Oxochromium(V) Species Formed with 2,3-Dehydro-2-deoxy-N

Oxochromium(V) Species Formed with 2,3-Dehydro-2-deoxy-N-acetylneuraminic or N-Acetylneuraminic (Sialic) Acids: An In Vitro Model System of ...
0 downloads 0 Views 230KB Size
Chem. Res. Toxicol. 2003, 16, 881-892

881

Oxochromium(V) Species Formed with 2,3-Dehydro-2-deoxy-N-acetylneuraminic or N-Acetylneuraminic (Sialic) Acids: An In Vitro Model System of Oxochromium(V) Species Potentially Stabilized in the Respiratory Tract upon Inhalation of Carcinogenic Chromium(VI) Compounds Rachel Codd* and Peter A. Lay Centre for Heavy Metals Research, School of Chemistry, University of Sydney, NSW 2006, Australia Received April 15, 2003

The human respiratory tract is rich in sialoglycoconjugates. These polyfunctional, oxygenrich molecules stabilize potentially genotoxic oxoCr(V) species, which has relevance to the metabolic fate of carcinogenic Cr(VI) compounds when inhaled as dusts, mists, and fumes during occupational exposure. A series of oxoCr(V) species formed from solutions of Cr(VI) and the reductant, glutathione (GSH), in the presence of an excess of small molecule ligands that model sialoglycoconjugates, either 2,3-dehydro-2-deoxy-N-acetylneuraminic acid (I) or N-acetylneuraminic acid (II), have been characterized as a function of pH value, using electron paramagnetic resonance (EPR) spectroscopy (Cr(V) ) d1). Where [I] or [II] is in large excess of [Cr(V)-GSH], oxoCr(V)-I or -II species form (giso ) 1.97-1.98) via ligand-exchange reactions in which oxygen donor atoms from the carboxylic acid (I), tert-2-hydroxycarboxylato(2-) (II), and/or diolato(2-) groups (I, II) feature in the oxoCr(V) coordination sphere. At pH ∼ 7.1 (where [Cr(V)-GSH]:[II] , 1), the oxoCr(V)-II complexes remain EPR active for 24 h (with t1/2 ∼ 380 min). The parent EPR spectrum from oxoCr(V)-I or -II solutions represents the sum of spectra from individual linkage isomers; using EPR spectral simulation procedures, coupled with empirical EPR spectroscopic data for previously characterized oxoCr(V)-hydroxycarboxylato and -diolato species, a signature EPR spectrum has been predicted for each oxoCr(V)-I or -II linkage isomer, with defining giso and 1H aiso values and signal multiplicities. At pH values > 7.0, the major EPR signal (giso ∼ 1.980) is a septet, indicating the presence of oxoCr(V)-I or oxoCr(V)-II species with coordination from two O8,O9-diolato(2-) chelates, with six magnetically equivalent protons (1H aiso ∼ 0.7 × 10-4 cm-1). At pH values < 7.0, where carboxylato binding competes with diolato binding, the species (giso ∼ 1.979) most likely feature mixed binding modes from one diolato(2-) and one tert-2-hydroxycarboxylato(2-) chelate (oxoCr(V)-II) or one diolato(2-) chelate, one monodentate carboxylato(1-) group, and aqua and/or hydroxo oxygen donors (oxoCr(V)-I). In I and II, the relative concentration of the oxoCr(V) linkage isomer featuring donation via the O7,O8 diolato group is about 1 order of magnitude smaller than the concentration of the O8,O9 linkage isomer. The dominance of oxoCr(V)-I or -II linkage isomers with O8,O9 donation correlates with the dihedral angle in these isomers (O9-C9-C8-O8 ∼ 65°) being more favorably disposed toward metal binding, as compared to the O7,O8 analogue (O7-C7-C8-O8 ∼ 163°). These results have important implications with respect to the nature of the oxoCr(V) species that may potentially form in the sialoglycoconjugate-rich environment of the human respiratory tract upon inhalation of carcinogenic Cr(VI) compounds.

Introduction Exposure to chromate in the workplace is a serious occupational hazard, since Cr(VI) is a documented human carcinogen (1). Workers in Cr(VI)-dependent industries, such as leather tanning, pigment production, and stainless steel welding, suffer a higher incidence of respiratory cancers as compared to the normal population and can also suffer from perforated nasal septa from the inhalation of chromates (2, 3). Although there remains uncer* To whom correspondence should be addressed. E-mail: r.codd@ chem.usyd.edu.au.

tainty about the exact mechanisms by which Cr(VI) causes cancer, oxoCr(V) and/or oxoCr(IV) species have been strongly implicated as playing central roles in the carcinogenic cascade, as detailed in a recent review (4). For example, in the absence of reducing agents, Cr(VI) compounds are unreactive toward DNA in vitro, while oxoCr(V) and oxoCr(IV) species are able to cleave DNA and cause other types of DNA damage under a variety of conditions (5-7). Oxochromium(V) species have also been found to be genotoxic in bacterial and mammalian cell assays (8, 9). In vitro studies of the biological chemistry of oxoCr(V) complexes are useful in order to

10.1021/tx034077n CCC: $25.00 © 2003 American Chemical Society Published on Web 07/01/2003

882

Chem. Res. Toxicol., Vol. 16, No. 7, 2003

understand better the nature of the species that are likely to be stabilized in vivo. Recent studies of this kind have focused upon oxoCr(V) species formed in solution with polyfunctional, oxygen-rich ligands, such as carbohydrates (10-12) and quinic acid (1,3,4,5-tetrahydroxycyclohexanecarboxylic acid), the latter of which models both tert-2-hydroxycarboxylato(2-) and carbohydrate (diolato(2-)) binding motifs (13). These studies led to the understanding that oxoCr(V) species formed with polyfunctional ligands show coordination preferences that are pH-dependent (13). At acidic pH values (pH ∼ 5) representative of conditions that exist during phagocytosis of insoluble chromates, oxoCr(V) species are preferentially stabilized by tert-2-hydroxycarboxylato(2-) groups, with diolato(2-) coordination predominating under neutral and alkaline conditions, representative of the prevalent pH environment during the uptake of soluble chromates. A polyfunctional, oxygen-rich ligand that may also stabilize oxoCr(V) species is N-acetylneuraminic (sialic) acid, an acidic carbohydrate that occurs as the terminal residue of glycoconjugates and, of the family of sialic acids, is the most predominant form found in humans (14). The negatively charged, cell-bound sialoglycoconjugate moieties are crucial for maintaining the integrity of cells and play important roles in cell recognition and cell coagulation events (15). The extracellular location of sialoglycoconjugates suggests that these molecules are likely to play a role in stabilizing hard metal ions via the hard oxygen donor atoms derived from the alcohol and/or carboxylato groups. This prediction was borne out by the observation of strong electron paramagnetic resonance (EPR) spectroscopic signals from mixtures of Cr(VI) and human saliva, which were assigned as oxoCr(V) species stabilized by salivary-derived sialoglycoconjugates, where the terminating sialic acid residue was coordinated to the oxoCr(V) center (16). Because of the polyfunctional nature of N-acetylneuraminic acid, a more detailed understanding of the oxoCr(V)-sialoglycoconjugate speciation profile warrants studies of species formed with ligands that model sialoglycoconjugates over a range of pH values. A ligand that models the conformation of N-acetylneuraminic acid when linked to the penultimate carbohydrate residue via the O2 hydroxyl atom, as it occurs as the terminal residue in sialoglycoconjugates, is 2,3-dehydro-2-deoxy-N-acetylneuraminic acid (I, naenH5). The enzyme, neuraminidase, acts upon specific types of sialoglycoconjugates, releasing the free ligand, N-acetylneuraminic acid (II, naH6). Both I and II are important ligands in human biology (17, 18) with relevance to the influenza cascade (19-21) and several other health disorders, marked by aberrant concentrations of II (15, 22). The interaction between II and Ca(II) has been studied by 1H NMR spectroscopy (23, 24), and more recently, the coordination chemistry between several transition metal ions and II has been examined (16, 25). Here, we present results from our studies of the oxoCr(V) species formed from solutions of Cr(VI) and GSH in the presence of excess concentrations of I or II and consider the implications with respect to oxoCr(V) species that may be formed in sialoglycoconjugate-rich environments of the human respiratory tract, following the inhalation of Cr(VI) dusts, mists, and fumes. It is important that the nature of the oxoCr(V)-I or -II species and their distributions are more fully understood, to provide the context for the interpretation of results from biological studies conducted in the future.

Codd and Lay

Experimental Section Caution: Chromium(VI) is a documented human carcinogen (1), and Cr(V) complexes are mutagenic and potentially carcinogenic (5, 26). Therefore, Cr(VI) and Cr(V) compounds and solutions should be handled with due caution and appropriate safety measures should be taken, such as the use of masks and gloves. Reagents. Ligand II was obtained from Calbiochem (g98%), and Na2Cr2O7‚2H2O was obtained from Merck (GR grade). All remaining chemicals (I (95%), reduced GSH (96%), tetrahydro2-furoic acid (97%), 2-furoic acid (98%), and tetrahydropyran2-methanol (98%)) were purchased from Sigma-Aldrich and used without further purification. Sample Preparation for EPR Spectroscopy. 1. Ligand I (naenH5). An aliquot (68.6 µL) of an aqueous Na2Cr2O7‚2H2O stock solution (0.1 M) was added to an aqueous I stock solution (260.7 µL of 0.132 M), and an aliquot of a GSH stock solution (13.7 µL of 50 mM) was added. An EPR spectrum was collected of the final solution ([Cr(VI)]:[GSH]:[I] ) [40 mM]:[2 mM]:[100 mM], pH 1.94), and subsequent spectra were acquired of the parent solution (reclaimed from the EPR flat cell (Wilmad)) with adjustments made to the pH value of the solution using small aliquots (1-5 µL) of NaOH (5 M) or HCl (1 M). The pH values were studied (in order of the sequence in which the parent solution was reclaimed/pH adjusted) as follows: 1.94, 5.17, 7.26, 9.44, 6.09, 3.50, 1.95, 2.61, and 8.24. In a second series of experiments, spectra were acquired (within 5-10 min of mixing the components) from freshly prepared aqueous solutions (500 µL) where [Cr(VI)]:[GSH]:[I] ) [10 mM]:[5 mM]:[10 mM] (pH 7.01), [Cr(VI)]:[GSH]:[I] ) [10 mM]:[5 mM]:[5 mM] (pH 7.14), or [Cr(VI)]:[GSH]:[I] ) [10 mM]:[5 mM]:[10 mM] (pH 4.01). For all experiments, the pH values of the solutions were monitored during spectral acquisition using a HANNA micro pH meter (HI 9023) and probe (HI 1083B) and found not to deviate from the reported values by more than 0.05 of a pH unit. 2. Ligand II (naH6). Solutions (500 µL) were prepared from aqueous stock solutions of II (400 mM), Cr(VI) (0.1 M Na2Cr2O7‚ 2H2O), and GSH (50 mM), and the pH value of the solution was adjusted with aliquots (0-15 µL) of NaOH (5 M) prior to being made up to volume with water. Spectra were acquired from freshly prepared solutions within 5-10 min of mixing the components. Series of experiments varied either the pH value (pH values studied as follows: 2.08, 3.14, 4.37, 5.47, 6.21, 7.18, 8.38, and 9.76) while the [Cr(VI)]:[GSH]:[II] was kept constant ([40 mM]:[2 mM]:[100 mM]) or the [Cr(VI)]:[GSH]:[II] ratio while the pH value was kept constant as follows: series 1 [Cr(VI)]:[GSH]:[II] ) [20 mM]:[1 mM]:[50 mM], [20 mM]:[1 mM]: [100 mM], [20 mM]:[1 mM]:[250 mM], or [20 mM]:[1 mM]:[400 mM] (average pH value 1.9); series 2 [Cr(VI)]:[GSH]:[II] ) [20 mM]:[1 mM]:[50 mM], [20 mM]:[1 mM]:[100 mM], [20 mM]:[1 mM]:[250 mM], or [20 mM]:[1 mM]:[400 mM] (average pH value 7.2); series 3 [Cr(VI)]:[GSH]:[II] ) [40 mM]:[2 mM]:[25 mM], [40 mM]:[2 mM]:[62.5 mM], [40 mM]:[2 mM]:[125 mM], or [40 mM]:[2 mM]:[250 mM] (average pH value 9.7); series 4 [Cr(VI)]: [GSH]:[II] ) [4 mM]:[2 mM]:[1 mM], [4 mM]:[2 mM]:[2 mM], [4 mM]:[2 mM]:[5 mM], or [4 mM]:[2 mM]:[10 mM] (average pH value 6.4). EPR spectra were acquired from a solution with the reactant ratios [Cr(VI)]:[GSH]:[II] ) [40 mM]:[2 mM]:[100 mM] at the following time points (pH value): 12 (7.06), 25 (7.06), 40 (7.06), 60 (7.06), 120 (7.07), 180 (7.07), 240 (7.07), 300 (7.06), 380 (7.06), and 1440 (7.15) min. EPR Spectroscopy and Spectral Simulation. Continuous wave EPR spectra were recorded at room temperature on a Bruker (EMX) EPR spectrometer at X-band frequency (ca. 9.6 GHz), linked to a Bruker field controller (EMX 032T) and

Oxochromium(V) Species in the Respiratory Tract

Chem. Res. Toxicol., Vol. 16, No. 7, 2003 883

gaussmeter (EMX 035M). Spectra were acquired using the following conditions: modulation frequency, 100 kHz; modulation amplitude, 0.4 G; conversion time, 10.24 ms; time constant, 2.56 ms; and microwave power, 2 mW. EPR spectra were simulated using the program, WinSIM, developed at the National Institute of Environmental Health Sciences (27, 28). This program fits a simulated isotropic EPR spectrum for up to 10 species (with user-defined EPR parameters (giso (g shift), number/type of nuclei, aiso values) to the observed spectrum, using the simplex optimization algorithm (number of restarts, 4; maximum iterations per restart, 600; fractional tolerance, n < 0.01). The EPR parameters determined by simulation methods of a specific oxoCr(V)-X (X ) I or II) species comprising part of the equilibrium mixture of, for example, parent spectrum A, were not allowed to vary significantly (giso (0.0001; 1H aiso (0.05 × 10-4 cm-1) when the same species was deemed a component of an alternative parent spectrum (e.g., parent spectrum B), acquired under different conditions. Final simulation parameters (correlation value g 0.99) of oxoCr(V)-I or -II species are included in the Supporting Information (Tables S1 and S2).

Results Ligand I (naenH5). The EPR spectra from solutions of Cr(VI), GSH, and I as a function of pH (Figure 1) show sharp signals (line width ∼ 0.5 G) with giso values ranging from 1.973 to 1.980. These signals are typical of five- or six-coordinate oxoCr(V) species (d1) with hard ligands, such as oxygen donor atoms, in the coordination sphere (29). The oxoCr(V)-I species are formed in situ from the reduction of Cr(VI) by GSH, to form Cr(V)-GSH species, followed by ligand exchange reactions with I (which is present in excess, i.e., [Cr(V)-GSH]:[I] , 1). In the absence of GSH, the signals of oxoCr(V)-I species are very weak. Under conditions where I is in excess of [Cr(V)-GSH] (which itself is limited by [Cr(VI)]:[GSH]), the concentrations of oxoCr(V)-GSH species present are negligible since the well-defined signals due to oxoCr(V)-GSH species (giso ) 1.9858 and giso ) 1.996) (30, 31) are not observed. This regime reflects work that established that the reduction of Cr(VI) by GSH is under kinetic rather than thermodynamic control, based upon the similar redox potentials of GSH and biologically relevant carboxylic acids (32). With increasing pH values in the oxoCr(V)-I system (where [Cr(V)-GSH]:[I] , 1), the giso value of the multiplet shifts from giso ∼ 1.979 to giso ∼ 1.980 (Figure 1), which indicates that the donor strength of the ligands in the complexes is increasing (4). At pH values > 6.0, the dominant oxoCr(V)-I signal (giso ∼ 1.980) is a (somewhat asymmetric) septet with a superhyperfine coupling constant (∼0.7 × 10-4 cm-1) similar in magnitude to that of oxoCr(V) species formed with acyclic diolato ligands, such as glycerol (giso ) 1.9800; 1H aiso ) 0.58 × 10-4 cm-1) or 1,2-ethanediol (giso ) 1.9801; 1H aiso ) 0.59 × 10-4 cm-1) (33). The superhyperfine coupling arises from the perturbation of the dxy orbital of the oxoCr(V) center (d1) by the electron density from the protons in the second coordination sphere (10, 11, 29). The values of the superhyperfine coupling constant are too small to be attributable to 14N aiso superhyperfine coupling (14N aiso ∼ 2.3 × 10-4 cm-1) (34) and are typical of those observed for acyclic aliphatic 1,2-diolato ligands (11). The signal multiplicity and the value of the 1H aiso superhyperfine coupling constant of the spectra from oxoCr(V)-I solutions at pH values > 6.0 support the assignment of the major species as a bis-diolato oxoCr-

Figure 1. Room temperature X-band EPR spectra from aqueous solutions of Cr(VI) (40 mM), GSH (2 mM), and I (100 mM) as a function of pH value, presented as the (a) first (upper trace ) observed; lower trace ) simulation) and (b) second derivative of absorption.

(V)-I complex, with I coordinated in a bis-chelate fashion via the O8,O9-diolato(2-) group (Figure 2). This species, [CrO(O8,O9-naenH2)2]3- (Ik), would (like glycerol) give rise to a septet in the EPR spectrum, since the complex has six protons in the second shell of the oxoCr(V) coordination sphere, which are magnetically indistinguishable by continuous wave EPR spectroscopy. The bischelate nature of the species is borne out by the signal multiplicity, since a monochelate complex, such as [CrO(O8,O9-naenH2)(OH2)2]0 or [CrO(O8,O9-naenH2)(OH)2]2-, would yield a quartet, rather than the observed septet. A poorer fit between the observed and the simulated data was obtained assuming the presence of an overlapping quartet (i.e., [CrO(O8,O9-naenH2)(OH)2]2-; three magnetically equivalent protons) and a triplet (i.e., [CrO(O7,O8naenH2)(OH)2]2-; two magnetically equivalent protons), which mitigates against the presence of oxoCr(V)-I species containing one diolato(2-) chelate, where [Cr(V)GSH]:[I] , 1. Under conditions where the [Cr(V)-GSH]:

884

Chem. Res. Toxicol., Vol. 16, No. 7, 2003

Codd and Lay

Figure 2. Proposed oxoCr(V)-I linkage isomers present at equilibrium in aqueous solutions of Cr(VI) (40 mM) and GSH (2 mM) in the presence of excess I (100 mM) between pH values 1.95 and 9.44.

[I] ratios g 1, at pH values ∼ 7.1, the signal due to the oxoCr(V)-GSH species at giso ) 1.9858 is observed, with additional minor signals appearing at giso ) 1.977 and giso ) 1.972 (Supporting Information, Figure S1, traces a-c). The giso values of the latter two signals correlate with those described for minor oxoCr(V)-GSH species (31); however, it is not clear whether these signals might also be due to oxoCr(V)-I species featuring only a single I ligand, since the decrease in the relative intensity of these signals with increasing [I] is also consistent with the presence of a oxoCr(V)-mono-I to oxoCr(V)-bis-I equilibrium. These high field signals (where [Cr(V)GSH]:[I] g 1) could be simulated as oxoCr(V)-mono-I species with coordination via the O8,O9 (quartet) or O7,O8 (triplet) region of I, with the coordination sphere completed with I ligands coordinated in a monodentate fashion via the carboxylato group and/or aqua and/or hydroxo donors. However, the coincidence of the giso values of these species with the values of oxoCr(V)-GSH species (the relative concentrations of which would also be expected to decrease with increasing I) precludes the unequivocal assignment of these species (giso ) 1.977 and giso ) 1.972) at this time. Where the [Cr(V)-GSH]:[I] ∼ 1 (at pH 4.01), there is a broad signal (line width ∼ 3.5 G) centered at giso ) 1.973, which is most likely due to a family of oxoCr(V)-I complexes with a single I ligand, since the intensity of the broad signal decreases with increasing [I] at a similar pH value (Figure S1, trace d). As distinct from the analogous experiment (i.e., ([Cr(V)GSH]:[I] g 1) conducted under alkaline conditions (where the possible presence of Cr(V)-GSH species at high field complicates the clear designation of a oxoCr(V)-mono-I to oxoCr(V)-bis-I equilibrium), at acidic pH values, the Cr(V)-GSH complexes are not stable (31); therefore, the separate EPR signals in acid conditions result from an equilibrium between oxoCr(V)-mono-I and oxoCr(V)bis-I species.

The major analyses here focus upon systems where [Cr(V)-GSH]:[X] (X ) I or II) , 1, in which the dominant oxoCr(V)-X species most likely have two ligands per oxoCr(V) center; the bis-chelate nature of the oxoCr(V)-I species at giso ) 1.98 (where [Cr(V)-GSH]:[I] , 1) is unambiguous, based upon the observed spectral multiplicity and the similarity to spectra from analogous systems, oxoCr(V)-bis-diolato systems, such as glycerol. Extensive data of EPR spectroscopic parameters that exist for oxoCr(V) species with biologically relevant ligands, such as diols and hydroxy acids, have been analyzed in terms of donor types, strengths, and coordination number, resulting in an empirical method for correlating giso values (and, in a somewhat less developed fashion, Aiso and 1H aiso values) with likely species, according to eq 1 (4).

giso (species) ) 2.0023 - Σ∆giso (donor type) (1) The method correlates donor atom strength to the deviation of the giso value from the free electron value (ge ) 2.0023) (4). For example, a five-coordinate bis-chelate oxoCr(V)-diolato(2-) species (∆giso (RO-) ) 0.00505 and ∆giso (O2-) ) 0.00210) would be predicted to have a giso value ) 1.9800, according to the calculation 2.0023 ((4 × 0.00505) + 0.00210) (4). Therefore, in addition to considering the plausibility of species in terms of the chemistry of the system, the EPR spectroscopic parameters obtained from simulation procedures should be consistent with empirical predictions; this approach provides a more rigorous framework for the assignment of species in complex systems, such as the oxoCr(V)-I or -II systems examined here. The giso value of the septet signal obtained from spectral simulation (giso ) 1.9801) is within experimental error to the value predicted for a five-coordinate oxoCr(V) species with four alkoxide donor groups (giso ) 1.9800), based on these empirical data (4).

Oxochromium(V) Species in the Respiratory Tract

The decreased multiplicity of the signals at pH values e 3.5 indicates the likelihood that only one diolato(2-) chelate feature in the oxoCr(V) coordination sphere and that the carboxylato(1-) motif of I may compete for coordination sites, as a monodentate donor. At higher pH values, the diolato(2-) binding motif will likely displace the carboxylato(1-) donor, which is consistent with the increasing signal multiplicity (i.e., indicative of diolato(2-) binding) with increasing pH values. Also, the giso values of the signals at pH values e 3.5 are lower than the values of signals observed under alkaline conditions, which in the light of the empirical data (∆giso (RCO2-) ) 0.00593) (4), supports the notion of the coordination of the weaker carboxylato donor. The signals acquired from oxoCr(V)-I solutions, where [Cr(V)-GSH]:[I] , 1, at pH values e 3.5, are consistent with the following series of complexes: [CrO(O1-naenH4)(O7,O8-naenH2)(L)]n (L ) OH2 (n ) 1-) or OH- (n ) 2-); isomers Ie or Ig, respectively) and [CrO(O1-naenH4)(O8,O9-naenH2)(L)]n (L ) OH2 (n ) 1-) or OH- (n ) 2-); isomers If or Ih, respectively), with one bis-chelate diolato(2-) group, one monodentate carboxylato(1-) donor, and aqua or hydroxo (OH2 or OH-) groups in the coordination sphere. Each of the complexes above (Ie-h) would also be in equilibrium with the complex in which the noncoordinated carboxylic acid remains protonated (e.g., Ie partner complex is [CrO(O1-naenH4)(O7,O8-naenH3)(OH2)]0), since the pKa value of I is approximately 3.90 ( 0.7 (35). With two (Ie, Ig, and partner complexes) or three (If, Ih, and partner complexes) magnetically equivalent protons in the second coordination sphere, these linkage isomers would yield a triplet or a quartet signal in the EPR spectrum, respectively. The pH independence of EPR signals from oxoCr(V)glycerol species (results not shown) provides strong evidence that the increase in the giso value of the signals from the oxoCr(V)-I system (where [Cr(V)-GSH]:[I] , 1) with increasing pH value is not attributable to the coordination of nondeprotonated I diolato groups at low pH values. At the most acidic pH values studied (pH 1.95 and pH 2.61), where the [Cr(VI)]:[GSH]:[I] concentrations prescribe oxoCr(V)-bis-I species, there are additional broad signals in the EPR spectra at giso ∼ 1.973, which are absent in the spectra acquired at pH 3.50 or higher. On the basis of studies of oxoCr(V) complexes formed with hydroxycarboxylato(2-) and/or diolato(2-) ligands (4, 13), these signals are expected to be six-coordinate oxoCr(V)-bis-I species, with two aqua or one aqua and one hydroxo ligand in the coordination sphere [CrO(O1naenH4)(O7,O8-naenH3)(L)(OH2)]n (L ) OH2 (n ) 0) or OH- (n ) 1-); isomers Ia or Ic, respectively) and [CrO(O1-naenH3)(O8,O9-naenH2)(L)(OH2)]n (L ) OH2 (n ) 0) or OH- (n ) 1-); isomers Ib or Id, respectively), in which the noncoordinated carboxylic acid group of I remains protonated, which is consistent with the pH dependence observed in the spectra and with the predicted pKa value of I (pKa ) 3.90 ( 0.7) (35). Although coordination to a metal ion depresses the pKa values of protonated metal donor atoms, the noncoordinated carboxylic acid group is too distant from the oxoCr(V) center for any significant effect. An alternate coordination regime in the oxoCr(V)-I linkage isomers, in which the I ligand coordinates via the carboxylato group and the endocyclic oxygen (Oendo) atom (thereby displacing one aqua/hydroxo ligand) (i.e., [CrO(O1,Oendo-naenH4)(O8,O9-naenH3)]0 at pH < pKa (I)

Chem. Res. Toxicol., Vol. 16, No. 7, 2003 885

or [CrO(O1,Oendo-naenH4)(O8,O9-naenH2)]- at pH > pKa (I)) is not likely, since previous studies of oxoCr(V)carbohydrate complexes found no evidence to support endocyclic oxygen donation (10-12). Additional experiments of oxoCr(V) species formed with simple ligands with endocyclic oxygen donor atoms (described in a subsequent section) also provide strong evidence that concentrations of oxoCr(V)-(O1,Oendo-I) linkage isomers will be very small. At pH 5.17 in the oxoCr(V)-I system, the EPR signal appears as a “transition point” (more clearly seen from the second derivative of absorption, Figure 1, column b), representing a combination of the binding modes described for pH values > 6.0 and e 3.50. Although the EPR signals from oxoCr(V)-bis-I species at alkaline pH values (where [Cr(V)-GSH]:[I] , 1) appear as septets, the signals cannot be simulated as pure septets. The asymmetries of the signals point to the presence of overlapping signals from different oxoCr(V)bis-I species, due to diolato coordination via the O8,O9 or via the O7,O8 region of the ligand (i.e., linkage isomerism). Each of these species (with bis-chelate coordination modes) would have a different number of protons in the second coordination sphere, which would result in EPR signals with distinct multiplicities. The EPR signal from a solution of Cr(VI), GSH, and I ([40 mM]:[2 mM]:[100 mM]) at pH 9.44 was best simulated, using the simplest model, as comprising a mixture of three linkage isomers, [CrO(O7,O8-naenH2)2]3- (Ii), [CrO(O7,O8-naenH2)(O8,O9naenH2)]3- (Ij), and [CrO(O8,O9-naenH2)2]3- (Ik). Spectral simulation indicates that the isomer featuring donation from the O8,O9-diolato group of I (Ik) is present in significantly greater concentration, relative to the O7,O8diolato binding isomer (Ii), with the mixed diolato binding mode (O7,O8; O8,O9) isomer (Ij) present at concentrations intermediate to those of Ii and Ik. This trend is also found in the linkage isomers present in oxoCr(V)-I solutions at pH values < 6.0, where the relative concentration of the linkage isomer with O7,O8-I diolato binding (Ia,c,e,g) is about half that of the concentration of the related O8,O9-I isomer (Ib,d,f,h, respectively). The EPR spectra at all pH values have been simulated (Figure 1, offset (lower) spectrum in a) as comprising a mixture of the linkage isomers Ia-k (Figure 2) and the EPR parameters (Table 1) show good agreement between observed and calculated giso values (4). Ligand II (naH6). The EPR spectra from solutions of Cr(VI), GSH, and II as a function of pH (Figure 3), where [Cr(V)-GSH]:[II] , 1, show sharp signals (line width ∼ 0.5 G) with giso values ranging from 1.978 to 1.980. The signals obtained in the oxoCr(V)-II system are different from those obtained from the oxoCr(V)-I system, most particularly under acidic conditions, which indicates that the tert-2-hydroxycarboxylato(2-) “head” of II is involved in the coordination of oxoCr(V)-II species, most likely in a bidentate motif. This is consistent with data from oxoCr(V) species formed with the related polyfunctional ligand, quinic acid (13), which has been isolated in the solid state and characterized by X-ray absorption spectroscopy (XAS) as a [CrO(tert-2-hydroxycarboxylato(2-))2]- complex (36). As for the oxoCr(V)-I system, the spectra from solutions of Cr(VI), GSH, and II were obtained under conditions that favored the formation of bis-chelate species (i.e., [Cr(V)-GSH]:[II] , 1); in these cases, signals due to oxoCr(V)-GSH species (30, 31) were not observed. However, where the [Cr(V)-GSH]:[II] > 1) (e.g., [Cr(VI)]:[GSH]:[II] ) [4 mM]:[2 mM]:[1 mM]), the

886

Chem. Res. Toxicol., Vol. 16, No. 7, 2003

Codd and Lay

Table 1. EPR Parameters of Species Formed between Cr(V) and I Determined from Spectral Simulation Ia Ib Ic Id Ie If Ig Ih Ii Ij Ik a

species

giso(expt)

giso(calcd)a

7 8 0 4)(O ,O -naenH3)(OH2)2] [CrO(O1-naenH4)(O8,O9-naenH3)(OH2)2]0 [CrO(O1-naenH4)(O7,O8-naenH3)(OH2)(OH)][CrO(O1-naenH4)(O8,O9-naenH2)(OH2)(OH)][CrO(O1-naenH4)(O7,O8-naenH2)(OH2)][CrO(O1-naenH4)(O8,O9-naenH2)(OH2)][CrO(O1-naenH4)(O7,O8-naenH2)(OH)]2[CrO(O1-naenH4)(O8,O9-naenH2)(OH)]2[CrO(O7,O8-naenH2)2]3[CrO(O7,O8-naenH2)(O8,O9-naenH2)]3[CrO(O8,O9-naenH2)2]3-

1.9723 1.9730 1.9735 1.9740 1.9788 1.9787 1.9794 1.9793 1.9799 1.9801 1.9801

1.9721 1.9721 1.9731 1.9731 1.9781 1.9781 1.9791 1.9791 1.9800 1.9800 1.9800

[CrO(O1-naenH

giso ) 2.0023 - ∆Σgiso (4).

b 1H

1H

aisob (σ)

0.70 (0.01) 0.65 (0) 0.69 (0.03) 0.67 (0.01) 0.70 (0.02) 0.65 (0.03) 0.69 (0.03) 0.72 (0.02) 0.66 (0.02) 0.67 (0.02) 0.66 (0.02)

1H

(eqv)

2 3 2 3 2 3 2 3 4 5 6

aiso units ) (× 10-4 cm-1); σ ) standard deviation.

Figure 3. Room temperature X-band EPR spectra from aqueous solutions of Cr(VI) (40 mM), GSH (2 mM), and II (100 mM) as a function of pH value, presented as the (a) first (upper trace ) observed; lower trace ) simulation) and (b) second derivative of absorption.

EPR signal due to Cr(V)-GSH species (giso ) 1.9858) is observed, which decreases in intensity as the [II]:[Cr(V)] increases (Figure S2). The signal-to-noise ratio of spectra acquired where [Cr(V)-GSH]:[II] g 1 (Figure S2) is quite poor; spectra with superior resolution were obtained where [Cr(V)-GSH]:[X] (X ) I or II) , 1, as shown in Figures 1 and 3. At low pH values ( 8.3, where [Cr(V)-GSH]:[II] , 1, the dominant signal is a septet, similar to the signal from oxoCr(V)-bis-I solutions at high pH values, which indicates that the dominant species are likely to be bis-diolato oxoCr(V) complexes featuring O8,O9 donation (with six magnetically equivalent protons in the second coordination sphere). Earlier (37) and more recent (12, 13) papers have shown the presence of geometrical isomers of oxoCr(V)-hydroxycarboxylic acid species, distinguished by the coordination of alkoxide and/or carboxylato groups in the apical or equatorial region of a trigonal bipyramid. Similarly, the two oxoCr(V)-II species with giso ) 1.9785 and giso ) 1.9792 are assigned as geometrical isomers of bis-chelate complex, [CrO(O1,O2-naH4)2]-, with the ligand coordinated via the tert-2-hydroxycarboxylato(2-) head of II. Bis-chelate tert-2-hydroxycarboxylato(2-) oxoCr(V) complexes are relatively stable and have been structurally characterized by X-ray crystallography (38, 39) and XAS techniques (29, 36). The giso values of the geometrical isomers of [CrO(qaH3)2]- (where qaH3(2-) is coordinated via the tert-2-hydroxycarboxylato(2-) region), which appear upon dissolution of solid K[[CrO(qaH3)2]‚H2O in water (giso ) 1.9787; giso ) 1.9791) (13), are similar to those obtained for IIa,b. Because the solutions described here comprise a mixture of linkage isomers of oxoCr(V)-I or -II species, attempts to isolate pure solid products have been largely unsuccessful; a crude preparation of an oxoCr(V)-II species yielded a solid that contained approximately 4% Cr(V), as determined using spin quantitation. The EPR spectra from a solution comprising oxoCr(V)-II species, however, show a remarkable consistency in the speciation profile, as a function of time (Figure 4) at the pH value studied (pH ∼ 7.1). After 24 h, the EPR signal from a solution of [Cr(VI)]:[GSH]:[II] ) [40 mM]:[2 mM]:[100 mM] maps onto the signal acquired from the same solution at t ) 12 min (Figure 4; panel a inset); the invariance of the pattern of the EPR signals with time shows that the system is at equilibrium when the initial spectrum is acquired at t ) 5-10 min. The height of the integral of the signal where t ) 24 h is 36% of the height of the integral where the signal reaches a maximum (t ) 40 min). With increasing pH value, the EPR signals of the oxoCr(V)-II system, where [Cr(V)-GSH]:[II] , 1, in-

Oxochromium(V) Species in the Respiratory Tract

Figure 4. Room temperature X-band EPR spectra from an aqueous solution (pH ∼7.1) of Cr(VI) (40 mM), GSH (2 mM), and II (100 mM) at times 12, 25, 40, 60, 120, 180, 240, 300, 380, and 1440 min after mixing the initial solution, presented as (a) the first derivative of absorption (inset ) spectra at t ) 12 min, 40 min, and 24 h) and (b) absorption (inset ) integral height (at giso ) 1.9795) vs time (min)).

Chem. Res. Toxicol., Vol. 16, No. 7, 2003 887

crease in giso values (from giso 1.979 to 1.980), which indicates the formation of complexes with stronger donor atoms, such as alkoxide groups, relative to the weaker donation from the carboxylato group. Also, the signals exhibit superhyperfine coupling with 1H aiso values ∼ 0.7 × 10-4 cm-1, similar in magnitude to the values observed in the oxoCr(V)-I system. To investigate the possibility that the oxoCr(V) complexes might feature binding via the endocyclic oxygen atoms of II (or I), EPR spectra were acquired from solutions of Cr(VI) and GSH in the presence of excess concentrations of ligands containing endocyclic oxygen atoms (data not shown). Although oxoCr(V) signals were observed (giso ) 1.9785) from solutions (pH 3.5) of tetrahydro-2-furoic acid, Cr(VI), and GSH, where [Cr(VI)]: [GSH]:[endocyclic oxygen-containing ligand] ratios ) [40 mM]:[2 mM]:[1 M], the signals were weak, relative to signals obtained from oxoCr(V)-II solutions (where [Cr(VI)]:[GSH]:[II] ) [40 mM]:[2 mM]:[100 mM]). No EPR signals were obtained from solutions of Cr(VI) (40 mM), GSH (2 mM), and an excess (1 M) of either 2-furoic acid or tetrahydropyran-2-methanol. Also, an equilibrium mixture of oxoCr(V) species with endocyclic ([CrO(Oendo,O1-naH5)2]+) or tert-2-hydroxycarboxylato(2-) binding ([CrO(O1,O2-naH4)2]-) would exhibit a pH dependence, which is not observed at pH values 2.08 or 3.14. In addition, previous work on species formed between oxoCr(V) and D-glucose (10) and other glycosides (12) did not yield evidence of endocyclic oxygen donation. The oxoCr(V)-II spectra were simulated as comprising equilibrium mixtures of linkage isomers (IIa-g, Figure 5), with II coordinated to the oxoCr(V) center via the tert2-hydroxycarboxylato(2-) head (O1,O2-II) and/or the diolato(2-) regions (O7,O8-II or O8,O9-II) of the glycerol tail

Figure 5. Proposed oxoCr(V)-II linkage isomers present at equilibrium in aqueous solutions of Cr(VI) (40 mM) and GSH (2 mM) in the presence of excess II (100 mM) between pH values 2.08 and 9.76.

888

Chem. Res. Toxicol., Vol. 16, No. 7, 2003

Codd and Lay

Table 2. EPR Parameters of Species Formed between Cr(V) and II Determined from Spectral Simulation species

a

giso(expt)

giso(calcd)a

IIa IIb IIc IId IIe IIf IIg IIh

[CrO(O1,O2-naH

4)2] [CrO(O1,O2-naH4)2][CrO(O1,O2-naH4)(O7,O8-naH3)]2[CrO(O1,O2-naH4)(O8,O9-naH3)]2[CrO(O7,O8-naH3)2]3[CrO(O7,O8-naH3)(O8,O9-naH3)]3[CrO(O8,O9-naH3)2]3[CrO(O2,O4,O7,O8-R-naH)]2-

1.9785 1.9792 1.9790 1.9792 1.9800 1.9800 1.9803 1.9796

1.9782 1.9782 1.9791 1.9791 1.9800 1.9800 1.9800 1.9800

IIi

[CrO(O2,O4,O8,O9-R-naH)]2-

1.9804

1.9800

giso ) 2.0023 - Σ∆giso (4).

b 1H

1H

aisob (σ)

N/Ac N/Ac 0.68 (0.02) 0.70 (0.02) 0.72 (0.02) 0.65 (0.01) 0.65 (0.05) 0.77 (0.01) 0.48 (0.01) 0.98 (0.01) 0.55 (0.02)

1H

(eqv)

N/Ac N/Ac 2 3 4 5 6 2 1 3 1

aiso units ) (× 10-4 cm-1); σ ) standard deviation. c N/A, not applicable.

(Figures 3 and 5, Table 2). The carboxylato group of II will most likely be deprotonated at all pH values studied, since pKa (II) ∼ 2.60 (40). The distribution of bis-diolato oxoCr(V)-II linkage isomers at pH values > 8.3, [CrO(O7,O8-naH3)2]3- (IIe, minor), [CrO(O7,O8-naH3)(O8,O9-naH3)]3-] (IIf, intermediate), and [CrO(O8,O9-naH3)2]3- (IIg, major), is similar to the distribution within the analogous oxoCr(V)-I family. It was necessary to include two additional species (giso ) 1.9796 and giso ) 1.9804) in simulations of oxoCr(V)-II spectra at pH values > 8.3. The inclusion of these minor species (totaling approximately 5-6% of the concentration) was not necessary in the oxoCr(V)-I system, which suggests that the minor species are related to the chemistry occurring at the tert-2-hydroxycarboxylato(2-) head of II, which is not present in I. It is postulated that the minor species found at alkaline pH values involve oxoCr(V) coordination from the R anomer of II. This is supported by the observation that showed that the protons on C3 undergo D2O exchange at a rate that increases with pD value in alkaline solution (41). In the R anomer of II, the O4 group becomes a viable donor atom and the species IIh ([CrO(O2,O4,O7,O8-R-naH)]2-) and IIi ([CrO(O2,O4,O8,O9-R-naH)]2-) may be formed. The rotationally constrained protons in the proximity of the oxoCr(V) center give rise to 1H aiso superhyperfine coupling constants greater in magnitude than those where the chelate ring is flexible (11, 13, 42), as is observed for these signals. The possibility of alkaline decomposition products from II (e.g., 2-carboxypyrrole) being formed, with the potential for stabilizing the oxoCr(V) center, is discounted, based upon the harsh conditions required (∼0.6 M OH-, 100 °C) to form these products (43).

Discussion Inhalation of carcinogenic Cr(VI) dusts, mists, and fumes is one of the main routes of Cr(VI) uptake in the workplace (2). Following inhalation, one of the first physiological environments encountered by the Cr(VI) species is the mucin-rich respiratory tract, which has a high local concentration of sialoglycoconjugates (44), micro pH environments (45), and varying types and concentrations of reductants (46). In the absence of exogenous reducing agents, stable oxoCr(V) species with sialoglycoconjugates are formed in vitro from mixtures of Cr(VI) and human saliva (16); hence, it may be possible that related oxoCr(V) species are formed in vivo in the respiratory tract following Cr(VI) inhalation. Therefore, in vitro studies of oxoCr(V) species formed with I and II that model sialoglycoconjugates has led to a better

understanding of the oxoCr(V) intermediates produced on inhalation of carcinogenic Cr(VI) compounds. Before studies of the genotoxicities of oxoCr(V)-I, -II, or -sialoglycoconjugate species are conducted, it is necessary to characterize their solution chemistries. EPR spectroscopic studies show that the oxoCr(V) center is stabilized by both ligands I and II. On the basis of the documented anomeric mix of II (R-II:β-II ) 5-8%:9592%) (23, 47), the major oxoCr(V)-II species are attributed to β-II species. The strong signals in the EPR spectra are characteristic of oxoCr(V) complexes with oxygen ligands coordinating via the tert-2-hydroxycarboxylato(2-) head (II) group, monodentate carboxylato(1-) donor groups (I), and/or dioloto(2-) groups (I or II). No reasonable fit between the observed and the simulated spectra was obtained assuming coordination from the amide nitrogen donor atom; the values of the superhyperfine coupling are representative of 1H aiso values. Also, on the basis of the present studies using model ligands containing endocyclic oxygen atoms and previous studies, there is no evidence for the coordination of endocyclic oxygen atoms to the oxoCr(V) center (10, 12). In addition, the poor fit between calculated giso values (assigning a value of ∆giso (0.00604 ) ∆giso of OH2) to an ether linkage, since this group is a weak donor, supports predicted giso values and the unlikely binding from the endocyclic oxygen atom in oxoCr(V)-I or -II species. On the basis of the independence of the EPR signals of the oxoCr-II system on increasing [II]:[Cr(V)-GSH] ratios at pH values ) 1.9, 7.2, or 9.7 (Figure S3) and the spectral multiplicity (septet) arising from oxoCr(V)-I or -II species (at high pH values) that must contain six magnetically equivalent protons (bis-O8,O9-I or -II) in the coordination sphere, the major oxoCr(V)-I or -II species formed, where [Cr(V)-GSH]:[X] (X ) I or II) , 1, are predicted to feature two ligands per oxoCr(V) center. The bis-diolato linkage isomers, with donation from one O7,O8 group and one O8,O9 group or two O7,O8 chelates, would be expected to yield a sextet or quintet, respectively, since five or four (respectively) magnetically equivalent protons exist in the coordination sphere. Using spectral deconvolution, a signature EPR spectrum has been assigned to each oxoCr(V)-I or -II linkage isomer, with an example of the sum of individual spectra comprising the oxoCr(V)-II equilibrium mixture at pH 7.18, given in Figure 6. The analysis of all spectra has enabled the relative concentration of each isomer be determined, as a function of pH value (Figure 7), together with the equilibrium constants of intramolecular ligand exchange (Kex) (Table 3), according to eqs 2-5. The reliability of the spectral deconvolution analysis is evidenced by the

Oxochromium(V) Species in the Respiratory Tract

Figure 6. Signature EPR spectra predicted for oxoCr(V)-II linkage isomers present at equilibrium at pH ) 7.18, where [Cr(VI)]:[GSH]:[II] ) [40 mM]:[2 mM]:[100 mM].

Figure 7. Relative concentrations determined from EPR spectral simulations of (a) oxoCr(V)-I [Ia, open circle; Ib, solid circle; Ic, open square; Id, solid square (Ia-d present only at pH values < 3.6); Ie, open up-triangle; If, solid up-triangle; Ig, open down-triangle; Ih, solid down-triangle; Ii, up-right cross; Ij, side-ways cross; and Ik, asterisk] or (b) oxoCr(V)-II species [IIa, open circle; IIb, solid circle; IIc, open square; IId, solid square; IIe, up-right cross; IIf, side-ways cross; IIg, and asterisk] present at equilibrium, as a function of the pH value.

reproducibility of the spectral parameters for each species under different experimental conditions and the pH dependencies of the relative intensities of the signals, which are in accord with eqs 2-5.

Chem. Res. Toxicol., Vol. 16, No. 7, 2003 889

If h Ix + H2O + 2H+ (x ) i, j, or k)

(2)

Ih h Ix + H2O + H+ (x ) i, j, or k)

(3)

IIb h IIx + H+ (x ) c or d)

(4)

IId h IIx + H+ (x ) e, f, or g)

(5)

The relative concentrations of the oxoCr(V)-II linkage isomers as a function of the pH value show that at low pH values (pH < 3.1), the tert-2-hydroxycarboxylato(2-) binding motif is favored, while at high pH values (pH > 7.0), binding via the diolato(2-) region of the II glycerol tail predominates. This is consistent with previous work on species formed between oxoCr(V) and the polyfunctional ligand, quinic acid (13), and has implications with respect to the types of oxoCr(V) species that are likely to form under micro pH environments present in the cytoplasm (pH ∼ 7.4) or during phagocytosis (pH ∼ 5). The pH dependence of the coordination modes of oxoCr(V)-I or -II species observed here has implications with respect to the inhalation of Cr(VI) compounds, since the pH value of the nasal passage varies from pH 5.2 to pH 8.1 (45). Also, GSH is an antioxidant present in the human respiratory tract (together with uric acid (major) and vitamin E (minor)) (46), providing conditions with the potential to reduce Cr(VI) prior to the stabilization of oxoCr(V) species with mucin-derived I-type and/or II chelates. The solution of [Cr(V)-GSH]:[II] ) [40 mM]:[2 mM]:[100 mM] at pH ∼ 7.1 comprising oxoCr(V)-bis-II species, resulted in a strong oxoCr(V) EPR signal, and even after 24 h, its intensity was 36% of the maximum value (t ) 40 min). Hence, oxoCr(V)-II and related sialoglycoconjugate species may persist in vivo over extended time periods and are likely to contribute to Cr(VI) genotoxicities of relevance to an important occupational health and safety problem. The relative concentrations of the oxoCr(V)-I or -II species where the ligand coordinates via the O8,O9 region are significantly greater (about 10-fold) than the corresponding isomer where coordination occurs via the O7,O8 region of the ligand. This is rationalized in the light of structural data, which shows that the O9-C9-C8-O8 dihedral angle is in a more favorable conformation with respect to metal chelation (∼60°), relative to the O8-C8C7-O7 dihedral angle (∼160°) (48). Crystallographic and NMR spectroscopic data have established the existence of a network of intramolecular hydrogen bonds in I (49), II (48, 50, 51), and derivatives (52), which affect the conformations of I and II. The actual network of hydrogen bonds observed depends on the state of the rotation about the C6-C7 and C5-N bonds; what is consistent within all of these studies, however, is the prescribed trans and gauche torsional angles for O7-C7-C8-O8 and O9-C9C8-O8, respectively. The implications with respect to the oxoCr(V)-I or -II systems are that there exists a higher energy cost for the formation of the O7,O8 linkage isomers, relative to the O8,O9 linkage isomers, since the formation of the former isomer necessitates a rotation about the C7-C8 bond, which is not required for the formation of the latter linkage isomer. This effect is of particular relevance to the ability of terminal sialic acid residues to stabilize metal ions found in vivo, since the glycerol tail of II is positioned into extracellular space. The oxoCr(V)-sialogylcoconjugate species formed from mixtures of Cr(VI) and human saliva at pH values ∼ 6-7

890

Chem. Res. Toxicol., Vol. 16, No. 7, 2003

Codd and Lay

Table 3. Equilibrium Constants of Intramolecular Ligand Exchange (Kex) for Linkage Isomers of OxoCr(V)-I and OxoCr(V)-II Species at Room Temperature species A If If If Ih Ih Ih IIb IIb IId IId IId

[CrO(O1-naenH4)(O8,O9-naenH2)(OH2)][CrO(O1-naenH4)(O8,O9-naenH2)(OH2)][CrO(O1-naenH4)(O8,O9-naenH2)(OH2)][CrO(O1-naenH4)(O8,O9-naenH2)(OH)]2[CrO(O1-naenH4)(O8,O9-naenH2)(OH)]2[CrO(O1-naenH4)(O8,O9-naenH2)(OH)]2[CrO(O1,O2-naH4)2][CrO(O1,O2-naH4)2][CrO(O1,O2-naH4)(O8,O9-naH3)]2[CrO(O1,O2-naH4)(O8,O9-naH3)]2[CrO(O1,O2-naH4)(O8,O9-naH3)]2-

Ii Ij Ik Ii Ij Ik IIc IId IIe IIf IIg

species B

Kex

[CrO(O7,O8-naenH2)2]3- a [CrO(O7,O8-naenH2)(O8,O9-naenH2)]3- a [CrO(O8,O9-naenH2)2]3- a [CrO(O7,O8-naenH2)2]3- b [CrO(O7,O8-naenH2)(O8,O9-naenH2)]3- b [CrO(O8,O9-naenH2)2]3- b [CrO(O1,O2-naH4)(O7,O8-naH3)]2- c [CrO(O1,O2-naH4)(O8,O9-naH3)]2- c [CrO(O7,O8-naH3)2]3- d [CrO(O7,O8-naH3)(O8,O9-naH3)]3- d [CrO(O8,O9-naH3)2]3- d

7.7 × 10-13 M2 3.4 × 10-12 M2 9.5 × 10-12 M2 3.6 × 10-7 M 1.6 × 10-6 M 4.5 × 10-6 M 8.7 × 10-6 M 6.8 × 10-5 M 3.4 × 10-9 M 8.8 × 10-9 M 2.6 × 10-8 M

a Calculated from a plot of 1/[H+]2 vs [Ii]/[If] (or [Ij]/[If] or [Ik]/[If]), with the value of the slope ) K (M2). b Calculated from a plot of ex 1/[H+] vs [Ii]/[Ih] (or [Ij]/[Ih] or [Ik]/[Ih]), with the value of the slope ) Kex (M). c Calculated from a plot of 1/[H+] vs [IIc]/[IIb] (or [IId]/[IIb]), with the value of the slope ) Kex (M). d Calculated from a plot of 1/[H+] vs [IIe]/[IId] (or [IIf]/[IId] or [IIg]/[IId]), with the value of the slope ) Kex (M). The values of the intercepts in all plots were close to the origin, as expected for these simple equilibria (refer to eqs 2-5). Water was omitted from Kex calculations.

were most likely bound as bis-diolato chelates (16), which is consistent with the studies of the free ligand system (I is a good model for a terminal-bound II residue (via the 2′-OH group), as it appears in sialoglycoconjugates) at similar pH values. At pH values < 6.0, I is able to coordinate to the oxoCr(V) via the carboxylato group (in a monodentate fashion), with the coordination sphere completed by diolato groups and aqua/hydroxo donors (Ia-h), giving rise to five- and six-coordinate oxoCr(V)-I species. There were no six-coordinate species (those appearing at relatively high field) observed for the oxoCr(V)-II system, where tert-2-hydroxycarboxylato(2-) (and diolato(2-)) binding (both modes being bidentate) has been proposed. Therefore, the steric bulk posed about the oxoCr(V) center from two bidentate II ligands would be greater than that posed from one bidentate and one monodentate ligand (as in I), thereby mitigating against the formation of six-coordinate oxoCr(V)-II species. This may also be due to electronic effects with weaker donors (such as a monodentate carboxylato donor) more likely to yield six-coordinate species, as observed in the oxoCr(V)-I system. The oxoCr(V)-I spectra at pH values > 6.0 are very similar to spectra of oxoCr(V)-glycerol species (37), which indicates that the oxoCr(V)-I (-II) or coordination unit of the diolato linkage isomers has a flexibility comparable to the sterically unhindered oxoCr(V)glycerol species. This flexibility of the oxoCr(V)-I (or -II) species renders the protons magnetically indistinguishable on the EPR time scale. Only where the protons in the second coordination sphere are rotationally restricted (as in a cyclic diol) is magnetic inequivalence observed in continuous wave EPR spectroscopy (11-13, 42). The magnetic inequivalence in oxoCr(V)-I or -II species may be detectable using electron nuclear double resonance spectroscopy, which has been used to detect distinct magnetic environments for other oxoCr(V)-hydroxycarboxylic acid species (53). In EPR spectral simulations, excellent fits were obtained with the majority of experimental data in the oxoCr(V)-I or -II systems, with the inclusion of one geometrical isomer per linkage isomer. Only in the case of [CrO(O1,O2-naH4)2]- was the presence of two geometrical isomers of one linkage isomer (IIa,b) observed in the EPR spectrum, with two sets of resolvable 53Cr Aiso satellites. Therefore, the distinction between the electronic structures of IIa,b is greater than the distinction between the electronic structures of oxoCr(V)-II

isomers involving diolato donation from the glycerol tail of II, as detectable by the limits of the EPR spectroscopic techniques used here. In oxoCr(V) species involving chelation from ring-bound (cyclic) diolato groups (e.g., quinic acid), good fits between experimental and simulated EPR spectroscopic data assumed the presence of two geometrical isomers per linkage isomer (13). Therefore, the electronic structure of oxoCr(V) species appears to be influenced by the donor type (diolato(2-), tert-2hydroxycarboxylato(2-)) and the cyclic (or acyclic) nature of the ligand. The nature of the biological damage potentially caused by oxoCr(V)-I or -II species that may form in the respiratory tract following Cr(VI) inhalation is uncertain. However, it has been shown that Cr(VI) is both genotoxic and cytotoxic in human gastric mucosa cells (54), and in another study, that a Cr(V) species is able to cleave the sialylated R1-acid glycoprotein (55). More recently, it has been shown that a Cr(III) complex modulated the aggregation behavior of pig gastric mucin (56). Because sialylated biomolecules are ubiquitous in the human body, it is important that the full implications of the formation of oxoCr(V)-I and -II species are further explored in a biological context.

Concluding Remarks The ligand, naH6, and the related ligand, naenH5, form stable complexes with the oxoCr(V) center. These results have particular relevance to the response of the human respiratory system toward inhalation of Cr(VI) compounds, since the respiratory tract is rich in mucins (sialoglycoconjugates) and antioxidants, which may provide an environment suited toward the stabilization of potentially genotoxic oxoCr(V) species. The illustration that I or II is able to stabilize the potentially genotoxic oxoCr(V) ion suggests that these chelates may play other important roles in the stabilization of alternative exogenous and endogenous transition metal ions in biological systems.

Acknowledgment. R.C. gratefully acknowledges the Gritton Postdoctoral Fellowship Award, and P.A.L. acknowledges an ARC Large Grant and RIEFP Grant for the EPR spectrometers. Dr. R. Thomson is gratefully acknowledged for useful discussions about the chemistries of I and II. Supporting Information Available: EPR spectra from solutions where [Cr(V)-GSH]:[X] (X ) I or II) g 1 (Figures S1

Oxochromium(V) Species in the Respiratory Tract (X ) I) and S2 (X ) II)). Tables of the EPR parameters from spectral simulations for oxoCr(V)-I and oxoCr(V)-II systems (Tables S1 and S2). EPR spectra from solutions where [Cr(V)GSH]:[II] (under conditions where [Cr(V)-GSH]:[II] , 1) at constant pH values (Figure S3). This material is available free of charge via the Internet at http://pubs.acs.org.

References (1) IARC (1990) Chromium, nickel and welding. In IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 49, pp 127-198, International Agency for Research on Cancer, Lyon, France. (2) Pellerin, C., and Booker, S. M. (2000) Reflections on hexavalent chromium. Environ. Health Perspect. 108, A402-A407. (3) Gibb, H. J., Lees, P. S. J., Pinsky, P. F., and Rooney, B. C. (2000) Lung cancer among workers in chromium chemical production. Am. J. Ind. Med., 115-126. (4) Levina, A., Codd, R., Dillon, C. T., and Lay, P. A. (2003) Chromium in biology: toxicology and nutritional aspects. Prog. Inorg. Chem. 51, 145-250. (5) Farrell, R. P., Judd, R. J., Lay, P. A., Dixon, N. E., Baker, R. S. U., and Bonin, A. M. (1989) Chromium(V)-induced cleavage of DNA: are chromium(V) complexes the active carcinogens in chromium(VI)-induced cancers? Chem. Res. Toxicol. 2, 227-229. (6) Levina, A., Barr-David, R., Codd, R., Lay, P. A., Dixon, N. E., Hammershøi, A., and Hendry, P. (1999) In vitro plasmid DNA cleavage by chromium(V) and -(IV) 2-hydroxycarboxylato complexes. Chem. Res. Toxicol. 12, 371-381. (7) Levina, A., Lay, P. A., and Dixon, N. E. (2000) Disproportionation and nuclease activity of bis[2-ethyl-2-hydroxybutanoato(2-)]oxochromate(V) in neutral aqueous solutions. Inorg. Chem. 39, 385-395. (8) Dillon, C. T., Lay, P. A., Bonin, A. M., Cholewa, M., and Legge, G. J. F. (2000) Permeability, cytotoxicity and genotoxicity of Cr(III) complexes and some Cr(V) analogues in V79 Chinese hamster lung cells. Chem. Res. Toxicol. 13, 742-748. (9) Dillon, C. T., Lay, P. A., Bonin, A. M., Dixon, N. E., and Sulfab, Y. (2000) DNA interactions and bacterial mutagenicity of some chromium(III) imine complexes and their chromium(V) analogues. Evidence for chromium(V) intermediates in the genotoxicity of chromium(III). Aust. J. Chem. 53, 411-424. (10) Barr-David, G., Charara, M., Codd, R., Farrell, R. P., Irwin, J. A., Lay, P. A., Bramley, R., Brumby, S., Ji, J.-Y., and Hanson, G. R. (1995) EPR characterisation of the Cr(V) intermediates in the Cr(VI)/Cr(V) oxidations of organic substrates and of relevance to Cr-induced cancers. J. Chem. Soc., Faraday Trans. 91, 12071216. (11) Codd, R., Irwin, J. A., and Lay, P. A. (2003) Sialoglycoprotein and carbohydrate complexes in chromium toxicity. Curr. Opin. Chem. Biol. 7, 213-219. (12) Rizzotto, M., Levina, A., Santoro, M., Garcı´a, S., Frascaroli, M. I., Signorella, S., Sala, L. F., and Lay, P. A. (2002) Redox and ligand-exchange chemistry of chromium(VI/V)-methyl glycoside systems. J. Chem. Soc., Dalton Trans., 3206-3213. (13) Codd, R., and Lay, P. A. (1999) Competition between 1,2-diol and 2-hydroxy acid coordination in Cr(V)-quinic acid complexes: implications for stabilization of Cr(V) intermediates of relevance to Cr(VI)-induced carcinogenesis. J. Am. Chem. Soc. 121, 78647876. (14) Angata, T., and Varki, A. (2002) Chemical diversity in the sialic acids and related R-keto acids: an evolutionary perspective. Chem. Rev. 102, 439-469. (15) Traving, C., and Schauer, R. (1998) Structure, function and metabolism of sialic acids. Cell. Mol. Life Sci. 54, 1330-1349. (16) Codd, R., and Lay, P. A. (2001) Chromium(V)-sialic (neuraminic) acid species are formed from mixtures of chromium(VI) and saliva. J. Am. Chem. Soc. 123, 11799-11800. (17) Corfield, A. P., and Schauer, R. (1982) Occurrence of Sialic Acids. In Sialic Acids. Chemistry, Metabolism and Function (Schauer, R., Ed.) Vol. 10, pp 5-50, Springer-Verlag, Vienna. (18) Schauer, R., Kelm, S., Reuter, G., Roggentin, P., and Shaw, L. (1995) Biochemistry and role of sialic acids. In Biology of the Sialic Acids (Rosenberg, A., Ed.) pp 7-67, Plenum Press, New York. (19) Burmeister, W. P., Ruigrok, R. W. H., and Cusack, S. (1992) The 2.2 Å resolution crystal structure of influenza B neuraminidase and its complex with sialic acid. EMBO J. 11, 49-56. (20) Kiefel, M. J., and von Itzstein, M. (2002) Recent advances in the synthesis of sialic acid derivatives and sialylmimetics as biological probes. Chem. Rev. 102, 471-490.

Chem. Res. Toxicol., Vol. 16, No. 7, 2003 891 (21) von Itzstein, M., Wu, W.-Y., and Jin, B. (1994) The synthesis of 2,3-didehydro-2,4-dideoxy-4-guanidinyl-N-acetylneuraminic acid: a potent influenza virus sialidase inhibitor. Carbohydr. Res. 259, 301-305. (22) Sillanaukee, P., Po¨nnio¨, M., and Ja¨a¨skela¨inen, I. P. (1999) Occurrence of sialic acids in healthy humans and different disorders. Eur. J. Clin. Invest. 29, 413-425. (23) Jaques, L. W., Brown, E. B., Barrett, J. M., Brey, J. W. S., and Weltner, J. W. (1977) Sialic acid. A calcium-binding carbohydrate. J. Biol. Chem. 252, 4533-4538. (24) Jaques, L. W., Riesco, B. F., and Weltner, W. J. (1980) NMR spectroscopy and calcium binding of sialic acids: N-glycolylneuraminic acid and periodate-oxidized N-acetylneuraminic acid. Carbohydr. Res. 83, 21-32. (25) Saladini, M., Menabue, L., and Ferrari, E. (2002) Binding ability of sialic acid towards biological and toxic metal ions. NMR, potentiometric and spectroscopic study. J. Inorg. Biochem. 88, 6168. (26) Dillon, C. T., Lay, P. A., Bonin, A. M., Cholewa, M., Legge, G. J. F., Collins, T. J., and Kostka, K. L. (1998) Permeability, cytotoxicity, and genotoxicity of chromium(V) and chromium(VI) complexes in V79 chinese hamster lung cells. Chem. Res. Toxicol. 11, 119-129. (27) Duling, D. R. (1994) Simulation of multiple isotropic spin-trap EPR spectra. J. Magn. Reson. B104, 105-110. (28) WinSIM (1995) EPR Calculations for MS-Windows, Version 0.96, National Institute of Environmental Health Sciences. (29) Codd, R., Dillon, C. T., Levina, A., and Lay, P. A. (2001) Studies on the genotoxicity of chromium: from the test tube to the cell. Coord. Chem. Rev. 216-217, 533-577. (30) Aiyar, J., Borges, K. M., Floyd, R. A., and Wetterhahn, K. E. (1989) Role of chromium(V), glutathione thiyl radical and hydroxyl radical intermediates in chromium(VI)-induced DNA damage. Toxicol. Environ. Chem. 22, 135-148. (31) Levina, A., Zhang, L., and Lay, P. A. (2003) Structure and reactivity of a chromium(V) glutathione complex. Inorg. Chem. 42, 767-784. (32) Connett, P. H., and Wetterhahn, K. E. (1985) In vitro reaction of the carcinogen chromate with cellular thiols and carboxylic acids. J. Am. Chem. Soc. 107, 4282-4288. (33) Bramley, R., Ji, J.-Y., and Lay, P. A. (1991) Ligand-exchange reactions of chromium(V): characterization of the ligand-exchange equilibria of bis(2-ethyl-2-hydroxybutanoato(2-))oxochromate(V) in aqueous 1,2-ethanediol and the solution structure of bis(1,2ethanediolato(2-))oxochromate(V). Inorg. Chem. 30, 1557-1564. (34) Headlam, H. A., and Lay, P. A. (2001) EPR spectroscopic studies of the reduction of chromium(VI) by methanol in the presence of peptides. Formation of long-lived chromium(V) peptides complexes. Inorg. Chem. 40, 78-86. (35) Advanced Chemistry Development (ACD) Software Solaris V4.67 (1994-2002). The larger pKa value of I, relative to II, is most likely due to the increased delocalization of the negative charge in II over the R-keto or the R-hemiacetal group (40). (36) Codd, R., Levina, A., Zhang, L., Hambley, T. W., and Lay, P. A. (2000) Characterization and X-ray absorption spectroscopic studies of bis[quinato(2-)]oxochromate(V). Inorg. Chem. 39, 990-997. (37) Bramley, R., Ji, J.-Y., Judd, R. J., and Lay, P. A. (1990) Solvent dependence of the EPR spectra of oxochromate(V) complexes: solution structures and the effects of hydrogen bonding between the solvent and the complex. Inorg. Chem. 29, 3089-3094. (38) Krumpolcˆ, M., DeBoer, B. G., and Rocek, J. (1978) A stable chromium(V) compound. Synthesis, properties, and crystal structure of potassium bis(2-hydroxy-2-methylbutyrato)oxochromate(V) monohydrate. J. Am. Chem. Soc. 100, 145-153. (39) Judd, R. J., Hambley, T. W., and Lay, P. A. (1989) Electrochemistry of the quasireversible bis[2-ethyl-2-hydroxybutanoato(2-)]oxochromate(V) and -(IV) and bis[2-hydroxy-2-methylbutanoato(2-)]oxochromate(V) and -(IV) redox couples and the crystal and molecular structure of sodium bis[2-ethyl-2-hydroxybutanoato(2-)]-oxochromate(V) sesquihydrate. J. Chem. Soc., Dalton Trans., 2205-2210. (40) Hurd, C. D. (1970) The acidities of ascorbic and sialic acids. J. Chem. Educ. 47, 481-482. (41) Friebolin, H., Schmidt, H., and Supp, M. (1981) Proton NMR study of sialinic acids. Part 3. Labeling of N-acetyl-D-neuraminic acid with deuterium. Tetrahedron Lett. 22, 5171-5174. (42) Signorella, S., Daier, V., Santoro, M., Garcı´a, S., Palopoli, C., Gonza´lez, J. C., Korecz, L., Rockenbauer, A., and Sala, L. F. (2001) The EPR pattern of [CrO(cis-1,2-cyclopentanediolato)2]- and [CrO(trans-1,2-cyclopentanediolato)2]-. Eur. J. Inorg. Chem., 18291833. (43) Gottschalk, A. (1955) 2-Carboxypyrrole: its preparation from and its precursor in mucoproteins. Biochem. J. 61, 298-307.

892

Chem. Res. Toxicol., Vol. 16, No. 7, 2003

(44) Castells, M. T., Ballesta, J., Pastor, L. M., Madrid, J. F., and Marin, J. A. (1990) Histochemical characterisation of glycoconjugates in the epithelium of the extrapulmonary airways of several vertabrates. Histochem. J. 22, 24-35. (45) Washington, N., Steele, R. J. C., Jackson, S. J., Bush, D., Mason, J., Gill, D. A., Pitt, K., and Rawlins, D. A. (2000) Determination of baseline human nasal pH and the effect of intranasally administered buffers. Int. J. Pharm. 198, 139-146. (46) Westerveld, G.-J., Dekker, I., Voss, H.-P., Bast, A., and Scheeren, R. A. (1997) Antioxidant levels in the nasal mucosa of patients with chromic sinusitis and healthy controls. Arch. Otolaryngol. 123, 201-204. (47) Dabrowski, U., Friebolin, H., Brossmer, R., and Supp, M. (1979) 1H NMR studies at N-acetyl-D-neuraminic acid ketosides for the determination of the anomeric configuration II. Tetrahedron Lett. 20, 4637-4640. (48) Flippen, J. L. (1973) The crystal structure of β-D-N-acetylneuraminc acid dihydrate (sialic acid). Acta Crystallogr., Sect. B B29, 1881-1886. (49) Furuhata, K., Sato, S., Goto, M., Takayanagi, H., and Ogura, H. (1988) Studies on sialic acids. VII. The crystal and molecular structure of N-acetyl-2,3-dehydro-2-deoxyneuraminic acid. Chem. Pharm. Bull. 36, 1872-1876. (50) Czarniecki, M. F., and Thornton, E. R. (1976) Carbon-13 spin lattice relaxation in the neuraminic acids. Evidence for an unusual intramolecular hydrogen bonding network. J. Am. Chem. Soc. 98, 1023-1025.

Codd and Lay (51) Brown, E. B., Brey, W. S., Jr., and Weltner, W., Jr. (1975) Cellsurface carbohydrates and their interactions. 1. NMR of Nacetylneuraminic acid. Biochim. Biophys. Acta 399, 124-130. (52) O’Connell, A. M. (1973) The crystal structure of N-acetylneuraminic acid methyl ester monohydrate. Acta Crystallogr., Sect. B B29, 2320-2328. (53) Branca, M., Dessi, A., Micera, G., and Sanna, D. (1993) EPR and proton ENDOR study of the solution equilibria of bis(2-ethyl-2hydroxybutanoato(2-))oxochromate(V) and bis(2-hydroxy-2-methylbutanoato(2-))oxochromate(V). Inorg. Chem. 32, 578-581. (54) Pool-Zobel, B. L., Lotzmann, N., Knoll, M., Kuchenmeister, F., Lambertz, R., Leucht, U., Schro¨der, H.-G., and Schmezer, P. (1994) Detection of genotoxic effects in human gastric and nasal mucosa cells isolated from biopsy samples. Environ. Mol. Mutagen. 24, 24-45. (55) Shrivastava, H. Y., and Nair, B. U. (2000) Cleavage of human orosomucoid by a chromium(V) species: relevance in biotoxicity of chromium. Biochem. Biophys. Res. Commun. 279, 980-983. (56) Shrivastava, H. Y., Dhathathreyan, A., and Nair, B. U. (2003) Insight into structural transition, aggregation and folding of glycoprotein in the presence of chromium(III) complexes by spectroscopic, microscopic and kinetic studies. Chem. Phys. Lett. 369, 534-540.

TX034077N