biochemistry - Is there an optimal composition and length for protein linkers in FRET?

I'm designing a protein that I'd like to use in FRET reporting. General idea on the shape is: FRETprotein1--Linker--CleavageSite--Linker--FRETprotein2. I would like to know what AA are best for the linker, or does it matter? I have heard SSG is a good choice, but when I started a lit review I become unconvinced. If it doesn't matter, I may try and put more convenient restriction sites there.

Is there a good way to theoretically determine optimal length? I realize that may change depending on which FRET protein pair is chosen. The cleavage sites are between 15-18 AA long. Unfortunately, I don't think I can give any more detail than that on the cleavage sites.

Response to answer: So far there have been no FRET systems or tests published with the cleavage sites I want to study. In fact, I haven't even heard of anyone else trying. If I'm forced into a positive FRET probe, which I'm worried that I am, does anyone have any FRET pairings that don't interfere with Texas-Red on the spectrum? My first choices were Venus and Cerulean, but if I have to make a positive probe then I don't think they are a good choice.

Answer

I am also about to undertake some FRET studies (this week in fact). FRET linkers are a thing of tinkering, unfortunately. Förster resonance energy transfer, or FRET, is a phenomena that decays with $ 1/{r^6} $, the radius between the donor and acceptor. When constructing FRET reporters, there are a few things to keep in mind:

• Length of linker. The length of the linker (including any cleavage sites), follows an concave optimization curve. A linker too long (> 18 amino acids) or too short (< 5 amino acids) and the FRET signal rapidly diminishes. When considering a FRET reporter in which no cleavage site is introduced, I've seen reporter an optimal linker length of 7-8 amino acids. When including a cleavage site, I've seen optimal liker lengths reporter of 18 amino acids. TO my knowledge, there is no theoretical way to predict the FRET efficiency based on linker sequence, and from scientists I've talked to, they conduct a trial-and-error approach using purified protein products to screen for an optimal length if it appears necessary to do so.


• Type of FRET reporter. Cleavable FRET reporters fall into two classes. Positive reporters, in which signal appears after cleavage, must use longer linkers to separate fluorophores enough to reduce the FRET phenomenon. Negative reporter, in which FRET signal is lost after cleavage, must necessarily employ both short cleavage sites and linkers, to maximize the FRET phenomenon prior to cleavage. Since you mention the cleavage site itself is 15-18 AA, I would certainly expect very little FRET when uncleaved, assuming that the linker and cleavage site are unstructured, as the distance between the two fluorophores is quite large for the FRET phenomenon. Alternatively, intramolecular FRET reporters are usually designed as positive FRET controls, using short linkers to maximize the FRET phenomenon. When tagging two proteins seperately for FRET, short linkers are also chosen, relying on some external interaction to bring them together, however, this requires a lot of optimization of linker length for both fluorophores for maximum efficiency.


• Choice of linker amino acids. This part here is somewhat subjective. The choice I often see in the literature are the tiny and small amino acids (especially Ser, Ala and Gly). The choice of these is mostly because they will likely have the least interaction among the linker structure and adjoining proteins and will be the most flexible. The linker should be chosen to avoid use of putative sites for modification (e.g., phosphorylation, glycosylation), rigid geometry (Pro, or bulky aromatics) and excessive charged amino acids.

I should note that I have found no comprehensive review of these three parameters for FRET construct design, and I have gleaned them from the literature. In my opinion, linker length is more critical to optimize than its amino acid sequence, provided that the linker does not lend itself to stable/rigid geometry and modification.

I would also suggest seeing if FRET constructs have been made using your cleavage site of interest already, and perhaps use the same linking sequences. From this, you could construct a slightly shorter and longer construct to see if one is more optimal over the other.