AMBER Archive (2004)

Subject: RE: AMBER: Molecular dynamic

From: Ross Walker (ross_at_rosswalker.co.uk)
Date: Mon Jul 12 2004 - 13:12:53 CDT

Dear Bo,

> dynamic. Before doing so, I tested the running condition with native
> structure to see if there are some changes. The protein is made up of
> standard amino acids. There was no problem for leap to load its .pdb
> file. The native protein contains 4 cycteine, but no S-S bond
> formed in its 3D-structure.

Ok, but what is the distance between the sulphur atoms of the various
cysteines? Could there be a bond here in reality? What do the
crystallographers suggest? Is there any speculation about the cysteine
protonation states in the pdb header?

> The template for the modeling contains 3
> 'MSE' and 1 'TAR'. MSE refers Selenomethionine, and TAR refers to
> tartaric acid. I don't know how these molecules got into the crystal
structure. And the
> auther did not mention anything about these unuausl molecules
> with their crystal structure.

Hmmm, this could be something to do with inhibiting the system to stop it
reacting, or also as a way of encouraging crystallisation. Note such
residues may well have an effect on the crystal structure and make it
unrepresentative of the solution structure. Your best bet though is just to
try out some simulations and see what happens. You definitely want to avoid
having to simulate a selenium compound if you can help it. It is very
strange that the authors make no mention of it, do they refer to it in the
pdb header?

> As Amber can't recognize these molecules.

Indeed, for the moment I assume you just removed them from the pdb...

> I did not include these moleculars with my model. The model
> went through
> energy minimization. The unfolding only occurred when I tried
> to run it

Very little structural change will occur during minisation since the amber
minimiser will only move down hill to the nearest 'local' minimium. As such
the minimisation will not move you far from the initial structure, it is
simply designed to remove the highest strains caused by steric clashes from
hydrogenation etc. Really it is just a precursor to avoid large initial
forces in your MD causing your system to blow up.

> with MD. It may worth to put them back to my model structure.

You may need to do this if they are essential to the structure. I assume the
protein itself though exists in solution without these compounds? What pH is
it typically catalytic at? Does it need Selenium (I doubt it) or some other
metal to either stabilise it's structure or for catalytic purposes? You
should look into this before going much further. Often crystallographers may
substitute a metal in a protein for a different one in order to make
crystallisation easier.

> When I solvated and add ions to the structure, the program
> automatically search the places to put the ions. And I have to depended
the program
> does the job properly.

This sould be fine, the important thing is that you are adding the correct
number of counter ions. The locations chosen by Leap are normally
sufficient. The problem that can sometimes occur is if you don't have an
exact integer charge due to rounding. E.g +1.9999. If you tell Leap to
neutralise this it will only add 1 ion. Thus it is often better to manually
specify the number of ions you want in the add ions command rather than just
ask Leap to neutralise.

> With all the .pdb file after the energy
> minimization, HIS residues have been changed to HID.

This is probably reasonable but if you can find out the correct protonation
states at neutral pH that would be better.

> The following is how the system solvated and neutralized:
>
> > solvateBox test WATBOX216 10
> Solute vdw bounding box: 50.000 48.604 54.035
> Total bounding box for atom centers: 70.300 68.604 74.035
> Solvent unit box: 18.774 18.774 18.774
> Total vdw box size: 73.553 71.808 77.076
> angstroms.
> Volume: 407096.264 A^3
> Total mass 197457.906 amu, Density 0.805 g/cc
> Added 9629 residues.
> > charge test
> Total unperturbed charge: -2.00
> Total perturbed charge: -2.00
> > addIons2 test Na+ 0
> 2 Na+ ions required to neutralize.
> Adding 2 counter ions to "test" using 1A grid
> Grid extends from solute vdw + 1.87 to 7.87
> Resolution: 1.00 Angstrom.
> grid build: 7 sec
> Calculating grid charges
> charges: 683 sec
> Placed Na+ in test at (-10.88, -2.12, 15.09).
> Placed Na+ in test at (-1.88, -5.12, 3.09).

Seems fine to me.

> For my understanding now, I should only allow the target
> regions to move during a MD run, while the rest should be in a fixed
> condition. Is this correct?

I'm note sure what you mean by this. Are you running targeted MD or just a
regular MD simulation of your system? Really you want to allow everything to
move unless you have any real reason for stopping things moving. I suspect
that if you are using restraints it is these restraints causing your system
to unfold rather than any real stability issues in the protein...

I would minsimise your whole system with ntb=1 and then run about 20ps of
heating fro 0 to 300K with ntb=1 (constant volume) allowing EVERYTHING to
move and then switch to constant temp 300K and ntb=2 (constant pressure) for
the remainder of your simulation. Note if you have any bad initial contacts
your system may blow up which can sometimes look (since bonds cannot be
broken in classical md) like the system unfolding. Best to take a look at
the first 500fs or so of you sim (ntwx=1) and see if anything rapidly moves
out of position. Also set vlimit to something low like 10 and see if it gets
exceeded at all. If it does then you may need a longer minimisation and
slower heating etc...

> To put the whole system in a MD run might generate faulse
> results.

Restraining parts of your system are more likely to generate false results..

> As a beginner for Amber program, I find that this mail system
> is great!.
> Thanks for providing such a place! Thank you very much for reading my
> question and giving me advices.

You are welcome. Have you tried out the tutorials on the Amber website as
well? amber.scripps.edu/tutorial/ these give examples of how to run
calculations..
 
All the best
Ross

/\
\/
|\oss Walker

| Department of Molecular Biology TPC15 |
| The Scripps Research Institute |
| Tel:- +1 858 784 8889 | EMail:- ross_at_rosswalker.co.uk |
| http://www.rosswalker.co.uk/ | PGP Key available on request |

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu