Introduction

The production cycles of industrialized products, commodities and agroindustry have led to the generation and accumulation of increasing amounts of liquid and solid waste. The release of these wastes into the environment without treatment can cause pollution and deplete natural resources (España-Gamboa et al. 2011; Fuess and Garcia 2014).

Vinasse is one of the most polluting liquid wastes (Ohgren et al. 2006; Buitrón and Carvajal 2010). For each litre of distillate produced, an estimated 6–17 litres of vinasse is generated, depending on the beverage or product (Van Haandel 2005; Cabello et al. 2009; CRT 2016). This waste is also generated during the production of bioethanol, which is produced in several countries, including China (813 million gallons/year), the European Union (1387) and Brazil (7093). Currently, the USA is the main bioethanol producer at 14.806 million gallons/year (RFA 2016). Vinasse also results from the production of the spirit cachaça in Brazil (1.2 billion litres cachaça/year) (IBRAC 2016). Overall, approximately 320 billion litres of vinasse (CONAB 2016) are generated per annum in Brazil. The processing of tequila, a spirit distilled from agave, also generates high amount of vinasse by-product (227.2 million litres of tequila/year) (CRT 2016), generating some 2272 million litters of vinasses (CNIT 2016).

Vinasse presents physical and chemical characteristics that characterize it as a highly polluting residue. Generally, vinasse has a strong odour, dark colouration, pH value < 3 and high values for BOD (≥ 5000 mg L−1) and COD (≥ 12,000 mg L−1) (WHO 1995; COPAM 2008; Silva et al. 2011).

Different methods for treating and reusing vinasse have been tested in several countries such as Colombia (Ortegón et al. 2016), Mexico (López-López et al. 2010), Spain (Bueno et al. 2009), Egypt (Haggag et al. 2015), China (Jiang et al. 2012), Iran (Sadeghi et al. 2016) and Brazil (Silva et al. 2011; Campos et al. 2014; Pires et al. 2016). One of the most economic and profitable applications of vinasse is the production of microbial protein (single-cell protein [SCP]) (Dhanasekaran et al. 2011; Gao et al. 2012; Garcia et al. 2014; Pires et al. 2016) using yeast strains. Yeasts are microorganisms commonly present in vinasse during its production and, therefore, are adapted to the physical and chemical characteristics of this residue. The selection of strains to obtain SCP should be based on the presence of nutritional and anti-nutritional factors that are influenced by the composition of the culture medium (Bekatorou et al. 2006). Yeasts also present high protein, lipid, vitamin, carbohydrate and amino acid content as well as low nucleic acid content (Anapuma 2000; Pires et al. 2016).

The use of vinasse to produce SCP as supplementation in animal feed has already been reported by Rajoka et al. (2006), Silva et al. (2011), Nitayavardhana et al. (2013) and Pires et al. (2016). In this context, produced biomass should also be of high enough quality for food supplementation. Using vinasse as a substrate can also led to the reduction of the polluting parameters present in vinasse, allowing it to be directly disposed of in water (Campos et al. 2014) after confirming its low toxicity using bioindicator organisms (APHA 2005).

The aim of this work was to propose an aerobic biological treatment for vinasse that would simultaneously produce yeast biomass (SCP) as a value-added product that could potentially be used for supplementing animal feed.

Materials and methods

Vinasse

Fresh vinasse was provided by a cachaça producer in the State of Minas Gerais (MG), Brazil, during the 2014 harvesting season. Samples were collected immediately after the distillation of fermented sugarcane juice, transported and stored in aseptic tanks at − 20 °C until use. Prior to use, vinasse was filtered and sterilized at 121 °C for 20 min. Before starting the experiment, the physicochemical characteristics of vinasse were determined according to standard procedure (item 2.2).

Physicochemical analysis of vinasse

A physicochemical analysis of fresh and spent vinasse was performed to identify organic contaminants. Total sugars, glucose, sucrose, BOD (5210B), COD (5220B), electric conductivity (2510B), turbidity (SM number 2130C), colour (2120C), dissolved oxygen (DO) (4500-OB), dissolved solids (2540C), sediment solids (2540E) and total nitrogen (4500 N) were determined according to the American Public Health Association (APHA 2005. Manganese, copper, zinc and iron content was analysed by atomic absorption spectrometry (Malavolta et al. 1997).

Microorganism and inoculum concentration

Yeasts were obtained from the Culture Collection of Agricultural Microbiology (CCMA), Federal University of Lavras (UFLA), Lavras, MG, Brazil. The isolates were (1) Saccharomyces cerevisiae CCMA 0137, CCMA 0185, CCMA 0186 and CCMA 0187 (formerly UFLA CA15, UFLA CA76, UFLA CA93 and UFLA CA155), obtained during cachaça fermentation; (2) Saccharomyces cerevisiae CCMA 0188, CCMA O189 and CCMA 0190 (formerly PE2, CAT1 and VR1), obtained during alcohol fuel production; (3) Candida parapsilosis CCMA 0544 (formerly UFLA YCN448 and CCMA 0191) and Pichia anomala CCMA 0192 (formerly UFLA CAF70) and Candida glabrata CCMA 0193 (formerly UFLA CAF119), both isolated from fermented coffee. These isolates were selected according to the results reported in Silva et al. (2011).

Yeast was grown in YPD broth with 0.6% yeast extract, 0.6% peptone and 1.2% glucose (w/v). Cells were centrifuged at 10,000 g for 5 min and washed with sterile distilled water. Once the population reached 108 CFU mL−1, a 1:10 ratio of each isolate to the total volume of the culture medium (100 mL) was inoculated into 250-mL flasks with the addition of fresh vinasse, according to the CCRD conditions. Cultures were incubated with shaking (150 rpm) at 28 °C and sampled every 24 h to determine pH and viable cell count in a Neubauer chamber using methylene blue dye. Samples were then analysed by fluorescence microscopy using propidium iodide (PI) staining methods. For each yeast mixture, 80 μL aliquots of yeast suspension were mixed with 20 μL of a 20 μg/mL PI stock solution to yield a final concentration of 40 μg/mL PI. Western-blotted yeast samples were analysed by epifluorescence microscopy (excitation 525 nm and emission 595 nm), using an Apotome system to obtain images. Cultivation was ended once cells entered the stationary phase or died, usually after 48 h. Microbial biomasses were determined at the end of the incubation period.

Central composite rotational design (CCRD)

Previous work (Silva et al. 2011) showed that peptone and vinasse concentration influenced biomass production, so both variables were contemplated in the CCRD. Eight CCRD experiments were generated from two independent variables, presenting one central point and two axial points in which X1 = vinasse concentration (0.18, 1.0, 3.0, 5.0 and 5.82 v/v) and X2 = peptone (21.8, 30.0, 50.0, 70.0 and 78.2 g L−1).

The experiment was validated by testing the best conditions obtained during the optimization process and comparing their corresponding values with the values predicted by the model. The experiment was done in triplicate, based on five points under the conditions of interest. The same experimental procedures were used to build the models. Statistical analyses and graphs were performed in Design Expert® version 8.0 software (Stat-Ease Inc., Minneapolis, MN, EUA).

Nutritional characteristics of biomass production

The biomass produced in the culture medium containing 70% vinasse was evaluated in terms of nitrogen and protein content, DNA/RNA ratio and amino acid composition. Nitrogen content was determined by the Kjeldahl method (APHA 2005), and crude protein was estimated using a conversion factor of 6.25 (N × 6.25). Microbial DNA and RNA contents were determined using the Insta-GeneTM kit (Bio-Rad, Hercules, CA), according to the manufacturer’s instructions, and quantified at 260 and 280 nm, respectively, using a NanoDrop®ND-1000 (Thermo Fisher Scientific, Waltham, MA). Total amino acid composition was determined using a Shimadzu HPLC (Kyoto, Japan) equipped with a fluorescence RF-20A/20AXS detector, Shim-pack Amino-Na separation column, Trap Ammonia Shim-pack IS C-30Na column and EX detection at 350 nm. Mobile phase A consisted of a 99.5% ethanol aqueous solution of sodium citrate dihydrate, trisodium salt p. a. (0.2 N) and perchloric acid (pH 3.22). Mobile phase B consisted of an aqueous solution of sodium citrate dihydrate, trisodium salt p. a. (0.2 N), sodium hydroxide p. a. and boric acid (0.2 M; pH 10.0). A 0.6 mL/min flow rate was applied during 45 min. Amino acids were quantified by normalizing the peak area of each amino acid with respect to standard peak areas. The latter were derived under the same conditions as the samples. Amino acid content was expressed in grams of amino acid per 100 g of protein.

Toxicity analysis

The toxicity of spent vinasse was evaluated using the microcrustacean Daphnia similis in samples of fresh and spent vinasse. Residues are considered suitable for environment disposal when the survival rate of D. similis is greater than 50% (APHA 2005). Six concentrations of fresh and spent vinasse (1, 2.5, 5, 10 and 50%) and distilled water (as a positive control) were used. Five young microcrustaceans (6–24 h old) were exposed to the vinasse and to the negative control (100% vinasse) at 20 ± 2 °C in a 12/12-h photoperiod. In each concentration, immobility and/or mortality of individuals was observed after an exposure period of 48 h.

The immobility and/or mortality data for the test organisms were used to calculate the LC50 using the trimmed Spearman–Karber method (Hamilton et al. 1977) for estimating toxicity in bioassays. Four replications of each concentration, considering 20 individuals, were analysed. The relationship between fresh and spent vinasse was used to calculate the percentage of toxicity reduction using the formula proposed by Isidori et al. (2003):

$$ \% {\text{TR}} = 1 - \frac{{{\text{LC}}_{50} {\text{fresh}}}}{{{\text{LC}}_{50} {\text{spent}}}} \cdot 100 $$

%TR: toxicity reduction.LC50fresh: Lethal concentration with fresh vinasse.LC50spent: Lethal concentration with spent vinasse.

Results and discussion

Central composite rotational design (CCRD)

The 10 strains used in this study (Table 1) were evaluated for biomass production. Four strains (S. cerevisiae CCMA0186, S. cerevisiae CCMA0188, C. parapsilosis CCMA0544 and C. glabrata CCMA0193) showed the highest biomass production.

Table 1 Two-variable central composite rotational design (CCDR) for studying the effects of peptone and vinasse concentrations on yeast biomass production (mg L−1) during 120 h of cultivation at 28 °C and 150 rpm
Full size table

The pH, total soluble solids and viable cell counts (data not shown) were monitored throughout the treatment period in the CCRD experiments. The pH ranged from 3.5 to 6 in all treatments. After 48 h, yeast viability reduced by 2 logs. Maximum cell growth was achieved in the C. parapsilosis CCMA0544 assay, reaching 2 log CFU mL−1. Biomass production ranged from 41 mg L−1 (assay 8) to 388 mg L−1 (assay 4), reflecting the combined influence of peptone and vinasse concentrations on biomass (Table 1).

The C. parapsilosis strain CCMA0544 showed the highest biomass production (388 mg L−1) and was selected for optimization using a CCRD. The experimental design was 22, and 11 trials were conducted, including four axial tests and three central points. The best conditions were observed in assay 4 (peptone, 5 g L−1 and vinasse, 70 v/v), which showed a maximum biomass production of 388 mg L−1 (Table 1). The model fit was assessed by the determination coefficients (R2). The resulting regression equation indicated that R2 = 0.9938, with predicted and fitted values of 0.9591 and 0.9875, respectively, suggesting that the model adequately fitted the quadratic experimental data, explaining 99.38% of the variability in the response. The experimental results were modelled with a second-order polynomial equation to explain the dependence of the microbial growth on the two analysed factors (peptone and vinasse concentrations).

$$ {\text{Y}}_{{{\text{CCMA}}0544}} = \, + 241.00 + 43.51* \, X_{1} + 45.17* \, X_{2} + 20.50* \, X_{1} * \, X_{2} + 24.44* \, X_{1}^{2} + 12.19* \, X_{2}^{2} $$

where Y is the estimated biomass production and X1 and X2 are coded values for peptone (g L−1) and vinasse (v/v) concentrations, respectively. Thus, the biomass production can be estimated on the basis of the quadratic effect of both factors. The statistical significance of the model was checked by an F test (Table 2). The analysis of variance (ANOVA) of biomass production showed that the regression model was significant, which was described by a mathematical model based on the significant variables. Accordingly, the mathematical model described the biomass production based on the significant variables.

Table 2 ANOVA analysis of CCRD for the experimental results of the C. parapsilosis CCMA0544 grown for 5 days
Full size table

The regression analysis between X1 and X2 (Table 2), evaluated after 5 days of growth, was significant at a confidence interval of 95% (p < 0.05). As noted, the factors of X1 and X2 showed positive effects (43.51 and 45.17, respectively).

The combined peptone and vinasse concentrations were able to induce a greater yeast biomass production, which was also noted by the quadratic effect of X1 and X2. The analysis of variance showed significant differences in the effects of each analysed factor. The prediction of the optimal operating conditions for biomass production was determined experimentally using the response surface methodology (RSM). The interaction effect of the parameters that significantly affected biomass production by C. parapsilosis CCMA0544 is shown in Fig. 2. The curve in the response surface was plotted against two independent variables (peptone and vinasse concentrations) for the predicted response Y (biomass production).

The RSM defined the model for biomass production after 5 days of testing. Table 2 indicates that maximum biomass production occurred in the presence of higher peptone (A) and vinasse concentrations (B) (Fig. 1). The optimal conditions for biomass production (214.9 mg L−1) were obtained at 5 g L−1 peptone and 70 v/v vinasse concentrations. A validation test was conducted with new experiments carried out under the optimum conditions uncovered by the CCRD. The biomass yield was 494 mg L−1 for C. parapsilosis CCMA 0544, confirming the proposed model.

Fig. 1
figure 1

Effect of peptone and vinasse concentrations on C. parapsilosis CCMA0544 (coded levels). Response surface for biomass production during the central composite rotational design (CCRD)

Full size image

Physicochemical analysis of vinasse

The physical and chemical analysis of fresh vinasse samples (Table 3) revealed a low content of total sugars (0.013%), glucose (0.015%) and mineral components, amongst other analysed variables. The COD and BOD levels indicated that vinasse is biodegradable, so microorganism action is possible. Due to the low concentration of nutrients, the addition of glucose, yeast extract and potassium phosphate was necessary.

Table 3 Physical and chemical parameters of vinasse before and after treatment in comparison to the permissible limits established by Brazilian environmental legislation
Full size table

According to the parameters pre-established by COPAM (2008), spent vinasse showed values that surpassed the limits established by environmental legislation for pH, COD and BOD (Table 3).

Of the 10 isolates tested, four strains were able to minimize the COD and BOD present in vinasse. When grown in the presence of C. parapsilosis CCMA0544, spent vinasse displayed 3.9% higher DO and 55.8% and 46.9% lower BOD and COD, respectively; these levels were still above those allowed by COPAM (2008), CONAMA (2005) and WHO (1995), whose parameters serve as a worldwide standard. Mineral concentrations in fresh vinasse were low and were not detected in spent vinasse (Table 3). Values for turbidity, colour, total nitrogen, dissolved solids and sediment solids decreased in spent vinasse (Table 3). Reductions in BOD and COD were used to measure the efficiency in the treatment of vinasse waste.

Cell viability was analysed by fluorescence microscopy during the treatment period (Fig. 2). Each fluorescence image is an overlay of the fluorescence signal from the FL channel 1 (viable), presented using red false colour. Yeast cell counts were performed in the Apotome software, which clearly distinguishes viable cells from non-viable cells. The viable cells of the yeasts were effectively stained with high fluorescence intensities. On average, 45% of visualized cells died after 5 days of treatment. Saccharomyces cerevisiae CCMA0186 showed the highest death rate after 5 days of culture, while C. parapsilosis CCMA 0191 presented the greatest number of live cells (65%) (Fig. 2).

Fig. 2
figure 2

Yeast viability assay stained with propidium iodide (PI), as described in the methods. Four samples are shown for live yeast (left panel), live and dead yeast (central panel) and dead yeast (right panel). aS. cerevisiae CCMA0186, bS. cerevisiae CCMA0188, cC. parapsilosis CCMA0544, and D) C. glabrata CCMA0193

Full size image

Nutritional quality of biomass

The amino acid content of yeast biomasses was compared with the reference standards of FAO (1989) (Table 4). The nutritional analysis for maximum biomass of C. parapsilosis CCMA 0544 showed 55.01% protein content. Meanwhile, C. glabrata CCMA 0193 had the highest anti-nutritional content (nucleic acid ≅ 2.76%) (Table 4). Protein composition analysis revealed the presence of almost all essential amino acids except threonine, valine, leucine, phenylalanine and tryptophane in S. cerevisiae CCMA 0188 and leucine and tyrosine in C. parapsilosis CCMA 0544 (Table 4). In addition, nine non-essential amino acids were quantified, of which proline was the most abundant (17.58%). The concentration of lysine was high in all yeasts strains, but C. parapsilosis CCMA 0544 (40.38%) and S. cerevisiae CCMA 0188 (30.7%) had the highest values (Table 4).

Table 4 Protein profile and total nucleic and amino acids of S. cerevisiae CCMA0188, S. cerevisiae CCMA 0187, C. parapsilosis CCMA 0544 and C. glabrata CCMA 0193 when cultivated with vinasse as a substrate
Full size table

Toxicity analysis

The toxicity tests were performed with fresh and spent vinasse treated with four yeasts. The LC50 of juveniles were sensitive to a range of vinasse concentrations, from 4.72 to 7.77 mg L−1. The yeast C. glabrata CCMA0193 showed the lowest death percentage, but S. cerevisiae CCMA0188 showed the highest toxicity reduction (Table 5). The positive control (distilled water) was found to have 100% survival.

Table 5 Lethal concentration (LC50) for young D. similis after 48 h of exposure to fresh and spent vinasse
Full size table

In 100% vinasse (negative control), a mortality of 100% of the population was observed. The low pH value of fresh vinasse (3.5) was also an interferent in the assay since the microcrustaceans showed pH sensitivity.

Disposal of effluents in the environment is regulated by environmental laws. In Brazil, CONAMA (2005) and COPAM (2008) are responsible for establishing maximum thresholds, while in the USA and Mexico, the DEC (Department of Environmental Conservation 2016) and the Mexican norms (Norma Mexicana [NMX] 2012) have established permissible limits for potential pollutants. The World Health Organization (WHO 1995) also sets similar guidelines. No consensus exists with respect to the best parameters for evaluating or describing the contamination risk of effluents and residues, although the most commonly considered parameters are the COD and BOD values (Arvanitoyannis 2008). Biological treatment with four selected yeast strains was able to reduce the COD and BOD values of vinasse (COD/BOD 2.5). The total solids content is another parameter that influences the BOD and COD values, and its reduction may be associated with decreases in these parameters. However, the COD value after vinasse treatment still surpassed the permissible threshold. In this case, the solids concentration may have influenced the below optimal COD reduction in addition to the presence and possible influence of oxidizable inorganic compounds.

According to the norms established by the COPAM/CERH-MG Joint Legislative Decree No. 1 of 5 May 2008, BOD values should be less than 60 mg L−1 or be reduced by 60% after treatment because the release of organic load from vinasse into freshwater stimulates increases of the algal growth, leading to depletion of the oxygen dissolved (DO). Moreover, the organic content in the waste is oxidized biochemically by microorganisms measured in BOD values. Added to this, the organic material in the waste undergoes chemical oxidation (COD values). That is, high organic content reflects high oxygen consumption that compromises the maintenance of aquatic life and, consequently, the death of living organisms. In this study, the maximum reduction (55.8%) was very close to the desired reduction in the case of C. parapsilosis CCMA0544 cultivated in vinasse. These results indicated that some adjustments, such as the use of physicochemical methods in subsequent treatment stages to reduce the organic load, might enable compliance with the standards established by legislation in order for vinasse to be discarded in the environment. A study conducted by Campos et al. (2014) showed that biological treatment using an 8-stage physicochemical method led to a BOD reduction of 96.7%. López-López et al. (2010) also observed that treatment with microorganisms represents a viable method for treating vinasse; however, the steps suggested in this study were far more complex than the ones proposed here. Considering the other physical and chemical parameters evaluated, the treatment of vinasse with 4 yeasts strains was efficient, and C. parapsilosis CCMA0544 showed the highest potential to treat vinasse. In this scenario, vinasse could be treated during the simultaneous production of high quality yeast biomass.

Based on the physical and chemical characteristics of the spent vinasse, the maximum permissible flow of effluent to be discarded in water bodies was calculated. For example, in a river with a water volume of 45 m3 s−1, vinasse can be discarded at a flow rate of around 0.8–1 m3 s−1 without exceeding the established limits for colour and turbidity (Tigini et al. 2011). However, due to the ecotoxicity of vinasse, as demonstrated by the lethality of Daphnia similis spent vinasse should be released at a concentration of less than 6%. Daphnia crustaceans are widely used in toxicity testing because of their wide distribution in rivers and ponds and their importance in the food chain (APHA 2005). Today, D. similis is the most widely used bioindicator organism for measuring toxicity due to its high sensitivity to toxic by-products (Vesela and Vijverberg 2007). Campos et al. (2014) proposed a combined physical, chemical and biological treatment for vinasse that presented a toxic effect to Daphnia sp., similar to the one found in this study, reinforcing the necessity of biological treatment to increase treatment effectiveness. Silva et al. (2011) and Pires et al. (2016) also studied vinasse as a substrate for yeast growth. Usually, vinasse presents a toxic effect due to its chemical composition, which stresses yeast cells and induces pseudomycelia formation (Silva et al. 2011), causing death after 96 h of exposure. However, even with the reduction of viable cells, this method for producing biomass is advantageous as a result of its low cost and the high nutritional value of the produced biomass, which can be destined for the supplementation of animal feed (Nytayavardhana et al. 2013).

The addition of peptone to vinasse at different concentrations affected the performance of biomass production in the four selected strains (CCMA0188, CCMA0187, CCMA0193 and CCMA0544), as expected. Yeasts require certain concentrations of carbon and nitrogen, which are involved in the regulation of cell growth (Schneper et al. 2004). Another advantage of using yeast is that it leads to an increase in pH without chemical additives, thereby avoiding additional costs and meeting the requirements that pH values of final effluent range from 6.0 to 9.0, as set by the Brazilian regulatory agency (CONAMA 2005) and the World Health Organization (WHO 1995).

The use of cheap and abundant raw materials to produce SCP using microorganisms still represents one of the best and most economical technologies for the commercialization of SCP (Pires et al. 2016). As a dietary supplement, microbial biomass should contain adequate levels of nutritional compounds (e.g. total nitrogen) and low levels of anti-nutrients (e.g. nucleic acid ≤ 10% dry mass) (Silva et al. 2011; Pires et al. 2016; FAO 2006).

Microbial production efficiency and microbial flow are determinants of the amount of microbial protein that reaches animals’ small intestine (Cavalcante et al. 2006). The synthesis of microbial proteins can result in an improved balance of amino acids, confirming that the production of microbial biomass (SCP) for feed is a good method for supplementing animal diets (Sumar et al. 2015).

Amino acid composition is the most important factor in food protein quality, followed by protein digestibility and bioavailability of amino acids (Wolfe et al. 2016). Most non-ruminant animals require the ingestion of essential amino acids such as arginine, histidine, isoleucine, lysine, methionine, phenylalanine, threonine, tryptophan and valine (FAO/WHO/UNU 1989). In the present study, the determination of the constituent amino acids of the protein present in the generated microbial biomasses revealed the presence of all essential amino acids except for methionine and cysteine. Overall, the biomass produced using vinasse as a substrate was an adequate and economically viable source of protein for feed supplementation. Additional studies should determine the amount of protein and amino acid that can be extracted from SCP in order to verify further the suitability of such generated proteins for animal diets (Ahmadi et al. 2010).

Conclusion

The proposal of the joint biological treatment of vinasse effluent and production of SCP as a value-added product was satisfactory from an economic and environmental perspective. Some modifications could be made to the biological treatment process, such as the use of physicochemical methods in subsequent treatment stages in order to increase the reduction of the organic load. The use of immobilized cells represents another possibility for increasing cell viability and consequently biomass productivity as well as for reducing the organic load. The biological treatments suggested herein are low cost, have low technological complexity and could easily be reproduced on a large scale. The efficiency of our proposed method could be increased in subsequent cycles during constant exposure of C. parapsilosis CCMA0544 to vinasse.