Introduction

The neurohypophyseal peptide hormone oxytocin (OT) is a cyclic nonapeptide with a disulfide linkage between two cysteine residues at positions 1 and 6. The N-terminal amino group is free, the C-terminus tail with the sequence Pro-Leu-Gly is amidated. Its highly potent deamino analogue (dOT) differs only in the absence of the N-terminal amino group. Among other methods, NMR was frequently used for the study of conformation of OT, dOT, vasopressin and several of their analogues (Hruby and Lebl 1987; Lebl et al. 1990; Marik et al. 2001; Budesinsky et al. 2005; Budesínsky et al. 2005; Sikorska et al. 2006; Rodziewicz-Motowidlo et al. 2008; Li et al. 2008; Zhou and Troy 2005; Zhou and Troy 2003).

First results of 1H- and 13C-NMR experiments with OT and arginine vasopressin (AVP) led to the consideration that the Cys6-Pro7 bond of OT and AVP exists exclusively in the trans conformation (Hruby and Lebl 1987). Later, 10% cis isomer population in water was detected when studying the amide isomer equilibrium of the Cys6-Pro7 bond in OT by one- and two-dimensional NMR spectroscopy (Larive et al. 1992). The relative quantity of the cis isomer was decreased when the solvent was changed from water to methanol (Larive and Rabenstein 1993) which agreed with the expectation that the rate of cis/trans interconversion may be faster in non-aqueous solvents (Harrison and Stein 1992). Cis peptide bonds were latter also observed in sarcosyl7-, N-methylalanyl7-, thiazolidine-4-carboxylic acid7 or 3,4-didehydroproline7-OT analogues (Grzonka et al. 1985; Grzonka et al. 1983; Rodziewicz-Motowidlo et al. 2002; Rosamond and Ferger 1976; Moore et al. 1977). Introduction of glycine into position 7 leads to a decrease of uterotonic activity though to an increase of selectivity (Hruby and Lebl 1987; Lowbridge et al. 1977). All the results indicate that proline in position 7 is an important conformational constraint necessary for biological activity.

Structure–activity studies concerning the structural features leading to agonism or antagonism of oxytocin effect on uterus prompted investigations on the hypothesis that antagonism is due to the cis Cys6-Pro7 isomer, whereas the trans isomer results in agonistic activity. For example, the cis Cys-Pro peptide bond was observed in the potent bicyclic antagonists of OT, [Mpa1,cyclo(Glu4,Lys8)]OT and [dPen1,cyclo(Glu4,Lys8)]OT, by NMR spectroscopy and computational analysis (Shenderovich et al. 1997). The X-ray structure of the potent OT agonist, [Mpa1]OT (dOT), showed the Cys-Pro imide bond in the trans conformation (Wood et al. 1986). These observations led Lubell and others to synthesize analogues of OT, [Mpa1]OT, and [dPen1]OT having replaced Pro7 by (2S,5R)-5-tert-butylproline, a proline analogue inducing up to 90% of the cis amide conformation in N-(acetyl)dipeptide-N′-methylamides as shown previously (Belec et al. 2000; Bélec et al. 2001).

Wittelsberger et al. (2005) on the other hand reported the synthesis and the conformational and biological properties of thiazolidine and oxazolidine derivatives in position 7 that served as proline analogues with increased proline-specific properties. The dimethyl-substituted derivatives, 2,2-dimethyl-1,3-thiazolidine-4-carboxylic acid and 2,2-dimethyl-1,3-oxazolidine-4-carboxylic acid, induce up to 95% of the cis conformation as determined by one- and two-dimensional NMR spectroscopy in DMSO and in water. The impact of the dimethyl moiety at 2-C was assessed by comparison with the corresponding dihydro compound, [Cys(φH,Hpro)]7OT. Comparison of the oxytocic activities of the cis-constrained compounds and OT showed that the agonistic potency increased proportionally to the trans content of the 6-7 peptide bond; however, no antagonistic activity was observed for the cis-constrained analogue, weakening the possibility that the cis conformation is necessary for antagonism. The results lead however to the hypothesis that the cis/trans conformational change is playing a role in OT receptor binding and activation. There was one interesting finding and thus that in the absence and presence of magnesium ions, a considerable change was observed in the pattern of the HR protons of residues 3 and 1, indicating conformational changes that influence activity (Wittelsberger et al. 2005). Also in the case of cyclic analogues, the structural analysis revealed β-turns around residues Tyr2 and Ile3, which differed from the previously discussed β-turn geometry around residues 3 and 4 that was ascribed to the conformation of OT agonists (Hruby and Lebl 1987; Oldziej et al. 1995).

Recently we have published the synthesis and biological activities of four new OT and dOT analogues having incorporated l- or d-1,2,3,4,-tetrahydroisoquinoline-3-carboxylic acid (l- or d-Tic) into position 7 (Fragiadaki et al. 2007), i.e., [l-Tic7]OT (1), [d-Tic7]OT (2), [Mpa1,l-Tic7]OT (3) and [Mpa1,d-Tic7]OT (4). As their biological activities have been very interesting (see Table 1) we decided to study their structure more deeply. The NMR study (determination and analysis of NOESY, TOCSY, and 1H–13C HSQC spectra) is presented in this paper.

Table 1 Biological activities of studied oxytocin analogues

Materials and methods

NMR spectroscopy

Data were acquired at 298 K on a Bruker Avance 600 MHz spectrometer. 1H 1D NMR spectra were recorded using spectral width of 12–17 ppm with or without presaturation of the H2O signal. 1H–1H 2D TOCSY (Braunschweiler and Ernst 1983; Bax and Davis 1985) were recorded using the MLEV-17 spin lock sequence using τ m = 80 ms, and 1H–13C HSQC (Bax and Grzesiek 1993; Bothner-By et al. 1984) with 200.791 ppm spectral width in F1. 1H–1H TPPI NOESY (Marion and Wüthrich 1983; Jeener et al. 1979) spectra were acquired using mixing time τ m = 300 ms applying water suppression during the relaxation delay and mixing time. All 2D spectra were acquired with 10.014 ppm spectral width, consisting of 2 K data points in the F2 dimension, 16–32 transients and 512–1,024 complex increments in the F1 dimension. Raw data were multiplied in both dimensions by a pure cosine-squared bell window function and Fourier-transformed to obtain 2,048 × 2,048 real data points. A polynomial base-line correction was applied in both directions. For data processing and spectral analysis, the standard Bruker software (XWinNMR 3.5) and XEASY (Eccles et al. 1991) program (ETH, Zurich) were used.

NOE constraints

313, 286, 225 and 440 NOESY cross-peaks were assigned in both dimensions for [d-Tic7]OT, [Mpa1,d-Tic7]OT, [Mpa1,l-Tic7]OT and [l-Tic7]OT, respectively, in DMSO. The number of unique cross-peaks were 151, 136, 124 and 216 (15, 14, 13 and 22 constraints per residue for [d-Tic7]OT, [Mpa1,d-Tic7]OT, [Mpa1,l-Tic7]OT and [l-Tic7]OT, respectively). Their intensities were converted into upper limit distances through CALIBA (Güntert et al. 1991). Sequential constraints, number and range of NOEs and chemical shift differences ({(Δδ)2 + (Δδ)2}1/2) are illustrated in Online Resources 1, 2 and 3.

Structure calculations and refinement

The NOE-derived structural information extracted from the analysis of NOESY spectra acquired in DMSO –d 6 solutions under identical experimental conditions for all three peptides was introduced to DYANA (Güntert et al. 1997; Wüthrich et al. 1983) software for structure calculations. The peptide models in the figures have been generated with MOLMOL (Pearlman et al. 1997). Structural calculations have been performed on IBM RISC6000 and xw4100/xw4200 HP Linux workstations. 3J(NH–Hα) coupling constants were determined by inverse Fourier transformation of in-phase multiplets from NOESY spectra using the INFIT routine of the XEASY software (Eccles et al. 1991).

Results

Proton assignment

TOCSY maps were first analyzed to assign the individual spin patterns of amino acids through scalar connectivities. Sequential, medium and long range connectivities were identified from NOESY maps acquired with τ m = 300 ms. Chemical shifts for the four peptides are reported in Tables 2, 3, 4, and 5. Characteristics TOCSY fingerprints regions are given in Fig. 1.

Table 2 Chemical shifts (in ppm) and 3J(NH–Hα) coupling constants (in Hz) of the residues in the [d-Tic7]OT peptide at 298 K (DMSO-d 6)
Table 3 Chemical shifts (in ppm) and 3J(NH–Hα) coupling constants (in Hz) of the residues in the [Mpa1,d-Tic7]OT peptide at 298 K (DMSO-d 6)
Table 4 Chemical shifts (in ppm) and 3J(NH–Hα) coupling constants (in Hz) of the residues in the [Mpa1,l-Tic7]OT peptide at 298 K (DMSO-d 6)
Table 5 Chemical shifts (in ppm) and 3J(NH–Hα) coupling constants (in Hz) of the residues in the [l-Tic7]OT peptide at 298 K (DMSO-d 6)
Fig. 1
figure 1

Characteristic TOCSY fingerprints regions of [l-Tic7]OT (left) and [Mpa1,l-Tic7]OT (right) extracted from 2D 1H 600-MHz NMR recorded in DMSO–d 6 at 298 K

[l-Tic7]OT (1)

The NOE pattern involving HN–HN, Hα–ΗΝ and Hβ–HN (i, i + 1) connectivities for this peptide is rather similar to the [Mpa1,l-Tic7]OT peptide (Online Resource 2). The Hα–HN and Hβ–HN (i, i + 2) NOEs network involves residues Tyr2 and Gln4. Moreover, Hα–HN and Hβ–HN of (i, i + 3) type connectivities are also identified between Tyr2 and Asn5. This finding is also supported by the formation of a (i, i + 3) hydrogen bond between residues Tyr2 and Asn5 observed in all 20 calculated models. The observed NOEs between Cys1 and Cys6 suggest that these residues are in close proximity in a similar way as in the other studied peptides. Among the noteworthy NOEs are those of the medium-range connectivities between Tyr2-l-Tic7, which are not detected in NOESY spectra of the other three peptides, [d-Tic7]OT, the [Mpa1,d-Tic7]OT and [Mpa1,l-Tic7]OT.

[d-Tic7]OT (2)

Numerous HN–HN, Hα–ΗΝ and Hβ–HN sequential connectivities are detected in the region Cys1-l-Tic7 and Leu8-NH 102 , while Hα–ΗΝ and Hβ–HN of (i, i + 2) type connectivities between Cys6 and Leu8 have also been observed. Furthermore, (i, i + 2) type connectivities have been identified between the Hα and Hβ proton of Cys1 with Hα and Ηδ proton of Ile3. The medium-range connectivities between the Hβ proton of Cys1 and the amide proton of Asn5 are among the characteristic NOEs that indicate the spatial proximity of these amino acids. Schematic representation of the sequential and medium range connectivities is given in Online Resource 1.

[Mpa1,l-Tic7]ΟΤ (3)

Numerous HN–HN sequential connectivities are detected in the region Mpa1-Ile3 and Gln4-l-Tic7 and Leu8-NH 102 , while Hα–HN and Hβ–HN sequential connectivities are also identified between all residues, except l-Tic7. Furthermore, an HN–HN of (i, i + 2) connectivity between Ile3-Asn5 has also been observed. A Hα–ΗΝ of (i, i + 2) type connectivity between Cys6-Leu8 has also been identified. The medium-range connectivities between the Mpa1 Hβ proton and the Hα proton of Cys6 are among the characteristic NOEs further supporting the spatial proximity of these residues. Schematic representation of the sequential and medium range connectivities is given in Online Resource 2.

[Mpa1,d-Tic7]OT (4)

The NOE pattern involving HN–HN, Hα–ΗΝ and Hβ–HN (i, i + 1) connectivities for this peptide is rather similar to the [d-Tic7]OT peptide. HN–HN and Hα–ΗΝ (i, i + 2) type connectivities are observed for the tripeptide comprised of Tyr2-Ile3-Gln4, while Hβ–HN (i, i + 2) type NOEs are also identified among residues Tyr2-Ile3-Gln4, Ile3-Gln4-Asn5 and Cys6-Pro7-Leu8. The observed NOE network in concern with HN-HN connectivity between Ile3-Cys6 suggests a turn-like structure comprised of these residues. Additionally, two medium-range NOEs of Hβ protons of Mpa1 with Hα and Hβ protons of Cys6 indicate that these residues of the peptide are in close proximity in a similar way as in [d-Tic7]OT peptide. Furthermore, the formation of a hydrogen bond between residues Ile3 and Asn5 is consistent in all 20 calculated models. Schematic representation of the sequential and medium range connectivities is given in Online Resource 1.

Chemical shift difference analysis

Two diagrams illustrating the 1H chemical shift differences between the peptide pairs were plotted. The variations in chemical shifts are potential indicators for the conformational changes imposed by the replacement of the amino acids at position 7 (Online Resource 3). These plots refer to the peptide pairs [d-Tic7]OT-[Mpa1,d-Tic7]OT and [l-Tic7]OT-[Mpa1,l-Tic7]OT. The largest chemical shift variation in the first peptide pair [d-Tic7]OT-[Mpa1,d-Tic7]OT (black) was observed for Cys6 (>0.40 ppm), while smaller variations were calculated for Asn5 and Leu8, suggesting conformational rearrangements between the two peptides. The introduction of Mpa1 in position 1, which is a non-protein residue and is less bulky than Cys1, has been shown to affect the chemical environment and consequently the nature of the disulfide bond linking positions 1 and 6. Furthermore, Asn5, which lies in the middle of the sequence, seems to be affected by modification in position 1.

In the other peptide pair [l-Tic7]OT-[Mpa1,l-Tic7]OT the largest chemical shift variations are calculated for Tyr2 and Cys6 (>0.60 ppm) affected by modification in position 1. A smaller variation is observed for Gly9 (>0.50 ppm), suggesting a different conformation for the tripeptide segment Pro7-Leu8-GlyNH 92 .

Another two diagrams illustrating the 1H chemical shift differences were plotted for peptide pairs [Mpa1,d-Tic7]OT-[Mpa1,l-Tic7]OT and [d-Tic7]OT-[l-Tic7]OT (Online Resource 3). As far as the [Mpa1,d-Tic7]OT-[Mpa1,l-Tic7]OT peptide pair is concerned, the largest chemical shift variation was observed for Leu8 (>0.50 ppm), while smaller variations were calculated for Ile3 and Gly9 (>0.30 ppm) suggesting conformational rearrangements between the two peptides (Online Resource 3). The chemical shift variation of Ile3 is justified by the NOE network involving amino acids Ile3and Cys6, present only in the [Mpa1,d-Tic7]OT peptide. As a conclusion, the modification in position 7 suggests small conformational rearrangements of the fragment Leu8-GlyNH 92 between the two peptides.

As far as the [d-Tic7]OT-[l-Tic7]OT peptide pair is concerned, the largest chemical shift variation was observed for Asn5 (>0.50 ppm), while smaller variations were calculated for Tyr2 and Leu8 (>0.30 ppm) suggesting conformational rearrangements between the two peptides. The chemical shift variations of Asn5 and Tyr2 are justified by the NOE network involving amino acids Cys1-Asn5 in the [d-Tic7]OT peptide, and by NOEs between Tyr2-l-Tic7, present only in the [l-Tic7]OT peptide. The chemical shift variation calculated for Leu8 (>0.30 ppm) is justified by the modification of the amino acid in position 7.

Structure calculations and conformational analysis

The average target function for the DYANA family of 20 calculated models was found to be 0.32 ± 1.34 × 10−5 Å2 for [d-Tic7]OT, 0.11 ± 0.003 Å2 for [Mpa1,d-Tic7]OT, 0.66 ± 1.06 × 10−7 Å2 for [Mpa1,l-Tic7]OT and 0.45 ± 0.23 Å2 for [l-Tic7]OT peptide models. No consistent violations existed at the final DYANA run and no constrained violation was found larger than 0.30 Å.

The DYANA family models for [d-Tic7]OT peptide exhibit pairwise rmsd values for all residues 0.10 ± 0.05 Å (BB), 1.26 ± 0.48 Å (HA) for the 20 structures. The rmsd values for [Mpa1,d-Tic7]OT ensemble are 0.36 ± 0.43 Å (BB), 0.99 ± 0.61 Å (HA) for the 20 models, while for [Mpa1,l-Tic7]OT are found 1.00 ± 0.44 Å (BB), 2.50 ± 0.73 Å (HA) for the 20 models. Furthermore, the DYANA family models for [l-Tic7]OT peptide exhibit rmsd values for all residues 0.69 ± 0.48 Å (BB), 1.62 ± 0.52 Å (HA) for the 20 structures.

3D solution structures

The NMR data for the four analogues indicate that the residue in position 1 (Cys1/Mpa1) remains in close spatial proximity with Cys6. Medium range NOEs between residues Cys1/Mpa1 and Cys6 are fully consistent with the fact that Cys1/Mpa1 and Cys6 are linked through a disulfide bond. In general, [d-Tic7]OT, [Mpa1,d-Tic7]OT and [Mpa1,l-Tic7]OT peptides exhibit a lower number of NOE cross-peaks (151, 136 and 124 cross-peaks, respectively) relative to the [l-Tic7]OT peptide (216 NOE cross-peaks).

Highly populated hydrogen bonds are identified through the analysis of the NMR solution models ensemble. Specifically, both analogues bearing LTic in position 7 of the peptide sequence form a hydrogen bond involving amino acids in positions 2 and 5. In the [LTic7]OT analogue, a strong backbone hydrogen bond is identified between Tyr2-NH and the Asn5-CO (1.42–1.55 Å). Additionally, a diagnostic tool for the determination of a turn structure is the Cα distance of the residues (Lewis et al. 1973). To this context, the Cα atoms of Tyr2 and Asn5 are found in close proximity (4.37–5.00 Å). In the [Mpa1,LTic7]OT analogue, a high populated H-bond is formed between the amide proton of Asn5 and the carbonyl oxygen of Tyr2 (1.57–2.09 Å), while the distance between the Cα atoms of these amino residues is measured 4.40–4.49 Å. The above data suggest regular β-turns for both [LTic7]OT and [Mpa1,LTic7]OT peptides comprised of Tyr2-Asn5 residues. Applying the same criteria as above, for the [Mpa1,DTic7]OT analogue, the turn formed bears the conformational features of a γ-turn. A very highly populated hydrogen bong, observed in 19 out of the 20 calculated structures of the ensemble, is formed among residues Asn5-NH and Ile3-CO (1.87–2.32 Å) in the [Mpa1,DTic7]OT analogue. The distance between the Cα atoms of Tyr2 and Asn5 is found to be 2.4–3.4 Å. Moreover, a second weaker H bond is identified between Hδ protons of Asn5 and the carbonyl oxygen of Gln4, present in 10 out of 20 calculated structures. Concerning the [DTic7]OT peptide, no highly populated hydrogen bonding is identified and no safe conclusion can be extracted from the analysis of the NMR data and calculated models for the assignment of this turn conformation to any of β- or γ-turns.

In the [l-Tic7]OT analogue, residues Cys1 and Cys6 are also found to be in close proximity as manifested by medium-range NOEs, such as the Hβ proton of Cys1 with the amide proton of Cys6, as well as the amide proton of Tyr2 with the amide proton of Asn5 and Cys6. The observed NOEs are indicative for a local conformation that favours the formation of a hydrogen bond between Tyr2 and Asn5 observed in the family of 20 best DYANA structures, while a turn is formed by residues Ile3-Gln4-Asn5. Furthermore, the Hα–ΗΝ of (i, i + 2) and Hβ–HN of (i, i + 2) type NOE connectivities for residues Tyr2 and Gln4 indicate the formation of a type II-β turn for this segment. Moreover, (i, i + 3) type connectivities between Tyr2 and Asn5 further support the backbone turn for this region (Fig. 2).

Fig. 2
figure 2

Mean structures calculated for the [d-Tic7]OT, [Mpa1,l-Tic7]OT, [l-Tic7]OT and [Mpa1,d-Tic7]OT analogues. Figures were generated with the MOLMOL program

According to the NMR analysis and structure calculations presented here, [l-Tic7]OT shares some similar conformational features with [Mpa1,l-Tic7]OT. Residues Mpa1 and Cys6 are also found in spatial proximity. The vicinity of these residues is supported by a significantly smaller amount of NOEs than in the case of the [l-Tic7]OT peptide, which leads to a less compact structure (Fig. 3).

Fig. 3
figure 3

Superimposition of the backbone of the two peptide pairs; [l-Tic7]OT-[Mpa1,l-Tic7]OT (left) as well as [d-Tic7]OT-[Mpa1,d-Tic7]OT (right). Figures are generated with MOLMOL

A backbone turn involving residues Ile3-Gln4-Asn5 is observed in the [Mpa1,l-Tic7]OT peptide, supported by (i, i + 2) type connectivities between residues Ile3-Asn5, similar to the [l-Tic7]OT analogue. In contrast, their side chains are oriented to opposite directions, possibly due to the observed NOEs between amino acids Cys6-Leu8, present only in the [Mpa1,l-Tic7]OT peptide (Fig. 3).

The NMR data for [d-Tic7]OT and [Mpa1,d-Tic7]OT peptides indicate that residues in positions 1 and 7 are in close spatial proximity (Fig. 3). Medium-range NOEs between Cys1 and Asn5 present in the [d-Tic7]OT peptide and NOEs between Mpa1 and Cys6 in the [Mpa1,d-Tic7]OT peptide support the vicinity of these peptide residues. The formation of a hydrogen bond between Ile3-Asn5 observed in all 20 calculated models of [Mpa1,d-Tic7]OT peptide further supports this finding. The overall conformation of the backbone for the segment Cys1/Mpa1-Cys6 reveals great structural similarities between the two analogues, but also some differences. NOE (i, i + 2) connectivities between Cys1 and Ile3 indicate the formation of a backbone bend for this segment in the [d-Tic7]OT peptide. On the other hand, NOE interactions between (1) Hβ protons of Tyr2 with the amide proton of Gln4 and (2) the amide proton of Ile3 with the amide proton of Asn5 and Cys6 lead the [Mpa1,d-Tic7]OT peptide to adopt a turn-like structure comprised of these residues.

As far as the fragment of Cys6-d-Tic7-Leu8-Gly9NH2 is concerned; both OT peptides exhibit a backbone turn due to the (i, i + 2) type connectivities of Hβ proton of Cys6 with the amide proton of Leu8, observed in both peptides. However, NOE connectivities of Hβ proton of Cys6 with the Hβ and the side chain protons of Leu8, present only in the [d-Tic7]OT peptide, lead the two amino residues to adopt a parallel orientation in space with the side chain protons of Leu8 lying in the same region in space with Hβ protons of Cys6. The segment comprised of Leu8-Gly9NH2 exhibits a rather extended backbone conformation for both [d-Tic7]OT and [Mpa1,d-Tic7]OT peptides.

As a conclusion, deamination in position 1 influences the conformational flexibility of Cys1/Mpa1–Cys6, which in turn affects the geometry of the disulfide bond and imposes a conformational difference among the peptides. On the other hand, the change of configuration of amino acid in position 7 (replacement of l-Tic7 by d-Tic7) is inducing smaller conformational changes in the backbone of the OT peptides in the case of deamino analogues than the amino analogues (Fig. 3). The N-terminal 7-residue fragment of the [Mpa1,l-Tic7]OT and [Mpa1,d-Tic7]OT peptides adopts the very same backbone conformation (Fig. 3). The backbone structure of the 3-residue fragment of the C-terminus is also similar among the peptides with the different orientation, resulting from replacement of l-Tic7 by d-Tic7, being the main difference among the [Mpa1,l-Tic7]OT and [Mpa1,d-Tic7]OT peptides. Finally, it is worth noting that the distribution of the amino acids in the tertiary structures leads the peptides to form a polar face and a more polar side. Tyr2, Gln4 and Asn5 lead the peptide to adopt a solvent exposed polar face, while the hydrophobic aminoacids Tic7 and Leu8 form the C-terminal tripetide.

Discussion

Literature data indicate that the neurohypophyseal hormone analogues lacking the N-terminal amino group are inactivated slower than the mother compounds and have enhanced biological activities both agonistic and antagonistic ones (Oldziej et al. 1995; Fragiadaki et al. 2007). The anti-oxytocic activity was significantly enhanced in the case of both analogues [Mpa1,l-Tic7]OT and [Mpa1,d-Tic7]OT, while the agonistic activity was only slightly affected (Bélec et al. 2001). Furthermore, the first amino acid seems to play a decisive role in receptor binding (Oldziej et al. 1995; Fragiadaki et al. 2007). In our case, deamination slightly increased the affinity of the analogues to human oxytocin receptor. It has been known that the proper orientation and the sequence of the C-terminal tripeptide are critical for obtaining high-potency OT analogues (Wittelsberger et al. 2005; Oldziej et al. 1995; Fragiadaki et al. 2007; Flouret et al. 2006). Moreover, the side chains of Ile3 and Pro7 are involved in the recognition and binding of the hormone by the uterine receptor (Wittelsberger et al. 2005; Oldziej et al. 1995; Fragiadaki et al. 2007; Braunschweiler and Ernst 1983). Therefore, structural modifications of the side chains in the C-terminal tripeptide might lead to analogues with variable biological properties at different OT issues. The great decrease in the agonistic potency of [l-Tic7]OT analogue as compared to the native hormone suggests that modification at position 7 by conformational restricted and bulky residue, such as l-Tic, induces such conformational changes in the peptide backbone that cause markedly different distribution of the elements necessary for the binding of agonists and intrinsic activity.

According to our structural analysis, the replacement of amino acid l-Tic7 by the more stereo-chemically constrained d-Tic7 seems to induce minor conformational changes in the backbone conformation, with the main difference being the altered side chain orientation of the C-terminal 3-residue fragment of the peptides. The presence of the d-Tic aminoacid in position 7 results in a more compact structure by limiting the local conformational flexibility, inducing a bend in the backbone conformation. This structural difference is reflected in the biological properties, with the d-counterpart being a pure antagonist with higher antagonistic potency than the l-counterpart. This finding is also confirmed in the [Mpa1,l-Tic7]OT-[Mpa1,d-Tic7]OT peptide pair, with the anti-oxytocin activity being significantly enhanced in the [Mpa1,d-Tic7]OT peptide. It is worth noting that the main conformational difference resulting from the substitution of l-Tic7 with d-Tic7 in this peptide pair is observed for the backbone conformation of the Leu8-Gly9 NH2 fragment. The studied analogues are also exhibiting remarkable conformational properties from the point of view of the cis/trans isomerization of the Cys6-Tic7 peptide bond. The orientation of this bond and its relationship to agonism and antagonism has been previously studied; however, no direct relationship was found (Shenderovich et al. 1997; Wood et al. 1986). The introduction of the non-proteinogenic and less bulky Mpa residue in position 1 is not affecting the nature of the disulfide bond linking residues in positions 1 and 6 in oxytocin. On the contrary, the introduction of the d-Tic isomer into position 7, when position 1 is occupied by Cys, alters the geometry of the Sγ and neighboring atoms of the S–S bond (Fig. 4). The rearrangement of the C-terminal 3-residue fragment can be focused on the variation of the geometry of the Cys6-Tic7 peptide bond. In both [d-Tic7]OT and [Mpa1,d-Tic7]OT analogues, a second set of peaks for most residues of the peptide sequence has been observed in all spectra used (NOESY, TOCSY and 13C-HSQC). The population of these species is ranging between 25 and 30%, suggesting the presence of cis-isomerization in the peptide bond between positions 6 and 7. On the other hand, in both cases where l-Tic occupies position 7 in the peptide sequence, a unique set of peaks is observed, suggesting that Cys6-l-Tic7 bond is found in trans conformation and no cis-isomers are present.

Fig. 4
figure 4

Schematic representation and projection of the disulfide bond in the [d-Tic7]OT, [l-Tic7]OT, [Mpa1,d-Tic7]OT and [Mpa1,l-Tic7]OT analogues. Figures are generated with MOLMOL

According to our structural investigation of the four OT analogues and bearing in mind that d-counterparts showed higher inhibitory potency than the l-counterparts, the reduction in agonistic activity is fully consistent with the reduction of the trans conformation form of Cys6-Tic7 peptide bond.