Beyond the Cuvette: Redox Indicators in Biological Experiments

Redox imaging using fluorescent dyes potentially surpasses more traditional methods (electrodes and immunoassays) due to the possibility of reporting, in a specific way, the production of reactive oxygen species (ROS) or the change in the state of a redox couple in living cells. Fluorescent dyes can be targeted to subcellular compartments (this is especially true for protein dyes but also applies to some extent to synthetic dyes—see the review by Dr. Ribou in this Forum), which confers a clear spatial resolution advantage compared with other methods. In addition, fluorescent redox dyes can be used in conjugation with other dyes for simultaneous report of several cell parameters by a careful choice of imaging wavelengths. Recent efforts in the development of new probes with higher specificity, signal over noise ratio, and red-shifted excitation and emission wavelength have given birth to a plethora of genetically encoded dyes, opening the possibility for in vivo imaging (4, 5).

The traditional path to validation of a newly engineered sensor is first, an in-depth characterization of the fluorescence properties and specificity of the sensor in a cuvette, second, a trial of the sensor in a relatively simple system (yeast and mammalian cell lines), and finally, the release of the new tool to the community. However, no matter how thoroughly the sensor has been characterized in vitro or in simple models, the uninformed researcher will face difficulties while trying to apply the new tool in his or her own experimental system. Before using a sensor to study cell signaling pathways, one must at least ask the two following questions: (a) How is the sensor interfering with the system being studied? For instance, a ROS probe could behave as an antioxidant. (b) How is the system interfering with the measurement? In other words, which other cell parameters could affect the sensor (pH, Cl⁻, and protease levels)? How the autofluorescence of the system impacts the sensor measurement? Is there any physical specificity with the model that might alter the detection? For instance, light scattering effects in thick tissues are wavelength dependent and might affect the ratio recorded with ratiometric dyes.

The Forum aims to address those questions and raise awareness on the problems that one might encounter during the use of ROS and redox sensors in biological systems. It is composed of four review articles, one rebuttal letter, and an original research article.

The review by Ribou introduces this Forum by giving a chemist point of view on several aspects of synthetic redox probes, commercially available or not, and providing practical handling considerations for rigorous measurements. As we talk about the versatility of genetically encoded dyes in terms of subcellular targeting, we often forget that synthetic dyes can also be, to some extent, addressed to specific cell compartments. Ribou details targeting strategies of synthetic dyes, with a special emphasis on mitochondrial probes.

The review by Wang et al. also concerns a mitochondrial targeted probe, cpYFP. cpYFP is, in a way, an excellent example of the benefits and downsides of the fluorescent dyes. On one hand, it highlights the potentially impactful results that can be obtained by the use of ROS and redox probes. On the other hand, it provides an important illustration of how genetically encoded indicators are prone to artifacts.

The review relates the story of the discovery of cpYFP as a superoxide sensor. While trying to measure relatively slow calcium events in mitochondria using the genetically encoded calcium dye pericam, the authors discovered 20 s long transient increases in fluorescence intensity that involved spatially restricted portions of the mitochondrial network (7). They soon realized that these events were not calcium transients, but involved detection of other parameters by the fluorescent protein moiety of the genetically encoded calcium dye, the cpYFP. The authors characterized the sensor in vitro and in several cell types, and identified the measured signals as transient increases in superoxide level. As a result, they named these events superoxide flashes (known now as mitochondrial flashes).

In addition to highlighting some subcellular functional organization of the mitochondrial network, the article brought the promise of a long awaited specific and reversible sensor for superoxide detection. However, like many other protein sensors, especially those based on circularly permuted proteins, cpYFP is sensitive to its environment. Unfortunately, its pK _a (around 8.5) makes it very responsive to mitochondrial pH changes. Hence, in their current review, the authors propose that mitochondrial flashes are a mix of superoxide and pH detection. Superoxide nature of mitochondrial flashes has been questioned soon after their publication (3), bringing others to consider that cpYFP does not detect superoxide at all and that flashes consist in transient alkalizations of mitochondrial matrix (6). In their rebuttal letter, Schwartzlander and Demaurex put forward several arguments in favor of a pH composition of mitochondrial flashes and propose some experiments to help solve the controversy.

The review by Nault et al. brings an additional perspective to this debate. By elegantly using the phagosome as a thread, the authors describe how extreme environmental constraints encountered in some cell compartments and microorganisms might interfere with ROS detection. The phagosome exhibits high protease content, low and fluctuating pH, which can alter the properties of genetically encoded indicators. The authors illustrate how H₂O₂ measurements with Hyper and perFret are affected in this challenging system. They also offer a solution: ongoing development of redox sensors based on proteins with low pK _a, which would be pH resistant under most physiological situations.

The use of sensors in vivo is illustrated in the work of Braeckman et al., which reviews the field of research of ROS signaling in the nematode Caenorhabditis elegans. C. elegans is an interesting model organism for the study of ROS and redox signaling as it contains an elaborate machinery of redox-related proteins. In addition, it is easy to genetically manipulate, and its small transparent body and simple, well-characterized anatomy allow identification and fluorescence imaging of several cell types within the intact animal (1). However, C. elegans also exhibits some downsides such as decrease in vitality in some reporter strains and interference of some autofluorescent compartments with the sensor signal. As a consequence, redox imaging in the live animal requires a careful choice and handling of the indicator. Using roGFP2 ORP1, the authors investigated the pattern of H₂O₂ production within the C. elegans reproductive system.

Finally, the research article by Mongeon et al. illustrates the measurement of NADH/NAD+ ratio in brain tissue using a genetically encoded indicator, Peredox, and biphoton fluorescence lifetime imaging (2P-FLIM). Complex brain functions such as synaptic transmission and plasticity are regulated by ROS and redox processes (2), and the study of such regulations requires detection of ROS and redox signals in brain preparations. However, the most common experimental model in neuroscience, the brain slice, is hard to image mostly due to the thickness of the tissue that generates light scattering effects. The authors elegantly overcome these limitations by using 2P-FLIM, a technique that allows quantitative measurement in thick tissue independently of dye concentration, coupled to the biosensor Peredox. They show that Peredox exhibits a large change in its fluorescence lifetime while sensing NADH/NAD+ ratio, which makes it suitable for FLIM experiments. Using this technique, they report differences in NADH/NAD+ ratio between two cell types (neurons and astrocytes) of a brain slice.

In summary, this Forum highlights the latest developments in ROS and redox imaging in complex systems, such as organelles, tissue explants, and in vivo. With the rise of microscopy techniques, engineering of ROS and redox fluorescent probes is a very active field of research, with new sensors being recurrently released. Each of them carries the promise of in vivo study of redox signaling. However, using these new tools to answer biological questions is not as straightforward as one might wish. As illustrated in each article of this Forum, adapting the technology to various experimental systems requires the pooling of expertise as diverse as microscopy, biophysics, cell biology, and chemistry. I hope that these examples will stimulate the discussion among scientists with different backgrounds and help promote the expansion of this exciting field of research.