AN AB INITIO STUDY OF INTERMOLECULAR INTERACTIONS OF GLYCINE, ALANINE AND VALINE DIPEPTIDE-FORMALDEHYDE DIMERS

Transcription

1 Journal of Undergraduate Chemistry Research, 2004, 1, 15 AN AB INITIO STUDY OF INTERMOLECULAR INTERACTIONS OF GLYCINE, ALANINE AND VALINE DIPEPTIDE-FORMALDEHYDE DIMERS J.R. Foley* and R.D. Parra Chemistry Department, DePaul University, Chicago, , Abstract Ab Initio calculations were performed on glycine, alanine and valine dipeptide-formaldehyde dimer systems. It was found that as the R-group of the amino acid dipeptide increases in size, the interaction energy remains fairly constant due to the stabilization effects of the hydrogen bonding in each dimer system. Such information is valuable to explain the prominence and structural properties of glycine, alanine and valine in globular proteins. Keywords: Ab Initio, Alanine, Conformation, Dipeptides, Glycine, Hydrogen bonding, Valine Introduction The hydrogen bond is one of the most important intra- and intermolecular interactions; many biological processes are highly dependent upon it. Without hydrogen bonds, our bodies could not function (1). Three types of hydrogen bonding will be considered in this study: conventional, non-conventional and three-centered (bifurcated). When a hydrogen bond results from the hydrogen atom approaching an electronegative atom, this is a conventional interaction. Later it was discovered that carbon, not a particularly electronegative atom, could form a non-conventional hydrogen bond. Lastly, a configuration showing evidence of an attractive force between the covalently bonded hydrogen atom of a donor X-H group and two acceptor atoms is termed a bifurcated interaction (1). There are two types of three-centered hydrogen bond interactions: one hydrogen atom interacting with two acceptor atoms (Y-HX), and one acceptor atom interacting with two hydrogen atoms (HXH) (Figure 1) (2). Figure 1. Bifurcated Hydrogen Bonding Due to their biological relevance, alanine, glycine and valine dipeptides were used in this study to assess different types of hydrogen bonding. Dipeptide amino acids are slightly modified, having two terminal methyl groups instead of the traditional amine and carboxylic acid terminal groups. Modified amino acid residues are useful building blocks in molecular engineering. They can be designed to control the fold of the peptide backbone and enhance resistance to biodegradation, while leaving their ability to bind and respond as the native peptide would (3). The key factor is that the side chains are not modified, and this is what is usually required for molecular recognition (3). Glycine and alanine dipeptides have shown conformational variations, which are similar to that of assembled proteins, and for that reason have been studied by various theoretical and experimental methods (4). Also, due to their conformational flexibility, they may be considered reasonable models of larger globular proteins (4). Despite the similarities, application of data is somewhat limited for glycine studies because of the absence of secondary structural features (4). The R-group side chain of just a hydrogen is simplistic and decreases the chances for significant interactions with other molecules. Additionally, valine is a particularly important residue to study because it has a higher percentage of occurrence in proteins than most other residues (3,5). Moreover, valine has a high tendency to form

2 Journal of Undergraduate Chemistry Research, 2004, 1, 16 β-sheets, so it is important in the study of the secondary structure of proteins (5,6). Figure 2. Numbering Schemes For Dipeptide-Formaldehyde Dimer Systems Forming dimer systems with a formaldehyde molecule allows the simple simulation of hydrogen bonding interactions. The use of glycine, alanine and valine dipeptides for this study will allow the extrapolation of some of the conclusions to β-sheet secondary structures and other aspects of proteins. These three amino acid dipeptides are among the most important and widely observed in protein structure and hold valuable information on protein interactions, particularly structural conformations. Computational Method All the computations were carried out using the GAUSSIAN 98 program (7). Glycine, alanine, valine and formaldehyde monomers and dimer systems were fully optimized using MP2/4-31G level of theory. Interaction energies were calculated by subtracting the individually calculated energies of the monomers from the single point energy of the dimer system as shown: 2a) Glycine Dipeptide-Formaldehyde Dimer E = E dimer - (E dipeptide + E formaldehyde ) Energy values were reported at both HF and MP2/ 4-31G levels for completeness and comparison of correlation effects. In addition, basis set superposition error (BSSE) corrections on interaction energy calculations were applied (8,9). Frequency values were calculated and used as an assurance that each conformation was a minimum on the potential energy surface. Gauss View was used as a graphic user interface to visualize the results. 2b) Alanine Dipeptide-Formaldehyde Dimer Results and Discussion Two conformational minima were studied for the glycine dipeptide: planar and non-planar. An interesting non-planar conformation was found to be a minimum at the MP2/3-21G* level of theory during another work by our group that warranted further investigation. In the planar glycine-formaldehyde dimer, there are two prominent hydrogen bonding interactions: one conventional, the other non-conventional. The conventional hydrogen bond involving N 6 has a calculated distance of Å and an angle of in the optimized structure (Figure 2a and 2c) Valine Dipeptide-Formaldehyde Dimer

3 Journal of Undergraduate Chemistry Research, 2004, 1, 17 Table 1. Hydrogen Bond Lengths and Angles for Glycine, Alanine and Valine DipeptideFormaldehyde Dimer Systems Dimer H-Bond Number Distance (H...A)(Å) Angle (X-H...A)( o ) Planar Glycine- N Formaldehyde C Non-planar N Glycine- C Formaldehyde C 4 -H...O=C C 4 -H...O Alanine- N Formaldehyde C C Valine- N 9 -H...O 11 =C Formaldehyde C 6 -H...O=C C 7 -H...N 9 -H C 7 -H...O 11 =C C C 4 -H...O 11 =C Table 1). Both of these calculated values are within the accepted hydrogen bond distance and angle ranges listed for conventional hydrogen bonds (Table 2). The non-conventional hydrogen bond involving C 9...O=C 2 has a calculated distance of Å, with an angle of , both within accepted ranges for weak hydrogen bond geometries (Tables 1 and 2). As mentioned before, a glycine-formaldehyde dimer was optimized at the MP2/321G* level of theory in an earlier project carried out by our group, which gave a nonplanar glycine conformation. This dimer was optimized at the MP2/4-31G level and it remained non-planar. Since it stayed non-planar, we decided to investigate this conformation of glycine further as a monomer. However, when the non-planar glycine monomer was optimized, it returned to a planar conformation with no imaginary frequencies, indicating that it was a minimum. Considering the fact that the glycine monomer for this particular study existed in a non-planar conformation only in a dimer with hydrogen bonding interactions, it seems that the presence of formaldehyde has a stabilizing effect, allowing the monomer to remain in a non-planar conformation. This conformation had a Φ rotational angle of and a Ψ rotational angle of (Table 3). All hydrogen bond distances and angles of the non-planar glycine dimer structure are within accepted ranges. A conventional hydrogen bond between N 6 has a calculated distance of Å and an N-H...O angle of (Figure 2a and Table 1). There is also a non-conventional hydrogen bond taking place between C 9, which had a calculated distance of Å and a C-H...O angle of , both within the non-conventional hydrogen bond distance and angle ranges (Tables 1 and 2). Interestingly, there also seems to be an additional interaction between one of the hydrogens on the C α and a carbonyl oxygen on the peptide backbone as well as the carbonyl oxygen of formaldehyde. A Table 2. Properties of Hydrogen Bonds a Property Conventional Non-conventional (Strong) Hydrogen Bond (Weak) Hydrogen Bond d(h...a) range (Å) (X-H...A) range ( o ) a Adapted from Desiraju and Steiner (10)

4 Journal of Undergraduate Chemistry Research, 2004, 1, 18 Table 3. Calculated Peptide Angles of Rotation for Glycine, Alanine and Valine Dipeptide Monomers Dipeptide ( o ) ( o ) Glycine-Planar (monomer) Glycine-Planar (in dimer) Glycine-Non-Planar Alanine Valine three-centered hydrogen bonding interaction is observed in this conformation. The C α hydrogen that is in a position to interact with the formaldehyde and an oxygen of a carbonyl group at C 2 of the backbone is found to have calculated hydrogen bond distances and angles consistent with those acceptable to identify the presence of a non-conventional hydrogen bonding interaction (Tables 1 and 2 and Figure 2a). In the interaction with formaldehyde, the calculated C 4 -H a...o 8 hydrogen bond distance was Å, with an angle of (Table 1). For the C-H...O interaction with the C 2 =0 of the peptide backbone, the calculated C 4 -H α...o=c 2 hydrogen bond distance was Å with an angle of (Table 1). The presence of this three-centered interaction is allowed by the bent conformation of glycine dipeptide and makes this conformation more stable than its planar counterpart (Table 4). Only a bent conformation (Φ = and Ψ = ) was studied for alanine since the presence of the methyl R-group makes it very difficult sterically to adopt a planar conformation (Table 3). In this alanine dipeptide-formaldehyde dimer, a three-centered hydrogen bonding interaction is taking place with the alanine dipeptide and the oxygen of the carbonyl of formaldehyde (Figure 2b). The three-centered hydrogen bond has a conventional and non-conventional component. The hydrogen bond N 6 has a length of Å and angle of , both of which are in the accepted ranges for strong hydrogen bond distances and angles, respectively (Tables 1 and 2). The other component of the three-centered interaction C 4 has a calculated distance of Å and an angle of , both of which were well within the parameters for weak hydrogen bonds (Tables 1 and 2). Another hydrogen bonding interaction is occurring between the alanine dipeptide and one of the hydrogens on formaldehyde (Figure 2b). In this case, the C 9 =H...O=C 2 interaction has a calculated distance of Å, with and angle of , which are in the accepted ranges for weak hydrogen bonds (Tables 1 and 2). Both of these interactions are contributors to the overall stability of the alanine dipeptide-formaldehyde dimer. In comparison to the BSSE corrected interaction energy calculated for the planar glycine dipeptide, that for the alanine dipeptide is almost 1 kcal mol -1 lower. However, the BSSE corrected interaction energies for the non-planar glycine dipeptideformaldehyde dimer and the alanine dipeptideformaldehyde dimer were comparable (Table 4). This indicates that the methyl group on the Ca has almost no energetic effect. Since the planar glycine dipeptide-formaldehyde dimer was higher in energy that the nonplanar alanine dipeptide-formaldehyde dimer, this may suggest there is some stability gained by adopting a bent conformation for these dipeptides since the methyl group doesn t seem to significantly affect the overall interaction energy of the molecule. As with alanine, only a bent conformation (Φ = and Ψ = ) of valine dipeptide was studied due to the steric hindrance induced by the large isopropyl R-group (Table 3). In one study involving retro-inverso valine dipeptides it was found that conformational preferences for this particular Table 4. Interaction Energies for X-Formaldehyde Dimers X-Formaidehyde Interaction Energy (kcal mol -1 ) Dimer HF BSSE MP2 BSSE Corrected Corrected Glycine-Planar Glycine-Non-planar Alanine Valine

5 Journal of Undergraduate Chemistry Research, 2004, 1, 19 modified valine dipeptide were determined by stabilizing intramolecular hydrogen bonds (3). Interestingly, this is also observed in this study. There is an intramolecular hydrogen bond from one of the methyl groups of the side chain to N 9 on the peptide backbone that is within the parameters for a conventional hydrogen bonding interaction (Figure 2c and Tables 1 and 2). The C 7 H...N 9 -H interaction has a calculated distance of Å and an angle of , in compliance with weak hydrogen bond ranges (Tables 1 and 2). Another intramolecular hydrogen bond between the other methyl group of the side chain and an oxygen from a carbonyl of the backbone is also observed. The C 6 -H...O=C 8 interaction has a calculated distance of Å and an angle of , which are within the accepted ranges for weak hydrogen bonding interactions (Tables 1 and 2). In another interesting interaction, hydrogens from C 4, N 9, and C 7 of the isopropyl R-group are all interacting with the oxygen of formaldehyde. Three hydrogen bonds are formed to one oxygen atom. The N 9 -H...O 11 =C 12 interaction has a calculated length of Å and an angle of , making it a representative strong hydrogen bond (Tables 1 and 2). The other two hydrogen bonds to the oxygen are weak C-H...O interactions. One involves one of the carbons of a methyl group from the side chain, and the other involves one of the carbons from the backbone. The C 7 -H...O 11 =C 12 hydrogen bond has a calculated distance of Å, and an angle of , which are both within the accepted ranges for the identification of weak hydrogen bonds (Tables 1 and 2). Similarly, the C 4 -H...O 11 =C 12 interaction is also in compliance with known weak hydrogen bond ranges with a distance of Å and an angle of (Tables 1 and 2). The valine dipeptide-formaldehyde dimer also has a hydrogen bond from the carbon of formaldehyde to a carbonyl oxygen of the peptide backbone. The C 12 interaction has a calculated distance of Å and an angle of (Table 1). This interaction is within known ranges for weak hydrogen bonds, but is almost in the parameters for strong hydrogen bonds. The distance of 2.3 Å is very close to the 2.2 Å upper limit for strong hydrogen bond distances (Table 2). This dipeptideformaldehyde dimer system has the most hydrogen bonding interactions occurring out of any other studied. The MP2 BSSE corrected interaction energy for the valine dipeptide-formaldehyde dimer was calculated to be kcal mol -1 and that for alanine dipeptide-formaldehyde dimer was kcal mol -1, while that for the comparable non-planar glycine dipeptideformaldehyde dimer was kcal mol -1 (Table 4). It appears that overall the nonplanar glycine dipeptide-formaldehyde dimer was the most stable. Conclusions The data involving the different dipeptides with increasingly bulky R-groups showed that there is a slight destabilizing effect due to the R-group on its interaction energy with another molecule. However, since the BSSE corrected interaction energies of the valine dipeptide-formaldehyde system and the alanine dipeptide-formaldehyde system only differed by about 0.28 kcal mol -1, this destabilization may have been overcome by the increased number of hydrogen bonding interactions in the valine system. Interestingly, the valine dipeptide-formaldehyde dimer only differed from the lowest-energy non-planar glycine-formaldehyde dimer by 0.41 kcal mol -1. The presence of an isopropyl group is not much energetically different than having a single hydrogen. This seems to suggest that the energetic hindrance of having a larger side chain is offset by the increased opportunity of formation of hydrogen bonds and the stabilization gained by these interactions. This is evidenced by the fact that the valine dipeptide-formaldehyde dimer is the system that has the most hydrogen bonding interactions out of all the dimer systems studied, which enables it to have the bulkiest R-group while maintaining comparable energy. Additionally, there seems to be an increased ability to form three-centered interactions when the dipeptides adopt non-planar conformations. The planar glycine dipeptideformaldehyde dimer is the only dimer system studied that did not exhibit a three-centered interaction. Furthermore, the larger R-groups allow for more participation in all types of hydrogen bonding discussed. This is evidenced by the increase in observed hydrogen bonding in the dipeptide-formaldehyde dimers as the R-group is

6 Journal of Undergraduate Chemistry Research, 2004, 1, 20 increased in size. The results from this investigation will be useful in further studies assessing the structural properties and intermolecular interactions of glycine, alanine and valine in globular proteins. Acknowledgement This research was funded by an Alexandroff summer research grant for the summer of 2002 from the DePaul University Chemistry Department and an Undergraduate Research Assistantship for winter and spring quarters 2003 from the College of Liberal Arts and Sciences. We send our special thanks to DePaul University for their support. References (1). G.A. Jeffrey and W. Saenger. Hydrogen Bonding in Biological Structures, Springer-Verlag, Berlin, Ger many, (1991) pp. 7, 36. (2). J. Yang and S.H. Gellman. J. Am. Chem. Soc., 1998, 120, (3). C. Aleman. J. Phys. Chem. B, 2001, 105, (4). T. Head-Gordon, M. Head-Gordon, M.J. Frisch, C.L. Brooks III and J.A. Pople. J. Am. Chem. Soc., 1991, 113, (5). W. Viviani, J. Rivail, A. Perczel and I.G. Csizmadia. J. Am. Chem. Soc., 1993, 115, (6). B.S. Kinnear, D.T. Kaleta, M. Kohtani, R.R. Hudgins and M.F. Jarrold. J. Am. Chem. Soc., 2000, 122, (7). M.J. Frisch, G.W. Trucks, H.B. Schlegel et al., Gaussian 98, Revision A.9, Gaussian, Inc., Pittsburgh, PA, (8). S. Scheiner. Hydrogen Bonding: A Theoretical Per spective, Oxford University Press, New York, USA, (1997), pp (9). S. Simon, M. Duran and J.J. Dannenberg. J. Phys. Chem. A, 1999, 103, (10). G.R. Desiraju and Thomas Steiner. The Weak Hydrogen Bond in Structural Chemistry and Biology, Oxford UP, New York, USA, (1999) Chapter 1.