Unexpected assembly machinery for 4(3H)-quinazolinone scaffold synthesis

Abstract

4(3H)-quinazolinone is the core scaffold in more than 200 natural alkaloids and numerous drugs. Many chemosynthetic methodologies have been developed to generate it; however, investigation of its native enzymatic formation mechanism in fungi has been largely limited to fumiquinazolines, where the two nitrogen atoms come from anthranilate (N-1) and the α-NH₂ of amino acids (N-3). Here, via biochemical investigation of the chrysogine pathway, unexpected assembly machinery for 4(3H)-quinazolinone is unveiled, which involves a fungal two-module nonribosomal peptide synthase ftChyA with an unusual terminal condensation domain catalysing tripeptide formation; reveals that N-3 originates from the inorganic ammonium ions or the amide of L-Gln; demonstrates an unusual α-ketoglutarate-dependent dioxygenase ftChyM catalysis of the C-N bond oxidative cleavage of a tripeptide to form a dipeptide. Our study uncovers a unique release and tailoring mechanism for nonribosomal peptides and an alternative route for the synthesis of 4(3H)-quinazolinone scaffolds.

Fig. 1. Representative 4(3H)-quinazolinone scaffold-containing drugs and their biosynthetic machinery in natural products.

a Three quinazolinone scaffolds and b the best-selling drugs containing a 4(3H)-quinazolinone scaffold. c 4(3H)-quinazolinone scaffold formation is catalysed by the NRPS C_T domain and d α-KGD-mediated rearrangement.

Fig. 2. Gene clusters and previously proposed pathway for the synthesis of 1.

a Three homologue clusters for the synthesis of 1 in different fungi. b In vivo gene overexpression and knockout experiments in P. chrysogenum and F. graminearum suggested that 3 or its derivatives and 2 are the possible precursors of 1.

Fig. 3. Confirmation of the ftchy cluster and heterologously produced products.

a LC-MS analyses of the A. nidulans transformant culture extracts. b Chemical structures of the compounds isolated from the AN-ftchyACDEHM transformant. c LC-MS analyses of the AN-ftchyA transformant culture extracts. d Chemical structure of 10 showing it is a linear γ-l-glutamyl-l-alanyl-anthranilate tripeptide.

Fig. 4. Biochemical confirmation of the unexpected assembly process of ftChyA for the synthesis of 10 and 3.

a Biochemical confirmation of ftChyA synthesizing 10 and 3. b Phylogenetic analysis of fungal NRPS C domains showing that ftChyA-C_T is separated into an independent clade. The C domain protein sequences used for phylogenetic analysis are listed in Source Data file. c ftChyA-C₁ and ftChyA-C_T domain mutations confirm that the C₁ domain is responsible for amide bond formation between l-Glu and l-Ala of 10 and that C_T is responsible for amide bond formation between l-Ala and Ant to release 10 and 3. d ATP-PPi release assay and MALDI-TOF MS analysis of the A domains for substrate recognition. Data are shown as the mean ± SEM of 3 independent experiments. ***P < 0.001, ****P < 0.0001 (ftChyA-A₁, P = 1.5e-6 between l-Glu and Ant, P = 4.5e-6 between l-Glu and l-Ala; ftChyA-A₂, P = 0.0008 between l-Ala and Ant), unpaired two-tailed Student’s t test. e Biochemical assays of ftChyA in H₂¹⁸O-Tris buffer confirming that water is not involved in the formation of 10 and 3. f In vitro assays of ftChyA with l-Glu, l-Ala and Ant-Me. The extracted ion chromatograms (EICs) were extracted at m/z 223 [M + H]⁺ for Ant-Me-3 and m/z 352 [M + H]⁺ for Ant-Me-10.

Fig. 5. Three assembly mechanisms of NRPS ftChyA were proposed for the synthesis of 10 and 3.

a Canonical assembly rule of NRPS. b Pass-back mechanism of NRPS. c Proposed mechanism in this study.

Fig. 6. Biochemical confirmation of the functions of ftChyD, ftChyE and ftChyM.

a ftChyD uses NH₄Cl or l-Gln to catalyse the amidation of 10 to form 11. b LC-MS analysis of the incorporation of ¹⁵N into 11. The EICs were extracted at m/z 338 [M + H]⁺ for 10 and ¹⁵N-labelled 11, and m/z 337 [M + H]⁺ for 11. c ftChyE catalyses the hydrolysis of 11 and 10 to form 2 and 3, respectively. d ftChyM is an α-KGD that catalyses the C-N bond oxidative cleavage of 11 to form 4 via a possible intermediate 12. e Conversion of 4 to 6 shows the alkaline-induced spontaneous C-2-N-3 bond closure.

Fig. 7. The proposed complex pathways for generating the 4(3H)-quinazolinone scaffold in 1 synthesis.

The primary pathway shows unexpected assembly machinery starting from a linear tripeptide and the salvage pathway depends on the promiscuous substrate selectivity of ftchy cluster enzymes.