Proteasomal Degradation of Zn-Dependent Hdacs: The E3-Ligases Implicated and the Designed Protacs That Enable Degradation

Abstract

Protein degradation by the Ubiquitin-Proteasome System is one of the main mechanisms of the regulation of cellular proteostasis, and the E3 ligases are the key effectors for the protein recognition and degradation. Many E3 ligases have key roles in cell cycle regulation, acting as checkpoints and checkpoint regulators. One of the many important proteins involved in the regulation of the cell cycle are the members of the Histone Deacetylase (HDAC) family. The importance of zinc dependent HDACs in the regulation of chromatin packing and, therefore, gene expression, has made them targets for the design and synthesis of HDAC inhibitors. However, achieving potency and selectivity has proven to be a challenge due to the homology between the zinc dependent HDACs. PROteolysis TArgeting Chimaera (PROTAC) design has been demonstrated to be a useful strategy to inhibit and selectively degrade protein targets. In this review, we attempt to summarize the E3 ligases that naturally ubiquitinate HDACs, analyze their structure, and list the known ligands that can bind to these E3 ligases and be used for PROTAC design, as well as the already described HDAC-targeted PROTACs.

Keywords: E3-ligase; HDACs; PROTACs; drug design; proteasomal degradation.

Figure 11

HECT E3 ligases. (a) Smurf2. I is the schematic representation of the assembly of the Smurf2 E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, HECT domain (HECT) in pink, WW3 domain in light blue, WW2 domain in orange, WW1 in yellow, C2 domain in green, and substrate protein in grey. Protein assembly II is the 3D representation of the C2 domain as found in PDB code 2JQZ. Protein assembly III is the 3D representation of the WW2 and WW3 domains in complex with a fragment of the substrate protein Smad7 as found in PDB code 2KXQ. Protein assembly IV is the 3D representation of the HECT domain as found in PDB code 1ZVD. (b) HUWE1. Schematic representation of the assembly of the Smurf2 E3 ligase adapted from Hunkeler M. et al. [162]: HECT domain (HECT) in pink, armadillo repeated as domain 1–4 (ARLD1–4) in shades of lilac, ubiquitin associated domain (UBA) in yellow, tryptophan and glutamic acid-rich domain (WWE) in orange, HUWE1 WWE module associated domain (HWA) in salmon pink, tower domain in dark blue, Bcl-2 homology region 3 (BH3) in light blue, and ubiquitin binding motif 1 (UBM1) in green. The 3D representation of the individually crystalized fragments of this assembly are represented next to their position in the scheme (UBA (PDB code 2EKK) yellow, HECT (PDB code 3H1D) pink, WWE (PDB code 6MIW) orange, and UBM1 (PDB code 2MUL) green).

Figure 11

HECT E3 ligases. (a) Smurf2. I is the schematic representation of the assembly of the Smurf2 E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, HECT domain (HECT) in pink, WW3 domain in light blue, WW2 domain in orange, WW1 in yellow, C2 domain in green, and substrate protein in grey. Protein assembly II is the 3D representation of the C2 domain as found in PDB code 2JQZ. Protein assembly III is the 3D representation of the WW2 and WW3 domains in complex with a fragment of the substrate protein Smad7 as found in PDB code 2KXQ. Protein assembly IV is the 3D representation of the HECT domain as found in PDB code 1ZVD. (b) HUWE1. Schematic representation of the assembly of the Smurf2 E3 ligase adapted from Hunkeler M. et al. [162]: HECT domain (HECT) in pink, armadillo repeated as domain 1–4 (ARLD1–4) in shades of lilac, ubiquitin associated domain (UBA) in yellow, tryptophan and glutamic acid-rich domain (WWE) in orange, HUWE1 WWE module associated domain (HWA) in salmon pink, tower domain in dark blue, Bcl-2 homology region 3 (BH3) in light blue, and ubiquitin binding motif 1 (UBM1) in green. The 3D representation of the individually crystalized fragments of this assembly are represented next to their position in the scheme (UBA (PDB code 2EKK) yellow, HECT (PDB code 3H1D) pink, WWE (PDB code 6MIW) orange, and UBM1 (PDB code 2MUL) green).

Figure 13

2D representation of the structures of the known HDAC-directed PROTACs [60,61,166,167,168,169,170,172] adapted from Smalley J. P. et al. [75].

Figure 1

(a) Schematic representation of the Ubiquitin-Proteasome System (UPS) for substrate degradation. (b) Classification of the E3 ligase families according to their ubiquitination mechanism: HECT-type E3 ligase (indirect), RING-type E3 ligase (direct), and RBR-type E3 ligase (indirect). (c) Schematic representation of the overall structure and mechanism of action of PROTACs.

Figure 2

Schematic representation of the Cullin-RING ligases regulation and ubiquitination cycle adapted from Cui D. et al. [41].

Figure 3

Cul3-KCTD. (a) Schematic representation of the assembly of the Cul3-KCTD E3 ligase adapted from Correale S. et al. [46]: adaptor protein KCTD in green, Recognition Protein (RP) in light blue, Substrate protein (S) in grey, Cullin-3 (Cul3) in magenta, Ring Box Protein (Rbx) in yellow, E2 ligase (E2) in dark blue, and ubiquitin (Ub) in purple. (b) 3D representation of the incomplete Cullin-3 as found in PDB code 4AP2.

Figure 4

CRL2^VHL (a) Schematic representation of the assembly of the CRL2^VHL E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring Box Protein 1 (Rbx1) in yellow, Cullin-2 (Cul2) in magenta, adaptor protein Elongin B (EloB) in orange and Elongin C (EloC) in green, substrate recognition protein VHL (VHL) in light blue, and substrate protein in grey. Protein assembly I is the 3D representation of the system made up of Cul2, EloB, EloC, and a small α-helix of VHL as found in PDB code 5N4W. Protein assembly II is the 3D representation of the N-terminal end of Cul2 bound to EloB, EloC, and VHL, as found in PDB code 4WQO. (b) 2D representation of the structures of the main VHL:HIF-1α inhibitors developed by Ciulli A. et al. [63,64]. The hydroxylated proline that mimics the activated HIF-1 α is represented in blue.

Figure 4

CRL2^VHL (a) Schematic representation of the assembly of the CRL2^VHL E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring Box Protein 1 (Rbx1) in yellow, Cullin-2 (Cul2) in magenta, adaptor protein Elongin B (EloB) in orange and Elongin C (EloC) in green, substrate recognition protein VHL (VHL) in light blue, and substrate protein in grey. Protein assembly I is the 3D representation of the system made up of Cul2, EloB, EloC, and a small α-helix of VHL as found in PDB code 5N4W. Protein assembly II is the 3D representation of the N-terminal end of Cul2 bound to EloB, EloC, and VHL, as found in PDB code 4WQO. (b) 2D representation of the structures of the main VHL:HIF-1α inhibitors developed by Ciulli A. et al. [63,64]. The hydroxylated proline that mimics the activated HIF-1 α is represented in blue.

Figure 5

CRL4^CRBN (a) Schematic representation of the assembly of the CRL4^CRBN E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring Box Protein 1 (Rbx1) in yellow, Cullin-4A (Cul4A) in magenta, adaptor protein DDB1 (DDB1) in green, substrate recognition protein CRBN (CRBN) in light blue, and substrate protein in grey. Protein assembly I is the 3D representation of the system made up of a fragment of DDB1, CRBN bound to lenalidomide (orange sticks) and the substrate protein CK1α as found in PDB code 5FQD. Protein assembly II is the 3D representation of the full sequence of DDB1, CRBN bound to lenalidomide (orange sticks) as found in PDB code 4CI2. Protein assembly III is the 3D representation of the Rbx1, Cul4A, and DDB1 as found in PDB code 2HYE. (b) 3D representation of the two different conformations adopted by CRBN. Open conformation when bound to Ikaros as found in PDB code 6H0F (left), and closed conformation when bound to CK1α as found in PDB code 5FQD (right). For the sake of clarity, neither the fragment of Ikaros nor the full CK1α protein bound to CRBN present in the PDB codes are shown.

Figure 5

CRL4^CRBN (a) Schematic representation of the assembly of the CRL4^CRBN E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring Box Protein 1 (Rbx1) in yellow, Cullin-4A (Cul4A) in magenta, adaptor protein DDB1 (DDB1) in green, substrate recognition protein CRBN (CRBN) in light blue, and substrate protein in grey. Protein assembly I is the 3D representation of the system made up of a fragment of DDB1, CRBN bound to lenalidomide (orange sticks) and the substrate protein CK1α as found in PDB code 5FQD. Protein assembly II is the 3D representation of the full sequence of DDB1, CRBN bound to lenalidomide (orange sticks) as found in PDB code 4CI2. Protein assembly III is the 3D representation of the Rbx1, Cul4A, and DDB1 as found in PDB code 2HYE. (b) 3D representation of the two different conformations adopted by CRBN. Open conformation when bound to Ikaros as found in PDB code 6H0F (left), and closed conformation when bound to CK1α as found in PDB code 5FQD (right). For the sake of clarity, neither the fragment of Ikaros nor the full CK1α protein bound to CRBN present in the PDB codes are shown.

Figure 6

Classical ImiDs and CELMoDs (a) 2D representation of the structures of the Classical ImiDs and CELMoDs reported to date that have reached clinical trials. (b) 3D representation of the binding mode to CRBN (blue) of the classical ImiDs: thalidomide (magenta), pomalidomide (yellow), and lenalidomide (orange) as found in PDB codes 4CI1, 4CI3, and 4CI2, respectively. Hydrogen bonds are shown as dashed lines.

Figure 7

Cul3^SPOP (a) Schematic representation of the assembly of the Cul3-SPOP E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring Box Protein 1 (Rbx1) in yellow, Cullin-3 (Cul3) in magenta, substrate protein (S) in grey, and adaptor protein SPOP shown as the three domains: MATH in light blue, BTB in green, and BACK in orange. Protein assembly I is the PyMOL cartoon representation of the three domains that make up SPOP as found in PDB code 3HDI. Protein assembly II is the 3D representation of the fragment of Cul3 bound to the BTB domain as found in PDB code 4EOZ. (b) 3D representation of the dimerization of SPOP as found in PDB code 3HDI.

Figure 8

Monomeric RING E3 ligases. (a) CHFR E3 ligase. I is the schematic representation of the assembly of the CHFR E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring domain (RING) in yellow, substrate recognition domain FHA (FHA) in light blue, substrate protein in grey, and the cysteine rich domain (CRD) in orange. Protein assembly II is the 3D representation of the monomeric and dimeric conformations of FHA as found in PDB code 1LGQ. Protein assembly III is the 3D representation of the CRD domain as found in PDB code 2XP0. (b) PIRH2 E3 ligase. I is the Schematic representation of the assembly of the PIRH2 E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring domain (RING) in yellow, the N-terminal domain (NTD) in turquoise, the C-termina domain (CTD) in light blue, and the substrate protein in grey. Protein assembly II is the 3D representation of the NTD as found in PDB code 2K2C. Protein assembly III is the 3D representation of the RING domain as found in PDB code 2JRJ. Protein assembly IV is the 3D representation of the CTD as found in PDB code 2K2D. (c) RLIM E3 ligase. I is the Schematic representation of the assembly of the RLIM E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring domain (RING) in yellow, the disordered N-terminal end in light blue, and the substrate protein in grey. Protein assembly II is the 3D representation of the RING domain bound to the E2 ligase Ube2e2 as found in PDB code 6W9A. Protein assembly III is the 3D representation of the RING domain bound to the E2 ligase Ube2d2 as found in PDB code 6W9D. Hydrogen bonds are shown as dashed lines and zinc ions as grey spheres.

Figure 8

Monomeric RING E3 ligases. (a) CHFR E3 ligase. I is the schematic representation of the assembly of the CHFR E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring domain (RING) in yellow, substrate recognition domain FHA (FHA) in light blue, substrate protein in grey, and the cysteine rich domain (CRD) in orange. Protein assembly II is the 3D representation of the monomeric and dimeric conformations of FHA as found in PDB code 1LGQ. Protein assembly III is the 3D representation of the CRD domain as found in PDB code 2XP0. (b) PIRH2 E3 ligase. I is the Schematic representation of the assembly of the PIRH2 E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring domain (RING) in yellow, the N-terminal domain (NTD) in turquoise, the C-termina domain (CTD) in light blue, and the substrate protein in grey. Protein assembly II is the 3D representation of the NTD as found in PDB code 2K2C. Protein assembly III is the 3D representation of the RING domain as found in PDB code 2JRJ. Protein assembly IV is the 3D representation of the CTD as found in PDB code 2K2D. (c) RLIM E3 ligase. I is the Schematic representation of the assembly of the RLIM E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring domain (RING) in yellow, the disordered N-terminal end in light blue, and the substrate protein in grey. Protein assembly II is the 3D representation of the RING domain bound to the E2 ligase Ube2e2 as found in PDB code 6W9A. Protein assembly III is the 3D representation of the RING domain bound to the E2 ligase Ube2d2 as found in PDB code 6W9D. Hydrogen bonds are shown as dashed lines and zinc ions as grey spheres.

Figure 9

SIAH2. (a) Schematic representation of the assembly of a monomer of the SIAH2 E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring domain (RING) in yellow, zinc finger domain 1 (ZNF1) in pink, zinc finger domain 2 (ZNF2) in green, substrate binding domain (SBD) in light blue, and substrate protein in grey. (b) 3D representation of the dimerization of SIAH2 by the SBD as found in PDB code 5H9M. One of the monomers of the homodimer is colored following the existing color code, and for clarity, the second monomer has been colored grey.

Figure 10

MDM. (a) Schematic representation of the assembly of a monomer of the MDM E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring domain (RING) in yellow, central domain (CD) in pink, N-terminal domain (NTD) in light blue, and substrate protein in grey. (b) Protein assembly I is the PyMOL stick and cartoon representation of the NTD of MDM in complex with a small α-helix of the p53 oncosuppressor protein as found in PDB code 4HFZ. Protein assembly II is the 3D representation of the CD as found in PDB code 2C6A. Protein assembly III is the 3D representation of the RING domain in complex with E2 ligase UbcH5B bound to ubiquitin as found in PDB code 7AI1. (c) I is the schematic representation of the assembly of the heterodimeric assembly of MDM2/MDMX E3 ligases: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring domain (RING) of MDM2 in yellow, RING domain (RING) of MDMX in green, central domain (CD) in pink, N-terminal domain (NTD) in light blue, and substrate protein in grey. Protein assembly II is the 3D representation of the heterodimeric assembly of MDM2/MDMX as found in PDB code 5MNJ.

Figure 10

MDM. (a) Schematic representation of the assembly of a monomer of the MDM E3 ligase: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring domain (RING) in yellow, central domain (CD) in pink, N-terminal domain (NTD) in light blue, and substrate protein in grey. (b) Protein assembly I is the PyMOL stick and cartoon representation of the NTD of MDM in complex with a small α-helix of the p53 oncosuppressor protein as found in PDB code 4HFZ. Protein assembly II is the 3D representation of the CD as found in PDB code 2C6A. Protein assembly III is the 3D representation of the RING domain in complex with E2 ligase UbcH5B bound to ubiquitin as found in PDB code 7AI1. (c) I is the schematic representation of the assembly of the heterodimeric assembly of MDM2/MDMX E3 ligases: ubiquitin (Ub) in purple, E2 ligase (E2) in dark blue, Ring domain (RING) of MDM2 in yellow, RING domain (RING) of MDMX in green, central domain (CD) in pink, N-terminal domain (NTD) in light blue, and substrate protein in grey. Protein assembly II is the 3D representation of the heterodimeric assembly of MDM2/MDMX as found in PDB code 5MNJ.

Figure 12

Schematic representation of the assembly of the E3 ligases that have been used for PROTAC design and the 2D representation of the structures of their respective ligands with and arrow marking the position for the functionalization. Adapted from Scheepstra M. et al. [38].