Dear Pengfei,
Thank you for your response - and sorry for my lack of clarity!
I have a protein that has four HIS residues, and thus can bind up to
four molecules of a Zn-containing ligand. The binding sites are all
equivalent. I performed the parametrization using MCPB.py with one
ligand molecule bonded to a HIS. Everything worked fine. Now, when I
wanted to study the system with two ligands, I did not repeat the whole
MCPB procedure. Instead, I modified the PDB file to have a second ligand
molecule, with the same residue name and atom types as the first ligand,
so they take their parameters from the same mol2/frcmod files. Then I
used LEaP to build the prmtop and inpcrd, using the script that resulted
from MCPB but slightly modified to account for the Zn bonds in the
second ligand molecule (please note that the protein is ignored in the
following, since it does not affect the problem):
source leaprc.gaff
addAtomTypes {
{ "M1" "Zn" "sp3" }
{ "Y2" "N" "sp3" }
{ "Y3" "N" "sp3" }
{ "Y4" "N" "sp3" }
{ "Y5" "N" "sp3" }
}
ZN1 = loadmol2 ZN1.mol2
PZ1 = loadmol2 PZ1.mol2
loadamberparams PPZ.frcmod
loadamberparams complex_mcpbpy.frcmod
mol = loadpdb complex.pdb
bond mol.1.ZN mol.2.N
bond mol.1.ZN mol.2.N1
bond mol.1.ZN mol.2.N2
bond mol.1.ZN mol.2.N3
bond mol.3.ZN mol.4.N
bond mol.3.ZN mol.4.N1
bond mol.3.ZN mol.4.N2
bond mol.3.ZN mol.4.N3
saveamberparm mol complex.prmtop complex.inpcrd
This /apparently /worked fine, but on closer inspection there are more
dihedral parameters involving the Zn atom of the second ligand, than for
the Zn atom of the first ligand. By splitting the prmtop in two (one for
each ligand), I can use cpptraj's comparetop to see the extra dihedrals
in the second ligand:
# Dihedrals
< :1.ZN - :2.N - :2.C2 - :2.C {M1 -Y2 -ce -c3 } Pk=0.8 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
< :1.ZN - :2.N - :2.C2 - :2.C3 {M1 -Y2 -ce -cf } Pk=0.8 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
< :1.ZN - :2.N1 - :2.C15 - :2.C13 {M1 -Y3 -cc -cc } Pk=4.75 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
< :1.ZN - :2.N1 - :2.C15 - :2.C19 {M1 -Y3 -cc -cf } Pk=4.75 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
< :1.ZN - :2.N1 - :2.C8 - :2.C3 {M1 -Y3 -cd -cf } Pk=4.75 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
< :1.ZN - :2.N1 - :2.C8 - :2.C9 {M1 -Y3 -cd -cd } Pk=4.75 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
< :1.ZN - :2.N2 - :2.C16 - :2.C12 {M1 -Y4 -cc -cf } Pk=4.75 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
< :1.ZN - :2.N2 - :2.C16 - :2.C20 {M1 -Y4 -cc -cc } Pk=4.75 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
< :1.ZN - :2.N2 - :2.C23 - :2.C21 {M1 -Y4 -cd -cd } Pk=4.75 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
< :1.ZN - :2.N2 - :2.C23 - :2.C28 {M1 -Y4 -cd -cf } Pk=4.75 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
< :1.ZN - :2.N3 - :2.C22 - :2.C19 {M1 -Y5 -cd -cf } Pk=4.75 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
< :1.ZN - :2.N3 - :2.C22 - :2.C24 {M1 -Y5 -cd -cd } Pk=4.75 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
< :1.ZN - :2.N3 - :2.C26 - :2.C27 {M1 -Y5 -cc -cc } Pk=4.75 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
< :1.ZN - :2.N3 - :2.C26 - :2.C28 {M1 -Y5 -cc -cf } Pk=4.75 Pn=2 Phase=3.14159 SCEE=1.2 SCNB=2
# Dihedral Parameters
To overcome the problem, what I did is:
- I modified the PDB file to give the two ligands different names
(ZN1/PZ1 for the first ligand, ZN2/PZ2 for the second).
- I copied and renamed the files generated by MCPB. In other words, I
took "ZN1.mol2", "PZ1.mol2", and "complex_mcpbpy.frcmod" and copied them
as "ZN2.mol2", "PZ2.mol2", and "complex_mcpbpy2.frcmod"
- To differentiate atom types, I renamed the MCPB-generated "M1" and
"Y1-Y5" for the second ligand as "M2" and "I1-I5".
- I modified the LEaP input script accordingly, resulting in:
source leaprc.gaff
addAtomTypes {
{ "M1" "Zn" "sp3" }
{ "Y1" "N" "sp3" }
{ "Y2" "N" "sp3" }
{ "Y3" "N" "sp3" }
{ "Y4" "N" "sp3" }
{ "Y5" "N" "sp3" }
{ "M2" "Zn" "sp3" }
{ "I1" "N" "sp3" }
{ "I2" "N" "sp3" }
{ "I3" "N" "sp3" }
{ "I4" "N" "sp3" }
{ "I5" "N" "sp3" }
}
ZN1 = loadmol2 ZN1.mol2
PZ1 = loadmol2 PZ1.mol2
ZN2 = loadmol2 ZN2.mol2
PZ2 = loadmol2 PZ2.mol2
loadamberparams PPZ.frcmod
loadamberparams complex_mcpbpy.frcmod
loadamberparams complex_mcpbpy2.frcmod
mol = loadpdb complex.pdb
bond mol.1.ZN mol.2.N
bond mol.1.ZN mol.2.N1
bond mol.1.ZN mol.2.N2
bond mol.1.ZN mol.2.N3
bond mol.3.ZN mol.4.N
bond mol.3.ZN mol.4.N1
bond mol.3.ZN mol.4.N2
bond mol.3.ZN mol.4.N3
saveamberparm mol complex.prmtop complex.inpcrd
Now the two Zn atoms have exactly the same number of dihedral
parameters. So, in a way, I was able to solve the problem. The thing is,
I have already ran several microseconds of MDs with the problematic
prmtops, that's why I wanted to hear the list's opinion before throwing
them away...
I hope this clarifies the problem. If not, I will be happy to provide
any necessary file or extra information.
Thank you for your help, and best regards,
Mariano
On 22/10/2020 23:19, Pengfei Li wrote:
> Hi Mariano,
>
> Can you specify more details about "treating the two ligands as different” and what is its difference from the previous settings you chose?
>
> Pengfei
>
>> On Sep 18, 2020, at 10:41 AM, Mariano Curti <mcurti.iciq.es> wrote:
>>
>> Dear all,
>>
>> I've found unexpected behavior from LEaP (Amber 18), thus would be
>> grateful if you help me determine whether this is a mistake from my
>> part, or a bug in the program.
>>
>> I'm using MCPB.py to build parameters for a metal-containing ligand that
>> is covalently attached to a protein. When using only one of these
>> ligands, everything works well. However, when I include two ligands,
>> while LEaP builds the dihedral parameters list correctly for the first
>> ligand, it (wrongly) includes additional dihedrals for the second ligand.
>>
>> A reduced tLEaP input that reproduces this behavior is as follows. Note:
>> I have derived the metal-related parameters using MCPB.py for one
>> ligand, but the second is identical, so I use the same set. In this
>> example the two first residues are the first ligand, while residues 3
>> and 4 are the second ligand. I omitted the protein since it does not
>> affect this problem.
>>
>>
>> source leaprc.gaff
>> addAtomTypes {
>> { "M1" "Zn" "sp3" }
>> { "Y2" "N" "sp3" }
>> { "Y3" "N" "sp3" }
>> { "Y4" "N" "sp3" }
>> { "Y5" "N" "sp3" }
>> }
>> ZN1 = loadmol2 ZN1.mol2
>> PZ1 = loadmol2 PZ1.mol2
>> loadamberparams PPZ.frcmod
>> loadamberparams complex_mcpbpy.frcmod
>> mol = loadpdb complex.pdb
>> bond mol.1.ZN mol.2.N
>> bond mol.1.ZN mol.2.N1
>> bond mol.1.ZN mol.2.N2
>> bond mol.1.ZN mol.2.N3
>> bond mol.3.ZN mol.4.N
>> bond mol.3.ZN mol.4.N1
>> bond mol.3.ZN mol.4.N2
>> bond mol.3.ZN mol.4.N3
>> saveamberparm mol complex.prmtop complex.inpcrd
>>
>>
>> When looking at the resulting prmtop, I notice that there are extra
>> dihedral parameters (with non-zero force constants) involving the metal
>> of the second ligand. At the end of the day, this seems to result in
>> weird MMGBSA results.
>>
>> Now, if I follow the same steps, but this time treating the two ligands
>> as different (i.e. renaming the residues so LEaP thinks they are
>> different), I get the expected prmtop, with matching dihedral parameters
>> for both ligands.
>>
>> My question is, then: am I doing something wrong, or this is a bug in
>> LEaP? I would be happy to provide any necessary additional information
>> or files.
>>
>> Many thanks in advance!
>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
--
Dr. Mariano Curti
Postdoctoral researcher - Romero Group
Institute of Chemical Research of Catalonia (ICIQ)
Av. Països Catalans 16 – 43007 Tarragona (Spain)
mcurti.iciq.es
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Oct 26 2020 - 02:30:03 PDT
