Progress in the Development of E3 Ligase VHL Ligand
The von Hippel-Lindau (VHL) Cullin RING E3 ligase is an essential enzyme in the ubiquitin-proteasome system, recruiting substrates for ubiquitination and subsequent proteasomal degradation of hypoxia-inducible factors. Designing high-quality small molecule ligands with strong binding affinity to the E3 ligase is the basis for PROTAC development. Small molecule ligands binding to VHL can function as inhibitors themselves or be applied in well-designed, effective PROTACs targeting various proteins.
VHL Gene and Protein
The von Hippel Lindau (VHL) gene was first discovered in 1993 through positional cloning. It is associated with VHL disease, a hereditary disorder caused by inactivating mutations in the vhl lineage, predisposing individuals to various types of tumors such as retinal hemangioblastomas and hemangioblastomas. These conditions were known long before the discovery of the vhl gene. The vhl gene, located on chromosome 3p25, encodes two major protein isoforms: a 213-amino acid long isoform (VHL1-213) and a 160-amino acid short isoform (VHL54-213), originating from an internal alternative translation initiation site (Met54) and lacking the N-terminal pentameric acidic repeat domain. These two isoforms, long and short, are commonly referred to as pVHL30 and pVHL19, based on their apparent molecular weights on gel electrophoresis. Both isoforms exhibit tumor suppressor functions, as demonstrated by functional complementation studies in mice. Importantly, both isoforms are widely expressed and show E3 ligase activity, targeting hypoxia-inducible factors (HIF) for oxygen-dependent degradation as substrates.
Small Molecule VHL Inhibitors
Driven by the advancement of PROTAC technology and its potential applications, the search for novel E3 ligase binders began. Peptide binders have provided useful tools for targeting protein-protein interactions (PPIs), which are necessary for targeting E3 ligases and can be optimized for high affinity and excellent selectivity at their target binding sites. Non-peptide binders typically consist of small molecule scaffolds with an average molecular weight less than 500 Da, thus possessing ideal drug-like properties. The physicochemical properties and biological activities of small molecule inhibitors can be finely tuned by adjusting their molecular structures, for example, by introducing lipophilic groups, controlling the number of hydrogen bond donors (HBD), or cleverly altering their electronics through substitution. However, due to their much smaller size and surface area compared to peptide binders, non-peptide small molecules have difficulty successfully targeting shallow protein surfaces and PPIs discovered outside the active site, compared to their peptide counterparts. Therefore, when targeting PPIs, small molecule non-peptide inhibitors generally have lower affinity and lower target selectivity compared to peptide counterparts. Ideally, small molecule E3 ligase binders should possess good physicochemical properties while maintaining high target binding affinity, selectivity, and cellular activity.
- First Generation VHL Inhibitors
Since both high-throughput screening (HTS) and virtual screening methods initially failed to identify genuine VHL binders, researchers from the Ciulli and Crews laboratories turned to nature to find starting points for rationally designed VHL small molecule ligands. They targeted the VHL protein surface at the known binding site of its natural substrate hydroxyl-HIF-1a, attempting to mimic key PPIs observed from co-crystal structures. Based on the elemental recognition of VHL for hydroxyl-HIF-1α, hydroxylation of proline Hyp564, originating from Pro564, was chosen as the initial core motif for de novo synthesis of hydroxyproline (Hyp) derivatives.
- Second Generation VHL Inhibitors
Despite successful design of first-generation VHL binders, these molecules still possessed only moderate binding potency in the single-digit micromolar range and low lipophilicity, potentially limiting cellular permeability and hence lacking cellular activity. Efforts to enhance binding affinity and lipophilicity led the Ciulli lab to follow a structure-guided VHL inhibitor design strategy over the ensuing years, using structural measurement methods and Isothermal Titration Calorimetry (ITC). As previous studies revealed that the a tBu functionality at the LHSα site was highly beneficial, further optimization efforts began with the tBu-Hyp fragment, which displayed a more balanced lipophilicity as an anchor ligand compared to the hydrophilic isoazole-Hyp. The crystal structure of 14 bound to VCB elucidated key interactions retained at the Hyp core and new LHS orientation, with the tBu group facing upward forming hydrophobic contacts with Phe91 and Trp88.
- Virtual Screening for VHL Inhibitors
Virtual screening methods have been employed to identify and develop new VHL binders. An early structure-based virtual screening against a library of 90,000 natural products and natural product-like molecules docked against VHL identified binder with an IC50 value of 2.3 mM. Molecular modeling suggests that binder occupies the Hyp-binding site of VHL, but its Hyp fragment interacts with Ser68 in the LHS pocket instead of Ser111, as reported for first-generation VHL binders. Although binder resulted in increased gene expression of downstream targets of HIF-1α and promoted angiogenesis in a zebrafish in vivo model.
VHL Ligands for PROTACs
Shortly after the reporting of the first cell-active VHL second-generation inhibitor VH032, several studies demonstrated successful incorporation of this small molecule VHL ligand as the E3 ligase recruiting moiety into PROTACs. With the development of VHL-recruiting PROTACs, several suitable VHL ligand scaffolds have been developed and employed for linker conjugation. Building upon the LHS peptide structure of the second-generation VHL inhibitor, the linker moiety can be easily attached to the N-terminus of VH032 by swapping the terminal acetyl group with an appropriate linker moiety. On the RHS of the inhibitor, analysis of the co-crystal structure of VH032 with VCB has led to the derivation of additional solvent-exposed atoms as potential interfaces for linker binding. Over 750 VHL-recruiting PROTACs have been published targeting over 50 different proteasome-degradation POIs.
- First VHL-Recruiting PROTAC
The Ciulli lab reported one of the first applications of small molecule inhibitors as the recruiting moiety of VHL in degradation techniques in 2015, integrating VH032 or ligand into the first series of PROTACs targeting Bromo and Extra-Terminal (BET) proteins. These BET-targeting PROTACs consisted of VH032 as the VHL-recruiting moiety, assembled with a polyethylene glycol (PEG) linker moiety, with one side bearing a carboxylic acid and the other a diazo group, and the pan-BET selective bromodomain inhibitor JQ1 as the BET ligand. In a two-step synthesis strategy, VH032's N-terminal free amine was coupled with the carboxylic acid of the linker through initial HATU-mediated acylation, followed by reduction of the diazo to a free amine and another amide bond formation, yielding three PROTACs with hydrolyzable carboxylates towards the BET ligand JQ1.
- N-Terminal VHL-Recruiting PROTACs
The majority of subsequent VHL-recruiting PROTACs utilize the LHS N-terminus as the junctional binding site. Following the disclosure of the BET degrader MZ1 by the Ciulli lab, a team at Arvinas disclosed the structurally related pan-selective BET degrader ARV-771 for Brd2/3/4. Using established amide coupling chemistry, a slightly shorter ether chain (eight atoms in ARV-771 versus ten in MZ1) was tethered to the VH032-modified structure, with an additional methyl at the benzyl position in RHS of Hyp. ARV-771 served as a potent degrader for Brd2, Brd3, and Brd4, with a DC50 value of approximately 5nM in 22Rv1 cancer cells, and resulted in tumor size reduction in mouse xenograft models. ARV-771, along with MZ1, has been widely used as a benchmark BET protein PROTAC degrader.
- Chemical Innovation in VHL-Based PROTACs
To aid in the development of PROTAC degraders for new target POIs and expand the chemical space sampled, exploration of alternative and unconventional linker chemistries and motifs is crucial. The Lokey and Ciulli groups analyzed the impact of structural features on the cellular permeability of previously reported VHL-recruiting degraders, finding that the environment around HBD, particularly intramolecular hydrogen bonds (IMHBs), greatly influences their lipophilicity and cellular permeability efficiency. Specifically, solvent-exposed amide functionalities were found to contribute to low cellular permeability. To alleviate this, attempts to replace amide functionalities with ester functionalities in the model VHL linker led to increased cellular permeability but decreased binding affinity for VHL when the ester was positioned near the Hyp binding site.
Summary and Outlook
Fragment-based design has proven crucial for the discovery of early VHL binders and subsequent rational structure-guided optimization into effective VHL inhibitors. Beyond their apparent applications as chemical probes disrupting the VHL/HIF-1α interaction and thus upregulating HIF-1α-dependent processes, VHL inhibitors have served as platforms for further chemical development, such as VHL-recruiting fluorescent or NMR probes for biophysical assays. Particularly, the development of highly affine and specific small molecule ligands targeting VHL has paved the way for PROTACs development, resulting in the establishment of numerous small molecule VHL-recruiting PROTACs effectively targeting degradation of a variety of different proteins.
Reference
- Diehl, C. J., and Ciulli, A., Discovery of small molecule ligands for the von Hippel-Lindau (VHL) E3 ligase and their use as inhibitors and PROTAC degraders, Chem. Soc. Rev., 2022, 51, 8216-8257.
