Citation
Zovko, S. (2010, June 22). Dynamics and structural features of the microtubule plus-ends in interphase mouse fibroblasts. Retrieved from https://hdl.handle.net/1887/15711
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/15711
Note: To cite this publication please use the final published version (if
applicable).
CHAPTER V
Role of CLIP-170 and CLIP-115 in microtubule dynamics and cellular morphology of interphase mouse fibroblasts
Work in progress
98
99
Role of CLIP-170 and CLIP-115 in microtubule dynamics and cellular morphology of interphase mouse fibroblasts
LIP-170 and CLIP-115 bind to and track growing microtubule (MT) plus-ends.
They function as recue factors in mouse embryonic fibroblasts (MEFs). Here we aimed to characterize MEFs deprived of CLIP-170/CLIP-115 regarding their MT network and morphological features by the means of fluorescence microscopy. In addition, we set up to elucidate MT plus-end conformation (s) which might be targeted/stabilized by CLIPs using electron microscopy. We have found that MEFs deprived of CLIP-170/CLIP-115 (DKO) exhibit a number of notable distinctions compared to the WT MEFs. We show here that CLIPs are required for proper distribution of MTs in interphase MEFs. Our data also suggest that both dynamic and stabile MTs (i.e. MTs enriched with acetylated and/or detyrosinated α-tubulin) are less resistant to MT depolymerizing agent nocodazole in cells missing CLIP proteins, compared to the WT cells. Moreover, CLIP-deficient cells show a defect in cell spreading and perturbed cell-extracellular matrix adhesion signaling. In addition, electron microscopy studies of the MT plus-ends in WT and DKO cells indicate that sheet-like conformation of the MT plus-end is significantly less present in the periphery of the CLIPs-deficient MEFs.
Whether the measured effect is a direct consequence of CLIPs absence and thus lack of stabilization of this particular plus-end structure, or an indirect effect through the perturbation of the MT dynamics in the DKO cells in general, needs to be elucidated yet.
C
100
IntroductionMicrotubule (MT) dynamics, spatial organization of MTs within the cell, and the association of MT plus-ends with various cellular structures are, for a large part, regulated by MT plus-end tracking proteins (+TIPs). +TIPs associate preferentially with the plus-ends of growing MTs. In general +TIPs are involved in stabilization of the plus- end by either promoting the shrinkage-to-growth (i.e. rescue) switch or by inhibiting the growth-to-shrinkage switch (i.e. catastrophe).
Cytoplasmic linker protein of 170kDa (CLIP-170) belongs to +TIPs group of proteins and is characterized by two conserved cytoskeleton-associated protein glycine-rich (CAP- GLY) domains situated in the N-terminal region. The two CAP-Gly domains (Cap-Gly1 and Cap-Gly2) have been proposed to interact with α-tubulin from two tubulin dimmers (Slep and Vale, 2007). Interestingly, only the tyrosinated C-terminus of α-tubulin is able to bind CLIP-170; detyrosination of α-tubulin abolishes this interaction (Peris et al., 2006). At the C-terminal region a metal binding domain, the Zn-finger domain, is localized, necessary for targeting of other +TIPs, i.e. LIS1, the dynactin subunit, p150Glued, and CLIP-115 (Holzbaur et al., 1991; Akhmanova et al., 2001; Coquelle et al., 2002). The Zn-finger domain is also able to interact with the CAP-GLY domain of the same molecule, resulting in intra-molecular folding and leading to auto-inhibition of CLIP- 170 (Mishima et al., 2007). The N- and C-terminal regions of CLIP-170 are separated by a coiled-coil domain which facilitates homodimerization of CLIP-170. This region is also responsible for binding another +TIP, namely CLIP-associated protein or CLASP.
The closest mammalian homologue of CLIP-170, CLIP-115, is also a +TIP (Hoogenraad et al., 2000). All domains present in CLIP-170 are also present in CLIP-115, except the Zn-finger domains at the C-terminus, indicating lack of ability of CLIP-115 for auto- inhibition (Coquelle et al., 2002).
Dragestein and colleagues (Dragestein et al., 2008) proposed a plus-end tracking model in which an excess of binding sites present at the polymerizing plus-end bind and release +TIPs a number of times before disappearing. The disappearance of these binding sites has been thought to be exponential suggesting the control of these binding sites by the conformation of the plus-end itself. +TIP has its highest intensity at the very tip of the plus-end exponentially decaying in the direction of the minus-end, generating a
>1µm long „comet-tail‟ (Bieling et al., 2008). Interestingly, in fission and budding yeast, the CLIP-170 homologues, Tip1p and Bik1p respectively, which also track the growing MT plus-ends, do not create the „comet-tail‟ appearance but rather appear as bright dots.
The exact features of the plus-end that are recognized by the +TIPs still need to be identified, although a recent study suggests that central +TIP, EB1, probably recognizes GTP-tubulin cap present at the growing plus-end, thereby not ruling out the possibility of the plus-end targeting in a non-nucleotide dependent manner (Zanic et al., 2009).
101
Recently, it has been shown in in vitro assay that mammalian CLIP-170 requires another +TIP, end binding protein 1 (EB1), to bind to plus-ends, in addition to sequences at the C-terminus of alpha-tubulin (Bieling et al., 2008). In cells, siRNA- mediated knockdown of the +TIP protein EB1 reduces CLIP-170 accumulation at the plus-end by accelerating its dissociation from the plus-end. Interestingly, when overexpressed in the cell, EB1 decorates the entire MT lattice thereby still allowing for the specific association of CLIP-170 with the plus-end, which indicates that CLIP-170‟s recognition of the binding sites at the plus-end is its intrinsic property (Komarova et al., 2002a). These in vitro data correlate well with findings based on experiments with cultured mammalian cells (Peris et al., 2006; Dragestein et al., 2008). Interestingly, yeast CLIP-170 homologue is delivered to the MT plus-end via motor proteins (Busch and Brunner, 2004; Carvalho et al., 2004).
In the cell interior MTs grow persistently, catastrophes are rare and rescue is rapid.
However, in the area near the cell edge the MT plus-ends are highly dynamic, stochastically oscillating between periods of slow growth and fast shrinkage. Some peripheral MTs shrink to their nucleation point–MTOC (Komarova et al., 2002b). The persistent growth in the cell interior is thought to be a consequence of the increase of free tubulin concentration resulting from the boundary-induced catastrophes (Gregoretti et al., 2006) and involves EB-mediated anti-catastrophe activity (Komarova et al., 2009).
In mammalian cells, the removal of CLIP-170 and CLIP-115 inhibits the rescue of MTs at the cell cortex (Komarova et al., 2002b; Arnal et al., 2004). Instead of having frequent fluctuations of growth and shrinkage near the cell edge, the MTs in these cells display persistent growth and persistent shortening (Komarova et al., 2002a). CLIP-170 recruitment to the plus-ends was found to be enhanced in CLIP-115 knock-out fibroblasts, suggesting that these proteins compete for binding sites at plus-ends (Hoogenraad et al., 2000). Eliminating Tip1p in fission yeast resulted in impaired MT growth, leading to abnormal cell morphology (Brunner and Nurse, 2000). While in mammalian cells CLIP-170 acts as a rescue factor, the role of Tip1 and Bik1 in fission and budding yeast respectively is that of an anti-catastrophe factor (Brunner and Nurse, 2000).
Dynamic behavior of the MT plus-end is thought to be essential for the targeting of MTs towards, and the capture by, specific structures, such as cell-matrix adhesions (CMAs) (Krylyshkina et al., 2003) and the cortical actin meshwork (Fukata et al., 2002). This
“search-and capture” model was originally proposed by Kirschner and Mitchison (Kirschner and Mitchison, 1986). Through the temporally and spatially regulated capturing of individual MTs at the cell cortex, the stabilization of the usually dynamic MTs takes place. By the stabilization of a subset of MTs, the asymmetry within the cell, cell polarization, is established, providing the basis for cell morphogenesis and directed cell migration.
102
MTs regulate the positioning of leading edge during cell migration; however, they are not required for cell motility as such (Grigoriev et al., 1999). Incorrect stabilization of leading- edge-directed MTs results in deficient directional migration. The role of CLIP-170 in migration is not certain. Iin some systems depletion of CLIP-170 disrupts polarization of cells in response to scratch wounding (Watson and Stephens, 2006b), whereas in another study knockout of CLIP-170 had no influence on the formation of stabilized MTs (Drabek et al., 2006). Interestingly, CLIP-binding protein CLASP has been demonstrated to specifically localize to the plus-ends directed towards the leading edge (Akhmanova et al., 2001).
It has been suggested that one of the consequences of the extended capture and/or stabilization of MTs is that α-tubulin in the lattice of the captured MTs becomes a target for post-translational modifications (PTMs) including acetylation (addition of acetyl group) and detyrosination (removal of C-terminal tyrosine). These processes are time- dependant: the “older” the stable MTs are (t1/2 can vary between ~1h to even 16h) (Webster and Borisy, 1989) the more enriched in acetylated and detyrosinated α-tubulin they become. The MTs enriched in acetylated α-tubulin will be referred to as acetyl-MTs, those enriched in detyrosinated α-tubulin are called glu-MTs (glu for the newly exposed C-terminal glutamate residue), and those without any PTMs, the dynamic ones, the tyr- MTs. In motile cells, glu- and acetyl-MTs are mainly localized at the leading edge, with their plus-end towards the direction of migration (Bulinski and Gundersen, 1991). It has been speculated that PTMs on stable MTs specify preferred tracks for certain motor proteins, thereby facilitating directed delivery of components to the leading edge of cells (Verhey and Gaertig, 2007). Moreover, a recent study on TTL (tubulin Tyr ligase;
enzyme that ligates a tyr group on the glu-alpha tubulin) knock out MEFs suggests that tubulindetyrosination generates transiently stabilized MTs that are actually resistant to motor-driven depolymerization (for example by kinesin-13 protein mitotic centromere- associated kinesin, MCAK) (Peris et al., 2009).
Cortical capture of MTs is thought to take place at the plus-end itself. CLIP-170, strategically positioned at the plus-end of the MT, was shown to directly interact with cortical actin binding protein IQGAP at the leading edge of migrating Vero cells (Fukata et al., 2002). Additionally, IQGAP acts as an activator of two Rho GTPases, Rac1 and Cdc42- signaling proteins that regulate cell morphology and migration though actin cytoskeleton rearrangements. Interestingly, increased Rac1 activity also correlates with enhanced MT polymerization upon nocodazole-washout (Waterman-Storer et al., 1999).
In this study we set out to characterize the effects of CLIP-170/CLIP-115 depletion on the MT dynamics, MT network and the functions of the cell as a whole, in particular on cell spreading and morphology. As model cell lines we used interphase mouse embryonic fibroblasts (MEFs) derived from CLIP-170 and CLIP-115 double knockout (DKO) mice. CLIP-170 knockout MEFs show no distinct phenotype, even though the mice from which the fibroblasts were derived, are not able to produce offspring due to
103
morphological deviations of the sperm and the subsequent sperm dysfunction (Akhmanova et al., 2005). CLIP-115 deficient mice display, beside a mild growth deficit, various brain abnormalities (Hoogenraad et al., 2002).
Firstly, we aimed to explore the effects of CLIP 170/115-deficiency on the formation of the MT array. We have found that CLIP170/115-deficiency results in perturbed distribution of MTs in MEFs. Moreover, using a MT disassembly inducing agent, nocodazole, we explored the degree of stability of the various MT subpopulations in both DKO and WT cells. Our fluorescence microscopy findings indicate a greater susceptibility to nocodazole of both dynamic and stable MTs of the DKO cells compared to the WT cells.
Secondly we have asked whether the CLIPs deficiency has an effect on the attachment and spreading of MEFs. Indeed we found that lack of CLIP-170 and CLIP-115 increases the extent of cell spreading in MEFs suggesting a role for CLIP 170/115 in the control of cell spreading process. These observations prompted us to investigate whether the CLIPs-deficiency affects the cell-matrix adhesions (CMAs) in MEFs. We have found that the CMAs in DKO cells were significantly larger, but less intense than CMAs in WT cells.
Although the DKO cells have larger surface area, we found the density of CMAs to be significantly lower in DKO than in WT cells. These alterations on the level of CMAs might have consequences for the motility and (directional) migration of MEFs. However, whether these effects on CMAs are only due to the disturbance of the MT dynamics, or due to a specific role of CLIPs, remains to be investigated.
We have also investigated the effect of CLIP170/115-deficiency on the MT plus-end morphology in detail using electron microscopy (EM). For this purpose we compared MT plus-end structure in the DKO MEFs to wild type (WT) MEFs. Since CLIPs are important MT dynamics regulators in the cell cortex (Komarova et al., 2002a) we expected to see a shift in the distribution of plus-end conformations in the CLIP-170/115 DKO cells in this region compared to WT MEFs. We found that eliminating the CLIPs results in a decreased frequency of growing MT plus-ends, which is in accordance with previous results obtained by live-cell fluorescence microscopy.
Material and methods Reagents
Collagen was Upstate Collagen Type I, rat tail 08-115 from Millipore. Primary antibodies used were against tyr-tubulin (rat monoclonal, clone YL1/2, Abcam, Cambridge, UK), detyrosinated tubulin (glu-tubulin; rabbit polyclonal, Chemicon), Acetylated tubulin (mouse monoclonal, Clone 6-11B-1, Sigma-Aldrich), anti-PY antibody for the staining of cell-matrix adhesion proteins (mouse monoclonal, Santa Cruz), paxillin, phosphoryated form of paxillin (phosphorylated at Ser118; rabbit polyclonal, Biosource) and nucleus (Hoechst staining). Secondary antibodies used were goat-anti-rat-Alexa 488, goat-anti-
104
mouse-Alexa 594 and goat-anti-rabbit Alexa 488 were (all 1:200) all from Molecular Probes. Rhodamine-phalloidin (actin staining) was used in dilution of 1:1000 in PBS.
Vectashield mounting medium was from Vector Laboratories. Nocodazole was from Sigma.
Cell culture
Experiments described in this chapter were performed on primary wild type (WT) and CLIP-115/-170 deficient, or double-knockout (DKO), mouse embryonic fibroblasts (MEFs), derived from E13.5 days embryos as previously described (Hoogenraad et al., 2002). MEFs were cultured in Dulbecco‟s MEM in a 1:1 ratio with Ham‟s F10 medium (complete medium), supplemented with 10% fetal bovine serum, 2mM L-glutamine and antibiotics, in a humidified atmosphere with 5% CO2 at 37˚C.
Specimen preparation for fluorescence microscopy on MTs
For immunefluorescent analysis WT and DKO CLIP-170/115 MEFs were grown in complete medium (see above) overnight on collagen-coated (10ug/ml) and BSA (1% in PBS) -treated glass coverslips, until reaching 50% confluence. Cells were fixed for 10 minutes with cold methanol, containing 1mM EGTA at -20°C and incubated in blocking buffer for 45 minutes at RT. Next, cells were incubated with primary antibodies for 1h at room temperature. The primary antibodies used were against tyr-tubulin, detyrosinated tubulin, acetylated tubulin and phospho-tyrosine (PY) containing proteins for the staining of cell-matrix adhesion proteins. Samples were then washed three times in PBS/0.05%
Tween-20 and incubated with rhodamine-phalloidin (actin staining), goat-anti-rat-Alexa 488, goat-anti-mouse-Alexa 594 and goat-anti-rabbit Alexa 488 secondary antibodies for 1 h at RT. Next, cells were washed three times in PBS/0.05% Tween-20, incubated 5 minutes with DAPI solution and again washed three times in PBS. Samples were then dipped for few seconds in 70% and 100% ethanol respectively, air dried and mounted onto the glass slide using Vectashield mounting medium.
Images of the samples were made using a Leica DMRXA fluorescence microscope with a Coolsnap K4 camera using ColorPro software.
Nocodazole assay
Nocodazole assay was performed by exposing the cells, grown on collagen-coated coverslips as described in §5.2.3, to 10µM nocodazole for either 5, 15 or 30 minutes. A part of the samples were treated with only medium without nocodazole (=control) and another part was left to recover for 30 minutes in the incubator after 30 minute long nocodazole treatment and a wash step with pre-warmed medium (=nocodazole washout, WO). The samples were then fixed and further processed as described in § 5.2.3.
105
Cell-surface measurements
Cells were grown overnight on collagen-coated, BSA-blocked coverslips, fixed and stained for PY proteins as described in §5.2.3. Images were acquired using Nikon Eclipse E600 fluorescence microscope at 10x magnification. Image Pro 6.2 software was used for further processing of the images and cell area measurements as follows. First, the intensity threshold was set as high as to be able to get the signal from the whole cell cytoplasm, covering the whole cell area. If the cells were touching each other, a line was drawn manually to separate the cells. The area was then automatically calculated for all the separate objects in the image. We have analyzed 217 WT and 148 DKO cells analyzed in two independent experiments.
Cell surface area of the cells plated on 96-wells glass bottom plate (see §5.2.6;
Sensoplate, Greiner Bio-One) was measured by manually drawing of the cell perimeter using tools in ImageJ software. In total 29 WT and 21 DKO cells were analyzed.
Cell-matrix adhesion (CMA) analysis
MEFs were seeded on a 96-wells glass-bottom plate (Sensoplate, Greiner Bio-One) in density of 65 000 cells/ml. Prior to seeding of the cells, the glass-bottom plate was coated with extracellular matrix. This was done by incubating the plate with collagen solution (10µg/ml) for 1 hour at 37°C. Next, the collagen solution was sucked off from the wells and the wells was then incubated with BSA (1% in PBS) for 1 hour at 37°C in order to block the aspecific binding of the cells. The wells were then washed with PBS and the cells were seeded in complete medium, grown overnight, then fixed with 4%
formaldehyde, blocked with blocking buffer and fluorescently stained for paxillin, phosphorylated form of paxillin (p-pax at Ser118) and nucleus (Hoechst staining).
Imaging of DKO and WT cells was performed on Nikon high-throughput Eclipse Ti screening microscope using 20x/0.75 NA objective, 3x magnification. The illumination (laser/gain) settings were the same for all the acquired images, for both WT and DKO cells. Images of 512x512 pixels were acquired randomly using an automated system and processed for segmentation of nuclei and cell-matrix adhesions (CMAs). The segmentation was performed using a new segmentation method termed „Watershed mask with remerging‟ using an ImageJ-based in house developed macro (Kuan Yan, manuscript in preparation). The CMAs segmentation was performed only on images of paxillin stained CMAs, since this antibody did not give high cytoplasmic and nuclear staining like phosho-paxillin antibody. An example of CMA and nuclei segmentation is given in Figure 7. CMAs of 29 WT and 21 DKO MEF cells were analyzed (~7000 and
~6400 CMAs, respectively). Parameters like CMA size, number CMAs per cell, CMAs intensity, distance from nucleus, distance from one CMA to the other (density) and number of cells measured, were automatically calculated using the above mentioned macro. Since parameters „size‟ and „intensity‟ were found not to follow normal distribution, the statistical analysis was performed using Kolmogrov-Smirnov (KS) test
106
(alpha= 0.005) (Paran et al., 2007). In order to calculate the CMA density in WT and DKO cells the surface area of these cells was measured by manual drawing of the cell edges using ImageJ software (see second paragraph in §5.2.5; see Table 2 in Results).
Specimen preparation for Transmission Electron Microscopy (TEM)
MEFs were grown overnight on 13nm Thermanox® coverslips at 37˚C in complete medium in a humidified atmosphere with 5% CO2, a time frame in which cells reached
~80% confluence. After blotting off the access of medium in an automated temperature (37˚C) and humidity regulated chamber, the cryo-immobilization of the sample was achieved by automatic plunging of the sample into liquid ethane chamber using the
„Leiden Vitrification System‟ (LVS) device. Cryo-immobilized samples were then transferred into liquid nitrogen and subsequently into tubes containing a frozen mixture of 0.1% OsO4 and 0.25% uranyl acetate (dissolved in methanol) in anhydrous acetone.
The samples were freeze-substituted using an automated freeze substitution apparatus (AFS System, Leica). Freeze-substitution was performed for either three days („long program‟) or one day („short program‟) at -90˚C. The temperature was then slowly (10˚C/hour) brought to -20˚C, where samples remained for 12 hours, and finally slowly (10˚C/hour) increased to 0˚C. Cells were washed twice with acetone for 15 minutes, and infiltrated with increasing concentrations of epon. The cells were then flat embedded in epon and left to polymerize for two days at 60˚C.
Thin sections (~100nm) were cut using a Reichert Ultracut S microtome. Sections were collected on Formvar-coated copper mesh grids and stained with 7% uranyl acetate in methanol and lead according to Sato. Cell sections were investigated in a Philips CM10 Electron microscope at 80kV.
Results
CLIPs are required for proper MT distribution in interphase MEFs and regulation of cell spreading
A typical MT array of interphase fibroblasts consists of radially organized MTs, running from MT organizing center (MTOC) toward the cell membrane thereby covering the whole cytoplasm. In order to investigate the effects of CLIP-170/ CLIP-115 depletion on MT network in MEFs we fluorescently immuno-stained MTs in WT and DKO cells using anti-tyr-tubulin antibody (Figure 1, Figure 2A). Staining against phosphorylated tyrosine residues of the proteins (PY) was used to visualize cell-matrix contacts, thus cell surface attached to the extracellular matrix.
While most of the WT cells displayed an extensive radial MT array, in a small number of WT cells, areas deprived of MTs were observed. Interestingly, the frequency of such cells with at least one MT-free area (Figure 2A and B) was significantly higher in DKO
107
population than in WT population, 0.102 ± 0.0819 for WT versus 0.587 ± 0.0180 for DKO cells (i.e. 10% of the WT cells showed MT-free areas versus ~60% of the DKO cells, see Figure 2C, here 93 WT and 72 DKO cells were analyzed). Moreover, in contrast to the WT population, the DKO population showed three and more of these MT-free zones within one cell (Figure 2D). The MT-free zones were previously found to contain aggregates of dynein/dynactin (personal communication N.Galjart).
Next to the MT-free zones, we also observed donut-like “holes” in the peripheral regions of both WT and DKO cells, highlighted by surrounding tubulin deposition (Figure 4D, asterisk). We assume that in these cases the cell cortex became too thin, generating a
“hole”.
Figure 1. Perturbed MT dynamics in CLIP 170/115 deficient MEFs result in an aberrant distribution of MTs. WT and DKO cells were stained for dynamic MTs using anti-tyr-tubulin antibody and for phosphorylated tyrosine-containing proteins (PY) to visualize the cell-extracellular matrix adhesion complexes. While most of the WT cells showed a typical MT array in which MTs covered the cytoplasm of the entire cell, in DKO cells often MT-free areas (indicated by the asterisks) were visible. Magnification used for all images was 60x
108
Figure 2. CLIP 170/115 deficient MEFs show increased formation of MT-free zones and enhanced cell spreading compared to the WT MEFs. In order to visualize the MT network and the whole cell surface simultaneously, WT and DKO MEFs were stained for tyr-tubulin (A) and phospho tyrosin (PY) containing proteins (i.e tyrosin-phosphorylated cell-matrix adhesion proteins) (B). Bar, 20μm.
For quantification of MT-free zones/cell, areas devoid from MTs or with a low number of MTs, as indicated in (A) with an asterisk, were counted in WT and DKO cells. In total 93 WT and 72 DKO cells were analyzed. The results are summarized in (C) and (D). 90% of WT and 40% of DKO MEFs exhibited a normal phenotype without MT-free zones. On average, 10 percent of the WT cells had 1 or 2 MT-free zones per cell (D), while 60% of the analyzed DKO cells had 1-6 MT-free zones per cell (C) and (D). (D) Depicts results from two independent experiments from (C). Actin staining as shown in (E) is used to quantify morphological differences between WT and DKO cells. Quantitative cell surface area measurements indicate that DKO cells are on average twice as large as WT cells (F). Cell area measurements were performed on images of actin-stained WT and DKO (E) cells using Image Pro 6.2 software by setting the intensity threshold as high as to be able to visualize the whole cell cytoplasm, covering the whole cell area. Cells that were touching each other were
109
manually separated by drawing a line between the cells. The area was then automatically calculated for all the separate objects (cells) in the image. In total 217 WT and 148 DKO cells were analyzed in two independent experiments. Bars represent standard deviation between the experiments.
Actin staining (and others) showed that both WT and DKO cells created rather heterogeneous populations, regarding cell size and morphology, which is typical for primary cells (Figure 2E). We have analyzed 217 WT and 148 DKO cells, stained for actin, in two independent experiments (in total 10 images for each cell line). The average cell-surface area of DKO cells was estimated to be twice as large as that of WT cells;
21071± 2014 pixels versus 13077 ± 347 pixels (Figure 2F). This is in agreement with previous findings (personal communication N. Galjart). These data suggest a role of CLIPs in the regulation of cell spreading.
Dynamic MTs of CLIP170/115-deficient MEFs are more susceptible to nocodazole- induced depolymerization than MTs in the WT MEFs
To gain more insight in the stability of the MT array we exposed WT and DKO cells for various time periods to the MT depolymerizing agent nocodazole. We then fixed and stained cells for dynamic MTs using anti-tyr-tubulin antibodies. WT MEFs showed retraction of MTs from the outermost peripheral area after 5 minutes of nocodazole treatment (Figure 3). Nevertheless, large parts of the dynamic (tyrosinated-α-tubulin enriched) MT network were still visible even after 15 minutes of nocodazole-exposure.
Remarkably, the 5 minute long nocodazole treatment caused a significant loss of MT array in DKO cells. This effect was even more prominent upon 15 minutes of exposure.
Peripheral patches of tyr-tubulin in these cells at 15‟ indicate deposition of soluble tyr- tubulin in this area, while the MTs are hardly to be found. These findings suggest that CLIPs are involved in stabilization of the MTs and protection from nocodazole induced depolymerization.
CLIPs are not essential for the formation of stable acetyl-tubulin and glu-tubulin enriched MTs
Since the effects of CLIPs-deprivation on the stability of the dynamic MTs were observed in DKO cells, we next investigated whether CLIP170/115-deficiency causes perturbations of the stable MT network. Stable MTs are known to be enriched in acetyl- tubulin and/or detyrosinated or glu-tubulin. Interestingly, staining for stable MTs using anti-acetyl-tubulin antibody showed no clear differences in WT versus DKO MEFs
110
Figure 3. MTs of CLIP 170/115 deficient MEF cells (DKO cells) are more susceptible to nocodazole-induced depolymerization than the MTs in WT MEFs.
DKO cells exposed to nocodazole lost a large part of their tyr- MT network already after 5‟ of exposure to 10µM nocodazole.
This effect was even more prominent upon 15‟ exposure.
Peripheral patches of tyr-tubulin in these cells (15‟) indicate deposition of tyr- tubulin in this area, while the MTs are hardly to be found.
Interestingly, nocodazole treated WT MEFs showed retraction of MTs from the peripheral area;
nevertheless, the large parts of the MT network were still visible after 5‟ and 15‟
of nocodazole- exposure. The images are representative for three independently repeated experiments.
extended lives. Moreover, these findings indicate that, if the longevity of these MTs is initialized and /or established though capturing and linking of the MT plus-end to the cortical actin and/or membrane, this process does not involve CLIP-170 and CLIP-115.
111
CLIPs deficiency abolishes nocodazole-resistance of the long-lived MTs
Next we investigated whether CLIPs affect the sensitivity of the long-lived MTs to nocodazole. It has been shown previously that acetylated and detyrosinated (long-lived, stable) MTs are nocodazole-resistant (Kreis, 1987), thus we wondered whether the nocodazole-resistant acetyl-MTs and glu-MTs would be as stable in WT as in DKO cells.
To test this we exposed WT and DKO cells to 10µM nocodazole for various time periods and stained them for α-tubulin, acetylated-α-tubulin and detyrosinated or glu-α-tubulin.
Upon 30‟ nocodazole-treatment the WT cells still showed a relatively extensive acetyl- MT array (Figure 4C). Surprisingly, the majority of the acetyl-MTs in the DKO cells disappeared upon 30‟ nocodazole exposure, leaving only a few short MTs in the cell center (Figure 4D). Nocodazole-washout (WO) lead to recovery of a normal acetyl-MTs arrays in both WT and DKO cells (Figure 4E and 4F respectively). In WO condition we observed no apparent difference in stable MTs arrays between the WT versus DKO cells.
These findings indicate that even though CLIPs themselves are not involved in direct capturing of the MT plus-ends in the periphery of the cells, the stability of these
“nocodazole-resistant” MTs is enhanced during nocodazole-induced shrinkage when CLIPs are present. This extra stabilization of the “long-lived” MTs could possibly be due to the binding of CLIPs onto the MT lattice or by stabilization of the conformation of the plus-end. Also CLIPs might recruit other factors that stabilize the plus-ends of the nocodazole-resistant MTs.
112
Figure 4.
CLIP170/115 deficiency enhances nocodazole sensitivity of the stable, Acetyl- tubulin enriched
MTs. MEFs
were either not- treated (control, A-B), or treated with nocodazole for 30 minutes (C-D), or treated with nocodazole for 30 minutes, washed with pre- warmed medium and left to recover for 30 minutes (E-F).
Antibody against acetylated α- tubulin was used to visualize stable MTs. (A) WT control cells and (B) DKO control cells both show a typical array of acetyl- MTs. (C-D)
Nocodazole caused a loss of acetyl-tubulin enriched MTs in both WT (C) and in DKO (D) cells.
However, the nocodazole induced MT depolymerization was much more prominent in DKO cells.
Insets represent zoom-in of the marked area of the cell, showing the typical wavy acetyl-MTs (arrows). In the periphery of DKO cells donut-like structures with deposition of acetylated-tubulin are visible (asterisks) which might represent the “holes” previously observed in these cells (Niels Galjart, personal communication). Recovery after nocodazole exposure allowed for the formation of a new acetyl-MTs network in both WT and DKO cells (E and F, respectively). No clear differences between the networks in WT and DKO cells were observed. Bar 50 µm. The images are representatives for two independently repeated experiments.
113
Nocodazole-resistant MTs in MEFs are enriched with both detyrosinated tubulin and acetylated-tubulin simultaneously
Nocodazole treated DKO cells were observed to lose most of the acetyl-MTs and glu- MTs (Figure 5). However, a small population of these MTs still remained after the 15 and 30 minute nocodazole treatment (Figure 4 and 5 respectively). We looked more closely to these nocodazole-resistant MTs (here double immuno-labeled for glu-tubulin and acetyl-tubulin). These MTs appeared to contain interspersed acetyl-tubulin- and glu- tubulin-enriched fragments. This was even more obvious in nocodazole-treated DKO cells (lower panel in Figure 5 and Figure 5G) because the MT density is low and individual MTs can easily be traced. Interestingly, it appears that the MTs containing fragments positive for both acetyl-tubulin and glu-tubulin are the ones that are the most nocodazole-resistant. It is plausible that the MTs containing both glu-Tub and Acetyl-Tub are less sensitive to nocodazole-induced depolymerization than the MTs enriched with only one of the two forms of modified α-tubulin possibly by attracting proteins that can bind to and stabilize the MT.
Figure 5. Acteyl-MTs and Glu-MTs in DKO cells in control situation (respectively A and B; C is merged image) and upon 15‟ exposure to nocodazole (D and E; F is merged image). Often the MTs that are left after the nocodazole-treatment are positive for both Glu-tubulin and Acetyl-tubulin (yellow color). Note the fragments of Acetyl-tubulin alternating with the Glu-tubulin fragments (magnified inset from F) in (G). Bar 30µm.
114
CLIPs stabilize plus-end conformations associated with the MT assembly
In CHO cells, CLIPs regulate dynamic instability of the MT plus-end by acting as rescue factors, inducing conversion from shrinkage to growth. In addition, dynamic instability in CLIP-170/115-deficient interphase MEFs has been found to be perturbed. Here the rescue frequency is five times lower compared to the WT MEFs, therefore upon catastrophe most MTs in CLIP-deficient cells shrink persistently, eventually ending up in the MTOC (Komarova et al., 2002a); personal communication N.Galjart). Despite the decrease of plus-end rescues and associated persistent MT shrinkage, CLIP-deficient cells are still able to form a MT network in steady state situation. We used electron microscopy (EM) to visualize and compare MT plus-ends in WT and DKO cells. Cryo- fixed/freeze-substituted, resin flat-embedded MEFs were sectioned parallel to the plane of cell attachment. Sections ~100nm thick were examined using EM. The periphery of MEFs is filled with individual actin fibers and more or less dense actin bundles (Figure 6A). Large parts of the periphery were framed by electron-dense bundles of cortical actin
>0.5 µm thick precluding scoring of plus-ends within this area (Figure 6B). Instead MT plus-ends were scored within a virtual peripheral ring of ~7 µm from the cell membrane toward the cell center, leaving out the areas covered by extensive actin bundles.
We scored in total 131 MT plus-ends in the cell periphery of WT cells, and 171 in DKO cells in three independent experiments. Plus-ends were subsequently classified with regard to their characteristic morphological features as previously described (this thesis;
(Zovko et al., 2008). We observed nine distinctive plus-end morphologies, previously also found in 3T3 cells. Figure 6C shows the frequency of various plus-end conformations found in the periphery of either DKO or WT cells. All conformations found in the WT cells were also present in DKO cells, indicating that CLIP-170 and CLIP-115 themselves do not participate in the design of a particular plus-end conformation, whether that may be through their co-polymerization with tubulin or through direct/indirect binding at the MT tips.
Furthermore, it appears that blunt plus-end conformation is rare in MEFs, which is consistent with previous findings in 3T3 fibroblasts (Zovko et al., 2008). If we look at the trends, the frequency of flared, forked and sheet-straight plus-end conformations were more prominent in WT than in DKO MEFs, while frayed and sheet-frayed were more prominent in DKO cells relative to WT. However, a significant difference was only found for sheet-straight ends (p=0.03). The other conformations were more or less present in similar amounts in WT relative to DKO cells. Although the data are not statistically significant, forked and flared plus-ends may also be influenced by the CLIPs. According to our previous kinetic analysis model the forked and sheet-straight ends are associated with growth, while frayed and sheet-frayed ends 1 and 2 are associated with shrinkage (Zovko et al., 2008). Flared ends were found to be a transition state from forked and early-frayed towards frayed end (Figure 4 in (Zovko et al., 2008). These findings confirm that CLIPs-170/115 deficiency leads to perturbed dynamic instability in the cell periphery,
115
resulting in a loss of growing MTs. Due to the reduced rescue levels, the shrinking plus- ends cannot be converted into the growing ones anymore, resulting in a significant decrease of sheet-straight plus-ends, and a relative increase of the shrinkage-associated plus-end morphologies.
116
Figure 6. CLIP-170/115 deficiency results in significant reduction of sheet-straight plus-end conformation. (A) Electron micrograph showing MTs (arrows) in the periphery of a MEF cell running towards the actin meshwork at the cell edge (asterisks). Bar, 500nm. (B) MTs (arrows) are running parallel to the dense actin bundle (asterisk) aligned with the cell membrane. Bar, 500nm. MT plus- ends in the proximity of cell membrane were visualized using electron microscopy and categorized with respect to their morphological features, as described previously (Zovko et al, 2008). (C) MT plus-end morphologies associated with growth (forked and sheet straight) are more prominent in WT then in DKO MEFs, while the morphologies associated with shrinkage ((early) frayed, sheet- frayed 1 and 2) are more frequent in DKO cells, compared to WT (mean±stdev). Significant difference was only found for sheet-straight ends (p=0.03; n=3).
117
CLIP170/115- deficiency leads to perturbed cell-matrix adhesion signaling in MEFs Processes like cell matrix adhesion and cell migration are mediated by integrin-based cell-matrix adhesion (CMA) structures. The cell attachment/ cell spreading of the CLIPs deficient cells (DKO) on collagen-coated glass surface was found to be altered when compared with the WT cells i.e the average surface area of the attached DKO cells was found to be significantly larger than that of WT cells (Figure 2). We hypothesized that the observed phenotype might be due to the altered CMA function caused by CLIPs absence. Thus, next we investigated the CMAs in WT and compared these to CMAs in DKO cells.
Figure 7.
Segmentation of nuclei and paxillin stained cell-matrix adhesions (CMAs) in MEF
cells. An
example of a raw image (A) containing information from three different channels (paxillin, p- paxillin and nuclei).
Segmentation of the image in nucleus-channel gives the mask of the cell nuclei. (C) Contrast- enhanced image showing only paxillin stained CMAs and the corresponding segmentation (D). Notice the presence of both classical CMAs (arrowhead) and elongated-fibrillar adhesions (FBs; arrow). Magnification 60x.
In order to investigate the role of CLIP-170 and CLIP-115 deficiency on the formation and distribution of CMAs, we stained WT and DKO cells for the CMAs proteins- vinculin, talin, phospho-Ser118-paxillin, paxillin, focal adhesion kinase(FAK), phospho-FAK as well as for phospho-tyrosinated proteins in general (antibody against PY). In contrast to all the other stainings, paxillin staining gave the lowest cytoplasmic signal and no nuclear signal allowing for the proper CMAs segmentation (Figure 7A and
118
C). Nuclear staining (Hoechst) was used in order to automatically count the cells within an image and to
Figure 8.
Analysis of the size and intensity of CMAs in WT versus DKO cells. Size (pixels) in WT (A) and DKO (B) cells follows a non-normal distribution.
Histogram of DKO cells shows less small adhesions (2 to
~30pixels) compared to the histogram of WT cells. Cumulative distribution function (CDF) of CMA size shows differences in size distribution between CMAs in WT vs. DKO cells. The difference between the two distributions (A vs. B) was shown to be significant (KS test, alfa=0.005), with WT cells having smaller CMA (see also Table 1). CDF of CMAs intensity showing different distributions for WT versus DKO cells. KS test (alfa=0.005) showed the intensity of the CMAs in DKO cells to be significantly lower than in WT cells. Analysis was performed on ~7000 and ~6400 CMAs present in 29 WT and 21 DKO cells respectively.
determine the distance of CMAs to the nucleus (Figure 7B). Images (512x512 pixels) were acquired randomly on Nikon high throughput screening microscope.
In total 29 WT and 21 DKO cells were analyzed (~7000 and ~6400 CMAs, respectively).
An example image of MEFs stained for paxillin and phospho-paxillin is shown in Figure 7A. The corresponding one-channel image of paxillin stained CMAs is shown in Figure 7C. Comparison of Figure 7C and D shows that the small dot-like adhesions or focal complexes (FCs), „classical‟ CMAs, also termed focal adhesions (FAs), and fibrillar adhesions (FB) are picked-up by our segmentation method, while the cytoplasmic noise is ignored.
Statistical analysis (KS test, alpha=0.005) of the size of CMAs showed that CMAs of DKO cells are significantly larger than CMAs in WT cells (Figure 8A and 8B, Figure 8C,
119
Table 1), while their intensity was significantly lower than in WT cells (Figure 8D, Table 1).
The distance of CMAs to the nucleus was significantly smaller in WT than in DKO cells (Table 1), while the CMAs density was higher in WT cells (parameter „closest CMA distance‟ is larger in DKO cells). Similar result for CMAs density is obtained when the total number of CMAs (normalized for the cell number in WT and DKO condition) was divided with the total investigated area (pixels) (Table 2).
Number of CMAs per cell was found to be significantly lower in WT cells compared to DKO cells (Table 2), but since the WT cells were found to be (significantly) smaller than DKO cells (Table 2, cell area: WT=31768 pixels and DKO=45981 pixels, p=0.00, KS test;
see also Figure 2E-F) we calculated the number of CMAs per area (in pixels) or CMAs density. The results indicate that the CMAs density in DKO cells is almost twice as low as in WT cells (0.14 versus 0.22, Table 2), even though the number of CMAs/cell in DKO is larger than in WT (Table 2).
Table 1. Various CMAs parameters derived from the analysis of 29 WT and 21 DKO cells. In all cases comparison of DKO and WT CMAs was performed using KS test with alfa=0.005). 1 stands for significantly „larger‟ or „higher‟ in DKO than in WT, -1 “lower” or “shorter” for DKO than in WT, and 0 stands for „similar‟ in DKO vs. WT (no significant difference at alfa=0.005) .
CMA 'area' CMA 'Intensity' 'CMA-Nucleus Distance'
'Closest CMA Distance'
DKO vs. WT 1 -1 1 1
p-value 0.0000 0.0000 0.0000 0.0000
average WT 14.69319 29.07683 93.35116 7.001336
average DKO 16.52411 24.89299 109.4811 7.31245
Table 2. CMA density in WT and DKO cells (number of CMAs per area (in pixels))
#FA/Cell total #FA in investigated area Total area investigated/#Cell #FA/Area
WT cells 246 7121 31768 0.224
DKO cells 305 6396 45981 0.139
120
Discussion and conclusionThe MT plus-end is an intrinsically dynamic structure, but also a scaffold utilized by a variety of proteins which associate with it and thereby modify its properties in order to specialize it for the use in a particular, often spatiotemporally regulated, process.
In mammalian interphase cells, MTs often grow persistently from the MTOC, which is situated near the cell nucleus, running with their plus-ends towards the cell membrane where they exhibit dynamic instability, i.e. an oscillation between phases of growth and shrinkage. This behavior results in a concentration of MT plus-ends in the cortical area.
CLIP-170 and CLIP-115 associate specifically with the growing plus-ends. In mammalian interphase fibroblasts, CLIPs regulate the dynamic instability of the plus-end by acting as rescue factors. Dynamic instability in the peripheral area of CLIP-170/CLIP-115 deficient interphase MEFs is perturbed in such a manner that the rescue frequency is five times lower compared to that in WT MEFs. Therefore, upon catastrophe most MTs of DKO cells shrink persistently (Komarova et al., 2002a); personal communication N. Galjart).
This behavior leads to a relative decrease of growing plus-ends in the cortical area.
In this study we used WT and DKO primary MEFs as a model in order to evaluate the effects of absence of CLIPs on MT plus-end structure, MT network and cell morphology.
Our findings are discussed below.
CLIPs and the structure of the MT plus-end
A MT plus-end oscillates between different dynamic states and these transitions are accompanied by defined conformational changes of the plus-end (Zovko et al., 2008).
Here we aimed to elucidate whether CLIP-170 and CLIP-115, that have been shown to modify the plus-end dynamics, also modify and influence plus-end morphology itself.
Therefore by means of electron microscopy (EM) we examined and scored plus-ends in the peripheral area of WT and compared them to CLIP-deficient MEFs. We hypothesized that absence of the CLIPs might result in a shift in frequency, or even a malformation, of particular plus-end conformation(s).
We observed nine conformations of plus-ends in MEFs, a finding consistent with those in previous study in 3T3 fibroblasts (Zovko et al., 2008). Interestingly, the frequencies of the plus-end conformations were found to deviate from those in 3T3 fibroblasts. This can be explained by the fact that the plus-end scoring area of MEFs differed from that in 3T3s, MEFs being partly obstructed by thick actin bundles in the region near the cell membrane. Moreover, unlike 3T3 cells, MEFs used here were primary cells. Metabolism and other processes in primary cells have often been shown to differ from those in growth-inhibited, „cancer-like‟ cell lines like 3T3. In addition, the different outcomes could possibly be due to the great morphological heterogeneity of the MEFs.
121
Moreover, recent papers show that MT plus-ends in interphase cells often target structures at the cell-substratum interface at the periphery of the cell. Sectioning of the embedded cells for the use in (tomo) EM often leads to cutting off of these plus-ends near the substratum. Cell periphery thickness inevitably has an effect on the angle of dipping of the MTs into e.g. cell-matrix adhesions, and thus on the probability that a near-substratum plus-end is eliminated from the sample due to the imperfections of the technique used. Whole-mount cryo(tomo) EM would therefore be the technique of choice to proceed with in order to investigate in more detail what the differences are between 3T3 and MEF cells are.
The finding that all nine plus-end conformations found in WT MEFs were also present in DKO MEFs indicates that CLIPs are not required for the design or modulation of the plus-end conformation/morphology as such; suggesting either that this is an intrinsic property of the plus-end, or that proteins other than CLIPs are involved in the formation of a particular plus-end conformation.
We observed a significantly lower frequency of sheet-straight plus-ends in CLIP-deficient compared to WT cells. Since sheet-straight plus-ends were previously shown to correlate with MT growth, this finding suggests that absence of CLIPs results in a decrease of growing MTs in cell periphery. This confirms that CLIP-170 and -115 promote MT growth in cells. It is hard to speculate about other differences in frequencies as the deviations between the three experiments were rather large (Figure 6C). These deviations could possibly be due to a large heterogeneity in size and morphology within both WT and DKO populations. Nevertheless, a trend is visible for the average frequencies of the conformations previously proposed to be associated with growth (forked and sheet- straight) (Zovko et al., 2008) : these conformations were at least thirty percent elevated in WT relative to DKO cells, while those associated with shrinkage (frayed, sheet-frayed 1) were slightly decreased in WT compared to the DKO cells. These observations can be explained by the fact that reduced rescue frequency in DKO cells results in more persistently shrinking MTs. Since other dynamic instability parameters (growth and shrinkage rate and catastrophe frequency) are not affected by CLIPs absence, a consequence is a relative decrease of growing MT plus-ends in the peripheral area of DKO cells.
Although we can conclude that CLIPs are not required for the construction of any of the nine plus-end conformations as such, we cannot rule out that the observed reductions in frequency of sheet-straight, forked and flared ends were partly due to the lack of binding of CLIPs to these conformations and their subsequent stabilization. Consequently CLIPs might associate with conformations present during growth, like forked and sheet-straight conformations (comet-tail structures at the growing MTs). However, the flared conformation is, in contrast to forked and sheet-straight, characterized by slightly curved protofilaments which probably have lost their connection with each other, which is a prerequisite for the onset of disassembly. Diminished frequency of the flared end in
122
CLIP-170/115-deficient cells implies that CLIPs might stabilize this conformation, preventing it from collapsing. In other words the flared end might be a transition conformation on which CLIPs could bind with higher affinity, and rescue it from its conversion to the frayed end (Zovko et al., 2008). This is supported by the findings that N-terminal part of CLIP-170, which contains the MT binding domains, was observed to bind to curved tubulin oligomers in assembly condition in vitro (Arnal et al., 2004). This indicates that CLIPs could have two modes of action: one would be the tracking of the growing plus-end, the other- stabilization and rescue of the flared conformation from transiting into a frayed end.
Although there is strong evidence that CLIP-170, and probably also CLIP-115, require EB1 protein to track MT plus-ends, the question “which structural features of the plus- end are recognized by EB1/CLIP-170” is still not answered. In cells, siRNA-mediated knockdown of EB reduces CLIP-170 accumulation at the plus-end by accelerating its dissociation from the plus-end. When over-expressed in the cell, EB1 decorates the entire MT lattice thereby still allowing for the accumulation of CLIP-170 specifically at the plus-end, which indicates that CLIP-170‟s recognition of the binding sites at the plus-end is due to CLIPs‟ intrinsic property and thus EB-independent (Komarova et al., 2005).
What do sheet-straight, forked and flared structures have in common that can be recognized by CLIPs? First, CLIPs might recognize the luminal side of the MT which is exposed in each of the three conformations. This would mean that also sheet-frayed 1 and 2 should be recognized by CLIPs, as they also have their luminal sides exposed.
This is, however, not supported by our findings. In addition, the signal of GFP-tagged CLIP-170 shows a few micrometers long comet-tail (Perez et al., 1999) yet none of these three structures has such a large luminal part exposed. Moreover, a recent publication on the mechanism of +TIP association with the plus-end suggests the process of binding and exchange to be rather rapid and controlled by cytoplasmic diffusion (Dragestein et al., 2008). All together these findings make it unlikely that CLIPs bind the lumen side of MTs.
Another recognition point for CLIPs could be straight protofilaments like those in sheet- straight and forked ends. Straight protofilaments are associated with the presence of GTP-tubulin, while curved are associated with GDP-tubulin. Although, unlikely, it is possible that smooth curving protofilaments of flared ends contain a few molecules (remnants) of GTP-tubulin, which would serve as recognition sites for CLIPs (Dimitrov et al., 2008). This is in agreement with a recent study that suggests that EB1 proteins discriminate between GTP- and GDP-bound tubuIin, with preferential binding to the GTP-tubulin (Zanic et al., 2009).
Interesting about forked and flared ends is that they have a funnel-like shape (Zovko et al., 2008) which indicates that the contacts between the protofilaments are not firm. It is plausible to say that the contacts between the protofilaments within an open sheet of sheet-straight ends are also not as tightly associated as in a MT. Hence, CLIPs might
123
recognize the sites between the two adjacent loosely connected relatively straight protofilaments. Each CLIP dimer could bind to the (tyrosinated) α-tubulin from the two adjacent protofilaments acting as a molecular zipper, or provide the binding site for other proteins which would bring the protofilaments together and “close” the gap. This fits well with the proposed “polymerization chaperone”-model in which multimerized CLIP-170 MTB domains (CAP-GLY) are suggested to be a prerequisite for plus-end localization and tracking by CLIP-170 and other +TIPs, since a single MTB domain does not localize at or promote growth of the plus-end (Slep and Vale, 2007). However, it should be noted that the recent findings on EB proteins demonstrate that the dimerization is not required for the tracking behavior of EB, but it is for its anti-catastrophe activity in cells (Komarova et al., 2009).
It is possible that CLIPs have multiple functions at the plus-end which are utilized in different situations. Firstly, CLIPs might bind as monomers to the growing (sheet-straight and forked) ends and track these through fast association-dissociation (Dragestein et al., 2008) via EB-binding, providing the stabilization of these structures. Secondly, CLIPs might bind to the ends which are likely to collapse, like flared end, in order to re-establish lateral connections between the protofilaments, with other words to “rescue” the plus-end from disassembly. This mode of action could possibly utilize CLIPs in dimerized form.
Role of CLIPs in organization of the MT network
Tyr-tubulin immunofluorescence staining of WT and DKO MEFs showed the existence of MT-free zones in the primary MEFs. Cells with MT-free zones were much more frequent in the DKO than in the WT population. Interestingly, in DKO cells these areas deprived of MTs were found to co-localize with dynein/dynactin aggregates (personal communication N.Galjart), indicating a role of CLIPs in inhibition of aggregation of these proteins. This is not surprising, as the dynactin subunit, p150Glued, which is also a +TIP, was previously shown to be recruited to the plus-ends by CLIP-170 (Lansbergen et al., 2004). Perhaps this recruitment is necessary to prevent aggregation.
An explanation for the large number of MT-free zones in DKO cells could be the fact that in these cells rescue is five times less frequent then in WT cells. MTs which shrink without being rescued shrink until they reach the MTOC, which leaves an area deprived of MTs. These collapsed areas probably do not allow MT re-growth, creating permanent MT-free zones.
The DKO cells were found to be twice as large as WT cells, while the free tubulin concentration was similar in WT and DKO cells (personal communication N. Galjart).
This implies that, since MTs in a large cell must be longer in order to reach the cortex, there are fewer MTs in DKO cells compared to the WT cells. The distribution of these
124
few MTs needs an optimal adjustment, which might result in the creation of MT-free zones.
According to our findings, the network of dynamic (tyr) MTs in CLIPs-deficient cells seems to be much more susceptible to nocodazole-induced depolymerization than the MT network of WT cells. Nocodazole causes almost immediate release of +TIPs from the growing plus-end (Perez et al., 1999). Since nocodazole binds free-tubulin, the MT plus-end rescue during nocodazole-treatment is highly implausible. Our findings suggest that CLIPs provide „stabilization‟ of the dynamic MTs not only through the plus-end rescue, but also trough other mechanism(s). One possibility would be through CLIP interaction with CLIP associated proteins (CLASPs), +TIPs known to bind directly to CLIPs, and to stabilize MTs by linking them to cell cortex via LL5β–ELSK, complex often found localized in vicinity of CMAs, and membrane-associated protein PIP3 (Akhmanova et al., 2001). In CLIPs deficient cells, CLASP staining is reduced at the plus-ends, however, CLASP still localizes at the cell edge. The presence of CLASP at the cell edge of DKO cells might explain the observation that the stabile MT network in CLIPs DKO cells is intact, even though the CLASP is reduced at the MT plus-ends in these cells (Drabek et al., 2006). This indicates that either CLIPs, or at the plus-ends localized CLASP, might be involved in the protection of the plus-ends from the nocodazole induced-depolymerization in WT vs DKO cells.
CMAs have been shown to stabilize and capture the MT plus-ends during nocodazole- induced depolymerization (Kaverina et al., 1998). It is possible that this capture occurs through interaction of CLIP-170/115, or CLASP, with the CMA proteins. Since in DKO cells both CLIPs as well as CLASPs are not present at the plus-end, the interaction with CMAs is diminished and so is the stabilization. This might explain the high sensitivity of the MTs in DKO cells to nocodazole treatment. On the other hand, since our findings indicate that the integrin-based signaling in DKO cells is altered with respect to WT cells, it is possible that the CMAs in DKO cells are not able to capture and stabilize the MT plus-ends during the nocodazole-induced depolymerization.
We have also shown that the susceptibility of the „nocodazole-resistant‟, long-lived (stable) MTs to nocodazole is enhanced in the absence of CLIPs. It seems that the long- lived MTs, identifiable by presence of acetyl- and glu-α-tubulin, can be formed in absence of CLIPs, however, these MTs are not nocodazole-resistant any longer; quite the opposite- they are much more sensitive to nocodazole than the long-lived MTs in WT cells. This indicates that CLIPs enhance the stability of the already captured MTs. CLIPs have previously been shown to be involved in formation of cortical capturing complexes by direct binding to the actin binding protein IQGAP1 in a Cdc42- and Rac1-dependent fashion, (Fukata et al., 2002). However, Fukata and colleagues have described the period that EGFP-CLIP-170 remained at the complex-sites as rather short, indicating that CLIPs rather „deliver‟ the plus-end than permanently settle at the site of capture. CLIPs may thus „deliver‟ the right conformation of the plus-end to a specific site in the cortex, or
125
recruit factors involved in the stabilization. CLASPs have previously been found to capture MTs at the leading edge though a membrane-bound complex made of LL5β–
ELSK –PIP3 (Lansbergen et al., 2006). As mentioned above, CLIP-deficiency influences the levels of CLASPs at the plus-end, but not at the cell edge, which could result in establishment of less solid cortical capture compared to the WT cells. These mechanisms, however, should be investigated in more detail. CLIPs themselves might also play a role in stabilization of the plus-end by integrin-mediated adhesions in the cell cortex though activation of focal adhesion kinase (FAK) and mDia-Rho pathway (Palazzo et al., 2004).
Dynein/dynactin have been implicated in the cortical capture of MT plus-ends (Carminati and Stearns, 1997) and at the sites of adherens junctions (Lee A. Ligon, 2007). Since p150Glued, large subunit of dynactin implicated in „search and capture‟ behavior of MT plus-end, is targeted to the plus-ends by CLIP-170 (Lansbergen et al., 2004; Watson and Stephens, 2006b) and CLIPs-deficiency leads to formation of cytoplasmic aggregates of dynactin (Akhmanova et al., 2005), it is possible that loss of functional p150Glued abolishes the interaction of the MT tips with the cell cortex, which would affect the stabilization of MTs at this site.
It is highly plausible that the effects of CLIP deficiency on the stable MT array might affect regulation of cell polarization and directed cell migration. Moreover, a study of the effects of CLIP-170 siRNA-mediated knock-down in the context of wound-healing assay showed that CLIP-170 depletion disrupts cell polarization within the wound (Watson and Stephens, 2006a).
Role of CLIP-170 and CLIP-115 in regulation of cell-matrix adhesion
Total internal reflection fluorescence microscopy (TIRF) studies with live cells have shown that the dynamic MTs repeatedly target a subset of CMAs at the cell periphery and that this local targeting leads to the disassembly of these specialized structures (Kaverina et al., 1999; Krylyshkina et al., 2002). Growing, GFP-tagged CLIP-170 decorated, MTs plus-ends were shown to make an intimate contact with the CMAs, coming in their proximity as close as 50nm (Krylyshkina et al., 2003). MT plus-ends are known to target CMAs causing their dissociation, possibly through sequestering of Rho- activator GEF-H1 from the site of CMA (Chang et al., 2008), however other mechanisms are not excluded. Here we showed that CLIPs-deficient cells have enlarged, less intense and fewer CMAs per cell area, while the cells have enlarged cell surface and extremely thin MT-free zones. This particular phenotype could no be ascribed to perturbations in MT dynamic instability alone, since taxol and nocodazole, both inducers of loss of MT dynamicity, cause an enlargement of CMAs size accompanied by an increased CMAs intensity, Rho-associated formation of thick stress fibers and smaller, contractile cell phenotype (Bershadsky et al., 1996; Enomoto, 1996); unpublished data). It is thus
126
possible that CLIPs deficiency might negatively affect the Rho activity, so the typical non- contractile phenotype of DKO cells could be the result of low Rho. This, however, does not clarify why the CMAs in DKO cells are enlarged. Observed perturbations of CMA dynamics might also be due to a loss of interaction of CLIP-170's (and possibly CLIP- 115 itself) with IQGAP, which binds and recruits active Rac1. High level of active Rac1 is associated with CMA disassembly.
Differences in integrin engagement due to the lack of CMA-CLIPs crosstalk in DKO cells might affect the activities of various RhoGTPases, or the balance between Rac1/Cdc42 and Rho levels, in these cells, resulting in enhanced cell spreading.
Recent work from Lewkowicz et al suggested that CLIP-170, through its binding to mDia1, might regulate MTs - actin crosstalk downstream of integrins. It is thus also possible that CLIP deficiency results in impaired cross-talk between these two major cytoskeletal networks leading to perturbed spreading, MT stabilization and CMAs alteration (Lewkowicz et al., 2008).
MT-induced CMA disassembly is highly prominent in the rear of a crawling cell, even under high Rho levels, in order to establish tail retraction and enable the cell to move forward (Broussard et al., 2008). The crawling DKO cells have been observed to leave traces of adhesion proteins and cell parts behind them, indicating that disassembly of the CMAs in the rear of a crawling DKO cells is not performed in a controlled manner, or not entirely (unpublished observations). This CMAs disassembly involves dynamin-2-driven endocytosis. Moreover, the above described tail phenotype, together with extreme cell flattening, corresponds with that found in dynamin-2 (MT binding protein required for proper dynamic instability of MTs; (Tanabe and Takei, 2009). silencing experiments in 3T3 cells (Ezratty et al., 2005). These observations support the idea that CLIPS might affect the disassembly of the CMAs in the cell rear by regulating dynamin-2 activity and/or localization. Moreover, the observed effects on cell spreading and adhesion might also be a consequence of the sum of the effects of CLIPs-deficiency in DKO cells.
Signaling between the MT plus-ends and CMAs is bidirectional (Kaverina et al., 2002).
Moreover, the crosstalk between CMAs and MTs is situation dependant. In normal conditions paxillin rich CMAs can trigger MT catastrophes (Efimov et al., 2008), while CMAs also capture and stabilize MT plus-ends during the nocodazole-induced MT depolymerization (Kaverina et al., 1998). Focal adhesion kinase (FAK), an important component and regulator of cell motility, as well as dynamin-2 , are required for the disassembly of CMAs during the nocodazole-washout induced MT re-growth (Ezratty et al., 2005). FAK activation via integrin-dependant signaling was found to be required for the Rho-mDia mediated stabilization of MTs in the leading edge, indicating a role of the integrin-signaling in the polarization of the MT network and thus polarization of the cell (Palazzo and Gundersen, 2002; Palazzo et al., 2004). Therefore, it would be useful in the future to investigate the status of other proteins including FAK, mDia, Rho GTPases and integrin expression and activity in the context of CLIPs-deficiency.
