ADCK3/COQ8A: the choice target of the UbiB protein kinase-like family
UbiB kinase-like proteins exist in archaea, bacteria and eukaryotes, and comprise ~25% of all microbial protein kinase-like (PKL) enzymes. The founding member of the UbiB family, UbiB from Escherichia coli, supports the biosynthesis of ubiquinone, also known as coenzyme Q (CoQ), by an unknown mechanism. In eukaryotes, UbiB homologues are found exclusively in mitochondria, where they likewise have been connected to CoQ biosynthesis (for a review, see Trends Biochem. Sci. 42, 824–843; 2017).
Mutations in two of the five human UbiB homologues, COQ8A (ADCK3) and COQ8B (ADCK4), give rise to neurological and kidney disorders, respectively, both associated with CoQ deficiency. To date, only UbiB, its Saccharomyces cerevisiae orthologue Coq8p and COQ8A (collectively referred to here as COQ8) have been studied in depth biochemically, and COQ8A is therefore the focus of this article.
COQ8 was originally speculated to be a protein kinase. However, subsequent structural work revealed that the COQ8 active site is sterically occluded by UbiB-specific domains, suggesting that it possesses unorthodox functionality in lieu of typical protein kinase activity (Mol. Cell 57, 83–94; 2015). COQ8 possesses a conserved ATPase activity that is activated by binding to membranes containing cardiolipin and by phenolic compounds that resemble CoQ pathway intermediates (Cell Chem. Biol. 25, 154–165; 2018). This ATPase activity appears to be essential for COQ8’s role in CoQ biosynthesis, where it supports the integrity of a biosynthetic complex of other CoQ-related proteins. Although its precise mechanism of action remains unclear, a current model proposes that COQ8 leverages its ATPase activity to access hydrophobic CoQ intermediates within the mitochondrial inner membrane (Fig. 1a).
Fig. 1 | COQ8A function and modulation. a | Model of COQ8A facilitating CoQ biosynthesis through its ATPase activity. COQ8A may use this activity to gain access to hydrophobic CoQ intermediates and deliver mature CoQ into the hydrophobic region of the mitochondrial inner membrane. The domain structure of COQ8A, as described in Mol. Cell 57, 83–94; 2015, is highlighted and other proteins involved in CoQ biosynthesis are indicated as circles. The structures of CoQ and its precursor have been simplified for clarity; for details, see Trends Biochem. Sci. 42, 824–843; 2017. b | Structure of the quinoline analogue UNC-CA157 with reported kinase inhibition values. GAK, G-associated kinase; NLK, Nemo-like kinase; Pi, inorganic phosphate; RIPK2, receptor-interacting serine/threonine protein kinase 2.
A chemical genetic knockout system in S. cerevisiae was established to probe COQ8 function. Two point mutations to cysteines in Coq8p enable it to accept a halomethyl ketone compound, which irreversibly inhibits the protein (Cell Chem. Biol. 25, 154–165; 2018).
However, literature tool compounds for the human protein, COQ8A, are sparse. Several very promiscuous compounds show reasonable biochemical potency, including dasatinib (dissociation constant (Kd) = 190 nM), PD-173955 (Kd = 910 nM) and R406 (Kd = 1100 nM), with only TG-100-115 (Kd = 94 nM) below 100 nM (Nat. Biotechnol. 29, 1046–1051; 2011). Owing to their promiscuity, these compounds are of very limited use for establishing a phenotype for COQ8A inhibition, although they could be starting points for medicinal chemistry efforts.
Recently, a chemical probe developed for cyclin G-associated kinase (GAK) known as SGC-GAK-1 showed COQ8A as an off-target (GAK Kd = 1.9 nM and COQ8A Kd = 190 nM) (J. Med. Chem. 62, 2830–2836; 2019). A quinoline analogue of this series (UNC-CA157) (Fig. 1b) demonstrated the most potent COQ8A binding for a small-molecule inhibitor in the literature to our knowledge (Kd = 66 nM) to date. This series more generally has also been shown to have a narrow-spectrum kinome inhibition profile (ChemMedChem 13, 48–66; 2018). The literature therefore provides a potential foundation to inform the design of a selective cell-active COQ8A inhibitor through combining information from multiple chemotypes.
This article is part of a series from the NIH Common Fund Illuminating the Druggable Genome (IDG) programme. The goal of IDG is to catalyse research on understudied proteins from druggable gene families by providing reagents, phenotypes and a mineable database, focusing on G protein-coupled receptors, kinases and ion channels. For more information, see https://druggablegenome.net/. Work was also supported by NIH R35GM131795 (D.J.P.) and NIH T32GM008505/NSF DGE-1747503 (N.H.M.).