Planta Med 2015; 81(15): 1309-1325
DOI: 10.1055/s-0035-1546055
Reviews
Georg Thieme Verlag KG Stuttgart · New York

Anti-infective Natural Products from Cyanobacteria

Timo Horst Johannes Niedermeyer
  • German Center for Infection Research (DZIF), Partner Site Tübingen and University of Tübingen, Interfaculty Institute for Microbiology and Infection Medicine, Tübingen, Germany
Further Information

Correspondence

Prof. Timo Niedermeyer
University of Tübingen, Interfaculty Institute for Microbiology and Infection Medicine
Auf der Morgenstelle 28
72076 Tübingen
Germany
Phone: +49 7 07 12 97 20 79   
Fax: +49 70 71 29 59 99   

Publication History

received 23 January 2015
revised 16 March 2015

accepted 13 April 2015

Publication Date:
17 June 2015 (eFirst)

 

Abstract

Cyanobacteria are a promising yet underexplored source for novel natural products with potent biological activities. While predominantly cytotoxic compounds have been isolated from cyanobacteria in the past, there are also a significant number of compounds known that possess anti-infective activities. As the need for novel anti-infective lead compounds is high, this manuscript aims at giving a concise overview on the current knowledge about anti-infective secondary metabolites isolated from cyanobacteria. Antibacterial, antifungal, antiviral, antiprotozoal, and molluscicidal activities are discussed. Covering up to February 2015.


#

Introduction

Anti-infective drugs have paved the way for modern medicine, and ended a world where the major cause of death was infectious diseases. They have been one of the important factors contributing to the rise in life expectancy in the past 75 years, and are surely the most important class of drugs in this regard [1], [2]. The majority of anti-infective drugs in current use are based on natural products [3]. However, pharmaceutical companies have neglected both the search for novel anti-infectives and natural product-based drug discovery over many years [4], resulting in the current situation that infections with antibiotic-resistant bacteria often cannot be treated adequately [5]. Especially infections with drug-resistant gram-negative bacteria such as Pseudomonas or Enterobacter are often a serious and life-threatening condition, and only a few novel drugs are in advanced development [6]. This problem is widely acknowledged today [7], [8], [9], and pharmaceutical and biotech companies as well as public institutions have reinforced or reinstalled research in this field. This is, for example, shown by the establishment of research centers such as the German Center for Infection Research (DZIF) and the Sanofi-Fraunhofer collaboration Center for Natural Product Research as well as by the acquisition of the antibiotics specialist company Cubist Pharmaceuticals by Merck & Co. in December 2014 and the “New Drugs for Bad Bugs” program of the Innovative Medicines Initiative (IMI) in which major pharmaceutical companies such as GlaxoSmithCline, Sanofi, AstraZeneca, Basilea, Janssen, and others as well as SMEs and public research institutes are involved.

The secondary metabolism of plants as well as microorganisms such as actinobacteria and fungi has been studied extensively over many decades, and many indispensable drug substances in many therapeutic areas have been approved that are based on compounds originally isolated from these organisms [3], [10], [11], [12]. In contrast, cyanobacteria have long been neglected by natural product scientists. Until the 1980s, they were mainly known for the toxins they produce [13], [14], [15]. In recent years, however, cyanobacteria have gained more attention. They are now recognized as a promising yet underexplored source for novel natural products with potent biological activities, and several reviews covering cyanobacterial metabolites have been published [16], [17], [18], [19], [20], [21], [22], [23]. The interest in secondary metabolites from cyanobacteria is rising ([Fig. 1]). While only about 200 cyanobacterial metabolites had been structurally characterized up until 1996 [24], this number has risen to about 1200 today [25]. However, despite the progress made in the past 15 years, this is still a small number of metabolites compared to the structures known from other microorganisms such as, e.g., the actinomycetes (> 9000 [25]).

Zoom Image
Fig. 1 SciFinder entries for “Natural product from cyanobacteri?” (09 Dec 2014).

Cyanobacteria, formerly known as blue-green algae and not recognized to be bacteria, are among the oldest organisms known and have been inhabiting the earth for more than three billion years [26]. They populate almost all habitats including extreme ones, and are highly diverse in terms of their morphology, physiology, and metabolism [27]. The synthesis of highly potent bioactive metabolites, in general, is one of the evolutionary strategies to cope with the dangers posed by planktivorous grazers or environmental rivals [28], [29].

Burja et al. noted in 2001 that a major portion of the approximately 200 marine cyanobacterial natural products described up until 1996 displayed cytotoxic activity [16]. Indeed, cytotoxicity is still today an often-observed bioactivity of cyanobacterial secondary metabolites. The most prominent examples for cytotoxic compounds from cyanobacteria are probably the cryptophycins [30], [31], [32] and the dolastatins [33], [34]. Especially the latter compounds are most interesting from a drug development point of view. Although the history of the discovery and development of the dolastatins to the drug substance Brentuximab vedotin, approved in 2011 for the treatment of patients with Hodgkinʼs lymphoma or with systemic anaplastic large cell lymphoma (ALCL), has been rather intricate [22], it can be said that this antibody drug conjugate is the first commercially available drug substance that is based on a cyanobacterial secondary metabolite.

Cyanobacterial secondary metabolites exhibit a high chemical diversity [35]. Even though compounds from many chemical classes have been isolated, peptide and polyketide structural elements are predominant among cyanobacterial metabolites [16], [21], [36]. The peptides comprise cyclic, branched, and linear structures as well as depsipeptides, lipopeptides, and peptides with uncommon modifications such as N- and O-methylation, sulfation, halogenation, glycosidation, oxidation, dehydration, heterocyclization, prenylation, ketide extensions, and others [21], [37]. Often, these compounds are synthesized via combined polyketide sythases and non-ribosomal peptide synthetases (PKS/NRPS) [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], resulting in a high prevalence of non-proteinogenic amino acids as building blocks of these compounds. Although often not recognized at first glance, ribosomally synthesized products play an important role among bioactive metabolites from cyanobacteria as well [48], [49], [50]. Peptidic structures, especially cyclic peptides, have been postulated as “privileged structures” for bioactivities, because they have a high probability of being able to mimic peptidic substrates or ligands of endogenous proteins such as enzymes or receptors [51], [52].

Although PKS and NRPS genes can be detected in all cyanobacteria orders, in particular the cyanobacteria from the orders Oscillatoriales and Nostocales seem to make extensive use of PKS/NRPS for natural product synthesis [21], [53]. The marine cyanobacterium Moorea producens (formerly often called Lyngbya majuscula [54]) is known for a very diverse product spectrum from both a chemical and a bioactivity point of view; more than 25 % of all secondary products known from cyanobacteria have been isolated from this species [16], [17], [21], [55], [56].

Often, several structural variants of one parent compound are found within one strain or related strains. For example, the compound family of the microcystins comprises more than 100 natural congeners [57], and about 25 natural microginins and 140 variants of the aeruginopeptin/micropeptin/cyanopeptolin/oscillapeptin/planktopeptin family are known [25]. This facilitates both the identification of structure-activity relationships (SAR) at early research stages and the semisynthetic modification of possible lead structures. The high natural variety within the compound families is due to the variability and flexibility of the various enzymes contained in the PKS/NRPS modules discussed above as well as transposition and recombination events of biosynthesis genes (“natural combinatorial biosynthesis”) [44], [58]. The already high diversity of these polyketide/peptides can be even more enhanced by biocombinatorial techniques [59], [60].

Cyanobacteria genomes are extraordinarily large, comprising almost 10 million base pairs in the case of Nostoc punctiforme PCC 73102. A large proportion of the genome seems to be dedicated to genes encoding the biosynthetic machinery for secondary metabolites [44], [61]. Given that cyanobacteria can be expected to have a biosynthetic capacity at least equal to that of other microorganisms such as myxobacteria [62], the potential of cyanobacteria for future drug discovery programs becomes clear. Indeed, the rate of rediscovery of already known compounds when working with cyanobacteria is significantly lower than for other better-studied organisms [35], [63]. The relative disregard of cyanobacteria in natural product research in the past, paired with the high chemical diversity of their secondary metabolites, makes them an attractive source of novel natural products today for pharmaceutical as well as other applications (e.g., agrochemicals, food, cosmetics).

While predominantly cytotoxic compounds have been isolated from cyanobacteria, there are also a significant number of compounds known that possess anti-infective activities. This manuscript aims at giving a concise overview on the current knowledge about anti-infective secondary metabolites isolated from cyanobacteria. Antibacterial, antifungal, antiviral, antiprotozoal, and molluscicidal activities will be covered.


#

Antibacterial Metabolites

Numerous compounds with antibacterial activities have been isolated from cyanobacteria. However, only a few metabolites with specific antimicrobial activity have been found. In many cases, antimicrobial activity is associated with general cytotoxicity, thus being of limited use for further development. Few compounds have been characterized in more detail (e.g., elucidation of the mode of action) after their initial description. To the best of my knowledge, no antibacterial secondary metabolite from cyanobacteria is currently under development for this indication.

Metabolites with direct antibacterial effects

In this section, only the most interesting compounds with a direct antibacterial effect will be discussed; these are mainly compounds that have at least some specificity for antibiotic effects. A concise summary including the reported potencies of all reported compounds with antibacterial activity can be found in [Table 1], the compounds discussed in the following are shown in [Fig. 2].

Zoom Image
Fig. 2 Selected cyanobacterial metabolites with antibacterial activity.

Table 1 Antibacterial and antimycobacterial compounds from cyanobacteria. Activity data are given as MIC (µM) or inhibition zone in an agar diffusion assay (mm inhibition zone at the amount of the compound applied to a disk).

Compound

From

Structure type

Active against

MIC/Inhibition zone

Additional bioactivity/comments

Ref.

Aeruginazole A

Microcystis sp.

cyclic peptide

B. subtilis

2.2 µM

inactive against E. coli and S. albus; also antifungal (MIC S. cerevisiae 43 µM); cytotoxic (IC50 of 22–41 µM); no protease inhibition at 40 µM

[88], [89]

Ambigol A–C

Fischerella ambigua

polychlorinated polyaromatic phenols

B. megaterium

8 mm @ 100 nmol

inactive against gram-negatives; inhibition of cyclooxygenase and HIV reverse transcriptase; molluscicidal; cytotoxic (IC50 about 75 µM); antialgal; antitrypanosomal

[90], [91]

Ambiguine isonitrils

Fischerella sp.

alkaloids

E. coli S. aureus B. subtilis C. albicans

6 µM 0.2 µM 0.8 µM 1.0 µM

activity data of congener I; compounds are also antifungal; most congeners also moderately cytotoxic (IC50 around 50–150 µM)

[76], [77]

Bromoanaindolone

Anabaena constricta

alkaloid

B. cereus

530 µM

anticyanobacterial; cytotoxicity not tested

[92]

Carbamidocyclophanes

Nostoc sp.

paracyclophanes

M. tuberculosis S. aureus E. faecalis S. pneumoniae

0.8–5.4 µM 0.1–100 µM 0.2–1.1 µM 0.2–2 µM

no activity against M. smegmatis, A. baumannii, E. coli, P. aeruginosa, K. pneumoniae; cytotoxic (IC50 of 0.5–12 µM); some congeners also antifungal

[93], [94], [95]

Carriebowlinol

yet unclassified

alkaloid

e.g. Vibrio sp.

< 1 µM

also antifungal (MIC < 0.5 µM)

[96]

Comnostins

Nostoc commune

diterpenes

B. cereus S. epidermidis E. coli

40–300 µM 10–80 µM 150–300 µM

also cytotoxic (EC50 of 1 µM) and molluscicidal (MIC of 50 µM)

[97]

Crossbyanol A–C

Leptolyngbya crossbyna

polybrominated polyaromatic phenols

S. aureus

3 µM

activity data of congener B; also brine shrimp (congener B; IC50 of 3 µM) and cytotoxicity (IC50 congener A 30 µM; congener B > 30 µM);

[98]

α-Dimorphecolic and coriolic acid

Oscillatoria redekei

unsaturated hydroxy fatty acids

B. subtilis M. flavus S. aureus

2–6 mm @ 200 nmol 0–9 mm @ 200 nmol 2–7 mm @ 200 nmol

activity comparable to that of linoleic acid

[99]

Eucapsitrione

Eucapsis sp.

anthraquinone

M. tuberculosis

3–6 µM

inactive against M. smegmatis, S. aureus, E. coli, and C. albicans at 55 µM, not cytotoxic at 28 µM

[83]

Hapalindoles

Hapalosiphon fontinalis; Fischerella sp., Westelliopsis sp.

indole alkaloids

S. aureus B. subtilis E. coli

36 mm 30 mm 31 mm

activity data of congener N; allelopathic activity; moderately cytotoxic (IC50 > 30 µM), antifungal, highly insecticidal

[68], [69], [70], [71], [72], [73], [74], [100], [101]

Kawaguchipeptins

Microcystis aeruginosa

cyclic peptides

S. aureus

0.7 µM

only tested against S. aureus

[102]

Lyngbyazothrins

Lyngbya sp.

cyclic peptides

B. subtilis E. coli

18 mm @ 16 µmol 18 mm @ 65 µmol

activity data of congeners C/D; antialgal; also inhibit 20S proteasome (IC50 of 7–19 µM)

[103], [104], [105]

Malyngamides

Lyngbya majuscula

fatty acid amides

S. aureus B. subtilis

?

cytotoxic (IC50 < 60 µM); feeding deterrents

[106], [107], [108]

Malyngolide

Lyngbya majuscula

fatty acid/δ-lactone

M. smegmatis S. pyogenes S. aureus B. subtilis

?

also interferes with quorum sensing; feeding deterrent

[64], [108]

Muscoride A

Nostoc muscorum

peptide alkaloid

B. subtilis

3–6 mm

amount on disc not given; inactive against E. coli

[109], [110]

Norabietanes

Microcoleous lacustris

diterpenes

S. aureus S. epidermidis S. typhi V. cholerae

45 µM 55 µM 150 µM 850 µM

not active against B. subtilis/cereus, E. coli, K. pneumoniae; cytotoxicity not tested

[111]

Nostocyclyne A

Nostoc sp.

polyketide

S. aureus B. subtilis

MIC @ 36 nmol MIC @ 30 nmol

no activity against Staphylococcus albus and E. coli; weak photosynthesis inhibition

[112]

Noscomin

Nostoc commune

diterpene

B. cereus S. epidermidis E. coli

75 µM 18 µM 300 µM

cytotoxicity not tested

[97]

Nostocarboline dimers

synthetic

alkaloid

S. aureus E. faecium S. pneumonia H. influenzae E. coli A. baumannii P. aeruginosa

0.35–0.7 µM 2.8 µM 5.7 µM 11.4 µM 11.4 µM 22.8 µM 45.6 µM

activity data of compound NCD9; synthetic derivatives based on nostocarboline from Nostoc; inactive against S. cerevisiae and C. albicans

Pitipeptolides A–F

Lyngbya majuscula/Moorea producens

cyclic depsipeptides

M. tuberculosis

0–30 mm @ 60 µmol

cytotoxic (IC50 11–100 µM), feeding deterrents

[113], [114], [115]

Schizotrin A

Schizotrix sp.

cyclic peptide

B. subtilis C. albicans

15 mm @ 7 nmol 7 mm @ 13 nmol

no activity against Gram-negatives; also antifungal; cytotoxicity not tested

[116]

Scytoscalarol

Scytonema sp.

sesterterpene

B. anthracis S. aureus E. coli M. tuberculosis

6 µM 2 µM 30 µM 110 µM

also antifungal and weakly cytotoxic (IC50 of 135 µM)

[81]

Unidentified

diverse genera

?

various microorganisms

[117], [118], [119], [120], [121]

The first antibiotic compound that has been described from a cyanobacterium is malyngolide (1), isolated in 1979 from a marine L. majuscula (today classified as M. producens) [64]. Subsequently, many total synthesis routes have been developed (e.g., [65], [66], [67]). Originally, the compound was described as being active against gram-positive microorganisms such as Mycobacterium smegmatis, Streptococcus pyogenes, Staphylococcus aureus, and Bacillus subtilis, and inactive against the gram-negative Salmonella enteritidis, Escherichia coli, and Pseudomonas aeruginosa. Recently, it has been found that malyngolide interferes with bacterial quorum sensing (see section “Compounds interfering with bacterial quorum sensing”).

The hapalindoles are a large compound family of related indole alkaloids that were first found in Hapalosiphon species [68], [69], [70], and were later also isolated from Fischerella [71], [72] and Westelliopsis [72]. Hapalindoles show activity against a range of microorganisms including S. aureus, B. subtilis, Salmonella gallinarum, and E. coli [73]. Interestingly, the various hapalindole congeners feature highly differing activity against a panel of microorganisms. The most potent compound, hapalindole A (2), shows MICs at sub-µM concentrations [72]. While having high antibacterial activity, they later have also been shown to be moderately cytotoxic [72] and highly insecticidal [74]. Their mode of action against eukaryotic cells is likely based on the modulation of sodium channels [75]. The ambiguine isonitrils are structurally related to the hapalindoles and have comparable bioactivity [76], [77].

Noscomin (3) and a structurally related metabolite isolated from Nostoc commune [78], [79] are terpenoid metabolites, a class of compounds that is rarely found to be produced by cyanobacteria. Noscomin was found to be active against Bacillus cereus, Staphylococcus epidermidis, and to a lesser extend against E. coli. It has not been described whether the compound has also been tested for other activities. Chemically closely related and of similar bioactivity are the comnostins isolated from the same Nostoc strain [80]. Another antibacterial terpenoid is the guanidine-sesterterpene scytoscalarol (4) from Scytonema sp. [81], again being more active against the gram-positive Bacillus anthracis and S. aureus than against E. coli. The compound also showed weak activity against Mycobacterium tuberculosis. Based on in silico docking studies, its mechanism of action against the latter microorganism has been proposed to be the inhibition of the arabinosyltransferase Mtb EmbC [82].

Eucapsitrione (5) is an anthraquinone derivative isolated from Eucapsis sp. [83]. It was found to selectively inhibit M. tuberculosis in assays testing both the activity against fast growing and nonreplicating persistent states. The compound was inactive against M. smegmatis, S. aureus, E. coli, and the yeast Candida albicans, and also not cytotoxic in the concentrations tested.

Nostocarboline from a Nostoc species is a quaternary indole alkaloid [84]. It was found to inhibit the enzyme butyrylcholine esterase. First thought to be suitable as a lead for the treatment of Alzheimerʼs disease, it was soon realized that the compound or derivatives of it are also algicidal against photosynthetic organisms [85], and active against M. tuberculosis and the malaria parasite Plasmodium falciparum [86]. Synthetic studies on nostocarboline resulted in several nostocarboline dimers (e.g., NCD9, 6) showing activity against S. aureus, E. coli, or C. albicans [87].


#

Compounds interfering with bacterial quorum sensing

The term “quorum sensing” (QS) has been coined for inter- or intraspecies cell-to-cell communication by chemical signals in bacteria [122], [123], [124]. This mechanism enables bacteria to sense other species or their own population density, thus making it possible for individual bacteria in populations to coordinate their behavior. As, for example, biofilm formation and the production and secretion of many virulence factors are controlled by QS, inhibition of QS is currently being discussed as a novel and promising target for antibacterial therapy [125], [126], [127], [128]. Different QS systems using specific chemical signals have been described, but the best-studied system is based on N-acyl homoserine lactones as signaling compounds (AHLs, general structure 7, [Fig. 3]) [125].

Zoom Image
Fig. 3 Cyanobacterial metabolites with quorum sensing inhibiting activity.

Although, to date, none of the typical genes encoding AHL biosynthesis have been found in cyanobacteria, and research on quorum sensing in cyanobacteria is still in its infancy, AHL-like compounds have been isolated from two strains: The first report, the isolation of N-butyryl homoserine lactone from M. producens, dates back to the 1970s, when it was not yet even known that AHLs were involved in bacterial communication. It is not known whether this compound is involved in QS in this cyanobacterium. Years later when awareness of bacterial communication had risen, it was found that Gloeothece sp. produces N-octanoyl homoserine lactone; the compound seems to have QS activity in this strain and is involved in carbohydrate and amino acid metabolism regulatory processes [129]. AHLs have also been reported to affect nitrogen fixation in Anabaena sp. PCC 7120 [130], which perhaps is the reason that this strain is able to produce the enzyme AHL-acylase that deactivates AHLs by opening the lactone ring mandatory for activity, and thus “detoxifies” them [131].

A possibility to interfere with QS of other bacteria, and in this way possibly helping to outcompete them, is the production of compounds inhibiting QS signaling molecules at the receptor site.

The first cyanobacterial metabolites for which inhibition of QS was observed were the tumonoic acids, isolated from Blennothrix cantharidosum [132]. Suspected to be QS-active due to their distant structural similarity to the AHLs, tumonoic acid F (8) was found to be the most active compound among the isolated congeners. It inhibits wild-type Vibrio harveyi bioluminescence with an IC50 of 62 µM, without affecting bacterial growth at this concentration. Interestingly, the compounds did not show cytotoxicity. However, the exact target of the compounds in the Vibrio QS machineries has not been elucidated.

Malyngolide (1), already mentioned above as an antibacterial compound, has been found in a directed screening for cyanobacteria extracts interfering with AHL-regulated violacein production of the reporter strain Chromobacterium violaceum CV017 [133]. The EC50 in this assay was determined to be 110 µM, while the compound had no effect on bacterial growth up to 220 µM. Compund 1 also blocked QS-dependent production of elastase by P. aeruginosa (EC50 of 10 µM). Subsequent studies on the mode of action of 1 suggested that the compound acts by blocking the expression of lasR (a homologue of the luxR gene encoding the AHL sensing receptor protein) but does not interfere with the AHL-binding domain of the respective protein.

Malyngamide C (9) was isolated in 1985 from L. majuscula (M. producens) [134]. In 2010, its 8-epi-isomer was isolated from the same species, and both compounds were found to reduce AHL signaling at concentrations not inhibiting bacterial growth (IC50 of about 1 mM). However, the compounds were cytotoxic at far lower concentrations (IC50 of about 10 µM).

The three QS-inhibiting compounds discussed above all feature a rather lipophilic, fatty acid-like side chain. This also holds true for the last two compounds discussed in this section that even more closely resemble fatty acids. Lyngbyoic acid (10) from L. majuscula (M. producens) directly acts on the AHL receptor proteins, especially LasR from P. aeruginosa, reducing the expression of important virulence factors in a wild-type strain [135]. Nothing has been reported about cytotoxic properties on eukaryotic cells. Pitinoic acid A (11) inhibits LasB and pyocyanin production in P. aeruginosa with an IC50 between 0.1 and 1 mM. Again, nothing has been reported about cytotoxicity.


#
#

Antifungal Metabolites

All reported cyanobacterial compounds with antifungal activity are summarized in [Table 2]. As has already been noted above for the antibacterial compounds, many of the identified antifungal compounds also possess general cytotoxicity. The most interesting compounds will be discussed in the following paragraphs, there structures are shown in [Fig. 4].

Zoom Image
Fig. 4 Selected cyanobacterial metabolites with antifungal activity.

Table 2 Antifungal compounds from cyanobacteria. Activity data are given as MIC (µM) or inhibition zone in an agar diffusion assay (mm inhibition zone at the amount of the compound applied to a disk).

Compound

From

Structure type

Active against

MIC/Inhibition zone

Additional bioactivity/comments

Ref.

Ambiguine isonitrils

Fischerella sp.

alkaloids

C. albicans

1.0 µM

activity data of congener I; also antibacterial; moderately cytotoxic

[76], [77]

Balticidins

Anabaena cylindrica

glycosylated lipopeptide

C. maltosa

9–18 mm @ 6 nmol

no antibacterial activity against B. subtilis, E. coli, P. aeruginosa; cytotoxicity not tested

[140]

Calophycin

Calothrix fusca

cyclic peptide

A. oryzae C. albicans P. notatum S. cerevisiae T. mentagrophytes

13 mm @ 1 nmol 7 mm @ 1 nmol 12 mm @ 1 nmol 12 mm @ 1 nmol 15 mm @ 1 nmol

also cytotoxic (IC50 of 0.2 µM, KB cells)

[155]

Carriebowlinol

yet unclassified

alkaloid

Fusarium sp. L. thalassiae D. salina

0.2 µM 0.4 µM 0.5 µM

IC50 given; also antibacterial against several marine bacteria (MIC < 1 µM)

[96]

Fischerellin A

Fischerella muscicola

other

U. appendiculatus E. graminis

100 % inh. at 0.6 mmol 100 % inh. at 2.5 mmol

also antialgal and herbicidal effects

[161]

Hapalindoles

Hapalosiphon fontinalis

indole alkaloids

C. albicans

0.7 µM

activity data of congener J; also cytotoxic (IC50 of 12–44 µM), antibacterial, highly insecticidal

[72]

Hassallidins A and B

Hassallia sp.; widely spread among filamentous cyanobacteria

glycosylated lipopeptide

A. fumigatus C. albicans

3.5 µM

activity data of congener A; also active against Fusarium, Ustilago, Penicillium; not active against Bacillus subtilis, Streptomyces versicolor, E. coli; cytotoxicity 10-fold higher

[136], [137]

Laxaphycins

Anabaena laxa, Moorea producens, Anabaena torulosa

cyclic peptides

A. oryzae

20 µM

also active against C. albicans, P. notatum, S. cerevisiae, and T. mentagrophytes; synergistic effects between congeners A and B; cytotoxicity about 0.2 µM

[141], [142], [143], [144]

Lobocyclamide A–D

Lyngbya confervoides

cyclic lipopeptides

C. albicans

10 µM

MIC of a synergistic mixture of congeners A and B

[163]

Lyngbyabellin B

Lyngbya majuscula/Moorea producens

cyclic depsipeptide

C. albicans

10 mm @ 150 nmol

also brine shrimp toxicity (LD50 of 4.4 µM); not active against P. aeruginosa, E. coli, S. cholerae-suis, B. subtilis, S. aureus

[158]

Majusculamide C

Lyngbya majuscula/Moorea producens

cyclic depsipeptide

R. solani P. aphanidermatum A. euteiches P. infestans

4 µM < 1 µM 2 µM 1 µM

ED50 values given; also cytotoxic (ED50/GI50 at 20–750 nM)

[146], [147], [148]

Majusculoic acid

Undefined

lipid

C. albicans

8 µM

cytotoxicity not tested; fluconazole-resistant Candida strains also resistant against majusculoic acid

[162]

Nostofungicidine

Nostoc commune

cyclic lipopeptide

A. candidus

1.5 µM

also cytotoxic (IC50 1.5 µM, NSF-60 cells)

[159]

Scytophycins and Tolytoxins

Scytonema sp. Tolypothrix sp.

macrolides

S. pastorianus N. crassa C. albicans P. ultimum R. solani S. homoeocarpa

24 mm @ 1.2 µmol 30 mm @ 1.2 µmol 23 mm @ 1.2 µmol > 30 mm @ 1.2 µmol 30 mm @ 1.2 µmol > 30 mm @ 1.2 µmol

activity given for scytophycin A; also cytotoxic (IC50 50–100 nM)

[149], [150], [151], [152], [153], [154]

Scytoscalarol

Scytonema sp.

sesterterpene

C. albicans

4 µM

also antibacterial and weakly cytotoxic (IC50 135 µM)

[81]

Tanikolide

Lyngbya majuscula/Moorea producens

fatty acid/δ-lactone

C. albicans

13 mm @ 350 nmol

also brine shrimp and snail toxicity (LD50 12 µM/32 µM)

[156], [157]

Tjipanazoles

Tolypothrix tjipanasensis Fischerella ambigua

indolocarbazoles

?

?

also weak cytotoxicity; no inhibition of protein kinase C at 1 µM

[164]

Tolybyssidins A/B

Tolypothrix byssoidea

cyclic peptide

C. albicans

22 and 42 µM

cytotoxicity not tested

[160]

Very interesting metabolites that were initially identified as antifungal agents but later shown to be extremely potent cytotoxins with IC50 values in the low pM (!) range are the cryptophycins from Nostoc [30], [31], [32]. Their story is told in more detail elsewhere [22].

The hassallidins, first isolated from Tolypothrix [136], [137], were found to be widespread among filamentous cyanobacteria [138]. They are structurally interesting non-ribosomal cyclic depsipeptides decorated with both a sugar and a dihydroxy fatty acid that can also be glycosylated. The MICs of hassallidins A (12) to D against various Candida species including C. albicans are in the range from 1.5 to 10.5 µM [137], [138]. Cytotoxicity of the hassallidins is about 10-fold higher [139]. The balticidins A–D from Anabaena cylindrica are structurally closely related compounds with a comparable antifungal activity spectrum [140]. Interestingly, the cyclic structure does not seem to be essential for activity, as balticidins cleaved at the ester bond of the cyclic depsipeptide showed comparable bioactivity.

The laxaphycins are well-studied antifungal and cytotoxic compounds. Initially isolated from Anabaena laxa [141], [142] and later also from L. majuscula/M. producens [143] and Anabaena torulosa [144], they were shown to be most active in a mixture, pointing at synergistic effects of the different laxaphycins [141]. The combination of congeners A and B had an MIC of about 20 µM against Aspergillus oryzae. It is also active against other fungi such as C. albicans, Penicillium notatum, Saccharomyces cerevisiae, and Trichophyton mentagrophytes. However, as the cytotoxicity of this mixture is about 100-fold higher, the compounds seem of limited use to serve as antifungal lead compounds. Thus, in recent years, these compounds have been studied more from a cytotoxicity point of view, hinting at different modes of actions for laxaphycins A and B, which could already be suspected from their synergistic activity [144], [145].

Majusculamide C, a metabolite isolated from L. majuscula/M. producens, has been found to be active against several plant pathogenic fungi such as Rhizoctonia solani, Pythium aphanidermatum, Aphanomyces euteiches, and Phytophthora infestans at low µM concentrations [146], [147]. However, the compound was later found to also be cytotoxic at even lower, nM concentrations [148]. Cytotoxicity in addition to antifungal activity has also been observed for the scytophycins and the related tolytoxin from, e.g., Scytonema and Tolypothrix, which are the most potent antifungal compounds isolated from cyanobacteria to date [149], [150], [151], [152], [153], [154]. Calophycin from Calothrix fusca is another compound with both antifungal and cytotoxic activity [155]. Tanikolide from L. majuscula/M. producens, chemically related to malyngolide (1) and active against C. albicans, also showed brine shrimp and snail toxicity [156], [157]. Interestingly, 1, differing from tanikolide by opposite stereochemistry and an additional methyl group at the lactone ring, showed no antifungal activity. Pronounced brine shrimp toxicity in addition to activity against C. albicans has also been found for lyngbyabellin B, a cyclic depsipeptide isolated from L. majuscula/M. producens [158]. Nostofungicidine from Nostoc commune has equal antifungal and cytotoxic activity (MIC of Aspergillus candidus and IC50 of NSF-60 cells 1.5 µM) [159]. The ambiguine isonitrils and the hapalindoles have already been discussed above due to their additional antibacterial activity.

A number of antifungal compounds were described without data on cytotoxic activity:

Moderate activity against C. albicans (MIC of 22 µM) has been found for tolybyssidin A (13) from Tolypothrix byssoidea [160].

Fischerellin A (14) from Fischerella muscicola, a structurally interesting compound featuring an enediyne and two heterocyclic moieties, displays antialgal and herbicidal effects in addition to its antifungal activity [161].

Majusculoic acid (15), isolated from an uncharacterized cyanobacterial mat assemblage, exhibited antifungal activity against C. albicans (MIC of 8 µM) [162]. Fluconazole-resistant strains, however, were also resistant against majusculoic acid. Due to its close chemical similarity with lyngbyoic acid (10), it is possible that 15 could also have QS-inhibiting activity.

The lobocyclamides are cyclic lipopeptides from Lyngbya confervoides. A mixture of lobocyclamides A and B had higher activity against a fluconazole-resistant C. albicans than the separate compounds (MIC of the mixture 10 µM) [163].

The tjipanazoles (e.g., tjipanazole A1 16) from Tolypothrix tjipanasensis and Fischerella ambigua were found to be active against a range of phytopathogenic fungi, which is in contrast to other indolocarbazoles that show only weak cytotoxicity and no inhibition of protein kinase C [164]. The alkaloid carriebowlinol (17) has recently been isolated from a yet unclassified cyanobacterium. It showed high antifungal activity (IC50 of 0.2–0.4 µM) and activity against marine bacteria [96]. Scytoscalarol (4) has already been discussed due to its antibacterial activity. It also possesses antifungal activity (MIC against C. albicans 4 µM; cytotoxicity > 30-fold lower) [81].


#

Antiviral Substances

A number of screening campaigns have identified cyanobacteria as a potential source for antiviral compounds [165], [166], [167], [168]. More detailed studies have been done on sulfoglycolipids and lectins.

Sulfoglycolipids

Antiviral sulfoglycolipids such as sulfolipid 1 (18, [Fig. 5]) were isolated from the genera Lyngbya, Phormidium, and Scytonema, and also identified in Anabaena, Calothrix, and Oscillatoria. These compounds as well as structurally related acylated diglycolipids from Oscillatoria and Phormidium show inhibition of the human immunodeficiency virus (HIV-1) via inhibition of the DNA polymerase function of HIV-1 reverse transcriptase [169], [170], [171]. Interestingly, the sulfoglycolipids, showing IC50 values as low as 25 nM, are an order of magnitude more active than the related glycolipids without the sulfonic acid group [170]. Esterification of the free hydroxyl groups of the sulfosugar with further fatty acids leads to a significant decrease of activity [171]. The presence of the fatty acid chains of the sulfoglycolipids are mandatory for activity. Alterations in the fatty acids (e.g., 16 : 0, 16 : 1, 18 : 1, 18 : 2, 18 : 3), however, have a neglectable effect on potency [169], [171].

Zoom Image
Fig. 5 Selected cyanobacterial metabolites with antiviral activity.

#

Lectins

Cyanovirin-N is a peptide lectin isolated from Nostoc ellipsosporum, comprising 101 amino acid residues [172], [173]. It targets N-linked, high-mannose glycans [174], [175], [176], and was found to be a fusion inhibitor, preventing infection with all HI virus types. Cyanovirin-N is active in the low nanomolar range and noncytotoxic at a thousandfold higher concentration. It is also strongly active against influenza A and B, respiratory syncytial virus, enteric viruses, and several coronaviruses [177], [178]. As the compound is readily available by heterologous expression in E. coli and can also be optimized by rational design [179], it is discussed as a promising template for antiviral lectins. Applications for cyanovirin-N are under current investigation [180], although safety issues such as the release of chemokines and stimulatory/mitogenic activity have been recognized [181]. Recently, microvirin, a lectin with a comparable pharmacophor but supposedly better safety profile, has been isolated from Microcystis aeruginosa [182]. A third antiviral lectin from cyanobacteria with similar properties is scytovirin, isolated from Scytonema varium [183].


#

Other antiviral compounds

Nostoflan is a complex acidic polysaccharide from Nostoc flagelliforme [184]. It inhibits the virus-cell interaction of enveloped viruses such as herpes simplex virus, human cytomegalovirus, and influenza A virus, whose cellular receptors are carbohydrates. Interestingly, it exhibits only a very low cytotoxicity and, in contrast to sulfated antiviral polysaccharides, does not show antithrombin activity. Serinol-derived malyngamides (e.g., derivative 19), isolated from an unidentified cyanobacterium, display weak anti-HIV activity [185]. Activity against influenza A virus has been found for the ichthyopeptins A (20) and B (IC50 of 12 µM), isolated from Microcystis ichthyolabe [186]. Other antiviral compounds show significant cytotoxicity. These compounds comprise some aplysiatoxin derivatives from Trichodesmium erythraeum, which are active against Chikungunya virus [187], as well as the β-carbolines bauerines A–C isolated from Dichothrix baueriana [188] and the indolocarbazoles isolated from Nostoc sphaericum [189], both active against herpes simplex virus type 2.


#
#

Antiprotozoal Compounds

Many compounds active against the malaria parasite Plasmodium as well as other protozoal parasites such as Trypanosoma (sleeping sickness or Chagasʼ disease) or Leishmania (leishmaniasis) have been reported from cyanobacteria [190]. However, as it has already been noted with antibacterial and antifungal compounds, many compounds found to be active in the antiprotozoal assays also display cytotoxicity, limiting their usability as drug leads. It is common to evaluate a compoundʼs selectivity index (IC50 against human cell lines vs. IC50 against parasites) to assess bioactivity.

The ribosomal cyclic peptide aerucyclamide B (21, [Fig. 6]) from M. aeruginosa is the most active antiplasmodial compound isolated from cyanobacteria to date [191]. Its IC50 of 0.7 µM against P. falciparum and an almost 200-fold lower cytotoxicity make it an interesting lead. The structurally closely related balgacyclamides, isolated from a different M. aeruginosa strain, show comparable activity [192]. A total synthesis for aerucyclamide B has been established, and the first work on the optimization of the anti-trypanosomal activity of the aerucyclamides has been described, leading to a compound active against T. brucei with a selectivity index of about 150 [193].

Zoom Image
Fig. 6 Selected cyanobacterial metabolites with antiprotozoal activity.

Synthetic optimization of the linear lipopeptide almiramides from L. majuscula/M. producens has resulted in derivatives (e.g., 22) active against Leishmania donovani in the low µM range with selectivity indices of up to 50 [194], [195].

Nostocarbolin has already been discussed in the section “Antibacterial Metabolites”. Of special interest concerning antiprotozoal activity is a synthetic nostocarboline dimer (23) that, in addition to having an IC50 of 18 nM against P. falciparum, showed a selectivity index against the parasite vs. rat myoblasts of > 2500. Other dimers were more potent against Trypanosoma or Leishmania [196].

The linear depsipeptide viridamide A (24) from Oscillatoria nigro-viridis, structurally distantly related to the almiramides, showed an IC50 in the low µM range against the parasites P. falciparum, Trypanosoma cruzi, and Leishmania mexicana. The cytotoxicity data presented do not allow a comparative assessment of cytotoxicity, but activity seems to be in favor of antiparasitic activity [197]. The same holds true for the cyclic depsipeptide companeramides from a yet unclassified filamentous cyanobacterium with moderate antiplasmodial activity [198].

Several compounds with antiprotozoal, but at the same time cytotoxic activities, have been described (selectivity indices ≤ 10). Among these compounds are the lipophilic phenolic ambigols from F. ambigua (already mentioned above) [91] and hierridin B from Phormidium ectocarpi and Cyanobium sp. [199], [200], the indolophenanthridine alkaloid calothrixins from Calothrix [201], the linear lipopeptides carmabin A, dragomabin, and dragonamide A [202], the cyclic depsipeptides lagunamide A–C [203], [204] the cyclodepside malyngolide dimer [205] from L. majuscula/M. producens, the cyclic peptides venturamide A and B from Oscillatoria [206], and the linear peptide gallinamide A from Schizothrix [207], which has recently been shown to be a potent inhibitor of human cathepsin L [208].


#

Molluscicidal Bioactivities

Schistosomiasis, caused by the parasitic flatworm Schistosoma, is one of the most prevalent parasitic infections worldwide. Hundreds of millions of people are infected and/or at risk of infection, especially in African countries [209], [210]. Schistosoma has a complex life cycle that requires both an aquatic snail host belonging to the genus Biomphalaria and a mammalian host to complete their reproductive cycle. As eradicating the disease in infected patients using anthelmintic drugs like praziquantel does not protect against the possibility of reinfection, the treatment of water containing the snail vector with molluscicides like niclosamide is seen as a viable way to protect the population from schistosomiasis. This strategy is widely practiced and nowadays a crucial part of schistosomiasis control, but has several drawbacks such as the price of the applied compounds as well as environmental concerns [211]. Thus, there is a demand for novel molluscicides. As cyanobacteria might produce molluscicidal compounds to protect themselves against snails feeding on cyanobacterial biomass, strains have been screened for molluscicidal activity, and subsequently, several compounds with this bioactivity have been found. Their structures are shown in [Fig. 7].

Zoom Image
Fig. 7 Selected cyanobacterial metabolites with molluscicidal activity.

The first compound for which molluscicidal activity has been found is barbamide (25) from L. majuscula/M. producens [212]. It has specific molluscicidal activity (LC100 of 21.6 µM), and does not show brine shrimp toxicity or ichthyotoxicity. Due to several intriguing chemical features, like a trichlormethyl group and a methyl enol ether of a β-keto amide, its biosynthesis has been studied in great detail [213], [214], [215], [216], [217].

Cyanolide A (26) from Lyngbya bouillonii is an unusual symmetric glycosidic macrolide. It is more active against Biomphalaria than barbamide (LC50 of 1.2 µM), but in addition it also shows brine shrimp toxicity (LC50 of 10.8 µM) [218]. As the compounds cytotoxicity is comparably low (nontoxic at 35 µM), its total synthesis has received considerable attention (e.g., [219], [220], [221], [222]).

Thiopalmyrone (27) and palmyrrolinone (28) have been isolated from the same marine cyanobacteria assemblage of Oscillatoria and Hormoscilla. Both compounds activities were in the low micromolar range (LC50s of 8.3 and 6.0 µM, resp.) [223]. As the compounds were not antifungal, antibacterial, or cytotoxic up to a concentration of 100 µM, they seem to be selectively molluscicidal and deserve closer attention.


#

Conclusions

To date, many compounds with anti-infective activities have been isolated from cyanobacteria. Most of the isolated compounds showed cytotoxicity in addition to the desired anti-infective activity, limiting their direct use as anti-infective drug substances. However, they might well serve as lead compounds for the development of derivatives with a lower toxicity, a common approach in pharmaceutical natural product chemistry.

Interestingly, despite intense screening campaigns searching for novel antibacterial compounds from cyanobacteria, these organisms have only been a poor source for selective antibacterial compounds, in contrast to other microorganisms such as the soil dwelling actinomycetes (of the about 7000 known compounds from the genus Streptomyces, about 2600 compounds have antibiotic activity) or microscopic fungi. An explanation for this observation might be the unique lifestyle of cyanobacteria compared with the aforementioned microorganisms. Photosynthetic cyanobacteria as autotrophs and primary producers serve as food for a plethora of micro- and macroorganisms [224]. They do not directly compete with heterotrophic bacteria for organic nutrients in their environment, but rather compete with other phototrophic organisms or eukaryotic organisms grazing on cyanobacteria. That might explain why the synthesis of antibacterial compounds might only be of limited evolutionary advantage, while the production of compounds with toxicity against other phototrophic organisms or eukaryotic organisms might be favored. This could suggest that searching for cyanobacterial compounds targeting eukaryotes such as fungi, protozoa, and mollusks might be more fruitful than searching for antibacterial metabolites in cyanobacteria. But, as has been noted above, cyanobacteria are still poorly researched. Thus, it can be expected that selective antibacterials will also be found in these organisms sooner or later. Another interesting aspect to note concerning the compounds found to have antibacterial activity is that they most often are not peptides, the structural class most often observed in cyanobacteria. Instead, mainly polyketides, terpenes, and alkaloids have been found to exert antibacterial activity. This might be due to the fact that peptides have physicochemical properties unsuitable for crossing cell walls. It is also worth mentioning that cyanobacterial secondary metabolites have more often been isolated from the biomass rather than from the cultivation medium. For example, amongst the antibacterial compounds summarized in this review, only bromoanaindolone, the comnostins, and noscomin have been described as being isolated from the cultivation medium. Although this finding is slightly biased due to the fact that cyanobacteria biomass has simply been better studied in the past compared to cyanobacteria cultivation media, it might indicate that the antibacterial compounds found to date have not evolved to act as antibacterials, as this would need export or diffusion into the extracellular environment. It might be sensible to intensify the examination of compounds secreted into the medium during cyanobacteria cultivation for the purpose of finding novel anti-infective secondary metabolites from these organisms.

A rather novel field is the search for cyanobacterial QS inhibitors as indirect antibacterials. Own screening results (not discussed here) confirm that cyanobacteria have a good potential in this area, and the isolation of QS-inhibiting natural compounds from hit extracts is under way in our laboratory. Antiviral lectins are the most advanced anti-infective compounds from cyanobacteria, although no compound has reached the market, yet.

Not discussed within the scope of this review is the potential of cyanobacterial protease inhibitors as anti-infective compounds. As many potent protease inhibitors have been isolated from cyanobacteria [22], and proteases are validated or currently discussed targets for treating viral or protozoal infections [225], it can be expected that cyanobacterial compounds will, in the future, play a role in this research field.

As discussed above, the pace of cyanobacterial natural product research has increased. We expect that the majority of cyanobacterial genera, species, and strains have not even been discovered yet, and that novel cyanobacterial metabolites with interesting bioactivities will continuously be described as more academic and industrial groups acknowledge the potential of their exploitation. In view of the relatively small overall number of cyanobacterial strains and isolated secondary metabolites investigated, it is remarkable that several of them have already entered clinical trials, and that one analog based on the dolastatins could successfully be developed into a marketed drug (brentuximab vedotin). A more thorough exploitation of cyanobacterial natural products for drug discovery will potentially increase the number of successful drug discoveries in the future, also in the field of anti-infective drugs.

Often, the suitability of cyanobacteria as producers of secondary metabolites is controversially discussed. However, great advances have been made in this field. Cyanobacteria as photosynthetic autotrophs do not require carbon or energy sources supplied in the cultivation media, but accept cultivation in inorganic salt solutions. This facilitates medium standardization and logistics when mass cultivations are needed. The downstream processing (the isolation of the secondary metabolites) is straightforward, as the products do not have to be separated from a complex organic medium. Furthermore, the use of inorganic medium ingredients reduces the costs and hampers the growth of contaminating heterotrophic organisms. However, cultivation facilities with artificial or natural illumination and photobioreactor systems have yet to be established before cyanobacteria can be utilized to produce active pharmaceutical ingredients (APIs) under GMP-compliant conditions.

The longer generation times compared to, e.g., E. coli (about 1 day vs. about 20 min), result in substantially longer cultivation times to reach a given amount of biomass. Also, production titers of secondary metabolites are still 2–3 orders of magnitude lower compared to the optimized industrial bacterial or fungal production systems. These considerations have led to several developments. Firstly, most programs that pursue leads from cyanobacteria make a switch from cyanobacteria cultivation/processing to total synthesis in order to produce multigram quantities of a drug substance. Secondly, research on the production of cyanobacterial secondary metabolites in heterologous hosts such as E. coli has been taken up. The proof-of-principle that complex ribosomal metabolites and more simple cyanobacterial metabolites can be produced in this way has successfully been brought forward [226], [227], [228], [229], [230]. However, due to the size and multifunctional character of PKS/NRPS biosynthesis complexes, the heterologous expression of non-ribosomal cyanobacterial peptides is a significant challenge, and no successful example has been published to date. A third possibility to overcome the lower relative productivity is the development of low-cost large-scale photobioreactors allowing for cost efficient mass cultivation of the cyanobacteria of interest. In this context, the significant progress made in regard to using microalgae for food as well as biofuels and chemical feedstock production will also support the development of novel low-cost photobioreactors for high-value products, such as pharmaceuticals.

Therefore, we advocate striving for intensified drug discovery efforts from cyanobacteria as well as for improved biotechnological and process engineering solutions for compound production.


#

Acknowledgement

I thank M. Preisitsch, University of Greifswald, Germany, and Dr. D. Kramer, Cyano Biotech GmbH, Berlin, Germany, for helpful discussions and critically reading the manuscript. Parts of this review have been published before in reference [22].


#
#

Conflict of Interest

The author has been employee at the company Cyano Biotech GmbH, Berlin, Germany, until 11/2013. He still serves in the academic advisory board of the company.


Correspondence

Prof. Timo Niedermeyer
University of Tübingen, Interfaculty Institute for Microbiology and Infection Medicine
Auf der Morgenstelle 28
72076 Tübingen
Germany
Phone: +49 7 07 12 97 20 79   
Fax: +49 70 71 29 59 99   


Zoom Image
Fig. 1 SciFinder entries for “Natural product from cyanobacteri?” (09 Dec 2014).
Zoom Image
Fig. 2 Selected cyanobacterial metabolites with antibacterial activity.
Zoom Image
Fig. 3 Cyanobacterial metabolites with quorum sensing inhibiting activity.
Zoom Image
Fig. 4 Selected cyanobacterial metabolites with antifungal activity.
Zoom Image
Fig. 5 Selected cyanobacterial metabolites with antiviral activity.
Zoom Image
Fig. 6 Selected cyanobacterial metabolites with antiprotozoal activity.
Zoom Image
Fig. 7 Selected cyanobacterial metabolites with molluscicidal activity.