Hi Rukmankesh,

Based on the knowledge of you system, you should try to manually place the ions at the protein region you want before creating your neutralized protein-bilayer system with CHARM-GUi. You can also use a docking program to couple your ions but I dont know how appropiate be this approach. Remember to consider these ions during the initial steps of bilayer construction to the program consider the position of these ions in your system.

Hi Martiniano,

Thank you for your response. Is it possible during system preparation in Gromacs to add some concentration of ions at specific location? gmx genion randomly places ions. Is there any other way?

Manually put it. You can use software such as vmd or text editor to put it close to the position you want.Then minimization and dynamics simulation would provide you an equilibrium structure.

Rukmankesh Mehra I use gmx insert-molecules to place ions where I want them. You use the "-ip " flag to pair the command with a positions.dat file like the one I've attached in which I placed 66 chloride ions at different places above and below a double bilayer system. The positions.dat file contains XYZ coords. You can then use the "-dr " flag which sets the allowable displacement.

To give you an example, say you have a 5 nm x 5 nm x 5 nm box and you want a to place an ion in the box. If I wanted to put it dead center, I would put the following in the positions.dat file (below) and give it in an allowable displacement of 0 nm by using this flag "-dr 0 0 0" which says it cannot be displaced from this position in the X, Y, or Z dimension.

# positions.dat file

# x y z

2.5 2.5 2.5

If I wanted the ion within 1 nm of the center of box, I would pass the same positions.dat file with the "-ip " flag and use a "-dr 1 1 1" to signify that the ion can be be placed up to 1 nm away (in X, Y, and Z) from the position I put in the positions.dat file.

So if you want to place an ion in your system. Figure out the coordinates. Decide on an allowable displacement. And then I'd just use trial and error until you get it right where you want it. Alternatively, I think VMD has a function where you can drag molecules around and place them where you like. That may be easier if you only need to place a few ions in certain places.

JD Tamucci Thank you very much for a very useful response! I will implement the same. Where to define the ion type (such as chloride) in gmx insert-molecules? What is the -ci flag? Do I need to use gmx editconf -center for the protein-membrane system?

Rukmankesh Mehra I think the gmx editconf -center is a matter of preference. If you'd rather set the center of your box to 0 0 0, then go right ahead. It may make it a bit simpler to understand the positions, but it's not necessary for the command.

For the gmx insert-molecules command, here's the command I used to insert those ions and an accompanying explanation of each flag.

gmx insert-molecules -f solvated_system.gro -ci CLA.pdb -ip CLA_equal_positions.dat -nmol 66 -try 10000 -dr 3.4 3.4 1 -o ionsolv.gro

-f solvated_system.gro: This is my double bilayer system with water that I wanted to insert ions into.

-ci CLA.pdb: This is a PDB of the chloride ion I intended to add.

For reference, I wanted to add 66 of them (hence -nmol 66) with an allowable displacement equivalent to my XY dimensions and the Z distance between the bilayers (-dr 3.4 3.4 1). I also told it to try to insert every one 10000 times (-try 10000) just because I didn't want it to insert fewer than 66. Having a high number of tries increases the odds of inserting the correct amount.

Dear JD Tamucci Thank you very much for sharing your experience and very useful information!

Zhaoxi Sun Thank you for your suggestion!

Note that while you can place ions in any part of the unit cell you like, there is no "extracellular" or "intracellular" compartment. Due to periodic boundary conditions, the water layer is continuous, so the ions are not going to show you any directionality. They will quickly mix.

Justin Lemkul brings up a good point Rukmankesh Mehra . If you're looking to model compartments with different ion concentrations, then you'd need a double bilayer system. But if you just want to model in an ion and restrain it to your protein, or place it inside the transmembrane region of your protein next to some oppositely charged residue, one bilayer may suffice.

I'd advise you to be careful with modelling compartments with different ion concentrations though because it can create spurious membrane effects. Having as small as a few ions charge difference between the compartments can cause electroporation, and not necessarily due to your protein, but rather just due to the membrane potential. This paper goes over it some.

Article Probing Lipid Bilayers under Ionic Imbalance