Species Diversity of Cyanobacteria and Desmids of a Drinking Water Source under Anthropogenic Pressure, and Their Implication in Toxin Production and Water Quality in Sub-Saharan Africa (Burkina Faso, Western Africa)

Drinking water sources in many African countries have been progressively degraded over the past decades. This degradation due to human activities leads to the proliferation of algae, especially toxin-producing cyanobacteria. The presence of toxigenic algae in water has adverse consequences on human and animal health. This study aimed to determine the diversity and density of Cyanobacteria and Desmids and to identify toxin-producing cyanobacteria and environmental variables that influenced the structure of these groups of microalgae in the Loumbila reservoir in Burkina Faso located in the western part of Africa. Algal samples were collected and physico-chemical parameters were measured. Plankton species were observed under a light microscope and identified using standard methods. Species density was determined by cell counting using a Fuchs-Rosenthal chamber. Kruskal Wallis and Pearson correlation tests were performed using R software. A canonical analysis was performed using CANOCO software. In total, 205 algal species were inventoried, of which 83 species composed of 37 species of Cyanobacteria and 46 species of Desmids were identified. Microcystis aeruginosa, Staurodesmus convergens and Cosmarium connatum var africanum had the highest presence index respectively (100%, 83.333% and 77.77%). Among cyanobacteria species, toxin-producing cyanobacteria was positively correlated (p < 0.05) with temperature, pH, dissolved oxygen, transparency, nitrates, and orthophosphates. However, at p < 0.05, desmids community was only correlated with dissolved oxygen, transparency, and conductivity. Furthermore, canonical analysis showed that temperature, dissolved oxygen, transparency, and orthophosphates influenced the density of both cyanobacteria and Desmids. These results reveal the high occurrence of toxin-producing cyanobacteria and certainly high toxins produced in the drinking water source. Basic tools should be developed for monitoring of cyanotoxins in drinking water sources and drinking water supplied to population to consider cyanotoxins during water treatment.


Introduction
Drinking water supply is a vital issue for people living in the African continent, as the continent is experiencing high population growth thereby increasing water needs [1]. Few developing countries have the capacity to control the quality of this water [2]. More than 80% of wastewater worldwide, and more than 95% in some developing countries, are still discharged into the environment without treatment [2] impacting water resources. In 2017, the number of people without access to basic water services was 785 million worldwide, including a worrying number of people drawing water directly from surface water points, estimated at 144 million [3]. Many countries face a gap between the growing demand for water and the availability of knowledge about the water quality of potential surface water resources [2].
In Burkina Faso, specifically in the city of Ouagadougou, the population has experienced an average annual growth rate of about 6.4% [4]. This galloping population growth is not without consequences for the growth of the city, creating other problems, notably that of drinking water supply [5]. The city is supplied by the "Office National de l'Eau et de l'Assainissement (ONEA)", the only public drinking water utility. This company obtains water from two reservoirs, one of which is the Loumbila reservoir to serve the city of Ouagadougou. The water from this reservoir is used for market gardening, watering animals, fishing and major public works. These activities, especially agricultural, have led to the advanced degradation of its quality. However, in this reservoir, which is so important for its various functions, few data are available on the microalgal communities, which represent one of the most important communities for assessing the quality of aquatic environments. The only study available on the microalgal community of this reservoir is the work of Zerbo [6] that was a major contribution to the knowledge of phytoplankton. In fact, the degradation of an aquatic ecosystem leads to algal blooms that considerably reduce the availability of oxy-gen and light in the water [7]. Some algal groups, such as Desmids, due to their low tolerance to inorganic salts are indicators of trophic status and water pollution [8] [9]. Cyanobacteria are highly adaptable to colonize terrestrial and aquatic ecosystems [10]. They can become abundant with high nutrient concentrations (nitrates and orthophosphates), making them important indicators of water quality [9]. Small variations in physico-chemical parameters such as pH, conductivity, dissolved oxygen and total hardness can contribute to the development of a diverse Desmid community [11]. Some species of Desmids are considered acidophilic, acid-neutrophic and oligotrophic [12]. In addition, the soil, the morphological characteristics, and the low water flow of the water body have an effect on the formation of a specific community of cyanobacteria [13] and desmids also.
Despite being indicators of water quality, some microalgae are clearly circumscribed in their ability to produce specific metabolites including toxins [14] [15]. Toxic cyanobacterial species have been reported in freshwaters in over 45 countries and in many brackish, coastal and marine environments [16] where some synthesize biotoxins and produce metabolites responsible for the bad taste of water and fish flesh and their death, unpleasant odors, for public health problems [17] [18] [19]. Considering species diversity of cyanobacteria and desmids, and their implication in toxins production and water quality of the Loumbila reservoir in Burkina Faso, this study aimed to characterize the pollution level of a water body under anthropogenic pressure and provide related scientific data for future protection of populations using drinking water supplied from such source. Specifically, the aim is to determine the diversity and density of the two groups of algae in Loumbila reservoir, to establish the relationship between them and water quality of the reservoir, to determine on the basis of available literature, cyanotoxin-producing species and related produced toxins in water, to show the importance of taking the elimination of these toxins into account in the drinking water treatment process.

Study Site
This study was carried out in the Loumbila reservoir ( Figure 1 and Figure 2) located about 20 km north-east of Ouagadougou. This reservoir is located between 12˚28' and 12˚35' North latitude and 1˚23' and 1˚30' West longitude. This locality is subject to a north Sudanian climate, with a dry season from October to May and a rainy season from June to October. The maximum of rain is observed in August. The average annual rainfall varies between 700 and 800 mm but can reach 1000 mm. The reservoir has an estimated current area of 1500 ha and a maximum depth of 6.6 m [20]. Three (03) sampling points were selected along a longitudinal gradient using a GPS (Global Position System) and physically marked during the first fieldwork. These three points were chosen considering some physical parameters such as transparency and position of the station, Journal of Water Resource and Protection  availability and accessibility of water during the study period (January 2014 to June 2015). Thus, the first station (point 1) is characterised by a higher water B A transparency than the other two. Its banks are protected by stone barriers, filter dikes and grass bards (anti-erosion devices). The second station (point 2) is characterised by agricultural activities on part of the water body in this area during the dry season. The third station (point 3) is characterised by the presence of a beach where activities such as swimming and the use of motorised boats are frequent, resulting in a high level of water agitation at this point.

Collection of Biotic and Abiotic Data
Samples were collected monthly over a period of 18 months going from January 2014 to June 2015. During the sampling days, they were collected between 11:00 am and 1:00 pm. Thus, two types of samples were taken. The first sample was collected with a cylindrical-conical plankton net of 20 μm mesh and 25 cm diameter. This sample was used for the qualitative analysis of microalgae. The second type of sample was collected with a 1 litre bottle and used for the quantitative analysis. After each collection, the contents of the net were emptied into 50 ml jars and the 1litre bottle content were immediately preserved with 8% formaldehyde. In addition, during the collection of the algae, pH and temperature were measured with HANNA handheld electronics (model HI98129), conductivity and dissolved oxygen with a conductivity meter model WTW 3110SET1 and an oximeter model WTW 3205SET3 respectively. Water was collected using 500 ml jars and stored according to the method described by [21] for analysis of the main algal nutrients (nitrates and orthophosphates) in the laboratory using a DR 3800 ® molecular absorption spectrophotometer.

Identification of Algal Species
Algal samples were examined, photographed using an OPTIKA B-350 microscope and a SONY camera. Species identification was carried out on the basis of the images taken using standard works [22] [23] [24] [25] [26] and following the taxonomical criteria of AlgaeBase (www.algaebase.org/search/species/).

Quantitative Analysis of Microalgae
The assessment of the algal biomass in the reservoir was done by determining the density of algae in sub-samples (5 ml/sample) obtained from initial samples.
The method described by Zongo et al. [7] was used to determine the density of algae in each subsample. For this, a Fuch-Rosenthal chamber was filled with a homogeneous subsample. For the particles to settle, this subsample was kept stable for 5 minutes before counting all the cells in the chamber up to four times for each sample. Thus, the following formula was used to determine the density of the algae [7].
D is the density of the algae; N is the number of individuals per chamber; V is the volume of the sample taken and v is the volume of formalin used for preser-vation

Classification of Cyanobacterial Species According to Their Toxin Production
This classification concerns toxin-producing cyanobacteria. As toxin analyses have not been carried out during this activity, identification of toxin-producing cyanobacteria was based on previous studies [16] [17] [27] [28]. Thus, according to Bernard [17], they were grouped into three families according to the cyanotoxic effects: hepatotoxins (main target organ: the liver), neurotoxins (target organ: the nervous system) and dermatotoxins (target organ: the skin). Depending on the mechanism of toxicity, the same toxin can be produced by different genera and a genus can produce several types of toxins [27].

Statistical Analysis of Data
To test the variability of physico-chemical parameters between the sampling sites defined in the reservoir, a Kruskal-Wallis test was performed to compare these parameters between sites. Cyanobacteria and Desmids indices were used to compare the presence of these algal groups at the same site and across the three sites. In addition, the presence indices of toxin-producing and non-toxin-producing cyanobacteria were also compared. For this purpose, Anova and Kruskal-Wallis tests were applied to these different data respectively. The software R version 3.6.1 was used for these different analyses.
Correlation analyses were performed with the same software to determine the correlation between environmental variables and the density of Desmids, cyanobacteria, toxic and non-toxic cyanobacteria. The pH, transparency, conductivity, dissolved oxygen, nitrate, orthophosphate and monthly density of these algal groups were used for this analysis. The correlation matrix was established to determine the correlation coefficients between the different parameters and the density of algal groups. In fact, the correlation coefficient makes it possible to determine and measure the intensity of the relationship between two variables. When this coefficient is equal to 1, it shows an absolute link between the variables and when it is equal to 0, it shows an absence of link. The calculation of this coefficient gives an idea of the possible relationships between parameters.
To study the distribution of phytoplankton genera in relation to environmental parameters in the Loumbila reservoir, a redundancy analysis (RDA) was performed. Matrices representing the monthly densities of the different genera and the monthly evolution of physico-chemical parameters were used for this analysis. It allows analyzing the links between environmental and biotic variables [9] and cross-reference monthly microalgae density data with monthly environmental data. Based on a logarithmic transformation log(x + 1) of the phytoplankton genus density, the Monte Carlos permutation test (permutation 499) was performed to highlight the environmental variables that best influence the genus density. The RDA was performed using CANOCO for Windows version 4.5. Journal of Water Resource and Protection Table 1 presents the spatial variations of the mean values of the environmental factors of the Loumbila water. The Kruskal-Wallis test shows that at p < 0.05, physico-chemical characteristics of the three sites were different except temperature (p > 0.05). Thus, except for the mean values for dissolved oxygen and transparency, the other parameters showed higher mean values at site 2 than at the other sites. However, the results in Table 2

Diversity of Cyanobacteria and Desmids
During this study, 205 algal taxa were identified in the Loumbila reservoir. Of these, Desmids and cyanobacteria had the highest number of taxa (46 and 37 species respectively). Due to the high number of these algal groups and their ecological importance in water quality, only these two algal groups were considered for the different analyses. The species of cyanobacteria and Desmids were divided into 12 and 6 genera respectively. Among these different genera identified, the genus Cosmarium (34.7%) was the most diverse (16/46 species) of the  Of the 12 genera of cyanobacteria studied, 10 genera composed of 30 species were identified as toxin producers ( Table 3). The number of species producing microcystins is the most important with 30 species producing this toxin ( Figure 4) out of 37 species in total. They are followed respectively by those producing anatoxins, cylindrospermopsins, saxitoxins, aphysiatoxins and lyngbyatoxins. Thus, the Kruskal-Wallis test carried out between the presence indices of toxin-producing Table 3. Cyanobacteria genera and their associated toxins in the Loumbila reservoir.   and non-toxin-producing cyanobacteria showed a significant difference (p < 0.00129). A variety of toxin-producing species of cyanobacteria was obtained ( Figure  5). The list of cyanobacterial species and their associated toxins (Table 4) indicates that the same toxins can be produced by several species of cyanobacteria and one species of cyanobacteria can also produce several types of toxins.
The canonical analysis of the environmental variable and the density of the different genera of Cyanobacteria and Desmids showed a strong correlation  between these variable ( Figure 6). Temperature (Temp) (p < 0.002), dissolved oxygen (Oxy) (p < 0.018), transparency (Transp) (p < 0.004) and orthophosphates (OrthoP) (p < 0.002) significantly influenced the density of Cyanophyceae

Occurrence of Cyanobacteria and Desmidiaceae in the Loumbila Reservoir
In the Loumbila reservoir, Cyanobacteria and Desmids showed a remarkable diversity with 37 and 46 species recorded respectively. Chlorophytes, more precisely Desmids, were the most diverse. The high diversity of species in this family is justified by their ability to adapt and grow under various physico-chemical and environmental conditions [23] [29]. In fact, Desmids are characteristic of freshwater [8]. The considerable diversity of this group of algae was observed in the Bagré reservoir [23] and in temporal ponds [9]. A strong predominance of desmids was found in the Bia River [24] and in the Bandama River [30] of Ivory Coast, in the aquatic systems of Brazil [31], in the high mountain lakes of the Artabel Lakes Natural Park (Gümüşhane, Turkey) [32]. The diversity of this algal group is marked by the genus Cosmarium which accounts for 34.7% of all Desmid species. This could be related either to the ecological characteristics of this genus or to the very high diversity of this genus. In fact, the highest diversity of Cosmarium species was recorded in dystrophic ecosystems by [29]. Furthermore, among Desmid species, Staurodesmus convergens (Ehr.) Teil. and Cosmarium connatum (Breb.) Ralfs var. africanum Fristch and Rich had the highest occurrence index. That indicates an adaptivity of the two species to the environmental conditions of the Loumbila reservoir. Species like Staurodesmus convergens can even regenerate its capsule in the absence of nutrients such as nitrates and orthophosphates [33].
The diversity of cyanobacteria also characterises the environmental conditions of the waters of the Loumbila reservoir. Indeed, the genus Oscillatoria was the most diverse (21%) compared to the other genera of cyanobacteria recorded in the Loumbila reservoir. This could be due to the environmental characteristics of the water in this reservoir. In fact, Oscillatoria species prefer to grow in water bodies containing hard, alkaline, polluted water rich in essential nutrients [34] [35]. Among the cyanobacteria species, Microcystis aeruginosa had the highest presence index (100%). This could be explained by the ability of this species to grow at different states of water transparency in the reservoir. In fact, Scott [36] showed that the ability of M. aeruginosa to grow well beyond a range of light intensities gives it a competitive advantage over other algae. In addition, temperatures above 25˚C are favourable for the development of this species [37] and, at the same time, favourable for the increase of toxins (microcystins) in the water [38]. Consequently, its excessive development in an aquatic environment can constitute a real health problem for water users (animals and humans) in this environment.
In the Loumbila reservoir, no significant difference (p > 0.5) was observed between the diversity of desmids and cyanobacteria at the same site. Physico-chemical quality of water in the reservoir seems to be favourable to the development of both Desmids and cyanobacteria groups. They both have high plasticity and occur in various freshwater habitats although Desmids have a preference for oligotrophic environments [11] [39] [40]. Furthermore, cyanobacteria are an extremely diverse group whose adaptability and ability to tolerate extreme conditions allow them to be found worldwide [40] [41]. However, Desmids are distributed in freshwaters at very low conductivity and nutrient concentration [42]. Furthermore, there was spatial variation in the diversity of these two algal groups (p > 0.05) between the three sampling sites. This indicates that the different sampling sites have different environmental characteristics. Thus, the spatial variation in phytoplankton diversity is due to the disparity in ecological distributions of organism types and variability in climatic and geographical conditions [43].

Physico-Chemical Variables and Algae Structure
The physico-chemical parameters of the water in the Loumbila reservoir vary greatly from one site to another, with exception of temperature. Temperature is mainly impacted by climatic conditions which are practically the same throughout the reservoir. Thus, variables such as air temperature, solar radiation, wind speed, flow rate, groundwater flow can have an impact on the water temperature in a river [44]. Compared to the Loumbila water temperature (Table 1 & Table   2), similar water temperatures have been noted by other authors in the warm tropics where they range from 25˚C to 32˚C [45] [46]. However, with the exception of the mean values for dissolved oxygen and transparency, the other parameters showed higher mean values at site 2 than at the other sites. This could be due to the consequences of the different anthropogenic activities carried out at this site. This is confirmed by the work of Kuffour et al., [47] who indicated that the nitrate level in Lake Bosomtwe in Ghana could be attributed to the use of nitrogen fertilizers along the lake shore by most farmers. In fact, this site is subject to very intense anthropic pressure with market gardening and water abstraction by public works companies. Numerous studies [48] [49] [50] have indicated that the intensification of agricultural activities is causing serious ecological problems and is the main cause of water source degradation. The low values of transpa-rency and dissolved oxygen at site 2 could be related either to irrigation backwash that clouds the water or to the presence of agricultural activities producing organic matter into the water. In fact, Sánchez-Colón and Schaffner [51] have indicated that the low transparency of the water is due to the high amounts of inorganic and organic matter in suspension. Also, the presence of oxidizable organic matter can lead to a decrease in dissolved oxygen concentration due to oxygen depletion through aerobic decomposition of organic waste by microorganisms [52].
Furthermore, physico-chemical parameters such as pH, transparency, dissolved oxygen, conductivity, nitrates, and orthophosphates content peaked in March. This could be explained not only by the degradation of organic matter from agricultural soil leaching but also by the clarity of the water at this time. In fact, some of these parameters are related to each other. Water clarity affects the amount of oxygen produced by the environment [52] which in turn is involved in oxidation reactions to transform ammonium into nitrates and phosphorus into orthophosphates. In the same way, Ouattara et al. [5] demonstrated that nitrates are produced by the oxidation of ammonium to nitrate and then to nitrite.
However, during the period of July and August characterized by the strongest rainfall, these parameters were observed to be low in water of the Loumbila reservoir. This is mainly related to the dilution of water by rainwater and the non-degradation of agricultural organic matter.
Overall, the conductivity of the water was low at all three sites (<100 µS·cm −1 ). This could be explained by the low concentrations of ions such as calcium (Ca 2+ ), magnesium (Mg 2+ ), sodium (Na + ), potassium (K + ), bicarbonate ions (HCO 3− ), sulphate ions ( 2 4 SO − ) and chloride ions (Cl − ) in the water [5]. In fact, the higher the concentration of these ions in water, the higher the conductivity of the water. Furthermore, the average pH values (6 -8)  Average nitrates and orthophosphate levels were respectively high than 0.3 mg·l −1 indicating high nitrates and phosphates concentrations in water of Loumbila reservoir. At certain content in water, nitrates and phosphates characterized a eutrophication of water and favor the proliferation of microalgae [23]. Their presence in river water favors the increase of primary productivity and their excess in surface water is considered as a warning for algal blooms [54]. This is confirmed by the results of the analyses (Pearson correlation test and canonical analysis) which showed that orthophosphates have an influence on the density of cyanobacteria. Phosphorus remains the predominant factor in water to increase productivity and the trophic status index [55] [56]. Indeed, [57] showed that phosphorus concentration was the key in the regulation of nitrogen-fixing cyanobacteria, such as Microcystis aeruginosa. This is supported by Djabourabi et al. [58] and Taffouo et al. [59] who observed a positive correlation of orthophosphates with Cyanobacteria.
Except for conductivity, other parameters such as pH, temperature, dissolved oxygen, transparency, nitrate and orthophosphate showed a positive correlation with cyanobacteria and toxin-producing cyanobacteria. This could be explained by the adaptability of cyanobacteria to extreme conditions. Indeed, cyanobacteria are known to be highly adaptable and have been found in a wide range of aquatic and terrestrial environments and even in every imaginable habitat [19] [40]. However, it is only dissolved oxygen, transparency and conductivity that have a positive correlation with the density of Desmidiaceae. This could be due to the ecological characteristics of this group of algae to grow in waters less polluted by nutrients (nitrates and orthophosphates) and transparent. The sensitivity of this group of algae to environmental conditions is often used in water quality assessments [60] [61].

Occurrence of Toxin-Producing Cyanobacteria in the Reservoir
The toxin producing cyanobacteria are becoming more and more common worldwide and pose a danger to drinking water [41].  [68]. As observed in reservoirs of Ghana [69], in most common waters of South Africa [70], M. aeruginosa species is found to be the most abundant species that could produce important microcystin in the water of the Loumbila reservoir. The higher occurrence of toxin-producing species compared to non-toxin-producing species indicates degradation conditions of water quality of a drinking water source favourable to toxin-producing species. A remarkable presence of toxin-producing cyanobacteria may lead to the increase of cyanotoxins in the water habitats and that is consequently a health problem for the populations supplied in drinking water from these water sources.

Considering Treatment of Algal Toxins in African Countries
The identification of toxic cyanobacterial species in Loumbila Reservoir is an important database that helps know the different types of toxins that could occur in the Reservoir. Consequently, we can imagine the impact of toxin producing species on water quality and thereby on the health of population using water. In fact, so far algal toxins are not considered in the drinking water treatment process in most developing countries. Of the 54 countries in Africa, only 21 have notable research information in the area of cyanobacterial blooms over the past decade and 11 of the 21 countries have recorded toxicity and physicochemical parameters related to the presence and abundance of toxic cyanobacteria and cyanobacterial blooms [67]. Although the fact that cyanotoxins are the most studied in freshwaters worldwide, and that there are undoubtedly many cases of problems related to these algal toxins, they remain largely hidden in water quality and drinking water supply records [71] [72]. Indeed, the presence of cyanobacteria and their toxins in water bodies used as drinking water source is a problem for the services that make these waters drinkable [73]. Cyanotoxins can be present in drinking water, depending on their concentration in surface waters and the effectiveness of the treatment method [41]. In countries without adequate water treatment, the population may be exposed to intracellular toxins and dissolved cyanotoxins [28]. These cyanobacterial toxins are structurally diverse and cause numerous health problems in humans and animals [ [77]. Therefore, given these adverse effects on human health, people who drink water containing these cyanotoxins may fall victim to these various diseases in countries where the removal of cyanotoxins is not taken into account during drinking water treatment process. It is imperative and even wise to identify the different types of cyanotoxins and to take them into account in the drinking water treatment process in African countries, especially sub-Saharan countries. Under these conditions, chlorine can be combined with adsorption with activated carbon to remove these cyanotoxins from these waters during drinking water treatment [41].
But the cost of advanced treatment for cyanotoxin removal is high and the chemicals used in the treatment process have health drawbacks [41]. Therefore, in developing countries where financial resources are often lacking, it would be important to focus on monitoring toxin-producing cyanobacteria in raw water while searching for low-cost treatment methods. For this purpose, it would be interesting to set up a water quality monitoring policy based on the analysis of physico-chemical parameters (especially nitrates and orthophosphates) and biological parameters. In addition, it would also be important to raise the awareness of residents around the reservoir about good hygiene practices and the control of the use of chemical and organic inputs in agriculture in order to reduce the quantity of water pollutants that favour the development of cyanobacteria.

Conclusion
The diversity and density of Cyanophyceae and Desmidiaceae in the water of the Loumbila reservoir were found to be important and characterize the water quality of this drinking water source under anthropogenic pressure. The results of this study showed that the different activities around the reservoir have impact on the physico-chemical quality of the water and consequently on the diversity of cyanobacteria and Desmids. In addition, most cyanobacteria species are found to be able to produce toxins, especially microcystins. Microcystis aeruginosa species, a microcystin producer was highly present in the reservoir compared to all other species. These results draw the attention of drinking water treatment services to consider the elimination of toxins by using chlorine combined with activated carbon adsorption in the treatment process before supplying water for drinking water supply to the populations. Furthermore, for the reduction of these cyanotoxins in the water, there is an urgent need to take measures to protect this reservoir by controlling activities that increase nutrients contents and cyanobacteria growth in the water.