V-ATPase primarily acidifies organelles of the endocytic and secretory pathway. The lumen of each of these organelles maintains a distinct pH, but how this diverse pH range is established is not well understood. Different modes of regulation of the V ATPase may help fine-tune their activity. The a-subunit from the membrane bound Vo subcomplex is a regulatory hub of V-ATPase which harbors many regulatory interactions in its cytosolic N-terminal (aNT) domain. The aNT domains of all isoforms consist of a proximal and a distal globular domain connected by a coiled-coil. Vo a-subunit exhibits organelle- and tissue-specific isoforms. Vph1 and Stv1 are the two organelle-specific a subunit isoforms in yeast. V-ATPase containing the Vph1 isoform reside in the lysosome-like yeast vacuole whereas V-ATPase containing the Stv1 isoform reside predominantly in the Golgi apparatus. The RAVE (Regulator of H+ -ATPase of vacuoles and endosomes) complex and phosphoinositide phospholipids (PIP lipids) are two important factors previously implicated in isoform-specific V-ATPase regulation. The a subunit isoforms vary in their dependence on the RAVE and PIP lipid. But where the information of RAVE and PIP lipid recognition reside in the aNT domain and how these two regulatory inputs are integrated to control V-ATPase function are not well understood. We hypothesize that the aNT domain contains distinct sequences for RAVE and PIP lipid recognition, and that the differential interactions of the a-subunit isoforms with RAVE and PIP lipids result in isoform-specific function and regulation of the V ATPase. To better understand the regulatory information present in the aNT domain, we generated chimeric aNT constructs by combining parts of Vph1NT and Stv1NT. Vph1NT was previously shown to bind to a RAVE subunit and Vph1 V-ATPases require the RAVE complex for their assembly. Stv1-containing V-ATPases on the other hand assemble independent of RAVE and Stv1NT does not bind to RAVE subunits in vitro. In chapter 2, we have shown that replacing the proximal domain of Vph1NT with the proximal end of Stv1NT fully restores the RAVE interaction, implicating this region of Vph1 in RAVE-dependent assembly. The two isoforms also exhibit preferences for distinct PIP lipids enriched in their organelle of residence; Stv1NT binds tightly to Golgi PI(4)P and Vph1NT binds to vacuolar PI(3,5)P2. A PI4P binding site in the proximal domain of Stv1NT was previously reported. Our analysis with chimeric constructs suggests that a 6-amino acid sequence containing this site is sufficient to transfer PI(4)P binding to Vph1NT. Our results also suggest that both the proximal and the distal ends of Stv1NT contain sequences that promote PI4P binding. Interestingly, when expressed in yeast as a full-length a-subunit, the chimera containing both PI4P and PI(3,5)P2 binding sites has wild-type level activity and assembly in isolated vacuoles even though it lacks a RAVE binding site. Although V-ATPases in this chimeric strain is fully functional there are consequences of their altered regulatory properties. V-ATPases in this chimera disassemble during glucose deprivation but do not reassemble efficiently after glucose re addition, consistent with a lack of RAVE binding. We also observed a delayed growth in media containing raffinose suggesting they cannot readily adjust during a transition to a less preferred carbon source. While the lack of RAVE binding gives rise to the growth defect of this chimera in a poor carbon source, increased PI(3,5)P2 binding of this mutant, on the other hand, contributes to growth benefit in alkaline pH stress. Together these data reveal the interplay between two mechanisms of V-ATPase regulation and suggest that aNT domains can functionally integrate multiple regulatory inputs.

Human cytomegalovirus (HCMV) is a highly prevent pathogen with seropositivity ranging from 40-90% in North America, and in excess of 100% in developing nations. Although HCMV infections are often mild to asymptomatic among immunocompetent hosts, HCMV represents a significant cause of morbidity and mortality for those with compromised immune systems. HCMV's systemic dissemination within the host is facilitated by infected blood monocytes. Within monocytes, HCMV establishes a quiescent infection, thus acting as "Trojan horses", delivering the virus to distant end organs all while remaining hidden from innate immune detection. Blood monocytes are inherently short-lived however, with an average lifespan of only 48 h. To circumvent this shortcoming HCMV aberrantly regulates the Akt/mTORC1 signaling axis. Our lab has previously shown that HCMV binding and entering into target monocytes elicits a unique action of Akt, characterized by the preferential phosphorylation at Serine 473 (S473). Critically, downstream substrate specificity of Akt signaling is dictated by the ratio of S473 to threonine 308 phosphorylation, suggesting this unique Akt activation would have biologic significance during infection. The studies of this thesis reveal that a major consequence of HCMV-mediated aberrantly regulated Akt is a robust increase in protein translation through the activation of mTORC1 and its downstream signaling substrates. Moreover, HCMV uniquely usurps the activity of heat shock factor 1 (HSF1) to bypass cellular stress response traditionally intended to limit protein synthesis through the direct interaction between HSF1 and mTOR. The tightly regulated signaling events downstream of Akt and mTORC1 lead to a reshaping of the host translatome to redirect protein synthesis towards antiapoptotic factors necessary for the induction of monocyte survival beyond their canonical 48 h lifespan. Elucidating the molecular mechanisms behind HCMV's control over protein translation may allow for the development of novel therapies that specifically target HCMV-infected monocytes, thus preventing viral dissemination and the establishment of disease.

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease linked to >sixty genes. However, the mechanistic details of onset are unknown and there is no known cure or effective treatment. Disease mechanisms linked to the dysfunction of the neuronal cytoskeleton are arguably the least explored, despite being involved in important cell processes. Eight cytoskeleton associated proteins are genetically linked to ALS onset, including tubulin-4A, spastin, kinesin-5A, dynactin subunit-1, neurofilament, peripherin, alsin, and profilin-1 (PFN1). Most of these genes are linked to microtubule dynamics, yet only PFN1 directly regulates actin assembly. Profilin is essential for many neuronal cell processes mediating the dynamics of lipids, nuclear signals, and the cytoskeleton in neurons. We performed a quantitative biochemical comparison of the impact of the eight ALS-associated profilin variants on actin assembly using single-molecule microscopy assays. The binding affinity of each ALS-related profilin for actin monomers was loosely correlated with the actin filament nucleation strength. The A20T, R136W, Q139L, and C71G variants failed to activate formin-based actin assembly, mostly explained by deficiencies in binding to poly-L-proline stretches in formin. In addition, chemical denaturation experiments suggest that the folding stability of some ALS profilin proteins impacts on actin assembly. Finally, to better understand the connection between the cytoskeleton and ALS we investigated the role of profilin and its direct binding ligand TDP-43 (TAR DNA-binding protein 43) on actin assembly. TDP-43 misregulation is a hallmark of nearly all forms of neurodegeneration, including an estimated 40% of familial ALS. Using advanced microscopy assays, we visualized purified TDP-43 forming biomolecular condensates. Actin was sequestered within TDP-43 condensates tempering total filament assembly. These observations were further exacerbated in the presence of profilin. These results indicate disruptions to actin assembly contribute to ALS and suggest therapeutic interventions targeting actin assembly may be a useful in treating ALS.

The nuclear hormone receptor and transcription factor, the androgen receptor (AR), is the main driver of prostate cancer growth and disease progression. Hallmark target genes of AR are used as biomarkers to indicate disease progression and treatment response. Furthermore, current clinical treatments for prostate cancer include androgen deprivation treatment and anti-AR which deplete ligand availability or directly target the androgen receptor, respectively. Transcription regulates key functions of living organisms in normal and diseases states, including cell growth and development, embryonic and adult tissue organization, and tumor progression. Here we identify a novel mechanism of transcriptional regulation by an actin regulatory and signaling protein, ABI1. As established by ChIP sequencing and DNA binding assays, ABI1 binds to chromatin through its intrinsically disordered DNA binding domain. Furthermore, ABI1 interacts with AR in vitro and in vivo and targets its activity to specific subset of genes. ABI1-AR driven transcription is dysregulated during prostate tumor progression. Additionally, anti-androgen and anti-AR treatments induce alterations in AR-mediated transcription which leads to downregulation of ABI1 expression and induces disruption of epithelial integrity. The results from this study indicate that ABI1 controls tumor plasticity through connecting actin cytoskeleton and cellular signaling to transcriptional regulation. We propose that ABI1 is a regulator of transcriptional homeostasis in prostate cancer.

Parkinson's disease (PD) is a progressive debilitating neurodegenerative disorder affecting 2% of subjects over age 65. The major motor symptoms of PD result from loss of midbrain dopamine-synthesizing neurons and include resting tremor, rigidity, slowed movements (bradykinesia), and postural instability. Although there are known genetic causes, most PD cases have undetermined etiology. Thus, there is considerable interest in identifying biomarkers of early-stage PD to develop effective intervention strategies that might modify the disease course. Studies completed in this dissertation were designed to identify and validate such biomarkers through comprehensive analysis of the oral microtranscriptome in human subjects, as well as in animal and cellular PD models. A total of 300 human subjects were recruited, including 178 with PD and 122 controls. Both groups contained subjects with comorbid or isolated conditions frequently associated with PD, including restless legs syndrome, dystonia, and essential tremor. Subjects completed questionnaires to assess risk factors, medication use, and motor and non-motor symptoms, and underwent formal neurological examination. Approximately half of the subjects also completed olfactory testing and computerized neuromotor, cognitive, and postural testing. Saliva samples were collected using a vial or swab. Levels of microRNA and microbiota were quantified using next generation sequencing. Saliva miRNA levels were also compared in mice expressing wildtype (WT) or A53T mutant copies of human alpha synuclein (SNCA), a known cause of PD. Performance of the mice was also assessed in motor and cognitive tasks during pre- and early-symptomatic stages of the model. Lastly, measurements of both extracellular miRNA and cellular mRNA, as well as oxidative stress and DNA damage, were completed in human iPSC-derived dopaminergic neurons expressing WT or A53T SNCA, combined with exposure to media or Paraquat, an environmental risk factor for PD. All PD subjects fell within the mild-moderate early stage category. Significant differences were observed for olfactory, postural, and more challenging cognitive tests. Profiling of salivary miRNAs revealed subsets highly-changed in early stage PD. The mouse and dopaminergic neuron data strongly supported the findings for two specific miRNAs - miR-103a-3p and miR-107 - and indicate a number of key cellular pathways are altered across the disease and models.

Scarring in the cornea obstructs the refraction of incoming light onto the retina causing visual disability. Both acute scarring and chronic fibrosis are characterized by an accumulation of disorganized extracellular matrix (ECM). Disorganized ECM is deposited into the wound by specialized cells termed myofibroblasts. Pathological myofibroblasts are characterized by the expression of the highly contractile alpha smooth muscle actin (a-SMA) and the av-family of integrins (avb1,b3, b5, b6). The persistence of myofibroblasts in a healing wound promotes an autocrine loop of TGFb activity, over contraction of tissue, deposition of fibrotic ECM proteins, and ultimately the generation of scar tissue. My work is focused on the relative contribute of the deubiquitinase, USP10 to scarring in the cornea. I found that after wounding an increase in the expression of USP10 leads to deubiquitination of integrins and a subsequent increase in integrin recycling and matrix deposition. Knockdown of USP10 in vivo after corneal wounding significantly reduced the presence of myofibroblasts and immune cells in the healing wound, and corneal scarring. Through a yeast 2-hybrid screen I also identified a novel USP10 interacting protein, the formin Daam1. I found that Daam1 sequesters USP10 on actin stress fibers inhibiting its activity. Under pathological conditions, the expression of both USP10 and Daam1 are increased. My data suggest that Daam1 acts as a cellular reservoir, adding a layer of homeostatic control over USP10 activity and integrin function. Although defects in protein degradation have been identified as a major contributor to many diseases, together, my studies indicate that protein degradation (ubiquitin) pathways need to be considered in the context of integrin biology and in the pathogenesis of fibrotic healing.

Actin and actin cytoskeleton associated proteins are important for maintenance of cellular homeostasis and cellular processes such as cell polarization, cell movement, and endocytosis. The long-tailed class one myosin, Myosin-1e (Myo1e), is an actin dependent molecular motor that is found to be expressed in and contribute to the function of epithelial cells. Specific mutations of Myo1e are associated with dysfunction of specialized kidney epithelial cells known as podocytes and increased incidence of kidney disease. We found that disease associated mutation of Myo1e affects the conformation of key functional domains within the Myo1e protein structure using molecular dynamic simulations (Chapter 2). This finding leads to a more mechanistic understanding of dysfunctional Myo1e protein conformational pathways that may be associated with kidney disease. We also find that Myosin-1e expression influences the behavior of mammary tumors in mice (Chapter 3). Specifically, mammary epithelial cell oncogenic transformation induced using the MMTV-PyMT oncogene (mouse mammary tumor virus, polyoma middle T-antigen) was phenotypically altered in mice lacking Myo1e expression compared to Myo1e expressing mice. The tumor cells in mice lacking Myo1e were more differentiated and expressed transcriptomic profiles associated with tumor suppression compared to the more oncogenic profiles of Myo1e expressing cells. In-vitro analysis revealed a cell autonomous effect of Myo1e expression on cell-cell junctional strength and gene expression associated with differentiation (Chapter 3). Appendix chapters 1 and 2 discuss experiments examining Myo1e expression on in-vitro cell migration and epithelial to mesenchymal transition (EMT). Overall, experiments discussed in this thesis have uncovered insight into the effects of Myo1e expression on breast cancer progression, cell differentiation and motility. Moreover, experiments in this thesis also contribute to further understanding of the dynamics of protein conformation associated with disease mutations of a class one myosin.

Dysfunction of the conventional outflow pathway (comprised of the trabecular meshwork (TM) and adjacent Schlemm's canal (SC)) is the principal cause of elevated intraocular pressure in primary open-angle glaucoma. Other in vitro TM model systems cannot accurately mimic the cell-extracellular matrix (ECM) interface, limiting their use for investigating glaucoma pathology. In this dissertation, we report a novel biomimetic hydrogel by mixing donor-derived human TM (HTM) cells with ECM proteins found in the native tissue. We demonstrated that this HTM hydrogel system allowed for investigation of actin arrangement, ECM remodeling, cell contractility, and HTM stiffness on a simulated tissue level (Chapter 2). Furthermore, we showed that TGFβ2-induced ERK signaling negatively regulates Rho-associated kinase-mediated phospho-myosin light chain expression and HTM cell contractility when cultured on soft ECM hydrogels but not on glass (Chapter 3). YAP and TAZ are important mechanotransducers implicated in glaucoma pathogenesis. We demonstrated that YAP/TAZ activity was upregulated by transforming growth factor beta 2 (TGFβ2) in both HTM and HSC cells cultured on/in ECM hydrogels (Chapters 4 and 5). It is widely accepted that the glaucomatous TM/SC interface is stiffer. To mimic the stiffness difference between diseased and healthy tissue, we utilized two different methods. In Chapter 4, riboflavin was used to facilitate secondary UV crosslinking of collagen fibrils and stiffen the matrix. We showed that ECM stiffening elevated YAP/TAZ activity in HTM cells through modulating focal adhesions and cytoskeletal rearrangement. In Chapter 6, we developed an ECM-alginate hybrid hydrogel system, which allowed for on-demand control over matrix stiffness during the culture of cells. We found that the stiffened matrix increased nuclear YAP and filamentous-actin fibers in HSC cells, which was completely reversed by matrix softening. We further demonstrated that YAP/TAZ inhibition could rescue HTM/HSC cell dysfunction induced by either TGFβ2 or stiff matrix (Chapters 4, 5, and 6). Finally, we showed that pharmacologic YAP/TAZ inhibition had promising potential to improve outflow facility in an ex vivo mouse eye perfusion model (Chapter 6). Collectively, we have developed bioengineered ECM hydrogels for modeling and investigation of conventional outflow cell-ECM interactions under normal and simulated glaucomatous conditions.

Engineered protein conformational switches have applications in cellular and in vitro biosensing, molecular diagnostics and artificial signaling systems in synthetic biology. They broadly consist of an input module and an output module that communicate via a conformational change. The overarching goal of this thesis is to tackle two major challenges in protein switch design - signal transduction, by coupling a target recognition domain to an output domain to produce a robust change in signal in addition to modularity, which allows the facile creation of sensors binding novel targets. Here, we attempted to test a rational design strategy that exploits two key protein engineering principles (1) loop entropy, by which long insertions into a loop of a host protein destabilizes the host due to an entropic cost associated with loop closure unless the inserted sequence adopts a folded structure; and (2) alternate frame folding (AFF), which allows a protein - green fluorescent protein variants(GFP), in this case - to switch between two mutually exclusive folds. Toward this goal, we first studied the effect of loop entropy at two different insertion sites in a GFP variant (chapter 2) using a well-characterized ribose binding protein as the input domain. We provide stability measurements using circular dichroism and fluorescence data to support our hypothesis of the application of the loop entropy principle in a GFP beta barrel scaffold. To provide a proof-of-concept of the combination of loop entropy and the AFF mechanism in a genetically encodable GFP scaffold, we chose an unstable, circularly permuted FK506-binding protein (cpFKBP) as the input recognition domain and inserted it in one of the two mutually exclusive folds of the GFP-AFF fusion protein (chapter 3). Upon addition of ligand, binding induced folding of the cpFKBP domain effects a conformational change in which the tenth beta strand of GFP exchanges, replacing Thr203 (green state) with Tyr203 (yellow state). We confirmed this mechanism in vitro by a ratiometric change in fluorescence output and observed that the process is slow and irreversible. We elucidate the biophysical principles underlying this mechanism by using denaturant and temperature to modulate the relative populations of the two folds in vitro. We also observed a faster and higher intensiometric response in mammalian cells which may be attributed to an alternate mechanism. We then harnessed this intensiometric response in a single fold of the fluorescent protein combined with a previously engineered monobody scaffold capable of binding a variety of targets (chapter 4). Altogether this work may have the potential to create a novel class of fluorescent protein biosensors comparable to existing single fluorescent protein-based biosensors currently available.