David J Kieber Barrie M Peake and Norman M Scully

Table of contents

Abstract 253

8.1 Introduction 253

8.1.1 Reaction kinetics 255

8.2 Formation and removal of ROS 256

8.2.1 Hydroxyl radical 256

8.2.2 Singlet oxygen 258

8.2.3 Superoxide radical 259

8.2.4 Hydrogen peroxide 261

8.2.4.1 Sources of H202 263

8.2.4.2 H202 removal pathways 267

8.2.4.3 Reactions of H202 with DOM 269

8.3 Other ROS 271

8.4 Impact of ROS on aquatic organisms 273

8.5 Conclusions 275

Acknowledgements 276

References 276

Abstract

The absorption of solar energy by dissolved organic matter (DOM) in natural waters results in a variety of photochemical transformations involving oxygen as a major reactant. These photochemical transformations generate a suite of reactive oxygen species (ROS) including the superoxide anion, the hydroxyl radical, singlet oxygen, alkoxy and peroxy radicals, the carbonate radical, and hydrogen peroxide. ROS cause numerous biogeochemical changes in aquatic ecosystems affecting the cycling of DOM, biological processes, and trace metal speciation.

In this chapter, we present a synthesis of research conducted in the study of ROS in marine and fresh waters, including a detailed discussion of recent evidence regarding the formation and removal of these compounds in the photic zone. Salient findings of this review are: (1) direct photolysis of DOM and reactions of DOM with trace metals and dissolved oxygen are one of the main sources of ROS in aquatic environments; (2) solar action spectra for ROS indicate that the highest production rates are between 290-400 nm; (3) removal pathways for most ROS are poorly known, but are likely to occur through reactions with DOM. Hydrogen peroxide is relatively unreactive towards DOM; however, it can affect DOM indirectly through photo-Fenton reactions; and (4) the impact of externally-generated ROS on aquatic organisms should be a function of their permeability through the cell membrane and inversely related to their reactivity in the water outside the cell.

8.1 Introduction

DOM absorbs nearly all of the ultraviolet radiation (UVR, 280-400 nm) in natural waters thereby controlling the attenuation of the UVR in the water column ([1-4] and Chapters 3, 6). From this perspective, DOM plays a fundamental role in controlling the ecology in the photic zone [5]. Absorption of UVR, in turn, provides the energy to break down and remineralize the DOM. It has been known for some time that when natural water samples are exposed to solar radiation their optical transparency increases due to the loss of chromo-phores in the organic matter [6,7]. In the case of organic-rich natural waters, it is easy to observe an increase in optical clarity and loss of the yellow-brown color of the water when it is exposed to solar radiation in the presence of dissolved oxygen. However, very little or no loss in color is observed when oxygen is removed prior to irradiation [8], which points to the important role that oxygen plays in the photooxidation of organic matter in the photic zone.

The photolysis of organic and inorganic constituents in natural waters is also an important mechanism for the production of free radicals [9]. Zafiriou and Dister [10] determined that the total production rate of radicals varied from 6 to 24 x 10-8 M h_1 during the summer in the Atlantic Ocean along the East Coast of the United States, employing a solar simulator with a spectral output and photon flux approximately equal to the noon time solar irradiance on a clear summer day in the Gulf of Maine. These rates translate to a very large, mean radical flux of approximately 0.7 mmol m~2 day-1 in surface waters during the summer.

The loss of yellow-brown color (and UV absorbance) in the water and free radical production are undoubtedly due to interactions of organic matter with solar UVR and visible radiation, which lead to a series of photochemical transformations involving oxygen as a major reactant (Figure 1). The importance of oxygen in the photooxidation of organic matter in natural waters is clearly evident in oxygen consumption studies, all of which show a substantial loss of dissolved oxygen when filter-sterilized natural water samples are exposed to solar radiation [11-16]. Oxygen plays a pivotal role as the initial scavenger of radicals that are produced during the irradiation of natural waters, forming an "oxygen wall" [9] and generating alkoxy and peroxy radicals (RO' and ROO', respectively) that eventually decay to stable oxygenated species. Some of the energy absorbed by DOM is dissipated through energy transfer reactions that yield singlet oxygen (^2), and electron transfer reactions with DOM that yield the superoxide radical (02~), a fraction of which disproportionates to form H202 [17]. These oxygen-containing compounds, together with the highly reactive hydroxyl radical (*OH), are collectively referred to as ROS.

Many of the UVR effects that have been observed in aquatic organisms and ecosystems occur either directly or indirectly from the production and subsequent reactions of ROS. It is therefore not surprising that many ROS have been intensively studied to understand their impact on chemical and biological processes in natural waters. For example, reactions of ROS can increase the nor-

Enzym a tic Decay

Figure 1 Schematic summary of the sources and removal pathways of ROS in natural waters including singlet oxygen (102), superoxide (02"), hydrogen peroxide (H202) and the hydroxyl radical ('OH). The main ROS are indicated by squares. Notation: FW, freshwater; SW, seawater; Me"+ or Me(n"1)+, metal in the n+ or (n-l) + oxidation state; NOx", the nitrate or nitrite anion; and ?, unknown pathway.

Enzym a tic Decay

Figure 1 Schematic summary of the sources and removal pathways of ROS in natural waters including singlet oxygen (102), superoxide (02"), hydrogen peroxide (H202) and the hydroxyl radical ('OH). The main ROS are indicated by squares. Notation: FW, freshwater; SW, seawater; Me"+ or Me(n"1)+, metal in the n+ or (n-l) + oxidation state; NOx", the nitrate or nitrite anion; and ?, unknown pathway.

mally slow rate of oxidation of some organic compounds in natural waters [18-24], change the redox state and speciation of trace metals [25-32], and cause oxidative stress to aquatic organisms [33,34]. Additionally, since ROS are involved in the photooxidation of DOM (Figure 1), they can influence the cycling of important biogeochemical elements in natural waters, as reviewed in Chapter 5 and elsewhere [6,7,35,36].

While the reaction of excited state organic matter with oxygen is the main source of many ROS in the photic zone, processes responsible for the removal of some ROS are less clear, especially when they involve DOM (Figure 1). In this chapter, we will discuss some of the processes that result in the formation and loss of ROS in marine and fresh water ecosystems, and the potential effects of externally generated ROS (i.e., photoproduced outside the cell in the surrounding water) on aquatic organisms. Intracellular production of ROS and the resultant oxidative stress that they impose on aquatic organisms is beyond the scope of this chapter and will not be discussed (see recent reviews by Josephy [37] and Vincent and Neale [34]). Analytical methods to detect these species will also not be discussed since they have been critically reviewed elsewhere [4,9,38],

8.1.1 Reaction kinetics

The concentration of a ROS measured after a given time t, [ROS]t, is equal to the sum of the initial concentration (i.e., at t = 0) and the concentration photo-produced over a given period of time less the concentration that has reacted (decayed) during that time:

[ROS]t= [ROS]t= o+ [ROS]t,production- [ROS],decay (1)

In the laboratory, experimental variables such as photon flux are controlled ensuring that ROS precursors are not appreciably depleted. In this case, [ROS]t rapidly attains a constant value called the steady state concentration corresponding to constant rates of formation and decay of the ROS. Invariably, this situation does not hold in natural aqueous environments involving solar irradiation because irradiance levels and biological processes (which may be a source of precursors or a sink of the ROS) will vary spatially and temporally. These variations will cause concentrations of ROS to undergo diurnal changes, as observed for H202 (vide infra). If the factors that control the rates of production and loss are known for a ROS, then spatial and temporal variations in their levels can be modeled [39].

More often than not, only the net rate of production (R) is measured, which is simply the capacity of a water sample to generate a specific ROS. This net rate (or accumulation rate) is the concentration of reactive oxygen species ([ROS]) produced during a given time (Ai) or more appropriately, photon exposure:

The rate of a photochemical reaction is the product of the probability that an incident photon is absorbed and the probability that the absorbed photon will bring about a reaction. These probabilities are measured by the wavelength-dependent absorbance coefficient (a^) (see Chapters 5 and 6) of the DOM and the quantum yield (O^), respectively. The quantum yield is the efficiency of a photochemical process, and is equal to the number of moles of species formed or photolyzed divided by the number of moles of photons (Einsteins) absorbed by the chromophore. However, the complex molecular composition of DOM in aquatic environments means that there are likely to be multiple electronic transition energies and multiple precursors involved in the formation of individual ROS in natural waters, and these precursors are generally not known. Therefore, the quantum yield for the photochemical formation of ROS in natural waters is defined in terms of the DOM absorbance and is referred to as an apparent quantum yield, which is invariably wavelength-dependent. To express this wavelength dependence, the product (e^) of the apparent quantum yield and the absorbance coefficient (a;) is plotted as a function of wavelength to yield a chemical action spectrum [40]. The formulation of an action spectrum is an important component of photochemical models, but it can be problematic due to uncertainties in DOM absorbance measurements and the assumption that wavelength-dependent apparent quantum yields are constant with photon exposure (photon exposure is the irradiance integrated over time of exposure) when in fact they can increase or decrease [40].

Apart from the inherent efficiency of the reactions leading to the light-induced formation of a ROS as summarized by the relevant apparent quantum yield and action spectrum, the observed rate of production will depend on other factors that affect the photon exposure including water column composition and depth (Chapter 3), time of day (i.e., solar zenith angle), season, latitude (Chapter 2), and physical transport processes (Chapter 4). For more details regarding the fundamental equations used to define the rates of primary and secondary photochemical reactions and their application to aquatic systems, the reader is referred to recent reviews on this topic [41,42].

8.2 Formation and removal of ROS

8.2.1 Hydroxy I radical

The hydroxyl radical (*OH) is perhaps the most important ROS detected in natural waters. It plays a central role in transformations of organic matter in the troposphere [43], but its biogeochemical role in natural waters is poorly understood. The main source of the 'OH radical in most natural waters is from the photolysis of DOM, nitrate and nitrite [44,45], with production rates in the low to high 10~9 M h_1 range [44,46,47]. Iron and H202 can also be an important source of the 'OH radical through photo-Fenton chemistry, although this will be largely limited to iron-rich, high H202 environments such as the Suwannee or Orinoco Rivers [48,49]. The production of the 'OH radical from the photolysis of DOM is quite surprising, as there are very few known sources for the 'OH radical reported in the basic chemical literature that involve specific organic compounds. Vaughan and Blough [45] have shown that the production of the 'OH radical in a Suwannee River fulvic acid isolate occurs directly from the photolysis of DOM through an oxygen-independent pathway. They also observed that the formation of the *OH radical occurred in the same wavelength range as the absorption band of benzoquinone. These results, along with evidence that quinones photoproduce 'OH (for review see [50]) and are a component of DOM [51,52], suggest that quinones are a potential source of the 'OH radical in natural waters.

Once formed, the 'OH radical is extremely unstable and it reacts rather indiscriminately with many organic or inorganic species at rates that are at or near the diffusion limit, either through an addition or H-atom abstraction pathway. As a result of its extreme reactivity, day time concentrations of the *OH radical are very low in surface waters, with estimates ranging from 10_19tol0~17 M [44,53]. While the 'OH radical is generally very reactive, as indicated by the extensive number of rate constants at or near the diffusion limit (108—10101 mol-1 s~') [54], there are some notable exceptions of species that react relatively slowly with the 'OH radical (e.g., borate, carbon dioxide, phosphate) [54]. Likewise, while rate constants are large in many cases, there are differences in reactivity that lead to the selective loss of the 'OH radical in natural waters through reaction with only a few reactants. For example, in seawater, the 'OH radical is primarily removed through its reaction with the bromide ion, while in fresh waters with high alkalinity the bicarbonate and carbonate ions are the principal reactants and DOM predominates in low alkalinity waters [41]. These reactions result in the formation of less reactive dibromide and carbonate radicals (Figure 1) whose fates in natural waters are still unknown.

While only a few reactants are expected to control the loss of the 'OH radical in the photic zone, this does not preclude the possibility that the 'OH radical can affect the cycling of other minor species, especially those that strongly bind to DOM. Currently, the role of the 'OH radical in the transformations and cycling of DOM is poorly understood. If an organic species is present at trace levels (ca. < 1 x 10-6 M) in natural waters, then kinetic calculations suggest that the 'OH radical will not effectively remove that compound from the dissolved phase unless it is long-lived and not otherwise reactive (i.e., it is not biologically or chemically reactive). Estimates of half-lives for most trace organic compounds based on their reaction with the 'OH radical are quite long in aquatic environments located at mid latitudes (approximately 20-210 days) [46]. Shorter half-lives (ca. 7-60 days) are predicted in nitrate-rich systems, which contain higher steady state concentrations of the 'OH radical compared to nitrate-depleted waters [19]. However, the reactivity of the 'OH radical towards an organic substrate may be much higher if that compound has a heterogeneous distribution in solution due to its binding to DOM, especially if that binding is at or near the site involved in the photochemical production of the 'OH radical.

The importance of binding was demonstrated for the photolysis of the hydrophobic chlorocarbon pesticide mirex, which is predominantly bound to DOM in natural waters [55,56]. When humic acid solutions of mirex were irradiated, the mirex was photo-reduced by a humic-generated hydrated electron, which would not have occurred if mirex did not bind strongly to the DOM through hydrophobic interactions. Evidence also indicates that electrostatic interactions can affect the binding of a species to DOM, thereby affecting its photochemical reactivity. Blough [57] demonstrated that the decrease in the EPR (Electron Paramagnetic Resonance) signal during the photolysis of a series of water-soluble nitroxides in humic acid solutions was fastest for the cationic nitroxide probe (i.e., higher rate of signal loss) compared to the neutral or anionic nitroxide probes. This trend in reactivity paralleled the degree of interaction with DOM, which increased from the anionic nitroxide to the cationic nitroxide. These results indicate that it should be possible to predict the reactivity of a species with the DOM-generated 'OH radical, or other ROS, based on property-reactivity analysis employing physicochemical parameters such as the octanol water partition coefficient (Kow) [58].

8.2.2 Single t oxygen

Dissolved oxygen is present in natural waters at relatively high concentrations, generally ranging between 2.0 and 3.0 x 10~4 M. It exists primarily in the ground state triplet (denoted by 302), with the two highest energy n electrons occupying separate molecular orbitals and having parallel spins. However, in the presence of UVR and appropriate photosensitizers, ground state oxygen is easily converted into its lowest excited singlet state OO2) through energy transfer reactions. These reactions are often quite facile because the lowest energy level of (specifically the 'Ag species) is only 94 kJ mol-1 above the triplet ground state. One of several detailed descriptions of the physical and chemical properties of l02 is given in Kearns [59].

There is considerable interest in the role of ^2 in the oxidation of DOM in natural waters because '02 is reactive towards a wide range of electron-rich organic compound classes such as alkenes, sulfides and phenols (Table 1). Due to its potential importance in natural waters, structure-activity models have been developed to predict reaction rate constants for the !02 oxidation of a series of environmentally-relevant, substituted phenols [22]. The reader is also referred to Wilkinson et al. [60] for an extensive compilation of quenching and reaction rate constants for reactions involving ^2, albeit most of the rate data are reported for solutions involving organic solvents. Singlet oxygen is also known to exert an oxidative stress in cellular systems causing toxicological effects such as lipid peroxidation and DNA damage [61].

In natural waters, '02 is a ubiquitous ROS in the photic zone, with midday surface concentrations ranging from 10~ 15-10~12 M, depending primarily on the concentration of DOM [62-64]. The primary source of '02 in natural waters is through energy transfer reactions involving excited state triplet DOM [63], with production rates ranging from 10~9 to 10~7 Ms-1 [62,63]. In humic isolates and natural water samples, quantum yields for '02 production (ca. 0.005-0.03) decrease with increasing wavelength in the UVR [8,65] a trend that has also been observed for many other photoproduced species in natural waters (for reviews

Table 1. Rate constants for reaction of l02 with selected organic compounds in aqueous solution at or near room temperature.3 Combined quenching and reaction rate constants are denoted by*.

Table 1. Rate constants for reaction of l02 with selected organic compounds in aqueous solution at or near room temperature.3 Combined quenching and reaction rate constants are denoted by*.

Compound

pH

kt (1 mol-'s-')

Reference

Olefins (crocin, bixin)

7.8

~108

[169]

1,4-Naphthalenedipropionate ion

NA

1.4 x 106

[170]

Hydroquinone

NA

2.5 x 107

[171]

2-Hydroxybenzoate

7.5

2.5 x 105

[172]

Phenol

8.0

1.0 x 106

[173]

4-Chlorophenol

NA

6.0 x 106

[22]

2,5-Dimethylfuran

7.0

8.2 x 108

[174]

Indole

7.2

7.0 x 107*

[175]

Alanine

8.4

2.0 x 106

[176]

Arginine

7.1

<1.0 x 106

[177]

Cytochrome B

7.4

1.4 x 108*

[178]

Superoxide dimutase

7.1

2.7 x 109

[179]

Histidine

7.0

9.0 x 107

[180]

Tryptophan

7.4

6.6 x 107

[181]

Diethyl sulfide

7.0

2.0 x 107

[174]

Methionine

7.0

2.1 x 107

[180]

Ascorbate

6.8

8.3 x 106*

[182]

DNA

7.0

5.1 x 105*

[183]

aNA: Solution pH not reported.

aNA: Solution pH not reported.

see [41,42]). Singlet oxygen rapidly decays back to ground state triplet oxygen almost entirely through physical quenching by water. This process effectively removes nearly all of the !0? that is formed and limits its lifetime to approximately 4 /is in water [66]. Ground state DOM can reduce the lifetime of !02 even further via quenching, but only at extremely high DOM concentrations [64],

Because water is so effective in removing '02, singlet oxygen is not expected to affect the concentrations of most organic or inorganic compounds in natural waters [67], even though reaction rate constants for these reactions can be large (Table 1). Only a few compounds have been shown to react with '02 at appreciable rates [68]. Singlet oxygen should not be important in the removal of most organic compounds from natural waters because they are present at trace levels (< 1 x 10-6 M), and will not effectively compete for '02 relative to its rate of physical quenching. For example, the reaction of *02 with dimethyl sulfide (DMS) is extremely slow in coastal and oligotrophic seawater (ca. 10-16 M s"1) yielding a turnover time of approximately 10 years, based on ambient concentrations of reactants (10"9 M DMS, 10"14 M ^2) and a rate constant of 2.0 x 1071 mol-1 s-1 for diethyl sulfide (Table 1). This turnover time is orders of magnitude longer than observed DMS turnover times (e.g. 1-5 days in Equatorial Pacific surface waters [69]).

8.2.3 Superoxide radical

The superoxide radical (02~)is one of the main ROS formed in sunlit natural waters, and it undoubtedly plays a key role in trace metal cycling and DOM

transformations in the water column. Early studies by Petasne and Zika [17] identified possible mechanisms for the production of 02~ in natural waters and revealed the need for research on the biogeochemical effects of 02~ in marine and fresh waters. The primary source of 02~ in natural waters is through the photolysis of DOM and the subsequent reduction of ground triplet state oxygen. Blough and Zepp [4], in their review of ROS in natural waters, outlined two mechanisms that may be responsible for 02~ production from DOM photolysis. One mechanism involves the direct transfer of an electron from an excited triplet state chromophore such as DOM (3DOM*) to ground triplet state oxygen [4]:

The other proposed mechanism involves the reduction of ground triplet state oxygen by an aqueous solvated electron generated by excited triplet state DOM [70-72]:

Recently, Thomas-Smith and Blough [73] determined that quantum yields for the production of the aqueous solvated electron in irradiated coastal DOM samples were too low to account for the production of 02~ (as determined from the yield of H202), suggesting that reaction (3) is the main source of 02~ in natural waters.

Typical production rates of02~ in seawater range from 2 x 10~12Ms_I in the open ocean to 2x 10~10 M s_1 in coastal water [74]. Once 02~ is formed, approximately 50 — 80% disproportionates to H202 with a second-order rate constant of 2.2 x 104 1 mol-1 s_1 measured in oligotrophic seawater at pH 8.3 [75]. The remaining 20-50% of 02~ that does not disproportionate to H202 may be removed through reactions with trace metals or DOM. Superoxide interacts with DOM and a number of environmentally important trace metals in both marine and fresh waters [28,30,31,76-78]. The presence of 02~, for example, can result in a significant accumulation of reduced iron (Fe(II)) in surface waters [28], which in turn may initiate further chemical transformations (vide infra). The reduction of iron by 02~ has biogeochemical implications through increasing the bioavailability of iron, particularly in some marine environments where iron limits phytoplankton growth [79,80].

It has recently been shown that organic Cu-complexes increase the decay rate of 02~ in natural waters. Voelker et al. [30] found that organically complexed copper significantly lowered steady state 02_ concentrations in marine waters. Goldstone and Voelker [78] also demonstrated that DOM contains a non-metallic, non-enzymatic fraction that can catalyze superoxide dismutation. When copper-DOM reactions are considered, estimated steady state concentrations of 02~ in coastal waters are 100- to 1000-fold lower than predicted concentrations, which only consider its decay through bimolecular dismutation [78]. Thus, the photochemical redox cycling of DOM via 02~ reactions may represent an essential step in the alteration of the optical and biochemical properties of DOM [78].

DOM is also likely to react with 02~ at appreciable rates to form oxygenated species. However, this will depend on the pH, which regulates the relative amount of the perhydroxyl radical, H02* and its conjugate base (02~) [81]. Generally, the perhydroxyl radical is much more reactive than 02~. For example, amino acids are approximately two orders of magnitude more reactive towards H02* than 02~ [82]. Since the pKa for the H02'/02~ equilibrium is 4.8 in pure water and 4.6 in seawater [75], the predominant species in most natural waters will be the less reactive superoxide anion. Nevertheless, 02~ is moderately reactive towards some inorganic and organic species (Table 2), but further studies are needed to determine possible reaction mechanisms and to identify specific organic molecules responsible for superoxide-DOM interactions in natural waters.

Table 2. Rate constants for the reaction of different organic compounds with 02 in aqueous solution at or near room temperature.3

Compound

pH

k( 1 mol's1)

Reference

Cysteine

7.0

> 5 X 104

[184]

L-Glutamic acid

8.7

<0.39

[185]

Glycine

8.8

<0.42

[185]

Glutathione

7.8

6.7 X 105

[186]

2-Oxoglutarate ion

10.1

<0.30

[187]

Carbonate radical ion

NA

4 X 108

[159]

Nitrite ion

NA

5 X 106

[188]

1,4-Benzoquinone

~7

9 X 108

[189]

3,4-Dihydroxybenzaldehyde

7.0

1.40 X 107

[190]

2,3-Dimethyl- 1,4-benzoquinone

7.0

4.5 X 108

[191]

Ethylene

NA

2 X 105

[192]

Ethylenediaminetetraaccetate ion

9.9

<0.01

[187]

Formate ion

10.1

<0.01

[187]

Linoleate ion

11-12

0.01-0.1

[81]

L-Malate ion

10.1

<0.11

[187]

D,L-Methionine

8.3

<0.33

[185]

2-Naphthylamine

NA

1.3 X 107

[194]

Peroxidase (horseradish)

5.5

-2.5 X 108

[195]

Superoxide dismutase

6.5-9.0

0.01-5.4 X 109

[82,193]

aNA: Solution pH not reported.

aNA: Solution pH not reported.

8.2.4 Hydrogen peroxide

Hydrogen peroxide is produced in all natural waters and it is one of the major products formed from the photolysis of DOM. Since Van Baalen and Marler [83] first detected H202 in the Gulf of Mexico, it has been intensively studied by numerous investigators because of its high concentrations relative to other ROS and due to its potential chemical and biological reactivity. Hydrogen peroxide can oxidize DOM through transformations involving the photo-Fenton reaction [84] and it can alfect the redox chemistry of trace metals such as iron, copper and manganese [85-87], making H202 an important chemical reactant in the aquatic environment. Hydrogen peroxide is also a known cellular oxidant [88,89] and organisms devote considerable metabolic energy to remove this ROS [90].

Numerous studies have been conducted to quantify the temporal and spatial variations in H202 concentrations in both fresh water and seawater, and to assess the factors that affect its production and loss in the photic zone [83,91-99]. Variations in production and removal rates for H202 in the photic zone yield daytime H202 concentrations in the 10~8 M range for a variety of natural waters (Table 3). Surface water concentrations in lakes and rivers vary from approximately 5.0 x 10~8 M in clear oligotrophic systems to nearly 5.0 x 10-7 M in humic-rich lakes [92,94,100-102]. The relatively low concentrations of DOM in most marine systems give rise to H202 concentrations that rarely exceed 2.0 x 10~7 M [83,91,96,97,103,104]. Temporal changes in production and decay rates lead to diurnal variations in the net (observed) H202 concentration in the photic zone. In a typical diurnal cycle, the concentration of H202 increases in surface waters after sunrise until it reaches a maximum in the early afternoon between 1200 to 1400 h (local time). As photochemical production rates decrease during the remainder of the day and cease in the evening, the net H202 concentration decreases due to its biological delay reaching a minimum in the early morning before sunrise [91,92,96,97,100,103,105-108]. These diurnal changes in H202 concentration closely match changes in solar radiation intensity throughout the day and are especially apparent when the H202 concentration is calculated on an areal basis (mg m~2). Both gradual and rapid changes in the solar irradiance (e.g., intermittent cloud cover) caused similar changes in the areal concentrations of H202 during diurnal studies conducted in a variety of Canadian shield lakes [100,108],

The half-life for the biological consumption of H202 in freshwater ecosystems ranges from a few hours for eutrophic and dystrophic systems to greater than 24 hours for oligotrophic waters [94,100,109]. In marine systems, the half-life for biological consumption of H202 can be much longer, ranging from 10 hours in coastal waters to 15 days in Antarctic seawater [99,110]. When the half-life for the biological consumption of H202 is relatively long (e.g., > 24 h), then vertical mixing will heavily influence its depth distribution of H202 in natural waters [39,96,103]. Depending on the intensity of vertical mixing, the residence time (i.e., the time scale for complete turnover) of H202 in the surface mixed layer can vary from minutes in large turbulent lakes to several hours in small, thermally stratified humic-rich lakes [101,102].

On a global scale, H202 concentrations generally decrease with increasing latitude in oligotrophic waters (ca. 1.0-2.0 x 10~7 M in subtropical regions to 3.0 x 10~8 M in polar regions) [99,111]. This trend can largely be explained in terms of latitudinal gradients in temperature and UVR, both of which decrease with increasing latitude. Temperature inversely affects rates, in part, because apparent quantum yields for the photoproduction of H202 decrease nearly two-fold per 10°C decrease in temperature [99,112], This temperature depend-

Location

Table 3. Daytime surface concentrations of H202 in natural waters. H,02 ( x 10 9 M) Reference

Polar Ocean

Antarctic Peninsula Paradise Harbor Weddell Sea

Open Ocean

Bermuda Equatorial Pacific Sargasso Sea

Coastal Ocean

Gulf of Mexico 90-240 [91,123]

Mediterranean Sea 90-130 [103]

Wadden Sea, intertidal zone 1500-4500 [197]

Caribbean Sea 140-470 [93]

Peru upwelling zone 8-50 [233]

Seto Inland Sea 10-400 [107]

Port Aransas, Open Gulf 97-161 [83]

BiscayneBay 150-275 [110]

Florida Current 55-65 [133]

Laurentian Great Lakes

Lake Erie Lake Ontario

66-220 38-122

Oligotrophic Lakes

Canadian Shield

152-210

Dystrophic Lakes

Canadian Shield Sub-Arctic Canadian Shield

121-444 104-620

Rivers

Patuxent Orinoco Shark

St. Lawrence

12-350 20-640 32-139 69-136

Geothermal waters

Yellowstone National Park

20-490

ent decrease in production is partly compensated by a concomitant decrease in biological H202 decay rates at low temperatures (vide infra).

8.2.4.1 Sources of H202

The main source of hydrogen peroxide in natural waters is through its photochemical production [113]. However, as will be discussed below, other processes also affect concentrations of H202 in the water column. These include dry and wet atmospheric deposition, and biological release.

Photochemical production. The photochemical production of H202 in natural waters occurs through photoreactions initiated by absorption of ultraviolet and visible radiation by DOM [4,94]. Excited state DOM reduces dissolved oxygen to form 02_, which then disproportionates to H202 (Figure 1) [17,93,94,114]. Apparent quantum yields for the photochemical formation of H202 are remarkably similar in diverse marine [41,93,104,112] and fresh [41,98,115] waters. With few exceptions, apparent quantum yields decrease exponentially from approximately 10~3 at 290 nm to 10-4 at 400 nm [112]. Based on apparent quantum yield measurements in marine waters, Yocis et al. [104] and Miller [112] showed that the production of H202 was primarily due to absorption of UV-B (280-320 nm) and UV-A (320-400 nm) radiation by DOM. However, a small fraction of the total production (ca. 10-20%) occurred at wavelengths greater than 400 nm in coaastal and oligotrophic seawater [112]. Since H202 photoproduction is partly UV-B dependent, increases in UV-B radiation through stratospheric ozone thinning should affect H202 production rates. In Antarctic waters, increased UV-B resulting from ozone depletion increased predicted H202 production rates in surface waters from 20-50% [104]. Similarly, results of a modeling study by Scully et al. [116] showed that, under enhanced UV-B conditions, the relative increase in H202 production rates was greater in low DOM, optically clear waters compared to high DOM lake waters.

In addition to apparent quantum yield measurements, studies have been conducted to determine midday photochemical production rates of H202 in fresh water [92,94,95,98,115] and marine environments [93,103,104,117]. These studies indicate that H202 production rates are dependent on the solar irra-diance, temperature (vide supra) and DOM concentrations. Cooper et al. [115] determined that H202 production rates varied significantly (0-74 x 10-7 M h_ *) when a variety of ground water samples were exposed to 6 h of solar radiation. Ground water production rates increased non-linearly with increasing dissolved organic carbon (DOC) content from 0.22 to 17.8 mg C l-1. A similar trend was observed by Scully et al. [98] who found a significant, non-linear correlation between production rates and DOC content in a series of Canadian Shield lakes. Their data were fitted to a power function, and it was shown that when production rates were normalized to the concentration of DOC, they were not constant, as would be expected from a simple linear fit, but rather they increased with increasing DOC content. This result is not surprising since the chromophoric fraction of DOC in lake waters has been demonstrated to increase with increasing DOC concentration [3],

Atmospheric input. In the atmosphere, the gas phase is an important reservoir of H202 [43,118], with mixing ratios in the marine boundary layer ranging from 0.1 to 5 ppb [119,120]. At these levels, natural waters are undersaturated with respect to H202. Therefore H202 is expected to undergo a net diffusion from the atmosphere into the surface layer of the water body. For example, the diffusion of ppb levels of H202 into oligotrophic seawater results in a H202 flux at the ocean's surface (1 m2 surface area by 1 m deep) of 0.09-4.0 x 10~9 M h-1, based on a transfer velocity of 0.6 cms-1 [121]. This air-to-sea flux is significant at the sea surface, corresponding to approximately 1-40% of summertime photo chemical production rates. However, this flux will be insignificant deeper in the water column due to the dilution of air-derived H202 in seawater. The air-to-sea flux of H202 does not take into account turbulent mixing and wave action, and is most likely an underestimation of the true flux. Additionally, fluxes will vary spatially and temporally due to changes in gas phase H202 concentrations, which are controlled by physical and photochemical processes in the troposphere.

Wet deposition can also be a significant source of H202 in surface waters [96,122-126]. The concentration of H202 in rain is typically 1.0 x 10"5 M [122], which is one to three orders of magnitude higher than surface concentrations in fresh water or seawater. Therefore, rain events can rapidly increase H202 concentrations in the water column. During a rainstorm in the Gulf of Mexico, Cooper et al. [123] reported that the in situ concentration of H202 increased from 8.5 to 18.5 x 10~8 M at a depth of 1 m over a period of 127 min. In another rain event in the Gulf of Mexico, the concentration at 1 m increased from 8.6 to 20.3 x 10~8 M in 25 min. Marine rain events can also increase H202 concentrations in the mixed layer down to 50 m or more [123]. Miller and Kester [96] and Kieber et al. [126] noted a 50-200% increase in surface H202 concentrations due to rain inputs into marine waters near Bermuda in the Sargasso Sea.

Other wet deposition sources include snow, melting glaciers, and run-off from snowmelt and sea ice. Snow samples collected from a remote location at Palmer Station, Antarctica had an average concentration of 5.6 x 10~6 M H202 (G.W. Miller and D.J. Kieber, unpublished results). Snowmelt run-off increased concentrations of H202 in surface seawater by more than a factor of two. In contrast, ice melt from sea ice was only slightly higher than surface seawater concentrations (G.W. Miller and D.J. Kieber, unpublished results).

Biological sources. While most if not all microorganisms actively decompose H202 (vide infra), it is somewhat surprising that some cyanobacteria and eukaryotic phytoplankton produce this compound in natural waters [105,127-131], One important pathway for the algal mediated-formation of H202 occurs under nitrogen limiting conditions when algae acquire nitrogen by cell surface enzymatic deamination of dissolved L-amino acids (or amines) to form the ammonium cation, which is subsequently taken up by the cell. By products of this reaction, including H202 and organic species such as a-keto acids, are released into seawater and not taken up by the algae [130,131]. Not all algae produce H202 under these conditions, indicating that this process is not universal [128]. The biological production of H202 has also been noted in cultures of the icthyotoxic flagellate Heterosigma akashiwo [132]. Twiner and Trick observed that this toxic phytoplankton produced substantial amounts of H202 (up to 7.6 pmol min-1 cell-1). The rate of peroxide production was stimulated by increasing temperature and was regulated by the availability of iron but was independent of light. Results of their study suggest that extracellular production of H202 occurs through metabolic pathways not directly linked to photosynthesis. These findings are very intriguing given that microorganisms generally decompose H202 to alleviate its toxicity.

Dark production rates for H202 were measured in the Sargasso Sea by Palenik et al. [105]. They collected samples from depths between 0 and 130 m and incubated them for 8 h in the dark. Seawater samples from the thermocline (40 to 60 m) yielded the highest H202 production rates within the first two hours (1.1 x 10"8 M h-1) and averaged 1.0 to 3.0 x 10-9 M h-1 for the entire incubation study. Presumably algae were the source of the H202, since hydrogen peroxide production was not observed when seawater samples were first filtered through a 1 /im filter. However, the depth corresponding to the maximum biological production rate for H202 (ca. 40-60 m) did not coincide with the chlorophyll maximum (90 m). It is interesting to note that the range of production rates reported by Palenik et al. [105] is comparable to the range of particle-mediated, dark production rates determined by Moffett and Zafiriou (0.8-2.4 x 10~9 M h"1) [106] in coastal seawater (Vineyard Sound, MA). Although microorganisms may be a potentially important source of H202 deeper in the water column (e.g., below the pycnocline) because photochemical production rates are comparatively small, these studies indicate that the biological production of H202 should be a minor source of this compound in surface waters compared to its photochemical production, which ranges from 9xl0~9Mh_1 in the Mediterranean Sea [103] to greater than 134 x 10~9Mh_1in high DOM coastal water from the Orinoco estuary [93].

Sources involving trace metal reactions. Hydrogen peroxide is formed from the oxidation of inorganic Cu(I) complexes through their reaction with 02~ [77]:

The observed rate constant for this overall reaction is approximately 2 x 1091 mol-1 s-1, depending on the type of Cu(I) complex considered. In all cases, the rate of the reverse reaction is negligible [77]. Midday concentrations of Cu(I) in surface seawater range from 1.0 to 1.3 x 10_1° M in coastal and oligotrophic waters, which accounts for no more than 10—15% of total Cu concentrations in these waters [133]. Since Cu(II) is the predominant Cu species present in the water column, a competing reaction is the reduction of inorganic Cu(II) complexes by 02~ (k = 0.1 x 109 1 mol-1 s-1) yielding molecular oxygen:

The rate for reaction (7) is approximately 30% faster than reaction (6) at a Cu(I)/Cu(II) ratio of 0.25.

Zafiriou et al. [77] observed that reactions (6) and (7) are more important in the removal of 02~ in seawater than its bimolecular disproportionation to H202. Their conclusion is similar to that of Petasne and Zika [17] who found that approximately 20-40% of 02~ did not disproportionate to H202 in seawater but rather decayed through other unknown reactions. Because kinetic data for representative organic complexes were not available, Zafiriou et al. [77] did not consider possible contributions from the reaction of organic copper complexes with 02~, even though these complexes are likely to control copper speciation in marine waters [25,85]. In a follow-up study, Voelker et al. [30] observed that naturally occurring organic copper(n)-complexes degraded 02 ~ at

a rate comparable to the inorganic copper(II) complexes previously studied by Zafiriou et al. [77],

8.2.4.2 H202 removal pathways

Biological removal. The major removal pathway for H2O2 in natural waters is through algal and bacterial consumption, presumably mediated by enzymatic processes as a detoxification mechanism. Petasne and Zika [110] observed that the biologically mediated loss of H202 was due to prokaryotic microorganisms and, to a lesser extent, eukaryotic microorganisms. The bacterium Vibrio pelgius, with estimated natural levels at 108 cells l-1, removed H202 in the dark with a second order rate constant of 8.81 x 10~u 1 cell-1 h_1. This translates to a H202 loss of 2x 10~9 M h-1 due to Vibrio pelgius at ambient H202 concentrations measured in Biscayne Bay, FL. This is a significant fraction of the total biological consumption (5 —16 x 10-9 M h-1) that was observed due to a natural assemblage of microorganisms in unfiltered Biscayne Bay seawater [110]. If samples were sterilized by autoclaving, no H202 loss was observed in the dark. In addition, H202 loss was not observed in 0.2 /¿m filtered seawater, but was observed after the addition of bacteria. The use of 0.2 fim filters to sterilize marine, lake or hydrothermal water samples almost completely stopped the loss of H202 in these samples [95,109]. These results are consistent with a biological rather than an abiological removal mechanism for H202. The residual loss of H202 that was observed in some filtered water samples may be due to H202 reactions with DOM or trace metals [95].

Zepp et al. [129] studied H202 cycling in nine different algal cultures, including cultures of cyanobacteria and green algae, and determined that the mean second order rate constant for the dark loss of H202 was 4 x 10~3 m3 (mg chl a)~l h-1. Using this rate constant, and typical concentrations of chlorophyll a of 1 and 10 mg m~3 (G.L. Boyer, personal communication) and H202 levels of 3 and 15 x 10-8 M in Antarctic [104] and coastal seawater [110], respectively, the calculated loss rate of H202 is 1.2 x 10-10 M h_1 in Antarctic seawater and 6xlO_9Mh_1in coastal seawater. Calculated loss rates are in good agreement with the average loss rate of5xl0_9Mh_1 determined at four coastal stations in the Caribbean Sea and Orinoco River outflow [93], and the loss rate of 7xl0-11Mh-1 determined at an oligotrophic Antarctic station [99]. These loss rates were all determined during non-bloom, low chlorophyll conditions. However, it is important to note that algal blooms will likely yield much faster removal rates for H202 in the water column. Likewise, since these biologically-mediated decay studies were conducted in the dark, any additional effect of light on the biological removal of H202 is unknown.

Trace metal reactions. Hydrogen peroxide is an important reactive redox intermediate in natural waters for reactions involving biologically important trace metals such as iron, copper and manganese [85,87,134,135]. Perhaps the most environmentally and biologically significant aspect of H202 is its capacity to react with trace metals to form the highly reactive hydroxyl radical. As noted previously in this chapter, the hydroxyl radical can rapidly oxidize organic matter, transform anthropogenic organic pollutants into either toxic or inert compounds [21,136], and damage cell membranes in aquatic microorganisms. The toxicological response of bacterioplankton to H2O2 in humic-rich lakes is probably due to the production and subsequent reactions of the 'OH radical [137,138], The high concentration of iron (> 200 fig l"1) and low pH (< 5.5) that are typical of some humic-rich lakes and rivers are ideal conditions for efficient 'OH radical formation. The 'OH radical is formed by the oxidation of reduced iron via the photo-Fenton reaction (so-called because the reactants Fe(II) and H202 are formed from photochemical processes):

As may be expected, this reaction is affected by pH, ionic strength and temperature [86]. An analogous Fenton-type reaction is also observed for Cu(I) and H202 (for review see [135]). The pseudo-first order rate constant for this reaction in seawater is 5 x 10~3 s"1 at 25°C and 1.0 x 10"7 M H202 [85]. This decomposition pathway is insignificant in oligotrophic seawater at subnanomo-lar concentrations of photochemically produced Fe(II) [139]. For example, at typical concentrations of 1.0 x 10 ~11 M Fe(II) and 1.0xl0~7M H202, the rate of loss of H202 is only 1.8 x 10~10 M h-1, which is over two orders of magnitude slower than H202 production rates in these waters. Molecular oxygen will also react with Fe(II) according to reaction (9) at rates comparable to those of reaction (8):

At a dissolved oxygen concentration of 2.1 x 10~4 M, the reported pseudo-first order rate constants for this reaction range from 5.8 x 10-4 s~1 to 2.2 x 10~3 s~1 [134,140-142], with an average value of 1.2 x 10-3 s~].

At low pH (ca. 3-5), H202 can be consumed through iron cycles associated with the oxidation of fulvic acids involving the photo-Fenton reaction (reaction 8) [84], As previously discussed, reactions (8) and (9) may be important in humic-rich natural waters at low pH such as in black-water rivers of the southeastern United States (e.g., Suwanee River, GA), but their importance in marine waters at higher pH (typically ca. pH 8.2) is unlikely except for some organic rich coastal environments (vide supra).

Hydrogen peroxide, produced from the photolysis of organic matter, has been shown to reduce insoluble synthetic manganese oxides to soluble Mn(n) in natural waters, with a second order rate constant of 40 1 mol-1 s_1 [87]. However, natural oxides were not significantly reduced by H202. Observed rates of photochemical formation of Mn(II) in the natural oxide samples were only one-sixth of those observed with synthetic oxides. These results coupled with DOM photolysis experiments and catalase addition studies indicate that H202 reduction is only a minor source of dissolved manganese in seawater, accounting for only 10-20% of the total Mn(II) signal. The main chromophores and reductant(s) responsible for Mn photoreduction are not known, but are speculated to arise from the bacterial manganese oxides themselves [87].

Primary photolysis. Even though biologically-mediated decay is generally

considered to be the main pathway for removal of H202 in seawater, direct photochemical loss also occurs:

The primary photolysis of H202 in aqueous solution yields the 'OH radical [143], However, in natural waters, H202 has very little absorption at wavelengths greater than 290 nm and, therefore, only a very small cross section with actinic solar radiation. As a result, the photochemical decay rate of H202 is slow in natural waters (vide infra) and 'OH radical yields are small compared to other photochemical sources of 'OH (e.g., DOM, nitrate, nitrite) [47]. In the field, H202 photolysis rates are less than 5% of photochemical production rates in seawater from temperate latitudes [49], and approximately 50% of midday production rates in polar regions, based on in situ drifter experiments [104]. These contrasting results are likely due to the much lower photochemical production rates of H202 in polar waters (ca. factor of five lower) compared to lower latitudes, primarily due to differences in temperature. These results are not expected to be due to differences in photolysis rates, since the rate of primary photolysis of H202 should not be affected by temperature and should therefore be similar in these waters (and only affected by differences in the scalar irra-diance).

8.2.4.3 Reactions of H202 with DOM

Based on standard state reduction potentials, H202 should be a strong oxidant in both acidic and basic solutions [144]:

However, experimental results indicate that H202 is not very reactive in aqueous solution in the absence of catalysts such as oxidases, Fe(II) or light [21]. Uncatalyzed redox reactions involving H202, especially those with organic compounds, proceed very slowly in aqueous solution even though they are thermodynamically favored in some cases. Presumably these reactions are slow due to a high kinetic barrier. For example, the standard state free energy for the oxidation of dimethyl sulfide to dimethyl sulfoxide by H202 is —223 kJ mol-1, indicating that this oxidation is thermodynamically quite favorable under most reaction conditions. However, this oxidation proceeds slowly with a small rate constant, 0.14 1 mol-1 s_1 or less (Table 4). To further illustrate this lack of reactivity, we compiled a selection of the relatively few rate constants that have been reported for the uncatalyzed oxidation of organic species by H202 in aqueous solution (Table 4). Indeed, although H202 reacts with a wide variety of organic compounds, rate constants for these reactions are quite small (ranging from 3.0 x 10~6 to 1021 mol-1 s_1) relative to comparable reactions involving other ROS (vide supra). A notable exception is the methyl radical [145], but even this rate constant (2.7 x 104 1 mol-1 s_1) is small relative to reactions of the methyl radical with other species. The poor reactivity of H202 in uncatalyzed

H202+ 2H+ + 2e-->2H20 £H° = 1-77V H02- + H20 + 2e--»30H- £H° = 0.87V

In general, the impact of ROS on aquatic organisms should be inversely related to the reactivity of the ROS. For example, the direct effect of externally-generated *OH radicals on aquatic microorganisms is expected to be minimal because it is extremely reactive, which results in an extremely short lifetime that does not allow for significant transport to the cell surface [160]. However, production of the OH radical outside the cell may exert an indirect effect on cellular systems through the production of longer-lived free radical species such as the carbonate radical (which can potentially affect algal uptake of inorganic carbon) and the dibromide radical ion (Figure 1). These longer-lived

Solar Power

Solar Power

Start Saving On Your Electricity Bills Using The Power of the Sun And Other Natural Resources!

Get My Free Ebook


Post a comment