Stu2 uses a 15-nm parallel coiled coil for kinetochore localization and concomitant regulation of the mitotic spindle.

XMAP215/Dis1 family proteins are potent microtubule polymerases, critical for mitotic spindle structure and dynamics. While microtubule polymerase activity is driven by an N-terminal tumor overexpressed gene (TOG) domain array, proper cellular localization is a requisite for full activity and is mediated by a C-terminal domain. Structural insight into the C-terminal domain's architecture and localization mechanism remain outstanding. We present the crystal structure of the Saccharomyces cerevisiae Stu2 C-terminal domain, revealing a 15-nm parallel homodimeric coiled coil. The parallel architecture of the coiled coil has mechanistic implications for the arrangement of the homodimer's N-terminal TOG domains during microtubule polymerization. The coiled coil has two spatially distinct conserved regions: CRI and CRII. Mutations in CRI and CRII perturb the distribution and localization of Stu2 along the mitotic spindle and yield defects in spindle morphology including increased frequencies of mispositioned and fragmented spindles. Collectively, these data highlight roles for the Stu2 dimerization domain as a scaffold for factor binding that optimally positions Stu2 on the mitotic spindle to promote proper spindle structure and dynamics.

INTRODUCTION

Stu2 is a yeast member of the XMAP215/Dis1 microtubule polymerase family (Gard and Kirschner, 1987). It was isolated in a screen for suppressors of a cold sensitive β-tubulin (tub2) mutation that caused a defect in mitotic spindle formation (Wang and Huffaker, 1997). Its depletion leads to reduced cytoplasmic microtubule dynamics, increasing the pause state and decreasing the frequencies of catastrophe and rescue (Kosco ). During mitosis, Stu2 localizes at kinetochores and spindle pole bodies and along the mitotic spindle itself (Chen ; Usui ; Gandhi ). Its activity is required for kinetochore microtubule dynamics and stability, centromere positioning, and anaphase spindle elongation (He ; Kosco ; Severin ; Pearson ).

XMAP215/Dis1 family members have a multi-domain architecture. Nonyeast members contain an array of three to five N-terminal tubulin-binding tumor overexrpessed gene (TOG) domains, a basic linker, and a C-terminal microtubule plus-end tracking domain that are collectively used to position the polymerase and drive microtubule polymerization (Gard ; Brittle and Ohkura, 2005; van der Vaart ; Li ). In contrast, yeast family members contain only two N-terminal TOG domains, a central basic linker, and a unique C-terminal dimerization domain predicted to be a coiled coil (Nabeshima ; Nakaseko ; Wang and Huffaker, 1997; Garcia ; Al-Bassam ). Stu2 microtubule polymerase activity is dimerization dependent and functionally operates with a pair of two TOG domains (four total). The orientation of the coiled coil (parallel vs. antiparallel) puts mechanistic restrictions on how these four TOG domains collectively engage tubulin subunits to promote their polymerization. The orientation of the coiled coil and whether it plays a role in Stu2 localization and polymerase function remains to be determined.

We present the crystal structure of the Stu2 coiled coil (Stu2CC) that reveals a 15-nm-long parallel coiled coil. The coiled coil has two conserved, surface exposed regions. Mutating determinants in either conserved region compromises yeast growth rates under conditions that stress microtubules. Mutations in these conserved regions affect Stu2 spindle localization and compromise spindle structure and dynamics. This work establishes distance constraints for the dimer’s N-terminal TOG domains, unveils roles for the coiled coil in proper Stu2 spindle localization, and implicates determinants along the coiled coil as spindle factor binding sites.

RESULTS

The Stu2 conserved C-terminal domain is a homodimeric α-helical domain

The Stu2 C-terminal domain is conserved across yeast species, is predicted to be a coiled coil (Figures 1, A and B), and is required for full Stu2-dependent activity (Al-Bassam ). We used size exclusion chromatography–multi-angle light scattering (SEC-MALS) to determine the oligomeric state of the Stu2 C-terminal domain. Stu2CC (residues 655–760) eluted off the gel filtration column as a single homodimeric peak (experimental mass: 27.6 kDa, formula weight: 13.5 kDa; Figure 1C). This agrees with previous studies showing that full-length Stu2 is dimeric and that the Stu2 C-terminal domain is necessary for oligomerization (De Wulf ; van Breugel ; Al-Bassam ).

FIGURE 1:

The Stu2 C-terminal domain is a dimeric coiled coil with two conserved regions. (A) Stu2 domain architecture, showing the two N-terminal TOG domains (TOG1 and TOG2, shown in light and dark orange, respectively), a central basic linker region implicated in microtubule binding (dark blue), and the C-terminal coiled-coil domain (rainbow). (B) Sequence alignment of the Stu2CC region across diverse yeast species. Residues with 80% identity and 80% similarity are highlighted in green and yellow, respectively. Conserved regions (CRI and CRII) are boxed and point mutations in these regions are indicated in red below the alignment. Residue numbers are indicated for Stu2. Heptad repeat positions are indicated in lowercase letters (a–f) above the alignment. Residues involved in the dimerization interface are colored magenta. The accessible surface area (ASA, black) and buried surface area (BSA, gray) for each residue in chain A is presented above the alignment, calculated using the ePISA server (Krissinel and Henrick, 2007). ASA is calculated based on the solvent accessibility of a residue in the absence of chain B and symmetry mates. Regions from the alignment shown in molecular detail in Figure 2, D–F, are boxed in cyan, green, and yellow, respectively. (C) SEC-MALS analysis of Stu2CC. Stu2CC forms a dimer in solution at pH 7.5 with an experimentally determined molecular weight of 27.6 kDa vs. its 13.5-kDa formula weight. (D) CD spectra of Stu2CC (11 μM) at 23°C, showing minima at 208 and 222 nm, indicative of α-helical secondary structure. (E) CD thermal melt of Stu2CC monitoring CD signal at 208 nm (gray trace) and 222 nm (black trace). The inflection point of each melt curve occurs at 51.5°C. (F) Cartoon diagram of the Stu2CC structure with both chains colored in an indigo to red spectrum from the N-terminus to the C-terminus. Image below is after a 90° rotation about the x-axis. (G) The Stu2CC structure shown in spherical format, colored gray with conserved resides colored in green and yellow as in Figure 1B. (H) Stu2CC electrostatics, contoured from –3 kBT/e (red) to +3 kBT/e (blue). The two orientations of Stu2CC in G and H correlate with the orientations in F.

FIGURE 2:

Stu2CC contains conserved and distinct structural features. (A) Stick diagram of the Stu2CC dimer, shown in an indigo to red spectrum from the N-termini to the C-termini. Regions shown in detail in D–F are boxed and their corresponding sequence is indicated at the top. (B) Standard homodimeric coiled-coil helical wheels showing the relative positions of heptad repeat residues a through f and a′ through f′. Positions a, d, a′, and d′ are located at the dimerization interface. (C) Interchain a–a′ (red) and d–d′ (cyan) distances (measured in Å) between homologous Stu2CC Cα atoms at the dimer interface are plotted. The asymmetric packing of the conserved W678 side chains underlies the helical flare that spans E675 to K688 (indicated). Red and cyan dashed lines represent the average Cα–Cα′ distance for the a–a′ (5.9 Å) and d–d′ (6.5 Å) measurements, respectively (averages were calculated excluding residues in the helical flare). (D) Zoom view of the CRI helical flare involving W678. W678 and W678′ are packed asymmetrically. W678′ is stabilized by a network of hydrogen bonds (dashed lines) involving the conserved residues E681 and R682′. An intrahelical hydrogen bond involving T679′ and the E675′ backbone carbonyl further stabilizes the region. Residues are shown in cyan stick format. (E) Zoom view of CRII spanning E702–R707 involving an extensive regional dimerization interface stabilized by a network of hydrogen bonds (dashed lines). Selected residues are shown in green stick format. Image is viewed out of the bore of the helices after a 90° rotation about the x-axis of the image shown in A. (F) Zoom view of the CRII segment containing the sequence LRSK (residues 724–727) shown in yellow stick format. The surface exposed R725 and S726 residues are not involved in dimer formation or intra- or interhelical hydrogen bonding. CRI and CRII residues mutated in this study are indicated in red in A, D, and F.

We next analyzed Stu2CC using circular dichroism (CD). The CD spectrum showed minima at 208 and 222 nm, indicative of α-helical secondary structure (Figure 1D). Thermal melt analysis revealed that Stu2CC melted cooperatively with an inflection point at 51.5°C (apparent Tm), indicative of a relatively stable structure (Figure 1E).

Stu2CC is a 15-nm parallel coiled coil

To determine the structure of Stu2CC, we crystallized selenomethionine-substituted Stu2CC and collected 2.5-Å-resolution single-wavelength anomalous dispersion data. Crystals belonged to the space group P21 and contained two protein molecules in the asymmetric unit that formed a single homodimeric coiled coil. The Stu2CC structure was built and refined to R and Rfree values of 25.7 and 29.2, respectively. Crystallographic statistics are presented in Supplemental Table S1.

The Stu2CC domain forms a parallel coiled coil over 99 residues (amino acids 657–755) that spans 14.8 nm (Figure 1F). The dimerization interface buries 2670 Å2 of solvent accessible surface area from each chain and involves van der Waals contacts as well as hydrogen bond networks. The domain has a left-handed super-coil that rotates each helix ∼270° about the coiled coil’s axis. Cortexillin I is the most similar structure currently available in the Protein Data Bank (PDB) (Dali server: 2.3 Å rmsd over 99 aligned Cα atoms, Z = 7.8, pdb 1D7M; Burkhard ; Holm and Rosenström, 2010). Our structure confirms the homodimeric nature of the Stu2 coiled coil.

Stu2CC has two distinct conserved regions

We next mapped conservation (as delineated in Figure 1B) and electrostatics on the Stu2CC structure. Two distinct areas of conservation were apparent: conserved region I (CRI), spanning residues 671–682, and conserved region II (CRII), spanning residues 702–745 (Figure 1G). Stu2CC’s electrostatic profile is polarized with a basic region concentrated within the N-terminal half of the coiled coil and an acidic region concentrated at the domain’s C-terminal region (Figure 1H). The coiled coil’s N-terminal basic region abuts a positively charged linker (not present in the structure) that bridges TOG2 and the coiled coil and is involved in microtubule binding (Wang and Huffaker, 1997; Al-Bassam ). The electrostatic surfaces of CRI and CRII are interspersed with hydrophobic content, suggesting that factors that engage these regions would bind using electrostatic interactions and van der Waals contacts.

At CRI, the Stu2 coiled coil locally flares, deviating significantly from canonical coiled-coil structure (Figure 2A, cyan box). The closest approach of equivalent Cα atoms along a coiled coil occurs at heptad repeat positions a–a′ and d–d′ (Figure 2B). Analysis of the a and d Cα–Cα′ interchain distances across residues 661–755 revealed a flare spanning residues E675–L685 (Figure 2C). Within this region, the conserved tryptophan W678 is packed at the dimerization interface. W678 and W678′ adopt distinct rotameric states and are rotated 180° relative to one another. This positions each tryptophan’s indole ring amine on the same face of the coiled coil (Figure 2D). While most of the coiled coil has approximate twofold symmetry, the asymmetric packing of W678 and W678′ creates local asymmetry. The asymmetry generates nonequivalent surfaces on each side of CRI which may limit the stoichiometry or binding affinity of factors that engage CRI.

Stu2CC CRII also displays a number of unique features. The first feature spans residues E702–R707 and involves a network of hydrogen bonds at the dimerization interface (Figure 2, A, green box, and E). At the interface, the asparagine N703 and N703′ side-chain amide groups are oppositely oriented and interact through a hydrogen bond between the N703 Oδ and the N703′ Nδ. Flanking these residues are the arginine-glutamate perimeter pairs R707–E702′ and R707′–E702. The complementary arginine and glutamate residues are differentially positioned to pair and stabilize each respective N703 residue. A second highly conserved feature within CRII is the sequence LRSKE (residues 724–728; Figures 1B and 2, A, yellow box, and F). While L724 and the aliphatic portion of the K727 side chain are positioned at the dimerization interface, R725, S726, E728, and the K727 side-chain amine are solvent accessible and are not involved in intra- or interhelical stabilization. The surface accessibility of these residues suggests that they are involved in factor binding.

The Stu2CC structure does not contain an apparent consensus trigger sequence found in many coiled coils that initiates coiled-coil formation (Burkhard ; Kammerer ; Steinmetz ). Trigger sequences employ intra- and interhelical hydrogen bonding to promote autonomous, regional helical folding, and concomitant homodimerization. The element that triggers Stu2CC coiled-coil formation remains to be determined.

Mutations in Stu2CC CRI and CRII yield growth defects under conditions that stress microtubules

We designed targeted point mutations at the STU2 genomic locus in CRI and CRII to test the role of these regions in Stu2 function. A W678L mutation in CRI (Stu2 CRImut) replaced the tryptophan with leucine, a residue common to coiled-coil interfaces. In a separate strain, a double R725E, S726A mutation in CRII (Stu2 CRIImut) was introduced (see Supplemental Table S2 for strain details). The viability of these strains contrasts with the lethality caused by stu2 deletion (Wang and Huffaker, 1997). We analyzed the ability of the stu2 strains to grow under normal growth conditions and under conditions that stress the microtubule cytoskeleton: decreased temperature (18°C) and/or benomyl treatment. (Figure 3A). In the absence of benomyl, all strains grew indistinguishably at 32° and 40°C. However, mutants placed at 18°C in the absence of benomyl displayed severe growth inhibition. Mutants also displayed enhanced benomyl sensitivity when grown on plates at 32° or 40°C or in liquid culture at 32°C (Figure 3, A–C). These results support a role for Stu2 CRI and CRII in regulating the microtubule cytoskeleton and cell division. In support, after overnight growth at 18°C, Stu2 CRImut and CRIImut cells grew massive in size (>2 × WT) while WT Stu2-myc cells had no gross cellular defect (unpublished data).

FIGURE 3:

Mutations in Stu2CC CRI and CRII lead to increased benomyl and cold sensitivity. (A) Yeast strains (WT Stu2-myc, Stu2-myc-CRImut, and Stu2-myc-CRIImut) were plated using a 10-fold serial dilution series on 0 or 10 μg/ml benomyl (Ben). Duplicate plates were put at 18°, 32°, and 40°C. Benomyl sensitivity is indicated by decreased growth at 32 and 40°C (no growth is observed even for WT yeast on benomyl plates at 18°C). Cold sensitivity is apparent for 18°C plates with no benomyl treatment. Data shown are representative of two independent experiments. (B) Yeast growth curves. Yeast strains were diluted to an OD660 of 0.1 in YPD and grown at 32°C shaking at 250 rpm. Growth curves plot OD660 over a 24-h time period. All strains grown in YPD at 32°C have similar growth rates. (C) Yeast strains were grown as in B, but YPD was supplemented with 10 μg/ml benomyl. Introducing benomyl slows the growth rate for both the CRImut and CRIImut strains. Growth experiments shown in B and C are from a single experimental day and are representative of three experimental replicates. (D, E) SEC-MALS analysis of Stu2CC CRImut (D) and CRIImut (E). Both constructs are homodimeric based on their experimentally determined masses. (F) CD spectra of WT and mutant Stu2CC constructs CRImut and CRIImut at 23°C all show signature minima at 208 and 222 nm, indicative that the mutations do not perturb the domain’s α-helical character. (G, H) CD thermal melts of mutant Stu2CC constructs CRImut (G) and CRIImut (H). Stu2CC CRImut and CRIImut have cooperative unfolding profiles with inflection points at 46.5 and 51.5°C, respectively. The CRI mutation W678L lowered the Tm by 5°C relative to the WT construct (see Figure 1E, 51.5°C, indicated in G and H by a green arrow). The CRII double mutations R725E and S726A did not change the Tm from the WT value. SEC-MALS and CD data are representative of two experimental replicates.

We analyzed mutant Stu2CC constructs using SEC-MALS and CD to determine if the CRI and CRII mutations we introduced compromised the domain’s stability or dimeric state. Stu2CC CRImut and Stu2CC CRIImut each eluted as a single peak off a gel filtration column with an experimentally determined mass of 27.0 kDa, indicative of homodimerization (Figure 3, D and E). Each mutant also produced an α-helical CD spectrum with minima at 208 and 222 nm (Figure 3F). Thermal melt analysis gave apparent Tm values of 46.5°C for Stu2CC-CRImut and 51.5°C for Stu2CC-CRIImut (Figure 3, G and H) that were similar to WT (51.5°C; Figure 1E). These results indicate that both mutants are stable and dimeric over the temperature range analyzed in vivo, giving us license to move forward with functional analyses.

Stu2 CRI and CRII determinants contribute to spindle morphology

We hypothesized that growth defects in the Stu2 CRImut and CRIImut strains arose from mitotic failure. We introduced mCherry-Tub1 into our Stu2 strains and analyzed spindle morphologies at 32°C. Severe spindle defects were observed in the Stu2 CRImut and CRIImut strains (Figure 4, A and B). Stu2 CRImut and CRIImut strains had significantly more mispositioned spindles as compared with WT (CRImut = 43%, CRIImut = 19%, WT = 6%). Unexpectedly, we observed many cells with fragmented spindles reminiscent of a temperature-sensitive stu2 mutation (Severin ). While spindle fragmentation was not observed in WT cells, it was observed in CRImut cells with a frequency of 15% and in CRIImut cells with a frequency of 27%. Collectively, abnormal spindles were observed in 58% of Stu2 CRImut mitotic cells, 46% of Stu2 CRIImut mitotic cells, and only 6% of WT Stu2 mitotic cells. The higher percentage of aberrant spindles in the Stu2 CRImut strain correlates with a more pronounced growth defect under benomyl treatment (Figure 3C).

FIGURE 4:

Mitotic spindle structure is disrupted in the Stu2-CRImut and Stu2-CRIImut strains. (A) Representative images of spindle morphology phenotypes scored including normal, mispositioned, and fragmented spindles. Images taken are from the Stu2-CRIImut strain grown at 32°C. Rows show phase, mCherry-Tub1, and merge. Scale bar is 1 μm. (B) Top: Cartoon representation of each mitotic spindle phenotype scored. Gray scale color correlates with the bars in the histogram shown at the bottom. Spindle morphologies were scored and binned for WT Stu2, Stu2-CRImut, and Stu2-CRIImut strains. Stu2-CRImut and Stu2-CRIImut strains have a significant increase in the frequency of aberrant spindle morphology (mispositioned or fragmented) as compared with the WT strain (p < 0.005; two-tailed unpaired Student’s t test). Error bars represent SD (n = 95 [WT], 102 [CRImut], 85 [CRIImut]). Data for each strain were collected across at least three independent experimental days.

CRI and CRII determinants regulate Stu2 positioning and spindle structure

We next analyzed Stu2-GFP localization in CRImut, CRIImut, and WT mitotic cells. Spindle poles were labeled using Spc29-RFP. In ∼90% of WT Stu2-GFP mitotic cells, two Stu2-GFP foci appeared internal to the two Spc29-RFP foci, correlating with the position of kinetochore microtubule (kMT) plus ends (Figure 5A). Surprisingly, the distribution and number of discreet Stu2-CRImut-GFP and Stu2-­CRIImut-GFP foci along the spindle axis was dramatically altered. Stu2-CRImut-GFP showed aberrant spindle localization in ∼50% of mitotic cells, grown at either 18° or 32°C (Figure 5C). Stu2-CRIImut-GFP showed aberrant spindle localization in 41% of mitotic cells grown at 32°C and 58% of mitotic cells grown at 18°C. When the Stu2 CRImut and CRIImut GFP strains displayed two foci, the foci were anisotropically elongated along the spindle axis. Moreover, the amount of Stu2 on the spindle (integrated GFP fluorescence intensity) in foci directly proximal to the spindle pole body was significantly increased in the Stu2-CRImut and Stu2-CRIImut strains compared with WT (Figure 5D). The dispersal of Stu2-CRImut and Stu2-CRIImut along the spindle axis could compromise spindle dynamics, potentially leading to the fragmented and/or broken spindles observed in these strains. Overall, the mislocalization of mutant Stu2-GFP indicates a role for CRI and CRII in proper Stu2 kinetochore localization.

FIGURE 5:

Mutations in the Stu2CC CRI and CRII regions alter Stu2 distribution and localization on the yeast mitotic spindle and affect spindle length. (A) Stu2-GFP localizes to the metaphase mitotic spindle primarily as two foci, just interior to the Spc29-RFP foci that label the two spindle poles. Yeast grown at 32°C, scale bar: 1 μm. Mutant Stu2 strains CRImut and CRIImut show altered distributions along the spindle. The Stu2-GFP CRImut and CRIImut strains show defects in localization, including aberrant numbers of foci along the metaphase spindle. (B) Intensity line scan across a normal metaphase spindle shown in A. Two Spc29-RFP maxima define the mother (major peak) and daughter (minor peak) spindle pole bodies. Two WT Stu2-GFP foci localize just interior to the Spc29-RFP foci and are relatively uniform in intensity and distribution. (C) Quantification of Stu2-GFP localization to metaphase spindles between Spc29-RFP-labeled poles for strains grown at 32°C or grown at 32°C and shifted to 18°C for 1 h prior to imaging. Normal Stu2-GFP localization appears as two uniformly round foci as seen in A (left panel). Anything that deviated from this localization was classified as abnormal. Stu2-CRImut and CRIImut strains had a significant increase in abnormal Stu2 localization relative to the WT strain at corresponding temperatures. (D) Stu2-GFP localization proximal to the spindle poles (32°C) is enhanced in the Stu2-CRImut and Stu2-CRIImut strains. **** Indicates p value < 0.0001. (E) Stu2-CRImut and Stu2-CRIImut strains show an increased apparent spindle length at 18°C. The black dotted line represents the WT mean at 32°C. All data with black errors bars are not statistically different from one another. The two gray scatter plots with red error bars indicate data that is statistically different from the black scatter plots (p value < 0.0001, two-tailed unpaired Student’s t test) but not statistically different from one another. Scatter dot plots in D and E indicate the mean plus and minus the SD. (F) Ndc80-mCherry localization and foci separation are not perturbed in the Stu2-GFP CRImut and CRIImut strains. Left: WT and mutant Stu2-GFP localization relative to Ndc80-mCherry foci. Right: mean distances between Stu2-GFP foci and Ndc80-mCherry foci for WT and mutant Stu2 strains. Yeast grown at 32°C, scale bar: 1 μm. Errors indicate SD. Data presented in C–F represent data collected from at least three independent experimental days per stain, per assay.

We next analyzed the effect of CRI and CRII mutations on metaphase spindle length and kinetochore positioning. No significant change in metaphase spindle length was observed at 32°C using Spc29-RFP as a pole marker. In contrast, metaphase spindle length was significantly longer in the mutant strains grown at 18°C (Stu2-CRImut: 1.18 μm, Stu2-CRIImut: 1.25 μm) as compared with WT (0.89 μm) (Figure 5E). This may indicate that the mutants could not properly transition from metaphase to anaphase. In support, we observed a larger number of late metaphase spindles in the mutant stu2 strains compared with WT (unpublished data). We inquired whether the Stu2 mutants compromised kinetochore organization using Ndc80-mCherry as a kinetochore marker. Neither mutant significantly perturbed the localization of Ndc80 at kinetochores, and sister kinetochore separation remained unchanged. Collectively, these data show that CRI and CRII play important roles in Stu2 localization, spindle architecture, and spindle dynamics.

DISCUSSION

We find that the Stu2CC domain forms a 15-nm parallel homodimeric coiled coil. This has implications for the relative positioning of the N-terminal TOG domains. Both TOG1 and TOG2 have αβ-tubulin binding activity and are required for Stu2’s microtubule polymerase activity (Al-Bassam ; Slep and Vale, 2007; Ayaz , 2014). The parallel coiled coil functionally permits the dimer’s four TOG domains to drive microtubule polymerization in proximity to one another, potentially by promoting longitudinal and/or lateral tubulin:tubulin contacts. If the coiled coil was anti-parallel, then the N-terminal TOG domain pairs would be spatially separated, limiting potential cooperative activity. Previous work using a monomeric TOG1-TOG2 construct that lacked the coiled coil formulated a TOG-dependent noncooperative tubulin-tethering mechanism that enhanced microtubule polymerization (Ayaz ). The Stu2CC structure primes the field to investigate whether full-length, dimeric Stu2 uses its four TOG domains to tether tubulin or to cooperatively bind tubulin and enhance lateral and/or longitudinal tubulin-tubulin interactions.

The Stu2CC structure highlights two conserved regions, CRI and CRII. CRI contains a conserved tryptophan (W678) that packs asymmetrically with its mate, causing a local flare in the coiled coil’s diameter. A W678L mutation lowers the apparent Tm by 5°C, indicating that the W678–W678′ interface (potentially in the context of the flanking residues) provides enhanced stability to the coiled coil. The conserved nature of the CRI tryptophan and the structurally unique helical flare it causes suggests that this region may be a site for factor binding. The asymmetric packing of W678 and W678′ suggests that a binding factor may engage it in a 1:2 factor:Stu2 stoichiometry. This contrasts with CRII, which exhibits twofold regional symmetry and likely engages binding factors in a 2:2 factor:Stu2 stoichiometry.

Structure-based mutagenesis within CRI and CRII has unveiled critical roles for the coiled coil in Stu2-dependent mitotic function. Mutating residues in either CRI or CRII compromised growth rates when yeast were challenged with cold or benomyl. Importantly, this was not because the mutations compromised Stu2 dimerization. Even in the absence of cold or benomyl treatment, these mutations led to an increased percentage of mispositioned and/or fragmented spindles and perturbed Stu2 localization along the spindle. While wild-type (WT) Stu2-GFP appeared as two foci along the spindle, CRImut and CRIImut constructs often appeared as multiple (>2) foci, anisoptropically distributed along the spindle axis. This localization pattern suggests an uncoupling of microtubule plus ends from kinetochores or a defocusing of the kinetochores. Analysis of Ndc80-mCherry localization in the Stu2 mutant strains did not reveal aberrant Ndc80 localization. This suggests that CRI and CRII determinants promote Stu2 kinetochore engagement and, when mutated, may uncouple microtubule plus ends from kinetochores. When yeast were challenged with cold (18°C), the metaphase spindle length increased in the CRImut and CRIImut strains but not in the wild-type strain. This suggests that a tension-sensing clutch mechanism, normally engaged in metaphase, became released in these mutants when challenged with cold. Recent work has identified Stu2/Dis1 as a Ndc80 complex binding factor involved in tension sensing (Measday ; Hsu and Toda, 2011; Tang ; Aravamudhan ; Miller ). Whether the Ndc80 complex engages CRI or CRII remains to be determined. Identifying the factors that do bind these regions will be critical for our understanding of kinetochore-microtubule attachment and dynamics. The Stu2CC structure we presented here and the conserved determinants we identified will be useful tools to resolve these questions.

MATERIALS AND METHODS

Cloning, expression, and purification of the Stu2 dimerization domain

STU2 DNA encoding residues 655–760 (Stu2CC) was cloned from Saccharomyces cerevisiae genomic DNA into pET28 (Novagen, Millipore Sigma, St. Louis, MO) yielding an N-terminal, thrombin-cleavable H6 tag. H6-Stu2CC was expressed in BL21 DE3 (pLysS) Escherichia coli and induced using 0.1 mM isopropyl-1-thio-β-d-galactopyranoside (Teknova, Hollister, CA) at 18°C for 16 h. Cells were harvested, resuspended in lysis buffer (25 mM Tris, pH 8.0 [Fisher Chemical, Thermo Fisher Scientific, Waltham, MA], 300 mM sodium chloride [Fisher Chemical], 10 mM imidazole [Fisher Chemical], 0.1% β-mercaptoethanol [β-ME] [Acros Organics, Thermo Fisher Scientific]) and lysed by sonication. After lysis, phenylmethlysulfonyl fluoride (BioSynth, Wahlkreis of Rorschach, St. Gallen, Switzerland) was added to a final concentration of 1 mM. Lysate was clarified by centrifugation at 23,000 × g, and the supernatant loaded onto a Ni2+-NTA (Qiagen, Venlo, Netherlands) column. H6-Stu2CC was eluted using a 250-ml 10–300 mM imidazole gradient. Eluted protein was pooled, calcium chloride (Fisher Chemical) was added to 1 mM final concentration, and 0.125 mg bovine α-thrombin (Haematologic Technologies, Essex Junction, VT) was added to cleave the N-terminal H6 tag. Stu2CC was filtered over 1 ml benzamidine sepharose (GE Healthcare, Marlborough, MA), exchanged into 25 mM Tris, pH 8.0, 0.1% β-ME using a Millipore Ultrafree 3k molecular weight cut-off concentrator, and loaded onto a Q Sepharose Fast Flow column (GE Healthcare). Stu2CC was eluted using a 0–1 M sodium chloride gradient over a 250-ml volume. Peak protein fractions were pooled, exchanged into storage buffer (25 mM Tris, pH 8.0, 100 mM sodium chloride, 0.1% β-ME), aliquoted, frozen in liquid nitrogen, and stored at -80°C. The final Stu2CC domain contained an N-terminal cloning artifact, residues: GSHM. Selenomethionine (SeMet) (Sigma Aldrich, St. Louis, MO)-substituted Stu2CC was produced in B834 E. coli as described previously (Leahy ) and purified as described above for the native construct.

Size exclusion chromatography–multi-angle light scattering

A 100-μl sample of 45 μM Stu2CC was injected onto a Superdex 200 10/300 GL gel filtration column (GE Healthcare) in 25 mM HEPES, pH 7.5 (Calbiochem, MilliporeSigma, St. Louis, MO), 300 mM sodium chloride, 0.1% β-ME, 0.2 g/l sodium azide (Fisher Chemical) and run in-line with a Wyatt DAWN HELIOS II light scattering instrument and a Wyatt Optilab T-rEX refractometer (Wyatt Technology Corp., Goleta, CA). Molecular weight was calculated with light scattering and refractive index data using the Wyatt Astra V software package (Wyatt Technology Corp.). SEC-MALS data presented are representative of experiments conducted in duplicate.

Circular dichroism

Circular dichroism spectra of Stu2CC WT and mutant constructs were collected at room temperature (23°C) using a Chirascan-plus CD spectrometer (Applied Photophysics, Leatherhead, UK). Stu2CC constructs were diluted to a final concentration of 11 μM in 10 mM sodium phosphate buffer (pH 7.5) (Fisher Chemical), 50 mM sodium fluoride (Fisher Chemical). Spectra were recorded from 260 to 185 nm with a step size of 0.5 nm using a 1-mm-path-length cuvette. The time per point was kept at 1.25 s. A base-line CD spectrum of the buffer alone was taken and subtracted from each Stu2CC spectrum. Spectra were smoothed in the Chirascan-plus software. CD melt spectra were obtained, monitoring 208 and 220 nm in 1°C steps from 20–94°C, with the time per point maintained at 1.25 s. CD data presented are representative of experiments conducted in duplicate.

Crystallization

Stu2 crystals were obtained by hanging drop vapor diffusion using 2 μl of purified SeMet-substituted Stu2CC (20 mg/ml stock) and 2 μl of the 1-ml well solution: 13% polyethylene glycol 8000 (Honeywell Fluka, Morris Plains, NJ), 200 mM magnesium acetate (Fisher Chemical), 100 mM sodium cacodylate, pH 5.75 (Fisher Chemical). Crystals grew at 20°C into thin, individual needles over a 1-wk period. Crystals were transferred into paratone-N (Hampton Research, Aliso Viejo, CA) and flash frozen in liquid nitrogen.

Diffraction data collection, structure determination, and refinement

Diffraction data were collected on a single SeMet-Stu2CC crystal at the Advanced Light Source beamline 8.3.1 at 100 K with 1° oscillations over 360° from a single crystal. Data were indexed, integrated, and scaled using HKL2000 (Otwinowski and Minor, 1997). The structure was determined using the AutoSol program (PHENIX crystallographic suite; Adams ). Ten selenium sites were identified in the heavy atom search in space group P21, accounting for the 10 ordered SeMet residues in the asymmetric unit. The model was built using AutoBuild (PHENIX; Adams ) and refined iteratively through manual builds in Coot (Emsley ) followed by refinement runs using phenix.refine (PHENIX; Adams ). Refinement statistics were monitored using a Rfree, calculated using 10% of the data, randomly excluded from refinement (Brünger, 1992). Refinement runs utilized a maximum-likelihood Hendrickson Lattman target function. The final model includes two Stu2 chains (chain A: residues 657–755; chain B: residues 658–755) and 25 water molecules. Crystallographic data and refinement statistics are presented in Supplemental Table S1. Atomic coordinates for the Stu2CC structure have been deposited to the PDB under accession code 6BL7.

Generation of STU2 targeting plasmids

A DNA fragment of STU2 corresponding to residues 357–888 was amplified from wild-type genomic DNA using a forward primer flanked with a BamHI site and a reverse primer flanked with a HindIII site. The PCR product was digested with BamHI and HindIII (New England Biolabs, Ipswich, MA) and cloned into a pFA6-13x-myc::NAT vector yielding a targeting plasmid with STU2-357-888-13x-myc::NAT. Within the targeting vector, specific point mutations (W678L or R725E, S725A) were introduced using site-directed mutagenesis and verified by sequencing. Primers to amplify and target the engineered STU2 sequence to the endogenous locus were designed using homology-flanking arms of sequence just upstream of Stu2 residue 357 and just downstream of the STU2 STOP codon. A similar protocol was used to generate STU2 strains with a C-terminal GFP tag and HIS selection. Western blot analysis was used to confirm normal expression levels of Stu2 in the WT and mutant STU2-13x-myc strains generated.

Generation of STU2 yeast strains

All strains in this study use YEF473A as the parental strain. PCR product with homology arms was gel extracted and transformed into competent WT yeast cells. NAT- or HIS-positive isolates were screened using a forward primer 401 base pairs upstream and a reverse primer 499 base pairs downstream of the STU2 coding sequence. Isolates containing the integration at the genomic locus were confirmed by PCR. Two positive isolates from each desired background (WT-STU2-13x-myc, STU2-W678L-13x-myc, STU2-R725E-S725A-13x-myc, WT-STU2-GFP, STU2-W678L-GFP, and STU2-R725E-S725A-GFP) were chosen to move forward with functional analyses. PCR product amplified off of genomic DNA was sent for sequencing to verify introduction of the desired mutation(s). To visualize spindles, pRS305:MET3p:mCherry-TUB1 (a gift from Wei-Lih Lee, Department of Biological Sciences, Dartmouth College, Hanover, NH) was digested with HindIII (New England Biolabs) and transformed into the STU2-13x-myc strains (Markus ). Strains were screened using fluorescence microscopy to confirm integration of the plasmid (Markus ). To determine Stu2 localization relative to the spindle pole body and the kinetochore, we generated independent STU2-GFP strains with either Spc29 (spindle pole body component; Elliott ) tagged at the C-terminus with RFP or Ndc80 (kinetochore component) tagged at the C-terminus with mCherry using a PCR cassette (Longtine ). Yeast strains used and generated in this study are presented in Supplemental Table S2.

Growth curves of yeast and benomyl sensitivity (liquid culture)

Yeast strains were diluted into either yeast extract–peptone–­dextrose (YPD) or YPD + 10 μg/ml benomyl (Sigma Aldrich) to an OD660 of 0.1 from an overnight culture. These cultures were placed at 32°C, shaking at 250 rpm. The OD660 was measured every 1–2 h. Cell culture density over time was plotted in Sigma Plot (Systat Software, San Jose, CA).

Yeast plate growth assays

Tenfold serial dilutions of overnight yeast cultures were pronged onto plates containing 0 and 10 μg/ml benomyl. Two replicate plates were put at 18°, 32°, and 40°C. Cells were allowed to grow on plates for 3–5 d. Each plate was scanned on day 3 and day 5. Images shown in Figure 3A are after 5 d.

Imaging and image analysis

All strains were imaged on a Nikon TE-2000E (Nikon Instruments, Melville, NY) inverted microscope with a 1.43 NA, 100× Plan-Apo objective controlled by Metamorph software (Molecular Devices, Sunnyvale, CA). Cells were grown shaking in liquid culture in log phase at 32°C. One hour prior to imaging, cells were either maintained at 32°C or shifted to 18°C. Two Stu2-GFP foci along a spindle axis were characterized as normal Stu2 localization, aberrant Stu2 localization was classified as either the loss of two distinct GFP foci resulting in one large focus or a line of Stu2-GFP fluorescence along the spindle or more than two foci that were at least 1/3 the intensity of Stu2-GFP at the kinetochores. Distances between foci were determined using Gaussian fits as described (Haase ). Stu2 focus intensity was determined by integrating background-subtracted intensities in a 9 × 9 pixel region of interest centered on the respective pole-proximal Stu2-GFP focus. Metaphase spindle length was determined from max projections by measuring the distance between metaphase spindle poles using Spc29-RFP as a marker. Accordingly, the spindle length measurement presented is the apparent length based on the max projection. Only spindles that had two resolvable Spc29-RFP foci were used for this analysis. Any collapsed spindles (where two foci could not be independently resolved) were not included in these measurements. Imaging for all yeast strains was conducted across at least three separate experimental days per assay. Mitotic cells were chosen for image analysis. Sample size was chosen to ensure a high theoretical statistical power (greater than 90%) to generate reliable p values.

Click here for additional data file.