A cyclodecapeptide ligand to vitamin B12

Article abstract of DOI:10.1039/b811234g

Libraries of cyclic decapeptides were screened with vitamin B₁₂ derivatives to give cyclic peptide ligands incorporating histidine and cysteine as coordinating residues and negatively charged amino acids. Two hits, cyclo-(HisAspGluProGlyIleAlaThrProdGln) and cyclo- (ValAspGluProGlyGluAspCysProdGln) were resynthesized in good yields for solution experiments. The peptides bind aquocobalamin with coordination of His or Cys to the cobalt with high affinities (K_a～10⁵ M ^-1). Additional interactions between the peptide side chains and the vitamin B₁₂ corrin moiety were determined by studying the ¹H NMR solution structure. The cyclopeptide-cobalamin complex with the histidine residue showed enhanced stability towards cyanide exchange, demonstrating the shielding effect of the ligand on the metal center.

Full text of DOI:10.1039/b811234g

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
A cyclodecapeptide ligand to vitamin B †
1
2
a
b
b
b
a
a
Vincent Dul e´ ry, Nicolas A. Uhlich, No e´ lie Maillard, Viviana S. Flux a´ , Julian Garcia, Pascal Dumy,
a
b
b
Olivier Renaudet,* Jean-Louis Reymond* and Tamis Darbre*
Received 1st July 2008, Accepted 8th August 2008
First published as an Advance Article on the web 16th September 2008
DOI: 10.1039/b811234g
Libraries of cyclic decapeptides were screened with vitamin B12 derivatives to give cyclic peptide ligands
incorporating histidine and cysteine as coordinating residues and negatively charged amino acids. Two
hits, cyclo-(HisAspGluProGlyIleAlaThrProDGln) and cyclo-(ValAspGluProGlyGluAspCysProDGln)
were resynthesized in good yields for solution experiments. The peptides bind aquocobalamin with
5
-1
coordination of His or Cys to the cobalt with high affinities (K ~10 M ). Additional interactions
1
a
between the peptide side chains and the vitamin B12 corrin moiety were determined by studying the H
NMR solution structure. The cyclopeptide–cobalamin complex with the histidine residue showed
enhanced stability towards cyanide exchange, demonstrating the shielding effect of the ligand on the
metal center.
7
Introduction
peptides to assemble so-called RAFT-proteins. We considered
the recently reported solid-supported RAFT synthesis starting
Proteins binding to porphyrin-type cofactors create a large shell
around them that affects their reactivity and stability for transport
and catalysis. Coordination of these cofactors to smaller peptide-
based models of these proteins provides useful systems to study
their essential properties. While many such model-systems have
been reported based on linear peptides and dendrimers, there
are only very few studies of cyclic peptide ligands, which only
concern heme so far. Herein we report the first cyclic peptide
complexes of vitamin B12, a corrin-containing cofactor bound
in vivo to transport proteins and enzymes. The complexes show
enhanced stability compared to small molecule analogs and a
reduced reactivity towards nucleophiles, as observed for related
8
with side-chain attachment at a D-glutamate residue for the solid-
phase synthesis of a split-and-mix combinatorial library. The D-
glutamate residue at position X is part of one of the two conserved
turn dipeptides (qP at X X and GP at X X ), separating two
variable tripeptides X X X and X X X (see Scheme 1). The
amino acids at the variable positions were assigned following our
recently reported “unique pair” algorithm called TAGSFREE.
This algorithm creates libraries in which peptide sequences can be
assigned from the amino acid composition analysis. The method
is particularly advantageous for cyclic peptides, which are not
9
1
1
2
6
7
3
4
5
8
9
10
1
2
10
3
4
11
amenable to Edman sequencing.
The 15 625 member “unique pair” combinatorial library with
5
1
dendritic peptide ligands. The H NMR-structure of the complex
shows a cobalt–cysteine residue bond and additional multiple
contacts between the cyclodecapeptide and the corrin ring. This
cyclic peptide features a remarkably simple and well-defined model
of B12-binding proteins. Peptide-basedligands for vitamin B12, such
as the cyclodecapeptides reported here, might find application as
adjuvants for B12-formulations for facilitating uptake by patients
5
amino acids at each of the 6 variable positions was designed
for potential B12-binding by coordination of the cobalt with
a histidine or cysteine side-chain, by placing these residues at
variable positions on either side of the RAFT framework (cysteine
at positions X and X and histidine at positions X and X ).
The other 13 amino acids allowed by the unique pair design
were distributed along the variable positions to cover a variety
of charged, hydrophobic, aromatic, small and polar residues.
The library was prepared using Fmoc-chemistry on a 1 g batch
3
9
8
10
6
deficient in this intrinsic factor.
-
1
Results and discussion
of Rapp Polymers TentaGel Macrobeads (0.3 mmol g , 1.6 ¥
5
10 beads per gram) allowing 10-fold coverage of sequence space.
Ligand library design and synthesis
The synthesis was completed by removal of the a-N-Fmoc group,
deallylation of the a-carboxyl group of the D-glutamate at position
X , and on-bead cyclization using PyBOP–DIEA in DMF (2 ¥
Cyclic decapeptides incorporating a pair of b-turn inducing
Pro-Gly dipeptides display a preorganized structure originally
designed as an anchoring point for the attachment of further
1
6
0 min). In the final step of the synthesis, side-chain protecting
groups were removed by acidic treatment, leaving the on-bead
library of cyclic decapeptides ready for activity screening.
a
D e´ partement de Chimie Mol e´ culaire – UMR-CNRS 5250 & ICMG FR
2
607, Universit e´ Joseph Fourier, BP 51, 38041, Grenoble Cedex 9, France
b
Department of Chemistry and Biochemistry, University of Berne, Freies-
trasse 3, CH-3012, Berne, Switzerland. E-mail: tamis.darbre@ioc.unibe.ch;
Fax: +41 31 631 80 57; Tel: +41 31 631 43 70
On-bead binding assay and ligand synthesis
The cyclodecapeptide library was assayed for aquocobalamin
binding. A portion of the library (10 mg) was first equilibrated
in aqueous PBS (20 mM phosphate, 150 mM NaCl, pH 7.4), and
then incubated with the vitamin, followed by extensive washing
†
Electronic supplementary information (ESI) available: Plots of exchange
kinetics with cyanide, HPLC traces of bead analysis, tables for non-hits
and sequences picked randomly, analytical HPLC profiles and +ESI-MS
data. See DOI: 10.1039/b811234g
4
134 | Org. Biomol. Chem., 2008, 6, 4134–4141
This journal is © The Royal Society of Chemistry 2008

View Article Online
-1
5
Scheme 1 Synthesis and structure of the cyclic peptide library. The solid support is Rapp TentaGel Macrobeads (0.3 mmol g , 1.6 ¥ 10 beads per g).
n
6
Using five a-N-Fmoc protected amino acids at positions X (n = 3–5, 8–10) gives a library of 5 = 15 625 members. Conditions: (a) 3 eq. Fmoc protected
amino acid, 3 eq. PyBOP, 6 eq. DIEA, DMF, 2 ¥ 60 min. (b) Ac O–CH Cl (1 : 1), 15 min. (c) 20% piperidine in DMF, 3 ¥ 10 min. (d) The resin batch
2
2
2
was suspended in DMF–CH
2
Cl
2
(2 : 1, v/v), mixed via nitrogen bubbling for 15 min, and the batch was split into five equal portions 1–5 by volume.
a
a
(
e) In each portion X
b
(b = 1–5): 3 eq. Fmoc-X
b
3 3 4 2 2
-OH, 3 eq. PyBOP, 6 eq. DIEA, DMF, 2 ¥ 60 min. (f) PhSiH , Pd(PPh ) , dry CH Cl , under argon,
2
¥ 30 min. (g) PyBOP, DIEA, DMF 2 ¥ 45 min (h) TFA–TIS–H
2
O–EDT (94 : 1 : 2.5 : 2.5, v/v) 3 h.
10
9
8
5
4
3
and visual detection of stained beads. Under optimized assay
conditions (100 mM aquocobalamin in PBS at 25 C for 2.5 hours),
Table 1 Hits from cyclodecapeptide library X X X -Pro-Gly-X X X -
◦
Pro-gln
only very few beads remained stained (less than 2%, Fig. 1).
Deep red colored beads were picked and subjected to amino acid
analysis. From 24 beads, 18 returned a readable sequence (Table 1).
While control beads picked at random from the library showed
no sequence preference (Table S3 in the ESI†), the B12-stained
beads were rich in anionic (Glu at X /X and Asp at X /X ) and
hydrophobic residues (Leu, Val, Ile). In twelve sequences cysteine
residues were present while histidine was found in one sequence
as the only possible coordinating residue and in two with cysteine.
Five beads gave no coordinating residues but incorporated the
anionic residues as well. On the other hand, cationic and aromatic
residues were almost absent (Table 1, Table 2). The obtained results
indicate the possible coordination of cysteine to cobalt as a binding
feature. Coordination to cobalt was also suggested by the absence
of hits when the library was screened with cyanocobalamin or
methylcobalamin, where the ligands in the upper side can not be
easily exchanged.
1
0
9
8
5
4
3
X
X
X
X
X
X
1
2
3
4
5
Ser
Ser
His
Val
Val
Val
Val
Val
His
Ser
Val
Val
Val
Val
Val
Val
Val
Val
Asp
Tyr
Asp
Asp
Cys
Ser
Ser
Asp
Cys
Ile
Glu
Leu
Glu
Glu
Glu
Glu
Leu
Glu
Thr
Glu
Glu
Glu
His
Glu
Glu
Ile
Glu
Glu
Ile
Arg
Asp
Ala
Asp
Leu
Asp
Asp
Asp
Leu
Trp
Asp
Asp
Ala
Ala
Asp
Leu
Ala
Asp
Cys
Cys
Thr
Cys
Thr
Thr
Ala
Thr
Ala
Cys
Cys
Ala
Cys
Cys
Cys
Thr
Cys
Cys
a
b
a
a
5
8
4
9
c
6a
a
d
7
8
9
10
11
1
1
1
Glu
Val
e
a
Glu
Glu
Glu
Glu
Glu
Glu
Glu
Glu
Glu
Glu
Ile
Ile
2
f
3a
4a
Asp
Asp
Ser
Ile
Ser
Ile
g
h
Leu
Leu
Glu
Tyr
15a
1
1
1
6 i
7aj
8a
Leu
a
b
Alternative sequences from decoding: SDYEKC; HITEDA (3b);
The cyclodecapeptide sequences 3a, 3b with His, 4, 11 with Cys,
c
d
e
f
SIEVDT; SDEVLA; HCLEAT;j HDEVAC, VCHEDA, HCEVDA;
9a with both coordinating residues and sequence 12 without His or
g
h
i
VCLEDA; SDEVLC; STEVAC; VDEILC.
Cys were selected as representative hits from the screening. They
were synthesized on Rink-amide resin using on-bead cyclization as
above, followed by acidic deprotection and HPLC purification of
the cleaved cyclic peptides, providing the material in good yields
and purity (Table 3). For comparison, the corresponding linear
sequences (4¢ and 4¢Ac) were also synthesized and isolated by
HPLC. In addition, we prepared the mutant 4His of peptide 4 in
which the coordinating cysteine residue is replaced by histidine.
Although the screening method gave mainly cyclic peptides with
cysteine, the His containing ligands better simulate the binding
mode of aquocobalamin in the transport protein transcobalamin,
where a histidine residue is coordinated on the upper phase of the
cofactor.
12
Binding affinities to B12
The binding affinities of the synthesized hits 3a, 3b, 4, 9a, 11 and
12 for aquocobalamin were determined by UV-titration (Fig. 2,
Table 4). The cyclic peptides 4 and 11 coordinate the cobalamin
with the cysteine residue as evidenced by the UV changes whereas
This journal is © The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 4134–4141 | 4135

View Article Online
Table 2 Hit percentage occurrence of the amino acid at the given position: H% (P%), H% is the percentage occurrence of the amino acid at the given
position in picked hits; the relative occurrence P% is the observed frequency of the amino acid at the given position across a random analysis of 24 beads
(
see ESI†)
10
9
8
5
4
3
No.
AA
X
X
X
X
X
X
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
Tyr
His
Phe
Cys
Glu
Ile
Lys
Asp
Ser
Val
Ala
Leu
Thr
Trp
Arg
6% (36%)
6% (25%)
3% (21%)
14% (18%)
0% (21%)
0% (21%)
50% (5%)
17% (20%)
31% (18%)
67% (14%)
72% (14%)
8% (38%)
0% (5%)
0% (25%)
36% (13%)
11% (14%)
47% (13%)
28% (7%)
58% (21%)
1
1
1
1
1
1
19% (18%)
17% (5%)
25% (32%)
22% (9%)
19% (16%)
6% (23%)
28% (39%)
0% (32%)
6% (18%)
6% (32%)
0% (25%)
Table 3 Sequences, isolated synthetic yields (mg, %) and +ESI-MS data
for the cyclic peptides and the linear controls, including the N-acetylated
linear and mutant 4His sequences
Table 4 Binding affinities for the cyclodecapeptides and linear analogs
to aquocobalamin, determined by UV-titration
5
-1
Peptide
Structure
a
K /10 [M ]
Yield/
+ESI-MS +ESI-MS
Peptide
Structure
mg (%) (calcd)
(found)
3a
c-(HDEPGIATPq)
c-(HITPGEDAPq)
c-(VDEPGEDCPq)
VDEPGEDCPGNH
AcVDEPGEDCPGNH
c-(HCTPGELAPq)
5.8 ± 1.7
6.5 ±.1.2
13 ± 5.1
16 ± 5.7
13 ± 8.0
32 ± 9.5
3
4
4¢
b
3
3
4
4
4
4
4
4
9
1
1
a
b
c-(HDEPGIATPq)
c-(HITPGEDAPq)
c-(VDEPGEDCPq)
c-(VDEPGEDHPq)
VDEPGEDCPGNH
AcVDEPGEDCPGNH
VDEPGEDHPGNH
41 (79)
24 (47)
46 (86)
32 (58)
46 (82)
40 (76)
35 (66)
38 (69)
18 (35)
48 (90)
39 (76)
1046.5
1046.5
1070.4
1104.5
1016.4
1058.4
1050.5
1092.5
1034.5
1068.5
1036.5
1046.6
1046.4
1070.6
1104.6
1016.4
1058.4
1050.4
1092.5
1034.5
1069.0
1036.5
2
4¢Ac
9a
2
His
¢
¢Ac
His¢
2
2
5,13
2
(
Table 4). Linear controls showed comparable binding affinities
His¢Ac AcVDEPGEDHPGNH
2
to the cyclic peptides, which possibly reflects the dominant effect
of the acidic amino acids and the cysteine residues present on
linear and cyclic peptides. Additionally, the b-turn inducing diad
Pro-Gly may induce some preorganization in the linear analogs.
Peptide 9a, with both His and Cys, shows coordination with the
cysteine thiol group as expected by the higher affinity of thiol
a
1
2
c-(HCTPGELAPq)
c-(VIEPGEDCPq)
c-(VIEPGEDAPq)
14
(
thiolate) for Co(III) when compared to imidazole (Fig. 2B).
Structure determination of the cobalamin–cyclodecapeptide 4
complex (Cbl–4)
As already observed in peptide dendrimers, amino acid sequence
discrimination was also found in the cyclic peptides, with gluta-
mate and aspartate residues prevailing in the hits (Table 2). We
postulate that the acidic residues could contribute to binding
by interactions with the peripheral side chains of corrin. Such
interactions could pre-organize the metal cofactor and peptide
or contribute to the binding affinity. In order to investigate the
occurrence of peripheral binding interactions, the structure of the
Cbl–4 complex was analyzed by NMR-spectroscopy.
Fig. 1 Structure of vitamin B12 derivatives, and on-bead binding assay
with aquocobalamin (a portion of the screened batch is seen).
3
a and 3b showed the shifts expected for imidazoyl coordination.
The amino acid signals of the cyclic peptide were assigned by
sequential attribution. The chemical shifts of the two cysteine Hb
protons in the B12 complex (1.50 and 0.46 ppm, Fig. 3) show
lower shielding when compared to the free cyclodecapeptide (3.28
and 2.96 ppm), supporting cobalt coordination at that residue.
The conformation of the cysteine side chain was determined by
distance constraints for Cys3-H /Cys3-H and Cys3-H /Cys3-
b a b
NH. Furthermore, five distance constraints were unambiguously
determined between the peptide and the cobalamin. Interestingly,
Addition of hit 12 did not induce changes in the UV-spectrum
of aquocobalamin, showing that cobalt coordination with side
chain carboxylic groups is not taking place, although peripheral
interactions of the peptide with the corrin cannot be excluded.
The cyclic peptides 4 and 11 showed affinities to aquocobalamin
comparable to the natural tripeptide glutathione and linear
controls, and lower than values previously reported for dendritic
ligands of B12, which are 2- to 5-fold higher than for glutathione
4
136 | Org. Biomol. Chem., 2008, 6, 4134–4141
This journal is © The Royal Society of Chemistry 2008

View Article Online
Fig. 4 Structural model for the cobalamin–cyclodecapeptide 4 complex
(
Cbl–4) and the structure of cyclic peptide 4. Dashed lines represent
H-bonds.
are established between the corrin carboxamide side chains and
several residues of the cyclodecapeptide. Intermolecular H-bonds
involve the carbonyl groups of D-glutamine q1, aspartate D4, and
proline P2, with H62, H40 and H29 of cobalamin, respectively. An
additional H-bond between the carboxyl group of glutamate E8
and O51 is also observed because the minimization was performed
with protonated side chain carboxylic acids. Van der Waals
contacts are furthermore predicted between the methyl group of
valine V10 and the methyl groups Me-54 and Me-46 of cobalamin.
The intermolecular interactions with acidic amino acids correlate
with the high frequency of glutamate residues in the hits. It is also
interesting that in the X-ray structure of the transport protein Cbl–
transcobalamin complex, hydrogen-bonding interactions with Gln
and Asp transcobalamin side chains and corrin carboxamide
Fig. 2 Representative UV-titration for binding of cyclodecapeptides
to aquocobalamin: (A) histidine-cyclodecapeptide 3a;(B) histidine and
cysteine-cyclodecapeptide 9a.
12
groups are also observed. The large number of additional
direct and solvent-mediated contacts in the transcobalamin–Cbl
complex probably explains the much higher affinity of the protein
to vitamin B12 compared with the cyclodecapeptides, also taking
into account the entropically favourable dissolvation of B12 in the
protein complex, an effect which is much more limited in the B12
–
cyclodecapeptide complex formation due to the limited surface
interaction.
Cyanide exchange kinetics
Fig. 3 NOESY-NMR spectrum of the Cbl–4 complex in H
90 : 10). Intramolecular NOEs are in red and intermolecular NOEs are
in blue.
2 2
O–D O
The stability of the B12-cyclodecapeptide complexes was evaluated
more closely in the case of cyclic peptide 4 and its histidine analog
(
4
His by determining the rate of ligand exchange against cyanide,
the intermolecular NOEs involve the side chains of the three
amino acids occurring in higher percentages in the 18 sequences
in Table 1; Glu8 is present in 10 hits, Glu5 in 15 and Val10 in
which is reduced in the presence of macromolecular coordination
15
such as by B12-transport proteins. The cysteine coordinating
peptide 4 and analogs showed exchange rates comparable to
the natural tripeptide glutathione, known to bind strongly to
cobalamins (data in the ESI†). In the case of the natural
proteins the cobalt center is coordinated to a histidine side chain
and therefore 4His is a better analog to the natural system.
1
3 sequences. Minimization under these constraints gave a stable
˚
model with no distance violations larger than 0.1 A (Fig. 4).
1
In the H NMR-model obtained, the cyclodecapeptide shields
the upper corrin face almost completely, and H-bonding contacts
This journal is © The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 4134–4141 | 4137

View Article Online
The histidine cyclodecapetide 4His reduced the rate of ligand
exchange by more than 10 fold (Fig. 5), showing that a significant
stabilization of the complex was achieved compared to imidazole,
suggesting significant interactions of the cyclodecapeptide with
the corrin peripheral groups besides the coordinating residue, in
agreement with the solution structure (Table 5). Exchanges with
linear analogs, with free or acetylated N-termini, were 6 times
slower than imidazole.
of Cys containing cyclic peptides as ligands for aquocobalamin
could have a thermodynamic and a kinetic origin. Thiolates bind
to Co(III) with higher affinity than imidazoles. It is also known that
the rate of complexation of nitrogen containing ligands to Co(III)
is lower than that of other ligands. In the case of the on-bead
screening, the kinetic factor may be predominant since the time of
incubation is not very long. Therefore, more cysteine containing
cyclopeptides were found in the screening assay, although thiolates
14
and imidazoles are both known ligands for Co(III).
The cyclodecapeptide–Cbl complexes show stability compara-
ble to the tripeptide glutathione–B12 complex. Glutamate residues
were present in the sequences obtained by on-bead screening, an
observation already made with peptide dendrimers that gave hits
5
with polyglutamate shells. However a slower exchange rate of
the water axial ligand on the beads was observed with the cyclic
peptides, with 2–3 h incubation time required. In comparison,
the peptide dendrimers gave Cys containing hits after 15 min.
Further studies of the exchange kinetics in the two systems are
1
being performed. The H NMR-structure of the complex revealed
hydrogen-bonding interactions with the peripheral carboxamides
on cobalamin and residues on the peptide ligand, besides metal
coordination. The cyclodecapeptide–cobalamin complex 4His–
Cbl displays enhanced kinetic stability against ligand exchange
by cyanide when compared to linear analogs and imidazole.
To date only very few model systems of B12-binding proteins
1
6
Fig. 5 Conversion of imidazole–Cbl and 4His–Cbl to cyanocobalamin.
have been reported in the literature. These include antibodies
◦
5
Absorption was measured at 560 nm at pH 7.5, 25 C. Initial rates
and dendritic peptide ligands. Cyclodecapeptide ligands repro-
duce essential features of previous model systems, yet display a
much simpler structure and are particularly easy to synthesize.
Screening combinatorial libraries of cyclic peptides with both
coordinating residues (Cys and His) enabled the discovery of
the first cyclopeptide ligands incorporating either Cys or His.
Further investigations on focused libraries with only histidine as
the binding residue might give new models for the metal–peptide
interaction present in the transport protein transcobalamin. The
cysteine containing cyclic peptides however are probably better
candidates as adjuvants for vitamin B12 due to their faster and
stronger complexation to cobalt ions. Further studies with both
systems are now in progress.
-1
v/mM min : 0.34 for the conversion of imidazole–Cbl and 0.02 for
His–Cbl. Both curves were measured until complete displacement of the
4
upper ligand by cyanide.
Conclusion
The experiments above describe the first complex between a
vitamin B12 derivative and a cyclic peptide. The discovery of
this cyclic peptide was made possible by the availability of an
efficient combinatorial chemistry protocol for library synthesis,
including the on-bead formation of cyclic peptide, a selective
binding assay, and a powerful bead-decoding protocol enabled
by the TAGSFREE algorithm for library encoding. The key
binding interaction occurs by coordination of a nucleophilic
residue such as cysteine or histidine to cobalt. The prevalence
Experimental
General methods
Table 5 Cyanide exchange rates with His-peptides in cobalamin
complexes
ꢀ
R
Protected amino acids and Tentagel Macrobeads resin were
obtained from Advanced ChemTech Europe (Brussels, Belgium),
Bachem Biochimie SARL (Voisins-Les-Bretonneux, France), and
France Biochem S.A. (Meudon, France). PyBOP was purchased
from France Biochem and other reagents were obtained from
either Aldrich (Saint Quentin Fallavier, France) or Acros (Noisy-
Le-Grand, France). RP-HPLC was performed on Waters equip-
ment with a 600 controller and a Waters 2487 dual absorbance
detector. The peptide derivatives were analyzed on an analytical
column (Macherey-Nagel Nucleosil 120 A 3 mm C18 particles,
0 ¥ 4.6 mm, flow rate of 1.3 mL min for 15 minutes gradient,
Nucleosil 100 A˚ 5 mm C particles, 250 ¥ 4.6 mm, flow rate of
mL min for 30 minutes gradient) using a linear gradient and the
following solvent system: water containing 0.1% TFA (solvent A);
acetonitrile containing 0.1% TFA and 9.9% H O (solvent B). UV
t
1/2 for 50% ligand
a
Ligand in Cbl complex
exchange
4
4
4
His
c-(VDEPGEDHPq)
VDEPGEDHPGNH
AcVDEPGEDHPGNH
2
7–8.5 h (87 mM)
2 h 46 min (87 mM)
3 h (87 mM)
30 min (87 mM)
<1 s (2.5 mM)
50 min (1.0 mM)
20 s (0.5 mM)
His¢
2
His¢Ac
Imidazole
b
H O
TC
IF
HC
2
b
Transcobalamin
Intrinsic factor
Haptocorrin
c
˚
c
1 min (0.1 mM)
-
1
3
a
Half-life t1/2 of the ligand–B12 complex in the presence of the indicated
1
8
◦
cyanide concentration (in parentheses). Conditions: [Cbl] = 25 mM; 37 C,
-1
1
◦
b
pH 7.5 for proteins; 25 C, pH 7.0–7.5 for other ligands. data from ref
c
1
4a; data from ref 14b.
2
4
138 | Org. Biomol. Chem., 2008, 6, 4134–4141
This journal is © The Royal Society of Chemistry 2008

View Article Online
absorbance was monitored at 214 nm and 250 nm simultaneously.
Bead analysis and decoding
TM
˚
Semi-preparative column chromatography (Delta-Pak 100 A
5 mm C18 particles, 200 ¥ 2.5 mm) was used to purify crude
peptides by using an identical solvent system at a flow rate of
Single-peptide-containing resin beads were hydrolyzed with aque-
1
◦
ous HCl (6 M) at 110 C for 22 h, and their amino acid composition
-
1
was determined quantitatively by HPLC after derivatization with
phenyl isothiocyanate (PITC). The sequences were then assigned
2
2 mL min . The electron spray ionization mass spectrometry
(
+ESI-MS) was recorded on a VG Platform II (Micromass). The
1
0
by TAGSFREE decoding.
analysis was performed in the positive mode for peptide derivatives
using 50% aqueous acetonitrile as the eluent.
Decapeptide SPPS
The on-bead decapeptides were synthesized on Rink Amide
Library synthesis
ꢀ
R
MBHA resin (0.1 mmol scale) with an automated synthesizer
433A from Applied Biosystem) using Fmoc/tBu chemistry. 10 eq.
excess of protected amino acids, HBTU coupling reagent and
0 eq. excess of DIEA were used. The FastMoc 0.1 mmol program
(
The on-bead cyclic-decapeptides library was synthesized on Tent-
ꢀ
R
-1
agel Macrobeads resin (1.0 g, loading: 0.3 mmol g ) using the
split-and-mix procedure. The resin was acylated with 3 eq. of N-
a-Fmoc amino acid in the presence of PyBOP (312 mg, 0.6 mmol,
2
TM
(
SynthAssist version 3.1) was used with a single coupling. Cyclic
peptides 3a, 4, and 4His were cyclized on-bead after Fmoc and
allyl deprotection as described above in the library synthesis
3
eq.) and diisopropylethylamine (DIEA) (260 mL, 1.5 mmol, 6 eq.)
in DMF. After 2 ¥ 60 min the resin was washed (3 ¥ 5 mL each) with
DMF, CH Cl , and MeOH and controlled with the Kaiser and
section. The resin was treated for 2–3 h with TFA–H
2
O–iPr
3
SiH
2
2
(
95 : 2.5 : 2.5, v/v) for 3a and 4His or TFA–H O–iPr
2
3
SiH–1,2-
TNBS (trinitrobenzenesulfonic acid) tests followed by acetylation.
The Fmoc protecting groups were removed by treatment with
piperidine, 20% in DMF (3 mL, 3 ¥ 10 min), and then washed with
ethanedithiol (94 : 2.5 : 1 : 2.5, v/v) for 4. After evaporation and
precipitation in diethyl ether, the cyclic peptides were purified by
semi-preparative RP-HPLC. Peptide 3a: 41 mg (79%); t
R
= 5.6 min
16 1046.5,
= 7.9 min
19S 1070.4,
DMF, CH
a-amine deprotection. The resin was suspended in DMF : CH
2 : 1, v/v), mixed via nitrogen bubbling for 15 min and distributed
2
Cl
2
, and MeOH. Split and mix was carried out after
(
5–100% B in 15 min); +ESI-MS: calcd for C45
H
68
N O
13
2
Cl
2
+
found: m/z 1046.6 [M + H] . Peptide 4: 46 mg (86%); t
R
(
(
5–40% B in 15 min); +ESI-MS: calcd for C43
H
64
N O
11
in five equal portions for the next coupling. After the last coupling
step, the Fmoc protecting group was removed. The resin bearing
the library of decapeptides (loading: 0.21 mmol g , quantified by
measuring the absorption of fluorenylmethylpiperidine adduct at
+
found: m/z 1070.6 [M + H] . Peptide 4His: 32 mg (58%); t
R
=
-
1
6.6 min (5–40% B in 15 min); +ESI-MS: calcd for C H N O
46
66
13 19
+
1
104.5, found: m/z 1104.6 [M + H] . Linear peptides 4¢ and 4His¢
were cleaved by treatment of half of the resin with TFA–H
2
O–
2
99 nm) was dried under vacuum for 1 h and swollen in a glass
reactor fitted with a sintered glass frit with dry CH Cl (10 mL, 2 ¥
5 min) and dry DMF (10 mL, 1 ¥ 15 min). The resin was treated
with PhSiH (2.6 mL, 21 mmol) in dry CH Cl (10 mL) for 5 min;
Pd(PPh (61 mg, 0.052 mmol) was then added and the resin was
stirred under argon for 20 min. The reagents were removed by
filtration, the resin washed with CH Cl (10 mL, 4 ¥ 1 min) and
DMF (6 mL, 3 ¥ 1 min) and the procedure was repeated once.
The resin was finally washed with CH Cl (10 mL, 2 ¥ 1 min),
dioxane–water (9 : 1, 10 mL, 2 ¥ 1 min), DIEA 5% in DMF (6 mL,
¥ 1 min) and DMF (10 mL, 2 ¥ 1 min). On-bead cyclization
iPr
iPr
3
SiH–1,2-ethanedithiol (94 : 2.5 : 1 : 2.5, v/v) or TFA–H
SiH (95 : 2.5 : 2.5, v/v) respectively. Linear acetylated peptides
2
O–
2
2
1
3
4¢Ac and 4His¢Ac were obtained by treatment of the second half of
3
2
2
resin with Ac O–pyridine–DMF (1 : 2 : 7) solution before cleavage.
2
)
3 4
After evaporation and precipitation in diethyl ether, the linear
peptides were purified by semi-preparative RP-HPLC. Peptide 4¢:
2
2
4
6 mg (82%); t
R
= 5.0 min (5–100% B in 15 min); +ESI-MS:
+
calcd for C40
H
62
N
11
O
18S 1016.4, found: m/z 1016.4 [M + H] .
2
2
Peptide 4¢Ac: 40 mg (76%); t
R
= 5.7 min (5–100% B in 15 min);
+
ESI-MS: calcd for C42
H
64
N O
11
19S 1058.4, found: m/z 1058.4
3
+
[
M + H] . Peptide 4His¢: 35 mg (66%); t
R
= 4.5 min (5–100%
18 1050.5, found:
= 6.0 min
19 1092.5,
was performed following the acylation procedure described above.
The protecting groups were removed on-bead by treatment with
B in 15 min); +ESI-MS: calcd for C44
H
66
N
12
O
+
m/z 1050.4 [M + H] . Peptide 4His¢Ac: 38 mg (69%); t
R
a TFA–H
solution for 2–3 h. This procedure was repeated once. The reagents
were removed by filtration and the resin washed with CH Cl
10 mL, 5 ¥ 1 min), MeOH (10 mL, 2 ¥ 1 min) and CH Cl (10 mL,
¥ 1 min).
2
O–iPr
3
SiH–1,2-ethanedithiol (94 : 2.5 : 1 : 2.5, v/v)
(
5–100% B in 15 min); +ESI-MS: calcd for C45
H
66
13
N O
+
found: m/z 1092.5 [M + H] . HPLC and +ESI-MS data for the
other peptides prepared and isolated using the same procedure are
found in the ESI.†
2
2
(
5
2
2
UV–vis titrations
On-bead B12-binding assay
0 mg of dry resin containing the cyclodecapeptide library was
1
000 mL of a ca. 20 mM buffered solution (exact concentration was
3
-1
-1
1
determined by measuring absorption: e351 = 25 000 dm mol cm
kept in PBS buffer (20 mM, 150 mM NaCl, pH 7.4) for swelling
for 30 minutes. The buffer was then removed and 4 mL of
aquocobalamin solution (100 mM) in PBS buffer were added. After
stirring for 2.5 hours, the aquocobalamin solution was removed
by filtration and the resin was washed six times with PBS buffer.
The resin was suspended in PBS, spread on a silica plate and
the most intensively red colored beads were picked for sequence
determination.
in 20 mM HEPES buffer, pH 7.0) of aquocobalamin was prepared
in a gas tight cuvette. The titration was carried out by adding
portions of a stock solution of peptide (1.25 mM) with a gas
tight syringe and waiting until no further change in the spectrum
took place. Data points were collected at 351 nm for Cys peptides
(equilibration time 1.5–3 h) and at 358 nm for the UV–vis titrations
of His containing peptides (4 h equilibration time) and fitted using
17
a single site binding equation.
This journal is © The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 4134–4141 | 4139

View Article Online
Cyanide exchange kinetics
coupling constants extracted from a DQF-COSY spectrum, and
interproton Ha–Hb distances determined from ROESY spectrum.
Restrained energy minimizations were performed by using the
Insight II/Discover software (Version 2005, Accelrys, San Diego,
CA, USA), using the set of distance and dihedral restraints
determined by NMR. The selected force field was ESFF, and, to
shorten the range of Coulomb interaction, a distance-dependent
The peptide complexes were formed prior to the exchange
measurement by dissolving the peptide (10 mM) in a solution of
aquocobalamin (1 mM in HEPES buffer pH 7.5). The imidazole
complexes were formed in the same way. All exchange experiments
◦
were performed at 25 C and in parallel with an imidazole
measurement for direct rate comparison.
relative dielectric constant, e
r
, was used (e
r
= 4r). The structure
For the exchange measurement HEPES buffer (975 mL, 20 mM,
pH 7.5) containing 87 mM NaCN was mixed with peptide complex
solution (25 mM) in a cuvette. The concentration of aquocobal-
amin, determined by measuring the absorption at lmax/nm: 352,
served as a basis for the starting concentrations of the complexes.
After the addition the conversion of the cobalamin complexes
to cyanocobalamin was detected by measuring the absorption
at 560 nm over time. Under these conditions, according to the
observed absorbance spectrum, only cyanocobalamin (lmax/nm:
was subjected to 2500 iterations of steepest descent minimization,
followed by 2500 iterations of conjugate gradient minimization
and the convergence of minimization was followed until the RMS
-1
derivative was less than 0.01 kcal mol .
Acknowledgements
This work was financially supported by the University of Berne,
the Swiss National Science Foundation, the Swiss Federal Office
for Science and Education, the “Universit e´ Joseph Fourier (UJF-
Grenoble), the “Centre National de la Recherche Scientifique”
361, 550, 522) is formed.
The initial rates of cyanocobalamin formation were obtained
by linear regression using the first 20 data points. The t1/2 of
cyanocobalamin formation was determined by calculating the
mean value between the absorbance at t = 0 and the absorbance at
cyanocobalamin saturation and reading the corresponding time.
Cyanide exchange kinetics of the Cys containing peptides
were performed in the same way with the following changes:
concentration of Cys-peptide: 2.5 mM; HEPES buffer pH 7.0;
(
CNRS) and the COST action D34. We are grateful to NanoBio
program for access to the facilities of the Synthesis platform.
Notes and references
1 (a) A. Lombardi, F. Nastri and V. Pavone, Chem. Rev., 2001, 101, 3165–
3
189; (b) F. V. Cochran, S. P. Wu, W. Wang, V. Nanda, J. G. Saven, M. J.
Therien and W. F. DeGrado, J. Am. Chem. Soc., 2005, 127, 1346–1347;
c) G. Ghirlanda, A. Osyczka, W. Liu, M. Antolovich, K. M. Smith,
P. L. Dutton, A. J. Wand and W. F. DeGrado, J. Am. Chem. Soc., 2004,
-
concentration of CN : 81 mM; every exchange experiment was
(
done in parallel with a glutathione measurement for direct rate
comparison; absorption was measured at 361 nm under oxygen-
free conditions.
1
26, 8141–8147; (d) H. K. Rau, N. DeJonge and W. Haehnel, Angew.
Chem., Int. Ed., 2000, 39, 250–253.
2
3
(a) M. Uyemura and T. Aida, J. Am. Chem. Soc., 2002, 124, 11392–
1
4
1403; (b) D.-L. Jiang and T. Aida, Prog. Polym. Sci., 2005, 30, 403–
22; (c) D. Paul, H. Miyake, S. Shinodo and H. Tsukube, Chem.–Eur. J.,
NMR-spectroscopy and modeling
2006, 12, 1328–1338.
(a) M. M. Rosenblatt, J. Wang and K. S. Suslick, Proc. Natl. Acad. Sci.
U. S. A., 2003, 100, 13140–13145; (b) F. V. Cochran, S. P. Wu, W. Wang,
V. Nanda, J. G. Saven, M. J. Therien and W. F. DeGrado, J. Am. Chem.
Soc., 2005, 127, 1346–1347; (c) M. M. Rosenblatt, D. L. Huffman, X.
Wang, H. A. Remmer and K. S. Suslick, J. Am. Chem. Soc., 2002, 124,
12394–12395; (d) D. L. Huffman, M. M. Rosenblatt and K. S. Suslick,
J. Am. Chem. Soc., 1998, 120, 6183–6184.
NMR spectra were obtained at 500 MHz with a Varian Unity Plus
spectrometer. Samples were dissolved in a mixture of 95% H
2
O–5%
2
2
◦
H
2
O or 99.9% H
2
O to a final concentration of 1 mM at 25 C. A
18
set of two-dimensional (2D) spectra, including TOCSY, DQF-
19
20
21
COSY, ROESY and NOESY were acquired with 2 s steady
state recovery times, mixing times (t ) of 60 ms for TOCSY and
50 ms for NOESY. Water suppression was achieved by appending
4
5
(a) Chemistry and Biochemistry of B12, ed. R. Barnejee, J. Wiley and
Sons, New York, 1999; (b) K. L. Brown, Chem. Rev., 2005, 105, 2075–
m
2
2
149; (c) W. Buckel, C. Kratky and B. T. Golding, Chem.–Eur. J.,
22
an excitation-sculpting module to the non-selective detection
pulse and with selective Gaussian-shaped pulses of 3–5 ms. Spectra
were acquired in the phase sensitive mode using TPPI, or States
2006, 12, 352–362; (d) E. Nexø, in Vitamin B12 and B12-Proteins, ed.
B. Kra u¨ tler, D. Arigoni and B. T. Golding, pp. 461–475, Wiley-VCH,
Weinheim, 1998.
P. Sommer, N. A. Uhlich, J.-L. Reymond and T. Darbre, ChemBioChem,
2008, 9, 689–693.
23
24
method. The spin-lock mixing of the TOCSY experiment was
25
obtained with DIPSI-2 pulse trains at gB
acquisitions were performed over a spectral width of 10 ppm in
both dimensions, with matrix size of 1024–2048 data points in t
and 256–512 points in t , and 32–128 scans/t . All spectra were
referenced with external TSP-d . Data processing and analysis
2
/2p = 9–10 kHz. The
6 L. R. Solomon, Blood Rev., 2007, 113–130.
7
For a selected review, see: D. Boturyn, E. Defrancq, G. T. Dolphin,
J. Garcia, P. Labb e´ , O. Renaudet and P. Dumy, J. Pept. Sci., 2008, 14,
2
2
24–240.
1
1
8 (a) O. Renaudet and P. Dumy, Org. Biomol. Chem., 2006, 4, 2628–2636;
(
b) E. Garanger, D. Boturyn, O. Renaudet, E. Defrancq and P. Dumy,
4
J. Org. Chem., 2006, 71, 2402–2410.
were performed using Felix software (version 2001, Accelrys, San
Diego, CA, USA) with shifted (60–90 degrees) square sinebell
apodization and polynomial baseline correction for NOESY data.
Approximate interproton distance restraints were calculated
9
(a) A. Furka, F. Sebesty e´ n, M. Asgedom and G. Dib o´ , Int. J. Pept.
Protein Res., 1991, 37, 487–493; (b) K. S. Lam, S. E. Salmon, E. M.
Hersh, V. J. Hruby, W. M. Kazmierski and R. J. Knapp, Nature, 1991,
3
54, 82–84; (c) R. A. Houghten, C. Pinilla, S. E. Blondelle, J. R. Appel,
C. T. Dooley and J. H. Cuervo, Nature, 1991, 354, 84–86; (d) K. S. Lam,
M. Lebl and V. Krchnak, Chem. Rev., 1997, 97, 411–448.
using the isolated two-spin approximation relationship, rij
=
1
/6
r
kl (skl/sij
)
, where sij and skl are the NOE intensities for
10 (a) J. Kofoed and J.-L. Reymond, J. Comb. Chem., 2007, 9, 1046–1052;
(b) J. Kofoed and J.-L. Reymond, Chem. Commun., 2007, 4453–4455.
1
the atom pairs i, j and k, l separated by distances rij and
1 V. S. Fluxa and J.-L. Reymond, Bioorg. Med. Chem., 2008, Jan 30,
article in press.
2 J. Wuerges, G. Garau, S. Geremia, S. N. Fedosov, T. E. Petersen and L.
Randaccio, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 4386–4391.
r
kl, respectively. Cross-peaks from cysteine Hb were used as
calibrants. Stereospecific assignment for the cysteine b-methylene
1
3
protons were achieved by combining qualitative analysis of JHa–Hb
4
140 | Org. Biomol. Chem., 2008, 6, 4134–4141
This journal is © The Royal Society of Chemistry 2008

View Article Online
1
1
3 For glutathione and dipeptide–cobalamin complexes see: R. K. Suto,
17 H. Bakirci, X. Zhang and W. M. Nau, J. Org. Chem., 2005, 70, 39–46.
18 L. Braunschweiler and R. R. Ernst, J. Magn. Reson., 1983, 53, 521–528.
19 U. Piantini, O. W. Sørensen and R. R. Ernst, J. Am. Chem. Soc., 1982,
104, 6800–6801.
20 C. Griesinger and R. R. Ernst, J. Magn. Reson., 1987, 75, 261–71.
21 J. Jeener, B. H. Meier, P. Bachmann and R. R. Ernst, J. Chem. Phys.,
1979, 71, 4546–4553.
N. E. Brasch, O. P. Anderson and R. G. Finke, Inorg. Chem., 2001, 40,
2
686–2692.
4 (a) J. M. Pratt, Inorganic Chemistry of B12, Academic Press, London,
972, pp. 139–147, 217–218; (b) L. Hannibal, S. D. Bunge, R. van Eldik,
D. W. Jacobsen, C. Kratky, K. Gruber and N. E. Brasch, Inorg. Chem.,
007, 46, 3613–3618.
1
2
1
1
5 (a) S. N. Fedosov, N. U. Fedosova, E. Nexø and T. E. Petersen, J. Biol.
Chem., 2000, 275, 11791–11798; (b) S. N. Fedosov, L. Berglund, N. U.
Fedosova, E. Nexø and T. E. Petersen, J. Biol. Chem., 2002, 277, 9989–
22 T. L. Hwang and A. J. Shaka, J. Magn. Reson., 1995, 112, 275–279.
23 D. Marion and K. W u¨ thrich, Biochem. Biophys. Res. Commun., 1983,
113, 967–974.
24 D. J. States, R. A. Haberkorn and D. J. Ruben, J. Magn. Reson., 1982,
48, 286–292.
25 A. J. Shaka, C. J. Lee and A. Pines, J. Magn. Reson., 1988, 77, 274–293.
9
996.
6 R. B. Hannak, R. Konrat, W. Sch u¨ ler, B. Kra u¨ tler, M.-T. M. Auditor
and D. Hilvert, Angew. Chem., Int. Ed., 2002, 41, 3613–3616.
This journal is © The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 4134–4141 | 4141