Evaluation of total sugar content in Camelina meal and optimization of enzymatic hydrolysis
The content of water, insoluble components, acetic acid, and sugars in Camelina meal was quantified following acid hydrolysis. Acetic acid and sugar were analyzed also after enzymatic hydrolysis, as no enough data exist in literature to assess the use of Camelina meal as substrate for microorganisms. As shown in Table 1, almost one third (31%) of Camelina meal was composed of sugars. Of these, glucose and fructose accounted for more than two-thirds (w/w) as revealed by high-performance liquid chromatography (HPLC) analysis. Even though acid hydrolysis can be suitable for saccharification, its use is limited by the low final pH, which needs to be neutralized before sugars are added to the cells, and by the release of inhibitory compounds such as furfurals [2,25]. These limitations negatively affect the sustainability of the overall biorefinery process in terms of use and disposal of acid solutions [25]. Therefore, to release monomeric sugars from their polymeric form, an enzymatic instead of an acid hydrolysis was performed under different test conditions (see below). The other main components of Camelina meal, as reviewed by [7], are crude proteins and crude fats, which account for 35.2–46.9% and 4.9–11.9% of total biomass, respectively, as well as micronutrients such as vitamins; whereas the insoluble fraction is composed mainly of lignin and ashes. Based on these data, we hypothesized that Camelina meal could be a suitable substrate for the growth of R. toruloides and carotenoid production.
Enzymatic hydrolysis can be performed under conditions that are generally more compatible with subsequent growth of mesophilic microbial cell factories. Moreover, it can take advantage of a broad range of commercially available enzymatic cocktails [26,27]. Here, this step was optimized by using the commercial cocktail NS22119 (Novozymes A/S), which can release both hexose and pentose sugars. Different initial concentrations of Camelina meal were tested to determine the effect of solids loading on sugar release. After autoclaving, the measured pH was 5.5, which was compatible with enzymatic catalysis, as NS22119 is supposed to retain up to 90% of its maximum activity at this pH, according to the indications of the producer. Remarkably, the pH remained constant until the end of hydrolysis, which reduced both the economic and environmental impact of the procedure, as neither neutralization or additional buffer were required. As shown in Figure 1A, pre-treatment of biomass by autoclaving resulted in the initial concentration of released sugars to range from 1.8 ± 0.03 g/L to 9 ± 0.3 g/L. The values reflected the amount of biomass loaded at the beginning of the experiment. After enzymatic treatment (11.9% w/wCamelina meal), the concentration of free sugars rose to at least double the initial amount, independently of the quantity of loaded biomass (Figure 1A). No additional release of sugar was detectable over time from negative control samples, in which 3% or 15% of the initial biomass but no enzyme was incubated in a shaking water bath at 50°C (Additional File 1, Figure S1). The sugar titer increased in the presence of enzymes in a linear manner (R2 = 0.98, p < 0,001, calculated with R) until 24 h in respect to the initial quantity of biomass. Hence, the yield of sugars released by enzymatic hydrolysis was constant regardless the concentration of Camelina meal (Figure 1B) and the maximum yield of sugars over total biomass was 20% after 24 h. Considering the original amount of carbohydrates, a sugars recovery of 65% was calculated (see Calculations section), which is in accordance with commonly reported values for lignocellulose enzymatic hydrolysis [28,29].
Based on these data, successive experiments were performed using Camelina meal at the maximum tested solids loading (15% w/v). To determine if sugar recovery could be further improved in spite of a possible inhibition of enzymatic activity by released products or by biomass itself, two different strategies were designed. In one, the initial quantity of enzymes was doubled from 11.9% to 23.8% w/wCamelina meal; in the other, the mixture was pulsed with a second dose of enzymes (11.9% w/wCamelina meal), thus doubling the total amount, after 6 h of hydrolysis. When the first strategy using double the amount of enzymes was applied, the quantity of sugars released from Camelina meal (Figure 2, black bars) did not differ significantly from that of a single enzyme dose (Figure 1A). A similar result was obtained when the second strategy, based on an additional pulse of enzymatic cocktail after 6 h of hydrolysis was applied (Figure 2, white bars). These findings indicate that incomplete saccharification is related more to the intrinsic accessibility of polysaccharides in the biomass than to limitations in catalytic activity. They also suggest that the initial procedure was the preferred one, as it minimized the use of enzymes and thus the overall cost of the process. The enzymatic cocktail exhibited greatest activity during the first hours of hydrolysis; prolonging incubation beyond 6 h to 24 h improved sugar titer by only 20%. Therefore, the conditions for enzymatic hydrolysis of Camelina meal were as follows: 15% w/w solids loading, 11.9% w/wCamelina meal of enzymatic cocktail NS22119, reaction time of 6 h, operative temperature of 50°C, and initial pH of 5.5. As underlined before, the pH of the reaction mixture remained constant over time and was conveniently closer to the optimum reported for carotenoids accumulation in R. toruloides (pH 5) than to the value suitable for lipid production (pH 4) [30].
The above settings allowed for about 25 g/L of monomeric sugars to be released and, with a sugar recovery of 53.3%. The fraction of residual non-hydrolyzed carbohydrates could be considered as an added value to the final product, as a feed with Camelina meal enriched in carotenoids by fermentation of R. toruloides would still contain fibers of nutritional value.
Inhibitory compounds in Camelina meal hydrolysate
Compared with traditional acid treatment, enzymatic hydrolysis is efficacious in releasing sugars from lignocelluloses and minimizing the accumulation of inhibitory compounds [31,32]. Nevertheless, there are some drawbacks related to other compounds detached from these complex matrices [2,20]. Acetic acid is the most common inhibitor released by hydrolysis of the hemicellulose fraction composing lignocellulosic biomasses. Acetic acid can easily impair microbial growth and metabolism due to its generic and specific toxicity [33], reducing the key performance indicators of the production process [2,33,34]. Nevertheless, the toxicity of acetic acid is greater at low pH; extracellular pH values higher than its pKa (4.76) reduce its diffusion across the membrane and, therefore, the cellular damage it could trigger [35,36]. In the present study, the operative pH (5.5) was higher than the pKa of acetic acid, thus lowering the detrimental effect of this molecule on cells. Moreover, R. toruloides has been shown to withstand acetic acid when added to defined media or even as the sole carbon source at up to 20 g/L at pH 6 [37–39]. During enzymatic hydrolysis, acetic acid titer increased , reaching 1.8 ± 0.01 g/L after 24 h from the start (Additional File 1, Figure S2). This amount has been described as bearable by diverse yeasts [34,37,38] including R. toruloides.
Considering the above constrains, Camelina meal hydrolysate appears to be a suitable feedstock for yeast cell factory-based biorefineries, which would enable the exploitation of yeast biodiversity and engineering strategies to obtain different products of interest.
Carotenoids production from Camelina meal hydrolysate in SHF and SSF processes
Having established a protocol for obtaining Camelina meal hydrolysate, carotenoids production by SHF was investigated. In the SHF experiment, medium consisted of clarified supernatant collected after 6 h of enzymatic hydrolysis. This medium was sufficient to sustain R. toruloides growth, as indicated by the accumulation of biomass and consumption of sugar (Figure 3A, dotted and dashed lines, respectively). The accumulation of carotenoids increased over time, reaching 5 ± 0.7 mg/L after 96 h of fermentation (Figure 3A, white bars), with a yield on consumed sugars of 0.034% (w/w) and on maximum quantity of sugars per biomass of 0.011% (w/w). These data are in accordance with previous reports that R. toruloides and carotenogenic microorganisms in general produce carotenoids mostly in response to stressful or sub-optimal conditions such as stationary phase [17,18,40,41]. The period of 96 h was chosen mainly to allow comparison with previous studies, whereby R. toruloides was provided with defined media or other/different residual biomasses (Table 2). After 96 h, fewer carotenoids could be extracted from the cells (Additional File 1, Figure S3), which could correlate with the export/release of carotenoids from the cells or with an imbalance of nutrients that might promote their consumption/corruption. The carotenoid production achieved here by SHF was competitive with R. toruloides grown in shake flasks and supplemented with other complex matrices (Table 2).
To overcome the need for clarifying the medium after enzymatic hydrolysis, we fed the cells the entire Camelina meal hydrolysate, including the water-insoluble solids (WIS) fraction left over after enzymatic hydrolysis. WIS may impair microbial growth and production because of the uneven homogenization of the liquid medium caused by the presence of solid components, as well as due to the toxicity of some of their components [28]. Under conditions termed here as “SSF + presaccharification”, R. toruloides was able not only of consuming sugars and producing carotenoids (Figure 3B), but it also achieved a higher titer of intracellular carotenoids, reaching 13 ± 2.6 mg/L after 96 h, with a yield on consumed sugars of 0.108% (w/w) and on maximum quantity of sugars per biomass of 0.028% (w/w). Given that the amount of carotenoids extracted from Camelina meal with and without the addition of enzymes remained constant over time, the carotenoids measured in this and in the following experiments in the presence of WIS were due to microbial metabolism (Additional File 1, Figure S4).
Often proposed as an alternative to SHF, SSF is characterized by a single combined hydrolysis and microbial fermentation step. The two processes have several pros and cons in terms of efficiency, time, presence/release of inhibitory molecules, and downstream final product [42,43]. SHF and SSF have been proposed and compared for several 2nd generation biorefineries that used Arundo donax, grass or wheat straw as feedstocks [28,44,45]. A potential drawback of incubating enzymes and cells in the same environment is the compromise that needs to be reached allowing optimum operating conditions for both of them. In the present study, because 50°C was not a viable temperature for R. toruloides, 30°C was selected as the operative temperature. Thus, increased shaking was intended to partially compensate for the reduced activity by augmenting the probability of interactions between the matrix and the enzymes. Remarkably, the release of sugars in these conditions was comparable to that obtained by SHF or SSF + presaccharification (Additional File 1, Figure S5). As shown in Figure 3C, after the first 6 h of hydrolysis, the amount of sugars in the medium was lower compared to that obtained by SHF, most likely due to the initial growth (and therefore sugar consumption) of R. toruloides. After 24 h, sugar consumption became clearly evident and was accompanied by the accumulation of carotenoids. After 96 h, the carotenoid concentration reached 16 ± 1.9 mg/L, with a maximum amount of sugars per biomass of 0.028% (w/w). In the case of SSF, it is not possible to measure the total sugar released during saccharification because of simultaneous fermentation. Importantly, the amount of carotenoids was significantly higher when WIS were left in the medium (SSF and SSF + presaccharification) compared to simple SHF (t-test p < 0.05). While sub-lethal concentrations of insoluble solids might impair microbial growth, they could also trigger the accumulation of metabolites important for the microalgae and yeasts own defense systems [41,46]. For example, β-phenol was shown to trigger carotenoid production in yeast [47].
The titers achieved by SSF indicate the efficacy of concurrent hydrolysis and fermentation, suggesting that a simplified procedure involving a single vessel could be used. Because productivity remained similar over time, the initial sugar released in the presence of cells did not seem to speed up the overall process. Overall, data from SSF and SSF + presaccharification reveal that the often mandatory detoxification step, indicated also for R. toruloides [48], is avoidable with this type of residual biomass. Moreover, the final product obtained by both SSF and SSF + presaccharification is a Camelina meal enriched with carotenoids, which can be used directly in the animal feed industry.
Therefore, different products, such as pure carotenoids and carotenoids-enriched Camelina meal can be recovered from the tested processes. Camelina meal in particular, would be an innovative product on the market, as carotenoids are commonly added to animal feed for nutritional and organoleptic reasons [3,12]. In addition, the production of carotenoids from a residual biomass of lower value may increase the economic attractiveness of the proposed process. Based on the logic of cascading [49,50], the present work paves the way for the use of Camelina meal as an alternative feedstock in 2nd generation biorefineries exploiting microbial cell factories to produce fine chemicals.