Addition of Hydrogen Sulfide to Juglone - Environmental Science

May 17, 2002 - After sparging, the solution was transferred to the glovebox for storage. At pH 7.0, solid white particles were observed in the solutio...
0 downloads 8 Views 127KB Size
Environ. Sci. Technol. 2002, 36, 2663-2669

Addition of Hydrogen Sulfide to Juglone

SCHEME 1. Michael Addition Reaction Mechanism of Bisulfide at C-2

J . A . P E R L I N G E R , * ,† V . M . K A L L U R I , † R. VENKATAPATHY,† AND W. ANGST‡ Michigan Technological University, 1400 Townsend Drive, Houghton, Michigan 49931, and Swiss Federal Institute for Environmental Science and Technology (EAWAG), CH-8600 Du ¨ bendorf, Switzerland

Evidence of the addition of hydrogen sulfide to 5-hydroxy-1,4-naphthoquinone (juglone) in aqueous solution was obtained by nuclear magnetic resonance spectrometry (NMR), electron paramagnetic resonance spectrometry (EPR), UV-visible absorbance spectroscopy, and kinetic measurements. Although numerous addition reactions of thiolated alkane and aromatic compounds to quinones have been previously reported, this study indicates that inorganic forms of S(-II) act as nucleophiles and electrophiles in addition reactions to the R,β-conjugated system of the quinone. The results obtained are consistent with competing Michael and radical addition reactions, with radical addition favored with increasing pH. The simplest structure that simulated the NMR spectrum was a sulfur molecule containing sulfur bonded between two juglone molecules at C-2 or C-3, while EPR measurements of aqueous reaction solutions indicated the presence of a stable semiquinone that contained a sulfur substituent at C-2 or C-3. Quinones are present in trace amounts in natural organic matter, and the addition of S(-II) has important implications with respect to transport and transformation of a variety of compounds that react with natural organic matter.

Introduction In sulfate-reducing environments, organosulfur compounds are formed by reaction of sulfur species with natural organic matter (NOM). Organosulfur compounds have been reported in marine sediments (1-3), in freshwater lake sediments (4, 5), in salt marsh sediments (6), in peat (7), in coal (8, 9), and in kerogens (10). Reactions of inorganic sulfur and NOM take place at points of unsaturation or at points of attachment of any functional group in an organic molecule (11). The reactions take place in the early diagenesis of sediments at ambient temperatures and pressures (12-14). The formation of organosulfur compounds has been studied extensively. However, a literature search showed no report of incorporation of inorganic sulfur into quinone moieties. Quinones are present in trace quantities in NOM, and the addition of sulfur to quinones is important from an environmental science point of view, in particular in terrestrial environments, where quinone concentrations in natural organic matter can be expected to be higher as compared to those in marine environments. Incorporation of inorganic * Corresponding author phone: (906) 487-3641; fax: (906) 4872943; e-mail: [email protected]. † Michigan Technological University. ‡ Swiss Federal Institute for Environmental Science and Technology. 10.1021/es015602c CCC: $22.00 Published on Web 05/17/2002

 2002 American Chemical Society

sulfur into quinones in NOM has implications with respect to the formation of organosulfur compounds, transport and transformation of metals and organic pollutants, and biogeochemical cycling of sulfur, nitrogen, and carbon. Along with phenolic groups, ketonic groups, and complexed transition metals in NOM, quinones are capable of carrying out the oxidation-reduction reactions in which NOM has been shown to be involved (15, 16). Incorporation of inorganic sulfur into quinone molecules may alter the oxidationreduction and acid-base properties of NOM significantly, in turn affecting the transformation and transport of the NOM, inorganic sulfur, nitrogen, metals, and other organic compounds. The objectives of this work were to determine the structure of addition products of S(-II) and one quinone, 5-hydroxy1,4-naphthoquinone, or juglone, the mechanism of their reaction, and to examine the subsequent reaction of mercaptoquinones with substances that react with NOM, including polyhalogenated alkanes. In an earlier paper, we reported that electrochemically reduced juglone was unreactive with respect to hexachloroethane transformation, whereas solutions containing juglone and sodium sulfide were reactive (17). Juglone is a naturally occurring compound that is known to be produced by 50 species of trees in seven genera native to North temperate and subtropical regions, south to the Andes, and eastern Asia (18, 19). Juglone was also interesting as a study compound because earlier works examined the addition of thiols to it (see, e.g., see review by Finley (20)). Thomson (21) presented the first experimental evidence of competing nucleophilic and electrophilic additions of thiols to juglone. These addition mechanisms (22) (Michael addition of bisulfide, a nucleophilic addition, and radical addition of sulfhydryl radical, an electrophilic addition) are schematized for juglone and inorganic S(-II). Through the use of property information, application of multiple spectrometric techniques, kinetic experiments, and computational chemistry, reaction product structure, abundance, and mechanisms of formation were examined.

Materials and Methods Chemicals. The following chemicals were used in experimentation: methanol (nanograde; Burdick & Jackson; Muskegon, MI), ethanol (200 proof dehydrated alcohol USP Punctilious; Quantum, Foster City, CA), glycine (99+%; Aldrich, Milwaukee, WI), 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES; ACROS, Belgium), iodine (USP; I.T. Baker), juglone (5-hydroxy-1,4-naphthoquinone, 97%; VOL. 36, NO. 12, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2663

SCHEME 2. Radical Addition Reaction Mechanism of Sulfhydryl Radical at C-3

Aldrich), potassium biiodate (Fisher Scientific, Pittsburgh, PA), potassium iodide (99+%; Aldrich), sodium chloride (certified ACS; Fisher Scientific), sodium hydroxide (50% w/w; Fisher Scientific), sodium thiosulfate salt (certified ACS; Fisher Scientific), sodium sulfide nonahydrate (98%; Aldrich), sulfuric acid (ACS reagent; Fisher Scientific), acetonitrile (99.9%; ACROS), benzidine ISOPAC (Sigma, St. Louis, MO), sodium metaperiodate (Fisher Scientific), toluene (certified ACS; Fisher Scientific), erythro-1,4-dimercapto-2,3-butandiol (dithioerythritol, 99%; Aldrich), 4-hydroxy-TEMPO (Aldrich), acetic acid glacial (99.9%; Fisher Scientific), deuterated chloroform (99.8 atom % D; Aldrich), ethyl acetate (99.9%; Fisher Scientific), diethyl ether (certified ACS; Fisher Scientific), and hexane (99.9%; Fisher Scientific). Preparation of Sodium Sulfide Stock Solutions. Two methods were used to prepare the sodium sulfide stock solution that was later added to reaction solutions. In Method 1, a known volume of Milli-Q water was sparged with nitrogen at the rate of 100 min/60 mL of water. Na2S crystals were rinsed with nitrogen-sparged water, weighed, and added to nitrogen-sparged Milli-Q water to obtain 0.5 M. The solution was immediately transferred to a glovebox filled with a 96%/ 4% mixture of N2/H2 containing a palladium catalyst in a ventilator (COY Laboratory Products Inc., Grasslake, MI) for storage. To adjust the pH of juglone-sodium sulfide solution to the required level, HCl was added to the juglone-buffer solution immediately followed by the addition of sodium sulfide stock solution. In Method 2, the 0.5 M Na2S stock solution was prepared by weighing Na2S crystals and dissolving them in nitrogen-sparged Milli-Q water, followed by the addition of 0.01 M sulfuric acid to lower the pH to the required level. This solution was sparged with N2(g) at the rate of 100 min/60 mL of solution. After sparging, the solution was transferred to the glovebox for storage. At pH 7.0, solid white particles were observed in the solution a few weeks after preparation. Preparation of Reaction Solutions for Thin-Layer Chromatography (TLC) and Nuclear Magnetic Resonance (NMR) Spectrometry. The juglone-sodium sulfide reaction solution was prepared by weighing the equivalent of 1 mmol of juglone in a round-bottom flask, and transferring into the glovebox. Then, 30 mL of 0.5 M glycine (for pH > 8.5) or 0.5 M HEPES (for pH e 8.5) buffer that had been sparged with nitrogen (30 min/100 mL of aqueous solution) was added to the juglone. Next, 30 mL of ethanol (sparged with nitrogen gas for 30 min and transferred to the glovebox) was added to enhance the solubility of juglone. Finally, 20 mL of 0.5 M standardized Na2S solution sparged with nitrogen prepared according to Method 2 was added to the mixture, and it was stirred overnight at room temperature in the glovebox. Final concentrations of each solute in this solution were the following: juglone (12.5 mM), sodium sulfide (125 mM), and HEPES/glycine (187.5 mM). 2664

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 12, 2002

Preparation of Reaction Solutions for Electron Paramagnetic Resonance Spectrometry (EPR) and UV-vis Spectrometry and HCE Kinetic Experiments. Method 1 was used to generate the sodium sulfide stock solution. Further details of the methods can be found in ref 17. Electrochemical Reduction of Juglone. Juglone was electrochemically reduced at a platinum net in a specially designed cell as follows. To a 300-mL buffer solution (pH 7.0, 50 mM HEPES buffer, 0.1 M ionic strength reached by adding KCl to the solution), 3 mL of 0.02 M juglone in air-equilibrated methanol was added. This solution was placed in the electrochemical cell and sparged with argon through the porcelain frit for 3 h while stirring. The electrochemical cell was transferred into the glovebox, and all electrodes were connected to the potentiostat outside the glovebox via wires that ran through the side of the glovebox. A potential difference of -200 mV versus standard hydrogen electrode (SHE) was applied and the solution was stirred. The reduction required approximately 3 h to completion at pH 7.0. The reduction was verified by the transfer of 3 mL of the solution to a cuvette fitted with a septum and a screwcap and the measurement of the UV-vis spectrum. TLC. DC-Fertigplatten Kieselgel 60/kieselgur, silica gelcoated, glass backed, 250 µm thick, 20 × 20 cm TLC plates, with and without fluorescence indicator at 254 nm (Merck KgaA, Darmstadt, Germany) were used as preparative plates for the separation of compounds to obtain samples for NMR spectrometry. The elution solvent was a 1:1 mixture of acetonitrile and toluene. The compounds from the reaction mixtures were extracted from aqueous solvent into an organic solvent (e.g., toluene) before application to the TLC plates. Separation was achieved by passing mobile phase through the silica gel in a developing chamber. Precautions taken to avoid oxygen included carrying out plate development in the glovebox using nitrogen-sparged solvents and plates that had been evacuated and stored overnight in the transfer chamber of the glovebox in the N2/H2 atmosphere. Visualization of the separated compounds was carried out by viewing the TLC plates under short (254 nm) and long (364 nm) wavelength UV light or by using iodine spray reagent (23). NMR. The reaction mixture was extracted into 40 mL of either N2-sparged ethyl acetate or diethyl ether in a separatory funnel in the glovebox and concentrated by evaporating the solvent using a rotoevaporator outside the glovebox, taking precautions to minimize contact with air. Then, 6 mL of N2sparged ethyl acetate or diethyl ether was added to this solution, and 500 µL of the sample was separated on a TLC plate. Twelve plates were developed to obtain the minimum mass (30 mg) required for 1H and 13C NMR spectrometry. The separated compounds were scraped off the TLC plate, dissolved in diethyl ether or ethyl acetate, and vacuum filtered into round-bottom flasks. The filtrate was rotoevaporated, and the sample was dissolved in ∼1 mL of CDCl3 and placed in NMR sample tubes. The round-bottom flasks were weighed before and after the sample was transferred to the sample tubes to obtain sample mass. The operating parameters for the UNITYInova 400 MHz Varian (Palo Alto, CA) instrument for 1H NMR and gradient heteronuclear multiple quantum coherence (gHMQC) NMR spectrometry were, respectively, a minimum sample mass of 5 and 30 mg, a duration of 1 min and 16 h, and the type of probe in both cases was indirectNMR. The yield of product in the 1H NMR spectrum of Fraction 3 was computed from published peak areas of protons in juglone, the mass of Fraction 3 collected relative to the total mass on the plate, and the peak areas of the 6.98 and 7.01 ppm peaks measured in the spectrum. Details of the calculation can be found in ref 24. EPR. Aqueous samples for EPR analysis were transferred by syringe from the serum flasks to the EPR thin layer quartz

sample holder in a glovebox filled with N2 gas. The instrumentation consisted of a Varian E-9 century series spectrometer with a 9-in. magnet, an E-102 microwave bridge, an E-204 low-frequency module, and a data aquisition system on an IBM-AT computer containing a 1 MB hard disk. Operating parameters were the following: field set, 3384.0 G; microwave power, 1 mW; microwave frequency, 9.45 GHz; time constant, 2 s.; modulation, 0.63 G; scan time, 30 min; calibration, 10 µM-1 mM 4-hydroxy-TEMPO in water. The relative standard deviation in replicates of standards was 20%. UV-vis Absorbance Spectroscopy. Aqueous sample was transferred by syringe from serum vials to 1-cm quartz cuvettes equipped with septum-lined screw-caps in the glovebox. Instrumentation consisted of a Hitachi U-2000 spectrophotometer connected to a Compac PC. The wavelength range was 190-700 nm, and the scan rate was 1200 nm min-1. Computation of Charge Distributions. Semiempirical calculations were calculated using CAChe Worksystem software (Oxford Molecular Group, U.K.) version 3.9, running on a PowerMac. The calculations involve optimizing the molecular structure using MOPAC with PM3 parameters followed by a population analysis of molecular orbitals of the optimized structure. Charges were found under the “net atomic charges and dipole contributions” section of the output. Computation of 1H NMR Shifts. Density functional calculations were carried out using the Gaussian 98 (Gaussian Inc., Carnegie, PA) suite of programs on a Sun UltraSparc 40 workstation. Computational methods of determination of proton NMR shifts involved optimizing the structures of any given molecule and a reference compound (tetramethylsilane (TMS)) using B3LYP/6-31g(d) theory/basis set. Later, NMR magnetic shielding tensors for each atom in a molecule were calculated by performing a single-point calculation using B3LYP/6-311++g(2d,p) theory/basis set and NMR)CSGT keyword for both the molecule and TMS. Values of 1H NMR shifts in a molecule were then obtained by subtracting the NMR shielding tensors for hydrogen atoms in the molecule from the tensors for hydrogen atoms in TMS.

Results Charge Distributions. Partial atomic charges on carbons in the various species of juglone, including consideration of hydrogen bonding (25), can be used to indicate which carbon atoms in the various species of juglone present at the various pH values and redox conditions studied are most susceptible to nucleophilic and electrophilic attack (Figure S1, Supporting Information). The relative susceptibility of C-2 and C-3 in the quinone species to nucleophilic attack was dependent on pH. C-2 was computed to be more susceptible to nucleophilic attack than C-3 at pH < pKa, whereas the opposite was true at pH > pKa. These predictions are corroborated by measured hyperfine couplings observed in EPR spectra of the radical anion of a molecule, which are related to the site of reaction of nucleophiles in quinones (26). Because a2 ) 3.30 G and a3 ) 3.05 G (27, 28), the site of nucleophilic attack is predicted to be C-2 at pH < pKa. C-7 was generally most susceptible to nucleophilic attack in all species of juglone except in the two hydroquinone species. In deprotonated hydroquinone, C-2 was most susceptible. C-6 carried the highest partial charge in all species except the hydroquinones, indicating that it was relatively most susceptible to electrophilic attack in those species. In the hydroquinone, C-3 and C-2 were computed to be most susceptible to electrophilic attack in the deprotonated and doubly deprotonated species, respectively. TLC and NMR. Direct evidence of addition to juglone was obtained through separation of the reaction mixture

FIGURE 1. TLC separation of a juglone-hydrogen sulfide reaction mixture at pH 8.5 for NMR measurements. The developing solvent was 1:1 acetonitrile/toluene. The Na2S stock solution was prepared according to Method 2. using TLC (Figure 1) and NMR (Figure 2). Fraction 1 in Figure 1 was colorless and nonpolar, and its 1H NMR spectrum contained no signals in the aromatic region. The iodineazide spray test indicated that this compound was a thiol or a disulfide. This compound eluted at the same time on the TLC plates from solutions that contained sodium sulfide prepared according to Method 2 but was absent in solutions containing juglone in buffer, leading to its identification as a polysulfide. Fraction 2 eluted at the same time as juglone quinone and had identical 1H (29, 30) and 13C (31) NMR spectra to juglone, indicating that it was juglone. Fraction 3 responded positively to the iodine-azide spray test. From the 1H NMR spectra, it was apparent that Fraction 3 was not pure. Nearly all of the samples contained a significant amount of juglone, possibly due to either tailing during separation of the reaction mixture or transformation of addition products during plate development. The lack of purity was confirmed in the finding that juglone was present on a TLC plate to which Fraction 3 in organic solvent was reapplied and developed. Although purified product may have been obtained by successively redeveloping TLC plates and isolating the product prior to NMR analysis, this procedure caused considerable loss of product in a preliminary test and was not performed. In addition, to prevent oxidation of -SH in the aqueous reaction solutions, we attempted to methylate the thiol using methyl trifluoromethanesulfonate and a modified procedure by Arnarp et al. (32). However, this approach was not successful. The presence of juglone in Fraction 3 made 13C NMR a poor tool in the determination of the structure of the product, but using 1H NMR, the juglone signals could be easily identified from published proton shifts. The proton NMR spectrum of Fraction 3 contained a singlet with a shift of 7.01 ppm that was not present in the proton NMR spectrum of juglone. This peak was downfield of the signal at 6.93 ppm, which corresponds to protons on the 2 and 3 carbons of juglone. The gHMQC spectrum in Figure 2 indicated that this proton is attached to a carbon with a shift in the region of the C-2 and C-3 shifts in the juglone spectrum. Whether the substituent is present on C-2 or C-3 cannot be deduced from the NMR spectra because influence of substitution on the shift of the neighboring atoms cannot be deduced from such spectra. No peak corresponding to a sulfhydryl proton was observed. Its absence may have been the result of oxidation of the -SH group attached at C-2 or C-3 to the disulfide species or to a higher oxidation state species. Also, the absence of a nonproton substitutent at C-6 leads to the conclusion that this carbon was not the site of electrophilic addition. Ratios of peak areas in the aromatic region of the proton NMR spectra of Fraction 3 indicated the presence of product. There are three sets of signals in the aromatic region of the published 1H spectrum of juglone (29, 30) and of our spectrum of pure juglone (not shown). On integration, they yielded a ratio of 2:1:2 when the five aromatic protons are taken into VOL. 36, NO. 12, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2665

FIGURE 2. 2D-NMR (gHMQC) spectrum (aromatic region) of Fraction 3 at pH 12.5. The Na2S solution was prepared according to Method 2. The spectrum indicates that the peak at 7.01 ppm is attached to either C-2 or C-3. account (C-7, C-8:C-6:C-2, C-3; 7.65/7.25/6.93 ppm). The shifts at 7.65 and 7.25 ppm remained unaffected in the spectrum of Fraction 3. A substituent at either C-2 or C-3 of juglone does not affect the shifts of the protons on C-6, C-7, and C-8 due to the presence of the intervening quinone group. In the 1H NMR of Fraction 3, this ratio was not 2:1:2, which further indicates the presence of a reaction product. A peak upfield of the 6.93 ppm peak at 4.98 ppm was observed in some Fraction-3 1H NMR spectra over the course of the 2-year period over which proton NMR spectra were obtained. This peak is of interest because it could indicate an addition product with a substituent at C-7, which is reasonable to expect given that C-7 was predicted from atomic charge calculations to be the most susceptible to nucleophilic attack in all juglone species except the hydroquinones. Laugraud et al. (33) showed that the β proton (the proton adjacent to the site of addition at either C-2 or C-3) of the 2-phenylthio derivative of juglone exhibited a β-proton shift of 5.98 ppm. A density functional calculation of this shift produced a value of 5.41 ppm. The difference in the measured and simulated shift (0.57 ppm) was larger than the difference in shifts for any of the protons in the simulated and measured shifts of pure juglone (max difference ) 0.46 ppm) but gives an indication of the magnitude of uncertainty in the simulations. 1 H NMR shift simulations of 19 possible structures with substitution at C-2, C-3, or C-7 using density functional theory revealed structures that did not contain peaks between 4.00 and 6.93 ppm corresponding to a shift of a β proton (simulated β-proton shifts appear in parentheses in the following discussion) and structures that did. None of the structures simulated produced β-proton shifts nearer than 1 ppm downfield of the observed 4.98 ppm peak. This observation leads us to conclude that this peak did not indicate a β proton. On the basis of the simulations, the simplest structure having an NMR spectrum similar to the observed spectrum is a sulfur molecule containing sulfur bonded to two juglone molecules 2666

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 12, 2002

FIGURE 3. Observed EPR spectrum of the juglone-sodium sulfide solutions. Aqueous solution contained 60 mM glycine buffer (pH 11), 10 mM Na2S (prepared according to Method 1), and 500 µM juglone. In identical solutions lacking Na2S, no EPR signal was detected. at C-2 or C-3 and at C-2′ or C-3′ (7.19 ppm; 7.24 ppm; avg, 7.25 ppm). However, given the uncertainty of the simulations, other structures are certainly possible. EPR. EPR spectrometry was also employed to detect paramagnetic semiquinones. EPR analysis of an aqueous juglone-sodium sulfide solution indicated the presence of radicals that were stable for at least 1 week. No EPR signal was obtained when solutions containing glycine buffer or glycine buffer with 10 mM Na2S were analyzed. A simulated spectrum of juglone semiquinone using published splitting constants (27, 28) was identical to the measured spectrum of juglone electrochemically reduced to the semiquinone form but was quite different from the spectrum of juglone solutions containing sodium sulfide (Figure 3). The measured spectrum was 5.75 G wide as compared to the EPR spectrum of juglone semiquinone, which is 9.87 G in width. The smaller width of the spectrum may have resulted from the formation of a stable radical through an addition reaction. The addition of an amino group to the 2 position of 1,4-naphthoquinone decreased a2 by 1.4 G and a3 by 2.9 G (27, 28). Similarly, the addition of glutathione to 1,4-naphthoquinone causes a decrease in a2 of 2.7 G and in a3 of 1.5 G (27, 28).

TABLE 1. Concentrations, Yields, and Pseudo-First-Order Rate Constants for Disappearance of Hexachloroethane in Solutions Containing Juglone and Sodium Sulfide

pH 7.0 7.5 8.0 8.5 8.9 9.9 12.5

NMR EPR UV-vis yielda [radical]b yieldb,c [mjug]d yieldd,e (%) (µM) (%) (µM) (%)

kobsb, f (10-6 s-1)

ndg

nd

2.0

2.0

2.1 ((0.094)

1.2 8.8 9.6 15

0.2 1.8 1.9 3.0

3.0

3.0

7.0 10

7.0 10

4.1 ((1.4) 7.2 ((3.4) 13 ((2.8) 22 ((2.5)

3.5 6.6 7.1

a Conditions: 188 mM HEPES/glycine buffer, 125 mM Na S, 12.5 2 mM juglone. b Conditions: 50 mM HEPES or 61-94 mM glycine buffer, 10 mM Na2S, 500 µM juglone. c Percent of the added juglone that was present as the semiquinone species. d Conditions: 50 mM HEPES or 61-94 mM glycine buffer, 10 mM Na2S, 100 µM juglone. The concentration was estimated from the absorbance at λ ) 415 nm divided by the absorption coefficient of juglone of 8000 M-1 cm-1. e Percent of the added juglone that was converted to mercaptojuglone. f Pseudofirst-order rate constants for disappearance of hexachloroethane in juglone-hydrogen sulfide solutions. Numbers in parentheses represent one standard deviation based on three replicates. g nd ) not detected.

FIGURE 4. UV-visible spectrum of 100 µM juglone in 50 mM HEPES buffer at pH 7.0. The reduced form was generated through the addition of 10 mM Na2S prepared according to Method 1 or by electrochemical reduction. The concentration of radical and percent yield of product increased with increasing pH at constant total juglone concentration (Table 1). At pH 7.0, no radicals were detected by EPR. The observation of radicals corresponded to the observation of particles in the solutions. At pH values equal to 8.5, the particles were grayish, and at pH values >8.5 the particles were dark brown. Particles may have resulted from the precipitation of elemental sulfur, quinhydrones, or other insoluble polymerization products from solution. UV-vis Absorbance Spectroscopy. Solutions containing juglone and sodium sulfide exhibited different absorbance spectra than those of juglone reduced electrochemically (Figure 4). In the spectrum of the juglone-sodium sulfide solution, the reduced juglone has maximum absorbance (λmax) at 343 nm at pH 7.0. In solutions containing 10 mM sodium sulfide, 99.9% of the juglone is expected to be reduced to hydroquinone (34). However, the measured spectrum contained a shoulder at 415 nm, corresponding to the λmax of juglone (Figure 4). This peak was found in all solutions having pH values of 7-10 containing juglone and sodium sulfide. The spectrum of electrochemically reduced juglone contained no shoulder (Figure 4); the shoulder in the spectrum occurred only in the presence of sodium sulfide. The presence of a sulfur atom in the β position of an R,βunsaturated carbonyl compound causes a bathochromic shift in the π f π* transition of 85 nm (35). Assuming that a similar shift occurs in the semi- or hydroquinone form of the reaction

FIGURE 5. Yields and pseudo-first-order rate constants for the disappearance of hexachloroethane in solutions containing juglone and sodium sulfide. Solution conditions can be found in Table 1. product, the λmax of this transition would be (343 + 85) or 428 nm. This value is close to the λmax of the shoulder at 415 nm, suggesting that the observed shoulder corresponds to the π f π* transition of 5-hydroxy-2-(or -3-)mercapto-1,4naphthoquinone and, by inference, the semiquinone and hydroquinone forms, which were more likely the species that were present at the high sodium sulfide concentrations in solution. Reaction Kinetics of Hexachloroethane (HCE) Disappearance. HCE disappearance kinetics were measured as described by Perlinger et al. (17). HCE was not transformed over 40 days in solutions containing electrochemically reduced juglone, whereas pseudo-first-order disappearance was observed when the solutions contained juglone and sodium sulfide. The kobs values increased with increasing pH, radical concentration, and total product concentration (Table 1; Figure 5). Above the pKa of juglone (8.5), kobs increased, suggesting that oxidation of the deprotonated thiol is more facile than that of the protonated product. The results suggest that the semi- or hydroquinone forms of the product were the reactive species with respect to HCE transformation in solutions containing juglone and sodium sulfide. Interestingly, the magnitude of the rate constant for HCE disappearance depended on the order in which the juglone and sodium sulfide solutions were added. Pseudo-first-order rate constants were 1.4-5.7 times higher if the sodium sulfide stock solution was added prior to the juglone stock solution between pH 10 and 12, suggesting greater yields of product in this case. The higher yields may be the result of a different mechanism of product formation under the two conditions or to the relative kinetics of the addition reaction and reduction of mercaptojuglone.

Discussion Addition Reaction Mechanism. On the basis of the results, competing Michael and radical addition of S(-II), schematized in the Introduction section, may explain these results. Nucleophilic addition likely occurs at lower pH through stabilization of reaction intermediates, as shown in Scheme 1. Because loss of a proton rather than gain of a proton is required for the formation of a stable species in the radical addition reaction (Scheme 2), such a reaction is favored with increasing pH. Although, in the more oxidized quinone and semiquinone species, radical addition was predicted from atomic charges on carbon to occur at C-6 and addition at C-6 was ruled out by NMR, in the hydroquinone radical, addition was predicted at C-3 or C-2. An initially formed thiolated product may react further as a nucleophile and form polymeric species that comprise the particles observed in reaction solutions. Preliminary X-ray photoelectron spectroscopy measurements of sulfur in precipitates from aqueous reaction mixtures (not shown) support this possibility. Thomson presented early experimental evidence documenting competing nucleophilic and electrophilic additions VOL. 36, NO. 12, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2667

to juglone (20, 21, 36). Thiols were found to add to juglone at C-3, supporting a radical addition reaction of the thiols under the experimental conditions of the study. Product isomerization has been suggested as another explanation for these observations (37). syn-Hydroxyl at C-5 acts as an electron donor, activating C-2 for nucleophilic attack, as deduced from comparison of ELU El values computed using SCF ab initio theory/STO-3G basis set and experimental results (37). Although bisulfide is, by far, the most abundant species in these solutions, radical addition can be expected to become more favorable than Michael addition of bisulfide with increasing pH. Alkaline autoxidation can form sulfhydryl radicals at higher pH, and this neutral species is capable of reacting with deprotonated species without the competing repulsion of like charges encountered by bisulfide. Quinone molecules such as juglone show an affinity toward radicals, and quinones are good antioxidants (i.e., radical quenchers (38)). EPR spectrometry studies have shown that hyperconjugation and resonance stabilization are extraordinarily effective in stabilizing radicals (39). Through this stabilization process, quinones react with odd-electron species to form stable radicals themselves. Free radical addition to multiple bonds is the most common reaction that a thiyl radical can undergo besides the formation of disulfides through dimerization (40). The radical addition mechanism requires that sulfhydryl radical be present in large enough quantities to react. The reduction potential of HS• to form bisulfide ion has been computed to be +1.15 V versus SHE (41), indicating that the reverse reaction, oxidation of bisulfide to sulfhydryl radical, is thermodynamically unfavorable. However, thiyl radicals are known to exist in the presence of certain oxidants including oxygen (40, 42) and exhibit a range in reduction potentials that includes that of sulfhydryl radical (43). In the sparged solutions investigated here, oxygen may have been introduced through the addition of reagent stock solutions to reaction mixtures. For example, the juglone stock solution was added in air-equilibrated methanol containing 2.06 × 10-3 M O2 at 25 °C (44). Higher product formation due to higher oxygen concentration in the region of the juglone molecules when this stock solution was added to sodium sulfide in buffer may explain the higher yield indicated by the 1.4-5.7 times higher rate of HCE transformation as compared to solutions in which sodium sulfide was added after the juglone stock solution. Because strict precautions against light were not taken, photolysis reactions may also be responsible for generation of thiyl radical. Traces of oxygen in sparged reaction solutions could result in the formation of thiyl radicals through photolysis of a disulfide molecule (40). Among the radicals that are generated through photolysis of oxygen in water (38), superoxide radical (O2•-) has been shown to react with thiols to form thiyl radicals (45). In the reaction mixture containing juglone and sodium sulfide, radicals detected by EPR at pH > 7.0 may indicate the reaction of superoxide with mercaptojuglone. Quinones react with superoxide at very high rates to form semiquinone (46). Thiyl radical may also have been generated during one-electron reduction of quinone or semiquinone, one-electron oxidation of semiquinone or hydroquinone, hydrogen abstraction from hydrogen sulfide by another radical, or reaction of an oddelectron species with bisulfide, disulfide, or polysulfide species. Formation of the hydroquinone and semiquinone forms of product through addition was suggested in the previous discussion and shown in Schemes 1 and 2, respectively, but subsequent reactions likely occurred. Alkaline autoxidation can result in the formation of disulfides from thiyl radicals such as thiolated semiquinone. Apart from oxidation of thiols 2668

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 12, 2002

to disulfides, thiols and disulfides can be oxidized to higher oxidation states through autoxidation in alkaline medium (42). These reactions and others including disproportionation, oxidation-reduction, acid-base reactions, and polymerization likely led to an equilibrium between quinone, semiquinone, hydroquinone, and particulate forms of mercaptojuglone and juglone in solution. Numerous researchers have demonstrated that thiols add to double bonds, and specifically to quinones (see reviews in refs 20, 47, and 48). Addition of thiols to double bonds occurs in the early diagenesis of natural organic matter in S(-II)-containing environments at ambient temperature and pressure, as discussed previously. The reactions presented in this study differ from many previous studies presented in the geochemistry literature in that (1) inorganic S(-II) rather than an organic thiol was the reactant; and (2) the intermediates proposed in Schemes 1 and 2 are stabilized through resonance of the aromatic molecule that is formed (the ability of quinones to assume an aromatic structure makes radical addition energetically more favorable than with many of the compounds previously shown to undergo additions); and (3) the results are consistent with competing Michael and radical addition reactions, with radical addition assuming predominance as pH increases. Nucleophilic addition alone is typically recognized in causing sulfurization of natural organic matter. Subsequent Reactions of Mercaptoquinones in NOM. Quinones are present in trace amounts in NOM. For example, the quinone content in Suwanee River Fulvic Acid was estimated by EPR to be 1-2 molecules in 1000 molecules of NOM with an estimated molecular weight of 1000 daltons (49). Radicals in dissolved natural organic matter (DOM) are usually attributed to the presence of semiquinones (50). These semiquinones may, in part, be stabilized through addition of nitrogen- or sulfur-containing compounds according to the reactions discussed previously. Radical stabilization may encourage further odd-electron-transfer processes to occur. Environments that will tend to favor such reactions occur when marine, estuary, lake, pond, or wetland sediments or surface waters of different origin mix at interfaces, such as between aerobic waters and DOM-containing anoxic sediment porewaters (e.g., ref 3), or between hydrogen sulfidecontaining waters and the photic zone of surface waters rich in DOM (e.g., ref 51). Reaction of inorganic sulfur with NOM also increases the polyfunctionality, polyelectrolytic character, reducing power, and nucleophilicity of NOM, and these changes likely affect the transport and transformation of other elements in the environment. For example, the second-order rate constant for the reaction of mercaptojuglone, computed based on yield at pH 7.0 from the UV-vis absorbance at 415 nm (see Table 1), was higher than that of polysulfides by more than 2 orders of magnitude (17, 52, 53). Also, because of the presence of sulfur in ortho position to carboxyl oxygen, thiol addition products of quinones can be expected to be stronger ligands than the parent quinone with respect to class B metal (bidentate) complexation (54). Through altered oxidationreduction and acid-base properties, the transport and transformation of metals and NOM will be influenced by such complexation reactions. In addition, sulfurization of quinones in natural organic matter has been implicated to outcompete the addition of nitrogen-containing compounds, including pollutants, to NOM (55-59). Prevention of binding will lead to greater bioavailability and transport of nitrogenous species. Therefore, these abiotic addition reactions may play a role in the biogeochemical cycles of sulfur, nitrogen, and carbon as well as metals.

Acknowledgments The authors gratefully acknowledge Dr. B. Mu¨ller (EAWAG, Kastanienbaum, Switzerland) for collaboration in design of the electrochemical cell, Prof. A. Schweiger (ETH, Zu ¨ rich, Switzerland), who facilitated use of the EPR instrument, and Drs. N. Urban (Michigan Tech) and U. Jans (CUNY) for valuable comments on the manuscript.

Supporting Information Available Figure S1. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Mopper, K.; Taylor, B. F. Thiols in coastal marine sediments. In Biogeochemical Cycling of Sulfur; Sohn, M. L., Ed.; ACS Symposium Series 305; 1986; pp 324-339. (2) Vairavamurthy, A.; Mopper, K. Nature 1987, 329, 623-625. (3) Henneke, E.; Luther, G. W. I.; de Lange, G.; Hoefs, J. Geochim. Cosmochim. Acta 1997, 61, 307-321. (4) Losher, A. The sulfur cycle in freshwater lake sediments and implications for the use of C/S ratios as indicators of past environmental change, Dissertation, ETH, Zu ¨ rich, Switzerland, 1989. (5) Urban, N. R.; Ernst, K. H.; Bernasconi, S. Geochim. Cosmochim. Acta 1999, 63, 837-853. (6) Ferdelman, T. G.; Church, T. M.; Luther, G. W. Geochim. Cosmochim. Acta 1991, 55, 979-988. (7) Brown, K. A. Soil Biol. Biochem. 1986, 18, 131-140. (8) Casagrande, D. J.; Ng, L. Nature 1979, 282, 598-599. (9) Casagrande, D. J.; Idowu, G.; Friedman, A.; Ricket, P.; Siefert, K.; Schlenz, D. Nature 1979, 282, 599-600. (10) Svel-Cercovecki, S.; Eskinja, I.; Alajbeg, A. J. Anal. Appl. Pyrolysis 1995, 34, 143-156. (11) Sinninghe Damste´, J. S.; Rijpstra, W. I.; Kock-Van Dalen, A. C.; de Leeuw, J. W.; Schenck, P. A. Geochim. Cosmochim. Acta 1989, 53, 1343-1355. (12) LaLonde, R. T.; Ferrara, L. M.; Hayes, M. P. Org. Geochem. 1987, 11, 563-571. (13) Kohnen, M. E. L.; Sinninghe Damste´, J. A.; De Leeuw, J. W. Nature 1991, 349, 775-778. (14) Schouten, S.; van Driel, G. B.; Sinninghe Damste´, J. S.; de Leeuw, J. W. Geochim. Cosmochim. Acta 1993, 57, 5111-5116. (15) Thurman, E. M. Organic Geochemistry of Natural Waters; Martinus Nijhoff/Dr W. Junk: Dordrecht, The Netherlands, 1985. (16) Helburn, R. S.; MacCarty, P. Anal. Chim. Acta 1994, 295, 263272. (17) Perlinger, J. A.; Angst, W.; Schwarzenbach, R. P. Environ. Sci. Technol. 1996, 30, 3408-3417. (18) Thomson, R. H. Naturally Occurring Quinones; Academic Press: London, U.K., 1971. (19) Schmidt, R. J. http://bodd.cf.ac.uk/BotDermFolder/BotDermJ/ JUGL.html (accessed 1999). (20) Finley, K. T. In The Chemistry of Quinonoid Compounds; Patai, S., Ed.; Wiley Interscience: New York, 1974; pp 877-1144. (21) Thomson, R. H. J. Org. Chem. 1951, 16, 1082-1090. (22) March, J. Advanced Organic Chemistry; Wiley-Interscience: New York, 1992. (23) Touchstone, J. C.; Dobbins, M. F. Practice of Thin-Layer Chromatography; Wiley: New York, 1978. (24) Kalluri, V. M. Reactions of S(-II) and NOM in Water, Thesis, Michigan Technological University, Houghton, MI, 1999. (25) Jaccard, G.; Lauterwein, J. Helv. Chim. Acta 1986, 69, 14691485. (26) Epiotis, N. D.; Cherry, W. J. Am. Chem. Soc. 1976, 98, 5432. (27) Fischer, H. Magnetische eigenschaften freier radikale. In LandoltBo¨rnstein. Zahlenwerte und Funktionen aus Naturwissenschaften und Technik; Madelung, O., Ed.; Springer-Verlag: Berlin, Germany, 1977.

(28) Pedersen, J. A. CRC Handbook of EPR Spectra from Quinones and Quinols; CRC Press: Boca Raton, FL, 1985. (29) Moore, R. E.; Scheuer, P. J. J. Org. Chem. 1966, 31, 3272-3283. (30) Parker, K. J. Org. Chem. 1981, 46, 3218-3223. (31) Bowden, B. F.; Cameron, D. W.; Crossley, M. J.; Feutrill, G. I.; Griffiths, P. G.; Kelly, D. P. Aust. J. Chem. 1979, 32. (32) Arnarp, J. L.; Kenne, B.; Lindberg, B.; Lonngren, J. Carbohydr. Res. 1975, 44, C5-C7. (33) Laugraud, S.; Guingant, A.; Chassagnard, C.; d’Angelo, J. J. Org. Chem. 1988, 53, 1557-1560. (34) Clark, W. M. Oxidation-Reduction Potentials of Organic Systems; Waverly Press: Baltimore, MD, 1960. (35) Pretsch, E.; Clerc, T.; Seibl, J.; Simon, W. Tables of Spectral Data for Structure Determination of Organic Compounds; SpringerVerlag: Berlin, Germany, 1983. (36) MacLeod, J. W.; Thomson, R. H. J. Org. Chem. 1960, 25, 36-42. (37) Rozenboom, M. D.; Tegmo-Larsson, I.-M.; Houk, K. N. J. Org. Chem. 1981, 46, 2338-2345. (38) Larson, R. A.; Weber, E. J. Reaction Mechanisms in Environmental Organic Chemistry; Lewis: Boca Raton, FL, 1994. (39) Abell, P. I. Addition to multiple bonds. In Free Radicals; Kochi, J., Ed.; Wiley: New York, 1973. (40) Kice, J. L. Sulfur-centered radicals. In Free Radicals; Kochi, J., Ed.; Wiley: New York, 1973; pp 711-740. (41) Surdhar, P. S.; Armstrong, D. A. J. Phys. Chem. 1986, 90, 5915. (42) Oae, S. Organic Sulfur Chemistry: Structure and Mechanism; CRC Press: Boca Raton, FL, 1991. (43) Wardman, P. J. Phys. Chem. Ref. Data 1989, 18, 1755. (44) Pospisil, J.; Luzny, Z. Collect. Czech. Chem. Commun. 1960, 25, 589-592. (45) Asada, K.; Kanematsu, S. Agric. Biol. Chem. 1976, 40, 1891. (46) Halliwell, B.; Gutteridge Free Radicals in Medicine and Biology; Clarendon Press: Oxford, U.K., 1989. (47) Finley, K. T. The addition and substitution chemistry of quinones. In The Chemistry of Quinonoid Compounds; Patai, S., Ed.; Wiley: New York, 1988; pp 537-717. (48) Kutyrev, A. A.; Moskva, V. V. Russ. Chem. Rev. 1991, 60, 72-88. (49) Saleh, F. Y.; Theriot, L. J.; Amani, S. K.; Kim, I. Electron-spin resonance of fulvic and humic acids from the Suwannee River. In Humic substances in the Suwannee River, Georgia: interactions properties, and proposed structures; Averett, R. C., Leenheer, J. A., McKnight, D. M., Thorn, K. A., Eds.; U.S. Geological Survey: Denver, CO, 1989; pp 119-134. (50) Steelink, C. Review of ESR spectroscopy of humics. In NMR of Humic Substances and Coal; Wershaw, R. L., Mikita, M. A., Eds.; Lewis: Chelsea, MI, 1987; pp 47-72, (51) Phillippe, P. A. E.; Albrecht, P. Geochim. Cosmochim. Acta 1998, 62, 265-271. (52) Perlinger, J. A.; Buschmann, J.; Angst, W.; Schwarzenbach, R. P. Environ. Sci. Technol. 1998, 32, 2431-2437. (53) Perlinger, J. A.; Venkatapathy, R.; Harrison, J. F. J. Phys. Chem. A 2000, 104, 2752-2763. (54) Stumm, W.; Morgan, J. J. Aquatic Chemistry; Wiley-Interscience: New York, 1996. (55) Parris, G. E. Environ. Sci. Technol. 1980, 14, 1099-1106. (56) Baughman, G. L.; Weber, E. J. Environ. Sci. Technol. 1994, 28, 267-276. (57) Weber, E. J.; Adams, R. L. Environ. Sci. Technol. 1995, 29, 11631170. (58) Weber, E. J.; Spidle, D. L.; Thorn, K. A. Environ. Sci. Technol. 1996, 30, 2755-2763. (59) Thorn, K. A.; Pettigrew, P. J.; Goldenberg, W. S.; Weber, E. J. Environ. Sci. Technol. 1996, 30, 2764-2775.

Received for review July 9, 2001. Revised manuscript received February 14, 2002. Accepted March 20, 2002. ES015602C

VOL. 36, NO. 12, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2669