Major Effect of Hydrogen Peroxide on Bacterioplankton Metabolism in the Northeast Atlantic

Reactive oxygen species such as hydrogen peroxide have the potential to alter metabolic rates of marine prokaryotes, ultimately impacting the cycling and bioavailability of nutrients and carbon. We studied the influence of H2O2 on prokaryotic heterotrophic production (PHP) and extracellular enzymatic activities (i.e., β-glucosidase [BGase], leucine aminopeptidase [LAPase] and alkaline phosphatase [APase]) in the subtropical Atlantic. With increasing concentrations of H2O2 in the range of 100–1000 nM, LAPase, APase and BGase were reduced by up to 11, 23 and 62%, respectively, in the different water layers. Incubation experiments with subsurface waters revealed a strong inhibition of all measured enzymatic activities upon H2O2 amendments in the range of 10–500 nM after 24 h. H2O2 additions also reduced prokaryotic heterotrophic production by 36–100% compared to the rapid increases in production rates occurring in the unamended controls. Our results indicate that oxidative stress caused by H2O2 affects prokaryotic growth and hydrolysis of specific components of the organic matter pool. Thus, we suggest that oxidative stress may have important consequences on marine carbon and energy fluxes.


Introduction
Reactive oxygen species such as hydrogen peroxide (H 2 O 2 ) are known to commonly cause oxidative stress in marine organisms (see [1] for review), and therefore, might have the potential to alter metabolic rates of marine prokaryotes, ultimately impacting the cycling and bioavailability of nutrients and organic carbon. H 2 O 2 commonly occurs in marine and freshwater environments where it can be found in concentrations as high as 10 mM [2], although in open oceans it usually varies between ,10-440 nM [3], with concentrations up to 1450 nM in Antarctic waters during autumn and winter periods [4]. Oceanic H 2 O 2 concentrations generally decrease with depth; with concentrations up to 15 nM and 6 nM at 100 m depth and in the bathypelagic zone, respectively [5]. In the subtropical North Atlantic (our study region), H 2 O 2 concentrations show a surface maximum of 75-220 nM, followed by a sharp decline at the shallow thermocline of ,40 m depth [6]. In deeper waters in this region H 2 O 2 concentrations are relatively stable, varying between 5-10 nM [6] which is in agreement with concentrations of up to 6 nM found at 150-5500 m in the subtropical North Pacific [5].
H 2 O 2 formation is traditionally attributed to photo-oxidation of chromophoric dissolved organic matter by molecular oxygen [7,8]. However, there is growing evidence that biological production could also be an important source of H 2 O 2 . Dark production of H 2 O 2 has been found in unfiltered water of the coastal Mediterranean, the Red Sea and the Baltic Sea [9] and in surface waters of the Sargasso Sea [10]. Moreover, overnight production of H 2 O 2 has been detected during in situ studies in the Atlantic [11,12] and the Pacific [13]. Yuan and Shiller [13] suggested that biological H 2 O 2 production might control the in situ H 2 O 2 concentrations measured in the Northwest Pacific. The detection of H 2 O 2 concentrations up to 6 nM at 5000 m depth was interpreted as an indication of its biological origin at depth since the short half-life of H 2 O 2 makes it highly unlikely that photochemically produced H 2 O 2 reaches this depth [5].
Recently, in situ H 2 O 2 concentrations were found to be close to a steady state between dark production and decay in samples from depths of $10 m, suggesting that the H 2 O 2 at those depths could be maintained primarily by particle-associated dark production and not by abiotic photochemical processes [14]. These authors suggested that extracellular H 2 O 2 could be a common byproduct of biological processes (as is intracellular H 2 O 2 ), or a product of a specific biochemical process, indicating that H 2 O 2 production rates might vary with environmental factors and species composition. Interestingly, it was recently hypothesized that abiotic processes such as photoxidation are able to produce peroxides (i.e., hydroperoxides associated with organic matter) that might inhibit bacterial activity by singlet oxygen transfer from phytodetritus to particle-attached bacteria [15]. The oxidative stress provoked by such particle-associated production of reactive oxygen species could be particularly relevant in microhabitats and in the deep ocean due to the preferential particle-related life strategy of dark ocean prokaryotes [16,17,18,19]. H 2 O 2 can readily diffuse across cytoplasmic membranes, influencing biological processes and damaging cellular constituents [20,21,22]. The damage includes formation of hydroxyl radicals inside the cell that might inhibit collagen gelation, modify amino acid residues of proteins and react with cell components, thereby producing organic peroxides which can attack DNA [23,24,25,26]. Lipid peroxides are formed when hydroxyl radicals interact with lipids, affecting the functioning of cell membranes, membrane-bound enzymes and other macromolecules [27]. Superoxide radicals and peroxides may also bind to DNA and alter or break the double helix structure [28].
Concentrations of ,100 nM H 2 O 2 influenced bacterioplankton by strongly reducing their abundance in microcosm experiments with water from the Gulf of Mexico [29], and causing oxidative stress in bacteria in Mediterranean waters [30]. In lake waters, the same H 2 O 2 concentration (100 nM H 2 O 2 ) caused a 40% reduction of the prokaryotic heterotrophic production in Lac Cromwell [22], and H 2 O 2 concentrations of 2000-3000 nM effectively inhibited prokaryotic production in Lake Fiolen [31]. Moreover, H 2 O 2 is an effective chemical algicide limiting cyanobacterial and green algal growth [32,33]. If H 2 O 2 influences bacterioplankton, then H 2 O 2 could indirectly affect the physicochemical characteristics of the environment (e.g., inorganic and organic nutrient quality and quantity, CO 2 fixation and respiration) concomitantly affecting microbial community composition.
Despite the ubiquitous presence of reactive oxygen species such as H 2 O 2 in the ocean and the crucial role played by marine prokaryotes in the marine biogeochemical cycles, the effect of the oxidative stress caused by reactive oxygen species on the metabolic rates of marine bacterioplankton and its ecological implications remain basically unknown. Here we studied the potential influence of H 2 O 2 on prokaryotic heterotrophic production and extracellular enzymatic activities (b-glucosidase, leucine aminopeptidase and alkaline phosphatase) in the water column of the subtropical Atlantic. Our results suggest that H 2 O 2 can have substantial effects on prokaryotic metabolism from the epipelagic to bathypelagic waters, which in turn might have profound implications on marine carbon cycling.

Ethics Statement
No specific permits were required for the described field studies. Sampling locations are not privately-owned or protected and sampling did not involve endangered or protected species.

Study Site and Experimental Setup
Two different sets of experiments were conducted with water collected in the subtropical northeast Atlantic Ocean during the MEDEA-I cruise with the RV Pelagia in October-November 2011 (Fig. 1). Seawater was collected from the bathypelagic in the core of the North Atlantic Deep Water (NADW, 2700-2800 m depth), the mesopelagic in the oxygen minimum layer (OML, 500-990 m depth) and the epipelagic at the base of the euphotic layer (100 m depth) using 25-L Niskin bottles mounted on a CTD (conductivitytemperature-depth) rosette sampler (Table S1).
In the first set of experiments, 100, 500 and 1000 nM of H 2 O 2 was added to seawater in duplicate 50 ml sterile conical tubes (Greiner Bio One). Blanks with no H 2 O 2 addition were simultaneously done. On these samples, triplicate measurements of extracellular enzymatic activity (EEA) were performed immediately after H 2 O 2 addition as explained below. These experi-ments were conducted at 9 stations (Stn. 6-16; see Fig. 1 (50,250, 500 nM at Stn. 18, and 10, 50, 100 nM at Stn. 20-21) to duplicate 1-L bottles containing water from the OML and 100 m depth. Amended samples were incubated together with duplicate unamended controls, in the dark at in situ temperature for 24 h. In this second set of experiments, EEAs were measured in triplicates at the beginning of the experiment (0 h) and after 24 h. Additionally, prokaryotic heterotrophic production (PHP) was measured at the start of the experiment, after 6 h and 24 h. The concentration of H 2 O 2 stock solutions was determined daily prior to adding H 2 O 2 to the samples using a spectrophotometer and the molar absorptivity of 38.161.4 M 21 cm 21 at 240 nm [34]. Measurements of in situ H 2 O 2 concentrations were attempted but failed due to an error in the preparation of the working solutions that were brought to the cruise.

Measurements of Prokaryotic Extracellular Enzymatic Activity (EEA)
The hydrolysis of the fluorogenic substrate analogs 4-methylcoumarinyl-7-amide (MCA)-L-leucine-7-amido-4-methylcoumarin, 4-methylumbelliferyl (MUF)-phosphate and MUF-b-Dglucoside was measured to estimate the potential hydrolytic activity of leucine aminopeptidase (LAPase), alkaline phosphatase (APase) and b-glucosidase (BGase), respectively [35]. All chemicals were obtained from Sigma. The procedure was followed as described previously [19,36]. Briefly, EEA was determined after substrate addition and incubation using a spectrofluorometer (Fluorolog-3) with a microwell plate reader (MicroMax 384, Horiba) at an excitation and emission wavelength of 365 and 445 nm, respectively. Samples (300 ml) were incubated in the dark at in situ temperature for 3-24 h. The linearity of the increase in fluorescence over time was checked on sets of samples incubated for 24 to 48 h, resulting in the same hydrolytic rates h -1 . Subsamples without substrate additions served as blanks to determine the background fluorescence of the samples. Previous experiments showed insignificant abiotic hydrolysis of the substrates [37,38,39]. The fluorescence obtained at the beginning and the end of the incubation was corrected for the corresponding blank. This increase in fluorescence over time was transformed into hydrolysis rates using standard curves established with different concentrations of the fluorochromes MUF and MCA added to 0.2 mm filtered sample water. A final substrate concentration of 10 mmol l -1 for BGase, 100 mmol l -1 for APase and 500 mmol l -1 for LAPase was used. These concentrations have been previously determined as saturating substrate concentrations [36], i.e., resulting in maximum hydrolysis rates. Consequently, the EEAs given throughout the paper represent potential hydrolysis rates. The substrates used in this study were previously shown to be unaffected by H 2 O 2 [40] thus, excluding the possibility of abiotic artifacts due to the added H 2 O 2 .

Prokaryotic Heterotrophic Production (PHP)
Bulk PHP was measured following the centrifugation method [41]. Triplicate 1 ml live-samples and TCA-killed blanks (5% final concentration) with 10 nM [ 3 H]-leucine (final concentration, specific activity 140 Ci mmol 21 ; Perkin Elmer) were incubated in temperature-controlled incubators in the dark at in situ temperature for 3-24 h. Incubation times were consistent for EEA and PHP incubations, and similar to those used in previous studies in the same region [18,19,36]. Incubations were terminated by adding TCA to a final concentration of 5% and live-samples and blanks were centrifuged at 12,000 g for 10 min. After aspirating the water, 1 ml of 5% TCA was added followed by a second round of centrifugation. Subsequently, the water was aspirated again, the vials dried and scintillation cocktail (1 ml Canberra-Packard Ultima-Gold) was added. After 18 h, the samples and blanks were counted in a liquid scintillation counter (Tri-Carb 3100TR, Perkin Elmer). The mean disintegrations per minute (DPM) of the TCA-fixed blanks were subtracted from the mean DPM of the respective samples, and the resulting DPM converted into leucine incorporation rates.

Effect of H 2 O 2 Concentration on EEA in the Northeast Atlantic Water Column
In the first set of H 2 O 2 enrichment experiments (i.e., 100, 500 and 1000 nM), LAPase exhibited the highest rates among the EEA measured followed by APase and BGase (Wilcoxon rank sum test, p,0.0001) ( Table 1 and Table S2    of experiments (i.e., 500 nM) initially reduced PHP by 70% and 97% in the OML and 100 m waters, respectively, and almost completely inhibited PHP at both depths after 24 h ( Fig. 2A, D). Interestingly, PHP increased after 24 h only in the unamended treatment (i.e., where H 2 O 2 was not added) in the 100 m waters experiments (see Fig. 3A, D, G).
In these experiments, the EEA inhibition pattern at time zero was similar to that found in the first set of experiments performed with waters collected at Stn. 6-16 (compare Fig. 3 and 4 to Table 1). Also in the second set of experiments, BGase was generally most strongly inhibited and LAPase was the least inhibited EEA. As for PHP, the inhibition of EEAs due to H 2 O 2 was stronger after 24 h than at the initial time in all the 100 m experiments (Wilcoxon rank sum test, p,0.001) (Fig. 3). This increased H 2 O 2 -induced inhibition after 24 h was not only found for BGase but also for APase and LAPase, the latter EEAs being reduced by 20-80% (Fig. 3C, F, I). In contrast, less inhibition was found in the OML experiments after 24 h than in the 100 m layer (Wilcoxon rank sum test, p,0.005). In fact, only in one out of the three OML experiments (St. 21; Fig. 4I), the inhibition of APase and LAPase (but not BGase) was stronger after 24 h (Wilcoxon rank sum test, p,0.005) (Fig. 4).

Discussion
In this study, a major inhibiting effect of H 2 O 2 on prokaryotic metabolism as deduced from measurements of EEA and PHP was found throughout the Northeast Atlantic water column. Few other studies have investigated the effect of H 2 O 2 on prokaryotes in aquatic environments, and to the best of our knowledge none have been done in open ocean waters. In our study region, H 2 O 2 concentrations range between 75-220 nM in surface waters but decrease with depth to 5-10 nM below 100 m depth [6], suggesting that the relative contribution of in situ H 2 O 2 should be very different between samples collected in the upper mixed layer and the ones from the deep ocean. Previous findings suggest that oxidative stress caused by H 2 O 2 may become important above the 100 nM level in surface seawater [29,30]. Our results suggest that H 2 O 2 concentrations lower than 100 nM still significantly retard prokaryotic metabolism in the subtropical North Atlantic water column reducing EEA and PHP by about 1-85% and 38-96%, respectively (Figs. 2, 3, 4).
We also found that the detrimental impact of H 2 O 2 on prokaryotic metabolism varies with time depending on the parameter investigated. For instance, although APase and LAPase were not strongly reduced by H 2 O 2 initially, they were greatly inhibited after 24 h of exposure to H 2 O 2 (Fig. 3). Recently, it was found that even short-term (,4 h) exposure to another reactive oxygen species (singlet oxygen) has profound effects on the prokaryotic community composition in a German lake [42]. This implies that reactive oxygen species can have a dual effect on the prokaryotic community: i) directly reducing the metabolic activity of the in situ community, and ii) altering the community structure and thus, indirectly affecting metabolic rates. We cannot confirm that a shift in community composition occurred in our experiments since we did not measure it.
It is also noteworthy that H 2 O 2 did not reduce all EEAs equally but preferentially inhibited BGase, thereby shaping the spectrum of EEAs. Although there are no previous studies investigating the effect of H 2 O 2 on EEA rates, the inhibition observed in the EEAs reflects the potential importance of reactive oxygen species such as H 2 O 2 in affecting the initial step in heterotrophic processing of polymeric organic matter in the ocean. This is critical since the activity of extracellular enzymes largely determines the supply of low molecular weight substrates for direct bacterial uptake [43]. The preferential inhibition of BGase over LAPase (Fig. 3) would result in an increase in the LAPase:BGase ratio. This can be interpreted as an enhanced degradation of protein over polysaccharides [44] which, in turn, ultimately affects the quality and quantity of the organic matter pool available to microbial communities.
Studies investigating the role of H 2 O 2 in lakes report contrasting results. Studying the effect of photodegradation of humic substances on prokaryotic metabolism in a Swedish lake, Anesio et al. [31] found that H 2 O 2 concentrations of about 2000-3000 nM were inhibitory for PHP. This range was similar to that formed during UV exposure of aquatic macrophyte leachates (4000-8000 nM H 2 O 2 ), suggested to inhibit growth of prokaryotes [45]. However, those H 2 O 2 levels are around one order of magnitude higher than reported by Xenopoulos and Bird [22] for a Canadian humic lake. These authors found that small amounts of added H 2 O 2 (,50 nM) inhibited PHP in that lake, and that 100 nM H 2 O 2 inhibited the PHP by as much as 40%. Although the photochemical production of low-molecular-weight substances from recalcitrant dissolved organic matter can stimulate prokaryotic activity and growth [46,47], photochemical reactions are known to inhibit prokaryotic activities [48,49,50] by the formation of inhibitory substances such as H 2 O 2 [51,52,53], but also by polymerization and condensation of labile compounds [54,55] and direct mineralization of dissolved organic matter [56,57]. Thus, these complementary effects of the photolysis of dissolved organic matter on prokaryotic activity might mask the direct effect of H 2 O 2 on prokaryotic metabolic rates in surface waters. This could potentially explain why in those photochemical studies where H 2 O 2 concentrations were raised by exposing samples to UV radiation, the inhibitory H 2 O 2 concentrations were higher compared to experiments where H 2 O 2 was directly added without exposure to UV radiation. H 2 O 2 concentrations have been shown to follow a diel periodicity in oceanic surface waters in our study region, ranging between ,50-220 nM with increasing concentrations until the mid afternoon [6]. According to our data and assuming that surface water prokaryotes would behave similarly to those at 100 m depth, the range in H 2 O 2 concentrations in surface waters could provoke a considerable PHP inhibition. Although speculative, our results could then suggest the possibility that the periodicity of PHP observed in the surface waters of the ocean (e.g. [58]), commonly interpreted as UV-induced [59], could be due, at least partly, to the accumulation of H 2 O 2 in surface waters during daytime.
Taken together, our results suggest a potentially relevant role of oxidative stress affecting bacterioplankton metabolism, reducing prokaryotic growth and the hydrolysis of specific components of the organic matter pool. Reactive oxygen species such as H 2 O 2 can significantly reduce the amount of organic carbon being channeled through prokaryotes, as well as alter the bioavailability of organic carbon and nutrients via the reduction of PHP and EEA rates and shifts in the EEA spectrum. This influence of oxidative stress could be potentially important throughout the entire oceanic water column where the dark biological production of H 2 O 2 has been recently shown to be significant [14]. These effects would be even more crucial in specific (micro)environments where elevated concentrations of biologically-produced reactive oxygen species can be expected, like in particle-attached communities. The recent finding of considerable concentrations of hydroperoxides in sinking particles collected in the Arctic Ocean [15] supports this idea. Due to the assumed preferential particle-associated way of life of dark ocean prokaryotes [16,17,18,19], the impact of oxidative stress on this community should be particularly significant. Taking into account the paramount role of openocean prokaryotes in the marine biogeochemical cycles, reactive oxygen species such as H 2 O 2 can have relevant repercussions in marine carbon fluxes.