Targeting Alp7/TACC to the spindle pole body is essential for mitotic spindle assembly in fission yeast.

The conserved TACC protein family localises to the centrosome (the spindle pole body, SPB in fungi) and mitotic spindles, thereby playing a crucial role in bipolar spindle assembly. However, it remains elusive how TACC proteins are recruited to the centrosome/SPB. Here, using fission yeast Alp7/TACC, we have determined clustered five amino acid residues within the TACC domain required for SPB localisation. Critically, these sequences are essential for the functions of Alp7, including proper spindle formation and mitotic progression. Moreover, we have identified pericentrin-like Pcp1 as a loading factor to the mitotic SPB, although Pcp1 is not a sole platform. Crown

Introduction

The microtubule cytoskeleton plays diverse roles in various cellular processes including cell motility, polarity, intracellular transport and chromosome segregation. Microtubules are hollow polymers consisting of α-/β-tubulin heterodimers. The distinct biophysical properties of α- and β-tubulins provide the intrinsic polarity of microtubules, with β-tubulin facing towards the dynamic plus end and α-tubulin defining the less dynamic minus end [1], [2], [3]. Nucleation of microtubules in vivo do not occur spontaneously; instead specialised structures called microtubule organising centres (MTOCs) [4] are required, in which the minus end of microtubules is embedded. In animal cells, the centrosome comprises a major MTOC, whilst in fungi the spindle pole body (SPB) plays an analogous role.

The Transforming Acidic Coiled Coil (TACC) family proteins were originally identified as a group of proteins implicated in human cancers [5], [6], [7], and the family is conserved throughout evolution. Human beings contain three TACC members (hTACC1-3) that localise to the centrosome and the spindle microtubules [8], [9]. A number of previous reports from different organisms indicate that centrosome/spindle microtubule localisation is the universal feature of individual TACC homologues, the members being crucial regulators of mitotic spindle assembly and dynamics [10]. Several lines of evidence show that TACC family members comprise adaptor proteins, by which these members act as a hub to interact with multiple effector molecules [11], [12], [13], [14]. Among the various binding effectors identified, the microtubule associated protein family, ch-TOG/Dis1/XMAP215, is a common interactor that is conserved throughout evolution from fission yeast to humans [10].

In fission yeast, Alp7/Mia1/TACC and Alp14/Mtc1/TOG form a complex, which is critical for mitotic and meiotic spindle assembly and proper chromosome segregation [15], [16], [17], [18]. Previous analysis of Alp7 and Alp14 uncovered functional distinctions between these two proteins. Alp7 is an SPB-targeting factor for this complex, which accumulates in the mitotic nucleus (yeast undergoes a closed mitosis) and specifically localises to the SPB in a RAN-GTPase dependent fashion [15], [19]. By contrast, Alp14/TOG acts as a microtubule stabiliser/polymerase [20], [21] only when it is targeted to the mitotic SPB via Alp7, resulting in localisation of this complex to the plus end of spindle microtubules including the kinetochore [15], [19], [22].

Intriguingly, despite the universal localisation to the centrosome/SPB, cis-acting sequences within the TACC proteins responsible for centrosomal/SPB recruitment remain unknown in any organism, although the localisation relies on the TACC domain [15], [23], [24]. Moreover, it is unclear how the TACC proteins interact with the centrosome/SPB; the identity of the receptor molecule(s) remaining elusive. In this work, we have addressed these two issues using fission yeast. We have determined the region within the TACC domain that is necessary and sufficient for SPB targeting, and further identified clustered five amino acid residues as critical sequences. In addition, an SPB component, conserved pericentrin-like Pcp1, physically interacts with Alp7 and is required for proper recruitment of Alp7 to the mitotic SPB.

Materials and methods

Yeast genetics, strains and general methodologies

Strains used in this study are listed in Supplementary Table S1. Standard methods for yeast genetics and molecular biology were used [25], [26], [27].

Fluorescence microscopy, time-lapse live cell imaging and quantification

Fluorescence microscope images were obtained using the DeltaVision microscope system (Applied Precision, Inc., Seattle, WA) with a cooled CCD camera CoolSNAP.HQ (Photometrics, Tucson, AZ, USA), which was equipped with a temperature-controlled chamber (Precision Control LLC, WA, USA). Live cells were imaged in a glass-bottomed culture dish (MatTek Corporation, Ashland, MA, USA) coated with soybean lectin at 27 °C or at 36 °C for ts mutants. The ts mutant cells were cultured in rich YE5S media until mid-log phase at 27 °C, and subsequently shifted to the restrictive temperature of 36 °C for further culture until observation. Images shown in Figs. 1B and 2A, B, D, E were obtained after fixing cells with methanol at −80 °C for 20 min and suspended in PBS.

Fig. 1

The 50 amino acid internal region within the TACC domain is essential for Alp7 targeting to the mitotic SPB. (A) Schematic representation of Alp7/TACC truncation mutants. Previous work showed that the C-terminal 45 amino acid residues (430–474) are required for binding to Alp14/TOG (Alp14-BD) [28], [29]. The N-terminal NLS is marked with a box (magenta). (B) Localisation of individual Alp7 mutants during mitosis. Representative images of mitotic cells in various truncation mutants are shown. Positions of the SPBs are marked with arrowheads and the cell peripheries and nuclear regions are marked with dotted outlines. Additional punctate signals of Alp-GFP located between the two SPBs correspond to kinetochores [15]. Note that only the Alp7-Δ319–368 cell failed to localise to the SPBs, although this truncation mutant protein accumulated in the nucleus. MT/Knt, Microtubule/Kinetochore. (C) The TACC region encompassing from 219 to 360, but not from 219 to 330, is enough for SPB targeting. Strains containing T-antigen-derived NLS and Alp7 (219–360)-GFP or Alp7 (219–330)-GFP, were introduced into wild type cells carrying an SPB marker Sfi1-mRFP [47], [48]. In each sample, more than 200 mitotic cells were observed. Scale bars, 5 μm.

Fig. 2

The internal 5 amino acids are specifically required for Alp7 targeting to the mitotic SPB. (A) Images showing localisation of three Alp7 mutants are presented. The peripheries of mitotic cells are marked with dotted outlines; none of the Alp7 mutant proteins localise to SPBs. Sequence comparison of various TACC proteins and substitutions of five amino acids to alanines (boxed) are shown at the bottom (hTACC3, human TACC3; Xl, Xenopus laevis, Xl-TACC3 = Maskin [10]; D, Drosophila melanogaster; Ce. C. elegans). (B) Alp7SPB colocalises with Alp14. Note that in wild type cells Alp14 colocalises with Alp7 to the SPBs, whilst it colocalises with Alp7SPB in the nucleus. (C) Levels of Alp7SPB are the same as those of wild type Alp7. Protein extracts (40 μg, two lanes on the left or 50 μg, two lanes on the right) were prepared from each GFP-tagged strain and immunoblotting was performed with anti-GFP or anti-Cdc2 antibodies. (D and E) Mitotic localisation of Alp4 and Pcp1. Representative images of wild type or alp7 cells containing Alp4-tdTomato (C) or Pcp1–2mRFP (D) are shown. Nuclear regions are presented. The positions of mitotic SPBs are marked with arrowheads. Scale bars, 5 μm. (F) Quantification of Pcp1 levels at the SPB. The box-and-whisker plot indicates 5–95 percentiles, the 25th and 75th percentiles, and the median. The sample numbers are also shown (n).

For quantification of SPB-localising Alp7, Pcp1, Sid4, Sfi1 or spindle microtubule levels, 12–14 sections were taken along the Z-axis at 0.3 μm intervals. After deconvolution, projection images of maximum intensity were obtained and maximum fluorescence intensities over the background intensity used for statistic data analysis. Statistical significance was determined by the student’s t-test. ∗P < 0.05, ∗∗∗P < 0.001, n.s. not significant.

Further information of Materials and Methods is available in on-line Supplementary Data.

Results

The C-terminal 30 amino acid residues within the TACC domain of Alp7 is required for SPB localisation during mitosis

Our previous work showed that the C-terminal region of Alp7 (219–474) composed of the TACC domain is sufficient for localisation to the SPB during mitosis [15], [28]. As a first step to determine the SPB-targeting region, a series of internal truncation mutants within this TACC domain were created, and the localisation of individual constructs was examined (GFP was added at the C-terminus, by which each construct was expressed under the endogenous promoter) (Fig. 1A). This systematic analysis uncovered that 50 amino acid residues spanning from 319 to 368 is essential for SPB localisation, though this region may not be sufficient (Fig. 1B).

We next addressed the minimal C-terminal region of Alp7 that is enough for SPB localisation. It is known that proper mitotic functioning of Alp7 requires a nuclear localisation signal (NLS) situated in the N-terminal region outside of the TACC domain (117PEFKHRKRNV126) [28], [29], [30] (Fig. 1A). To compensate for the loss of the authentic NLS, we added a canonical NLS sequence (PKKKRKV) [28], [31] to the C terminus of these constructs (Fig. 1C). It was found that peptides encompassing 219–360 (219–360-NLS-GFP) were capable of localising to the mitotic SPB (Fig. 1C, bottom). In sharp contrast, shorter peptides lacking the C-terminal 30 amino acid residues (219–330-NLS-GFP) failed to accumulate at the SPB, although this construct localised to the nucleus (top). Taken together, we concluded that 140 amino acid resides (219–360) are required and sufficient for SPB localisation of Alp7 during mitosis and furthermore, 30 amino acid residues (331–360) are indispensable for SPB targeting.

Amino acid substitutions within the SPB-targeting region abrogate Alp7 recruitment to the SPB

Further deletion analysis within the 331–360 region showed that Alp7 that lacked the first 20 amino acids did not localise to the mitotic SPB (Alp7-Δ331–350-GFP, Fig. 2A). The TACC domain consists of coiled coil motifs [10], [11] in which hydrophobic and/or charged residues are critical for α-helixes-mediated protein–protein interactions [32]. To interfere with potential protein–protein interactions, we substituted alanines for three basic (K341, R346 and K347) and two hydrophobic amino acid residues (Y344 and Y348) within the 331–350 region (K341A, Y344A, R346A, K347A and Y348A, designated Alp7SPB). Then we introduced the alp7 gene into the genomic locus under the endogenous promoter. Similar to the aforementioned deletion construct, the Alp7SPB mutant protein was also incapable of localising to the mitotic SPB (Fig. 2A). Defective SPB localisation is not simply ascribable to non-specific protein unfolding, as Alp7SPB retained its binding ability to its partner Alp14/TOG [15], as Alp14 also colocalised in the nucleus (Fig. 2B); note that Alp14 on its own cannot accumulate in the mitotic nucleus in the absence of Alp7, as this protein lacks an intrinsic NLS [15], [28], [29], [30]. Alp7SPB protein levels were indistinguishable from those of wild type (Fig. 2C). In addition, the failure of Alp7 to localise to the mitotic SPB did not impact on the localisation of other SPB proteins including a component of the γ-tubulin complex (γ-TuC), Alp4/GCP2 [33] and pericentrin-related Pcp1 [34], [35] (Fig. 2D–F). These data confirmed that the internal 20 amino acid residues (331–350), in particular five amino acids (K341, Y344, R346, K347 and Y348), are specifically required for mitotic targeting of Alp7 to the SPB.

Targeting of Alp7 to the mitotic SPB is critical for bipolar spindle assembly and proper mitotic progression

Having identified five amino acids essential for mitotic SPB targeting, we next addressed whether these residues are required for Alp7 functions. Alp7 is needed to organise bipolar spindle assembly and its deletion results in delayed and compromised spindle formation and activation of the spindle assembly checkpoint (SAC) [15], [17], [19], [29], [36]. Time-lapse live imaging analysis of mitotic alp7 cells exhibited the extended duration during the initial stage of bipolar spindle assembly, called phase I [37], which corresponds to the period from prophase to prometaphase. In wild type cells, this duration was less than 6 min (at 27 °C, Fig. 3A, left panels and Fig. 3B, top), whilst alp7 cells showed a substantially longer duration with some cells spending more than 30 min in the phase I period (Fig. 3A, middle panels and Fig. 3B, middle). Importantly, during delayed phase I, alp7 cells displayed monopolar-like, V-shaped spindle morphologies, which are reminiscent of alp7-deleted (Δalp7) cells [15] (Fig. 3A. right panels). In addition, in alp7 cells, the duration of phase II, corresponding to prometaphase to anaphase A [37], was also extended (5–7 min in wild type vs >10 min in alp7 cells, Fig. 3B). Consistent with mitotic delay and similar to the deletion strain, the alp7 mutant strain showed synthetic lethal genetic interactions with deletions of Dis1 (another member of the ch-TOG/XMAP215 family in fission yeast) [15], [16], [38] and the SAC component Mad2 [36] (Fig. 3C). We therefore, concluded that the targeting of Alp7 to the mitotic SPB is essential for proper spindle assembly and mitotic progression.

Fig. 3

alp7 cells are defective in proper spindle assembly and delayed in mitotic progression. (A) Time-lapse live imaging of alp7 cells. Wild type (left), alp7 (middle) or alp7-deleted (Δalp7. right) cells containing Sid4-mRFP and mCherry-Atb2 (an SPB marker and α2-tubulin, respectively [40], [49]), were grown at 27 °C and mitotic cells were recorded under a fluorescence microscope. Note that alp7 and Δalp7 cells exhibited extended mitotic duration and the appearance of monopolar-like V-shaped spindles (2–20 min). (B) Delayed mitotic progression of alp7 cells. Mitotic progression of wild type (top, n = 12), alp7 (middle, n = 39) or Δalp7 (bottom, n = 29) strains used in (A) was plotted. Mitosis in fission yeast is composed of the following three distinct phases, phase I exhibits the duration of initial spindle elongation coincident with the process of SPB separation, Phase II of constant spindle length and Phase III of anaphase B when spindle microtubules elongate polewards [37]. Vertical lines are placed to clarify the timing and duration of mitotic phase transition. These positions were assigned based upon the average values of the duration corresponding to each phase. In alp7 or Δalp7 mutants (shown in dotted lines), some of mutant cells showed extremely extended phase I and II, which made the precise assignment of phase transition incomprehensible. (C) Genetic interaction. Tetrad dissection was performed between an alp7 strain (alp7, G418-resistant) and dis1-deleted (dis1::ura4) or mad2-deleted strains (mad2::ura4). An example of tetratype is shown at the top; a double mutant of alp7 is inviable. A summary of genetic interaction is shown at the bottom.

Pericentrin-like Pcp1 is involved in Alp7 localisation to the mitotic SPB but is not the sole factor required for its recruitment

Our previous analysis showed that the conserved pericentrin-like protein Pcp1 is a loading factor for the γ-TuC and polo-like kinase (Plo1) to the SPB specifically during mitosis [35]. It is proposed that in Drosophila melanogaster, pericentrin-like Centrosomin is necessary for proper recruitment of D-TACC (the fly homologue of TACC) and the γ-TuC [39]. Given these preceding results, we examined whether Pcp1 is required for the mitotic localisation of Alp7/TACC to the SPB. We previously isolated a number of temperature-sensitive (ts) pcp1 mutants [35]. Among these ts mutants, pcp1–14 mutant cells (containing two point mutations, K848R and L1187P, Fig. 4A and Supplementary Fig. S1) showed reduced protein levels even under the permissive condition (27 °C) (reduced by >60%, Fig. 4B) [35], [40]. Consistently, pcp1–14 mutant cells were hypersensitive to the anti-microtubule drug, thiabendazole (TBZ) with lower viability even at 27 °C (Fig. 4A). Under the restrictive condition (4 h at 36 °C), Pcp1–14 protein levels became even lower, to ∼20% (Fig. 4B). Using this depletion-type pcp1 mutant allele, we observed the Alp7 localisation patterns. Intriguingly, careful quantification of Alp7-GFP intensities located at the mitotic SPBs showed that Alp7 signals were significantly reduced [35], [40] (P < 0.05 at 27 °C, Fig. 4C and D); however, clearly a substantial amount of Alp7 was still retained at the SPB under both permissive and restrictive conditions. The reduction of Alp7 at the SPB is not attributable to the dominant-negative effect of this pcp1–14 mutant, as there is no wild type Pcp1 protein is present in this strain.

Fig. 4

Pericentrin-like Pcp1 is involved in Alp7 recruitment (A) Serial dilution spot assay of pcp1 ts mutants. Mutants were spotted (10-fold serial dilutions) onto rich YE5S media plates (far left) or rich media containing indicated concentrations of the anti-microtubule drug, thiabendazole (TBZ). (B) Pcp1 protein levels are substantially reduced in the pcp1–14 mutant, but not in pcp1–15 or pcp1–18. Protein extracts were prepared from indicated strains grown at 27 °C (0 h) or 36 °C (4 h) and immunoblotting was performed with affinity-purified anti-Pcp1 (top) or anti-α-tubulin antibody (bottom). Relative values of Pcp1 intensities that were calculated using α-tubulin as loading control are also shown. (C and D) Localisation patterns of Alp7 and quantification of its intensities in the pcp1–14 mutant. Alp7-GFP signals (also Sid4-mRFP and mCherry-Atb2) were observed in cells grown at 27 or 36 °C (C, only images of cells grown at 27 °C are shown) and relative GFP intensities at the mitotic SPB were quantified (D). The number of cells counted in each sample is also shown (42, 46, 20 and 38). (E, F) Localisation patterns of Alp7 and quantification of its intensities in the pcp1–15 and pcp1–18 mutants. Alp7–3GFP signals (also Sfi1-mRFP and mCherry-Atb2) were observed and quantified as in (C) and (D). Statistical significance was determined by the student’s t-test; ∗P < 0.05, ∗∗∗P < 0.001 and n.s., not significant (P > 0.05). Note that Alp7–3GFP intensities decreased after 4 h incubation at 36 °C even in wild type cells, which is ascribable to fluorescence bleaching. The number of cells counted in each sample is also shown (26, 26, 20, 28, 32 and 28). Scale bars, 5 μm. (G) Alp7 and Pcp1 physically interact. Immunoprecipitation was performed using a temperature sensitive cut9–665 strain containing Alp7-myc. Cut9 is a subunit of the Anaphase Promoting Complex/Cyclosome [50] and cut9–665 cells arrest in mid mitosis at the restrictive temperature [51]. Protein extracts were prepared from cultures incubated at either 27 °C (asynchronous, lanes 1–5) or 36 °C (for 2.5 h, mitotic, lanes 6–10). Reciprocal immunoprecipitation was performed using anti-myc (lanes 4 and 9) or anti-Pcp1 antibodies (lanes 5 and 10). WCE (whole cell extract, 5 or 10 μg, lanes 1, 2, 6 and 7) and precipitates in the absence of primary antibodies (bead, lanes 3 and 8) were also loaded. Asterisks denote degradation products of Alp7-myc and Pcp1 [22]. (H) Quantification of coimmunoprecipitation. Band intensities of IP:bead shown in (G, lanes 3 and 8) were used as background and subtracted from those of IP:myc (lanes 4 and 9) or IP:pcp1 (lanes 5 and 10) band intensities. Band intensities in asynchronous cultures were set as 1, and those of mitotic samples were calculated accordingly.

We also examined the mitotic localisation of Alp7 in two other pcp1 ts alleles, pcp1–15 and pcp1–18. Our previous work showed that pcp1–15 cells are specifically defective in the recruitment of the γ-TuC, whilst pcp1–18 mutants impair the recruitment of the polo kinase Plo1, but not the γ-TuC [35]. Unlike Pcp1–14, protein levels were not decreased in either Pcp1–15 or Pcp1–18 at the permissive temperature (Fig. 4B). Interestingly, however, signal intensities of Alp7 at the mitotic SPB were reduced by <50% in the pcp1–18 mutant but not in pcp1–15 at the permissive temperature (Fig. 4E and F). This result substantiates the notion that Pcp1 is, at least partly, responsible for the mitotic recruitment of Alp7 to the SPB and may imply that Plo1 is involved in this process.

Finally physical interaction between Alp7 and Pcp1 was addressed with coimmunoprecipitation. We found that these two proteins weakly interact under the asynchronous growing condition (Fig. 4G, lanes 4 and 5) and importantly, the extent of interaction was increased during mitosis (lanes 9 and 10 and Fig. 4H). Although the possibility of the existence of the third intermediate molecule is not excluded, Pcp1 forms a complex with Alp7. Collectively, we envisage that Pcp1 plays an important role in the recruitment of Alp7 to the SPB as a structural platform, albeit being not a sole molecule.

Discussion

In this work, we have identified the internal region within the TACC domain that is essential for Alp7 loading to the mitotic SPB. Centrosomal TACC proteins interact with a cohort of proteins, where they play a critical role in proper bipolar spindle assembly [10], [11]. Despite its importance however, the detailed structural and functional dissection of the TACC proteins remains to be explored, except for their binding with Alp14/ch-TOG/XMAP215 and Clathrin molecules [10], [14], [15], [41]. Accordingly, it is currently unknown as to how TACC proteins are recruited to the mitotic centrosome. Nor has the physiological significance of the TACC region corresponding to Alp7/TACC (331–350) been examined in other TACC members. Therefore, it would be of great interest to address this issue in higher eukaryotes, particularly potential interactions between the TACC members and the pericentrin-related proteins.

Previous work shows that Alp7 interacts with not only Alp14/TOG or the SPB, but also the outer kinetochore component Ndc80, thereby ensuring proper kinetochore-microtubule attachment [22]. All these interactors bind the TACC domain [15], [22]. We envision that coiled coil motifs within the TACC domain represent multiple structural subdomains, each of which interacts with distinct proteins. Further analysis of the Alp7/TACC domain in order to narrow down the kinetochore binding sequence will provide deeper insight into the composite structures of the TACC domain and the multiple roles of Alp7/TACC in bipolar spindle assembly, kinetochore-microtubule attachment and genome integrity.

In human cells, either downregulation or abnormal upregulation of the TACC proteins is intimately linked to tumourigenesis, and their expression profiles are used for cancer prognosis [5], [11]. Furthermore, centrosome abnormalities associated with defective spindle assembly are one of the hallmarks of cancer [6], [7]. The molecular understanding of the mechanisms underlying centrosomal recruitment of TACC proteins in higher eukaryotes is critical for the aetiology of cancer and other human diseases [42], [43], [44]. Although Pcp1 is implicated in Alp7 loading, our data strongly suggests the existence of other platforms on the SPB. As several fission yeast SPB components other than Pcp1 are required for mitotic spindle formation [45], [46] and vertebrates contain multiple pericentrin-related members [39], it would be important to examine the requirement of these molecules for TACC loading to the mitotic centrosome/SPB. Our work is, therefore, instrumental for the further characterisation of the TACC proteins.

Competing financial interests

The authors declare no competing financial interests.