660 nm. Due to its strong absorbance at 605 nm, the highest concentration of DCPIP used was 0.10 mM. A standard curve of DCPIP was linear up to 0.10 mM; this ensures that decreases in absorbance are in the linear range (Supplemental Figures S1 and S2). As irradiation time increased, DCPIP absorbance at 605 nm decreased. In the absence of MB or EDTA, no decrease in absorbance was observed.

MB absorbs at 665 nm; therefore, after 6 min of light and when the DCPIP absorbance had decreased by ~90%, the MB peak became visible (Figure 1A). Only 2 μM MB was used; therefore, its absorbance did not interfere with DCPIP quantification at 605 nm. The MB peak observed after 6 min of light is essentially equal to that of the dark control, confirming that little to no MB photobleaching occurred.

DCPIP absorbance at 605 nm for each time point in Figure 1A was used to calculate concentrations shown in Figure 1B (in blue). As in our prior work using benzoquinones, HRP was added to determine if H₂O₂ formed and if reduced DCPIP was a co-substrate for HRP. Upon addition of HRP, the blue color of oxidized DCPIP was regenerated, indicative of H₂O₂. If catalase was added prior to HRP, thereby consuming H₂O₂, no increase at 605 nm was observed.

Multiple photoreduction experiments were performed to determine H₂O₂ concentration for each time interval. For example, a separate sample was irradiated for 2 min, scanned to measure photoreduction from the decrease at 605 nm, treated with HRP and rescanned to determine the H₂O₂ concentration from the increase at 605 nm. H₂O₂ concentration values determined by this method are shown in Figure 1B (orange).

According to our proposed mechanism, for each quinol that reacts with ¹O₂, one H₂O₂ molecule is formed. Because the typical concentration of dissolved O₂ at 20–22 °C is ~0.30 mM, the maximum concentration of H₂O₂ that could form is 0.30 mM if no additional O₂ enters the reaction vessel. The narrow dimensions of the cuvette limit O₂ from mixing with the solution. Figure 1B shows that H₂O₂ increased to ~0.09 mM at 6 min when 90% of the DCPIP was photoreduced.

If irradiation was extended beyond 6 min, additional H₂O₂ formed as a result of continued cycles of ¹O₂ generation and reoxidation of reduced DCPIP because O₂ was not yet depleted (Scheme 3 and Scheme 4). In this case, the concentration of H₂O₂ exceeded that of the reduced DCPIP, which functions as the HRP co-substrate. This was evident because upon addition of HRP, the blue color of oxidized DCPIP formed quickly but faded to pale pink within 1 min.

We suspected that the phenol ring of DCPIP shown in Scheme 2 could be further oxidized to yield a colorless product. To address this, we mixed DCPIP with H₂O₂ and HRP. As added H₂O₂ increased, absorbance at 605 nm decreased (Supplemental Figure S3) and only a weak absorbance ~500 nm (pink) remained. Therefore, both forms of DCPIP in Scheme 2 are co-substrates for HRP. Consequently, H₂O₂ concentrations in Figure 1B are only estimates, especially at 6 min.

The role of dissolved O₂ during DCPIP photoreduction is explored in Figure 1C. Increasing concentrations of histidine were included in DCPIP photoreduction reactions with MB and EDTA identical to those in Figure 1A,B. Histidine is a well-characterized ¹O₂ scavenger that reacts with it to yield an oxygenated species, oxo-histidine, thereby depleting O₂ in the solution [31,32,33]. Figure 1C shows that DCPIP photoreduction increased at 4 min for all concentrations of histidine tested. For 1.5 mM histidine, DCPIP photoreduction was also greater at 2 min than for the untreated or lower histidine concentrations. By scavenging ¹O₂, histidine limited both reoxidation of reduced DCPIP and H₂O₂ generation. Decreased H₂O₂ for histidine-treated samples was confirmed because when HRP was added to those samples after irradiation, the resultant increase in A₆₀₅ was always less than that of the untreated samples.

We tested azide as an ¹O₂ scavenger in DCPIP photoreduction assays but did not observe the same effect as for histidine. Although azide reacts avidly with ¹O₂, additional azide-derived radical intermediates form and O₂ is regenerated [34,35]. For these reasons, we did not pursue azide studies further.

Because catalase regenerates O₂ from H₂O₂ according to Equation (2), it favors the formation of more ¹O₂ that may subsequently reoxidize reduced DCPIP. However, because only 0.09 mM H₂O₂ was detected after 6 min (Figure 1B), ~0.20 mM dissolved O₂ remained. Because O₂ was not limited, catalase had no effect on DCPIP photoreduction under the assay conditions employed.

We also compared the time course of DCPIP photoreduction in different experimental formats. MB, EDTA and DCPIP were premixed and then aliquoted into a cuvette (1 mL) or plate wells (100 μL or 200 μL each). Figure 1D shows that the reaction format affected DCPIP photoreduction, with the fastest rate observed in the cuvette, followed by the 200 μL reactions, and the slowest with only 100 μL. The difference in rates was more pronounced at the longer reaction times when more DCPIP had been photoreduced and it could be reoxidized by ¹O₂. The dimensions of the plate well allow more O₂ to enter the solution as it is consumed. In addition to increased ¹O₂, newly reduced DCPIP is also exposed to O₂ at the well/air interface and may be directly oxidized independent of light-generated ¹O₂. HRP was added to select wells and A₆₀₅ values increased, indicative of H₂O₂ formation.

Although the data in Figure 1C support ¹O₂-mediated reoxidation of reduced DCPIP as the source of H₂O₂, an alternative path to generate H₂O₂ exists. The tertiary amines that we employ as electron donors also react with ¹O₂, though their rate constants are quite low (~10⁵ M⁻¹ s⁻¹) [36,37]. According to Equation (3), ¹O₂ oxidation of a tertiary amine also produces a superoxide anion that becomes H₂O₂.

R₃N + ¹O₂ → R₃N^+• + O₂^−•

(3)

A solution of 2 μM MB and 4 mM EDTA at pH 7.4 in a cuvette (no DCPIP) was irradiated for up to 6 min. Aliquots of this solution were combined with HRP and tetramethylbenzidine (TMB) as a co-substrate. At 6 min, only 0.04 mM H₂O₂ was produced from EDTA, which is less than the H₂O₂ concentration when DCPIP was included (0.09 mM). Greater H₂O₂ with DCPIP is supportive of redox cycling where photoreduced DCPIP reacts with ¹O₂. At 2 min, only 0.014 mM H₂O₂ was detected from EDTA. At this earlier time, prior to more complete DCPIP photoreduction, ¹O₂-mediated oxidation of EDTA may be the source of H₂O₂.

The intense color of DCPIP is attributed to the extended conjugation of its phenolate form (reported pK_a = 5.9). Absorbance scans of DCPIP in PB at pH 6.4 and 4.5 show the decrease in absorbance at 605 nm and the shift to a less intense peak at ~514 nm (Supplemental Figures S4–S8). Because the extended conjugation of the phenolate is likely to impede its reduction, we examined DCPIP photoreduction from pH 6.0 to 8.0. The pK_a values of the EDTA amines are also crucial; a tertiary amine must be deprotonated to act as an electron donor. The pK_a values of the EDTA amines are 6.16 and 10.24. In the pH range from 6.0 to 8.0, the EDTA pK_a of 6.16 is important since its protonation will also vary.

DCPIP (0.10 mM), 2 μM MB and 4 mM EDTA were combined in 10 mM phosphate buffers of pH 6.0, 6,4, 6.9, 7.4 and 8.0 (cuvette). Figure 2A shows a clear pH dependence, with photoreduction increasing as pH decreased. The increase in DCPIP photoreduction between pH 7.4 and 6.9 was especially striking. Because DCPIP photoreduction rates at pH 6.0, 6.4 and 6.9 were nearly identical in Figure 2A, lower concentrations of both MB and EDTA were assayed. With 1 μM MB and 2 mM EDTA, little to no photoreduction was observed at pH 8.0 after 7 min of irradiation. DCPIP photoreduction at the other pH values is shown in Figure 2B. Under these conditions, differences in photoreduction rates are discernible for all pH values.

DCPIP photoreduction was also performed in a 96-well plate using 2 μM MB and 4 mM EDTA at each pH. As in Figure 2A,B, photoreduction increased as pH decreased (Supplemental Figure S9). In the 96-well plate, pH 6.9 was noticeably slower than pH 6.0 or 6.4, though all were complete after 3 min. Consistent with Figure 1D, exposure to O₂ in the plate well slowed DCPIP photoreduction.

Turnover numbers (mol DCPIP reduced/mol MB) and irradiation times for each pH and set of MB/EDTA conditions are summarized in

Table 1. pH dependence of DCPIP photoreduction by MB and EDTA.

pH	2 μM MB and 4 mM EDTA		1 μM MB and 2 mM EDTA
	Mol DCPIP/ mol MB	Irradiation Time (min)	Mol DCPIP/ mol MB	Irradiation Time (min)
6.0	46 ± 2	2	91 ± 4	4
6.4	47 ± 3	2	90 ± 5	5
6.9	45 ± 2	2	78 ± 3	7
7.4	41 ± 2	5	26 ± 2	7
8.0	28 ± 3	5	6 ± 3	7

for the cuvette reactions. Data for 1 μM MB and 2 mM EDTA highlight the difference across this pH range, especially the three-fold increase in turnovers at pH 6.9 vs. 7.4.

Further data analysis for DCPIP photoreduction was performed to obtain rate constants and half-lives for the reaction conditions and experimental configurations presented in Figure 1B–D and Figure 2A. These values are summarized in

Table 2. Summary of DCPIP photoreduction rate constants and half-lives.

pH	Treatment	Rate Constant (min−1)	Half-Life (min)
7.4	none	0.36 1	2.0
	0.5 mM his	0.38	1.8
	1.0 mM his	0.72	1.0
	1.5 mM his	0.82	0.85
	Plate-200 μL	0.22	3.1
	Plate-100 μL	0.15	4.5
8.0	none	0.15	4.5
6.9	none	1.49	0.46
6.4	none	1.57	0.44
6.0	none	1.59	0.44

¹ Rate constants were calculated from DCPIP concentration changes over time in Figure 1B–D and Figure 2A. All DCPIP photoreduction reactions contained 2 μM MB and 4 mM EDTA.

. Clear differences in photoreduction rate constants are evident at the varying pH values tested, with the largest rate constant and consequently shortest half-life at pH 6.0.

Photoreduction of CoQ₀, a benzoquinone frequently used in our studies, was performed using the same buffers, MB and EDTA. CoQ₀ showed no pH dependence (Supplemental Figure S10); therefore, we attribute the pH dependence shown in Figure 2A,B to the properties of DCPIP, not the protonation states of MB or EDTA. Given the EDTA amine pK_a of 6.16, one might have expected a slower rate at lower pH instead of the equal rates for CoQ₀ from pH 6.0 to 8.0.

Attempts to directly photoreduce DCPIP with pheoA as the photosensitizer and TEOA as the electron donor in 20% DMF were unsuccessful. These reaction conditions have been used successfully to photoreduce benzoquinones and naphthoquinones [9,10]. However, because the pK_a of TEOA ~7.8, we reported a two-fold increase in quinone photoreduction at pH 8.0 relative to pH 7.4 [9]. When pheoA was assayed with EDTA at pH 6.4 (also with 20% DMF), no DCPIP photoreduction was detected.

Direct photoreduction of DCPIP was only one goal of this work. We were interested in DCPIP as an electron acceptor from a reduced quinone (Scheme 5). Although we routinely use absorbance scans prior to and after irradiation to study photoreduction, quantitation is challenging because many quinone substrates only absorb in the UV range. In some cases, even absorbances of their oxidized and reduced forms overlap in the UV [10]. Further, chlorophyll metabolites like pheoA and chlorin e6 absorb at the same wavelengths as many quinones of interest.

Previously, we used DTNB derivatization after irradiation to detect benzoquinones because this method produced intensely colored adducts at ~550 nm [9]. However, this method required a 30 min incubation, during which some of the newly photoreduced quinols were likely to air oxidize. In addition, naphthoquinones are generally too large to react completely with excess DTNB in 30 min.

We also reported that the tetrazolium salt MTT reacted with photoreduced benzoquinones and naphthoquinones. The reduced MTT product, a formazan, was quantitated based on its characteristic absorbance ~600 nm [38,39]. Two limitations of this method were the time required (up to 15 min) and the limited solubility of the formazan product.

To test DCPIP as an electron acceptor from photoreduced quinols, we chose several benzoquinones and optimized their direct photoreduction in a 1 mL cuvette using both the combination of MB and EDTA as well as pheoA with TEOA as the electron donor. In Figure 3A, spectra of oxidized (dark) and reduced CoQ₀ (0.40 mM) are shown, as well as that of MB at 665 nm. As irradiation time increased, absorbance at 405 nm decreased. A full spectrum from 250 to 750 nm in Supplemental Figure S11 shows a peak at 270 nm (4 min), indicative of full reduction. Chemical reduction of CoQ₀ by sodium ascorbate or NaBH₄ yielded identical spectral changes.

Figure 3B illustrates the challenge of studying CoQ₀ photoreduction by pheoA because both absorb ~400 nm. Notably, the control pheoA spectrum in Figure 3B contained all reaction components except CoQ₀. This is important because 20% DMF, which is necessary for optimal pheoA-mediated photoreduction, affects pheoA absorbance [9]. A second, smaller pheoA peak at 665 nm was monitored for photobleaching. Figure 3B shows the decrease in CoQ₀ absorbance as irradiation time increased. At 4 min, all CoQ₀ was photoreduced and a peak at 270 nm formed, also indicative of photoreduction.

Regardless of the photoreduction conditions (Figure 3A or B), photoreduced CoQ₀ was assayed for electron transfer to DCPIP (Scheme 5). Control DCPIP (80 μM) and DCPIP combined with CoQ₀ (60 μM) labeled “dark” have identical spectra in the visible range (Figure 3C). Below 280 nm, the dark sample absorbs more than the DCPIP alone. Supplemental Figure S12 shows the full spectrum down to 250 nm, where the dark sample is off scale but DCPIP alone is not.

When aliquots of reduced CoQ₀ were combined with a slight excess of DCPIP (1.33:1), absorbance at 605 nm decreased. CoQ₀ that had been irradiated for 2 min yielded less DCPIP reduction than the 4 min sample. The reaction between photoreduced CoQ₀ and DCPIP was rapid and DCPIP absorbance became stable within 1 min of mixing. Because excess DCPIP was present, we did not expect absorbance at 605 nm to decrease to zero. Photoreduced CoQ₀, prepared in a cuvette as shown in Figure 3A, was also stable. Up to three aliquots of it were mixed with DCPIP sequentially in a separate cuvette to obtain DCPIP absorbance changes. CoQ₀ that had been photoreduced using pheoA/TEOA and 20% DMF reacted with DCPIP in the same manner. A “dark” control was essential because DMF (only 3% final) affected DCPIP absorbance and TEOA affected the final pH.

We also developed a plate reader assay where the same DCPIP and CoQ₀ solutions were combined as for Figure 3C (200 μL total). DCPIP absorbance changes at 605 nm in both formats were used to calculate the photoreduced CoQ₀ concentration and it was compared to time courses based on direct photoreduction (Figure 3A). Direct photoreduction is indicated by “abs405” in Figure 3D and shows nearly complete photoreduction of 0.40 mM CoQ0 at 4 min. From the CoQ₀ concentration during electron transfer (60 μM), we used the dilution factor to estimate photoreduced quinone in the 1 mL samples. The DCPIP electron transfer assay in a cuvette “DCPIP-cuvette” showed full reduction of 0.40 mM CoQ0 at 4 min. This slight difference may be explained by the greater sensitivity of DCPIP absorbance (ε = 19,100 M⁻¹ cm⁻¹) vs. that of CoQ₀ (ε = 740 M⁻¹ cm⁻¹) at pH 7.4. The 2 min time samples for “abs405” and “DCPIP-cuvette” were identical.

For the plate assay “DCPIP-plate”, less reduced CoQ₀ was detected at both 2 and 4 min, even though the CoQ0 photoreduction reactions were identical and CoQ₀ and DCPIP were mixed in the same manner. A challenge of this format is that air can both oxidize newly photoreduced quinol and oxidize DCPIP, which served as the electron acceptor. Nonetheless, the rapid reaction between photoreduced CoQ₀ and DCPIP (1 min) is noteworthy.

Electron transfer from photoreduced CoQ₀ to DCPIP was also rapid in PB at both pH 6.4 and 4.5. At pH 4.5, the protonated DCPIP peak at 515 nm was monitored (Supplemental Figures S6–S8). CoQ₀ photoreduction reactions were also performed with catalase to ensure that no H₂O₂ was available to interfere with electron transfer from reduced CoQ₀ to DCPIP. No difference in reactivity was observed.

Direct photoreduction of methoxybenzoquinone (methoxyBQ) by both MB/EDTA (Figure 4A) and pheoA/TEOA (Figure 4B) was performed as described above for CoQ₀. Figure 4A shows distinct oxidized and reduced peaks for methoxyBQ and the characteristic MB peak at 665 nm. The decrease in absorbance for oxidized methoxyBQ at 366 nm coincided with an increase in absorbance at 286 nm for the reduced form. Comparison with CoQ₀ in Figure 3A,B supports equivalent photoreduction by equal concentrations of MB and EDTA in 4 min.

As for CoQ₀, photoreduction by pheoA with TEOA as the electron donor is complicated by pheoA absorbance, which overlaps with oxidized methoxyBQ (Figure 4B). Absorbance of pheoA alone also increased the baseline where reduced methoxyBQ absorbs at 286 nm. The peak at 286 nm did not reach the same absorbance (~1.5) as that of the methoxyBQ photoreduced by MB/EDTA in Figure 4A; therefore, we conclude that methoxyBQ was not fully reduced under the pheoA/TEOA conditions employed.

When photoreduced methoxyBQ (using MB/EDTA) was combined with DCPIP, absorbance at 605 nm decreased rapidly and leveled off within 1 min. Figure 4C summarizes the electron transfer results. Gratifyingly, absorbance at 366 nm “abs366” shows full reduction of 0.40 mM methoxyBQ at 4 min, as does “DCPIP-cuvette”. Plate values “DCPIP-plate” were ~10% lower at both 2 and 4 min compared to the other assay formats. It is noteworthy that the majority of the methoxyBQ photoreduction occurred in the first 2 min (~80%).

Photoreduction of methyl benzoquinone with MB and EDTA shows a decrease in the broad, weak absorbance ~320 nm and a sharp peak at 286 nm (Figure 5A). This example, more so than CoQ₀ or methoxyBQ, highlights the challenges of relying on absorbance alone to quantitate photoreduction even by MB/EDTA. The baseline at 286 nm where reduced methylBQ absorbs is not zero due to MB and pheoA absorbance (Figure 5A,B). Comparison with chemically reduced methylBQ supports full photoreduction of 0.40 mM methylBQ after 6 min with 2 μM MB and 2 mM EDTA.

When photoreduced methylBQ (60 μM) was combined with DCPIP (80 μM), methylBQ reacted more slowly with DCPIP relative to CoQ₀ and methoxyBQ, requiring up to 3 min before DCPIP absorbance remained constant. Further, the extent of electron transfer to DCPIP was not consistent with complete methylBQ reduction, like that for “abs286” (Figure 5C). For the 1.33:1 ratio of DCPIP to methylBQ, only ~50% of the photoreduced methylBQ was detected (4 min sample) relative to the value for the direct absorbance measurement at 286 nm. Because we suspected that the electron transfer reaction had reached equilibrium, we increased the ratio of DCPIP to methylBQ to 2:1. Because DCPIP absorbs so strongly, we could not increase its concentration but rather decreased that of methylBQ to 40 μM. Figure 5C shows that more reduced methylBQ was detected at this higher ratio (adjusting for a different dilution factor). Ratios of DCPIP to methylBQ higher than 2:1 were not attempted because we would have had to decrease the methylBQ concentration even more, resulting in increased opportunity for air oxidation.

The slower and incomplete reaction presented in Figure 5C suggests that reduced methylBQ is more stable than reduced CoQ₀ or methylBQ and less willing to donate electrons to DCPIP. For comparison, we mixed DCPIP with hydroquinone (HQ), the stable and commercially available reduced form of 1,4-benzoquinone. When HQ was mixed with DCPIP under identical conditions, it did not react with DCPIP, as evidenced by little to no decrease in DCPIP absorbance. This was attempted at pH 6.4 and 4.5 as well as with incubation times up to 10 min. At most, DCPIP absorbance decreased by 10%, with any change detected in the first 2–3 min.

The direction of the electron transfer was reversed by combining photoreduced DCPIP, generated as described in Figure 1 and Figure 2, with each of the quinols. No increase in DCPIP absorbance occurred, which would have indicated DCPIP oxidation.

Lastly, we mixed photoreduced 1,4-naphthoquinone (1,4-NQ) with DCPIP as described for benzoquinones. In previous work, we had optimized photoreduction conditions for several naphthoquinones including 1,4-NQ. The reaction of DCPIP with reduced 1,4-NQ was immediate, with the lowest DCPIP absorbance reached upon mixing. Because photoreduced 1,4-NQ reoxidizes more readily than benzoquinones, the calculated reduced 1,4-NQ concentration from the change in DCPIP absorbance was less than those obtained from direct absorbance changes.

4. Discussion

Our results confirm that DCPIP, a quinone imine, is a substrate for direct photoreduction using MB as the photosensitizer and EDTA as the tertiary amine electron donor (Figure 1A). The reaction conditions are nearly identical to those employed for other quinone substrates that we have studied [9,10]. DCPIP photoreduction was very pH-dependent across the range from 6.0 to 8.0, with increased rates as pH decreased (Figure 2). We attribute this to protonation of the phenol of DCPIP, which has a pK_a of 5.9 (Scheme 2). In its deprotonated form, the negative charge on the phenolate is delocalized throughout the entire structure; DCPIP reduction disrupts this extended conjugation. The pH dependence is not due to the protonation states of MB or EDTA because no pH dependence was observed when CoQ₀ was photoreduced using the same MB, EDTA and buffer solutions (Supplemental Figure S10).

DCPIP photoreduction turnover numbers (mol DCPIP reduced per mol MB) are lower than those we reported for other quinones. Table 1 shows that the highest turnover number for DCPIP approached 100 at pH 6.0. For benzoquinones herein (Figure 3, Figure 4 and Figure 5) and in prior work, turnover numbers were ~200 and the ratio of EDTA electron donor to MB was also lower at 1000:1 vs. 2000:1 for DCPIP.

Clearly, DCPIP is more difficult to photoreduce because attempts to photoreduce it with pheoA, our most extensively studied chlorophyll metabolite, were unsuccessful. Optimized conditions using pheoA and TEOA (ratio of TEOA:pheoA = 2000) as the electron donor in 20% DMF consistently results in ~50 turnovers for multiple benzoquinone and naphthoquinone substrates [9,10]. We suspect this is at least partially due to the pH dependence shown in Figure 2. The higher pH required for TEOA electron donation is at odds with the greatly enhanced DCPIP photoreduction at lower pH values [9]. EDTA does not serve as an electron donor with pheoA due to charge repulsion and/or differences in polarity [8].

Azobenzenes are the only other class of molecules besides quinones reported to be photoreduced by chlorophyll metabolites. Using tin-substituted pyropheophorbide A, electron donors and red light, Dutta et al. reported azobenzene photoreduction on a nucleic acid template [40]. In their work, only five turnovers were detected.

Multiple experiments support a role for ¹O₂ in our DCPIP photoreduction studies with MB and EDTA (Figure 1B). H₂O₂ was detected using HRP, an enzyme that uses H₂O₂ to oxidize a co-substrate. Our results confirm that photoreduced DCPIP (colorless) served as a co-substrate in the HRP reaction (Figure 1B) and that blue DCPIP (Scheme 2) was further oxidized by HRP when H₂O₂ exceeded the concentration of photoreduced DCPIP. Because HRP oxidizes a broad range of organic molecules, including phenol, this result was not unexpected. It was confirmed in an independent control experiment where stable, blue DCPIP was mixed with HRP and known amounts of H₂O₂ (Supplemental Figure S3).

While this second DCPIP oxidation by HRP complicated H₂O₂ quantitation in Figure 1B, it identified DCPIP as a potentially useful HRP co-substrate. With its molar absorptivity of 19,100 M⁻¹ cm⁻¹ at 605 nm and pH 7.4 (Supplemental Figures S1 and S2), DCPIP is nearly as sensitive as TMB, a commonly used HRP co-substrate that absorbs strongly at 652 nm or at 450 nm under acidic conditions [41].

Consistent with prior work using quinones, histidine increased DCPIP photoreduction and decreased H₂O₂ formation by scavenging ¹O₂ (Figure 1C). Histidine competes with photoreduced DCPIP for ¹O₂; this limits reoxidation of DCPIP and, consequently, the photochemical redox cycling that is depicted in Scheme 3 and Scheme 4. Further, histidine oxidation to oxo-histidine by ¹O₂ removes O₂ from the solution so that it is not available to react further. This differs from ¹O₂ scavenging by azide, where O₂ is regenerated [34,35].

Published rate constants for the reaction of quinols with ¹O₂ are ~10⁹ M⁻¹ s⁻¹, which are several orders of magnitude greater than for the reaction of ¹O₂ with tertiary amines (~10⁵ M⁻¹ s⁻¹) [37,42]. Once reduced DCPIP or a quinol are available, they are the primary targets for ¹O₂ and the likely source of H₂O₂. When photoreduction is first initiated and quinols are absent, any ¹O₂ generated would react with tertiary amine electron donors, according to Equation (3).

Inspired by the Hill reaction, where DCPIP acts as an artificial electron donor in studies of intact chloroplasts, we explored electron transfer from photoreduced quinones to DCPIP. This proved to be a rapid and sensitive method to detect quinols. The DCPIP absorbance change at 605 nm is far removed from those of quinone substrates and chlorophyll metabolites at 300–400 nm. Figure 3B, Figure 4B and Figure 5B highlight the challenges of using pheoA for photoreduction of quinones because their absorbances overlap. Chlorin e6, another chlorophyll metabolite of interest, has an even more intense absorbance at ~400 nm than pheoA [43]. Tertiary amines are not physiologically relevant electron donors; if NADH or a nicotinamide derivative were a physiological electron donor, it would be difficult to detect photoreduction because they absorb at similar wavelengths [44].

Electron transfer between photoreduced CoQ₀ or methoxyBQ and DCPIP was successful and resulted in detection of 95–100% of the photoreduced quinones (Figure 3D and Figure 4C). The cuvette format for mixing DCPIP and quinol was more accurate than the 96-well plate, even though the same solutions were mixed. Both photoreduced quinol and reduced DCPIP are susceptible to air oxidation, a process that is more likely in a plate well. Further, the reaction of DCPIP and quinol worked well regardless of the photosensitizer and electron donor combination employed for reduction. This is important because this assay is needed for photoreduction processes mediated by chlorophyll metabolites, not MB.

Incomplete electron transfer from methylBQ to DCPIP (Figure 5C) was initially confusing; however, a review of reduction potentials for substituted benzoquinones provided some clarity. Unsubstituted 1,4-benzoquinone is most easily reduced based on its reduction potential of 0.10 V at pH 7 [26]. MethylBQ has a reduction potential of 0.023 V but those of ubiquinone and more substituted benzoquinones are negative [26]. Reduction potentials of naphthoquinones are generally more negative than benzoquinones [45]. Based on these values, the more difficult it is to reduce a quinone (more negative E°_red), the more readily it reacts with DCPIP.

Because methylBQ, with its slightly positive reduction potential, is easier to reduce relative to CoQ₀ and methoxyBQ, once reduced, it is less likely to react with DCPIP (Figure 5C). By increasing the ratio of DCPIP to reduced methylBQ, more electron transfer occurred but it was still incomplete (Figure 5C). Additionally, this explains why HQ did not react with DCPIP regardless of pH and after incubation for up to 10 min. Conversely, while the DCPIP reactions of reduced CoQ₀ and methoxyBQ were complete within 1 min, that of reduced 1,4-naphthoquinone occurred within the mixing time (5−10 s). Thus, the rate and extent of electron transfer to DCPIP provides a method to rank quinone reduction potentials in addition to its goal of reduced quinone quantitation [45].

The results herein expand our understanding of the photoreduction capabilities of MB and pheoA and the tertiary amine electron donors that are required. DCPIP is the first substrate we have identified that can be photoreduced by MB, but not by pheoA. Successful photoreduction by MB and EDTA confirms a common mechanism of concurrent photoreduction and photo-oxidation of newly reduced quinol by ¹O₂ that generates H₂O₂.

Electron transfer from photoreduced quinols to DCPIP is a rapid and sensitive method to investigate quinone photoreduction by chlorophyll metabolites. In particular, this will be effective as we seek to identify a physiologically relevant electron donor and to test substrates other than quinones. In vivo, we do not expect photocatalytic turnovers in the 50–200 range as we have reported for pheoA and MB. Rather, even one or two cycles of photoreduction mediated by dietary chlorophyll metabolites may affect health outcomes by recycling catechol- or quinone-based antioxidants.

5. Conclusions

Photoreduction of DCPIP and its subsequent reoxidation by ¹O₂ was studied using the photosensitizer methylene blue and EDTA as an artificial electron donor. The involvement of ¹O₂ was confirmed using histidine, a well-characterized ¹O₂ scavenger that limited the reoxidation step, thereby increasing photoreduction rates. DCPIP redox cycling is consistent with our prior photoreduction work with biologically relevant benzoquinones and o- and p-naphthoquinones. Rates of DCPIP photoreduction were sensitive to oxygen exposure and pH-dependent, with 3–4-fold faster rates at pH 6.9 vs. 7.4.

DCPIP served as an effective electron acceptor from several photoreduced benzoquinones. For photoreduced CoQ₀ and methoxy-benzoquinone, electron transfer to DCPIP was rapid and stoichiometric, whereas for reduced methyl benzoquinone and hydroquinone, it was incomplete, likely due to differences in reduction potential and equilibrium position. Electron transfer from photoreduced quinols to DCPIP is a rapid and sensitive method to investigate quinone photoreduction by chlorophyll metabolites.