**1. Introduction**

Fungi produce a plethora of chemically diverse natural products (NPs) with various bioactivities and a wide variety of applications in medicine and agriculture [1]. Because of this, fungal NPs have been historically recognized as an invaluable source of inspiration for the development of drug leads for the treatment of infections and cancer, as well as the prevention of crop damage, such as lovastatin (the cholesterol-lowering drug), mycophenolic acid (the immunosuppressive drug), and pyripyropene A (the insecticide) [2–4]. In addition, fungal NPs with novel molecular scaffolds provide excellent templates for the chemical synthesis of new bioactive compounds [5].

In fungi, the natural products are commonly synthesized by the genes arranged in a contiguous fashion as a biosynthetic gene cluster (BGC) [6]. The core biosynthetic enzymes encoded by the BGCs mainly include polyketide synthases (PKSs), non-ribosomal peptide synthetases (NRPSs), terpene synthases, and ribosomally synthesized and posttranslationally modified peptide (RiPP) biosynthetic enzymes to enlarge the variety of carbon skeletons of the products. Tailoring enzymes such as the monooxygenase, hydroxylase, and methyltransferase further increase the diversity and complexity of the molecules [7]. With the development of genome sequencing technology, bioinformatic methods, genome mining algorithms, and scalable expression platforms, the genomicdriven approaches have revolutionized NPs discovery, greatly expand the access to the

**Citation:** Dai, G.; Shen, Q.; Zhang, Y.; Bian, X. Biosynthesis of Fungal Natural Products Involving Two Separate Pathway Crosstalk. *J. Fungi* **2022**, *8*, 320. https://doi.org/ 10.3390/jof8030320

Academic Editors: Tao Feng and Frank Surup

Received: 27 February 2022 Accepted: 17 March 2022 Published: 21 March 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

chemical repertoire of fungal-derived NPs, and provide unprecedented opportunities to investigate their biosynthetic mechanisms [8]. Several vital factors that regulate the expression of fungal BGCs, including the environmental signals, transcriptional regulation, and epigenetic regulation, have been summarized by Keller [9], which might be instructive for genome mining and activation of the silent BGCs.

In most reported studies, the microbial NPs biosynthetic pathways are generally completed by one single BGC. However, in some interesting but rare cases, the products have been demonstrated to be constructed via the hybridization of two precursors, which are biosynthesized by two separate BGCs. For instance, the antithrombotic myxadazoles, a family of novel chimeric compounds isolated from *Myxococcus* sp. SDU36, consists of Nribityl 5,6-dimethylbenzimidazole and a linear fatty acid chain endowed with an isoxazole ring. Interestingly, a non-canonical PKS/NRPS biosynthetic pathway and the vitamin B12 metabolism pathway were proved to interwind together through the construction of isoxazole-benzimidazole hybrids [10]. Tasikamides, recently isolated from *Streptomyces tasikensis* P46, contain a rare hydrazone group that joins the cyclic peptide scaffold to an alkyl 5-hydroxylanthranilate (AHA) moiety. This research also addressed that the biosynthesis of tasikamides required the coupling of two separate gene clusters, with one BGC encoding a NRPS pathway for assembling the cyclic peptide scaffold, and the other BGC encoding the AHA-synthesizing pathway [11]. These works illustrate that functional crosstalk between two different biosynthetic pathways is not only of considerable value in increasing the structural diversity of NPs, but also a more effective way to construct new drug leads of natural origin.

Several key issues need to be considered when we investigate the biosynthetic mechanism of two separate BGC co-participation in NP biosynthesis: (a) identification of each BGC and characterization of the necessary gene function; (b) isolation of sufficient quantity of key intermediates through gene knockout, enzymatic catalyzation, and heterologous expression; and (c) determination of whether the coupling process of the two separate BGCs is spontaneous or enzymatic. Thus, in-depth understanding of the biosynthetic mechanism for the convergence of two distinct biosynthetic pathways will provide an alternative to accelerate the discovery of NPs with novel skeletons, as well as shed new light on which enzyme(s) could catalyze the formation of C-C, C-N, or N-N bonds that link two biosynthetic precursors together. In this mini-review, we summarized some fungal NPs with novel skeletons produced by the crosstalk between two discrete biosynthetic gene clusters, mainly including polyketides, meroterpenoids, and non-ribosomal peptides. We highlighted their biosynthetic processes, which might provide valuable insights into the coupling mechanism of two separate BGCs in fungal NP biosynthesis.

#### **2. Fungal Polyketide Biosynthesis Involving Two Separate Pathway Crosstalk**

#### *2.1. The Biosynthesis of Penilactones A and B*

The representative examples about non-enzymatic Michael addition mediated the coupling process of polyketide–polyketide hybrids were highly oxygenated fungal polyketides penilactones A (**1**) and B (**2**), which were firstly isolated from an Antarctic deep-sea derived fungus *Penicillium crustosum* PRB-2 by Li and co-workers [12]. The biosynthetic pathway was proposed to be originated from the hybridization of one *o*-quinone methide unit (**5**) and a γ-butyrolactone moiety through 1,4-Michael addition to complete the carbon skeleton construction of **1** and **2**. A biomimetic total synthesis was subsequently achieved to confirm the biosynthetic hypothesis [13]. Considering the enzymes for Michael addition involved in **1** and **2** biosynthesis have not been reported yet, Li and co-workers identified two separate BGCs (termed as *cla* and *tra* BGCs in this review) responsible for the biosynthesis of **1** and **2** through the gene deletion and heterologous expression in *Aspergillus nidulans* [14]. The core non-reducing polyketide synthase (NR-PKS) ClaF in the *cla* BGC is responsible for the biosynthesis of crucial intermediate clavatol (**3**) (Figure 1). The nonheme FeII/2-oxoglutarate-dependent oxygenase ClaD oxidized **3** into hydroxyclavatol (**4**), which spontaneously underwent dehydration into the crucial intermediate *o*-quinone methide (**5**). For *tra* BGC, a hybrid polyketide synthase-nonribosomal peptide synthetase (PKS-NRPS) TraA and a *trans*-acting enoyl reductase (ER) TraG collaboratively catalyzed the formation of crustosic acid (**6**) precursor. The nonheme FeII/2-oxoglutarate-dependent oxygenase TraH subsequently catalyzed the oxidative decarboxylation of **6** into dehydroterrestric acid, the terminal double-bond of which was finally reduced by the flavin-dependent oxidoreductase TraD into an important intermediate terrestric acid (**7**) [15]. The results of precursors feeding experiments in ∆*traA* mutant confirmed that both **6** and **7** were the on-pathway intermediates, and could be transformed into 5-carboxylmethyltetronic acid (**8**) and 5-methyltetronic acid (**9**), respectively. The fascinating issue associated with the biosynthesis of **1** and **2** is which enzyme(s) could catalyze the Michael addition to couple two building blocks together. Surprisingly, incubation of **8** with **4** in water at 25 ◦C led to the formation of penilactone D (**10**) as the major product, and **2** as the minor product (Figure 1). Incubation of **9** with **4** generated peniphenone D (**11**) as the major product as well as **1** as the minor product. Both **10** and **11** were separately incubated with **4**, the formation of **1** and **2** could be observed (Figure 1). These results unambiguously indicated that the Michael addition involving in the biosynthesis of **1** and **2** is nonenzymatic and could happen spontaneously.

**Figure 1.** The biosynthetic pathway of fungal polyketides penilactones A (**1**) and B (**2**). The Michael addition that triggers the coupling of *cla* and *tra* BGCs is nonenzymatic.

In general, the combination of enzymatic and nonenzymatic reactions originated from the crosstalk between two separate biosynthetic pathways significantly enriched the structural diversity of fungal NPs. The biosynthesis of fungal polyketides penilactones A and B provides an excellent example to investigate that how two different BGCs interwind at a gene cluster level.

#### *2.2. The Biosynthesis of Dalmanol A and Acetodalmanol A*

The mantis-associated fungus *Daldinia eschscholzii* TL01 is known to produce novel polyketides including dalmanol A (**12**) and acetodalmanol A (**13**) with immunosuppressive bioactivity [16–18]. The structural characteristic of **12** and **13** implied that their carbon skeletons construction involved in co-participation of two building blocks naphthalene and chromane. Tan and co-workers conducted the pioneering work to identify two separate BGCs (termed as *chr* and *nap* BGCs in this review) by gene deletion as well as heterologous expression [19]. The *chr* BGC is responsible for chromane biosynthesis and the *nap* BGC biosynthesizes naphthalene. This work addressed the two-gene-cluster crosstalk based biosynthetic pathways of dalmanol A and acetodalmanol A (Figure 2).

**Figure 2.** The biosynthetic pathway of fungal polyketides dalmanol A (**12**) and acetodalmanol A (**13**).

In plants and fungi, the assembly of chromane-based aromatic polyketides have been reported to be biosynthesized by both type III PKS and partially reducing type I PKS (PR-PKS) [20,21]. The results of the quantitative reverse transcription PCR analysis, targeted gene deletion, and heterologous expression experiments supported that the PR-PKS ChrA and the keto-reductase ChrB are indeed responsible for the formation of vital intermediate 1-(2,6-dihydroxyphenyl)but-2-en-1-one (PBEO,**14**) (Figure 2). Thus, the genes encoding PR-PKS ChrA, the keto-reductase ChrB, and a transporter constitute the *chr* BGC, which is located on Scaffold\_36 in the genome (Figure 2). The NR-PKS gene *pksTL* in *Daldinia eschscholzii* TL01 has been confirmed to be responsible for the biosynthesis of naphthalene-based polyketide 1,3,6,8-tetrahydroxynaphthalene (4HN, **15**) by gene deletion [22]. Co-expression of the NR-PKS pksTL and 4HN reductase 4HNR in *A*. *oryzae* led to the accumulation of 1,3,6-tetrahydroxynaphthalene (3HN, **16**), a biosynthetic precursor for the assembly of dalesconols, which are polyketides also isolated from *Daldinia eschscholzii* TL01. The genes encoding the NR-PKS pksTL, 4HN reductase 4HNR, two transcription factors, and a laccase constitute the *nap* BGC, which is within Scaffold\_20 in the genome and locates at least 493 kb away from the *chr* BGC. Only when *pksTL*, *chrA*, and *chrB* were co-introduced into *A*. *oryzae*, the polyketides 12 and 13 could be successfully produced. The critical biosynthetic process for the coupling of *chr* and *nap* BGCs was proposed to be the epoxidation of PBEO (14) to the proposed intermediate 17 (Figure 2). The results

of enzymatic activity inhibition experiments revealed an unspecific P450 monooxygenase located elsewhere in the genome of *A*. *oryzae* host might be responsible for the epoxidation of 14 to 17, which triggered the cross-cluster coupling of both *chr* and *nap* BGCs.

Overall, this work further illustrated that two separated BGCs crosstalk is a promising access to improve the structural diversity of fungal NPs. Understanding the regulatory mechanism of the multiple-gene-cluster coupling is of great significance in establishing the synthetic biology approaches to discover NPs with novel skeletons and potential biological activities.

### *2.3. The Biosynthesis of Azasperpyranone A*

Azaphilones, a group of structurally related fungal polyketides, contain a highly oxygenated bicyclic pyrone quinone moiety, and exhibit a broad range of bioactivities including anticancer, antifungal, and antiviral activities [23]. Azasperpyranone A (**18**), recently isolated from *A*. *terreus*, contains a highly oxygenated pyranoquinone moiety possessing a 6/6/6/6 tetracyclic ring system, and shows potential anticancer activity [24]. Scrutiny of the structural feature of azasperpyranone An (**18**) revealed that two building blocks 5-methyl orsellinic aldehyde (**19**) and preasperpyranone (**20**) are the biosynthetic precursors. Lu and co-workers have identified two separate BGCs by gene deletion, including the BGC A responsible for polyhydric phenol formation and the BGC B involving the azaphilonoid scaffold construction (Figure 3) [24].

**Figure 3.** The biosynthetic pathway of fungal polyketide azasperpyranone A (**18**).

In a previous study, the full-length, intron-free open reading frames of two core genes *ATEG*\_*03629* and *ATEG*\_*03630* from BGC A, which encode a NR-PKS and a NRPS-like enzyme, respectively, were co-transformed into *Saccharomyces cerevisiae* to produce the intermediate **19** (Figure 3) [25]. Then the FAD-dependent monooxygenase (FMO) encoded by *ATEG*\_*03635* gene catalyzed the hydroxylation of **19** to afford the intermediate **21**, which was subsequently oxidized into the crucial precursor **22** by the P450 monooxygenase encoded by *ATEG*\_*03631* gene. In BGC B, the two core genes *ATEG*\_*07659* encoding a highly reducing PKS (HR-PKS) and *ATEG*\_*07661* encoding a NR-PKS were heterologously co-expressed in *A*. *nidulans* to produce **23** [26,27]. Compound **23** was rapidly converted into an important precursor preasperpyranone **20** by the FMO encoded by gene *ATEG*\_*07662*. The remaining gap for the entire biosynthetic pathway of azasperpyranone A is that which enzyme could couple the two vital intermediates **20** and **22** together by catalyzing the

formation of C-C and C-O bonds. Deletion of the gene *ATEG\_03636* with unknown function abolished the production of **18**, but accumulated two precursors **19** and **20**, which implied the formation of **18** was more likely catalyzed by ATEG\_03636 rather than caused by the spontaneous reaction between **20** and **22** (Figure 3). The authors also found that the ATEG\_07667 in BGC B could indirectly regulate the cluster-specific regulators ATEG\_03638 in BGC A and ATEG\_07666 in BGC B to collaboratively synthesize the anti-cancer compound **18**. This interesting collaborative model in the fungal NPs biosynthesis provides new clues for the investigation of regulatory mechanism for other novel natural products which were biosynthesized by two separate BGCs crosstalk.
