DEFINING THE ROLE OF EGFR ACETYLATION IN CELLULAR PROCESSES: CLINICAL IMPLICATIONS

Epidermal growth factor receptor (EGFR) is a cell membrane tyrosine kinase receptor and plays a pivotal role in regulating cell growth, differentiation, cell cycle, and tumorigenesis. Deregulation of EGFR causes many diseases including cancers. Intensive investigation of EGFR alteration in human cancers has led to profound progress in developing drugs to target EGFR-mediated cancers. While exploring possible synergistic enhancement of therapeutic efficacy by combining EGFR tyrosine kinase inhibitors (TKI) with other anti-cancer agents, we observed that suberoylanilide hydroxamic acid (SAHA, a deacetylase inhibitor) enhanced TKI-induced cancer cell death, which further led us to question whether SAHA-mediated sensitization to TKI was associated with EGFR acetylation. What we know so far is that SAHA can inhibit class I and II histone deacetylases (HDACs), which could possibly preserve acetylation of underlying HDAC-targeted proteins including both histone and non-histone proteins. In addition, it has been reported that an HDAC inhibitor, TSA, enhanced EGFR phosphorylation in ovarian cancer cells. EGFR acetylation has also been reported to play a role in the regulation of EGFR endocytosis recently. These observations indicate that there might be an intrinsic correlation between acetylation and phosphorylation of EGFR. In other words, the interplay between EGFR acetylation and phosphorylation may contribute to HDAC inhibitors (HDACi)-augmented EGFR phosphorylation. In this investigation, we showed that CBP acetyltransferase acetylated EGFR in vivo. In response to EGF stimulation, CBP rapidly translocated from the nucleus to the cytoplasm. We also demonstrated protein-protein interaction between CBP and EGFR as well as the enhancement of EGFR acetylation by CBP. Moreover, EGFR acetylation enhanced EGFR tyrosine phosphorylation and augmented its association with Src kinase. Acetylation-deficient EGFR mutant (EGFR-K3R) significantly reduced the function and activity of EGFR. Furthermore, ectopic expression of EGFR-K3R mutant abrogated its ability to respond to EGF-induced cell proliferation, DNA synthesis, and anchorage-independent growth using cell-based assays and tumor growth in nude mice. In addition, we demonstrated that EGFR expression was associated with SAHA resistance in the treatment of cancer cells that overexpress EGFR. The knockdown of EGFR in MDA-MB-468 breast cancer cells could sensitize the cells to respond to SAHA. The overexpression of EGFR in SAHA-sensitive MDA-MB-453 breast cancer cells rendered the cells resistant to SAHA. Together, these findings suggest that EGFR plays an important role in SAHA resistance in breast carcinoma cells that we tested. The combination therapy of HDACi with TKI has been proposed for treating cancers with aberrant expression of EGFR. The evidence from pre-clinical or clinical trials demonstrated significant enhancement of therapeutic efficacy by using such a combination therapy. Our in vivo study also demonstrated that the combination of SAHA and TKI for the treatment of breast cancer significantly reduced tumor burden compared with either SAHA or TKI alone. The significance of our study elucidated another possible underlying molecular mechanism by which HDACi mediated sensitization to TKI. Our results unveiled a critical role of EGFR acetylation that regulates EGFR tyrosine phosphorylation and may further provide an experiment-based rationale for combinatorial targeted therapy.

KEAP1 E3 ligase-mediated downregulation of NF-kappaB signaling by targeting IKKbeta.

IkappaB kinase beta (IKKbeta) is involved in tumor development and progression through activation of the nuclear factor (NF)-kappaB pathway. However, the molecular mechanism that regulates IKKbeta degradation remains largely unknown. Here, we show that a Cullin 3 (CUL3)-based ubiquitin ligase, Kelch-like ECH-associated protein 1 (KEAP1), is responsible for IKKbeta ubiquitination. Depletion of KEAP1 led to the accumulation and stabilization of IKKbeta and to upregulation of NF-kappaB-derived tumor angiogenic factors. A systematic analysis of the CUL3, KEAP1, and RBX1 genomic loci revealed a high percentage of genome loss and missense mutations in human cancers that failed to facilitate IKKbeta degradation. Our results suggest that the dysregulation of KEAP1-mediated IKKbeta ubiquitination may contribute to tumorigenesis.

Crystal structures of progressive Ca2+ binding states of the Ca2+ sensor Ca2+ binding domain 1 (CBD1) from the CALX Na+/Ca2+ exchanger reveal incremental conformational transitions.

Na(+)/Ca(2+) exchangers (NCX) constitute a major Ca(2+) export system that facilitates the re-establishment of cytosolic Ca(2+) levels in many tissues. Ca(2+) interactions at its Ca(2+) binding domains (CBD1 and CBD2) are essential for the allosteric regulation of Na(+)/Ca(2+) exchange activity. The structure of the Ca(2+)-bound form of CBD1, the primary Ca(2+) sensor from canine NCX1, but not the Ca(2+)-free form, has been reported, although the molecular mechanism of Ca(2+) regulation remains unclear. Here, we report crystal structures for three distinct Ca(2+) binding states of CBD1 from CALX, a Na(+)/Ca(2+) exchanger found in Drosophila sensory neurons. The fully Ca(2+)-bound CALX-CBD1 structure shows that four Ca(2+) atoms bind at identical Ca(2+) binding sites as those found in NCX1 and that the partial Ca(2+) occupancy and apoform structures exhibit progressive conformational transitions, indicating incremental regulation of CALX exchange by successive Ca(2+) binding at CBD1. The structures also predict that the primary Ca(2+) pair plays the main role in triggering functional conformational changes. Confirming this prediction, mutagenesis of Glu(455), which coordinates the primary Ca(2+) pair, produces dramatic reductions of the regulatory Ca(2+) affinity for exchange current, whereas mutagenesis of Glu(520), which coordinates the secondary Ca(2+) pair, has much smaller effects. Furthermore, our structures indicate that Ca(2+) binding only enhances the stability of the Ca(2+) binding site of CBD1 near the hinge region while the overall structure of CBD1 remains largely unaffected, implying that the Ca(2+) regulatory function of CBD1, and possibly that for the entire NCX family, is mediated through domain interactions between CBD1 and the adjacent CBD2 at this hinge.

The effect of v-{\it mos\/} expression on the regulation of the {\it fos\/} promoter in 490N3T cells

The v-mos oncogene acquired by Moloney murine sarcoma viruses by recombination with the c-mos proto-oncogene encodes a 37kD cytoplasmic serine/threonine protein kinase which can phosphorylate tubulin and vimentin, as well as the cyclin B component of the maturation promotion factor complex (MPF). Our earliest experiments asked whether the v-mos protein could activate the transcription of transin. Since the transcription of transin was known to be mediated by both fos-dependent and fos-independent pathways, it seemed possible that the induction of transin transcription by v-mos might be mediated by p55$\sp{\rm c-}\sp{fos}$. Surprisingly, when we examined the effect of v-mos on the fos promoter, we observed a significant inhibition of transcription in 49ON3T cells, a subclone of N1H3T3 mouse fibroblasts.^ In this thesis we show that in mouse 49ON3T cells, transcription from the fos promoter is up to 10-fold repressed in the presence of v-mos. Moreover, in this cell line several other transforming constructs (v-ras, v-src, neu) also cause repression of the fos promoter. Interestingly, nontransforming oncogenes (e.g. myc) do not repress fos transcription. The repressive effect was lost in v-mos mutants lacking in ATP-binding or kinase domain, arguing that the effect on fos transcription was mediated by v-mos transforming kinase activity. As mos is a cytoplasmic protein, it was assumed that transcriptional repression was mediated by conversion of a transcriptional regulator to a repressor by mos-induced phosphorylation. As a first approximation of the identity of this factor, we mapped the position of the mos effect on the fos promoter using reporter (CAT) constructs. We found that repression was mediated by regions $-$221 to $-$106 and $-$122 to $-$65 relative to the fos transcriptional start site, both of which regions regulate baseline fos transcription. There are direct repeats containing E2F transcriptional activator/repressor recognition motifs in these regions which bind similar nuclear proteins independently of v-mos presence or absence. Our data show that the contribution of the direct repeat to baseline fos transcription is mediated by these E2F sites with perhaps some contribution from the overlapping retinoblastoma control element (RCE). We have shown that there is a separate DNA protein interaction in the direct repeat which is more pronounced in the presence of v-mos. The recognition site for this protein, which we speculate mediates the mos-induced downregulation of fos transcription, overlaps but is distinct from the E2F and RCE binding sites. (Abstract shortened by UMI.) ^

Studies of subcellular localization and complex formation of the Agrobacterium tumefaciens transport ATPase VirB11

The VirB11 ATPase is an essential component of an Agrobacterium tumefaciens type IV bacterial secretion system that transfers oncogenic nucleoprotein complexes to susceptible plant cells. This dissertation investigates the subcellular localization and homo-oligomeric state of the VirB11 ATPase in order to provide insights about the assembly of the protein as a subunit of this membrane-associated transfer system. Subcellular fractionation studies and quantitative immunoblot analysis demonstrated that $\sim$30% of VirB11 partitioned as soluble protein and $\sim$70% was tightly associated with the bacterial cytoplasmic membrane. No differences were detected in VirB11 subcellular localization and membrane association in the presence or absence of other transport system components. Mutations in virB11 affecting protein function were mapped near the amino terminus, just upstream of a region encoding a Walker 'A' nucleotide-binding site, and within the Walker 'A' motif partitioned almost exclusively with the cytoplasmic membrane, suggesting that an activity associated with nucleotide binding could modulate the affinity of VirB11 for the cytoplasmic membrane. Merodiploid analysis of VirB11 mutant and truncation derivatives provided strong evidence that VirB11 functions as a homo- or heteromultimer and that the C-terminal half of VirB11 contains a protein interaction domain. A combination of biochemical and molecular genetic approaches suggested that VirB11 and the green fluorescence protein (GFP) formed a mixed multimer as demonstrated by immunoprecipitation experiments with anti-GFP antibodies. Second, a hybrid protein composed of VirB11 fused to the N-terminal DNA-binding domain of bacteriophage $\lambda$ cI repressor conferred immunity to $\lambda$ superinfection, demonstrating that VirB11 self-association promotes dimerization of the chimeric repressor. A conserved Walker 'A' motif, though required for VirB11 function in T-complex export, was not necessary for VirB11 self-association. Sequences in both the N- and the C-terminal halves of the protein were found to contribute to self-association of the full length protein. Chemical cross-linking experiments with His$\sb6$ tagged VirB11 suggested that VirB11 probably assembles into a higher order homo-oligomeric complex. ^

Molecular architecture of basement membranes

Basement membranes are specialized extracellular matrices with support, sieving, and cell regulatory functions. The molecular architectures of these matrices are created through specific binding interactions between unique glycoprotein and proteoglycan protomers. Type IV collagen chains, using NH2-terminal, COOH-terminal, and lateral association, form a covalently stabilized polygonal framework. Laminin, a four-armed glycoprotein, self-assembles through terminal-domain interactions to form a second polymer network, Entactin/nidogen, a dumbbell-shaped sulfated glycoprotein, binds laminin near its center and interacts with type IV collagen, bridging the two. A large heparan sulfate proteoglycan, important for charge-dependent molecular sieving, is firmly anchored in the basement membrane and can bind itself through a core-protein interaction to form dimers and oligomers and bind laminin and type IV collagen through its glycosaminoglycan chains. Heterogeneity of structure and function occur in different tissues, in development, and in response to different physiological needs. The molecular architecture of these matrices may be regulated during or after primary assembly through variations in compositions, isoform substitutions, and the modifying influence of exogenous macromolecules such as heparin and heparan sulfate.

Probing interactions of NMD factors in a distance‐dependent manner by BioID

The understanding of molecular mechanisms requires the elucidation of protein-­‐protein interaction in vivo. For large multi-­‐factor complexes like those assembling on mRNA, co-­‐immunoprecipitation assays often identify many peripheral interactors that complicate the interpretation of such results and that might conceal other insightful mechanistic connections. Here we address the protein-­‐protein interaction network for key factors in the nonsense-­‐mediated mRNA decay (NMD) pathway in a distant-­‐dependent manner using BioID1,2. In this novel approach, the mutant E. coli biotin-­‐protein ligase BirAR118G is fused to the bait protein and biotinylates proximal proteins promiscuously. Hence, interactors positioned close to the bait in vivo are enriched by streptavidin purification and identified by mass spectrometry or western blotting. We present a validation of the BioID assay and preliminary results for close interactors of UPF1 and other key players in NMD.

Ubiquitylation of voltage-gated sodium channels.

Ion channel proteins are regulated by different types of posttranslational modifications. The focus of this review is the regulation of voltage-gated sodium channels (Navs) upon their ubiquitylation. The amiloride-sensitive epithelial sodium channel (ENaC) was the first ion channel shown to be regulated upon ubiquitylation. This modification results from the binding of ubiquitin ligase from the Nedd4 family to a protein-protein interaction domain, known as the PY motif, in the ENaC subunits. Many of the Navs have similar PY motifs, which have been demonstrated to be targets of Nedd4-dependent ubiquitylation, tagging them for internalization from the cell surface. The role of Nedd4-dependent regulation of the Nav membrane density in physiology and disease remains poorly understood. Two recent studies have provided evidence that Nedd4-2 is downregulated in dorsal root ganglion (DRG) neurons in both rat and mouse models of nerve injury-induced neuropathic pain. Using two different mouse models, one with a specific knockout of Nedd4-2 in sensory neurons and another where Nedd4-2 was overexpressed with the use of viral vectors, it was demonstrated that the neuropathy-linked neuronal hyperexcitability was the result of Nav1.7 and Nav1.8 overexpression due to Nedd4-2 downregulation. These studies provided the first in vivo evidence of the role of Nedd4-2-dependent regulation of Nav channels in a disease state. This ubiquitylation pathway may be involved in the development of symptoms and diseases linked to Nav-dependent hyperexcitability, such as pain, cardiac arrhythmias, epilepsy, migraine, and myotonias.

Probing interactions of NMD factors in a distance-dependent manner by BioID

The understanding of molecular mechanisms requires the elucidation of protein-protein interaction in vivo. For large multi-factor complexes like those assembling on mRNA, co-immunoprecipitation assays often identify many peripheral interactors that complicate the interpretation of such results and that might conceal other insightful mechanistic connections. Here we address the protein-protein interaction network for key factors in the nonsense-mediated mRNA decay (NMD) pathway in a distant-dependent manner using BioID1,2. In this novel approach, the mutant E. coli biotin-protein ligase BirAR118G is fused to the bait protein and biotinylates proximal proteins promiscuously. Hence, interactors positioned close to the bait in vivo are enriched by streptavidin purification and identified by mass spectrometry or western blotting. We present a validation of the BioID assay and preliminary results for close interactors of UPF1 and other key players in NMD.

Probing interactions of NMD factors in a distance-dependent manner by BioID

The understanding of molecular mechanisms requires the elucidation of protein-protein interaction in vivo. For large multi-factor complexes like those assembling on mRNA, co-immunoprecipitation assays often identify many peripheral interactors that complicate the interpretation of such results and that might conceal other insightful mechanistic connections. Here we address the protein-protein interaction network for key factors in the nonsense-mediated mRNA decay (NMD) pathway in a distant-dependent manner using BioID1,2. In this novel approach, the mutant E. coli biotin-protein ligase BirAR118G is fused to the bait protein and biotinylates proximal proteins promiscuously. Hence, interactors positioned close to the bait in vivo are enriched by streptavidin purification and identified by mass spectrometry or western blotting. We present a validation of the BioID assay and preliminary results for close interactors of UPF1 and other key players in NMD.

Genetic association study of QT interval highlights role for calcium signaling pathways in myocardial repolarization

The QT interval, an electrocardiographic measure reflecting myocardial repolarization, is a heritable trait. QT prolongation is a risk factor for ventricular arrhythmias and sudden cardiac death (SCD) and could indicate the presence of the potentially lethal mendelian long-QT syndrome (LQTS). Using a genome-wide association and replication study in up to 100,000 individuals, we identified 35 common variant loci associated with QT interval that collectively explain ∼8-10% of QT-interval variation and highlight the importance of calcium regulation in myocardial repolarization. Rare variant analysis of 6 new QT interval-associated loci in 298 unrelated probands with LQTS identified coding variants not found in controls but of uncertain causality and therefore requiring validation. Several newly identified loci encode proteins that physically interact with other recognized repolarization proteins. Our integration of common variant association, expression and orthogonal protein-protein interaction screens provides new insights into cardiac electrophysiology and identifies new candidate genes for ventricular arrhythmias, LQTS and SCD.

A study in the molecular and cellular functions of GSK3beta in prostate adenocarcinoma

Kinases are part of a complex network of signaling pathways that enable a cell to respond to changes in environmental conditions in a regulated and coordinated way. For example, Glycogen Synthase Kinase 3 beta (GSK3β) modulates conformational changes, protein-protein interaction, protein degradation, and activation of unique domains in proteins that transduce signals from the extracellular milieu to the nucleus. ^ In this project, I investigated the expression and function that GSK3β exhibits in prostate cells. The capacity of GSK3β to regulate two transcription factors (JUN and CREB), which are known to be inversely utilized in prostate tumor cells, was measured. JUN/AP1 is constitutively activated in PC-3 cells; whereas, CREB/CRE activity is ∼20 fold less than the former. GSK3β overexpression obliterates JUN/AP1 activity. With respect to CREB GSK3β increases CREB/CRE activity. Cellular levels of active GSK3β can determine whether JUN or CREB is preferentially active in the PC-3s. Theoretically, in response to a particular cellular context or stimulus, a cell may coordinate JUN and CREB function by regulating GSK3β.^ A comparison of various prostate cell lines showed that active GSK3β is less expressed in normal prostate epithelial cells than in tumor cells. Differentially expressed active (GSK3β) may correlate with progression of prostate carcinoma. If a known marker associated with carcinoma of the prostate could be shown to be regulated by GSK3β then, further study of GSK3β may lead to a better understanding of both possible prevention of the disease and improved therapy for advanced stages. ^ The androgen receptor (AR) is an intriguing phosphoprotein whose regulation is potentially determined by a variety of kinases. One of these is (GSK3β) I found that (GSK3β) is a regulator of the androgen receptor in both the unliganded and liganded states. It can inhibit AR function as measured by reporter assays. Also, GSK3β associates with the AR at the DNA binding domain because deletion constructs expressing either the n-terminus or the c-terminus (both having the DBD in common) immunoprecipitated with GSK3β. Increased understanding of how GSK3β functions in prostate cancer would provide clues into how (1) certain signal pathways are coordinated and (2) the androgen receptor may be regulated. ^

Novel molecular insights into human tooth agenesis

The development of dentition is a fascinating process that involves a complex series of epithelial-mesenchymel signaling interactions. That such a precise process frequently goes awry is not surprising. Indeed, tooth agenesis is one of the most commonly inherited disorders in humans that affects up to twenty percent of the population and imposes significant functional, emotional and financial burdens on patients. Mutations in the paired box domain containing transcription factor PAX9 result in autosomal dominant tooth agenesis that primarily involves posterior dentition. Despite these advances, little is known about how PAX9 mediates key signaling actions in tooth development and how aberrations in PAX9 functions lead to tooth agenesis. As an initial step towards providing evidence for the pathogenic role of mutant PAX9 proteins, I performed a series of molecular genetic analyses aimed at resolving the structural and functional defects produced by a number of PAX9 mutations causing non-syndromic posterior tooth agenesis. It is likely that the pathogenic mechanism underlying tooth agenesis for the first two mutations studied (219InsG and IIe87Phe) is haploinsufficiency. For the six paired domain missense mutations studied, the lack of functional defects observed for three of the mutant proteins suggests that these mutations altered PAX9 function through alternate mechanisms. Next, I explored further the nature of the partnership between Pax9 and the Msx1 homeoprotein and their role in the expression of a downstream effector molecule, Bmp4. When viewed in the context of events occurring in dental mesenchyme, the results of these studies indicate that the Pax9-Msx1 protein interaction involves the localized up-regulation of Bmp4 activity that is mediated by synergistic interactions between the two transcription factors. Importantly, these assays corroborate in vivo data from mouse genetic studies and support reports of Pax9-dependent expression of Bmp4 in dental mesenchyme. Taken together, these results suggest that PAX9 mutations cause an early developmental defect due to an inability to maintain the inductive potential of dental mesenchyme through involvement in a pathway involving Msx1 and Bmp4. ^

Functional Characterization of AtxA, the Bacillus anthracis Virulence Regulator

Coordinated expression of virulence genes in Bacillus anthracis occurs via a multi-faceted signal transduction pathway that is dependent upon the AtxA protein. Intricate control of atxA gene transcription and AtxA protein function have become apparent from studies of AtxA-induced synthesis of the anthrax toxin proteins and the poly-D-glutamic acid capsule, two factors with important roles in B. anthracis pathogenesis. The amino-terminal region of the AtxA protein contains winged-helix (WH) and helix-turn-helix (HTH) motifs, structural features associated with DNA-binding. Using filter binding assays, I determined that AtxA interacted non-specifically at a low nanomolar affinity with a target promoter (Plef) and AtxA-independent promoters. AtxA also contains motifs associated with phosphoenolpyruvate: sugar phosphotransferase system (PTS) regulation. These PTS-regulated domains, PRD1 and PRD2, are within the central amino acid sequence. Specific histidines in the PRDs serve as sites of phosphorylation (H199 and H379). Phosphorylation of H199 increases AtxA activity; whereas, H379 phosphorylation decreases AtxA function. For my dissertation, I hypothesized that AtxA binds target promoters to activate transcription and that DNA-binding activity is regulated via structural changes within the PRDs and a carboxy-terminal EIIB-like motif that are induced by phosphorylation and ligand binding. I determined that AtxA has one large protease-inaccessible domain containing the PRDs and the carboxy-terminal end of the protein. These results suggest that AtxA has a domain that is distinct from the putative DNA-binding region of the protein. My data indicate that AtxA activity is associated with AtxA multimerization. Oligomeric AtxA was detected when co-affinity purification, non-denaturing gel electrophoresis, and bis(maleimido)hexane (BMH) cross-linking techniques were employed. I exploited the specificity of BMH for cysteine residues to show that AtxA was cross-linked at C402, implicating the carboxy-terminal EIIB-like region in protein-protein interactions. In addition, higher amounts of the cross-linked dimeric form of AtxA were observed when cells were cultured in conditions that promote toxin gene expression. Based on the results, I propose that AtxA multimerization requires the EIIB-like motif and multimerization of AtxA positively impacts function. I investigated the role of the PTS in the function of AtxA and the impact of phosphomimetic residues on AtxA multimerization. B. anthracis Enzyme I (EI) and HPr did not facilitate phosphorylation of AtxA in vitro. Moreover, markerless deletion of ptsHI in B. anthracis did not perturb AtxA function. Taken together, these results suggest that proteins other than the PTS phosphorylate AtxA. Point mutations mimicking phosphohistidine (H to D) and non-phosphorylated histidine (H to A) were tested for an impact on AtxA activity and multimerization. AtxA H199D, AtxA H199A, and AtxA H379A displayed multimerization phenotypes similar to that of the native protein, whereas AtxA H379D was not susceptible to BMH cross-linking or co-affinity purification with AtxA-His. These data suggest that phosphorylation of H379 may decrease AtxA activity by preventing AtxA multimerization. Overall, my data support the following model of AtxA function. AtxA binds to target gene promoters in an oligomeric state. AtxA activity is increased in response to the host-related signal bicarbonate/CO2 because this signal enhances AtxA multimerization. In contrast, AtxA activity is decreased by phosphorylation at H379 because multimerization is inhibited. Future studies will address the interplay between bicarbonate/CO2 signaling and phosphorylation on AtxA function.

Network-based analysis of genome-wide scans of natural selection

The genomic era brought by recent advances in the next-generation sequencing technology makes the genome-wide scans of natural selection a reality. Currently, almost all the statistical tests and analytical methods for identifying genes under selection was performed on the individual gene basis. Although these methods have the power of identifying gene subject to strong selection, they have limited power in discovering genes targeted by moderate or weak selection forces, which are crucial for understanding the molecular mechanisms of complex phenotypes and diseases. Recent availability and rapid completeness of many gene network and protein-protein interaction databases accompanying the genomic era open the avenues of exploring the possibility of enhancing the power of discovering genes under natural selection. The aim of the thesis is to explore and develop normal mixture model based methods for leveraging gene network information to enhance the power of natural selection target gene discovery. The results show that the developed statistical method, which combines the posterior log odds of the standard normal mixture model and the Guilt-By-Association score of the gene network in a naïve Bayes framework, has the power to discover moderate/weak selection gene which bridges the genes under strong selection and it helps our understanding the biology under complex diseases and related natural selection phenotypes.^