The toxicity and enzyme activity of a chlorine and sulfate containing aeruginosin isolated from a non-microcystin-producing Planktothrix strain

Abstract

The toxicity of six different Planktothrix strains was examined in acute toxicity assays with the crustacean Thamnocephalus platyurus. The presence of toxicity in two strains could be explained by the occurrence of microcystins. The other four Planktothrix strains were not able to produce microcystins due to different mutations in the microcystin synthetase (mcy) gene cluster. In these strains, toxicity was attributed to the presence of chlorine and sulfate containing compounds. The main representative, called aeruginosin 828A, of such a compound in the Planktothrix strain 91/1 was isolated, and structure elucidation by 2D-NMR and MS methods revealed the presence of phenyllactic acid (Pla), chloroleucine (Cleu), 2-carboxy-6-(4’-sulfo-xylosyl)-octahydroindole (Choi), and 3-aminoethyl-1-N-amidino-Δ-3-pyrroline (Aeap) residues. Aeruginosin 828A was found to be toxic for Thamnocephalus platyurus with a LC₅₀ value of 22.4 µM, which is only slightly higher than the toxicity found for microcystins. Additionally, very potent inhibition values for thrombin (IC₅₀ = 21.8 nM) and for trypsin (IC₅₀ = 112 nM) have been determined for aeruginosin 828A. These data support the hypothesis that aeruginosins containing chlorine and sulfate groups, which were found in microcystin-deficient Planktothrix strains, can be considered as another class of toxins.

1. Introduction

Climate change as well as the anthropogenic input of nutrients into freshwaters has resulted in an increase of cyanobacterial biomasses in many lakes and rivers throughout the world during the last decades (Paerl and Huisman, 2008). These organisms are often found to accumulate in very high densities also in lakes that are intensively used as recreational areas or as drinking water reservoirs (Posch et al., 2012). However, cyanobacteria are known to produce and store intracellularly a wide variety of bioactive secondary metabolites, and a collapse of such a bloom might liberate high amounts of these compounds into the water. In recent years, substantial progress has been made to identify possible harmful compounds from different cyanobacterial genera. Most of these substances could be assigned to distinct chemical oligopeptide classes, e.g. microcystins, cyanopeptolins, anabaenopeptins, aeruginosins, cyclamids, microginins and microviridins (Welker and von Döhren, 2006). These peptides exhibit various biological activities such as the inhibition of different phosphatases and proteases (Welker and von Döhren, 2006). Microcystins (MCs) have been found to be responsible for countless animal poisonings all over the world, and even for human toxicity (Sivonen and Jones, 1999), and cyanopeptolins and aerucyclamides received some attention due to their toxicity to Thamnocephalus platyurus (Blom et al., 2003; Gademann et al., 2010; Portmann et al., 2008a; Portmann et al., 2008b).

There is intense discussion about the possible ecological role of MCs in the environment; however, MCs are still regarded to be the primary defence mechanism of cyanobacteria against grazers (Blom et al., 2001; Kurmayer and Jüttner, 1999). MC-deficient Planktothrix genotypes are typically found in rather low proportions of the total cyanobacterial population (Ostermaier and Kurmayer, 2009), but might occasionally exceed 50% of the total abundances. Reasons for the lack of MC production might be various point mutations (e.g. insertions or deletions) or the loss of genes encoding the MC synthetase (mcy) gene cluster (Christiansen et al., 2006; Christiansen et al., 2008). However, the loss or dysfunction of the mcy gene cluster did not seem to be disadvantageous for the overall success of the Planktothrix genotypes. This is supported by the linear relation of MC deficient genotypes to the total population density of bloom forming Planktothrix populations in European lakes (Ostermaier and Kurmayer, 2009). In order to explain these findings it was suggested that an alternative peptide or peptide class might functionally compensate for the lack of MCs. Here we report on a peptide class, the chlorine and sulfate containing aeruginosins, which were found in MC-deficient strains, and which exhibited acute toxicity against the crustacean T. platyurus.

2. Material and methods

2.1. General Experimental Procedures

¹H-NMR spectra were recorded on a Bruker Avance III 600 MHz with a 5 mm BBFO plus probe or a Bruker Avance III Ascend 700 MHz spectrometer with a 5 mm TCI (H-C/N-D) cryo probe at room temperature. Chemical shifts (δ-values) are reported in ppm, spectra were calibrated related to the solvent residual proton chemical shift (DMSO, δ = 2.50). The coupling constants are specified in Hz. HRMS spectra were recorded on a Bruker maXis 4G instrument. Identification and purification of aeruginosin 828A (1) was performed on a Shimadzu 10AVP HPLC system equipped with an automated sample injector, a thermostated column compartment, and photo diode array detector. Mass spectra were recorded on a combined LC-MS (LCQ Duo mass spectrometer, Finningan Thermoquest, USA) equipped with an electrospray ionization source (ESI-MS).

2.2. Culture, Extraction, and Bioassay-guided fractionation

A culture collection of six Planktothrix strains was established in the laboratory at the beginning of this study. All cyanobacteria were grown in 300 mL Erlenmeyer flasks at 20 °C under constant light conditions at an irradiation of 6 μmol m^-2 s^-1 from fluorescent tubes (Osram 930; Lumilux Delux; Warm White 3000K) in 120 mL mineral medium described by Jüttner and co-workers (Jüttner et al., 1983). Two of the six Planktothrix strains were capable of producing MCs, four strains lacked the MC production either due to point mutations or complete loss of the mcy gene cluster (summarized in Table S1, Supplementary material). The analysis of the mcy gene cluster in Planktothrix strains was carried out as part of another study as described elsewhere (Christiansen et al., 2008).

Frozen biomass of all six Planktothrix strains was extracted twice with 50 % MeOH (10 mL per gram of wet cell biomass) for 2 h in the dark. After centrifugation (25’700 g for 15 min) the supernatants (‘crude extracts’) were fractionated by HPLC equipped using a reversed phase column (Hydrosphere C18, YMC, 4.6 x 250 mm, Stagroma, Reinach, CH) using two solvents: UV-treated deionised water and acetonitrile. The solvents were free of trifluoroacetic acid (TFA) to avoid undesired toxic side effects and unwanted isomerisation products (Blom et al., 2001). A linear increase was applied (acetonitrile from 20 % to 70 % in 50 min, 70 to 100 % in 2 min, isocratic for additional 10 min). Fractions were collected every three minutes; the solvents were evaporated by vacuum centrifugation (SPEEDVAC Plus® SC110A, Savant Instruments Inc. USA). The fractions containing the different cyanobacterial compounds were transferred to a 24 well plate, and re-dissolved in 1 % MeOH (500 µL). The toxicity of the fractions of the six crude extracts was tested in a 24-h acute toxicity assay performed using instar II-III larvae (Thamnotoxkit F; MicroBioTests Inc.) of the sensitive crustacean Thamnocephalus platyurus, which can be easily hatched from cysts (Blom et al., 2003; Blom et al., 2001). About 20 to 30 crustaceans were transferred to each well; after 24 h, the mortality rate was determined for each fraction. Each fraction represented the extract of an equivalent of 0.05 µM chlorophyll a of the biomass of the six Planktothrix strains.

2.3. Isolation of aeruginosin 828A

Fresh biomass (32 g) of Planktothrix strain 91/1 was extracted with 50 % MeOH. Crude extracts were obtained after centrifugation and were separated by HPLC using a reversed phase column (Hydrosphere C18, YMC, 4.6 x 250 mm, Stagroma, Reinach, CH) under the following conditions: solvent A was UV-treated deionised water (+ 0.05 % TFA), solvent B: HPLC-grade acetonitrile (+ 0.05 % TFA); a linear increase was applied (as described above). Under the conditions applied aeruginosin 828A (1) eluted after 16.0 min. TFA was removed before evaporating the solvents to prevent undesirable isomerization products by applying sequentially the combined HPLC fractions on a C18 cartridge (10 g; Mega Bond Elute, Varian, Agilent Technologies, Basel, CH; conditioned with 10 % MeOH). The cartridge was flushed with water to remove the TFA, and aeruginosin 828A (1) was eluted with 80 % MeOH. Subsequently, the aqueous methanol was evaporated (35 mbar, 40 °C) to achieve a colourless, amorphous solid. After purification 533 µg of pure aeruginosin 828A (> 99 % HPLC) could be achieved.

Aeruginosin 828A (1): UV (47% acetonitrile in water with 0.05 % TFA λ_max 277 nm; ¹H and ¹³C NMR data (DMSO-d₆), see Table 1; HRMS-ESI: calcd. for C₃₆H₅₃³⁵ClN₆O₁₂SNa⁺ [M+Na]⁺: 851.3023; found: 851.3019.

Table 1.

NMR spectroscopic data for aeruginosin 828A.

Aeruginosin 828A (700MHz, DMSO-d6, 298K)
Residue	Position	1H (J in Hz)	13C (from 2D)	HMBCa	NOE
Xyl	1eq	4.94 d (3.8)	95.0	Xyl 2,3,5, Choi 6	Choi 6, Xyl 2,4',7,7'
	2ax	3.28 m	71.8	Xyl 3,5	Xyl 1,4
	3ax	3.57 ddd (2.6, 9.0, 9.1)	71.4	Xyl 2,4,5	Xyl 4,5ax, 3-OH
	4ax	3.93 ddd (5.8, 9.1, 10.6)	74.7	Xyl 3,5	Xyl 1,2,5ax,5eq,3-OH
	5ax	3.36 dd (10.6, 10.7)	59.3	Xyl 1,3,4	Xyl 4,5eq
	5eq	3.67 dd (5.8, 10.7)		Xyl 1,3,4	Xyl 4,5ax
	2-OH	4.41 (7.3)	-	Xyl 1,2	-
	3-OH	4.96 (2.6)	-	Xyl 2,3,5	Xyl 3,4
Choi	1	-	171.1	-	-
	2	4.18 dd (9.4, 8.2)	59.5	Choi 1,3	Aeap NH, Choi 3,3',3a,7
	3	2.01 ddd ( 12.6, 7.3, 7.3)	30.5	Choi 3a,7a	Choi 2,3a,3',4',6,7',7a
	3'	1.81 ddd ( 12.6, 12.6, 9.7)	-	Choi 1,2,3a	Choi 2,3,3a,4',5',7
	3a	2.25 m	35.6	Choi 3,4,5,7,7a	Choi 2,3',4',7,7a
	4	2.14 m	19.1	Choi 3,3a,5	Choi 3a,4',5,5'
	4'	1.48 m	-	Choi 5,6,7a	Choi 3a,4,5',6
	5	1.49 m	24.4	Choi 6,7a	Choi 6,Cleu 2
	5'	1.54 m	-	-	Choi 3a,4,5,6,7'
	6	3.83 m	68.4	-	Choi 5,5',7,7', Xyl 5ax
	7	1.58 dd (11.9, 12.9)	28.4	Choi 7a	Cleu 2, Choi 6,7'
	7'	2.25 m	-	Choi 5,7a	Choi 2,6,7',7a, Cleu 2
	7a	4.32 ddd (11.9, 6.4, 6.4)	54.1	Choi 2,3,3a,7, Cleu 1	Choi 3a,4,4',7,7', Cleu 2,5
Pla	1	-	172.5	-	-
	2	4.18 (br)	71.5	-	Pla 3, 3'
	3	2.79 dd ( 14.0, 7.6)	39.7	Pla 1,2,4,5,9	Pla 2,3'
	3'	2.96 dd (14.0, 3.7)	-	Pla 1,2,4,5,9	Pla 2,3
	4	-	137.9	-	-
	5,9	7.23 m	129.5	Pla 3,5,7,9	Pla 3,3',7
	6,8	7.26 m	127.7	Pla 4,6,8	Pla 5,7,9
	7	7.18 m	125.8	Pla 5,9	-
	2-OH	-	-	-	-
Cleu	1	-	167.4	-	-
	2	4.93 dd (10.7,8.7)	50.9	Cleu 1,3, Pla 1	Cleu 3,4,5, Choi 7a,7,7'
	3	4.00 dd (10.7, 1.8)	68.6	Cleu 1,2,4,5,5'	Cleu 2,4,5'
	4	1.71 dsept. (6.6, 1.8)	27.3	Cleu 5,5'	Cleu 2,3,5
	5	0.87 3H, d (6.6)	15.3	Cleu 3,4,5'	Cleu 2,3,4,l Choi 7',7a
	5'	0.86 3H, d (6.6)	20.6	Cleu 3,4,5	Cleu 2,3,4,l Choi 7',7a
	NH	7.68 d (8.7)	-	-	Cleu 3
Aeap	1	3.16 dddd (13.0, 6.5, 6.5, 5.7)	36.3	Aeap 2,3, Choi 1	Aeap 2,4,5,6,NH, Choi 2
	1'	3.23 dddd (13.0, 6.5, 6.5, 5.7)	-	Aeap 2,3, Choi 1	Aeap 2,4,5,6,NH, Choi 2
	2	2.25 2H, m	28.1	Aeap 1,3,4,6	Aeap 1,1',4,5,6,NH
	3	-	136.1	-	-
	4	5.61 t (1.6)	119.0	Aeap 2,3,5,6	Aeap 1,1',2,5,6,NH
	5	4.07 2H, d (1.6)	53.6	Aeap 3,4	Aeap 1,1',2,4
	6	4.07 2H, s	54.9	Aeap 3,4	Aeap 1,1',2,4
	NH	8.00 dd (5.7, 5.7)	-	Aeap 1, Choi 1	Aeap 1,1',2,4,5,6,Choi 2,3,3'
	8	-	154.9	-	-

^aHMBC correlations are given from proton(s) stated to the indicated carbon atom.

2.4. Acute toxicity of aeruginosin 828A

The highly purified aeruginosin 828A was tested in a 24 h acute toxicity assay with T. platyurus in six concentrations ranging from 0.5 to 100 μM in triplicates. For every concentration, 20 to 30 animals were used. The nonlinear regression analysis as well as the LC₅₀ value were calculated using Graph Pad Prism 5 for Windows.

2.5. Enzyme inhibition assays

The inhibition of trypsin (0.04 U/ 200 µL; No 9471 Fluka, Buchs, Switzerland), and thrombin (5 nM; IHTa Innovative Research, Peary, USA) was tested in microtiter plates. Boc-Gln-Ala-Arg-aminomethylcoumarin (50 µM; Bachem AG, Bubendorf, Switzerland) served as substrate for trypsin, and Boc-Phe-Ser-Arg-aminomethylcoumarin (100 µM; Bachem AG, Bubendorf, Switzerland) as substrate for thrombin according to previously established protocols (Blom et al., 2006; Gademann et al., 2010). All fluorescent substrates were dissolved in H₂O (5% DMSO). The reaction solution consisted of 140 µL Tris buffer (50 mM Tris-HCl buffer pH 8.0, 150 mM NaCl, 1 mM CaCl₂ and 0.1 mg/mL BSA), 10 µL enzyme solution, 30 µL toxin solution, and 20 µL fluorescent substrate; samples were measured at 37 °C for 20 min in a Fluorescence plate reader (SpectraMAX, GeminiXS, Molecular Devices Corp., USA) with the excitation wavelength at 380 nm and the emission wavelength at 440 nm.

3. Results and Discussion

3.1. Bioassay-guided fractionation

Methanolic extracts of the biomasses of six Planktothrix strains (Fig. 1) were fractionated and assayed against the sensitive freshwater crustacean T. platyurus for acute toxicity. Overall, more than 35 compounds were present in these fractions, and out of total 72 fractions, twelve were found to be highly toxic. In these fractions, three compounds were assigned to known MCs, seven to unknown (sulfate-containing) cyanopeptolins, and ten were sulfate and chlorine containing peptides of the size of aeruginosins (Fig. 1).

Fig. 1.

HPLC-chromatograms (220 nm; 0.05 µM chlorophyll a equivalent) of all six Planktothrix strains (AB = Anabaenopeptin, AG = Aeruginosin, Cl^--AG = Cl^--Aeruginosin, Cl^--SO₃-AG = Cl^--SO₃-Aeruginosin, CP = Cyanopeptolin, SO₃-CP = SO₃-Cyanopeptolin, MC = Microcystin, PC= Planktocyclin; x = fraction was toxic to T. platyurus).

3.2. Structure elucidation of aeruginosin 828A

In order to elucidate the structure of the unknown toxin, the major toxic oligopeptide of P. rubescens strain 91/1 was purified and analysed by NMR spectroscopy. The putative molecular ions at m/z = 851.3019 [M+Na]⁺ and m/z = 853.3003 [M+Na]⁺ in the HRMS spectrum support a molecular formula of C₃₆H₅₃³⁵ClN₆O₁₂SNa⁺ or of C₃₆H₅₃³⁷ClN₆O₁₂SNa⁺ for the respective Na adduct. In addition, the mass spectrum suggests the presence of a sulfate group by the fragment with m/z = 749, as well as the presence of one Cl atom with its characteristic isotope pattern (Fig.S1, Supplementary material). The ¹H–NMR spectrum of aeruginosin 828A (1) (Table 1; Fig. S2 Supplementary material) displays two sets of characteristic peptide resonances, among others. The constitution of each of the building blocks was assigned using COSY, HSQC, HMBC and NOESY experiments, and the chemical shifts as well as key HMBC and NOE correlations are reported in Table 1. These 2D-NMR experiments and comparison to literature data led to the identification of the following building blocks: phenyllactic acid (Pla); chloroleucine (Cleu); 2-carboxy-6-(4’-sulfo-xylosyl)-octahydroindole (Choi) fragment and 3-aminoethyl-1-N-amidino-Δ-3-pyrroline residue (Aeap). The position of the Cl atom in Cleu was assigned based on the characteristic chemical shift of the corresponding Cβ atom of the Leu moiety (δ_c = 68.6 ppm). The presence of a carbohydrate residue could be detected due to the low field shifted carbon atoms (δ_c ≈ 70 ppm), which indicated the presence of oxygen substituents. The xylose was identified based on the typical proton and hydroxy group substitution pattern. The position of the sulfate group was established by the typical chemical shift of Xyl C-4 (δ_c = 74.7 ppm), by the COSY correlations between H-2 (δ_H = 3.28 ppm) and Xyl OH-2 (δ_H = 4.41 ppm), as well as between Xyl H-3 (δ_H = 3.57 ppm) and Xyl OH-3 (δ_H = 4.96 ppm) (Table 1; Fig. S3 and S4, Supplementary material). The relative configuration of the carbohydrate moiety was assigned on the basis of the NOESY spectrum and J-coupling constants. The α-anomer of the sugar moiety was assigned by its small coupling constant (J = 3.8 Hz), which results from the (+)-sc arrangement between Xyl H-1 to Xyl H-2. The configuration of the Choi moiety was assigned using the relevant NOE correlations between Choi H-3’ (δ_H = 1.81 ppm) and Choi H-4’ (δ_H = 1.48 ppm), as well as Choi H-4 (δ_H = 2.14 ppm) and Choi H-3a (δ_H = 2.25 ppm), Choi H-3 (δ_H = 2.01, 1.81 ppm) to Choi H-3a (δ_H = 2.25 ppm) and, additionally, to Choi H-6 (δ_H = 3.83 ppm) and Choi H-7’ (δ_H = 2.25 ppm). The configuration at the C-6 stereogenic center of Choi was assigned on the basis of the line width of its signal (total ~ 16 Hz), which implies that no trans-diaxial relationship of the Choi H-6 to its coupling partners is present. Due to the lack of commercially available, authentic standards for the Cleu and Choi residues, we did not perform hydrolysis and subsequent analysis on chiral stationary phases. The observed NOE between Choi H-3a and Choi H-7a supports the relative cis-configuration, which is further substantiated by numerous literature examples of identical configurations. The tentatively assigned relative configuration of aeruginosin 828A (1) is further corroborated by comparison with data reported for the structurally similar aeruginosin 205B (Table S2, Supplementary material) (Hanessian et al., 2009). This natural aeruginosin differs only by an agmatine moiety instead of the Aeap residue and shows almost identical ¹H and ¹³C shifts. We therefore assign the relative configuration for the Pla and Cleu building blocks as shown for compound 1. Yet, we are aware that its final structure elucidation will require the use of total chemical synthesis.

The assembly of the different fragments was established by HMBC and NOESY experiments (Fig.S5 and S6, Supplementary material). The linkage between central Choi and the Cleu was established by a HMBC correlation between Choi H-7a (δ_H = 4.32 ppm) and Cleu C-1 (δ_c = 167.4 ppm). In a similar fashion, the connection between the Choi and the Aeap residues was determined by the coupling between Aeap H-1 and H-1’ (δ_H = 3.16 ppm and 3.23 ppm) and Choi C-1 (δ_c = 171.1 ppm). The HMBC correlation from Cleu H-2 (δ_H = 4.93 ppm) to Pla C-1 (δ_c = 172.5 ppm) unambiguously demonstrated the connection of these moieties. In addition, the connection of the Xylose unit with the Choi moiety was elucidated by the NOE correlation of the Xyl H-1 (δ_H = 4.94 ppm) to Choi H-6 (δ_H = 3.83 ppm) as well as by the HMBC correlation of the Xyl H-1 (δ_H = 4.94 ppm) to Choi C-6 (δ_c = 68.4 ppm).

3.3. Protease inhibitory activities

Aeruginosin 828A (1) belongs to a large group of linear peptides that feature the (hydroxyl)phenyllactic acid (Hpla/Pla) or glyceric acid at the N-terminus (position 1), followed by a variable amino acid in position 2, the Choi moiety in position 3, and an arginine derivative at the C-terminus in position 4 (Welker and von Döhren, 2006). These peptides have been evaluated for high inhibitory activity against trypsin and trypsin-like serine proteases such as thrombin. Aeruginosin 828A (1) exhibited strong inhibition towards these proteases: an IC₅₀ value of 21.8 nM was measured for thrombin, an IC₅₀ value of 112 nM for trypsin (Fig. 1). Thus, aeruginosin 828A (1) is one of the strongest inhibitors of trypsin-like enzymes among the aeruginosins, together with chlorodysinosin A, and the aeruginosins 205A and B (Hanessian et al., 2006; Shin et al., 1997; Toyooka et al., 2003). Common to these four aeruginosins is the presence of a Cleu in position 2 that was so far unknown for natural products (Hanessian et al., 2006), and of either a sulfated glyceric acid derivative at the N-terminus or a sulfated Choi subunit. A comparison of the biological activity of other aeruginosins led to the hypothesis that both, chlorine and sulfate moieties are necessary for strong inhibition and toxicity of these peptides.

It was proposed that the Choi sulfate group would increase the selectivity for trypsin rather than for thrombin (Sandler et al., 1998), and that the sulfate group in the N-terminal position would make a significant contribution upon inhibiting thrombin (Carroll et al., 2004). This is supported by the hydrogen bonding network of the dysinosin A-thrombin complex (Carroll et al., 2002). The N-terminated sulfate group of dysinosin A interacted with arginines 208 and 263 of the thrombin with several hydrogen bonds, whereas the 5,6-dihydroxyoctahydroindole group did not appear to have any hydrogen bonding interactions with thrombin. A structure and activity comparison of different aeruginosins revealed that the inhibition values for trypsin of Choi sulfated aeruginosins were lower than the IC₅₀ values for thrombin and that the presence of the N-terminal sulfate group considerably decreased the IC₅₀ values for thrombin (Table 2, groups 2 and 3). The IC₅₀ values for the inhibition of trypsin are typically in the low micromolar range. However, for a strong inhibition of thrombin in the low nanomolar range, the N-terminal sulfate group seems to be essential (Table 2, groups 2 and 3). The inhibition values for aeruginosin 828A seem to be in opposition to this more general trend. However, other factors, such as the arginine derivative in position 4 might have an additional influence on the inhibition potential towards trypsin-like proteases as it was shown for the guanidino group of dysinosin A that strongly bonded to aspartic acid 229 of thrombin (Carroll et al., 2002). Future experiments by using derivatives of aeruginosin 828A might address this conundrum.

Table 2.

Structure characteristics and enzyme inhibition of aeruginosins (that are characterized by the following structure: Pla/Hpla/Glyceric acid – Amino acid – Choi – Arg derivative).

Group	Aeruginosin example (number of peptides evaluated)		Modifications (Cl, chlorination; Br, bromination; SO3, sulfation)			Enzyme inhibition (IC50) 7
			Pla/Hpla/Glyceric acid	Amino acid	Choi	Trypsin [µM]	Thrombin [µM]
1	Chlorodysinosin A1	(1)	SO3	Cleu	---	0.037	0.0057
	Aeruginosin 828A2	(3)	---	Cleu	SO3	0.09-0.112	0.022-1.9
2	Aeruginosin 89A3	(2)	Cl, SO3	Leu	---	0.6-9.2	0.04-0.07
	Aeruginosin 98A4	(6)	Cl, Cl2, Br, Br2	Ile	SO3	0.1-18.2	3.6- >45.5
3	Aeruginosin 102-A5	(6)	SO3	Tyr / Val / Leu	---	0.27-1.5	0.05-0.14
	Aeruginosin 98B4	(1)	---	Ile	SO3	0.9	15.3
4	Aeruginosin GE 6866	(5)	Cl, Cl2, Br, Br2	Ile/Leu	---	2.2-8.5	12.8- >45.5

¹(Hanessian et al., 2006);

²(This study);

³(Ishida et al., 1999);

⁴(Murakami et al., 1995);

⁵(Matsuda et al., 1996);

⁶(Elkobi-Peer et al., 2012);

⁷IC₅₀ values reported in mg mL^-1 were converted to µM.

The absence of chlorine was thought to be responsible for the comparatively high concentration of aeruginosin 98B that was required for plasmin and thrombin inhibition (Sandler et al., 1998). However, the comparison of different aeruginosins revealed that a chlorine at the N-terminus does not significantly contribute to the inhibition of trypsin-like proteases (Table 2, groups 2 and 4). Only the chlorine of Cleu in position 2 might indicate an additional positive effect upon inhibiting proteases such as trypsin and thrombin, as indicated by the lowest IC₅₀ values for these aeruginosins (Table 2, group 1).

3.4. Toxicity of aeruginosin 828A

Aeruginosins have been long known for their inhibition potential towards trypsin-like serine-proteases, however, toxicity was not described to date (Ersmark et al., 2008; Nagarajan et al., 2013). Assessment of acute toxicity of aeruginosin 828A (1) towards T. platyurus revealed a LC₅₀ value of 22.4 µM. This value is comparable to those determined for MCs, e.g. a LC₅₀ value of 10.8 µM was measured for MC-LR (Blom and Jüttner, 2005). This clearly shows that aeruginosin 828A (1) represents another toxin in P. rubescens strain 91/1, a strain that has lost the ability to produce MCs due to a point mutation within the mcy gene cluster. Such toxicity might be also unveiled for other aeruginosins. Bioassay-guided fractionations showed strong toxicity of aeruginosins only, if they contain both, chlorine and sulfate (Fig. 1). Aeruginosins bearing only sulfate, or containing neither sulfate nor chlorine did not seem to be toxic to T. platyurus in the concentrations tested (P. agardhii strain 79, fraction 2; P. rubescens strain 91/1, fraction 3; Fig. 1).

Water-bloom forming cyanobacteria such as Planktothrix are producing several bioactive substances, most prominently MCs that are the most abundant toxins in freshwater. Besides the fact that MCs strongly inhibit the protein phosphatases 1 and 2A, the ecological function of MCs is still not clear. However, it is widely agreed that MCs reduce significantly the survival rate of grazers such as crustaceans (Blom et al., 2006; Kurmayer and Jüttner, 1999; Rohrlack et al., 1999). Blooms of cyanobacteria are typically composed of microcystin- and non-microcystin-producing strains that evolved independently (Christiansen et al., 2008). In this study, we investigated the toxicity of aeruginosin 828A that was found in a MC-deficient strain. However, MCs and aeruginosins clearly inhibit a different spectrum of enzymes pointing to a different mode of action. In future, it will be necessary to consider not only the MCs but also compounds such as aeruginosin 828A to forecast the toxicity of cyanobacterial blooms and to unravel their impact on various aquatic organisms.

4. Conclusions

In this note, we have reported the structure elucidation of aeruginosin 828A (1) isolated from P. rubescens strain 91/1, which inhibits trypsin and thrombin in the low nanomolar range. Additionally, aeruginosin 828A (1) was found to be toxic for the freshwater crustacean T. platyurus, and its toxicity was determined to be only slightly lower than MCs that are the prototypical examples of peptide toxins in cyanobacteria. This report thus suggests the presence of alternative toxins in MC-deficient Planktothrix strains and encourages further evaluation of the presence and toxicity of modified aeruginosin variants in Planktothrix strains.

Supplementary Material

Supplementary Material Available:

The NMR spectra, key NOE correlations and additional data for aeruginosin 828A (1) are available free of charge via the Internet at http://pubs.acs.org.

Fig. 2.

Trypsin and thrombin inhibition curves of aeruginosin 828A (1).