Objective and Overview

The objective of this tutorial is to illustrate usage of CHARMM GBSW (generalized Born with a smoothed switching function. Specifically, we will:

• Setup an implicit solvent protein simulation using an optimized GBSW implicit solvent force field.
• Generate the pentameric phospholamban from PDB:1ZLL and orient its transmembrane (TM) domain along Z
• Solvate the protein with GBSW implicit solvent
• Minimize and equilibrate the system in GBSW membrane
• and/or use 5-fold symmetry in the membrane simulation

Phospholamban is a transmembrane (TM) protein which regulates Ca2+ ATPase. The number of simulation steps is limited in this tutorial due to the time constraint, but can be extended to explore the stability and dynamics of the protein.

The TAR/GZ file contains all the necessary files, including a README file for a more detailed description of associated files. CHARMM version 33a2 (33b1) or later is required.

To run the sample simulations:

 $CHARMMEXEC < INPUTSCRIPT_FILENAME > LOG_FILENAME

or (tee bifurcates the output to screen and file),

 $CHARMMEXEC < INPUTSCRIPT_FILENAME | tee LOG_FILENAME

---

1. GBSW implicit solvent simulation of protein G B1 domain

  gbsw_peptide.inp

This example illustrates the usage of GBSW implicit solvent simulation of a small protein. Several key differences from a typical explicit water simulation exist.

• GBSW-specific CMAP: the peptide backbone torsion potentials have been optimized consistently with the input radii for GBSW to accurately describe peptide conformational equilibria.

    read  rtf card name @toppar/top_all22_prot_cmap.inp
    read para card name @toppar/par_all22_prot_gbsw.inp
  

• Nonbonded interactions: no need to setup periodic boundary conditions (unless you have good reasons). Note that GBSW actually supports image (see example below). Also note that cton=ctof. This is because the switiching is already included in GBSW.

    NBOND atom switch cdie vdw vswitch -
          ctonnb 16.0 ctofnb 16.0 cutnb 20.0
  

• Setup the input radii: the intrinsic atomic radii define the location of the solute-solvent boundary and are one of the most important phyiscal parameters for GBSW (and any GB/PB models that allow radii input).

    stream @toppar/radius_gbsw.str

  The file "radiius_gbsw.str" contains a set of input radii that have been optimized for GBSW peptide simualtion (Nina et al, 1997; Chen et al, JACS 2006). Note that this radii set should always be used together with the GBSW-specific CMAP for best accuracy in peptide conformational equilibria.

• "Solvate" using GBSW: invoking GB solvent such as GBSW is very simple

    gbsw sgamma 0.005 nang 50 

  "sgamma" is the effective surface tension coefficient (0.005 is recommended for all GBSW simulations). "nang" is the number of angular integration points. Larger nang increases the accuracy, but the calculation becomes more expensive. nange=50 is necessary (and mostly sufficient).

Once GBSW is setup, the rest is same as any "regular" simulations, except for a few small differences. For example, there is no pressure control (most GB models including GBSW are parameterized to mimic 1 atm pressure and this is not adjustable) and PME is not applicable. Either Nose-Hoover thermostat or Langevin dynamics is typically used for temperature control.

2. GBSW implicit membrane simulation of phospholamban

  gbswmemb1.inp

This input illustrates a typical setup for simulations in GBSW membrane without images. In this example, we will

1. The raw PDB file needs to pre-processed by MMTSB/convpdb.pl (or by another tool), mainly to extract the desired chain and/or model and to add "segid" column.

     convpdb.pl -chain A -model 1 1ZLL.pdb >! 1zll_monomer.pdb
     convpdb.pl -model 1 -segnames 1ZLL.pdb >! 1zll_model1.pdb
   

2. Generate the PSF for the pentameric channel (5X repeat operations fro each chain) and read the initial coordinates. Many atoms (mostly hydrogens) will still miss coordinates after reading, mainly due to atom naming differences. CHARMM has built-in facilities to build missing coordinates from known coordinates based on the covalent geometry (defined either in IC tables or in parameter files).

     ic param
     ic build
     hbuild
   

3. Orient the TM domains along Z and adjust their position in the membrane (such that the C-terminal locates at the membrane interface)

     coor orient
     coor stat
     coor rotate ydir 1.0 phi 90.0
     coor trans zdir -?zmin
     coor trans zdir -17.0
   

4. Set up input radii and nonbond options (see above)
   
   
5. Request GBSW implicit membrane:

     gbsw sgamma 0.005 nang 50 tmemb 30.0 msw 2.5

   Compared to GBSW implicit solvent, the new keyword "tmemb" activates the implicit membrane option of GBSW module. "msw" is the half of switching length over which hydrophobic environment is changed to solvent region. Thus, "tmemb-msw*2" is the hydrophobic thichness of the membrane (25 A with the above).

6. Minimize, equilibrate and run a very short production. You are welcome to apply your (newly acquired) VMD and CHARMM analysis skills to examine the results.

3. GBSW implicit membrane simulation with images

  gbswmemb2.inp

This input illustrates how to use GBSW membrane with images. We will attempt to simulate the same system as in "gbswmemb1.inp". Only one monomer is present, but a 5-fold sysmmetry is imposed to describe the pentermic channel:

  READ IMAGE
  * 5-fold symmetry
  *

  IMAGE RT2
  ROTATE 0.0 0.0 1.0  72.0

  IMAGE RT3
  ROTATE 0.0 0.0 1.0 144.0 

  IMAGE RT4
  ROTATE 0.0 0.0 1.0 216.0

  IMAGE RT5
  ROTATE 0.0 0.0 1.0 288.0

  END

The images must be set up before invoking GBSW. GBSW is compatable with any kinds of images in solvent or membrane. The rest of the script is exactly the same as "gbswmemb1.inp".