A common problem encountered when using fluorescence detection in real samples analysis is that the matrix may contain compounds that autofluoresce or that can be excited at the wavelengths of commonly employed fluorescent reporter molecules. This causes an increase in background fluorescence, which in turn tends to compromise the detection limits of the system. To address this issue, we investigated the use of a reporter enzyme that produces fluorescent compounds, which can be excited at wavelengths that are not commonly encountered in compounds present in real samples. For that, a whole cell-based sensing system for arsenite that employs cobA as the reporter gene was developed. The system utilizes genetically engineered bacteria that incorporate the specificity of the ars operon with the sensitivity of the cobA gene. The cobA gene codes for uroporphyrinogen III methyltransferase that converts the substrate uroporphyrinogen (urogen) III into two fluorescent compounds sirohydrochlorin and trimethylpyrrocorphin. Urogen III is ubiquitous within the cell, however, because the cells use it for vitamin B-12 and siroheme biosynthesis, this sensing system is limited by substrate availability. By supplementing the media with ALA, a precursor of urogen III, a more stable and reproducible response was obtained. We observed three excitation maxima at 357, 378, and 498 nm, with a single emission maximum at 605 nm. Excitation at 498 nm was selected because it results in less background interference as most endogenous substances are not active at this wavelength. Advantages and limitations of using the cobA gene in whole-cell sensing applications are presented. (c) 2006 Wiley Periodicals, Inc.