tumor suppressor nf2/merlin is a microtubule stabilizer · tumor and stem cell biology tumor...

Tumor and Stem Cell Biology

Tumor Suppressor NF2/Merlin Is a Microtubule Stabilizer

Zlatko Smole1, Claudio R. Thoma1, Kathryn T. Applegate2, Maria Duda1, Katrin L. Gutbrodt1,Gaudenz Danuser2, and Wilhelm Krek1

AbstractCancer-associated mutations in oncogene products and tumor suppressors contributing to tumor

progression manifest themselves, at least in part, by deregulating microtubule-dependent cellular pro-cesses that play important roles in many cell biological pathways, including intracellular transport, cellarchitecture, and primary cilium and mitotic spindle organization. An essential characteristic of micro-tubules in the performance of these varied cell processes is their ability to continuously remodel, aphenomenon known as dynamic instability. It is therefore conceivable that part of the normal functionof certain cancer-causing genes is to regulate microtubule dynamic instability. Here, we report the resultsof a high-resolution live-cell image-based RNA interference screen targeting a collection of 70 humantumor suppressor genes to uncover cancer genes affecting microtubule dynamic instability. Extraction andcomputational analysis of microtubule dynamics from EB3-GFP time-lapse image sequences identifiedthe products of the tumor suppressor genes NF1 and NF2 as potent microtubule-stabilizing proteins.Further in-depth characterization of NF2 revealed that it binds to and stabilizes microtubules throughattenuation of tubulin turnover by lowering both rates of microtubule polymerization and depolymer-ization as well as by reducing the frequency of microtubule catastrophes. The latter function appears tobe mediated, in part, by inhibition of hydrolysis of tubulin-bound GTP on the growing microtubule plusend. Cancer Res; 74(1); 353–62. �2013 AACR.

IntroductionCell division, polarization, signaling, receptor recycling,

and invasion are among the morphologic processes affect-ed during cancer progression. All these processes dependat least partially on the microtubule cytoskeleton. Thatregulation of microtubule dynamics and cancer progres-sion are linked is evidenced by the fact that productsof human tumor suppressor genes, for example, adeno-matous polyposis coli (APC; ref. 1), AXIN (2), and vonHippel–Lindau (VHL; ref. 3) have microtubule-regulatingfunctions. Among the tumor suppressor gene products,pVHL is well characterized with respect to its quantitativeeffects on microtubule dynamic instability (3–7). It stabi-

lizes microtubules by dampening tubulin turnover duringboth growth and shrinkage phases, and by loweringthe rate of catastrophe and increasing the rate of rescue(6). The latter two functions are by virtue of pVHL'sproperties as an inhibitor of GTP hydrolysis resultingin longer microtubule-end protecting GTP-tubulin capsand denser GTP remnants along the microtubule lattice(6). The APC and AXIN gene products also work as med-iators of microtubule cytoskeleton formation, APC asa promoter of microtubule assembly (8), and AXIN as apromoter of de novo microtubule nucleation at the cen-trosome (2).

Given the broad role that microtubules play in cancer-related cellular processes, we sought to systematicallyidentify genes encoding tumor suppressors, whose func-tional inactivation is invariably associated with the devel-opment of various aspects of the malignant phenotype, onthe basis of their ability to affect microtubule dynamicinstability. We undertook a focused siRNA screen targeting70 tumor suppressors and assessed the functional impactof downregulation of these genes on microtubule dynamicinstability using a previously established live-cell imagingassay that relies on the tracking of a microtubule plus tip(þTIP) binding protein as a proxy for both microtubulegrowth and shrinkage dynamics (9). Among the hits of thissiRNA screen, the neurofibromatosis 1 and 2 (NF1 and NF2)genes emerged with high confidence as robust stabilizers ofmicrotubule dynamics. The choice of NF2 for further ana-lysis was rooted in the fact that NF2 has been found

Authors' Affiliations: 1Institute of Molecular Health Sciences, ETH Zurich,Zurich, Switzerland; and 2Department of Cell Biology, The ScrippsResearch Institute, La Jolla, California

Z. Smole, C.R. Thoma, K.T. Applegate, andM. Duda contributed equally tothis work.

Current address for K.L. Gutbrodt: Institute of Pharmaceutical Sciences,ETH Zurich, Wolfgang-Pauli-Str. 10, 8093 Z€urich, Switzerland; and currentaddress for G. Danuser: Department of Cell Biology, Harvard MedicalSchool, 240 Longwood Ave, Boston, MA 02115.

Corresponding Authors: Wilhelm Krek, ETH Zurich, Schafmattstrasse22, Zurich 8093, Switzerland. Phone: 41-44-633-3447; Fax: 41-44-633-1357; E-mail: [email protected]; and Gaudenz Danuser, E-mail:[email protected].

doi: 10.1158/0008-5472.CAN-13-1334

�2013 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org 353

on September 25, 2020. © 2014 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 26, 2013; DOI: 10.1158/0008-5472.CAN-13-1334

http://cancerres.aacrjournals.org/

10 200

A (i)

0 50 100 150

100

300

500

700

Frame number

Co

me

t n

um

be

r

500

600

700

800

900

Co

me

t n

um

be

r

B (i)

(ii)

100 200 300 400 500 600 700 800 900 1,000 1,1006

8

10

12

14

16

18

20

22

24

Movie index

Me

dia

n M

T g

row

th s

pe

ed

(μm

min

–1) 1 2 3 4 5 6 7 8 9 10 1112 13 1415 16 17 18 19 20 21 22 23 24 25

D

7/70

14/70

49/70

0, Excellent

1, Good

CConsistency between

RNAi oligos

AXIN

BARD1BCL2

BRCA1

LATS1

LATS2

NF1NF2

P63TSC1

TSC2VHL

20

15

10

5

0

5

10

15

EAXIN

BARD1

BCL2BRCA1

LATS1

LATS2

NF1NF2

P63 TSC1TSC2

VHL

–lo

g(P

valu

e

(|ta

rget–

Alls

tar|

))

–lo

g(P

valu

e

(|ta

rget–

Negctr

l|))

Growth speedGrowth lifetime

Pause frequencyShrinkage speed

Reference P value levels: |Allstar–Negctrl| same day

CFLAR

CFLAR

Days

(ii) (iii)

≥2, Poor

Figure 1. Screening of tumor suppressors with microtubule-regulatory functions. A, extraction of microtubule (MT) dynamics by automated tracking of EB3comets; (i) raw image of EB3-GFP tagged comets associated with growing microtubule ends (contrast inverted; bar, 10 mm); (ii) microtubule growthtracks color coded by speed (mm*min�1); (iii) initiation points of forward gaps in growth trajectories (yellow, inferred pause events) and backward gaps (red,inferred, rescued shrinkage events). B, elimination of low-quality movies based on stability in the comet number; (i) acceptable (inlier) movies contain aconstant number of detectable comets over the entire duration; (ii) rejected (outlier) movies display a significant drift in the comet number. C, summary ofinteroligo heterogeneity quantified by the number of parameters (out of 4), where P < 0.01 between any oligo pair (3 oligos per target). D, variationof median microtubule growth speed across all 1,143 movies of the primary screen. Vertical lines, days of filming. Orange and red blocks, cellsthawed from one master stock. Blue dots, Allstar control movies; black dots, Negctrl control movies; green, red, cyan, magenta, yellow circles, moviesper color belonging to 1 of 70 siRNA-targeted genes; black crosses overlaid to dots/circles, eliminated movies due to low image quality (see B); redcrosses overlaid to dots/circles, eliminated movies due to group inconsistency (see Materials and Methods). (Continued on the following page.)

Smole et al.

Cancer Res; 74(1) January 1, 2014 Cancer Research354




mutated in human clear-cell renal cell carcinoma (ccRCC)in a manner that appears to be mutually exclusive with VHL(10) and that pVHL is known to quantitatively affectmultiple parameters of microtubule dynamic instability(6), making one wonder about possible parallels in theirrole as tumor suppressors in certain tumor cell biologiccontexts.NF2 encodes NF2/Merlin, a ubiquitously expressed mem-

ber of the ERM (Ezrin-Radixin-Moesin) protein family.Mutations in the NF2 gene cause neurofibromatosis type-2, a cancer syndrome characterized by the development oftumors of the nervous system (11). NF2 tumor suppressor isalso dysfunctional in diverse nonnervous malignancies suchas mesotheliomas (12), colorectal cancer (13), melanomas(14), and ccRCC (10). NF2 exerts its tumor suppressionactivity through multiple distinct pathways including acti-vation of the Hippo pathway (15), induction of a growthsuppressive program through inhibition of CRL4 E3 ubiqui-tin ligase (16), establishment of epithelial adhesion andpolarity via Par3 and aPKC (17), insurance of proper cen-trosome positioning and spindle orientation (18), and main-tenance of Rac-dependent anterograde microtubule-basedtrafficking of exocytic vesicles (19). Finally, the Drosophilahomolog of NF2 has been detected in cytoplasmic particlesthat move along microtubules (20) and in vitro associationof recombinant NF2 with polymerized microtubules hasbeen observed (21).Our screen and detailed analysis of NF2 function now

establishes the NF2 gene product as a potent stabilizer ofmicrotubules both by attenuation of tubulin turnover throughreduction of the rates of microtubule polymerization anddepolymerization and by reducing the frequency of microtu-bule catastrophes.

Materials and MethodsBiochemical and cell biological methodsBiochemical and cell biologic materials and methods are

described in the Supplementary Information.

Live cell imagingMovies were acquired with an Olympus IX70 Delta Vision

Spectris microscope (Applied Precision) at 37�C using a�601.4NA DIC Oil PlanApo objective: exposure time 200 ms,frame rate 1.25 Hz. For the screen, four movies wereacquired per well, 150 frames per movie. A total of 1,143movies were acquired for the 70 genes plus controls over aspan of 25 days. For the validation screen, an additional220 movies were acquired for 12 selected genes plus con-trols over a span of 4 days (Supplementary Fig. S1B). For

NF2 validation, the acquisition protocols were the sameexcept that we acquired >8 movies per well. For nocodazoleexperiments, a first movie was acquired in 300 mL of normalgrowth medium and then 300 mL of growth medium con-taining 80 nmol/L nocodazole was added (final concentra-tion 40 nmol/L). Video sequences were acquired before(0) and 20, 120, and 360 seconds after nocodazole additionas described.

Analysis of microtubule dynamics from EB3-GFP cometdata and tracking data validation

An analysis of Microtubule Dynamics from EB3-GFP cometdata tracking and data validation methods is described in theSupplementary Information.

ResultsNF1 and NF2 emerge as high confidence hits among 70tumor suppressor genes screened for effects onmicrotubule dynamics

To identify tumor suppressors with previously unappreci-ated roles in microtubule regulation, we designed a live-cellimaging screen for effects on the dynamic behavior of micro-tubules upon siRNA-mediated depletion. We employed RPE-1cells stably expressing a GFP-fusion to the þTIP trackingprotein EB3 (Fig. 1A, i). The screen was conducted in 8-wellchambers. Per target gene, three siRNA oligos were transfectedindependently in three wells. In every chamber, we furthertransfected RPE-1 cells in separate wells with two siRNAcontrol oligos: Allstar and Negctrl (Supplementary Data file,"siRNA information"). Thus, per chamber, two genes weretargeted with corresponding negative controls. A total of1,143 EB3-GFP movies were acquired over 25 days.

Image time-lapse sequences were then analyzed using theplusTipTracker software (22). The package offers modulesfor detection and tracking of EB3-GFP comets (Fig. 1A, ii)and for inference of pauses and shrinkage events, in whichthe comets temporarily disappear (Fig. 1A, iii). Thus, weobtained in a fully automated and unbiased fashion statisticsof microtubule dynamic behaviors that included theirgrowth, pausing, and shrinkage characteristics. The softwaregenerates >80 metrics to quantify the full spectrum ofmicrotubule dynamics. Fifty of those are documented foreach movie (Supplementary Data file, "Metrics" and "MovieDocumentation").

Before analyzing putative effects of siRNA oligos on theseparameters, we performed a thorough assessment of thedata quality (see Materials and Methods). First, we excludedmovies (12 in total) in which a high variation of detectedcomets over time suggested inconsistent image acquisition,

(Continued.) E, evaluation of effects on 4 analyzed parameters: growth speed (red), growth phase lifetime (magenta), pause frequency (blue), shrinkagespeed (black); for a detailed description of all parameters and effects see Supplementary Data file. Shown are the �log P values of a permutation t testevaluating the difference between microtubule dynamics in movies of cells transfected with targeting siRNA and microtubule dynamics in movies ofcells transfected with nontargeting Allstar (y-axis north) or nontargeting Negctrl (y-axis south) siRNAs. Dashed lines, the �log P values testingthe difference in the same parameters between movies of cells transfected with Allstar and movies of cells transfected with Negctrl siRNA. Asterisks,targets with parameters where the difference between target effect and control data is more significant than the difference between the two controls.Gray bars, 12 targets selected for a validation screen.

NF2 Regulates Microtubule Dynamic Instability

www.aacrjournals.org Cancer Res; 74(1) January 1, 2014 355




for example, due to focus drift or photobleaching (Fig. 1Band Supplementary Movie S1). Second, we tested the con-sistency of effects between the three oligos targeting thesame gene by pairwise comparison of oligos in four metrics(Supplementary Data file, "Oligo Variation"). Quite remark-ably, 49 of the 70 genes showed no difference in any of the 12pairwise tests (Fig. 1C). This high level of consistencysuggests two possibilities: either all three oligos affectmicrotubule dynamics in an identical fashion, or they donot cause any detectable change in microtubule dynamics.Another 20% of the genes (14/70) showed a difference in onlyone of the 12 tests, whereas the final 10% (7/70) showed adifference in two or more. None of the genes had a single,unambiguously deviant oligo across the four metrics, whichwould have indicated transfection or targeting failure. Thus,we felt confident in pooling the movies for one target into asingle group of typically 12 movies. Third, we tested theintra-group consistency among these movies and eliminatedan additional 42 movies as outliers.

Variation analysis of individual metrics across all moviesdisplayed significant heterogeneity in the data. This is anotorious problem of live-cell screens where even smallchanges in cell culture conditions penetrate into the readouts(23). In our case, it was obvious that the cell populationsystematically drifted in particularmetrics,most likely coupledto the increasing number of passages (Fig. 1D). To identifytumor suppressor genes with significant contributions to theregulation ofmicrotubule dynamics, it was therefore necessaryto compare the validated set of movies for a target gene to thecontrol siRNA movies filmed on the same day. We performedtwo-tailed permutation t tests between these movie groups foreach of the 50 metrics (Supplementary Data file, "TargetAnalysis"). Figure 1E shows the P values of these tests for thefour metrics: growth speed, growth lifetime, pause frequency,and shrinkage speed (solid lines). An analogous test wasperformed between the Allstar and Negctrl movies for eachday (Fig. 1E, dashed lines; Supplementary Data file, "ControlData Analysis"). This latter analysis revealed that two nom-inally nontargeting oligos can produce differences with highstatistical significance in many metrics, some of which werecomparable to the differences between the effects of aparticular target knockdown and Allstar. To avoid a highrate of false positives, we therefore reasoned that valid hitsshould produce an effect upon knockdown that is moresignificant than the difference between the two controls(Allstar and Negctrl) on the same day. For instance in Fig. 1ENF1 fulfills this criterion for the growth speed, which is alsodocumented in Supplementary Fig. S1A, where the siRNAcondition shows an increase in growth velocity that isgreater than the difference in growth velocity measuredbetween Allstar and Negctrl on the same day. The sameapplies to the shrinkage speed for NF1 and NF2 (the latter isillustrated in Supplementary Fig. S1A). In contrast, theCASP8 and FADD-like apoptosis regulator (CFLAR) presentsan example where the difference between target knockdownand Allstar is highly significant (P < 10�12). However, nosignificant difference is observed between the knockdownand Negctrl.

In view of these differences between control conditions,we identified hits by testing for each target which of the 50metrics had P values <0.05 both in the target/Allstar and inthe target/Negctrl comparisons and smaller than the P valuein the associated Allstar/Negctrl test (Supplementary Datafile, "Target Analysis"). The target genes were then ranked byhow many metrics were significant according to this crite-rion (Supplementary Data file, "Target Ranking"). Becausesome metrics are redundant, we reduced the set for the finalranking to 23 nearly orthogonal parameters. The highestranked genes with more than 50% of the metrics fulfillingthe criterion for an affected behavior were NF1, TGFBR2,SCRIB, BARD1, SHH, and SMARCB1. Other high-rankinggenes with more than 30% of the metrics changed includedLATS1, LATS2, NF2, P63, and BRCA1.

Surprisingly, two other genes known to affect microtu-bule dynamics, VHL and AXIN, did not rank particularlyhigh in our screen. To understand this result, we examinedthe control movies on the corresponding days (days 1 and10, respectively) and noticed a particularly wide discrep-ancy between the Allstar and Negctrl populations in meangrowth speed (Supplementary Fig. S1B) and in several othermetrics (Supplementary Data file, "Control Data Analysis").Therefore, for validation, we decided to repeat the screenfor 12 targets of high interest (highlighted in Fig. 1E,without CFLAR). Ten genes were chosen on the basis oftheir oligo consistency (or lack thereof) and target ranking,plus VHL and AXIN. The Supplementary Data file, "Com-parison First-Second" summarizes the results for each ofthe 12 selected target genes in the primary and secondaryvalidation screen. Consistency between the two screens wasperformed again under the requirement that siRNA effectsin both screens significantly changed a metric relative toNegctrl and Allstar, and that the changes were greater thanthe differences between the two control conditions. Tominimize the risk of false positives, we increased thestringency of these tests to a 99% confidence level, i.e., alltwo-sided permutation t tests had to return a P value <0.01to pass.

NF1 and NF2 were the only two genes to show such highconsistency between primary and secondary validation screen.Of these two tumor suppressor genes, NF2 shares notablefeatures with VHL with regards to tumor suppression in thecontext of kidney cancer. Whole-exome sequencing revealedthat both genes are mutated in human ccRCC (10). It isconceivable that loss of function in either VHL or NF2 con-tributes to ccRCC development through the dysregulation ofsimilar cellular processes, including microtubule regulation.Hence, we focused next on further analysis of the effect of NF2on microtubule dynamics.

Validation of NF2 as a microtubule stabilizing proteinWe engineered two additional sets of cell lines in which

NF2 protein levels were manipulated and EB3-GFP cometstracked. First, we depleted NF2 in EB3-GFP–expressingRPE-1 cells by stably expressing shRNAs targeting NF2.Two distinct shRNAs, shNF2.1 and shNF2.2, reduced NF2protein levels to similar degree; a nonsilencing scrambled

Smole et al.





shRNA control (ns) did not demonstrate such effects(Fig. 2A). Second, NF2-negative mesothelioma cancer cellsalready stably expressing EB3-GFP (referred to as Meso33-EB3-GFP) were engineered to stably express wild-type NF2or vector control. NF2 production in these cells wasconfirmed by immunoblotting (Fig. 2B).Using the plusTipTracker package, we then analyzed

microtubule dynamics in live-cell images (SupplementaryData file, "NF2 Validation Experiments"). As shown in Fig.2C and D, EB3-GFP tracking in these cells supports thefinding from the screen that NF2 is a strong microtubulestabilizer. Both NF2 depletion in RPE-1 cells and reconsti-tution of NF2 gene expression in Meso33-EB3-GFP cellsshow highly consistent, complementary effects on micro-tubule dynamics. First, NF2 decreases both microtubulepolymerization and depolymerization rates as evidencedby the changes in "growth and shrinkage speeds," respec-tively, indicating that it influences the net turnover oftubulin dimers at the microtubule plus end in both phases.Second, NF2 increases the time of EB3-GFP presence at the

microtubule plus end as expressed by "growth lifetime"metric, suggesting that NF2 protects the growing plus endof microtubules and therefore reduces catastrophe rates asdefined by 1/tgrowth. The net distance of microtubule growth(growth length) shows no significant changes, indicatingthat the prolonged growth phases are compensated bylower growth rates. Third, NF2 influences microtubuleshrinkage not only by reducing shrinkage speeds, but alsonet distance of shrinkage (shrinkage length). Since the"shrinkage lifetime" (tshrink) does not show consistentresults and does not change significantly, the shorter netdistance of shrinkage in presence of NF2 is primarilyassociated with the lower disassembly rate. Fourth,although NF2 dampens growth and shrinkage speeds andprolongs growth times and therefore has a net stabilizingeffect on dynamic microtubules, it significantly reduces"pause lifetime" in which pause is measured as the timebetween loss of an EB3-GFP signal and regain of growth,resulting in a forward gap closure (22). This observationsuggests that NF2 might influence reinitiation of

Growth speed

22

18

0

26

******

NF2

Tub 50

70

B

A

18

16

22

20

***

0

D

******7

6

0

8

Growth lifetime

s

9

*** ***

Shrinkage speed40

30

35

0

25

1.4

1.2

Shrinkage length

0

0.8

1.0

Growth speed

50

30

40

0

60***

Shrinkage speed

20

1.2

0

1.6

Shrinkage length

µm

0.8

C

14

2.0

1.80

2.2

Growth length

µm

2.0

1.50

2.5

Shrinkage lifetime

1.8

1.7

1.6

0

Growth length

µm

1.5

0

2.0

1.8

Shrinkage lifetime

s

1.6

4.6

4.2

0

4.4

Pause lifetime

** **

*

4.6

4.2

0

4.8

4.4

Pause lifetime

s

4.0

***

***

**

ns

shNF21 2

nsshNF2

1 2 nsshNF2

1 2 nsshNF2

1 2

nsshNF2

1 2 nsshNF2

1 2 nsshNF2

1 2 nsshNF2

1 2

ctrl NF2 ctrl NF2

ctrl NF2 ctrl NF2 ctrl NF2 ctrl NF2

ctrl NF2

NF2

Tub 50

70

0

7.0

6.0

8.0**

Growth lifetime

s5.0

ctrl NF2

μm m

in-1

μm m

in-1

μm m

in-1

µmss

μm m

in-1

Figure 2. NF2 stabilizes microtubules in vivo. A, Western blot showing depletion of NF2 in RPE-1 cells using lentiviral expression (LKO-1) of twodifferent shRNAs against NF2 (shNF2.1 and .2) and expression of a nonsilencing shRNA as control (ns); anti-NF2 (top) and antitubulin (bottom). B,Western blot analysis showing stable reexpression (LKO-1-CMV) of NF2 using lentiviral expression in NF2-deprived mesothelial cancer cells (Meso33),expression of the empty vector serves as control (ctrl); anti-NF2 (top), antitubulin (bottom). C, microtubule dynamic parameters in RPE-1 cellsmeasured by EB3-GFP comet tracking. Box and whisker plots representing tracks for growth speed, growth lifetime, growth length, pauselifetime, shrinkage speed, shrinkage lifetime and shrinkage length are shown from three independent experiments: per condition over 60,000 tracks (N)were measured from a minimum of 8 movies. Boxes indicate 25, 50 (median), and 75% quantiles; whiskers extend to 1.5� the interquartilerange; red dots, outliers beyond this range. Notched boxes, uncertainty of the median. Boxes whose notches do not overlap indicate that the mediansof the two clusters differ at the 5% significance level (��, P < 0.0001; ��, P < 0.001; �, P < 0.05; Student t test). D, microtubule dynamic parameters inMeso33 cells as described in C.






microtubule plus end polymerization by faster recruitmentof þTIPs including EB3-GFP.

NF2 colocalizes and interactswith tubulin andacetylatedtubulin

Given the prominent effects of NF2 on microtubule dynam-ics, we next sought for cell biologic and biochemical evidencefor interactions of NF2 with microtubules in vivo. As shownby indirect immunofluorescence microscopy, endogenousNF2 associates with the microtubule network in RPE-1 cells(Fig. 3A). Immunofluorescence stainings for NF2 either inthe presence of bacterially purified glutathione-S-transferase(GST)-NF2 fusion proteins or in cells that had been pre-viously depleted for NF2 by shNF2 expression revealed thatthe staining and thus the detected colocalization of NF2with microtubules is specific (Supplementary Fig. S2A–S2D).

To corroborate this finding, we immunoprecipitated tubulinfrom RPE-1 cells ectopically producing NF2 and observedspecific coimmunoprecipitation of NF2 (Fig. 3B). Also, endog-enous NF2 coimmunoprecipitated specifically with tubulinantibody (Fig. 3C). These results provide biochemical evidencethat NF2 and tubulin interact in vivo.

The notion, that NF2 mainly colocalizes to microtubulebundles (Fig. 3A, zoom) and not to all microtubules presentin the cell let us wonder whether NF2 interacts with modifiedforms of tubulin. Indeed, immunofluorescence stainings showthat NF2 colocalizes with the acetylated form of tubulin (Fig.

3D). Consistent with this, we also found that ectopicallyproduced NF2 coimmunoprecipitated with an acetyl-tubu-lin–specific antibody in RPE-1 cells (Fig. 3E). Since tubulinmodifications are indicative for a differential turnover andstability (reviewed in ref. 24), we additionally investigatedthe state of tubulin tyrosination in cells depleted for NF2.Neither in Meso-33 cells reconstituted with wild-type NF2,nor in RPE-1 cells depleted for NF2, we detected differencesin tubulin tyrosination state as assessed by Western blottingor immunofluorescence microscopy (Supplementary Fig. S3).

NF2 is a microtubule protector and stabilizer byincreasing GTP-tubulin frequency along the lattice andon the plus end

To further characterize NF2's microtubule regulatory func-tion, we asked whether NF2 has the capability of protectingmicrotubules from the depolymerizing activity of nocodazole(25). We first determined in an initial experiment with low(1 mmol/L) and high (10 mmol/L) concentrations of nocoda-zole, the amount of nocodazole needed in RPE-1 cells depletedfor NF2 to obtain a measurable effect (Supplementary Fig.S4A). While NF2-depleted cells displayed statistically signifi-cant tendency of higher sensitivity towards nocodazole treat-ment already at 1 mmol/L, robust effects were only found at10 mmol/L concentration. As illustrated in Fig. 4A, at thisconcentration the number of cells displaying stable microtu-bules was significantly reduced for NF2-depleted RPE-1 cells

Anti-acetyl-tubulin Anti-NF2 Merge

A

Anti-NF2Antitubulin Merge

TubIgG

Em

pty

2F

NEm

pty

2F

N

50

70NF2

Tub

pLKO

CMV

Em

pty

2F

N

Input

B

C

Acetyl

-tub *

IgG

Em

pty

2F

NEm

pty

2F

N

50

70NF2

pLKO

CMV

Em

pty

2F

N

Input

Acetyl

-tub

TubIgGInput

NF2

Tub 50

70

E

Zoom

D

Zoom

Figure 3. NF2 colocalizes to andinteracts with tubulin andacetylated tubulin. A,immunofluorescence costaining inRPE-1 cells using antibodiesagainst tubulin (red) and NF2(green). Zoomed areas shownbelow are indicated by a dashedsquare. Bar, 10 mm. B, Whole-cellextracts of RPE-1 cells transientlytransfected with pLKO1-CMV NF2or the empty vector as control(input) were subjected toimmunoprecipitation with control(IgG) and antitubulin (Tub)antibodies, separated by SDS–PAGE, and immunoblotted withanti-NF2 and antitubulin antibodies.C, immunoprecipitation of NF2from whole-cell extracts of RPE-1cells using antibodies as describedin B. D, immunofluorescencecostainings in RPE-1 cells usingantibodies against acetyl-tubulin(red) and NF2 (green). Zoomedareas shown below are indicated bya dashed square. Bar, 10 mm. E,immunoprecipitation as describedin B with control (IgG) and anti-acetylated-tubulin antibodies wereprocessed for immunoblotting withanti-NF2 and anti acetylated-tubulin antibodies. �, cross-reactionof the IgG heavy chain in theacetylated-tubulin blot.

Smole et al.





compared with control cells. Conversely, reexpression of NF2in Meso-33 cells significantly increased the number of positivecells with stable microtubules after nocodazole treatment

(Fig. 4B). Quantification of these data revealed a strong pro-tection function of NF2 towards nocodazole-induced micro-tubule depolymerization (Fig. 4C and D).

Un

tre

ate

dN

oco

da

zo

le

Ce

lls w

ith

in

tact

MT

(p

erc

en

tag

e)

0

20

40

60

80

*** ***100

0.9

Re

l. g

row

th s

pe

ed

Re

l. s

hrin

ka

ge

sp

ee

dctrl NF2

Meso33

RPE-1 Meso33 ***

0

20

40

60

80

100

Ce

lls w

ith

in

tact

MT

(p

erc

en

tag

e)

A B C D

E

0.8

0.70

1.0

0.5

0

1.0

0 20 120 360 0 20 120 360Re

l. c

om

et

nu

mb

er

0.5

1.0

00 20 120 360

Time after nocodazole treatment (s)

Untreated Nocodazole

ns

shNF21 2

shNF2.1ns NF2ctrl

Untreated Nocodazole

NF2ctrl

0

0.75

1.50

2.25

3.00

GT

P r

em

na

nts

dis

tan

ce

(µ

m)

No

rma

lize

d G

TP

tubulin

sig

na

l in

ten

sity

F

***

***

0.1

0.2

0.3

0

*

*

Meso33RPE-1Meso33RPE-1 Meso33RPE-1

GT

P-c

ap

pe

d m

icro

tub

ule

s

(pe

rce

nta

ge

)

25

50

75

0

100

G

**

**

ctrl NF2

H

I

ns

shNF21 2 NF2ctrl ns

shNF21 2 ns

shNF21 2NF2ctrl NF2ctrl

**

***

*

00.51

1.52

2.53

3.54

MAP-T

ub

GST-X

a

NF2-

Xa

Are

a u

nd

er

cu

rve

* **

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0 5 10 15 20 25 30 35 40

OD

35

0

Time (min)

* *

GST-XaNF2-Xa

J K

Figure 4. NF2 protects microtubules from nocodazole-mediated depolymerization, increases net in vitro polymerization rate and GTP-tubulin frequency.A and B, representative pictures from RPE-1 cells expressing control (ns) or shRNA against NF2 (shNF2.1) and Meso33 cells expressing a control (ctrl)or NF2 (NF2) processed for immunofluorescence using an antitubulin antibody (green) before (top) or after (bottom) treatment with 10 mmol/Lnocodazole for 20 minutes. DAPI was used to stain the DNA (blue). Bar, 10 mm. C and D, quantification of cells treated as described in A and B. Cellswere scored for the presence of stable microtubules before (blue bars) and after (green bars) treatment with nocodazole. The bars represent meanvalues from three slides per condition from two independent experiments (N > 1,000 cells scored per condition; ��, P < 0.0001, Student t test). E, live-cell imaging and EB3-GFP particle tracking before (0 seconds) and after the addition of 40 nmol/L nocodazole to Meso33 ctrl and Meso33 NF2 cells(see Fig. 2B, iii). Loss of EB3-GFP signals and slowdown of growth and shrinkage speeds at 20, 120, and 360 seconds post-nocodazole addition isshown as a ratio to the same parameters measured before drug addition. For every time point the mean � SD of three movies from three differentplates are shown (> 8,000 tracks per condition). F, polymerization curves of MAP-purified tubulin in the presence of NF2 or GST alone. For every timepoint the mean �SD of three independent experiments is shown; �, P < 0.05. G, bar graphs depicting the area under curves from F, backgroundsubtracted mean � SD; �, P < 0.05; ��, P < 0.01. H, immunofluorescence costaining with anti-a-tubulin (green) and anti-GTP-tubulin (red) antibodies inMeso33 cells reconstituted with NF2 or the empty vector as control (ctrl). White arrowheads, GTP-tubulin remnants on the microtubule lattice;yellow arrows, GTP-capped microtubule plus ends. Bar, 10 mm. I, GTP-tubulin remnant signal intensity relative to tubulin intensity in RPE-1 cells with astable NF2 knock down (left) or Meso33 cells reconstituted with NF2 (right). The bar graphs represent mean � SD (�, P < 0.05; Student t test)from three independent experiments of each 15 cells. J, quantification of the distance between GTP-tubulin remnants on microtubules in cellsdescribed in I. The bar graphs represent mean � SD (��, P < 0.0001; Student t test) from three independent experiments (each experiment minimum12 cells and N > 500 distance measurements). K, percentage of the GTP-tubulin presence at the microtubule tips in cells described in I. The bar graphsrepresent mean � SD (��, P < 0.001; Student t test) from three independent experiments (each experiment minimum 5 cells and N > 100 tips).






To obtain more detailed insights into the mechanismunderlying NF2-mediated protection of microtubules fromnocodazole-induced depolymerization, we treated Meso33(ctrl) and Meso33(NF2) cells with substoichiometric con-centrations of nocodazole (40 nmol/L) and measured thetransient effects of this agent on EB3-GFP particle dynamicsover a time course of 6 minutes after perfusion. As previouslyshown, this treatment causes increased EB3-GFP disappear-ance over time (6). Meso33(NF2) cells displayed a signifi-cantly lower decay in particle number as Meso33(ctrl) cells(Fig. 4E, panel 1). Relative growth and shrinkage speeds werealso differentially affected by NF2 status (Fig. 4E, panels 2and 3). Notably, NF2's effect on relative growth and shrink-age rate reduction is most critical during the first 20 secondsof nocodazole-induced depolymerization. It is conceivablethat once nocodazole reaches an equilibrium in the system(that is when it binds to its cognate binding site on micro-tubules; ref. 5), decreases in relative growth and shrinkagespeeds are very similar. A possible explanation for thisphenomenon might be that NF2 reduces nocodazole bindingto microtubules in a noncompetitive manner. However, asdiscussed before, the relative comet numbers show a con-stant decrease over time in Meso33(ctrl) compared withMeso33(NF2) cells. This implies that at least over the timeperiod of our analysis, NF2 continually stabilizes the micro-tubule plus end during nocodazole treatment.

An earlier report demonstrated that NF2 enhances poly-merization of tubulin in vitro, although to a smaller extentwhen compared with adding microtubule-associated proteins(MAP) to the pure tubulin fraction (21). To further validatethis finding and get additional insight into whether NF2enhances also microtubule polymerization in vitro in thepresence of MAPs, we performed tubulin polymerizationassays of MAP-purified tubulin fractions (containing 30%MAPs) in the presence of purified NF2 (Supplementary Fig.S4B). Our results show that even in the presence of MAPs, NF2leads to enhanced tubulin polymerization when comparedwith controls (Fig. 4F and G).

It is well established that microtubule turnover is stronglyaffected by the presence of the GTP-form of tubulin. The latterstabilizes microtubules at the plus end by decreasing catas-trophe rates and, along the entire microtubule lattice, byincreasing rescue rates (26). We have previously observed astrong effect of pVHL on GTP-remnant frequency and GTP-capping (6). Therefore, we investigated whether NF2 wouldalso influence the localization, intensity, and frequency ofGTP tubulin on the microtubule lattice by indirect immuno-fluorescence microscopy. In both cell systems, we costainedtubulin and GTP-tubulin with respective antibodies (Fig. 4H)and quantified signal intensities, frequencies, and distancesbetween GTP-remnants. Interestingly, analysis of the net in-tensities of the GTP-tubulin signal showed that NF2 presenceleads to enhanced signals in both cell systems (Fig. 4I). More-over, distance measurements between GTP-remnants alongthe microtubule lattice as well as quantification of micro-tubule plus end GTP-capping frequencies demonstrated thatboth parameters are significantly affected depending on NF2status (Fig. 4J and K). This observation is in line with the

previous findings, that NF2 presence prolongs the growthlifetime (correlates with more GTP capping) and reduces theshrinkage length distance (correlates with GTP remnantsdistances). Collectively, these data strongly imply that NF2is influencing the intrinsic GTP hydrolysis rate of tubulinto GDP and Pi, which is translated in changed dynamicparameters.

Finally, given the striking parallels between pVHL and NF2in microtubule dynamics control, we were wondering wheth-er these two tumor suppressor proteins display any coop-erativity with respect to effects on microtubule dynamics. Weemployed a nocodazole protection assay in VHL-negativeccRCC cell line RCC4 (RCC4 ctrl) and the correspondingcounterpart engineered to reexpress wild-type pVHL (RCC4VHL30). As these cells are wild-type for NF2 expressionwe depleted in both of these cell lines NF2 with shNF2.1 orused a nonsilencing shRNA (ns) for control (SupplementaryFig. S4C). While single depletion of the tumor suppressorproteins under investigation showed reduced protectionfrom nocodazole-mediated microtubule depolymerizationtheir combined absence did not enhance this effect (Supple-mentary Fig. S4D), implying a lack of cooperation betweenNF2 and pVHL loss-of-function in the context of their micro-tubule regulatory function.

DiscussionOur live-cell imaging analysis of the effects of RNAi-medi-

ated knockdown of a collection of 70 tumor suppressor geneproducts identified NF2 as a key regulator of microtubuledynamics, besides 10 other genes, including LATS2 and BRCA1,which have known cytoskeleton interactions, at least duringmitosis at the centrosome (27, 28). These hits passed stringentselection criteria requiring robust shifts of a high-dimensionalparameter set relative to variations between internal controlsiRNAs. AXIN and VHL, two other tumor suppressor geneproducts with known effects on microtubule dynamics(2, 6), however, were not identified as potential hits. In thesecases the drop-out criteria were strongly influenced by het-erogeneous effects of control siRNA in the primary screen.Given the established roles these two tumor suppressor pro-teins play in microtubule regulation, it is possible that ourscreen contains additional false negatives. It also sheds light onthe challenges of interpreting the outcome of live-cell screensbased on sensitive, multivariate readouts of subcellular beha-viors (23, 29). Actual targeting effects in the multidimensionalreadout metric are confounded by variations in penetrance ofthe perturbation and in the basal states of the cell populationthat can easily reach statistically significant differencesbetween repeats. Thus, statistical tests alone are insufficientto identify true hits, but a series of fairly subjective choicesmust be made for the extraction of cell biologically meaningfultargets.

NF2 is mutated in a mutually exclusive manner with VHLin ccRCC and like pVHL affects multiple parameters of micro-tubule dynamic instability, which in combination result innet stabilization of microtubule dynamics. First, NF2 reducesnet tubulin turnover at the growing as well as at the shrinking

Smole et al.





microtubule plus end resulting in slower growth and shrinkagespeeds. Second, due to prolonged phases of growth, NF2reduces catastrophe rates, which may be the result of highermicrotubule GTP-cap frequencies observed in NF2-positivecells. Third, besides stabilizing effects in microtubule growthphases, NF2 also reduces the shrinkage distance after a catas-trophe event, which is in line with a higher GTP remnantfrequency along the microtubule lattice in NF2-expressingcells. GTP remnants coincide with microtubule rescue sites,promoting a switch from shrinkage to growth (6, 26). Bio-chemically, these effects might be directly linked to NF2interaction with microtubules as evidenced by coimmunopre-cipitation with tubulin and acetyl-tubulin. This interactionmay be either direct, mediated through two previously iden-tified tubulin-binding sites in NF2 (21), or indirect, intercededby kinesin or dynein motor proteins (20). Irrespective, lossof NF2 microtubule stabilizing effects might translate, forexample, into altered vesicle trafficking (19), failure of contactinhibition in concert with the actin cytoskeleton remodeling(30–32), or changes in centrosome positioning (18).While NF2 and pVHL demonstrate striking similarities

with respect to their effects on microtubule dynamic instabil-ity, we failed to observe cooperativity between NF2 and pVHLloss-of-function in the protection from nocodazole-mediatedmicrotubule depolymerization. This supports the view thatthese tumor suppressors may mediate their stabilizing effectsthrough similar and probably overlapping, mechanisms.pVHL's microtubule regulatory function has been linked tothe maintenance of the microtubule-based axoneme in theprimary cilium and suppression of renal cysts (33), precursorlesions of ccRCC. Deletion of Nf2 in the proximal convolutedepithelium of the kidney in mice produces also neoplasia thatprogressed to invasive carcinoma (34). Given this and the fact

that both pVHL and NF2 display microtubule regulatoryfunctions, it is conceivable that loss of NF2 in ccRCC maypredispose to renal pathology. Thus, disruption ofmicrotubuledynamics may be a common mechanism of renal cysts for-mation induced by loss of different tumor suppressor proteinsthat share common tumor suppressor mechanisms.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: C.R. Thoma, K.T. Applegate, G. Danuser, W. KrekDevelopment of methodology: C.R. Thoma, K.T. ApplegateAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): Z. Smole, C.R. Thoma, M. Duda, K.L. GutbrodtAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): Z. Smole, C.R. Thoma, K.T. Applegate, M. Duda,W. KrekWriting, review, and/or revision of the manuscript: Z. Smole, C.R. Thoma,M. Duda, G. Danuser, W. KrekAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): Z. Smole, C.R. Thoma, K.T. Applegate,W. KrekStudy supervision: C.R. Thoma, G. Danuser, W. Krek

AcknowledgmentsThe authors thank all members of our laboratories for discussions, A. Ittner

for providing pLKO.1-CMV vector, S. Jhanwar, F. Perez, and F. Giancotti forreagents, and the LMC ETH Zurich for support with microscopy.

Grant SupportW. Krek is supported by a grant from the Swiss National Science Foundation.

G. Danuser is supported by NIH grant U01 GM67230.The costs of publication of this article were defrayed in part by the

payment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicate thisfact.

Received May 9, 2013; revised September 13, 2013; accepted October 31, 2013;published OnlineFirst November 26, 2013.

References1. Munemitsu S, Souza B, Muller O, Albert I, Rubinfeld B, Polakis P. The

APC gene product associates with microtubules in vivo and promotestheir assembly in vitro. Cancer Res 1994;54:3676–81.

2. Fumoto K, Kadono M, Izumi N, Kikuchi A. Axin localizes to thecentrosome and is involved in microtubule nucleation. EMBO Rep2009;10:606–13.

3. Hergovich A, Lisztwan J, Barry R, Ballschmieter P, KrekW. Regulationof microtubule stability by the von Hippel–Lindau tumour suppressorprotein pVHL. Nat Cell Biol 2003;5:64–70.

4. Lolkema MP, Mehra N, Jorna AS, van Beest M, Giles RH, Voest EE.The von Hippel–Lindau tumor suppressor protein influencesmicrotubule dynamics at the cell periphery. Exp Cell Res 2004;301:139–46.

5. Schermer B, Ghenoiu C, Bartram M, Muller RU, Kotsis F, Hohne M,et al. The von Hippel–Lindau tumor suppressor protein controlsciliogenesis by orienting microtubule growth. J Cell Biol 2006;175:547–54.

6. Thoma CR, Matov A, Gutbrodt KL, Hoerner CR, Smole Z, Krek W,et al. Quantitative image analysis identifies pVHL as a keyregulator of microtubule dynamic instability. J Cell Biol 2010;190:991–1003.

7. Thoma CR, Toso A, Gutbrodt KL, Reggi SP, Frew IJ, Schraml P, et al.VHL loss causes spindle misorientation and chromosome instability.Nat Cell Biol 2009;11:994–1001.

8. Nakamura M, Zhou XZ, Lu KP. Critical role for the EB1 and APCinteraction in the regulation of microtubule polymerization. Curr Biol2001;11:1062–7.

9. Matov A, Applegate K, Kumar P, Thoma C, Krek W, Danuser G, et al.Analysis of microtubule dynamic instability using a plus-end growthmarker. Nat Methods 2010;7:761–8.

10. Dalgliesh GL, Furge K, Greenman C, Chen L, Bignell G, Butler A, et al.Systematic sequencing of renal carcinoma reveals inactivation ofhistone modifying genes. Nature 2010;463:360–3.

11. Asthagiri AR, Parry DM, Butman JA, KimHJ, Tsilou ET, Zhuang Z, et al.Neurofibromatosis type 2. Lancet 2009;373:1974–86.

12. Bianchi AB, Mitsunaga SI, Cheng JQ, Klein WM, Jhanwar SC, Sei-zinger B, et al. High frequency of inactivating mutations in the neuro-fibromatosis type 2 gene (NF2) in primary malignant mesotheliomas.Proc Natl Acad Sci U S A 1995;92:10854–8.

13. Rustgi AK, Xu L, Pinney D, Sterner C, Beauchamp R, Schmidt S, et al.Neurofibromatosis 2 gene in human colorectal cancer. Cancer GenetCytogenet 1995;84:24–6.

14. Murray LB, Lau YK, Yu Q. Merlin is a negative regulator of humanmelanoma growth. PLoS ONE 2012;7:e43295.

15. Hamaratoglu F, Willecke M, Kango-Singh M, Nolo R, Hyun E, Tao C,et al. The tumour-suppressor genes NF2/Merlin and Expanded actthrough Hippo signalling to regulate cell proliferation and apoptosis.Nat Cell Biol 2006;8:27–36.






16. Li W, You L, Cooper J, Schiavon G, Pepe-Caprio A, Zhou L, et al.Merlin/NF2 suppresses tumorigenesis by inhibiting the E3ubiquitin ligase CRL4(DCAF1) in the nucleus. Cell 2010;140:477–90.

17. Gladden AB, Hebert AM, Schneeberger EE, McClatchey AI. The NF2tumor suppressor, Merlin, regulates epidermal development throughthe establishment of a junctional polarity complex. Dev Cell 2010;19:727–39.

18. Hebert AM, DuBoff B, Casaletto JB, Gladden AB, McClatchey AI.Merlin/ERM proteins establish cortical asymmetry and centrosomeposition. Genes Dev 2012;26:2709–23.

19. Hennigan RF, Moon CA, Parysek LM, Monk KR, Morfini G, Berth S,et al. The NF2 tumor suppressor regulates microtubule-based vesicletrafficking via a novel Rac, MLK and p38(SAPK) pathway. Oncogene2013;32:1135–43.

20. Bensenor LB, Barlan K, Rice SE, Fehon RG, Gelfand VI. Microtubule-mediated transport of the tumor-suppressor protein Merlin and itsmutants. Proc Natl Acad Sci U S A 2010;107:7311–6.

21. Muranen T, Gronholm M, Lampin A, Lallemand D, Zhao F, GiovanniniM, et al. The tumor suppressor merlin interacts with microtubules andmodulates Schwann cell microtubule cytoskeleton. Hum Mol Genet2007;16:1742–51.

22. Applegate KT, Besson S, Matov A, Bagonis MH, Jaqaman K,Danuser G. plusTipTracker: Quantitative image analysis softwarefor the measurement of microtubule dynamics. J Struct Biol 2011;176:168–84.

23. Mohr SE, Perrimon N. RNAi screening: new approaches, understand-ings, and organisms. Wiley Interdiscip Rev RNA 2012;3:145–58.

24. Westermann S, Weber K. Post-translational modifications regulatemicrotubule function. Nat Rev Mol Cell Biol. 2003;4:938–47.

25. Sale WS, Besharse JC, Piperno G. Distribution of acetylated alpha-tubulin in retina and in vitro-assembled microtubules. Cell Motil Cyto-skeleton 1988;9:243–53.

26. Dimitrov A, Quesnoit M, Moutel S, Cantaloube I, Pous C, Perez F.Detection of GTP-tubulin conformation in vivo reveals a role for GTPremnants in microtubule rescues. Science 2008;322:1353–6.

27. Joukov V, Groen AC, Prokhorova T, Gerson R, White E, Rodriguez A,et al. The BRCA1/BARD1 heterodimer modulates ran-dependentmitotic spindle assembly. Cell 2006;127:539–52.

28. Toji S, Yabuta N, Hosomi T, Nishihara S, Kobayashi T, Suzuki S, et al.The centrosomal protein Lats2 is a phosphorylation target of Aurora-Akinase. Genes Cells 2004;9:383–97.

29. Pepperkok R, Ellenberg J. High-throughput fluorescence microscopyfor systems biology. Nat Rev Mol Cell Biol 2006;7:690–6.

30. Okada T, Lopez-Lago M, Giancotti FG. Merlin/NF-2 mediates contactinhibition of growth by suppressing recruitment of Rac to the plasmamembrane. J Cell Biol 2005;171:361–71.

31. Kadir S, Astin JW, Tahtamouni L, Martin P, Nobes CD. Microtubuleremodelling is required for the front-rear polarity switch during contactinhibition of locomotion. J Cell Sci 2011;124:2642–53.

32. Stramer B, Moreira S, Millard T, Evans I, Huang CY, Sabet O, et al.Clasp-mediated microtubule bundling regulates persistent motilityand contact repulsion in Drosophila macrophages in vivo. J Cell Biol2010;189:681–9.

33. Thoma CR, Frew IJ, Hoerner CR, Montani M, Moch H, Krek W. pVHLand GSK3beta are components of a primary cilium-maintenancesignalling network. Nat Cell Biol 2007;9:588–95.

34. Morris ZS, McClatchey AI. Aberrant epithelial morphology and per-sistent epidermal growth factor receptor signaling in amousemodel ofrenal carcinoma. Proc Natl Acad Sci U S A 2009;106:9767–72.

Smole et al.





2014;74:353-362. Published OnlineFirst November 26, 2013.Cancer Res Zlatko Smole, Claudio R. Thoma, Kathryn T. Applegate, et al. Tumor Suppressor NF2/Merlin Is a Microtubule Stabilizer

Updated version

10.1158/0008-5472.CAN-13-1334doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2013/11/26/0008-5472.CAN-13-1334.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/74/1/353.full#ref-list-1

This article cites 34 articles, 12 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/74/1/353.full#related-urls

This article has been cited by 2 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/74/1/353To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-13-1334

http://cancerres.aacrjournals.org/content/suppl/2013/11/26/0008-5472.CAN-13-1334.DC1

http://cancerres.aacrjournals.org/content/74/1/353.full#ref-list-1

http://cancerres.aacrjournals.org/content/74/1/353.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/74/1/353


tumor suppressor nf2/merlin is a microtubule stabilizer · tumor and stem cell biology tumor...

Documents