Geometrical Control of Actin Assembly and Contractility

Anne-Cécile Reymann; Christophe Guérin; Manuel Théry; Laurent Blanchoin; Rajaa Boujemaa-Paterski    Laboratoire de Physiologie Cellulaire et Végétale, Institut de Recherches en Technologies et Sciences pour le Vivant, iRTSV, CNRS/CEA/INRA/UJF, Grenoble, France

Introduction

Actin is a major cell component and the actin cytoskeleton is fundamental for many cellular processes ranging from morphogenesis to cellular motility (Pollard & Borisy, 2003). The actin network in cells displays distinct architectures specialized for well-defined functions. These architectures can be divided into three main classes. (i) A branched entangled and dense actin network localized underneath the plasma membrane, generating protrusive force production (Svitkina & Borisy, 1999; Urban, Jacob, Nemethova, Resch, & Small, 2010; Vinzenz et al., 2012; Weichsel, Urban, Small, & Schwarz, 2012). (ii) Antiparallel actin bundles running across the cell cytoplasm. These are contractile structures hosting myosin molecular motors and linked to the plasma membrane via focal adhesions (Gardel et al., 2008; Hotulainen & Lappalainen, 2006; Naumanen, Lappalainen, & Hotulainen, 2008; Takeya, Taniguchi, Narumiya, & Sumimoto, 2008; Vicente-Manzanares, Ma, Adelstein, & Horwitz, 2009). (iii) Parallel actin bundles are present in the dorsal area of cells and in the “linear protrusions” of the plasma membrane called filopodia (Nicholson-Dykstra & Higgs, 2008; Yang & Svitkina, 2011). They are not contractile structures (Block et al., 2008; Hotulainen et al., 2009; Mattila & Lappalainen, 2008; Schirenbeck, Bretschneider, Arasada, Schleicher, & Faix, 2005; Steffen et al., 2006; Yang et al., 2007). All these specialized actin structures overlap within the cell volume and are controlled by more than 70 regulatory protein families (Pollard & Borisy, 2003). However due the extraordinary cellular complexity (Xu, Babcock, & Zhuang, 2012), it is difficult to study in detail and decouple the biochemical and physical laws that govern the dynamics of actin networks. During the past decades, in vitro simplified reconstituted systems, in bulk solution (Pollard & Borisy, 2003) or using pathogens and functionalized particles (Cameron, Footer, Van Oudenaarden, & Theriot, 1999; Frishknecht et al., 1999; Giardini, Fletcher, & Theriot, 2003; Loisel, Boujemaa, Pantaloni, & Carlier, 1999; Noireaux et al., 2000; Pantaloni, Boujemaa, Didry, Gounon, & Carlier, 2000; Reymann et al., 2011; Romero et al., 2004) overcame the cell complexity and significantly improved our understanding of actin dynamics. Unfortunately, these reconstituted systems were unable to investigate to which extent geometrical confinement may be a determinant for actin network self-organization into specific architectures.

In vivo studies have reported that actin substructures are often associated with a specific cortege of regulatory factors, which are yet available in the entire cell volume. Even if it is now well established that the recruitment of many actin regulators is under tight biochemical control including the nucleotide state of actin subunits or the exclusive interactions of some actin regulators (Pollard & Borisy, 2003), the extent to which network structural organization plays a role in these selective interactions is still unclear. More specifically, recent cell studies reported the crucial role of myosin in actin network turnover during cell motility. Although the motor protein is available in the entire cell, myosin had been shown to localize, contract and disassemble specific actin substructures (Burnette et al., 2011; Vallotton, Gupton, Waterman-Storer, & Danuser, 2004; Wilson et al., 2010). Therefore, further investigations remain to be carried out to shed light on the biochemical or physical factors controlling selective recruitment of regulators on defined actin organizations.

Here, we describe the first method to assemble structured actin networks on micropatterns and analyze how geometrical confinement governs actin network organization (Reymann et al., 2010). Micropatterning allows grafting of nucleation promoting factors (NPFs) on a wide repertoire of geometries. These innovative biomimetic systems enable the reconstitution of branched, parallel, and antiparallel actin organizations observed in cells (Reymann et al., 2010). Furthermore, as micropatterning reconstitutes these three actin organizations separated in space, it interestingly permits the investigation of the extent geometrical actin filament organization can govern the selective recruitment of specific regulatory proteins. Here, we describe the method we used to assess myosin-induced selective contraction and the cross-linking by alpha-actinin of defined actin structures.

2.1 Deep UV Micropatterning Method: Geometry Rules for Actin Network Organization

We present here a simple and straightforward method allowing the polymerization in a reproducible fashion of actin on micropatterned features. It consists in printing adhesive micropatterns on a uniformly repellent layer. These patterns are then selectively coated with NPFs triggering actin filament assembly while repellent areas remain bare of actin nucleation (Fig. 2.1). This protocol is an adapted version of a former one published by Azioune and collaborators describing in details the procedures used to micropattern glass surfaces for cell adhesion (Azioune, Storch, Bornens, Thery, & Piel, 2009). While keeping the same technique for the design of the chrome mask, we hereafter describe precisely a novel protocol specifically adapted to design functionalized micropatterned motifs required for actin polymerization in vitro.

2.1.1 Designing the chrome mask and nucleation geometries

Photomasks. The creation of a proper photomask starts with some inventiveness and a careful design of individual motifs and their judicious juxtaposition. There are several software packages available on the market and any of them permitting a drawing with precise and clear indications of size will allow the mask producing company to convert the file into the proper format prior to its production. We use Clewin Software. For deep UV radiation, the photomask needs to be transparent to wavelengths below 200 nm. The material required for that is fused silica or synthetic quartz. The photomask has a specific resolution provided by the company that usually is around 1 μm. This resolution and the quality of the tight contact between the surface to pattern and the photomask during the UV radiation are two parameters that control the size of the features to be printed onto the photomask (Azioune et al., 2009). We typically draw features of 3 μm in width, which allows a remarkable size and shape reproducibility with accuracy.

Designing features. First of all, bear in mind that the choice of NPFs and experimental conditions are also to be taken into account while choosing the width of motifs. A large nucleation region will need several primer nucleation events to trigger the covering of the whole surface, such history could impact both the timing and homogeneity of the nucleation geometry as well as modifying the outcome network (Achard et al., 2010). In opposite a very thin/small zone might not be very reproducibly covered.

Second, the surface area prone to actin nucleation (transparent zone) should not be designed too large, in order to avoid...