A Novel Fluorescence Spectral Deconvolution Method to Characterize Membrane 
Mimetic Systems

Kerry K. Karukstis and Tyrel M. McQueen*

Surfactant molecules self-assemble into a variety of membrane-mimetic 
microstuctures in aqueous solution.  To characterize the structural 
aggregates formed, we use a fluorescent probe that distributes into 
multiple microregions of the aggregates present. Deconvolution of the 
overall fluorescence emission spectrum into a sum of overlapping Gaussian 
functions (the typical shape of spectral curves) enables an interpretation 
of the probe location within discrete microdomains.  We have developed an 
alternative spectral deconvolution method that does not require an 
assumption of the spectral band shape but fits the overall spectrum to a 
sum of individual known spectra collected for various surfactant 
aggregates. The fitting routine is based on fuzzy linear programming and 
sets two constraints that any "good" solution must satisfy, namely that (1) 
the sum of the areas of any solution must be very close to (or exceed) the 
area of the target curve (thus placing a lower bound on the solution set), 
and (2) at every point the sum of the solution values must not exceed the 
target value by more than a set amount.  In mathematically defining these 
two constraints, a common fitting parameter is minimized to find the "best" 
solution.  We are currently optimizing this fitting algorithm for several 
spectral probes used in investigations of surfactant systems.

Kerry K. Karukstis
Professor of Chemistry
Harvey Mudd College
301 East Twelfth Street
Claremont, CA 91711
E-mail: Kerry_Karukstis@hmc.edu
Phone:  (909) 607-3225
Fax:    (909) 607-7577