Link two proteins in Pymol

Question

[Hansen_869]

I have two pdb-files containing two different protein structures. I need to combine these structures in Pymol. with a linker-sequence combining them. The linker I will need to create myself in Pymol. I also need to make some signal-peptides in the ends of each of the proteins. The linker and the signal-peptides are not in a pdb-file, as these will just be some random small sequences (the goal is to just come up with a reasonable illustration, it doesn't have to be exact).

How do I get both PDB-files into the same project in Pymol and how do I arrange them? And then, how do I create these custom sequences? Thanks!

Answer 1

[Mensur Dlakic]

Pymol is a visualization program. Even though it has some rudimentary superposition and modeling capabilities, it is not meant for any of the things you want.

I suggest you try this with Modeller or using SwissModel if you prefer web interface. First you extract the sequences from the two PDB files and make a chimeric sequence where they are connected by a linker of your choice. Next, align the first structure to its sequence, and the second to its own, and use both as templates for modeling. This will essentially copy the coordinates from the two structures and do free modeling for the linker. It may not be the most accurate way, but it will be a reasonable illustration.

Template 1 sequence:
SEQV
Linker sequence:
LINKER
Template 2 sequence:
ENCE

Alignment:
SEQVLINKERENCE   #This is target
SEQV----------   #This is template 1
----------ENCE   #This is template 2

[Hansen_869]

Thanks for your response, I think I will check out Swiss-model for this. What exactly will aligning the PDB structures to their own sequences accomplish? Or am I missing something? As I understand it, I make a fasta-file with "SEQVLINKERENCE" and then align both PDB files to this fasta-file?

[Mensur Dlakic]

Yes, that's the idea. I see now that SwissModel may not be able to use multiple templates, so Modeller may be your only choice. Aligning the sequences to their own structure essentially directs the program to copy the coordinates from those regions and do light minimization. Since there will be no template for the linker and the alignment above dictates that those two structures need to be connected by it, the program will translate coordinates of those two structures in such a way that they are close enough for the linker to connect them. There will still be very many possible conformations depending on linker length and the modeling will not be realistic in terms of exhaustive sampling, but you said the illustration is what matters here.

[Hansen_869]

Alright. Do you know if Modeller is available for free anywhere? I'm not in a position where I can buy it at this moment, unfortunately. I see they have an academic version, but I'm not currently a student.