EB3 Regulates Microtubule Dynamics at the Cell Cortex and Is Required for Myoblast Elongation and Fusion

Summary During muscle differentiation, myoblasts elongate and fuse into syncytial myotubes [1]. An early event during this process is the remodeling of the microtubule cytoskeleton, involving disassembly of the centrosome and, crucially, the alignment of microtubules into a parallel array along the long axis of the cell [2–5]. To further our understanding on how microtubules support myogenic differentiation, we analyzed the role of EB1-related microtubule-plus-end-binding proteins. We demonstrate that EB3 [6] is specifically upregulated upon myogenic differentiation and that knockdown of EB3, but not that of EB1, prevents myoblast elongation and fusion into myotubes. EB3-depleted cells show disorganized microtubules and fail to stabilize polarized membrane protrusions. Using live-cell imaging, we show that EB3 is necessary for the regulation of microtubule dynamics and microtubule capture at the cell cortex. Expression of EB1/EB3 chimeras on an EB3-depletion background revealed that myoblast fusion depends on two specific amino acids in the calponin-like domain of EB3, whereas the interaction sites with Clip-170 and CLASPs are dispensable. Our results suggest that EB3-mediated microtubule regulation at the cell cortex is a crucial step during myogenic differentiation and might be a general mechanism in polarized cell elongation.


Cloning of Expression and shRNA Constructs
Depletion constructs were based on pSUPERneo.gfp (Oligoengine). The following RNAi target sequences were used: dEB3, GTGAGAA TCTGAGTCGCCA; dEB1, CAGACAAGGTCAAGAAACT; dEB1+EB3, TGGACATGCTCTTCCCTGG; and luciferase, CGTACGCGGAATACT TCGA, which was used as a control. For obtaining the pSuper variants expressing GFP-tubulin, the neomycin restriction cassette was replaced with the tubulin open reading frame (ORF) from pEGFP-tubulin (Clontech) with BsrGI and BamHI. Long and short forms of EB3 cDNA, as well as the EB1 coding region, were amplified from random primed cDNA, obtained from C2C12 cells that were differentiated for 3 days (see above), and subsequently cloned into AgeI and NotI of pEGFP-N1 (Clontech), thereby replacing the eGFP open reading frame (ORF). For the rescue constructs, five silent point mutations were introduced to the EB3 cDNA by a three-step PCR [S1], resulting in the sequence GTGAGAACCT CTCACGCCA. Single point mutations H29Q and Y32L were introduced to EB3 with the same strategy. The double FLAG tag (DYK-DADLDKDDDDK) was introduced in two parts by PCR to the upstream and downstream parts of EB1 and EB3 and ligated with a BglII site within the tag. This site was also used for domain swapping at in the chimeric constructs. Domain swapping at position 53 was done with a SapI site that is conserved in the nucleotide sequences of EB1 and EB3. For C-terminal deletion in EB3, a stop codon followed by a NotI site was introduced at the respective sites by PCR.

Cell Culture and Transfection Experiments
Mouse myoblasts C2C12 were cultured in DMEM (Invitrogen) supplemented with 10% fetal bovine serum (FBS) (Perbio), 2 mM L-glutamine, and antibiotics. For shRNA depletion during differentiation, cells were transfected with Lipofectamine Plus (Invitrogen) and shifted to differentiation medium (Dulbecco's modified Eagle's medium [DMEM] + 0.1% FBS + 2 mM L-glutamine + 5 mg/ml insulin + 5 mg/ml transferrin + antibiotics) when cells had grown to confluency. The differentiation medium was exchanged daily, and analysis of shRNA depleted cells was consistently performed 72-76 hr after transfection. To comply with that, cell density at the time of transfection was adjusted to the desired differentiation time. Because cell density is crucial for progression of myoblast fusion, caspase inhibitor Z-VAD-FMK (Calbiochem) was added after transfection to prevent reduction in cell density due to apoptosis. Skeletal myoblasts of H-2K b -tsA58 transgenic mice [S2] were grown in DMEM (Invitrogen) supplemented with 20% FBS (Perbio), 2% chicken embryo extract (SLI), 20 mg/ml interferon-g (Sigma), and antibiotics at 33 C and 6% CO 2 . For differentiation, cells were incubated in DMEM supplemented with 2% FBS, 0.2% chicken embryo extract, and antibiotics at 39 C and 5% CO 2 .
Live-Cell Imaging and Image Analysis Cells grown on acid-washed coverslips were mounted in a closed chamber on a heated stage insert (ALA Scientific Instruments) in medium supplemented with 25 mM HEPES. Imaging was done with a Deltavision system. Acquisition and image analysis were performed with Softworx (Applied Precision). Images were exported to Adobe Photoshop and Corel Draw for presentation. Statistical analysis was performed with Origin 6.0 (Microcal). The statistical significance of the observed differences was evaluated with ANOVA tests, and in all plots, standard deviation (SD) is indicated.

Cell Sorting and Western Analysis
Cells were detached from culture dishes with trypsin-ethylenediamine tetraacetic acid (EDTA). For cell sorting, cells were transferred to PBS + 10% FBS (1% for differentiated cells), and GFP-expressing cells were separated with a DakoCytomation MoFlo MLS flow cytometer. Cells were extracted in protein sample buffer and boiled for 5 min. Amounts of extracts equal to 15,000 or 20,000 GFPexpressing cells were loaded onto 10% or 12.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels. Primary antibodies were used as for immunofluorescence and detected with goat anti-mouse pAb HRP (Promega), donkey anti-rabbit IgG pAb horseradish peroxidase (HRP), and anti-rat pAb HRP (both Amersham), and ECL (Amersham) or SuperSignal West Dura Extended substrate (Pierce) followed.

Cell-Spreading Experiments
Coverslips were coated with 10 mg/ml fibronectin, collagen, laminin, or poly-D-lysine (all from Sigma) for 24 hr at room temperature (RT), washed with water, and allowed to dry. GFP-expressing C2C12 cells were enriched by FACS sorting (see above) 1 day after transfection with pSuper constructs and seeded at high density to allow shifting to differentiation conditions the day after sorting. Three days after transfection, cells were trypsinized, and an equal amount of cells was diluted in 1 ml DMEM + 1% FBS and allowed to settle on coverslips coated with different substrates for 70 min.

3D-Structure Analysis
Structural data for EB1 and EB3 were obtained from protein data bank at http://www.rcsb.org/pdb under accession numbers 1PA7 and 1WYO, respectively. To compare EB1 and EB3 structures, the data file for EB3 was modified with SYBYL 6.9.1., thereby removing the first seven amino acids that were a cloning artifact and the last 12 amino acids, which were not included in the EB1 structure and appeared to be highly flexible. Both structures were superimposed and surface representation calculated with SYBYL. Other structure representations were done with Protein Explorer 2.78 (http://www. proteinexplorer.org) on the modified file for 1WYO.     (D) Cell-length distribution in differentiated myoblasts after 40 hr, cultured on either glass coverslips only or coverslips coated with fibronectin, collagen, or laminin. Diamond-shaped boxes represent 25% and 75% at the tips and the median in the middle, and whiskers 10% and 90%. Mean values are indicated by a white dot.
(E) Myoblast-fusion efficiency was determined on cells cultured on fibronectin versus glass only, 3 days after transfection with hairpin RNA constructs and after 54 hr under differentiation conditions. Columns represent the percentage of cells containing two or more nuclei, averaged from two experiments, and error bars indicate the SD.

Figure S5. Rescue Constructs
Schematic representation of constructs used to rescue fusion defects of EB3 depletion. All constructs contain an internal double FLAG tag (abbreviated as 2FLAGi) that is inserted just before the alternative splice site in EB3 and at the respective site in EB1. The polka-dot pattern highlights sequence stretches that were derived from EB1. Figure S6. Tip Tracking and Localization of Rescue Constructs (A) Stills from time-lapse movies of GFP-tubulin and various rescue constructs fused to mCherry, demonstrating plus-end tracking ability for all constructs shown. The relative time is indicated in the upper-right corner, and arrowheads highlight the microtubule plus end in each frame. The scale bar represents 2 mm.

S4
(B) Immunolocalization of rescue constructs in undifferentiated C2C12 cells with FLAG M2 antibodies. Colocalization with microtubules is shown at higher magnification in the lower row. Arrowheads highlight signals at microtubule tips. Scale bars represent 10 mm in the upper row and 2 mm in the lower row.