2.4.1. Compounds 1–3 with One Msa Residue
The SRIF analogs containing only one Msa insertion (in position 6, 7 or 11, analogs
1–
3) were analyzed by NMR. Unlike somatostatin, which populates several conformations in aqueous solution, the two dimensional TOCSY and NOESY homonuclear experiments [
45] showed a major set of NOE peaks. The well-defined 2D spectra of compounds
1–
3 enabled us to characterize their main conformation in solution using the software Crystallography & NMR System (CNS) [
46]. To generate the list of experimental restraints required for the calculation, the volume of all assigned peaks was integrated, and then converted into distances. Three calculations (120 structures each) were run until the best match between the NMR assignments and final structures was obtained. Based on these results we concluded that under the experimental conditions used, compounds
1–
3 were sufficiently structured to obtain defined families of structures that are in good agreement with the experimental data (
Figure 4,
Figure 5 and
Figure 6).
Figure 4.
Structures of the lowest energy conformers of [L-Msa6]-SRIF (1) (A), [L-Msa7]-SRIF (2) (B) and [L-Msa11]-SRIF (3) (C). Hydrogen atoms have been omitted for clarity.
Figure 4.
Structures of the lowest energy conformers of [L-Msa6]-SRIF (1) (A), [L-Msa7]-SRIF (2) (B) and [L-Msa11]-SRIF (3) (C). Hydrogen atoms have been omitted for clarity.
The 3D structure of [L-Msa6]-SRIF (
1) showed a singular aromatic ring cluster among Msa6, Phe7 and Phe11 as well as a hairpin (
Figure 4A). This arrangement of aromatic rings was defined by a number of NOE’s among the aromatic protons of these three residues. However, the absence of an additional stabilizing effect in the hairpin area increased the conformational mobility of this peptide. Thus, our NMR data suggested that other minor conformations were present in solution. As shown in
Table 2, compound
1 binds to all receptors although with weaker affinity than somatostatin. Furthermore, the presence of only one unnatural residue in its sequence did not increase its stability in serum; its half-life is slightly lower than that of the natural hormone.
The NMR data of [L-Msa7]-SRIF (
2) clearly showed that this compound was conformationally more rigid than
1 in solution. The 3D structure of the lower energy conformers showed a highly structured region from residues
6 to
11, with a clear aromatic interaction between Phe6 and Phe11 (
Figure 4B) which can be defined as an
edge-to-face interaction
. The Msa residue in the seventh position does not participate in the aromatic interaction, lying flat at the opposite face of the molecule. However, it probably plays an essential role in conformer stabilization, helping the aromatic rings of Phe6 and Phe11 to attain the optimal geometry. NOE crosspeaks of Lys9-Trp8 interaction were weak and difficult to identify. Peptide
2 was highly selective towards receptor SSTR2, with a
Ki of 0.019 nM. Interestingly, and perhaps due to its conformational stability, the stability of peptide
2 in serum (5.2 h) is almost double to that of SRIF, and has the highest stability of the three analogs
1–
3 with only one residue modification.
Figure 5.
NOESY (600 MHz, D2O, 200 ms) of the aromatic ring-long range interaction region of the natural hormone, [D-Trp8]-SRIF, [L-Msa7]-SRIF (2) and [L-Msa7,D-Trp8]-SRIF (5). NMR data were acquired at 285 K, using trifluoroacetate as a counter-ion at pH 1.5.
Figure 5.
NOESY (600 MHz, D2O, 200 ms) of the aromatic ring-long range interaction region of the natural hormone, [D-Trp8]-SRIF, [L-Msa7]-SRIF (2) and [L-Msa7,D-Trp8]-SRIF (5). NMR data were acquired at 285 K, using trifluoroacetate as a counter-ion at pH 1.5.
The set of low energy conformers calculated for peptide [L-Msa11]-SRIF (
3) showed a remarkable level of convergence in the majority of geometrical parameters as a result of the high number of experimental restraints observed and the intrinsically high conformational rigidity in the molecule. In this analog, the Msa amino acid at position 11 participates in a π-π aromatic interaction, with the phenyl ring of Phe6 oriented in an
offset-tilted arrangement on one side of the molecular plane (
Figure 4C). Moreover, the Phe7 ring lies on the other side of the molecular plane as it occurs in the structure of molecule
2, but in
3 the orientation is almost perpendicular to the plane.
Figure 6.
Structures of the lowest energy conformers of [L-Msa6,D-Trp8]-SRIF (4) (A), [L-Msa7,D-Trp8]-SRIF (5) (B) and [L-Msa11,D-Trp8]-SRIF (6). (C). Hydrogen atoms have been omitted for clarity.
Figure 6.
Structures of the lowest energy conformers of [L-Msa6,D-Trp8]-SRIF (4) (A), [L-Msa7,D-Trp8]-SRIF (5) (B) and [L-Msa11,D-Trp8]-SRIF (6). (C). Hydrogen atoms have been omitted for clarity.
As observed in peptide 2, the interaction between the side chains of Lys9 and Trp8 is instrumental in defining the formation of the hairpin. However, these interactions are not strong enough to fix the L-Trp8 in a fixed rotamer, with some conformations of peptide 3 bringing the L-Trp8 indole side-chain in close proximity to the benzyl side-chain of Phe7. Remarkably, the binding profile of this compound reflects that it is highly selective toward SSTR2, despite having a dissociation constant larger than that of compound 2. However, its stability in serum is poor, displaying a half-life even shorter than that of natural SRIF.
2.4.2. Compounds 4–6 with One Msa Residue and D-Trp8
The discovery of unique conformations in peptides
1–
3, which allowed us to determine their 3D structures by NMR in aqueous solution, prompted us to concentrate our efforts on increasing the stability of these analogs while maintaining their binding properties. Pioneering work by Rivier and co-workers in 1975 [
39] showed that [D-Trp8]-SRIF exhibits an excellent binding profile toward all receptors but higher stability than its [L-Trp8]-SRIF natural counterpart. Indeed, several studies [
39,
40] have demonstrated that the Lys9 side chain is more effectively shielded by [D-Trp8] because the D-configuration of tryptophan favors an orientation where the indole ring is in close proximity to the aliphatic side chain of Lys9. These contacts not only maintain the Lys9 side chain in a defined orientation, but also induce the formation of the hairpin centered at Trp8-Lys9.
Our aim was to investigate whether the introduction of the [D-Trp8] residue would increase the stability of the peptides containing Msa side-chains while maintaining the conformational structure and the binding profile. Therefore, we prepared three new analogs
4–
6 carrying Msa and [D-Trp8] residues by SPPS and determined whether these substitutions affected either the structure or stability of the peptides. In addition, we also prepared the [D-Trp8]-SRIF analog and compared its NMR properties with that of the natural [L-Trp8]-SRIF. A section of the 2D NOESY spectra of the natural hormone, [D-Trp8]-SRIF, [L-Msa7]-SRIF (
2) and [L-Msa7,D-Trp8]-SRIF (
5), is shown in
Figure 5 to illustrate the improvement in the NMR data. As it can be observed, [D-Trp8]-SRIF maintains the intrinsic flexibility of the natural hormone, whereas
2 and
5 show an increasing amount of NOE signals.
The conformational flexibility of both SRIF and [D-Trp8]-SRIF accounts perfectly for its functional versatility against all receptors (SSTR1-5) (
Table 2). In both cases, the coexistence of several different conformations prevented us from carrying out definitive structural studies. In contrast, the 2D spectra of compounds
4–
6 were extremely well-defined, enabling us to characterize their main conformation in solution from the NMR restraints and using the CNS software. As before, the volume of all assigned peaks was integrated, transformed into distances and used to generate the list of experimental restraints for calculation. Three sets of calculations (120 structures each) were run until the best match between assignments and final structures was obtained. As expected, the generated structures (from the experimental NMR data) of these new peptides showed a clear increase in convergence not only with respect to SRIF and [D-Trp8]-SRIF, but also with respect to monosubstituted analogs
1–
3.
Peptide [L-Msa6,D-Trp8]-SRIF (
4) gave a well-defined conformation in solution (
Figure 6A) in which the two aromatic rings of Msa6 and Phe11 are markedly proximal due to the enhanced aromatic interaction, while Phe7 is also participating in the aromatic cluster. This cluster of three aromatic rings is similar to the one found in [L-Msa6]-SRIF (
1). It is apparent that the region containing residues
5–
11 is much more structured than the rest of the molecule. The pharmacophore region of the most stable conformations matched perfectly (residues
6 to
11). On the other hand, the rest of the molecule is less rigid. The binding profile of peptide
4 is also similar to peptide
1, although its affinity towards receptors SSTR3 and SSTR5 is comparable to that of the natural hormone, with an affinity 20–30 times more potent than octreotide (
Table 2). The lack of affinity of both peptides
1 and
4 towards receptor SSTR2 (which both have a Msa residue in the sixth position) does not support the idea that residue
6 interacts via a π-donation with an amino acid side chain in SSTR2, as suggested in Hirschmann’s hypothesis [
47].
Peptide [L-Msa7,D-Trp8]-SRIF (
5) also showed a strong interaction between Phe6-Phe11 and a well-defined hairpin at the pharmacophore region (
Figure 6B). This hairpin was also observed in the [L-Msa7]-SRIF peptide (
2). Again, the calculated structures display a good degree of convergence, being the pharmacophore region (residues
5–
11) more structured than the remaining part of the molecule. In contrast to peptide
2, a large set of interactions between Trp8 and Lys9 can be clearly observed in compound
5. In this case, the Lys-Trp interaction is reflected in the upfield shift γ protons of Lys9 which are shielded by the aromatic indole ring. Furthermore, the aromatic-aromatic interactions between residues Phe6 and Phe11 in peptides
2 and
5 are very similar as depicted in
Figure 7, (the backbone is shown as a cartoon representation and the Phe6 and Phe11 residues as sticks). The molecules were rotated with respect to previous figures to provide a close view of these interactions. The aromatic pair association can be classified as an
edge-to-face type [
48,
49]. The shortest inter-residue carbon-carbon distances (SICD) are 3.3 and 2.7 Å respectively. The angle formed between the ring centroid to centroid segment and the z-axis of an axial system centered on the centroid of the reference ring (θ) is very small (16° and 13° for
2 and
5 respectively). The ring planes are almost perpendicular with interplanar angles (P) of 81° and 72° respectively. The mesityl ring in peptide
5, as in peptide
2, is not involved in any aromatic interactions. The aromatic ring is lying flat at the other side of the molecule, probably facilitating the interaction between both Phe6 and Phe11 by steric repulsion. The lowest energy conformation is fairly similar to
2, but the synergistic stabilizing effect due to the presence of the D-Trp8 leads to the least flexible 14-residue SRIF analogue described to date. Its outstanding affinity for SSTR2 probably correlates with having a well-defined structure close to the conformation that fits best in the structure of the SSTR2 receptor, which is so far uncharacterized. Its inhibition constant was 22-fold lower than that of octreotide and similar to SRIF. Peptide
5, unlike octreotide, also exhibited an impressive affinity toward SSTR1, at the same level than SRIF.
Figure 7.
Schematic representation of the aromatic interactions in peptides 2, 3, 5 and 6.
Figure 7.
Schematic representation of the aromatic interactions in peptides 2, 3, 5 and 6.
The analogue [L-Msa11,D-Trp8]-SRIF (
6) also had remarkable conformational stability. The most stable conformers showed a strong π-π aromatic interaction between the Phe6 and Msa11 residues as displayed in
Figure 6C. It was also apparent that peptides
3 and
6 showed a remarkable similarity to one another. The geometric values of the aromatic interaction in peptides
3 and
6 are shown in
Figure 7 showing only the backbone and the two aromatic rings of the most stable conformer. In this case, the position of the two aromatic rings fits as an
offset-tilted interaction, with the θ angles wider than in previous analogs (30° and 34° for
3 and
6 respectively). In this case, the interplanar angles (P) are 53° and 61° (
Figure 7). Peptide
6 showed high affinity towards the SSTR5 receptor, although it displayed a lower level of selectivity, since it also binds to SSTR1, SSTR2 and SSTR3 receptors with comparable affinities. Its stability in serum is very high (41 h), which is higher than any of the analogs prepared carrying one Msa and the D-Trp8 residues.
Overall, the structures of peptides with L-Trp8 (compounds 1–3) are remarkably similar to those with D-Trp8 (compounds 4–6). Nearly all of the observed NOE’s for compounds 1–3 were also present in the spectra of compounds 4–6. This suggests that in peptides 4–6, the D-Trp8 residue shifts the equilibrium toward conformations that are already populated in a considerable manner in peptides 1–3. The NMR data of compounds 4–6 showed a higher increase in number and intensity of the contacts between Trp8-Lys9 than peptides 1–3. Furthermore, the contrast between the γ protons of Lys9 in peptides 1–3 (which showed a similar shift to those in [L-Trp8] SRIF), and analogs 4–6 (which were significantly upfield shifted) could be attributed to the efficient interaction of the Lys9 and D-Trp8 side chains.