A simple and nonradioactive gel retardation assay was established in order to separate weakly biotinylated nucleic acids according to the number of biotin molecules with which they are labeled. The method uses streptavidin to retard biotinylated RNA (Bio-RNA) in nondenaturing polyacrylamide gels; the streptavidin-Bio-RNA complexes are detected by silver staining. From the quantitative analysis of the band distribution, the average number of biotinylated nucleotides incorporated in a particular preparation may be calculated. The degree of biotinylation was analyzed with this method by studying the incorporation of Bio-4-UTP or Bio-11-UTP into RNA by in vitro transcription with T7 RNA polymerase. We found that both Bio-UTPs were incorporated into RNA by T7 polymerase three to four times less efficiently than unmodified UTP and that the average degree of biotinylation depended in a simple and calculable way on the Bio-UTP to total UTP ratio during transcription. Therefore, RNA transcripts with a defined average degree of biotinylation can be synthesized by T7 RNA polymerase simply by setting the Bio-UTP to total UTP ratio to the appropriate value. The gel retardation assay was also used to determine the average number and distribution of biotinyl residues in photobiotinylated RNA.