Highly selective inhibition of histone demethylases by de novo macrocyclic peptides
Aberrant histone modifications are associated to several chronic and highly prevalent worldwide inflammatory and cardiovascular diseases, as well as cancer. One of the primary reasons leading to this scenario is the mutation or dys-regulation of proteins responsible to read and write post-translational histone modifications, particularly methyltransferases and demethylases, commonly known as KDMs.
Several KMDs proteins (KDM1/LSD, JmjC-KDMs, KDM2 – KDM7) or related proteins (2OG oxygenases) have been identified, sharing similar biological function but distinguished by the mechanism of action. Despite the interest in tightly controlling the regulation of such KDMs, only a few selective inhibitors have been able to modulate the function of some of these proteins, as reported in the past decade. One the main challenging targets is KDM4/A-C protein, belonging to the KDM4 subfamily. The protein is constituted by a catalytic JmjC-domain and shares high similarity on the remaining binding site with other proteins of the subfamily KDM4s (KDM4/D-E).
To target KDM4/A-C, a RAndom nonstandard Peptide Integrated Discovery (RaPID) approach was developed. First, an mRNA library consisting on AUG-(NNK)4–12-UGC; where: NNK are random nucleotides. Then, a puromycin linker is added, to obtain a puromycin-mRNA fused library. Afterwards, the mRNA sequences are translated using a reprogrammed ribosome system, producing a library of peptides-fused to mRNA through the puromycin linker; and later the mRNA is reverse transcripted to DNA. Afterwards, a spontaneous cyclization is produced, generating a thioether-macrocyclic library.
The library is then screened against binding to KDM4. Macrocycles with binding to the target are finally identified through PCR.
Figure 1. Generation of macrocyclic library and screening of binders.
By using this approach, the authors identified two cyclic peptides (CP2 and CP4) able to selectively bind to KDM4/A-C with high potency (IC50 < 50 nM) while the potency against KDM4/D-E exhibited a 100-fold activity reduction, as detailed in Table 1.
Table 1. IC50 inhibition values of CP1-5 cyclic peptides when targeting several KDMs families. IC50 values were determined using AlphaScreen assay while binding constants were measured using biolayer interferometry.
To gain a better understanding of the CP2 selectivity (14-mer peptide), the complexed KDM4A / CP2 system was crystallized (2.7 Ả resolution). The structure demonstrates that CP2 binds to the histone-binding groove of KDM4A, occupying the substrate binding site and not the 2OG-binding pocket. Analysis of the inhibitor conformation revealed that CP2 adopts a twisted β-sheet fold with a type-1 β-turn at the active site, and engages in an extensive hydrogen bonding network that induce a notable conformational change in the surrounding KDM4A
Figure 2. (a) Binding site of KDM4A complexed with CP2. (b). Overlay of CP2 with the backbone KDM4A substrate. (c) B-sheet fold conformation of CP2 when bound in the histone substrate binding site.
The crystallographic structure also allowed to follow a structure-guided optimization approach. The former identification of key elementary amino acids (Arg6 of CP2, whose mutated analogs exhibited a 50-fold activity reduction; IC50 > 2400 nM), was supplemented by a) the introduction of a poly-arginine sequence (CP2(polyR), IC50 < 2 nM) with increased cellular intake and b) improvement of proteolytic stability through selective backbone N-methylation at different amino acid positions that were evidenced to not participate directly in the interaction with KDM4A. The previous modifications led to a highly stable macrocycle able to stabilize the protein target conformation and thus alter the histone methylation pattern.
Selective inhibition of targets that share high similarity among family proteins remains a challenge in the field of drug discovery. Thus, an intense research is being conducted since the last years.The use of peptide, and more especifically, macrocyclic peptides, is gaining a predominant role when dealing with selectivity issues as demonstrated by the authors.
As reported, CP2 peptide, a 14-mer cyclic structure optimized in terms of cell permeability and proteolytic stability, has been proven to be a potent and selective modulator of KDM4/A-C. At Iproteos we master the design of cyclic and non-cyclic peptidomimetic structures to modulate challenging targets, as the one reported in the present article. Our propietary IPROTech platform is able to provide peptide-like molecules with excellent DMPK profiles, and highly potent activity and selectivity. We have applied IPROTech to our pipeline in diffcult targets, such as CNS proteins or PPIs, and to partnership projects as well. If you want to know more, do not hesitate to contact us.