Nickel(III) oxidation of its glycylglycylhistamine complex

Article abstract of DOI:10.1039/b409929j

The doubly-deprotonated Ni(III) complex of Gly₂Ha (where Ha is histamine) undergoes base-assisted oxidative self-decomposition of the peptide. At ≤ p[H⁺] 7.0, a major pathway is a two-electron oxidation at the a-carbon of the N-terminal glycyl residue. Major products (up to 73%) of this two-electron oxidation are glyoxylglycylhistamine and ammonia. Glyoxylglycylhistamine will decay to give isocyanatoacetylhistamine and formaldehyde. Two-electron oxidations of the second glycyl and histamine residues occur as minor pathways (12% of the total possible reaction). Above p[H⁺] 8.5, two Ni(III)-peptide complexes form an oxo bridge in the axial positions to give a reactive dimer species. This proximity allows the resulting Ni(III)-peptide radical intermediates to undergo peptide-peptide cross-linking at the N-terminal glycyl residues. The products found below p[H⁺] 7.0 are observed above p[H⁺] 8.5 as well, although in lower yields. In contrast to this work, Ni^III(H- ₂Gly₂HisGly) undergoes a four-electron oxidation at the N-terminal glycyl residue. Oxidation at the internal glycyl and histidyl residues are not observed. The reactivity of Ni^III(H _-2Gly₂Ha)⁺ is also different than Cu ^III(H_-2Gly₂Ha)⁺, which undergoes a two-electron oxidation at the histamine group with no peptide-peptide cross-linking in basic solution.

Full text of DOI:10.1039/b409929j

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
Nickel(III) oxidation of its glycylglycylhistamine complex†
Brandon J. Green, Teweldemedhin M. Tesfai and Dale W. Margerum*
Department of Chemistry, Purdue University, West Lafayette, IN, 47907-2084, USA
Received 30th June 2004, Accepted 2nd September 2004
First published as an Advance Article on the web 20th September 2004
The doubly-deprotonated Ni(III) complex of Gly2Ha (where Ha is histamine) undergoes base-assisted
oxidative self-decomposition of the peptide. At  p[H+] 7.0, a major pathway is a two-electron oxidation at
the a-carbon of the N-terminal glycyl residue. Major products (up to 73%) of this two-electron oxidation are
glyoxylglycylhistamine and ammonia. Glyoxylglycylhistamine will decay to give isocyanatoacetylhistamine and
formaldehyde. Two-electron oxidations of the second glycyl and histamine residues occur as minor pathways
(12% of the total possible reaction). Above p[H+] 8.5, two Ni(III)–peptide complexes form an oxo bridge in
the axial positions to give a reactive dimer species. This proximity allows the resulting Ni(II)–peptide radical
intermediates to undergo peptide–peptide cross-linking at the N-terminal glycyl residues. The products found
below p[H+] 7.0 are observed above p[H+] 8.5 as well, although in lower yields. In contrast to this work,
NiIII(H−2Gly2HisGly) undergoes a four-electron oxidation at the N-terminal glycyl residue. Oxidation at the
internal glycyl and histidyl residues are not observed. The reactivity of NiIII(H−2Gly2Ha)+ is also different than
CuIII(H−2Gly2Ha)+, which undergoes a two-electron oxidation at the histamine group with no peptide–peptide
cross-linking in basic solution.
Table 1 Molar absorptivities for NiII(H−2Gly2Ha) and NiIII(H−2-
Gly2Ha)+
Introduction
Many biological molecules contain histidine residues that are
important binding sites for Ni(II). Human serum albumin,
which binds nickel(II) at the N-terminal AspAlaHis residues,
is believed to be the primary nickel transport protein in
blood.1 Nickel is also present in a number of enzymes2 and it
is known that both Ni(II) ions in urease contain two histidine
donors.3 Dervan and co-workers showed that site-specific
DNA cleavage occurred after binding GlyGly–L-His to the
amine glycyl terminus of a DNA binding agent and adding
Ni(II) with oxidizing agents.4 High-valent nickel–peptide
complexes that contain a histidine group in the third residue
have been used in site-specific DNA cleavage,4–6 RNA
cleavage,7 protein–protein cross-linking,8–11 DNA–peptide
cross-linking,12 and protein affinity labeling.13 Ni(III) inter-
mediates are proposed in some cases.8,9,13 Knowledge of the
reaction mechanisms of these species is of interest because of
their biochemical importance. We have recently studied the
decomposition products of NiIII(H−2Gly2HisGly) in neutral
and basic solution14 as well as the decomposition kinetics
of the Gly2HisGly and Gly2Ha (where Ha is histamine)
complexes of Ni(III).15 In the present work, we examine the
sites of peptide oxidation and the decomposition products
of NiIII(H−2Gly2Ha)+. Below p[H+] 7.0, oxidative self-decom-
position of NiIII(H−2Gly2HisGly) and NiIII(H−2Gly2Ha)+
occurs at the a-carbon of the N-terminal glycyl residue to
give an intermediate that releases ammonia. Above p[H+]
8.5, decomposition kinetic studies of both Ni(III)–peptides
show a change in reaction order that supports the formation
of oxo-bridged dimer intermediates15 followed by the forma-
tion of cross-linked peptides.14 However, the major pathways
of these two Ni(III)–peptide complexes differ significantly
below p[H+] 7.0. A significant amount of NiIII(H−2Gly2Ha)+
undergoes a two-electron oxidation at the peptide, as
opposed to the predominant four-electron oxidation seen in
NiIII(H−2Gly2HisGly).14 NiIII(H−2Gly2Ha)+ also reacts much
differently than other Ni(III)16 and Cu(III)–peptides,17–19 which
undergo oxidation at the third residue and not the N-terminal
residue.
Species
k/nm
e/M−1 cm−1
NiII(H−2Gly2Ha)
NiIII(H−2Gly2Ha)+
NiIII(H−2Gly2Ha)+
424
247
334
138(1)a
1.16(2) × 104
3.92(6) × 103
a Reference 15.
Experimental section
Reagents
Methods for preparation and standardization of HClO4,
NaClO4, Ni(ClO4)2, as well as sodium phosphate and acetate
buffers are given previously.14 Glycinamide hydrochloride
(Aldrich) and diglycinamide hydrochloride (Bachem) were used
as received. Gly2Ha was synthesized and purified by Dr H. D.
Lee17 (Calcd. for C9H15N5O2·1/4 H2O: C, 47.05%; H, 6.80%; N,
30.48%. Found: C, 47.40%; H, 6.66%; N, 30.15%.). Ammonium
formate and ammonium acetate buffers were used in samples
for mass spectrometric analysis. The pH was adjusted with
ammonia, acetic acid, or formic acid. Oxone® (2KHSO5·K2SO4·
KHSO4·5H2O),20 obtained from Aldrich, was standardized as
previously described.14
Methods
NiII(H−2Gly2Ha) was prepared by adding the peptide in 5–10%
excess to a solution containing 1 mM Ni(ClO4)2. Solutions
were adjusted to pH 7.5 to ensure that the complex was fully
formed.
The trivalent metal peptide complex was prepared by either
electrochemical oxidation via a flow-through bulk electrolysis
apparatus21 or via chemical oxidation with Oxone. Oxidation via
the flow-through bulk electrolysis apparatus was carried out at
800–820 mV (vs. Ag/AgCl) with a flow rate of 0.43 mL min−1.
Ni(III) formation via chemical oxidation was carried out by
rapid-flow mixing of a buffered NiII(H−2Gly2Ha) solution with
a limiting quantity of Oxone. In p[H+] ranges where the Ni(II)
and Ni(III) complexes are unstable (below 6.4 and above 7.2,
respectively), unbuffered NiII(H−2Gly2Ha) (pH 7.5) was mixed
with Oxone and the resulting Ni(III) solution (pH 6.8–7.2) was
† Electronic supplementary information (ESI) available: Tables S1–S7
and Figs. S1–S10. See http://www.rsc.org/suppdata/dt/b4/b409929j/
3 5 0 8
D a l t o n T r a n s . , 2 0 0 4 , 3 5 0 8 – 3 5 1 4
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4

View Article Online
Table 2 Gly2Ha recoveries from NiIII(H−2Gly2Ha)+
immediately collected in a buffered solution. Oxone selectively
oxidizes Ni(II) to Ni(III) (99+% yield, Figs. S1 and S2, see ESI†).
Molar absorptivities of NiIII(H−2Gly2Ha)+ (Table 1) were deter-
mined from UV-vis spectral data with iodometric determination
of Ni(III).14
Atmospheric conditions
p[H+] (l = 0.1 M)
% Gly2Ha Recovery
Air-Saturateda
Argon-Saturateda
Air-Saturatedb
Air-Saturatedc
Air-Saturateda
Air-Saturateda
6.7
6.7
6.7
67(2)
64(2)
63(2)
69(2)
36(1)
32(1)
An Orion 720A digital pH meter and Orion Ross® E Com-
bination Electrodes, stored in 0.1 M or 1.0 M NaClO4, were
used to measure pH values. In cases where the ionic strength
(l) was controlled, the pH values were corrected for activities
and liquid junction potentials to give −log[H+] (i.e., p[H+])
values. Linear fits of −log[H+] vs. measured pH, obtained from
NaOH titrations of HClO4 solutions at l = 0.1 and 1.0 M
(NaClO4), were used to convert electrode responses to p[H+].
An ammonia-sensing ion selective electrode (Accumet) was
used for NH3 determinations. After acid-catalyzed hydrolysis in
the absence of oxygen in 6 M HCl at 100 °C, the sample pH was
adjusted to 12 with NaOH. All measurements were carried
out at 25.0 °C.
5.4
9.2d
10.3d
a Samples were analyzed 2 h after Ni(III) generation. Ni(III) samples
generated in basic solution were originally 0.5 mM and acidified to pH
b
2.5 immediately after Ni(III) decay. Sample was analyzed 3 days after
c
Ni(III) generation. Sample was analyzed 24 h after Ni(III) generation.
d l = 1.0 M.
UV-vis spectra were obtained from a Perkin-Elmer Lambda-9
UV-vis-NIR spectrophotometer with 1.00 cm or 5.00 cm cells.
Stopped-flow kinetic data were obtained on an Applied
Photophysics Stopped-Flow Spectrophotometer with a cell path
of 0.962 cm at 25.0 °C.
Products of Ni(III)–peptide decomposition were separated by
strong cation-exchange chromatography with either a Whatman
Partisil 10 lm or a Dionex CS-12A analytical column (both
4.6 × 250 mm). Separations were performed on a Varian 5020
Liquid Chromatograph with a Hewlett-Packard 1050 Diode
Array Detector or a Dionex DX-500 HPLC system with an
AD20 single-wavelength absorbance detector.
Fig. 1 UV-vis spectra of NiII(H−2Gly2Ha) at p[H+] 7.5 and
NiIII(H−2Gly2Ha)+ at p[H+] 2.0 (l = 0.1 M).
Peptide recoveries from their Ni(III) complexes were deter-
mined chromatographically by acidifying the reaction products
to the eluent pH prior to injection. Glycinamide, diglycinamide,
and Gly2(a,b-dehydro-Ha) were quantified in a similar manner,
using eqn. (1) to account for the reaction stoichiometry. The pre-
viously-reported molar absorptivity of Aib2(a,b-dehydro-Ha)17
was used to determine Gly2(a,b-dehydro-Ha) at 260 nm. The
yield of glyoxylglycylhistamine is expressed as the total possible
yield of ammonia (eqns. (1) and (2) minus the contributions from
glycinamide and diglycinamide hydrolysis. All uncertainties are
reported as the standard deviation of three measurements.
Products analyzed by electrospray ionization mass
spectrometry (ESI-MS) were isolated via strong cation-
exchange HPLC with a formic acid/ammonium formate
eluent. The reaction solutions were either free of inorganic
salts prior to chromatographic isolation of the products or
were chromatographically separated from the products during
the fraction collections. The samples were dried on a VirTis
Freezemobile 12 lyophilizer. All ESI-MS analyses were carried
out with a Finnigan MAT LCQ mass spectrometer system in
the Purdue Medicinal Chemistry and Molecular Pharmacology
Department.
of the N-terminal glycyl residue. In this two-electron oxidation,
2 mols of NiIII(H−2Gly2Ha)+ react to form one mol of oxidized
peptide and one mol of NiII(H−2Gly2Ha) (eqn. (1)).
2 NiIII(H−2Gly2Ha)+ →
NiII(Oxidized Peptide) + NiII(H−2Gly2Ha)
(1)
NiII(Oxidized Peptide) + H2O →
NiII(H−2glyoxylglycylhistamine) + NH4+ (2)
We propose that the oxidized peptide intermediate is (a-
OH–Gly)GlyHa, but this is a transitory species that reacts
readily to release ammonia and form glyoxylglycylhistamine
(I, Scheme 1 and eqn. (2)). Glyoxylglycylhistamine can decay
to give isocyanatoacetylhistamine (II, Scheme 1). Chromato-
graphic data show other minor pathways including oxidation
at the internal glycyl residue (7.7(3)%) and the histamine group
(4.3(2)%, Scheme 2). In each case, two NiIII(H−2Gly2Ha)+ react
to give one oxidized peptide product. The reaction mixture was
hydrolyzed in the presence of strong acid and heat to force
ammonia release from glycinamide (GlyNH2) and diglycinamide
(Gly2NH2, Scheme 2). The total ammonia yield was determined
to be 85(8)% of what is possible from a two-electron oxidation.
After subtracting the ammonia contribution from GlyNH2 and
Gly2NH2 hydrolysis (12%, Scheme 2), the pathway outlined
in eqns. (1) and (2) was found to account for up to 73(6)% of
the total possible reaction. Gly2Ha recoveries of 63–67% were
obtained at p[H+] 6.7 (Table 2), as opposed to the 50% yield
predicted by eqn. (1). It is possible that some four-electron oxida-
tions occur, as is the case for NiIII(H−2Gly2HisGly) where peptide
recoveries are as high as 75%. We were not able to identify all of
the oxidation fragments of NiIII(H−2Gly2Ha)+ decomposition.
Results and discussion
Characterization of NiII/III(H−2Gly2Ha)0/+
Protonation and stability constants of the Ni(II) complex of
Gly2Ha have been reported.22 Nickel(II) and Ni(III) complexes
of Gly2Ha were characterized in the present work via UV-
vis spectroscopy. NiII(H−2Gly2Ha) exhibits a d–d transition
at 424 nm and NiIII(H−2Gly2Ha)+ has strong ligand-to-metal
charge-transfer bands at 247 and 334 nm. UV-vis spectra of
the Ni(II) and Ni(III) complexes are shown in Fig. 1 and molar
absorptivities of these bands are summarized in Table 1. The
spectral characteristics of both complexes are very similar to
Identification of products from p[H+] 5.4–7.0
those of NiII/III(H−2Gly2HisGly)−/0 14
.
A chromatogram of NiIII(H−2Gly2Ha)+ products originally at
p[H+] 6.7 is shown in Fig. 2 and an ESI mass spectrum of an
isolated fraction of glyoxylglycylhistamine (I) is shown in Fig. 3.
Along with the [I + H]+ adduct at m/z 225, [I + MeOH + H]+
and [I + MeOH + Na]+ adducts are also present at m/z 257
Decomposition reactions from p[H]+ 5.4–7.0
The predominant pathway for oxidative self-decomposition is a
general-base assisted,15 two-electron oxidation of the a-carbon
D a l t o n T r a n s . , 2 0 0 4 , 3 5 0 8 – 3 5 1 4
3 5 0 9

View Article Online
Scheme 1 The major two-electron oxidative decomposition pathway of NiIII(H−2Gly2Ha)+ from p[H+] 5.4–7.0.
Scheme 2 Minor NiIII(H−2Gly2Ha)+ two-electron oxidative decompo-
sition pathways at the internal glycyl residue (a) and histamine
group (b).
Fig. 3 ESI mass spectrum of glyoxylglycylhistamine (I) from Fig. 2.
acetonitrile/water mixture for analysis, the m/z 257 and 279
peaks were absent. Additionally, the MS–MS spectrum of m/z
257 exhibits an intense (100% relative abundance) peak at m/z
225, indicating that it is an adduct of I.
Isocyanatoacetylhistamine (II, Fig. 2) is proposed to form
from the decay of Ni(II)glyoxylglycylhistamine. Mass spectro-
metric and chromatographic data show that this species is
formed in appreciable amount from p[H+] 5.4 to 7.0. Isocya-
nates have been observed as products of Fe(II)-catalyzed protein
oxidation where oxygen is present.23,24 In these cases, atmospheric
O2 inserts into an a carbon radical to form a peroxypeptide
intermediate that reacts further to form an isocyanatopeptide.
However, atmospheric O2 does not play a role in isocyanate
formation in our work. If O2 were to react with oxidized peptide
intermediates, the resulting peroxypeptide could react with
additional NiII(H−2Gly2Ha), resulting in much lower Gly2Ha
recoveries. Our previous study of NiIII(H−2Gly2HisGly) showed
similar parent peptide recoveries and Ni(III) decomposition
Fig. 2 Chromatogram of 0.5 mM NiIII(H−2Gly2Ha)+ products
originally at p[H+] 5.4. Chromatographic conditions: Strong cation-
exchange separation in 125 mM [formate]T (pH 4.40) at 0.5 mL min−1,
100 lL injection, and kdet = 225 nm. Glyoxylglycylhistamine (I), and
isocyanatoacetylhistamine (II) were identified via ESI-MS. The latter
portion of the chromatogram containing unreacted Gly2Ha is omitted
for clarity.
and 279, respectively. A methanol/water solution was used
to dissolve the sample for mass spectrometric analysis, so the
presence of methanol in the mass spectrum is not surpris-
ing. When the sample was dried again and re-dissolved in an
3 5 1 0
D a l t o n T r a n s . , 2 0 0 4 , 3 5 0 8 – 3 5 1 4

View Article Online
Table 3 Yields of ammonia and other productsa from Ni(III)–peptides
Product
% Yields from NiIII(H−2Gly2Ha)+
% Yields from NiIII(H−2Gly2HisGly)
Ammoniab
43(4)
73(6)
7.7(3)
4.3(2)
1(1)
0
18(1)
0
0
0
0
72(2)
2(1)
Glyoxylglycyl–histaminec,d
GlyNH2c
Gly2NH2c
Gly2(a,b-dehydro–Ha)c,e
oxamylglycylhistidyl–glycyinec,f
CO2c,f
0
a Conditions: p[H+] 6.7, l = 0.1 M, 25.0 °C. b Ammonia yields are expressed in terms of [Ni(III)–peptide]total. c Other product yields are expressed as a
percent of the total possible product of a two-electron oxidation for NiIII(H−2Gly2Ha)+ and the total possible product of a four-electron oxidation of
NiIII(H−2Gly2HisGly).9 d Determined from the total ammonia yield minus ammonia contributions from oxidation products at the internal glycyl residue
and histamine group. e p[H+] 7.5–8.9. f Reference 14.
kinetics when the complex was in the presence and absence of
O2.14 Gly2Ha recoveries from its trivalent Ni complex at p[H+]
6.7 do not deviate significantly either (Table 2). The recoveries
of 67(2)% and 64(2)% in the presence and absence of O2,
respectively, are statistically equivalent. When the sample is
left at p[H+] 6.7 for three days in the presence of atmospheric
O2, the Gly2Ha recovery is 63(2)%. These results indicate that
isocyanate formation does not occur via an O2 oxidation path-
way. HPLC, MS, and IR data of NiIII(H−2Gly2HisGly) products
suggest the presence of isocyanatoacetylhistidylglycine,14 which
is analogous to species II.
Oxidations at the internal glycyl residue (7.7(3)%, Scheme 2a)
and the histamine group (4.3(2)%, Scheme 2b) were quantified
by chromatographic measurement of their amide products
(glycinamide and diglycinamide, respectively). The formation of
both amides is complete within 24 h at p[H+] 6.7 and the percent
oxidation at each residue is summarized in Table 3.
electron-transfer reactions from NiIII(H−2Gly2Ha)+, even though
NiII(H−2(a-OH–Gly)GlyHa) might be sufficiently long-lived for
this to occur. Stopped-flow kinetic data show that even at p[H+]
5.4, where NiII(H−2Gly2Ha) is unstable, the rate of dissociation
of Gly2Ha from Ni(II) is not instantaneous (k = 0.174 s−1).
Although the NiII(H−2(a-OH–Gly)GlyHa) complex may be
sufficiently stable kinetically to undergo an additional oxida-
tion to Ni(III), this does not occur. The two-electron oxidation
products observed in this work are not observed as products of
NiIII(H−2Gly2HisGly) decay. We propose that the high standard
potential of NiIII(H−2Gly2HisGly) (E0 = 0.960(4) V)15 allows it
to undergo a four-electron oxidation, while the lower potential
of NiIII(H−2Gly2Ha)+ (E0 = 0.928(4) V)15 may at least partially
restrict the decomposition to a two-electron oxidation.
The Ni(III)– and Cu(III)–peptide complexes have different
sites of ligand oxidation. We have proposed14 that this is due to
the fact that Cu(III)–peptide complexes can have deprotonated
amine groups17,18,21 with multiple bonding to the metal21 while
Ni(III)–peptide complexes do not.15 Cu(III)–peptide complexes
are not oxidized at the N-terminal glycyl residue and we believe
that the Cu(III)–N multiple bond character at this residue may
be responsible for this.
At p[H+] 7.5–8.9, NiII(H−2Gly2(a-OH–Ha)) can also undergo
dehydration to form an alkene peptide (Scheme 2b). Gly2(a,b-
dehydro–Ha) is a very minor product, accounting for only
1(1)% of the total possible product. Both Gly2(a-OH–Ha) and
Gly2(a,b-dehydro–Ha) are isolated chromatographically and
identified via ESI-MS (Figs. S4 and S5, respectively, see ESI†).
A fraction of co-eluting Gly2(a-OH–Ha) and Gly2(a,b-dehydro–
Ha) afforded a mass spectrum containing a [Gly2(a,b-dehydro–
Ha) + H]+ adduct at m/z 224, a [Gly2(a,b-dehydro–Ha) + Na]+
adduct at m/z 246, and a [(Gly2(a,b-dehydro–Ha)2 + Na]+ dimer
at m/z 469. The mass spectrum also includes a species that corres-
ponds to a [Gly2(a-OH–Ha) + Na]+ adduct at m/z 264. At p[H+]
8.9, the formation of NiII(H−2Gly2(a,b-dehydro–Ha)) is seen at
307 nm. This is in agreement with the spectroscopic data of the
oxidation products of the decarboxylated NiIII(H−2Gly2His)
product.16
Decomposition reactions from p[H]+ 8.5–10.3
As is the case for NiIII(H−2Gly2HisGly),14 the decomposition
kinetics of NiIII(H−2Gly2Ha)+ changes from first- to second-
order above p[H+] 8.5.15 An additional pathway exists in basic
solution where two Ni(III)–peptide complexes form an oxo-
bridged dimer species via Ni(III)–O–Ni(III) bonding.14,15 If
each peptide undergoes a one-electron oxidation, the resulting
radicals can then recombine at their a-carbons to give a cross-
linked peptide (Species III). The reaction is outlined in Scheme 3
and eqn. (3). A peptide product is formed that has approximately
twice the molecular weight of the monomeric Gly2Ha.
Comparison of reactivity with similar Cu(III) and Ni(III)
complexes
NiIII(H−2Gly2Ha)+ reacts much differently than its Cu(III)
analog.17 The main pathway of CuIII(H−2Gly2Ha)+ decom-
position is a two-electron oxidation at the a carbon of the
histamine group to form CuIII(H−2Gly2(a-OH–Ha))+. This
species then dehydrates to form the alkene peptide Gly2(a,b-
dehydro–Ha).17 An analogous alkene is also the major product
of CuIII(H−2Aib2Ha)+ decay.17 However, chromatographic and
UV-vis data show that Gly2(a,b-dehydro–Ha) is only a minor
product of NiIII(H−2Gly2Ha)+ decomposition.
NiIII(H−2Gly2Ha)+ also reacts differently than NiIII(H−2Gly2-
His), which undergoes decarboxylation at the histidyl residue,16
while NiIII(H−2Gly2Ha)+ cannot undergo oxidative decarboxyl-
ation. NiIII(H−2Gly2Ha)+ reacts similarly to NiIII(H−2Gly2HisGly)
in that both undergo oxidation at the N-terminal glycyl resi-
due.14 However, below p[H+] 7.0, NiIII(H−2Gly2HisGly) primar-
ily undergoes a four-electron oxidation, while NiIII(H−2Gly2Ha)+
mainly undergoes two-electron oxidations. No products are
observed that result from additional oxidation of the proposed
NiII(H−2(a-OH–Gly)GlyHa) intermediate (Scheme 1) via rapid
Scheme 3 Proposed reaction sequence for double one-electron oxida-
tion at pH 9 to give cross-linking. After acidification, species III and IV
are isolated.
D a l t o n T r a n s . , 2 0 0 4 , 3 5 0 8 – 3 5 1 4
3 5 1 1

View Article Online
Table 4 Comparison of peptide recoveries from NiIII(H−2Gly2Ha)+ and NiIII(H−2Gly2HisGly) and the percent shifts to the double one-electron
oxidation pathway in basic solution
Complex
p[H+] (l = 0.1 M)
% Peptide recovery
% Shift to double 1 e− oxid.a
NiIII(H−2Gly2HisGly)
6.7
9.2
10.3
6.7
9.2
10.3
72(2)
37(1)
33(1)
67(2)
36(1)
32(1)
0
49(2)
54(2)
0
46(2)
52(2)
NiIII(H−2Gly2Ha)+
a The percent shifts to the double one-electron oxidation pathways were calculated based on the peptide recoveries at p[H+] 6.7. All [Ni(III)–
peptide] = 0.5 mM.
-2 H+
+ 
2 NiIII (H-2Gly2Ha)(H2O)2

[NiIII (H-2Gly2Ha)(H2O)]2O →
+
Ni2II(cross-linked peptide) H→ Species III + 2 Ni2+
(3)
Chromatographic, mass spectrometric, and spectroscopic
data show that the 2,3-diaminobutanedioic acid derivative (III)
slowly oxidizes in the presence of atmospheric O2 at p[H+] 4.40
to yield a 2,3-diaminobutenedioic acid derivative (IV, Scheme 3).
Recoveries of Gly2Ha from its Ni(III) complex are significantly
lower at p[H+] 8.5–10.3 than at p[H+] 5.4–7.0 and the reac-
tion shifts to the double one-electron oxidation pathway to
approximately the same extent as NiIII(H−2Gly2HisGly) (Table 4).
The two-electron oxidation pathway (eqn. (1)) also occurs at
p[H+] 8.5–10.3.
Ni(III) complexes have been proposed14,15 to link axially via
oxo bridges. A similar structure with a bent geometry of the
oxo bridge has been reported.25 This type of structure gives
close proximity of the peptides and their carbon radical inter-
mediates, where radical recombination gives a C–C bond, to
form a peptide–peptide cross-linked species.
Identification of species III
The chromatogram in Fig. 4 shows a product (III) that elutes
5 min later than Gly2Ha (not shown). ESI-MS data of 3
show [M + H]+ and [M + Na]+ adducts at m/z 449 and 471,
respectively (Fig. 5). This species corresponds to the cross-linked
peptide with a molecular weight that is two mass units less than
twice the mass of the parent peptide (MW = 225). In addition,
formylatedadductsareobservedat m/z477and499. Formylation
of the diamino product of III occurs during lyophilization of the
sample in formate buffer.26 MS–MS data of these peaks (Tables
S1–S3, see ESI†) show the loss of neutral molecules, such as
water and amino acid residues. A small amount of formylation
Fig. 5 ESI mass spectrum of cross-linked Gly2Ha product (III) and
its N-formylated derivatives from Fig. 4. Some of species IV has formed
from exposure to atmospheric O2 (where the peaks at m/z 447 and 469
are their [IV + H]+ and [IV + Na]+ adducts, respectively).
at the histamine group is also observed in the MS–MS frag-
mentation of the m/z 499 peak (Fig. S3, see ESI†). Supporting
evidence for the regiochemistry of peptide–peptide cross-linking
will be discussed later.
Peptide–peptide cross-linking becomes more significant as
p[H+] increases, as shown by lower recovery of the original
peptide at higher p[H+] (Tables 2 and 4) and the formation of
a higher molecular weight product. The double one-electron
oxidation pathway is not significant below p[H+] 7.0.15 As p[H+]
increases, the reaction shifts more towards the double one-
electron pathway (Scheme 3). Table 4 shows that approximately
half of the reaction shifts to the double one-electron pathway at
p[H+] 10.3. Under similar conditions, the percent shifts of the
NiIII(H−2Gly2HisGly) and NiIII(H−2Gly2Ha)+ complexes to their
respective cross-linking pathways are comparable (Table 4). The
peptide–peptide cross-linking observed with Ni(III) is not found
in the Cu(III) analog,17 as the square-planar geometry of Cu(III)
precludes the formation of an oxo bridge between two Cu(III)
species.
Identification of species IV
Fig. 4 Chromatograms showing the loss of III and the formation of
IV from NiIII(H−2Gly2Ha)+, originally at p[H+] 9.0. The reaction mixture
was held at pH 4.40 in the absence of light under atmospheric conditions
and chromatograms were taken 0, 1, 2, 3, and 4 h after Ni(III) generation.
Chromatographic conditions: Strong cation-exchange separation in
Fig. 6 shows an ESI mass spectrum of the 2,3-diaminobutene-
dioic acid derivative (species IV in Fig. 4). MS–MS data of this
species (Tables S4 and S5, see ESI†) show similar fragmentation
to III. As is the case for its Gly2HisGly analog,14 species IV is
formed by atmospheric O2 oxidation of III (Scheme 3, Fig. 4). It
+
25 mM [formate]T, 150 mM [NH4 ]T, pH 4.4 at 2.0 mL min−1, 100 lL
injection, and kdet = 225 nm.
3 5 1 2
D a l t o n T r a n s . , 2 0 0 4 , 3 5 0 8 – 3 5 1 4

View Article Online
is unlikely that Ni(III) is involved in this oxidation. The doubly-
deprotonated Ni(II)–peptide complex, needed for Ni(III) forma-
tion, is not stable under these conditions (pH 4.0). Therefore,
we propose that atmospheric O2 is solely responsible for the
oxidation of III to IV. The mass spectrum of isolated fractions
of III (Fig. 5), where no nickel is present, shows a small amount
of IV when atmospheric O2 is the only possible oxidant. The
2,3-diaminobutenedioic acid derivative has three conjugated
double bonds that will stabilize the molecule. In addition, the
lone pair electrons from the enamine nitrogens can participate
with the CC double bond and adjacent carbonyl groups to
provide additional stability via resonance forms. MS–MS data
of III show that the cross-linking C–C bond breaks readily in
the gas phase. The MS–MS spectrum of [III + H]+ has a large
(48% relative abundance) peak at m/z 226 (Table S1, see ESI†).
This mass correlates to a [Gly2Ha + H]+ adduct, indicating that
a gas phase rearrangement involving the cross-linked a carbons
has occurred. These data suggest that the cross-linking C–C
bond in III is relatively weak. However, the MS–MS spectrum
of [IV + H]+ does not show a similar adduct of appreciable
(>5%) relative abundance (Table S4, see ESI†). These data are
consistent with the presence of stronger bonding (a double
bond) between the a carbons in IV. As is the case with the
corresponding Gly2HisGly analog,14 the UV-vis spectrum of IV
has an absorbance at 275 nm.
ammonium formate buffer during the chromatographic fraction
collection/lyophilization steps. Since the conjugated enamino
nitrogens of IV are less nucleophlic than the corresponding
amines28 of III, they should be less reactive towards formylation.
As shown in Fig. 6, there is only a very weak (<5% relative abun-
dance) [formylIV + Na]+ adduct at m/z 497. This is probably due
to formylation at the imidazole p nitrogen. MS–MS data show
that the imidazole nitrogen of III formylates to a small extent,
as a fragment at m/z 360 is present which corresponds to a
[III + Na–formylHa]+ adduct (42% relative abundance, Table
S3, see ESI†). On the other hand, the mass spectrum of III
shows very strong [formylIII + H]+ and [formylIII + Na]+
adducts at m/z 477 and 479, respectively (Fig. 5). Formylation
of III occurs at one of the vicinal amino groups.
As is the case with the NiIII(H−2Gly2HisGly) complex,14
cross-linking between two carbon radicals is proposed.
Nitrogen–nitrogen radical linkage, followed by oxidation would
give an azo compound that is not supported by our spectro-
scopic data.29 Carbon–nitrogen radical formation followed by
oxidation to an imine is also unlikely, as oxidation at nitrogen
atoms has not been observed.14,17–19 For these reasons, peptide
oxidation and radical recombination are believed to occur at the
a-carbon atoms.
Conclusions
A major reaction of NiIII(H−2Gly2Ha)+ self-decomposition
from p[H+] 5.4–7.0 is a two-electron oxidation at the N-
terminal glycyl residue to give NiII(H−2glyoxylglycylhistamine).
By contrast, the NiIII(H−2Gly2HisGly) complex undergoes
a four-electron oxidation at the same site. This difference in
reactivity may be attributed to the lower standard potential
of NiIII/II(H−2Gly2Ha)+/0. Another major difference is that
NiIII(H−2Gly2Ha)+ oxidizes its ligand at all peptide residues, so
it is much less site-specific than the corresponding Gly2HisGly
complex. Above p[H+] 8.5, NiIII(H−2Gly2Ha)+ more closely
mimics the reactivity of NiIII(H−2Gly2HisGly) by forming an
oxo-bridged Ni(III) dimer intermediate that reacts to form a
cross-linkedpeptideproduct. Thereactivityof NiIII(H−2Gly2Ha)+
also significantly differs from that of CuIII(H−2Gly2Ha)+ and
CuIII(H−2Gly2HisGly), which undergo oxidation exclusively at
the third residue and cannot form oxo-bridged species in basic
solution.
Acknowledgements
This work was supported by National Science Foundation
Grant CHE-0139876. The authors are grateful for the synthesis
of Gly2Ha by Dr H. D. Lee and the donation of a lyophilizer
by Mary Bower of the Biochemistry Department. The authors
also wish to thank Karl V. Wood and Mark A. Lipton for helpful
conversations.
Fig. 6 ESI mass spectrum of oxidized crosslinked Gly2Ha
product (IV) from Fig. 4. The small peak at m/z 497 is proposed to be a
[formylIV + Na]+ adduct from formylation at the histamine group.
References
1 H. L. Drake, in The Bioinorganic Chemistry of Nickel, ed. Lancaster,
J. R., Jr., VCH, New York, 1988, ch. 6, pp. 112–113.
2 S. W. Ragsdale, Curr. Opin. Chem. Biol., 1998, 2, 208.
3 M. A. Pearson, L. O. Michel, P. R. Hausinger and P. A. Karplus,
Biochemistry, 1997, 36, 8164.
4 D. P. Mack and P. B. Dervan, J. Am. Chem. Soc., 1990, 112, 4604.
5 J. G. Muller, R. P. Hickerson, R. J. Perez and C. J. Burrows, J. Am.
Chem. Soc., 1997, 119, 1502.
6 Q. Liang, D. C. Ananias and E. C. Long, J. Am. Chem. Soc., 1998,
120, 248.
7 W. Bal, J. Lukszo and K. S. Kasprzak, Chem. Res. Toxicol., 1997, 10,
915.
8 I. J. Brittain, H. Xiaofen and E. C. Long, Biochemistry, 1998, 37,
12113.
9 G. Gill, A. A. Richter-Rusli, M. Ghosh, C. J. Burrows and S. E.
Rokita, Chem. Res. Toxicol., 1997, 10, 302.
10 K. C. Brown,Z. Yu,A. L. BurlingameandC. S. Craik, Biochemistry,
1998, 37, 4379.
Regiochemistry of peptide–peptide cross-linking
Radical recombination in the formation of the cross-linked
Gly2Ha species is proposed to occur at the a-carbons of the N-
terminal glycyl residues, as is the case for NiIII(H−2Gly2HisGly).14
The proposed site of oxidation at p[H+] 8.5–10.3 is similar to
the site of oxidation at p[H+] 5.4–7.5. The lower retention time
of species IV compared to species III (Fig. 4) suggests that the
pKa value of at least one of the amino groups is significantly
decreased. Cross-linking at the N-terminal glycyl residue,
followed by oxidation at this site (Scheme 3) would lower the pKa
values of the resulting enamino groups. Conjugated enamines
exhibit pKa values as low as 2.8.27 Such a pKa change could
account for the lower retention time of IV compared to III on
a strong cation-exchange column. Finally, ESI-MS data show
that III formylates much more readily than IV in the presence of
11 M. D. Person, K. C. Brown, S. Mahrus, C. S. Craik and A. L.
Burlingame, Protein Sci., 2001, 10, 1549.
D a l t o n T r a n s . , 2 0 0 4 , 3 5 0 8 – 3 5 1 4
3 5 1 3

View Article Online
12 A. J. Stemmler and C. J. Burrows, J. Am. Chem. Soc., 1999, 121,
6956.
13 F. Amini, T. Kodadek and K. C. Brown, Angew. Chem., Int. Ed.,
2002, 41, 356.
14 B. J. Green, T. M. Tesfai, Y. Xi and D. W. Margerum, Inorg. Chem.,
2004, 43, 1463.
15 T. M. Tesfai, B. J. Green and D. W. Margerum, Inorg. Chem., in
press.
20 R. J. Kennedy and A. M. Stock, J. Org. Chem., 1960, 25, 1901.
21 T. A. Neubecker, S. T. Kriksey, K. L. Chellappa and D. W.
Margerum, Inorg. Chem., 1979, 18, 444.
22 T. Gajda, B. Henry, A. Aubry and J. J. Delpuech, Inorg. Chem., 1996,
35, 586.
23 B. S. Berlett and E. R. Stadtman, J. Biol. Chem., 1997, 272, 20313.
24 E. R. Stadtman and B. S. Berlett, Chem. Res. Toxicol., 1997, 10,
485.
16 W. Bal, M. I. Djuran, D. W. Margerum, E. T. Gray, Jr., M. A.
Mazid, R. T. Tom, E. Nieboer and P. J. Sadler, J. Chem. Soc., Chem.
Commun., 1994, 1889.
17 M. R. McDonald, W. M. Scheper, H. D. Lee and D. W. Margerum,
Inorg. Chem., 1995, 34, 229.
18 M. R. McDonald, F. C. Fredericks and D. W. Margerum, Inorg.
Chem., 1997, 36, 3119.
19 S. K. Burke, Y. Xu and D. W. Margerum, Inorg. Chem., 2003, 42,
5807.
25 B. Bag, N. Mondal, G. Rosair and S. Mitra, Chem. Commun., 2000,
1729.
26 R. Viville, A. Scarso, J. P. Durieux and A. Loffet, Pept.: Struct.
Funct., Proc. 8th Am. Pept. Symp., 1983, 747.
27 H. Leismann, G. Marzolph, H.-D. Scharf and M. Behruzi, Chem.
Ber., 1983, 116, 2591.
28 G. Haeflinger and H. G. Mack, in The Chemistry of Enamines, Part 1,
ed. Z. Rappoport, John Wiley & Sons, New York, 1994, ch. 1, pp. 2–3.
29 H. Rau, Angew. Chem., Int. Ed., 1973, 12, 224.
3 5 1 4
D a l t o n T r a n s . , 2 0 0 4 , 3 5 0 8 – 3 5 1 4