Reactive Oxygen Species: Not Omnipresent but Important in Many Locations

Reactive oxygen species (ROS), such as the superoxide anion or hydrogen peroxide, have been established over decades of research as, on the one hand, important and versatile molecules involved in a plethora of homeostatic processes and, on the other hand, as inducers of damage, pathologies and diseases. Which effects ROS induce, strongly depends on the cell type and the source, amount, duration and location of ROS production. Similar to cellular pH and calcium levels, which are both strictly regulated and only altered by the cell when necessary, the redox balance of the cell is also tightly regulated, not only on the level of the whole cell but in every cellular compartment. However, a still widespread view present in the scientific community is that the location of ROS production is of no major importance and that ROS randomly diffuse from their cellular source of production throughout the whole cell and hit their redox-sensitive targets when passing by. Yet, evidence is growing that cells regulate ROS production and therefore their redox balance by strictly controlling ROS source activation as well as localization, amount and duration of ROS production. Hopefully, future studies in the field of redox biology will consider these factors and analyze cellular ROS more specifically in order to revise the view of ROS as freely flowing through the cell.

O 2 − quickly dismutates to H 2 O 2 , which is more, although not freely, diffusible for cellular membranes (Bienert et al., 2006;Wang et al., 2020;Chauvigne et al., 2021), which questions saturation of the cell with H 2 O 2 to fulfill signaling functions in compartments, which are not in direct proximity to the ROS source (Beretta et al., 2020;Sies, 2021). Communication between cellular compartments can be achieved by aquaporins, which facilitate a controlled passage of H 2 O 2 over membranes (Bienert and Chaumont, 2014;Wang et al., 2020). H 2 O 2 has a longer cellular half-life (∼1 ms) with concentrations of ∼10 −7 M under cellular homoeostatic conditions (D'Autreaux and Toledano, 2007). Because of these properties, it functions as an important signaling molecule involved in many different cellular processes (Kamata et al., 2005;Tonks, 2005;Rhee, 2006;Marinho et al., 2013;Holmstrom and Finkel, 2014;Romero et al., 2014;Jones et al., 2016;Short et al., 2016;Zhang et al., 2016;Herb et al., 2019b;Sies and Jones, 2020;Chauvigne et al., 2021). H 2 O 2 -mediated signaling is mainly based on the oxidation of cysteine residues of proteins (Chiarugi et al., 2001;Rhee, 2006;Herscovitch et al., 2008;Romero et al., 2014;Jones et al., 2016;Short et al., 2016;Herb et al., 2019b). These cysteine residues have a low pKa, are exposed to the cytosol and deprotonated to thiolate groups (Finkel, 2011;Poole, 2015). An increase to nanomolar concentrations (∼100 nM) of H 2 O 2 is sufficient to induce reversible oxidation. This can lead to allosteric protein changes that alter the enzymatic function of the target proteins in many ways (Lee et al., 1998;Meng et al., 2002;Kamata et al., 2005;Tonks, 2005). ROS-mediated oxidation can also lead to covalent linkage of cysteine residues by disulfide bonds (Herscovitch et al., 2008;Zhou et al., 2014;Herb et al., 2019b). Since these H 2 O 2 -mediated protein oxidations can be reversed by the antioxidant defense system, they represent important redox switches involved in various cellular processes (Barford, 2004;Holmstrom and Finkel, 2014). Excessive H 2 O 2 production, however, leads to further oxidation of the oxidized cysteines, which is an irreversible process and results in protein malfunction (Winterbourn and Hampton, 2008).
Unfortunately, a lot of studies, which show the important role of ROS in various cellular processes, often suggest that ROS are produced in excess, saturate the cell and react randomly with redox-sensitive targets. This is mainly due to experimental setups that might lead to misinterpretation of the location of ROS production in cells.
Many studies use only one type of ROS probe, but do not provide an explanation for the choice, such as specificity for a cellular compartment or a defined type of ROS subspecies.
Also the combined use of only globally working ROS scavengers in combination with ROS probes that detect total cellular ROS can lead to results, which suggest that ROS are present in the whole cell after diffusion from the location of their production. With NAC as most prominent globally working ROS scavenger (Patriarca et al., 2005;Aldini et al., 2018;Ezerina et al., 2018) only the general involvement of ROS in the cellular process of interest can be investigated, but no compartment-specific ROS production can be analyzed. More examples of globally working ROS scavengers are Tempol (4-Hydroxy-Tempo) (dismutation of O 2 − into H 2 O 2 ) (Bernardy et al., 2017;Herb et al., 2019b), Tiron (a global O 2 − scavenger) (Krishna et al., 1992;Hein and Kuo, 1998;Manzano et al., 2000), Trolox (globally scavenges OOH and OOR) (Davies et al., 1988;Dugas et al., 2000) and ebselen (effectively removes H 2 O 2 and ONOO − ) (Nakamura et al., 2002;Matsushita et al., 2004;Mugesh, 2013). All of the scavengers mentioned above are diffusible (Davies et al., 1988;Krishna et al., 1992;Hein and Kuo, 1998;Haj-Yehia et al., 1999;Dugas et al., 2000;Manzano et al., 2000;Rak et al., 2000;Nakamura et al., 2002;Matsushita et al., 2004;Mugesh, 2013;Herb and Schramm, 2021). Assessment of specific removal of ROS subspecies and therefore their involvement in cellular processes is possible with these substances, but they cannot be used to identify the specific compartment in which the ROS exert their function.
Not only the location of ROS production, but also their various sources and their activation, regulation and termination is of major importance for the understanding of the complex redox maintenance in cells. For the identification of ROS sources it is not always possible to provide genetic evidence with a knock-out system or by siRNA usage. ROS source inhibitors are in these cases an option to block ROS production and analyze possible ROS sources. There are a lot of specific ROS source inhibitors commercially available and the choice is continuously expanded (Murphy, 2009;Wind et al., 2010;Altenhofer et al., 2015;Herb and Schramm, 2021).
In mitochondria, the complexes of the ETC not only are essential for energy generation of the cell, but are also ROS production sites (Nohl et al., 2003;Lambeth and Neish, 2014). Inhibition of the complexes for analysis of ROS production might also result in energy deprivation and the energy status of the cell has to be checked every time these inhibitors are used. Typically used inhibitors are rotenone (Stowe and Camara, 2009;Heinz et al., 2017;Scialo et al., 2017), which inhibits complex I and increases ROS production inside the mitochondrial matrix (St-Pierre et al., 2002;Lambert and Brand, 2004;Panov et al., 2005;Stowe and Camara, 2009;Sena et al., 2013) and antimycin A (Murphy, 2009;Bleier and Drose, 2013), which inhibits complex III and increases ROS production into the intermembrane space (IMS) (Chen et al., 2003;Han et al., 2003;Al-Mehdi et al., 2012;Quinlan et al., 2012;Herb et al., 2019a). The most commonly used ROS probe for detection of mitochondrial ROS is MitoSOX, which measures O 2 − exclusively inside the mitochondrial matrix (Robinson et al., 2006;Mukhopadhyay et al., 2007;Ernst et al., 2021). However, since the ETC complexes show compartment-specific differences concerning ROS production (Fridovich, 1997;Murphy, 2009;Brand, 2010;West et al., 2011b;Herb and Schramm, 2021), this probe can only be used to measure ROS production inside mitochondria and, therefore, other cellular compartments should always be analyzed in addition. In healthy, undamaged mitochondria, ROS cannot escape the mitochondrial matrix because of the very effective antioxidative defense system (Roca and Ramakrishnan, 2013;Briston et al., 2017;Hos et al., 2017;Hernansanz-Agustin et al., 2020;Lin et al., 2020;Wang et al., 2020;Zhao et al., 2020). Only after prolonged overproduction or when the structure of the mitochondrial membranes is ruptured, either by opening of the mitochondrial permeability transition pore or direct damage, e.g., by pathogenic toxins, ROS can escape from the matrix into the cytosol (Koterski et al., 2005;Stavru et al., 2011;Roca and Ramakrishnan, 2013;Briston et al., 2017;Hos et al., 2017;Zhang Y. et al., 2019;Zhao et al., 2020). Nevertheless, the general term "mitochondrial ROS" is used in many studies, which often is synonymous for matrix-located mitochondrial ROS production measured by MitoSOX. ROS measurements in other cellular compartments as well as an explanation if and how the mitochondrial ROS escape from the matrix and fulfill their role in the cell, with a few exceptions (Koterski et al., 2005;Zhou et al., 2011;Roca and Ramakrishnan, 2013;Briston et al., 2017;Hos et al., 2017;Herb et al., 2019b;Zhao et al., 2020), are often not provided. Additionally, the usage of inhibitors of the ETC, like rotenone or antimycin A, which have compartment-specific effects on ROS production in combination with diffusible ROS probes can also lead to misinterpretations of the performed ROS measurements. For further reading concerning ROS scavengers and inhibitors, we like to point to other reviews (Wind et al., 2010;Altenhofer et al., 2015;Herb and Schramm, 2021).

ROS PRODUCTION: THE DOSE MAKES THE POISON
A model that involves an uncontrolled increase in total cellular ROS levels implies that cells take into account the collateral damage that ROS can inflict while enroute to their redox-sensitive target, that can be at a completely different cellular location (Bulua et al., 2011;Nazarewicz et al., 2013;Kelly et al., 2015;Garaude et al., 2016;Kim et al., 2017; Figure 1). But oxidative distress (Buczynski et al., 2013;Li et al., 2013;Bhattacharyya et al., 2014;Niki, 2016;Sweeney and McAuley, 2016;Yang et al., 2016;Sies and Jones, 2020) is a situation for the healthy cell that has to be avoided. Tightly controlled production of ROS in direct vicinity of a redox-sensitive target (Tai et al., 2009;Finkel, 2011;Nathan and Cunningham-Bussel, 2013;Reczek and Chandel, 2014;Herb et al., 2019b) requires much less ROS production and hence results in much less collateral damage, while fulfilling important cellular functions, i.e., oxidative eustress (Niki, 2016;Sies and Jones, 2020; Figure 2). FIGURE 1 | Several studies suggested that ROS are produced in excess, saturate the cell and find their redox-sensitive targets at random. Usage of diffusible ROS probes, globally working ROS scavengers and unspecific inhibitors often place the suggested ROS source at a completely different location than the redox-sensitive target, which might lead to the interpretation that cells "take into account" the damage that ROS can inflict on their way to the target molecule. The importance of ROS in general for various cellular processes was shown by many excellent studies (Bulua et al., 2011;Nazarewicz et al., 2013;Kelly et al., 2015;Garaude et al., 2016;Kim et al., 2017), however, diffusible ROS probes or only one ROS probe are often used to determine ROS production in cells, which might lead to the suggestions, e.g., that (1) ROS escape from the mitochondrial matrix and regulate expression and secretion of cytokines (Bulua et al., 2011;Kelly et al., 2015), (2) extracellular Nox2-derived ROS modify enzyme activity in the mitochondrial matrix or matrix-derived ROS modulate Nox2 activity (Nazarewicz et al., 2013;Garaude et al., 2016) or (3) ROS produced by ER-located Nox4 reach the phagosome for inactivation of phagocytosed parasites .
Frontiers in Cell and Developmental Biology | www.frontiersin.org FIGURE 2 | Growing evidence supports the hypothesis that cellular compartments show big differences and tight regulation of their redox status. The induction of ROS production is controlled by the cell in terms of location, source, duration and amount. The localized and timely controlled ROS production in the direct vicinity of the redox-sensitive target reduces the induced damage to cellular components and results in beneficial consequences for the cell, a condition termed as oxidative eustress (Sies, 2021). Examples of localized ROS production are (1) the production of antimicrobial ROS by Nox2 (Craig and Slauch, 2009;Gluschko et al., 2018) or mitochondria, which are recruited to pathogen-containing phagosomes (West et al., 2011a;Geng et al., 2015), (2) the recruitment of the redox-regulated target to ROS-producing mitochondria for NLRP3 inflammasome activation (Zhou et al., 2011) or the relocation of mitochondria to the nucleus for ROS-mediated nuclear signaling (Al-Mehdi et al., 2012) and (3) ROS production by ER-localized Nox4 during formation of mitochondria-associated membranes for regulation of calcium signaling (Beretta et al., 2020).

CONCLUDING REMARKS
In recent years, more and more studies supported the model-and highlighted the importance-of localized cellular ROS production in direct vicinity of the redox target (Meinhard and Grill, 2001;Veal et al., 2007;Go and Jones, 2008;Craig and Slauch, 2009;West et al., 2011a;Wink et al., 2011;Zhou et al., 2011;Al-Mehdi et al., 2012;Naviaux, 2012;Allan et al., 2014;Romero et al., 2014;Geng et al., 2015;Gluschko et al., 2018;Herb et al., 2019b;Acin-Perez et al., 2020;Beretta et al., 2020;Chanin et al., 2020;Miller et al., 2020;Sies and Jones, 2020;Herb and Schramm, 2021;Ligeon et al., 2021b;Sies, 2021;Wong et al., 2021). Therefore, the model of ROS molecules as "omnipresent and freely diffusing throughout the cell" should always be interpreted carefully in the context of research and highly depends on the proper use of ROS probes, scavengers and inhibitors. In healthy cells, ROS should be considered as molecules, whose production is tightly controlled in terms of stimulus, source, location, duration and amount.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

AUTHOR CONTRIBUTIONS
MH and MS: conceptualization. MH: writing-original draft preparation and visualization. MH, AG, and MS: writing-review and editing and funding acquisition.
MS: supervision and project administration. All authors have read and agreed to the published version of the manuscript.

FUNDING
This work was supported by the Deutsche Forschungsgemeinschaft (DFG) grant SCHR 1627/2-1 to MS, by the Köln Fortune grant 278/2019 to AG, and by the Köln Fortune grant 302/2020 to MH.