Show simple item record

dc.contributor.authorBaccile, Joshua Andrew
dc.date.accessioned2017-04-04T20:27:37Z
dc.date.available2017-04-04T20:27:37Z
dc.date.issued2017-01-30
dc.identifier.otherBaccile_cornellgrad_0058F_10124
dc.identifier.otherhttp://dissertations.umi.com/cornellgrad:10124
dc.identifier.otherbibid: 9906037
dc.identifier.urihttps://hdl.handle.net/1813/47790
dc.description.abstractRecent advances in genomic sequencing technology have facilitated relatively inexpensive, high-fidelity sequencing of microbial genomes. Increased genome sequencing capacity has been accompanied by developments in bioinformatics, which enable rapid homology searching to predict gene products, such as enzymes involved in the production of small-molecules. The recent surge in genomic information revealed that only a small percentage of genes involved in small-molecule biosynthesis had previously been characterized. In fact, it appears that we have only scratched the surface in understanding the full biosynthetic potential of microorganisms. The current void in small-molecule discovery is in part because of the particular challenges associated with identifying new compounds and in part because of the nature of microbial small-molecule production. Perhaps the largest obstacle in the path to characterizing new small-molecules is the absence of a templated structure. For instance, in genomics, the number of building blocks is extremely limited, just four different bases, which fit together predictably through phosphodiester bonds. Moving up in complexity to proteomics, the number of building blocks increases to roughly twenty amino acids, however still assembled in a predictable manner, through peptide bonds. Additionally, protein primary sequence can be derived directly from the corresponding genetic sequence, which coupled with homology searching enables quick judgments as to whether or not a detailed characterization is warranted. The situation in metabolomics is starkly more complicated as the number of building blocks increases to over one-hundred chemical entities that can be assembled in virtually any imaginable way that obeys the laws of chemistry. Moreover, small-molecule structural insights simply cannot be attained from the genetic sequence alone, meaning every compound must be individually characterized. In addition to the challenges of small-molecule characterization, the very nature of microbial biosynthesis has significantly hindered more thorough annotation of the metabolome. Expression of genes encoding biosynthetic enzymes is tightly regulated and responsive to various natural environments. Simple laboratory culturing conditions usually fail to reproduce the necessary cues to promote small-molecule production. While media screening methods or so called fermentation optimization is an option to overcome this barrier, the power of genetic engineering offers a far more straightforward solution and has the added benefit of precision. The production of small-molecules can be turned on or eliminated in a controlled manner by creating over-expression or knock-out mutant microbial strains. Comparison of mutant and wild-type strains using differential analysis by 2D-NMR spectroscopy (DANS) and LC-MS-based comparative metabolomics enables a systems biology prospective of metabolic changes resulting from genetic engineering. Described herein is the utilization of DANS and LC-MS-based comparative metabolomics to discover novel metabolites from orphan microbial biosynthetic gene clusters. Applied to the has gene cluster in Aspergillus fumigatus DANS revealed the biosynthesis of a novel iron(III)-complex, which was shown to increase A. fumigatus’ virulence. Results from the has study led to an investigation of diketopiperazine formation in the notorious gliotoxin biosynthetic pathway. Gliotoxin biosynthesis was shown to be dependent on a cluster-encoded cyclization domain, which ultimately functioned to cyclize a modified dipeptide into the diketopiperazine core of the gliotoxin structure. In another A. fumigatus gene cluster, called fsq, DANS and LC-MS-based comparative metabolomics revealed the elusive fungal isoquinoline formation pathway, which despite having been described in plants more than twenty years ago, remained unknown in fungi until now. The fungal isoquinoline pathway was shown to be conserved through the characterization of another isoquinoline producing gene cluster in the plant pathogenic fungus A. flavus. Lastly, the identification and characterization of an additional set of fsq-dependent metabolites, called the fumizinones, is described. The fumizinones derive from an alternative biosynthetic mechanism, resulting in pyrazinone formation, rather than isoquinoline formation, as observed for the main products of the fsq pathway.
dc.language.isoen_US
dc.subject2D NMR
dc.subjectaspergillus
dc.subjectbiosynthesis
dc.subjectlc-ms
dc.subjectmetabolomics
dc.subjectnrps
dc.subjectChemistry
dc.titleUncovering hidden microbial metabolism using 2D NMR- and LC-MS-based comparative metabolomics
dc.typedissertation or thesis
thesis.degree.disciplineChemistry and Chemical Biology
thesis.degree.grantorCornell University
thesis.degree.levelDoctor of Philosophy
thesis.degree.namePh. D., Chemistry and Chemical Biology
dc.contributor.chairSchroeder, Frank
dc.contributor.committeeMemberAye, Yimon
dc.contributor.committeeMemberCrane, Brian
dcterms.licensehttps://hdl.handle.net/1813/59810
dc.identifier.doihttps://doi.org/10.7298/X47H1GJZ


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record

Statistics