Issues with incorporating ligands in multichain protein
Hi there, I am trying to model a multichain protein (homodimer) using 2 templates. I have two issues, Currently my .pir file looks like this >P1;seq1 sequence:seq1:1:A:1224:A:::: xxxx/ xxxx* >P1;template1.pdb structureX:template1.pdb:40:A:651:B:41:B:648:xxxx/xxxx* >P1;template2.pdb structureX:template2:3:B:291 :B::Thermus thermophilus HB8:1.45:xxxx/-------/ >P1;template2.pdb structureX:template2:3:B:291 :B::Thermus thermophilus HB8:1.45:-------/xxxx*
Firstly, my template 1 I am using both chain A and chain B, they have slightly different missing residues in the template but thats fine.For template 2 I cant put them under the same header, modeller doesn't seem to understand that I want to use chain B of template 2 for both chain A and B of my target so I specified it separately - it seems to work fine though with this setup. Unfortunately when I try add HETATM information ".." at the end of my target and template 1 and 2 gaps "--" for template 2 modeller says the whole sequence is out of alignment suddenly. I have specific HETATM=true in my script so it reads that part fine. My alignment looks like this but does not work? Am I missing some obvious?
>P1;seq1 sequence:seq1:1:A:1224:A:::: xxxx../ xxxx..* >P1;template1.pdb structureX:template1.pdb:40:A:651:B:41:B:648:xxxx../xxxx..* >P1;template2.pdb structureX:template2:3:B:291 :B::Thermus thermophilus HB8:1.45:xxxx--/---------/ >P1;template2.pdb structureX:template2:3:B:291 :B::Thermus thermophilus HB8:1.45:---------/xxxx--* RegardsKyle
On 6/14/24 4:57 AM, Kyle- richard Dawson via modeller_usage wrote: > Firstly, my template 1 I am using both chain A and chain B, they have > slightly different missing residues in the template but thats fine. > For template 2 I cant put them under the same header, modeller doesn't > seem to understand that I want to use chain B of template 2 for both > chain A and B of my target so I specified it separately - it seems to > work fine though with this setup.
You are almost right here. To use the same template twice for two different chains you need to include it twice in the alignment (as you have done) but the two entries need to have different codes (the text after P1;), e.g.
>P1;template2_A structureX:template2.pdb:3:B:291 :B::Thermus thermophilus HB8:1.45: xxxx/ ----* >P1;template2_B structureX:template2.pdb:3:B:291 :B::Thermus thermophilus HB8:1.45: ----/ xxxx*
Then in your Python script you would use knowns = ['template1', 'template2_A', 'template2_B']
If you just duplicate the template without changing the code, one of them will just be ignored.
> Unfortunately when I try add HETATM information ".." at the end of my > target and template 1 and 2 gaps "--" for template 2 modeller says the > whole sequence is out of alignment suddenly. I have specific HETATM=true > in my script so it reads that part fine.
Nobody can tell for sure without seeing the real alignment file but it looks like you have the range specified incorrectly here:
>>P1;template1.pdb > structureX:template1.pdb:40:A:651:B:41:B:648: > xxxx../ > xxxx..*
You can only specify a single range, not "40:A:651:B:41:B:648". See a recent discussion for more: https://salilab.org/mm/hyperkitty/list/modeller_usage@salilab.org/thread/ACZ...
Ben Webb, Modeller Caretaker
Thanks,
Seems to be working well now, the issue was related to the order of my cofactors within my PDB. Lastly, I would just like to know about how modeller tackles the protein interface. Is it able to model the subunit interactions if I only use duplicate chain templates? Since my protein is a homodimer and the subunits are identical, I was hoping to use just the best chain of each template but I worry that modeller might not be able to discern the subunit interactions if I do so. RegardsKyle
On Saturday, June 15, 2024 at 04:52:54 AM GMT+2, Modeller Caretaker via modeller_usage modeller_usage@salilab.org wrote:
On 6/14/24 4:57 AM, Kyle- richard Dawson via modeller_usage wrote: > Firstly, my template 1 I am using both chain A and chain B, they have > slightly different missing residues in the template but thats fine. > For template 2 I cant put them under the same header, modeller doesn't > seem to understand that I want to use chain B of template 2 for both > chain A and B of my target so I specified it separately - it seems to > work fine though with this setup.
You are almost right here. To use the same template twice for two different chains you need to include it twice in the alignment (as you have done) but the two entries need to have different codes (the text after P1;), e.g.
>P1;template2_A structureX:template2.pdb:3:B:291 :B::Thermus thermophilus HB8:1.45: xxxx/ ----* >P1;template2_B structureX:template2.pdb:3:B:291 :B::Thermus thermophilus HB8:1.45: ----/ xxxx*
Then in your Python script you would use knowns = ['template1', 'template2_A', 'template2_B']
If you just duplicate the template without changing the code, one of them will just be ignored.
> Unfortunately when I try add HETATM information ".." at the end of my > target and template 1 and 2 gaps "--" for template 2 modeller says the > whole sequence is out of alignment suddenly. I have specific HETATM=true > in my script so it reads that part fine.
Nobody can tell for sure without seeing the real alignment file but it looks like you have the range specified incorrectly here:
>>P1;template1.pdb > structureX:template1.pdb:40:A:651:B:41:B:648: > xxxx../ > xxxx..*
You can only specify a single range, not "40:A:651:B:41:B:648". See a recent discussion for more: https://salilab.org/mm/hyperkitty/list/modeller_usage@salilab.org/thread/ACZ...
Ben Webb, Modeller Caretaker
On 6/17/24 12:31 AM, Kyle- richard Dawson wrote: > Lastly, I would just like to know about how > modeller tackles the protein interface. Is it able to model the subunit > interactions if I only use duplicate chain templates? Since my protein > is a homodimer and the subunits are identical, I was hoping to use just > the best chain of each template but I worry that modeller might not be > able to discern the subunit interactions if I do so.
Short answer: no, not the way you're doing it. Modeller extracts distances from each template. If you only give it monomeric templates, it has no information about the interface between the two chains. You'll need to provide at least one template which spans the interface, or provide other distance restraints. See also https://salilab.org/modeller/10.4/manual/node29.html
Ben Webb, Modeller Caretaker
participants (2)
-
Kyle- richard Dawson -
Modeller Caretaker