The microalgae strain D. subspicatus was provided by the laboratory of Ecotoxicology of the University of Santa Cruz do Sul (UNISC), State Rio Grande do Sul, Brazil. The microalgae strain of Chlorella sp. was generously provided by the Federal University of Santa Catarina (UFSC), State Santa Catarina, Brazil. D. subspicatus and Chlorella sp. were cultivated in photobioreactors and 1 L Erlenmeyer flasks, respectively. The strains were acclimated to a solution of 3 g L-1 of N:P:K (18:6:18) fertilizer (Gressler et al., 2014), and CO₂ was supplied to the cultures through diaphragm pumps with a flow rate of 0.22 vvm. A solenoid valve was used to control the injection of CO₂ into the system, which was interspersed in periods of 3 min. every 2 hours. Continuous artificial illumination with fluorescent light was applied (2500 lux, 24 hours day^-1), and the temperature was maintained at 25°C. The cell density was monitored daily for each species through spectrophotoscopy at 682 nm, and the results were plotted as calibration curves. Experiments were conducted in triplicate.

Filamentous fungi strains examined in the present study were transferred from the Culture Collection of Microorganisms at Industrial Microbiology from UNISC. A total of 11 strains of filamentous fungi were grown on plates containing potato dextrose agar medium for 7 days at 25°C. The cultures were maintained in slants with Sabouraud agar medium.

The electroflotation method was applied for the biomass separation of D. subspicatus and Chlorella sp. A power supply with direct current (Instrutherm, FA – 3003) was used. The electrode was constructed using 6 iron plates (100 x 60 mm), 3 mm thickness; the distance between the plates was 0.4 mm, and the current density was 1.0 mA cm^-2 as shown in previous work (Baierle et al., 2015). The biomass dry weight was measured for each microalgae species.

Lipids were extracted from the microalgae biomass using the method of Bligh and Dyer (1959) using chloroform: methanol (2:1, v v^-1) as solvents and using ultrasound (Unique USC-1400) for cell disruption, carried out with 25 kHz and 4 cycles of 30 min. After extraction, the solvent was evaporated using a rotary evaporator (Quimis, Q344M), and the residual biomass was dried at 50°C. These residues were stored under refrigeration temperature for subsequent tests of fungal hydrolysis.

The carbohydrate content was determined in microalgae biomass after lipid extraction using the method described by Sluiter (2008). A portion of 300 mg of the residual biomass was weighed and placed into a pressure tube with 3 mL of 72% sulfuric acid. The tube was incubated for 60 min. in a water bath at 30°C, followed by the addition of 84 mL deionized water. The samples were then autoclaved at 121°C for 1 hour.

The dry biomass residue (20 mg) was placed in test tubes, and 2 mL deionized water was added. These mixtures were sterilized through autoclaving at 121°C for 20 min. The spores (approximately 1×10⁸ spores mL^-1) of 11 fungal strains were inoculated, and the tubes were incubated at 30°C for 7 days. The fungal growth on microalgae biomass was monitored daily through visual evaluation and the fungi strains were identified as: (-) without micellar growth, (+) little micellar growth, and (++) high micellar growth.

The fungi mycelium was removed and aliquots of 100 µL of the supernatant from each culture were transferred to 1.5 mL microtubes containing 100 mg potassium bromide, lyophilized for 2 hours at 1.10^-4 Torr (Labcomco) and ground with a pestle. The infrared spectrum was acquired in triplicate using a diffuse reflectance device (Pyke Technologies, Madison) coupled to a Fourier Transform Infrared Spectroscopy (FTIR; Perkin Elmer model 400) at a range of 4000 to 400 cm^-1, with 16 scans and a resolution of 4 cm^-1. The spectral data (intensity of absorbance for a given frequency associated with specific bands) were evaluated to check for the presence of characteristic bands that indicated that the hydrolysis step was effective with the selected fungal strains. The analysis was performed in order to observe the presence of FTIR bands that could reflects the potential of hydrolysis, such as the bands in approximately 3300 (O-H stretching vibrations), 733 (C=O stretching originating from esters or amides), 1100 (C-O-C indicating carbohydrates as saccharides and polysaccharides) and 1600 cm^–1 (NH₂ deformation corresponding to proteins or amino acids).

The normalized average spectra were pre-processed through mean centering, and subsequently analyzed for a hierarchical cluster analysis and principal component analysis using PIROUETTE 3.1 software (Infometrix^TM). Fungi strains were chosen considering the results found through the analysis with FTIR and by the evaluation of the results obtained by PCA and HCA. The PCA was applied in the exploratory analysis to identify similarities and differences between different fungi strains species, after data compression. Hierarchical Cluster Analysis was applied to the auto-scaled data and the Euclidean distances generated a dendrogram for the samples.

For the assays, 1 g of residual biomass of both microalgae and 100 mL deionized water were added to 500 mL Erlenmeyer flasks and sterilized at 121°C for 20 min. The selected fungal strains were grown on Potato Dextrose Agar (PDA) at 25°C during 7 days. A saline solution containing the spores was analyzed on a Neubauer chamber and the inoculum was adjusted to 1×10⁸ spores mL^-1, approximately. The samples were incubated for 7 days at 30°C in an orbital shaker (20 rpm). The hydrolysis reactions were performed in quintuplicate for all fungal strains selected.

Sugar determination in the hydrolysates was conducted using an HPLC equipped with refractive index detector RID-20A, model LC-20AD from Shimadzu (USA). Software LabSolutions (Thermo Scientific, USA) was used for data acquisition and processing. The compounds from the sugar mixture (glucose and xylose) were separated in a Rezex RHM monosaccharide H⁺ column (8% cross-linked sulfonated styrene-divinylbenzene, 300 × 7.8 mm) from Phenomenex. The analysis was performed with isocratic flow using ultrapure water. The total chromatographic run time was 20 min. The optimum flow rate was 0.8 mL min.^-1 and the injection volume was 20 μL. A standard solution at 10 mg mL^-1 of each sugar was subjected to chromatographic analysis and the linearity was evaluated by the coefficient of determination (r²) from the analytical curves containing a mixture of the selected sugars at the levels 0.1; 0.2; 0.3; 0.4 and 0.5 mg mL^-1.

Microalgae cell density was evaluated as a function of days of cultivation, and D. subspicatus and Chlorella sp. presented similar growth, as shown in Figure 1. Chlorella sp. showed a slightly higher cell density (28.7×10⁶ cells mL^-1) compared to Desmodesmus subspicatus (15.8×10⁶ cells mL^-1) and the dry weight obtained was 0.36 and 0.28 g L^-1, respectively. Coefficients of determination were satisfactory since both microalgae presented values of r² > 0.99. The maximum growth was observed on the 10th day of culture for both microalgae.

A visual analysis for the ability of 11 filamentous fungal strains to hydrolyze microalgae biomass in submerged culture was performed. As listed in Table 1, we assessed the development of 11 fungal strains with different rates of mycelial growth, indicating the consumption of the substrate compared with a control sample (B).

After lipid extraction, the lipid content obtained in the separation of biomass through electroflotation was 6.07% for D. subspicatus and 4.72% for Chlorella sp. Mallick, Bagchi, Koley, and Singh (2016), indicate that microalgae lipids can exceed 50-60% of dry cell weight. On the other hand, Silva and Bertucco (2016) describes levels in a range of 8-15%. According to Schneider et al. (2012), the lipid content can range from 7 to 77%, depending on the species and growth conditions. Considering that lipids can be used in biodiesel production (Doan, Sivaloganathan, & Obbard, 2011), we applied the biomass residue to examine the hydrolytic capacity of fungal strains using FTIR.

The total carbohydrate content was obtained considering the amount of microalgae biomass after lipid extraction. As a result, D. subspicatus and Chlorella sp. presented 16.9 and 23.1% of total carbohydrates, respectively.

Figure 1.
Growth curve of Chlorella sp. and Desmodesmus subspicatus in N:P:K (3 g L^-1) medium and CO₂ supply during 12 days.

After evaluating hydrolysis, the biomass of both microalgae was analyzed through infrared spectroscopy. The same profile was obtained for both microalgae, as illustrated in Figure 2.

The vibration frequency depends on the linkage power involved in the functional groups. Most of the absorption bands observed in the hydrolysates reflect the presence of various functional groups in the medium (Giordano et al., 2001). The biomass hydrolysis produced a complex mixture of compounds derived from proteins, lipids, nucleic acids, and carbohydrates, such as carboxylic acids, amino acids and oligosaccharides (Stehfest, Toepel, & Wilhelm, 2005), which are observed as an increasing O-H band area. Despite the difficulty in distinguishing certain bands, characteristic stretches of microalgae composition have been observed in other studies. The corresponding wavelengths and bands found in microalgae under study are reported in Table 2. The spectra recorded for both microalgae in this study were identified using data reported in previous studies.

Table 1.
Selection of microorganisms for microalgal biomass hydrolysis.

Figure 2.
Spectrum profile obtained after hydrolysis of Desmodesmus subspicatus and Chlorella sp.

Table 2.
Characteristics bands revealed in studies on microalgae biomass.

In the present study, similarities between the microalgae compositions were observed after comparing the spectral profiles, highlighting the bands indicative of sugars (saccharides and polysaccharides), proteins and lipids. Indeed, spectroscopy is widely used for microalgae analysis, because this technique facilitates the characterization and quantification of the compounds present in the samples. Considering the main bands obtained through infrared spectroscopy (Table 2), the multivariate analysis was performed.

In this study, the analytical methods used were Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA) (Tan, Aziz, & Aroua, 2013; Liu, Mukherjee, Hawkes, & Wilkinson, 2013). After hydrolysis, the main components of the microalgae D. subspicatus and Chlorella sp. were assessed (Table 3).

The results showed that the first component described 99.22% of the data obtained for D. subspicatus. The analysis of the principal components of Chlorella sp. biomass showed a similar patterns of biomass uptake, where 99.5% of the data can be described using a single component.

PCA score plots were obtained for both microalgae to distinguish common parameters through an assessment of the similar components between samples. In this case, we used PCA analysis (PC1 x PC3) to demonstrate the grouping based on similarities between fungal strains as shown in Figure 3. In Desmodesmus subspicatus, the formation of three major distinct groups on the score graph was observed in Figure 3(1). In Chlorella sp., a better separation was achieved compared to D. subspicatus, with the distinction of two groups, as shown in Figure 3(2). This separation indicated that more fungal strains are useful for hydrolysis, considering that only one strain grouped with the control sample.

Subsequently, the HCA of the D. subspicatus and Chlorella sp. biomass was performed (Figure 4) to check for a profile consistent with the PCA results and determine which fungal strains were most favorable for the hydrolysis step. HCA was run in order to emphasize the natural clusters and similar patterns between the samples. The obtained dendrogram and the distance between the samples were performed to verify the similarity of fungal strains.

Table 3.
Principal components obtained for Desmodesmus subspicatus and Chlorella sp.

Figure 3.
Plots of PCA scores (PC1 x PC3) of (1) Desmodesmus subspicatus and (2) Chlorella sp., showing the best groups (not inserted in B group) of fungal strains to be selected for the hydrolysis step. B (control sample), AF (A. fumigatus USP-174), AG (A. gigantus MI 2012 G12-10), AT (A. terreus 233-3), AN (A. niger USP 9), PG1 (Penicillium MI 2012 G1), PG4 (Penicillium MI 2012 G4 A5), PG7 (Penicillium MI 2012 G7), PG9 (Penicillium MI 2012 G9), PG12 (Penicillium MI G12 A), TA (Trichoderma auriculares).

Figure 4.
Dendrogram generated from the analysis of (A) Desmodesmus subspicatus and (B) Chlorella sp. biomass hydrolyzed with different fungal strains showing two main groups. B (control sample), AF (A. fumigatus USP-174), AG (A. gigantus MI 2012 G12-10), AT (A. terreus 233-3), AN (A. niger USP 9), PG1 (Penicillium MI 2012 G1), PG4 (Penicillium MI 2012 G4 A5), PG7 (Penicillium MI 2012 G7), PG9 (Penicillium MI 2012 G9), PG12 (Penicillium MI G12 A), TA (Trichoderma auriculares).

The HCA analysis was useful to demonstrate that the fungi strains that were grouped together presented similar patterns of biomass uptake. The groups of fungi strains that resemble with the control sample (water and microalgae biomass, without fungi strains) were considered unsuitable to the further tests of hydrolysis. The dendrogram arrangement of Desmodesmus subspicatus in Figure 4(1) showed the formation of two distinct groups: Group 1, comprising the control and other fungal strains considered less favorable to interact with microalgal biomass; and Group 2, comprising only Penicillium strains, grouped according to similarities in the components present in this genus and showing a suitable profile for pretreatment considering the distance to the control sample.

For Chlorella sp., the HCA analysis in Figure 4(2) showed the distinction of two main groups: Group 1, containing most of the fungal strains capable of hydrolyzing biomass; and Group 2, containing only one fungal strain (Penicillium MI 2012 G7) and the control sample which, in this case, would not be useful for hydrolysis.

To demonstrate the ability of the fungi examined and the efficiency of FTIR for screening, we selected two fungal strains from the dendrogram analysis. The Penicillium G12 strain was selected for the hydrolysis of D. subspicatus, and T. auriculares, for Chlorella sp.

The HPLC analysis was successfully applied and results demonstrated that analytical curves presented good linearity with r² ≥ 0.99 for the evaluated sugars. The concentration of glucose and xylose per gram of biomass after lipid extraction and hydrolysis with fungal strains obtained by D. subspicatus was 1.9 and 1.3 mg mL^-1, which represents a yield of 9.4 and 6.6%, respectively. For Chlorella sp., the concentration of glucose and xylose was 2.6 and 1.9 mg mL^-1, which represents a yield of 12.9 and 9.6%, respectively.

The determination of sugars from raw materials demonstrated that pretreatment is crucial to improve the release of sugars, as this treatment effectively degrades complex carbohydrates (Sun & Cheng, 2002; Harun, Danquah, & Forde, 2010). In the present study, pretreatment was not performed, considering that several studies have revealed other alternatives for the release of sugars; for example, the addition of enzymes (Aikawa et al., 2013).

The use of different microorganisms for hydrolysis of biomass has primarily been evaluated at a low cost. The hydrolytic potential of microalgae biomass was identified in a group of fungi examined in the present study. Nevertheless, additional studies using commercial enzymes are required to optimize the hydrolysis conditions of these competitive strains with respect to time and rate of biotransformation.

The development of an easy and rapid method to select hydrolytic fungal strains is the main highlight of our research. FTIR followed by multivariate analysis was successfully applied to screen these microorganisms, distinguishing groups based on hydrolytic capacities.

michelehoeltz@gmail.com