Introduction

GPAs are a family of natural and semisynthetic glycosylated non-ribosomal peptides that comprise tricyclic or tetracyclic polypeptide scaffold and display antibacterial activity against Gram-positive organisms through binding to the C-terminal D-alanyl-D-alanine of the lipid II bacterial cell-wall precursor via H-bonds, preventing crosslinking of the peptidoglycan and ultimately inhibit the synthesis of cell wall (Butler et al. 2014; Vimberg et al. 2019). However, recent research discovered that GPA oritavancin inhibited bacterial activity by jointly blocking the biosynthesis of wall teichoic acid and peptidoglycan (Singh et al. 2017). In addition, the new actional mechanism of the type V GPAs complestatin, GP6738, and corbomycin was discovered, which presented that these antibiotics could inhibit the action of autolysins by binding to peptidoglycan and overcome the D-Ala-D-Lac GPA resistance (Culp et al. 2020). Therefore, these discoveries of new action mechanisms provide the possibility for the development of new GPAs.

Generally, the antibacterial spectrum of GPAs excludes Gram-negative species due to their physiochemical properties precluding transit through the outer membrane and blocking target to the lipid II in Gram-negative pathogens (Schaenzer and Wright 2020). In fact, it had been demonstrated that the vancomycin analogues with the characteristics of lipophilic and cationic modification could strongly destroy bacterial membranes and thereby overcome the inherent resistance of Gram-negative pathogens (Yarlagadda et al. 2016). Research showed that the cold stress cultivation made Gram-negative Escherichia coli susceptible to glycopeptide antibiotics by altering outer membrane integrity (Stokes et al. 2016). In addition, many studies had confirmed the inhibitory effect of glycopeptide antibiotics on viruses, including Ebola, MERS-Cov, and SARS-Cov virus (Acharya et al. 2022; Zhou et al. 2016).

GPAs are the last line of defense against Gram-positive drug-resistant bacteria. However, with the clinical application of first- and second-generation GPAs, drug-resistant bacteria have emerged. Antibiotic resistance continues to be a severe threat to modern medicine. Resistance to antibiotics occurs through a variety of molecular strategies. In particular, resistance to glycopeptide antibiotics manifests through the activity of enzymes such as dipeptidases VanH, VanA, and VanX that chemically alter elements of the cell envelope required for antibiotic binding (Frasch et al. 2015; Schaenzer and Wright 2020). In addition, resistance to GPAs can also be performed through intermediate strategies which are found in Vancomycin-intermediate S. aureus (VISA) (Hu et al. 2016).

GPAs are divided into five distinct structural subclasses (I–V) according to the substituents and the type of residues at positions 1 and 3 of the polypeptide (Butler et al. 2014) (Fig. 1). Type I GPAs, represented by vancomycin, contain three side-chain crosslinked rings, including an A-B ring, a C-O-D ring, and a D-O-E ring. Type II GPAs (exemplified by avoparcin) possess the same crosslinked structure as Type I GPAs, unlike Type I, AA1 and AA3 residues of Type II GPAs are modified by aromatic amino acids rather than aliphatic groups. Type III and IV GPAs include an additional crosslinked F-O-G ring between aromatic groups at AA1 and AA3 residues compared to Type I/II GPAs. The difference between Type III and IV GPAs is that there is an acyl chain substitute in the structural scaffold of Type IV, but there isn’t in Type III. Type V GPAs contain a typical tryptophan (Trp) moiety linked to the central residue Hpg to form a conserved Trp-Hpg-(m)Tyr motif and are not generally glycosylated. The core scaffold of GPAs, known as the aglycone, requires the essential crosslinked rings, including A-B, C-O-D, and D-O-E, which are necessary for GPAs to present their antibacterial activity (Fig. 1; Table 1).

Fig. 1
figure 1

Schematic representation of GPAs grouped into five (I − V) structural subtypes. The core structural scaffold of GPAs (I–V) is marked in red. The typical chemical structure of different type GPA is presented using red core scaffold plus specific tailoring moieties marked in the enclosing ring. Type I: red scaffold plus black circle; Type II: red scaffold plus blue ellipse; Type III: red scaffold plus pink dashed ellipse; Type IV: red scaffold plus pink dashed ellipse and green dashed ellipse; Type V: red scaffold plus cyan ellipse

Full size image
Table 1 Structural diversities of natural GPA family
Full size table

Diversity and conservation of GPA BGCs

Generally, GPA BGCs are modularly composed of the genes required for the amino acid precursor supply, peptide assembly, and scaffold crosslinking by non-ribosomal peptide synthetases (NRPSs) and cytochrome P450s, tailoring enzymes (halogenation, sulfation, glycosylation, methylation, and acylation), transport, regulation, and resistance. Recently, Wright’s group revealed that the abilities of glycopeptide biosynthesis and resistance in Actinobacteria emerged approximately 150–400 million years ago by phylogenetic analysis based on the sources of 71 GPA BGCs. Especially noted in this research, the precursors of GPA biosynthesis are much older than other components (Waglechner et al. 2019).

NRPSs are a class of ribosome-independent large modular enzymes involved in the biosynthesis of many different peptide-derived natural products (Fig. 2). A typical linear NRPS structure consists of multiple modules, and each module on the assembly line is responsible for transferring and inserting an amino acid monomer into the final peptide. Typically, these modules include adenylation (A)-domain (responsible for amino acid selection and activation), condensation (C)-domain (responsible for peptide bond formation), thioesterase (TE)-domain (responsible for cleaving the mature peptide from the NRPS), epimerization (E)-domain (responsible for the altering the stereochemistry of peptide residues), methyltransferase (MT)-domain (responsible for methylation of amino acid), and P450 recruitment (X)-domain (responsible for interaction with P450 oxygenases). Except above module, peptidyl carrier proteins (PCPs) act as the attachment point for all biosynthetic intermediates during peptide assembly (Izore and Cryle 2018). In NRPS modules, the specific combination of A domain, C domain, and E domain determines the corresponding peptide sequence, which can be as a vital machinery to guide the development of new GPAs.

Fig. 2
figure 2

The amino acid constitution and modular assembly of typical GPAs (Type I-V). A adenylation domain, C condensation domain, PCP peptidyl carrier protein domain, TE thioesterase domain, E epimerization domain,  X P450 recruitment domain. The specific amino acids are highlighted by red

Full size image

A domains, requiring ATP activation, are composed of the larger N-terminal subdomain and the smaller C-terminal subdomain and can recognize and activate more than 500 different substrates for peptide synthesis (Walsh et al. 2013). Previous studies had demonstrated that redesigning the A domain could extend a range of non-natural monomers and produce different novel peptides (Weist et al. 2004). C domains are responsible for peptide bond formation and the stereochemical selectivity of the substrates with the L- or D-configuration, indicating a tight relationship between the C domains and E domains (Chen et al. 2016). Generally, A domains mainly select and activate L-configured monomers, requiring E domains to epimerize the substrate from the L to the D form. TE domains in NRPS modular are typically located at the end of the NRPS assembly line, which is responsible for the cleavage of thioester products and for the peptide cyclization (Owen et al. 2016). The conserved X domains without catalytical activity are located in the final module of all GPA NRPS lines. They are responsible for recruiting P450s to the NRPS-bound peptide to perform the crosslink of linear peptide (Haslinger et al. 2015). Additionally, a previous study revealed that the P450 recruitment abided by a continuous association/dissociation model between the P450 enzymes and the NRPS, namely the P450 scanning (Peschke et al. 2016).

Ingenious biosynthetic cascades of GPAs

Preparation of amino acids and assembly of linear polypeptide

Within the biosynthesis of GPAs, the amino acids as monomers include non-proteinogenic amino acids and common amino acids. In general, non-proteinogenic amino acids are derived from the classic 20 proteinogenic amino acids, but there are some special cases, which are produced de novo according to the corresponding modules in BGCs of GPAs.

The common amino acid precursors in the GPA family mainly include Asn, Leu, Ala, Phe, Glu, Trp, and Tyr with different epitopic configurations. The other non-proteinogenic amino acids are hydroxyphenylglycine (Hpg), 3,5-dihydroxyphenylglycine (Dpg), and β-hydroxytyrosine (Bht) (Fig. 2). The three amino acids share common aromatic amino acid biosynthetic pathways containing the crucial enzymes 3-deoxy-D-arabinoheptulosonate-7-phosphate synthase (DAHPS), chorismate mutase (CM), and prephenate dehydrogenase (PDH) (Waglechner et al. 2019).

During the NRPS-mediated biosynthesis of GPAs, A domains, C domains, and Type-II TE domains can act collaboratively to ensure the incorporation of the correct amino acid to the polypeptide line, even if the A domain displays weak specificities for substrate selection. Furthermore, the amino acid modification in trans can be selectively controlled during non-ribosomal peptide biosynthesis through the exchange of selective C-domains (Kaniusaite et al. 2019). In addition, the C domain has a broad tolerance to changes to the amino acids. It can select both L- and D-configured peptides and significantly prefer linear peptides over crosslinked ones. Although the C domain is regarded as a stereochemical gatekeeper during GPA synthesis, its selectivity depends on the presence or absence of the adjacent E domain. Furthermore, the peptide bond formation as an essential role for the C domain is tightly related to the rate of monomer activation exerted by the neighboring A domain (Schoppet et al. 2019).

Crosslinking of GPAs guided by P450s 

Oxidative crosslinking of aromatic side chains during the biosynthesis of GPAs is a vital modification depending on the catalytic action of cytochrome P450 enzymes (P450s). P450s are a superfamily of monooxygenases given their diverse oxidative reactions and abundant catalytic substrates by a cysteine thiolate-ligated heme iron moiety which is crucial for their ability to generate extremely powerful oxidization. During the biosynthesis of GPAs, different P450s generally catalyze the respective oxidative crosslinking of aromatic side chains. Those known as “Oxy enzymes” play a crucial role in the biosynthesis of GPAs (Haslinger and Cryle 2016).

During the scaffold crosslinking of GPAs, the P450s are recruited to the PCP-bound peptide substrate by binding the same site on the X domain surface. Therefore, these cytochrome oxygenases should compete the binding site of X domain. The related research had verified that the P450s enzymes performed the reactions as the scanning model depending on the experimental evidence of the continuous association/dissociation of P450s with the peptide substrates (Peschke et al. 2016). In addition, the P450 proteins have many Tyr residues to maintain the typical heme orientation and avoid potential oxidative damage based on crystallographic experiments (Greule et al. 2022).

In the first step of substrate selection in the crosslinking oxidation, different P450s present distinct binding abilities to the stereochemical substrate. For example, OxyBtei has a higher tolerance to the seventh amino acid’s stereochemical change than OxyAtei. During the complicated crosslinking process, there exist one or more intermediates of the P450-substrate complex that perform a continuous cycle of combination and dissociation with the NRPS. In vivo and in vitro, previous studies presented that the conserved P450s conducted the oxygenation in a specific order. Generally, the P450s OxyA, OxyB, OxyC, and OxyE act sequentially on the polypeptide precursor in OxyB-(OxyE)-OxyA-OxyC to produce the GPAs (Forneris and Seyedsayamdost 2018; Peschke et al. 2016; Tailhades et al. 2020; Zhao et al. 2020). In another case of teicoplanin biosynthesis, the TE domain located in the last module of Tcp12 is responsible for acting as a logic gate to ensure the formation of mature linear peptide during NRPS biosynthesis. Furthermore, once the Oxy cascade is completed, the finished aglycone is rapidly hydrolyzed to be released (Peschke et al. 2018).

Here, we summarized the classification and protein structure of P450s involved in the oxidational biosynthesis of GPAs (Fig. 3 and Table 2). In general, the D-O-E ring is catalyzed by the OxyA and its homologous enzymes; the OxyB enzymes oxidize the C-O-D ring; the A-(O)-B ring is formed by the OxyC and its homologous counterparts; the F-O-G ring is catalyzed by the OxyE and its homologous enzymes. In particular, oxygenase OxyAkis conducts the formation of a D-E ring in the biosynthesis of kistamicin belonging to type V GPAs. As well, the P450 OxyCkis oxidizes the A-O-B ring (Fig. 3). These obvious differences in P450s indicate that the Type V GPAs are atypical and structurally diverse. In another special case, the enzyme OxyD involved in vancomycin biosynthesis performs hydroxylation (Fig. 3). Additionally, the catalytic activity of the OxyCkis enzyme is flexible and can install two crosslinks of A-O-B ring and C-O-D ring in GPA kistamicin (Greule et al. 2019). Furthermore, the Oxy enzymes have broad tolerance for modification to the N-terminal peptide of GPAs and result in generating diverse derivatives. For instance, a method of click chemistry was used to diversify the cyclic peptide based on the wide tolerance of the enzyme OxyB/A for substituted amino acids at position 3 of GPA (Tailhades et al. 2020).

Table 2 Comparison of different Oxy enzymes in GPA biosynthesis
Full size table
Fig. 3
figure 3

P450 oxygenases involved in the crosslinking scaffold formation during GPAs biosynthesis. The crystallographic structures of some P450s loaded in the Protein Data Bank (PDB) database are showed. NA means no available of crystallographic information from PDB. kis kistamicin, Oxy vancomycin, Sta A47934, tei teicoplanin, Cep chloroeremomycin, bal balhimycin, Dbv A40926

Full size image

Glycosylation of GPAs

Glycosylation is one of the essential modifications which are performed by glycosyltransferases (GTs) found in different BGCs of GPAs. Generally, glycosylation and acylation are necessary for the antimicrobial activities of some GPAs, such as teicoplanin and A40926, belonging to Type IV GPAs. Furthermore, it should be noted that all types of GPAs are modified by glycosylation except for Type V and Type III GPA A47934 (Fig. 4) (Pootoolal et al. 2002; Xu et al. 2022).

Fig. 4
figure 4

The monosaccharide moieties and their corresponding glycosyltransferases of GPAs. Solid ellipses contain the different sugars; dashed ellipses encircle glycosyltransferases. B Balhimycin, V Vancomycin, Dbv A40926, K Keratinimicin A, C Chloroeremomycin, Orf Ristocetin A, Pek Pekiskomycin, Tei Teicoplanin, Auk UK-68,597

Full size image

As for sugar donors, their resources are from primary metabolism or non-conventional metabolism. In detail, the primary metabolite sugar, namely conventional sugars, are D-glucose, D-mannose, D-arabinose, N-acetylglucosamine (GlcNAc), and L-rhamnose. For example, Type IV GPAs, including teicoplanin and A40926, are decorated with sugars deriving from primary metabolites such as α-D-mannose and GlcNAc. Non-conventional sugar residues include L-vancosamine, L-epivancosamine, L-4-oxovancosamine, L-ristosamine, and L-actinosamine. For instance, Type I GPAs vancomycin, balhimycin, and chloroeremonycin have sugar moiety of non-conventional monosaccharides such as L-vancosamine, L-4-oxovancosamine, and L-epivancosamine. These non-conventional monosaccharides are biosynthesized by enzymes from the BGCs of GPAs. Furthermore, all GTs involved in the glycosylation belong to two families GT1 and GT39, according to the Carbohydrate-Active enZYmes Databse (CAZy, http://www.cazy.org) (Drula et al. 2022). The GT1 family is responsible for transferring non-conventional aminosugars and conventional D-glucose, D-arabinose, GlcNAc, and L-rhamnose, and requires the dTDP- or UDP-activated sugar substrates. These enzymic proteins characterize a unique two-domain scaffold, including C domains as the recognition sites for the donor NDP-activated sugars and N domains as the acceptor binding sites for polypeptide aglycones (Zhang et al. 2020). As for the second GT family, the installation of conventional sugar D-mannose is the only case performed by the GT39 family, which requires D-mannosyl-1-phosphoundecaprenol as a donor substrate.

Generally, the GT1 enzymes prefer the glycosylation sites at aa-4 (Hpg) and aa-6 (Bht) of the aglycone or add additional monosaccharides to already existing mono/di/trisaccharides at aa-4. The other family GT39 is responsible for attaching mannose at aa-7 (Dpg).

Halogenation of GPAs

Halogenation is also an important tailoring modification for the stability and activity of natural GPAs, in which chlorination is the most common halogenation modification. For example, dechlorinated vancomycin showed a decrease in antibacterial activity (Wadzinski et al. 2016). Similarly, the halogenation of A47934 is not only necessary for antimicrobial activity but also for avoiding the induction of resistance mechanisms in some pathogens (Yim et al. 2018). In contrast, the antimicrobial activity of dechlorinated A40926 is similar to that of A40926 (Beltrametti et al. 2003).

So far, the known GPA BGCs contain up to two halogenases, and can lead to six chlorinations at most in the known GPA structures. Generally, one halogenase gene is responsible for one to two chlorinations. If there exist two halogenases in the BGCs, three or four chlorinations will be installed in the GPA structure. As an exception, StaI encoding only one halogenase in the sta cluster of A47934 presents three halogenations at different moieties of amino acids on the A47934 scaffold (Yim et al. 2018). In another specific case, atypical GPA complestatin contains six chlorine atoms derived from three different amino acid monomers of 1,3-dichloro-Hpg, 3,5-dichloro-Hpg, and 3,5-dichloro-hydroxybenzoylformate catalyzed by nonheme metal-free halogenase ComH (Chiu et al. 2001).

During the halogenation, halogenase recognizes the amino acid substrate loaded to the corresponding PCP domain, which is necessary and contributes some role in the procedure. Additionally, the halogenation of GPAs can improve oxidative processing by assisting the P450s to maintain higher activity. This reveals that GPA chlorination occurs before the oxidative crosslinking cascade (Peschke et al. 2017).

Acylation of GPAs

Compared with Types I-III GPAs, Type IV GPAs include an emblematic modification of N-lipidated glucosamine on the residue aa4. This particular feature inspired the development of novel semisynthetic GPA analogues such as oritavancin and telavancin. Therefore, these GPAs featured a long aliphatic acyl side chain and are also named lipoglycopeptide antibiotics. This type of antibiotic is powered by the N-acyltransferases (NAT), which introduce the N-acylation on glucosamine at the central residue of pseudoglycone. The X-ray crystallographic analysis of Orf11 protein (N-acyltransferase from tei cluster) demonstrates that it can dock the bulky or lengthy acyl moiety and thus allows diversity in novel derivatives. Based on the above characteristics of NAT, many teicoplanin analogues with different acyl modifications are developed and show excellent antibacterial activities (Lyu et al. 2014).

New development of GPAs analogues

Over the past 50 years, GPAs have been regarded as the last resort to cure serious Gram-positive infections. However, drug resistance has always been a thorny problem accompanying the clinical application of GPAs. The emergence of glycopeptide resistance has resulted in developing the synthetic and semisynthetic glycopeptide analogues to target acquired resistance. Recently, many methodologies, including chemistry, biochemistry, enzymology, microbiology, and the mixing method based on the above methods, have been applied to develop new GPAs with higher activity and more broad antibacterial spectra. Here, we summed up the progress of some GPA analogues as described below.

Vancomycin analogues

Although more novel semisynthetic GPAs such as telavancin, dalbavancin, and oritavancin have been used in clinic, vancomycin, a representation of the first-generation GPA, has been a mainstream antibiotic against serious Gram-positive infections. Over the past few years, many analogues of vancomycin have been developed via different modification strategies, including C-terminal modification (membrane targeting approaches) (Blaskovich et al. 2018), core scaffold modifications (Okano et al. 2017), aryl ring functionalization (Guan et al. 2018), hydroxyl modifications (Han and Miller 2013), N-terminal modifications(Chang et al. 2013), glycopeptide dimers (Silverman et al. 2017) and conjugates (Varisco et al. 2014). For example, a series of compounds named vancapticins 7 is developed by modifying the C-terminus of vancomycin with a modular assembly containing a bis-amine linker moiety, a cationic peptide, and a hydrophobic cap. These analogues are designed to target bacterial membranes with the aid of hydrophobic caps and cationic peptides. The results of antibacterial tests show that the combined modifications present an over 100-fold increase in potency against MRSA (Blaskovich et al. 2018). Similarly, the other studies reported that the modification of vancomycin C-terminus including the zinc-binding dipicolyl moiety, the benzoxaborole group, the quaternary ammonium propylamine, and positively-charged amino acids (such as lysine and arginine) contribute to the increase of antibacterial activity and the extend of antimicrobial spectra (Antonoplis et al. 2018; Antonoplis et al. 2019).

Teicoplanin analogues

Teicoplanin, which belongs to the second-generation and Type IV GPA, has a typical long-chain N-acyl moiety, resulting in teicoplanin being a mixture of five related derivatives and cannot form non-covalent dimers because of the apolar tag. In order to develop new teicoplanin analogues, teicoplanin generally is treated to remove the one or two β-D-GlcNAc substituents and further form new analogues by lipidation, N-terminal guanidalization, and dimerization. These new teicoplanin analogues show high activity against VanA teicoplanin-resistant enterococci (Bereczki et al. 2022a; Szucs et al. 2019, 2020). In addition, the latest research reported that teicoplanin derivatives bearing hydrophobic or super-basic side chain showed activity against severe acute respiratory syndrome coronavirus (SARS-CoV-2) in human Calu-3 cells and HCoV-229E in HEL cells (Bereczki et al. 2022b).

A40926 analogues

A40926 is the precursor of the third-generation GPA dalbavancin and is biosynthesized by actinobacteria ATCC 39,727. A previous study reported that a series of alkylated derivatives were synthesized on the deacyl A40926 scaffold. However, the derivatives harboring only the lipophilic mono- or dialkylation of the amino groups showed poor antibacterial activity. Interestingly, further modification of the two carboxylic acids contributed to the increase in antibiotic activity (Maffioli et al. 2005). In addition, a research had reported that a double mutant lacking dbv8 and dbv23 was constructed to produce A40926 intermediates in order to obtain the single A40926 component by chemical synthesis method in vitro (Alt et al. 2019). Recently, a novel GPA A50926 was discovered from Nonomuraea coxensis DSM45129 and shared a similar structure with A40926, lacking the carboxyl group on the N-acylglucosamine moiety (Yushchuk et al. 2021). The semisynthetic antibiotic dalbavancin as the derivative of A40926 is proven to has a high affinity to human angiotensin-concerting enzyme 2 (ACE2), which is the docking site of the SARS-COV-2 spike protein, indicating that dalbavancin is a promising anti-COVID-19 drug candidate (Wang et al. 2021).

Discovery of novel natural polypeptide

The NRPS-encoded polypeptides, including the known GPAs, have always been an appealing source of novel antibiotics. The genome sequencing technology dramatically improves the discovery of NRPS BGCs encoding potential natural lipopeptides from bacteria, especially actinomycetes. Based on the bioinformatic analysis, a new lipopeptide cilagicin containing ten amino acid residues encoded by the cil cluster has been discovered recently. This novel antibiotic apparently resists antibiotic-resistant pathogens (Wang et al. 2022a). Furthermore, macolacin, an analogue of colistin, is chemically synthesized depending on bioinformatic analysis of identified BGC from sequenced bacterial genomes. The colistin congener is effective against pathogens that express the mcr-1 (phosphoethanolamine transferase) resistance gene (Wang et al. 2022b). As the above strategy, a similar route for discovering and verifying new GPAs is developed and described as the GPA Heterologous expression (GPAHex) platform, which is a synthetic biology tool to exploit the cryptic non-expressed GPA under laboratory conditions. With the aid of GPAHex platform, the novel Type V GPAs corbomycin and GP6738 are biosynthesized successfully and effectively (Xu et al. 2020). Furthermore, a series of new Type V GPAs rimomycin (A/B/C) and misaugamycin (A/B) are discovered by combining phylogeny-guided genome mining and the synthetic biology platform GPAHex. These new Type V GPAs have novel chemical scaffolds that expand this type of GPAs’ diversity. In detail, rimomycin shares a similar chemical structure with kistamicin, but presents an N-methyl-tyrosine group at aa6 residue and two Hpg moieties at aa1 and aa3 residues, and while misaugamycin contains an unprecedented N-C ring between aa2 and aa4 and sole N-terminal acylation (Xu et al. 2022).

Perspectives

Few compound classes play dramatically essential roles in human health protection as antibiotics. As typical representatives, GPAs are a group of complicated molecules used to treat serious Gram-positive bacterial infections, which mainly are obtained by the sole pathway of actinomycetes fermentation directly or indirectly. GPA researches continue to generate new discoveries and new entity molecules for antibiotic development. For instance, the reassembly of the NRPS line is performed to generate novel potential GPAs based on synthetic biology. A recent study presented that a total of thirteen tailoring enzymes from seven GPA BGCs in different combinations were introduced into the model actinomycete S. coelicolor, which had priorly loaded a heterogenous minimal scaffold of teicoplanin, which tried to explore the criteria for GPA modification in vivo and to expand GPA chemical diversity (Yim et al. 2016). For another example, a team from Princeton University developed a universal high-throughput platform to screen the secondary metabolome based on an imaging mass spectrometric technology. They found a new glycopeptide substance with potent inhibitory activity against a virus (Xu et al. 2019).

In a word, we believe that with the development and interdisciplinary integration of synthetic biology, NGS, and AI, more GPAs with new chemical structures and action mechanisms will constantly be emerging.