Peptide Mode Tutorial

Overview:

In peptide mode, you will need to provide a receptor, a peptide motif and its corresponding peptide sequence. The input receptor should be prepared as described below (See Preparation of Receptor Structures). The selection of the motif is a critical step of our protocol (See Buildup of Motif). Once the motif has been designed and the set of fragments has been extracted, the amino acid sequence is changed back to the actual peptide sequence (using a backbone-dependent rotamer library).

Preparation of Receptor Structures:

Wobbly tails should be removed from the input receptor. The structure of the free receptor is represented as an independent binding unit that is defined as either a single domain, or repeated, non-decomposable domains. Should a protein consist of repeated domains (of the same class), it should be treated as a single unit. If the protein instead consists of different domains, they should be split. Multi-domain proteins may have numerous peptide regulatory sites per domain. For this reason, only the selected domain should be used for docking. An example of domain splitting is shown below (PDB ID: 1EG3). The structure consists of three different domains. Shown in yellow is the single domain used for docking.

Buildup of Motif:

  1. Start. Begin with the peptide sequence of interest and a known, reported motif (e.g. as defined in the ELM database). To allow for a motif of 4 and more residues and one or more wild cards if necessary, start with a peptide sequence of minimal length of 5 residues. If needed, extend the reported motif by including additional positions from the peptide sequence. The preferential direction of extension is defined based on the type of residues, according to the following priority: Polar, Aromatic, and other residues. Small amino acids (GSTA) are not considered for extension, except as a bridge to the next extended residue. Perform a preliminary fragment search using the Advanced Search function in the PDB (select Sequence Motif Query Type). This initial sequence will usually result in very few fragment hits in the PDB.
  2. Expand. Insert wildcards back (X, or redundant positions of the motif), starting with the smallest residues. Refrain from introducing adjacent wild cards if possible, and do not introduce X at the termini of the peptide (which would be equivalent to shortening the motif). Each time, return to the PDB Sequence Motif search and assess the number of resulting hits.
  3. Large. If more than 1000 hits to PDB structures are found, introduce specific residues back into the motif, starting with the largest residues. If this does not help, try to extend into the same direction, if possible.
  4. Stop. When there are between 100 and 1000 hits to PDB structures (or more if further extension of motif is not possible)
  5. Complement F/Y. F & Y show very similar conformational preferences in the backbone dependent rotamer libraries.

Sample motif selection for the peptide HTLKGRRLVFDN with the reported motif RXL.

  1. The reported motif is only three residues, so we initially extend the peptide motif corresponding to the peptide sequence in the n-terminal direction towards a polar residue, to a pentamer KGRRL. Using the PDB Sequence Motif search we identify the number of motif hits in the PDB.
  2. Since this motif is not found frequently enough in the PDB (9x), we make it more general by introducing a wild-card at the smallest residue, G (KXRRL). This motif is found frequently enough to proceed (in 485 PDB structures, homologs of 2CCH excluded).

The peptide motif can also be determined by using the 'Build Motif' button. To use, enter full peptide sequence, e.g. HTLKGRRLVFDN and the reported motif, e.g. RXL. After clicking the 'Build Motif' button, the peptide and motif fields will be generated.

Fragment Upload:

Instead of inputing a motif to utilize our motif-based fragment search, users can upload their own fragment set of peptide structures. Fragment sets should contain no more than 25 pdb files. We recommend that each peptide structure be between 5 and 8 residues long. All pdb files should have the same number of atoms. Fragment sets can be uploaded as a zip or tar file.

Processing Result Structures:

After docking, resulting clusters are ranked according to cluster size. In cases where the receptor structure was extracted from a multi-domain protein, solutions that overlap with domain-domain interfaces should be removed. Shown below are predictions for peptide placement on a single domain of a receptor (yellow). The full multi-domain protein (purple) is aligned to the receptor, and results evaluated for domain interference. The predictions which interfere with the other domains are shown in red, and would be disregarded as potential binding sites for the peptide. Valid predictions are displayed in green.

Benchmark Examples:

Our lab has developed this protocol using the PeptiDB dataset and the following sequences, motifs, and exclusion lists. Click on the receptor name to download the pdb file. Docking results may be viewed by clicking the corresponding job id. It should be noted that the linked results will contain predictions that have not been processed for domain interference.

Receptor/Chain Motif Sequence Exclusion List Results
1EG3/A_1 RXPPX[YF] RSPPPY 1EG4 View Results
1D1Z/A TI[YF]XX[VI] TIYAQV 1D4T 1D4W 1I3Z 1M27 View Results
1OOT/A PXMPXR PAMPAR 1SSH 1Q2V 2B3H 2B3K 2B3L 2BER 2BPA 2BZD 2G6P 2GZ5 2J2F 2NQ6 2NQ7 2X6U 2X6V 2XPY 2XPZ 2XQ0 2XZ0 2XZ1 3T2B 3T2C 3T2D 3T2E 3T2F 3T2G View Results
2H3L/A LDVXV LDVPV 1MFG 1MFL View Results
2H14/A ARTKQ ARTKQ 2CNX 2CO0 2G99 2G9A 2H13 2H6K 2H6N 2H6Q 2H9M 2H9N 2H9P 2O9K 2XL3 3PSL 4A7J View Results
2F1W/A R[PA]HXS RAHSS 2F1X 2FOJ View Results
2HPJ/A DXL[YF]G DDLYG 2HPL View Results
1CA4/A PXQXXDD PQQATDD 1CZY 1ZMS View Results
1JD4/B A[VTI][AP][YF][YF] AIAYF 1JD5 1JD6 View Results
1B9K/A_1 WXX[FY]E WVTFE 1W80 2VJ0 View Results
1B9K/A_2 [FY]XDN[FY] FEDNF 1KY7 2VJ0 View Results
2J2I/B_2 RXRHXS RRRHPS 2C3I 3CXW 3CY2 3CY3 4GW8 View Results
1H1R/B KXRRL KGRRL 2CCH 2CCI 2J90 3BHY 3BQR View Results
1RWZ/A QXX[LVI]XXW[FY] QATLERWF 1RXZ 2IZO View Results
4APE/A H[LVI][LVI][LVI][YF] HLLVY 1ER8 View Results
1JWF/A DXDLL DEDLL 1JPL 1JWG 1LF8 3G2U 3G2V View Results

A pymol session containing the best results for each of the benchmark cases can be downloaded here: Download Benchmark Results

If you have any questions, please look to see if it is addressed on the help page. If you have any suggestions, please contact us.

ClusPro should only be used for noncommercial purposes.
Vajda Lab and ABC Group
Boston University and Stony Brook University