Competitive Effects of Calcium and Magnesium Ions on the

Article pubs.acs.org/est

Competitive Effects of Calcium and Magnesium Ions on the Photochemical Transformation and Associated Cellular Uptake of Iron by the Freshwater Cyanobacterial Phytoplankton Microcystis aeruginosa Manabu Fujii,† Anna C. Y. Yeung,‡ and T. David Waite*,‡ †

Department of Civil Engineering, Tokyo Institute of Technology, 2-12-1-M1-4 Ookayama, Tokyo 152-8552, Japan School of Civil and Environmental Engineering, The University of New South Wales, Sydney, NSW 2052, Australia

‡

Downloaded by UNIV OF CAMBRIDGE on September 8, 2015 | http://pubs.acs.org Publication Date (Web): July 15, 2015 | doi: 10.1021/acs.est.5b01583

S Supporting Information *

ABSTRACT: Photochemical reduction of iron and iron uptake by Microcystis were investigated in a freshwater medium (pH 8) containing a range of calcium (Ca) and magnesium (Mg) ion concentrations (0.002−20 mM). In a medium containing the chelator ethylenediaminetetraacetic acid (EDTA), 50-fold increases in net photochemical formation rates of unchelated ferrous iron (Fe(II)′) were observed as the concentration of calcium or magnesium metal (Me) was increased to exceed the concentration of EDTA. Kinetic modeling of iron transformation processes indicated that the facilitated Fe(II)′ formation is attributed to Me-promoted photoreductive dissociation of the ferric iron-EDTA complex. In the medium containing Suwanee River fulvic acid, in contrast, the competitive effect of Me on photochemical Fe(II)′ formation appears to be negligible due to the weak binding affinities of fulvic acid to Me. The cellular iron uptake rate in the EDTA-buffered system increased by ∼3-fold in the excess Me condition where the increased rate of photochemical Fe(II)′ formation was observed, whereas the presence of Me resulted in a decrease in iron uptake rate in the fulvic acid system (by up to 5-fold). The decrease in iron uptake is likely caused by Me binding to iron transporters and other entities involved in intracellular iron transport. The findings of this study indicate a significant effect of Ca and Mg concentrations in natural waters on iron uptake by Microcystis, with the magnitude of effect depending strongly on ligand type.

1. INTRODUCTION Iron is one of the important trace metal nutrients for the growth of microorganisms, including phytoplankton.1,2 In aerobic natural surface waters at circumneutral pH, a majority of dissolved Fe is present as organically complexed ferric iron (Fe[III])3−7 due to the low solubility of inorganic Fe(III) (e.g., ∼10−11 M at pH 7.5−9)8 and rapid oxidation of ferrous iron (Fe[II]).9 However, organically complexed Fe is generally considered to be inaccessible for acquisition by phytoplankton, except for those possessing transport systems that are specific for particular Fe complexes, including Fe(III) siderophores.10−13 Recent studies have consistently suggested that Fe uptake in cyanobacteria may occur via siderophore-independent processes due to the lack of siderophore-associated genes in many prokaryotic phytoplankton14 and the predominance of the uptake of unchelated iron by many species of freshwater and coastal cyanobacteria (e.g., Microcystis,15 Anabaena,16 Synechocystis,17,18 and Lyngbya19), even under Fe stress. Therefore, the concentration of unchelated Fe species (Fe′), potentially in both ferric (Fe(III)′) and ferrous (Fe(II)′) states, is likely a significant determinant of Fe availability by cyanobacteria in surface natural waters. © 2015 American Chemical Society

There is now substantial accumulated evidence that reductive dissociation of organically compelxed Fe(III) to Fe(II)′ facilitates Fe uptake by phytoplankton in both seawater19−23 and freshwaters.17,18,24,25 This process may include the extracellular reduction via photochemical and biological processes mediated by ligand-to-metal charge transfer (LMCT),15,18,20 cellular plasma−membrane reductase,21,22 and superoxide,19,26 with the significance of these various processes possibly depending on the phytoplankton species, type of Febinding ligand, and other medium conditions. While the reduction of Fe(III) generates Fe(II)′, the steady-state Fe(II)′ concentration achieved is also influenced by other competing reactions, including re-complexation by Fe-binding ligands15 and oxidation by oxygen27 when present. The oxidation of Fe(II)′ to Fe(III)′ by dissolved oxygen may be critical in Fe acquisition by oceanic eukaryotic phytoplankton.21,28 Received: Revised: Accepted: Published: 9133

March July 1, July 1, July 1,

28, 2015 2015 2015 2015 DOI: 10.1021/acs.est.5b01583 Environ. Sci. Technol. 2015, 49, 9133−9142

Article


Environmental Science & Technology

Figure 1. Fe uptake model for Microcystis aeruginosa in the absence and presence of major divalent metals (Me). Under illuminated condition, the photochemical reductive dissociation of ferric-ligand complex (FeIIIL) generates unchelated Fe(II) (i.e., Fe(II)′). The generated Fe(II)′ can be taken up by cells via the Fe transporter. In the EDTA system, the photochemical Fe(II)′ generation (khν) is slower in the absence of Me (reaction depicted as bluecolor arrow). In contrast, Fe(II)′ generation (khν,Me) is facilitated in the presence of Me (reaction depicted as red-color arrow). In the fulvic acid system, Fe(II)′ generation is relatively comparable in the absence and presence of Me. Regarding cellular Fe uptake, the half-saturation constant substantially increases in the presence of Me (i.e., KS,X < S,MeX) due to the association of Me with plasma−membrane transporter, and the maximum Fe uptake max decreases (i.e., ρmax S,X >ρS,MeX) due possibly to the physiological or physicochemical (adverse) effects of Me on the intracellular Fe transport. In the abiotic Fe transformation, ferrozine (FZ) also significantly competes with Fe(II) complexation by L both in the absence and presence of Me. Solid arrows represent major reactions, whereas dashed arrow shows relatively minor reactions. Note that, in the absence of FZ, the FeIIL formation becomes a significant reaction that influences steady-state Fe(II)′ concentration. Kinetic rate constants depicted near the arrows correspond to those listed in Table 1.

cyanobacteria, likely incorporates concentration gradient dependent passive diffusion through nonspecific transmembrane channels (porins) in the outer-membrane followed by intracellular transport processes.18,37,38 Free or membrane-anchored periplasmic Fe-binding proteins (FutA1 and FutA2) and membrane transporters (FutB, FutC, and FeoB) are known to be involved in intracellular iron uptake.18,38,39 While Fe uptake by many of these Fe transporters is likely to be highly specific, the inhibitory effect of nontoxic levels of Cu on Fe uptake by the Gram-negative proteobacterium Helicobacter pylori40 suggests that competition with other metals may influence intracellular Fe transport. Although the physiological and physicochemical processes involved remains unclear, the effect of competing metals on cellular Fe uptake is likely an important consideration in developing a comprehensive understanding of Fe uptake by cyanobacterial phytoplankton. In our previous studies15,25,41 we have consistently suggested that Fe availability by cyanobacterial phytoplankton (such as Microcystis aeruginosa) is strongly or moderately controlled by photoreductive dissociation of organically complexed Fe(III) (Figure 1). However, limited studies have been undertaken to examine the impact of the major divalent metals Ca and Mg on either the photochemical transformations of Fe or the associated uptake of Fe by phytoplankton. Given the highly variable concentrations of these metals in natural waters, we have investigated the effect of Ca and Mg on Fe uptake by the freshwater cyanobacterium M. aeruginosa. This organism was selected because the Fe uptake kinetics is otherwise well understood.

Ethylenediaminetetraacetic acid (EDTA) is the most common organic ligand used in the culturing of algae in both seawater and freshwater media.29 This trace metal buffer binds thermodynamically stable Fe(III) with a relatively high affinity under freshwater conditions30 but only with moderate binding strength in seawater31,32 with the lower binding strength in seawater due to the presence of high concentrations of the divalent metals (Me) calcium (Ca) and magnesium (Mg) which compete with Fe(III) for EDTA binding sites.33 Specifically, the concentration of unchelated Fe(III) is substantially higher in media exhibiting a high concentration ratio of Me relative to EDTA (where most EDTA that is not bound to Fe is present as an EDTA-Me complex) because the exchange reaction between Fe(III) and the EDTA-Me complex is very slow compared to that for free EDTA.30,32 In contrast to EDTA, humic and fulvic acids, major Fe-binding ligands in freshwater34 and coastal seawaters,6,7 have less affinity for Ca and Mg than does EDTA.30 Consequently, it was suggested in a previous study30 that the rate of Fe(III) complexation by Suwanee River fulvic acid (SRFA) at pH 8 only slightly decreases in the presence of Ca and Mg, even at the high concentrations found in seawater. These findings suggest that the effect of Ca and Mg (i.e., the elements most responsible for water hardness) on Fe transformation kinetics may vary substantially depending on the relative affinity of Me to the Fe-binding ligands present. In addition to any effects on abiotic transformation kinetics, the presence of competing metals may directly influence the uptake of iron and other nutrient metals by phytoplankton. Previous studies by Sunda and Huntsman,35,36 for example, indicated that the uptake of Mn in a diatom and a green alga was inhibited by high concentrations of the divalent metals Cu, Zn, and Cd. Metal transport in Gram-negative bacteria, including 9134

DOI: 10.1021/acs.est.5b01583 Environ. Sci. Technol. 2015, 49, 9133−9142

Article



2. EXPERIMENTAL METHODS Full descriptions of experimental procedures, including preparation of chemical stocks, culturing medium, pH adjustment, cell culturing, and photochemical and biological Fe uptake experiments, are provided in the Supporting Information (SI.1). Comments on potential analytical concerns, including contamination and redox speciation (thermal reduction) of Fe, are also provided in SI.2. Cell culturing, photochemical experiments, and Fe uptake assays were performed by employing the methods described elsewhere15,25 but with some modifications. Briefly, a batch culture of M. aeruginosa PCC7806 was grown in pH 8 Fraquil* medium (containing 26 μM EDTA and 100 nM Fe) at a temperature of 27 °C and 14 h:10 h light:dark cycle, where chemical transformations of Fe15,25 and biological responses41,42 are well documented. In the photochemistry and uptake experiments, modified Fraquil* media were also prepared at different Fe, ligand (EDTA or SRFA), and Me (Ca or Mg) concentrations. The photochemical experiments were conducted by using the ferrozine (FZ) competitive assay15 where the time course of ferrous-ferrozine (FeIIFZ3) complex formation was spectrophotometrically monitored at a wavelength of 562 nm following the addition of FZ stock to the Fraquil* medium containing pre-equilibrated Fe-ligand and Me. Final concentrations of chemicals in the samples were 5−20 μM Fe, 40 μM EDTA, 50 mg·L−1 SRFA, 2 μM −20 mM Me, and 1 mM FZ. In the case of Fe uptake experiments, the short-term incubational assays were initiated by spiking the pre-equilibrated radiolabeled 55 Fe-ligand stock solution to the cell culture. The test culture was prepared by resuspending the cells (harvested at exponential growth phase) to Fe- and ligand-free Fraquil* which was previously amended with different Me concentrations ranging from 2 μM - 20 mM. The 55Fe-ligand stock was added at final concentrations of 0.05−1 μM Fe, 20 μM EDTA, and 10 mg·L−1 SRFA. The final cellular concentration was adjusted to ∼2 × 106 cell·mL−1. After 2 h incubation under the light (where 55Fe accumulation is linear with respect to time), cells were harvested onto a membrane filter followed by washing with an EDTA/ oxalate chelate solution in order to remove any Fe (oxyhydroxides) adsorbed to the cell surface.43 Subsequently, cellular radioactivity was determined using a liquid scintillation counter. In all incubations and experiments, light was horizontally supplied by cool-white fluorescent tubes with total radiation intensity of 157 μmol·m−2·s−1.15 In this study, the formation rate constant for the FeIIEDTA complex was also determined as described in SI.3.

It should be noted that Fe(III) bound to SRFA is also continuously reduced even in the dark, most likely due to the presence of redox-active moieties in fulvic acids (e.g., hydroquinones) as reported previously,47,48 while thermal reduction of FeIIIEDTA is negligible under the conditions examined here. Thus, the reduction rate in the photochemical experiment using SRFA includes the thermal reduction which was estimated to account for ∼30% of total reduction rate, as noted in SI.2.3. The generated Fe(II)′ may subsequently form complexes with the ligand (L) and FZ, when present: k f‐ L

Fe(II)′ + Lf ⎯→ ⎯ Fe IIL k f‐FZ

Fe(II)′ + 3FZ ⎯⎯⎯→ Fe IIFZ3

khν

(3)

where kf‑L (M−1·s−1) and kf‑FZ (M−3·s−1) are rate constants for Fe(II)′ complexation by L and FZ, respectively. Lf indicates free ligand that is not bound to Fe or any other metals. The steadystate Fe(II)′ concentration is significantly affected by the oxidation process in some conditions,27 because Fe(II)′ is oxidized at appreciable rates primarily by dissolved oxygen (O2) in air-saturated and circumneutral pH waters.9,27 However, the effect of oxidation on the Fe(II)′ concentration is expected to be minimal under our experimental conditions using freshwater medium, where the Fe(II)′ oxidation rate is calculated to be at least ∼103-fold smaller than the rates of complexation of Fe(II)′ by EDTA and FZ (i.e., Fe(II)′ complexation by organic ligands outcompetes Fe(II)′ oxidation by oxygen), as described in SI.2.3 and in a previous study.15 In addition to the two complexation reactions, the photochemically generated Fe(II)′ can interact with ligand−metal complex (MeL) in the presence of Me, as follows: k f‐MeL

Fe(II)′ + MeL ⎯⎯⎯⎯→ Fe IIL + Me

(4)

where kf‑MeL (M−1·s−1) is the rate constant for Fe(II) complexation by MeL. Given the slower process of metal exchange,30,32 the rate of complexation between Fe(II)′ and MeL is also expected to be slower than that between Fe(II)′ and Lf. Indeed, the Fe(II) complexation rate was determined to decrease with increasing Me concentration, as described in detail below (see section 4.1). However, an observed sigmoidal increase in FeIIFZ3 formation with increase in Me concentration could not be solely accounted for by the decreased rate of the FeII-L complexation reaction. Since the photochemical experiment employed high concentration of FZ (i.e., FZ complexation was found to be the dominant reaction in Fe(II)′ removal in most cases where Me was present), the reduction in extent of FeIIL formation as a result of Me competition for ligand binding sites was calculated to be relatively small (see section 4.1). Therefore, to account for the observed increase in FeIIFZ3 formation with increasing Me concentration, we introduced the adjunctive metal-exchange reaction, which becomes a significant factor in FeIIFZ3 formation at higher Me concentration, as follows:

3. MODEL 3.1. Fe(II) Formation in the Photochemical Experiment. The photochemical transformation of Fe(III)-ligand complexes ultimately produces Fe(II)′ and photo-oxidized ligand (Lox) most likely via ligand-to-metal charge transfer (LMCT)27,44 and/ or other redox processes mediated by secondary generated entities during the photolysis, which may include superoxide45 and organic radicals.46 While light-mediated Fe transformations are recognized to involve a number of complex reactions, the photoreductive dissociation of organically complexed Fe(III) (FeIIIL) resulting in formation of Fe(II)′ can be empirically described by a pseudo-first-order reaction (rate constant khν, s−1) under constant illumination,15,27 as follows: Fe IIIL → Fe(II)′ + Lox

(2)

khν ,Me

Fe IIIL−Me ⎯⎯⎯⎯→ Fe(II)′ + MeLox −1

(5) III

where khν,Me (s ) is the rate constant for Fe L photoreduction in the presence of Me. We assume that photochemical FeIIFZ3 formation is facilitated by the adjunctive association of Me with the Fe-ligand complex to form a ternary complex (FeIIIL-Me). Although there is limited evidence for the FeIIIL-Me ternary complex from previous studies, the work by Nowack and Sigg in 1996 indicated that metal−ligand complexes, including

(1) 9135


Article

Environmental Science & Technology Table 1. Kinetic and Thermodynamic Rate Constants Used in This Study EDTA parameters

no Me

Ca

Mg

k*hν

s−1

1.6 × 10−7 b,c

5.9 × 10−6 b,c 6.4 × 10−6 e

khν, khν,Me kf‑L, kf‑MeL

s−1 M−1·s−1

2.5 × 10−7 f 7.1 × 106 c,h

KMeL KFeL‑Me

M−1 M−1

6.0 × 10−6 f 2.7 × 105 c,h 2.1 × 106 k 1.3 × 108 l 1.0 × 107 m

ρmax S,X ,


units

ρmax S,MeX

KS,X, KS,MeX KMeX

−1

zmol·cell · h−1 M−1

660

o

M−1

(i) Fe Transformation Kinetics 5.2 × 10−6 b,c 3.1 × 10−5 b

5.3 × 10−6 f 2.7 × 105 c 4.8 × 105 h 2.0 × 106 l 5.0 × 104 m

3.1 × 10−5 f 4.5 × 104 i

p

120

1.0 × 10

no Me

(ii) Cellular Fe Uptaken 36 660o

p

−14o

SRFA

−13 o

−13 o

4.1 × 10

4.1 × 10

3.4 × 104 q

5.9 × 103 q

−14 o

Ca

Mg

3.1 × 10−5c,d 1.1 × 10−5b 2.9 × 10−5e N.D.g 4.5 × 104d

3.1 × 10−5 c,d 2.9 × 10−5 b

4.7l N.D.g

16l N.D.g

120p

36p −13 o

1.0 × 10

FZ

N.D.g 4.5 × 104 d

related equationsa 10, 11

3.1 × 1011d,j

1, 5, 9 2−4, 9 6−9 9

14, 15 −13 o

4.1 × 10

4.1 × 10

3.4 × 104 q

5.9 × 103 q

14 13

Equations described in the text. k*hν values were determined by fitting eq 11 to the Fe FZ3 time course data in the absence of Me and presence of excess Me (Figure S1). cValues used in this study, as these values provided the best fit to the experimentally determined data. dEffect of Me on the complexation of Fe(II) by fulvic acid and FZ was assumed to be negligible under the conditions examined. eReported photoreduction rate constant15 determined in the Fraquil* (pH 8) containing Ca and Mg concentrations of 1.5 × 10−4 M and 1.0 × 10−3 M, respectively, at constant fluorescence illumination of 157 μmo·m−2·s−1. fkhν and khν,Me values were determined from eq 9 in the absence of Me (αMe = 0) and presence of excess Me (αMe = 1), respectively. Associated parameters and FeIIFZ3 time course data at given conditions were used for the calculation. gValues were not determined due to the negligible association of Me and fulvic acid. hValues were independently determined in this study by using ligand competition methods as described in SI.2. In the calculation of steady-state Fe(II) concentration using eq 9, the Fe-binding capacity of 260 nmol·mg−1 for SRFA61 was used. i Reported value.62 jReported value52 with unit of M−3·s−1. kValue reported in previous work15 at Fraquil* (pH 8) containing Ca and Mg concentrations of 1.5 × 10−4 M and 1.0 × 10−3 M, respectively. lReported values.30 mValues from best fit of model eqs 9 and 10 to the experimental data shown in Figure 2A. nUptake parameters are identical in the EDTA- and SRFA-buffered systems. oValues were determined from best fit of model to the 55Fe uptake data in Figure 3B via nonlinear regression using software R. The value for Mg was assumed to be equal to that for Ca. p Values were determined by using 55Fe uptake rate in the SRFA system containing 20 mM Me (Figure 3C), where Fe uptake is assumed to be saturated. qValues from best fits to the 55Fe uptake data in the SRFA system (Figure 3C). a

b

II

For the ternary complex FeIIIL-Me, conditional stability constant KFeL‑Me = [FeIIIL-Me]/([FeIIIL][Me]) was similarly introduced, yielding [FeIIIL-Me] = KFeL‑Me[Me][FeIIIL*]/(1 + KFeL‑Me[Me]) = αMe[FeIIIL*] and [FeIIIL] = [FeIIIL*] − [FeIIIL-Me] = (1 − αMe)[FeIIIL*], where [FeIIIL*] is the total concentration for Fe(III) bound to ligand (i.e., [FeIIIL*] = [FeIIIL] + [FeIIIL-Me]) and αMe is a fraction of the ternary complex FeIIIL-Me relative to FeIIIL* (where this parameter is a function of [Me]) (0 ≤ αMe ≤1). At steady state, where the time-dependent change in Fe(II)′ concentration is considered to be invariant (i.e., d[Fe(II)′]/dt ≈ 0), the rate law for FeIIFZ3 formation can be written as

CaEDTA, can form ternary complexes with Fe on the surface of goethite (α-FeOOH).49 Regarding the interaction between Me and L, concentrations of MeL, Lf, and Me at given total Me and L concentrations were calculated by assuming a pseudo-equilibrium (side) reaction between Me and L, as described by Me + Lf = MeL with conditional stability constant of KMeL = [MeL]/[Me][Lf].30 Provided that the concentrations of free ligand and its complex with Me ([L*] = [Lf] + [MeL]) are constant over the duration of measurement (due to the relatively small changes in concentrations of ligand complexes with Fe and other trace metals present in the Fraquil* medium), the equilibrated concentrations of MeL, Lf, and Me were determined by KMeL with known total concentrations of Me ([Me]T) and [L*], where KMeL = [MeL]/ {([L*] − [MeL])([MeL]T − [MeL])}, yielding

d[Fe IIFZ3] = k f‐FZ[FZ]3 [Fe(II)′]SS = dt III ⎛ k [FeIIIL] + k ⎞ hν hν ,Me[Fe L‐Me] ⎟= k f‐FZ[FZ]3 ⎜⎜ 3⎟ ⎝ k f‐L[Lf ] + k f‐MeL[MeL] + k f‐FZ[FZ] ⎠ ⎛ ⎞ khν(1 − αMe) + khν ,MeαMe ⎟[Fe IIIL*] k f‐FZ[FZ]3 ⎜ 3 ⎝ k f‐L[Lf ] + k f‐MeL[MeL] + k f‐FZ[FZ] ⎠

⎧ ⎪ 1 − [MeL] = ⎨[L*] + [Me]T + KMeL ⎪ ⎩ ⎫ 2 ⎛ ⎪ 1 ⎞ ⎜[L*] + [Me]T + ⎟ − 4[L′][Me]T ⎬ KMeL ⎠ ⎪ ⎝ ⎭

(9)

2

where [Fe(II)′]SS is the steady-state Fe(II)′ concentration. Under the given conditions, the FeIIFZ3 formation can be further simplified with pseudo-first-order rate constant (k*hν, s−1) as described below:

(6)

[Lf ] = [L*] − [MeL] [Me] =

[MeL] KMeL[Lf ]

(7)

d[Fe IIFZ3] = kh*ν[Fe IIIL*] dt

(8) 9136

(10) DOI: 10.1021/acs.est.5b01583 Environ. Sci. Technol. 2015, 49, 9133−9142

Article

Environmental Science & Technology Approximations of [FeIIIL*] ≈ [Fe]T − [FeIIFZ3] (where [Fe]T denotes total Fe concentration in the system) due to the low [Fe(II)′] (e.g., [Fe(II)′]/[FeIIFZ3] < 1.4 × 10−7 in the presence of 1 mM FZ) and [FZ] ≈ [FZ]T due to the high [FZ]T (e.g., [FZ]T = 1 mM ≫ 20 μM = [Fe]T) followed by integration yields the following relationship between FeIIFZ3 concentration and time (t):


⎛ ⎞ [Fe]T ⎟ = kh*νt ln⎜ II ⎝ [Fe]T − [Fe FZ3] ⎠

Fe for cellular uptake. This effect is best described by a reduced affinity of the plasma−membrane transporters for Fe when the transporter forms a complex with Me, as follows: X + Me = MeX,

(11)

ρS = =

max [S] ρS,X

KS,X + [S] k up,X[X][S] KS,X + [S]

+

+

max [S] ρS,MeX

KS,MeX + [S] k up,MeX[MeX][S] KS,MeX + [S]

⎛ k up,X(1 − β )[S] k up,MeXβMe[S] ⎞ Me ⎟⎟[X]T + = ⎜⎜ KS,X + [S] KS,MeX + [S] ⎠ ⎝ (14)

ρSmax [S] KS + [S]

(13)

where X represents free Fe transporters that are not bound to Me, MeX represents the Me−transporter complex, and KMeX is the conditional stability constant for complexation between X and Me. Given that the total concentration of transporter ([X]T) is a sum of X and MeX concentrations (i.e., [X]T = [X] + [MeX]), [X] and [MeX] can be described as [X] = (1 − βMe)[X]T and [MeX] = βMe[X]T, where, similar to the photochemical model, βMe (= KMeX[Me]/(1 + KMeX[Me])) is the fraction of [MeX] relative to total transporter concentration (0 ≤ βMe ≤ 1). Assuming that Fe transport is mediated by concurrently occurring reactions between Fe(II)′ and X or MeX, the total Fe uptake in the presence of Me may be described as below:

The detailed procedure of parameter determination in the photochemical Fe(II)′ generation process is described in SI.4.1. Briefly, by using time course data of [FeIIFZ3] collected in the photochemical experiment, khν * was experimentally determined from the slope of the linear regression line in the plot of time versus ln([Fe]T/([Fe]T − [FeIIFZ3])) (Figure S1). Subsequently, khν and khν,Me were determined by substitutions of khν * determined in the absence of Me (αMe = 0) and in the presence of excess Me (αMe = 1), respectively, to eqs 9 and 10. Then, KFeL‑Me was determined from the best fit of the model eqs 9 and 10 to the khν * data as a function of Me concentration (see section 4.1 for details), where concentrations of MeL, Lf, and Me were determined by using eqs 6−8 with reported values of KMeL (Table 1).30 3.2. Cellular Fe Uptake. A number of previous studies on trace metal (including Fe) uptake by phytoplankton1,50 have consistently suggested that cellular uptake follows Michaelis− Menten-type saturation theory:

ρS =

KMeX = [MeX]/[X][Me]

where KS,X and KS,MeX are half-saturation constants for X and max MeX, respectively. ρmax S,X (= kup,X[X]) and ρS,MeX (= kup,MeX[MeX]) represent rates of maximum Fe uptake in the absence and presence of Me, respectively. kup,X and kup,MeX are first-order rate constants for Fe uptake in the absence and presence of Me, respectively. To satisfactorily describe the experimental observation in section 4.2, the latter term of eq 14 needs to dominate at high concentrations of Me (i.e., βMe = 1), while the value of ρmax S,MeX is lower than ρmax S,X (i.e., the maximum uptake occurs at a slower rate at high [Me]). Under saturated uptake (i.e., KS,X, KS,MeX ≪ [S]), the total Fe uptake rate can be described by the maximum Fe uptake rate:

(12)

where [S] indicates the steady-state concentration of the biologically available portion of Fe in the extracellular environment (i.e., [Fe(II)′]SS in our system)15 and KS and ρmax S represent the half-saturation constant and the maximum uptake rate under the conditions examined, respectively. ρmax S can be described by a product of the first-order rate constant for Fe transport and the total concentration of Fe transporter. In this study, we define the term “Fe uptake” as an intracellular accumulation of Fe which is not removed from the cells by the EDTA/oxalate wash. In contrast, “Fe transport” is used to refer to the intracellular specific transport process, which represents the Fe transport from periplasmic to cytoplasmic spaces by free or membraneanchored periplasmic Fe-binding proteins (e.g., FutA1 and FutA2) and membrane transporters (FutB, FutC, and FeoB).18,38,39 Intracellular transport may follow concentrationgradient-dependent passive diffusion of Fe from the extracellular environment into the periplasmic space through transmembrane channels in the outer membrane.37 The Michaelis−Menten theory is only applicable when uptake is not limited by the diffusive flux of available metals.1 However, given that the diffusion layer thickness of metal complexes in aqueous solution is generally on the order of tens of micrometers,51 it is likely that the effect of physical diffusion on the metal flux in proximity of the cell surface is relatively small for small-sized phytoplankton such as M. aeruginosa (cellular radius is ∼3 μm). Under such conditions, the substrate concentration at the outer surface of the cellular membrane can be approximated to that of the bulk environment. According to the observations described below (section 4.2), Me present in the culture medium can effectively compete with

max max ρS = ρS,X + ρS,MeX = (k up,X(1 − βMe) + k up,MeXβMe)[X]T

(15)

Detailed description of the procedure used in parameter determination is described in SI.4.2. Briefly, in the Fe uptake max model, five parameters (KS,X, KS,MeX, ρmax S,X , ρS,MeX, and KMeX) were considered to be fitting parameters. The first two parameters, KS,X and KS,MeX, were determined from a nonlinear fit of the Michaelis−Menten equation to the 55Fe uptake data measured over a range of Fe(II)′ concentrations in the absence of Me (βMe = 0) and presence of excess Me (βMe = 1), respectively. Fe(II)′ concentrations was calculated according to the procedure described in SI.5. The maximum Fe uptake rates in the absence max and presence of excess Me (i.e., ρmax S,X and ρS,MeX, respectively, in 55 eq 15) were determined from the Fe uptake data in the SRFA system, where Fe uptake rate is saturated. The affinity of transporter to Me (KMeX) was determined by fitting eq 15 to the 55 Fe uptake data for all Me concentrations in the SRFA system. To this end, values of βMe and [Me] were calculated by βMe = KMeX[Me]/(1 + KMeX[Me])) and eqs 6−8, respectively. 9137


Article


The large increase in FeIIFZ3 formation (k*hν) around the threshold values is primarily due to the higher net generation of Fe(II)′, given that (i) the FZ concentration was kept constant over the range of Me concentrations used and (ii) the effects of hardness cations on the rate of Fe(II) complexation by FZ is negligible.52,53 The higher rate of Fe(II)′ generation can be associated with (i) the masking effect of Me on re-complexation of generated Fe(II)′ by EDTA and/or (ii) the increased photoreduction rate due to the competitive effect of Me on dissociation of the photochemically generated Fe(II)-ligand complex, as described in eqs 4 and 5, respectively. However, the former reaction is likely to have a minor effect on the steady-state Fe(II)′ concentration under the condition examined (i.e., [FZ]T = 1 mM, [L]T = 20 μM, and [Me]T = 0−20 mM) because the third term in the denominator of eq 9 (i.e., kf‑FZ[FZ]3 = 310 s−1) is calculated to be higher than the sum of the first and second terms (i.e., kf‑L[L*] + kf‑MeL[MeL] = 5.3−163 s−1). Thus, the increase in FeIIFZ3 formation due to the masking effect of Me on re-complexation of Fe(II)′ and ligand was determined to be small (e.g., by 1.5-fold at maximum) as Me increases from 0 to 20 mM, regardless of the facts that (i) thermodynamic calculations using the stability constants listed in Table 1 indicated that almost all of the free EDTA (>99%) forms complexes with Me at concentrations greater than 21 μM for Ca and 70 μM for Mg (also see Figure S2A,B), and (ii) an independent experiment suggests that FeIIEDTA complexation decreased by 15−26-fold in the presence of excess Me (Figure S3A). It is more likely that the latter reaction (i.e., enhanced photoreduction) substantially contributed to the observed increase in FeIIFZ3 formation. As described in eq 5, the presence of the ternary complex (FeIIIL-Me) was introduced to account for the Me-promoted FeIIFZ3 formation at higher Me concentrations with assumptions that (i) reductive dissociation of Fe from FeIIIL-Me is faster than that from the FeIIIL complex and (ii) its fraction relative to FeIIIL increases with increasing Me concentration. Under this assumption, the photochemical reduction rates were determined to be khν = 2.5 × 10−7 s−1 for FeIIIEDTA and khν,Me = (5.3−6.0) × 10−6 s−1 for the FeIIIEDTAMe complex by substituting the time-course data of FeIIFZ3 formation in the absence of Me (i.e., αMe = 0) and presence of excess Me (αMe = 1) in eq 9. The results indicate that the reduction rate for the FeIIIEDTA-Me complex was an order of magnitude greater (by 21−24-fold) than that for FeIIIEDTA (Figure S3B). Furthermore, the model fit of eq 9 to all the data shown in Figure 2A yielded stability constants for the ternary complexes (KFeL‑Me) that are 1 or 2 orders of magnitude lower than the stability constant of the MeL complex (KMeL, Table 1). The relatively lower value of KFeL‑Me is consistent with the notion that the metal-binding ligand in the ternary complex has a smaller number of sites available for coordination with incoming metals (i.e., Me).30,54 In addition, the lower stability of the ternary complex suggests that FeEDTA has a significant interaction with excess Me only when free EDTA is saturated with Me (i.e., [Me]T > [Lf], otherwise Me preferentially binds to free EDTA; also see Figure S2C,D). Overall, the proposed model provided reasonable fits to the observed data over the range of Ca and Mg concentrations examined (Figure 2A). Furthermore, the greater sensitivity of photochemical FeIIFZ3 formation for Ca compared to that for Mg is consistent with the thermodynamically higher affinity of Ca for EDTA (Table 1). In contrast to EDTA, Me-promoted FeIIFZ3 formation was not observed in the system buffered by SRFA (Figure 2B),

Details of the calculation procedure of substrate concentration, including sensitivity analysis (SI.5.1 and SI.5.2), comprehensive reaction set of Fe transformation kinetics (Tables S2 and S3), and definitions of all the parameters used in this study (Table S8), are provided in the Supporting Information. The significant reactions in the Fe uptake assay, extracted via the sensitivity analysis (SI.5.1), are depicted in Figure 1.

4. RESULTS AND DISCUSSION


4.1. Photochemical Fe(II) Formation. The photochemical experiments revealed that FeIIFZ3 formation from the FeIIIEDTA complex was significantly affected by Me concentration (Figure 2A). As the Me concentration increased, the FeIIFZ3 formation rate (k*hν) increased sharply (up to ∼50-fold) and reached a constant maximum value when the Me concentration reached a concentration comparable to or greater than the concentration of free EDTA (i.e., 20 μM).

Figure 2. Effects of Me concentration on the photochemical FeIIFZ3 formation in Fraquil* buffered by (A) EDTA (Ca, black squares; Mg, gray triangles; no Me, open diamond) and (B) SRFA. The photochemical experiments were performed under the conditions of total concentrations of 20 μM for Fe, 40 μM for EDTA, and 0−20 mM for Ca and Mg in the EDTA system and 5 μM for Fe, 50 mg·L−1 SRFA, and 0− 20 mM for Ca and Mg in the SRFA system. By assuming the first-order reaction as described by [FeIIFZ3]/dt = k*hν[FeIIIL*] (eq 10), the * , s−1) was photochemical FeIIFZ3 formation rate constant (khν determined via linear regression analysis in a plot of ln([Fe]T/([Fe]T − [FeIIFZ3])) versus time (where [Fe]T ≈ [FeIIIL] and [FeIIFZ3] are time course data of FeIIFZ3 concentration). Symbols and error bars represent average value and standard deviation from triplicate experiments. In panel A, solid and dotted lines indicate model fits of eq 9 to the data for the Ca and Mg cases, respectively. In the EDTA case, data plotted for log [Me] = −7 were used for the “no Me” case, while data plotted for log [Me] > −4 were considered as the “excess Me” case. 9138


Article



Figure 3. (A) 55Fe uptake rate in a range of Me concentrations. 55Fe uptake rates were measured in EDTA-buffered Fraquil* in the absence and presence of Me (Ca, black bars; Mg, gray bars). Total concentrations were adjusted to 0.5 μM for 55Fe, 20 μM for EDTA, 0−20 mM for Me, and ∼2 × 106 cell· mL−1 for cells. Averaged data with standard deviation from triplicate runs are shown. (B) 55Fe uptake rate as a function of unchelated Fe(II) concentration. The short-term 55Fe uptake assays were conducted in Fraquil* buffered by EDTA in the absence of Me (open diamond) and presence of excess Ca (total Ca concentration of 200 μM, open triangle). Total concentrations were adjusted to 0.05−1 μM for 55Fe, 20 μM for EDTA, and ∼2 × 106 cell·mL−1 for cells. In addition, 55Fe uptake rates measured in the SRFA-buffered system (0.5 μM for 55Fe and 10 mg·L−1 for SRFA) were also plotted, as indicated by closed symbols. The closed diamond and closed triangle represent Fe uptake rates for SRFA systems in the absence of Ca and presence of excess Ca (total Ca concentration of 200 μM), respectively. Symbols and error bars indicate average and standard deviation from triplicate experiments. Solid and dotted lines represent model fits of eq 14 to the data measured in the absence and presence of Me, respectively. As noted in SI.5 (eq S9), Fe(II)′ concentration was determined by the balance of light-mediated reduction of FeIIIL (including thermal reduction in the case of SRFA) and complexation of photogenerated Fe(II)′ by L. (C) Maximum Fe uptake as a function of Me concentration. 55Fe uptake rates were measured in the SRFA-buffered Fraquil* in the absence (open diamonds) and presence (Ca, black squares; Mg, gray triangles) of Me. Total concentrations were adjusted to 0.5 μM for 55Fe, 10 mg·L−1 for SRFA, 0−20 mM for Me, and ∼2 × 106 cell·mL−1 for cells. Symbols and error bars represent average and standard deviation from triplicate runs. Solid and dotted lines represent model fits of eq 15 to the data for the Ca and Mg cases, respectively. The arrows indicate Ca (gray arrow) and Mg (black arrow) ion concentrations in the Fraquil* medium in which cells were initially acclimated.

followed by a slight decrease at 20 mM Ca. In the SRFA system, Fe uptake rates were determined to be higher (by up to ∼10fold) than observed in the EDTA system, particularly at lower Me concentration (Figure S4). The uptake rate, however, monotonically decreased with increasing Me concentration by 5fold for Ca and 18-fold for Mg compared to the rate observed in the absence of Me. As noted below, the 55Fe uptake rates in the EDTA system were in most cases calculated to be below the saturated rate of uptake (e.g., by 1.3−12-fold for the Ca system, depending on the Me concentration), whereas the Fe uptake in the SRFA system is expected to be saturated. The Michaelis−Menten-type uptake theory indicates that uptake rate increases proportionally as substrate concentration increases in cases where [S] ≪ KS (e.g., ρS1/ρS2 = [S1]/[S2], where ρS = (ρmax S /KS)[S]), unless the Fe uptake parameters vary otherwise. Therefore, the relatively small increase in the cellular iron uptake rate compared to the increase in the Fe(II) formation rate (in the nonsaturated EDTA system) in addition to the decreased uptake at the higher Me

suggesting that (i) the rate of metal exchange between Fe(II)′ and Me complexed by SRFA is comparable to that for free SRFA and/or (ii) Me does not competitively bind to the Fe-binding sites of SRFA. Thermodynamic calculations (using KMeL values shown in Table 1) indicated that Me complexes with fulvic acids account for only a small portion of total fulvic acids, even at high concentrations of Me (e.g.,

Competitive Effects of Calcium and Magnesium Ions on the

Recommend Documents