Name: Ramkrishna Singh
Department: CTARA (IITB Monash)
Name of supervisor: Prof. Amit Arora
The human gastrointestinal tract hosts an army of billions in the form of microbes in the gut. These microbes, belonging to about 1000 different types live in symbiotic association with the human host. These “good bacteria” uses undigested protein, non-digestible carbohydrates, and other undigested fraction of food reaching the distal part of the gut as an energy source. The presence of a healthy number of the good bacteria and the metabolites produced by them confer health benefits to the host like good overall gut health, reduced risk of pathogenic infection in the gut, improved metabolic control, improved immunity, etc. Non-digestible carbohydrates such as fructooligosaccharides, galactooligosaccharides, inulin, xylooligosaccharides, termed as prebiotics, can be as an energy source by gut bacteria. Xylooligosaccharides, a prebiotic, is obtained from hemicellulose, a component of the plant cell wall. Interestingly, billions of tonnes of residue result from agricultural, forestry, industrial and allied practices, comprising of cellulose, hemicellulose, and lignin along with other bioactive. The surplus quantum of such residue scores millions of tonnes. Also, with increased emphasis and incentive for cellulosic ethanol production, the availability of residue may see a further rise. Thus, xylooligosaccharides, has the advantage of being produced from abundantly available, cheaper starting material.
In this work, underutilized sources such as arecanut husk and almond shell were used for their potential valorization into prebiotic xylooligosaccharides. The alkaline pretreatment process was studied for maximal recovery of hemicellulose while achieving delignification of husk. The hydrothermal treatment (1 h) of biomass soaked in 20% NaOH was able to recover about 82% of available hemicellulose while causing 69% delignification. When the biomass suspended in alkali solution was treated with hydrothermally for 1 h (without soaking), only 75% of hemicellulose was recovered with 65% lignin removal. Further, when the hydrothermal time was increased to 1.5 h (without soaking of biomass), 83-86% of hemicellulose could be recovered with 15% w/v NaOH. Further optimization of alkaline pretreatment process was performed by evaluating incubation time (8, 16, 24 h), incubation temperature (25, 50, 65oC), alkali concentration (5, 10, 15, 20% w/v) and assistance of hydrothermal treatment. The results indicate that 10% alkali can recover more than 90% of available hemicellulose when the biomass is incubated at 65oC for 8 h and then treated hydrothermally for 1 h. However, when microwave irradiation was used as a heat source, 52% hemicellulose could be recovered within 3 min at 900 W with 15% alkali. Also, the alkali treated husk residue was concentrated in cellulose and could be enzymatically converted to glucose. Upon enzymatic treatment of residue with cellulase and β- glucosidase, 75-79% reducing sugar was obtained containing 69% glucose. Thus, the husk could be fractionated into hemicellulose and cellulose using alkaline pretreatment.
For the almond shell, the effect of particle size on hemicellulose recovery and alkali concentration was evaluated. The shell was subjected to ball milling or grinding to reduce the particle size. Upon alkali treatment at 4 and 8% w/v, the observation indicates that irrespective of the method of grinding, more than 90% hemicellulose can be recovered with 8% alkali when the particle size is less than 120 µm. The optimum reaction condition was 121oC, 1 h. Thus, the energy-intensive process of ball milling can be replaced with grinding while still achieving near complete hemicellulose recovery. It has been suggested that alkali pretreatment is effective for biomass with low lignin content, and thus 12-16% alkali has been reported for above 90% hemicellulose recovery. However, in this work, we have demonstrated that optimization of process, or reduction in particle size can achieve more than 90% hemicellulose at alkali concentration below 10%, even when lignin content is about 25%. As enzymes are an important cost determining factor of a process, in this work, efforts were made to optimize enzymatic reactions. An iron oxide-based magnetic nanoparticle was prepared and used as support for endoxylanase immobilization. The immobilized enzyme was found to be at par with the free enzyme in terms of pH and temperature profile, reaction kinetics. The immobilized enzyme, however, gave a lower yield of XOS as compared to free enzyme. However, when the enzymatic reaction was conducted at 50 and 60oC, immobilized enzyme gave similar XOS yield, whereas the yield of free enzyme decreased at 60oC as compared to 50oC. Interestingly, the immobilized enzyme has limited reusability, as the XOS yield was reduced to 41% in cycle 3, as compared to the first cycle. This may be due to longer reaction time necessary to maximize XOS yields. The response surface methodology based optimization of enzyme dose and substrate concentration indicates that 10 U of the enzyme is optimum to maximize XOS yield. The analysis of RSM data indicates that substrate concentration below 2% is optimal, as an increase in substrate concentration decreases the desirability of higher XOS yield.
The enzymatically produced XOS was refined using membrane assisted separation and Diaion WA 30 resin treatment. The filtration using 10 kDa membrane allowed permeation of more than 90% of xylobiose, xylotriose, xylose, and acetic acid while decreasing the colour intensity of liquor. This is due to the retention of higher molecular weight impurities including enzyme and higher molecular weight hemicellulose or undissolved hemicellulose. The filtration with 1 kDa did not produce a significant difference in terms of composition or colour intensity. Finally, the liquor was refined using 150 Da membrane, which retains 79% of xylobiose, 41% xylotriose along with 36-40% of xylose and acetic acid. At the end of membrane filtration, about 80% of initial low DP XOS was recovered. The resin treatment could remove colour impurities as indicated by a decrease in absorption at 230 and 280 nm from 3.8 to 1. The final colourless XOS concentrate was composed of 77% xylose based molecules, out of which 72% was present as XOS (xylobiose and xylotriose). The concentrate also contains about 10% acetic acid.
The almond shell XOS was similar to commercial corncob derived XOS in term of fermentation. The in-vitro fermentation using human fecal samples shows that XOS is rapidly fermented to produce short chain fatty acids and gases. Among, the acids, acetate was predominantly produced in case of XOS as compared to longer chain oligosaccharides such as fructooligosaccharides or inulin. The total gas produced upon XOS fermentation was not significantly different from that obtained by fermentation of commercial XOS, inulin or FOS. The in-vitro bacterial fermentation using Lactobacillus, suggest that XOS can be used as an energy source by the bacteria, as the bacterial population increases ten-fold in 48 h. Also, similar to human fecal fermentation, acetate was predominant SCFA. It was observed that the bacteria utilized xylobiose more as compared to xylotriose and xylotetrose. Thus, this study suggests that the almond shell can be used for the production of prebiotic XOS and justifies production of low DP.
As a green approach, autohydrolysis of the almond shell was studied for XOS production. For this a sequential autohydrolysis, enzymatic hydrolysis and membrane purification process was developed to obtain a low degree of polymerization xylooligosaccharides. The treatment of shell at 200oC for 5 min, corresponding to severity factor of 3.64, could hydrolyze 65% of xylan to produce oligosaccharides or 10.97g XOS/ 100g biomass was obtained. However, the percentage of low DP XOS (Xylobiose and xylotriose) was 3.5% of the biomass. Upon enzymatic hydrolysis, the contribution of low DP XOS increases to 8.5% of the biomass. Further membrane separation using 1 kDa and 250 Da membrane could recover 69% of low DP XOS.
Thus, the present work provides two approaches for the production of low DP XOS. A techno-economic evaluation of the developed process will help to realize the most economical and suitable process.