瘤胃菌酵素系統於纖維素生質轉換之應用

The application of rumen bacterial enzyme systems to convert cellulolytic biomass

陳栢元(Bo-Yuan Chen)

指導教授 ： 徐濟泰

國立臺灣大學/生物資源暨農學院/動物科學技術學研究所/博士(2021年)

https://doi.org/10.6342/NTU202100842

摘要

本論文以台灣本地所產之木質纖維生質（禾本科）為來源基質，經不同前處理後其成分變化的改變及後續酵素水解模式的搭配加以探討，且希望利用瘤胃細菌來尋找有效之水解酵素比例增水解效率。試驗分為三階段，第一階段，以稻草（台梗9號、台農14號）、狼尾草、盤固草與五節芒為低成本之木質纖維生質來源，藉由氨水（liquid ammonia）、氫氧化鈣（lime, Ca(OH)2）與過氧化氫-鹼處理（alkaline peroxide, NaOH-H2O2），進行不同片段大小(1-2 cm與20 mesh)、處理濃度(氨水3%或5%；氫氧化鈣 0.1%或0.5 %)與溫度(25℃或50℃)之測試，了解其組成變化，以尋求各種不同生質合適之前處理程序，作為後續調整酵素添加之組合與微生物培養組合之依據。前處理結果，氨水處理在較高濃度(5%)及較高溫度(50℃)作用下，對於半纖維素之去除有最佳之效果(降低30-40%)，但是對於木質素之處理效果不明顯。氫氧化鈣處理對於五節芒與台農14號具有較明顯之反應，而溫度與氫氧化鈣濃度具有加成效應。乾式過氧化氫-鹼處理時間過長(8小時以上)，將影響纖維素之殘留量，濕式處理則可有效移除半纖維素40%，並可移除50%之木質素。木質纖維生質成分組成明顯受到前處理方法的影響。過氧化氫-鹼處理對於生質組成改變較明顯，但於後續考量，希望保留較多之纖維及半纖維素之考量下，以石灰及氨進行處理，會是較好之選擇。
第二階段根據第一階段之結果，挑選適合之木質纖維生質及前處理搭配不同酵。此階段以氨水，石灰兩種鹼處理方式為基礎，針對五節芒、狼尾草及台梗九號在不同酵素水解模式下的產糖性能進行評估。試驗中，石灰處理加上混合酵素模式水解，其顯著提高了纖維生質的降解，尤其是在五節芒和台梗9號。但單一酵素使用（纖維分解或半纖維分解酵素），其對糖的轉化有一定的能力限制。纖維生質轉換的結果表示，未經處理之狼尾草，在混合酵素模式下，有相當高的產糖量。但鹼處理對狼尾草的葡萄糖轉換沒有進一步的影響。增加混合酵素的活性從2.89到10.68 FPU g-1，其增進氨處理後纖維生質之葡萄糖的產量從3.4到4.4倍，石灰處理後之纖維生質之糖產量從2.8到3.3倍。酵素混合模式明顯的改進了纖維生質的糖產量。相較於未處理，在五節芒及台梗9號稻草，因前處理後影響纖維生質的組成特性，使得後續酵素水解有明顯之效果。因此為提高纖維生質的水解效率，其前處理方法與酵素添加模式應同時考慮。此外，前處理方法應根據生質的纖維組成來加以使用。並針對單一纖維生質尋找合適的前處理方法與混合酵素的搭配組合。
第三階段試驗企圖以瘤胃纖維分解菌分泌之酵素模式搭配不同前處理生質，以改善木質纖維生質的轉化效率。試驗中，瘤胃纖維分解菌Fibrobacter succinogenes S85為一測試平台，以了解微生物對於不同基質來源刺激後，其酵素分泌之狀況。並以商業用酵素進行模擬試驗，了解糖的轉化效率。試驗中，不同前處理的生物質刺激F. succinogenes S85表現出不同程度的纖維素分解與木聚醣分解酵素活性。經由SDS-PAGE及酶譜(zymogram)分析，了解不同的前處理方法對於蛋白質分泌上沒有明顯的差別。而相關的纖維分解酵素之分布位置，亦表示了相類似的結果。說明木質纖維生質可能是主要影響F. succinogenes S85蛋白質分泌的因素。使用商業纖維分解酵素模擬F. succinogenes S85酵素分泌的比例進行生質之水解。當細菌分泌模型應用於生質降解，經氨處理之五節芒（SG）和台梗9號（TK）顯著提高了葡萄糖產量（分別高於等量酵素模式1.5和1.7倍）。另外，相較於等量酵素模式，細菌分泌模式亦增加經石灰處理之五節芒1.4倍及狼尾草1.95倍。不同前處理的生物質刺激F. succinogenes S85表現出不同程度的纖維素分解與木聚醣分解酵素活性。細菌之分泌酵素模式在生質水解的轉化率上亦有明顯之效果。因此，瘤胃纖維分解細菌（F. succinogenes S85）酵素分泌模式利用商業纖維素分解酵素進行組合，可用來作為一個高轉化率之酵素系統的參考。以合適的前處理搭配最佳纖維分解酵素模式來提高生質之糖產量。

關鍵字

木質纖維生質 ； 鹼前處理 ； 纖維分解酵素模式 ； 酵素水解 ； 瘤胃細菌

並列摘要

Five kinds of low cost cellulose biomass (Japanese silvergrass, Chinese Pennisetum, Pangolagrass and two rice straw) were applied in this study to test the optimal pretreatment procedure. All biomass were treated by liquid ammonia, lime and alkaline peroxide. The different particle size, concentration and reaction temperature were also tested during treatment procedure. The result showed that 5% liquid ammonia treatment under 50℃ removed 30-40% hemicellulose from biomass, but no significant effect on lignin. The lime treatment had high efficiency on Japanese silvergrass and rice straw, the reaction condition test also showed additive effect between temperature and lime concentration. The alkaline peroxide treatment result indicated that the long time (over 8 hr) dry treatment may reduced the cellulose content in biomass, over 40% hemicellulose and 50% lignin were removed after wet treatment procedure. All treatment in this study indicated that reduced the biomass particle size was beneficial to treatment efficiency.
Two commonly alkaline pretreatment processes base on aqueous ammonia and lime under different enzyme hydrolysis models were evaluated to provide comparative sugar production performance from silvergrass, napiergrass and rice straw. The chemical composition variation of all biomass were nearly in stable after 4 weeks pretreatment under room temperature and recovery of the cellulose fraction was >90% by both pretreatment methods, the silvergrass recovered more dry matter than other biomass after pretreatment. Compared with other combination of pretreatment and enzyme model, mixed enzyme model after lime pretreatment significantly enhanced the biomass degradation especially in silvergrass and rice straw, but single enzyme supplement (cellulase or hemicellulase) result in limited sugar yield in this study. The biomass conversion result showed that considerable sugar yield from untreated napiergrass under the mixed enzyme model. However, alkaline pretreatment had no positive effect on glucose conversion from napiergrass. Increasing the mixed enzyme activity from 2.89 to 10.68 FPU/g improved the glucose yield from 3.4 to 4.4 times and from 2.8 to 3.3 times after ammonia and lime pretreatment, respectively. The findings of this study suggest that pretreatment methods and enzyme loading model should be considered simultaneously to enhance cellulosic biomass degradation. Furthermore, the pretreatment method should be applied according to the fiber composition of the biomass. The suitable pretreatment process and constituent of enzyme mixture for individual cellulosic biomass is a promising line of inquiry.
In this study, we attempted from the enzyme secretion model of F. succinogenes S85 with different alkaline pretreatment biomass to understand the enzyme synergy and ratio. Further, determining the effect of enzyme model on cellulosic biomass conversion with commercial enzyme. In order to building a method to find the optimal cooperation model of biomass enzyme loading for maximum sugar yield. The cellulolytic biomasses were including silvergrass (S), napiergrass (N) and Taikeng 9 straw (T). Using lime (0.1 g / g, L) and ammonia (3%, A) treated biomass as the carbon source of culture media for F. succinogenes S85 growthing.
From the results of SDS-PAGE and zymogram analysis, the different pretreatment methods seem less influence on protein secreted pattern. Indicating that kinds of biomass may be the main factors to influence F. succinogenes S85 secreted protein. Further, according to the correlations between secretion enzyme activity and biomass components. The fiber chemical composition seems to provide limited information for suitable enzyme pattern of degradation. It suggested that fiber structure, morphology and enzyme binding situation might affect the enzyme degradation ability. Using commercial cellulosic enzymes to simulate the F. succinogenes S85 enzyme secreted. When bacterial secreted model was applied to biomass degradation, the ammonia pretreatment significantly improved the glucose production from SG and TK9 (1.5 and 1.7 times higher than equal mixed enzyme model, respectively). Furthermore, bacterial secretion model increased glucose production that 1.95 times (3791 μg/mL) from SG and 1.4 times (2958 μg/mL) from NG after lime pretreatment than equal mixed enzyme model (1940 and 1992 μg/mL, respectively).
Using microorganism to find the ratio model of cellulolytic enzyme cocktail by different biomass stimulate which might provide the way to find the optimize multi cellulolytic enzyme ratio to enhance the hydrolysis efficiency. The secreted enzyme ratio form rumen cellulolytic bacteria stimulated by different pretreated biomass could be a useful reference platform to understand how cellulolytic bacteria degraded the biomass at high efficiency.

並列關鍵字

Lignocellulosic biomass ； Alkaline pretreatment ； Cellulytic enzyme model ； Enzyme hydrolysis ； Rumen bacteria

參考文獻

王翰聰。2004。瘤胃細菌纖維及蛋白質分解酵素之生產與利用。國立台灣大學畜產學研究所博士論文。

許福星等。1994。芻料作物生產與利用。行政院農委會畜產試驗所。

謝文彰。2006。有機芻料生產與利用之研究。農委會畜產試驗所科技計畫。

Abou-El-Enin, O.H., J.G. Fadel, and D.J. Mackill. 1999. Differences in chemical composition and fibre digestion of rice straw with, and without, anhydrous ammonia from 53 rice varieties. Animal feed science and Technology. 79:129–136.

Anthon, G.E., and D.M. Barrett. 2002. Determination of Reducing Sugars with 3-Methyl-2-benzothiazolinonehydrazone. Analytical Biochemistry. 305:287–289.