Investigations carried out on Lake Dojran, fYR Macedonia, during the spring-autumn seasons in 2015 have been focused on detecting the degree of eutrophication in the lake, successive algal flora changes in the plankton communities and eventual presence of cyanotoxins (free microcystins) in the water. The obtained results revealed a co-existence of nine Microcystis species in the lake (M. aeruginosa, M. botrys, M. flos-aquae, M. ichthyoblabe, M. novacekii, M. protocystis, M. smithii, M. viridis and M. wesenbergii), with domination of the pan and neo-tropical species M. protocystis, again confirmed in a European lake. Results also corroborate the necessity to change the accepted morphospecies concept into separation of Microcystis taxa as distinct species which are clearly delimited according to their constant morphological features. Toxicity analyses demand for a specific and targeted investigation, since the toxin production and presence depends on many factors, and the toxin dynamics including the highest peaks may be easily overlooked if other issues are in the focus of the performed monitoring. Detected values for free microcystins in the water reached 2.84 μg L–1 microcystin-LR equivalents.

Dojran Lake is situated on the border between two countries, fYR Macedonia and Greece; 2/3 of the open lake waters belonging to the former while the same amount of the catchment area belongs to the later country (Fig. 1). It is considered as a remnant of a much larger tectonic/volcanic Peonic Lake (Plio-Pleistocene) (Stojanov and Micevski, 1989).

Located at 144 m asl, Dojran Lake has been labeled as a stable eutrophic lake (Stojanovski et al., 1996) with intensive primary production (up to 760 g m–2 as periphytic growth on Phragmites spp.; Stojanov, 1986), rich with various microflora and a high fish productivity – the richest lake with fish in south-east Europe. Nevertheless, due to irrigation purposes the lake has been subjected to an intensive water outlet through deepening of the former channel in Greece (Griffiths et al., 2002; Zacharias et al., 2002) what has produced a huge oscillation and decrease in water level which has culminated in 1993, with a total water column decrease of 7 m. This situation has forced an accelerated eutrophication process in this small and quite shallow lake that has been recorded in the chemical, biological and stable isotope records of the surface sediments (Franke et al., 2013).

While the rapid decrease of water level due to abstraction combined with a period of prolonged dry conditions (1988-1998) has devastated the lake ecosystem, the anthropogenic pressure has also intensified in form of ever increasing waste water input and agricultural diffuse pollution. Consequently, the ecological status of the lake has turned rapidly towards hypertrophy, clearly reflected and documented in its microflora communities (Stojanovski et al., 1997). Namely, at the first instance in period 1988-1990 the hitherto scarce populations of the periphytic Gloeotrichia natans Rabenhorst ex Bornet & Flahault have exploded into mass domination in periphytic and planktic (in their later stages) communities with extensive biomass. This event has caused a rapid depletion of oxygen in the whole water column and a mass extinction of the bottom dwelling mollusks [Dreissena presbensis (D.polymorpha)], but also a massive succession in microflora assemblages towards domination of Asterionella gracillima (Hantzsch) Heiberg, Dolichospermum flosaquae (Brébisson ex Bornet & Flahault) P.Wacklin, L.Hoffmann & J.Komárek and Dolichospermum sigmoideum (Nygaard) Wacklin, L.Hoffmann & Komárek (Stojanovski et al., 1997). Many of detected taxa belonging to blue-green algae (like Chroococcus minutus (Kützing) Nägeli, Merismopedia glauca (Ehrenberg) Kützing, M. punctata Meyen and many others) and green algae (like various Pediastrum, Scenedesmus, Ankistrodesmus, Golenkinia, Tetraedron, Crucigenia, or desmid taxa) have disappeared from the system, while the diatom community has evidently shifted towards taxa composition indicative for higher levels of trophy and salinity. Many of the persistent and alarming scientific publications and public information (Stojanovski et al., 1997) or propositions for helping the Dojran Lake and minimizing the lethal effects of the forced ecocide (Stojanovski et al., 1995; Krstić et al., 1996) have been in vain.

For a long time, the lake Dojran has been left without any monitoring system in place and with a very doubtful attempt to bring additional water supply originating from River Vardar by means of artificial system of channels. As a result, the water level rose at 2.5 m below the optimum while the ecological parameters of the water got worse due to heavy impact of the River Vardar’s highly polluted waters. The deep layer of organic mud, at places reaching to 7 m in depth, and the continuous input of untreated waste waters have resulted in dominating cyanobacterial blooms consisted of Microcystis aeruginosa (Kützing) Kützing, Microcystis ichthyoblabe (G.Kunze) Kützing, Microcystis wesenbergii (Komárek) Komárek ex Komárek and Planktolyngbya contorta (Lemmermann) Anagnostidis & Komárek, with some remnants the previous microflora like Aphanizomenon flos-aquae Ralfs ex Bornet & Flahault, Dolichospermum scheremetieviae (Elenkin) Wacklin, L.Hoffmann & Komárek, Aphanocapsa elachista West & G.S.West or Pediastrum boryanum (Turpin) Meneghini still present in very limited numbers (Krstić, 2011). At this point the dominant cyanobacterial community has produced the highest hitherto detected dissolved microcystins concentration in a grab water sample for fYR Macedonia detected by ELISA - 270 μg L–1 microcystin-LR equivalents.

Therefore, during 2015, we aimed to investigate the intensity of the eutrophication processes and the present microalgal communities in Dojran Lake, with a special focus on the taxonomically important genus Microcystis and the morphological characters of the detected Microcystis species. The results of the determined cyanobacterial community structure, plankton species composition and successions in relation to water quality parameters, as well as the detected free microcystins concentrations in the water are presented in this study.

Filed sampling campaigns were conducted on 6th of May, 5th of July and 28th of September 2015, on three sampling sites (Fig. 1): two near shore locations in vicinity of the major settlements and one pelagic open water site near the border with Greece.

Field measurements included basic physicochemical parameters (temperature, pH, dissolved oxygen, oxygen saturation, conductivity) by means of portable equipment Senso Direct 150 multimeter, while the measurements of basic nutrients and chlorophyll a were performed according to laboratory protocol standard methods (APHA, 1998) using spectrophotometer Lovibond Tintometer®. The measurement of the basic physicochemical parameters, nutrients and chlorophyll a was performed on an integrated water sample using Ruttner water sampler. For the measurement of dissolved nutrients, the water sample was firstly filtrated through 0.45 μm membrane filters. The maximal depth was measured with weighted marked line.

The plankton material was collected by means of a planktic net (pore size 10 μm) by slowly dragging using a motor boat, or by means of vertical column mixing. Collected material was preserved in 4% formaldehyde or in Lugol’s solution for further processing in the laboratory (WHO, 1999). All plankton material was analyzed by standard light microscopy, but for better visualization of the mucilaginous envelope, the material was also stained with China Ink and analyzed in details. Microphotographs were taken both from native materials and stained colonies with Nikon eclipse E800M LM with Nikon Coolpix 4500 camera. The identification of the species was performed using standard literature (Komárek and Anagnostidis, 1998; Šejnohová and Maršálek, 2012) as well as the main report for European Microcystis morphospecies (Komárek and Komárková, 2002).

For the first and second campaign we have also performed toxin analyses. For this purpose, 50 mL of the integrated water sample were filtered in situ through glass-fiber filters 47 mm (GF/C) using vacuum filtration device. The filtrated samples were transported on cold and kept frozen at -20°C till the day of the analysis. The quantitative detection of free microcystins in the samples was performed using the Microcystins-ADDA ELISA kit from Abraxis (Product No. 520011) according to the manufacturer’s instructions. All analyses were performed in triplicate. The assay was calibrated with microcystin-LR and the results were expressed as microcystin-LR equivalents.

The results obtained from the basic physicochemical analyses of the lake’s water as well as the principal nutrients are presented on Tab. 1. These results clearly reflect the seasonal changes in the water chemistry and they also point out several important trends:

It seems that the pollution pressure was intensive, constan and prolonged during the investigated period (Tab. 1), with a clear increase of the observed parameters over the summer-autumn period. This ecological situation should have been reflected in the plankton communities and algal flora successions what has been documented in this study.

Our results concerning the planktic communities for the first sampling in May are shown in Fig. 2. On the first site, a domination of the zooplankton was recorded. The phytoplankton was composed of Microcystis species in association with Melosira varians C.Agardh. We did not detect any green algae. Six Microcystis species were determined, with co-dominance of M. protocystis Crow and M. aeruginosa. Site 2 showed almost complete domination of the zooplankton (90.6%). In the phytoplankton, high domination of Sphaerocystis sp. (probably S. schroeteri) was detected and low abundance of the genus Microcystis. Two Microcystis species were identified: M. aeruginosa and M. flos-aquae (Wittrock) Kirchner. However, we have detected domination of the phytoplankton in Site 3, with massive presence of Sphaerocystis sp. and low abundance of the genus Microcystis. Six Microcystis species were identified: M. aeruginosa, M. protocystis, M. flos-aquae, M. botrys Teiling, M. ichthyoblabe andM. novacekii (Komárek) Compère in association with Phormidium interruptum Kützing ex Forti, Aphanothece stagnina A. Braun and low presence of Pediastrum boryanum and Scenedesmus sp.

Our results from the second sampling in July are shown on Fig. 3. On Site 1 we have detected domination of the phytoplankton with co-dominance of Aphanizomenon gracile Lemmermann, M. protocystis and M. ichthyoblabe. Six Microcystis species were present on this site from which M. aeruginosa, M. botrys, M. novacekii and M. smithii Komárek & Anagnostidis were present with low abundance in association with colonies of Volvox aureus Ehrenberg. Presence of the endogloeic Pseudanabaena mucicola (Naumann & Huber-Pestalozzi) Schwabe was detected in the mucilage of Microcystis spp. The second site (Site 2) was characterized by massive domination of the phytoplankton (completely opposite from the first sampling in May), with M. ichthyoblabe as clearly dominant species, with fully developed, and sometimes macroscopic (gigantic) colonies. M. protocystis was also well developed and the rest two Microcystis species (M. aeruginosa and M. novacekii) were poorly present. Microcystis species were detected in association with Aphanizomenon gracile and Volvox aureus. Our results have also shown high presence of the endogloeic Pseudanabaena mucicola in the mucilage of the Microcystis spp. However, on the third site (Site 3), we have determined clear domination of the zooplankton (61.3%), because of the massive presence of the Dreissena sp. larvae. The phytoplankton was not dominant, but nevertheless quite rich with cyanobacterial taxa. Eight Microcystis species were identified. M. protocystis showed dominance, with Microcystis ichthyoblabe as subdominant. The rest Microcystis spp. (M. novacekii, M. aeruginosa, M. botrys, M. flos-aquae, M. wesenbergii and M. viridis (A. Braun) Lemmermann) were not abundant. The detected Microcystis species were identified in coincidence with several other cyanobacteria: Gloeothece membranacea (Rabenhorst) Bornet, Aphanizomenon gracile, Phormidium sp., Planktolyngbya contorta and Dolichospermum circinale (Rabenhorst ex Bornet & Flahault) P. Wacklin, L. Hoffmann & J. Komárek. The high presence of the endogloeic Pseudanabaena mucicola was again confirmed.

The results from the third sampling in September are presented in Fig. 4. On the first site the phytoplankton was dominant, with clear dominance of Aulacoseira granulata (Ehrenberg) Simonsen in association with massive presence of Microcystis colonies. All nine species of the genus Microcystis were present, with high abundance of M. protocystis and M. aeruginosa (always with big and well developed colonies). The rest Microcystis spp. (M.novacekii, M. flos-aquae, M. botrys, M. wesenbergii, M. smithii, M. ichthyoblabe and M. viridis) showed low abundance. The Microcystis spp. were found in coincidence with Aphanizomenon gracile and Dolichospermum circinale. The second site was characterized by complete domination of the phytoplankton, in which M. protocystis showed clear dominance, with M. ichthyoblabe as a subdominant species. Six Microcystis spp. were present, from which M. novacekii, M. aeruginosa,M. wesenbergii and M. smithii were poorly present. Green algae and the diatom Aulacoseira granulata were not detected on this site. Site 3 showed domination of the phytoplankton, with massive presence of Microcystis colonies. Seven species of the genus Microcystis were present, with massive domination of M. protocystis and big and well developed colonies of M. aeruginosa. The rest of the Microcystis species were not present in high abundance (M. smithii, M. botrys, M. novacekii, M. ichthyoblabe and M. flos-aquae). The colonies of Microcystis were in clear association with Aulacoseira granulata, as well as with Dolichospermum circinale and Aphanizomenon gracile. High presence of the endogloeic Pseudanabaena mucicola in the mucilage of the Microcystis colonies was also detected in all three sites in this campus.

In all, the obtained results from this investigation showed a distinctive pattern of plankton dynamics over the sampling period. Namely, spring samplings revealed the dominance of zooplankton over phytoplankton in near shore locations, while the pelagic waters have clear dominance of phytoplankton taxa with limited presence of six Microcystis taxa (M. aeruginosa, M. botrys, M. flosaquae, M. ichthyoblabe, M. novacekii and M. protocystis). Plankton communities collected in the second sampling campaign in July 2015 had a completely different structure than in May. If the dominance of Dreissena sp. larvae in pelagic zone is excluded (due to the period of reproduction of mussels), the plankton was fully dominated by phytoplankton communities, clearly consisted of eight Microcystis taxa (M. aeruginosa, M. botrys, M. flosaquae, M. ichthyoblabe, M. novacekii, M. protocystis, M. viridis and M. wesenbergii) where M. ichthyoblabe and M. protocystis took the dominant role, with significantly declined abundance of Sphaerocystis sp. Finally, in September 2015, M. protocystis and the well-developed colonies of M. aeruginosa have characterized the plankton community, with frequent observations of long Aulacoseira granulata chains (replacing the previously detected Melosira varians). All nine Microcystis taxa (M. aeruginosa, M. botrys, M. flos-aquae, M. ichthyoblabe, M. novacekii, M. protocystis, M. smithii, M. viridis and M. wesenbergii) were detected in the same time.

The levels of free microcystins (expressed as microcystin- LR equivalents) as well as the relations between the concentration of free microcystins and the domination of species are shown on Tab. 2.

According to Komárek & Komárková (2002) and Šejnohová & Maršálek (2012), the species concept within cyanobacteria, and especially Microcystis, is still problematic and doubtful. The wide variations in the basic differential characteristics among species, like form of the colony, mucilage structure, cell diameter, the organization and structure of cells within the colony, pigment content and life cycles have resulted in application of the taxonomic term ‘morphospecies’ which expresses significant overlap of the limiting criteria (Cronberg & Komárek, 1994). Consequently, the numerous unidentifiable colonies, atypical stages or transient forms of Microcystis have resulted in a proposition that all main morphospecies (M. aeruginosa, M. ichthyoblabe, M. viridis, M. novacekii, M. wesenbergii) should be taxonomically unified (Otsuka et al., 2001) under the Bacteriological Code of nomenclature. Additionally, according to the results provided by molecular sequencing, a lot of Microcystis species were found genetically similar, so the question if all of them should be considered as one species was raised. Nevertheless, Šejnohová and Maršálek (2012) point out that studies using molecular and biochemical criteria have been proven contradictory and should be used with caution.Having obtained the presented results on Microcystis spp. presence and dominance in Lake Dojran, we believe that new insights can be highlighted on this controversial subject – morphospecies vs species within this cyanobacterial genus. Namely, during the investigated period we have detected fine and gradual succession of Microcystis taxa in this relatively small ecosystem. In general, M. protocystis (Plate 6) and M. aeruginosa (Plate 1) have dominated in spring and autumn, while in summer M. ichthyoblabe (Plate 4) has taken the sub-dominant position, mostly well developed in the site with the highest recorded waste water pressure (Site 2, Fig. 4). In spring, a total of six Microcystis taxa were recorded (M. aeruginosa, M. botrys, M. flos-aquae, M. ichthyoblabe, M. novacekii, M. protocystis), while in summer period all nine taxa were present in the lake (M. aeruginosa, M. botrys, M. flos-aquae, M. ichthyoblabe, M. novacekii, M. protocystis, M. smithii, M. viridis and M. wesenbergii). First appearance of M. viridis and M. wesenbergii was recorded in the pelagic zone, while M. smithii was detected only in the littoral zone. It is important to emphasize that all of the nine detected taxa (Plates 1-9) have been recorded in their full development, meaning that all of the morphological criteria were visible and could be clearly and distinctly characterized without transition. These results inevitably lead to a conclusion that all nine detected Microcystis taxa have distinctive features of separate species (Komárková et al., 2005) that have not been transformed over time from another morphospecies. Their ecological preferences might be slightly different since in our study we have not detected the full development of all nine taxa until the overall domination of high eutrophic conditions and temperature during the summer period, which extended towards autumn (Tab. 1). In this context, the ecology of the different species must represent clear and quite important criterion for taxonomic classification (Komárek, 2013).

The taxonomic confusion in the line morphospeciesspecies within the genus Microcystis could have been a result of several problems associated with important biological features of this complex genus. Namely, the full development of characteristic colonies might be expected after fulfillment of the optimal environmental conditions. The symbiosis with different heterotrophic bacteria seems to be detrimental for a proper development of the protective mucilage that actually determines the colony form (Xie et al., 2016), which on the other hand is essential for the correct taxonomical determination of the species. The vegetation cycle (the stage of development and decomposition of the colonies) in the time of sampling might also have a critical influence on the final taxonomy. On the technical side, it is important always to use China Ink staining in order to properly identify some of the taxa (ex. M. botrys is frequently misinterpreted as M. aeruginosa – Sant’Anna et al., 2004).

In summary, it seems appropriate to point out the basic morphological and eco-physiological characters that ought to be included in every taxonomical work on this genus. These embrace: i) cell size; ii) pigment content and cell color; iii) mucilage structure, outline and margins; iv) form of the colonies; v) density and organization of the cells in the colony; vi) toxicity and production of bioactive peptides; and vii) ecological preferences; most importantly, it should be stressed that all of these characters usually appear in strict association in the defined types. For example, Microcystis ichthyoblabe (Plate 4) is characterized with small cells (never more than 3.5 μm in diameter), usually more or less reddish and the margin of the mucilage never exceeds the margin of the cell clusters in the colony. Contrarily, Microcystis aeruginosa (Plate 1) has always larger cells (4-7 μm), which are green to dark green, with wide mucilaginous margin around the colony. In all, we believe that the mutual co-existence of nine Microcystis taxa in the same lake in their fully developed state is an important finding in regard to the controversial opinion by the molecular biologists (Otsuka et al., 2001). This finding, as well as the observed diacritical characters that always appear in association in the defined morphotypes, should represent a strong argument against their taxonomic unification.

In the line of previous arguments, we believe that it is important to stress the co-occurrence of nine Microcystis morphospecies/species in one small and shallow lake, like Lake Dojran, which is quite rare according to the literature findings. For example, Šejnohová (2008) in her PhD thesis reports that eight morphospecies have been described in the whole Czech Republic. According to Loudiki et al. (2002), six Microcystis species are present in Morocco. Similarly, six Microcystis species were reported in São Paulo State, Brazil, after careful and several years’ biodiversity studies of planktic cyanobacteria of 14 reservoirs (Sant’Anna et al., 2004). Closest to our results is the report for five Microcystis morphotypes (M. aeruginosa, M. flos-aquae, M. ichthyoblabe, M. viridis and M. wesenbergii) in Lake Taihu, China (Hu et al., 2016).

Furthermore, we consider that our results regarding the M. protocystis dominance in Lake Dojran deserves a special attention. In our previous paper (Krstic and Aleksovski, 2016), we have reported the presence of M. protocystis for the first time in a European lake (Dojran Lake). In the present study, we have performed detailed analyses on the same ecosystem, and the presence of this species was again confirmed with a marked dominance. M. protocystis was originally described by Crow (1923) and the type locality is Sri Lanka, freshwater plankton of the inland waters of Ceylon. Later, this species was characterized as pan-tropical (Komárek & Anagnostidis, 1999; Komárek & Komárková, 2002) and probably one of the commonest pan-tropical species. Till now, this species was reported as a true element of plankton in karst environments of the Yucatán peninsula, Mexico (López- Adrián & Barrientos-Medina, 2007; Tavera et al., 2013) which are rich in tropical species of cyanobacteria. This species was also reported on several localities in Brazil: in São Paulo State (Sant’Anna et al., 2004), in the Cordeiro, Camalau and Acauã Reservoirs (Vasconcelos et al., 2013) and in Minas Gerais State (Magalhães et al., 2014). It was also found in the neo-tropical, eutrophic reservoir Riogrande II in Colombia (Palacio et al., 2015), but still, up to our knowledge, it was never reported for Europe. However, in one of these papers the authors suggest that the sparse cell disposition gives M. protocystis the aspect of old or senescent colonies of different Microcystis species and maybe, for this reason, M. protocystis distribution should be wider than normally referred in the literature (Sant’Anna et al., 2004). It seems obvious that Dojran Lake is experiencing the accelerated eutrophication processes (Svirčev et al., 2014) due to excessive nutrient input, oxygen depletion and the intensive decomposition on the lake’s bottom. Additionally, the evident dominance of this pan/neo-tropical species characteristic for Brazil, Columbia and Sri Lanka in a European lake is an important finding which might have a valuable contribution for the global climate change investigations and may be important for the future monitoring programs.

Last but not least, in the list of arguments on the morphospecies/ species subject, the issue of toxicity of the specific taxa becomes both taxonomically and ecologically very important. During our presented investigations, the focus of the research has been pointed to the taxa successions and dominance rather than on the toxicity itself (Tab. 1). The obtained values for this period are quite lower than the values reported in a previous study (Krstić, 2011) when a maximum value of 270 μg L–1 has been detected. Detected values for Lake Dojran around 2-5 μg L–1, similar to the findings in our research, have been also reported in some previous studies (Papadimitriou et al., 2010; Gkelis et al., 2015), what underlines the continuity of predominantly microcystin group of toxins presence in the lake. These discrepancies in detected toxin levels for Lake Dojran, or any other ecosystem, would be primarily a result of time of sampling and analyses. As documented in this research, several Microcystis taxa dominance successions have been recorded over the investigated period, meaning that taxa with low/high toxin production potency have changed their dominance and consequently the final toxin (microcystins) concentration in water. Also, some of the dominant taxa like M. protocystis, or sub-dominant M. ichthyoblabe, have not been hitherto reported to produce any toxins or produce other toxins than microcystins, usually neurotoxins. Thus, the detection of toxins in any ecosystem has to be a specific scientific task which demands a toxin orientated sampling frequency, usually with much denser sampling activities, and also analyses of a broad spectrum of toxins in order to reveal the real toxin dynamics and dominance in an ecosystem. In light of previous findings (Krstic, 2011), it is possible that Dojran Lake undergoes several periods of very high toxin presence (peaks) over the year which may last for couple of days only, and then the ecosystem falls back to the basic microcystins presence of only a few μg L–1. However, the limitations of the ELISA technique should also be considered (mostly used as a screening tool and without possibility to identify the different microcystin variants) as well as our primary focus on extracellular toxins, so future investigations with complementary methods (e.g., LC-MS) are foreseen.

Having in mind the importance of previous arguments, it seems appropriate to list some of the basic toxin properties of dominant taxa detected in Dojran Lake. Also, the toxin potency and genetic ability of different taxa make the argument on morphospecies even more obsolete since it is very hard to imagine that a taxon genetically equipped for a toxin production will ‘transform’ into a taxon not capable of any toxin production, or vice versa. These basic properties are:

In conclusion, seasonal investigations on Lake Dojran during 2015 on presence, dominance, successions and toxin production of Microcystis spp. have enabled the following statements: