• Allele-specific polyadenylation of dihydrofolate reductasemRNA in Chinese hamster lung cells

      Yang, Honghao; Melera, Peter W. (1993)
      Two polymorphic alleles encoding dihydrofolate reductase (DHFR), exist in Chinese hamster lung (CHL) cells. One allele codes for a 21 KD enzyme, while the other codes for a 20 KD enzyme. Upon selection in methotrexate (MTX), either allele, but not both, may be amplified, resulting in a MTX-resistant phenotype. Three major DHFR transcripts 1000, 1650 and 2150 nt in length are expressed from each allele. The expression of these mRNAs is allele-specific, i.e. the 2150 nt mRNA is the most abundant species derived from the 21 K allele, while the 1000 nt mRNA is the most abundant species derived from the 20 K allele. DC- 3F/MQ8, a cell line overexpressing the 20 K allele, has a unique RNA expression profile. In this cell line, both the 1000 nt and the 1650 nt mRNA species are overexpressed to the same extent in contrast to what is seen with other 20 K-overexpressors in which only the 1000 nt species is predominantly overexpressed. Previous studies suggested that the allele specific DHFR RNA expression in CHL cells might be related to differential utilization of the three poly(A) sites present in each allele. As a first step, to study the unique RNA expression profile in DC- 3F/MQ8, we have cloned and sequenced a 2.1 kb BamH1 fragment extracted from DC-3F/MQ8 cells that contained the three major DHFR poly(A) sites. In comparing this fragment with the same fragment from another 20 K-overexpressing cells, DC-3F/MQ19, we found no sequence differences. Thus, the unique RNA expression profile in DC-3F/MQ8 could not be related to sequence changes in the 3{dollar}\sp\prime{dollar}UTRs of the DHFR mRNAs. Using PCR-mediated RNase protection assays, we have mapped the 3{dollar}\sp\prime{dollar} ends of the three major DHFR mRNAs from each allele. While the first two poly(A) sites were mapped to the same positions in the two alleles, the third poly(A) site was located at the different positions between the two alleles. The 2.1 kb BamH1 fragment containing the three poly(A) sites from each allele was then cloned into a SV40 based transfection vector downstream of a CAT reporter gene, and the vectors thus constructed were transiently expressed in COS cells. As a result, an allele-specific CAT RNA expression profile was reproduced. A four-base deletion has been found at the 3{dollar}\sp\prime{dollar} end of the third poly(A) signal sequence in the 20 K allele so that the consensus sequence AAUAAA is altered to AAUAAU in this allele. Site-directed mutagenesis studies have demonstrated that this four-base deletion is primarily responsible for the dramatic alterations in 3{dollar}\sp\prime{dollar} polyadenylation between the two alleles. Finally, our studies indicated that in spite of the pronounced differences in 3{dollar}\sp\prime{dollar} polyadenylation of DHFR mRNA between the two alleles, DHFR gene expression between the two alleles was essentially unchanged. Thus, the allelic differences in 3{dollar}\sp\prime{dollar} processing of DHFR mRNAs can not be the basis for the three-fold selective preference of the 21 K allele over the 20 K allele observed in CHL cells during long term MTX selection.