Complexation of uranium(VI) with peptidoglycan

0 downloads 0 Views 206KB Size Report
Oct 22, 2008 - sites of glutamic acid and diaminopimelic acid (pKa = 4.55 ± 0.02 and ... pKa for hydroxyl and amino groups (which are not distinguishable) ...
PAPER

www.rsc.org/dalton | Dalton Transactions

Complexation of uranium(VI) with peptidoglycan† Astrid Barkleit,* Henry Moll and Gert Bernhard Received 22nd October 2008, Accepted 31st March 2009 First published as an Advance Article on the web 21st May 2009 DOI: 10.1039/b818702a We investigated the interaction of UO2 2+ with peptidoglycan (PG), the main part of the outer membrane of Gram-positive bacteria, by potentiometric titration and time-resolved laser-induced fluorescence spectroscopy (TRLFS) over a wide pH (2.0 to 9.0) and concentration range (10-5 to 10-4 M U(VI), 0.01 to 0.2 g L-1 PG). With potentiometry two different dissociation constants for the carboxyl sites of glutamic acid and diaminopimelic acid (pK a = 4.55 ± 0.02 and 6.31 ± 0.01), and one averaged pK a for hydroxyl and amino groups (which are not distinguishable) (9.56 ± 0.03) and the site densities could be identified. With potentiometry three different uranyl PG complexes were ascertained: two 1 : 1 uranyl carboxyl complexes R–COO–UO2 + , one with the glutamic acid carboxyl group (log b 110 = 4.02 ± 0.03), which has a very small formation ratio, and one with the diaminopimelic acid carboxyl group (log b 110 = 7.28 ± 0.03), and a mixed 1 : 1 : 1 complex with additional hydroxyl or amino coordination, R–COO–UO2 (+) –Ai –R (Ai = NH2 or O- ) (log b 1110 = 14.95 ± 0.02). With TRLFS, also three, but different, species could be identified: a 1 : 1 uranyl carboxyl complex R–COO–UO2 + (log b 110 = 6.9 ± 0.2), additionally a 1 : 2 uranyl carboxyl complex (R–COO)2 –UO2 (log b 120 = 12.1 ± 0.2), both with diaminopimelic acid carboxyl groups, and the mixed species R–COO–UO2 (+) –Ai –R (Ai = NH2 or O- ) (log b 1110 = 14.5 ± 0.1). The results are in accordance within the errors of determination.

Introduction The influence of microorganisms on the transport behaviour of heavy metals in the environment is significant. Bacteria can immobilize or mobilize metal ions through diverse accumulation and complexation mechanisms.1 Bacteria have the highest surface to volume ratio of any life form.2 Therefore, soluble metal ions will interact mainly with bacterial surfaces. These surface interactions of uranyl ions could be demonstrated for various bacteria strains by transmission electron microscopic studies.3–7 The main binding sites of the bacterial cell surfaces for heavy metal ions are phosphoryl, carboxyl, hydroxyl, and amino groups. It is wellknown that the uranium ion has a high affinity to phosphoryl groups,4–12 but even so carboxyl binding has been observed.5,7,12 Francis et al.7 and Merroun et al.5 detected, besides phosphoryl coordination, also uranyl carboxyl complexation with EXAFS at pH about 5.0 and 4.5, respectively. Kelly et al.12 observed it even at lower pH (from 3.2 to 4.8), also with EXAFS. But at pH below 3 they could detect only phosphoryl complexation.12 In contrast, with TRLFS up to now only uranyl phosphoryl complexation could be observed.4,8,10 It is essential to investigate the complexation mechanisms on a molecular level to completely understand the surface interaction of biota like bacteria in relation to the biogeochemical cycle. There have been no systematic studies of reactions with isolated cell wall compartments on a molecular level so far. Therefore, we investigate the complexation of uranium with single cell wall constituents, Institute of Radiochemistry, Forschungszentrum Dresden-Rossendorf e.V., P.O Box 510119, D-01314, Dresden, Germany. E-mail: [email protected]; Fax: +49 351 260 3553; Tel: +49 351 260 2148 † Electronic supplementary information (ESI) available: Summary of all potentiometric titration and luminescence measurements, additional spectroscopic data. See DOI: 10.1039/b818702a

This journal is © The Royal Society of Chemistry 2009

which are assembled like a mosaic to get a basis for a general understanding of the whole complex biological system. After our previously published study dealing with lipopolysaccharide, a main constituent of Gram-negative bacteria,13 in this paper we want to present the uranium interaction with peptidoglycan, the main constituent of Gram-positive bacteria. Because of the high resistance of their spores, especially Grampositive bacterial strains from the genus Bacillus can be found in a large variety of natural habitats, including uranium waste piles.10 The cell wall of these Gram-positive bacteria is composed primarily of peptidoglycan (PG) which is in contact with the outer environment of the microbe. It includes teichoic acids, glycerol phosphate polymers which provide mainly phosphoryl groups for metal ion binding. The cytoplasm cellular membrane forms the inner layer of the cell wall. PG is the real framework of the cell wall and its dry weight can account for 30–50% of the dry weight of the wall.14 The PG macromolecule consists of repeating b(1–4)-linked N-acetylglucosaminyl-N-acetylmuramyl (NAG, NAM) dimers with a short stem of four amino acids (Fig. 1).15 The peptide chains are cross-linked through D-alanyl-diaminopimelyl bonds. The average degree of peptide cross-linking is 30–35%.15 Glutamic acid and diaminopimelic acid contain the free carboxyl groups of interest for metal ion complexation. Furthermore, amino groups of non-crosslinked amino acids or even amides and also hydroxyl groups from the sugars NAG and NAM are potential coordination sites for metal ions. Until now, either isolated cell walls16 or even whole cells16–19 have been used to determine deprotonation constants of its functional groups. These investigations can only possess composite values for a summary of similar functional groups which is not sufficient for identifying and characterising the absorption of metal ions on cell surfaces, and especially its transport into the cells on a molecular level. In general, stability constants of heavy metal Dalton Trans., 2009, 5379–5385 | 5379

the uranyl PG complex species at trace metal concentrations near environmental conditions. With TRLFS, uranyl speciation can be detected down to 10-8 M.27 Normally, the mean amount of uranium in ground water is approximately 10 mg L-1 (4.2 ¥ 10-8 mol L-1 ), but mining-related waters can contain up to 7 ¥ 10-5 mol L-1 uranium.28 In this study, we used a uranium concentration of 10-5 mol L-1 for TRLFS investigations.

Experimental section Solutions and reagents

Fig. 1 Structure of a peptidoglycan chain, based on the literature.43 NAM = N-acetylmuramic acid; NAG = N-acetylglucosamine.

complexes with cell surfaces are mostly determined from adsorption studies.17,18,20 Only a few studies have been published which contain stability constants of uranyl adsorption on bacteria.20,21 The authors determined uranyl carboxyl and phosphate surface complexes onto Bacillus subtilis20 and Shewanella putrefaciens.21 However, the stability constants indeed represent average constants containing all uranyl interactions at the whole cell surface. Neither potentiometric investigations of uranyl complexation with cell walls nor with isolated cell wall compartments of bacteria are known, except our previously published study about uranyl lipopolysaccharide interaction.13 Herein we determined the coordination behaviour with potentiometric titration and timeresolved laser-induced fluorescence spectroscopy (TRLFS). Three different uranyl phosphoryl complexes and one uranyl carboxyl complex could be identified and characterized. Investigations that provide TRLFS data are published only for the interaction of uranyl ions with whole cells.4,8,10,22 And, unfortunately, with this technique only uranyl phosphoryl interactions could be observed until now, due to the specific very strong luminescence behaviour of uranyl phosphate complexes.23 To get an understanding of the binding behaviour of uranium to other groups than phosphoryl, we examined the complexation reaction with isolated phosphate-free PG. Even early studies could show that isolated PG has a great potential for metal binding.24 Investigations with Gram-positive bacterial cells or cell walls could clearly demonstrate interactions of the uranyl ion in vivo with carboxyl groups of PG.9,12,20,25 One study, so far, has determined the dissociation constant of isolated PG, calculating only one average pK a of 7.70.26 In this paper we present the results of complexation studies between uranium(VI) and PG from Bacillus subtilis, studied by potentiometric titration and TRLFS over a wide pH range. While potentiometric titration provides the deprotonation constants of the functional groups of PG and stability constants of the uranyl complex species, TRLFS offers the possibility to detect 5380 | Dalton Trans., 2009, 5379–5385

Peptidoglycan (PG) from B. subtilis was purchased from Fluka and used without further purification. In water it is relatively poorly soluble, which restricted its experimental concentration to 0.2 g L-1 . UO2 2+ stock solutions were made from UO2 (ClO4 )2 ·6H2 O (natural uranium with 99.27% U-238, 0.72% U-235, 0.0054% U-234; Merck, p.A.). The ionic strength was kept constant for all experiments at 0.1 M by adding stock solutions from NaClO4 ·H2 O (Merck, p.A.). All solutions were prepared with carbonate-free deionized water. For it, the deionized water was boiled for 3 h to outgas the CO2 and then stored in an inert gas glove box under nitrogen. All experiments were carried out under inert gas atmosphere (nitrogen) excluding CO2 from air to avoid unwanted carbonate complexation. Potentiometric titrations The potentiometric titration experiments were carried out in a glove box under inert gas atmosphere (nitrogen) at 25 ± 1 ◦ C and a constant ionic strength of 0.1 M (NaClO4 ). In each case, the starting volume was 30 mL. For each ligand titration (5 measurements), 3 mg PG was dissolved in carbonate-free water, resulting in a PG concentration of 0.1 g L-1 . For complexation titrations (3 measurements), additionally 3 mmol UO2 2+ was added, resulting in a concentration of 10-4 M U(VI). Some of the mixtures were acidified with 6 mmol HClO4 (2 ¥ 10-4 M) to obtain a starting pH of about 4, others were used with the self-contained near-neutral pH. The solutions were titrated up with 10-3 M NaOH (standardized, Merck, Titrisol) and down with 10-3 M HClO4 (Merck, suprapure; exact concentration analyzed with standardized NaOH) in the pH range from about 4 to 9 (PG) and 8 (uranyl with PG), respectively. The pH values were measured with a BlueLine 16 pH electrode (Schott). The electrode was calibrated, for each experiment, with NBS buffers (4.01 and 6.86, Schott). All samples were titrated with an automatic titrator (TitroLine alpha, Schott) and monitored by the accompanying software (Titrisoft 2.11, Schott). The titration procedure was a dynamic titration with the highest available precision (titration speed “slow”) and a delay of at least 20 s at each titration point before measuring. The experimental conditions of all titrations are summarized in Table 1. The titration curves were analyzed using the program HYPERQUAD2006.29 The pK a values of the uranyl hydroxide species, corrected for I = 0.1 M, were included in analyzing the complex mixtures.30 The titration curves and fit results are to be found in the electronic supplementary information ESI, Fig. S1 and S2.† TRLFS measurements TRLFS spectra were measured with a fixed uranyl concentration of 10-5 M. Three measurement series were carried out: First This journal is © The Royal Society of Chemistry 2009

Table 1 Experimental conditions of the potentiometric titrations Sample

pH range

PG 1 (without HClO4 ) 2 (without HClO4 ) 3 (without HClO4 ) 4 (with HClO4 ) 5 (with HClO4 )

6.56–9.34 5.28–9.34 5.28–3.66 3.92–9.33 3.75–9.34

PG with uranyl 6 (with HClO4 ) 7 (with HClO4 ) 8 (with HClO4 )

4.44–7.50 4.52–7.46 4.33–7.52

Table 2 Calculated pK a values and site densities from potentiometric titration for PG from B. subtilis (25 ± 1 ◦ C, 0.1 M NaClO4 )

Data points pK a

Site density/mmol g-1

Sites

75 107 69 165 173

4.55 ± 0.02 6.31 ± 0.01 9.56 ± 0.03

0.65 ± 0.17 0.76 ± 0.02 1.45 ± 0.23

Carboxyl (glutamic acid) Carboxyl (DAPa ) Amine/hydroxyl

101 91 103

where R is the PG polymer with the attached functional groups Ai . The corresponding proton binding constant K a can be written as

a

at fixed PG concentration (0.1 g L-1 ) as a function of the pH (2.0–9.0), second and third at fixed pH = 2.5 ± 0.1 and pH = 4.0 ± 0.1, respectively, as a function of the PG concentration (0.01–0.20 g L-1 ). The samples were prepared in a glove box under inert gas atmosphere (nitrogen). Necessary pH adjustments were made with a BlueLine 16 pH electrode (Schott) using HClO4 or NaOH with an accuracy of 0.02 units. The solutions were allowed to equilibrate for 24 h. Before TRLFS measuring, the pH was determined again. The real pH values of all measured samples are to be found in the ESI, Fig. S3, S4, and S5.† The exact uranyl concentration at all measurements was detected by ICP-MS (inductively-coupled-plasma mass-spectrometry) with an accuracy of 5%. The spectra were recorded at 25 ± 1 ◦ C using a pulsed Nd:YAG laser system (Continuum Minilite Electro-Optics, Inc., Santa Clara, USA) with a fast pulse generator (FPG/05, EG & G Princeton Instruments, NJ, USA), and a digital delay generator (Uniblitz, model VVM-D1, Vincent Associates, NY). The excitation wavelength of the uranyl fluorescence was 266 nm with pulse energy of 0.2–0.5 mJ. To exclude possible energy absorption of the biomacromolecule, UV/vis measurements were made with 0.1 g L-1 PG at several pH values in the range between 190 and 350 nm (Cary 50, Varian Co.). The main absorption is around 210 nm. There is no absorption at 266 nm. UV/Vis spectra are to be found in the ESI, Fig. S9.† The TRLFS spectra were measured with a diffraction grating of 100 mm-1 from 371 to 674 nm, averaging three spectra with 200 laser shots each, and a gate delay of 100 ns. The time-resolved fluorescence emission was detected using an iHR 550 spectrograph (Horiba Jobin Yvon, Germany), controlled by the accompanying software LabSpec5 (Horiba Jobin Yvon, Germany). The baseline corrections, lifetime and peak maxima determinations of the spectra were carried out with Origin.31 The complex stability constants were determined using SPECFIT.32 Again the pK a values of the uranyl hydroxide species, corrected for I = 0.1 M, were included in analyzing the complex mixtures.30

Results and discussion Potentiometric titrations Peptidoglycan. The data of the pure ligand titrations were analyzed based on the deprotonation of discrete monoprotic acids according to the reaction R–Ai H  R–Ai - + H+ , This journal is © The Royal Society of Chemistry 2009

(1)

DAP = diaminopimelic acid.

(2) [R–Ai - ] and [R–Ai H] represent the concentrations of the deprotonated and protonated form of the functional group Ai , respectively, and [H+ ] represents the proton concentration in the solution. The following deprotonation reactions are possible in the aqueous PG solution: R–COOH  R–COO- + H+ ,

(3)

R–NH3 +  R–NH2 + H+ ,

(4)

R–OH  R–O- + H+ .

(5)

The best fit for all titration curves was obtained with a threesite model (see ESI, Fig. S1†). The results are summarized in Table 2. The pK a values of 4.55 ± 0.02 and 6.31 ± 0.01 with nearly equal site densities can both be dedicated to carboxyl groups. The PG molecule offers two different free carboxyl groups, from the glutamic acid and the diaminopimelic acid. The pK a values of the second carboxyl groups of glutamic acid (4.1533 ) and pimelic acid (5.0833 ) are within the same range. Electrostatic potentials, originating from charged groups nearby, the so-called polyelectrolyte effect, can influence the dissociation of the functional groups and cause a shift in dissociation constants in comparison to the single components.16 The free carboxyl group of non-crosslinked D-alanine units can be neglected in this study. Because the pK a value of free alanine is 2.33,33 it is not expected to be detected in the investigated pH range. The third pK a of 9.56 ± 0.03 can be dedicated to both, amino and hydroxyl groups. Related pK a (NH3 ) of glutamic acid (9.5833 ) or diaminohexanoic acid (lysine, 9.15 and 10.6633 ) and the pK a (NH3 ) for S. putrefaciens (10.0421 ) or pK a (OH) for B. subtilis cell walls (9.417 ) are within the same range. Hence, we assume that in contrast to the two specifiable carboxyl groups the amino and hydroxyl groups are not distinguishable. We can only determine an average value. Hence the relatively high site density is the sum of both functional groups. Uranium(VI) and peptidoglycan. The titration curves of the uranyl PG complex solutions were analyzed based on the formal complex formation equation for discrete binding sites xUO2 2+ + yR–Ai - + zH+  [(UO2 )x (R–Ai )y Hz ]x(2x-y+z)+

(6)

and the appropriate mass action law, which represents the complex stability constant log K: Dalton Trans., 2009, 5379–5385 | 5381

Fig. 2 Luminescence spectra of 10-5 M U(VI) with 0.1 g L-1 PG in dependency of pH.

(7)

As initial data, we used the pK a values and site densities from potentiometric titration of PG (see Table 2; for titration curves with fit and residuals see ESI, Fig. S2†). If we analyzed the titration data only up to pH 6.0, we could identify two different uranyl carboxyl complexes with glutamic and diaminopimelic acid. The 1 : 1 uranyl complex R–COO–UO2 + with the carboxyl group of glutamic acid with a stability constant of log b 110 = 4.02 ± 0.03 has only a very small formation ratio (see Fig. 4). The one with the carboxyl group from the diaminopimelic acid with the stability constant of log b 110 = 7.28 ± 0.03 is formed in higher amounts. If we increase the analyzing pH range up to about 9.3, the best fit is reached with an additional mixed uranyl complex with a carboxyl group from diaminopimelic acid and an amino or hydroxyl group in the coordination sphere. This 1 : 1 : 1 complex R–COO–UO2 (+) – Ai –R (with Ai = NH2 or O- ) has a stability constant of log b 1110 = 14.95 ± 0.02. All determined constants are summarized in Table 3. Up to now, only one comparable stability constant of a bacterial uranyl carboxyl complex has been determined, the surface complex R–COO–UO2 + with log K = 5.4 ± 0.2 for uranyl adsorption onto B. subtilis.20 This value lies between ours, indicating that it is an averaged value from different binding carboxyl groups.

TRLFS measurements Fig. 2 depicts one measurement series of the uranyl PG system. The luminescence was measured at fixed uranyl (10-5 M) and PG concentrations (0.1 g L-1 , equal to about 7 ¥ 10-5 M of each carboxyl group and 1. 5 ¥ 10-4 M amino and/or hydroxyl groups, see Table 2) between pH 2.0 and 9.0. At very low pH no change of the luminescence of the free uranyl ion was observed, which implies that the complexation is insignificant. Between pH 2.0 and 3.0 a slight decrease of the luminescence intensity, connected with a strong red shift of the peak maxima of about 8 nm at pH 3.0, can be observed (Fig. 2, left). Above pH 3.0 up to pH 5.6 the luminescence intensity increases again with the same constant red shift of the peak maxima. From pH about 5.6, the luminescence intensity decreases once more (Fig. 2, right). For comparison, the luminescence spectra of uranyl at pH 2.0 (UO2 2+ (aq)) (left) and pH 6.2 and 9.0 (right) are included. The luminescence intensity of the main uranyl hydroxide species at pH between 6 and 7, (UO2 )3 (OH)5 + (about 70%, spectroscopically characterized by Sachs et al.34 ) is very high, even in comparison to the free uranyl ion. Then, the luminescence intensity of the U(VI) hydroxides decreases up to pH 9 (main species: 50% (UO2 )3 (OH)7 - ,

Table 3 Calculated complex species and stability constants of the uranyl PG system (25 ± 1 ◦ C, 0.1 M NaClO4 ) Species

x y1 y2 y3 za log b xyz

R–COO–UO2 + (glutamic acid) R–COO–UO2 + (DAPb )

11000 10100

(R–COO)2 –UO2 (DAPb ) 10200 (R–COO)–UO2 (+) –Ai –R (COO- 1 0 1 1 0 b from DAP; Ai = NH2 or O )

4.02 ± 0.03 7.28 ± 0.03 6.9 ± 0.2 12.1 ± 0.2 14.95 ± 0.02 14.5 ± 0.1

Method Potentiometry Potentiometry TRLFS TRLFS Potentiometry TRLFS

x = metal ion (UO2 2+ ); y = ligand (y1 = COO- from glutamic acid, y2 = COO- from diaminopimelic acid, y3 = amin or hydroxyl); z = proton (H+ ). b DAP = diaminopimelic acid. a

5382 | Dalton Trans., 2009, 5379–5385

Fig. 3 Bi-exponential luminescence decay in the UO2 2+ /PG system at two different pH. At lower pH (3.0) the longer lifetime dominates, at higher pH (5.6) the shorter one dominates.

This journal is © The Royal Society of Chemistry 2009

50% UO2 (OH)3 - ), but is even higher than those of the uranyl PG complex mixtures. The spectra of the U(VI) mixed species at the same conditions, but without PG, are depicted in ESI, Fig. S7.† Fig. S8† shows the speciation of U(VI) over the measured pH range. The changes in the peak maxima and luminescence intensities of the uranyl PG system, which are very different from the uranyl hydroxides, imply that three complexes are formed, one which occurs at low pH with a lower luminescence intensity than the free uranyl ion, a second one with a higher luminescence intensity which dominates up to pH 5.6, and a third one which appears from pH 5.6 and shows no luminescence properties. The third complex prevents the formation of uranyl hydroxide species at least up to pH 7. In the slightly basic pH range the uranyl hydroxides appear, as is recognizable in the shape of the spectra. Two measurement series at fixed pH and with varying PG concentration (0.01–0.2 g L-1 ) confirm these observations. In the series at pH 2.5 (ESI, Fig. S3†), significant changes in the peak maxima could be observed only at PG concentrations greater than 0.1 g L-1 . At lower PG concentrations, only a slight decrease of the luminescence intensity without shift could be observed, possibly due to a static quench effect of the biomacromolecule. An appropriate series at pH 4.0 (ESI, Fig. S4†) showed an increase of the luminescence intensity, connected with a red shift of the peak maxima of about 8 nm, starting even at 0.025 g L-1 PG. The time-resolved measurements give information about the lifetimes of the exciting state of the luminescent species in the mixture. Thus, it provides further information on the number of the luminescent species. To evaluate the number of the luminescent species and their lifetimes, the integrated luminescence signal is fitted to a sum of exponential decay functions:

E (t ) =

 E exp(-1 / t ) i

i

(8)

E is the total luminescence intensity at the time t, E i the luminescence intensity of the species i at t = 0, and t i the corresponding lifetime. The spectra up to pH 3 show mono-exponential decay with lifetimes between 1.2 and 1.8 ms, which can be clearly identified as the free uranyl ion (1.7 ± 0.5 ms35 ). From pH 3, we observe mostly bi-exponential decay. At pH 3, a longer lifetime (averaged: 7.3 ± 1.4 ms) appears first and can be detected up to pH about 6.2. Its intensity is highest at pH 3, followed by a strong decrease. A second shorter lifetime (averaged 0.7 ± 0.1 ms) was detected from pH about 3.5 with a strong increasing intensity. It can be observed up to pH about 6.8. Afterwards only lifetimes of uranyl hydroxides are detectable: A shorter one of about 3.0 to 3.5 ms can be dedicated to (UO2 )4 (OH)7 + (literature: 4.2 ± 0.4 ms34 ), and a longer lifetime between 15 and 19 ms can be assigned to (UO2 )3 (OH)5 + (literature 19.8 ms34 ), the two main uranyl hydroxide species in near neutral to slightly basic pH. The progression of the lifetimes (depicted in Fig. 3) gives the same information as the development of the peak maxima and luminescence intensities: one uranyl PG complex with characteristic luminescence properties begins to form at pH 3, but will be overlapped soon by a second uranyl PG complex with stronger luminescence intensity. A third complex, which appears at pH about 6, shows no luminescence properties. The quantitative investigation of the luminescence spectra to calculate complex stability constants is based on the formal complex formation reaction (6) with the appropriate mass action law (7) for discrete binding sites in the biomacromolecule. As initial data we used again the pK a values and site densities of PG, calculated from the potentiometric titrations (see Table 2), and the pK a values for the uranyl hydroxide species.30 The best fit for the measurement series at pH 4.0 was reached with the 1 : 1 uranyl carboxyl complex (R–COO–UO2 + ) and the 1 : 2 uranyl carboxyl complex (R–COO)2 –UO2 , both with carboxyl groups from diaminopimelic acid. Because of the small formation of the uranyl carboxyl species from glutamic acid (below 5%; see Fig. 4)

i

Fig. 4 Speciation of the uranyl PG system with 10-4 M U(VI) (left), 10-5 M U(VI) (right), and 0.1 g L-1 PG (equal to 6.5 ¥ 10-5 M glutamic acid, 7.6 ¥ 10-5 M diaminopimelic acid, 1.45 ¥ 10-4 M amine and/or hydroxyl groups). Left: Potentiometry conditions with a slight deficit of binding sights; right: TRLFS conditions with a slight excess of binding sites. U–PG 1 = R–COO–UO2 + (glutamic acid), U–PG 2 = R–COO–UO2 + (diaminopimelic acid), U–PG 3 = (R–COO)2 –UO2 , and U–PG 4 = R–COO–UO2 (+) –Ai –R (Ai = NH2 or O- ). The right diagram contains the uncertainty (in the form of error lines) resulting from the uncertainties of the stability constants determined from TRLFS.

This journal is © The Royal Society of Chemistry 2009

Dalton Trans., 2009, 5379–5385 | 5383

Table 4 Summary of measured luminescence data Species Uranyl (aq) species UO2 2+ pH 2.0 (100% UO2 2+ ) U(VI) pH 6.2 72% (UO2 )3 (OH)5 + 11% (UO2 )4 (OH)7 + 11% UO2 OH+ (UO2 )3 (OH)5 + 34 (UO2 )4 (OH)7 + 34 UO2 OH+ 38 Uranyl PG complexes R–COO–UO2 + (R–COO)2 –UO2 R–COO–UO2 (+) –Ai –R (Ai = NH2 or O- )

Peak maxima/nm 470.8

Lifetime/ms 488.8 496

533.0 533

559.0 557

586.3 583

484

498

514

534

557

583

482

498

519

543

570

599

539(2)

566(1)

595(3)

466.0(8) 481.6(6) 498.1(3) 518.0(6) No luminescence at room temperature

and the accuracy of TRLFS, which is restricted to 5% at best, it was not possible to detect this complex. The formation ratio of this complex is below the detection limit at these conditions. For the series with varying pH between 2.0 and 9.0 the best fit was reached with the two uranyl carboxyl species R–COO–UO2 + and (R–COO)2 –UO2 from diaminopimelic acid plus the mixed complex (R–COO–UO2 (+) –Ai –R with Ai = NH2 or O- ), which was found by potentiometry, too. The average stability constants were calculated to be log b 110 = 6.9 ± 0.2 and log b 120 = 12.1 ± 0.2 for the two uranyl carboxyl complexes, and log b 1110 = 14.5 ± 0.1 for the mixed complex (see Table 3). The values of the 1 : 1 complex and the mixed species are in good accordance with those determined by potentiometry. Because of the stoichiometry used for potentiometry (a slight deficit of the complexation sites), the probability of the formation of a 1 : 2 species is low, therefore we could not detect this species with potentiometry which we found with TRLFS. The speciation for both systems is depicted in Fig. 4.36 The structures of these complexes possibly follow the proposed structure from Texier et al. for Ln3+ ions bound in peptidoglycan.37 The uranyl ion is embedded between two PG strings, surrounded by the carboxyl groups of the respective diaminopimelic acids and partly with one non-crosslinked amino group of a diaminopimelic acid. A combined carboxyl and amino coordination is sterically more feasible than with a hydroxyl group, though a coordination of the uranyl ion with an amino group seems to be more unlikely than hydroxyl coordination, because it could never be detected in bacterial systems until now. Only uranyl hydroxyl coordination was considered by Fowle et al.20 in their adsorption model with a stability constant value of log K between 13 and 14, but this species could not be safely identified by the authors. We can assign the observed luminescence properties to the uranyl PG complex species. The 1 : 2 uranyl carboxyl PG complex (R–COO)2 –UO2 dominates around pH 6. It has strong luminescence intensity, causes a red shift of the peak maxima of about 8 nm and has a luminescence lifetime of about 0.7 ms. The 1 : 1 uranyl carboxyl PG complex R–COO–UO2 + appears between pH 3 and 5 (see Fig. 4, right) but is already at pH about 4.5 covered by the 1 : 2 uranyl carboxyl PG complex. The 1 : 1 complex has a lower luminescence intensity than UO2 2+ (aq) and a longer lifetime of about 7.3 ms. Because of its significantly lower luminescence intensity and the quite strong covering by the 1 : 2 complex and the free uranyl ion, the deconvolution of the 5384 | Dalton Trans., 2009, 5379–5385

510.0 512

1.4 ± 0.1 15.9 ± 1.5 3.2 ± 0.9 30.0 ± 4.8 19.8 ± 1.8 4.2 ± 0.4 35.0 ± 2.0 7.3 ± 1.4 0.7 ± 0.1

measured sum spectra (with SPECFIT) could not result in two clearly distinguishable single spectra (see ESI, Fig. S6†). These two uranyl carboxyl coordinated PG complexes have the same shift of the peak maxima. They are only specifiable by the different lifetimes. The uranyl PG complex with the mixed coordination R–COO–UO2 (+) –Ai –R (Ai = NH2 or O- ) shows no luminescence behaviour at room temperature. The 1 : 1 uranyl carboxyl complex with glutamic acid coordination could not be observed with TRLFS. The spectroscopic data of the uranyl PG complex species are summarized in Table 4. These different luminescence effects have already been reported in the literature. Uranyl complexes with organic ligands which contain carboxyl groups can show luminescence and effect a red shift of the peak maxima, as was observed for the uranyl complex systems with malonate,38 a-aminoisobutyrate,39 glycine,40 or L-threonine.41 Other compounds have no luminescence properties, like the uranyl carboxyl complexes with glycolate,39 a-hydroxyisobutyrate,39 or L-cysteine.40

Conclusions Interactions of uranium with bioligands can be conveniently examined in aqueous solution with a combination of potentiometry and luminescence spectroscopy over a wide pH and concentration range. TRLFS, in particular, allows the investigation of the speciation of actinides in complex natural systems up to trace concentrations. The complexation of the uranyl cation with peptidoglycan (PG) from Bacillus subtilis has been studied using potentiometric titration and time-resolved laser-induced fluorescence spectroscopy (TRLFS). As a result, four uranyl complexes could be detected in combination of both methods within a broad uranyl (10-5 to 10-4 M) and ligand concentration range (0.01 to 0.2 g L-1 ) and also over a wide pH range (2.0 to 9.0). Three complexes are only carboxyl coordinated. With potentiometry, we could distinguish between coordination of glutamic acid and diaminopimelic acid of the PG molecule. A fourth complex species has an additionally coordination via amino or hydroxyl groups, which appears at slightly higher pH. The determined stability constants of the complexes which could be detected with both methods are within the error limits. In addition to the overlapping area, both methods provide special conditions to detect more, completing the overall picture. This journal is © The Royal Society of Chemistry 2009

The calculated speciations (Fig. 4) show the dominance of the uranyl PG complexes at higher uranyl concentration and a deficit of binding sites (left) as well as at lower, more environmentally relevant, uranyl concentrations with a slight excess of binding sites (right) over a wide pH range. For the first time, the luminescence behaviour of an isolated phosphate-free cell wall compartment could be studied. It is wellknown that inorganic and organic phosphates strongly enhance the luminescence intensity of the uranyl ion.41,42 Consequently, phosphate coordination can cover the luminescence of uranyl complexes coordinated with carboxyl or other functional groups, as was determined for the uranyl lipopolysaccharide system.13 Here we could show that uranyl carboxyl complexes do have luminescence properties. The determined spectra and luminescence characteristics (lifetimes, peak maxima, and luminescence intensities) provide a basis for future TRLFS studies at complex biological systems to identify carboxylic bound uranyl species as well. The determined dissociation constants of the biomacromolecule and the appropriate stability constants of the uranyl complexes are important data for a better description of the interaction of bacterial surfaces with heavy metals in the biogeochemical cycle.

Acknowledgements This work was funded by the BMWi under contract number 02E9985. Furthermore, we would like to thank Juliane Schott and Peggy J¨ahnigen for carrying out most of the experimental work, and U. Schaefer for ICP-MS measurements.

13 14 15 16 17 18 19

20 21 22 23 24 25 26 27 28 29 30

31

References 1 T. J. Beveridge and R. J. Doyle, Metal ion and bacteria, John Wiley & Sons, Inc., New York, N.Y., 1989. 2 T. J. Beveridge, Can. J. Microbiol., 1988, 34, 363–372. 3 S. Krueger, G. J. Olson, D. Johnsonbaugh and T. J. Beveridge, Appl. Environ. Microbiol., 1993, 59, 4056–4064; F. Jroundi, M. L. Merroun, J. M. Arias, A. Rossberg, S. Selenska-Pobell and M. T. GonzalezMunoz, Geomicrobiol. J., 2007, 24, 441–449. 4 M. L. Merroun, G. Geipel, R. Nicolai, K. H. Heise and S. SelenskaPobell, BioMetals, 2003, 16, 331–339. 5 M. L. Merroun, J. Raff, A. Rossberg, C. Hennig, T. Reich and S. Selenska-Pobell, Appl. Environ. Microbiol., 2005, 71, 5532–5543. 6 M. Merroun, M. Nedelkova, A. Rossberg, C. Hennig and S. SelenskaPobell, Radiochim. Acta, 2006, 94, 723–729; L. E. Macaskie, K. M. Bonthrone, P. Yong and D. T. Goddard, Microbiology (Reading, U. K.), 2000, 146, 1855–1867. 7 A. J. Francis, J. B. Gillow, C. J. Dodge, R. Harris, T. J. Beveridge and H. W. Papenguth, Radiochim. Acta, 2004, 92, 481–488. 8 R. Knopp, P. J. Panak, L. A. Wray, N. S. Renninger, J. D. Keasling and H. Nitsche, Chem.–Eur. J., 2003, 9, 2812–2818. 9 Y. Andres, H. J. MacCordick and J.-C. Hubert, FEMS Microbiol. Lett., 1994, 115, 27–32. 10 P. Panak, J. Raff, S. Selenska-Pobell, G. Geipel, G. Bernhard and H. Nitsche, Radiochim. Acta, 2000, 88, 71–76. 11 M. Merroun, C. Hennig, A. Rossberg, T. Reich and S. Selenska-Pobell, Radiochim. Acta, 2003, 91, 583–591. 12 S. D. Kelly, K. M. Kemner, J. B. Fein, D. A. Fowle, M. I. Boyanov, B. A. Bunker and N. Yee, Geochim. Cosmochim. Acta, 2002, 66, 3855–3871.

This journal is © The Royal Society of Chemistry 2009

32 33 34 35 36 37 38 39 40 41 42

43

A. Barkleit, H. Moll and G. Bernhard, Dalton Trans., 2008, 2879–2886. T. J. Beveridge, Annu. Rev. Microbiol., 1989, 43, 147–171. A. D. Warth and J. l. Strominger, Biochemistry, 1971, 10, 4349–4358. A. vanderWal, W. Norde, A. J. B. Zehnder and J. Lyklema, Colloids Surf., B, 1997, 9, 81–100. J. B. Fein, C. J. Daughney, N. Yee and T. A. Davis, Geochim. Cosmochim. Acta, 1997, 61, 3319–3328. C. J. Daughney and J. B. Fein, J. Colloid Interface Sci., 1998, 198, 53– 77; S. Markai, Y. Andres, G. Montavon and B. Grambow, J. Colloid Interface Sci., 2003, 262, 351–361. J. S. Cox, D. S. Smith, L. A. Warren and F. G. Ferris, Environ. Sci. Technol., 1999, 33, 4514–4521; R. E. Martinez, D. S. Smith, E. Kulczycki and F. G. Ferris, J. Colloid Interface Sci., 2002, 253, 130– 139; J. B. Fein, J. F. Boily, N. Yee, D. Gorman-Lewis and B. F. Turner, Geochim. Cosmochim. Acta, 2005, 69, 1123–1132. D. A. Fowle, J. B. Fein and A. M. Martin, Environ. Sci. Technol., 2000, 34, 3737–3741. J. R. Haas, T. J. Dichristina and R. Wade, Chem. Geol., 2001, 180, 33–54. P. Panak, S. Selenska-Pobell, S. Kutschke, G. Geipel, G. Bernhard and H. Nitsche, Radiochim. Acta, 1999, 84, 183–190. Y. Kato, G. Meinrath, T. Kimura and Z. Yoshida, Radiochim. Acta, 1994, 64, 107–111; C. Moulin, P. Decambox and L. Trecani, Anal. Chim. Acta, 1996, 321, 121–126. T. H. Matthews, R. J. Doyle and U. N. Streips, Curr. Microbiol., 1979, 3, 51–53; B. D. Hoyle and T. J. Beveridge, Can. J. Microbiol., 1984, 30, 204–211. T. J. Beveridge and R. G. E. Murray, J. Bacteriol., 1980, 141, 876–887. Z. Vajtner and B. Suskovic, Acta Pharm. Jugosl., 1988, 38, 3–10. C. Moulin, Radiochim. Acta, 2003, 91, 651–657. G. Bernhard, G. Geipel, V. Brendler and H. Nitsche, J. Alloys Compd., 1998, 271–273, 201–205. P. Gans, A. Sabatini and A. Vacca, Talanta, 1996, 43, 1739–1753. R. Guillaumont, T. Fangh¨anel, J. Fuger, I. Grenthe, V. Neck, D. A. Palmer and M. H. Rand, Update on the Chemical Thermodynamics of Uranium, Neptunium, Plutonium, Americium and Technetium, Elsevier, Amsterdam, 2003. OriginPro 7.5G SR6, OriginLab Corporation, Northhampton, MA, USA, 2006. ¨ R. A. Binstead, A. D. Zuberbuhler and B. Jung, SPECFIT Global Analysis System, Version 3.0.37, Spectrum Software Associates, Marlborough, MA, USA, 2005. A. E. Martell and R. M. Smith, NIST – Critical stability constants, U.S. Department of Commerce, Gaithersburg, MD, USA, 1998. S. Sachs, V. Brendler and G. Geipel, Radiochim. Acta, 2007, 95, 103– 110. M. Rutsch, G. Geipel, V. Brendler, G. Bernhard and H. Nitsche, Radiochim. Acta, 1999, 86, 135–141. P. Gans, A. Sabatini and A. Vacca, HySS2006 Hyperquad Simulation and Speciation, version 3.2.24, Protonic Software, Leeds, UK, 2006. A. C. Texier, Y. Andres, M. Illemassene and P. Le Cloirec, Environ. Sci. Technol., 2000, 34, 610–615. A. Brachmann, G. Geipel, G. Bernhard and H. Nitsche, Radiochim. Acta, 2002, 90, 147–153. H. Moll, G. Geipel, T. Reich, G. Bernhard, T. Fanghanel and I. Grenthe, Radiochim. Acta, 2003, 91, 11–20. ¨ A. Gunther, G. Geipel and G. Bernhard, Polyhedron, 2007, 26, 59–65. ¨ A. Gunther, G. Geipel and G. Bernhard, Radiochim. Acta, 2006, 94, 845–851. A. Koban, G. Geipel, A. Rossberg and G. Bernhard, Radiochim. Acta, 2004, 92, 903–908; A. Koban and G. Bernhard, J. Inorg. Biochem., 2007, 101, 750–757; A. Koban and G. Bernhard, Polyhedron, 2004, 23, 1793– 1797; G. Geipel, Coord. Chem. Rev., 2006, 250, 844–854; I. Bonhoure, S. Meca, V. Marti, J. De Pablo and J. L. Cortina, Radiochim. Acta, 2007, 95, 165–172. K. J. Johnson, R. T. Cygan and J. B. Fein, Geochim. Cosmochim. Acta, 2006, 70, 5075–5088.

Dalton Trans., 2009, 5379–5385 | 5385