15267
2008, 112, 15267–15268 Published on Web 11/01/2008
Peptide CrDr Stretch Frequencies in a Hydrated Conformation Are Perturbed Mainly by Cr-Dr · · · O Hydrogen Bonding Noemi G. Mirkin and Samuel Krimm* LSA Biophysics, UniVersity of Michigan, 930 North UniVersity AVenue, Ann Arbor, Michigan 48109-1055 ReceiVed: September 25, 2008; ReVised Manuscript ReceiVed: October 21, 2008
We have shown (J. Phys. Chem. A 2004, 108, 10923; 2007, 111, 5300) that the CRDR stretch frequency, ν(CD), can discriminate between uniform RR, β, and polyproline II conformations of isolated peptides. Similar results for such peptides to which explicit waters are hydrogen bonded exhibit shifts in ν(CD) from those of the isolated structures. We demonstrate that the main source of these frequency shifts is the formation of CR-DR · · · O hydrogen bonds to water. Taking into account C-H · · · O(water) hydrogen bonding, together with the traditional bonding of peptide groups to water, can be expected to increase our understanding of the interaction of proteins with their aqueous environment. In a previous communication,1 we demonstrated that the frequency of the peptide CRDR stretch mode, ν(CD), is capable of discriminating between the RR, β, and polyproline II conformations of the isolated CR-deuterated alanine dipeptide (DR-ADP), CH3[CONH]1CRDR(CH3)[CONH]2CH3. In a subsequent paper,2 we extended this ab initio analysis to longer uniform alanine conformations through the octapeptide, and showed that at least three spectral properties enabled such structural discrimination. The initial publication1 also gave the shifted ν(CD) for the DR-ADP(H2O)4 species, and the latter paper2noted that hydrogen bonding also influences the ν(CD) of longer peptides. These results have now been analyzed,3 and they indicate that the main perturbing contribution to the values of the isolated-molecule ν(CD) for any given conformation is the pattern of water structure in the vicinity of DR that leads to the formation of specific CR-DR · · · O hydrogen bonding, with the associated decrease in the CR-DR bond length, r(CD), and the resulting increase in ν(CD). In order to illustrate this effect, we give the results for the β conformer (φ ) -134°, ψ ) 145°) of DR-ADP(H2O). As before,1,2 our calculations were done at the B3LYP/6-31+G* level, with unscaled ν(CD) being obtained for the otherwise fully optimized structures. (Test MP2 and PCM calculations3 give qualitatively the same results, thus supporting our use of the B3LYP protocol.) In Figure 1a, we show the structure in which the single water has optimized to the position cis to the terminal methyl group, designated CO1a. In the absence of the water, the isolated molecule has values of r(CD) ) 1.0948 Å (see Table 1) and ν(CD) ) 2276 cm-1. With this hydrogenbonded water, the values are r(CD) ) 1.0944 Å and ν(CD) ) 2281 cm-1. With a second water bonded to NH1, the values are r(CD) ) 1.0939 Å and ν(CD) ) 2284 cm-1, a small effect on ν(CD) of the adjacent fully hydrogen-bonded peptide group. In Figure 1b, we show the structure in which the water has optimized to the position cis to NH1, CO1b (it was actually * To whom correspondence should be addressed. Phone: 734-763-8081. Fax: 734-764-3323. E-mail:
[email protected].
10.1021/jp808515t CCC: $40.75
Figure 1. Structure of DR-ADP(H2O) with water in two different optimized positions: (a) water cis to CH3 across CO1; (b) water cis to NH1 across CO1.
initially placed in a hydrogen-bonding position at NH2, but it shifted toward CO1 during the optimization). The new values are r(CD) ) 1.0876 Å and ν(CD) ) 2348 cm-1, with a DR · · · O distance of 2.223 Å, a value at the lower end of the range found for multiply hydrated structures.3 The decrease in r(CD) and the large increase in ν(CD) must result from a specific interaction of the water with the CRDR bond, since the H · · · O hydrogen-bond interaction with CO1 hardly changes: r(CO1) ) 1.2398 Å in the structure of Figure 1a and 1.2406 Å in the structure of Figure 1b (the value in the isolated molecule is 1.2310 Å). Also, the r(CD) and ν(CD) values hardly change by adding a surrounding water continuum in the calculation.3 It might be thought that the bond contraction (and thus the frequency increase) is due mainly to a repulsive interaction resulting from the same water also hydrogen bonding to NH2. 2008 American Chemical Society
15268 J. Phys. Chem. B, Vol. 112, No. 48, 2008
Letters
TABLE 1: Some Geometrical Parametersa and Frequenciesa of βb-Alanine Dipeptidec Structures DR-ADP (H2O) DR-ADP r(CO1) r(O1 · · · Hh) r(Ow-H) r(Ow-Hh) r(NH1) r(CD) r(DR · · · Ow) r(CO2) r(NH2) θ(CN1H1) θ(CN2H2)
CO1ad
CO1be
r, θ
ν
r, θ
ν
r, θ
ν
1.2310
1734
1703
3593 2276
1.2319 1.0107 122.6 118.5
1749 3627
1.2316 1.0108 120.2 118.8
1744 3626
1.2406 1.8848 0.9681 0.9834 1.0139 1.0876 2.2230 1.2323 1.0143 120.3 117.0
1693
1.0138 1.0948
1.2398 1.8497 0.9678 0.9825 1.0143 1.0944
3587 2281
3592 2348 1746 3582
a Obtained at the B3LYP/6-31+G* level. Bond lengths, r, in Å, bond angles, θ, in degrees, and frequencies, ν, in cm-1. b φ ) -134°, Ψ ) 145°. c ADP: CH3[CONH]1CRDR(CH3)[CONH]2CH3. d Water cis to CH3 across CO1. e Water cis to NH1 across CO1.
However, the evidence does not support this. (1) The (N)H2 · · · O distance is 2.79 Å, much longer than the typical ∼1.9 Å distances found for the obvious NsH · · · O hydrogen bonds in the more fully hydrated peptide structures.3 (2) The value of r(NH2) ) 1.0143 Å (ν ) 3582 cm-1) is much shorter than values of up to r(NH2) ) 1.0277 Å (ν ) 3366 cm-1) found for the typical NsH · · · O bonds in hydrated structures3 (the nonbonded value is 1.0108 Å with ν ) 3626 cm-1). (3) In a β-DR-ADP(H2O)4 structure, we find two DR · · · O distances of 2.689 and 2.773 Å, yet r(CD) ) 1.0891 Å, hardly different from the CO1b water value even though the O atoms are more than 0.5 Å farther away. At most, these results are indicative of a weak (N)H2 · · · O interaction, likely a compromise with the formation of the combined CdO · · · H and CRsDR · · · O interactions. The ν(CD) shift is clearly due to the formation of a CRsDR · · · O hydrogen bond, in addition to the expected CdO · · · H bond, and is consistent with the results of earlier studies of such bonds in non-zwitterionic amino acids.4 Although they had not been thought to occur in peptides, the presence of C-H · · · O hydrogen bonds was indicated from early infrared studies of polyglycine II.5–8 Such bonds are now a wellestablished attractive interaction,9 being recognized in amino acids4 and indeed found in proteins.10–12 The uncommon bond contraction and stretch frequency increase are now also understood to result from the dominant contracting force generated by the antiparallel orientation of the C-H dipole derivative and the acceptor electric field13–15 (in distinction to the parallel orientation, with bond extension and frequency decrease, in traditional hydrogen bonds such as N-H · · · O), and more generally to depend on properties of the charge density derivatives of the donor.16,17 An estimate of the energy of the C-H · · · O hydrogen bond in the β-DR-ADP(H2O) system can be obtained from the difference in energies of the structures in parts a and b of Figure
1. The energy of the CO1b structure is lower by 2.31 kcal/mol than that of the CO1a structure (after taking into account, through a CO1b-structure-constrained calculation of CO1a, the small differences in the structures of the two peptides, Table 1). This is a maximum value, considering the possible (probably small) stabilization due to the [N]H2 · · · O interaction, but it agrees well with the value obtained for the amino acid, viz., 2.10 kcal/mol.4 Being about half the energy of typical peptide-group hydrogen bonds, C-H · · · O hydrogen bonds should not be considered to be of negligible importance. They demonstrably determine the resulting r(CD) and thus ν(CD) (as well as having an impact on the charge distribution in the peptide group3). Therefore, they clearly can have an important influence on the conformations of peptides in aqueous solution. The determination of the correct distribution of r(CD), and from this the ab initio-derived ν(CD),2 will ultimately be best achieved through molecular dynamics simulations with the most physically accurate force fields,3 but the structural discrimination of the method is expected to be maintained.3 It should be noted that such computations with constrained X-H bond lengths18 cannot be expected to capture the complete impact of the peptide-water interactions. Perhaps one of the more important consequences of recognizing the presence of C-H · · · O hydrogen bonds will be in the added insight it will provide to a comprehensive understanding of the interactions of proteins with their aqueous environment, which has heretofore been analyzed only in terms of hydrogen bonding of the peptide groups to water.19 Also, such additional bonding could help in interpreting the experimentally observed heterogeneous water dynamics near protein surfaces.20 Acknowledgment. This research was supported by NSF grant CHE-0517905. References and Notes (1) Mirkin, N. G.; Krimm, S. J. Phys. Chem. A 2004, 108, 10923– 10924. (2) Mirkin, N. G.; Krimm, S. J. Phys. Chem. A 2007, 111, 5300–5303. (3) Mirkin, N. G.; Krimm, S. To be published. (4) Scheiner, S.; Kar, T.; Cu, Y. J. Biol. Chem. 2001, 276, 9832–9837. (5) Krimm, S. Nature 1966, 212, 1482–1483. (6) Krimm, S.; Kuroiwa, K.; Rebane, T. Conformation of Biopolymers; Academic Press: London, 1967; p 439. (7) Krimm, S.; Kuroiwa, K. Biopolymers 1968, 6, 401–407. (8) Dwivedi, A. M.; Krimm, S. Biopolymers 1982, 21, 2377–2397. (9) Barnes, A. J. J. Mol. Struct. 2004, 704, 3–9. (10) Derewenda, Z.; Lee, L.; Derewenda, U. J. Mol. Biol. 1995, 252, 248-262. (11) Senes, A.; Ubarretxena-Belandia, I.; Engelman, D. M. Proc. Natl. Acad. U.S.A. 2001, 98, 9056–9061. (12) Manikandan, K.; Ramakumar, S. Proteins 2004, 56, 768-781. (13) Hermansson, K. J. Phys. Chem. A 2002, 106, 4695–4702. (14) Qian, W.; Krimm, S. J. Phys. Chem. A 2002, 106, 6628–6636. (15) Qian, W.; Krimm, S. J. Phys.Chem. A 2002, 106, 11663–11671. (16) Qian, W.; Krimm, S. J. Phys. Chem. A 2005, 109, 5608–5618. (17) Qian, W.; Krimm, S. THEOCHEM 2006, 766, 93–104. (18) Feig, M. J. Chem. Theory Comput. 2008, 4, 1555–1564. (19) Klotz, I. M. Protein Sci. 1993, 2, 1992–1999. (20) Malardier-Jugroot, C.; Johnson, M. E.; Murarka, R. K.; HeadGordon, T. Phys. Chem. Chem. Phys. 2008, 10, 4903-4908.
JP808515T
