Individuals with higher metabolic rates have lower levels of reactive oxygen species in vivo

There is increasing interest in the effect of energy metabolism on oxidative stress, but much ambiguity over the relationship between the rate of oxygen consumption and the generation of reactive oxygen species (ROS). Production of ROS (such as hydrogen peroxide, H2O2) in the mitochondria is primarily inferred indirectly from measurements in vitro, which may not reflect actual ROS production in living animals. Here, we measured in vivo H2O2 content using the recently developed MitoB probe that becomes concentrated in the mitochondria of living organisms, where it is converted by H2O2 into an alternative form termed MitoP; the ratio of MitoP/MitoB indicates the level of mitochondrial H2O2 in vivo. Using the brown trout Salmo trutta, we tested whether this measurement of in vivo H2O2 content over a 24 h-period was related to interindividual variation in standard metabolic rate (SMR). We showed that the H2O2 content varied up to 26-fold among fish of the same age and under identical environmental conditions and nutritional states. Interindividual variation in H2O2 content was unrelated to mitochondrial density but was significantly associated with SMR: fish with a higher mass-independent SMR had a lower level of H2O2. The mechanism underlying this observed relationship between SMR and in vivo H2O2 content requires further investigation, but may implicate mitochondrial uncoupling which can simultaneously increase SMR but reduce ROS production. To our knowledge, this is the first study in living organisms to show that individuals with higher oxygen consumption rates can actually have lower levels of H2O2.


Introduction
Oxidative stress occurs when the generation of reactive oxygen species (ROS) exceeds the capacity of antioxidant defence and repair mechanisms, thereby generating oxidative damage to lipids, DNA and proteins [1]. Given their potential role in cellular senescence, ROS are proposed as being one of the main mediators of life-history trade-offs [2,3]. Most of the ROS present in cells are produced within the mitochondria as natural by-products of aerobic respiration [1]. This has led to the pervasive idea that increased energy expenditure towards one life-history trait will result in greater ROS production, leading to accelerated senescence [4][5][6]. However, it is still unclear whether higher aerobic respiration actually alters in vivo ROS levels [7,8].
Some of the oxygen consumed by the mitochondria (mtVO 2 ) is subsequently reduced to superoxide and other ROS such as hydrogen peroxide (H 2 O 2 ) and hydroxyl radicals [1]. Despite the long-held belief that individual organisms consuming more oxygen have higher ROS production, the relationship between mtVO 2 and in vitro ROS production is unclear, with studies reporting positive, negative or no correlation between mtVO 2 and ROS among individuals [8]. However, several recent studies have raised reservations over the measurement of ROS production in vitro [8][9][10], since artificially high levels of metabolic substrates [11] and very high partial pressures of oxygen (20% in contrast to approx. 5% in vivo [10]) make extrapolations of in vitro results to the in vivo situation potentially problematic [7,8]. This is especially true when examining among-individual variation in mitochondrial traits, given that the in vitro conditions standardize the mitochondrial environment among individuals, which may hide the actual sources of variability in the relationship between oxygen consumption in vivo and mitochondrial ROS production [11].
To avoid some of these potential confounding effects in vitro, a ratiometric probe called MitoB has recently been developed to infer the level of mitochondrial H 2 O 2 in vivo [12]. When this artificial probe compound is administered to the organism, it becomes concentrated within the mitochondria, where it is converted to its alternative form MitoP by H 2 O 2 . The level of mitochondrial H 2 O 2 can then be expressed as the rate at which MitoB is converted to MitoP [12]. Here, we use the MitoB probe to directly quantify variation in H 2 O 2 content in living animals and relate this to variation in whole animal oxygen consumption among individuals of similar age and under identical environmental and nutritional states, using brown trout (Salmo trutta), a species known to exhibit consistent individual differences in oxygen consumption [13].

Material and methods
Juvenile brown trout S. trutta were collected from the wild and then kept in individual compartments for 22 weeks under standard conditions of temperature (mean + actual range: 11.5 + 18C) and photoperiod (12 L : 12 D), as described for these same individuals in [14]. Forty fish were randomly assigned to eight batches of five and fed daily with a specific ration calculated for each fish based on its weight [15].
At week 21, the standard metabolic rate (SMR), defined as the oxygen consumption of a resting and post-absorptive ectotherm at a given temperature, was measured over a 20 h period using flow-through respirometry (further details in [14] and electronic supplementary material). The SMR was calculated as the mean of the lowest 10th percentile of oxygen consumption measurements after controlling for body mass and is described hereafter as the residual SMR (rSMR), in mg O 2 h 2 1 .
Each fish was allowed a week of recovery after its SMR measurement before being injected with 50 nmol of MitoB (initial concentration of MitoB: 5.44 + 0.21 nmol g 2 1 of fish). The fish were then culled after 24 h, and aliquots of their liver were immediately flash frozen for subsequent extraction and quantification of the amounts of MitoB and MitoP [12]. The content in MitoP and MitoB was determined by high performance liquid chromatography-tandem mass spectrometry and used to estimate H 2 O 2 levels as the ratio of MitoP/MitoB.
Citrate synthase (CS) and cytochrome c oxidase (COX) activities were measured to determine liver mitochondrial density [16]. We analysed the link between the rSMR and MitoP/MitoB ratio using a general linear mixed model approach. The model included the MitoP/MitoB ratio as the dependent variable and rSMR, CS, COX and initial concentration of MitoB in the fish as continuous predictors, with batch as a random effect (see the electronic supplementary material for details of all assay protocols and statistical analyses).

Discussion
Measurement of ROS levels within living organisms make it possible to avoid potential biases introduced when using in vitro assays [17]. Development of the MitoB probe has enabled in vivo estimations of H 2 O 2 levels over a period of several hours in mice, Drosophila and Caenorhabditis elegans [12]. We now demonstrate that this method can also be used successfully in fish over a 24 h period. Our findings reveal for the first time that individuals with a high SMR have a lower level of H 2 O 2 . The negative relationship between SMR and the level of H 2 O 2 was independent of mitochondrial density estimates.
The MitoP/MitoB ratio represents the H 2 O 2 content within the mitochondria and reflects the balance between rsbl.royalsocietypublishing.org Biol. Lett. 11: 20150538 the H 2 O 2 generated by mitochondria during aerobic respiration and that scavenged by antioxidants such as mitochondrial glutathione peroxidase (GPx) [1,3]. Consequently, lower H 2 O 2 levels may be attributed to a lower rate of mitochondrial H 2 O 2 generation and/or greater antioxidant scavenging capacity [1,3]. The synthesis of GPx, an endogenous antioxidant, may be costly in term of resources [3,6]. Food intake in this experiment was limited and similar between individuals, so it is feasible that individuals with a lower H 2 O 2 level may have allocated more resources towards antioxidant defences. Alternatively, they may have had a lower rate of ROS production. H 2 O 2 arises in the mitochondria from the enzymatic conversion of superoxide anions that are produced by the respiratory chain. In order for the mitochondria to reduce oxygen to the superoxide anion, the mitochondrial respiratory chain must be in a highly reduced state [7,8]. Previous studies have shown that natural variation in mitochondrial function can have a significant influence on mitochondrial H 2 O 2 production, but also on the relationship between oxygen consumption and H 2 O 2 generation [7,8,18]. One key parameter capable of influencing such a relationship is the degree of mitochondrial uncoupling: higher uncoupling can lead to lower membrane potentials and greater rates of electron and oxygen flow in the respiratory chain (so making it less reduced) [7]. Uncoupling is known to simultaneously increase SMR and decrease ROS generation in vitro [7,19,20]. Our study is the first to report a negative relationship between oxygen consumption and H 2 O 2 levels in vivo, but measurements of H 2 O 2 production and H 2 O 2 scavenging are now required in order to understand the mechanisms underlying variability in H 2 O 2 levels.
The SMR of brown trout varies considerably among individuals of the same age and size within a population [13]; this level of interindividual variation in oxygen consumption is common across a broad range of taxa [21]. We demonstrate that a lower level of mitochondrial H 2 O 2 is associated with a higher respiration rate, which must increase the rate at which energy substrates are oxidized. A lower H 2 O 2 may therefore carry a cost in terms of a decrease in resources available for other traits such as body reserves or growth [2,3,6]. However, alleviation of the accumulation of oxidative damage may be a selective advantage for individual trout that have a higher rate of aerobic respiration and less H 2 O 2 [7,19,20]. If oxidative damage is an important contributor to ageing [1], then those individuals with a high SMR may benefit through slower ageing but at a cost of reduced resource allocation to other traits, although this requires further study [22].