Special Issue on RNA-Based Catalysts that Revolutionized the Discovery of Bioactive Peptides

Abstract

This special issue of the Israel Journal of Chemistry is in celebration of the Wolf Prize awarded to Hiroaki Suga in 2023 “for pioneering discoveries that illuminate the functions and pathological dysfunctions of RNA and proteins and for creating strategies to harness the capabilities of these biopolymers in new ways to ameliorate human diseases”. In this issue we collect contributions from former trainees, collaborators, and beyond to profile how his work has impacted the fields of bioorganic chemistry, synthetic biology, and drug discovery. Focusing on his development of RNA-based aminoacylation catalysts, his share of the prize “For developing RNA-based catalysts that revolutionized the discovery of bioactive peptides” is emphasised here in a collection of nine review articles and one research article that span aspects of oligonucleotide acylation catalysis, reprogramming of the genetic code, and applications of this in peptide drug discovery.

The RNA-based catalysts developed by Suga are called flexizymes. These were initially developed using the oligonucleotide selection platform ‘SELEX’ (systematic evolution of ligands by exponential enrichment). The in vitro selection scheme was designed to select for self-acylating activity by evolving a 5’ extension on tRNA, mimicking a key step in a potential transition between RNA-based life and modern protein-dominant life. RNA molecules that were able to attach a biotinylated amino acid to their own 3’ end were enriched by streptavidin pull-down, connecting survival to catalytic activity. After several more selection campaigns aiming to improve the level of activity and scope of the 5’ extension to independently aminoacylate added tRNA in trans, the modern flexizymes eFx, dFx, and aFx were born (discussed in detail in several reviews elsewhere).^{1, 2} While this work began in fundamental biochemistry, the practical applications of a catalyst that is largely agnostic about which amino acid it attaches to which tRNA soon became apparent. However, as outlined in the contribution by Cho, Lee, and coworkers,³ an understanding of the precise working of flexizymes at the molecular level is still lacking. While a crystal structure of the oligonucleotide has been available for some time,⁴ the amino acid component is poorly resolved and so this leaves open questions about the positioning of the substrate ester and contributions of nucleotides in the catalytic pocket. They describe attempts to profile the rules for design of good flexizyme substrates, and how this has expanded research into novel bio-based polymers. One consequence of having relatively easy access to tRNA acylated with amino acid-like non-canonical building blocks is the ease with which translational space could be explored, and this fed back into further fundamental insights into the translation process itself.

Flexizymes have wide-ranging applications in ‘genetic code reprogramming’ of in vitro translation by providing tRNA pre-loaded with amino acids. However, the full potential of engineering translation in a test tube is yet to be reached. Jones and Hartman⁵ describe efforts to reduce the degeneracy of the genetic code, with a long-term goal of increasing the number of assignable codons and thereby the total number of building blocks that can be translated at once. In their work they give a detailed break down per codon box of the decoding process and how tRNA modifications influence this, as well as how evolved ribosomes can shift this balance. Further approaches on how to shift the balance between native decoding and reprogramming is discussed in work by Hopstaken, Große-Wichtrup, and Jongkees,⁶ where they cover this challenge from the perspective of the competing factors throughout the translation process (beyond tRNA base pairing with mRNA in the ribosome). Included here are options for removing competition by targeting aaRS enzymes, amino acids, tRNA, release factors, and initiation elements. Finally, within the set of articles describing aminoacylation and genetic code reprogramming approaches, Watkins, Kavoor, and Musier-Forsyth⁷ profile in vitro and cellular strategies for detecting enzymatic aminoacylation. Covered here are radioisotope assays, coupled enzyme assays for detecting pyrophosphate release or AMP production, ATP consumption, and acylation-dependent biotinylation for application in vitro. However, these assays typically use in vitro transcribed tRNA that lack modifications. Also described are cell-based assays that use mature tRNA, profiling isoacceptor specificity arising from aaRS-related diseases through tRNA capture and analysis of the released amino acid or by oxidation of unacylated tRNA and poly-A extension of the remaining pool. These approaches can further be combined with high-throughput sequencing of the tRNA pool to correlate acylation levels with modification profiling.

Combining acylation of tRNA with non-canonical amino acids (using flexizymes, chemoenzymatic synthesis or endogenous/evolved aaRS enzymes) with approaches to remove competing factors (vacating sense or stop codons) allows the reassignment of the genetic code. As a result, a vast pool of non-standard building blocks can be incorporated into peptides and proteins by the ribosome. Zhou and Obexer⁸ profile this breadth of possibilities in their contribution. They outline the functionality accessible with this diverse building block pool, including contributions from artificial enzymes with novel catalytic cores and natural product-like peptides. Next to genetic code reprogramming, further diversification in the latter can be accessed using chemical and/or enzymatic modification following translation (which can synergise with reprogramming by building on novel reaction handles). In a research contribution from Zhang, Goto, and Suga⁹ such an approach is described, showing engineering of a geranyltransferase to broaden its peptide specificity. Using their engineered enzyme, C-lipidation of histidine can be achieved in an efficient manner that is more tolerant of a broad peptide context, an important consideration in library building where sequence context is difficult to control.

The application of genetic code reprogramming and (chemo)enzymatic diversification into display approaches for drug discovery is profiled by Yin and Hipolito.¹⁰ The combination of genetic code reprogramming using flexizymes in the context of in vitro translation (coined flexible in vitro translation or the FIT system) together with mRNA display of the resulting peptides, together termed the Random non-standard Peptides Integrated Discovery (RaPID) system, has proven highly effective at generating hits against a diverse array of targets that are difficult to target using traditional modalities. In this review, a number of representative hits, targets, and structures are showcased. Also discussed are approaches for macrocyclisation of the libraries used to generate such hits, a critical consideration in allowing application of small peptides in a biological setting. Further development of initial hits is also discussed, covering two prominent cases wherein such hits have been developed further from the initial hit to a clinical candidate with oral bioavailability. An alternate but complementary modality, the highly knotted cysteine-rich plant-derived peptides ‘knottins’ (exemplified by the cyclotides), are discussed in the contribution by Xie, Craik, and coworkers.¹¹ These are composed of canonical amino acids, but their highly complex cyclic and disulfide cross-linked structure makes them highly stable in biological settings, while the loops between these critical cysteines are amenable to engineering or replacement with target-binding sequences (‘grafting’). Further discussed are approaches to display of such molecules in genetically encoded libraries, including suitable locations to break the head-to-tail macroycle for display and which loops are most amenable to grafting. This closes with a case study wherein this scaffold was engineered by mRNA display, giving 350-fold improved affinity.

Rounding out the collection are two articles focusing on specific applications of several of the above technologies in drug discovery efforts. In the contribution by Ullrich and Nitsche,¹² they describe drug discovery campaigns against targets related to the SARS-CoV-2 pandemic. These focused on the main protease and the spike protein as targets, and 8 hits are discussed in more detail. These include some with particularly challenging building blocks such as gamma amino acids targeting the main protease and with novel binding sites targeting the spike protein. Sakai, Sato, and Matsumoto¹³ present efforts to modulate activity of cytokine receptors using diverse modalities, including small molecules, oligonucleotide aptamers, antibody fragments, DARPins, and peptides. For this last category, an important contribution from the RaPID system was highlighted, wherein a potent binder of the Met receptor was dimerised to be able to induce receptor dimerisation, thereby converting it into a novel agonist. These same peptides could also be grafted into loops on recombinant proteins while maintaining binding of the peptide motif and function of the host protein (similar to the approach described above for cyclotides). These grafted proteins can host multiple loops, allowing tuned avidity and combinatorial binding, while at the same time expanding serum half life based on host protein re-uptake (when based on an antibody or Fc domain). This new approach to generation of a unique class of biologicals may again open new directions in therapeutic development.