Pilot search for cylindrospermopsin-producers in nine shallow Bulgarian waterbodies reveals nontoxic strains of Raphidiopsis raciborskii, R. mediterranea and Chrysosporum bergii

Abstract

This paper presents results from the first application of molecular-genetic analysis combined with conventional light microscopy (LM) for identification of cylindrospermopsin (CYN) producers in field phytoplankton samples from Bulgaria. In total 68 cyanoprokaryotes were found by LM, out of which 18 were considered potential CYN-producers according to the literature. They occurred in different abundance (0-40%) in five waterbodies. The presence of CYN-producers was sought by application of the primer pair cynsulfF/cylnamR specific for the sulfotransferase gene (cyrJ) considered the best genetic marker for detection of CYN-toxigenic strains. Although CYN was detected by enzyme-linked immunosorbent assay (ELISA) in two waterbodies (lake Vaya and reservoir Mandra), we did not find cyanoprokaryotes with cyrJ gene. Therefore, Raphidiopsis raciborskii, despite being identified by LM in both waterbodies where CYN occurred, was not the producer of this toxin and belonged to the European nontoxic population of the species. The same was true for the less widely spread Raphidiopsis mediterranea and Chrysosporum bergii (both found in small amounts only in lake Vaya). Our results on CYN occurrence without the assignment of its producer in the collected samples suggest that: 1) CYN remains stable in water after disappearance of the toxigenic species, or, more probably, 2) besides the already known species, there are others that are associated with the production of CYN. Because revealing of novel CYN producers is of ultimate importance for health risk assessment and water management, we propose more analyses to be performed in future studies.

Keywords:

• algae
• cyanobacteria
• cyanoprokaryotes
• cyanotoxins
• phytoplankton
• toxins

Introduction

Cyanoprokaryota/Cyanobacteria is a group of microscopic photoautotrophs widely distributed in different habitats. However, mainly in aquatic environments they have become of great concern for both human and ecosystem health, and for drinking water management in particular because of their ability to form dense blooms, which decrease the water quality, and to produce different toxins. These toxins, named cyanotoxins, are commonly classified as hepatotoxins (microcystins, nodularins), neurotoxins (anatoxins, saxitoxins), cytotoxins (cylindrospermopsin) and dermotoxins [1, 2]. In the past two decades, the knowledge on cyanotoxins has advanced significantly, with increasingly documented data on their findings and threats as enhanced by the climate changes with global warming, rapidly growing human population and eutrophication [3, 4]. So far, only the specific group of microcystins (MCs) can be outlined as routinely analyzed in the monitoring of recreational and drinking waters [5]. However, another toxin, cylindrospermopsin (CYN), is gaining strong attention currently. The growing interest in CYN is explainable by its high liberation to surrounding water, which poses more hazards to water users and is important for water management [6]. Moreover, CYN is an alkaloid with cyto-, hepato-, neuro- and genotoxic properties and potential carcinogenic effects [7–10], which has also been linked to death of domestic animals [11]. Besides the liver, CYN can affect other organs, such as the eyes, kidneys, lung, heart, thymus, spleen, adrenal glands and the intestine [12–19]. Probably, because of its genotoxic and mutagenic properties, CYN is even more hazardous to human and animal health than MCs [18, 20]. The toxicity of CYN has been associated mainly with protein synthesis inhibition and oxidative stress [10] but other mechanisms could be involved, including indirect effects after some metabolic modifications. Human exposure to CYN may occur during recreational activities (by dermal contact during bathing) and by consumption of contaminated drinking water or food (vegetables, fish, crustaceans, etc.) (for details see [10]). Therefore, a provisional Tolerable Daily Intake (TDI) of 0.03 µg kg⁻¹ body weight (b.w.) and a guideline value of 1 µg L⁻¹ in drinking water were proposed for CYN [21]. At present, five analogues of CYN are known [22].

In Bulgaria, there have been records of the occurrence of potentially toxic cyanoprokaryotes and their blooms in inland and coastal wetlands since the beginning of the twentieth century (for details see [23, 24]), but it was in 2006 when the presence of cyanotoxins (mainly MCs) was proved [25]. Much later, in 2018, CYN was detected for the first time in Bulgaria in two large, but shallow waterbodies situated at the Black Sea coastline: reservoir Mandra and lake Vaya [26]. The purposive analysis for this cyanotoxin was based on the knowledge of the increased spread in the country of the first species reported as a CYN-producer [7]: Raphidiopsis raciborskii (Woloszynska) Aguilera, Berrendero Gómez, Kastovsky, Echenique & Salerno (Syn. Cylindrospermopsis raciborskii (Woloszynska) Seenayya & Subba Raju) [23, 24, 27–36]. It is broadly accepted that this tropical species benefited from climate changes and became globally invasive with actual multi-continental spread [37–42], which strongly suggested its current classification as a cosmopolitan [42]. Therefore, R. raciborskii has been long considered as a major problem for water management on a global scale [40].

In the context of climate changes with global warming, more studies of the distribution and toxicity of both native and invasive species became necessary worldwide [43] and by years more cyanoprokaryotes capable of CYN-production were identified: Anabaena lapponica Borge, Aphanizomenon flos-aquae Ralfs ex Bornet & Flahault, Aphanizomenon gracile Lemmermann, Chrysosporum bergii (Ostenfeld) E. Zapomelová, O. Skácelová, P. Pumann, R. Kopp & E. Janecek (Syn. Anabaena bergii Ostenfeld), Chrysosporum ovalisporum (Forti) E. Zapomelová, O. Skácelová, P. Pumann, R. Kopp & E. Janecek (Syn. Aphanizomenon ovalisporum Forti), Dolichospermum mendotae (W. Trelease) Wacklin, L. Hoffmann & Komárek, Dolichospermum planctonicum (Brunnthaler) Wacklin, L. Hoffmann & Komárek (Syn. Anabaena planctonica Brunnthaler), Microseira wollei (Farlow ex Gomont) G. B. McGregor & Sendall ex Kenins (Syn. Lyngbya wollei (Farlow ex Gomont) Speziale & Dyck), Oscillatoria sp. PCC6506, Raphidiopsis curvata F. E. Fritsch & M. F. Rich, Raphidiopsis mediterranea Skuja, and Umezakia natans M. Watanabe (for details see [5, 22, 44–46]). A proposal to focus on Aphanizomenon klebahnii Elenkin ex Pechar as a putative CYN producer was also published [47]. Besides the above-mentioned species, a few other potential CYN-producers have remained identified only at the genus level, e.g. Aphanizomenon sp., Anabaena sp. [48, 49]. Similarly to R. raciborskii, some of these potential CYN-producers were broadly distributed in the phytoplankton of more than 100 Bulgarian waterbodies during the period 2000-2015: Aphanizomenon flos-aquae, Aphanizomenon gracile, Aphanizomenon klebahni, and R. mediterrannea [23]. However, up to now, there has been no molecular-genetic investigation to prove the production of CYN by any of these species. Such a study is necessary because most of the known CYN-producing species have both toxic and nontoxic strains [22]. There are a growing number of reports of CYN occurrence worldwide without assignment of the causative agent. Thus, it is possible to suggest that, besides the species already known, there are others that are associated with CYN production [9, 48]. The suggestion that the list of potential CYN-releasing cyanoprokaryotes probably still remains incomplete also comes from the accumulating evidence that the production of CYN is not species specific [22].

The importance of revealing all species responsible for CYN production is strongly related to the risk management, especially considering the fact that the blooms of R. raciborskii generally appeared to be non-toxic in Europe [40, 50–52]. It was even assumed that CYN-producing strains of this species do not appear in water bodies of temperate climate [53]. Data on species phylogeography, spreading routes and toxicity are quite challenging [42] because R. raciborskii can produce at least two types of potent toxins, CYN or saxitoxins (STX), with biogeographical differences between strains [54]. There is a general indication that American strains are able to produce STX but not CYN, Australian strains produce CYN but not STX, only some Asian strains are CYN-positive, whereas African strains are non-toxic (for details see [54]). In Europe, SXT production by R. raciborskii was also reported [55]. In addition, a possibility for this species to produce a third type of toxin, MC, could be supposed because its MCs-positive strain was isolated from Greece [56]. Although abundant data have accumulated, in some way the occurrence of the toxic strains of R. raciborskii has remained underestimated [42]. It is believed that at present state-of-art studies of this species related to CYN production, especially in Europe, are essential [57]. Therefore, in this study, a purposive molecular-genetic analysis based on the specific cynsulfF/cylnamR primer pair (with cyrJ gene as a genetic marker for detection of CYN-producing strains [22, 58]) was conducted on the phytoplankton material from nine shallow Bulgarian waterbodies considered as threatened by harmful algal blooms (HABs) [23, 25, 59].

Materials and methods

Sites and sampling

The study was carried-out in June 2018 in nine shallow waterbodies situated in Central and Eastern Bulgaria (Table 1, Figure 1). Detailed descriptions on the morphometry, historical development, usage, conservational status and biodiversity of each of the waterbodies are provided in the Inventory of Bulgarian wetlands and their biodiversity [60]. Therefore, in Table 1 the unique inventory number of each waterbody from this database (IBWXXXX) is provided.

Figure 1. Map of Bulgaria showing the sampling sites (modified after http://www.ginkgomaps.com and Google Maps, accessed 6 November 2019).

Table 1. Sampling sites and their environmental parameters with types of cyanotoxins found in Bulgarian waterbodies (WBs) (after [26,59]): WBN – name of the water body; IBW – number in the Inventory of Bulgarian Wetlands [60]; SAN – site abbreviation and number; Alt – altitude; WT – water temperature (° C), SD – Secchi depth (m); TTB – total transparent to bottom; CND – conductivity (µS); TDS – total dissolved solids (µg L^-1); DO – oxygen concentration (mg L^-1); TP – total phosphorus (mg L^-1); TN – total nitrogen (mg L^-1); CT - cyanotoxins, MCs – microcystins, CYN – cylindrospermopsin, SXT – STXitoxins.

WBN and IBW	SAN	Alt	Latitude	Longitude	WT	pH	SD	CND	TDS	DO	TP	TN	CT
Reservoir Sinyata Reka (IBW1890)	SR1	317	42^o28.1480'	24^o42.217	27.4	9.72	0.5	470	305	9.36	25.00	4.8	MCs
	SR2	317	42^o28.1473'	24^o42.2175	26.7	9.36	0.6	468	306	9.21	27.00	4.3
Vaya Lake (IBW0191)	VA1	−2	42^o30.5940'	27^o22.075	26.9	9.65	0.25	2588	1682	12.51	13.00	5.4	CYN
	VA2	0	42^o28.4540'	27^o25.482	28.28	8.86	0.25	1183	768	11.94	11.00	3.7
	VA3	6	42^o29.1850'	27^o26.531	23.7	9.5	0.25	1024	665	7.01	12.00	4.6
Reservoir Mandra (IBW1720)	MN1	12	42^o24.0463'	27^o26.1120'	25.88	8.28	0.4	649	421	6.81	3.00	3.0
	MN2	13	42^o24.0670'	27^o19.1310'	26.2	8.2	0.2	663	461	5.89	6.00	4.0	CYN
	MN3	9	42 26.1420'	27^o26.5860'	24.9	8.48	0.3	639	415	7.91	4.00	3.3
Uzungeren Lake (IBW0710)	UZ	7	42^o26.1782'	27^o27.1998'	25.9	8.06	0.4	14.38	9351	7.83	5.00	2.8
Reservoir Poroy (IBW3038)	PR	41	42^o43.0190'	27^o37.3160'	25.10	8.33	1.2	762	495	9.45	1.00	2.8
Reservoir Aheloy (IBW3032)	AH	144	42^o42.8230'	27^o30.9740'	25.4	8.51	1.10	614	399	8.92	1.00	4.1
Ezerets (IBW0233)	EZ	−2	43^o35.2770'	28^o33.2290'	26.4	8.35	TTB	1084	0	9.94	0.50	5.3
Shabla (IBW0219)	SH	−2	43^o33.8180'	28^o34.1860'	27.1	8.46	TTB	1087	0706	9.97	0.10	5.1
Durankulak Lake (IBW0216)	DR1	6	43^o40.3240'	28^o32.0470'	24.03	8.54	1	1111	722	7.35	21.00	2.8
	DR2	6	43^o40.3340'	28^o32.0220'	24.7	8.21	1	1094	711	7.79	20.00	4.0	MCs, SXT
	DR3	4	43^o40.5300'	28^o32.9930'	24.6	8.49	1	1075	698	6.19	24.00	3.9	MCs
	DR4	3	43^o40.6950'	28^o32.6000'	26.5	8.53	1	1087	706	9.6	20.00	3.2

The sampling was preceded by sending a drone (DJI Mavic Pro, Model: M1P GL200A) supplied with a photo camera to observe and document the whole waterbody and potential hot spots with visible differences in the color as indicators of cyanoblooms. The spots/areas of different color were chosen for sampling or, in case of visible water homogeneity, the sites from our previous studies were repeated for each waterbody (for details see [26]). All the chosen 17 sites were reached by inflatable boats. The site coordinates, altitude, water temperature, pH, water hardness (TDS), oxygen content (DO) and conductivity were measured in situ by Aquameter AM-200 and Aquaprobe AP-2000 from Aquaread water monitoring instruments, 2012 Aquaread Ltd. Total nitrogen (TN) and phosphorus (TP) were measured ex situ using Aqualytic AL410 Photometer from AQUALYTIC®, Dortmund, Germany. The water transparency was measured using Secchi disk. All results, together with detailed data on cyanotoxins found were published [26, 59].

Phytoplankton samples for taxonomic identification and for molecular-genetic studies (each in a volume of 0.5 L) were collected from the water surface (0–20 cm). The samples for light microscopy processing were fixed immediately with 2% formalin and thus transported to the lab, where they were further processed by sedimentation method [26, 59]. The samples for polymerase chain reaction (PCR) studies were filtered within several hours after collection, and the obtained filters, kept in sterile plastic tubes, were transported to the lab in a box with dry ice [26, 59].

Phytoplankton identification by conventional light microscopy (LM)

In the lab, the microscopic work was done mainly under magnification 100x and immersion on 52 non-permanent slides on a Motic BA 4000 microscope with a Moticam 2000 camera, and later on 22 non-permanent slides on a Motic B1 microscope with a Moticam 2.0 mp camera. Both cameras were supplied by Motic Images 2 and 3 Plus software program, respectively [59].

Taxonomic identification of cyanoprokaryotes followed the standard European taxonomic literature [61–65] with updates from AlgaeBase [66], CyanoDB [67] and relevant modern taxonomic papers. Traditional morphological features for distinguishing nostocalean taxa (to which most CYN-producers belong) include the cell form and size, heterocyte shape and size, akinete shape and size in combination with their position in the trichomes in relation to heterocytes, and facultative or obligatory occurrence of aerotopes [61–65].

Quantitative contribution of the different species to the phytoplankton biomass was estimated using the Thoma-counting chamber and the method of the stereometrical approximations [23, 68].

Molecular studies

The molecular study was conducted in order to detect CYN-producing strains or species based on the presence of the sulfotransferase gene (cylJ) using a primer set specific for this gene, cynsulfF/cylnamR [58]. The targeted gene is a part of the gene cluster, 43 kb in size, which encodes amidinotransferase, peptide synthetase (PS), polyketide synthase (PKS) and tailoring enzymes involved in CYN production [69]. The cluster has been described in R. raciborskii [58], R. curvata [70], Aphanizomenon sp. [49] and Oscillatoria sp. [44]. The primer set used in this study is specific for the strains of R. raciborskii, R. curvata, R. mediterranea and Microseira wollei. The expected length of amplified fragments is 585 bp [58].

Water samples were collected from the nine water bodies described above and a few hours afterwards were filtered through 45 µm cellulose filters Whatman NC45 ST/Sterile EO. The filters were kept under -20 °C and used for total DNA extraction. The protocol of Genomic DNA Purification Mini Kit (Sigma) was followed. DNA was amplified using primer combination cynsulfF (5’ACTTCTCTCCTTTCCCTATC3’) x cylnamR (5′ GAGTGAAAATGCGTAGAACTTG3’) according to [58]. The PCR amplification was performed in a 25 µL volume containing 10 pmol primers; 0.16 mmol L⁻¹deoxynucleoside triphosphates dNTPs; 1.25 units of Taq polymerase, 0.75 mmol L⁻¹ MgCI₂ and 10xPCR buffer supplied by FastStart High Fidelity PCR System (Roche). The amplification of DNA was done in a thermal cycler QB-96 (Qianta Biotech) under the following PCR conditions: denaturation at 95 °C for 5 min, 35 cycles of denaturation (30 s at 95 °C), annealing at 55 °C for 30 s, extension at 72 °C for 40 s, and a final extension at 72 °C for 5 min. The resulting PCR products around 500 bp in size were purified using Sigma Gel GenElute Gel Extraction Kit (Sigma), following the manufacturer’s instructions. The amplification products obtained were subsequently cloned using a CloneJET PCR Cloning Kit (Thermo Scientific). Recombinant plasmids were isolated using Sigma Plasmid Miniprep Kit. For each sampling site six clones were selected and sequenced by Macrogen Inc. The obtained sequences were processed with Vector NTI 11.5 software and used for search by Basic Local Alignment Search Tool (BLAST) [71] in the National Centre for Biotechnology Information (NCBI) data base [72].

Results and discussion

Phytoplankton species composition, obtained by light microscopy (LM)

In total, more than 240 species from different taxonomic groups were identified using LM in the phytoplankton of all studied waterbodies. The reliably determined cyanoprokaryotes were 68, or 28% of all taxa. Considering CYN-producers as target of the study and the results from the PCR analysis provided below, here we report only the presence of genera and species known as toxigenic for CYN and their close relatives in the studied waterbodies.

Using traditional morphological diagnostic features [61–65], we identified 18 taxa which could be possibly associated with CYN-production (Table 2). Among them, there were three species from the genera Raphidiopsis (2) and Chrysosporum (1) well-known for their ability to produce CYN. We found also one representative of the CYN-producing genus Raphidiopsis (unidentified to species level because of the lack of heterocytes and akinetes) and a representative of the genus Oscillatoria unidentified to the species level, from which a benthic strain PPC 6506 positive for CYN production was reported [73]. The other fourteen taxa have not been listed among CYN-producers, but were taken into consideration and included in Table 2 because of their close relations or belonging to the genera Anabaena/Dolichospermum and Aphanizomenon/Chrysosporum, from which CYN-positive strains were found, as already described in the Introduction.

Table 2. Distribution of potential CYN-producing taxa in the studied Bulgarian waterbodies: VA – lake Vaya, MN - reservoir Mandra, PR – reservoir Poroy, DK – lake Durankulak, UZ – lake Uzungeren, SR – reservoir Sinyata Reka; d – dominance (>25% of the biomass); a – abundance (5–25% of the biomass); x – occurrence (<5% of the biomass); r - rare occurrence in single trichomes. Bright grey lines indicate the species which have been reported as CYN producers (for details see the text). Taxa are enlisted in alphabetical order.

Taxon/Waterbody	VA	MN	PR	DK	UZ	SR
Anabaena cf. oscillarioides Bory ex Bornet & Flahault	x
Anabaena cf. torulosa Lagerheim ex Bornet & Flahault	x
Anabaena sp. ster. 1 (straight trichomes, different heterocytes)	r
Anabaena sp. ster. 2 (spirally coiled trichomes)	x					r
Anabaena sp. ster. 3 (straight short trichomes, no gas vesicles)	r
Aphanizomenon cf. yezoense M. Watanabe				a
Chrysosporum bergii (Ostenfeld) E. Zapomelová, O. Skácelová, P. Pumann, R. Kopp & E. Janecek	x
Cuspidothrix issatschenkoi (Usachev) P. Rajaniemi, Komárek, R. Willame, P. Hrouzek, K. Kastovská, L. Hoffmann & K. Sivonen (Syn. Aphanizomenon issatschenkoi (Usacev) Proshkina-Lavrenko)	r	r	r
Cuspidothrix tropicalis (Horecká & Komárek) P. Rajaniem, J. Komárek, R. Willame, P. Hrouzek, K. Kastovská, L. Hoffmann & K. Sivonen (Syn. Aphanizomenon tropicalis Horecká & Komárek)	r		r	x
Dolichospermum crassum (Lemmermann) P. Wacklin, L. Hoffmann & J. Komárek (Syn. Anabaena crassa (Lemmermann) Komárková-Legnová & Cronberg)	x
Dolichospermum compactum (Nygaard) P. Wacklin, L.Hoffmann & J. Komárek (Syn. Anabaena compacta (Nygaard) Hickel)	a
Dolichospermum flosaquae (Brébisson ex Bornet & Flahault) P. Wacklin, L. Hoffmann & J. Komárek (Syn. Anabaena flosaquae Brébisson ex Bornet & Flauhault)	a
Dolichospermum perturbatum (H.Hill) Wacklin, L.Hoffmann & Komárek (Syn. Anabaena perturbata H.Hill, Anabaena flosaquae f. aptekariana Elenkin, Anabaena spiroides var. tumida Nygaard)	d			x
Dolichospermum sp. ster. (spirally coiled trichomes)	x
Oscillatoria sp.	x
Raphidiopsis mediterranea Skuja	x
Raphidiopsis raciborskii (Woloszynska) Aguilera, Berrendero Gómez, Kastovsky, Echenique & Salerno (Syn. Cylindrospermopsis raciborskii (Woloszynska) Seenayya & Subba Raju, Anabaena raciborskii Woloszynska)	x		a		a
Raphidiopsis sp. (initial trichomes without heterocytes and akinetes)				x

The differences in the distribution of potential CYN-producers by waterbodies are shown in Table 2. CYN-producers had different quantitative role in the total phytoplankton biomass: ca. 40% in lake Vaya and reservoir Poroy, ca. 20% in lake Durankulak and reservoir Mandra, 4% in lake Uzungeren, <1% in reservoir Sinyata Reka, and were not found in the lakes Shabla and Ezerets, and in reservoir Aheloy as well.

Results from PCR analysis for cylindrospermopsin-producing strains

The sulfotransferase (cyrJ)-gene-specific primer pair was used to identify CYN-producing genotypes in the phytoplankton samples from 17 studied sites of nine Bulgarian waterbodies. As a result, PCR fragments ca. 500 bp in length were obtained from the lakes Durankulak (sites 1 and 2) and Vaya (sites 1 and 2), and from the reservoirs Mandra (site 2) and Poroy. Аlthough the amplified fragments were shorter than the expected length of 585 bp, all PCR products were extracted from the gel and cloned. From each site six individual clones were sequenced. The results obtained showed no homology with the annotated sequences from CYN-producing cyanobacteria available in the NCBI [72].

Discussion on the results obtained by light microscophy and genetic-molecular methods

The sulfotransferase gene (cyrJ) required for tailoring the reaction to complete the biosynthesis of the CYN [58] was applied to assess the toxigenic potential of 17 water samples collected in nine shallow Bulgarian lakes and reservoirs. In four of them (Vaya, Mandra, Durankulak and Uzungeren) three well-known potential CYN producers were identified by LM: Raphidiopsis raciborskii, R. mediterranea and Chrysosporum bergii (Table 2), and in two of them (Vaya and Mandra) the presence of CYN at the lowest detection limit (0.4 µg L⁻¹) was discovered [26]. However, in all investigated samples, no amplification of the sulfotransferase gene was found.

On the background of the results obtained by LM, this absence of the cyrJ gene, which is considered a good marker for CYN toxicity [22, 45, 58, 74, 75], clearly shows that the first species shown to be a CYN-producer, Raphidiopsis raciborskii [7], is practically incapable of CYN production in the studied Bulgarian waterbodies. This suggestion is additionally supported by the different distribution of this species in both waterbodies, Mandra and Vaya, where CYN was detected ([26], Table 1): it was never found in reservoir Mandra ([34], Table 2) but since 2000 it has been developing periodically forming water blooms in lake Vaya [31, 33–35]. During this study it was found to occur there but not as a dominant or abundant species (Table 2). Moreover, R. raciborskii was more abundant in two other waterbodies, reservoir Poroy and lake Uzungeren (Table 2), where CYN was not detected ([26], Tables 1 and 2). The absence of cyrJ and hence of the ability to produce CYN is in accordance with the earlier opinion of Stoyneva [34] that R. raciborskii appeared in Bulgaria from the European populations along the Danube and is in accordance with the generally accepted idea for geographically distinct strains of this species spread in the world.

During the studied period Chrysosporum ovalisporum was not identified in the samples either by LM, or by molecular-genetic methods. This species is regarded as a major source of CYN not only in Australia, the United States (Florida) and Israel, but also in Mediterranean Europe, where it can be considered the main CYN producer to date, which has a great spreading potential (for details see [9]). We note especially the absence of this well-known CYN-producer in our samples because it once more confirms the need to identify the real CYN causative agents in the studied Bulgarian waterbodies. With the same consideration we have to point out also the absence of all other known CYN producers [5, 22, 44–46]: Aphanizomenon flos-aquae, Aphanizomenon gracile, Anabaena lapponica, Dolichospermum mendotae, Dolichospermum planctonicum, Microseira wollei, Oscillatoria sp. PCC6506, Raphidiopsis curvata and Umezakia natans.

Besides the powerful cyr genes [53, 58] the activity of other genera and species in the production of CYN was previously observed as based only on identification of aoaA gene encoding an amidinotransferase [76], or cyrA/aoaA gene [49], on ps gene [48] or on pks/ps genes putatively associated with the CYN production [58, 77–79]. Therefore, it is possible to suppose that the detected CYN production [26] sought in the present study only by cyrJ, was caused by strains and species bearing other genes, so further PCR investigations have to be performed to reveal the real CYN producers.

At present, by the primer applied, we cannot assign to any species or strain the CYN recorded in low amounts in the coastal waterbodies Vaya and Mandra [26]. Here, we have to mention also that still relatively little work has been done on CYN-detection methods [22]. CYN in the studied waterbodies was discovered by ELISA [26], which was confirmed as a good screening method for CYN with a sensitive level of detection [47]. However, there are some opinions that ELISA is nonselective for CYN analogues and that cross-reactivity may exist [45, 47, 77].

The discrepancy which at first glimpse occurs between the molecular-genetic results, which showed no presence of known CYN-producers, and our ELISA data, which indicated the presence of CYN in Mandra and Vaya [26], needs more discussion. One possible explanation could be the effect of environmental conditions on CYN production. This effect has been only partly resolved, and optimal conditions for toxin production may not be optimal for the growth of the CYN producer [22]. Although laboratory experiments have shown effects of temperature, nitrogen source, phosphate availability and light intensity on CYN production [80–85], it was supposed that the influence of environmental conditions on CYN production may vary among different producing species and strains, especially those isolated from different geographical regions [22, 86, 87]. Therefore, considering also the strong extracellular release of CYN (which can reach up to 90% of the total toxin) [22], it is possible to suppose that the strains responsible for CYN production were developed in very low amounts during the sampling period. Taking into account the stability of CYN in waters [22, 88, 89], we cannot completely exclude the possibility that the low amounts of CYN found by us in Mandra and Vaya [26] represented remnants from former blooms.

Overall, our results are in accordance with other studies where the source of CYN remained uncertain due to still scarce knowledge of organisms able to produce it [47, 57, 90–92]. CYN occurrence cannot always be clearly linked with the presence of its putative producing species in the waterbodies (e.g. [77, 91–95]). For example, despite several indications of R. raciborskii occurrence in European lakes, its strains had a minor contribution to the CYN production [47] or were not found to be involved in it [92, 96, 97], with the exception of one strain from Hungary [98]. Similar results were obtained for Aphanizomenon flos-aquae: despite being the only reliably identified CYN producer in German waters [48, 94], its presence could not always explain the occurrence of the toxin [94]. Therefore, an explanation was sought in the co-occurrence of CYN-producing and non-producing strains of the same taxon or in the presence of additional, so far unidentified CYN-producing species, most likely of the genera Aphanizomenon/Chrysosporum or Anabaena/Dolichospermum [45, 94, 99]. Here we would like to underline also the occurrence of CYN in relation with Microcystis and Microcystis-dominated blooms in Greek freshwaters [55, 100]. Considering these data, we have to note that seven Microcystis species were detected by our team in the studied waterbodies during the same investigation period and in the same samples [59]. Therefore, in the proved absence of cyrJ-bearing cyanobacteria but detected CYN [26], we should not exclude the possibility to find some link between presence of Microcystis and CYN-production in the coastal waterbodies Vaya and Mandra.

Conclusions

The results obtained from application of LM combined with molecular-genetic studies allowed us to determine that the most commonly suspected species, the invasive Raphidiopsis raciborskii, was not the CYN-producer in the studied shallow Bulgarian waterbodies. This shows its general affiliation to the European non-toxic gene pool of this species, confirming also the wide-spread opinion that CYN synthesis appears to be linked to biogeographic patterns. The lack of positive signal for the cyrJ gene demonstrated also the nontoxic character of the other potential CYN-producing species found during the study, Chrysosporum bergii, Raphidiopsis mediterranea and Raphidiopsis sp. By application of cyrJ (based on the single cynsulfF/cylnamR primer set) we could not identify any other species responsible for the presence of CYN in two of the waterbodies (Vaya and Mandra). The occurrence of fourteen other species closely related to the known CYN-producers and identified by LM shows that further more complex research based on genetic analyses and cyanobacterial cultures should be performed in order to reveal the CYN-producers in Bulgarian waterbodies.

Author contributions

Conceptualization and writing - MSG, GG and BU; sampling – BU, GG, MSG, MR; formal analysis and investigation of molecular genetic data - KS and MR; investigation by light microscopy - MSG, BU, and GG; funding acquisition – BU, MR, GG; project administration - BU.

Acknowledgments

The authors would like to acknowledge SRF-MESB for project granting and the European Cooperation in Science and Technology, COST Action ES 1105 ‘CYANOCOST - Cyanobacterial blooms and toxins in water resources: Occurrence, impacts and management’ for adding value to this paper through networking and knowledge sharing with European experts and researchers in the field.

Disclosure statement

No conflict of interest was reported by the author(s).

Additional information

Funding

This research was funded by Scientific Research Fund of the Ministry of Education and Science of Bulgaria (SRF-MESB) KP-06-OPR03/18/19.12.2018.