Practical 7: MD simulation of a membrane protein

Bert de Groot

Contents
Setting up of a membrane protein simulation Analysis of water permeation in aquaporin-1 Mechanism of proton exclusion References

Setting up of a membrane protein simulation

First of all, download the bovine aquaporin-1 structure (PDB code: 1J4N) currently available from the Protein Data Bank.

Have a look at the structure using pymol: 

wget http://www3.mpibpc.mpg.de/groups/de_groot/compbio/p5/1J4N.pdb
pymol 1J4N.pdb

Before we can feed the structure to gromacs, we first have to delete from the PDB file a detergent molecule that was co-crystallized with the aquaporin. We do this with the unix 'grep' command. The '-v BNG' option tells 'grep' to print all lines except for the ones containing the string 'BNG': 

grep -v BNG 1J4N.pdb > t.pdb 
mv t.pdb 1J4N.pdb

now we can generate hydrogen positions and a gromacs topology using pdb2gmx: 

gmx pdb2gmx -f 1J4N.pdb -o 1J4N_H.pdb

When asked for a force-field, select "15" (the OPLS/AA-L force-field), and select the TIP4P water model. Again examine the structure using pymol: 

pymol 1J4N_H.pdb

You may remember from the lecture that aquaporins are active as tetramers, so if we want to study water permeation, we should simulate an aquaporin tetramer. Since we only have the structure of one monomer so far, we have to generate the coordinates of the other three monomers based on symmetry. Usually, this can be done with software like WHAT IF, but in this case it is more convenient to use a small script that can be downloaded from our site. Before we can use the script, we must make it executable: 

wget http://www3.mpibpc.mpg.de/groups/de_groot/compbio/p5/maketet_i4.csh
chmod +x maketet_i4.csh

and now we can use it to generate the tetramer: 

./maketet_i4.csh 1J4N_H.pdb tet.pdb

Have a quick view on the structure with pymol: 

pymol tet.pdb

Now we'll merge the protein system with a pre-equilibrated POPE membrane to yield the final simulation system. Download the membrane from us, and have a look at it with pymol: 

wget http://www3.mpibpc.mpg.de/groups/de_groot/compbio/p5/pope.pdb
pymol pope.pdb

Before we can do the actual merge, we have to position the protein to match the lipid coordinates: 

gmx editconf -f tet.pdb -o conf.pdb -box 11.485 11.383 10.394

and now we can merge the two, removing all lipids and water molecules that would overlap with the protein: 

 gmx solvate -cp conf.pdb -cs pope.pdb -o box.pdb

Analyse the result with pymol: 

pymol box.pdb

You now see the (almost) complete simulation system: a protonated aquaporin-1 tetramer embedded in a solvated POPE lipid bilayer. Before a production run can be started (to study water permeaion), the system would first have to be energy-minimised, counter-ions would have to be placed to compensate the net charge of the protein before a dynamical equilibration can be started.

Question: Aquaporins are passive water channels, driven by osmotic gradients across the membrane. If we want to simulate such a process, how could we generate a concentration gradient across the membrane? Wouldn't that dissipate across the periodic boundaries of the simulation box rather than by permeation through the channel?

Question: Would it be possible to study permeation rates also without applying an osmotic gradient? What could we learn from simulations in which water molecules diffuse through the channel, without a driving force defined by the osmotic gradient?

Since an actual simulation of water permeation would take too long for this practical, we now make a leap in time, and pretend we have already carried out the simulation.


Go back to Contents

Analysis of water permeation in aquaporin-1

First of all, download the trajectory (fragment) full.xtc and a reference PDB file ref.pdb and create a small animation: 

wget http://www3.mpibpc.mpg.de/groups/de_groot/compbio/p5/full.xtc
wget http://www3.mpibpc.mpg.de/groups/de_groot/compbio/p5/ref.pdb
gmx trjconv -s ref.pdb -f full.xtc -o movie.pdb -e 4100

Select "0" when prompted for a group. 

pymol movie.pdb

Press the "play" button on the lower right of the main pymol button to see the animation. You will appreciate the difficulty of analysing water permeation by simple visual inspection. This is why specialised analysis tools have been developed to follow all water molecules during the simulation, and to assess which water molecules participate in permeation. Gromacs works with the concept of "index groups" which are basically groups of atoms with a particular property. One could for example define an index group containing all CA atoms of the protein, and then use this index group to filter all CA atoms from an MD trajectory for subsequent analysis. Such an index file has been generated that contains a selection of water molecules that participate in a particular permeation event. Download the index file and extract the protein plus a selection of water molecules from the trajectory like this: (First remove the old movie file, as it is quite big) 

rm -f movie.pdb
wget http://www3.mpibpc.mpg.de/groups/de_groot/compbio/p5/index.ndx
gmx trjconv -s ref.pdb -f full.xtc -o movie.pdb -n

and select e.g. group 21.
Now view with pymol: 

 pymol movie.pdb

highlight one monomer of the aquaporin and the water molecules by typing (on the pymol prompt): 

hide all
select aqua, res 1-227
show cartoon, aqua
select water, resn SOL
show spheres, water

Now highlight one of the water molecules by colouring it yellow: 

select perm, res 3358
color yellow, perm

and press the "play button". You now see one water molecule that finds its way from the extracellular face of the membrane to the cytoplasmic side. Note that it might have gone just the opposite direction as well, as no driving force (like an osmotic pressure) was applied during the simulation. All permeation events, therefore, rest on pure diffusion of water molecules through the aquaporin pores.
It is always important to test whether the simulation results agree with experimental observations, whenever possible. In the aquaporin case, we fortunately have an excellent opportunity to directly compare our simulation results to experiments, namely to analyse whether the permeation rate in the simulations is similar to the experimentally determined one. In total, 16 full permeation events were observed during the 10 ns simulation, in good agreement with the experimental permeation rate.

Question: Can you identify specific regions in the pore that you expect to be particularly critical to make this channel a specific water channel?

Question: Would you expect other small molecules like glycerol to pass through as well? What about CO2?

Go back to Contents

Mechanism of proton exclusion

It is critically important that no protons leak across the membrane, as proton gradients are essential for bioenergetics, like for example in ATP synthesis. Therefore, it is crucial that protons do not pass through aquaporins, alongside with water molecules. This is a challenging task, as usually water, or water filled pores, conduct protons efficiently. How can aquaporins, therefore, accomplish the challenging task of being on the one hand an efficient water pore, whereas on the other hand, actively block protons from passing?

Therefore, let us carefully check the behaviour of water molecules in the pore to see if we can discover something about a possible mechanism of proton blocking.

Watch the last movie again and carefully check the behaviour of individual water molecules inside the pore carefully.

Question: Would you say that the water molecules are oriented randomly inside the pore? Hint

Note that water molecules carry an electrical dipole:

In an external electric field, water molecules will rotate to align with the external field, to optimise their interaction with the field.

Question: Look again at the behaviour of water molecules in the pore. How are the dipoles distributed? Could this be due to an electric field inside the protein? How can the dipole behaviour be explained in terms of interactions between the protein and passing water molecules? How will this behaviour affect the permeation of water? How will it affect the permeation of positively charged ions and protons? What about negatively charged ions?

Have a close look at the distribution of amino acids in the pore. In pymol, clicking on an atom gives you the residue number and type. Have a look at this table with amino acid types (best opened in a new window).

Question: Which amino acids do you think are responsible for the effect observed above?

In addition to the interactions with the individual amino acids, note that alpha helices carry a macro dipole , with the N-terminus carrying a positive charge and the C-terminus carrying a negative charge.

Question: Have a close look at the aquaporin structure again. In pymol, choose the "cartoons" representation. How do you think the two short helices, that end in the middle of the pore, could contribute to the electrostatic properties of the channel? Hint: look at the location of the N-termini of these two helices.


Go back to Contents

Optional

Here we're going to have a go at protein design. Have another look at this table with amino acid types.

Further references and advanced reading
For questions or feedback please contact Bert de Groot / bgroot@gwdg.de