Protein-Phosphopeptide per-residue decomposition MM/GBSA

[Samuele Di Cristofano]

Hello,

I'm running a MM/GBSA per-residue decomposition analysis on a protein-phosphopeptide complex.

I'm trying to figure out why the energy decomposition of phosphoSerine 6 (SEP:6), is positive.

Since this is the main binding hotspot, I expect an high negative value for this interaction.

By visually inspecting the trajectory, the SEP:6 residue forms H-bonds with surroundings residues and everything looks fine.

Do we need to add forcefield parameters for phosphorylated residue?

This is the .in file I'm using:

&general
sys_name="test",
startframe=1, endframe=751, interval=50, verbose=2
forcefields="oldff/leaprc.ff14SB,leaprc.gaff,leaprc.phosaa14SB"
/

&gb
igb=5, saltcon=0,
/
&decomp
idecomp=2, dec_verbose=3,
print_res="within 3"

This is the per-residue decomposition output (TDC):
L ACE   1 -0.27
L SER   2 -0.69
L GLY   3 -0.69
L PRO   4 -2.06
L ARG   5 -6.20
L SEP   6 16.94
L PRO   7 -2.21
L SER   8 -1.20
L GLY   9 -1.30
L NHE  10 -1.33
R ARG  14 -0.44
R SER  16 -8.26
R ARG  17 -14.1
R SER  18 -6.93
R TYR  23 -3.09
R PHE  25 -1.95
R ASN  30 -0.54
R SER  32 -1.96
R TRP  34 -2.23
R GLU  83 -0.81
R LEU  86 -1.09
R GLU  87 -0.59
R ASN  90 -1.02
R ILE  93 -0.47
R GLN  94 -0.63
R LYS  97 -0.10
R PRO 149 -1.06
R VAL 150 -0.41
R PHE 151 -0.97

Any comments or suggestion are welcome!

Thanks in advance,
Samuele

[Mario Sergio Valdes]

Hi Samuele.

Can you zip this test folder and send it to me with all the results for review?

The parameterization of this residue may not be properly converted.

Mario S.

[Mario Sergio Valdes]

Summary of the offline discussion:
Basically to be expected given that this residue is charged. In this case, it is recommended to use the igb=8 model and vary the dielectric constant (intdiel) to avoid overestimating the electrostatic component and the polar component of the solvation. It requires a literature search for the most commonly used parameters for similar systems or an exhaustive test according to your study objective.

[Krish]

Hi All

I am calculating MM/GBSA+IE for a protein-peptide. My peptide contain phosphoserine and phosphothreonine. Can anybody tested for optimal parameters for calculations. I have tried all igb and suggested intdel by couple for papers used MMGBSA.py (Amber). The energy is positive after considering IE term. Although the MM/GBSA energy is negative, but SD is very high. Any suggestion/advise are welcome. Thanks 

[marioe911116]

Convergence of protein-peptide systems energies tends to be a little bit difficult to achieve as the peptide is much bigger (with a lot of polar groups and different kinds of interactions) and more mobile than the traditional small molecule drugs. This leads to higher standard deviations in the internal energy and that's why your IE is not probably useful in this case. Check this paper for more details on interaction entropy and standard deviation of the internal energy reported by gmx_MMPBSA:

https://pubs.acs.org/doi/full/10.1021/acs.jctc.1c00374

On the other hand, high SD in the MM/GBSA energy could mean that you need more simulation time to achieve convergence... How about the SEM (standard error of the mean) for the MM/GBSA energy? This is traditionally what's reported... 

[Krish]

Hi Mario)

Thank you for information. I already read this paper. I took 500 snap shot from 500ns to 1000ns trajectories. I tried igb 1, 5 and 8 with different PBRadii and for extdiel=80 and intdiel=1. The results were unrealistic other than igb=1.  According to https://link.springer.com/article/10.1186/2046-1682-4-12#Sec2. 

igb=1, PBRadii=2 gave me some acceptable energy for mmgbsa but the SD is higher i.e ~20 and SME bellow 1.  

Do you recommend some other entropy than IE? Do you recommend other PBRadii or dielectric or any other parameters to care about?

Thanks & Regards

Krishna

[marioe911116]

If you are working with phosphopeptide, probably a higher internal dielectric value will help as it should be a highly charged binding site