[gmx-users] to neutralize the system

[Justin Lemkul]

On 6/19/16 11:03 AM, Alexander Alexander wrote:
> Hello Mark,
>
> I still need help choosing the right method of neutralization of my system.
> I am wondering how the final states of the surface-binding process can shed
> light on choosing the feasible method of neutralization, here, either
> protonation (deprotonation) of amino acid and peptide or adding ions Na+
> (Cl-)?
>

In the generic "charge neutralization" discussions here in recent days, people 
are talking about artifacts associated with using non-neutral simulation systems 
and relying on a uniform background charge from PME to "neutralize" these 
systems.  In some cases, the artifacts can be quite pronounced so counterions 
are a simple fix.

Neutralizing side chains has important physical consequences; you should choose 
the protonation state based on the actual physical properties of the system and 
the chemical environment (e.g. solution, the calculate pKa values of the 
species, etc).  Simply neutralizing side chains to fit a preconceived notion 
that the solute should be neutral (conflating the PME issues with real physical 
issues) is a dangerous road to go down.

> For example, in one sample, I neutralized the Glutamic(zwitterionic form)
> by protonating it and then in simulation, it gradually absorbed firstly to
> the surface via one of the Hydrogen(C-H) in the end of the side chain, and
> then laid steady down while NH3 group was pointing up.
> In another sample, again I neutralized the Arginine(zwitterionic form) by
> deprotonating it and then in simulation, it absorbed firstly to the surface
> via the COO(-) group in the backbone, and then laid steady down similar to
> the Glutamic.
>

Welcome to the challenge of using fixed charge/fixed topology simulations. 
Protonation states, in real systems, can vary.  Using traditional MD, they 
cannot.  There are methods to get around this (e.g. constant-pH simulations, but 
those are not - to my knowledge - readily available in GROMACS, though people 
have published wrapper-type approaches to doing this).

What you need to simulate is the most probable form of your peptides under 
whatever the solution condition is.  Deprotonating Arg requires crazily basic 
conditions, which may or may not be relevant to what you're studying.  Perhaps 
the protonation state does vary during the binding mechanism.  That's a 
challenge for you to address based on your knowledge of your specific system and 
published studies of similar systems.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================