Introduction: Despite many advances in therapeutic and surgical techniques for glioblastoma multiforme (GBM), this form of brain cancer still remains incurable. A hallmark feature of GBM is the ability of the glioma cells to infiltrate surrounding brain tissue. The invasive nature of glioma cells is a key challenge in considering treatment for patients with GBM. Certain members of the matrix metalloproteinase (MMP) family play a role in tumor cell invasion and metastasis (Coussens, et al., 2002). A functional SNP resulting from an additional guanine at position -1607 in the MMP-1 promoter creates an erythroblastosis twenty six transcription factor protein (ETS) DNA consensus binding site, which results in significantly higher transcriptional activity of MMP-1 (Rutter et al., 1998). Several published studies show the incidence of this 2G allele is significantly higher in aggressive and metastatic tumors. Binding of an adjacent transcription factor DNA consensus site, activator protein -1 (AP1) site at -1607 has been shown to cooperate with ETS binding to activate transcription of the MMP-1 gene. We have reported a significant increase in the 2G/2G MMP-1 genotype in glioblastomas (pPurpose: To determine if a novel SNP decoy can inhibit the 2G genotype-dependent increase in MMP-1 transcriptional activity, three specific aims were tested: one, to verify specificity of binding of a transcription factor decoy designed to mimic the -1607 SNP site within the MMP-1 promoter; two, to determine the effect of transcription factor decoy ODN on transcriptional activity of an MMP-1 promoter containing the 2G SNP at -1607; and three, to assess the effect of the transcription factor decoy ODN on MMP-1 mRNA and protein expression in treated glioma cells. Methods: Modified and unmodified decoys were designed to mimic position -1607 to -1593 of the MMP-1 promoter. The SNP decoy contains both ETS and AP1 DNA consensus sites and MMP-1 flanking sequences. We first determined optimal binding conditions with electromobility shift assays (EMSAs). The EMSA assays were used to determine the presence of Ets-1 and AP1 DNA binding activity within the glioma cell lines, T98 and U87. EMSAs were also used to determine if these transcription factors could bind to the MMP-1 promoters with and without the SNP. Lastly, EMSAs were done to determine the binding characteristics of the two modified SNP decoys (LNA-locked nucleic acid, and a PS-phosphothioate modification). The effect of the decoy on MMP-1 transcriptional activity was assessed using a Dual-Luciferase Reporter Assay. The effect of the SNP decoys on mRNA was assessed using quantitative RT-PCR, and on protein expression using a sandwich enzyme-linked immunoassay (ELISAs). Statistical analysis was done using a two-way ANOVA to evaluate the effect of the decoy on MMP-1 transcriptional activity, and protein expression. Results: EMSA results indicate that Ets-1 and AP1 probes, and MMP-1 promoter probes effectively bind proteins from glioma cell nuclear extracts. Addition of excess decoy was able to inhibit protein interactions with the 2G MMP-1 promoter probe and to a lesser extent the 1G promoter probe. The scrambled decoy had no effect. Promoter studies showed a significant increase in transcriptional activity of the 2G promoter and addition of 5 mm PS-SNP decoy could effectively prevent the increase in activity (pConclusions: U87 and T98 cell lines contain DNA binding activity of the transcription factors of interest, namely ETS-1 and AP1. The candidate transcription factors can bind to the MMP-1 promoter in the presence or absence of the 2G. Both the LNA and PS-SNP modified decoys can inhibit nuclear proteins from binding to the MMP-1 2G promoter. The PS-SNP decoy was able to inhibit MMP-1 (2G) gene transcription in a dose dependent manner, whereas the control decoy showed a consistent non-specific effect. The PS-SNP decoy inhibited MMP-1 mRNA and protein expression in glioma cells containing the 2G genotype, and to lesser extent in glioma cells containing the 1G genotype. The results presented here support the conclusion that the chimeric SNP decoy can selectively inhibit the MMP-1 promoter containing the 2G genotype.

Elevated expression of proteins, such as those involved in native antibiotic resistance pathways or introduced to enable biosynthesis of a metabolic engineering target, frequently leads to increased fitness cost. This can result in reduced growth and places selective pressure on cells. In conditions where there is diversity in expression within the population, this can result in cells with higher fitness out-competing their low-fitness counterparts. In the antibiotic resistance context, differential fitness costs caused by antibiotic resistance machinery can be exploited to select against resistant bacteria. However, in biotechnology applications, introducing burdensome synthetic constructs often requires additional engineering to increase genetic stability and maintain production. In this thesis, we investigate the origin of fitness costs and strategies for either exploiting or reducing it, focusing on specific examples related to antibiotic resistance and metabolic engineering. In the resistance work, we study the multiple antibiotic resistance activator MarA and related proteins in Escherichia coli. We quantify the differential fitness cost impacts of salicylate on E. coli antibiotic resistance variants. We demonstrate that salicylate, the natural inducer of MarA, imposes a higher fitness cost on resistant cells compared to the susceptible counterparts, making it possible to bias bacterial population membership towards those cells that are susceptible. In a second study, we focus on the role of salicylate in antibiotic tolerant persister cell formation, finding that salicylate induces reactive oxygen species and consequently persistence. In the metabolic engineering parts of the thesis we first review the mechanisms of fitness cost and existing strategies to ameliorate cost and cell-to-cell variation. Next, we present a technique for reducing fitness cost while maintaining production that takes advantage of transcription factor decoy sites to regulate biosynthesis in E. coli. Using arginine production as a model system, the transcription factor decoy is able to increase production by 16-fold without detectable growth differences. Together, the thesis provides an understanding of the origins and mechanisms of fitness cost in the context of antibiotic resistance and metabolic engineering. It also introduces strategies to exploit fitness costs to select against resistant bacteria and engineering strategies to ameliorate cost while increasing production and genetic stability.