Mutational analysis of the transcript 3' end signal of the CYC1 gene of yeast

This thesis describes an attempt to define the nucleotide sequences involved in specifying the location of the 3' ends of transcripts of the CYCl gene of the yeast Saccharomyces cerevisiae. The boundaries of the 3' end signal were defined with the aid of two sets of plasmid-borne deletion variants of CYCl. Mutants in each set lacked sequences extending for varying distances toward the CYCl transcript 3' end site from a fixed site either upstream or downstream. The properties of yeast strains carrying the altered genes suggested that the CYCl 3' end signal occupied no more than 50 base pairs, 119-168 base pairs downstream of the coding sequence and therefore immediately upstream of the 3' end site itself. A fragment carrying the intact 3' end signal was inserted between the CYCl promoter and the E. coli lacZ gene on an autonomously-replicating plasmid called a pAll plasmid. Sequences within the inserted fragment were capable of suppressing the expression of lacZ from the plasmid in yeast. The distal boundary of sequences responsible for this effect coincided with the distal boundary of the 3' end signal itself. I tentatively concluded that the suppression of lacZ expression depended on 3' end signal function. Mutations were introduced throughout the 3' end signal using a new procedure of in vitro mutagenesis relying on limited primer extension on a gapped heteroduplex and subsequent misincorporation of an excision-resistant α-thionucleotide. The mutant 3' end signal fragments were introduced into pAll plasmids and assayed for their effects on lacZ expression in yeast. Several of the plasmids carrying mutant fragments supported lacZ expression at greater levels than the parental, "wild-type" plasmid, suggesting that the mutations interfered with 3' end signal activity. Those 3' end signal fragments which suppressed lacZ expression from pAll plasmids were also capable of causing the synthesis of truncated cycl transcripts when inserted into a site within the CYCl coding sequence, indicating that they indeed retained 3' end signal activity. Those fragments which allowed elevated lacZ expression from pAll plasmids did not cause production of truncated cycl transcripts from analogous constructs. These observations confirmed that lacZ expression from a pAll plasmid could be a useful screening device for 3' end signal defects. Seme 3'-extended transcripts were produced from all of the truncated genes suggesting that maximally efficient 3' end generation required sequences outside the region defined by deletion analysis. Five of six mutations tested which introduced GC base pairs into the 3' end signal impaired its activity, suggesting that a high overall AT content may be an important feature of the 3' end signal. "Terminator" sequences recognized by Zaret and Sherman in 1982 [Cell 28: 563-573] and Henikoff and colleagues in 1983 [Cell 33: 607-614] indeed seem to be involved in 3' end generation in CYCl transcripts, although neither completely describes the sequence requirements of the 3' end signal.Medicine, Faculty ofBiochemistry and Molecular Biology, Department ofGraduat