Fluorescent Labels in Proteins: How Random is Random Labeling

Abstract

Most proteins have one or more lysine residues exposed on their surface that can react with fluorescent probes directed to ε-amino groups. If a protein contains multiple lysine residues, labeling is often considered random assuming equal probability of the probe to interact with any of available lysines. Working with human Neutrophil Gelatinase Associated Lipocalin (NGAL), we noticed that the fluorescence of randomly labeled protein is strongly quenched even at low incorporation of the probe (AlexaFluor488 sulfodichlorophenol ester). According to X-ray data NGAL contains sixteen lysine residues, eleven of which are exposed to the solvent. To understand the labeling and quenching mechanisms, we performed labeling at multiple probe-to-protein ratios and different pH. The dye incorporation was determined by both absorption spectra and mass spectrometry. Positions of the dye's attachments were identified by protein enzymatic fragmentation followed by mass spectrometric characterizations. Surprisingly, at low incorporation ratios, the probe was always attached to either of two lysine residues, Lys 126 and 135, which are located in the calyx of NGAL in close proximity to Trp 32 and Trp 80. This finding may explain why the fluorescence of labeled NGAL is greatly quenched. At incorporation ratios 6:1 and greater the probe was predominantly linked to five lysines: Lys 31, 63, 126, 135 and 150. The activity of lysine residues at all other positions is much lower. Thus, our study shows that labeling of NGAL is not random and depends on the local microenvironment of targeted ε-amino groups.