Yeast single cell protein production from a biogas co-digestion substrate

Biogas plants serve as hubs for the collection and utilization of highly nutritious waste streams from households and agriculture. However, their outputs (biogas and digestate) are of relatively low economic value. Here, we explore the co-production of yeast single cell protein, a potentially valuable feed ingredient for aquaculture and other animal producing industries, with biogas on substrate collected at a co-digestion biogas plant, using three yeast species well suited for this purpose (Wickerhamomyces anomalus, Pichia kudriavzevii, and Blastobotrys adeninivorans). All yeasts grew rapidly on the substrate, yielding 7.0–14.8 g l−1 biomass after 12–15 The biomass crude protein contents were 22.6–32.7 %, with relatively favorable amino acid compositions mostly deficient in methionine and cysteine. Downstream biomethanation potential was significantly different between yeast species, with the highest product yielding species (Blastobotrys adeninivorans) also yielding the highest biomethanation potential. Highlights All yeasts grew well on the biogas substrate, with high growth rates. Produced biomass was of high nutritional value for use in fish feed formulations. Downstream effects on methane potential were strain-dependent. Yeast biomass may be a viable biogas co-product.

amino acid compositions mostly deficient in methionine and cysteine. Downstream 23 biomethanation potential was significantly different between yeast species, with the 24 highest product yielding species (Blastobotrys adeninivorans) also yielding the 25 highest biomethanation potential. 26 27 Highlights 28 • All yeasts grew well on the biogas substrate, with high growth rates. 29 • Produced biomass was of high nutritional value for use in fish feed 30 formulations. 31 • Downstream effects on methane potential were strain-dependent. 32 • Yeast biomass may be a viable biogas co-product. Organisms suitable for SCP production on OFMSW should, first and foremost, be 112 able to utilize a large array of substrate molecules. Other desirable characteristics 113 include phytase production, as this may improve nutritional quality of the feed if 114 ingredients of vegetable origin are included in the formulation (Cao et al., 2007), as 115 well as the ability to outcompete other organisms due to the non-sterile nature of the 116 substrate. In this study, we evaluated three yeast species with suitable properties for 117 SCP production on typical co-digestion AD substrate: Wickerhamomyces anomalus, 118 a metabolically versatile species which has shown robustness to difficult growth 119 conditions, has been evaluated in fish feeding trials and which is known for its 120 biocontrol properties as well as being a phytase producer (Huyben et  adeninivorans CBS 7377), stored in 50% glycerol stocks at -80°C, were inoculated 141 onto YPD agar (10 g l -1 yeast extract (BD, Le Pont-de-Claix, France), 20 g l -1 142 bacterial peptone (BD, Le Pont-de-Claix, France), 20 g l -1 D-glucose (Merck,  143 Darmstadt, Germany), and 20 g l -1 agar (BD, Le Pont-de-Claix, France)). Inoculum 144 cultures were prepared using the same medium, without agar, in 125-ml baffled 145 Erlenmeyer flasks (Thomson Ultra-Yield, Thomson Instrument Co., Carlsbad, CA, 146 USA), and cultivated on a rotary shaker for 24 h. Cells were harvested at 3000 × g for 147 5 min and washed with saline (NaCl, 9 g l -1 ) using the same settings. 148 149

150
The biogas substrate was obtained directly from the inlet to the digester at a biogas 151 plant in Sweden, and consisted mainly of source-separated household waste, 152 organic waste from municipal kitchens, and liquid agricultural waste (swine and cattle 153 manure). Metals and plastics had been mechanically removed at the biogas plant, 154 and the substrate had been hygienized at 70°C for >1 h. This substrate will be 155 referred to as native substrate (NS). 156

157
To be able to separate yeast biomass after culturing, and to reduce the risk of 158 contamination as substrate was collected through a non-sterile sampling port at the 159 biogas plant, the substrate was sterile-filtered. This was accomplished using an To assess downstream effects of yeast cultivation on biogas performance, spent 197 medium was collected from cultivations of each yeast. Cultivations were performed 198 largely as described in Section 2.3. In order to minimize confounding factors, all 199 cultivations were terminated at the same time (i.e., the time was determined by the 200 growth performance of the slowest growing yeast), and were run at the same 201 temperature (30°C). This was needed to ensure that evaporative losses of volatile 202 energy carriers, such as short-chain carboxylic acids, were similar between the 203 treatments. At the end of cultivation, yeast biomass was collected as described in 204 Section 2.3, and the supernatants, referred to as spent media, collected. 205

206
The BMP assay was conducted largely according to Angelidaki et al. (2009). In brief, 207 total solids (TS) and volatile solids (VS) of NS and spent medium (supernatants, post 208 yeast-treatment) were determined by drying the substrates at 105°C and incinerating 209 at 550°C in aluminum containers, noting the weights after each step. TS was 210 calculated as the quotient of dry matter divided by initial weight. VS was determined 211 as the difference between TS and ash content. 212

213
The assay was performed using untreated NS and fresh inoculum (collected from the 214 same biogas plant and degassed for 3 days at 37°C), contributing approximately 1.2 215 g VS and 3.6 g VS, respectively. The substrate control treatment contained only NS 216 and inoculum. Spent medium was added in the remaining treatments, so that the 217 mixtures contained, by weight, NS:spent medium in ratios of 10:1, 10:3, 2:1, and 1:1, 218 which corresponded to 9-50% spent medium in the final AD slurry. Due to the low 219 VS content of the spent medium, increases in VS due to supernatant additions were 220 modest, at most 20%, and it was assumed that this slight change in 221 inoculum:substrate ratio would not affect inoculum performance. Inoculum and 222 cellulose process controls were included. The inoculum control, used for 223 determination of background methane production, consisted of inoculum contributing  (Table 1). 292 293 Final biomass concentrations ranged from 7.0 g l -1 (22.6% CP) for W. anomalus to 294 14.8 g l -1 (30.5% CP) for B. adeninivorans. Final biomass productivities ranged from 295 0.53 g l -1 h -1 for P. kudriavzevii to 0.99 g l -1 h -1 for B. adeninivorans (see Table 1 for 296 the kinetic parameters and gross nutritional characterization of the yeast biomass). 297 Protein content according to the AA analysis was largely in agreement with CP 299 values, indicating that most nitrogen present in the biomass originated from protein 300 (Table 2). Yeast biomass AA composition was similar for all three species, with W. 301 anomalus notably having a somewhat lower content of methionine and cysteine, 302 sulfur-containing EAA, compared to the other two yeasts. In general, yeast biomass 303 was deficient in arginine and in the sulfur-containing AA relative to the requirements 304 of finfish, whereas other AA where either in excess or at similar levels ( Table 2)

322
Filtration of the substrate removed a considerable amount of nutrients, presumably 323 contained in particulate matter. Compared to NS, filtered substrate contained 5-fold 324 less VS, 2.1-fold less total N, and 4.5-fold less total C (Table 3). Compared to the 325 filtered supernatant, spent medium contained less TS, VS, C and N, and generally 326 reduced levels of micronutrients. The exceptions were Na, which increased in all 327 treatments, and P, which increased in two out of three treatments. This is most likely 328 due to the use of NaOH and H3PO4 for pH control. Treatment with P. kudriavzevii 329 yielded the lowest reduction in TS and VS. 330 331 The BMP assay was terminated after 40 days. The methane potential of native 336 substrate with addition of spent substrate was significantly different depending on 337 strain choice (p = 0.0009), with no effect of the dose of spent medium added or 338 interaction of strain and dose observed (Table 4). Native substrate with spent 339 medium had BMP levels of 342.0-380.5 ml CH4/g VS ( In this study, we have produced yeast biomass from the liquid fraction of a substrate 352 from a Swedish co-digestion biogas plant. Depending on the yeast strain, biomass 353 concentrations reached 7.0-14.8 g/l after 12-15 h. By including the spent medium 354 from yeast fermentation in the AD slurry, we were also able to assess the relative 355 effects of each yeast strain on downstream methane potential. 356 357 For yeast-based SCP to be a viable co-product at a biogas plant, the organism must 358 be able to utilize the substrate used at the plant. The species evaluated here grew 359 well on the biogas substrate, with the highest final biomass productivity being 0.99 g 360 l -1 h -1 , with only minor efforts made to optimize cultivation conditions. Initial 361 experiments, carried out in shake flasks and in aerated 96-well deep-well plates, 362 were conducted to determine suitable fermentation parameters, but no subsequent 363 optimization was conducted. Furthermore, fermentations were performed at the same 364 pH for all strains, in order to get representative spent substrates for the BMP assay. It 365 is worth noting that none of the yeast strains were able to grow at a pH of below 6, 366 likely due to the presence of weak acids in the substrate. The substrate otherwise 367 proved non-toxic to the yeast and no dilution was necessary. This is important for a 368 biogas plant, as further addition of water would lower the organic loading rate and 369 hydraulic retention time, compromising methane production. Notably, lysine levels were close to the requirements of finfish. Lysine is commonly a 379 limiting amino acid, especially in feed ingredients of vegetable origin but also in many 380 species of yeast . With further optimization, it is likely that the 381 protein content could be increased. Rajoka et al. (2006), in a study using Candida 382 utilis, found that true protein contents increased during the first 24 hours of 383 cultivation. Likewise, for Saccharomyces cerevisiae it was shown that protein 384 contents were highest after 36 h, for two out of three strains evaluated (Novak, 385 2007). In the present study, however, cultivations were terminated after 12-15 h, 386 suggesting an obvious route for optimization. 387

388
The choice of yeast strain proved to be important both for product yields as well as 389 for downstream BMP. The best-performing yeast strain, B. adeninivorans, produced 390 approximately twice the amount of biomass compared to the other two species, with 391 a similar protein content. Accordingly, final biomass productivity was also 392 approximately double compared to the two other strains. It is possible that this is due 393 in part to the higher cultivation temperature used for this strain. Rajoka et al. (2006) 394 found that temperature had a large impact on crude protein contents of yeast 395 biomass, which they attributed to increased transport of nutrients over the cell 396 membrane. Interestingly, B. adeninivorans also had the highest BMP when mixing its 397 spent substrate with native biogas substrate, suggesting a synergistic effect of this 398