University of Groningen Exploring the metabolic potential of Penicillium rubens Viggiano, Annarita

(1)

University of Groningen

Exploring the metabolic potential of Penicillium rubens

Viggiano, Annarita

DOI:

10.33612/diss.126598491

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2020

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Viggiano, A. (2020). Exploring the metabolic potential of Penicillium rubens. University of Groningen. https://doi.org/10.33612/diss.126598491

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

(2)

Exploring the metabolic potential

of Penicillium rubens

(3)

The work described in this thesis was carried out in the Molecular Microbiology Group of the Groningen Biomolecular Sciences and Biotechnology Institute (GBB) of the University of Groningen, the Netherlands. The research leading to these results has received fund-ing from the People Programme (Marie Curie Actions) of the European Union’s Seventh Framework Programme FP7/2007-2013/under REA agreement no. [607332].

Cover and layout: Lovebird design.

www.lovebird-design.com

Printed by: Eikon+ Book

ISBN: 978-94-034-2709-6 Ebook

ISBN: 978-94-034-2710-2

Copyright @A.Viggiano, Groningen, the Netherlands, 2020

(4)

Exploring the metabolic potential of

Penicillium rubens

PhD thesis

to obtain the degree of PhD at the University of Groningen

on the authority of the

Rector Magnificus Prof. C. Wijmenga and in accordance with the decision by the College of Deans. This thesis will be defended in public on

Friday 12 June 2020 at 11.00 hours

by

Annarita Viggiano born on 5 May 1989

(5)

Supervisors

Prof. A.J.M. Driessen Prof. R.A.L. Bovenberg

Assessment Committee

Prof. G.J. Poelarends Prof. I.J. van der Klei Prof. V. Meyer

(6)

CHAPTER 1 Introduction

Annarita Viggiano1_{, Roel A.L. Bovenberg}2,3_{, Arnold J.M. Driessen}1

1_{Molecular Microbiology, Groningen Biomolecular Sciences and Biotechnology Institute,}

University of Groningen, Groningen, The Netherlands

2_{DSM Biotechnology Centre, Delft, The Netherlands}

3_{Synthetic Biology and Cell Engineering, Groningen Biomolecular Sciences and}

Biotechnology Institute, University of Groningen, Groningen, The Netherlands

(9)

(10)

1

1. Secondary metabolites

Plants, fungi and bacteria produce a wide range of diverse molecules, named secondary metabolites. While primary metabolites are essential for growth and survival, secondary metabolites confer the ability to adapt in a specific environment and show a variety of biological func-tions (1, 2). Pigments protect the organism from harsh environment conditions such as UV radiation (3); siderophores transport and store iron compounds, but they can also be related to virulence (4); other molecules are used for communication or competition, being toxins or having antimicrobial and antifungal activity (5). Due to their properties, some metabolites can be harmful for men. Many toxins are produced by food-associated molds and therefore represent a risk for men and animals. In the Middle Ages, the contamination of rye and other cereals by the fungus Claviceps, which produces ergot alkaloids, caused the epidemic fungal disease ergotism (6). The aflatoxins are powerful toxins produced by Aspergillus flavus. They were discovered in 1960s when thousands of turkeys died in London and surroundings after eating the same peanut meal, which was contaminated by fungi. Later studies led to the identification of the aflatoxins B1, B2, G1 and G2, which are carcinogenic in animals and men (6).

On the other hand, many secondary metabolites are of great interest for applications in medicine, agriculture, biotechnology. Penicillin is the best example of the great impact that these compounds can have on our health, as it significantly improved the quality and expectancy of life (7). Since its discovery, many other antimicrobial compounds have been characterized, such as cephalosporin and streptomycin (8). Be-sides antibiotics, secondary metabolites found also other applications in medicine, for example as cholesterol lowering agents (lovastatin) (9), immunosuppressants (cyclosporin) (10), anti-osteoporosis (orsell-inic acids and derivatives) (11) or anticancer molecules (paclitaxel) (12). Moreover, several secondary metabolites, mainly produced by plants, can be used in cosmetics (camphor, kojic acid) (13), as food pigments (curcumin, anthocyanin) (14), as natural insecticides (pyrethrum), insect repellents (citronellal) and antifungals (cinnamaldehyde) (15).

Today over 500.000 secondary metabolites have been described and about 12% show biological activities (2, 16). Among them, many compounds became drugs or provided the core chemical structure for

(11)

1: I

ntr

oduc

tion

semi-synthetic or synthetic drugs (16, 17). However, metagenomics and genomics analyses suggest that only a small portion of these molecules has been identified so far (18–20). Considering the large number of species not yet characterized and that many secondary metabolite biosynthetic genes are not expressed under laboratory conditions, it is clear that there is still a lot to discover. The potential of secondary metabolites for human applications appears to be huge and can be untapped by a deeper understanding of the biosynthetic mechanisms. Here, we will focus on the secondary metabolites produced by filamen-tous fungi, describing the incredible chemical variety of these mole-cules and presenting different approaches for genetic and metabolic engineering to allow the identification of novel compounds as well as to increase the structural diversity or improve production. In the end, we will focus on Penicillium rubens, previously named P. chrysogenum, as a platform for the characterization of secondary metabolites and production of heterologous compounds.

1.1. Fungal secondary metabolites

Fungi play a crucial role in the environment. Together with other or-ganisms as invertebrates and bacteria, fungi are responsible for the decomposition processes, degrading organic material and therefore recycling nutrients in the ecosystem (21); they can be part of the food chain, pathogens or they can establish symbiotic relations with other organisms, such as plants (mycorrhyzae) or algae/cyanobacteria (lichens) (22). Moreover, since antiquity, yeast and many filamentous fungi have been exploited for their fermentation properties, making cheese, brew-ing or bakbrew-ing, and for the production of pigments, which were used as food colorants (23).

In their 900 millions years of history (24), fungi have colonized almost every environment on earth, from marine and freshwaters to the des-erts, from indoor places to the poles (25). The success in adapting to such a variety of conditions might be also attributed to the production of a broad range of secondary metabolites (2). The fungal kingdom makes about 47% of all microbial bioactive secondary metabolites currently known, with the filamentous fungi being the main produc-ers (2). Filamentous fungi belong to the division Ascomycota. Among them, the genus Penicillium and Aspergillus (Class Eurotiomycetes, Or-der Eurotiales, Family Trichocomaceae) and TrichoOr-derma and Fusarium

(12)

1

(Class Sordariomycetes, Order Hypocreales, Family Hypocreaceae and Nectriaceae respectively) are abundant producers (26).

Many secondary metabolites have interesting properties for human applications. Besides the aforementioned examples, fungal metabolites include the antibiotic pleuromutillin, the antifungal griseofulvin (2) and several pigments attractive for the textile or food industry (27, 28). For an overview of the main bioactive fungal metabolites which found ap-plications as drugs, agrochemicals and cosmetics see Table 1. Although extremely diverse for their biological activities and structures, the most abundant fungal secondary metabolites belong to few chemical classes: non ribosomal peptides (NRPs), polyketides (PKs), polyketide-amino acid hybrid molecules, terpenes and alkaloids (6). The scaffold molecules are synthesized by complex and large enzymes and are often modified by tailoring enzymes (6). The biosynthetic genes tend to be grouped in biosynthetic gene clusters (BGC), which can span some thousands bp (29, 30) and are often located in the sub-telomeric regions of the chro-mosomes (31, 32). These gene clusters are often co-regulated and include a transcriptional factor. It has also been shown that there is occasional cross-talk between secondary metabolite gene clusters. For example, the regulator scpR controls the expression of two neighboring NRPSs genes and a PKS cluster located on another chromosome in

A. nidulans (33). Moreover, epigenetic regulation plays an important

role in the activation or deactivation of secondary metabolite genes, as first demonstrated by Shwab et al., who showed that the deletion of the histone deacetylase hdaA in A. nidulans resulted in an increased production of penicillin and sterigmatocystin (34). In P. rubens, the de-letion of hdaA has a broad effect on secondary metabolism: while the expression of the sorbicillinoids BGC is enhanced, the chrysogine BGC and the PKS responsible for the production of dihydroxynaphtalene (DHN)-melanin are significantly downregulated (35).

With the increased number of genomes sequenced, several bioinfor-matics tools have been developed to identify secondary metabolite gene clusters, like the “Secondary Metabolite Unique Regions Finder – SMURF” (36) and the “Antibiotics and Secondary Metabolite Analysis Shell – AntiSMASH”. AntiSMASH provides also the domain annotation of the core enzymes in the cluster and a recent version of this software allows a more precise prediction of gene cluster boundaries as well as compounds produced and substrates specificity (37). The information

(13)

1: I

ntr

oduc

tion

Table 1 Main fungal secondary metabolites which found applications in health, agriculture cosmetic. Adapted from Bills and Gloer, 2016 (2).

Metabolites Producing

organisms Classification Indications

Penicillins G and V Penicillium rubens NRP Treatment of bacterial

infections

Cephalosporin C Acremonium

chrysogenum NRP Treatment of bacterial infections

Pleuromutilin Clitopilus spp Diterpene Topical treatment of

infections from Gram-pos-itive bacteria, treatment of Gram-positive bacterial infections in livestock

Fusidic acid Acremonium

fusidioides Triterpene Topical antibiotic for Gram- positive bacterial infections including methicillin-resis-tant Staphylococcus aureus; under development for prosthetic joint infections Strobilurins

A–D and other strobilurins and oudemansins

Strobilurus

tenacellus PK Fungicides for broad range of plant

Griseofulvin Penicillium

griseo-fulvum and other Penicillium spp

PK Treatment of fungal

infec-tions of the skin, hair and nails

Pneumocandin B0 Glarea lozoyensis NRP acylated

to PK Treatment of systemic fungal infections

FR901379 Coleophoma

cylindrospora NRP acylated to fatty acid Treatment of systemic fungal infections

Echinocandin B Aspergillus

pachycristatus NRP acylated to fatty acid Treatment of systemic fungal infections

Enfumafungin Hormonema

carpetanum Triterpene glycoside Treatment of systemic fungal infections Lovastatin

(Mona-colin K) Aspergillus terreus, Monascus

purpureus

PK Treatment of

hypercholes-terolemia to reduce risk of cardiovascular disease Compactin

(Mevastatin) Penicillum citri-num, Penicillum

solitum and other Penicillium spp

PK Treatment of

hypercholes-terolemia to reduce risk of cardiovascular disease

Cyclosporin A Tolypocladium

inflatum NRP Prevention of organ transplant and tissue graft rejection

Mycophenolic acid Penicillium brev-icompactum and other Penicillium spp

Meroterpenoid Prevention of organ rejection following kidney, liver or heart transplant

(14)

1

which is nowadays available about secondary metabolite gene clusters is collected in the database Minimum Information about a Biosynthetic Gene cluster (MIBiG) (38, 39).

1.2 Chemical diversity of fungal secondary metabolites

The advances in the discovery and characterization of secondary metab-olites and BGCs show that filamentous fungi are an extraordinary source

Metabolites Producing

organisms Classification Indications

Myriocin (ISP-I) Isaria sinclairii Amino acid

lipid Treatment of multiple sclerosis

Ergotamine Claviceps

pur-purea, Claviceps fusiformis and Claviceps paspali

Prenylated

NRP Vasoconstrictor used as antimigraine agent, also

combined with belladonna and phenobarbital for relief from menopausal hot flashes Ergometrine

(ergonovine) Claviceps pur-purea, Claviceps

fusiformis and Claviceps paspali

Prenylated

NRP Treatment of postpartum haemorrhage

Ergocryptine Claviceps

pur-purea, Claviceps fusiformis and Claviceps paspali

Prenylated

NRP Treatment of reproductive disorders as amenorrhoea

Mizoribine Penicillium

brefeldianum Imidazole nucleoside Immunosuppressant used for renal transplant in Korea, Japan, China

Kojic acid Aspergillus oryzae,

Aspergillus tamarii, Aspergillus flavus Pyrone derived from glucose Antioxidant in cosmetic products

PF1022A Rosellinia spp NRP Combined with

praziquan-tel, another anthelmintic drug, to treat roundworm, hookworms, and tapeworms in cats

Fumagillin Aspergillus

fumigatus Meroterpenoid Control of nosema disease in honey bees

Gibberellic acid Fusarium fujikuroi Diterpene Plant growth hormone for

plant tissue culture and applications for certain high value crops

α-Zearalanol

(Zeranol) Fusarium spp PK An anabolic agent used to increase the rate of weight

gain and feed conversion in cattle in North America

(15)

1: I

ntr

oduc

tion

of bioactive molecules. The compounds produced have an incredible variety of chemical structures, derived from the wide range of build-ing blocks used for their synthesis as well as the number of chemical modifications performed by the biosynthetic enzymes, which include oxidations, isomerizations and alkylations. Although the filamentous bacteria of the phylum Actinobacteria are the richest source of microbial secondary metabolites, an interesting paper from 2012 reports about 30.000 fungal natural compounds, of which half are bioactive. These molecules are mainly produced by Ascomycetes, but a significant part is made by Basidiomycetes (16). Figure 1 shows some structures of fungal secondary metabolites, which are mainly NRPs, PKs, PKS-NRPS hybrid molecules and terpenes.

1.2.1 Non ribosomal peptides (NRPs)

NRPs are formed in a ribosome-independent way by large enzymes named non ribosomal peptide synthetases (NRPSs). Besides the 20 pro-teinogenic amino acids, NRPSs can incorporate modified amino acids such as hydroxy-, methyl-, β-amino acids (40), as well as fatty acids and α-hydroxy acids (41), therefore generating complex and diverse compounds. Β-hydroxylation is a frequent modification in amino acids and offers an anchor for further modifications like glycosylations (40). For example, the lipoglycopeptide antibiotic ramoplanin A2 contains a β-OH-Asn residue, besides other modified amino acids like the aryl-glycines (42).

NRPSs have a modular structure and each module is involved in the activation of a specific building block and its incorporation into the final molecule (43) (Fig.2A). These functions are carried on by different domains. The adenylation domain (A) recognizes the specific substrate and activates it, by forming an aminoacyladenylate which is later trans-ferred to the thiolation (T) or peptidyl carrier protein (PCP) domain of the same module (44). The PCP domain contains a conserved serine residue which is post-translationally modified by a phosphopanteth-eine transferase (PPTase). The phosphopantethphosphopanteth-eine moiety covalently binds the activated substrate and transfers it to the following domain. The condensation (C) domain is responsible for the formation of the

peptide bond between two building blocks. The release of the product is catalyzed by the thioesterase (TE) domain. Additional domains can be present and contribute to generate a large variety of NRPs. Among

(16)

1

N S O OH H HN O O O O NH+ O -O HO O NH O NH+ O -N H H N O O + H N O -O OH O + H N O -N H HN HN H N NH NH O O O O O O NH+ O -O NH+ O -O O OH HO N H N O OH O O HO OH OH OH N S O OH O N H O O HO NH2 O O H HO O O _O HO H O HO O O H O N N N N N H N N N H N N H H N O O O O O O O O O O O OH HO H H O OH O O HO HN H N O NH O N O N H_HO O O HO O OH O O O O O O H H N O OH OH OH O O OH O O O OH N H O HO OH O N N S S O O OH OH O O O O O O O O O N O N O N O O O O O O coprogen ferrichrome sorbicillin chrysogine YWA1 cephalosporin C pleuromutilin penicillin G lovastatin cyclosporin A gibberellic acid ergotamine kojic acid aflatoxin B1 tenellin mycophenolic acid aspyridone A gliotoxin beauvericin verrucarin J

(17)

1: I

ntr

oduc

tion

them, the epimerization (E) domain changes the stereochemistry of the compound bound to the PCP domain, while the methyltransfer-ase (MT) domain methylates the substrate. The MT domain, which is incorporated in the A domain, is mostly responsible for N-methylation, although also O-, S- and C-methylations are possible (40).

Besides the variety of substrates and domains of the NRPSs, the di-versity of the non ribosomal peptides also depends on the number of modules of the NRPS and their mechanism, which define the length and

SAT KS AT ACP

Starter unit Termination

B

Module n (iterative)

PCP C A PCP other

Starter unit Module n (iterative) A Te Polyketide synthase MT E PPT +ATP A AT ACP S O NH2 R1 S O NH R2 R1 O NH2 S O O OH R R R R Termination Te Te

Non ribosomal peptide synthetase

+Amino Acid MT E KR DH ER KS AT DH MT KR ACP C A PCP Te PKS-NRPS hybrid C Termination PKS NRPS S O ER O O OH HN O R O O OH S S +Malonyl CoA +SAM +ATP +Amino Acid PPT

Fig.2 Structure of NRPSs (A), PKSs (B) and PKS-NRPS hybrids (C). Abbreviations of domains: A, Adenylation; PCP, Peptidyl carrier protein; C, Condensation; MT, Methyltransferase; E, Epimerase; PPT, 4'-Phosphopantetheine transferase; Te, Thioesterase; ACP, Acyl carrier pro-tein; AT, Acyltransferase domain; KS, Ketosynthase; KR, Ketoreductase; DH, Dehydratase domain; ER, Enoyl reductase. SAM is S-adenosylmethionine. Adapted from Brakhage, 2012, Boettger and Hertweck, 2013, (30, 134, 135).

(18)

1

complexity of the molecules. In Type A-Linear NRPSs, each module is used once and incorporates one single substrate, while Type B-Iterative NRPSs reuse the modules multiple times. Enniatins, found in Fusarium species, are an example of NRPs produced by this mechanism. They are a large class of cyclic compounds with antimicrobial and anticancer activities and are derived from 3 iterations of a D-hydroxycarboxylic acid and an amino acid (45).

At last, the NRPs released by the NRPSs can be modified by enzymes encoded by genes of the same biosynthetic cluster, like oxygenases, oxidoreductases, dehydrogenases and transferases (46). In this way, starting from a single scaffold molecule, a large number and variety of compounds can be formed.

1.2.2 Polyketides (PKs)

PKs are the most abundant group of secondary metabolites (6) They are synthesized by the polyketide synthases (PKSs), which use short-chain organic acids like acetyl coenzyme A (CoA) and malonyl CoA (47). Like the NRPSs, the PKSs are organized in modules, that are further divided in domains (Fig.2B). The acyltransferase (AT) domain selects the substrate; the acyl carrier protein (ACP) domain, activated by the transferring of a phosphopantetheine moiety on a conserved serine residue, binds the molecule and shuttles it to the downstream domains. The ketosynthase (KS) domain is responsible for the polyketide chain

elongation through a decarboxylative Claisen condensation, resulting in a β-ketothioester, and for its translocation to the following module (47). Optional domains can further modify the scaffold molecule. The ke-toreductase (KR) domain reduces the β-ketothioester, the dehydratase (DH) domain dehydrates it, the MT domain methylates the thioesther, the enoyl reductase (ER) domain reduces the substrate, the TE domain catalyzes the release of the product as a linear polyketide or a macro-lactone, formed through intramolecular cyclizations (47).

Although their building blocks are not as various as the ones used in NRPs biosynthesis, the PKSs are able to generate a considerable diversity of compounds through the modifications introduced by the optional domains and through the iterative mechanism. Moreover, like the NRPs, also the PKs formed by the PKSs can be modified by other enzymes. In A. parasiticus, the biosynthetic gene cluster for the myco-toxin aflamyco-toxin consists of a PKS gene and 14 flanking genes (including

(19)

1: I

ntr

oduc

tion

reductases, monooxygenases, O-methyltransferases), besides genes involved in fatty acids biosynthesis and sugar utilization (48).

Sometimes the PKS gene clusters are involved in the biosynthesis of different bioactive compounds. In P. marneffei, the PKS involved in the production of the red pigments monascorubrin and rubropunctatin is also responsible for citrinin biosynthesis. In particular, the mycotoxin citrinin originates from acetyl CoA and malonyl CoA through an iterative mechanism, but the same PKS can also form the so called compound 1. This molecule reacts with 3-oxo-octanoic acid and 3-oxo-decanoic acid

synthesized by other enzymes of the pathway, resulting in the formation of a mixture of red pigments (49).

1.2.3 Polyketide-amino acid hybrid molecules

A chemically diverse and interesting class of metabolites is represented by the polyketide-amino acid hybrid compounds, which are synthe-sized by PKS-NRPS hybrids. The PKS portion of these large enzymes functions with an iterative mechanism and has the same domain orga-nization of the so called highly reducing PKSs (HR-PKSs) (50) (Fig.2C), having KS, AT, DH, MT, KR and ACP domains. The ER domain, optional in the HR-PKSs, is not active in the PKS-NRPS hybrids. Its function is not necessary for the formation of the compound or it can be performed by ER proteins encoded from the same biosynthetic cluster or elsewhere (51). The polyketide chain formed in the PKS is then transferred to the NRPS part, which condenses it with an amino acid and releases it (51). The first fungal PKS-NRPS hybrid was discovered in 2004 in Fusarium

monoliforme and F. venenatum, where it is responsible for the

biosyn-thesis of the mycotoxin fusarin C (52). Since then, many other hybrid molecules have been described, as the toxin aspyridone (53) and the immunosuppressant agent pseurotin A (54).

1.2.4 Terpenes

Terpenes derive from dimethylallyl diphosphate (DMAPP) and isopente-nyl diphosphate (IPP) molecules, produced in the mevalonate pathway. DMAPP and IPP are combined to form linear chains of different lengths, which are the direct precursors of terpenes. The terpene synthases (TSs) are responsible for the reactions of dephosphorylation and cyclization of the linear precursors. Like for other secondary metabolites, the scaffold molecule can be also modified by other enzymes (55). Moreover, hybrid

(20)

1

terpenes can be found. The meroterpenes are a group of compounds derived from terpene and polyketide precursors, like austinol and de-hydroaustinol, which are the product of two different gene clusters in

A. nidulans (56).

Depending on the lengths of the precursor molecules, different classes of terpenes are generated: sesquiterpenes (C15), diterpenes (C20) and triterpenes (C30). Sesquiterpenes include mycotoxins like the trichoth-ecenes or the PR toxin, produced by Fusarium species and P. roqueforti respectively (57), but also compounds that are potentially useful for human applications. For example, the terrecyclic acid from A. terreus shows antitumor activities (58). Examples of diterpenes are the gibber-ellins, plant hormones produced by several Fusarium species (59). Among the triterpenes, the helvolic acid shows antimicrobial activity (58). In A.

fumigatus, the biosynthetic cluster consists of the terpene synthase, a

dehydrogenase/reductase, a cytochrome P450 monooxygenase and acyltransferases (60).

2. Genetic and metabolic engineering of secondary

metabolism for the discovery, characterization and

production of bioactive compounds

Fungi represent an incredible reservoir of secondary metabolites, many of which have interesting properties for human applications. Although natural products have dramatically contributed to improve the quality of our lives, novel molecules are needed to face the challenges of now-adays’ society. These include the increasing antimicrobial resistance, the fight against cancer and aging associated-diseases, as well as the emergence of plant pathogens which threaten agriculture.

Of the millions of fungal species estimated to exist, only 100.000 have been identified (22). The undiscovered fungi could be an immense source for bioactive molecules. In this context, fungi from extreme environments are potentially great candidates for the discovery of novel metabolites. Since they adapted to harsh conditions such as high salinity, temperature or pressure, extreme pH or radiations, they are expected to produce unique and unusual molecules. Some examples from the past years include new cytotoxic compounds from an endophytic fungus of the Sonoran Desert (61), a Penicillium species from the deep sea (62)

(21)

1: I ntr oduc tion Native prom oter Inducible/con stitutive promoter

Heterochromatin Silenced BGC Euchromatin Activated BGC

Promoter replacement Overexpression of regulators Chromatin remodeling Global regula tor Specific reg ulator Gene deletion Protein engineering - O O OH - O O OH - O O OH Feeding or mutasynthesis

Heterologous gene expression

Discovery and production of novel secondary metabolites

Fig.3 G ene tic and me tabolic engineering appr oaches for the disc ov ery and pr oduction of no vel sec ondary me tabolit es. A dap ted fr om B rakhag e and S chr oeckh, 2011, B rakhag e, 2012, (19 , 30 , 135).

(22)

1

and an halotolerant Aspergillus (63). Analyzing the genome sequences of three fungi from extreme environments, Chávez et al. found unique secondary metabolite gene clusters, having no homology with the ge-nomes in the AntiSMASH database (64). These could specify for novel chemical structures.

Besides classical screening methods, new metabolites can be dis-covered or produced by genetic and metabolic engineering (Fig.3). The increasing availability of genomic data and advances in

bioinfor-matics analysis have shown that a considerable part of the secondary metabolites clusters is low expressed or silent under laboratory con-ditions (18–20). This trend applies to all fungi, even more in the case of organisms living in harsh environments, since very specific stimuli might be needed for gene activation. Below, we will present the main strategies to awaken these silent genes and discuss the importance of pathway elucidation and engineering of the core biosynthetic enzymes for structural diversification.

2.1 Main strategies to awaken the secondary metabolite gene clusters

The regulation of secondary metabolite gene clusters is complex and dependent on specific stimuli and conditions such as pH, nitrogen or carbon sources and signals from other microorganisms. Moreover, there are different levels of regulations, mediated by gene cluster specific factors, as well as global and epigenetics regulators (30). For these reasons, many secondary metabolite gene clusters are not expressed under standard laboratory conditions. These are so-called “sleeping”, silent or cryptic gene clusters. Awakening the silent genes by promoter replacement, manipulation of regulators and chromatin remodeling or finding environmental conditions that result in gene expression has led to the discovery of several novel compounds.

Promoter replacement. The exchange of the native promoter with a well characterized promoter is a powerful strategy, especially when applied for overexpressing regulators that can control multiple genes. After the genome of A. nidulans was sequenced in 2006, Bergmann et al. noticed the presence of a silent hybrid PKS-NRPS gene, which was putatively part of a cluster containing the transcription factor apdR. In order to investigate the function of these genes, the regulator apdR was overexpressed by using the strong inducible alcohol dehydrogenase alcA

(23)

1: I

ntr

oduc

tion

promoter. This approach resulted in the activation of the cluster and the production of novel compounds, the aspyridones (53). Few years afterwards, investigating the eight NR-PKSs of A. nidulans which were not yet characterized, Ahuja et al. replaced their native promoters with

alcA, identifying their products and seven novel metabolites (65). As only

the promoter of the PKSs was exchanged and the other genes of the cluster were not overexpressed, the final compounds of the pathways were most likely not formed. However, this strategy allows to identify the function of the core enzyme in the cluster and could represent the first step in elucidating the biosynthetic pathways.

Manipulation of specific or global regulators. Overexpressing the regulator scpR of a silent gene cluster, it has been possible to express the two neighboring NRPSs genes in A. nidulans. Interestingly, a PKS gene cluster located on another chromosome was also upregulated, resulting in the production of a novel metabolite, the asperfuranone. This cluster contained itself a transcription factor, which was proposed

to be upregulated by scpR (33). Interfering with global regulators also can lead to awakening silent or low-expressed biosynthetic genes, since they have a broad effect on secondary metabolism. Analyzing the gene expression profile in A. nidulans strains deleted of LaeA or overexpressing it, Wok et al. could characterize the gene cluster for the antitumor compound terrequinone A (66). In bacteria, this approach had an impressive effect in the activation of many silent genes. For example, the heterologous expression of the pleiotropic regulator

absA1 from S. coelicolor affected the secondary metabolite production

in many Streptomyces host strains, as shown by an increased antimi-crobial activity. Moreover, a molecule known to be produced in other organisms was found for the first time in S. flavopersicus, the potent antibiotic pulvomycin (67).

Chromatin remodeling and post-translational modifications. Modifi-cation of chromatin regulation also led to the discovery of new metab-olites. The COMPASS (complex associated with Set1) complex alters the chromatin landscape by methylation of lysine 4 of histone 3 in eukaryotes (68). An important member of this complex was identified in Saccharomyces cerevisiae and has a putative ortholog in A. nidulans, CclA. The deletion of CclA led to the expression of two silent clusters, one responsible for the biosynthesis of monodictyphenone, emodin and emodin derivatives, and the other one for two anti-osteoporosis

(24)

1

molecules. Although already known to be produced by other organisms, these metabolites were discovered in A. nidulans for the first time (11). Another known compound was discovered to be produced in A.

nidu-lans by interfering with sumoylation, a post-translational modification

which is also involved in regulation processes (69). The deletion of the only sumoylation gene in A. nidulans affected the production of different secondary metabolites and resulted in a dramatic increase of a molecule, which was then recognized as asperthecin. Moreover, the asperthecin biosynthetic cluster could be identified and a preliminary pathway was proposed (70).

2.2 Elucidation of secondary metabolite biosynthetic pathways: methods and significance

Secondary metabolite genes are usually located in clusters, which en-code for a core enzyme and several tailoring enzymes. This means that a single gene cluster forms a variety of molecules in a multi-step process, a biosynthetic pathway. Elucidating pathways deepens the knowledge and understanding of the biosynthetic mechanisms, which leads to more accurate predictions of other gene clusters, the functions of the respective enzymes and the chemical structures produced. Moreover, it opens the possibility for further engineering of the organisms, for example by deleting or overexpressing specific genes in the cluster, in order to redirect the flux towards a particular compound of interest and therefore increase the production. In addition, the pathway character-ization allows to identify new molecules and develop new analogs or unnatural compounds with the desired properties. Secondary metab-olites are a precious source for new drugs (17). However, they might have limitations for human applications, such as bioavailability, stability or toxicity, and therefore they might need optimizations. Apart from organic synthesis or semi-synthesis, which is chemical modifications of a natural product, new molecules with novel or improved biological properties can be produced by engineered microorganisms, for example by building combinatorial biosynthetic pathways having enzymes from different organisms (71). This approach allows even the production of complex chemical structures, which cannot be obtained by organic synthesis or semi-synthesis. Filamentous fungi are confirmed to be an incredible and versatile source of unknown and complicated molecules, which have a great potential for novel applications.

(25)

1: I

ntr

oduc

tion

Gene deletion. A valuable approach to elucidate a pathway is the deletion of the putative biosynthetic genes and comparative analysis of the metabolite profile. In linear metabolic pathways, each compound produced is the substrate of the following enzyme. By deleting one of the biosynthetic genes, the enzymatic reactions will be altered, causing the accumulation of the compounds that are upstream of the missing enzyme and the absence of all downstream metabolites. The sequence of the reactions can be then reconstructed by analyzing the metabolic profile of individual deletion strains by HPLC or LC-MS. The compounds can then be isolated and their structure can be

deter-mined by NMR. With this approach, the pathway for the red pigment aurofusarin in F. graminearum was elucidated. The PKS12 was known to be responsible for the production of this pigment. By silencing the putative transcription activator located close by PKS12, ten surrounding genes were downregulated and therefore considered to be part of the same cluster. The deletion of the single genes in the cluster and the analysis of the metabolites produced by the mutant strains allowed to identify the intermediate molecules in the pathway, including the yellow pigment rubrofusarin, and to determine the function of each biosynthetic gene (72).

More complex is the elucidation of branched biosynthetic pathways, as the intermediates are converted into different molecules and the same enzymes can be involved in multiple steps. In this case, the deletion of a gene does not necessary lead to the accumulation of the upstream metabolites, as other enzymes could use them as substrates. This means that the series of reactions is not necessary interrupted: one branch of the pathway could be favored or the same final compounds could be still produced in alternatives paths. Bioinformatics analyses can help to interpret the metabolite profiling of the deletion strains, providing information about the possible function of the enzymes. In P. rubens, two branched pathways have been described, one responsible for the non ribosomal peptide roquefortine (73) and the other for the polyketides sorbicillinoids (74). In both cases, the direct products of the NRPS roqA and PKSs sorA and sorB are substrates of different enzymes, which imme-diately split the pathways in two branches. However, the two branches reconverge to make the same final molecules. A complex pathway re-sponsible for the biosynthesis of chrysogine in P. rubens will be described in this thesis (chapter 2). The chrysogine pathway is a relevant example of

(26)

1

how enzymes encoded by genes outside the cluster can also be involved in the biosynthetic reactions. In this case, a transaminase catalyzes an early step in the pathway and a transferase can partly replace the func-tion of ChyE, which is part of the chrysogine gene cluster (75).

Expression of biosynthetic genes in non-producing organisms. An-other approach for pathway elucidation is the heterologous expression of the biosynthetic genes in an organism which does not produce the metabolites of interest. This strategy allows to identify the interme-diate compounds in a pathway, as the biosynthetic genes are usually expressed one by one or in different combinations. The intermediates are often novel molecules, not described in the native organism, which mostly accumulate only the final metabolites. As previously described, the production of aspyridone in A. nidulans has been achieved by over-expressing the transcriptional regulator of the putative gene cluster, which consists of a PKS-NRPS hybrid and seven ORFs (76). For the char-acterization of the pathway, Wasil et al. expressed the genes in A. oryzae, under the control of strong promoters. The expression of the PKS-NRPS hybrid and the enoyl reductase resulted in the production of three new compounds, the early intermediates of the pathway. By introducing different combinations of the genes, eight new metabolites have been discovered (77). A similar approach has been used for the elucidation of the citreoviridin biosynthetic pathway. Citreoviridin is a toxin that inhibits the mitochondrial oxidative phosphorylation. In this case, the biosynthetic genes were unknown, but they were putatively indicated by identifying the gene that confers self-resistance to citreoviridin. In fact, toxin-producing organisms often have the resistance gene within the toxin biosynthetic gene cluster, for example an extra copy of the target of the compound. The heterologous expression of the putative cluster from A. aureus var. aureus into A. nidulans resulted in the identification of novel metabolites and the genes that are involved in the pathway (78). In the previous examples, the biosynthetic genes have been expressed in heterologous hosts. However, this strategy can be also applied in the native organism, if it is genetically deleted of the gene cluster of interest. In chapter 2, we describe how the expression of the single NRPS chyA in a P. rubens chrysogine cluster deleted strain allowed the identification of the early compounds of the pathway (75).

Feeding experiments. Feeding the precursor molecules can help to identify their role in the pathway. By feeding the substrates of the

(27)

1: I

ntr

oduc

tion

NRPS roqA to P. rubens, Ali et al. showed that tryptophan is limiting in the production of roquefortines. In fact, increasing concentrations of tryptophan resulted in an increased production of the metabolites in the pathway, while no significant effect was evident when histidine was fed (73). Feeding experiments can be also used to restore the me-tabolite production in mutant strains and confirm the characterized pathway (75, 79). Moreover, feeding can be a strategy to obtain novel compounds. In the mutasynthesis approach, a key enzyme in a partic-ular pathway is deleted, blocking the production of the downstream molecules. However, by feeding the engineered organism with mod-ified versions of the natural intermediates, new compounds can be produced. Cyclooligomer depsipeptides like bassianolide, beauvericin and enniatins have antimicrobial, insecticidal, cell migration inhibitory activities and antiproliferative function against human cell lines. The biosynthetic pathway is therefore of high interest for the production of drugs or insecticides. In Beauveria bassiana these molecules are origi-nated from a common precursor, D-hydroxyisovalerate (D-Hiv). After identification of the gene responsible for the formation of D-Hiv, Xu et al. deleted it and used the mutant strain for a combinatorial feeding of diverse D-Hiv analogs. This resulted in production of fourteen new beauvericin-like compounds, showing different antiproliferative and cell migration inhibitory activities (80).

2.3 Engineering the core biosynthetic enzymes for chemical diversification

NRPSs, PKSs and hybrids are arranged in modules and domains, each of them having a specific role in the formation of the final compound. Therefore, altering this organization might result in new chemical

struc-tures. However, this approach has some limitations, as the structural modifications might affect the protein folding as well as the intra- or intermolecular interactions. The engineering of the core biosynthetic geneshas been successfully achieved in bacteria, leading to the bio-synthesis of novel variants of the natural products. Module exchange in the NRPS involved in daptomycin biosynthesis resulted in new mol-ecules. Replacing the D-alanine selective module 8 with the D-serine selective module 11 and vice versa, the expected novel analogs were produced. Although the production was lower than in the parental strain, the antimicrobial activity was retained. The exchange was successful

(28)

1

even when an heterologous module from a similar enzyme was used (81). Besides module or domain swapping, novel compounds can be obtained by altering the selectivity of the A domain, through target amino acid mutations or direct evolution (81). The major achievements have been accomplished in the engineering of bacterial enzymes. How-ever, a few significant examples are present in fungi. In A. nidulans, the domain responsible for the selection of the starter units in the asper-furanone PKS was swapped with the one from the sterigmatocystin PKS, resulting in the formation of a new compound (82). A. terreus has five NRPS-like genes with a single module of A-T-TE domains. Three of them have been associated with a product: ApvA produces aspulvi-none, BtyA forms butyrolactone IIa and PgnA phenguignardic acid. To investigate the possibility to make novel molecules, hybrids of these known NRPS-like enzymes have been expressed heterologously in

A. nidulans. Exchanging the A domain of BtyA with the A domain of

ApvA still resulted in the production of butyrolactone IIa, since both enzymes utilize the same precursor but perform different cyclizations. However, the hybrid carrying the A domain of PgnA produced a new molecule, phenylbutyrolactone IIa.

3. Heterologous production of secondary metabolites

The first examples of heterologous expression of secondary metabolite biosynthetic genes go back to the 80’s. In 1985, the gene cluster for the biosynthesis of the antibiotic actinorhodin from S. coelicolor was introduced into medermycin-producing Streptomycetes strains, resulting in the production of the hybrid antibiotics mederrhodins A and B (83). Since then, the heterologous expression has become easier and more accessible, due to the development of new tools for the engineering, the improvement of the host strains as well as the decrease in the cost of synthetic genes. Expressing the chosen biosynthetic genes into a heterologous host is ideal when the native organism is pathogenic or not characterized, as there is no or little knowledge about transformation procedures, genetic engineering, growth and fermentation conditions for product formation. Even when the native producer is well known, the heterologous expression offers several advantages, such as over-coming the endogenous gene regulation, increasing/controlling the

(29)

1: I

ntr

oduc

tion

production by using strong/inducible promoters, revealing new bio-chemistry. Hereby, examples of heterologous hosts and results achieved by heterologous gene expression are summarized.

3.1 Bacteria and yeast as heterologous hosts for secondary metabolite production

Escherichia coli and Streptomyces species are commonly used for

ex-pression of bacterial genes, due to the developed tools available for their engineering. These organisms have been also successfully used for the production of secondary metabolites. However, since E. coli does not have an extensive endogenous secondary metabolism, further engineering is needed such as expression of a heterologous phospho-pantetheine transferase to activate NRPSs and PKSs or optimization of precursors pathways (84). The precursor of the polyketide erythromycin from Saccharopolyspora erythraea was produced in E. coli in 2001 (85) and few years later further engineering resulted in the biosynthesis of the final active antibiotics erythromycin C (86) and A as well as novel designed analogs (87). In 2018, E. coli was used as platform organism for the production of the carbapenem antibiotic (5R)-carbapen-2-em-3-carboxylic acid from Pectobacterium carotovorum (88). E. coli successfully produced also fungal and plant metabolites (84), as 6-methylsalicylic acid from P. patulum (89) and precursors of the anticancer paclitaxel from the yew tree (90). Many Streptomyces species have been used as cell factories for the production of valuable compounds from other Streptomycetes (91), as the antibiotic daptomycin (92). Streptomycetes naturally produce several secondary metabolites, which makes them good host candidates, as the precursors are already available in the cell. To avoid competition among different biosynthetic pathways, some species have been deleted of the major secondary metabolite gene clusters (93, 94). For example, a region of about 1.4 Mbp was removed from the 9.02 Mbp genome of the industrial S. avermitilis species, generating a strain with a reduced metabolic profile and still able to produce higher amounts of heterologous streptomycin and cephamycin C, when compared to the full-size genome parental strain (94).

S. cerevisiae is a versatile organism for heterologous gene expression.

Thanks to its natural ability to recombine several DNA fragments hav-ing short homologous regions, it is possible to rapidly and efficiently assemble complex constructs like biosynthetic pathways. Moreover,

(30)

1

many genetic tools are available for engineering, such as the clustered regularly interspaced short palindromic repeats (CRISPR)-associated RNA-guided DNA endonucleases (Cas9) (95) or for controlling gene expression, like sets of inducible promoters, diverse plasmids, synthetic transcription factors (96–98). In addition, the lack of a significant endog-enous secondary metabolism may allow a better detection of the novel compounds and reduce possible interferences and competitions among pathways (96). In 2006, S. cerevisiae was engineered for the production of artemisinic acid, the precursor of the antimalaria agent artemisinin, a plant-derived sesquiterpene (99). More recently, the penicillin gene cluster from P. rubens was introduced in the baker’s yeast, resulting in the production of active penicillin G. The amount of antibiotics pro-duced was furthermore optimized by a combinatorial cloning with promoters of different strengths (100). An impressive work by Harvey et al. established a platform for rapid and inducible expression of fungal secondary metabolite gene clusters in S. cerevisiae (101). Large genetic constructs can be built on plasmids by in vivo homologous recombina-tion, transforming the genes of interest with standard parts, including markers and a set of newly characterized inducible promoters. In this way, the researchers expressed 41 gene clusters with unique features, selected through extensive bioinformatics analysis, from Ascomyce-tes and BasidiomyceAscomyce-tes. 22 clusters produced detectable compounds, many of which were never described before. This platform represents an exceptional tool for the discovery of novel metabolites and strongly confirms S. cerevisiae as remarkable candidate for the expression of secondary metabolite genes and production of compounds.

3.2 Heterologous secondary metabolite production in filamentous fungi

Filamentous fungi are used as host for heterologous production of secondary metabolites. Being able to produce a wide range of diverse and complex molecules, filamentous fungi are extremely versatile and most likely already possess the substrates needed for the production of the compounds of interest. Nonetheless, the extensive endogenous secondary metabolism means there is an intricate net of biosynthetic pathways. In order to decrease the complexity of the metabolite profiles and make new or low-produced compounds easy to detect, Chiang et al. deleted eight highly expressed BGCs in A. nidulans, for a total of about

(31)

1: I

ntr

oduc

tion

244 kbp. The new strain showed a reduced secondary metabolites background and it was possible to detect a new compound, named aspercryptin. The researchers could also identify the biosynthetic genes and propose a pathway (102).

Heterologous gene expression in filamentous fungi allowed to identify the function of genes, especially from organisms which are difficult to handle, being harmful or not well characterized. In the case of patho-genic organisms, the characterization of the biosynthetic genes and pathways could be of particular importance for understanding the mechanism of pathogenicity. Dermatophytes from the Trichophyton and

Arthroderma genera have a PKS and four flanking genes, including a

tran-scription factor, that are highly conserved among different species (103). Moreover, they also show homology with genes from the pathogenic fungi A. fumigatus and Neosartorya fischeri, which are responsible for the biosynthesis of the immunosuppressant neosartoricin (104). In order to assign a function to these genes from T. tonsurans, Yin et al. expressed them into A. nidulans, which is a well known, safe model organism often used for genetic studies and it does not produce neosartoricin. The expression of the transcription factor was driven by the strong gpdA promoter from A. nidulans itself, to guarantee the overexpression of the other heterologous genes. The engineering resulted in the production of a novel compound, which had the same UV absorption as neosartoricin but a smaller m/z. NMR characterization showed that this compound was similar to neosartoricin but without an acetyl group, probably because A. nidulans does not have or does not express the required acetyltransferase. The compound was named neosartoricin B. Moreover, researchers showed that the metabolite could convert into two similar compounds, neosartoricin C and D. The heterologous expression in A.

nidulans demonstrated that the PKS and the four flanking genes from T. tonsurans are responsible for neosartoricin B biosynthesis. Furthermore,

by expressing additional flanking genes which had no homologs in A.

fumigatus or N. fischeri, the researchers could exclude their involvement

into further modifications of neosartoricin B (103).

In 2017 Zhang et al. showed that the PKS PfmaE from the endophytic fungus Pestalotiopsis fici was involved in spore pigmentation and de-velopment (79). However, although they used different methods for metabolites extraction and analysis, the researchers could not identify the compounds associated with the PKS, probably because they were

(32)

1

produced in low amount or in polymers, difficult to detect. Therefore, the PKS gene was expressed into A. nidulans, resulting in the production of two compounds, from which one was the known pigment flaviolin. The heterologous expression of the entire putative gene cluster led to

the identification of another known metabolite, scytalone. In order to define the role of the biosynthetic genes into the pathway, they were individually deleted from the heterologous recombinant host strain. This approach allowed for the identification of the essential

biosyn-thetic genes and the transcriptional regulator. Overexpression of the transcription factor resulted in the upregulation of the gene cluster and stronger pigmentation, due to the accumulation of flaviolin.

The heterologous gene expression in known organisms can lead to higher production compared to the native strain. In 2010 Heneghan et al. rebuilt the tenellin biosynthetic gene cluster from the entomopatho-genic fungus B. bassiana into A. oryzae. It was already known that tenellin is produced in a short pathway by a PKS-NRPS hybrid, a trans-acting enoyl reductase and two P450 oxidases. The four biosynthetic genes were expressed into A. oryzae by replacing the native promoters with the starch-inducible promoter PamyB. Due to the usage of a strong promoter, endogenous of the host organism, the engineered strain was able to produce tenellin in an amount five times higher than the native organ-ism (105). An even more surprising result has been achieved by Bailey et al. in 2016. In their work, for the first time, the researchers engineered

A. oryzae with a secondary metabolite gene cluster from a

basidiomy-cete fungus. Basidiomybasidiomy-cetes produce several secondary metabolites, but the lack of genetic engineering tools and the dikariotyc nature of these organisms makes it difficult to identify the biosynthetic genes or improve the production. Pleuromutilin is a diterpene produced by

Clito-pilus passeckerianus and it has antimicrobial activity. Its derivatives are

used as antibiotics for treating animal and human infections, therefore there is great interest in understanding the biosynthetic mechanisms and increasing the production. After identifying the pleuromutilin gene cluster, Bailey et al. attempted to overexpress it in C. passeckerianus itself, by using native and strong heterologous promoters. However, the engineered strains showed no increase in pleuromutilin produc-tion or even lower producproduc-tion compared to the parental strain. On the contrary, the heterologous expression of the biosynthetic genes under constitutive promoters from A. oryzae resulted in a ten times increased

(33)

1: I

ntr

oduc

tion

Table 2 List of NRPSs, PKSs and PKS-NRPS hybrids and their related products in P. rubens. YWA1, naphtho-γ-pyrone; DHN, 1,8-dihydroxynaphthalene; 6-MSA, 6-methyl salicylic acid; DMOA, 3,5-dimethylorsellinic acid.

Gene ID Products

NRPSs Pc13g05250 Ferrichrome (siderophore)

Pc13g14330

-Pc16g03850 Coprogen/Fusarinine (siderophore)

Pc16g04690 Fungisporin and related Hydrophobic Cyclic Tetrapeptides

Pc21g01710

-Pc21g10790

-Pc21g12630 Chrysogine and related metabolites

Pc21g15480 Roquefortines/Meleagrine Pc22g20400 Fusarinine (siderophore) Pc21g21390 Penicillins PKSs Pc12g05590 -Pc13g04470 -Pc13g08690 -Pc16g00370 6-MSA/Yanuthones Pc16g03800 -Pc16g04890 -Pc16g11480 -Pc21g00960 -Pc21g03930 -Pc21g03990 -Pc21g04840 -Pc21g05070 Sorbicillinoids Pc21g05080 Sorbicillinoids Pc21g12440 -Pc21g12450 -Pc21g15160 -Pc21g16000 YWA1/DHN-Melanin Pc22g08170 6-MSA/Patuline Pc22g22850 DMOA/Andrastin A Pc22g23750 -hybrids Pc14g00080 -Pc16g13930

(34)

-1

production, besides reducing the time of sampling, as the fermentation of C. passeckerianus requires a long preculture (106).

4. Secondary metabolism in P. rubens

The filamentous fungus P. rubens is largely known for producing peni-cillins at a commercial scale (7), but it is able to synthetize several other secondary metabolites (107). The natural production of β-lactam anti-biotics has been dramatically increased by classical strain improvement (CSI), consisting of several rounds of mutagenesis and selection (108). This led to many genetic modifications, including amplification of the

biosynthetic penicillin gene cluster, alteration of metabolic flux towards the precursors of penicillin, proliferation of microbodies, where the last step of β-lactams biosynthesis occurs (107). Moreover, other secondary metabolite clusters were affected, resulting for example in the loss of pigmentation (109, 110). Besides improvements of penicillin yields, also fermentation features were enhanced, making these strains ideal for further studies and applications.

From the ancestor strain NRRL1951, many lineages have been ob-tained, some of them showing an increase of β-lactams production of at least three orders of magnitudes (107, 109). An important interme-diate in the CSI process is the Wisconsin 54-1255 mutant, which was sequenced in 2008 and became the reference strain for P. rubens. The genome sequencing revealed the presence of 10 NRPS, 20 PKS and 2 PKS-NRPS hybrid genes (107) (Table 2), demonstrating the metabolic potential of this organism. Transcriptomic analysis showed that only few of these genes were expressed under standard laboratory conditions. The endogenous secondary metabolism was investigated in an

indus-trial strain deleted of the penicillin gene cluster (DS68530), in order to facilitate the detection of new molecules or compounds produced in low amount. By deletion of two highly expressed NRPSs, the roque-fortine and hydrophobic cyclic tetrapeptides biosynthetic genes were identified (73, 111). The roquefortine NRPS was suggested to be part of a cluster together with six further genes, which were co-regulated. The biosynthetic pathway was then elucidated by generating single gene deletion strains and analyzing their metabolite profile (73). Another up-regulated NRPS was linked to the pigment chrysogine, which is

(35)

1: I

ntr

oduc

tion

produced in significant amounts in P. rubens (109). Three other NRPSs were predicted to be siderophore synthetases by bioinformatics analysis (107). Later, their function was confirmed experimentally, by growth in iron-limiting conditions and metabolite analysis of the respective deletion strains (112). Furthermore, another siderophore synthetase was identified by overexpressing a silent NRPS gene (112).

Among the 20 PKS genes present in P. rubens genome, only few of them have been characterized and associated to a metabolite. The biosynthetic gene cluster for the yellow pigments sorbicillinoids has been extensively studied. During the CSI, the two neighboring PKS genes (PKS12 and PKS13) have accumulated mutations which resulted in the absence of pigmentation. In particular, a single amino acid change in the KS domain of PKS13 is responsible for the inactivation of the core enzyme in the biosynthetic pathway (110). When the functional

PKS13 gene was deleted from the ancestor strain NRRL1951, the mutant

was not able to produce the yellow pigments; the restoration of the mutation in the industrially improved strain DS68530 led to the pro-duction of sorbicillinoids (110). Moreover, the biosynthetic pathway was elucidated and a complex mechanism of regulation was characterized (74). Regarding the two PKS-NRPS hybrids, no extensive studies have been performed yet.

4.1 Genetic toolbox development in P. rubens

The CSI program that ran for several decades has generated strains with excellent fermentation capabilities and that are optimized for the biosynthesis of penicillins. Therefore, because of the generic way in which fermentation and metabolic features have been tailored in these strains, industrial P. rubens strains may be ideal hosts for the production of other valuable compounds from other organisms. Nonetheless, to achieve this goal, it is important that these fungi are well accessible for genetic engineering. In recent years, the genetic toolbox for P. rubens has been greatly enhanced. Hereby, we report the major advances in the field.

Strain development. By the deletion of the hdfA gene (113, 114), in-volved in the non-homologous end joining (NHEJ) DNA repair pathway, homologous recombination (HR) is favored, which facilitates more effective in vivo gene inactivation and replacement methodologies as well as construction of complex pathways. Furthermore, the removal

(36)

1

of multiple copies of penicillin biosynthetic genes from the CSI strains allows an easier detection of compounds that are produced in low quantities. Besides the β-lactams, several other metabolites are pro-duced in high levels. Recently, in addition to the β-lactam gene cluster, the highly expressed NRPS clusters responsible for the production of roquefortine, chrysogine and hydrophobic cyclic tetrapeptides were removed genetically. This strain shows a reduced background of en-dogenous metabolites and may therefore be a favorable host strain precluding possibly competition among endogenous and exogenous biosynthetic pathways (Pohl et al., submitted).

Genome editing. A further major advance has been the introduction of the CRISPR/Cas9 technology for genome editing in P. rubens. With the CRISPR/Cas9 genome editing (115, 116), DNA double strand breaks (DSB) can be efficiently created in a specific locus, awakening the machinery for DNA repair. In a strain deleted of the hdfA gene, the DSB is mainly fixed through the HR system. Therefore, the insertions of the desired genetic parts can be achieved by creating homology in the DNA sequences. In this way, multiple fragments can be delivered and recombine more efficiently even with short homologous regions (100 bp instead of 1 kbp as previously used), thus large gene clusters can be built. The Cas9 endonuclease and the single guide RNA (sgRNA), which leads it to the target locus, can be transiently expressed on an AMA1 (autonomous maintenance in Aspergillus) based plasmid or delivered during transformation (116).

Expression systems. For gene expression, the use of different pref-erentially inducible promoters is desired. A set of constitutive and in-ducible promoters of different strengths is now available for P. rubens (117) (Büttel et al., unpublished), allowing to optimize the expression of the specific genes of interest (GOI). Regulation systems can also tune the GOI expression. A new transcription control device is based on the Q-regulatory system from Neurospora crassa and has been ap-plied in different organisms (118–121). For the application in P. rubens, a two component synthetic transcription factor QF has been created, consisting of the N. crassa DNA binding domain, which recognizes the repeated upstream activation sequences (UAS) and a Herpes simplex virus activation domain, which recruits the RNA polymerase on a core promoter, driving the expression of the GOI. The level of expression is proportional to the number of UAS upstream the core promoter, making

(37)

1: I

ntr

oduc

tion

the Q system versatile and tunable. The Q system has been success-fully used for regulating the expression of the penicillin gene cluster (122). A high gene expression level can be also achieved by integrating the target gene on a plasmid. In P. rubens, the AMA1 based plasmids (123) are able to replicate autonomously and they can be maintained in variable copy numbers. The AMA1 plasmids require a marker for selection and continuous selection pressure to prevent loss during cell divisions. The addition of the essential gene Tif35 (translation initiation factor 35) makes the vector stable, if the same gene is simultaneously deleted from the genome of the host organism (124). The stable plasmid has been used for the assembly of the penicillin pathway in a penicillin cluster free strain, resulting in the restoration of penicillin production, as shown by bioactivity assays (125).

Selection markers. Markers are required for an efficient selection of the transformants. Two selection markers have been extensively used since the 80s: the acetamidase amdS gene from A. nidulans allows for growth in the presence of acetamide as sole nitrogen source (126), while the phleo gene from Streptoalloteichus hindustanus confers resistance against the antibiotic phleomycin (127). More recently, the overexpres-sion of the squalene epoxidase-encoding ergA was shown to confer resistance against its inhibitor terbinafine (128). The auxotrophic marker histidine, already used for A. niger (129), resulted to be efficient also for

P. rubens (unpublished). A larger set of markers facilitates the genetic

engineering, as for examples multiple deletions can be performed in the same strain. However, it is also possible to use a marker transiently, for example by co-transforming a deletion cassette for a target gene with an AMA1 plasmid on which the marker can recombine. In this way, when the selection pressure is released, the plasmid is lost and the strain is deleted of the gene of interest but does not have any marker integrated in the genome.

4.2 Industrial P. rubens strains as platform for the production of heterologous secondary metabolites

Besides the toolbox development for genetic engineering, P. rubens has proved to be a versatile organism for the expression of heterologous genes of bacterial and fungal origin and for the production of chemical diverse molecules of industrial interest. Interestingly, codon optimiza-tion was not needed for adequate gene expression.

(38)

1

Exploiting the first enzymes in the biosynthesis of penicillins, ceph-alosporin C and other different intermediates in the semi-synthetic production of cephalosporins have been produced in P. rubens. The intermediate adipoyl-7-aminodeacetoxy-cephalosporanic acid (ad-7-ADCA) has been obtained by expressing the deacetylcephalosporin expandase/hydroxylase cefEF from the filamentous fungus

Acremom-ium chrysogenum (130), which acts on the adipoyl-6-APA (ad-6-APA)

produced in the penicillin pathway. Through the further expression of the deacetylcephalosporin acyltransferase cefG, ad-7-ADCA can be converted into adipoyl-7-aminocephalosporanic acid (ad-7-ACA), from which an important intermediate in the semi-synthesis of cephalospo-rins is derived, 7-ACA (130). Cephalospocephalospo-rins can be also synthesized from adipoyl-7-amino-3-carbamoyloxymethyl-3-cephem-4carboxylic acid (ad-7-ACCCA), which has been produced in P. rubens by the expres-sion of cefEF and the 3’ hydroxymethylcephem-O-carbamoyltransferase

cmcH from S. clavuligerus (131). Moreover, intracellular

deacetylceph-alosporin C and cephdeacetylceph-alosporin C, from which 7-ACA can be obtained, was accumulated in a strain expressing genes from A. chrysogenum (the isopenicillin N-CoA synthetase cefD1, the IPN-CoA epimerase cefD2,

cefEF and cefG) (132).

P. rubens has been also successfully engineered for the production

of the cholesterol-lowering drug pravastatin (133). Pravastatin is ob-tained from the hydroxylation of the natural compound compactin, which is produced by P. citrinum. The expression of the heterologous biosynthetic gene cluster resulted in the production of compactin in amounts higher than the native organism. Furthermore, by introducing an evolved cytochrome P450 from the actinomycete Amycolatopsis

orientalis, the specific hydroxylation of compactin into pravastatin was

achieved. In 10 L fed-batch fermentations, the engineered P. rubens strain was able to produce more than 6 g/L of pravastatin.

5. Concluding remarks

Fungal secondary metabolites have a wide range of biological activi-ties, which make them highly attractive for human applications. Many natural compounds are already used in medicine, industry or agricul-ture. However, the advances in genome sequencing and bioinformatics

University of Groningen Exploring the metabolic potential of Penicillium rubens Viggiano, Annarita