No abstract

The bond dissociation enthalpies (BDE) of all of the amino acid residues, modeled by HC(O)NHCH(R)C(O)NH(2) (PH(res)), were determined at the B3LYP/6-31G//B3LYP/6-31G level, coupled with isodesmic reactions. The results for neutral side chains with phi, psi angles approximately 180 degrees, approximately 180 degrees in ascending order, to an expected accuracy of +/-10 kJ mol(-)(1), are Asn 326; cystine 330; Asp 332; Gln 334; Trp 337; Arg 340; Lys 340; Met 343; His 344; Phe 344; Tyr 344; Leu 344; Ala 345; Cys 346; Ser 349; Gly 350; Ile 351; Val 352; Glu 354; Thr 357; Pro-cis 358; Pro-trans 369. BDEs calculated at the ROMP2/6-31G//B3LYP/6-31G level exhibit the same trends but are approximately 7 kJ mol(-)(1) higher. All BDEs are smaller than those of typical secondary or tertiary C-H bonds due to the phenomenon of captodative stabilization. The stabilization is reduced by changes in the phi,psi angles. As a result the BDEs increase by about 10 kJ mol(-)(1) in beta-sheet and 40 kJ mol(-)(1) in alpha-helical environments, respectively. In effect the alpha C-H BDEs can be "tuned" from about 345 to 400 kJ mol(-)(1) by adjusting the local environment. Some very significant effects of this are seen in the current literature on H-transfer processes in enzyme mechanisms and in oxidative damage to proteins. These observations are discussed in terms of the findings of the present study.

A previous experimental study of the αC−H bond dissociation energies (BDEs) of amines indicated a significant decrease in BDE (or increase in radical stabilization energy, E s) in the series primary, secondary, and tertiary. However, this was not supported by theoretical investigations. The αC−H BDEs of trimethylamine ((CH3)3NH), triethylamine ((C2H5)3NH), and tri-n-butylamine ((C4H9)3NH) and of the cyclic secondary amines piperidine, piperazine, morpholine, and pyrrolidine were therefore determined by photoacoustic calorimetry in benzene solvent. Ab initio procedures, which incorporated isodesmic reactions to minimize residual correlation errors, were used to obtain the BDEs of several of these for direct comparisons. Also the BDEs of methylamine (CH3NH2), ethylamine (C2H5NH2), isopropylamine ((CH3)2CHNH2)), and dimethylamine ((CH3)2NH) were calculated as a check on the earlier results. The experimental BDEs in kJ mol-1 at 298 K (±10 kJ mol-1), estimated from the photoacoustic calorimetric measurements, were as follows:  trimethylamine 372, triethylamine 381, tri-n-butylamine 381, piperidine 385, piperazine 385, morpholine 389, and pyrrolidine 377. The ab initio results were in excellent agreement with these values. From earlier work and the present calculations the α-to-N C−H BDE of methylamine was estimated to be 388 ± <10 kJ mol-1, corresponding to a radical stabilization energy, E s, of ∼51 kJ mol-1. Contrary to the previous experimental finding, both theory and experiment showed that the increase in Es on alkylation either at N or C is expected to be less than 4 kJ mol-1. Values of for the α-C radicals of the smaller aliphatic amines, except that of methylamine, must therefore be revised. The three-electron two-orbital π-like interaction, which causes the αC radical stabilization, is maximized when the singly occupied sp n orbital of C and the nonbonded doubly occupied sp n orbital of N are anticoplanar to each other. Alkylamines preferably adopt a conformation in which at least one αC−H bond is anticoplanar to the lone pair on nitrogen, and the most stable carbon centered α-to-N free radical is that derived by abstraction of this H atom. In the five-membered pyrrolidine ring the radical adopts an envelope conformation with the C5 carbon atom at the vertex. This accommodates the favorable alignment of the sp n orbitals of C• and N but has no C−H eclipsing interactions like those which occur in the parent. Thus, in effect, there is a reduction of strain on formation of the radical, and the BDE is lowered by ∼8 kJ mol-1 below that of typical secondary amines.

Ab initio computations (B3LYP/6-31G(D), coupled with isodesmic reactions) were used to predict bond dissociation energies (BDEs) of αC−H (D α CH) and other bonds of cysteine, both as free neutral amino acid (AH(Cys)) and as a residue in a model peptide (PH(Cys)). The latter was intended to mimic the environment in proteins. Transition structures were located for intermolecular and intramolecular H atom transfer to a thiyl radical (RS•) from a sulfhydryl group (RSH) or the αC−H bond. The predicted BDEs, at 298 K, in kJ mol-1 to an estimated accuracy of 10 kJ mol-1 for the fully optimized system are (AH(Cys)) D α CH = 322, D β CH = 390, D α CC = 264, and D SH = 373 and (PH(Cys)) D α CH = 346, D β CH = 392, D α CC = 287, and D SH = 367. In PH(Cys) with torsional angles constrained to simulate β-sheet and α-helical secondary structure, D α CH rises to 359 and 376, respectively. Cystine in the peptide environment was modeled by replacing −SH by −SSCH3, PH(CysSCH3), D α CH = 330. Enthalpies of activation for intermolecular H transfer to RS• were found to be low:  from RSH, 12 kJ mol-1; from αC−H, about 25 kJ mol-1, the latter being consistent with reaction rates on the order of 105 M-1 s-1. The enthalpic barrier for intramolecular H transfer from αC−H to −S• within a single cysteine residue is too high (83−111 kJ mol-1) for this to be a competitive process.