Production of palmitoleic acid by oleaginous yeast Scheffersomyces segobiensis DSM 27193 using systematic dissolved oxygen regulation strategy

Background: Recently, a newly isolated oleaginous yeast strain Scheffersomyces segobiensis DSM 27193 was identied with the capability of accumulating high contents of palmitoleic acid (POA) in its intracellular lipid (13.8%). However, high amounts of ethanol as the main by-product were also produced in this system. To eliminate ethanol formation and to further improve POA production, process optimization focused on dissolved oxygen (DO) regulation was conducted in this study. Results: As a result, agitation control was found to be highly benecial for cell growth and lipid production. When keeping pO 2 >40% through dynamic agitation control, signicant improvements toward lipid (70.12% increase) and POA (1.44 fold) production have been obtained, and the accumulation of ethanol was decreased to 2.3 g/L. While increasing the aeration rate to 2 vvm, no signicant decrease has been found in ethanol accumulation. Further application of a two-stage agitation regulation (600 rpm in the rst 48 h fermentation then adjusted the agitation speed to 1000 rpm) gave 10.23 g/L of lipid (50% increase) with 2.02 g/L of POA (1.15 fold) production. At last, under the best condition of 1000 rpm of agitation and 1 vvm of aeration, no ethanol was detected during the whole fermentation process, while a dry biomass concentration of 44.8 g/L with 13.43 g/L of lipid and 2.93 g/L of POA was achieved. Transcription analysis revealed that ethanol synthetic pathway was down regulated under the condition of high agitation, while the expression of the key enzymes responsible for lipid and POA accumulation were enhanced. Conclusion: This study introduces the oleaginous yeast chassis S. segobiensis DSM 27193 and an effective pO 2 regulation strategy for ecient biological production of POA.


Background
Palmitoleic acid (C16:1△9) (POA), as an important monosaturated fatty acid (MUFA), has attracted great interest due to its high value in medical and chemical applications [1]. It is reported that POA can reduce nonalcoholic fatty liver disease caused by liver in ammation, decrease the progression of atherosclerosis in obese mice, lower the production of in ammatory cytokines and prevent lifestyle diseases such as diabetes and cardiovascular diseases [1][2][3][4]. POA-rich diets have also been shown to be bene cially affects serum lipids/lipoproteins, resulting in the reduction of total and LDL cholesterol [5]. In nature, sea buckthorn and macadamia nut oil contain 20%-40% POA, making them a possible natural source for POA [6]. However, the rare natural occurrence of these plants and their low total lipid content (e.g., 3.5% for the sea buckthorn) limit the commercial development of POA production [1,7].
Currently, a number of oleaginous microorganisms, including microalgae, yeasts and lamentous fungi, have been screened for their capability of accumulating relatively high amounts of POA [6,8]. Among those microorganisms, microalgae and cyanobacterium have attracted much attention for their abundant POA content [9]. For example, Tribonema minus and Oscillatoria sp. can accumulate more than 50% of POA in their intracellular lipids [10,11]. Some other microalgae like Dunaliella primolecta and Nannochloris sp. can also accumulate POA in the range of ~ 8% − 25 % of lipid [8]. However, growth rates of microalgae are relatively low, and the lipid production process requires su cient light and large space, limiting the fermentation process by weather, climate, site, etc. [12]. In addition, lipid extraction of microalgae is very energy-consuming, and gene editing of microalgae is relatively di cult emphasizing the disadvantages of developing microalgae for POA production. In our previous study, the oleaginous yeast Scheffersomyces segobiensis DSM 27193 has been recognized for accumulating high amounts of POA (approximately 16% in its intracellular lipid) [13]. Compared with microalgae, oleaginous yeasts possess the advantages of easy cultivation, fast growth rates resulting in high biomass concentrations, high lipid content, wide substrate spectrum, and strong adaptability in industry [14], which endow oleaginous yeasts great potential in functional MUFA production.
The level of dissolved oxygen (DO) is one of the most important factors affecting the cell growth and lipid accumulation during the fermentation process. Besides, The degree of unsaturation can also be regulated by the variation of DO, due to the oxygen requirements of key enzymes responsible for regulating lipid unsaturation [15]. The requirements of DO concentration seems to be strain-related and appear to be variable among oleaginous yeasts. For example, low DO concentration led to a high nal lipid content in Yarrowia lipolytica, while its cell growth was enhanced at high DO concentration [16]. The same phenomenon has also been observed in Rhodotorula glutinis [17]. In contrast, low lipid yields have been obtained by Apiotrichum curvatum under lower DO level [18]. In addition, the in uence of DO concentration on lipid unsaturation also needs to be investigated for every strain. For example, low DO conditions decreased the lipid unsaturation in Lipomyces starkeyi lipid [15], while it increased the lipid unsaturation in Saccharomyces cerevisiae, since the expression of desaturase was induced under low DO condition [19].
Aeration and agitation are the two common ways for regulating the DO concentration in medium. Higher aeration rates are usually bene cial for cell growth by improving mass transfer of substrate, product and oxygen and enhancing the oxygen availability since oxygen is an electron acceptor [20]. Agitation can also facilitate the transfer of mass and heat in the reactor, but high agitation rate will inevitably cause stronger shear forces, leading to morphological changes and might even damage the cell structure [21,22]. In this study, the effect of different DO concentrations expressed as partial pressure of oxygen (pO 2 ) regulated by agitation and aeration on cell growth, lipid production, POA accumulation and ethanol generation were comprehensively investigated. Furthermore, the transcription analysis indicated that the key enzymes of POA production were signi cantly activated under the best condition of 1000 rpm of agitation and 1 vvm of aeration, which will provide more information and experience for POA, even more MUFAs biological production.

Results And Discussion
Characteristics of S. segobiensis DSM 27193 fermentation using fed-batch culture Fig. 1 displays the cell growth and lipid accumulation of S. segobiensis DSM 27193 using fed-batch culture with glucose as sole carbon source. The fermentation process of general oleaginous yeast was usually be divided into two phases: cell growth phase and lipid accumulation phase [23]. Surprisingly, S. segobiensis DSM 27193 kept the cell growth during whole the fermentation time. At the end of the rst phase, 9 g/L of biomass with only 5.78% of lipid has been obtained. Nitrogen depletion induced the lipid accumulation phase at ~50 h. Finally, 28.6 g/L of dry biomass and 19.2% per dry biomass of lipid has been obtained, corresponding to 5.49 g/L of lipids. Fatty acid pro le analysis revealed that S. segobiensis DSM 27193 contained 13.8% of POA in its intracellular lipid. At present, high content of POA is usually reported in Kluyveromyces polysporus and S. cerevisiae, however, the total lipid content in these non-oleaginous yeasts is much lower than S. segobiensis DSM 27193 [24]. In contrast, oleaginous yeasts like R. glutinis and Trichosporum cutaneum can accumulate higher lipid amounts, but their POA content is much lower, as <5% of POA have been detected in their intracellular lipid [24]. The high POA content in the lipid pro le of S. segobiensis DSM 27193 endows it promising potential in developing POA commercial production.
In addition to intracellular lipids, 6.26 g/L of ethanol has been detected as the main by-product in the fermentation process. As shown in Fig 1, ethanol began to accumulate at ~48 h, when pO 2 decreased to 0%.
The limited oxygen availability might have stimulated the production of ethanol, characterizing this yeast as Crabtree negative [25], contrary to Crabtree positive yeast like S. cerevisiae, which only produces ethanol under the condition of excess carbon source [26]. Ethanol production might lower the carbon ux toward lipid production. Therefore, process optimization strategies should focus on DO regulation to improve lipid and POA production and to decrease ethanol accumulation.
Effect of dissolved oxygen concentration on cell growth and lipid accumulation DO concentration plays a key role in the production of lipid and is also re ected by cell growth during the fermentation process. In the early phase of cultivation, pO 2 of the medium decreased rapidly even to zero and maintained at a low level throughout the exponential phase (data not known), which inhibited cell growth and lipid production, while promoting ethanol formation [13]. Therefore, su cient oxygen supply in the early phase (approximately at 48 h) should be helpful for enhancing cell growth and lipid accumulation.
In this study, three different pO 2 levels of pO 2 >20%, pO 2 >40% and without pO 2 control (constant agitation at 600 rpm) were carried out through dynamic agitation regulation. The initial agitation speeds were set at 600 rpm. Fast cell proliferation led to rapid pO 2 decrease. Within 30 h and 72 h cultivation, pO 2 level decreased to 40% and 20%, respectively. Subsequently, the dynamic agitation regulation was launched. After around 2 days of cultivation, 850 rpm and 750 rpm were necessary to maintain pO 2 >40% and pO 2 >20%, respectively. In terms of cell growth, 33.9% improvement in cell growth has been achieved by maintaining pO 2 >40% compared to no pO 2 control, indicating the limitation of DO in the medium during the experiment without pO 2 control; however, no obvious difference has been found between pO 2 >20% (28.6 g/L) and without pO 2 control (29.55g/L) ( Fig.   2A).  (Table.2). Thereof, POA and C18:1 were the major components constituting 13.8% and 60.1% of the total lipid, respectively. POA ratio could be increased from 13.8% to 16.4% and 19.9% when maintaining pO 2 >20% and pO 2 >40% compared with the without pO 2 control, respectively. In contrast, C18:1 decreased from 60.1% to 54.7% when keeping pO2>40%. We also noted that C16:0, the precursor of POA slightly increased from 12.95% to 16.4% once keeping pO 2 >40%. The increase of DO concentration regulated by agitation might prevent the transformation of C16:0 to C18:0, so that more POA could be accumulated. The phenomenon that variation in pO 2 levels affects the fatty acid pro le has also been observed in Lipomyces starkeyi, Y. lipolytica and Rhodotorula gracilis [15,27,28]. While in these studies, fatty acid desaturases were considered to be oxygen-related enzymes and oxygen-rich conditions re ected relative activity of desaturases and nally led to an increase in the degree of unsaturation of fatty acids. Y L/g (g lipid/g glucose) was estimated by the formula X L / g = Lipid production/glucose consumption Y L/X (g lipid/g biamass) was estimated by the formula X L / X = Lipid production/biomass In terms of ethanol production, high DO concentration lowered ethanol production, as the highest ethanol concentrations of 6.26 g/L, 5.76 g/L and 2.3 g/L have been detected under the conditions of constant agitation at 600 rpm (no pO 2 control), pO 2 >20% and pO 2 >40%, respectively (Fig. 2B). Interestingly, the highest ethanol accumulation was found in the middle period of fermentation process, and the produced ethanol was reassimilated in the later fermentation stage in all experiments. Speculatively, on one hand, the decrease of oxygen demand in the later stage of fermentation led to the DO increase in the fermentation medium, therefore, the prerequisites of DO limitation for ethanol production has been rescinded, hence limiting the further production of ethanol; while on the other hand, the increased oxygen supplied to medium caused an increasing oxidation of the ethanol produced even in the presence of glucose in the fermentation medium. The same ethanol reassimilation phenomenon has also been described in Candida tropicalis and Pachysolen tannophilus [29,30]. Despite the fact that high DO control decreased the ethanol production, 2.3 g/L of ethanol was still produced at 69 h under the condition of pO 2 >40%. Therefore, more considerate DO regulation strategy was required.

Effect of aeration rate on cell growth and POA production
Aeration control is another important DO regulation strategy. Compared with agitation, aeration regulation imposes lower shear stress to the cells and energy consumption. In this study, three different aeration rates of 0.1 vvm, 1 vvm and 2 vvm were tested with a constant agitation speed of 600 rpm. Fig.3A shows the cell growth under different aeration conditions. When aerating with 0.1 vvm, S. segobiensis DSM 27193 grew slowly, and only 8.8 g/L of dry biomass were obtained at the end of fermentation process. In contrast, high aeration invigorated cell growth throughout the fermentation process, and maximal dry biomass concentration of 46.8 g/L was obtained under the condition of 2 vvm, whereas 33 g/L of dry biomass has been reached with 1 vvm aeration.
As expected, higher aeration rates did also increase the lipid accumulation, as 18.91%, 20.05% and 28.26% lipid per cell dry mass was reached when employing aeration rates of 0.1 vvm, 1 vvm and 2 vvm, respectively (Fig.   3C). In addition, lipid accumulation was accelerated when using the highest aeration rate. Lipid composition analysis revealed, that the accumulation of POA and C18:2 was slightly higher at the lowest aeration rate of 0.1 vvm, while the amount of C16:0 and C18:1 decreased from 12.95% to 7.91% and from 60.10% to 52.47%, respectively (Fig. 5A). Low DO was speculated to induce the expression of Δ-9 fatty acid desaturase, which is responsible to form a double bond in saturated fatty acyl-CoA substrates, thereby increase the unsaturation level of lipid, which has been con rmed in yeast S. cerevisiae [19]. However, POA and linoleic acid content was more or less constant regardless of the aeration rate.
In terms of ethanol production, high ethanol levels up to 20 g/L were produced with the lowest aeration rate (0.1 vvm) (Fig.3B), which seems not to be reassimilated in the later fermentation stages. Instead, it stagnated at a high level. This might be due to the low oxygen availability at the lowest aeration rate used. A similar phenomenon is known for P. stipites, where the volumetric oxygen transfer coe cient (KLa) of less than 11.7 h -1 inhibited ethanol assimilation [31]. However, even at higher aeration rates, about 6 g/L (1 vvm) and 7,4 g/L (2 vvm) of ethanol was produced around 120 h fermentation time, indicating DO limitation. Therefore, higher applying aeration rates are seemingly not su cient for decreasing ethanol production. Further combination of both aeration and agitation regulation might not only favor cell growth and lipid production, but might also decrease ethanol production, and was therefore tested in the next experiments.
Two-stage agitation regulation to enhance POA production Considering that higher agitation speed did signi cantly improve cell growth and lipid production without exerting damaging stress to the yeast cells, and the high DO demand in the early fermentation period, two high agitation regulation strategies were applied: constant high agitation speed of 1000 rpm; two-stage agitation regulation with 600 rpm in the early fermentation phase which was switched to 1000 rpm after 48 h to avoid reduction of cell growth. For both strategies 1vvm was selected as the aeration rate, when taking all factors (cell growth, lipid and POA production, ethanol production) into consideration.
As shown in Fig. 4A, high agitation speed had a positive effect on cell growth as expected, as the biomass concentration of 44.8 g/L and 42.1 g/L were obtained under the condition of 1000 rpm and the two-stage agitation regulation, respectively. Additionally, nearly no ethanol was detected throughout the fermentation when using constantly high agitation speed of 1000 rpm (Fig. 4B). In contrast, with the two-stage agitation regulation ~2 g/L of ethanol were produced at 45 h, which was quickly reassimilated after the agitation speed increase. Constant agitation at 1000 rpm could keep pO 2 level always >40%, while in the two-stage agitation regulation, pO 2 was decreased to 10% during the early fermentation period.
In terms of lipid production, 1000 rpm and two-stage agitation regulation increased lipid content to 30.07% and 25.3%, respectively (Fig. 4C). In this case, high pO 2 condition had both positive effects on cell growth and lipid accumulation for S. segobiensis DSM 27193. In addition, POA content was greatly improved from 13.8% to 21.78% with constant high agitation of 1000 rpm compared to the control condition (Fig. 5B), and the nal POA production of 2.93g/L was achieved, corresponding to a 2-fold increase.
The synthesis pathway of POA in S. segobiensis DSM 27193 was displayed in Fig.6A. Transcription analysis of the key enzymes in ethanol production pathway (ALDH1, ADH1, PDC1) and POA synthetic pathway (ACC, DGAT1, DGAT2, OLE1, OLE2) in the early fermentation phase were displayed in Fig.6B. ALDH1, ADH1, PDC1 were strongly induced at 600 rpm compared with high agitation condition. In detail, the transcription levels of ADH1 at 600 rpm was 3.6 times higher than that at 1000 rpm; Also, the transcription of PDC1 under the condition of 600 rpm displayed 3 times higher than that achieved in 1000 rpm, indicating ethanol synthetic pathway was signi cantly activated at low DO condition in the early fermentation phase. Unsurprisingly, the transcription of the key enzymes such as DGAT1 and OLE2 responsible for lipid synthesis and desaturation in S. segobiensis DSM 27193 were increased at higher DO condition in the early fermentation phase. In speci c, the transcription of DGAT1 was up-regulated by 5.4 times and OLE2 was up-regulated by 3.76 times under the condition of 1000 rpm. This may explain why increasing the agitation speed can effectively increase the total lipid production and the proportion of POA.

Conclusion
In this work, S. segobiensis DSM 27193 was evaluated as a candidate for lipid and POA production. The study revealed that DO concentration has a great impact on the cell growth, POA accumulation, as well as the ethanol production in the strain DSM 27193. Applying the optimized conditions, a moderate POA production (2.93 g/L) was achieved. To our knowledge, this is the rst report of POA as the target product in a wild-type yeast strain.
Current advances in yeast engineering for POA production will establish S. segobiensis DSM 27193 as a reliable source for further improvement in POA production.

Strains and media
Scheffersomyces segobiensis (syn. Pichia segobiensis) DSM 27193 used in this study was deposited at DSMZ culture collection in Braunschweig, Germany. The glycerol stocks were stored at −80°C in the in-house culture collection. For plate cultivation, Yeast Malt (YM) Agar plates were used (g/L): yeast extract 3.0, malt extract 3.0, peptone 5.0, glucose 10.0, agar 20.0. For lipid and POA production, a mineral salt medium was used as described by Ines Schulze et al. [13]. 50 g/L of glucose was used initially, from the 2 nd day, glucose was replenished to a maximum concentration of 90 g/L after determining the actual concentration.
Batch fermentation in bioreactor with controlled dissolved oxygen levels A single colony of yeast cells was picked from YM agar plates and then inoculated into 20 mL mineral salt medium in a 200 ml ba ed shake ask as pre-culture I. Pre-culture II was prepared with 200 ml using the same medium in 2 L shake asks and inoculated with pre-culture to obtain an initial OD600 of 0. The initial fermentation condition (control) was set at 600 rpm and 1 vvm without control of pO 2 . The pO 2 value was measured constantly by a probe. The condition of pO 2 >20% and pO 2 >40% was automatically maintained by continuously increasing the agitation if pO 2 dropped below the set-point. In the two-stage strategy, fermentation was started with 600 rpm and the agitation was enhanced to 1000 rpm after the rst 48 h.

Determination of biomass
The cell concentration in the culture medium was determined by the optical density at 600 nm (OD600) with the initial culture medium as blank. All samples were diluted to an optical density of 0.2-0.8 at 600 nm for measurement.
Biomass was analyzed gravimetrically. 1 mL aliquot of the culture broth was transferred into a pre-dried and pre-weighed 1.5 mL reaction tube and centrifuged at 13,000 rpm for 5 min. The supernatant was collected and used for the determination of glucose. The cell pellet was washed with 1 mL saline (0.9% NaCl), dried at 60 °C for 24 h and weighed.

Lipid analysis
The extraction of cellular lipids and determination of fatty acid composition were modi ed according to the method of Qian et al [23].

Quantitative PCR
Total messenger RNA (mRNA) was extracted using Vazyme's FastPure Cell/Tissue Total RNA Isolation Kit. Total mRNA (2 μg) was then treated with TURBO DNase (Ambion) to remove genomic DNA. Total mRNA (2 μg) was then treated with TURBO DNase (Ambion) to remove genomic DNA. cDNA was synthesized from mRNA using Tiangen's QuantiTect Reverse Transcription kit. Primers were designed using Primer Premier 6 and are listed in Table S1.
The ITS gene was used as the reference gene for internal control. Standard curves were generated using serial dilutions of Scheffersomyces segobiensis DSM 27193 genomic DNA as the template. All data points were collected from three biological replicates.