RNA structure is a critical determinant of poly(A) site recognition by cleavage and polyadenylation specificity factor

Sequence conservation among mammalian poly(A) sites is limited to the sequence AAUAAA, coupled with an amorphous downstream U- or GU-rich region. Since these sequences may also occur within the coding region of mRNAs, additional information must be required to define authentic poly(A) sites. Several poly(A) sites have been shown to contain sequences outside the core elements that enhance the efficiency of 3' processing in vivo and in vitro. The human immunodeficiency virus type 1, equine infectious anemia virus, and adenovirus L1 3' processing enhancers have been shown to promote the binding of cleavage and polyadenylation specificity factor (CPSF), the factor responsible for recognition of AAUAAA, to the pre-mRNA, thereby facilitating the assembly of a stable 3' processing complex. We have used in vitro selection to examine the mechanism by which the human immunodeficiency virus type 1 3' processing enhancer promotes the interaction of CPSF with the AAUAAA hexamer. Surprisingly, RNAs selected for efficient polyadenylation were related by structure rather than sequence. Therefore, in the absence of extensive sequence conservation, our results strongly suggest that RNA structure is a critical determinant of poly(A) site recognition by CPSF and may play a key role in poly(A) site definition.