of nucleotides in the cell are known to block DNA Department of Genetics, Harvard Medical School, 200 Longwood elongation (Vassilev and Russev, 1984; Friedberg et al Hartwell, 1995;Longhese et al., 1996). Here we rqh1 mutants. rqh1 ⍣ , previously known as hus2 ⍣ , demonstrate that reversible S phase arrest also requires encodes a putative DNA helicase related to the Escherprotective functions that are distinct from cell cycle ichia coli RecQ helicase, with particular homology to checkpoint controls. the gene products of the human BLM and WRN genesWe have previously studied S phase arrest in the fission and the Saccharomyces cerevisiae SGS1 gene. BLM and yeast Schizosaccharomyces pombe, by isolating mutants WRN are mutated in patients with Bloom's syndrome that are sensitive to hydroxyurea (HU), a drug that blocks and Werner's syndrome respectively. Both syndromes DNA replication by depleting deoxynucleotides. Our are associated with genomic instability and cancer screening strategy was based on the observation that susceptibility. We show that, like BLM and SGS1, checkpoint-defective cells undergo an aberrant mitosis rqh1 ⍣ is required to prevent recombination and that ('cut') when treated with HU (Enoch and Nurse, 1990). in fission yeast suppression of inappropriate recombinUnder the same conditions normal cells cease DNA ation is essential for reversible S phase arrest.synthesis and arrest cell division, displaying an elongated Keywords: cell cycle/hus2/recombination/RecQ DNA cell morphology. By screening for mutants that 'cut' in helicase/Schizosaccharomyces pombe HU, a number of checkpoint-defective HU-sensitive (hus) mutants were identified (Enoch et al., 1992). Several of the mutants were found to be allelic to previously known

The binding of cyclin A to p34cdc2 and p32cdk2 and the protein kinase activity of the complexes has been measured by cell-free translation of the corresponding mRNA in extracts of frog eggs, followed by immunoprecipitation. A variety of mutant cyclin A molecules have been constructed and tested in this assay. Small deletions and point mutations of highly conserved residues in the 100-residue "cyclin box" abolish binding and activation of both p34cdc2 and p32cdk2. By contrast, large deletions at the N-terminus have no effect on kinase binding and activation, until they remove residues beyond 161, where the first conserved amino acids are found in all known examples of cyclin A. At the C-terminus, removal of 14 or more amino acids abolishes activity. We also demonstrate that deletion of, or point mutations, in the cyclin A homologue of the 10-residue "destruction box," previously described in cyclin B (Glotzer et al., 1991) abolish cyclin proteolysis at the transition from M-phase to interphase.

The specific and rapid destruction of cyclins A and B during mitosis is their most remarkable property. A short peptide motif of approximately 10 amino acids near the N‐terminus, known as the destruction box, is absolutely required for programmed proteolysis. In this paper we show that although the destruction box is necessary for the degradation of cyclin A, it is not sufficient. Mutant versions of cyclin A that cannot form complexes with p34cdc2 are stable, which we interpret to mean that this cyclin must bind to p34cdc2 in order to undergo programmed proteolysis. Thus, N‐terminal fragments of cyclin A containing little more than the destruction box and its surroundings are indestructible. p34cdc2 binding also appears to be required for the destruction of cyclin B2. In contrast, cyclin B1 does not require p34cdc2 binding for specific proteolysis. The systems for the proteolysis of cyclins A, B1 and B2 thus appear to show important differences in the way they recognize their substrates.

The destruction of mitotic cyclins by programmed proteolysis at the end of mitosis is an important element in cell cycle control. This proteolysis depends on a conserved motif of nine residues known as the ‘destruction box’, which is located 40–50 residues from the N‐terminus. The sequences of the A‐ and B‐type destruction boxes are slightly different, which might account for the differences in timing of their destruction. When the cyclin A‐type destruction box was substituted for the normal one in cyclin B1 or B2, however, the resulting constructs were unexpectedly stable, although the converse substitution of B‐type destruction boxes in cyclin A permitted normal degradation. We compared the ubiquitination of various cyclin constructs, and found that whereas mutation of the highly conserved residues in the destruction box strongly reduced the level of ubiquitinated intermediates, the stable destruction box ‘swap’ constructs did form such adducts. Thus, while ubiquitination is probably necessary for cyclin destruction, it is not sufficient. We also found that poly‐ubiquitinated cyclin derivatives are still bound to p34cdc2, which is not detectably ubiquitinated itself, raising the questions of how cyclin and cdc2 dissociate from one another, and at what stage, in the process of degradation.