Fmoc protected amino acids were purchased from Novabiochem (Laufelfingen, Switzerland) and from Iris Biotech (Marktredwitz, Germany), H-l-Thr(tBu)-ol-2-chlorotrityl resin and HATU were purchased from Iris Biotech. Second generation Grubbs catalyst was obtained from Aldrich (St. Louis, MO, USA). Fmoc-Agl-OH was purchased from Polypeptide Laboratories (Strasbourg, France). Peptide grade DMF was from Scharlau (Barcelona, Spain). TSP was purchased from MSD Isotopes (Montreal, Canada). ²H₂O was obtained from Aldrich. SDS-d₂₅ was obtained from Cambridge Isotope Laboratories, Inc. (Andover, MA). All the other solvents and reagents used for SPPS were of analytical quality and used without further purification. The solid phase syntheses and DOTA-conjugations were performed using a semi-automatic synthesizer from MultiSynTech (Witten, Germany). Analytical RP-HPLCs were performed on a Waters instrument equipped with a UV detector on a Phenomenex Jupiter C18 column (5 μm, 250 × 4.6 mm) using a flow rate of 1 mL/min, with the following solvent system: 0.1% TFA in H₂O (A), 0.1% TFA in MeCN (B). Semi-preparative RP-HPLC were performed on the same instrument using a flow rate of 4 ml/min with the same solvent system, on a Phenomenex Jupiter C18 column (10 μm, 250 × 10 mm). Mass spectra were registered on a Thermo-Finnigan ESI LCQ Advantage mass spectrometer (Waltham, MA, USA). LC-ESI-MS analyses were performed on a Phenomenex Jupiter C18 column (5 μm, 150 × 2.0 mm) using a flow rate of 500 μL/min on a ThermoFinnigan Surveyor HPLC system coupled to ESI-MS, using the solvent system: 0.1% TFA in H₂O (A), 0.1% TFA in MeCN (B).

Peptides were synthesized following the method reported in the preceding paper (Di Cianni et al., 2010). Briefly, the peptides were prepared using the general Fmoc-SPPS strategy on preswelled H-l-Thr(tBu)-ol-2-chloro-trityl resin. Couplings were performed by adding 2 equivalents of protected amino acid activated by HATU and 4 equivalents of NMM in DMF. Each coupling was monitored by the qualitative ninhydrin (Kaiser) test (Kaiser et al., 1970). The cyclization was performed on-resin by second generation Grubbs catalyst (0.5 mol eq. calculated on the basis of 0.5 mmol/g of peptide). After swelling, NH₂ terminal Fmoc-Agl was deprotected and coupled with Fmoc-d-Phe-OH, affording the on-resin peptides. At this point, all the on-resin peptides were divided in two equal amounts, one portion was deprotected with 20% piperidine in DMF and cleaved [3a,b with TFA/H₂O/phenol (96:2:2, 3 h); 4a,b, and 5a,b with TFA/H₂O/phenol (70:28:2, 2.30 h)]. The aqueous solutions of the peptides 3a–5a were pre-purified by solid phase extraction and after subjected to the purification by semipreparative RP-HPLC and subsequently characterized by ESI-MS. Analytical RP-HPLC and ESI-MS analysis of the crude compounds revealed two chromatographic peaks with the same MW for compounds 3a–5a, corresponding to the geometric isomers (Z/E ratio ≈ 90:10). Compounds were then purified by semipreparative RP-HPLC, and the most abundant chromatographic peaks were collected. For all the products, HPLC purity was 97%. The other portion with the deprotected on-resin peptides 3a–5a were coupled with the azamacrocycle DOTA, by adding 2 equivalents of protected DOTA-tris-tBu ester activated by 2 equivalents of HATU and 4 equivalents of NMM in DMF. Also in this case the coupling was monitored by the Kaiser test. The crude peptides 3b–5b were cleaved from the solid support and purified as already described.

Determination of Somatostatin receptors affinity was performed at CEREP (Le Bois l'Evêque, B. P. 3,001, 86,600 Celle l'Evescault, France). Cell membrane pellets were prepared from human sst₂-expressing endogeneous IMR32 cells, and sst₅-expressing human CHO cells. For each of the tested compounds, complete displacement experiments with the universal SRIF radio-ligand [Leu⁸, D-Trp²², ¹²⁵I-Tyr²⁵]-SRIF-28 (125I-[LTT]-SRIF-28) (2000 Ci/mmol; Anawa, Wangen, Switzerland) were performed. As control, unlabeled seglitide and SRIF-14 were run in parallel, using the same increasing concentrations, with sst₂ and sst₅ subtypes, respectively. IC₅₀ values were calculated by non-linear regression analysis of the competition curves generated with mean replicate values using Hill equation curve. The analysis was performed using software developed at CEREP (Hill software) and validated by comparison with data generated by the commercial software SigmaPlot 4.0 for Windows.

Analytical RP-HPLC was run on a Thermo Finnigan Surveyor HPLC equipped with a Phenomenex aqua C18 column 300 Å 5 μm (250 mm × 4.6 mm). The solvent systems used for gradients were A (0.1% TFA in H₂O) and B (0.1% TFA in CH₃CN). The flow rate was 1.0 mL/min, with a linear gradient from 40 to 90% of B in 20 min. The chromatographic peaks were monitored with a PDA detector at 254 nm.

A solution of each compound (1 mg/mL) was prepared in HPLC-grade water and 20 μL of solution were injected into the instrument. Each compound was repeated in triplicate and the average retention time was then calculated.

The samples for NMR spectroscopy were prepared by dissolving the appropriate amount of peptide in 0.55 mL of ¹H₂O (pH 5), 0.05 ml of ²H₂O to obtain a concentration 1–2 mM of peptides and 200 mM of SDS-d₂₅. TSP was used as internal chemical shift standard. The water signal was suppressed by gradient echo (Hwang and Shaka, 1995). NMR experiments were recorded on a Varian Inova-Unity 700 MHz at 308.1 K. 1D ¹H NMR spectra of the new compounds 3b–5b are reported in Figure S1 of the Supplementary Material. Complete ¹H NMR chemical shift assignments were effectively achieved for all the analyzed peptides (Tables S3–S5) according to the Wüthrich procedure (Wüthrich, 1986) via the usual systematic application of TOCSY (Braunschweiler and Ernst, 1983) and NOESY (Jeener et al., 1979) experiments recorded in the phase-sensitive mode using the method from States (States et al., 1982).

Typical data block sizes were 2,048 addresses in t₂ and 512 equidistant t₁ values. Before Fourier transformation, the time domain data matrices were multiplied by shifted sin² functions in both dimensions. A mixing time of 70 ms was used for the TOCSY experiments. NOESY experiments were run with mixing times of 100 and 200 ms. The qualitative and quantitative analyses of TOCSY and NOESY spectra were obtained with the support of the XEASY software package (Bartels et al., 1995).

The NOE-based distance restraints were obtained from NOESY spectra collected with the mixing time of 100 ms. The NOE cross peaks were integrated with the XEASY program and were converted into upper distance bounds using the CALIBA program incorporated into the program package DYANA (Güntert et al., 1997). Only NOE derived constraints (Tables S6–S8) were considered in the annealing procedures. In a first calculation run, all the upper distance bounds were used, generating an ensemble of 100 structures with the simulated annealing standard protocol of the program DYANA. For all peptides, a number of consistent (i.e., in all calculated structures) violated upper limit constraints (>0.1 Å) were observed (Tables S6–S8). These violations were discarded in a subsequent MD run. This step was repeated until no violation was observed (two runs were enough for all peptides). Thus, we obtained a first family of structures (family I). In a second MD cycle, the violated upper limit constraints of the first cycle were up-weighted (10-fold) for the contribution to the target energy function of DYANA. Hence, we obtained a new set of violated constraints which were discarded in the subsequent MD runs. After two MD runs, no violations were observed. In the final calculation run, we applied the same weight to the not discarded constraints and obtained a second family of structures (family II). Since, the two sets of violations had no common member we did not repeat further the described procedure.

Finally, 20 structures for each family of peptides 3b–5b were chosen, whose interprotonic distances best fitted NOE derived distances, and then refined through successive steps of restrained and unrestrained energy minimization calculations using the Discover algorithm (Accelrys, San Diego, CA) and the consistent valence force field (CVFF) (Maple et al., 1988).

The minimization lowered the total energy of the structures. The final structures were analyzed using the InsightII program (Accelrys, San Diego, CA). Graphical representations were carried out with the UCSF Chimera package (Pettersen et al., 2004). The root-mean-squared-deviation analysis between energy-minimized structures was carried out with the program MOLMOL (Koradi et al., 1996).

The synthesis of dicarba-analogs followed the procedure described in our previous article (Di Cianni et al., 2010). All the syntheses take advantages from the Solid Phase Peptide Synthesis protocol, allowing to rapidly afford the products in the range from a μM to mM scale and with high purity level.

The syntheses of the linear peptides were performed on a 2-chloro-trityl resin pre-loaded with the protected aminoalcool H-l-Thr(tBu)-ol and following the Scheme 1. The peptide chain elongation, according to the selected amino acid sequence, was achieved by the coupling of a pre-activated Fmoc-protected amino acid with HATU and NMM. All the couplings were checked by the Kaiser test (Kaiser et al., 1970). Once the syntheses of the linear peptides were completed, the corresponding dicarba-cyclopeptides were obtained by RCM reactions catalyzed with the second generation Grubb's catalyst in anhydrous DCM for 48 h at 50°C.

At this point, the last amino acid residue d-Phe, was coupled to the three on-resin cyclopeptides and the resins were divided in two halves. One of these was cleaved in order to obtain the free dicarba-cyclopeptide 3a–5a, while the other underwent to the DOTA coupling on the free amino-group of the d-Phe residue with the DOTA-tris-tBu ester, suitably activated with HATU and NMM. Also in this case, the compounds were cleaved from the solid support, affording the crude products 3b–5b. The cleavage of compounds 3a,b was carried out with a mixture of TFA/H₂O/phenol (96:2:2) for 3 h at r.t., while compounds 4a,b, and 5a,b, that needed some milder conditions since the OBz must be retained on the side-chain of the Tyr residue, were treated with TFA/H₂O/phenol (70:28:2) for 2.5 h. All compounds obtained were pre-purified by SPE. The concentrated solutions of the raw products were adsorbed on the SPE and eluted with an increased percentage of CH₃CN in H₂O (from 0 to 100%). The enriched fractions were then purified by semipreparative RP-HPLC and characterized by ESI-MS (see Tables S1, S2). For each peptide, the HPLC chromatogram showed two peaks with the same MW, corresponding to the geometric isomers (Z/E ratio ≈ 90:10). In particular, the Z structure of the C-C = C-C tether of the sample, eluted at higher R_t, was recovered and the cis-structure ascertained by ¹H NMR inspection. In particular, the geometry of the double bond of compounds 3b–5b was confirmed as cis (E) from the coupling constant value (³J_CH=CH = 11 Hz) between the two olefinic protons of the bridge (Figure S2) and the relative strong NOE between the same olefinic H_γs. Since, the signals of the two olefinic protons of all compounds 3b–5b overlapped each other in the NMR spectra acquired in SDS solutions, previous parameters were extracted from spectra acquired in D₂O solution. The HPLC purity of each compound was >97%, and the isolated structure showed unique Z configuration, confirmed by NMR analysis. No oligomeric by-products were observed.

The three DOTA conjugated compounds 3b, 4b, and 5b were tested for their ability to bind to the sst₂ and sst₅ receptors subtypes in complete displacement experiments using the universal somatostatin radioligand [¹²⁵I]-[Leu⁸,d-Trp²²,Tyr²⁵]-somatostatin-28. IC₅₀ values were calculated after quantification of the data using a computer-assisted image processing system (Table 1). Receptor subtypes sst₂ and sst₅ were chosen for preliminary binding assays because of their overexpression in some tumor types (Miller et al., 1995). In Table 1, the affinities of the parent unconjugated compounds 3a, 4a, and 5a are also reported for comparison (Di Cianni et al., 2010).

Bearing in mind that affinity values >100 nM cannot permit to consider the related compounds of pharmacological interest, from data reported in Table 1 it clearly appears that 3b maintains a fairly good affinity for sst₂ while it loses affinity for sst₅ subtype. The opposite case is reported for the Tyr(OBz)¹⁰ containing derivative 4b that shows a moderate reduction of the affinity at the sst₅ receptor but a total loss of sst₂ affinity. In the case of compound 5b, after the conjugation with DOTA there is roughly no effect on sst₂ activity while the sst₅ affinity is reduced (2.4-fold). Nevertheless, the binding ability with sst₅ still remains of significant interest.

RP-HPLC retention time (t_R) measurements can give an idea of the difference in hydrophobicity of our peptides after the conjugation with DOTA (Table 2; Tachi et al., 2002; Hossain et al., 2011).

Compound 3a is, as expected, the less hydrophobic analog having the –OH group of the Thr¹⁰ side chain in the place of the lipophilic Tyr(OBz)¹⁰ residue of 4a and 5a. On the other hand, it is not surprisingly that 5a is still more hydrophobic than 4a because of the outcome of the aromatic naphtyl side chain in position 7. The introduction of the DOTA moiety at the amino end of these analogs enhanced the affinity of the entire molecule for the stationary phase through a slight increment of the chromatographic retention times, at least in our experimental conditions but did not alter the hydrophobic/hydrophilic nature of the molecules to any significant extent. Noticeably, the contribution to the hydrophobicity carried by the DOTA moiety is almost the same for 3b–5b molecules. Because the lipophilicity = hydrophobicity—polarity (Giaginis and Tsantili-Kakoulidou, 2008), these finds seem to suggest that changes in hydrophobicity and then also in lipophilicity, introduced by the conjugation with the DOTA group, show a very similar trend along the three structures.

NMR analysis of the analogs 3b–5b was performed in SDS-d₂₅ micelles solution. The employment of SDS micelles to investigate the conformational properties is justified on the basis of their interaction with a membrane receptor. For peptides that bind membrane receptors, such as GPCR, the use of membrane mimetic solution is suggested, hypothesizing a membrane-assisted mechanism of interactions between the peptides and their receptors (Sargent and Schwyzer, 1986). Hence micelle solutions have been extensively used for conformational studies of peptide hormones and neuropeptides (Grieco et al., 2011; Carotenuto et al., 2013, 2015).

For compound 3b, NMR data resemble those of the corresponding parent 3a (Table S3) with the main differences located at N-terminal residues 2-3-7 (Note: numbering of the residues follows that of native SRIF, Figure 2). As for 3a (Di Cianni et al., 2010), NOESY spectra of 3b showed, simultaneously, both diagnostic connectivities consistent with folded structures and NOE contacts characteristic of extended regions (Table S6). Considering incompatible NOEs separately in different calculation cycles (Di Cianni et al., 2010), two families of conformation were obtained (Figure 3); family I with a type II' β-turn spanning residues d-Trp⁸-Lys⁹, followed by a short 3₁₀-helix along residues 10-14-15 (Figure 3A) and family II which differed from the first mainly in that C-terminal residues were in extended conformations (Figure 3B). Compared to the conformations obtained for peptide 3a, the principal difference in both families is a better definition of d-Phe² side chain which is predominately gauche- oriented in 3b, and a higher conformational freedom for 1-Nal⁷ side chain which populates both trans and gauche- conformers starting (in 3a) from a preferred trans conformation. d-Phe2 reorientation is probably due to attractive interactions between d-Phe and DOTA moieties, while 1-Nal⁷ movements should be a consequence of the first. Interactions between d-Phe² residue and N-terminal conjugate moieties were already observed in other octreotide dicarba-derivatives (Barragán et al., 2012).

These conformational modifications have to account for the loss of activity at receptors, mainly at the subtype sst₅, because probably the reorientation of the 1-Nal⁷ induces the loss of an important interaction. Differently from NMR spectra of 3a, those of Tyr(OBz) containing compounds (4b, and 5b) show significant differences regarding all residues compared to the parents 4a and 5a, respectively (Tables S4, S5). Some NOE interactions between Tyr(OBz) side chain and both Phe⁷ (1-Nal⁷) and d-Trp⁸ could be observed (Tables S7–S8). Structure calculation explained the experimental NMR data. In fact, apart the gauche- side chain orientations of d-Phe² and of Phe⁷ (1-Nal⁷) already described for 3b, Tyr(OBz) side chain of both extended (family I) and folded (family II) structure clusters of 4b and 5b had a gauche⁺ orientation (Figures 4, 5). Henceforth, the long Tyr(OBz) side chain is inserted between the aromatic systems of Phe⁷ (1-Nal⁷) and d-Trp⁸. This tight packing blocks the rotation of both Phe⁷ (1-Nal⁷) and d-Trp⁸. As a consequence of this loss of rotational freedom d-Trp⁸ indole moiety is closer, on average, to Lys⁹ residue compared to what happens in the unconjugated peptides; this is in accordance with the significant upfield shifts of all the proton signals of Lys⁹. This compact conformation is favorable to the sst₅ binding of 5b (and partially of 4b) since this compound recovers the affinity toward sst₅ compared to 3b.

At the same time, 4b and 5b show a moderate/slight reduction of the affinity toward sst₅ compared to 4a and 5a, respectively, again the non-perfect orientation of Phe⁷ (1-Nal⁷) can explain this worsening. Considering the sst₂ receptor, a slight (5b) or total loss (4b) of affinity is observed. Unfortunately, due to their poor overall affinities for sst₂ (both >100 nM), neither compound, 4b or 5b, can be considered as a good candidate for eliciting pharmacological activity on this receptor subtype. Consequently, the “cross-arm” orientation of Tyr(OBz) in the context of the β-hairpin structure (Figures 4B, 5B), that is considered the binding conformer at the sst₂ (Grace et al., 2006), results to be not suitable for sst₂ receptor.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.