The functional components of the mouse beta-globin poly(A) signal
Mutational analysis of the mouse $\beta$-globin poly(A) signal function in COS cells revealed a striking bipartite structure of the downstream element, located downstream of the poly(A) addition site, and one novel cleavage site determinant for cleavage site selection. From analysis of the unidirectional Ba131 deletions, the 3$\sp\prime$ boundary of the downstream element is best defined as +22 (22 nucleotides downstream from the poly(A) addition site). The 5$\sp\prime$ boundary of the downstream element was successfully defined as +5 by cluster mutations, in which groups of adjacent bases were replaced with a random CA-containing sequence in a manner that did not alter spacing. Thus, the sequence between +5 and +22 contains the major portion of the downstream element. Analysis of the internal sequences of the downstream element clearly demonstrates that the downstream element consists of two major functional components, approximately located from +5 to +10 (component I) and from +17 to +22 (component II). The length of either component I or II is longer than 3 nucleotides, and the U residues appear to be critical for poly(A) signal function. Further analysis of downstream component II with 3-base cluster mutations reveals that this component is functionally redundant. Thus, various definitions of downstream component II are possible and the spacing between components I and II are somewhat flexible. Lack of function of several random GU sequences for component II indicates that the concept of GU-richness as a description of downstream element sequences must be revised. The major role of the downstream element is in modulation of the efficiency of polyadenylation. A portion of component II, the sequence between +17/+19, appears to influence both efficiency and accuracy of polyadenylation. However, the major component determining the accuracy of polyadenylation is the dinucleotide (CA) right upstream of the cleavage site. One novel cleavage site determinant, GA, is identified, and it is half as effective as CA. Comparison between the mouse $\beta$-globin poly(A) signal and other poly(A) signals reveals some common features, and a general poly(A) signal structure is proposed.
Chen, Jingshan, "The functional components of the mouse beta-globin poly(A) signal" (1994). ETD Collection for Fordham University. AAI9416664.