In our recent study we have demonstrated an important role of the carbonyl H-bonds for a -proton abstraction. Their major contribution to the catalysis is lowering the pKa of the a -proton². The effect of H-bond donors on the relative stability of the enolate ion resulting from a -proton abstraction was predicted from their respective proton affinities (Table 1) estimated from B3LYP/6-31+G(d,p) total energies. Naked thioenolate anion, [CH₃CH(C=O)SCH₃]^-, has a much higher PA than acetate anion (367.9 vs 352.9 kcal/mol), which means, that in the absence of H-donors, acetate anion is too weak a base to abstract an a -proton. For this reason, neither transition structure nor product cluster have been located². Two H-donors to the model substrate reduce the a -proton affinity to the value 345.3 kcal/mol which is 7.6 kcal/mol lower than PA of the base acetate anion. The effect of one H-donor, methylformamide, is not so pronounced; the PA of acetate anion is slightly lower (by 1.4 kcal/mol) than the PA of CH₃CH(C=O)SCH₃• HNCH₃(C=O)H. This H-bond mediated reduction of the energy needed for a -proton abstraction from the thioester leads to a significant decrease (from 10.4 to 5.8 kcal/mol) in the reaction barrier (Scheme 2).

In order to get an initial idea on how the flavin proximity found in the X-ray active site structure may affect the a -proton abstraction, we calculated proton affinities for a simplified sandwich-like system consisting of thioenolate,CH₃CH(C=O)SCH₃, and a bicyclic oxidized flavin model (Figure 2). The distance (d, Å) was varied between 3 and 5.0 without post-optimization. The starting complexes (neutral and anionic) were pre-optimized at the B3LYP/6-31+G(d,p) with the two C-C distances (see Figure 2) fixed to 3 Å and two dihedral angles fixed to +/-90° in order to keep the planes of the flavin [defined by {N5,C4a,C10a}] and substrate [defined by {O=C1Ca }] parallel.

The results of these estimations are presented in Figure 3. They suggest that the proximity of the flavin to the substrate may have a strong acidifying effect upon the a -hydrogen of the thioester, so that bases even weaker than acetate anion can be effective.

Using the fragment of the enzyme active site X-ray structure we selected a more extended base (CHO)NHCH₂CH₂CH₂COO to mimic GLU376. Its PA was estimated to be 345.1 kcal/mol (Table 1). Thus, comparing this value with the proton affinities given in Table 2, one may expect a significant reduction of the reaction barrier when a flavin molecule is docked in a sandwich-like structure with the distance d=3 Å (which is slightly less then sum of van der Waals radii of two carbons).

First, we found that the reaction barrier for the a -proton abstraction in the model system (Figure 3) without flavin is 7.7 kcal/mol.

Then the tricyclic oxidized flavin molecule was placed with the orientation appropriate to the crystal structure of the enzyme. The pre-reaction complex, transition structure and product complex have been optimized with two C-C distances (C1-C10a and Ca -C4a) constrained to 3 Å. The resulting reaction barrier was found to be about 2-fold lower (3.6 kcal/mol, Figure 4). The B3LYP/6-31G(d) calculations performed for the system with contact distance d=4Å, result in the reaction barrier 2.3 kcal/mol higher than that for the d=3Å. Remarkably, the 3Å-constrained pre-reaction complex is about 12 kcal/mol higher in energy than the 4Å-constrained pre-reaction complex. Thus, one may definitely conclude that flavin proximity lowers the reaction barrier for a -proton abstraction. The increase of contact distance by 1 Å may result in the 50-fold decrease in the reaction rate.

The anionic form of thioester substrate and oxidized flavin form a stable sandwich-like complex without any geometrical constraint (Figure 5). The estimated total energy of complexation is 22.2 kcal/mol. When the ribityl moiety, represented by the —CH₂CH₂OH group, is substituted by CH₃ (lumiflavin) in order to prevent the contribution of the H-bond stabilization, the complexation energy drops to 16.3 kcal/mol, which we assign to the charge-transfer interaction energy. In spite of the decrease in complexation energy upon H-bond removal, the interacting flavin and thioenolate come into closer contact [C₁C_10a=3.111 Å Ca C_4a=2.897Å vs. 3.269 Å and 2.934 Å, Figure 5]. Interestingly, when the thioester substrate is replaced by the 3-thia-butanoyl-CoA model substrate (Figure 6), the complexation energy is reduced significantly [22.2 vs 15.6 kcal/mol, Figure 5(A) and Figure 6], and the distance between substrate and flavin is markedly increased.

Spectroscopic studies of the acyl-CoA dehydrogenase model with the redox-inactive thioester analog 3-thia-octanoyl-CoA (that cannot do the hydride transfer to the flavin) have indicated the appearance of an intense charge-transfer band (centred at 800 nm) generated upon removal of the weakly acid a -proton.^4,5

We performed TD DFT (B3LYP) calculations on the charge-transfer model complex with the CH₃SCH(C=O)SCH₃^- model anionic form of the substrate (Figure 6). This charge transfer complex has three major absorption bands characterized by the calculated Franck-Condon (FC) transitions at 957, 408 and 337 nm [TD B3LYP/6-31G(d)//B3LYP/6-31+G(d,p),Table 2]. The first transition can be assigned to a charge-transfer band, whereas the two last transitions belong to flavin as indicated by the summary of TD B3LYP calculations for flavin (Table 3). We also examined the solvent effect (COSMO, Table 2) and found no significant changes except some increase in the oscillator strengths. For a comparison, the absorption spectrum of lumiflavin in water shows two characteristic bands at 446 nm and 370 nm, followed by an intense band at 270 nm.¹² Thus, as also noted in ref. 13, the theoretically predicted spectrum is somewhat blue-shifted with respect to the experimental one.

It should be noted that generally the TD DFT (B3LYP) approach predicts the excitation energies for low-lying excited states for single molecules remarkably well (very often within a few nm). As indicated in Table 2, the predicted CT band position is highly dependent upon the level of theory at which the complex has been optimized. The complex optimized with smaller basis set (6-31G(d)) is tighter and has a shorter wavelength of the CT transition (859 nm), which better agrees with the experimental observation. Since we do not include in our model all the residues surrounding the substrate/Fl_ox complex in studies of the acyl-CoA dehydrogenase with the redox-inactive thioester analog 3-thia-octanoyl-CoA,⁵ one should not necessary expect excellent agreement with experiment. The contact distance in the CT complex can be affected by external steric and electrostatic interactions.

The LUMO orbital of flavin consist of bonding/antibonding interactions of p (C8C7), p (C9C9a), lp(N10), p *(C4aN5), p *(C10aN1), p *(C2=O), p *(C4=O) orbitals (Figure 7).

The HOMO and (HOMO-3) molecular orbitals are mostly different combinations of p orbitals (Figure 8).

HOMO

p (C6C7C8),p (C5aC9aC9),p (C4aN10),lp(N1),lp(O_C2),lp(O_C4)

HOMO-3

p (C7C6C5a),p (C8C9C9a),p (C4aN5),lp(N10), lp(O_C4), lp(O_C2, lp(N1), lp(N3)

The LUMO orbital of the oxidized flavin (Figure 8) has a clearly distinguished p character of the C4a-C10a bond. We suggest that this may result in a charge-transfer interaction (partial electron transfer) between the substrate and Fl_ox, in particular, if the substrate has a complementary p bond with an excess of electrons.

In the CH₃SCH(C=O)SCH₃^-/Fl_ox model complex discussed above (Figure 6), the HOMO and LUMO orbitals responsible for generation of the CT band are the corresponding HOMO and LUMO of isolated CH₃SCH(C=O)SCH₃^- and Fl_ox(Figure 9).

The same observation was made for CH₃CH₂CH₂(C=O)SCH₃/Fl_ox (substrate/flavin) complexes which have been optimized with and without constraint of the C1C10a and Ca C4a distances. The CH₃CH₂CH(C=O)SCH₃/Fl_ox complex optimized without constraints is shown in Figure 5(A). Then, distances d= C1C10a= Ca C4a were fixed to 3 Å and re-optimized. Two other complexes with the distances increased to 3.5 and 4 Å were taken for excited state energy calculation without preliminary optimization. The results summarized in Table 4 suggest that the CT band moves to the low-energy region and loses its intensity with increase in contact distance d, whereas the two other flavin bands are not affected so significantly. The decrease in contact distance leads to an increase in negative charge on Fl_ox up to about —0.7 e in the CT complex which is 70% of whole "-1" charge of the complex initially provided by thioenolate.

 HOMO LUMO

5. Conclusions.

The present computational study, in conjunction with available published experimental data on acyl-CoA-dehydrogenases, suggests that the developing enolate can be subjected to strong charge-transfer interaction with flavin (16.3 kcal/mol). Analysis of the charge distribution in thioenolate/Fl_ox complexes indicates significant charge transfer to flavin (about 70%). This type of complex is characterized by an intense charge-transfer absorption band (644 nm) which is assigned to p à p * transfer from the highest occupied molecular orbital of thioenolate to the lowest virtual orbital of flavin. The energy of this charge-transfer vertical (Franck-Condon) excitation depends strongly on the contact distance but appears less sensitive to the polarity of microenvironment.

The B3LYP calculations of two model systems (with and without flavin) designed for simulation of a -proton abstraction in acyl-CoA-dehydrogenases show that proximity of flavin may have a strong impact on the reaction efficiency leading to a 2-fold decrease in the reaction barrier. It is suggested that flavin proximity due to a charge-transfer interaction causes a significant acidifying effect on the a -proton.

The available X-ray structures of acyl-CoA-dehydrogenases suggest that the geometrical constraint imposed by the active site of the enzyme as well as H-bonding with the substrate carbonyl oxygen forces the thioester-substrate to approximate within van der Waals contact with the flavin prosthetic group. This, according to our computational simulations, should acidify the a -proton of the substrate and accelerate base-catalyzed proton abstraction.

The effect of flavin proximity on the a -proton abstraction step as well charge-transfer interaction between thioenolate and flavin raises the possibility of a significant impact upon the reaction-dependent molecular motion on the overall reaction kinetics.

Acknowledgment. This work was supported by the National Science Foundation (CHE-0138632), by NIH GM26643 and partially supported by National Computational Science Alliance under CHE990021N and utilized the NCSA SGI Origin2000 and University of Kentucky HP Superdome.

1. (a) Schopfer, L. M.; Massey, V.; Ghisla, S.; Thorpe, C. Biochemistry 1988, 27, 6599. (b) Thorpe, C.; Kim, J-J. P. FASEB J. 1995, 9, 719. (c) Rudik, I.; Ghisla, S.; Thorpe, C. Biochemistry 1998, 37, 8437. (d) Engst, S.; Vock, P.; Wang, M.; Kim, J-J. P.; Ghisla, S. Biochemistry 1999, 38, 257; and references therein. (e) Johnson, J. K.; Srivastava, D. K. Biochemistry 1993, 32, 8004.

2. Bach, R.D.; Thorpe, C.; Dmitrenko, O. J.Phys.Chem.A. 2002, 106, 4325-4335.

3. (a) Kim, J-J.P; Wang, M.; Passchke, R. Crystal structures of medium chain acyl-CoA dehydrogenase from pig liver mitochondria with and without substrate. Proc. Natl. Acad. Sci. USA 1993, 90, 7523. (b) Djordjevic, S.; Pace, C. P., Stankovich, M. T.; Kim, J. J. Biochemistry 1995, 34, 2163; (c) PDB Ref.Code is 3MDE for medium-chain Acyl-CoA dehydrogenase and 1BUC for butyryl-CoA dehydrogenase; Berman, H.M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T.N.; Weissig, H.; Shindyalov, I.N.; Bourne, P.E. The Protein Data Bank. Nucleic Acids Research, 2000, 28. 235.

4. (a) Lau, S.-M.; Brantley, R.K.; Thorpe, C. Biochemistry 1988, 27, 5089. (b) Tamaoki, H.; Nishina, Y.; Shiga, K.; Miura, R. J. Biochem. 1999, 125, 285.

5. Rudik, I.; Thorpe, C. Arch. Biochem. Biophys. 2001, 392 (2), 341.

6. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone,V.; Cossi, M.; Cammi,R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98, Revision A.7, Gaussian, Inc., Pittsburgh PA, 1998.

9. (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (b) Stevens, P. J.; Devlin, F. J.; Chabalowski, C. F. and Frisch, M. J. J. Phys. Chem. 1994, 98, 11623-11627.

10. (a) Bauernschmitt, R. ; Ahlrichs R. Chem. Phys. Lett. 1996, 256, 454—464. (b) Casida, M. E.; Jamorski, C.; Casida, K. C.; Salahub, D. R. J. Chem. Phys. 1998,108, 4439. (c) Stratmann, R. E.; Scuseria, G. E. ; Frisch, M. J. J. Chem. Phys. 1998, 109, 8218—8224.

11. Barone, V.; Cossi, M.; Tomasi, J. J. Comp. Chem. 1998,19, 404.

12. Dudley, K. H.; Ehrenberg, A. ; Hemmerich P. ; Müller F. Helv.

Chim. Acta 1964,47,1354—1383.

13. Neiß, C.; Saalfrank, P. ; Parac M.; Grimme S. J Phys. Chem A. 2003,107: (1) ,140-147.