CO-OCCURRENCE OF MICROCYSTIN AND CYLINDROSPERMOPSIN IN HYPEREUTROPHIC MAHAKANADARAWA AND NACHCHADUWA RESERVOIRS

Microcystins (MCs) and cylindrospermopsin (CYN) are the most abundant toxins produced by cyanobacteria in tropical freshwaters. We studied the spatial distribution of MC and CYN in two multipurpose reservoirs, Mahakanadarawa and Nachchaduwa in Anuradhapura district in Sri Lanka in September 2020. Fourteen water quality parameters, phytoplankton composition, chlorophyll- a , MC and CYN concentrations were analyzed in triplicate in 25 sampling sites from each reservoir. Both reservoirs were at hypereutrophic status. Microcystis was the dominant cyanobacteria with 0-3.75 x 10 3 cell/mL in Mahakanadarawa and 1-7 x 10 3 cell/mL in Nachchaduwa. Besides Microcystis , no other potential MC-producing cyanobacteria were observed. In Mahakanadarawa, MC was detected in the range of 0.11-1.63 µg/L which was above the WHO permissible level (1.0 µg/L) for drinking water. Although comparatively high Microcystis cell density was present in Nachchaduwa, its MC concentration was low (0.06-0.17 µg/L). The CYN concentration in Nachchaduwa was above the WHO permissible level (0.7 µg/L) for drinking water. It was 0.20-1.02 µg/L in Nachchaduwa and 0.03-0.08 µg/L in Mahakanadarawa. We did not observe any potential CYN-producing cyanobacteria in either of the reservoirs. There was no relationship between the spatial distribution pattern of MC and Microcystis cell density in both reservoirs. Although the majority of physico-chemical properties of water indicated suitability for drinking, co-occurrence of high concentrations of MC and CYN indicated their unsuitability for drinking. Hence, this study highlights the necessity for routing detection of cyanotoxins in both reservoirs. Further, our findings alarm potential health risks for the local community that relies on Mahakanadarawa and Nachchaduwa reservoirs for drinking, irrigation and fisheries.


INTRODUCTION
Cyanobacterial blooms in freshwaters can cause serious ecological and health problems worldwide due to their ability to produce toxic secondary metabolites which are known as cyanotoxins. Cyanotoxins are classified according to their chemical structure mainly into cyclic peptides, alkaloids, lipopeptides, non-protein amino acids and lipoglycans (Du et al. 2019). Some of the cyanotoxins such as microcystin (MC), nodularin (NOD) and cylindrospermopsin (CYN) are hepatotoxins (Benamara et al. 2021). Moreover, CYN is also a genotoxin (Žegura et al. 2011). Anatoxin and saxitoxin are neurotoxins (Testai et al. 2016) while lyngbyatoxin and aplysiatoxin lead to dermatotoxicity (Kaebernick and Neilan et al. 2001). Repeated exposure to cyanotoxins may cause organ failures such as liver, kidney, lung and thyroid and even death (Buratti et al. 2017). Based on the health impacts of cyanotoxins, the Word Health Organization (WHO) proposed a provisional permissible concentration limit of 1.0 µg/L for MC (WHO 2020) and 0.7 µg/L for CYN in drinking water (WHO 2019).
MC and its chemical variants are cyclic peptides produced by various species of cyanobacteria such as Anabaena, Anabaenopsis Microcystis, Nostoc, Oscillatoria and Planktothrix in the major orders of cyanobacteria viz. Chroococcales, Oscillatoriales, Nostocales (Nowruzi and Porzani 2021;Kesari et al. 2022). MC-producing cyanobacteria are found primarily in freshwaters with a wide geographic and ecological distribution worldwide. Hence, MCs are the most frequently reported cyanobacterial toxins. To date, more than 250 different chemical variants of MCs have been identified and of all, the most common and toxicologically most studied MCs is the MC-LR variant (Spoof and Catherine 2017;Bouaïcha et al. 2019). Next to MCs, CYN is the most widely found cyanotoxin worldwide in freshwaters (Antunes et al. 2015), which is chemically an alkaloid with few naturally occurring variants (Ohtani et al. 1992;Kokocinski et al. 2017). CYNs are produced M. H. M. A. S. V. Gunawardana et al. 69 by various strains of cyanobacterial species, primarily in the order Nostocales (Anabaena, Aphanizomenon, Cylindrospermopsis, Raphidiopsis, Umezakia) and Oscillatoriales (Lyngbya, Oscillatoria) (De la Cruz et al. 2020). In Asia, Australia and New Zealand, Cylindrospermopsis raciborskii is a major CYN producer while in Europe and the Americas it is Aphanizomenon and Dolichospermum (formerly Anabaena) (Antunes et al. 2015). In Sri Lanka, MCs and CYN are the most common and frequently reported cyanotoxins (Kulasooriya 2017) as they are the most frequently studied cyanotoxins in Sri Lanka and not much attention was paid on other cyanotoxins.
MCs and CYN occur naturally in surface waters. Therefore, if untreated or insufficiently treated, drinking water is the most likely route of exposure to cyanotoxins (WHO 2020). In addition, bathing, irrigation and consumption of toxin-accumulated food items may also be other exposure pathways. Since recreational activities such as bathing typically take place near-shore where surface blooms usually accumulate, unintentional swallowing of water sometimes can lead to exposure to extremely high concentrations of toxins. In natural water, MCs are mostly confined to intracellular in viable cells (Juttner and Lüthi 2008). Therefore, MC concentrations in water coincide with the abundance of MCproducing cyanobacterial species. MCs can be released from cells by senescence of cells or induced cell lysis for example by cyanophage activity, unfavorable conditions for growth or algicidal treatment (WHO 2020). In contrast to MCs, up to 90% of the CYN in water are released from viable cells (Preußel et al. 2009). Therefore, the highest MC and CYN concentrations occur during the peak of blooms. Previous studies have reported cooccurrence of two or more cyanotoxins in freshwater reservoirs. For instance, co-occurrence of CYN, MC and nodularin in Lake Victoria in Tanzania (Mchau et al. 2021), CYN, MC and saxitoxin in Lake Taihu in China (Li et al. 2022) and anatoxin-a and MC in the Lake Garda in Italy (Cerasino and Salmaso 2020). In Sri Lanka, although there are many reports on the co-occurrence of toxigenic cyanobacteria in freshwater reservoirs, to our knowledge, the co-occurrence of two or more cyanotoxins has not been reported to date.
In the environment, MCs are moderately resistant to chemical and microbial breakdown whereas CYN is highly resistant to degradation (Chiswell et al. 1999;Massey and Yang 2020). Hence, even if the cell-bound CYN concentration is low, the concentration of dissolved CYN can be high. It can remain long after the CYN-producing cyanobacteria have disappeared (WHO 2019). Therefore, careful monitoring of water is required to identify the factors that can affect the growth of toxin-producing cyanobacteria such as nutrient status and other physico-chemical parameters that favors cyanobacterial growth.
In this study, we selected Mahakanadarawa and Nachchaduwa reservoirs in the North-Central province in Sri Lanka. Both reservoirs provide water for drinking, fisheries and irrigation. The National Water Supplies and Drainage Board (NWSDB) provides longterm and stable drinking water supplies for the local community from these two reservoirs.
In addition, Mahakanadarawa reservoir supplies raw water for a recently implemented largescale drinking water project that aims to resolve long-felt water scarcity in rural areas in the Anuradhapura district (NWSDB, Sri Lanka). The main transmission pipeline extends to 89 km supplying water for the population of around 75,000 in 30 Grama Niladhari divisions in the Medawachchiya, Rambewa and Mihintale Divisional Secretariats. Therefore, a large community in the Anuradhapura district utilize water from both reservoirs for their domestic needs, agriculture and fisheries. Hence, the assessment of water quality and cyanotoxin contamination levels in the above two reservoirs have significant importance. We studied the spatial distribution of MC and CYN cyanotoxins in both Mahanadarawa and Nachchaduwa reservoirs along with their phytoplankton composition and physico-chemical properties to understand potential health risks to the local community.

MATERIALS AND METHODS
Site description: Mahakanadarawa and Nachchaduwa are medium-scale, man-made reservoirs with an area at a full capacity of 6100 and 7000 acres respectively (Irrigation Department, Sri Lanka). These two reservoirs are situated in the Anuradhapura district in the North-Central province (8°23'27.83" N & 80°32'47.65" E and 8°14'56.29" N &80°28'59.24"E respectively) ( Figure 1). Mahakanadarawa is mainly fed by rainwater and used for aquaculture and fisheries. Malwathu Oya which makes the second-largest catchment of the river basins in the country provides water to the Nachchaduwa reservoir.
Sampling was conducted in September 2020 on two-clear sunny days during the daytime between 9.00 am to 2.00 pm. Twenty-five sites from each reservoir were selected for on-site water quality assessment and laboratory analyses ( Figure 1). Sampling was not carried out in some areas of Mahakanadarawa reservoir due to the presence of submerged tree trunks and dense growth of Potamogeton sp. in Nachchaduwa reservoir. Therefore, those areas were not included as sampling sites in this study. On-site assessment of water quality: Temperature (°C), pH, oxidation-reduction potential (ORP, mV), electrical conductivity (EC, µS/cm), total dissolved solids (TDS, ppm), salinity (psu), pressure (psi), dissolved oxygen (DO, ppm) and turbidity (FNU) were measured in situ in triplicate using a multi-parameter water quality meter (HI9829-Hanna Instruments, Romania) and water transparency was estimated using a Secchi disk.
Laboratory analysis of water quality: Water samples for the laboratory analysis were taken with three replicates from the same sites where in-situ water quality parameters were measured. For the estimation of chlorophyll-a, 50 L of water was filtered through a 30 µm plankton net (Hydro-Bios, Germany). In addition, water samples were collected for the determination of NO3 --N, NH4 + -N, PO4 3--P and total suspended solids (TSS). All the samples were collected from the surface of the water into 250 mL glass bottles. Samples were transported to the laboratory following standard protocols of the United States Environmental Protection Agency (US-EPA 2013). All chemical reagents used were from standard manufacturers (Sigma-Aldrich, USA; BDH, UK and Oxoid, UK) unless otherwise mentioned.
The concentrations of NO3 --N (ppm) and NH4 + -N (ppm) were measured using a nitrate probe (HI7609829-12) and an ammonium probe (HI7609829-10) from Hanna Instruments Inc., Romania respectively. The concentration of PO4 3--P (ppm) in water was determined using PhosVer3 ® phosphate reagent powder pillows having 0.00 to 2.50 ppm PO 4 3--P detection range (Hatch, USA) following the manufacturer's protocol which is equivalent to the Ascorbic acid method in the US-EPA protocol 365.3 (1978). Absorbance was measured at 881 nm wavelength using a UV/visible spectrophotometer (Genesys 150, Thermo-Scientific, USA). A gravimetric method was used to estimate TSS following the Method 2540D of the American Public Health Association (APHA 1999).
The chlorophyll-a concentration of field-collected water samples was immediately measured after returning to the laboratory following the method of Arnon (1949) with slight modifications. Briefly, phytoplankton biomass was collected by low-speed centrifugation at 5000 rpm for 10 minutes. Chlorophyll was extracted into 80% (v/v) acetone. Absorbance was measured at 645 nm and 663 nm wavelengths using a UV/Visible spectrophotometer (Genesys 150, Thermo-Scientific, USA). Chlorophyll-a concentration was calculated using the following formula, where OD is the optical density (absorbance), v is the volume of acetone, and V is the volume of the tested water sample (L). The path length of light (cm) is represented by d.
Chlorophyll-a (ppb)= [(12.7 Water samples that were filtered through the plankton net were used for the identification and enumeration of phytoplankton. Cyanobacteria, diatoms and green algae were identified up to the genus or species level with the aid of standard identification keys (Baker et al. 2012;Nienaber and Steinitz-Kannan 2018). The cell density of phytoplankton was estimated following the method described by LeGresley and McDermott (2010).
Briefly, water samples were observed within 24 h after collection and enumerated using a Neubauer hemocytometer (Marienfeld, Germany) under the 400x magnification of a brightfield microscope (Nikon Ci-L, Japan) and average cell density was determined from three replicates collected from each sampling site. In order to release intracellular MC, water samples were subjected to three cycles of freezing (-20 °C) and thawing (30 °C) (US-EPA 2016). Thereby, both intracellular and extracellular MC were taken for the measurement. As up to 90% of the total CYN in the surface water is in the dissolved fraction (WHO 2019), water samples were directly measured for CYN without subjecting to cell lysis. Toxin concentrations were determined using MC and CYN plate kits (96-well), which provide a 0 ppb-2 ppb min-max detection range (Beacon, USA).
All the reactions were performed following the manufacturer's protocols. The absorbance of MC and CYN standard solutions (provided with the kit) and test samples were measured at 450 nm wavelength using a microplate reader (Multiscan FC, Thermo-Scientific, USA).
Based on the measured MC and CYN toxin concentrations and cell count of Microcystis for each sampling site, their spatial distribution in two reservoirs was mapped using AchGIS 10.6.1 software (Environmental Systems Research Institute Inc., Redlands, CA, USA).

Determination of trophic status: Trophic status of the two reservoirs was determined as
Total Trophic State Index (TTSI) with the aid of Secchi depth and chlorophyll-a concentration using the following formulae in which TTSI 2.0-2.7, oligotrophic, 3.3-3.7, mesotrophic, 4.5-5.0 eutrophic, and 6.0 ≤, hypereutrophic (Burns 1999).
TSI for Secchi depth (SD) = 5.10 + 2.60 x log (1/SD -1/40) (meters) TSI for chlorophyll-a (CA) = 2.22 + 2.54 x log (CA) (µg/L) (4) Data analysis: Data were statistically analyzed by using a standard statistical software package MINITAB version 18. The measured parameters were presented as arithmetic means ± SD of three independent measurements. Pearson product correlation coefficient (r) was calculated to determine the relationship between water quality parameters, toxin concentration and phytoplankton cell density. The statistical significance was defined at the 0.05 level.

RESULTS
Water quality and trophic status: An algal bloom was not observed at the time of sampling in September 2020 in both Mahakanadarawa and Nachchaduwa reservoirs. The quality of water as indicated by +physico-chemical and biological properties is summarized in

DISCUSSION
In this study, we report the co-occurrence of two hepatotoxic cyanotoxins, MC and CYN in Mahakanadarawa and Nachchaduwa reservoirs. Unlike in temperate countries where cyanobacterial blooms are likely to be seasonally present, a greater risk is possible for communities in tropical regions with persistent cyanobacterial blooms throughout the year.
Therefore, our findings alarm health risks for the local community that rely on Mahakanadarawa and Nachchaduwa for drinking, irrigation and aquaculture due to the potential long-term exposure to the above toxins. Although most measured water quality parameters fell within the favorable range for ambient water quality standards according to as Phormidium and Plantothrix (Magonono et al. 2018). We found that the patterns of the spatial distribution of two toxins in the two reservoirs were highly variable (Figure 4). In Mahakanadarawa, MC accumulated toward the South-shore (Figure 4a) which was opposite to the highly concentrated areas of chlorophyll-a were present (Figure 4e). This might be due to the buoyancy of surface blooms such as Microcystis and their dislocation at temporary locations following wind direction (Xingqiang et al. 2019). Therefore, it is highly likely that the observed spatial distribution of Microcystis in Mahakanadarawa corresponds to the direction of wind toward the North-shore although its density was not high enough to form surface scum. However, MC did not show a distribution pattern following the downwind.
Therefore, it appears that the dissolved fraction of MC resulting from the cell lysis may have contributed more to the detected high concentrations of MC at sites near the South shore.
That dissolved fraction of MC may have originated from a previous Microcystis bloom or any other MC-producing cyanobacterial community (Magonono et al. 2018;Preece et al. 2021). In contrast, we detected comparatively low levels of MC in Nachchaduwa throughout the water body although it had comparatively high densities of Microcystis (Figure 4f).
Therefore, it seems that the dissolved fraction of MC had contributed less to the detected concentrations in Nachchaduwa and most likely toxin was intracellular within intact cells.
Senescence of those high densities of intact cells could elevate MC in the water body after some time when cells undergo natural or induced lysis releasing their intracellular toxins.
Therefore, high Microcystis cell density is an indication of a future MC outbreak.
Of the physico-chemical parameters that we measured, MC and CYN in Mahanadarawa and Nachchaduwa reservoirs positively correlated only with temperature and turbidity whereas negatively correlated with Secchi depth transparency. If nutrient and mixing conditions permit, warmer water bodies can sustain more stable toxin (such as MCs) production because thermal stratification is favored by buoyant phytoplankton species such as cyanobacteria (Rigosi et al. 2014). Further, water column stability is an important determinant of the dominancy of cyanobacteria and their toxin production (Persaud et al. 2015). Therefore, in countries like Sri Lanka where the warm temperature is prevalent throughout the year, stable cyanotoxin production could be expected. Moreover, according to our data, water temperature within the reservoir in a single day showed a positive correlation with MC and CYN concentrations probably due to the spatial and diurnal variation in temperature where warmer areas accumulated more toxins. The DO is M. H. M. A. S. V. Gunawardana et al. 83 considered as a potential indicator for the onset of cyanotoxins derived from lysed cells in a blooming population (Hartnell et al. 2020). However, DO of both Mahakanadarawa and Nachchaduwa reservoirs did not show a significant correlation to either MC or CYN.
Previous studies have shown that a decrease in DO have a close association with the presence and concertation of MCs (O'Neil et al. 2012;Paerl and Otten 2013). In parallel to our results, a positive correlation between turbidity and MC has been reported in several previous studies (Te and Gin 2011;Graham et al. 2004). Turbidity of surface water is increased with abundance growth of phytoplankton including cyanobacteria such as Microcystis resulting in the accumulation of more intracellular MCs (Te and Gin 2011;Lee et al. 2015). Regular monitoring of turbidity could be used as a proxy for MC concentrations in reservoirs, hence perturbation in turbidity is an indication of water contamination (Cunha et al. 2018). It is obvious that turbidity reduces the Secchi depth transparency and it could be the reason for the negative correlation that we obtained in this study for MC and CYN.
Based on the currently available literature, toxin-producing Microcystis and Cylindrospermopsis were first reported in Sri Lanka in 1907 in northern dry zone reservoirs as non-dominant genera in association with green algae and diatoms (Fritsch 1907a;Fritsch 1907b). Later, between the 1995-1996 periods, Cylindrospermopsis has been identified from Tisawewa as the dominant cyanobacteria and in Basawakkulama wewa as an occasionally present species (Weerakoon et al. 1998). At the same time, Mycrocystis was identified as a dominant genus in northern dry zone reservoirs . Several later research reported toxigenic genera Cylindrospermopsis raciborskii as the dominant cyanobacterial community in several northern dry zone reservoirs and its co-occurrence with Microcystis only in Unnichchai tank (Wanigatunge et al. 2014;Kulasooriya 2017 andYatigammana andPerera 2017). Further, several other toxin-producing cyanobacteria such as Planktolyngbya, Pseudoanabaena, Anabaena aphanizomenoides (Silva and Wijeyaratne 1999), Phormidium, Oscillatoria (Liyanage et al. 2014) and Lyngbya (Shihana et. al. 2012) have been reported from the reservoirs in the northern dry zone. However, we did not observe any of those species in both reservoirs understudy. When a lentic water body is eutrophicated, phytoplankton that are sensitive to nutrient pollution gradually disappear and more resistant species survive and tend to become dominant (Carmichael 1992 Further, they reported Anabaena as the dominant cyanobacteria in Nachchaduwa and in contrast, we observed Microcystis as the dominant species and Anabaena was not found at the time of sample collection. Therefore, this is good evidence for variation in composition and diversity of cyanobacteria over time while indicating the importance of frequent monitoring of cyanobacteria and cyanotoxins in these reservoirs. According to our observations, the spatial distribution pattern of CYN in Mahakanadarawa reservoir is approximately uniform while in Nachchaduwa it showed slight accumulation toward the direction of the South (Figures 4b and 4d). Although we did not observe any of the CYN-producing cyanobacterial species in either of the reservoirs, Wanigatunge et al. (2014) and Yatigammana and Perera (2017) have reported the presence of potential CYN-producing cyanobacterial species Cylindrospermopsis raciborskii in Nachchaduwa. However, in this study, we did not observe C. raciborskii or any other CYNproducing species probably due to two reasons, one is the presence of very low densities of such filaments at the time of sampling due to their disappearance. It is known that CYN is persistent in water for a long time after the toxin-producing cyanobacteria have disappeared (Smith et al. 2008). The second reason could be the presence of other CYN-producing species such as Lyngbya and Oscillatoria in the reservoir (Seifert et al. 2007;Bormans et al. 2014). Hence, those were not captured when collecting water samples from the surface for the analysis of phytoplankton. As such, the mere presence or absence of toxin-producing cyanobacterial species in the water samples is not a reliable indicator for the presence of CYN.
Further, the co-occurrence of multiple toxins in freshwaters can arise the risk of their bioaccumulation in the tissues of aquatic organisms to levels exceeding those of the surrounding water (Mchau et al. 2021). Therefore, the human being at the highest level of the food chain is highly vulnerable to the bioaccumulation of cyanotoxins. As different toxins have different exposure routes, the co-occurrence of multiple toxins enhances the possibility of multiple exposures to the human population through ingestion of contaminated drinking water, consumption of aquatic organisms, fishing and recreational activities (Rastogi et al. 2015). Moreover, the consumption of agricultural products fed with toxincontaminated irrigational water and meat from livestock that consumes toxin-contaminated water could increase the risk of exposure (Mchau et al. 2021). Not only that, but even in laboratory experiments, it has been proven that the combined effect of MC and CYN induces higher toxicity rather than pure MC or pure CYN (Gutiérrez-Praena et al. 2019). Even at relatively low doses, repetitive and multiple toxin exposures, could result in cumulative liver damage, which may lead to chronic liver diseases in long-term exposure (Chorus et al. 2000).
Therefore, the co-occurrence of multiple toxins can impose high health risks. Therefore, frequent monitoring and implementation of strategies for controlling potential toxigenic cyanobacterial blooms and their accumulation in water supplies and food are essential.

CONCLUSIONS
We detected co-occurrence of hepatotoxic MC and CYN in two multipurpose reservoirs in the Anuradhapura district in Sri Lanka. MC in Mahakanadarawa and CYN in Nachchaduwa were found above the maximum permissible levels for drinking water defined by WHO. Microcystis was dominant in both reservoirs and appears to be the MC producing genera. However, we did not observe any potential CYN-producing species in either of the reservoirs. The spatial distribution of two toxins in water bodies was highly variable and not correlated to Microcytis cell density, chlorophyll-a concentration or any other physicochemical parameter of water. Most physico-chemical parameters of both reservoirs fell within the favorable range for drinking according to the Ambient Water Quality Standards of Sri Lankan and WHO standards. Although, MC and CYN levels exceeded WHO guidelines in both reservoirs and which indicated the unsuitability of water for drinking without treatment for toxins. Therefore, our findings highlight the necessity of routing monitoring for cyanotoxins in tropical freshwaters to minimize health risks that can be resulted due to long-term exposure to toxins. Further, we recommend more sampling within a year to better understand the spatial distribution pattern of MC and CYN toxins.