Metabolic Engineering of Non-Ribosomal Peptides

The diverse structures and promising pharmacological activities of non-ribosomal peptides (NRPs) have attracted widespread research interest in developing novel NRPs and their derivatives. Complex and diverse fungal NRPs are one of the most important sources for innovative drug research. For instance, β-lactam antibiotics have been used as frontline anti-infective drugs for nearly a century, while CsA, beauvericin (BEA), enniatin (ENN), cyclosporins, and ergot alkaloids play important roles in treating fungal or bacterial infections, organ transplantation, tumors, and neurological diseases.

Discovering new NRP through fermentation and separation technology is one of the means to obtain new NRP drugs. In contrast, the use of chemical modifications to alter active NRP scaffolds suffers from challenges such as the lack of non-proteinogenic amino acids, low peptide cyclization efficiency, poor functional group compatibility, lengthy synthesis times, and difficulty controlling structure-activity relationships. Therefore, it is crucial to develop efficient biosynthetic NRP derivatives.

Among the effective methods for obtaining NRPs, biological approaches include novel NRP mining techniques guided by genomics (such as overexpression of transcriptional regulators and heterologous expression of gene clusters), as well as drug molecule remodeling based on biosynthetic theories, precursor-directed biosynthesis, mutational biosynthesis, and enzymatic methods.

Heterologous Expression

For fungal NRPS gene clusters silenced under conventional laboratory culture conditions, overexpression can be achieved if they contain global regulators (hdaA, VeA, or LaeA), thus activating the gene clusters to obtain new peptide products. For example, knocking out the gene hdaA encoding histone deacetylase (HDAC) in Calcarisporium arbuscular successfully activated 42 gene clusters, resulting in the production of two new cyclic peptides, arbumycin and arbumelin. Additionally, utilizing heterologous expression systems for fungal NRPs production is an important approach for exploring novel NRPs or enhancing yields. This includes prokaryotic expression systems (Escherichia coli), yeast expression systems (Saccharomyces cerevisiae), and filamentous fungal expression systems (A. nidulans, A. oryzae, A. niger, and Neurospora crassa). Heterologous expression of NRPS-like (PgnA) from Aspergillus terreus in A. nidulans produced the dipeptide product phenguignardic acid. In A. niger, heterologous expression of Esyn1 from Fusarium oxysporum successfully increased the yield of enniatin to 5 g/L.

Precursor-Directed Biosynthesis

Precursor-directed biosynthesis involves the addition of chemically synthesized precursor analogs to the production strains, thus directing the synthesis of peptide analogs. In 1989, feeding the non-proteinogenic amino acids DL-α-allylglycine, L-β-cyclohexylalanine, and D-Ser to T. inflatum successfully synthesized three CsA analogs, namely [Allygly2]CsA, [MeCyclohexylala1]CsA, and [D-Ser8]CsA, with [D-Ser8]CsA exhibiting strong immunosuppressive activity. Similarly, feeding L-Leu to V. hemipterigenum accumulated the products enniatin H and enniatin I, while feeding L-Ile produced MK1688. Furthermore, feeding 30 D-2-Hiv and L-Phe analogs to B. bassiana yielded six novel BEA analogs, including beauvericin H1-3 and beauvericin G1-3, with beauvericin G1-3 showing reduced cell migration inhibition and cytotoxicity, while beauvericin H1-3 enhanced cytotoxicity.

Mutational Biosynthesis

Precursor-directed biosynthesis faces challenges such as competition between fed precursors and natural substrates, resulting in low target product yields and difficulties in isolation, making large-scale production challenging. However, mutational synthesis technology disrupts the production of natural substrates and is one of the most successful methods in combinatorial biosynthesis, which has been applied to NRP biosynthesis. In vivo knockout of the gene kivR in B. bassiana eliminated the production of the natural precursor D-Hiv, and feeding four non-homologous hydroxycarboxylic acids (D-Hbu, D-Hmv, L-2-F-Phe, and L-3-F-Phe) yielded 14 non-natural BEA analogs, some of which exhibited potent antiproliferative activity. Additionally, feeding different chain-length fatty acid substrates, including myristoyl, pentadecanoyl, and palmitoyl, to the mutant strain GlareaG lozoyensis-△GLPKS4 successfully synthesized four pneumocandin analogs of different chain lengths, pneumocandins H-K, with pneumocandin I showing similar hemolytic activity to pneumocandin B but higher antifungal activity. Introducing three genes (papA-C) catalyzing the synthesis of p-aminophenylpyruvate from Streptomyces venezuelae into the branched-chain deficient fungus Rosellinia sp. resulted in the synthesis of four PF1002 analogs, including PF1022-220, PF1022-260, PF1022-268, and PF1022-269, via conversion of p-aminophenylpyruvate to p-amino-phenyllactate and p-nitrophenyllactate.

Figure 1 Metabolic Engineering of NRPs.

Enzyme Engineering

Based on the functional domains and assembly mechanisms of NRPSs, protein structure remodeling approaches include (1) replacement of A domains or A-T domains; (2) directed mutagenesis of amino acid residues in the substrate-binding pockets of A domains; (3) replacement, deletion, and insertion of C-A, C-A-T, or T-C-A domains. Through enzyme engineering, the synthesis enzymes PSYN (Rosellinia abscondita) for PF1022 were modified by swapping module 1, recognizing β-OH-Ala and β-OH-Phe, with module 1 from Esyn1 and BbBEAS, recognizing D-Hiv, to generate artificial synthesis enzymes hPESYN and hPBEAS, thereby synthesizing four ENN and two BEA analogs, including [PheLac]-ENN and [PheLac]-BEAS, etc. The C-terminal condensation domain CT catalyzes the cyclization of peptides with different chain lengths, thus swapping CT domains of different NRPSs can achieve the synthesis of NRPs with different numbers of monomers. Substituting the CT of Bassianolide synthase (BaSYN) with that of BbBEAS-CT formed BaSYN-BaCT synthase and BaSYN-BaT2bCT synthase, both catalyzing the synthesis of Octa-beauvericin (Hiv-Phe×4). Furthermore, replacing the entire BaSYN with the CT of BbBEAS yielded BeBaSYN2, which also produced Octa-beauvericin. Replacing the CT of Esyn1 with BaSYN produced EnSYN-BaCT and EnSYN-BaT2bCT, catalyzing the formation of Octa-enniatin B (Hiv-Val×4).

Recombinant and engineered NRPS biosynthetic pathways can effectively synthesize complex derivatives of NRPs. Eleven new BEA analogs were synthesized using either in vitro enzymatic synthesis or whole-cell biocatalysis with E. coli expressing bbBeas. Utilizing the substrate promiscuity of the NRPS-TE domain (TycC-TE) of tyrocidine, over 300 linear tyrocidine analogs were successfully synthesized. In vitro recombination of PKS-NRPS from Aspergillus terreus separately with different amino acids and free thiols as substrates synthesized over 60 thiopyrazine compounds. The KR domain in NRPS modules catalyzes the formation of cryptophycin analogs through chemical synthesis of SNAC precursors with α-ketoisovaleric acid. The NRPS-like enzyme IvoA (A-T-E-R) catalyzes the stereoisomerization of various substituted tryptophan analogs on the indole ring to produce D-configured products.

Summary and Prospects

Currently, elucidating complete biosynthetic pathways of numerous fungal NRP drugs has been achieved through genomic mining, targeted gene knockout, isotope labeling, in vivo chemical feeding, heterologous host expression, and in vitro enzymatic characterization. With continuous research on fungal NRP biosynthesis, many novel biocatalysts with catalytic activities have been revealed from natural sources, such as the heme-independent iron oxygenase PcbC, the α-KG-dependent oxidase CefEF, the sulfotransferase McfS, and the oxidoreductase EasG. The discovery of novel biocatalysts provides more enzyme selectivity for the enzymatic synthesis of novel NRP derivatives, facilitating the development of new bioactive drugs or drug lead compounds with improved efficacy. Additionally, the discovery of new functional biocatalysts helps address reactions that cannot be completed through chemical synthesis. Through directed evolution of biocatalysts, overcoming rate-limiting enzyme restrictions, and combining combinatorial biosynthesis remodeling pathways, green synthesis and efficient creation of NRP derivatives can be achieved.

A large number of potential new peptide compound resources remain untapped. In the future, bioinformatics analysis tools can be used to predict products of NRPs gene clusters, enabling the targeted development of NRPs with specific structures or activities. Synthetic biology technologies such as overexpression of transcriptional regulators, promoter engineering, and heterologous system expression can activate silent gene clusters, thus exploring and developing novel natural NRPs.