Arthrospira platensis biomass with high protein content cultivated in continuous process using urea as nitrogen source

Authors

  • I. Avila-Leon,

    1. Department of Biochemical and Pharmaceutical Technology, Faculty of Pharmaceutical Sciences, University of São Paulo, São Paulo, Brazil
    Search for more papers by this author
  • M. Chuei Matsudo,

    1. Department of Biochemical and Pharmaceutical Technology, Faculty of Pharmaceutical Sciences, University of São Paulo, São Paulo, Brazil
    Search for more papers by this author
  • S. Sato,

    1. Department of Biochemical and Pharmaceutical Technology, Faculty of Pharmaceutical Sciences, University of São Paulo, São Paulo, Brazil
    Search for more papers by this author
  • J.C.M. de Carvalho

    1. Department of Biochemical and Pharmaceutical Technology, Faculty of Pharmaceutical Sciences, University of São Paulo, São Paulo, Brazil
    Search for more papers by this author

João Carlos Monteiro de Carvalho, Department of Biochemical and Pharmaceutical Technology, Faculty of Pharmaceutical Sciences, University of São Paulo, Av. Prof. Lineu Prestes, 580, B-16, 05508-900, São Paulo, Brazil. E-mail: jcmdcarv@usp.br

Abstract

Aims: Arthrospira platensis has been studied for single-cell protein production because of its biomass composition and its ability of growing in alternative media. This work evaluated the effects of different dilution rates (D) and urea concentrations (N0) on A. platensis continuous culture, in terms of growth, kinetic parameters, biomass composition and nitrogen removal.

Methods and results: Arthrospira platensis was continuously cultivated in a glass-made vertical column photobioreactor agitated with Rushton turbines. There were used different dilution rates (0·04–0·44 day−1) and urea concentrations (0·5 and 5 mmol l−1). With N0 = 5 mmol l−1, the maximum steady-state biomass concentration was1415 mg l−1, achieved with = 0·04  day−1, but the highest protein content (71·9%) was obtained by applying D = 0·12 day−1, attaining a protein productivity of 106·41 mg l−1 day−1. Nitrogen removal reached 99% on steady-state conditions.

Conclusions:  The best results were achieved by applying N0 = 5 mmol l−1; however, urea led to inhibitory conditions at  0·16 day−1, inducing the system wash-out. The agitation afforded satisfactory mixture and did not harm the trichomes structure.

Significance and Impact of the Study:  These results can enhance the basis for the continuous removal of nitrogenous wastewater pollutants using cyanobacteria, with an easily assembled photobioreactor.

Introduction

Cyanobacteria have an enormous industrial and economic potential as important sources for pharmaceuticals, carotenoids (Spolaore et al. 2006; Guedes et al. 2011), vitamins (Becker 1981), fine chemicals and biofuels (Chisti 2007). In this context, Arthrospira platensis represents an alternative source of proteins and a very promising food supplement or additive (Vonshak 1990), because its biomass contains up to 70% of protein in its dry weight (Cohen 1997) and good digestibility owed to low nucleic acids content (Ciferri and Tiboni 1985).

Arthrospira platensis has been commercialized since the seventies (Belay, 1997) and, according to Vonshak (1997), more than 70% of the produced biomass is destined for human consumption, mainly as health food. It is currently cultivated in several countries like USA, China, Taiwan, Japan, Israel and Thailand and commercialized as powder, pills and bars (Spolaore et al. 2006).

Although the Arthrospira genus has been officially included in Bergey’s Manual of Systematic Bacteriology (Castenholz 1989), the species A. maxima and A. platensis, widely cultivated in industrial scale, are often referred to as Spirulina maxima and Spirulina platensis. The sight of Arthrospira and Spirulina as two distinct genera has been repeatedly confirmed based on many morphological and genetic characteristics (Tomaselli 1997).

The conventional nitrogen sources are potassium or sodium nitrates (Schlösser 1982; Paoletti et al. 1985), even though alternative sources such as ammonium salts or urea have recently been successfully exploited to reduce costs (Danesi et al. 2002; Sassano et al. 2007; Ferreira et al. 2010). Although urea is a cheap nitrogen source, its use could be hindered in batch process, because part of the ammonia released from its hydrolysis in alkaline conditions (Danesi et al. 2002) and/or the release of urease (Donoso et al. 1982) could be lost by off-gassing (Carvalho et al. 2004), restraining microbial growth. On the other hand, an excessive release of ammonia could inhibit the growth at high levels (Sassano et al. 2007); therefore, the control of urea supply during the cultivation is a fundamental requisite (Matsudo et al. 2009), and the use of continuous process is a good alternative to adjust the nitrogen supply and to avoid inhibitory levels.

Previous studies with urea in fed-batch allowed biomass concentrations and productivities comparable to those obtained with batch process using nitrate salts (Danesi et al. 2002; Rangel-Yagui et al. 2004; Matsudo et al. 2009), suggesting that continuous cultures could control the ammonia levels in the medium, tendering a feasible alternative for A. platensis cultivation using urea-containing media. In fact, Kim et al. (2000) reported a continuous process to simultaneously reduce several nitrogenous compounds in complex medium (swine waste) and to produce feed resource. Conventionally, the cyanobacteria or microalgae wastewater treatment systems could be employed as postsecondary treatment process to remove nutrients from wastewater (Park et al. 2010), basically because in aerobic and anaerobic digestion, the most common methods of wastewater treatment (Markou and Georgakakis 2010), the concentration of nitrogen compounds remains practically constant (Campos et al. 2009), which could allow algal blooms and other environmental and ecological impacts because of water eutrophication (Lee and Lee, 2002). It is known that outdoor cultures of microalgae suffer from contamination by other micro-organisms such as bacteria and other microalgae genera (Markou and Georgakakis 2010). However, it has been observed that supplementation with sodium bicarbonate (Lincoln et al. 1996) and high pH values (Shapiro 1990) avoids chlorophytes growth (e.g. Chlorella), one of the main contaminants present in Arthrospira cultivation. Furthermore, at optimal pH values for Arthrospira growth (9·0–9·5), concentration of ammonia above 2·5 mmol l−1 can be toxic for many algae (Abeliovich and Azov 1976) but not for this cyanobacterium (Belkin and Boussiba 1991). In terms of scaling up, the cultivation of Arthrospira on digested waste encounters the same constraints as those of commercial algal production as, for example, the availability of carbon, nitrogen and phosphorus, optimum temperature and light conditions (Richmond 1990). Another important drawback of the cultivation of Arthrospira on wastes, especially if the emphasis is put on the efficiency of inorganic nutrient removal, is the high dilution of the digested waste needed for the growth of this cyanobacterium, regarding the ammonia toxicity, which is certainly the major factor explaining the need for considerable dilution of the waste (Lalibertéet al. 1997).

The present study is focused on the continuous cultivation of Arthrospira platensis in a simple vertical glass-made column photobioreactor as an alternative for wastewater treatment. The effects of different dilution rates and urea concentrations on parameters like steady-state biomass concentration, cell productivity, nitrogen-to-cell conversion factor, protein content and urea removal were evaluated. Such information could be of great significance not only in clean methodologies for the cultivation of this cyanobacterium for its use as protein supplement, but also in the field of wastewater treatment.

Material and Methods

Micro-organism and cell maintenance

Arthrospira platensis UTEX 1926 was obtained from the University of Texas Culture Collection. This micro-organism was maintained at 25°C using a standard medium (Schlösser 1982), containing the following nutrients (g l−1): NaHCO3, 13·6; Na2CO3, 4·03; NaCl, 1·00; K2SO4, 1·00; NaNO3, 2·50; K2HPO4, 0·50; MgSO4·7H2O, 0·20; CaCl·2HO, 0·04. All nutrients were dissolved in distilled water containing (per litre): 6 ml of metal solution (97 mg FeCl·6 H2O, 41 mg MnCl2·4H2O, 5·0 mg ZnCl2, 2 mg CoCl2·6H2O, 4·0 mg Na2MoO4·2H2O) and 1 ml of micronutrient solution (50·0 mg Na2EDTA, 618 mg H3BO3, 19·6 mg CuSO4·5H2O, 44·0 mg ZnSO4·7H2O, 20·0 mg CoCl2·6H2O, 12·6 mg MnCl2·4H2O, 12·6 mg Na2MoO4·2H2O).

Inoculum preparation

Cells grown in the above maintenance medium were transferred into 500 ml Erlenmeyer containing 200 ml of the same medium and kept for 6–8 days on rotary shaker at 100 min−1, 30°C and photosynthetic photon flux density of 72 μmol photons m−2 s−1 (Carvalho et al. 2004). Biomass was harvested during the exponential growth, filtered and washed with physiological saline solution (NaCl 0·9% w/v) to remove the original nitrogen source (NaNO3).

Culture conditions

To investigate the effect of continuous supply of urea as nitrogen source, the medium was altered by substituting NaNO3 by urea and using a different proportion of inline image/inline image to achieve pH 9·3, but keeping the same overall carbon amount. The resulting medium contained 14·28 g l−1 of NaHCO3, 3·18 g l−1 of Na2CO3 and 0·5 or 5 mmol l−1 of urea; the concentration of the other nutrients was unaltered.

There were used a glass-made vertical column photobioreactor, with 20·7-cm internal diameter, 46 cm in height and 0·5 cm glass thickness (allowing a working volume of 9·0 l), with four 2·0-cm-wide chicanes, ensuring optimum medium homogeneity (Sassano et al. 2007). The mechanical agitation was provided by three-flat-blade turbines at 140 rev min−1. A peristaltic pump was used to feed the fresh medium as well as to simultaneously remove the exhausted broth. The reactor volume was kept constant using an inlet flow rate lower than the outlet one and positioning the outlet tube at a selected upper height. The homogeneity of the bioreactor was evaluated in accordance with the methodology described by Koshimizu et al. (1983).

A. platensis cultivation

The filtered and washed biomass was inoculated in the photobioreactor with an initial biomass concentration of 400 mg l−1. The fed-batch startup process was performed in the same photobioreactor subsequently utilized for the continuous tests at 30°C (Carvalho et al. 2004) by regulating the room temperature at 26°C, and it was carried out by daily intermittent addition of 0.145 mmol l−1 urea (CH4N2O; Synth- Brazil, Diadema, Brazil). The continuous cultivations were performed by adding urea-containing medium (either 0·5 or 5 mmol l−1) at different dilution rates (D) (Table 1). For both fed-batch startup and continuous operations, the light intensity was regulated at 108 μmol photons m−2 s−1. The selected light intensity was measured with an illuminance meter (Minolta T, Osaka, Japan) at the surface of the bioreactor exposed to six 20 W-fluorescent lamps. Steady-state conditions were assumed to take place when nitrogen and biomass concentrations as well as pH in the medium are kept almost constant (coefficient of variation ≤ ±5%) for at least 1 week (Sassano et al. 2007).

Table 1.   Experimental conditions and results of continuous A. platensis cultivations
RunN0* (mmol l−1)D (day−1)XS (mg l−1)NF§ (mmol l−1)pHPX** (mg l−1 day−1)YX/N†† (mg mg−1)PTN‡‡ (%)PPTN§§ (mg l−1 day−1)YPTN/N¶¶ (mg mg−1)
  1. *Urea concentration in the feed.

  2. Dilution rate.

  3. Biomass concentration under steady-state conditions.

  4. §Ammonia concentration in the medium during steady-state.

  5. ¶pH in the cultivation medium during steady-state.

  6. **Biomass productivity.

  7. ††Nitrogen-to-biomass conversion factor.

  8. ‡‡Protein content of biomass.

  9. §§Protein productivity.

  10. ¶¶Nitrogen-to-protein conversion factor.

  11. !–: System wash-out.

 10·50·08683 ± 10·20·0001 ± 0·000110·1254·448·7918·610·119·07
 20·50·12649 ± 5·90·0004 ± 0·00019·997846·3619·214·978·9
 30·50·16625 ± 6·80·0012 ± 0·00019·9399·9644·6422·522·4910·04
 40·50·24475 ± 7·60·0035 ± 0·00099·811433·9322·525·657·63
 50·50·32414 ± 4·50·0044 ± 0·00229·56132·629·5726·729·77·9
 60·50·4421 ± 6·40·0003 ± 0·00019·7616830·0726·845·028·06
 70·50·44458 ± 5·00·0004 ± 0·00019·75201·632·7127·254·838·9
 850·041415 ± 5·60·0034 ± 0·000210·3156·610·1128·316·012·86
 950·081235 ± 7·90·0013 ± 0·00069·8798·818·8256·856·125·01
1050·121235 ± 9·00·0251 ± 0·0083101488·8271·9106·416·34
1150·164·6797 ± 0·9631
1250·325·3147 ± 0·8248

Analytical techniques

Dry cell mass concentration was determined, as described by Leduy and Therien (1960), by optical density (OD) measurements at 560 nm using a calibration curve. The concentration of free ammonia was determined in cell-free medium samples, after preliminary pH adjustment at 13·0 with NaOH 1·5 mol l−1, using a potentiometer and a selective ammonia ion electrode (Orion, Beverly, MA, USA) (LeDuy and Samson 1982). The residual urea was determined in the same way, with previous treatment with H2SO4 at 200°C (Cezare 1998). After steady-state achievement, all the biomass contained in the bioreactor was centrifuged, washed twice with distilled water, dried in furnace at 55°C with air circulation for 12 h and then stored in glass flasks at −20°C for posterior cell mass composition analyses. Protein content was analysed in accordance with the method of the Association of Official Analytical Chemists (1984).

Calculation of the response variables of A. platensis cultivation

Biomass productivity, PX (mg l−1 day−1), was calculated as the product of D (day−1) and the steady-state biomass concentration, XS (mg l−1). Proteins productivity, PPTN (mg l−1 day−1), was calculated as the product of PX and the protein content in dry biomass, PTN. Steady-state nitrogen-to-biomass conversion factor, YX/N (g g−1), was calculated as the ratio of XS to the difference between feed and outlet total ammonium concentration (mg l−1). The steady-state nitrogen-to-protein conversion factor, YPTN/N (g g−1), was calculated as the product of YX/N to PTN.

Results

Taking into account that mixing is one of the most significant factors influencing cyanobacteria growth, ensuring cell suspension as well as homogeneity of nutrients, temperature and light intensity to the culture (Becker 1981), the homogeneity of the bioreactor was tested in accordance with Koshimizu et al. (1983). When the bioreactor works under homogeneous conditions, the biomass concentration falls exponentially (data not shown). The experimental value of D calculated by this method (0·0087 day−1) was only 3·6% higher than the theoretically estimated (0·0084 day−1); hence, the mechanical agitation proved to be adequate for the homogeneity of the system and suitable for continuous process.

Two cultivations with standard Schlösser medium were performed to determine the daily nitrogen requirement of A. platensis, and consequently, the urea values to be added during the fed-batch startup culture, which preceded the continuous runs. The daily nitrogen requirement was considered as 7% of the mean productivity obtained (56·7 ± 8·06 mg l−1 day−1), which corresponds to the nitrogen incorporated to the overall biomass (Bezerra et al. 2008). This daily concentration of urea supplied during the fed-batch startup process was established in 0·145 mmol l−1 day−1.

Table 1 lists the results of continuous cultures performed at dilution rate increasing from 0·04 to 0·44 day−1, using either 0·5 or 5 mmol l−1 urea concentrations in the feed medium. In general, after approx. 1 week, the biomass concentration needed was achieved for the beginning of the continuous process. After a few days, it was possible to reach the steady-state condition for most of the runs, as it can be seen in Fig. 1.

Figure 1.

 Biomass concentration (X: ■) and pH (○) vs time in run 9. N0 = 5 mmol l−1; D = 0.08 day−1.

The residual ammonia values were calculated by adding the free ammonia concentration value (NF) (Table 1 and Fig. 2) to the ammonia coming from urea, the latter after digestion from the exhausted broth. When N0 = 5 mmol l−1 and the steady-state was achieved, it was observed that nitrogen removal was as high as 99%.

Figure 2.

 Logarithm of steady-state available ammonia concentration (LogNF) vs dilution rate (D). (■): N0 = 0.5 mmol l−1; (⋄): N0 = 5 mmol l−1.

Figure 3 represents the protein contents in A. platensis biomass. The protein content increased along the raise of the dilution rate. In runs with N0 = 5 mmol l−1, the protein content was up to 72%, for = 0·12 day−1 almost three fold higher than those with N0 = 0·5 mmol l−1 and highest D value (0·44 day−1) (Table 1).

Figure 3.

 Protein content vs dilution rate. (inline image): N0 = 0.5 mmol l−1; (■): N0 = 5 mmol l−1.

The biomass concentrations and the kinetic parameters are shown in Table 1. As it can be seen, in runs 11 and 12, the system wash-out occurred. The influence of the dilution rate (D) and urea concentration (N0) on steady-state biomass concentration (XS), biomass productivity (PX) and nitrogen-to-cell conversion factor (YX/N) is illustrated in Figs 4, 5 and 6, respectively. In Fig. 4, it was shown that XS had a decreasing behaviour when the D value increased. Figure 5 shows that D had a positive influence on biomass productivity (PX), considering the runs in which steady-state condition was achieved. On the other hand, nitrogen-to-biomass conversion factor (YX/N) decreased as the D value increased, as it can be seen in Fig. 6.

Figure 4.

 Steady-state biomass concentration (XS) vs dilution rate (D). (▲): N0 = 0.5 mmol l−1; (■): N0 = 5 mmol l−1.

Figure 5.

 Steady-state biomass productivity (PX) vs dilution rate (D). (▲): N0 = 0.5 mmol l−1; (■): N0 = 5 mmol l−1.

Figure 6.

 Steady-state values of nitrogen-to-biomass conversion factor (YX/N). (▲): N0 = 0.5 mmol l−1; (■): N0 = 5 mmol l−1.

Discussion

The radial flow with Rushton turbines seemed to be appropriate for cell growth, and microscopic observations of the cells did not permit to observe any noticeable injurious effect on the trichomes structure, thus evidencing that the homogeneity of the photobioreactor can be obtained without shear stress. This fact suggests that high height tubular bioreactors can be used for the cultivation of A. platensis without using air to maintain the homogeneity of the cell suspension. Consequently, less quantity of volatile compounds would be lost by off-gassing using this type of agitation process, in contrast with the air-lift system.

During the fed-batch startup stage, in all the runs, the pH increased as expected (Fig. 1), because of the bicarbonate uptake (Goldman et al. 1982). Because bicarbonate is the natural form of carbon assimilated by this cyanobacterium, a feed medium with pH 9·3 was necessary to avoid excess alkaline conditions, which could have affected biomass productivity (Jimenez et al. 2003). Under steady-state conditions, the lowest pH values were obtained with the highest D values, because of the withdrawal of the exhausted broth that avoided carbonate accumulation and thus alkaline conditions. When N0 = 5 mmol l−1, the pH values were above 10, as a result of a higher growth rate, which carried the highest values of biomass concentration.

It was observed that biomass concentration on steady-state (XS) was a decreasing function of the free ammonia concentration (NF) in the bioreactor, as evidence of its assimilation by the cyanobacterium. As observed in Fig. 2, the NF values were an increasing function of D; when  0·12 day−1 and N0 = 5 mmol l−1, the higher flow rates overcame the ammonia assimilation by A. platensis, accumulating this molecule in the system and granting inhibitory concentration levels. Similar results were observed by Sassano et al. (2007), in which, feeding with ammonium chloride, the use of  0·15 day−1 led to the system wash-out, achieving (NF) values up to 6 mmol l−1, considered to be inhibitory levels in A. platensis fed-batch cultivation (Carvalho et al. 2004).

In this study, when > 0·12 day−1, and applying 5 mmol l−1 of urea in the feed, the NF reached values up to 4·7 mmol l−1 and the system wash-out was observed. This inhibition, occurred in such a lower NF value, could be explained as a result of a prolonged exposure of the cyanobacteria to the accumulated ammonia during the continuous process, reminding that the inhibitory value considered by Carvalho et al. (2004) applies for fed-batch process, because the contact with this molecule is less pronounced. Furthermore, in these runs, where > 0·12 day−1, cell death is because of the incapability of A. platensis to equalize the specific growth rate with the dilution rate, which bestowed the ammonia accumulation in the system.

Biomass colour was affected by lower nitrogen feeding rate. In runs 1–3, when N0 = 0·5 mmol l−1 and  0·16 day−1, the biomass became pale and lost its typical blue-green colour. This alteration was also observed by Walach et al. (1987) and Sassano et al. (2007), the latter suggesting that the nitrogen shortage could have affected even the chlorophyll synthesis. Another fact to be considered is that when nitrogen deprivation occurs in A. platensis culture, the cyanobacteria use the stored phycocyanin as a nitrogen source (Ciferri 1983; Richmond 1986), contributing to the loss of colour in the culture.

As commented earlier, when N0 = 5 mmol l−1 and the steady-state was achieved, it was observed that nitrogen removal was as high as 99%, revealing that the nitrogen removal was a decreasing function of the dilution rate, as a consequence of less biomass values once the D value raised. On the other hand, runs with N0 = 0·5 mmol l−1 did not displayed a correlated function for nitrogen removal, probably due to the lower biomass levels reached under this urea concentration. Lee and Lee (2002) did in fact conclude that higher biomass densities assured the effectiveness of wastewater treatment, especially when microalgae were used for wastes with high nitrogen content. As expected, the results in this study ensured the suitability of using continuous process in A. platensis cultivation for wastewater treatment, alongside previous works (Kim et al. 2000; Sassano et al. 2007), which attained similar removal percentages treating organic and inorganic nitrogen, respectively.

The protein content in A. platensis biomass was certainly influenced by the dilution rate. When the nitrogen supply increased, regarding the rise in the D value, it released greater amounts of ammonia by the hydrolysis of urea, permitting a higher assimilation of this compound. According to Syrett (1962), nitrogen sources are used for cellular growth as well as a storage material, permitting its accumulation in an organic form. Sassano et al. (2007) achieved higher protein contents up to 72% in continuous process, feeding with ammonium chloride and D = 0·08 day−1. In the present research, a similar value was obtained under a higher dilution rate (D = 0·12 day−1), certainly due to the necessity of the urea hydrolysis for nitrogen assimilation and consequently a higher feeding rate.

There are reports in the literature regarding wastewater with high amounts of urea and other nitrogen compounds. Vidal et al. (1999) reported wastewater containing high loads of urea and ammonia, with a total nitrogen concentration of almost 1 g l−1, treated in anaerobic and anoxic systems. Rittstieg et al. (2001) treated an industrial wastewater containing 1·2 g l−1 of urea, along with high amount of sulfate, proposing the use of the aerobic process to avoid the production of sulfide. Likewise, there are several studies emphasizing the need of diluting wastewater for the use of cyanobacteria and microalgae as treatment agents. Park et al. (2010) using the microalga Scenedesmus sp. for nitrogen removal of an anaerobically digested swine effluent wastewater diluted 1/10 upon the addition to the culture, with a total nitrogen concentration of 1·2 g l−1 in pure wastewater, being mostly NH4-N. Also, Chaiklahan et al. (2010) found that the effluent from an anaerobic sludge blanket from a pig farm was a suitable medium for the growth of A. platensis, obtaining an enriched biomass with up to 57% protein and employing 20% of this effluent in the culture media, which contained 144 mg l−1 of ammonia and 19·9 mg l−1 of nitrate. Taking this into account, the urea removal system in the current study, using A. platensis in an easily assembled photobioreactor, is presented as a relevant exploration in postsecondary wastewater treatment, despite the fact of requiring the wastewater dilution for avoiding ammonia toxicity, as remarked before. The urea concentrations employed in this study (300 and 30 mg l−1 for 5 and 0·5 mmol l−1, respectively) are in the range of study of cyanobacteria cultivation with diluted digested wastewater.

In relation to the kinetic parameters studied, in runs 11 and 12, the system wash-out occurred, as evidence that the dilution rate was higher than the maximum specific growth rate (μmax). It also indicates that the μmax decreased when the urea values raised, probably due to the inhibitory effect of the available ammonia in the medium, attaining higher values in runs 11 and 12 (Fig. 2).

Figure 4 shows that steady-state biomass concentration (XS) decreased with the increase in D value. This effect was awaited, because the biomass withdrawal rose with higher dilution rates, as well as an excess in the ammonia concentration. As remarked before, increasing cell removal by the outlet stream is the reason for lower biomass concentrations and consequently lower nitrogen requirements when using higher dilutions rates (Sassano et al. 2007). The same authors evidenced that higher dilution rates led to larger discharge of cells, but this effect was partly counterbalanced by the higher growth rate because of larger nitrogen source supply. According to that, the effect of the nitrogen supply rate (NR) was also analysed as a product of the dilution rate and the urea concentration applied. Higher NR values, obtained in runs 8–10, influenced the biomass concentration, permitting higher amounts of free ammonia and, therefore, upper XS values when compared to runs 1–7. Still, in runs 11 and 12, the NR values allowed inhibitory levels of ammonia (Table 1) and consequently the system wash-out.

Biomass productivity (PX) gradually increased as the D value increased in all the runs with steady-state conditions (Fig. 5). This could be explained as an effect of the D value on the nitrogen supply rate, which enlarged the nitrogen available for cell growth. There were expected higher PX values with higher N0 values. Nevertheless, the runs performed at N0 = 0·5 mmol l−1 exhibited higher cell productivities, attaining a maximum value of 201·6 mg l−1 day−1 at D = 0·44 day−1, which is higher than those attained in continuous process with ammonium chloride (PX = 92 mg l−1 day−1) (Sassano et al. 2007) and in fed-batch cultures with urea (PX = 69·5 mg l−1 day−1) (Sánchez-Luna et al. 2007). On the other hand, the cell productivity was not optimized when N0 = 0·5 mmol l−1, because the protein content at the highest D value was below the desired, and it could be considered as nonviable in terms of protein productivity. Considering that in run 10 (N0 = 5 mmol l−1 and D = 0·12 day−1) the protein productivity was tenfold higher in comparison with the same D and N0 = 0·5 mmol l−1 (Table 1), it was then inferred that higher nitrogen supply rate not only enabled higher biomass concentration but also provided adequate nitrogen supply for protein synthesis and accumulation. These results justify the fact that there were no runs with D higher than 0·44 day−1, when N0 = 0·5 mmol l−1.

A decrease in the nitrogen-to-biomass conversion factor (YX/N) was observed as the D value increased (Fig. 6). This result was awaited, because the continuous process theory argues that the higher the dilution rate, the higher the cell removal, and thereby a lower biomass concentration inside the bioreactor (Aiba et al. 1973). Furthermore, the increase in D implied a consequent raise in the amount of added nitrogen that led to a decrease in YX/N, as observed previously for fed-batch culture with ammonium chloride (Carvalho et al. 2004) and continuous cultures with urea (Sassano et al. 2007; Matsudo et al. 2011) for A. platensis. Comparing YX/N values (Table 1) with the protein content accumulated in the biomass (Fig. 3), it can be inferred that in runs with N0 = 0·5 mmol l−1 (runs 1–7), most of the nitrogen incorporated was assimilated to cellular growth and survival of the micro-organism in the system conditions, but the formation of organic nitrogen in the form of proteins was affected by the limitation of substrate, as a consequence of a low nitrogen supply rate. On the other hand, as proteins have almost the same nitrogen content (Llewellyn et al. 1932), the nitrogen-to-protein conversion factor (YPTN/N, Table 1) was less influenced by the different conditions tested. In runs 8–10, the adequate nitrogen supply conditions were responsible for an increase in the protein content of biomass where the steady-state condition was attained, which is also consistent with the observation by Matsudo et al. (2011).

Applying N0 = 5 mmol l−1, when steady-state condition was reached, high protein content and nitrogen removal up to 99% were achieved. The results in this study confirm the suitability of A. platensis for nitrogen sequestration in wastewater treatment, by using an easily assembled photobioreactor. The study of urea removal, along with the reduction of biological oxygen demand by the oxygen produced in the photosynthesis, presents A. platensis cultivation as a good outlook in wastewater treatment. Furthermore, the achievement of an enriched biomass in this study presents the feasibility of manipulating cyanobacterial biomass’ composition, by controlling the cultivation conditions, for different purposes, such as an animal feed supplement or highly valuable biomolecules.

Acknowledgements

The authors acknowledge the financial support from the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) of Brazil.

Ancillary