Help

About SynLinker

SynLinker is an integrated system that provides linker sequences for constructing fusion proteins. It contains natural linkers from a non-redundant multi-domain PDB structures, as well as natural and artificial linkers in empirical recombinant fusion proteins collected from literature and patent information. For a demonstration, refer to the video below.



Framework

SynLinker takes a series of searching criteria and outputs linker candidates having the desired properties. A linker can be selected and used to connect two uploaded domain structures. A structure of the fusion protein will be generated by a backend program "fusedomains".


Searching SynLinker

A series of searching criteria serve as filters for SynLinker. The user may specify PDB ID, protein name, the number of residues in a linker, the end-to-end distance of a linker conformation in Angstroms, to search SynLinker. Addionally, the user can choose linker starting residue, ending residue, residue enrichment, residue depletion and sensitivity to protease as the desired linker properties. The query types can be used separately or together. After the searching criteria are input, SynLinker will output a list of linker candidates. The properties of the linkers are displayed in a table. The user can choose to modify the query types. To visualize a linker's conformation, click the "Show" button in its entry, which brings up a JSmol viewer. A few presentation styles are available to show the 3D conformation of the linker, whose PDB format can be downloaded at the bottom of the JSmol viewer.


BLAST SynLinker

The user is allowed to input a query linker sequence with a minimum length of five amino acids to BLAST against SynLinker. The BLAST result shows a list of linker sequences in SynLinker that are most similar to queried sequence. To visualize a linker's conformation, click the "Show" button in its entry, which brings up a JSmol viewer window.


Fusion Protein Structure Prediction Overview

After selecting a linker conformation, the user is allowed to upload two domain structures to predict a fusion protein structure via a backend program "fusedomains". The following figures show snapshots of using SynLinker to build a fusion protein "HAS-ANF" (human serum albumin-atrial natriuretic factor) from de Bold et al 2012. You can view the example output here.


Step 1     On the search page, linker length is set to be 6 residues and linker starting residue is Gly.
 
Step 2 A list of linker candidates with their accession ids, properties and conformations are shown. The artificial linker FGSR_6-1 (G6) is chosen.
 
Step 3 Navigate to the "Protein Fusion" interface. "FGSR_6-1 (G6)" is in the selection list. Click "Show" button to bring up a JSmol viewer displaying its conformation.
 
Step 4 Choose from a local directory to upload two domain structures in PDB format. Here, human serum albumin [PDB:1N5U] and atrial natriuretic factor [PDB:1YK0] are uploaded, which are appended at the N and C terminals of the fusion protein.
 
Step 5 Select a chain in the uploaded domain structures.
 
Step 6 Select the start and end residues in the uploaded domain structures to be fused together.
 
Step 7 The website server will call up a backend program "fusedomains" to fuse the domains and the selected linker.
 
Step 8 A possible conformation of the fusion protein "HAS-ANF" is generated and shown in the JSmol viewer.
 
Step 9 The hydropathicity profile of the fusion protein "HAS-ANF" is calculated and shown.

  Upload domain structures and PDB format

The user can upload two domain structures, whose PDB files must conform to the following requirements. Users are advised to provide structure files without missing residues. The PDB file must have a header line as follows:

HEADER    IMMUNE SYSTEM                           06-JUN-00   1F3R


The structure must be described by "ATOM" lines with the last column containing the element symbol as following:

ATOM      1  N   MET     1      28.516  21.281 101.835  1.00100.00           N


For the information about PDB format, please further refer to here.


  Select Chain

For an uploaded PDB structure containing multiple chains, the user is allowed to select a chain for usage. By default, the first chain of the first model is selected.


  Select Start and End Residue

After selecting the chain in the uploaded PDB structure, a user is allowed to specify a particular portion of the structure that corresponds to the interested domain. Be default, the entire selected chain is selected.


  Fusedomains

Once you have specified the two domain structures, they will be processed and joined by the selected linker conformation via a backend program "fusedomains". A perl script called "fusedomains" was developed for constructing a fusion protein. This program utilizes the BioPerl package. Given the solved atomic structure of each domain in the standard PDB format, a fusion protein model is created by recursively joining the first domain, a linker and the second domain by following the sub-functions: 1) coordinates translation, 2) clash checking, 3) coordinates rotation, 4) domain fusion and 5) energy minimization. In the rotation process, the first protein structure is kept rigid. All the coordinates of the second protein structure rotate together around the axis of Ca-C bond of the last residue in the first protein structure by an angle a, whose value is randomly generated.


The framework of "fusedomains" is shown below.


     

The user can wait for the system to finish generating the fusion protein. A possible conformation of the fusion protein predicted by "fusedomains" will be shown in another JSmol viewer within the same window/tab of the selected linker conformation, along with a structure download link, potential enegy and hydropathicity results. The running time depends on the size of the domain structures. If the domain structures are in large size, the user may provide an email address, to where a link to the output file will be sent when the results are ready.


  Predicted structure of the fusion protein

The predicted structure of the fusion protein can be visualized in parallel to the selected linker conformation. The two domains of the fusion proteins are colored in red and blue and the linker is colored in black. A hydropathicity profile of the fusion protein is also generated. The fusion protein sequence and hyropathicity values can be downloaded as well.


Our software will attempt to run GROMACS energy minimization, but if GROMACS cannot work on the structure (e.g. if there are missing residues), SynLinker will still provide the non-minimized fusion protein. In order to prepare GROMACS topology files, users are advised to provide structure files without missing, incomplete or invalid residues. Protein structures can be prepared using software like DeepView etc, in which waters, ions and ligands are suggested to be removed.


Clarifications

  For linker sequences that do not have 3D coordinates in PDB database, one of their extended conformations is generated ab initio using the software package TraDES (Trajected Directed Ensemble Sampling) by Feldman & Hogue 2002.
The predicted fusion protein structure can be validated via available software for example PROCHECK (Laskowski, et al., 1993) and used for further analysis based on the user’s discretion. Conformational sampling and domain orientations will be covered in future implementation. Molecular dynamics simulation can be applied for a better understanding of the fusion protein systems.

References
Chen, X., J.L. Zaro, and W.C. Shen, Fusion protein linkers: Property, design and functionality. Adv Drug Deliv Rev, 2012.
de Bold M.K., W.P. Sheffield, A. Martinuk , V. Bhakta, L. Eltringham-Smith, and A. J. de Bold. Characterization of a long-acting recombinant human serum albumin-atrial natriuretic factor (ANF) expressed in Pichia pastoris. Regulatory peptides 2012 175(1-3):7-10
Feldman, H.J. and C.W. Hogue, Probabilistic sampling of protein conformations: new hope for brute force? Proteins, 2002. 46(1): p. 8-23.
Laskowski, R.A., MacArthur M.W., Moss D.S., and Thornton J.M. PROCHECK - a program to check the stereochemical quality of protein structures. Journal of Applied Crystallography, 1993 26, 283-291.
© 2013-2019 SynCTI, National University of Singapore (NUS). All Rights Reserved.
SynLinker is free for academic and non-commercial use. For a commercial license please contact us.
The project is partly supported by the Next-Generation BioGreen 21 Program (SSAC, No. PJ01109405), RDA, Republic of Korea.
This page requires Chrome 32, Firefox 26, or Internet Explorer 9 (without compatibility mode).