Valorization of Solid Food Waste as a Source of Polyunsaturated Fatty Acids Using Aurantiochytrium sp. L3W

This study aimed at valorizing solid food waste containing docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA). Aurantiochytrium sp. L3W that produces DHA and EPA was cultivated on eight types of solid food waste: sake lees (SL), crown daisy, Japanese mustard spinach (JMS), soy sauce residue, lemon peel (LP), orange peel, grape skin, and Hiroshimana old pickle (HOP). The biomass mixture of the remaining food waste and strain L3W was analyzed for DHA and EPA. To characterize the types of food waste, the leachability of dissolved organic carbon (DOC) and dissolved nitrogen (DN) was compared. The strain L3W grew on both pasteurized and unsterilized food waste such as SL and JMS. Elution of DOC and DN from the food waste might be a factor affecting the growth of strain L3W. However, the strain L3W might utilize solid-state organic compounds in JMS. Despite the unsterile conditions, the biomass mixture of SL contained both DHA and EPA, whereas DHA was found in the biomass mixtures of JMS, LP and HOP, thereby confirming the valorization of these types of solid food waste. Unsterile mass cultivation of the strain L3W using SL and HOP in a 200 L tank also produced a biomass mixture containing 12.6 mg-DHA/g and 0.217 mg-EPA/g. These DHA and EPA contents were 1500-times and 37-times higher, respectively, than that in commercial poultry feed, indicating that the biomass mixtures could be used as an additive in poultry feed.


Introduction
Fish oil contains polyunsaturated fatty acids (PUFAs) such as docosahexaenoic and eicosapentaenoic acids (DHA and EPA) [1], which are essential for many marine species [2] including cultured fish such as salmon [3] and red sea bream [4]. Therefore, fish oil has been used as an important component of fish feed. The benefits of DHA and EPA for human health such as improving cardiovascular, visual, and neurological functions have been recognized [5], and DHA and EPA supplements for human consumption [6] and DHA-and EPA-enriched eggs [7] have been produced from fish oil. Because of the functions of DHA and EPA in the aquaculture and food industries, the global usage of fish oil has been increasing [3]. However, limitations exist on the production of PUFAs because of issues relating to the sustainability of wild fisheries and constraints on aquaculture systems [8,9]. Therefore, alternative sources of PUFAs are required to meet the increasing demand for fish oil. Thraustochytrids such as Aurantiochytrium, Schizochytrium, and Thraustochytrium are halophilic and heterotrophic microalgae that produce valuable PUFAs such as DHA and EPA [1,10]. The biomass of Schizochytrium limacinum is already commercially available [11]. The biomass of S. limacinum and other Schizochytrium spp. has been demonstrated to be a good source of DHA and EPA for culturing fish such as Atlantic salmon [11,12], totoaba [13] and trout [14]. For the cultivation of thraustochytrids, using artificial media may not be feasible because of the expensive growth substrates. Because of their high contents of organic compounds and low contents of harmful substances, utilizing wastewater from food processing, as well as liquid and solid food waste as alternative substrates may be a good option for producing thraustochytrid biomass as an alternative to fish oil.
Many studies have investigated the usability of wastewater and liquid waste from the food industry for cultivating thraustochytrids and producing PUFAs: examples include bean boiling water [15] and waste sauce [16] for Aurantiochytrium sp. L3W, soybean curd wastewater for Aurantiochytrium sp. ATCC PRA-276 [17], and waste oil for Aurantiochytrium limacinum SR21 [18]. Solid waste from food processing, cafeteria food waste [19], poultry waste [20] and brewery spent yeast [21] have also been demonstrated to be useful but an hydrolysis pretreatment was required before they could be used. However, only a limited range of information is available on the applicability of solid-state waste for cultivating thraustochytrids to produce PUFAs.
Many strains of thraustochytrids have been isolated from mangrove leaves [10,22]. Certain thraustochytrids are known to produce cellulase [23,24], indicating that solid-state waste might be usable as a substrate. Taoka et al. [25] used distillery lees (Shochu kasu) for culturing Schizochytrium aggregatum and Schizochytrium sp. TM02Bc and found that the contents of PUFAs such as DHA and EPA were enriched in the resultant Shochu kasu mixture but no details on the growth of thraustochytrids or the PUFAs contents in the biomass were provided. Neither was the possibility of growing thraustochytrids on the dissolved organic compounds released from the Shochu kasu investigated, although these might contain soluble organic compounds. If solid food waste can be used for cultivating thraustochytrids, a mixture of the remaining food waste and thraustochytrid biomass may contain PUFAs and used as a source of PUFAs for fish aquaculture and in the food industry for poultry feed. To investigate the feasibility of this valorization scenario, it is necessary to screen and characterize the usability of different types of solid waste for culturing thraustochytrid. For practical application, pretreating the food waste by sterilization and hydrolysis would preferably be avoided but adjusting the pH might be an option to mitigate any adverse effects of contaminating microorganisms on the production of PUFAs by a thraustochytrid under unsterile conditions [16].

3
The present study aimed to investigate the valorization of different types of solid food waste as a substrate without any hydrolysis by cultivating thraustochytrid containing PUFAs such as DHA and EPA. Aurantiochytrium sp. L3W was cultured on different types of pasteurized solid food waste, and any suitable types were characterized in terms of their leachability of dissolved organic carbon (DOC) and dissolved nitrogen (DN). The solid food waste was used for culturing the thraustochytrids under unsterile conditions at pH values of 4 and 7 to investigate the effect of pH adjustment on thraustochytrid growth and the production of PUFAs. Finally, the thraustochytrids will be cultivated using unsterilized solid food waste in a 200 L mass cultivation tank to investigate the feasibility of valorizing solid food waste as a PUFAs source. The usability of the valorized solid food waste as sources of PUFA for the aquaculture and poultry industries was finally discussed.

Thraustochytrid, Solid Waste, and Commercial Feeds
A thraustochytrid, Aurantiochytrium sp. L3W, was used in the present study [10]. This strain was preincubated at 25°C for 24 h with rotation at 70 rpm in a 790By + medium consisting of 5.0 g of d(+)-glucose, 1.0 g of polypeptone, and 1.0 g of yeast extract in 1.0 L of sand-filtered seawater. The sand-filtered seawater (30 psu) was collected at the Takehara Fisheries Research Station of Hiroshima University, Hiroshima, Japan. Two types of solid food waste, lees from Japanese sake brewing and soy sauce residue, were purchased from Gekkeikan Sake Co., Ltd., Kyoto, Japan and Maruhide, Saga, Japan, respectively. Two vegetable samples, the leaves of crown daisy (Shungiku) and Japanese mustard spinach (Komatsuna) were collected from a supermarket in Hiroshima, Japan then washed with tap water. To obtain samples of orange and lemon peels and grape skin, the fruits were purchased from a supermarket, then the edible parts were removed by hand after washing with tap water. In addition, we tested waste old pickles of Hiroshimana green (Hiroshimana) supplied from Yamatoyo, Hiroshima, Japan, because the Hiroshimana pickle is one of the specialty products of Hiroshima Prefecture, Japan. A commercial fish feed (Nova EP-O) was obtained from Hayashikane Sangyo Co. Ltd., Yamaguchi, Japan and poultry feed (Hi egg 17S) from the West Japan Feed Co., Okayama, Japan to allow a comparison of DHA and EPA contents with those of the valorized solid food waste.

Cultivation Using Various Types of Solid Waste
To investigate the use of solid food waste as substrates for the strain L3W, the waste samples except for the Hiroshimana old pickle were pasteurized by heating at 105 °C for 2 min. The reason for the exception was that the Hiroshimana old pickles were wet fermented food waste containing a seasoning liquid, and therefore the seven types of pasteurized food waste were applied for cultivation of the strain L3W. After drying, the two vegetable and three fruit samples were chopped to a size of less than 5 mm using a blender (IFM-FR10-R, Iwatani Corp., Osaka, Japan), but the sake lees and soy sauce residue were crushed manually. The seven solid food waste samples were then dispersed at a range of concentrations from 1 to 10 g-dry/L in 200 mL of sandfiltered seawater in 500-mL Erlenmeyer flasks that had been previously sterilized at 120 °C for 15 min. After inoculation with the precultured strain L3W at 10 4 cells/mL, the flasks were incubated in triplicate for 72 h at 25 °C with rotation at 70 rpm. The 790By + medium was also used in the control experiment. The pH values in the culture media to which the solid waste was added were at a range from 6.4 to 7.2. Because this is within the optimal range for the strain L3W [10], pH adjustment was not carried out prior to cultivation.
Of the seven pasteurized food waste samples, the sake lees, Japanese mustard spinach, and lemon peel were best at growing the strain L3W, as described later. Therefore, these three waste samples were further tested as well as the Hiroshimana old pickle chopped to sizes of less than 5 mm for culturing the strain L3W without being pasteurized. Each type of waste sample was dispersed at 5 g-dry/L in 200 mL of sand-filtered seawater in 500-mL Erlenmeyer flasks, followed by adjusting the initial pH to 4 or 7 using 0.1 M NaOH and 0.1 M HCl. The strain L3W was then inoculated and cultured as mentioned above.
During the cultivation period, the cell numbers were periodically measured by microscopic observation at a magnification of 400 X using a Thoma hemocytometer (0.1 mm depth, Matsuyoshi, Tokyo, Japan). To measure the content of fatty acids in the biomass mixture consisting of the remaining food waste and the biomass of the cultured strain L3W, 10 mL of the culture solution were centrifuged at 8232 g for 15 min then the pellet was washed twice with pure water and lyophilized.

Mass Cultivation
For the mass cultivation of the strain L3W, the sake lees was chosen as a substrate because it produced good contents of DHA and EPA under both pasteurized and unsterile conditions (Tables 1 and 2). Alongside the sake lees, the Hiroshimana old pickle was used, possibly for branding the aquaculture and poultry products resulting from feeding the 1 3 biomass mixture to fish and poultry. For example, mackerels fed with sake lees are marketed at a high price as the branded one, "Yopparai Saba (drunk mackerel)" in Fukui Prefecture Japan. Briefly, 6 kg of crude salt (Namishio, Diasalt Corp., Sapporo, Japan) were dissolved in 200 L of tap water in a 200-L mass cultivation tank equipped with a mixing paddle ( Fig. S1) then 1 kg-dry of each of the crushed sake lees and chopped Hiroshimana old pickle were dispersed in the solution. The strain L3W was inoculated at 3 × 10 4 cells/ mL in the aerated tank and incubated at room temperature. During the 144 h of the cultivation period, the cell growth was monitored as well as the water temperature, pH and dissolved oxygen (DO). Finally, the culture solution was filtered through a 1-µm mesh filter cloth to obtain the biomass mixture from which three samples were randomly taken to obtain a composite sample for analysis of fatty acids.

Elution of Dissolved Organic Carbon and Nitrogen
The different types of solid food waste were characterized as a useful substrate for cultivating the strain L3W in terms of the elution of dissolved organic carbon (DOC) and dissolved nitrogen (DN). The waste samples were subjected to a leaching test for DOC and DN in a similar manner to that for cultivating the strain L3W. Briefly, the crushed sake lees and soy sauce residue, the two chopped vegetable, and three chopped fruit samples were dispersed in 200 mL of sterilized sand-filtered seawater in 500-mL Erlenmeyer flasks at 5 g-dry/L then the flasks were rotated at 70 rpm at 25 ℃ for 48 h. In this experiment, the Hiroshimana old pickle was not tested, because it contained the seasoning liquid. The reason for using a leaching test lasting 48 h was to ensure that the time was sufficient for the DOC and DN concentrations to reach the maximum, as shown in Fig. S2.

Instrumental Analyses
To measure DOC and DN, the eluate samples were filtered through a glass fiber filter (GF/F, 0.7 µm-pore size, Whatman, Maidstone, UK) and then injected into a total organic carbon analyzer (TOC-VSCN, Shimadzu, Kyoto, Japan) with an attached total nitrogen chemiluminescence measurement unit (TNM-1, Shimadzu). To analyze the fatty acids (FAs), 500 mg of the lyophilized biomass mixture and commercial fish and poultry feeds were spiked with 0.1 µg of undecanoic acid in 100 µL of n-hexane as an internal standard (Humaidah et al. 2020) and then subjected to extraction and methylation using a methylation kit (06482-04, Nacalai Tesque, Kyoto, Japan). After purifying the methylated sample with a Simplified Liquid Extraction tube (Strata, Phenomenex, Torrance, CA, USA), 1-µL sample aliquots were analyzed using a gas chromatograph equipped with a flame ionization detector (GC7820A, Agilent, Santa Clara, CA, USA) and a DB-17 column (30 m × 0.250 mm × 0.5 µm, Agilent) to which helium was supplied as the carrier gas. The injection port was operated in the splitless mode at 300 °C. The column temperature was set to 40 °C for 5 min followed by a gradual increase at 4 °C/min to 250 °C then at 30 °C/min to 280 °C. The temperature was then held at 280 °C for 2 min [15]. A 37-component fatty acid methyl ester mixture (Sigma-Aldrich, St Louis, MO, USA) was used as the fatty acid methyl ester standard. As well as the fatty acids content, the carbon and nitrogen content of the biomass of strain L3W cultured in the 790By + medium was analyzed using an elemental analyzer (2400 Series II CHNS/O, PerkinElmer Co., Ltd., Waltham, MA, USA).

Statistical Analysis
The statistical significance of differences between mean values of the growth of strain L3W using the seven types of pasteurized food waste and between mean values for each food waste treated under pasteurized or unsterile conditions was determined by one-way analysis of variance followed by a Tukey test. A t-test was used to compare the growth of the strain L3W on the Hiroshimana old pickle between initial pHs of 4 and 7. The statistical analyses were done using the add-in software (Excel-tokei 2012, Social Survey Research Information Co., Ltd. Tokyo, Japan) for the Microsoft Excel 2013, and significant differences were assumed at p < 0.05.

Screening of Solid Food Wastes for Cultivation
The strain L3W grew on the pasteurized sake lees (Fig. 1) and although the growth after 72 h was lower than that in the control, this confirmed that the sake lees could be used as a substrate for growing the strain. In a previous study, S. aggregatum and Schizochytrium sp. TM02Bc were cultured on distillery lees for 15 days [25]. Although the amount of the distillery lees decreased after cultivation, the growth of these strains was not investigated further. The present study has directly confirmed that Aurantiochytrium sp. strain L3W can grow on sake lees and showed that a long cultivation period of 15 d may be unnecessary for thraustochytrids. Regarding the growth of the strain L3W after 72 h, the cell numbers were significantly higher when using 5 g/L of dried sake lees (p = 0.04). We therefore tested other food waste samples as a substrate for growing the strain L3W at this same concentration. As for the reason for less growth at 10 g/L, it might be due to the possible negative impact of the components of sake lees such as ethanol.
The growth of the strain L3W on pasteurized food waste after 72 h was normalized to that of the control experiment and then compared (Fig. 2). Of the seven waste samples, the sake lees and Japanese mustard spinach exhibited the highest growth of the strain L3W (p < 0.001), with the soy sauce residue and lemon peel as the next highest. In contrast, the strain L3W did not grow on the crown daisy and grape skin. These results confirmed that certain types of solid food waste can be used to cultivate thraustochytrids. Because the growth potential of the strain L3W depended on the type of food waste, we measured the elution of DOC and DN from the seven food wastes to characterize them. Figure 3 summarizes the relationship between the elution of DOC and DN from the pasteurized waste samples and the normalized growth of the strain L3W. Except for the Japanese mustard spinach (indicated by solid black circles), the eluted DOC and DN concentrations exhibited exponential relationships with the normalized growth values, thus indicating that elution of DOC and DN might be one of the factors dominating the growth of the strain L3W cultivated on solid food waste. Because the coefficient of determination for DOC was about ten times higher than that for DN, the elution of DN might affect the growth of strain L3W more than that of DOC.
The reason for exception of the Japanese mustard spinach may be the possible utilization of solid-state organic compounds by the strain L3W as the substrates. The Japanese mustard spinach eluted DOC and DN at only 53.2 and 2.33 mg/L, respectively, but exhibited the highest normalized growth of the seven waste samples. In contrast, the 790By + medium exhibiting the normalized growth (100%) was reported to contain much higher DOC and DN values of 4200 and 350 mg/L, respectively [15,16]. The biomass production of the strain L3W in the 790By + medium was about 1600 mg/L [15,16]. Because of the difficulty in separating the remaining Japanese mustard spinach and cells of the strain L3W, it was impossible to measure the biomass production of the strain L3W. However, the normalized growth based on cell numbers (Fig. 2) suggests that the biomass production on Japanese mustard spinach might be 59.4% of 1600 mg/L, i.e., 950 mg/L. The carbon and nitrogen contents of the biomass of strain L3W determined by the elemental analyzer were 60.5 ± 0.4% and 6.01 ± 0.09%, respectively, (n = 3). These values suggest that the carbon and nitrogen immobilized by the strain L3W might be about 60.5% and 6.01% of 950 mg/L, i.e., 575 and 57.1 mg/L, respectively. Because these estimated amounts of carbon and nitrogen were much higher than those for DOC and DN eluted from the Japanese mustard spinach, these results indicate the possible utilization of its solid-state carbon and nitrogen by the strain L3W.
Raghukumar et al. [26] reported that a thraustochytrid, Schizochytrium mangrovei produced cellulase using crystalline cellulose as a substrate. This thraustochytrid was recently reassigned to the genus Aurantiochytrium by Yokoyama and Honda [27]. However, a later study demonstrated the existence of extracellular cellulase in 14 of 19 strains of thraustochytrids but not in the five strains of Aurantiochytrium spp. using carboxymethylcellulose. Although the production of cellulase by the strain L3W was not investigated in the present study, one possible explanation for its growth on Japanese mustard spinach might be its ability to utilize dissolved and solid organic carbon and nitrogen. Comparing the characteristics of Japanese mustard spinach and crown daisy from which DOC and DN were released at the same level (Fig. 3), the former contains more water and ash (Table S1). However, the strain L3W did not grow on the crown daisy (Fig. 2). These imply that the Japanese mustard spinach may have loose structure as compared to the crown daisy and that the characteristic may be advantageous for the strain L3W to utilize it as the substrate. Future studies should address the characteristics of solid food waste suitable for consumption by thraustochytrids.
In addition to the crown daisy, the strain L3W did not grow on the grape skin (Fig. 2). The grape skin exhibited the excellent DOC elution more than 800 mg/L; however, the DN elution was the second lowest (Fig. 3). As the possible reason for the lack of growth of the strain L3W, less degradability of the food wastes and/or growth inhibitors, such as polyphenols in the grape skin, might also be considered. The present results provided evidence that the leachability of DOC and DN and the usability of solid-state organic carbon and nitrogen may be the factors determining the suitability of using particular types of solid food waste as substrates for cultivating thraustochytrids.

Production of DHA and EPA from Pasteurized Solid Food Waste
The FAs contents of the four types of food waste, sake lees, Japanese mustard spinach, soy sauce residue and lemon peel and of the resultant biomass mixtures and the biomass of strain L3W obtained in the control experiment were analyzed. Table 1 summarizes the four types of food waste that contained neither DHA nor EPA, whereas the pure biomass of strain L3W exhibited the highest DHA, EPA and total FAs contents. Because all biomass mixtures contained DHA and showed increase in total FAs contents, these results confirmed the feasibility of valorizing these four types of food waste regarding their DHA and/or EPA contents.
The sake lees was the only type of food waste from which both DHA and EPA were produced, at a ratio of DHA to EPA (20.6:1) similar to that in the pure biomass of strain L3W (29.1:1). One possible explanation for this might be its composition of organic compounds. For example, Osada et al. [28] reported that three samples of sake lees contained not only glucose at 12.5-16.8% but also free amino acids at 1900-3500 mg/g. Glucose is one of the carbon sources of the 790By + medium to which polypeptone containing free amino acids was added. Although the glucose and free amino acids in the tested sake lees were not analyzed, it would be reasonable to expect that these might have promoted the growth of the strain L3W.
As shown in Fig. 2, the growth of strain L3W on the Japanese mustard spinach was higher than that on the sake lees. However, the DHA content of the former biomass mixture was about half that of the latter biomass mixture, possibly because the Japanese mustard spinach was an inadequate substrate for the production of DHA by the strain L3W. The growth of strain L3W on the Japanese mustard spinach may involve possible the utilization of its solid carbon and nitrogen (Fig. 3). The substrate composition may naturally affect the synthesis of organic compounds by microorganisms. For example, Quilodrán et al. [29] reported that the production of DHA by the two thraustochytrids depended on the carbon sources. Regarding Aurantiochytrium spp., previous studies confirmed that squalene can be produced as well as DHA by Aurantiochytrium mangrovei [30] and Aurantiochytrium sp. T66 [31] and that xanthophylls can be produced by Aurantiochytrium sp. KH105 [32]. In order to find the unknown additional value of the biomass mixture, further studies should analyze organic compounds adding to DHA and EPA produced by the strain L3W growing on the Japanese mustard spinach. Figure 4 shows the growth curves of strain L3W on pasteurized and unsterilized Japanese mustard spinach samples. With no pasteurization, the growth of the strain L3W was suppressed at an initial pH of 7 (p = 0.0007). However, by adjusting the initial pH to 4, the growth increased significantly (p = 0.0035). This was possibly because of the inactivation of the contaminating microorganisms by the pHshock load, whereas the strain L3W is resistant to the acidic pH of 4 [10]. Figure 5 compares the normalized growth of strain L3W between pasteurized and unsterilized food waste, where any reduction due to unsterile conditions was insignificant at an initial pH of 4, not only for the sake lees (p = 0.89) but also for unsterilized lemon peel at an initial pH of 4 where the growth was higher than that on the pasteurized lemon peel (p = 0.354). The growth of strain L3W on unsterilized Hiroshima pickle at an initial pH of 4 was higher than that at a pH of 7 (p = 0.049), similar to that on Japanese mustard spinach. These results confirmed that adjusting the pH is a promising method for improving the growth of the strain L3W under unsterile conditions. When lemon peel was used, the normalized maximum growth at the initial pH of 4 under unsterile conditions was, surprisingly, higher than that under pasteurized conditions, possibly because of the hydrolysis of organic compounds to produce hydrolysates available for the strain L3W and/or the suppression of the autoxidation of antioxidants such as ascorbic acid and polyphenols contained in the lemon peel [33]. It has been reported that the autoxidation of polyphenols may induce an anti-algal effect through the production of radicals [34], as the suppression of autoxidation by ascorbic acid and polyphenols under acidic condition is well-known. Table 2 summarizes the DHA, EPA, and total FAs contents in the biomass mixtures produced using the four types of unsterilized solid food waste. Except for the DHA in the sake lees mixture at a pH of 4, the DHA, EPA, and total FAs contents in the biomass mixtures of sake lees, the Japanese mustard spinach and lemon peel exhibited lower values under unsterile conditions than under conditions (Table 1). However, the biomass mixtures of these three food wastes did contain DHA. The biomass mixture of sake lees also contained EPA. The Hiroshimana old pickle mixture also contained DHA, but its raw material did not (Fig. 6). The reason for the higher content of DHA in the sake lees mixture at an initial pH of 4 might be due to the possible suppressed degradation by the contaminating microorganisms. These results clearly showed that valorizing solid food waste containing DHA and/or EPA can be achieved by cultivating the strain L3W without pasteurization.

Application of Unsterilized Solid Food Waste for Production of DHA and EPA
The growth of strain L3W is critical for increasing the contents of DHA (C22:6n3) and EPA (C20:5n3) in the biomass mixtures, and the three most abundant FAs were DHA, palmitic (C16:0) and docosadienoic (C22:2) acids in the biomass of strain L3W (Fig. 6). These FAs formed the majority of the total FAs content in the resultant biomass mixtures of sake lees obtained under pasteurized and unsterile conditions at an initial pH of 4, where the normalized growth of the strain L3W was about 50% (Fig. 5). Under unsterile conditions, the growth of L3W on the sake lees at an initial pH of 7 was suppressed (Fig. 5), and the proportion of oleic acid increased (Fig. 6), possibly due to contaminating microorganisms. These three FAs were found in the biomass mixtures of Japanese mustard spinach and lemon peel obtained under pasteurized conditions. However, the proportion of palmitic acid was highest under pasteurized conditions, possibly because of the different substrate compositions. On lemon peel, the growth of strain L3W at an initial pH of 4 under unsterile conditions was greater than that under pasteurized conditions (Fig. 5). However, under unsterile conditions, the proportion of the other FAs increased (Fig. 6), possibly because of the greater production of FAs from the lemon peel caused by hydrolysis and/or the activity of contaminating microorganisms. Under unsterile conditions, the normalized growth values of strain L3W on the Japanese mustard spinach at an initial pH of 7 and on the Hiroshimana old pickle at an initial pH of 4 and 7 were the three lowest (Fig. 5), and therefore the proportion of DHA in the biomass mixtures was also lowest (Fig. 6).

Mass Cultivation for Producing Sources of DHA and EPA
The strain L3W grew successfully on a mixture of sake lees and Hiroshimana old pickle (Fig. 7a). The time to achieve maximum growth was 96 h, whereas 48 h was sufficient for growth on sake lees at 25 °C and pH 4 during flask-scale cultivation under unsterile conditions. This was not surprising because the temperature for mass cultivation was between 16 and 21 °C (Fig. 7b), and the optimal temperature for the growth of strain L3W is between 20 and 25 °C [10]. The strain L3W is an obligate aerobe, and during the first 24 h of cultivation, the DO was less than 3 mg/L (Fig. 7b), which was lower than that during the flask-scale cultivation of more than 6 mg/L. The resultant biomass mixture collected after 120 h contained DHA and EPA at 12.6 and 0.217 mg/g, respectively, thus confirming the valorization of solid food waste in terms of DHA and EPA contents on the mass scale unsterile cultivation. These values were expected because the biomass mixtures of sake lees produced in the flask-scale cultivation contained DHA at 26.9 mg/g (pH 4) and 13.1 mg/g (pH 7) and EPA at 0.155 mg/g (pH 4) and 0.205 mg/g (pH 7) (Table 2), whereas the DHA content in the biomass mixture of Hiroshimana old pickle was 2.23 mg/g (pH 7) and 1.16 mg/g (pH 4), respectively ( Table 2).

Utilization of the Biomass Mixture as a Source of DHA and EPA
The biomass mixture of sake lees and Hiroshima old pickle produced in the mass cultivation tank contained DHA and EPA at 12.6 and 0.217 mg/g, 1500 times and 37 times higher, respectively, than those in the samples of commercial poultry feed (DHA, 0.0084 mg/g; EPA, 0.0058 mg/g), thus indicating that the biomass mixtures are potentially usable as an additive in poultry feed to produce DHA and EPA-enriched products such as eggs. In contrast, the commercial fish feed contained DHA and EPA at 5.73 and 2.13 mg/g, respectively. Although the three sake lees biomass mixtures contained DHA at levels of 26.9, 13.1 and 12.6 mg/g, more than twice that of commercial fish feed, a higher margin may be necessary because the biomass mixtures cannot serve as a complete fish feed but only as an additive for fish feed. The EPA contents of the biomass mixtures (0.155, 0.205 and 0.217 mg/g) were lower than those of the commercial fish feed. This means that a greater enrichment of DHA and EPA contents in the biomass mixtures would be necessary for use as an additive in fish feed.
Optimizing the operational parameters for cultivating the strain L3W is one option for increasing the production of DHA and EPA. Jakobsen et al. [35] found that limiting O 2 increased the DHA content of the biomass of Aurantiochytrium sp. T66 as well as decreasing the palmitic acid content, which suggested that O 2 limitation hindered the O 2 -dependent desaturase(s) and favored the O 2 -independent polyunsaturated fatty acid synthase. In terms of energy consumption for cultivation, it is also reasonable to optimize the aeration time. A later study based on a fermentation strategy for producing DHA by Aurantiochytrium limacinum SR21 used intermittent oxygen feeding in a fed-batch culture system [36]. An increase in the DHA content in the biomass of A. limacinum SR21was associated with a decrease in its C16:0 content [35]. As well as the DO, pH may also affect the production of DHA and EPA by the strain L3W [10]. However, the value set for the pH may depend on the type of food waste in real applications because of the possible influence of pH on the utilization of the food waste and the activities of contaminating microorganisms in the biomass and FAs production ( Fig. 5; Table 1). Future research should focus on optimizing the cultivation processes or thraustochytrids using food waste to increase the DHA and EPA contents of the resultant biomass mixture.

Conclusion
In the present study, we investigated the valorization of eight types of solid food waste: sake lees, crown daisy, Japanese mustard spinach, soy sauce residue, lemon and orange peels, grape skin, and Hiroshimana old pickle by using them as a substrate for cultivating Aurantiochytrium sp. L3W to produce PUFAs such as DHA and EPA. The strain L3W grew on both pasteurized and unsterilized food waste such as sake lees and Japanese mustard spinach, thereby confirming that the solid food waste can be used as a substrate. The elution of DOC and DN from the food waste might be one of the factors affecting the growth of the strain L3W but on Japanese mustard spinach, the strain might utilize solid-state organic compounds. For cultivating the strain L3W using unsterilized food waste, adjusting the pH was a promising method to mitigate any reduction in its growth because of unsterile conditions. Despite unsterile conditions, both DHA and EPA were present in the biomass mixture of sake lees, whereas those of Japanese mustard spinach, lemon peel and Hiroshimana pickle contained only DHA. These results confirmed that food waste can be valorized in terms of DHA and/or EPA contents by their use as substrates for cultivating the strain L3W under unsterile conditions.
Of the types of food waste examined, the sake lees was the best for producing DHA and EPA. Under unsterile conditions, the biomass mixtures of sake lees contained DHA and EPA. The biomass mixture of sake lees and Hiroshima old pickle obtained by mass cultivation contained DHA and EPA at 12.6 and 0.217 mg/g, respectively. These DHA and EPA contents were 1500 times and 37 times higher, respectively, than those of the commercial poultry feed. This indicated that the biomass mixtures are potentially usable as an additive for poultry feed to produce DHA and EPA-enriched products such as eggs. For utilizing the biomass mixtures as a supplement for fish feed, the further enrichment of DHA and EPA contents is necessary. Future research should focus on optimizing the operational parameters such as DO in the cultivation processes for thraustochytrids using food waste to increase the DHA and EPA contents in the resultant biomass mixture.