Antibodies and antibody fragments are the largest class of biotherapeutics in development with many products already available in the clinic. Antibodies are promising due to their naturally high affinity and specificity for biological targets. A key stumbling block to biopharmaceutical development compared to small molecule drugs is the general requirement for a stable liquid formulation, which is often difficult to obtain due to issues with aggregation, phase separation, particle formation, and chemical instabilities. Aberrant solution behaviour limits the production, storage and delivery of the monoclonal antibody. Biopharmaceutical solution behaviour is determined by weak, transient protein-protein and protein-solvent interactions. An attractive interaction potential between proteins in solution can lead to association. Irreversible association occurs when proteins undergo large scale structural changes and aggregate. Reversible association is less severe, but can lead to undesirable solution properties such as high viscosity, phase separation and opalescence, which can lead to difficulties throughout the downstream processing and formulation steps. These problems can become exacerbated during formulation of antibodies when trying to achieve high protein concentrations often required for effective antibody dosage. Firstly, we studied the domains of the Fab fragment using statistical models and continuum electrostatic calculations and found that the CH1 domain is more soluble than the other domains and has properties of intrinsically disordered like proteins which is supported by observations in the literature. We then investigated the immunoglobulin superfamily and found 11 proteins which may have a similarly disordered nature. We present a new web server for predicting protein solubility from primary sequence using an in-house algorithm that weighs the contribution of various sequence properties for predicting solubility. Lastly, we conducted physical characterisation of an antibody and human serum albumin in pharmaceutically relevant buffers and found that the interaction potential can be modelled using spherical models from low to high protein concentration. We hope that the work outlined in this thesis will contribute to the theoretical understanding and modelling of protein solution behaviour.

Despite therapeutic advances, colorectal cancer still has a 45% mortality rate, and one of the most crucial problems is the development of acquired resistance to treatment with anticancer drugs. Thus the aims of this project are to develop drug-resistant colon cancer cell lines in order to identify mechanisms of resistance for the most commonly drugs used in colorectal cancer: 5-fluorouracil, oxaliplatin, and irinotecan. Following evaluation of drug sensitivity to these agents in an initial panel of eight colorectal cancer cell lines, 3 lines (DLD-1, KM-12 and HT-29) were selected for the development of 5-FU (3 lines), oxaliplatin (2) and irinotecan (1) resistant sublines by continuous drug exposure, with resistance confirmed using the MTT assay. Consistently resistant sublines were subject to a „stable isotope labelling with amino acids in cell culture‟ (SILAC) approach and a MudPIT proteomics strategy, employing 2D LC and Orbitrap Fusion mass spectrometric analysis, to identify novel predictive biomarkers for resistance. An average of 3622 proteins was quantified for each resistant and parent cell line pair, with on average 60-70 proteins up-regulated and 60-70 down-regulated in the drug resistant sublines. The validity of this approach was further confirmed using immunodetection techniques. These studies have provided candidate proteins which can be assessed for their value as predictive biomarkers, or as therapeutic targets for the modulation of acquired drug resistance in colorectal cancer.

Hepatoma-derived growth factor (HDGF) is commonly over-expressed in human cancer cells. It was able to stimulate cell growth. The expression level of HDGF was reported to correlate with poor prognosis of cancer therapy. It was found that HDGF is over-expressed in the fractionated gamma radiation conditioned HepG2 cells, which have higher growth rate, lower radiosensitivity and higher drug sensitivity. The aim of the present study was to investigate the role of HDGF in mediating these changes in human cancer cells and the underlying mechanisms. The results indicate that transfection of HDGF cDNA carrying vector stimulated the growth of cancer cells while knock-down of HDGF by transfection of HDGF antisense oligos not only suppressed the growth but also triggered apoptosis in human cancer cells. It suggests that HDGF stimulates cancer cell growth and acts as a survival factor for human cancer cells. Mechanistic study showed that knock-down of HDGF may trigger apoptosis through the regulation of the apoptotic pathways. The apoptosis induced by HDGF knock-down was mediated by the BAD regulated intrinsic apoptotic pathway and the Fas regulated extrinsic apoptotic pathway. The HDGF knock-down induced apoptosis was also mediated by the changes in the activity of the cell survival pathways, including the Ras/Raf/MEK/ERK, PI3K/Akt, NFkappaB and Jak/STAT pathways. In addition to the growth promoting function, HDGF was found to regulate the radiosensitivity and chemosensitivity of cancer cells. Overexpression of HDGF reduced the radiosensitivity and the level of apoptosis induced by gamma radiation. On the contrast, overexpression of HDGF increased the chemosensitivity and the level of apoptosis induced by anti-cancer drugs, including Taxol, doxorubicin (Dox) and tamoxifen. The results indicated that HDGF may stimulate the growth, reduce the radiation sensitivity and increase the drug sensitivity of cancer cells. HDGF may also be responsible for the changes in cancer cell properties after fractionated gamma radiation treatment. The present findings suggest that HDGF may be a potential target for cancer therapy. / Tsang, Tsun Yee. / Adviser: Tim Tak Kwok. / Source: Dissertation Abstracts International, Volume: 70-06, Section: B, page: 3497. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2008. / Includes bibliographical references. / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstracts in English and Chinese. / School code: 1307.

This thesis describes structural studies on the interactions between the fucosylated chondroitin sulfate (fCS) oligosaccharides and human proteins known as selectins. fCS is a carbohydrate obtained from sea cucumbers, that can be classified as a branched glycosaminoglycan (GAG). It has attracted much attention due to its anti-coagulant, anti-inflammatory, antimetastatic and anti-HIV properties and its structure was previously determined by NMR. Selectins constitute a family of proteins involved in cell adhesion processes, such as inflammation, attachment of viral particles and migration of tumour cells. fCS oligosaccharides have been shown to bind to selectins, which is likely a reason behind their biological activity. However, the mechanism of this interaction is currently unknown. The initial part of the thesis describes the experimental work on expression and purification of the recombinant L- and P-selectin constructs in Pichia pastoris, Escherichia coli and HEK 293 cells. The aim of these experiments was to produce two constructs for each selectin, a single domain construct, consisting of the C-type lectin domain only, and a double domain construct, consisting of both the C-type lectin and the EGF-like domains. The intention was that the recombinant proteins would be labelled with 13C and 15N to allow for the in-depth structural NMR studies on the fCS-selectin interaction. Various experimental approaches have been explored, including the use of different cell lines, modifications to construct design, as well as alterations to expression and purification conditions. Although it was not possible to produce soluble selectin constructs in either bacterial or yeast cells, protein expression tests in HEK293 cells, performed in collaboration with the Oxford Protein Production facility (OPPF), led to production of a soluble L-selectin construct, consisting of the L-selectin C-type lectin domain. The produced L-selectin construct, as well as two commercially available constructs of the Land P-selectin extracellular domains, were used in the Saturation Transfer Difference (STD) NMR experiments to provide new information about the nature of the fCS-selectin binding. The STD experiments allowed to identify the regions within the fCS oligosaccharides that are in direct contact with the protein and likely play an important role in this interaction. Experiments on different protein constructs allowed the comparison of fCS binding to P-selectin and to two different recombinant constructs of L-selectin. Results of these studies suggest that the binding occurs via a similar mechanism for both L- and P-selectins and that the fCS oligosaccharides bind to one-domain L-selectin construct with similar affinity as to a larger construct, consisting of the entire extracellular region of the protein. Alongside the experimental work, theoretical in silico studies on the fCS-selectin binding were undertaken as part of this project. The existing X-ray structures of selectin complexes were subjected to Molecular Dynamics (MD) simulations, which allowed to explore the dynamic behaviour of E-selectin upon binding to sialyl Lewis x (sLex). It was found that sLex forms a more favourable interaction with the extended conformation of E-selectin and that the protein in this conformation is characterised by a high degree of interdomain flexibility, with a new type of interdomain movement observed in the MD studies on this complex. In further in silico studies, the fCS oligosaccharides were docked to the existing P-selectin structures. The docking tests were performed on the computationally produced fCS trisaccharides with fucose branches either 2,4 or 3,4-sulfated. Results were evaluated with MD simulations and analysed in the light of current knowledge of selectin-ligand binding and the STD NMR experimental results. The in silico studies allowed to identify a subset of P-selectin residues that are likely involved in the interaction with fCS oligosaccharides in vivo. The conformational behaviour of P-selectin upon binding to fCS was also explored and it was found that the interdomain hinge is flexible during this interaction and allows transition from bent to extended conformational state. Finally, a new NMR method was developed to facilitate the studies of complex carbohydrates, incorporating the concepts of G-matrix Fourier Transform (GFT) NMR into 2D HSQC and 2D HSQC-TOCSY experiments. The method allows to separate peaks in the regions of high spectral overlap, providing information that can simplify the assignment process. The new experiments facilitated the structural evaluation of a sample containing a mixture of oligosaccharides resulting from the depolymerisation of fCS polysaccharide.

STRAP (Stress responsive activator of p300) is a 440 amino acid protein, predicted to have 6 TPR (Tetra-Tri-Co-Peptide Repeats) motifs, known to mediate protein-protein interactions. STRAP has been shown to form a complex with proteins p300 and JMY (Junctional Mediatory Protein), and is implicated in the DNA damage, heat shock response pathway, regulation of the Glucocorticoid receptor and in the function of p53.The aims of this project were to clone, express and purify full length and truncated human STRAP (hSTRAP) variants in high quantities. Full length and shorter hSTRAP fragments, which contain different combinations of the predicted TPR motifs and hence cover different regions, would be then structurally characterised by various structural and biophysical experiments. Another important aim was to identify interacting partners of hSTRAP in breast cancer and to map the position of their interaction sites to different parts of the protein. To this direction GST- and His- tagged full length hSTRAP, as well as His- tagged truncated hSTRAP protein variants have been successfully cloned, expressed and purified. Independent and reproducible biochemical pull-down assays have been carried out in MCF7 breast cancer cells, followed by mass spectrometry-based proteomics analysis which identified 25 hSTRAP-interacting partners from various signaling pathways such as regulation of the actin cytoskeleton and translation. In addition, crystallization trials were carried out using pure His-hSTRAP(1-440) protein, which were unfortunately un-successful. Various hSTRAP protein variants have been characterized by CD, showing that hSTRAP(1-150), His-hSTRAP(1-440), hSTRAP(1-219), hSTRAP(151-284) and hSTRAP(285-440) comprise of alpha and β structures, but the hSTRAP protein variants show no clear cooperative unfolding transitions, suggestive of molten globule states. NMR on hSTRAP(1-219), hSTRAP(1-150) and hSTRAP(151-284) have shown these proteins are not folded at a tertiary structure level. We conclude that a protocol has been established to clone, express and purify various hSTRAP variants and the thermal and secondary structure characteristics of each have been determined, although the 3D structure could not be solved. Pull-down assays followed by proteomic analysis have shown that hSTRAP is implicated in many aspects of cellular regulation.

Ten-Eleven-Translocation (TET) proteins form a family of enzymes responsible for active DNA demethylation by oxidation of 5-methylcytosine. TET proteins play a key role in genomic reprogramming in vitro and in vivo. Although TET proteins are expressed in embryonic stem cells (ESCs), their role in regulating pluripotency remains unclear. In addition, the mechanisms by which TET proteins are recruited to chromatin are largely unknown. To visualise TET protein dynamics during pluripotency and differentiation, the endogenous Tet1/2/3 alleles were fused to epitope tags in ESCs using CRISPR/Cas9. Characterisation of these cell lines showed that TET1 is the highest expressed TET protein in both naïve and primed pluripotent cells. In contrast, TET2 is expressed heterogeneously in ESCs and marks cells with a high self-renewal capacity. To assess the function of Tet genes in pluripotent stem cells, the endogenous Tet1/2/3 ORFs were removed using CRISPR/Cas9. Comparative analysis of single and combined Tet gene knockout ESC lines indicated that Tet1 and Tet2, but not Tet3, play redundant roles to promote loss of pluripotency. Furthermore, Tet-deficient cells retained a naïve morphology in differentiating conditions, suggestive of a LIF-independent self-renewal phenotype. To characterise physiological TET1 protein-protein interactions, TET1 protein partners were identified in ESCs by mass spectrometry and co-immuno-precipitations. This revealed that TET1 interacts with multiple epigenetic and pluripotency-related factors in ESCs. Moreover, detailed characterisation of the interaction between TET1 and NANOG identified three regions of TET1 involved in protein-protein interactions that are conserved in evolution. To investigate TET1 chromatin binding in ESCs, both at the molecular and cellular levels, TET1 was characterised by ChIP-seq analysis and live imaging experiments. Interestingly, TET1 is targeted to chromatin by two different mechanisms, involving distinct protein regions. The interaction with multiple protein partners, including NANOG, might enable TET1 to be targeted to specific chromosomal locations. Additionally, TET1 has the unusual ability to bind mitotic chromatin through its N-terminus, independently of its interaction with NANOG. Together these analyses provide a new understanding of the role of TET proteins in pluripotent cells, as well as a detailed map of TET1 residues involved in protein-protein interactions and mitotic chromatin binding.

骨骼肌分化是由一个包括转录因子、表观遗传调控子和非编码RNA在内的复杂网络共同调控的。YY1能够通过募集PRC2抑制一系列肌肉结构基因的表达，进而抑制肌肉分化。miRNA是一组转录后调控基因表达的小片段非编码RNA，miRNA与转录因子的相互作用已经被广泛证实。在本次研究中，我们证实了一个YY1和肌肉特异性miRNA（miR-1,miR-133和miR-206）的调控回路。实验证实，YY1通过肌肉特异性miRNA增强子区域的YY1结合位点募集PRC2来抑制肌肉特异性miRNA的表达。YY1调控miR-1在体外和体内肌肉分化均被证实有重要意义。另外，我们还证实miR-1能够负反馈作用于YY1，抑制YY1的表达。 / 为了阐述YY1在基因组转录中的作用，我们做了肌肉中YY1的ChIP-seq。测序结果表明在C2C12肌肉母细胞中有1820个YY1结合位点，其中很大部分位于基因间的区域。进一步研究发现，基因间YY1的结合可能调控一些lincRNA，而这些lincRNA在肌肉发育的作用目前尚不清楚。进一步研究这些可能受YY1调节的lincRNA，我们证实了YY1能够正调控两个新的lincRNA，YAM-1和YAM-2。YAM-1在肌肉分化过程中逐渐下调，并且通过正调控他的临近基因miR-715，抑制肌肉分化，而YAM-2能够促进早期的肌肉分化。 / 总之，我们第一次在肌肉细胞中进行了YY1的ChIP-seq，并且证实在肌肉分化过程中转录因子和非编码RNA相互作用的重要性和普遍性。 / Skeletal muscle cell differentiation is a process orchestrated by a complex network of transcription factors, epigenetic regulators and non-coding RNAs. As a repressor of myogenesis, Yin Yang 1 (YY1) silences a number of muscle structural genes through recruiting Polycomb repressive complex2 (PRC2) in proliferating myoblasts. microRNAs (miRNAs) are small non-coding RNAs that regulate gene expression post-transcriptionally, and mounting evidences support the prevalence and functional significance of their interplay with transcription factors (TFs). Here we describe the identification of a regulatory circuit between muscle miRNAs (miR-1, miR-133 and miR-206) and Yin Yang 1 (YY1). The subsequent experimental results demonstrate that YY1 indeed represses muscle miRs expression in myoblasts and the repression is mediated through multiple enhancers and recruitment of Polycomb complex to several YY1 binding sites. YY1 regulating miR-1 is functionally important for both in vitro and in vivo myogenesis. Furthermore, we demonstrate that miR-1 in turn targets YY1, thus forming a negative feedback loop. / To elucidate its role on genome-wide regulation of transcription, here in the second part of this study we performed ChIP-Seq for YY1 in muscle cells. Our results revealed 1820 YY1 binding peaks genome-wide in myoblasts, with a large portion residing in the intergenic region. A close analysis of the intergenic region bound by YY1 uncovered that YY1 may regulate a large number of lincRNAs (Long Intergenic non-coding RNAs), whose roles in skeletal myogenesis have not been explored yet. As further elucidation of the functional roles of YY1-lincRNA regulation, we identified two novel lincRNAs, YAM-1 and YAM-2 as positively regulated by YY1. YAM-1 was found to be down-regulated upon myogenic differentiation and acts as an inhibitor of myoblast differentiation. We further demonstrated that YAM-1 functions by its in cis regulation on a downstream gene, miR-715 which promotes differentiation. YAM-2, on the other hand, appears to promote myogenesis. / Together, our studies not only provide the first genome-wide picture of YY1 association in muscle cells but also uncovered novel regulatory circuits required for skeletal myogenesis and reinforce the idea that regulatory circuitry involving non-coding RNAs and TFs is essential components of myogenic regulatory network. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Lu, Leina. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2012. / Includes bibliographical references (leaves 144-167). / Abstract also in Chinese. / Abstract / 摘要 / Acknowledgement / Publications / List of figures / List of tables / Abbreviations / Table of content / Chapter Chapter 1: --- INTRODUCTION / Chapter 1.1 --- Skeletal Myogenesis --- p.1 / Chapter 1.2 --- Transcriptional Regulation of myogenic differentiation --- p.3 / Chapter 1.2.1 --- Transcriptional regulatory network in myogenic differentiation --- p.3 / Chapter 1.2.2 --- YY1 as a transcription factor in myogenic differentiation --- p.5 / Chapter 1.3 --- Epigenetic Regulation during skeletal muscle differentiation --- p.6 / Chapter 1.4 --- microRNA: Post-transcriptional regulation on myogenic differentiation --- p.11 / Chapter 1.4.1 --- Muscle specific miRNAs in skeletal myogenic differentiation --- p.15 / Chapter 1.4.2 --- Non-muscle specific miRNAs in skeletal myogenic differentiation --- p.20 / Chapter 1.4.3 --- miRNAs and skeletal muscle diseases --- p.23 / Chapter 1.5 --- Long Non-coding RNAs --- p.26 / Chapter 1.5.1 --- Long Non-coding RNAs and lincRNAs --- p.26 / Chapter 1.5.2 --- LincRNAs in muscles --- p.30 / Chapter Chapter 2: --- MATERIALS AND METHODS / Chapter 2.1 --- C2C12 cell line --- p.32 / Chapter 2.2 --- Primary Myoblast isolation and in vitro culture --- p.32 / Chapter 2.3 --- Animal studies --- p.33 / Chapter 2.4 --- RNA extraction --- p.34 / Chapter 2.5 --- RT-PCR and Real-Time RT-PCR --- p.35 / Chapter 2.6 --- Transfection and infection --- p.37 / Chapter 2.7 --- Oligonucleotides --- p.38 / Chapter 2.8 --- Dual-luciferase reporter assay --- p.43 / Chapter 2.9 --- Immunofluorencence staining --- p.44 / Chapter 2.10 --- Antibodies --- p.45 / Chapter 2.11 --- Protein extraction and Western blotting --- p.46 / Chapter 2.12 --- DNA constructs --- p.48 / Chapter 2.13 --- Mutagenesis --- p.49 / Chapter 2.14 --- RNA-Fluorescence In Situ Hybridization (RNA-FISH) --- p.51 / Chapter 2.15 --- C2C12 cells with YY1-stably knocked down --- p.52 / Chapter 2.16 --- Rapid Amplification of cDNA Ends (RACE) --- p.53 / Chapter 2.17 --- Chromatin Immunoprecipitation (ChIP) --- p.55 / Chapter 2.18 --- ChIP-PCR --- p.58 / Chapter 2.19 --- ChIP-sequencing --- p.58 / Chapter 2.20 --- Northern blotting --- p.59 / Chapter 2.21 --- Prediction of miRNA targets --- p.60 / Chapter 2.22 --- Statistical analysis --- p.60 / Chapter Chapter 3: --- Results / Chapter 3.1 --- YY1-miR-1/133 regulatory circuitry in skeletal myogenesis --- p.61 / Chapter 3.1.1 --- YY1 decreases miR-1/133 during skeletal muscle differentiation --- p.61 / Chapter 3.1.1.1 --- Negative correlation between YY1 and miR-1/133 during C2C12 differentiation --- p.61 / Chapter 3.1.1.2 --- Negative correlation between YY1 and miR-1/133 in primary cell differentiation --- p.63 / Chapter 3.1.1.3 --- Negative correlation between YY1 and miR-1/133 in postnatal muscle development and mdx mouse model --- p.65 / Chapter 3.1.1.4 --- Deletion of YY1 upregulates miR-1/133 both in C1C12 and primary myoblast --- p.68 / Chapter 3.1.1.5 --- Deletion of YY1 upregulates miR-1/133 at the transcriptional level --- p.70 / Chapter 3.1.2 --- YY1 represses miR-1/133 by binding to 4 enhancers --- p.72 / Chapter 3.1.2.1 --- Four enhancers of miR-1/133 with potential YY1 targeting sites --- p.72 / Chapter 3.1.2.2 --- YY1 represses the four enhancers’ activities --- p.75 / Chapter 3.1.2.3 --- Depletion of YY1 up-regulates the four enhancers’ activities --- p.77 / Chapter 3.1.2.4 --- YY1 directly binds to the putative binding sites and mediates the repression on miR-1/133 --- p.79 / Chapter 3.1.2.5 --- YY1 recruits Ezh2 to the enhancers which subsequently causes histone modification --- p.82 / Chapter 3.1.3 --- YY1 repressing miR-1/133 is functionally significant in myogenesis --- p.84 / Chapter 3.1.3.1 --- Negative correlation between YY1 and miR-1/133 in CTX induced muscle regeneration model --- p.84 / Chapter 3.1.3.2 --- Depletion of YY1 in CTX induced muscle regeneration model promotes miR-1/133 expression --- p.87 / Chapter 3.1.3.3 --- Depletion of YY1 in CTX induced muscle regeneration model promotes muscle differentiation --- p.89 / Chapter 3.1.4 --- miR-1 can target YY1 forming a feedback loop --- p.92 / Chapter 3.1.5 --- miR-1 can repress Pax7 by targeting two binding sites on 3’UTR --- p.95 / Chapter 3.1.5.1 --- miR-1 targets Pax7 by binding to two target sites --- p.95 / Chapter 3.1.5.2 --- miR-1 represses Pax7 forming an YY1-miR-1-Pax7 regulating circuitry in skeletal myogenesis --- p.98 / Chapter 3.1.6 --- Conclusion: YY1-miR-1-Pax7 regulatory circuitry in skeletal myogenesis --- p.100 / Chapter 3.2 --- ChIP-seq reveals YY1-lincRNA regulation in skeletal myogenesis --- p.102 / Chapter 3.2.1 --- ChIP-seq uncovered a large number of genes under YY1 regulation --- p.102 / Chapter 3.2.2 --- ChIP-seq reveals that YY1 associates with lincRNA loci --- p.105 / Chapter 3.2.2.1 --- YY1 associates with lincRNA-YAM loci --- p.105 / Chapter 3.2.2.2 --- YY1 positively regulates YAM-1 and YAM-2 both in vitro and in vivo --- p.107 / Chapter 3.2.3 --- YY1-YAM-1-miR-715 regulatory pathway in muscle differentiation --- p.109 / Chapter 3.2.3.1 --- Genomic organization and cellular localization of YAM-1 --- p.109 / Chapter 3.2.3.2 --- Expression of YAM-1 decreases during myogenic differentiation --- p.112 / Chapter 3.2.3.3 --- YAM-1 represses myogenic differentiation both in vitro and in vivo --- p.115 / Chapter 3.2.3.3.1 --- YAM-1 inhibits C2C12 differentiation --- p.115 / Chapter 3.2.3.3.2 --- YAM-1 inhibits muscle differentiation in vivo --- p.117 / Chapter 3.2.3.4 --- A functional YY1-YAM-1-miR-715 regulatory axis in skeletal myogenic differentiation --- p.119 / Chapter 3.2.3.4.1 --- miR-715 is down-regulated during muscle differentiation --- p.119 / Chapter 3.2.3.4.2 --- miR-715 is under the regulation of YY1-YAM-1 --- p.122 / Chapter 3.2.3.4.3 --- miR-715 represses muscle differentiation forming a YAM-1-miR-715 regulatory axis during muscle differentiation --- p.124 / Chapter 3.2.4 --- YAM-2 promotes early myogenic differentiation --- p.126 / Chapter 3.2.4.1 --- Genomic organization and cellular localization of YAM-2 --- p.126 / Chapter 3.2.4.2 --- YAM-2 is regulated during myogenic differentiation --- p.129 / Chapter 3.2.4.3 --- YAM-2 promotes early myogenic differentiation --- p.131 / Chapter Chapter 4: --- DISCUSSION / Chapter 4.1. --- YY1-miRNA regulatory circuit in skeletal myogenesis --- p.133 / Chapter 4.2 --- YY1 mediates epigenetic modification in skeletal myogenesis --- p.135 / Chapter 4.3 --- miRNAs in skeletal myogenesis --- p.136 / Chapter 4.4 --- YY1 regulates long intergenic non-coding RNAs in skeletal myogenesis --- p.138 / Chapter Chapter --- 5: SUMMARY AND FUTURE WORK --- p.142 / REFERENCE --- p.144

3D proteomics combines chemical cross-linking with mass spectrometry to study the structure of protein assemblies and protein-protein interactions both in vitro and in vivo by providing distance constraints that indicate which residues are in close spatial proximity. I addressed the main bottleneck of this technology: the reliable identification of cross-linked peptides. Reporter ion signatures for cross-linked peptides were developed, by fragmenting model compounds containing two lysine residues joined by a cross-linker backbone or a lysine residue modified with a hydrolysed cross-linker. The reporter ion signatures showed 97% specificity at 90% sensitivity and segregated cross-linked from modified and linear peptides. They decreased the false discovery rate of the identification of cross-linked peptides from 5% to 1% in a large dataset. The signatures permit data sorting during and after mass spectrometry acquisition. The advanced 3D proteomics workflow was applied to study the protein-protein interactions in Mycoplasma pneumoniae cells. In lysates of the bacterium we identified 128 protein-protein interactions (of which 24 are novel) and obtained in vivo topological data on 208 proteins, even for cases where high-resolution structures are not yet available. We showed that our data are in excellent agreement with crystal structures of proteins and complexes where available. We defined a network of ribosomal and RNA polymerase proteins that reveals an intricate link between transcription and translation in bacteria. We demonstrated that the method is suitable for identification of homomultimeric protein complexes by exploiting peptide pairs of identical amino acid sequence. The technology has the potential to provide a complete protein interaction network map after the selective enrichment of cross-lined peptides is achieved. The method was next applied to investigate the structure of condensin and cohesin complexes, which play a crucial role in stabilization of chromosome structure during mitosis. The complexes were purified, cross-linked and their linkage map created. The condensin coiled coil cross-linked on the entire length was modeled. The information was used to direct the analysis of in situ cross-linked condensin in intact chromosomes. I found two high confidence linkages between SMC2 and SMC4 coiled coils and identified H2A as a potential condensin receptor on chromosomes.

Le cytosquelette bactérien des homologues d’actine (protéines de la famille MreB) joue un rôle majeur dans la morphogénèse cellulaire. Des homologues de MreB sont retrouvés chez la plupart des espèces bactériennes non sphériques, où ils sont essentiels pour la viabilité cellulaire. Les bactéries à Gram-positif ont généralement plusieurs isoformes. L’organisme modèle Bacillus subtilis en possède trois : MreB, Mbl et MreBH, tous trois impliqués dans la détermination de la forme de la cellule. Le postulat actuel est une organisation, des complexes de synthèse du peptidoglycane, le long des parois latérales par les filaments hélicoïdaux des MreB-like. Cependant, les mécanismes moléculaires et les protéines effectrices impliqués dans cette fonction ne sont pas encore élucidés. Par analogie avec les rôles de l’actine eucaryote, des implications dans d’autres processus cellulaires cruciaux et la présence de partenaires protéiques sont également attendus pour les actines procaryotes. Afin d’explorer les rôles des protéines MreB chez B. subtilis nous avons généré, par des criblages génomiques double hybride chez la levure, un réseau d’interaction protéine-protéine centré sur MreB, Mbl et MreBH. Une vérification systématique et drastique de toutes les interactions obtenues lors des criblages a été réalisée afin d’éliminer les faux positifs. Les interactions identifiées révèlent des liens entre les protéines MreB-like et seize protéines issues de catégories fonctionnelles variées ou de fonction inconnue. Une étude exploratoire a été menée pour huit des protéines partenaires par des approches in silico et in vivo et nous a permis de sélectionner une seule interaction à caractériser plus en détail. Nous nous sommes principalement intéressés à l’interaction physique et directe entre MreB et DapL, une protéine essentielle vraisemblablement impliquée dans la voie de biosynthèse des précurseurs du peptidoglycane, par analogie à DapE d’E. coli. La caractérisation approfondie de DapL a confirmé son essentialité dans la synthèse du peptidoglycane. Bien que l’interaction MreB-DapL ait été confirmée biochimiquement, son rôle biologique exact n’a pas été élucidé. Cependant, nous avons mis en évidence d’autres interactions entre MreB et DapG, LysA et MurE, des enzymes également impliquées dans les étapes précoces de la synthèse du peptidoglycane. L’existence de telles interactions renforce le rôle du cytosquelette MreB de B. subtilis dans l’orchestration des machineries de synthèse de la paroi cellulaire. / Bacterial actin homologues (MreB proteins) play a major role in cell morphogenesis in non-spherical bacteria. The prevailing model postulates that helical, membrane-associated MreB-like filaments organize elongation-specific peptidoglycan-synthesizing complexes along the sidewalls. However, the mechanistic details, as well as the effector proteins of MreBs morphogenetic function, remain to be elucidated. MreB proteins are also involved in DNA segregation, cell polarity, cell motility and, by analogy to eukaryotic actins, possibly in other functions that require the targeting and accurate positioning of proteins and molecular complexes in the cell. Gram-positive bacteria usually have more than one MreB isoform. Our model organism, Bacillus subtilis, has three called MreB, Mbl and MreBH. To explore the roles of the MreB cytoskeleton in B. subtilis, we used genome-wide yeast two-hybrid screens to identify proteins that physically associate with MreB, Mbl and MreBH. Stringent specificity assays were systematically performed to remove false positives and confirm the specificity of all potential interactions identified in the screens. A protein-protein interaction network centered on the three MreBs was generated which includes 16 protein partners. This interaction network provides insights into the links of MreB proteins with proteins belonging to several functional categories as well as proteins of unknown function. An exploratory study was conducted in silico and in vivo for 8 of the partner proteins identified in the network and allowed us to select one interaction for a more in-depth analysis. We next focused in the physical interaction between MreB and DapL, an essential protein presumably involved in the early steps of peptidoglycan biosynthesis. The characterization of DapL confirmed its essential role in cell wall synthesis. The MreB-DapL interaction was confirmed biochemically and we showed that MreB also associates with other proteins involved in the synthesis of the PG precursors (DapG, LysA and MurE). Together, these results suggest that B. subtilis MreB orchestrates the PG biosynthetic cytosolic machineries to achieve and maintain its rod shape.

Drugs exert their (therapeutic) effects via molecular-level interactions with proteins and other biomolecules. Computational prediction of drug-protein interactions plays a significant role in the effort to improve our current and limited knowledge of these interactions. The use of the putative drug-protein interactions could facilitate the discovery of novel applications of drugs, assist in cataloging their targets, and help to explain the details of medicinal efficacy and side-effects of drugs. We investigate current studies related to the computational prediction of drug-protein interactions and categorize them into protein structure-based and similarity-based methods. We evaluate three representative structure-based predictors and develop a Protein-Drug Interaction Database (PDID) that includes the putative drug targets generated by these three methods for the entire structural human proteome. To address the fact that only a limited set of proteins has known structures, we study the similarity-based methods that do not require this information. We review a comprehensive set of 35 high-impact similarity-based predictors and develop a novel, high-quality benchmark database. We group these predictors based on three types of similarities and their combinations that they use. We discuss and compare key architectural aspects of these methods including their source databases, internal databases and predictive models. Using our novel benchmark database, we perform comparative empirical analysis of predictive performance of seven types of representative predictors that utilize each type of similarity individually or in all possible combinations. We assess predictive quality at the database-wide drug-protein interaction level and we are the first to also include evaluation across individual drugs. Our comprehensive analysis shows that predictors that use more similarity types outperform methods that employ fewer similarities, and that the model combining all three types of similarities secures AUC of 0.93. We offer a first-of-its-kind analysis of sensitivity of predictive performance to intrinsic and extrinsic characteristics of the considered predictors. We find that predictive performance is sensitive to low levels of similarities between sequences of the drug targets and several extrinsic properties of the input drug structures, drug profiles and drug targets.