|Home | About | Journals | Submit | Contact Us | Français|
Cytoplasmic linker protein (CLIP)-170 is a microtubule (MT) plus-end-tracking protein that regulates MT dynamics and links MT plus ends to different intracellular structures. We have shown previously that intramolecular association between the N and C termini results in autoinhibition of CLIP-170, thus altering its binding to MTs and the dynactin subunit p150Glued (J. Cell Biol. 2004: 166, 1003–1014). In this study, we demonstrate that conformational changes in CLIP-170 are regulated by phosphorylation that enhances the affinity between the N- and C-terminal domains. By using site-directed mutagenesis and phosphoproteomic analysis, we mapped the phosphorylation sites in the third serine-rich region of CLIP-170. A phosphorylation-deficient mutant of CLIP-170 displays an “open” conformation and a higher binding affinity for growing MT ends and p150Glued as compared with nonmutated protein, whereas a phosphomimetic mutant confined to the “folded back” conformation shows decreased MT association and does not interact with p150Glued. We conclude that phosphorylation regulates CLIP-170 conformational changes resulting in its autoinhibition.
The microtubule (MT) cytoskeleton plays an essential role in numerous fundamental processes, including cell division, migration, differentiation, morphogenesis, and intracellular trafficking. A variety of accessory factors control spatial and temporal organization of the MT cytoskeleton (for reviews, see Maccioni and Cambiazo, 1995 ; Howard and Hyman, 2003 , 2007 ). Among these factors is a class of proteins designated plus-end-tracking proteins (+TIPs) that specifically bind to growing MT plus ends, control MT dynamics, and transiently link MT plus ends to the actin cytoskeleton and other intracellular structures (for reviews, see Schuyler and Pellman, 2001 ; Akhmanova and Steinmetz, 2008 ).
Cytoplasmic linker protein (CLIP)-170, the first described +TIP (Perez et al., 1999 ), was discovered as a linker between endocytic vesicles and MTs (Rickard and Kreis, 1990 ; Pierre et al., 1992 ). Structural analysis of CLIP-170 shows that it is an elongated molecule dimerized through its central α-helical rod domain (Scheel et al., 1999 ; Lansbergen et al., 2004 ). The N-terminal region of CLIP-170 contains two MT-binding cytoskeleton-associated protein glycine-rich (CAP-Gly) domains flanked by three serine-rich stretches; the C-terminal region contains two zinc-knuckle motifs (Pierre et al., 1992 ). Structural studies revealed the presence of positively charged basic grooves in both CAP-Gly domains involved in tubulin binding; the second CAP-Gly domain furnished with a more basic groove directly binds the acidic tail of α-tubulin (Mishima et al., 2007 ). This CAP-Gly domain is thought to be necessary and sufficient for CLIP binding to growing MT ends in cells (Gupta et al., 2009 ). A second mammalian CLIP, CLIP-115, has a similar structural organization but lacks the C-terminal zinc-knuckle motifs (De Zeeuw et al., 1997 ). In contrast to the yeast homologues, which are delivered by kinesin to MT tips (Busch et al., 2004 ; Carvalho et al., 2004 ), CLIP-170 rapidly exchanges at the growing MT ends by recognizing composite end-binding protein 1/tubulin binding sites (Bieling et al., 2008 ; Dragestein et al., 2008 ).
CLIP family members are multifunctional proteins that are involved in distinct MT-dependent pathways. In general, they are “positive” regulators of MT growth (Brunner and Nurse, 2000 ; Komarova et al., 2002 ; Arnal et al., 2004 ; Carvalho et al., 2004 ). Both mammalian CLIPs directly interact with microtubule-stabilizing factors clip-associated proteins (Akhmanova et al., 2001 ; Mimori-Kiyosue et al., 2005 ) and with IQGAP1, a Rac1/Cdc-42 binding factor (Fukata et al., 2002 ), and may play a role in coordinating MT and actin network remodeling during cell migration.
The C-terminal zinc-knuckle motifs, which distinguish CLIP-170 from CLIP-115, directly bind to LIS1 and p150Glued, the large subunit of dynactin complex (Coquelle et al., 2002 ; Tai et al., 2002 ; Goodson et al., 2003 ; Lansbergen et al., 2004 ), suggesting a role of CLIP-170 in the dynein–dynactin pathway. During mitosis, an interaction with dynein–dynactin complex targets CLIP-170 to kinetochores (Dujardin et al., 1998 ; Wieland et al., 2004 ; Tanenbaum et al., 2006 ). The C-terminal part of CLIP-170 also binds to the formin homology 2 domain of mDia1; this interaction plays a role in the regulation of actin polymerization during phagocytosis (Lewkowicz et al., 2008 ).
In our previous study, we demonstrated that the intramolecular interaction between the C and N termini of CLIP-170 inhibits CLIP-170 association with the MT lattice and C-terminal–specific binding partners (Lansbergen et al., 2004 ). Detailed analysis suggested that the first zinc-knuckle of CLIP-170 and α-tubulin bind to overlapping but not identical sites covering the basic groove of the second CAP-Gly domain of CLIP-170 (Mishima et al., 2007 ).
CLIP-170 can be phosphorylated on multiple residues in cells and in vitro (Rickard and Kreis, 1991 ; Choi et al., 2002 ; Yang et al., 2009 ). Phosphorylation on Thr287 by cdc2 positively regulates binding of CLIP-170 to growing MT ends in the G2 phase and the G2/M transition, although it is not essential for CLIP-170 function during other phases of the cell cycle (Yang et al., 2009 ). Another study indicates that multiple kinases are involved in positive and negative control of CLIP-170 binding to MTs (Choi et al., 2002 ); however, the biological significance of differential phosphorylation of CLIP-170 is yet to be addressed. We hypothesized that phosphorylation of sites located in the serine-rich regions adjacent to the second CAP-Gly domain regulates the intramolecular interactions of CLIP-170. Using site-directed mutagenesis, we mapped critical sites of phosphorylation in the third serine-rich stretch and demonstrated their significance in regulating CLIP-170 conformational changes, as well as its interaction with MTs and dynactin. Furthermore, by using phosphoproteomic analysis, we determined that S309 and S311 of CLIP-170 are phosphorylated in cells and mapped S311 as a protein kinase A (PKA) phosphorylation site.
We used mouse monoclonal antibody (mAb) against hemagglutinin (HA; Covance Research Products, Princeton, NJ), green fluorescent protein (GFP; Invitrogen, Carlsbad, CA), α-tubulin (DM1A; Sigma-Aldrich, St. Louis, MO), p150Glued and dynamitin (p50) (BD Biosciences, San Jose, CA), rabbit anti-GFP (Invitrogen), and rat anti-α-tubulin mAb (a gift from J. V. Kilmartin; Laboratory of Molecular Biology, Cambridge, United Kingdom). Secondary antibodies were horseradish peroxidase-conjugated donkey anti-mouse and anti-rabbit (Jackson ImmunoResearch Laboratories, West Grove, PA), tetramethylrhodamine B isothiocyanate- and fluorescein isothiocyanate-conjugated donkey anti-mouse and anti-rabbit and Cy5-conjugated anti-rat (Jackson ImmunoResearch Laboratories).
HA-tagged expression construct was described by Komarova et al. (2002) , and GFP-CLIP-170-H1S (4-284) and GFP-CLIP-170-H2 (4-391) were produced by polymerase chain reaction (PCR). Yellow fluorescent protein (YFP)-CLIP-170 was described by Komarova et al. (2002) . Cyan fluorescent protein (CFP)-YFP tandem and YFP-CLIP-170-CFP were described by Lansbergen et al. (2004) . All alanine and glutamic acid mutations were introduced by using in vitro site-directed mutagenesis kit (Stratagene, La Jolla, CA). RNA interference (RNAi) expression vector for CLIP-170A was described by Lansbergen et al. (2004) . The RNAi cassette was inserted into the AseI site of pECFP-C1. Rescue constructs were prepared by a PCR-based strategy, by introducing triple silent substitutions (underlined) in the target site resulting in a sequence GGAGAAGCAACAACACATC.
For coIP experiments, COS-1 cells were lysed in an immunoprecipitation (IP) buffer (20 mM Tris-HCl, pH 7.5, 100 mM NaCl, 0.5% NP-40, and phosphatase inhibitor cocktail 1 [1:100; Sigma-Aldrich]) as described previously (Komarova et al., 2002 ). CoIPs and Western blots were performed as described by Hoogenraad et al. (2000) . For coIP, we used mouse anti-GFP mAb (Roche Applied Science, Indianapolis, IN) and protein G beads (Invitrogen). The bound proteins were dissolved in SDS-polyacrylamide gel electrophoresis (PAGE) sample buffer.
Chinese hamster ovary (CHO)-K1 cells were grown in F-10 medium, and COS-1 and NIH 3T3 cells wee grown in DMEM/F-12 (1:1 mixture) supplemented with 10% fetal bovine serum and antibiotics. Cells were transfected by FuGENE 6 (Applied Science).
Drug treatment was used in some experiments before coIP. COS-1 cells were incubated with 1 μM okadaic acid (OA) or 50 μM staurosporine (Calbiochem, San Diego, CA) for 1 h at 37°C or with 20 μM forskolin (Thermo Fisher Scientific, Waltham, MA) or 200 nM H-89 (Calbiochem) for 30 min at 37°C. Extracts were incubated with calf intestinal phosphatase (20 U; Roche Applied Science) for 1 h at 37°C.
NIH 3T3 and CHO-K1 cells were treated with 10 μg/ml Taxol (Sigma-Aldrich), 20 μM forskolin, and 100 nM or 1 μM OA for 1 h or with 40 μM forskolin, 200 nM H-89, and 10 μg/ml Taxol for 2 h.
Cell fixation, staining, and analysis were performed as described by Komarova et al. (2002) . In brief, cells were fixed in cold methanol (−20°C), postfixed with 3% formaldehyde, and permeabilized with 0.15% Triton X-100. The samples were imaged by fluorescence deconvolution microscopy using a DeltaVision microscope system (Applied Precision, Issaquah, WA). Images were prepared using Photoshop (Adobe Systems, Mountain View, CA). Linescan analysis, measurements of fluorescence intensity, and densitometry analysis of Western blots were performed using MetaMorph software (Molecular Devices, Sunnyvale, CA). The integrated fluorescence intensities above internal signal were measured within a rectangles covering p50 positively stained tips. Approximately 200 MTs were analyzed in 10–20 control or depleted cells for each experimental condition. Data handling was performed using SigmaPlot software (SPSS, Chicago, IL).
Analysis of fluorescence decay was performed as described by Komarova et al. (2005) and Dragestein et al. (2008) . Cells were observed at 36°C on a Diaphot 300 inverted microscope (Nikon, Tokyo, Japan) equipped with a Plan 100×, 1.25 numerical aperture objective using YFP filter set. Time-lapse series were acquired with stream acquisition mode. YFP intensity decay was analyzed on 16-bit depth images after subtraction of external background. Curve fitting was applied to determine the decay constant (kd). Photobleaching was negligible (kbl = 0.01 ± 0.004). Analysis of CLIP-170 fluorescent recovery was performed as described previously (Dragestein et al., 2008 ), with some modifications (Supplemental Data).
Taxol-stabilized MTs were prepared from bovine brain tubulin as described previously (Peloquin et al., 2005 ). High-speed supernatant extracts were obtained as described by Rickard et al., 1990 . An MT-pelleting assay was conducted as described previously (Choi et al., 2002 ). In brief, COS-1 cells were transfected with YFP-CLIP-170 and its mutants, washed twice with phosphate-buffered saline (PBS), and once with the PEM buffer [0.1 M piperazine-N,N′-bis(2-ethanesulfonic acid)-KOH, pH 6.8, 2 mM EGTA, 1 mM MgSO4 and 1 mM dithiothreitol; no exogenous ATP was added] at 4°C. The cells were scraped in PEM buffer (2 mM EGTA and 1 mM MgSO4) containing protease inhibitors (Roche Applied Science) at 0°C, Dounce-homogenized, and pelleted by centrifugation sequentially at 40,000 × g for 10 min and at 150,000 × g for 90 min at 4°C. The high-speed supernatants were incubated with 20 μM Taxol at 37°C for 30 min, and the endogenous MTs were depleted by centrifugation at 30,000 × g for 30 min at 20°C. For MT-pelleting assays, 30 μg of total protein of the high-speed supernatant was incubated with Taxol-stabilized MTs in PEM buffer containing 20 μM Taxol at 37°C for 15 min and then centrifuged at 30,000 × g for 30 min at 20°C. Resulting pellets were washed in the PEM buffer. Comparable amounts of supernatants and pellets were subjected to SDS-PAGE analysis. YFP-tagged proteins were detected by anti-GFP antibody on Western blot, and tubulin was stained by silver-staining kit (Bio-Rad Laboratories, Hercules, CA).
Fluorescence resonance energy transfer (FRET) measurements were performed using a fluorescence spectrometer (PC1 photon counting spectrofluorometer; ISS, Champaign, IL) as described previously (Lansbergen et al., 2004 ). COS cells were lysed in a buffer consisting of 20 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1% Triton X-100, 10% glycerol, and protease inhibitors (Roche Applied Science). No exogenous ATP was added. Excitation was performed at 425 nm (1 nm bandwidth), and emission spectra were collected in the range 450–600 nm (1 nm bandwidth).
A tandem fusion construct of rat CLIP-170 head (aa 4-354) with glutathione-S- transferase (GST) at N terminus and 6×His at C terminus were generated in pGEX-4T3 by PCR-based cloning. The proteins were expressed in Escherichia coli BL21 and purified by sequential affinity columns on nickel-nitrilotriacetic acid agarose (QIAGEN, Valencia, CA) and glutathione-Sepharose 4B (GE Healthcare, Little Chalfont, Buckinghamshire, United Kingdom) as described by Hegeman et al., 2004 . Purified protein (30 μg) was incubated in a manufacturer's reaction buffer and 1 mM ATP with PKA (2500 U; New England Biolabs, Ipswich, MA) for 3–18 h at 30°C. The proteins were denatured by 8 M urea, diluted with 50 mM ammonium bicarbonate, pH 7.5, and digested with 0.6 μg of trypsin for 14 h as described by Hegeman et al. (2004) .
Tryptic digests were solid phase extracted (using C18 SPEC-PLUS-PT pipette tips; Varian, Lake Forest, CA) and analyzed by microcapillary liquid chromatography-tandem mass spectrometry (MS/MS) using a Micromass Q-TOF2 hybrid quadrupole/orthogonal time of flight mass spectrometer (Waters, Milford, MA). Chromatography of peptides before mass spectral analysis was accomplished using C18 reverse phase high-performance liquid chromatography columns, made in-house, from which eluted species are directly microelectrosprayed. Columns were made using lengths of fused silica tubing (365 μm o.d., 100 μm i.d.) with pulled tips (1-μm orifice) that were packed to 12 cm with Zorbax Eclipse XDB-C18 (Agilent Technologies, Santa Clara, CA), 5-μm, 300Å-pore size media. An Agilent 1100 series HPLC delivered solvents A: 0.1% (vol/vol) formic acid in water, and B: 95% (vol/vol) acetonitrile, 0.1% (vol/vol) formic acid at either 1 μl/min, to load sample, or 150–200 nl/min, to elute peptides over a 180 min 10% (vol/vol) to 70% (vol/vol) B gradient. Voltage was applied upstream of the column through a platinum wire electrode into the fluid path via a PEEK T-junction. As peptides eluted from the HPLC-column/electrospray source, MS/MS spectra were collected from 400 to 2200 m/z; redundancy was limited by dynamic exclusion. Collision energy profiles were empirically predetermined for different peptide charge states. MS/MS data were converted to pkl file format by using Micromass Protein Lynx Global Server version 2.1.5 (Waters). Resulting pkl files were used to search a nonredundant (nrNCBI) Rattus norvegicus amino acid sequence database by using Mascot Search Engine (Matrix Science, London, United Kingdom), with methionine oxidation; glutamic and aspartic acid deamidation; carbamidomethylation of cysteines and tyrosine; and serine and threonine phosphorylation as variable modifications. Putative modifications identified by Mascot were confirmed using manual assignments of MS/MS spectra.
For analysis of CLIP-170 phosphorylation sites in cells, CLIP-170 was purified by using avidin-biotin affinity-based separation. The tryptic and chymotryptic digests of CLIP-170 were extracted with 50% acetonitrile (MeCN), 5% formic acid (FA) and used for mass spectra analysis (Supplemental Data). The probability of correct phosphorylation site localization was determined for every site in each peptide using the Ascore algorithm (Beausoleil et al., 2006 ).
The two CAP-Gly domains of CLIP-170 are flanked by three serine-rich stretches. In our previous study, we demonstrated that the N-terminal domain of CLIP-170 (H1; aa 4-309) lacking a part of the third serine-rich stretch binds to the C-terminal domain (Lansbergen et al., 2004 ). Here, we showed that deletion of the third serine-rich stretch [4-284 mutant, H1S(short); Figure 1A] increased the binding of N to C termini compared with H2 mutant, which contains all three stretches (4-391; Figure 1B), suggesting that the third serine-rich region might regulate the intramolecular association within CLIP-170. Inhibition of Ser/Thr-specific phosphatases with OA increased the binding of H2 to ΔH but had no effect on the H1S mutant (Figure 1B), indicating that the third serine-rich stretch might contain Ser/Thr residues phosphorylation of which induces intramolecular interactions within CLIP-170.
We sought to determine the critical residues involved in regulation of CLIP-170 intramolecular association. The third serine-rich stretch of CLIP-170 contains 17 serine residues, 16 of which are highly conserved among mammals (Figure 2A). By using site-directed mutagenesis, we generated phosphorylation-deficient (Ser→Ala) and phosphomimetic (Ser→Glu) mutants of YFP-CLIP-170 and assessed their binding to MT tips and to Taxol-stabilized MTs in cells (Table 1). Because intramolecular association within CLIP-170 significantly reduces its affinity for MTs (Lansbergen et al., 2004 ), we reasoned that mutants predominantly possessing an “open” conformation may display increased MT affinity. In line with this idea, a mutation in the first zinc-knuckle (Goodson et al., 2003 ; Lansbergen et al., 2004 ) that alters the intramolecular associations, induced binding of protein to MTs in Taxol-treated cells (unpublished data). The distribution of YFP-CLIP-170 mutants was assessed by live cell imaging (Supplemental Figure 1) and by immunofluorescence (Figure 2B). Because overexpression alters CLIP-170 behavior in cells, thus resulting in the longer labeled stretches at the MT ends (Goodson et al., 2003 ), we selected only the cells expressing relatively low levels of exogenous proteins. In particular, we determined the relationship between the length of CLIP-170–positive comets and the average intensities of YFP signals at the MT ends by using the wild-type (wt) protein (Supplemental Figure 1) for which we previously estimated the average length in the CHO-K1 stable cell line (Komarova et al., 2005 ). Based on this analysis, we then selected the cells expressing different CLIP-170 mutants showing the average intensities of YFP signals within the same range as for the wt protein (Supplemental Figure 1). For immunofluorescence staining, the expression levels were controlled by the YFP/tubulin integrated intensity ratio (Figure 2C). The colocalization of YFP-CLIP-170 mutants with MTs in Taxol-treated cells was determined by immunofluorescent staining and expressed by colocalization coefficient (Figure 2D). The S309A, S311A, S313A, S319A, and S320A mutants displayed both a statistically significant increase in the length of YFP-CLIP-170–positive stretches at the MT tips (Table 1 and Supplemental Figure 1) and the binding along the MTs in Taxol-treated cells (Figure 2B). Mutation of the same residues for glutamic acid led to shorter YFP-CLIP-170 structures at MT tips (Table 1). Similar to wild-type CLIP-170, phospho-mimetic mutants displayed no MT binding in Taxol-treated cells (Figure 2B; shown for S309–311E). The mutation of S347 to either Ala or Glu had no effect on CLIP-170 distribution at MT tips; however, both mutations induced moderate binding of CLIP-170 to Taxol-stabilized MTs in cells, suggesting that phosphorylation of this residue might not be directly involved in modulation of MT affinity (Table 1 and Figure 2, B–D).
Next, we determined whether mutation of the six identified residues regulates affinity of N to C termini. HA-CLIP-170-ΔH was coexpressed with either phosphomimetic or phosphorylation-deficient CLIP-170-H2 mutants; coexpression with H1 or nonmutated H2 was used as controls. Individual mutations of S309, S311, S319, and S320 to glutamic acid markedly increased the amount of coimmunoprecipitated ΔH, whereas mutation of S313 had a marginal effect (Figure 3A). Consistent with this observation, mutation of the same residues to alanine greatly diminished the amount of coimmunoprecipitated ΔH, except for S309 and S313, which were similar to nonmutated H2 (Figure 3A). However, mutation of both S309 and S311 to alanine decreased the amount of immunoprecipitated ΔH significantly more than S309 or S311 alone, suggesting that dephosphorylation of S309 and S311 might have an additive effect on establishing an open CLIP-170 conformation. Mutation of S311 led to the most significant increase in association of N and C termini, suggesting that phosphorylation of S311 may be critical for establishing the “folded back” conformation of CLIP-170. Furthermore, mutation of all five serine residues to alanine (designated as A-5 mutant) abolished the interaction between N and C termini, whereas substitution of the same residues for glutamic acid (E-5) greatly increased the amount of coimmunoprecipitated ΔH (Figure 3A). Overall, coimmunoprecipitation assay results were consistent with binding of these mutants to MT tips and Taxol-stabilized MTs.
To confirm that phosphorylation of identified residues regulates the conformation of CLIP-170, we performed FRET analysis on extracts of cells expressing a YFP-CLIP-170-CFP fusion (Lansbergen et al., 2004 ) in which all five serine residues were replaced with either alanine or glutamic acid. As a negative control, we used a mixture of extracts prepared from cells expressing either CFP or YFP; CFP-YFP tandem fusion was as a positive control. A mixture of cell extracts containing equimolar CFP and YFP proteins displayed no significant emission of the YFP acceptor after excitation of the CFP donor, giving a fluorescence ratio at 527/475 nm similar to that obtained for CFP alone (Figure 3, B and C). In contrast, the CFP-YFP tandem displayed a prominent peak of YFP emission (Figure 3B). A smaller but significant YFP-sensitized emission peak was detected for nonmutated YFP-CLIP-170-CFP and its E-5 mutant (Figure 3B). FRET expressed as the ratio of fluorescence obtained at 527 and 475 nm for wild-type CLIP-170 was consistent with published data (Lansbergen et al., 2004 ) and was indistinguishable from FRET observed for the E-5 mutant (Figure 3C). This means that a significant proportion of wild-type CLIP-170 might possess a folded back conformation in cells; alternatively, our FRET assay might be insensitive to the increase of CLIP-170 self-folding beyond a certain point due to its potentially transient character (Lansbergen et al., 2004 ). The YFP-CLIP-170A-5-CFP mutant produced no significant emission from the acceptor after excitation of the donor; its FRET signal was similar to the negative control (Figure 3D), indicating that phosphodeficient CLIP-170 possesses an open conformation. FRET results suggest that the phosphorylation state of the identified serine residues may determine conformational changes of CLIP-170.
We proposed previously that conformational changes of CLIP-170 control its interaction with the MT lattice (Lansbergen et al., 2004 ). Using an MT copelleting assay, we observed a dramatic difference in the binding affinities of phosphomimetic and phosphorylation-deficient CLIP-170 to the MT lattice. The CLIP-170A-5 mutant was completely sedimented with Taxol-stabilized MTs at a tubulin concentration of 0.5 μM (Figure 4, A and B), whereas <30% of the CLIP-170E-5 mutant was sedimented at 20 μM tubulin (Figure 4, A and B; unpublished data).
Next, we used live imaging to assess kinetics of CLIP-170A-5 and E-5 mutants at the growing MT ends. WT and both mutants bound to the MT growing ends, suggesting that dephosphorylation of these five residues is not required for MT tracking (Figure 4, C–D, and Table 1). The analysis of the relationship between intensities and the lengths of CLIP-170 structures suggested that phosphorylation might regulate affinity of CLIP-170 to growing MT ends (Figure 4E; n = 300 MTs for each group). The phosphorylation-deficient CLIP-170 mutant associated with longer MT segments (3.04 ± 0.6 μm; p < 0.0001), whereas the phosphomimetic mutant labeled only outer most MT segments (0.7 ± 0.15 μm; p < 0.0001).
The distinct length distribution of mutants predicted the difference in fluorescence decay at MT ends. Consistent with published results (Folker et al., 2005 ; Komarova et al., 2005 ; Dragestein et al., 2008 ), the decrease of YFP-CLIP-170 fluorescence at MT tips could be fitted to an exponential decay curve. The fluorescent decay constant (kd) for wild type was 0.45 ± 0.13 s−1 corresponding to an apparent half-life of 1.6 ± 0.5 s (n = 30) for the CLIP-170 binding sites (Figure 4F). The phosphorylation-deficient mutant displayed a considerably reduced decay constant of 0.24 ± 0.09 s−1 (half-life 3.23 ± 1.19 s; n = 37; Figure 4F), whereas the phosphomimic mutant had an increased decay constant of 1.05 ± 0.43 s−1 (half-life 0.76 ± 0.29 s; n = 40; Figure 4F). The expression of CLIP-170 mutants had no effect on the rate of MT growth. The instantaneous growth rates were 23.2 ± 4.9, 25.7 ± 5.1, and 24.6 ± 7.2 μm/min for wild-type, phosphomimetic, and phosphodeficient mutants, respectively (n = 60 MTs for each experimental condition), demonstrating that the distinct length distribution of CLIP-170 mutants was not due to the changes in MT growth rates. Therefore, we address whether CLIP-170 mutants might have a distinct affinity to the growing MT ends. The kinetics of CLIP-170 was determined previously by a fast fluorescence recovery after photobleaching (FRAP; Dragestein et al., 2008 ). Here, we used an improved protocol using total internal reflection fluorescence (TIRF) microscopy. The YFP-positive structures were detected by real-time TIRF imaging, and tips of the comets were bleached within circular spots. An apparent krecovery was 3.49 s−1 for wild-type YFP-CLIP-170 (Figure 4G and Supplemental Figure 3), which is close to the value obtained previously for GFP-CLIP-170 (Dragestein et al., 2008 ). Interestingly, the phosphodeficient mutant of CLIP-170 displayed a krecovery of 1.25 s−1 (Figure 4G), suggesting that it exchanges slower than wild-type CLIP-170. These data indicate that CLIP-170 is largely phosphorylated in cells and are in agreement with FRET analysis.
Previously, we proposed a regulatory role for CLIP-170 conformational changes in recruitment and release of the dynactin complex at MT plus ends (Lansbergen et al., 2004 ). Here, we tested whether CLIP-170 mutants display a difference in recruitment of dynactin to MT tips. We replaced endogenous CLIP-170 either with A-5 or with E-5 mutants and evaluated the distribution and amount of endogenous p50, a subunit of the dynactin complex, at MT tips. Expression of YFP-CLIP-170 was used as a control. The endogenous CLIP-170 was depleted with plasmid-based RNAi (Brummelkamp et al., 2002 ). The RNAi cassettes containing a previously described target region (Lansbergen et al., 2004 ) were inserted into pECFP-C1 vector for detection of transfected cells; YFP-CLIP-170 or its mutants in which we introduced three silent substitutions in the siRNA target region were used for rescue experiments. Cells expressing both the CFP reporter and YFP-tagged proteins were analyzed.
In nontransfected CHO-K1 cells, p50 displayed “dot”-like structures, with an average length of ~1 μm at MT tips (Figure 5, A and D, enlarged 1). Expression of YFP-CLIP-170 slightly increased p50 accumulation at the MT distal ends (Figure 5A, enlarged 2), resulting in ~1.5-μm-long segments (Figure 5E). Expression of the YFP-CLIP-170 A-5 mutant led to even longer ~4-μm segments of p50 (Figure 5, B and F, enlarged 2). The YFP-CLIP-170 E-5 mutant had the opposite effect (Figure 5C, enlarged 2), reducing both the length of p50 structures and the peak intensity (Figure 5G). To evaluate the relative amount of dynactin bound to the MT tips, we quantified the integrated fluorescence intensity of the p50 signal within rectangles circumscribing the entire positively stained tip (Figure 5H). We found that accumulation of p50 compared with surrounding nontransfected cells was slightly enhanced to 127 ± 36% by expression of YFP-CLIP-170 and slightly reduced to 74 ± 30% by expression of the phosphomimetic mutant. The most dramatic change was observed after expression of the phosphorylation-deficient mutant that increased the recruitment of p50 by nearly threefold (258 ± 67%). These data support the view that targeting of dynactin to MT tips might depend on CLIP-170 phosphorylation state.
CLIP-170 interacts directly with p150Glued, the large subunit of dynactin complex, and this interaction occurs via the second zinc knuckle of CLIP-170 and the CAP-Gly motif of p150Glued (Goodson et al., 2003 ; Lansbergen et al., 2004 ; Weisbrich et al., 2007 ). YFP-CLIP-170 and the phosphomimetic mutant weakly interacted with endogenous p150Glued, whereas the A-5 mutant strongly bound to p150Glued as demonstrated by coimmunoprecipitation study (Figure 5I). We conclude that phosphorylation might regulate CLIP-170 affinity for dynactin either directly or indirectly via modulation of CLIP-170 conformation.
Next, we tested whether CLIP-170 phosphorylation contributes to dynein-dependent transport of membrane organelles. We found no mislocalization of the Golgi apparatus in cells transiently coexpressing CFP-tagged β-1,4-galactocyltransferase (CFP-GalT) and YFP-tagged mutants in steady state (Supplemental Figure 2). However, expression of CLIP-170 phosphorylation-deficient mutant caused a significant delay in Golgi reassembly (up to 4 h) after nocodazole-induced scattering (Figure 5J and Supplemental Figure 2), whereas the majority of control cells or cells expressing YFP-CLIP-170 or E-5 mutant showed Golgi recovery within 30 min. Our results demonstrate that CLIP-170 dephosphorylation induces increased binding to dynactin and interferes with minus-end–directed membrane trafficking, and we suggest that CLIP-170 phosphorylation status might be important for efficient dynein function.
We used phosphoproteomic analysis to determine phosphorylation sites of CLIP-170 in cells. CLIP-170 was phosphorylated on multiple serine and threonine residues in cells pretreated with the phosphatase inhibitor calyculin A (Supplemental Figure 4). We confirmed phosphorylation of S309, S311, S314, and S347 in cells. The effect of these sites on CLIP-170 function was described above. In addition, we found several phosphorylation sites in the second serine-rich stretch, which are also adjacent to the second CAP-Gly domain (Supplemental Figure 4). To determine whether phosphorylation of these residues might regulate binding of CLIP-170 to MTs, we substituted S146, S149, T188, S192, S194, and S203 with alanine and determined the length distribution of these mutants at the MT tips in 3T3 fibroblasts (Table 1). We found that the S146A mutant displayed a significant increase in the comet length at the MT ends and could bind to Taxol-stabilized MTs in cells. This finding is in an agreement with coimmunoprecipitation data (Supplemental Figure 5), suggesting that mutation of S146 for Ala reduces the affinity between the N and C termini. In contrast to S146, mutation of Thr188 for A resulted in reduced binding to MT tips. Moderate binding to MTs in Taxol-treated cells also was observed for S194A and S203A, whereas the other mutations had no effect. These data suggest that S146 and T188 also might be involved in regulating CLIP-170 binding to MTs.
We aimed to determine the kinase responsible for phosphorylation of S311 because mutation of this residue had the most profound effect on CLIP-170 intramolecular association. A database search (Scansite [http://scansite.mit.edu/] and NetPhos 2.0 [http://www.cbs.dtu.dk/services/NetPhos/]) determined S311 as a site for 90-kDa ribosomal S6 kinase (RSK) and PKA. We focused on CLIP-170 regulation by PKA because it is known to phosphorylate numerous MT accessory proteins including p150Glued (Maccioni and Cambiazo, 1995 ; Gradin et al., 1998 ; Vaughan et al., 2002 ; Sengupta et al., 2006 ).
First, we determined whether PKA directly phosphorylates S311 in vitro. Bacterially expressed and purified rat CLIP-170 head (aa 4-354) was phosphorylated by PKA in vitro, and tryptic digests were analyzed by mass spectrometry. The mass of the aa 309-331 tryptic peptide from CLIP-170 treated with PKA differed from the expected mass of the untreated peptide by a phosphate group (80 Da). A collision-induced dissociation spectrum of this peptide showed a loss of H3PO4 (98 Da) for the y21 and y22 ions leaving a dehydroalanine at position 311. This confirmed the loss of phosphate at S311 upon fragmentation, suggesting that PKA phosphorylates S311 (Supplemental Figure 6).
Next, we tested whether PKA induces the interaction between N and C termini of CLIP-170 in cells. PKA activation with forskolin increased the binding of HA-CLIP-170-ΔH to GFP-CLIP-170-H2, whereas the PKA inhibitor H-89 abolished this interaction (Figure 6B). Consistent with this observation, pretreatment of CHO-K1 cells with H-89 resulted in binding of YFP-CLIP-170 to Taxol-stabilized MTs (Figure 6C). Although forskolin alone was not sufficient to induce redistribution of YFP-CLIP-170 at the MT tips (Figure 6C), forskolin in combination with 100 nM OA (a concentration 10 times lower than the IC50) significantly diminished the accumulation of YFP-CLIP-170, restricting its binding to the very distal part of MT ends (Figure 6C). This distribution was similar to that displayed by the phosphomimetic mutant. Treatment with OA at the IC50 concentration (1 μM) yielded a similar YFP-CLIP-170 distribution (Figure 6C), suggesting that fully phosphorylated CLIP-170 still binds to MT tips. To confirm the site-specificity of PKA phosphorylation in cells, we performed the above-mentioned treatments with CLIP-170 mutants (Supplemental Figure 7). YFP-CLIP-170 A309,311 or YFP-CLIP-170 E309,311 mutants were insensitive to activation or inhibition of PKA, confirming the role of PKA in S311 phosphorylation in cells. We conclude that PKA phosphorylates CLIP-170 on S311 in cells and controls its conformational changes.
Our results demonstrate that CLIP-170 on and off states are modulated by phosphorylation (Figure 7, model). The phosphorylation induces conformational changes in CLIP-170 by increasing the affinity of the N for C terminus, thus resulting in inhibition of protein function. These data are in line with our previous work suggesting that CLIP-170 open and folded back conformations represent active and inactive modes of the protein, respectively (Lansbergen et al., 2004 ).
Here, we establish the role of five conserved serine residues located in the third serine-rich region in regulating conformational changes as well as CLIP-170 interaction with MTs and p150Glued. Based on our analysis of phosphomimetic and phosphorylation-deficient mutations, which have opposite effects on CLIP-170 affinity for MTs and p150Glued, we extend the evidence for CLIP-170 autoinhibition.
We demonstrated that CLIP-170 can be phosphorylated in cells on multiple residues and that S311, a PKA phosphorylation site, is essential for establishing intramolecular associations. Phosphorylation of S309 might have an additive effect but is not sufficient per se to induce CLIP-170 autoinhibition. S313, S319, and S320 are potentially important sites for switching CLIP-170 off, although, we did not confirm their phosphorylation in cells. Whether the phosphorylation of these sites might occur during morphogenesis of highly specialized cells such as spermatids, in which CLIP-170 represents an immobile structural component of spermatid manchettes (Akhmanova et al., 2005 ), or at the onset of mitosis, when CLIP-170 predominantly binds to kinetochores and facilitates the formation of kinetochore–microtubule attachments (Tanenbaum et al., 2006 ), is to be determined.
Our data indicate that majority of endogenous CLIP-170 molecules in the cell possess a folded back conformation; therefore, phosphorylation might provide an essential mechanism for the regulation of steady-state “activity” of CLIP-170. This becomes critical in light of our data suggesting that phosphorylation might prevent nonproductive interaction of CLIP-170 with dynactin–dynein complex. The phosphorylation-deficient mutant of CLIP-170 displays strong binding affinity for p150Glued, resulting in a significant delay in Golgi reassembly after nocodazole-induced scattering. The loss of CLIP-170 function, however, has no effect on membrane trafficking in our model, and this finding is in agreement with several reports indicating that plus-end localization of p150Glued is not required for transport of diverse membrane organelles (Watson and Stephens, 2006 ; Kim et al., 2007 ; Lomakin et al., 2009 ).
CLIP-170 but not p150Glued is essential for efficient aggregation of pigment granules, melanosomes, in specialized cells such as Xenopus melanophores (Lomakin et al., 2009 ). It should be noted that aggregation of melanosomes is associated with inhibition of PKA in melatonin-treated cells (Reilein et al., 1998 ; Rodionov et al., 2003 ; Kashina and Rodionov, 2005 ). There is an attractive possibility that dephosphorylation of CLIP-170 might be critical for productive interaction of CLIP-170 with melanosomes at the MT tips, thus ensuring efficient redistribution of thousands of pigment granules in response to extracellular stimuli (Daniolos et al., 1990 ). In this respect, an accumulation of CLIP-170 at growing MTs also might be important. CLIP-170 rapidly exchanges at the MT tips with diffusion as the rate-limiting factor (Dragestein et al., 2008 ). The phosphorylation-deficient mutant of CLIP-170 displays much slower turnover at the growing MT ends, suggesting that phosphorylation also might regulate CLIP-170 affinity to the MT lattice (Rickard and Kreis, 1991 ).
Overall, our data demonstrate that on and off states of CLIP-170 are regulated by phosphorylation-induced autoinhibition. The complex pattern of CLIP-170 phosphorylation in cells might suggest that its function is finely modulated depending on the cellular context.
We thank Mark Ginsberg (University of California–San Diego, La Jolla, CA) for reagents and useful discussions. This work was supported by National Institutes of Health grant GM-25062 (to G.G.B.); Netherlands Organization for Scientific Research grants (to A. A. and N. G.); a Cancer Genomics Centre grant (to J.v.H.); and Presidential Program of Russian Academy of Sciences and RFBP grant 05-04-4915 (to E.S.N.).
This article was published online ahead of print in MBoC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E09-12-1036) on June 2, 2010.