Toxicity and genotoxicity of water and sediment from streams on dotted duckweed ( Landoltia punctata )

Most rivers are used as a source to supply entire cities; the quality of water is directly related to the quality of tributaries. Unfortunately men have neglected the importance of streams, which receive domestic and industrial effluents and transport nutrients and pesticides from rural areas. Given the complexity of the mixtures discharged into these water bodies, this study aimed to evaluate the quality of water and sediment of ten tributaries of Pirapó River, in Maringá, Paraná State, Brazil. To this end, the free-floating macrophyte Landoltia punctata (G. Meyer) Les & D.J.Crawford was used as test organism in microcosm, and the toxicity of water and sediment samples was evaluated by the relative growth rate, dry/fresh biomass ratio, and genotoxic effects (comet assay). Samples of water and sediment of each stream were arranged in microcosms with L. punctata. Seven days later, plants were collected for analysis. Nutrient levels were higher than the reference location, indicating eutrophication, but the results indicated a toxic effect for only three streams, and a genotoxic effect for all streams.


Introduction
Pollutants can affect organisms in different ways.Many of them are toxic, and some of them may cause loss of genomic stability by altering the balance in chromatin metabolism.They are classified as genotoxic (Barlow et al., 2006;Marple et al., 2004).Several substances can react directly with DNA, causing lesions (Butterworth, 2006), or even preventing the DNA repair (Marple et al., 2004), subsequently causing mutations, chromosomal aberrations and/or changes in chromosome number (Butterworth, 2006).Some genotoxic agents have carcinogenic potential (Butterworth, 2006).
Because of the large number of pollutants, environments are usually affected by mixtures, which can be evaluated by ecotoxicological analysis, in order to recognize the harmfulness to biota of chemical substances, isolated or in form of mixtures (Knie and Lopes, 2004).Analysis are carried out with biological models that have the advantage of responding to levels below the detectable limit by methods of chemical analysis and possible synergism between different agents (Knie and Lopes, 2004).
Currently, for genotoxic evaluation, the comet assay is well defined.It has shown to be a valuable tool for being simple, versatile, of rapid execution, visual and highly sensitive to detect DNA lesions (Dhawan et al., 2009).This assay has been largely employed in genotoxic tests and can be applied to different models, from bacteria to man, to different tissues and cells, aiming to evaluate the genotoxic potential of substances and/ or environmental conditions: terrestrial, aquatic and aerial (Dhawan et al., 2009;Díaz et al., 2009;Prá et al., 2005;Kammann et al., 2001Kammann et al., , 2004;;Sriussadaporn et al., 2003;Poli et al., 1999).
In this context, this study aimed to evaluate the environmental quality of ten streams.They are important for being tributaries of Pirapó River, that supplies the city of Maringá.In this evaluation, in addition to physical and chemical factors such as pH, temperature, conductivity, dissolved oxygen, total nitrogen and phosphorus, responses were analyzed from the free-floating aquatic macrophyte Landoltia punctata (G.Meyer) Les & D.J.Crawford, considering its relative growth rate, dry/fresh biomass ratio as indicative of eutrophication or toxic effect.Potential genotoxic of water and sediment samples were also measured by the comet assay technique.It is expected that the present study contribute as a source of data to warn authorities and for proposals of improvement measures to these water bodies.

Study area
Pirapó River basin is in the third plateau of the Paraná State, with a catchment area of about 5.076 Km 2 , located between the latitudes 22°30' and 23°30' S and longitudes 51°15' and 52°15' W. It encompasses 28 towns; at least 14 of them are supplied by this river, including Maringá (Peruço, 2004), which extends over the watershed ridge between Pirapó and Ivaí rivers basins, and because of this, numerous streams arise out within the urban perimeter of the city, being under the urban influence on their physical, chemical and biological characteristics (Cunico, 2010).For this study, were sampled 10 tributary streams of Pirapó River, situated on the same soil type, within the city of Maringá (Figure 1).These streams were classified into rural and urban, according to Kühl (2009) (Table 1).

Sampling
Samples of water and sediment were collected in November 2010, at nearby sites of 10 streams.In order to obtain results about the effect of pollutants in relation to a non-impacted area, we used a reference location instead of artificial means.The headwaters of Tagaçaba Stream were chosen as a reference location for being located inside a forest and for not receiving effluents or runoff from agricultural areas.Samplings were performed within three days without rainfall (Table 1).
Samples of water and sediment were collected in November 2010, at intermediate sites of 10 streams.The headwaters of Tagaçaba Stream was chosen as a reference location for being located inside a forest and for not receiving effluents or runoff from agricultural areas.Samplings was performed within three days without rainfall (Table 1).
The method employed to collect samples was based on Knie and Lopes (2004).Water was taken from under the water surface using two plastic vials of 3 liters.Sediment was collected and placed in glass vials with a capacity of 1 kg.These samples were subjected to tests within five hours after collection, in order to minimize possible effects caused by physic-chemical and biochemical changes.

Physical and chemical parameters
Simultaneously to water and sediment sampling, conductivity, pH, dissolved oxygen (DO), dissolved oxygen saturation and temperature (Table 2) were measured.One of the water samples of each stream was taken to the Laboratory of Agrochemistry and Environment from the State University of Maringá for the quantification of total nitrogen (mg/l) and total phosphorus (mg/l).

Test organism
The free-floating aquatic macrophyte L. punctata was selected for tests for being a native species, easily found in the study area, it is a non-endemic species and has wide distribution in the Paraná State and Atlantic Forest (Coelho, 2010).Like other duckweeds, it is tiny, allowing its growth in microcosms.Besides, the predominance of vegetative reproduction enables to obtain a genetically homogeneous population of plants, which is ideal for genotoxic analysis.It also has a very rapid growth, providing significant results in a short time when subjected to phytotoxicity tests, and allowing efficient estimates when subjected to comet assay (Babić et al., 2009;Radić et al., 2010).
Plants used in the present study came from an initial sample collected in nature, in Maringá, six months before the experiments and grown in plastic tanks for propagation.

Microcosms
Plastic trays (20x10x4cm) comprised the microcosms.Ten trays were used for each stream, with five repetitions for water and five for sediment.In each tray, 400 ml water of each stream were added.Sediment was distributed in a way to reach the height of 1 cm, and distilled water was added to achieve a water depth of 1 cm.All microcosms received one gram of macrophyte previously acclimated to the test conditions.Microcosms were randomly arranged in a greenhouse (Oikos, UEM) with a photoperiod of 14 hours, protected from direct sunlight with a 2mm mesh screen to control the temperature and loss of water by evaporation.The temperature varied between 30 and 25 °C.

Mass assay
Initial fresh biomass (1 g) was added to each microcosm, after slightly drying with paper towel.In the same way, after 7 days macrophytes were removed from microcosms and weighed again, in order to obtain the final fresh biomass; they were then wrapped in paper envelopes and oven-dried until achieving constant weight, the final dry biomass (DB f ).The initial dry biomass (DB i ) was estimated by the proportion of the initial fresh biomass and dry biomass obtained from a portion of macrophyte of the start of testing.We then calculated the relative growth rate (RGR) = (lnDB f -lnDB i )/t, where t is the period of 7 days.Ratio  between final dry biomass and final fresh biomass (DB f / FB f ) was also calculated, which can also be an indicative of plant stress (Perfus-Barbeoch et al., 2002).

Genotoxicity assay
Ten fronds were randomly collected from each microcosm, rinsed with distilled water and dried with paper towel to be used in comet assay, according to procedure of Navarrete et al. (1997) with modifications.Plants were not cut, but crushed with a metallic aste against the wall of the Eppendorf tubes with 0.5 mL PBS, of which 120 µl was taken to another Eppendorf tube with 120µl low melting point agarose at 1%.After homogenization, the volume was distributed onto two slides previously coated with normal melting point agarose at 1.5% and then a coverslip was placed over each slide to spread the suspension of nuclei and taken to refrigerator for five minutes to gel low melting point agarose; after this time interval, coverslips were removed.Afterwards, the processes of electrophoresis, neutralization and fixation were carried out, based on Navarrete et al. (1997).
At the time of analysis with a fluorescence microscope under 200X magnification (Zeiss Axioskop2 Plus), slides were stained with 2% ethidium bromide.Random fields were examined on the slide until totaling 50 nuclei or more.Nuclei were classified into five levels (0-4) from which was calculated the damage index, which is the sum of nuclei multiplied by the level that they occupy, divided by total nuclei multiplied by 100, so that this score ranges from 0 to 400, if all nuclei have had damage 0 or 4, respectively.The damage frequency was also calculated, it corresponds to the percentage of damaged nuclei (1-4) divided by the total nuclei (0-4) multiplied by 100 (Collins et al., 1997).

Statistical analysis
In order to evaluate possible differences between growth rate of L. punctata treated with only water or sediment of each stream, a two-way analysis of variance was run, and a post-hoc Dunnett's test was applied to compare with the reference location.Data obtained in the damage index of comet assay and dry/fresh biomass ratio were transformed as proposed by Dias and Barros (2009) to meet normality and homoscedasticity.These analyses were performed using the software Statistica 7.1 (StatSoft).
To check correlations between abiotic data and growth rate/comet assay with L. punctata, a Principal Component Analysis (PCA) was employed.With the Broken-Stick criterion, axes 1 and 2 were retained for interpretation.This analysis was run on the software PC-ORD 5.0 (MjM Software).

Abiotic variables
The lowest value of conductivity was observed for the reference location (Tagaçaba).The highest values were found for streams Nazaré, Guaiapó, Queçaba and Mandacaru.In relation to saturation of dissolved oxygen, Romeira, Nazaré, Guaiapó, Água da Roseira and Água do Pirapó have stood out by lower values, from 70 to 84%.Other streams (Tagaçaba, Água Queçaba, Remo, Zaúna, Mandacaru and Miosótis) had values around 90%. Higher concentrations of total nitrogen were registered in streams Miosótis, Água do Pirapó and Água da Roseira.Values of total phosphorus were lower for streams Mandacaru (around 40% from reference location value), Romeira, Zaúna and Miosótis (around 60% from reference location value), Água do Pirapó and Guaiapó (around 75% from reference location value).Higher values of this parameter were verified in the streams Nazaré and Água da Roseira with 139% and 264% compared with the reference location (Table 2).

Relative growth rate (RGR) and dry/fresh biomass ratio
According to ANOVA results, mean values of RGR were significantly different between water and sediment, with higher values in this latter (F 10;88 =14.146; p<0.001).A significant difference was found between RGR values of  2).Mean values of the dry/fresh biomass ratio were significantly different between water and sediment, with higher values for water (F 10;88 =30.57; p<0.001).Also, mean values were different between streams (F 10;88 =20.82; p<0.001), and according to Dunnett's test (d 10;88 ), water from streams Água da Roseira and Água do Pirapó and sediment from streams Água da Roseira, Miosótis and Água do Pirapó, have been significantly higher than the reference location (Figure 3).

Comet assay
In genotoxic evaluation, the damage index of comet assay tested by a two-way ANOVA evidenced significant interaction between variables stream and water-sediment (F 10;198 =2.98; p<0.01).Mean values were not significantly  different between water and sediment (F 10;198 =1.28; p>0.05), but were statistically different between streams (F 10;198 =9.14; p<0.001).According to Dunnett's test (d 10;198 ), only sediment from streams Zaúna and Água do Pirapó were not different from the reference location, the others were significantly higher than the reference location (Figure 4).
The damage frequency to water and sediment diverged among themselves in relation to streams (significant interaction F 10;198 =2.569; p=0.0061).Mean values were not significantly different between water and sediment (F 10;198 =0.080; p=0.778), however they were significantly different among streams (F 10;198 =11.915; p<0.001).In order to identify streams different from the reference location, the Dunnett's test(d 10;198 ) was applied.It showed that the water of streams Água Queçaba, Mandacaru, Miosótis and Nazaré, and sediment from streams Remo, Zaúna, Miosótis and Água do Pirapó were not statistically different from the reference location; the others were significantly greater in relation to the reference location (Figure 5).

Principal component analysis
Subsequently, to identify possible correlations between variables and to identify the most important characteristic for the differentiation among streams, a principal component  analysis (PCA) was performed with abiotic and biotic variables, followed by ANOVA on axes retained for analysis.Firstly, the ordination presented correlations somewhat consistent with the literature, data from the Água da Roseira Stream have caused this divergence since they presented the highest concentration of phosphorus, but the plant growth was low in this sample, probably a toxic effect.In this way, these data were removed and PCA was run again (Figure 6).For this new analysis, axes 1 and 2 were retained for interpretation and explained 33.4% and 21.3% of variance, respectively.This time the total phosphorus presented the expected correlation with the growth of macrophyte as well as temperature and dissolved oxygen.The three variables pH, total nitrogen and biomass index were the most negatively influencing the ordination for the PCA1.The RGR and total phosphorus have positively influenced the ordination of the PCA1.The concentration and saturation of dissolved oxygen were the variables that most influenced negatively the ordination of PCA2, and temperature, positively.ANOVA evidenced significant difference between streams, for both PCA 1 (F 9;40 =242.5401;p<0.001) and PCA 2 (F 9;40 =677.8497;p<0.001).After this, the Dunnett's test (d 9;40 ) at a significance level of 0.05 has indicated that, for PCA 1, the streams Água do Pirapó, Mandacaru, Miosótis and Zaúna have negatively differed from the reference location, with higher pH, total nitrogen and dry/fresh biomass ratio.Nazaré, Remo and Romeira streams have differed positively from the reference location, with higher RGR and total phosphorus.For the PCA 2, Água do Pirapó, Guaiapó, Miosótis, Romeira and Nazaré streams have differed positively from the reference location, with higher temperature values.Queçaba, Remo and Zaúna streams have differed negatively from the reference location, with higher concentration and saturation of oxygen.

Discussion
Although high values of conductivity may be an indirect indicative of pollution (Graça and Coimbra, 1998), as expected, the reference location showed low conductivity, but this parameter did not influence linearly the biological parameters tested on L. punctata.In agreement with Scheffer and Busch (2010), dissolved oxygen values below the saturation are especially related to increased organic matter in the water body; in this way it can indicate sewage discharge.In this aspect, the Romeira Stream presented the lowest saturation, despite being a rural stream, possibly with little or without domestic sewage, and this result was probably due to the discharge of pig farm waste.Nazaré Stream presented the second lowest saturation probably owed to the effect of domestic sewage, because, according to Cunico (2010) this stream has 100% urbanization in its microbasin.The other streams presented oxygen saturation closer to the reference location and despite the stream Mandacaru has characteristics of waste discharge similar to the Nazaré Stream (Kühl, 2009) with 82.5% urbanization in its microbasin (Cunico, 2010), it presented an oxygen saturation higher than the Nazaré Stream, probably because of a rapid stretch near the sampling site.Regarding the concentration of dissolved oxygen, all streams are categorized into the class 1, according to resolution 357/2005 of CONAMA (>6mg/l) (Brasil, 2006).According to Esteves (1998), low concentrations of nitrogen compounds can work as a limiting factor for aquatic macrophytes, but in this study, total nitrogen has negatively influenced the RGR of macrophytes, and positively the biomass index, indicating a toxic effect for L. punctata, which was possibly associated with the effect of nitrogen in form of ammonia.The relative growth rate of duckweeds linearly decreases with increasing ammonia, leading to death in concentrations above 8mg/l (Körner et al., 2003).Onaindia et al. (1996) reported that ammonia nitrogen usually derives from discharges and is commonly associated with absence of macrophytes.Moreover, ammonia nitrogen is the reduced form of nitrogen and indicates recent pollution of the water body (Von Sperling, 1996).
The pH also had a negative correlation with RGR and positive with DB f /FB f , but the pH should not have had a direct influence on these parameters as all streams were within the range of 5-9 indicated as suitable for the growth of duckweeds by Radić et al. (2010).The positive correlation of pH with total nitrogen suggests once again the predominance of the ammonia form, and in agreement with Dickson (1981) ammonia tends to make the pH alkaline.On the other hand, alkalinity shifts the ammonium/ammonia balance, increasing the concentration of ammonia that is toxic to duckweeds even at low concentrations, justifying its negative but indirect effect of pH on the macrophyte.
The results of RGR of L. punctata were positively influenced by phosphorus, corroborating the study of Kobayashi et al., (2008) who evidenced the phosphorus as a limiting factor for the growth of a floating macrophyte species.Although the stream Água da Roseira had presented the highest concentration of total phosphorus, this stream along with Mandacaru, Miosótis and Água do Pirapó streams, have presented mean RGR significantly lower than the reference location, for both water and sediment; these streams had higher concentration of total nitrogen and higher pH.Mean values of RGR in sediment were higher than in water, and besides that there was no interaction between variables stream and water-sediment, pointing out that the sediment acts as a pool of nutrients that can be transferred to the water column similarly among streams.According to the resolution 357/2005 of CONAMA (Brasil, 2006) relative to total phosphorus, streams Remo and Nazaré are classified into class 3 and Água da Roseira Stream into class 4, and the others into class 1.
Considering the biomass index, the sediment from Miosótis Stream, water and sediment from Água da Roseira and Água do Pirapó streams presented mean values significantly higher than the reference location.According to Perfus- Barbeoch et al. (2002) this negative effect on the plant is the result of osmotic stress that reduces the growth and increases the dry/fresh biomass ratio.The other streams have not differed from the reference location, and in general, assays with water presented indices significantly higher than assays with sediment, perhaps due to the availability of nutrients by sediment, which occurs similarly among streams (without significant interaction).
Parameters RGR and DB/FB in this study were similar, but the RGR was more sensitive, indicating one more stream (Mandacaru) significantly different from the reference location.
As for the genotoxicity, almost all streams presented damage index of comet assay significantly higher than the reference location, except the sediment from Zaúna and Água do Pirapó streams.Results presented interaction between streams and water-sediment, indicating that streams presented possible genotoxic agents different in relation to the dynamics between water column and sediment.Although not significant, apparently, tests with water from rural streams tended to be more genotoxic than with sediment, and for urban streams, sediment tended to be more genotoxic than water.According to Esteves (1998), various compounds may concentrate in the sediment, this way, the sediment is important for assessing the ecosystem, also giving responses to previous contamination.
In relation to the damage frequency, water from four out of the ten streams studied (Água Queçaba, Mandacaru, Miosótis and Nazaré), and sediment of four out of the ten streams (Remo, Zaúna, Miosótis and Água do Pirapó) have not differed from the reference location.In the present study, the damage index showed to be more sensitive than damage frequency.This is because the damage frequency only considers the number of nuclei with damage, and in some cases, some cells can also have more severe damage, fact that is taken into account in the damage index.
The results obtained with the comet assay are very important because, according to Xu et al. (1998) and Wang and Freemark (1995), it has been demonstrated premutagenic and mutagenic activities in higher plants, and in algae they occur through mechanisms similar to those found in mammals.Zagatto and and Bertoletti (2006) stated that depending on the amount and location of mutations, they can cause diseases in individual and offspring, and can increase the incidence of cancer, hereditary, developmental, neurobehavioral and heart diseases, virulence of pathogens, in addition to the negative effect on biodiversity.
In this work the comet assay proved to be a fast, economic and reliable too; it and can be well employed in aquatic ecotoxicological studies, and applied to other organisms such as algae (Aoyama et al., 2003), flatworms (Guecheva et al., 2001) and fish (Grisolia et al., 2009).
These results indicate a risk for aquatic biota and for man, since these streams flow into the Pirapó River, which in turn, is responsible for supplying Maringa and several others cities that make up its basin (Peruço, 2004).However, further studies are necessary to verify possible contaminants and their origins specifically in these locations.

Figure 1 .
Figure 1.Location of the ten studied streams.

Table 1 .
Streams, geographic coordinates and altitude of sampling sites (C= control, R= rural, U= urban).