Hello, 

The molecule tagged by dye A is just an arbitrary lipid bilayer component or is it a specific target (such as PIP2) you expect your molecule of interest (the one tagged with B) to interact with?

In the first case (a random membrane component tagged with A) there will be probably no cross-correlation in FCCS as the two molecules will most likely diffuse in the lipid bilayer independently of each other. You can still use FCS to detect association of your molecule of interest with the bilayer. The diffusion of a molecule in the bilayer is slower than its diffusion in solution. When measuring FCS for the dye B you should then get correlation function with two diffusion times - one corresponding to the molecule in solution and the other to the molecule in the bilayer. The fraction of the molecules bound to the bilayer can be deduced from the ratio of the amplitudes of the two components of the correlation function. See for example https://www.researchgate.net/publication/26713367_Simultaneous_Characterization_of_Lateral_Lipid_and_Prothrombin_Diffusion_Coefficients_by_Z-Scan_Fluorescence_Correlation_Spectroscopy

In the latter case (a specific target) FCCS can measure the affinity of the molecules. you should be, however, careful in interpreting the data, since the system is more complex than when interaction of two molecules in solution is measured by FCCS. In this case you have to take into account the presence of three population of the molecule of interest 1) in solution, 2) in bilayer associated with the target and 3) in bilayer not associated with the target. Only population 2) contributes to the cross-correlation.

Article Simultaneous Characterization of Lateral Lipid and Prothromb...

Thank you D.Radek Machan.

Here dye A is just a lipid tagged with atto mixed in a low fraction with other untagged lipids. So we are looking at a non-specific interaction (no specific sites). I thought only when  molecule with dye B binds to the membrane, we will see overlap between A and B signal resulting in cross correlation.And as more and more B binds to the membrane(both diffusing together as a complex), this coincidence should increase resulting in increase in amplitude of cross correlation. So, we will get to know how G(0) evolves with time. For ex, if it can be fitted to a first order or second order reaction pulling out a rate constant.

Can you please comment on the flaw here 

Hello,

Cross-correlation in FCCS arises from co-diffusion of molecules tagged with the two different dyes. The diffusion of all molecules in a lipid bilayer is restricted to a single plane; however, within that single plane the molecules can diffuse independently of each other. There are still infinitely many possible possible trajectories in the 2-dimensional space. Positive cross-correlation is observed only if members of the two molecular species diffuse along the bilayer plane together. This requires a selective interaction between the two membrane components and is indeed used in FCCS to study specific interactions between molecules in lipid membranes. To use a trivial everyday parallel - two people confined to the same floor of a building do not necessarily follow each other.

So if A is just an arbitrary fluorescently labelled lipid, A and B are expected to diffuse independently in the bilayer, regardless of the concentration of B in the bilayer, and there will be no cross-correlation. Look at the Stefl et al. paper to see how you can use FCS to study binding of your molecule to lipid bilayers.