Using the nuclear run-on assay we found that in proliferating cells the transcription rate in the 5' end of the murine dihydrofolate reductase (dhfr) gene was approximately ten-fold higher than in the 3' end of the gene, suggesting transcriptional attenuation within the dhfr gene. However, when the transcription rate was measured by pulse-labeling, the rate was uniform throughout the gene, and the 5' dhfr signal was approximately ten-fold lower relative to a control gene signal than in the run-on assay. Previously, the activity of a dhfr promoter linked to a luciferase reporter gene was shown to increase about ten-fold at the G1/S-phase boundary following stimulation of serum-starved cells. To determine if the run-on procedure would detect growth regulation of the endogenous dhfr gene, serum-starved and -stimulated NIH 3T3 cells were analyzed. Using a dhfr 5' end probe no difference in transcription rate between these growth states was detected and the dhfr 3' end probe did not detect signal above background. In a cell line that was amplified at the dhfr locus, the transcription rate in the 5' end of the gene increased less than two-fold in stimulated cells, but the rate in the 3' end of the gene increased five- to seven-fold. Therefore, the dhfr gene is growth regulated at the level of transcription, but the nuclear run-on assay was only able to detect a difference in transcription rate in the 3' end of the gene in amplified cells. We suggest that isolation of nuclei may activate dhfr transcription complexes that normally are activated only at the G1/S-phase boundary.