Rad9 functions in the DNA-damage checkpoint pathway of Saccharomyces cerevisiae. In whole-cell extracts, Rad9 is found in large, soluble complexes, which have functions in amplifying the checkpoint signal. The two main soluble forms of Rad9 complexes that are found in cells exposed to DNA-damaging treatments were purified to homogeneity. Both of these Rad9 complexes contain the Ssa1 and/or Ssa2 chaperone proteins, suggesting a function for these proteins in checkpoint regulation. Consistent with this possibility, genetic experiments indicate redundant functions for SSA1 and SSA2 in survival, G2/M-checkpoint regulation, and phosphorylation of both Rad9 and Rad53 after irradiation with ultraviolet light. Ssa1 and Ssa2 can now be considered as novel checkpoint proteins that are likely to be required for remodelling Rad9 complexes during checkpoint-pathway activation.

We find budding yeast Rad9 in two distinct, large, and soluble complexes in cell extracts. The larger (> or =850 kDa) complex, found in nondamaged cells, contains hypophosphorylated Rad9, whereas the smaller (560 kDa) complex, which forms after DNA damage, contains hyperphosphorylated Rad9 and Rad53. This smaller Rad9 complex is capable of catalyzing phosphorylation and release of active Rad53 kinase, a process requiring the kinase activity of Rad53. However, Mec1 and Tel1 are no longer required once the 560 kDa complex has been formed. We propose a model whereby Mec1/Tel1-dependent hyperphosphorylation of Rad9 results in formation of the smaller Rad9 complex and recruitment of Rad53. This complex then catalyzes activation of Rad53 by acting as a scaffold that brings Rad53 molecules into close proximity, facilitating Rad53 in trans autophosphorylation and subsequent release of activated Rad53.

The Saccharomyces cerevisiae RAD9 checkpoint gene is required for transient cell‐cycle arrests and transcriptional induction of DNA repair genes in response to DNA damage. Polyclonal antibodies raised against the Rad9 protein recognized several polypeptides in asynchronous cultures, and in cells arrested in S or G2/M phases while a single form was observed in G1‐arrested cells. Treatment with various DNA damaging agents, i.e. UV, ionizing radiation or methyl methane sulfonate, resulted in the appearance of hypermodified forms of the protein. All modifications detected during a normal cell cycle and after DNA damage were sensitive to phosphatase treatment, indicating that they resulted from phosphorylation. Damage‐induced hyperphosphorylation of Rad9 correlated with checkpoint functions (cell‐cycle arrest and transcriptional induction) and was cell‐cycle stage‐ and progression‐independent. In asynchronous cultures, Rad9 hyperphosphorylation was dependent on MEC1 and TEL1, homologues of the ATR and ATM genes. In G1‐arrested cells, damage‐dependent hyperphosphorylation required functional MEC1 in addition to RAD17, RAD24, MEC3 and DDC1, demonstrating cell‐cycle stage specificity of the checkpoint genes in this response to DNA damage. Analysis of checkpoint protein interactions after DNA damage revealed that Rad9 physically associates with Rad53.

There is increasing evidence for a central role in mammalian apoptosis of the interleukin-1β– converting enzyme (ICE) family of cysteine proteases, homologues of the product of the nematode “death” gene, ced-3. Ced-3 is thought to act as an executor rather than a regulator of programmed cell death in the nematode. However, it is not known whether mammalian ICE-related proteases (IRPs) are involved in the execution or the regulation of mammalian apoptosis. Moreover, an absolute requirement for one or more IRPs for mammalian apoptosis has yet to be established. We have used two cell-permeable inhibitors of IRPs, Z-Val-Ala-Asp.fluoromethylketone (ZVAD.fmk) and t-butoxy carbonyl-Asp.fluoromethylketone (BD.fmk), to demonstrate a critical role for IRPs in mammalian apoptosis induced by several disparate mechanisms (deregulated oncogene expression, ectopic expression of the Bcl-2 relative Bak, and DNA damage–induced cell death). In all instances, ZVAD.fmk and BD.fmk treatment inhibits characteristic biochemical and morphological events associated with apoptosis, including cleavage of nuclear lamins and poly-(ADP-ribose) polymerase, chromatin condensation and nucleosome laddering, and external display of phosphatidylserine. However, neither ZVAD.fmk nor BD.fmk inhibits the onset of apoptosis, as characterized by the onset of surface blebbing; rather, both act to delay completion of the program once initiated. In complete contrast, IGF-I and Bcl-2 delay the onset of apoptosis but have no effect on the kinetics of the program once initiated. Our data indicate that IRPs constitute part of the execution machinery of mammalian apoptosis induced by deregulated oncogenes, DNA damage, or Bak but that they act after the point at which cells become committed to apoptosis or can be rescued by survival factors. Moreover, all such blocked cells have lost proliferative potential and all eventually die by a process involving cytoplasmic blebbing.

1046 base‐pairs (bp) of genomic DNA spanning the first exon of the human alpha/beta‐interferon (IFN)‐inducible gene 6‐16 have been analysed for their role in induction. The whole gene or 5′‐flanking deletion derivatives of it were assayed for inducibility in populations of stably transfected mouse cells. 5′‐Flanking DNA fragments were assayed for their ability to confer inducibility on a reporter gene in stably and transiently transfected mouse and human cells. The data suggest that a 39 bp sequence is sufficient to confer transcriptional inducibility and can account in large part for the response of 6‐16. Two copies of this sequence, one of which contains a dinucleotide insert, are located in tandem 88 bp upstream of the 6‐16 transcriptional initiation site. For at least one of the repeat units the 5′ limit of a subregion required for induction lies in the sequence GGGAAAAT. The motif GGAAA occurs in several well characterized enhancers. Furthermore, one residue 3′ of the GGAAA there is a second motif, TGAAACT, which is conserved in the regulatory regions of other IFN‐induced genes. In gel retardation assays the oligonucleotide GGGAAAATGAAACT competes with the repeat element for binding to IFN‐modulated protein(s) but a mutated oligonucleotide, GGGAAAATGACACT does not. These results identify an alpha/beta IFN response element partially homologous to those described previously for the genes of the MHC complexes.

An intact interferon‐inducible gene has been isolated from a cosmid library of human genomic DNA. The gene (designated 6‐16) encodes a mRNA of approximately 1 kb which is induced well by alpha‐ and beta‐ but poorly by gamma‐interferons. Genomic and cDNA sequences indicate that the gene contains five exons, and that the mRNA encodes a hydrophobic polypeptide of 130 amino acids with a putative NH2‐terminal signal sequence. The 5′ end has been identified by primer extension. The corresponding genomic DNA contains a TATA box 20 nucleotides upstream of the putative transcription initiation site. After transfection of the human genomic cosmid into mouse Ltk‐ cells, human 6‐16 mRNA is expressed in response to mouse alpha‐ and beta‐ but not gamma‐interferons with the same kinetics and dose‐response as in the human cells. No such expression is observed in response to human interferons. It can be concluded that the human cosmid DNA contains all of the sequences necessary for alpha‐ and beta‐interferon‐induced gene expression and that the mechanisms governing such expression are conserved between murine and human cells.