Fluorescent In Situ Hybridization (FISH) allows detection of both cultivable and uncultivable microbes in the environment.  In conventional FISH protocols, the number of microbial groups that can be distinguished simultaneously is limited by the filter sets employed.  The objective of our study is to develop a technique that enables differentiation of distinct but overlapping reporter spectra, thereby increasing the number of bacterial groups that can be detected simultaneously.  By applying a spectral imaging technology and a linear unmixing algorithm, we successfully distinguished bacterial cells labeled with 8 different fluorescent reporter molecules.  As a proof of principle, E. coli cells were hybridized with unique binary combinations of 8 distinct fluorescent probes in 28 separate tubes.  In this experiment, we used fluorescent molecules conjugated to the oligonucleotide Eub338 that targets the 16S rRNA of most bacteria.  After hybridization, cells were mixed into one tube and mounted on a slide for image analysis.  By resolving spectral information within each cell in the acquired images, we distinguished 28 E. coli groups that were labeled with different binary combinations of the 8 reporter molecules.  To demonstrate the applicability of the method in a more realistic situation, a mixture of several bacterial species was incubated with a mixture of species-specific probes using binary combinations of fluorescent reporter molecules.  By applying the same spectral imaging protocol and linear unmixing algorithm, we visually distinguished the bacterial species within the artificial community.  
Results/Conclusions

Our data show that this method allows both qualitative and quantitative analysis of microbial community dynamics.  Potential applications of this protocol are numerous.  Once further optimized, this FISH-based combinatorial spectral imaging technique can be used to simultaneously visualize and monitor a large number of distinct microbes in the environment. Supported by a grant from the Sloan Foundation.