Antifungal and Antileukemic Compounds from Cyanobacteria : Bioactivity, Biosynthesis, and Mechanism of Action

(1)

ANTIFUNGAL AND ANTILEUKEMIC COMPOUNDS FROM CYANOBACTERIA:

BIOACTIVITY, BIOSYNTHESIS, AND MECHANISM OF ACTION

Anu Humisto

Department of Microbiology Faculty of Agriculture and Forestry

University of Helsinki, Finland

ACADEMIC DISSERTATION

To be presented, with the permission of the Faculty of Agriculture and Forestry of the University of Helsinki, for public examination in Auditorium I at Info Centre Korona,

Viikinkaari 11, on 23^rd November 2018, at 12 o'clock noon.

Helsinki 2018

(2)

Supervisors Professor Kaarina Sivonen Department of Microbiology University of Helsinki, Finland Professor Lars Herfindal Department of Clinical Science University of Bergen, Norway Docent Jouni Jokela

Department of Microbiology University of Helsinki, Finland

Reviewers Docent Päivi Tammela

Division of Pharmaceutical Biosciences Faculty of Pharmacy

University of Helsinki, Finland Professor Lena Gerwick

Center for Marine Biotechnology and Biomedicine Scripps Institution of Oceanography

University of California San Diego, USA Thesis evaluation committee Professor Per Saris

Department of Microbiology University of Helsinki, Finland Academy Fellow Miia Mäkelä Department of Microbiology University of Helsinki, Finland

Opponent Professor Jeanette H. Andersen

Norwegian College of Fishery Science University of Tromsø, Norway

Custos Professor Kaarina Sivonen

Department of Microbiology University of Helsinki, Finland

Dissertationes Schola Doctoralis Scientiae Circumiectalis, Alimentariae, Biologicae

Cover: Lake Löytänejärvi (Pori, Finland), light micrograph ofNostoc sp. UHCC 0450, and transmission electron micrograph of MOLM-13 cell organelles. All figures by Anu Humisto.

ISSN 2342-5423 (print) ISSN 2342-5431 (online)

ISBN 978-951-51-4658-8 (paperback)

(3)

List of original publications

This thesis is based on the following publications:

I Humisto A, Herfindal L, Jokela J, Karkman A, Bjørnstad R, Choudhury RR, Sivonen K. 2016. Cyanobacteria as a source for novel anti-leukemic compounds.Current Pharmaceutical Biotechnology, 17(1):78–91.

II Shishido TK,Humisto A, Jokela J, Liu L, Wahlsten M, Tamrakar A, Fewer DP, Permi P, Andreote APD, Fiore MF, Sivonen K. 2015. Antifungal compounds from cyanobacteria.Marine Drugs, 13(4):2124–2140.

III Humisto A, Jokela J, Liu L, Wahlsten M, Wang H, Permi P, Machado JP, Antunes A, Fewer DP, Sivonen K. 2018. The swinholide biosynthesis gene cluster from a terrestrial cyanobacterium, Nostoc sp. strain UHCC 0450.

Applied and Environmental Microbiology, 84(3):e02321-17.

IV Humisto A, Jokela J, Teigen K, Wahlsten M, Permi P, Sivonen K, Herfindal L. 2018.

Characterization of the membrane interaction of the antifungal and cytotoxic cyclic glycolipopeptide hassallidin. Submitted manuscript.

The publications are referred to in the text by their roman numerals.

Author's contribution

I Anu Humisto participated in the design of the study, performed the experiments, participated in the data analysis, and wrote the first article draft.

II Anu Humisto participated in the design of the study, performed part of the experiments, and participated in the writing of the manuscript.

III Anu Humisto participated in the design of the study, performed the experiments, participated in the data analysis, and wrote the first article draft.

IV Anu Humisto participated in the design of the study, performed the experiments, participated in the data analysis, and wrote the first article draft.

(6)

List of related publications

V Pancrace C, Jokela J, Sassoon N, Ganneau C, Desnos-Ollivier M, Wahlsten M, Humisto A, Calteau A, Bay S, Fewer DP, Sivonen K, Gugger M. 2017.

Rearranged biosynthetic gene cluster and synthesis of hassallidin E in Planktothrix serta PCC 8927.ACS Chemical Biology, 12(7):1796–1804.

(7)

Abbreviations

A Adenylation domain

ACP Acyl carrier protein AML Acute myeloid leukemia

AT Acyltransferase domain

C Condensation domain

cf. sp. Uncertainly defined species

DH Dehydratase

E Epimerization domain

EC50 Half-maximal effective concentration

ER Enoyl reductase

HGT Horizontal gene transfer

IC50 Half-maximal inhibitory concentration IPC-81 Rat acute myeloid leukemia cell line

kb Kilo base pairs

KS Ketosynthase

KS⁰ Nonelongating ketosynthase

KR Ketoreductase

LC-MS Liquid chromatography mass spectrometry

Mb Mega base pairs

MIC Minimum inhibitory concentration

MOA Mechanism of action

MOLM-13 Human acute myeloid leukemia cell line

MT Methyl transferase

NMR Nuclear magnetic resonance NRK Normal rat kidney epithelial cell line NRPS Nonribosomal peptide synthetase

OMT O-methyl transferase

ORF Open reading frame

OX Monooxygenase domain

PCP Peptidyl carrier protein PKS Polyketide synthase

PS Pyran synthase

scp Scytophycin gene cluster

sp. Species

swi Swinholide gene cluster

TE Thioesterase domain

UHCC University of Helsinki Cyanobacteria Culture Collection

(8)

Abstract

Nature is a treasury of bioactive natural products that are developed into pharmaceuticals, cosmetics and other industrial applications. Natural products have an impact on all of us, for instance, in the form of antibiotics. Most of the drugs sold today are natural products or their derivatives. However, the need for new natural products has not decreased but is rather increasing. The incidences of certain diseases such as cancers are rising and resistance to treatment is a major problem. Especially prokaryotes and plants produce these intriguing natural products, which display several different bioactivities. Cyanobacteria are photosynthetic prokaryotes that belong to the most prolific sources of bioactive compounds.

This study expands the knowledge of cyanobacterial natural products, including their activities, biosynthesis, and mechanisms of action. The University of Helsinki Cyanobacteria Culture Collection was utilized in this study. Cultured cyanobacteria were screened for antileukemic and antifungal activity using cell assays and disk diffusion analyses. Bioactive compounds were identified with spectrometric methods. The screenings revealed several bioactive hits, including antifungal, antileukemic and cytotoxic activities. Novel compound candidates and known compounds from new habitats or genera were found. In addition, novel variants of known compounds were identified.

The results from screening were used to select cyanobacterial strains for whole genome sequencing. Genomic analysis was used to identify the biosynthesis gene clusters of the cytotoxic compounds swinholide and scytophycin. In addition, the production of swinholide was confirmed for the first time using axenic cyanobacteria. In the last part of the study, the mechanism of action of hassallidin was determined using cellular assays, im- aging, and lipid models (in vitro andin silico). Hassallidin D purified from Anabaena sp.

showed cholesterol-dependent lytic activity against eukaryotic cells.

The development of a natural product into a drug or other application is a long pro- cess, but natural products are needed also in the future. This thesis contributes to this effort by increasing the understanding of several bioactive cyanobacterial natural compounds at both the functional and molecular level.

(9)

Tiivistelmä

Luonto on kuin aarreaitta, josta ammennamme uusia bioaktiivisia yhdisteitä lääkkeiden, kosmetiikan ja muun teollisuuden käyttöön. Luonnon bioaktiivisilla yhdisteillä on merkit- tävä vaikutus meidän kaikkien elämään, esimerkiksi antibioottien muodossa. Suurin osa myydyistä lääkkeistä onkin luonnonyhdisteitä tai niitä mukailevia synteettisiä yhdisteitä.

Uusien luonnonyhdisteiden löytäminen on kuitenkin edelleen tärkeää, sillä joidenkin sai- rauksien on todettu lisääntyneen. Esimerkiksi syöpien määrä on kasvussa, ja lisäksi lääke- resistenssi on vakava ja kasvava ongelma. Erityisesti prokaryootit ja kasvit tuottavat yhdis- teitä, joilla on useita erilaisia bioaktiivisuuksia ja potentiaalia muun muassa lääkeaineiksi.

Fotosynteettiset prokaryootit syanobakteerit kuuluvat parhaimpiin bioaktiivisten aineiden tuottajiin.

Tässä väitöskirjatutkimuksessa tutkittiin syanobakteerien bioaktiivisia yhdisteitä, mu- kaan lukien niiden aktiivisuus, biosynteesi ja toimintamekanismi. Tutkimuksessa hyödyn- nettiin Helsingin yliopiston syanobakteerikantakokoelmaa. Syanobakteerikantoja seulottiin erityisesti niiden antileukeemisten ja antifungaalisten ominaisuuksien osalta käyttäen solu- viljelytestejä. Massaspektrometrisiä menetelmiä käytettiin yhdisteiden tunnistamiseen.

Työssä havaittiin useita aktiivisia syanobakteerikantoja, joilta löydettiin mahdollisia uusia yhdisteitä. Lisäksi löydettiin useita tunnettuja yhdisteitä ja niiden aiemmin tunnistamatto- mia variantteja uusilta syanobakteerikannoilta tai uusista elinympäristöistä.

Seulontatutkimusten perusteella valittiin syanobakteerikannat kokogenomisekven- sointiin. Genomianalyysin avulla pystyttiin selvittämään biosynteesireitti tunnetuille syto- toksisille yhdisteille nimeltä swinholidi ja skytofysiini. Lisäksi swinholidin tuotto pystyttiin osoittamaan ensimmäisen kerran laboratoriossa kasvatetussa puhtaassa syanobakteerikan- nassa. Väitöskirjan viimeisessä osatyössä tutkittiin tunnetun yhdisteen hassallidiinin toimintamekanismia, missä hyödynnettiin solutestejä, kuvantamista ja lipidimallinnusta.

Anabaena-syanobakteerista puhdistetun hassallidiinin havaittiin hajottavan kolesterolia si- sältävien eukaryoottisolujen solukalvon.

Luonnonyhdisteen kehittäminen lääkkeeksi tai muuksi lopputuotteeksi on hyvin hi- dasta, mutta tarpeellista tulevaisuuden kannalta. Tämä väitöskirjatutkimus lisää ymmärrys- tämme useista syanobakteerien tuottamista bioaktiivisista yhdisteistä sekä toiminnallisella että molekulaarisella tasolla.

(10)

(11)

1 Introduction

1.1 Natural products and their importance

Natural products are compounds produced in nature by micro- or macroorganisms. They are present from the bottom of the ocean to your backyard. A huge variety of places, sources and specific organisms have been studied for novel natural products, but even more striking and varied are the discovered chemical structures and features of these molecules.

They vary from small to large, linear to cyclic, and certain molecule type to hybrids of several types, and many of them possess biological activities. Natural products have activities against many targets, including multiple mammalian and cancer cell lines, bacteria, viruses, parasites and more. Consequently, natural products are an important source of pharmaceuticals (Newman & Cragg 2016). In addition, natural products have found their way into many agricultural and nutritional applications as well as into chemical probes in research and industry (Cantrellet al. 2012; Singhet al. 2017).

The era of natural product discovery started to flourish by the finding of antibiotics and the success of penicillin in the 1940s (Davies & Davies 2010; Giddings & Newman 2013).

The discovery of penicillin proved the value of natural products and raised interest in finding more. Today, most of the drugs on the market are natural products or derived from them (Newman & Cragg 2016). These include a wide variety of drugs, not just antibiotics. From the year 1981 to 2014, 174 new anticancer and 32 antifungal drugs with novel chemistry were approved to the market, and among these, most of the anticancer drugs were natural product derived compounds (17 being unmodified natural product structures) (Newman &

Cragg 2016). In contrast, the majority of accepted antifungal agents were synthetic in origin (Newman & Cragg 2016). Finding cures to diseases has been a principal goal in natural product discovery, with human interest guiding the research. The strong interest to find new drugs does not show signs of diminishing but rather the opposite, as antibiotic resistance is increasing as well as the incidence of cancers with insufficient treatment options. In recent years, producing antibody conjugate drugs, where a natural product is attached to an antibody that targets a certain tissue or tumor cells, has gained considerable interest (Becket al. 2017). Such combination techniques provide new prospects also for those natural prod- ucts that, for instance, display unacceptable levels of toxicity.

Over the years, scientists and the industry have shown interest in natural products not only for use as pharmaceuticals but also for use in research and biotechnological applications. Nevertheless, their role in nature is also worthy of attention. Natural products, also called secondary metabolites or specialized metabolites, are not usually considered essen- tial for the growth or reproduction of the producing organism. However, they can play an important role in the survival and distribution of the organism. For instance, they may act as signaling molecules between organisms or with the environment, or represent defense or predation mechanisms for the organism. Information regarding produced compounds, the targets of attack or defense, and the functioning of compounds in their own environ-

(12)

ment, is often lacking. However, knowledge regarding ecological aspects may help to identify active parts of the compound, improve productivity or even find novel compounds. In addition to providing benefits to producer organisms, natural products may be useful to surrounding organisms. Many examples of such cooperation exist, such as the symbiosis within sponges or lichens.

Natural products are assembled from small building blocks, often via ribosomal or nonribosomal biosynthesis pathways. The biosynthesis pathways are composed of individ- ual enzymes, which together produce the chemical structure and properties of a natural product and further enable variation and production of multiple variants of the natural product (Firn & Jones 2000; Hertweck 2009; Marahiel 2016). The identification of biosynthesis pathways and sequencing whole genomes have greatly altered the focus of natural product discovery. Currently, genome-mining projects are undertaken alongside traditional bioactivity screening studies. Genetic tools have reaccelerated natural product discovery after experiencing a slump, including abandonment by several pharmaceutical companies.

However, finding active hits is still a difficult task as only a fraction of compounds are active or useful (Firn & Jones 2000). Nevertheless, the omics era has proven that there is still much to discover. For example,Actinobacteria have been an important source of natural products for a long while but genomic studies have shown that they still harbor previously unidenti- fied biosynthesis gene clusters (see e.g. Bentleyet al. 2002). Other prolific bacterial sources includeProteobacteria,Firmicutes, andCyanobacteria but also fungi belonging toAscomy- cota contain vast amounts of natural product biosynthesis pathways (Wanget al. 2014, 2015). Thus, microbes have remained a key target for searching natural products. One mi- crobial species may produce a number of different compounds and have many biosynthesis pathway genes in its genome. Furthermore, bioinformatic methods can reveal the evolutionary origins of natural product diversity in nature and horizontal gene transfer (HGT) events which have been shown in many cases (Jensen 2016; Ziemertet al. 2016).

Altogether, natural products play a crucial part in our lives as well as in the lives of microbes. Vast amounts of different bioactivities have been described from these natural products but many compounds are yet to be discovered. In addition to scientific curiosity, the human utility perspective drives natural product discovery. Important aspects of natural product research include identifying chemical structures and biosynthesis genes, and eval- uating and testing the activities and mechanisms of action of the compounds.

1.2 Cyanobacteria and their natural products

Cyanobacteria are oxygenic photosynthetic prokaryotes comprising a single phylogenetic branch in the domain Bacteria (Komáreket al. 2014; Castenholz 2015). These photoauto- trophic bacteria evolved early in Earth's time scale, and are believed to have influenced the

(13)

cells to branching filaments (Whitton & Potts 2012; Komáreket al. 2014; Castenholz 2015).

A number of cyanobacteria live as planktonic organisms in oceans, seas and fresh waters (Whitton & Potts 2012). These water environments are inhabited also by benthic cyanobacteria, which are attached to surfaces such as rocks, plants and animals. Moreover, cyanobacteria are found from terrestrial, symbiotic and extreme habitats including hot springs, and hot and cold deserts (Whitton & Potts 2012; Ciréset al. 2017). Besides the oxygenic photosynthetic properties of cyanobacteria, many of them can fix nitrogen (Castenholz 2015). Cyanobacteria have specialized cells, heterocytes, for nitrogen fixation. Cyanobacte- ria may also produce other specialized cells such as akinetes (resting cells) or hormogonia (chain of migrating cells) (Castenholz 2015). Furthermore, cyanobacteria produce various natural products (Burjaet al. 2001; Welker & von Döhren 2006; Singhet al. 2011).

Cyanobacteria have gained widespread interest in recent years. Their photosynthetic machinery, various natural products and adaptation to various environments provide many avenues for applications. Cyanobacteria are sold as a health food supplement with great success, and they are used in feedstock and in agriculture as a soil additive (Abedet al. 2009;

Mazardet al. 2016; Singhet al. 2017). One of the main areas of interest is currently biofuel production (Mazardet al. 2016; Singhet al. 2017). The current interest in cyanobacteria, however, started from toxins. Cyanobacterial blooms (example in Figure 1) are a considerable risk for humans and animals all around the globe, including the Baltic Sea and fresh water environments (Sivonen et al. 1989, 2007; Carmichael 2001; Codd et al. 2005a,b).

Unfortunately, toxin production cannot be easily distinguished from blooms. Toxic cyanobacteria cause human and animal poisonings, shellfish poisoning, problems in fishery, and skin irritation (swimmer's itch) to recreational users (Codd et al. 2005a; Sivonen 2009).

Cyanobacteria also cause aesthetic harm owing to water odour and taste problems (Watson 2003; Coddet al. 2005b). The human liver toxin microcystin, which is often associated with bloom forming cyanobacteria, is the most extensively studied cyanobacterial natural product (Carmichaelet al. 1988; Sivonen 2009). Other cyanobacterial toxins include anatoxin, aplysiatoxin, cylindrospermopsin, nodularin, and saxitoxin (Codd et al. 2005b; Sivonen 2009).

Figure 1. Cyanobacterial bloom in Finnish coastal area of Baltic Sea (Laajasalo, Helsinki, July 2018).

(14)

Besides toxins, many other bioactive natural products have been found from cyanobacteria (Burjaet al. 2001; Singhet al. 2011; Gerwick & Moore 2012), and already by the 1980s, many compounds had been described from cyanobacteria, including dolastatin and scytophycin (Pettitet al. 1981; Ishibashiet al. 1986). Cyanobacterial natural products are structurally variable, and include compounds such as peptides, polyketides, alkaloids, and ter- penes (Welker & von Döhren 2006; Joneset al. 2009; Dittmannet al. 2015; Yamadaet al.

2015). Modern genomic techniques support the conclusion that cyanobacteria are rich in natural products, and enable an increasing number of compounds to be identified (Calteau et al. 2014; Wanget al. 2014, 2015; Micallefet al. 2015). Especially marine cyanobacteria have been a target of interest (Burjaet al. 2001; Tan 2007; Nunneryet al. 2010; Bluntet al.

2017). One of the most extensively studied marine cyanobacterial species rich in natural products isMoorea(previouslyLyngbya sp.), which produces compounds such as curacin, jamaicamide, and apratoxin (Kleigrewe et al. 2016; Leaoet al. 2017). Fresh and brackish water species and terrestrial cyanobacteria produce bioactive molecules as well (Welker &

von Döhren 2006; Chlipalaet al. 2012). These include the trypsin inhibitors nostosin (Liuet al. 2014c), pseudospumigin (Jokelaet al. 2017), and suomilide (Fujiiet al. 1997); the cyto- toxic compounds (i.e. compounds toxic to cells) anabaenolysin (Jokela et al. 2012) and scytophycin (Ishibashiet al. 1986); and the antifungal compound hassallidin (Neuhofet al.

2005).

1.3 Bioactivity

The numerous natural products produced by cyanobacteria vary greatly in structure and therefore also in biological activity. Cyanobacterial natural compounds exhibit antibacterial, antiprotozoal, antiviral, antifungal, and cytotoxic activity (Burjaet al. 2001; Tan 2007; Singh et al. 2011; Swainet al. 2017). In addition, some have protease or ion channel inhibitory activity. However, certain natural products have no described activity. These observations have led to a debate regarding whether natural products confer a fitness advantage to the host (Firn & Jones 2000; Jensen 2016). There are likely several evolutionary paths creating the numerous natural products. The discussion can be further complicated by limited knowledge regarding the activity of natural products. The identified activity is only based on the bioactivity assay performed, and the more the compound is tested, the more activities may be revealed. For instance, anticancer compounds show commonly also antifungal activity, thus being toxic or inhibitory to all eukaryotic cells.

Many bioactive natural compounds from cyanobacteria have been tested in cell assays under specific conditions, and the ecological roles of the active compounds remain unknown. Surprisingly little is known about why cyanobacteria produce bioactive compounds.

However, some advantage for the producer is expected (Méjean & Ploux 2013). One hy-

(15)

components, indicating that the production of this cytotoxic compound is a defense mechanism against co-occurring fungal species (Patterson & Bolis 1997). In addition, cyanobacteria have been found to inhibit the growth of other cyanobacteria, micro- and macro-algae and some plants, suggesting that certain compounds may have allelopathic activity (Leãoet al. 2009). Cyanobacteria live in numerous environments in complex food webs with other organisms, including fungi and viruses, which could indicate that cyanobacteria need defense mechanisms but also molecules for communication between organisms. Cyanobacte- ria are often found in symbiosis, for example, with lichens and marine sponges. However, only photoprotective agents such as scytonemin have thus far been shown to have a clear role in symbiosis, protecting cyanobacteria and their symbionts from harmful UV light (Garcia-Pichel & Castenholz 1991; Nguyenet al. 2013a).

Regardless of their ecological functions, the diverse and potential activities of cyanobacterial natural products have caused them to gain interest as drug candidates. The following chapters provide examples of cyanobacterial anticancer and antifungal natural products.

1.3.1 Anticancer activity

New anticancer drugs are needed due to the increasing prevalence of cancer and the diversity of the cancers detected. Furthermore, tumor cells are developing resistance against drugs, similar to bacteria against antibiotics. In many cases, cancer treatment involves heavy chemotherapy, and there is a pressing need for new drugs. A number of cyanobacterial natural products show anticancer or cytotoxic activity (see examples in Figure 2 and Table 1).

Some cyanobacterial anticancer compounds or analogues of them have entered clinical trials but only one compound has thus far made it to the market. In 2011, FDA approved Adce- tris® (Brentuximab vedotin, Seattle Genetics®) for the treatment of Hodgkin's lymphoma and systemic anaplastic large cell lymphoma (with certain regulations) (Denget al. 2013).

In 2012, the drug was also approved in Europe, and in 2018, the spectrum of treated Hodg- kin's lymphomas was expanded. The antibody-drug conjugate Adcetris® consists of an anti- CD30 monoclonal antibody and the cytotoxic monomethyl auristatin E, which is a synthetic analogue of dolastatin 10 (Figure 2) (Denget al. 2013). Dolastatins were first described from the sea hareDolabella auricularia (Pettitet al. 1981, 1987) but the actual producer was later found to be the cyanobacteriumSymploca sp. (Lueschet al. 2001a). Dolastatins show activity against several cancer cell lines, the variants dolastatin 10 and 15 showing the highest potential (Pettitet al. 1987, 1989). Several variants and analogues of dolastatins have been tested and entered clinical trials but the monomethyls auristatin E and F seem to be the most potent drug candidates, and there are several ongoing studies on these variants as antibody-drug conjugates against different cancer types (Mayer 2018).

Another anticancer peptide from cyanobacteria is apratoxin A (Figure 2) (Lueschet al.

2001b). It was isolated fromLyngbya majuscula and shows activity against several solid tumor cell lines (Lueschet al. 2001b). In addition, other apratoxin variants show cytotoxic activity (Lueschet al. 2002; Gutiérrez et al. 2008). ManyMoorea sp. (previously named Lyngbya sp.) have been found to produce bioactive compounds (Engeneet al. 2012; Leao

(16)

et al. 2017), such as curacin A (Figure 2) (Gerwicket al. 1994). Curacin A shows antimitotic activity but is also a potent brine shrimp toxin (Gerwicket al. 1994). However, the development of curacin A into a drug has been hindered by its insolubility.

The macrolide compounds cryptophycins (Figure 2) were found fromNostoc sp. GSV 224 (Schwartzet al. 1990; Trimurtuluet al. 1994). Cryptophycin A shows antifungal activity and is active against several cancer cell lines at the picomolar range (Schwartzet al. 1990;

Smithet al. 1994; Trimurtuluet al. 1994). Synthetic variants of cryptophycin have been produced and extensively studied, with cryptophycin 52 being a particularly promising drug candidate (Edelmanet al. 2003). Another group of cyanobacterial macrolide compounds is the scytophycins (Figure 2), which include several variants such as tolytoxin. Scytophycins have been isolated fromScytonema spp.,Cylindrospermum, andNostocsp. (Ishibashiet al.

1986; Carmeliet al. 1990; Junget al. 1991; Tomsickovaet al. 2014). Scytophycins show potent cytotoxic effects against human cancer cell lines as well as against fungi (Patterson &

Carmeli 1992). Scytophycins are structurally similar to swinholides, which are associated with symbiotic sponge species although their actual producer is unknown. However, the swinholide variants ankaraholides have been described from a field sample containing the cyanobacterium Geitlerinema sp. (Andrianasoloet al. 2005). Swinholides show cytotoxic and antifungal activity similar to scytophycins (Carmeli & Kashman 1985; Kobayashiet al.

1990).

Many of the anticancer compounds mentioned above were screened with the US Na- tional Cancer Institute (NCI) 60 human tumor cell line anticancer drug screen (Shoemaker 2006). This panel includes a few leukemia cell lines but no acute myeloid leukemia (AML) cell lines. AML treatment is also in urgent need of new drugs as the drugs currently used to treat acute myeloid leukemia were developed half a century ago (Rowe 2013). Not many screening studies have been carried out with cyanobacteria and AML cells. However, a potential for apoptosis (i.e. programmed cell death typically involving e.g. blebbing or nuclear fragmentation) inducing activity in AML cells has been detected in compounds produced by cyanobacterial strains from certain locations and habitats including the Baltic Sea and lichen associated cyanobacteria (Herfindalet al. 2005; Oftedalet al. 2010, 2011; Liuet al. 2014a).

(17)

Figure 2. Chemical structures of some anticancer and/or cytotoxic natural products produced by cyanobacteria.

(18)

Table 1. Examples of anticancer (or cytotoxic) natural products found from cyanobacteria.Most of the example compounds have several variants, often separated by different letters.

Compound Chemical structure Cyanobacterial taxa Reference Almiramide Linear lipopeptide Lyngbya majuscula Sanchezet al.2010 Ankaraholide Glycosylated polyketide Geitlerinemasp. Andrianasoloet al.2005 Apratoxin Cyclodepsipeptide (mixed NRPS-

PKS)

Lyngbyasp. Lueschet al.2001b

Aurilide Cyclodepsipeptide (mixed NRPS- PKS)

Lyngbya majuscula Suenagaet al.1996; Han et al. 2006

Belamide Linear tetrapeptide Symploca sp. Simmonset al.2006

Bisebromoamide Linear heptapeptide Lyngbyasp. Teruyaet al.2009 Calothrixin Indole alkaloid Calothrixsp. Rickardset al.1999 Carmaphysin Linear tripeptide Symplocasp. Pereiraet al.2012 Caylobolide Macrolactone polyketide Lyngbya majuscula MacMillan & Molinski

2002

Coibamide Cyclodepsipeptide Leptolyngbyasp. Medinaet al.2008 Cryptophycin Macrolactone (mixed NRPS-PKS) Nostocsp. Schwartzet al.1990 Curacin Linear peptide (mixed NRPS-PKS) Lyngbya majuscula Gerwicket al.1994 Dolastatin Linear peptidic compounds Symplocasp. Pettitet al.1981; Luesch

et al. 2001a

Dragonamide Linear lipotetrapeptide Lyngbyaspp. Jiménez & Scheuer 2001;

Gunasekeraet al. 2008 Hectochlorin Cyclic mixed NRPS-PKS Lyngbya majuscula Marquezet al.2002 Hoiamide Cyclic lipodepsipeptide (mixed

NRPS-PKS)

L. majuscula, Phormidium gracile

Pereiraet al.2009 Jamaicamide Linear lipopeptide (mixed NRPS-

PKS)

Lyngbya majuscula Edwardset al.2004 Lagunamide Cyclodepsipeptide Lyngbya majuscula Tripathiet al.2010 Largazole Cyclic lipodepsipeptide Symplocasp. Taoriet al.2008 Lyngbyabellin Cyclic lipodepsipeptide Lyngbya majuscula Lueschet al.2000 Malevamide Cyclic lipodepsipeptide, linear pep-

tides

Symploca laete-viridis Horgenet al.2000 Malyngamide,

isomalyngamide

Small linear amides Lyngbya majuscula Mynderse & Moore 1978;

Changet al. 2011 Minutissamide Cyclic lipodecapeptide Anabaena minutissima Kanget al.2011 Obyanamide Cyclodepsipeptide Lyngbya confervoides Williamset al.2002 Palmyramide Cyclodepsipeptide Lyngbya majuscula Taniguchiet al.2010 Pseudodysidenin Linear depsipeptide Lyngbya majuscula Jiménez & Scheuer 2001 Santacruzamate Small linear diamides Symplocasp. Pavliket al.2013 Scytophycin Macrolactone (mixed NRPS-PKS) Scytonema sp. Ishibashiet al.1986 Somocystinamide Linear mixed NRPS-PKS compound Lyngbya majuscula Nogle & Gerwick 2002 Symplocamide Ahp-cyclodepsipeptide Symplocasp. Liningtonet al.2008 Tubercidin Purine ribonucleoside Tolypothrix byssoidea Barchiet al.1983 Veraguamide Cyclic lipodepsipeptide Oscillatoria margaritifera Meverset al.2011

(19)

1.3.2 Antifungal activity

A variety of cyanobacteria produce natural products that exhibit antifungal activity (Abed et al. 2009; Chlipalaet al. 2012; Swainet al. 2017). The chemical structures of these com- pounds are variable, including alkaloids, aromatic compounds, polyketides and different kinds of peptides. Examples of specifically antifungal natural compounds include fischerin- dole (Park et al. 1992), laxaphycin (Frankmölle et al. 1992), and welwitindolinone (Stratmannet al. 1994b). Furthermore, there are even more compounds that exhibit antifungal activity along with other activities. There are natural compounds showing, for instance, antifungal and cytotoxic activity against mammalian cells, such as scytophycins and swinholides (Carmeli & Kashman 1985; Kobayashiet al. 1990; Patterson & Carmeli 1992), or against other cyanobacteria, such as fischerellin (Grosset al. 1991).

One interesting group of compounds is the glycolipopeptides hassallidins, which consist of a peptide ring of eight amino acids with one additional amino acid, fatty acid and one to three sugar moieties (Tables 2, 3 and 4). Hassallidin A, the first variant described, was isolated from the terrestrial epilithic cyanobacteriumTolypothrix sp. (basionymHassallia) (Neuhofet al. 2005). Hassallidin A has only one sugar in its structure. It was shown to exhibit a minimum inhibitory concentration (MIC) of 4.8 μg ml^-1 (3 μM) againstCandida albicans andAspergillus fumigatus(Neuhofet al. 2005). In addition, clear inhibition zones were seen in plate diffusion assays against other fungi, including Aspergillus niger,Candida glabrata, Fusarium sambucium, Penicilliumsp., andUstilago maydis (Neuhofet al. 2005). Hassallidin B (isolated from the same strain), in turn, has two carbohydrate units, which have little ef- fect on the activity of the compound (Neuhofet al. 2006b). The MIC values against several strains of Candida sp. and Cryptococcus neoformans were shown to range from 8 to 16 μg ml^-1 (5 to 10 μM) for hassallidin B and from 4 to 16 μg ml^-1 (3 to 12 μM) for hassallidin A (Neuhofet al. 2006b). Hassallidin A and B were subsequently patented, and additional antifungal testing was carried out in addition to cytotoxicity assays (Neuhofet al. 2006a).

Hassallidin B showed a half-maximal inhibitory concentration (IC50) of 0.2 μg ml^-1 (0.1 μM) against the Jurkat ATCC-TIB-152 (human acute T cell leukemia) cell line, and both hassallidin A and B showed high activity against L 929 (murine aneuploid fibrosarcoma) cells (resazurin assay) (Neuhofet al. 2006a).

The biosynthesis genes for hassallidin production were described from the Baltic Sea isolateDolichospermum sp. UHCC 0090 (previouslyAnabaenasp. 90), which led to the discovery of many new producer genera and locations (Table 3) (Wanget al. 2012; Vestolaet al. 2014). Furthermore, the main variants hassallidin C and D were described and shown to have similar antifungal activity (Vestola et al. 2014). For hassallidin D, an MIC value of 2.8 μg ml^-1 (2 μM) and IC50 values of 0.55 to 1.86 μg ml^-1 (0.3 to 1 μM) were measured against differentC. albicansstrains (Vestolaet al. 2014). The most recently described main variant is hassallidin E, which was isolated from the nonheterocystous cyanobacterium Planktothrix serta PCC 8927 (Pancraceet al. 2017b). By then, hassallidins and their variants balticidins had been found only from heterocystous species such asAnabaena/Dolichosper- mum,Aphanizomenon,Cylindrospermopsis,NostocandTolypothrix spp. (Table 3) (Neuhof et al. 2005; Buiet al. 2014; Vestolaet al. 2014). In addition, many other variants had been

(20)

detected but not described as main variant categories (Vestolaet al. 2014). Interestingly, hassallidins are now being increasingly found from across the world due to expanding genomic data (Abreuet al. 2018). The terms hassallidin or balticidin are used for these glycolipopeptides when found from cyanobacteria but structurally identical or highly similar compounds are found with inconsistent naming from other bacteria (Table 4), including herbicolin (Aydin et al. 1985), jagaricin (Graupner et al. 2012), chromobactomycin (Kimet al.

2014), and Sch 20561 and 20562 (Afonsoet al. 1999a,b). They exhibit antifungal activities highly similar to hassallidins. Moreover, similar to hassallidins, no antibacterial activity has been shown for these compounds (Neuhofet al. 2005), with the exception of herbicolin A, which displays activity against mycobacteria (Freundt & Winkelmann 1984).

Cyanobacteria produce surprisingly many peptides with lipid moieties of varying lengths attached to them (Burjaet al. 2001). Many of these lipopeptides have cytotoxic activities. For instance, the linear lipopeptide malyngamide exhibits cytotoxic activity (Appletonet al. 2002; Sabryet al. 2017; Jianget al. 2018) and dragonamide shows antipar- asitic activity (Balunaset al. 2010). Furthermore, besides the hassallidins, cyanobacteria also produce other cyclic lipopeptides (examples in Table 5). Anabaenolysins (Jokelaet al.

2012), calophycins (Moonet al. 1992), laxaphycins (Frankmölleet al. 1992), lobocyclamides (MacMillanet al. 2002), lyngbyacyclamides (Maruet al. 2010), muscotoxins (Tomeket al.

2015), and puwainaphycins (Hrouzeket al. 2012) are all examples of cyclic lipopeptides ex- tracted from cyanobacteria and exhibiting cytotoxic or antifungal activities. For example, anabaenolysins comprise a group of compounds found uniquely from benthicAnabaena strains in coastal areas (Gulf of Finland, Baltic Sea) (Jokelaet al. 2012; Shishidoet al. 2015).

(21)

Table 2. Hassallidin family glycolipodepsinonapeptides. Ac, acetyl; Dhb, dehydrobutyric acid; NMe, N-methyl * The balticidin amino acids 7 and 10 are likely in opposite order as both chemical and biosynthesis gene cluster evidence show that D-aa is in position 7. ** The sequence of amino acids 5 to 10 was changed to match hassallidin structure after reanalyzing HMBC data.

Fatty acidAmino acidsMonosaccharides Name12345678910M1 M2M3[M+H]+ Reference Hassallidin A2,3-OH-C14Thr ThrThrTyrDhb Gln GlyNMeThrGln--Man1382,69 Neuhofet al. 2005 Hassallidin B2,3-OH-C14Thr ThrThrTyrDhb Gln GlyNMeThrGln-RhaMan1528,75 Neuhofet al. 2006b Hassallidin C2,3-OH-C16L-Thr Thr D-allo-ThrD-Tyr Dhb D-GlnGlyNMeThrLysGlcNAc AradiAcMan 1829,90 Vestolaet al. 2014 Hassallidin D2,3-OH-C16L-Thr Thr D-allo-ThrD-Tyr Dhb D-GlnGlyNMeThrL-TyrGlcNAc AradiAcMan 1864,87 Vestolaet al. 2014 Hassallidin E2,3-OH-C16Thr ThrThrTyrDhb Gln GlyNMeThrGln--Hex1410,73 Pancraceet al. 2017b Balticidin A*2,3-OH-13-Cl-C16Thr ThrThr Tyr(3-OH)Dhb L-GlnGlyNMeThrD-Gln-OHGalA AraMan1786.77 Buiet al. 2014 Balticidin B2,3-OH-13-Cl-C16Thr ThrThr Tyr(3-OH)Dhb L-GlnGlyNMeThrD-GlnGalA AraMan1768,76 Buiet al. 2014 Balticidin C2,3-OH-C16Thr ThrThr Tyr(3-OH)Dhb L-GlnGlyNMeThrD-Gln-OHGalA AraMan1752.80 Buiet al. 2014 Balticidin D2,3-OH-C16Thr ThrThr Tyr(3-OH)Dhb L-GlnGlyNMeThrD-GlnGalA AraMan1734,79 Buiet al. 2014 Sch 20561D-3-OH-C14 E-Dhb L-ThrD-allo-ThrD-Tyr E-DhbD-GlnGlyL-NMe-allo-ThrL-His---1195,64 Afonsoet al. 1999b Sch 20562D-3-OH-C14 E-Dhb L-ThrD-allo-ThrD-Tyr E-DhbD-GlnGlyL-NMe-allo-ThrL-His---D-Glc1357,69 Afonsoet al. 1999a Herbicolin A(R)-3-OH-C14 DL-DhbL-ThrD-allo-ThrD-Leu Gly D-GlnGlyL-NMe-allo-ThrL-Arg-D-Glc-1300,74 Aydinet al. 1985 Herbicolin B(R)-3-OH-C14 DL-DhbL-ThrD-allo-ThrD-Leu Gly D-GlnGlyL-NMe-allo-ThrL-Arg---1138,68 Aydinet al. 1985 Jagaricin(R)-3-OH-C14Dhb L-ThrD-allo-ThrD-Tyr Dhb D-GlnGlyL-allo-ThrL-His---1181,62 Graupner et al. 2012 Chromobactomycin3-OH-C14Dhb ThrThrTyrDhb Gln GlyNMeThrHis---1195,64 Kimet al. 2014**

(22)

rains producing hassallidins A to E, balticidins or other variants.The variants are listed to provide an overview of ical or related to main variants). In addition, the identification of production through genes or chemical analysis is shown (+, positive is; -, no variants found; NA, data not available). OriginVariants(identicalor relatives)GenesChemical analysis Reference ermum sp. Lake Vesijärvi, FinlandHassallidin C, D, others+*+Vestolaet al. 2014 Lake Tuusulanjärvi, FinlandHassallidin C, D, others++Vestolaet al. 2014 Lake Vesijärvi, Finland+-Vestolaet al. 2014 Lake Vesijärvi, FinlandHassallidin C, D, others++Vestolaet al. 2014 Lake Tuusulanjärvi, FinlandHassallidin C, D, others++Vestolaet al. 2014 Lake Tuusulanjärvi, FinlandHassallidin C, D++Vestolaet al. 2014 Lake Tuusulanjärvi, FinlandHassallidin C, D++Vestolaet al. 2014 Lake Tuusulanjärvi, FinlandHassallidin C, others++Vestolaet al. 2014 Lake Tuusulanjärvi, FinlandHassallidin D++Vestolaet al. 2014 Lake Tuusulanjärvi, FinlandHassallidin C, B, others++Vestolaet al. 2014 The Gulf of Finland, FinlandOthersNA+This study The Gulf of Finland, Vuosaari, FinlandBalticidin B++Vestolaet al. 2014 Kobben, Hanko, Baltic Sea, FinlandOthersNA+This study Lake Frøylandsvatnet, NorwayHassallidin C, others++Vestolaet al. 2014 Lake Frøylandsvatnet, NorwayHassallidin C, others++Vestolaet al. 2014 Lake Knud, Denmark+-Vestolaet al. 2014 Lake Tuusulanjärvi, FinlandHassallidin C, D, others++Vestolaet al. 2014 Lake Tuusulanjärvi, FinlandHassallidin C, D++Vestolaet al. 2014 Lake Kotojärvi, FinlandHassallidin C, D, others++Vestolaet al. 2014 Porkkala Cape, Baltic Sea coast, FinlandBalticidin B++Vestolaet al. 2014 Porkkala Cape, Baltic Sea coast, FinlandBalticidin B, hassallidin C++Vestolaet al. 2014 Porkkala Cape, Baltic Sea coast, FinlandOthers++Vestolaet al. 2014 Porkkala Cape, Baltic Sea coast, FinlandHassallidin C, others++Vestolaet al. 2014

(23)

Strain and nameOriginVariants(identicalor relatives)GenesChemical analysis Reference Bio33Baltic Sea, Rügen Island, GermanyBalticidin A, B, C, DNA+Buiet al. 2014 Aphanizomenon gracile Heaney/Camb 1986 140 1/1Freshwater, Lough Neagh, IrelandBalticidin B++Vestolaet al. 2014 Cylinrdrospermopsis raciborscii ATC-9502Lake Balaton, HungaryOthers++Vestolaet al. 2014 CENA 303Theobaldo Dick Lake, Lajeado, RS, Brazil+NAAbreuet al. 2018 CS-505Freshwater, Solomon Dam, Q, AustraliaOthers++Vestolaet al. 2014 CS-508Pond, Townsville, NQ, Australia+NAAbreuet al. 2018 CR12Freshwater lake, Singapore+NAAbreuet al. 2018 Nostocsp. 6 sf CalcDobre Pole, Czech RepublicOthersNA+This study 113.5Lichen associated, FinlandHassallidin C, balticidin B, others++Vestolaet al. 2014 UHCC 0159 (159)Lake Haukkajärvi, FinlandBalticidin B++Vestolaet al. 2014 CENA 219Morro Branco, CE, BrazilOthersNA+This study Planktothrix serta PCC 8927Sewage plant, Berre-le-Clos, FranceHassallidin E++Pancraceet al. 2017b Tolypothrix sp. PCC 9009Watkins Glen State Park, NY, USABalticidin B, Others++Vestolaet al. 2014 PCC 7101Soil, Borneo+-Vestolaet al. 2014 PCC 7504Aquarium, Stocholm, Sweden+-Vestolaet al. 2014 B02-07Epilithic, Bellano, ItalyHassallidin ANA+Neuhofet al. 2005 B02-07Epilithic, Bellano, ItalyHassallidin BNA+Neuhofet al.2006b *Mutation inhas genes has resulted in inability to produce hassallidin.

(24)

ther bacteria. Bacterial strainAssociated environmentGenes Chemical analysis Reference Chromobacterium sp. C61RhizobacteriaNA+Kimet al. 2014 Erwinia herbicolaA111Plant pathogensNA+Aydinet al. 1985 Erwinia herbicolaA112Plant pathogensNA+Aydinet al. 1986 Janthinobacterium agaricidamnosum Agaricussp. mushroom pathogenNA+Graupneret al. 2012 Aeromonas sp. W-10 NRRL B-11053Human pathogensNA+Afonsoet al. 1999a Aeromonas sp. W-10 NRRL B-11053Human pathogensNA+Afonsoet al. 1999b

(25)

Table 5. Examples of cyanobacterial cyclic lipopeptides.Most of the example compounds have several variants, often separated by different letters. CompoundActivityCyanobacteriaReference AnabaenolysinCytotoxicAnabaena sp.Jokelaet al. 2012; Oftedal et al. 2012 CalophycinAntifungalCalothrix fuscaMoonet al. 1992 HapalosinCytotoxicHapalosiphon welwitschiiStratmannet al. 1994a HassallidinCytotoxicTolypothrix sp.Neuhofet al. 2005 HoiamideCytotoxicLyngbya majuscula, Phormidium gracilePereiraet al. 2009 HormothamninCytotoxic, antimicrobialHormothamnion enteromorphoidesGerwicket al. 1992 Largamide HUnknownOscillatoria sp.Plaza & Bewley 2006 LargazoleAnticancerSymploca sp.Taoriet al. 2008 LaxaphycinAntifungalAnabaena laxaFrankmölleet al. 1992 LobocyclamideAntifungalLyngbya confervoidesMacMillanet al. 2002 LyngbyabellinCytotoxicLyngbya majusculaLueschet al. 2000 LyngbyacyclamideAnticancerLyngbya sp.Maruet al. 2010 LyngbyazothrinAntibacterialLyngbya sp.Zainuddinet al. 2009 MalevamideCytotoxicSymploca laete-viridisHorgenet al. 2000 MinutissamideAnticancerAnabaena minutissimaKanget al. 2011 MuscotoxinCytotoxicDesmonostoc muscorumTomeket al. 2015 NostofungicidineAntifungalNostoc communeKajiyamaet al. 1998 NostopeptolideAntitoxinNostocsp.Golakotiet al. 2000; Liuet al. 2014b PahayokolideCytotoxic, antimicrobialLyngbya sp.Berryet al. 2004 PrecarriebowmideCytotoxicMoorea producensMeverset al. 2013 PuwainaphycinCytotoxicAnabaenasp.Mooreet al. 1989; Gregsonet al. 1992 Schizotrin AAntimicrobialSchizotrix sp.Pergament & Carmeli 1994

(26)

1.4 Biosynthesis

A biosynthetic gene cluster is a group of genes that encode the production of a natural product and are often genomically located close to each other (hence the word cluster).

However, a biosynthetic gene cluster does not directly encode a natural product. Instead, the product is post-translationally modified or the genes encode enzymes that build the actual natural products. Cyanobacterial natural products are produced through ribosomal or nonribosomal biosynthesis pathways (Kehr et al. 2011). Nonribosomal peptides and polyketides are biosynthesized by nonribosomal peptide synthetases (NRPS), polyketide synthases (PKS) or hybrid pathways (NRPS-PKS). These pathways constitute the majority of identified cyanobacterial natural product biosynthetic pathways (Dittmann et al. 2015;

Micallefet al. 2015). They are composed of multi-domain enzyme complexes that build the natural products from small precursor molecules into functional compounds by chain elongation (Fischbach & Walsh 2006). These pathways consist of modules, where each module produces one component to the growing molecule and moves it to the next module. The modules consist of smaller units called domains that are the enzymes modifying and moving the compound from one enzyme to another. Ribosomally synthetized peptides (RiPPs) are cleaved as precursor peptides, and the core peptide is modified post-translationally (Arnisonet al. 2013). For instance, cyanobactins is a large family of cyanobacterial RiPPs (Sivonenet al. 2010). The biosynthetic gene clusters are often large and contain repetitive parts, and are therefore more prone to changes than the primary metabolites of the genome, which might explain their tremendous diversity (Medemaet al. 2014; Ziemertet al.

2016). They are enriched in insertions, deletions and duplications, and horizontal gene transfer plays an important role in their evolution (Medemaet al. 2014; Ziemertet al. 2016).

Identification of natural product biosynthesis gene clusters is conceptually akin to genome mining, as the classical definition of genome mining is the search for specific biosynthetic genes for the production of a certain compound. The biosynthesis clusters found from cyanobacteria, which have genome sizes ranging from 1.4 to 12 Mb (e.g.Dolichospermum sp. UHCC 0090 has a 5.3 Mb genome (Wang et al. 2012)), vary in size from 2 to 58 kb (Méjean & Ploux 2013; Dittmannet al. 2015). The number of biosynthesis gene clusters correlates with genome size, such that cyanobacteria with smaller genomes may have less biosynthesis clusters (Shihet al. 2013). For instance,Cylindrospermum raciborskii genomes, which are the smallest known (approx. 3 Mb) among multicellular cyanobacteria, were found to have only a few biosynthesis gene clusters (Abreuet al. 2018).

Rapidly increasing sequencing capacity and decreasing costs have enabled a huge in- crease in the availability of whole genomes in public databases and a preference for whole genome sequencing over targeting only biosynthesis genes. This has also lead to the development of multiple tools for the annotation of genomes or genes. The analysis of genetic

(27)

National Center for Biotechnology Information (NCBI) database of the U.S. National Library of Medicine. They also provide tools such as the Basic Local Alignment Search Tool (BLAST).

One of the most well-known tools in natural product research today is antiSMASH that can be used to identify, annotate and analyze biosynthesis gene clusters from genomes (Medemaet al. 2011; Blinet al. 2017). The development of genomic methods and tools has provided a major boost to natural product discovery. Genome mining is used to find, for instance, PKS, NRPS, RiPPs and terpenoid biosynthesis clusters using their conserved structures, but specific enzymes can also be used as search criteria. Nevertheless, results from genome mining are often limited to known biosynthesis gene cluster types, and it is still difficult to identify the genes responsible for the production of new structural classes. Het- erologous expression or knockout mutants are used to confirm that the predicted genes are responsible for producing the natural product. Genome mining and biosynthesis gene cluster identification often still depend on cultured strains, as metagenomes rarely produce the full biosynthesis gene cluster. Improvements in metagenomic analyses and methods such as sequence tagging and single-cell genomics are attempting to overcome this problem.

1.4.1 Polyketide synthesis

Polyketides include compounds such as macrolides, polyphenols and polyenes. For example, macrolides consist of a macrocyclic lactone ring with additional side chains. Cyanobac- terial macrolides include compounds such as cryptophycin (Trimurtuluet al. 1994), lyngbyabellin (Lueschet al. 2000), and scytophycin (Ishibashiet al. 1986). Polyketide biosynthesis by a modular polyketide synthase resembles fatty acid synthesis (Figure 3). The common precursors for PKS include acetyl-CoA, malonyl-CoA, and propionyl-CoA (Hertweck 2009).

The minimal PKS module consists of an acyltransferase (AT), an acyl carrier protein (ACP), and ketosynthase (KS) domains (Hertweck 2009; Kehret al. 2011). The acyltransferase acti- vates the acyl carrier protein and selects the precursor molecule. The acyl carrier protein moves the precursor into contact with ketosynthase, which is needed to produce the actual chain lengthening Claisen-condensation reaction. The chain-lengthening reaction is possibly followed by the action of modifying domains such as ketoreductase (KR), dehydratase (DH), and enoyl reductase (ER). PKSs are commonly divided into three (I–III) subgroups (Hertweck 2009; Helfrichet al. 2014). Cyanobacterial PKSs belong into modular type I, which includes cis- andtrans-AT PKSs. Thecis-AT PKSs follow common co-linearity rules of biosynthesis pathways and they have iterative acyltransferases (AT) in each module. In contrast, the trans-AT PKSs have non-iterative acyltransferases, such that they use, for instance, only one AT multiple times (Hertweck 2009; Piel 2010). Other functions may also be provided itera- tively intrans-AT gene clusters, and they may have unusual domain orders, unique domains, non-elongating modules, or split modules (Piel 2010; Helfrich & Piel 2016). Interestingly, it has been proposed thatcis- andtrans-PKSs evolved independently from each other (Nguyen et al. 2008). Type II and III PKSs work in a completely iterative manner, and the products they typically synthetize are aromatic compounds (Hertweck 2009).

(28)

Figure 3. Modular polyketide synthase builds polyketides by a chain-elongation system.The loading module inserts the precursor (here, propionyl-CoA) and the following modules elongate the molecule.

Additional enzymes, such as ketoreductase here, modify the growing molecule. Thioesterase often terminates chain lengthening and may cyclize the compound. AT, acyltransferase; KR, ketoreductase;

KS, ketosynthase; TE, thioesterase; •, acyl carrier protein.

Despite the large amount of polyketides found from cyanobacteria, only a few biosynthesis gene clusters and their products have been identified (Wang et al. 2014; Dittmannet al.

2015). The cyanobacterial neurotoxins anatoxin-a and homoanatoxin-a are produced by type Icis-AT PKSs (Méjeanet al. 2009, 2014; Rantala-Ylinenet al. 2011). The firsttrans-AT PKS gene cluster from cyanobacteria was identified from the symbioticNostoc sp. `Peltigera membranacea cymbiont` producing nosperin (Kampa et al. 2013). Furthermore, the to- lytoxin gene cluster from Scytonema sp. PCC 10023 and luminaolide gene cluster from Planktothrix paucivesiculata PCC 9631 are produced bytrans-AT PKSs (Ueokaet al. 2015).

Recently, the biosynthesis gene clusters for phormidolide from Leptolyngbya sp. and nosperin-related cusperin fromCuspidothrix issatsechenkoi were shown to be produced by trans-AT PKSs (Bertinet al. 2016; Kustet al. 2018).

1.4.2 Nonribosomal peptide synthesis and hybrid synthesis

Nonribosomal peptides and peptide-like compounds are biosynthesized through nonribosomal peptide synthetases (NRPS) or hybrid pathways of NRPS and PKS. Each NRPS module adds a single amino acid to the growing amino acid chain (Fischbach & Walsh 2006; Kehret al. 2011; Marahiel 2016). The NRPS module needs an adenylation (A) domain for the acti- vation and identification of the amino acid, a peptidyl carrier protein (PCP) to move the

Antifungal and Antileukemic Compounds from Cyanobacteria : Bioactivity, Biosynthesis, and Mechanism of Action