FEBS Letters 582 (2008) 2875–2881

Atomic force microscopy reveals binding of mRNA to microtubulesmediated by two major mRNP proteins YB-1 and PABP

Konstantin G. Chernova,b,1, Patrick A. Curmib,1, Loic Hamonb, Alain Mechulamb,Lev P. Ovchinnikova, David Pastreb,*

a Institute of Protein Research, Russian Academy of Sciences, Pushchino, Moscow Region 142290, Russiab Laboratoire Structure-Activite des Biomolecules Normales et Pathologiques, INSERM/UEVE U829, Evry 91025, France

Received 17 April 2008; revised 25 June 2008; accepted 11 July 2008

Available online 22 July 2008

Edited by Michael R. Bubb

Abstract A significant fraction of mRNAs is known to be asso-ciated in the form of mRNPs with microtubules for active trans-port. However, little is known about the interaction betweenmRNPs and microtubules and most of previous works were fo-cused on molecular motor:microtubule interactions. Here, wehave identified, via high resolution atomic force microscopyimaging, a significant binding of mRNA to microtubules medi-ated by two major mRNP proteins, YB-1 and PABP. This inter-action with microtubules could be of critical importance foractive mRNP traffic and for mRNP granule formation. A similarrole may be fulfilled by other cationic mRNA partners.� 2008 Federation of European Biochemical Societies. Publishedby Elsevier B.V. All rights reserved.

Keywords: Microtubule; RNA; RNP; Atomic forcemicroscopy; YB-1; PABP

1. Introduction

In the cytoplasm, mRNAs interact with numerous protein

partners [1–5], which enable the packaging of mRNAs into

mRNPs and participate in the mRNA translational regulation,

protection and anchoring to cytoskeleton [6]. Because of hin-

dered diffusion in the cytoplasm [7], the efficient traffic of

mRNPs is possible only through active transport on the cyto-

skeleton and particularly along microtubules by molecular mo-

tors [1]. In agreement with this, fluorescence tracking of

mRNA shows probabilistic movements made of pauses, direc-

tional slidings and periods of free diffusion after detachments

[8–10]. As detachments from microtubules slow down trans-

port kinetics, transported mRNPs are thought to stay within

a short distance from microtubules by an unclear mechanism

[9]. An interesting explanation is that several motors interact

with microtubules to ensure a longer lifetime of mRNP bind-

ing [9,11,12]. However, the implication of other mechanisms

remains to be tackled.

Recently, a large scale identification of tubulin and microtu-

bule binding proteins showed that 21% of them were known

*Corresponding author. Fax: +33 1 69 47 02 19.E-mail addresses: david.pastre@univ-evry.fr, dvpastre@yahoo.fr

(D. Pastre).

1Equal contribution.

0014-5793/$34.00 � 2008 Federation of European Biochemical Societies. Pu

doi:10.1016/j.febslet.2008.07.019

mRNA binding proteins like PABP [13]. Previous studies also

demonstrated that mRNA binding partners were good candi-

dates for tubulin binding [14]. Most of the mRNP proteins,

especially the core mRNP proteins like YB-1, are highly cat-

ionic (pI � 9.5 for YB-1) [15]. Consequently, they are prone

to interact with highly anionic microtubules (charge per tubu-

lin at neutral pH being about 20–30e�). In regard to this, we

wondered whether cationic mRNP proteins can induce a rele-

vant mRNP binding to microtubules. Here, we show by AFM

imaging that two major core mRNP proteins (YB-1 and

PABP) provoke an attraction between mRNPs and microtu-

bules. The mechanism of this attraction was also further ana-

lyzed by gel shift and co-precipitation assays. We then suggest

that mRNA binding partners other than molecular motors

could facilitate the active transport of mRNPs and mRNA

granule formation.

2. Materials and methods

Unless stated otherwise, all chemicals used in this work were pur-chased from Sigma–Aldrich (France).

2.1. YB-1 and PABP purificationRecombinant YB-1 and N-terminal His-tagged PABP were ex-

pressed in Escherichia coli and purified as previously described in[16,17], respectively. Purified proteins were dialyzed against 200 mMNaCl, 20 mM HEPES–KOH, pH 7.6, 1 mM DTT and stored at�80 �C.

2.2. Synthesis of mRNAPlasmids pET-28a-a-globin and pSP72-2Luc were used as templates

for synthesis by T7 polymerase of a-globin (660 nt) and 2Luc mRNAs(3000 nt), respectively [18]. After transcription, unincorporated NTPswere removed by gel-filtration through a NAP-5 column (GE Health-care) and mRNAs were further isolated with RNAble (Eurobio) fol-lowing manufacturer�s recommendation.

2.3. Tubulin purification, microtubule preparationTubulin was purified from sheep brain using the method by Castoldi

and Popov [19] and stored at �80 �C in 50 mM MES–KOH, pH 6.8,0.5 mM dithiothreitol, 0.5 mM EGTA, 0.25 mM MgCl2, 10% glycerol,and 0.1 mM GTP. Before use, tubulin stock was thawed and an addi-tional cycle of polymerization was performed.

For microtubule assembly, 20 lM tubulin was incubated in 200 ll ofassembly buffer (50 mM MES–KOH, pH 6.8, 1 mM EGTA, 5 mMMgCl2, 20% glycerol, 1 mM GTP) containing 20 lM taxol for15 min at 37 �C. Microtubules were then pelleted at 25000 · g for10 min at 37 �C and gently resuspended in a starting volume of assem-bly buffer.

blished by Elsevier B.V. All rights reserved.

2876 K.G. Chernov et al. / FEBS Letters 582 (2008) 2875–2881

2.4. RNP sample preparationYB-1:mRNA or PABP:mRNA complexes were first preformed in

the assembly buffer for 10 min at 37 �C. Preformed microtubules werethen added to the samples at various concentrations, and the mixtureswere incubated for 10 min at 37 �C.

2.5. Atomic force microscopyAtomic force microscopy is a powerful tool to study biomolecules,

especially DNA [20] and RNA [21]. Its application to the imaging ofmicrotubules is less common but some interesting studies were al-ready published [22,23]. No special requirements are necessary to ob-tain the results presented in this work. Ten microliters of each samplewas deposited on freshly cleaved mica and dried for AFM imaging asdescribed by Pastre et al. [24]. The electrostatic adsorption of both

Fig. 1. (I) High resolution AFM imaging of RNP:microtubule association:secondary structure of RNA. (b) RNP particles obtained by mixing 5 lg/mlthe YB-1:RNA complexes was easily distinguished. (c) 2Luc RNA (5 lg/ml)that microtubules and mRNA did not significantly interact with each other.polymerized tubulin). RNPs clearly co-localized with microtubules and tendemicrotubules to form bundles. (II) Co-sedimentation assays with a-globin mYB-1 (R = 1/10). (e) mRNA staining with Sybr Green II. (f) Protein stainingthe pellet in a dose dependent manner, which clearly indicated that RNPs werbuffer (50 mM MES–KOH, pH 6.8, 1 mM EGTA, 5 mM MgCl2, 20% glyce

protein:mRNA complexes on mica was then mediated by divalentmagnesium cations. All AFM experiments were performed in inter-mittent mode with a multimode AFM instrument (Digital Instru-ments, Veeco, Santa Barbara, CA) operating with a Nanoscope IIIacontroller. We used AC160TS silicon cantilevers (Olympus, Ham-burg, Germany) with resonance frequencies of around 300 kHz. Theapplied force was minimized as much as possible. Images werecollected at a scan frequency of 1.5 Hz and a resolution of512 · 512 pixels.

2.6. Native agarose gel electrophoresisSamples were loaded onto 0.6% agarose gels prepared using

microtubule-stabilizing buffer (50 mM MES–KOH, pH 6.8, 1 mMEGTA, 2 mM MgCl2, 10% glycerol) to avoid microtubule

(a) 2Luc mRNA (5 lg/ml). A proper spreading on mica revealed the2Luc RNA and 1.5 lM YB-1. The typical beads-on-string structure ofin the presence of microtubules (0.8 lM polymerized tubulin) showing(d) RNPs (5 lg/ml of RNA) in the presence of microtubules (0.8 lM

d to cluster on microtubule walls. In addition RNPs can also cross-linkRNA (5 lg/ml) and microtubules in the presence or absence of 1.5 lMwith Coomassie Blue. Microtubules increased the amount of RNPs ine bound to microtubules. The experiments were performed in assemblyrol, 1 mM GTP).

K.G. Chernov et al. / FEBS Letters 582 (2008) 2875–2881 2877

depolymerization during migration. The gel was run at 4 V/cm for 4 hat 10 �C, and mRNA was revealed with Sybr Green II for 30 min. Ina second time, gel was stained with Coomassie blue to analyze pro-teins.

Fig. 2. Influence of mRNA saturation with YB-1 on the interaction of RNPs(50 mM MES–KOH, pH 6.8, 1 mM EGTA, 5 mM MgCl2, 20% glycerol, 1 mM3) or with varying concentrations of YB-1 (lanes 4–12) to obtain different mtubulin (lanes 2, 5, 8, and 11) or taxol-stabilized microtubules (lanes 3, 6, 9, ananalyzed by native agarose gel electrophoresis. (a) RNA detection with Sybconditions, microtubules remained trapped in the well due to their size (paneand 10). The presence of non-polymerized tubulin or microtubules did not sThe presence of non-polymerized tubulin increased the complex mobility (cstabilized microtubules produced a similar effect at R = 1/30 and R = 1/15 (lan15, a fraction of RNPs was retained in the well (panel a, lane 9), while at R =that YB-1 promotes mRNA binding to microtubules at high R values ( P 1/(panel b, lanes 7 and 10, migration towards cathode). (II) AFM imaging ofRNPs (2Luc mRNA) with microtubules occurred only at high YB-1:mRNA

2.7. Co-sedimentation assaysRNP complexes were preformed in assembly buffer and preformed

microtubules were added at different concentrations. Each sample(400 ll) was centrifuged at 25000 · g for 10 min at 37 �C to pellet

with microtubules. The experiments were performed in assembly bufferGTP). (I) 50 lg/ml of a-globin mRNA was incubated alone (lanes 1–

olar ratios of YB-1 per RNA base, R. After incubation with YB-1,d 12) was added to a final concentration of 1 lM tubulin. Samples werer Green II. (b) Proteins detection with Coomassie blue staining. In alll b). Mobilities of RNPs alone decreased with increasing R (lanes 4, 7

ignificantly influence the mobility of mRNA (cf. lanes 2 and 3 with 1).f. lane 4 with 5; lane 7 with 8; lane 10 with 11), and the presence ofes 6, 8, and 11). Interestingly, in the presence of microtubules at R = 1/

1/7.5, RNA totally bound to microtubules (lane 12). This demonstrates15). We also noticed the presence of free YB-1 under such conditionsRNP:microtubule association at various R values. The association ofratio, in agreement with the results of gel shift assays.

2878 K.G. Chernov et al. / FEBS Letters 582 (2008) 2875–2881

microtubules. Supernatants were collected, and pellets were resus-pended in the starting volume of assembly buffer. Twenty microlitersfrom each fraction were analyzed on Coomassie stained 9% SDS–PAGE gel. For experiments requiring mRNA quantification, mRNAwas extracted with RNAble from the remaining of each sample(380 ll) and loaded on 2% agarose formaldehyde denaturing gel. Gelswere equilibrated in 50 mM Tris–borate, pH 8.0, run at 4 V/cm for 1 hthen stained with Sybr Green II.

3. Results

3.1. YB-1 mediates mRNA binding to microtubules

We first performed high resolution imaging of YB-1:RNA

complexes on mica by atomic force microscopy. As shown pre-

viously [18], YB-1 leads to the packaging of mRNA into small

RNP particles containing several nucleoprotein globules,

resulting in a beads-on-string structure (Fig. 1I). mRNA alone

did not interact with microtubules whereas RNP particles were

Fig. 3. (I) YB-1:microtubule interaction is of electrostatic origin. The bindingsedimentation assays. YB-1 started to be released from the microtubule surfbinding can take place under physiological conditions. However, as the bounmost probably electrostatic. I was adjusted by adding KCl to the assembneutralization of microtubules. Microtubule co-sedimentation assays revealecationic molecules like poly-LL-lysine (MW between 15 and 30 kDa for this eximplicated in the attraction of YB-1 to microtubules. (III) Influence of ionicYB-1-containing RNPs (2 lg/ml a-globin mRNA at R = 1/4) and microtubuionic strength led to a progressive release of RNPs from microtubules, in ag

clearly associated with microtubules (Fig. 1I(c) and (d)). The

distribution of RNPs along microtubules was not uniform,

in agreement with the tendency of YB-1 containing RNPs to

self-aggregate, especially when mRNA strands are saturated

with YB-1 (see Fig. 4).

To further explore the RNP binding to microtubules, co-pre-

cipitation assays of RNPs with microtubules were performed

(Fig. 1II). In control, mRNA did not co-sediment with micro-

tubules, and only a small fraction of RNPs was found in the

pellet in the absence of microtubules. By comparison, RNPs

clearly co-sedimented with microtubules with an efficiency

increasing with microtubule mass.

3.2. Influence of the saturation of mRNA with YB-1 on its

binding to microtubules

To better understand the mechanism of interaction between

microtubules and RNPs, we varied the YB-1:globulin mRNA

of YB-1 to microtubules at various ionic strengths was analyzed by co-ace at ionic strengths (I) larger than 200 mM, which indicates that thisd proteins were nearly completely removed at I = 0.6 M, this binding isly buffer. (II) Competition between YB-1 and poly-LL-lysine for thed that YB-1 can be displaced from the microtubule surface by highlyperiment). Therefore, non-specific electrostatic interactions are mostlystrength on RNP:microtubule interaction. Co-sedimentation assays ofles (0.8 lM polymerized tubulin) at various ionic strengths. Increasingreement with an electrostatic interaction mediated by YB-1.

K.G. Chernov et al. / FEBS Letters 582 (2008) 2875–2881 2879

base ratio, R, from 1/30 to 1/7.5, corresponding to undersatu-

rated and oversaturated mRNA, respectively [18]. The interac-

tion between RNPs and microtubules was then analyzed by

non-denaturing gel electrophoresis (Fig. 2I). With mRNA alone

no clear differences in the migration response were observed in

the presence or absence of tubulin or microtubules, indicating

that mRNA did not interact either with tubulin or with microtu-

bules. Addition of increasing amounts of YB-1 gradually re-

duced the complex mobility (cf. lanes 4, 7 and 10) probably

due to mRNA neutralization by YB-1 and to an increase in com-

plex size. Interestingly, the presence of free tubulin increased the

complex mobility (cf. lanes 4 and 5, 7 and 8, 10 and 11), and the

presence of stabilized microtubules produced a similar effect at

R = 1/30 and R = 1/15. This indicates that either free or poly-

merized tubulins can compete with mRNA for YB-1 binding.

Most importantly, in the presence of microtubules, at R = 1/

15, a small RNP fraction stayed in the well in agreement with

its binding to microtubules (lane 9). For R = 1/7.5, this effect

is more pronounced and all RNPs were then bound to microtu-

bules and thus remained in the well (lane 12). Finally, we de-

tected the presence of free YB-1 at high R values (Fig. 2I(b),

lanes 7 and 10, migration towards cathode), which should play

an important role in the RNA:microtubule binding.

The influence of mRNA saturation with YB-1 was also inves-

tigated by AFM (Fig. 2II). Increasing YB-1 concentration also

led to a more efficient binding of RNPs to microtubules. At

high R values (R = 1/7.5), RNPs were strongly associated to

microtubules on which they tended to form large aggregates.

Fig. 4. (I) Microtubules favor RNP aggregation. 2Luc mRNA (12 lg/ml) wadiluted in assembly buffer to indicated mRNA concentrations. In the abseconcentrations P 6 lg/ml, whereas in the presence of 0.25 lM polymerized tmRNA concentration. The apparition of very large RNP aggregates on micgranules. (II) RNP-microtubule binding mediated by PABP RNPs assemblewithout (a) or with (b) microtubules (0.25 lM tubulin). We also observed th

3.3. Electrostatic origin of RNP:microtubule interaction and

other mRNA partners

We first tested whether the binding of YB-1 to microtubules,

which is necessary for the binding of mRNP to microtubule, is

specific. Co-sedimentation assays indicate that the YB-

1:microtubule interaction is weakened at high ionic strength

(Fig. 3I) and that bound YB-1 can be removed from the micro-

tubule surface by poly-LL-lysine (Fig. 3II), a strong cationic

competitor. The YB-1:microtubule interaction is then mainly

non-specific and electrostatic, but the point is that the binding

of YB-1 to microtubules remains unaffected at physiological

concentrations of monovalent salt (I < 0.2 M). Regarding the

binding of RNPs to microtubules, we found that, even though

it was gradually reduced when the ionic strength increased, the

binding is still significant at physiological concentrations of

monovalent salt (Fig. 3III). The electrostatic origin of YB-

1:microtubule and RNP:microtubule interactions suggests that

YB-1 can form a salt bridge between microtubule and mRNA

anionic surfaces.

To decipher whether the ability of YB-1 to mediate microtu-

bule:RNA association was unique or shared by other cationic

mRNA binding proteins, we evaluated the possibility that

PABP, a major core RNP protein can also mediate such bind-

ing (Fig. 4II). We found that PABP also provoked RNP bind-

ing to microtubules at a high PABP:RNA ratios. This indicates

that other cytoplasmic partners of mRNA may play a similar

role that of YB-1 and participate to the binding of mRNPs to

microtubules.

s incubated with 3.6 lM YB-1 to form saturated RNPs. Samples werence of microtubules, large RNP aggregates only occured at mRNAubulin, large aggregates of RNPs were formed on microtubules at lowrotubules indicates that microtubules promote the formation of RNPd in the presence of 5 lg/ml 2Luc mRNA and 1 lM PABP for 10 mine formation of RNP granules on microtubules.

Fig. 5. Schematic representation of the mechanisms of RNA:microtubule attraction mediated by cationic mRNA binding proteins. (a) There is noattraction between RNPs and microtubules when RNA is not saturated with protein partners like YB-1. (i) Proteins partners are then strongly boundto RNA as described in [18]. (b) Extra protein partner mediates RNP:microtubule attraction via (ii) salt bridges or (iii) protein:protein interactions.These mechanisms can also promote RNP self-aggregation on microtubules (iv).

2880 K.G. Chernov et al. / FEBS Letters 582 (2008) 2875–2881

3.4. RNP granule formation

One tempting assumption is that microtubules may serve as

tracks to assemble RNP particles into larger RNP aggregates

like RNP granules observed in vivo [25]. RNP aggregation oc-

curred at high R values, i.e. when extra YB-1 was available for

RNP:RNP self-association. Indeed, it is known that the aggre-

gation of RNPs is a result of a dynamic equilibrium [26] and,

below a critical concentration of self-attracting particles, mul-

timolecular aggregation hardly occurs [27]. In agreement with

this, we observed here by AFM that highly diluted RNPs

(3 lg/ml mRNA) were not likely to form high molecular

weight aggregates (Fig. 4I), while an increase of the RNP con-

centration (�6 lg/ml RNA) triggered the apparition of large,

fractal-like structures. In the presence of microtubules, RNP

aggregates were observed at a significantly lower concentration

of RNPs (�2 lg/ml RNA) and were mainly co-localized with

microtubules, suggesting that microtubules promote RNP

aggregation on their surfaces.

4. Discussion

We observed for the first time a direct binding of RNPs to

microtubules mediated by two major core mRNP proteins,

YB-1 and PABP, using co-sedimentation, gel shift assays and

high resolution imaging by AFM. The RNP:microtubule inter-

action seems to be mainly of electrostatic origin and related to

the positive charges of these mRNA partners, which can neu-

tralize both anionic tubulin and mRNA. However, our results

show that the binding of RNPs to microtubules is preserved at

physiological salt concentrations and thus could be significant

in vivo. Protein:protein interactions can also participate in the

RNA: microtubule attraction, as many mRNA partners, like

YB-1, are able to self-aggregate via multimerization domains.

In agreement with this proposal, we observed that RNP par-

ticles should be saturated with YB-1 to induce a significant

RNP binding to microtubules. At low YB-1:RNA ratios, most

of these proteins are bound via high affinity sites to RNA and

buried inside RNPs which most probably make difficult their

binding to microtubules. However, with mRNA over-satu-

rated with YB-1, extra YB-1 are highly mobile and thus avail-

able both for electrostatic interactions via salt bridges between

RNP and microtubules and (or) for YB-1:YB-1 interactions

between free RNPs and those adsorbed on microtubules

(Fig. 5).

We finally captured by AFM the tendency of RNP particles

to self-associate on microtubules. First some microtubules

were depleted of RNPs, whereas others were densely coated

(Fig. 2II). Second microtubules led to the formation of large

RNP aggregates (Fig. 4), in accordance with a previous report

showing that RNA granule formation is microtubule depen-

dent [25]. Together, our results show that the binding of

RNP particles on microtubules, eventually followed by self-

aggregation upon colliding with one another on microtubules

via random diffusion, can be an interesting model of facilitated

aggregation (Fig. 5). Besides, direct RNPs binding to microtu-

bules, even though weaker than molecular motor:microtubule

interactions, could keep RNPs on the microtubule surface

when molecular motors are transiently unconnected to micro-

tubules and thus contribute to RNP trafficking.

Acknowledgements: We thank Dr. Elena Nadezhina for comments onthe manuscript. The study was supported in part by a grant on‘‘Molecular and Cellular Biology’’ Program from Russian Academyof Sciences (K.G.C. and L.P.O.). The authors thank INSERM, TheConseil Regional d�Ile de France; the ‘‘Service pour la Science, la Tech-nologie et l�Espace (SSTE)’’ from the French Embassy at Moscow,Russia, Genopole Evry for constant support, and the Association pourla Recherche sur le Cancer (ARC).

References

[1] Palacios, I.M. and St Johnston, D. (2001) Getting the messageacross: the intracellular localization of mRNAs in higher eukary-otes. Annu. Rev. Cell Dev. Biol. 17, 569–614.

[2] Czaplinski, K. and Singer, R.H. (2006) Pathways for mRNAlocalization in the cytoplasm. Trends Biochem. Sci. 31, 687–693.

[3] Anderson, P. and Kedersha, N. (2006) RNA granules. J. CellBiol. 172, 803–808.

K.G. Chernov et al. / FEBS Letters 582 (2008) 2875–2881 2881

[4] Kiebler, M.A. and Bassell, G.J. (2006) Neuronal RNA granules:movers and makers. Neuron 51, 685–690.

[5] Kanai, Y., Dohmae, N. and Hirokawa, N. (2004) Kinesintransports RNA: isolation and characterization of an RNA-transporting granule. Neuron 43, 513–525.

[6] Davis, I. (2004) A helicase that gets Oskar�s message across. Nat.Cell Biol. 6, 285–287.

[7] David-Cordonnier, M.H., Payet, D., D�Halluin, J.C., Waring,M.J., Travers, A.A. and Bailly, C. (1999) The DNA-bindingdomain of human c-Abl tyrosine kinase promotes the interactionof a HMG chromosomal protein with DNA. Nucleic Acids Res.27, 2265–2270.

[8] Fusco, D., Accornero, N., Lavoie, B., Shenoy, S.M., Blanchard,J.M., Singer, R.H. and Bertrand, E. (2003) Single mRNAmolecules demonstrate probabilistic movement in living mamma-lian cells. Curr. Biol. 13, 161–167.

[9] Carson, J.H., Cui, H. and Barbarese, E. (2001) The balance ofpower in RNA trafficking. Curr. Opin. Neurobiol. 11, 558–563.

[10] Chang, L., Shav-Tal, Y., Trcek, T., Singer, R.H. and Goldman,R.D. (2006) Assembling an intermediate filament network bydynamic cotranslation. J. Cell Biol. 172, 747–758.

[11] Welte, M.A. (2004) Bidirectional transport along microtubules.Curr. Biol. 14, R525–R537.

[12] Bullock, S.L., Zicha, D. and Ish-Horowicz, D. (2003) TheDrosophila hairy RNA localization signal modulates the kineticsof cytoplasmic mRNA transport. Embo J. 22, 2484–2494.

[13] Chuong, S.D., Good, A.G., Taylor, G.J., Freeman, M.C.,Moorhead, G.B. and Muench, D.G. (2004) Large-scale identifi-cation of tubulin-binding proteins provides insight on subcellulartrafficking, metabolic channeling, and signaling in plant cells.Mol. Cell Proteomics 3, 970–983.

[14] Hamill, D., Davis, J., Drawbridge, J. and Suprenant, K.A. (1994)Polyribosome targeting to microtubules: enrichment of specificmRNAs in a reconstituted microtubule preparation from seaurchin embryos. J. Cell Biol. 127, 973–984.

[15] Minich, W.B., Maidebura, I.P. and Ovchinnikov, L.P. (1993)Purification and characterization of the major 50-kDa repressorprotein from cytoplasmic mRNP of rabbit reticulocytes. Eur. J.Biochem. 212, 633–638.

[16] Evdokimova, V., Ruzanov, P., Imataka, H., Raught, B., Svitkin,Y., Ovchinnikov, L.P. and Sonenberg, N. (2001) The majormRNA-associated protein YB-1 is a potent 5 0 cap-dependentmRNA stabilizer. EMBO J. 20, 5491–5502.

[17] Khaleghpour, K., Kahvejian, A., De Crescenzo, G., Roy, G.,Svitkin, Y.V., Imataka, H., O�Connor-McCourt, M. and Sonen-berg, N. (2001) Dual interactions of the translational repressorPaip2 with poly(A) binding protein. Mol. Cell Biol. 21, 5200–5213.

[18] Skabkin, M.A., Kiselyova, O.I., Chernov, K.G., Sorokin, A.V.,Dubrovin, E.V., Yaminsky, I.V., Vasiliev, V.D. and Ovchinnikov,L.P. (2004) Structural organization of mRNA complexes withmajor core mRNP protein YB-1. Nucleic Acids Res. 32, 5621–5635.

[19] Castoldi, M. and Popov, A.V. (2003) Purification of brain tubulinthrough two cycles of polymerization-depolymerization in a high-molarity buffer. Protein Expr. Purif. 32, 83–88.

[20] Sorel, I., Pietrement, O., Hamon, L., Baconnais, S., Cam, E.L.and Pastre, D. (2006) The EcoRI–DNA complex as a model forinvestigating protein–DNA interactions by atomic force micros-copy. Biochemistry 45, 14675–14682.

[21] Hansma, H.G., Golan, R., Hsieh, W., Daubendiek, S.L. andKool, E.T. (1999) Polymerase activities and RNA structures inthe atomic force microscope. J. Struct. Biol. 127, 240–247.

[22] Vater, W., Fritzsche, W., Schaper, A., Bohm, K.J., Unger, E. andJovin, T.M. (1995) Scanning force microscopy of microtubulesand polymorphic tubulin assemblies in air and in liquid. J. CellSci. 108 (Pt 3), 1063–1069.

[23] Elie-Caille, C., Severin, F., Helenius, J., Howard, J., Muller, D.J.and Hyman, A.A. (2007) Straight GDP-tubulin protofilamentsform in the presence of taxol. Curr. Biol. 17, 1765–1770.

[24] Pastre, D., Hamon, L., Landousy, F., Sorel, I., David, M.O.,Zozime, A., Le Cam, E. and Pietrement, O. (2006) Anionicpolyelectrolyte adsorption on mica mediated by multivalentcations: a solution to dna imaging by atomic force microscopyunder high ionic strengths. Langmuir 22, 6651–6660.

[25] Ivanov, P.A., Chudinova, E.M. and Nadezhdina, E.S. (2003)Disruption of microtubules inhibits cytoplasmic ribonucleopro-tein stress granule formation. Exp. Cell Res. 290, 227–233.

[26] Kedersha, N., Cho, M.R., Li, W., Yacono, P.W., Chen, S., Gilks,N., Golan, D.E. and Anderson, P. (2000) Dynamic shuttling ofTIA-1 accompanies the recruitment of mRNA to mammalianstress granules. J. Cell Biol. 151, 1257–1268.

[27] Bentz, J. and Nir, S. (1981) Aggregation of colloidal particlesmodeled as a dynamical process. Proc. Natl. Acad. Sci. USA 78,1634–1637.