Reversible oxidative addition and reductive elimination of fluorinated disulfides at gold(I) thiolate complexes: A new ligand exchange mechanism

Article abstract of DOI:10.1021/ja805266r

Reaction of HAuCl₄·3H₂O with excess HSAr (Ar = C₆F₅ or C₆F₄H) in ethanol, followed by addition of [Et₄N]Cl, produced [Et₄N][Au(SAr) ₄] (Ar = C₆F₅ (1a) or C₆F ₄H (1b)) as red crystalline solids in high yield. These complexes are rare examples of homoleptic gold(III) thiolate complexes. The crystal structures 1 show square planar geometry at the gold center with elongated Au-S bonds. Both complexes undergo reversible reductive elimination/oxidative addition processes in solution via thermal and photochemical pathways. Equilibrium constant and photostationary state measurements indicate that the relative importance of the two pathways depends on the nature of the aromatic groups. The metal-containing reductive elimination products, [Et ₄N][Au(SAr)₂] (Ar = C₆F₅ (2a) or C₆F₄H (2b)), were confirmed by both independent synthesis and crystallographic characterization. Cross-reactions between either 1 or 2 and various disulfides led to ligand exchange via an addition-elimination process, a previously unknown reaction pathway for ligand exchange at gold(I) centers.

Full text of DOI:10.1021/ja805266r

Published on Web 10/01/2008
Reversible Oxidative Addition and Reductive Elimination of
Fluorinated Disulfides at Gold(I) Thiolate Complexes: A New
Ligand Exchange Mechanism
Robert E. Bachman,*,§ Sheri A. Bodolosky-Bettis,‡ Chelsea J. Pyle,§ and
Margaret Anne Gray§
Departments of Chemistry, The UniVersity of the South, 735 UniVersity AVenue, Sewanee,
Tennessee 37383, and Georgetown UniVersity, Box 571227, Washington, D.C. 20057-1227
Received July 8, 2008; E-mail: rbachman@sewanee.edu
Abstract: Reaction of HAuCl4 ·3H2O with excess HSAr (Ar ) C6F5 or C6F4H) in ethanol, followed by addition
of [Et4N]Cl, produced [Et4N][Au(SAr)4] (Ar ) C6F5 (1a) or C6F4H (1b)) as red crystalline solids in high yield.
These complexes are rare examples of homoleptic gold(III) thiolate complexes. The crystal structures 1
show square planar geometry at the gold center with elongated Au-S bonds. Both complexes undergo
reversible reductive elimination/oxidative addition processes in solution via thermal and photochemical
pathways. Equilibrium constant and photostationary state measurements indicate that the relative importance
of the two pathways depends on the nature of the aromatic groups. The metal-containing reductive
elimination products, [Et4N][Au(SAr)2] (Ar ) C6F5 (2a) or C6F4H (2b)), were confirmed by both independent
synthesis and crystallographic characterization. Cross-reactions between either 1 or 2 and various disulfides
led to ligand exchange via an addition-elimination process, a previously unknown reaction pathway for
ligand exchange at gold(I) centers.
antiarthritic treatments for almost 80 years.5 More recently, there
have been reports suggesting that gold complexes containing
Introduction
As can be seen in the widely varied occurrences of gold
thiolate species in the literature, sulfur-based ligands play a
central role in the chemistry of gold, particularly in its reduced
states of gold(I) and gold(0).1 For example, thiols are commonly
used to passivate and/or modify gold nanoparticles2 and to form
robust self-assembled monolayer assemblies on gold surfaces.3
Thiols are also a key component in the formulation of gilding
inks.4 In essentially all of these applications, ligand substitution
reactions are critical for selective patterning of surfaces or for
tailoring other physical properties of the final system.2-4 Despite
the widespread use of such reactions, little is known about the
mechanism of the substitution process.
gold(III) as well as gold(I) display antitumor and/or antimalarial
behavior.6 Interestingly, despite their long clinical use and
intensive research by numerous groups, the precise mode of
action of gold-based drugs in ViVo remains elusive and likely
involves diverse targets including transcription factors and
cellular redox modulators.7 One generally accepted paradigm
is that many of the administered gold(I) thiolate complexes are
rapidly undergoing ligand exchange processes with proteins,
such as serum albumin, that possess available cysteine residues
via a simple nucleophilic displacement mechanism (eq 1).
Gold thiolates also play a central role in the biological
chemistry of gold. Gold(I) thiolate species have been used as
While there is substantial evidence to support this mechanism,
there are some intriguing reports in the literature that hint that
§ The University of the South.
‡ Georgetown University.
(1) Shaw, C. F., III. The Biochemistry of Gold. In Gold: Progress in
Chemistry, Biochemistry, and Technology; Schmidbaur, H., Ed.; Wiley:
Chichester, 1999; pp 259-308.
(6) (a) Arizti, M. P.; Garcia-Orad, A.; Sommer, F.; Silvestro, L.; Massiot,
P.; Chevallier, P.; Gutierrez-Zorrilla, J. M.; Colacio, E.; Martinez de
Pancorbo, M.; Tapiero, H. Anticancer Res. 1991, 11, 625–628. (b)
Navarro, M.; Pe´rez, H.; Sa´nchez-Delgado, R. A. J. Med. Chem. 1997,
40, 1937–1939. (c) Pillarsetty, N.; Katti, K. K.; Hoffman, T. J.; Volkert,
W. A.; Katti, K. V.; Kamei, H.; Koide, T. J. Med. Chem. 2003, 46,
1130–1132. (d) Messori, L.; Abbate, F.; Marcon, G.; Orioli, P.;
Fontani, M.; Mini, E.; Mazzei, T.; Carotti, S.; O’Connell, T.; Zanello,
P. J. Med. Chem. 2000, 43, 3541–3548.
(2) (a) Dolamic, I.; Gautier, C.; Boudon, J.; Shalkevich, N.; Burgi, T. J.
Phys. Chem. C 2008, 112, 5816–5824. (b) Tao, A. R.; Habas, S.; Yang,
P. D. Small 2008, 4, 310–325.
(3) (a) Heimel, G.; Romaner, L.; Bredas, J. L.; Zojer, E. Langmuir 2008,
24, 474–482. (b) Ulman, A. Acc. Chem. Res. 2001, 34, 855–863. (c)
Schlenoff, J. B.; Li, M.; Ly, H. J. Am. Chem. Soc. 1995, 117, 12528–
12536.
(7) (a) Urig, S.; Fritz-Wolf, K.; Reau, R.; Herold-Mende, C.; Toth, K.;
Daoid-Charvet, E.; Becker, K. Angew. Chem., Int. Ed. 2006, 45, 1881–
1886. (b) Gunatilleke, S. S.; Barrios, A. M. J. Med. Chem. 2006, 49,
3933–3937. (c) Larabee, J. L.; Hocker, J. R.; Hanas, J. S. Chem. Res.
Toxicol. 2005, 18, 1943–1954. (d) Chircorian, A.; Barrios, A. M.
Bioorg. Med. Chem. Lett. 2004, 14, 5113–5116. (e) Stoyanov, J. V.;
Brown, N. L. J. Biol. Chem. 2003, 278, 1407–1410. (f) Handel, M. L.;
Watts, C. K. W.; DeFazio, A.; Day, R. O.; Sutherland, R. L. Proc.
Natl. Acad. Sci. U.S.A. 1995, 92, 4497–4501.
(4) Bishop, P. T. Gold Bull. 2002, 35, 89–98.
(5) (a) Forestier, J. J. Lab. Clin. Med. 1935, 20, 827–840. (b) Smith, W. E.;
Reglinski, J. Gold Drugs Used in the Treatment of Rheumatoid
Arthritis. In PerspectiVes on Bioinorganic Chemistry Vol. 1; Dilworth,
J. R., Hay, R. W., Nolan, K. B., Eds.; JAI Press Ltd.: Stamford, CT,
1991; pp 183-208. (c) Fricker, S. P. Gold Bull. 1996, 29, 53–59. (d)
Smith, W. E.; Reglinski, J. Metal-Based Drugs 1994, 1, 497–507. (e)
Shaw, C. F. Chem. ReV. 1999, 99, 2589–2600.
9
10.1021/ja805266r CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 14303–14310 14303

A R T I C L E S
Bachman et al.
this might be an incomplete picture. For example, Sadler and
co-workers reported finding the mixed Cys-STg (STg )
thioglucose) disulfide species in preparations of Auranofin and
bovine serum albumin.8 Shaw and co-workers have also
observed reactivity between cationic gold(I) complexes and
disulfide-protected albumin.9 They have also recently reported
the formation of the disulfide GSSG as a byproduct of the
glutathione reduction of [Au(CN)4]- via the heteroleptic gol-
d(III) complex [Au(CN)2(SG)2]- (eq 2).10
addition of thiols. However, in a pair of little-noticed papers,16
Muller et al. and Peach reported preliminary evidence suggesting
that highly electron-deficient thiols might be sufficiently less
capable of reducing gold(III) to allow for the isolation of
homoleptic gold(III) thiolate complexes. In this report, we build
on their initial observations by isolating and structurally
characterizing two examples of homoleptic gold(III) thiolate
complexes with highly fluorinated ligands. To our knowledge,
only one other homoleptic gold(III) thiolate complex has been
isolated to date. This work also provides the first unambiguous
example of the reductive elimination of disulfide from a gold(III)
center and, perhaps more significantly, the reverse process of
oxidative addition of a disulfide to a single gold(I) center. In
light of these findings, we propose a novel addition-elimination
mechanism for ligand substitution reactions at gold(I) centers,
which is likely to have relevance to the reactivity of gold(I)
complexes in a variety of settings, including biological ones.
[Au(CN)4]- +
G-SH
8 [Au(CN)2(SG)2]3- f [Au(CN)2]- + GSSG2-
-2HCN
(2)
Additionally, multiple groups have also demonstrated the
cleavage of disulfide linkages by gold(I) in the gas phase.11
Finally, Ashraf and Isab have shown that both (R3P)2Au+ and
[Au(I)STm]n (STm ) thiomalate) are capable of cleaving
disulfides or diselenides in solution, with the mixed species RSe-
STm observed in the latter case.12
Given this wide array of results, it is intriguing that Bruce
and co-workers have demonstrated that gold thiolate complexes,
including currently used gold drugs, are capable of mediating
thiol-disulfide exchange reactions.13 In their work, they have
hypothesized that gold(III) species may form as reaction
intermediates, which suggests that gold(III) species may play a
significant role in the chemistry of gold thiolates generally.
Numerous other groups have also found evidence for the
formation of gold(III) species in ViVo.14 For example, De Wall
et al. have reported that gold(III) is highly effective at altering
the conformation of MHC proteins, which are known to play
an important role in immune function.15
Experimental Procedures
Materials. All experiments were carried out in air unless
otherwise noted. Solvents were reagent grade or better and used as
received from the vendor. HAuCl4 ·3H2O (Strem Chemicals),
C6F5SH (Aldrich), 2,3,5,6-C6F4HSH (Aldrich; referred to hereafter
as HSC6F4H-p), Et4NCl·XH2O (Aldrich), bentonite clay (Acros),
and trifluoroacetic acid (Acros) were purchased from the noted
suppliers and used as received. Me2SAuCl was prepared by direct
reaction of Me2S and HAuCl4 · 3H2O in ethanol. The disulfides,
RSSR (3), were prepared by the method described by Meshram.17
1
Physical Measurements. H, 19F, and 13C NMR spectra were
recorded in the solvent specified on a Varian Gemini spectrometer
1
operating at 300 (1H), 282 (19F), and 75.45 MHz (13C). The H
and 13C chemical shifts were internally referenced to TMS (δ ) 0
ppm), while the 19F chemical shifts were externally referenced to
neat trifluoroacetic acid (δ ) 0 ppm). UV-visible spectra for
routine characterization were recorded on a Hewlett-Packard 8453
diode array spectrophotometer in the solvents indicated. Irradiation
experiments to determine the photostationary states were performed
using an Osram XBO 450 W lamp connected to a PTI monochro-
mator, model 101/102. Slit widths of entrance and exit were both
set to 5 mm, corresponding to 20 nm of resolution. Sample
temperature was maintained by use of a water filter. Following
irradiation, UV-visible spectra were recorded on a Cary 300 Bio
UV-vis dual-beam scanning spectrophotometer. Electrochemical
measurements were performed using a CH Instruments model 630A
potentiostat system with a glassy carbon working electrode, a
platinum auxiliary electrode, and a nonaqeuous (CH3CN) Ag/
AgNO3 reference electrode (590 mV vs NHE). The measurements
were carried out at room temperature in acetonitrile solutions
containing ca. 0.1 M [(n-Bu)4N][PF6] as the supporting electrolyte.
All processes were electrochemically irreversible, and as such, the
values reported are the peak currents, adjusted so as to be relative
to NHE, for the process in question.
An inherent difficulty in studying the behavior of gold(III)
thiolate species, as well as the possible role of gold(III) as an
intermediate in the reactions of gold(I), is the facile and rapid
reduction of gold(III) to gold(I) that is typically observed upon
(8) Orla, M. N. D.; Sadler, P. J.; Tucker, A. J. Am. Chem. Soc. 1992,
114, 1118–1120.
(9) (a) Isab, A. A.; Hormann, A. L.; Hill, D. T.; Griswold, D. E.;
DiMartino, M. J.; Shaw, C. F., III Inorg. Chem. 1989, 28, 1321–1326.
(b) Roberts, J. R.; Xiao, J.; Schliesman, B.; Parsons, D. J.; Shaw, C. F.,
III Inorg. Chem. 1996, 35, 424–433.
(10) Yangyuoru, P. M.; Webb, J. W.; Shaw, C. F., III J. Inorg. Biochem.
2008, 102, 584–593.
(11) (a) Lioe, H.; Duan, M.; O’Hair, R. A. J. Rapid. Commun. Mass
Spectrom. 2007, 16, 2727–2733. (b) Gunawardena, H. P.; O’Hair,
R. A. J.; McLuckey, S. A. J. Proteome Res. 2006, 9, 2087–2092.
(12) (a) Ahmad, S.; Isab, A. A.; Wazeer, M. I. M. Inorg. React. Mech.
2002, 1-2, 95–102. (b) Ashraf, W.; Isab, A. A. J. Coord. Chem. 2004,
57, 337–346.
Determination of Thermal Equilibrium Constants. “Normal”
equilibrium constantssthose determined under ambient lightings
were determined by preparing solutions with known concentrations
of 1, or 1 and the corresponding disulfide 3, in acetone-d6. The
mixtures were monitored by 19F NMR until the integrals of the
signals achieved constant values. Alternatively, the concentration
of 1 in solution could be monitored by UV-vis spectrometry at
470 nm until a constant value was observed. To determine “dark”
equilibrium constants (i.e., in the absence of ambient light), solutions
of known concentrations of 1 were prepared in acetone-d6 in
photographic dark room (red light) conditions. Aliquots (1 mL) of
(13) (a) Mohamed, A. A.; Chen, J.; Bruce, A. E.; Bruce, M. R. M. Inorg.
Chem. 2003, 42, 2203–2205. (b) Mohamed, A. A.; Abdou, H. E.; Chen,
J.; Bruce, A. E.; Bruce, M. R. M. Comm. Inorg. Chem. 2002, 23,
321–334. (c) Chen, J.; Jiang, T.; Wei, G.; Mohamed, A. A.;
Homrighausen, C.; Bauer, J. A. K.; Bruce, A. E.; Bruce, M. R. M.
J. Am. Chem. Soc. 1999, 121, 9225–9226. (d) DiLorenzo, M.; Ganesh,
S.; Tadayon, L.; Chen, J.; Bruce, M. R. M.; Bruce, A. E. Metal-Based
Drugs 1999, 6, 247–253.
(14) (a) Shaw, C. F., III; Schraa, S.; Gleichmann, E.; Grover, Y. P.;
Dunemann, L; Jagarlamundi, A. Metal-Based Drugs 1994, 1, 351–
362. (b) Takhashi, K.; Griem, P.; Goebel, C.; Gonzalez, J.; Gleich-
mann, E. Metal-Based Drugs 1994, 1, 483–496. (c) Verwilghen, J.;
Kingsley, G. H.; Gambling, J.; Panayi, G. S. Arthritis Rheum. 1992,
35, 1413–1418.
(15) De Wall, S. L.; Painter, C.; Stone, J. D.; Bandaranayake, R.; Wiley,
D. C.; Mitchison, T. J.; Stern, L. J.; DeDecker, B. S. Nature Chem.
Biol. 2006, 2, 197–201.
(16) (a) Muller, M.; Clark, R. J. H.; Nyholm, R. S. Transition Met. Chem.
1978, 3, 369–377. (b) Peach, M. E. Can. J. Chem. 1968, 46, 2699.
(17) Meshram, H. M. OPPI Briefs 1993, 25 (2), 232–233.
9
14304 J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008

Ligand Exchange by Addition-Elimination at Au(I) Centers
A R T I C L E S
the solutions were transferred to amber-colored NMR tubes and
wrapped in foil. The 19F NMR spectrum of each sample was taken
once daily, and the integrals of the signals were monitored until a
constant value was established.
Determination of Photostationary States. Solutions of known
concentrations of 1 were prepared in acetone under red light
conditions. A small aliquot (∼3 mL) of the solution was transferred
to a 1 cm quartz cuvette and capped. The sample was placed 1 cm
from the monochromator opening and irradiated for 2 h at room
temperature with light of either 470 or 350 nm. After 2 h, the
absorbance at 470 nm was monitored periodically (typically every
15 min) until it remained constant. The determination of this
photostationary state was repeated three times for each sample at
both 470 and 350 nm.
X-ray Crystallography. All X-ray data were collected on a
Bruker-AXS SMART CCD diffractometer at -100 °C using
methods previously described elsewhere.18 The essential experi-
mental details for all structures are provided in the Supporting
Information. Final unit cell parameters were calculated by least-
squares refinement of 4546 (1a), 6868 (2a), 5958 (1b·EtOH), and
1220 (2b) strong reflections taken from the entire data sets. The
data were corrected for Lorentz and polarization effects, and an
absorption correction was applied on the basis of equivalent
reflection measurements using Blessing’s method as incorporated
into the program SADABS.19 The structures were solved using
direct methods and refined against F2 by full-matrix least-squares
methods using the programs incorporated into the SHELXTL/PC
suite and XSEED.20 All non-hydrogen atoms were refined aniso-
tropically, while hydrogen atoms were added in calculated positions
using a standard riding model. The solvent molecule located in the
lattice of 1b·EtOH is severely disordered around an inversion
center. Given the difficulty in establishing reliable carbon and
oxygen atom positions, no attempt was made to model the hydrogen
atom positions of the solvent molecule. The less than ideal model
of this solvent molecule does not affect the chemically significant
structural information, as the bonding parameters for 1b remain
unchanged within error when the solvent positions are completely
omitted from the model.
water and small aliquots of ether (3 × 2 mL). The solid was dried
1
in air. Yield: 619 mg (70%). H NMR (acetone-d6, ppm): 7.789
3
4
(tt, 1H, JH-F ) 10.2 Hz, JH-F ) 7.7 Hz), 3.473 (t, 2H, JH-H
)
8.55 Hz), 1.843 (m, 2H), 1.448 (m, 2H), 0.989 (t, 3H, JH-H ) 7.20
Hz). 19F{1H} NMR (acetone-d6, ppm): -52.79 (m, 2F), -63.76
(m, 2F). 13C NMR (acetone-d6, ppm): 147.8 (o-C, JC-F ) 241
1
1
2
Hz), 146.7 (m-C, JC-F ) 234 Hz), 125.2 (i-C, JC-F ) 22 Hz),
100.0 (p-C, 1JC-H ) 170 Hz, 2JC-F ) 24 Hz), 52.5 (CH2, 1JC-H
)
)
1
144 Hz), 7.2 (CH3, JC-H ) 129 Hz). UV-vis (acetone): λmax
473 (ꢀ ∼4100 cm-1 M-1), 355 nm (sh). Ered ) -279 mV vs NHE.
X-ray quality crystals were grown by slow evaporation of ethanolic
solutions of 1b in air at room temperature. This salt crystallizes as
an ethanol solvate, which slowly loses solvent when removed from
the mother liquor. Addition of Ph4PBr rather than Et4NCl · H2O
yields the corresponding [Ph4P]+ salt. Elemental analysis, calcd
for C48H24F16S4PAu: C, 45.71; H, 1.91. Found: C, 45.48; H, 1.91.
Preparation of [Et4N][Au(SC6F5)2] (2a). Method 1. F5C6SH
(0.068 mL, 0.510 mmol) was added to a solution of KOH (0.029
g, 0.510 mmol) in methanol (1 mL). This solution was then added
to a suspension of Me2SAuCl (0.075 g, 0.255 mM) in H2O (∼5
mL). Additional methanol (∼10 mL) was added to dissolve any
remaining solids. The solution was filtered, and Et4NCl·H2O (0.042
g, 0.255 mmol) was added to the filtrate to precipitate a white
crystalline solid. The solid was washed with water (∼10 mL) and
ether (∼1 mL). Yield: 0.0847 g (46%). 1H NMR (acetone-d6, ppm):
3.50 (q, 2H, JH-C ) 6.90 Hz), 1.40 (tt, 3H, JH-C ) 6.90, 1.80 Hz).
19F{1H} NMR (acetone-d6, ppm): -55.62 (dd, 2F, JF-F ) 26.97,
5.79 Hz), -88.54 (m, 3H). 13C NMR (acetone-d6, ppm): 147.8 (o-
1
1
C, JC-F ) 240 Hz), 137.9 (m-C, JC-F ) 232 Hz,), 119.0 (i-C,
2JC-F ) 23 Hz), 138.4 (p-C, 1JC-F ) 247 Hz), 53.0 (CH2, 1JC-H
)
)
1
144 Hz), 7.7 (CH3, JC-H ) 129 Hz). UV-vis (acetone): λmax
282 nm (ꢀ ∼17 000 cm-1 M-1). Eox ) 1.35V vs NHE. Elemental
analysis, calcd for C20H20F10S2NAu: C, 33.11; H, 2.78; N, 1.93.
Found: C, 33.25; H, 2.85; N, 1.81.
Method 2. A solution of 1a in acetone was allowed to stand
until it changed from red to essentially colorless, typically at least
1 day. After the solution had turned colorless, ether was layered
on top, and the solution was placed in a freezer (-20 °C) for
approximately 1 week, after which time a small amount of colorless
X-ray quality needle-shaped crystals of 2a formed. All spectral
features of these crystals are identical to those reported in method
1 above. Attempts to isolate and separate the gold complexes and
disulfide remaining in solution led to intractable mixtures of the
gold(III) salt and the disulfide.
Preparation of [Et4N][Au(SC6F5)4] (1a). F5C6SH (0.45 mL, 3.4
mmol) was added dropwise to a solution of HAuCl4 ·3H2O (290
mg, 0.74 mmol) in ethanol (15 mL). Immediately after completion
of the thiol addition, 320 mg of Et4NCl·XH2O in water (5 mL)
was added to the resulting red solution to produce a red microc-
rystalline precipitate that was isolated by filtration, washed with
1
water, and dried in air. Yield: 579 mg (70%). H NMR (acetone-
Preparation of [Et4N][Au(SC6F4H)2] (2b). Method 1. Me2SAuCl
(0.077 g, 0.262 mmol) was suspended in methanol (10 mL), and a
solution of p-HF4C6SH (0.062 mL, 0.523 mmol) in basic methanol
(0.029 g, 0.523 mmol of KOH, in 5 mL of methanol) was added
dropwise over a period of 2-3 min. After being stirred for 1 h, the
colorless, transparent solution was filtered using a Whatman
microfiber glass filter to remove small amounts of metallic gold
and/or other insoluble impurities. Et4NCl·XH2O (0.087 g. 0.523
mmol) was added to the filtrate, and the solvent was removed in
vacuo to yield a cream-colored solid. The solids were washed with
water and then dried in air. Yield: 0.129 g (71%). 1H NMR
(acetone-d6, ppm): 7.77 (tt, 1H, 3JH-F ) 10.2 Hz, 4JH-F ) 7.7 Hz),
3.51 (q, 2H, JH-H ) 7.50 Hz), 1.37 (tt, 3H, JH-H ) 7.50, 1.80 Hz).
19F{1H} NMR (acetone-d6, ppm): -57.35 (m, 2F), -66.41 (m, 2F).
13C NMR (acetone-d6, ppm): 147.6 (o-C, 1JC-F ) 235 Hz), 146.7
d6, ppm): 3.509 (q, 2H, J ) 7.20 Hz), 1.391 (tt, 3H, J ) 7.20, 1.80
Hz). 19F{1H} NMR (acetone-d6, ppm): -51.925 (dd, 2F, JF-F
)
25.98, 7.63 Hz), -80.254 (t, 1F, JF-F ) 21.18 Hz), -87.129 (m,
1
2F). 13C NMR (acetone-d6, ppm): 148.3 (o-C, JC-F ) 240 Hz),
1
2
146.7 (m-C, JC-F ) 250 Hz), 119.3 (i-C, JC-F ) 23 Hz), 140.7
1
1
(p-C, JC-F ) 256 Hz), 52.5 (CH2, JC-H ) 144 Hz), 7.6 (CH3,
1JC-H ) 129 Hz). UV-vis (acetone): λmax ) 470 nm (ꢀ ∼3600
cm-1 M-1), 353 nm (ꢀ ∼29 000 cm-1 M-1). Ered ) -302 mV vs
NHE. Elemental analysis, calcd for C32H20F20S4NAu: C, 34.19; H,
1.78; N, 1.25. Found: C, 34.46; H, 1.76; N, 1.41. X-ray quality
crystals were grown by slow evaporation of a methanolic solution
of 1a in air at room temperature.
Preparation of [Et4N][Au(SC6F4H-p)4] (1b). p-HF4C6SH (0.50
mL, 3.85 mmol) was added dropwise to a solution of HAuCl4 ·3H2O
(330 mg, 0.84 mmol) in ethanol (15 mL). Et4NCl·XH2O (350 mg)
dissolved in water (5 mL) was immediately added to the resulting
transparent red solution, resulting in a slight turbidity. The mixture
was then placed in a freezer (-20 °C) for several days, during which
time a red crystalline solid formed. The product was isolated by
decanting off the remaining solvent and washing the solids with
1
2
(m-C, JC-F ) 244 Hz), 125.7 (i-C, JC-F ) 23 Hz), 100.1 (p-C,
2
1
1JC-H ) 170 Hz, JC-F ) 24 Hz), 53.0 (CH2, JC-H ) 144 Hz),
7.7 (CH3, 1JC-H ) 129 Hz). UV-vis (acetone): λmax ) 298 nm (ꢀ
∼ 29 000 cm-1 M-1). Eox ) 1.59 V vs NHE. Elemental analysis,
calcd for C20H20F10S2NAu: C, 34.84; H, 3.22; N, 2.03. Found: C,
35.10; H, 3.05; N, 2.32.
Method 2. A solution of 1b in ethanol was allowed to stand
until it changed from red to pale orange or colorless, typically at
least 1 day. Once the system had equilibrated, ether was layered
on top, producing a small amount of X-ray quality crystals of 2b
(18) Bachman, R. E.; Andretta, D. F. Inorg. Chem. 1998, 37, 5657–5663.
(19) Blessing, R. H. Acta Crystallogr. 1995, A51, 33–38.
(20) (a) SHELXTL-PC, Version 5.10; Bruker AXS: Madison, WI, 1998.
(b) Barbour, L. XSEED, 1999; http://www.x-seed.net/index.html.
9
J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008 14305

A R T I C L E S
Bachman et al.
Table 1. Average Equilibrium and Photostationary State Constants
for 1a
Scheme 1. Synthetic Routes for Preparing 1 and 2
1a / 2a + 3a
thermal equilibrium, dark
0.12b
thermal equilibrium, room light
photostationary state, 350 nm
photostationary state, 470 nm
0.13 (0.02)
0.16 (0.01)
0.19 (0.03)
1b / 2b + 3b
thermal equilibrium, dark
ND
thermal equilibrium, room light
photostationary state, 350 nm
photostationary state, 470 nm
1.09 (0.12)
0.62 (0.10)
1.30 (0.10)
a The standard deviation is provided in parentheses. b Experiment
only performed once due to extremely slow reaction kinetics.
Table 2. Reaction of 2 and [PhSAuSPh]- with Various Disulfides
[ArS-Au-SAr]-
RSSR
products
2a
+
+
+
3a
3b
PhSSPh
p-FC6H4SSC6H4F
p-MeC6H4SSC6H4Me
(AcMeCys)2
f
f
f
1a
[Au(SC6F5)2(SC6F4H)2]-
2a + RSSR
2a + RSSR
2a + RSSR
[Au(AcMeCys)]n + 3a
red solution from which [Et4N][Au(SAr)4] (1: a, Ar ) F5C6; b,
Ar ) p-HF4C6) (Scheme 1) can be isolated in high yield
(typically 70% or better) via the addition of [Et4N]Cl. Addition
of other organic cations, such as Ph4P+, results in the formation
of the corresponding salts. These complexes are soluble in most
moderately polar organic solvents and insoluble in water and
nonpolar organic solvents. As expected, given the large amount
of HCl byproduct generated in the reaction, the final solution
medium is quite acidic (pH e 1 by indicator paper). Interest-
ingly, 1a and 1b appear to be unreactive toward this high
concentration of acid, most likely due to the low basicity of
the electron-deficient thiolates. We have measured the pKa’s of
F5C6SH and p-HF4C6SH as 4.18 and 4.21, respectively;
moreover, it is reasonable to assume that these values are even
lower when the thiol is bound to a metal ion.21 For comparison,
the pKa of thiolphenol is 6.6, while that of aliphatic thiols is
approximately 8.0.22
While stable in the solid-state, 1a and 1b undergo a reductive
elimination to form their respective gold (I) complexes,
[Au(SAr)2]- (2), along with the corresponding disulfide, ArSSAr
(3). Since complexes 1 are red and complexes 2 are colorless,
the extent of this reaction can be qualitatively monitored by
simple visible inspection; quantitative measurements can be
made by UV-vis spectroscopy. Interestingly, even when a
solution appears colorless, all three species (1, 2, and 3) can
still be clearly identified in 19F NMR spectra. Furthermore,
evaporation of the solvent from such colorless mixtures results
in the quantitative re-formation of 1, by oxidative addition of 3
across the gold atom of 2, yielding a red solid (Scheme 1).
While 2 may not be isolated from the reductive elimination
reaction by solvent removal, these complexes can be isolated
by addition of a nonpolar solvent, such as ether, to the mixture
resulting from the reductive elimination process. Alternatively,
samples of 2 can be prepared directly via the reaction of a
suitable gold(I) precursor, such as Me2SAuCl, with a stoichio-
metric quantity of the desired thiolate anion (see Scheme 1).
2b
3a
3b
PhSSPh
p-FC6H4SSC6H4F
p-MeC6H4SSC6H4Me
(AcMeCys)2
[Au(SC6F5)2(SC6F4H)2]-
1b
2b + RSSR
2b + RSSR
2b + RSSR
[Au(AcMeCys)]n + 3b
[PhSAuSPh]-
3a
3b
2a + PhSSPh
2b + PhSSPh
as colorless needles. All spectral features of these crystals are
identical to those reported in method 1 above. Attempts to isolate
and separate the gold complexes and disulfide remaining in solution
led to intractable mixtures of the gold(III) salt and the disulfide.
Reaction of 2a or 2b with ArSSAr: Typical Procedure. A
small amount (0.03-0.04 mmol) of 2a or 2b was dissolved in
approximately 5 mL of methanol. A solution of an equimolar
amount of the disulfide dissolved in methanol (5 mL) was added
in one portion, and the mixture was allowed to stand for a
minimum of 30 min. The solvent was removed in Vacuo, and
the residue was taken up in acetone-d6 and immediately
characterized by NMR spectroscopy. The results of these
reactions are summarized in Table 2.
Reaction of 2a with N-Acetyl-L-cystine Methyl Ester
((AcMeCys)2). A small sample of 2a or 2b (15-25 mg) was
dissolved in approximately 5 mL of methanol. A solution containing
an excess (25-55 mg) of (AcMeCys)2 (AcMeCys ) N-acetyl-L-
cysteine methyl ester) in 5 mL of methanol was added in one
portion. The reaction was mixed and left to stand for 10 min, during
which time a white, highly luminescent precipitate of [Au(AcMe-
Cys)]n formed. The solid was isolated by filtration, washed with
methanol and ether, and dried in air. The filtrate was concentrated
in Vacuo to yield a residue of C6F5SSC6F5 or C6F4HSSC6F4H, which
was characterized by 19F NMR. 19F NMR (acetone-d6, ppm) for
reaction with 2a: -54.61 (dd, 2F), -71.21 (t, 1F), -82.91 (m, 2F).
Elemental analysis, calcd for C6H10AuNO3S: C, 19.39; H, 2.70.
Found: C, 19.44; H, 2.76.
Results
(21) (a) Bodolosky-Bettis, S. Gold Thiolates: Luminescent Properties and
Reductive-Elimination/Oxidative Addition Processes. Ph.D. Thesis,
Georgetown University, 2003. (b) Tachtman, M.; Markam, G. D.;
Glusker, J. P.; George, P.; Bock, C. W. Inorg. Chem. 2001, 40, 4230–
4241.
Synthesis. Unlike the results observed with most thiols, the
reaction of [AuCl4]- with either F5C6SH or p-HF4C6SH does
not lead immediately to the formation of oligomeric gold(I)
thiolate complexes typically seen in reactions of gold(III) salts
with thiols. Rather, the initial result is the formation of a bright
(22) (a) Jencks, W. P.; Salvesen, K. J. Am. Chem. Soc. 1971, 93, 4433–
4436. (b) Mathews, C. K.; van Holde, K. E.; Ahern, K. G. Biochem-
istry, 3rd ed.; Addison Wesley: San Fransciso, 2000; p 128.
9
14306 J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008

Ligand Exchange by Addition-Elimination at Au(I) Centers
A R T I C L E S
Figure 1. Structures of the anion portions of 1a (left) and 1b (right), including atom labeling scheme in the vicinity of the metal atom.
Figure 2. Structures of the anionic portions of 2a (left) and 2b (right), with the atom labeling scheme near the metal atom.
Reaction of 2 prepared in this manner with the appropriate
disulfide (3) regenerates the characteristic red color of the
gold(III) species, 1.
explanation for this difference is that the electron-deficient
nature of the thiolate ligands in 1 results in slightly longer,
and presumably weaker, metal-sulfur interactions. Alterna-
tively, it is possible that the longer bonds are caused by the
larger steric bulk of four fluorinated aromatic rings. The bond
distances and angles within the ligands, including the
aromatic C-C distances, are all within the expected ranges
and show no significant or systematic variations between
gold(I) and gold(III) or between degrees of fluoridation on
the ring (see Supporting Information for detailed bond
distances and angles).
Structure. Slow evaporation of methanolic or ethanolic
solutions of 1 provided X-ray quality crystals of the red four-
coordinate gold(III) complexes (Figure 1). In the case of 1b,
the salt crystallizes as an ethanol solvate. As expected for Au(III)
complexes, the gold atoms in 1 are situated in a square planar
coordination environment. The Au-S bond lengths in 1a and
1b are essentially identical (averages 1a, 2.352 Å; 1b, 2.349
Å). In both cases, they are also similar to those seen in the only
other reported Au(III) complex with four monodentate thiolate
ligands, [Ph4P][Au(Smetetraz)4]·0.5H2O (Smetetraz ) 1-meth-
yl-1,2,3,4-tetrazole-5-thiolate) (average 2.354 Å).23 Interestingly,
these distances are slightly longer than those observed in the
mixed ligand complex [Au2Cl4(µ-SPh)2] (average 2.334 Å), even
though in this latter case the thiolates are bridging ligands.24One
X-ray quality crystals of 2 were prepared by dissolving 1
in acetone, allowing the reductive elimination process to
proceed, and then isolated by layering the reaction mixture
with ether and cooling to -20 °C. Crystallographic analysis
of the resulting colorless needles confirmed the structures of
2 (Figure 2). As expected for two-coordinate gold(I) com-
plexes, the geometry of the gold atoms in 2 is essentially
linear (S-Au-S ) 177.2° and 180°).25 The Au-S distances
(average 2a, 2.2861 Å; 2b, 2.2852 Å) are slightly longer than
those observed in the corresponding perhydro complex,
[Ph4P][Au(SPh)2] (average 2.266 Å), but within the range
observed (2.276-2.292 Å) for gold(I) thiolate complexes
(23) Lang, E. S.; Dahmer, M.; Abram, U. Acta Crystallogr. 1999, C55,
854–856.
(24) Wang, S.; Fackler, J. P., Jr. Inorg. Chem. 1990, 29, 4404–4407.
(25) (a) Bartczak, T. J. Acta Crystallogr. 1985, C41, 865–869. (b) Nomiya,
K.; Noguchi, R.; Sakurai, T. Chem. Lett. 2000, 274–275. (c) Watase,
S.; Nakamoto, M.; Kitamura, T.; Kanehisa, N.; Kai, Y.; Yanagida, S.
J. Chem. Soc., Dalton Trans. 2000, 3585–3590.
9
J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008 14307

A R T I C L E S
Bachman et al.
Figure 4. UV-vis spectra of 1b over the course of 30 min showing the
disappearance of the peaks at 350 and 470 nm as the reductive elimination
occurs. Inset: plot of [1b] vs time at λ ) 468 nm.
Figure 3. UV-vis spectra of 1a over the course of 30 min, showing the
disappearance of the peaks at 350 and 470 nm, as well as growth of a
shoulder at 275 nm, as the reductive elimination occurs. Inset: plot of [1a]
vs time at λ ) 468 nm.
Scheme 2. Thermal and Photochemical Pathways for Reductive
Elimination/Oxidative Addition Reactions
generally.26 The slightly longer bonds observed for 2, in
relation to its perhydro derivative, are consistent with the
idea put forth above regarding the impact of electron
deficiency on metal-ligand bonding. Interestingly, the Au-S
distances seen in 2 are significantly (0.065 Å) shorter than
those in 1, while the C-S and C-C bond lengths are
essentially identical. The difference in the observed Au-S
lengths for these two complexes is counter to expectations
based on the relative sizes of the Au(III) and Au(I) centers.
The most likely explanation for this result is the greater steric
crowding of the ligands in the four-coordinate Au(III)
complexes. This overall picture is also consistent with the
lower stability of Au(III) thiolates.
Reductive Elimination. The solution state UV-visible spectra
(Figure 3) of 1a display two strong bands in the visible or near-
UV region (468 nm, ꢀ ∼3600 cm-1 M-1; 350 nm, ꢀ ∼29 000
cm-1 M-1). We have tentatively assigned the former transition
to a metal-to-ligand charge transfer state, while the latter is likely
a ligand-based transition. The only other features in the spectra
are extremely intense ligand-centered transitions below 300 nm,
which are common to all three speciess1, 2, and 3sand hence
of little diagnostic value. The spectrum of 1b is similar (Figure
4) but with the higher 350 nm band overlapped with the ligand-
centered bands and now appearing as a shoulder on the ligand-
centered bands. This overlap is due to a red shift of the ligand-
centered bands of both 2b and 3b associated with the loss of a
fluorine atom (see Supporting Information). As 1 reductively
eliminates 3 to form 2 (Scheme 2), both of the bands at 470
and 350 nm are seen to decrease in intensity. The presence of
an isosbestic point at 310 nm in the spectra of 1a supports the
one-step nature of the reaction. As might be expected, kinetic
data under these conditions are consistent with a first-order
reaction for the reductive elimination process (Figures 3 and 4)
with an apparent rate constant of 1.61 × 10-3 s-1 for 1a and
1.58 × 10-3 for 1b.
In contrast to the earlier reports by Muller et al.,16a the
reductive elimination process appears to be independent of the
presence of oxygen, as solutions of 1 rapidly convert to 2 in a
rigorously oxygen-free environment (glovebox with <10 ppm
O2). However, the rate at which samples of 1 reach equilibrium
was found to depend strongly on the presence of light. Under
illumination by standard fluorescent lights, solutions of 1a reach
equilibrium in a few hours. In contrast, samples of 1a kept in
the dark require more than 10 days to achieve equilibrium. The
impact of light is even more pronounced for 1b. Samples
exposed to room illumination reach equilibrium within a few
hours, while those kept in the dark show almost no change,
either visibly or by 19F NMR spectroscopy, over the course of
several weeks. These results suggest that the reductive elimina-
tion process involving 1 occurs via both a thermal and a
photochemical pathway and, moreover, that the relative impor-
tance of these two pathways depends on the identity of the
thiolate ligands.
In order to begin to probe the impact of illumination on the
equilibria, we examined the equilibrium constant in the presence
and in the absence of room light. For 1a the value was
essentially identical in both casessKlight ) 0.13 and Kdark
)
0.12. While it was not possible to measure the equilibrium
constant for 1b under dark conditions, Klight was found to be
1.09. These results are consistent with electrochemical measure-
ments. It is slightly more difficult to reduce 1a than 1b (-302
and -279 mV vs NHE, respectively) and, conversely, slightly
easier to oxidize 2a than 2b (1.35 and 1.59 V vs NHE). Both
the equilibrium measurements and the electrochemical results
are consistent with the idea that greater electron deficiency at
the thiolate ligand enhances the thermodynamic stability of the
Au(III) complex. Indeed, we have been unable to isolate
gold(III) thiolate complexes for any thiol with fewer than four
(26) For example, see: (a) Hunks, W. J.; Jennings, M. C.; Puddephatt, R. J.
Inorg. Chem. 2000, 39, 2699–2702. (b) Tzeng, B.-C.; Schier, A.;
Schmidbaur, H. Inorg. Chem. 1999, 38, 3978–3984. (c) Bishop, P.;
Marsh, P.; Brisdon, A. K.; Brisdon, B. J.; Mahon, M. F. J. Chem.
Soc., Dalton Trans. 1998, 675–682. (d) Tzeng, B.-C.; Chan, C.-K.;
Cheung, K.-K.; Che, C.-M.; Peng, S.-M. Chem. Commun. 1997, 135–
136. (e) Schneider, W.; Bauer, A.; Schmidbaur, H. Organometallics
1996, 15, 5445–5446. (f) Forward, J. M.; Bohmann, D.; Fackler, J. P.,
Jr.; Staples, R. J. Inorg. Chem. 1995, 34, 6330–6336.
9
14308 J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008

Ligand Exchange by Addition-Elimination at Au(I) Centers
A R T I C L E S
Scheme 3. Equilibrium System of Oxidative Addition and Reductive Elimination Reactions Involving Gold(I) Dithiolate Complexes and
Disulfides
fluorine atoms, as such systems rapidly convert to either a
bisdithiolate complex analogous to 2 or an insoluble oligomeric
material with a generic formula of [AuSR]n.21a Even though
1a is thermodynamically more stable than 1b, its more rapid
equilibration with its reduced form (2a), particularly in the
absence of light, implies a lower thermal activation energy for
the more fluorinated complex (1a).
ratio (which would be expected for simple oxidative addition
process) observed in this case clearly illustrates the dynamic
nature of the addition-elimination process as well as the
preferential stabilization of the gold(III) species by the more
electron-deficient ligand.
Attempts to add more electron-rich disulfides across the Au(I)
center of 2a yielded no apparent reaction with one significant
exception, N-acetyl-L-cystine methyl ester ((AcMeCys)2). In this
case the result was the rapid formation of an insoluble white
solid consistent with the formula [Au(AcMeCys)]n by elemental
analysis. This solid also displayed an intense red luminescence,
consistent with similar oligomeric gold(I) species previously
prepared in our laboratories.28 In keeping with the preference
for gold to selectively bind the more electron-deficient ligands,
reactions of [Au(SPh)2]- with 3a or 3b resulted in rapid and
complete ligand exchange to yield 2a or 2b, respectively.
Reactions of [Au(SPh)2]- with other disulfides, such as (p-
FH4C6S)2 or (p-CH3H4C6S)2, have yielded complex mixtures,
including oligomeric materials with a nominal formula of
[AuSAr]n, that have eluded complete identification to date.21a
To further probe the role of light in these reactions, we also
examined the photostationary state (PS) constants for 1a/b at
two wavelengths: 470 and 350 nm. The results of these
experiments are summarized in Table 1. In the case of 1a, the
PS constant obtained at both wavelengths is identical within
experimental error to both Klight and Kdark. For 1b, irradiation
at 350 nm results in a PS constant that is roughly half that seen
for irradiation at 470 nm, while Klight is intermediate between
the two values. The overall picture that can be developed from
these results (Scheme 3) is that both the reductive elimination
reaction of 1 and the reverse process involving oxidative addition
of 3 across the gold center in 2 occur via both photochemical
and thermal pathways. Furthermore, the more significant
wavelength dependence of 1b suggests that the photochemical
pathway is more important in this case. This conclusion is
consistent with the exceedingly slow rate of reaction observed
for 1b in the absence of light. While the precise structural/
electronic reasons for the different behavior of 1a and 1b are
not entirely clear, it has been suggested that the relative energy
of π*-based and σ*-based excited states of fluorinated aromatic
rings depends on the number of fluorine atoms attached to the
ring.27 Increasing numbers of fluorine atoms stabilize the σ*
states such that they become the more stable excited state when
at least five fluorine atoms are present. Hence, it seems
reasonable to conjecture that similar differences in excited states
lead to the reactivity patterns seen in this system.
Oxidative Addition and Ligand Exchange. The outcomes of
mixing 2 or the closely related species [(PhS)2Au]- with a
variety of disulfide species are summarized in Table 2. As
expected, mixing 2a with 3a resulted in the re-formation of 1a,
while mixing 2b with 3b regenerated 1b. The cross reactions
of 2a with 3b or 2b with 3a resulted in a scrambling of the
ligands at gold. These results were confirmed by 19F NMR
spectroscopy, which shows signals for both the symmetric (3)
and asymmetric (F5C6SSC6F4H) disulfides, as well as a crystal
structure of an Au(III) complex in which the para positions of
the aromatic rings are a superposition of fluorine and hydrogen
atoms in an approximately 2:1 ratio.21a The greater than 1:1
Discussion
The surprising stability of the gold(III) tetrathiolate complexes
(1) reported here is certainly due to the electron deficiency of
the highly fluorinated thiolates employed. While such stability
has been suggested before,16 the formation of these complexes
by oxidative addition of a disulfide to a gold(I) fragment has,
to our knowledge, no precedent in the literature. Previously
reported examples of oxidative addition at gold(I) typically
involve addition of halogens or alkyl halides. Moreover, many
of these examples involve the distinctly different process of
addition across a pair of gold(I) centers to produce gold(II) (eq
3),29 rather than addition across a single metal atom to yield a
gold(III) center, as is seen in this work.
(28) Bachman, R. E.; Bodolosky-Bettis, S. A.; Glennon, S. C.; Sirchio,
S. A. J. Am. Chem. Soc. 2000, 122, 7146–7147.
(29) For recent examples, see: (a) Schuster, O.; Schmidbaur, H. Inorg. Chim.
Acta 2006, 359, 3769–3775. (b) Schneider, D.; Schuster, O.; Schmid-
baur, H. Dalton Trans. 2005, 1940–1947. (c) Abdou, H. E.; Mohamed,
A. A.; Fackler, J. P., Jr. Inorg. Chem. 2005, 44, 166–168. (d) Abdou,
H. E.; Mohamed, A. A.; Fackler, J. P., Jr. Z. Naturforsch. B: Chem.
Sci. 2004, 59, 1480–1482.
(27) Zgierski, M.; Fujiwara, T.; Lim, E. C. J. Chem. Phys. 2005, 122,
144312–144316.
9
J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008 14309

A R T I C L E S
Bachman et al.
Comparing the reactivity of 1a and 1b, the loss of a fluorine
atom in 1b results in a significant increase in electron density
at the ipso carbon of the aromatic ring. The latter can be easily
observed by the significant shifts seen for the appropriate 13C
NMR signals: 20 ppm for 1a and 1b; 6 ppm for 2a and 2b. It
is reasonable to assume that the increased electron density at
the ipso carbon translates directly to an increase in electron
density at the attached sulfur atom and, by extension, at the
gold center. Such an increase in electron density should
destabilize the higher gold oxidation state and lead to more facile
reduction. This picture is consistent with the rapid reduction to
gold(I) typically seen upon addition of thiols to gold(III) salts
as well as our inability to isolate gold(III) complexes of di- or
trifluorobenzenethiolates. In such cases, we observe rapid
reduction to gold(I) and isolation of either a complex analogous
to 2 or an insoluble oligomeric complex with a 1:1 metal-to-
ligand ratio; the precise outcome depends on both the identity
of the thiol and the reaction conditions.21a These results illustrate
the sensitivity of the gold center to relatively small changes in
the electronics of the ligand periphery.
are considered (i.e., the process designated by equilibrium
constant K3). Additionally, a new reaction pathway (A or B in
Scheme 3) becomes operable if the metal-bound sulfur atom is
a strong enough base to compete for a proton with other solution
species, including the solvent. Such a pathway, if operable, leads
to loss of a ligand and the formation of insoluble oligomeric
[RSAu]n. The resulting removal of gold species from the solution
equilibrium system drives the overall process toward the
oligomeric product, as required by Le Chatelier’s principle. In
the cases of the tetra- or pentafluorinated thiols, the very low
basicity at sulfur (pKa ∼4) implies that protonation is not
significant under the conditions employed and little to no
oligomer should be observed. In contrast, if [Au(AcMeCys)2]-
forms at any point, it will be protonated to a much greater extent
(pKa ∼8) and hence efficiently form an oligomer, as is indeed
observed. In intermediate cases such as [(PhS)2Au]-, the
outcome is likely to be highly sensitive to the reaction
conditions, particularly with respect to the presence of a proton
source.
Conclusion
While the complete mechanistic details of the addition and
elimination processes have yet to be elucidated, the rapidity and
reversibility of these processes under normal room illumination
suggest an alternative mechanism for ligand substitution reac-
tions at gold(I) centers generally, one that explains the seemingly
anomalous behavior seen for NAMeCys. Conceptually, the
addition of any disulfide to a solution of a gold(I) thiolate
complex can generate an equilibrium between the gold(I)
complex and the corresponding gold(III) species. The equilib-
rium constant for this process is set largely by the electron
deficiency of the organic fragments of the disulfide and the
ligands on the gold center. Greater electron deficiency on either
group yields a larger equilibrium constant for the oxidative
addition (or a smaller equilibrium constant for the reductive
elimination). Moreover, as shown by the reactions of [Au(S-
Ph)2]- with 3, the reductive elimination process favors elimina-
tion of the most electron-rich thiolates, an interpretation
completely consistent with the results reported by Shaw for the
conversion of auricyanide to aurocyanide by glutathione.10
Hence, multiple equilibria will be established in any system
containing gold(I) thiolate complexes and disulfides, at least in
principle (Scheme 3). However, gold(I) dithiolate complexes
are also known to undergo loss of a ligand, probably via
protonation of the lost ligand at sulfur, to form oligomeric
gold(I) thiolate species, [Au(SR)]n. This latter species is
essentially insoluble and, hence, removed from the equilibrium
system as it is formed. Therefore, a series of competitions are
created, with the observed outcome decided by both the relative
magnitude of the various equilibrium constants involved and
Le Chatelier’s principle.
The use of highly fluorinated electron-deficient ligands has
allowed us to demonstrate, for the first time, the role that
oxidative addition and reductive elimination processes play in
the reactivity of gold thiolate complexes. Given the facile nature
of the observed reactions, it is plausible that ligand substitution
reactions at gold(I) in a variety of systems, including biological
systems, undergo substitution processes by such an addition-
elimination process via transient gold(III) intermediates rather
than by a direct substitution process as has been previously
hypothesized. Furthermore, these results demonstrate that fine
control over reactivity at the gold center is possible via tuning
of ligand electronics. Such information should prove useful in
the further development of catalytically active gold complexes,
an area of much current attention.30
Acknowledgment. The authors thank The University of the
South and Georgetown University for financial support of this work.
Funding for the purchase of the diffractometer was provided by
the National Science Foundation and Georgetown University.
R.E.B. thanks Amy Barrios as well as Alice and Mitchell Bruce
for fruitful discussions.
Supporting Information Available: Complete tables of crystal
data, positional parameters, bond distances and angles, and
anisotropic displacement parameters for 1 and 2 as well as
UV-vis data for 2 and 3; X-ray crystallographic data, in CIF
format, for 1a,b and 2a,b. This information is available free of
charge via the Internet at http://pubs.acs.org. The X-ray crystal-
lographic files have also been deposited with CCDC (ref.
668528-668531).
For example, if K1 is large and K2 small, the oxidative
addition product will dominate. In contrast, if K1 is small and
K2 large, the outcome will be determined by the relative
magnitudes of K2 and K1-1: for K2 > K1-1, a ligand exchange
to produce a new gold(I) complex will be seen; for K1-1 > K2,
the result would be no apparent reaction, at least over the
relatively short time periods of the study, as the dominant
solution species would be the starting gold(I) complex. A
slightly more complex picture arises if mixed ligand systems
JA805266R
(30) (a) Della Pina, C.; Dimitratos, N.; Falletta, E.; Rossi, M.; Siani, A.
Gold Bull. 2007, 40, 67–72. (b) Pichugina, D. A.; Kuzµmenko, N. E.;
Shestakov, A. F. Gold Bull. 2007, 40, 115–120. (c) Richard, L. L.;
Gagosz, F. Organometallics 2007, 26, 4704–4707. (d) Ray, L.; Shaikh,
M. M.; Ghosh, P. Organometallics 2007, 26, 958–964. (e) Witham,
C. A.; Mauleon, P.; Shapiro, N. D.; Sherry, B. D.; Toste, F. D J. Am.
Chem. Soc. 2007, 129, 5838–5839.
9
14310 J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008