High concentration and yield production of mannose from açaí (Euterpe oleracea) seeds via diluted-acid and mannanase-catalyzed hydrolysis

The açaí berry’s seed corresponds to 85–95% of the fruit’s weight and represents ~1.1 million tons of residue yearly accumulated in the Amazon region. This study confirmed that mannan is the major component of mature seeds, corresponding to 80% of the seed’s total carbohydrates and about 50% of its dry weight. To convert this high mannan content into mannose, a sequential process of diluted acid and enzymatic hydrolysis was evaluated. Diluted-H2SO4 hydrolysis (3%-acid, 60-min, 121°C) resulted in a 30% mannan hydrolysis yield and 41.7 g/L of mannose. Because ~70% mannan remained in the seed, a mannanase-catalyzed hydrolysis was sequentially performed with 2–20% seed concentration, reaching 146.3 g/L of mannose and a 96.8% yield with 20% solids. As far as we know, this is the highest reported concentration of mannose produced from a residue. Thus, this work provides fundamental data for achieving high concentrations and yields of mannose from açaí seeds. Highlights Mannan was confirmed as the major component (~50%) of açaí seeds. Diluted-H2SO4 hydrolysis had a limited effect on mannan conversion into mannose. Enzymatic hydrolysis was sequentially performed with a high seed concentration. Mannan was efficiently hydrolyzed by mannanases, producing a 96.8% yield. Mannose production of 146.3 g/L was obtained with mannanase-catalyzed hydrolysis. Graphical abstract

hand, mannan-degrading enzymes could also be applied for the release of free mannose

99
For example, mannose can be easily reduced to mannitol-a specialty chemical 100 with a wide variety of uses in the pharmaceutical industry-in a process with a 90% 101 yield (Mishra and Hwang, 2013). However, because of the lack of abundant and low-102 cost sources of mannose, the conventional industrial processes for mannitol production 103 are based on the chemical hydrogenation of fructose or inverted sucrose, which produce 104 low yields of about 25% and 50%, respectively (Makkee et al., 1985). 105 Considering that açaí seeds can be a potential rich source of mannan, its high 106 abundance in Brazil, and the limited current knowledge of mannan depolymerization, 107 the aim of the current study was to confirm the carbohydrate composition of açaí seeds 108 and to evaluate acidic-and enzymatic-catalyzed strategies to maximize mannose 109 production.  119 Two lots of açaí seeds were characterized in the current study. Samples from lot 120 1 were received as shown in Figure 1a and were noted as "whole seeds" (Fig. 1a), 121 6 while lot 2 contained samples already milled (Fig. 1d). For characterization, the whole 122 seeds samples from lot 1 were processed with a knife mill as received. Alternatively, 123 for lot 1 samples, the external fiber layers (Fig. 1c) were manually separated from the 124 inner core stones (Fig. 1b), and the two fractions were milled separately for further 125 chemical characterization. In parallel, the masses of 35 whole seed samples were 126 measured using an analytical balance (0.001 precision). Subsequently, the external 127 fiber layers were manually removed from the core stones and then both were weighted 128 separately to determine their percentage in relation to the total mass of the whole seed. In natura milled açaí seeds underwent an extraction process (Sluiter et al., 2008) 136 with some modifications. Approximately 2 g of the biomass were weighted into 137 cellulose thimbles and extracted with water, which was followed by a 95% ethanol 138 extraction; each extraction step was performed for at least 12 h. The procedure was 139 carried out using six Soxhlet apparatus in parallel. By the end of the extraction, three of 140 the thimbles were put in a 105 ºC drying oven overnight to calculate the extractives by 141 weight difference, while the other three were put in a 40 ºC drying oven to be used in 142 the following chemical characterization step. Then, 0.3 g of the dried, extractive-free in 143 natura and acid-treated açaí seed samples were submitted to an acid hydrolysis process 144 7 in two steps in triplicate (Sluiter et al., 2012). In the first step, the samples were mixed 145 with 3 mL of a 72% sulfuric acid solution in round-bottom hydrolysis tubes and put in 146 a 30 °C water bath for 1 h under constant stirring. In the second step, 84 mL of 147 deionized water were added to the tubes, which were autoclaved for 1 h at 121 °C.

148
After this, the solutions were vacuum filtered through dried, preweighted Gooch 149 crucibles. The acidic liquors were neutralized with CaCO 3 and went through HPLC and 150 HPAEC-PAD analysis, which is described below, for carbohydrate quantification. The 151 crucibles containing the remaining solids were dried overnight in an oven at 105 ºC, 152 and the dry weight was recorded for acid insoluble solids (AIS) quantification using the 153 difference in weight. The insoluble ash content was also measured using the difference 154 in weight after the same crucibles were put overnight in a furnace at 575 ºC. Four sulfuric acid concentrations were evaluated for the diluted-acid hydrolysis 164 step, corresponding to 1.5%, 3.0%, 3.5%, and 4.5% (% w/w). Each of these solutions 165 was evaluated at a 30-and a 60-min residence time at 121 °C. Each condition was 166 performed in at least four replicates in round-bottom hydrolysis tubes containing 4 g 167 (dry weight) of the milled açaí seeds and 16 mL of the corresponding diluted-acid 168 solution, resulting in a solid:liquid ratio of 1:4. The tubes were put in an autoclave for 169 8 30-or 60-min at 121 ºC and then cooled in an ice bath. After this, 64 mL of water were 170 added to the tubes, which were agitated for homogenization, and samples of the liquid 171 streams were withdrawn, being then filtrated, neutralized, and prepared for HPLC and 172 HPAEC-PAD analysis, to determine the sugar and degradation products, as described 173 below.

174
The solid contents of two of the four tubes were filtrated in preweighted fiber 175 glass filters and put in an oven at 105 ºC overnight to calculate the amount of mass 176 transferred to the acidic liquid phase. The solid contents of the other two tubes were 177 filtered and stored in the refrigerator until further use either for characterization of the 178 chemical composition or for enzymatic hydrolysis assays. Prior to the characterization 179 assays, the samples were dried at 40 °C until reaching less than 10% moisture and then 180 used for the determination of AIS, ash, and carbohydrates, as previously described.

181
The combined severity factor was calculated for each diluted-acid hydrolysis 182 condition, which was evaluated based on the severity factor R 0 , which accounts for the 183 effect of the temperature, residence time, and pH of the hydrolysates after the reaction,  1). By botanical definition, the external fiber layer is not considered part of the seed; 265 however, we denominated the seed as the residue generated after the depulping and 266 sieving of açaí berries (fibers + seed) because-for the sake of brevity-it is 267 improbable that any large-scale commercial use of this residue will separate those 268 fractions.

269
The average weight of the whole seeds was 0.78 g ± 0.12, ranging from 0.56 g to 270 1.06 g, and the mass percentage of fiber in relation to the whole seed was equivalent to 271 5.97% ± 1.45. These data are in close agreement with a previous study that reported 272 that the whole seeds average weight was 0.72 ± 0.04 g and that the fibrous layer 273 corresponded to 6.50% of the whole seed weight (Pessoa et al., 2010).

274
The literature data regarding the açaí seed composition so far is conflicting.

275
Therefore, to better evaluate the seed's uses, a confirmation of its chemical 276 composition is crucial to design the most suitable processing methods for sugar 277 recovery. In the current study, the characterization of two distinct seed samples lots 278 was performed, as well as an analysis of different seed fractions.

300
The composition of the core stones (Fig. 1b) showed a high similarity with the 301 whole seed, as expected, considering that this fraction corresponds to almost 94% of  1,4-linked β-D-mannopyranosyl residues that contain less than 5% galactose and small 384 amounts of other polysaccharides (Aspinall, 1959). Therefore, considering the 385 monosaccharide's profile obtained from the acid hydrolysis of açaí seeds (Table 1)  further studies to elucidate the carbohydrate structure will be necessary.    1999). Additionally, it has been shown that at room temperature, glucose and mannose 465 present a lower carbonyl percentage than galactose, xylose, and arabinose (Hayward 466 and J. Angyal, 1977). These data correlate to the low concentration of the sugar 467 degradation products found in the hydrolysates and to the fact that water-insoluble and

492
Nevertheless, in this study it was decided to avoid their formation during hydrolysis. 493 Therefore, a sequential process of enzymatic hydrolysis with the mannanases of the 494 recovered solids was evaluated to attempt to further release the mannose. The native açaí seed sample was highly recalcitrant to the enzymatic attack, 506 resulting in a less than 3% mannose yield. Nonetheless, after the material had been 507 partially digested by sulfuric acid, it became much more susceptible to the attack of the 508 mannanases, resulting in a 90% mannose yield for samples that were treated for 60 min 509 with 3%, 3.5%, and 4.5% sulfuric acid. Consequently, the recovery of mannose could 510 be substantially increased through a sequential process of diluted-acid hydrolysis   (Düsterhöft et al., 1991), which is the same structure that we hypothesized for the 531 mannan in açaí seeds. However, in the present study, native açaí seeds were poorly 532 hydrolyzed by mannanases, reaching only 3% conversion of mannan to mannose, 533 25 suggesting that these residues have distinctive mannan structures and/or the mannan is 534 less accessible in açaí seeds because of interactions with other structural and 535 nonstructural components.

536
The set of experiments presented in Figure 4 were performed with a 2% açaí 537 seed content (w/w), which led to high yield but also to hydrolysates containing low 538 concentrations of mannose of about 11 g/L of at the best conditions. To have an 539 effective industrial process, it is of the utmost importance to work on concentrated 540 media to reduce the capital cost of equipment and the use of water. Therefore, the 541 effect of solids loading on the enzymatic hydrolysis was evaluated in a range of 2-20% 542 ( Figure 5). Samples treated with 3% acid for 60 min were selected for the assays 543 because the seeds treated with 3%, 3.5%, and 4.5% of sulfuric acid were equally 544 susceptible to mannanase attack (Figure 4), and this condition has a lower impact on 545 the use of H 2 SO 4 , formation of acidic effluents, and degradation products.

564
At 72 h, the glucose concentrations reached 0.9, 2.8, 7.3, 9.2, and 13.7 g/L for the 565 assays containing 2%, 5%, 10%, 15%, and 20% of acid-hydrolyzed açaí seed, 566 respectively. It is very interesting to note that roughly, a glucose:mannose ratio of 1:10 567 could be observed (See Suplementary Material). Considering that the enzyme used is a 568 mannanase with nearly no cellulase activity, we hypothesize that the glucose released 569 27 during enzymatic hydrolysis is derived from the mannan structure. The 570 glucose:mannose ratio of 1:10 derived from mannan hydrolysis is in agreement with 571 the definition of a "true" mannan, which relates to polysaccharides with more than 85-572 95% mannose content and a high degree of uniformity in the structure (Aspinall, 1959;573 Stephen, 1983).

591
(2016), reported a direct increase in the production of simple sugars with an increase of 592 the PKC content, which indicated that no substrate inhibition effect was taking place. 593 However, the authors did not present a hydrolysis profile over time or the PKC 594 characterized, which restricts other comparisons with the present study.

595
A recent study has shown that 15 g of mannose could be obtained for every 100 596 g of spent coffee ground (SCG) after the removal of the non-saccharide content after 597 delignification and defatting of SCG (Nguyen et al., 2019). At the present study, 57.5 g 598 of mannose could be obtained for every 100 g of in natura açaí seed, with a total 599 mannose recovery of 98.6%. Figure 6 presents the mass balance of the overall process 600 for the mannose release from açaí seeds considering the sequential process of diluted 601 acid hydrolysis and an enzymatic hydrolysis step with 20% solids. The results 602 presented in the current study demonstrate that mannan from açaí seeds could be a low-

616
In this pioneer study, a sequential process of diluted-acid and enzymatic 617 hydrolysis of açaí seeds was developed to convert its high mannan content into 618 mannose. Mannanases-catalyzed hydrolysis of acid-treated seeds resulted in 146 g/L of 619 mannose and a 96.8% yield. To the best of our knowledge, this is by far the highest