Modeling amino acid insertions

Here, we predict a structural model for the structurally unresolved amino acid residues 162 to 177 of chain B of 4OCM. Create the folder domain_insertion for the files generated in this section.

Structure prediction with Rosetta

For the domain insertion folding, you can take the correctly numbered full length model that was generated in section 3.1 and continue. Again, create a new folder named insertion, create a folder named full_length_model in it and copy the already generated full length PDB to it.

1

Prediction of an ensemble of possible structures:
The recommendation from the literature is to predict between 5,000 and 20,000 structures. In our test case we predict only 100 structures for demonstration purpose. We use the topology Broker as direct input to the classic Rosetta de novo protocol, wrapped by the ModelMaker package, to predict a possible ensemble of structural models. As in section 3.1, we create a configuration file fold_insertion_rpn11.tcl. Copy your configuration file with the already adapted variables to the working directory. To perform a de-novo structure prediction of the missing domain, we only need to change the analysis tasks and the commands to run Rosetta:

• line 22: set comps [list [list ss 138 157 "A"] [list cluster 138 157 "A" 2]]
  In this case, we want to analyze the outcoming secondary structure from residue 138 to 157 and cluster the best 25 structures with the Rosetta cluster application. The cluster analysis task is configured by the second list in $comps, where the single elements stand for: [list cluster $\langle$start residue ID$\rangle$ $\langle$end residue ID$\rangle$ $\langle$chain$\rangle$
  $\langle$max. cluster number$\rangle$]
• line 24: start_rosetta_insertion ...
  As in section 3.1, we call a ModelMaker function to run Rosetta with the given arguments (Tab. 4).
• line 26: analyze_abinitio ...
  The analysis arguments are to be obtained from Tab. 3. In this case, Rosetta renames the outcoming structure files, so that we need to append the original name of the template PDB file at the end and append _S to the full length name. The whole line should now be:

  analyze_abinitio "rpn11_insertion" "rpn11_yeast_23-306_complete_S"
  $bestN $nstruct 0 "resid 1 to 137 or resid 158 to 284"
  "resid 1 to 137 or resid 158 to 284" $comps "rpn11_yeast_23-306_complete"

Execute the configuration file in VMD text mode and wait for the tasks to finish. Depending on the number of structures to generate, this may take a while.

vmd -dispdev text -e fold_rpn11_insertion.tcl

If no error occurs, go to the folder called rosetta_output_rpn11_insertion containing the results of your run.

Table 4: Rosetta domain insertion procedure arguments

Interactive fitting to a cryo-EM density with iMDFF

Follow the steps explained in section 3.2.

1

Aligning the predicted model with the cryo-EM density map: See step 1 in in section 3.2.

2

Generating a density map file for MDFF:
See step 2 in in section 3.2.

3

Crop the density map:
See step 3 in in section 3.2.

4

Fitting the modeled part to the cryo EM density: See step 4 in in section 3.2.