Optimization of an oligonucleotide microchip for microbial identification studies: a non-equilibrium dissociation approach

The utility of a high-density oligonucleotide microarray (microchip) for id entifying strains of five closely related bacilli (Bacillus anthracis, Baci llus cereus, Bacillus mycoides, Bacillus medusa and Bacillus subtilis) was demonstrated using an approach that compares the non-equilibrium dissociati on rates ('melting curves') of all probe-target duplexes simultaneously. Fo r this study, a hierarchical set of 30 oligonucleotide probes targeting the 16S ribosomal RNA of these bacilli at multiple levels of specificity (appr oximate taxonomic ranks of domain, kingdom, order, genus and species) was d esigned and immobilized in a high-density matrix of gel pads on a glass sli de. Reproducible melting curves for probes with different levels of specifi city were obtained using an optimized salt concentration. Clear discriminat ion between perfect match (PM) and mismatch (MM) duplexes was achieved. By normalizing the signals to an internal standard (a universal probe), a more than twofold discrimination (> 2.4x) was achieved between PM and 1-MM dupl exes at the dissociation temperature at which 50% of the probe-target duple xes remained intact. This provided excellent differentiation among represen tatives of different Bacillus species, both individually and in mixtures of two or three. The overall pattern of hybridization derived from this hiera rchical probe set also provided a clear 'chip fingerprint' for each of thes e closely related Bacillus species.