The University of Electro-communications Taki Laboratory

Mission/Aim in brief:

Biological targeted covalent inhibitors (bioTCI), made of biologics, show stringent target specificity; we believe that bioTCI can show prolonged duration of drug action at low doses and greatly reduce the risk of adverse drug effects. Among biologics, middle molecules such as peptides and nucleic acids have problems of low stability in blood due to renal clearance and enzymatic degradation, however, we found that such middle-molecular bioTCIs escape the enzymatic degradation after forming a covalent bond with the target protein. These results suggest that bioTCIs may be used in vivo, and we would like to develop basic science and technology for the potential clinical application of this modality. We would like to conduct research that will shift from a modern society in which large quantities of drugs are continually supplied and consumed to one in which better drugs are supplied and consumed in smaller quantities.

Background and our achievement:

Targeted Covalent Inhibitor (TCI) is a substance that forms a covalent bond to a target biomolecule to semi-permanently inhibit the target,and the number of TCI-related publications and patents has increased rapidly in recent years(Bethany, Chem. Eng. News, 28, 2020). Since 2013, we have been conducting research and development focusing on mid-sized TCIs (i.e., bioTCIs), which can be targeted by strict multi-point molecular recognition like antibodies and can be easily chemically synthesized like small molecule inhibitors. Especially, we have reported for the first

(1)Peptidic TCI: a matchmaking reaction microenvironment formed during combinatorial selection of the bioTCI.

In order for TCIs to covalently bind to target proteins in the presence of a variety of other biomolecules, it is necessary to create highly bioorthogonal TCIs by incorporating an extremely weak electrophilic reactive group (warhead) at a specific position of a ligand with target specificity (Copeland, Methods Biochem. Anal., 340, 2013) In other words, fine-tuning the electrophilicity of warheads (Ojida, Bioorg. Med. Chem., 116386, 2021) determines TCI performance. Among the warheads reported so far, sulfur-fluorine exchange (SuFEx) reaction types are known to exhibit microenvironment-dependent reactivity (Sharpless, PNAS, 18808, 2019). Among them, fosylate group (aryl fluorosulfate; aryl-OSO2F), also called as a latent warhead, iscompletely inert that it isnot hydrolyzed in water, but becomes active only in a matchmaking (i.e., enzyme-like) microenvironment created between the target protein and an appropriate peptide during the reactivity & affinity-based co-selection process of the T7 phage display (Figure). The fosylate warhead minimized the promiscuous reaction during the library’s construction/selection, and a peptidic bioTCI was obtained with only 2 rounds of bio-panning. Non-specific and non-covalent interactions between target- unrelated biomolecules were eliminated during a very harsh washing step with a urea and SDS containing buffer, while the robust T7 phage still kept its infectivity（Chem. Commun., 5378, 2021; hot and front cover article）.

Serum stability of bioTCI:

We had the idea from the beginning to increase the molecular weight of the peptidic bioTCI by covalently binding it to the target to inhibit renal clearance, but in the course of our experiments, we unexpectedly found that simply covalently binding the peptidic bioTCI to the target improved stability by allowing it to escape protease degradation. Specifically, after covalently binding the combinatorially selected peptidic bioTCI to the target GST protein, the TCI was not hydrolyzed even after incubation in serum at 37°C for 24 hours.

Early history of combinatorial selection of peptidic TCI:

Theoretically, peptidic TCIs can be obtained via the combinatorial screening. Practically, control of the warhead reactivity during the library construction and selection is difficult, and the warhead in the library often forms promiscuous covalent bonds between biomolecules. To get around this problem, an indirect combinatorial method was implemented to first select for a targeting peptide using a mock-warhead-introduced peptide library on the T7 phage. After the selection and peptide-sequence identification, the desired target-selective covalent binding was observed when the unreactive mock warhead was replaced by a reactive warhead (Figure A). Later, in 2021, direct combinatorial screening via the phage display was independently reported by Bogyo’s group and us (Figure B, C)

(2) DNA-aptameric TCI: Neutralization (i.e, detoxification) of drug effects after irreversible binding

Peptidic TCI has a higher reaction specificity than small-molecule TCI.Although peptidic TCIs have higher specificity compared to small molecular TCIs and are thought to greatly reduce the risk of adverse reactions associated with TCIs, they are still irreversible bonds that are difficult to neutralize if an adverse reaction occurs unexpectedly after TCI dosing or if the drug effect is too strong. We have also developed a nucleotidic TCI that can neutralize the drug effect after covalent binding. For the antidote, an addition of a complementary-strand (CS) to a nucleic acid aptamer to form a helical double-stranded molecule eliminates the complex higher- order structure and neutralizes the drug's efficacy of the nucleotidic TCI even after covalent inhibition of the target protein. We created a tethered-TCI (TeTCI; Figure) since the protein domain forming the covalent bond is outside the actual aptamer docking domain. The long linker between the warhead and the aptamer serves as a chain that tethers the aptamer to the target protein. The TeTCI is conceptually different from most of the small molecule or even the peptidic TCI where the covalently binding residue is within or near the docking domain. As expected, the addition of the CS oligonucleotide against the aptamer sequence as an antidote reversed the protein inhibition, and the CS antidote rendered the protein-conjugated TeTCI nuclease- sensitive. The reversal by the CS antidote was swift, probably because the relatively long tether placed between the warhead and the aptamer did not interfere with either the double-strand (DS) formation between the aptamer and the CS, or the exposure of the DS towards the outside of the binding pocket on the target protein accessible to nuclease digestion (Figure). We have applied the same technology and confirmed the creation of a TeTCI targeting SARS-CoV-2 S-protein RBD domain from a previously reported DNA aptamer. Multiple warhead introduction into a single aptamer showed greater inhibition than the corresponding monoadduct.The aforementioned stability in serum has also been confirmed for the nucleotidic TeTCI, which are even less resistant to hydrolysis in vivo than peptides. Specifically, TeTCI acquired nuclease resistance after covalent binding when incubated in serum containing exo- and endonuclease (Figure Inset).

(3) Main directions for the near-future research:

・Establishment of a direct selection method for DNA-aptameric TCI
・Improvement of stability of bioTCIs before target binding
・Elucidation of the reaction mechanism in the matchmaking microenvironment formed on the target protein/bioTCIs
・Expansion and improvement of the overall molecular structure (i.e., local modality) of bioTCI
etc

Artificial molecule-library peptide conjugate on phage:

　We found, for the first time, that cysteine in the library peptide on T7 phage can be selectively modified, while the other cysteines on phage proteins remain intact (Fukunaga, Mol. BioSyst., 2013). Even bachelor students can make the artificial molecule-library peptide conjugate by JUST MIX the library phage and Cysteine (i.e., SH group) reactive reagents, which are widely synthesized and utilized for protein modification. Selected from the artificial library, we can make different 'intelligent targeted' biologics, such as turn-on fluorophores, covalent binders, macrocycles, etc.

　We have established a novel concept for detection of a target protein by using a keep-on-type fluorescent pharmacophore which was discovered from a dynamic combinatorial library of Schiff bases (Tabuchi, ABC, 2018). As shown in the figure below, the keep-on-pharmacophore exhibits bright fluorescence when irradiated by a UV hand lamp, and the presence of the target protein can be unambiguously detected by naked eye. When the target protein is absent, the keep-on-pharmacophore was degraded by hydrolysis, resulting that we can see no fluorescence.

Figure. Target protein (i.e. HSA)-specific sensing by a keep-on-type fluorescent pharmacophore. HSA was fixed in agarose matrix, incubated with the keep-on-pharmacophore, and UV light of 365 nm was irradiated (dotted-line inset; right panel), and the image was taken by a smartphone.

　N-terminal specific peptide/protein modification via the NEXT-A (N-terminal Extension of protein by Transferase and Aminoacyl-tRNA synthetase) reaction has been invented by us (e.g., Chem. Commun., 2011) . Perhaps, the most useful modification is azide (i.e., N3) introducion to the N-terminus of complicated proteins such as glycosilated antibodies (e.g., Hirasawa, Bioconj. Chem., 2019). This is the fastest enzymatic reaction where 'artificial' substrates can be recognized (Amino Acids, 2015); even radioactive PET probes with short half-life (ca. 1 hour) can be incorporated precisely to the N-terminus of peptide / protein (e.g., CTMC, 2015).

Topic 1: Combinatorially Screened Peptide as Targeted Covalent Binder

Brief summary
　Peptide-type covalent binders for a target protein were obtained by combinatorial screening of nonreactive bait-conjugated peptide libraries on T7 bacteriophage via the 10BASEd-T, followed by structural alteration of the bait into reactive warheads.

Motivation
　Finding targeted covalent binders is one of the cutting-edge disciplines such as biomedical sciences / chemical biology / pharmaceutical fields. Such binders can form permanent bonds to target proteins, and eternally deactivate them. Thus, they would be potentially useful as future medicines (i.e., covalent drugs as antibody-drug substitutes). Here is the first proof-of-concept study for finding peptide-type targeted covalent binder by a combinatorial screening, instead of a rational designing. As shown in the figure, we design a concept for finding peptide covalent binders from solvatochromic bait-modified peptide library[1,2] on T7 bacteriophage constructed via the gp10-based thioetherification (10BASEd-T),[3] followed by structure alteration into reactive warheads[4,5] as bioisosteres.

What has been achieved?
　First, specific introductions of the solvatochromic bait fragments into a designated Cys on displaying library peptides on a capsid protein (i.e., gp10) of T7 bacteriophage were conducted. This 10BASEd-T was carried out without side reactions or loss of phage infectivity.
　Second, from the bait-conjugated peptide library, target-protein (i.e., glutathione S-transferase; GST) specific binders were selected by biopanning; the peptide sequences were analyzed by using DNA sequencers, and consensus sequences for bait-conjugated peptides around the designated Cys were deduced.
　Third, to obtain GST-specific covalent binders, the bait fragments in the consensus peptides were altered into reactive warheads whose structures are shown in Fig. 1.
　Lastly, the chemically-synthesized covalent binders were incubated with GST to facilitate the conjugation.
　In case of using a photo-reactive warhead, ultraviolet (365 nm) irradiation must be needed for the conjugation between the covalent binder and GST; the irradiation simultaneously crosslinks the warhead to the ligand binding site and uncages the fluorescence property of the warhead by forming an intramolecular charge transfer (ICT) structure, which facilitates a rapid confirmation of the successful crosslinking using SDS-PAGE / fluorescence imaging.[5]
　In case of using an always-reactive warhead, such irradiation is unnecessary, and the warhead should theoretically react with limited numbers of amino acids (i.e., nucleophilic ones such as Tyr, Lys, Ser).[4]
　In both cases, site- and position-specific conjugation toward GST was successfully confirmed; the binder-conjugated GST in the SDS-PAGE gel was excised and digested with trypsin, followed by tandem mass spectrometry analysis of the fragment. The specific conjugation was also rationalized by molecular docking simulations of the covalent binders to GST using sievgene of myPresto.[4,5]

References
　1. Taki, M., Inoue, H., Mochizuki, K., Yang, J., and Ito, Y. (2016) Anal. Chem., 88, 1096-1099.
　2. Uematsu, S., Midorikawa, T., Ito, Y., and Taki, M. (2017) AIP Conf. Proc., 1807, 020028.
　3. Fukunaga, K., Hatanaka, T., Ito, Y., Minami, M., and Taki, M. (2014) Chem. Commun., 50, 3921-3923.
　4. Uematsu, S., Tabuchi, Y., Ito, Y., and Taki, M. (2018) Bioconj. Chem., 29, 1866-1871.
　5. Yatabe, K., Hisada, M., Tabuchi, Y., and Taki, M. (2018) Int. J. Mol. Sci., 19, 3682.

Topic 2: Half-life extension of targeted biologics inside of animal bodies

Motivation
　The use of middle molecules (i.e., targeted biologics) in therapeutic applications was limited because of their short half-lifved of these compounds inside bodies, and lack of a suitable bioconjugation method.

What has been achieved?
　We developed a semi-synthetic method of producing Fc-fusion compounds that, unlike conventional recombinant methods, can incorporate artificial middle molecules. The synthesis involves the introduction of an azide group (i.e., N3) to the Fc protein via the N-terminal extension (NEXT-A) reaction developed by our group, and bioconjugation of the middle molecule via strain-promoted azide-alkyne cycloaddition (i.e., SPAAC reaction). The Exenatide-Fc fusion produced through this method not only exhibited a longer plasma half-life, but also retained the biological activity of the original drug through optimization of the length of the spacer between the N-terminus of the protein and the Fc fragment. The method was also successfully applied to a peptide containing non-natural amino acids, a cyclic peptide, and DNA aptamers (Bioconj. Chem., 2019).

Topic 3: Turn-on / keep-on fluorescent molecules as targeted binders

Motivation
　We'd like to establish a general methodology for detection of target proteins by fluorescence-based sensing.

What has been achieved?
　By using an extended phage-display system namely 10BASEd-T [1], peptide-conjugated solvatochromic probes were combinatorially selected (Fig.1A, upper); a conjugation of an optimized peptide to a fluorophore strengthened both specificity and affinity to the target protein along with the solvatochromism, and consequently created a target-specific turn-on sensor [2]. Meanwhile, the fluorophore on the peptide-conjugated probe was altered to a novel photo-crosslinker possessing a caged-fluorescence property (Fig. 1A, lower). The crosslinker worked as a bioisostere of the solvatochromic fluorophore, and eventually, the peptide-conjugated crosslinker bound to the target protein. At this stage, irradiation with UV light simultaneously conjugated the crosslinker with the target at a specific site and uncaged the fluorescence property by forming an intramolecular charge transfer (ICT) structure [3]. Alternatively, as shown in Fig. 1B, a low-molecular-weight pharmacophore, as a targeted fluoroprobe, was also selected from a dynamic combinatorial library of Schiff bases by using size-exclusion chromatography. The identified pharmacophore retained its fluorescence when bound to the hydrophobic site of the target, whereas it lost because of hydrolysis when unbound. We defined it as ?keep-on? type fluorescence probe because the fluorescent pharmacophore is only kept intact when bound to the target [4].

Keywords: targeted biologics, pharmacophore, combinatorial screening, bioisostere.

References
　1. Chem. Commun., 50, (2014) 3921; inside cover article.
　2. Anal. Chem., 88, (2016) 1096; AIP Conf. Proc., 1807, (2017) 020028.
　3. Int. J. Mol. Sci., 19, (2018) 3682.
　4. Anal. Bioanal. Chem., 410, (2018) 6713.