2024 Volume 72 Issue 9 Pages 804-809
Protein-based enzymes are among the most efficient catalysts on our planet. A common feature of protein enzymes is that all catalytic amino acids occupy a limited, narrow space and face each other. In this study, we created a theoretical novel biomimetic molecule containing different multiple catalytic peptides. Although single peptides are far less catalytically efficient than protein enzymes, Octopus-arms-mimicking biomolecules containing eight different peptides (Octopuzymes) can efficiently catalyze organic reactions. Since structural information for extant protein enzymes, predicted enzymes based on genome data, and artificially designed enzymes is available for designing Octopuzymes, they could in theory mimic all protein enzyme reactions on our planet. Moreover, besides L-amino acids, peptides can contain D-amino acids, non-natural amino acids, chemically modified amino acids, nucleotides, vitamins, and manmade catalysts, leading to a huge expansion of catalytic space compared with extant protein enzymes. Once a reaction catalyzed by an Octopuzyme is defined, it could be rapidly evolvable via multiple amino acid substitutions on the eight peptides of Octopuzymes.
Ribozymes may have evolved during the prebiotic era1) (Fig. 1A(a)). Protein enzymes have catalytic amino acids in their pockets (Fig. 1A(b)). A common feature of the ribozyme and protein enzyme is a single and/or multiple catalytic core of atoms in a limited, narrow space (Figs. 1B(a), (b)). Proteins such as antibodies (Fig. 1C(a)) recognize target antigens (Fig. 1C(b)) using the same principle (Fig. 1B). Similarly, artificial enzymes2) (Fig. 1D(a)), phage display proteins3) (Fig. 1D(b)), and artificial drug binding proteins4) (Fig. 1D(c)) also harbor pockets (Fig. 1B).
(A) Catalysts. (a) Ribozyme purely composed by ribonucleotides (black curved line). Gray hexagons represent catalytic nucleotides. (b) Extant protein enzyme. The brown curved line represents the polypeptide. Two catalytic amino acids, Glu and Asp, reside near each other in the catalytic center. (B) Key and keyhole relationship. (a, b) Catalytic pocket in the form of a keyhole made of poly-ribonucleotide (black), and polypeptide (brown). Symbols for nucleotides, and amino acids are the same as in Fig. 1A. The substrate (key) interacts specifically with the catalyst (keyhole). (C) Interactions between antibodies and antigens. (a, b) Antibodies (keyhole) interact specifically with antigens (key). (D) Artificial keyhole and key. (a) Artificial enzyme. (b) Phage display. (c) Artificial drug binding protein. (E) Interaction between a key and a keyhole as described in Figs. 1A–D. (F) Human hand. (a, b) our hands can grasp many completely different target objects. (G) Different types of interactions. (a) Postulated ancestor of squid, Nectocaris pteryx (500 million years ago), with two arms. (b) Extant squid with ten arms. (H) (a, b) The eight arms of extant octopus can grasp numerous completely different target objects. (c) Robotic arms mimicking the arms of an octopus can grasp targets. (d) IDR of p53 interacts with a variety of p53-interacting proteins. (e) Nucleosome with eight N-tails. Modified histone residues on one and/or two tails can interact with recognizer proteins. However, there is no report that eight N-tails can simultaneously and cooperatively interact with a target recognizer.
Interactions between catalysts and substrates (Figs. 1A–1D) resemble key and keyhole relationships (Fig. 1E). By contrast, our hands can grasp target objects of any shape (Figs. 1F(a), (b)) in a completely different manner to that of a key and keyhole relationship. Although extinct ancient squid possessed only two arms (Fig. 1G(a)), extant squid and extant octopus have more than eight arms (Figs. 1G(b), 1H(a), (b)). Compared with the two arms of ancient squid, more than eight arms should be evolutionary beneficial for survival of extant cephalopods. Indeed, like our hands, octopus can efficiently grasp target objects of any shape using their arms (Figs. 1H(a), (b)). Artificial robotic arms mimicking those of an octopus can grasp target objects5) (Fig. 1H(c)). At the nano scale, the intrinsically disordered region (IDR) of the p53 protein recognizes a variety of differently shaped p53-interacting proteins6) (Fig. 1H(d)). Thus, IDRs can mimic Octopus arms at the nano scale. Although nucleosomes possess eight IDRs (N-terminal histone tails of H2A × 2, H2B × 2, H3 × 2, H4 × 2) like the eight arms of an octopus, only one or two of the eight IDRs participate in target recognition7) (Fig. 1H(e)). Thus, there is no extant natural biomolecule mimicking the coordinated action of eight IDRs like the arms of an octopus.
In this study, we propose a hypothetical framework for artificially generating novel biomolecules mimicking Octopus arms, which we call “Octopuzymes” (Figs. 2, 3). In theory, Octopuzymes can catalyze any biochemical reactions on our planet (Fig. 4). Moreover, they possess high evolvable potential for integrating all knowledge derived from our planet and the human brain (Fig. 5).
(A) Single peptide. Single peptides (brown curved line) containing one or several catalytic amino acids (colored dots) can catalyze organic reactions. Unlike protein enzymes, single peptide catalysts cannot efficiently grasp substrates, explaining their poor catalytic ability. (B) Peptide with cofactor. Peptides conjugated with cofactors such as metals, nucleotides, or vitamins can catalyze organic reactions more effectively than single peptides. Colored dots represent catalytic amino acids. Colored stars represent cofactors. (C) Multiple peptide catalysts. Multiple peptides (brown curved lines) containing catalytic amino acids (colored dots) can be anchored to Au atoms. Since the location of catalytic amino acids in each peptide differ, some cooperative and/or cascade reactions can be detected using multiple peptide-conjugated Au. (D) Structure of the catalytic pocket. Substrates (irregular closed black line) are surrounded by the catalytic pockets of protein enzymes. Interfaces (colored thick curved lines; numbers ①–⑧ represent different peptides) of enzymes interacting with substrates can be structurally characterized. Colored dots represent catalytic amino acids. (E) Theoretical biomolecule mimicking Octopus arms. Each of the eight peptides (numbers ①–⑧, colored differently and anchored to the yellow circle) has a different amino acid sequence. Some peptides carry catalytic amino acids (colored dots). Since the eight peptides differ and cooperatively mimic the catalytic pocket of a protein enzyme, the novel biomolecule mimicking Octopus arms can efficiently grasp substrates and catalyze organic reactions. These theoretical biomolecules are named Octopuzymes. (F) Octopuzyme grasping both cofactor and substrate. Octopuzymes can interact with cofactors, like protein enzymes. (G) The active site in the β-barrel structure of pyruvate kinase (PDB; 1PKN), which was referenced for generating the structures of Octopuzymes. The β-barrel structure is shown in green. (H) The predicted structure of an Octopuzyme, which consists of the β-barrel structure extracted from pyruvate kinase and the foundational cyclic peptide. Detailed computer simulation is described in Supplementary Figs. S1–S3.
(A) Parallel synthesis of the eight arms. Each protected peptide is prepared in parallel by solid-phase peptide synthesis. Briefly, starting from Lys-Gly resin with Nα-position by allyloxycarbony (Alloc), i.e., εLys(αAlloc)-Gly resin, protective peptide resins are constructed for the domains of unit ① or other units using the 9-fluorenylmethyloxycarbonyl (Fmoc) solid-phase peptide synthesis. Parallel synthesized units ②–⑧, each containing either Lys-Gly or D-Lys-Gly at the C-terminus, were cleaved from the solid-phase resin and utilized as their respective carboxy components. (B) Conjugation of protected peptide segments to scaffolds. After removing the Alloc group from the Lys residue of unit ①, the protected unit ② is condensed onto the free amino group on the resin. Subsequently, the Alloc group is removed, and each unit is sequentially condensed. The condensation of unit ⑧ leads to the construction of a branched protective peptide consisting of units ①–⑧, attached in sequence to the Lys side chain on the dKGKGdKGKGdKGKGdKGKG (dKGKG backbone) resin.27) (C) Macrocyclic lactamization of multi-branched peptides and deprotection. Following Fig. 3B, the branched protective peptide is cleaved from the resin, and the amino and carboxyl groups at the ends of the dKGKG backbone undergo dehydration-condensation to macrocyclize the foundational skeleton.28) Finally, the protecting groups are removed to obtain the desired Octopuzyme.
(A) Designing Octopuzymes based on the catalytic pocket structure of protein enzymes. (a) Structure of the catalytic pocket of an extant enzyme and an artificially designed enzyme deposited in the PDB. (b) The catalytic pocket of a protein enzyme could be simulated and reconstructed by cooperative actions among the eight peptides of an Octopuzyme. (B) Catalytic pocket predicted by AlphaFold/AlphaFold2/AlphaFold3. (a) Based on genome data, AlphaFold/AlphaFold2/AlphaFold3 can predict unknown enzyme structures. (b) Octopuzymes can mimic organic reactions catalyzed by unknown protein enzymes. (C) Metabolic map in vitro. (a) Sequential biochemical reactions form a metabolic map in vivo. Organic reactions catalyzed by each protein enzyme in the metabolic map can be sequentially replaced with the corresponding Octopuzyme. Thus, multiple Octopuzymes could be used to reconstruct a metabolic map in vitro. (b) Synthetic biology in vitro. Complex compounds (right column) can be produced via synthetic biology using transformation of multiple genes of particular species (left column) into recipient cells. The process for producing complex compounds could be reconstructed using multiple Octopuzymes. (D) Biosphere in vitro. Various cells (open rectangles) form biofilms. Since the survival of cells (colored rectangles) depends on compounds produced by other species, each of the colored cells cannot growth by itself. Various organic compounds (black dots) could be produced by an ensemble of biofilm cells (black curved line). Thus, the whole biochemical process (black curved line) could be reconstructed using multiple Octopuzymes in vitro.
(A) Expanded genetic code. (a) The standard genetic code only utilizes 20 standard amino acids. Genetic manipulation of the genetic code expands it to utilize non-standard L-amino acids (colored squares) as well as standard amino acids. (b) Since they are synthesized in vitro, Octopuzymes mimicking an expanded genetic code could be easily achieved using non-standard amino acids (colored squares). (B) Utilizing artificial residues. Since Octopuzymes are synthesized in vitro, a variety of components such as D-amino acids, chemically modified amino acids, nucleotides, and cofactors could be incorporated into some of the arms (peptides). (C) Mirror image enzymes. (a) Mirror image proteins composed of D-amino acids have been produced. In principle, mirror image protein enzymes catalyze opposite chiral compounds to their normal counterparts. (b) Octopuzymes mimicking mirror image protein enzymes could be created in vitro. (D) Interactions between proteins and other macromolecules. Besides small substrates recognized by protein enzymes, proteins can interact with other proteins, DNA, RNA, and many other types of macromolecules. Open circles represent such target macromolecules. Multiple gray and pink dots in the circle represent the interfaces of target molecules interacting with protein A and B, respectively. Interfaces of protein A and B interacting with the macromolecule are represented by black and red dots, respectively. Topological reconstruction of black and red dots using Octopuzymes (Figs. 2E, 3) could create novel affinity materials mimicking the interacting interfaces of protein A and B, respectively. (E) Multiple mutations. Simultaneous amino acid substitutions on the eight different peptides could rapidly improve the original enzymatic activity. (F) Nano-robots mimicking human chemists. Biochemistry is one field among many including inorganic-, organic-, and geo-chemistry. Since Octopuzymes are synthesized in vitro, all chemical knowledge could be incorporated into some of the arms (peptides). The four colored symbols (suits of cards) represent these different chemistry knowledge bases. Thus, Octopuzymes have high potential as nano-robots mimicking human chemists.
A two-step process was employed to construct a 3D model of Octopuzyme: (i) molecular dynamic (MD) simulations were performed on the eight peptides corresponding to the β-barrel of the pyruvate kinase active site, and (ii) a scaffold GK ring (Fig. 3C) was conjugated to the eight peptides. Two-step modeling was necessary because the Octopuzyme has multiple N-termini and no C-terminus, and thus it is not covered by the current force field used for MD simulation.
The initial coordinates of the eight peptides with pyruvic acid for MD simulations were based on those of the reported X-ray crystal structures of pyruvate kinase (PDB ID: 1PKN) using BIOVIA Discovery Studio 2021. The CHARMM36 force field was applied.8) The parameter settings for the force field were performed using CHARMM-GUI.9) The simulation time was set to 50 ns. The dimensions of simulation system were 68 × 68 × 68 Å, ensuring that the distance between periodic images of the peptides was 18 Å. Explicit water molecules were placed for each coordinate of the structure (number of water molecules: 9444), and TIP3P was used as the force field template.10) In order to neutralize the system, counter ions (K+ and Cl−) were added. The entire structure was minimized using the steepest descent algorithm, with 10000 steps taken to eliminate any distortion. Subsequently, heating was conducted under an nVT ensemble at 500 K. Following this, production was performed under an nPT ensemble at 300 K with a time step of 2 fs. Throughout the simulation, the root-mean-square-deviation (RMSD) values for each C-terminal amino acid residue of the eight peptides were constrained to be less than 2 Å from the initial structure. GENESIS 1.7.0 (RIKEN CCS, Kobe, Japan) was employed for the MD simulation, RMSD analysis (all atom of the peptides), and hydrogen bonding analysis. Hydrogen bonding patterns were analyzed for the entire trajectory from 0 to 50 ns, and those with a formation frequency exceeding 90% were listed.
Subsequently, the simulated peptides coordinates were utilized to conjugate the GK ring, thereby constructing a 3D structural model of Octopuzyme. The coordinates of the GK ring were generated using MOE 2022.0211) and BIOVIA Discovery Studio 2021. The amino acid sequence of the GK ring is shown below:
To conjugate eight peptides to the GK ring (Fig. 3C), each peptide C-terminus was modified with a methylamino group using CHARMM-GUI. The GK ring was coupled with the eight peptides by setting the nitrogen of the Lys side chain of the GK ring to the nitrogen of the methylamino group at the C-termini of the eight peptides. Finally, the GK ring part was simulated with the molecular mechanics method using BIOVIA Discovery Studio 2021, resulting in a 3D structural model of Octopuzyme.
Although 224 types of reactions catalyzed by peptides are listed in a recent review,12) the catalytic abilities of single peptides are far inferior to those of protein enzymes (Fig. 2A). Even when peptides are combined with metals, nucleotides, and/or vitamins, they can catalyze some reactions13–15) (Fig. 2B) but the catalytic ability remains poor. It has been reported that multiple peptides anchored on a gold (Au) surface can catalyze reactions16) (Fig. 2C). Although cooperative and cascade reactions among multiple peptides on Au have been detected,16) the catalytic ability remains inferior compared to protein enzymes.
Novel Theoretical Artificial Enzymes Mimicking Octopus Arms: Eight Different Peptides Anchored in a Defined OrderA huge quantity of structural data for enzyme-substrate complexes has been deposited in the Protein Data Bank (PDB). Thus, amino acid sequences of multiple peptides surrounding substrates can be extracted from the PDB (Fig. 2D).
In principle, more than three arms are required to grasp a substrate in a three-dimensional structure. Indeed, an artificially designed drug binding protein has been reported containing four α-helixes4) (Fig. 1D(c)). Although a single finger cannot grasp a targeted object, the cooperative action of five fingers anchored to the palm can grasp it efficiently. Moreover, ten fingers of both hands are more efficient for grasping than five fingers of one hand. Taking into account these considerations and the evolutionary advantages conferred by eight arms in squid and octopus (Figs. 1G(b), 1H(a), (b)), we tentatively suggest that the eight different peptides are sufficient for substrate grasping.
If each of the eight different peptides has the ability to interact with different parts of a substrate (Figs. 2D, 2E), eight different peptides anchored in a defined order may be able to coordinately grasp substrates. If so, catalytic amino acids embedded in several peptides in a cis and trans manner may efficiently catalyze organic reactions (Fig. 2E). Such putative novel biomolecules mimic the arms of an octopus, hence we named them ‘Octopuzymes’ (Fig. 2E). In theory, Octopuzymes can simultaneously grasp both substrates and cofactors (metals, nucleotides, vitamins; Fig. 2F), and cofactors could be covalently linked to one of the peptides.
For example, some enzymes, such as pyruvate kinase, have a β-barrel structure made up of eight β-strands that also serve as the active site (Fig. 2G). Our proposed Octopuzymes have eight arms. Achieving the desired structure may be difficult if the number of peptide chains on the cyclic peptide is too low. Using computer simulation, it is anticipated that forming hydrogen bonds between the eight peptides, similar to the β-barrel structure of pyruvate kinase, and then anchoring this structure to the base will enable the formation of the desired structure to some extent (Fig. 2H). Currently, it is difficult to simulate interaction between Octopuzymes and substrates (Figs. 4A, B), however, this technical problem will be solved in future.
Like proteins, it is undeniable that Octopuzymes are unstable under extreme temperature or pH conditions. However, due to its unique structure, it is expected to tolerate not only physiological conditions but also slight changes in temperature or pH.
As shown in Fig. 3, conventional peptide synthesis in vitro could be used to construct Octopuzymes (details shown in the legend of Fig. 3). As an example of Octopuzymes synthesis, we present a route for the chemical synthesis of cyclic peptides that employs a fragment condensation strategy with Lys-Gly and D-Lys-Gly repeating structures as scaffolds to attach eight different units to the amino groups of their Lys/D-Lys side chains. The Lys-Gly and D-Lys-Gly repeats are designed so that the Lys/D-Lys side chain is assembled on one side of the cyclic peptide. Protein expression using genetic recombination technology causes problems such as the formation of inclusion bodies or inactive proteins due to inadequate folding of the protein. In the case of Octopuzymes, not only do the eight peptides not fold correctly, but even if they achieve the desired conformation, the amino acid sequence may make them less water soluble. Fortunately, peptides are chemically synthesized and it is believed that this problem can be addressed by replacing some of the amino acids with hydrophilic amino acids.17,18)
Importantly, peptides can be automatically synthesized industrially, enabling the automatic construction any newly designed Octopuzymes within a day.
Mimicking All Biochemical Reactions on Our PlanetAn enormous number of structurally defined enzymes have been deposited in the PDB (Fig. 4A(a)). Moreover, many artificially designed enzymes have been reported19) (Fig. 4A(a)). This structural information can be used to construct a variety of Octopuzymes (Fig. 4A(b)).
AlphaFold320) can predict interactions between substrates and the structure of any enzyme translated from genome data (Fig. 4B(a)), which could lead to the generation of Octopuzymes capable of catalyzing any biochemical reaction (Fig. 4B(b)).
A series of Octopuzymes could be embedded into a microfluidic device,21) mimicking biochemical reactions of any metabolic map (Fig. 4C(a)). Currently, synthetic biology uses both a recipient cell (such as Escherichia coli or Saccharomyces cerevisiae) and dozens of external genes22) (Fig. 4C(b)). In principle, instead of synthetic biology, any compounds such as taxols, cannabinoids, artemisinins, and morphines could be synthesized in vitro by a combination of Octopuzymes.
Metagenome analyses has revealed that the majority of cells forming biofilms are non-cultivated bacteria and non-cultivated archaea.23) The survival of each cell essentially depends on compounds produced by other species. Thus, a series of Octopuzymes could create compounds produced by elements of the biosphere such as biofilms (Fig. 4D).
Incorporating Non-standard Amino Acids into OctopuzymesThe extant protein enzyme is composed of glycine and 19 standard L-amino acids. Expansion of the genetic code enables the incorporation of non-standard L-amino acids into the protein24) (Fig. 5A(a)). Indeed, a boronic-acid-containing designer enzyme was recently reported.25) Thus, this expansion can be easily applied to Octopuzymes containing non-standard amino acids (Fig. 5A(b)). Since each peptide is produced in vitro, Octopuzymes containing chemically modified amino acids, D-amino acids, nucleotides, and cofactors (Fig. 2B) could be easily constructed (Fig. 5B).
Opposite Chiral CompoundsProtein enzymes made from L-amino acid usually bind only one chiral substrate and generate one chiral product (Fig. 5C(a), left panel). Since the mirror image T7 RNA polymerase successfully produces the mirror image of 5S, 16S, and 23S ribosomal RNA,26) any mirror image protein enzyme made from D-amino acids can produce opposite chiral compounds (Fig. 5C(a), right panel). Thus, mirror image Octopuzymes containing peptides made from D-amino acids anchored onto a mirror image base could generate mirror image extant biomolecules (Fig. 5C(b)).
Affinity Materials Mimicking Octopus ArmsInteractions between antibodies and antigens reflect a key and keyhole relationship (Fig. 1E). The interaction interface between a macromolecule and a protein (protein A, black dots; protein B, red dots) has been deposited in the PDB (Fig. 5D(a)). Thus, affinity materials mimicking the interface of protein A (Fig. 5D(b)) and protein B (Fig. 5D(c)), which can be generated via the same principle as that of Octopuzymes, should be able to grasp target macromolecules.
Octopuzymes Possess High Potential EvolvabilitySince the eight peptide arms of Octopuzymes are made from amino acids, multiple and simultaneous amino acid substitutions could rapidly and dramatically improve their original biochemical activities, as occurs in extant protein enzymes through evolution (Fig. 5E).
The theoretical Octopuzymes in this study (Fig. 2) could be produced experimentally in vitro (Fig. 3). Although Octopuzymes do not yet exist, when created they could revolutionize all fields of chemistry and biochemistry. All knowledge on inorganic-, organic-, geo-, and bio-chemistries could be incorporated and integrated to improve Octopuzyme activity (Fig. 5F). Thus, Octopuzymes could be considered nano-robots with the potential to mimic all known chemistries. In future work, we intend to generate the first Octopuzymes, then mass-produce and apply them in an attempt to solve various global chemistry-related issues.
Since the theoretical research was based on several studies done by past undergraduate students, we thank the following students: S. Shiraiwa, Y. Nagao, and Y. Kan-no. We thank Ms. M. Seki for illustrating some of the figures, and Re Suyong (NIBIOHN) for his invaluable assistance with MD simulations. The MD simulations were conducted using the supercomputing resources at the Cyberscience Center, Tohoku University. This study was financially supported by the Tohoku Medical and Pharmaceutical University, whose founding spirit is “We will open the gate of truth.” We extend our gratitude to Dr. M. Takayanagi, chairman of the University.
This research received no external funding.
All conceptional ideas were discussed by AA and MS. Computer simulation and model building were performed by SO and YY. AA and SO equally contributed to this study, and thus are listed as co-first authors. The manuscript was drafted by AA and MS, and all authors contributed to subsequent editing. All authors read and approved the final manuscript.
The authors declare no conflict of interest.
All data are included in the published article.
This article contains supplementary materials.
