Gold-phosphine binding to de novo designed coiled coil peptides

Article abstract of DOI:10.1016/j.jinorgbio.2012.05.010

The coordination of the therapeutically interesting [AuCl(PEt₃)] to the de novo designed peptide, TRIL23C, under aqueous conditions, is reported here. TRIL23C represents an ideal model to investigate the binding of [AuCl(PEt₃)] to small proteins in an effort to develop novel gold(I) phosphine peptide adducts capable of mimicking biological recognition and targeting. This is due to the small size of TRIL23C (30 amino acids), yet stable secondary and tertiary fold, symmetric nature and the availability of only one thiol binding site. [AuCl(PEt₃)] was found to react readily with the Cys side chain in a 1:1 ratio as confirmed by UV-visible, ³¹P NMR and mass spectrometry. Circular dichroism confirmed that the coiled coil structure was retained on coordination of the {Au(PEt₃)}⁺ unit. Redesign of the exterior of TRIL23C based on a biologically relevant recognition sequence found in GCN4, did not alter the coordination chemistry of [AuCl(PEt₃)]. To the best of our knowledge, this represents the first report on the coordination of gold(I) phosphine compounds to de novo designed peptides, and could lead to the generation of novel gold(I) phosphine peptide therapeutics in the future.

Full text of DOI:10.1016/j.jinorgbio.2012.05.010

Journal of Inorganic Biochemistry 117 (2012) 298–305
Contents lists available at SciVerse ScienceDirect
Journal of Inorganic Biochemistry
journal homepage: www.elsevier.com/locate/jinorgbio
Gold-phosphine binding to de novo designed coiled coil peptides
a
b
b
Anna F.A. Peacock a, , Gemma A. Bullen , Lee A. Gethings , Jonathan P. Williams ,
⁎
Frederik H. Kriel c, Judy Coates c
a
School of Chemistry, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK
Waters Corporation, Atlas Park, Simonsway, Wythenshawe, Manchester M22 5PP, UK
Biomed, Advanced Materials Division, Mintek, Private Bag X3015, Randburg 2125, South Africa
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
The coordination of the therapeutically interesting [AuCl(PEt3)] to the de novo designed peptide, TRIL23C,
under aqueous conditions, is reported here. TRIL23C represents an ideal model to investigate the binding
of [AuCl(PEt3)] to small proteins in an effort to develop novel gold(I) phosphine peptide adducts capable
of mimicking biological recognition and targeting. This is due to the small size of TRIL23C (30 amino
acids), yet stable secondary and tertiary fold, symmetric nature and the availability of only one thiol binding
site. [AuCl(PEt3)] was found to react readily with the Cys side chain in a 1:1 ratio as confirmed by UV-visible,
31P NMR and mass spectrometry. Circular dichroism confirmed that the coiled coil structure was retained on
coordination of the {Au(PEt3)}+ unit. Redesign of the exterior of TRIL23C based on a biologically relevant rec-
ognition sequence found in GCN4, did not alter the coordination chemistry of [AuCl(PEt3)]. To the best of our
knowledge, this represents the first report on the coordination of gold(I) phosphine compounds to de novo
designed peptides, and could lead to the generation of novel gold(I) phosphine peptide therapeutics in the
future.
Received 13 March 2012
Received in revised form 3 May 2012
Accepted 21 May 2012
Available online 2 June 2012
Keywords:
De novo peptide
Coiled coil
Protein design
Gold phosphine
Auranofin
© 2012 Elsevier Inc. All rights reserved.
1. Introduction
cell penetrating peptides (CPPs) and tumour targeting sequences such
as RGD.
The search for new and improved metal based drugs is fueled pri-
marily by the success of the first platinum-based chemotherapeutic
drug, cisplatin, and the second generation analogues such as car-
boplatin and oxaliplatin. Both the success of platinum drugs and the
severe limitations associated with their use in the clinic have driven
extensive research into alternative metal complexes for drug devel-
opment. Gold is a promising non-platinum metal which has been in-
vestigated for the development of new therapeutics, and for which
promising results are increasingly being reported. Notably the phos-
phine gold(I) complex auranofin (marketed as Ridaura) is currently
used in the clinic for the treatment of rheumatoid arthritis. Further-
more, promising anticancer and anti-HIV activity have been reported
for auranofin and related gold(I) complexes [1,2].
Various studies suggest that the phosphine gold(I) unit in auranofin is
important for antitumor activity, but that the thiol carbohydrate moiety is
replaced in vivo [2]. It has been recognized that this offers an opportunity
to incorporate a bioactive ligand in its place [3–7]. One attractive class of
bioactive and biocompatible ligands are peptides, which offer targeting
and therapeutic opportunities, e.g. nuclear localization sequences (NLS),
Reports exist in the literature of short peptide gold(I) phosphine
complexes, which have been synthesized through the incorporation
of the gold(I) phosphine unit into a non-natural amino acid. These
are generally derivatives of the thiol containing amino acid cysteine
(Cys), and are subsequently used in solid phase peptide synthesis as
a means of incorporating the gold(I) phosphine into the peptide se-
quence [5]. This approach has been successfully employed; however,
these synthetic procedures can potentially be time consuming, re-
stricted by the harsh chemical synthesis conditions required and ulti-
mately would be limited to peptide sequences that can readily be
synthesized.
Long peptides and proteins can achieve biomolecular recognition
through the formation of many favorable non-covalent interactions
between amino acid side chains and the target. In addition to the
primary structure, it is the correct secondary structure, such as the
α-helix or β-sheet, which is crucial for the correct positioning of
these side chains, thereby achieving strong and selective binding
to the target. Not surprisingly biological recognition is achieved on
a larger length scale (nm) than traditional small molecule chemistry
(Å). It is therefore our intention to investigate the coordination of
gold(I) triethylphosphine based on auranofin, to protein-type sequences
with well defined secondary structure elements (nm). Ultimately the
goal would be to develop targeted derivatives of auranofin, with im-
proved biodistribution and bioactivity.
⁎
Corresponding author. Tel.: +44 1214144362; fax: +44 1214144403.
E-mail address: a.f.a.peacock@bham.ac.uk (A.F.A. Peacock).
0162-0134/$ – see front matter © 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.jinorgbio.2012.05.010

A.F.A. Peacock et al. / Journal of Inorganic Biochemistry 117 (2012) 298–305
299
A potential challenge associated with this approach is the large
size and complexity of the natural proteins we are intending to
mimic in our design. Therefore our intention is to exploit the advan-
tages offered by de novo peptide design (from “first principles”),
which provides an opportunity to study the coordination of gold(I)
phosphine to simplified protein folds in aqueous solution. Due to
the high affinity of gold(I) for soft ligands such as the thiol of Cys, it
should be possible to readily incorporate the gold(I) phosphine into
a peptide sequence under mild aqueous conditions, as has previously
been reported for natural proteins including albumin [8], zinc fingers
[9–11], glutathione reductase [12] and cyclophilin [13]. In this manu-
script we will focus on de novo designed α-helical coiled coil peptides
based on TRI, Ac-G(LKALEEK)4G-NH2, developed by Pecoraro, DeGrado
and co-workers [14,15]. Cys derivatives of these have been shown to
readily coordinate soft heavy metal ions such as mercury, cadmium, bis-
muth, arsenic and lead [14–23]. However, to the best of our knowledge,
the study reported here represents the first on the coordination of
gold(I) phosphine compounds to de novo designed peptides with well
defined secondary and tertiary structure. This approach could ultimate-
ly result in the development of a new generation of gold(I) phosphine
peptide therapeutics.
prior to, and at the end of, the titration using the Ellman's test
[26,27]. This confirmed 30 μM of available thiols prior to the addition
of any [AuCl(PEt3)]. However, once the absorbance at 250 nm had
reached its plateau, the Ellman's test indicated that no available thiols
were present in solution. This observation is further supported by an
analogous titration performed with a peptide sequence which does
not contain a Cys residue, TRIL2W.1 The titration of a stock solution
of [AuCl(PEt3)] into a 30 μM solution of TRIL2W monomer in 10 mM
phosphate buffer pH 8, did not result in any increase in the absor-
bance at 250 nm (Fig. 1B).
These results and the ratio of 1 Au(I): 1 TRIL23C monomer are con-
sistent with substitution of the chloride group with the deprotonated
thiol of Cys, so as to generate [Au(PEt3)(TRIL23C-S)]. In contrast, soft
metal ions such as mercury, cadmium, bismuth, arsenic and lead are ca-
pable of binding in the interior of the three stranded coiled coil,
(TRIL23C)3, to all three equivalent Cys side chains, one from each pep-
tide chain [14–23]. Despite reports of three and four coordinate
gold(I) complexes [2], linear two coordinate gold(I) is more common
and we observe the replacement of the highly labile chloride ligand
by a single thiol from one peptide chain. This suggests that the tri-
ethylphosphine ligand remains bound (vide infra) and prevents the
gold(I) from being accommodated in the interior of the coiled coil, as
a result of steric clashes.
2. Results and discussion
2.3. pH dependent binding
2.1. Peptide design
In order to assess the pH dependence of gold(I) complexation, the
pH of an aqueous solution containing 30 μM [AuCl(PEt3)] and 30 μM
TRI is a 30 amino acid peptide based on the heptad repeat approach,
Ac-G(LaKbAcLdEeEfKg)G-NH2, which spontaneously self-assembles in
water to generate a two stranded coiled coil at low pH, and a three
stranded coiled coil at higher pH (>5.5) [14]. This self-assembly is the
result of a hydrophobic core of leucine residues (a and d sites) and
interhelical salt bridges (between e and g sites). Gold(I) complexes
have been reported to bind to both histidine (His) and Cys side chains,
though with a greater affinity for the latter. As such, a thiol metal binding
site was designed in the interior of the coiled coil by replacing a leucine
(Leu) residue (at the 23 position) with a Cys, to generate the previously
reported TRIL23C, Ac-G-LKALEEK-LKALEEK-LKALEEK-CKALEEK-G-NH2
[23]. A control peptide which lacked the necessary thiol group required
for gold(I) coordination, was prepared. Tryptophan (Trp) was incorpo-
rated into this peptide sequence, Ac-G-WKALEEK-LKALEEK-LKALEEK-
LKALEEK-G-NH2 (TRIL2W), to aid in concentration determination.1
These sequences were synthesised, purified and characterized by rou-
tine methods [24].
2.2. UV-visible spectroscopy
Starting with the more reactive chloride derivative of auranofin,
[AuCl(PEt3)], its ability to bind to TRIL23C under mild conditions and
in aqueous solution, was investigated by UV-visible spectroscopy. Ali-
quots of a stock solution of [AuCl(PEt3)] were titrated into a 30 μM solu-
tion of TRIL23C monomer in 10 mM phosphate buffer pH 8. This
resulted in the steady increase in the absorbance at 250 nm (Fig. 1A),
which plateaus on addition of 30 μM [AuCl(PEt3)] (Fig. 1B), indicating
a 1:1 ratio between the gold(I) complex and the peptide monomer
TRIL23C. The UV-visible spectra obtained (Fig. 1A) are not dissimi-
lar to those previously reported for [AuCl(PEt3)] binding to cyste-
ine, penicilamine and small model peptides based on zinc finger
proteins [10,25]. The total metal concentration was used to calcu-
late an extinction coefficient of λ250 2935 M− 1 cm− 1 for the com-
plex [Au(PEt3)(TRIL23C-S)]. This compares well to similar species
in the literature [25].
In order to establish if binding was indeed through the intended
Cys side chain, the concentration of available thiols was measured
Fig. 1. (A) UV-vis spectra for a titration of [AuCl(PEt3)] into a solution containing 30 μM
TRIL23C in 10 mM phosphate pH 8. (B) A plot of the absorbance at 250 nm as a function
of [AuCl(PEt3)]/μM titrated into a 30 μM solution of TRIL23C and TRIL2W, respectively.
1
In the absence of a Cys thiol side chain, a single Trp residue is required to deter-
mine the peptide concentration by monitoring the absorbance at 280 nm.

300
A.F.A. Peacock et al. / Journal of Inorganic Biochemistry 117 (2012) 298–305
following the reaction of [(η6-bip)Ru(en)Cl]+, a potential anticancer
complex, to the neurotransmitter Substance-P peptide (RPKPQQFFGLM)
using ETD [29].
Initially, TRIL23C (2 μM) was analyzed denatured in an aqueous so-
lution of 50% methanol (Figure S1). Abundant multiply charged ions
were detected from [TRIL23C+4 H]4+ − [TRIL23C+8 H]8+ over the
m/z acquisition range shown (annotated as A4–A8, Figure S1). The
inset in Figure S1 shows the Maximum Entropy deconvoluted ESI
mass spectrum which provided the molecular weight of 3411.0 Da (se-
quence average mass of 3411.1 Da). A dimeric species, (TRIL23C)2, is
also detected (annotated as B6–B11, Figure S1), and can be attributed
to oxidation of Cys side chains in the denatured peptide, due to the ox-
idizing conditions of ESI [30,31].
Following the reaction of 100 μM TRIL23C with 1000 μM [AuCl(PEt3)]
(1:10), abundant multiply charged ions were detected for the
{Au(PEt3)}+-peptide complex from [TRIL23C+{Au(PEt3)}+2 H]3+
−
[TRIL23C+{Au(PEt3)}+6 H]7+ as shown in Figure S2A. The Maximum
Entropy deconvoluted ESI mass spectrum of the {TRIL23C+10 equiva-
lents [AuCl(PEt3)]} spectrum is shown in Figure S2B. The spectra con-
tain adducts, labelled A-F, which are assigned as TRIL23C, {Au}
TRIL23C, {Au(PEt3)}TRIL23C, {Au(Au(PEt3))}TRIL23C, {Au(PEt3)}2-
TRIL23C and (TRIL23C)2 respectively. The high ratio of [AuCl(PEt3)]:
TRIL23C (10:1) used in these experiments, is responsible for the for-
mation of species containing multiple gold(I) units. The formation of
{Au(PEt3)}2(Cys-S) has previously been reported for reaction of ex-
cess [AuCl(PEt3)] with albumin [32].
Fig. 2. UV-visible pH titration of a 30 μM solution of TRIL23C in the presence of 30 μM
[AuCl(PEt3)]. Absorbance monitored at 250 nm as a function of pH.
TRIL23C was gradually raised from ca. 4 to 9.5, and monitored by UV-
visible spectroscopy. This increase in pH was accompanied by a
steady increase in the absorbance at 250 nm, with a final spectrum
which resembles that shown earlier in Fig. 1A. A plot of the absor-
bance at 250 nm, associated with coordination to the Cys, as a func-
tion of pH is shown in Fig. 2. The absorbance at 250 nm does not
reach a plateau until ca. pH 9.5, suggesting complete formation of
[Au(PEt3)(TRIL23C-S)] at this pH. An apparent pKa of 7.3 has been
extracted from the pH profile shown in Fig. 2, for the deprotonation
of the thiol side chain on gold(I) coordination. A reverse pH titration
was performed by lowering the pH gradually from ca. 9.5 to 4.5. This
resulted in a drop in absorbance at 250 nm, suggesting reversible
binding of [AuCl(PEt3)] to TRIL23C as a function of pH. The observa-
tion of reversible binding is further supported by the identical elution
times of [Au(PEt3)(TRIL23C-S)] and TRIL23C under the acidic conditions
of our reversed phase HPLC experiments (data not shown).
In theory the precise site of binding of {Au(PEt3)}+ to TRIL23C can
be determined by both ETD and CID fragmentation methods. The
choice of method will depend on the multiply charged precursor ion
detected and thus selected to undergo fragmentation. Following CID of
[TRIL23C+{Au(PEt3)}+4 H]5+ and [TRIL23C+{Au(PEt3)}+6 H]7+
no sequence-specific ions containing the gold(I) based complex were
observed. The only peak containing the gold(I) complex was that of
{Au(PEt3)}+, a direct result from the loss of bound TRIL23C (data not
shown). The reverse was found following ETD-generated mass spectra
of the [TRIL23C+{Au(PEt3)}+4 H]5+ and [TRIL23C+{Au(PEt3)}+
6 H]7+ multiply charged precursor ions. Fig. 3 shows the ETD spec-
trum obtained from [TRIL23C+{Au(PEt3)}+4 H]5+. This spectrum
shows efficient ETD from which the precise site of binding of the
{Au(PEt3)}+ unit could be determined upon the Cys residue at posi-
tion 23 of the peptide TRIL23C.
The structural information inferred from the amount of peptide se-
quence coverage obtained using ETD has previously been shown to be
charge-state dependent [33]. It is known that in comparison to doubly
charged peptide ions, triply charged ions provide greater sequence cov-
erage upon fragmentation using ETD. The same observation has been
obtained in these experiments. Limited sequence coverage was obtained
for the triply and quadruply charged {Au(PEt3)}+-peptide adducts.
These ions produced abundant charge-reduced species which did not
allow the site of the gold(I) binding to be precisely determined. Howev-
er, the sequence coverage obtained by fragmentation of the 5+ and 7+
{Au(PEt3)}+-peptide adduct allowed the site of gold(I) binding to be lo-
cated upon the Cys residue at position 23, as expected. The percentage of
sequence coverage was 97% and 100% for the 5+ and 7+ respectively.
The 6+ ion was not selected as a precursor ion to undergo fragmenta-
tion since an interfering ion was observed on the shoulder of the isotopic
distribution. These MS experiments confirm binding of {Au(PEt3)}+ to
the thiol side chain of Cys in position 23, as originally intended.
These experiments suggest that binding of {Au(PEt3)}+ to the
thiol Cys side chain of TRIL23C is a pH dependent process, and that
[Au(PEt3)(TRIL23C-S)] is only fully formed under basic conditions.
This implies that at biological pH only a portion of the {Au(PEt3)}+
would remain bound to the peptide. However, as the {Au(PEt3)}+
unit is thought to be important for the antitumor activity of auranofin,
its release from the targeting peptide, triggered by a pH drop, could
be considered an attractive activation mechanism. Notably, cancerous
cells are thought to be more acidic than healthy cells, and this feature
could subsequently be responsible for triggering the release of more
{Au(PEt3)}+ than in healthy cells.
2.4. Mass spectrometry characterisation
Electrospray ionisation (ESI)-mass spectrometry (MS) has been
used to examine the reaction of the gold(I) phosphine complex
[AuCl(PEt3)] with our de novo designed peptide, TRIL23C. The inher-
ent sensitivity, speed and specificity of MS-based techniques can pro-
vide significant insight into both the design and mechanism of action
of metallodrugs. Here we investigate the application of a tandem MS
approach, electron transfer dissociation (ETD), to allow rapid deter-
mination of the binding site for [AuCl(PEt3)] reacted with TRIL23C.
ETD [28] is a powerful fragmentation technique, complementary
to collision-induced dissociation (CID), and was proved particularly
useful for determining sites of labile post-translational modifications
(PTMs) of peptides and proteins which can be difficult to characterize
using conventional CID methods. ETD is a radical-driven fragmenta-
tion technique and results in cleavage of the peptide N–Cα bond to
give c and z● type peptide product ions (cf. cleavage of the amide
bonds producing b and y" type ions using CID). It has previously
been demonstrated that the binding site could rapidly be obtained
2.5. 31P NMR characterisation
The substitution of the trans chloride of [AuCl(PEt3)] with the pep-
tide, TRIL23C, can readily be monitored by 31P NMR spectroscopy. The
addition of 1/3 equivalents of [AuCl(PEt3)] to a 4.96 mM solution of
TRIL23C monomer in 10% D2O at pH 2.3, resulted in one major reso-
nance at 48.3 ppm and a very small resonance at 37.4 ppm (Fig. 4A).

A.F.A. Peacock et al. / Journal of Inorganic Biochemistry 117 (2012) 298–305
301
Au(PEt )
3
c" 1 2 3 4 5
6
7 8 9 10 1112 13 14 15 1617 18 19 20 21 22 23 24 2526 27
GLKALEEKLKALEEKLKALEEKCKALEEKG
z'
29 28
26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7
6
5 4 3 2
1864.2
1484.0
100
100
(b)
(a)
c''23 2+
1485.0
1355.0
c''12
c''13
1483.0
1400.3
1354.0
z'22 2+
1399.8
z'9
1303.7
z'23 2+
1464.4
z'10
1432.7
c''22
1276.8
1351.3
1327.0
c''3
358.3
1392.4
0
1250
m/z
1300
1350
1400
1450
c''6
671.5
z'17
2244.3
z'18
2373.3
1658.0
3669.4
3370.1
3056.9
0
m/z
500
1000
1500
2000
2500
3000
3500
Fig. 3. ETD mass spectrum of the precursor ion [TRIL23C+{AuPEt3} +4H]5+, detected at m/z 746, showing 97% sequence coverage and identifying the attachment of the gold(I)
triethylphosphine to the cysteine residue in the 23 position.
The nature of the resonance at 48.3 ppm is not clear, but could be due to
formation of {Au(PEt3)2}+ [34,35]. The UV-visible pH titration (vide
supra) has demonstrated that basic conditions are required for depro-
tonation of the thiol and binding of the {Au(PEt3)}+ unit to the peptide,
confirmed by MS to be through the Cys side chain. The pH of the NMR
solution was therefore raised to 8.1, which resulted in the disappearance
of the original resonances, and the appearance of a new resonance at
39.2 ppm (accounting for 57% of the overall 31P signal), with two smaller
resonances at 39.9 (32%) and 40.5 ppm (11%), respectively (Fig. 4B).
These three signals all fall close to the range expected for {Au(PEt3)}+
bound to thiols, with previous reports citing resonances ranging be-
tween 34.2 and 39.1 ppm [8,34–37,39]. Under these experimental con-
ditions we have less {Au(PEt3)}+ than available thiol ligands (1:3) and
as such a mixture of species will be formed. The three resonances ob-
served are most likely due to coiled coils with differing numbers of
thiol bound to {Au(PEt3)}+. For example, one resonance could be due
to a coiled coil with all Cys side chains coordinated to {Au(PEt3)}+, e.g.
([Au(PEt3)(TRIL23C-S)])3. The two remaining resonances are likely
the result of coiled coils containing a mixture of metallated and un-
metallated peptide strands, e.g. (TRIL23C)2([Au(PEt3)(TRIL23C-S)])
and (TRIL23C)([Au(PEt3)(TRIL23C-S)])2, respectively. The affinity
of gold(I) for a thiol group has previously been demonstrated to be
inversely proportional to its pKSH [38]. Similarly a relationship was
reported between the 31P NMR chemical shift and the thermody-
namic stability of the resulting [Au(PEt3)(SR)]+ complex [39].
Based on these correlations it would seem that the Cys side chain
in TRIL23C has a high affinity for gold(I) forming stable complexes
with the {Au(PEt3)}+ unit.
resonance at 66.3 ppm (b1%) most likely due to formation of PEt3O
[39], there were no additional resonances detected over the range
240 to −50 ppm.
An analogous experiment performed on addition of [AuCl(PEt3)] to a
10% D2O solution (resulting concentration would have been 6.8 mM) in
the absence of TRIL23C, resulted in precipitation of the [AuCl(PEt3)]
complex and no resonances were detected in the 31P NMR spectrum.
The insolubility of [AuCl(PEt3)] in the absence of TRIL23C compared to
On further addition of [AuCl(PEt3)], to achieve a final ratio of 0.8
equivalents [AuCl(PEt3)] per TRIL23C peptide monomer, the reso-
nance at 39.2 ppm grew in intensity, accounting for 83% of the overall
31P signal. This is consistent with this signal being responsible for the
fully metallated coiled coil, ([Au(PEt3)(TRIL23C-S)])3. The addition of
more [AuCl(PEt3)] reduces the number of un-metallated peptide
strands and the two peaks at 39.9 and 40.5 ppm reduce in intensity
and correspond to significantly less of the overall 31P signal (b2
and b14%, respectively) (Fig. 4C). Except from an extremely small
Fig. 4. 31P{1H} NMR spectra from the reaction of [AuCl(PEt3)] and TRIL23C in 10% D2O.
(A) [AuCl(PEt3)] and TRIL23C in a 1:3 ratio at pH 2.3, (B) in a 1:3 ratio at pH 8.1 and
(C) in a 0.8:1 ratio at pH 8.1.

302
A.F.A. Peacock et al. / Journal of Inorganic Biochemistry 117 (2012) 298–305
Fig. 5. Ribbon diagrams of models of {Au(PEt3)}+ bound to the Cys residue at the alpha-helical interface of either a three stranded coiled coil, ([Au(PEt3)(TRIL23C-S)])3, or two
stranded coiled coil, ([Au(PEt3)(TRIL23C-S)])2. Shown are the main chain atoms represented as helical ribbons and the Cys side chains in stick form with the thiol group and
the Au(I) shown as spheres (orange and grey, respectively), as a (A or C) top-down view from the C-terminus and a (B or D) side-on view of the coiled coil.
experiments performed in the presence of TRIL23C, for which no
precipitation was observed, would be consistent with the coordination
of [AuCl(PEt3)] to TRIL23C.
accommodated in the hydrophobic interior of the coiled coil struc-
ture, but could be accommodated at the α-helical interface of a coiled
coil (Fig. 5). An alternative aggregation state (n) of the gold(I) phos-
phine coiled coil, ([Au(PEt3)(TRIL23C-S)])n, cannot be ruled out.
Despite this, it is clear these designed peptides retain their secondary
structure on binding the {Au(PEt3)}+ unit, and it is this designed sec-
ondary structure element which is ultimately going to be important
for achieving biomolecular recognition.
2.6. Circular dichroism
Although the UV-visible, MS and 31P NMR results confirm binding of
one {Au(PEt3)}+ unit to the Cys side chain of TRIL23C in a 1:1 ratio, they
are unable to provide information concerning the folding state of the
peptide. One option is that the designed coiled coil structure can toler-
ate the binding of {Au(PEt3)}+ to the Cys side chain. This would require
the thiol side chain to adopt a rotamer which is directed towards the α-
helical interface of the coiled coil. Potential models which satisfy this
scenario are proposed, and are shown in Fig. 5. Alternatively, the data
would also satisfy the scenario where binding of the {Au(PEt3)}+ unit
is accompanied by peptide unfolding, resulting in one {Au(PEt3)}+
unit bound through the Cys side chain of an unstructured peptide
monomer. However, our goal has been to generate gold(I) phosphine
peptide adducts with well defined secondary structure elements
which could ultimately be exploited for biomolecular recognition and
targeting. This therefore warranted further investigation by circular di-
chroism (CD), so as to evaluate the secondary structure of the peptide in
the absence and presence of increasing concentrations of [AuCl(PEt3)]
at pH 8.
2.7. Coiled coils based on natural protein recognition motifs
The complexation of [AuCl(PEt3)] to TRIL23C is a simple easy to
interpret system with which one can probe the binding of the
{Au(PEt3)}+ unit through a single Cys side chain of a peptide with de-
fined secondary and tertiary structure in aqueous solution. However,
ultimately we are interested in the coordination of {Au(PEt3)}+ to
peptides capable of biomolecular recognition as a result of both the
primary amino acid sequence and the α-helical secondary structure.
We therefore investigated the binding of [AuCl(PEt3)] to a designed
peptide, TRIGCN4, which similar to TRIL23C folds into a three stranded
coiled coil with a single Cys residue in the 23 position. However, its
A 30 μM solution of TRIL23C in 10 mM phosphate buffer at pH 8 was
monitored by CD and yielded the characteristic coiled coil spectra with
negative minima at 208 and 222 nm, see Fig. 6. The 222 nm molar ellip-
ticity (Θ222) value of −34 948 deg dmol−1 cm2 is consistent with a well
folded coiled coil structure (>95%). Aliquots of a stock solution of
[AuCl(PEt3)] were titrated into the 30 μM solution of TRIL23C, so that
the final concentration of [AuCl(PEt3)] was 5, 10, 15 and 30 μM, respec-
tively. No significant change was observed in the CD spectra, see Fig. 6,
indicating that the secondary α-helical structure is maintained on
addition of up to 1 equivalent of [AuCl(PEt3)] per peptide monomer.
The CD results suggest that on binding {Au(PEt3)}+, the peptide
remains well folded, with no noticeable change to the CD spectra.
This includes the ratio of Θ222/Θ208 (1.05), which is indicative of a
coiled coil [40,41]. This scenario rules out peptide denaturation on
binding {Au(PEt3)}+, but is consistent with {Au(PEt3)}+ bound to
the exposed Cys side-chains at the α-helical interface of either a
three- or a two-stranded coiled coil (proposed models are shown in
Fig. 5), both of which are adopted by the TRI peptide family under
certain conditions [14]. The steric bulk of PEt3 will certainly not be
Fig. 6. CD spectra showing the peptide region for a solution of 30 μM TRIL23C in a
0.1 cm quartz cuvette in 10 mM phosphate buffer at pH 8, in the presence of increasing
concentrations (5, 10, 15 and 30 μM) [AuCl(PEt3)]. The intensity at 222 nm remains
unchanged on increasing the concentration of [AuCl(PEt3)] and represents a well
folded coil (>95%).

A.F.A. Peacock et al. / Journal of Inorganic Biochemistry 117 (2012) 298–305
303
precise site of binding of {Au(PEt3)}+ to TRIGCN4 was determined from
ETD-generated mass spectra of the 7+ ion, [TRIGCN4+{AuPEt3}+
6 H]7+, which was selected as a precursor ion to undergo fragmenta-
tion, see Figure S4. This spectrum provides 80% sequence coverage,
however, the precise site of binding of the {Au(PEt3)}+ unit can still
be determined to be the Cys residue at position 23 of TRIGCN4
.
Both the UV-visible and MS results importantly verify that our
approach is not limited to TRIL23C. It can be used to coordinate
{Au(PEt3)}+ through the Cys side chain of peptides with primary
amino acid sequences and secondary structures, which have been
designed for targeting and biomolecular recognition.
3. Conclusions
We report our investigations into complexation of a gold(I) phos-
phine complex based on the drug auranofin, [AuCl(PEt3)], with syn-
thetic de novo designed peptide structures. These simplified protein
folds offer advantages over natural proteins as one can more readily
establish structure-function relationships through systematically cor-
relating chemical observations with modifications to the designed se-
quence. The peptide TRIL23C possesses a single thiol site for gold(I)
phosphine coordination, as well as a well folded α-helical coiled coil
structure in aqueous solution. It is the well defined secondary and ter-
tiary structure which can ultimately be exploited for biomolecular
recognition. UV-visible, 31P NMR and mass spectrometry confirm
that coordination readily occurs in aqueous solution in a 1:1 ratio
though the single Cys side chain. Binding is pH dependent with
basic conditions required for complete coordination. This process is
reversible and could potentially offer an interesting trigger or activa-
tion mechanism for delivery of the active {Au(PEt3)}+ unit in vivo.
The coiled coil structure of our peptide is retained on coordinating
the gold(I) phosphine, and possible models consisting of gold(I)
bound to the Cys side chain at the α-helical interface of either a
two- or a three-stranded coiled coil, have been proposed (Fig. 5).
Our findings suggest that this approach can be extended to the design
of peptide sequences based on biological recognition motifs of natural
proteins, e.g. the bacterial transcription factor GCN4, as redesign of
the exterior of TRIL23C based on this sequence, did not alter the
gold(I) phosphine coordination chemistry. Ultimately these studies
could lead to the development of a new generation of gold(I) phos-
phine peptide therapeutics.
Fig. 7. (A) UV-vis spectra for a titration of [AuCl(PEt3)] into a solution containing 30 μM
TRIGCN4 in 10 mM phosphate pH 8. (B) A plot of the absorbance at 250 nm as a function
of [AuCl(PEt3)]/μM, indicating a ca. 1:1 binding ratio.
exterior has been modified to resemble the DNA binding basic
domain of the bacterial transcription factor, GCN4.2
4. Experimental
UV-visible spectroscopy was employed to investigate the reaction
of the gold(I) phosphine complex, [AuCl(PEt3)], based on auranofin,
to the synthetic coiled coil, TRIGCN4, based on the transcription factor
GCN4. The absorbance at 250 nm was monitored for a titration of in-
creasing concentrations of [AuCl(PEt3)] into a solution containing
30 μM TRIGCN4 in 10 mM phosphate buffer pH 8. This resulted in the
formation of near identical spectra to that obtained with TRIL23C,
with the absorbance at 250 nm beginning to plateau at a ca. 1:1
ratio of [AuCl(PEt3)] to TRIGCN4, see Fig. 7.
4.1. Peptide synthesis and purification
All peptide synthesis reagents were purchased from AGTC
bioproducts. Peptides were synthesized on a CEM Liberty 1 automat-
ed peptide synthesizer on Rink amide MBHA resin (0.25 mmol
scale), using standard Fmoc-amino acid solid-phase peptide synthe-
sis protocols [24]. Iterative coupling and deprotections, separated by
washings (30–40 mL DMF), of all Fmoc-L-amino acids were performed
using a standard coupling programme (25 W, 75 °C, 300 s), except for
the coupling of Fmoc-L-Cys(Trt)-OH (0 W, 50 °C, 120 s followed by
25 W, 50 °C, 240 s) to prevent racemisation, and deprotection pro-
gramme (35 W, 75 °C, 180 s). The peptide N-terminus was acetylated
manually (20% acetic anhydride, 20% DIPEA in DMF), and then cleaved
from the resin (90% trifluoroacetic acid (TFA), 5% thioanisole, 2% anisole,
3% 1,2-ethanedithiol for 2 h at 298 K) with simultaneous removal of
side-chain protecting groups, to yield the C-terminal amide. Resin was
removed by filtration and crude peptides precipitated out of solution
on addition of cold diethyl ether. Peptides were purified by preparative
reversed phase C18 HPLC with a solvent mixture altered with a linear
gradient from 0.05% TFA in water to 0.05% TFA in CH3CN/H2O (2:1)
over 50 min. Peptides were identified by electrospray ionisation and
MALDI-MS. Stock solutions of the apopeptides were prepared in
ESI-MS was used to confirm {Au(PEt3)}+ binding to TRIGCN4 and to
determine the location of binding. Following the reaction of 100 μM TRI-
GCN4 with 1000 μM [AuCl(PEt3)] (1:10), abundant multiply charged ions
were detected for the peptide ([Au(PEt3)(TRIGCN4-S)]) complex from
[TRIGCN4 +{Au(PEt3)}+2 H]3+ - [TRIGCN4 +{Au(PEt3)}+7 H]8+ as
shown in Figure S3A. The Maximum Entropy deconvoluted ESI
mass spectrum of the {TRIL23C+10 equivalents [AuCl(PEt3)]} spec-
trum is shown in Figure S3B. The spectrum contains adducts, labeled
A-C, observed at mass 3577, 3891 and 4206 Da, and are assigned as TRI-
,
{Au(PEt3)}TRIGCN4 and {Au(PEt3)}2TRIGCN4, respectively. The
GCN4
2
Sequence Ac-G LKRLRNR LAALESR LRKLERK CKALEEK G-NH2. Unpublished results
A. F. A. Peacock, P. Holmes.

304
A.F.A. Peacock et al. / Journal of Inorganic Biochemistry 117 (2012) 298–305
distilled water, and their concentrations determined by calculating
the available thiol concentration using the Ellman’s test [26] or by
measuring the Trp absorbance at 280 nm in the presence of a chemical
denaturatant (guanidinium.hydrochloride).
mass spectrometer (Synapt-G2, Waters Corporation, Manchester, UK)
fitted with electron transfer dissociation (ETD) functionality. The instru-
ment has been described in detail elsewhere [44]. In brief, the instru-
ment comprises three consecutive, gas filled, travelling wave (T-Wave)
RF stacked ring ion guides prior to the ToF mass analyser. For ETD type
fragmentation, a sub-ambient pressure (~2 mbar) glow discharge
anion source [29] was used to fill the Trap (first) T-Wave cell with quad-
rupole mass selected ETD reagent anions formed from para-nitrotoluene
(m/z 137). During an acquisition, the source polarity and quadrupole set
mass are switched to allow multiply charged cations formed from ESI of
the peptides to interact with stored reagent anions in the Trap T-Wave.
This interaction allows an ion–ion type reaction resulting in ETD product
ions which if desired can be separated according to their ion mobilities in
the second cell (IMS T-Wave cell). (The mobility section of the instru-
ment was switched off during these studies.) Upon exiting the IMS cell
the ions enter the (third) Transfer T-Wave cell which can be used to
transfer ions into the ToF or optionally provide supplemental activation
(CID) prior to the ToF. For efficient ETD, within the Trap T-Wave cell,
the bath gas used was helium set to a pressure of 0.05 mbar. The Transfer
T-Wave cell was pressurized to 0.005 mbar with argon. The Trap T-Wave
speed and amplitude which influence the ion–ion interaction time as
well as the reaction rate were set to 300 m/s and 0.2 V respectively. In
collision induced dissociation (CID) mode selected peptide precursor
ions were induced to fragment by collisions with Argon. The collision
energy was optimized on the precursor ion charge state selected. Data
acquisition and processing were carried out using MassLynx (V4.1).
Maximum Entropy based deconvolution software was used to establish
the approximate mass from which the charge states were derived, for
the species investigated. Solutions were infused into the source region
of the MS at a flow rate of 5 μL/min.
4.2. Synthesis of HAuCl4.4H2O [42]
Gold metal (4.50 g) was added to freshly prepared aqua regia
(HClConc:HNO3Conc; 3:1) (40 ml) and stirred on a hotplate (mild
heating) for 15 h or until all the gold had dissolved. The aqua regia
was evaporated at 40–45 °C under vacuum to yield a bright yellow
residue. This residue was washed with 2×20 ml portions of HClConc
,
and dried at 40–45 °C under vacuum after each wash. The yellow res-
idue of HAuCl4.4H2O was thoroughly dried under vacuum and stored
under nitrogen. Yield: 9.14 g, 97%.
4.3. Synthesis of [AuCl(PEt3)] [43]
Excess triethylphosphine (PEt3) was added drop-wise to a yellow
solution of HAuCl4.4H2O (0.20 g, 0.49 mmol) in 5 ml ethanol until
colourless. The solution was stirred at room temperature for 1 h.
The solution was filtered over silica gel (2 cm) and the solvent re-
moved from the filtrate to yield a white precipitate. Yield: 151 mg,
89%. 1H NMR (CDCl3, 400 MHz): δH 1.14 (dt; 9 H; J=18.8, 7.6;
CH3CH2), 1.79 (dq; 6 H; J=10.0, 7.6; CH3CH2); 31P NMR (CDCl3,
162 MHz): δP 31.4 (s).
4.4. UV-visible spectroscopy
UV-Visible spectra were recorded in 1 cm pathlength quartz cuvettes
at 298 K on a Cary 5000 (TRIL23C) or Cary 50 (TRIL2W and TRIGCN4
)
spectrometer. Aliquots of a fresh 2.5 mM stock solution of [AuCl(PEt3)]
in (1:1) water:methanol were titrated into a 30 μM solution of peptide
monomer in 10 mM phosphate buffer pH 8 and the UV-Vis spectra
recorded after 10 min equilibration.
4.8. NMR spectroscopy
1H and 31P NMR spectra for the characterisation of [AuCl(PEt3)]
in CDCl3 were collected on a Bruker AVANCE (400 MHz 1H and
162 MHz 31P) spectrometer equipped with a 5 mm probe. All 31P NMR
spectra of [AuCl(PEt3)] in the presence of TRIL23C in 10% D2O were col-
lected at room temperature on a Bruker DRX500 (500 MHz 1H and
202 MHz 31P) spectrometer equipped with a 5 mm probe. Concentrated
stock solutions of [AuCl(PEt3)] (0.5 M) were freshly prepared in DMSO-d6
and aliquots added to the aqueous 10% D2O solution in the presence or
absence of TRIL23C, so that the final concentration of DMSO-d6 was less
than 1%. All spectra were recorded at 300 K; chemical shifts (δ) are
given in parts per million (ppm) downfield of 85% phosphoric acid
(H3PO4) at zero ppm. Data was processed using Bruker Topspin.
4.5. pH titrations
pH titrations of aqueous solutions of TRIL23C (30 μM) in the pres-
ence of 1 equivalent of [AuCl(PEt3)] were performed at ambient tem-
perature on a Cary 50 UV-visible spectrometer. The pH was measured
using a mini-glass combination pH electrode coupled to a Jenway
3510 pH meter, and titrated by adding small aliquots of concentrated
solutions of HNO3 and KOH, respectively. Both forward and reverse ti-
trations from ca. pH 9.5 to 4.5 were recorded. UV-visible spectra were
recorded from 200 to 500 nm and the absorbance at 250 nm plotted
as a function of pH.
Abbreviations
DMF
Fmoc
NLS
CCPs
ESI
MS
CD
ETD
CID
PTMs
TFA
dimethylformamide
4.6. Circular dichroism spectroscopy
9-fluorenylmethyloxycarbonyl
nuclear localisation sequences
cell penetrating peptides
electrospray ionisation
mass spectrometry
CD spectra were recorded in 1 mm pathlength quartz cuvettes at
298 K on a Jasco J-715 spectropolarimeter. The observed ellipticity
in millidegrees was converted into molar ellipticity, (Θ), and is
reported in units of deg dmol−1 cm2. Aliquots of a 2.5 mM stock
solution of [AuCl(PEt3)] in (1:1) water:methanol, were titrated into
a 30 μM TRIL23C solution in 10 mM phosphate buffer pH 8 and the
CD spectra recorded after 10 min equilibration.
circular dichroism
electron transfer dissociation
collision induced dissociation
post translational modifications
trifluoroacetic acid
4.7. Mass spectrometry (MS)
Aqueous solutions of peptide (100 μM) and the gold complex
[AuCl(PEt3)] (1000 μM) were prepared and mixed to give a peptide:
[AuCl(PEt3)] molar ratio of 1:10. The samples were incubated at
room temperature for 45 min. The mixture was then diluted with 50%
aqueous methanol to a concentration of 10 μM prior to MS analysis.
ESI-MS was performed on a hybrid quadrupole/ion mobility/oa-ToF
Acknowledgments
AFAP thanks the School of Chemistry at the University of
Birmingham, The Royal Society and EC COST group D39 for financial
support, along with Mintek for hosting a visit and for the use of their
facilities. Some equipment used in this research was obtained through

A.F.A. Peacock et al. / Journal of Inorganic Biochemistry 117 (2012) 298–305
305
[18] M. Matzapetakis, D. Ghosh, T.-C. Weng, J.E. Penner-Hahn, V.L. Pecoraro, J. Biol.
Inorg. Chem. 11 (2006) 876–890.
[19] A.F.A. Peacock, M. Stachura, L. Hemmingsen, V.L. Pecoraro, Proc. Natl. Acad. Sci.
U. S. A. 105 (2008) 16566–16571.
[20] V.L. Pecoraro, A.F.A. Peacock, O. Iranzo, M. Luczkowski, Bioinorganic chemistry,
in: E.C. Long, M.J. Baldwin (Eds.), ACS Symposium Series 1012, Understanding
the Biological Chemistry of Hg(II) Using a De Novo Design Strategy, American
Chemical Society, Washington, 2009, pp. 183–197.
Birmingham Science City: Innovative Uses for Advanced Materials
in the Modern World (West Midlands Centre for Advanced Materials
Project 2), with support from Advantage West Midlands (AWM) and
part funded by the European Regional Development Fund (ERDF).
FHK and JC wish to acknowledge the Department of Science and
Technology FP Seed Funding Programme.
[21] A.F.A. Peacock, J.A. Stuckey, V.L. Pecoraro, Angew. Chem. Int. Ed. 48 (2009)
7371–7374.
[22] S. Chakraborty, D.S. Touw, A.F.A. Peacock, J.A. Stuckey, V.L. Pecoraro, J. Am. Chem.
Soc. 132 (2010) 13240–13250.
Appendix A. Supplementary data
[23] O. Iranzo, D. Ghosh, V.L. Pecoraro, Inorg. Chem. 45 (2006) 9959–9973.
[24] W.C. Chan, P.D. White (Eds.), Fmoc Solid Phase Peptide Synthesis: A Practical
Approach, Oxford University Press, New York, 2000.
Supplementary data to this article can be found online at http://
dx.doi.org/10.1016/j.jinorgbio.2012.05.010.
[25] D.H. Brown, G. McKinlay, W.E. Smith, J. Chem. Soc., Dalton Trans. (1977)
1874–1879.
[26] G.L. Ellman, Arch. Biochem. Biophys. 82 (1959) 70.
References
[27] R. Rubbiani, I. Kitanovic, H. Alborzinia, S. Can, A. Kitanovic, L.A. Onambele, M.
Stefanopoulou, Y. Geldmacher, W.S. Sheldrick, G. Wolber, A. Prokop, S. Wölfl, I.
Ott, J. Med. Chem. 53 (2010) 8608–8618.
[28] J.E.P. Syka, J.J. Coon, M.J. Schroeder, J. Shabanowitz, D.F. Hunt, Proc. Natl. Acad. Sci.
U. S. A. 101 (2004) 9528–9533.
[1] E.R.T. Tiekink, Inflammopharmacology 16 (2008) 138–142.
[2] C.F. Shaw III, Chem. Rev. 99 (1999) 2589–2600.
[3] C.K. Mirabelli, R.K. Johnson, D.T. Hill, L.F. Faucette, G.R. Girard, G.Y. Kuo, C.M. Sung,
S.T. Crooke, J. Med. Chem. 29 (1986) 218–223.
[4] I. Ott, X. Qian, Y. Xu, D.H.W. Vlecken, I.J. Marques, D. Kubuat, J. Will, W.S.
Sheldrick, P. Jesse, A. Prokop, C.P. Bagowski, J. Med. Chem. 52 (2009) 763–770.
[5] C.P. Bagowski, Y. You, H. Scheffler, D.H. Vlecken, D.J. Schmitz, I. Ott, Dalton Trans.
(2009) 10799–10805.
[6] J. Caddy, U. Hoffmanns, N. Metzler-Nolte, Z. Naturforsch. 62b (2007) 460–466.
[7] J. Lemke, A. Pinto, P. Niehoff, V. Vasylyeva, N. Metzler-Nolte, Dalton Trans. (2009)
7063–7070.
[8] M.T. Coffer, C.F. Shaw III, M.K. Eidsness, J.W. Watkins II, R.C. Elder, Inorg. Chem. 25
(1986) 333–339.
[9] J. Larabee, J.R. Hocker, J.S. Hanas, Chem. Res. Toxicol. 18 (2005) 1943–1954.
[10] M.A. Franzman, A.M. Barrios, Inorg. Chem. 47 (2008) 3928–3930.
[11] F. Mendes, M. Groessl, A.A. Nazarov, Y.O. Tsybin, G. Sava, Isabel Santos, P.J. Dyson,
A. Casini, J. Med. Chem. 54 (2011) 2196–2206.
[29] J.P. Williams, J.M. Brown, I. Campuzano, P.J. Sadler, Chem. Commun. 46 (2010)
5458–5460.
[30] H. Deng, G.J. Van Berkel, Anal. Chem. 71 (1999) 4284–4293.
[31] G.J. Van Berkel, G.E. Giles, J.S. Bullock, L.J. Gray, Anal. Chem. 71 (1999) 5288–5296.
[32] J. Xiao, C.F. Shaw III, Inorg. Chem. 31 (1992) 3706–3710.
[33] S.J. Pitteri, P.A. Chrisman, J.M. Hogan, S.A. Mcluckey, Anal. Chem. 77 (2005)
1831–1839.
[34] A.A. Isab, A.L. Hormann, D.T. Hill, D.E. Griswold, M.J. DiMartino, C.F. Shaw III,
Inorg. Chem. 28 (1989) 1321–1326.
[35] W. Ashraf, A.A. Isab, J. Coord. Chem. 57 (2004) 337–346.
[36] C.F. Shaw III, M.T. Coffer, J. Klingbeil, C.K. Mirabelli, J. Am. Chem. Soc. 110 (1988)
729–734.
[37] E.M. Kinsch, D. Stephan, Inorg. Chim. Acta 91 (1984) 263–267.
[38] A.A. Isab, P.J. Sadler, J. Chem. Soc., Dalton Trans. (1982) 135–140.
[39] M.T. Coffer, C.F. Shaw III, A.L. Hormann, C.K. Mirabelli, S.T. Crooke, J. Inorg. Biochem.
30 (1987) 177–187.
[12] S. Urig, K. Fritz-Wolf, R. Réau, C. Herold-Mende, K. Tóth, E. Davioud-Charvet, K.
Becker, Angew. Chem. Int. Ed. 45 (2006) 1881–1886.
[13] J. Zou, P. Taylor, J. Dornan, S.P. Robinson, M.D. Walkinshaw, P.J. Sadler, Angew.
Chem. Int. Ed. 39 (2000) 2931–2934.
[40] N.E. Zhou, C.M. Kay, R.S. Hodges, J. Biol. Chem. 267 (1992) 2664–2670.
[41] N. Choy, V. Raussens, V. Narayanaswami, J. Mol. Biol. 334 (2003) 527–539.
[42] A. Haas, J. Helmbrecht, U. Niemann, in: G. Brauer (Ed.), Handbuch der Präparativen
Anorganischen Chemie, Verlag Stuttgart, Stuttgart, 1978, p. 1014.
[43] F.G. Mann, A.F. Wells, D. Purdie, J. Chem. Soc. (1937) 1828–1836.
[44] K. Giles, J.P. Williams, I. Campuzano, Rapid Commun. Mass Spectrom. 25 (2011)
1559–1566.
[14] G.R. Dieckmann, D.K. McRorie, D.L. Tierney, L.M. Utschig, C.P. Singer, T.V.
O'Halloran, J.E. Penner-Hahn, W.F. DeGrado, V.L. Pecoraro, J. Am. Chem. Soc. 119
(1997) 6195–6196.
[15] G.R. Dieckmann, D.K. McRorie, J.D. Lear, K.A. Sharp, W.F. DeGrado, V.L. Pecoraro,
J. Mol. Biol. 280 (1998) 897–912.
[16] A.F.A. Peacock, O. Iranzo, V.L. Pecoraro, Dalton Trans. 13 (2009) 2271–2280.
[17] B.T. Farrer, C.P. McClure, J.E. Penner-Hahn, V.L. Pecoraro, Inorg. Chem. 39 (2000)
5422–5423.