Master of Science / Department of Biochemistry / Michal Zolkiewski / Early-onset dystonia (EOTD, also known as DYT1 or Oppenheim’s dystonia is the most severe and common form of hereditary dystonia, a neurological disorder characterized by abnormalities in the control of movement. It is linked to the deletion of a single GAG codon in the gene DYT1 that leads to the loss of a single glutamic acid residue in the C-terminal region of the protein torsinA (ΔE-torsinA). It is not known how the GAG deletion alters the torsinA structure and function. In this thesis, the expression and purification of recombinant torsinA variants from E. coli is reported. Wild type torsinA is not soluble after its expression in E. coli, possibly due to misfolding caused by cysteine. We produced Cys-less torsinA, and established a purification procedure to produce this mutant torsinA. Furthermore, because of the critical role likely to be played by the C-terminal domain of torsinA that contains the glutamate deletion, we produced fragments encoding the C-terminal domain of torsinA, and attempted to purify it. However, we failed to obtain appreciable amount of active proteins by both of the strategies. A novel SUMO fusion technology was also used in this study. We demonstrated that SUMO, when fused with torsinA variants, was able to enhance its expression and solubility in E. coli. A satisfactory yield of the fusion protein was successfully purified. Once we get appreciable quantities of folded torsinA variants, it is our future goal to study their function by using biochemical and high-resolution structural approaches.

Doctor of Philosophy / Department of Biochemistry / Thomas E. Roche / Specific mutants were developed to evaluate the roles of several residues in [Alpha]8 helix of the regulatory (R) domain of human pyruvate dehydrogenase kinase 2 (PDHK2) in the linkage between the Regulatory (R) and catalytic (Cat) domain (Q144A), dichloroacetate (DCA)/pyruvate inhibition (R154C, R158A, I157F) and stimulation by reductive acetylation (L160A, R154C/L160A). All mutants, with the exception of L160A, were active, and were bound to and had their activities enhanced by dihydrolipoyl acetyltransferase (E2). The cross arms between subunits are anchored by W383. Based on the studies on the W383F mutant, W383 provided majority of the intrinsic Trp fluorescence; and ligand(s) binding quenched primarily (pyruvate) or exclusively (ADP or ATP) the fluorescence of W383. The Q144 mutation in the R domain caused 14-fold weaker K[superscript]+ binding with ATP in the Cat domain but did not alter the weaker K[superscript]+ binding with ADP unless Pi was included. Similarly, with 100 mM K[superscript]+, the Q144A mutant had weaker ATP binding but the affinity for ADP was not changed even in the presence of Pi, which enhanced the binding of ADP to kinase by 2-fold. R154 and R158 were shown to be important residues in the inhibition by pyruvate, DCA and Cl[superscript]-. The R154C, R154C/L160A and R158A mutations reduced the inhibition by DCA or pyruvate using E1 or E1[dot in middle of line]E2 as the substrates. Pyruvate plus ADP did not significantly hinder the binding of GST-L2 to these mutants in AUC studies. Cl[superscript]- appears to bind to kinase at the same site as DCA/pyruvate based on lack of Cl[superscript]- effects with above mutants and evidence that Cl[superscript]- weakened the inhibition by DCA or pyruvate of native PDHK2. Q144 and L160 may play important roles in the signal transmission from the lipoyl group-binding site to the active site. Q144A and R154C/L160A mutants were less stimulated by reductive acetylation than other mutants and native PDHK2. Nov3r binds to where lipoyl goup binds PDHK2. Using E1 alone as substrate, Nov3r binding caused a 20% increase of kinase activity at low levels. Nov3r binding also reduced the inhibition of DCA/pyruvate with elevated K[superscript]+ plus Pi.

Master of Science / Department of Biochemistry / Gerald R. Reeck / Armet is a bifunctional protein widely distributed in animal species, vertebrate and invertebrate. It is an evidently part of the Unfolded Protein Response (UPR) and promotes survival in cells that are under endoplasmic-reticulum (ER) stress. It has also been found as a secreted protein with neurotrophic activity. The crystal and solution structures of human Armet show it is a helix-rich protein with two domains linked through a flexible linker region. In this study, immunofluorescence staining was used to verify Armet’s localization in ER and Golgi apparatus in MBA-MD-231 cells. Evidence for calcium binding by Armet was obtained by circular dichroism spectroscopy (the binding of calcium appeared to decrease helix content), by differential scanning calorimetry (binding of calcium resulted in a less structured protein) and two-dimensional (1H-15N HSQC) nuclear magnetic resonance spectroscopy. A difference HSQC spectrum of Armet, with and without calcium, showed peaks of increased intensity, of decreased intensity and of perturbed chemical shift. There were about 30 such peaks in total. Several of these affected amino acid residues appeared to form a cluster of negatively charged side chains that could possibly form a binding site for a calcium ion. Heterogeneity of three types was observed in recombinant Armet expressed in E. coli cells. Two bands of slightly different mobility were observed in SDS gels run in the absence of reducing agent. These may represent alternate arrangements of disulfide bonds, as previously reported by other investigators but not explained. Further, in the absence of reducing agent, a faint ladder was formed by human Armet, indicating formation of disulfides between Armet molecules. Oligomers with sedimentation coefficient greater than the monomeric protein, in the absence of reducing agent, disappeared in the presence of a reducing agent. Finally, minor species of mass differences of 98 and 180 with respect to the main protein component were observed by MALDI-TOF mass spectrometry. These studies provide a more thorough characterization of Armet than has been previously available and set the scene for future investigations of the binding of organic ligands to the protein.

Master of Science / Department of Anatomy and Physiology / Lei Wang / Jishu Shi / Transforming growth factor beta (TGF-β) and interleukin-1β (IL-1β) are both up-regulated in high grade gliomas and their elevated activities have been associated with prognosis in glioma patients. It is known that TGF-β is involved in proliferation and maintenance of glioma stem cells. In this study, I evaluated whether IL-1β also plays an important role in glioma stem cell development. Glioma stem cells are usually identified by using a sphere assay where glioma stem cells proliferate as neurospheres in serum free medium (SFM) in the presence of two growth factors: EGF and bFGF. However, LN229, a human glioblastoma cell line does not form neurospheres in SFM, suggesting that LN229 cells contain very few stem cells. I found that combination of IL-1β and TGF-β, but not IL-1β or TGF-β alone induced LN229 cells to grow as neurospheres in SFM. Furthermore, quantitative RT-PCR analyses show that the expression of stem cell markers (Nestin, Bmi1, Notch2, and LIF), cytokines (IL-1β, IL-6 and IL-8) and invasive genes (SIP1, β-integrin and N-Cadherin) are significantly enhanced in IL-1β /TGF-β induced spheres compared to the control. Using an invasion assay, drug resistance test, and colony assay, I found that LN229 sphere cells induced by IL-1β and TGF-β are more invasive, have increased drug resistant ability, and are more oncogenic in comparison to the control. Together, these results suggest that IL-1β cooperates with TGF-β to induce glioma stem-like cell phenotype.

Master of Science / Graduate Biochemistry Group / Michael R. Kanost / Prophenoloxidase-activating proteinase (proPAP) activates prophenoloxidase when bacteria or fungi invade Manduca sexta. Upon activation, phenoloxidase initiates synthesis of melaninin, which can encapsulate the invaders and kill them. M. sexta contains three proteases that can activate prophenoloxidase, proPAP1, proPAP2, and proPAP3. The study of proPAP function has been slowed by the difficulty of expressing the proteins in recombinant systems. ProPAP1 contains one clip domain and one serine proteinase domain, a simpler structure than proPAP2 and proPAP3, which have two clip domains. For this reason, proPAP1 was selected for this investigation, to develop an improved system for expression of recombinant proPAP zymogens. In past experiments proPAP1 had a low expression level in insect cells using a baculovirus vector. In Escherichia coli, proPAP1 was expressed as an insoluble protein that could not be refolded successfully. The Bacillus subtilis expression system offers a potential improvement for expression of recombinant clip domain proteases because it can secrete recombinant proteins into the medium, it is a Biosafety Level 1 organism that is easy to handle, and it is less expensive to culture than insect cells. Four constructs for expression of proPAP1 and proPAP1 mutants were produced in the plasmid shuttle vector pHT43, which is compatible with both E. coli and B. subtilis. Experiments were carried out to test and optimize expression and purification of proPAP1 in B. subtilis. Conditions were optimized for IPTG (isopropyl β-D-1-thiogalactopyranoside) concentration, IPTG induction time, growth medium and induction temperature. Results showed that 0.5mM IPTG with 20 hours induction at 37°C in 2xYT medium was the optimum condition for proPAP1 production in the B. subtilis system. The recombinant proPAP1 was precipitated from the medium in 50% saturated ammonium sulfate and partially purified by nickel affinity chromatography. In addition to the full length proPAP1 protein, degradation of proPAP1 was also observed. Further experiments should be done to try to solve this problem. With purified protein, future work can be aimed at study of the structure and function of proPAP1.

Doctor of Philosophy / Graduate Biochemistry Group / Michal Zolkiewski / ClpB, a bacterial chaperone that belongs to the AAA+ protein family, cooperates with the Hsp70/40 system (DnaK, DnaJ and GrpE in E.coli) in the reactivation of aggregated substrates by translocating them through the central channel of its hexameric form. ClpB is essential for survival of bacteria under heat shock and plays an important role in the infectivity of pathogenic microorganisms. However the detailed mechanism of ClpB disaggregation activity is still not clear. ClpB is a multi-domain protein, which consists of two nucleotide binding domains (NBD1 and NBD2) connected by the middle domain (M domain), and the N-terminal domain connected to the rest of the protein by a flexible linker. In this work, mutations were introduced into the linker region to modify the mobility of the N-terminal domain. It was found that without altering the proper folding and oligomerization of ClpB, all the mutants had deficiencies in aggregate reactivation, possibly due to the weaker binding to aggregated substrates in the initial step of disaggregation. This led to the conclusion that the flexible attachment of the N-terminal domain supports substrate binding and controls the disaggregation by ClpB. Moreover, partial inhibition of the ClpB chaperone activity was observed for all the linker variants, suggesting that the linker sequence might have been optimized by selective pressure to maintain the optimal efficiency of aggregate reactivation. To study the substrate translocation of ClpB, a BAP (ClpB-ClpA P-loop) variant that binds to the protease ClpP was constructed. A FRET-based experiment was designed and the fluorescently-labeled ClpB substrates were produced. This work sets the stage for further studies on the mechanism of aggregate recognition by ClpB. ClpB also plays important roles in pathogenic bacteria invasion and virulence. Recombinant ClpB from Ehrlichia chaffeensis, a pathogenic bacterium that causes human monocytic ehrlichiosis, was purified to study its biochemical properties. Ehrlichia ClpB (Eh_B) and E.coli ClpB (Ec_B) sequences are highly conserved in the nucleotide binding region and poorly conserved in the N-terminal and M domain. The oligomerization, ATPase activity, chaperone activity and substrate binding of the recombinant Eh_B were tested. Recombinant Eh_B was able to reactivate aggregated proteins in the presence of HSP70 from E.coli with equal efficiency as Ec_B. However, the mechanism of Eh_B interactions with substrates and/or substrate specificity may be different from that of E. coli ClpB.

Master of Science / Department of Biochemistry / Jianhan Chen / Paul E. Smith / Intrinsically disordered proteins have been shown to be important in many physiological processes, including cell signaling, translation, and transcription. They are also associated with cancer, and neurodegenerative diseases. The tumor suppressor p53 contains several disordered regions, including the C-terminal negative regulatory domain (NRD). In cancer the function of p53 has been shown to be repressed by S100B(ββ) binding to p53-NRD. Binding of S100B(ββ) blocks acetylation and phosphorylation sites in the p53-NRD, which leads to tetramer dissociation and prevents p53 activation. NMR studies have shown that p53-NRD binds S100B(ββ) in a stable α-helix conformation. Interestingly, despite the well-converged and apparent rigid nature of the NMR structure ensemble, a majority of intermolecular NOEs used to calculate the NMR ensemble are very weak (≥6 Å). The final NMR structures also contains unsatisfied buried charged residues at the binding interface. It’s plausible that the p53-S100B(ββ) complex is more dynamic than previously believed. The goal of the study is to determine the potential conformational heterogeneity in p53-S100B(ββ) complex using molecular modeling. For this, five diverse structures were selected from the 40-member NMR ensemble. For each initial conformation, we performed 100 ns molecular dynamic simulations in explicit solvent to explore the structure and dynamics of the p53-NRD in complex with S100B(ββ). Several analytical tools were used to characterize the p53-NRD conformation, including root-mean squared deviation (RMSD), root-mean squared fluctuation (RMSF), and residue helicity. The accuracy of the simulations was mainly assessed by comparing with experimental NOEs. The results show that, even though the ensemble is heterogeneous it satisfies 82% of the experimental NOEs. Clustering analysis further suggests that many conformational sub-states coexist for this complex, and individual clusters appear to satisfy only subsets of NOE distances. Importantly, the buried surface analysis demonstrates that the heterogeneous ensemble generated from MD provides similar shielding of key residues, which include post-translational modification residues needed for p53 activation. This study also demonstrates that atomistic simulations can provide important insights into structure and dynamics of IDPs for understanding their biological function.

Doctor of Philosophy / Department of Biochemistry / John M. Tomich / Peptide-based delivery systems show great potential as safer drug delivery vehicles. They overcome problems associated with lipid-based or viral delivery systems, vis-a-vis stability, specificity, inflammation, antigenicity, and tune-ability. We have designed and synthesized a set of 15 and 23-residue branched, amphiphilic peptides that mimic phosphoglycerides in molecular architecture. They undergo supramolecular self-assembly and form solvent-filled, bilayer delineated spheres with 50-150 nm diameters (confirmed by TEM and DLS). Whereas weak hydrophobic forces drive and sustain lipid bilayer assemblies, these structures are further stabilized by β-sheet hydrogen bonding and are stable at very low concentrations and even in the presence of SDS, urea and trypsin as confirmed by circular dichroism spectroscopy. Given sufficient time, they fuse together to form larger assemblies and trap compounds of different sizes within the enclosed space. They are prepared using a protocol that is similar to preparing lipid vesicles. We have shown that different concentrations of the fluorescent dye, 5(6)-Carboxyfluorescein can be encapsulated in these assemblies and delivered into human lens epithelial cells and MCF-7 cells grown on coverslips. Besides fluorescent dyes, we have delivered the plasmid (EGFP-N3, 4.7kb) into N/N 1003A lens epithelial cells and observed expression of EGFP (in the presence and absence of a selection media). In the case of large molecules like DNA, these assemblies act as nanoparticles and offer some protection to DNA against certain nucleases. Linear peptides that lacked a branching point and other branched peptides with their sequences randomized did not show any of the lipid-like properties exhibited by the branched peptides. The peptides can be chemically decorated with target specific sequences for use as DDS for targeted delivery.

Doctor of Philosophy / Biochemistry / Michael R. Kanost / Insect innate immune responses include prophenoloxidase activation and antimicrobial peptide production. These responses involve extracellular serine proteinase cascades that are regulated by serpins. This work involved the study of serine proteinase 16 (HP16) and serpin-3 from hemolymph of the tobacco hornworm, Manduca sexta. HP16 has an amino-terminal domain with no similarity to any characterized protein and a carboxyl-terminal S1 family serine proteinase domain. HP16 levels in plasma were highest during the wandering, prepupal, and pupal stages. HP16 mRNA levels in fat body were highest at the wandering stage. Injection of bacteria into fifth instar larvae stimulated HP16 expression. To further characterize and investigate the biological function of HP16, recombinant proteins for proHP16, two HP16 mutants, the amino-terminal domain (NT16), and three NT16 mutants were purified. Recombinant HP16 was cleaved at the predicted activation site during expression, and its amino-terminal and catalytic domains remained connected by a disulfide bond. ProHP16 in plasma was apparently activated in the presence of the microbial elicitor, zymosan. Recombinant HP16 formed a complex with serpin-1Z, indicating that it was catalytically active, but no other natural or artificial substrates were identified. Analysis of NT16 and NT16 mutants led to the discovery that multiple disulfide bond arrangements were formed in the recombinant amino-terminal domain of HP16. This work furthered the understanding of HP16 and laid a foundation for subsequent experiments involving the proteolytic activity, regulation, and biological function of HP16. Active serine proteinases in insect hemolymph are often regulated by serpins. Immunoaffinity chromatography was used to identify plasma proteinases that are inhibited by serpin-3. Four serpin-3-proteinase complexes purified from plasma were identified by immunoblot analysis as serpin-3 complexes with HP8, PAP-1, PAP-2, and PAP-3. MALDI-TOF/TOF or ESI-MS/MS analysis after separation by 1D- or 2D-PAGE confirmed serpin-3 complex formation with HP8, PAP-1, and PAP-3. ProHP8 in plasma was activated by exposure to the β-1,3-glucan curdlan and inhibited by serpin-3. Purified recombinant serpin-3 and active HP8-Xa formed an SDS-stable complex in vitro. Identification of serpin-3-proteinase complexes in plasma provides insight into proteinase targets of serpin-3 and extends the understanding of serpin/proteinase function in the immune response of M. sexta.

Doctor of Philosophy / Department of Grain Science and Industry / Yong-Cheng Shi and Jeff D. Wilson / Sorghum (Sorghum bicolor [L.] Moench) is the fifth most important cereal grain grown in the world. Sorghum is an important cereal crop for both animal feed and biofuel production in the United States. The genetic, environmental, and agronomic management influences on sorghum starch and protein chemistry and functionality were evaluated. A method was developed to determine amylose content in cereal starches that achieved the same level of accuracy and precision as traditional methods, but had the capability of analyzing 50 samples per day or approximately a 10-fold increase in throughput. The effect of kernel maturity on sorghum starch properties was conducted by collecting grain from two hybrids at various stages throughout kernel development. The samples ranged from 16.3% amylose in 10 days after anthesis (DAA) to 23.3% amylose in 35 DAA. Starch thermal properties were also altered due to DAA, most notably the ΔH was 16.1 J/g at 14 DAA and 9.45 J/g at 56 DAA. In a separate study using the same developmental samples the protein and starch digestibility was analyzed. The kernel maturity had a notable effect on digestibility with the maximum digestibility occurring at 17 DAA with 82.44% digestible protein. In another study a diverse set of 19 sorghums was grown in three locations in Kansas to evaluate the genetic, location, and genetic x location effect on grain quality attributes. The physical and chemical properties of the sorghums were greatly affected by the genotype, environment, and the GxE interaction. Protein content ranged from 11.09% to 15.17% and digestibility ranged from 45.58% to 62.05% due to genotype. The final study investigates the role of agronomic management on sorghum grain quality. A sorghum hybrid was grown on plots with varying nitrogen fertilization rates and cover cropping systems that are currently used by Kansas producers. Grain attributes such as hardness and size were variable due to the treatments but negative impacts to protein digestibility were not seen due to cropping system. Sorghum grain quality is affected by many variables and a better understanding of the variables will lead to a higher quality product.