Growth and lipid accumulation by L. starkeyi in SBP hydrolysates
When lignocellulosic hydrolysates are used as substrate for microbial productions, it is necessary to find a compromise between the need to start from a high sugar concentration and the necessity to minimize the toxic effect of inhibitors. With the aim to determine the proper concentration of SBP to be used, L. starkeyi was grown in SBP hydrolysates originated from 1 to 7 % of initial total solid (TS). Poor or no growth was observed at low (1 and 2 % TS) and high (7 % TS) concentration of SBP hydrolysates (Figure 1A); in the case of the lowest TS percentages, sugars quantity is probably not enough to sustain growth, whereas in the highest, lactic and acetic acid are very likely reaching an inhibitory concentration (Figure 1B). The presence of lactic and acetic acids can be mainly ascribable to the initial contamination of the SBP with acid-producing bacteria, as reported by Kühnel and colleagues [22]. 3, 4 and 5 % TS supported yeast growth without apparent impairment due to inhibitors, since the OD reached after 72 hours increases with the increase of TS percentage. This means that, in the considered TS range, the growth correlates with the amount of sugars released by enzymatic hydrolysis (Figure 1A). In 6 % TS, we observed a very long lag phase (Figure 1A), which can be ascribable to the time necessary to cells to rescue from the stress caused by organic acids present at moderately toxic concentrations.
Because of the heterogeneous nature of this raw substrate, the use of diverse SBP stocks can generate hydrolysates with quit different amount of sugars and inhibitors, despite the preparation protocol is standardized. To avoid possible growth delay or inhibition due to this heterogeneity, and consequent discrepancies, we selected 3 % SBP hydrolysate as suitable medium for reproducible L. starkeyi cultivation. The capability of L. starkeyi to metabolize glucose, arabinose, lactic acid and acetic acid present in SBP hydrolysates was determined by analyzing the concentration of these compounds at different time points. As previously observed [23-25], L. starkeyi is able to co-consume glucose and acetic acid, but arabinose and lactic acid were not assimilated throughout the cultivation time (Figure 1C).
Under these conditions, cells accumulated about 19.2 % of their dry weight in intracellular oils, leading to a production of 0.5 g/L of lipids after 144 hours (Table 3). These values are low if compared with the results obtained by cultivating the same strain on other residual substrates [26-28]: very likely this depends on the low C/N ratio of the media, which is not suitable to efficiently redirect the yeast metabolism towards lipid biogenesis.
Growth and lipid accumulation in SBP hydrolysate blended with molasses
To unbalance the C/N ratio, we evaluated the effect of the addition of different concentrations of molasses to 3% SBP hydrolysate on L. starkeyi growth and lipid accumulation. As shown in Figure 2A, the final OD reached by cells directly correlated with the concentration of molasses. Therefore, acetic and lactic acids also contained in molasses (Figure 2E, F) were never reaching inhibitory concentrations in any of the quantity tested. In respect to main metabolites, it can be observed that cells co-consumed glucose and acetic acid from about 24 hours of cultivation and started to also co-consume sucrose after about 48 hours (Figure 2B, C, E). As observed in SBP hydrolysate alone, arabinose and lactic acid were not used by cells as carbon sources (Figure 2D, F). The lag phase of growth was longer in the presence of 6 % molasses compared to the other concentrations (Figure 2A, square symbol); this delay could be caused by the presence of moderately-toxic amount of acetic and lactic acids (Figure 2E, F, square symbol) and/or by the high sucrose concentration that might cause osmotic stress (Figure 2C, square symbol).
Lipid accumulation in intact cells of L. starkeyi grown in SBP hydrolysate added with molasses was monitored over time by Fourier Transform Infrared (FTIR) micropectroscopy. FTIR is an efficient tool for rapidly following up microbial lipid accumulation at different stages of growth [29, 30], since it can analyze intact cells identifying specific molecular groups by their absorption bands. Starting from the spectra obtained sampling the cultivations over time, Figure 3 illustrates the temporal evolution of the CHx stretching band area, between 3050 and 2800 cm−1, and of the ester carbonyl band area, between 1760 and 1730 cm−1, after normalization for the total protein content, given by the amide I band area (Additional file 1: Figure S1). The analysis showed that the addition of molasses at concentrations higher than 1.25 % significantly improved the final intracellular lipid production (Figure 3). In the presence of 6 % molasses, intracellular lipids were the 47.2 % of cellular dry weight and lipid production reached 9.3 g/L after 144 hours (Table 3).
Ammonium sulfate is one of the compounds commonly used to fulfil the deficiency of nitrogen in the molasses [12, 13]. SBP hydrolysate was therefore compared with ammonium sulfate blended with molasses in respect to L. starkeyi growth and lipid accumulation. The growth curves and lipid accumulation in media composed of 6 % molasses and (NH4)2SO4 at three different concentrations (0.5, 1 and 2 g/L) are shown in Figure 4A. Growth was significantly reduced in the presence of 2 g/L compared to 0.5 and 1 g/L of (NH4)2SO4 (Figure 4A), leading to lower values of cell dry weight and lipid production after 144 h of cultivation (Table 2). Interestingly, growth and lipid production were reduced in the presence of ammonium sulfate compared with SBP hydrolysate, which therefore resulted to be superior in supporting growth and lipogenesis in L. starkeyi when blended with molasses.
Growth and lipid accumulation in molasses pulse-fed batch cultures
FTIR data on lipid accumulation in SBP hydrolysate added with molasses showed that lipid accumulation was faster at low concentrations of molasses (1 % and 2 %), or even in its absence, during about the first 48 h of growth. Because of the existence of a correlation between lipid accumulation and cell density that was independent from the concentration of molasses (Additional file 1: Figure S2), the delayed accumulation of lipid observed in the presence of 6 % molasses was a consequence of the slowed growth shown in Figure 2A and discussed above.
The use of pulsed feeding fermentations has the advantage to avoid the possible stress imposed by high concentrations of substrates and to temporally separate cell growth and lipid accumulation, which has been described as a successful method to obtain high lipid production [31, 32]. Considering this, we cultivated cells in SBP hydrolysate with pulse feeding of molasses at different time intervals (0, 24 h and 48 h) and low concentrations (1 %), followed by a pulse feeding at 72 h with a higher molasses concentration (3 %). The application of molasses pulse feeding allowed a better growth of L. starkeyi during the first 72 h of cultivation compared with the batch culture (Figure 5A). After 144 h of growth, cells consumed almost all sucrose, glucose and acetic acid and reached a comparable value of biomass to the batch culture (Figure 4A; Table 2; Additional file 1: Figure S3).
FTIR measurements showed that the pulse-feeding cultivation allowed a higher intracellular lipid accumulation compared to the batch culture throughout all the fermentation (Figure 5B and C). Interestingly, the ratio between the lipid accumulation measured by FTIR and cell density was higher in fed-batch compared with batch cultures (Additional file 1: Figure S2), confirming that molasses pulse feeding is effective in increasing the intracellular lipid content.
The difference in lipid accumulation among batch and fed-batch cultures was not evident when lipid content was measured as the ratio of their weight to dried cell biomass (lipid content %, Table 2), but we cannot exclude that lipid extraction was not completely efficient. In fact, compared to lipid extraction technologies, spectroscopic analysis have the advantage to evaluate lipid accumulation in intact cells avoiding possible lipid loss [30]. Interestingly, the addition of molasses increased not only lipid production but also lipid yield (g/g), and this improvement was more pronounced in pulse-fed compared to batch cultivations (Table 3).
Molasses fed-batch fermentations were also applied using ammonium sulfate as external nitrogen source. Under these conditions, the growth of L. starkeyi was better than in batch cultures and was not affected by the concentration of ammonium sulfate (Figure 4B, Table 2). After 144 h of growth, cell dry weight, lipid production and lipid content were not affected by the different concentrations of ammonium sulfate. These performances were higher than in batch cultures, but lower than in molasses pulsed-fed cultivations with SBP hydrolysate (Table 2).
The lipids produced by L. starkeyi in SBP hydrolysate without and with molasses in batch and fed-bath cultures were transmethylated by alkaline catalysis and the resulting fatty acid methyl esters (FAMEs) were determined by gas chromatography. The main fatty acids produced by L. starkeyi in SBP hydrolysate with or without molasses were oleic (18:1) and palmitic (16:0) acids (Table 3). If compared with the composition of fatty acids of cells grown in SBP hydrolysate alone, the addition of molasses increased the percentage of palmitic acid (16:0) and reduced the percentage of linoleic acid (18:3), overall increasing and reducing the content of monounsaturated and polyunsaturated fatty acids, respectively (Table 3).
If on one hand microbial lipids are emerging as important platforms for the sustainable production of biodiesel, on the other hand immobilized lipases are considered as alternative catalyst of transesterification reactions that are consistent with the development of green processes [33-35]. For this reason, we evaluated the FAMEs profile deriving from lipids produced by L. starkeyi in SBP hydrolysate fed with molasses also using the immobilized lipase Novozym 435. Due to the inhibitory effect of methanol on lipase activity [35, 36], a small amount of methanol (2.5 %) was added to the reaction every 24 h for four times (Additional file 1: Figure S4). Compared to the alkaline catalysis, the enzymatic reaction was less efficient and more time consuming (144 h to reach about the 65% of FAMEs; Additional file 1: Figure S4). Interestingly, the use of lipase reduced the percentage of transesterification of linoleic acid (18:2) and increased that of stearic acid (18:0) (Table 3).