HSP70 is required for the proper assembly of pericentriolar material and function of mitotic centrosomes
At the onset of mitosis, the centrosome expands and matures, acquiring enhanced activities for microtubule nucleation and assembly of a functional bipolar mitotic spindle. However, the mechanisms that regulate centrosome expansion and maturation are largely unknown. Previously, we demonstrated in an immortalized human cell line CGL2 and cancer cell line HeLa that the inducible form of heat shock protein 70 (HSP70) accumulates at the mitotic centrosome and is required for centrosome maturation and bipolar spindle assembly.
In this study, we further show that HSP70 accumulated at the spindle pole in a PLK1-dependent manner. HSP70 colocalized with pericentrin (PCNT), CEP215 and γ-tubulin at the spindle pole and was required for the 3D assembly of these three proteins, which supports mitotic centrosome function. Loss of HSP70 disrupted mitotic centrosome structure, reduced pericentriolar material recruitment and induced fragmentation of spindle poles. In addition, HSP70 was necessary for the interaction between PCNT and CEP215 and also facilitated PLK1 accumulation and function at the spindle pole. Furthermore, we found that HSP70 chaperone activity is required for PCNT accumulation at the mitotic centrosome and assembly of mitotic spindles.
Our current results demonstrate that HSP70 is required for the accurate assembly of the pericentriolar material and proper functioning of mitotic centrosomes.
KeywordsHSP70 Mitotic centrosome Pericentriolar material Spindle pole
ground state depletion
heat shock protein 70
MT organizing center
Polo like kinase 1
structured illumination microscopy
The centrosome is a structurally complex and functionally diverse organelle that regulates various cellular processes. It consists of a pair of barrel-shaped structures called centrioles and a surrounding protein complex, which is referred to as the pericentriolar material (PCM). The PCM contains hundreds of proteins, including signaling molecules, cell cycle regulators and crucial microtubule (MT)-nucleating factors. The combined function of these proteins makes the centrosome the major MT-organizing center (MTOC), which coordinates all MT-related functions, including cell shape, cell polarity, mobility, intracellular trafficking and cell division . The centrosome is duplicated during S phase along with DNA replication, yielding two centrosomes that nucleate extensive MT arrays and form the two poles of the mitotic spindle . Importantly, from late S phase to mitotic onset, the two fully duplicated centrosomes exhibit enhanced recruitment of PCM components that are essential for MT nucleation, a process termed centrosome maturation. In this process, PCM components accumulate at the existing centrosome and expand into a greatly enlarged and intermingled matrix structure to enhance MT nucleation capacity and promote the assembly of a functional bipolar spindle [3, 4, 5]. Inaccurate accumulation of PCM components at the centrosome results in a functionally compromised mitotic centrosome with multipolar or disorganized spindles that may further promote mitotic arrest, cell death and/or aneuploidy [6, 7].
Previous studies using subdiffraction resolution microscopy and fluorescence recovery after photobleaching have revealed that PCM components adopt either a concentric toroidal or an extending fiber-like organization to form a thin layer of PCM at the interphase centrosome; these structures then undergo extensive rearrangement, continuous replenishment, local trafficking, massive accumulation and expansion, and occupy distinct subdomains within the dramatically enlarged mitotic PCM matrix [3, 4, 8, 9, 10, 11, 12, 13]. While the mechanisms that precisely regulate the behaviors of PCM components are not clearly understood, several major contributors have been identified. For example, Polo like kinase 1 (PLK1) is a master mitotic kinase that is known to phosphorylate multiple PCM substrates and drive several steps of centrosome maturation [14, 15, 16]. Additionally, pericentrin (PCNT) and CEP215 and their respective homologues in various model organism are putative scaffold proteins that are thought to physically interact at the maturing centrosome, providing a proper platform to recruit other PCMs [15, 17, 18, 19]. In the PCM matrix, PCNT phosphorylation by PLK1 was shown to increase recruitment of MT nucleating factors, such as NEDD1 and γ-tubulin, whose accumulation is also PLK1-dependent [15, 17]. Notably, PLK1 activity persists throughout mitosis to ensure continuous replenishment of PCM components, including PLK1 itself, suggesting a highly dynamic and interdependent orchestration of PCM components at the functional mitotic centrosome [20, 21]. Despite these major advances in understanding centrosome maturation, the highly organized regulatory mechanisms governing the concentration and exchange of these and other proteins at the mitotic centrosome are still largely unknown. Improper accumulation of γ-tubulin has been implicated in abnormal spindle formation and human malignancies [22, 23]. Moreover, blocking PLK1 activity in the mitotic cells with already matured centrosomes reduces PCNT and γ-tubulin accumulation , while mutant PLK1 with an altered exchange rate at the mitotic centrosome induces mitotic arrest . Mutant PCNT that cannot be phosphorylated by PLK1 fails to recruit multiple PCM components, including PLK1 itself , while mutations that abolish interaction between PCNT and CEP215 disrupt the recruitment of themselves and γ-tubulin, inducing the formation of defective mitotic centrosomes and spindles . Together, these data imply that delicate proteostatic control of PCM components is important for determining the proper function of a fully matured mitotic centrosome.
Molecular chaperones, such as heat shock proteins (HSPs), are major orchestrators of the cellular proteostatic mechanisms that control protein quality through folding, oligomerization, trafficking, disaggregation, and proteasomal or autophagic degradation. Among the wide variety of HSPs, HSP70 family members form a central hub of the chaperone networks that cooperatively utilize various chaperones and co-chaperones to regulate all aspects of cellular proteostasis, especially protein quality control in certain organelles [24, 25]. HSP70 members in S. cerevisiae were proposed to control the oligomeric state of Cin8 motor protein to regulate spindle length . In mammalian cells, HSP70 mainly exists in constitutive form (HSC70) and inducible form. The inducible HSP70 (thereafter HSP70) is phosphorylated by NEK6, which targets HSP70 to the mitotic spindle where it maintains kinetochore-MT stability and supports centrosome clustering [27, 28]. Additionally, HSP70 was also shown to protect centrosome and spindle integrity when overexpressed, likely by preventing proteasomal degradation of centrosomal proteins in the heat-shocked cells [29, 30]. Thus, the importance of HSP70 in the mitotic centrosome is well established, and some evidence suggests that it may exert chaperone activities on centrosome components to regulate centrosome maturation.
We have previously demonstrated that HSP70 (encoded by genes HSPA1A and HSPA1B) localizes to the mitotic spindle poles and is required for maintenance of centrosome integrity, MT nucleation, mitotic spindle assembly, mitotic progression, and cell viability . In this study, we further employed subdiffraction resolution microscopy to examine the roles of HSP70 in the structure and function of the mitotic centrosome. We found that HSP70 cooperates with PLK1, PCNT, and CEP215 to support their accumulation and proper 3D assembly, and this action appears to be required for complete maturation of a fully functional mitotic centrosome.
HSP70 accumulates at the spindle pole in a PLK1-dependent manner and colocalizes with PCNT, CEP215 and γ-tubulin
Loss of HSP70 induces spindle pole fragmentation and disrupts the accumulation of PCNT and CEP215 at the mitotic centrosome
HSP70 is required for the 3D assembly of PCNT, CEP215 and γ-tubulin at the mitotic centrosome
HSP70 is required for the interaction between PCNT and CEP215 at the mitotic centrosome
HSP70 facilitates PLK1 accumulation at the spindle pole and ameliorates PLK1 interference-induced spindle abnormalities
HSPA1A chaperone activity is required for PCNT accumulation at the mitotic centrosome and assembly of mitotic spindles
The accumulation of the PCM components at the centrosome during maturation is a delicately regulated process that is essential for the ability of the mitotic centrosome to enhance MT nucleation and assemble a well-functioning bipolar spindle. In this study, we reveal that HSP70 associates with critical PCM components, including PCNT, CEP215 and γ-tubulin, to ensure appropriate 3D assembly of the mitotic centrosome and permit subsequent MT nucleation and bipolar spindle assembly.
The importance of HSP70 in the function and integrity of mitotic centrosomes and in mitosis progression has long been recognized. For example, Sse1, a member of the yeast HSP70 family, was reported to modulate the oligomeric state of the MT motor Cin8 to regulate the length of mitotic spindle . Furthermore, NEK6 phosphorylates HSP70 and targets it to the mitotic spindle to facilitate chTOG–TACC3 complex recruitment, thus promoting kinetochore microtubule stability . Overexpressed HSP70 was also shown to repair the heat shock-induced defects in the mitotic centrosome and facilitate subsequent progression through mitosis . Additionally, HSP70 was shown to accumulate at the centrosomes of the heat shocked cells and prevent the proteasomal degradation of centrosome proteins . These studies all suggested the possibility that HSP70 may regulate centrosome functions to support the mitotic progression, and we previously confirmed that HSP70 localizes to the mitotic spindle pole and is required for centrosome integrity, MT nucleation and spindle assembly . However, the question of how HSP70 regulates the functions of the mitotic centrosome remained unanswered. In this study, we utilized the enhanced resolution of 3D-SIM and GSD microscopy to demonstrate that HSP70 colocalizes with PCM components, including PCNT, CEP215 and γ-tubulin. Loss of HSP70 significantly reduces the recruitment of these PCM components and disrupts their 3D assembly at the mitotic centrosome, leading to decreased MT nucleation. Since the proper functioning of the mitotic centrosome requires PCM components to be recruited in a highly organized manner, our results reveal that HSP70 permits MT nucleation and bipolar spindle assembly by promoting the recruitment and accurate 3D assembly of PCM components at the mitotic centrosome.
It has been shown that the Hspa1a−/−/Hspa1b−/− double knock out (KO) mice can grow to adulthood with only mild cardiac hypertrophy , indicating that HSP70 may be not essential for cell cycle progression and division during normal mouse embryo development. However, the early-stage embryonic cells exhibit a very rapid cell cycle progression and their divisions involve a gradual transition from the acentrosomal spindle assembly to the centrosomal spindle assembly during embryo development [54, 55, 56]. These findings indicate that the embryonic cells might control the cell cycle progression, centrosome organization, and spindle assembly during development in ways different from those in differentiated somatic cells and thus might not depend on HSP70. Alternatively, it has been shown that the embryonic mouse hearts exhibit a robust regeneration after ablation of up to 60% of cardiac progenitor cells at embryonic day 7.5, which permits embryo development and survival . Thus, the injuries and cell loss caused by HSP70 KO might induce a compensatory proliferation of fetal cardiomyocytes and the animal could eventually tolerate and adapt to HSP70 deficiency. Another study showed that the Hspa1a−/−/Hspa1b−/− mice were on average 12% lighter in weight than wild-type newborns and had elevated levels of spontaneous genomic instability . The lighter weight of the Hspa1a−/−/Hspa1b−/− mice could result from reduced cell growth or cell loss as the authors demonstrated that the embryonic fibroblasts (MEF) prepared from these mice exhibited decreased cell growth and enhanced genomic instability. Our previous study  has demonstrated that inhibition of HSP70 induced mitotic arrest and reduced cell viability. In the current study, the malfunctioning mitotic centrosomes or the multipolar spindles induced in HSP70-depleted cells might initiate cell cycle arrest and/or cell death, which may also result in reduced cell growth or cell loss. Additionally, the MEFs from the mutant mice exhibited enhanced genomic instability, elevated micronuclei formation and chromosome damages, all of which could result from mitotic errors [59, 60]. Thus, the phenotypes exhibited in Hspa1a−/−/a1b−/− mice could be partially explained by the effects of HSP70 depletion-induced defects on mitotic centrosomes and mitosis.
Recently, the NEK6-HSP70 axis was reported to facilitate centrosome clustering in cancer cells with amplified centrosomes, promoting mitotic progression through the formation of a pseudobipolar spindle . Moreover, cancer cells with supernumerary centrosomes, which enter mitosis with multiple MTOCs, fail to resolve the multipolar spindle into a pseudobipolar spindle under HSP70 inhibition. Together with the previously reported function of NEK6-HSP70 in the recruitment of the chTOG–TACC3 complex, these data point to a role for HSP70 in regulating MT dynamics to facilitate centrosome clustering and bipolar division. In our study using an immortalized cell line, loss of HSP70 impaired the 3D structure of mitotic centrosomes and reduced PCM accumulation, an effect that was accompanied by the formation of extra spindle poles comprising disorganized PCM components. Thus, HSP70 depletion may specifically disrupt the stability of protein-rich PCM networks, and the PCM components that fail to be properly organized at the mitotic centrosome may contribute to new MT nucleating centers, leading to multipolar spindle formation. Interestingly, in the live imaging of mitotic progression in HSP70-depleted cells, we observed spindle pole fragmentation, which supports this hypothesis. Thus, in addition to its roles in regulating MT dynamics and centrosome clustering, our data provide another possible function of HSP70 in maintaining centrosome integrity and spindle bipolarity. As such, HSP70 appears to directly regulate the organization of mitotic PCM components, thus supporting their assembly in a bipolar spindle.
PCNT and CEP215 are putatively considered to be scaffold proteins based to their large sizes and predicted coil-coiled domains. Moreover, these proteins were proposed to physically interact with each other and polymerize around the maturing centrosome to form a matrix structure [17, 61, 62]. In drosophila, polo kinase, the PLK1 homolog, phosphorylates centrosomin, the CEP215 homolog, driving its structural rearrangement and assembly into a matrix structure; whereas in mammalian cells, the phosphorylation of PCNT and presumably CEP215 by PLK1 is prerequisite for the further recruitment of PCM components, promoting centrosome maturation [12, 15, 17, 19]. Based on these reports, it is currently believed that the recruitment of PLK1, PCNT and CEP215 to the centrosome is an interdependent process that occurs at the early phase of centrosome maturation [3, 4, 5]. Importantly, recruitment of PCM components is highly dynamic. PLK1 activity persists throughout mitosis and drives the continuous exchange of proteins, which are thought to be first incorporated in close vicinity to the centriole and then gradually extended outward into the PCM matrix [8, 12, 13, 20, 21, 63]. HSP70 has been proposed to be a mitotic substrate of PLK1 , which together with its observed localization at the spindle pole, implied a potential role for HSP70 in regulating centrosome maturation. Our results show that PLK1 may positively regulate HSP70 spindle pole accumulation, consisting with a previous report . Intriguingly, we also observed that loss of HSP70 reduced PLK1 accumulation at the spindle pole and exacerbated mitotic defects induced by PLK1 inhibition, in addition to diminishing the PCNT-CEP215 interaction and accumulation at the spindle pole of the mitotic cells that had not yet formed multipolar spindles. These results suggest that HSP70 is recruited to the spindle pole by PLK1 and is also required for PLK1 recruitment and function at the spindle pole. We therefore hypothesize that HSP70 is a critical constituent of the maturing centrosome that cooperates with PLK1, PCNT and CEP215 to establish a functional mitotic centrosome in CGL2 cells.
The effects of the mutant HSP70s in spindle assembly suggest that HSP70 may regulate the accumulation of PLK1, PCNT and CEP215 through its chaperone activities, including prevention of substrate protein aggregation by direct binding (holding), ATP-fueled unfolding/refolding of polypeptides to regulate (alter)natively folded or oligomeric states, and the timely targeting of the substrate proteins for proteasomal or autophagic degradation [25, 32, 49, 64, 65]. In our study, HSP70-GG and -GY, two HSP70 mutants with disrupted SBD, failed to rescue the spindle defects induced by depletion of endogenous HSP70, implying the importance of the substrate binding in this process. Additionally, HSP70-K71E, which has disrupted ATPase activity but contains intact ability to bind substrates, showed a partial rescue effect, while HSP70-E175S, which is locked in an ADP-bound conformation and therefore lacks both ATPase and refolding activities, exhibited no rescue effect. These results imply that in addition to the substrate binding/holding activity of HSP70, ATP-ADP exchange is also required, and therefore, the repetitive refolding of substrates by HSP70 is likely to be required for a fully functioning mitotic centrosome. We hypothesize that during centrosome maturation, when abundant PCM components accumulate and form a highly protein-rich environment, HSP70 holding activities may provide a ‘buffering effect’ by maintaining certain PCM proteins in a properly folded state, preventing them from misfolding and aggregation. Concomitantly, HSP70 actions also require ATP-ADP cycling, which fuels repetitive processing of substrates to support the local transportation, oligomerization or interactions of the PCM proteins. This regulation of PCM components further supports the maturation and function of the mitotic centrosome. PCNT and CEP215, both of which play critical roles in centrosome maturation, tend to form polymerized complexes, and their homologs in drosophila were proposed to adopt specific conformations around the centrosome [11, 19]. Therefore, HSP70 could potentially maintain the proper conformation of these complexes via its client holding activity. In addition, the subcellular localization and kinase activities of PLK1 depend on the proper conformation of its polo-box domain [66, 67], and this conformation might also be maintained by HSP70. In addition to the importance of the SBD and ATPase domain, HSP70-CTD, in which the C-terminal CHIP/HOP-interacting IEEVD motif is deleted, also exhibited no rescue effects. Since we previously reported that this mutant hardly interacts with PCM components , this observation suggests a requirement for HSP70 interactions with the co-chaperone such as CHIP, which is particularly important in the chaperone-mediated proteasomal degradation, or HOP, which mediates the cooperation between HSP70 and HSP90 to achieve substrate folding [32, 65, 68, 69]. Since numerous PCM components undergo rapid turnover at the mitotic centrosome [12, 20, 63] and proteasomal activities were reported to control the centrosome protein level [30, 70, 71], and HSP90 has been reported to be a core member of Drosophila centrosome that regulates polo kinase stability and the function of mitotic centrosome [72, 73], we propose that HSP70-CHIP-mediated proteosomal degradation or HSP70-HOP-HSP90-mediated substrate folding may control the turnover of certain PCM components to facilitate mitotic PCM assembly. Notably, the alternate binding of HSP70 IEEVD motif to CHIP or HOP, and therefore the alternate substrate processing, is decided by phosphorylation near the C-terminal region of HSP70 , which is very close to the reported PLK1 phosphorylation site . It would be interesting to investigate whether PLK1 phosphorylation of HSP70 modulate its interaction with the co-chaperones to achieve alternate substrate processing in the mitotic centrosome. Future identification of the HSP70 substrate proteins in the mitotic centrosome will be important to test this hypothesis.
In conclusion, we have extended our previous finding that HSP70 is required for the functions of the mitotic centrosome in MT nucleation, bipolar spindle assembly and mitotic progression by showing that HSP70 acts through it chaperone activity to facilitate proper 3D assembly of PCM components at the mitotic centrosome. Since HSP70 activities and centrosome aberrations are well correlated with pathologies of cancer, age-related diseases, degeneration, and developmental defects, our findings may have broad-ranging applications in these conditions and provide information for future development of therapeutic strategies based on HSP70 or centrosome functions.
Cell culture and drug treatments
CGL2 and HeLa-S3 cells were obtained and cultured as previously described . HeLa-S3 cells stably expressing EYFP-fused α-tubulin (HeLa-S3-tub-EYFP) were established by transfecting cells with pEYFP-tubulin vector (Clontech, Mountain View, CA, USA) and selecting stable transfectants in medium containing 1 mg/ml G418. Cells were routinely maintained in Dulbecco’s Modified Eagle’s Medium (Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (Invitrogen), 0.37% sodium bicarbonate, 100 U/ml of penicillin, and 100 μg/ml of streptomycin at 37 °C in an humidified incubator with 10% CO2; cells were passaged twice a week. PLK inhibitor III (Merck, Darmstadt, Germany), BI2536 (Selleckchem, Houston, TX, USA) and Pifithrin-μ (PES, Tocris Bioscience, Bistro, UK), were dissolved in dimethyl sulfoxide and stored at − 20 °C in aliquots. Treatment concentrations are given in Figure Legends, and unless otherwise indicated, all drugs were treated 30 min prior to cell fixation in order to examine immediate effects on the mitotic cells.
Depletion and overexpression of HSP70
The shRNAs targeting HSPA1A (TRCN8757 and TRCN342860) and HSPA1B (TRCN8760 and TRCN 8763) were obtained from the National RNAi Core Facility Platform (Genomic Research Center, Academia Sinica). HSP70 in CGL2 or HeLa-S3-tub-EYFP cells were depleted as described . The pFB-Neo vector containing FLAG-tagged wild-type (WT) HSPA1A was prepared as previously described . Mutations of Lys71 to Glu, Glu175 to Ser, Ala403 to Gly, Val438 to Gly, and Val438 to Tyr were generated by site-directed mutagenesis. WT or mutant HSPA1A were ectopically overexpressed in cells as previously described . For rescue experiments, cells stably overexpressing WT or mutant FLAG-HSPA1As were established under 1 mg/ml G418 selection and then transduced with shRNA targeting the 5′-untranslated region of HSPA1A (TRCN342860) to transiently deplete endogenous HSP70. Cells were also transduced with empty-vector pLKO.1 and pFB-Neo to provide controls for depletion and overexpression of HSP70, respectively.
Cells were seeded onto coverslips, incubated for 20 h before drug treatment or viral transduction, and then fixed in PTEMF buffer containing 20 mM PIPES, 4% paraformaldehyde, 0.2% Triton-X, 10 mM EGTA and 1 mM MgCl2 for 15 min. Primary antibodies included anti-HSP70 (GeneTex, Hsinchu, Taiwan or StressMarq, Victoria, Canada), anti-α-tubulin (GTX112141, GeneTex or T5168, Sigma, St. Louis, MO, USA), anti-γ-tubulin (T6557 or T3559, Sigma), anti-EB1 (E3406, Sigma), anti-pericentrin (ab4448 or ab28144, Abcam, Cambridge, UK), anti-CEP215 (IHC-00063, Bethyl, Cambridge, UK), anti-CEP164 (HPA037606, Sigma), anti-CEP152 (ab183911, Abcam), anti-PLK1 (No. 33-1700, Invitrogen), and anti-phospho-T210-PLK1 (PA1-126, ThermoFisher, Waltham, MA, USA). Alexa-Fluor 488-, 568-, 633-, or 647-conjugated goat anti-mouse or anti-rabbit IgG were purchased from Invitrogen, and Atto-488 goat anti-mouse IgG was from Sigma. Samples were mounted in Fluoromount-G from SouthernBiotech (Birmingham, AL, USA) for confocal or structured illumination microscopy (SIM) imaging. For ground state depletion (GSD) imaging, samples were mounted in DPBS (Sigma) containing 5% glucose, 100 mM 2-mercaptoethylamine (Sigma), 0.8 mg/ml glucose oxidase and 40 μg/ml catalase, and sealed with Twinsil (Picodent, #13001000, Wipperfürth, Germany).
Confocal microscopic image acquisition and analysis
Confocal images of the stained cells were obtained from a Leica TCS-SP5 microscope with a HCX PLAPO 63×/1.4 objective. Images were acquired with 300 Hz scanning speed in 15 μm image stacks with a 0.5 μm step size. For the time-lapse images of mitosis progression, HeLa-S3-tub-EYFP cells were seeded in chambered coverslips (μ-slide 4 well, ibidi GmbH, Germany) and imaged on a Leica TSC-SP5 inverted microscope with a HCX PLAPO 40×/0.85 objective, 30 μm image stack, 0.5 μm step size and 10 min imaging span under a 37 °C and 10% CO2 environment. Laser intensity and HyD gain were fixed for each independent experiment, and the 3D confocal images were processed and projected into 2D with ImageJ. Quantitative analyses of confocal images were conducted with Imaris software. To determine the accumulation of HSP70 and other centrosome markers, a “Spot” of 1 or 2 μm in diameter surrounding the centrosome/spindle poles was created, and the intensity sum within the spot was measured. To determine the size of the centrosome, a “surface” covering the fluorescence signal of a centrosome marker was created, and the signal volume was measured. The conditions for immunostaining, imaging, and spot/surface creation were consistent for each independent experiment. Values for intensity and volume were normalized to markers that were stained and imaged in each experiment.
3D-SIM image acquisition and analysis
3D-SIM imaging was performed on a Zeiss ELYRA PS.1 LSM780 system with Plan-APOCHROMAT 63×/1.4 objective. A 28 μm grating for the 488-nm channel and 34 μm grating for the 568-nm channel with five rotations and five phases per z-section were used to image a stack covering the spindle pole region with 0.11 μm intervals in z-direction. Channels were aligned on ZEN software according to the reference landmark of multispectral beads (100 nm in diameter, Life technology). Percentage of colocalization between HSP70 and centrosome markers was then measured on Imaris software. To assess the relative spatial distribution of HSP70 and other centrosome markers from the SIM images, 3D image alignment and average were utilized on the Chimera software from UCSF [11, 41]. Briefly, all centrosome markers were co-stained with γ-tubulin, which served as a reference during the alignment process. The initial “top view” of γ-tubulin images were selected as the reference, as judged by the patterns of both γ-tubulin and the co-stained markers (e.g., CEP164). Other γ-tubulin images were then aligned to the reference to preserve the spatial relationship with the co-stained marker (CEP164). Once all the images were aligned, the γ-tubulin and CEP164 images were averaged separately to generate respective model images. Iterations of this process were then performed for other centrosome marker/γ-tubulin pairs to obtain the final 6-channel model images displaying relative distribution of all the centrosome markers. Images of HSP70, which was co-stained with γ-tubulin, CEP215 and PCNT, were then added by aligning the three PCMs to the model image. These 3D model images were then maxima-projected and channels were merged on ImageJ software.
GSD image acquisition and analysis
GSD images were acquired from Leica SR GSD superresolution system with a HCX APO 100×/1.47 objective. Laser power was gradually increased until the repetitive photoswitching (blinking) of the single fluorophore molecule was observed; then the power was fixed, and the threshold was set to collect serial images of the fluorescence blinking. All single-molecule blinking events within ~ 10,000 serial images were localized and reconstructed into a high-resolution centrosome image. Laser power, threshold, and exposure time were kept constant for each independent experiment and all image acquisitions started from the 647-nm channel and proceeded to the 488-nm channel. To compare the spatial distribution and accumulation of PCNT and CEP215 at the mitotic centrosomes in control and HSP70-depleted cells, five lines intersecting at the center of PCNT GSD image were drawn and applied to the CEP215 channel for each GSD image, and the intensity along each line was measured, normalized, and averaged with ImageJ to obtain the mean distribution curve. For each line, the maximum intensity (peak) values within the 600 nm regions flanking the center were identified and the distance between two peaks was determined to verify the diameter of PCNT and CEP215 toroid. The maximum intensity (peak) values were also divided by the values at the center to obtain the peak-center intensity ratio to verify the accumulation of PCNT and CEP215. PLK1 recruitment to the spindle pole was measured with ImageJ, according to its intensity in the GSD images. A circle with a 1.2 μm diameter surrounding the spindle pole region was drawn. Normalized PLK1 intensity was then determined as the total intensity inside the circle divided by the total intensity outside the circle.
Cell lysis and immunoblotting were carried out as described . Specific proteins were detected using antibodies against HSP70 (GeneTex), GRP78 (610979, BD Biosciences, San Jose, CA, USA), HSC70 (sc-7298, Santa Cruz Biotechnology, CA), CEP215 (Bethyl), γ-tubulin (T6557, Sigma), EB1 (E3406, Sigma), GM130 (610822, BD Biosciences), FLAG (F3165, Sigma), respectively, and GAPDH (GTX100118, GeneTex) for use as loading controls.
Proximity ligation assay (PLA)
PLA was used to test whether HSP70 regulates the interaction between PCNT and CEP215. Cells with or without HSP70 depletion were fixed and subjected for PLA with Duolink red starter kits (DUO92101, Sigma) as described previously . Afterward, cells were mounted in Fluoromount-G containing DAPI and imaged with a Leica TCS-SP5 microscope. The PLA signal intensity was obtained by Imaris following the procedures described above.
The authors thank the Core Facility of the Institutes of Cellular and Organismic Biology and Institute of Molecular Biology, Academia Sinica, for their assistance with confocal microscopy, GSD, SIM and image analysis.
CTF and HHK performed the experiments. SCH performed the image analysis of SIM data. CTF and LHY analyzed and interpreted the data, and drafted the manuscript. LHY contributed to the conception and design of the work. All authors read and approved the final manuscript.
This work was supported by grants from Academia Sinica and the Ministry of Science and Technology (MOST 103-2320-B-001-023-MY3 and 106-2320-B-001-003-), Taiwan.
Ethics approval and consent to participate
Consent for publication
The authors declare no competing interests.
- 52.Fontaine SN, Rauch JN, Nordhues BA, Assimon VA, Stothert AR, Jinwal UK, et al. Isoform-selective genetic inhibition of constitutive cytosolic Hsp70 activity promotes client tau degradation using an altered co-chaperone complement. J Biol Chem. 2015;290(21):13115–27.PubMedPubMedCentralCrossRefGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.