Correct Regulation of the Septation Initiation Network in Schizosaccharomyces pombe Requires the Activities of par1 and par2

Corresponding author: Richard L. Hallberg, 411 Lyman Hall, 108 College Pl., Department of Biology, Syracuse University, Syracuse, NY 13244., hallberg{at}syr.edu (E-mail)

Communicating editor: P. RUSSELL

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In Schizosaccharomyces pombe, the initiation of cytokinesis is regulated by a septation initiation network (SIN). We previously reported that deletion of par1 and par2, two S. pombe genes encoding B' regulatory subunits of protein phosphatase 2A, causes a multiseptation phenotype, very similar to that seen in hyperactive SIN mutants. In this study, we examined the genetic interactions between par deletions and mutations in the genes encoding components of SIN and found that deletion of par1 and par2 suppressed the morphological and viability defects caused by overproduction of Byr4p and rescued a loss-of-function allele of spg1. However, par deletions could not suppress any mutations in genes downstream of spg1 in the SIN pathway. We showed further that, in suppressing the lethality of a spg1 loss-of-function allele, the correct localization of Cdc7p to the spindle pole body (SPB), which is normally lost in spg1 mutant cells, was restored. The fact that par mutant cells themselves exhibited a symmetric localization of Cdc7p to SPBs indicated a hyperactivity of SIN in such cells. On the basis of our epistasis analyses and cytological studies, we concluded that par genes normally negatively regulate SIN at or upstream of cdc7, ensuring that multiple rounds of septation do not occur.

CYTOKINESIS, the separation of one cell into two, is the final event in cell division. It happens immediately after mitosis and ensures that each daughter cell receives one copy of the genome as well as other essential organelles (recently reviewed in HALES et al. 1999 ; FIELD et al. 1999 ). The fission yeast Schizosaccharomyces pombe has been used extensively as a model system to study cytokinesis because, as in animal cells, S. pombe uses an actin-myosin-based contractile ring for this process. Many cytokinesis mutants have been made available through genetic screens, and studies of such mutants together with the cloning and characterization of the corresponding genes have begun to elucidate the mechanisms behind this complicated process.

In S. pombe, the initiation of cytokinesis is regulated by a novel signal transduction pathway—the septum initiation network (SIN; MCCOLLUM and GOULD 2001 ; previously designated as the Sid pathway). Mutations in many of the components in this pathway (including spg1, cdc7, cdc11, cdc14, sid1, sid2, and sid4) give rise to a septation initiation defective (sid) phenotype, in which a medial actomyosin ring forms but fails to constrict, and the septum does not form, so cells are unable to divide. However, nuclear division and growth continue, resulting in elongated, multinucleated cells (NASMYTH and NURSE 1981 ; BALASUBRAMANIAN et al. 1998 ). Combined genetic, cytological, and molecular analyses have revealed that SIN is an elaborate signal transduction network, containing a GTPase (Spg1p; SCHMIDT et al. 1997 ), three protein kinases (Cdc7p, Sid1p, and Sid2p; FANKHAUSER and SIMANIS 1994 ; SPARKS et al. 1999 ; GUERTIN et al. 2000 ), and several other proteins (Cdc14p, Sid4p, Mob1p; FANKHAUSER and SIMANIS 1993 ; CHANG and GOULD 2000 ; GUERTIN et al. 2000 ; HOU et al. 2000 ; SALIMOVA et al. 2000 ), each of which is essential for vegetative growth. Epistatic analysis and cytological data indicated the order of the SIN pathway as follows: Spg1p Cdc7p Sid1p/Cdc14p Sid2p/Mob1p (reviewed in GOULD and SIMANIS 1997 ; LE GOFF et al. 1999 ; BALASUBRAMANIAN et al. 2000 ; MCCOLLUM and GOULD 2001 ).

The spg1 gene encodes a small Ras-like GTPase, whose activation is thought to trigger the SIN pathway. Loss-of-function alleles of spg1 display a sid phenotype, whereas overexpression of spg1 leads to uncontrolled septation (SCHMIDT et al. 1997 ). The GTPase activating protein (GAP) for Spg1p is composed of two proteins, Cdc16p and Byr4p (FURGE et al. 1998 ). Together, they act as inhibitors of this pathway. Loss-of-function mutations in byr4 or cdc16 lead to multiple rounds of septation in one cell cycle, while increased level of Byr4p, causing inactivation of Spg1p, results in a lack of septation (FANKHAUSER et al. 1993 ; SONG et al. 1996 ). The guanine-nucleotide exchange factor (GEF) for spg1, responsible for activating this pathway, has not yet been identified. Cdc7p kinase is the immediate downstream target of GTP-Spg1p. It physically associates with GTP-Spg1p, and this association is important for septation, because an spg1 effector mutant, which perturbs a region of spg1 used by many Ras family GTPases necessary for interaction with their targets, associates poorly with Cdc7p and does not cause septation when overproduced (SCHMIDT et al. 1997 ).

Several studies suggest that the SPB plays an important role in the signaling of the SIN pathway, as the SIN gene products localize to SPB(s) during at least some portion of the cell cycle. Localization of all the known components of the SIN to the SPB depends on a novel protein, Sid4p, which itself becomes associated with the SPB (CHANG and GOULD 2000 ). Spg1p is present on the SPB as an inactive GDP-bound form during interphase, as an active GTP-bound form on both SPBs in early mitosis, and then as a GTP-bound form on only one SPB in late anaphase, while the other SPB contains the GDP-Spg1p (SCHMIDT et al. 1997 ).

The level and kinase activity of Cdc7p do not vary during the cell cycle, but the protein is spatially regulated by Spg1p. The active GTP-Spg1p recruits Cdc7p to SPB(s), so that in late anaphase Cdc7p becomes localized to only one SPB, the one with the GTP-Spg1p (SOHRMANN et al. 1998 ). This asymmetric localization pattern is also true of Sid1p kinase and Cdc14p: in late anaphase they localize to only the SPB with Cdc7p (GUERTIN et al. 2000 ). Sid1p and Cdc14p physically associate with each other, and they act downstream of Cdc7p in the SIN pathway (GUERTIN et al. 2000 ). Sid2p kinase and Mob1p are the most downstream components identified to date in this pathway. They physically associate with each other, localize to the SPB(s) throughout the cell cycle, and transiently localize to the cleavage site during medial ring constriction and septation (SPARKS et al. 1999 ; HOU et al. 2000 ; SALIMOVA et al. 2000 ), requiring the function of the SIN pathway. Thus, Sid2p and Mob1p might serve as a link between the signal for initiating septation (possibly emanated from the SPB) and the actual septum formation at the cleavage site.

In this study, we have investigated the role of par1 and par2, two S. pombe B' regulatory subunits of protein phosphatase 2A, in regulating the SIN pathway. Protein phosphatase 2A (PP2A) is a major serine/threonine type phosphatase found in all eukaryotic cells and is highly conserved throughout evolution. Its activity has been linked with numerous cellular processes as diverse as DNA transcription, RNA translation, cell cycle regulation, stress responses, and signal transduction (reviewed in SCHONTHAL 1998 ; MILLWARD et al. 1999 ; JANSSENS and GORIS 2001 ). Biochemical studies have shown that the PP2A holoenzyme consists of three different subunits: a catalytic C subunit, a structural A subunit, and a regulatory B subunit that can be encoded by at least four different families of genes (designated as B, B', B'', and B''') in mammalian cells. The A subunit serves as a scaffold to bring together the B and C subunits, and it is increasingly clear that the B subunits direct PP2A activity toward different substrates or to different intracellular locations.

In fission yeast, the catalytic subunit of PP2A is encoded by two closely related genes, ppa1 and ppa2, and strains in which both genes are disrupted are inviable (KINOSHITA et al. 1990 ). The A (PR65) subunit in S. pombe is encoded by an essential gene, paa1 (KINOSHITA et al. 1996 ). The B-type (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 (KINOSHITA et al. 1996 ). We recently identified two B'-type (PR61) subunit genes, par1 and par2 (JIANG and HALLBERG 2000 ). Neither gene is essential, but together they are required for growth at both high and low temperatures, for growth under a number of stress conditions, and for normal septum positioning and cytokinesis (JIANG and HALLBERG 2000 ).

par1 par2 cells exhibit a complicated mixture of abnormal morphologies (JIANG and HALLBERG 2000 ; also see Fig 1B and Fig 3B): aberrant cell shape, misplacement of septa (septa not centrally located), multiple septation (two or more septa separate the divided nuclei), and a cell separation defect (septa are formed but cannot be cleaved, resulting in multicompartment cells, with each compartment having one or two nuclei). One of these morphological defects, the multiple-septation phenotype, is reminiscent of that observed in several other mutant strains such as cdc16-116 cells at their restrictive temperature (MINET et al. 1979 ; FANKHAUSER et al. 1993 ), byr4^- cells (LI et al. 2000 ), cells overproducing spg1 (SCHMIDT et al. 1997 ), and cells overproducing cdc7 (FANKHAUSER and SIMANIS 1994 ). In all these cases it was shown that the SIN pathway becomes hyperactive. To determine whether the par1 par2 phenotype was also caused by a hyperactive SIN pathway, we tested genetic interactions between par1 par2 and other SIN genes. Here we show that par1 par2 could suppress the byr4 overproduction effect, that par1 alone rescued a spg1-106 mutant allele, but that par1, par2, and par1 par2 could not suppress mutations in genes downstream of spg1 in the SIN pathway. We also found that in some par1 cells, Cdc7p loses its asymmetric localization pattern and is observed on both SPBs in late anaphase. We also show that the capacity of par1 cells to suppress both the temperature sensitivity and the sid morphology induced by the spg1-106 allele correlated with the restoration of normal Cdc7p localization onto SPBs, thus permitting the signal for initiating septation to be transduced to the downstream components in the SIN pathway. On the basis of our epistatic analyses and cytological data, we conclude that par1 and par2 are normally required to prevent the SIN pathway from becoming hyperactive and that the point of action of this activity is most likely at or upstream of cdc7.

View larger version (308K): In this window In a new window Download PPT slide   Figure 1. Deletion of par1 and par2 suppresses the effects of byr4 overproduction. Wild-type (FY527), par1, par2, and par1 par2 cells were transformed with either an empty Rep42 vector or a Rep42-byr4 plasmid. Twenty milliliters of WT + Rep42-byr4 (A) and par1 par2 + Rep42-byr4 (B) cells were grown in EMM + thiamine media until early log phase, 10 ml of which was fixed with formaldehyde, stained with DAPI (see MATERIALS AND METHODS), and then examined microscopically. The other 10 ml of cells was washed three times with EMM - thiamine media, resuspended in EMM - thiamine media, and grown for 16 hr before fixation and visualization. (i and ii) DAPI staining of the cells. (iii and iv) DIC images of the cells. Arrows, broken arrows, and arrowheads indicate cells described in the text. (C) To test their growth properties, cells were grown in EMM + thiamine liquid media until early log phase and half of them were then diluted in EMM + thiamine media to OD600 = 0.1. Identical amounts of serial dilutions (1/10) of the cell suspensions were applied to EMM + thiamine plates and incubated at 30° for 3 days. The other half of the cells were washed three times with EMM - thiamine media, diluted in EMM - thiamine media to OD600 = 0.1. Identical amounts of serial dilutions (1/10) of the cell suspensions were applied to EMM - thiamine plates and incubated at 30° for 3 days. (1) WT + Rep42; (2) WT + Rep42-byr4; (3) par1 + Rep42; (4) par1 + Rep42-byr4; (5) par2 + Rep42; (6) par2 + Rep42-byr4; (7) par1 par2 + Rep42; (8) par1 par2 + Rep42-byr4. (D) To determine levels of Byr4p in different strains, cells were grown in EMM + thiamine media at 30° until early log phase, and half were harvested for total protein extraction. The other half were washed three times with EMM - thiamine media, resuspended in EMM - thiamine media, grown for 16 hr, and then harvested for total protein extraction. The total proteins were then separated by SDS-PAGE (10%) gel. The upper half of the gel was transferred to nitrocellulose filter and subjected to immunoblotting with anti-Byr4p antibody (top; see MATERIALS AND METHODS), whereas the lower half of the gel was stained with Coomassie blue to show the loading in each lane (bottom panel). Lane 1: WT + Rep42-byr4 in EMM + thiamine; lane 2: WT + Rep42-byr4 in EMM - thiamine; lane 3: par1 par2 + Rep42-byr4 in EMM + thiamine; lane 4: par1 par2 + Rep42-byr4 in EMM - thiamine.

View larger version (137K): In this window In a new window Download PPT slide   Figure 2. Deletion of par1 suppresses a spg1-106 ts allele. (A) WT, spg1-106, par1, and par1 spg1-106 cells were grown on YE5S-rich plates at 25° for 3 days. The cells of each strain were then streaked onto two YE5S plates, one of them was incubated at 25° for 3 days, and the other one was incubated at 36° for 3 days. Three independent par1 spg1-106 isolates from the cross were tested. (B) par1 (i and ii), spg1-106 (iii and iv), and par1 spg1-106 (v and vi) cells were grown at 25° in YE5S liquid media until early log phase and then fixed with formaldehyde and stained with DAPI (see MATERIALS AND METHODS). (i, iii, and v) DAPI stainings; (ii, iv, and vi) DIC images. Arrows indicate septa, and arrowheads indicate cells described in the text.

View larger version (111K): In this window In a new window Download PPT slide   Figure 3. par1, par2, or par1 par2 do not suppress the sid2-250 ts allele. (A) WT, par1 par2, sid2-250, par1 sid2-250, par2 sid2-250, and par1 par2 sid2-250 cells were grown on YE5S-rich plates at 25° for 3 days. The cells of each strain were then streaked onto two YE5S plates, one of them was incubated at 25° for 3 days, and the other one was incubated at 36° for 3 days. (B) WT (i and ii), sid2-250 (iii and iv), par1 par2 (v and vi), and par1 par2 sid2-250 (vii and viii) cells were grown at 25° in YE5S liquid media until early log phase and half of the cultures were fixed with formaldehyde and stained with DAPI (i, iii, v, and vii; see MATERIALS AND METHODS), while the other half was shifted to 36° for another 5 hr before being fixed and stained with DAPI (ii, iv, vi, and viii). a, b, c, and arrows indicate cells described in the text.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Yeast strains, growth media, and standard genetics and molecular methods:
All the S. pombe strains used in this study are listed in Table 1. Strains obtained from others have been crossed with either FY527 or FY528, thus introducing a his3-D1 marker into each of them. To generate double or triple mutants, individual mutants were crossed with a par1 par2 strain carrying a plasmid containing a wild-type par1 gene as the par1 par2 strain has poor mating efficiency compared to wild-type cells. Subsequent sporulation and tetrad dissections were performed, and tetrad analyses yielded the strains with the expected genotypes. Cells were grown in rich YE5S or Edinburgh minimal medium (EMM), supplemented with appropriate amino acids. MEA plates were used for mating and sporulation. Standard yeast genetic (MORENO et al. 1991 ) and molecular methods (SAMBROOK et al. 1989 ) were used except where noted. Rep41-byr4 was obtained from Dr. C. Albright. The byr4 gene in this plasmid was then subcloned into Rep42 to create a Rep42-byr4.

spg1 knockout and analyses:
A spg1 knockout construct was generated as follows: based on the database information (Sanger Center, SPAC1565.00701), SPG1-sense (5'-ACTTACTTATGAGCTCAAGAATTACG-3'; SacI site underlined) and SPG1-antisense (5'-CAGGTTTACTCGAGAAATACATATAGGGGTTA-3'; XhoI site underlined) primers were used to amplify a 1.7-kb spg1 sequence (containing all the coding regions) from wild-type S. pombe genomic DNA. The PCR product was digested with SacI and XhoI, ligated into a pBSK vector that was digested with SacI and XhoI, thus creating pBSK-spg1. SPG1-sense #2 (5'-CAATGAAGTTGAAGAGCTCCAATCCTG-3'; SacI site underlined) and SPG1-antisense #2 (5'-CTCTGCTTGAACCCCGGCAGTAGATG-3') were used to amplify a 994-bp fragment of spg1-5' sequence from wild-type S. pombe genomic DNA. This fragment was then digested with SacI and BlgII to obtain a 810-bp fragment, which was then ligated into pBSK-spg1 and digested with SacI and BlgII, thus creating pBSK-spg1(#2). This plasmid contains the genomic sequence from -1361 to 1283 with regard to the start codon of spg1. URA4 (BglII) sense (5'-GTCAGCAGATCTAGCTACAAATCCCACTGGC-3'; BglII site underlined) and URA4(HindIII)-antisense (5'-CGTGATCCCGGGAAGCTTGTGATAT TGACG-3'; HindIII site underlined) were used to amplify a 1.8-kb ura4 gene from the plasmid pTZ-ura4 (a gift from Dr. S. Forsburg), and the PCR product was digested with BglII and HindIII. The 1.8-kb BglII-HindIII ura4 gene was then ligated into pBSK-spg1(#2) that had been digested with BglII and HindIII to create pBSK-spg1::ura4. In this plasmid, most of the coding region for spg1 was replaced with a ura4 gene.

A par1 par2 diploid strain (WJY191) was generated by crossing WJY101 with WJY104 and screening the resulting diploids for histidine and leucine auxotrophs. One-step gene disruption (ROTHSTEIN 1983 ) was performed to disrupt spg1 in WJY191 as follows: pBSK-spg1::ura4 was digested with SacI and XhoI, and the resulting 3.1-kb linear fragment containing spg1::ura4 was used to transform WJY191. Stable ura⁺ transformants were selected, and genomic DNAs were extracted from both WJY191 and the stable transformants. Southern blots were performed to verify the disruption of one copy of spg1 in the genome in nine of the stable transformant diploid strains obtained. Several of the these strains were then sporulated and tetrad analyses were performed. Out of 187 tetrads dissected, only one or two viable spores were obtained in each tetrad, none of which was ura⁺. All of the viable spores showed one of the following genotypes according to their markers: par1, par2, par1 par2, and wild type (the wild-type genotype was rarely observed compared to the others). Random spore analyses (http://pingu.salk.edu/users/forsburg/diploids.html#spores) were also performed on these strains after sporulation, and no ura⁺ spores were recovered. From the above analyses, we confirmed the finding that spg1 is essential for vegetative growth (SCHMIDT et al. 1997 ) and concluded that this essentiality cannot be bypassed by the deletion of either par1 or par2 or both.

Microscopy methods:
For differential interference contrast (DIC), 4',6-diamidino-2-phenylindole (DAPI), and Calcofluor stainings, S. pombe cells were grown in liquid media until early log phase, fixed by adding 1/10 volume of 37% (w/v) formaldehyde and incubated 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. Before visualization under the microscope, cells in PBS-sorbitol were harvested and resuspended in mount solution (0.1% phenylene diamine in 1x PBS, 90% glycerol) and then stained with either 0.1 µg/ml DAPI (Sigma, St. Louis) or 20 µg/ml Calcofluor (fluorescent brightener no. 28; Sigma) to visualize DNA and septa, respectively. For visualization of Cdc7-GFP, cells were grown in liquid media at different temperatures, and 2–3 µl of cells were spotted on the slide and observed under the microscope with a green fluorescent protein (GFP) filter. The procedure for indirect immunofluorescence microscopy was as described (JIANG and HALLBERG 2000 ). Photographs were taken using an Olympus BX60 microscope equipped with an Olympus color digital cooled-CCD camera and an Olympus UplanApo x100/1.35 oil objective. The images were analyzed by MagnaFire and Adobe Photoshop softwares.

Total protein extraction and Western analysis:
Total proteins were isolated from yeast cells and separated by SDS-PAGE (10% gels) as previously described (SHU and HALLBERG 1995 ). The proteins were then transferred to nitrocellulose membranes, probed with a monoclonal antibody directed against the HA epitope (12CA5; Boehringer Mannheim, Indianapolis) or a polyclonal antibody against Byr4p (a generous gift from Dr. C. Albright), followed by a secondary antibody of goat-anti-mouse IgG (alkaline phosphatase conjugated or horseradish peroxidase conjugated; GIBCO BRL, Grand Island, NY) or goat-anti-rabbit IgG (alkaline phosphatase conjugated; GIBCO BRL).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

par1 par2 rescues the deleterious effects of byr4 overexpression:
The multiseptation phenotype (two or more septa separating divided nuclei) observed in par1 par2 cells prompted us to ask whether this was caused by a hyperactive SIN pathway. Overproducing byr4 (byr4-OP) in wild-type cells causes inhibition of this pathway, leading to a sid phenotype (multinucleated, very elongated cells unable to form a septum) and cell death (SONG et al. 1996 ). As this is essentially the opposite of the par1 par2 phenotype, we decided to examine what the effects might be of overproducing byr4 in our par1 par2 strain. To test this, we introduced a Rep42-byr4 plasmid into both wild-type and par1 par2 cells, grew them in thiamine-containing media until early log phase, and then removed thiamine from the media to induce byr4 expression. As reported (SONG et al. 1996 ), after 16 hr of induction, wild-type cells displayed the typical sid phenotype, becoming highly elongated, multinucleated, and lacking obvious septa (Fig 1A, cells indicated by arrows in ii and iv). Calcoflour staining also revealed no septa deposition, confirming that no septation occurred in these cells (data not shown). In contrast to the wild-type situation, no sid phenotype was observed in par1 par2 cells when byr4 was overproduced (Fig 1B, iv). In fact, the morphology of par1 par2 cells remained unchanged after the induction of byr4 expression (compare iii and iv in Fig 1B), and the typical abnormalities associated with a par1 par2 strain were observed (JIANG and HALLBERG 2000 ): aberrant cell shape (Fig 1B, cells indicated by broken arrows in iii), septum misplacement (Fig 1B, cells indicated by arrows), and defective cell separation (Fig 1B, cells indicated by arrowheads). Thus, introducing par1 par2 into the genome completely suppressed the byr4-OP morphological defect. Deletion of par1 alone showed a weak suppression of the byr4-OP effect: the percentage of cells showing the sid phenotype was slightly decreased when compared with wild-type cells (data not shown). par2 alone had no effect on byr4-OP morphology (data not shown).

We then tested the growth of the various strains. As reported (SONG et al. 1996 ), overproducing byr4 in wild-type cells inhibited cell growth due to the inability of such cells to initiate cytokinesis (Fig 1C, lane 2). However, deletion of par1 and par2 completely abolished this inhibition and restored growth to a wild-type level (Fig 1C, lane 8). On media lacking thiamine, par1 par2 cells with Rep42-byr4 grew as well as both the wild-type and par1 par2 cells with Rep42 alone (Fig 1C, compare lane 8 with 1 and 7). This is consistent with our finding that the sid morphology caused by byr4-OP was also rescued by deleting par1 and par2. We also noticed that par1 itself can weakly suppress the growth defect of byr4-OP (Fig 1C, compare lanes 4 and 2), whereas par2 has little or no suppressing effect at all (Fig 1C, compare lanes 6 and 2).

There is the possibility that in a par1 par2 background Byr4p is unstable and cannot accumulate to sufficiently high levels. This could explain why par1 par2 can suppress the byr4-OP defects. To test this directly, we did a Western analysis using a polyclonal anti-Byr4p antibody (a generous gift from Dr. C. Albright) on cell extracts from various strains. We found that when par1 par2 cells with Rep42-byr4 were grown in the media without thiamine, Byr4p was indeed overproduced (Fig 1D, lanes 3 and 4), and its level was actually higher than that of the wild-type cells with Rep42-byr4 under the same conditions (Fig 1D, lanes 2 and 4). Thus, decreased amounts of Byr4p in par1 par2 cells cannot be the reason for the phenotypic suppression.

On the basis of the results from the above experiments, we concluded that in the absence of par1 and par2, S. pombe cells become insensitive to the hyperactivation of spg1 GAP activity, which would normally inhibit cytokinesis and cause cell death.

par1 rescues spg1-106 lethality at high temperature:
The above observations prompted us to carry out further genetic analyses on the interaction between par1, par2, and other SIN pathway components. We first asked whether par1 par2 could suppress spg1-106 (previously called sid3-106). This mutant allele, identified from a genetic screen for cytokinesis mutants, causes cells to be unable to grow at 36° and has been shown to be a loss-of function allele of spg1 (BALASUBRAMANIAN et al. 1998 ). We generated a double mutant of par1 spg1-106 and tested its growth and morphological characteristics at both the permissive (25°) and nonpermissive (36°) temperatures for spg1-106. We found that while spg1-106 cells did not grow at 36°, a par1 spg1-106 double mutant grew as well as both the wild-type and par1 cells (Fig 2A). Thus, par1 is a strong suppressor of the temperature sensitivity of a spg1 loss-of-function mutant allele.

We then looked at the morphology of both the single and the double mutant strains. At 25°, par1 cells had a phenotype similar to that of par1 par2 cells with respect to morphogenesis, septum formation, and cytokinesis, the only difference being that the percentage of abnormal cells was lower and the phenotype was less severe (JIANG and HALLBERG 2000 ; data not shown). When grown at 25°, spg1-106 cells looked normal because spg1 function was largely intact (data not shown), and par1 spg1-106 cells displayed a par1-like phenotype (data not shown). When shifted to 36° for 4 hr, par1 cells looked the same as when they were grown at 25°, with many of them having visible septa (Fig 2B, arrows in i and ii indicate the septa). spg1-106 cells showed the distinctive sid phenotype: multinucleated, elongated cells without septum formation (Fig 2B, iii and iv). Ninety percent of par1 spg1-106 double mutant cells displayed a par1-like phenotype, with septa formed in the majority of the cells (Fig 2B, arrows in v and vi indicate the septa). Less than 10% of cells showed the sid phenotype (Fig 2B, arrowheads), suggesting that par1 cannot completely bypass the spg1-106 morphological defect. Nonetheless, the suppression allows the cells to continue division and growth, thus rescuing the lethality displayed by the spg1 mutant allele.

We tried to generate a par2 spg1-106 strain as well as a par1 par2 spg1-106 strain via mating and sporulation, but were unsuccessful because of the close linkage between par2 and spg1. This was confirmed by our tetrad analysis following sporulation as well as information from the S. pombe genome project (http://www.sanger.ac.uk/Projects/S_pombe/). However, since Par2p is 10 times less abundant than Par1p in the cell, and par1 and par2 can functionally substitute for each other (JIANG and HALLBERG 2000 ), we would predict that par2 alone could not suppress the spg1-106 defects and, if par2 were to be introduced into par1 spg1-106 cells, the sid phenotype might be completely rescued, similar to the results of the byr4-OP experiment.

par1 par2 cannot rescue mutants of the downstream components in the SIN pathway:
Because par1 par2 cells could rescue the effect of byr4-OP and par1 alone is a potent suppressor of spg1-106, we tested the genetic interactions between par1 par2 and mutants of the components that are shown to be downstream of spg1. Sid2p kinase is the most downstream component identified to date, and sid2-250 is a ts (temperature sensitive) allele that showed the sid phenotype when shifted to high temperature (BALASUBRAMANIAN et al. 1998 ). If par mutants can suppress the temperature sensitivity of this allele, then this would indicate that par1 and par2 normally act at or downstream of sid2 in this pathway.

We generated par1 sid2-250, par2 sid2-250, and par1 par2 sid2-250 strains and tested their growth and morphology at different temperatures. In contrast to the byr4-OP and spg1-106 results, par1, par2, and par1 par2 could not rescue the sid2-250 temperature sensitivity at 36° (Fig 3A). When observed microscopically, wild-type cells had normal septation and cytokinesis at both temperatures (Fig 3B, i and ii). sid2-250 cells appeared wild type at 25° (Fig 3B, iii), but had a typical sid phenotype when shifted to 36° for 5 hr (Fig 3B, iv). par1 par2 cells at both temperatures displayed the typical abnormalities associated with this strain (Fig 3B, v and vi). For example, cell a in Fig 3B, v, had a cell separation defect as well as double septa formed between two nuclei (arrows indicate the septa, which define a small compartment without a nucleus). par1 par2 sid2-250 cells at 25° showed a mixed morphology, with the sid phenotype showing some penetration (Fig 3B, vii, cells b and c had four nuclei but managed to form a septum, which is indicated by an arrow). When shifted to 36° for 5 hr, the majority of the triple mutant cells displayed a sid phenotype (Fig 3B, viii). As with par1 par2, single mutants of par1 or par2 could not rescue the morphological defects of sid2-250 cells. The above results indicate that par1 and par2 most likely act upstream of sid2 in the pathway.

To identify the gene(s) that par1 and par2 normally act upon in the SIN pathway, we crossed par1 par2 with mutants of the genes that are shown to be downstream of spg1 but upstream of sid2, namely, cdc7-24, cdc14-118, and sid1-239. All these mutants are temperature sensitive and display sid phenotypes at 36° (NASMYTH and NURSE 1981 ; BALASUBRAMANIAN et al. 1998 ). We generated strains with each mutant gene together with par1, par2, or par1 par2 and tested their growth and morphological properties at both 25° and 36°. The results are summarized in Table 2. We found that par1, par2, and par1 par2 could not rescue the temperature sensitivity or the sid phenotype of any of these mutants.

We also asked whether par1 par2 could rescue another sid mutant: sid4-SA1. sid4 has recently been cloned and characterized (CHANG and GOULD 2000 ). It encodes an essential protein and is required for the proper localization of all the components of the SIN pathway. We generated par1 sid4-SA1, par2 sid4-SA1, and par1 par2 sid4-SA1 strains. None of these strains grew at 36° (the nonpermissive temperature for sid4-SA1 cells; Table 2). When shifted from 25° to 36° for 5 hr and observed by microscopy, all of the above strains showed the sid phenotype (Table 2). Thus, sid4-SA1 defects at high temperature could also not be rescued by deletion of par genes.

par1 par2 could not rescue spg1-null lethality:
The above genetic analyses demonstrated that par1 par2 could suppress the byr4-OP effect and par1 alone could strongly suppress a spg1-106 ts allele, but deletion of these PP2A subunits could not rescue any of the mutants in the downstream components of the SIN pathway. This places the action of par1 and par2 in this pathway on either cdc7 or spg1.

The activity of the SIN pathway is essential because each of the known genes in this pathway is essential for vegetative growth. In par1 par2 cells, two possible scenarios exist: (1) Cdc7p is constitutively active causing a hyperactive SIN pathway, and this hyperactivity of Cdc7p does not require spg1 function. In this case, the defects of spg1-106 as well as byr4-OP are suppressed because both affect only the event upstream of cdc7, namely, spg1 activity. spg1 could become hyperactive in par1 par2 cells, leading to the increased activity of this pathway. If the first situation is correct, then par1 par2 cells would not need spg1 to survive, and spg1 can be deleted from this strain background. This follows from the fact that spg1-null lethality could be rescued by overproducing Cdc7p kinase (SOHRMANN et al. 1998 ). However, if the latter alternative is the case, then par1 par2 would not be expected to bypass the spg1 essentiality.

We tested these alternatives in the following way: par1::his3⁺ par2::LEU2 cells were crossed with a wild-type strain of the opposite mating type to generate a diploid par1 par2 heterozygous strain. One copy of most of the spg1 coding sequence was replaced with a ura4⁺ marker in this strain (see MATERIALS AND METHODS). The disruption of spg1 in the genome was verified by Southern blot analysis. Subsequent sporulation and tetrad analyses of the resulting diploid strain yielded only one or two viable spores in each tetrad. We were able to obtain viable spores with wild-type, par1, par2, or par1 par2 genotypes, but never isolated any with the spg1 genotype. The viable spores were all ura^-. The fact that out of almost 200 tetrads analyzed, not a single ura⁺ spore survived confirmed that spg1 is essential for vegetative growth, and this essential function cannot be bypassed by inactivating par1 and par2 by deletion.

Cdc7p levels and localization in par mutants:
In par1 and par1 par2 strains, a small percentage of cells displayed a multiseptation phenotype, resembling the mutant phenotype of cells in which the spg1-cdc7 signaling cascade is hyperactive. Our genetic analyses showed that par1 par2 can rescue certain strains in which the SIN pathway is kept inactive (spg1-106 cells presumably have defective spg1 activity, and overproducing byr4 keeps spg1 in its inactive GDP-bound state). These results are consistent with our hypothesis that in par1 par2 cells, this pathway is hyperactive. Since our epistatic genetic data indicated the necessity for par1 and par2 function at Cdc7 or upstream, we examined further the effects of deleting these two genes on Cdc7p.

The fact that overproducing Cdc7p could rescue spg1-null lethality prompted us to ask whether par1 par2 cells have a higher level of Cdc7p compared to that in wild-type cells, as an increased amount of Cdc7p could explain both the phenotype of par1 par2 cells and its suppression of a loss-of-function mutant allele of spg1. To do this, we crossed a par1 par2 strain with a WT-Cdc7HA strain (a gift from Dr. C. Albright) and generated par1 Cdc7HA, par2 Cdc7HA, and par1 par2 Cdc7HA strains. These strains contained Cdc7HA in their genome as the sole copy of cdc7. We grew these cells at 30°, extracted the total proteins, and quantitated Western blots probed with anti-HA antibody. This analysis showed that the level of Cdc7-HAp in par1, par2, or par1 par2 cells was the same as that in wild-type cells (data not shown). Thus, it is not the elevated Cdc7p level that accounts for the capacity of par1 to suppress spg1-106.

While the level of Cdc7p does not change in a par1 par2 mutant background, it may be that the intracellular localization of Cdc7p has been altered. It has been shown that neither the protein level nor the kinase activity of Cdc7p changes throughout the cell cycle (SOHRMANN et al. 1998 ). The regulation of Cdc7p activity occurs via its localization. In interphase cells, no discrete Cdc7p localization is observed. As cells enter mitosis, Cdc7p translocates to both SPBs in an spg1-dependent manner. During late anaphase, Cdc7p disappears from one SPB, but is maintained at the other SPB until cell separation (SOHRMANN et al. 1998 ). While the molecular bases of this asymmetric SPB localization of Cdc7p is unknown, SPB asymmetry appears essential in ensuring that only a single septum is formed. This is demonstrated by the fact that, in cdc16-116 and byr4^- mutants that undergo repeated rounds of septation, Cdc7p is found at both SPBs in late mitosis (SOHRMANN et al. 1998 ; LI et al. 2000 ). Therefore, the symmetric localization of Cdc7p to both SPBs during late anaphase is a strong indication of the hyperactivity of this signaling pathway.

To determine the subcellular localization pattern of Cdc7p in our mutant cells, we obtained a Cdc7-GFP strain (D. McCollum) and crossed it with a par1 par2 strain, generating par1 Cdc7-GFP and par2 Cdc7-GFP strains (see MATERIALS AND METHODS). In these strains, Cdc7-GFP is again the only copy of the cdc7 in the genome. We grew cells at 30° and observed the Cdc7-GFP signal using fluorescence microscopy. In wild-type cells, as reported (SOHRMANN et al. 1998 ), we did not observe any discrete Cdc7-GFP signal in interphase cells (cell a in Fig 4A, iii and iv). However, Cdc7-GFP was seen on both SPBs in early mitosis (cell b in Fig 4A, i and ii) and on only one SPB in late anaphase cells (cells c, d, and e in Fig 4A, i–iv). We did not observe any late mitotic cells (where nuclei have been well separated but a septum is not yet visible) with Cdc7GFP on both SPBs.

View larger version (90K): In this window In a new window Download PPT slide   Figure 4. Localization of Cdc7GFP in WT, par1, and par2 cells. (A) WT-Cdc7GFP and par1-Cdc7GFP cells were grown in YE5S media at 30° until early log phase and then examined with a fluorescence microscope. (i–iv) WT-Cdc7GFP; (v–viii) par1-Cdc7GFP. (i, iii, v, and vii) Cdc7GFP; (ii, iv, vi, and viii) DIC images. a, b, c, d, e, arrows, arrowheads, and broken arrow indicate cells described in the text. (B) Quantitation of Cdc7GFP localization. WT-Cdc7GFP, par1-Cdc7GFP, and par2-Cdc7GFP cells were grown in YE5S media at 30° until early log phase and then examined by fluorescence microscopy. The fourth and the fifth column also include cells with a visible septum and that showed Cdc7GFP on SPBs. For each strain, at least 300 cells were scored. The small solid circles represent the SPB, and the shaded ovals represent nuclei.

We then examined Cdc7-GFP in par1 cells. Again, we saw no interphase cells with discrete Cdc7-GFP staining (cells indicated by asterisks in Fig 4A, vii). Early mitotic cells showed double SPB staining, and some late anaphase cells had the normal asymmetric SPB localization (cell indicated by a broken arrow in Fig 4, v). However, a portion of par1 cells had both SPBs stained with Cdc7-GFP even in late anaphase (Fig 4A, vii is a representative field showing this phenotype). Most of these cells had equal intensity of the Cdc7-GFP signal at the two SPBs (cells indicated by arrows in Fig 4A, v and vii), while some of them showed stronger staining on one SPB than the other (cells indicated by arrowheads in Fig 4A, vii). In par2 cells, the localization pattern of Cdc7-GFP appeared to be normal, and we did not observe any late anaphase cells with Cdc7-GFP signals on both SPBs (data not shown).

The quantitation of the above results is given in Fig 4B. Approximately one-quarter (23.9%) of wild-type cells showed Cdc7-GFP staining on SPB(s), with 1.7% being in early mitosis (one nucleus with two SPBs stained) and 22.2% in late anaphase (binucleated with only one SPB stained). par2 cells have a similar distribution when compared with wild-type cells, with 23.9% of cells showing Cdc7-GFP signal at the SPB(s). In par1 cells, a total of 29.5% showed Cdc7-GFP SPB staining, with 3.3% in early mitosis, 15.0% having only one SPB signal in late anaphase, and 10.4% having two SPBs stained in late anaphase. The percentage of the symmetric localization of Cdc7-GFP on SPBs in late anaphase corresponds to the percentage of par1 cells that showed the multiseptation phenotype (JIANG and HALLBERG 2000 ). This result shows that in par1 cells, the Cdc7p localization pattern is indeed altered, and this alteration presumably reflects the hyperactivation of the SIN pathway, which in turn gives rise to the mutant phenotype.

We attempted to generate a par1 par2 Cdc7-GFP strain but were not successful. However, when we used indirect immunofluorescence to look at Cdc7-HAp localization in par mutant strains, we obtained results similar to those we observed using Cdc7-GFP. In a small percentage of both par1 and par1 par2 cells, Cdc7-Hap was localized to both SPBs in binucleated cells (data not shown), and there did not appear to be any qualitative difference between par1 and par1 par2 with regard to Cdc7-HAp localization. Quantitation of those cells was not as reliable as that using Cdc7-GFP (data not shown). Given the fact that Par1p is the major form of the PP2A B' subunit in S. pombe (JIANG and HALLBERG 2000 ), we speculate that there would not be any qualitative difference between par1 par2 and par1 cells with respect to the localization of Cdc7GFP, but we might see a slightly higher percentage of symmetric Cdc7p localization in par1 par2 cells compared to par1 cells.

par1 rescues spg1-106 by restoring Cdc7p localization:
We have demonstrated that in late mitotic par1 par2 cells, Cdc7p was localized to both SPBs, and this leads to a hyperactive SIN pathway and, in turn, the multiseptation phenotype. As deletion of par1 suppresses the mutant growth phenotype of spg1-106 cells, an obvious question is: does it do so by affecting Cdc7p localization in these cells?

Accordingly, we examined Cdc7-GFP localization in spg1-106 cells. At 25°, these cells have normal Cdc7-GFP localization pattern, with 22.1% cells having Cdc7-GFP on the SPB (Fig 5A, i and ii; Fig 5C). However, when shifted to 36° for 5 hr, only 3.6% of spg1-106 cells showed Cdc7-GFP signal on the SPB (Fig 5C), and a typical sid phenotype was observed (Fig 5A). This is consistent with the notion that without Cdc7p SPB localization, the SIN pathway is turned off, so no septum is formed and the cells eventually die.

View larger version (194K): In this window In a new window Download PPT slide   Figure 5. Localization of Cdc7GFP in spg1-106 and par1 spg1-106 cells. spg1-106-Cdc7GFP cells (A) and par1 spg1-106-Cdc7GFP cells (B) were grown in YE5S media at 25° until early log phase; half were then examined with a fluorescence microscope and the other half were shifted to 36° and grown for another 5 hr before microscopic observation. (A, i and ii) spg1-106-Cdc7GFP cells at 25°; (iii and iv) spg1-106-Cdc7GFP cells at 36°. i and iii show the Cdc7GFP, while ii and iv show the corresponding DIC pictures. (B, i–iv) par1 spg1-106-Cdc7GFP cells at 25°; (v–xii) par1 spg1-106-Cdc7GFP cells at 36° for 5 hr. i, iii, v, vii, ix, and xi show the Cdc7GFP, while ii, iv, vi, viii, x, and xii show the corresponding DIC pictures. (C) The quantitation of the results in A and B. The third and the fifth columns also include cells with visible septum that showed Cdc7GFP on SPBs. For each strain, at least 200 cells were scored. The small solid circles represent the SPB, and the shaded ovals represent nuclei.

We then examined Cdc7-GFP in par1 spg1-106 cells at 25° and 36°, respectively. At 25°, these cells showed a par1-like morphology, and Cdc7-GFP was detected on SPBs in 28.6% of cells (Fig 5C). Cdc7-GFP was seen on either one or two SPBs in binucleated cells that did not yet have visible septa (Fig 5B, arrows and arrowheads in i and iii). When par1 spg1-106 cells were shifted to 36° for 5 hr, SPBs still showed Cdc7-GFP staining (Fig 5C). Cdc7GFP was found on either one or two SPBs in 27.4% of those cells (arrows in Fig 5B, v–xii). This is in contrast to spg1-106 cells at 36° where <4% of the cells showed Cdc7-GFP at a SPB (compare Fig 5B v–xii with Fig 5A, iii and iv). Furthermore, septum formation was clearly visible in these cells, and a minor fraction even had multiple septa (e.g., cell a in Fig 5B, xi and xii).

In summary, our data showed that spg1-106 cells at 36° are unable to localize Cdc7p to the SPB. However, deleting par1 rescues spg1-106 morphological and growth defects, and it does so by restoring the localization of Cdc7p onto SPBs, thus allowing proper signal transduction to the downstream components in the SIN pathway. We take this as an indication that the par1 suppression effect is specific to the SIN pathway and is not achieved through an alternative pathway.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Cytokinesis is a crucial event in cell division. It needs to occur at the right time, in the right place, and only once during a single cell cycle. As such, it must be highly regulated. We have presented evidence showing that par1- and par2-directed PP2A activity is required for proper functioning of the S. pombe SIN pathway. This activity is required at the level of the Ras-like GTPase, Spg1p, whose function is situated at the beginning of this signaling pathway. It appears that the role played by par1 and par2 is to negatively regulate SIN, thereby ensuring that multiple rounds of septation do not occur.

The suppression of spg1-106 by par1 is specific to the restoration of the SIN pathway function:
As a strain with a loss-of-function ts allele of spg1, spg1-106 cells displayed a typical sid phenotype when shifted to 36°, their restrictive temperature. Deleting par1 in this strain strongly suppressed both the morphological and growth defects of spg1-106 cells. When par1 spg1-106 double mutant cells were shifted to 36°, they grew as well as both par1 and wild-type cells, and most of the cells showed a par1-like morphology, with the sid phenotype occurring in <10% of the cells. It is still possible that this suppression is nonspecific to the SIN pathway, but due rather to the restoration of some other aspects of spg1 function. Although the nature of the spg1-106 mutant is not clear, we showed that Cdc7p failed to localize to SPBs in this strain at 36°, and this failure of localization apparently leads to the sid phenotype. By contrast, in par1 spg1-106 cells at 36°, Cdc7p localized to the SPBs, which would account for the suppression of the spg1-106 defects. Furthermore, no countersuppression was seen in this double mutant; i.e., spg1-106 did not suppress par1 defects. The double mutant cells did not grow at 37° (which is the restrictive temperature for par1 cells; JIANG and HALLBERG 2000 ; data not shown), nor would they grow in high salt, high ethanol, or hypertonic media (par1 cells are sensitive to these stress conditions; JIANG and HALLBERG 2000 ; data not shown). Also, the morphology of the double mutant was par1-like, with many cells still showing the septum positioning and cell separation defects. Finally, the fact that another independent event, byr4-OP, could also be suppressed by deletion of par1 and par2 strengthens our conclusion that par1 rescues spg1-106 defects through specific restoration of the SIN pathway function.

Point of action of par genes in the regulation of the SIN pathway:
We tested genetic interactions between par deletions and mutants of the SIN pathway components and found that par1 par2 rescues the byr4-OP defects, par1 rescues spg1-106 ts allele, but par mutants could not rescue cdc7-24, sid1-239, cdc14-118, or sid2-250, all of which are ts mutants of the components downstream of spg1 in this pathway. We also found that Cdc7-GFP was altered in par mutant cells, consistent with par genes exerting their effect at or upstream of cdc7. In a sid4-SA1 mutant strain, the localization of all of the SIN pathway components to SPBs is abolished, so it is not surprising to find that par deletions could not suppress this allele. On the basis of our epistasis analyses as well as cytological studies, we concluded that par genes most likely act at or upstream of cdc7 in the SIN pathway.

The exact point of regulation by par genes is not yet clear. The fact that par mutants suppressed two independent mutants upstream of cdc7 led us to hypothesize that par1 and par2 might act upon cdc7 and that this action is independent of spg1. If this were true, then in par1 par2 cells, spg1 function would become dispensable and, regardless of what happens to spg1, Cdc7p could still be localized to SPBs. This would explain why these two strains are suppressed by par mutants. Also, if this were true, we should be able to eliminate spg1 in par1 par2 cells. The results of such experiments, however, showed that spg1 function was indispensable for par1 par2 cell survival. Taken together with the fact that Cdc7p levels did not change in par mutants, these data indicate that the par genes probably regulate Cdc7p localization through a component upstream of cdc7 in the SIN pathway.

In a small portion of par mutant cells, Cdc7p was found at both SPBs in late anaphase. Since Cdc7p normally is recruited to the SPB via the active Spg1p, we hypothesize that in par mutant cells, it is the increased activity of Spg1p that leads to the symmetric Cdc7p localization. We suggest that excess Spg1p activity in par mutants leads to a breakdown of the normal asymmetry of GDP-Spg1p and GTP-Spg1p localization in some cells, thus accounting for our results.

As a member of the Ras GTPase superfamily, the nucleotide state of Spg1p is regulated by a two-component GAP and a yet-to-be-identified GEF. The two-component GAP for Spg1p consists of Byr4p and Cdc16p (FURGE et al. 1998 ), which is localized to the SPBs that do not contain Cdc7p. In byr4^- mutants, Cdc7p localizes to interphase SPBs and only symmetrically localizes to mitotic SPBs. In contrast, byr4-OP prevents Spg1p and Cdc7p localization to SPBs (CERUTTI and SIMANIS 1999 ; LI et al. 2000 ). These results strongly suggest that Byr4p localization to SPBs maintains Spg1p in an inactive form and that loss of Byr4p from mitotic SPBs increases the fraction of active Spg1p and consequently Cdc7p binding. Thus, the symmetric Cdc7p localization in par mutants could be caused by a decreased GAP activity, either by a decreased level of Byr4p or Cdc16p or by their decreased localization to SPBs.

The symmetric Cdc7p localization in par mutants could also be caused by an increased GEF activity, if in fact a GEF for Spg1p exists in S. pombe. A well-conserved signaling pathway, the mitotic exit network, has recently been identified in the budding yeast (HOYT 2000 ; MCCOLLUM and GOULD 2001 ). Tem1p, a GTP-binding protein, is the S. cerevisiae homolog of Spg1p (SHIRAYAMA et al. 1994 ; SHOU et al. 1999 ), while LTE1 encodes the GEF for Tem1p (BARDIN et al. 2000 ). No sequence homologous to LTE1 has been identified in the S. pombe genome as yet. Interestingly, when LTE1 is overproduced in wild-type S. cerevisiae cells, only 10% show a premature exit of mitosis (BARDIN et al. 2000 ). Considering that a similar percentage of par mutant cells show a multiseptation phenotype, it is certainly possible that par genes normally inhibit the GEF for Spg1p and that in par mutants the GEF activity is increased, resulting in a hyperactive Spg1p. It remains to be determined if a Spg1p GEF actually exists in S. pombe.

All of the above possible points of regulation by par genes could be either direct or indirect. It is possible that Par1p, Par2p, or both interact directly with some of the SIN components such as Spg1p, Byr4p, Cdc16p, etc., to regulate their activities and/or localizations. While all these SIN components localize to the SPB, Par1p and Par2p were not seen at the SPBs in our immunofluorescence assays (JIANG and HALLBERG 2000 ). However, we cannot exclude the possibility that a small fraction of Par1p or Par2p is associated with the SPB. As Par1p is very abundant in the cell, its SPB localization might have been masked by the overall staining, or the fixation procedure in the indirect immunofluorescence assay somehow disrupted the Par proteins' association with the SPB. Interestingly, the S. cerevisiae homolog of the par genes, Rts1p (SHU et al. 1997 ), has been observed at the SPB in premitotic cells (M. GENTRY and R. HALLBERG, unpublished data). On the other hand, it is also possible that Par proteins regulate the SIN pathway through some mediator proteins without being localized to the SPBs themselves.

The hyperactivity of the SIN pathway in par mutant cells:
As the major signaling pathway that regulates the onset of septum formation and cytokinesis in fission yeast, the SIN pathway has to be tightly monitored so that its activity can be carefully controlled. Both the inactivation and the hyperactivation of this pathway are lethal to the yeast cell. When the SIN pathway is inactivated, as seen in all of the sid mutants (spg1-106, spg1-B8, cdc7-24, sid1-239, cdc14-118, sid2-250, sid4-SA1, cdc11-123) as well as in byr4-OP cells, no septum is formed and cells cannot divide after nuclear division. On the other hand, hyperactivation of this pathway causes uncontrolled septation, as observed in byr4^-, cdc16-116, spg1-OP, and cdc7-OP cells where multiple septa were formed in a single cell cycle, which also leads to cell death.

Although the biochemical characterization of this pathway is not complete, genetic as well as cytological data support the notion that the control over the intracellular localization of the individual components plays a pivotal role in transducing the signal along this pathway. Cdc7p kinase is a good example of this. Although the kinase activity is required for its function, Cdc7p is regulated neither by the fluctuation of the protein level nor by the change in its kinase activity, but rather by its localization within the cell (SOHRMANN et al. 1998 ). Normally, in late anaphase Cdc7p is localized to only one SPB, ensuring that the signal for septation is transduced to the downstream components in the SIN pathway, and only once. If Cdc7p does not associate with the SPB, then no septation occurs. However, if Cdc7p localizes to both SPBs, then multiple rounds of septation are observed. The localization pattern of Cdc7p is then a very good, if not the best, indication of SIN pathway activity available at present.

The multiseptation phenotype observed in par mutant cells prompted us to test whether this is caused by a hyperactive SIN pathway and, as expected, we found that Cdc7p indeed localized to both SPBs in late anaphase cells. Furthermore, the percentage of cells showing mislocalized Cdc7p corresponds to the percentage of par mutant cells that show the multiseptation phenotype. We concluded that in par mutant cells the SIN pathway becomes abnormally hyperactive. This could be caused by one of two mechanisms. Either the activity of SIN becomes higher than that in wild-type cells at a certain point in the cell cycle or the activity of SIN is normal but is maintained for a prolonged period of time. The latter could also be termed a "failure of inactivation." Although we cannot differentiate these two possibilities at present, either could cause a hyperactive SIN pathway during late anaphase when it should be low in wild-type cells.

In an attempt to find other evidence that the SIN pathway is hyperactive in par mutant strains, we examined the localization of Sid2-GFP in par mutant cells. Sid2p is the most downstream component in the SIN pathway identified to date, and it is seen on SPBs throughout the cell cycle and transiently at the cleavage site during septation. The kinase activity of Sid2p is required for its function, and it peaks during medial ring constriction and septation (SPARKS et al. 1999 ). We created a par1 par2 Sid2-GFP strain and localized the Sid2-GFP signal. As seen in WT cells, Sid2-GFP in par1 par2 cells was localized to SPBs as well as the cleavage site during septation (in multiseptated cells, Sid2-GFP was only seen with one septum per cell; data not shown). However, both the localization and the kinase activity of Sid2p remain to be determined in mutant strains where the SIN pathway is hyperactive, such as cdc16-116, spg1-OP, and cdc7-OP cells. It is possible that in these cells Sid2p localization is normal but its kinase activity is higher, causing the hyperactivation of the pathway. Once the physiological substrates of Sid2p have been identified, then it should be possible to test whether the Sid2p in vivo kinase activity changes in par mutant cells.

Another question we addressed was: are the par-directed functions on the SIN pathway dosage dependent? spg1 and cdc7 are both dosage-dependent activators of this pathway, as shown by the fact that deletion or mutation of either gene causes the inactivation of septation, while overproduction leads to multiple septation (FANKHAUSER and SIMANIS 1994 ; SCHMIDT et al. 1997 ). Since par mutant cells showed a multiseptation defect, we wondered whether a high dosage of par genes would have the opposite effect. We thus overproduced either par1 or par2 in wild-type cells (both genes are under the control of the nmt1 promoter). Overproducing par1 was lethal, while overexpression of par2 had no effect on cell growth (data not shown). We also examined the morphology of these cells when they were shifted to media without thiamine for 16 hr to induce par1 or par2 expression and did not observe any obvious sid phenotype (data not shown). Since we showed that par1 is involved in a number of other cellular functions (JIANG and HALLBERG 2000 ), it is not surprising to see that its overproduction is detrimental to the cell. However, this might be achieved through some cellular events other than the inhibition of the SIN pathway or, alternatively, the sid effect might simply be masked by the other cellular events.

	ACKNOWLEDGMENTS

We are indebted to Dr. C. Albright for providing us with a collection of yeast mutants as well as Rep41-byr4 plasmid and anti-Byr4p antibody. We thank Dr. D. McCollum and Dr. S. Forsburg for yeast strains and plasmids. We thank Scott Erdman for his comments on the manuscript and also thank members of the Hallberg lab and members of the Upstate Medical University/Syracuse University yeast group for their thoughtful criticisms and suggestions. This research has been supported by National Science Foundation grant MCB-9603733 (R.L.H.).

Manuscript received March 28, 2001; Accepted for publication May 7, 2001.

	LITERATURE CITED

TOP