As a part of our construct it is necessary to build linkers to connect the intein halves with the target receptor. The challenge in developing linkers for the system is that they must be of a specific length that will allow association of the intein halves in the bound conformation of the receptor but will not allow association of the intein halves in the unbound conformation of the receptor. In addition, the flexibility of the linkers must be adjusted for the same purpose. In comparison, the Buskirk et al. (2004) paper from which our project draws its inspiration has a much simpler time developing linkers as the change in confirmation of their chosen receptor created a significantly larger change in distance compared to the receptor that we are looking at. As the project progressed many different attempts at creating linkers were made.
The initial linkers that we wanted to test were from the scientific paper that was used for our project inspiration. As there is a high degree of homology between the receptor used in the paper and the receptor that we were interested in we decided to try the exact linkers used in the paper. This approach was flawed however as it overlooked the differences in the receptor conformation changes.
While testing the initial design it was also decided to test other linkers at the same time. These linkers were generated by using common synthetic linkers, found through research, and testing to see if an association event was viable by modelling the system on PyMOL. This approach generated linkers that were too long as we were focussed on making linkers long enough for the interaction but forgot about the need for the linkers to prevent interaction in the unbound conformation.
New linkers were developed taking into consideration flexibility and secondary structure.
To try to make more informed linkers we determined the shortest path around the molecule that would be ideal for the association event. We were also provided with web databases that contain the structure and information for synthetic and natural linkers. In PyMOL the angstrom distances on the newly chosen path were determined and linkers with the appropriate length and flexibility were made.
Linker Molecular Dynamics
To model our target system and to get a more realistic view on how the linkers would perform, the constructs were modelled on PyMOL with some of the initial linker designs. These models were then sent to the cluster computers here at Queen’s University to undergo a molecular dynamics simulation. Dynamic modelling of molecular constructs allows for the molecule to simulate how it would vibrate in space.
This allows for qualification of the linkers’ length and flexibility. By examining the proximity between the two halves of the intein we can determine if an association event is possible. If the linkers do not appear to reach or if there appears to be excess length causing the halves of the inteins to overshoot each other the linker can be qualified as either too short or too long. The flexibility can be qualified by watching to note how easily and often the two halves of the intein are able to come close to each other. If the halves of the intein are easily able to reach each other and do so more than once this indicates that the linker is probably at an appropriate flexibility for our purposes. If the halves easily reach each other but also flail around more than expected the linkers are likely too flexible. Conversely if the halves do not easily come into proximity than the linker is likely too inflexible.
To further the work done to design linkers a software program that interacts with PyMOL is being developed to find the shortest path for the user and design a linker that will span the space.