Roco Proteins: GTPases with a Baroque Structure and Mechanism

Abstract

Mutations in leucine-rich repeat kinase 2 (LRRK2) are a common cause of genetically inherited Parkinson's Disease (PD). LRRK2 is a large, multi-domain protein belonging to the Roco protein family, a family of GTPases characterized by a central RocCOR (Ras of complex proteins/C-terminal of Roc) domain tandem. Despite the progress in characterizing the GTPase function of Roco proteins, there is still an ongoing debate concerning the working mechanism of Roco proteins in general, and LRRK2 in particular. This review consists of two parts. First, an overview is given of the wide evolutionary range of Roco proteins, leading to a variety of physiological functions. The second part focusses on the GTPase function of the RocCOR domain tandem central to the action of all Roco proteins, and progress in the understanding of its structure and biochemistry is discussed and reviewed. Finally, based on the recent work of our and other labs, a new working hypothesis for the mechanism of Roco proteins is proposed.

Keywords: GTPase mechanism; Parkinson’s disease; Roco proteins; leucine-rich repeat kinase 2; unconventional G-protein.

Figure 1

Classification of Roco proteins based on domain topology. Roco proteins are characterized by the presence of a Roc domain (red) and COR domain (blue). Based on their domain topology, Roco proteins can be divided into three groups [29]. Here some representatives of each group are shown. The first group only contains an LRR domain (yellow) preceding the RocCOR. Representatives of this group are found in metazoa, plants, archaea, and bacteria. The second group contains, in addition to this LRR–Roc–COR domain topology, a C-terminal kinase domain (green). Moreover, members of this group contain several protein–protein interaction and regulatory domains like an Ankyrin repeat (ANK, pink), armadillo repeats (ARM, light blue), a N-terminal motif of RasGEF (N-GEF, light blue), a cyclic nucleotide binding domain (cNB, light grey), a Rab-like GTPase activators and myotubularins domain (GRAM, blue), a Ras Guanine Exchange Factor domain (RasGEF, beige), a N-terminal myotubularin-related domain (myotub, orange), a protein tyrosine phosphatase domain (PTP, light green), a Dishevelled domain, and an Egl-10 domain and Pleckstrin domain (DEP, grey). The last group has one human family member, DAPK1. DAPK1 lacks an LRR domain and is characterized by its C-terminal death domain (DD, brown).

Figure 2

Crystal structure of the swapped dimer of the LRRK2 Roc domain bound to GDP (pdb: 2zej) [104]. The two swapped G-domains are shown in orange and pink, respectively. For one active site the P-loop or G1 motif (cyan), the switch I or G2 motif (light pink), the switch II or G3 motif (yellow), the G4 (red), and G5 motif (blue) are highlighted. Each active site is composed of the G1–3 motif of one protomer and the G4–5 motif of the other. GDP is depicted as a green stick model an Mg²⁺ as a yellow sphere.

Figure 3

The C-terminal half of the COR domain is important for dimerization (pdb: 3DPU, 4WNR and 6HLU). (a) The crystal structure of the RocCOR∆C domain construct from the Roco2 protein of Methanosarcina barkeri reveals a monomeric conformation [33]. (b) The crystal structure of the RocCOR domain construct from the Roco protein of Chlorobium tepidum shows dimer formation via the C-terminal COR domains. The COR domain contains two subdomains: an N-terminal, mainly α-helical domain with a short antiparallel β-sheet (N-COR), and a C-terminal domain with a β-sheet surrounded by α-helices and a β-hairpin involved in dimerization (C-COR). Only one Roc domain is resolved in the structure, presumably due to the flexibility of the other Roc domain. The putative site of the second Roc domain is indicated with a black dotted line [33]. (c) The crystal structure of the LRR–RocCOR domain construct from CtRoco. This structure reveals the second Roc domain and the orientation of the LRR domains with respect to the other domain. Apart from the COR–COR interactions, also Roc–Roc and Roc–COR interactions contribute significantly to the dimer interface [17].

Figure 4

G-protein cycle of conventional G-proteins versus G-proteins activated by dimerization (GADs). (a) Conventional G-proteins have very high nucleotide affinities resulting in a low nucleotide dissociation rate. Binding of guanine nucleotide exchange factors (GEFs) decreases the affinity for the bound nucleotide allowing it to be released from the protein. Due to the higher cellular GTP concentration, GTP then binds to the protein and the protein switches to its active GTP-bound state, where it interacts with downstream effectors. GTPase activating proteins (GAPs) can then bind to the G-protein, stabilize and/or complement its catalytic machinery and in this way, increase its intrinsically low GTP hydrolysis rate. The G-protein then switches to its GDP-bound inactive state. (b) GADs have micromolar nucleotide affinities, leading to high nucleotide dissociation rates. Therefore, they do not require GEFs to cycle from their inactive GDP- to their active GTP-bound state. Following GTP binding, the G-domains of GADs dimerize. In this way, both subunits complement each other’s active site and are able to hydrolyse GTP. The GADs then cycle back to their inactive, GDP-bound state. In conclusion, GADs possess all the necessary components for GTP-binding and hydrolysis and cycle between an active and inactive state without the aid of GEFs and GAPs [108].

Figure 5

A newly proposed working mechanism for LRRK2. Cytosolic GTP-bound monomeric LRRK2 is recruited to the cell membrane by binding via its N-terminal domains to GTP-bound Rab proteins that are located at the membrane. At the cell membrane, GTP is hydrolysed, and the protein dimerizes. Meanwhile, the LRRK2 kinase domain is activated, and Rab proteins are phosphorylated. The low affinity of LRRK2 for GDP then leads to fast GDP release. Due to the higher GTP concentration present in the cell, this results in rebinding of GTP, monomerization of LRRK2 and return to the cytosol.