Far-red and sensitive sensor for monitoring real time H2O2 dynamics with subcellular resolution and in multi-parametric imaging applications

H2O2 is a key oxidant in mammalian biology and a pleiotropic signaling molecule at the physiological level, and its excessive accumulation in conjunction with decreased cellular reduction capacity is often found to be a common pathological marker. Here, we present a red fluorescent Genetically Encoded H2O2 Indicator (GEHI) allowing versatile optogenetic dissection of redox biology. Our new GEHI, oROS-HT, is a chemigenetic sensor utilizing a HaloTag and Janelia Fluor (JF) rhodamine dye as fluorescent reporters. We developed oROS-HT through a structure-guided approach aided by classic protein structures and recent protein structure prediction tools. Optimized with JF635, oROS-HT is a sensor with 635 nm excitation and 650 nm emission peaks, allowing it to retain its brightness while monitoring intracellular H2O2 dynamics. Furthermore, it enables multi-color imaging in combination with blue-green fluorescent sensors for orthogonal analytes and low auto-fluorescence interference in biological tissues. Other advantages of oROS-HT over alternative GEHIs are its fast kinetics, oxygen-independent maturation, low pH sensitivity, lack of photo-artifact, and lack of intracellular aggregation. Here, we demonstrated efficient subcellular targeting and how oROS-HT can map inter and intracellular H2O2 diffusion at subcellular resolution. Lastly, we used oROS-HT with other green fluorescence reporters to investigate the transient effect of the anti-inflammatory agent auranofin on cellular redox physiology and calcium levels via multi-parametric, dual-color imaging.


Introduction
Oxidative stress is often a key component of many disease progressions.Tremendous efforts have been made to develop therapeutic approaches to target the excessive presence of oxidants and their source.
However, the unsatisfying results of antioxidative therapy call for a more nuanced understanding of cellular oxidants, antioxidative defense networks, and their effect on the cellular system with precision and speci city to improve rationales on antioxidative therapeutics 1 .
H 2 O 2 is a major oxidant in redox biology that can also act as a pleiotropic secondary messenger in various cellular signaling processes [2][3][4][5][6] .Its precursor superoxide is a natural byproduct of aerobic metabolism, which rapidly gets converted to H 2 O 2 naturally or by superoxide dismutase (SOD) 7 .The level of intracellular H 2 O 2 is tightly regulated by peroxide-reducing mechanisms 8,9 .Although peroxide is considered less reactive than other cellular oxidative agents, its excessive accumulation is often observed in pathology, with growing evidence of its causal role in the progression of diseases [10][11][12] .The engineering of genetically encoded H 2 O 2 indicators (GEHI, e.g.OxyR-based sensors [13][14][15] , peroxidasebased sensors [16][17][18] ) has been a signi cant step towards understanding the role of peroxide in redox biology by enabling real-time monitoring of peroxide dynamics in a wide array of biological hosts 19 .One advantage of GEHIs over redox-sensitive uorescence dyes is their spatiotemporal exibility: they can be targeted to speci c cell types or various cellular compartments for extended periods when coupled with proper expression systems (e.g.promoters and tra cking/export tags).Speci cally, red-uorescent GEHIs facilitate multiparametric analysis of peroxide dynamics along with other key biomolecules or processes considering a large number of green uorescent sensors for biological molecules and processes (e.g.Ca 2+ , pH, voltage, redox potential, etc.) 20,21 .Nevertheless, current red-shifted GEHIs exhibit slow kinetics, a bottleneck for real-time peroxide imaging.Most importantly, blue-light-induced photochromic artifacts commonly associated with red FP based sensors makes unobstructed multiparametric analysis alongside green uorescent sensors di cult 22 .Lastly, aggregation tendency and low brightness are also observed among red uorescent proteins 23 thus affect the utility of existing red GEHIs.
In this study, we coupled the bacterial OxyR peroxide sensor with a rhodamine-HaloTag-based chemigenetic reporter system to create a rst-in-class, far-red indicator for H 2 O 2 : oROS-HT 635 (optogenetic hydRogen perOxide Sensor with HaloTag with JF635).We developed a rational engineering strategy based on structural information derived from experimentally resolved structures and computational methods (ColabFold) 24 .oROS-HT 635 has excitation and emission wavelengths of 640 nm and 650 nm.We validated it in various biological host systems, including stem cell-derived cardiomyocytes in vitro and primary neurons ex vivo.Moreover, we found that the fast oROS-HT 635 kinetics allows the observation of intracellular diffusion of peroxide.Also, oROS-HT 635 is free from photochromic artifacts, allowing multiparametric analysis of contextual peroxide dynamics.As a proofof-concept, we showed the acute effect of the anti-in ammatory agent aurano n on peroxide with the context of changes in cellular redox potential in HEK293 cells and Ca 2+ in human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs), demonstrating intriguing multi-parametric effects of acute antioxidant system perturbation.

Results
Structure-guided engineering of oROS-HT 635 : a bright far-red optogenetic sensor for H 2 O 2 .
OxyR is a bacterial transcription activator with high speci city and sensitivity toward H 2 O 2 with low peroxidative capability (i.e. the protein exhibits high sensitivity towards peroxide with limited catalytic activity) 25 .Existing red-shifted GEHI, such as HyPerRed 14 and SHIRMP 15 utilize ecOxyR-LBD (regulatory domain of OxyR from Escherichia coli), as their sensing domain.However, both red GEHIs show slower kinetics (10s to 100s seconds for full activation under saturation and half an hour for reduction) than the innate kinetics and sensitivity reported for ecOxyR itself [26][27][28] .Speci cally, rate of ecOxyR oxidation is at a sub-second scale, and its reduction takes 5 ~ 10 minutes, implying that the insertion of the uorescence reporter domain may have slowed down the activation and deactivation of ecOxyR.Our engineering strategy aimed to maintain the exibility of the protein loop that drives the conformational change in the sensing domain (i.e.ecOxyR -LBD) in the derived sensors as we previously described for a GFP-based oROS-G sensor 29 .Speci cally, ecOxyR contains a hydrophobic pocket that forms the active center for peroxide interactions.Upon binding, peroxide forms a hydrogen bonding network with adjacent residues, bringing residues C199 and C208 into close proximity to form a disul de bridge.By analyzing the Bfactors of ecOxyR-LBD structures, we observed an evident high exibility peak in the 199-208 region [Fig.1A].We reasoned that preserving this exibility is necessary for e cient OxyR activation by peroxide 25,30,31 .Thus, inserting a bulky uorescent reporter between C199 and C208, as in HyPerRed and SHIRMP, may signi cantly slow OxyR's activation, and we explored alternatives outside this region [Fig.1B].Furthermore, red uorescent proteins pose challenges for versatile use involving optical multiparametric analysis or neuron expression.For example, cpmApple, used in HyPerRed, exhibits a false positive photochromic artifact induced by blue light commonly used to excite green uorescent proteins (e.g.488 nm) 22 and neuronal aggregation 32 .
Deo et al. proposed a chemigenetic solution for designing optogenetic sensors incorporating a selflabeling enzyme (HaloTag) with an irreversible conjugation of rhodamine-based Janelia Fluorophores (JF) 33,34 .Red to far-red shifted JFs exhibit exceptional photophysical characteristics such as brightness, and photostability, which surpasse existing red FPs.We aimed to engineer a new class of GEHIs using cpHaloTag labeled with the far-red uorescent JF635 as a reporter domain.Insertion of cpHaloTag into multiple positions outside of the C199-C208 loop in ecOxyR was well tolerated, and we identi ed a prototype sensor variant 213-214 with a robust response to bolus 300µM H 2 O 2 (∆F/Fo%: -38.23%; ci = [-40.36,-36.18]) [Fig.1C].Interestingly, we observed inverse responses (e.g. increase in peroxide level leads to decreased uorescence) to peroxide in all insertional variants.Thus, we aimed to improve the brightness, guided by structure predicted from ColabFold (AlphaFold2 with MMseqs2 for multiple sequence alignment) 24 .The prediction yielded a highly con dent structure of variant 213-214, which is exempli ed by a dimeric interface of the sensing domain that closely resembles the dimeric interface of reduced ecOxyR resolved by crystallography [Supp.Figure 1A, B].We superimposed the cpHaloTag-JF635 structure from [PDB: 6U2M] to identify the putative position of JF635 with the sensing domain of variant 213-214 [Supp.Figure 1C].The predicted position of OxyR sensing domain was oriented away from JF635 rather than covering the JF635 uorophore [Fig.1D], increasing the potential in uence of interdomain linker regions on the uorophore's local environment.This con guration is consistent with the spatial con guration of the chemigenetic calcium indicator HaloCaMP 33 .
Consequently, random mutagenesis of interdomain linker residues (XX-cpHaloTag-X, X indicates mutagenesis targets) affected both the sensor brightness and dynamic range [Fig.1E].From the linker variant library, we found a variant with 4.9-fold increased resting brightness and a 41% increase in dynamic range induced by 300µM H  1F]. In addition to the structural hypothesis of the interdomain linker's in uence on both sensor dynamics and brightness, we also identi ed F209 to be a putative mutational site for the uorophore local environment tuning, resulting in a more than a 3-fold difference in resting brightness between the dimmest variant (F209L) and the brightest variant (F209R) and the trend was also consistent when the sensor was labeled with ligand JF585.Unfortunately, the mutational bene t of MS-cpHaloTag-N and F209R was non-synergistic, which led us to exclude mutation F209R for our nal variant [Supp.Figure 1D-G].
We rst characterized oROS-HT 635 by exogenously applying H 2 O 2 to cells expressing the sensor and second by applying menadione, which induces intracellular peroxide generation.Menadione generates H 2 O 2 through various redox cycling mechanisms [35][36][37][38] [Fig.2A].Saturation of oROS-HT 635 induced by 300µM H 2 O 2 revealed a fast sub-second activation that could capture the extracellular H 2 O 2 diffusion across the imaging eld of view.It implies that the kinetic e ciency of the sensor passed a milestone of no longer being reaction-limited in this scenario.Intriguingly, the response amplitude of oROS-HT 635 at 10µM external peroxide was − 58.69% ∆F/Fo (ci = [-59.18,-58.18]), which is 87% of the amplitude at saturation upon 300µM peroxide (-67.27%∆F/Fo; ci = [-67.64,-66.91])), demonstrating the exceptional sensitivity of the sensor [Fig.2B] compared to previously reported red GEHIs.Previous studies showed the intracellular H 2 O 2 concentrations in HEK293 cells are at approximately 10 and 300 nM under these external conditions, respectively 14,39 .Furthermore, oROS-HT 635 allowed the monitoring of titrated peroxide levels in HEK293 cells induced by 10, 20, and 50µM of menadione.We oROS-HT 635 also displayed robust expression in various mammalian tissues (e.g.primary rat cortical neurons and ex vivo rat brain tissue) and human stem cell-derived models (e.g.cardiomyocytes and cortical neurons) [Fig.2F].Many experimental studies of intracellular peroxide often assume well-mixed uniformity of peroxide concentrations 19,39 .However, a previous model for cytosolic H 2 O 2 also showed spatial peroxide gradients in mammalian cells can emerge upon external peroxide stimulation 39 .Exceptional kinetics of oROS-HT 635 revealed spatial peroxide diffusion at ≈ 10µm/s in cardiomyocytes when exposed to bolus 300µM H 2 O 2 [Fig.2G].For the rst time, we optically monitored the in ux of H 2 O 2 into hiPSC-CMs with subcellular resolution, demonstrating that the sensor dynamics re ect the diffusion event.
Optimized biophysical properties and versatility of oROS-HT 635 under varying conditions.
We envision users of oROS-HT 635 studying peroxide dynamics under varying conditions.Thus, we further characterized notable features of oROS-HT 635 that demonstrate its environmental resiliency.oROS-HT 635 could be repeatedly activated and reduced back to baseline by serial peroxide stimulation and washout, demonstrating the reversibility of the sensor.Thus, the sensor is able to track real-time uctuations of intracellular peroxide [Fig.3A, B].Most beta-barrel uorescent proteins in sensor designs require oxygen for their uorophore maturation 40,41 .In addition, it was reported that GFP undergoes photoconversion under hypoxic conditions, where the excitation/emission spectra shift and become similar to RFP 42 .In contrast, the HaloTag-Rhodamine-based chemigenetic sensors incorporate synthetic uorophores which don't require oxygen for the protein maturation.To demonstrate oxygen independence during maturation, we engineered a loss-of-function mutation of oROS-HT 635 (C199S), a sensor variant insensitive to peroxide [Fig.3C].As a negative control, oROS-HT 635 -C199S can re ect any environmental effect on the level of uorescence that is not associated with the sensor function 14 .HEK293 cells transfected with oROS-HT 635 -C199S did not signi cantly differ in uorescence level when matured under normoxic or hypoxic conditions [Fig.3D-F].Red-shifted GEHIs are often limited for multiparametric use with green sensors due to a photochromic false positive artifact in response to blue light.However, oROS-HT 635 lacks this artifact, rendering oROS-HT 635 ideally compatible with green reporters [Fig.3G].Harnessing its multiplexing capability, we co-expressed oROS-HT 635 or oROS-HT 635 -C199S with a GFP-based pH indicator SypHer3s to demonstrate the low pH sensitivity of oROS-HT and its functionality under pH change with sequential events of 1.) acidic pH insult (pH 6) and 2.) 10µM menadione-induced peroxide increase.oROS-HT 635 did not respond to the initial change in pH but detected the menadione-induced increase in cytosolic peroxide, exemplifying its robust functionality under changing cellular pH environments [Fig.3H].As a benchmarking comparison, we compared pH-dependent uorescence change of oROS-HT-C199S and HyPerRed-C199S under neutral pH (pH 7.44) in response to pH shift to either 9 (basic) or 6 (acidic).oROS-HT 635 -C199S exhibited no signi cant uorescence change to either condition in contrast to responses of HyPerRed-C199S at equivalent conditions, demonstrating that the oROS-HT's uorescence is largely insensitive to physiological pH uctuation in contrast to HyPerRed [Fig.3I, J].Multiparametric analysis of the acute effect of aurano n on H 2 O 2 , redox potential, and Ca 2+ .
Acute effect of aurano n on cellular H 2 O 2 level and redox potential.
Grx1-roGFP2 is an indicator sensitive to glutathione redox potential (E GSH ).It is a fusion between glutaredoxin1 (grx1) and the redox-sensitive green uorescent protein roGFP2.Multiplexed imaging of oROS-HT 635 with Grx1-roGFP2 could enable peroxide imaging with augmented information about the redox cellular environment.Here, we monitored both sensors simultaneously in HEK293 cells upon 10µM H 2 O 2 exposure.We revealed sequential events of intracellular peroxide increase followed by a decrease in glutathione redox potential E GSH (peak oROS−HT to peak Grx = 3.12 s) as indicated by the respective sensor responses [Fig.4A].In contrast, inhibition of cellular redox potential with Trx/Grx (Thioredoxin/Glutaredoxin) inhibitor aurano n (1µM) showed rapid decay of E GSH followed by a slow increase of intracellular peroxide level.Interestingly, aurano n-induced peroxide build-up was transient, as we observed the elevation in peroxide level for 45 minutes after the application, followed by a recovery to the baseline within the following 60 min [Fig.4B], potentially due to stress-induced antioxidative capacity increase.Consistent with the previous reports 43,44 , we observed increased translocation of Nrf2 into the nucleus in HEK 293 cells within 30 minutes of exposure to 1µM Aurano n [Fig.4C, Supp. Figure 2].In conclusion, the multiplexed use of Grx1-roGFP2 with oROS-HT 635 exempli es the peroxide monitoring capability of oROS-HT 635 in the context of the cellular redox environment.

Acute effect of aurano n on peroxide and calcium dynamics in hiPSC-CM.
There is growing evidence of a mutual interplay between redox and Ca 2+ dynamics in biological systems 45 .Ca 2+ is functionally critical in excitable cells such as neurons and cardiomyocytes.Still, simultaneous real-time observations of oxidative stress and Ca 2+ in the same cell with a temporal resolution that can capture dynamic Ca 2+ transients (CaT) have been limited.Here, we performed multiplexed imaging of H 2 O 2 and CaT using oROS-HT 635 with Fluo-4, a Ca 2+ -sensitive green uorescent dye in hiPSC-CMs [Fig.4D].It is widely accepted that oxidative stress perturbs key Ca 2+ transporters like ryanodine receptors (Sarcoplasmic reticulum Ca 2+ leak) 46 , L-type calcium channels (ICaL, inward Ca 2+ current) 47 , and sarcoplasmic reticulum calcium ATPase pumps (SERCA, decreased Ca 2+ reuptake) [48][49][50] .Functional in uence of these perturbations can manifest as changes in speci c CaT phenotypes such as baseline Ca 2+ level, CaT amplitude, Time-to-Peak (TtP, on-kinetics), and Calcium Transient Duration 90% (CaTD90, completion of 90% of one CaT period).We explored how the aurano n-induced acute oxidative stress perturbs these transporters and affects Ca 2+ dynamics in detail.Previous studies reported aurano n-induced Ca 2+ increases in some cell types 51,52 .Indeed, aurano n (5µM) induced peroxide increase [Fig.4E] during the 20-minute imaging period, accompanied by an increase in basal Ca 2+ level [Fig.4F].Next, we extracted the CaT pro le from the Fluo-4 imaging data to further characterize the effect of aurano n [Fig.4G, H].Compared to the vehicle control, CaTs of aurano n-treated hiPSC-CM exhibited the following phenotypes: elevated CaT peak amplitude and prolonged TtP and CaTD90 [Fig.4I].
Modeling effect of perturbed Ca 2+ transport on cytosolic Ca 2+ levels in silico.
To investigate whether the oxidative insult and their effects on Ca 2+ transporters would lead to the observed changes in the CaT phenotypes, we simulated the intracellular Ca 2+ level dynamics using a preexisting computational model for CaT in iPSC-CMs 53 .Aligned with the reported effect of oxidative stress on the Ca 2+ transporters discussed above, we modi ed parameters corresponding to the cytosolic Ca 2+ e ux via SERCA, the SR Leak amplitude, and the conductance of the L-type Ca 2+ channel (ICaL) to model oxidative stress.The trend in simulation aligned with observed CaT phenotypes: decreased SERCA uptake simulated a pronounced increase in intracellular Ca 2+ baseline, delay of TtP and CaTD90, while higher ICaL conductance showed a pronounced increase in intracellular Ca 2+ baseline and CaT amplitude.Interestingly, increased SR Leak did not noticeably affect the aforementioned CaT phenotypes [Supp.Figure 3], re ecting the hiPSC-CMs electrophysiological immaturity.Speci cally, CaTs in hiPSC-CM models are often mostly governed by L-type Ca 2+ channel activities due to functional immaturity associated with SR-associated Ca 2+ transporters [54][55][56][57][58] , which may explain the observed CaT insensitivity to the increased SR leak.This result is further supported by our 11x10 synergistic perturbation simulation of ICaL and SERCA [Supp.Figure 4A].The baseline Ca 2+ level showed pronounced elevation with a focal point at SERCA 0.5x and ICaL 2.0x activity levels (relative to the starting conditions).In contrast, the CaT amplitude showed an elevated focal point around SERCA 0.25x, ICaL 2.0x activity levels.The focal point for TtP and CaTD90 elevation lies near SERCA 0.1x, ICaL 1.0x activity levels.We calculated a CaT in uence map derived from an additive weighing of the normalized individual phenotype arrays.It revealed the biased in uence of ICaL over SERCA for the phenotypic changes, implying that observed CaT phenotypes in the study may be the result of the biased effect of ICaL over SERCA for Ca 2+ handling [Supp.Figure 4B].The result acknowledges the potential intricate nature of effect of oxidative stress on Ca 2+ dynamics in cardiomyocytes, which calls for systemic studies on the in uence of oxidative stress on speci c Ca 2+ transport and their synergistic outcome.
Multiparametric imaging of intracellular and extracellular peroxide dynamics.oROS-HT 635 could be targeted to cellular sub compartments, including the mitochondrial matrix, mitochondrial intermembrane space, actin cytoskeleton, and intracellular side of the plasma membrane, and more [Fig.5A].Intracellular H 2 O 2 generation is potentially localized and functionally differentiated in aerobic organisms 59 , which calls for monitoring of H 2 O 2 in a spatially resolved manner (e.g.cellular subcompartments) 19 .Growing evidence demonstrates the signi cant contribution of NADPH oxidasesourced superoxide and peroxide in redox signaling and disease progression [60][61][62][63][64] .The oxidase generates H 2 O 2 on the extracellular side of the cellular plasma membrane 65 , constituting an extracellular pool of H 2 O 2 66 .Furthermore, its intracellular distribution is achieved through autocrine (aquaporin-mediated diffusion of peroxide 67,68 into cells) and paracrine 69 mechanisms.oROS-HT 635 fused to PDGFR transmembrane domain-based tra cking sequence (pDisplay vector, invitrogen) showed robust membrane localization of oROS-HT 635 , and its co-expression with oROS-G, a sensitive and fast green variant of oROS we previously reported 29 , was well tolerated in HEK293 [Fig.5B].Here, we measured 25µM menadione-induced H 2 O 2 increase in both extracellular and intracellular space.Intriguingly, we found that the extracellular peroxide response detected by oROS-HT 635 (inverse response sensor) was faster than oROS-G (direct response sensor).This supports previous observations that menadione increases H 2 O 2 in the extracellular space, potentially via NADPH oxidase-sourced peroxide [70][71][72][73] [Fig.5C].

Discussion
This study introduces a novel bright far-red chemigenetic indicator for peroxide, oROS-HT 635 .To fully harness the brightness of JF635 rhodamine dye, this inverted response sensor was further optimized for higher brightness and dynamic range while exhibiting unrivaled sensitivity and kinetics compared to existing red shifted GEHIs.Since oROS-HT 635 maintains bright uorescence in the sensor activation range (e.g.partially oxidized state), it detects high-delity signal at physiological peroxide levels.By incorporating chemigenetic reporter system (cpHaloTag-JF635), we could achieve oxygen-insensitive, pHresistant, and photochromic artifact-free imaging that vastly extends its application range.Guided by the crystal structures of OxyR, we optimized the peroxide sensing e ciency of oROS-HT 635 , implying the design avoids disruption of the exible protein region critical for H 2 O 2 -induced disul de bridging.
Harnessing oROS-HT 635 's exceptional multiplexing capability, we performed imaging paired with green uorescence-based redox potential and Ca 2+ reporters, allowing monitoring of peroxide level, along with changes in redox potential or Ca 2+ .Aurano n, a treatment for rheumatoid arthritis, is gaining attention from the cancer community as a potential therapeutic candidate due to its dose-and-cell-dependent multifaceted mode of action 74,75 .As a Trx/Grx inhibitor, it attenuates the intracellular antioxidant capacity, which increases oxidative stress.Intriguingly, recent studies to repurpose aurano n as a potential cancer therapeutics revealed a more nuanced role of aurano n as increasing cellular oxidative stress can activate regulators such as Nrf2 to boost cellular antioxidative capacity 74,75 .Here, we showed, in real-time, how low-dose aurano n initiates transient oxidative stress, followed by a Grx-independent reversal of H 2 O 2 levels.The time course of the reversal correlated with increased Nrf2 translocation into the cell nucleus in HEK293 cells, supporting observations from previous studies.Aurano n also altered dynamic Ca 2+ transients in hiPSC-cardiomyocytes, correlating with an increased level of H 2 O 2 .These observations were consistent with our computational simulation of the effect of oxidative stress on key Ca 2+ transporters.They con rmed previous studies identifying tight coupling between oxidative stress and Ca 2+ transport in various cells and tissues 45,46,50 .
Users can also exploit the remarkable subcellular targeting of oROS-HT 635 to monitor peroxide with higher spatial resolution near its sources.GEHIs have been pivotal in unraveling cellular peroxide topology by enabling optical monitoring of peroxide dynamics in spatially resolved manner in cytoplasmic and mitochondrial spaces 13,76 .oROS-HT can aid users to study peroxide biology by delineating the topology of peroxide from mitochondria, plasma membrane spaces, and paracrine peroxide 69 , which is critical for understanding the systemic propagation of peroxide build-up in tissues and organisms.Speci cally, membrane-tagged oROS-HT 635 provides new opportunities to investigate peroxide topology proximal to the plasma membrane, which is well demonstrated by the result that rehighlights the potential involvement of plasma membrane NADPH oxidases in menadione-induced peroxide production [70][71][72][73] .
The next iteration of oROS-HT 635 could be optimized for other JF dyes with shifted emission spectra ranging from (494 nm to 722 nm), further enhancing its exibility in multiplexed optogenetic applications.
Another possible avenue for future oROS-HT 635 development is maximizing its in vivo application capability.As a trade-off to its exceptional uorogenicity, the bioavailability of JF635 dye can be a challenge for animal application.We envision two paths for optimizing the use of oROS-HT 635 in live animals.First, introducing the dye into brain tissue can be aided with engineered solutions such as injection cannulas or drug delivery systems 77,78 .Alternatively, optimization of the oROS-HT 635 with highly bioavailable dyes (e.g.JF669) 79,80 can be explored for e cient animal applications.
In conclusion, oROS-HT 635 enables the monitoring of peroxides with high spatiotemporal resolution, offering unparalleled exibility in its multiplexed application with other optogenetic tools.The rapid kinetics and robust subcellular targeting capabilities of oROS-HT 635 , particularly at the outer and inner surfaces of the plasma membrane, render it an invaluable tool for investigating peroxide topology near the plasma membrane.When used with uorescent sensors for various analytes, oROS-HT 635 facilitates a dynamic, multidimensional analysis of peroxide changes and environmental responses in real-time, enhancing the contextual understanding of peroxides in biological systems.

Declarations
Health O ce of Laboratory Animal Welfare (OLAW), is registered with the United States Department of Agriculture (USDA, certi cate #91-R-0001), and is accredited by American Association for Accreditation of Laboratory Animal Care International.

Methods
Molecular Biology oROS-HT variants were all cloned based on the pC1 plasmid backbone from pC1-HyPer-Red (Addgene ID: 48249).Primers for point mutations or fragment assembly required to generate the oROS-HT screening variants were designed for In Vitro Assembly cloning (IVA) technique 81 , and they were ordered from Integrated DNA Technologies (IDT).All gene fragment ampli cations were done using Super -II polymerase (Invitrogen; 12368010).Ampli cation of the DNA fragment was veri ed with agarose gel electrophoresis.30 minutes of DpnI enzyme treatment were done on every PCR product to remove the plasmid template from PCR samples.Circulaization or assembly of the PCR products was achieved with the IVA technique, while the linear DNA products were transformed into competent E.Coli cells (DH5 or TOP10) and grown on agar plates that contain kanamycin selection antibiotic (50 µg/mL).Upon colony formation, single colonies were picked and grown in 5mL cultures containing LB Broth (Fisher BioReagents; BP9723-2) and selection antibiotic (/kanamycin; 50 µg/mL) overnight (37°C, 230 RPM).DNA was isolated using Machery Nagel DNA prep kits (Machery Nagel; 740490.250).Sanger sequencing (Genewiz; Seattle, WA) or Whole plasmid nanopore sequencing (Plasmidsarus; Eugene, OR) of the isolated plasmid DNA was used to con rm the presence of the intended mutation.Genes encoding the nal variants were cloned into a CAG-driven backbone, pCAG-Archon1-KGC-EGFP-ER2-WPRE (Addgene; #108423), using the methods above.All subsequences were veri ed with Sanger sequencing (Genewiz; Seattle, WA) or Whole plasmid nanopore sequencing (Plasmidsarus; Eugene, OR).

Protein structure prediction and analysis
Protein structure analysis and plotting were performed using Chimera-X-1.2.1.Oxidized [PDB:1I6A] and reduced [PDB:1I69] crystal structures of ecOxyR were imported from the Protein Data Bank (PDB).Pairwise residue distance between reduced and oxidized ecOxyR structure was achieved by aligning both structures using a matchmaker algorithm that superimposes protein structures by creating a pairwise sequence alignment and then tting the aligned residue pairs to derive pairwise residue distances.The and removing supernatant + resuspending in 10 mLs of Neuronal Basal Media (Invitrogen; 10888022) supplemented with B27 (Invitrogen; 17504044) and glutamine (Invitrogen; 35050061) (NBA++).After nal wash spin and supernatant removal, cells were resuspended in 10 mLs of NBA++ before counting.Just before neurons were plated, matrigel was aspirated from the wells.Neurons were plated on the prepared culture plates at the desired seeding density.Twenty-four hours after plating, 1µM AraC (Sigma; C6645) was added to the NBA++ growth media to prevent the growth of glial cells.Plates were incubated at 37°C and 5% CO2 and maintained by exchanging half of the media volume for each well with fresh, warmed Neuronal Basal Media (Invitrogen; 10888022) supplemented with B27 (Invitrogen; 17504044) and glutamine (Invitrogen; 35050061) every three days.
AAV transduction and confocal imaging: After 1 day in vitro (DIV), crude AAV9-CAG-oROS-HT prep was added to the slices to be expressed.At the end of the 3-day incubation, 1 μM JF635-HTL was added to the slices for an additional 48 hours.OWH brain slices were transferred to 35mm confocal dishes (VWR, 75856-742).Confocal images were acquired with 10x (Nikon Plan Apo 10x Objective, 0.45 numerical aperture) and 20x (Nikon Plan Apo 10x Objective, 0.75 numerical aperture) magni cations (Nikon Corporation, Minato City, Tokyo, Japan).Brain slice tile scans were obtained with the Cy5 channel before multiple representative images were taken from both the cortex and striatum of each slice.Image acquisition settings were kept consistent before and after the 300µM H 2 O 2 stimulation.
Immuno staining uorescence staining performed for Nrf2 translocation study were done using polyclonal Nrf2 antibody (PA5-27882, Invitrogen) and Donkey anti-Rabbit IgG Alexa Fluor 488 (A21206, Invitrogen).HEK293 cells for each condition were xed in 4% paraformaldehyde for 15 minutes and permeabilized in 0.2% Triton-x solution for 1 hour.After blocking the xed cells for 1 hour with 0.5% Bovine Serum Albumin (BSA) blocking buffer in TBST, Cells were then incubated with primary antibodies diluted in the blocking buffer overnight at 4°C.The next day, cells were washed 3 times with PBS.They were then incubated in a secondary antibody solution containing secondary antibodies diluted in 0.5% BSA in PBS overnight at 4°C.Counterstaining was performed with Vectashield containing DAPI (Vector Labs).

Microscopy
Imaging experiments described in this study were performed as follows unless speci cally noted.Hypoxic oROS-HT sensor maturation in HEK293 2-day post-seeding of HEK293 cells in 24 well plates (150,000 cells/well), culture media was swapped from complete DMEM media (as mentioned above) to complete Fluorobrite DMEM (A1896701, Gibco) with 20mM HEPES.After 2-hour of acclimation, cells were transfected (Lipofectamine-based, as described above) with pC1-oROS-HT-C199S (Loss-of-function), with 100nM JF635-HTL.Immediately after the transfection, transfected cells were either incubated at 37°C in an atmospheric environment or under hypoxic conditions.For hypoxic conditions, culture plates were transferred into a sealable chamber.The chamber was ushed with N 2 for 10 min at a ow rate of 10 L/min before being placed into the incubator.
Approximately 18 hours after, epi uorescence imaging were performed as described earlier.
Multiplexed experiments oROS-HT/SypHer3s: HEK293 cells were co-transfected with pC1-oROS-HT/pC1-SypHer3s or pC1-oROS-HT-C199S/pC1-SypHer3s as described above.2 days post-transfection, both sensors expressed in HEK293s were imaged using epi uorescence microscope.pH change experiment for oROS-HT-C199S were performed with HEK293s in PBS (10010001, Gibco) prepared at pH of 6, 7.44, and 9. Fluorescence level for GFP and Far-red pro le were captured every 1.5 seconds.Sequential pH-changes plus Menadione applications were performed with HEK293s in PBS (pH 7.44), which was changed to PBS (pH 6) followed by menadione stimulation prepared in PBS (pH 6).Fluorescence level for GFP and Far-red pro le were captured every 2 seconds.oROS-HT/Grx1-roGFP2: HEK293 cells were co-transfected with pC1-oROS-HT and pC1-Grx1-roGFP2 as described above.2 days post-transfection, both sensors expressed in the cells with live cell imaging solution with 10mM glucose (LCIS+, Gibco, A14291DJ) were imaged using an epi uorescence microscope.For the sequential response of oROS-HT/Grx1-roGFP2 to 10µM H 2 O 2 , uorescence level for GFP and Far-red pro le were captured every second.For the response to Aurano n, uorescence levels for GFPratio, GFP, and Far-red pro les were captured every minute.oROS-HT/Fluo-4: hiPSC-CMs were transfected with pCAG-oROS-HT as described above.2 days posttransfection, cells were incubated with Fluo-4 (Invitrogen, F14201) at 5µM and JF635-HTL in RPMI + B27+insulin for 1 hour prior to imaging.For the response to Aurano n, uorescence level of GFP pro le (10Hz) and Far-red (0.1Hz) pro le were acquired every 10 seconds for hiPSC-CMs in HEPES-buffered RPMI + B27+insulin.

Analysis
Analysis of cell uorescence imaging data was done by FUSE, a custom cloud-based semi-automated time series uorescence data analysis platform written in Python.First, the cell segmentation quality of the selected Cellpose 92 model was manually veri ed.For the segmentation of cells expressing cytosolic uorescent indicators, model 'cyto' was selected as our base model.If the selected Cellpose model was low-performing, we further trained the Cellpose model using the Cellpose 2.0 human-in-the-loop system 93 .
Using an "optimized" segmentation model, uorescence time-series data is extracted for each region of interest.This allows for unbiased extraction of change in cellular uorescence information for a complete set of experimental samples.Extracted uorescence data is normalized as speci ed in the text using a custom Python script.

Computational Cell Scale Modeling
We used an existing model of iPSC-CM membrane kinetics 53 with one modi cation.Based on experimental observations, the spontaneous beating of the iPSC-CMs was observed to be around 0.5 Hz.
To re ect this observation in our computational simulations, we increased the maximal value of the inward recti er potassium (I K1 ) by a factor of 1.71484375.This change resulted in a decrease in spontaneous beating rate from 1.1 Hz to 0.5 Hz.To simulate ROS effects on iPSC-CMs, we ran simulations in which we modi ed parameters corresponding to maximal e ux via the SR Ca 2+ ATPase (SERCA2a), SR Leak amplitude, and maximal conductance of the L-type Ca 2+ channel (g CaL ).The perturbation factor for SERCA e ux varied from 0.1 to 1.0 in steps of 0.1.SR Leak amplitude and g CaL were both increased from the default level (1.0) to 2.0 in steps of 0.1.Simulations of bioelectrical activity were conducted using openCARP 94 , a cardiac electrophysiology modeling software that is freely available for non-commercial reuse (see: http://opencarp.org/).Stimulated Ca i values were post-analyzed with custom-written python scripts.Scripts and les used to run all simulations can be found at the Github depository.

Figures
Page 24/28    See image above for gure legend.

Figure 1 See
Figure 1

Figure 2 See
Figure 2

Figure 3 See
Figure 3