Buck Lab. Research Brief

Molecular Biophysics of small GTPase-protein interactions in neuronal cell signaling.

We use a wide range of experimental and computational methods
to determine the basic biophysical mechanisms by which proteins transmit signals in cells.

The methods include molecular biology, circular dichroism, fluorescence and NMR spectroscopy, as well as microcalorimetry and computational modeling (activated dynamics and free energy simulations). We determine the structures of proteins and characterize their interactions with binding partners in order to understand how particular protein features are used in
transmitting signals.

Principal project: The plexin transmembrane receptor in axon guidance.
Upcoming projects: BORGS; conformational transitions in GTPases and their binding partners,
and effect of binding interactions and cellular environment on GTPase function.

Understanding the molecular basis of events that constitute mental activity is one of the most exacting challenges for science this century. Signal transduction proteins are central to most of the processes involved, ranging from the assembly of neural circuits during development and learning, the acquisition of short and long term memory to the manifestation of emotions and neurological disease[1]. Although each of these processes is highly complex, characterization of the protein components involved and of their interactions will provide important mechanistic details into how signaling and its regulation are accomplished.

Proteomics and other screening efforts will in the foreseeable future identify most gene products relevant to cellular processes in neuronal cells and the interaction partners of these proteins will be catalogued. However, an increased understanding of the structural biophysics of protein-protein interactions and of the propensity of structures to undergo conformational change will be of critical importance, particularly in the case of proteins involved in cell signaling[2]. Such biophysical studies will provide fundamental insights into protein structure and dynamics, explain how these features are used for specific signaling purposes and how the proteins function in distinct cellular environments. In addition, this basic knowledge will eventually allow us to design polypeptides or other agents that can be used to monitor and manipulate cell signaling events.

The aim of this laboratory is to use a range of tools, including Molecular Biology, Molecular Modeling and Dynamics Simulation, Nuclear Magnetic Resonance (NMR) and other spectroscopic techniques to gain an insight into the biophysical features of neurospecific proteins. Following the determination of the polypeptide regions and of the protein structures that are involved in inter-protein contacts, thermodynamic and kinetic studies must be carried out in order to characterize the structural and conformational basis for their signaling function[e.g. 3]; for example, how do the binding energy and the details of the interaction surfaces relate to protein conformational changes that form the hallmark of many signaling events (project area 1)? Such characterizations can be tricky since some of the signaling domains are un- or only partially structured prior to interactions with their binding partners (see project area 2,[4]); others undergo substantial conformational change, structural destabilization and possibly complete unfolding on complex formation. Furthermore, the location of signaling proteins in specific cellular environments, such as proximity to the plasma membrane, is likely to affect their function (project area 3, [5]).

Figure 1: Schematic depiction of a conformational change in a signaling protein accompanying a binding event (Blue Protein). Which part of the protein is involved in the interaction? How is conformational change transmitted/coupled to another part of the protein (region in red) - both structurally and energetically? In fact, what is the biophysics basis for a modular architecture of the signaling protein?

Figure 2: Thresholds for conformational change. A single binding event may not supply sufficient energy to effect the conformational change and multiple interaction partners may be required. How do these and surrounding modules (yellow and red) synergize in order to transmit a signal?

Figure 3: Influence of the cellular environment. How does molecular crowding, co-localization and proximity to the membrane, for example, effect the mechanism of signal transduction? How does the cellular environment provide additional levels of regulation?

NMR spectroscopy is one of the most powerful methods to investigate each of these aspects of protein biophysics in solution. Ample experience with highly structured, partially unfolded and non-native states of proteins [6] has allowed us to develop a set of tools, most recently employing hydrogen exchange as well as spectroscopic measurements, to characterize these states and their conformational equilibria[7 & 8]. If the structures of the states are known, or if they can be modeled, computational approaches can be extremely useful in providing deeper insights into experimental data and in making predictions for futher experiments. Thus, we are currently exploring the use of activated dynamics and free energy calculations in order to map the energy surface between different conformational states; conformational states that may play a role in signaling. A wide range of additional tools ranging from molecular biology to biophysics are used in a problem centered approach. Development and application of new computational and NMR methods for the study of protein-protein interactions and protein conformational dynamics are also long-term activities in the laboratory.

Biological questions arise from the biophysical basis of intracellular signaling and concern the ultimate function of the processes. The specificity of the signaling and its regulation rely on protein-protein and protein-ligand recognition events. Thus a cornerstone for the identification of the molecules, involved in a biological process, is formed by details of the interactions between them. In the first instance we are working on several systems involving small GTPases and their interacting partners. Focusing on GTPases and their interaction with other molecules that function in neuronal cells is particularly exciting since these systems are most likely to involve specific interactions with novel proteins[9]. Despite the fact that GTPases act as ubiquitous switches in cell signaling, the structural features which allow proteins to bind specific GTPases are just beginning to be completely understood[10] and the role of the interaction may differ depending on the detailed structures involved. Like other GTPases and G-proteins, the molecules are understood as an on/off switch for signaling depending on the nature of the nucleotide that is bound. Conventionally only the active, GTP loaded form, is able to bind effector molecules. Regulatory proteins bind to the GTPase and effect the rate of their nucleotide hydrolysis (GTP to GDP), catalyze or inhibit the re-exchange GTP for GDP. It is these G TPase A ctivating P roteins (GAPs), G uanine E xchange F actors (GEFs) and G uanine D issociation I nhibitors (GDIs) that are regulated in turn by upstream signaling events. Thus a molecule that interacts with a GTPase, may serve to sequester it away from other sites of activity, may function as a GAP, GEF, GDI or in fact slow intrinsic nucleotide hydrolysis (anti-GAP). The three systems (one principal, two currently in their early stages) which we investigate are examples of different conformational states of proteins and how they are involved in GTPase signaling.

Growth, guidance and branching of axons are critical events which generate precise connections in the developing nervous system. Small Rho-family GTPases (Rac1, Cdc42 and Rho ) are signal transduction molecules that remodel the actin cytoskeleton and are thus intimately involved with path finding and maintenance of neuronal cells[11]. The Rho-GTPases function by signaling to components that regulate the nucleation and breakdown of actin cytoskeletal structures (e.g. [12]). Ligands, receptors and an intracellular machinery that connects to the cytoskeletal dynamics via Rho-GTPases is needed for the cell to make decisions whether to grown towards and contact another nerve cell or retreat away from it. For example, guidance cues, such as neuropilin and semaphorin, are ligands that are detected by the plexin receptor, located in the tip of the growing axon. By contrast to other guidance receptors, a direct binding interaction between the cytoplasmic domain of several members of the plexin family (human-B1, -A1 and A2) and the GTPase, Rac1, has been demonstrated [13,14] but the role of this interaction for the plexin signaling mechanism is not clearly understood. Current models propose that the principal function of plexin is to bind and sequester active Rac1, removing it from the pool of GTPases that can interact with effector molecules, thereby reducing actin turnover and contributing to growth cone collapse. However, it has also been reported that the cytoplasmic region of plexin has an active effect on GTPase function. The plexin-Rac interaction may either directly, or indirectly via another molecule, promote nucleotide hydrolysis or exchange in other GTPases [15, 16].

Figure 4: Signaling mechanisms of axon guidance by the plexin transmembrane receptor.

Events (adapted from reviews by [17-19]). In Plexin B1 ligand Semaphorin4D (light blue) binds (panel B) and causes a conformational change and stimulates Rac1.GTPase (red) binding (C), which in turn enhances binding of RhoA PDZ-GEF (light and blue-green) binding (D), possibly activating the exchange factor (E). Conformational changes may also activate GAP or other binding functions in regions (orange) surrounding the Rac binding domain (F). Rnd1 GTPase competes for the Rac1 binding site (not shown). In plexin A1 (not shown), we also detect affinity for Rac1.GDP. Furthermore, L1 and Semaphorin3A cooperate as extracellular ligands for plexin A1 and MICAL is another binding partner for its cytoplasmic domain. Interactions of plexin A1 with L1 may change the biological response to semaphorin ligands from a growth repulsive to an attractive one.

Questions: How is binding specificity for Rac1.GTP vs. Rac1.GDP and Rnd1 determined? How does PDZ domain and MICAL-1 binding affect affinity for the GTPases? Is there a role for the cytoplasmic domains of neuropilin-1, semaphorin4D or L1? Does the cytoplasmic domain of plexins have a regulatory function (GEF, GAP or GDI) for GTPases? How do the different guidance receptors together orchestrate well define responses during neuronal growth, positional development and maintenance?

Projects under development:

Recent reports suggest that septins are involved in processes ranging from vesicle budding/exocytosis to cellular and axonal migration. Septins contain a GTP binding region and some of their functionality relates to those of GTPases. A binding partner, Borg appears to regulate certain septins in conjunction or competition with Cdc42 binding [20, 21]. Although the septin containing filaments that finally result are large, they appear to have important roles also in cancer and Parkinson's disease [22]. Our preliminary study shows that Borg proteins are largely unstructured in absence of binding partners. Thus it will be of interest to see whether a folding transition is at the heart of this signaling mechanism.

Rin belongs to a novel calmodulin binding GTPase family involved in growth factor dependent signaling in mature nervous cells[23]. It has been shown that Rin-calmodulin interactions are required to induce neurite outgrowth through Rac/Cdc42, thus linking GTPase and calcium mediated signaling in neuronal cells[24]. p21 or H-Ras, is the classical small GTPase and has been the subject to an astonishing amount of research since mutations in Ras are found in ~30% of human cancers. We are using Ras as an example for a non-Rho family GTPase and as part of a study of the effect of protein environment (using mimics for membranes and molecular crowding) on GTPase biophysics.

Selected references and background reading:

[1] Albright, T.D. et al., (2000) Neural Science Review Suppl. to Cell 100, S5-55
[2] Kobe , B. & Kemp, B.E. (1999) Nature 402, 373-6
[3] Buck, M. & Rosen, M.K. (2001) Science 291, 2329-30.
[4] Dyson, H.J. & Wright, P.E. (2002) Curr. Opin. Struct. Biol. 12, 54-60.
[5] Casey, P.J. (1995) Science 268, 221-5
[6] Buck, M. (1998) Quarterly Reviews of Biophysics 31, 297-355.
[7] Buck, M. et al., (2001) Biochemistry 40, 14115-14122.
[8] Buck, M. et al., (2004) J. Mol. Biol. 338, 271-285.
[9] Luo, L. (2000) Nat. Reviews Neuroscience 1, 173-80
[10] Vetter, I.A. & Wittinghofer, A. (2001) Science 294, 1299-304.
[11] Ng, J. et al., (2002) Nature 416, 442-447.
[12] Etienne-Manneville, S. & Hall, A. (2002) Nature 420, 629-635.
[13] Vikis, H.G. et al., (2000) Proc. Natl. Acad. Sci. U S A 97, 12457-12462.
[14] Driessens, M.H. et al., (2000) Curr. Biol 39, 1243-50.
[15] Oimuma, I. et al., (2004) Science 294, 1299-1304.
[16] Turner, L.J. et al., (2004) J. Biol. Chem, 279, 33199-33205.
[17] Liu, B.P. & Strittmatter, S.M. (2001) Curr. Opin.CellBiol.13, 619-626.
[18] Pasterkamp, R.J. & Kolodkin, A.L. (2003) Curr. Opin. Neurobiol. 13, 79-89.
[19] Fiore, R. & Pueschel, A.W. (2003) Front. Biosci. 8, S484-499.
[20] Sheffield , P.J. et al., (2003) J.Biol.Chem. 278, 3483-3488.
[21] Joberty, G. et al., (2001) Nat. Cell Biol. 3, 861-866.
[22] Hall, P.A. & Russell, S.E. (2004) J. Pathol, 204, 489-505.
[23] Spencer, M.L. et al., (2003) J. Biol. Chem. 277, 17605-17615.
[24] Hoshino, M. & Nakamura, S. (2003) J. Cell Biol. 163, 1067-1076.

Sponsors: The Buck lab. is grateful to the following for financial support:

American Heart Association	March of Dimes Birth Defects Foundation
American Cancer Society	National Institutes of Health