Stu1 is the Saccharomyces cerevisiae member of the CLASP family of microtubule plus-end tracking proteins and is essential for spindle formation. A genomewide screen for gene deletions that are lethal in combination with the temperature-sensitive stu1-5 allele identified ldb18Δ. ldb18Δ cells exhibit defects in spindle orientation similar to those caused by a block in the dynein pathway. Consistent with this observation, ldb18Δ is synthetic lethal with mutations affecting the Kar9 spindle orientation pathway, but not with those affecting the dynein pathway. We show that Ldb18 is a component of dynactin, a complex required for dynein activity in yeast and mammalian cells. Ldb18 shares modest sequence and structural homology with the mammalian dynactin component p24. It interacts with dynactin proteins in two-hybrid and co-immunoprecipitation assays, and comigrates with them as a 20 S complex during sucrose gradient sedimentation. In ldb18Δ cells, the interaction between Nip100 (p150Glued) and Jnm1 (dynamitin) is disrupted, while the interaction between Jnm1 and Arp1 is not affected. These results indicate that p24 is required for attachment of the p150Glued arm to dynamitin and the remainder of the dynactin complex. The genetic interaction of ldb18Δ with stu1-5 also supports the notion that dynein/dynactin helps to generate a spindle pole separating force.
A combination of microtubule motor and nonmotor proteins are involved in generating forces required for mitotic spindle formation and orientation (Sharp et al. 2000; Rogers et al. 2005). In the budding yeast Saccharomyces cerevisiae, spindle formation relies on the kinesin-5 proteins, Cin8 and Kip1, and on Stu1, a member of the CLASP family of microtubule plus-end tracking proteins (Hoyt et al. 1992; Roof et al. 1992; Yin et al. 2002). Kinesin-5 proteins crosslink antiparallel polar microtubules and slide them past one another to generate a spindle pole separating force (Sharp et al. 1999). CLASP proteins are believed to stimulate microtubule assembly at kinetochores to help maintain spindle pole separation (Maiato et al. 2005). In temperature-sensitive cin8-3 kip1Δ and stu1-5 strains, spindle pole separation is blocked, and metaphase spindles collapse with their previously separated spindle poles being drawn together.
Nonspindle proteins also exert force on the spindle poles by acting on astral microtubules. The primary role of these proteins is to orient the spindle in the cell. In yeast, the spindle is oriented via two sequential pathways. Prior to anaphase, the spindle moves to the bud neck and becomes aligned along the mother-bud axis. This process involves a complex of the microtubule plus-end binding protein Bim1, Kar9, and the myosin V motor Myo2 that transports microtubule tips along actin filaments toward the bud neck, as well as the kinesin-8 protein Kip3 that promotes microtubule depolymerization (Yin et al. 2000; Hwang et al. 2003; Gupta et al. 2006). Shortly after the onset of anaphase, the elongating spindle is pulled through the mother-bud neck. In this process, Bik1, Pac1, and Ndl1 load dynein onto the plus end of an astral microtubule (Lee et al. 2003; Sheeman et al. 2003; Li et al. 2005). When dynein contacts the bud cortex, its minus-end-directed motor slides the microtubule across the cortex, pulling the spindle pole through the bud neck (Eshel et al. 1993; Adames and Cooper 2000; Heil-Chapdelaine et al. 2000; Lee et al. 2005). Although both Kar9 and dynein pathway mutants show defects in spindle orientation, they are viable; however, loss of both pathways is lethal.
Dynein activity requires the dynactin complex. Dynactin was originally isolated as a multisubunit complex required for dynein function in vitro (Gill et al. 1991; Schroer and Sheetz 1991). Dynactin contains a 40-nm rod composed of the actin-related protein Arp1 (see Figure 2A). One end of the rod is capped by the actin capping protein CapZ, while the other end contains a second actin-related protein, Arp11. Proteins p25, p27, and p62 are localized at the pointed end of the Arp1 filament. A shoulder-sidearm projection, consisting of p150Glued, dynamitin, and p24, binds the rod and projects away from the filament (Schroer 2004). When the dynactin complex is disrupted, these proteins can be isolated together as a subcomplex with a p150Glued:dynamitin:p24 stoichiometry of 2:4:2 (Eckley et al. 1999). p150Glued is the subunit responsible for binding dynein and contains a CAP-Gly motif for microtubule binding. Homologs for key dynactin subunits in budding yeast have been identified, including Nip100 (p150Glued), Jnm1 (dynamitin), Arp1 (Arp1), and Arp10 (Arp11) (McMillan and Tatchell 1994; Muhua et al. 1994; Kahana et al. 1998; Clark and Rose 2006). Loss of any one of these proteins results in a spindle-orientation defect similar to that observed for dynein mutants. No homologs in yeast have yet been characterized for the pointed-end complex proteins (p25, p27, and p62), or p24.
To understand the forces acting on spindle poles, we set out to identify proteins that interact genetically with Stu1. One of these is the relatively uncharacterized Ldb18 that we show is the yeast homolog of the p24 dynactin subunit.
MATERIALS AND METHODS
Yeast strains and sequence analysis:
The yeast strains used in this study are listed in Table 1. Gene deletion strains were obtained from the Genomic Deletion Collection (Winzeler et al. 1999). Epitope-tagging of Ldb18 was performed by integrating plasmids pAH44 and pALM79 into wild-type strains; pAH44 and pALM79 contain a C-terminal segment of Ldb18 fused to 13Myc and 3GFP, respectively. Epitope-tagged versions of Ldb18 were deemed functional because they did not increase the percentage of binucleate cells and were not synthetic lethal with kar9Δ. Strains containing epitope-tagged versions of Nip100, Jnm1, and Arp10 (MY8895, MY8896, MY8912, MY8913, and MY8938) were provided by Mark Rose (Princeton, NJ) (Clark and Rose 2006). A strain expressing mCherry-Tub1 (Khmelinskii et al. 2007) was provided by Elmar Schiebel (Universität Heidelberg, Heidelberg, Germany).
The AlignX module of VectorNTI suite (Invitrogen, Carlsbad, CA) was used to align sequences Ldb18 and p24 (human dynactin 3 NP_009165) using the Clustal W algorithm. To determine the statistical significance of the observed percentage identity between p24 and Ldb18, we performed a permutation test by randomly shuffling the p24 sequence 10,000 times and determining the identity between each of the shuffled p24 sequences and Ldb18. The (one-tailed) P-value was calculated as the number of shuffled sequences with percentage identity equal to or greater than the observed percentage identity, divided by the number of permutations done. Secondary structures were predicted using PHD (Rost and Sander 1993). Coiled-coil domains were predicted using the MTIDK matrix and 2.5-fold weighing function of the COILS algorithm (Lupas et al. 1991).
Synthetic genetic array analysis:
The temperature-sensitive allele stu1-5 was crossed with nonessential gene deletions and double mutants were screened for viability in triplicate at 26° and 30° as done previously (Tong et al. 2001). Select results were confirmed by manual tetrad dissection.
Two-hybrid assays were performed as described previously (Wolyniak et al. 2006). Vectors containing Jnm1 (MR4212), Arp1 (MR4187), and Arp10 (MR5393) fused to the Gal4-activation domain were a gift from Mark Rose (Princeton, NJ). pAH34 (Gal4-BD LDB18), pAH35 (Gal4-AD LDB18), pAH50 (Gal4AD-NIP100), and pAH51 (Gal4AD-NIP100 (400-600)) were constructed in this study.
Co-immunoprecipitation experiments were performed as described previously (Wolyniak et al. 2006) using antibodies against Myc (9E10; Covance, Emeryville, CA), HA (16B12; Covance), Jnm1 (gift from Kelly Tatchell, Louisiana State University, LA), or Arp1 (gift from Mark Rose). Either goat anti-mouse IgG light chain (Jackson ImmunoResearch, West Grove, PA) or goat anti-rabbit (Bio-Rad, Hercules, CA) was used as a secondary antibody.
Sucrose gradient sedimentation:
Cell lysates (500 μl of 5 mg/ml) were sedimented on 5–20% sucrose gradients as described previously (Clark and Rose 2006). Fractions (1 ml) were collected and trichloroacetic acid precipitated (Schuyler and Pellman 2002) before SDS–PAGE and immunoblotting. Yeast alcohol dehydrogenase (7.4 S) and thyroglobulin (19.4 S) were run as standards (Sigma, St. Louis).
DAPI staining of cells was carried out as described previously (Sheeman et al. 2003) and imaged on a Axioplan 2 imaging microscope (Zeiss, Thornwood, NY) using Openlab software (Improvision, Lexington, MA). Live cells expressing GFP and mCherry constructs were imaged on a spinning disk confocal imaging system (UltraVIEW, Perkin-Elmer, Wellesley, MA).
RESULTS AND DISCUSSION
Ldb18 plays a role in the dynein pathway:
To identify genes that play roles in spindle stability, we carried out a synthetic genetic array (SGA) screen looking for gene deletions that are lethal in combination with the temperature-sensitive stu1-5 allele (Tong et al. 2001). Many of the identified genes are involved in a relatively small number of processes including tubulin folding (CIN1, CIN2, GIM3, GIM4, GIM5, PAC2, PAC10, and YKE2), dynein/dynactin activity (ARP1, DYN1, DYN2, DYN3, JNM1, NIP100, PAC1, and PAC11), microtubule and spindle function (ASE1, BIM1, KIP3, KAR9, TUB3, and VIK1), spindle checkpoint activity (BUB1, BUB3, MAD1, MAD2, and MAD3), kinetochore structure (CHL4, CTF3, CTF19, IML3, MCM21, and MCM22), chromosome cohesion (CTF8 and CTF18), and nuclear transport (APQ12, MOG1, NPL3, NUP170, NUP188, SAC3, SOY1, and TEX1). Genes previously isolated in an earlier nonsystematic synthetic-lethal analysis with stu1-5 (PAC10, GIM3, and KEM1) (Brew and Huffaker 2002) emerged in this SGA, demonstrating the effectiveness of identification of stu1-5 synthetic-lethal interactions. One additional gene from this SGA screen, the relatively uncharacterized LDB18 (YLL049w) had previously been shown to be synthetic lethal with a variety of other genes involved in spindle elongation, nuclear migration, and tubulin folding (Tong et al. 2004). ldb18Δ strains also bind less of a cationic dye used to identify mutants defective in oligosaccharide modification (Corbacho et al. 2005); hence the name LDB for low dye binding.
To investigate whether Ldb18 is involved in spindle function, we observed microtubules in live ldb18Δ cells expressing GFP-Tub1. While anaphase spindles in wild-type cells extend through the bud neck, elongated spindles in ldb18Δ cells are frequently observed entirely within the mother cell (Figure 1A). To quantify the spindle orientation defect, we stained cells with DAPI to visualize chromosomal DNA. In wild-type cells, segregated chromosomes are always located in mother cell and bud, respectively. However, in ∼20% of ldb18Δ cells, segregated chromosomes reside entirely within the mother cell (Figure 1, B and C). This number increases to >60% when cells are grown at 12°. This phenotype, and its increased penetrance at low temperatures, is typical of mutations that inhibit dynein (Figure 1, A–C) and other components of the dynein-mediated spindle orientation pathway (Eshel et al. 1993; McMillan and Tatchell 1994; Muhua et al. 1994; Kahana et al. 1998).
Because the Kar9 and dynein pathways are redundant for cell viability, mutations in the dynein pathway will be lethal in combination with mutations in the Kar9 pathway, but not with other mutations in the dynein pathway. To test whether Ldb18 acts in the dynein pathway, ldb18Δ was crossed to a number of Kar9 and dynein pathway mutations. ldb18Δ is synthetic lethal with all tested deletions of Kar9 pathway genes (BIM1, BNI1, KAR9, and KIP3), but not with any tested deletions of dynein pathway genes (ARP1, BIK1, DYN1, JNM1, NIP100, NUM1, and PAC11). Thus, both phenotypic and genetic evidence place Ldb18 in the dynein pathway. Analysis by others using available genomic synthetic-lethal and two-hybrid data also suggests a role for Ldb18 in the dynein pathway (Kelley and Ideker 2005; Ye et al. 2005).
Ldb18 is the p24 homolog in the yeast dynactin complex:
Available genomewide two-hybrid data identified Jnm1 as a protein that interacts with Ldb18 (Ito et al. 2001; Ye et al. 2005). Jnm1 is a component of the yeast dynactin complex, a homolog of the mammalian dynamitin protein, and its loss produces a phenotype similar to what is observed for ldb18Δ (McMillan and Tatchell 1994). To test if Ldb18 is associated with dynactin, the dynactin proteins Nip100, Jnm1, Arp1, and Arp10 were tested for their interaction with Ldb18. In the yeast two-hybrid assay, Ldb18 interacts strongly with Jnm1 and Nip100 and weakly with Arp1 and Arp10 (data not shown). The Ldb18-binding region of Nip100 was narrowed down to residues 400–600 that lie between its two coiled-coil domains. Ldb18 also interacts with itself, suggesting that it may dimerize (data not shown). These results indicate that Ldb18 is a component of the dynactin complex that interacts directly with the shoulder-sidearm components Jnm1 and Nip100.
Interactions between Ldb18 and dynactin components were also assessed by co-immunoprecipitation assays. Immunoprecipitation of Ldb18 pulls down Nip100, Jnm1, Arp1, and Arp10 (Figure 2B). Similarly, immunoprecipitations of Nip100, Jnm1, and Arp1 pull down Ldb18 (data not shown). Additionally, Ldb18 migrates with dynactin proteins during sucrose gradient sedimentation as a ∼20 S complex (Figure 2C). Immunoblots of Ldb18 detect three bands; the lower two bands are near the predicted molecular weight for Ldb18-13Myc, while the upper band is ∼10 kDa higher. Interestingly, only the upper band comigrates with the dynactin complex on sucrose gradients, indicating that post-translational modification of Ldb18 may be needed for its incorporation into the dynactin complex. The nature of this Ldb18 modification is under investigation.
Yeast homologs have not been identified for several mammalian dynactin components, including the shoulder-sidearm protein p24 and the pointed-end proteins p25, p27, and p65. BLAST searches with Ldb18 do not identify any metazoan proteins. However, the size of Ldb18 (21 kDa), and its interaction with the yeast homologs of dynamitin (Jnm1) and p150Glued (Nip100), suggest that Ldb18 may be the homolog of p24. An alignment of the Ldb18 and human p24 sequences show that they share 29.8% similarity and 16.9% identity (Figure 3A).
To demonstrate that this percentage identity is significant and not due to chance alone, we did a permutation analysis by randomizing the p24 sequence and calculating the percentage identity between each permuted sequence with Ldb18 (n = 10,000). Overall, the mean percentage identity of the randomly permuted sequences with Ldb18 is lower than the actual identity (6.2 vs. 16.9%) (P-value < 0.0001, permutation test). The maximum percentage identity observed between Ldb18 and the permuted p24 sequences was 14.6%, still lower than the actual value. The percentage identity and similarity between Ldb18 and p24 is comparable to the percentages between other known and accepted dynactin homolog pairs—Nip100-p150Glued (21.9% similar and 12.7% identical), Jnm1-dynamitin (26.7% similar and 15.8% identical), and Arp10-Arp11 (24% similar and 12.2% identical). In addition, the two proteins share similar secondary structure. Both p24 and Ldb18 are predicted to contain long stretches of α-helical structure (Karki et al. 1998; Pfister et al. 1998). By contrast, the other two dynactin proteins of similar molecular weight, p25 and p27, are predicted to adopt a left-handed β-helix, a motif not commonly found (Parisi et al. 2004). Sequence analysis using the COILS algorithm suggests that both p24 and Ldb18 contain coiled-coil domains near their amino termini and possibly at their carboxyl termini (Figure 3B). It is the amino termini of these proteins that share the most similarity. Hence, based on sequence alignment, similar secondary structure, and strong interactions with Nip100 and Jnm1, Ldb18 is likely the yeast homolog of the dynactin component p24.
Yeast dynactin localizes to the spindle pole bodies and to the distal ends of astral microtubules (McMillan and Tatchell 1994; Kahana et al. 1998; Grava et al. 2006). To determine the localization of Ldb18 in live cells, we fused three tandem copies of GFP to the carboxyl terminus of the protein. Ldb18-3GFP was visualized in live cells also expressing mCherry-Tub1 to determine its localization relative to microtubules. Ldb18 localizes near the spindle pole body (SPB), often on the daughter-bound pole (Figure 4A). Asymmetric localization on the daughter-bound SPB has been reported for Dyn1 and to a lesser extent for Jnm1 (Grava et al. 2006). Additionally, Ldb18 is observed at the distal ends of astral microtubules (Figure 4, B and C). Thus, Ldb18 localization is consistent with its being a component of the dynactin complex.
Ldb18 is required for dynactin integrity:
When the dynactin complex is disrupted, a subcomplex of p24, dynamitin, and p150Glued can be isolated (Eckley et al. 1999); however, the precise nature of the interactions among these shoulder-sidearm proteins is not known. The phenotype of ldb18Δ cells suggests that Ldb18 is required for dynactin function. To test whether Ldb18 is needed for the integrity of the dynactin complex, we measured the coprecipitation of Jnm1 with Nip100 or Arp1 in wild-type and ldb18Δ cells. Loss of Ldb18 does not affect the interaction between Jnm1 and the Arp1 filament (Figure 5A). However, the interaction between Jnm1 and Nip100 is disrupted in the absence of Ldb18 (Figure 5B), indicating that p24 is important in mediating the p150Glued–dynamitin interaction. Similar results were obtained by measuring the coprecipitation of Arp1 or Nip100 with Jnm1 (data not shown). The amount of Jnm1 that is precipitated with Nip100 in the ldb18Δ strain is ∼ of the Jnm1 precipitated in the wild-type strain (Figure 5C). Thus, loss of Ldb18 reduces the Jnm1–Nip100 interaction by ∼95%. This result is consistent with Ldb18 being part of the shoulder-sidearm dynactin subcomplex. Earlier studies demonstrated reduced Nip100 binding to the Arp1 filament in jnm1Δ cells (Kahana et al. 1998), indicating Jnm1 was responsible for sidearm attachment. In ldb18Δ cells, Jnm1 still interacts with Arp1 and only Nip100 dissociates from the complex. Thus, Ldb18 is essential for dynactin function because it is required to tether the microtubule and dynein-binding Nip100 arm to the dynactin complex through its interaction with Jnm1.
Our genetic and biochemical data indicate that Ldb18 is the yeast homolog of the mammalian dynactin protein p24 and is essential for Nip100 (p150Glued) attachment to dynactin. We identified ldb18Δ, as well as deletions of a number of other genes in the dynein pathway, through a SGA screen with stu1-5. The simplest explanation for synthetic lethality of a double-mutant combination is redundancy of function. Stu1 is localized on spindle microtubules and is necessary for SPB separation. Dynein/dynactin, on the other hand, exerts force on astral microtubules at the cell cortex that pulls the daughter-bound SPB through the mother-bud neck. Assuming the mother-bound SPB is somehow tethered in the mother cell, dynein/dynactin activity could also provide a SPB-separating activity. Previous evidence indicates that dynein does play a role in SPB separation. First, deletions of a number of genes encoding dynein/dynactin proteins (dynΔ, jnm1Δ, ldb18Δ, and nip100Δ) are synthetic lethal with loss of Cin8, a kinesin-5 protein that is also required for SPB separation (Geiser et al. 1997; Tong et al. 2001). In addition, anaphase SPB separation depends on both Cin8 and Dyn1 (Saunders et al. 1995). Thus, we favor the view that synthetic lethality between stu1-5 and loss of dynein activity is due to their overlapping roles in SPB separation.
We thank Tim Sackton for help with statistics on the protein homology and Aster Legesse-Miller for constructing the Ldb18-3GFP strain. We also thank Mark Rose (Princeton, NJ) for generously providing Arp1, Jnm1, and Arp1 yeast two-hybrid vectors, Arp1 antibody, and dynactin epitope-tagged strains, Kelley Tatchell (Louisiana State University, LA) for the Jnm1 antibody, and Elmar Schiebel (Universität Heidelberg, Germany) for the mCherry-Tub1 strain. This work was supported by grants from the Canadian Institute of Health Research, Genome Canada, and Genome Ontario (to C.B.) and by National Institutes of Health grant GM-040479 (to T.C.H.). I.A.A. is funded by National Institutes of Health pre-doctoral grant GM-073576.
Communicating editor: S. Dutcher
- Received July 19, 2007.
- Accepted December 16, 2007.
- Copyright © 2008 by the Genetics Society of America