Molecular recognition of a tris(histidine) ligand

Article abstract of DOI:10.1039/a706711i

Design and synthesis of a tri-Hg²⁺ complex to selectivity recognize a tris(histidine) ligand is presented.

Full text of DOI:10.1039/a706711i

View Article Online / Journal Homepage / Table of Contents for this issue
Molecular recognition of a tris(histidine) ligand
Shuguang Sun, Jennifer Saltmarsh, Sanku Mallik*†‡ and Kathryn Thomasson
Department of Chemistry, Box: 9024, University of North Dakota, Grand Forks, ND 58202, USA
cyclam
Design and synthesis of a tri-Hg2+ complex to selectivity
recognize a tris(histidine) ligand is presented.
NH HN
i–iv
v
R (88%)
cyclam
NH HN
We are interested in the design and synthesis of transition metal
ion based receptors for histidine-containing peptides.1 Histi-
1
4 (34%)
cyclam
vi, vii
dine-to-metal ion interactions (Cu2+, Ni2+, Zn2+ etc.) have been
NH2
L R = CH2NH2 (48%)
C R = H (63%)
used for various applications, e.g. protein purification,2 cross-
linking,3 targeting proteins to lipid bilayers.4 It is these strong,
directed metal ion-to-histidine interactions that we are using as
the basis of the recognition process.5
R
R
viii
L′ R = CH2NH2 (45%)
C′ R = H (60%)
5
As a model system for peptide recognition, we have chosen
the ligand L to position three (S)-histidines 12 Å apart (Fig. 1).
The compound C, with one histidine, served as the control for
our studies. LA and CA were tested as histidine-mimetic ligand
and control, respectively. Three-dimensional structures for the
receptor and ligands were constructed using the molecular
modelling software INSIGHT II and DISCOVER (ver. 95.0,
BioSym Technologies/MSI, San Diego, CA) and energy-
minimized in the gas phase using the consistent valence force
field (cvff).
CH2Br
Boc
BocHN
HO2C
N
EtO2C
N
N
N
BrCH2
CH2Br
6
7
8
Scheme 1 Reagents and conditions: TsCl, Et3N, CHCl3; ii, 6, K2CO3,
MeCN, sonication; iii, HBr, AcOH, 70 °C; iv, ion exchange; v,
Hg(ClO4)2.3H2O, MeOH–MeCN; vi, N-hydroxysuccinimide, DCC, Et3N,
THF, 7; vii, NaOH; viii, 8, EtOH, sonication
Syntheses of the receptor R, ligands L, LA and the control C,
CA are shown in Scheme 1. Selective protection of three
nitrogens of the cyclam (1,4,8,11-tetraazacyclotetradecane)
ring 1 was carried out following a literature procedure.6
Reaction of cyclam tritosylate 2 with 1,3,5-tris(bromo-
methyl)benzene 67 proceeded smoothly in MeCN (using
powdered K2CO3 as the base) under sonication (8 h). The crude
product was purified by flash chromatography, using 5%
MeOH–CH2Cl2 as the solvent (Rf 0.3). This reaction was found
to yield a complex mixture of products under reflux. Removal of
the tosyl groups was carried out in HBr–AcOH at 70 °C (10 h).
Free ligand 4 was isolated by ion-exchange chromatography
(IRA-400 column, hydroxide form) using water as eluent.
Receptor R was synthesized by adding a solution of the free
ligand 4 in MeCN to a methanolic solution of Hg(ClO4)2.3H2O.
Receptor R was isolated as the air-stable perchlorate salt after
addition of Et2O to the reaction mixture. (Note: we did not
observe any explosive tendency for this compound.)§
H
BocHN
HN
H
N
N
NHBoc
O
N
H
N
H
H
BocHN
O
N
H
HN
O
HN
N
L
H
H
N
Ligand L is known in the literature8 and control C was
synthesized by an analogous procedure. Reactions for the
synthesis of LA and CA gave higher yields under sonication
compared to refluxing conditions. Control CA was purified by
recrystallization from CHCl3–hexane and LA was purified by
flash chromatography using MeOH as the eluent (Rf = 0.3).§ In
the binding studies, a diamagnetic metal ion (Hg2+) with strong
affinity for histidine ( > 103 m21) was used so that the binding
could be monitored by 1H NMR spectroscopy. Since the
receptor R contains the embedded distance information, cyclam
was used to hold the metal ions. Cyclam has very high affinities
( > 1020 m21) for transition metal ions. This ensures that the
metal ions will not get displaced from R at high histidine
concentrations. Cyclam also gives us the flexibility to synthe-
size the receptor with a variety of transition metal ions and to
optimize the recognition properties.
N
N
O
HN
HN
N
N
H
O
HN
H
N
N
L′
O
H
N
H
BocHN
H
N
H
N
N
N
O
O
C′
C
NH HN
Hg2+
N
HN
Recognition studies were conducted in highly polar
[2H6]DMSO, and were followed by 1H NMR spectroscopy. The
C-2-H of the imidazole moieties (indicated in Fig. 1) of C and
CA were found to be shifted downfield (0.80 ppm for C; 0.87
ppm for CA) upon complexation with the Hg2+ ions of the
receptor R (10 mm in R, 0–60 mm in C or CA). Both of these
controls were in fast exchange with the receptor and an average
signal was observed in each case. Resultant titration curves are
NH HN
Hg2+
NH HN
Hg2+
N
HN
NH
N
R
Fig. 1 Structures of the tris(histidine) ligand L, histidine mimetic ligand LA,
controls C, CA and the designed receptor R. Hydrogens monitored in the
titration studies are circled.
Chem. Commun., 1998
519

View Article Online
1.0
0.8
0.6
0.4
0.2
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
[R] / [His]
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
[Receptor]
t
t
/ [Im]
t
t
Fig. 2 Titration curves for (.)C, («)CA and (2) LA. The curves indicate the
calculated titration curves with the reported binding constants.
Fig. 3 Fraction of bound (5) C and (:) CA in the competition experiments.
The corresponding fractions bound in the titration experiments for (2) C
and («) C’ are also plotted.
shown in Fig. 2. The turning of the titration curves at a 3 :1 ratio
of [R]t/[His]t indicated a 3:1 stoichiometry of binding between
R and C (or CA).9
Studies are currently underway to optimize the selectivity by
changing the spacer of the tris(histidine) ligand L.
This work is supported by grants from the NIH (1R15 GM/
OD 54321-01) and from the NSF (CAREER award to S. M.).
For data analysis, the three metal ions were taken as
interacting independently. Also shown in Fig. 2 are the
calculated titration curves with the best-fit estimates of the
binding constants. Non-linear regression analysis of the binding
data following a previously-developed procedure5 (SIGMA
PLOT 4.0 for Windows, Jandel Scientific Inc.) provided the
Notes and References
† E-mail: mallik@badlands.nodak.edu
‡ Present address: Department of Chemistry, North Dakota State Uni-
versity, Fargo, ND 58105, USA
value of the binding constants (KRC = 1.1 3 104 m21
,
,
§ All new compounds gave satisfactory characterization data. Selected data
for 4 (J values in Hz): glassy solid (80%); mp 88–90 0C; dH(D2O) 7.15 (s,
3H), 3.47 (s, 6H), 2.71–2.41 (m, 22H), 1.76–1.61 (m, 4H), 1.55–1.42 (m,
4H). HRMS (M+) Calc. for C39H78N12: 714.6472. Found: 714.6471. For R:
white solid (88%); mp 165–167 0C; dH([2H6] DMSO): 7.09 (s, 3H), 5.36 (br
s, 2H), 5.25 (br s, 2H), 4.58 (br s, 2H), 4.19 (d, 2H, J 15.0), 3.85 (d, 2H, J
15.0); the rest of the hydrogens appear as multiplets between 3.05–2.99,
2.88–2.83, 2.47–2.20, 1.90–1.77 and 1.72–1.67. Calc. for
C39H78N12Hg3(ClO4)6.3H2O: C, 23.80; H, 4.30; N, 8.54. Found: C, 24.01,
H, 4.29; N, 8.51%. For L: white foamy solid; dH([2H6]DMSO) 7.55 (s, 3H,
Im-C2H), 6.99 (s, 3H, Im-C5H), 6.87 (s, 3H, Ar-H), 4.21 (3H, Ca-H), 4.17
(6H, ArCH2), 2.80 (m, 6H, His- b-CH2), 1.35(s, 27H, But); dC([2H6]DMSO)
171.5, 155.2, 139.2, 134.7, 124.3, 78.1, 54.6, 42.1, 33.4, 28.2. HRMS M+
Calc. for C42H60N12O9: 876.4605. Found: 876.4610. For C: white solid
(85%); TLC (Rf 0.24, 3% MeOH–CH2Cl2); mp 150–152 °C; [a]2D1 +58
(MeOH, c 7.6); dH([2H%]DMSO) 7.53 (s, 1H, Im-C2H), 7.20 (m, 5H, Ar),
6.75 (s, 1H, Im-C5H), 4.24 (s, 2H, ArCH2N), 4.18 (m, 1H, Ca-H), 2.82 (m,
2H, His-b-CH2), 1.35 (s, 9H, But); dC([2H6DMSO) 171.6, 155.1, 139.4,
134.6, 128.1, 126.8, 126.5, 78.1, 54.6, 41.9, 28.1; HRMS (MH+) Calc. for
C18H24N4O3: 345.1926. Found 345.1911.
[R]t = 10.25–6.5 mm; [C]t = 0–32 mm; KRC’ = 104 m21
[R]t = 9.6–6.6 mm; [C’]t = 0–32 mm; in both regressions,
error: < 10%). The regression analysis converged to these
numbers starting from either a smaller or a larger value as the
initial estimate.
Ligand LA was also in fast exchange with the receptor (Fig. 2).
The sharp turning of the titration curve at [R]/[LA] = 1 indicated
a 1:1 stoichiometry of the complex and a high affinity. Due to
the high affinity, only a lower limit of K can be estimated from
the binding data (KRLA > 105 m21).
Similar titration experiments (10 mm in R, 3–30 mm in L)
showed that R interacts differently with L compared to C. The
ligand L was found to be in slow exchange with R.10 Two
different C-2-H signals were observed for the free (d 7.67) and
bound (d 8.605) ligand. The amounts of free L (measured by the
integration of bound and free C-2-H resonances) were very
small up to 1:1 stoichiometry and then the amount of free L
increased rapidly. The aromatic hydrogens of L were shifted
upfield by 0.8 ppm in the presence of the receptor R. These
observations indicated that R is forming a 1:1 complex with L
and that the benzene rings of R and L are stacking. This was
corroborated by the observance of a cross-peak between the
aromatic ring hydrogens of the two benzene rings of R and L in
a NOESY spectra and by molecular modeling. The binding
constant was estimated from the integration of bound and free
signals of L.11 Owing to inherent errors in the integration of
very small peaks in 1H NMR spectra, only a lower limit of KRL
can be obtained (KRL > 105 m21).
In order to determine the binding selectivity of R and L (or
LA) over C (or CA), competitive titration experiments were
conducted (Fig. 3).5 A 10 mm solution of R.L (or R.LA) was
titrated with C or CA (1–30 mm). The C-2–H chemical shift of C
(or CA) was followed to measure the concentration of C (or CA)
bound to R. The fraction of R bound to L (or LA) compared to
the fraction of R bound to C (or CA) was taken as the measure
of selectivity. R was found to be selective for L compared to C
by a factor of 20; its selectivity for LA over CA was 10. This
difference in selectivity may be due to the difference in inter-
imidazole distances of L and LA and the resultant strain in the
complexes arising from non-optimal distance matching be-
tween the receptor and the ligand. We are currently probing this
by synthesizing tris(histidines) with varying inter-imidazole
distances.
1 S. Mallik and I. Mallik, Synlett, 1996, 734. For examples of amino acid
recognition based metal–ligand interactions, see K. Konishi, K. Yahara,
H. Toshishige, T. Aida and S. Inoue, J. Am. Chem. Soc., 1994, 116,
1337; T. Mizutani, T. Ema, T. Yoshisa, Y. Kuroda and H. Ogoshi, Inorg.
Chem., 1993, 32, 2072.
2 A. V. Terskikh, J. M. Ledoussal, R. Crameri, I. Fisch and J. P. Mach,
Proc. Natl. Acad. Sci. USA, 1997, 94, 1663; A. Gambero, L. T. Kubota,
Y. Gushikem, C. Airoldi and J. M. Granjeiro, Colloid Interfacial Sci.,
1997, 185, 313.
3 D. A. Fancy, K. Melcher, S. A. Johnston and T. Kodadek, Chem. Biol.,
1996, 3, 551.
4 K. M. Malony, D. R. Shnek, D. Y. Sasaki and F. H. Arnold, Chem. Biol.,
1996, 3, 185; L. Schmitt, T. M. Bohanon, S. Denzinger, H. Ringsdorf
and R. Tempw, Angew. Chem., Int. Ed. Engl., 1996, 35, 317.
5 S. Mallik, R. D. Johnson and F. H. Arnold, J. Am. Chem. Soc., 1994,
116, 8902.
6 I. M. Helps, D. Parker, J. R. Morphy and J. Chapman, Tetrahedron,
1989, 45, 219.
7 W. P. Cochrane, P. L. Pauson and T. S. Stevenes, J. Chem. Soc. (C),
1968, 630.
8 D. Goldfarb, J.-M. Fauth, Y. Tor and A. Shanger, J. Am. Chem. Soc.,
1991, 113, 1941.
9 R. S. Macomber, J. Chem. Ed., 1992, 69, 375; Z.-X. Wang, N. Ravi
Kumar and D. K. Srivastrava, Anal. Biochem., 1992, 206, 376.
10 This may or may not indicate a very strong association constant.
A. Bosti, B. Perley and E. Hadjoudis, J. Chem. Soc., Perkin Trans. 2,
1997, 89.
11 K. A. Connors, Binding Constants, Wiley, New York, 1987, pp.189–
216.
Thus, the designed receptor R is indeed selective for pattern
matched tris(histidine) ligand L compared to the control C.
Received in Corvallis, OR, USA, 10th September 1997; 7/06711I
520
Chem. Commun., 1998