I recently had some success in this area. My approach was to get lucky
when I guessed the linker. In this case, my linker became ordered
which seemed to help my fusion construct settle down enough to exhibit
density. My linker sequence came was (approximately) alternating
hydrophilic and hydrophobic residues. It ordered as part of a sheet
that mediated inter-protein contacts. I purposely did not use a low-
complexity motif for reasons I'll explain below.
While staring at this structure and allowing my imagination to wander,
I came up with the idea of an "anything goes" linker motif. Find one
that can form an alpha helix or an extended conformation with equal
(in a subjective sense) regularity. The idea is that you want a
segment that can "adapt" to arbitrary crystal packing environments. I
think many natural segments that display multiple conformations have
been observed. Check out 1owr for a natural motif that can be alpha
helical or extended/disordered. (1owr is not the chimeric construct I
refer to above which is unpublished.) I think the natural linker in
1owr would make an ideal linker in a chimeric construct.
Although I base this advice on nothing but a hunch, stay away from
poly-gly or other absurdly simple motifs. The idea is purely
mathematical: think of choosing a linker as a guessing game, and you
want to guess the right linker. Let's pretend you want a 7 aa linker.
What are the chances that the optimal linker is going to be any random
sequence of residues? 20**7. So poly-gly, from a purely combinatorial
perspective, has a 1 in 20**7 chance to be the optimal sequence. But
the chances are probably lower than 1/20**7, because we know that poly-
gly is going to be disordered and the optimal sequence will probably
be ordered. In other words, common protein structure sense and
mathematics suggest that you will be shooting yourself in the foot if
you choose a poly-gly motif as a linker.
An exception to the low complexity "rule" is the poly-A motif used in
fusion constructs with MBP. The poly-A is usually ordered and alpha
helical in these constructs.
James
On Apr 2, 2009, at 1:11 PM, Raji Edayathumangalam wrote:
Hi People,
Could anyone point me to successful examples for two unrelated
proteins that have been stitched together into one single
polypeptide chain with flexible amino acids to create a functional
chimera that was subsequently crystallized. I've looked up a few.
I am particularly interested in understanding all the important
considerations while designing a flexible linker even though many of
these factors might be case dependent might be variable, obvious and
commonsensical. Either way, I'd like to hear what folks have to say.
Thanks very much.
Raji
