Canadian Journal of Microbiology

Transcription

1 Recovery of glucose from dried distiller's grain with solubles using combinations of solid-state fermentation and insect culture Journal: Canadian Journal of Microbiology Manuscript ID cjm r2 Manuscript Type: Article Date Submitted by the Author: 04-May-2018 Complete List of Authors: Howdeshell, Timothy; University of Saskatchewan College of Agriculture and Bioresources, Food and Bioproduct Sciences Tanaka, Takuji; University of Saskatchewan College of Agriculture and Bioresources, Food and Bioproduct Sciences Keyword: Is the invited manuscript for consideration in a Special Issue? : Insect bioprocessing, Recovery of fermentable sugars, Black soldier fly larvae, Aspergillus, Lactobacillus N/A

2 Page 1 of 34 Canadian Journal of Microbiology 1 2 Recovery of glucose from dried distiller's grain with solubles using combinations of solid- state fermentation and insect culture 3 4 Timothy Howdeshell, Takuji Tanaka * 5 6 Department of Food and Bioproduct Sciences, 7 College of Agriculture and Bioresources, 8 University of Saskatchewan 9 51 Campus Drive, Saskatoon, SK S7N 5A8 10 Canada 11 Phone: Fax: takuji.tanaka@usask.ca 14 * Correspondence should be addressed. 15 1

3 Page 2 of Abstract: A bioethanol byproduct (dried distiller s grains with solubles; DDGS) contains a high cellulose and starch. We hypothesized combinations of solid-state fermentation (SSF) and digestion by black soldier fly larvae (BSFL) (Hermetia illucens) could increase the recovery of glucose from this byproduct through concentrating and loosening cellulose matrix by their activities. DDGS was individually fermented with Aspergillus niger, Aspergillus fumigatus, Trichoderma koningii, Phanerochaete chrysosporium, or Lactobacillus plantarum. The fermented DDGS was fed to BSFL, and glucose recoveries from spent feeds were conducted. SSF increases lipid and protein contents, supporting BSFL growth, as well as weakening cellulosic matrix. BSFL use nutrients in SSF-DDGS, further concentrating and weakening cellulose, i.e., DDGS becomes a half without changing cellulose contents. For example, Lactobacillus plantarum SSF with BSFL culture concentrates the cellulose content from 9.7% to 26.5% of spent feed. Glucose recovery was determined using three sequential processes (free glucose determination, weak-acid hydrolysis of amorphous cellulose and enzymatic hydrolysis of micronized crystalline cellulose). Total glucose obtained from 100g of DDGS increases from 4.8 g to 10.7g. These results show that the combinations of SSF and BSFL would provide additional fermentable sugars (and insect biomass) from bioethanol byproducts, suggesting a high productivity from the same feedstock Keywords: Insect bioprocessing; recovery of fermentable sugars; black soldier fly larvae; Aspergillus; Lactobacillus 2

4 Page 3 of 34 Canadian Journal of Microbiology Introduction: Biofuels, such as bioethanol, have been considered to reduce the reliance on and consumption of petroleum-based fuels. First-generation bioethanol is made from food crops such as starchy grains and sugarcane. Second-generation bioethanol, such as lignocellulosic ethanol, is made from renewable non-food commodities which does not compete with food supply chains, making it an attractive alternative to first generation biofuels. However, utilization of lignocellulosic ethanol remains elusive due to the recalcitrant nature of lignocellulose. Lignocellulose is a complex mixture of lignin, hemicellulose, and cellulose which comprises a large amount of the dry matter of plants wherein it helps to give structural integrity. Cellulose polymers join together to form cellulose fibres through hydrogen bonding with unbound hydroxyl groups on adjacent glucose units, leading to the formation of crystalline regions of cellulose. Crystalline cellulose is more resistant to both chemical and enzymatic hydrolysis. This is due in part to the tightly packed structure which prevents the access of liquids and small molecules such as enzymes from the interior of the crystalline cellulose particle. In general, the recalcitrant nature of lignocellulosic biomass means that pre-treatment steps are necessary to achieve optimum conversions of lignocellulose into biofuel. Pre-treatment of lignocellulosic biomass must meet the following requirements: 1) improve the ability to liberate sugars; 2) avoid degradation or reduction in the carbohydrate; 3) avoid the formation of by-products which are inhibitory to downstream processes; and 4) be cost effective (Sun & Cheng, 2002). Pre-treatment of lignocellulosic biomass serves to dissociate and expose cellulose from its original matrix, thereby increasing the efficiency of acid and enzymatic hydrolysis of cellulose. 3

5 Page 4 of Solid-state fermentation (SSF) has been utilized as a means of lignocellulose pretreatment via enzymatic action. SSF is a means of introducing degrading enzymes into a substrate, and fungal SSF in particular has been used to hydrolyze and dissociate the components of lignocellulose: hemicellulose, lignin, and cellulose. Fungal organisms achieve this through a secretion of a cocktail of extracellular enzymes which can act synergistically to disrupt not only lignocelluloses, but also other biomass constituents such as starches and protein networks. In addition to fungal SSF, bacterial SSF can provide as a means of biological pre-treatment in biomass. However, SSF alone may not be sufficient to make polymeric carbohydrates susceptive to hydrolysis in order to recover sugars Many species and strains of organisms have been used in fermentation, but filamentous fungi are the best organisms for SSF due to their low water activity requirements and ability to penetrate into and between substrate particles. Compared to submerged fermentation SSF has a lower capital cost, lower energy requirements, needs simpler fermentation substrate, and requires a less rigorous control of the fermentation parameters (Yang et al. 2012). Wood rot fungi secrete much higher levels of extracellular enzymes than yeast or bacteria, and white-rot fungi in particular are able to degrade all plant cell wall components, including lignin (Yang et al. 2012). Phanerochaete chrysosporium is a white-rot fungus which secrete of a variety of peroxidases and oxidases which act non-specifically on the heteropolymer lignin through the generation of lignin free-radicals, which then undergo cleavage reactions (Singh & Chen, 2008). Aspergillus fumigatus is a saprotroph which is fairly ubiquitous throughout nature, and is often found in decaying matter, such as in compost heaps (Adav et al. 2015). It produces a diverse variety of extracellular enzymes including cellulases and other glycoside hydrolases, amylases, hemicellulases, lignin degrading enzymes, poly-peptide degrading enzymes, chitinases, lipases, 4

6 Page 5 of 34 Canadian Journal of Microbiology and phosphatases which are useful bioprocessing of lignocellulosic materials. Trichoderma koningii is another fungal strain which is ubiquitous throughout nature, being frequently isolated from soil samples. It produces a variety of extracellular enzymes, but is most frequently noted for its cellulase enzyme production (Wood & McCrae, 1978). Aspergillus niger is a common contaminant of food and produces black mould on some fruits and vegetables. It has GRAS status by the U. S. FDA under the Food, Drug, and Cosmetic Act (Schuster et al. 2002). It produces a wide variety of extracellular enzymes, and thus is used in a wide variety of industrial processes such as: the production of glucoamylase for sugar production; pectinase production used for juice and wine clarification; production of proteases which are used to reduce the gluten content of beer; and more. Relevant to this project is the production of cellulase complexes. Aspergillus is known to produce high amounts of β-amylase particularly (Bansal et al. 2012) The use of insect larvae in agricultural and food systems has recently gained popularity and sparked the rise of a cottage industry in North America. Of special interest is the wastedegrading insect Hermetia illucens, also known as the black soldier fly (BSF). BSF are a nonpest, tropical and subtropical insect, commonly associated with decaying animal, plant and other organic matter. The insect has four life stages: egg, larvae, pupa, and adult. Under optimized conditions, the larvae can reach the pre-pupal stage in 2-3 weeks. Captive, large-scale, indoor breeding of the BSF has been achieved, making them suitable for both small scale and commercial applications (Sheppard et al. 2002). The BSF larvae (BSFL) are high in lipid, as well as protein (40-44% dry matter (DM)), calcium (5-8% DM), and phosphorous ( % DM) (Makkar et al. 2014). The amount of lipid is highly variable and dependent upon the feed source of the larvae. Literature reported lipid values ranged from 15% on poultry manure diet to as high as 49% on oil rich diets (Makkar et al. 2014). They have an amino acid profile comparable to 5

7 Page 6 of fishmeal, and are sold commercially for use in aquaculture, poultry and pig feeding where they have shown high acceptability in feeding as compared to traditional feeds (Makkar et al. 2014). The protein and lipid in BSFL can be separated for use in animal feed and conversion into high quality biodiesel respectively (Li et al. 2011; Zheng et al. 2012a; Zheng et al. 2012b). One interesting nature of the larvae is their ability to liquefy the feed. These omnivorous larvae can degrade much organic matter into paste. This phenomenon is partly due to their powerful mastication, micronizing the particles in the feed. Dried distiller s grains with solubles (DDGS) is a by-product of grain ethanol fermentation. The chemical composition of DDGS varies with processing conditions and crop variations, but on the average for wheat DDGS protein, oil, ash, and total carbohydrates are 38.9%, 5.1%, 5.3%, and 50.7% (dry weight basis; dwb) respectively (University of Saskatchewan, 2010). Cellulose and starch constitute 18.7% and 8.3% of the carbohydrates in wheat DDGS (University of Saskatchewan, 2010). The SSF treatment loosens the structure of lignocellulose and makes available nutrients in DDGS more accessible to BSFL. The microorganisms also add nutrients to the substrate which might be inadequate in the untreated sample for the BSFL s nutritional requirements, such as lipid. Since BSFL can masticate these weakened cellulose materials into even smaller particles, we hypothesize that treating DDGS with both SSF and BSFL can increase the efficiency in the sugar recovery from DDGS through exposing cellulose for acid and enzymatic hydrolysis, and other valuable nutrients in DDGS can be recovered in the form of insect biomass. In this study, we aimed to examine the combinations of SSF and BSFL culture in order to concentrate cellulose and to modify susceptivity of cellulose to the hydrolysis. DDGS was treated with SSF for 7 days, followed by BSFL digestion for 12 days. The fermented DDGS was fed to BSFL, and the spent feeds before and after BSFL 6

8 Page 7 of 34 Canadian Journal of Microbiology digestion were analyzed for glucose yields using three sequential analyses involving dilute acid hydrolysis followed by enzymatic hydrolysis. 7

9 Page 8 of Materials and Methods: Materials: DDGS was kindly provided by NorthWest Bioenergy (Unity, SK). BSFL (Hermetia illucens) were purchased from Supercricket (Prince Albert, SK) and Bugs4Pets.com (Vancouver, BC), and were maintained for several generations in a greenhouse breeding enclosure at the University of Saskatchewan (Saskatoon, SK). All chemicals used in this study were commercially available ACS grade, and were purchased from VWR International (Edmonton, AB) and Fisher Scientific (Ottawa, ON). Extraction of bacterial DNA was done using the EZ-10 Spin Column Plasmid DNA Kit, and was purified after PCR using the EZ-10 Spin Column PCR 139 Purification Kit (Bio Basic Inc., Markham, ON). DNA sequencing was performed by the Plant Biotechnology Institute, National Research Council (Saskatoon, SK). All fungal strains were obtained from the ARS Culture Collection (Bacterial Foodborne Pathogens and Mycology Research Unit; National Center for Agricultural Utilization Research; Peoria, IL USA): Aspergillus niger BRRL 322, A. fumigatus BRRL 163, Trichoderma koningii BRRL 54330, Phanerochaete chrysosporium BRRL Glucose concentration was analyzed using the Glucose (HK) Assay Kit from Sigma Aldrich (St. Louis, MO, USA). Isolation of Lactobacillus plantarum: L. plantarum was isolated from spent feed beds of BSFL. Dilute suspensions of spent larval feed were again plated on MRS media agar plates which were incubated at 37 C in aerobic and anaerobic conditions. Colonies possessing different oxygen requirements or colony morphologies were streaked onto MRS plates in order to obtain a pure bacterial isolate. These isolates were grown in liquid MRS broth and identified through 16S rdna sequencing. L. 8

10 Page 9 of 34 Canadian Journal of Microbiology plantarum was the most ubiquitous isolate obtained and was utilized in SSF of DDGS alongside fungal SSF. Solid State Fermentation: Wheat DDGS was knife milled and fractionated using sieves in order to reduce the effect of particle size on analysis. A fraction with an average particle size of 595µm was sterilized via autoclave at 120 C for 15 min. Three-day seed cultures of bacterial and fungal strains were individually prepared by inoculation of 50mL of MRS Broth and potato dextrose media respectively in 250 ml Erlenmeyer flasks, followed by 72 hr incubation at 30 C and shaking at 150 rpm. Ground DDGS was autoclaved, and urea was added at 1.2 % (w/w; dry weight basis) to it in order to enhance cellulose and particularly xylanase productions by filamentous fungi (Yang et al. 2012). Fifty grams of urea-fortified, sterile, ground DDGS were individually inoculated with 16 ml of seed culture of a single strain, and the moisture was adjusted to 75% moisture. The samples were fermented at 30 C for 7 days, and were mixed once per day using aseptic technique. A control sample was prepared identically to fermented DDGS, but without seed culture inoculation. The samples were prepared inside of containers with small holes placed all around the lid to allow moisture and gas transfers during SSF. These containers were all placed into a larger container, ensuring to not obstruct ventilation holes. Small containers of water were also placed inside of a larger plastic container to maintain high humidity. The relative humidity was 75% (±7% according to manufacturer s specifications), which was measured using Accu-Temp Brand Humidity Meter. At the conclusion of SSF, all of the samples were frozen at -4 C. Samples were thawed prior to larval addition by allowing equilibration with room temperature. BSFL Digestion: 9

11 Page 10 of BSFL were purchased from Supercricket (Prince Albert, SK) and Bugs4Pets.com (Vancouver, BC), and maintained in the closed cage for a few generations. One hundred grams (wet weight basis (wwb); approximately larvae, 5~10 days after hatching), of larvae were placed into the thawed SSF treated and control DDGS. At days 0, 3, 6, 9, and 12 of larval digestion a portion of feed beds was removed, and larval biomass was separated from the spent feeds. This spent feed was tested with the followings: moisture content of the substrate; acidic and enzymatic hydrolysis of the extracted substrate for sugar liberation; and cellulose content. The moisture content was measured by placing ~1 g of substrate into pre-weighed microfuge tubes, and drying overnight at 80 C to a constant weight. The loss in weight of the sample is equal to the moisture content of the sample. Cellulose Content Assay: The cellulose content was measured according to the methods of Brendel et. al. with minor modifications (Brendel et al. 2000). Approximately 1.5 g of substrate (wwb) was loaded into pre-weighed 25 ml centrifuge tubes with caps. Acetic acid (80%; v/v) was added at 2.0 ml/ g (wwb) substrate, and nitric acid (69%; v/v) was then added at 0.2 ml/ g (wwb) substrate. These acids served to de-lignify the substrate and hydrolyze non-cellulosic compounds. The method has been shown to produce little to no cellulose degradation. The hydrolyzed substrate was then cooled, and mixed with 2.5 ml of 95% ethanol. The tubes were then sealed, mixed by vortexing, and centrifuged for 5 minutes at 5000 rpm (Sorval SS-34 rotor) to pelletize the cellulose. After the initial ethanol wash, the following washes were used to remove the excess nitrogen content from the nitric acid, as well as the remaining hydrolyzed substrate constituents: 5 ml 95% ethanol, 5 ml deionised water, 5 ml 95% ethanol, and finally 5 ml acetone. Each wash was done in two 2.5 ml parts. The first volume was added, the substrate was mixed via 10

12 Page 11 of 34 Canadian Journal of Microbiology vortexing, and then the second wash was used to wash the walls of the vessel free of substrate. After each wash, the sample was centrifuged, and the supernatant was removed via vacuum suction. Following the acetone wash, the tubes containing the pelletized sample of α-cellulose and the cap were dried to a constant weight overnight at 80 C. The weight of the dried sample was equal to the cellulose content in the starting material. Proximate Analysis of DDGS and BSFL: Proximate analysis was conducted on BSFL, DDGS, post-ssf DDGS, and DDGS at 6 and 12 days of larval digestion by measuring the moisture content, crude lipid, crude protein, ash content, and total carbohydrates by difference. The moisture content was measured by placing ~ g of substrate into pre-weighed microfuge tubes, and drying overnight at 80 C to a constant weight. The loss in weight of the sample is equal to the moisture content of the sample. Crude lipid was measured according to AACC method , Crude Lipid in Grain and Stock Feeds (AACC International, 2009). The samples were dried to 0% moisture prior to analysis, and g of sample was assayed in duplicate using Goldfisch extraction for 6 hr with petroleum ether as the extraction solvent. The crude lipid was equal to the weight of the lipids extracted after removal of the solvent. The ash content was measured according to AACC method , Ash-Basic Method (AACC International, 2009a). Triplicates of g of sample at 0% moisture was incinerated in a muffle furnace at 550 C for 24 hrs. The ash content was equal to the weight of the non-combusted material after total combustion and cooling of the sample. Crude protein content was measured according to AACC method , Crude Protein- Combustion Method (AACC International, 2009c). Ssample (0.2 g) was combusted in duplicate, and the resulting N was measured. A conversion factor of 6.25 was used to calculate total protein in the sample. 11

13 Page 12 of Glucose Yields from DDGS: Overall flow of SSF-BSFL treatments and glucose determination is summarized in Figure 1. The free glucose amount was measured by suspending the spent feed at 60 mg /ml in water. The suspension was vortexed for 1 min and centrifuged to recover the supernatant. The acid hydrolysis conditions were as follows: the pellet from free sugar extraction was suspended at 60 mg/ml in 2% HCl, and the suspension was incubated at 90 C for 3 hours. The suspension was centrifuged (12,000 g, 20 min, room temperature) to separate the suspension into liquid and solid parts. The amount of glucose was measured via hexokinase/glucose-6-phosphate dehydrogenase system (Worthington, 1993). The NADH generated by this process is linear in concentration to amount of glucose and was measured through spectrophotometric absorption at 340 nm. The results indicate glucan hydrolysis; mainly the amount of amorphous cellulose on the cellulose particles and starch. The solid matter after acid treatment was washed with water several times to a neutral ph. The solid parts represent the remaining crystalline cellulose which has been micronized from the BSFL mastication. The solid part was loaded into flasks and suspended 1% (wt. substrate/v.) in sodium acetate buffer (ph 4.8). Cellulase was then added at 20 mg enzyme/g (wet-weight basis) of solid. Tetracycline was added for bacterial control at 40 µg/ml. The hydrolysis was carried out in an incubator shaking at 50 C and 150 rpm for forty-eight hours. The glucose yields from cellulose processing were again measured via the hexokinase/glucose-6-phosphate dehydrogenase system. This enzymatic step liberates the glucose from the crystalline fraction of the cellulose particle. The sugar amounts of the two stages in addition to free pre-hydrolysis sugars are considered as the total amounts of sugar that can be obtained through the BSFL culture. 12

14 Page 13 of 34 Canadian Journal of Microbiology

15 Page 14 of Results: The length of SSF is set from the results of preliminary experiments of solid-state fermentation for 25 days of period with a series of moisture contents. At time intervals, the DDGS fermentation was observed via visual inspection for the following: DDGS coverage, hyphal growth, sporulation presence, the moisture content, cellulose content, crystallinity index of the extracted cellulose, and acidic and enzymatic sugar liberation (data not shown). Based on the results of the preliminary SSF experiment Trichoderma koningii, Aspergillus niger, Aspergillus fumigatus, and Phanerochaete chrysosporium fungal strains, and Lactobacillus plantarum isolated from the spent larval feeds were individually used to ferment DDGS at 30 C for 7 days, then the individual lot of fermented DDGS was introduced with BSFL. The analyses of nutrient profiles changes were conducted for fresh SSF-DDGS and for spent feeds after SSF-DDGS was fed to BSFL. The items analyzed included: 1) proteins, lipids, carbohydrate and ash contents, 2) the amount of DDGS before and after BSFL culture, and 3) amounts of recoverable glucose from SSF-DDGS before and after BSFL culture. The proximate are given in the nutrient ratios in each sample, and sugar recovery was converted to the amount recoverable from 100g of SSF-DDGS. Table 1 shows changes in the nutrient compositions of fresh SSF-DDGS before larval digestion treatments. Control DDGS in Table 1 is the DDGS treated under the same conditions (i.e., at 30 C for 7 days) without inoculation of SSForganisms. Comparison of each freshly fermented DDGS shows the SSF effects with different strains on nutrient components. SSF treatments had negligible effects of the ash content of DDGS, but resulted in a decrease in the relative carbohydrate content from 56.1% (Control DDGS) to between 43.8% (A. niger SSF) and 51.2% (L. plantarum SSF). Among the organisms studied, bacterial SSF consumed the lowest amount of DDGS carbohydrates. SSF increased the 14

16 Page 15 of 34 Canadian Journal of Microbiology relative protein content from 37.3% (Control DDGS) to between 40.7% (P. chrysosporium SSF) and 46.2% (A. niger SSF). SSF also increased the relative lipid content of DDGS (1.9%) to between 2.7% (A. niger SSF) and 5.3% (P. chrysosporium and T. koningii SSF). These results indicated that SSF on DDGS used a part of carbohydrates and increased the nutrients to support BSFL growth. The effect of SSF on DDGS was dependent upon the SSF organism used. Figure 2 shows the residual nutrient profiles in the spent feeds after 6 and 12 days of larval culture with the estimated amounts of residual feeds. As shown by the height of bar graphs in Figure 2, the dry matter of DDGS was reduced in every sample during BSFL digestion, indicating that BSFL could use SSF-DDGS as nutritious feeds no matter which strains were used for SSF. This change in utilization rates corresponded to the growth of BSFL during 12 days of BSFL culture where BSFL went into a prepupal stage between 9 to 12 days. The proximate in spent feeds are shown as percentages with in the residual amount. The total amount of spent feed decreased through BSFL culture because proteins, lipids and starch were assimilated while lignocellulose and ashes were left by BSFL. Residual nutrient profiles show that ratios of carbohydrates increased through the reduction of protein and lipid contents, i.e., concentrating carbohydrates. The rates of consumption differ among types of digestible nutrients, resulting in their shifting relative abundances over the digestion period. As these changes of ratio depended on SSF treatments, it can be concluded that the choice of SSF organism can change the digestive capabilities of the BSFL. The ash ratio in spent feeds was increased during larval digestion in all samples except A. fumigatus. Since ashes could not be assimilated as much as organic nutrients, this increase is a reflection of utilization of nutrients in DDGS feed by BSFL. In fact absolute amounts of ash stayed similar during BSFL culture. In every sample the protein contents decreased during larval digestion. T. koningii, A. niger, and L. plantarum treatments allowed 15

17 Page 16 of BSFL to utilize more portions of proteins in the feed compared to that of the control DDGS sample, indicating the likely ligno-cellulosic matrix were loosened to release proteins for larval digestion systems. For example, T. koningii-treated DDGS contained 42.2% protein before BSFL introduction (i.e., Day 0), and it reduced to 23.0% during 12 days BSFL growth; whereas unfermented DDGS samples showed the consumption of proteins from 37.3% of Day 0 to 26.0% of Day 12. The lipid ratio in SSF-treated DDGS did not shift considerably or predictably during the twelve-day larval digestion. The carbohydrate content of residual feeds increased during larval digestion except P. chrysosporium treated DDGS. While proteins, lipids, and carbohydrate profiles changed as above, cellulose is indigestive and remained in the spent feeds in a similar manners as ashes. The BSFL possess a wide range of digestive enzymes, primarily in the gut, which facilitate the hydrolysis and digestion of lipids, starch, and proteins with strong activity. They possess comparably weaker cellulolytic ability (β-glucosidase) (Li et al. 2011). This uneven digestion of DDGS nutrients resulted in a gradual concentration of cellulose during larval digestion as seen in Figure 2. The absolute amount of cellulose were unchanged during BSFL, thus the relative cellulose ratio of DDGS feed was significantly increased during BSFL digestion. Trichoderma koningii SSF- DDGS, for example, increased cellulose contents from 11.3% to 33.5% during 12 days of BSFL feeding. All combinations of SSF and twelve-day BSFL digestion resulted in a higher final DDGS cellulose concentration compared to those before BSFL digestion. The spent feeds were then analyzed for the recoverable glucose with three-step sugarrecovery (Figure 1): free-glucose, diluted acid hydrolysis and enzymatic hydrolysis. The recovery of glucose is calculated as the amount of glucose recovered from 100g of SSF-DDGS to allow direct comparison among samples. SSF gave an increase in the free-glucose in DDGS 16

18 Page 17 of 34 Canadian Journal of Microbiology (free sugar in Table 2). All SSF-processing (Days 0 in Table 2) resulted in higher free-glucose compared to the Control DDGS sample, except for T. koningii, which had nearly identical values to Control DDGS. Larval digestion significantly increased the amount of free glucose in DDGS over the twelve-day digestion period. This increase was the fastest and greatest for A. niger treated DDGS (1.0 to 4.3 g of 100g SSF-DDGS). Dilute acid hydrolysis is capable of liberating glucose from starch and some cellulose structure particularly susceptive to dilute acid hydrolysis. Due to the tight packing and ordered structure of crystalline cellulose regions, crystalline cellulose is not very susceptive to acid hydrolysis; whereas amorphous cellulose is primarily hydrolyzed among cellulose fibres during dilute acid pre-treatment (Noureddini & Byun, 2010). The Acid-liberated column of Table 2 shows the results of dilute acid hydrolysis of SSF treated and BSFL digested DDGS. The acidic hydrolysis glucose yields in the Control DDGS samples were significantly decreased by the third day of BSFL digestion (3.4 to 0.9 g/100g SSF-DDGS; Table 2). After SSF treatment (Day 0 in Table 2), all strains except L. plantarum showed lower glucose yields than the Control DDGS samples in sugar recovery using acid hydrolysis. It indicated that all strains except L. plantarum utilize starch or amorphous cellulose during the SSF period. The sugar yields with acid hydrolysis was not changed through BSFL culture lengths for those fungal SSF treated DDGS since a majority of amorphous cellulose has already digested in SSF-treatments as discussed above. These observations indicated that the acid-hydrolysis susceptive glucans (presumably mostly starch, which was already consumed in SSF processes) in Control DDGS were rapidly digested by the insect, leaving little susceptive glucans to acid hydrolysis at Day 3. Moreover, despite an increase in the cellulose content of the Control DDGS sample during BSFL digestion, 17

19 Page 18 of no significant increases in acidic hydrolysis during this time were seen in Control DDGS (Table 2). The sugar yields from enzymatic hydrolysis of SSF-DDGS before and after BSFL digestion are also shown in Table 2. Since amorphous cellulose and free glucose were removed in the preceding analysis, the results of this enzymatic hydrolysis correspond to the hydrolysis of crystalline cellulose. The amount of glucose liberated from Control DDGS did not significantly increase during larval digestion (Day 3 ~ 12). SSF treated DDGS samples substantially increased the susceptibility of cellulose to enzymatic hydrolysis which resulted in higher glucose yields. P. chrysosporium SSF treatment increased glucose yields (5.2 g/100g SSF-DDGS), which were greater than seven times of the glucose yield from the Control DDGS (0.7 g/100g SSF-DDGS). All SSF treated DDGS (Day 0 before BSFL introduction) had higher glucose yields than the control DDGS sample indicating that SSF reduced the crystal packing of cellulose and increased the porosity of cellulose particles allowing for greater enzyme access. BSFL mastication further increased in the susceptivity of cellulose. Physical disruption by BSFL digestion has been shown to open feed structural matrices and increase the surface area of celluloses thereby increasing susceptibility towards enzymatic hydrolysis (Lynd, 1996). As expected, BSFL digestion of SSF treated DDGS generally increased enzymatic hydrolysis glucose yields up to Day 6, where after a slight decrease was seen until Day 12. This change in the utilization rate corresponded to the growth of BSFL during 12 days of BSFL culture where BSFL went into prepupal stage between 9 to 12 days. A summation of the glucose recovered from free-glucose, acidic hydrolysis, and enzymatic hydrolysis is shown in the total sugar recovery of Table 2. BSFL digestion of Control DDGS (non-ssf treated DDGS) results in a slight increase (4.8 to 6.5 g/100g SSF-DDGS) in 18

20 Page 19 of 34 Canadian Journal of Microbiology the total glucose yields from DDGS and most of the increases are due to free glucose and an increase in the cellulose content which provided greater enzymatic hydrolysis glucose yields. SSF alone increased total glucose yields compared to Control DDGS (4.8g) for P. chrysosporium (8.8 g) and L. plantarum (10.7 g) from 100 g of SSF-DDGS. While effects of P. chrysosporium and L. plantarum SSF were mainly associated with SSF itself (accumulation of free glucose), T. koningii, A. fumigatus, and A. niger allowed higher susceptivity of cellulose after BSFL digestion (micronization of cellulose) as indicated by sugar yields from acid- and enzymatichydrolysis. Except for A. fumigatus day 12, combination SSF and BSFL digestion surpassed the comparable days control sample in total glucose yields on days 6, 9, and 12. The greatest total glucose yields were seen with a combination of SSF and BSFL digestion: SSF using the companion bacteria L. plantarum and six days of larval digestion resulted in the single greatest glucose yield of 10.7 g/ 100g SSF-DDGS which is a 2.3-time higher yield than untreated Control DDGS

21 Page 20 of Discussions: The proximity analysis of spent feeds (Figure 2) showed decreases in residual nutrients through BSFL culture, i.e., consumed by BSFL to build their body mass. These results indicated that changes in nutritional profiles during SSF (Table 1) allowed the efficient bioconversion of DDGS organic substances into larval biomass. Trichoderma koningii, A. niger, and L. plantarum fermentation increased in proteins and lipids, and resulted in higher degrees of protein consumption compared to untreated DDGS after twelve days of larval digestion. As we hypothesized, combination of SSF and BSFL digestion represent a viable means of concentrating cellulose (Figure 2) which may be used for glucose recovery (Table 2); while simultaneously bioprocessing organic matter from DDGS into insect biomass. In addition to concentrating cellulose in DDGS, SSF should weaken the cellulose structure, allowing for a greater yield of glucose from DDGS following BSFL digestion. Hoskins and Lyons (Hoskins & Lyons, 2009) showed that adding small amounts of SSF treated DDGS back into the saccharification process during original bioethanol production can significantly increase the final ethanol yield from the grain. They discussed it was due to the increase of cellulose and starch exposure to saccharification. Also we observed DDGS became more slurry-like appearance as BSFL consume SSF-DDGS, suggesting micronization of particles happened through larval mastication. These factors suggested that more glucose should be recovered from spend DDGS feeds after the combined treatment of SSF and BSFL. In order to examine the susceptivity of cellulose in spent feeds, we conducted the sugar recovery from spent feeds. We considered the forms of recoverable glucose in three states in the spent DDGS feed: free glucose, amorphous cellulose and crystalline cellulose. The amount is calculated to show the amount recoverable from 100 g of SSF-DDGS. We assumed that the SSF treatment and larval 20

22 Page 21 of 34 Canadian Journal of Microbiology digestion produce the cellulase/amylase activities to hydrolyze free cellulose and residual starch to yield glucose. The cellulose structure is micronized through SSF and larval mastication which together, change the ratio between amorphous and crystalline cellulose. Amorphous cellulose should be susceptive to diluted-acid hydrolysis, and crystalline cellulose that is micronized by larval mastication should be more susceptive to enzymatic hydrolysis than hard-packed crystalline cellulose in the untreated DDGS. Thus we analyzed the glucose liberation in three steps: free glucose, diluted acid hydrolysis for amorphous cellulose, and enzymatic hydrolysis of crystalline cellulose (Figure 1). Asperigillus niger is regarded as an excellent producer of β-glucosidase, an enzyme which cleaves the glycosidic bond between two β-d-glucose monomers resulting in the production of glucose (Lynd, 1996). In general, increases in DDGS free-glucose concentrations are likely the result of physical shearing of glucans (both starch and cellulose) during micronization of the material by the powerful mouthparts of the BSFL. During SSF and BSFL culture, more susceptive amorphous cellulose and residual starch are likely degraded and hydrolyzed by SSF strain enzyme and BSFL digestive system. As a result of these biological treatments, DDGS have free glucose, presumably through hydrolysis of more susceptive amorphous regions of cellulose. Since L. plantarum did not utilize starch or amorphous cellulose during the SSF period, the high glucose yields in L. plantarum-ddgs and Control DDGS (3.6 and 3.6 g/100 g SSF- DDGS, respectively; Table 2) were the results of acidic hydrolysis of these constituents unused during SSF periods. Meanwhile fungal SSF treated DDGS produced lower amounts of glucose (1.0 to 1.9 g/100 g SSF-DDGS) from acidic hydrolysis, which indicates that fungal organisms hydrolyzed a great deal of starch during SSF (prior to BSFL introduction). This also explains 21

23 Page 22 of why Control DDGS had lower free-glucose concentrations than SSF treated samples. During the SSF period, fungi hydrolyzed starch into its component monomer glucose increasing the freeglucose in the sample to a level above the Control DDGS where no hydrolysis of starch is expected. Enzymatic-hydrolysis could recover a lower amount of glucose from control DDGS sample (Table 2). It indicated that mastication of Control DDGS by the BSFL did not micronize crystalline or convert crystalline cellulose into amorphous cellulose at a rate which exceeded the digestive abilities of the BSFL to remove those amorphous regions. Thus it is likely SSFprocesses can loosen the ligno-cellulosic matrix. Acid-hydrolysis of L. plantarum SSF-DDGS yielded much higher sugar than the Control DDGS before BSFL digestion (Table 2). These higher glucose yields from acid hydrolysis indicate that L. plantarum SSF increased the susceptibility of cellulose and starch to acidic hydrolysis. In contrast, fungal SSF treated samples generally showed an increase in acidic hydrolysis glucose yields during BSFL digestion. When comparing across digestion days it is seen that the acidic hydrolysis glucose yields from SSF treated samples exceeded the control DDGS sample yields during larval digestion. This was due partially to differences in cellulose content, but largely to disruption of the cellulose structures during SSF. These observations indicated that the fungal growth made hard cellulose matrices susceptive to BSFL mastication and made crystalline cellulose acid-hydrolysis susceptive, i.e., into amorphous cellulose. The enzymatic hydrolysis of spent feeds yielded increase in the glucose up to Day6 and then decrease towards Day 12 (Table 2). It indicates that enzymes remaining from the SSF process continue to have effect on DDGS fibre during larval digestion, but are eventually inactivated through larval action or environmental changes such as ph. SSF and BSFL digestion 22

24 Page 23 of 34 Canadian Journal of Microbiology of DDGS produced greater glucose yields than the control DDGS samples at every sample and day of BSFL digestion except for day 12 of A. fumigatus. These observations showed that the combinations of SSF and BSFL can increase in the recoverable glucose. The recovery of sugar was increased by 2.3-folds through utilization of BSFL on top of SSF. Utilization of DDGS to recover glucose which can be turned into bioethanol increases the efficiency of ethanol production from the starting grain. This can help to alleviate concerns over the utilization of crops for industrial purposes as opposed to foods. This result reflects that the combination of insect with SSF can use ligno-cellulosic materials (e.g., vast agro-byproducts) to recover glucose. We speculate, combined with the results reported by Hoskins and Lyons (Hoskins & Lyons, 2009), that the insect spent feeds could be thrown into the bioethanol fermentation to increase the total ethanol production per feedstock. Additionally it should be noted that the BSFL biomass contained protein (36.8%), lipid (23.4%), carbohydrate (32.9%), and ash (6.91%) (Table 1). This content of insect biomass indicated that a good part of nutrients in DDGS were recovered as insect biomass, which can be used as feeds or potentially as food source. The creation of multiple product streams from DDGS may have an economic incentive, especially BSFL biomass has good nutritive properties such as high protein contents in easy-to-digest forms. We are investigating further optimization of SSF-BSFL processes and are examining the fermentation of recovered sugar from SSF-BSFL treatment to prove the concept of ethanol production efficiency. We also look into the possibilities of nutrient recovery in the form of insect biomass to maximize the valorization of waste materials

25 Page 24 of References: AACC International (2009a) Ash-Basic Method: AACC International Approved Method doi: /aaccintmethods AACC International (2009b) Technical, A. in AACC International Approved Methods 1 2. doi: /aaccintmethod AACC International (2009c) Technical, A. in AACC International Approved Methods 1 2. doi: /aaccintmethod Adav SS, Ravindran A, Sze SK (2015) Quantitative proteomic study of Aspergillus Fumigatus secretome revealed deamidation of secretory enzymes. J. Proteomics 119(Supplement C): Bansal N, Tewari R, Soni R, Soni SK (2012) Production of cellulases from Aspergillus niger NS-2 in solid state fermentation on agricultural and kitchen waste residues. Waste Management 32(7): Brendel O, Iannetta PPM, Stewart D (2000) A rapid and simple method to isolate pure α- cellulose. Phytochem Anal 11(1): Hoskins B, Lyons M (2009) Improving Bioethanol Yield: The Use of Solid-State Fermentation Products Grown on DDGS. J. Institute of Brewing 115(1):

26 Page 25 of 34 Canadian Journal of Microbiology Li Q, Zheng L, Qiu N, Cai H, Tomberlin JK, Yu Z (2011) Bioconversion of dairy manure by black soldier fly (Diptera: Stratiomyidae) for biodiesel and sugar production. Waste Management 31(6): Lynd LR (1996) OVERVIEW AND EVALUATION OF FUEL ETHANOL FROM CELLULOSIC BIOMASS: Technology, Economics, the Environment, and Policy. Ann Review Energy Environ. 21(1): Makkar HPS, Tran G, Heuzé V, Ankers P (2014) State-of-the-art on use of insects as animal feed. Animal Feed Sci Technol 197(Supplement C): Noureddini H, Byun J (2010) Dilute-acid pretreatment of distillers' grains and corn fiber. Bioresource Technol 101(3): Schuster E, Dunn-Coleman N, Frisvad J, van Dijck P (2002) On the safety of Aspergillus niger - a review. Appl. Microbiol. Biotechnol. 59(4): Sheppard DC, Tomberlin JK, Joyce JA, Kiser BC, Sumner SM (2002) Rearing methods for the black soldier fly (Diptera: Stratiomyidae). J Med Entomol 39(4): Singh D, Chen S (2008) The white-rot fungus Phanerochaete chrysosporium: conditions for the production of lignin-degrading enzymes. Appl. Microbial. Biotechnol. 81(3):

27 Page 26 of Sun Y, Cheng J (2002) Hydrolysis of lignocellulosic materials for ethanol production: a review. Bioresource Technol 83(1): University of Saskatchewan s Wheat DDGS Nutrient Profile. (2010). at < Wood TM, McCrae SI (1978) The cellulase of Trichoderma koningii. Purification and properties of some endoglucanase components with special reference to their action on cellulose when acting alone and in synergism with the cellobiohydrolase. Biochem. J. 171(1): Worthington Enzyme Manual : Enzymes and Related Biochemicals (ed. Worthington, V.) (Worthington Biochemical Company, 1993) Yang S, Lio J, Wang T (2012) Evaluation of Enzyme Activity and Fiber Content of Soybean Cotyledon Fiber and Distiller' Dried Grains with Solubles by Solid State Fermentation. Appl Biochem Biotechnol 167(1): Zheng L, Hou Y, Li W, Yang S, Li Q, Yu Z (2012a) Biodiesel production from rice straw and restaurant waste employing black soldier fly assisted by microbes. Energy 47(1):

28 Page 27 of 34 Canadian Journal of Microbiology Zheng L, Li Q, Zhang J, Yu Z (2012b) Double the biodiesel yield: Rearing black soldier fly larvae, Hermetia illucens, on solid residual fraction of restaurant waste after grease extraction for biodiesel production. Renewable Energy 41(Supplement C):

29 Page 28 of Figure legends 537 Figure 1: Process Flow Chart for the Bioconversion of DDGS Solid state fermentation (SSF) and larval digestion by the Black Soldier Fly larvae both produce fermentable sugars through the hydrolysis of starch, and the hydrolysis of amorphous cellulose from cellulose particles. This increases the glucose concentration of High Cellulose Larval Waste here. Dilute acid hydrolysis produces glucose via the hydrolysis of starch, and of amorphous cellulose. Free glucose is subtracted from the glucose yields from dilute acid hydrolysis, so acid hydrolysis is primarily a measure of amorphous cellulose, and/or starch in the sample. Following dilute acid hydrolysis, enzymatic hydrolysis produces glucose by enzymatically cleaving it from crystalline cellulose. As free glucose, starch, and amorphous cellulose were removed from the substrate in the preceding steps, the enzymatic hydrolysis step is a measure of the sugars which can be recovered from the crystalline fractions of cellulose under the current system. The digested glucose, and other biomass components such as proteins are partially recovered in the form of high quality larval biomass Figure 2: The Changes of SSF-DDGS Feed During BSFL Culture The overall length of each bar graph shows the residual amount of spent feeds relative to the initial amount of SSF-DDGS feed before BSFL culture. Each bar is divided according to the ratio of nutrient contents: dotted area, carbohydrates; white area, proteins; black area, lipids; and hatched area, ashes. Gray bar insets in the carbohydrate contents (dotted area) indicate the ratios of cellulose, i.e., a carbohydrate ratio includes a cellulose amount in calculation. Each panel presents the results from: A, unfermented DDGS (control DDGS); B, Phanerochaete 28

30 Page 29 of 34 Canadian Journal of Microbiology chrysosporium-ssf; C. Trichoderma koningii-ssf; D. Aspergillus niger-ssf; E, Lactobacillus plantarum-ssf; and F, A. fumigatus-ssf

31 Page 30 of 34 1 Table 1. Residual nutrient proximity analysis in larval culture after solid state fermentation Days Ash Proteins Lipids Carbohydrates (% dwb) (% dwb) (% dwb) (% dwb) Total Cellulose Control DDGS 4.8 ± ± ± ± ± 0.9 Phanerochaete chrysosporium 5.4 ± ± ± ± ± 0.3 Trichoderma koningii 4.9 ± ± ± ± ± 1.4 Aspergillus niger 7.3 ± ± ± ± ± 0.9 Lactobacillus plantarum 5.4 ± ± ± ± ± 1.2 Aspergillus fumigatus 4.2 ± ± ± ± ± 3.0 Ref: Young Larvae 6.9 ± ± ± ± 4.7 1

32 Page 31 of 34 Canadian Journal of Microbiology Table 2. Sugar recovery from larval spent feed Sugar liberation (g from 100g DDGS) Day Free sugar *1 Acid- Enzymatic- Total liberated *2 liberated *3 sugar recovery Control DDGS a ± b ± a ± ± b ± a ± a ± ± b ± a ± a ± ± b ± a ± a ± ± c ± a ± a ± ± 1.6 Phanerochaete chrysosporium a ± ab ± c ± ± b ± a ± c ± ± bc ± ab ± c ± ± bc ± ab ± a ± ± c ± b ± b ± ± 0.6 Trichoderma koningii a ± a ± a ± ± ab ± a ± a ± ± c ± b ± a ± ± b ± bc ± a ± ± c ± c ± a ± ± 1.3 Aspergillus niger a ± a ± a ± ± b ± b ± bc ± ± c ± a ± c ± ± c ± ab ± ab ± ± c ± b ± bc ± ± 1.1 Lactobacillus plantarum a ± c ± a ± ± b ± b ± a ± ± c ± a ± a ± ± b ± ab ± a ± ± bc ± a ± a ± ± 1.9 Aspergillus fumigatus a ± a ± bc ± ± b ± a ± ab ± ± ab ± a ± c ± ± c ± a ± c ± ± c ± a ± a ± ± 1.9 Each data is a mean of three measurement, and means followed by different letters are significantly different at α = 0.05 by Fisher s LSD within groupings.