Anal. Chem. 2005, 77, 7335-7341
Thorium Complexation by Hydroxamate Siderophores in Perturbed Multicomponent Systems Using Flow Injection Electrospray Ionization Mass Spectrometry Miranda J. Keith-Roach,*,† Marta Vetri Buratti,†,‡ and Paul J. Worsfold†
School of Earth, Ocean and Environmental Sciences, University of Plymouth, Drake Circus, Plymouth PL4 8AA, U.K., and Faculty of Pharmacy, University of Bologna, Bologna, Italy
Flow injection electrospray ionization mass spectrometry has been shown to produce simple, characteristic m/z signals for Th-hydroxamate siderophore (desferrioxamine and ferrichrome) complexes, with Th complexed as a simple 4+ ion in the environmentally relevant pH range investigated (pH 5-9). All species of interest for this study were identified optimally in the positive mode; thus, multiple species were analyzed concurrently in a single spectrum. Complexation of Th by the two siderophores was rapid in 1:1 molar aqueous solution, reaching equilibrium before the first measurement was possible at 2 min. However, a significant proportion of the equimolar siderophore remained uncomplexed. Both siderophores rapidly exchanged Th for Fe when equimolar Fe(III) was added to the Th complexes, and only a small proportion of each siderophore remained complexed with Th at equilibrium (7-30 min). The results show a difference in the affinities of the two siderophores for the metals; ferrichrome has a 5-fold higher affinity than desferrioxamine for Th and a 5-fold lower affinity than desferrioxamine for Fe. Also, siderophore-complexed Th interacted strongly with a cation-exchange resin suggesting that, even when complexed by trianionic siderophores, Th mobility will be impeded by interactions with negatively charged binding sites in subsurface environmental matrixes. These results have important implications regarding siderophore-enhanced actinide(IV) mobility in the terrestrial environment. Siderophores are important microbial iron(III)-binding ligands, thought to be produced under the stress of low iron availability.1 Iron-siderophore complexes have extremely high stability constants; for example, iron(III) has a stability constant of 1030 with desferrioxamine B,2 orders of magnitude higher than those for iron with other environmentally important complexes, such as * To whom correspondence should be addressed. E-mail: mkeith-roach@ plymouth.ac.uk. Fax: +44 1752 233035. † University of Plymouth. ‡ University of Bologna. (1) Neilands, J. B. J. Biol. Chem. 1995, 270, 26723-26726. (2) Brainard, J. R.; Strietelmeier, B. A.; Smith, P. H.; Langston-Unkefer, P. J.; Barr, M. E.; Ryan, R. R. Radiochim. Acta 1992, 58-9, 357-363. 10.1021/ac051069y CCC: $30.25 Published on Web 10/13/2005
© 2005 American Chemical Society
fulvic or humic acids.1 Siderophores have high affinities for tetravalent actinides, for example, enterobactin for plutonium(IV)3 and desferrioxamine B for thorium(IV),2 that are almost as high as the affinities of the siderophores for iron(III). Tetravalent actinides display very low environmental solubilities, with high distribution coefficients between the solid and solution phases. Therefore, complexing agents that aid solubilization of these hazardous elements may be important even if the complexing agent is present at a low environmental concentration; i.e., they may significantly increase the total concentration of mobile actinide present.2-4 Understanding the stability and mobility of these complexes is therefore important both environmentally, for example, predicting migration from waste repositories or contaminated land, and industrially, where analogues of siderophores may be of use. In natural soils, there is an abundance of negatively charged binding sites, and to our knowledge, association of positively charged siderophore complexes with the solid phase has not been investigated. These issues highlight the complexity of siderophore-actinide interactions and the need to understand them in more detail and under real environmental conditions. Electrospray ionization mass spectrometry (ESI-MS) is used increasingly in the analysis of equilibrated metal complexes because it generates simple mass spectra as a result of the gentle ionization method.5-8 Compounds or complexes tend not to fragment during ionization and typically produce a single diagnostic peak. Other advantages include the relatively high sensitivity of the technique and the structural information given by both the m/z signal of a complex ion and investigative MSn fragmentation. Applications of ESI-MS to date include the determination of metal EDTA complexes,5,6,9 uranyl citrates,8 thorium hydrolysis species,10 and hydroxamate siderophore complexes of iron 7 and (3) Harris, W. R.; Carrano, C. J.; Raymond, K. N. J. Am. Chem. Soc. 1979, 101, 2722-2727. (4) Banaszak, J. E.; Rittmann, B. E.; Reed, D. T. J. Radioanal. Nucl. Chem. 1999, 241, 385-435. (5) Baron, D.; Hering, J. G. J. Environ. Quality 1998, 27, 844-850. (6) Collins, R. N.; Onisko, B. C.; McLaughlin, M. J.; Merrington, G. Environ. Sci. Technol. 2001, 35, 2589-2593. (7) Gledhill, M. Analyst 2001, 126, 1359-1362. (8) Pasilis, S. P.; Pemberton, J. E. Inorg. Chem. 2003, 42, 6793-6800. (9) Mollah, S.; Pris, A. D.; Johnson, S. K.; Gwizdala, A. B.; Houk, R. S. Anal. Chem. 2000, 72, 985-991. (10) Moulin, C.; Amekraz, B.; Hubert, S.; Moulin, V. Anal. Chim. Acta 2001, 441, 269-279.
Analytical Chemistry, Vol. 77, No. 22, November 15, 2005 7335
Figure 1. Molecular structures of ferrichrome (A) and desferrioxamine (B).
uranium.11 The technique therefore has diverse applications and has the potential to aid a better understanding of siderophoreenhanced actinide mobility. The ability to measure uranyl desferrioxamine complexes has been demonstrated by Groenewold et al.,11 and uranyl (UO22+) is, as expected, bound as a dioxo cation as determined by the m/z of the complex. Since siderophores have a higher affinity for tetravalent actinides, thorium is a good choice for experiments with its single oxidation state and relatively low radiological and chemical hazards. This study aimed to accomplish the following: (1) establish fundamental data regarding the analysis of Th-hydroxamate siderophore complexes by flow injection ESI-MS, including the diagnostic signals, i.e., the form in which Th is bound, the potential for quantification, and signal suppression; (2) observe Th complexation by siderophores and subsequent metal exchange on addition of Fe over relatively short (minute) time scales; (3) assess the relative affinities of the siderophores for Th and Fe; and (4) examine interactions of these Th-siderophore complexes with a cation-exchange resin as a model of soil binding sites. Two hexadentate hydroxamate siderophores, desferrioxamine and ferrichrome (Figure 1), were selected for this study on the basis that they could be easily purchased and because they are common in terrestrial environments.12 These siderophores contain three ionizable hydroxamate groups, thus having a maximum charge of 3-. Overall, this study aimed to test the hypothesis that ESIMS is a useful tool in determining complexation processes and competitive interactions in perturbed multicomponent systems moving toward equilibrium. EXPERIMENTAL SECTION Instrumentation. A Finnigan Mat LCQ ion trapmass spectrometer was used in this study, coupled to a P580 HPLC pump (Dionex, Softron GmbH, Germering, Germany). Samples were acquired and processed with Xcalibur 1.0 sp1 software in a m/z range from 100 to 2000. Samples were introduced to the instrument using flow injection, and the samples were drawn into the 5-µL sample loop of the injector through suction, to avoid contact with the iron needle of the syringe. Flow injection was selected as the most appropriate sample introduction approach as it (11) Groenewold, G. S.; Van Stipdonk, M. J.; Gresham, G. L.; Chien, W.; Bulleigh, K.; Howard, A. J. Mass Spectrom. 2004, 39, 752-761. (12) Witter, A. E.; Hutchins, D. A.; Butler, A.; Luther, G. W. Mar. Chem. 2000, 69, 1-17.
7336
Analytical Chemistry, Vol. 77, No. 22, November 15, 2005
minimizes any changes within the sample as a result of handling, contamination, or change in solvent/mobile phase. The solvent flow rate was 150 µL/min, the source voltage ( 4.5 kV, the capillary voltage 15 V, the capillary temperature 220 °C, the tube lens offset -10 V, the sheath gas flow rate 60 arbitrary units, and the auxiliary gas flow rate 20 arbitrary units. The instrument can analyze positive or negative ions, and this is controlled by the sign of the source voltage. The positive mode generates a positive ion spectrum. Material and Methods. Deionized water (18.2 MΩ) was obtained from a MilliQ (MQ) water purification system (Millipore) or G Chromasolv water (Fluka) was used. Methanol was LCMS grade (Riedel de Hae¨n), and all other reagents analytical grade. All glassware and equipment were soaked in 2% Decon overnight, rinsed with MQ water, soaked in 10% hydrochloric acid overnight, rinsed with MQ water, and dried in a laminar flow hood. All sample preparation and handling was performed under laminar flow to minimize airborne metal contamination. Stock aqueous standards (1.52 mM) were prepared from desferrioxamine mesylate salt (95%, Sigma-Aldrich) and Fe-free ferrichrome from Ustilago sphaerogena (Sigma-Aldrich). Working 300 and 150 µM stock solutions of Fe and Th in 2% HNO3 were prepared from ICP standards (Fisher Chemicals and BDH Chemicals, respectively). Other Fe stock solutions were prepared freshly from FeCl3 (Fisher Chemicals) in deionized water. Metals and complexing agents were mixed in a variety of molar ratios to investigate their interactions. Dilute ammonia was used to adjust the pH, which was checked with small range pH paper. Samples were prepared in triplicate to investigate the following: Identification, Quantification, and Peak Suppression. The 1:1 mixtures (15 µM) of desferrioxamine and ferrichrome with Th were prepared at pH 5, 7, and 9, equilibrated for 24 h, and analyzed to establish the signals associated with the species formed in the positive and negative modes of the instrument. Quantitation and detection limit were evaluated from calibration graphs. The 15 µM standards of each siderophore were prepared containing 0-32 mM HNO3, representative of the range of nitric acid that was introduced through the addition of acidified metal standards. These were neutralized and analyzed to assess the effect of the nitrate on peak suppression. Complexation. The rate at which Th is complexed by each siderophore was investigated in both 100% water (adjusted to pH
Figure 2. ESI-MS mass spectrum showing the characteristic m/z signals of Th-ferrichrome (m/z 916) and Fe-ferrichrome (m/z 741) complexes.
7 and 9) and 50% water, 50% methanol (pH 5, 7, and 9). Water was used as an environmentally relevant solvent while the 50% methanol was used to investigate whether there were differences in the complexation rates in a typical solvent used in the separation of siderophores,13 since a long-term aim is to use chromatography to allow investigation of complex environmental samples. The metal was added to produce a final 15 µM concentration of metal and siderophore at time zero. There was a delay from mixing to delivery into the instrument of 2-3 min, and measurements were recorded for up to 50 h. Competitive Interactions between Siderophores and Metals. Equimolar (15 µM) Fe was added to aqueous 1:1 Thsiderophore solutions at pH 5, 7, and 9, and the change in the complexes present was monitored over time. Thorium, ferrichrome and desferrioxamine, and Fe, ferrichrome and desferrioxamine, were mixed in 1:1:1 ratios (15 µM) and equilibrated overnight to investigate whether the two metals are preferentially bound by one siderophore over the other. Interactions of Siderophores and Complexes with a Cation-Exchange Resin. Since Th forms positively charged complexes with desferrioxamine and ferrichrome, there are questions relating to the environmental mobility of these complexes. A cation-exchange resin, Dowex 50W-X8 (20-50 U.S. mesh; BDH Chemicals Ltd.), was used to investigate the interactions of the complexes with static sulfonate groups, as a model of an environmentally relevant soil binding site.14 Glass Pasteur pipets were packed to 1 mL, and the void volume was determined to be 500 µL with KMnO4. Initially, each column was preconditioned with several column volumes of 3 M HCl to remove any residual Fe from the column 15 and then rinsed with water at the desired pH until the effluent was at the correct pH. One void volume of 150 µM siderophore-Th aqueous solution at a given pH was added to a column, followed by a void volume of freshly prepared 150 µM Fe solution at the same pH and four additional void volume washes. Each experiment was carried out in triplicate. Flame AAS analysis (SpectrAA 400 plus, Varian) was used to (13) McCormack, P.; Worsfold, P. J.; Gledhill, M. Anal. Chem. 2003, 75, 26472652. (14) Kertesz, M. A. FEMS Microbiol. Rev. 2000, 24, 135-175. (15) Saito, N. Pure Appl. Chem. 1984, 56, 523-539.
Table 1. m/z Signals and Species Assignmentsa m/z 561 281 585 614 789 688 712 741 916
species detected DFOH+ DFOH22+ DFOAlH+ DFOFeH+ DFOTh+ FCH+ FCAlH+ FCFeH+ FCTh+
description of m/z 560 + 1 (560 + 1)/2 560 - 3 + 27 + 1 560 - 3 + 56 + 1 560 - 3 + 232 687 + 1 687 - 3 + 27 + 1 687 - 3 + 56 + 1 687 - 3 + 232
a Desferrioxamine is denoted DFO and ferrichrome FC. The metal complexes involve triply deprotonated siderophores, as shown in the description of m/z. The same species were detected at pH 5, 7, and 9.
assess the interactions of Fe with the resin in the absence of siderophores and less than 10% was eluted from the column, with all detectable Fe eluted in the two void volumes following application to the column. RESULTS AND DISCUSSION 1. Identification, Quantification and Peak Suppression. Identification and Speciation. A major advantage of using ESIMS for the direct analysis of complexes, for example, using flow injection rather than chromatography, is the ability to identify several different complexes simultaneously. This was demonstrated particularly well in this work since all species of interest were detected optimally as positive ions and could be compared in a single spectrum (see Figure 2 and Table 1). Table 1 includes Al(III) complexes, since Al was identified as a contaminant (see below) in some samples. All of the species studied form predominantly singly charged ions in the ion source, although doubly charged desferrioxamine ions also form in variable proportions and both the singly and doubly charged species need to be monitored. The predominant m/z peaks (Table 1) were independent of pH between pH 5 and 9. It is important to note that there were no significant peaks in the higher m/z range (1000-2000); Analytical Chemistry, Vol. 77, No. 22, November 15, 2005
7337
Figure 3. Signal suppression from the addition of nitric acid and subsequent neutralization (pH 7) for (A) ferrichrome and (B) desferrioxamine. The signals from all of the particular siderophore species were summed. Error bars show one standard deviation for three replicate solutions.
thus, 1:1 Th-siderophore complexes predominate under these conditions. When the instrument is in positive mode, the Fe(III) and Al(III) complexes are protonated in the ion source because the 3+ ions form neutral complexes with the deprotonated siderophore ions. The Th complexes are not protonated in the ion source because the complexes are naturally charged ([Th4+ ligand3-]+). Therefore, the spectra show that Th is bound as a simple Th4+ ion, Fe as Fe3+, and Al as Al3+ in these complexes, regardless of the pH. Thorium hydrolyses above pH 1, has been found to be complexed by perchlorate as a hydrolyzed species,10 and so would be expected to be hydrolyzed at all of the pH values studied. The interaction with hydroxamate siderophores is therefore sufficiently favorable for complete ligand exchange to occur, within the sensitivity of the technique. Uranyl (U(VI)O22+) has been shown to be complexed as a dioxo ion by desferrioxamine,11 resulting in a peak at m/z 829 (U (238) + 2O (32) + singly deprotonated desferrioxamine (559)). If U(IV) was bound as a simple ion, analogously to Th, it would produce a signal at m/z ) 795 (U4+ (238) + triply deprotonated desferrioxamine (557)), and hydrolysis products would form intermediate signals (e.g., U(OH)22+ (272 + singly deprotonated desferrioxamine (559) ) 831). Therefore, both oxidation state and extent of hydrolysis can be examined simultaneously by ESI-MS. X-ray crystallography of Pu(IV)-desferrioxamine E has shown the presence of three aqua ligands in the complex.16 In this study, there was no evidence of coordinated aqua ligands. The capillary temperature used (220 °C), following Gledhill7 and general practice, may cause loss of aqua ligands, and thus, the data do not discount the possibility of aqua ligands coordinated to the metal center of the complex at room temperature. Contamination. Iron and Al contamination were apparent in all samples despite their preparation under laminar flow and the rigorous cleaning protocol used. This highlights an area of caution when relatively low concentrations of actinides and complexing agents are investigated using techniques with less transparent ability to detect contamination. Peak Suppression and Calibration. The addition of nitric acid to siderophore solutions, followed by pH adjustment to pH7, demonstrated the extent of signal suppression from nitrate (Figure 3). The 15 µM standards prepared using a Th standard in 2% nitric acid contained 26 mM nitrate. Therefore, the ferrichrome signal (16) Neu, M. P.; Matonic, J. H.; Ruggiero, C. E.; Scott, B. L. Angew. Chem., Int. Ed. 2000, 39, 1442.
7338
Analytical Chemistry, Vol. 77, No. 22, November 15, 2005
would be suppressed by 1 order of magnitude, and the desferrioxamine signal 5-fold. After correction for nitrate suppression (from Figure 3), excellent (R2 ) 0.999) linear calibration graphs of Th-siderophore complexes can be constructed. Therefore, to achieve quantification, suppression should be accounted for indirectly, as achieved here, by matrix-matching samples and calibrants or by standard addition. This is relevant for natural water and soil porewater samples, which may have high nitrate loadings. Detection Limits. The typical detection limit for the Thsiderophore complexes (blank plus 3σ) was 0.5-1 mg L-1, which corresponds to 1-2 µM. This is at the higher end of soil siderophore concentrations.17,18 2. Complex Formation. Complexation of Th by the Siderophores. In general, when Th was added to separate solutions of the siderophores, both siderophores complexed the Th rapidly (Table 2) with the complexes forming within the 2-3 min before it was possible to take the first reading. The exception to this was in 1:1 methanol-water at pH 9, where the kinetics were much slower, and for two replicates, only half of the equilibrium complex concentration was reached by the time of the first reading. Since the kinetics were rapid in all other cases, this suggests that the combination of methanol and high pH slows the rate of reaction, and the outlier may reflect the low precision of the pH adjustment. However, when samples are prepared in water, the complexation kinetics are rapid and there is no need to allow an extended period for equilibrium to be reached. This is in agreement with the fast complexation rates previously determined for Fe with hydroxamate siderophores.12 Complexation at Equilibrium. At equilibrium, thoriumdesferrioxamine comprised 52 ( 8% of the total desferrioxamine signal, and thorium-ferrichrome comprised 57 ( 5% of the total ferrichrome signal over the range of conditions studied. The standard deviations over these different experiments were comparable to those between the three replicates of each set ((510%), suggesting that the pH values and solution matrixes investigated do not have a significant affect on the final equilibrium position. Despite some Fe and Al contamination, the free siderophore was the second most dominant peak in every spectrum after the Th complex, showing that Th does not form a complex with (17) Buyer, J. S.; Kratzke, M. G.; Sikora, L. J. Appl. Environ. Microbiol. 1993, 59, 677-681. (18) Crowley, D.; Reid, C.; Szaniszlo, P. In Iron transport in microbes, plants and animals; Winklemann, G., van der Helm, D., Neilands, J., Eds.; VCH: New York, 1987; pp 375-386.
Table 2. Complexation of Th by Ferrichrome (FC) and Desferrioxamine (DFO)a
siderophore
pH
equilibrated after 2-3 min?
% DFO or FC signal associated with Th at equilibrium
5
yes
56 ( 4
7 9
yes no
62 ( 5 60 ( 5
H2O
7 9 5
yes yes yes
58 ( 9 50 ( 10 52 ( 6
1:1 MeOH-H2O
7
yes
62 ( 6
9
no
47 ( 3
7 9
yes no
43 ( 6 58 ( 10
solvent
FC 1:1 MeOH-H2O
DFO
H2O
comment
gradual exchange from 30 to 40% (3 min) to 40-50% (400 min) to 55% (3000 min) for two replicates, and one replicate reached equilibrium by 9 min
gradual exponential exchange from 10 to 25% (2-3 min) to 45-52% by 270-440 min reaction complete by 3-4 min
a The proportions of the Th complex are described in terms of the Th-siderophore signal divided by the sum of the signals from all complexes involving the siderophore. The uncertainties show one standard deviation between three replicate experiments.
Table 3. Extent and Rate of Metal Exchange between the Siderophore-Th Complexes in the Presence of Fea
siderophore ferrichrome
desferrioxamine
solvent 1:1 MeOH-H2O
H2O 1:1 MeOH-H2O
H2O
% siderophore signal from Fe complex by first measurement
time to reach equilibrium (min)
% siderophore signal from Fe complex at equilibrium
5
74 ( 1
between 7 and 34
91 ( 2
7 9 7 5
76 ( 1 80 ( 3 80 ( 2 84 ( 1
24 26 between 8 and 100 between 7 and 45
91 ( 2 92 ( 1 96 ( 1 97 ( 1
7 9 7
86 ( 3 91 ( 9 81 ( 2
between 7 and 20 between 7 and 20 between 8 and 125
97 ( 2 99 ( 1 98 ( 1
pH
a The signal from all species of a given siderophore were summed to give the total siderophore signal. The first measurement was taken within 3 min of mixing Fe with the complex. The uncertainties show one standard deviation between three replicate experiments.
all of the available siderophore. This is surprising given a published stability constant of Th-desferrioxamine of >1027.2 3. Competitive Interactions between Siderophores and Metals. Addition of Iron to Thorium-Siderophore Complexes. Before an equimolar amount of Fe was added to equilibrated 1:1 molar Th-siderophore mixtures, 54 ( 15% of the total desferrioxamine signal and 77 ( 5% of the total ferrichrome signal resulted from the Th complexes. Background Fe contamination was present in the Th-siderophore mixtures, representing 13 ( 5% of the desferrioxamine signal and 2 ( 1% of the ferrichrome signal, which resulted in the final concentration of iron being in slight molar excess of the siderophore and Th. There was also Al contamination in the desferrioxamine thorium mixture, which comprised 99.2% of the Th associated with desferrioxamine, and >97.6% of the Th associated with ferrichrome, were retained over the remaining washes. These data show that either the charged complexes interact with the resin or the sulfonate groups on the resin extract Th from the siderophore chelate. The pH 5 ferrichrome results provide support for the argument that Th is released from the siderophore complex, since 100 ( 42% of the ferrichrome eluted from the column at pH 5. However, this was not observed to the same extent at higher pH values or with desferrioxamine, and the cumulative uncertainties associated with the quantities eluted make definitive interpretation difficult. Overall, the results show that Th in charged siderophore complexes is not readily transported through a medium rich in negatively charged binding sites, which relates to a soil, and that the Th could be retained either as a complex or as the metal ion, following exchange. CONCLUSIONS This study has demonstrated the power of FI-ESI-MS to measure siderophores and their metal complexes at micromolar concentrations and, thus, to observe complex formation and competitive interactions in perturbed systems at environmentally realistic concentrations. The simple spectra provide informative data, which allow the oxidation state of the metal and extent of hydrolysis to be deduced quickly and easily. This study has provided novel information on the interactions between Th and two siderophores in isolation and in the presence of Fe. The data show that there is significant, but not complete, complexation in 15 µM equimolar mixtures of siderophore and Th and that Fe rapidly replaces a large proportion of the Th in the complexes. These interactions were largely independent of pH from pH 5 to 9. Ferrichrome had a higher affinity for Th and lower affinity for Fe than desferrioxamine, and therefore, a larger, although still
small, fraction remained complexed to Th in the presence of Fe. Both Th complexes interacted with sulfonate groups on a cationexchange resin, which suggests that formation of charged Thsiderophore complexes does not result in a proportionate increase in Th mobility in soils. In conclusion, ESI-MS has been shown to be a powerful analytical tool that can be used to gain a greater understanding of actinide-siderophore interactions in multicomponent, competitive systems both at equilibrium and following perturbations. Development of a chromatographic separation of the complexes from bulk soil porewater constituents, including the suppressant nitrate, will allow this technique to be applied to the analysis of Th-siderophore complexes in real world samples either directly or following preconcentration.
ACKNOWLEDGMENT We thank the Nuffield Foundation for financial support, Prof. Stefano Girotti and the University of Bologna for organizing and funding M.V.B.’s exchange to Plymouth, Martha Gledhill for advice on analyzing siderophore complexes by ESI-MS and stimulating discussion, and Paul McCormack and Pippa CurtisJackson for help troubleshooting instrumental problems.
Received for review June 17, 2005. Accepted September 19, 2005. AC051069Y
Analytical Chemistry, Vol. 77, No. 22, November 15, 2005
7341
