able for calixarene complexation) for Phe and Tyr, and possibly
for Trp (which is too large to fit the cavity of SC[4]A), Gly
was used as a blank. Ala was also chosen in view of the
future possibility of the construction of a soluble model of
the α-helical peptides known to bind heparin, thus permitting
“model” studies of the heparinoid behaviour of SC[4]A. The
structures and ionisations states at pH 2 and 8 of the amino-
acids used in this study are given in Fig. 1.
As shown in Table 1, Tyr and Phe, the two amino acids
with simple aromatic side chains exhibit in mobile phase A
essentially identical KA values of 433 and 448 MϪ1 respectively.
Complexation of mono- and para-di-substituted benzene rings
within the cavity of calix[4]arene is well known15 and the com-
plex of SC[4]A with the trimethylanilinium ion is known, with
an apparent KA value of 5600 MϪ1.6 Direct comparison with
this value is difficult due to the difference in the pH values
between the two measurements. For Trp, in which the aromatic
side chain is sterically much larger, the KA value observed is, as
expected, smaller than those of Phe and Tyr in mobile phase A:
282 MϪ1. The two positively charged amino acids Lys and Arg
show the highest KA values of 1209 and 1149 MϪ1 respectively. It
is apparent that in mobile phase A, the favourable electrostatic
interactions between the positively charged amino acids and the
negatively charged sulfonatocalix[4]arene lead to tight ion-pair
binding in the complex. As expected in the case of Asp where
unfavorable electrostatic interactions will be present, the lowest
KA was observed, 171 MϪ1. A KA of 541 MϪ1 is observed for
His, intermediate between Pro and the positively charged amino
acids. The amino acid Gly, chosen as a negative blank as
it contains no side chain, shows a low KA of 180 MϪ1 by
RP-HPLC. Surprisingly in mobile phase A, Ala shows quite
strong binding to SC[4]A, with a KA of 648 MϪ1.
The KA and capacity factors values for mobile phase B are
summarised in Table 2. Comparing with the values obtained
in mobile phase A, the complexation behaviour of the amino
acids can be divided into two groups; firstly, those for which the
KA value remains the same or decreases slightly Gly (128 vs. 180
MϪ1), Phe (588 vs. 448 MϪ1), Tyr (210 vs. 433 MϪ1) and Lys
(1221 vs. 1209 MϪ1) and those for which a very strong increase
in KA is observed, Trp (1518 vs. 282 MϪ1), Pro (1138 vs. 285
MϪ1) and Arg (2587 vs. 1149 MϪ1). It is evident that the inter-
action mechanism for these last three amino acids must be
different from the other amino acids and also change between
the two mobile phases. Looking at the structural similarities
between Trp, Pro and Arg (Fig. 1) it is tempting to ascribe these
differences to the presence in these three amino acids of a
secondary amino function. However, it is probably better to
treat Pro differently from Arg and Trp. For Arg and Trp a
double interaction involving both the secondary amine and π–π
interactions between the aromatic groups of the SC[4]A macro-
cycle with, respectively, the guanidinium function of Arg and
the aromatic ring of Trp, are probably responsible for the
changes in KA, with the change in solvent polarity as the driving
force.
Fig. 4 Model of proline–SC[4]A complexation.
In the case of His, a much smaller variation in KA is observed
between the two mobile phases (541 MϪ1 in mobile phase A and
717 MϪ1 in mobile phase B). This may be explained by the more
polar nature of the imidazole ring compared to the Pro five-
membered ring.
Two of the amino acids bear simple aromatic side chains: Phe
and Tyr. In the case of Phe essentially no change in the KA value
occurs between the two mobile phases; here the driving force
for inclusion will be simple aromatic–aromatic interactions.
However, for Tyr the KA value decreases in the more polar,
H-bonding solvent of mobile phase B. This must arise from
solvation of the phenolic OH group of Tyr, reducing the
capacity for inclusion.
For Lys it may be that a balance in the energy between ion-
pair interactions arising in mobile phase A and hydrophobic
interactions coupled with weakened electrostatic interactions in
mobile phase B leads to an apparent lack of change in the
observed KA values (1209 MϪ1 and 1221 MϪ1). As expected for
Gly, which does not possess a side chain capable of interaction
with the cavity of SC[4]A a small decrease in KA is consistent
with a diminution in the importance of ion-pair effects in
mobile phase B.
Again relatively strong binding is observed for Ala which has
a KA of 675 MϪ1. The change between phase A and phase B
is similar to that observed for Phe, suggesting a simple hydro-
phobic association of the short alkyl side chain is the driving
force for binding. The crystal structure of the Lys–SC[4]A
complex12 shows that the Lys side chain is folded on binding
into the cavity. Therefore, it might be expected that Pro, in
which the five-membered ring forms the side chain, could mimic
the folded Lys conformation, and thus bind tightly to SC[4]A.
1H NMR studies of the complexation of Pro with SC[4]A show
that this amino acid indeed interacts with the host at pH 2 and
pH 8. The association constants determined at pH 2 and 8 are
respectively 100 and 70 MϪ1. The KA is somewhat higher than
those observed by Arena for other amino acids, but is an order
of magnitude lower than those values previously observed by us
for Lys. For the three amino acids studied by NMR at pH 2 and
pH 8, the KA values at pH 2 are always higher. This observation
arises from the removal of repulsive interactions between the
anionic carboxylate group and the anionic sulfonate functions
of the calixarene.
For Pro in this more polar medium, the hydrophobic inter-
actions between the folded conformation of the cyclic amino
acid and the SC[4]A cavity give rise to a KA similar to that of
Lys (1138 vs. 1221 MϪ1), which suggests that in this medium,
the role of the electrostatic interactions is less important than
that of the hydrophobic interactions. Interestingly we have
recently observed18 that for the two mixed dipeptides Arg-Lys
and Lys-Arg, the Lys is in both cases prefentially included in
the SC[4]A cavity, in spite of introduction of an unfavourable
electrostatic interactions for Arg-Lys, in which the Lys frag-
ment carries a terminal carboxylate group; again this results
from the hydrophobic interactions in the side chain. Once more
in mobile phase B, the weakest complexation is with the acidic
Splitting of the non-equivalent δ protons of Pro was
observed in the NMR experiment at pH 2. This implies that Pro
adopts a specific conformation in the cavity of SC[4]A (Fig. 4).
The splitting increases with increasing concentration of SC[4]A
implying that on inclusion into the calixarene cavity, the
δ protons are situated in quite different environments. In con-
trast, the β protons show a decrease in the chemical inequiv-
alence (Fig. 5). If in the case of lysine it was the δ and ε protons
that had the highest chemical shifts, then in the case of proline,
these are now the γ and δ protons. Comparison with the X-ray
structure of the lysine complex shows that one of the δ protons
1
amino acid Asp with a KA of 113 MϪ1. H NMR experiments
at pH 2 show no shift in the protons’ signals, giving an effective
KA of 0 MϪ1.
J. Chem. Soc., Perkin Trans. 2, 2001, 258–263
261
Kalchenko, Olga I.
