Adsorption of Enterobactin to Metal Oxides and the Role of

ARTICLE pubs.acs.org/Langmuir

Adsorption of Enterobactin to Metal Oxides and the Role of Siderophores in Bacterial Adhesion to Metals Hamish G. Upritchard,† Jing Yang,‡ Philip J. Bremer,§ Iain L. Lamont,† and A. James McQuillan‡,* †

Departments of Biochemistry, ‡Chemistry, and §Food Science, University of Otago, P.O. Box 56, Dunedin, New Zealand ABSTRACT: The potential contribution of chemical bonds formed between bacterial cells and metal surfaces during biofilm initiation has received little attention. Previous work has suggested that bacterial siderophores may play a role in bacterial adhesion to metals. It has now been shown using in situ ATR-IR spectroscopy that enterobactin, a catecholate siderophore secreted by Escherichia coli, forms covalent bonds with particle films of titanium dioxide, boehmite (AlOOH), and chromium oxide
hydroxide which model the surfaces of metals of significance in medical and industrial settings. Adsorption of enterobactin to the metal oxides occurred through the 2,3-dihydroxybenzoyl moieties, with the trilactone macrocycle having little involvement. Vibrational modes of the 2,3-dihydroxybenzoyl moiety of enterobactin, adsorbed to TiO2, were assigned by comparing the observed IR spectra with those calculated by the density functional method. Comparison of the observed adsorbate IR spectrum with the calculated spectra of catecholate-type [H2NCOC6H3O2Ti(OH)4]2
and salicylate-type [H2NCOC6H3O2HTi(OH)4]2
surface complexes indicated that the catecholate type is dominant. Analysis of the spectra for enterobactin in solution and that adsorbed to TiO2 revealed that the amide of the 2,3-dihydroxybenzoylserine group reorientates during coordination to surface Ti(IV) ions. Investigation into the pH dependence of enterobactin adsorption to TiO2 surfaces showed that all 2,3-dihydroxybenzoyl groups are involved. Infrared absorption bands attributed to adsorbed enterobactin were also strongly evident for E. coli cells attached to TiO2 particle films. These studies give evidence of enterobactin
metal bond formation and further suggest the generality of siderophore involvement in bacterial biofilm initiation on metal surfaces.

’ INTRODUCTION The adhesion of bacteria to metal surfaces leading to formation of bacterial cell communities (biofilms) represents a major problem in diverse settings. Biofilms contribute significantly to infections associated with surgical implants,1,2 are a source of contamination in food processing environments, and can enhance corrosion and/or fouling in industrial and aquatic environments.3,4 The development of biofilms following surface colonization has been quite well characterized.5,6 In contrast, initiation of biofilms is poorly understood, and in particular, the potential contribution of chemical bond formation between bacterial cell components and metal surfaces requires investigation. In Gram-negative bacteria, the complex mixture of surficial bacterial cell components includes lipopolysaccharides, membrane proteins, polymeric proteins, and siderophores. Among these species, the bacterial siderophores have the most marked ability to sequester metal ions and are expected to have great affinity for metal oxide surfaces. Bacterial adhesion has generally been described using Derjaguin
Landau
Verwey
Overbeek (DLVO) theory,7 based on Lifshitz
van der Waals attractive forces and electrostatic interactions.8 While the initial DLVO model has been further expanded to include steric and hydrophobic interactions,9 other relevant interactions such as covalent bonding are yet to be included. The absence of these important interactions may in part explain why experimental results r 2011 American Chemical Society

have not been in agreement with bacterial adhesion kinetics modeled on electrostatic and dispersive interactions alone.8 In general, metal surfaces in contact with the atmosphere or in aqueous environments, including titanium, iron, aluminum, and stainless steel, are covered with a thin film of metal oxide. Therefore, contact between bacteria and metal surfaces generally involves interaction with metal oxides. The interfacial chemistry of bacteria and wet metal oxide colloidal particle films can be studied in situ using attenuated total reflectance infrared (ATRIR) spectroscopy,10
12 and this technique is the primary experimental approach in the present work. Escherichia coli bacteria are ubiquitous in the environment, and pathogenic strains are notable for their ability to cause morbidity in humans and animals.13,14 E. coli biofilms on metal surfaces are of increasing concern in the food processing environment and have been linked to infection outbreaks and sporadic illness worldwide.15,16 The ability of E. coli to survive and grow in diverse habitats is dependent in part on an efficient iron-acquisition system centered on a siderophore, enterobactin. Siderophores are secreted by many bacteria and are organic chelators with very high specific affinity for iron(III) (association constants, K > 1020 L mol
1).17 The resulting ferri
siderophore molecules re-enter Received: January 27, 2011 Published: July 11, 2011 10587

dx.doi.org/10.1021/la202352j | Langmuir 2011, 27, 10587–10596

Langmuir

Figure 1. Schematic structure of the enterobactin siderophore comprised of a cyclic triester of 2,3-dihydroxy-N-benzoyl-L-serine. The structure shown was adapted from Raymond et al.27.

the bacteria via specific cell
surface receptor proteins. The metabolism of this potent iron chelator is tightly controlled by the Fur (Ferric uptake regulator) protein.18,19 Siderophores display a high structural diversity, but most bind Fe3+ ions through hydroxamate, catecholate, or hydroxycarboxylate groups.17,20 Following its isolation in 1970,21,22 enterobactin has been extensively studied with regard to solution thermodynamics, microbial transport, catecholate siderophore synthesis, and receptor recognition.20,23 Enterobactin forms with iron the most stable complex reported (Kf = 1049) with an associated pM value of 35.5 at pH 7.4,20 where pM =
log[Fe3+] at total Fe3+ and ligand concentrations of 1  10
6 and 1  10
5 mol L
1, respectively. The enterobactin molecule has 3-fold symmetry and is composed of three 2,3-dihydroxybenzoic acid groups, with each having an appended L-serine moiety. The three serines form a trilactone macrocyclic ring (Figure 1). Metal coordination at neutral pH occurs through the six catecholate oxygens in a Δ-cis complex, creating a metal-centered hexadendate complex.24 Raymond et al.25
27 proposed that enterobactin is predisposed for metal binding with the o-hydroxy proton of the free ligand hydrogen bonded to the amide oxygen atom. Upon deprotonation (or metal complexation), this conformation adopts the trans form, in which the amide proton hydrogen bonds to the ohydroxy oxygen atom. The intramolecular hydrogen bonds consequently serve to lock the catecholate groups into one of the two rigid conformations.27 Following investigations into ferric enterobactin protonation, an alternative geometry has been proposed in which the coordination shifts from catecholate to salicylate geometry around the metal center. In this mode of bonding, following protonation at the m-hydroxyl oxygen the Fe(III) shifts from the two catecholate oxygens to the o-hydroxyl oxygen and the amide oxygen.23,28 Gutierrez et al.29
31 also used 1 H NMR and computer modeling to study the conformation of enterobactin and its derivatives in solvent systems. From these analyses, the iron-free enterobactin (apo-enterobactin) was shown to adopt either a pseudoaxial or a pseudoequatorial conformation in solution. For the initial binding and active transport of enterobactin
metal complexes across the outer membrane, E. coli bacteria express an outer membrane transporter FepA.32,33 The crystal structure of FepA shows a remarkable structural similarity to FpvA, the pyoverdine outer membrane transporter of Pseudomonas aeruginosa.34,35 In each structure there are two domains: a 22stranded C-terminal β barrel and a cork domain formed by approximately 150 N-terminal residues.36 The beta strands are connected by solvent-accessible extracellular loops that extend

ARTICLE

30
40 Å above the outer membrane to facilitate the initial binding of ferric
siderophores from the environment.36 For P. aeruginosa it has also been shown that pyoverdine without Fe3+ (apo-pyoverdine) is not only secreted but also held at the bacterial surface by the extracellular loops of FpvA.37 In an earlier study using ATR-IR spectroscopy it was shown that pyoverdine held by P. aeruginosa cells forms covalent bonds with surface metal ions primarily through the catecholate group of the molecule.38 Formation of these strong interactions and initiation of the adhesion process could have significant implications for biofilm formation. It is therefore likely that FepA also holds apoenterobactin at the E. coli cell surface and would therefore be available to coordinate surface metal ions. There have been few infrared spectroscopic studies of E. coli interactions with metal oxide films. However, Nadtochenko et al.39 studied the adsorption of E. coli to TiO2 using IR spectroscopy in the context of TiO2 photocatalysis and bacterial disinfection. In the Nadtochenko et al. study there was no evidence for adsorption of enterobactin to the TiO2, most likely due to utilization of an iron-replete bacterial growth medium. Such an in vitro growth environment would likely result in inhibition of enterobactin biosynthesis through Fur-mediated repression. In related research, infrared spectroscopic measurements of hydroxyaromatic compounds such as catechol have been shown to interact strongly with TiO2, forming inner-sphere complexes.40
42 Recently, the interaction of catechol with chromium(III) oxide (Cr2O3), iron(III) oxide (Fe2O3), and TiO2 was evaluated as a function of pH and ionic strength using ATRIR spectroscopy.43 This IR spectroscopic study also indicated that catechol binds predominately as an inner-sphere complex on the metal oxides. Related IR spectroscopic studies38,44,45 of the adhesion to metal oxides of P. aeruginosa have shown that pyoverdine is involved in the initial stages of bacterial biofilm formation. These studies are also important in understanding the role of siderophores in bacterial iron sequestration via iron oxide mineral dissolution.46
50 There have been other recent IR spectroscopic studies of bacterial adhesion to iron oxides which have shown formation of covalent bonds involving phosphoryl moieties in bacterial surface polymers.51,52 The bacteria in these recent studies were grown in iron-replete media, which repress siderophore biosynthesis, and therefore, siderophore adsorption to metal oxides would not have been expected. The present study extends our previous work on the role of pyoverdine in adhesion of P. aeruginosa bacteria to metal (oxide) surfaces. Here, using ATR-IR spectroscopy we examine the binding to metal oxides of enterobactin, both from solution and bound to E. coli bacteria. The metal oxides examined, titanium dioxide (TiO2), boehmite (γ-AlOOH), and chromium(III) oxide
hydroxide (CrOOH), are present at metal surfaces relevant to medical and industrial environments. To investigate the adsorption mode of enterobactin on a TiO2 surface, catecholatetype and salicylate-type surface complex simulation models were constructed and IR frequencies calculated using density functional theory (DFT). These investigations describe in detail the adsorptive coordination of enterobactin to surface metal ions and its occurrence during E. coli cell adhesion to wet metal oxide surfaces.

’ MATERIALS AND METHODS Materials. Enterobactin was isolated from an E. coli Δfur strain,53 and the enterobactin was purified via adsorption and gel-permeation 10588

dx.doi.org/10.1021/la202352j |Langmuir 2011, 27, 10587–10596

Langmuir chromatography, as previously described.54 Matrix-assisted laser desorption/ionization-time-of-flight (MALDI-TOF) mass spectrometry was used to confirm the identity of enterobactin. Briefly, a small aliquot (0.5 μL) of the sample was mixed with R-cyano-4-hydroxycinnamic acid (0.5 μL of 10 mg mL
1 in 0.1% trifluoroacetic acid, 60% acetonitrile), spotted onto a MALDI target, and allowed to dry. Mass data were collected in the positive-ion mode, with an acceleration voltage of 20 kV on a Finnigan Lasermat 2000 (Thermo Finnigan, San Jose, CA). The E. coli Δfur strain had been constructed by introducing a Tn10::kan insertion within the fur gene to disrupt gene expression.53 The Δfur strain was kindly provided by Dr. D. Touati, University of Paris, France. HCl (BDH, AR) and NaOH (Merck, AR) was used as received. All water for solutions and cell washing was deionized (Millipore, Milli-Q). Solution pH measurements were made with a Mettler-Toledo MP220 meter to a precision of (0.1. Preparation of Biomass. E. coli strains used in this study were cultivated at 37 °C in Vogel Bonner minimal medium.55 Cells were harvested in early stationary growth phase (16 h) by centrifugation at 10 000 g for 10 min, washed three times, and resuspended in water. Suspensions used in flow cell experiments had viable cell concentrations of ∼108 cfu mL
1, as determined by plating onto Luria
Bertani agar.56 Preparation of Metal Oxide Sols. Titanium dioxide (TiO2), boehmite (γ-AlOOH), and chromium(III) oxide
hydroxide (CrOOH) sols were used to model the surfaces of titanium, aluminum, and stainless steel. Titanium(IV) oxide sol (0.01 mol L
1) was prepared by hydrolysis of TiCl4 (Riedel de Haen) in water as previously reported.40 These films contained amorphous TiO2 with an isoelectric point pH ≈ 5.57 The TiO2 particle size is 6 and are less significant for pH < 6, where there are indirect influences of surface charge on the adsorption of neutral species. The evidence shown in Figure 4 that enterobactin adsorbs to TiO2 at pH > 6 confirms a coordinative adsorption rather than electrostatic adsorptive interactions. Figure 4 shows the IR spectra of enterobactin adsorbed to TiO2 over the pH range from 3.2 to 12.0. The adsorbed enterobactin spectrum at pH 3.2 shows prominent bands between 1800 and 1200 cm
1 with peaks at 1740, 1615, 1591, 1551, 1464, 1447, 1402, 1255, and 1226 cm
1. Bands at these wavenumbers for spectra collected between pH 3.2 and 7.2 revealed little 10592

dx.doi.org/10.1021/la202352j |Langmuir 2011, 27, 10587–10596

Langmuir

Figure 4. IR spectra showing the pH dependence of enterobactin adsorption to TiO2 from a 1  10
4 mol L
1 enterobactin aqueous solution, recorded in sequence from low pH to high pH.

change in shape, which suggests the adsorption mode of enterobactin is relatively stable in this pH range. Significant band changes within the spectrum across the pH range may have indicated a varied degree of participation of the three catecholate ligands in surface binding. The lack of spectrum change with pH could therefore suggest that all of the catechol units are bound throughout this pH range. The enterobactin peak absorbances reach their maxima about pH 7
8, close to the reported pKa values of enterobactin, which is also consistent with the behavior of catecholate ligand adsorption on TiO2.70 At pH > 8 the adsorbed enterobactin spectrum shows a number of peak relative intensity changes. These include a relative decrease in the 1615 cm
1 amide I absorption, a relative increase in the peaks at ∼1590 and 1460 cm
1, and a shift of the 1255 cm
1 band to 1263 cm
1. Interestingly, although the 1615 cm
1 band has the most dramatic change in intensity the 1551 cm
1 band (Amide II) has a relatively slower decrease at pH > 7, which may indicate a conformation change of the amide bonds as pH increases. From pH 7 to 10, the intensity of the trilactone band at 1740 cm
1 has a slight decrease, suggesting that the trilactone ring is relatively stable. Above pH 10 there is a significant absorbance decrease in the band at 1740 cm
1 arising from the CdO stretching mode of the ester groups, and this is also accompanied by an increase in the 1401 cm
1 absorption. While this absorbance decrease may in part be due to conformational change it may also be partly attributable to some hydrolysis of adsorbed enterobactin under alkaline conditions. However, in comparison with the spectrum of enterobactin in solution (Figure 2, spectrum b) it appears that enterobactin adsorbed to TiO2 retains its structural integrity over a longer period under the same pH conditions and is more resistant to hydrolysis. The increased resistance to hydrolysis may be due to the different conformation that enterobactin has to adopt upon adsorption to TiO2 from solution. A recent conformation study of free enterobactin indicates that the conformation preference of enterobactin is sensitive to the donor/acceptor strength of the solvent or even cations in solution.31 Therefore, the pH dependence of enterobactin in the higher pH range may be related to hydrolysis and a conformation change of the adsorbed species. Two different adsorption modes may assist in explaining this potential for a conformational change. In the lower pH range, the enterobactin molecule approaches the TiO2 surface either in an axial, propeller-like conformation or in an equatorial conformation.

ARTICLE

Figure 5. Adsorption isotherm of enterobactin adsorbed on TiO2 from 1  10
6 to 1  10
4 mol L
1 enterobactin in 0.01 mol L
1 KCl solutions at pH = 3. Data points collected at 1742, 1553, 1448, 1395, 1265, and 1227 cm
1 with isotherm shown as a Langmuir fitted curve.

For the axial conformation enterobactin molecules may have two of the catechol moieties chelating to separate Ti sites with a free catechol moiety group being mobile. For the equatorial conformation, three of the catechol moieties are chelated on three separate Ti sites. For the axial adsorption mode with pH increase, the linkage to the uncoordinated catechol group(s) may undergo hydrolysis with the ligand(s) being released from the surface. It is also likely that the trilactone ring would be attacked by alkaline solution, which would lead to a decrease of the 1740 cm
1 band absorbance. For the equatorial conformation, the trilactone part may not be so readily hydrolyzed. Changes to the trilactone ring conformation and the amide bonds of the adsorbed enterobactin may still occur, accompanied by observed spectral changes in the higher pH range. Adsorption Isotherm for Enterobactin Adsorption to TiO2. To further investigate the nature of enterobactin binding on TiO2, the binding affinity of this ligand was evaluated by measuring an adsorption isotherm from the concentration dependence of the IR spectra of the adsorbed species. The measured adsorption isotherm data over the 1  10
6
1  10
4 mol L
1 concentration range at pH 3 is shown in Figure 5. The adsorption isotherm data is expected to be similar across the pH range from 3 to 9, where a single adsorbed species is observed for catechol on TiO2.42,71 The simple Langmuir model was used to analyze the adsorption behavior and obtain an adsorption equilibrium constant. While adsorption of many ligands to TiO2 surfaces is not expected to conform closely to the Langmuir isotherm model, such data has been widely reported to provide comparative data. From the nonlinear least-squares fitting of the absorbance versus concentration data a Langmuir adsorption constant of 3  105 L mol
1 was determined. The Langmuir adsorption constant for catechol on the same substrate was also determined as 1.2  104 L mol
1 (data not shown). This value is also close to that of 8.2  103 mol L
1 for catechol adsorption onto nanocrystalline TiO2 (∼80% anatase, ∼20% rutile) obtained by Rodriguez and others.71 The comparison of these adsorption constants is consistent with enterobactin, with three catecholate ligands, being more strongly bound to TiO2 than catechol alone. Comparison of Langmuir adsorption constants for other adsorbates on the same TiO2 substrate such as 3  103 L mol
1 for lysine which 10593

dx.doi.org/10.1021/la202352j |Langmuir 2011, 27, 10587–10596

Langmuir

Figure 6. ATR-IR spectra of enterobactin adsorbed to (a) TiO2, (b) AlOOH, and (c) CrOOH particle films from 1  10
4 mol L
1 aqueous solution at pH 6 after 60 min solution flow.

adsorbs relatively weakly through outer-sphere interactions and 1.4  106 L mol
1 for the photosensitizer ruthenium(II) bipyridyl dicarboxylic acid which adsorbs strongly via two carboxylic acid groups can also be made.72 IR Spectra of Enterobactin Adsorbed to TiO2, AlOOH, and CrOOH. Enterobactin has been previously shown to have a high affinity for Al3+ and Cr3+ in aqueous solution, and it is therefore likely this hexadentate ligand can adsorb to metal oxides other than TiO2. Figure 6 shows the IR spectra of enterobactin adsorbed at pH 6 to particle films of TiO2, CrOOH, and AlOOH, where the spectrum from TiO2 is the same as spectrum c in Figure 2. The spectra show several prominent bands in the range between 1800 and 1200 cm
1. While there are recognizable similarities in the spectral features from the different substrates, there are also notable differences between the three spectra. Many of these differences may arise from the intrinsic differences in substrate geometry which impose conformational constraints on enterobactin as it seeks to adsorb to the maximum number of sites in order to minimize its energy. All spectra show a band at about 1740 cm
1 which indicates the presence of ester groups from the triserine lactone ring. Comparing spectra from the three substrates, the ester bands are somewhat different in absorbance relative to the other major peaks in the enterobactin spectrum and have slightly different peak wavenumbers. These spectral differences may be due to the surface structures of the chemically distinct substrates inducing altered adsorbate conformations. It is also possible that the differences are partly due to adsorption of some hydrolyzed enterobactin or to the chemically different substrates having a variable influence on the tendency of adsorbed enterobactin to hydrolyze. Spectrum b for enterobactin on AlOOH shows a broadened ester peak, suggesting a composite band arising from more than one adsorbate conformation. The three component peaks in the band spreading from about 1650 to 1500 cm
1 show minor relative intensity differences and wavenumber changes. In the 1480
1440 and 1290
1200 cm
1 regions there are characteristic and prominent bifurcated peaks arising from the binding of the catecholate ligands to the surface metal ions which confirm the chemical basis of the enterobactin adsorption. The twin peaks in the 1480
1440 cm
1 region are well resolved for spectra from AlOOH and CrOOH in comparison with the spectrum from TiO2. In the 1290
1200 cm
1

ARTICLE

Figure 7. ATR-IR spectra of E. coli cells attached to (a) ZnSe prism and (b) ZnSe prism coated with TiO2. Spectra were collected following 120 min of bacterial flow across the surface. The bacteria were suspended in H2O. Approximate absorption peak IR wavenumber is indicated.

region there is more divergent behavior in relative band intensity and peak wavenumber across the three substrates with that from CrOOH being most different. The divergent behavior is mainly in the higher wavenumber part of this range, while the lower wavenumber peaks are observed within a narrow range from 1226 to 1221 cm
1. This could be due to geometry hindrance limiting the full coordination of the hexadentate ligand to a surface Cr(III) site so that only some of the catecholate groups are involved in chelation with surface Cr3+ ions. The mode of enterobactin adsorption onto CrOOH may be similar to that of enterobactin adsorbed to TiO2 The variable presence of minor absorptions ∼1400 cm
1 are attributed to the CH and CH2 bending modes of the trilactone group and may have arisen from carboxylate groups of some partially hydrolyzed enterobactin present in the adsorbate solution. The results described above confirm that adsorption of enterobactin to a variety of metal oxides occurs via formation of coordinate covalent bonds via the catecholate ligands. Siderophore
Metal Bond Formation during Attachment of E. coli to TiO2. Enterobactin bound at the cell surface by the FepA protein may potentially adsorb to TiO2, thereby initiating bacterial biofilm formation. Therefore, to investigate if enterobactin at the cell surface forms covalent bonds with metal oxides we examined the attachment of E. coli cells to TiO2 particle films using in situ ATR-IR spectroscopy. Representative spectra after 120 min of bacterial flow over ZnSe and TiO2/ZnSe are shown in Figure 7. For each substrate the spectral peak intensities increased throughout the 120 min period but the relative peak intensities remained constant. The time dependence of the spectral absorbances corresponded to an increasing number of cells attaching to the surface with increased time but with a decreasing attachment rate as surface coverage grew.73 The absorption peaks in the spectrum recorded using the bare ZnSe crystal (Figure 7a) are typical of bacteria attached to ATRIR crystals and have been commonly assigned to the vibrational modes of various bacterial components.74 The most prominent features in this spectrum are peaks at 1632/1627 and 1549 cm
1 arising from amide vibrations and the composite band at ∼1080 cm
1 which contains contributions from carbohydrate polymers as well as phosphate groups.75 For E. coli on the TiO2 film (Figure 7b), additional peaks are evident at 1743, 1591, 1464, 1448, 1254, and 1227 cm
1 and are readily recognized as arising 10594

dx.doi.org/10.1021/la202352j |Langmuir 2011, 27, 10587–10596

Langmuir from adsorbed enterobactin. The prominent absorption peak at 1743 cm
1 is consistent with the presence of ester groups from the triserine lactone ring of the intact enterobactin molecule at the cell surface. The 1700
1500 cm
1 region of the spectrum is noticeably altered from that of enterobactin adsorbed on TiO2 in that the amide absorptions at 1627 and 1549 cm
1 from the bacterial protein are dominant over the siderophore contributions. The strong band at 1448 cm
1 with a shoulder at 1464 cm
1 and the closely adjacent 1254 and 1227 cm
1 bands correspond closely to those observed in the IR spectrum of enterobactin adsorbed on TiO2 (Figure 2c). These absorptions, from the CdC ring vibrations and phenolic C
O stretch modes of the 2,3-dihydroxybenzene part of enterobactin signal formation of coordinate covalent bonds with the titanium(IV) ions in the TiO2 surface. The strong binding of apo-enterobactin to metal oxide surfaces may compete with the binding of enterobactin to the cell membrane. The absorbance of the adsorbed enterobactin IR peaks suggests that there is more adsorbed enterobactin present than might be expected from the adhesion of the E. coli with enterobactin at the same time bound to the cell membrane via a receptor. Thus, it is possible that with cell adhesion some of the cell wall-bound enterobactin may be detached from the cell membrane due to the strong interaction with the metal oxide surface. However, such a process will facilitate adhesion to some extent and could be involved in the sequestration of metal ions from the surfaces of minerals such as iron oxides.

’ SUMMARY AND CONCLUSIONS This study provides direct spectroscopic evidence that enterobactin binds covalently through the 2,3-dihydroxy-N-benzoylserine groups of the molecule to particle films of TiO2. Analyses of the spectra for enterobactin in neutral solution conditions when compared with that adsorbed on TiO2 revealed a significant conformation change of the amide bonds upon adsorption. This change in conformation occurs by an interconversion where the carbonyl of an amide group reorientates during coordination to surface Ti(IV) ions. For this to occur, the hydrogen bonds between the amide carbonyl and the o-hydroxyl of the catechol unit are cleaved with rotation of the catechol unit allowing it to form coordinative bonds with surface Ti sites. Raymond et al.27 previously reported such an interconversion occurs in solution and can be triggered by enterobactin deprotonation/metal complexation. Our experiments have also confirmed that the structurally dominant trilactone ring forming the central scaffold part of the enterobactin molecule had little if any involvement in the adsorption process. The calculated IR spectra of catecholate and salicylate-type model surface complexes were compared with the observed IR absorption bands of the adsorbed enterobactin. From this simulation data we established that catecholate
surface-type complexes are more favored. The Langmuir adsorption constant for enterobactin was determined from IR adsorption isotherm data and compared to that of catechol on the same substrate. The strength of the binding confirmed a bidendate binding mode and revealed that more than one of the 2,3-dihydroxy-N-benzoylserine ligands are involved in the coordination to metal ions at the TiO2 surface. The pH dependence of IR spectra of adsorbed enterobactin would indicate that adsorption of enterobactin is most favored near neutral pH (7
8), well within the physiological pH range of E. coli bacteria investigated here. The lack of significant band

ARTICLE

changes observed within the pH range of 3.2
12.0 supports the conclusion that all 2,3-dihydroxy-N-benzoylserine ligands are involved in enterobactin’s adsorption to the Ti surface sites. Therefore, while enterobactin can adopt either an axial or an equatorial adsorption mode, in which all three catecholate groups are free to coordinate metal ions at the TiO2 surface, the latter mode appears more favored. In addition, enterobactin adsorbed to the TiO2 surface was observed to be very stable and resistant to hydrolysis in alkaline conditions as compared to enterobactin in solution. Adsorption studies of enterobactin to other metal oxides such as AlOOH and CrOOH particle films by in situ ATR-IR spectroscopy suggest that enterobactin also forms catecholate-type surface complexes on these substrates. However, intrinsic differences in spectra were also observed, which demonstrates different conformations of adsorbed enterobactin can occur due to the substrates’ chemical nature and geometry. These results also suggest that enterobactin is likely to adsorb to metal oxide surfaces in general. apo-Enterobactin secreted from E. coli cells and bound to the outer membrane via its association with the outer loops of the FepA receptor provides an opportunity for the siderophore to assist in the initiation of cell attachment to surfaces. The IR spectrum of enterobactin adsorbed to TiO2 accompanying adhesion of washed E. coli cells to TiO2 strongly suggests that enterobactin is involved in the cell adhesion process. Collectively, the present results suggest a general mechanism by which siderophores associated with the cell surface form covalent bonds in coordinative adsorption to metal (oxide) surfaces during initiation of biofilms on these surfaces.

’ AUTHOR INFORMATION Corresponding Author

*Phone: 64-3-479-7928. Fax: 64-3-479-7906. E-mail: jmcquillan@ chemistry.otago.ac.nz.

’ ACKNOWLEDGMENT This research was supported by grant UOO203 from the New Zealand Marsden Fund. ’ REFERENCES (1) Costerton, J. W.; Stewart, P. S.; Greenberg, E. P. Science 1999, 284, 1318–1322. (2) Darouiche, R. O. Clin. Infect. Dis. 2001, 33, 1567–72. (3) Costerton, J. W.; Cheng, K. J.; Geesey, G. G.; Ladd, T. I.; Nickel, J. C.; Dasgupta, M.; Marrie, T. J. Annu. Rev. Microbiol. 1987, 41, 435–64. (4) Hamilton, W. A. Biofouling 2003, 19, 65–76. (5) Stoodley, P.; Sauer, K.; Davies, D. G.; Costerton, J. W. Annu. Rev. Microbiol. 2002, 56, 187–209. (6) O’Toole, G. A. J. Bacteriol. 2003, 185, 2687–2689. (7) Hermansson, M. Colloids Surf., B 1999, 14, 105–119. (8) Poortinga, A. T.; Bos, R.; Norde, W.; Busscher, H. J. Surf. Sci. Rep. 2002, 47, 1–32. (9) Van Oss, C. J. Cell Biophys. 1989, 14, 1–16. (10) Harrick, N. J. In Internal Reflection Spectroscopy, 3rd ed.; Harrick Scientific Products, Inc.: New York, 1987. (11) Mirabella, F. M., Jr. Appl. Spectrosc. Rev. 1985, 21, 45–178. (12) McQuillan, A. J. Adv. Mater. 2001, 13, 1034–1038. (13) Levine, M. M. Escherichia coli infections. In Bacterial Vaccines; Germanier, R., Ed.; Academic Press: Orlando, 1984; pp 187
235. (14) Kaper, J. B.; Nataro, J. P.; Mobley, H. L. T. Nat. Rev. Microbiol. 2004, 2, 123–140. 10595

dx.doi.org/10.1021/la202352j |Langmuir 2011, 27, 10587–10596

Langmuir (15) Wilks, S. A.; Michels, H; CW., K. Int. J. Food. Microbiol. 2005, 105, 445–54. (16) Silagyi, K.; Kim, S. H.; Lo, Y. M.; Wei, C. I. Food Microbiol. 2009, 26, 514–9. (17) Boukhalfa, H.; Crumbliss, A. L. BioMetals 2002, 15, 325–339. (18) Escolar, L.; Perez-Martin, J.; de Lorenzo, V. J. Bacteriol. 1999, 181, 6223–9. (19) Hantke, K. Curr. Opin. Microbiol. 2001, 4, 172–7. (20) Stintzi, A.; Raymond, K. Siderophore Chemistry. In Molecular and Cellular Iron Transport; Templeton, D. M., Ed.; Marcel Dekker, Inc.: New York, 2002; pp 273
319. (21) O’Brien, I. G.; Gibson, F. Biochim. Biophys. Acta 1970, 215, 393–402. (22) Pollack, J. R.; Neilands, J. B. Biochem. Biophys. Res. Commun. 1970, 38, 989–92. (23) Abergel, R. J.; Warner, J. A.; Shuh, D. K.; Raymond, K. N. J. Am. Chem. Soc. 2006, 128, 8920–8931. (24) Thulasiraman, P.; Newton, S. M. C.; Xu, J.; Raymond, K. N.; Mai, C.; Hall, A.; Montague, M. A.; Klebba, P. E. J. Bacteriol. 1998, 180, 6689–6696. (25) Scarrow, R. C.; Ecker, D. J.; Ng, C.; Liu, S.; Raymond, K. N. Inorg. Chem. 1991, 30, 900–6. (26) Stack, T. D. P.; Karpishin, T. B.; Raymond, K. N. J. Am. Chem. Soc. 1992, 114, 1512–14. (27) Raymond, K. N.; Dertz, E. A.; Kim, S. S. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 3584–3588. (28) Pecoraro, V. L.; Harris, W. R.; Wong, G. B.; Carrano, C. J.; Raymond, K. N. J. Am. Chem. Soc. 1983, 105, 4623–33. (29) Gutierrez, C. G.; Burton, J. L.; Ramirez, R. J. A.; Amador, R. B. Book of Abstracts, 218th ACS National Meeting, New Orleans, LA, Aug 22
26, 1999; American Chemical Society: Washington, DC, 1999; CHED-152. (30) Garcia, A., Jr.; Ramirez, R. J. A.; Gutierrez, C. Abstracts of Papers, 226th ACS National Meeting, New York, NY, Sept 7
11, 2003; American Chemical Society: Washington, DC, 2003; INOR-473. (31) Godinez, C. E.; Oganesyan, A.; Amador, R.; Garcia, A.; Gutierrez, C. G. Abstracts of Papers, 233rd ACS National Meeting, Chicago, IL, Mar 25
29, 2007; American Chemical Society: Washington, DC, 2007; INOR-798. (32) Ecker, D. J.; Matzanke, B. F.; Raymond, K. N. J. Bacteriol. 1986, 167, 666–73. (33) Hoegy, F.; Celia, H.; Mislin, G. L.; Vincent, M.; Gallay, J.; Schalk, I. J. J. Biol. Chem. 2005, 280, 20222–30. (34) Buchanan, S. K.; Smith, B. S.; Venkatramani, L.; Xia, D.; Esser, L.; Palnitkar, M.; Chakraborty, R.; Van Der Helm, D.; Deisenhofer, J. Nat. Struct. Biol. 1999, 6, 56–63. (35) Cobessi, D.; Celia, H.; Folschweiller, N.; Schalk, I. J.; Abdallah, M. A.; Pattus, F. J. Mol. Biol. 2005, 347, 121–34. (36) Chakraborty, R.; Lemke, E. A.; Cao, Z.; Klebba, P. E.; van der Helm, D. Biometals 2003, 16, 507–18. (37) Schalk, I. J.; Hennard, C.; Dugave, C.; Poole, K.; Abdallah, M. A.; Pattus, F. Mol. Microbiol. 2001, 39, 351–360. (38) Upritchard, H. G.; Yang, J.; Bremer, P. J.; Lamont, I. L.; McQuillan, A. J. Langmuir 2007, 23, 7189–7195. (39) Nadtochenko, V. A.; Rincon, A. G.; Stanca, S. E.; Kiwi, J. J. Photochem. Photobiol., A 2005, 169, 131–137. (40) Connor, P. A.; Dobson, K. D.; McQuillan, A. J. Langmuir 1995, 11, 4193–5. (41) Araujo, P. Z.; Mendive, C. B.; Rodenas, L. A. G.; Morando, P. J.; Regazzoni, A. E.; Blesa, M. A.; Bahnemann, D. Colloids Surf., A. 2005, 265, 73–80. (42) Lana-Villarreal, T.; Rodes, A.; Perez Juan, M.; Gomez, R. J. Am. Chem. Soc. 2005, 127, 12601–11. (43) Gulley-Stahl, H.; Hogan, P. A.; Wall, S. J.; Buhrlage, A.; Bullen, H. A. Environ. Sci. Technol. 2010, 44, 4116–4121. (44) McWhirter, M. J.; Bremer, P. J.; Lamont, I. L.; McQuillan, A. J. Langmuir 2003, 19, 3575–3577. (45) Comeau, J. W.; Pink, J.; Bezanson, E.; Douglas, C. D.; Pink, D.; Smith-Palmer, T. Appl. Spectrosc. 2009, 63, 1000–7.

ARTICLE

(46) Kraemer, S. M.; Cheah, S.-F.; Zapf, R.; Xu, J.; Raymond, K. N.; Sposito, G. Geochim. Cosmochim. Acta 1999, 63, 3003–3008. (47) Cocozza, C.; Tsao, C. C. G.; Cheah, S.-F.; Kraemer, S. M.; Raymond, K. N.; Miano, T. M.; Sposito, G. Geochim. Cosmochim. Acta 2002, 66, 431–438. (48) Cervini-Silva, J.; Sposito, G. Environ. Sci. Technol. 2002, 36, 337–342. (49) Cheah, S.-F.; Kraemer, S. M.; Cervini-Silva, J.; Sposito, G. Chem. Geol. 2003, 198, 63–75. (50) Kraemer, S. M. Aquat. Sci. 2004, 66, 3–18. (51) Parikh, S. J.; Chorover, J. Langmuir 2006, 22, 8492–8500. (52) Omoike, A.; Chorover, J. Geochim. Cosmochim. Acta 2006, 70, 827–838. (53) Compan, I.; Touati, D. J. Bacteriol. 1993, 175, 1687–96. (54) Berner, I.; Greiner, M.; Metzger, J.; Jung, G.; Winkelmann, G. Biol. Met. 1991, 4, 113–8. (55) Vogel, H. J.; Bonner, D. M. J. Biol. Chem. 1956, 218, 97–106. (56) Sambrook, J.; Fritsch, E. F.; Maniatis, T. In Molecular Cloning: A Laboratory Manual; Nolan, C., Ferguson, M., Eds.; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY, 1989; Vol. 3, p A.1. (57) Dobson, K. D.; Connor, P. A.; McQuillan, A. J. Langmuir 1997, 13, 2614–2616. (58) Degenhardt, J.; McQuillan, A. J. Chem. Phys. Lett. 1999, 311, 179–184. (59) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502. (60) Loomis, L. D.; Raymond, K. N. Inorg. Chem. 1991, 30, 906–11. (61) Raymond, K. N.; Dertz, E. A. Biochemical and Physical Properties of Siderophores. In Iron Transport In Bacteria; Crosa, J. H., Mey, A. R., Payne, S. M., Eds.; ASM Press: Washington, DC, 2004; pp 3
17. (62) Leslie, A. D.; Daneshfar, R.; Volmer, D. A. J. Am. Soc. Mass Spectrom. 2007, 18, 632–641. (63) Meyer, M.; Telford, J. R.; Cohen, S. M.; White, D. J.; Xu, J.; Raymond, K. N. J. Am. Chem. Soc. 1997, 119, 10093–10103. (64) Smith, B. C. Infrared Spectral Interpretation: A Systematic Approach; CRC Press: Boca Raton, 1998. (65) Yang, J. The roles of siderophores in bacterial adhesion to metals and iron transport. Ph.D. Thesis, University of Otago, 2009. (66) Llinas, M.; Wilson, D. M.; Neilands, J. B. Biochemistry 1973, 12, 3836–43. (67) Steiner, G.; Tunc, S.; Maitz, M.; Salzer, R. Anal. Chem. 2007, 79, 1311–6. (68) Greaves, S. J.; Griffith, W. P. Spectrochim. Acta, Part A 1991, 47A, 133–40. (69) Shanzer, A.; Libman, J. Biomimetic Siderophores. In CRC Handbook of Microbial Iron Chelates; Winkelmann, G., Ed.; CRC Press, Inc.: Boca Raton, FL, 1991; pp 309
338. (70) Martin, S. T.; Kesselman, J. M.; Park, D. S.; Lewis, N. S.; Hoffmann, M. R. Environ. Sci. Technol. 1996, 30, 2535–2542. (71) Rodriguez, R.; Blesa, M. A.; Regazzoni, A. E. J. Colloid Interface Sci. 1996, 177, 122–131. (72) Duffy, N. W.; Dobson, K. D.; Gordon, K. C.; Robinson, B. H.; McQuillan, A. J. Chem. Phys. Lett. 1997, 266, 451–455. (73) Kang, S. Y.; Bremer, P. J.; Kim, K. W.; McQuillan, A. J. Langmuir 2006, 22, 286–91. (74) Jiang, W.; Saxena, A.; Song, B.; Ward, B. B.; Beveridge, T. J.; Myneni, S. C. B. Langmuir 2004, 20, 11433–11442. (75) Le Gal, J. M.; Manfait, M.; Theophanides, T. J. Mol. Struct. 1991, 242, 397–407.

10596

dx.doi.org/10.1021/la202352j |Langmuir 2011, 27, 10587–10596