ATAD2
September 2022
ATPase Family AAA Domain Containing 2
ATAD2 (ATPase Family AAA Domain Containing 2) is a nuclear protein involved in chromatin remodeling. It is overexpressed in many cancer types, making it a broad oncology target. ATAD2 possesses an ATPase associated with diverse cellular activities (AAA+) domain and a bromodomain (Image 1), which is the pharmacological target of many small molecule inhibitors currently under investigation.
Image 1. Image is done with the Annotation Browser of the 3decision software. ATAD2 (PDB ID: 5LJ0) possesses an ATPase domain (red rectangle in Pfam line) and a bromodomain (blue rectangle in Pfam line). These annotations are coming from the Pfam database. The protein structure is colored corresponding to the selection.
However, bromodomain is very common even among other therapeutic targets such as the bromodomain and extra-terminal (BET) protein family. In their recent paper, scientists from GSK studied the structural basis of the selectivity of their ATAD2 inhibitor over an extensively investigated BET protein - BRD4 (if you want to know more about BRD4, check out our POM of April 2021 or a dedicated use case ).
Image 2. Structure of compound 6, sketched with 3decision molecule editor (powered by ChemAxon).
From a previous study, they identified compound 6 (Image 2), which showed high selectivity towards the bromodomain domain of ATAD2, compared to BRD4 bromodomain 1 (BD1). The molecular basis of this selectivity is given by the difference in polarity of the two binding sites, which triggers a conformational change of the ligand.
In ATAD2, the polar substituent at the C3’ position assumes an equatorial position inside the pocket, which has a higher polar character (Arg1007 and Arg1077), while
In BRD4, the same substituent is unable to form interactions due to more hydrophobic residues of BD1 (Trp81, Pro92, Phe93; also called“WPF shelf”), thus pointing towards the solvent. (Image 3)
This difference in the binding mode provides 500-fold more selectivity towards ATAD2.
Image 3. Compound 6 bound to ATAD2 (in orange, PDB ID: 5LJ0) and to BRD4 (in blue, PDB ID: 5LJ1). The differences between the two structures are highlighted using Full Structure highlight mode, 3decision software. The binding site of BRD4 is represented by a blue mesh. The three residues constituting the WPF shelf of BRD4 BD1 are in white sticks.
Based on those structural characteristics, GSK scientists made a hypothesis to even further increase the ATAD2 selectivity. They decided to rigidify the piperidine ring by replacing it with tropane - its constrained analog. This would potentially “lock” the C3’ substituent to the equatorial orientation also inside the BRD4 pocket, forcing it to point to the hydrophobic region which would disfavor binding to BRD4 but preserve the affinity to ATAD2.
But even if furtherly optimized, none of the synthetized compounds reached the level of selectivity of the piperidine compound 6.
Image 4: Structures of compound 6 and compound 26, its tropane analogue (circled in blue). Positions of C3’ and C5’ are indicated on the structures. Chemical structures were sketched with 3decision molecule editor (powered by ChemAxon).
By analyzing the structures of the two ligand-series (compound 6 and compound 26, Image 4) complexed with both protein targets, they provided two structural explanations for the superior selectivity of 6:
Contrary to expectations, the C3’ substituent on tropane-compound26 is well tolerated (and not at all “locked”) in the WPF shelf of BRD4 due to the local conformational rearrangement of the Trp81 residue (Image 5), which allows the allocation of the substituent and does not impair the binding.
Image 5. Compound 6 and compound 26 bound to BRD4 (PDB ID: respectively 5JL1 in blue and 6HDQ in light grey), relocation of tryptophane W81 is highlighted.
The C5’ methoxy substituent on the piperidine ring (compound 6) is detrimental to BRD4 binding due to steric reasons. The side chain of the residue leucine 94 on a loop in the bromodomain of BRD4 does not leave enough space to allocate the methoxy moiety. On contrary, in ATAD2, the pocket is bigger at this position due to valine 1018, enabling binding. Since compound 26 is not substituted at that position, it can easily fit in both pockets (Image 6). This gives a molecular explanation of the lower selectivity observed for tropane compounds.
Image 6. Compound 6 bound to ATAD2 (orange, PDB ID: 5LJ0) and compound 26 bound to BRD4 (blue, PDB ID: 6HDQ). Location of leucine Leu94 close to C5’ position is highlighted. Val1018 is less bulky and leaves more space for ligands (see molecular surface of ATAD2 in orange).
This work reveals an unexpected way in which a protein can adapt itself to minimize unfavorable interactions. These structural insights can suggest novel strategies for further optimization and to address selectivity issues.
Reference paper:
Aiming to Miss a Moving Target: Bromo and Extra Terminal Domain (BET) Selectivity in Constrained ATAD2 Inhibitors; Paul Bamborough, Chun-wa Chung, Rebecca C. Furze, Paola Grandi, Anne-Marie Michon, Robert J. Watson, Darren J. Mitchell, Heather Barnett, Rab K. Prinjha, Christina Rau, Robert J. Sheppard, Thilo Werner, and Emmanuel H. Demont, Journal of Medicinal Chemistry 2018 61 (18), 8321-8336