【作者】 邹伟；

【导师】 陈坚；

【作者基本信息】 江南大学 ， 发酵工程， 2013， 博士

【摘要】 本论文以维生素C工业生产菌株生酮基古龙酸菌(Ketogulonicigenium vulgareWSH001)和巨大芽孢杆菌(Bacillus megaterium WSH002)组成的人工微生物生态系统为研究对象，结合目前关于两菌生理生化水平和组学水平(基因组学、蛋白组学、代谢组学)研究结果，通过构建基因组规模代谢网络模型(Genome scale metabolic model, GSMM)和基于约束的算法(constraints-based methods)，从系统生物学水平解析K. vulgare的生理特性及其与B. megaterium之间的相互作用机制，主要研究结果如下：1.基于RAST、KAAS、PRIAM、本地BLASTp注释K. vulgare基因组，共有834个蛋白注释出EC号且具有较高可信度。根据KAAS和本地BLASTp搜索TCDB数据库注释K. vulgare基因组共预测476个转运蛋白。基于本地BLASTp和比较基因组学显示K. vulgare RAST新注释的231个蛋白中186个具有功能注释或在K. vulgareY25中存在对应同源蛋白；B. megaterium RAST新注释的219个蛋白中153个具有功能注释或在B. megaterium QM B1551或DSM319中存在对应同源蛋白；2.通过代谢网络自动构建服务器Model SEED与KAAS，结合文献调研K. vulgare相关的生化信息、公共数据库、已报道实验结果等构建了K. vulgare的GSMM，命名为iWZ663，该模型包含663个基因、649个代谢物和830个反应。模型iWZ663中基因覆盖率达21.4%，反应可分为14个代谢亚系统，其中转运系统、碳水化合物代谢、氨基酸代谢所占比例最大，分别为16.5%、15.3%、15.2%；代谢物中与能量相关(ATP、ADP、NADP、NADPH、NAD、NADH)和氮代谢相关(谷氨酸、胺离子、甘氨酸)的代谢物在网络中占较大的连通度。模型注释K. vulgare山梨糖代谢途径发现山梨糖不仅可以转化成2-KLG或维生素C，也可进入中心碳代谢，为细胞生长提供能量和骨架；3.运用基于约束的算法结合Cobra工具箱在MATLAB平台上对K. vulgare GSMMiWZ663系统分析，发现模型iWZ663中116个基因被预测为生长必需基因，且均位于染色体上；153个反应被预测为生长必需反应。K. vulgare单独生长微弱的原因在于：(1)天冬酰胺、半胱氨酸、甲硫氨酸、生物素、烟酸、焦磷酸硫胺素、二氢叶酸的合成途径不完整；(2) K. vulgare代谢山梨糖主要通过ED (Entner-Doudoroff)途径和戊糖磷酸途径而不是酵解途径，其中戊糖磷酸途径流量仅占碳流的5.7%，低水平的戊糖磷酸途径影响细胞还原力(NADPH)和碳骨架(戊糖等)的供给；(3) K. vulgare不能还原硫酸盐生成亚硫酸盐，阻碍了半胱氨酸和甲硫氨酸的合成并进一步影响胞内辅酶A和谷胱甘肽的合成。K. vulgare具有强大的多肽转运和水解能力，K. vulgare能代谢天冬酰胺、天冬氨酸、谷氨酸、谷氨酰胺、甘氨酸、丙氨酸、脯氨酸、丝氨酸、苏氨酸参与TCA循环或核苷酸代谢；4.整合RAST新注释基因、代谢组学数据、碳氮源生长数据、亚细胞定位信息，将前期实验室构建的B. megaterium代谢网络模型iMZ992精炼升级为iMZ1055，该模型包含1055个基因、1137个代谢反应、1011个代谢物。统一K. vulgare与B. megateriumGSMM格式得到修饰后的两菌模型iWZ663a和iMZ1055a，二者共有453个反应和548代谢物相同。比较两菌GSMMs发现iMZ1055a的代谢能力更为多样，表现在：(1)具有15个特有的代谢亚系统而iWZ663a只有一个；(2)相比于iWZ663a，在组氨酸代谢、缬氨酸、亮氨酸和异亮氨酸生物合成、丙氨酸、天冬氨酸和谷氨酸代谢、核黄素代谢中也具备特殊的代谢功能；(3)必需反应分析显示iWZ663a中必需反应所占模型总反应比例是iMZ1055a的2倍多，iMZ1055a有51个非必需共有反应在iWZ663a是必需的，主要分布在嘌呤代谢、嘧啶代谢、核黄素代谢、泛酸与辅酶A生物合成等；(4) iMZ1055a可以合成并转运到胞外的代谢物有78种，而iWZ663a仅能合成和分泌22种代谢物。另外两者GSMMs在果糖与甘露糖代谢、泛醌与其他类萜醌生物合成、硫代谢和苯甲酸降解、转运系统中存在不同的代谢机制；5.整合模型iWZ663a和iMZ1055a构建维生素C二步发酵两菌代谢互作模型，命名为iWZ-KV-663-BM-1055，包含1718个基因、1583个代谢物和1910个反应。FBA与鲁棒性分析显示两菌间既存在共生也存在竞争的生理关系。FVA与必需反应分析显示B. megaterium伴生下：42个K. vulgare单菌模型时无流量通过的代谢反应有流量通过；36个在K. vulgare单菌模型时的必需反应转变为非必需反应；38个B.megaterium单菌模型中的非必需反应，主要参与合成K. vulgare的必需营养物质如焦磷酸硫胺素、生物素、烟酸、泛酸、二氢叶酸等，可以影响K. vulgare在混菌系统中达到最大生长速率。基于iWZ-KV-663-BM-1055分析两菌相互作用机制为：(1)K. vulgare与B. megaterium之间的相互作用主要通过胞外代谢物和反应相联系；(2)B. megaterium分泌23种代谢物到K. vulgare，其中15种营养物质K. vulgare自身也能合成；(3) K. vulgare自身的多数氨基酸、维生素和辅酶的合成途径无流量通过，但不同核苷酸之间的回补反应、亮氨酸、异亮氨酸、缬氨酸、脯氨酸的合成却有流量通过；(4)两菌代谢相互作用受两菌转运系统注释水平的影响。更多还原

【Abstract】 This dissertation focuses on the elucidation of physiological chrateristics and interactionof an artificial microbial ecosystem (AME) consisting of Ketogulonicigenium vulgare andBacillus megaterium, which was used in the vitamin C industrial production. This can berealized by reconstruction of genome scale metabolic model (GSMM) of these two bacteriaon the basis of information of biochemical and omics studies (such as genomics, proteomics,metabolomics). With GSMMs, constraints-based methods can be used to investigate thephysiological features and ineractions with B. megaterium. The main results are as follows:1. The genome of K. vulgare WSH001was annotated by four approaches: RAST, KAAS,PRIAM, and a local sequence similarity search (BLASTp) of UniProtKB/Swiss-Protdatabase. Totally,834proteins were annotated and assigned with an EnzymeCommission (EC) number, which can be further used in the metabolic reconstruction.For transport proteins, KAAS annotation and a BLASTp with TCDB database wereperformed and476transportors were annotated. In addition, new annotated genesidentified during the RAST annotation were rechecked by a BLASTp ofUniProtKB/Swiss-Prot database and a comparative genomics analysis:186genes out of231new genes in K. vulgare and153genes out of219new genes in B. megaterium wereproved because of a high similarity or having homologous protein(s) in other strains ofits species.2. We constructed the GSMM of K. vulgare, iWZ663, on the basis of Model SEED, KAAS,literature mining, public databases, and experimental data. It consists of663genes,649metabolites and830reactions. The gene coverage of model iWZ663is21.4%. Themodel reactions were divided into14metabolic subsystems, among which Transportingsystem, Carbohydrate metabolism, and Amino acid metabolism occupy the largestproportions:16.5%,15.3%, and15.2%, respectively. Metabolites that associated withenergy generation and nitrogen metabolism have the most extensive connectivities iniWZ663. Annotation of L-sorbose metabolic pathway indicated that L-sorbose can notonly be converted into2-KLG or vitamin C, but also enter into the central carbonmetabolism for the production of energy and biomass precursors.3. Model iWZ663was comprehensive analyzed with Cobra toolbox on MATLAB. Essentialgene analysis was performed and116genes were predicted to be essential, and all theessential genes are located on chromosome. In addition,153reactions were predicted asessential reactions. Flux balance analysis (FBA) was carried out to investigate thereasons for the poor growth of K. vulgare, and three main reasons were included:(1) K.vulgare could not de novo biosynthesize asparigine, L-cysteine, L-methionine, biotin,nicotinate, thiamine diphosphate, and dihydrofolate (DHF);(2) the carbon flux mainlyenter into the ED pathway and only5.7%into the PP pathway, resulting a lower level ofreducing power (NADPH) and shortage of carbon backbones such as ribose;(3) thedefect in sulfate metabolism hampering the syntheses of L-cysteine, L-methionine,coenzyme A (CoA), and glutathione. Meanwhile, K. vulgare has aboundant peptide transportors and peptidase and was capable of assimilating L-asparagine, L-aspartic acid,L-glutamic acid, L-glutamine, glycine, L-alanine, L-proline, L-serine, and L-threonineinto TCA cycle or purine metabolism.4. The previous B. megaterium GSMM iMZ992was reifined by integrating the informationof RAST new annotations, metabolomics data, growth phenotype data on differentcarbon and nitrogen sources, and protein sublocation information. The new model namediMZ1055, comprised of1055genes,1137reactions, and1011metabolites. Then, the twomodels iMZ1055and iWZ663were reconciled (named iMZ1055a and iWZ663a,respectively) and compared. It was found that they share453reactions and548metabolites. Comparison of the two GSMMs suggested that B. megaterium has a morediversity metabolism, because:(1) iMZ1055a has15unique metabolic subsystems whileiWZ663a has only one;(2) iMZ1055a has some special metabolic functions in Histidinemetabolism, Valine, leucine and isoleucine biosynthesis, Alanine, aspartate andglutamate metabolism, and Riboflavin metabolism;(3) the percentage of essentialreactions in iWZ663a was more than two folds of that in iMZ1055a, and iMZ1055a has51non-essential shared reactions which were essential for. These51reactions mainlydistributed in Purine metabolism, Pyrimidine metabolism, Riboflavin metabolism,Pantothenate and CoA biosynthesis;(4) iMZ1055a could biosynthesize and transport out78metabolites, while K. vulgare only22metabolites. In addition, the two models havedifferent metabolic mechanisms in Fructose and mannose metabolism, Ubiquinone andother terpenoid-quinone biosynthesis, Sulfur metabolism, Benzoate degradation, andTransporting system.5. A two-species metabolic interaction model of AME in vitamin C production wasconstructed by integrating iWZ663a and iMZ1055a. The resulted model, namediWZ-KV-663-BM-1055, consists of1718genes,1583metabolites, and1910reactions.FBA and robustness analysis found both mutualism and competition exist between K.vulgare and B. megaterium. Futher, FVA and essential reactions analysis revealed thatwith the companion of B. megaterium: forty-two reactions that could not carry flux iniWZ663a were capable of carrying flux; thirty-three reactions that were essential iniWZ663a became non-essential in iWZ-KV-663-BM-1055; thirty-eight non-essentialreactions in B. megaterium could affect K. vulgare reach its maximum growth rate.Further, analysis of metablic interactions between K. vulgare and B. megaterium shows:(1) metabolism of K. vulgare and B. megterium were connected mainly by extracelluarmetabolites and reactions;(2) B. megaterium could secrets23to K. vulgare, amongwhich15metabolites can be biosynthesized by K. vulgare;(3) most of reactions inamino acids, vitamins and cofactors pathways has no flux, excepting several reactions innucleotide salvage pathway and the biosynthesis of leucine, isoleucine, valine, and proline;(4) predictions of metabolic interaction were affected by the accuracy oftransporter annotations in iWZ-KV-663-BM-1055.更多还原