Research

An overview of our current projects

Our lab takes a multi-disciplinary approach to studying the cytoskeleton, a highly dynamic network of interconnected fibers found in all living cells (eukaryotic and prokaryotic). We are focused on actin filaments, one of the three principal systems of cytoskeletal fibers, and how their polymerization dynamics and spatial organization drives key cellular processes, such as endocytosis, intracellular transport, polarized cell growth and cytokinesis.

Specifically, we use genetics, biochemistry, cell imaging and structural techniques to dissect the functions of large biological machines that control actin cytoskeletal dynamics. These include protein complexes that nucleate actin filament assembly (Arp2/3 complex and formins) and promote rapid turnover of actin filaments (Srv2/CAP, cofilin, Aip1 and profilin).

To read more about a specific area of our research, select one of the links below or scroll down the page.

Research overview | Regulation networks | Arp2/3 complex | Actin turnover | Formin proteins

Introduction and Broad Goals

Model of the actin cytoskeleton All of the diverse actin structures found in cells are assembled from the same basic building blocks (globular actin monomers) into filaments, which are then organized by actin associated proteins into specialized arrays with distinct architectures and dynamic properties. Our long-term goal is to obtain a comprehensive view of the mechanisms controlling the assembly and turnover of actin filament networks.

One approach that we are using to this end is to define the activities and interactions among all proteins that comprise the S. cerevisiae actin cytoskeleton. In many organisms, gaining a comprehensive view of actin regulation represents a daunting challenge, due to the vast numbers of actin-associated factors, the presence of multiple isoforms of each, and the limited genetic tools available for in vivo analyses. However, Saccharomyces cerevisiae (budding yeast) is an ideal model organism for this endeavor, as yeast has a core set of actin-associated proteins (some of which are shown in the figure at right), the majority of which have obvious mammalian counterparts (view table). Working in budding yeast also allows us to combine advanced genetic and biochemical approaches, so we are able to define protein mechanisms and functions while testing their physiological relevance in parallel.

In recent years, incredible progress has been made towards understanding actin assembly and turnover in yeast. Dual-label imaging studies in live cells have begun to define the spatially and temporally ordered steps in maturation of cortical actin patches, actin cables and the cytokinetic ring. These studies show that actin structures are assembled in a hierarchal manner, and that different actin structures in cells are nucleated by the two known actin nucleators, Arp2/3 complex and formins. Our lab is focused on determining the precise biochemical mechanisms and molecular interactions underlying these events. Importantly, while we are actively dissecting the functions of such central players as the Arp2/3 complex, cofilin and formins, we are also committed to defining the functions of all yeast actin-associated proteins, including those that are non-essential (e.g. Crn1, Twf1, Abp1, Aip1, Scp1). This is necessary to obtain a complete understanding of this complex biological machine.

Overhauling the Yeast Actin Cytoskeleton

Actin associated proteins - MonoQ An important tool that seeded many of the research projects in our lab was development of an assay that reconstitutes actin assembly in soluble yeast extracts (Goodman et al., in preparation). This offers unprecedented biochemical access to a system that until now has been investigated predominantly through genetics, and allows us to tie together the strengths of genetic and biochemical approaches.

We have used the assay to isolate native actin-associated protein (AAP) mixtures and identify all the components using tandem mass spectrometry (in collaboration with Dr. John Yates of the Scripps Research Institute), to identify new activities arising from interactions, and to delineate pathways regulating actin assembly upstream of the Arp2/3 complex. The strength of this approach is that it allows us to test protein function (quantitatively) within complex cellular mixtures, where in vivo binding partners (view map) are present at physiological ratios. By fractionating the mixtures of actin-associated proteins on gel filtration columns, we have identified new functional complexes, including Abp1-Arp2/3 (Goode et al., 2001; Quintero et al., submitted), Crn1-Arp2/3 (Humphries and Balcer et al., 2002), and Srv2/CAP-actin complex (Balcer et al., 2003; Mattila et al., 2004).

An actin monomer A second approach we are taking to understand actin regulation on a broad scale is to define all possible genetic interactions between 25 mutated surfaces on yeast actin (shown at right) and 17 overexpressed yeast actin binding proteins. This screen has revealed many new points of contact between actin and actin binding proteins and/or novel functional interactions among actin binding proteins. We are testing these interactions biochemically using purified actin mutants and purified actin binding proteins. Through this combined genetic and biochemical approach, we are defining a map of the overlapping binding surfaces of different actin binding proteins on actin. This map will be instrumental for understanding how the activities of numerous actin binding proteins are coordinated to control actin dynamics.

Mechanism and regulation of Arp2/3 complex

Crystal structure of the Arp2/3 complex How new actin filaments are nucleated is a central question in cell biology, since actin filaments must be generated rapidly in response to cellular cues to direct such processes as cell migration, endocytosis and cytokinesis. Cells require nucleation machinery to direct de novo actin filament assembly, and this machinery is under tight spatial and temporal regulation. Only two actin nucleators have been identified to date, the Arp2/3 complex and formins, and both are highly conserved across eukaryotes.

The actin-related protein (Arp)2/3 complex is comprised of seven conserved subunits, two of which structurally resemble actin and likely serves as the direct template for actin polymerization. Purified Arp2/3 complex alone, however, is relatively inactive for filament nucleation and requires interaction with a nucleation-promoting factor (NPF) to achieve efficient actin nucleation. Many different NPFs have now been identified, but members of the WASp/SCAR/WAVE family remain the best characterized.

Abp1, a novel NPF of the Arp2/3 complex

Abp1 crystal structure One of the first discoveries stemming from our reconstituted actin assay was the identification of Abp1 as an NPF for Arp2/3 complex (Goode et al., 2001). Previously, Abp1 was considered to be a static linker between actin filaments and specific components of the endocytic machinery. We found that Abp1 activates the Arp2/3 complex by a mechanism distinct from class I NPFs such as Las17/WASp; now Abp1 and its mammalian relative cortactin define class II NPFs (Welch and Mullins, 2002) Recently, our collaborators (Dr. Steve Almo, Einstein College of Medicine) solved the crystal structure of the actin-binding domain of Abp1. The structure is very similar to ADF/cofilins, despite the very different effects on actin dynamics these two proteins have. We have used this structure to dissect the Abp1-Arp2/3 complex activation mechanism by mutational analyses and demonstrate its importance in vivo (Quintero et al., submitted).

Coronin, the first direct inhibitor of the Arp2/3 complex

Figure from Avi's paper Coronin is a widely conserved actin binding protein that localizes to sites of dynamic actin assembly in cells. Coronin mutants in Dictyostelium have defects in cell migration, endocytosis, and cytokinesis (deHostos et al., 1993), and expression of dominant fragments of coronin in Xenopus cell lines causes defective cell migration and spreading (Mishima and Nishida, 1999). We identified and characterized the S. cerevisiae homolog of coronin, defining its biochemical effects on actin and genetic interactions with other cytoskeletal factors (Goode et al., 1999). Collectively, these studies suggested that coronin played some role in regulating actin dynamics, but the precise role remained unclear.

A key breakthrough came when we discovered that coronin binds to and directly inhibits actin nucleation by the Arp2/3 complex (Humphries and Balcer et al., 2002), as measured in kinetic assays for actin assembly (view kinetics). The inhibition activity by coronin was overridden by the addition of preexisting actin filaments. This suggested that coronin may help restrict actin nucleation by the Arp2/3 complex to the sides of filaments, thereby promoting formation of highly branched networks.

This activity of coronin defines a novel mechanism for regulating the Arp2/3 complex, which we have been dissecting further. Instrumental in these efforts has been our collaboration with Dr. Niko Grigorieff (Brandeis University), using EM and single particle analysis to solve the coronin-bound Arp2/3 structure. These images show that coronin associates with the p35 subunit of the Arp2/3 complex (consistent with two hybrid data) to stablilize Arp2/3 complex in an open (inactive) conformation (view data).

Many open questions about coronin mechanism and function remain: (1) How does the coronin coiled coil domain interact with the p35 subunit of the Arp2/3 complex to stabilize its open conformation? (2) How does coronin affect Arp2/3-dependent branching of filaments? (3) Does coronin have direct effects on actin filaments?

A molecular dissection of Arp2/3 complex regulation

Arp2/3 complex
crystal structure We have used collections of mutant alleles in three different subunits of the complex, Arp2, p35, and p40, to dissect Arp2/3 complex mechanism and regulation. Our studies on Arp2 alleles have led to the isolation of a novel negative regulator of Arp2/3 and revealed an antagonistic relationship between specific NPFs in vivo (D'Agostino et al., in preparation). Our studies on p40 alleles have defined four functional surfaces on this subunit, each independently required for cell viability. This points to a complex mechanism of action for p40 during actin nucleation, involving coordination of two other subunits and actin (Balcer et al., in preparation).

Figure from Avi's paper Our analyses of p35 mutants have shown that this subunit plays a crucial role in mediating the conformational changes in Arp2/3 complex leading to actin nucleation (Rodal et al., 2004). We defined the activities of Arp2/3 complex isolated from yeast strains expressing 17 different temperature sensitive alleles of p35. The mutant Arp2/3 complexes each showed one of two distinct classes of biochemical defects: impaired in WASp-induced actin nucleation or constitutive ("leaky") for actin nucleation in the absence of any NPF.

We used EM and single particle imaging techniques to examine structure and found that both the yeast and bovine Arp2/3 complex exist in equilibrium among three conformational states: open, intermediate and closed. The inactive crystal structure of Arp2/3 complex docked well into 3D reconstructions of the open structure. WASp bound over the cleft between Arp2 and Arp3, and all WASp-bound complexes were closed. The inhibitor coronin bound to p35, and all coronin-bound complexes were open. "Leaky" mutants (e.g., arc35-5) were skewed in distribution towards the closed conformation, whereas "nucleation impaired" mutants (e.g., arc35-6) were skewed towards the open conformation. We conclude from this work that WASp and p35 drive closure of the complex, bringing Arp2 and Arp3 together to nucleate actin (Rodal et al., 2004).

Control of actin filament turnover

Model of filament
turnover Maintenance of an assembly-competent cellular pool of actin monomers for new filament growth, requires older actin filaments to be rapidly disassembled, and the released ADP-actin monomers recycled to the ATP-bound state. This process is referred to as filament turnover. Two conserved regulators of actin turnover are cofilin, which severs and depolymerizes filaments, and profilin, which promotes nucleotide exchange on actin monomers. Upon filament disassembly, there is rapid accumulation of cofilin-bound ADP-actin monomers, which are inhibited from undergoing nucleotide exchange.

Srv2, a high molecular weight middleman for filament turnover

Profilin has low affinity for ADP-actin and thus cannot compete efficiently with cofilin for ADP-actin binding. This poses a dilemma for the rapid recycling of actin monomers from the cofilin-bound ADP state to the profilin-bound ATP state. However, we recently identified Srv2/CAP (cyclase associated protein), a highly conserved and ubiquitously expressed cellular factor that acts as a "middleman" in this process. From isolated native actin-associated protein mixtures, we purified a 600kDa complex, 30-35 nm in length by rotary shadowing EM, comprised of only Srv2 and actin in equal molar ratios. We demonstrated that the Srv2 complex is linked to actin filaments via Abp1. We also showed that Srv2 inhibits pointed end growth and promotes cofilin-dependent actin turnover by recycling cofilin from ADP-actin monomers in vitro and in vivo (Balcer et al., 2003). In close collaboration with the lab of Dr. Pekka Lappalainen (University of Helsinki, Finland), we defined the actin-binding site on Srv2 and showed that Srv2 binds with 100-fold higher affinity to ADP-G-actin than ATP-G-actin and competes with cofilin for actin binding (Mattila et al., 2004). Thus, Srv2 functions as a middleman in actin turnover, between profilin and cofilin.

Next, we must determine how Srv2 and actin molecules (estimated to be six of each) are organized into this enormous, symmetrical structure. Further, we must determine how this complex interacts with at least four other actin-binding proteins (Abp1, profilin, cofilin and Aip1) to coordinate rapid actin turnover. We are attempting to define the atomic structure of the complex and using specific mutations in Srv2, cofilin, profilin and Aip1 to uncouple their network of physical/functional associations.

Aip1: Sharpening the cofilin shears

Aip1 crystal structure Aip1 is another protein that plays a role in actin turnover. In 1999, two groups found both in vivo and in vitro evidence that Aip1 enhances cofilin-mediated actin filament disassembly in vitro (Rodal et al. 1999, Iida and Yahara 1999). Aip1 is thought to co-localize with cofilin-coated actin filaments, and may play a role in directing cofilin's activity; cofilin mislocalizes to actin cables in aip1 null cells.

One possible mechanism of Aip1 action which our lab is investigating is that Aip1 caps filaments that have recently been severed by cofilin (as shown in Okada et al. 2002), serving to inhibit both the re-annealing of filaments and new filament growth at those ends. In this way, Aip1 would preserve the effects of cofilin activity and therefore appear to enhance the rate of severing. There is also evidence that Aip1 may directly enhance the severing and/or depolymerizing activity of cofilin and/or have some severing activity by itself (Ono et al., 2004).

Our lab is currently using a mutagenesis approach to map the binding sites for cofilin and actin on Aip1. We are also using this approach to dissect the mechanism by which cofilin and Aip1 work together to sever and cap actin filaments, and to test the specific functions of Aip1 in vivo, in combination with mutations to cofilin and Srv2.

Mechanism and regulation of formins

Formin Model Another recent area of research is on formin proteins, which assemble polarized actin cables and cytokinetic rings. This work began as a collaboration with Dr. Isabel Sagot and Dr. David Pellman (Dana Farber Cancer Institute). Through biochemical assays performed in our lab, we discovered that C-terminal fragments of the yeast formin protein Bni1 directly nucleate actin assembly (Sagot et al., 2002). Similar discoveries were made in parallel by the labs of Boone, Zigmond and Bretscher (Pruyne et al., 2002). This co-discovery inspired researchers working in diverse organisms to investigate the role of formins in actin assembly and has led to the establishment of formins as a novel class of actin nucleators.

Crystal structure of a formin In a subsequent study, we found that formins dimerize and tightly associate with the fast-growing ends of actin filaments, allowing rapid insertion of actin subunits while protecting ends from capping proteins (Moseley et al., 2003). This study also demonstrated that the mouse formin mDia1 works by a mechanism similar to yeast Bni1, in agreement with studies performed by Li and Higgs (2003). Subsequently, through a close collaboration with Dr. Mike Eck (Dana Farber Cancer Institute), we reported the first crystal structure of an active formin molecule (Xu et al., 2003). This revealed the active state of formins to be a flexibly-tethered dimer and suggested a mechanism by which formins processively cap the growing end of an actin filament to allow insertional growth. We have also identified the cell polarity factor Bud6 as a Bni1 activator, the mechanism of which we have dissected further (Moseley et al., submitted).

Among the open questions we are now addressing are: (1) How do formins "ride" the barbed end of a growing actin filament? (2) What is the activity of full-length formin molecules? (3) How are they regulated by ligand binding and post-translational modification?

Please notify the webmaster of any broken links, missing graphics or incorrect information. All contents copyright 2000-2007, Goode Research Laboratory, Brandeis University. All rights reserved. Last modified February 9, 2007. This site was designed and created by Jacqueline Hendries.