In budding yeast, contraction of the actomyosin ring and formation of the primary septum at cytokinesis depend on each other

In budding yeast, cytokinesis starts with the invagination of the plasma membrane at the neck between the mother and daughter cell. Concomitantly, the polysaccharide chitin is extruded into the invagination as the latter advances toward the center of the mother-daughter channel (Fig. 1B) (Cabib et al., 1996). Eventually, this process results in pinching off of the membranes and separation of the two cells by a chitin disk, the primary septum (Fig. 1C). Subsequently, secondary septa are added from both mother and daughter cells, so that the complete septum has a trilaminar structure, with the chitin disk in the middle (Fig. 1D). The formation of primary septum chitin is catalyzed by one of the three yeast chitin synthases, ChsII. Its putative catalytic subunit, Chs2p, is encoded by the CHS2 gene (Silverman et al., 1988). Null mutants of CHS2 have abnormal and very thick septa but are able to survive (Bulawa and Osmond, 1990; Shaw et al., 1991).

Recently, it has been found that the invagination of the plasma membrane at cytokinesis is brought about by the contraction of an actomyosin ring (Epp and Chant, 1997; Lippincott and Li, 1998; Bi et al., 1998), as is the case for division of mammalian cells. Mutants in MYO1, which encodes the type II myosin required for function of the contractile ring, show abnormal septation (Rodríguez and Paterson, 1990) but are still able to complete cytokinesis (Bi et al., 1998).

Several aspects of the septation process remain obscure: for instance, can synthesis of the primary septum take place in the absence of contractile ring closure and vice versa? Is at least one of these two processes required for viability? What is the nature of the remedial septation that takes place in chs2 and myo1 mutants?

We have reinvestigated the events that occur during cytokinesis in different mutants and have come to the conclusion that contractile-ring closure and formation of the chitin primary septum are interdependent phenomena. The cell can survive even when both mechanisms are not functional. The remedial septa formed under those conditions are probably a modification of the secondary septa laid down at the end of normal cell division. Another chitin synthase, ChsIII, is required for the successful execution of the repair mechanism.

Yeast strains are listed in (Table 1). Culture media for Escherichia coli were described in Ausubel et al. (Ausubel et al., 1994). E. coli DH10B (Gibco-BRL) was used as plasmid host. Yeast cells were grown either in YEPD (2% glucose, 2% peptone, 1% yeast extract) or in SD (2% glucose, 0.7% Difco Yeast Nitrogen Base). Generation time was determined by monitoring the OD₆₆₀.

Yeast transformation was performed according to Gietz et al. (Gietz et al., 1992). For kanamycin selection, the standard transformation protocol (Ausubel et al., 1994) was followed. After heat shock, the cells were suspended in 5 ml of YEPD medium and incubated overnight at room temperature before being spread on YEPD plates containing 250 mg/L G418 (Gibco-BRL). After a three-day incubation at 30°C, the colonies were replicated onto fresh G418 plates and incubated for another two days at 30°C before resistant colonies were picked.

General methods of DNA manipulation were used as described in Ausubel et al. (Ausubel et al., 1994) except when following procedures recommended by the manufacturers of the enzymes and DNA isolation kits used (New England Biolabs; Qiagen). To obtain URA3-containing vectors with genes SPA2-3xHA, MYO1-c-myc and MYO1-GFP, a 2.9 kb PvuI fragment from pRS316 was ligated to the large fragment of the PvuI-digested plasmids pBU4 (M. Snyder), pLP7 and pLP8 (Lippincott and Li, 1998), respectively. This exchanged the LEU2 markers of the original plasmids with the URA3 gene and yielded plasmids pMS53, pMS54 and pMS55 (Table 2).

To construct pEC28 for the detection of synthetic lethality with chs3 mutants, plasmid pHV7 (Valdivieso et al., 1991) containing the CHS3 gene was cut with ClaI and PvuII. The digest was treated with Klenow fragments to remove the overhang; the 5.2-kb fragment containing CHS3 was isolated and ligated to the ADE2-harboring vector pRS412 (Brachmann et al., 1998) that had been cut with SmaI. For construction of pMS52, used in the detection of synthetic lethality with chs2 mutants, pMS11, containing the CHS2 gene in the BamHI/HindIII sites of pRS316 ,was cut with KpnI/BamHI. The 3.9-kb fragment was purified and ligated to KpnI/BamHI-digested pRS412.

Disruption of CHS2 and CHS3 was performed according to Crotti et al. (Crotti et al., 2001). For disruption of MYO1, a 6.2-kb fragment containing the MYO1 ORF and 200 bp each of 5′ and 3′ flanking sequences was amplified by PCR from chromosomal DNA with oligonucleotides 5′-CCCTCCTTGGAATTTGG-3′ and 5′-GACTTTATTTCCGCTGC-3′. The fragment was then treated with Klenow fragments and T4 polynucleotide kinase and ligated to EcoRV-linearized pBluescript II KS+ vector (Stratagene). From the resulting plasmid a 3.3-kb MfeI fragment was cut out and replaced with a blunt-ended 1.0-kb TRP1 or 1.2-kb URA3 fragment, both generated by PCR with subsequent Klenow filling-in and phosphorylation. This yielded plasmids pMS48 and pMS50, respectively. Plasmids were cut with EcoRI/ClaI (pMS48) or EcoRI (pMS50) before transformation of yeast strains.

PCR-based C-terminal tagging of Bud3p, Bud4p and Chs2p was performed as described by Longtine et al. (Longtine et al., 1998). Primers consisted of a 40-bp sequence homologous to the target gene (followed, in the case of CHS2, by a linker encoding a stretch of 8 glycines) and a 20-bp sequence homologous to the pFA6a-GFP(S65T)-kanMX6 plasmid. The oligonucleotide sequences were as follows:

BUD3UP: 5′-GAACTGCGCTGTTGGTGGCCCAGAGAAACTGAAATTTTATCGGATCCCCGGGTTAATTAA-3′

BUD3DOWN: 5′-TTGCATTAAATTAAAAAGAAAAAAAAAATCAATAAAACACGAATTCGAGCTCGTTTAAAC-3′

BUD4UP: 5′-GACTCGAACAACCGGTCATAATTTGAAGCAAGATTTTAATCGGATCCCCGGGTTAATTAA-3′

BUD4DOWN: 5′-TCATCAGATTATATGCTGTTTTCATTCCATTAATCACCTTGAATTCGAGCTCGTTTAAAC-3′

CHS2UP: 5′-CTCTAAATTAGACTTACCAAATGTTTTCCACAAAAAGGGCGGTGGAGGTGGAGGTGGAGGTGGACGGATCCCCGGCTTAATTAA-3′

CHS2DOWN: 5′-AAAGAGGGAATGACGAGAAATTAGCTGAAAAATACTGGCAGAATTCGAGCTCGTTTAAAC-3′

PCR conditions were: one minute at 94°C, then 30 cycles of one minute at 94°C, one minute at 55°C, two minutes at 72°C, followed by 15 minutes at 72°C with Takara ExTaq polymerase (Takara) and the supplied buffers. After transforming yeast with the resulting fragments, G418-resistant transformants were selected (see above).

Cells were fixed and embedded essentially as described by van Tuinen and Riezman (van Tuinen and Riezman, 1987). Aliquots of cells were washed in 0.1 M sodium phosphate buffer, pH 7.2 (PB) and fixed by suspension in 3% formaldehyde with 0.5% glutaraldehyde for two hours at room temperature. The cells were subsequently washed twice in PB, treated for 30 minutes with aqueous 1% metaperiodate, rinsed in PB and quenched for 30 minutes in 50 mM NH4Cl in PB. After rinsing in PB, the cells were enrobed in low melting point agarose, cut into small blocks and dehydrated in cold ethanol. The cells were embedded in LR White resin (Electron Microscopy Science). Thin sections were stained in aqueous 1% uranyl acetate, followed by lead citrate and carbon-coated before examination in the electron microscope.

GFP fluorescence was observed with a Zeiss Axioskop microscope equipped with a Chroma 41018 filter set (exciter 470 nm, dichroic 495 nm, emitter 500 nm). To enhance fluorescence intensity, cultures were saturated with oxygen for 20 minutes at room temperature before observation.

For assessment of synthetic lethality, segregation of an ADE2-containing plasmid from strains carrying an ade2-101 mutation was monitored by the appearance of red sectors. The screen for mutants synthetically lethal with chs3::LEU2 was performed in strain ECY46-4-1B complemented by a plasmid harboring the CHS3 gene in addition to the ADE2 marker (pEC28). Correspondingly, for detection of synthetic lethality with chs2::TRP1, YMS11 was complemented by a plasmid containing ADE2 and CHS2 (pMS52). After disruption of the genes under investigation, synthetic lethality with either chs2 or chs3 mutations leads to an inability of the strain to segregate the ADE2-containing plasmid, hence yielding white colonies devoid of red sectors when grown on medium containing 8 mg/L of adenine.

For the determination of budding pattern, strains were grown overnight in YEPD medium at 30°C to 10⁷ cells/ml, then spread on YEPD plates at room temperature. Cell pairs were then separated with the use of a micromanipulator, and the position of emerging new buds was recorded at 30-minute intervals.

0.1-1.0 L cultures of yeast were grown to 10⁷ cells/ml, centrifuged at 3,000 g and washed with ice-cold water. Cell pellets were frozen at –80°C. 50 μl of wet cells were mixed with 100 μl of water and 150 μl of 2× sample buffer (63 mM Tris-HCl, pH 6.8, 10% glycerol, 2% SDS [sodium dodecyl sulfate], 0.0025% bromphenol blue) with 1mM phenylmethylsulfonyl fluoride and 1% 2-mercaptoethanol. After heating the cell suspension at 95°C for five minutes, 200 μl glass beads were added and the cells were broken by vortexing for three 30-second periods. The liquid was aspirated with a Pasteur pipette and centrifuged at 13,000 g for two minutes. Samples of the supernatant were loaded on the polyacrylamide gel. In the case of Pma1p, both the addition of 2-mercaptoethanol and the heating at 95°C were omitted.

40 μg of protein were loaded on Tris-glycine gels (thickness 1 mm, 4-12% or 4-20% polyacrylamide, Invitrogen). The electrophoresis was run for two hours in 25 mM Tris, 192 mM glycine and 0.1% SDS at 125 V. Proteins were transferred onto PVDF membranes (Invitrogen) at 30 V for two hours in 12 mM Tris, 96 mM glycine, 0.1% SDS and 20% methanol. After transfer, membranes were blocked for one hour in 5% nonfat dry milk in TTBS/0.9% NaCl (0.1 M Tris chloride, pH 7.5, 0.9% NaCl, 0.1% Tween 20) and incubated overnight at 4°C in 20 ml of the same buffer, containing 10 μl of one of the following antibodies: a) anti-c-myc 9E10 (Santa Cruz Biotechnology) for detection of Myo1-c-mycp; b) anti-HA (Roche) for detection of Spa2-HAp; c) anti-Hsp104p (Stressgen); d) anti-Rho1p (Drgonová et al., 1999); e) anti-Cdc42 y-191 (Santa Cruz Biotechnology); f) anti-Bud3p (S. Sanders); g) anti-Bud4p (S. Sanders); or h) 2 μl of anti-Pma1p (C. Slayman).

The blots were washed three times for 15 minutes with TTBS/0.5M NaCl (0.1 M Tris-chloride, pH 7.5, 0.5 M NaCl, 0.1% Tween 20) and incubated for two hours at room temperature with these secondary antibodies: 10 μl/20 ml of anti-mouse IgG-POD/Fab fragments-HRP (Roche) for a) and b) or 2 μl/20 ml of anti-rabbit-HRP (Pierce) for c)-h) in TTBS/0.9% NaCl. The blots were then washed again three times for 15 minutes each time with TTBS/0.5 M NaCl, rinsed with water and developed with the ECL detection kit (Pierce) and Kodak BioMax ML film.

As mentioned above, cytokinesis in budding yeast starts with invagination of the plasma membrane and simultaneous deposition of chitin in the invagination (Fig. 2A). Continuation of this process results in the pinching off of the membrane, while a thin chitin disk, the primary septum, separates the mother and daughter cell (Fig. 2B). Subsequent deposition of secondary septa, both from the mother and daughter side, generates the complete, trilaminar septum (Fig. 2C) [for electron microscope images showing in detail the steps in septum formation see Cabib et al. (Cabib et al., 1974)]. In chs2 mutants, which are deficient in chitin synthase II (the enzyme responsible for the formation of primary septum chitin), cytokinesis does occur but in an aberrant manner. No localized invagination of the membrane can be seen. Rather, a wide area of the cell wall at the neck between the mother and daughter cell thickens gradually, squeezing the membrane (Fig. 2D) and finally causing its pinching off. In this case, the closure of the channel is an irregular process, as shown by the lacunae that often remain in the thick septa formed in the chs2 strains (Fig. 2E) (Shaw et al., 1991). Note that although the cell wall pushes the plasma membrane in, the wall growth occurs by secretion and extrusion of material from the interior of the cell through the membrane itself.

Cells defective in Myo1p, an essential component of the contractile ring (Lippincott and Li, 1998), are similarly able to complete cytokinesis (Bi et al., 1998) despite abnormalities in septation (Rodríguez and Paterson, 1990). In fact, we have examined myo1Δ and chs2Δ mutants by light and electron microscopy and found their cytokinesis defect to be indistinguishable. Both mutants are characterized by the formation of large clumps (Fig. 3), inward growth of cell walls at the neck (Fig. 2D,F,G) and thick septa, occasionally enclosing lacunae (Fig. 2D,E,G).

These observations suggest that ring contraction and primary septum formation are processes that depend on one another. Thus, interference with one function would also block the other. A prediction of this hypothesis is that myo1 chs2 double mutants should have the same phenotype as the single mutants. That turned out to be the case. By using a variation of the white-red selection devised by Bender and Pringle (Bender and Pringle, 1991) we found that null mutations of CHS2 and MYO1 are not synthetically lethal. In fact, the double mutant hardly differs from the single mutants, both in morphology and in generation time (Fig. 3C).

Both Chs2p (Chuang and Schekman, 1996) and Myo1p (Lippincott and Li, 1998; Bi et al., 1998) localize at the neck between mother cell and bud, although they reach that area at different points in the cell cycle. Because of the close relationship between their function, we asked whether either protein is localized correctly in the absence of the other. Chs2p-GFP and Myo1p-GFP fusion proteins were used in these experiments. Morphological observations showed that Myo1p-GFP totally complemented a MYO1 defect, whereas a strain carrying an integrated copy of CHS2-GFP was largely normal but still exhibited some clumping. Myo1p rings were easily visible both in the wild-type and in the chs2Δ mutant (Fig. 4A,B,C,D). In the wild-type, 26% of the cells showed rings (536 cells counted, mother cell and bud counted as two cells); in a chs2Δ strain, 21% of the cells had rings (404 cells counted). Chs2p-GFP could be seen in the wild-type in only 1.7% of the cells (700 cells counted, Fig. 4E,F). This is not surprising because Chs2p is apparently synthesized just before septum formation and is degraded immediately afterwards (Choi et al., 1994; Chuang and Schekman, 1996). If one assumes that Chs2p will be at the neck only during contraction of the actomyosin ring, its presence there would last only approximately five minutes, about 5% of the cell cycle (Bi et al., 1998). In myo1Δ strains, Chs2p-GFP could also be detected (Fig. 4G,H) in about the same proportion as in wild-type (1.5%; 700 cells counted). This result indicates that Chs2p does not need a functional contractile ring to attain its localization.

By using time-lapse fluorescent microscopy, we were able to show that at least some of the myosin rings in a chs2 mutant (as in Fig. 4C) do contract (results not shown). However, this result is not easily interpretable, because it is not possible to ascertain whether the contraction was caused by closure of the aberrant septa or whether it took place independently but with detachment of the ring from the plasma membrane.

What is the nature of the remedial septation observed both in chs2Δ and myo1Δ mutants? The septa formed in those strains, although thicker than the normal ones, do not show a trilaminar structure, that is, there is no indication that secondary septa are formed in these cases (Fig. 2). This observation suggests that the thickening of the cell wall seen at cytokinesis in the mutants is itself a manifestation of secondary septum formation. Here, rather than having two portions of membrane facing each other across the primary septum (Fig. 1C), we have a single cylindrical portion of membrane parallel to the axis connecting mother and daughter cell (Fig. 1A). Therefore, when appropriate signals elicit the formation of secondary septa, a single such septum would arise, as an annular structure, growing at 90° from the normal direction (Fig. 1E,F) and finally closing the channel between mother and daughter cell (Fig. 1G).

For this salvage process to take place, another chitin synthase, ChsIII, appears to be necessary. As already reported, chs2 and chs3 mutants are synthetically lethal (Shaw et al., 1991). We now find, by a modification of the red-white screen (Bender and Pringle, 1991), that myo1 and chs3 mutants are also synthetically lethal by that test. Despite this result, we were able to isolate the myo1 chs3 double mutant by deletion of the MYO1 gene in a chs3 background. However, these strains grew very slowly and gave rise to extremely large clumps (Fig. 5A). Staining with methylene blue showed that many cells in the clump were dead (results not shown). Frequently, cells were enlarged, lysed and practically empty (Fig. 5A, arrows). Furthermore, the double mutant could not grow at all at 37°C (Fig. 5B).

Observation of some of the mutants used in this study suggested an abnormal budding pattern. However, accurate determinations of the budding pattern by the usual methods were practically impossible, owing to the clumping manifested by those strains (Fig. 3). Therefore, an alternative procedure was used. Cells with a single bud, rare but still present in the mutants, were micromanipulated with a spore-dissecting apparatus to established positions on a plate. The plate was incubated at room temperature and observation of the cells was carried out every 20-30 minutes. Thus, the budding pattern (Chant and Herskowitz, 1991) for two generations could be determined. However, a polar pattern could not be easily distinguished from a random one with this method; therefore the patterns are recorded in (Table 3) as axial or non-axial, a distinction that could easily be made. Both chs2Δ and myo1Δ mutants exhibited an abnormal, non-axial budding pattern, with a more pronounced defect in chs2Δ mutants. A chs2Δ myo1Δ double mutant was not more defective than a chs2Δ mutant (Table 3). Some abnormality in budding pattern of myo1 mutants was reported by Rodríguez and Paterson (Rodríguez and Paterson, 1990), although the defect was less noticeable in their strains, which contained partial deletions of MYO1.

We also examined the levels of Bud3p and Bud4p, two proteins required for the axial budding pattern (Chant and Herskowitz, 1991), in chs2 and myo1 mutants. Both proteins were severely reduced in the mutants, especially in the myo1 strains (Fig. 6A). Another protein showing a low level in the mutant strains is Spa2p (Fig. 6), which localizes at bud tips and is involved in cell polarization (Sheu et al., 1998). In contrast, other proteins usually localized at the cell cortex, such as Rho1p, Cdc42p and the plasma membrane ATPase Pma1p, were not diminished in the mutants (Fig. 6A). This result prompted us to find out whether the localization of Bud3p and Bud4p in double rings at the junction between mother cell and bud (Chant et al., 1995; Sanders and Herskowitz, 1996) was maintained in chs2 and myo1 mutants. GFP-fusions of Bud3p and Bud4p were constructed for this purpose. Surprisingly, fluorescence microscopy showed that the rings were indeed present both in wild-type and mutants (Fig. 7). An interesting observation was that the double rings showed much greater separation in the mutants than in wild-type (Fig. 7), perhaps as a consequence of the thicker septa present in the mutant strains. In contrast with the results obtained with native Bud3p and Bud4p (Fig. 6A), the level of the GFP fusion proteins was not much diminished in the mutants (Fig. 6B). However, the budding pattern of the mutant harboring the Bud4p-GFP fusion in a chs2 background was as different from wild-type as that of the chs2 strain containing the unmodified Bud4p (Table 3). Therefore, the low level of the Bud proteins in chs2 and myo1 mutants does not appear to be the cause of the abnormal budding pattern. Rather, both the aberrant budding and Bud-protein depletion seem to result somehow from the abnormal septation of the chs2 and myo1 mutants. We will return to this point in the Discussion.

The recent discovery that cytokinesis in S. cerevisiae includes the contraction of an actomyosin ring prompted us to re-examine the process of septation in this organism. Mutants unable to construct a chitin primary septum (as chs2 mutants) or a contractile ring (as myo1 mutants) exhibit a remarkable similarity in their aberrant mechanism of cytokinesis and septation. As judged from observations by electron microscopy, in both cases invagination of the plasma membrane and the pinching off of the membrane in a restricted area does not take place. Instead, the lateral cell wall at the neck between mother cell and bud grows inward over a wide region, gradually closing the gap between the two cells and giving rise to a thick septum lacking the normal three layers. In this process, a wide portion of plasma membrane becomes involved in the closing of the septum, rather than the very small area pinching off in normal cytokinesis. The thick aberrant septum, as it grows, may pinch off the membrane at more than one point, thus giving rise to the cytoplasm-enclosing lacunae that are frequently observed (Fig. 1G; Fig. 2E,F). As mentioned in the Results, we interpret the salvage mechanism of septation found in chs2 and myo1 mutants as the growth of secondary septa at 90° of the normal direction. This is because of lack of the orientation, which is usually provided by the primary septum (Fig. 1E-G).

Thus, it appears that primary-septum formation and actomyosin ring contraction are parts of the same process and are interdependent. This concept is consistent with the finding that chs2 and myo1 mutants are not synthetically lethal. In fact, the double mutant does not differ significantly from the single ones, either in morphology or growth rate. In contrast, chs2 and chs3 mutations are synthetically lethal (Shaw et al., 1991), and a myo1 chs3 double mutant grows extremely slowly, forming very large clumps, which contain many lysed cells (Fig. 5). Thus, CHS3 seems to be necessary for the salvage route that leads to the abnormal but ultimately successful cytokinesis found in myo1 or chs2 mutants. In this connection, it is interesting that we observed a high concentration of chitin, as judged from wheat germ agglutinin-gold staining in the thick septa of chs2 mutants (Shaw et al., 1991). Those strains also showed a higher level of chitin than wild-type in their cell wall (Shaw et al., 1991) and an enhanced activity of chitin synthase III (Crotti et al., 2001). An increase in chitin content (Rodríguez-Medina et al., 1998) and in the activity of chitin synthases (Cruz et al., 2000) has also been reported in myo1 mutants. The extra chitin may be needed to reinforce the growing aberrant septa by cross-linking to β-glucans (Kollár et al., 1995; Kollár et al., 1997). This reinforcement would not be needed for normal secondary septa, whose function is not to effect cytokinesis but to fill in bud and birth scars. These fillings protect the cells after a chitinase hydrolyzes part of the primary septum chitin to effect cell separation (Kuranda and Robbins, 1991). Indeed, chs3 mutants assemble trilaminar septa, which lacks only the external chitin ring laid down at early budding (Shaw et al., 1991).

It is interesting to compare the process of septation in S. cerevisiae with that of the fission yeast, Schizosaccharomyces pombe. S. pombe also form a contractile ring (Gould and Simanis, 1997), as well as primary and secondary septa, except that in this case the primary septum appears to consist of β(1→3)glucan rather than chitin. It has been recently reported that ring contraction and formation of the primary septum are both required for cytokinesis, as we found in S. cerevisiae (Le Goff et al., 1999; Liu et al., 1999). The difference with budding yeast is that in S. pombe, a defect in either structure is lethal, because there is no alternate route for cytokinesis. This lethality may be caused by the lack of chitin in the wall of vegetative cells (Arellano et al., 2000) or by the larger distance that the salvage septa would have to cover to close the gap in the cigar-shaped cells.

In addition to their septation defects, chs2 and myo1 mutants show abnormalities in their budding pattern (Table 3). Concomitantly, they are depleted in some proteins necessary for axial budding, such as Bud3p and Bud4p, and in a protein localizing at the tip of a new bud, Spa2p (Fig. 6A). It would be tempting to assume that the decrease in the Bud proteins is the cause of the abnormal pattern. However, Bud3p and Bud4p double rings were normally formed in myo1 and chs2 mutants (Fig. 7), albeit with more separation between them than in the wild-type. Furthermore, the strains in which the rings were observed harbored Bud3p-GFP or Bud4p-GFP fusion proteins, which were maintained at much higher levels than the original Bud3p and Bud4p (Fig. 6B). Nevertheless, those strains had the same axial budding defect as the others. Therefore, it is not the depletion of the proteins in itself that causes the aberrant budding pattern. Rather, the reason for this behavior may reside in the abnormal septation of chs2 and myo2 strains. As shown in the scheme of Fig. 1B, the wide area along which repair septa make contact may result in the loss of a fairly large portion of plasma membrane, some of it remaining entrapped in lacunae. This part of the plasma membrane may contain binding sites where certain proteins, such as Bud3p, Bud4p or Spa2p, would attach in the next budding event. Elimination of those sites may lead to the abnormal budding pattern and to enhanced degradation of those proteins with a resulting decrease in their cellular level.

Finally, why are formation of the chitin primary septum and contraction of the actomyosin ring interdependent processes? First of all, this is not due to delocalization of the proteins of interest in the mutants used in this study. Chs2p localized correctly in myo1 mutants and so did Myo1p in chs2 mutants (Fig. 4). Second, Chs2p and Myo1p do not seem to interact directly. This notion is supported by the lack of effect of one protein on the localization of the other. Furthermore, we were unable to detect two-hybrid interaction between the two proteins (results not shown). Finally, Chs2p is located in the plasma membrane and in secretory vesicles, whereas Myo1p is in the cytoplasm. Fractionation of cell extracts by centrifugation in density gradients showed that the two proteins are in fractions of very different density (results not shown). It is not difficult to understand why formation of the chitin septum requires contraction of the ring. Chitin synthase II, the enzyme responsible for the formation of primary septum chitin, is localized in the plasma membrane (Sburlati and Cabib, 1986) at the neck between the mother cell and the bud (Fig. 4). Invagination of the plasma membrane coupled to continuing extrusion of chitin from the center of the invagination will automatically generate an inwardly growing ‘doughnut’ of the polysaccharide that will ultimately evolve into a disk. If there is no invagination, some chitin may build up at the neck but no disk will result.

The reciprocal need for chitin septum formation in order that the ring may contract is less easily understood. We propose an explanation on the basis of the high turgor pressure of fungal cells. Because of the elevated intracellular osmolarity, relative to that of the medium, water tends to enter the cell and the resulting turgor pressure pushes the plasma membrane against the cell wall. In fact, even a small hole or weak area in the cell wall results in cell lysis and in the blowing out of cell membrane through the opening (Cabib et al., 1989). Thus, in order to have contraction, the force developed by the actomyosin ring must exceed that generated by the turgor pressure. Here is where the extrusion of chitin may help, both by pushing the membrane inward and, through a sort of ratcheting action, by preventing it from moving back towards the cell wall. Therefore, in our conception, primary septum formation and actomyosin ring contraction are integral parts of the same cytokinetic process. Note that mammalian cells complete cytokinesis without the need of a septum. This is consistent with our hypothesis, because mammalian cells are devoid of turgor pressure.

We thank S. Sanders for antibodies against Bud3p and Bud4p and C. Slayman for antibodies against Pma1p. We are grateful to M. Snyder for plasmid pBU4 and to J. Lippincott for plasmids pLP7 and pLP8. We are also indebted to A. Minton for useful discussions and to O. Cohen-Fix and T. Roberts for a critical reading of the manuscript.