MD simulations

This protocol was used to generate structural models of the protein SRSF1 bound to the RS8 peptide. The final model was refined using the FF19SB forcefield in AMBER20.

The XPLOR-NIH scripts used in this protocol have been uploaded to our github page (https://github.com/taliafargason/XPLOR_peptides_step1 and https://github.com/taliafargason/XPLOR_peptides_step2). To use the script on a different system, it is necessary to format the starting structure as mentioned below and format NMR data following the examples on the github page. It may also be wise to contact the developers of XPLOR NIH, who helped us write custom scripts for our system.

We used the Xplor-NIH 3.4 installment (which can be downloaded from https://nmr.cit.nih.gov/xplor-nih/download.cgi). 

The overall procedure involved three major steps:

1. Generate a starting structure
2. Generate an ensemble of peptide-bound structures
3. Sort structures according to best fits

List of input files:

A. PDB starting structure files:

• Step1: 220C_only.pdb
  • The initial starting structure was downloaded from the Alphafold database as AF-Q07955-F1-mod.pdb.
  • Mutagenesis was performed in PyMol to convert native cysteines to serines and insert the cysteine labeling site (N220C).
  • To allow XPLOR to place the MTSL label correctly, the name of the cysteine residue was modified from “CYS” to “CYSP” by opening the pdb file as a text file and manually changing the name of the residue, making sure to keep the columns intact.
• Step2: best_single_struct+RS8b.pdb
  • The best fit single structure was obtained from step 1
  • RS tail (197-248) was removed in PyMol
  • A single peptide (sequence SRSRSRSRC) was built in place of the tail using the builder tool in PyMol
  • To allow XPLOR to place the MTSL label correctly, the name of the cysteine residue in the peptide was modified from “CYS” to “CYSP” by opening the pdb file as a text file and manually changing the name of the residue, making sure to keep the columns intact.

B. TBL files with NMR restraints:

(NMR data was manually converted into the tbl format by exporting it from an excel spreadsheet and changing the file extension.)

• Step1 inputs
  • 220CPRE_Data.tbl (intramolecular PRE data)
  • CSP_high_confidence_simplified.tbl
    • This data table contains artificial NOE values:
      • CSP values were used to assign ambiguous H-bonding restraints (within 4Å).
      • Bleached PRE values were also incorporated as ambiguous restraints (within 12 Å).
  • RDC_unphos_sup_conf_struct.tbl (RDC values)
  • talosn.tbl
    • Dihedral angle restraints
    • STAR file from NMR assignment was submitted to the Talos online server (https://spin.niddk.nih.gov/bax/nmrserver/talos/) to obtain prediction files (pred.tab or predall.tab)
    • The XPLOR command “convertTalos” was used to convert this into an XPLOR- friendly format
• Step2 inputs:
  • RS8PRE.tbl (PRE data in which the MTSL tag is placed on the peptide)
  • RDC_unphos_sup_conf_struct.tbl (RDC values)
  • CSPdata.tbl
    • This data table contains artificial NOE values:
      • CSP values were used to assign ambiguous H-bonding restraints (within 4Å).
      • Bleached PRE values were also incorporated as ambiguous restraints (within 12 Å).
  • talosn.tbl
    • Dihedral angle restraints
    • STAR file from NMR assignment was submitted to the Talos online server (https://spin.niddk.nih.gov/bax/nmrserver/talos/) to obtain prediction files (pred.tab or predall.tab)
    • The XPLOR command “convertTalos” was used to convert this into an XPLOR- friendly format

Procedure:

Call scripts for steps 1 and 2:

Note that the script is the same for both steps, but the inputs have been modified.

After installing Xplor-NIH, the following command can be used to run the script:

XPLOR -py RefineF.py

To also designate an output file:

xplor -py RefineF.py -o outB.txt

Generate a list of fitting parameters:

The above will generate 100 structures.  On a linux operating system the following command can be used to obtain a txt file with the PRE fitting for each structure:

grep -A 1 Q-fac * >Q.out

Structures were sorted according to Q-factors using an Excel spreadsheet.