Cleavage of the peptide bond of β-alanyl-L-histidine (carnosine) induced by a CoIII-amine complexes: Reaction, structure and mechanism

Article abstract of DOI:10.1039/b402571g

Cleavage of the peptide bond occurs when β-alanyl-L-histidine (carnosine) reacts with [Co(tren)Cl₂] (tren = tris(2-aminoethyl) amine) to give [Co(tren)(histidine)]^2- 1 and [Co(tren)(β- alanine)]^2- 2. [Co(tren)(histidine)]^2- 1 crystallizes in the enantiomorphic space group P2₁2₁2₁ and 2 crystallizes in the P2₁/c space group. The mechanism of the cleavage reactions were studied in detail for the precursor [Co(tren)Cl₂] ⁺ and [Co(trien)Cl₂]⁺, which convert into [Co(tren)(OH)₂]⁺/[Co(tren)(OH)(OH₂)] ²⁺ and [Co(trien)(OH)₂]⁺/[Co(trien)(OH)(OH ₂)]²⁺ in water at basic pH [trien = 1,4,7,10- tetraazadecane). At a slightly basic pH, the initial coordination of the substrate (β-alanyl-L-histidine) is by the carboxylate group for the reaction with [Co(tren)Cl₂]⁺. This is followed by a rate-limiting nucleophilic attack of the hydroxide group at the β-alanyl-L-histidine carbonyl group. In a strongly basic reaction medium substrate, binding of the metal was through carboxylate and amine terminals. On the other hand, for the reaction between [cis-β-Co(trien)Cl ₂]⁺ and β-alanyl-L-histidine, the initial coordination of the substrate takes place via an imidazole ring nitrogen, independently, and followed by a nucleophilic attack of the hydroxide group at the β-alanyl-L-histidine carbonyl group. The circular dichroism spectrum for 1 suggests that a very small extent of racemization of the amino acid (L-histidine) takes place during the cleavage reaction between [Co(tren)Cl ₂]⁺ and β-alanyl-L-histidine. Reaction between [cw-β-Co(trien)Cl₂]⁺ and β-alanyl-L-histidine also causes cleavage of the peptide bond, producing a free β-alanyl molecule and a cationic fragment [cis-α-Co(trien)(histidine)]²⁺ 3 that crystallizes in the optically active space group P2₁2 ₁2₁. Unlike the previous case an appreciable degree of racemization of the L-histidine takes place during the reaction between [cis-β-Co(trien)Cl₂]⁺ and β-alanyl-L-histidine. Crystals containing L-histidine and D-histidine fragments in the [cis-α-Co(trien)(histidine)]^2- moiety were crystallographically documented by mounting a number of randomly selected crystals.

Full text of DOI:10.1039/b402571g

View Article Online / Journal Homepage / Table of Contents for this issue
Cleavage of the peptide bond of ꢀ-alanyl-L-histidine (carnosine)
induced by a CoIII–amine complexes: reaction, structure and
mechanism
Manas K. Saha, Uday Mukhopadhyay and Ivan Bernal*
Department of Chemistry, University of Houston, Houston, Texas 77204, USA
Received 19th February 2004, Accepted 17th March 2004
First published as an Advance Article on the web 31st March 2004
Cleavage of the peptide bond occurs when β-alanyl--histidine (carnosine) reacts with [Co(tren)Cl2]ϩ (tren =
tris(2-aminoethyl)amine) to give [Co(tren)(histidine)]2ϩ 1 and [Co(tren)(β-alanine)]2ϩ 2. [Co(tren)(histidine)]2ϩ
1
crystallizes in the enantiomorphic space group P212121 and 2 crystallizes in the P21/c space group. The mechanism
of the cleavage reactions were studied in detail for the precursor [Co(tren)Cl2]ϩ and [Co(trien)Cl2]ϩ, which convert
into [Co(tren)(OH)2]ϩ/[Co(tren)(OH)(OH2)]2ϩ and [Co(trien)(OH)2]ϩ/[Co(trien)(OH)(OH2)]2ϩ in water at basic
pH (trien = 1,4,7,10-tetraazadecane). At a slightly basic pH, the initial coordination of the substrate (β-alanyl--
histidine) is by the carboxylate group for the reaction with [Co(tren)Cl2]ϩ. This is followed by a rate-limiting
nucleophilic attack of the hydroxide group at the β-alanyl--histidine carbonyl group. In a strongly basic reaction
medium substrate, binding of the metal was through carboxylate and amine terminals. On the other hand, for the
reaction between [cis-β-Co(trien)Cl2]ϩ and β-alanyl--histidine, the initial coordination of the substrate takes place
via an imidazole ring nitrogen, independently, and followed by a nucleophilic attack of the hydroxide group at the
β-alanyl--histidine carbonyl group. The circular dichroism spectrum for 1 suggests that a very small extent of
racemization of the amino acid (-histidine) takes place during the cleavage reaction between [Co(tren)Cl2]ϩ
and β-alanyl--histidine. Reaction between [cis-β-Co(trien)Cl2]ϩ and β-alanyl--histidine also causes cleavage
of the peptide bond, producing a free β-alanyl molecule and a cationic fragment [cis-α-Co(trien)(histidine)]2ϩ
that crystallizes in the optically active space group P212121. Unlike the previous case an appreciable degree of
racemization of the -histidine takes place during the reaction between [cis-β-Co(trien)Cl2]ϩ and β-alanyl--
3
histidine. Crystals containing -histidine and -histidine fragments in the [cis-α-Co(trien)(histidine)]2ϩ moiety
were crystallographically documented by mounting a number of randomly selected crystals.
crystals are not readily obtained. The structural characteriz-
Introduction
ation of the product is important to further elucidate the
mechanistic pathway1 and to understand the stereochemical
preferences10 of the reaction. Our study shows that, depending
on the amine ligands attached to the CoIII ion, different types of
products occur and different stereochemical characteristics
result. Also, depending upon the dipeptide molecule, different
extents of racemization of amino acids were observed during
the course of the cleavage reactions.
Recently, we published the isolation and characterization of an
unprecedented trans-product in the cleavage of a peptide bond
with a CoIII(amine) complex.1 A few years back, during an
investigation of the conglomerate crystallization phenomenon,
some of us noticed the cleavage of amide bonds mediated by
CoIII(amine) complexes.2 This study prompted us to further
investigate the scope of peptide bond cleavages. Metal com-
plexes that can cleave peptide bonds under mild conditions are
useful for structure–activity studies of proteins, investigation of
protein structural domains and in converting large proteins into
smaller fragments that are more amenable for sequencing.3,4
In addition such reagents are also useful for the chemical
modification or heavy atom labeling of proteins.5 Metal
complex-assisted cleavage reactions follow either an oxidative
or a hydrolytic pathway and the latter is preferred because the
oxidative process involves the high risk of damaging the protein
chain. For better understanding of these hydrolytic cleavage
reactions, metal complex-promoted cleavage of small peptide
molecules are considered to be useful model studies. CoIII com-
plexes are known to induce the hydrolysis of peptides and
model studies were carried out to investigate the mechanism of
peptide cleavage reactions by using small peptide molecules.6
Those studies involved cis-CoIII complexes, at a slightly basic
pH, which is followed a cis-coordination by the peptide mole-
cule. Subsequently, the peptide bond undergoes cleavage by the
attack of hydroxide anions.7 The reactions proceed through the
removal of N-terminal amino acid residues,8 and in favorable
cases the amino acids of small peptide molecules can be cleaved
in a sequential manner.4 Despite a significant advance in the
studies of these type of reactions, structural characterization of
the cleaved products are not often available,9 due to the fact that
the reactions produce a mixture of different fragments of
peptides and the parent CoIII complex from which suitable
Results and discussion
When β-alanyl--histidine was treated with an equimolar
amount of [Co(tren)Cl2]ϩ at pH ∼7.5 and 45 ЊC, [Co(tren)-
(histidine]2ϩ 1 was obtained. The reaction mixture was spotted
on a silica gel plate and eluted with CH3OH–H2O (1 : 1). The
cobalt containing species remained on the starting line but
the β-alanine moved in the normal manner. The fragmented
β-alanine was separated through column chromatography
and was subjected to electron spray mass spectrometry, which
shows a single peak at m/z 90.1. Almost no spot for histidine
is observed, which illustrates the sole formation of [Co(tren)-
(histidine)]2ϩ 1. When treated with two equivalents of [Co(tren)-
Cl2]ϩ both 1 and [Co(tren)(β-alanine)]2ϩ 2 were obtained which
demonstrates the stepwise nature of the reaction. Both the
products were crystallographically analyzed; 1 consists of a
CoIII center bonded to five nitrogen atoms and one oxygen
atom, as shown in Fig. 1.
The initial binding of the substrate β-alanyl--histidine to the
CoIII complex would be important to elucidate the mechanism
of the hydrolysis reaction. β-Alanyl--histidine can bind to a
transition metal via its carboxylate oxygen, amide nitrogen or
oxygen and terminal amine group.11,12 The mode of binding
depends strongly on the metal ion, the parent amine ligand
1
skeleton and the substrate. H NMR spectroscopy is not very
D a l t o n T r a n s . , 2 0 0 4 , 1 4 6 6 – 1 4 7 3
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
1466

View Article Online
Fig. 1 Perspective drawing of 1 showing the local coordination
scheme of the cobalt atom. The large open circles represent carbon
atoms and the small open circles represent hydrogen atoms.
Illustrations of the two isomeric dangling moieties are shown in the
right hand lower corner. The inset shows the two possible geometric
isomers depending on the coordination of the histidine residue.
sensitive to β-alanyl--histidine coordination to a metal com-
plex. On the other hand the 13C NMR (165–185 ppm) and
the IR spectra (1500–1600 cmϪ1) of β-alanyl--histidine carb-
oxylate groups are very sensitive to metal coordination. The
signals for β-alanyl--histidine are well documented in the
literature.13 As the data presented here indicate, addition of
[Co(tren)Cl2]ϩ to an slightly basic (pH ∼ 7.5) aqueous solution
(temperature maintained between 5 and 10 ЊC to arrest any
cleavage reaction) of β-alanyl--histidine shifts the 13C reson-
ance of the carboxylate group from 179.1 to 181.6 ppm. The
positions of the other resonances do not change significantly.
The downfield shift of the carboxylate 13C resonance is char-
acteristic of metal ion coordination.14 (It should be noted that
as the 13C resonance positions of β-alanyl--histidine depend
on the pH,13 thus the pH of the solution should be readjusted
on addition of [Co(tren)Cl2]ϩ.) Fig. 2(a) presents the 13C NMR
spectra in the range 165–180 ppm for free β-alanyl--histidine
while Fig. 2(b) illustrates the downfield shift of the carboxylate
resonance when [Co(tren)Cl2]ϩ was added to an aqueous
solution of β-alanyl--histidine (pH ∼ 7.5). The IR spectrum
of the solid, isolated from the solution by addition of a large
volume of acetone, and washed with a small amount of ice cold
(water ϩ methanol) shows a 10 cmϪ1 shift of the ν(COO)
maxima. The blue shift as well as the position of νas(COO) at
1563 cmϪ1 are consistent with monodentate coordination of
the carboxylate group to the metal center.15 The 13C NMR and
IR characteristics demonstrate that the carboxylate oxygen of
β-alanyl--histidine attacks the CoIII center and, thus, the
intermediate [Co(tren)(β-alanyl--histidine)(OH)]ϩ is formed.
The formation of this intermediate was further confirmed
by ESI-mass spectrometry (see inset in Scheme 1). The reaction
mixture shows a strong peak at m/z 447.4 for the intermediate
[Co(tren)(β-alanyl--histidine)(OH)]ϩ along with three weak
signals at m/z 125.5, 226.2 and 240.1. These correspond to
[Co(tren)(H2O)(OH)]2ϩ, β-alanyl--histidine and [Co(tren)-
(OH)2]ϩ, respectively.
Fig. 2 The CO region (165–185 ppm) 13C NMR spectra of (a) free β-
alanyl--histidine, (b) [Co(tren)Cl2]Cl and β-alanyl--histidine system
at pH 7.5 and temperature 5 ЊC, (c) [Co(tren)Cl2]Cl and β-alanyl--
histidine system at pH 7.5 and temperature raised to 45 ЊC, allowed to
react for 15 min and reaction stopped by lowering the temperature to
5 ЊC, (d) after 30 min of reaction, (e) after 45 min of reaction.
as II rather than I (Fig. 1, inset). The crystal structure of
cleavage product 1 reinforces this assignment, as the struc-
tural study reveals attachment of a carboxylate oxygen at the
position trans to the primary amine, instead of trans to the
tertiary amine nitrogen of tren.
The temperature of the reaction mixture was brought to
45 ЊC with an aim to investigate whether any further inform-
ation regarding the cleavage pathway can be available from 13
C
NMR spectroscopy. As a function of time, the intensity of the
amide group 13C resonance12 at 172.9 ppm diminishes, with a
concomitant generation of a new resonance. The newly gener-
ated resonance appeared at 176.1 ppm, which is almost in the
same position as the 13C resonance for the carboxylate carbon
of free β-alanine at a similar pH.17 This spectral characteristics
indicate cleavage of the peptide bond and formation of a free
β-alanine molecule. Fig. 2(c)–(e) show the formation of the
resonance at 176.1 ppm, and depletion of the resonance at
172.9 ppm, when the spectra were recorded at 15 min time
intervals. Prior to the each of those measurements, the reac-
tion’s progress was quenched by lowering the temperature
below 5 ЊC. After ca. 45 min, the resonance at 172.9 ppm
had almost completely disappeared, and the growth of the
resonance at 176.1 ppm was complete.
The carboxylate oxygen of the substrate β-alanyl--histidine
can approach the Co complex in two possible orientations: (i)
the position trans to the primary amine and (ii) the position
trans to the tertiary amine group of the parent tren ligand. It
has been clearly illustrated in the literature that the chloride ion
trans to the primary amine of tren in [Co(tren)Cl2]ϩ should be
more easily aquated, because the elimination of this chloride
ion would release steric strain from the dichloride complex.
When this occurs, a rather stable trigonal bipyramidal inter-
mediate is generated prior to aquation.16 The water molecule
can then be readily displaced by the carboxylate oxygen of
β-alanyl--histidine. If so, hydrolysis would then determine
the structure of the amino acid complex [Co(tren)(histidine)]2ϩ
Through a joint investigation of the 13C resonance charac-
teristics, the IR spectra and the single crystal structure of the
cleavage product it is proposed that the substrate β-alanyl--
D a l t o n T r a n s . , 2 0 0 4 , 1 4 6 6 – 1 4 7 3
1467

View Article Online
Scheme 1
histidine initially coordinates to the metal centre through the
carboxylate oxygen atom in a monodentate fashion. This is
followed by an attack of metal bound hydroxide group to the
peptide carbonyl center, occurring simultaneously by an
approach of the nitrogen atom or oxygen atom of the amide
group to the metal center. Depending upon the mode of coord-
ination (through nitrogen or oxygen atom of the amide group)
two possible different pathways of the cleavage of peptide bond
occur, which are described in Scheme 1. The exclusive form-
ation of [Co(tren)(histidine)]2ϩ 1 from a equimolar reaction
between [Co(tren)Cl2]ϩ and β-alanyl--histidine can only be
explained by pathway 1, where the second attack has to
be through the nitrogen atom of the peptide linkage. On the
other hand, a second attack through the oxygen atom of the
peptide linkage, pathway 2, would form [Co(tren)(histidine)-
(β-alanine)]ϩ. However, the extra stability that can be achieved
through chelate ring formation may drive out either a histidine
or a β-alanine molecule from the coordination sphere, in which
case [Co(tren)(histidine)(β-alanine)]ϩ could generate both [Co-
(tren)(histidine)]2ϩ and [Co(tren)(β-alanine)]2ϩ simultaneously.
However, experimentally, the exclusive formation of [Co(tren)-
(histidine)]2ϩ rules out the possibility of pathway 2, and
establishes pathway 1 as the mechanism of this reaction.
cleavage product 1 and [Co(tren)(histidine)]2ϩ, obtained directly
by reacting the (tren)CoIII-salt with -histidine. The spectra are
shown in Fig. 3. It is evident from the spectra that the rotation
for [Co(tren)(histidine)]2ϩ, directly obtained is greater than that
of the cleavage product 1 by a very small, but measurable,
extent (peak ratio ∼1 : 0.9).
The crystal structural study of [Co(tren)(histidine)]2ϩ
1
reveals that the aromatic imidazole ring is dangling and the
chirality at the optically active carbon is , even though we
began with an -amino acid. This is a crucial observation
since it affects the mechanism of the reaction inasmuch as (a)
hydrolytic cleavage occurred with some inversion of the chiral-
ity of the active carbon atom and (b) the resulting material
underwent conglomerate crystallization in the perchlorate salt
form. Fortunately, the crystal we accidentally selected for X-ray
diffraction studies happened to be the -amino acid derivative;
and, thus, the inversion mechanism was revealed to us. Judicious
selection of the additional crystals revealed the presence of an
analogous species with the unchanged -amino acid derivative
in its crystalline form. For further elucidation of the extent of
racemization we recorded the circular dichroism spectra for the
Fig. 3 CD spectra of the cleavage product 1 (dotted line) obtained by
the reaction of [Co(tren)Cl2]Cl and β-alanyl--histidine at pH ∼ 7.5 and
the product obtained by the direct reaction between [Co(tren)Cl2]Cl
and -histidine (full line).
It is known18 that the most likely mechanism for the inversion
of the chirality of amino acids is the momentary formation of
a C᎐C bond caused by the temporary, reversible, ionization of a
᎐
proton. The temporary species resulting from such an ioniz-
ation step is planar as a result of its double bond character.
Return of the proton can occur on either side of the double
bond, resulting in the racemization of the amino acid. The
extent of the racemization depends on the nature of dangling
groups.18,19 The existence of the large histidine group inhibits
D a l t o n T r a n s . , 2 0 0 4 , 1 4 6 6 – 1 4 7 3
1468

View Article Online
Fig. 4 Perspective view of the coordination of the cobalt atom in the single-crystal structure of 2. The inset shows the two possible geometric
isomers depending of the coordination of β-alanine residue.
the attack from one particular side; thus, only a small amount
of racemization may be possible in this case.
A structural view of 2 is illustrated by Fig. 4. All bond
distances and angles are consistent with other Co() amine
amino acid complexes reported in the literature.20 The β-alanine
residue forms a six-membered ring with the metal center and
the ring is in the chair form, as expected.
For understanding the progress of the cleavage of β-alanyl--
histidine by [(tren)CoCl2]Cl, at alkaline pH, we studied the time
resolved absorption spectra of the reaction mixture. The time
resolved spectra at pH 8.5 are shown in Fig. 5. The spectra show
that the intensity of the peaks at 468 and 385 nm gradually
decrease with time while the formation of new peaks is seen to
occur at 550 and 347 nm. The extent of the increase of the peak
at 550 nm was taken as a measure of the formation of the
cleaved product. To further elucidate the mechanism, we carried
out the reaction at different pH values, in the range 7.5–12. It is
interesting to note that at pH 7.5–9.5 there is no substantial
change in the rate of formation of the cleaved product; but,
when the pH is above 10, the reaction rate was found to be
faster (Fig. 6). Unlike the 13C NMR spectra described earlier, for
the reaction at relatively low pH, the 13C NMR at pH higher
than 9.5 do not appear to be conclusive. This is due to the
formation of more than one intermediate, which may be
explained by the fact that the carboxylate group is fully ionised
at any pH above 6, whereas the amino group is mostly proton-
ated at pH below 10. Thus, under pH 9.5, reaction is solely
controlled by initial attack from the carboxylate end but at
pH > 10, the reaction takes place by initial coordination of the
Fig. 6 Increase of the peak at 550 nm (as a measure of cleavage
reaction) with time for the reaction mixture of [Co(tren)Cl2]Cl and β-
alanine--histidine at different pH. Different symbols and line style
designates different pH: (᭿, —) = 7.5, (ꢀ, - -) = 8.5, (᭹, ؒ ؒ ؒ ) = 9.5,
ؒ
(᭢, ؒ ؒ ؒ ) = 10.5 and (*, - -) = 12.
ؒ
substrate to metal center through both ends, namely carb-
oxylate as well as the amino group. Considering this fact, we
decided to subject a reaction mixture of pH 11 to analysis by a
chromatographic column hoping to isolate, and crystallize, the
amino group cleaved product. The band split initially into four
components: orange red, red, purple and orange, in that elution
order. The third band turned to be the cation [(tren)Co(H2O)-
(OH)]2ϩ from its absorption and chromatographic behavior.
The other bands were collected and crystals grown from the
first and fourth band were found to be compounds 2 and 1,
respectively. Despite several attempts, we could not isolate
single crystals from the red band. The analytical data suggest
the red band has the same composition as that of 2, which
suggests that the orange–red (first band) and the red (second
band) bands are the two isomers of the [Co(tren)(β-alanine)]2ϩ
cation, shown in the inset of Fig. 4. The existence of the two
isomers, red and orange–red, of the β-alaninato complex were
confirmed by the absorption spectra which are analogous to
those for the corresponding orange and red isomers of the
glycinato and alaninato complexes.21 It has been mentioned in
the literature that the red isomer has a lesser stability owing
to unfavorable, non-bonded, interactions between the amino
protons of the amino acid ring and the adjacent protons on the
carbon and nitrogens of tren. We believe that the instability of
this isomer is the reason of our failure to crystallize it despite
multiple attempts. We also observed that the formation ratio of
Fig. 5 Time-resolved electronic spectra of the reaction mixture of
[Co(tren)Cl2]Cl and β-alanine--histidine at pH ∼ 7.5.
D a l t o n T r a n s . , 2 0 0 4 , 1 4 6 6 – 1 4 7 3
1469

View Article Online
the red isomer to the orange–red one was rather low. It is
noteworthy that if the reaction mixture has a pH < 10, we do
not observe the existence of the second red band on the
chromatographic column. The structural elucidation of 1 and
2, the 13C NMR resonances at pH less than 10, the time
resolved absorption spectral studies, along with the isolation
and characterization of both the isomers of [Co(tren)-
(β-alanine)]2ϩ at pH higher than 10, allow us to postulate that,
at pH lower than 10, the cleavage reaction proceeds by an
initial attack at cobalt center solely by the carboxylate end of
β-alanyl--histidine. However, at pH values higher than 10, the
reaction is initiated by the attack of both the carboxylate end as
well as the amino terminus of the dipeptide molecule.
Reaction between [cis-β-Co(trien)Cl2]ϩ and β-alanyl--
histidine in an 1 : 1 molar ratio produces [cis-α-Co(trien)-
(histidine)]2ϩ 3. This was formed and crystallized as [cis-α-
Co(trien)(histidine)](ClO4)2. A TLC study of the reaction
mixture indicates the formation of β-alanine which was
chromatographically characterized and isolated. A sample was
then subjected to electron spray mass spectrometry [(M ϩ 1) =
90.1]. Interestingly, this reaction is noticeably faster than
the other cleavage reactions described here. Even at 0 ЊC, the
reaction takes place in an appreciable faster manner, whereas
the other reactions hardly proceed at room temperature
(∼20 ЊC) as evidenced by TLC. This is noteworthy, as milder
conditions for the cleavage reactions are always an important
criterion. However, when we carried out the reaction in a 2 : 1
molar ratio of [Co(trien)Cl2]Cl and β-alanyl--histidine, we
were unable to isolate the [Co(trien)(β-alanine)]2ϩ species.
Rather, column chromatographic and spectral work shows the
presence of uncoordinated β-alanine and excess [Co(trien)-
Cl2]Cl. To check the possibility of formation of [Co(trien)(β-
alanine)]2ϩ we reacted [Co(trien)Cl2]Cl directly with β-alanine
and we are able to synthesise [Co(trien)(β-alanine)]2ϩ with
perchlorate as the counter ion. However, attempts to obtain
usable single crystals were not successful; rather, we observed
the deposition of metallic cobalt which indicated the instability
of this product. The possible reaction path for the decom-
position of [Co(trien)(β-alanine)]2ϩ is not yet understood. The
electronic spectra of the β-alanine species show the gradual
disappearance of its characteristic bands, and formation of
new bands corresponding to [Co(trien)(OH/OH2)]2ϩ. It has
been observed that the product is readily decomposed in the
chromatographic column and on the TLC plate. Thus, it is
proposed that a 2 : 1 reaction essentially produces [Co(trien)-
(β-alanine)]2ϩ. However, the unstable nature of the product
meant that only free β-alanine and the amine Co complex were
isolable.
Fig. 7 Perspective view of 3. Note that the parent ligand trien is in a
cis-α stereo-arrangement even though the starting cobalt cation was cis-
β. Obviously, the reaction proceeds with isomerization at the metal
center. We have observed the same rearrangement (cis-β to cis-α) before,
however, it involved a light-induced redox reaction. For details see
ref. 30.
(histidine)]2ϩ 1, where the imidazole ring remains dangling.
Possibly, the substitution of a hydroxyl ion, or of a water
molecule, was initiated by the ring nitrogen whose bulkiness
forces the system to be in an cis-α-orientation. The chiral
carbon atom is in the -form. The circular dichroism spectrum
for 3 was recorded and is shown in Fig. 8, along with that of the
same species obtained by the direct reaction of the (trien)CoIII-
salt and -histidine. There is an appreciable amount of racemiz-
ation (peak ratio 1 : 0.75) in this case. This is quite a different
observation from that of the reaction between [Co(tren)Cl2]ϩ
and β-alanyl--histidine. This difference may be accounted by
the fact that there is a –CH2–imidazole dangling group in
[Co(tren)(histidine)]2ϩ 1 that hinders the racemization process.
By contrast, owing to the formation of a six-membered ring,
via imidazole and an –NH2 nitrogen, 3 contains a dangling
carboxylate. Such an arrangement may not be as effective in
hindering the re-entry of the proton from a specific side. The
time-resolved electronic spectra at different pH for this reaction
are unique. They show no pH dependence and the rate of the
cleavage is much faster than the cleavage reaction described
above. In fact a complete change in spectral bands was achieved
within ∼15 min (Fig. 9). The reason for this faster reaction rate
is not clear at the moment.
To investigate the initial coordination of substrate to the
metal complex, we carried out a similar 13C NMR study by
adding the substrate to a slightly basic solution of [cis-β-Co-
(trien)Cl2]Cl. Interestingly, the 13C resonance for carboxylate
at 179.1 ppm shows no significant shift which indicates the
carboxylate oxygen atoms remain non-coordinated in this
case. Rather the three carbon atoms of the imidazole show
appreciable downfield shifts (137.0 to 148.9 ppm, 134.5 to 135.3
ppm and 118.7 to 127.6 ppm) which suggest an initial co-
ordination through the imidazole nitrogen atom. Due to the
faster reaction kinetics, even at 0 ЊC, characterization of the
intermediate involved in this case was not possible. However,
in time the formation of free β-alanine was observed with
the generation of a new 13C resonance at 176.1 ppm, as well as
the depletion of that at 172.9 ppm as observed earlier, but at a
relatively faster rate.
Surprisingly, the crystal structure of 3 shows a cis-α-
oriented22 arrangement (Fig. 7) of the amine ligand. Moreover,
the carboxylate group was found to be dangling, as indicated by
the 13C resonance, whereas one of the two nitrogens of the
imidazole ring is coordinated to the metal ion. This is contrary
to the coordination arrangement for the cation [Co(tren)-
Fig. 8 CD spectra of the cleavage product 3 (dotted line) obtained by
the reaction of [cis-β-Co(trien)Cl2]Cl and β-alanine--histidine at
pH ∼ 7.5 and the product obtained by the direct reaction between
[cis-β-Co(trien)Cl2]Cl and -histidine (full line).
D a l t o n T r a n s . , 2 0 0 4 , 1 4 6 6 – 1 4 7 3
1470

View Article Online
17.56. Found: C, 25.57; H, 4.60; N, 17.42%. Yield, 1.42 g (81%).
If instead of subjecting the solution to the column for isolation
of free β-alanine, another equivalent (1 g, 0.003 mol) of
[Co(tren)Cl2]Cl is added to the reaction mixture and then eluted
with 5% HClO4 on the cation exchange column, [Co(tren)-
(β-alanine)]Cl(ClO4)ؒ2H2O 2 was isolated as a very fast moving
band. Anal. Calc. for [Co(tren)(β-alanine)]Cl(ClO4)ؒ2H2O 2:
C9H28N5O8Cl2Co: C, 23.29; H, 6.08; N, 15.09. Found: C, 23.52;
H, 5.92; N, 15.30%. Yield, 0.83 g (60%). Yield percentages were
calculated based on the amount of the substrate β-alanyl--
histidine used for reaction.
Reaction in 2 : 1 molar ratio. Reaction between 2 g (0.006
mol) of [Co(tren)Cl2]Cl and 0.7 g (0.003 mol) of β-alanyl--his-
tidine was carried out in a similar manner as mentioned above.
The solution was transferred to a column (30 × 800 mm)
containing cation exchange resin (Bio Rad AG 50w-X2 50–100
mesh, Hϩ form). After the column had been swept with 3 dm3
of water the adsorbed band was eluted with 5% HClO4. The
initial band separated into three bands: orange–red, purple and
orange, in that elution order. The second purple band is from
unreacted starting metal complex which was characterized by
electronic spectroscopy. The eluted solutions, from the other
two bands, were concentrated to a few ml in a vacuum evap-
orator at 30–35 ЊC. To this concentrated solution acetone was
added to deposit the crude complexes. Crystals were collected
for both the bands by re-crystallizing the crude product from
water at room temperature. The first band (orange–red) was
confirmed to be [Co(tren)(β-alanine)]Cl(ClO4)ؒ2H2O from the
crystal structure study. Anal. Calc. for [Co(tren)(β-alanine)]-
Cl(ClO4)ؒ2H2O 2, C9H28N5O8Cl2Co: C, 23.29; H, 6.08;
N, 15.09. Found: C, 23.38; H, 6.12; N, 15.00%. Yield, 1.2 g
(85%). Yield percentages were calculated based on the amount
of the substrate β-alanyl--histidine used for reaction.
Fig. 9 Increase of the peak at 550 nm (as a measure of the extent of
the cleavage reaction) with time for the reaction mixture of
[Co(trien)Cl2]Cl and β-alanine--histidine at different pH. Different
symbols and line styles designates different pH. (᭿, —) = 7.5, (ꢀ, - -) =
ؒ
9.5, (᭹, ؒ ؒ ؒ ) = 13.
Conclusions
It is evident that the CoII(amine) complexes effectively cleave
the peptide bond of carnosine. Depending upon the parent
ligand backbone, and the nature of the dipeptide molecule, the
reaction gives rise to products with different stereochemical
characters. In the case of the reaction between [Co(tren)Cl2]ϩ
and the carnosine molecule, initial coordination between the
metal complex and the substrate occurs by carboxylic oxygen at
a pH lower than 9.5. At a higher pH the isolation of an extra
isomer of [Co(tren)(β-alanine)]2ϩ indicates the possibility of
simultaneous approach of the substrate through both the
carboxylate and amino group.
The other band (orange) was found to be [Co(tren)(histi-
dine)](ClO4)2 1, with an similar yield as for the 1 : 1 reaction.
The analytical results are also similar to the 1 : 1 reaction.
Within a pH range of 7.5–10, the isolation of the same
products as mentioned above was observed. However, with
pH in the range of 10–12 a new isomer of the [Co(tren)-
(β-alanine)]2ϩ cation was isolated. The reaction and the product
isolation details are as follows.
In the case of [Co(trien)Cl2]ϩ and β-alanyl--histidine the
reaction is possibly initiated by the imidazole nitrogen atom.
This reaction progresses at a faster rate even at low temperature
(0 ЊC), whereas the earlier (tren) reaction eventually ceased at
that temperature.
During the course of the cleavage reaction, racemization of
an amino acid of the dipeptide molecule is often observed. It is
not clear why the cleavage product experiences a higher extent
of racemization than that for the compound when is directly
prepared by mixing the metal amine salts with optically pure
amino acids. It appears that extent of racemization depends on
the nature of the dangling group attached to the bound amino
acid.
Reaction at pH range 10–12. 1 g (0.003 mol) of [Co(tren)-
Cl2]Cl was dissolved in 60 ml of water and the pH was adjusted
to ∼11 using a freshly prepared 10% solution of lithium
hydroxide. 0.7 g (0.003 mol) of β-alanyl--histidine was added
to this solution. The solution was subjected to a column
(30 × 800 mm) containing cation exchange resin (Bio Rad AG
50w-X2, 50–100 mesh, Hϩ form). After washing with water
5% HClO4 was added for elution. The band split into four
bands, whereas at lower pH three bands were observed. The
new band is red and found to be second in the elution order. All
the bands were eluted and recrystallized using the method
mentioned earlier. The new band was found to be the isomer of
2 in which the amine group is trans to the primary amine, as
evidenced by electronic spectroscopy.21 The analytical values
obtained fit well for [Co(tren)(β-alanine)]Cl(ClO4)ؒ2H2O. Anal.
Calc. C9H28N5O8Cl2Co: C, 23.29; H, 6.08; N, 15.09. Found: C,
23.41; H, 5.93; N, 15.23%. Despite numerous attempts we
were unsuccessful in obtaining single crystal for this isomer.
The other bands have similar crystallographic, spectroscopic
and analytical characteristics as mentioned for [Co(tren)-
(β-alanine)]Cl(ClO4)ؒ2H2O (first orange–red band), unreacted
metal complex (third purple band) and [Co(tren)(histidine)]-
(ClO4)2 1 (fourth orange band).
Experimental
Syntheses: [Co(tren)Cl2]Cl23 and [Co(trien)Cl2]Cl24 were
prepared by known methods.
Reaction of [Co(tren)Cl2]Cl and ꢀ-alanyl-L-histidine
Reaction in 1 : 1 molar ratio. 1 g (0.003 mol) of [Co(tren)-
Cl2]Cl were dissolved in 60 ml of water and the pH was adjusted
to ∼7.5–8 using a freshly prepared 10% solution of lithium
hydroxide. 0.7 g (0.003 mol) of β-alanyl--histidine was added
to this solution. The solution was stirred at room temperature
for 15 min followed by heating at 45 ЊC for 2 h. After cooling
to room temperature, 0.50 g of NaClO4ؒ2H2O was added to
the solution. The solution was allowed to stand at room
temperature from which crystals of [Co(tren)(histidine)](ClO4)2
were isolated. The solution was reduced to 10 ml for further
crystallization of [Co(tren)(histidine)](ClO4)2. The crystals were
filtered off, and 10 ml of water added to the solution. The solu-
tion was subjected to a silica column and unbound β-alanine
was isolated by eluting with water. Anal. Calc. for [Co(tren)-
(histidine)](ClO4)2 1: C12H26N7O10Cl2Co: C, 25.82; H, 4.69; N,
Yield for red isomer of 2, 0.55 g (40%), yield for orange
red isomer of 2, 0.95 g (65%); yield percentage calculated based
on the amount of the substrate β-alanyl--histidine used for
reaction, Yield for 1, 1.33 g (77%).
D a l t o n T r a n s . , 2 0 0 4 , 1 4 6 6 – 1 4 7 3
1471

View Article Online
Table 1 X-Ray crystallographic data for 1(ClO4)2, 2Cl(ClO4)ؒH2O and 3(ClO4)2ؒ2H2O
Empirical formula
T /K
C12H26N7O10Cl2Co (1)
293(2)
C9H28N5O8Cl2Co (2)
293(2)
C12H30N7O12Cl2Co (3)
293(2)
Space group
P212121
P21/c
P212121
Unit cell dimensions
a/Å
b/Å
c/Å
10.4622(20)
11.0346(50)
18.3206(60)
9.1026(50)
13.6506(20)
15.1994(40)
102.499(30)
1843.86(126)
4
1.67
1.267
951.8
4374
10.8844(40)
11.4734(60)
18.1038(50)
β/Њ
V/Å3
2115.04(120)
4
1.75
1.131
1151.7
3373
3095
2260.82(160)
4
1.75
1.070
1231.7
2954
Z
Dc/g cmϪ3
µ/mmϪ1
F(000)
Reflns. collected
Indep. Reflns.
Absorption correction
Refinement method
Data/restraints/params.
GOF on F 2
Final R indices [I > 2σ(I )]
3905
2771
Empirical DIFABS
Full-matrix least squares on F 2
3095/0/298
1.018
R1 = 0.041
wR2 = 0.109
R1 = 0.055
wR2 = 0.118
0.516 and Ϫ0.306
3905/1/227
1.023
R1 = 0.066
wR2 = 0.155
R1 = 0.162
wR2 = 0.191
0.648 and Ϫ0.589
2771/0/311
1.060
R1 = 0.042
wR2 = 0.116
R1 = 0.064
wR2 = 0.126
0.691 and Ϫ0.499
R Indices (all data)
Largest diff. peak and hole/e ÅϪ3
Reaction of [cis-ꢀ-Co(trien)Cl2]Cl with ꢀ-alanyl-L-histidine
employed graphite-monochromatized Mo-Kα radiation, the
ω–2θ scan mode and were collected at room temperature
(ca. 21 ЊC). During data collection, the intensity of three
standard reflections were monitored every 3000 s of X-ray
exposure time. They showed no significant decay. The data
were processed using the WinGx25 package of programs and
an empirical absorption correction (DIFABS scans) were
employed for all the crystals. Details of data collection and
refinement are summarized in Table 1.
Reaction in 1 : 1 molar ratio. The same reaction was carried
out as described for that between [Co(tren)Cl2]Cl and β-alanyl-
-histidine. The solution was allowed to stand at room
temperature and crystals of [Co(trien)(histidine)](ClO4)2ؒ2H2O
were isolated. The solution was reduced to 10 ml for further
crystallization of [Co(trien)(histidine)](ClO4)2ؒ2H2O. The
crystals were filtered off and 10 ml of water added to the solu-
tion. Using a silica column, and eluting with water, unbound
β-alanine was isolated which was further characterized by
analytical results and electron spray mass spectrometry. Anal.
Calc. for [Co(trien)(histidine)](ClO4)2ؒ2H2O 3: C12H30N7O12-
Cl2Co: C, 24.25; H, 5.09; N, 16.50. Found: C, 24.30; H, 5.00; N,
16.61%. Yield, 1.66 g (90%), yield percentage calculated based
on the amount of the substrate β-alanyl--histidine used for
reaction)
Anal. Calc. for β-alanine: C3H7NO2: C, 40.44; H, 7.92; N,
15.72. Found: C, 40.50; H, 7.97; N, 17.63%. Yield, 0.33 g (84%),
yield percentage calculated based on the amount of the sub-
strate β-alanyl--histidine used for reaction)
The reaction with 2 : 1 molar ratio and reactions at higher pH
between [Co(trien)Cl2]Cl and β-alanyl--histidine were also
performed in which a cation exchange resin (Bio Rad AG
50w-X2, 50–100 mesh, Hϩ form) was used. Unlike the reaction
between [Co(tren)Cl2]Cl and β-alanyl--histidine, [Co(trien)-
(histidine)](ClO4)2ؒ2H2O 3 and free β-alanine were obtained
irrespective of the molar ratio and pH of the solution.
The structures were solved by direct methods using SIR-9226
and SHELXS-8627 and refined, by full-matrix least squares
based on F 2, using SHELXL-97.28 All non-H atoms were
refined anisotropically by unit-weighted full-matrix least
squares methods. The hydrogen atoms were assigned on the
basis of geometric considerations, and treated according to the
riding model, with isotropic displacements corresponding to
the atom they are linked to. No hydrogen atoms were assigned
to the oxygen atoms of the solvent water molecules, unless
they were located experimentally. 1(ClO4)2 and 3(ClO4)2ؒ2H2O
crystallize in the enantiomorphic space group, P212121, with
refined Flack parameters29 of 0.01(3) for both the compounds.
CCDC reference numbers 208929, 208930 and 208933.
See http://www.rsc.org/suppdata/dt/b4/b402571g/ for crystal-
lographic data in CIF or other electronic format.
Elemental analyses of carbon, hydrogen and nitrogen
were obtained from Galbraith Analytical Laboratories, 2323
Sycamore Drive, Knoxville, TN, 37921–1750.
IR spectra were measured on a Perkin Elmer 983 G spectro-
photometer using KBr pellets.
In control experiments, aqueous solutions of β-alanyl--
histidine were kept at similar conditions (at pH ∼ 7.5 and 45 ЊC
for 6 h) without [Co(tren)Cl2]Cl or [Co(trien)Cl2]Cl. No cleav-
1H and 13C NMR spectra were recorded at 300 MHz on a GE
QE-300 spectrometer in deuterated water. The solvent signal
1
age products were formed as evidenced by H and 13C NMR
1
(δ 4.8 ppm) was used as the H NMR internal standard. The
spectroscopy, TLC and electron spray mass spectrometry. To
check the possibility of a peptide bond cleavage, under the
influence of silica gel or cation exchange resin, used for column
chromatography, we ran the dipeptide solutions through a
13C chemical shifts were measured relative to benzene (in a
capillary) and converted into chemical shifts from TMS by
means of the relation δTMS = δbenzene Ϫ 128.5 ppm. The spectra
as a function of time were measured in 15 min intervals after
quenching the cleavage reaction by lowering the temperature of
the reaction mixture to 5 ЊC.
1
column and eluted them with water or 5% HClO4. H and 13C
NMR spectroscopy shows an identical spectrum as that known
for β-alanyl--histidine13 and was further checked by electron
spray mass spectrometry.
Electron spray mass spectra in aqueous solution were
recorded with a LCQ DECA XP mass spectrometer.
Circular dichroism spectra were recorded by AVIV
Lakewwod, NJ, CD Spectrometer Model 62A DS.
X-Ray data collection and structure refinement
Suitable crystals were glued on glass fibers and aligned on an
Enraf Nonius CAD-4 diffractometer for data collection. Unit
cell parameters were determined on the basis of 25 reflec-
tions within the range 9 < 2θ < 20Њ. Intensity measurements
Electronic spectra were recorded in aqueous solutions on
a Shimadzu model MPS-2000 multipurpose spectrophoto-
meter. Hydrolytic cleavage reactions of β-alanyl--histidine
(carnosine) in aqueous solution at pH 8 were followed by
D a l t o n T r a n s . , 2 0 0 4 , 1 4 6 6 – 1 4 7 3
1472

View Article Online
8 D. A. Buckingham, J. P. Colman, D. A. R. Happer and
UV-vis spectroscopy in the presence of the metal complexes.
Initial concentrations for the complex and β-alanyl--histidine
were both of the order of 10Ϫ3. Each time 2 cm3 aliquots
were withdrawn from the reaction mixture and immediately
frozen with liquid nitrogen to stop the progress of the reaction.
The temperature was then raised to 2 ЊC and the spectra
recorded.
L. G. Marzilli, J. Am. Chem. Soc., 1967, 89, 1082.
9 D. A. Buckingham, P. A. Marzilli, I. E. Maxwell, A. M. Sargeson,
M. Fehlmann and H. C. Freeman, Chem. Commun., 1968, 488.
10 L. G. Marzilli and D. A. Buckingham, Inorg. Chem., 1967, 6, 1042.
11 A. Torreggiani, M. Tamba and G. Fini, Biopolymer, 2000, 57, 149;
A. Torreggiani, M. Tamba and G. Fini, Biopolymer, 2000, 57, 352.
12 K. Okamoto, T. Yasui, H. Kawaguchi, T. Ama and J. Hidaka, Chem.
Lett., 1988, 2, 335; C. E. Brown and W. E. Antholine, Biochem.
Biophys. Res. Commun., 1979, 88, 529.
13 J. O. Friedrich and R. E. Wasylishen, Can. J. Chem., 1986, 64, 2132.
14 E. Gaggelli and G. Valensin, J. Chem. Soc., Perkin Trans. 2, 1990,
401.
15 K. Nakamoto, Infrared and Raman Spectra of Inorganic and
Coordination Compounds; Part B, Wiley-Interscience, New York, 4th
edn., 1978.
Acknowledgements
We thank the Robert A. Welch Foundation for support of these
studies. We also thank the Institute for Mexican American
Studies (UH) for their support. Drs John S. Olson and Susan
Cates (Rice University) kindly assisted us with the CD spectral
studies.
16 S. K. Madan, W. M. Reiff and J. C. Bailar, Jr., Inorg. Chem., 1965, 4,
1366.
17 H. W. E. Rattle, Annu. Rep. NMR Spectrosc. A, 1981, 11, 2.
18 W. E. Keyes, R. E. Caputo, R. D. Willett and J. I. Legg., J. Am.
Chem. Soc., 1976, 98, 6939.
19 D. H. Williams and D. H. Busch, J. Am. Chem. Soc., 1965, 87, 4644.
20 J. Cai, X. Hu, X. Feng, W. Shao, L. Ji and I. Bernal, Eur. J. Inorg.
Chem., 2000, 2199; F. M. Akhter, M. Hirotsu, I. Sugimoto,
M. Kojima, S. Kashino and Y. Yoshikawa, Bull. Chem. Soc. Jpn.,
1996, 69, 643.
21 Y. Mitsui, J. Watanabe, Y. Harada, T. Sakamaki, Y. Iitaka, Y. Kushi
and E. Kimura, J. Chem. Soc., Dalton Trans, 1976, 2095.
22 cis-α-trien is the isomer of an octahedral complex in which the two
terminal –NH2 binding points are trans to each other; cis-β-trien is
that isomer in which the terminal –NH2 binding points are cis to
each other.
23 D. A. Buckingham, P. J. Cresswell and A. M. Sargeson,
Inorg. Chem., 1975, 14, 1485.
24 F. Basolo, J. Am. Chem. Soc., 1948, 70, 2634.
25 L. Farrugia, J. Appl. Crystallogr., 1999, 32, 837.
26 A. Altomare, G. Cascarano, C. Giacovazzo and A. Gualardi,
J. Appl. Crystallogr., 1993, 26, 343.
27 G. M. Sheldrick, SHELXS-86, University of Göttingen, Göttingen,
Germany, 1986.
References
1 M. K. Saha and I. Bernal, Chem. Commun., 2003, 612.
2 H. Chun, B. J. Salinas and I. Bernal, Eur. J. Chem., 1999, 723.
3 P. B. Dervan, Nature, 1992, 359, 87; A. H. Kortz, L. Y. Kuo,
T. P. Shields and J. K. Barton, J. Am. Chem. Soc., 1993, 115, 3877;
C. C. Cheng, S. E. Rokita and C. J. Burrows, Angew. Chem., Int. Ed.
Engl., 1993, 32, 277; N. Gupta, N. Grover, G. A. Neyhart, P. Singh
and H. H. Thorp, Inorg. Chem., 1993, 32, 310; D. P. Mack and
P. B. Dervan, Biochemistry, 1992, 31, 9399; S. F. Singleton and P. B.
Dervan, J. Am. Chem. Soc., 1992, 114, 6957.
4 (a) A. Schepartz and B. Cuenoud, J. Am. Chem. Soc., 1990, 112,
3247; (b) D. Hoyer, H. Cho and P. G. Schultz, J. Am. Chem. Soc.,
1990, 112, 3249; (c) T. M. Rana and C. F. Meares, J. Am. Chem. Soc.,
1990, 112, 2457.
5 (a) N. Ettner and W. Hillen, J. Am. Chem. Soc., 1993, 115, 2546;
(b) N. Ettner, J. W. Metzger, T. Lederer, J. D. Hulmes, C. Kisker,
W. Hinrichs, G. A. Ellestad and W. Hillen, Biochemistry, 1995, 34,
22; (c) S. A. Strobel and P. B. Dervan, Science, 1991, 254, 1639.
6 (a) D. A. Buckingham and J. P. Colman, Inorg. Chem., 1967, 6, 1803;
(b) E. Kimura, S. Young and J. P. Colman, Inorg. Chem., 1970, 9,
1183.
7 D. A. Buckingham and C. R. Clark, in Metal ions in Biological
Systems, ed. A. Siegel and H. Siegel, Marcel Dekker, Basel, New
York, 1976, vol. 38, pp. 43–98.
28 G. M. Sheldrick, SHELXL-97, University of Göttingen, Göttingen,
Germany, 1997.
29 H. D. Flack, Acta. Crystallogr., Sect. A, 1993, 39, 876.
30 I. Bernal and James Cetrullo, Inorg. Chim. Acta, 1986, 120, 109.
D a l t o n T r a n s . , 2 0 0 4 , 1 4 6 6 – 1 4 7 3
1473