Symbiotic Co-Culture of Scenedesmus sp. and Azospirillum brasilense on N-Deficient Media with Biomass Production for Biofuels

The treatment of nitrogen-deficient agriculture wastewater, arising from the vegetable and fruit processing, is a significant problem that limits the efficiency of its biological treatment. This study evaluates the effectiveness of the symbiotic co-culture of Azospirillum brasilense and Scenedesmus sp., under two nitrogen levels (8.23 mg L−1 and 41.17 mg L−1) and mixing systems (aeration and magnetic stirring), aiming to simultaneously use the N-deficient media for their growth while producing biomass for biofuels. Microalgae growth and biomass composition, in terms of protein, carbohydrate and fatty acid contents, were evaluated at the end of the exponential growth phase (15 days after inoculation). Results show that the symbiotic co-culture of microalgae-bacteria can be effectively performed on nitrogen-deficient media and has the potential to enhance microalgae colony size and the fatty acid content of biomass for biofuels. The highest biomass concentration (103 ± 2 mg·L−1) was obtained under aeration, with low nitrogen concentration, in the presence of A. brasilense. In particular, aeration contributed to, on average, a higher fatty acid content (48 ± 7% dry weight (DW)) and higher colony size (164 ± 21 μm2) than mechanical stirring (with 39 ± 2% DW and 134 ± 21 μm2, respectively) because aeration contribute to better mass transfer of gases in the culture. Also, co-culturing contributed in average, to higher colony size (155 ± 21 μm2) than without A. brasilense (143 ± 21 μm2). Moreover, using nitrogen deficient wastewater as the culture media can contribute to decrease nitrogen and energy inputs. Additionally, A. brasilense is approved and already extensively used in agriculture and wastewater treatment, without known environmental or health issues, simplifying the biomass processing for the desired application.


Introduction
Microalgae are photosynthetic microorganisms that can grow relatively fast, even under harsh environmental conditions [1], and have the capability to produce biomass rich in lipids, production [31,32] and food applications [33]. There are some previous publications about the co-culturing of Scenedesmus with different bacteria in wastewater, showing that it increases biomass production due to a higher CO 2 concentration in the media [34], or its suitability to degrade complex pollutants such as crude oil [35]. However, to the best of the authors' knowledge, there is no previous work about the effect of A. brasilense co-cultured with Scenedesmus sp. in suspended cultures, even though Scenedesmus sp. is one of the most promising alternatives for large-scale cultivation for biomass production, wastewater treatment and other applications. As this could be the first experimental study involving the co-culture of A. brasilense and Scenedesmus sp., in this work an exploratory experimental design was performed, in which two mixing systems (aeration and magnetic stirring), two nitrogen levels (8.23 mg L −1 and 41.17 mg L −1 ) and the presence or absence of bacteria were evaluated in order to analyze their effects on the co-culture. The purpose of selecting two different mixing systems for the experiment, aerated versus stirred systems, was to provide two different oxygen levels to be tested in this study: a higher oxygen level provided by the aerated system and a lower oxygen level provided by the magnetic stirring. This is because it is known that the A. brasilense nitrogenase activity can be inhibited by oxygen, affecting its performance as MGPB [36]. Also, two different nitrogen concentrations were tested in the experiment, as it is known that nitrogen concentration has a key role in the establishment of a symbiosis between bacteria and microalgae, and strongly influences and nitrogen fixation of A. brasilense [37]. Biomass composition in terms of proteins, carbohydrates and fatty acids was also evaluated to identify a possible metabolic response of Scenedesmus sp.

Culture Conditions and Inoculum Preparation
Scenedesmus sp. (UTEX 1589, University of Texas, Austin, TX, USA) and A. brasilense (ATCC 29145, Manassas, VA, USA) were used in all the experiments. A modified BBM was used for microalgae cultures, i.e. to the original UTEX BBM it was added vitamin B12 solution, biotin, and thiamine [38]. The pH was initially adjusted to 6.5-6.6. BBM macronutrients are (in g·L −1 ): NaNO 3  A. brasilense cells were grown in 500 mL Erlenmeyer flasks containing 450 mL of AZ medium. Each flask was kept in thermostatic water bath at 36 ± 1 • C under continuous agitation by magnetic stirring (around 150 rpm) at room conditions. Scenedesmus sp. cells were grown in 1 L Erlenmeyer flasks containing 800 mL of BBM at room temperature, light/dark cycles of 12/12 h and light intensity of 15 µmol photons·m −2 ·s −1 . Each flask was mixed using an 8 cm magnetic stirrer at about 60 rpm, and continuously aerated using a 5 mm glass tube diffuser placed at the bottom of the flask at a flow rate of about 500 mL·min −1 .
For the inoculum preparation, Scenedesmus sp. was harvested by centrifugation (Centurion 4000r, Chichester, West Sussex, UK) at 2500 rpm for 10 min, and washed three times with saline solution (0.8% NaCl). Microalgae inoculum was adjusted to an optical density at 750 nm (OD 750 ) of 0.8 (approximately 8 × 10 6 cells) with saline solution. A. brasilense was harvested by centrifugation at 5000 rpm for 15 min, and washed three times with a saline solution. A. brasilense inoculum was adjusted to an optical density at 540 nm (OD 540 ) of 1.0 (approximately 1 × 10 9 cells) with a saline solution.
A volume of 8 mL of Scenedesmus inoculum along with 2 mL of A. brasilense inoculum were added to 450 mL of sterile medium in each flask. Final bacteria/microalgae ratio was maintained as 30 cells/cells.

Experimental Design and Statistical Analysis
In this study, three factors were evaluated: A. brasilense presence, mixing system, and initial nitrogen concentration in the culture media, with two levels each, corresponding to a 2 3 factorial experimental design.
A total of 8 experimental treatments were performed to assess all the possible combinations of factors levels as shown in Table 1. Four replicates per treatment (n = 4) were performed. In the A. brasilense column, (+) means presence and (-) absence of bacteria. In the mixing system column, (+) means aeration and (-) means magnetic stirring. In the nitrogen column, (+) means high nitrogen concentration and (-) means low nitrogen concentration.
Results were evaluated using one-way analysis of variance (ANOVA) using Minitab 17 and Minitab Express software, in order to determine which factors are statistically significant. Then, to compare the different treatments the Fisher test was applied. An alpha level of 5% was considered significant for all comparisons.

Biomass Growth
Optical density was measured daily after inoculation at a wavelength of 750 nm to minimize the interference by pigments [39]. Counting of microalgae cells was performed with a Neubauer chamber (Auxilab 30100011, Beriain, Navarra, Spain).
For cell counting a 20 µL sample was placed in the Neubauer chamber, allowed to repose for 2 min, and counted under 40× objective in a bright-field microscope equipped with a digital camera (Scopetek Opto-Electric DCM 200, Hangzhou, China). Counting was performed in duplicate. Correlation curves of OD 750 and dry weight (DW) were done according to Griffiths et al. [39].
For the biomass concentration estimation data from the correlation curves of OD 750 was used. To analyze the total biomass production, 50 mL of each replicate were filtered at the end of the exponential growth phase (at about 15 days from the inoculation) using pre-weighed 0.45 µm cellulose nitrate membranes, rinsed with distilled water and dryed at 105 • C for 8 h.
Growth monitoring was stopped when the cultures finished the exponential growth phase. Biomass was harvested by centrifugation (Centurion 4000r, Chichester, West Sussex, UK) at 2500 rpm for 10 min and freeze-dried (VirTis BT6K, Warminster, PA, USA) for biomass analysis.

Cell Size
Scenedesmus sp. is typically found in colonies of four cells inside a parental mother wall named coenobium. The coenobium area was determined at the end of the runs by image analysis using the AutoCAD2015 software to measure cell length and width.
To calibrate measurements, Neubauer chamber small squares (50 µm) were used as scale.
For statistical analysis, 25 cells of each treatment replicate were considered.

Biomass Analysis
For biomass analysis, biomass was harvested by centrifugation (Centurion 4000r, Chichester, West Sussex, UK) at 2500 rpm for 10 min and freeze-dried (VirTis BT6K, Warminster, PA, USA). The biomass analysis was performed for the whole co-culture of microalgae and bacteria.
Carbohydrates were determined from the freeze-dried biomass (5 mg), according to Dubois et al. [40], using glucose as standard.
Protein content was determined on the freeze-dried biomass (5 mg), according to Lowry et al. [41], using bovine serum albumin as standard. Determinations were performed in duplicate.
Fatty acids were determined on the freeze-dried biomass (10 mg) in a simultaneous extraction/methylation process following the methodology of Ryckebosch et al. [42]. 2 mL of chloroform/methanol (2:1 v/v) were used as extraction reagent. Later, 1 mL of Arachidic acid C20:0 (25 mg/100 mL chloroform) was added to the samples as an internal standard, followed by the addition of 1 mL of methanol. Sulfuric acid (300 µL) was used as catalyst for the esterification reaction. The mixture was placed under a water bath at 100 • C for 10 min, and then cooled to room temperature. Later, 1 mL of water was added, and the lower layer (organic phase) was extracted, filtered (Whatman 1) over anhydrous sodium sulfate, and transferred to a new glass tube. The solvent was evaporated using N 2 , and the extract was re-dissolved with 1.5 mL hexane for gas chromatography analysis. The fatty acids (FA) profile was analyzed by gas chromatography (GC) with cold on-column injection and a flame ionization detector (FID) (Trace GC Ultra, Thermo Scientific, Belgium). An EC Wax column of 30 m length, 0.32 mm internal diameter, and 0.25 µm film (Grace, Belgium) was used with the following time-temperature program: 70-180 • C (5 • C·min −1 ), 180-235 • C (2 • C·min −1 ), and 235 • C (9.5 min). Peak areas were quantified with ChromCard for Windows software (Interscience, Belgium). Fatty acids standards (Nu-check, USA) containing a total of 35 different fatty acids were analyzed for peak identification. Fatty acids determinations were performed in duplicates.

Biomass and Growth Analysis
To study the effect of A. brasilense in algal biomass productivity, 8 experimental treatments were performed, as shown in Table 1, corresponding the set of treatments to a 2 3 experimental design. For all the 8 treatments, biomass growth was monitored from the inoculum to day 15 of the exponential growth phase, as shown in Figure 1. Comparing the various growth curves, it can be seen that although qualitatively similar, the experimental treatments containing Azospirillum (T1 to T4) reached on day 15, a higher average biomass concentration of 96 ± 5 mg·L −1 , than treatments without bacteria (T5 to T8) of 91 ± 9 mg·L −1 . Also, differences on biomass productivity are mainly observed after day 10 of the exponential growth phase.
Among treatments with Azospirillum, treatment T2, under aeration and with low nitrogen concentration, reached the highest average biomass concentration of 103 ± 2 mg·L −1 on day 15 of exponential growth, and had the highest average biomass productivity of 6 ± 4 mg·L −1 ·day −1 , considering all the 15 days, and of 11 ± 7 mg·L −1 ·day −1 , just considering the last 5 days of exponential growth, after day 10. Concerning the cultures without Azospirillum, treatment T6, also under aeration and with low nitrogen concentration, reached the highest average biomass concentration of 98 ± 11 mg·L −1 on day 15 of exponential growth, and had the highest average biomass productivity of 6 ± 5 mg·L −1 ·day −1 , considering all the 15 days, and of 11 ± 8 mg·L −1 ·day −1 , just considering the last 5 days of exponential growth. Therefore, aeration and low nitrogen concentration are two relevant factors influencing microalgae growth, with and without Azospirillum. Biomass concentration expressed as dry weight is shown in Figure 2A. Results show that the presence of A. brasilense is statistically significant (p-value: 0.034). On average, the experimental treatments containing Azospirillum had higher biomass concentration (96 ± 5 mg·L -1 ) than treatments without bacteria (91 ± 9 mg·L -1 ). The lowest biomass concentration (83 ± 7 mg·L -1 ) was achieved in treatment 8, an experiment without A. brasilense and under magnetic stirring ( Figure 2A). Normally, higher biomass production is related to higher cell concentration in microalgae cultures [43]. When working with MGPB, microalgae cells are in contact with substances that can enhance their growth, cell size and biomass composition [44]. A. brasilense co-culturing with cyanobacteria species has also been reported to increase biomass growth. Synechococcus elongatus immobilized in alginate beads with A. brasilense increased its specific growth rate and biomass production during wastewater treatment [27]. However, immobilized cultures represent higher costs. Biomass concentration expressed as dry weight is shown in Figure 2A. Results show that the presence of A. brasilense is statistically significant (p-value: 0.034). On average, the experimental treatments containing Azospirillum had higher biomass concentration (96 ± 5 mg·L −1 ) than treatments without bacteria (91 ± 9 mg·L −1 ). The lowest biomass concentration (83 ± 7 mg·L −1 ) was achieved in treatment 8, an experiment without A. brasilense and under magnetic stirring (Figure 2A). Normally, higher biomass production is related to higher cell concentration in microalgae cultures [43]. When working with MGPB, microalgae cells are in contact with substances that can enhance their growth, cell size and biomass composition [44]. A. brasilense co-culturing with cyanobacteria species has also been reported to increase biomass growth. Synechococcus elongatus immobilized in alginate beads with A. brasilense increased its specific growth rate and biomass production during wastewater treatment [27]. However, immobilized cultures represent higher costs.  The nature of the mixing system is also statistically significant (p-value: 0.000), in particular aeration contributed to higher biomass production than the mechanical stirring. Notably, aeration in the presence of bacteria yielded the highest biomass production (103 ± 2 mg·L −1 ) in treatment 2. This is explained because the mixing system directly affects the mass transfer of gases, in particular oxygen and carbon dioxide in the culture, and it is already known that the optimization of these factors can enhance Scenedesmus growth in photobioreactors [45].
Nitrogen is a critical nutrient for microalgae growth because it is used by the cells to grow through the production of proteins and nucleotides present in biomass [46]. Based on cell counting, Scenedesmus sp. did not show any significant difference on its specific growth rate when co-cultured with A. brasilense under different mixing conditions and nitrogen levels, as shown in Figure 2B. According to other studies, high nitrogen concentration increased biomass production of Scenedesmus [46]. A. brasilense, due to its nitrogen fixing capacity, can reduce the effect of low exogenous nitrogen concentration, as observed in the treatments with bacteria and low nitrogen concentration (treatments 2 and 4). In addition, nitrogen is one of the most important nutrients for microalgae growth. Therefore, microalgae biomass production using A. brasilense can reduce the need to add exogenous nitrogen loads in cultures, one of the main drawbacks in microalgae culture cost-effectiveness [47,48].

Growth in terms of Colony Size
As shown in Figures 2C and 3, coenobium size of Scenedesmus sp. was affected by the presence of A. brasilense (p-value: 0.044). Co-culturing contributed in average, to higher colony size (155 ± 21 µm 2 ) than without A. brasilense (143 ± 21 µm 2 ). Also, colony size was affected by the mixing system (p-value: 0.000). Aeration contributed to, in average, higher colony size (164 ± 21 µm 2 ) than mechanical stirring (134 ± 21 µm 2 ), because aeration contribute to better mass transfer of gases in the culture. The same number of cells (four) per colony, but bigger colony sizes, were found in treatment 1 (172 ± 32 µm 2 ) with A. brasilense under continuous aeration and with the highest nitrogen concentration, while the lowest coenobium area was found in treatment 7 (130 ± 25 µm 2 ), without A. brasilense, under magnetic stirring and with the highest nitrogen concentration.
Similar results were found with Chlorella sp. co-culturing with A. brasilense in immobilized cultures [23]. In the mentioned study, microalgae cells increased their size, but contrary to our results, A. brasilense enhanced Chlorella sp. growth rate in terms of cell number per volume of culture. Although considering a different co-culturing mode, in particular Scenedesmus sp. with A. brasilense in suspended cultures, the results obtained from cell size analysis are similar to the results obtained with other species such as Chlorella sorokiniana [24], Synechococcus elongatus [27] and Chlorella vulgaris [22] co-cultured with A. brasilense, but immobilized.
In order to identify the possible mechanisms involved in increasing cells size, previous reports suggest that phytohormones (such as IAA) produced by MGPB can induce significant changes in the microalgae metabolism, such as enhancing cell size [23], and are responsible for the ammonium uptake and assimilation in microalgae [49]. The exchange of IAA is independent of nitrogen fixation, and for the present study, it can explain the differences in cell size among treatments, even though we cannot see a significant difference in growth rate." IAA, an auxin responsible of cell elongation and cell division, can be one of the substances implicated in the increment of cellular biomass production, therefore producing bigger cells [22]. Previous experiments reported that the addition of this phytohormone into Chlorella vulgaris media without A. brasilense emulates the effects of the bacteria in increasing microalgae biomass production in terms of dry weight, total number of cells, and cells size [22]. IAA produced by A. brasilense might be involved in ammonium uptake and its assimilation, leading to higher biomass production [49]. Since ammonium requires less energy from microalgae cells for its assimilation than other nitrogen sources [50], energy saved from nitrogen uptake process can be used to support other metabolic reactions, and can explain the higher colony size and biomass production.

Analysis of Biomass Composition
Growth under nitrogen deprivation is believed to induce decreases in protein productivity and to increase the productivity of lipids and carbohydrates of microalgae. However, the intensity of these processes greatly varies depending on the species tested and the experimental conditions [30,51,52]. The graph in Figure 4, presents the biomass analysis results in terms of protein, carbohydrate and fatty acids contents. Results of the Fisher test are shown by letters (a, b, c, d) placed at the top of each bar. Dry weight content of protein, carbohydrate or fatty acids was separately analyzed with the Fisher test. Therefore, results shown compare averages among treatments for a given biomass component (protein, carbohydrate or fatty acids). Treatments not sharing a common letter mean that the corresponding average values significantly differ from each other (p < 0.05 Fisher).

Analysis of Biomass Composition
Growth under nitrogen deprivation is believed to induce decreases in protein productivity and to increase the productivity of lipids and carbohydrates of microalgae. However, the intensity of these processes greatly varies depending on the species tested and the experimental conditions [30,51,52]. The graph in Figure 4, presents the biomass analysis results in terms of protein, carbohydrate and fatty acids contents. Results of the Fisher test are shown by letters (a, b, c, d) placed at the top of each bar. Dry weight content of protein, carbohydrate or fatty acids was separately analyzed with the Fisher test. Therefore, results shown compare averages among treatments for a given biomass component (protein, carbohydrate or fatty acids). Treatments not sharing a common letter mean that the corresponding average values significantly differ from each other (p < 0.05 Fisher).

Protein Content
In this study, protein content was affected by the presence of A. brasilense (p-value: 0.019). Coculturing treatments had an average protein content of 21 ± 2% of dry cell weight (DW), while treatments without MGPB had an average of 23 ± 1% DW. The lowest protein content (19 ± 4% DW) was found in treatment 1, with A. brasilense, under aeration and sufficient nitrogen concentration, while the highest protein content (24 ± 1% DW) was found in treatment 8, without A. brasilense, under magnetic stirring and with low nitrogen concentration in the culture media. These protein content values agree with the ones reported in literature for Scenedesmus sp. of 24.87-35.17% DW [30]. Similarly, treatments with low nitrogen concentration had in average higher protein content (23 ± 1% DW) than those with sufficient nitrogen (21 ± 2% DW). Although a sufficient nitrogen medium may lead one to infer that the protein content will also be high, this and other literature studies [30,53] have shown the opposite, reporting for Scenedesmus a higher protein content when the nitrogen concentration in the media was lower.
When co-culturing with A. brasilense, higher nitrogen fixing activity and therefore higher protein levels are expected due to the absence of nitrogen deprivation. However, the mechanisms involved in nitrate uptake and protein synthesis are more complex and depend on several factors, greatly varying depending on the species tested and experimental conditions. It is already known that A. brasilense is able to establish cell to cell interactions with microalgae to sustain mutualistic symbiotic relations [17,54]. Also, significant increments in enzymatic activities were observed in literature studies [55] during the co-culturing of A. brasilense and Chlorella sorokiniana. This shows that the coculturing of these two species involve other mechanisms different than the abiotic stresses, such as nitrogen deprivation. Therefore, its effect could modify metabolic pathways, thus changing the biomass composition in both microorganisms involved.
The two main factors affecting protein content in co-culturing are the production of growth promoters and the biological nitrogen fixation from A. brasilense [26]. Based on growth promoters

Protein Content
In this study, protein content was affected by the presence of A. brasilense (p-value: 0.019). Co-culturing treatments had an average protein content of 21 ± 2% of dry cell weight (DW), while treatments without MGPB had an average of 23 ± 1% DW. The lowest protein content (19 ± 4% DW) was found in treatment 1, with A. brasilense, under aeration and sufficient nitrogen concentration, while the highest protein content (24 ± 1% DW) was found in treatment 8, without A. brasilense, under magnetic stirring and with low nitrogen concentration in the culture media. These protein content values agree with the ones reported in literature for Scenedesmus sp. of 24.87-35.17% DW [30]. Similarly, treatments with low nitrogen concentration had in average higher protein content (23 ± 1% DW) than those with sufficient nitrogen (21 ± 2% DW). Although a sufficient nitrogen medium may lead one to infer that the protein content will also be high, this and other literature studies [30,53] have shown the opposite, reporting for Scenedesmus a higher protein content when the nitrogen concentration in the media was lower.
When co-culturing with A. brasilense, higher nitrogen fixing activity and therefore higher protein levels are expected due to the absence of nitrogen deprivation. However, the mechanisms involved in nitrate uptake and protein synthesis are more complex and depend on several factors, greatly varying depending on the species tested and experimental conditions. It is already known that A. brasilense is able to establish cell to cell interactions with microalgae to sustain mutualistic symbiotic relations [17,54]. Also, significant increments in enzymatic activities were observed in literature studies [55] during the co-culturing of A. brasilense and Chlorella sorokiniana. This shows that the co-culturing of these two species involve other mechanisms different than the abiotic stresses, such as nitrogen deprivation. Therefore, its effect could modify metabolic pathways, thus changing the biomass composition in both microorganisms involved.
The two main factors affecting protein content in co-culturing are the production of growth promoters and the biological nitrogen fixation from A. brasilense [26]. Based on growth promoters produced by A. brasilense it is known that during co-culturing with Chlorella sorokiniana, the microalgae provides tryptophan and thiamine to bacteria in exchange for IAA [55]. The exchange of amino acids to produce growth promoters and the effect of growth promoters itself can affect biomass growth and composition, including protein content, thus explaining the differences between protein content with and without A. brasilense in the culture. Further research is required to better understand changes in the proteome of Scenedesmus sp. in co-culturing with A. brasilense in suspended cultures.

Carbohydrate Content
Carbohydrate content of Scenedesmus species has been reported to differ, depending on culture conditions and strains [56]. Different strains of Scenedesmus obliquus were reported to have carbohydrate content between 6 and 47% of DW depending on light intensities and nutrient deprivation [50]. Scenedesmus sp. and Chlorella sp. carbohydrate content has been reported to depend on the nitrogen concentration in the culture media as higher carbohydrate contents were obtained under lower nitrogen concentration, similar to this study results [51].
In this study, carbohydrate content followed the same pattern already observed for protein content in terms of A. brasilense effect (p-value: 0.000). This can be seen in Figure 4 that shows that co-culturing treatments had an average carbohydrate content of 18 ± 2% of dry cell weight (DW), while treatments without MGPB had an average of 23 ± 1% DW. Similarly, nitrogen contributed significantly to carbohydrates content (p-value: 0.001). Treatments with low nitrogen concentration had in average higher carbohydrate content (21 ± 1% DW) than those with sufficient nitrogen (19 ± 1% DW). This explained because when microalgae cells are nitrogen deprived, metabolism changes to the accumulation of lipids or carbohydrates [51].
The highest carbohydrate content of 26 ± 1% DW (treatment 8, Figure 4) was obtained in culture without bacteria, low nitrogen concentration, and under magnetic stirring, while the lowest carbohydrate content of 16 ± 3% DW (treatment 1, Figure 4), was obtained in co-culture with bacteria, sufficient nitrogen and continuous aeration, supporting the conclusion that A. brasilense allows the microalgae to obtain enough nitrogen and limiting their need to produce carbohydrates.

Fatty Acid Content and Profile
High nitrogen levels are related with low lipid accumulation in Scenedesmus species [52], because higher nitrogen levels were shown to be enough to avoid nutrient stress, yielding lower fatty acid content. In this study, results show that the fatty acids content was affected by the nitrogen concentration in the media (p-value: 0.027). Treatments with low nitrogen concentration had in average higher fatty acid content (47 ± 3% DW) than those with sufficient nitrogen (40 ± 6% DW). Similarly, co-culturing treatments had an average carbohydrate content of 42 ± 4% DW, while treatments without MGPB had an average of 45 ± 6% DW. The highest fatty acid content of 57 ± 3% DW was obtained in treatment 6, without bacteria and low nitrogen concentration, followed by treatment 2 (51 ± 5% DW) with bacteria and low nitrogen concentration. This can be explained because in treatment 2, in co-culturing, N is expected to be supplied through the nitrogen fixation relation. It is reported for Scenedesmus species, lipid contents of 50-80% measured gravimetrically [28], when using nitrogen and phosphorus deprivation strategies [31]. Thus, the values of fatty acid content, obtained in the present study, are within the literature range. Results also show that the fatty acids content was affected by the different mixing systems (p-value: 0.015). Comparing the mixing systems, aeration led in average to higher fatty acid content (48 ± 7% DW) than magnetic stirring (39 ± 2% DW), which can be explained since the carbon dioxide mass transfer is increased by aeration. This response in fatty acids accumulation by Scenedesmus cells has been previously reported, indicating its capacity for carbon fixation [52,57]. Table 2 presents the fatty acid profile of biomass obtained for each experimental treatment, showing that the unsaturated fatty acids (UFA) represent more than 80% of the total fatty acids (FA) content in all treatments, in which UFA with 18 carbons were predominant, as previously reported for Scenedesmus species [31,56,58]. Higher content of saturated fatty acids (SFA) was obtained in treatments with magnetic stirring. Further metabolic research is required to understand how carbon dioxide mass transfer affects fatty acid metabolism.

Conclusions
Microalgae biomass production for different applications, such as wastewater treatment and biofuels, is limited by several factors, such as nitrogen availability and mass transfer limitations. This work demonstrated that Scenedesmus sp. and A. brasilense co-culturing can be effectively performed on nitrogen-deficient media and in suspended cultures, without requiring immobilization of cultures, contributing to decrease nitrogen and energy inputs, while producing biomass for biofuels. Results show that in average, cultures with aeration reached higher biomass production, colony size and fatty acid content. The highest biomass concentration was obtained under aeration, with low nitrogen concentration, in the presence of A. brasilense. In particular, aeration contributed to higher fatty acid content and higher colony size than mechanical stirring. Also, co-culturing contributed in average, to higher colony size than without bacteria. Since A. brasilense is already approved and extensively used in agriculture and in wastewater treatment, without known environmental or health issues, there is no need to separate the bacteria from the resulting biomass, simplifying the cultivation and biomass processing for the desired final application such as biofuels production. Future research on long-term co-culture is expected to better explain the mutualist relationship between bacteria and microalgae and improve performance of co-culture systems.