Dynamic of Filamentous Cyanobacteria in the Dam Ain Zada (North of Algeria)

At present, harmful cyanobacterial efflorescence in Algerian water bodies used for drinking water are dominant throughout the year and their management requires a better knowledge of the cyanobacteria dynamics and the environmental parameters related to their dominance. The Ain Zada reservoir is a home of 5 genera. The occurrence frequency estimation of the identified genera shows the constancy of Planktothrix, the regularity of Aphanizomenon, Pseudanabaena and Cylindrospermopsis and the rarity of Oscillatoria. The dominant genus Planktothrix is represented exclusively by Planktothrix agardhii; this species showed the highest proportion rates in winter (95.61%), spring (94.80%) and autumn (80.29%) and the lowest in summer (29.16%). Planktothrix agardii abundances show positive relation with suspended solids and chlorophyll a and negative relation with the depth of Secchi disc. The blooms of the potential toxinogenic Planktothrix agardhii and Cylindrospermopsis are related negatively. The results from cyanotoxins and pigments characterization showed a strong positive relation of Planktothrix agardhii abundance with microcystins (MCs) and phycocyanin (PC) concentrations. Indeed, in Ain Zada dam, we found that MCs contents have positive relation with water temperature and pH and a negative relation with NO3. Our results show that – except for the months of March, October and November – the monthly abundances of cyanobacterial population recorded in Ain Zada dam exceeded the alert level 2. The filamentous species found in this dam are of great concern, as they are able to synthesize toxins harmful to aquatic and terrestrial organisms. Further research is needed to confirm the presence of other cyanotoxins (e.g. cylindrospermopsin) in Ain Zada dam.


INTRODUCTION
Cyanobacteria are a diverse well-adapted group of organisms; they can be found in a great variety of environments [Whitton et al., 2012].Their strong biomasses are associated with anthropogenic eutrophication (urban and domestic waste, agricultural practices and erosion of fertile soils [Hamilton et al. 2016].According to many authors, climate change could contribute to increasing their frequency and persistence [ According to Paerl & Fulton [2006], the growth and dominance of the bloom-forming cyanobacteria are controlled by abiotic and biotic factors; the factors that promote a species with specific morphology, physiology and ecology will not necessarily promote another in the same way.The ecostrategies applied by cyanobacteria allow them to become dominant.Reynolds [1996], reported that water trophy contributes to the qualitative and quantitative composition of cyanobacterial population; filamentous forms increased and become dominant in the eutrophic or hypertrophic lakes, but their concentration decreased in the mesotrophic and oligotrophic lakes.The distribution of the dominant cyanobacteria could be linked to the specificity of the water body. The cyanobacterial blooms in drinking water sources and at recreational sites pose a serious threat for humans and livestock [Affsa/ Afsset, 2006].Some species of cyanobacteria produce secondary metabolites called cyanotoxins known to provide different effects; microcystins are hepatotoxic, anatoxins and saxitoxins are neurotoxic and cylindrospermopsin is cytotoxic or dermatotoxic.According to Chorus, [2001], 25 to 75% of cyanobacteria blooms are toxic.The previous studies showed that temperature, nutrients and pH contribute indirectly to cellular MC production and cyanobacteria growth rates [Orr & Jones, 1998], which can account for 3 to 4 fold variation in total MC concentrations [Kurmayer et al., 2002[Kurmayer et al., , 2003]].Many studies reported that a climate change and increasing eutrophication would increase the toxic bloom frequency [Dokulil & Teubner, 2011 In Algeria, water is threatened in its quality and its quantity as well as the predicted increasing needs of the population (compared to water shortages) due to the climate changes [Remini et al., 2010].In Algerian lakes and reservoirs, several studies reported the presence of various potentially cyanotoxin-producing species.
In Oubeira lake, Bouaïcha & Nasri, [2004] noted the co-occurence of Microcystis spp and Cylindrospermopsis raciborskii (density estimated to be 43 10 5 trichomes/L).In this water body, Boussadia et al. [2015] reported that cyanobacterial diversity was dominated mainly by Aphanizomenon, Oscillatoria, Cylindrospermopsis, and Microcystis.During a bloom of Microcystis Nasri et al. [2007], noted that the concentrations of Microcystins found in the liver and in the viscera were 1192.8 and 37.19 mg MCYST-LR equivalent/g dw, respectively; this finding supports the possibility that cyanobacterial microcystins is implied in the death of the turtles.Lately, Amrani et al. (2014) noted the presence of MC in lake water of Oubeira at the concentrations ranging from 0.028 to 13.4 g equivalent MC-LR/l; they also reported the accumulation of MC-LR in intestine, hepatopancreas and muscle of common carp (Cyprinus carpio) and European eel (Anguilla anguilla).The presence of Microcystis spp bloom is reported in another natural lake 'lac des oiseaux' of the Northeast of Algeria by Bouhadada et al. (2016); according to these authors, the microcystins analyzed was composed of 21 putative congeners but MC-RR was the major.
In Cheffia dam, Nasri et al. [2007] noted a bloom of Microcystis sp and found in raw water, MCYST-LR equivalent concentrations ranged from 28.8 to 50.8 ng/L.In Mexa reservoir, Saoudi et al. [2015] reported the constancy of Microcystis and the regularity of Oscillatoria; However, Microcystis predominates with an average density higher than 500 000 cells/ml.
In found that Microcystis represent an average proportion of 43% to global cyanobacteria population; this species was followed by Woronichinia 21% and Planktothrix 16%.
The current study was conducted in a hypereutrophic freshwater body subjected to Planktothrix agardhii blooms.Thus, this study aims at i) characterizing the principal environmental factors interacting with the Planktothrix agardhii dynamic in the Ain Zada reservoir, ii) determining the production of microcystins in the Ain Zada reservoir.

Study area
The Ain Zada reservoir is a freshwater body located at 36°10'27.73"N 5°08'58.00"E with an average altitude of 800 m, covering an area of 2080 km 2 with a mean depth of 26 m and total capacity of 121 10 6 m 3 [ANB BBA, 2013].This reservoir (Fig. 1) is fed by three wadis (Boussellem, Taghrout and Kharoua); It is located in region characterized by a semi-arid climate (i.e.harsh winter and dry hot summer) where rainfall varies from 300 to 600 mm and air temperature ranges between 0° and 38°C; prevailing winds are northwest; sirocco is more recurrent in summer.
This reservoir provides drinking water for cities of Sétif and Bordj Bou Arreridj.It is also used for irrigation and as an extensive aquaculture of the royal carp (Cyprinus carpio).

Sample collection, field measurements and chemical analysis
From December 2014 to November 2015, sampling was conducted monthly at sub central part of the reservoir (Fig. 1); the water samples were collected from depth of 0.5 m below water surface.
Cyanobacteria sampling was performed using a phytoplankton net (with 20 μm of mesh size) for cyanobacteria identification and 1 L plastic bottle for cyanobacteria counting and chemical analysis.The in situ measurements of the water temperature (TW), dissolved oxygen (DO), pH, and conductivity (Con) were performed using a 3420 IDS multiparameter probe (WTW, Germany).Transparency was taken with Secchi disc.
The nutrient analyses including Nitrates (NO 3 ), nitrites (NO 2 ), ammonium (NH 4 ) and orthophosphate (PO 4 ) were performed in the laboratory from the water samples stored at low temperatures for the application of spectrophotometric methods [Aminot & Chaussepied, 1983; ISO/TC, 1994].
In order to measure the concentration of Suspended Solids matter (SS), a heat-treated (450°C, 30 min) glass fiber filters of 0.45 μm nominal porosity (47 mm diameter, Whatman GF/CTMTM, Germany) were pre-weighed and then used for the filtration of the surface water samples.After filtration, the filters were dried at 70°C for 48h.The SS concentration was determined from the difference between the weight of the filter before and after filtration [Aminot & Chaussepied 1983].
Cyanobacteria identification was carried out on the basis of microscopic observation of the morphological characters according to the identification keys used by Castenholz, [2001]; [2005].For the cyanobacteria quantification, we used 100 ml of filtrate preserved in formaldehyde (37%) and Nageotte counting chamber as described in Brient et al., [2001].
The Phycocyanin content was determined with the freeze-thaw method [Bennett & Bogorad, 1973] as described by Horváth et al., [2013] in which the samples were frozen at -20°C.Biomass was harvested by filtration on GF/C (Whatman) filters; pigment was extracted with 10 ml of phosphate buffer (pH 7.0) and its concentration was determined using a spectrophotometer at 620 nm 650 nm and 750 nm.

Cyanotoxin contents analysis
For MCs analysis water samples were filtered through microfiber glass filter of 0.45 μm nominal porosity (47 mm diameter, Whatman, Germany), the filters were then kept frozen (-20°C) until toxins analysis.The intracellular MCs concentrations were quantified using a commercial enzyme-linked immunosorbent assay (ELISA) kit (ABRAXIS, USA) with a mean lower detection limit of 0.15 ppb of MCs.After extracting MCs with a methanol/water solution (0.8V:0.2V).The essay was conducted in 96 wells plate and readings were taken at 450 nm wavelength using a microplate ELISA photometer (MindrayMR-96A)

Statistical analysis
For the statistical analysis of the data, we first applied the Shapiro-Wilk test then the nonparametric tests of Spearman and Kruskal-Wallis.Finally, a PCA was performed using the R facto package Miner.All statistical analyses were performed using R software (3.1.2).

Temporal variation of the environmental variables in the dam Ain Zada
In Ain Zada dam, the temperatures of water vary from 5.6 to 27°C (mean T°= 17.19°C ±7.20); temperatures of more than 18°C were recorded from April to October.The waters of Ain Zada showed mean dissolved oxygen contents of 10.30 mg/l (±5.12); period extending from December to July showed more than 10 mg/l of dissolved oxygen contents ((Fig.2a).In Ain Zada, pH is alkaline (mean pH value of 8.73 (±0.49); however, the pH values of more than 9 were recorded from April to July (Fig. 2b).The conductivity values in Ain Zada waters ranged from 900 to 1394 μs/cm (mean value of 1033 μs/cm±169.11);values of less than 1000 μs/cm were witnessed during the period between March and October.The values of suspended solids ranged between 3.2 mg/l and 185.55 mg/l (mean value = 34.91 mg/l ±50.26);Peaks were recorded in January (36.8 mg/l), February (47.60 mg/l), April (185.56 mg/l) and June (53.33mg/l) (Fig. 2e).Transparency (Secchi disc) values ranged from 0.35 to 1.45 m; transparency values of less than 0.50 m were recorded from February to July (Fig. 2e).
The ammonium values in Ain Zada dam water (Fig. 2d) ranged between 0.04 and 3.66 mg/l (mean of 0.81 mg/l ±1.03); the ammonium content (NH 4 ) did not exceed 0.6 mg/l except in July (1.27 mg/l), September (3.65 mg/l) October (1.57mg/l) and November (1.11 mg/l).NO 3 showed values ranged from 0.56 to 4.87 mg/l (mean of 2.95 mg/l ±1.47); the values greater than 4 mg/l were recorded from September to November and in February (Fig. 2c).The value of NO 2 ranged between 0.01 and 0.28 mg/l (mean of value = 0.1 mg/l±0.08); the NO 2 values remained below 0.11 mg/l except in March (0.18 mg/l), April and May (0.27 and 0.24 mg/l respectively).The orthophosphate contents (PO 4 ) ranged from 0.01 to 0.30 mg/L (mean value = 0.12 mg/l ±0.07); values of more than 0.11 mg/l were recorded in March and April and during the autumnal period (Fig. 2d).
The Kruskal Wallis test showed the existence of significant month-to-month differences (p <0.05) for each of the variables measured.

The frequency occurrence of the identified cyanobacteria in the dam of Ain Zada
The microscopic characterization of the cyanobacteria communities of Ain Zada dam revealed the presence of five genera belonging to the order of Oscillatoriales and Nostocales.The occurrence frequency estimation of the identified genera (Table 1) showed the constancy of Planktothrix, the regularity of Aphanizomenon, Pseudanabaena and Cylindrospermopsis and the rarity of Oscillatoria.The main representatives of the cyanobacterial community were Planktothrix, Cylindrospermopsis and Aphanizomenon (Table 1).

Temporal abundance of the identified cyanobacteria in the dam of Ain Zada
The monthly variation of the total cyanobacteria population recorded in the reservoir of Ain Zada dam showed that the abundances exceeded the alert level 2 throughout the year except for the months of March, October and November (Fig. 3).
In Ain Zada dam, the monthly Planktothrix abundance values ranged from undetected in September to 2 444 800 cells/ml in June with a mean value of 674 181 cells /ml.This genus was strongly present from April to July where abundance values exceeded one million cells/ml (Fig. 4).
The monthly abundance values of Cylindrospermopsis genus varied from 0 to 1 694 933 cells/ml with a mean value of 144 848 cells/ml.The highest value was recorded in September and three peaks of 9 485, 13 637 and 13 201 cells/ ml were noted in August, July and December, respectively (Fig. 4).Aphanizomenon showed that the monthly abundance values ranged from 0 to 854 400 cells/ ml (mean value of 132 679 cells/ml).This genus showed four peaks of 28 800, 278 400, 406 400 and 854 400 cells/ml respectively in May, April, August and September (Fig. 4).
Spearman correlation test showed a strong negative correlation between Pseudanabaena and DO (r=-0.76;p<0.01), and positive correlation between Pseudanabaena and NH 4 (r=0.66;p<0.05) and conductivity (r=0.64;p<0.05).During this study, Oscillatoria was present only in November with the abundance value of 5600 cells/ml.

Temporal variation of the Chlorophyll a and phycocyanin
The monthly chlorophyll a contents ranged from 13.25 to 172.79 µg/l.The highest values of chlorophyll a contents (more than 100 µg/l) were recorded from April to June; values ranged from 50 to 100 µg/l were noted in January, July, August, September and December.The lowest values (less than 50 µg/l) were recorded in February, March, October and November (Fig. 5).The monthly phycocyanin contents ranged from 39.97 (October) to 387.39 µg/l (June).This pigment showed highest levels in spring and summer and the lowest in winter (Fig. 5).
The non-parametric Spearman correlation showed a strong positive correlation between the chlorophyll a contents and the abundances of cyanobacteria (r=0.75;p<0.01), and a very high positive correlation between the chlorophyll a contents and phycocyanin contents (r=0.9;p<0.001).

Temporal variation of the intracellular microcystins content
The microcystins levels recorded monthly in the water of Ain Zada dam showed the values of more than 1 µg/ml from April to July and in November (Fig. 6); two peaks of 4.77 and 3.79 µg/l were recorded in June and July, respectively.During the rest of the year, microcystin contents did not exceed 0.5 μg/l.

Relationship between environmental variables and cyanobacteria in the dam of Ain Zada
The first two main components (Table 2-3, Fig. 7-8) of the PCA carried out on the

DISCUSSION
The microscopic characterization of the cyanobacteria communities in the water of Ain Zada dam revealed the exclusive presence of filamentous cyanobacteria belonging to the orders of Oscillatoriales and Nostocales.The occurrence frequency estimation of the identified genera showed the constancy of Planktothrix, the regularity of Aphanizomenon, Pseudanabaena and Cylindrospermopsis and the rarity of Oscillatoria.The dominant genus Planktothrix was represented exclusively by P. agardii.Blooms of this species were observed from winter to summer with a proportion rate exceeding 80% from the whole cyanobacteria.
According to Scheffer et al. [1997] the growth of Planktothrix agardhii is favored by the conditions of greater turbid waters with a high disturbance frequency.In addition, Nixdorf et al. [2003] place these species in the phytoplankton functional group S1 [sensu Reynolds et al., 2002] of photoadapting filamentous cyanobacteria common in shallow and wind-exposed lakes with strong interactions between sediment and water and having low light levels.These findings were supported by the high positive correlation of P. agardhii abundances and the measures of suspended solids observed in the reservoir of Ain Zada.
According to Dokulil and Teubner [2000] and Kokocinsky et al. [2010], in temperate lakes, low temperature did not stop the growth of P. agardhii; Poulickova et al. [2004] noted , that P. agardhii in eutrophic lakes grow even in winter.This could explain the low positive correlation of P. agardhii abundances with the water temperature in the dam Ain Zada; Moreover, the cells abundances of this species were correlated positively with chlorophyll a and negatively with the depth of Secchi disk.
According to the OCDE classification scheme, the chlorophyll a and transparency values noted in Ain Zada dam showed that this water body fluctuated between the eutrophic and hypereutrophic status [Vollenweider and Kerekes, 1982].
The lower values of transparency observed in the Ain Zada dam explained the mass occurrence of Planktothrix agardii throughout the year.In Ain Zada dam, high levels of phosphorus were observed during the Planktothrix agardii bloom.However, no correlations were noted with phosphorus.In Zit Emba reservoir, Touati et al. [2019] reported the positive correlation between the PO4 -P concentrations and the Planktothrix abundance suggesting that the growth rate of this genus was increased by high orthophosphates concentrations.This corroborates the results reported by Bidi-Akli et al. [2017] in Zeralda dam (Algeria).
According to Kokocinski et al. [2010], Planktothrix agardhii and Cylindrospermopsis raciborskii could grow under different environmental requirements.This finding is in good concordance with the dynamics of these two cyanobacteria in Ain Zada reservoir.Indeed, a bloom of Cylindrospermopsis was observed in September and the presence of this species overlapped with the highest transparency and highest ammonium values and the absence of Planktothrix agardhii.Our results show the impact of the light on the composition of the phytoplankton community; we noted, indeed, a community was dominated by P. agardhii in very turbid waters, in relation to more diverse communities including the invasive Cylindrospermopsis in clearer waters.Several studies indicated that C. raciborskii is generally more restricted to dry periods with low rainfall [Chellappa and Costa, 2003] and high temperatures [Bouvy et al., 2006].In the reservoirs, during dry periods, the long period of water retention provide good conditions of temperature and irradiation for the dominance of the species C. raciborskii [Bouvy et al., 2000].
In Ain Zada dam, the results from cyanotoxins and pigments characterization showed a strong positive correlation of P. agardhii abundance with the concentrations of microcystins (MCs) and phycocyanin (PC); this led us to conclude that the detected MC and PC in the collected samples derived from this cyanobacterium.The species P. agardhii is known for synthesizing a variety of microcystin [Carmichael, 1994;Chorus, 2001;Kurmayer et al., 2004;Welker et al., 2004;Rohrlack et al., 2008;Solis et al., 2012], while C. raciborskii is able to produce the neurotoxin 'saxitoxin' [Lagos et al., 1999;Castro et al., 2004] and the hepatotoxin 'cylindrospermopsin' [Ohtani et al., 1992;Li et al., 2001].According to Cirés and Ballot (2016), several members of the genus Aphanizomenon have the ability to develop blooms and to synthesize microcystins (hepatotoxins), anatoxins and saxitoxins (neurotoxins) and cylindrospermopsin (cytotoxins).Pseudanabaena species could represent a potential risk for exposed fauna, because the presence of neuro and hepatotoxins in extract and neurotoxin in growth medium of Pseudanabaena galeata has been shown (Teneva et al., [2009]).
According to Codd [2000], the fact that toxigenic cyanobacteria are present in these drinking water production reservoirs involves regular monitoring of cyanobacterial populations and cyanotoxins in raw water.This monitoring should be undertaken, due to its potential health risk for drinking water and for bioaccumulation of cyanotoxins in the flesh of fish [Jia et al., 2014;Amrani et al., 2014].The establishment of a wastewater treatment could help to reduce nutrients loads discharged in the watershed in order to prevent eutrophication and development of toxic cyanobacterial blooms.

CONCLUSION
Blooms of the potential toxinogenic P. agardhii and Cylindrospermopsis were related negatively.The occurrence of blooms of P. agardhii in the dam of Ain Zada was observed from winter to summer.In turn, the blooms of the invasive Cylindrospermopsis were observed in autumn.This succession was explained by their possible different responses to environmental constraints resulting in contrasting strategies.The most important environmental variables probably leading to this succession was the water transparency and the suspended solid matter (SS).Thus, P. agardhii is favored by turbid conditions, while Cylindrospermopsis prefers less turbid waters.

Fig. 3 .Fig. 4 .
Fig. 3. Monthly variation of the total cyanobacteria population recorded in the dam Ain Zada

Table I .
Occurrence and proportional rates of identified cyanobacteria in Ain Zada dam

Table 3 .
Spearman correlation coefficients [Kuiper-Goodman et al., 1999;Xie et al., 2007; an increase in MCs concentrations which overlapped with a strong increase in P. agardhii.Moreover, in PolandSolis et al. (2012)found a correlation between MC-RR level and P. agardhii abundance in the eutrophic reservoirs of Tomaszne.According toBarco et al. [2004];Kurmayer et al. [2004];Welker et al. [2004], in "Planktothrix lakes", the demethylated MC-RR should be the principal microcystin variant.According toKurmayer and Christiansen [2009], temperature, nutrients and pH can induce changes in the cyanotoxin content in individual strain, by a factor of no more than three to four.Indeed, in the Ain Zada dam, the MCs contents are correlated positively with water temperature (r=0.64) and pH (r=0.72) and negatively with NO 3 (r=-0.63).Except for the months of March, October and November the monthly abundance of cyanobacterial populations recorded in Ain Zada dam exceeded the alert level 2 [Affsa/Afsset, 2006].The strong and permanent dominance of P. agardhii over other toxic species impacted the water quality negatively.The filamentous species met in the Ain Zada dam were of great concern because they can synthesize toxins harmful to aquatic and terrestrial fauna[Kuiper-Goodman et al., 1999;Xie et al., 2007; Pawlik-Skowronska et  al., 2012]