|Home | About | Journals | Submit | Contact Us | Français|
Cell morphogenesis is a complex process that relies on a diverse array of proteins and pathways. We have identified a transglutaminase-like protein (Cyk3p) that functions in fission yeast morphogenesis. The phenotype of a cyk3 knockout strain indicates a primary role for Cyk3p in cytokinesis. Correspondingly, Cyk3p localizes both to the actomyosin contractile ring and the division septum, promoting ring constriction, septation, and subsequent cell separation following ring disassembly. In addition, Cyk3p localizes to polarized growth sites and plays a role in cell shape determination, and it also appears to contribute to cell integrity during stationary phase, given its accumulation as dynamic puncta at the cortex of such cells. Our results and the conservation of Cyk3p across fungi point to a role in cell wall synthesis and remodeling. Cyk3p possesses a transglutaminase domain that is essential for function, even though it lacks the catalytic active site. In a wider sense, our work illustrates the physiological importance of inactive members of the transglutaminase family, which are found throughout eukaryotes. We suggest that the proposed evolution of animal transglutaminase cross-linking activity from ancestral bacterial thiol proteases was accompanied by the emergence of a subclass whose function does not depend on enzymatic activity.
Transglutaminases (TGases) are a family of enzymes that catalyze intramolecular or intermolecular protein cross-linking through isopeptide bond formation between lysine (or polyamines) and glutamine residues. The activity of the enzymes plays an important role in various intracellular and extracellular processes. For example, keratinocyte TGase functions in the terminal differentiation of keratinocytes and formation of the cornified cell envelope (Rice and Green, 1978 ; Thacher and Rice, 1985 ); tissue TGase (TG2) is found in many cell types and is involved in inflammation, apoptosis, cell adhesion, and cancer and other human diseases (Fesus and Szondy, 2005 ; Mangala and Mehta, 2005 ; Mehta et al., 2010 ; Facchiano et al., 2006 ; Zemskov et al., 2006 ). Factor XIII probably represents the best-studied TGase; it participates in blood clotting, tissue repair, and wound healing (Pisano et al., 1968 ; Schwartz et al., 1973 ; Mosher and Schad, 1979 ; Sakata and Aoki, 1980 ; Knox et al., 1986 ). A hallmark of these enzymes is a catalytic triad made up of conserved cysteine, histidine, and aspartate residues, each of which is essential for activity and defines the catalytic core (Figure 1A; Hettasch and Greenberg, 1994 ; Micanovic et al., 1994 ; Pedersen et al., 1994 ; Yee et al., 1994 ).
The increasing abundance of protein sequence information has revealed a new inactive class of TGases that possesses a conserved catalytic domain lacking the catalytic triad (Makarova et al., 1999 ). For example, a red blood cell protein (band 4.2) possesses a TGase domain that lacks the catalytic cysteine and histidine residues (Figure 1A) and consequently lacks enzyme activity (Korsgren et al., 1990 ; Cohen et al., 1993 ). Naturally occurring mutations in the human band 4.2 gene cause congenital spherocytic anemia characterized by sphere-shaped (rather than biconcave disk–shaped) cells that are more prone to hemolysis (Yawata, 1994 ; Bruce et al., 2002 ; Dahl et al., 2004 ). This phenotype highlights the physiological relevance of proteins with inactive TGase domains, and may reflect a structural role at the cell membrane in the case of band 4.2 (Satchwell et al., 2009 ). However, nothing is known regarding the cellular role (if any) of the inactive TGase domains themselves. In this study, we examined the role of an inactive TGase (Cyk3p) from the fission yeast Schizosaccharomyces pombe.
Cyk3p was originally identified in the budding yeast Saccharomyces cerevisiae (Korinek et al., 2000 ), and (based on available sequence data) represents a highly conserved fungal protein. Cyk3p participates in cytokinesis in S. cerevisiae (Korinek et al., 2000 ), and was recently found to be essential for cytokinesis and growth in the pathogenic budding yeast Candida albicans (Reijnst et al., 2010 ). Several studies in S. cerevisiae recently shed light on the molecular mechanisms governing Cyk3p function. Localization of Cyk3p at the division site depends on its N-terminal SH3 domain (Jendretzki et al., 2009 ), which mediates an interaction with a proline-rich region at the C-terminus of the C2 domain protein (Inn1p), a critical cytokinetic factor (Sanchez-Diaz et al., 2008 ; Jendretzki et al., 2009 ; Nishihama et al., 2009 ). Cell cycle–specific phosphorylation may regulate such complexes, because the division site localization of Cyk3p and Inn1p relies on the mitotic exit network (MEN; Meitinger et al., 2010 ), a conserved signaling pathway essential for the coordination of mitotic progression and cytokinesis (McCollum and Gould, 2001 ). Relatively little is known about fission yeast Cyk3p, although preliminary studies have pointed to a role in cytokinesis. Similar to several other fission yeast cytokinetic mutants, a cyk3Δ mutant was found to be sensitive to loss of the cytokinetic checkpoint mediated by the Cdc14 family phosphatase (Mishra et al., 2004 ). In addition, fission yeast Cyk3p was recently found to coprecipitate from cell extracts with F-BAR protein Cdc15p (Roberts-Galbraith et al., 2010 ), an essential component of the contractile ring (Fankhauser et al., 1995 ).
In this study, we define roles for Cyk3p in fission yeast cytokinesis and morphogenesis and identify an inactive C-terminal TGase motif as the critical functional element. Overall, our study provides new insights into the mechanisms of Cyk3p function, cytokinesis, and cell morphogenesis in fungi, and more broadly identifies an important cellular role for the inactive subclass of the TGase family.
Like other known fungal members of this protein family (Korinek et al., 2000 ; Reijnst et al., 2010 ), fission yeast Cyk3p possesses an N-terminal SH3 domain (Figure 1B). In addition, we identified a TGase-like domain in the C-terminal half of the protein (Figure 1B), a domain preserved in Cyk3p sequences from other fungi. This 111–amino acid domain showed homology to TGase domains in S. cerevisiae Cyk3p (30% identical/51% similar) and animal TGases (e.g., 17%/41% vs. human factor XIII). The catalytic core of the TGase domain can be separated into three motifs centered on the conserved cysteine, histidine, and aspartic acid active-site residues that form the catalytic triad. Interestingly, the catalytic triads of the fungal Cyk3p proteins are all incomplete, containing the conserved histidine and aspartic acid residues in motifs II and III but lacking the active site cysteine in motif I (Figure 1A).
Deletion of cyk3 had no obvious effect on morphology or growth at 25–32°C. However, cyk3Δ cells exhibited temperature-sensitive defects in cell separation at 36°C, as reflected by the appearance of elongated cells with multiple septa and multiple nuclei (Figure 2A). In addition, cyk3Δ cells showed a greater tendency to adopt a rounded/swollen morphology, as opposed to the typical cigar shape of fission yeast (Figure 2A). Consistent with a significant role for Cyk3p in cytokinesis, the cyk3Δ mutation showed synthetic defects when combined with mutations in genes encoding known contractile ring components, such as Myo2p (myosin II), Cdc4p (essential light chain), Cdc12p (formin), and Cdc15p (F-BAR protein) (Figure 2, B and C, and Supplemental Figure S1). Although the contractile rings typically assembled in such double mutants, they showed obvious defects in cytokinesis (Figure 2D), suggesting a role for Cyk3p in ring constriction.
Surprisingly, unlike myo2-E1 and many other cytokinetic mutations (Bezanilla et al., 1997 ; Motegi et al., 1997 ), cyk3Δ did not exhibit synergistic cytokinetic defects when combined with a myp2 null (Figures 2E and S1). Myp2p is a nonessential myosin II required for normal cytokinesis (Bezanilla et al., 1997 ; Motegi et al., 1997 ), and the lack of an additive phenotype suggests that Cyk3p and Myp2p may share a specific function in cytokinesis. Myp2p is known to associate with chitin synthase (Chs2p), an inactive form of the enzyme that localizes to the contractile ring and contributes to septum formation in fission yeast (Martin-Garcia et al., 2003 ; Martin-Garcia and Valdivieso, 2006 ). We therefore also examined a cyk3Δ chs2Δ double mutant. Interestingly, loss of Chs2p completely suppressed the cytokinetic defects associated with loss of Cyk3p (Figures 2F and S1), suggesting a functional relationship between these two proteins at the septum.
We used gene replacement to generate single and triple chromosomal green fluorescent protein (GFP) fusions to examine the subcellular localization of endogenous Cyk3p. The fusion proteins were functional based on their ability to fully support Cyk3p function in a cdc4-8 background. During vegetative growth, Cyk3p localized to three distinct sites: contractile rings, division septa, and cell tips (Figure 3A). The Cyk3p signal was most prominent at contractile rings. Time-lapse analysis revealed that Cyk3p joins the ring at the final stages of its assembly, is present at both rings and across growing septa during ring constriction, and remains as a band spanning the septum following ring disassembly (Figure 3B). This pattern was particularly clear when Cyk3p was colocalized with the myosin II regulatory light chain Rlc1p, which assembles earlier and disappears following ring contraction (Figure 3C and Supplemental Movie S1).
On contractile ring disassembly, Cyk3p localization at the septum was relatively faint, but it became brighter a short time later, forming a tight spot at the center of the septum just before cell separation and remaining at the new cell poles following separation (Figure 3, A and D). This unipolar tip localization became bipolar as Cyk3p accumulated at the old cell poles later in the cell cycle (Figure 3D). Although the bipolar tip localization was relatively faint when the tagged Cyk3p was expressed at endogenous levels, it became more conspicuous when Cyk3p-GFP was expressed from a multi-copy plasmid (Figure 3E).
In summary, Cyk3p localizes in a manner consistent with roles in cell morphogenesis during both division and polarized cell growth.
Given its prominent localization at the actomyosin ring, we tested whether Cyk3p was truly a component of this structure. Fluorescence recovery after photobleaching (FRAP) studies have shown that ring components exchange rapidly and that myosin II Myo2p exchanges with different kinetics in nonconstricting and constricting rings (Sladewski et al., 2009 ; Stark et al., 2010 ). Examination of Cyk3p-GFP exchange revealed almost identical kinetics to those of Myo2p in both nonconstricting (t½ = 27 s vs. ~37 s for Myo2p), and constricting (t½ = 12 s vs. ~13 s for Myo2p) rings (Figure 4, A and B; Sladewski et al., 2009 ). Depolymerization of the actin cytoskeleton with latrunculin A blocked the formation of both myosin II and Cyk3p rings and led to a buildup of both proteins as ring precursors at the incipient medial division sites (Figure 4C). Taken together, these results suggest that Cyk3p is indeed a component of the actomyosin ring.
To test for a possible role of Cyk3p in contractile ring dynamics, we first monitored contractile ring behavior, using Rlc1p-GFP as a marker, in relation to mitotic progression, as judged by the spindle-pole body (SPB) protein Sad1p-GFP (Figure 5A, left box). Using the beginning of SPB separation as a reference point, we found that ring assembly time and constriction rate were affected little or not at all in a cyk3 mutant grown at 25°C, whereas the preconstriction “dwell” time was significantly lengthened (Figure 5, A and andB,B, and Table 1). Interestingly, the delay was paralleled by a delay in the completion of spindle elongation: although ring constriction began ~8–9 min following spindle breakdown in both wild-type and cyk3Δ cells, spindle breakdown (like ring constriction) was delayed ~6 min in the mutant relative to wild-type (Figure 5B). The reason for this delay is not clear, but it appeared to be due mostly to a slower spindle elongation during the phase of contractile ring assembly (Figure 5B).
The ability of cyk3 cells to grow at 36°C allowed a further informative analysis in which cells grown at 36°C were subsequently imaged at 23°C. In this case, the assembly, dwell, and constriction phases were all significantly affected, with the largest effect on the dwell time (Figure 5C and Table 1), probably because the absence of Cyk3p during growth at 36°C produced changes in the levels of Cyk3p-interacting proteins that could affect ring dynamics even after a return to lower growth temperature. In summary, although Cyk3p appears to be involved in all phases of contractile ring function, its most important role seems to be in promoting ring constriction.
The localization of Cyk3p to the septal region (Figure 3) and multi-septate phenotype of cyk3 mutants (Figure 2A) suggested Cyk3p might play a role in the maturation of the septum, its splitting at cell division, or both. Consistent with such a role(s), cyk3 cells showed increased percentages of septated cells (Figure 6A) and of cells that apparently failed to separate (Figure 6B), as well as an increase in the average time from septum completion to cell separation (Figure 6, C and D), during growth at 36°C. These effects were not seen at 25°C (Figure 6D), and even at 36°C, septum completion appeared to coincide with the completion of contractile ring constriction, as in wild-type cells (Figure 6C). Thus, in addition to a role in actomyosin ring constriction, Cyk3p also contributes to cell separation.
When cells expressing Cyk3-GFP grew to stationary phase, they displayed a pattern of localization that appeared to be distinct from the division site and cell tip localization seen during vegetative growth. In particular, Cyk3p was found randomly throughout the cortex as discrete, dynamic puncta (Figure S2 and Movies S2 and S3). Although these puncta were similar in size and distribution to endocytic actin patches, Cyk3p did not colocalize with the actin-patch component Fim1p (fimbrin) during stationary phase (Figure S2), and the Cyk3p puncta had a considerably longer average lifetime (~4 min) than did Fim1p or actin patches (~20 s; Figure S2; Sirotkin et al., 2010 ). In addition, unlike typical endocytic patches, the Cyk3p puncta did not depend on actin (Figure S2) and often showed considerable lateral motility (Movies 2 and 3). These observations suggest that Cyk3p may play a more general role in cell-surface organization in addition to its roles at specific sites during the vegetative cell cycle.
To explore further the roles of Cyk3p in vegetative cells, we overexpressed it in otherwise wild-type cells. On overexpression, cells displayed gross morphological defects characterized by multinucleate cells with abnormal septa and rounded/swollen shapes that had lost the typical cigar shape of fission yeast (Figures 7 and S3). A combination of both phenotypes was evident in some cells (Figure 7), but time-lapse observations revealed that the phenotypes were often independent of one another (Figure S3), so that cell swelling was not simply a secondary effect of cytokinetic failure. These phenotypes paralleled those of cyk3Δ cells (Figure 2A), presumably reflecting roles for Cyk3p in both cytokinesis and cell shape determination.
To gain insight into Cyk3p function, we created plasmids expressing proteins that lacked the SH3 domain or contained mutations in the transglutaminase domain (see Materials and Methods). Surprisingly, Cyk3p-SH3 appeared to retain full function during cytokinesis (Figure 8, A and andB)B) and to support normal polarized growth (Figure 8A). Moreover, like wild-type, this truncated form localized normally (Figure 8C), produced cytokinetic defects and morphological abnormalities when overexpressed (Figure 8D), and was expressed at normal levels (Figure 8E). In contrast, mutation of the conserved aspartate residue of the catalytic triad (in the TGase core) led to a seemingly complete loss of Cyk3p function (Figure 8, A, ,B,B, and andD),D), even though the protein localized normally (Figure 8C) and was expressed at normal levels (Figure 8E). We further tested the importance of the TGase domain using another mutation (His-577-Ala) in the catalytic triad (Figure 1). This mutant protein was expressed effectively, localized normally, and rescued the growth of a cdc4-8 cyk3Δ mutant (Figures 8E and S4). However, closer inspection revealed that Cyk3p-His-577-Ala retained only partial function: it was unable to fully relieve the morphological and cytokinetic defects associated with loss of Cyk3p, and it produced relatively mild morphological defects when overexpressed (Figure S4). In summary, the TGase domain, but not the SH3 domain, is critical for Cyk3p function in fission yeast.
Contractile ring constriction and septum formation are coupled in yeast. The unique localization of Cyk3p in both rings and septa prompted us to further examine its roles in these structures. Because S. cerevisiae Cyk3p division site localization depends on the MEN pathway (Meitinger et al., 2010 ), we tested whether the homologous pathway (the septation initiation network, or SIN) governed Cyk3p localization in fission yeast. The SIN is essential for initiating septum formation and ring constriction (McCollum and Gould, 2001 ). Use of a temperature-sensitive sid2-250 mutant demonstrated that localization of Cyk3p to rings was not dependent on the SIN (Figure S5). Thus, as with other contractile ring proteins, the formation of Cyk3p rings is actin-dependent (Figure 4C) but does not rely on septum formation. This indicates that Cyk3p is a core component of the ring that must interface with the leading edge of the trailing septum during cytokinesis, presumably leading to its incorporation across the septum (Figure 3C).
We turned to electron microscopy (EM) to directly assess Cyk3p's role in ring constriction and septation, and, in particular, how the contractile ring defects of cyk3 mutants affected growth of the septum. Although septation was generally normal in the absence of Cyk3p (Figure 9A, compare a–d with e–g; Figure 9C), a few cyk3 cells (3 of 114 cells examined) contained an apparently complete septum with a gap in the electron-translucent central layer (Figures 9Ah and and9C)9C) that was never observed in wild-type cells. Actomyosin ring mutants also displayed defects in septum structures (Figure 9B, a, b, f, and g), which became much more severe in double mutants lacking cyk3 (Figure 9B, c–e and h–j). Although these double mutants were viable, their septa were typically abnormally thickened and often appeared misdirected across the division plane (Figure 9, B and andC).C). Consistent with the genetic data indicating that Cyk3p functions in the same pathway as Myp2p (Figures 1E and S1), the EM analysis revealed little or no exacerbation of the myp2 septation defect by loss of Cyk3p (Figures 9C and S6). Taken together with the localization of Cyk3p, these direct observations of the septa imply a role for Cyk3p in coupling ring constriction and septum growth during cytokinesis.
Our work highlights the role of a TGase-related protein in cell morphogenesis. Fission yeast Cyk3p participates in actomyosin ring constriction, septation, and cell separation during cytokinesis, and also plays a relatively minor role in cell shape and integrity during vegetative growth and stationary phase. Inactive TGases such as Cyk3p are found throughout eukaryotes and our results demonstrate that the TGase domains of such proteins can have key roles in the cell.
Cyk3p functions in cytokinesis, promoting ring constriction and subsequent cell separation. Correspondingly, Cyk3p appears in two distinct structures, the actomyosin ring and the division septum. Cyk3p is an integral component of the ring and is incorporated late in ring assembly. Around this time cyk3Δ cells exhibit a cell cycle delay reflected by a brief stall in anaphase B and a delay in the initiation of ring constriction. While the defects are subtle, cyk3Δ mutants (like other cytokinetic mutants) invoke lethality in the absence of the Cdc14 family phosphatase Clp1p/Flp1p (Mishra et al., 2004 ). When ring integrity is compromised Clp1p activates a cytokinetic checkpoint, during which Clp1p activity is also mobilized to stabilize ring structures (Mishra et al., 2004 ). Thus actomyosin rings lacking Cyk3p rely on compensatory maintenance to support cytokinesis and growth.
Although we do not yet understand the molecular mechanisms by which Cyk3p carries out its role at the ring, they probably involve three other factors: the F-BAR protein Cdc15p, the nonessential myosin II Myp2p, and the chitin synthase Chs2p. Cyk3p coprecipitates with Cdc15p from cell lysates and (along with other ring components) was found to accumulate at the division site upon premature self-assembly of a constitutively dephosphorylated form of Cdc15p (Roberts-Galbraith et al., 2010 ). Our genetic analysis showed that myp2Δ was epistatic to cyk3Δ. Myp2p stabilizes rings as they constrict and ensures normal septation (Bezanilla et al., 1997 ; Motegi et al., 1997 ; Mulvihill and Hyams, 2003 ). Correspondingly, direct analysis of septation by EM in myo2-E1 cyk3Δ and cdc4-8 cyk3Δ double mutants indicated a role for Cyk3p in coordinating septum deposition with ring constriction. Chs2p is a transmembrane protein involved in septum formation and maintaining contractile ring integrity during the latter stages of constriction (Martin-Garcia et al., 2003 ; Martin-Garcia and Valdivieso, 2006 ). The ability of the chs2Δ mutation to suppress cytokinetic defects associated with loss of Cyk3p suggests that Chs2p and Cyk3p work in a specific pathway with Myp2p. The actin-independent accumulation of Cyk3p at the division site probably relies on Cdc15p at ring precursors (nodes), given that Myp2p and Chs2p only localize to the division site once rings have formed (Bezanilla et al., 2000 ; Martin-Garcia and Valdivieso, 2006 ).
The actomyosin ring disassembles once ring constriction and septum formation are completed (Rajagopalan et al., 2003 ). Completion of cytokinesis and cell separation occur a short time later upon septum maturation and subsequent digestion of the primary septum by the Eng1p and Agn1p endoglucanases (Martin-Cuadrado et al., 2003 ; Dekker et al., 2004 ). Cell separation was significantly slower in cyk3Δ cells, suggesting an important role throughout this process for Cyk3p, consistent with its accumulation across the trailing septum during ring constriction and its polarization at the center of the septum immediately prior to cell separation. Outside of cytokinesis, Cyk3p localized at growing cell tips and redistributed throughout the cortex during stationary phase. These localizations and other results suggest that Cyk3p is mobilized for polarized growth during cell elongation and maintenance of cortical integrity upon cell cycle arrest and entry into the dormant phase. One consistent feature of Cyk3p localization is its appearance at the cortex during cell wall remodeling, both in growing cells and during stationary phase (when reinforcement and thickening of the wall occur to maintain cell integrity; Herman, 2002 ; Rincon et al., 2006 ). Thus we hypothesize that Cyk3p is generally involved in facilitating cell wall synthesis and remodeling at relevant sites on the cortex.
At a mechanistic level, how does Cyk3p contribute to morphogenesis? The SH3 domain of S. cerevisiae Cyk3p is important, mediating its localization and association with the C2-domain protein Inn1p at the division site (Jendretzki et al., 2009 ; Nishihama et al., 2009 ). Similarly, the SH3 domain of the S. cerevisiae F-BAR protein Hof1p also propagates an interaction with Inn1p (Jendretzki et al., 2009 ; Nishihama et al., 2009 ; Meitinger et al., 2010 ). However, whereas loss of Inn1p typically results in lethality or extremely poor growth (Sanchez-Diaz et al., 2008 ; Jendretzki et al., 2009 ; Nishihama et al., 2009 ), cyk3Δ or hof1Δ cells show relatively minor defects in cytokinesis and growth (Kamei et al., 1998 ; Lippincott and Li, 1998 ; Korinek et al., 2000 ). Moreover, reminiscent of the lethal phenotype of a cyk3Δ hof1Δ double mutant, a double mutant possessing truncated forms of both Cyk3p and Hof1p (lacking their SH3 domains) does not grow (Jendretzki et al., 2009 ). In contrast, in fission yeast, the Cyk3p SH3 domain is not critical for function and does not share a redundant role with the SH3 domains of the F-BAR proteins Cdc15p and Imp2p. Irrespective of the status of Cyk3p, removal of the SH3 domains from both F-BAR proteins is enough to prevent recruitment of the C2-domain protein Fic1p to the division site (Roberts-Galbraith et al., 2009 ).
Although its SH3 domain is dispensable, the fission yeast Cyk3p TGase domain is required for function (an issue that does not appear to have been investigated for the S. cerevisiae protein). Each of the conserved active-site residues cysteine, histidine, and aspartic acid, which make up the catalytic triad at the core of TGases, is essential for enzymatic activity (Hettasch and Greenberg, 1994 ; Micanovic et al., 1994 ; Pedersen et al., 1994 ; Yee et al., 1994 ), and mutagenesis of the corresponding aspartic acid (Asp-592) or histidine (His-577) residues of fission yeast Cyk3p led to complete or partial loss of function, respectively. Although these results demonstrate the importance of the TGase domain, it cannot function enzymatically because it lacks the cysteine residue to complete the catalytic triad (Figure 1A). The nearest cysteine residue (Cys-554) is 18 amino acids downstream from where the catalytic cysteine would be expected (Figure 1A). Nonetheless, this cysteine is conserved among Cyk3p homologues (Figure 1A) and could theoretically form a catalytic triad with His-577 and Asp-592. However, analysis of a Cys-554-Ala point mutant indicated that this was not the case. The Cys-554-Ala protein was expressed normally and behaved just like wild-type Cyk3p in vivo (Figures 8E and S4). In addition, the fact Cyk3p-His-577-Ala (but not -Asp-592-Ala) retains partial function (as opposed to complete loss or full function) also argues against an enzymatic role for the TGase domain in Cyk3p function.
The inactive nature of the Cyk3p TGase domain may reflect a structural/cytoskeletal role for this domain in cell wall remodeling, in which the charged His-577 and Asp-592 residues could help mediate electrostatic interactions between the TGase domain and Cyk3p binding partners. Similarly, a structural role has been proposed for band 4.2, the inactive TGase implicated in human disease that functions at the membrane of red blood cells (Satchwell et al., 2009 ). Alternatively, the inactive Cyk3p TGase domain may be a mimic that functions as a dominant-negative regulator of other TGases. In this case, the localization of Cyk3p to sites of cell wall remodeling may promote cell wall dynamics and plasticity by sequestering substrates and thus limiting their cross-linking and stabilization by active endogenous TGases. This hypothesis is supported by the observations that TGase-mediated cross-linking contributes to cell wall organization in S. cerevisiae (Iranzo et al., 2002 ), C. albicans (Ruiz-Herrera et al., 1995 ), and the green alga Chlamydomonas reinhardtii (Waffenschmidt et al., 1999 ). Inhibition of cell wall cross-linking by Cyk3p might explain why its overexpression leads to the loss of polarity and ballooning of cells that could accompany loss of rigidity/excessive plasticity within the cell wall.
In conclusion, our work on fission yeast Cyk3p highlights the importance of the inactive subclass of the TGases that are found throughout eukaryotes. Animal TGases are thought to have evolved from bacterial thiol proteases, because they utilize the same catalytic triad and share the same core structural fold (Pedersen et al., 1994 ; Yee et al., 1994 ). Thus, in addition to adaptations favoring substrate cross-linking (via a reversion of the proteolytic reaction), adaptations associated with loss of enzymatic activity also appear to have been selected during the course of evolution.
Strains were grown in EMM (Edinburgh minimal media) or YE5S (yeast extract plus supplements) medium, and standard genetic and cell biology protocols were used (Moreno et al., 1991 ). Table 2 lists the strains used in this study. Gene deletion mutants and strains expressing fusion proteins tagged with GFP, 3GFP, or Cherry were constructed using genomic integrations with the relevant kanR or natR cassettes (Bahler et al., 1998 ). Cyk3p-GFP and Cyk3p-3GFP fusions were functional based on their ability to fully support Cyk3p function in a cdc4-8 background. Plasmids used in this study are listed in Table 3. The cyk3 open reading frame (ORF) was amplified from a genomic library using the primers: 5′ XhoI-cyk3 CTCGAGATGTCCATTCCTAAACAACTACCATGC; 3′ NotI-cyk3 GCGGCCGCCAACCGCCTGCCAGGTTGCATAGC. The ORF was ligated into XhoI/NotI-linearized pDS572a (overexpression vector with the nmt1 high-strength promoter) and pDS572-81 (nmt1-81 weak-strength promoter vector for plasmid-based complementation). A Cyk3p N-terminal truncation construct lacking amino acids 3–65 (spanning the entire SH3 domain and seven upstream residues) was made using the 5′ primer: XhoI-cyk3-SH3Δ: CTCGAGATGAGTGATATTCCCACGGTACGACCTGGC. A Cyk3p Asp-592-Ala point mutant was constructed using overlap extension (Ho et al., 1989 ) to amplify a mutated 3′ 800–base pair fragment of the cyk3 ORF. Overlapping subfragments were generated using the following primers: 1) 5′ SspI-cyk3: CTGTAGGTACTAATATTCATAACATG and 3′ Asp-592-Ala: GGCAAAACTAGCAGCAATTAGACG; and 2) 5′ Asp-592-Ala: CGTCTAATTGCTGCTAGTTTTGCC and 3′ HpaI-cyk3: GTTTCCCAGAAAGCCTGTGTTAACGTG. The mutant fragment was amplified from the overlapping fragments (using the 5′ SspI and 3′ HpaI primers) and subcloned into a Topo-cyk3 clone via the unique SspI and HpaI sites (which reside in the 3′ region of the cyk3 ORF). The mutated cyk3 ORFs were inserted into the pDS572a and pDS572-81 vectors. The same strategy was used to generate Cyk3p-Cys-554-Ala and -His-577-Ala mutants with the following primers (and their 3′ complements): 5′ Cys-554-Ala: GCACTAGATTTATGGGCTGAGGTTATCG and 5′ His-577-Ala: CCAGAGATATTAATATAAATGCTGCTTGGAATG. The fidelity of cyk3 sequences was confirmed by DNA sequencing.
Differential interference contrast (DIC) and epifluorescence cell images were captured using a Nikon (Melville, NY) TE2000-E2 inverted microscope with motorized fluorescence filter turret and a Plan Apo 60×/1.45 numerical aperture (NA) objective. For fluorescence, an EXFO X-CITE 120 illuminator was utilized. NIS Elements software was used to control the microscope, two Uniblitz shutters, a Photometrics CoolSNAP HQ2 14-bit camera, and auto-focusing. Time-lapse movies of cells monitored Cyk3p and/or Rlc1p contractile ring dynamics (every 2–3 min for 2–3 h) and Cyk3p and/or Fim1p cortical patch lifetimes (by capturing images every 5–10 s for 5–10 min) using appropriate filters. For ring movies, autofocusing was performed on the DIC channel before each image capture. Cell suspensions (3 μl) were mounted on flat 30-μl media pads (solidified by 1% agarose) prepared on the slide surface. VALAP (1:1:1 vasoline:lanolin:paraffin) was used to seal slides and coverslips. For simultaneous tracking of SPBs (Sad1p-GFP) and rings (Rlc1p-GFP), a Z-stack of six images (taken every 0.75 μm spanning the depth of the cell) was collected every 2–3 min for 90–120 min. Images were captured using Nikon (Melville, NY) ND software, and analysis of ring and patch dynamics was performed using Image J, Microsoft Excel, and KaleidaGraph software. Ring dynamics were quantified by assessing individual phases: assembly was time taken for Rlc1p-GFP to compact into a mature ring following its appearance as a broad band of nodes; dwell was time from completion of ring compaction until initiation of constriction; constriction was change in ring circumference over time. Dwell times and constriction initiation were discerned by plotting ring diameter over time for each ring, and constriction rates were derived from the slopes of these plots.
FRAP experiments used confocal laser-scanning microscopy with a Zeiss LSM 510 META system equipped with an argon laser, META detector, and a Plan-Apo 100×/1.4 NA objective. Cells were mounted on 1% agarose pads (as described above) prior to microscopy at room temperature. A region of interest (ROI) was selected on Cyk3p-GFP rings for directed bleaching. Photobleaching iterations were performed briefly at high laser power, resulting in ~90–100% signal loss. Signal recovery was monitored by time-lapse analysis at low laser power, with images collected every 5 s postbleach until the recovery signal plateaued (~2 min). The criterion for determining whether Cyk3p rings were constricting or not lay in their diameters: if rings were clearly narrower than the cell diameter, we scored them as constricting; if they were the same diameter as the cell width, we viewed them as nonconstricting. The LSM 510 software (version 4.2) was used to collect images and perform data analysis (see Figure 4B legend). Recovery curves of Cyk3p-GFP signal versus time were plotted and fit using KaleidaGraph software. Data sets for each trace were corrected for any additional bleaching encountered during time-lapse imaging by a control ROI (derived from an unbleached ring in the same field of cells). To facilitate curve fitting, zero signal intensity was set for each trace by subtracting any residual Cyk3p-GFP signal (detected at the first time point postbleach, 0 s) from all trace values. The t½ values (± SD) represent the mean generated from the fits of each individual FRAP experiment.
EM was performed as previously described (Nishihama et al., 2009 ), with minor modifications. Cells were cultured overnight in YE5S medium at 25°C to mid-log phase, diluted to an OD600 of 0.1, and then regrown at 27.5°C to an OD600 of 0.5 (or at 37°C for 4 h). The cells were harvested, fixed with glutaraldehyde and potassium permanganate, stained with uranyl acetate, and embedded in LR White resin (Fluka, St. Louis, MO). Thin-section samples were poststained with uranyl acetate and lead citrate and were observed using a JEM1230/JEOL microscope (Tokyo, Japan) equipped with an ORIUS SC1000A cooled-CCD camera (Gatan, Pleasanton, CA).
cyk3Δ cells harboring pGFP (vector alone), cyk3-GFP, cyk3-Asp-592-Ala-GFP, cyk3-His-577-Ala-GFP, cyk3-Cys-554-Ala-GFP, or cyk3-SH3Δ-GFP plasmids were grown to an OD595 of 1 in 200 ml of EMM Ura− medium. Cells were then harvested and washed once in water and once in ice-cold lysis buffer (750 mM KCl, 25 mM Tris-HCl, pH 7.4, 4 mM MgCl2, 20 mM Na4P2O7, 2 mM ethylene glycol tetraacetic acid, and 0.1% Triton X-100). Pellets were resuspended in an equal volume of ice-cold lysis buffer with additives consisting of 1 mM dithiothreitol, 4 mM ATP, 2 mM phenylmethylsulfonyl fluoride, and complete EDTA-free protease inhibitors (Roche, Indianapolis, IN). From this point forward, all work was performed at 4°C and samples were stored on ice. Cells were lysed by glass bead beating with a Fastprep (MP Biochemicals, Solon, OH). Lysates were normalized for total protein using a Bradford mix (Bio-Rad, Hercules, CA), mixed 1:1 with 2X SDS–PAGE loading buffer, and boiled for 10 min. Samples were run on a 10% SDS–PAGE gel, transferred to nitrocellulose, and immunoblotted (Sambrook et al., 1989 ) using anti-GFP (Clontech, Mountain View, CA) and anti-actin (Chemicon, Temecula, CA) antibodies diluted 1:1000 in PBS containing 0.1% Tween-20. Horse radish peroxidase-conjugated secondary antibodies were used (diluted 1:3000).
We thank Mohan Balasubramanian (Temasek Life Sciences Laboratory, Singapore) for the cdc4-8, cdc12-112, cdc15-127, myo2-E1, and sid2-250 strains; Jian-Qiu Wu (Ohio State University, Columbus, OH) for the cdc25-22 strain; Henar Valdivieso (University of Salamanca, Spain) for the chs2Δ strain; and Tom Pollard (Yale University, New Haven, CT) for the myp2Δ strain. We thank Wei-Lee Li (University of Massachusetts, Amherst, MA), Volodia Sirotkin (SUNY Upstate Medical University, Syracuse, NY), Jian-Qiu Wu, and Susan Forsburg (University of Southern California, Los Angeles, CA) for pFA6a derivatives and fission yeast plasmids. We are very grateful to Tom Pollard, within whose laboratory this work was initiated. Work in the Lord laboratory is funded by an American Heart Association Scientist Development Grant (0835236N) and a National Institutes of Health grant (GM097193); yeast work in the Pringle laboratory has been funded by the National Institutes of Health grant GM31006.
This article was published online ahead of print in MBoC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E11-07-0656) on May 9, 2012.