Figure 7.

The influence of membrane bending on computing the biological hydrophobicity scale and the interplay of electrostatics and nonpolar forces. (A) Amino acid insertion energies for 17 residues calculated using our membrane-bending model (green bars) and compared with the translocon scale (Hessa et al., 2005a) (red bars) and a scale developed from MD simulations of lone amino acids (MacCallum et al., 2007) (blue bars). All three scales were shifted by a constant factor to set the insertion energy of alanine to zero (1.97 kcal/mol green, −0.11 kcal/mol red, 2.02 kcal/mol blue). The insertion energy of charged residues is reduced by 25–30 kcal/mol by permitting membrane bending. Calculations similar to those in Fig. 6 B indicate that only charged residues result in distorted membranes. (B) The energy difference between very polar amino acids and charged amino acids is quite small; however, our model predicts that the physical scenarios are quite different. The insertion penalty for asparagine is primarily electrostatic, while the nonpolar component stabilizes the amino acid (left bars). Conversely, there is very little electrostatic penalty for inserting arginine and most of the cost is associated with the nonpolar energy required to expose the TM domain to water (right bars). For comparison, we show that the classical flat membrane gives rise to a huge electrostatic penalty and a significant 6 kcal/mol membrane dipole potential penalty for arginine (middle bars).