This is a pre-print of an article published in Waste and biomass valorization (Ed. Springer). The final authenticated version is available online at: https://doi.org/10.1007/s12649-018-0327-5 1 Optimization of down-stream for cellulases produced under solid-state 2 fermentation of coffee husk 3 4 5 6 Maria Marín, Adriana Artola, Antoni Sánchez* 7 Composting Research Group 8 Department of Chemical, Biological and Environmental Engineering 9 Universitat Autònoma de Barcelona 10 Edifici Q, Campus de Bellaterra, 08193- Cerdanyola del Vallès, Spain 11 12 *Corresponding author: Antoni Sánchez 13 Tel.: +34 935811809; Fax: +34 935812013 14 Email address: antoni.sanchez@uab.cat 15 16 17 18 Acknowledgements 19 The authors would like to thank the Spanish Ministerio de Economía y Competitividad, 20 which gave financial support (Project CTM2015-69513-R) to this work. Maria Marín 21 also thanks the Universitat Autònoma de Barcelona for a predoctoral scholarship. 22 1 23 Abstract 24 This work systematically studies the downstream process of the solid state fermentation 25 (SSF) of a mixture of coffee husk and wood chips, inoculated with compost, for 26 cellulase production. Downstream of SSF (at pilot scale) remains as one of the less 27 studied stages of the process, being critical in technical, environmental and economic 28 terms. In this study, the specific downstream points considered were: i) enzyme 29 extraction yield, in terms of extraction ratio solid:solvent, agitation mode and solvent 30 type; ii) enzymatic activity recovery of the lyophilised extract and iii) efficiency of 31 consecutive extractions. Results indicate a maximum activity recovery of 108 ± 30% in 32 the extraction performed at ratio 1:5 solid-solvent, in static mode and with distilled 33 water. Statistical analysis revealed a high dispersion of the results and needs to be 34 considered to extract consistent conclusions in any downstream of SSF. Lyophilisation 35 demonstrated to be an adequate technology for enzymatic activity preservation. 36 Regarding consecutive extractions, yield recovery in the first and second extraction 37 maintain a similar value. In a framework of a zero-waste enzyme production process, 38 different strategies have been tested for the remaining solid after extraction. 39 Respirometric tests reveal that it is possible to aerobically stabilize the remaining solid 40 obtaining a compost like material, whereas anaerobic digestion resulted in low methane 41 yields (51±3 mL methane g-1 VS). 42 Keywords: Cellulase extraction, activity recovery, solid state fermentation, down- 43 stream, zero waste process. 44 2 45 Statement of novelty 46 47 Nowadays, most of the studies carried on in the field of enzyme production through 48 solid state fermentation are focused on optimizing the production yield at a laboratory 49 scale. Extraction is mainly performed using a high ratio solid:solvent in the range of 1 50 to 15 mL per g, suitable for a laboratory scale but non-viable at an industrial one. The 51 present work differs from the others published so far as it focuses on the aspects related 52 to the downstream stage to be implemented at an industrial scale, like the optimization 53 of the extraction conditions in order to save resources, management of the remaining 54 solid after the extraction and conservation of the final product in low-cost mode. 55 The fact that the production process itself presents many advantages, as it has been 56 proved to be reproducible at a pilot scale, requires a low investment due to the use of a 57 residue as raw material, the equipment is simple and the production yield is high, makes 58 the whole process interesting for implementation at an industrial scale. 59 The scale of the experiments performed in this work allow to apply the results to the 60 design of an industrial plant for the production of low cost cellulase from coffee husk 61 and to its economic assessment, as well as the conservation of the cellulases by standard 62 ways. The remaining waste is also considered as a source of biogas or compost in a zero 63 waste global strategy. 64 3 65 Introduction 66 Almost 85% of the energy consumed worldwide was obtained from fossil fuels during 67 2016 [1]. The excessive use of fossil fuels during the last years has produced an 68 increment of CO2 level in the atmosphere, which is nowadays reaching an historical 69 maximum of 409 ppm. This rise has a pernicious impact on the global climate creating a 70 greenhouse effect and increasing temperatures around the globe. In the search for new 71 renewable sources of energy, bioproducts like bioethanol are gaining relevance. 72 However, its production process must be optimized in order to achieve a major use 73 compared to fossil fuel. Among the different stages of bioethanol production, hydrolysis 74 of the lignocellulosic material can be performed chemically, by an acid or alkaline 75 reagent, or enzymatically using cellulases, which results in less harsh operational 76 conditions, since most of the enzymes require mild operational conditions, avoiding 77 great energy expenses and corrosion issues [2]. In fact, one of the main uses for 78 cellulases is the conversion of cellulosic and lignocellulosic materials to bioethanol [3, 79 4]. For this reason, finding a low-cost and robust process for cellulase production is a 80 key point in bioethanol manufacturing. 81 Cellulases have been in the market for more than 30 years, and they are used in 82 all type of industries. As an example, cellulases have a role in textile industry as 83 biopolishers of fibres or in the pulp and paper field in enzymatic deinking and to 84 improve brightness of the product. Microbial production is the most used method in 85 industry for cellulase production. A wide variety of microorganism like fungi [5], yeast 86 or gram-positive [6] and gram-negative bacteria can be used in the production of 87 cellulase. As a result, the morphology and specificity of the enzyme will depend on the 88 substrate, operational conditions and microorganism used in the production process [7]. 4 89 Two main methods are used for cellulase production, submerged fermentation 90 (SmF) and solid state fermentation (SSF). Both methods show advantages and 91 disadvantages. No gradients of temperature and media composition are found in 92 submerged fermentation. However, solid state fermentation requires smaller and 93 cheaper equipment, which makes it economically attractive; besides, the option of using 94 a lignocellulosic solid waste as a nutrient source and support for the growing of the 95 microorganism increases the profit obtained in an industrial process and reduces the 96 waste disposed into the environment. Several types of agro industrial wastes have been 97 reported as sources of cellulose: Sun et al. [8] used banana peel obtaining a cellulase 98 activity filter paper unit (FPU) of 5.56 FPU per gram of dry matter; Dhillon et al. [9] 99 also reported a cellulase production of 133.68 U per gram of dry substrate from apple 100 pomace using Aspergillus niger NRRL-567 as inoculum. Other residues used as 101 substrate are wheat straw, corn fibre or seaweed [10, 11]. The mentioned cases were all 102 reported at laboratory scale. Few processes have been developed at a scale large enough 103 to be used for industrial purposes, especially when considering all the downstream 104 process. Cerda et al. [12] reported a cellulase production of 9 FPU of enzyme activity 105 per gram of dry matter from coffee husk, a residue obtained after the thermal peeling of 106 the coffee bean. In that case, solid state fermentation was carried on under non-sterile 107 conditions and no temperature control at a 4.5 L scale, allowing the consortia of 108 microorganisms present in the residue to develop. Maximum production was observed 109 after 48 hours of process. Also, Cerda et al. [12] developed a strategy to obtain a 110 continuous stable production of cellulase from the mentioned residue. This process was 111 tested at pilot scale, obtaining higher temperatures, which caused a decrease in the 112 production, but showing that this strategy was valid for scaling up. These results 113 confirmed coffee husk as a promising low-cost source of cellulase, as it cannot be used 5 114 as animal feedstock, according to local regulation, and the only possible use would be 115 composting, less profitable than enzyme production. 116 One of the higher expenses required in solid state fermentation is the extraction 117 of the final product from the solid matrix. In this specific case there is no information 118 regarding recommended extraction agent, ratio solid:liquid for the extraction or 119 agitation method. The optimization of these parameters is not a key point at laboratory 120 scale but at industrial scale will determine the profitability of the process. Also, after the 121 extraction, the solid waste remains, which still can be used in a more profitable way 122 than landfilling. Thus, considering these facts, the current work aims to study different 123 aspects of the downstream process for the production of cellulase from coffee husk, 124 including optimal extraction conditions, taking into account different extraction agents, 125 agitation methods and solid:liquid ratios. Also, the dispose of the solid residue after 126 extraction has been studied, considering its use as feedstock for anaerobic digestion and 127 biogas production, along with the composting of the waste to obtain an organic 128 amendment. Storage of the extracted enzyme was also studied by determining the 129 activity loss of the lyophilised product stored at 4 °C compared to its initial activity. 130 Materials and methods 131 Raw material 132 Coffee husk was provided by Marcilla S.A (Mollet del Vallés, Barcelona, Spain). 133 Coffee husk or silverskin was produced during coffee grains roasting. The main 134 physico-chemical characteristics of this waste are summarized in Table 1. The residue 135 was collected within 24h after its production and stored at -20 °C until needed. During 136 SSF, coffee husk was mixed in a proportion 9:1 (w:w) with compost provided by the 137 waste biological treatment plant of the Consorci per a la Gestió dels Residus del Vallès 6 138 Oriental (Granollers, Barcelona, Spain). Compost was added to increase the initial 139 microbial population. 140 The mixture of compost and coffee husk was mixed with wood chips as bulking 141 agent to provide an adequate porosity to the final mixture (CH). Wood chips were 142 obtained from Trabede Jorba composting plant (Jorba, Barcelona, Spain), and mixed in 143 a ratio 1:1 (v:v). Moistening of the mixture was not necessary since its value was 144 always near 60% in weight. The suitability of this industrial residue as raw material for 145 cellulase production through SSF has been described for the first time by Cerda et al. 146 [12], where details of the process and cellulase production range are thoroughly 147 provided. 148 149 SSF Materials and experimental set up 150 SSF was carried out in 4.5 and 10 L (working volume) adiabatic reactors, as described 151 in detail in previous works [13, 14]. Both types of reactors were identical in operation. 152 Briefly, air flow was adjusted by an air flow controller (Bronkhorst Hitec, The 153 Netherlands), provided to the bottom of the reactor and evenly distributed by a net 154 placed below the biomass. The oxygen content of output gases was measured by an 155 electrochemical O2-A2 oxygen sensor (Alphasense, United Kingdom) and data were 156 collected using a personal computer HP Compaq LA1951g Intel core ISvpro (Hewlett- 157 Packard, USA), equipped with Ubuntu operation system. Data analysis was carried on 158 by a non-commercial tailor-made software. Temperature was also measured 159 (Transmisor CCPI/T-120, Seneca, Italy). Air flow was supplied to the reactors assuring 160 that oxygen was always above 10% in output gases through a feedback control system 161 in which two oxygen content set points were fixed: 11.5% as a minimum and 12.5% as 162 maximum. Two values of air flow are set by the user, when oxygen concentration is 7 163 lower than 11.5%, the system selects the maximum air flow, when oxygen content is 164 higher than 12.5%, the system selects the minimum air flow. Reactors capacity (mass of 165 mixture coffee husk, compost and bulking agent) was 2.7 kg for 10 L reactors and 1.5 166 kg for 4.5 L reactors. 167 168 Cellulase activity recovery – Downstream 169 Cellulase activity determination – Reference value 170 To obtain the total cellulase activity, an extraction in agitated mode at 1:15 solid: 171 solvent ratio and buffer citrate mono hydrated 0.05M solution at pH 4.8 (BC) as solvent 172 was performed during half an hour. Pirota et al. [15] performed different experiments 173 for cellulase extraction using ratios of 1:3, 1:6 and 1:9 finding that 1:9 was the optimal 174 extraction ratio. Also Chandra et al. [16] found an optimum extraction ratio for β- 175 endoglucanase at 1:4 ratio w:v from fermented wheat bran. In the case of coffee husk 176 we have observed that an extraction ratio of 1:15 w:v guarantees a good contact 177 between solid and liquid to ensure a total extraction of the enzyme [12]. 178 Activity recovery was calculated referring the activity obtained in the extraction 179 at given conditions as a percentage of the activity in the standard extraction, performed 180 at 1:15 ratio, using BC and orbital agitation, according to equation 1. 181 100 x Cellulase activity (U g-1 DM) 182 % Activity recovery = (Eq 1) -1 Standard Cellulase activity (U g DM) 183 184 Since various replicates of extractions under same experimental conditions (6 or 5 185 depending on the case) were performed, the rule of 25% was applied, discarding the 186 values that differ more than 25% from the average. 8 187 188 Extraction parameters 189 Extractions were performed to evaluate the recovery of cellulase activity under different 190 experimental conditions. 191 The experimental parameters assayed were solvent type, extraction mode and 192 w:v solid:solvent (w:v) ratio. The variations of each parameter were: 193 a) Solvent type: Two different solvents were used to perform extractions, distilled 194 water (DW) and BC. 195 b) Extraction mode: The extraction modes were three: 1) no agitation (static mode), 196 ii) orbital agitation at 120 rpm, and iii) circulation of the solvent through a 197 column packed with the waste at 140 mL min-1 (8.4 kg h-1), which resulted in the 198 renewal of the whole volume of liquid 7 times during the extraction time. 199 c) Waste to solvent ratio (weight:volume): Assayed ratios of solid:solvent were 1:2, 200 1:3 and 1:5. 201 The different combinations of these tested variables are shown in Table 2. 202 Between 150 and 200 g of solid from SSF after 48h of fermentation were used in 203 every extraction. The solid and the solvent were in contact during half an hour under the 204 different regimes stated above and then the liquid phase decanted and centrifuged 205 during 10 minutes at 10000 rpm. Cellulase activity of the supernatant was measured and 206 referred to the value considered the total cellulase activity produced. 207 Experimental set up for static and agitated mode extraction consist of a set of 208 plastic containers of equal dimension and shape where solvent and solid were placed. 209 Static extraction was performed placing the fermented solid and the chosen volume of 210 solvent in the plastic container and waiting for half an hour. To obtain the extracted 211 liquid in agitated mode, the same procedure was followed but plastic containers were 9 212 placed in a Sanyo orbital shaker IOC402.XX1.C (Sanyo UK) incubator for half an hour 213 at 120 rpm and room temperature. 214 The experimental set up used in liquid recirculation extraction was composed of 215 a peristaltic pump Watson-Marlow 403U/L2 Ultra compact twin channel pump with 216 variable rotor speed from 0 to 99.9 rpm (Watson-Marlow Alitea, England) and a 2 L 217 plastic vessel with two adaptors, one at the top and one at the bottom, connected by 218 plastic tubes to the pump input and output. Inside the vessel, a device was coupled to let 219 the water fall over the biomass in a drop-shower mode, in order to achieve a 220 homogenous contact with the solid in a percolation mode. Fermented material was 221 placed inside the vessel and each solvent added. The flowrate was set up in the pump 222 and liquid circulation lasted for half an hour. 223 224 Number of extraction stages 225 Enzyme recovery from consecutive extraction stages was assayed performing three 226 consecutive extractions to the same fermented material. Between extractions, the 227 biomass was drained to reduce its moisture as much as possible. 1:2 solid:solvent (w:v 228 ratio) was chosen for these experiments. Extractions were performed using the lowest- 229 cost agitation mode and extraction solvent, which are static mode and DW as solvent, 230 respectively, in order to evaluate if this strategy implied an increase in activity recovery 231 with no extra cost. 232 233 Lyophilisation 234 Supernatant from extractions at 1:2 and 1:5 solid:solvent ratios, in static mode and using 235 BC as extraction agent were used in this experiment. According to Farinas et al. [17] 10 236 optimum cellulase pH regarding stability is between 4 and 5.5, so BC was selected to 237 preserve the enzyme stability. 238 Samples of 10 ml of the supernatant obtained from the centrifuged extract were 239 frozen at -80°C inside plastic falcons and stored at same temperature. On the other 240 hand, 500 mL of same supernatant were frozen at -80°C into two 250 mL beakers and 241 lyophilized using a Virtis 5L sentry lyophilizer 248627 (Virtis, Gardiner, USA) 242 connected to an Edwards vacuum pump RV5 A653_01_903 (Edward, United 243 Kingdom). The beakers used in the experiment were weighted before and after the 244 removal of the powder and the dry cleaning of the beaker. The obtained powder was 245 easily removed from the walls of the flasks so there were no perceptible losses. After 246 that, the remaining powder was mixed and pulverized using a ceramic mortar in order to 247 completely homogenize the solid. Mass of powder obtained from the 500 mL extract 248 was calculated by weigh difference, and the equivalent quantity to 10 mL of original 249 supernatant was dissolved in DW. The remaining powder was stored at 4 °C. 250 Cellulase activity for the redisolved solid and for an unfrozen sample of 10 mL 251 was measured for comparison using filter paper units (FPU) of activity for the 252 supernatant obtained after extraction for both 1:2 and 1:5 ratios. 253 254 Zero waste strategies 255 Biomethane Potential (BMP) 256 Biomass fermented for the production of cellulase was tested after extraction as 257 feedstock for biogas production through anaerobic digestion. Anaerobically digested 258 sludge from a municipal WWTP in Sabadell (Barcelona) was used as inoculum. 11 259 The methodology used was a modified method of the protocol described by the 260 German Institute for Standardization and reported by the Ordinance on environmentally 261 compatible storage of waste from human settlements [18], detailed by Ponsá et al. [19]. 262 According to this standard methodology, BMP tests were carried out under mesophilic 263 conditions and during approximately 20-30 days. Inoculum and biomass from SSF were 264 mixed in a feed to inoculum (F:I) final ratio of 0.5 and placed in hermetic closed bottles. 265 Each sample was tested in triplicate. The ratio F:I was calculated based on substrate and 266 inoculum initial amounts of volatile solids (VS). According to Ponsá et al. [19], this 267 ratio is the optimum to maximize biogas production. The content of volatile solids in 268 fermented CH and anaerobic inoculum were measured, obtaining a percentage of 29.8% 269 and 1.1% on a wet basis, respectively. A triplicate with only inoculum was also tested 270 as control and its biogas production subtracted from the sample tests. 271 All the bottles were placed in an incubator Memmert In750 (Memmert, 272 Germany) working at 37ºC for 21-25 days. The amount of biogas produced was 273 calculated from the biogas pressure, measured by an ISE 30A-01-P vacuum switch 274 (SMCpneumatics.com, USA), temperature and headspace volume. The bottles were 275 manually agitated before and after measurement, and biogas was periodically released 276 to avoid overpressure. 277 Representative measures of methane percentage in the biogas were taken at 278 different days of incubation. Percentages of methane and carbon dioxide in the biogas 279 were measured using a gas chromatograph 5890A with a column 17066_F ParcpackQ 280 (250C), support 100/120, tube 3m 1/8”x 5.5 mm. Initial oven temperature was 70°C, 281 final oven temperature was 120°C, determination time was 8 min, injector temperature 282 was 150°C and detector temperature was 180°C. Methane peak was detected at 0.8 283 minutes. 100 µl of gas sample were required for the analysis. 12 284 285 Composting assays 286 Solid material after SSF and cellulase extraction was stabilized in order to test its 287 suitability as soil organic amendment. Experiments were carried out in 10 L working 288 volume reactors. Reactors have been previously described [13, 20]. The composting 289 process was performed with air supplied under OUR control [14], which ensures 290 maximum O2 consumption during the experiment. 291 292 Analytical methods 293 Cellulase activity determination 294 Cellulase activity was measured using filter paper analysis as described by Ghose [21]. 295 The extract was centrifuged 10 min at 10000 rpm and the supernatant was 296 analysed. An Activity Unit (U) or Filter paper unit (FPU) is defined as 1 µg of glucose 297 released in one hour under the assay conditions. 298 299 Respirometric tests 300 In this case, respirometric tests were performed on the residues before and after SSF in 301 order to determine their stability (Ordinance on Environmentally Compatible Storage of 302 Waste from Human Settlements and on Biological Waste-Treatment Facilities [18]). A 303 dynamic respirometer constructed following the method described by Adani et al. [22] 304 was used. The experimental device was described by Pognani et al. [23]. According to 305 Gea et al. [24] the test was carried out at a constant temperature of 37°C. All the 306 experiments were performed in triplicate. Cumulative oxygen demand at 4 days AT4 (g 13 307 O2 kg-1 dry matter) and Dynamic Respirometric Index, DRI24 (g O2 kg-1 dry matter h-1), 308 were calculated [24, 25]. 309 OUR (g O2 kg-1 dry matter h-1) was calculated during solid state fermentation as 310 a respirometric indicator of the biological activity of the mixture [14]. When obtaining 311 DRI24, fermentations are carried on with temperature control at 37 °C, while for OUR 312 calculation, the fermentation is performed with no temperature control. OUR was 313 expressed as the average value after an hour (OUR1h). 314 315 Routine analytical methods 316 pH was calculated by soaking 10 g of sample in 50 mL of distilled water. After 30 min, 317 pH value was measured with a pH meter Crison micro pH 2001. Volatile solids (VS), 318 moisture content, total organic carbon (TOC), total Kjeldahl nitrogen, (TKN) and 319 Soluble N-NH4+ were measured according to TMECC [26]. 320 TOC (Total Organic Carbon) was determined using an O.I. Analytical Solid 321 TOC Analyser/Win TOC Solids v3.0, and TKN was measured using a Bloc Digester 6 322 (with six tubes capacity) (J.P. Selecta S.A., Barcelona, Spain) and a Büchi Distillation 323 Unit K‐355 (Flawil, CH). Fibre content (lignin, cellulose and hemicellulose) was 324 determined by the method of Van Soest et al. [27]. Reducing sugars were determined 325 using the method described by Miller et al. [28]. Air filled porosity was calculated using 326 an air pycnometer, as described by Ruggieri et al. [29]. 327 328 Statistics 329 t-student mean analysis comparison was performed to compare the mean values of 330 activity recovery obtained under different experimental conditions. Also, three way 14 331 ANOVA analysis was performed. Sigmaplot 11.0 (Systat software, Inc) was used for all 332 the calculations. Linear regression of the data using Microsoft Excel 2013 was also 333 performed. 334 335 Results and discussion 336 Waste characterization 337 Table 1 shows the characterization of the waste. Moisture was adequate for solid state 338 fermentation since it falls within the recommended values of 40-70%. Krishna et al. 339 [30] reported a range as wide as 20% to 70% for fungi growth and higher than 80% for 340 bacteria, so no extra moistening was needed. Air filled porosity (AFP) was also 341 adequate to ensure total aeration during the process. Ruggieri et al. [29] reported that the 342 recommended AFP value is highly dependent on the material, and can vary from 30% to 343 60%. Alburquerque et al. [31] found that due to the wide variety of methods for 344 measuring AFP, the same measurement can vary between 26% and 61%, so 77% 345 obtained for coffee husk and bulking agent is acceptable. As the objective of the work 346 was the production of cellulase, coffee husk fibres content was determined. According 347 to literature, measured cellulose content in some lignocellulosic substrates for 348 bioethanol production was within 32-47%, 34-45% and 42% for rice straw, wheat straw 349 and corn straw [32], respectively, so the percentage of fibrous material present in coffee 350 husk makes it suitable for cellulase production. As for the biodegradability of the 351 residue, initial respirometric values on Table 3 show that the raw material is not stable 352 and it can be aerobically degraded [18]. 353 354 Solid State Fermentation 15 355 Figure 1 shows an example of the temperature and OUR1h profiles of a fermentation 356 process performed in 10 L working volume reactors. At 48 hours roughly, according to 357 Cerda et al. [12], the maximum enzymatic activity was reached. Cerda et al. [12] also 358 proved that maximum OUR1h and maximum cellulase activity were achieved 359 simultaneously. Temperatures reached 70 °C in the thermophilic range. Maximum 360 OUR1h achieved was 9 mg O2 h-1 g-1 DM. It was observed that the fermentation process 361 is highly reproducible, as the differences in temperature profiles among all replicates 362 were not higher than 6%, presenting an average value of 98.7 °C day (area below 363 temperature curve). Oxygen consumption, on the other hand, presented a maximum 364 deviation of almost 50% of the mean value, which was 1.5 mg O2 h-1 g-1 wet matter 365 (WM). This parameter was calculated for each solid state fermentation process as total 366 amount of oxygen consumed (mg) divided by process time (h) and initial quantity of 367 wet matter inside the reactor (g). Mean value of oxygen consumption was calculated as 368 an average of the values obtained in the performed fermentations. Values of oxygen 369 consumption seem to be related to the obtained cellulase activity extracted in standard 370 conditions (1:15 ratio, BC, agitated mode). The highest cellulase activity value was 371 obtained in fermentation processes presenting the highest oxygen consumption, which 372 were 1.3 ± 0.2 FPU g-1 DM corresponding to 2.3 mg O2 h-1 g-1 WM. Jimenez-Peñalver 373 et al. [33] reported a correlation between sophorolipids production yield by Starmerella 374 bombicola in winterization oil cake and oxygen consumption. Among the fermentations 375 performed, a case was found where correlation between cellulase activity and oxygen 376 consumption was not observed (0.45 ± 0.07 FPU g-1 DM corresponding to 1.7 mg O2 h-1 377 g-1 WM). However, when extractions under different conditions using this fermented 378 solid were performed and the cellulase activity was compared to the standard, the 379 highest percentage of activity recovery was obtained, indicating that probably extraction 16 380 at 1:15 (w:v) did not have the proper efficiency in this case. This can be due to the 381 heterogeneity of the material and the size of the sample extracted as standard. 382 383 Cellulase activity recovery - Downstream 384 Extraction method 385 After 48 hours of fermentation the process was interrupted and different extraction 386 strategies were performed on quantities of material between 160 and 200 g, under 387 different conditions. Table 2 shows the combination of the different extraction 388 parameters and the activity recovery obtained, as well as the pH and conductivity of 389 each extract. t-student test was performed comparing the means of the activity recovery 390 values for each experimental condition. Results of the test are also presented in Table 2. 391 As observed, the most favourable extraction mode, according to Table 2, was the 392 one performed at 1:5 ratio in static mode using DW as extraction solvent with an 393 activity recovery of 108 ± 30 %. The less favourable conditions, according to Table 2, 394 are at a ratio 1:5 and DW as extractant but in agitated mode. As t-test showed, no 395 differences due to extraction method can be appreciated between the means of 5-Static- 396 DW, 5-Agitated-CB, 3-StaticDW, 3-Agitated-DW, 2-Static-DW and 5-Static-CB, being 397 all equal to 5-Static-DW. This group of extraction methods that present higher activity 398 recovery include the combinations in which DW and static agitation mode were used, 399 which would be the most favourable conditions from an economic point of view. Linear 400 regression carried out with the obtained results showed a very low correlation 401 coefficient. Coefficients for each variable indicate that the most relevant extraction 402 parameter is the extraction ratio, being 1:5 the most favourable one. On the other hand, 403 solvent type is the less influencing parameter. This result agrees with that deduced from 17 404 t-student test. According to these results, it is clear that solubility of the produced 405 cellulase is not diminished by the difference in pH observed between BC extract and 406 DW extract. On the contrary, higher pH seems to have a positive influence in the 407 extraction yield. Rezaei et al. [34] reported buffer containing 50 mM sodium acetate at 408 pH 5 to have a positive effect on the recovery of commercial fungal-derived cellulase 409 amended to switchgrass and a negative effect in the recovery of the same enzyme from 410 solid state fermentation samples colonized by the bacterium Acidothermus 411 cellulolyticus. 412 In the particular case of the extraction of cellulase from a solid matrix of 413 fermented coffee husk it is clear that BC has no positive effect compared to DW 414 extraction. Singh et al. [35] asserted that the ideal buffer would be selective and 415 preferable of the same pH of the fermented substrate. As observed in Table 2, pH of the 416 solid matrix, which corresponds to 5-Agitated-DW experiment, is very similar to the pH 417 achieved in all extractions with DW, where pH value depended only on the electrolytes 418 from the solid matrix, contrarily to what happens with BC extract, which pH is close to 419 6 in all cases, due to the buffer effect of citrate. Comparing initial pH of DW and BC, 420 DW pH is closer to that value. Regarding this, Bera et al. [36] presented the extraction 421 of amylase at different pHs ranging from 2 to 12 and observed no statistical significant 422 differences among the results . 423 According to the data obtained, both agitated and static methods provide an 424 adequate contact between solid and solvent. This can be due to the small size of the 425 coffee husk particles, and consequently no aggregates were formed once in contact with 426 the liquid. Singh et al. [35] reported no differences between static and agitated (220 427 rpm) extraction modes for pectinase from bran, and only an increase of viscosity and 428 colour in the extract. On the other hand, Shata et al. [37] reported a yield in milk- 18 429 clotting (proteins) extraction from also bran of 3000 U g-1 in agitated extraction in front 430 of 500 U g-1 in static extraction. Therefore, agitation requirements seem to be highly 431 dependent on the type of enzyme. In the case of coffee husk and cellulase extraction, 432 agitation mode has a low influence in the activity recovery yield. 433 Three independent extractions were performed in recirculated mode with DW as 434 solvent and ratio 1:3 solid:liquid, with 7 times the recirculation of the whole volume of 435 liquid. Also, another three independent extractions were performed using fermented 436 coffee husk from the same process in agitated mode at same ratio and with DW as 437 solvent. Activity recovery for both extraction modes was calculated, obtaining a 98 ± 1 438 % of activity recovery for the agitated mode and a 93.0 ± 0.7 % for the recirculated 439 mode, showing that no improvement is achieved with recirculation. 440 441 Number of extraction stages 442 Three extractions in static mode, ratio 1:2 and DW as a solvent were performed and the 443 cellulase activity compared to the standard in order to calculate activity recovery. Figure 444 2 shows cellulase activity recovery in those extractions. As it can be seen, although 445 fermented material came from the same process, one of the replicates differs 446 significantly from the others. One explanation could be the heterogeneity of the 447 material, with an irregular distribution of the bulking agent. A decrease in the activity 448 recovery through the different extraction stages is observed, but a total activity recovery 449 after three extractions of 215% was observed in two of the replicates and a 144% for the 450 remaining one. Activity recovery in the second extraction represents a percentage 451 between 20% and 27% of the activity recovered in the first extraction. In the case of the 452 second replicate. Pirota et al. [15] performed consecutive endoglucanase extractions 453 over the same bran with DW reporting a 75% of recovery in the first extraction and 19 454 almost a 25% in the second and third. Compared to the results obtained in this work, 455 first extraction has a similar yield than in this case. Diaz et al. [38] reported consecutive 456 extractions of exo-polygalacturonase from fermented grape pomace obtaining almost 457 the same extraction yield up to the fourth extraction, needing at least 6 extractions in 458 order to recover all the enzyme. Although enzyme can be recovered up to third 459 extraction, the enzyme obtained will be much diluted already at the second extraction, 460 which represents increasing costs in processing the extract for lesser yield. According to 461 these data, it is recommended to perform only one extraction. 462 463 Lyophilisation and activity preservation 464 Enzymatic activity conservation in the lyophilized product was assessed during 115 465 days. The activity of the resuspended lyophilized enzyme and that of frozen samples 466 were compared to the activity of fresh extract. The weight of remaining powder 467 obtained was 10.881 g for 500 mL of 2-Static-BC lyophilized extract and 8.413 g for 5- 468 Static-BC. 469 Figure 3 shows the activity recovery of lyophilized and frozen extract compared 470 to the activity of the fresh one. For the samples obtained in the 1:2 (w:v) extraction 471 ratio, the activity recovery presented an initial drop of 20% and afterwards presented 472 values overpassing the 100% activity recovery in the last measures. Activity losses can 473 appear in every step from freezing to rehydration. Hédoux et al. [39] reported structural 474 changes in protein lyophilisation after drying step, although the changes in secondary 475 structure were reversible after rehydration. Also, other parameters of the lyophilisation 476 process can affect the subsequent activity recovery, Passot et al. [40] reported that 477 nucleation control by different techniques can improve activity preservation. 20 478 In our case, values from lyophilised and unfrozen samples can be considered 479 equal, thus, there is no activity loss during the dehydratation process. The observed 480 initial losses can be due to the freezing process and to the error associated to the 481 cellulase determination method, as they are observed in both frozen and lyophilised 482 samples. When comparing lyophilisation and freezing as conservation methods, 483 lyophilisation presents some advantages which are: the concentration of the enzyme 484 facilitating its use and the reduction of the space required for storing and the storage at 485 not very low temperatures for its conservation. For this reason, lyophilisation should be 486 chosen as a preservation method. 487 488 Zero waste strategies 489 Biomethane potential 490 Bio methane potential (BMP) was assayed using extracted CH as feedstock to evaluate 491 the suitability of the residue after cellulase extraction for the production of biogas. 492 Experiments were carried out during 21 days and biogas produced and methane content 493 were measured. Although recommended ratio of inoculation according to Ponsá et al. 494 [19] is 1:2 F:I (1 g of VS of feedstock per 2 g of VS of inoculum), a first experiment 495 was carried out using different ratios F:I. As the anaerobic biodegradability of this 496 residue had not been measured before, it was necessary to verify that a higher or lower 497 F:I ratio was not needed in order to have an adequate biogas production. Ratios assayed 498 were four; 1:1; 1:1.5; 1:2; 1:2.5 (g of VS of feedstock: g of VS of Inoculum). After 21 499 days of incubation, biogas production was 96 ± 10, 107 ± 7, 128 ± 10 and 138 ± 6 mL 500 biogas g-1 VS respectively for the mentioned ratios. In view of the results obtained, ratio 501 1:2 (also according to Ponsá et al. [19]) was selected for future experiments. In the next 21 502 set of biogas potential tests, fermented CH after and before extraction from two 503 different fermentation processes was used as anaerobic feedstock in a ratio 1:2 F:I. Tests 504 were performed in triplicate. Biogas and methane productions are shown in Table 4. 505 After 27 days of anaerobic digestion at 37 °C biogas production for fermented material 506 after extraction was 104 ± 7 mL biogas g-1 VS in both cases, and 124 ± 6 and 100 ± 11 507 mL biogas g-1 VS for the fermented material before extraction, respectively. As for 508 methane production, extracted samples after 27 days produced 40 ± 8 and 51 ± 3 mL 509 methane g-1 VS and for non-previously extracted samples the methane production was 510 52 ± 3 and 54 ± 6 mL methane g-1 VS respectively. Considering the standard deviation, 511 methane production can be considered equal for all samples. As shown in Figure 4, in 512 all cases, more than the 85% of the production was measured during the first 15 days of 513 digestion, which indicates that the length of the process is adequate, since production 514 remarkably decreased from that moment. Regarding biogas production, comparing this 515 material to fresh lignocellulosic materials used as a feedstock for anaerobic digestion 516 such as rice straw, hazelnut skin and cocoa shell, they have been reported to produce 517 207 ± 22.1, 223 ± 25.1 and 199 ± 22.4 mL CH4 g-1 VS [41], respectively, whereas other 518 lignocellulosic materials as ley, straw or blue mussels produced between 190 and 330 519 mL CH4 g-1VS [42]. The low production of methane by coffee husk can be due to the 520 presence of bulking agent in the mixture, which is not easily biodegradable and does not 521 contribute to the production of biogas. Also, coffee husk is generated during the 522 roasting process of coffee grains, which can lead to an easily biodegradable matter loss. 523 Fresh coffee husk itself produced 241 ± 11 mL biogas g-1 SV and 141 ± 7 mL CH4 g-1 524 SV after 28 days of incubation, half of the amount produced by fresh lignocellulosic 525 wastes with no previous roasting. Lee et al. [43] reported different profile in volatile 22 526 gases during fermentation of coffees with different degree of roasting. Moreover, loss of 527 soluble biodegradable mater also occurs during the extraction of cellulase. 528 529 Composting 530 Once the extraction was carried out, the remaining solid waste was stabilized in order to 531 obtain a compost-like material adequate as agricultural amendment, closing the cycle 532 and following a zero waste strategy. 533 As described previously, solid state fermentations lasted 48 hours before being 534 stopped and extraction performed. The remaining solid after extraction was pressed 535 manually in order to remove the maximum amount of moisture. However, the final 536 measured moisture varied between 71% and 75% in weigh, higher than the 537 recommended values for composting, which should be in the range of 40%-60% [44]. 538 Three replicates of the composting process were performed, using 10 L working volume 539 reactors. Also, a 10 L reactor was run containing fermented material without extraction 540 in order to compare both materials. Oxygen consumption for the non-extracted sample 541 was 12.5 mg O2 g-1 DM day-1, higher than that of the extracted sample of the same 542 fermentation experiment, which was 4.3 mg O2 g-1 DM day-1. Similar values were 543 measured in the rest of the composting processes, being 4.57 and 4.43 mg O2 g-1 DM 544 day-1. Hygenization temperatures were reached, maximum temperature measured was 545 around 40 °C. However, non-extracted material reached temperatures above 50°C for 24 546 hours approximately. Differences in biological activity of non-extracted material and 547 extracted material can be due to the loss of soluble biodegradable material during the 548 extraction process, since particle size of coffee husk was measured, presenting values 549 between 1.6 mm and 710 µm. Also, biological activity could be hampered by the high 550 moisture of the material after extraction [45]. Table 3 shows the respirometric values 23 551 measured during the different steps of the process. It is observed that the stability of the 552 waste was reached after 15 days of composting with values of DRI24 of 0.33 ± 0.04 g O2 553 kg-1 DM h-1 and AT4 of 29 ± 5 g O2 kg-1 DM. These values were very similar to the ones 554 obtained from a mature compost sample [46]. Waste used in SSF with no extraction was 555 also stable after 7 days of fermentation with values of DRI24 and AT4 of 0.7 ± 0.3 g O2 556 kg-1 DM h-1 and 59 ± 21 g O2 kg-1 DM respectively. pH was measured for one replicate, 557 presenting values between 8 and 8.5, which are like those of final compost. 558 559 Comparison of anaerobic digestion and composting 560 561 According to coffee production companies, approximately 1% of the weight of final 562 product ends up as coffee husk. In the case of the coffee husk supplier of this study, 563 around 180 t/y of coffee husk are produced. According to our results, anaerobic 564 digestion of this fermented coffee husk would represent a production of approximately 565 3.6 kW of electric power (considering the methane content of the biogas, yield of biogas 566 conversion to electricity, etc.). Although in the case of composting no electricity is 567 generated and there is a net consumption of energy, Zulkepli et al. [47] compared the 568 investment required by an anaerobic digestion plant and by a composting plant for the 569 treatment of municipal solid waste finding that capital cost was for the anaerobic 570 digestion plant is 20 times higher than for the composting plant. In view of these facts, 571 composting would be recommended in the present study as a low cost technology given 572 that the production of electricity from biogas is lower than those reported for other 573 organic wastes. However, anaerobic co-digestion could be considered if other wastes 574 with higher biogas production are available to be treated with coffee husk fermented 575 residues. 24 576 577 Conclusion 578 Different aspects of the downstream stage in the production of cellulase through solid 579 state fermentation of coffee husk were assessed in this work. Extraction parameters as 580 solid:solvent ratio, agitation mode and type of solvent seem to have low influence in the 581 total activity recovery. Buffering effect of citrate solution does not present any 582 advantage over distillate water on enzyme solubility. Contact between solid and solvent 583 appears to be adequate independently of the agitation provided. Although maximum 584 activity recovery was 108 ± 30 % performing the extraction under 1:5 w:v ratio, static 585 mode and distilled water as solvent, statistically equal result was obtained using 1:2 w:v 586 ratio. 587 Activity loss after lyophilisation is not observed, as values oscillate around 588 100% recovery during the first 50 days of lyophilized material storage. For the aerobic 589 stabilization of the extracted material, water can hamper the process, thus, a previous 590 drying step is recommended. At the end of the stabilization process a compost-like 591 product is obtained, which can be used as organic amendment. Biogas production 592 resulted in low values compared with other agricultural wastes. Therefore, anaerobic 593 digestion of the residue is recommended in co-digestion with higher anaerobically 594 biodegradable residues. 595 In summary, a detailed process to optimize the production of cellulase from 596 lignocellulosic wastes is presented, considering the overall process from SSF to a final 597 lyophilized product. The strategies to manage the spent material after extraction are also 598 presented for a zero waste process. 599 25 600 References 601 602 1. British Petroleum plc.: http://www.bp.com/es_es/spain/prensa/notas-de-prensa 603 /2016/bp-statistical-review world-energy-2016.html Accessed 10 July 2017 604 2. Kuhad, R.C., Gupta, R., Singh, A.: Microbial Cellulases and Their Industrial 605 Applications. Enzyme Research DOI 10.4061/2011/280696 (2011) 606 3. Cherian, E., Dharmendirakumar, M., Baskar. G.: Immobilization of Cellulase onto 607 MnO2 Nanoparticles for Bioethanol Production by Enhanced Hydrolysis of 608 Agricultural Waste. Chinese J Catal. 36 (8), 1223–1229 (2015) 609 4. Idris, A.S.O., Pandey, A.; Rao, S.S., Sukumaran, R.K.: Cellulase Production through 610 Solid-State Tray Fermentation, and Its Use for Bioethanol from Sorghum Stover. 611 Bioresource Technol. 242, 265-271 (2017) 612 5. Li, Y.H., Zhang, X.Y., Xiong, L., Mehmood, M.A., Zhao, X.Q., Bai, F.W.: On-Site 613 Cellulase Production and Efficient Saccharification of Corn Stover Employing cbh2 614 Overexpressing Trichoderma Reesei with Novel Induction System. Bioresource 615 Technol. 238, 643–649 (2017) 616 6. Oke, A.M., Suffian, M., Annuar, M., Simarani, K.: Mixed Lignocellulosic Biomass 617 Degradation and Utilization for Bacterial Cellulase Production. Waste Biomass 618 Valori. 8(3), 893–903 (2017) 619 7. Saini, A., Aggarwal, N.K., Yadav, A.: Cost-Effective Cellulase Production Using 620 Parthenium Hysterophorus Biomass as an Unconventional Lignocellulosic Substrate. 621 3 Biotech. 7(1), 1-11(2017) 622 8. Sun, H.Y., Li, J.H., Zhao, P.J., Peng, M.: Banana Peel: A Novel Substrate for 623 Cellulase Production under Solid-State Fermentation. Afr J Biotechnol. 10(77), 624 17887-17890 (2011) 625 9. Dhillon, G.S., Kaur, S., Brar, S.K., Verma, M.: Potential of Apple Pomace as a Solid 626 Substrate for Fungal Cellulase and Hemicellulase Bioproduction through Solid-State 627 Fermentation. Ind Crop Prod. 38, 6–13 (2012) 628 10. Yoon, L.W., Ang, T.N., Ngoh, G.C., Chua, A.S.M.: Fungal Solid-State 629 Fermentation and Various Methods of Enhancement in Cellulase Production. 630 Biomass Bioenerg. 67, 319–38 (2014) 631 11. Trivedi, N., Reddy, C.R.K., Radulovich, R., Jha, B.: Solid State Fermentation 632 (SSF)-Derived Cellulase for Saccharification of the Green Seaweed Ulva for 633 Bioethanol Production. Algal Res. 9, 48–54 (2015) 26 634 12. Cerda, A., Gea, T., Vargas-Garcia, M.C., Sanchez, A.: Towards a Competitive Solid 635 State Fermentation: Cellulases Production from Coffee Husk by Sequential Batch 636 Operation and Role of Microbial Diversity. Sci Total Environ. 589, 56–65 (2017) 637 13. Maulini-Duran, C., Puyuelo, B., Artola, A., Font, X., Sanchez, A., Gea, T.: VOC 638 Emissions from the Composting of the Organic Fraction of Municipal Solid Waste 639 Using Standard and Advanced Aeration Strategies. J Chem Technol Biot. 89(4), 579- 640 586 (2014) 641 14. Puyuelo, B., Gea, T., Sanchez, A.: A New Control Strategy for the Composting 642 Process Based on the Oxygen Uptake Rate. Chem Eng J. 165(1), 161–169 (2010) 643 15. Pirota, R.D.P.B., Miotto, L.S., Delabona, P.S., Farinas, C.S.: Improving the 644 Extraction Conditions of Endoglucanase Produced by Aspergillus Niger under Solid- 645 State Fermentation. Braz J Chem Eng. 30(1), 117–23 (2013) 646 16. Chandra, M.S., Viswanath, B., Reddy, B.R.: Optimization of extraction of β- 647 endoglucanase from the fermented bran of Aspergillus Niger. Indian J Microbiol. 648 50(1), 122-126 (2010) 649 17. Farinas, C.S., Loyo, M.M., Baraldo Junior, A., Tardioli, P.W., Bertucci Neto, V., 650 Couri, S.: Finding stable cellulase and xylanase: evaluation of the synergistic effect 651 of pH and temperatura. New Biotechnol. 27 (6), 810-815 (2010) 652 18. German Federal Ministry for the Environment, Nature Conservation and Nuclear 653 Safety, Ordinance on environmentally compatible storage of waste from human 654 settlements and on biological waste treatment facilities of 20 February (2001) 655 URL:http://www.bmu.de/files/pdfs/allgemein/application/pdf/ablagerungsverordnun 656 g.pdf Accessed 25 July 2017 657 19. Ponsá, S., Gea, T., Sánchez, A.: Anaerobic co-digestion of the organic fraction of 658 municipal solid waste with several pure organic co-substrates. Biosyst Eng. 108(4), 659 352-360 (2011) 660 20. Sayara, T., Sarrà, M., Sánchez, A.: Effects of Compost Stability and Contaminant 661 Concentration on the Bioremediation of PAHs-Contaminated Soil through 662 Composting’, J Hazard Mater. 179 (1–3), 999–1006 (2010) 663 21. Ghose, T.K.: Measurement of Cellulase Activities. Pure Appl Chem. 59(2), 257– 664 268 (1987) 665 22. Adani, F., Gigliotti, G., Valentini, F., Laraia, R.: Respiration Index Determination: 666 A Comparative Study of Different Methods. Compost Sci Util. 11(2), 144–151 667 (2003) 27 668 23. Pognani, M., Barrena, R., Font, X., Adani, F., Scaglia, B., Sanchez, A.: Evolution 669 of Organic Matter in a Full-Scale Composting Plant for the Treatment of Sewage 670 Sludge and Biowaste by Respiration Techniques and Pyrolysis-GC/MS. Bioresource 671 Technol. 102(6), 4536–4543 (2011) 672 24. Gea, T., Barrena, R., Artola, A.,Sanchez, A.: Monitoring the Biological Activity of 673 the Composting Process: Oxygen Uptake Rate (OUR), Respirometric Index (RI), and 674 Respiratory Quotient (RQ). Biotechnol Bioeng. 88(4), 520–527 (2004) 675 25. Almeira, N., Komilis, D., Barrena, R., Gea, T., Sánchez, A.: The importance of 676 aeration mode and flowrate in the determination of the biological activity and 677 stability of organic wastes by respiration indices. Bioresource Technol. 196, 256-262 678 (2015) 679 26. The U.S. Department of Agriculture and The U.S. Composting Council (2001) Test 680 Methods for the Examination of Composting and Compost, Edaphos International, 681 Houston 682 27. Van Soest, P., Robertson, J., Lewis, B.: Methods for dietary fiber, neutral detergent 683 fiber, and non-starch polysaccharides in relation to animal nutrition. J Dairy Sci. 684 74(10), 3583-3597 (1991) 685 28. Miller, G.: Use of Dinitrosalicylic Acid Reagent for Determination of Reducing 686 Sugar. Anal Chem. 31(3), 426-428 (1959) 687 29. Ruggieri, L., Gea, T., Artola, A., Sanchez, A.: Air Filled Porosity Measurements by 688 Air Pycnometry in the Composting Process: A Review and a Correlation Analysis. 689 Bioresource Technol. 100(10), 2655–66 (2009) 690 30. Krishna, C.: Solid-State Fermentation Systems: An Overview. Crit Rev Biotechnol. 691 25(1–2), 1–30 (2005) 692 31. Alburquerque, J.A., McCartney, D., Yu, S., Brown, L., Leonard, J.J.: Air space in 693 composting research: A literature review. Compost Sci Util. 16(3), 159-170 (2008) 694 32. Sarkar, N., Ghosh, S.K., Bannerjee, S., Aikat, K.: Bioethanol Production from 695 Agricultural Wastes: An Overview. Renew Energ. 37, 19–27 (2012) 696 33. Jimenez-Penalver, P., Gea, T., Sanchez, A., Font, X.: Production of Sophorolipids 697 from Winterization Oil Cake by Solid-State Fermentation: Optimization, Monitoring 698 and Effect of Mixing. Biochem Eng J. 115, 93–100 (2016) 699 34. Rezaei, F., Joh, L.D., Kashima, H., Reddy, A.P., VanderGheynst, J.S.: Selection of 700 Conditions for Cellulase and Xylanase Extraction from Switchgrass Colonized by 701 Acidothermus Cellulolyticus. Appl Biochem Biotech. 164(6), 793–803 (2011) 28 702 35. Singh, S.A., Ramakrishna, M., Appu Rao, A.G.: Optimisation of Downstream 703 Processing Parameters for the Recovery of Pectinase from the Fermented Bran of 704 Aspergillus Carbonarius. Process Biochem. 35(3–4), 411–17 (1999) 705 36. Bera, M.B., Panesar, P.S., Panesar, R., Singh, B.: Application of Reverse Micelle 706 Extraction Process for Amylase Recovery Using Response Surface Methodology. 707 Bioproc Biosyst Eng. 31(4), 379–84 (2008) 708 37. Shata, H.M.A.: Extraction of Milk-Clotting Enzyme Produced by Solid State 709 Fermentation of Aspergillus Oryzae. Pol J Microbiol. 54(3), 241–47 (2005) 710 38. Díaz, A.B., Caro, I., de Ory, I., Blandino, A.: Evaluation of the conditions for the 711 extraction of hydrolitic enzymes obtained by solid state fermentation from grape 712 pomace. Enzyme Microb Tech. 41, 302-306 (2007) 713 39. Hédoux, A., Paccou, L., Achir, S., Guinet, Y.: In Situ Monitoring of Proteins during 714 Lyophilization Using Micro-Raman Spectroscopy: A Description of Structural 715 Changes Induced by Dehydration. J Pharm Sci. 101(7), 2316–26 (2012) 716 40. Passot, S.,Trelea, I.C., Marin, M., Galan, M., Morris, G.J., Fonseca, F.: Effect of 717 Controlled Ice Nucleation on Primary Drying Stage and Protein Recovery in Vials 718 Cooled in a Modified Freeze-Dryer. J Biomech Eng. 131(7), 74511 (2009) 719 41. Mancini, G., Papirio, S., Lens, P.N.L., Esposito, G.: Effect of N - 720 Methylmorpholine- N -Oxide Pretreatment on Biogas Production from Rice Straw, 721 Cocoa Shell, and Hazelnut Skin. Environ Eng Sci. 33(11), 843–50 (2016) 722 42. Ammenberg, J., Feiz, R.: Assessment of Feedstocks for Biogas Production, Part II: 723 Results for Strategic Decision Making. Resour Conserv Recy. 122, 388–404 (2017) 724 43. Lee, L.W., Tay, G.Y., Cheong, M.W., Curran, P., Yu, B., Liu, S.Q.: Modulation of 725 the Volatile and Non-Volatile Profiles of Coffee Fermented with Yarrowia 726 Lipolytica : II. Roasted Coffee. Lebensm-Wiss Technol. 80, 32–42 (2017) 727 44. Kim, E., Lee, D.H., Won, S., Ahn, H.: Evaluation of Optimum Moisture Content for 728 Composting of Beef Manure and Bedding Material Mixtures Using Oxygen Uptake 729 Measurement. Asian Austral J Anim. 29(5), 753–58 (2016) 730 45. Richard, T.L., Hamelers, H.V.M., Veeken, A., Silva, T.: Moisture Relationships in 731 Composting Processes. Compost Sci Util. 10(4), 286–302 (2002) 732 46. Komilis, D., Kanellos, D.: A modified dynamic respiration test to assess compost 733 stability: Effect of sample size and air flowrate. Bioresource Technol. 117, 300-309 734 (2012) 29 735 47. Zulkepli, N.E., Muis, Z.A., Mahmood, N.A.N., Hashim, H., Ho, W.S.: Cost Benefit 736 Analysis of composting and anaerobic digestion in a community: A review. Chem 737 Eng Trans. 56, 1777-1782 (2017) 30 Tables Table 1. Characterization of wastes used as substrates in SSF. Parameter Coffee Husk Compost Mixture** pH 6.4 ± 0.1 7.6 ± 0.5 6.7 ± 0.4 CE (mS cm-1) n.m. 6.3 ± 0.2 n.m. Moisture (%, wb) 60.2 ± 0.6 35 ± 1 61 ± 1 Dry mater (%, wb) 40.1 ± 0.4 64 ± 1 39 ± 1 Organic matter* (%, db) 90.21 ± 0.01 n.m. 90 Total C* (%, db) 80.1 n.m. n.m. Total N* (%, db) 3.5 n.m. n.m. C/N ratio 22.9 ± 0.1 n.m. n.m. Bulk density* (g L-1) 238.1 n.m. 358 Air filled porosity* (%) 78.9 n.m. 77.2 Reducing sugars (%, db) 0.65 ± 0.01 n.m. n.m. Glucose (%, db) 0.02 ± 0.01 n.m. n.m. Cellulose (%, db) 25.7 ± 0.2 10 ± 1 26 ± 3 Hemicellulose (%, db) 14.6 ± 0.1 10.2 ± 0.1 13.25 ± 0.07 Lignin (%, db) 17.6 ± 0.5 14 ± 1 21.1 ± 1.0 wb: wet basis. db: dry basis; n.a.: not available. AFP: Air filled porosity (v/v, percentage in volume). Values are the average of independent experiments and its standard deviation. (*) s.d.<6% n.m.: not measured. (**) Mixture of compost and coffee husk at 1:1 weight ratio and bulking agent in a volume ratio 1:1 31 Table 2. Summary of extraction experiments, 30 min extraction time. Extraction method pH Conductivity (mS cm-1) Activity recovery (%) 2-Static-DW a, b, e, g 9.0 ± 0.2 3.6 ± 0.2 84 ± 22 2-Agitated-DW c, d, e, g 9.0 ± 0.2 3.8 ± 0.3 56 ± 14 2-Static-BC b, c, f, g 6.1 ± 0.5 7.8 ± 0.5 58 ± 6 2-Agitated-BC c, d, e, f, g 6.2 ± 0.9 7.9 ± 0.6 56 ± 12 3-Static-DW a 9.12 ± 0.05 2.6 ± 0.2 95 ± 13 3-Agitated-DW a, b, d 9.11 ± 0.07 2.8 ± 0.1 84 ± 15 3-Static-BC c, d, e 5.6 ± 0.2 7.2 ± 0.5 61 ± 3 3-Agitated-BC b, c, g 5.9 ± 0.4 7.6 ± 0.5 69 ± 12 5-Static-DW a 9.21 ± 0.09 1.5 ± 0.1 108 ± 30 5-Agitated-DW g 9.1 ± 0.1 1.9 ± 0.1 50 ± 7 5-Static-BC a, b 5.2 ± 0.1 6.9 ± 0.5 83 ± 18 5-Agitated-BC a 5.4 ± 0.1 7.2 ± 0.5 103 ± 17 BC: Buffer citrate mono hydrate pH 4.8; DW: distilled water; 2: extraction ratio 1:2 (w:v), 3: extraction ratio 1:3 (w:v), 5: extraction ratio 1:5 (w:v). Subscript indicates the groups of extractions that are considered equal according to t- student test. Values are the average of independent experiments. 32 Table 3. Result of respirometric assay. Final SSF values are initial composting values. AT4 (O2 g-1 DM) DRI24 (O2 g-1 DM h-1) Fresh material 86 ± 23 1.6 ± 0.2 48 hours of SSF - Before extraction 86 ± 17 1.3 ± 0.3 7 days of SSF - Before extraction 59 ± 21 0.7 ± 0.3 48 hours of SSF - After extraction 47 ± 7 0.70 ± 0.06 Stabilized (previous extraction) 29 ± 5 0.33 ± 0.04 Stabilized (no previous extraction) 20.0 ± 0.7 0.29 ± 0.01 All values are the average of independent experiments. 33 Table 4. Biogas production of coffee husk anaerobic digestion. For the materials before and after extraction, two independent fermentations were evaluated for biogas and methane production. Fresh material Before extraction After extraction Fermentation Fermentation Fermentation Fermentation 1 2 1 2 Biogas production 241 ± 11 124 ± 6 100 ± 11 104 ± 7 104 ± 7 (mL g-1 VS) Methane production 141 ± 7 52 ± 3 54 ± 6 40 ± 8 51 ± 3 (mL g-1 VS) All values are the average of triplicate measurements. 34 Figure captions Fig. 1 Temperature (solid line) and OUR 1h (dotted line) profiles of CH solid state fermentation Fig. 2 Percentage of activity recovery of three consecutive extractions over the same solid. Three independent experiments are showed Fig. 3 Percentage of activity recovery of: lyophilised samples of CH extracted in Static mode, using BC and ratio 1:2 (w:v) (circles), 1:5 (w:v) (diamonds) and unfrozen samples of CH extracted in Static mode, using BC and ratio 1:2 (w:v) (triangles), 1:5 (w:v) (squares) Fig. 4 Biogas production from anaerobic digestion of fresh coffee husk (squares), fermented coffee husk before extraction (triangles) and fermented coffee husk after extraction (circles). For materials before and after extraction, two independent fermentations were evaluated. 35 Fig. 1 80 Extraction point 14 Temperature (48 hours roughly) 70 OUR 1h 12 OUR 1h (O2 mg h g dm) 60 -1 Temperature (°C) 10 -1 50 8 40 6 30 4 20 2 10 0 0 10 20 30 40 50 60 Time (hours) 36 Fig. 2 200 180 Replicate 1 160 Replicate 2 Replicate 3 140 % Activity recovery 120 100 80 60 40 20 0 1ª 2ª 3ª Extraction number 37 Fig. 3 250 200 % Activity recovery 150 100 Unfrozen 1:2 50 Unfrozen 1:5 Liophilised 1:2 Liophilised 1:5 0 0 20 40 60 80 100 120 Time (days) 38 Fig. 4 300 Post extraction Pre extraction Biogas production (mL biogas g VS) 250 Fresh material -1 200 150 100 50 0 0 5 10 15 20 25 30 Time (days) 39