Dear MODELLER Users,
I would like to model a short nucleotide sequence as a substrate for an enzyme. In the top_heav.lib file, I notice that the listing for URA (a uridine mono-phosphate molecule) has bond angle entries that include '+' and '-' prefixes on some of the atoms (referring to atoms in molecules adjacent in a chain sequence). I was under the impression that these entries were enough to form a connected sequence of molecules (in this case, an oligonucleotide). In my alignment file, I have two of these URA molecules situated next to each other (separated from the amino acids of my enzyme by a chain break character '/'), but in the final model, they do not appear to be bonded to each other. As a test, I substituted two amino acids for the two URA molecules and noticed that in the final model, they were not joined by a bond either. Since the amino acids of my enzyme were all clearly bonded to each other, this result made me wonder if MODELLER was not recognizing my two substrate molecules as being members of a chain themselves (albeit two subunits long).
- Do I need to, or can I, explicitly define these two (nucleotide) molecules as being members of a chain such that MODELLER will create a phosphodiester bond between them?
- Is it possible that they would normally be bonded, but during the simulation, the molecules are being "torn apart"?
- My other option appears to be to impose a "special_patches" restraint (that I would have to add to the top_heav.lib file) to explicitly join them. However, seeing that the URA residue listing already appears to be set up for nucleotide chains, this seems like it would be unnecessary.
Thanks for your time, Sincerely, -Andy Torelli
Ph.D. Student University of Rochester, Department of Biochemistry and Biophysics