本研究分為兩部份來進行。第一部份是利用快速分子生物檢測方法，確認預先經過自然酸篩處理的污泥中可以篩選出生成內孢子的高產氫菌群，且以Clostridium菌種為主。進一步利用各兩座體積分別為4升及1.5升的連續流式反應槽(CSTR)，進行不同污泥來源、基質馴養及水力停留時間(HRT)差異的操作，並比較其菌群結構。利用16S rRNA序列資料庫(clone library)建立法，成功地將其中三個樣品（LG：污水廠污泥，基質:glucose，HRT=10hrs和HRT=6hr，以及LS：污水廠污泥，基質:sucrose，HRT=5hrs）中菌群結構解析出，並與DGGE細菌菌群指紋譜進行比對相對位置。比對結果指出，於不同污泥來源(污水廠、酒廠及豆類污泥)、相同基質(蔗糖)與HRT (14.4小時)的操作下，厭氧生物產氫系統中的微生物菌群結構相似，與污泥來源關係不大，是屬於low G+C， gram-positive，與Clostridium pasteurianum親緣相近的菌群，產氫速率維持在107~137 (ml H2/g VSS-hr)左右。此外，相同污泥來源，不同基質(葡萄糖或蔗糖)及不同HRT操作下，隨著HRT的增加(2小時~10小時)，厭氧生物產氫系統中的微生物菌群多樣性顯著提高，比對位置結果顯示主要有三類菌群。其中兩類是屬於Gram-positive，low G+C bacteria，一類是與Clostridium pasteurianum親緣相近的菌群，而另一類是與Clostridium ramosum親緣相近的菌群。而第三類是屬於γ-Proteobacteria，與Pseudomonas親緣相近的菌群。在高HRT時，有較高的基質轉化率，各為1.65 mol H2/mol-glucose及3.94 mol H2/mol-sucrose；相對於低HRT時，則有較高的產氫速率，各為20.4 mmol H2/g VSS-hr (基質為glucose)及35.0 mmol H2/g VSS-hr (基質為sucrose)。進一步結合定量FISH以及水質分析結果，顯示屬於γ-Proteobacteria菌群幾乎可以忽略。另外LG反應器的菌群在不同HRT下有明顯的變動情形。其中，無法與Chis 150以及其他探針進行雜交的菌群，比例達到32.5~41.7%，甚至多到71%時，有較高的產氫速率，較低的基質轉化率，代謝產物主要為丙酸及乙醇，而乙酸和丁酸的濃度相對較低。相反地，屬於Gram-positive ，low G+C，與Chis150 probe可以進行雜交的菌群佔優勢時(94.5~97.5%)，代謝產物主要為乙酸及丁酸，而丙酸和乙醇濃度相對較低。故推論上述兩種菌群，也就是無法與Chis 150以及其他探針進行雜交的菌群，以及屬於Gram-positive ，low G+C，與Chis150 probe可以進行雜交的菌群，應是屬於基質競爭的關係，且這兩個族群應該是扮演主要產氫的角色。至於LS反應器，其菌群組成比例較為穩定。故基質種類、HRT、代謝產物和菌群種類皆會影響到厭氧產氫的效率。此外，探針測試結果顯示37℃、25% 的formamide濃度，是Lg10-6 probe的最佳雜交條件。進一步得到Lg10-6 probe與Chis150 probe 對於不同HRT操作下的污泥，其FISH偵測結果幾乎完全疊合(>99%)。 第二部份同樣是以16S rDNA為分析指標，來探討於另一個污泥來源(食品廠污泥)，更複雜的基質(peptone)，經過多次轉植馴養後之厭氧生物產氫微生物菌群多樣性。定量FISH結果顯示，主要有四大類菌群實際存在於peptone為基質之馴養試程，它們是屬於δ, α -Proteobacteria，以及屬於Gram-positive，low G+C，與Chis150 probe可以或不能進行雜交的菌群，分別佔細菌群菌群的30.4 ± 11.0 %、19.0 ± 10.1 %、11.0 ± 3.2 %和38 %。它們參與複雜的代謝peptone產氫反應。 By using molecular biological techniques, the microbial community structures of two hydrogen producing CSTRs fed with glucose and sucrose respectively were investigated. In the reactor fed with glucose, cloning results indicated that the ecology contained members of Gram Positive Low G+C (LGC) group and beta-Proteobacteria. Otherwise, the microbial community of sucrose feeding reactor was composed of bacteria affiliated with LGC group and gamma-Proteobacteria. Fluorescence in situ hybridization further revealed that three morphotypes of bacteria existed in both microbial communities. One is long rod bacteria which can be targeted by Chis150 probe designed to hybridize LGC bacteria (Franks ei al., 1998). Another type is curved rod bacterium not detected by Chis150. The other morphotype is short rod bacteria phylogenetically familiar with Pseudomonas spp. Moreover, the quantitative results of FISH clearly showed that Chis150 targeted cells and curved rod bacteria dominated both microbial community but in different ratio, whereas the amount of gamma-Proteobacteria rods were negligible. Interestingly, the population ratios of the two predominant groups reflected different operational performance of the two reactors such as hydrogen producing rates, substrate turnover rates and metabolites compositions. Therefore, a competition mode of the two dominant bacteria groups was hypothesized. On the other hand, microbial community of hydrogen producing batch reactor fed with peptone was also studied. FISH indicated that the community was more diversified than those of reactors fed with glucose or sucrose. Short rods akin to the δ-Proteobacteria subdivision and bacteria untargeted by probes used in this study accounted for 30.4% and 38% of EUB338 stained cells respectively. Besides, members of α-Proteobacteria and rods can be targeted by Chis150 probe were also abundant in the microbial community (19.0 ± 10.1 % and 11.0 ± 3.2 % each). The diversity of the society structure may be due to the complex components of peptone substrate. In general, feeding substrate played an important role in determining the bacterial community composition.
