Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2013 Sep 17.
Published in final edited form as: Inorg Chem. 2012 Aug 30;51(18):9883–9892. doi: 10.1021/ic301324u

Formation, Spectroscopic Characterization, and Solution Stability of an [Fe4S4]2+ Cluster Derived from a β-Cyclodextrin Dithiolate

Wayne Lo a, Ping Zhang b, Chang-Chun Ling b, Shaw Huang a, R H Holm a,*
PMCID: PMC3483649  NIHMSID: NIHMS404668  PMID: 22934734

Abstract

The formation and solution properties, including stability in mixed aqueous-Me2SO media, have been investigated for an [Fe4S4]2+ cluster derived from a β-cyclodextrin (CD) dithiolate. Clusters of the type [Fe4S4(SAr)4]2− (Ar = Ph, C6H4-3-F) are generated in Me2SO by redox reactions of [Fe4S4(SEt)4]2− with 2 equiv of ArSSAr. An analogous reaction with the intramolecular disulfide of 6A,6D-(3-NHCOC6H4-1-SH)2-6A,6D-dideoxy-β-cyclodextrin (14), whose synthesis is described, affords a completely substituted cluster formulated as [Fe4S4{β-CD-(3-NHCOC6H4-1-S)2}2]2− (15). Ligand binding is indicated by a circular dichroism spectrum and also by UV/visible and isotropically shifted 1H NMR spectra and redox behavior convincingly similar to [Fe4S4(SPh)4]2−. One formulation of 15 is a single cluster to which two dithiolates are bound, each in bidentate coordination. There being no proven precedent for this binding mode, we show that the cluster [Fe4S4(S2-m-xyl)2]2− is a single cubane whose m-xylyldithiolate ligands are bound in a bidentate arrangement. This same structure type was proposed for a cluster formulated as [Fe4S4{β-CD-(1,3-SC6H4-S)2}2]2− (16; Kuroda et al., J. Am. Chem. Soc. 1988, 110, 4049–4050) and reported to be water-stable. Clusters 15 and 16 are derived from a similar ligands differing only in the spacer group between the thiolate binding site and the CD platform. In our search for clusters stable in aqueous or organic-aqueous mixed solvents that are potential candidates for reconstitution of scaffold proteins implicated in cluster biogenesis, 15 is the most stable cluster we have thus far encountered under anaerobic conditions in the absence of added ligand.

INTRODUCTION

We have described elsewhere1 the potential use of iron-sulfur clusters in treating human diseases arising from impaired cluster biogenesis that disrupts cellular iron homeostasis and leads to mitochondrial failure. Friedreich's ataxia is the most prominent of these diseases25 An untested approach involves delivery of intact iron-sulfur clusters to organelles for the purpose of reconstituting scaffold proteins involved in cluster biogenesis as part of a process that restores mitochondrial function.4 The clusters required for implementation of this approach must be soluble and stable in water or mixed organic-aqueous media and, ideally, largely impervious to destructive reactions at physiological dioxygen concentrations. Given that active mitochondrial aconitase and cytosolic aconitase IRP1 both contains [Fe4S4] clusters,6,7 the problem devolves to a search for suitable clusters of this type.

Within the large family of known [Fe4S4(SR)4]2− clusters (R = alkyl, aryl),8 nearly all of which have been isolated as quaternary ammonium salts, very few have been prepared that are water-soluble. This property is enhanced by hydrophilic substituents such as R = CH2CH2OH,1,9 CH2CH(OH)Me,10 and CH2CH2CO2.11 For stability in water, such clusters require excess ligand to diminish the rate of solvolytic decomposition.12 Certain dendritically encapsulated clusters are described as soluble in water and Me2SO/water media; stability information was not reported.13

A recent approach in this laboratory has utilized clusters derived from doubly deprotonated α-cyclodextrindithiol, 1 in Figure 1.1 The hydrophilic nature of cyclodextrins (CDs), which contain 6–8 d-glucopyranoside units linked in a toroidal topology,14 together with [Fe4S4]2+ core shielding by a large ligand, are expected to promote aqueous solubility and stability. This ligand and the related α-cyclodextrin monodithiolate are generated in solution by base hydrolysis of the corresponding thioesters. Cluster binding occurs by ligand substitution of a preformed cluster, for example [Fe4S4Cl4]2−, in a polar solvent such as Me2SO, and was demonstrated by uv-visible and circular dichroism spectra and by 1H NMR signals paramagnetically shifted into the downfield range (10–16 ppm) diagnostic of SCH2 protons in clusters with the core oxidation state [Fe4S4]2+.1 These clusters were formulated as [Fe4S4{α-CD-CH2S}4]2− and [Fe4S4{α-CD-(CH2S)2}2]2− to emphasize complete substitution by thiolate, but formation of oligomeric species (more than one [Fe4S4]2+ unit per molecule) could not be eliminated. The clusters showed modestly improved stability compared to, e.g., [Fe4S4(SCH2CH2OH)4]2−, in aqueous Me2SO solvent mixtures with no added ligand..

Figure 1.

Figure 1

Schematic structures of cyclodextrin dithiols: 6A,6D-(CH2 SH)2-6A,6Ddideoxy-α-cyclodextrin (1), 6A,6D-(3-SC6H4-1-SH)2-6A,6D-dideoxy-β-cyclodextrin (2), 6A,6D-(3-NHC(O)C6H4-1-SH)2-6A,6D-dideoxy-β-cyclodextrin (3). Abbreviations of doubly deprotonated forms: α-CD-(CH2S)2, β-CD-(1,3-SC6H4S)2, β-CD-(3-NHC(O)C6H4-1-S)2. CD = cyclodextrin.

The investigation of α-CD thiolate clusters was motivated in part by the work of Kuroda et al.15, who prepared the β-CD bis(3-thiophenyl-1-thiol) 2 (Figure 1) and the related monothiol. They proposed the formation of [Fe4S4{β-CD-(1,3-SC6H4S)}4]2− and [Fe4S4{β-CD-(1,3-SC6H4S)2}2]2− by reaction of [Fe4S4(SBut)4]2− with, apparently, four and two equiv of monothiol and dithiol, respectively. The UV-visible spectrum and 2-/3-redox potential of the dithiol reaction product are entirely consistent with formation of a cluster [Fe4S4(SAr)4]2−. However, the formulation [Fe4S4{β-CD-(1,3-SC6H4S)2}2]2− requires that the dithiolate function as a bidentate chelating ligand to the same cluster core, a rare, if not unprecedented, behavior in [Fe4S4] cluster chemistry. The molecular weight measured by light scattering was consistent with this formulation. It was also reported that this cluster was remarkably stable in aqueous phosphate buffer (pH 7.0) with no added ligand, displaying a half-life of greater than 120 hours. In the context of water-soluble clusters, the structural and stability features of the cluster derived from 2 have engaged our attention. Our results and conclusions for a related anaerobic [Fe4S4]2+ cluster system based on the β-CD bis(3-carboxamidobenzene-1-thiol) 3 are described here.

EXPERIMENTAL SECTION

Preparation of Compounds

(Et4N)2[Fe4S4(SR)4] (R = Et, Ph)16 and (Et4N)2[Fe4S4(S2-m-xyl)2]17 (S2-m-xyl = m-xylyl-α,α'-dithiolate(2-)) were prepared as previously described. The method of synthesis of the desired ligand in oxidized form (14, β-cyclodextrin disulfide--an oxidation product of dithiol 3) is set out in Scheme 1.

Scheme 1.

Scheme 1

Scheme for the synthesis of β-cyclodextrin disulfide 14 via intermediates 48, 13, and 14. Abbreviations: DCC = dicyclohexylcarbodiimide, EDC = N-ethyl-N'-(3-dimethylaminopropyl)carbodiimide, NH3 · H2O = conc. ammonia, Piv = pivaloyl.

2A,2B,2C,2D,2E,2F,2G,3A,3B,3C,3D,3E,3F,3G,6B,6C,6E,6F,6G-Nonadeca-O-acetyl-6A,6D-diazido-6A,6D-dideoxy-β-cyclodextrin (7)

To a solution of β-cyclodextrin (4, 10.0 g, 8.8 mmol) in anhydrous pyridine (100 mL) was added dropwise a solution of 4,4'-biphenyldisulfonyl chloride18 (3.09 g, 8.8 mmol) in anhydrous pyridine (30 mL), and the reaction mixture was stirred at 50°C for 2 h. Acetic anhydride (40 mL) was added and the reaction was continued overnight to cause conversion of 5 to 6. The mixture was concentrated under reduced pressure; the residue was dissolved in ethyl acetate (250 mL) and sequentially washed with aqueous HCl (2 N, 2 × 100 mL), saturated NaHCO3 (2 × 100 mL), and 10% brine (1 × 100 mL). After being dried with anhydrous Na2SO4, the organic solution was evaporated to afford a mixture containing the capped disulfonate 6. The residue was dissolved in DMF (100 mL), and NaN3 (5.7 g, 88 mmol) was added. After heating at 70 °C overnight, the solution was concentrated under reduced pressure; the residue was extracted with ethyl acetate (250 mL), the solution was washed with saturated brine (1 × 100 mL), dried over anhydrous Na2SO4, and evaporated. The crude mixture was purified by column chromatography on silica gel using a mixture of 35% ethyl acetate – hexane as the eluent to afford the diazide 7 (5.42 g, 31%) as a white solid. [α]D20: +71° (c 0.82, MeOH). 1H NMR (400 MHz, CDCl3): δ5.39 – 5.19 (m, 7H, 7 × H-3), 5.18 – 5.10 (m, 4H, 4 × H-1), 5.08 (d, J = 3.8 Hz, 1H, H-1), 5.04 (d, J = 3.5 Hz, 1H, H-1), 5.03 (d, J = 3.7 Hz, 1H, H-1), 4.88 – 4.74 (m, 7H, 7 × H-2), 4.66 – 4.51 (m, 5H, 5 × H-6a_CHaHbOAc), 4.34 – 4.00 (m, 12H, 7 × H-5 + 5 × H-6b_CHaHbOAc), 3.84 – 3.62 (m, 11H, 7 × H-4 + 2 × H-6a_CHaHbN3 + 2 × H-6b_CHaHbN3), 2.18 – 2.02 (m, 57H, 19 × OAc). 13C NMR (100 MHz, CDCl3) δ 170.86, 170.81, 170.73, 170.68, 170.56, 170.56, 170.52, 170.48, 170.43, 170.35, 170.27, 169.35 (CO), 97.11, 96.89 (× 2), 96.80 (× 2), 96.72, 96.41 (C-1), 77.59, 77.23 (× 2), 76.74, 76.61, 76.54, 76.22 (C-4), 71.30, 71.24, 71.13 (× 2), 71.05, 70.80, 70.66, 70.62, 70.54, 70.45, 70.24 (× 2), 70.17 (× 2), 70.11, 69.91, 69.66, 69.63, 69.57 (× 2), 69.34 (C-2, C-3, C-5), 62.71, 62.65, 62.61 (× 2), 62.46 C-6_CHaHbOAc), 50.82, 50.68 (C-6_CHaHbN3, 20.89-20.65 (19 × OAc). High-resolution MS (ESI): calcd m/z for C80H106N6O52Na (M+Na)+: 2005.57268; found: 2005.56980.

6A,6D-Diazido-6A,6D-dideoxy-β-cyclodextrin (8)

A solution of the per-O-acetylated diazide 7, (1.0 g, 0.50 mmol) in methanol (10 mL), water (2.0 mL) and concentrated ammonia (2.0 mL) was heated to 50 °C for 24 h, and then concentrated under reduced pressure. The residue was purified by reverse-phase chromatography on C18 using a gradient of water-methanol to afford the fully deacetylated diazide (8) which was freeze-dried (0.573 g, 96%). [α]D20: +55° (c 0.45, MeOH). 1H NMR (400 MHz, D2O) δ 5.14 - 5.08 (m, 7H, 7 × H-1), 4.08 – 3.80 (m, 27H, 7 × H-3 + 7 × H-5 + 5 × H-6a_CHaHbOH + 5 × H-6b_CHaHbOH + 2 × H-6a_CHaHbN3), 3.75 - 3.55 (m, 16H, 7 × H-2 + 7 × H-4 + 2 × H-6b_CHaHbN3). 13C NMR (100 MHz, D2O) : δ 101.81 (× 3), 101.78 (× 2), 101.56 (× 2), 81.98 (× 2), 81.23 (× 2), 81.17, 81.16, 81.10 (C-4), 72.99, 72.94, 72.76, 71.96, 71.89, 71.70, 70.46 (C-3, C-5, C-2), 60.32, 60.26 (× 2), 51.00 (× 2) (C-6). High resoluton MS (ESI): calcd m/z for C42H68N6O33Na (M+Na)+: 1207.37195; found: 1207.37022. This compound has been briefly described previously.19,20

6A,6D-Diamino-6A,6D-dideoxy-β-cyclodextrin acetate (9)

A solution of the diazide (8, 200 mg, 0.169 mmol) and triphenylphosphine (177 mg, 0.675 mmol) in pyridine (8.0 mL) and water (2.0 mL) was heated to 50 °C overnight. The solvent was removed under reduced pressure, and coevaporated with water to remove the residual amount of pyridine. The residue was re-dissolved in H2O (10 mL), neutralized with a few drops of acetic acid. The aqueous solution was washed with dichloromethane (3 × 5 mL), and freeze-dried to afford the previously known 6A,6D-diamino derivative,21, isolated as the acetate salt (9, 189 mg, ~96% yield). 1H NMR (400 MHz, D2O): δ 5.02 (m, 3H, 3 × H-1), 4.98 (m, 4H, 4 × H-1), 3.99 – 3.69 (m, 24H), 3.65 – 3.38 (m, 14H), 3.32 (dd, J = 13.6, 2.7 Hz, 2H, 2 × H-6a_CHaHbNH3+), 3.07 (dd, J = 13.7, 7.6 Hz, 2H, 2 × H-6a_CHaHbNH3+), 1.82 (s, 6H, 2 × Ac). 13C NMR (100 MHz, D2O) δ 101.80 (× 3), 101.77 (× 3), 101.36 (C-1), 82.74, 82.74 (× 2), 81.17 (× 3), 81.09, 80.72 (C-4), 73.03, 72.91, 72.58, 72.01, 71.91, 71.79, 68.97 (C-2, 3, 5), 60.41, 60.36, 60.26, 60.20 (× 2), 40.42 (× 2), 23.26 (OAc). High-resolution MS (ESI): calcd m/z for C42H73N2O33 (M+H)+: 1133.40956; found: 1133.40617.

3-(Pivaloylthio)benzoic acid (11)

A solution of 3,3'-dithiodibenzoic acid (10, 3.0 g, 9.8 mmol) and triphenylphosphine (2.83 g, 10.78 mmol) in THF (30 mL) and water (1.5 mL) was stirred for 1 h at room temperature, and evaporated to dryness under reduced pressure. The residue was re-dissolved in a mixture of THF (30 mL) and anhydrous pyridine (15 mL), and treated with pivaloyl chloride (7.23 mL, 58.7 mmol) for 30 min. To hydrolyze the formed anhydride, water (15 mL) was added and the mixture was stirred for 1 h at 50°C. The reaction mixture was concentrated under reduced pressure and the residue was partitioned using a mixture of ethyl acetate (100 mL) and 2 N HCl (100 mL); the organic solution was separated, dried over anhydrous Na2SO4, and concentrated. The desired acid (3.45 g, 74%) was obtained by column chromatography on silica gel using a mixture of 0.5% methanol-dichloromethane as the eluent. 1H NMR (400 MHz, CDCl3): δ 11.97 (br, 1H, COOH), 8.21 – 8.11 (m, 2H, H-2 + H-6), 7.65 (ddd, J = 1.6, 1.6, 7.8 Hz, 1H, H-4), 7.50 (dd, J = 8.0, 8.0 Hz, 1H, H-5), 1.35 (s, 9H, (CH3)3CCO). 13C NMR (100 MHz, CDCl3) δ 171.31 (CO), 147.99, 140.34, 136.59, 130.80, 130.22, 129.17 (Ar), 47.10 ((CH3)3CCO), 27.35 ((CH3)3CCO). High-resolution MS (ESI): calcd m/z for C12H14O3SNa (M+Na)+: 261.05618; found: 261.05558.

2,5-Dioxopyrrolidin-1-yl 3-(pivaloylthio)benzoate (12)

The protected acid 11 (565 mg, 2.37 mmol) and N-hydroxysuccinimide (273 mg, 2.37 mmol) were dissolved in ethyl acetate (10 mL) with heating, and N,N-dicyclohexylcarbodiimide (689 mg, 3.28 mmol) was added to the hot solution. After stirring for 3 h at room temperature, the precipitate was filtered off and washed with ethyl acetate. The combined solution was evaporated to dryness. The residue was recrystallized from isopropanol to afford the desired ester 12 (620 mg, 78%). 1H NMR (400 MHz, CDCl3): δ 8.14 (m, 2H, H-2 + H-6), 7.71 – 7.66 (ddd, J = 1.3, 1.7, 7.8 Hz, 1H, H-4), 7.56 (dd, J = 7.8, 7.8 Hz, 1H, H-5), 2.87 (br s, 4H, NHS), 1.31 (s, 9H, (CH3)3CCO). 13C NMR (100 MHz, CDCl3): δ 203.27 ((CH3)3CCOS), 169.13 (ArCO), 161.20 (CO_NHS), 141.47, 136.63, 130.98, 129.87, 129.51, 126.12 (Ar), 47.11((CH3)3CCO), 27.30 (NHS), 25.64 ((CH3)3CCO). High-resolution MS (ESI): calcd m/z for C16H17NO5SNa (M+Na)+: 358.07196; found: 358.07147.

β-Cyclodextrin Disulfide (14)

Diazide 8 (200 mg, 0.169 mmol) was reacted with triphenylphosphine (177 mg, 0.675 mmol) in a mixture of pyridine (8.0 mL) and water (2.0 mL) at 50 °C overnight. The activated N-hydroxysuccinimide ester 12 (170 mg, 0.506 mmol) was added, and the mixture was maintained at 50°C for 24 h. The reaction mixture (presumably containing the intermediate diamide 13) was concentrated by evaporation. Methanol (3.0 mL), water (1.0 mL), and concentrated ammonia (1.0 mL) were added to remove the pivalate thioester. After heating at 50 °C overnight, the solution was concentrated under reduced pressure, and the residue was purified by column chromatography, first on silica gel using a mixture of isopropanol–water–aqueous ammonia (7/1/1 v/v/v) as the eluent, followed by reverse phase chromatography on a C18 Sep-Pak cartridge to provide disulfide 14, which was freeze-dried (102 mg, 43%). [α]D20: +38.5° (c 0.19, MeOH). 1H NMR (600 MHz, CD3OD): δ 8.08 (t, J = 1.7 Hz, 1H, H-2’_Ar), 7.85 (t, J = 1.7 Hz, 1H, H-2’_Ar), 7.76 (m, 2H, 2 × H-6’_Ar), 7.72 (ddd, J = <1, <1, 7.6 Hz, 1H, H-4’_Ar), 7.69 (ddd, J = <1, 1.7, 7.8 Hz, 1H, H-4’_Ar), 7.51 (dd, J = 7.8, 7.8 Hz, 2H, 2 × H-5’_Ar), 5.05 (d, J = 3.8 Hz, 1H, H-1), 5.04 (d, J = 3.6 Hz, 1H, H-1), 5.03 (d, J = 3.8 Hz, 1H), 5.01 – 4.99 (m, 4H, 4 × H-1), 4.42 (dd, J = <1, 12.9 Hz, 1H, H-6a_CHaHbNH), 4.32 (dd, J = <1, 12.4 Hz, 1H, H-6a_CHaHbNH), 4.10 (ddd, J = 2.4, 2.4, 9.9 Hz, 1H, H-5), 3.98 (dd, J = 12.4, 2.8 Hz, 1H, H-6a_CHaHbOH), 3.95 – 3.49 (m, 33H, 7 × H-2 + 7 × H-3 + 6 × H-5 + 4 × H-6a_CHaHbOH + 4 × H-6b_CHaHbNH + 5 × H-4), 3.47 (dd, J = 1.5, 12.2 Hz, 1H, H-6b_CHaHbOH), 3.28 (dd, J = 9.3, 9.3 Hz, 1H, H-4), 3.24 (dd, J = 9.3, 9.3 Hz, 1H, H-4), 3.08 (dd, J = 14.0, 10.4 Hz, 2H, 2 × H-6b_CHaHbNH). 13C NMR (150 MHz, CD3OD) δ 168.24 (CO), 167.94 (CO), 137.74, 137.57, 135.54, 135.02, 132.15, 129.25, 129.15, 129.05, 126.65, 126.32, 125.80, 124.99 (Ar), 102.94 (× 2), 102.40, 102.31, 102.22, 102.17, 102.14 (C-1), 85.17, 84.99, 81.25, 81.11, 81.05, 80.91, 80.64 (C-4), 73.33 (× 3), 73.27, 73.24, 73.18, 73.07, 72.93, 72.81, 72.65 (× 2), 72.55, 72.51, 72.38, 71.98, 71.92, 71.89, 71.74 (× 2), 71.17, 70.56 (C-2, C-3, C-5), 60.17, 60.11, 59.81, 59.79, 59.67 (C-6_CHaHbOH), 41.85, 41.58 (C-6_CHaHbNH). High-resolution MS (ESI): calcd m/z for C56H78N2O35S2Na (M+Na)+: 1425.37187; found: 1425.36793.

X-ray Structure Determination

The compound (Et4N)2[Fe4S4(S2-m-xyl)2] was crystallized from DMF/ether. Owing to the very small size of crystals (ca. 0.02 × 0.02 × 0.001 mm), data were acquired at Argonne National Laboratory utilizing the Advanced Photon Source. Data were collected at 95 K with a high-brilliance synchroton X-ray source and a Bruker-Apex II CCD detector. The unit cell was determined with Bruker SMART software. The crystal did not show any significant decay during data collection (ca. 1 h). Data were corrected for Lorentz and polarization effects (Bruker SAINT V7.46A) and absorption corrections were applied (SADABS). The space group was assigned based on symmetry analysis and systematic absences (XPREP). The structure was solved by direct methods and refined by least-squares on F2 (SHELXL-9722). All non-hydrogen atoms were located in difference-Fourier maps and refined anisotropically. Hydrogen atoms were introduced at calculated positions using a riding model. Crystallographic data are collected in Table 1.23

Table 1.

Crystallographic Data for (Et4N)2[Fe4S4(S2-m-xyl)2]

formula C32H56Fe4N2S8
formula weight M 948.70
crystal system monoclinic
space group Cc
a (Å) 18.2095 (6)
b (Å) 11.2828 (4)
c (Å) 21.8275 (7)
α = γ (deg) 90
β (deg) 111.866 (1)
V (Å3) 4161.9 (2)
dcalc (g/cm3) 1.514
Z 4
GOF 1.256
Rb (wR2)c 0.0403 (0.142)
a

Synchrotron radiation source (λ = 0.39364 Å), T = 95 K.

b

R1 = E | |Fo| – |Fc| |/E |Fo|.

c

wR2 = {E[w(Fo2Fc2)2/E(Fo2)2]}1/2

Other Physical Measurements

All measurements were made under anaerobic conditions. Absorption spectra were measured on a Varian Cary 50 Bio spectrophotometer and circular dichroism spectra (at 20°C) on a Jasco J-715 spectropolarimeter. NMR spectra were recorded on a Varian 500 or a Varian 600 MHz spectrometer. High-resolution mass spectra (ESI-TOF) were obtained on an Agilent 6520-Q-TOF spectrometer. Electrochemical measurements were made with a Bioanalytical Systems potentiostat/galvanostat in Me2SO solutions using a glassy carbon working electrode, 0.1 M (Bu4N)(PF6) supporting electrolyte, and an SCE reference electrode. Mössbauer spectra were determined with a constant acceleration spectrometer; isomer shifts are referenced to iron metal at room temperature.

RESULTS AND DISCUSSION

Ligand Synthesis

Initially we wished to reexamine the results of Kuroda et al.15 involving Fe4S4 cluster binding to doubly deprotonated 2. With reference to Scheme 1, the 6A,6D-capped disulfonate 5 was first prepared by a published procedure.18 Unfortunately, several attempts to synthesize the previously reported 6A,6D-diarylthiol 2 from 5 by nucleophilic attack with excess benzene-1,3-dithiol monoanion were unsuccessful. Consequently, the alternate synthetic target 3 was designed which employs the carboxamide linkage to connect a 6A,6D-diamino-substituted β-CD to the arylthio groups.

We first sought the 6A,6D-diamino scaffold 9 using 5 as a key intermediate. This compound was obtained from β-CD 4 and 4,4'-biphenyldisulfonyl chloride in pyridine, and was isolated in pure form (15–20%) by reverse-phase chromatography on C18 silica gel. We found that if direct acylation was carried out after sulfonylation, the presumed intermediate 6 could be directly reacted with NaN3 in DMF (70°C) to afford the diazide 7. This compound was isolated by normal phase chromatography on silica gel (31%); the acetyl protecting groups were smoothly removed in methanol-water by concentrated ammonia (50°C) to provide 8 (96%). The azido groups were simultaneously reduced with Ph3P in pyridine-water (50°C) to obtain the diamine 921 which was isolated in the acetate salt form. Unfortunately, the EDC-mediated coupling of 9 with the diacid 10 to give the desired 14 was unsuccessful despite numerous attempts. Given the possibility that the difficulty might originate with dual reactive sites in both 9 and 10, it was decided to reduce the disulfide bond of 10 to provide a carboxylic acid with one reactive group.

The disulfide linkage of 10 was reduced with Ph3P,24 and the intermediate thiol was directly protected with pivaloyl chloride to afford 11 (74%). This compound was converted to the activated N-hydroxysuccinimide ester 12 (78%). Attempts to functionalize 9 with either 11 (using EDC as a reagent) or 12 in the presence of base (Et3N, 4-N,N-dimethylaminopyridine) were unproductive. The ultimately successful method utilized reduction of the diazide 8 to the diamine followed by treatment in situ with ester 12 to generate S-pivaloyl protected 13 (not isolated) and deprotection with ammonia. Normal phase column chromatography followed by reverse phase chromatography on C18 silica gel yielded the 6A,6D-disubstituted β-CD disulfide 14 (43%), the oxidized form of dithiol 3 formed under the aerobic purification conditions.

The structure of 14 was confirmed by several methods, including 1D 1H and 13C as well as 2D COSY and HSQC NMR experiments.23 The two phenyl groups were found to be magnetically inequivalent by the appearance of eight aromatic proton and twelve aromatic carbon signals in CD3OD and (CD3)SO solutions. The 1H NMR spectrum in the downfield region in (CD3)2SO, displayed in Figure 2, reveals two well-resolved NH resonances, two p-H signals from different rings, three o-H signals (four are resolved CD3OD) and a m-H doublet of doublets. This spectral region is relevant to cluster formation. The cyclic voltammograms of two phenyldisulfides, one containing electron-withdrawing substituents, and 14 are presented in Figure 3. Irreversible reduction (2ArS → ArSSAr + 2e) and oxidation (reverse reaction) steps are observed, a behavior typical of compounds of this type in aprotic solvents.25,26 The 3-fluoro substitutents cause a positive shift of Epc, in agreement prior observations of 4-substituted compounds.27 The same redox pattern is observed for 14 at nearly the same potentials as for PhSSPh, confirming the presence of an S-S bond. The structure of 14 was further confirmed by high resolution mass spectrometry.

Figure 2.

Figure 2

Schematic structure of the β-cyclodextrin disulfide and designation of aromatic protons, and the 1H NMR spectrum (600 MHz) of the downfield region in (CD3)2SO. Signal assignments are indicated; solv = DMF.

Figure 3.

Figure 3

Cyclic voltammograms (100 mV/s) in Me2SO of diphenyl disulfide (blue, upper) and bis(3-fluorophenyl)disulfide (red, upper) and β-cyclodextrin disulfide (lower). Peak potentials are indicated.

Thiolate Cluster Ligation

Variation of terminal thiolate ligation in [Fe4S4(SR)4]2− clusters is conventionally accomplished by displacement of halide (e.g., in [Fe4S4Cl4]2−) with thiolate or by substitution of bound thiolate by reaction with another, usually more acidic, thiol.8,28. To avoid conversion of disulfide 14 to dithiol 3 required for either method, we have utilized redox reaction 1 which finds a near-precedent in the formation of [Fe4S4(SePh)4]2− from [Fe4S4(SBut)4]2− and PhSeSePh.29 Reactions with both disulfides have been monitored by 1H NMR as seen in Figure 4. At n = 1 and 2 equiv, multiple downfield signals (12.5–13.5 ppm) due to methylene protons in the species

[Fe4S4(SEt)4]2+2 ArSSAr[Fe4S4(SAr)4]2+2 EtSSEt (1)

[Fe4S4(SEt)4-n(SAr)n]2− are evident and the characteristically contact-shifted aromatic ring proton signals appear at 5.0–8.3 ppm. At n = 2.1–2.2 equiv and up to 36 h reaction time reaction 1 is complete for both disulfides.

Figure 4.

Figure 4

1H NMR spectra (600 MHz) in the downfield region for the reaction between 14 mM [Fe4S4(SEt)4]2− and n equiv of (a) diphenyl disulfide and (b) bis(3-fluorophenyl)disulfide in (CD3)2SO at ~ 25°C. Signal assignments of bound arylthiolate ligands are indicated; solv = DMF.

In the following sections, CD-derived clusters are designated as indicated. See also Figure 1. These formulations are not intended to imply molecularity.

[Fe4S4{β-CD-(1,3-NHCOC6H4S)2}2]2− 15
[Fe4S4{β-CD-(1,3-SC6H4S)2}2]2− 1615
[Fe4S4{α-CD-(CH2S)2}2]2− 171
[Fe4S4{α-CD-(CH2S)}4]2− 181

With the feasibility of reaction 1 demonstrated, reaction 2 was conducted in order to generate cluster 15 in Me2SO solution. The reactions were followed by 1H NMR with variable equiv of 14. The spectra in Figure 5 are similar to those of reaction 1; one mixed ligand cluster (13.05 ppm) was detected. As n equiv of the disulfide increase, the three ring

[Fe4S4(SEt)4]2-+2β-CD-(1,3-NHCOC6H4S)2[Fe4S4{β-CD-(1,3-NHCOC6H4S)2}2]2+2 EtSSEt (2)

proton resonances emerge, and at n = 2.0 and 16 h the limiting spectrum, ascribed to cluster 15, is reached. Resonances of m-H and p-H were correlated by a COSY experiment. Shifts of m-H are close to those of [Fe4S4(SAr)4]2−, thus assuring cluster binding. Shifts of o-H and p-H indicate smaller contact contributions to their chemical shifts than in the arylthiolate clusters. This arises from differences in spin delocalization possibly due to phenyl group orientations controlled by ligand structure and by the carboxamide substituent. The spectrum of 16 would be an interesting comparison but it was not reported.15

Figure 5.

Figure 5

(a) 1H NMR spectra (600 MHz) in the downfield region for the reaction between 2.3 mM [Fe4S4(SEt)4]2− and n equiv of β-cyclodextrin disulfide at various times in (CD3)2SO at ~25°C. (b) COSY spectrum of the n = 2.0 equiv product at 16 h reaction time showing the correlation between Hm and Hp.

Because a crystalline reaction product has not been isolated from reaction 2, additional evidence for cluster binding to reduced 14 has been sought from other observations. Relevant absorption spectra in Me2SO are collected in Figure 6. The spectrum of the product of reaction 2 has a prominent band at 440 nm which is the counterpart of the 457 nm feature of [Fe4S4(SPh)4]2−. These spectra distinguish arylthiolate from alkylthiolate clusters such as [Fe4S4(SEt)4]2−max 297, 419 nm). Treatment of the reaction product with ca. 25 equiv of benzenethiol in reaction 3 generates a spectrum identical with that of authentic [Fe4S4(SPh)4]2− and corresponds to 98% of the initial cluster concentration in reaction 2.

[Fe4S4{β-CD-(1,3-NHCOC6H4S)2}2]2+4 PhSH[Fe4S4(SPh)4]2+2β-CD-(1,3-NHCOC6H4SH)2 (3)

We conclude that (barring an unlikely core dissembly/assembly event) reaction 2 proceeds with retention of the [Fe4S4]2+ core of the starting cluster. Note also the circular dichroism spectra in Figure 7 where the effect is induced by one and two equiv of the β-CD disulfide. Cluster 18 with four α-CD monothiolate ligands also shows a strong circular dichroism signal.1 Disulfide 14, as CD itself, does not exhibit a circular dichroism spectrum in the same region. The circular dichroism spectrum constitutes further evidence for the occurrence of reaction 2, whose chromophoric product is necessarily chiral.

Figure 6.

Figure 6

Absorption spectra in Me2SO of [Fe4S4(SEt)4]2− (red), the reaction product of [Fe4S4(SEt)4]2− and β-CD disulfide (reaction 2, blue), the reaction product plus 25 equiv of benzenethiol (orange), and β-CD disulfide (green).

Figure 7.

Figure 7

Circular dichroism spectra of the reaction product between 0.51 mM [Fe4S4(SEt)4]2− and 1.0 equiv (blue) and 2.0 equiv (red) of β-cyclodextrin disulfide in Me2SO. The disulfide (green, 1.4 mM) did not produce an effect at 270–770 nm.

Lastly, the voltammetry of the 2-/3- redox couples of [Fe4S4(SPh)4]2− and the product of reaction 2 are compared in Figure 8. Note that both are chemically reversible (ipc/ipa ≈ 1) and differ by only 30 mV, a result that requires arylthiolate coordination in the reaction product. Alkylthiolate clusters such as [Fe4S4(SEt)4]2− (−1.33 V) exhibit 2-/3- couple as substantially more negative potentials.30

Figure 8.

Figure 8

Cyclic voltammograms(100 mV/s) of [Fe4S4(SEt)4]2− (upper) and the reaction product between 1.2 mM [Fe4S4(SEt)4]2− and 2.1 equiv of β-cyclodextrin disulfide after 16 hr (reaction 2, lower). The inset diagrams show the redox potentials obtained by differential pulse voltammetry (pulse width 50 ms, pulse period 200 ms). Voltammograms were recorded in Me2SO solutions at ~25°C; peak potentials are indicated.

Structural Considerations

Results from four physical techniques not only establish a reaction between the β-CD disulfide and [Fe4S4(SEt)4]2− in Me2SO solution but require arylthiolate binding to the iron sites, rather than any other form of interaction. These results do not demonstrate the molecularity of the reaction product nor is the property known in the solid state for we have not yet achieved diffraction-quality crystals. The dianion reduction product of 14 can bind to the [Fe4S4]2+ core as a bridging or terminal ligand in the stoichiometry of 15. The formation μ2-SR bridges would lead to oligomeric species, of which there is one documented example.31 Binding of one thiolate at an iron site would afford a single cluster with two bidentate ligands in chelate-like ligation, as proposed for 16 apparently based on a molecular weight from light scattering.15 This type of binding is found with the Fe protein of nitrogenase, which consists a cluster ligated by two cysteinate residues from each of two α-subunits.32,33 We have sought an example of this binding mode in a synthetic cluster closer to the problem at hand.

Attention has turned to the previously reported cluster [Fe4S4(S2-m-xyl)2]2−, formulated on the basis of absorption and 1H NMR spectral features and redox behavior but lacking structure proof.17,30 In the original work, the m-xylyldithiolate ligand was selected because its bite distance appeared appropriate to span two iron sites. We verified that the cluster possesses the [Fe4S4]2+ oxidation state from its Mössbauer parameters (δ = 0.41 mm/s, ΔEQ = 1.05 mm/s at 90 K) whose values are typical of that state.8 As is evident in Figure 9, cluster has idealized S4 symmetry. The two aromatic rings are held above opposite Fe2S2 core faces at a dihedral angle of 20.74°; the least-mean-squares planes of these faces are disposed at 15.00°. The shortest ring carbon distance to a face is 3.14 Å. The core unit has distorted tetrahedral iron sites and a compressed tetragonal arrangement with 4 short (2.236(5) Å) + 8 long (2.313(8) Å) Fe-S bonds, a frequently observed distortion from Td symmetry.8 Mean bond lengths are very close to those of [Fe4S4(SCH2Ph)4]2−.16 The S≅≅≅S distance (6.25 Å) between terminal ligands is smaller than the values in the latter cluster (6.30, 6.49 Å) because of constraints of the bidentate ligand structure.

Figure 9.

Figure 9

The structure of [Fe4S4(S2-m-xyl)2]2− with 50% probability ellipsoids and the atom numbering scheme. The 4 short (2.253(1)–2.268(1) Å) and 8 long (2.297(1)–2.318(1) Å) Fe-S bond distances define the indicated ranges. The mean of Fe-Fe distances (2.698(1)–2.756(1) Å) is 2.73(3) Å.

The structure of [Fe4S4(S2-m-xyl)2]2− provides the first demonstration of the unidentate terminal binding mode by a bidentate dithiolate. At the least, it offers a precedent for a similar binding arrangement in 15 and well as 16 and 17. We note that the S≅≅≅S separations in 15 are expected to be similar to those of [Fe4S4(S2-m-xyl)2]2− and comparable to the cavity diameter of β-CD (6.0–6.5 Å).34 While this size range does not strongly constrain thiolate sulfur positions, benzenethiolate substituents at 6A,6D positions on the rim of the structure can accommodate the S≅≅≅S distances of a synthetic cluster.

Solution Stability

In the examination of stability, reaction 2 was conducted in Me2SO solution after which there were added varying volumes of aqueous phosphate buffer to give solutions with the indicated solvent compositions (v/v). The spectrum in aqueous buffer alone was obtained by removing Me2SO and dissolving the residue in buffer. Spectra are assembled in Figure 10 for anaerobic solutions containing 80-0% Me2SO and no excess disulfide 14 and measured at times up to 150 hours after the initial spectra. For each system, spectra were recorded at 4–6 intermediate times as well, but for clarity only the intial and final spectra and those at 60 min and 12 hours are shown. Spectra in the different solvent media were measured in the concentration ranges given in the figure. Stability is qualitatively assessed by departure from the spectrum in pure Me2SO (Figure 6).

Figure 10.

Figure 10

Absorption spectra of the product of reaction 2 in specified Me2SO/aqueous buffer solutions (v/v) in which the aqueous component is 3.55 mM phosphate buffer (pH 7.5–7.6). Spectra were recorded under anaerobic conditions at 0 to 120 or 150 hr after reaction at the following concentrations: 80% Me2SO/20% aq., 0.27–0.41 mM; 60% Me2SO/40% aq., 0.20–0.31 mM; 40% Me2SO/60% aq., 0.14–0.21 mM; 20% Me2SO/80% aq.; 100% aq., 0.39 mM. Band maxima or shoulders are indicated.

The spectrum of cluster 15 in 80% Me2SO is virtually identical to that in the pure solvent. The cluster is completely stable in this medium at 72 hr and at 150 hr and evidences only ca. 6% decrease in intensity of the 440 nm band at 150 hr. At 60% Me2SO, the cluster is fully stable at 60 min and slowly decays at 440 nm to 80% of the original intensity after 150 hr. The spectra at 40% and 20% Me2SO are further progressions but with significant increases in absorbance in the 470–770 nm region. Increased visible absorbances are observed for α-CD thiolate clusters 17 and 18 and also for [Fe4S4(SCH2CH2OH)4]2− and [Fe4S4(SCH2CH2CO2)4}6− at .50% Me2SO. However, their appearance is a certain sign of (partial) cluster degradation. CD spectra at 270–770 nm (not shown) confirm stability at 60–100% Me2SO for up to 12 hr, and reveal substantial changes in ellipticity and several new features at higher aqueous buffer content after 1 hr. Comparison of spectra at 80–20% Me2SO indicate a somewhat improved stability of 15 in these media at 60 min to 12 hr compared to clusters 17 and 18, with the former somewhat more stable than the latter.1

Clusters 15 and 16, derived from β-CD ligands 3 and 2, respectively, are potentially capable of chelate-type ligation. Both possess essentially identical visible absorption features having maxima at 440 nm (Me2SO) and 441 nm (DMF)15 with µm ≈ 17,500, suggestive of related structures. Conceivably, the cluster structures in their entirety differ primarily by the connector (-S-, -NHCO-) from the 6-CH2 group of rings A and D to the 3-C6H4S binding sites. We have not attempted to determine the stability half-life of 15 in aqueous buffer as has been reported for 16 (120 hours) on the basis of an absorbance decrease at 420–470 nm. We do note that the absorbance of 15 at 440 nm is unchanged at 1 hour and decays by 15% over 120 hours with an increase in absorbance at longer wave lengths. The latter aspect, not just the absorbance at or near the band maximum, should be monitored to detect cluster decay. The results when compared to those for [Fe4S4(SCH2CH2OH)4]2− and [Fe4S4(SCH2CH2CO2)4}6− disclose a moderate stability enhancement owing to the presence of the large, apparently protective α- and β-CD ligand platforms. Thus we concur with Kuroda et al.15 that an unusual property of cyclodextrin-based clusters is their aqueous stability. In our hands, 15 is the most stable cluster produced thus far.

Supplementary Material

1_si_001

ACKNOWLEDGMENTS

This research was supported by Freidreich's Ataxia Research Alliance and NIH Grant GM 28856. We thank Dr. T. Rouault for useful discussion. C.-C. L. and P.Z. thank the Alberta Glycomics Centre, the Government of Alberta and the University of Calgary for the financial support. We also thank Dr. Shao-Liang Zheng at Harvard University and Dr. Yuk-Sheng Chen at ChemMetCARS, Advanced Photon Source, for their assistance with X-ray data collection. ChemMetCARS Section 15 is principally supported by the National Science Foundation/Department of Defense under Grant NSF/CHE-0822838. Use of the Advanced Photon Source is supported by the U. S. Department of Defense, Office of Science, Office of Basic Energy Sciences, under contract no. DE-AC02-06CH11357.

REFERENCES

  • 1.Lo W, Scott TA, Zhang P, Ling C-C, Holm RH. J. Inorg. Biochem. 2011;105:497–508. doi: 10.1016/j.jinorgbio.2010.12.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Lill R, Muhlenhoff U. Annu. Rev. Biochem. 2008;77:669–700. doi: 10.1146/annurev.biochem.76.052705.162653. [DOI] [PubMed] [Google Scholar]
  • 3.Rouault TA, Tong WH. Trends in Genetics. 2008;24:398–407. doi: 10.1016/j.tig.2008.05.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Ye H, Rouault TA. Biochemistry. 2010;49:4945–4956. doi: 10.1021/bi1004798. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Rouault TA. Disease Models Mech. 2012;5:155–164. doi: 10.1242/dmm.009019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Robbins AH, Stout CD. Proc. Natl. Acad. Sci. USA. 1989;86:3639–3643. doi: 10.1073/pnas.86.10.3639. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Dupuy J, Volbeda A, Carpentier P, Darnault C, Moulis J-M, Fontecilla-Camps JC. Structure. 2006;14:129–139. doi: 10.1016/j.str.2005.09.009. [DOI] [PubMed] [Google Scholar]
  • 8.Rao PV, Holm RH. Chem. Rev. 2004;104:527–559. doi: 10.1021/cr020615+. [DOI] [PubMed] [Google Scholar]
  • 9.Hill CL, Renaud J, Holm RH, Mortenson LE. J. Am. Chem. Soc. 1977;99:2549–2557. doi: 10.1021/ja00450a024. [DOI] [PubMed] [Google Scholar]
  • 10.Davies SC, Evans DJ, Henderson RA, Hughes DL, Longhurst S. J. Chem. Soc. Dalton Trans. 2002:3470–3477. [Google Scholar]
  • 11.Job RC, Bruice TC. Proc. Natl. Acad. Sci. USA. 1975;72:2478–2482. doi: 10.1073/pnas.72.7.2478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Bruice TC, Maskiewicz R, Job R. Proc. Natl. Acad. Sci. USA. 1975;72:231–234. doi: 10.1073/pnas.72.1.231. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Sharma AK, Kim N, Cameron CS, Lyndon M, Gorman CB. Inorg. Chem. 2010;49:5072–5079. doi: 10.1021/ic1002447. [DOI] [PubMed] [Google Scholar]
  • 14.Harata K. Chem. Rev. 1998;98:1803–1827. doi: 10.1021/cr9700134. [DOI] [PubMed] [Google Scholar]
  • 15.Kuroda Y, Sasaki Y, Shiroiwa Y, Tabushi I. J. Am. Chem. Soc. 1988;110:4049–4050. [Google Scholar]
  • 16.Averill BA, Herskovitz T, Holm RH, Ibers JA. J. Am. Chem. Soc. 1973;95:3523–3534. doi: 10.1021/ja00792a013. [DOI] [PubMed] [Google Scholar]
  • 17.Holm RH, Phillips WD, Averill BA, Mayerle JJ, Herskovitz T. J. Am. Chem. Soc. 1974;96:2109–2117. doi: 10.1021/ja00814a020. [DOI] [PubMed] [Google Scholar]
  • 18.Tabushi I, Yamamura K, Nabeshima T. J. Am. Chem. Soc. 1984;106:5267–5270. [Google Scholar]
  • 19.Fujita K, Yamamura H, Imoto T. J. Org. Chem. 1985;50:4393–4395. [Google Scholar]
  • 20.Cotner ES, Smith PJ. J. Org. Chem. 1998;63:1737–1739. [Google Scholar]
  • 21.Bonomo RP, Pedotti S, Vecchio G, Rizzarelli E. Inorg. Chem. 1996;35:6873–6877. doi: 10.1021/ic951447p. [DOI] [PubMed] [Google Scholar]
  • 22.Sheldrick GM. Acta Crystallogr. 2009;A64:112–122. doi: 10.1107/S0108767307043930. [DOI] [PubMed] [Google Scholar]
  • 23.See paragraph at the end of this article for Supporting Information available.
  • 24.Humphrey RE, Hawkins JM. Anal. Chem. 1964;36:1812–1814. [Google Scholar]
  • 25.Persson B. J. Electroanal. Chem. 1978;86:313–323. [Google Scholar]
  • 26.Christensen TB, Daasbjerg K. Acta Chem. Scand. 1997;51:307–317. [Google Scholar]
  • 27.Daasbjerg K, Jensen H, Benassi R, Taddei F, Antonello S, Gennaro A, Maran F. J. Am. Chem. Soc. 1999;121:1750–1751. [Google Scholar]
  • 28.Holm RH. In: Bio-coordination Chemistry. Que L Jr, Tolman WA, editors. Oxford: Elsevier; 2004. pp. 61–90. [Google Scholar]
  • 29.Que L, Jr, Bobrik MA, Ibers JA, Holm RH. J. Am. Chem. Soc. 1974;96:4168–4178. doi: 10.1021/ja00820a018. [DOI] [PubMed] [Google Scholar]
  • 30.DePamphilis BV, Averill BA, Herskovitz T, Que L, Jr, Holm RH. J. Am. Chem. Soc. 1974;96:4159–4167. doi: 10.1021/ja00820a017. [DOI] [PubMed] [Google Scholar]
  • 31.Rao PV, Bhaduri S, Jiang J, Hong D, Holm RH. J. Am. Chem. Soc. 2005;127:1933–1945. doi: 10.1021/ja040222n. [DOI] [PubMed] [Google Scholar]
  • 32.Georgiadis MM, Komiya H, Chakrabarti P, Woo D, Kornuc JJ, Rees DC. Science. 1992;257:1653–1659. doi: 10.1126/science.1529353. [DOI] [PubMed] [Google Scholar]
  • 33.Schlessman JL, Woo D, Joshua-Tor L, Howard JB, Rees DC. J. Mol. Biol. 1998;280:669–685. doi: 10.1006/jmbi.1998.1898. [DOI] [PubMed] [Google Scholar]
  • 34.Saenger W, Jacob J, Gessler K, Steiner T, Hoffman D, Sanbe H, Koisumi K, Smith SM, Takaha T. Chem. Rev. 1998;98:1787–1802. doi: 10.1021/cr9700181. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1_si_001

RESOURCES