- 1 -

The formin homology protein mDia1 regulates dynamics of microtubules and

their effect on focal adhesion growth

Christoph Ballestrem,* Natalia Schiefermeier,*ƒ Julia Zonis,* Michael Shtutman,* Zvi

Kam,* Shuh Narumiya, Arthur S. Alberts, ⁄ and Alexander D. Bershadsky*

*Department of Molecular Cell Biology, The Weizmann Institute of Science, Rehovot

76100, Israel; Department of Pharmacology, Kyoto University Faculty of Medicine,

Kyoto, Japan; ⁄Van Andel Research Institute, Grand Rapids, MI, USA.

ƒThis author made significant contribution to this paper

Address correspondence to:

Alexander Bershadsky

Department of Molecular Cell Biology

The Weizmann Institute of Science

P.O. Box 26, Rehovot 76100, Israel

Tel.: 972-8-9342884

Fax: 972-8-9344125

E-mail: alexander.bershadsky@weizmann.ac.il

Total characters: 59107

Running Title: mDia1 regulates dynamics of microtubules

Keywords: mDia, formin homology protein, microtubule, focal adhesion, actin

- 2 -

Abstract

The formin homology protein, mDia1, is a major effector of Rho controlling, together

with the Rho-kinase (ROCK), the formation of focal adhesions and stress fibers. Here we

show that a constitutively active form of mDia1 (mDia1∆N3) affects the dynamics of

microtubules at three stages of their life. We found that in cells expressing mDia1∆N3,

(1) the growth rate at the microtubule plus-end decreased by half, (2) the rates of

microtubule plus-end growth and shortening at the cell periphery decreased while the

frequency of catastrophes and rescue events remained unchanged, and (3) mDia1∆N3

expression in cytoplasts without centrosome stabilized free microtubule minus-ends. This

stabilization required the activity of another Rho target, ROCK. Interestingly, mDia1∆N3

as well as endogenous mDia1, localized at the centrosome. The changes in microtubule

behavior in the mDia1∆N3-expressing cells increased both microtubule targeting toward

focal adhesions, and their inhibitory effect on focal adhesion growth. Thus, mDia1-

induced alterations in microtubule dynamics augment the microtubule-mediated negative

regulation of focal adhesions.

- 3 -

Introduction

The assembly of integrin-mediated cell-matrix adhesions (known as focal adhesions or

focal contacts) depends on cell contractility, is coupled with the formation of associated

actin filament bundles (stress fibers), and is modulated by microtubules working as local

negative regulators (Geiger and Bershadsky, 2001; Small et al., 2002). Signal

transduction pathways triggering the formation of the focal adhesions require the activity

of the small GTPase RhoA (Ridley and Hall, 1992), which operates via its two major

targets, the Rho-associated kinase (ROCK) and the formin homology protein mDia1

(Watanabe et al., 1999). While the main function of ROCK in the formation of focal

adhesions is the regulation of myosin II activity (Kimura et al., 1996; Totsukawa et al.,

2000), the functions of mDia1 in this process are less clear. mDia1 belongs to a

conserved family of Diaphanous-related formins (DRF), present in most eukaryotic cells

(Evangelista et al., 2003; Wallar and Alberts, 2003). Interaction of these proteins with

active Rho or, in some cases, with other Rho family GTPases leads to a conformational

change - opening - that exposes formin homology (FH) domains 1 and 2 (Alberts, 2001;

Alberts, 2002; Watanabe et al., 1999). Truncated constructs containing FH1 and FH2

domains in the deregulated conformation are usually constitutively active (Evangelista et

al., 1997; Watanabe et al., 1999; Tominaga et al., 2000). One known activity of the

formin family molecules is the promotion of actin polymerization. The proline-rich FH1

domain binds to profilin (Chang et al., 1997; Evangelista et al., 1997; Imamura et al.,

1997; Watanabe et al., 1997), while the FH2 domains bind actin (Pring et al., 2003;

Pruyne et al., 2002). With assistance of profilin, budding yeast formins Bni1p and Bnr1p

(Pring et al., 2003; Pruyne et al., 2002; Sagot et al., 2002) and fission yeast formin

- 4 -

Cdc12p (Kovar et al., 2003) nucleate actin filaments that grow rapidly from their barbed

ends. Formins may work as leaky cappers that associate with barbed filament end but

still allow filament elongation even in the presence of tight capping proteins (Zigmond et

al., 2003). It has been recently shown that FH2-containing constructs of mDia1 are even

more potent actin nucleators than the yeast formins (Li and Higgs, 2003).

Besides these effects on actin, there is evidence that formins affect also the

microtubular system (for a review, Gundersen, 2002). In fission yeast, gene disruption

and the overexpression experiments demonstrated that For3 formin controls not only

actin cytoskeleton, but also cytoplasmic microtubule organization (Feierbach and Chang,

2001; Nakano et al., 2002). In mammalian cells, the active form of mDia1 induces

alignment of microtubules along the axes of bipolar cells. This effect was shown to

depend on the FH2 domain of mDia1 (Ishizaki et al., 2001). Furthermore, expression of

an active form of the closely related mDia2 formin or its Diaphanous-autoregulatory

domain which binds to and activates endogenous mDia1 (Alberts, 2001), increases the

fraction of microtubules containing detyrosinated α-tubulin, with Glu instead of Tyr at its

C-terminus (Palazzo et al., 2001); mDia2 was also found to associate with Taxol-

stabilized microtubules. Since α-tubulin detyrosination occurs mainly on long-living

microtubules (Bulinski and Gundersen, 1991; Kreis, 1987; Webster et al., 1987; Webster

et al., 1990), these data suggest that active form of formins may induce microtubule

stabilization. The fraction of microtubules enriched in detyrosinated (Glu) tubulin,

however, is usually rather small and may depend on factors other than microtubule

longevity (Mialhe et al., 2001). Therefore, these alterations do not necessarily provide

information about formin effects on the bulk of microtubule population.

- 5 -

It has been shown previously that focal adhesions can be targeted by microtubules

(Kaverina et al., 1998) and, often, are down regulated by this targeting (Kaverina et al.,

1999), while disruption of microtubules leads to focal adhesion growth (Bershadsky et

al., 1996). We demonstrate here for the first time that mDia1 governs multiple aspects of

the microtubule dynamics and their targeting to and regulation of focal adhesion growth.

At the plus ends, mDia1 decreased microtubule elongation and shortening velocities, but

did not affect the frequencies of transition from growth to shrinking phases and vice

versa. This effect did not require assistance of ROCK. In addition, mDia1 localized to the

centrosome, and expression of its active form efficiently protected microtubules in the

centrosome-lacking cytoplasts from depolymerization, stabilizing the microtubule minus-

ends. This minus-end effect required also the ROCK activity and was abolished by the

inhibitor, Y-27632. These effects of mDia1 on microtubule dynamics shed new light on

the regulation of microtubule interactions with focal adhesions. We propose that mDia1-

mediated changes in microtubule dynamics provide a mechanism coordinating the Rho-

and ROCK-dependent focal adhesion maturation with microtubule-dependent focal

adhesion turnover.

- 6 -

Results

The active form of mDia1 induces reorganization of both the actin and microtubular

cytoskeleton

Transfection of CHO-K1 cells with a constitutively active mutant of mDia1 (mDia1∆N3)

induces a dramatic reorganization of the actin cytoskeleton (in line with (Watanabe et al.,

1999)). Cells expressing the active mDia1 display a characteristic bipolar shape with

finger-like actin-rich projections at both ends (Fig. 1 A and B). Numerous thin actin

bundles extend from pole to pole and fill the entire cytoplasm (Fig. 1 B). The amount of

polymerized actin, estimated by intensity of rhodamine-phalloidin fluorescence, was

about five-fold higher in the mDia1∆N3 transfected cells than in control cells (Fig. 1 C).

Active mDia1 expressed in HeLa cells induces microtubule alignment along the

cell axis (Ishizaki et al., 2001). We made similar observations using CHO-K1 cells,

where, in contrast to the radial microtubule distribution in control cells (Fig. 2A)

mDia1∆N3-expression lead to alignment along the cell axis (Fig. 2 B and C). In order to

better analyze the mDia1 effect on microtubules we used a recently developed software

that allows to quantify the orientation and length of fibrillar structures (Lichtenstein et al.,

2003). We found that the degree of microtubule alignment increased about four-fold in

the cells expressing mDia1∆N3, compared to control cells (Fig. 2 D). In addition, active

mDia1 in CHO-K1 cells induced a moderate but a statistically significant (p<0.01)

increase in total microtubule length (Fig. 2 E), suggesting either an enhanced microtubule

polymerization or an increase of microtubule stability in these cells.

- 7 -

Active mDia1 slows down microtubule plus-end dynamics

For a more detailed analysis of changes in the microtubule system of cells expressing

active mDia1 we first studied its influence on microtubule dynamics in living cells using

GFP-EB1 chimeric protein (Mimori-Kiyosue et al., 2000). This approach is particularly

useful since EB1 marks only growing microtubule plus ends, which allows analysis of the

microtubule growth rate without disturbing background fluorescence. To quantify the

velocity we superimposed successive images of the GFP-EB1-labeled microtubule ends

and measured the displacement of the EB1-positive microtubule tips during the given

time period (Fig. 3 A-D; Fig3 video 1-4).

In CHO-K1 cells expressing GFP-EB1 only microtubules grew with a rate of

about 23 µm/min from the centrosome towards the cell periphery (Fig. 3 A-C; video 1

and 2). In contrast, growth rates in cells expressing active mDia1 were decreased by half

(Fig. 3 G and J; video 2 and 3). In addition, in these cells the speeds of individual

microtubule tips were not constant, in contrast to those in control cells, but varied slightly

during the period of microtubule growth (Fig. 3 F, G, H).

Active mDia1 decreases microtubule elongation and shortening velocities at the cell

periphery

Close to the cell edge, microtubule plus-ends oscillate, transitioning between growth and

shortening phases, a process known as dynamic instability (Cassimeris et al., 1988;

Komarova et al., 2002; Sammak and Borisy, 1988; Shelden and Wadsworth, 1993). Since

GFP-EB1 immediately disappears from the ends when microtubules begin to shorten

(Mimori-Kiyosue et al., 2000), it is not useful for assessing parameters such as

- 8 -

microtubule shortening. To monitor dynamic instability events of microtubules at the cell

periphery we used B16 cells that were stably transfected with GFP-tubulin (Ballestrem et

al., 2000) and transiently co-expressed mDia1∆N3 together with DsRed as a marker to

identify transfected cells (Fig. 4 A, B; video 5 and 6). As seen in CHO-K1 cells,

microtubules in mDia1∆N3 expressing B16 cells were aligned along the cell axis (Fig. 4

B and 9 B), while microtubules in control cells expressing GFP-tubulin only or GFP-

tubulin and DsRed had typically a curved shape (Fig. 4 A and 9 A). To better visualize

and measure dynamic instability parameters of microtubules in cells, two successive

images of time-lapse recordings taken in 5s intervals were first labeled in red and green

and then were superimposed. De novo polymerized or de-polymerized microtubule ends

appear as red or green microtubule ends, respectively (Fig. 4A, B video 5 and 6).

Structures remaining at identical location at the two different time-points appeared

yellow. Comparing the typical parameters of dynamic instability in mDia1∆N3-

transfected and control B16 cells (Table 1) we found that the velocities of microtubule

elongation and microtubule shortening were both reduced in mDia1∆N3-transfected cells,

compared to those in control cells, resulting in smaller amplitudes of microtubule end

oscillations at the periphery of the mDia1∆N3-transfected cells (Fig. 4 C and D). On the

other hand, the frequency of catastrophes and rescue events as well as the time that

microtubules spent growing, pausing and shortening were similar in control and

mDia1∆N3-expressing cells (Table 1).

Active mDia1 inhibits buckling of microtubules

- 9 -

Time-lapse recordings and the superimposition of successive images of GFP-

tubulin-labeled cells (Fig. 4 A and B; video 5and 6) show that the contours of many

microtubules in control cells changed significantly during the period of observation. The

numerous green- and red-labeled parts of microtubules seen in the inner part of the

lamella (Fig 4A and video5) shows that microtubules undergo frequent lateral

displacements caused by continuous buckling and unbuckling. In cells expressing active

mDia1, microtubule buckling was almost entirely abolished. Microtubules in these cells

remained relatively straight and did not change lateral positions (Fig 4B, video 6).

Centrosomal localization of mDia1.

In spite of its effect on plus end dynamics, we did not detect specific mDia1 localization

on the microtubules or their tips. Surprisingly, we detected GFP- mDia1∆N3 co-

localizing at the centrosome with the centrosomal marker γ-tubulin (Fig 5A-C). The FH2

domain seems to be sufficient for centrosomal localization since truncated form of mDia1

comprising only the FH2 domain fused to GFP was also found at the centrosome (Fig

5D). Furthermore, immunofluorescence staining using a rabbit antibody against mDia1

(Tominaga et al., 2000) shows the centrosomal localization of endogenous mDia1 in B16

cells (Fig. 5 G-I). These data suggest that mDia1 may directly or indirectly associate with

the microtubule minus ends.

Active mDia1 stabilizes microtubules in centrosome-free cytoplasts

In view of the enrichment of active mDia1 at the centrosome we looked for a possible

role of mDia1 at the microtubule minus end and therefore studied microtubule dynamics

- 10 -

in the cytoplasts lacking centrosomes. Previous studies demonstrated that the total

microtubule length in centrosome-free cytoplasts is several times lower than in intact

cells (Karsenti et al., 1984; Rodionov et al., 1999). This difference was attributed to a

rapid depolymerization of microtubules from their non-protected minus ends (Rodionov

et al., 1999). Indeed, also in our experiments, non-centrosomal cytoplasts produced from

control, GFP expressing, CHO-K1cells (Fig. 6 A e-g, B) or from CHO-K1cells

expressing mDia1∆N3 bearing a triple point mutation in its FH2 domain (KA3) (Ishizaki

et al., 2001) contained only a little amount of mostly treadmilling microtubules, as

compared to centrosome-containing cytoplasts, or intact cells (Fig. 6 A i-k, B, and data

not shown). Microtubules in these (usually, non-polarized) cytoplasts were randomly

organized and not aligned along a specific axis. In contrast, cytoplasts from CHO-K1

cells expressing active mDia1 (mDia1∆N3) had essentially the same morphology as

intact mDia1∆N3-expressing cells. Irrespective of the presence of the centrosome, they

were bipolar (Fig. 6 Aa), with parallel thin actin bundles throughout the entire cytoplasts

(not shown). Similarly to mDia1∆N3-expressing intact cells, mDia1∆N3-expressing

cytoplasts demonstrated a high degree of microtubule alignment along the bipolar cell

axis (Fig. 6 A b-c, B).

Measurements of total microtubule length revealed that centrosome-free

cytoplasts from control cells have a significantly lower amount of microtubule polymer

than centrosome-containing cytoplasts (not shown). On the other hand, centrosome-free

cytoplasts expressing mDia1∆N3 had an approximately four-fold increase in the amount

of microtubule polymer, as compared to centrosome-free cytoplasts expressing GFP

alone or mDia1∆N3KA3 (Fig. 6 C).

- 11 -

To test whether this increase in microtubule total length can be explained by

changes in microtubule plus-end dynamics induced by mDia1∆N3, we compared

microtubule growth velocities in control and mDia1∆N3-expressing cytoplasts. The

cytoplasts were prepared from CHO-K1 cells transfected with either GFP-EB1 alone

(control), or GFP-EB1 together with mDia1∆N3. The values of microtubule plus-end

velocity in the cytoplasts prepared from control and mDia1∆N3 cells were the same as in

intact cells (Fig. 6 C). Thus, in the cytoplasts, similarly to intact cells, mDia1∆N3

decreases the rate of microtubule polymerization at the plus end irrespective of the

presence of the centrosome (Fig. 6 C). Therefore, the increase in microtubule mass

observed in centrosome-free cytoplasts expressing mDia1∆N3 cannot be explained by an

enhanced polymerization of microtubules at the plus ends; instead it should be attributed

to the stabilization of the minus ends.

ROCK is required for the mDia1 induced stabilization of microtubule minus ends, but

for the mDia1 effect on microtubule plus-ends

mDia1 and ROCK are both activated by Rho-GTP, and cooperate in the formation of

actin bundles and focal adhesions (Tominaga et al., 2000; Watanabe et al., 1999). We

tested whether ROCK cooperates with mDia1 in the control of microtubule dynamics. To

investigate this we blocked ROCK activity using the specific inhibitor, Y-27632 (Uehata

et al., 1997). Treatment of mDia1∆N3-expressing cells with this inhibitor abolished cell

polarization and formation of the parallel arrays of actin fibers (not shown). Despite the

treatment, microtubule plus end growth rate was still diminished in mDia1∆N3-

- 12 -

expressing cells (Fig. 7 A). This indicates that the effect of mDia1 on the dynamics of

microtubule plus ends does not require cooperation of mDia1 with ROCK.

We further tested whether the effect of active mDia1 on the stabilization of

microtubule minus ends requires ROCK activity. Interestingly, here the ROCK-inhibitor

Y-27632 abolished the mDia1∆N3-induced increase of the microtubule mass in the

centrosome-free cytoplasts (Fig. 7 B). At the same time, Y-27632 did not affect the total

microtubule length in either centrosome-containing cytoplasts, or in control or

mDia1∆N3-expressing intact cells (data not shown). Furthermore, the inhibition of

ROCK had no effect on the increase of actin polymerization levels observed in

mDia1∆N3-expressing cells (Fig.7 C). Thus, mDia1-induced stabilization of microtubule

minus ends in centrosome-free cytoplasts, unlike the effect of mDia1 on the microtubule

plus end dynamics, required the activity of ROCK.

Active mDia1 augments microtubule targeting to focal adhesions and their response to

microtubule disruption

One of the more recently described microtubule functions is the regulation of focal

adhesions. Microtubules are often targeted to focal adhesions and this targeting promotes

focal adhesion turnover (Small et al., 2002). Since the active form of mDia1 causes

pronounced changes in microtubule dynamics and stabilization, we tested whether these

changes influenced focal adhesions. To visualize focal adhesion targeting by

microtubules we again used B16 cells stably expressing GFP-tubulin (Ballestrem et al.,

2000), and transfected them by a construct encoding a focal adhesion marker, RFP-zyxin

(Bhatt et al., 2002), with or without mDia1∆N3. As depicted in Fig 8 A and B (video 7

- 13 -

and 8) microtubules target focal adhesions in both control cells and mDia1∆N3-

transfected cells. However, the mode of the microtubule-focal adhesion interaction was

different in control and mDia1∆N3-transfected cells. To characterize this difference, we

counted the frequencies of microtubule entry into and exit from a 2 µm size square boxes

enclosing randomly selected individual focal adhesions (Fig 8 A and B, video 7 and 8).

The total number of microtubules associated with these boxes was, on the average,

similar in control and mDia1∆N3-transfected cells (Table 2). However, in control cells

microtubule tips usually went in and out the box several times during the 1.5 min period

of observation, while in mDia1∆N3-transfected cells the microtubule tips as a rule

remained inside the box for the entire 1.5 minutes (Fig. 8 A B; video 7 and 8). Thus, the

average time individual microtubule tips spent in the proximity of the focal adhesion

increased upon mDia1 activation.

It is well known that microtubule disruption leads to focal adhesion growth

(Bershadsky et al., 1996; Enomoto, 1996; Liu et al., 1998; Pletjushkina et al., 1998;

Kirchner et al., 2003). This phenomenon provides a tool that allows comparing the degree

of microtubule-dependant negative regulation of focal adhesions in control and Dia1

expressing cells. To investigate this, we transfected CHO-K1 cells YFP-paxillin as a

focal adhesion marker, or YFP-paxillin together with mDia1∆N3 and measured the total

area of paxillin-labeled focal adhesions per cell before and after disruption of

microtubules with nocodazole. In control cells, expressing only YFP-paxillin (Fig. 9 A),

microtubule disruption induced statistically significant (p<0.05) increase of focal

adhesion size (Fig. 9 B and E). In addition, expression of active mDia1 led to an increase

of focal adhesions (Fig. 9C and E). Notably, disruption of microtubules in the cells

- 14 -

expressing mDia1∆N3 induced a dramatic increase of focal adhesion area (Fig. 9 D and

E), which apparently exceeded the sum of the effects produced by microtubule disruption

and mDia1∆N3 expression separately. This synergism can be interpreted as an evidence

that active mDia1 enhances the inhibitory effect of microtubules on the focal adhesion

growth. Thus, active mDia1-induced alterations in microtubule dynamics modulate

microtubule effect on focal adhesions.

- 15 -

Discussion

In this study we investigated the role of mDia1 in the precisely controlled dynamics of

microtubules and their influence over focal adhesions. Dynamics of the microtubule plus

ends at the cell periphery can be described as a process of stochastic transitions between

growth and shortening phases ( dynamic instability ). It was demonstrated recently that

in budding and fission yeast (Brunner and Nurse, 2000) and in mammalian cells

(Komarova et al., 2002), that microtubules nucleated at the centrosome grow with a

constant speed until they reach the cell periphery. It is only at this site that microtubule

plus ends demonstrate oscillatory behavior, with alternating periods of shrinking and

growth. These observation suggest that the localized cellular environment defines the

plus-end microtubule dynamics are not uniform over the cytoplasm. Instead, regulation

depends upon local cues.

Minus end regulation is not any less important. It was recently shown that

microtubules are often released and detached from the centrosome and that the dynamics

of the free minus ends of these non-centrosomal microtubules are also strictly controlled

(Abal et al., 2002; Chausovsky et al., 2000; Keating et al., 1997; Rodionov et al., 1999;

Vorobjev et al., 1999) (see for a review (Dammermann et al., 2003)). Therefore it is

possible that cellular factors that direct/influence overall microtubule dynamics could act

at one or both of these sites. Thus, to characterize the mode of action of any putative

regulatory factor, information about its effects on all these aspects of microtubule

dynamics should be provided.

Our results show that the Diaphanous-related formin mDia1 is a master regulator

regulator of microtubule dynamics at plus and minus ends. Using GFP-EB1 (Mimori-

- 16 -

Kiyosue et al., 2000) as a marker of the growing microtubule plus-ends, we showed that

expression of active mDia1 results in a two-fold decrease in the rate of microtubule

growth. This result clearly shows that mDia1 can affect microtubule assembly in the

phase of persistent growth, although it cannot be excluded that exogenous GFP-EB1 is

not an innocent marker, as it can by itself modulate microtubule dynamics (Ligon et al.,

2003; Nakamura et al., 2001; Tirnauer et al., 2002). Furthermore, analysis of microtubule

behavior at the periphery of GFP-tubulin expressing cells revealed that active mDia1

diminished both microtubule elongation and shortening rates. Unlike to the majority of

known microtubule regulators, it does not affect probability of catastrophe and rescue

events. In addition, the active form of mDia1 affects the mechanical characteristics of

entire microtubule by suppressing spontaneous microtubule buckling, which may explain

augmented microtubule orientation and alignment in the affected cells. Finally, when we

evaluated microtubule minus end dynamics in cytoplasts devoid of centrosomes, we

found that active mDia1 protects microtubules from rapid depolymerization, by

stabilization of the minus ends.

The net effect of expression of activated mDia1 is that the average microtubule

survival time increases, thereby increasing the likelyhood of post-translational

modifications that accompany stabilization. This is consistent with the results of (Palazzo

et al., 2001) who observed an increased fraction of detyrosinated (Glu) microtubules in

the cells with active mDia1. Our data significantly extend these results by showing that

active mDia1 affects the dynamics and mechanics of all microtubules rather than

stabilizes a relatively small subpopulation of them.

- 17 -

For the explanation of the effects of mDia1 on microtubule dynamics two major

types of mechanism can be proposed. First, since it is well documented that several DRFs

including mDia1 efficiently promote actin polymerization (Evangelista et al., 2003; Li

and Higgs, 2003), the effects of mDia1 on microtubules can be a consequence of its

effect on the actin cytoskeleton, and do not necessarily require direct interaction of

mDia1 with microtubule components. The second possibility is that mDia1, like mDia2

(Palazzo et al., 2001; Peng et al., 2003), associates with microtubules, and is influencing

microtubule associated protein(s) that control microtubule dynamics.

The actin-dependent mechanism in its simplest form can be based on the fact that

the microtubule growth velocity and dynamic instability parameters in vitro can be

affected by mechanical barriers (Dogterom and Yurke, 1997; Janson et al., 2003). Such

barriers resist the pushing forces developed by the growing microtubule ends and limit

the rate of tubulin subunit addition (Janson et al., 2003). Similarly, in the cell, actin

structures may mechanically modulate microtubule growth, and mDia1 effects on

microtubules might be then a consequence of alterations in actin cytoskeleton rigidity

and/or spatial organization. Obviously, the actin cytoskeleton is more than a system of

mechanical obstacles for microtubules. Several types of protein links between

microtubules and the actin cytoskeleton have been discovered (reviewed in (Goode et al.,

2000; Rodriguez et al., 2003)), including plakin family members (Karakesisoglou et al.,

2000; Lee and Kolodziej, 2002; Leung et al., 2002; Subramanian et al., 2003), coronin

and coronin-like protein Dpod-1 (Goode et al., 1999; Rothenberg et al., 2003), or the

yeast Bim1-Kar9-Myo2 complex (Hwang et al., 2003). These links may affect

microtubule dynamics more specifically than just inert barriers, so that changes in the

- 18 -

actin polymerization induced by mDia1 could be efficiently translated into observed

changes in the microtubule dynamics and organization. There is strong evidence that

formin effects on microtubules in budding yeast are caused by alterations of the actin

cytoskeleton (Bretscher, 2003). Detailed elaboration of this hypothesis in higher

eukaryotic cells, however, requires additional information on the localization and mode

of action of these actin-microtubule links.

It is tempting to explain mDia1-induced decrease in microtubule growth and

shrinking velocity and, perhaps, suppression of buckling by an actin-dependent

mechanism, but events such as the minus end stabilization, appear to require additional

assumptions. Several observations suggest a direct interaction of mDia1 with the

microtubule system. First, we found that mDia1, as well as its constitutively active

mutant mDia∆N3 and isolated FH2 domain, co-localizes with γ-tubulin at the

centrosome. This result is consistent with the recent study that showed centrosomal

localization of mDia2 (Peng et al., 2003). The centrosomal localization of mDia1

suggests that it might interact with some proteins involved in the microtubule minus-end

anchorage and stabilization. Interaction with such proteins could explain the observed

mDia1 effect on the microtubule stabilization also in the centrosome-lacking cytoplasts.

It is worth noting that another major target of Rho, ROCK, was recently shown to be

localized to the centrosome too (Chevrier et al., 2002). Furthermore, in the present study

we have demonstrated that the ROCK inhibitor, Y-27632, prevented the stabilizing effect

of mDia1 on the microtubules in the centrosome-free cytoplasts. At the same time, this

inhibitor did not reduce the stimulating mDia1 effect on the actin polymerization. This

result suggests that mDia1-induced enhancement of actin polymerization is by itself not

- 19 -

sufficient for the stabilization of non-centrosomal microtubules. Cooperation with ROCK

and, perhaps, interaction with certain centrosome/microtubule-associated targets seem to

be required for mDia1 effect on microtubules.

It appeared recently that a group of proteins that specifically localizes to

microtubule tips is very important for the regulation of microtubule dynamics and their

targeting to the cell cortex (reviewed in Carvalho et al., 2003). Among these proteins are

EB1 and its partner APC (Mimori-Kiyosue et al., 2000; Nakamura et al., 2001), CLIP-

170 and its partners of the CLASP family (Akhmanova et al., 2001; Perez et al., 1999),

dynein and its receptor dynactin (in particular its major component p150 Glued)

(Vaughan et al., 1999), and Lis1 protein that can interact with dynein and with CLIP-170

(Coquelle et al., 2002). Remarkably, many, if not all of these microtubule tip proteins are

also found at the centrosome, where they may affect microtubule minus-end dynamics

and anchorage (Askham et al., 2002; Quintyne et al., 1999; Rehberg and Graf, 2002). It is

possible that mDia1 interacts with one or even several of these regulatory proteins to

modulate their effect on plus- and minus-ends of microtubules. Systematic studies of

mDia1 interactions with the microtubule tip proteins are required to elucidate the

mechanism of its effect on microtubules.

Finally, it cannot be excluded that mDia1 modulate microtubule dynamics neither

directly via microtubule associated proteins, nor via actin reorganization, but by

activation of certain signaling pathway(s) upstream to microtubule regulation. In

particular, recent data suggest that mDia1 can trigger Rac activation (Tsuji et al., 2002).

Active Rac, in turn, can affect microtubule dynamics (Wittmann et al., 2003) via several

- 20 -

pathways, including Pak1-Op18/stathmin pathway (Daub et al., 2001), and IQGAP1-

CLIP-170 pathway (Fukata et al., 2002).

Irrespectively to specific mechanism of mDia1 action on the microtubule

dynamics, the resultant alterations apparently augment the microtubule effect on the focal

adhesions. It is well established that in many cell types microtubules function as triggers

of focal adhesion turnover necessary for cell migration (reviewed in (Small et al., 2002)).

In particular, direct dynamic observations revealed that even transient contacts with

microtubule tip often result in gradual disassembly of focal adhesion (Kaverina et al.,

1999). Consistently, disruption of microtubules with nocodazole or colchicine promotes

focal adhesion growth (Bershadsky et al., 1996; Enomoto, 1996; Kirchner et al., 2003;

Liu et al., 1998). Microtubules are, in general, targeted to focal adhesions (Kaverina et

al., 2002; Kaverina et al., 1998; Krylyshkina et al., 2003), and induce their disassembly,

most probably, by local inhibition of myosin II-driven contractility (Elbaum et al., 1999;

Geiger and Bershadsky, 2001; Small et al., 2002). In the present study we have shown

that mDia1-induced changes in microtubule dynamics promote microtubule interactions

with focal adhesions. In cells expressing the active form of mDia1 the amplitude of

microtubule tip oscillations is lower, than in controls, and microtubule tips spend

significantly more time in a close proximity to focal adhesions. Moreover, our

experiments with microtubule disruption by nocodazole showed that the inhibitory effect

of microtubules on focal adhesion was significantly enhanced in cells expressing active

mDia1. Thus, mDia1 plays a dual role in the focal adhesion regulation. First, it is

necessary for the focal adhesion growth induced by force (Riveline et al., 2001). Most

probably, this function of mDia1 is related to its ability to nucleate linear (non-branching)

- 21 -

assembly of actin filaments (Li and Higgs, 2003). Mechanism of force-induced focal

adhesion assembly (Bershadsky et al., 2003; Geiger and Bershadsky, 2002) contains,

however, a positive feedback loop that could lead to unlimited growth of these structures

(growth of focal adhesion induces the growth of associated stress fiber, which generates

more force, which promotes further growth of focal adhesion, and so on ). To avoid

unlimited growth, the second mode of mDia1 action seems to be critical. Namely,

mDia1-induced changes in microtubule dynamics promote microtubule targeting to focal

adhesions, which may locally inhibit myosin II-driven contractility and interrupt the

above mentioned loop. Taken together our data show that mDia1 function coordinates the

activities of two major cytoskeletal systems, actin and microtubules, in the process of

formation and turnover of focal adhesions.

- 22 -

Materials and methods

Cell culture and transfection

CHO-K1 were obtained from American Type Tissue Culture (ATCC, Rockville, MD,

USA) and cultured in Ham s F12 medium supplemented with 10% fetal calf serum

(FCS), 2mM glutamine, and antibiotics (complete medium). B16 F1 cells stably

transfected with tubulin-GFP (Ballestrem et al., 2000) were cultured in DMEM with 10%

FCS, 2mM glutamine, and antibiotics. Transient transfections were performed with

LipofectAMINE PLUS (Rhenium, Jerusalem, Israel), according to the manufacturer s

instructions.

DNA constructs and antibodies

Paxillin (cDNA kindly provided by K. Nakata, S. Miyamoto and K Matsumoto, National

Institute of Dental and Craniofacial Research, NIH, Bethesda, MD) was cloned into

ECFP-C1 and EYFP-C1 (Clontech, Palo Alto, CA, USA). The plasmids pFL-mDia∆N3,

pEGFP-mDia∆N3, pEGFP-mDia∆N3KA3 encoding FLAG-tagged and GFP-fused

constitutively active mutants of mDia1 were described in (Watanabe et al., 1999). EB1-

GFP was obtained from Dr. S. Tsukita s lab (Mimori-Kiyosue et al., 2000)

The rabbit antibody against mDia1 were characterized in (Tominaga et al., 2000).

Antibodies for α-tubulin (DM-1A) and γ-tubulin (monoclonal) were from Sigma (Sigma,

St. Lois, MO, USA), the polyclonal anti γ-tubulin was obtained from Dr. G. G. Borisy

(Northwestern University, Chicago, IL).

- 23 -

Secondary antibodies used for our studies coupled to Cy-3, Cy-5 were purchased from

Jackson Laboratories (West Grove, PA, USA), Alexa-350 was from Molecular Probes

(Molecular Probes Inc., Eugene, OR, USA).

Reagents

The ROCK inhibitor Y27632 was purchased from Calbiochem (Merk Eurolab,

Darmstadt, Germany), nocodazole, cytochalasin D, rhodamine-phalloidin, and fibronectin

were all purchased from Sigma (Sigma, St.Louis, MO, USA). Tissue culture media,

antibiotics (penicillin, streptomycin) and glutamine were obtained from Gibco (Rhenium

Ltd., Jerusalem, Israel), and the FCS was from Biological Industries (Kibbutz Beit

Haemek, Israel).

Preparation of cytoplasts

Cytoplasts were prepared as described (Rodionov et al., 1999). Briefly, cells were plated

on fibronectin coated glass coverslips and cultured in medium containing nocodazole

(1µg/ml, Sigma) and cytochalasin D (1.25µg/ml, Sigma) for 90min. Coverslips were then

centrifuged upside down at 10,000g for 25min. After centrifugation, coverslips were

placed in fresh medium or medium containing Y-27632 (10 µM, Calbiochem) for 1-2

hour to allow the recovery and reassembly of microtubules.

Fluorescence Microscopy

For immunofluorescence, cells were fixed with glutaraldehyde. After fixation cells were

rinsed with with PBS and permeabilized with 0.5% of Triton X-100 for 5 minutes. For

- 24 -

actin staining cells were subsequently incubated with 100nM Rodamine-Phalloidin, for

antibody stainings cells were incubated with primary antibodies diluted in PBS. Cells

were then times three washed in PBS and stained with secondary antibodies. After three

final washes fluorescent labeling was analyzed using an Axiovert 100 TV inverted

microscope (Zeiss, Oberkochen, Germany).

Video Microscopy

Images were recorded on an Axiovert 100 TV inverted microscope (Zeiss, Oberkochen,

Germany) equipped with a Box & Temperature control system from Life Imaging

Services (Switzerland, www.lis.ch/), a 100W mercury lamp, a 100X/1.4 plan-Neofluar

objective (Zeiss, Oberkochen, Germany), excitation and emission filter wheels, and a

CCD Camera (CH300/CE 350, Photometrics, Tucson, AZ) with KAF1400 CCD chip,

controlled by a DeltaVision system (Applied Precision Inc., Issaquah, WA, USA). Filters

for detection of GFP, YFP, and dsRed were from Chroma (Chroma Technology Corp,

Rockingham, VT, USA)

For EB1-GFP recordings, CHO-K1 cells transfected with EB1-GFP or EB1-GFP

together with mDia1∆Ν3 were plated on fibronectin coated glass-bottom dishes (MatTek

corporation, Ashland, MA, USA). After overnight incubation (24-36 hours post

transfection) cells were placed in HEPES (25mM) buffered complete Ham s F12 under

the microscope. Time-lapse recordings of transfected cells were performed in 3s

intervals. Only cells that were weakly expressing EB1-GFP were chosen for recordings.

For recording of microtubule dynamics B16F1 cells stably expressing tubulin-GFP were

super-transfected with mDia1∆Ν3 plus dsRed2 (in a 2:1 ratio, for identification of

- 25 -

transfected cells), or dsRed2 only (control cells). Cells were plated as described above

and time-lapse images were recorded in 10s intervals.

Image analysis

Images were analyzed using Openlab software (Openlab, Improvision, UK). For

microtubule velocity analysis two subsequent images taken in 3s intervals were

superimposed in alternating colors, red and green. Distances between the dots were

measured using the measurment tools of the Openlab program. Numbers were collected

and analysed in Microsoft Exel.

For the microtubule dynamics measurements, two successive images from time-

lapse recordings taken in 5s intervals were labeled in red and green and merged to

visualize the relative displacement of MT during time (see, e.g., Fig. 4). Growth and

shrinkage of microtubules were measured using measurement tools of the Openlab

software. Data were then transferred to Microsoft Excel files to calculate the indicated

dynamic instability parameters. The parameters were calculated as previously described

by (Wittmann et al., 2003). Only microtubule length changes exceeding the optical

resolution of 0.2 µm per frame were considered as growth or shortening events.

Transitions from growth to shortening and pauses to shortening (only if growth was

preceded the pause) were considered as catastrophic events, transition from shortening to

growth and shortening to pauses (only if shortening preceded the pause) were considered

as rescue events. Catastrophe (or rescue) frequencies were calculated as number of their

events divided by the time of growth (or shortening).

- 26 -

Microtubule alignment and total fiber length were calculated using the fiber score

algorithm (Lichtenstein et al., 2003). The algorithm was implemented in Priism

environment and enabled the evaluation of fiber length, density and co-alignment.

For calculations of FA area per cell, fluorescent images were subtracted from

background and a threshold was set to then apply the water algorithm as previously

described (Zamir et al., 1999). All the individual areas of focal adhesions in a cell were

summed to obtain total focal adhesion area per cell. Statistical significance of the

differences between the samples was determined using Wilcoxon rank-sum test

(Montgomery and Runger, 1999).

- 27 -

Acknowledgements

We thank Benny Geiger for stimulating discussions and critical reading of the

manuscript. This work was supported in part by grants from the Minerva Foundation,

Israesl Sciences Foundation, and USA-Israel Binational Science Foundation to Alexander

Beshadsky. We also acknowledge support from Yad Abraham Center for Cancer

Diagnostics and Therapy. A. Bershadsky holds the Joseph Moss Professional Chair in

Biomedical Research. Christoph Ballestrem was supported by Minerva.

- 28 -

References

Abal, M., M. Piel, V. Bouckson-Castaing, M. Mogensen, J.B. Sibarita, and M. Bornens.2002. Microtubule release from the centrosome in migrating cells. J Cell Biol.159:731-737.

Akhmanova, A., C.C. Hoogenraad, K. Drabek, T. Stepanova, B. Dortland, T. Verkerk,W. Vermeulen, B.M. Burgering, C.I. De Zeeuw, F. Grosveld, and N. Galjart. 2001.Clasps are CLIP-115 and -170 associating proteins involved in the regional regulationof microtubule dynamics in motile fibroblasts. Cell. 104:923-935.

Alberts, A.S. 2001. Identification of a carboxyl-terminal diaphanous-related forminhomology protein autoregulatory domain. J Biol Chem. 276:2824-2830.

Alberts, A.S. 2002. Diaphanous-related Formin homology proteins. Curr Biol. 12:R796.Askham, J.M., K.T. Vaughan, H.V. Goodson, and E.E. Morrison. 2002. Evidence that an

interaction between EB1 and p150(Glued) is required for the formation andmaintenance of a radial microtubule array anchored at the centrosome. Mol Biol Cell.13:3627-3645.

Ballestrem, C., B. Wehrle-Haller, B. Hinz, and B.A. Imhof. 2000. Actin-dependentlamellipodia formation and microtubule-dependent tail retraction control-directed cellmigration. Mol Biol Cell. 11:2999-3012.

Bershadsky, A., A. Chausovsky, E. Becker, A. Lyubimova, and B. Geiger. 1996.Involvement of microtubules in the control of adhesion-dependent signaltransduction. Curr Biol. 6:1279-1289.

Bershadsky, A.D., N.Q. Balaban, and B. Geiger. 2003. Adhesion-dependent cellmechanosensitivity. Annu Rev Cell Dev Biol. 19:677-695.

Bhatt, A., I. Kaverina, C. Otey, and A. Huttenlocher. 2002. Regulation of focal complexcomposition and disassembly by the calcium-dependent protease calpain. J Cell Sci.115:3415-3425.

Bretscher A. 2003. Polarized growth and organelle segregation in yeast: the tracks,motors, and receptors. J Cell Biol. 160:811-816.

Brunner, D., and P. Nurse. 2000. CLIP170-like tip1p spatially organizes microtubulardynamics in fission yeast. Cell. 102:695-704.

Bulinski, J.C., and G.G. Gundersen. 1991. Stabilization of post-translational modificationof microtubules during cellular morphogenesis. Bioessays. 13:285-293.

Carvalho, P., J.S. Tirnauer, and D. Pellman. 2003. Surfing on microtubule ends. TrendsCell Biol. 13:229-237.

Cassimeris, L., N.K. Pryer, and E.D. Salmon. 1988. Real-time observations ofmicrotubule dynamic instability in living cells. J Cell Biol. 107:2223-2231.

Chang, F., D. Drubin, and P. Nurse. 1997. cdc12p, a protein required for cytokinesis infission yeast, is a component of the cell division ring and interacts with profilin. JCell Biol. 137:169-182.

Chausovsky, A., A.D. Bershadsky, and G.G. Borisy. 2000. Cadherin-mediated regulationof microtubule dynamics. Nat Cell Biol. 2:797-804.

Chevrier, V., M. Piel, N. Collomb, Y. Saoudi, R. Frank, M. Paintrand, S. Narumiya, M.Bornens, and D. Job. 2002. The Rho-associated protein kinase p160ROCK is requiredfor centrosome positioning. J Cell Biol. 157:807-817.

Coquelle, F.M., M. Caspi, F.P. Cordelieres, J.P. Dompierre, D.L. Dujardin, C. Koifman,P. Martin, C.C. Hoogenraad, A. Akhmanova, N. Galjart, J.R. De Mey, and O. Reiner.

- 29 -

2002. LIS1, CLIP-170's key to the dynein/dynactin pathway. Mol Cell Biol. 22:3089-3102.

Dammermann, A., A. Desai, and K. Oegema. 2003. The minus end in sight. Curr Biol.13:R614-624.

Daub, H., K. Gevaert, J. Vandekerckhove, A. Sobel, and A. Hall. 2001. Rac/Cdc42 andp65PAK regulate the microtubule-destabilizing protein stathmin throughphosphorylation at serine 16. J Biol Chem. 276:1677-1680.

Dogterom, M., and B. Yurke. 1997. Measurement of the force-velocity relation forgrowing microtubules. Science. 278:856-860.

Elbaum, M., A. Chausovsky, E.T. Levy, M. Shtutman, and A.D. Bershadsky. 1999.Microtubule involvement in regulating cell contractility and adhesion-dependentsignalling: a possible mechanism for polarization of cell motility. Biochem Soc Symp.65:147-172.

Enomoto, T. 1996. Microtubule disruption induces the formation of actin stress fibers andfocal adhesions in cultured cells: possible involvement of the rho signal cascade. CellStruct Funct. 21:317-326.

Evangelista, M., K. Blundell, M.S. Longtine, C.J. Chow, N. Adames, J.R. Pringle, M.Peter, and C. Boone. 1997. Bni1p, a yeast formin linking cdc42p and the actincytoskeleton during polarized morphogenesis. Science. 276:118-122.

Evangelista, M., S. Zigmond, and C. Boone. 2003. Formins: signaling effectors forassembly and polarization of actin filaments. J Cell Sci. 116:2603-2611.

Feierbach, B, and F. Chang. 2001. Roles of the fission yeast formin for3p in cell polarity,actin cable formation and symmetric cell division. Curr Biol. 11:1656-1665.

Fukata, M., T. Watanabe, J. Noritake, M. Nakagawa, M. Yamaga, S. Kuroda, Y.Matsuura, A. Iwamatsu, F. Perez, and K. Kaibuchi. 2002. Rac1 and Cdc42 capturemicrotubules through IQGAP1 and CLIP-170. Cell. 109:873-885.

Geiger, B., and A. Bershadsky. 2001. Assembly and mechanosensory function of focalcontacts. Curr Opin Cell Biol. 13:584-592.

Geiger, B., and A. Bershadsky. 2002. Exploring the neighborhood: adhesion-coupled cellmechanosensors. Cell. 110:139-142.

Goode, B.L., D.G. Drubin, and G. Barnes. 2000. Functional cooperation between themicrotubule and actin cytoskeletons. Curr Opin Cell Biol. 12:63-71.

Goode, B.L., J.J. Wong, A.C. Butty, M. Peter, A.L. McCormack, J.R. Yates, D.G.Drubin, and G. Barnes. 1999. Coronin promotes the rapid assembly and cross-linkingof actin filaments and may link the actin and microtubule cytoskeletons in yeast. JCell Biol. 144:83-98.

Gundersen, G.G. 2002. Evolutionary conservation of microtubule-capture mechanisms.Nat Rev Mol Cell Biol. 3:296-304.

Howard, J., and A.A. Hyman. 2003. Dynamics and mechanics of the microtubule plusend. Nature. 422:753-758.

Hwang, E., J. Kusch, Y. Barral, and T.C. Huffaker. 2003. Spindle orientation inSaccharomyces cerevisiae depends on the transport of microtubule ends alongpolarized actin cables. J Cell Biol. 161:483-488.

Imamura, H., K. Tanaka, T. Hihara, M. Umikawa, T. Kamei, K. Takahashi, T. Sasaki,and Y. Takai. 1997. Bni1p and Bnr1p: downstream targets of the Rho family small G-

- 30 -

proteins which interact with profilin and regulate actin cytoskeleton inSaccharomyces cerevisiae. Embo J. 16:2745-2755.

Ishizaki, T., Y. Morishima, M. Okamoto, T. Furuyashiki, T. Kato, and S. Narumiya.2001. Coordination of microtubules and the actin cytoskeleton by the Rho effectormDia1. Nat Cell Biol. 3:8-14.

Janson, M.E., M.E. de Dood, and M. Dogterom. 2003. Dynamic instability ofmicrotubules is regulated by force. J Cell Biol. 161:1029-1034.

Karakesisoglou, I., Y. Yang, and E. Fuchs. 2000. An epidermal plakin that integratesactin and microtubule networks at cellular junctions. J Cell Biol. 149:195-208.

Karsenti, E., S. Kobayashi, T. Mitchison, and M. Kirschner. 1984. Role of thecentrosome in organizing the interphase microtubule array: properties of cytoplastscontaining or lacking centrosomes. J Cell Biol. 98:1763-1776.

Kaverina, I., O. Krylyshkina, K. Beningo, K. Anderson, Y.L. Wang, and J.V. Small.2002. Tensile stress stimulates microtubule outgrowth in living cells. J Cell Sci.115:2283-2291.

Kaverina, I., O. Krylyshkina, and J.V. Small. 1999. Microtubule targeting of substratecontacts promotes their relaxation and dissociation. J Cell Biol. 146:1033-1044.

Kaverina, I., K. Rottner, and J.V. Small. 1998. Targeting, capture, and stabilization ofmicrotubules at early focal adhesions. J Cell Biol. 142:181-190.

Keating, T.J., J.G. Peloquin, V.I. Rodionov, D. Momcilovic, and G.G. Borisy. 1997.Microtubule release from the centrosome. Proc Natl Acad Sci U S A. 94:5078-5083.

Kimura, K., M. Ito, M. Amano, K. Chihara, Y. Fukata, M. Nakafuku, B. Yamamori, J.Feng, T. Nakano, K. Okawa, A. Iwamatsu, and K. Kaibuchi. 1996. Regulation ofmyosin phosphatase by Rho and Rho-associated kinase (Rho- kinase). Science.273:245-248.

Kirchner, J., Z. Kam, G. Tzur, A.D. Bershadsky, and B. Geiger. 2003. Live-cellmonitoring of tyrosine phosphorylation in focal adhesions following microtubuledisruption. J Cell Sci. 116:975-986.

Komarova, Y.A., I.A. Vorobjev, and G.G. Borisy. 2002. Life cycle of MTs: persistentgrowth in the cell interior, asymmetric transition frequencies and effects of the cellboundary. J Cell Sci. 115:3527-3539.

Kovar, D.R., J.R. Kuhn, A.L. Tichy, and T.D. Pollard. 2003. The fission yeastcytokinesis formin Cdc12p is a barbed end actin filament capping protein gated byprofilin. J Cell Biol. 161:875-887.

Kreis, T.E. 1987. Microtubules containing detyrosinated tubulin are less dynamic. EmboJ. 6:2597-2606.

Krylyshkina, O., K.I. Anderson, I. Kaverina, I. Upmann, D.J. Manstein, J.V. Small, andD.K. Toomre. 2003. Nanometer targeting of microtubules to focal adhesions. J CellBiol. 161:853-859.

Lee, S., and P.A. Kolodziej. 2002. Short Stop provides an essential link between F-actinand microtubules during axon extension. Development. 129:1195-1204.

Leung, C.L., K.J. Green, and R.K. Liem. 2002. Plakins: a family of versatile cytolinkerproteins. Trends Cell Biol. 12:37-45.

Li, F., and H.N. Higgs. 2003. The mouse Formin mDia1 is a potent actin nucleationfactor regulated by autoinhibition. Curr Biol. 13:1335-1340.

- 31 -

Lichtenstein, N., B. Geiger, and Z. Kam. 2003. Quantitative analysis of cytoskeletalorganization by digital fluorescent microscopy. Cytometry. 54A:8-18.

Ligon, L.A., S.S. Shelly, M. Tokito, and E.L. Holzbaur. 2003. The microtubule plus-endproteins EB1 and dynactin have differential effects on microtubule polymerization.Mol Biol Cell. 14:1405-1417.

Liu, B.P., M. Chrzanowska-Wodnicka, and K. Burridge. 1998. Microtubuledepolymerization induces stress fibers, focal adhesions, and DNA synthesis via theGTP-binding protein Rho. Cell Adhes Commun. 5:249-255.

Mialhe, A., L. Lafanechere, I. Treilleux, N. Peloux, C. Dumontet, A. Bremond, M.H.Panh, R. Payan, J. Wehland, R.L. Margolis, and D. Job. 2001. Tubulin detyrosinationis a frequent occurrence in breast cancers of poor prognosis. Cancer Res. 61:5024-5027.

Mimori-Kiyosue, Y., N. Shiina, and S. Tsukita. 2000. The dynamic behavior of the APC-binding protein EB1 on the distal ends of microtubules. Curr Biol. 10:865-868.

Montgomery, D. and G. Runger, 1999. Applied statistics and probability for engineers.Second Edition. New York: John Wiley & Sons.

Nakamura, M., X.Z. Zhou, and K.P. Lu. 2001. Critical role for the EB1 and APCinteraction in the regulation of microtubule polymerization. Curr Biol. 11:1062-1067.

Nakano, K., J. Imai, R. Arai, E.A. Toh, Y. Matsui, and I. Mabuchi. 2002. The smallGTPase Rho3 and the diaphanous/formin For3 function in polarized cell growth infission yeast. J Cell Sci. 115:4629-4639.

Palazzo, A.F., T.A. Cook, A.S. Alberts, and G.G. Gundersen. 2001. mDia mediates Rho-regulated formation and orientation of stable microtubules. Nat Cell Biol. 3:723-729.

Peng, J., B.J. Wallar, A. Flanders, P.J. Swiatek, and A.S. Alberts. 2003. Disruption of theDiaphanous-Related Formin Drf1 Gene Encoding mDia1 Reveals a Role for Drf3 asan Effector for Cdc42. Curr Biol. 13:534-545.

Perez, F., G.S. Diamantopoulos, R. Stalder, and T.E. Kreis. 1999. CLIP-170 highlightsgrowing microtubule ends in vivo. Cell. 96:517-527.

Pletjushkina, O.J., A.M. Belkin, O.J. Ivanova, T, Oliver, J.M. Vasiliev, and K. Jacobson1998. Maturation of cell-substratum focal adhesions induced by depolymerization ofmicrotubules is mediated by increased cortical tension. Cell Adhes. Commun. 5:121-35

Pring, M., M. Evangelista, C. Boone, C. Yang, and S.H. Zigmond. 2003. Mechanism offormin-induced nucleation of actin filaments. Biochemistry. 42:486-496.

Pruyne, D., M. Evangelista, C. Yang, E. Bi, S. Zigmond, A. Bretscher, and C. Boone.2002. Role of formins in actin assembly: nucleation and barbed-end association.Science. 297:612-615.

Quintyne, N.J., S.R. Gill, D.M. Eckley, C.L. Crego, D.A. Compton, and T.A. Schroer.1999. Dynactin is required for microtubule anchoring at centrosomes. J Cell Biol.147:321-334.

Rehberg, M., and R. Graf. 2002. Dictyostelium EB1 is a genuine centrosomal componentrequired for proper spindle formation. Mol Biol Cell. 13:2301-2310.

Ridley, A.J., and A. Hall. 1992. The small GTP-binding protein rho regulates theassembly of focal adhesions and actin stress fibers in response to growth factors. Cell.70:389-399.

- 32 -

Riveline, D., E. Zamir, N.Q. Balaban, U.S. Schwarz, T. Ishizaki, S. Narumiya, Z. Kam,B. Geiger, and A.D. Bershadsky. 2001. Focal contacts as mechanosensors: externallyapplied local mechanical force induces growth of focal contacts by an mDia1-dependent and ROCK-independent mechanism. J Cell Biol. 153:1175-1186.

Rodionov, V., E. Nadezhdina, and G. Borisy. 1999. Centrosomal control of microtubuledynamics. Proc Natl Acad Sci U S A. 96:115-120.

Rodriguez, O.C., A.W. Schaefer, C.A. Mandato, P. Forscher, W.M. Bement, and C.M.Waterman-Storer. 2003. Conserved microtubule-actin interactions in cell movementand morphogenesis. Nat Cell Biol. 5:599-609.

Rothenberg, M.E., S.L. Rogers, R.D. Vale, L.Y. Jan, and Y.N. Jan. 2003. Drosophilapod-1 crosslinks both actin and microtubules and controls the targeting of axons.Neuron. 39:779-791.

Sagot, I., A.A. Rodal, J. Moseley, B.L. Goode, and D. Pellman. 2002. An actin nucleationmechanism mediated by Bni1 and profilin. Nat Cell Biol. 4:626-631.

Sammak, P.J., and G.G. Borisy. 1988. Direct observation of microtubule dynamics inliving cells. Nature. 332:724-726.

Shelden, E., and P. Wadsworth. 1993. Observation and quantification of individualmicrotubule behavior in vivo: microtubule dynamics are cell-type specific. J CellBiol. 120:935-945.

Small, J.V., B. Geiger, I. Kaverina, and A. Bershadsky. 2002. Opinion: How domicrotubules guide migrating cells? Nat Rev Mol Cell Biol. 3:957-964.

Subramanian, A., A. Prokop, M. Yamamoto, K. Sugimura, T. Uemura, J. Betschinger,J.A. Knoblich, and T. Volk. 2003. Shortstop recruits EB1/APC1 and promotesmicrotubule assembly at the muscle-tendon junction. Curr Biol. 13:1086-1095.

Tirnauer, J.S., S. Grego, E.D. Salmon, and T.J. Mitchison. 2002. EB1-microtubuleinteractions in Xenopus egg extracts: role of EB1 in microtubule stabilization andmechanisms of targeting to microtubules. Mol Biol Cell. 13:3614-3626.

Tominaga, T., E. Sahai, P. Chardin, F. McCormick, S.A. Courtneidge, and A.S. Alberts.2000. Diaphanous-related formins bridge Rho GTPase and Src tyrosine kinasesignaling. Mol Cell. 5:13-25.

Totsukawa, G., Y. Yamakita, S. Yamashiro, D.J. Hartshorne, Y. Sasaki, and F.Matsumura. 2000. Distinct roles of ROCK (Rho-kinase) and MLCK in spatialregulation of MLC phosphorylation for assembly of stress fibers and focal adhesionsin 3T3 fibroblasts. J Cell Biol. 150:797-806.

Tsuji, T., T. Ishizaki, M. Okamoto, C. Higashida, K. Kimura, T. Furuyashiki, Y.Arakawa, R.B. Birge, T. Nakamoto, H. Hirai, and S. Narumiya. 2002. ROCK andmDia1 antagonize in Rho-dependent Rac activation in Swiss 3T3 fibroblasts. J CellBiol. 157:819-830.

Uehata, M., T. Ishizaki, H. Satoh, T. Ono, T. Kawahara, T. Morishita, H. Tamakawa, K.Yamagami, J. Inui, M. Maekawa, and S. Narumiya. 1997. Calcium sensitization ofsmooth muscle mediated by a Rho-associated protein kinase in hypertension. Nature.389:990-994.

Vaughan, K.T., S.H. Tynan, N.E. Faulkner, C.J. Echeverri, and R.B. Vallee. 1999.Colocalization of cytoplasmic dynein with dynactin and CLIP-170 at microtubuledistal ends. J Cell Sci. 112:1437-1447.

- 33 -

Vorobjev, I.A., V.I. Rodionov, I.V. Maly, and G.G. Borisy. 1999. Contribution of plusand minus end pathways to microtubule turnover. J Cell Sci. 112:2277-2289.

Wallar, B.J., and A.S. Alberts. 2003. The formins: active scaffolds that remodel thecytoskeleton. Trends Cell Biol. 13:435-446.

Watanabe, N., T. Kato, A. Fujita, T. Ishizaki, and S. Narumiya. 1999. Cooperationbetween mDia1 and ROCK in Rho-induced actin reorganization. Nat Cell Biol.1:136-143.

Watanabe, N., P. Madaule, T. Reid, T. Ishizaki, G. Watanabe, A. Kakizuka, Y. Saito, K.Nakao, B.M. Jockusch, and S. Narumiya. 1997. p140mDia, a mammalian homolog ofDrosophila diaphanous, is a target protein for Rho small GTPase and is a ligand forprofilin. Embo J. 16:3044-3056.

Webster, D.R., G.G. Gundersen, J.C. Bulinski, and G.G. Borisy. 1987. Differentialturnover of tyrosinated and detyrosinated microtubules. Proc Natl Acad Sci U S A.84:9040-9044.

Webster, D.R., J. Wehland, K. Weber, and G.G. Borisy. 1990. Detyrosination of alphatubulin does not stabilize microtubules in vivo. J Cell Biol. 111:113-122.

Wittmann, T., G.M. Bokoch, and C.M. Waterman-Storer. 2003. Regulation of leadingedge microtubule and actin dynamics downstream of Rac1. J Cell Biol. 161:845-851.

Zamir, E., B.Z. Katz, S. Aota, K.M. Yamada, B. Geiger, and Z. Kam. 1999. Molecular

diversity of cell-matrix adhesions. J Cell Sci. 112:1655-1669.

Zigmond SH, Evangelista M, Boone C, Yang C, Dar AC, Sicheri F, Forkey J, Pring M.

2003. Formin leaky cap allows elongation in the presence of tight capping proteins.

Curr. Biol. 13:1820-1823.

- 34 -

Table 1: Microtubule dynamic instability parameters in control and mDia1∆N3

expressing B16F1 cells

Parameter control mDia1(∆N3)

Elongation rate (µm/min) 6.43 – 2.42 4.06 – 1.66

Shortening rate (µm/min) 11,13 – 9.32 5.61 – 3.48

Catastrophe frequency (s-1) 0.084 0.089

Rescue frequency (s-1) 0.104 0.105

Time spent Growing (%) 41 – 11 35 – 17

Pausing (%) 28 – 12 36 – 18

Shortening(%) 31 – 9 29 – 16

MT survival for 90 s (%) 65 95

Parameters characterizing microtubule dynamics were measured at the periphery of cellsstably expressing GFP-tubulin. MT survival denotes the fraction of the microtubules,which remained in the same field of view from the beginning to the end of the 90 secondobservation period. The data are presented as mean – standard deviation (SD).

- 35 -

Table 2. Microtubule interactions with focal adhesions

Cell type FA # MT number in outControl 1 5 11 8

2 2 4 4

3 3 5 3

4 3 6 5

5 3 6 5

6 2 4 3

7 2 4 4

8 4 8 6mDia1∆N3- 1 2 2 0transfected 2 2 2 0

3 3 3 1

4 3 3 0

5 2 2 0

6 3 3 0

7 2 2 0

8 2 3 0

9 1 1 0

Acts of encounter and separation between microtubule tips and individual focal adhesionsfor a period of 90 s are presented. Each row corresponds to an individual focal adhesionselected in control or mDia1∆N3-transfected B16F1 cells expressing GFP-tubulin andRFP-zyxin. MT number corresponds to the total number of different microtubules thatoverlapped with the 2µm size square box enclosing the given focal adhesion during theperiod of observation (see Fig 8). in means number of microtubule tips that were foundin this box in the beginning of observation plus number of microtubule entering eventsregistered during the 90 s observation period. out denotes the number of timesmicrotubule tips went out from the box during the period of observation.

- 36 -

Figure Legends

Fig 1. Active mDia induces actin polymerization. (A) CHO cells were transfected withcDNA encoding a constitutively active form of mDia1 (mDia1(∆N3)) fused to GFP and(B) stained with rhodamine-phalloidin to visualize polymerized actin (F-actin). Note thatthe cell expressing GFP-mDia1(∆N3) has a bipolar shape with aligned actin fibers alongthe long axis. (C) The contents of F-actin estimated by the intensity of rhodamine-phalloidin fluorescence is dramatically enhanced in mDia expressing cells compared tothe surrounding non-transfected control cells.Bar, 20 µm.

Fig 2. Microtubule total mass and alignment is enhanced in mDia1(∆N3) expressingcells. (A) Control CHO cell. (B, C) CHO cells expressing GFP-mDia1(∆N3). (A,B) Anti-tubulin staining, (C) GFP mDia1(∆N3). Bar, 8 µm. (D) Quantification of microtubulealignment demonstrates an about four-fold increase in alignment of microtubules inmDia1(∆N3) expressing cells as compared to control cells . (E) The total length ofmicrotubules in mDia1(∆N3) expressing cells is about 30% higher than in control cells.

Fig 3. Microtubule velocities decrease in mDia1(∆N3) cells (see also video 1-4). Tomeasure the microtubule growth rates CHO cells were transfected with EB-1-GFP(control) or EB1-GFP plus mDia1(∆N3). EB1-GFP localizes at growing microtubule tipsin control (A) and mDia1(∆N3) expressing cells (D). Superimposition of time-lapseimages of EB1-GFP in alternating colors allows detailed analysis of microtubulebehavior (B; C; E; F). (C) and (F) are magnified inserts of (B) and (E), respectively. Barin (B) 8µm, in (C) 3µm. Microtubules in control cells grow with a rate of about 23µm/min and it is decreased by about 50% in cells expressing active mDia (G). The graphsof microtubule end displacement versus time (G) show that in control cells individualmicrotubules grow with a fairly constant growth rate, whereas microtubules in mDia1expressing cells grow with alternating speeds. Growth rate distributions in mDia1(∆N3)expressing cells (H) is shifted to smaller values and more broad than in control cells.

Fig 4. Microtubule oscillations at the cell periphery (see also videos 5 and 6). B16F1cells stably expressing tubulin-GFP were transfected with control plasmids (A) or cDNAencoding active mDia1 (B); the transfected cell in (B) is indicated by the white arrowwhile the non-transfected cell is indicated by the black arrow. To visualize microtubuleelongation, shortening and lateral movements two images of time-lapse recordings takenin 5s intervals were labeled in green and red and then superimposed using Openlabsoftware. Structures that remained at the same place during the 5s interval are seen inyellow. Elongation of microtubules at the cell periphery is seen in red and shortenings ingreen (arrows in A). When microtubules shift laterally during the time of observation thecontour of the same microtubule can be seen in green and red (see control cells in (A) and(B)). Note that microtubule elongation, shortening and lateral movements are lesspronounced in cells expressing active mDia1 than in control cells (see transfected cell in(B)). Bar, 3 µm. Examples of microtubule tip displacement versus time for fourmicrotubules from control and active mDia1 expressing cells are shown in (C) and (D).

- 37 -

Note that the amplitude of microtubule oscillations is decreased in cells with activemDia1 (D) as compared to those in control cells (C).

Fig 5. mDia localization at the centrosome. CHO cells were transfected with cDNAencoding indicated forms of mDia1 fused to GFP (A; D). 24 hours after transfection cellswere fixed and centrosomes were stained with antibodies against γ-tubulin (B; E). Notethat mDia1(∆N3) and mDia1(FH2) co-localize with γ-tubulin at the centrosome (A-F).(G) Native mDia1 localizes at the centrosome in B16 melanoma cells as visualized usingantibodies against mDia1(G) and γ-tubulin (H). Overlay in (I). Bar, 10 µm. Inserts showcentrosomal region of cells in higher magnification.

Fig 6. Increased amount and enhanced alignment of microtubules in centrosome-freecytoplasts from cells expressing actvive mDia1. Cytoplasts from CHO cells transfectedwith indicated plasmids (A: a, e, i) were prepared and microtubules were stained usingantibodies against α-tubulin (A:b, f, j). Cytoplasts used were centrosome-free asmanifested by stainings for γ-tubulin (A: d, h, l). Images of microtubule staining wereprocessed (A: c, g, k) and quantification of microtubule alignment and total microtubulelength was done as described in materials and methods. Note that microtubules alignalong a bipolar axis in cytoplasts expressing GFP-mDia1(∆N3) (A:a-c) but not incytoplasts expressing EGFP (e-g) or a mutated form of the active mDia1, GFP-mDia1(∆N3KA3) (i-k). Quantification reveals an increased microtubule alignment (B)and total microtubule legth (C) in mDia1(∆N3) expressing cytoplast. In cytoplastscontaining the mDia1 mutant (∆N3KA3), microtubule alignment and total microtubulelength were similar to those in control cells. (D) Microtubule plus end growth decreasedby about 50% in cytoplasts containing active mDia1, irrespective of the presence of thecentrosome.

Fig 7. ROCK together with active mDia1 influence minus end stability of microtubulesbut does not affect plus end dynamics. (A) Active mDia1 induces a decrease ofmicrotubule growth rate as revealed by GFP-EB1 marker and this decrease is notabolished by inhibition of ROCK by Y27632. (B) In centrosome-free cytoplastsmDia1(∆N3) increases total mass of microtubules. Inhibition of ROCK in these cells leadto a decrease of microtubule length to levels found in control cells. (C) The presence ofactive mDia1 leads to a dramatic increase of actin polymerization, which is not affectedby inhibition of ROCK.

Fig 8. Active mDia1 alters the mode of microtubule interactions with focal adhesions.B16F1 cells stably expressing tubulin-GFP were transfected with (A) dsRed-Zyxin or (B)dsRed-Zyxin plus mDia1(∆N3). Sequence of microtubule images presented in highermagnification in the lower part of the figure outline the dynamics of the region of boxesin (A) and (B). Time lapse images of 10s intervals show that microtubules inmDia1(∆N3) expressing cells remain for longer times in proximity (box, 2 µm size) offocal adhesions than highly oscillating microtubules of control cells (see videos 7 and 8).

- 38 -

Fig 9. Effect of mDia on focal adhesions. CHO cells were transfected with paxillin-YFPonly (A and B) or paxillin-YFP together with mDia1(∆N3) (C and D). Images ofuntreated cells (A, C) or cells treated with 10 µM nocodazole for 45 min (B, D) weretaken 24 hours post transfection. Focal adhesions in control cells expressing onlypaxillin-YFP(A) increase moderately in size when microtubules were disrupted (B).Focal adhesions increased in size upon co-expression of mDia1(∆N3) (C) anddramatically further increase when microtubules in these cells were disrupted (D). Bar,10 µm. Quantification of total focal adhesion area per cell (E).

Video Supplement

Video1: Microtubule plus end growth in CHO K1 cells transfected with EB1-GFP.Images were taken in 3s intervals for the period of 120 seconds and played back at 5frames/s. For still image, see Fig3A

Video2: Microtubule plus end growth in CHO K1 cells transfected with EB1-GFP. Time-lapse images taken in 3s intervals were staked in alternating colors. Stacking of imagesoutlines the microtubule plus end displacement in a period of 120s. For still image, seeFig 3B and C

Video3: Microtubule plus end growth in CHO K1 cells transfected with EB1-GFP andmDia1(∆N3). Images were taken in 3s intervals for the period of 120 seconds and playedback at 5 frames/s. For still image, see Fig3D

Video4: Microtubule plus end growth in CHO K1 cells transfected with EB1-GFP andmDia1(∆N3). Time-lapse images taken in 3s intervals were staked in alternating colors.Stacking of images outlines the microtubule plus end displacement in a period of 120s.Compare with video2. For still image, see Fig 3B and C.

Video5: Microtubule dynamics in B16F1 cells. B16F1 cells expressing GFP-tubulin weresupertransfected with dsRed and microtubule dynamics were recorded in 5s intervals forthe period of 90 seconds. To outline the dynamic behavior of microtubules two images 5sapart were colored in red and green and merged. Thus, yellow indicates microtubuleoverlay, red polymerization events (or lateral movement), and green depolymerizationevents (or lateral movement). Images were played back at 5 frames/s. For still image, seeFig 4A. Compare with microtubule dynamics of mDia1(∆N3).

Video6: Microtubule dynamics in B16F1 cells expressing active mDia1. B16 F1 cellsexpressing GFP-tubulin were transfected with mDia1(∆N3) together with dsRed toidentify transfected cells. Microtubule dynamics were recorded in 5s intervals for theperiod of 90 seconds. First frame of movie indicates microtubules in green andmicrotubules in red to identify active mDia1 expressing and control cell in the field ofview. To outline the dynamic behavior of microtubules two images 5s apart were coloredin red and green and merged. Thus, yellow indicates microtubule overlay, redpolymerization events (or lateral movement), and green depolymerization events (or

- 39 -

lateral movement). Images were played back at 5 frames/s. For still image, see Fig 4B.Note the dramatic decrease in oscillations and lateral movements of microtubules in cellexpressing active mDia1 versus the control cell (compare also with Video5).

Video7: Microtubule targeting to focal adhesions in B16F1 cells. Cells stably expressingGFP-tubulin were transfected with cDNA encoding dsRed-Zyxin, a marker of focaladhesions. Time-lapse images were recorded in 5s intervals for the period of 90 seconds.Images were played back at 5 frames/s. Note that microtubules oscillate in highfrequency in the proximity of focal adhesions. This is outlined in the section depicted inthe upper left corner. The small box in this section is 2µm large. Microtubules enter andleave this zone with high frequency.

Video8: Microtubule targeting to focal adhesions in B16F1 cells expressing activemDia1. Cells stably expressing GFP-tubulin were transfected with cDNA encoding anactive form of mDia1 and dsRed-Zyxin, a marker of focal adhesions. Time-lapse imageswere recorded in 5s intervals for the period of 90 seconds. Images were played back at 5frames/s. Note that microtubules remain relatively stable in the proximity of focaladhesions. This is outlined in the section depicted in the upper left corner. The small boxin this section is 2µm large. Compare with high frequency microtubule oscillations in theproximity of focal adhesions in control cells (video7).

- 1 -

Video Supplement

Video1: Microtubule plus end growth in CHO K1 cells transfected with EB1-GFP. Images were taken in 3s intervals for the period of 120 seconds and played back at 5 frames/s. For still image, see Fig3A

Video2: Microtubule plus end growth in CHO K1 cells transfected with EB1-GFP. Time-lapse images taken in 3s intervals were staked in alternating colors. Stacking of images outlines the microtubule plus end displacement in a period of 120s. For still image, see Fig 3B and C

Video3: Microtubule plus end growth in CHO K1 cells transfected with EB1-GFP and mDia1(∆N3). Images were taken in 3s intervals for the period of 120 seconds and played back at 5 frames/s. For still image, see Fig3D

Video4: Microtubule plus end growth in CHO K1 cells transfected with EB1-GFP and mDia1(∆N3). Time-lapse images taken in 3s intervals were staked in alternating colors. Stacking of images outlines the microtubule plus end displacement in a period of 120s. Compare with video2. For still image, see Fig 3B and C.

Video5: Microtubule dynamics in B16F1 cells. B16F1 cells expressing GFP-tubulin were supertransfected with dsRed and microtubule dynamics were recorded in 5s intervals for the period of 90 seconds. To outline the dynamic behavior of microtubules two images 5s apart were colored in red and green and merged. Thus, yellow indicates microtubule overlay, red polymerization events (or lateral movement), and green depolymerization events (or lateral movement). Images were played back at 5 frames/s. For still image, see Fig 4A. Compare with microtubule dynamics of mDia1(∆N3).

Video6: Microtubule dynamics in B16F1 cells expressing active mDia1. B16 F1 cells expressing GFP-tubulin were supertransfected with mDia1(∆N3) together with dsRed to identify transfected cells. Microtubule dynamics were recorded in 5s intervals for the period of 90 seconds. First frame of movie indicates microtubules in green and microtubules in red to identify active mDia1 expressing and control cell in the field of view. To outline the dynamic behavior of microtubules two images 5s apart were colored in red and green and merged. Thus, yellow indicates microtubule overlay, red polymerization events (or lateral movement), and green depolymerization events (or lateral movement). Images were played back at 5 frames/s. For still image, see Fig 4B. Note the dramatic decrease in oscillations and lateral movements of microtubules in cell expressing active mDia1 versus the control cell (compare also with Video5).

Video7: Microtubule targeting to focal adhesions in B16F1 cells. Cells stably expressing GFP-tubulin were supertransfected with cDNA encoding dsRed-Zyxin, a marker of focal adhesions. Time-lapse images were recorded in 5s intervals for the period of 90 seconds. Images were played back at 5 frames/s. Note that microtubules oscillate in high frequency in the proximity of focal adhesions. This is outlined in the section depicted in

- 2 -

the upper left corner. The small box in this section is 2µm large. Microtubules enter and leave this zone with high frequency.

Video8: Microtubule targeting to focal adhesions in B16F1 cells expressing active mDia1. Cells stably expressing GFP-tubulin were supertransfected with cDNA encoding an active form of mDia1 and dsRed-Zyxin, a marker of focal adhesions. Time-lapse images were recorded in 5s intervals for the period of 90 seconds. Images were played back at 5 frames/s. Note that microtubules remain relatively stable in the proximity of focal adhesions. This is outlined in the section depicted in the upper left corner. The small box in this section is 2µm large. Compare with high frequency microtubule oscillations in the proximity of focal adhesions in control cells (video7).