Isolation and Characterization of par1 and par2: Two ...

Copyright  2000 by the Genetics Society of America

Isolation and Characterization of par1ⴙ and par2ⴙ: Two Schizosaccharomyces pombe Genes Encoding Bⴕ Subunits of Protein Phosphatase 2A Wei Jiang and Richard L. Hallberg Department of Biology, Syracuse University, Syracuse, New York 13244 Manuscript received July 14, 1999 Accepted for publication November 11, 1999 ABSTRACT Protein phosphatase 2A (PP2A) is one of the major serine/threonine phosphatases found in eukaryotic cells. We cloned two genes, par1⫹ and par2⫹, encoding distinct B⬘ subunits of PP2A in fission yeast. They share 52% identity at the amino acid sequence level. Neither gene is essential but together they are required for normal septum positioning and cytokinesis, for growth at both high and low temperature, and for growth under a number of stressful conditions. Immunofluorescence microscopy revealed that Par2p has a cell-cycle-related localization pattern, being localized at cell ends during interphase and forming a medial ring in cells that are undergoing septation and cytokinesis. Our analyses also indicate that Par1p is more abundant than Par2p in the cell. Cross-organism studies showed that both par1⫹ and par2⫹ could complement the rts1⌬ allele in Saccharomyces cerevisiae, albeit to different extents, in spite of the fact that neither contains a serine/threonine-rich N-terminal domain like that found in the S. cerevisiae homolog Rts1p. Thus, while Schizosaccharomyces pombe is more similar to higher eukaryotes with respect to its complement of B⬘-encoding genes, the function of those proteins is conserved relative to that of Rts1p.

P

ROTEIN phosphatase 2A (PP2A) is a major serine/ threonine phosphatase whose activity is associated with a wide variety of cellular processes such as DNA replication, RNA transcription, signal transduction, metabolic regulation, and cell cycle progression (reviewed in Mumby and Walter 1993; Walter and Mumby 1993; DePaoli-Roach et al. 1994; Shenolikar 1994; Wera and Hemmings 1995; Stark 1996; Parsons 1998; Schonthal 1998; Millward et al. 1999). Extensive biochemical characterization of purified PP2A, isolated primarily from mammalian sources, has shown that this enzyme is composed of three subunits: a 36kD catalytic subunit (C), a 65-kD structural subunit (A), and a variable regulatory subunit (B) whose mass can range from 50 to 130 kD. The B regulatory subunits apparently confer the substrate specificity and/or subcellular localization of the PP2A heterotrimer (Kamibayashi et al. 1994; McCright et al. 1996; Strack et al. 1998). In mammalian cells, there are four types of proteins that can serve the function of a B regulatory subunit [viz., B (B55), B⬘ (B56), B″ (B72)] and certain viral proteins (e.g., simian virus 40 small T antigen). Notably, there are no amino acid sequences shared by these different classes of B-subunit proteins, yet they all bind to overlapping domains of the A subunit (Ruediger et al. 1994; Kremmer et al. 1997). Recently, however, individual amino acid residues that appear to be required for the discrimination of B-type subunit

Corresponding author: Richard L. Hallberg, Department of Biology, 411 Lyman Hall, 108 College Place, Syracuse University, Syracuse, NY 13244. E-mail: [email protected] Genetics 154: 1025–1038 (March 2000)

binding have been identified in the A subunit (Ruediger et al. 1999). Within each B subunit class for a given organism, multiple isoforms usually exist. Molecular cloning has shown that in most cases the different isoforms are encoded by separate, nonallelic genes (McCright and Virshup 1995; Csortos et al. 1996; Tehrani et al. 1996; Haynes et al. 1999), but in some cases (Hendrix et al. 1993; McCright et al. 1996), alternative RNA splicing leads to the production of different isoforms. Studies in mammals (rabbits and humans) have shown (Hatano et al. 1993; Zolnierowicz et al. 1994; McCright and Virshup 1995; Csortos et al. 1996; McCright et al. 1996; Tehrani et al. 1996; Strack et al. 1998) that certain PP2A B subunit isoforms have a tissue-specific and/ or developmental stage-specific expression pattern. In addition, the cellular distribution of different isoforms can vary (McCright et al. 1996; Strack et al. 1998). For instance, in the human B⬘ (B56) family, indirect immunofluorescence microscopy has demonstrated distinct patterns of intracellular locations of some B56 isoforms, with B56␣, B56␤, and B56ε being found in the cytoplasm, while B56␦, B56␥1, and B56␥3 are concentrated in the nucleus. Heterogeneity of B subunit isoforms has also been shown for plants (Latorre et al. 1997; Haynes et al. 1999), with four different B⬘ class genes and two different B″ class genes having been identified already in Arabidopsis thaliana. The B-type genes appear to be expressed in all organs examined but at varying levels. Furthermore, mRNA levels for three of the genes increase in response to heat shock, but to various extents. The localization of the different pro-

1026

W. Jiang and R. L. Hallberg

teins encoded by these genes has not yet been reported. Recently genes potentially encoding multiple B, B⬘, and B″ proteins in C. elegans have been reported, indicating the universality of the origin of this complex protein phosphatase “system.” The above observations, together with the in vitro biochemical data showing that different B-type subunits impart subunit-specific enzymatic properties on the catalytic activity of the PP2A heterotrimer, makes it imperative that we define better the in vivo roles that these various proteins play. While the complexity of the PP2A system in higher organisms is quite daunting, the complexity of this system in Saccharomyces cerevisiae is much simpler. In this yeast, PP2A catalytic (C) subunits are encoded by two highly related genes, PPH21 and PPH22 (Ronne et al. 1991). The A subunit is encoded by a single gene, TPD3 (van Zyl et al. 1989, 1992). There are only single family members of two B subunit classes: CDC55, which belongs to the B (B55) class, and RTS1, which belongs to the B⬘ (B56) class. Genetic analyses have shown that Cdc55p and Rts1p are required for the regulation of distinctly different sets of cellular functions (Healy et al. 1991; Shu and Hallberg 1995; Shu et al. 1997; Evangelista et al. 1996; Minshull et al. 1996; Wang and Burke 1997). CDC55 is required for maintaining normal morphology at low temperature and maintaining control of a functional spindle assembly checkpoint, while RTS1 is required for regulating responses to a number of physiologically stressful conditions. RTS1 is also required for proper cell cycle progression from G2 to M phase, especially at a high temperature. Finally, on the basis of an analysis of its genome using the PSI-BLAST program (http://www.ncbi.nlm.nih. gov/cgi-bin/BLAST/nph-psi_blast), S. cerevisiae has no apparent gene encoding a B″-type subunit. S. cerevisiae is not only unique, at least thus far, in possessing a single member of the PP2A B⬘ class, but the protein encoded by that gene, Rts1p, is unique among all its homologues (⬎25) thus far described. While other B⬘ genes encode proteins ranging in size from 429 to 602 amino acids, Rts1p contains 757 residues, and the sequence of its first 250 amino acids, which contains ⵑ20% serine and threonine residues, shows no homology to any other B⬘ sequences nor to any other protein sequences. Rts1p is a multiply phosphorylated protein with the modified residues being localized almost entirely within its 250-amino-acid terminus. Cells expressing a gene encoding a truncated form of Rts1p missing its first 216 amino acids exhibit normal stress responses but show the same cell cycle deficiencies that rts1-null cells possess (Shu et al. 1997). Thus, we seem to have identified a functional domain of Rts1p. These observations prompted the following questions. If PP2A B⬘ subunits are responsible for the regulation of similar cellular processes in all organisms, and if other organisms produce more than one type of B⬘ subunit, would it be possible to find one in which a B⬘ gene

is required for stress response regulation and another required for cell cycle control? Second, do other unicellular eukaryotes also have only one B⬘-encoding gene in their genome? To answer these questions we chose to study a distantly related fission yeast, Schizosaccharomyces pombe. This yeast is not only tractable for molecular genetic analyses, but a number of the genes encoding PP2A subunits have already been described (Kinoshita et al. 1990, 1993, 1996). Two catalytic subunits of PP2A are encoded by two closely related genes, ppa1⫹ and ppa2⫹, and strains in which both genes are disrupted are inviable. The A (PR65) subunit is encoded by an essential gene, paa1⫹. The B (PR55) subunit gene pab1⫹ is not essential, but pab1-null cells grow poorly at both high and low temperatures. They are morphologically abnormal and show defects in cell wall synthesis, sporulation, and cytoskeletal distribution. No gene encoding a B⬘ or B″ subunit has yet been described in S. pombe. Accordingly, we asked: (1) Does S. pombe express a gene encoding a PP2A B⬘ protein? (2) If so, is there more than one B⬘ protein and are they the products of different genes? (3) Do any S. pombe Rts1p homologues have an N terminus homologous to the Rts1p N terminus? (4) If S. pombe expresses multiple Rts1p homologues, what are their respective functions, and how do those functions relate to the function of Rts1p in S. cerevisiae? Herein we report the cloning of two genes encoding B⬘ subunits of PP2A in S. pombe, par1⫹ and par2⫹, that are ⵑ50% identical to one another and also to RTS1. Neither gene is essential, but together they are required for normal cytokinesis, for growth at high temperature, and for growth under a variety of normally stressful conditions. The phenotype of a par1-null par2-null strain is quite similar to an rts1-null S. cerevisiae strain. Our genetic analyses show that while Par1p and Par2p may possess some ovelapping functions, they, nonetheless, have distinctive functional capacities. MATERIALS AND METHODS Yeast strains, media, genetics, and molecular biology: S. pombe haploid strains FY527 (h⫺ his3-D1 ade6-M216 ura4-D18 leu1-32) and FY528 (h⫹ his3-D1 ade6-M210 ura4-D18 leu1-32) were used in this study (both were obtained from Susan Forsburg). Fission yeast cells were grown in rich YES or EMM minimal medium supplemented with appropriate amino acids. MEA plates were used for mating and sporulation. The S. cerevisiae ␣W303 rts1⌬::His3 (Shu and Hallberg 1995) strain was used in cross-organism studies. S. cerevisiae strains were grown in rich YPD medium or in synthetic minimal medium (SD) supplemented as required. Standard yeast genetic (Moreno et al. 1991) and molecular methods (Sambrook et al. 1989) were used except where noted. Cloning of par1ⴙ and par2ⴙ cDNAs: Oligos SPRTS1-SENSE (5⬘- GCCGAACTGTTGGAAATCCTCGGGAG-3⬘) and SPRTS1ANTISENSE (5⬘-CAACTTGTGGAAGAGCGGCACCTGG-3⬘) were designed on the basis of a S. pombe genomic sequence (GenBank accession no. Z70720; SPAC1B9.03). They were used to amplify an expected 382-bp fragment from wild-type

S. pombe PP2A Regulatory Subunits genomic DNA. SPRTS1(2)-SENSE (5⬘-CTCCTGACTTTAG AGCCAG-3⬘) and SPRTS1(2)-ANTISENSE (5⬘-CTCCACAG ACATAAGGCAC-3⬘) were designed on the basis of a partial cDNA sequence (GenBank accession no. D89107). The corresponding genomic sequence was later published in the GenBank database (under the accession no. Z98533; SPAC6F12.12). A 668-bp fragment was obtained using this pair of primers against wild-type genomic DNA. Genomic Southern blots using either the 382- or the 668-bp PCR product under stringent conditions verified the existence of the designated par1⫹ and par2⫹ genes, respectively. Both pairs of primers were also used for a PCR analysis of a ␭-ZapII S. pombe cDNA library (obtained from Alison Pidoux; Pidoux et al. 1996). A predicted 330-bp (par1⫹) fragment and a 613-bp (par2⫹) fragment were obtained, respectively. The ␭-ZapII S. pombe cDNA library was screened using either the 330- or the 613-bp fragment as a probe after the standard hybridization procedure under stringent conditions (Sambrook et al. 1989). Seven positive clones were isolated with the 330-bp probe; five were isolated using the 613-bp probe. The ExAssist helper phage system (Stratagene, La Jolla, CA) was used to excise plasmids containing the cDNAs in the pBluescript vector. Under stringent conditions, no crosshybridization occurred between the two groups of plasmids. Plasmids pa-2-1 and pb-4-1, which contain the longest inserts, were then sequenced by the Cornell University sequencing center. pa-2-1 contains all but the 28 bp of the putative open reading frame (ORF) of par1⫹ plus a 24-bp 3⬘ untranslated region (UTR). Plasmid pb-4-1 contains the ORF of par2⫹ except for the first 56 bp and a 3⬘ UTR of 178 bp. Generation of the full-length ORFs for par1ⴙ and par2ⴙ: The complete ORF of par1⫹ was generated by PCR using pa-2-1 as a template. The 5⬘ primer SPRTS1A-5⬘END (5⬘GGTACCGCTAGCATGAAAGGGATTAAAAGCAAAATG GTTTCTCGAGGGAAATCTCAAGATACCC-3⬘) contains the missing 28-bp sequence from the pa-2-1 plasmid (in boldface) and a NheI site (underlined) before the start codon. The 3⬘ primer PRA-3⬘END-ANTI (5⬘-CTGCAGTGATCACTAACCA TTGGTGTAGTCAAGTG-3⬘) contains a BclI site (underlined) after the stop codon. They were used to amplify a 1673-bp complete ORF of par1⫹ from pa-2-1. This PCR product was then gel purified, digested with NheI and BclI, and subcloned into a puc18 vector. A 1.3-kb XhoI-HincII fragment was then replaced with a similar fragment from the original cDNA from pa-2-1, leaving only a 0.3-kb PCR product in this plasmid. The complete ORF of par2⫹ was generated similarly by PCR using pb-4-1 as template. The 5⬘ primer SPRTS1B-5⬘END (5⬘-G CG C AC GA G C TC CA TA T GA A A GG A TTA A G GA G TAAATTTGTAAAAGCACTTTCGTTGAAGGAT GAACAAGGGTCACATAAAAATGGT-3⬘) contains the missing 56-bp sequence from plasmid pb-4-1 (boldface) and a SacI (italic) and a NdeI site (underlined) before the start codon. The 3⬘ primer SPRTS1B-3⬘END (5⬘-GCCGACGCATGCGGAT CCTTATCTTAAATAATCAGTGGG-3⬘) contains a BamHI (underlined) and a SphI site (italic) after the stop codon. A 1916-bp fragment was obtained by PCR using pb-4-1 as a template. This PCR product was then gel purified, digested with SacI and SphI, and subcloned into a puc18 vector. A 1.0-kb HincII-NsiI fragment was then replaced with a similar fragment from the original cDNA from pb-4-1. Gene disruption: One-step gene disruption (Rothstein 1983) by homologous recombination was performed to disrupt par1⫹ in FY528. par1⌬::his3⫹ was constructed as follows: plasmid pa-2-1 contains par1⫹ cDNA in pBSK⫺. A 200-bp BglII fragment in the coding region in pa-2-1 was replaced by a 1.8kb BglII fragment of the S. pombe his3⫹ marker. The resulting plasmid was digested with SacI and ApaI, and a 3.2-kb linear fragment containing the par1⌬::his3⫹ construct was trans-

1027

formed into FY528. Stable his⫹ transformants were selected, and disruption of par1 in the chromosome was verified by genomic Southern analysis. par2⌬::LEU2 was also constructed by one-step gene disruption. A 713-bp SacI-HindIII fragment containing the promoter of par2⫹ was obtained by PCR, with a HindIII site engineered at position ⫺156 with respect to the par2⫹ start codon AUG (positon 0). This PCR product was then digested with SacI and HindIII, creating a 693-bp fragment. In plasmid pb-4-1, which contains the partial cDNA of par2⫹, a 0.6-kb SacI-HindIII fragment of par2⫹ ORF was replaced with the above 693-bp fragment containing the par2⫹ 5⬘ UTR sequence, creating plasmid pb-5⬘. A 2.2-kb LEU2 DNA was then inserted into the HindIII site in pb-5⬘. The resulting plasmid contained a disruption construct in which part of the promoter and the first three exons of par2⫹ were replaced with LEU2 gene. This plasmid was then digested with SacI and ApaI, and the 4.1kb linear fragment was transformed into FY527. Stable Leu⫹ transformants were selected, and genomic Southern analysis confirmed the disruption of par2 in the genome. Creating a strain with both par1 and par2 disrupted was accomplished by crossing h⫹ his3-D1 ade6-M210 ura4-D18 leu132 par1⌬::his3⫹ with h⫺ his3-D1 ade6-M216 ura4-D18 leu1-32 par2⌬::LEU2. After tetrad dissection, segregants that could grow on his⫺ leu⫺ medium were selected. Genomic Southern analysis confirmed the disruption of both genes in the genome. S. pombe plasmid constructs: Plasmid vectors pIRT2 and pTZ-ura4 were provided by Susan Forsburg. pIRT2(ura4) was generated by replacing the LEU2 marker in pIRT2 with a ura4⫹ marker. To generate the par genes behind their own promoters, pfu (Stratagene) was utilized to amplify the upstream and downstream sequences of par1⫹ and par2⫹, using wild-type S. pombe genomic DNA as template. A forward primer with a SacI site and a reverse primer with an NheI site were used to amplify a 653-bp 5⬘ promoter of par1⫹. The 312-bp 3⬘ region of par1⫹ was obtained by PCR using a forward primer containing an XmaI site and a reverse primer containing a SphI site. A triple-HA tag was added in frame at the C-terminal end of the par1⫹ ORF, which has an NheI site prior to the start codon and an XmaI site following the stop codon. The above three fragments were then ligated into a pIRT2 vector to create pIRT2-par1-3HA. pIRT2(ura4)-par1-3HA was obtained by subsequent subcloning. The same approach was used to put par2-3HA in the vector pIRT2(ura4). A 840-bp 5⬘ promoter region of par2⫹ was generated by PCR using a forward primer containing a SacI site and a reverse primer containing a XbaI site. A 383-bp 3⬘ UTR of par2⫹ was generated by PCR using a forward primer that has an XmaI site and a reverse primer that has a SphI site. A tripleHA tag was added at the C-terminal end of par2⫹ ORF, and par2-3HA had an XbaI site immediately ahead of the start codon and an XmaI site following the stop codon. These three fragments were then ligated and cloned into pIRT2(ura4) vector, and pIRT2(ura4)-par2-3HA was generated. To swap the promoters of par1⫹ and par2⫹, a 1.65-kb NheINcoI fragment containing the par1⫹ ORF in pIRT2-par1-3HA was replaced with a 1.9-kb XbaI-NcoI par2⫹ ORF (XbaI and NheI have compatible cohesive ends that can be religated), giving rise to plasmid pIRT2-aB, in which the ORF of par2⫹ is driven by par1’s promoter. Similarly, a 1.9-kb XbaI-NcoI par2⫹ ORF in pIRT2(ura)-par2-3HA was replaced with a 1.65-kb NheI-NcoI par1⫹ ORF, resulting in pIRT2(ura4)-bA, which has the par1 gene under the control of par2’s promoter. Heterologous plasmid constructs: To express par1⫹ and par2⫹ in S. cerevisiae, heterologous constructs were generated in which the ORF of par1⫹ or par2⫹ is driven by RTS1’s promoter and 3⬘ flanking sequence. To do this, pRS315RTS1-

1028


3HA was digested with SacI and BamHI, and a 6.5-kb fragment (#1) containing pRS315 plus the RTS1 3⬘ flanking sequence was obtained. A sense primer containing a SacI site and an antisense primer containing an NheI site were used to do PCR against the pRS315RTS1-3HA plasmid, and the 688-bp PCR product was digested with SacI and NheI to obtain the RTS1 5⬘ promoter sequence (fragment #2). The ORF of par1⫹, which had an engineered NheI site prior to the start codon and a BclI site following the stop codon, was digested with NheI and BclI to obtain fragment #3. Fragments #1–3 were then ligated together to obtain pRS315par1. The ORF of par2⫹ was also engineered to have an XbaI site prior to the start codon and a BamHI site following the stop codon (fragment #4). Fragments #1, 2, and 4 were ligated together, and pRS315par2 was obtained. YEP352par1 and YEP352par2 were obtained by subsequent subcloning. A triple HA tag was added in frame at the C-terminal end of either par1⫹ or par2⫹ ORF to construct pRS315par1-3HA and pRS315par2-3HA. This was achieved as follows: an NcoI site was engineered at the end of either par1’s or par2’s ORF (prior to the stop codon) in pRS315. A 0.6-kb fragment containing NcoI-3HA-TAA and the RTS1 3⬘ sequence was generated by PCR. pRS315par1 was digested with HincII (inside the par1 ORF) and ApaI (in the multiple cloning site of pRS315), and a 7.9-kb pRS315 plus RTS1 promoter and part of the par1 ORF was obtained. A 0.3-kb HincII-NcoI fragment (the other part of the par1 ORF) and the 0.6-kb NcoI-ApaI fragment (containing the triple HA tag and the 3⬘ RTS1 sequence) were then religated into the 7.9-kb vector. The final product was pRS315par1-3HA. pRS315par2 was digested with NsiI (inside the par2 ORF) and ApaI (in the multiple cloning site of pRS315) to give rise to a 8.1-kb vector. A 0.3-kb NsiI-NcoI fragment (the other part of par2 ORF) and the 0.6-kb NcoIApaI fragment (containing the triple HA tag and the 3⬘ RTS1 sequence) were then religated into the 8.1-kb vector. The final product was pRS315par2-3HA. YEP352par1-3HA and YEP352par2-3HA were obtained by subsequent subcloning. Microscopy and protein method: For immunofluorescence microscopy, S. pombe cells were fixed by adding 1/10 volume of 37% (w/v) formaldehyde and incubating at the growth temperature for 45 min. Cells were then washed once with phosphate-buffered saline (PBS) and twice with PBS-sorbitol buffer (1.2 m sorbitol in PBS), harvested, and resuspended in PBS-sorbitol. Cells were stained with 0.1 ␮g/ml DAPI (4⬘,6diamidino-2-phenylindole; Sigma, St. Louis) and 20 ␮g/ml Calcofluor (fluorescent brightener 28; Sigma) to visualize DNA and septa, respectively. To visualize triple HA-tagged Par1p and Par2p by immunofluorescence, exponentially growing cells were fixed by 3.7% formaldehyde for 30 min in EMM medium and washed once with PBS and twice with PBS-sorbitol buffer. They were digested for 30 min at 30⬚ in PBS-sorbitol containing 2.5 mg/ml Zymolase 20T (ICN Biomedicals, Costa Mesa, CA), washed once in PBS plus 1% Triton X-100 and three times in PBS, and resuspended in PBAL (PBS plus 0.1 m lysine hydrochloride, 1% BSA, and 0.1% sodium azide). Cells were applied to multiwell slides and incubated at 4⬚ for 16 hr with rat monoclonal anti-HA highaffinity antibody (clone 3F10; Boehringer Mannheim, Mannheim, Germany) diluted 1:10 in PBAL. After three 10-min washes with PBAL, FITC-conjugated goat anti-rat IgG secondary antibodies (Zymed, South San Francisco, CA) diluted 1:200 in PBAL were applied for 2 hr at room temperature in the dark. Cells were again washed three times in PBAL and then observed in mounting medium containing 0.1 ␮g/ml DAPI. Photographs were taken using a Nikon Eclipse TE300 microscope equipped with a Nikon cooled CCD camera and an ⫻100/1.3 oil objective. The images were analyzed by Winview and Adobe Photoshop softwares.

Total proteins were isolated from yeast cells and separated by SDS-PAGE (10% gels) as described previously (Shu and Hallberg 1995). The proteins were then transferred to nitrocellulose membranes, probed with a monoclonal antibody directed against HA epitope (12CA5; Boehringer Mannheim) or a polyclonal antibody against ␤-F1 ATPase (as a control), followed by a secondary antibody of goat anti-mouse IgG (alkaline phosphatase-conjugated or horseradish peroxidase-conjugated; GIBCO BRL, Rockville, MD) or goat anti-rabbit IgG (alkaline phosphatase-conjugated; GIBCO BRL).

RESULTS ⴙ

Isolation of par1 and par2ⴙ from S. pombe: To isolate copies of putative S. pombe RTS1 homologs, we used published database genomic sequences (SPAC1B9.03, GenBank accession no. Z70720, and SPAC6F12.12, GenBank accession no. Z98533) that appeared to encode two distinctly different Rts1-like proteins. Oligonucleotides were designed and used for PCR generation of what should have been internal gene fragments. Genomic Southern analyses using these PCR products as probes verified the existence of the two potential genes that we named par1⫹ and par2⫹ (PP2A RTS1 homolog 1 and 2). Using the PCR products as probes, we screened an S. pombe ␭-ZapII cDNA library (a gift of Alison Pidoux; Pidoux et al. 1996) and obtained several cDNA clones of both par1⫹ and par2⫹ (see materials and methods). The longest of these cDNAs were sequenced and compared with the database genomic sequences. Plasmid pa-2-1 contained all but 28 bases of the putative ORF of par1⫹ and a 3⬘ UTR of 24 bases. Plasmid pb-4-1 contained the ORF of par2⫹ minus the first 56 bases and a 3⬘ UTR of 178 bases. Sequence analysis showed that both of these two cDNAs lacked the introns exactly as predicted by the S. pombe genome project (http:// www.sanger.ac.uk/Projects/S_pombe/). Sequence comparisons with other PP2A Bⴕ regulatory subunits: As predicted by the published genomic sequences, the par1⫹ ORF is 1647 bases, encoding a protein of 549 amino acids. The par2⫹ ORF is 1881 bases, encoding a 627-residue protein. When Par1p and Par2p sequences are compared with other PP2A B⬘ sequences, it is clear that the core internal one-third to one-half of all these proteins is highly conserved (ⵑ70% identities; Figure 1), with the N- and C-terminal regions showing little, if any, homologies (data not shown). The two S. pombe proteins are 52% identical overall. This difference in B⬘ sequence identity within an organism is also shown for C. elegans, where its two (so far) B⬘ proteins are 56% identical; for A. thaliana, where its four B⬘encoding gene products share identities of 55–82%; and for Homo sapiens, where six different sequences share identities of 47–85%. While the S. pombe Par1p and Par2p sequences are most closely related to Rts1p, neither contains the elongated N terminus of Rts1p with its 22% serine plus threonine content. In this respect and with the fact that it contains multiple B⬘ genes, S.

S. pombe PP2A Regulatory Subunits

Figure 1.—Comparison of amino acid sequences among the core regions of PP2A B⬘ subunit homologs. The deduced amino acid sequences of par1⫹ and par2⫹ were aligned with B⬘ homologous sequences from S. cerevisiae (Rts1p), C. elegans (CE1), Drosophila melanogaster (DM), A. thaliana (AT B⬘ alpha), and human (B56 alpha1). The alignment was performed using the DNASTAR MegAlign program. Identical residues are boxed and shaded.

1029

1030


Figure 2.—Growth properties of par1⌬, par2⌬, and par1⌬ par2⌬ strains at different temperatures. Wild-type strains (FY527 and FY528) as well as the deletion strains were grown in YES-rich medium at 30⬚ to mid-log phase and diluted in YES to OD600 ⫽ 0.1. Identical amounts of serial dilutions (one-eighth) of the cell suspensions were applied to YES plates and incubated at 20⬚, 25⬚, 29⬚, 34⬚, or 37⬚ for 3 days.

pombe is much more like higher eukaryotes. S. cerevisiae remains unique in having a single gene encoding a considerably longer B⬘ regulatory subunit. Characteristics of par1⌬ and par2⌬ null strains: Onestep gene replacement was used to create strains with either a par1⌬::his3⫹ gene or a par2⌬::LEU2 gene (see materials and methods). Both strains were viable (their disruptions were confirmed by genomic Southern analyses). The double-null strain, par1⌬ par2⌬, was generated by mating the two single-null strains and subsequent tetrad analysis. The confirmed double-disruption strain was also viable. The growth properties of the three mutant strains were checked at temperatures between 20⬚ and 37⬚ (Figure 2). The par2⌬ cells were indistinguishable from wild-type cells in their temperaturedependent growth properties. The par1⌬ cells were indistinguishable from wild-type cells in growth at all temperatures except 37⬚, where they showed a significant temperature sensitivity. The double-null strain showed growth defects at all temperatures and inviability at 37⬚. Thus, like pab1⌬ cells, deleted for a gene encoding another PP2A B-like regulatory protein (Kinoshita et al. 1996), cells missing both genes encoding B⬘ regulatory subunits are viable but both cold and high temperature sensitive. S. cerevisiae rts1⌬ cells have a number of stress-related as well as cell-cycle-related defects (Shu and Hallberg 1995). As S. pombe has two different B⬘-encoding genes, we wondered if the functions of Par1p and Par2p might be distinct and whether the phenotypes of the two null strains might reveal that fact. Consequently, we tested both null strains for their growth properties when exposed to hyperosmotic and hyperionic stress and to increased ethanol concentration, and for their ability to grow using glycerol as a primary carbon source. As

seen in Figure 3, neither null strain showed increases in growth-related sensitivity relative to controls with respect to increased salt, increased osmolarity, or elevated ethanol concentrations. The par1⌬ strain did, however, grow more poorly on YPEG plates than did the wildtype or par2⌬ cells. In all, neither deletion strain was particularly compromised in its ability to withstand a number of physiological stressful conditions. By contrast, the double-null strain showed growth defects under all conditions tested. With regard to these deficiencies, the overall stress-related phenotype of the double-null S. pombe strain was essentially identical to the phenotype exhibited by rts1⌬ cells. We concluded that both par1⫹ and par2⫹ are required for S. pombe cells to efficiently cope with stress conditions and that both genes may be equally important in that regard, but with respect to growth at 37⬚, par1⫹ appears to be far more important than par2⫹. The effects of deleting par1ⴙ and par2ⴙ on cell morphology: Microscopic examination of early to mid-log par2⌬ cells growing at 30⬚ revealed no morphological abnormalities (these being defined with regard to cell shape, septum number and localization, and number of nuclei). About 10% of par1⌬ cells growing under the same conditions exhibited abnormalities, while ⵑ50% of the double-null strain showed abnormalities (see Figure 4A). There was no single abnormal phenotype associated with the double-null cells; a complicated mixture of all possible abnormalities was displayed in the cell population. For instance, 23% of the double-null cells had a septum that was not centrally located, while 12% of the population had multiple septa (two to six per cell), and in these cases, such cells usually had three or more nuclei. Three percent of the population had septa not at right angles to the long axis of the cell. In most


1031

Figure 3.—Growth properties of par1⌬, par2⌬, and par1⌬ par2⌬ strains under various stress conditions. Wild-type (FY527 and FY528) and deletion strains were grown at 30⬚ in YES medium until mid-log phase. The cultures were diluted to OD600 ⫽ 0.1 and then serially diluted as described in Figure 2. Equal numbers of cells were spotted on YPD, YPEG, YPD ⫹ 0.7 m KCl, YPD ⫹ 1.2 m sorbitol, or YPD ⫹ 4% ethanol plates as indicated. The plates were incubated at 30⬚ for 2.5 days.

Figure 4.—par1⌬ par2⌬ cells have aberrant cell shapes and multiple defects in cytokinesis. (A) Cells were grown in rich YES medium at 30⬚ until log phase and were fixed with formaldehyde to a final concentration of 3.7%. They were then stained with DAPI and Calcofluor to visualize DNA and septa, respectively. (B) par1⌬ par2⌬ cells were grown in rich medium at 30⬚ for 3 days, at which time they had reached stationary phase.

cases of cells with multiple septa, one or more interseptal compartments lacked a nucleus, with as many as three septa separating adjacent nuclei, and in some cases a septum appeared to intersect a nucleus. Finally, a large majority of the cells, even those with no septal defects, were irregularly shaped (Figure 4A, DIC images), not exhibiting the normal rod shape of wild-type S. pombe. The complexity of the mutant phenotype of par1⌬ par2⌬ cells indicated that both genes were required for the normal regulation of cytokinesis and morphogenesis. Morphological abnormalities are exacerbated in stationary phase cells: When S. pombe cells enter the plateau stage of growth, they normally accumulate as small, unseptated, Go cells. About 5% of par2⌬ cells in stationary phase possessed septa, and all were localized normally (⬍1% of controls had septa). By contrast, 32% of par1⌬ and 72% of double-null cells in plateau showed septations. In both cases, not only did a large fraction of the cells show septation, indicating an inability of such cells to have a normal Go arrest, but in the case of par1⌬ cells, ⵑ40% of the septated cells had misplaced and/ or multiple septa. Abnormalities in par1⌬ par2⌬ cells appeared in 75% of the septated cells, with 50% showing multiple septa (Figure 4B). These data reinforce the conclusion that Par1p and Par2p are essential for the proper regulation of morphological aspects of the cell cycle. Furthermore, as null cells do not properly arrest in stationary phase, these proteins probably play roles in cell cycle checkpoint controls as well. Levels of Par1p and Par2p in S. pombe cells: One possible explanation for the difference in phenotypes

1032


Figure 5.—Par1-3HAp levels are much higher than Par23HAp levels in S. pombe cells. Wild-type cells were transformed with a plasmid vector [pIRT2(ura4)] containing either par13HA or par2-3HA driven by their own promoters. Cells were grown in EMM ura⫺ medium at 30⬚ until mid-log phase. Total proteins were extracted, separated by SDS-PAGE gel, transfered to nitrocellulose filters, and subjected to immunoblotting with 12CA5 antibody to detect either Par1-3HAp or Par23HAp (see materials and methods). Identical amounts of total proteins were loaded in each lane. (A and B) Lanes 1–3 are samples taken from three independent yeast transformants containing par1-3HA; lanes 4–6 are samples taken from three independent yeast transformants containing par2-3HA. Lane M contains molecular weight markers. Enhanced chemiluminence system was used to detect the proteins. B shows the longer exposure of the film in A. (C) Western blot of Par13HAp and Par2-3HAp before (lanes 1–4) and after (lanes 5–12) promoter “Swap.” Strains were grown exponentially at 30⬚ in minimal selective medium, and total proteins were extracted as described above. (Lanes 1 and 2) Two independent transformants of par1⌬, with par1-3HA driven by its own promoter. (Lanes 3 and 4) Two independent transformants of par2⌬ carrying par2-3HA driven by its own promoter. Lane M contains molecular weight markers. (Lanes 5–8) Four different transformants of par1⌬ with par1-3HA driven by par2’s promoter. (Lanes 9–12) Four different transformants of par2⌬ carrying par2-3HA driven by the promoter of par1. In all cases the genes were carried on a pIRT2(ura4) plasmid.

of par1⌬, par2⌬, and par1⌬ par2⌬ cells was that Par1p and Par2p both perform essentially identical functions in the cells, but that Par1p is naturally more abundant than Par2p. Thus, the phenotypes of the different null strains simply reflect the differing degrees in loss of this PP2A regulatory subunit. To test this idea, we epitope tagged (3 ⫻ HA) both genes (as cDNAs), placed them behind their own promoters (see materials and methods), and introduced them on plasmids into both wildtype and null cells. When the proteins from wild-type cells carrying such plasmids were subjected to Western analyses, we reproducibly found that the cellular levels of Par1p were at least 10 times higher than for Par2p (Figure 5, A and B). Similar experiments in which the same two plasmids were introduced into cells null for

the plasmid-borne gene gave identical results (data not shown). It is important to note that the functionality of both HA-tagged genes was checked by transforming each one separately into the double-null strain. par1⌬ par2⌬ cells transformed with a plasmid carrying par1-3HA had a wild-type phenotype (the same as a par2⌬ cell), while the same cells transformed with par2-3HA retained their mutant morphological phenotype but were now not as temperature sensitive at 37⬚. Thus, they exhibited a partial par1⌬-like phenotype. These latter cells not having a fully par1⌬ phenotype we ascribe to the fact that the plasmids carrying the tagged genes do not contain centromeres and exist as multiple copies. Having found that ⬎90% of the B⬘ subunit in S. pombe was Par1p [assuming there is not a third B⬘ gene (see discussion)], the question now was whether this difference in protein quantity was the result of transcriptional or posttranscriptional mechanisms (or both). Accordingly, the PCR fragments used to screen cDNA libraries were used as probes for Northern analyses. We found that the levels of mRNA transcripts for the two genes were essentially the same in log-phase cells (data not shown), suggesting that protein differences arose from differing post-transcriptional regulations. To test that directly, we placed the par1-3HA cDNA behind the par2⫹ promoter and vice versa. Proteins from cells transformed with plasmids carrying these chimeric genes were subjected to Western analyses, and it was found that the levels of Par1p and Par2p were indistinguishable from what had been seen earlier (Figure 5C), when the genes were expressed from their own promoters. Thus, the differences in levels of Par1p and Par2p in the cell are regulated at a post-transcriptional step, but we can not tell from these experiments whether this is due to a difference in the efficiency of translation of the two mRNAs or whether the degradation rates of the two proteins are substantially different. Whatever the reason for the differences in Par1p and Par2p levels in growing cells, as these proteins are required for proper stress responses and cell cycle controls, the levels of either or both of these proteins could possibly vary as a function of cell cycle, stage of growth, or under different physiological conditions. By Western analyses, we found that Par1p cellular levels showed no variation regardless of the state of growth (Figure 6A), cell cycle, or exposure to stress (data not shown). By contrast, levels of Par2p did show variability during different physiological stresses, increasing or decreasing some two- to fourfold (data not shown). By far the most striking finding was that Par2p levels significantly decreased in mid-stationary-phase cells while Par1p levels remained unchanged (Figure 6B). What causes any of these protein levels to change and what the biochemical consequences of such changes are remain to be determined. Suffice to say that under all conditions tested, cells possess ⬎90% of their B⬘ PP2A regulatory subunit


Figure 6.—The amount of Par1-3HAp does not change during different growth stages, while Par2-3HAp levels greatly reduce during stationary phase. (A) par1⌬ cells transformed with a plasmid carrying par1-3HA and (B) par2⌬ cells transformed with a plasmid carrying par2-3HA were grown in EMM ura⫺ medium at 30⬚. Cells were harvested for total protein extraction at different growth stages. Western blots were performed using anti-HA antibody to detect either Par1-3HAp (A) or Par2-3HAp (B). The filters were then reprobed with anti-F1 ATPase ␤ subunit antibody as an internal control for protein loading. Lanes 1–6 correspond to samples taken at OD600 ⫽ 0.3, 0.8, 1.5, 3.0, 7.0, and 25. The filter in B was overexposed to compare the reduction of the amount of Par23HAp in stationary phase vs. that of the log phase, so the absolute amount of Par2-3HAp in B and Par1-3HAp in A are not comparable.

as Par1p and this is regulated by post-transcriptional mechanisms. As suggested above, from these results it remains possible that Par1p and Par2p have essentially indistinguishable functions and that the array of phenotypes seen in the single- and double-deletion strains is simply a reflection of the overall levels of B⬘ regulatory subunits that each strain possesses. Localization of Par1p and Par2p: The morphological and cytokinetic defects of par1⌬ par2⌬ cells prompted us to investigate the intracellular localization of Par1p and Par2p. To this end, we introduced a plasmid carrying either par1-3HA or par2-3HA driven by its own promoter into the cell null for the plasmid-borne gene. As stated above, the growth rates and cell morphologies of the resulting strains were indistinguishable from those of wild-type strains (data not shown), indicating that the tagged genes were functional. Exponentially

1033

Figure 7.—Localization of Par1-3HAp and Par2-3HAp. (A) Strain par1⌬ ⫹ pIRT2(ura4)-par1-3HA was grown in selective medium at 30⬚ until mid-log phase. The cells were then stained with DAPI and HA-specific antibodies. (a) Interphase cells. (b and c) Mitotic cells. (d) Cells undergoing septation. (B) Strain par2⌬ ⫹ pIRT2(ura4)-par2-3HA was grown in selective medium at 30⬚ until mid-log phase. The cells were then stained with DAPI and HA-specific antibodies. (a–c) Interphase cells. (d and e) Mitotic cells. (f–h) Cells undergoing septation.

growing cells were then fixed and stained with HAspecific antibodies. Immunofluorescence microscopy revealed that Par1-3HAp was found throughout the cytoplasm, usually with the nuclear region less stained (Figure 7A). In late mitotic cells and cells undergoing septation, Par1-3HAp was sometimes seen more highly concentrated at the cell center (Figure 7A, b). By contrast, Par2-3HAp showed a strong cell-cycle-related localization pattern. In newly formed daughter cells, Par23HAp was found at only one end of the cell (Figure 7B, a). In progressively longer cells, Par2-3HAp was found at both ends of the cell, although one end usually had a stronger signal than the other (Figure 7B, b and c).

1034


The cytoplasm was also weakly stained in those cells. In cells undergoing nuclear division, the signal of Par23HAp was very weak, if any (Figure 7B, d and e). Since Western analyses showed that the Par2-3HAp level did not vary thoughout the cell cycle (data not shown), this must have been due to the dispersion of this protein, thus making it harder to detect. In late anaphase cells and cells undergoing septation, Par2-3HAp was observed as a strong band at the cell center, and it always spanned the full cell diameter (Figure 7B, f and g). By examining microscopic views at several focal planes, we also found that Par2-3HAp was in a ring structure. This ring-like structure persisted after the septum had formed, and it disappeared before the formation of two individual daughter cells (Figure 7B, h). These differences in cellular localization certainly suggest normally dissimilar functions, at least in part, for Par1p and Par2p. Expressing par1ⴙ and par2ⴙ in S. cerevisiae: Another way of uncovering possible functional differences (should they exist) between the two S. pombe RTS1 homologues is to ask what functional capabilities each exhibits when expressed in an rts1⌬ strain. If both partially complemented the rts1-null allele, would those aspects of the overall mutant phenotype suppressed be the same in both cases, or would a different subset of deficiencies be alleviated by the two different genes? To answer these questions we generated constructs in which the par1⫹ and par2⫹ ORFs replaced the ORF of RTS1 (see materials and methods). That is, both would be driven by the RTS1 promoter and their mRNAs would contain the RTS1 5⬘ and 3⬘ UTR, maximizing the chances that the regulation of expression of the two genes would be the same relative to each other and possibly even to RTS1 iteslf. These constructs were introduced into rts1null cells either on single copy (pRS315) or 2␮ (YEp352) plasmids, and the phenotypes of the resultant transformants were assessed. With regard to the temperature sensitivity of rts1⌬ cells, expressing par1⫹ on a CEN plasmid suppressed the ts mutant phenotype very little, if at all (Figure 8), while a similarly expressed par2⫹ strongly suppressed it (but not as well as the RTS1 gene itself). Coexpression of both genes suppressed no better than par2⫹ alone. When either gene was expressed from the 2␮ plasmid, wild-type growth at 37⬚ resulted. The ability of par1⫹ and par2⫹ to suppress, in single copy, stress-related deficiencies of rts1⌬ cells was tested (Figure 8), and again, par2⫹ was functionally superior to par1⫹ and comparable to RTS1 itself. Coexpression of the two genes gave a suppression identical to par2⫹ expression alone and overexpression of either gene gave complete suppression. Finally, with respect to slow growth on glycerol, both par1⫹ and par2⫹ had equal suppressing capabilities, which, nonetheless, were not as great as that of RTS1 itself. The ability of the two genes to alleviate the cell-cyclerelated defects of rts1⌬ cells was also examined. As shown earlier (Shu et al. 1997), when growing rts1-null

Figure 8.—par1⫹ and par2⫹ can suppress rts1⌬ growthrelated phenotypes. S. pombe par1⫹ and par2⫹ were expressed in S. cerevisiae as described (see materials and methods). rts1⌬ cells transformed with various plasmids were grown in selective medium at 30⬚ until mid-log phase and were diluted in YPD to OD600 ⫽ 0.1. Serial dilutions were carried out as in Figure 2, and equal number of cells were spotted on different plates. YPD plates were then incubated for 3 days at either 30⬚ or 37⬚ as indicated. YPD ⫹ 4% ethanol, YPD ⫹ 1 m KCl, and YPG plates were incubated for 3 days at 30⬚.

cells are shifted from 30⬚ to 37⬚, the growth rate of such cells drops dramatically, and large-budded, 2N cells with unelongated spindles accumulate (G2/M arrest). To test the ability of par1⫹ and par2⫹ to complement this defect, rts1⌬ cells transformed with various plasmids and growing at 30⬚ were arrested with hydroxyurea (HU) for 3 hr. This treatment causes all cells to arrest with a large bud with a single nucleus. The HU was then washed out and the cells were shifted to 37⬚ for another 3 hr. Following such treatment (Figure 9, column A), 93% of rts1⌬ cells remained large budded, indicating a continued inability to complete cell division, compared with 53% for wild-type controls, indicating that some of these cells had begun to divide at 37⬚ following the release from HU arrest. Both par1⫹ and par2⫹ on singlecopy plasmids partially suppressed the G2/M arrest phenotype. Overproduction of either or coexpression of both brought no further reduction in percentage of large-budded cells that remained at 37⬚. Examination of the large budded cells in all populations showed that for rts1⌬ cells, 84% still contained a single nucleus (79% of the entire population), 15% a partially divided nucleus, and only 1% had a nucleus in both mother and daughter cells (Figure 9). In wild-type cells, most (55%) of the large-budded cells contained two nuclei with only


1035

Figure 9.—par1⫹ and par2⫹ can suppress the cell cycle defects of rts1null cells. ␣W303rts1⌬::His3 was transformed with different plasmids, and the resulting strains were grown in selective medium at 30⬚ until early log phase. Cells were then arrested with hydroxyurea (at a final concentration of 0.2 m) for 3 hr, washed with TE solution three times, harvested and resuspended in selective medium, and shifted to 37⬚ for another 3 hr. They were then fixed with 70% ethanol, stained with DAPI to observe DNA, and subjected for microscopic analysis. The percentage of largebudded cells in the total cell population was determined (A). Large-budded cells were further divided into three subpopulations by their nuclear morphology: cells with a single nucleus (D), nucleus in between the mother-bud neck (C), and two separate nuclei (B), and the percentage of each subpopulation of the large-budded cells was determined for each strain. In D, the numbers in parentheses indicate the percentages of cells with this phenotype as a whole. In each case, 200 cells were counted. sc, single copy (genes are carried on CEN-based plasmid pRS315, except for line 4, where par1 was carried on pRS316); hc, high copy (genes are carried on 2␮ plasmid YEP352).

21% showing a single nucleus in the mother cell (11% of the entire population). Expressing par1⫹ and par2⫹ either in single or multiple copy partially alleviated this aspect of the rts1-null phenotype. While the differences in suppression were small, high expression of par2⫹ clearly had the greatest positive effect. In summary, while both genes in single copy partially complemented the rts1-null allele, par2⫹ suppressed temperature sensitivity and high salt and ethanol sensitivity while par1⫹ did not. In the case of growth on glycerol and the 37⬚ cell cycle delay, both had similar, partial complementation capacities. While we took steps to ensure that the levels of Par1p and Par2p would most likely be the same in S. cerevisiae, it was possible (e.g., because of codon biases) that we did not achieve this and that the differences in complementing activity of the two genes was due to differing amounts of the two proteins accumulating in transformed S. cerevisiae cells rather than there being intrinsic differences in the functions of the two proteins. To address this question, we epitope tagged (3HA) both RTS1/par⫹ genes and repeated the complementation experiments. In all ways tested, the HA-tagged genes produced suppressing activities indistinguishable from their non-tagged counterparts (data not shown). Western analysis of the proteins of exponentially growing cells transformed with RTS1/par1-3HA and RTS1/par23HA genes showed (Figure 10) that, as in S. pombe, the levels of Par1p were many time higher than Par2p levels. If the cause of suppression differences were due only to quantitative differences, we would then have expected to find more Par2p than Par1p. That we did not indicates that with respect to function in S. cerevisiae Par2p is much more like Rts1p than is Par1p. In addi-

tion, given the manner in which the two proteins were expressed, it is probably much more likely that the differences in protein levels we see are the result of decidedly different turnover rates of the two proteins. DISCUSSION

We cloned two S. pombe genes, par1⫹ and par2⫹, encoding two PP2A B⬘ regulatory subunits, and we described an initial characterization of deletion mutant phenotypes. Neither par1⫹ nor par2⫹ is essential, but together they are required for several important cellular func-

Figure 10.—In S. cerevisiae, the level of Par2p is much less than the level of Par1p when both genes were expressed from the RTS1 promoter. rts1⌬ cells transformed with different plasmids were grown in selective medium to log phase. Total proteins were extracted from each culture, separated by SDSPAGE gel, transferred to a filter, and immunoblotted with 12CA5 antibody to detect HA-tagged proteins. (Lane 1) Molecular weight marker; (lane 2) rts1⌬ ⫹ pRS315; (lane 3) rts1⌬ ⫹ pRS315RTS1-3HA; (lane 4) rts1⌬ ⫹ pRS315par1-3HA; (lane 5) rts1⌬ ⫹ pRS315par2-3HA; (lane 6) rts1⌬ ⫹ YEP352par1-3HA; (lane 7) rts1⌬ ⫹ YEP352par2-3HA. Lanes 8 and 9 show a longer exposure of lanes 3 and 4, respectively.

1036


tions. These include the capacity to grow at both high and low temperatures as well as under a number of stress conditions, and the regulation of normal cell morphology, septum positioning, and cytokinesis. Although at this point we do not have physical evidence that Par1p and Par2p associate with the other known PP2A subunits in S. pombe, we would strongly argue that they do encode B⬘ subunits of PP2A on the basis of the fact that either can partially replace the loss of a known B⬘ subunit gene (RTS1) in S. cerevisiae. The phenotypes of par1⌬ par2⌬ cells are quite different from those of paa1 (encoding the structural subunit of PP2A) -null or pab1 (encoding the B regulatory subunit of PP2A) -null cells. paa1⫹ was essential for cell viability while pab1⫹ was required for growth at both low (22⬚) and high (36⬚) temperatures. pab1⌬ cells were pear or round shaped, and premature septation occurred when cells were grown in EMM2 medium. They were also defective in cell wall synthesis and sporulation at permissive temperatures (Kinoshita et al. 1996). These observations further support the notion that different regulatory subunits direct PP2A to carry out different cellular functions. S. pombe resembles higher eukaryotes more than S. cerevisiae with regard to the PP2A Bⴕ gene structure and composition: Par1p and Par2p share a high degree of identity at the amino acid sequence level with the other known B⬘ subunit proteins. Thus, unlike S. cerevisiae, whose genome encodes only one B⬘ subunit of PP2A (RTS1), S. pombe has at least two, resembling the situation in higher eukaryotes. Also, unlike Rts1p in S. cerevisiae, neither Par1p nor Par2p contains the additional N-terminal serine/threonine stretch of 250 amino acids. Although Par2p is 627 amino acids long, at least 100 amino acids longer than most other B⬘ proteins, it does not share homology with the Rts1p N terminus. Furthermore, when par genes were tagged with a triple-HA epitope at either the N (data not shown) or C terminus and expressed from their own promoters, they appeared as homogeneous bands with predicted electrophoretic mobilities, unlike Rts1p, which runs significantly slower than predicted and with much heterogeneity due to phosphorylation (Shu et al. 1997). We conclude from these facts that with respect to PP2A B⬘ subunit composition and structure, S. pombe is more like higher eukaryotes than S. cerevisiae. Nonetheless, when overproduced in S. cerevisiae, both par genes in S. pombe can complement most aspects of the rts1⌬ defects. This strongly argues that par1⫹ and par2⫹ encode structural and functional homologs of RTS1 in S. pombe. Par1p is more abundant in S. pombe cells than Par2p: We conclude that par1⫹ is responsible for the majority of PP2A B⬘ function on the basis of the following observations: when par1⫹ is deleted from the genome, cells become temperature sensitive. They also have morphological abnormalities and cytokinetic defects, although all these phenotypes are less severe than the double deletion of par1 and par2. By contrast, par2⌬ cells do

not show any observable defects in any of our analyses. Furthermore, assuming no other B⬘-encoding gene, Par1p constitutes at least 90% of the B⬘ proteins in the cell. To date, we have not detected any qualitative difference in functionality between par1⫹ and par2⫹ in S. pombe cells. In other words, the two B⬘ genes seem to have overlapping functions. Nevertheless, the data obtained by expressing the S. pombe genes in S. cerevisiae suggests that there might be an intrinsic difference between the two proteins. We demonstrated that par2⫹ was a better suppressor of the rts1-null phenotype with respect to growth-related defects, even though the concentration of Par2p was only 1/10 that of Par1p. This argues even more strongly that par2⫹ may be functionally closer to RTS1 than par1⫹. The subcellular localization of Par1-3HAp and Par23HAp: Indirect immunoflurorescence microscopy revealed that Par1-3HAp is found throughout the cytoplasm, while Par2-3HAp shows a more defined, cellcycle-related subcellular localization pattern. The localization of Par2-3HAp in interphase cells is reminiscent of that of actin patches: both are found at one end of the newly formed cell and at both ends in progressively grown longer cells (Gould and Simanis 1997). It will be of interest to see whether Par2-3HAp colocalizes with actin patches, and furthermore, whether the Par2p localization is actin dependent. In mitotic cells, a Par23HAp ring was observed after the daughter nuclei have become well separated, and it persisted in cells undergoing septation. This is quite different from the medial actomyosin ring with regard to cell cycle timing (Gould and Simanis 1997). Thus, detailed double-localization experiments with actin and/or other medial ring components are crucial for the further understanding of Par2p’s function. The role of par1ⴙ and par2ⴙ in regulation of cytokinesis: The abnormal morphology observed for par1⌬ par2⌬ cells is strikingly complex. Cells are apparently unable to maintain a polarized growth pattern, as judged by their aberrant shapes. They also have multiple defects in cytokinesis. Judging from the phenotypes of par1par2 double-null cells, we postulate that PP2A, as directed by its B⬘ subunits, is involved in several important processes. First, as one-fourth of the par1par2 double-null cells have a misplaced septum, this suggests that the B⬘-specific PP2A is involved in septum positioning and orientation. Genes such as mid1/dmf1, plo1, pom1, and the pos genes are known to be involved in the regulation of this process (Ohkura et al. 1995; Edamatsu and Toyoshima 1996; Sohrmann et al. 1996; Bahler and Pringle 1998). Mutants defective in any of these genes show a frequent mislocalization and/or misorientation of septa, although the position of the centrally located nucleus is normal prior to nuclear division. It is known that Mid1p is nuclear during interphase and relocates to form a cortical medial ring during mitosis; this relocalization seems to correlate with hyperphosphorylation


of Mid1p (Sohrmann et al. 1996). Other studies have shown that Plo1p, a member of a conserved family of ser/thr kinases, the polo-like kinases, is required for Mid1p to exit the nucleus and form a ring (Bahler et al. 1998). Pom1p, another putative protein kinase, is required for proper placement of the Mid1 ring (Bahler et al. 1998). Since phosphorylation seems likely to be involved in this regulation, and as our immunolocalization experiments showed that Par2p localizes to the middle of the dividing cell in the form of a central ring, it is tantalizing to hypothesize that par genes are involved in either the plo1-mid1 or the the pom1 pathway. Further genetic and immunolocalization experiments should help elucidate the role of the par genes in septum positioning. Secondly, ⬎10% of the par1-null par2-null cells have multiple septa separating adjacent nuclei. This phenotype has been observed for cdc16 and byr4 mutants and for cells overproducing Spg1 GTPase or Cdc7 kinase (Fankhauser et al. 1993; Fankhauser and Simanis 1994; Song et al. 1996; Schmidt et al. 1997). Cdc16 and Byr4 form a two-component GTPase-activating protein for Spg1 (Furge et al. 1998), acting as the negative regulators of cytokinesis and ensuring that only one septum is formed during each cell cycle. Spg1 is a dosage-dependent inducer of septum initiation and Cdc7 is a downstream factor in the pathway. One possibility is that the loss of par1 and par2 causes an overactive Spg1 pathway, resulting in the multiple-septa phenotype. The third step that might require the par genes is cell separation after septum formation. A fraction of par1par2 double-knockout cells are elongated, having multiple septa, with each compartment containing one nucleus. This phenotype has been observed for mutants of several genes, including ppb1⫹, which encodes a phosphatase 2B catalytic subunit (Yoshida et al. 1994), and sep1⫹, which encodes a putative transcription factor (Sipiczki et al. 1993; Ribar et al. 1997). As the biochemical pathways involved in cell separation are still unclear, our findings that the absence of PP2A B⬘ regulatory subunit genes negatively affects this process may help further reveal the underlying molecular mechanisms. We thank S. Forsburg (Salk Institute) for yeast strains and plasmids, A. Pidoux (UCSF) for S. pombe ␭-ZapII cDNA library, D. Amberg (SUNY HSC at Syracuse) for technical assistance on immunofluorescence microscopy, and S. Erdman for generous access to his CCDequipped Nikon microscope as well as valuable discussions. We thank E. Hallberg, H. Yang, and M. Gentry for technical assistance with some experiments and for critical reading of the manuscript. We also thank members of SUNY Health Science Center at Syracuse/Syracuse University yeast group for their thoughtful criticism on this work. This research has been supported by National Science Foundation grant no. MCB-9603733 (R.L.H.).

LITERATURE CITED Bahler, J., and J. R. Pringle, 1998 Pom1p, a fission yeast protein kinase that provides positional information for both polarized growth and cytokinesis. Genes Dev. 12: 1356–1370.

1037

Bahler, J., A. B. Steever, S. Wheatley, Y. I. Wang, J. R. Pringle et al., 1998 Role of polo kinase and Mid1p in determining the site of cell division in fission yeast. J. Cell Biol. 143: 1603–1616. Csortos, C., S. Zolnierowicz, E. Bako, S. D. Durbin and A. A. DePaoli-Roach, 1996 High complexity in the expression of the B⬘ subunit of protein phosphatase 2A0. Evidence for the existence of at least seven novel isoforms. J. Biol. Chem. 271: 2578–2588. DePaoli-Roach, A. A., I. K. Park, V. Cerovsky, C. Csortos, S. D. Durbin et al., 1994 Serine/threonine protein phosphatases in the control of cell function. Adv. Enzyme Regul. 34: 199–224. Edamatsu, M., and Y. Y. Toyoshima, 1996 Isolation and characterization of pos mutants defective in correct positioning of septum in Schizosaccharomyces pombe. Zool. Sci. 13: 235–239. Evangelista, C. C. Jr., R. A. M. Torres, M. P. Limbach and R. S. Zitomer, 1996 Rox3 and Rts1 function in the global stress response pathway in baker’s yeast. Genetics 142: 1083–1093. Fankhauser, C., and V. Simanis, 1994 The cdc7 protein kinase is a dosage dependent regulator of septum formation in fission yeast. EMBO J. 13: 3011–3019. Fankhauser, C., J. Marks, A. Reymond and V. Simanis, 1993 The S. pombe cdc16 gene is required both for maintainance of p34cdc2 kinase activity and regulation of septum formation: a link between mitosis and cytokenesis? EMBO J. 12: 2697–2704. Furge, K. A., K. Wong, J. Armstrong, M. Balasubramanian and C. F. Albright, 1998 Byr4 and Cdc16 form a two-component GTPase-activating protein for the Spg1 GTPase that controls septation in fission yeast. Curr. Biol. 8: 947–954. Gould, K. L., and V. Simanis, 1997 The control of septum formation in fission yeast. Genes Dev. 11: 2939–2951. Hatano, Y., H. Shima, T. Haneji, A. B. Miura, T. Sugimura et al., 1993 Expression of PP2A B regulatory subunit beta isotype in rat testis. FEBS Lett. 324: 71–75. Haynes, J. G., A. J. Hartung, J. D. Hendershot III, R. S. Passingham and S. J. Rundle, 1999 Molecular characterization of the B⬘ regulatory subunit gene family of Arabidopsis protein phosphatase 2A. Eur. J. Biochem. 260: 127–136. Healy, A. M., S. Zolnierowicz, A. E. Stapleton, M. Goebl, A. A. DePaoli-Roach et al., 1991 CDC55, a Saccharomyces cerevisiae gene involved in cellular morphogenesis: identification, characterization, and homology to the B subunit of mammalian type 2A protein phosphatase. Mol. Cell. Biol. 11: 5767–5780. Hendrix, P., R. E. Mayer-Jackel, P. Cron, J. Goris, J. Hofsteenge et al., 1993 Structure and expression of a 72-kDa regulatory subunit of protein phosphatase 2A. Evidence for different size forms produced by alternative splicing. J. Biol. Chem. 268: 15267– 15276. Kamibayashi, C., R. Estes, R. L. Lickteig, S. I. Yang, C. Craft et al., 1994 Comparison of heterotrimeric protein phosphatase 2A containing different B subunits. J. Biol. Chem. 269: 20139–20148. Kinoshita, N., H. Ohkura and M. Yanagida, 1990 Distinct, essential roles of type 1 and 2A protein phosphatases in the control of the fission yeast cell division cycle. Cell 63: 405–415. Kinoshita, N., H. Yamano, H. Niwa, T. Yoshida and M. Yanagida, 1993 Negative regulation of mitosis by the fission yeast protein phosphatase ppa2. Genes Dev. 7: 1059–1071. Kinoshita, K., T. Nemoto, K. Nabeshima, H. Kondoh, H. Niwa et al., 1996 The regulatory subunits of fission yeast protein phosphatase 2A (PP2A) affect cell morphogenesis, cell wall synthesis and cytokinesis. Genes Cells 1: 29–45. Kremmer, E., K. Ohst, J. Kiefer, N. Brewis and G. Walter, 1997 Separation of PP2A core enzyme and holoenzyme with monoclonal antibodies against the regulatory A subunit: abundant expression of both forms in cells. Mol. Cell. Biol. 17: 1692–1701. Latorre, K. A., D. M. Harris and S. J. Rundle, 1997 Differential expression of three Arabidopsis genes encoding the B⬘ regulatory subunit of protein phosphatase 2A. Eur. J. Biochem. 245: 156– 163. McCright, B., and D. M. Virshup, 1995 Identification of a new family of protein phosphatase 2A regulatory subunits. J. Biol. Chem. 270: 26123–26128. McCright, B., A. M. Rivers, S. Audlin and D. M. Virshup, 1996 The B56 family of protein phosphatase 2A (PP2A) regulatory subunits encodes differentiation-induced phosphoproteins that target PP2A to both nucleus and cytoplasm. J. Biol. Chem. 271: 22081–22089. Millward, T. A., S. Zolnierowicz and B. A. Hemmings, 1999 Regu-

1038


lation of protein kinase cascades by protein phosphatase 2A. Trends Biochem. Sci. 24: 186–191. Minshull, J., A. Straight, A. D. Rudner, A. F. Dernburg, A. Belmont et al., 1996 Protein phosphatase 2A regulates MPF activity and sister chromatid cohesion in budding yeast. Curr. Biol. 6: 1609–1620. Moreno, S., A. Klar and P. Nurse, 1991 Molecular genetic analysis of fission yeast Schizosaccharomyces pombe. Methods Enzymol. 194: 795–823. Mumby, M. C., and G. Walter, 1993 Protein serine/threonine phosphatases: structure, regulation, and functions in cell growth. Physiol. Rev. 73: 673–699. Ohkura, H., I. M. Hagan and D. M. Glover, 1995 The conserved Schizosaccharomyces pombe kinase plo1, required to form a bipolar spindle, the actin ring, and septum, can drive septum formation in G1 and G2 cells. Genes Dev. 9: 1059–1073. Parsons, R., 1998 Phosphatases and tumorigenesis. Curr. Opin. Oncol. 10: 88–91. Pidoux, A. L., M. Ledizet and W. Z. Cande, 1996 Fission yeast pkl1 is a kinesin-related protein involved in mitotic spindle function. Mol. Biol. Cell 7: 1639–1655. Ribar, B., A. Banrevi and M. Sipiczki, 1997 sep1⫹ encodes a transcription-factor homologue of the HNF-3/forkhead DNA-binding-domain family in Schizosaccharomyces pombe. Gene 202: 1–5. Ronne, H., M. Carlberg, G. Z. Hu and J. O. Nehlin, 1991 Protein phosphatase 2A in Saccharomyces cerevisiae: effects on cell growth and bud morphogenesis. Mol. Cell. Biol. 11: 4876–4884. Rothstein, R. J., 1983 One-step gene disruption in yeast. Methods Enzymol. 101: 202–211. Ruediger, R., M. Hentz, J. Fait, M. Mumby and G. Walter, 1994 Molecular model of the A subunit of protein phosphatase 2A: interaction with other subunits and tumor antigens. J. Virol. 68: 123–129. Ruediger, R., K. Fields and G. Walter, 1999 Binding specificity of protein phosphatase 2A core enzyme for regulatory B subunits and T antigens. J. Virol. 73: 839–842. Sambrook, J., E. F. Fritsch and T. Maniatis, 1989 Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Schmidt, S., M. Sohrmann, K. Hofmann, A. Woollard and V. Simanis, 1997 The Spg1p GTPase is an essential, dosage-dependent inducer of septum formation in Schizosaccharomyces pombe. Genes Dev. 11: 1519–1534. Schonthal, A. H., 1998 Role of PP2A in intracellular signal transduction pathways. Front. Biosci. 3: D1262–D1273. Shenolikar, S., 1994 Protein serine/threonine phosphatases—new avenues for cell regulation. Annu. Rev. Cell Biol. 10: 55–86. Shu, Y., and R. L. Hallberg, 1995 SCS1, a multicopy suppressor

of hsp60-ts mutant alleles, does not encode a mitochondrially targeted protein. Mol. Cell. Biol. 15: 5618–5626. Shu, Y., H. Yang, E. Hallberg and R. L. Hallberg, 1997 Molecular genetic analysis of Rts1p, a B⬘ regulatory subunit of Saccharomyces cerevisiae protein phosphatase 2A. Mol. Cell. Biol. 17: 3242–3253. Sipiczki, M., B. Grallert and I. Miklos, 1993 Mycelial and syncytial growth in Schizosaccharomyces pombe induced by novel septation mutations. J. Cell Sci. 104: 485–493. Sohrmann, M., C. Fankhauser, C. Brodbeck and V. Simanis, 1996 The dmf1/mid1 gene is essential for correct positioning of the division septum in fission yeast. Genes Dev. 10: 2707–2719. Song, K., K. E. Mach, C. Y. Chen, T. Reynolds and C. F. Albright, 1996 A novel suppressor of ras1 in fission yeast, byr4, is a dosagedependent inhibitor of cytokinesis. J. Cell Biol. 133: 1307–1319. Stark, M. J., 1996 Yeast protein serine/threonine phosphatases: multiple roles and diverse regulation. Yeast 12: 1647–1675. Strack, S., J. A. Zaucha, F. F. Ebner, R. J. Colbran and B. E. Wadzinski, 1998 Brain protein phosphatase 2A: developmental regulation and distinct cellular and subcellular localization by B subunits. J. Comp. Neurol. 392: 515–527. Tehrani, M. A., M. C. Mumby and C. Kamibayashi, 1996 Identification of a novel protein phosphatase 2A regulatory subunit highly expressed in muscle. J. Biol. Chem. 271: 5164–5170. van Zyl, W. H., N. Wills and J. R. Broach, 1989 A general screen for mutants of Saccharomyces cerevisiae deficient in tRNA biosynthesis. Genetics 123: 55–68. van Zyl, W., W. Huang, A. A. Sneddon, M. Stark, S. Camier et al., 1992 Inactivation of the protein phosphatase 2A regulatory subunit A results in morphological and transcriptional defects in Saccharomyces cerevisiae. Mol. Cell. Biol. 12: 4946–4959. Yoshida, T., T. Toda and M. Yanagida, 1994 A calcineurin-like gene ppb1⫹ in fission yeast: mutant defects in cytokinesis, cell polarity, mating and spindle pole body positioning. J. Cell Biol. 107: 1725–1735. Walter, G., and M. Mumby, 1993 Protein serine/threonine phosphatases and cell transformation. Biochem. Biophys. Acta 1155: 207–226. Wang, Y., and D. J. Burke, 1997 Cdc55p, the B-type regulatory subunit of protein phosphatase 2A, has multiple functions in mitosis and is required for the kinetochore/spindle checkpoint in Saccharomyces cerevisiae. Mol. Cell. Biol. 17: 620–626. Wera, S., and B. A. Hemmings, 1995 Serine/threonine protein phosphatases. Biochem. J. 311: 17–29. Zolnierowicz, S., C. Csortos, J. Bondor, A. Verin, M. C. Mumby et al., 1994 Diversity in the regulatory B-subunits of protein phosphatase 2A: identification of a novel isoform highly expressed in brain. Biochemistry 33: 11858–11867. Communicating editor: P. G. Young