RE: energy difference between HSE and HSD

From: Radak, Brian K (bradak_at_anl.gov)
Date: Tue May 09 2017 - 11:12:14 CDT

I based that on the free energy of transformation between the HSE and HSD topologies in solution, which is ~16 kcal/mol when NMR gives a result more like ~-1 kcal/mol (note that the sign is also different).

You raise a good point. Should the force field model even produce such a result? The covalent bond energies are probably extremely important here and the force field has no physical description of those effects. One could try looking up tabulated bond energies and making an adjustment, but that seems rather ad hoc.

Brian Radak
Postdoctoral Appointee
Leadership Computing Facility
Argonne National Laboratory

9700 South Cass Avenue, Bldg. 240
Argonne, IL 60439-4854
(630) 252-8643
brian.radak_at_anl.gov
________________________________
From: Jérôme Hénin [jerome.henin_at_ibpc.fr]
Sent: Tuesday, May 09, 2017 10:51 AM
To: Namd Mailing List; Radak, Brian K
Cc: Hao Dong
Subject: Re: namd-l: energy difference between HSE and HSD

I must missing something here. How is a FF of this type supposed to reproduce a tautomerization energy? How do you calculate a tautomer ratio from C36?

Jerome

On 9 May 2017 at 17:43, Radak, Brian K <bradak_at_anl.gov<mailto:bradak_at_anl.gov>> wrote:
The HSE and HSD parameters were clearly not designed to reproduce tautomerization energies. The C36 force field does in fact get the tautomer ratio wrong in solution (by a LOT).

Maybe this will be improved in future force field versions?

Brian Radak
Postdoctoral Appointee
Leadership Computing Facility
Argonne National Laboratory

9700 South Cass Avenue, Bldg. 240
Argonne, IL 60439-4854
(630) 252-8643
brian.radak_at_anl.gov<mailto:brian.radak_at_anl.gov>
________________________________
From: owner-namd-l_at_ks.uiuc.edu<mailto:owner-namd-l_at_ks.uiuc.edu> [owner-namd-l_at_ks.uiuc.edu<mailto:owner-namd-l_at_ks.uiuc.edu>] on behalf of Hao Dong [donghaonj_at_gmail.com<mailto:donghaonj_at_gmail.com>]
Sent: Monday, May 08, 2017 11:48 PM
To: namd-l_at_ks.uiuc.edu<mailto:namd-l_at_ks.uiuc.edu>
Subject: namd-l: energy difference between HSE and HSD

Dear all,

I built two pdb files to calculate the energy difference between residues HSE and HSD, that is Histidine in different protonated states (the structures are "ACE-HSD-CT2" and “ACE-HSE-CT2”).

I tried Charmm36 force field and got E(HSD)-E(HSE)=13kcal/mol. Then I used Gaussian(b3lyp/6-311g(d,p)) and Amber force field to calculate the energy, it came out that: E(HSD)-E(HSE)=1-2kcal/mol.

Obviously, 13kcal/mol is too large for this situation. I checked the NAMD output files and found great gap in ELECT energy, which made me very confused. Why the energy difference between the two forms were overestimated by using Charmm36 force field ?

Here is my input file:
minimization on
numsteps 0

set targetemp 310.0
set coorfile hsd_1.pdb
coordinates $coorfile
structure hsd_1.psf

outputname myout1
restartname myrst1
outputEnergies 1
binaryoutput no
xstFreq 1
DCDFreq 1
restartFreq 1

COMmotion no
timestep 1.0 ;# 2fs/step
nonbondedFreq 1
fullElectFrequency 2
temperature $targetemp

paraTypeCharmm on
parameters par_all36_prot_wat_ions.prm

exclude scaled1-4
1-4scaling 1.0
dielectric 1.0
switching on
switchdist 10.0
cutoff 12.0
pairlistdist 14.0 ;# cutoff + 2
stepspercycle 10

This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:21:16 CST