The 5'-sequences flanking the human MUC1 gene have been analyzed for t
heir ability to direct expression of a reporter gene (the chlorampheni
col acetyltransferase gene (CAT)) in cell lines that normally express
or do not express the MUC1 gene. A construct containing 2.9 kilobase p
airs of MUC1 5'-flanking sequence sequence showed expression of CAT in
breast and pancreatic cell lines but not in the non-epithelial cell l
ines HT-1080, SK23, and HTB96. Deletion analysis showed that maximum e
xpression was obtained in ZR-75 (breast cancer line) and HPAF (pancrea
tic cancer line) with only 743 base pairs of 5'-flanking sequence. Seq
uences within 1.6 kilobase pairs of the transcriptional start site sho
wed enhancing activity in a vector carrying an enhancerless SV40 promo
ter. Analysis of proximal 5'-sequences in a promoterless CAT vector ca
rrying the SV40 enhancer showed that sequences between -60 and -150 we
re crucial for tissue-specific expression. An Spl site at -99/-90 and
an E box (E-MUC1) at -84/-64 in this region were shown by mutational a
nalysis to play a role in the regulation of transcription. Gel shift a
nalysis with oligonucleotides and nuclear extracts of ZR-75 showed pro
tein binding to both of these sites. Sp1 binding activity was similar
in ZR-75 and HT1080 cells, whereas binding of factors to the E-MUC1 ol
igonucleotide revealed quantitative and qualitative differences betwee
n epithelial and non-epithelial cells.