With the flag -inter, you can decide interactively the different options: C and N terminal state, protonation state of LYS, ARG, ASP, GLU, GLN and HIS, and the state of disulfide bridge.

Running at different pHs is much more difficult than running at different temperatures. Fundamentally, this is because proton exchange is a quantum effect and can't be easily handled in classical simulations. The usual way to do it is to run the simulation with different protonation states of the protein's amino acids, although this might not correspond exactly to any real pH (near the pKa of the titratable group, the protonation state is a fluctuating quantity).

Check out this Gromacs webpage:

http://www.gromacs.org/Documentation/How-tos/Constant_pH_Simulation

There is also a multiple replica algorithm that looks promising.

http://pubs.acs.org/doi/abs/10.1021/ct401042b

Two options are possible in GROMACS: 1) if you add ACE and NHE group  to the terminal group, all protein becomes neutral. some proteins neutral at low pH. 2) if you do not add ACE and NHE group it becomes positively of negatively charged depending on number of  amino acids in whole protein.

I think I can do now. I am using H++ Server to increase/Decrease pH of the system.

But, I want to try by different methods. Do you have any script/protocol step by step.

How to add ACE/NHE?

Protonation states can be changed using the proper command-line options in pdb2gmx.

I do not follow this argument about ACE and NHE; these are merely capping groups that neutralize termini, often used in studying peptide fragments where charged termini would lead to artifacts.  They have no effect on the net charge of the protein.  If termini are normally charged, they offset and contribute zero charge.  If they are capped, they still contribute zero charge.

@Justin Lemkul:

Please explain that to me, what option can I use in pdb2gmx. can you please write one command line for me. I am using this1,

pdb2gmx -f XXX.pdb -o XXX_processed.gro -water spce

With the flag -inter, you can decide interactively the different options: C and N terminal state, protonation state of LYS, ARG, ASP, GLU, GLN and HIS, and the state of disulfide bridge.

Florencia's answer is correct.  Note that you can discover such options yourself for all Gromacs commands by invoking the -h argument, otherwise everything is documented in the print manual.

I suggest using PROPKA tool to determine pKa's of amino acids. It is available freely as a standalone tool. PROPKA is also implemented as a part of Schrodinger's Maestro package; there you can use the tool to prepare the protein for MD. The protein prep seems to be more sophisticated than GROMACS' pdb2gmx, because of fancier pKa calculator and the ability to flip His/Asn/Gln to create a better hydrogen bonding network.The Maestro interface is more tailored for the Schrodinger's own MD program Desmond, but nothing prevents you from taking the coordinates from the prepped protein and feed in into pdb2gmx, just be careful about what you are doing.

I am using Ds server successfully. Can any1 please prepare the PDB at ph 3, 4, 5, 6, 7, 8 using different server.

Thanks

The Propka is a great tool to look the protonation states according to the chemical microenvironment of the molecule, It's guiding my choice when gromacs asks for the protonation states of the residues. But, with Amber (sander) its more difficult, so the H++ is a good option, even then that  top and crd files don't work..  The H++ generates a pdb file in a predetermined protonation state (under pH value). But, if someone knows how to make just a simple dynamics run simulation that changes the protein's protonation states... please tell me :)  

I used H++, but results are same for all :(

Hi Faez, take a look at the ProteinPrepare web application (www.playmolecule.org). It uses PROPKA3.1 for protonation prediction plus PDB2PQR for H-bond network optimization. You can then download results and optimized structure for simulation.