Fig. 2.3.8. A tripeptide-based library of cationic guanidinio-carbonyl pyrrole receptors 7 designed for the binding of Val-Val-Ile-Ala, a tetrapeptide representing the C-terminus of Ab.

peptide (Ab) which is responsible for the formation of protein plaques within the brain of patients suffering from Alzheimer's disease [19]. This specific peptide sequence is thought to promote the formation of self-aggregated b-sheets of Ab stabilized through a combination of H-bonds and hydrophobic interactions [20]. An artificial receptor which effectively binds to the model tetrapeptide Ac-Val-Val-Ile-Ala-OH can therefore enable us to learn more about the molecular basis of the self-aggregation of the amyloid-peptide.

Our general design of a potential receptor 7 for this substrate is shown in Figure 2.3.8. The ion pair between the carboxylate and the guanidiniocarbonyl pyrrole serves as a starting point for complex formation. An additional tripeptide unit attached to the pyrrole provides further binding sites for the formation of a hydrogen-bonded antiparallel b-sheet with the backbone of the tetrapeptide substrate (Chapter 2.4 contains more about the use of such b-sheet mimics). In addition to these multiple electrostatic interactions, hydrophobic contacts between the amino acid side-chains both in the substrate and the receptor should, especially in aqueous solvents, further stabilize the complex and also guarantee the necessary substrate selectivity. To identify which amino side-chains in the receptor will be most efficient for this purpose a combinatorial approach was used.

A solid phase bound library of 512 different but structurally related receptors 7 was therefore synthesized using a standard Fmoc-procedure and a split-mix approach [21] (Boxes 11 and 25). In each of the three coupling steps eight different amino acids were used; these were selectively chosen to provide a range of structurally varying hydrophobic or steric interactions. In the second step each of the various 512 different tripeptides thus obtained was coupled with the guanidinio-carbonyl pyrrole binding motif. The advantage of such a solid-phase-bound combinatorial receptor library is, besides the fast and time-saving synthesis, that the whole library can be tested for a specific feature, in this case its binding properties towards the tetrapeptide substrate, in a single experiment [22]. For this purpose a fluorescence label in form of a dansyl group was attached via a water-soluble spacer

2.3.4 Binding of Small Oligopeptides | 149 labeled substrate ^ UV h9ht

2.3.4 Binding of Small Oligopeptides | 149 labeled substrate ^ UV h9ht

receptor library bound on solid support

Fig. 2.3.9. Fluorescence binding assay.

receptor library bound on solid support

Fig. 2.3.9. Fluorescence binding assay.

to the N-terminus of the tetrapeptide substrate. After incubation of the library with this labeled substrate, a simple UV-assay can be used to identify efficient receptors (Figure 2.3.9). Only those beads on which the attached receptor is capable of bind the peptide can show the characteristic fluorescence of the dansyl group. All the other receptors which do not bind the peptide under the specific experimental conditions remain dark.

Such binding assays showed that our one-armed cationic receptors 7 are indeed capable of efficient binding of the tetrapeptide in polar solutions, even in water [23]. Interestingly, the uncharged methyl ester of the substrate binds only weakly and rather unselectively to the receptor library, suggesting that side-chain interactions alone are not strong enough to form a stable complex. The negatively charged carboxylate substrate is, however, selectively bound only by some and not all of the receptors, although ion pairing with the guanidiniocarbonyl pyrrole unit is the same for all the different receptors within the library. Hence, the binding of the tetrapeptide by the receptors requires both electrostatic and hydrophobic interactions. Neither of these interactions alone is sufficiently strong to ensure complex formation in polar solution (Figure 2.3.10).

One must, of course ensure that the fluorescence activity observed in the assay is really due to a selective complexation of the tetrapeptide substrate by the receptor. This was done by appropriate control experiments:

1. Because the labeled tetrapeptide does not bind to the unmodified solid support Amino-TentaGel, the observed fluorescence activity is not because of unspecific interaction with the solid support itself.

2. The dansylated spacer alone does not bind to the receptor library, showing that the binding indeed occurs between the peptide part of the substrate and the receptor.

3. The percentage of receptors that bind the substrate is concentration-dependent - at high concentrations nearly all of the library members bind the substrate, which shows that the observed binding specificity is not a result of selective quenching of the dansyl fluorescence rather than selective binding.

Fig. 2.3.10. Binding assay in water.

These qualitative assays show that one-armed cationic guanidiniocarbonyl pyrrole receptors can indeed effectively bind tetrapeptides even in water. Molecular modeling studies suggest a complex structure as shown for one specific example, the receptor Val-Val-Val-CBS, in Figure 2.3.11. Receptor and substrate form a hydrogen bonded ¿-sheet which is further stabilized by additional hydrophobic interactions between the apolar groups in the side-chains. Recognition of the tetrapep-tide thus seems to be controlled by a fine balanced interplay between electrostatic and hydrophobic interactions.

Another advantage of such a combinatorial binding assay is that it can also be performed quantitatively, enabling direct determination of the binding constants of the different receptors [24]. The association constants for each receptor can be calculated from the fluorescence intensity of the substrate in solution before and after incubation and the loading of the resin. Even though such binding constants determined on a solid support are not the same, and less accurate than data obtained in solution (for example from NMR or UV titration experiments), a comparison of relative data within a series of related receptors can, at least, help rationalize aspects such as complex structure, stability and selectivity on a molecular basis. One can identify structural features that are associated with strong or weak bind-

Fig. 2.3.11. Proposed structure for the complex between receptor (top) and tetrapeptide (below).

ing. Which parts of our modular receptors are most important for binding or selectivity? What kind of binding sites, electrostatic or hydrophobic, in the various positions of the receptor are needed? In other words a supramolecular structure-binding relationship can be derived from binding data obtained on a solid support.

We have so far performed a detailed thermodynamic analysis for binding of the tetrapeptide in methanol [23]. As these data show the binding is exceptionally strong with approximate association constants of 104 for the best receptors. In this assay the binding is measured relative to the formate counter-ion of the receptors, which also binds to the guanidiniocarbonyl pyrrole motif, although to a lesser extent than other carboxylates. Therefore, the interaction between the tetra-peptide and the receptors is actually even much stronger than suggested by these numbers. Hence, these one-armed hosts are among the most efficient peptide receptors in polar solvents yet reported. The selectivity of the receptors towards the tetrapeptide substrate is, furthermore, surprisingly high. The association constants for the various receptors differ by a factor of more than 100 among the library! Even small changes in the structure of the receptor have obviously pronounced effects on the binding properties. This also proves that even within such a small combinatorial library of only limited structural diversity the binding selectivity can be rather high - a necessary prerequisite for also achieving selective binding of different tetrapeptides by this general receptor class.

A closer look at the binding data enables correlation of complex stability and receptor structure. For example, the quantitative binding constants suggest that hydrophobic interactions of the receptor with the first amino acid residue of the substrate (Val) are most important. Exchange of valine in the position opposite of this residue in the receptor for N-Boc-protected lysine reduces the binding affinity by a factor of 10. The side-chain of lysine is probably too small to provide enough hydrophobic shielding of the isopropyl group of the substrate in the complex. This is excellent confirmation of previous studies of the Ab self-aggregation which have shown that hydrophobic interactions with Val 39, which corresponds to the first amino acid of our tetrapeptide substrate, are especially important [25]. This again emphasizes the feasibility of using carefully chosen small bioorganic models for the analysis of more complex natural systems. First screening experiments with a larger library have already shown that, even in water, receptors of this general type efficiently bind the model tetrapeptide with association constants of the order of 104 m^1 (in tris-buffer of pH 6).

0 0

Post a comment