Measurement of H2O2 within Living Drosophila during Aging Using a Ratiometric Mass Spectrometry Probe Targeted to the Mitochondrial Matrix

Summary Hydrogen peroxide (H2O2) is central to mitochondrial oxidative damage and redox signaling, but its roles are poorly understood due to the difficulty of measuring mitochondrial H2O2 in vivo. Here we report a ratiometric mass spectrometry probe approach to assess mitochondrial matrix H2O2 levels in vivo. The probe, MitoB, comprises a triphenylphosphonium (TPP) cation driving its accumulation within mitochondria, conjugated to an arylboronic acid that reacts with H2O2 to form a phenol, MitoP. Quantifying the MitoP/MitoB ratio by liquid chromatography-tandem mass spectrometry enabled measurement of a weighted average of mitochondrial H2O2 that predominantly reports on thoracic muscle mitochondria within living flies. There was an increase in mitochondrial H2O2 with age in flies, which was not coordinately altered by interventions that modulated life span. Our findings provide approaches to investigate mitochondrial ROS in vivo and suggest that while an increase in overall mitochondrial H2O2 correlates with aging, it may not be causative.


Figure S4. Uptake of MitoP by Energized Mitochondria
An electrode selective for the TPP cation was inserted in a stirred 3 ml chamber containing KCl medium (pH 7.2) supplemented with rotenone (4 µg/ml) and nigericin (100 nM) at 30ºC and calibrated by five sequential additions of 1 µM MitoP (arrowheads). Rat liver mitochondria (2 mg protein/ml) were then added, followed by succinate (10 mM) and the uncoupler FCCP (500 nM).

Figure S5. Uptake and Metabolism of MitoB by Mitochondria Within Flies
(A) Distribution of MitoB within flies. Wild-type females (7 d) were injected with MitoB and incubated for 3 h. Following snap freezing, the flies (cohorts of 20) were dissected into body parts, and the amount of MitoB in the homogenates from head (H), thorax (T) and abdomen (A) was measured by LC/MS/MS. Data are means ± SEM of three determinations. The % of total MitoB in each body part relative to the sum in all is indicated (mean ± SD, n = 3). (B) Concentration of MitoB within females flies. Data from (A) were divided by the wet weight of the individual body parts. The average MitoB concentration was determined by dividing the total amount of MitoB in the three body parts by the sum of the wet weight of the three body parts. Data are means ± SEM of three determinations. (C) Distribution of mitochondria within wild-type female flies (7 d). Total citrate synthase activity was measured in homogenates prepared from whole flies, or from flies dissected into body parts (cohorts of 10). Data are means ± SEM of three separate measurements. The % of total citrate synthase activity in each body part expressed relative to the sum in all is indicated (mean ± SD, n = 3). The specific activities for citrate synthase in the whole body, head, thorax and abdomen respectively were: 39 ± 7; 38 ± 10; 84 ± 14; 7 ± 1 nmol citrate/min/mg protein (means ± SEM, n = 3). (D) Western blots showing mitochondrial labelling with iodoacetamide-TPP (IAM-TPP). Left panel -control mitochondria (50 µg protein) were energized with glycerol-3phosphate (10 mM) and incubated at 25ºC for 15 min with IAM-TPP (5 µM). This resulted in TPP-labelling of mitochondrial proteins on cysteine residues that was visualized using antiserum against TPP. Blocking cysteine thiols with Nethylmaleimide (NEM, 100 µM) prevented IAM-TPP binding. Uncoupling with FCCP (1 µM) prevented the ∆ψ m -dependent accumulation of IAM-TPP inside mitochondria and thereby decreased labelling of proteins. Right panel -live flies (~150 7 d females) were injected with 75 pmol IAM-TPP per fly in the same way as was done for MitoB. After incubating the injected flies for 3 h, mitochondria were isolated and 50 µg mitochondrial protein was assessed for TPP labelling. The banding pattern obtained for TPP labelling of the mitochondria from flies that had been injected with IAM-TPP was very similar to that of control mitochondria that were incubated with IAM-TPP in vitro. These findings confirm the mitochondrial localization of IAM-TPP in vivo. IAM-TPP labelling was not detectable in cytosolic fractions from injected flies (data not shown). (E) Metabolites of MitoB Within Flies Assessed by Tandem Mass Spectrometry. Ten cohorts of 10 flies were injected with MitoB, incubated for 6 h and then homogenized and extracted without internal standards. The extracts were dried, then resuspended and combined in a total volume of 200 µl 40% acetonitrile/0.1% formic acid. Samples of the MitoB solution that was injected into the flies (traces a), extracts from the MitoB injected flies (traces b) and extracts of a control cohort of uninjected flies (traces c) were analyzed for molecules that contained the TPP moiety. To do this the samples were directly infused into the mass spectrometer at 5 µl/min for 1 min, and the m/z of parent ions that fragmented to give a daughter ion of m/z = 183.0 (left), or of m/z = 261.1 (right), both diagnostic of the TPP moiety, were identified. For traces A and B each trace is normalized to the highest total ion count peak within that trace, hence baseline noise is magnified in the traces for the daughter ion of m/z = 261.1 relative to that for the daughter ions of m/z = 183.0 as the total ion count is ~10-fold higher in the latter. The background traces (c) for the extracts from uninjected flies are normalized to the highest total ion count peak within the corresponding trace (b) to equalize baseline noise and to facilitate comparison. The prominent parent ion of MitoB (m/z 397.1) and a few contaminants present in the injectate were all detectable in the injected flies (dashed lines). The only new peak present in the injected flies that was not present in the injectate was at m/z = 369.1 corresponding to MitoP. Therefore within flies MitoB is only metabolized to MitoP.

Figure S6. Using MitoB to Assess Mitochondrial ROS Production in C. elegans and Mice
(A) Effect of paraquat on the MitoP/MitoB ratio within worms. Wild-type C. elegans (N2 strain) were incubated for 1 h with 10 µM MitoB and 50 U catalase ± 50 mM paraquat (PQ), then the MitoP/MitoB ratio in the worm pellet was quantified by LC/MS/MS. Data are means ± SEM of four samples, and were corrected for the ratio at t = 0. There was also a statistically significant increase in the MitoP/MitoB ratio in the worm incubation medium following treatment with 100 mM paraquat (data not shown). Statistical significance was determined by a two-tailed Student's t-test: * p < 0.05. (B,C) To see if the MitoP/MitoB ratio could be measured within mammalian tissues, mice were infused intravenously with 180 nmol of MitoB over 6 h, or infused with 180 nmol of MitoB over 2 h followed by 4 h saline infusion with no MitoB. The mice were then sacrificed and the MitoB and MitoP content of tissue samples was assessed by extraction followed by LC/MS/MS analysis relative to deuterated standards. Data are means ± SEM of three-four mice for each condition. After 6 h MitoB infusion, the amount of MitoB in the blood was 22 ± 7 pmol MitoB/50 mg blood (n = 3, mean ± SEM), far lower that the amount present within tissues. By 4 h after the 2 h MitoB infusion, there was no MitoB detectable in the blood, despite large amounts in the tissues. These findings are consistent with the rapid uptake of MitoB from the circulation into tissues, as has been found for other TPP cations (Porteous, 2010;Smith et al., 2003). (D) Mice were administered with MitoB (180 nmol of MitoB over 6 h) and the levels of MitoB and MitoP in the urine were measured. Data are means ± range for two mice. (E) Ratio of MitoP/MitoB in mouse tissues. Mice were administered MitoB and the ratios of MitoP/MitoB in the tissues were determined after MitoB infusion for 6 h or after MitoB infusion for 2 h followed by 4 h saline infusion with no MitoB. Data are means ± SEM of three-four mice for each condition. The MitoP/MitoB ratio in the tissues showed that the 4 h period post-infusion resulted in a 1.6 to 2.5-fold increase in the MitoP/MitoB ratio relative to continuous infusion, consistent with conversion of MitoB to MitoP by mitochondrial ROS in vivo.

Calculation of the Mitochondrial Content of MitoB In Vivo Within Living Flies
To estimate the amount of MitoB present within flies that was inside mitochondria, we first estimated the proportion of the MitoB that was present within cells inside the flies (α). To do this, we consider the known content and distribution of water within flies: female adult flies, which weigh approximately 1.5 mg, contain about 66% by weight total body water of which about 6-10% is hemolymph with the remaining 90-94% being cellular water (Folk et al., 2001). Thus, we take the proportion of MitoB within the fly that is intracellular, α, as 0.9. This is likely to be an underestimate as MitoB in tissues is concentrated into the cells driven by the plasma membrane potential (∆ψ p ) with subsequent uptake into the mitochondria driven by the mitochondrial membrane potential (∆ψ m ). This is consistent with the findings in mice administered MitoB (this paper) and other TPP compounds (Porteous et al., 2010) in which all TPP compounds were rapidly removed from the circulation and concentrated within the tissues.
The next factor to be determined is the proportion of intracellular MitoB that is present within the mitochondria (β). This value depends on the concentration of MitoB in the mitochondrial and cytoplasmic compartments ([MitoB] mito and [MitoB] cyto , respectively), and the volumes of those compartments (Vol mito and Vol cyto , respectively), and is given by: (1) β =

Volmito[MitoB]mito Volcyto[MitoB]cyto + Volmito[MitoB]mito
This can be arranged to: (2) β = The ratio of MitoB concentrations in the mitochondrial and cytosolic compartments can be calculated from the mitochondrial membrane potential (∆ψ m ) using the Nernst equation, allowing for binding corrections for MitoB in the mitochondrial (a mito ) and cytosolic (a cyto ) compartments (Brand, 1995). The distribution of MitoB between the cytosol and mitochondria is then determined by the Δψ m which in mV at 25ºC is: Therefore the proportion of intracellular MitoB that is within the mitochondria, β, is given by: The volume fraction of the Drosophila flight muscle that is mitochondrial has been determined by stereology of electron micrographs as 35-40% (Magwere et al., 2006). The average mitochondrial density within whole male flies has been estimated as ~30% from measuring the ratio of the specific activities of four mitochondrial enzymes in isolated mitochondria to their activities in whole fly homogenates (Magwere et al., 2006).
Therefore it is reasonable to take the average mitochondrial volume within cells inside flies as 30% so Vol mito /Vol cyto = 0.428. The mitochondrial binding correction factor for MitoB (a mito ) can be assumed to be the same as that of the closely related molecule methylTPP calculated for fly mitochondria: a mito = 0.24 (Brand et al., 2005). The cytosolic binding correction factor for MitoB (a cyto ) can be assumed to be the same as for methylTPP in the cytosol of hepatocytes: a cyto = 0.21 (Brand, 1995). Therefore a cyto /a mito = 0.875. Entering these values into equation (5) gives: If we take a reasonable value for the average ∆ψ m within flies of 140 mV, then β = 0.98. If we vary the ∆ψ m to 120 or 160 mV, then β ranges between 0.97 and 0.99 respectively. Therefore we take the fraction of the intracellular MitoB that is mitochondrial (β) as 0.98. Multiplying β by the fraction of MitoB within flies that is present within cells, α, which is 0.9 giving the fraction of total MitoB in the fly that is in the mitochondria: α x β = 0.98 x 0.9 = 0.88.

Calculation of Average Mitochondrial [H 2 O 2 ] In Vivo Within Living Flies
From equations 7 and 9 we can derive the relevant rate equations for the change in [MitoP] total : Solving equation 10 for [MitoP] total gives: As the rates of elimination of MitoB and MitoP from the whole fly are the same ( Figure 5A) k 1 = k 3 . Consequently the equation then simplifies to: Rearranging gives: Mol(MitoB) mito /Mol(MitoB) total is the same as the fraction of total MitoB in the fly that is mitochondrial (= α x β) which we earlier estimated to be 0.88. Vol total /Vol mito is the reciprocal of the product of the proportion of the total fly body water that is cellular (= 0.90 (Folk et al., 2001)) by the proportion of the cell that is mitochondrial, which is 0.30 (Magwere et al., 2006). Together these give γ = 3.26.

Synthesis of MitoB and MitoP
All reactions under an inert atmosphere were carried out using oven-dried or flamedried glassware. Solutions were added by syringe. Acetonitrile, tetrahydrofuran and toluene were dried where necessary using a solvent drying system, Puresolv TM .
Reagents were obtained from commercial suppliers and used without further purification unless otherwise stated. 1 H, 31 P and 13 C NMR spectra were obtained on a Bruker DPX/400 spectrometer operating at 400, 162 and 100 MHz respectively. All coupling constants are measured in Hz. 31 P and 13 C NMR spectra are protondecoupled. DEPT (distortionless enhancement by polarization transfer) was used to assign the signals in the 13 C NMR spectra as C, CH, CH 2 or CH 3 . Signals for deuterated carbons of perdeuterated aromatic rings appear as narrow multiplets and are given the assignment CD. ESI-MS were carried out on a Thermofisher LTQ Orbitrap XL at the University of Swansea, other mass spectra (MS) were recorded on a Jeol JMS700 (Mstation) spectrometer. Infrared (IR) spectra of compounds in either solid or liquid form were obtained on a Shimadzu FTIR-8400S using attenuated total reflectance (ATR). Microanalyses were obtained on an Exeter Analytical CE440. This instrument cannot distinguish between hydrogen and deuterium and the percentages of D and H found in deuterated compounds are calculated from raw data with any deviation from theoretical appearing in the percentage D.  (Howard et al., 2001), but no spectral data were provided.

Characterization of MitoB and MitoP
The MitoB pinacol ester and d 15 -MitoB pinacol ester were synthesized from 3bromotoluene 1 ( Figure S1A). The triphenylphosphine group is attached to a benzylic position in MitoB to allow easy synthesis and is meta to the boronate group, rather than ortho or para, so that fragmentation does not occur upon oxidation.
Since Chang used the pinacol esters of boronic acids for the original H 2 O 2selective fluorescent probes (Dickinson and Chang, 2008;Miller et al., 2007), we first assessed the stability of MitoB pinacol ester by mass spectrometry. We found that the pinacol ester (m/z = 479.23) was very rapidly hydrolyzed to the arylboronic acid, MitoB itself (m/z = 397.15) in the 0.1% formic acid used for the LC/MS analysis, consistent with previous reports (Xu et al., 2006). As pinacol esters are also hydrolyzed by the silica matrix of the column during RP-HPLC (Xu et al., 2006) (Dickinson and Chang, 2008;Miller et al., 2007), Peroxynitrite (ONOO -) and decomposed ONOOwere made as described previously (Packer and Murphy, 1994). To assess the effects of ONOOon MitoB, 100-250 µM MitoB in KCl medium, pH 8.0 at 37ºC was incubated with ONOO -(100-250 µM) and compared with the effects of decomposed ONOOsolutions. The reaction products were analyzed by UV/Vis spectrophotometry and by RP-HPLC.
MitoP did not react with 1 mM H 2 O 2 (data not shown).
Linoleic acid peroxide was made as described (Ohkawa et al., 1978). For this linoleic acid (Sigma, 0.1 g) was dissolved in 10 ml ethanol and mixed with 100 ml 50 mM sodium borate buffer (pH 9.0) to which was added 2 mg soybean lipoxygenase (BioChemika/Sigma; 7.9 U/mg). The mixture was incubated at 23ºC until the absorbance at 233 nm due to the linoleic acid diene peroxide had reached a stable maximum (~70 min). At this point the mixture was adjusted to pH ~3.0 with 1 M HCl, and the linoleic peroxide was extracted into diethyl ether, dried over sodium sulfate, and the solvent was evaporated under vacuum. The residue was dissolved in ethanol, the concentration of the linoleic acid diene peroxide was determined at 233 nm (ε = 27.4 x 10 3 M -1 cm -1 ; Tappel et al., 1952) and a stock solution of ~7 mM was prepared in ethanol.

Mitochondrial Preparation and Experiments
Rat liver and heart mitochondria were prepared by homogenization and differential centrifugation in ice-cold 250 mM sucrose, 10 mM Tris, 1 mM EGTA, pH 7.4, with addition of 0.1% w/v BSA for heart mitochondria. The protein concentration was determined by the biuret assay. Mitochondria were routinely incubated in KCl medium (120 mM KCl, 10 mM HEPES, 1 mM EGTA, pH 7.2) supplemented with 10 mM succinate and 4 µg/ml rotenone. An electrode selective for the TPP moiety of MitoB and MitoP was prepared and used as described (Asin-Cayuela et al., 2004).

Cell Experiments
The toxicity of MitoB and MitoP to C2C12 cells was assessed by incubating cells in 96 well plates with MitoB or MitoP from 0 to 100 µM for 40 h, then the medium was replaced with fresh medium containing 0.05% 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide (MTT) for 2 h at 37ºC. This was replaced with 20% SDS in 50% dimethylformamide and the absorbance at 570 nm was measured relative to the background absorbance at 650 nm. This MTT assay showed that over 40 h there was negligible toxicity of MitoB or MitoP up to 10 µM, with mild toxicity apparent at 25 µM and above (data not shown).
For MitoB uptake studies, Jurkat cells were incubated at 3 x 10 6 cells/ml in medium supplemented with 5 µM MitoB (Ross et al., 2008). Samples (1 ml) were taken at various time points and cells were pelleted by centrifugation (16,000 x g for 2 min). To measure the effect of uncoupling by FCCP on MitoB uptake, Jurkat cells (3 x 10 6 /ml) were incubated in PBS supplemented with 1 mM pyruvate ± 500 nM FCCP for 1 h with 5 µM MitoB and then 950 µl of the cell suspension was pelleted by centrifugation (16,000 x g for 2 min), the supernatant was discarded, and the pellets snap frozen for LC/MS/MS analysis.
To measure the uptake of [ 3 H]-methylTPP by Jurkat cells, cells were incubated in PBS with 1 mM pyruvate and 500 nM [ 3 H]-methylTPP (100 nCi/ml, American Radiolabelled Chemicals Inc.) at 37ºC. After 1, 2 and 3 h, a 0.5 ml aliquot of the cell suspension was sampled in triplicate and pelleted by centrifugation (16,000 x g for 2 min). The supernatant was removed, and the amount of [ 3 H]-TPMP in the pellet was determined by scintillation counting (Ross et al., 2008). The presence of FCCP decreased the amount of pellet uptake by 47, 53 and 55% at 1, 2 and 3 h respectively.
To extract MitoB and MitoP from adherent C2C12 cells in a 6-well culture dish, the cell medium was removed and the cell layer was incubated with 750 µl 95% ACN/0.1% FA with rocking for 5 min, then scraped and transferred to an eppendorf tube, and the well washed with 2 x 375 µl 95% ACN/0.1% FA. The combined extracts were spiked with IS (62.5 pmol each of d 15 -MitoB and d 15 -MitoP), vortexed, incubated at -20ºC for 30 min and then centrifuged (10 min at 16,000 x g).

Fly Culture and Experiments
The To determine the proportion of mitochondria in each fly body segment, the activity of citrate synthase (CS), a mitochondrial matrix enzyme, was measured (Magwere et al., 2006). Homogenates were prepared from whole flies, heads, thoraces and abdomens, using cohorts of 10 flies. Each sample was homogenized on ice in 300 µl of CS assay buffer (100 mM Tris-HCl, 0.1% Triton X-100, pH 8.0), using three 2-s bursts of a T8 Ultra-Turrax homogenizer (Jencons-PLS), at 1 min intervals.
Homogenates were then centrifuged twice at 2000 x g for 15 s, the volume of supernatant measured, and the CS activity assayed. To measure CS enzyme activity, samples were added to assay buffer at 25ºC containing 100 mM Tris-HCl, 0.1% Triton X-100, 0.1 mM DTNB and 0.4 mM acetyl-CoA. Change in absorbance at 412 nm was measured for 2 min to determine background activity due to acetyl-CoA hydrolysis, then 0.4 mM oxaloactetate was added to initiate reaction, and absorbance at 412 nm followed for a further 2 min. An extinction coefficient of 13,600 M -1 cm -1 was used to calculate CS enzyme activity. The protein concentration of each extract was determined in parallel using the bicinchoninic acid assay, and used to calculate the CS specific activity.
Fly mitochondria were isolated (from ~150 7 d females) and incubated as described (Miwa et al., 2003). Mitochondria (50 µg protein) were separated on a 12.5% SDS-PAGE gel and a Western blot was performed using antiserum raised against the TPP moiety (Porteous et al., 2010).

Worm Culture and Experiments
Wild-type C. elegans (N2 Bristol strain) was grown and manipulated as published (Sulston and Hodgkin, 1988 Blood and urine samples were collected by direct cardiac and bladder puncture; heart, liver and kidneys were removed and all were snap frozen on dry ice and stored at -80ºC. In both cases the cumulative dose of MitoB was 180 nmol/mouse, ~7 µmol/kg.

Extraction of MitoB and MitoP from Worms and Mice
To

LC/MS/MS Analysis of MitoB and MitoP
The LC/MS/MS system consisted of a Waters Quattro Ultima triple quadrupole mass spectrometer attached to a binary pump (model 1585; Jasco) and an HTC-PAL autosampler (CTC-Analytics). Samples and standards in autosampler vials were placed in a refrigerated holder (4ºC) while awaiting injection by the autosampler.
Liquid chromatography was performed at 30ºC using a Luna 5 µ Phenyl-Hexyl column (1 x 50 mm, 5 µm) with a Phenyl-Hexyl guard column (2 x 4 mm) (both from Phenomenex). The mobile phase consisted of 0.1% FA in water (buffer A) and 95% ACN/0.1% FA (buffer B) delivered as a linear gradient as follows: 0-2 min, 5% B; 2-3 min, 5-25% B; 3-5 min, 25-75% B; 5-7 min, 75-100% B; 7-10 min 100% B; 10-12 min, 100-5% B; 12-20 min, 5% B. The flow rate was 50 µl/min and a 30 µl volume was injected into a 20 µl sample loop. An in-line divert valve was used to divert eluant away from the mass spectrometer from 0-5 min and 16-20 min of the acquisition time. For mass spectrometry, electrospray ionization in positive ion mode was employed. The instrument parameters were: source spray voltage, 3 kV; cone voltage 100 V; ion source temperature, 80ºC; collision energy, 50 V. Nitrogen was used as the curtain gas and argon as the collision gas. Multiple reaction monitoring (MRM) in positive ion mode was used to detect the compounds. To determine the best fragmentations to use for quantification, direct infusion into the mass spectrometer at 2 µl/min of 1 µM d 15 -MitoP, MitoP, d 15 -MitoB pinacol ester or MitoB pinacol ester was used. Published fragmentation patterns for alkylTPP ions (Claereboudt et al., 1993;Denekamp et al., 1999;Denekamp et al., 2003) and the presence of the deuterium atoms solely on the TPP moiety were used to determine the fragmentation ( Figure S2).  Figure 3A). Data were acquired and analyzed with MassLynx software.
To determine whether there were any products of MitoB other than MitoP, solutions of fly extract were directly infused into the mass spectrometer at 5 µl/min.
The daughter ion m/z 183 was used to identify parent ions that fragmented to give this product, as this daughter contains the TPP ion and should therefore be present in any MitoB product. Background was determined using solutions extracted from flies with no exposure to MitoB, and MitoB products were determined using solutions extracted from flies that had been incubated with MitoB for 3 h following injection. Data were accumulated over a period of 1 min, and were normalized to total ion count.