Disorder-to-Order Markers of a Cyclic Hexapeptide Inspired from the Binding Site of Fertilin β Involved in Fertilization Process.

Synthetic peptides mimicking the binding site of fertilin β to its receptor, integrin α6β1, were shown to inhibit sperm-egg fusion when added to in vitro media. In contrast, the synthetic cyclic hexapeptide, cyclo(Cys1-Ser2-Phe3-Glu4-Glu5-Cys6), named as cFEE, proved to stimulate gamete fusion. Owing to its biological specificity, this hexapeptide could help improve the in vitro fertilization pregnancy rate in human. In an attempt to establish the structure-activity relationship of cFEE, its structural dynamics was herein analyzed by means of ultraviolet circular dichroism (UV-CD) and Raman scattering. The low concentration CD profile in water, containing mainly a deep minimum at ∼202 nm, is consistent with a rather unordered chain. However, an ordering trend of the peptide loop has been observed in a less polar solvent such as methanol, where the UV-CD signal shape is formed by a double negative marker at ∼202/215 nm, indicating the presence of a type-II' β-turn. Raman spectra recorded in aqueous samples upon a 100-fold concentration increase, still showed an important population (∼30%) of the disordered structure. The structural flexibility of the disulfide bridge was confirmed by the Raman markers arising from the Cys1-Cys6 disulfide bond-stretch motions. Density functional theory calculations highlighted the formation of the type-II' β-turn on the four central residues of cFEE (i.e., -Ser2-Phe3-Glu4-Glu5-) either with a left- or with a right-handed disulfide. The structure with a left-handed S-S bond, however, appears to be more stable.

Introduction

fertilin β, also known as ADAM2, is part of a disintegrin and metalloprotease (ADAM) family of proteins. ADAMs are widely expressed with various functions in cell–cell and cell–matrix interaction. Their common feature is to express a disintegrin-like domain that could bind to receptors like integrins,[1] a family of heterodimeric transmembrane receptors composed of covalently linked and β subunits. Fertilin β is located on spermatozoa, whereas its binding receptor the integrin 6β1 is located on oocytes. Numerous studies have highlighted the importance of the binding between fertilin β and integrin 6β1 in the sperm–egg binding process, allowing their fusion.[2,3] Indeed, in vitro studies showed that sperm–egg fusion is greatly reduced when the interaction between fertilin β and integrin 6β1 is specifically inhibited. Such inhibition was performed by adding to in vitro media, synthetic peptides mimicking the binding site of the fertilin β (e.g., the disintegrin domain), so that they act as a competitor.[3−7] Therefore, the binding between fertilin β and integrin 6β1 was identified as a crucial step for egg–sperm fusion. However, attempts based on using gene knockout experiments have revealed that several sperm factors identified from the in vitro system are in fact not essential for in vivo fertilization.[8] Indeed, fertilization was proven to be still possible when genes of fertilin β are deleted.[9−11] Fertilin β is actually believed to play a role in sperm transport in the oviduct rather than in sperm–oocyte fusion.[8] Nevertheless, regardless of its activity concerning the in vivo fertilization process, it is clear that the inhibition of the fertilin β binding to α6β1 inhibits egg–sperm fusion in vitro, thus becoming crucial importance upon the consideration of the in vitro fertilzation (IVF) approach.

The binding sequence of fertilin β is species specific. In human, this sequence corresponds to -Phe-Glu-Glu- (or FEE).[12] Ziyyat et al.[13] evaluated the role of fertilin β in human gamete fusion, by using a peptide reproducing its binding site to integrin 6β1. As the binding domain of fertilin β is localized at the top of a hairpin loop, to reproduce its natural conformation, a cyclic hexapeptide, NH2-CSFEEC-COOH (cyclic part underlined), hereafter referred to as cFEE, was used. Unexpectedly, while most peptides and recombinant forms of fertilin inhibit sperm–egg fusion, cFEE was shown to lead to an increase of the gamete fusion.[13] In fact, cFEE was shown to bind to the oocyte membrane, simulating spermatozoon contact by inducing the displacement of the adhesion protein to the oocyte surface. Similar results were observed upon a clinical trial on mice by the use of the mouse sequence equivalent to cFEE (cQDE), leading to the improved IVF pregnancy rate in mice.[14] It has been thought that this specific biological effect of cFEE in vitro might be explained by both its amino acid composition and cyclic structure maintained by a disulfide bridge between its two terminal cysteines.[13,14] Nevertheless, a convincing explanation has not yet been provided.

Therefore, it appeared to us interesting to analyze the structural dynamics of this cyclic hexapeptide to help understand its specific action (vide supra) that can be envisaged in IVF pregnancy rate improvement in human. On the other hand, the knowledge of the conformational features of cFEE might lead to design other peptides with higher IVF success rates. Herein, at a first stage of these investigations, the structural analysis of cFEE has been undertaken by means of optical spectroscopic techniques such as circular dichroism (CD) and Raman scattering. During the past decade, the joint use of these approaches has rendered possible the analysis of the structural properties of both linear and cyclic peptides in aqueous media as a function of their length, concentration, as well as of the ionic strength, pH, and time.[15−21] Furthermore, in many cases, the role of the surrounding solvent, in going from high to low dielectric media, favoring the intermolecular interactions rather than those from the intramolecular type has also been elucidated.[15−18] As concerns about the cyclic peptides, while CD spectra give rapid and general information on their folding type, Raman spectra provide details on the rate of different secondary structural elements (disordered, turn, and strand) appearing in their closed part.[19−22]

Results and Discussion

LC-MS/MS Analysis

The characterization based on reverse phase liquid chromatography hyphenated to high-resolution tandem mass spectrometry (LC-MS/MS) was made with the aim of providing orthogonal information regarding the integrity of the cyclic peptide and to eventually identify the presence of interfering impurities. Considering the obtained chromatogram, the presence of the hexapeptide cFEE was confirmed. [M + H]+ was observed at 715.2051 (Figure S1A, Supporting Information), with a mass deviation from the theoretical value not exceeding 2.5 ppm. This represents conventionally the instrumental performance. The present data with previous ones[24] and the resolution provided by the MS instrument allowed elucidating the isotopic distribution of cFEE, being in complete correlation with its cyclic structure.

Concomitantly, during the analysis, fragmentation by collision-induced dissociation (CID) was performed to investigate more precisely the structure of the compound. As illustrated in Figure S1B (Supporting Information), MS/MS spectra corresponding to cFEE mainly demonstrated the detection of the fragments corresponding to the b-ion series. The identified fragments demonstrated to be particularly consistent with the amino acid sequence of the cFEE peptide. Except cFEE, the investigation of LC-MS/MS data did not reveal the presence of any additional compound, thus excluding the presence of probable impurities.

UV-CD Markers

UV-CD spectra recorded at 300 μM in methanol and water are displayed in traces A and B of Figure , respectively. Upon comparison, the effect of the solvent type on the observed spectra can be clearly seen on CD spectra. In both media, all the spectra are mainly characterized by a deep minimum at ∼202 nm. However, in methanol (Figure A), a distinct shoulder at ∼215 nm was detected at the long wavelength side of the main minimum. On the contrary, the pH effect (from acidic toward neutral pH, Figure B) is reflected by an overall decrease of the negative band at ∼202 nm, as well as an increase of the positive signal at ∼220 nm, without changing the overall dichroic shape of the CD spectra. It can be reminded that the pKa corresponding to the Glu side-chain protonation/deprotonation is ∼4.1. As a result, the two selected pH values favor as major population either the protonated (pH 3) or the deprotonated (pH 7) side chains of Glu4 and Glu5 residues.

Figure 1

Room-temperature UV-CD spectra of cFEE displayed in the 190–280 nm region. (A) Data recorded in methanol at 200 μM. (B) Data recorded in aqueous samples at 300 μM as a function of pH.

The influence of the solvent polarity on the peptide structural features has been detailed previously.[15,16] Some linear peptides were reported to undergo a β-strand → α-helix transition from water (a polar solvent) to another having a lower dielectric constant (alcohols). This effect has been interpreted by the capacity of a polar solvent to reinforce the intermolecular H-bonds, rather than the intramolecular ones.[15] Beyond the solvent polarity, the influence of peptide length, concentration, and eventually its exposure time to the solvent appeared to act as the key structuring elements.[16−18] In cyclic peptides, the closure constraint (imposed for instance by a disulfide bond), the length, and the amino acid composition of its closed part were shown to be the key elements in their structural behaviors.[19−21]

Focusing on UV-CD structural markers, a random chain is routinely characterized by a deep negative dichroic band at ∼198 nm,[25,26] whereas a loosely structured loop (i.e., subjected to a rapid interconversion of several instable turns) provides a slightly red-shifted negative CD signal to ∼202 nm.[19] As a consequence, the characteristic CD fingerprint of cFEE in water, independently of the pH value (Figure B), is consistent with the formation of an unordered/weakly ordered turn, whereas the double negative band shape reveals the presence of a structured loop formed in methanol (Figure A). Previously, upon an extensive study of the CD and NMR data obtained from 10 gramicidin S cyclic peptides, all adopting a reverse type-II′ β-turn, it has been highlighted that their CD profile is composed of a double negative bands at ∼202 and ∼215 nm. Furthermore, the dichroic ratio [Φ2]/[Φ1], where [Φ2] and [Φ1] are the normalized ellipticities corresponding to the higher (∼215 nm) and lower (∼202 nm) wavelength components of the characteristic double negative CD fingerprint, respectively, can be taken as an indicator of their β-turn stability.[27] More precisely, while a compact (or highly stable) type-II′ β-turn gives a [Φ2]/[Φ1] ratio close to unity,[21] a less stable turn of the same family provides a lower (<1) ratio. Following these evidences, the CD band shape of cFEE (Figure A), with a [Φ2]/[Φ1] ratio of ∼0.4, is consistent with a rather instable (mobile) type-II′ β-turn in methanol. Other types of β-turns provide, noteworthy, a CD signal different from those observed in cFEE. For instance, a type-I β-turn is characterized by double minima at ∼208/222 nm, resembling that traditionally assigned to an α-helix, a type-II β-turn gives a positive band located within the 190–210 nm region, and finally, a reverse type-I′ β-turn provides a broad negative band at ∼215 nm.[25]

Raman Markers

Raman spectra of cFEE, recorded at 30 mM in the middle wavenumber region in aqueous samples, are displayed in traces A and B of Figure corresponding to pH 3 and 7, respectively. To achieve a reliable assignment of the observed Raman bands, we resorted to those from their building blocks. For this, the Raman spectra of free Ser, Phe, and Glu are shown in Figure A–D. Because of their full analysis in both solid and solution samples,[28] the Raman data of cysteine (i.e., Cys-Cys dimer with a disulfide bond) were not reported here. Selected pH values of the solution samples of free AAs favored their zwitterionic species (with NH3+/COO– backbone end groups). This fact can be verified for instance by the broad band at ∼1602 cm–1 corresponding to the carboxylate (COO–) asymmetric stretching mode (Figure A,C,D). Because of the absence of any internal reference, to facilitate their comparison, all spectra were normalized to their intensities in the 1500–1400 cm–1 region, arising from the aliphatic modes (angular bending motions of CH2 groups). The strongest Raman bands naturally correspond to the Phe aromatic markers located at 1606, 1586, 1207, 1032, 1004, and 622 cm–1 (Figure B). It is interesting to note that the spectrum of Glu0 species (with a protonated side chain) recorded at an acidic pH (2.9) clearly reveals the presence of the broad band at ∼1723 cm–1 assignable to the carbonyl (C=O) bond-stretch marker resulting from the side-chain head group (COOH) (Figure D). At pH 6, upon the predominance of Glu1– species in the solution (with a deprotonated side-chain head group COO–), the carbonyl marker expectedly vanishes (Figure C). The comparison between the spectra of Glu0 and Glu1– species permits locating the markers affected by the side-chain protonation/deprotonation within three distinct regions, that is, 1730–1700, 1350–1300, and 950–850 cm–1 (bands marked by pink arrows in Figure C,D). For the detailed assignment of the Raman bands, see refs (29) and (30) for Phe and ref (31) for Ser. As far as Glu is concerned, we present in Table S1 (Supporting Information) the theoretical wavenumbers of the main Raman bands along with their assignments from the presently performed DFT calculations on the lowest energy conformers of both Glu1– and Glu0 species (Figure G,H) (see also Results and Discussion for details). In particular, the Raman markers shown by pink arrows in Figure C,D are assigned to a mixture of side-chain/backbone vibrational motions, thus explaining their sensitivity to the side-chain head group COOH → COO– conversion.

Figure 2

Room-temperature Raman spectra recorded from the aqueous samples of cFEE. The spectra obtained at (A) pH 3 and (B) pH 7. The tentative assignment of the main Raman bands is made on the basis of the data corresponding to free AAs (Figure ). Amide (I and III) and disulfide bond regions are noted. See Figures and 5 for band decomposition of these particular regions. The characteristic Raman bands sensitive to the Glu side-chain protonation/deprotonation are marked by pink arrows. Phe and bk designate the Raman bands assigned to phenylalanine residue and backbone, respectively.

Figure 3

Room-temperature Raman spectra obtained from the aqueous samples of free AAs together with their lowest energy zwitterionic species. Raman spectra of (A) Ser, (B) Phe, (C) Glu1– (Glu with a deprotonated side chain), and (D) Glu0 (Glu with a protonated side chain). The characteristic Raman bands sensitive to the Glu side-chain protonaton/deprotonation are marked by pink arrows in (C) and (D). Lowest energy conformers of (E) Ser, (F) Phe, (G) Glu1–, and (H) Glu0. Side-chain torsion angles values and global orientations are reported at the right side of the graphical representations (E–H). See text for details. Theoretical level: DFT/B3LYP/6-311++G(d,p).

Figure 4

Band decomposition in the amide regions of Raman spectra. (A, B) Amide I and (C, D) amide III are decomposed at pH 7 (top) and pH 3 (bottom). Observed spectra are in red (pH 3) and blue (pH 7) colors. Band components are drawn in gray. Empty circles correspond to the sum of the components. See text for details and Table for normalized contributions and their assignments to different secondary structural elements (random, β-strand, and turn). Note that the component at 1721 cm–1 (pink color in (A)) arises from the stretching mode of the carbonyl bond appearing upon the Glu side-chain protonation at pH 3.

Figure 5

Band decomposition in the disulfide bond-stretch region of Raman spectra. Observed spectra are in (A) red (pH 3) and (B) blue (pH 7) colors. Band components are drawn in gray color. Empty circles correspond to the sum of components. See text for details.

The Raman bands observed in free AAs allowed us to propose in Figure A,B the tentative assignment of cFEE spectra. Whatever pH, the Raman intensity is expectedly dominated by that arising from Phe3. Nevertheless, the relative intensity of Glu markers is strong enough to permit appreciation of the changes occurring upon side-chain protonation/deprotonation (see medium bands marked by pink arrows in Figure A,B). For instance, the appearance of a shoulder at ∼1727 cm–1 at pH 3 should be noticed (Figure A), vanishing at pH 7 (Figure B). In the 950–900 cm–1 region, the broad and dissymmetric band peaking at ∼941 cm–1 at pH 7 (Figure B) is transformed to a set of partially resolved bands at ∼933, 922, and 906 cm–1 at pH 3 (Figure A).

The comparison between the Raman spectra of cFEE (Figure A,B) and those from free AAs (Figure A–D) also allows analyzing the vibrations arising from the peptide backbone, referred to as amide I (1750–1625 cm–1) and amide III (1320–1225 cm–1) modes. Both amide regions involve broad and barely resolved bands, consistent with the presence of several secondary structural elements in cFEE. Band decomposition in amide (I and III) regions (Figure ) leads us to achieve detailed structural information. The protocol used for band decomposition was established on the basis of the previous works on linear[16−18] and cyclic[19−22] peptides. Table shows the normalized contributions (as expressed in percent) of the band components at pH 3 (Figure A,C) and pH 7 (Figure B,D). Comparable results were obtained from amide (I and III) regions, confirming the reliability of band decomposition and corresponding assignments. As far as the influence of pH is concerned, apart from a slight decrease of ordered elements (turn and β-strand) versus a slight increase of random chain contribution, no other perceptible effect can be noticed in going from an acidic toward a neutral pH value (Table ).

Table 1

Normalized Contributions of Different Secondary Structural Elements As Determined by the Analysis of Amide (I and III) Regionsa

amide I (components)	turn (%)	β-strand (%)	random (%)
pH 3
1695			27
1682	41
1664		22
1644	10
pH 7
1690			32
1675	24
1660		20
1643	24

The components are referred to by their maximum wavenumber expressed in cm–1. See Figure for band decomposition of amide (I and III) regions.

The 550–480 cm–1 region of the Raman spectra involving the S–S bond stretching vibrations is zoomed in Figure . Independently of the pH value, band decomposition makes clearly appear two components at ∼509 and ∼525 cm–1, assignable to two different types of rotamers around the S–S linkage in cFEE (see for details ref (28)); the normalized area corresponding to these two components is estimated as ∼80% (for the ∼509 cm–1 component) versus ∼20% (for the ∼525 cm–1 component). Although the rotamer relative to the ∼509 cm–1 component should correspond to the major disulfide rotamer in aqueous solution, the presence of the second component at ∼525 cm–1 with lower intensity reveals the conformational flexibility of the S–S linkage.

Type-II′ β-Turn Structural Models for cFEE

UV-CD spectra led us to presume that a loosely structured type-II′ β-turn can be formed in methanol (see above UV-CD Markers for details). It is worth emphasizing that there exist two reverse type β-turns, referred to as type-I′ and type-II′. Both of them are considered as frequently occurring β-turns in proteins.[36,37] Briefly, a β-turn is composed of four residues, generally numbered as i, i + 1, i + 2, and i + 3 in which the chain direction change is made possible by the special values of the backbone torsion angles (φ, ψ) assigned to the two middle i + 1 and i + 2 residues.[32,33] Particularly, in a type-II′ β-turn, the mentioned angles fluctuate around the following mean values (φ ∼ +60°, ψ ∼ −120°, φ ∼ −80°, ψ ∼ 0°).

To prepare the initial conformers of cFEE, we took an advantage of the extensive structural data corresponding to a synthetic cyclic octapeptide, named octreotide (or Sandostatin). Octreotide with the primary sequence Nter-D-Phe1-cyclo(Cys-Phe-D-Trp-Lys-Thr-Cys)-Thr(ol)8 is one of the rare therapeutic peptides for which crystal,[34] NMR,[35] CD, and Raman[20,22] data, as well as DFT calculations,[36] are available. All these data confirm a type-II′ β-turn formed on the four central residues (-Phe3-D-Trp4-Lys5-Thr6-) of this peptide. Supposing that a similar turn can also be formed on the four central residues of cFEE (-Ser2-Phe3-Glu4-Glu5-), the backbone torsion angles (φ, ψ, ω) of octreotide[36] were transferred to the initial structure of cFEE. As far as the torsion angles defining the orientation of the side chains of Ser2, Phe3, Glu4, and Glu5 are concerned, we resorted to the DFT data existing on free AAs reported either in the previous investigations on Phe[37] and Ser[31] or derived from the presently performed calculations on Glu. The lowest energy conformers of Ser, Phe, Glu0, and Glu1– are displayed in Figure E–H, respectively. At the right side of each conformer, the values of their side-chain torsion angles, that is, (χ1, χ2) for Ser and Phe and (χ1, χ2, χ3) for Glu. As far as the cysteine (Cys1-Cys6) conformation is concerned, we were inspired from a systematic structural analysis on the functional disulfide bridges in proteins.[38−41] The conclusion of this study was that a set of five torsion angles (χ1, χ2, χ3, χ2′, χ1′) defined around the five chemical bonds (−Cα–Cβ–S–S–Cβ′–Cα′−) along the cysteine moiety are necessary to accurately define the conformational feature of a given S–S bridge. The torsion angles involved in the optimized structure of octreotide were (χ1 = −166°, χ2 = +73°, χ3 = −97°, χ2′ = +111°, χ1′ = −179°).[36] It is to be noted that this set of torsion angles corresponds to a left-handed S–S bridge because of the negative sign assigned to the χ3 angle defined around the S–S bond. To analyze the effect of the S–S bridge handedness, we have also constructed a second initial conformer having a right-handed S–S linkage (having a positive χ3 angle).

The optimized geometries of cFEE having either a left-handed (L) or a right-handed (R) disulfide linkage are reported in Figure A,B, respectively. Note that in both structures, Glu4 and Glu5 have a protonated side chain, reflecting a low dielectric constant environment such as methanol. For the same reason, the peptide backbone end groups were supposed to be neutral (NH2/COOH). The optimized torsion angles of these conformers are reported in Table . As it can be seen, the type-II′ β-turn can be stabilized both with a left- or a right-handed disulfide. Nonetheless, a higher stability (with a lower relative energy) was predicted for the (L) structure (Figure ). The interactions that stabilize both (L) and (R) structures are to be noticed as follows: (i) a quite short contact in the peptide backbone, (Ser2)N–H···O–C(Glu5), with a length varying between 2.02 and 2.20 Å, acting as the β-turn closing the H-bond and (ii) a set of three other H-bonds involving the side-chain head groups of Ser2 (O–H), Glu5 (COOH), as well as the backbone carbonyl C=O (Ser2) and N–H (Glu5) groups (with the lengths varying between 1.99 and 2.23 Å). To better appreciate the (L) and (R) rotamers of the S–S bridge, the cysteine moiety (Cys1-Cys6) is zoomed in Figure , along with the values of their characteristic torsion angles (χ1, χ2, χ3, χ2′, χ1′). Following the original terminology initiated by Hogg and co-workers,[38] the optimized left-handed disulfide bridge can be considered as a “-LHStaple”-type rotamer (Figure C), whereas the right-handed one is rather a “-RHHook”-type rotamer (Figure D). In these notations, “LH” and “RH” designate left-handed and right-handed disulfides, respectively. “Staple” and “Hook” refer to their spatial shapes, and finally, the sign “-” reflects the negative values assigned to the terminal χ1 and χ1′ torsion angles. It is worth noting that there also exists a traditional type of notation for defining the S–S bridge conformation, based on the g (gauche) and t (trans) orientations of the three middle torsion angles (χ2, χ3, χ2′), without mentioning their signs. Following these notations, the left-handed S–S bridge corresponds to a “ggg” conformer, whereas the right-handed one can be classified as a “ggt” conformer. For further use of the optimized structures, the corresponding atomic Cartesian coordinates are reported in Table S2.

Figure 6

Graphical representation of the optimized structures of cFEE in a polarizable continuum mimicking methanol. The structures having (A) a left-handed and (B) a right-handed disulfide bridge. The locations of the type-II′ β-turn and disulfide bridge are noted. Intramolecular H-bonds are drawn in green broken lines, and their lengths (in Å) are reported. The energy difference between the two structures (ΔE in kcal/mol) is mentioned. Theoretical level: DFT/B3LYP/6-31G(d).

Table 2

Backbone and Side Chain Torsion Angles of cFEE Structures As Determined by DFT calculationsa

DFT calculation at the B3LYP/6-31G(d) level on the hexapeptide embedded in a polarizable continuum mimicking methanol.

For the graphical representations, see Figure . Torsion angles are expressed in degrees. φ, ψ, and ω refer to the backbone torsion angles of the residues involved in the structure of the hexapeptide.χ1, χ2, and χ3 torsion angles allow the orientation of the side chain of a given residue to be determined. Asterisk (*) corresponds to side-chain torsion angles of the cysteine (Cys1-Cys6) moiety that are reported in Figure along with the corresponding graphical representation of the disulfide bridge.

Figure 7

Zoom on the disulfide linkage of the optimized conformers of cFEE. Two views of each structure are reported. (A, C) left-handed disulfide bridge and (B, D) right-handed disulfide bridge. The distance between the Cα atoms belonging to Cys1 and Cys6 residues is reported in green broken lines of which the length (in Å) is also mentioned. The values of the five torsion angles defining the conformation of each disulfide bridge are mentioned at the right side of the representations (A) and (C). See text for details.

Comparison with Other Cyclic Peptides

To better understand the structural behavior of cFEE, we briefly compare it to that of other short cyclic peptides having preferentially an identical loop size. The natural hormone oxytocin, Nter-cyclo(Cys-Tyr-Ile-Gln-Asn-Cys)-Pro7-Leu8-Gly9-CONH2, was provided in TFE (εr = 8.55),[42] a UV-CD profile similar to that obtained from cFEE in methanol (Figure A). In water (pH 7.3), its CD signal was rather consistent with an unordered chain.[43] Another structurally similar hormone, Arg-vasopressin, Nter-cyclo(Cys-Tyr-Phe-Gln-Asn-Cys)-Pro7-Arg8-Gly9-CONH2, was shown to be unstructured in an aqueous environment (pH 6.9) as characterized by a deep negative CD profile peaking below 200 nm.[44] In other words, both mentioned neuropeptides presented a structural dynamics close to that of cFEE. Taking into account the fact that all three peptides possess a loop formed by six residues, their common structural features should undoubtedly arise from their amino acid composition and especially from those of the four loop residues, that is, -Tyr2-Ile3-Gln4-Asn5- (in oxytocin), -Tyr2-Phe3-Gln4-Asn5- (in Arg-vasopressin), and -Ser2-Phe3-Glu4-Glu5- (in cFEE). It is worth noting that all three sequences contain at least three residues with a pronounced hydrophilic character (selected among Ser, Tyr, Asn, Gln, and Glu residues). Recent DFT calculations highlighted the propensity of Phe and Tyr residues to be hydrated through the stabilization of short contacts between water hydrogen atoms and aromatic π-electron systems.[37] Nevertheless, the main structural difference between cFEE and the two mentioned neuropeptides seems to be in the handedness of their S–S bonds. In fact, previous Raman optical activity (ROA) measurements led to the conclusion that oxytocin and Arg-vasopressin contain both a right-handed S–S bond,[45] whereas the presently reported DFT calculations in methanol are consistent with a left-handed S–S bond as the major rotamer of cFEE (see above for details).

Structural Argument Put Forward for Explaining the Particular Biological Activity of cFEE

As the culture media for IVF is an aqueous media with a pH at 7.3, the reason behind the particular biological behavior of cFEE (see Introduction for details) might be dependent on its unordered character. To support this hypothesis on the structure–activity relationship of therapeutic peptides, we can mention the example of the natural hormone somatostatin-14 (SST-14), acting particularly as an inhibitor of the growth hormone secretion through its binding to one of the five G-protein-coupled receptors, referred to as SSTRi (i = 1,...,5).[46] SST-14 is a cyclic tetradecapeptide with a large size loop formed by 12 residues. This hormone was shown to form an unordered chain, as confirmed by CD, Raman,[19,21,22] and NMR data.[47−49] Apart from its loop size, the conformational flexibility of SST-14 can be related to the presence of seven intraloop hydrophilic residues. Because of the plasma short half-life of SST-14, during the past decades, a series of synthetic analogues were elaborated and used in various somatostatin-based clinical therapies.[50,51] Among them, one can notice two widely spread cyclic octapeptides, octreotide and lanreotide, both forming a stable type-II′ β-turn on their four central residues.[20,21] Interestingly, these two analogues have shown a selective binding affinity toward two SST-14 receptors, that is, SSTR2 and SSTR5. The difference between the activities of SST-14 and its synthetic analogues might be better understood by their structural features. More explicitly, while SST-14 with its loose turn can equivalently interact with its five receptors presumably through an induced fit process,[52] the conformational restriction leads to considerably modulate its analogue activity.

Concluding Remarks

Because of the rapidity in acquisition and post-treatment of their data, the joint use of UV-CD and Raman spectra is now considered as a powerful method to probe the solution structural dynamics of the peptides, especially when they are unordered or barely structured. Herein, the application of this methodology evidenced that the cyclic therapeutic peptide cFEE adopts a rather unordered loop in aqueous media, without a perceptible structural transition from an acidic to a neutral pH value, being responsible for the side-chain protonation/deprotonation of its two adjacent Glu residues. Furthermore, Raman markers analyzed upon a ∼100-fold concentration increase still revealed that an important population of the random chain persists in cFEE. Nevertheless, a structural ordering trend has been observed in methanol through the appearance of type-II′ β-turn UV-CD markers. DFT calculations confirmed the formation of this reverse type β-turn in a molecular model embedded in a polarizable continuum mimicking methanol. It appeared that an intraloop H-bond network might contribute to maintain the structured loop in a low dielectric constant medium. It has thus been concluded that the disordered character of cFEE in water originates presumably from the preponderance of intermolecular H-bonds between the three hydrophilic residues (Ser2, Glu4, and Glu5) with surrounding water molecules. Beyond the static QM calculations described in the present analysis, further molecular dynamics simulations in the presence of explicit water molecules would be necessary to highlight (i) the loop conformational flexibility within a sufficiently extended timescale and (ii) the residence time of water molecules on the loop residues. The latter point would lead us to a better understanding of the loop structural invariability either with protonated or with deprotonated Glu residues, probably caused by the solvent screening of their electrostatic interactions.

Material and Methods

Sample Preparation

The lyophilized sample of the cyclic peptide cFEE (Scheme 1, Supplementary Information) was purchased from Synprosis SA Laboratory (Fuveau, France). Solution samples were prepared by dissolving the hexapeptide in water taken from a Millipore filtration system (Merck, Molsheim, France). Stock solution of the hexapeptide was prepared at 30 mM, that is, ∼21.5 mg/mL. These samples were used in Raman spectroscopic measurements. Further dilution to lower concentrations was made for other experiments. Upon dissolution, the pH value of the peptide sample was ∼3; it was adjusted at higher values by adding NaOH (1 N) to aqueous samples. Free amino acids (AAs), Ser, Phe, and Glu, were purchased from Sigma-Aldrich (Saint-Quentin-Fallavier, France). The concentration of their aqueous samples used for Raman spectroscopy was 50 mM. The pH of Glu samples was adjusted to render possible the analysis of the side-chain protonation/deprotonation by Raman markers. Other used chemicals were also provided by Sigma-Aldrich.

Liquid Chromatography-Tandem Mass Spectrometry Analysis (LC-MS/MS)

The hexapeptide was characterized by high-resolution mass spectrometry hyphenated to reverse phase liquid chromatography (RPLC) on a Waters BEH C18 (1.7 μm, 2.1 mm × 150 mm) column using an ultrahigh-performance liquid chromatography system (Waters ACQUITY UPLC; ACQUITY column manager-type ACQCM 1.40, binary solvent manager-type ACQ-BSM 1.65, sample manager-type ACQ-SM 1.65; Manchester, U.K.). The mobile phases were composed of 0.1% formic acid (FA) in water (mobile phase A) and 0.1% FA in acetonitrile (mobile phase B). The LC-MS/MS analysis was performed using a mobile phase flowrate of 100 μL/min and a column temperature of 40 °C. Peptide separation was carried out using a gradient from 5 to 80% B for 38 min followed by 80% B maintained for 3 min. Then, the column was reconditioned using 5% mobile phase B during 10 min. A volume of 10 μL, corresponding to 2.5 μg of cFEE, was injected and analyzed online with an LTQ-Orbitrap XL mass spectrometer (Thermo-Fisher Scientific, Manchester, U.K.) hyphenated through the intermediate of a heated electrospray ionization probe (HESI-II, Thermo-Fisher Scientific). Data acquisition was controlled by Xcalibur software (Thermo-Fisher Scientific). ESI source parameters were set as follows: ESI voltage of −4.0 kV while sheath gas and auxiliary gas flowrates were set to 40 and 12, respectively. The source heating temperature was set to 300 °C, capillary temperature was set to 320 °C, and capillary voltage to a value of 35 V. The tube lens voltage was set to 90 V. MS/MS was realized in an m/z targeting approach. Fragmentation was performed using collision-induced dissociation (CID). The parameters were set to a normalized collision energy of 35%, an activation time of 30 ms, and an isolation width of 2 Th. The mass/charge (m/z) range was 150–2000 in MS and 100–2000 in MS/MS. Using those parameters, the mean resolution provided by the instrument was 30,000 in MS (m/z = 715.2062).

Spectroscopic Setups and Post-Record Data Analysis

Room-temperature UV-CD spectra were analyzed on a JASCO J-810 spectrophotometer (Lisses, France) within the 190–300 nm spectral range. A path length of 1 mm and a spectral resolution of 0.2 nm were selected. Each spectrum corresponding to an average of five scans was recorded with a speed of 100 nm/min. To facilitate the comparison of the CD spectra recorded in different conditions, the measured ellipticity for each sample, referred to as [ϕ]obs, was further normalized to obtain the so-called mean residue ellipticity, [ϕ], by using the expression: , where n, c, and l are the numbers of residues in the peptide, the molar concentration, and the optical path length of the sample, respectively.[23] The normalized ellipticity was expressed in deg cm2 dmol–1.

Room-temperature Stokes Raman spectra were analyzed at the right angle on a Jobin-Yvon T64000 spectrometer (Longjumeau, France) at a single spectrograph configuration, with a 1200 grooves/mm holographic grating and a holographic notch filter. Raman data corresponding to a 1200 s acquisition time for each spectrum were collected on a liquid nitrogen-cooled CCD detection system (Spectrum One, Jobin-Yvon). The effective slit width was set to 5 cm–1. Solution samples were excited by the 488 nm line of an Ar+ laser (Spectra Physics, Evry, France), with 200 mW power at the sample. Buffer subtraction of the observed spectra was performed using the GRAMS/AI Z.00 package (Thermo Galactic, Waltham, MA, USA). Final presentation of Raman spectra was done by means of the SigmaPlot package 6.10 (SPSS Inc., Chicago, IL, USA).

Quantum Mechanical Calculations

Energetic, geometrical, and vibrational data of cFEE and its building blocks were estimated by the density functional theory (DFT) approach,[53] using the hybrid B3LYP functional.[54,55] Taking into account the structural complexity of the hexapeptide (86 atomic centers), a cost effective, reasonable size Gaussian-type atomic basis set, that is, polarized double zeta enriched with d orbital functions on C, N, and O atoms, referred to as 6-31G(d), was used. Two starting conformers of cFEE (see Results and Discussion for details) were placed in a polarizable continuum model (PCM),[56,57] mimicking methanol (εr = 32.63). In contrast, the geometry optimization on free zwitterionic AAs was performed in a solvent continuum with εr = 78.39 (corresponding to water) by means of a more extended (triple zeta) basis set additionally equipped with diffuse functions on all atoms, as well as p orbitals on H atoms, referred to as 6-311++G(d,p). Geometry optimization was followed by the harmonic vibration calculation. The absence of any imaginary frequency proved the correspondence of the optimized conformer to a local minimum in the molecular energy landscape. The energy order of the optimized geometries of a given molecular species was determined on the basis of their total energy (Etot), where Etot = Ee + free energy correction. Each optimized conformer was identified by its relative energy (ΔE), that is, its energy difference with that corresponding to the lowest energy one for which ΔE = 0. All quantum mechanical calculations were made using the Gaussian09 package.[58]