conformation in oxytocin analogs

The nonapeptide oxytocin (OT) is used in medicine to help begin and/or continue childbirth. Its analogs can be also used to control bleeding following fetus delivery. The main function of oxytocin is to stimulate contraction of uterus smooth muscle and the smooth muscle of mammary glands, thus regulating lactation. This paper describes theoretical simulations of the distribution of the torsional angles chi1 in the non-standard methylated phenylalanine residues of three oxytocin analogs: [(Phe)(2)o-Me]OT, [(Phe)(2)m-Me]OT, [(Phe)(2)p-Me]OT. The conformations of the oxytocin analogs were studied both in vacuum and in solution. We found some correlations between the biological activity of the considered peptides and the side-chain conformations of amino-acid residues 2 and 8.

Currently, preterm birth is fairly frequent (about 10% of all births in USA) (Wyatt et al., 2001) and remains the main cause of neonatal mortality and morbidity.On the other hand, prolonged retardation of labor can be dangerous for the fetus health and life.Therefore studying novel, effective tools for controlling parturition is of utmost importance.Searching among analogs of the natural hormone oxytocin, involved in childbirth seems reasonable approach.Oxytocin is a nonapeptide involved in the induction of labor by stimulating uterine smooth muscle contractions and in stimulation of the milk ejection by mammary gland and of specific sexual, affiliative and maternal behaviors.The hormone and its analogs are used in medicine to help begin and/or continue childbirth or to control bleeding following fetus delivery.Oxytocin is produced in the supraoptic and paraventricular nuclei in the hypothalamus and is released to blood by posterior pituary gland.Human oxytocin has the following sequence: H-Cys-Tyr-Ile-Gln-Asn-Cys-Pro-Leu-Gly-NH 2 Oxytocin initiates its physiological activity by interacting with the G-protein-coupled receptor (GPCR) known as oxytocin receptor (OTR).
All members of GPCR superfamily, including OTR, consist of seven hydrophobic transmembranes-helices joined by alternating intra-and extracellular loops (Barberis et al., 1998).Since the work of Law and Vigneaud in 1960 (Law & Vigneaud, 1960) a large number of oxytocin analogs have been synthesized and many have been reported as oxytocin agonists or antagonists often more potent than the natural hormone (Toth et al., 1999).Nevertheless, no fully medical agents are available at present to treat woman in preterm labor.This is so mainly because of the low orall activity or undesirable maternal or neonatal side effects of the synthesized analogs (Higby & Suiter, 1999).Therefore a research for new effective antagonists blocking oxytocin receptor and preventing preterm labor appears necessary.It is believed the critical role in the agonistic-antagonistic properties of the neurohormon analogs is played by amino-acid residues in position 2 and 8 (Bankowski et al., 1980;Manning et al., 1984;1995;Hruby et al., 1994).Therefore main modifications concern these positions.In this work we investigated the structure of oxytocin analogs modified at the second position: Their interesting feature is the difference in the ago-B.Śmiech and others nistic and antagonistic properties which depends on the site of the methyl group substitution in the phenyl ring of phenylalanine.Their activity also strongly depends on the presence of magnesium ions.In their absence oxytocin analogs containing the ortho and meta isomer are antagonists, while that containing the para isomer is an agonist.In the presence of magnesium ions the peptide containing the meta isomer shows antagonistic activity, while the analogs with the ortho and para isomers are agonists (Slaninova et al., 2001).Oxytocin contains a ring created by six amino-acid residues connected by a disulfide bond between residues Cys 1 and Cys 6 , and a "tail" of three amino-acid residues.Structural studies of oxytocin and its analogs (Urry et al., 1970;Wood et al., 1986;Budesinsky et al., 2005) reveal high flexibility of the "tail" and low flexibility of the peptide ring stabilized by two H-bonds.The relatively rigid ring serves as a frame for the amino-acid side chains, whose spatial arrangement is believed to play a cru-cial role in the peptide-receptor interactions.The torsional angle χ1 describes rotation around the C α -C β bond in amino-acid residues and changes of the χ1 value have a significant influence on the conformation of a particular side chain.The main goal of this paper was to obtain the distributions of the dihedral angle χ1 values using various theoretical methods, including quantum mechanics and semiemipirical methods at the setup level, with molecular mechanics playing the leading role.The conformations of oxytocin analogs were studied both in vacuum and in solution.
A great number of structures of the examined peptides were needed, which could be obtained in two ways: through molecular dynamics (e.g. using AM-BER package (Pearlman et al., 1995)) or through the EDMC method implemented in ECEPP force field.
We regard the EDMC method as superior in searching the conformational space in comparison to the high-temperature molecular dynamics.We decided to use the EDMC method, which being stochastic, was more suitable for our purpose.We needed tools for the production of as many as possible different conformations forming the statistical population of the peptides.The nonstandard residues geometry was created using phenylalanine structure from the ECEPP/3 database as a template by simple substitution of a hydrogen atom by the methyl group in the phenyl ring.These modifications were performed using the program Molden (Schaftenaar & Noordik, 2000).The distribution of charges on the nonstandard residue atoms was taken from the Mopac calculations using MNDO/3 method (Bingham et al., 1975).Partial atomic charges were obtained by fitting point charges to the MNDO electrostatic potential.It is the standard procedure of developing new residues in the ECEPP/3 force field.Tables 1, 2 and 3 (in the z-matrix form) include partial atomic charges multiplied by factor 0 . 2 0 .322 (needed for appropriate conversion to energy units of kcal/mol (Nemethy et al., 1992)) at the modified residues introduced into ECEPP/3 database of amino-acid residues.
After adding the nonstandard residues to the ECEPP/3 database we searched for low energy conformations using the EDMC method.We carried out the same calculation in vacuum and in a water environment.The range of temperatures during Monte Carlo simulations was from 500 K to 10 000 K. Two thousand conformations of each oxytocin analog in vacuum and two thousand in water solution were found.These conformations were used as an input data for dihedral angle χ1 distribution analysis in ECEPP/3 force field.

ResulTs
Two thousand conformations of each oxytocin analog in two environments (vacuum or water) were taken into consideration during preparation of the angle distributions.In Fig. 1 the lowest energy members of each set of conformations are presented.Figure 2 shows the distribution of dihedral angle χ1 value in the modified residue [Phe 2 ] in the oxytocin analogs as a result of the ECEPP/3 calculations in water.
Figure 2a shows the distribution of the dihedral angle χ1 in the nonstandard amino-acid residue in [(Phe) 2 o-Me]OT in water.Two sharply outlined maxima on the graph are readily noticeable.They are situated within the range: (-180°; -165°) and (-75°; -45°) and represent 38% and 57.7% of the whole angle population, respectively.The remaining area is occupied by just over 2% of the conformations.A similar situation is observed in the case of the calculations performed in vacuum (Fig. 3a).There are no significant shifts of the maxima of the preferred values in comparison with the first graph, except for some broadening of the maximum for the χ1 angle around 180° and appearance of a small maximum in the interval (-150°; -135°).The maxima are located in the intervals: (165°; -165°), (-150°; -135°) and (-75°; -45°) and represent 47.3%, 8.9% and 42.3% of the whole angle population, respectively.Figures 2b and 3b show the distributions of the dihedral angle χ1 in the nonstandard amino-acid residue in [(Phe) 2 m-Me]OT in water and in vacuum, respectively.As one can see, there are two distinct maxima in plot 2b: one in the range (-75°; -45°) comprising 68.4% of the population and the second occupying edges of the plot with 30% of the population.The first one practically overlaps analogous maxima of the distribution for o-methylphenylalanine in water (see Fig. 2a).The results obtained in vacuum suggest that most (82.5%)dihedral angle χ1 values are contained in the interval (-75°; -45°).The remaining conformations form the second distinct maximum in the range (165°; -150°).Therefore the distribution is similar to the one of o-methylphenylalanine.The distributions of the dihedral angle χ1 in the nonstandard amino-acid residue in [(Phe) 2 p-Me]OT are shown in Fig. 2c (water) and Fig. 3c (vacuum).The distribution in water is similar to the analogous distributions for the ortho and para isomers.Two sharply outlined maxima are placed in the ranges: (165°; -165°) and (-75°; -45°), representing 39.7% and 60.4% of the population, respectively.The distribution for structures calculated in vacuum is similar to the previous cases as well.There are two main maxima in the intervals: (-75°; -45°) and (165°; -165°), representing 18.8% and 73.1% of the population, respectively, a third, not so high in the interval (45°; 75°) representing 4.3% of the population.The distributions obtained from the vacuum calculations differ in the level of occupation of the particular interval.In the case of the distributions obtained from the calculations in the water, the populations of the conformations were divided quite evenly between two intervals, whereas the situation for vacuum looks as follow: conformations with the meta isomer prefer interval around -60° (82% of all conformations), conformations with the para isomer grouped in this interval constitute only 18.8% of the whole population, and conformations with the ortho isomer do not prefer distinctly either of the two intervals.

B. Śmiech and others
The distributions of the torsional angle χ1 for residue 8 in water (Fig. 4) are very consistent and individual plots do not differ significantly from each other.The placement of the methyl group in the phenyl ring of residue 2 does not appear to influence the spatial arrangement associated with the changes of χ1 value of the residue 8 side chain.The maxima of the plots are placed within the same regions: the first (greater) within the interval (165°; -165°) and the second (smaller) within the interval (-90°; 45°).The global maximum contains 82.4% of all conformations for the ortho isomer, 88.9% for meta and 70.85% for the para one.All the remaining conformers are confined to the second maximum.The preferred intervals overlap those found for residue 2.
In vacuum another maximum appears for isomers meta and para in the range of (-105°; -60°).It is present along side the two intervals known from the former distributions (see Fig. 5) There occur two main maxima -at (165°, -150°) and around -90°on all three plots.The first interval contains 91.6%, 25.1% and 39.7% of the population for the ortho, meta and para isomers, respectively.The second interval contains 7.93%, 61.8% and 55.9% of all conformations of the corresponding analogs.These intervals overlap those from the other distributions.
Differences occur in the level of occupation of the intervals considered.The situation is very simi-lar for the plots of distributions for isomers meta and para (Fig. 5b and 5c).More than half of all conformations is located in the interval around -90°.In the case of isomer para only 7.9% of all conformations prefer this region.

dIsCussIOn
The torsional angle χ1 at residues 2 and 8 in free oxytocin analogs shows a strong tendency to adopt values from strictly defined, narrow intervals.The intervals preferred by the χ1 angle values at residue 2 in all three analogs seem to be independent of the environment (vacuum or water solution).This observation agrees with the fact that angle χ1 in peptides, generally, adopts one of three rotameric states: gauche(+), trans or gauche(-) (Marcus et al., 1996;Bhargavi et al., 2003).It is interesting to note that the conformation gauche(-) is practically absent on all plots.Nevertheless the level of the particular interval occupation seems to be dependent on both the environment and the MePhe isomer substituted in position 2 of oxytocin analog.This can be easily observed in the distributions from the simulations in vacuum environment (Fig. 3).Individual rotameric states are preferred with a different affinity by the individual analogs.The analog containing the ortho  Influence of modification at position 2 in oxytocin analogs isomer prefers both occupied rotameric states approximately equally.The analog with the meta isomer evidently favors conformation gauche(+), whereas the analog containing the para isomer clearly prefers the trans conformation.The character of the biological activity of the peptides considered in this paper is presented below (Slaninova et al., 2001): If one attempts to arrange the individual analogs according to their growing agonistic properties depending on magnesium presence, the following order results: We can easily notice that this sequence follows the growing tendency to adopt conformation trans by the angle χ1 in residue 2 in the individual analogs: Apparently, adoption of the trans conformation by angle χ1 in residue 2 is correlated with the agonistic properties of the peptide and analogously adoption of the gauche(+) conformation can be correlated with antagonistic properties.
Moreover, the placement of the methyl group in the phenyl ring of the second residue seems to influence the side chain conformation of residue 8, which is believed to be correlated with the antagonistic activity of the oxytocin analogs (Bankowski et al., 1980;Manning et al., 1984;1995;Hruby et al., 1994).This phenomenon is most clearly seen in the vacuum environment.If one arranges the individual plots according to the growing contributions of the trans conformers into the overall χ1 angle distribution, the following order will be obtained: [(Phe) 2 m-Me]OT (25.1%) ® [(Phe) 2 p-Me]OT (39.7%) ® [(Phe) 2 o-Me]OT (91.6%).
It is not identical with the sequence observed in the case of residue 2, but we can conclude that the antagonistic properties can be correlated with the adoption of the gauche(+) conformation by angle χ1 in residue 8.
Summarizing, the conformations gauche(-) seems to be very unfavorable in all considered analogs and residues.The agonistic properties of the peptides can be correlated with the adoption of conformation trans by the angle χ1 in residue 2, whereas agonistic properties can be correlated with the adoption of conformation gauche(+) by the angle χ1 in residue 8.

Figure 1 .
Figure 1.stereoview of the lowest energy conformations of the considered analogs of oxytocin in vacuum and in water.a) and b) [(Phe) 2 o-Me]OT in water and in vacuum, respectively; c) and d) [(Phe) 2 m-Me]OT in water and in vacuum, respectively; e) and f) [(Phe) 2 p-Me]OT in water and in vacuum, respectively.All calculations were performed using EDMC method implemented in the ECEPP/3 force field.The lowest energy conformations were selected from two thousand conformations each of peptide in both environments.

Figure 2 .
Figure 2. distributions of the dihedral angle χ1 in methylphenylalanine in water.a) [(Phe) 2 o-Me]OT; b) [(Phe) 2 m-Me]OT; c) [(Phe) 2 p-Me]OT.The plots are drawn on the basis of the EDMC calculations in the ECEPP/3 force field using continuum solvent model.

Figure 3 .
Figure 3.The distributions of the dihedral angle χ1 in methylphenylalanine in vacuum.a) [(Phe) 2 o-Me]OT; b) [(Phe) 2 m-Me]OT; c) [(Phe) 2 p-Me]OT.The plots are drawn on the basis of the EDMC calculations in the ECEPP/3 force field in vacuum environment.

Figure 4 .
Figure 4.The distributions of the dihedral angle χ1 in amino-acid residue at position 8 in vacuum.a) for [(Phe) 2 o-Me]OT; b) for [(Phe) 2 m-Me]OT; c) and for [(Phe) 2 p-Me]OT.The plots are drawn on the basis of the EDMC calculations in the ECEPP/3 force field using continuum solvent model.

Figure 5 .
Figure 5.The distributions of the dihedral angle χ1 in amino-acid residue at position 8 in solvent.a) for [(Phe) 2 o-Me]OT; b) for [(Phe) 2 m-Me]OT; c) and for [(Phe) 2 p-Me]OT.The plots are drawn on the basis of the EDMC calculations in the ECEPP/3 force field in vacuum environment.