Optimizing the Enzymatic Hydrolysis of Microchloropsis salina Biomass for Single-Cell Oil Production

Abstract

There is an increasing industrial demand for sustainable resources for lipid-based biofuels and platform chemical production. A promising, CO₂-efficient resource is autotrophically cultivated microalgae, either for direct single-cell oil (SCO) production or as a biomass substrate for fermentative SCO production via organisms like yeasts. Regarding the latter, chemical biomass hydrolysis typically results in high sugar yield and high salt concentrations due to the required neutralization prior to fermentation. In contrast, enzymatic hydrolysis is often lacking in mass efficiency. In this study, the enzymatic hydrolysis of both nutrient-replete and lipid-rich autotrophic Microchloropsis salina biomass was optimized, testing different pre-treatments and enzyme activities. Hereby, the protease treatment to weaken the cell wall integrity and the dosing of the Cellic CTec3 was identified to have the highest effect on hydrolysis efficiency. Sugar yields of 63% (nutrient-replete) and almost 100% (lipid-rich) could be achieved. The process was successfully scaled-up in mini bioreactors at a 250 mL scale. The resulting hydrolysate of the lipid-rich biomass was tested as a substrate of the oleaginous yeast Cutaneotrichosporon oleaginosus in a consumption-based acetic acid fed-batch setup. It outperformed both the model substrate and the glucose control, demonstrating the high potential of the hydrolysate as feedstock for yeast oil production. The presented sequential and circular SCO-producing value chain highlights the potential for mass- and space–time-efficient biofuel production, combining the autotrophic cultivation of oleaginous algae with decoupled yeast oil fermentation for the first time.

1. Introduction

In light of imminent climate change and with greenhouse gas emissions in the focus of the general public, the mobility and energy sectors specifically need to identify sustainable alternatives to conventional fossil-based fuels and to reduce carbon dioxide emissions. To this end, biofuels are a renewable and sustainable alternative to fossil-based fuels, aiming for carbon-neutrality or at least reduced greenhouse gas emissions. While technological alternatives to biofuels, such as battery electric drive and hydrogen fuel cells, are becoming increasingly important in the automotive sector, the aviation industry remains dependent on liquid fuels for the foreseeable future. Additionally, the global demand for jet fuel is continuously increasing while the aviation industry is simultaneously aiming for a reduction in carbon dioxide emissions. This certainly highlights the importance of sustainable biofuel supply.

Biofuels can be categorized into four generations, which are distinguished based on the source of biomass and conversion technology applied. The first generation of biofuels is based on biogenic lipids, sugars or proteins produced from edible food crops, rendering them rather unsustainable due to competition with food and feed supply. The second generation is derived from agricultural waste or waste streams of the food industry, while the third generation utilizes microalgae as well as cyanobacteria for, e.g., production of algal biodiesel or jet fuels. The fourth generation involves biofuels produced through genetically modified organisms (GMOs) [1,2]. First-generation biofuels, such as biokerosene and biodiesel, are commonly derived from biogenic oils and fats, like sunflower, linseed or palm oil. Alternatively, waste cooking oil, for instance, is considered a second-generation biofuel [2,3,4,5]. Beyond the use of biogenic oils and lipids as biofuels, they are also considered to be important platform chemicals with various applications in bioenergy, food, feed, detergent, cosmetic, organic synthesis, and pharmaceutical industries.

Biogenic fats and oils, more specifically triacylglycerides (TAGs), are mainly derived from vegetable oils and animal fats. TAGs can be further derivatized via processes such as glycerolysis, hydrolysis, and transesterification to yield products such as diacylglycerides (DAGs), monoacylglycerides (MAGs), fatty acids (FAs), fatty acid esters, and glycerol, leading to a plethora of industrially relevant products [6]. For instance, TAGs can be modified for direct use in the production of fully bio-based epoxy thermosets dependent on the degree of desaturation [7] or for hydrolysis leading to free fatty acids (FFAs) and glycerol. FFAs (e.g., oleic acid) can then be used in biofuel production or further modified to produce, i.e., 10-hydroxy stearic acid (10-HSA) as a sustainable alternative to 12-HSA for food, cosmetic, and pharma applications or for estolide production [8,9,10,11]. Glycerol, especially from enzymatic hydrolysis, can subsequently be converted to allyl alcohol and acrylonitrile, opening new pathways towards the production of bio-based carbon fibers with a performance similar to those of petroleum-derived equivalents [12].

To this end, single-cell oil (SCO) from oleaginous microorganisms (e.g., microalgae or yeast) can be a more sustainable source of TAGs compared to vegetable oils, which are commonly associated with intensive monocultures, increased land use, deforestation, and competition with feed and food production, as well as negative impacts on biodiversity [13,14,15]. Oleaginous microalgae, for example, show a higher and thus beneficial space–time–biomass and lipid yield compared to conventional plant-based oil crops, due to a higher photosynthetic efficiency [16]. A promising candidate for biomass and lipid production in this regard is the oleaginous microalga Microchloropsis salina, formally known as Nannochloropsis salina. It is a unicellular seawater alga exhibiting comparatively high growth rates and a lipid content in excess of 50% of its dry cell weight [17,18]. In addition, this microalga can be grown in seawater, thereby avoiding competition with freshwater resources, and at a relatively high pH, limiting contamination, which improves process stability [19]. Oil accumulation in oleaginous organisms is commonly induced by nutrient limitations, e.g., nitrogen, phosphorus, and sulfur, which are also correlated with growth inhibition. This limitation leads to lipid accumulation in M. salina but also affects the cell wall and biomass composition [20,21]. The utilization of the residual biomass after oil extraction is still an unsolved issue for algae-based biofuels, considering their high demand. Hence, to further improve the overall process and mass efficiency and thus space–time SCO production yields, new process routes to utilize the residual biomass are imperative to generate an economically viable process. In this context, it has already been shown that a hydrolysate of lipid-extracted M. salina biomass could be used as a substrate for ethanol production using Saccharomyces cerevisiae [22]. This highlights the potential of residual algal biomass as feedstock in subsequent bioprocess steps. In the context of SCO production for energy-dense lipid-based biofuels, it would thus be beneficial for the overall SCO yield to include a further SCO-producing value chain.

In that respect, another promising candidate for biofuel and TAG production is the oleaginous yeast Cutaneotrichosporon oleaginosus, which is able to utilize a wide range of different sugars, reaching lipid contents in excess of 80% of the dry cell weight [23,24,25]. In addition, it has been demonstrated that C. oleaginosus can grow on different complex substrates, like lignocellulosic or microalgae hydrolysates, for lipid production [26,27,28,29]. It was recently demonstrated that consumption-based fed-batch fermentation using acetic acid as feed induces lipid accumulation without nutrient limitation, enabling the efficient use of nutrient-rich complex hydrolysates [25,28]. Hence, oleaginous yeasts are very promising candidates for industrial-scale bioprocessing since they are, firstly, able to accumulate a high amount of intracellular lipids and, secondly, as the bioprocess can be scaled to industrial level site-independently. Another advantage is that the intracellular lipids can be easily recovered after enzymatic hydrolysis just by centrifugation without the use of any toxic compounds for lipid extraction [25]. To this end, the combination of carbon fixation through microalgae and harnessing the accumulated biomass as a versatile nutrient source for highly efficient lipid production through oleaginous yeast is a promising step towards sustainable and economically viable production of biofuels.

Therefore, the enzymatic hydrolysis of nutrient-replete as well as lipid-rich M. salina biomass grown under nitrogen limitation was optimized to reach a high saccharification efficiency and lipid recovery for the lipid-rich biomass. The resulting hydrolysate of the lipid-rich biomass was tested as a substrate for the fermentation of C. oleaginosus for additional lipid production, thereby showcasing a two-step SCO production value chain for enhanced biofuel space–time yields.

2. Materials and Methods

2.1. Algae Biomass

The different types of Microchloropsis salina biomass were produced by the Chair of Biochemical Engineering at the Technical University of Munich using open thin-layer cascade photobioreactors at the TUM AlgaeTec Center [20,30]. Microchloropsis salina (SAG 40.85) was cultivated under nitrogen limitation to induce lipid accumulation or nutrient-replete conditions. After cultivation, the biomass was harvested by centrifugation and frozen at −80 °C (nitrogen limitation) or −20 °C (nutrient-replete).

2.2. Enzyme Formulation—Hydrolases

The following enzyme formulations covering the most important key activities for cell wall degradation were tested during biochemical pre-treatment and the following hydrolysis step (

Table 1. Overview of the enzyme formulations used for enzymatic hydrolysis.

Name	Supplier	Activities	Activity Units	Recommended Dosing
Protease from Aspergillus saitoi	Sigma-Aldrich	Protease, glucanase	≥600 U/g	Not available
Cellic CTec3	Novozymes	Cellulase	≥1000 BHU-2-HS/g	1–6% (w/w)
Rohament CEP	AB Enzymes	Cellulase	≥100,000 ECU/g	10–100 g/t
Rohalase GMP	AB Enzymes	Mannanase, ß-glucanase	≥600,000 MNU/g	5–50 g/t
Viscozym L	Novozymes	Arabanase, cellulase, ß-glucanase, hemicellulase, xylanase, pectinase	≥120 FBGU/mL	100–250 mL/t

).

2.3. Experimental Setup for Enzymatic Hydrolysis

The enzymatic hydrolysis was initially optimized at a 5 mL scale using 15 mL tubes incubated in a shaking incubator and later scaled up in a 250 mL DASbox mini bioreactor system (Eppendorf AG, Hamburg, Germany). For each experiment, the required amount of biomass was thawed and suspended in demineralized water to an approx. biomass load of 120 g_DCW/L.

The pre-treatment of the suspended biomass covered (1) physical, (2) chemical, and (3) biochemical methods (Figure 1). (1) For the physical methods, namely different thermal treatments, a concentrated buffer solution was added to the biomass suspension to reach a final concentration of 50 mM NaOAc at pH 5.0, and the sterilization was carried out at 121 °C for 20 min, 121 °C for 60 min, or at 134 °C for 20 min. (2) The chemical pre-treatment was performed by adding 0.5% (v/v) acetic acid, 1% (v/v) acetic acid, 2% (v/v) acetic acid, 0.25% (w/v) citric acid, 0.5% (w/v) citric acid, 1% (w/v) citric acid, or a combination of 1% (v/v) Tween 20 with a concentrated buffer solution to reach a final concentration of 50 mM NaOAc at pH 5.0. Afterwards, the biomass suspension was sterilized at 121 °C for 20 min and for the acid treatments the pH was adjusted to 5.0 with sodium hydroxide after sterilization. (3) For the biochemical treatment, specifically a protease digestion, a concentrated buffer solution that reached a final concentration of 50 mM NaOAc at pH 3.5 as well as protease powder 1% (w/w) were added to the suspended biomass and incubated over night at 37 °C, 120 rpm in a shaking incubator. Thereafter, the pH was changed to 5.0 with sodium hydroxide and the sterilization was performed at 121 °C for 20 min. The same procedure without the addition of the protease was used as a control.

For the enzymatic hydrolysis, 3% (w/w) Cellic CTec3 was added to the pre-treated biomass and the volume was adjusted using sodium acetate buffer (50 mM, pH 5.0).

For the DoE experiments, the biomass was pre-treated using the same procedure as for the biochemical pre-treatment in the screening experiments, with protease concentrations ranging from 0.1% to 5% (w/w). The enzymatic hydrolysis in the DoE experiments was performed by the addition of a mix of different hydrolases, namely Cellic CTec 3 (1–6% (w/w), Rohament CEP (10–100 g/t), Rohalase GMP (5–50 g/t), and Viscozym L (100–250 mL/t) in different ratios (

Table A1. DoE parameters and results for nutrient-replete biomass.

Run	Protease [wt%]	Cellic Ctec3 [wt%]	Viscozym L [mL/t]	Rohament CEP [g/t]	Rohalase GMP [g/t]	Pre-Treatment [% Chemical Hydrolysis]	Enzymatic Hydrolysis [% Chemical Hydrolysis]	Total Sugar [% Chemical Hydrolysis]
1	0.1	3.5	175	55	50	39	10	50
2	0.1	1	175	55	27.5	39	8	48
3	0.1	3.5	175	55	5	39	11	51
4	0.1	3.5	175	10	27.5	39	12	51
5	0.1	3.5	250	55	27.5	39	12	52
6	0.1	3.5	175	100	27.5	39	12	51
7	0.1	3.5	100	55	27.5	39	12	52
8	0.1	6	175	55	27.5	39	13	52
9	2.55	6	175	10	27.5	44	12	56
10	2.55	1	175	55	5	44	7	51
11	2.55	1	100	55	27.5	44	7	51
12	2.55	3.5	175	55	27.5	44	10	55
13	2.55	3.5	175	55	27.5	44	9	53
14	2.55	6	175	55	5	44	11	55
15	2.55	3.5	175	10	50	44	9	54
16	2.55	3.5	100	55	50	44	9	54
17	2.55	3.5	175	10	5	44	10	54
18	2.55	3.5	250	100	27.5	44	10	54
19	2.55	6	100	55	27.5	44	12	56
20	2.55	1	175	55	50	44	7	51
21	2.55	3.5	100	100	27.5	44	9	53
22	2.55	6	175	100	27.5	44	11	56
23	2.55	3.5	175	55	27.5	44	11	55
24	2.55	1	175	100	27.5	44	7	51
25	2.55	3.5	250	10	27.5	44	9	54
26	2.55	1	175	10	27.5	44	7	51
27	2.55	3.5	100	55	5	44	10	54
28	2.55	1	250	55	27.5	44	7	52
29	2.55	3.5	100	10	27.5	44	10	55
30	2.55	3.5	175	55	27.5	44	10	55
31	2.55	3.5	175	55	27.5	44	11	55
32	2.55	3.5	175	55	27.5	44	10	54
33	2.55	3.5	175	100	50	44	10	54
34	2.55	3.5	175	100	5	44	9	54
35	2.55	3.5	250	55	50	44	10	54
36	2.55	3.5	250	55	5	44	10	55
37	2.55	6	250	55	27.5	44	11	56
38	2.55	6	175	55	50	44	12	56
39	5	3.5	175	100	27.5	50	12	62
40	5	3.5	175	55	5	50	13	63
41	5	3.5	175	10	27.5	50	12	62
42	5	3.5	100	55	27.5	50	12	63
43	5	3.5	175	55	50	50	12	62
44	5	3.5	250	55	27.5	50	12	62
45	5	1	175	55	27.5	50	9	59
46	5	6	175	55	27.5	50	13	63

and

Table A2. DoE parameters and results for the lipid-rich biomass.

Run	Protease [wt%]	Cellic Ctec3 [wt%]	Viscozym L [mL/t]	Rohament CEP [g/t]	Rohalase GMP [g/t]	Pre-Treatment [% Chemical Hydrolysis]	Enzymatic Hydrolysis [% Chemical Hydrolysis]	Total Sugar [% Chemical Hydrolysis]	Lipid [% Chemical Hydrolysis]
1	0.1	3.5	175	55	50	46	43	89	85
2	0.1	1	175	55	27.5	46	46	92	90
3	0.1	3.5	175	55	5	46	44	90	87
4	0.1	3.5	175	10	27.5	46	44	91	91
5	0.1	3.5	250	55	27.5	46	44	90	87
6	0.1	3.5	175	100	27.5	46	43	89	95
7	0.1	3.5	100	55	27.5	46	44	90	86
8	0.1	6	175	55	27.5	46	40	86	94
9	2.55	6	175	10	27.5	49	41	89	89
10	2.55	1	175	55	5	49	46	95	95
11	2.55	1	100	55	27.5	49	46	94	89
12	2.55	3.5	175	55	27.5	49	45	94	89
13	2.55	3.5	175	55	27.5	49	45	94	88
14	2.55	6	175	55	5	49	40	88	88
15	2.55	3.5	175	10	50	49	45	93	91
16	2.55	3.5	100	55	50	49	44	93	93
17	2.55	3.5	175	10	5	49	44	93	89
18	2.55	3.5	250	100	27.5	49	44	92	97
19	2.55	6	100	55	27.5	49	40	89	87
20	2.55	1	175	55	50	49	37	86	95
21	2.55	3.5	100	100	27.5	49	43	92	92
22	2.55	6	175	100	27.5	49	41	90	93
23	2.55	3.5	175	55	27.5	49	43	92	93
24	2.55	1	175	100	27.5	49	46	95	94
25	2.55	3.5	250	10	27.5	49	44	93	86
26	2.55	1	175	10	27.5	49	45	93	87
27	2.55	3.5	100	55	5	49	45	94	88
28	2.55	1	250	55	27.5	49	47	95	91
29	2.55	3.5	100	10	27.5	49	44	92	90
30	2.55	3.5	175	55	27.5	49	43	91	90
31	2.55	3.5	175	55	27.5	49	40	89	93
32	2.55	3.5	175	55	27.5	49	45	94	93
33	2.55	3.5	175	100	50	49	45	94	85
34	2.55	3.5	175	100	5	49	46	94	87
35	2.55	3.5	250	55	50	49	45	94	89
36	2.55	3.5	250	55	5	49	44	93	99
37	2.55	6	250	55	27.5	49	41	89	90
38	2.55	6	175	55	50	49	41	90	93
39	5	3.5	175	100	27.5	54	42	96	84
40	5	3.5	175	55	5	54	42	96	91
41	5	3.5	175	10	27.5	54	43	96	85
42	5	3.5	100	55	27.5	54	48	102	99
43	5	3.5	175	55	50	54	44	98	91
44	5	3.5	250	55	27.5	54	44	98	90
45	5	1	175	55	27.5	54	44	97	89
46	5	6	175	55	27.5	54	42	95	87

).

Samples for sugar analysis and lipid extraction were taken before and after pre-treatment, and before and after enzyme addition, as well as after 72 h or over time. All experiments were carried out in biological triplicate and the results are presented as means with standard deviations.

Sugar and lipid yields were calculated using the following formulas:

$$\mathit{\text{sugar}} \mathit{\text{yield}} \left(\mathit{\text{pre}}-\mathit{\text{treatment}}\right)={\displaystyle \frac{c(\mathit{\text{after}} \mathit{\text{pre}}-\mathit{\text{treatment}})}{c(\mathit{\text{chemical}} \mathit{\text{hydrolysis}})}}\times 100$$

$$\mathit{\text{sugar}} \mathit{\text{yield}} \left(\mathit{\text{hydrolysis}}\right)={\displaystyle \frac{c(\mathit{\text{after}} \mathit{\text{hydrolysis}})-c(\mathit{\text{after}} \mathit{\text{enzyme}} \mathit{\text{addition}})}{c(\mathit{\text{chemical}} \mathit{\text{hydrolysis}})}}\times 100$$

$$\mathit{\text{lipid}} \mathit{\text{yield}}={\displaystyle \frac{c(\mathit{\text{after}} \mathit{\text{hydrolysis}})}{c(\mathit{\text{chemical}} \mathit{\text{hydrolysis}})}}\times 100$$

2.4. Response Surface Methodology and Further Statistical Analysis—Design of Experiment

The experimental design and the analysis of the resulting data were performed using Design-Expert Software, Version 22.0.2 (Stat-Ease, Inc. Minneapolis, MN, USA), applying a Box–Behnken model. If available, the dosing range given by the supplier was used for the model. The effect of the concentration of different enzyme formulations covering different activities on the sugar yield was analyzed and validated using analysis of variance (ANOVA).

The resulting model was used to identify two optimized conditions, one for maximized sugar yield within the tested enzyme concentration ranges and one for maximized sugar yield while simultaneously aiming for minimized enzyme usage. The predicted results were subsequently validated in an independent experiment.

2.5. Fermentation

The fermentation of Cutaneotrichosporon oleaginosus ATCC 20509 (DSM-11815) was performed in a 250 mL DASbox mini bioreactor system (Eppendorf AG, Hamburg, Germany) according to Rerop et al. [28]. The pre-culture was prepared in YPD medium (10 g/L yeast extract, 20 g/L peptone, and 20 g/L glucose) incubated at 28 °C and 120 rpm in a shaking incubator for 72 h.

The base medium for fermentation contained 0.9 g/L Na₂HPO₄, 2.4 g/L KH₂PO₄, 2 g/L MgSO₄·7H₂O, 0.5 CaCl₂·2H₂O, 0.00055 mg/L ZnSO₄·7H₂O, 0.024 mg/L MnCl₂·6H₂O, 0.025 mg/L CuSO₄·5H₂O, 0.027 mg/L C₆H₈O₇·Fe·H₃N, 1 g/L urea, 3 g/L peptone, and 2 g/L yeast extract. For the fermentation on algae hydrolysate, the water phase after enzymatic hydrolysis was separated by centrifugation (20,000 rcf for 20 min). All media components were directly dissolved in the hydrolysate, and the resulting fermentation medium was sterile filtered after pH adjustment using sodium hydroxide. A model substrate, based on the sugar and acetic acid content of the algae hydrolysate, namely 20 g/L glucose and 18 g/L CH₃COO·Na, as well as a medium containing 30 g/L glucose and 4.5 g/L CH₃COO Na were used as references.

The fermentation was performed using a starting volume of 150 mL at a constant pH of 6.5, a temperature of 28 °C, and dissolved oxygen level of 50%. The pH was kept constant by the addition of 50% (v/v) acetic acid. Samples for the determination of OD₆₀₀, dry cell weight, sugar content, FAMEs, and lipid titer were taken over time. All experiments were carried out in biological triplicates and data represent mean values with standard deviation.

2.6. Sugar Analysis

The samples from the enzymatic hydrolysis as well as fermentation were centrifuged for 5 min at 16,000 rcf. The supernatant was transferred to 10 kDa spin columns and filtered at 20 °C and 13,000 rcf until the liquid passed the filter completely.

As a reference, 4 mL of the biomass suspension in water was mixed with concentrated sulfuric acid in water to reach a final concentration of 2% (v/v) sulfuric acid. After hydrolysis at 121 °C for 60 min, the samples were neutralized using calcium carbonate. The supernatant was frozen at −20 °C and filtered through a 0.2 µm Phenex-NY syringe filter (Phenomenex, Torrance, CA, USA) after thawing.

Further, 98.5 µL of each sample were mixed with 1.5 µL of a 0.5 M Na₂EDTA solution and analyzed using an Agilent 1260 Infinity II HPLC system equipped with a diode array detector (DAD) and refractive index detector (RID) using a Phenomenex Rezex ROA-organic H+ 8% column with 0.5 mM sulfuric acid as the mobile phase [31]. The run time for the samples from enzymatic hydrolysis as well as fermentation was 40 min, while the run time for the chemical hydrolysis was 60 min.

2.7. Fatty Acid Analysis

The cell disruption during enzymatic hydrolysis was monitored by lipid extraction. Therefore, a sample of the enzymatic hydrolysis was mixed with hexane (1:1 (v/v)) and incubated in a shaking incubator at 20 °C over night. Then, 100 µL of this extract was transferred to glass vials and dried.

The fatty acid profile and lipid content of the fermentation were monitored by weighing in about 2 mg of the biomass used for dry cell weight determination into glass vials.

The lipids released during enzymatic hydrolysis of the lipid-rich biomass were extracted using hexane and quantified by FAME analysis using transesterification in a Multi-Purpose Sampler MPS robotic (Gerstel, Linthicum Heights, MD, USA), followed by GC-FID measurements on a GC-2025 coupled to an AOC-20i Auto Injector and AOC-20s Auto Sampler (Shimadzu, Duisburg, Germany), according to Engelhart-Straub et al. [32]. C12 TAG was used as an internal standard for algae samples, while C19 TAG was used for yeast samples. FAMEs were quantified using a marine oil fatty acid methyl ester (FAME) mix (20 components from C14:0 to C24:1; Restek GmbH, Bad Homburg, Germany).

2.8. Dry Cell Weight

The dry cell weight of the algae biomass suspended in water for hydrolysis experiments was determined by transferring 500 µL of the suspension into pre-weight reaction tubes and drying at 60 °C for 24 h. The dry cell weight was determined in triplicates.

To determine the dry cell weight of the yeast fermentation, samples of 3 mL were transferred in pre-weight tubes, mixed with ethanol (1:1 (v/v)), and centrifuged at 4400 rcf and 20 °C for 45 min. After washing two times with 50% (v/v) ethanol in water, the pellet was frozen in liquid nitrogen and lyophilized. The determination of dry cell weight for each biological replicate was performed in technical duplicates.

2.9. Optical Density at 600 nm

The optical density was measured at 600 nm using a photo spectrometer. If necessary, the samples were diluted in phosphate-buffered saline to measure in the range between 0.1 and 1. Dilutions were performed in technical triplicates for each biological replicate.

2.10. Determination of Lipid Titer

Samples of 7 mL were processed using a French press (EmulsiFlex-B15, AVESTIN Europe GmbH, Mannheim, Germany) for cell disruption, frozen, and lyophilized. Afterwards, the lipids were extracted using a modified Bligh and Dryer protocol with methanol–chloroform (2:1 v/v) and quantified gravimetrically after drying [33,34]. The lipid titer was determined in technical duplicates for each biological triplicate.

3. Results and Discussion

3.1. Enzymatic Hydrolysis

The cell wall of Nannochloropsis mainly consists of an outer algaenan-based layer and an inner cellulose-based layer [35]. Algaenan is a group of non-hydrolysable, mainly aliphatic, biomolecules biosynthesized by several marine microalgae to prevent enzymatic hydrolysis of the cell wall. In the case of Microchloropsis salina, formally known as Nannochloropsis salina, algaenan constitutes about 1–2% of the biomass and consists mostly of C28–C34 linear aliphatic chains linked by ether bonds [36]. Under nitrogen limitation, the biomass composition is subject to significant changes, which are well documented in the literature [20]. Specifically, the lipid and carbohydrate contents increase from 16% (w/w) and 26% (w/w) to 46% (w/w) and 35% (w/w), respectively, while the protein content decreases from 48% (w/w) to just 12% (w/w). Furthermore, the thickness of the cell wall is 1.54-fold higher than that under nutrient-replete conditions [20,21]. This swelling of the cell wall under nitrogen limitation seems to be mainly due to an increase in the cellulose-based inner layer and comes alongside an increase in cell size [37].

Since the properties of the M. salina biomass vary depending on nutrient availability, the optimization of the enzymatic hydrolysis was carried out for both types of biomasses separately, with the lipid-rich biomass grown under nitrogen limitation and the biomass produced under nutrient-replete conditions. For the lipid-rich biomass produced under nitrogen limitation, both sugar yield and lipid extraction were analyzed to quantify hydrolysis efficiency as well as cell disruption, while only sugar yield was used for the nutrient-replete biomass due to its low lipid content.

Thus, the sugar yield after enzymatic hydrolysis was significantly higher for the biomass grown under nitrogen limitation. A total sugar yield of up to 100% relative to a chemical hydrolysis could already be reached in the first trials, while the maximum reached for the nutrient-replete biomass was around 60% (Figure 2). Similar results were previously shown for enhanced biogas productivity for Chlamydomonas reinhardtii, Parachlorella kessleri and Scenedesmus obliquus under nitrogen starvation [38]. This could be attributed to a higher content of lipids and/or starch accumulated in the biomass and thus renders these biomasses as suitable substrates for fermentative biogas production [38]. Further, the metabolic changes during nitrogen limitation, especially regarding the cell size and cell wall composition, are not limited to a higher vulnerability towards shear stress, as reported in the literature, but seem to lead to a better digestibility of the cell wall and higher saccharification efficiency during enzymatic hydrolysis as well [37].

3.1.1. Pre-Treatment

In the first step, different pre-treatment methods covering physical (heat), chemical (acid, detergent), and biochemical (protease) methods and their effect on the efficiency of the enzymatic hydrolysis were investigated using just one cellulase formulation (Cellic CTec3 (3% w/w DCW)) in a defined volume for all samples in the hydrolysis step. As both types of algal biomass were produced in open thin-layer cascade photobioreactors under non-sterile conditions, the biomass might contain environmental contaminants, like bacteria or molds, that could interfere with sugar release used as a measure of hydrolysis efficiency. Therefore, the biomass was sterilized prior to enzymatic hydrolysis for saccharification, e. g., by autoclaving. Alternatively, if autoclaving would have a negative effect on the hydrolysis, antibiotics have to be used in the hydrolysis step to avoid microbial growth [39]. As autoclaving seemed to have a synergistic effect on the enzymatic hydrolysis of Nannochloropsis oceanica and does not limit the application of the hydrolysate in fermentation processes, an autoclaving step of commonly 20 min at 121 °C was performed in all pre-treatment methods [40]. Biomass autoclaved in water was used as a control for the different pre-treatment methods. While there were almost no sugars detectable when just suspending the biomass in water, about 30% of the sugars released by chemical hydrolysis for the nutrient-replete biomass and 25% for the lipid-rich biomass could be mobilized by the autoclaving step. Interestingly, an increase in glucose during the autoclaving step was not reported for the thermal pre-treatment of N. oceanica using algal slurry directly without the addition of water or buffer [40]. The reason for the differences in the sugar release during the autoclaving step comparing the different biomasses might thus be the addition of water before autoclaving.

Temperature is hypothesized to be a main driver of cell degradation of algal biomass for biogas production when comparing thermal treatments and hydro-thermal treatments using steam [41]. To investigate the effect of temperature and incubation time during the autoclaving process, an incubation time of 60 min at 121 °C and a temperature of 134 °C for 20 min during autoclaving were carried out. All conditions showed comparable results in the sugar release during the pre-treatment as well as in the hydrolysis step afterwards (Figure 2).

Besides the temperature and incubation time, the addition of organic acids as well as Tween 20 as a detergent during the sterilization were tested. In contrast to inorganic acids, the use of diluted organic acids, like citric acid, showed synergistic effects on enzymatic hydrolysis and could even be metabolized by oleaginous yeasts [22,25,42,43]. It should be noted that inorganic acids are typically associated with inhibitory effects on microbial growth due to the high salt concentration resulting from the neutralization after hydrolysis, even though their respective sugar release is relatively high. While the acetic acid had no effect on the sugar release during pre-treatment or the enzymatic hydrolysis of the nutrient-replete biomass, citric acid showed an increased sugar release with increasing concentration during the pre-treatment but did not affect the efficiency of the enzymatic hydrolysis. In the case of the lipid-rich biomass, the concentration of sugars even decreased with the concentration and strength of the acid used (Figure 2). The decrease in sugars released during the pre-treatment compared to the other pre-treatments might be due to cell lysis and the correlated dilution of the reaction mixture. In addition, the use of detergents could also increase the efficiency of the enzymatic hydrolysis of algal biomass [44,45,46,47]. This effect might be explained by a better binding of the hydrolytic enzymes to the solid surface [48,49]. Moreover, Tween 20 was demonstrated to increase biomass, lipid, and carotenoid production, as well as glycerol consumption in the oleaginous yeast Rhodotorula glutinis [50]. In this study, the addition of Tween 20 had only a minor effect on sugar release during the pre-treatment and enzymatic hydrolysis of nutrient-replete biomass but had a slightly greater effect on lipid-rich biomass, especially during the pre-treatment step (Figure 2). Unfortunately, the addition of Tween 20 caused the formation of a gel-like matter during hexane-based lipid extraction, which interfered with lipid recovery and quantification during the downstream processing of the algal biomass. This effect is used in production of stable nanoemulsions, e.g., using oil–hexane mixtures in phosphate buffer with the addition of Tween 20 [51].

In addition to the physical and chemical methods, the biochemical method, protease treatment, was tested. For both types of biomasses, this kind of pre-treatment using an acidic protease with glucanase side activity resulted in the highest overall yields in sugar, as well as high oil release during the hydrolysis of the lipid-rich biomass (Figure 2). The protein content in the isolated cell wall fraction of Nannochloropsis sp. is around 6% of the dry mass [35]. Furthermore, the use of protease could weaken the cell wall integrity of Nannochloropsis gaditana, leading to a release of intracellular proteins [52]. in accordance with the literature, pre-treatment using a protease from Aspergillus saitoi showed an increase in sugar release, especially during the pre-treatment step [39]. The enzyme formulation used for the pre-treatment is known to have ß-glucosidase side activity, which can explain the sugar release during the pre-treatment step [53]. In addition, the protease seems to weaken the cell wall integrity and make it more accessible for other hydrolytic enzymes, as the protease treatment of mechanically disrupted M. salina biomass increased the saccharification efficiency by only 1% [39].

3.1.2. Optimization of the Enzyme Mix

As the protease treatment was identified as the most promising candidate for biomass pre-treatment, in the next step, the dosing of the protease as well as the enzyme mix used for further digestion of the carbohydrates was optimized using a design of experiment approach applying a Box–Behnken model. Therefore, a combination of four different enzyme formulations covering key activities involved in cell wall degradation, like cellulase, mannanase, or hemi-cellulase, was carried out using different concentrations of each enzyme formulation in the hydrolysis step (Table 1). Due to the complex composition of the cell wall of algae, a single enzyme type is often insufficient for an efficient hydrolysis [54]. Common activities for the enzymatic hydrolysis of microalgae include cellulase, different hemicellulases like mannanase, galactanase, xylanase, and esterase, pectinase, protease, and lysozyme [21,22,54,55].

Based on the cell wall and carbohydrate composition, two different cellulase formulations (Cellic CTec3, Rohament CEP), one mannanase formulation (Rohalase GMP) and one formulation covering many hemi-cellulase and pectinase activities (Viscozym L), were tested in the dosing range given by the supplier [20,35]. To avoid the degradation of hydrolases by proteases, protease was used as a pre-treatment and inactivated by the autoclaving step afterwards. The total sugar yield during the whole enzymatic hydrolysis process for the nutrient-replete biomass ranged from 48% to 63% of the chemical hydrolysis, while the one for the lipid-rich biomass produced with nitrogen limitation was in the range from 86% to 100%. In general, these sugar yields significantly surpassed previous results reported in the literature [21,22,39]. The sugar release during pre-treatment with different protease concentrations ranging from 0.1 to 5% (w/w) was in a similar range of 40% to 50% for both types of biomasses but slightly lower for the nutrient-replete biomass. The enzymatic hydrolysis afterwards showed a total sugar release of 7% to 13% for nutrient-replete biomass and 37% to 48% for the lipid-rich biomass (Table A1 and Table A2 and Figure A1 and Figure A2). The analysis of the data indicated that the concentration of the protease during pre-treatment had the highest effect on the total sugar yield, followed by the dosing of Cellic CTec3 during hydrolysis. DoE experiments using two different commercially available cellulase mixtures, Cellic CTec 3 and Rohament CEP, revealed no significant influence of Rohament CEP on sugar release when used in concentrations according to the supplier’s information. However, it should be noted that concentrations of Rohament CEP exceeding the supplier’s information by approx. 500 times where shown to have beneficial effects on saccharification [39]. For economic reasons, the supplier’s recommendations were used in this experimental design. Moreover, the addition of pectinase and hemi-cellulase, specifically Viscozyme L, seems to have less effect on the saccharification in general [39,54], while mannanase had an effect on the sugar yield as well as lipid recovery [39,55]. The impact of the pre-treatment was much higher for the nutrient-replete biomass compared to the lipid-rich biomass (Figure 3).

3.1.3. Validation and Scale-Up

The model was used to predict two optimization outcomes for each type of biomass, one aiming to maximize the total sugar yield and the other aiming to maximize the total sugar yield while minimizing the enzyme concentrations (

Table 2. Suggestions of the response surface methodology for the optimization of the protease pre-treatment and the enzymatic hydrolysis afterwards with the aim of maximized sugar yield using enzyme concentrations in the tested range (in range) and while minimizing the enzyme usage (min), giving estimated sugar yields and desirability for each suggestion.

		Suggestions
Biomass	Enzyme Dosing	Protease [wt%]	Cellic CTec3 [wt%]	Viscozym L [ml/t]	Rohament CEP [g/t]	Rohalase GMP [g/t]	Sugar Yield	Desirability
lipid-rich	in range	5	1	175	55	27.5	0.983	0.785
	min	2.996	1.482	137.89	29.366	13.841	0.95	0.67
nutrient-replete	in range	4.981	5.204	243.923	54.884	47.898	0.633	1
	min	3.282	2.352	100	10	5	0.554	0.695

). To validate the results, the suggestions were tested experimentally. In addition, the sugar yield was monitored over time during these trials.

The total sugar yields for the lipid-rich and nutrient-replete biomass fit the model well, with slightly lower sugar yields than predicted but in a similar range. The differences in the total sugar yield between the two suggestions from the model were not significant. In addition, most of the sugar was mobilized in the first 12 to 24 h of hydrolysis after pre-treatment and sterilization (Figure 4a). Similar results were previously reported for lipid-extracted M. salina biomass, where the enzymatic hydrolysis was largely completed after five hours [22].

Based on these results, a scale-up experiment was performed in a 250 mL DASbox mini bioreactor system using minimized enzyme concentrations while maximizing the total sugar yield with a hydrolysis time of 24 h after pre-treatment and sterilization. As expected, the scale-up of the enzymatic hydrolysis from lab scale to the mini bioreactor system worked out well [12]. This experiment showed a comparable performance to the small-scale trials and resulted in a total sugar concentration in the hydrolysate of 7.0 g/L for the nutrient-replete biomass and 19.3 g/L for the lipid-rich biomass. The acetic acid concentration added to the reaction from the sodium acetate buffer used for hydrolysis was between 11.0 g/L for the nutrient-replete biomass and 13.2 g/L for the lipid-rich biomass (Figure 4a).

3.2. Algae Hydrolysate-Based Fed-Batch Fermentation with Consumption-Based Acetic Acid Feed

In the next step, the resulting hydrolysate was used as a substrate for C. oleaginosus for SCO production using a consumption-based acetic acid feed. The fermentation was performed in a 250 mL DASbox mini bioreactor system using the same parameters as recently used for the conversion of lignocellulosic biomass into SCO (28 °C, pH 6.5, 50% DO) [28]. The oleaginous yeast C. oleaginosus is able to utilize a wide range of different sugars [23,24]. In addition, it was demonstrated that C. oleaginosus can grow on different complex hydrolysates while achieving high lipid production [26,27,28,29]. In addition, the ability to utilize acetic acid in a consumption-based fed-batch fermentation enables high lipid productivity without encountering growth reduction due to nutrient limitation renders C. oleaginosus an ideal SCO production candidate when using M. salina hydrolysate as a sole fermentation substrate [25,28]. Commonly, nutrient limitations, like nitrogen or phosphorus limitation, are applied to induce lipid accumulation in oleaginous yeast. While phosphorus can be removed using precipitation, nitrogen limitation cannot be applied to biomass hydrolysates, especially after enzymatic hydrolysis. The consumption-based fed-batch fermentation enables the use of nitrogen-rich hydrolysates without any other treatments like phosphorus precipitation.

Although the majority of the carbon incorporated into biomass and lipids during the fermentation originates from acetic acid rather than from the hydrolysate, the sugars provided by the hydrolysate are particularly critical during the early stages. These sugars play a pivotal role, particularly at the onset of fermentation, where they provide the readily accessible energy and carbon required to initiate robust cell growth, enabling the culture to reach a high cell density and thereby creating the basis for efficient lipid accumulation in later fermentation stages [25].

Due to the low sugar content of the algae hydrolysate coming from the nutrient-replete biomass, the fermentation was just carried out using the hydrolysate from the lipid-rich biomass. After hydrolysis, the residual biomass was separated by centrifugation, and all medium components were dissolved in the hydrolysate directly and sterile filtered. A model substrate, based on the sugar and acetic acid concentrations of the hydrolysate, and a control containing 30 g/L glucose and 4.5 g/L sodium acetate, were used for comparison.

The highest optical density (OD₆₀₀) of about 140 as well as the highest dry cell weight (DCW) of about 76 g/L was obtained with the algae hydrolysate as a substrate, followed by the control with a maximum OD₆₀₀ of about 115 and a dry cell weight of about 49 g/L, while the model substrate with a maximum OD₆₀₀ of about 105 and a dry cell weight of about 37 g/L showed the lowest conversion efficiency (Figure 5a). Moreover, sugars were depleted after 24 h in the algae hydrolysate, while there were still sugars left in the model substrate and the control (Figure 5c). While the lipid content calculated by the quantification of the fatty acids reaching a maximum around 0.75 g/g_DCW was in a comparable range for all three conditions, the lipid titer was maximal for the algae hydrolysate with around 60 g/L, followed by the control around 40 g/L and the model substrate around 30 g/L (Figure 5b). The good growth and high lipid productivity when using the algae hydrolysate came alongside a high acetic acid consumption of 135 ± 7 mL acetic acid (50% v/v), compared to 52 ± 9 mL for the model substrate and 62 ± 6 mL for the standard condition. Considering the initial sugar concentration in the batch phase, the ratios between carbon coming from sugars and carbon coming from the acetic acid feed (C_sugar/C_AA) were 0.004 for the hydrolysate, 0.028 for the model substrate, and 0.108 for the standard condition. The fatty acid profile was comparable for the model substrate and the control. However, for the hydrolysate, it slightly shifted to C18 with a higher degree of desaturation for the hydrolysate (Figure 5d). The shift from C16:0 to C18:0, together with the increased proportion of C18:1 in the fatty acid profile obtained with the hydrolysate, may enhance the properties for biodiesel production. However, the slightly higher content of polyunsaturated fatty acids could, in turn, have a negative impact on these properties [56].

C. oleaginosus was able to grow and accumulate lipids under all tested conditions. The biomass yield of about 40 g/L after 72 h and about 49 g/L after 96 h and especially the lipid titer of about 40 g/L after 96 h of the standard condition used as control (30 g/L glucose and 4.5 g/L sodium acetate) was higher in comparison to the literature [28,34]. Further, the model substrate showed a higher lipid titer while the dry cell weight was in a comparable range [28,34]. The algae hydrolysate showed comparable results, with the same shift towards C18 to the nitrogen-rich conditions in acetic acid fermentation, and outperformed all other reported cultivation conditions [25]. This indicates that the nitrogen content of the hydrolysate coming from the biomass itself as well as the enzyme formulations had an impact on the growth performance and lipid production in C. oleaginosus. In addition, other important nutrients like essential amino acids, trace elements, vitamins, and growth factors coming from the biomass or the enzyme formulations could also have a beneficial effect on growth behavior. Comparing the performance of a lignocellulosic hydrolysate to the standard condition (30 g/L glucose and 4.5 g/L sodium acetate), the increase in productivity was in the same range. These findings highlight M. salina hydrolysate as a highly suitable substrate for SCO production [28].

3.3. Overview of the Overall Process

The optimized overall process consists of algae cultivation, harvesting of the algal biomass by centrifugation, pre-treatment using protease from Aspergillus saitoi, autoclaving, enzymatic hydrolysis using Cellic CTec 3, Viscozym L, Rohament CEP, and Rohalase GMP, lipid extraction, sterile filtration of the water phase after hydrolysis and lipid extraction, and yeast oil fermentation as well as downstream processing (Figure 6). Following this workflow and considering the biomass composition of M. salina under nitrogen limitation, about 37 kg algae oil with a fatty acid profile of about 49% C16:0, 30% C18:0, 7% C18:1, and 6% C20:5, as well as 32 kg fermentable sugars can be mobilized per 100 kg dried algae biomass [20]. The sugar-rich algae hydrolysate can be used as a batch medium in a consumption-based acetic acid fed-batch process using C. oleaginosus to produce yeast oil with a fatty acid profile of about 11% C16:0, 29% C18:0, and 50% C18:1, as well as a final lipid titer of about 60 g/L of the fermentation volume. This effectively demonstrates a technically feasible process for single-cell oil production and direct carbon dioxide capture through the utilization of microalgae M. salina as a carbon source on a laboratory scale. However, the sugar concentration after pre-treatment and enzymatic hydrolysis of about 20 g/L is quite low compared to lignocellulosic hydrolysates, respectively. For further optimization and to render this process economically viable, there are several steps that should be considered. Firstly, further optimization of the hydrolysis on a larger scale, as well as of the biomass load for the pre-treatment and hydrolysis step, might increase the sugar concentration in the hydrolysate, enabling increased carbon intake from microalgae biomass. Secondly, further bioprocess optimization steps like media or water recycling, recovery of the enzymes used for hydrolysis, or optimization of the yeast fermentation towards a higher share of carbon from the hydrolysate, as well as shorter residence time, can improve the economic feasibility of the overall process.

4. Conclusions

Protease treatment was identified to be the most effective pre-treatment method for weakening the cell wall integrity, while the concentration of the cellulase formulation Cellic CTec3 had the greatest impact on sugar release during enzymatic hydrolysis. The complex cell wall composition of M. salina appears to change during lipid accumulation under nitrogen limitation, resulting in a much higher hydrolysis efficiency using the same enzyme activities and resulting in sugar yields during enzymatic hydrolysis of 63% (nutrient-replete) and almost 100% (lipid-rich) for the different types of biomass. Further investigations into cell wall changes along with optimization of the hydrolysis regarding biomass load and retention time could further improve the hydrolysis of both types of biomass.

The resulting microalgae hydrolysate was demonstrated to be a great substrate for single-cell oil production using C. oleaginosus in a consumption-based acetic acid fed-batch fermentation setup. This resulted in a dry cell weight of 76 g/L with the algae hydrolysate as a substrate, followed by the control with a dry cell weight of about 49 g/L, and the model substrate with a dry cell weight of about 37 g/L, with a comparable lipid content of around 0.75 g/g_DCW for all conditions. Further investigation into the composition of the hydrolysate itself could provide valuable insights into what is causing the increase in performance and productivity. Optimization regarding media composition, like nitrogen content or the waiver of yeast extract as well as tryptone/peptone, could significantly increase the economic viability of the entire process chain.