Co-occurrence of anatoxin-a and microcystins in Lake Garda and other deep subalpine lakes

Abstract

Cyanotoxins are a global concern in freshwaters and eutrophication and climate changes can have synergistic effects in exacerbating the problem. The deep perialpine lakes are a group of lakes of huge economic and naturalistic importance located at the border of the Alps. At the southern border of the Italian and Swiss Alps, the largest waterbodies include the lakes Garda, Iseo, Como, Lugano and Maggiore (Deep Subalpine Lakes, DSL). Together with eutrophication (during the 1960s and 1970s) and re-oligotrophication (from the 1990s onward) these lakes have been experiencing warming and increase of the water column stability. These changes had a strong impact on the phytoplankton (including cyanobacteria) community. Four DSL (lakes Garda, Iseo, Como and Lugano) have been studied with the aim of comparing their toxic potential. For one of them (Lake Garda) an 8 years survey was conducted, allowing a long-term trend analysis. Toxin analysis was conducted on a monthly basis by targeted LC-MS/MS. A screening for anatoxins, cylindrospermopsins, saxitoxins, microcystins (MCs) and nodularins was carried out. Among all the listed toxins, only one anatoxin and five MCs were detected in the lakes. In particular, the alkaloid anatoxin-a (ATX) was found dominant in lakes Garda, Iseo and Como, and absent in Lake Lugano; the MC-[D-Asp3]RR was found as the most abundant MC in all four lakes. Four other less abundant MCs were also found. The two major toxins are produced by two different cyanobacteria, Tychonema bourrellyi (J.W.G. Lund) Anagnostidis & Komárek and Planktothrix rubescens (De Candolle ex Gomont) Anagnostidis & Komárek, which share however a number of ecological traits. Peaks of these toxins occurred in warmer months (typically between May and September) in the thermocline layer (around 20 m, in the considered lakes). In summer 2016, the highest concentrations of ATX and total MCs were registered in Lake Iseo (1100 and 430 ng L–1, respectively), while in the other lakes values were approximately twice lower. In the lakes where it was present, ATX peak levels were much higher than MCs, thus highlighting the necessity of including ATX in the procedures of risk assessment. The importance of ATX is expected to further grow in the future with respect to MCs, as demonstrated by the long-term trend analysis carried out in Lake Garda that showed a clear decline for MCs from 2009 till 2016 and a relative constancy of ATX. No n-c om me rci al us on ly L. Cerasino and N. Salmaso 12 rubescens (De Candolle ex Gomont) Anagnostidis & Komárek (MCs producer) was observed, which was partially replaced by Tychonema bourrellyi (J.W.G.Lund) Anagnostidis & Komárek (ATXs producer). The two cited toxic species are part of the cyanobacterial populations of DSL, which also comprise Microcystis aeruginosa (Kützing) Kützing, Aphanizomenon flos-aquae Ralfs ex Bornet & Flahault, and Dolichospermum lemmermannii (Richter) P.Wacklin, L.Hoffmann & J.Komárek (Cerasino et al., 2017; Salmaso, 2019; Salmaso et al., 2018a). Analysis conducted on strains isolated form DSL (Cerasino et al., 2017) have shown that M. aeruginosa and P. rubescens could produce MCs, T. bourrellyi anatoxin-a (ATX), while A. flos-aquae and D. lemmermannii resulted to be not toxic. A successive study (Capelli et al., 2017) showed also that the lack of toxicity of D. lemmermannii strains isolated in DSL was due to the lack of both MCs and ATXs encoding genes. A field study was conducted in 2009 in order to evaluate the toxic potential of cyanobacterial populations in DSL (Cerasino and Salmaso, 2012); the study showed that i) MCs were present in all DSL, with a dominance of demethylated variants; and ii) ATX was present in four out of five lakes, the exception being represented by Lake Lugano. Moreover, in the study, the differences in toxin concentrations among lakes were interpreted in terms of different lake characteristics; more specifically, MCs levels were positively related to nutrients’ concentrations, while ATX levels were more dependent on the water temperature, thus suggesting eutrophication and climate change as factors shaping the cyanotoxins diversity in different ways. Climatic factors have been also recently reported as the main driver of cyanotoxins distribution at European level (Mantzouki et al., 2018b). Since 2009, cyanotoxins in Lake Garda have been regularly measured, thus allowing following the changes of the toxic potential of the lake. If the effects of the oligotrophication of Lake Garda and the other DSL is certainly protective against cyanobacteria proliferation, the effects of climate change are more difficult to assess, due to the many ways they can operate, for example increasing the water temperature and water column stability, increasing CO2, and altering the hydrology (Callieri et al., 2014; Pareeth et al., 2017; Visser et al., 2016). The availability of a time series for cyanotoxins data collected with uniform methodology is a valuable tool for understanding past changes and foreseeing future trajectories. This information is of great importance for the management of the lake, as the risks posed by MCs are different from those posed by ATX, and also the measures for contrasting these risks are different (Ibelings et al. 2014). Moreover, considering the high chemical diversity of cyanotoxins, the analytical method should be adequate to describe as many toxins as possible. The objectives of this contribution include: i) critical evaluation of the results obtained from the analysis of cyanotoxins conducted with a monthly frequency in Lake Garda from 2009 till 2016; ii) the comparison of the longterm data collected in Lake Garda with those collected in other three DSL (lakes Iseo, Como and Lugano) in 2016. Data collected allow highlighting changes in the toxic profile in the considered lakes and contributing to define correctly the risks related to possible cyanobacterial blooms.