Cyanobacterial Blooms and Microcystins in Southern Vietnam

Studies on cyanobacteria in Vietnam are limited and mainly restricted to large reservoirs. Cyanobacterial blooms in small water bodies may pose a health risk to local people. We sampled 17 water bodies in the vicinity of urban settlements throughout the Mekong basin and in southeast Vietnam. From these, 40 water samples were taken, 24 cyanobacterial strains were isolated and 129 fish, 68 snail, 7 shrimp, 4 clam, and 4 duck samples were analyzed for microcystins (MCs). MCs were detected up to 11,039 µg/L or to 4033 µg/g DW in water samples. MCs were detected in the viscera of the animals. MC-LR and MC-RR were most frequently detected, while MC-dmLR, MC-LW, and MC-LF were first recorded in Vietnam. Microcystis was the main potential toxin producer and the most common bloom-forming species. A potential health hazard was found in a duck–fish pond located in the catchment of DauTieng reservoir and in the DongNai river where raw water was collected for DongNai waterwork. The whole viscera of fish and snails must be completely removed during food processing. Cyanobacterial monitoring programs should be established to assess and minimize potential public health risks.


Introduction
When sufficient resources are available, cyanobacteria may proliferate and bloom in reservoirs, lakes, rivers, estuaries, and coastal systems, where they may cause a multitude of water quality concerns, such as producing malodor, causing nocturnal oxygen deficiency leading to fish kills, disrupting the aquatic ecosystem, and posing a health hazard to wildlife, game, and humans, because of their potential to produce strong toxins (cyanotoxins) [1,2]. Cyanobacteria can cause acute or chronic toxicity to animals and humans via different exposure routes, such as contaminated drinking water, fish or shell fish, through crops irrigated with cyanobacteria-infested waters, or through recreational exposure. Numerous animal poisonings associated with exposure to cyanobacteria have been reported Figure 1. Daily activities of local people potentially exposed to cyanobacterial toxins in the Mekong Delta: Traditional fisheries in cyanobacteria bloom-water ponds (a,b) and bathing and washing (c,d).

Physical and Chemical Characteristics and Cyanobacterial Blooms
The results of the physical and chemical characteristics of 17 water bodies at the moment of our sampling in southern Vietnam are summarized in Table 1. The water temperature was high and ranged from 30 to 37.7 °C, which was due to the sampling that mainly took place during the hot season in Vietnam from January to June in 2015 and 2016. The high temperature was a consequence of the heatwaves recorded in recent years in the Indochina peninsula [26]. Salinity ranged from 0.5 to 7.6‰ at the sampling points. High concentrations of TN and TP indicated that the water bodies were eutrophic to hyper-eutrophic [27].
Cyanobacterial blooms were observed in ponds, reservoirs, canals, and rivers in southern Vietnam. The blooms in small ponds, where fish or ducks were cultivated, appeared more dense than the blooms found in reservoirs and rivers that are the main water supplies for domestic purposes, irrigation, and aquaculture. High cyanobacteria chlorophyll-a concentrations from 1437 to 5100 μg chl-a/L were measured in several carp ponds in HCMC, in a duck-fish pond and in extensive catfish ponds (Table 1). Lower concentrations from 46 to 420 μg chl-a/L were measured in the reservoirs, Tri An, Dau Tieng, and Binh Thieng ( Table 1). The abundance of cyanobacteria in terms of chl-a was strongly correlated with TN (r = 0.709) and TP (r = 0.676) (Appendix A, Figures A1 and A2). Microcystis, Planktothrix, Oscillatoria, and Cylindropermopsis were the dominant genera. Daily activities of local people potentially exposed to cyanobacterial toxins in the Mekong Delta: Traditional fisheries in cyanobacteria bloom-water ponds (a,b) and bathing and washing (c,d).

Physical and Chemical Characteristics and Cyanobacterial Blooms
The results of the physical and chemical characteristics of 17 water bodies at the moment of our sampling in southern Vietnam are summarized in Table 1. The water temperature was high and ranged from 30 to 37.7 • C, which was due to the sampling that mainly took place during the hot season in Vietnam from January to June in 2015 and 2016. The high temperature was a consequence of the heatwaves recorded in recent years in the Indochina peninsula [26]. Salinity ranged from 0.5‰ to 7.6‰ at the sampling points. High concentrations of TN and TP indicated that the water bodies were eutrophic to hyper-eutrophic [27].
Cyanobacterial blooms were observed in ponds, reservoirs, canals, and rivers in southern Vietnam. The blooms in small ponds, where fish or ducks were cultivated, appeared more dense than the blooms found in reservoirs and rivers that are the main water supplies for domestic purposes, irrigation, and aquaculture. High cyanobacteria chlorophyll-a concentrations from 1437 to 5100 µg chl-a/L were measured in several carp ponds in HCMC, in a duck-fish pond and in extensive catfish ponds (Table 1). Lower concentrations from 46 to 420 µg chl-a/L were measured in the reservoirs, Tri An, Dau Tieng, and Binh Thieng ( Table 1). The abundance of cyanobacteria in terms of chl-a was strongly correlated with TN (r = 0.709) and TP (r = 0.676) (Appendix A, Figures A1 and A2). Microcystis, Planktothrix, Oscillatoria, and Cylindropermopsis were the dominant genera.

MC Content in Cyanobacterial Field Samples and Isolated Strains
There were 41 field samples collected from the 17 water bodies to determine the MC concentrations. MCs were detected in 28 out of 41 water samples collected in 3 reservoirs, 9 small ponds, 1 canal, 2 lakes, and 2 rivers during bloom events. The MC concentrations of field samples were determined in µg/L and/or µg/g DW and are shown in Table 2. The MC concentrations ranged from under the detection level to 11,039 µg/L or to 4033 µg/g DW. The maximum concentrations of total MC in the water samples were determined in the duck-fish pond in TraVinh province, Mekong delta, where the fish climbing gouramies (Anabas testudineus) and ducks were cultivated. A higher MC concentration was detected in 18 samples collected from blooms with Microcystis dominance than in the three samples with Oscillatoria, Planktothrix dominance. Twenty-four cyanobacterial strains were isolated from sampled blooms that were also tested for MC concentration in both µg MC/L and µg MC/g DW and shown in Tables A1 and A2 (Appendix B). MCs were found in only nine Microcystis strains and was not found in four other Microcystis strains. In the isolated Anabaena, Anabaenopsis and Planktothrix strains, no MCs were detected (Appendix B). MC-LR and MC-RR variants were most frequently detected in 24 and 20 out of 28 MC-containing field samples, respectively. They were also the most abundant MC variants in the samples, in which MC-RR contributed from 44% to 100% of the total MC-pool in 17 samples and MC-LR contributed from 43% to 100% of the total MC-pool in the remaining 11 samples. Our study contributed to the diversity of MC variants as three MC variants including MC-dmLR, MC-LW, and MC-LF were recorded for the first time in Vietnam (Table 3).

MCs Accumulation in Animals
The MC content was determined in 212 organ and tissue samples of aquatic animals, including 129 fish, 68 snail, 7 shrimp, 4 clam, and 4 duck samples collected in water bodies suffering from cyanobacterial blooms. MCs were detected in 36 samples (17% of total samples), including 23 snail, 12 fish, and 1 shrimp samples ( Table 4, Appendix C).
The MC content in the animals varied among fish species, organs, tissues, and seemed to be influenced by the sampling sites where they had been collected (Table 4). On average, the MC content was higher in fish than in shrimp and snail with 24.1, 15.2, and 1.1 µg/g DW in fish, shrimp, and snail, respectively (Table 4). MCs were mainly found in the viscera including the visceral mass, liver, and gut in apple snails, tilapias, suckermouth catfish, goldfish, common carp, ganges river sprat fish, snakeskin gourami fish, and white leg shrimp. MCs were also found in 4 flesh samples, in 3 apple snails, and 1 suckermouth catfish (Table 4). No MCs were found in clam and duck. The highest MC content of 115.95 µg/g DW was detected in the gut of the omnivorous tilapia collected in a fish pond in BinhChanh, which was followed by 108.28 µg/g DW found in the gut of suckermouth catfish collected in the same pond. A lower MC content was found in the gut of goldfish (18.57 µg/g DW) collected in BinhChanh and in the gut of tilapia collected in TriAn, DauTieng reservoirs (11.55 and 0.76 µg/g DW).
A trace amount of 0.36 µg MC/g DW was found in the gut of common carp (Carassius sp.) collected in BinhChanh.
MCs were also detected in the livers of tilapia (13.39 µg/g DW) and suckermouth catfish (18.97 µg/g DW) collected in BinhChanh. In our survey, the visceral mass in small tilapia, snakeskin gourami fish, ganges river sprat fish, white leg shrimps, and apple snails was collected to determine the MC content, because it was very hard to separate gut or liver organs from the rest visceral organs of these small animals. In addition, the animals were eaten whole by local consumers, thus the MC content in the visceral mass would be more reliable for warning the local consumers. The MC content in the visceral mass in snakeskin gourami fish and white leg shrimp was 13. 86 and 15.21 µg/g DW; those in tilapia, ganges river sprat fish, and apple snail were 2.87, 2.38, and 0.21-2.90 µg/g DW, respectively. MC-LW and MC-LF seem to be more toxic than MC-LR that is currently used for risk calculations and assessments [39,40]. These MC variants were found for the first time in Vietnam, in TriAn reservoir and Mekong, DongNai rivers. Since different MC-variants have different toxicity as determined by bioassays [11,39,40], the contribution of each variant to the overall MC toxicity of the samples was estimated by multiplying its concentration by a toxicity factor as mentioned in Table 1 in Faassen and Lürling 2013 [41]. As a consequence of the lower toxicity factor of MC-RR in comparison with MC-LR (see Table 1 in Faassen and Lürling 2013 [41]), the overall toxicity of MC-RR-dominated samples became less. For instance, the MC concentration in the sample collected in a duck-fish pond in Travinh was 11,039 µg/L, but the overall estimated toxicity of the sample was equal to only 963 µg/L MC-LR equivalents. Samples containing MC-LW and MC-LF were estimated to increase in toxicity when expressed as MC-LR equivalents, because these two variants have higher toxicity factor in comparison with MC-LR.
MC-LR occupied, on average, 81% of total toxicity in 13 out of 18 samples containing high MC concentrations ( Figure 2). MC-RR and MC-LF contributed 81% and 44%, respectively, to the total toxicity in two other samples.

Occurrence of Cyanobacterial Blooms
There are few reports on cyanobacterial blooms and MCs in southern Vietnam, such as in TriAn and DauTieng reservoirs [22,25,30,42,43]. In addition to these relatively large water bodies, our study also points out that MC-producing cyanobacterial blooms occur in small ponds, small reservoirs and rivers in southern Vietnam. In fact, this is the first report on cyanobacterial blooms and MCs in the Vietnamese Mekong delta (Table 3). There are thousands of small ponds or water bodies in southern Vietnam, which are cultivating fish and ducks in a combination model to provide daily food for local markets. Pellet feeds containing high levels of nitrogen and phosphorus are used as food for fish and ducks leading to high concentrations of nitrogen and phosphorus from uneaten pellet feeds and feces of fish and ducks. Additionally, chicken manure is often applied to enhance primary productivity during the initial phase of fish cultivation. Our monitoring indicated that the ponds were highly eutrophic or hyper-eutrophic, as TN and TP ranged from 3.98 to 60.7 mg/L and from 0.25 to 1.75 mg/L, respectively ( Table 1). The excessive nutrients, high water temperature, and rather stable water column resulted in cyanobacterial blooms in the ponds. These blooms were evidently linked to TN and TP concentrations as the higher the TN and TP concentrations, the higher the cyanobacteria biomass in water (Appendix A). Microcystis strains isolated from the ponds expressed better growth rates under warming temperature [44]. Hence, excessive nutrients and high temperature may support the cyanobacterial blooms observed during the dry, hot seasons in 2015 and 2016 in southern Vietnam.

MC Content in Cyanobacterial Field Samples and Isolated Strains
Our study confirmed that Microcystis was the main potential toxin producer and the most common bloom-forming species in southern Vietnam, which is in accordance with previous studies [21,22,24,[28][29][30]41,43]. The other bloom-forming species, such as Oscillatoria and Planktothrix, can also accumulate high biomass in the water bodies, but the MC concentrations were below the detection level or found at trace levels (<0.5 μg/L). This finding is supported by the study of Nguyen et al. [28] in several water bodies in Hue, middle Vietnam, where the MC concentrations were high (47.8 μg/L) in bloom samples with Microcystis dominance, while the MC concentrations were much lower ranging from below the detection level to 0.05 μg/L in bloom samples with Arthrospira, Merismopedia dominance, or 1.31 μg/L in bloom samples with Jaaginema, Oscillatoria dominance [28].

MC Accumulation In Animals
In general, the MC content detected in the gut, liver and visceral mass of fish in our study was high when compared to other studies (Table 4). For instance, the MC content in the visceral mass of golden carp collected in BinhChanh was nine times higher than that found in the intestine of golden carp collected in lake Taihu, China [32]. The MC content in the liver of tilapia collected in BinhChanh

Occurrence of Cyanobacterial Blooms
There are few reports on cyanobacterial blooms and MCs in southern Vietnam, such as in TriAn and DauTieng reservoirs [22,25,30,42,43]. In addition to these relatively large water bodies, our study also points out that MC-producing cyanobacterial blooms occur in small ponds, small reservoirs and rivers in southern Vietnam. In fact, this is the first report on cyanobacterial blooms and MCs in the Vietnamese Mekong delta (Table 3). There are thousands of small ponds or water bodies in southern Vietnam, which are cultivating fish and ducks in a combination model to provide daily food for local markets. Pellet feeds containing high levels of nitrogen and phosphorus are used as food for fish and ducks leading to high concentrations of nitrogen and phosphorus from uneaten pellet feeds and feces of fish and ducks. Additionally, chicken manure is often applied to enhance primary productivity during the initial phase of fish cultivation. Our monitoring indicated that the ponds were highly eutrophic or hyper-eutrophic, as TN and TP ranged from 3.98 to 60.7 mg/L and from 0.25 to 1.75 mg/L, respectively ( Table 1). The excessive nutrients, high water temperature, and rather stable water column resulted in cyanobacterial blooms in the ponds. These blooms were evidently linked to TN and TP concentrations as the higher the TN and TP concentrations, the higher the cyanobacteria biomass in water (Appendix A). Microcystis strains isolated from the ponds expressed better growth rates under warming temperature [44]. Hence, excessive nutrients and high temperature may support the cyanobacterial blooms observed during the dry, hot seasons in 2015 and 2016 in southern Vietnam.

MC Content in Cyanobacterial Field Samples and Isolated Strains
Our study confirmed that Microcystis was the main potential toxin producer and the most common bloom-forming species in southern Vietnam, which is in accordance with previous studies [21,22,24,[28][29][30]41,43]. The other bloom-forming species, such as Oscillatoria and Planktothrix, can also accumulate high biomass in the water bodies, but the MC concentrations were below the detection level or found at trace levels (<0.5 µg/L). This finding is supported by the study of Nguyen et al. [28] in several water bodies in Hue, middle Vietnam, where the MC concentrations were high (47.8 µg/L) in bloom samples with Microcystis dominance, while the MC concentrations were much lower ranging from below the detection level to 0.05 µg/L in bloom samples with Arthrospira, Merismopedia dominance, or 1.31 µg/L in bloom samples with Jaaginema, Oscillatoria dominance [28].

MC Accumulation In Animals
In general, the MC content detected in the gut, liver and visceral mass of fish in our study was high when compared to other studies (Table 4). For instance, the MC content in the visceral mass of golden carp collected in BinhChanh was nine times higher than that found in the intestine of golden carp collected in lake Taihu, China [32]. The MC content in the liver of tilapia collected in BinhChanh was almost double that measured in the liver of tilapia collected in Funil and Furnas reservoirs, Brazil [35] and 57 and 87 times higher than the MC content found in the liver of tilapia collected in lake Mburo and lake Victoria, Uganda [36]. Both carp and tilapia are omnivorous fish; when living in ponds with a high biomass of toxic Microcystis, they likely consume Microcystis via their daily food and consequently may accumulate MCs. Fecal pellets of the fish in BinhChanh fish pond were collected during our sampling and microscopy revealed that the fecal pellets contained undigested Microcystis and empty rotifer ( Figure 3). Although we cannot estimate how much Microcystis was ingested and digested by the fish, the microscopy indicated that Microcystis was ingested by the fish.
was almost double that measured in the liver of tilapia collected in Funil and Furnas reservoirs, Brazil [35] and 57 and 87 times higher than the MC content found in the liver of tilapia collected in lake Mburo and lake Victoria, Uganda [36]. Both carp and tilapia are omnivorous fish; when living in ponds with a high biomass of toxic Microcystis, they likely consume Microcystis via their daily food and consequently may accumulate MCs. Fecal pellets of the fish in BinhChanh fish pond were collected during our sampling and microscopy revealed that the fecal pellets contained undigested Microcystis and empty rotifer (Figure 3). Although we cannot estimate how much Microcystis was ingested and digested by the fish, the microscopy indicated that Microcystis was ingested by the fish.  (Table 4), which was also found by Pham et al. [25]. However, covalently bound MCs were found in the muscle tissue of tilapia collected in DauTieng reservoir [25]. The hot methanol extraction applied in our study was not able to detect the bound MCs, which may be one of the reasons why MCs were undetected in muscle tissue in the fish in our study.

Risk Assessment for Drinking Water Supplies
DongNai, SaiGon rivers and DauTieng reservoir are the main water supplies for HoChiMinh city (hereafter HCMC) [45,46]-a megacity with 10 million inhabitants. Our monitoring indicated that MC concentrations ranged from 0.09 to 2.22 μg/L-equivalent to 0.01-0.78 μg MC-LR equivalent/L in raw water collected at five sampling points in DauTieng reservoir, which is below the WHO guideline value of 1.0 μg MC-LR/L for drinking water [1]. Additionally, extracellular MCs in raw water in DauTieng reservoir were lower than the level of detection. DauTieng reservoir was a safe source of drinking water at the time of our monitoring, in accordance with WHO guidelines. However, the occurrence of MCs in this reservoir should be paid attention due to their potential health risks. In addition, concentrations of 30 μg/L or 8.55 μg MC-LR equivalent/L and 485 μg/g DW or 47.73 MC-LR equivalent/g DW were detected in samples of a duck-fish pond located in the catchment of DauTieng reservoir. These concentrations were much higher than the WHO guideline value for drinking water. As these duck and fish cultivations occurred in the catchment of DauTieng reservoir, discharged water from these ponds could not only inoculate or increase blooms of cyanobacteria, but also imply a potential health hazard for drinking water supply.
An MC concentration of 664 μg/g DW was detected at HoaAn in the Dongnai river, where raw water was collected for DongNai waterwork. This raw water is treated two times with chlorine 1-2 mg/L and 2-4 mg/L to kill algae and moss as in the typical process scheme of a water drinking plant in Vietnam (Figure 4) [47]. Chlorine is known as an algaecide and chlorination has a strong action on membrane disruption, which can induce cyanobacterial cell lysis. For instance, 97% of M. aeruginosa  (Table 4), which was also found by Pham et al. [25]. However, covalently bound MCs were found in the muscle tissue of tilapia collected in DauTieng reservoir [25]. The hot methanol extraction applied in our study was not able to detect the bound MCs, which may be one of the reasons why MCs were undetected in muscle tissue in the fish in our study.

Risk Assessment for Drinking Water Supplies
DongNai, SaiGon rivers and DauTieng reservoir are the main water supplies for HoChiMinh city (hereafter HCMC) [45,46]-a megacity with 10 million inhabitants. Our monitoring indicated that MC concentrations ranged from 0.09 to 2.22 µg/L-equivalent to 0.01-0.78 µg MC-LR equivalent/L in raw water collected at five sampling points in DauTieng reservoir, which is below the WHO guideline value of 1.0 µg MC-LR/L for drinking water [1]. Additionally, extracellular MCs in raw water in DauTieng reservoir were lower than the level of detection. DauTieng reservoir was a safe source of drinking water at the time of our monitoring, in accordance with WHO guidelines. However, the occurrence of MCs in this reservoir should be paid attention due to their potential health risks. In addition, concentrations of 30 µg/L or 8.55 µg MC-LR equivalent/L and 485 µg/g DW or 47.73 MC-LR equivalent/g DW were detected in samples of a duck-fish pond located in the catchment of DauTieng reservoir. These concentrations were much higher than the WHO guideline value for drinking water. As these duck and fish cultivations occurred in the catchment of DauTieng reservoir, discharged water from these ponds could not only inoculate or increase blooms of cyanobacteria, but also imply a potential health hazard for drinking water supply.
An MC concentration of 664 µg/g DW was detected at HoaAn in the Dongnai river, where raw water was collected for DongNai waterwork. This raw water is treated two times with chlorine 1-2 mg/L and 2-4 mg/L to kill algae and moss as in the typical process scheme of a water drinking plant in Vietnam (Figure 4) [47]. Chlorine is known as an algaecide and chlorination has a strong action on membrane disruption, which can induce cyanobacterial cell lysis. For instance, 97% of M. aeruginosa cells was lysed within the first minute when exposed to 3 mg/L chlorine [12]. Chlorine application of 2-4 mg/L during pre-chlorination in many drinking water plants in Vietnam can result in cell lysis and consequently cyanotoxins can be liberated into drinking water. Although there was no information on cyanobacteria concentration in bloom events in HoaAn in January 2016, the greenish appearance and Microcystis dominance in water with an MC content of 664 µg/g DW or 554 µg MC-LR equivalent/g DW should be considered as a potential health hazard for water consumers.
Toxins 2018, 10, x FOR PEER REVIEW 10 of 21 cells was lysed within the first minute when exposed to 3 mg/L chlorine [12]. Chlorine application of 2-4 mg/L during pre-chlorination in many drinking water plants in Vietnam can result in cell lysis and consequently cyanotoxins can be liberated into drinking water. Although there was no information on cyanobacteria concentration in bloom events in HoaAn in January 2016, the greenish appearance and Microcystis dominance in water with an MC content of 664 μg/g DW or 554 μg MC-LR equivalent/g DW should be considered as a potential health hazard for water consumers. Rivers, lakes, and ponds in the Mekong delta are among the main water supplies for household daily use. These water supplies provided 36% for households' daily demands (besides drinking and cooking) and 25% for the demands of drinking and cooking [48,49] for 17.4 million people. The available surface water from these water supplies can be directly used for domestic purposes, especially in the dry season and cyanobacteria accumulation in the surface water has not been recognized as a potential health risk to local people who settle in villages and cities along and above the rivers, lakes and ponds. Additionally, raw water for local water supply stations in the Mekong delta collected from the rivers and canals in this area is treated via a simple process-rock and sand filters in combination with chlorinated disinfection-prior to reaching the local communities [14]. The piped-water provided by these local water supply stations may be contaminated by MCs liberated from cyanobacterial cell lysis due to the chlorination in the water process. Therefore, additional studies should examine the processed drinking water in the presence of MCs.

Risk Assessment for MC-Contaminated Foods
We assessed public health risk for the MC-contaminated apple snails and fish by calculating the estimated daily intake (EDT) for digestion of apple snails and fish ( Table 5). The estimated daily intake (EDI) of MC-LR equivalents (μg/kg/day) is based on the tolerable daily intake (TDI) of 0.04 μg MC-LR per kg of body weight per day over the lifetime of an individual weighing 60 kg and eating 300 g snail or 100 g fish per day, as recommended by the WHO [1]. DW of fish in our study was converted to WW by a conversation factor of 0.311 [34,37,38]. WW of apple snails was estimated by DW and 83% water content in live apple snails [50]. Rivers, lakes, and ponds in the Mekong delta are among the main water supplies for household daily use. These water supplies provided 36% for households' daily demands (besides drinking and cooking) and 25% for the demands of drinking and cooking [48,49] for 17.4 million people. The available surface water from these water supplies can be directly used for domestic purposes, especially in the dry season and cyanobacteria accumulation in the surface water has not been recognized as a potential health risk to local people who settle in villages and cities along and above the rivers, lakes and ponds. Additionally, raw water for local water supply stations in the Mekong delta collected from the rivers and canals in this area is treated via a simple process-rock and sand filters in combination with chlorinated disinfection-prior to reaching the local communities [14]. The piped-water provided by these local water supply stations may be contaminated by MCs liberated from cyanobacterial cell lysis due to the chlorination in the water process. Therefore, additional studies should examine the processed drinking water in the presence of MCs.

Risk Assessment for MC-Contaminated Foods
We assessed public health risk for the MC-contaminated apple snails and fish by calculating the estimated daily intake (EDT) for digestion of apple snails and fish ( Table 5). The estimated daily intake (EDI) of MC-LR equivalents (µg/kg/day) is based on the tolerable daily intake (TDI) of 0.04 µg MC-LR per kg of body weight per day over the lifetime of an individual weighing 60 kg and eating 300 g snail or 100 g fish per day, as recommended by the WHO [1]. DW of fish in our study was converted to WW by a conversation factor of 0.311 [34,37,38]. WW of apple snails was estimated by DW and 83% water content in live apple snails [50]. The golden apple snails, Pomacea canaliculata and Pomacea maculata, are invasive freshwater snails in Vietnam. The snails were introduced for culture as a food source in 1988, but the snails then spread in rice fields, canals, and rivers, becoming a harmful pest in wetland rice culture and other crops [51,52]. The apple snails are rich protein and mineral sources, hence, they are promoted for use as food for humans and as live-food in aquaculture and agriculture. For example, the apple snails are applied as food for striped catfish (Pangasianodon hypophthalmus) fingerlings [53] or as a protein supplement to replace soya bean meal in the diets of ducks in the Mekong delta [54]. The apple snails are being sold on many markets throughout Vietnam and have become a common food for the Vietnamese. Our study indicated that apple snails collected from water bodies with a cyanobacterial bloom contained MC both in their visceral and muscle tissue. The calculated EDI indicated that the viscera of apple snails had accumulated MCs (as MC-LR equivalent) 3 to 50 times higher than the TDI value recommended by the WHO. Therefore, it is important to test the viscera of apple snails collected in water bodies with toxic cyanobacteria on accumulated MCs, since evidently apple snails may pose a health hazard for human consumption in southern Vietnam. MC-LR equivalents found in the muscle tissues of apple snails collected in TriAn reservoir were almost equal to the TDI value. Hence, the consumption of apple snails in southern Vietnam is safer for humans if their viscera is completely removed and no more than 300 g flesh of the snails is consumed.
Ganges river sprat, sknakskin gourami fish, and tilapia are common food for local people in southern Vietnam. Ganges river sprat and sknakskin gourami fish are small-sized, high-value fish and well known as delicious specialties in the Mekong Delta and TriAn reservoir. The whole body of fresh and dried Ganges river sprat and sknakskin gourami fish are eaten by consumers, because time-consuming and labour-intensive actions are required to remove the viscera of these small fish. Moreover, the lipids accumulated in the viscera of these fish also contribute to its delicious flavour. However, the EDI calculated from MC-LR equivalents found in the viscera of Ganges river sprat and sknakskin gourami fish were 16 and 179 times higher than the recommended TDI value, respectively. The EDI values of the visceral mass in tilapia collected in DauTieng and TriAn reservoir were also 3 and 10 times higher than the recommended TDI value, respectively. The EDI value of the gut in tilapia collected in BinhChanh fish pond was 1440 times higher than the recommended TDI value.
The viscera of the fish can account for 10-18% of the whole body weight [55]. Although MCs were not detected in the flesh, the high MC concentrations in the viscera of the fish could possibly pose a health risk to consumers eating the whole fish. It is highly recommended that the viscera and especially, the gut of ganges river sprat, sknakskin gourami fish, and tilapia are removed before the fish is consumed.
The suckermouth catfish is an ornamental fish, but it is also known as an invasive species in fresh and brackish water systems in Vietnam. The invasion of suckermouth catfish in the Mekong Delta has been reported by public media since 2004 [56]. The suckermouth catfish has been sold on local markets as food since 2011 [57]. In particular, the suckermouth catfish was recently used as food for patients with diabetes due to local people believing that the flesh of the fish can reduce diabetic signs; this resulted in more serious diabetes, as reported by Faculty of Endocrinology at Hospital of 115 in HCMC [58]. The MC concentration found in the gut of suckermouth catfish was very high, leading to an EDI that was 1305 times higher than the recommended TDI value. Nevertheless, the gut of suckermouth catfish is completely removed during processing and thus no health hazard is expected from ingesting it. However, the flesh and liver of the catfish showed an EDI 48 and 236 times higher than the recommended TDI value and thus it could be considered a high potential hazardous food.

Cyanobacteria Control in Southern Vietnam
There is currently no management strategy to control cyanobacteria in southern Vietnam. This is due to the fact that cyanobacterial blooms have been considered as harmless phenomena. No information on the acute and long-term impacts of MCs on the health of the local public has been reported in this area. However, this and other studies [22,25,30] show that MCs were found in water and in aquatic animals with concentrations and contents that could be assessed as potentially hazardous to the health of the local public. Therefore, we suggest that cyanobacteria-monitoring programs should be established to identify the spatial and temporal variability of cyanobacterial contamination, to assess the potential hazards to the health of the local public and to warn local people who are at significant potential risk of exposure to cyanobacterial toxins. With predicted climatic changes in southern Vietnam, this becomes even more important, especially as warming temperature can boost the high accumulation of cyanobacterial biomass in water bodies in southern Vietnam.

Conclusions
Eutrophication and hyper-eutrophication with the accumulation of high cyanobacterial biomass were observed in several water bodies (reservoirs, rivers and small ponds) in southern Vietnam. Eutrophication and cyanobacterial blooms in small ponds, where fish and/or ducks were cultivated, were more serious than those in reservoirs and rivers. Microcystis was the main potential toxin producer and the most common bloom-forming species in southern Vietnam. The MC concentrations ranged from <LOD to 11,039 µg/L or to 4033 µg/g DW in field samples. MCs were only found in isolated Microcystis strains and were <LOD in other isolates, including Anabaena, Anabaenopsis, and Planktothrix strains MC-LR and MC-RR variants were most frequently found and the most abundant MC variants in MC-containing field samples. Three MC variants-MC-dmLR, MC-LW, and MC-LF-were recorded in Vietnam for the first time. MC-LR, MC-RR, and MC-LF significantly contributed to the total toxicity of MC-containing samples.
The MC content in fish was higher than in shrimp and snail. MC was mainly found in the visceral mass, liver, and gut, so consuming whole MC-containing fish and snails is not safe. It is strongly recommended that the whole viscera of fish and snails must be completely removed during food processing, especially when the animals are collected from water bodies with a high cyanobacteria biomass. The suckermouth catfish should be considered as an ornamental fish, while it is not a safe food source.
Cyanobacterial monitoring programs should be established to assess and minimize potential public health risks.

Sampling
Seventeen water bodies including rivers, lakes, ponds, canals, and reservoirs in the vicinity of urban settlements throughout the Mekong basin, and in Southeast Vietnam ( Figure 5) were sampled once during the dry season (December-May/June) to assess blooms, potential eutrophication effects and resulting cyanobacteria and MCs. At each collecting site, temperature, salinity, and pH were measured by pH/Cond 340i meter (WTW, Weilheim, Germany). Cyanobacterial-chlorophyll-a was measured with the bbe AlgaeTorch which is a lightweight instrument for the simultaneous quantification of the chlorophyll-a content of cyanobacteria and the total chlorophyll content of microalgae in water. (bbe Moldaenke GmbH, Schwentinental, Germany). Samples from sampling sites where Chl-a was higher than 200 μg/L were measured in a bucket after the dilution of collected scum material with tap water to maintain the advised measuring range for the AlgaeTorch. Cyanobacterial scum samples were also collected for isolation and filtration; sub-samples were preserved (in Lugol's iodine) for microscopic analysis (dominant cyanobacterial species). In addition, 1 to 300 mL of surface water from water-bloom sites was filtered through GF/C filters. The filters and filtrates were stored at −20 °C upon MC analysis. Animals in infested water bodies with cyanobacterial bloom were collected to determine MCs in their tissues.
Samples for nutrient analysis were kept on ice and transported within 24 h to the laboratory of Water Quality, Institute for Environment and Resources where nutrients were analyzed colorimetrically with a spectrophotometer (DR/2010, Hach, Loveland, CO, USA) using the following APHA (2005) [59] methods: Nitrate 4500NO3 − , ammonium 4500NH4 + , total nitrogen (TN) Kjeldahl 4500N, phosphate, and total phosphorus (TP) 4500P. The detection limits of the equipment for these parameters were 0.02 mg/L (nitrate), 0.04 mg/L (ammonium), 0.06 mg/L (TN Kjeldahl), and 0.05 mg/L for both TP and phosphate.

Strains Isolation
In the laboratory, single Microcystis, Planktothrix, Anabaenopsis, Anabeana cells or colonies were picked out of the collected scum material by the micropipette-washing method [60]. These isolates were grown in small glass tubes with a few mL modified WC medium (Woods Hole modified CHU10-medium) [61] for several months at 25 °C, under a 14:10 h light/dark cycle at a light intensity of 70 μmol photon/m/s. When isolates reached a greenish appearance, they were transferred into 50 mL Erlenmeyer flasks and subsequently into 250 mL flasks. In total, there were 24 isolated strains (Tables A1 and A2, Appendix B). At each collecting site, temperature, salinity, and pH were measured by pH/Cond 340i meter (WTW, Weilheim, Germany). Cyanobacterial-chlorophyll-a was measured with the bbe AlgaeTorch which is a lightweight instrument for the simultaneous quantification of the chlorophyll-a content of cyanobacteria and the total chlorophyll content of microalgae in water. (bbe Moldaenke GmbH, Schwentinental, Germany). Samples from sampling sites where Chl-a was higher than 200 µg/L were measured in a bucket after the dilution of collected scum material with tap water to maintain the advised measuring range for the AlgaeTorch. Cyanobacterial scum samples were also collected for isolation and filtration; sub-samples were preserved (in Lugol's iodine) for microscopic analysis (dominant cyanobacterial species). In addition, 1 to 300 mL of surface water from water-bloom sites was filtered through GF/C filters. The filters and filtrates were stored at −20 • C upon MC analysis. Animals in infested water bodies with cyanobacterial bloom were collected to determine MCs in their tissues.
Samples for nutrient analysis were kept on ice and transported within 24 h to the laboratory of Water Quality, Institute for Environment and Resources where nutrients were analyzed colorimetrically with a spectrophotometer (DR/2010, Hach, Loveland, CO, USA) using the following APHA (2005) [59] methods: Nitrate 4500NO 3 − , ammonium 4500NH 4 + , total nitrogen (TN) Kjeldahl 4500N, phosphate, and total phosphorus (TP) 4500P. The detection limits of the equipment for these parameters were 0.02 mg/L (nitrate), 0.04 mg/L (ammonium), 0.06 mg/L (TN Kjeldahl), and 0.05 mg/L for both TP and phosphate.

Strains Isolation
In the laboratory, single Microcystis, Planktothrix, Anabaenopsis, Anabeana cells or colonies were picked out of the collected scum material by the micropipette-washing method [60]. These isolates were grown in small glass tubes with a few mL modified WC medium (Woods Hole modified CHU10-medium) [61] for several months at 25 • C, under a 14:10 h light/dark cycle at a light intensity of 70 µmol photon/m/s. When isolates reached a greenish appearance, they were transferred into 50 mL Erlenmeyer flasks and subsequently into 250 mL flasks. In total, there were 24 isolated strains (Tables A1 and A2, Appendix B).

MC Analysis
The frozen filters stored at −20 • C were transferred to 8 mL glass tubes and dried for two hours in a freeze-drier (Alpha 1-2 LD, Martin Christ Gefriertrocknungsanlagen GmbH, Osterode am Harz, Germany). Tissue and scum samples were also dried for several hours in the freeze-drier and 5 to 8 mg freeze-dried material was then transferred to 2 mL Eppendorf vials.
The filters, the tissue and scum samples were extracted three times at 60 • C in 2.5 and 0.5 mL 75% methanol and 25% Millipore water (v/v). The extracts were then dried in the Speedvac (Savant SPD121P, Thermo Scientific, Waltham, MA, USA) and subsequently reconstituted in 900 µL 100% methanol. The reconstituted samples were transferred to 2 mL Eppendorf vials with a cellulose-acetate filter (0.2 µm, Grace Davison Discovery Sciences, Deerfield, IL, USA) and centrifuged for 5 min at 16,000× g (VWR Galaxy 16DH, VWR International, Buffalo Grove, IL, USA). Filtrates were then transferred to amber glass vials for LC-MS/MS analysis. If needed, samples with high MC concentrations were diluted in methanol before re-analysis.
Concentrations of eight MC variants (dm-7-MC-RR, MC-RR, MC-YR, dm-7-MC-LR, MC-LR, MC-LY, MC-LW, and MC-LF) and nodularin (NOD) were determined by LC-MS/MS as described in [5,62]. LC-MS/MS analysis was performed on an Agilent 1200 LC and an Agilent 6410A QQQ (Agilent Technologies, Santa Clara, CA, USA). The MCs were separated on an Agilent Eclipse 4.6 × 150 mm, 5-µm column. Hereto, a 10 µL sample was injected; the flow rate was 0.5 mL/min; the column temperature was 40 • C. Eluents were Millipore water with 0.1% formic acid (v/v, Eluent A) and acetonitrile with 0.1% formic acid (v/v, Eluent B) that were run using an elution program of 0-2 min 30% B, 6-12 min 90% B, with a linear increase of B between 2 and 6 min and a 5-min post run at 30% B. Detailed information on MS/MS settings for each MC can be found in [41]; information on the recovery, repeatability, limit of detection, and limit of quantification of the analysis is given in [5]. MCs were quantified against certified standards that were obtained from DHI LAB Products (Hørsholm, Denmark).