Binding energy calculation

Calculation of electrostatic binding energies for the PKA-balanol system requires higher-resolution electrostatics calculations than could be easily visualized in the previous section. We will use focusing to generate the finer data from the coarser calculations performed in the earlier calculations. This procedure is outlined in the following APBS input file which is available for download here.

Example 2. Balanol-PKA energy calculation APBS input


read 
    mol pqr bx6_7_lig_apbs.pqr  # Balanol
    mol pqr bx6_7_apo_apbs.pqr  # Protein kinase A
    mol pqr bx6_7_bin_apbs.pqr  # Complex
end

elec name bal                # BALANOL ENERGY CALCULATION
    mg-auto   
    dime 65 65 65            # Grid dimensions
    cglen 70 70 70           # Coarse grid lengths
    cgcent mol 3             # Coarse grid centered on the complex
    fglen 16 16 16           # Fine grid lengths
    fgcent mol 1             # Fine grid centered on the complex
    mol 1                    
    lpbe                     # Linearized PB
    bcfl sdh                 # Monopole boundary condition
    ion 1 0.000 2.0          # Zero ionic strength
    ion -1 0.000 2.0         
    pdie 2.0                 # Solute dielectric
    sdie 78.00               # Solvent dielectric
    chgm spl0                # Linear charge discretization
    srfm smol                # Smoothed molecular surface
    srad 0.0                 # Solvent probe radius
    swin 0.3                 # Surface spline window (not used)
    sdens 10.0               # Sphere density
    temp 298.15              # Temperature
    gamma 0.105              # Surface tension (not used)
    calcenergy total
    calcforce no
end

elec name pka                # PROTEIN KINASE A CALCULATION
    mg-auto   
    dime 65 65 65
    cglen 70 70 70
    cgcent mol 3
    fglen 16 16 16
    fgcent mol 1
    mol 2
    lpbe
    bcfl sdh
    ion 1 0.000 2.0
    ion -1 0.000 2.0         
    pdie 2.0       
    sdie 78.00     
    chgm spl0      
    srfm smol      
    srad 0.0       
    swin 0.3       
    sdens 10.0 
    temp 298.15    
    gamma 0.105   
    calcenergy total
    calcforce no
end

elec name complex            # COMPLEX CALCULATION
    mg-auto   
    dime 65 65 65
    cglen 70 70 70
    cgcent mol 3
    fglen 16 16 16
    fgcent mol 1
    mol 3
    lpbe
    bcfl sdh
    ion 1 0.000 2.0
    ion -1 0.000 2.0         
    pdie 2.0       
    sdie 78.00     
    chgm spl0      
    srfm smol      
    srad 0.0       
    swin 0.3   
    sdens 10.0    
    temp 298.15    
    gamma 0.105   
    calcenergy total
    calcforce no
end

print energy complex - pka - bal end

quit
        

Note

This APBS input file represents a "non-canonical" method for computing energies. In any electrostatic energy calculation, care needs to be taken to remove self-interaction terms. This removal is usually accomplished via a solvation energy calculation (cf. the Born example). However, in this example, we remove the self-interaction energies via a series of calculations which use exactly the same grid spacings, lengths, and centers. Since all atoms in the complex are present in the apo protein and ligand, the self-interaction terms are effectively removed.

This example can be run by typing (from the system prompt):

        $ apbs apbs-energy.in
      
(assuming the downloaded input file was name apbs-energy.in). This example takes roughly 2 minutes no a 2.2 GHz Intel Pentium 4 and should give a binding energy of 5.8 kJ/mol -- which is in good agreement with the reported results.

One question that should arise is: why is the binding unfavorable? The Coulombic contribution to the binding energy is favorable (you can use the APBS program coulomb to evaluate this); but there is a substantial polar desolvation penalty for binding of the substrate to the protein. However, our current calculation has neglected apolar contributions to the energy. These terms, which usally scale with the surface area, favor the binding. Using the APBS program acc, we can determine that over 1000 A2 of solvent-accessible surface area is lost when the complex is formed. Using even conservative values of the apolar energy surface tension (which usually ranges between 0.105 to 210 kJ/A2), we can see that this buried surface area provides stabilization to the complex.