Rab guanosine triphosphatases (GTPases) are key players in the regulation of eukaryotic intracellular trafficking events such as the formation, motility, targeting, and docking of vesicles. [1] The reversible association of Rab GTPases with intracellular membranes and other proteins is essential for their function, and is mediated by geranylgeranyl (GG) group(s) covalently attached to C-terminal cysteine residues. This modification, known as prenylation, is mediated by the enzyme Rab geranylgeranyl transferase (RabGGTase) in concert with the accessory Rab escort protein (REP). [2] The overexpression of RabGGTase and its substrates such as Rab5a, Rab7, and Rab25 in several cancer types [3] and the finding that RabGGTase inhibition leads to p53-independent apoptosis [4] validate RabGGTase as a promising anticancer target. Whereas numerous inhibitors of the related CaaX prenyl transferases, farnesyl transferase (FTase), [5] and geranylgeranyl transferase I (GGTase I) [6] have been developed and have even reached clinical trials, [7] in most cases their cellular targets are not known with certainty. Very few RabGGTase inhibitors are known to date, [8] many of which are either nonselective with respect to FTase or show only moderate inhibition of cellular Rab prenylation.
The most potent of the currently available RabGGTase inhibitors, BMS3, was originally designed as an FTase inhibitor. [4] Since BMS3 lacks selectivity with respect to FTase, both in vitro [4] and in cells, [9] its pro-apoptotic effect could only be attributed to RabGGTase inhibition indirectly. Selective inhibitors of RabGGTase would be valuable for analysis of the effects of selective inhibition of Rab trafficking on the proliferation of transformed cells, for the elucidation of side effects related to FTase inhibition, and, in general, for the analysis of Rab-mediated cellular processes. With these goals in mind, we chose to generate cocrystal structures of RabGGTase and FTase with BMS3 and use this structural information to design RabGGTase-specific inhibitors.
Herein we report the structure-guided design of selective RabGGTase inhibitors with potent cellular activities. The inhibitors target the RabGGTase-specific tunnel adjacent to the GGPP binding site (TAG tunnel), which was recently found in a cocrystal structure of an RabGGTase-peptidebased inhibitor. [8b] Since this TAG tunnel is a unique feature that distinguishes RabGGTase from FTase and GGTase I, it provides a rational molecular basis to achieve selectivity.
To enable rational design of the inhibitor, we synthesized BMS3, soaked it into both RabGGTase and FTase, and solved the binary RabGGTase:BMS3, the ternary RabGGTase:BMS3:GGPP, and the ternary FTase:BMS3:FPP cocrystal structures (Figure 1).
The binding modes and conformations of BMS3 in these structures are highly similar. In both enzymes the imidazole ring coordinates to the catalytic zinc ion, whilst the phenyl ring of the tetrahydrobenzodiazepine (THB) core p stacks with the Tyr361 or the Phe289 residue. The 3-benzyl moiety extends toward the lipid binding sites in both enzymes and is involved in extensive T stacking with hydrophobic residues of the enzymes (Figure 1). These common interaction patterns result in the twisted form of the central 3-benzyl-THB unit that is observed in both enzymes. The anisylsulfonyl substituent adopts a pseudoaxial position and is involved in internal p-stacking interactions. In RabGGTase, an additional