The MSN 1 and NHP 6 A Genes Suppress SWI 6 Defects in Saccharomyces cerevisiae Julia Sidorova and Linda Breeden

Ankyrin (ANK) repeats were first found in the Swi6 transcription factor of Saccharomyces cerevisiae and since then were identified in many proteins of eukaryotes and prokaryotes. These repeats are thought to serve as protein association domains. In Swi6, ANK repeats affect DNA binding of both the Swi4/Swi6 and Mbp1/Swi6 complexes. We have previously described generation of random mutations within the ANK repeats of Swi6 that render the protein temperature sensitive in its ability to activate HO transcription. Two of these SWI6 mutants were used in a screen for high copy suppressors of this phenotype. We found that MSN1, which encodes a transcriptional activator, and NHP6A, which encodes an HMG-like protein, are able to suppress defective Swi6 function. Both of these gene products are involved in HO transcription, and Nhp6A may also be involved in CLN1 transcription. Moreover, because overexpression of NHP6A can suppress caffeine sensitivity of one of the SWI6 ANK mutants, swi6-405, other SWI6-dependent genes may also be affected by Nhp6A. We hypothesize that Nhp6A and Msn1 modulate Swi6-dependent gene transcription indirectly, through effects on chromatin structure or other transcription factors, because we have not been able to demonstrate that either Msn1 or Nhp6A interact with the Swi4/Swi6 complex. THE Swi6 protein of Saccharomyces cerevisiae is inHenkel et al. 1992; Lambert and Bennett 1993; volved in the regulation of dozens of genes that are Schneider et al. 1994). The Swi6 ANK repeats are crititranscribed at the G1/S transition of the cell cycle. These cal for its function, but their role is unclear. Our modelinclude the genes encoding the HO endonuclease, G1 ing analysis of Swi6 ANK repeats (Ewaskow et al. 1998) cyclins (Nasmyth and Dirick 1991; Ogas et al. 1991; suggests that they may represent a new protein fold that Dirick et al. 1992; Measday et al. 1994), and many of supports active conformation of Swi6 complexes and/ the genes involved in DNA replication (McIntosh et or properly displays protein association interfaces. al. 1988; Lowndes et al. 1991, 1992; Dirick et al. 1992). We have generated numerous ANK repeat mutations Swi6 associates with at least two DNA-binding factors, in Swi6 that render a transcriptionally inactive protein Swi4 and Mbp1. These associations are mediated by (Sidorova and Breeden 1993; Ewaskow et al. 1998). the C-terminal domains of the proteins (Andrews and At the biochemical level, these mutations can lead to a Moore 1992b; Primig et al. 1992; Koch et al. 1993; reduced DNA binding, but most also cause a significant Sidorova and Breeden 1993), and the N termini of shift in the mobility of Swi4/Swi6 complexes in band Swi4 and Mbp1 confer the DNA-binding specificity to shift gels (Ewaskow et al. 1998). The latter suggests the these complexes. The Swi4/Swi6 complex binds to SCB possibility that the DNA-bound Swi4/Swi6 complex can (CACGAAA) elements in CLN2, HO, and PCL1 (PHO85 undergo a significant conformational change. FurtherCycLin) and MCB-like elements of CLN1 (Andrews and more, it is possible that there are accessory proteins Herskowitz 1989; Andrews and Moore 1992a; Primig that may modulate this change to affect transcriptional et al. 1992a; Partridge et al. 1997). The Mbp1/Swi6 activity of the Swi4/Swi6 complex. complex binds to MCB (ACGCGTnA) elements (McInIn this work we used two of the temperature-sensitive tosh et al. 1991; Lowndes et al. 1992; Koch et al. 1993). ANK alleles of SWI6 in a high copy suppressor screen Swi4, Mbp1, Swi6, and the S. pombe homologues Res1, to find proteins needed for the full activity of the Swi4/ Res2, and Cdc10 contain four ankyrin (ANK) repeats. Swi6 complex. We report the results of this screen and These repeats are degenerate 33 amino acid motifs that the further analysis of two of these high copy suppresare found in tandem in many different proteins in both sors, NHP6A and MSN1. eukaryotes and prokaryotes (Bork 1993). In several cases ANK repeats have been shown to be involved in MATERIALS AND METHODS protein-protein association (Thompson et al. 1991; Strains and plasmids: The BY600 MATa swi6::TRP1 ade ho::lacZ ura3 his3 leu2-3,112 trp1-1 can1-100 met2 and BY606 MATa swi4::LEU2 ade ho::lacZ ura3 his3 leu2-3,112 trp1-1 can1Corresponding author: Linda Breeden, Fred Hutchinson Cancer Re100 met2 strains are described in Sidorova and Breeden search Center, A2-168, 1100 Fairview Ave. N., Seattle, WA 98109. E-mail: [email protected] (1993). The BY1998 strain MATa his3D200 ura3-52 lys2-801 Genetics 151: 45–55 ( January 1999) 46 J. Sidorova and L. Breeden ade2-101 msn1D1::TRP1 trp1 was kindly provided by M. Carlson the URA3 gene and GAL-GST-NHP6A, into EcoRV-cut pBD2055; the resulting construct is called pBD2063. (Estruch and Carlson 1990). The BY2036 strain MATa ura352 trp1-289 his3D1 leu2-3 gal2 gal10 nhp6a-D3::URA3 nhp6bFor the GST-MSN1 fusions (pBD2061 and 2062), first the 1.5-kb SacI fragment from pBD2049 was substituted for the D3::HIS3 was kindly provided by M. Snyder (Costigan et al. 1994). BY28 MATa ade2 his3 leu2-3,112 trp1-1 ura3 is W303SacI fragment in pBD2058. Then the 1.4-kb BamHI fragment of the resulting pBD2060 containing MSN1 was cloned into 1a. The plasmid pSWI6 (pBD1378) was described previously (Sidorova et al. 1995). pSWI6-406 (pBD2046) and pSWI6-405 BamHI-cut pBD1905, to give rise to pBD2061, or into BamHIcut pGex2T (Pharmacia), to give rise to pBD2062. (pBD2031) are analogous to pBD1378, but contain mutated alleles of SWI6. Positions of these and other mutations in the To purify Gst fusions from E. coli, pBD2064 and pBD2062 were transformed into DH5a cells and the resulting strains SWI6 alleles used in this study are listed in Table 2. Strains BY1954 swi6D LEU2::swi6-405, BY1956 swi6D LEU2::swi6-406, were treated according to Pharmacia Biotech Gene Fusion System protocols. Bacterial cultures were grown to OD 0.6, and and analogous strains bearing other ANK mutations were constructed by integrative transformation of BY600 strain (Sidorfusion protein expression was induced by 0.1 mm isopropyl thiogalactoside for 2 hr. Cells were then harvested, sonicated, ova and Breeden 1993) with linearized pRS305 plasmids carrying SWI6 DNA (HindIII to SmaI fragments). High copy centrifuged, and extracts were incubated with glutathione Sepharose 4B beads (Sigma, St. Louis) for 30 min at 48. To suppressor subclones were generated in pRS426 or pZUC12 2m vector backgrounds. Subcloned suppressors were sedetermine if the Gst fusions were capable of interacting with Swi6, these glutathione beads with fusion proteins immobiquenced with M13 universal or reverse primers and library clones borne on YEp24 vector were sequenced with primer lized on them were incubated with recombinant Swi6 or in vitro-translated Swi4/Swi6 complex, washed, boiled, and BL146 59ACTACGCGATCATGG39. The plasmid pBD2068 was a gift from M. Snyder and contains an HA-tagged NHP6A loaded onto SDS PAGE. To obtain Gst fusions from yeast, pBD2057, pBD2063, and on the YEp352 vector (Costigan et al. 1994). The plasmids pDK267 and pDK268 were provided by D. Kolodrubetz. pBD2061 were transformed into W303-1a strain. The resulting strains were grown in selective media with raffinose overnight pDK267 contains NHP6B sequence flanked by z700 bp of genomic sequence on either side, cloned into EcoRI-HindIIIand then expression of the fusions was induced by galactose for 3–4 hr. Cells were harvested, and protein extracts were digested YEp352. pDK268 carries a 1.6-kb EcoRI-PstI fragment containing the NHP6A gene and flanking sequences cloned prepared as described before (Sidorova et al. 1995) and incubated for 1 hr with glutathione beads in GST buffer containing into YEp352. protease inhibitors (100 mm Tris HCl pH 8.0, 100 mm NaCl, Growth conditions: All rich (YEPD) and minimal (YC) me0.2% NP40 with 1 mm phenylmethylsulfonyl fluoride, 1 mg/ dia and growth conditions were as described previously ml leupeptin, 1 mg/ml pepstatin A). Beads were then washed (Breeden and Mikesell 1991). Temperature-sensitive ANK in three or four changes of GST buffer. To elute fusion promutant strains were cultivated at 308 and shifted to 378 for teins from the beads, the beads were resuspended in 50 ml 8–12 hr when grown in liquid media. When grown on plates, of glutathione buffer, prepared according to the Pharmacia they were incubated at 378 for the whole period of growth. Biotech protocol. Fusions were eluted for 15 min at room DNA, RNA, and protein analysis: FACS analysis of yeast cells temperature. To determine if Swi6 copurified with any of the was done as described in (Heichman and Roberts 1996) and Gst fusions from yeast, glutathione eluates were loaded onto data were analyzed using CellQuest software. Procedures for SDS PAGE, and Western blots were performed with Swi6 antiRNA isolation and S1 protection were performed as described bodies. Alternatively, extracts of yeast cells expressing fusion previously (Breeden and Mikesell 1991). Protein extract proteins were subjected to immunoprecipitation with Swi6 or preparation, immunoprecipitation, and Western blotting were Swi4 antibodies. Immunoprecipitates were loaded onto SDS done as described before (Sidorova and Breeden 1993; SidoPAGE, Western blotted, and probed with Gst antibodies (Santa rova et al. 1995). Cruz). In vitro transcription and translation: The plasmid pBD972 Thrombin cleavage of the Nhp6A from the Gst-Nhp6A fuwas used for in vitro translation of Swi4 (Ewaskow et al. 1998). sion, bound to glutathione beads, was performed according to pBD972 was added to a TNT rabbit reticulocyte lysate coupled Pharmacia Biotech protocols. Glutathione beads were mixed transcription translation system (Promega, Madison, WI) along with 38 ml of PBS and 2 ml of thrombin solution (1 unit/ml with 20–50 ng of the recombinant Swi6 purified from Escherichia thrombin in PBS), incubated overnight at room temperature, coli (Sidorova and Breeden 1993). Reactions were carried out and centrifuged. Supernatants were used directly for DNAaccording to manufacturer’s recommendations with cold amino binding reactions. acids. Reaction products were added directly to HO promoter Gel retardation: Gel retardation analysis was performed exDNA-binding reactions or loaded onto SDS PAGE. actly as described (Sidorova and Breeden 1993; Ewaskow Gst fusion and purification from yeast or bacterial cells: To et al. 1998). When the in vitro-translated Swi4/Swi6 complex construct GST fusions, MSN1 and NHP6A were generated by was bound to DNA, little (0.2–0.5 mg) or no nonspecific compolymerase chain reaction (PCR) from pM-4 (pBD2050) and petitor dI-dC was added. The binding pattern was the same, pN-5 (pBD2055), respectively, using M13 reverse primer and regardless of whether dI-dC was present in the reaction or BL138 59GGATCCATGGTCACCCCAAGAG39 primer for not. Thrombin or glutathione eluates of the Gst-Nhp6A fusion NHP6A, and M13 reverse and BL137 59CCGGATCCATGG were directly added to DNA-binding reactions with HO proCAAGTAACC39 primers for MSN1. PCR fragments were moter fragment. No dI-dC competitor was used in these reaccloned into pCRII (Invitrogen, Carlsbad, CA) generating tions since Nhp6A is a nonspecific DNA binder (Paull and pBD2056 and pBD2059, respectively. For construction of the Johnson 1995) and can be competed from HO DNA by dIGST-NHP6A fusion, the 2.3-kb BamHI fragment with the dC ( J. Sidorova, unpublished results). NHP6A open reading frame out of pBD2056 was cloned into BamHI-digested pBD1905, which bears GST under the control of GAL promoter in pRS316 vector, to give rise to pBD2057, or into BamHI-cut pGex2T (Pharmacia, Piscataway, NJ) to RESULTS make pBD2064. The GAL-GST-NHP6A cassette was also reScreen for high copy suppressors of temperaturecloned into a 2m vector pBD2055 by insertion of the 2.9-kb EcoRV fragment of pBD2057, containing a portion of sensitive ho::lacZ expression phenotype of swi6-405 and 47 Suppressors of Swi6 Transcription Factor