Iron is an essential element for nearly all forms of life because of its ubiquitous role in redox enzymes, especially in the context of respiration and photosynthesis. However, although it is the fourth most abundant element in the Earth's crust, since the great oxygenation event, [1] it is present at near neutral pH only in the form of extremely insoluble minerals like hematite, goethite, and pyrite or as polymeric oxide-hydrates, -carbonates, and -silicates. Thus, the bioavailability of this element is severely restricted with soluble forms of iron in the environment generally found in the nM to pM range while most organisms require µM levels to survive. To deal with this low bioavailability, bacteria and fungi have evolved sophisticated systems based on high affinity Fe(III)-specific binding compounds called siderophores to acquire, transport, and process this essential metal ion. Their major role is the extracellular solubilization of iron from minerals and/or organic substrates and its specific transport into microbial cells. Many hundreds of siderophores, whose biosyntheses are repressed by high iron levels, are now known and their numbers continue to grow. Extensive studies of their isolation, structure, transport, and molecular genetics have been undertaken in the last three decades and have been comprehensively reviewed many times. [2][3][4][5][6][7][8][9][10][11][12][13] Indeed, the number of papers with the term "siderophore" in the abstract or title has seen exponential growth. Much of this interest has been driven by their importance in pathogenicity and virulence and the battle between microbe and host for this vital element which has been dubbed a biological arms race. [11,[14][15][16][17][18][19][20][21][22][23][24][25] In this review however we focus on a unique subset of siderophores that has only been recognized in the last 20 years, [26,27] namely those whose iron complexes display photoactivity. This photoactivity, which typically results in the photooxidation of the siderophore ligand with concomitant reduction of Fe(III) to Fe(II) seemingly upsets the siderophore paradigm of forming and transporting only extremely stable Fe(III) complexes into microbial cells. Here we review the structure, synthesis, photochemistry, photoproduct coordination chemistry of these siderophores and explore the potential biological and ecological consequences of this photoactivity.
Siderophores are generally classified into four groups depending on the major iron binding ligands employed. Thus, there are the catecholates, the hydroxamates, the carboxylate and mixed ligand siderophores (Figure 1). As the photoactivity of the iron complexes of siderophores stems from the presence of a-hydroxy acid groups in the ligand, in this review we are concerned only with the carboxylate siderophores and those mixed ligand complexes with a-hydroxy carboxylate coordination (Table 1).
Organism Known or Suspected Reactive Group Ref.
Acinetobacter haemolyticus suspected central citrate [28] Aerobactin E. coli; Vibrio sp. DS40M5; etc known central citrate [27] Petrobactin Marinobacter hydrocarbonoclasticus known central citrate [29] Ochrobactins Marine vibrio suspected central citrate [30] Schizokinen Bacillus megaterium suspected central citrate [31] Staphyloferrin Staphylococcus hyicus known central citrate [32] Synechobactins Synechococcus PCC 7002 known central citrate [33] Achromobactins Erwinia chrysanthemi suspected terminal/central citrate [34] Rhizobactin Rhizobium meliloti suspected terminal citrate [35] Rhizoferrin Rhizopus microsporus suspected terminal citrate [36] Vibrioferrin Vibrio parahaemolyticus known terminal citrate [37] Woodybactin Shewanella woodyi known terminal citrate [38] Alterobactins Alteromonas luteoviolacea known β-hydroxyaspartate [39] Aquachelins Halomonas aquamarina known β-hydroxyaspartate [26] Chrochelin Azotobacter chroococcum suspected β-hydroxyaspartate [40] Corrugatin Pseudomonas corrugata suspected β-hydroxyaspartate [41] Cupriachelin Cupriavidus necator H16 known β-hydroxyaspartate [42,43] 1, Figures 1 and2). The photoproducts and subsequent iron binding capabilities of each of these classes differ, leading to potentially different biological and ecological effects which we consider later.
Siderophores are generally synthesized within the bacterial cell and secreted into the environment as iron-free compounds. Their synthesis is linked to pathways involving nonribosomal peptide synthetases (NRPS) and NRPS-independent synthetases (NIS). The NRPS have multiple domains where adenylation, thiolation and condensation reactions take place in a coordinated, sequential fashion to link amino acids via thiol-based intermediates. While the details of NRPS pathways are highly nuanced, they share a common modular biosynthetic logic, utilizing covalent acyl-enzyme intermediates, iterative cycles of condensation and tailoring of monomer units to produce great structural and functional diversity. NRPS siderophores include for example yersiniabactin, vibriobactin and mycobactins, with characteristic oxazoline or thiazoline rings. While NRPS enzymology and structural biology has been extensively investigated in the past 20 years, [61,62] research on the other important type of siderophore biosynthesis enzymes, the so-called NRPSindependent siderophore synthetases (NIS), has only very recently picked up.
The photochemistry of Fe(III) with small-molecule hydroxy acids has been well studied for a variety of compounds. One of the oldest known photochemical reactions involves Fe(III) and citrate (ferric ammonium citrate) where photo-reduction to Fe(II) is the first step in the cyanotype or blueprint process. [69,70] After light excitation into a ligand to metal charge transfer (LMCT) absorbance band, there is an electron transfer from the a-hydroxy acid ligand to the Fe(III) (Scheme A).
This electron transfer gives rise to a carboxyl radical complex that typically undergoes bond cleavage of the C-C bond to release CO2 and form a C-centered radical. Further reaction of this radical leads to generation of a variety of ketone and hydroxyl photoproducts, depending on the structure of the carboxylate ligand as well as pH (Scheme A). [71][72][73][74] This reaction occurs in the solid state (Figure 3b), as well as aqueous solution, and the color change can be easily monitored via UV-Vis spectroscopy, as irradiation leads to depletion of the LMCT band (Figure 3a). In addition, the generation of CO2 can be monitored in the solid state via IR spectroscopy (Figure 4c). The efficiency or quantum yield (QY or F) for reduction of the Fe(III) to Fe(II) can also be quantified, [75] and is dependent on both the structure of the ligand, and the stoichiometry and stereo/regiochemistry of the Fe(III) species. [76,77] For example, the QY of Fe(III)-lactic acid is relatively high, at 0.137 ± 0.002 (Table 2). [78] When the hydroxy group from the alpha carbon is moved, however, the QY drops significantly, such that the QY of Fe(II) formation after irradiation of Fe(III) with a-hydroxy butyric acid is only 0.009 ± 0.005. [78] This same trend is seen for malic acid compared to succinic acid, where the presence of the hydroxy group in malic acid gives a larger QY (F 0.26) compared to that of succinic acid, which lacks the hydroxy group and has a lower QY (F 0.13). [72] It is expected that this is primarily due to the stabilization of
hv the radial by the O bound to the alpha carbon in lactic acid and malic acid, whereas this stabilization would be absent in the β-hydroxy butyric acid or succinic acid. It is clear the hydroxy group contributes to the efficiency of this reaction as well, since the simple Fe(III) acetate is not photoreactive. [72] The stereochemistry of the hydroxy group also plays a role in the photoreaction, where changes in the location and projection of the hydroxy group changes the QY. The stereochemical effect on quantum yield is particularly evident in the photolysis of ferric tartaric acid where the chiral L(+)-tartaric acid isomer exhibits a much larger QY (F 0.58) than the achiral meso-tartaric acid (F 0.37) (Table 2). While the simple hydroxy acids have been well studied, it is now clear that carboxylatecontaining polysaccharides or polyuronic acids, can also form photoactive complexes with Fe(III). [77][78][79] The presence of the oxygen in the furanose ring can also stabilize the radical, as with the hydroxy group, leading to efficient photochemistry and reduction of Fe(III) to Fe(II). As with the hydroxy acids, the efficiency of the photoreaction (QY) changes depending on the stereochemistry of the uronic acids (Table 3). [78] This trend is especially clear in the case of alginate, a polyuronic acid from brown algae that is a co-polymer of mannuronate (M) and guluronate (G) uronic acid units. [80] The G and M units are just epimers of each other, but there is still a change in the QY, even with small changes in the %M in the polymer chain (Table 3).
The alginate with the highest mannuronate (65% M) has a QY of 0.109, where the alginate with the lowest mannuronate (33%M) has an order of magnitude lower QY of 0.018 (Table 3).
Depending on the amount, and location of the uronic acids, the QY is different as well. Pectate, a homopolymers of galacturonic acid shows in general a lower QY (1.7%) than the alginate, where only the lowest %M content alginates have similar QY. When comparing pectate (100% carboxylate content) with that of pectin, where ~20% of the carboxylates are esterified, there is actually a significant increase in the QY, with an efficiency of photoreaction of 1.7% for pectate and 2.7% for pectin. A variety of D-glucuronic acid-containing polysaccharides like Xanthan gum, hyaluronic acid, and Arabic gum have also been investigated. Especially when comparing Xanthan gum and Arabic gum, the differences in QY are apparent, even for approximately the same uronic acid content (only 0.11% for Xanthan gum and nearly 1% for Arabic gum.
* a unpublished data, pH 4.5, I0= 3.7x10 -8 [81] For all of the polyuronic acids, the QY decreased when going to lower concentration of Fe(III), but keeping the ratio of polysaccharide to Fe(III) the same (Table 3).
The different values of QY within the same family of alginates indicate that stereochemistry is also not the only factor that influences the efficiency of the photoreaction. The differences in reactivity are hypothesized to be due to changes in the interaction of the polysaccharide with Fe(III) ions. [81] Mössbauer spectroscopy reveals differences in the Fe(III) species when the Fe(III)-alginate (61%M) gels were compared to the Fe(III)-pectate gels. [77] The most abundant Fe(III) species for both the alginate and pectate was Fe(III)-oxo nanoclusters, and these nanoclusters were the first to undergo photoreduction to Fe(II). [77] There were differences, however, in the overall speciation of alginate vs. pectate gels, where alginate gels showed 60% Polysaccharide (FFe(II) ± SD)x10 2 Uronate content (%)
1% PS/ 0.9 mM Fe(III) 0.25% PS/0. [77] This suggests that differences in the QY can also explained by changes in the Fe(III) speciation, and how it is coordinated to the carboxylate and hydroxyl groups.
[81] The concentration of Fe(III) and pH are also very important then in the final photoreactivity, since these can both change the speciation (tris vs. bis Fe(III) complexes) as well as coordination environment (protonated vs. deprotonated carboxylate).
Siderophores containing citric acid and other a-hydroxyl carboxylic acids show this same photochemistry, with reduction of Fe(III) to Fe(II). Since the quantum yield and photoproducts of this reaction depend on the structure and Fe(III) speciation, it is expected then that with the diversity of siderophore structures, each siderophore will have different sensitivities to light, as well as different potential photoproducts, depending on overall structure, stereochemistry, pKa, aqueous pH, and concentration of Fe(III).
While it is clear that both mechanistically and phenomenologically the photolysis of the Fe(III) complexes of the photoactive siderophores leads to oxidative decarboxylation of the ligand with concomitant reduction of Fe(III) to Fe(II), the exact chemical nature of the photoproduct has only been determined in a few cases. This is important in that the structure and iron binding ability (or lack thereof) of the ligand has important ramifications with respect to their ultimate biological fate. [26,82] It has been found that the structure of the ligand photoproduct depends strongly on whether the a-hydroxy acid group is present in the terminal or bridging position and whether it is present as a citryl or b-hydroxyaspartate unit.
The best studied and characterized examples come from the siderophores that have a central or bridging citrate group such as aerobactin (Figure 1) and petrobactin (Figure 2). Initially, based solely on mass spectral (MS) data, the decarboxylation of the central citryl moiety in photolyzed petrobactin was proposed to produce a central a-ketoglutarate moiety. [29] A more detailed NMR and MS study of the related aerobactin photoproduct confirmed the transformation of the citryl moiety into an 3-ketoglutarate.
[27] However, while the keto form predominates in DMSO solution, in water the data indicated a keto-enol tautomerization with the enol form being the most important. The presence of the predominant enol form in water with a pKa around 10.8 results in the photoproduct being still able to coordinate Fe(III) in a stable octahedral geometry through both the two terminal hydroxamate groups but also by the central anionic, delocalized five member chelate ring structure formed, in a coordination mode reminiscent of acetylacetonate coordination.
This structure (Figure 4) confirms earlier speculation i.e. "that an enolate moiety of the 3ketoglutarate fragment in the photoproduct could also coordinate Fe(III), given that the metalcoordinative properties of α-ketoenolates are well known". [27] Measurements of the Fe(III) ligand stability constants give a value of log KML for the photoproduct of 28.6 while that of the starting aerobactin complex is 27.6. [27] The surprising similarity between the two numbers is a result of the replacement of one coordinating moiety, i.e., the citrate group in the parent siderophore with a new coordinating ketoenolate group in the photoproduct. The similarity between the two groups is exemplified by their similar pKa values both of which are near 11. It is remarkable that photolysis of the iron complex of aerobactin leads to a photoproduct that is actually a better Fe(III) chelator than the parent. This conclusion is further supported by the results from studies of the related siderophore, petrobactin, where essentially the same photochemistry occurs leading to a Fe(III) photoproduct that is again more stable than the parent (pM for petrobactin of 24.4, while that of petrobactin* is 23.0). [83] In contrast, when the a-hydroxyacid(s) are in a terminal position as in vibrioferrin (Figure 2), or rhizoferrin (Figure 1) the results are very different. Although only for vibrioferrin were details of the photoproduct structure and its (lack) of Fe(III) binding capacity examined in detail. [84] Vibrioferrin was first isolated and characterized by Yamamoto et al. from Vibrio parahaemolyticus and consists of coupled 2-ketoglutaric acid, alanine, ethanolamine and citric acid fragments. [37] In water it exists in both open chain and cyclized forms with the latter predominating. It binds iron via the two a-hydroxy acids and the free citrate carboxyl oxygen which together however are insufficient to satisfy the six-fold octahedral environment preferred by high spin Fe(III). The remaining iron binding site is presumably occupied by water. The only pentadentate coordination by the ligand thus leads to relatively weak iron binding as compared to other documented siderophores (Table 4). The Fe-vibrioferrin complex was shown to be unusually photoreactive with a half-life under some conditions of less than 6 min and a rate of photolysis some 10 to 20 times faster than other photoactive siderophores measured (Table 5). [84] The photoproduct was isolated and characterized and its structure is shown in Figure 5 where only the hydroxy acid that was part of the original five membered a-ketoglutarate ring having been decarboxylated and leaving behind a keto group as part of a succinyl ring. The terminal citrate group remained untouched even after extensive photolysis. only free terminal a-hydroxy acid groups, such as rhizoferrin, would lead to photoproducts that do not bind iron as rhizoferrin photochemistry seems not to have been investigated so far. The related siderophore staphyloferrin B, which has both a central and terminal a-hydroxy acid groups, has been studied and is postulated to undergo decarboxylation only at the central citryl unit. [92] Unfortunately neither structural data for the photoproduct nor its Fe(III) coordination properties or stability were reported leaving this intermediate situation unclear.
At this juncture it is worth pointing out that different authors use differing conventions for expressing iron siderophore stabilities making comparisons at times difficult. However, the three most common measures of Fe-Ligand stability are the formal stability constants log KML or alternatively β110, which depend on the convention that the ligand is in its fully deprotonated form.
This in turn requires knowledge of all the ligand pKa values, and the conditional stability constants log K cond (FeL, Fe') which measure stability under specific conditions, i.e. pH, and also consider the dissolved ferric hydrolysis species often referred to as Fe(III)' under these conditions, a terminology preferred by the oceanographic community. More generally there is the so called pM(etal) or pFe value which is a calculation of -log [Fe], the amount of free iron equilibrium under a specific set of conditions typically pH 7.4, [L] = 10 µM and [Fe] = 1 µM. The advantage of this latter measure is that it is independent of the stoichiometry of the complex formed and thus allows comparison between siderophores that form 1:1 complexes with iron and those that do not.
The photochemistry of siderophores containing the b-hydroxyaspartate moiety is more complicated than that for the citrate siderophores which undergo more or less simple oxidative decarboxylation of the ligand. Photolysis of the Fe(III) complexes of the b-hydroxyaspartate containing peptidic siderophores generally leads to the cleavage of the peptide backbone, although this cleavage likely results from multiple photooxidation events. [93] In a few cases an initial decarboxylation of b-hydroxyaspartate is observed, which is similar to photooxidation of citratecontaining siderophores. Upon sunlight photolysis of Fe(III) variochelin A (Figure 2), [60] two photoproducts are observed. One photoproduct corresponds to loss of CO2 from b-OH Asp forming the aldehyde. The other photoproduct of photolysis of Fe(III)-variochelin A is the formamide comprised of the fatty acid, Arg and Ala. [60] As observed for other b-hydroxyaspartate siderophores, [42,43,58] sunlight driven photolysis of Fe(III)-aquachelins leads to loss of the entire b-hydroxyaspartate group, along with the fatty acid tail giving a more hydrophilic peptidic photoproduct characterized only by MS and UV-Vis.
[26] The proposed structure of the initial photoproduct is shown in Figure 4 as determined by tandem MS. Importantly, it is capable of rebinding Fe(III) via at least the two remaining hydroxamate groups. What other ligands (if any) might be involved in the coordination of the iron photoproduct is not known. However, the conditional stability constant of the photoproduct with Fe(III) of log K = 11.5 is somewhat less than that that of the parent aquachelin C of 12.2, suggesting in contrast to the citrate siderophores aerobactin and petrobactin, a significant loss of stability, likely through the loss of some chelating groups. [26] Continued photolysis under aerobic conditions in the presence of iron leads to further degradation of the ligand into as yet unidentified products. Baldwin et al in studying model compounds has suggested that the differing photochemical outcomes between the citrate and b-hydroxyaspartate containing siderophores is likely due to differences in susceptibility of the decarboxylation product to air oxidation in the presence of iron. [94] Decarboxylation of the citrate moiety produces a ketone, which is resistant to further air oxidation even in the presence of iron. However, decarboxylation of the βhydroxyaspartate moiety produces an aldehyde, which, in the presence of iron and oxygen, is oxidized to a carboxylate, which can undergo further photolysis when coordinated to Fe(III). The authors predict that if the β-hydroxyaspartate-based marine siderophores were irradiated under anaerobic conditions, simple decarboxylation of the alpha-hydroxy acid group to the aldehyde would be observed. However, the reported studies on photolysis of marine siderophores all appear to have been conducted in air.
Three possible outcomes follow photolysis of photoactive siderophores: a) formation of a photoproduct that binds Fe(III) as well or better than the parent, b) formation of a photoproduct that binds Fe(III) more weakly than the parent, and c) formation of a photoproduct that no longer has any affinity for iron. These outcomes were shown to depend both on the nature of the a-hydroxy acid present and the location of the a-hydroxy acid in the structural formula. One could ask the question: is incorporation of these groups into siderophores an evolutionary "accident" or are there "good reasons" for their presence? Of course, for bacteria living in dark environments where photolysis is absent or unlikely, these siderophores are no different than "normal" siderophores and can be expected to function via the usual paradigm, i.e., acting to solubilize and transport iron into the cell. However, in environments exposed to light, the metabolic cost of producing a molecule that is likely to be degraded or converted into a functionally different product requires a strong ecological/evolutionary pressure to do so. [95] [96] There is evidence in the marine environment at least that both strategies are used by different populations of bacteria. Thus, for example, vibrioferrin biosynthetic genes are found to be highly expressed in bacteria living both near the ocean surface where photolysis is extremely likely and at depths below the photic zone where such photolysis is impossible. [97] It was found that despite production of the same vibrioferrin siderophore, the producing bacterial community statistically differed in the photic and dark zone of the marine water column, suggesting very different environmental pressures on marine bacteria. In this section we discuss the different photochemical outcomes and examine their potential biological, ecological or evolutionary effects or functions. Notable possibilities include increased or altered metal binding affinity and/or specificity, a strategy to prevent "cheaters", the conversion into signaling molecules or other ecological outcomes.
Formation of photoproducts that bind iron(III) as well or better than the parent siderophore. While carboxylate ligands are generally less effective as iron chelators than the more commonly employed hydroxamates or catecholates, their incorporation into siderophores may have originally been driven by the fact that they are widely available metabolites. In addition, as the pH of the environment is lowered, their stability disadvantage compared to the other ligating groups begins to disappear. Nevertheless, in general, intact a-hydroxy carboxylic acid containing siderophores remain among the weaker siderophores (Table 4). Thus, formation of photoproducts that bind iron better than the parent siderophore can be viewed as a metabolically "inexpensive", environmentally induced, post translational modification that turns a weak siderophore into a better one, which process would seem to be evolutionarily advantageous. Such an advantage presupposes of course that the photoproduct iron complex is recognized and transported into the microbe as efficiently as the parent Fe(III)-siderophore. Unfortunately, there has been surprisingly little work on the receptor recognition, transport and iron release from photoproduct iron complexes.
However, in the few cases we are aware of, it was shown that both Fe-aerobactin and its Fephotoproduct are transported equally well by Vibrio sp. DS40M5, an aerobactin producer. [27] Both, Fe(III)-aerobactin and Fe(III)-photoproduct, displayed kinetics consistent with receptor mediated iron uptake with similar Km (ca. 0.5 µM) and Vmax values and inhibited each other's uptake in an apparently competitive manner. [27] While no crystal structure of the aerobactin photoproduct iron complex (or indeed that of any siderophore photoproduct) has been forthcoming, molecular mechanics and DFT structure optimizations show that while the iron coordination sites are very similar, the backbone structures are different, suggesting that receptor recognition mainly involves the iron center (Figure 5) [Cooksy, Carrano, Gardner unpublished].
Additionally, the petrobactin photoproduct has been shown to be taken up by the producing bacteria and to bind to the parent receptor, YclQ, with only slightly reduced affinity (Kd FePB = 113 nM Kd FePB* = 142 nM). [98] Formation of a photoproduct that no longer has any affinity for iron. At the opposite end of the spectrum, siderophores such as vibrioferrin are both extremely photosensitive and result in a photoproduct that no longer can bind iron. Here there must be a strong ecological/environmental/evolutionary pressure at work since biosynthesis of a siderophore is a metabolically expensive process and its primary function, only to be rapidly diminished, is more difficult to place into an ecological context. What then could be the reason for a bacterium whose environment is in the photic zone for producing such a siderophore?
Many bloom-forming algal species, including climatically important dinoflagellates and coccolithophores, are known to form beneficial or obligatory close associations with certain bacteria, but the reasons for these associations remain obscure. [99][100][101] The bacteria that closely associate with phytoplankton are believed to be involved in a wide range of reciprocal interactions, including metabolite/nutrient uptake, provision and remineralization, cell differentiation, as well as algicidal or bacterioprotective effects. Among the most intriguing factors that may involve algal-bacterial interactions is iron acquisition since eukaryotic phytoplankton are not known to produce siderophores or in general to be able to directly take up bacterially derived Fe(III)siderophore complexes. [102] Indeed, iron is the limiting factor of primary productivity in large regions of the world's oceans, the so-called high nutrient low chlorophyll (HNLC) zones. [103] However, some eukaryotic phytoplankton are able to access iron from weak siderophores or other chelates via ferrireductases and adjacent Fe(II) transporters on their outer cell membranes. [108] In recent studies it was found that vibrioferrin producers dominated among the algalassociated clades of the genus Marinobacter, and it was proposed that the bacteria provide the algae with bioavailable iron through sunlight activated photolysis of vibrioferrin whereas the algae provide the bacteria fixed carbon in the form of leaked photosynthate in a form of iron-for-carbonmutualism. [82] It was subsequently shown that while intact iron-vibrioferrin is not accessible to the bloom-forming dinoflagellate Scrippsiella trochoidea in the dark, iron is rapidly released from the siderophore complex and taken up in the presence of sunlight. At the same time, it was observed that the bacterium itself, which grew in unamended seawater only in the presence of the algae, could take up iron both from the parent vibrioferrin as well as from photolyzed vibrioferrin.
Surprisingly it took up iron from VF* at a rate 70% higher than observed with unphotolyzed Fe(III)-vibrioferrin! [82] Such observations are consistent with the carbon-for-iron-mutualism hypothesis. [83] It should be noted that Fe(II) initially produced by vibrioferrin photolysis is rapidly oxidized to Fe(III) in aerobic seawater and subsequently hydrolyzed into increasingly refractory mineral phases (iron hydroxides) that are no longer soluble and nor bioavailable. Thus when a solution of 55 Fe(III)-vibrioferrin is completely photolyzed and then aged in artificial seawater at pH 8.1 for various lengths of time, uptake of iron by Scrippsiella trochoidea was found to decrease rapidly with increasing time of "aging", suggesting that the initially formed Fe(II) or Fe(III)' became increasingly bio-unavailable. [Amin and Carrano unpublished] This implies that for the iron-for-carbon-mutualism mechanism to be operative, a close association of the bacteria and algae is needed lest the bioavailable iron diffuse away and becomes remineralized.
Because the photolysis completely destroys vibrioferrin, and the vibrioferrin photoproduct shows no significant affinity for either Fe(II) or Fe(III), the vibrioferrin-producing bacteria must be able to acquire iron from photochemically released Fe(II) or Fe(III)'. This provides another intriguing ecological concept involving the interplay between bacterial producers of "public goods" and iron consumers that "cheat" on the siderophore-producing bacteria in complex communities. In microbial cooperation, public goods can be loosely defined as compounds or functions that provide a collective benefit, generally through release into the extracellular environment. [104,105] The defining characteristic of public goods is that they are costly to produce for the individual and are beneficial to all members present in the local ecological niche.
However, public goods can be exploited by "cheaters," which do not bear the cost of production but nonetheless can reap the benefits. Siderophore production can be considered a prime example of a public good and has been widely studied in this regard. [104,105] It has been noted that natural populations of bacteria involve coexisting non-clonal but close relatives with similar resource preferences and can have hundreds of unique genes per genome. [104] It has been found that some of these related bacteria have lost the ability to produce siderophores but still retain the receptor(s) and uptake mechanisms for iron-siderophore complexes. Because any cell that carries specific outer-membrane transporters can take them up irrespective of siderophore production, "cheaters" become part of the population. One potential way to thwart siderophore cheating or pirating would be to change the produced siderophore into a form where it could not be taken up by cells that only expressed its parent cognate receptor. Photolysis of vibrioferrin could be an example where just having the receptor, pvuA, would not allow for cheating since the photoproduct does not resemble the parent siderophore and no longer binds iron or presumably the vibrioferrin receptor. The uptake of iron from photolyzed vibrioferrin thus requires that any cheater bacteria possess a siderophore independent Fe(III) uptake system, such as an Fe(II) uptake system like the Feo system [106] or the FbpABC transporter, [107] which utilizes dissolved ferric hydrolysis species, and other partially chelated forms of ferric ion, potentially making "cheating" more difficult and therefore less likely.
Formation of a photoproduct that binds iron weaker than the parent. The situation as represented by the many b-hydroxyaspartate-containing siderophores, exemplified by the aquachelins, is in many ways the most puzzling. However, it was this class of photoactive siderophore that was first proposed to be involved in iron cycling and increasing its bioavailability in the ocean. [26] Dissolved Fe(III) in the upper oceans occur almost entirely in the form of complexes with organic ligands. [108] Although the importance of organic ligands to aquatic iron cycling has become clear, their nature has remained obscure. Nevertheless, based on similar conditional stability constants, siderophores have long been proposed to be one of the components.
In this regard it was shown that photolysis of aquachelins was able to increase the bioavailability of iron to a natural assemblage of phytoplankton. [26] As expected, the intact aquachelin B complex was largely unavailable to the natural assemblage while uptake of iron from previously photolyzed aquachelin B was indistinguishable from incubations in which inorganic iron was added. This is a surprising result since the aquachelin photoproduct can still bind iron reasonably well, the Fe(II) produced by the reaction will be rapidly oxidized and recomplexed. [26] Thus the Fe(II) produced would be expected to be bioavailable for only a short period of time assuming that the Fe-siderophore photoproduct itself is not able to be taken up. However, while the Fesiderophore photoproduct could be expected to be structurally similar to the parent siderophore, the loss of the b-hydroxyaspartate-chelating moiety results in a lowered stability of the complex which in turn makes its redox potential less negative. This is because there is an inverse relationship between the stability of Fe(III) siderophore chelates and their ease of reduction as high chelate stability energetically favors the maintenance of the bound iron in the +3 oxidation state and discourages its reduction to Fe(II). Since many phytoplankton acquire iron from the environment via a cell surface reductase mechanism, [109] anything that decreases the negative redox potential of an iron-siderophore would make it more accessible to this class of organisms.
Thus, a weakened siderophore is a more accessible one. While it is easy to see what the natural assemblage of organisms gains by photolysis the aquachelin, it is much less obvious how it is of benefit to the producing bacterium. The three possible outcomes of photolysis of siderophores and their effects on both the producer and other cohabitaing organisms are summarized in Figure 6.
Another potentially important ramification resulting from the photolysis of the many bhydroxyaspartate containing siderophores that exist as suites with various length fatty acid tails so prevalent among marine siderophores, is the loss of the of the fatty acid tail, when the fatty acid resides next to b-hydroxyaspartate. Although the exact purpose of the fatty acid tail remains somewhat of a mystery, it is clear that the loss of the tail after photolysis will convert a relatively non-polar parent siderophore into one that is much more polar, a result that will almost certainly profoundly affect the biological fate of the photoproducts.
Finally, the more complex photochemistry that accompanies the photolysis of bhydroxyaspartate containing siderophores also leads to the possibility of formation of other radical species such as OH . which can themselves have important biological consequences. Indeed, such has been demonstrated in cloud waters, a unique ecological niche where photoactive siderophores have been detected and whose photochemistry has the potential to affect the oxidizing capacity of the atmosphere. [110] Clouds play a key role in the atmospheric system and may influence the troposphere composition and thereby have a strong impact on the climate.
While photoactive siderophores have been recognized now for some decades there remain very large gaps in our knowledge. For example, amazingly little work has been done thus far with regard to the potential uptake, transport and release of iron from photoproducts by the producing organisms. This is clearly an area of need. Likewise the role of these photoproducts and the released Fe(II) to other cohabitating organisms is an area ripe for more detailed study, particularly
within the context of the potential biological effects of ocean acidification (OA) and climatechange. Photoactive siderophores, which were first proposed to affect the photochemical cycling of iron in the ocean and changing the biological availability of this vital element, are likely be affected by the projected 0.4 pH unit decrease projected over the next few decades. Since photolysis rates increase sharply as pH decreases, OA is likely to substantially increase photochemical iron cycling and thus affect the speciation and bioavailabilty of iron to nonsiderophore producing organisms and hence deserves further study. Another potential area of research is in exploring the possible role of siderophore photoproducts as signaling molecules or infochemicals in the context of within-and cross-kingdom signaling. [111,112] Siderophores themselves have recently been found to have roles beyond just iron binding and transport with increasing awareness of their function in a signaling capacity primary among them. [111][112][113] While the fate of siderophores whose photoproducts lead to better iron chelators than the parent is obvious, what happens to the photoproducts that either bind iron more poorly than the parent or even more so those that do not rebind iron at all, remains largely unexplored. However, it seems clear that these metabolically expensive photoproducts must have some important biological role as it is hard to imagine that they would just go to waste! Thus, a signaling role seems likely and should be explored. Finally, exploration of the biological and or chemical effects of siderophore photochemistry in unique ecological niches such as cloud and rain waters are also deserving of further study.
While the photoreactivity of many Fe(III)-siderophores has been known for at least 20 years and given their startling prevalence in the marine environment, our understanding of the structures and more importantly the biological activity of their photoproducts remains in its infancy. In this review, we highlight some of these issues and draw attention to the many exciting research possibilities available in the future.