Upcycling waste organic acids and nitrogen into single cell protein via brewer's yeast

,


Introduction
The population on earth is anticipated to reach 9.7 billion in 2050 (United Nations, 2015).Human diets and agricultural practices barely keep up with the population boom in the next 30 years, requiring double the amount of food production (Bahar et al., 2020).In order to coexist with nature, a global agricultural revolution and diet transition are underway.Protein is an essential part of the human diet that traditional agriculture provides in animal protein (i.e., livestock, fish, dairy) and plant protein (i.e., beans).However, both require massive arable land and water, emit enormous greenhouse gases during cultivation, and show extremely low substrate-protein mass conversion ratio (less than 5%) (Pimentel and Pimentel, 2003;Ravindra, 2000).Cellular agriculture is an emerging approach that enables a large decrease in the environmental impact of food production with the aid of microorganisms.Single cell protein (SCP) is a protein product of cellular agriculture harvested from fungi, algae, or bacteria whose dry biomass contains high levels of protein.The advantages of SCP over traditional food proteins are the short harvest intervals, independence from seasons or climates, and the ability to use cost-effective substrates.With a revenue of $5.3 billion in 2017 and forecast to reach $8.7 billion by 2023, the worldwide SCP market is flourishing without a doubt (Prescient and Strategic Intelligence, 2018).
Yeast is an ambitious SCP candidate because consumers have widely accepted it in a long history as animal and human food addictive.In addition to rich protein (30-50%) and wide amino acid spectrum (Jach and Serefko, 2018), SCP from yeast contains relatively less nucleic acid (5-12%) than bacteria-derived SCP (8-14%) (Jach and Serefko, 2018), which diminishes health threat and curtails the cost of downstream treatment.Moreover, yeast is rich in vitamin group B (Gervasi et al., 2018).It is also possible to use its whole-cell as probiotics and its extract as immunostimulant in animal feed (Hoseinifar et al., 2018;Jach and Serefko, 2018;Mohan et al., 2019).The benefits above will add more value to the SCP product from yeast.The yeast species Saccharomyces cerevisiae has been widely studied for different applications, including SCP production (Jones et al., 2020;Molitor et al., 2019).
Under the principle of circular economy, commercial SCP products have to reduce environmental impact and the dependence on natural resources.Many attempts have been made to use waste or wastewater as feed substrate since raw materials account for 62% of the total SCP production cost (Stanbury et al., 2016).Given the extraordinary ability of sugar fermentation, food waste, or agriculture residues, such as, sugar beet bagasse, fruit waste, etc. could be placed as the first choice for yeast SCP synthesis (Table A4).However, the process of manufacturing SCP from food and agricultural waste involves a series of pretreatments, which add cost to the process (Table A4).Examples include shredding and filtering to remove solids, followed by heat treatment, acid hydrolysis, or enzyme hydrolysis that converts pulp into soluble reducing sugars.Besides that, sugar-based organics could not be entirely converted into SCP, the fermentation broth still requires downstream treatment.These obstacles need to be resolved to make the overall process efficient and cost-effective.Many waste streams, on the other hand, are rich in organic acids, including short-chain volatile fatty acids and long-chain fatty acids.Still, the capability of S. cerevisiae in utilizing these organic acids and the protein content and biomass composition is far from fully understood.Furthermore, given the complex components in wastewater, the effect of crucial water components such as nitrogen and anionic species on SCP synthesis by S. cerevisiae should be elucidated.In addition, it is desirable to address available strategies to adapt raw wastewater as a suitable substrate for S. cerevisiae and thus maximize SCP production.
Therefore, this study first investigated the capability of S. cerevisiae for maximal SCP production using acetate-ammonia substrate along with pH regulation.The potential of using various organic acids for S. cerevisiae to produce SCP was then explored.Subsequently, the influence of inorganic anions and cations on SCP production was examined.Furthermore, the amino acid profile was revealed to evaluate the quality of SCP.In brief, the study expanded the substrate spectrum for S. cerevisiae, and managed to maximize the SCP production at a high organic load with an acidic pH control strategy.It helps to fill the knowledge gap in yeast-driving waste-to-protein conversion process, and to bring high impacts and positive implications for waste valorization and future feed and food supply.

Growth medium and cultivation
Saccharomyces cerevisiae (DSM 70424), from Leibniz Institute DSMZ-German collection of microorganisms and cell cultures GmbH, was routinely cultivated in yeast extract-peptone-dextrose (YPD) broth (ATCC medium 1245) at 25 • C (Table A1).Cell passaging was performed by inoculating 0.5 mL culture from the exponential phase into a 50 mL fresh medium.YPD agar plate was also used to periodically check the purity of S. cerevisiae and preserved at 4 • C in case of rejuvenation.
The acetate-ammonia-based (AAB) broth was tested to domesticate S. cerevisiae (Table A1).A polypropylene film was used to cover each conical flask, with a round membrane (0.22 μm pore size) in the middle, allowing air exchange and preventing contamination.The medium was adjusted to an initial pH of 5.6 by 3 M sodium hydroxide (HCl) before autoclaving at 121 • C for 20 min.MgSO 4 and vitamin solutions were added through a mixed cellulose esters membrane filter (0.45 μm).
S. cerevisiae was precultured in 50 mL AAB mediums for two generations ahead of all batch experiments.1.5 mL sample was daily taken to measure optical density (OD) at 600 nm and pH.Microscopy ( × 400 magnification) was used to examine culture contamination before subculturing (Fig. A1).A typical growth curve of S. cerevisiae in both YPD and AAB medium is shown in Fig. A2.

The effect of substrate concentration and pH regulation on S. cerevisiae growth
Different concentrations of sodium acetate were examined to see the substrate uptake capability of S. cerevisiae within a typical cell cycle.The baseline of acetate concentration (86 mM, 6 g-COD/L) was set according to the average concentration found in waste streams such as anaerobic digestion effluent (Jin et al., 2017).Acetate concentrations of 2.6, 6, 8, 10, and 20 g-COD/L were employed in 50 mL broth.Other compositions remained the same as the AAB medium.The initial pH was adjusted to 5.6.Each batch was inoculated with 2 mL parent AAB culture from the exponential phase to have similar initial ODs of 0.03 and then incubated for 6 days.1.5 mL broth was daily taken for OD and pH monitoring.Biomass and protein concentrations were tested on the last day of incubation.This sampling method also applies to Section 2.3-2.4.
In order to avoid high pH inhibition on S. cerevisiae, acetic acid was used to replace sodium acetate as organic carbon, with the same COD concentration at 2.6 g/L.Another attempt was to use phosphate buffer (KH 2 PO 4 /K 2 HPO 4 , molar concentration ratio: 17) or acetate buffer (NaAc/HAc, molar concentration ratio: 9) in AAB medium.KH 2 PO 4 and K 2 HPO 4 were 3.34 and 0.25 g/L, respectively, to keep the same phosphorus concentration as AAB medium.In acetate buffer, 2.98 g/L NaAc and 0.22 g/L HAc were used to achieve the same acetate concentration as AAB medium.The initial pHs were all adjusted to 5.6 by 3 M HCl.To see the maximal SCP production of S. cerevisiae, at 20 g-COD/L of acetate, 3 M HCl was manually added to bring the pH down from 8 to around 5.6.

Short-chain VFAs
Propionate, butyrate and lactate, which are commonly found in waste streams such as anaerobic digestion effluent, were both investigated to identify whether they can be used by S. cerevisiae besides acetate.Experiments ran for 6 days and were divided into two groups, i.e., using sole organic carbon or using hybrid organic carbons.In the sole organic carbon group, sodium propionate, sodium butyrate and sodium lactate were used as sole organic in broth instead of sodium acetate, at the same COD concentration of 2.6 g/L.In the hybrid organic carbon group, propionate, butyrate and lactate were respectively used in broth as a co-carbon source at 1 g-COD/L, along with 2.6 g-COD/L of acetate.

Long-chain VFAs
Oleate and linoleate are unsaturated long-chain fatty acids (C18), commonly detected in the treated effluent from anaerobic digestion of lipid-rich waste such as food waste (Chan et al., 2018).So the possibility for S. cerevisiae to utilize long-chain fatty acids was investigated.Oleate and linoleate were first separately added as sole carbon source at 2.6 g-COD/L.They were subsequently tested as hybrid organic carbons along with 2.6 g-COD/L acetate at 1 g-COD/L oleate, or linoleate.Other To examine the K + necessity for S. cerevisiae, K + was first excluded by replacing 26 mM Na 2 HPO 4 with the same molar concentration of K 2 HPO 4 ('Exclude K + (liquor) batch').Given the trace residual of K + in the liquor inoculum, a single colony of S. cerevisiae from agar plate was used to inoculate in a fresh medium, named as 'Exclude K + (agar) batch', which completely removed K + from mediums.
NH 4 Cl is the nitrogen source of S. cerevisiae in AAB medium.Different concentrations of ammonia nitrogen (NH 4 -N), i.e., 0.26, 0.6, 0.8, 1, and 1.2 g N/L, were tested to investigate their impact on SCP productions.Other components remained the same as the AAB medium.

Anions
Nitrate and nitrite nitrogen (NO 3 -N, NO 2 -N) were examined as alternative nitrogen sources.In the AAB medium, NH 4 -N was replaced with 18.69 mM NO 3 -N (1.6 g/L NaNO 3 ) and NO 2 -N (1.3 g/L NaNO 2 ), respectively.Different concentrations of Cl − , NO 3 − and SO 4 2− were investigated to examine the adaptability of S. cerevisiae in mediums that will possibly be derived from wastewater with fluctuant anion concentrations.The concentration of KCl, K 2 SO 4, and KNO 3 added in the AAB medium was 20, 80, 250, and 500 mM, respectively.Other components of the medium remained unchanged.

Modified Gompertz model
The OD of S. cerevisiae was daily measured at 600 nm absorbance by spectrophotometer (Varian-Cary 50, Varian, USA) during cultivation.The modified Gompertz model (Equation ( 1)) was applied to fit the OD-Time curve, and estimate the maximal OD (OD max ), maximal growth rate (μ max ), and lag phase (λ) for each batch (Li et al., 2012;Zwietering et al., 1990).
Where OD is the daily OD (obs), t is the specific sampling time (day).
OD max , μ max and λ can be solved based on the mathematical fitting result.

Biomass composition test
At the end of incubation, 40 mL broth was taken to measure dry cell weight.The broth was first centrifuged at 10000 rpm for 10 min, and the supernatant was discarded.The retained biomass was washed twice with deionized (DI) water to remove dissolved inorganic salts from biomass.The biomass of 40 mL broth was transferred into a 2 mL clean tube with a known weight.The biomass was lyophilized in a freeze dryer (CoolSafe, ScanVac, Denmark) for 24 h and weighed the dry biomass.10 mg of the dehydrated biomass was further used to determine the amino acid profile.The sample was first hydrolyzed in 500 μL 6N HCl with the assistance of a microwave digestion system (3000 SOLV, Anton-Paar®, Austria).The hydrolysis lasted for 30 min at 130 • C. HPLC-MS-MS ( 1290Infinity II-6470 QQQ, Agilent Technologies, USA) was then applied to analyze the amino acids for the hydrolysate (Yang et al., 2021).The rest dry biomass was placed in a muffle furnace at 600 • C for 2 h.The weight of residual ash divided by dry biomass weight is the ash content in biomass.
The biomass of 2 mL broth was resuspended by 2 mL 1 M NaOH, followed by 10 min boiling in a glass tube.Both alkali treatment and boiling of biomass aim to dissolve protein into the liquor well.The protein concentration of the liquor was measured according to Lowry protein assay (Lowry et al., 1951) at 550 nm.The standard protein solution was prepared by bovine serum albumin, using 1 M NaOH instead of DI water for gradient dilution.
The biomass of 400 mL broth was resuspended in 20 mL DI water for cell disruption in a microwave (Lee et al., 2010).The microwave lasted for 1 min at 2450 MHz but stopped every 10 s, at which point the tube was immediately inserted into ice to avoid explosive boiling.The 20 mL biomass was transferred into a thin tube and ready for lipid extraction and measurement by a modified Bligh-Dyer Method (Bligh and Dyer, 1959).Briefly, 20 mL chloroform and 40 mL methanol were first added to 20 mL biomass and vigorously blended for 2 min.Another 20 mL chloroform was then added and blended for 30 s.The homogenate was filtered through a filter paper and settled in a measuring cylinder.It would separate into two layers overnight, with the chloroform layer at the bottom and the methanol layer on top.The target chloroform layer  with lipid was transferred into a clean tube and dried at 60 • C until constant weight.The remnant was the lipid in biomass.The remnant weight (g) divided by biomass volume (mL) is lipid concentration (g/mL). 1 mL broth was taken for carbohydrate measurement by the Phenol-Sulfuric acid Method (Dubois et al., 1956).In short, 1 mL 5% phenol, followed by 5 mL H 2 SO 4 , was added to a 1 mL liquor sample and cooled down to room temperature.The absorbance was tested at 490 nm.Glucose was used to make the standard curve for carbohydrates.

Nitrogen and carbon balance measurement and calculation
NH 4 Cl was the only nitrogen source in the AAB medium for S. cerevisiae growth.At the end of incubation, the nitrogen was mainly distributed into two parts, i.e., residual NH 4 -N in broth and SCP in biomass.The initial and residual NH 4 -N in broth was tested by a commercial kit (Ammonium Test Kit 100683, Merck, Germany).The nitrogen content of protein was estimated by protein concentrations multiplied by 16%.
The carbon influx was derived from acetate as the carbon source of the substrate, and carbon efflux included microbial assimilation into biomass and CO 2 release out of the broth.To capture CO 2 , sealed serum bottles were used instead to cultivate S. cerevisiae.At the beginning and end of the incubation (Day 1, 6), 200 μL headspace gas was extracted to measure CO 2 and O 2 contents by a gas chromatograph (Trace 1310 GC-TCD, Thermo Fisher, Denmark).The GC was equipped with Trace PLOT TG-BOND Q column with helium as carrier gas.Element carbon was determined by an elemental analyzer (Vario MACRO cube, Elementar, Germany).Carbon distributions in biomass and gas were calculated by respective carbon weight divided by the total carbon consumption of acetate.
VFA concentrations including acetic acid, propionic acid, and butyric acid were tested by gas chromatography (GC, TRACE 1300, Thermo Scientific, US.) equipped with a flame ionization detector and a DB-FFAP fused silica capillary column.Lactic acid was measured by highperformance liquid chromatography (HPLC, Agilent, US.) with a refractive index detector and Bio-Rad Aminex HPX-87H (300 × 7.8 mm) column.

The impact of acetate concentrations on S. cerevisiae and pH regulation approach
In general, a relatively higher acetate concentration from 2.6 to 20 g-COD/L improved OD max (Fig. 1a), biomass, and protein concentrations (Fig. 1d) of S. cerevisiae.Based on the modeling result (Fig. 1c), OD max achieved 1.4, 1.7, 1.6, and 2 fold of the control (2.6 g-COD/L) under 6, 8, 10, 20 g-COD/L, respectively.μ max had a dramatic increment and reached 2.32 obs/day at 20 g-COD/L, which was four times of control.It could also be seen from Fig. 1a, a sharp OD increase on Day 2-3 at 20 g-COD/L.However, at the highest COD, the lag phase was 1.6-fold longer than control, suggesting that S. cerevisiae required more time to acclimate to the hyperosmotic pressure due to high organic concentration.S. cerevisiae was reported to decline its cell volume and viability under high osmotic pressures of sorbitol (Pratt et al., 2003).The cell surface also shrank with a crenate envelope (Pratt et al., 2003).A high osmolarity glycerol signaling system was involved when S. cerevisiae was exposed to osmotic shock, where glucose as organic carbon was stimulated to convert into glycerol in order to increase intracellular osmotic pressure and confront the hyperosmosis (Klipp et al., 2005).A similar but unknown osmoregulation system may also exist when using acetate as organics instead of glucose, which caused the more extended lag phase.After Day 4, OD reached a stationary phase in all scenarios, which was attributed to the high pH above 8. pH always increased during incubation when acetate was gradually consumed.It is because acetate was transported across S. cerevisiae membrane by proton symporters, which means proton and acetate were synchronously transported by the symporter from medium into cytoplasm.The transmembrane pH difference functions as a thermodynamic driving force for acetate transport, which is the crucial step of acetate utilization by S. cerevisiae.The stoichiometry of proton and acetate is 1:1 in the symporter (Casal et al., 1996).Acetate only crosses the membrane in anionic form and via the proton symporter, rather than in an undissociated acid form and across by diffusion (Casal et al., 1996).The broth pHs were above 8 on Day 4-6, and S. cerevisiae stopped growing at pH 8 (Peña et al., 2015).The growth stagnation may be attributed to the minor intra-and extra-cellular pH difference and deficient transmembrane driving force for acetate transport.Acetate was fully consumed in 6 days at 2.6-10 g-COD/L, while 53% acetate was remanent at 20 g-COD/L (Fig. A3), indicating the potential of more SCP production if pH inhibition can be relieved.The acetate consumption rate at high CODs was around 1.5 times compared with low CODs, increased from 62-67 to 96-101 mg/(L⋅h) when COD increased from 2.6-6 to 10-20 g/L (Fig. A3).Maximal biomass weight of 2 g/L was achieved at 20 g-COD/L, which was double the control.Protein concentration improved from 0.37 to 0.75 g/L when COD increased from 2.6 to 20 g/L, with a constant protein proportion of 35-42% (Fig. 1d).
To solve the problem of alkaline inhibition, phosphate buffer (KH 2 PO 4 /K 2 HPO 4 ), acetate buffer (NaAc/HAc), or acetic acid replacing sodium acetate were employed (Fig. A4).Neither OD improvement nor mitigation of high pH was achieved by acetate buffer since ionic acetate was continuously consumed by S. cerevisiae.On the other hand, phosphate buffer ensured an acidic environment (pH < 6.72) during incubation.Nevertheless, OD max was much lower than control, which means less SCP production if phosphate buffer was applied.Interestingly, when the same COD of HAc replaced NaAc, S. cerevisiae could not grow at all, which indicated the importance of Na + in the AAB medium.NaAc could provide both organic carbon and Na + source for S. cerevisiae metabolism.
At 20 g-COD/L, acetate could not be consumed entirely due to the strong alkalization of broth.So 3 M HCl was externally added to regulate the pH.OD, pH, acetate concentration, and biomass results are shown in Fig. 2. S. cerevisiae with acid addition showed 2.6 times OD max of the scenario without acid addition (Fig. 2a and d).With acid addition, biomass yield was 1.57 g/L, including 0.94 g/L protein, an increase of 57% in biomass compared to control (2.6 g-COD/L).Besides, a rise of 25% in biomass and 67% in protein compared to the scenario without acid addition were also observed (Fig. 2e).HCl addition provided a mild acidic condition for S. cerevisiae and ensured more reliable acetate consumption (Fig. 2c).From Day 9, OD began to decrease because daily sampling (2 mL per day) consumed a portion of medium and left limited space for S. cerevisiae.So SCP production will be even higher if a fresh medium is supplemented.In addition, online pH control is preferable for SCP enhancement and recommended in future studies, rather than manually adding acid in this study.
Therefore, it is feasible for S. cerevisiae to utilize acetate and ammonia as carbon and nitrogen sources instead of sugar and peptone.Furthermore, high acetate concentrations can improve SCP production as long as S. cerevisiae is cultivated and maintained in a mild acidic condition.In this study, S. cerevisiae achieved maximal 1.57 g/L biomass and 0.94 g/L protein at 20 g-COD/L of acetate when the broth pH was daily regulated.

Short-chain VFA
The effect of propionate, butyrate and lactate on S. cerevisiae growth in terms of OD, VFA consumption, biomass, and SCP are shown in Fig. 3. S. cerevisiae could not grow on sole propionate, butyrate, or lactate as organics (Fig. 3a).However, it survived when acetate was added as well.Acetate was consumed within three days in all scenarios of dual VFAs.In particular, lactate was concomitantly consumed with acetate despite a   slower rate.88% lactate was rapidly consumed in the first three days and remained steady after that (Fig. 3b).As a result of extra organic carbon from lactate, OD, biomass, and protein concentration increased to 2.1, 1.06 g/L, and 0.6 g/L, respectively (Fig. 3d).Propionate concentration decreased by 39% during six days, while butyrate was not consumed at all (Fig. 3b).Both propionate and butyrate mildly impeded the growth of S. cerevisiae, in terms of OD max and μ max (Fig. 3c).It indicated that S. cerevisiae was unable to take in butyrate, while it took in propionate without assimilation.It was elucidated that propionate is a nonmetabolizable analog of acetate for S. cerevisiae in a glucose-free medium and will compete with acetate since they share the same proton symporter (Casal et al., 1996).The competitive relationship explained the decreased OD max and μ max when propionate was added.Lactate uses a different proton symporter, which is induced to be functional as long as lactic acid exists in the medium (Casal et al., 1996).Thus, acetate and lactate were in parallel transported into cytoplasm and utilized by S. cerevisiae, which contributed to more SCP production.In contrast, butyrate inhibition was not caused by proton symporter competition but by suppressing histone deacetylase in S. cerevisiae (Nguyen et al., 2011) which regulates genetic transcription.The result is in line with previous research, where butyrate was reported to have antifungal activity (Nguyen et al., 2011) and blocked fungus germination (Hoberg et al., 1983).
Hence, S. cerevisiae can utilize the mixture of acetate and lactate as organics.Wastewater rich in lactic acid may be another target for SCP synthesis by S. cerevisiae, e.g., from dairy plants, food processing factories, starch extracting industry, etc. (Andreani et al., 2019;Fernández-Nava et al., 2010;Meng et al., 2020).But it is noteworthy that sole lactate cannot support its growth, which indicates the importance of acetate for S. cerevisiae to metabolize lactate.The mechanism is unfortunately still uncharted.It could be speculated that acetate was necessary and possibly triggered the pathway as an inducer by activating some critical enzymes involved in lactate catabolism.Propionate and butyrate can neither be consumed by S. cerevisiae, and they negatively influence its growth and thus slightly reduce SCP production.So wastewater with acetic acid or lactic acid as the main component rather than other VFAs will be a desirable substrate for S. cerevisiae SCP synthesis.

Long-chain VFA
Compared with specific oleaginous yeast strains such as Candida maltosa and Yarrowia lipolytica, it was reported that S. cerevisiae can but poorly grow with oleate as the sole carbon source (Schüller, 2003).It can utilize oleate but requires the proliferation of peroxisomes (organelles in eukaryotic cells as an exclusive site for β-oxidation) and enzyme import for β-oxidation (Schüller, 2003).S. cerevisiae grew well when transferred from synthetic dextrose to oleic acid-containing medium (Ohdate and Inoue, 2012).However, in this study, S. cerevisiae completely stopped growth when it was shifted from AAB medium to new media with oleate or linoleate as the sole carbon source (data was not shown).It indicated that S. cerevisiae might lose the ability to transmembrane transport or degrade long-chain VFA via β-oxidation when it has been acclimated to utilize acetate.A protein complex called Pxa1p-Pxa2p is responsible for long-chain VFA transport (Hiltunen et al., 2003).Given the decreasing macroscopic opacity of broth from Day 1 to Day 2, oleate and linoleate were transported into the cytoplasm.So it was in high possibility that the β-oxidation process in peroxisomes of S. cerevisiae was suppressed with sole oleate or linoleate as organic carbon, and induced in the presence of acetate.Key enzymes in peroxisomals, such as carnitine acetyltransferase (CAT) and isocitrate lyase (ICL) which was involved in lipid degradation (Holdsworth et al., 1988), may be inactivated without acetate.
When mixed carbon sources (acetate & oleate, namely Ace-Ole, or acetate & linoleate, namely Ace-Lino) were applied, they were both transported to cytoplasm, as indicated by the absence of turbidity due to insoluble oleate and linoleate.Higher OD (1.59, 1.34) and biomass (1.25, 1.29 g/L) were achieved under Ace-Ole and Ace-Lino scenarios than control (OD 1.22, biomass 0.99 g/L).SCP production using Ace-Ole and Ace-Lino was 1.8-and 1.6-fold of control.Carbohydrate concentration in biomass was also increased by 2.9 and 2.1 times.The corresponding proportion variations in biomass are shown in Fig. A5.Peroxisomal proliferation in cells stimulated by lipid (Sibirny, 2016) and accordingly activated β-oxidation related enzymes (Hiltunen et al., 2003) are expected to cause biomass improvement.

Cation
The effect of Na + , K + and NH 4 + on OD, biomass, protein concentrations and nitrogen consumptions are shown in Fig. 4. As shown in Fig. 4a, S. cerevisiae stopped growing when NaAc was replaced by HAc, whereas it grew when Na + was supplemented in the form of Na 2 HPO 4 , and HAc was well consumed.S. cerevisiae also died of K + shortage when K + was fully removed from the AAB medium.The result confirms the necessity of both ions for S. cerevisiae maintenance.Even so, K + requirement (Approx.0.12 g/L) was 8 times lower than Na + (Approx.0.92 g/L).Despite growing when trace K + remained in the broth derived from liquor inoculum (2 mL inoculum into 50 mL fresh medium), it was unable to grow devoid of K + by inoculating a single colony from agar plate into fresh medium.Ion homeostasis is fundamental for any living cell.Intracellular K + is normally at a much higher concentration of 200-300 mM than external environment (2 μM-2 M) for microorganisms (Navarrete et al., 2010).K + /H + exchange can maintain the membrane potential and a relative constant intracellular pH (Ramos et al., 1990).In the YPD medium, S. cerevisiae produces H + during fermentation and exports them from cells.Accordingly, a net K + influx takes place to balance the charge and control the intracellular pH.The K + transport mechanism is unknown using the AAB medium.In this study, a sharp increase in K + concentration was observed from an initial 65 mg/L to 1471 mg/L on Day 2, and decreased to 68 mg/L on Day 6.It indicated that, contrary to the scenario in the YPD medium, intracellular K + was pumped out of cells along with H + uptake in the AAB medium when H + and acetate were co-transported into cells via proton symporters.Thus, less K + was required for S. cerevisiae when acetate was applied as sole organic carbon.However, the active K + influx was still underway but prevailed by K + efflux.K + starvation will eventually diminish intracellular K + and jeopardize the cell, which explained the death of S. cerevisiae when completely devoid of K + .Under different NH 4 -N concentrations, a small amount of NH 4 -N was observed to release from cells on Day 2 except at 1.6 g/L NH 4 -N (Fig. 4d, Table A3).The ammonia may derive from the dead cell decomposition of the inoculum.NH 4 -N was then rapidly consumed on Day 2-3, corresponding to the exponential phase of S. cerevisiae growth.During incubation, higher NH 4 -N concentrations up to 1.6 g/L did not benefit more SCP production and protein content in biomass because S. cerevisiae always consumed similar amounts of NH 4 -N (67-136 mg/L) in all cases.As shown in Figs.4c and 74%-92% of the nitrogen in broth was wasted at the end of incubation at 0.6-1.6 g/L NH 4 -N.Given the actual consumption of NH 4 -N, the nitrogen assimilation ratio of S. cerevisiae varies between 51 and 96% at different NH 4 -N concentrations.Given the high nitrogen conversion ratio, S. cerevisiae will enable efficient bulk production of SCP in the future.In contrast, S. cerevisiae was inferior in carbon conversion.It was observed that 56% of carbon was lost in the air as CO 2, and only 40% of carbon from acetate was converted into biomass (Fig. A6).
Therefore, 0.26 g/L NH 4 -N in the AAB medium was enough for SCP synthesis by S. cerevisiae in a 6-day batch, at the COD/N ratio of 10 (gCOD/gN).SCP production was limited by carbon source rather than nitrogen source because both acetate shortage and pH inhibition prevented the further improvement of OD (Fig. 2b and c).Overall, S. cerevisiae is an attractive candidate using acetate for SCP synthesis and thus applicable in future acetate-rich wastewater treatment containing slight ammonia.

Anion
As shown in Fig. A7, S. cerevisiae survived only using NH 4 -N as the nitrogen source.Neither NO 2 -N nor NO 3 -N was consumable by S. cerevisiae.Although S. cerevisiae grew and achieved OD max of 0.48 when it was first transferred to a fresh NO 3 -N medium, it failed to grow in the second generation.The poor growth of the first generation was attributed to the remnant ammonia in the inoculum.NO 3 -N concentration also revealed no consumption by S. cerevisiae.So NH 4 -N is the best inorganic nitrogen source for S. cerevisiae.
The response of S. cerevisiae to adding 20-500 mM additional Cl − , NO 3 − and SO 4 2− in aspects of OD, pH, biomass and protein concentrations are shown in Fig. 5. S. cerevisiae was capable of reaching the same OD max in 6 days under various concentrations of Cl − , NO 3 − and SO 4 2− , albeit with different lag phases (Fig. 5a).20-500 mM Cl − , NO 3 − and SO 4 2− prolonged lag phase by 25-62%, 12-88%, 8-60%, respectively (Fig. 5d).The most serious case was exposing S. cerevisiae to 500 mM NO 3 − , whose lag phase was 2.9 days, compared with 1.6 days for the control.μ max was also affected by high concentrations of anions.500 mM Cl − and SO 4 2− inhibited μ max by 46% and 43%, respectively.S. cerevisiae was more sensitive to NO 3 − and began to decelerate growth by 31% at 250 mM NO 3 − , and even 69% at 500 mM NO 3 − (Fig. 5d).Adding a certain amount of anions initially helped improve biomass, but the biomass would decline when anions were beyond the range (Fig. 5c).The hyperosmotic pressure caused by anions resulted in the biomass decrease.It was in line with previous research, where biomass reduced nearly 50% at an osmotic shock of 50 g/L NaCl (Yang et al., 2014).When exposed to high osmotic pressure, S. cerevisiae initiated an osmotic stress response that converted carbon flux to produce intracellular metabolites (i.e., erythritol) to resist water loss instead of accumulating biomass.SCP productions were insusceptible and varied within 0.33-0.48g/L, accounting for 38-57% of biomass (Fig. 5c).In short, S. cerevisiae is tolerant and adaptable to Cl − , NO 3 − and SO 4 2− .20-500 mM of these anions were not harmful for SCP production, although a longer lag phase was required under high anion concentration.

Biomass compositions and SCP amino acid profile
The main components of biomass that were tested in this study included protein, carbohydrate, lipid, and ash.As seen in Table 1, there was consistently 3-4% lipid in either AAB medium or YPD medium.However, in AAB medium, the biomass contained a bit lower protein (47%) and carbohydrate (13%) than the scenario in YPD medium with 57% protein and 19% carbonhydrate.Instead, more ash was found in the biomass, which accounted for 37% compared with 7% using YPD medium.The ash was derived from Mg 2+ in medium, which formed Mg (OH) 2 precipitation when pH continuously increased during incubation.Other metal ions as trace elements, i.e., Ca 2+ , Zn 2+ , Fe 2+ , Mn 2+ , could also cause precipitation due to broth alkalization, but they were all less than 0.05 mM, compared to 6.2 mM Mg 2+ .We also ruled out these possibilities by alkali titration into solutions with equivalent trace elements and MgSO 4 , respectively, and only the latter one ended up with white turbidity.
Theoretically, Mg 2+ begins to precipitate into Mg(OH) 2 at pH 7.98 until complete precipitation at pH 10.87 at 25 • C. In this case, the pH gradually increased and reached a high level around 7.5-9 on Day 3-6.Most Mg 2+ would precipitate in the form of Mg(OH) 2 .The inorganic Mg (OH) 2 precipitation was also collected along with biomass when the broth was centrifuged.As a result, high ash content was observed in S. cerevisiae biomass.The actual protein content (%) in biomass will be even higher because biomass was overestimated due to inorganic  precipitation.Therefore, acidic pH control is highly recommended for S. cerevisiae cultivation.It helps improve OD and SCP production of S. cerevisiae and will be an excellent way to avoid inorganic salt precipitation and get more pure SCP products.
The amino acid profile of S. cerevisiae biomass is evaluated in Table 2 and Fig. A8.Lower amino acid contents in SCP were observed when S. cerevisiae grew in the AAB medium.But it still contained wellbalanced amino acids, the same as the scenario in the YPD medium.Compared with the suggested amino acid requirement in high-quality animal protein for adults by the Food and Agriculture Organization of the United Nations (FAO) (FAO/WHO Expert Consultation, 1989), SCP of S. cerevisiae in AAB medium has met most of the amino acid requirement, especially outstanding in leucine, lysine, threonine, valine, phenylalanine, and tyrosine.Hestidine and sulfur-containing amino acids (Methionine and Cysteine) were the only exceptions slightly below the FAO recommendation.

Significance and perspectives
The results will offer insights into utilizing waste streams as the alternative substrate of S. cerevisiae for SCP production.For instance, one potential source could be anaerobic digestate produced by food manufacturers, often containing acetate and lactate (Bühlmann et al., 2021).Another option could be anaerobic digestate from oil mills containing long-chain VFAs such as oleate and linoleate (Chan et al., 2018).The anaerobic digestates containing VFAs will have an advantage over previous sugar-based waste substrates because they can be directly utilized by S. cerevisiae with less pretreatment (Table A4).Based on the carbon-nitrogen ratio for S. cerevisiae's assimilation (10 g-COD/g-NH 4 -N) obtained from this study, it would be possible to evaluate the feasibility of any other waste stream as the substrate.S. cerevisiae can also adapt to concentration fluctuations of inorganic salts, ensuring a robust performance of fermenters in practical use.Furthermore, maintaining an acidic environment in yeast fermenters was necessary to enable higher biomass, less salt precipitation, and maximize SCP production.
With 20 g-COD/L of acetate in the substrate under acidic pH control, we achieved the maximal SCP production (1.57g/L biomass with 65% protein).Future studies could focus on whether S. cerevisiae can use a higher organic load of VFAs.We also recommend a large-scale fermenter that operates continuously to simulate real-world applications better.In that case, fermenters can be optimized from different aspects for robust SCP production in the long term, for example, pH, temperature, hydraulic retention time, and harvest interval.The only costs for our labscale flask fermenters were the chemical addition in substrates, the electricity used for magnetic stirring, and the temperature control.However, for large-scale fermenters, a comprehensive economic assessment and food safety analysis of the waste stream-derived SCP product is necessary.

Conclusions
This study delved into the ability of S. cerevisiae to produce SCP using organic acids and nitrogen that are readily available from wastewater.The effect of various organic acids on the growth dynamics, biomass composition, and amino acid profile of S. cerevisiae was revealed.The harvested SCP had a decent portfolio of essential amino acids.Moreover, the effect of inorganic ions was also investigated, considering the future application of wastewater with a complex matrix.In addition, a high acetate concentration along with pH control was found to be a potential way to improve SCP production of S. cerevisiae, avoid Mg 2+ precipitation and thus ensure product purity.Among various nitrogen sources, ammonia is the best inorganic nitrogen source for S. cerevisiae.The appropriate carbon/nitrogen ratio was 10 g-COD acetate/1 g NH 4 -N.Furthermore, S. cerevisiae can catabolize lactic acid along with acetate and results in more SCP production.Oleate and linoleate as long-chain fatty acids are also digestible for S. cerevisiae along with acetate and contribute to high protein content in biomass.No apparent inhibition on S. cerevisiae SCP production was caused by Cl − , NO 3 − and SO 4 2− in the range of 20-500 mM.The results of this study will enable an insightful understanding of S. cerevisiae to maximize SCP production using wastewater-derived substrates and provide a guideline for future researchers to couple efficient SCP synthesis and resource recovery from wastewater.We envision that the substrate for S. cerevisiae SCP synthesis can be expanded to all sorts of wastewater containing acetate, lactate and lipids (oleate and linoleate), e.g., domestic wastewater, anaerobic digestion effluent, slaughterhouse wastewater, dairy factory wastewater, and edible oil factory wastewater.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 2 .
Fig. 2. Adding acid as a strategy for pH control at high acetate load (20 g-COD/L) (a) OD, (b) pH, (c) acetate concentration, (d) modeling results of OD max , μ max and lag phase, (e) biomass and protein concentration of S. cerevisiae.

Fig. 3 .
Fig. 3. Short-chain fatty acid impact on S. cerevisiae growth.(a) OD, (b) VFA concentrations, (c) modeling results of OD max , μ max and lag phase, (d) biomass and protein concentration of S. cerevisiae.

Fig. 4 .
Fig. 4. Cation effect on S. cerevisiae growth.(a) OD, (b) biomass and protein concentration of S. cerevisiae, (c) nitrogen balance at different ammonia concentrations, (d) specific ammonia concentration variations in substrate.
D.Zeng et al.

Table 1
Biomass compositions of S. cerevisiae enriched in different mediums.

The effect of anion and cation on S. cerevisiae 2.4.1. Cations Na + and K + ions were excluded from the AAB medium to investigate their necessity for S. cerevisiae. Detailed compositions for each batch are shown in Table A2. To examine Na + effect, Na + was first completely excluded by replacing NaAc & Na 2 HPO 4 with HAc & K 2 HPO 4 respec
tively, named as 'HAc (exclude Na) batch'.Then, to see if S. cerevisiae can use HAc instead of NaAc, HAc was applied, and 19.93 mM Na 2 HPO 4 was added to keep an equivalent inorganic Na + concentration.This group was assigned as 'HAc & Na 2 HPO 4 batch'.Since HPO 4

Table 2
Amino acid distributions in SCP of S. cerevisiae compared with recommendations in FAO standard.