University of Southern California Dissertations and Theses

Three advancements in biotechnology: new tools for synthetic biology and next generation sequencing

(USC Thesis Other)

Three advancements in biotechnology: new tools for synthetic biology and next generation sequencing

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content THREE ADVANCEMENTS IN BIOTECHNOLOGY New Tools for Synthetic Biology and Next Generation Sequencing by David Ryan Tyrpak A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Pharmaceutical Sciences) December 2020 ii Dedicated to my wife Christina, for her endless love and support iii ACKNOWLEDGMENTS I would like to thank my PI Dr. Andrew Mackay for guiding and training me during this PhD. Over the past five years in his lab I saw him lead by example and he taught me much about good leadership, effective management, and constructive criticism, in addition to science. I would also like to thank Dr. Brooke Hjelm, who has also served as my mentor and given me great support and guidance. I would like to thank my committee members Dr. Curtis Okamoto, Dr. Sarah Hamm- Alvarez, and Dr. Bangyan Stiles for their generous advice and support. I also want to appreciate Yue Wang, Hugo Avila, Mincheol Park, Anh Truong, Hao Gou, Zhe Li, Siqi Lei, Yaocun Li, Stav Grossfeld, and Michelle Webb for contributing data to my thesis. I would like to acknowledge all the funding resources for my research: RO1 GM114839 to Dr. MacKay, my F31 fellowship (grant F31DK118881) and the USC Provost fellowship, P30 EY029220 to the USC Ophthalmology Core Grant in Vision Research, P30 CA014089 to the USC Norris Comprehensive Cancer Center, P30 DK048522 to the Liver Histology Core of the USC Research Center for Liver Diseases, the USC Ming Hsieh Institute, the Gavin S. Herbert Endowed Chair of Pharmaceutical Sciences, the Translational Research Laboratory at USC School of Pharmacy, and the USC Cell and Tissue Imaging Core. I also want to extend special thanks to the NABEC autopsy donors and family members for their contributions to chapter 5 of my thesis. Finally, I want to thank my family and my wife. Their support and encouragement carried me this far. iv TABLE OF CONTENTS DEDICATION ii ACKNOWLEDGMENTS iii LIST OF TABLES ix LIST OF FIGURES x PROLOGUE 1 CHAPTER 1 Stimuli responsive proteins as intracellular switches 1.1 Stimuli responsive proteins 5 1.2 Chemical stimuli 6 1.3 Optogenetics 8 1.4 Magnetic stimuli 13 1.5 Thermally responsive proteins 15 1.6 Conclusion 16 CHAPTER 2 Caveolin elastin-like polypeptide fusions mediate temperature-dependent assembly of caveolar microdomains 2.1 Abstract 17 2.2 Introduction 18 2.3 Materials and methods 21 2.3.1 Cell culture 21 2.3.2 Plasmid construction and amino acid sequences 21 2.3.3 Western blot 24 2.3.4 Live cell imaging and immunofluorescence 25 2.3.5 CAV1-ELP/CAV1-WT colocalization with 26 Cholera Toxin Subunit B (CXTB) 2.3.6 Cavin-1 immunofluorescence 28 2.3.7 CTXB Binding quantification 28 2.3.8 Statistics 30 2.4 Results 30 2.4.1 Construction and confirmation of a library of CAV1-ELP fusions 30 2.4.2 CAV1-ELP transition temperature (Tt) can be visually 33 v determined in live cells 2.4.3 CAV1-V96 microdomains colocalize with CTXB 36 2.4.4 Inducing CAV1-V96 microdomain formation before 39 CTXB labeling reduces CTXB signal 2.4.5 Cavin 1 exhibits broad colocalization with CAV1, regardless of 42 ELP sequence or temperature 2.5 Discussion 44 2.6 Conclusion 45 CHAPTER 3 Tunable assembly of protein-microdomains in living vertebrate embryos 3.1 Abstract 46 3.2 Introduction 46 3.3 Materials and methods 48 3.3.1 Zebrafish husbandry and care 48 3.3.2 Plasmid construction 49 3.3.3 mRNA preparation and microinjection 49 3.3.4 ELP purification and physicochemical characterization 50 3.3.5 Immunoblotting 50 3.3.6 Zebrafish imaging 50 3.3.7 Image processing 51 3.3.8 Statistics 51 3.3.9 Author contributions 52 3.4 Results 53 3.4.1 ELP expression and microdomain assembly in a zebrafish embryo 53 3.4.2 ELPs can co-assemble different proteins in vivo 56 3.4.3 Transition temperature in vivo can be controlled through 58 ELP sequence and concentration 3.4.4 Duration of in vivo ELP assembly can be tuned 62 3.4.5 ELPs do not affect survival or embryonic development 65 3.5 Conclusion 66 CHAPTER 4 Method for determining the intracellular transition temperature of elastin- like polypeptide fusion proteins 4.1 Abstract 67 4.2 Introduction 68 vi 4.3 Materials 4.3.1 Cell culture 73 4.3.2 Transfection 73 4.3.3 Live cell imaging 73 4.3.4 Temperature control during live cell imaging 74 4.3.5 Image analysis 74 4.3.6 Fixed cell imaging to confirm GFP-V60 colocalization 74 with non-fluorescent ELP fusion 4.4 Methods 75 4.4.1 Cell culture 75 4.4.2 Transfection 76 4.4.2.1 Preparing MatTek plates for cell culture 76 4.4.2.2 Transfection 77 4.4.3 Live cell imaging 78 4.4.3.1 Imaging settings 78 4.4.3.2 Temperature control during live cell imaging 81 4.4.4 Image analysis 84 4.4.4.1 Blinding and randomizing imaging data 84 4.4.4.2 Visual determination of intracellular Tt 85 4.4.4.3 Microdomain analysis 86 4.4.4.4 Background corrected integrated density (BCID) 87 4.4.4.5 Standard deviation determination of intracellular Tt 88 4.4.5 Fixed cell imaging to confirm GFP-V60 colocalization 92 with non-fluorescent ELP fusion 4.4.5.1 Transfection 92 4.4.5.2 Temperature incubation, fixation, and immunostaining 93 4.4.5.3 Imaging 94 4.5 Validation and Expected Results 95 4.5.1 Visual determination of intracellular transition temperature 95 of CAV1-ELP library 4.5.2 The effect of intracellular GFP-V60 concentration on the 99 intracellular Tt of CAV1-ELP fusion proteins 4.5.3 Fixed cell imaging to confirm GFP-V60 colocalization 101 with non-fluorescent CAV1-V96 4.5.4 Visual determination of intracellular transition temperature 102 of ELP-CLC library 4.5.5 The effect of intracellular GFP-V60 concentration on the 106 intracellular Tt of ELP-CLC fusion proteins 4.6 Conclusion 108 CHAPTER 5 Mitochondrial DNA deletions and copy number in whole genome seecig (WGS) daa: aae f agig ad Paki dieae vii 5.1 Abstract 109 5.2 Introduction 110 5.3 Materials and methods 113 5.3.1 General bioinformatics and statistical software 113 5.3.2 Whole genome sequencing (WGS) 113 5.3.3 Splice-Break analysis of WGS 114 5.3.4 mtDNA copy number determination with fastMitoCalc 114 5.3.5 Unique deletions per 10K coverage 115 5.3.6 Analysis of mtDNA deletions in small open reading frames 115 (sORFs) encoding mitochondrial derived peptides 5.3.7 Statistics 116 5.4 Results 116 5.4.1 Description of NABEC control cohort 116 5.4.2 mtDNA deletions are positively correlated with age in both 118 Frontal Cortex and Cerebellum, with the Frontal Cortex accumulating deletions at a greater rate 5.4.3 Splice Break with WGS detected 1,728 unique deletions, 120 with the great majority attributed to the frontal cortex 5.4.4 The majority of mtDNA deletions fall within the major arc 121 of the mitochondrial genome 5.4.5 mtDNA deletions accumulate with age inside regions coding 123 for mitochondrial-derived peptides 5.4.6 Splice Break WGS vs. Splice Break LR-PCR exhibit 125 biases in discovered mtDNA deletions 5.4.7 Measuring mtDNA copy number from WGS with fastMitoCalc 127 5.4.8 The Frontal Cortex and Cerebellum display significant 131 differences in mtDNA copy number 5.4.9 Decipion of he Pakinon Dieae e age mached 134 healthy control (PDVHC) cohort 5.4.10 PD cerebellums display higher mtDNA copy number than 136 age-matched healthy controls 5.4.11 PD cerebellums appear to maintain copy number with age 138 5.4.12 Cerebellum mtDNA copy number is increased in samples 140 with more severe subtsantia nigra depigmentation 5.5 Discussion 142 5.6 Conclusion 144 viii CONCLUSION AND FUTURE DIRECTIONS 145 REFERENCES 147 ix LIST OF TABLES 2.1 Nomenclature, amino acid sequence, and phase behavior of expressed proteins. 31 3.1 Nomenclature, amino acid sequence, and phase behavior of expressed proteins. 54 3.2 Statistics for Figure 3.3 59 4.1 Nomenclature, amino acid sequence, and phase behavior of expressed proteins. 71 5.1 mtDNA (rCRS NC_012920.1) coordinates of sORFs in the 115 mitochondrial genome 5.2 Summary of the NABEC control cohort across brain regions. 117 5.3 Summary of the PDVHC cohort across disease status 135 5.4 Summary of the PDVHC cohort sample sizes across brain banks 135 x LIST OF FIGURES Chapter 1 Figure 1.1 The ideal intracellular switch. 5 Chapter 2 Figure 2.1 Construction and confirmation of a library of CAV1-ELP fusions. 32 Figure 2.2 CAV1-ELP Transition Temperature (Tt) can be visually determined 35 in live cells. Figure 2.3 CAV1-V96 microdomains colocalize with CTXB. 37 Figure 2.4 Temperature mediated self-assembly of CAV1-V96 induces 38 internalization of fluorescent Cholera Toxin Subunit B (CTXB). Figure 2.5 CAV1-V96 self-assembly removes a fraction of CTXB binding 41 sites from the plasma membrane. Figure 2.6 Cavin 1 exhibits broad colocalization with CAV1, regardless of 43 ELP sequence or temperature. Chapter 3 Figure 3.1 ELP expression and microdomain assembly in a zebrafish embryo. 55 Figure 3.2 ELPs can co-assemble different proteins in vivo. 57 Figure 3.3 mRNA injection amount is the only significant predictor of GFP-ELP 60 protein levels, as measured by whole field integrated density of confocal microscope images. Figure 3.4 ELP self-assembly temperature can be tuned both in vitro and in vivo. 61 Figure 3.5 ELPs can be tuned for short- or long-term microdomain assembly. 63 Figure 3.6 At late blastula stages (4-5 hpf), embryos injected with GFP-V96 64 display more microdomains/particles than embryos injected with GFP-SI. Figure 3.7 ELPs do not affect survival or embryonic development. 65 xi Chapter 4 Figure 4.1 Intracellular Tt of non-fluorescent ELP fusion proteins can be 70 visually determined in live cells through co-assembly with a fluorescent ELP fusion protein. Figure 4.2 Stepwise schematic of the experimental procedure. 72 Figure 4.3 Equipment configuration during live cell imaging. 83 Figure 4.4 The standard deviation of intracellular pixel values during a 91 temperature ramp provides an alternative method for determination of intracellular transition temperature. Figure 4.5 Cells transfected with CAV1-V96 + GFP-V60 exhibit a reduced 97 transition temperature and large microdomains compared to transfection with GFP-V60 alone or CAV1-A96 + GFP-V60. Figure 4.6 Changes in intracellular transition temperature (Tt), microdomain 98 size, and microdomain number can be quantified from live cells transfected with different ELP fusions. Figure 4.7 Intracellular transition temperature (Tt) and GFP-V60 integrated 100 density follow a log-linear relationship [CAV1-ELP dataset]. Figure 4.8 Indirect staining in fixed cells confirms that CAV1-V96 co-assembles 101 with GFP-V60. Figure 4.9 Cells transfected with ELP-CLC + GFP-V60 display microdomains 104 which are similar to GFP-V60 alone. Figure 4.10 Changes in intracellular transition temperature (Tt) can be 105 quantified from live cells transfected with different ELP-CLC fusions. Figure 4.11 Intracellular transition temperature (Tt) and GFP-V60 integrated density 107 follow a log-linear relationship [ELP-CLC dataset]. Chapter 5 Figure 5.1 The number of unique mtDNA deletions increases with age in both 119 Frontal Cortex and Cerebellum. Figure 5.2 After filtering, Splice Break with WGS identified 1,728 unique 120 mtDNA deletions. xii Figure 5.3 The majority of the 1,728 mtDNA deletions fall within the 122 major arc, between OL and OH. Figure 5.4 mtDNA deletions accumulate with age inside regions coding 124 for mitochondrial-derived peptides. Figure 5.5 Splice Break WGS vs. LR-PCR exhibit biases towards 126 small and large deletions, respectively. Figure 5.6 Frontal Cortex and Cerebellum display similar levels of 129 autosomal DNA coverage but different levels of mtDNA coverage. Figure 5.7 fastMitoCalc performs well even at low autosomal 130 sequencing coverage. Figure 5.8 Frontal Cortex displays elevated mtDNA copy number 132 compared with Cerebellum. Figure 5.9 mtDNA copy number significantly decreases with age in the 133 frontal cortex but not the cerebellum. Figure 5.10 PD cerebellums display higher mtDNA copy number than 137 age-matched healthy controls. Figure 5.11 PD cerebellums appear to maintain copy number with age. 139 Figure 5.12 Cerebellum mtDNA copy number is increased in samples 141 with more severe subtsantia nigra depigmentation. 1 PROLOGUE This dissertation describes original work in two different scientific fields: 1) Synthetic biology, which is concerned with controlling and mimicking biological systems. 2) Bioinformatics, which is focused on the use and development of computer programs to analyze complex biological data. The work described in this thesis is mainly concerned with innovative methodologies for use in biomedical research. However, in Chapter 5 some novel biological findings regarding mtDNA dynamics in regard to aging and Parkinson’s Disease status are discussed. Note that Chapters 2 and 3 are nearly verbatim quoted from my published manuscripts. 1,2 The work detailed in Chapter 3 was performed in conjunction with Zhe Li, and we are co-first authors on the published manuscript. Chapter 1 gives a broad overview of the use and application of stimuli responsive proteins for intracellular switching. This chapter also introduces Elastin-like polypeptides. Chapters 2 and 3 describe two different projects which are unified by the use of ELPs for the control of protein assembly inside of biological systems. ELPs are biocompatible, environmentally responsive polymers inspired from the connective tissue tropoelastin. The fundamental unit of the ELP is a pentameric amino acid repeat of (VPGXG)n, where the X residue can be selected to control hydrophobicity, and the number of repeats, n, determines the molecular weight. ELPs can rapidly (within seconds to minutes) phase-separate and self-assemble into protein-rich domains in response to heating, whereby they form a secondary aqueous phase known as a coacervate. In cells, individual coacervate droplets are typically several hundred nanometers to several micrometers in diameter. This phase separation is thermodynamically reversible, and ELP 2 coacervates quickly resolubilize into bulk water upon cooling. The temperature of self-assembly, termed the transition temperature (Tt), is determined largely by the hydrophobicity of the guest residue and the ELP molecular weight. The modular design of ELPs, combined with their stimuli-responsive self-assembly and biocompatibility, makes them an excellent candidate for biological studies requiring the assembly of macromolecular complexes. Since so many intracellular processes are governed by the assembly of such complexes, previous researchers in the MacKay lab have exploited ELP fusion proteins (ELPs cloned to an effector protein) to modulate Epidermal Growth Factor Receptor signaling and reversibly inhibit Clathrin-mediated endocytosis. This unique approach to controlling intracellular events is termed “Intracellular Switching”, because the intracellular self-assembly of ELP fusion proteins rapidly activates or inactivates a cellular response. Although previous studies used ELP fusion proteins to control Clathrin-mediated endocytosis, this approach had not been extended to Caveolae-mediated endocytosis, the other major endocytic pathway. Chapter 2 describes the first use of thermo-responsive CAV1-ELP proteins (Caveolin 1 cloned to an ELP) to modulate CAV1. When CAV1-ELPs self-assemble, they form intracellular caveolar microdomains which internalize cholera toxin subunit B, a commonly used marker of Caveolae-mediated endocytosis. This chapter supports the hypothesis that ELPs can be used to create thermo-responsive mutants of CAV1 that can be thermally induced to undergo a process reminiscent of caveolae-mediated endocytosis. 3 The previous “intracellular switching” projects in the MacKay lab, including the use of CAV1-ELPs, were performed in cell culture. Chapter 3 tests the hypothesis that fluorescent ELP fusions reversibly form tunable protein microdomains inside the single cells of living vertebrate embryos. This proof-of-principle hypothesis lays the groundwork for future in vivo intracellular switching projects. By tuning ELP length, sequence, and concentration, fluorescent ELP protein microdomains can be assembled at different temperatures, in varying sizes, or for desired periods of time. For the first time in a multicellular living embryo, these studies demonstrate that temperature mediated ELP assembly can reversibly manipulate assembly of subcellular protein complexes, which may have applications in the study and manipulation of in vivo biological functions. Chapter 4 describes the live cell imaging technique I developed to determine the intracellular Tt of non-fluorescent ELP fusions proteins. To assist with many of the image analysis techniques described in this chapter (and encountered throughout my PhD work), I developed an easily installed FIJI plugin named SIAL (Simple Image Analysis Library), which contains programs for image randomization and blinding, phenotype scoring, and ROI selection. These tasks are important parts of the protocol detailed in this chapter but are also commonly employed in other image analysis workflows. Chapter 5 describes a new bioinformatic technique for quantifying deletions in the mitochondrial genome. Deletions in the mitochondrial genome, where hundreds to thousands of base pairs are deleted, have been described in the literature for decades and are correlated with a litany of degenerative illnesses. However, existing methods for detecting mitochondrial DNA 4 deletions are low throughput and not suitable for de novo discovery. It was recently reported that next generation sequencing and existing bioinformatic tools could be used to detect and quantify mitochondrial DNA deletions. However, this technique required isolation and amplification of mitochondrial DNA, throwing away autosomal DNA in the process. Chapter 5 explores the hypothesis that whole genome sequencing and existing bioinformatic tools can be used to quantify and discover mitochondrial DNA deletions. Because this technique retains autosomal sequencing data, it can be used with existing computational tools to quantify mitochondrial genome copy number (which requires autosomal sequencing data as a control). Thus, this new technique allows researchers to probe autosomal DNA variation, mitochondrial DNA deletions, and mitochondrial genome copy number in the same sample. In addition to describing this new technique, it is also shown that the cerebellums of patients with Parkinson’s Disease (PD) display elevated mitochondrial DNA copy number. This novel finding is in agreement with previous neuro-imaging studies suggesting that PD cerebellums hyper-activate to compensate for the loss of muscle control caused by the death of dopamine producing neurons in the midbrain. 5 CHAPTER 1 Stimuli responsive proteins as intracellular switches 1.1 Stimuli responsive proteins Cells rely on the selective formation and disassembly of proteins to achieve desired outcomes. For example, RNA transcription, endocytosis, and signaling cascades are all accomplished by the selective assembly and disassembly of proteins in response to stimuli. 3-5 Because stimuli responsive protein assembly is so prevalent inside of cells, when molecular biologists seek to understand the intracellular function of a protein, they will often try to inhibit that protein’s assembly behavior by either reducing its expression or by inhibiting its ability to assemble with its functional protein partners. Conversely, an experimenter might decipher a protein’s function by upregulating that protein’s expression or its assembly activity with its functional partners. In both cases, the goal is to selectively and precisely affect the protein of interest and the corresponding pathway of interest. Based on this rationale, the ideal tool for affecting protein function and assembly should be rapid, accurate, reversible, and adaptable. Such a tool would behave as a “intracellular switch”, allowing researches to turn off and on targeted protein function and assembly in order to delineate biological impact. (Figure 1.1). Figure 1.1 The ideal intracellular switch. A technology which can affect intracellular protein function like a light switch would be useful in a broad spectrum of biological studies. 6 Currently, the most widely tools for altering protein activity are small molecules and genetic silencing methods (siRNA, gene editing). Although much good work has been accomplished with both methods, small molecules often display widespread off-target effects, limited reversibility, and poor spatiotemporal resolution, owing to their passive diffusion across the cell membrame. 6-9 Genetic silencing methods also suffer from off-target activity, and in addition are either irreversible or require hours to days for effect, thus allowing compensatory cellular pathways to be activated. 10-13 In both cases, the side-effects of these methods may confound experimental results. To circumvent these limitations, control over a protein’s function could be handled more directly by creating a stimuli-responsive mutant of that protein. In the absence of external stimuli, the mutant protein would behave like its wild-type counterpart, while in response to external stimuli, the mutant version would rapidly assemble to activate or inactivate its pathway. Suitable stimuli might include chemicals, light, magnetic fields, or heat. In the below sections I summarize the first three strategies before introducing thermally responsive fusion proteins. 1.2 Chemical stimuli Owing to their ease of use, biomedical research has long relied on small molecule inhibitors and agonists to study cellular pathways. While the exogenous application of chemicals to study cell biology is convenient and typically rapid acting, the field has long realized that small molecules display off-target effects, limited reversibility, and poor spatiotemporal resolution. 6-9 7 Chemically induced dimerization (CID) was developed to overcome these limitations. The basic logic of CID relies on creating mutant proteins which only dimerize in the presence of a small molecule or enzymatic dimerizing agent. 14-16 By selecting dimerization mutants that localize to specific regions in the cell and which perform specific functional activities, a well-designed CID system can alter protein activity with high spatial resolution. The first CID system was developed in 1993 by Spencer and collagues. 17 As a dimerization agent they synthesized FK1012, a bivalent derivative of the immunosuppressive drug FK506/tacrolimus, which binds to FKBP (i.e. FK506 Binding Protein). They then molecularly cloned chimeric T-Cell receptors (TCRs) that lacked extracellular and transmembrane regions but expressed the intracellular TCR z chain, as well as an FKBP domain. Addition of FK1012 induced intracellular dimerization of the chimeric TCRs and activated target gene expression. Spencer and colleagues were the first to successfully demonstrate the CID concept, but they didn’t examine reversibility. In the twenty-five years since their landmark paper, a variety of reversible CID modalities have been created. Reversible strategies employing FKBP enjoyed serendipitous advancement in 2000, when it was discovered that a single point mutation can produce a mutant FKBP which naturally forms homodimers but rapidly dissociates in the presence of ligand. 18 In 2016, Barrero and colleagues developed an improvement of the reversibly dimerizing mutant of FKBP. 19 Orthogonal approaches for reversible CID have also been developed. In 2013, Inoue and co-workers created a two-step reversible CID system to control cellular processes. 20 This approach 8 first relied on rapamycin to induce dimerization of target proteins at the plasma membrane, and then used a second dimerizer, GA3-AM, to sequester the plasma membrane complexes to other parts of the cell. Although this translocation approach did successfully induce reversible changes in plasma membrane morphology, it did not actually reverse the initial rapamycin induced dimerization and it did not allow for repeated cycles of activation and deactivation. Another orthogonal approach for reversible CID was created by Liu and colleagues, who employed a competitive binder of their dimerization agent to rapidly reverse dimerization. 21,22 CID systems continue to enjoy broad use in biomedical research. Recent advancements have focused on improving spatiotemporal resolution by using “caged dimerizers”, which are activated or deactivated by light, as well as photoswitchable chemically induced dimerization (psCID). 22,23 1.3 Optogenetics Optogenetics is the use of light and genetically encoded light sensitive proteins to alter cellular behavior. The idea of using light to control biological systems was proposed by Francis Crick in his 1979 Scientific American article “Thinking about the brain” and later in his 1999 Kuffler Lectures at the University of San Diego. 24,25 Crick noted that a major challenge in neuroscience was the inability, at the time, to selectively activate specific neurons within the brain. In his 1999 lecture, Crick proposed that infrared light could be conceivably used to activate cell types engineered to be light-sensitive, but he noted that the idea itself was “far-fetched”. 9 By 2002, Zemelman and colleagues had reified Crick’s far-fetched idea when they demonstrated the use of light to selectively activate cultured genetically modified neurons. 26 Their approached relied on the co-expression of Drosophila photoreceptor genes encoding arrestin-2, rhodopsin (formed by liganding the photoreceptor protein opsin with the small molecule retinal), and the a subunit of the cognate heterotrimeric G protein. By expressing these three components inside of vertebrate neurons they were able to induce neuronal electrical activity using light. Zemelman and colleagues’ approach had relied on the ectopic expression of several components to induce light activated sensitivity. However, in the early 1970s microbiologists Walhter Stoeckenius and Dieter Oesterhelt discovered that Halobacterium halobium expressed a dual ion channel/photoreceptor encoded by a single gene. They named this microbial opsin protein bacteriorhodopsin, and analogues of bacteriorhodopsin were soon discovered throughout the microbial world. 27,28 However, neuroscientists resisted using these single component microbial opsins to induce light sensitivity in vertebrate neurons, largely because the idea seemed unlikely to work. In addition, all the microbial opsins discovered required the use of all-trans retinal as chemical co-factor to induce light sensitivity, thus they were not a purely single component optogenetic strategy. 29 The requirement of all-trans retinal notwithstanding, an important step in single component optogenetics also occurred in 2002, only a few months after Zemelman and colleagues’ paper. At the time, Peter Hegeman and Georg Nagel were studying the green alga Chlamydomonas reinhardtii, which moves in response to light. In their 2002 publication in Science, they reported that the light induced responses of the alga are mediated by an opsin-related protein. They termed 10 this protein Channelrhododopsin-1 (ChR1). 30 By expressing ChR1 inside Xenopus laevis oocytes, in the presence of all-trans retinal, they were able to change the membrane potential of the oocytes by light illumination. Based on these studies, they concluded that ChR1 must be also be a light sensitive ion channel (i.e. a photoreceptor that is also an ion channel). Although ChR1 required all-trans retinal, it could induce a light activated phenotype in vertebrate tissue. In 2005, Karl Deisseroth’s group at Stanford reported that another member of the Channelrhododopsin family, Channelrhodopsin-2 (ChR2), could induce rapid light activated electrical activity in cultured mammalian neurons. 31 Importantly, although ChR2 does require all- trans retinal to function properly, Deisseroth’s group found that Chr2 imparted light sensitivity to the neurons without the addition of all-trans retinal to the culture media. They concluded that the retinyl acetate in their culture media must have been sufficient. In 2009, Deisseroth’s group moved beyond cell culture when they successfully used ChR2 to optogenetically control the neurons of living mice. 32 Remarkably, they were able to condition mouse behavior through optogenetic stimulation of a subset of dopaminergic producing neurons within the mouse brain. In addition to activating neurons, optogenetics has been employed to modulate cell signaling. In 2005, Kim et. al described the creation of a chimeric rhodopsin, in which the cytoplasmic portion of the rhodopsin is replaced with the cytoplasmic portion of the b2- adrenergic receptor, a G-protein-coupled receptor (GPCR). They then demonstrated light-dependent activation of a b2-adrenergic receptor signaling pathway in cells transfected with the chimeric receptor. 33 11 In a 2009, Deisseroth’s group expanded on this work when they described the creation of chimeric genes encoding the cytoplasmic regions of different vertebrate G-protein-coupled receptors (GPCRs) fused to an extracellular domain of bovine rhodopsin. Deisseroth and colleagues termed these GPCR/rhodopsin chimeras optoXRs. 34 Similar to Kim et. al, they demonstrated that cells transfected with optoXRs and stimulated with light exhibit activation of their respective signaling pathways, comparable to that achieved by traditional pharmacological stimulation. However, they also demonstrated optogenetic control of specific signaling pathways within the neurons of optoXR transgenic mice, and by varying light stimulation of these neurons, they were able to achieve optical control of reward behavior in the mice. While microbial opsins and vertebrate rhodopsins have found broad use in optogenetics, particularly in neuroscience, plant and fungal derived light sensitive proteins have been widely applied in studies requiring light induced protein homodimerization, heterodimerization, and oligomerization. In 2002, Shimizu-Sato et al. successfully demonstrated light-switchable gene expression switch in yeast via the use of chimeric genes encoding the plant photoreceptor phytochrome bound to the Gal4 transcription factor. Under light stimulation, the fusion proteins heterodimerized to initiate gene transcription. 35 In 2012, Wang et al. employed a similar technique by encoding chimeric gene fusions of the light sensitive fungal protein VVD with the Gal4 transcription factor. Using this approach they were able to induce light stimulated gene expression via dimerization of the Gal4-VVD proteins in both cultured mammalian cells and in transgenic mice. 36 In addition to these highlighted works, 12 a variety of research groups have employed similar approaches, using plant phytochromes, to induce spatiotemporal control of transcription or cell signaling via protein heterodimerization. 37-40 Although protein homodimerization and heterodimerization are important modes of protein behavior, many events inside the cell are regulated by the oligomerization of higher-order structures. In 2013, Bugaj et al reported the optogenetic control of protein oligomerization. 41 Their approach relied on the use of the Arabidopsis thaliana cryptochrome 2 (Cry2), which oligomerizes in response to blue light. 42 Bugaj et al. cloned Cry2 to mCherry and demonstrated that the fluorescent Cry2-mCherry fusion proteins rapidly and reversibly oligomerized in mammalian cells in response to light exposure. Building upon this proof-of-principle experiment, they next determined whether Cry2-mCherry oligomerization could be employed to control intracellular signaling. To achieve this goal, they cloned Cry2-mCherry to a c-terminus fragment of the LRP6 receptor, which oligomerizes in response to Wnt ligand binding. LRP6 oligomerization induces canonical Wnt/b-catenin signaling, resulting in translocation of b-catenin to the nucleus, whereby it behaves as a transcription factor for specific genes. With the use of luciferse assays and western blots, Bugaj et al. successfully demonstrated that light stimulated Cry2-LRP6c fusion proteins activate Wnt/b-catenin signaling to levels comparable with pharmacological stimulation. Buraj and colleagues then demonstrated the modularity of this approach by cloning Cry2-mCherry to the RhoA GTPase, which regulates cytoskeleton dynamics. With this approach, they were able to show that light-induced clustering of Cry2-mCherry-RhoA produced cytoskeleton reorganization. 13 Oligomerization can also be used to sequester proteins away from their respective pathways, resulting in pathway inhibition. Utilizing this logic, Taslimi et al. demonstrated the Cry2-CLC fusion proteins could be oligomerized to reversibly inhibit Clathrin-mediated endocytosis. 43 Independent from this work, Lee et al. employed a Cry2 strategy titled LARIAT (Light-activated reversible inhibition by assembled trap) that also inhibits protein function by reversibly sequestering targets into oligomers. 44 In contrast to the work of Taslimi and colleagues, LARIAT doesn’t rely on the direct oligomerization of Cry2 fusion proteins. It consists of two modules: a multimeric protein (MP) and a Cry2 based heterodimerizer which binds to the MP in response to light stimulation, producing higher order structures. Cry2 based approaches for protein oligomerization have proven to be highly modular, and in the past several years a number of independent groups have created seemingly endless variations of Cry2 optogenetic tools to probe a variety of intracellular pathways. 1.4 Magnetic stimuli In contrast to optogenetics, which relies on naturally found light sensitive proteins to induce clustering, magnetic approaches for controlling protein assembly generally require the use synthetic magnetic nanoparticles (MNPs). MNPs are decorated with a functional protein or ligand which may be targeted for an extracellular receptor or an intracellular target. Protein assembly is then mediated by the application of a magnetic field which induces clustering of the functionalized MNPs. In 2008, Mannix and colleagues successfully demonstrated that MNPs decorated with monovalent ligands for individual IgE/FceR extracellular receptor complexes could activate 14 intracellular calcium signaling via magnetic field induced clustering of the MNP bound receptors. 45 In 2013, Bharde et al. adopted a similar approach to control EGFR signaling. 46 Cho and colleagues successfully extended this approach in vivo when they demonstrated magnetic field induced clustering of cell death receptors and subsequent apoptotic morphological changes in zebrafish embryos. 47 Because nanoparticles, including MNPs, can be internalized by cells magnetic switches have also enjoyed success in controlling intracellular targets. Using retinal ganglion cells, Steketee and colleagues were able to induce rapid endocytosis of superparamagnetic iron oxide nanoparticles (SPIONs) decorated with an agonist for the neurotrophic Tropomyosin receptor kinase B (TrkB). They then demonstrated that in the absence of a magnetic field, the endocytosed SPIONs were trafficked along the growth cone of the neurite where they induced axon growth. However, application of a magnetic field reversibly disrupted this trafficking and halted axon growth. 48 Although some MNPs can be internalized naturally, researchers have also relied on artificial means for internalization. For example, Etoc and colleagues employed microinjection of their TIAM-MNPs to control morphology in cultured NIH 3T3 and cos7 cells. 49 Hoffman et. al devised a cell free assay to study the potential of MNPs to control intracellular pathways. 50 Their approach employed Xenopus laevis egg extracts reconstituted with MNPs conjugated to RanGTP proteins. They demonstrated that in the presence of a magnetic field, these functionalized MNPs could induce microtubule fibers to assemble into asymmetric arrays of polarized fibers inside the reconstituted egg extracts. Additionally, the orientation of the fibers was dictated by the direction 15 of the magnetic force. Researchers have also begun to develop genetically encoded magnetically responsive proteins, opening the door for new strategies for magnetic-based control of intracellular protein assembly. 51 1.5 Thermally responsive proteins Thermally responsive proteins are a recent addition to the repertoire of existing strategies to control intracellular proteins and alter cell biology. In 2012, the MacKay group reported that thermally responsive elastin-like polypeptides (ELPs) reversibly assembly coacervates inside mammalian cells. 52 Inspired from human tropoelastin, ELPs are biomimetic polypeptides composed of a pentameric amino acid repeat of (VPGXG)n. The X residue can be selected to control ELP hydrophobicity, and the number of repeats, n, determines the ELP molecular weight. ELPs can rapidly phase-separate into protein-rich domains in response to heating, whereby they form a secondary aqueous phase known as a coacervate. In cells, individual coacervate droplets are typically several hundred nanometers to several micrometers in diameter. This phase separation is thermodynamically reversible, and ELP coacervates quickly resolubilize into bulk water upon cooling. The temperature of self-assembly, termed the transition temperature (Tt), is determined largely by the hydrophobicity of the guest residue and the ELP molecular weight. 53,54 By cloning ELPs to a functional protein, a temperature responsive fusion protein can be created that reversibly assembles in response to temperature changes. Below the Tt, the fusion protein is soluble and able to perform its wild-type duties. However, temperature increase above 16 the Tt induces rapid self-assembly. Depending on the context, this self-assembly can produce pathway activation or inactivation. For example, by creating ELP fusions of Clathrin light chain (CLC), the Mackay group was able to reversibly inhibit Clathrin mediated endocytosis (CME). 55 In this particular system, incubating cells above the Tt of the CLC-ELP fusion proteins induced their self-assembly and sequestration away from the CME pathway, thereby inhibiting it. In contrast, by creating ELP fusions of the epidermal growth factor receptor (EGFR), the MacKay group was able to reversibly activate the EGFR pathway. In this context, temperature triggered self-assembly of the EGFR-ELP proteins induced EGFR cross phosphorylation, receptor internalization, and phosphorylation of ERK1/2. 56 Recent studies in the MacKay group report that a similar approach can be used to control the enigmatic membrane protein CAV1 (Chapter 2). 1 In addition, ELPs have recently been applied for tunable assembly of protein microdomains within the single cells of living zebrafish embryos. 2 1.6 Conclusion Because so many cellular pathways are mediated by protein assembly, the ability to control cellular protein assembly is akin to creating an intracellular switch. To this end, the scientific field has exploited a variety of stimuli responsive proteins to achieve intracellular switching. Commonly used stimuli include chemicals, light, magnetism, and more recently, heat. Challenges with intracellular switching include the spatiotemporal control of the stimuli, tunability of the stimuli, delivery of the stimuli, and adaptability. Because each of the above described systems display pros and cons in regard to these factors, exciting advancements in the field include the combination of orthogonal stimuli, such as chemicals and light, to refine intracellular switching. 17 Chapter 2 Caveolin elastin-like polypeptide fusions mediate temperature-dependent assembly of caveolar microdomains 2.1 Abstract Caveolae are membrane organelles formed by submicron invaginations in the plasma membrane, and are involved in mechanosensing, cell signaling, and endocytosis. Although implicated broadly in physiology and pathophysiology, better tools are required to elucidate the precise role of caveolar processes through selective activation and inactivation of their trafficking. Our group recently reported that thermally-responsive elastin-like polypeptides (ELPs) can trigger formation of ‘genetically engineered protein microdomains (GEPMs)’ functionalized with either Clathrin-light chain or the epidermal growth factor receptor. In this chapter I show that this strategy can modulate caveolin-1 (CAV1). By attaching different ELP sequences to CAV1, mild heating can be used to self-assemble CAV1-ELP microdomains inside of cells. The temperature of self- assembly can be controlled by tuning the ELP sequence. The formation of CAV1-ELP microdomains internalizes Cholera Toxin Subunit B, a commonly used marker of caveolae mediated endocytosis. CAV1-ELPs also colocalize with Cavin 1, an essential component of functional caveolae biogenesis. With the emerging significance of caveolae in health and disease and the lack of specific probes to rapidly and reversibly affect caveolar function, CAV1-ELP microdomains are a new tool to rapidly probe caveolae associated processes in endocytosis, cell signaling, and mechanosensing. 18 2.2 Introduction The plasma membrane is a highly dynamic organelle comprised of domains that facilitate signaling, mechanosensing, and exchange of materials with the external environment. One of the most morphologically distinct of these domains is constituted by caveolae, 50-100 nm bulb-shaped pits in the plasma membrane. The integral membrane protein Caveolin 1 (CAV1) is the essential structural protein constituent of caveolae, with individual caveolae containing approximately 140 oligomerized CAV1 proteins. 57-59 More recently, the peripheral membrane protein Cavin 1 has been shown to be crucial for functional caveolae regulation and biogenesis. 60-62 Although CAV1 is an integral membrane protein, with its transmembrane domain completely buried in the lipid bilayer, both its N and C terminus face the cytosol. This unique molecular structure makes direct extracellular labeling or surface biotinylation of CAV1 impossible. 63 In addition, caveolar abundance varies widely across different cell types and tissue. In adipocytes, muscle cells, and endothelial cells, caveolae can account for up to 50% of the plasma membrane surface area. In contrast, caveolae are present at low levels in the liver and apparently not at all in kidney proximal tubules. 58,59,64 Furthermore, caveolae exhibit significant versatility in the structures they form, from individual caveolae to rosettes that extend deep into the cytoplasm. 65,66 The considerable heterogeneity of caveolar abundance and structural versatility, combined with the lack of a direct extracellular domain for labeling, has made it difficult to discern the functions that caveolae play in health and disease. However, it is now clear that caveolae perform roles in signaling, plasma membrane dynamics, and trafficking. They regulate signal transduction 19 by organizing signaling partners into discrete regions near the plasma membrane. 67-69 Not unlike the suspension system on an automobile, they can absorb mechanical or osmotic strain by flattening or assembling. 58,65,70-72 Perhaps most famously, caveolae can pinch off from the plasma membrane to endocytose and traffic cargo in a process termed caveolae-mediated endocytosis (CAVME). 73-75 ELPs are biocompatible, biodegradable polypeptides composed of a pentameric repeat of (VPGXG)n. The X residue can be selected to control ELP hydrophobicity, and the number of repeats, n, determines the ELP molecular weight. ELPs can rapidly phase-separate into protein- rich domains in response to heating, whereby they form a secondary aqueous phase known as a coacervate. In cells, individual coacervate droplets are typically several hundred nanometers to several micrometers in diameter. This phase separation is thermodynamically-reversible, and ELP coacervates quickly resolubilize into bulk water upon cooling. The temperature of self-assembly, termed the transition temperature (Tt), is determined largely by the hydrophobicity of the guest residue and the ELP molecular weight. 53,54 Our group previously reported that ELPs can be fused to effector proteins to reversibly assemble cytosolic fluorescent microdomains both in live cells and in the single cells of living zebrafish embryos. 2,52 In addition, we have cloned ELPs with Clathrin Light Chain (CLC), and used temperature mediated self-assembly of the resulting CLC-ELPs to reversibly inhibit Clathrin- mediated endocytosis. 55 Most recently, we reported that fusion of an ELP to the transmembrane receptor tyrosine kinase known as epidermal growth factor receptor 1 promoted temperature- dependent receptor internalization and phosphorylation of ERK1/2. 56 20 In addition to work performed by our group, Ge et al. reported that recombinant ELP-GFP proteins can form microcompartments inside E. Coli or tobacco cells. 76 More recently, Jang et al. have reported that recombinant ELPs can be used for the tunable creation of protein vesicles containing functional, globular domains. 77 Relatedly, Voegle et al. have reported that elastin-like peptides can create self-assembled peptide vesicles. Furthermore, the growth of these vesicles can be genetically encoded by encapsulating a cell-free transcription-translation system together with the peptide DNA template inside the vesicles. 78 Such studies demonstrate how ELPs and related materials can be used to probe and to model important biological questions in a surprising variety of systems. In this chapter I report that ELPs can be cloned to the C-terminus of CAV1 to create thermally-responsive CAV1-ELP fusion proteins. By changing ELP length and sequence, a library of CAV1-ELPs were created that display different transition temperatures. To determine the precise transition temperature of these CAV1-ELPs in live cells, I refined a novel technique developed in our lab to visualize and quantify their assembly through coassembly with coexpressed fluorescent ELPs. 56 I found that when CAV1-ELPs are incubated above their Tt, they self-assemble intracellular microdomains which partially overlap with internalized Cholera Toxin Subunit B (CTXB), a commonly used marker of caveolae-mediated endocytosis. In addition, CAV1-ELPs display broad colocalization with Cavin1 at all temperatures, which is consistent with their cellular biology. 21 2.3 Materials and methods 2.3.1 Cell Culture HEK293T cells (#CRL-3216, ATCC, Manassas, VA) were cultured in DMEM (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% FBS (#35-011-CV, Corning, NY) in a humidified incubator with 5% CO2 at 37 °C. Cells were transfected with Lipofectamine 3000 (L3000008, Life Technologies, Carlsbad, CA), using the manufacturer’s protocol, with the exception that after applying the transfection mixture to each well, cells were placed in a 30˚C incubator with 5% CO2 and were incubated for 48 hrs. before subsequent experiments. This lower temperature was chosen to inhibit premature ELP self-assembly as the exogenous proteins are expressed during the transfection incubation period. 2.3.2 Plasmid Construction and Amino Acid Sequences ELP expression vectors were synthesized using recursive directional ligation in pET25b (+) plasmid (Millipore Sigma, Burlington, MA), as previously reported. 79 CAV1-ELPs and CAV1-WT were constructed from human Caveolin 1 in pCMV6 plasmid (Origene, Cat # RC210274, Rockville, MD) which was modified by removing bases 1623-1670 through digestion with EcoRV and SacII followed by insertion of DNA nucleotides with EcoRV/Hind III digestion sites (Top strand: 5’ - gcGATATCCgccgAAGCTTatcgc - 3’; Bottom strand: 5’- gcgatAAGCTTcggcGGATATCgca - 3’). Finally, ELP sequences were inserted into this modified plasmid by digesting it with EcoRV/ HindIII and by digesting the pET25b (ELP) plasmid with MslI/HindIII. All sequences were verified by diagnostic digests and DNA sequencing (Retrogen, San Diego, CA). GFP-V60 was cloned as previously described. 52 Amino acid sequences of the open reading frames (ORF) of each construct are shown below: 22 CAV1-WT: MSGGKYVDSEGHLYTVPIREQGNIYKPNNKAMADELSEKQVYDAHTKEIDLVNRDPKH LNDDVVKIDFEDVIAEPEGTHSFDGIWKASFTTFTVTKYWFYRLLSALFGIPMALIWGIY FAILSFLHIWAVVPCIKSFLIEIQCISRVYSIYVHTVCDPLFEAVGKIFSNVRINLQKEITRTR PLEQKLISEEDLAANDAISAEACGHSCFLNRSRVASL* CAV1-V72: MSGGKYVDSEGHLYTVPIREQGNIYKPNNKAMADELSEKQVYDAHTKEIDLVNRDPKH LNDDVVKIDFEDVIAEPEGTHSFDGIWKASFTTFTVTKYWFYRLLSALFGIPMALIWGIY FAILSFLHIWAVVPCIKSFLIEIQCISRVYSIYVHTVCDPLFEAVGKIFSNVRINLQKEITRTR PLEQKLISEEDLAANDAMGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPG VGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGV GVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVG VPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGV PGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVP GVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPG VGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVG* CAV1-V96: MSGGKYVDSEGHLYTVPIREQGNIYKPNNKAMADELSEKQVYDAHTKEIDLVNRDPKH LNDDVVKIDFEDVIAEPEGTHSFDGIWKASFTTFTVTKYWFYRLLSALFGIPMALIWGIY FAILSFLHIWAVVPCIKSFLIEIQCISRVYSIYVHTVCDPLFEAVGKIFSNVRINLQKEITRTR PLEQKLISEEDLAANDAMGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPG 23 VGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGV GVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVG VPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGV PGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVP GVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPG VGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGV GVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVG VPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVG* CAV1-A96: MSGGKYVDSEGHLYTVPIREQGNIYKPNNKAMADELSEKQVYDAHTKEIDLVNRDPKH LNDDVVKIDFEDVIAEPEGTHSFDGIWKASFTTFTVTKYWFYRLLSALFGIPMALIWGIY FAILSFLHIWAVVPCIKSFLIEIQCISRVYSIYVHTVCDPLFEAVGKIFSNVRINLQKEITRTR PLEQKLISEEDLAANDAMGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPG AGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGA GVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAG VPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGV PGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVP GAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPG AGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGA GVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAG VPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGVPGAGY* 24 GFP-V60: MASKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDATYGKLTLKFICTTGKLPVPWP TLVTTFSYGVQCFSRYPDHMKRHDFFKSAMPEGYVQERTISFKDDGNYKTRAEVKFEG DTLVNRIELKGIDFKEDGNILGHKLEYNYNSHNVYITADKQKNGIKANFKIRHNIEDGSV QLADHYQQNTPIGDGPVLLPDNHYLSTQSALSKDPNEKRDHMVLLEFVTAAGITHGMD ELYKSGSGPVLAIGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPG VGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGV GVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVG VPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGV PGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVP GVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPG VGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGV GVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVG VPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVGVPGVG* 2.3.3 Western Blot HEK293T cells were grown in 6 well plates and transfected as described above. 48 hrs. after transfection, cells were washed 1X with ice-cold DPBS and were placed at 4˚C for 1 hr. to completely resolubilize CAV1-ELPs. Cells were then lysed in CHAPS lysis media (30 mM CHAPS, 1% SDS, 20 mM Tris-HCl, 150 mM NaCl, pH=8.0). Cell lysates were left at 4˚C with constant agitation for 2 hrs., followed by sonication in ice cold water for 5 min. Lysate was then spun down at 15500 rpm for 10 min at 4˚C. Supernatant was removed and proteins were separated by SDS-page using PAGEr EX 4–12% gradient gel (#59722, Lonza, Morristown, NJ) and then wet-transferred overnight in a 4˚C fridge onto a 0.45 μm PVDF membrane. The membrane was 25 blocked with fluorescent blocking buffer (#MB-070, Rockland, Limerick, PA) for 1 hr. at room temperature. After blocking, the membranes were shaken overnight with primary antibodies to CAV1 rabbit mAb (1:1000 dilution in blocking buffer, #3267, Cell Signaling Technology, Danvers, MA) and GAPDH rabbit mAb (1:1000 dilution in blocking buffer, #5174, Cell Signaling Technology, Danvers, MA) at 4˚C. The next day, the membranes were washed 3 times, each 5 min, with 0.2% Tween containing Tris-buffered saline (#T9039, Sigma-Aldrich, St. Louis, MO). The membranes were than incubated in secondary antibody IRDye 680RD donkey anti-rabbit IgG (1:5000 dilution in blocking buffer, 926-68073, LI-COR Biotechnology, Lincoln, NE) at room temperature for 1 hr. After washing with 0.2% Tween containing Tris-buffered saline again, membranes were imaged with Odyssey Licor imaging system (LI-COR Biotechnology, Lincoln, NE). 2.3.4 Live Cell Imaging and immunofluorescence Prepared HEK293T cells were grown on 35 mm glass bottom dishes (MatTek Corporation, Ashland, MA) coated with Poly-D-Lysine (P6407, Sigma-Aldrich, St. Louis, MO) and transfected as described above. A Nikon DIAPHOT 300 epifluorescent microscope equipped with a 10X 0.25 numerical aperture objective, X-Cite 120 LED light source (Lumen Dynamics, Ontario Canada), and DS digital camera (Nikon Instruments, Melville, NY) was used to examine GFP-V60 self- assembly vs. GFP-V60 coassembly with CAV1-V96 or CAV1-V72. During imaging, the temperature was increased from 4°C to 45°C at a rate of 1°C/min by using an Instec HCS60 stage attached to an Instec mK1000 temperature controller (Denver, CO). The actual temperature inside the medium of the cell culture dish was monitored using a thermometer coupled with a probe from Sper Scientific (Scottsdale, AZ). Image stacks of each temperature ramp run were then blinded 26 using a python script I wrote which is available from github: http://doi.org/10.5281/zenodo.3247836. 80 Cells were scored by eye to determine the frame/temperature at which they produced microdomains. GFP-V60 cells which did not produce microdomains by the end of the temperature were assigned a Tt of 50˚C. FIJI (version 2.0.0) was used to determine average microdomain size and number within each cell. 81 To confirm CAV1- ELP colocalization with GFP-V60 using immunofluorescence, CAV1-ELPs were probed with myc-tag mouse mAb (9B11 #2276S, Cell Signaling Technology, Danvers, MA) and Alexa Fluor 488 goat anti-mouse IgG (A-11001, ThermoFisher Scientific, Waltham, MA) while GFP-V60 was probed with rabbit GFP pAb antibody (ab290, Abcam, Cambridge, MA) and Alexa Fluor 647 chicken anti-rabbit IgG (A-21443, ThermoFisher Scientific, Waltham, MA). In all immunofluorescent (IF) experiments, blocking was performed with a 90-minute room temperature incubation in 1% BSA (A9647, Sigma-Aldrich, St. Louis, MO). Antibodies were also diluted in this solution. “No primary-antibody controls”, consisting of labeling with secondary antibody without primary antibody incubation, were used in all IF experiments to determine non-specific binding and imaging settings. 2.3.5 CAV1-ELP/CAV1-WT colocalization with Cholera Toxin Subunit B (CXTB) HEK293T cells were plated on glass coverslips coated with Poly-D-Lysine (P6407, Sigma- Aldrich, St. Louis, MO). The next day, cells were transfected with either CAV1V96, CAV1A96, or CAV1WT as described above. 48 hrs after transfection, cells were taken out of the incubator and were incubated at 4˚C for 45 min to resolubilize CAV1-ELPs and inhibit endogenous 27 endocytosis. Cells were then washed 1X with ice cold DPBS before labeling with Alexa Fluor 555 conjugated Cholera Toxin Subunit B (CXTB-AF555) using the Vybrant™ Alexa Fluor™ 555 Lipid Raft Labeling Kit (V34404, Life Technologies, Carlsbad, CA) according to the manufacturer’s protocol. Note that for this study we did use the kit’s rabbit anti-Cholera Toxin antibody. However, I found this antibody to be unnecessary for proper lipid raft labeling and did not use it in other studies involving lipid raft labeling. After lipid raft labeling, cells were then incubated for an additional 45 min at 4˚C and were then either incubated at 37˚C or 4˚C for an additional 50 min. Cells were then immediately fixed and immunostained for CAV1-ELP or CAV1-WT using myc-tag mouse mAb (9B11 #2276S, Cell Signaling Technology, Danvers, MA) and Alexa Fluor 488 goat anti-mouse IgG (A-11001, ThermoFisher Scientific, Waltham, MA). Cells were incubated with DAPI prior to mounting. Airy scan confocal images were captured on a Zeiss LSM 800 with a 1.4 numerical aperture 63X objective. DAPI, CAV1-ELPs, and CTXB- AF555 were excited with the 405, 488, or 561 laser line, respectively. To ensure unbiased results during imaging, CAV1-ELP transfected cells were randomly selected without observing the CTXB-AF555 channel. Pixel based colocalization was measured using Zen Blue Software. After selecting CAV1- ELP/CAV1-WT transfected cells, a Costes automatic threshold was applied to determine the appropriate thresholds in both channels. 82 Since the transfected CAV1-ELP/CAV1-WT proteins were stained with a green protein and the surface labeled Cholera toxin was conjugated to a red protein, a Mander's Colocalization Coefficient (MCC) for the green channel was calculated using the following Equation (2.1) ! !"##$ = ∑ ! ' !, $%&%$'& ∑ ! ' ! (2.1) 28 where Gi,colocal = Gi if Ri > 0 and Gi,colocal = 0 if Ri = 0. Ri and Gi are the intensity values of the red and green fluorophores in individual pixels above the Costes’ threshold. 2.3.6 Cavin-1 Immunofluorescence Mouse Cavin-1 was a gift from Paul Pilch (Addgene plasmid # 5044). HEK293T cells were plated on glass coverslips coated as described above. The next day cells were co-transfected with either CAV1-V96, CAV-1A96, CAV1-WT and also mouse Cavin-1. Lipid rafts were labeled with CXTB-AF555 as described above. After labeling and incubation at 4˚C for 45 minutes, cells were then incubated at either 4˚C or 37˚C for 50 minutes before fixation. Transfected CAV1- ELP/CAV1-WT was detected using mouse anti-myc (Cell Signaling Technology) and Alexa Fluor 488 goat anti-mouse (ThermoFisher Scientific). Transfected mouse Cavin-1 was detected using Rabbit anti-HA antibody (C29F4 #3724, Cell Signaling Technology) and Alexa Fluor 647 chicken anti-rabbit antibody (A-21443, ThermoFisher Scientific). Cells were incubated with DAPI prior to mounting. Airy scan confocal images were captured on a Zeiss LSM 800 with a 1.4 numerical aperture 63X objective. DAPI, CAV1-ELPs, CTXB-AF555, and Cavin 1 were excited with the 405, 488, 561, or 640 laser line, respectively. 2.3.7 CTXB Binding Quantification To quantify total cellular CTXB-AF555 binding after CAV1-V96 self-assembly (Figure 2.5), HEK293T cells were transfected as described above. Before CTXB-AF555 labeling, transfected cells were placed at 4 ˚C for 45 min to resolubilize CAV1-ELPs and inhibit endogenous 29 endocytosis. Cells were then either left at 4˚C or incubated at 37˚C for 50 min to induce CAV1- V96 self-assembly. Cells were then washed 1X with ice cold DPBS and were labeled with CTXB- AF555, washed 3X with ice cold DPBS, and were then immediately fixed and immunostained for CAV1-ELP as described above. To ensure unbiased results during imaging, CAV1-ELP transfected cells were randomly selected without observing the CTXB-AF555 channel. To further ensure unbiased imaging, the DAPI channel was used to determine the top and bottom of the z- stack for each cell. Cells were imaged on a Zeiss LSM880 using Airy Scan Fast with a 1.4 numerical aperture 63X objective. DAPI, CAV1-ELPs, and CTXB-AF555 were excited with the 405, 488, or 561 laser line, respectively. During image acquisition, a lateral pixel size of 40nm by 40nm was used. To compensate for cell to cell variability in axial width, we chose to slightly oversample the axial direction by collecting sixty focal planes/slices for each image, which ensured at least Nyquist sampling in the axial direction for both CAV1-ELP (488nm) and CTXB-AF555 (561nm laser line) channels regardless of cell width. Note that on average, we found cells to have an axial width of 9 µm. Based on the calibrated software of the microscope and the advice of the manufacturer, our conservative estimate of the axial resolution for the microscope when using the 561nm laser line is 450nm. However, we note that exact values for resolution must be empirically determined by imaging the PSF of sub-resolution fluorescent beads. To quantify total cellular CTXB-AF555 fluorescence from each z-stack, I developed an open source image processing pipeline which adapts a commonly used method for measuring cellular fluorescence from microscopy images. 83-86 The code and detailed instructions for this pipeline are available from github: http://doi.org/10.5281/zenodo.3247836. 80 Briefly, this pipeline calculates background corrected integrated densities of CTXB fluorescence from sum projections of each z-stack. To account for heteroscedasticity, skew, and non-normality in this data, the measurements were then 30 log-transformed and point estimates and 95% confidence intervals for 4°C vs. 37°C treatments were obtained. The anti-log of these estimates was then taken to convert them onto the ratio scale, using equations 2.2 and 2.3 shown below: 87,88 log () (()*+ * )− log () (()*+ +, )= log () . -./0 ( -./0 )* / (2.2) 10 log 10 1 +,-. ( +,-. )* 2 = -./0 ( -./0 )* (2.3) where CTXB4 and CTXB37 are the background corrected integrated density of CTXB fluorescence at 4°C and 37°C, respectively. 2.3.8 Statistics All plots and statistics were calculated using R. 89 The R package nlme was used to calculate statistics for the CTXB binding in Figure 2.5. 90,91 To determine the appropriate statistical analysis for each experiment and if transformations of the raw data were warranted, all datasets were examined with diagnostic plots to visually examine outliers, non-normality, heteroscedasticity, or pronounced patterns in the residuals. All statistical tests were two-sided and an alpha of 0.05 was chosen as the significance criteria. In Figure 2.2E, Boxplots whiskers extend to ± 1.5X the interquartile range or to the max/min data point, whichever is closest. This is the default in R. 2.4 Results 2.4.1 Construction and confirmation of a library of CAV1-ELP fusions To determine whether temperature mediated ELP self-assembly could produce caveolar microdomains, a library of different CAV1-ELPs was first cloned (Figure 2.1A, Table 2.1), and their identity was confirmed using Western-blotting (Figure 2.1B). In this western blot, smaller bands are seen which are consistent with blots for detergent resistant membrane proteins and with the degradation reported for other CAV1 fusion proteins. 92 By changing ELP length and sequence, CAV1-ELPs were designed with transition temperatures above and below a physiological 31 temperature of 37 ºC. Based on previous studies, three ELP sequences were chosen: V96, V72, and A96. CAV1-V96 was intended to display near maximal self-assembly at 37 °C. Purified A96 displays a Tt above 80˚C, thus CAV1-A96 is intended as a negative control (i.e. a CAV1-ELP with no assembly at physiological temperatures). Based on our previous work with membrane bound EGFR, it was estimated that CAV1-V72 would self-assemble within 5°C above the transition temperature of CAV1-V96. 56 As a control for overexpressed CAV1, wild-type CAV1 with a C- terminus myc epitope was also developed. Table 2.1. Nomenclature, amino acid sequence, and phase behavior of expressed proteins. Not applicable (NA). CAV1-A96 does not self-assemble at physiological temperatures. a) ORF amino acid sequences are found in the Methods section. b) myc-epitope amino acid sequence: EQKLISEEDL c) Estimated molecular weight from open reading frame and confirmed by western blot (Figure 2.1B). d) Median T t obtained from live cell imaging technique described in Figure 2.2. e) Mean T t (95% Confidence Interval) obtained from live cell imaging technique described in Figure 2.2. f) The percentage of cells with a T t ≤ 37°C. Percentages calculated from data shown in Figure 2.2E. Protein Label Amino acid Sequence a MW c [kD] Median T t in cells d [°C] Mean T t in cells e [°C] Percent of cells showing assembly f [%] Intended behavior CAV1- V96 CAV1-myc b - (VPGVG)96Y 62.5 36.3 35.9 (34.2, 37.7) 67 temperature- responsive CAV1- V72 CAV1-myc- (VPGVG)72Y 52.7 38.5 36.5 (34.8, 38.2) 43 temperature- responsive CAV1- A96 CAV1-myc- (VPGAG)96Y 59.8 NA NA NA temperature- insensitive CAV1- WT CAV1-myc 24.9 NA NA NA temperature- insensitive GFP-V60 GFP-(VPGVG)60Y 67.2 42 42.5 (41.6, 43.5) 6.4 live-cell imaging tool 32 Figure 2.1. Construction and confirmation of a library of CAV1-ELP fusions. A) Genes encoding three different ELP sequences (V72, V96, A96) were cloned to Caveolin 1. A myc-tag was placed in-between CAV1 and ELP sequences for convenient immunofluorescent detection. Each ELP sequence confers a unique transition temperature (Tt). CAV1 sequence with myc-tag and no ELP sequence (bottom row) is referred to as CAV1-WT. B) CAV1-ELPs were expressed in HEK 293T cells and their size was confirmed using a caveolin 1 antibody. GAPDH is displayed at the bottom. NT = No transfection. The entire gel picture for the GAPDH protein loading control can be found in the supporting information (Supplementary Figure S1) of my publication. 1 33 2.4.2 CAV1-ELP Transition Temperature (Tt) can be visually determined in live cells Because CAV1-ELPs are not fluorescent, their intracellular Tt cannot be directly visualized without fixation and indirect immunofluorescence. To circumvent this limitation, I refined a live cell imaging technique which I previously applied to visualize the self-assembly of EGFR-ELP microdomains. 56 This technique exploits the co-assembly of non-fluorescent ELP fusion proteins with fluorescent GFP-V60 to visualize ELP microdomain formation in live cells (Figure 2.2). This fluorescent fusion reporter protein consists of the ELP V60 fused to the C-terminus of Green fluorescent Protein (GFP). GFP-V60 typically undergoes phase separation above physiological temperature. Below the intracellular Tt of GFP-V60, this protein is therefore soluble and GFP fluorescence is detected homogenously throughout the cytoplasm. However, as temperature is increased above the Tt of GFP-V60, ELP mediated self-assembly produces fluorescent GFP microdomains. With the aid of an epifluorescent microscope, these fluorescent microdomains are readily visible. Accordingly, by tracking live cells undergoing a temperature ramp, a blinded observer can visually determine the temperature at which microdomain assembly occurs for individual cells. This temperature is defined as the intracellular Tt. Critically, when cells are dually transfected with either CAV1-V96 or CAV1-V72 in addition to GFP-V60, microdomain assembly occurs at a lower temperature in comparison to cells transfected solely with GFP-V60. In addition, co-transfected cells display GFP microdomains that are fewer in number and larger relative to those produced by GFP-V60 alone, which allows the distinction of singly- transfected cells from dually- transfected cells. Individual cells transfected with only GFP-V60 were evaluated for their microdomain Tt, microdomain number per cell, and average area per microdomain (Figure 2.2B-E). Using this technique, cells transfected with GFP- 34 V60 alone displayed a median intracellular Tt of 42°C. In contrast, cells transfected with both GFP- V60 and either CAV1-V96 or CAV1-V72 exhibit a significant reduction of intracellular Tt to 36.3°C or 38.5°C, respectively. (Figure 2.2 E, Table 2.1). Although the CAV1-V72 + GFP-V60 group did exhibit a trend towards a higher intracellular Tt compared to CAV1-V96 + GFP-V60, the difference between the groups was not statistically significant (p-value = 0.85). In contrast, our previous work did find differences in Tt between V72 and V96 fusion proteins of GFP, Clathrin light Chain, and EGFR. 2,55,56 This result highlights how the cellular location (e.g. membrane vs. cytosolic) and idiosyncrasies of the wild-type protein affect the transition temperature of its ELP fusion. In future work, experimenters may further fine tune the CAV1-ELP Tt by selecting pentameric repeats and guest residues beyond what we have examined. As there was no significant difference between CAV1-V96 and CAV1-V72, and because CAV1-V96 had a median Tt below 37°C, all subsequent characterization was performed using CAV1-V96. 35 Figure 2.2. CAV1-ELP Transition Temperature (Tt) can be visually determined in live cells. A) Schematic of the live cell imaging technique which was used to determine CAV1-ELP T t. B) Live cells transfected with either GFP-V60 or GFP-V60 + CAV1-V96 and subjected to a temperature ramp to assemble distinct microdomains (also termed coacervates) at different temperatures. Images were acquired with epifluorescence microscopy. Scale bar is 10 µm C) Indirect immunofluorescence using Airy scan confocal microscopy in fixed cells confirms that CAV1-V96 (myc antibody) co-assembles with GFP-V60 (GFP antibody) at 37°C. These CAV1-V96 + GFP-V60 microdomains also colocalize with internalized fluorescent cholera toxin subunit B (CTXB-AF555). Scale bar is 5 µm. D) Cells dually transfected with GFP-V60 + CAV1-V96 or CAV1-V72 + GFP-V60 display microdomains which are larger and fewer in number compared to cells transfected with GFP-V60 alone. Each circle is data for one cell above its T t. Horizontal dashed line is at 23 µm 2 , the max size recorded among all GFP-V60 only transfected cells. This line served as the filter for distinguishing true dual transfected cells. F) Boxplots of T t between different transfections after filtering shown in panel D. Each dot is the T t for a single cell. P-values calculated from a two-sided Tukey’s HSD post hoc-test following a statistically significant one-way ANOVA.***p-value < 0.001. 36 2.4.3 CAV1-V96 microdomains colocalize with CTXB Using a confocal laser scanning microscope equipped with an Airy scan super-resolution detector and an indirect immunofluorescence approach, we next determined whether CAV1-V96 self-assembly would disrupt internalization of cholera toxin subunit B (CTXB), a commonly used marker of caveolar lipid rafts and caveolar-mediated endocytosis. 93,94 HEK293T cells were transfected with either CAV1-V96, CAV1-A96, or CAV1-WT. After an initial incubation at 4°C, which was necessary to resolubilize any already assembled CAV1-V96 and also to broadly inhibit ongoing caveolar endocytosis, cells were surface-labeled with fluorescent CTXB at 4°C, washed, and then incubated at 37°C or kept at 4°C for 50 minutes. Surprisingly, at 37°C CAV1-V96 self-assembled into microdomains which directly overlapped with CTXB (Figures 2.1C, 2.3, 2.4). The formation of these microdomains produced a statistically significant increase in colocalization with CTXB at 37°C vs. 4°C (p-value = 0.0014). In contrast, CAV1-WT and temperature insensitive CAV1-A96 did not assemble microdomains and did not exhibit a statistically significant increase in colocalization with CTXB at 37°C compared to 4°C. These results suggest that CAV1-V96 self-assembly internalizes CTXB in a mechanism perhaps like caveolae-mediated endocytosis. To further explore if CAV1-V96 self-assembly was responsible for internalization of CTXB, 60 slice z-stacks were collected in fixed cells using indirect immunofluorescence. Representative images are shown in Figure 2.4, where the axial views reveal that CAV1-V96 microdomains overlap with CTXB detected inside the cell, in agreement with CAV1-V96 self- assembly internalizing CTXB. 37 Figure 2.3. CAV1-V96 microdomains colocalize with CTXB. A) HEK293T cells transfected with indicated constructs were labeled with CTXB-AF555 at 4°C and then placed at the indicated temperatures for 50 min before fixation and immunostaining via myc epitope. Airy scan confocal microscopy reveals that temperature sensitive CAV1-V96 self-assembles into microdomains which colocalize with fluorescent CTXB (2 nd from bottom row and corresponding magnified panel). This colocalization was not present in cells transfected with CAV1-WT or temperature insensitive CAV1-A96. Scale bar is 5 µm. Images shown are representative from dozens of images acquired within each treatment group. B) CAV1-V96 colocalization with fluorescent CTXB is temperature dependent. Dozens of transfected cells in each group were imaged to quantify colocalization of CAV1-ELP/CAV1-WT with CTXB in response to temperature increase from 4°C to 37°C. Each diamond represents one magnified image containing 1-3 transfected cells. Error bars are standard deviation. Pairwise p-values were calculated from a two-sided Tukey’s HSD post hoc-test following a statistically significant one-way ANOVA. * p-value < 0.05, ** p-value < 0.01, ***p- value < 0.001 38 Figure 2.4. Temperature mediated self-assembly of CAV1-V96 induces internalization of fluorescent Cholera Toxin Subunit B (CTXB). Z-stacks from Airy scan confocal microscopy reveal that CAV1-V96 self-assembly internalizes surface labeled CTXB (top row). CAV1-ELP/CAV1-WT signal is from indirect immunofluorescence via myc tag. Lateral scale bar is 5 µm. Axial scale bars are 2 µm. Images shown are representative from dozens of images. 39 2.4.4 Inducing CAV1-V96 microdomain formation before CTXB labeling reduces CTXB signal The above results suggest that CAV1-V96 self-assembly does internalize CTXB, but the specificity of microdomain assembly to internalize caveolar lipid rafts remained unknown. Since CTXB binds to the ganglioside GM1, which is preferentially found in caveolar and non-caveolar lipid rafts, I hypothesized that if CAV1-V96 self-assembly was responsible for internalization of caveolar lipid rafts, then cells transfected with CAV1-V96 and induced to self-assemble before CTXB labeling would display a fractional decrease in CTXB binding. This decrease would be due to CAV1-V96 meditated internalization of caveolar lipid rafts and subsequent reduction of extracellular GM1 available for CTXB binding. To answer this hypothesis, HEK293T cells were transfected with either CAV1-V96 or temperature insensitive CAV1-A96. After an initial incubation at 4°C to resolubilize CAV1-ELPs, cells were incubated at either 4°C or 37°C for 50 minutes and then labeled with fluorescent CTXB- AF555 at 4°C before fixation and analysis of indirect immunofluorescence to detect CAV1-ELPs. I then collected 60 slice Airy scan confocal z-stacks of over 100 cells and measured the CTXB- AF555 fluorescence from these z-stacks (Figure 2.5). Figure 2.5B displays 95% confidence intervals for the ratio of fluorescent CTXB signal at 4°C vs 37°C within each transfection. Indeed, CAV1-V96 displayed a statistically- significant temperature dependent reduction of CTXB signal (p = 0.00675). Figure 2.5B reveals that for CAV1-V96, a pre-incubation at 4°C vs 37°C leads to a an approximate 1.8-fold change in CTXB signal/binding (i.e. equivalently, for CAV1-V96 transfected cells pre-incubation at 37°C leads to a ~44.4% decrease in CTXB binding compared to a 4°C pre-incubation). In contrast, temperature insensitive CAV1-A96 does not display a statistically significant temperature dependent difference in CTXB binding, as its 95% confidence interval contains a ratio of 1, the 40 null value. This result occurs because CAV1-A96 does not self-assemble microdomains and therefore does not internalize caveolar GM1. Representative cells from each transfection and temperature are shown in Figure 2.5A. The cells chosen are from images at the approximate 50% percentile of CTXB signal from within each treatment. The CAV1-V96 37°C cell (second row) displays a staining pattern consistent with CAV1-V96 microdomain formation, however there is limited overlap with CTXB. This result is consistent with CAV1-V96 self-assembly internalizing caveolar GM1, thus making these regions unavailable for subsequent CTXB binding. However, while pre-incubating CAV1-V96 transfected cells at 37°C fractionally decreases CTXB binding, it does not abolish it. This result is consistent with reports in the literature of CTXB entry into cells via non-caveolar routes, the fact that caveolar GM1 represents a fraction of the GM1 pool found on the plasma membrane, and my observation that non-transfected HEK293T cells, which are CAV1 negative, display robust CTXB binding. 93- 95 41 Figure 2.5. CAV1-V96 self-assembly removes a fraction of CTXB binding sites from the plasma membrane. A) HEK293T cells transfected with CAV1-V96 or CAV1-A96 were incubated at 4°C or 37°C for 50 minutes and were then labeled with CTXB-AF555, fixed, and immunostained for CAV1-ELPs via myc epitope. Z-stacks were then collected and fluorescent CTXB signal was quantified. Cells shown are from images at the approximate 50% percentile of fluorescent CTXB signal from within each treatment. The CAV1-V96 37°C cell (second row) displays likely microdomain formation but without overlapping CTXB signal, consistent with CAV1-V96 self-assembly internalizing caveolar GM1, thus making these regions unavailable for subsequent CTXB binding. Scale bar is 5 µm. B) 95% confidence intervals for the ratio of fluorescent CTXB signal after a 4°C vs. 37°C pre-incubation. Dashed vertical line is at 1, the null value (i.e. no difference in signal). Corresponding p-values are also shown. For CAV1-V96, pre-incubation at 4°C vs. 37°C leads to a ~ 1.8-fold increase in CTXB binding (i.e. pre-incubation at 37°C leads to a ~44.4% decrease in CTXB binding compared to a 4°C pre-incubation). Twenty-seven (n = 27) z-stack images were acquired within each temperature/transfection combination, totaling 108 quantified images. 42 2.4.5 Cavin 1 exhibits broad colocalization with CAV1, regardless of ELP sequence or temperature Since Cavin 1 is required for functional caveolar biogenesis and because it colocalizes with wild-type CAV1 near the plasma membrane, it was next determined whether CAV1-ELPs would colocalize with Cavin 1 and whether this colocalization was temperature-dependent. 96 Accordingly, HEK293T cells were transfected with Cavin 1, in addition to either CAV1-V96, CAV1-A96, or CAV1-WT (wild type CAV1 with a C-terminal myc tag). The representative images shown in Figure 2.6 reveal broad overlap between transfected Cavin 1 and CAV1, regardless of ELP sequence or temperature. These results are consistent with CAV1-ELPs retaining partial wild-type CAV1 functionality and plasma membrane localization. 43 Figure 2.6. Cavin 1 exhibits broad colocalization with CAV1, regardless of ELP sequence or temperature. HEK293T cells co-transfected with Cavin 1 and either CAV1-V96, CAV1-A96, or CAV1- WT were labeled with CTXB-AF555 and then incubated at either 4°C or 37°C for 50 minutes before fixation and indirect immunofluorescence. Broad colocalization between Cavin 1 and CAV1-ELPs/WT is evident in all treatments. Note the CAV1-V96 cell at 37°C (top row) displays stereotypical microdomain formation with CTXB, which also colocalizes with Cavin 1. Transfected Cavin 1 was detected with an HA antibody for C-terminus HA epitope. CAV1 constructs were detected with a myc antibody. 44 2.5 Discussion In the near future, our research group envisions a wide variety of recombinant polypeptide fusions as synthetic biological triggers. Among these, the ELPs are particularly well-suited due their lack of association with other proteins and their abrupt, reversible phase separation. It is likely that other polypeptides could provide similar control; furthermore, these may have diverse applications to perturb and study many cellular pathways. I foresee CAV1-ELPs primarily as a tool for cell culture studies of endocytosis, cell signaling, and mechanosensing. As this chapter reports, the formation of CAV1-ELP microdomains induces a process reminiscent of endocytosis. Accordingly, CAV1-ELP microdomain formation could be used to study the internalization of receptors, pathogens, and drugs. I also note that by using viral transfection methods, the range of CAV1-ELP transfectable cells could be extended to primary cells or patient derived iPSCs to explore the role of CAV1 in a variety of pathological states. In addition to endocytosis, caveolae also regulate signal transduction by organizing signaling partners into discrete regions near the plasma membrane. 67-69 Accordingly, CAV1-ELP microdomains may cluster caveolar cell surface receptors. The ability to indirectly control receptor clustering on the plasma membrane could be exploited by the cell biology community to control and dissect cell signaling. In recent years it has become clear that caveolae act as suspension system for the plasma membrane by flattening and assembling in response to mechanical and osmotic pressure. 58,65,70-72 45 Therefore, another potential application would be the use of CAV1-ELP microdomains to control plasma membrane curvature and surface area. This would be of great practical use in the emerging field of mechanobiology. In combination with near infra-red laser-irradiation of nanoparticles or dyes, such as the FDA-approved indocyanine green, it may be possible to locally trigger transfected CAV1-ELP phase separation in vivo. 97,98 The ability to precisely control temperature at the tissue or organism level does represent a challenge for this technique. However, given our previous results with zebrafish embryos, in which ELP length and sequence, incubation temperature, and mRNA injection concentration were controlled to achieve tunable ELP assembly, it is likely that this model system could be used for in vivo studies of CAV1-ELPs. 2 Given the emerging significance of CAV1 in health and disease and the remaining major questions regarding caveolar function, CAV1-ELPs will provide the life science community with an innovative method to control CAV1. 2.6 Conclusion This chapter demonstrates, for the first time, a tunable temperature-dependent strategy to self-assemble intracellular CAV1 using ELPs. Self-assembly of CAV1-ELPs induces internalization of CTXB in a mechanism reminiscent of caveolae-mediated endocytosis. The temperature of CAV1-ELP self-assembly can be fine-tuned by adjusting the ELP sequence. CAV1-ELPs also retain the colocalization with Cavin 1 that is characteristic of endogenous CAV1. 46 Chapter 3 Tunable assembly of protein-microdomains in living vertebrate embryos 3.1 Abstract Subcellular events such as trafficking and signaling are regulated by self-assembled protein complexes inside the cell. The ability to rapidly and reversibly manipulate these protein complexes would likely enhance studies of their mechanisms and their roles in biological function and disease manifestation. 99,100 This chapter reports that thermally-responsive elastin-like polypeptides (ELPs) linked to fluorescent proteins can regulate the self-assembly and disassembly of protein microdomains within the individual cells of zebrafish embryos. By exploring a library of fluorescent ELP proteins, my colleague Zhe Li and I demonstrate that ELPs can co-assemble different fluorescent proteins inside of embryos. By tuning ELP length and sequence, we show that fluorescent protein microdomains can be assembled at different temperatures, in varying sizes, or for desired periods of time. For the first time in a multicellular living embryo, these studies demonstrate that temperature-mediated ELP assembly can reversibly manipulate assembly of subcellular protein complexes, which may have applications in the study and manipulation of in vivo biological functions. 3.2 Introduction A major goal of synthetic biology is the design of biocompatible materials that can emulate the sub-micrometer reversibly assembled structures formed within cells. 101-103 Subcellular structures (e.g. transcription machinery, the nuclear envelope, signaling complexes, endosomes) must assemble and disassemble in a regulated fashion for proper function. Since this fine-tuned 47 reversible assembly is so important for cell biology, technologies that can undergo reversible self- assembly inside of cells may yield greater control and understanding of intracellular pathways and their consequent roles in physiology and disease. In previous work our group has shown that thermally responsive elastin-like polypeptides (ELPs) can be fused to effector proteins to reversibly assemble fluorescent cytosolic protein microdomains and to reversibly inhibit Clathrin-mediated endocytosis. 52,55,104 ELPs are biocompatible, biodegradable polypeptides consisting of a pentameric repeat of (VPGXG)n, where X determines the ELP hydrophobicity and n determines the molecular weight. 105,106 ELPs abruptly phase separate into protein-rich domains in response to heating, whereby they form a secondary aqueous phase known as a coacervate. Individual coacervates are typically several hundred nanometers to several micrometers in diameter. This phase separation is a thermodynamically reversible process, and ELP coacervates quickly resolubilize into bulk water upon cooling. The temperature of self-assembly, termed the transition temperature (Tt), is specified by the hydrophobicity of the guest residue and the number of repeats. When fused to an effector protein, such as GFP or clathrin light chain, this temperature sensitive behavior is retained, producing a temperature sensitive mutant that rapidly cycles between self-assembly and disassembly in response to temperature, enabling the reversible control of protein clustering. While this system has been proven robust in cell culture, it remains unknown whether the same strategy will work in a multicellular organism. Zebrafish (Danio rerio) is a small freshwater fish that is commonly used for the study of vertebrate developmental biology. 107,108 Several features make zebrafish an attractive model 48 organism, including low cost of husbandry, high reproductive rates, and the ex vivo development of optically clear embryos. Zebrafish also exhibit a cellular and developmental biology which more closely mimics mammalian systems than commonly used invertebrate models, such as C. elegans or Drosophila melanogaster. 109-111 Importantly, zebrafish can tolerate a reasonable range of temperatures found in their native environments, allowing specimens to be maintained above and below the Tt of optimized ELP fusion proteins. 112 This chapter shows that GFP-ELPs (GFP cloned to amino-terminus of ELP) can be easily expressed in zebrafish, where they undergo temperature-dependent self-assembly and disassembly of protein microdomains (Figure 3.1). Moreover, by tuning the ELP sequence it is possible to design microdomains that form at different temperatures, have distinct sizes, or remain assembled for extended durations. These findings suggest that ELPs are a potentially powerful strategy for controlled protein assembly and disassembly within zebrafish. 3.3 Materials and methods 3.3.1 Zebrafish husbandry and care Zebrafish were raised and maintained at 28.5 °C in a circulating system according to standard protocols that are in accordance with Children’s Hospital Los Angeles IACUC animal care protocol. 113 Wild type AB fish were set up for breeding the day before to obtain embryos for microinjection. 49 3.3.2 Plasmid construction GFP-ELPs in mammalian expression vector pcDNA3.1 were cloned as previously described. 52 To engineer GFP-ELPs for bacterial expression, GFP gene was amplified using PCR and inserted to the N terminus of ELPs in pET25b (+) vectors using the BseRI digestion site. The sequences were verified by DNA sequencing (Retrogen, San Diego, CA). 3.3.3 mRNA preparation and microinjection To linearize the plasmids before in vitro transcription, the GFP plasmid was cut with XbaI and the GFP-ELP plasmids were cut with EcoRI. Linearized GFP and GFP-ELP plasmids were then transcribed into full length capped mRNAs with the T7 mMESSAGE mMACHINE Ultra Transcription Kit (AM1345, Life technologies, Carlsbad, CA) using the manufacturer’s protocol. This mRNA was then purified with the MEGAclear™ Transcription Clean-Up Kit (AM1908, Life technologies, Carlsbad, CA). We followed this kit’s protocol with one major caveat: in the final step we eluted mRNA from the silica filters using room temperature elution buffer, instead of heated elution buffer. This is because the high temperatures suggested in the protocol degraded the mRNA. mRNA degradation was assessed by running aliquots on a denaturing bleach gel. 114 mRNA purity was determined by measuring 260/280 and 260/230 values on a NanoDrop™ 2000 (ND-2000, ThermoFisher SCIENTIFIC, Waltham, MA). mRNAs with 260/280 and 260/230 values near 2.0 were selected for microinjection . Purified in vitro transcribed mRNA was then injected into the cytoplasm of one cell stage zebrafish embryos following standard injection protocols. 50 3.3.4 ELP purification and physicochemical characterization pET25b (+) vectors encoding GFP-ELPs were transformed into BLR (DE3) Escherichia coli competent cells (Novagen Inc., Milwaukee, WI) for protein expression. Inverse transition cycling (ITC) was used to purify ELP samples from the bacteria lysates as previously published. 115 To characterize the phase behaviors of GFP-ELPs, optical density (OD) at 310 nm was monitored using DU800 UV-Vis spectrometer while the temperature was ramped from 15 °C to 45 °C at a rate of 1 °C/min, and then plotted as a function of temperature. The maximum first derivative of the curve was defined as the transition temperature. 3.3.5 Immunoblotting Zebrafish embryos injected with GFP-ELPs were lysed with RIPA buffer containing protease/phosphatase inhibitor (#5872, Cell Signaling Technology, Danvers, MA) at 24 hpf and electronically separated on a PAGEr EX 4-12% gradient gel (#59722, Lonza, Walkersville, MD). Proteins were then transferred onto a nitrocellulose membrane using iBlot2 dry blotting system (Life Technologies, Carlsbad, CA) and probed for GFP (ab290, 1:5000, Abcam, Cambridge, MA) or GAPDH (ab8245, 1:5000, Abcam, Cambridge, MA). An HRP-conjugated anti-rabbit secondary antibody (#7074, 1:5000, Cell Signaling Technology, Danvers, MA) was used for visualization. 3.3.6 Zebrafish imaging Embryos at early time points (4~5 hpf) were placed on a 35-mm glass bottom dish (MatTek, Ashland, MA) and imaged using a LSM800 confocal microscope (Carl Zeiss Microscopy, Thornwood, NY) with a 40x 1.3 NA oil objective. For live fish imaging, embryos were incubated on an HCS60 microscope hot and cold stage (Instec, Boulder, CO) attached to a 51 mK2000 high precision temperature controller (Instec, Boulder, CO) and imaged using a Fluoview FV3000RS confocal microscope (Olympus, Waltham, MA) equipped with a 40x 1.25 NA oil objective. For fixed fish imaging, embryos were incubated at different temperatures and fixed overnight with 2% paraformaldehyde (Alfa Aesar, Tewksbury, MA) at the same incubation temperature before imaging. Fish at late time points (24, 48 hpf) were imaged live on a Leica MZFLIII Stereo/Dissection Microscope (Buffalo Grove, IL). 3.3.7 Image processing 16 bit grayscale images with a resolution of 1024 by 1024 pixels (4~5 hpf) or 1344 by 1024 pixels (24 hpf and 48 hpf) were analyzed with FIJI (version 2.0.0). 81 The FIJI macros and python scripts I wrote to calculate microdomain area and number can be found on github: https://doi.org/10.5281/zenodo.3950514 3.3.8 Statistics Data presented are representative curves or mean ± 95% confidence interval. Statistical analysis was performed in R (version 3.4.2). Outlier detection was performed through visual inspection of histograms of the dataset, summaries of the dataset (i.e. min, max, median, quartiles), and by directly identifying which embryos produced the maximum and minimum for the dataset. If a potential outlier was identified, the original image which produced the data was then inspected to determine if the embryo was out of focus or otherwise abnormal due to experimental error. Although the variability in the imaging datasets was large, this was expected as in vivo experiments tend to have large variability due to the intrinsic variability of each individual organism. This is especially true with zebrafish, as they are more genetically diverse than other commonly used 52 vertebrate model systems. 116-118 In all statistical tests an alpha of 0.05 was chosen as the threshold for statistical significance. All tests were two-sided. Non-parametric tests were chosen for datasets which exhibited pronounced deviations from the assumptions of parametric tests (i.e. homogenous variances and normality) or when the median was a better indication of the central tendency of the dataset. In Figure 3.4c, a two-sided Wilcox rank sum test was used to determine statistically significant differences in median particle size per embryo between the different treatments (p- value = 2.6e-07). To obtain the median particle size per embryo, the area of each particle from the embryo’s field of view was measured and the median was directly calculated from these measurements (25 embryos were injected with GFP-V96 (n = 25) and 23 embryos were injected with GFP-SI (n = 23). This procedure was carried out for each embryo in the GFP-V96 and GFP- SI groups. The Wilcox rank sum test was then applied to this data. A two-sided Welch’s two- sample t-test determined statistically significant differences in particle number per embryo between the different treatments at 24 hpf and 48 hpf (Figure 3.4e). In the 24hpf groups (p value = 0.03), 6 embryos were injected with GFP-V96 (n = 6) and 5 embryos were injected with GFP- SI (n = 5). In the 48hpf groups (p value = 0.010), 3 embryos were injected with GFP-V96 (n = 3) and 3 embryos were injected with GFP-SI (n = 3). 3.3.9 Author Contributions Zhe Li and I contributed equally to this work and are co-first authors on the published manuscript. 2 Unless otherwise noted, I prepared mRNA, while Zhe injected the mRNA into the embryos. I performed all image analysis and corresponding statistics, while Zhe performed all immunoblotting, DLS, and UV-Vis. Zhe and I both performed imaging. Writing of the manuscript was a collaborative effort between Zhe Li, myself, Ching-Ling Lien, and J. Andrew MacKay. 53 3.4 Results 3.4.1 ELP expression and microdomain assembly in a zebrafish embryo A library of GFP-ELPs were first expressed and characterized within zebrafish embryos (Table 3.1, Figure 3.1). While there does appear to be a band of free GFP in the GFP-ELP injected embryos, this level is ~10% of the total integrated density. Since these GFP fragments lack an ELP, they are not expected to significantly participate in setting the phase behavior of intact GFP- ELPs; furthermore, their relative abundance is subject to the nonlinearity inherent in the development of a Western Blot. Since the large majority of the fusion remains intact, it is unlikely that cleaved GFP will contribute significantly to interpretation of fluorescence imaging studies. The optimal temperature used for zebrafish husbandry is 28.5 °C; however, larvae can survive for at least several days at temperatures ranging from 21 to 33 °C. 119,120 Therefore, GFP-V96 was selected for this study as its in vitro Tt is within this range (Figure 3.1d). To visualize the temperature-dependent microdomain assembly, a live fish embryo injected with 400 pg mRNAs encoding GFP-V96 was imaged at late blastula stages (4~5 hours post fertilization, hpf) while the temperature was increased from 20 to 40 °C (Figure 3.1e). The late blastula stage was selected because sufficient GFP became visible to image single cells using confocal laser scanning microscopy. At 20 °C, below the Tt, GFP-V96 remains soluble throughout the embryo (3.1a). Within 5 min of ramping the temperature, microdomain formation was observed in every single cell. Once ELP coacervates have formed within the cytosol, their size and number remain constant 54 over short durations. This may occur because the protein-rich phases are separated from each other by high concentrations of other cytosolic macromolecules. This differs from how purified ELPs concentrate into a single, continuous phase based on their high density with respect to buffer, which is accelerated using centrifugation. Immediately after heating, the fluorescent microdomains within the embryo were completely resolubilized by cooling to 15 °C. These results demonstrate the efficient temperature-mediated self-assembly and disassembly of proteins within the single cells of a zebrafish embryo. Short durations of heating (~5 min) suggest the high temporal resolution of this system. Table 3.1. Nomenclature, amino acid sequence and phase behavior of expressed proteins. Protein Label Amino acid sequence a MW [kD] b Tt [°C] in PBS c Tt [°C] in zebrafish d Intercept, b [°C] e Slope, m [°C] [log10(µM)] -1 GFP GFP 30.0 NA NA NA NA GFP- V60 GFP-(VPGVG)60Y 52.4 36 38 43.8 [42.7 to 44.8] 5.3 [4.5 to 6.2] GFP- V96 GFP-(VPGVG)96Y 67.1 30 28 35.4 [34.8 to 36.1] 3.4 [2.8 to 4.0] GFP-SI GFP- (VPGSG)48(VPGIG)48Y 67.2 22 23 27.0 [26.3 to 27.6] 3.4 [2.8 to 3.9] Not applicable (NA) a) GFP indicates Green Fluorescent Protein with amino acid sequence: “MASKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDATYGKLTLKFICTTGKLPVPW PTLVTTFSYGVQCFSRYPDHMKRHDFFKSAMPEGYVQERTISFKDDGNYKTRAEVKFE GDTLVNRIELKGIDFKEDGNILGHKLEYNYNSHNVYITADKQKNGIKANFKIRHNIEDGS VQLADHYQQNTPIGDGPVLLPDNHYLSTQSALSKDPNEKRDHMVLLEFVTAAGITHGM DELYK” b) Estimated molecular weight from open reading frame confirmed by western blot (Figure 3.1c). c) Transition temperature of purified GFP-ELPs suspended in phosphate buffered saline at 25 µM. d) Observed transition temperature in zebrafish embryos at late blastula stages (molar amount of mRNA equal to 400pg GFP-V96 mRNA injected at one-cell stage) using fixed-zebrafish imaging. e) The Intercept b and slope m, were derived from the log-linear regression analysis for transition temperature vs. concentration (Figure 3.1d) fit to the equation Tt = b - m Log 10[C ELP]. Data represent the mean [95% confidence interval]. 55 Figure 3.1. ELP expression and microdomain assembly in a zebrafish embryo. a) Genes encoding three ELPs (V60, V96, SI; Table 3.1) were fused to the carboxy-terminus of GFP gene for direct visualization. b) Schematic representation of reversible microdomain assembly in a zebrafish embryo. Below their T t, GFP-ELPs are uniformly distributed inside the single cells of the embryo. Upon heat stimulation, the protein polymers reversibly phase separate and assemble microdomains. c) Expression of different GFP-ELPs in zebrafish embryos were confirmed using an anti-GFP antibody. d) The in vitro temperature-concentration phase diagrams for all purified GFP-ELPs follow a log-linear relationship (Table 3.1). e) ELP microdomain assembly in a zebrafish embryo is rapid and reversible. Live embryo imaging was conducted at late blastula stages (4~5 hpf) to visualize the temperature-dependent microdomain assembly from 20 to 40 °C. These polypeptide microdomains can be rapidly resolubilized by decreasing the temperature to 15 °C. Images are taken from the same live embryo at indicated temperatures. The insets are magnified 3.5x with respect to each panel. 56 3.4.2 ELPs can co-assemble different proteins in vivo Signaling and trafficking events inside living cells require the co-assembly of protein complexes composed of distinct proteins. Accordingly, we examined whether ELPs fused with different fluorescent proteins may co-assemble when induced to phase separate (Figure 3.2a). mRNAs encoding GFP-V96 and RFP-V96 were co-injected into single-cell zebrafish embryos. After 4 to 5 hours, blastula embryos were incubated above and below the Tt of V96, fixed, and then imaged. Figure 3.2b reveals that at 23 °C, both RFP-V96 and GFP-V96 remain soluble within the single cells of the embryo. However, at 30 °C both constructs form distinctive microdomains which readily colocalize with each other. While we cannot rule out the possibility that RFP-V96 has a different phase behavior than GFP-V96, the images clearly suggest that selection of an identical ELP tag, V96, leads to direct co-assembly above the Tt. This result is consistent with our previous in vitro results and serves as proof-of-principle that ELP mediated self-assembly can be used to co-assemble different proteins within a multicellular organism. However, we note two limitations with this proof-of-principle study. First, in Figure 3.2b there appears to be a preponderance of signal inside the nucleus compared with the cytosol. This difference in signal intensity may be an artifact of fixation, as it not nearly as apparent in live embryos (Figure 3.1e). However, future studies should determine the subcellular location of these or other ELP fusion proteins. This could be achieved through whole mount immunofluorescence to stain different organelles, by probing ELP fusions with secondary antibodies, or through the use of transgenic zebrafish lines which express tissue or organelle specific reporters. The second limitation is that the relative concentrations of GFP-V96 and RFP-V96 were not determined. The proper formation of many intracellular complexes (e.g. a signaling complex, an endosome, transcription machinery) depends on a proportional relationship between their various protein components. Accordingly, to 57 properly self-assemble such a complex using ELP fusion proteins may require a means to calibrate their relative concentrations in vivo. To address this our next experiments, detailed section 3.4.3, reveal that ELP protein levels can in fact be controlled by adjusting mRNA injection amount. Figure 3.2. ELPs can co-assemble different proteins in vivo. a) Schematic of triggered co-assembly of fluorescent ELP fusions in a zebrafish embryo. GFP-V96 and RFP-V96 are soluble throughout the cytosol below T t and co-assemble into mixed microdomains above the T t. b) Confocal microscopy imaging of GFP- V96 and RFP-V96 in zebrafish embryos. Fish embryos were co-injected with mRNAs encoding GFP-V96 and RFP-V96 at one-cell stage and then incubated either below (23 °C) or above (30 °C) T t until fixation at last blastula stages (4~5 hpf). 58 3.4.3 Transition temperature in vivo can be controlled through ELP sequence and concentration Prior studies have shown that the in vitro Tt can be fine-tuned by varying ELP length and concentration. 121 To examine this effect in vivo, GFP-V60, which contains 60 repeats of VPGVG, was compared with GFP-V96. GFP alone served as a non-switchable negative control. Given that mRNA injection amount was the only significant predictor of protein expression level (Figure 3.3, Table 3.2), fish embryos were injected with equimolar amounts of mRNA (GFP:150 pg, GFP- V60: 300 pg, GFP-V96: 400 pg) and incubated at various temperatures for 30 mins prior to overnight fixation at the same temperature. At 23 °C, nearly all GFP-V96 remained soluble, while at temperatures above 28 °C there was extensive GFP-V96 self-assembly (Figure 3.4 a, c,d). In contrast, fish embryos injected with GFP-V60 mRNA did not show any microdomain assembly until heated up to 38 °C, suggesting that the in vivo Tt increased with decreased polymer length. In addition, the effect of ELP concentration on the Tt was examined by injecting different amounts of mRNAs (80 pg, 400 pg, 2000 pg) encoding GFP-V96 (Figure 3.4b). The change of in vitro Tt as a function of ELP concentration was modeled in Figure 3.1d, 3.4e and Table 3.1, showing that a 10-fold decrease of GFP-V96 concentration increased Tt, by 3.4 °C with a log normal relationship between ELP concentration and Tt. Similarly, the in vivo Tt increased by 7.2 °C with a 10-fold decrease in the mRNA dose (Figure 3.4b, f). While the degree of concentration-dependence differs slightly, both purified and cytosolic GFP-V96 follow the same log-linear relationship between ELP amount and Tt. In addition, the size and irregularity of the GFP-V96 puncta appear to increase with increased mRNA concentration (Figure 3.4b). This phenomenon is most likely due to a dramatic increase in intracellular GFP-ELP amount and is consistent with data that reveals that GFP-V96 integrated density increases with injected mRNA amount (Figure 3.3). These results suggest that mRNA injection amount can be adjusted to control protein levels, transition 59 temperature, and perhaps coacervate size. To further explore these possibilities, future studies should confirm in vivo ELP expression levels through a combination of fluorescent measurements and biochemical measurements taken from lysed whole embryos. Table 3.2 Statistics for Figure 3.3 Comparison a Unadjusted p value b Adjusted p value GFP-SI (1000pg) vs. GFP-V60 (300pg) 0.432912580 1.00000000 GFP-SI (1000pg) vs. GFP-V96 (1000pg) 0.044941147 0.26964688 GFP-V60 (300pg) vs. GFP-V96 (1000pg) 0.553327664 1.00000000 GFP-SI (1000pg) vs. GFP-V60 (300pg) 0.063437802 0.38062681 GFP-V60 (300pg) vs. GFP-V96 (400pg) 0.030766406 0.18459844 GFP-V96 (1000pg) vs. GFP-V96 (400pg) 0.001472875 0.00883725 a) Pairwise comparisons of whole field integrated density values obtained from confocal images of embryos injected with different GFP-ELP constructs. Note that GFP-V60: 300 pg, and GFP-V96: 400 pg are equimolar. b) Unadjusted and adjusted p values were calculated using Dunn's test of multiple comparisons using rank sums with Bonferroni correction (requires R packages dunn.test and FSA). Note that adjusted p value = (unadjusted p value) * 6. If adjusted p value is below 0.05, result is statistically significant at an alpha of 0.05. 60 Figure 3.3. mRNA injection amount is the only significant predictor of GFP-ELP protein levels, as measured by whole field integrated density of confocal microscope images. a) At the single cell stage, embryos were injected with varying amounts of mRNAs encoding GFP or GFP-ELPs. Note that GFP:150 pg, GFP-V60: 300 pg, and GFP-V96:400 pg are equimolar. Five hours after microinjection, embryos were imaged on a confocal microscope. Whole field integrated density was calculated from 8-bit images using FIJI (version 2.0.0). A Kruskall-Wallis rank sum test returned a p value of 0.0146 (significant at alpha = 0.05). To determine pairwise statistically significant differences between all GFP-ELP groups, a Dunn multiple comparison test with Bonferroni correction was computed. Only GFP-V96 (1000 pg) vs. GFP- V96 (400 pg) display a statistically significant Bonferroni adjusted p value (* adjusted p value = 0.009). Note that GFP-SI (1000 pg) vs. GFP-V96 (1000 pg) does not exhibit a statistically significant difference (adjusted p value = 0.3). These results imply that the different in vivo transition temperatures for different constructs are not caused by differences in intracellular concentration (Figure 3.4a and 3.5). GFP 150 pg (n = 7), GFP-V60 300 pg (n = 6), GFP-V96 400 pg (n = 5), GFP-SI 1000 pg (n = 18), GFP-V96 1000 pg (n = 21). Pairwise statistical results for this data are shown in Table 3.2 b) Whole field integrated density exhibits a log-linear relationship with GFP-V96 mRNA injection amount. This analysis indicates that the in vivo concentration GFP-V96, which regulates in vivo transition temperature, can be controlled with mRNA injection amount (Figure 3.4b). At the single cell stage, embryos were injected with varying amounts of GFP-V96. Five hours after microinjection, embryos were fixed and then imaged on a confocal microscope. Whole field integrated density was calculated from 8-bit images using FIJI (version 2.0.0). A linear regression of log base 10 of integrated density vs. mRNA amount was fit to the data. Bars indicate 95% CIs for indicated concentrations. 85pg (n = 5), 400pg (n = 5), 1000pg (n = 21), 2000pg (n = 7). Adjusted R- squared = 0.76. Estimated slope = 5.5e-04. Slope p value = 5.2e-13. a b 61 Figure 3.4. ELP self-assembly temperature can be tuned both in vitro and in vivo. Temperature- dependent microdomain assembly was characterized by varying either a) molecular weight (400 pg mRNAs encoding GFP-V96 vs. 300 pg mRNAs encoding GFP-V60) or b) concentrations (80 pg, 400 pg, 2000 pg mRNAs encoding GFP-V96) of ELPs. Zebrafish embryos at late blastula stages (4~5 hfp) were incubated at different temperatures, fixed, and imaged using a Zeiss confocal microscope. *indicates microdomain assembly observed. c) Representative in vitro optical density profiles for GFP-V60 and GFP-V96 at 25 µM as a function of temperature. d) Quantification of the in vivo particle number per field as a function of temperature for GFP-V60 and GFP-V96. Mean ± 95% confidence interval (n=3~20). e) Representative in vitro optical density profiles for GFP-V96 at various concentrations (5 µM, 25 µM, 75 µM) as a function of temperature. f) Quantification of the in vivo particle number per field as a function of temperature for embryos injected with different amounts (80pg, 400pg, 2000pg) of mRNAs encoding GFP-V96. Mean ± 95% confidence interval (n=3~16). 62 3.4.4 Duration of in vivo ELP assembly can be tuned Having demonstrated the tunability of temperature-dependent ELP-mediated microdomain assembly in zebrafish embryos, we next explored the duration of assembly for mRNA-encoded ELPs. Since mRNA is injected into the embryo immediately after fertilization, the levels of mRNA are expected to decrease during the development. Eventually, this will reduce protein expression too low to promote assembly at a physiological temperature. Based on this, fusions with phase transition temperatures below 28 °C (Figure 3.1d) GFP-V96 and GFP-SI were compared. GFP- SI has a similar molecular weight and to GFP-V96 but transitions at a lower temperature (Table 3.1, Figure 3.1c-d, 3.5a). In vivo both GFP-SI and GFP-V96 exhibit significant microdomain assembly at physiological temperatures at late blastula stages (4~5 hpf) (Figure 3.5a, Figure 3.6), with GFP-SI forming microdomains (median area = 0.146 µm 2 ) slightly smaller than GFP-V96 (median area = 0.341 µm 2 ) (Figure 3.5c). Once fish developed to prim-5 (24 hpf) or long-pec (48 hpf) stages, microdomain assembly at 28 °C was readily observed in fish expressing GFP-SI, but not GFP-V96 (Figure 3.5b, d). These results are consistent with GFP-SI having a lower in vivo Tt than GFP-V96. They also suggest that by selecting constructs with low transition temperatures, microdomain assembly can be retained in zebrafish embryos over periods of at least days despite the loss and dilution of the encoding mRNA. However, we note that the variability in particle number was quite high for the 24 hpf GFP-SI embryos (max = 729, min = 83) (Figure 3.5d). In our studies, we noticed that each embryo displayed an intrinsic ability to translate injected mRNA into protein. This is likely due to the high genetic variation within zebrafish, which exhibit greater genetic diversity, even within the same strain, than other inbred vertebrate model systems. 116-118 This biological variability is further compounded by the great technical difficulty associated with 63 microinjecting each single celled embryo with the same amount of mRNA. Future studies should aim to mitigate these sources of biological and technical variability. Figure 3.5. ELPs can be tuned for short- or long-term microdomain assembly. a) Short-term microdomain assembly at physiological temperatures was observed in both GFP-SI and GFP-V96. Fish embryos injected with 1000 pg of mRNAs encoding either GFP-SI or GFP-V96 were fixed at late blastula stages (4~5 hpf) and imaged using a confocal microscope. b) Long term microdomain assembly at physiological temperatures was observed in GFP-SI, but not GFP-V96. Live fish embryos expressing GFP- SI or GFP-V96 were visualized using a dissecting scope at 24 hpf. c) Quantification of microdomain areas inside of fixed embryos at late blastula stages (4~5 hpf). GFP-SI (23 embryos, 6692 total particles, median area = 0.146 µm 2 ). GFP-V96 (25 embryos, 12078 total particles, median area = 0.316 µm 2 ). p value for difference in median particle area = 2.6e-07. d) Quantification of particle number per fish at 24 and 48 hpf, showing that GFP-SI assembled significantly more particles than GFP-V96 at longer time periods. Mean ± 95% confidence interval. 24 hpf: GFP-SI n = 5, GFP-V96 n = 6, two-sided t-test p value = 0.03. 48 hpf: GFP-SI n = 3, GFP-V96 n = 3, two-sided t-test p value = 0.009. 64 Figure 3.6. At late blastula stages (4-5 hpf), embryos injected with GFP-V96 display more microdomains/particles than embryos injected with GFP-SI. GFP-V96 (25 embryos, 12078 total particles, median number of particles per cell = 12). GFP-SI (23 embryos, 6692 total particles, median number of particles per cell = 6). A two-sided Wilcox rank sum test was used to determine if there was a statistically significant difference in particles per cell between the two treatments (p value = 0.001, significant at alpha = 0.05). Note that these microdomains were counted from the same embryos used to calculate the particle area histogram in Figure 3.5c. At the single cell stage, embryos were injected with ~1000 pg of GFP-V96 or GFP-SI mRNA. Approximately 4~5 hpf, embryos were incubated at physiological temperatures for 30 minutes before fixation and subsequent imaging on a confocal microscope. Particle per cell was calculated by counting the number of microdomains in the entire image divided by the number of cells in the image. *** 65 3.4.5 ELPs do not affect survival or embryonic development Finally, all constructs were compared for their effect on embryo survival at 48 hours, which show that ELP-mediated phase separation alone induces no significant loss of viability (Figure 3.7a). In addition, none of the ELP fusions produced gross developmental changes (Figure 3.7b), which suggests that ELP fusions are well tolerated in zebrafish embryos. Figure 3.7. ELPs do not affect survival or embryonic development. a) Survival rate of zebrafish embryos with different treatments at 48 hpf. Mean ± 95% confidence interval (n=3, 20 to 50 fish embryos injected with 150~1000 pg mRNAs were evaluated within in each experiment). b) ELP expression does not affect the embryonic development of zebrafish. Fish embryos expressing GFP or GFP-ELPs were maintained at 28 °C and imaged using a dissection scope at 24 and 48 hpf. 66 3.5 Conclusion In summary, for the first time my colleagues and I report a temperature-dependent strategy for controlled protein assembly in a vertebrate embryo. This approach is rapid, reversible, and highly tunable. In addition, it is biocompatible and can be used to assemble one or more protein species together inside the single cells of a multicellular organism. In future work, this approach may aid developmental biology studies by allowing tunable assembly of functional proteins inside of zebrafish embryos. 67 Chapter 4 Method for determining the intracellular transition temperature of elastin-like polypeptide fusion proteins 4.1 Abstract: Elastin-like polypeptides (ELPs) are modular stimuli-responsive materials that self-assemble into protein-rich microdomains in response to temperature change. By cloning ELPs to effector proteins, thermo-responsive fusion proteins can be expressed which modulate cellular pathways via temperature induced assembly. The temperature at which phase separation and microdomain formation occurs is called the transition temperature (Tt), and a critically important step in engineering any ELP fusion protein is to determine its intracellular Tt. In this chapter, I describe a simple live cell imaging technique to estimate the intracellular Tt of non-fluorescent ELP fusion proteins by co-transfection with a less temperature sensitive fluorescent ELP marker. Intracellular assembly and microdomain formation can then be visualized in live cells through the co-assembly of the non-fluorescent and fluorescent ELP fusion proteins. Critically, if the two different ELP species exhibit differences in temperature sensitivity, the resulting mixture of ELPs exhibit a decreased Tt that corresponds to the Tt of the more temperature sensitive ELP. In addition to changes in Tt, co-assembled ELP microdomains may also exhibit pronounced differences in size or number, compared to single transfected treatments. These features make it possible to employ relatively simple live cell imaging experiments and image analysis to determine the intracellular Tt of a given ELP fusion library. As a case study, I focus on the CAV1-ELP library but analyze a new and distinct dataset from that described in chapter 2. Furthermore, I also provide preliminary data regarding the ELP-CLC library, which has never been analyzed with this method. To assist with many of the image analysis techniques described in this chapter, I developed an easily installed FIJI plugin named SIAL (Simple Image Analysis Library), which contains programs for 68 image randomization and blinding, phenotype scoring, and ROI selection. These tasks are important parts of the protocol detailed here but are also commonly employed in other image analysis workflows. 4.2 Introduction Elastin-like polypeptides (ELPs) are thermally responsive peptides constructed out of a pentameric amino acid repeat of (VPGXG)n. In this modular design, the X residue can be selected to alter ELP hydrophobicity, and the number of repeats, n, determines the ELP molecular weight. In response to heat, ELPs can rapidly (within seconds to minutes) phase-separate and self-assemble into protein-rich domains, forming a secondary aqueous phase known as a coacervate. If the ELPs are assembled inside of a cell, these coacervates are commonly referred to as microdomains. Individual microdomains can be hundreds of nanometers to several micrometers in diameter. Phase separation is a thermodynamically reversible process, and ELP coacervates quickly resolubilize into bulk water upon cooling. The temperature of self-assembly, termed the transition temperature (Tt), is directly encoded in the ELP sequence, because it is determined largely by the hydrophobicity of the guest residue and the ELP molecular weight. 53,54,122 Importantly, these stimuli responsive characteristics are retained when ELPs are cloned to an effector protein, which allows researchers to create libraries of ELP fusion proteins to control their cellular pathway of interest. Based on this approach, we have previously reported that thermo-responsive ELP fusion of clathrin light chain can be used to reversibly inhibit clathrin mediated endocytosis. 55 More recently, we cloned ELPs to epidermal growth factor receptor (EGFR-ELPs) to control epidermal growth factor signaling, and have also created thermo-responsive fusions of Caveolin 1 and ELPs (CAV1-ELPs) which internalize cholera toxin via the formation of CAV1-ELP microdomains 1,56 69 A critically important step in engineering any ELP fusion protein is the determination of its intracellular Tt. This task cannot be directly accomplished in live cells, because ELP fusion proteins are typically not fluorescent or otherwise directly visible (unless an ELP is fused to a fluorescent protein). An alternative approach could involve incubating ELP-transfected cells at select temperatures followed by fixation and indirect staining with antibodies to determine if the ELPs have phase transitioned. Realistically, however, only a few different temperatures can be selected before this approach becomes unfeasible, and accordingly, the precise Tt is unlikely to be determined with this approach. Lastly, one could clone a fluorescent reporter onto their ELP fusion protein. Although this method would permit direct visualization of ELP self-assembly inside of a cell, the resulting fluorescent-ELP-fusion protein would likely be several times larger than the wild type protein. The creation of such a large genetically modified fusion protein raises valid questions of how similar the biological function of the fusion protein is to not only the wild type protein (without ELP or fluorescent reporter) but also how similar its transition temperature is to the non- fluorescent ELP fusion protein. An ideal method would permit visual confirmation of intracellular ELP fusion protein assembly in live cells, in real time, as they are subjected to a temperature ramp. Although direct visual confirmation of non-fluorescent ELP fusion protein self-assembly in live cells is not possible, indirect visual confirmation can be achieved by exploiting the fact that ELPs co-assemble with other ELP species. Thus, by co-transfecting cells with a non-fluorescent ELP fusion protein, and a secondary fluorescent ELP fusion protein, assembly can be visualized in live cells through the co-assembly of the non-fluorescent and fluorescent ELP fusion proteins. Importantly, if the 70 two different ELP species exhibit differences in temperature sensitivity, the resulting mixture of ELPs exhibit a decreased Tt which corresponds to the Tt of the more temperature sensitive ELP (Figure 4.1). In addition to changes in Tt, co-assembled ELP microdomains may also exhibit pronounced differences in size or number, compared to single transfected treatments (Figures 4.5- 4.6). These features make it possible to employ relatively simple live cell imaging experiments and image analysis to determine the intracellular Tt of a given ELP fusion library. Figure 4.1. Intracellular T t of non-fluorescent ELP fusion proteins can be visually determined in live cells through co-assembly with a fluorescent ELP fusion protein. If the fluorescent ELP fusion protein (e.g. a GFP-ELP) displays a higher T t than the non-fluorescent ELP fusion protein (as shown in this figure), then cells that are transfected with both constructs will on average display fluorescent microdomains at lower temperatures than cells transfected with only the fluorescent ELP fusion protein. This average lower temperature is the T t of the non-fluorescent ELP fusion protein. Although I have previously reported this method to study ELP fusions of epidermal growth factor receptor and Caveolin 1 (i.e. EGFR-ELPs and CAV1-ELPs), this chapter describes the technique in more detail and provide suggestions for improvement. 1,56 As a case study, I use the focus on the CAV1-ELP library but analyze a new and distinct dataset from that described in chapter 2. In addition, I also provide preliminary data regarding the ELP-CLC library, which has never been analyzed with this method. To assist with many of the common image analysis tasks 71 that are involved in the described methods, I developed an easily installed FIJI plugin which can be used to randomize and blind imaging data, score phenotypes, and record regions of interest (ROIs) from images. These tasks are an important part of the methods described here and are also commonly employed in other image analysis workflows. Table 4.1. Nomenclature, amino acid sequence, and phase behavior of expressed proteins. Protein Label Amino acid Sequence a MW c [kD] Mean Tt in cells e [°C] Median Tt in cells d [°C] Percent of cells showing assembly f [%] Intended behavior CAV1- V96 +GFP- V60 CAV1-myc b - (VPGVG)96 62.5 29.2 / 28.8 29.1 / 27.3 100 / 100 temperature- responsive CAV1- A96 +GFP- V60 CAV1-myc- (VPGAG)96 59.8 37.3 37.6 39 temperature- insensitive GFP- V60 g GFP-(VPGVG)60 67.2 35.5 35.5 71 live-cell imaging tool GFP- V60 h GFP-(VPGVG)60 36.6 37.4 42.3 live-cell imaging tool GFP- V60 i GFP-(VPGVG)60 36.0 36.4 57.4 live-cell imaging tool V96-CLC +GFP- V60 (VPGVG)96-myc-CLC j 33.7 33.5 92.3 temperature- responsive A96-CLC +GFP- V60 (VPGAG)96-myc-CLC j 34.3 35.9 71.4 temperature- insensitive a) ORF amino acid sequences. b) myc-epitope amino acid sequence: EQKLISEEDL c) Estimated molecular weight from open reading frame d) Median T t obtained from data in Figures 4.6 and 4.10. Unfiltered / Filtered e) Mean T t obtained from data in Figures 4.6 and 4.10. Unfiltered / Filtered f) Percentage of cells with a T t ≤ 37°C. Calculated from data in Figures 4.6 and 4.10. Unfiltered / Filtered g) GFP-V60 from the CAV1-ELP dataset (shown in Figures 4.5-4.7) h) GFP-V60 from the ELP-CLC dataset (shown in Figures 4.9-4.11) i) Combined GFP-V60 from the CAV1-ELP and ELP-CLC datasets. Note that the CAV1-ELP and ELP- CLC datasets did not have a significant difference in GFP-V60 T t (p-value = 0.149) j) Full length amino acid sequences are found in the supplementary materials of the original ELP-CLC paper. 55 72 Figure 4.2. Stepwise schematic of the experimental procedure. HEK293T cells are passaged onto 35mm glass bottom dishes. 24hrs after passaging, when the cells have achieved ~70% confluency, they are transiently transfected with either GFP-V60 alone or dual transfected with GFP-V60 and another non- fluorescent ELP fusion protein. After a 72hr incubation at 30°C and 5% CO 2, cell culture media is replaced with ice cold live cell imaging solution. Cells are then immediately time lapsed imaged during a temperature bath ramp from 15 °C to 60 °C to determine the temperature at which fluorescent GFP-V60 microdomains assemble. This temperature serves as the estimate of the intracellular transition temperature (T t) of the co- transfected non-fluorescent ELP fusion protein. 73 4.3 Materials 4.3.1 Cell culture ● HEK293T cells (#CRL-3216, ATCC, Manassas, VA) ● T75 Cell Culture Flask (4616, Laguna Scientific, Laguna Niguel, CA) ● Trypsin-EDTA (0.05%) (25300054, Thermo Fisher Scientific, Waltham, MA) ● Dulbecco's Modified Eagle Medium (DMEM) (11995065, Thermo Fisher Scientific, Waltham, MA) ● Fetal Bovine Serum (FBS) (#35-011-CV, Corning, NY) ● 1X dPBS (#25-508, Genesee, San Diego, CA) 4.3.2 Transfection ● Lipofectamine™ 3000 Transfection Reagent (L3000008, Life Technologies, Carlsbad, CA) ● Opti-MEM™ Reduced Serum Medium (31985070, Thermo Fisher Scientific, Waltham, MA) ● Poly-D-Lysine (P6407, Sigma-Aldrich, St. Louis, MO) ● 35mm glass bottom dishes (P35G-0-10-C, MatTek Corporation, Ashland, MA) 4.3.3 Live cell imaging ● Live Cell Imaging Solution (A14291DJ, Thermo Fisher Scientific, Waltham, MA) ● LSM 880 (Carl Zeiss, Oberkochen, Germany) ● LD LCI Plan Apochromat 25x/0.8 numerical aperture objective for oil/water/glycerol/silicone immersion (420852-9871-000, Carl Zeiss, Oberkochen, Germany) ● Stage attachment Z PIEZO WSB 500 (D) (432339-9000-000, Carl Zeiss, Oberkochen, Germany) 74 ● Stage insert Z PIEZO WSB 500 for Heating Inserts P S1 / Mxx S1 (D) (432339-9050- 000, Carl Zeiss, Oberkochen, Germany) ● Immersol W (444969-0000-000, Carl Zeiss, Oberkochen, Germany) ● Argon 488 laser line (Carl Zeiss, Oberkochen, Germany) 4.3.4 Temperature control during live cell imaging ● Heating Insert P Lab-Tek TM S1 (#131-800 029, PeCon GmbH, Erbach, Germany) ● Control Sensor T S1 (#880-800242, PeCon GmbH, Erbach, Germany) ● AP07R-40 Refrigerating / Heating Bath (89202-982, VWR, Radnor, PA) 4.3.5 Image analysis • FIJI • SIAL (a FIJI plugin for image randomization and blinding. Can be installed by adding https://sites.imagej.net/D-tear/ to ImageJ Updater) • Template Matching (an optional FIJI plugin for stabilizing lateral drift in time lapse images. Can be installed by adding http://sites.imagej.net/Template_Matching/ to ImageJ Updater) • measureStack.ijm (an optional ImageJ macro for automated measuring of multiple ROIs on an image stack. This macro can be found in the supplementary materials.) • R and RStudio (optional; if pixel value standard deviation analysis is desired) • batch_plot (optional R function for analyzing pixel value standard deviation data. R notebook and example data are provided here: https://doi.org/10.5281/zenodo.3950576 ) 4.3.6 Fixed cell imaging to confirm GFP-V60 colocalization with non-fluorescent ELP fusion ● LSM 800 (Carl Zeiss, Oberkochen, Germany) 75 ● Plan Apochromat 63x/1.40 numerical aperture Oil DIC M27 objective (420782-9900-000, Carl Zeiss, Oberkochen, Germany) ● Immersol 518F (444970-9000-000, Carl Zeiss, Oberkochen, Germany) ● 640nm (CAV1-V96), 488nm (GFP-V60), and 405nm (DAPI) laser lines ● 12-Well Cell Culture Plates (25-106, Genessee Scientific, San Diego, CA) ● VWR® Micro Cover Glasses, Round, No. 1 (48380-046, VWR, Radnor, PA) ● Torrey Pines Echotherm IC25 Fully Programmable Chilling/Heating Dry Bath (IC25, Torrey Pines Scientific, Carlsbad, CA) ● 10X PBS Dry Pack (#MB1001, Biopioneer, San Diego, CA) ● Paraformaldehyde (PFA) (43368, Alfa Aesar, Tewksbury, MA) ● Ammonium chloride (NH4Cl) (Alfa Aesar, Tewksbury, MA) ● Bovine Serum Albumin (BSA) (A9647, Sigma-Aldrich, St. Louis, MO) ● myc-tag mouse mAb (9B11 #2276S, Cell Signaling Technology, Danvers, MA) [1:1000 dilution] ● rabbit GFP pAb antibody (ab290, Abcam, Cambridge, MA) [1:500 dilution] ● Chicken anti-Mouse IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor 647 (#A-21463, ThermoFisher Scientific, Waltham, MA) [1:500 dilution] ● Donkey anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 488 (#A-21206, ThermoFisher Scientific, Waltham, MA) [1:500 dilution] ● ProLong™ Glass Antifade Mountant (P36984, ThermoFisher Scientific, Waltham, MA) 4.4 Methods 4.4.1 Cell culture 76 HEK293T cells were grown in T75 flasks with DMEM supplemented with 10% FBS in a humidified incubator with 5% CO2 at 37 °C. During subculture, cells were passaged using Trypsin- EDTA (0.05%) and washed with dPBS. In general, cells were maintained and subcultured under the guidelines recommended by ATCC. 4.4.2 Transfection 4.4.2.1 Preparing MatTek plates for cell culture To prepare the MatTek 35mm glass bottom dishes for adherent cell culture and subsequent transfection, a 0.1 mg/mL solution of Poly-D-lysine was added to each MatTek dish. To make this solution, we followed the manufacturer’s suggested protocol and added 50 mL of sterile tissue culture grade water to 5 mg of poly-D-lysine. 2ml of this solution was then applied, under sterile conditions, to each MatTek plate. After 5 minutes, the solution was removed through aspiration and gently rinsed with sterile dPBS. Critical Points and Comments: Poly-D-lysine improves cell attachment to the glass bottom of the MatTek dishes. An exact amount of 2ml isn’t required, but it is critically important that enough of the solution is applied to cover the bottom of the dish, especially the glass coverslip. When not in use, the Poly-D-lysine solution should be stored at 4°C. Also note that as long as aseptic conditions are used, we found that the Poly-D-lysine solution can be pipetted up and reused in future experiments with no discernible effect on cell attachment or health. Cell passaging onto MatTek plates Before transfection, cells were subcultured from the T75 flask, resuspended in complete media at a density of 1 x 10 6 cells/ml, and 2ml of this solution was then placed into each poly-D- 77 lysine coated MatTek dish. Plated cells were then incubated for 24hrs in the 37 °C, 5% CO2 incubator before transient transfection 24 hrs later. Critical Points and Comments: Seeding densities should always be empirically determined, especially if another cell type is used. In our work, we aimed for at least 70% confluency 24hrs after seeding. 4.4.2.2 Transfection Cells were transfected with the Lipofectamine 3000 kit. All transfection reagents and plasmid DNA were diluted in Opti-MEM™ Reduced Serum Medium media. In preparing the lipofectamine master mix, we mimicked the suggested protocol for a single 6-well plate. The transfection mix for a single 35mm dish involved 125µl of the diluted DNA/P3000 mixture (containing 2.5µg total of plasmid DNA, the P3000 Reagent at a concentration of 2µg/µl of DNA, and Opti-MEM up to 125µl) + 125µl of the Lipofectamine mixture (120µl of Opti-MEM + 5µl of Lipofectamine 3000). Each individual mixture was thoroughly mixed before being added together to create the 250µl master mix. This master mix was then incubated for 15 minutes at room temperature. During this time, the cell seeded MatTek dishes were removed from the incubator, washed one time with room temperature dPBS, and were then replenished with 2ml of Opti-MEM. After the 15-minute incubation, the 250µl master mix was pipetted drop-by-drop into the 2ml of Opti-MEM media. Cells were then placed in a 30˚C incubator with 5% CO2 and were incubated for 72 hrs before subsequent imaging. This lower temperature was chosen to help inhibit premature ELP self-assembly as the exogenous proteins are expressed during the transfection incubation period. 78 For dual transfected treatments (CAV1-V96 + GFP-V60, CAV1-A96 + GFP-V60, V96- CLC+ GFP-V60, A96-CLC + GFP-V60), 1.25ug of each DNA plasmid was added to the transfection reaction, for a total of 2.5µg DNA per MatTek dish. In the single transfection treatment, 2.5µg of GFP-V60 plasmid DNA was added to the transfection reaction. Each MatTek dish received either CAV1-V96 + GFP-V60, CAV1-A96 + GFP-V60, V96-CLC+ GFP-V60, A96- CLC + GFP-V60, or GFP-V60 alone. 4.4.3 Live cell imaging 4.4.3.1 Imaging settings Live cell imaging of ELP intracellular assembly was accomplished with a LSM880 laser scanning confocal microscope equipped with an airy scan detector and definite focus module. As the cells were immersed in aqueous live cell imaging solution during imaging, a plan apochromat 25x/0.8 numerical aperture water immersion objective was chosen for image acquisition. Immersol W was used as the immersion media, as it exhibits the same refractive index of water but does not evaporate easily. To excite GFP-V60, the 488-argon laser line was selected, and all images were acquired using the airy scan fast mode with the definite focus module activated. The laser power selected was 17% and laser light was passed through a 488nm main beam splitter (MBS) before illuminating the cells. Excited light emitted from the cells was then passed through a dual bandpass (495nm-550nm) + long pass (>= 570nm) filter before reaching the airy scan detector. Other imaging parameters include a 1.84µs pixel dwell time, with a lateral sampling of 0.149µm, no averaging, 625 master gain, 1.00 digital gain, and a total image size of 338.29 by 338.29 microns at 2272x2272 pixels. All images were 16-bit. Using the Zeiss Black “Time Series” module, 90 79 images were acquired over the course of the temperature ramp, with image acquisition every twenty seconds. This corresponded to a total imaging time of 30 minutes. Critical Points and Comments: The laser power, pixel dwell time, and master gain settings described above were used for the CAV1-ELP library. The ELP-CLC library was primarily imaged at an earlier period of time and although some of the images were acquired with the same settings as the CAV1-ELP library, a majority were acquired using a variety of laser powers, pixel dwell times, and master gain settings, as I was trying to optimize imaging settings. Although imaging settings should be held constant over the course of a study, the exact imaging settings used are not necessarily critical for determining intracellular Tt. For example, in my previous work with EGFR- ELPs and CAV1-ELPs (chapter 2), I employed an older epifluorescent microscope equipped with lower resolution non-immersion/air objectives. 1,56 Although this imaging setup was inferior to the setup currently employed, I was still able to detect significant differences in transition temperature and quantify differences in intracellular microdomain size and number for different ELP fusion proteins. Although this method’s results are not objectively linked to one particular imaging setting, I do recommend a few important guidelines 1) Employ water or oil immersion objectives, ideally water immersion. The buildup of condensation during temperature ramp imaging makes work with non-immersion/air objectives extremely challenging. Water immersion objectives are ideal over oil because water (or immersol W) will match the refractive index of the imaging solution 2) If using a water immersion objective, use immersol W (or a similar solution) instead of DI water as the immersion solution. Immersol W has the same refractive index as water but does not evaporate as easily. This is important because as the temperature is increased during imaging, a small droplet of water placed on the objective can easily evaporate. When evaporation happens, images lose substantial quality. 3) Balance magnification with numerical aperture 80 strength. Although higher numerical aperture objectives provide superior resolution and light collecting ability, they are typically restricted to higher magnification objectives. Magnification power is important because temperature changes can produce focal shift in the axial and lateral planes, although in our experience drift in the lateral plane is typically much more substantial. With high magnifications/small fields of view, lateral drift means that some cells may actually drift out of the field of view during the temperature ramp. Lower magnifications circumvent this problem by having a larger field of view. Larger fields of view/lower magnifications also mean that more cells can be imaged at a time. In addition, given two objectives with the same numerical aperture strength but different magnifications, the objective with the lower magnification will produce brighter images. This can help avoid photobleaching, because lower laser powers can be used. 4) Balance image quality and acquisition speed with photobleaching. On a confocal microscope, once an objective is selected, laser power, pixel dwell time, and averaging are positively correlated with photobleaching. Photobleaching is especially a concern during live cell imaging when numerous images of the same field of view are repeatedly collected. However, once an appropriate laser power is selected, increased pixel dwell time and averaging can improve image quality by increasing signal to noise (SNR), although increasing either of these options will slow acquisition time. Rapid acquisition times are important in live cell imaging of dynamic events, such as intracellular ELP assembly. In general, experimenters will have to empirically determine their optimal imaging settings which maximize SNR and image acquisition speed while reducing toxic photobleaching. 5) Acquire 16-bit images. Due to cell to cell variability in differences of GFP-ELP protein expression, cells within the same culture dish may exhibit order of magnitude differences in fluorescence intensity (Figure 4.7 quantifies how these differences in fluorescence intensity/ELP concentration affect the intracellular Tt). Because of this large dynamic range, it is 81 very important that experimenters acquire images with maximal bit depth. We recommend 16-bit images. 4.4.3.2 Temperature control during live cell imaging Before imaging, a single 35mm MatTek dish, corresponding to a single transfection condition, was removed from the 30°C incubator, washed one time with dPBS, and then replenished with 2.5ml of 4°C live cell imaging solution. These ice-cold temperature steps were taken to help resolubilize any intracellular ELPs which may have phase transitioned during the 72hr incubation at 30°C, as well as to stall any ongoing intracellular trafficking. The 35mm dish was then placed on ice and carried to the microscope room where it was placed on a PeCon Heating Insert P Lab-Tek TM S1stage attached to circulating temperature bath which was held at 4°C. To accurately measure the temperature of the live cell imaging solution, a PeCon Control Sensor T S1 was placed directly into the live cell imaging solution (Figure 4.3). Once all equipment was in position and the microscope objective was in focus, the temperature bath was set to a ramp rate of 15°C to 60°C over twenty minutes, and once 60°C was achieved, the temperature bath was held at 60°C. During the ramp, ninety images were acquired, with image acquisition every twenty seconds. This resulted in a total imaging duration of thirty minutes. The temperature of the live cell imaging solution at each image acquisition was manually recorded. Importantly, although the temperature bath was set from 15°C to 60°C, the actual measured temperature of the live cell imaging solution in the 35mm dish changed from approximately 15°C to 42°C during the 30- minute period of image acquisition. Critical Points and Comments: Accurately measuring the temperature of the live cell imaging solution inside the 35mm dish is one of the most challenging and important aspects of 82 this entire method. It’s important that experimenters never rely alone on their temperature controllers (e.g. temperature baths, hotplates, heating stages, etc.) to determine the temperature of the solution inside the imaging dish. Due to thermal lag and loss of heat to the ambient surroundings, the temperature of the imaging solution during temperature ramps will almost always be different than the temperature displayed on the temperature controller. This leads to two important points 1) In order to obtain accurate measurements, a temperature probe must be directly submerged in the imaging solution. 2) In order to get the solution to a certain temperature, 37°C for example, the temperature controller will typically have to be set to a higher point than 37°C. We chose a temperature ramp from 15°C to 60°C as this allowed us to heat 2.5 ml of imaging solution from approximately 15°C to 42°C within thirty minutes. However, as other labs may use different equipment or volumes of imaging solution, we recommend that experimenters empirically determine the appropriate end points for their temperature ramps. 83 Figure 4.3. Equipment configuration during live cell imaging. To rapidly heat live cells during imaging, transfected cells are cultured on a 35mm glass bottom MatTek dish which is placed on a Pecon microscope temperature stage attached to a circulating bath. To accurately measure the temperature of the imaging solution, a special lid is used that allows a temperature probe to be directly inserted into the imaging solution inside the glass bottom dish. A) Side view of experimental setup. Note that the Pecon stage is placed inside the Zeiss piezo stage, which assists in fine-tuning definite focus during imaging. B) Arial view of experimental setup. The attached circulating bath hoses are seen on the back-right side of the temperature stage. C) The circulating bath is housed on the ground underneath the anti-vibration table. Note the use of a bucket and the covering placed around the hoses on the table which prevent condensation from damaging sensitive electrical equipment. 84 4.4.4 Image analysis 4.4.4.1 Blinding and randomizing imaging data All image analysis was performed using FIJI. 81 It is recommended that any image analysis requiring human intervention, whether it involves directly identifying a cell’s Tt , drawing ROIs, or selecting thresholds for subsequent microdomain analysis, be performed on blinded and randomized imaging data. To blind and randomize our datasets, we developed a Java program which copies a specified directory of images, renames each image with a random three-digit number, and then places the renamed images in a new directory, along with a key which matches each original filename with its corresponding three digit number. This program is called File Randomizer and is part of a larger plugin named Simple Image Analysis Library (or SIAL). SIAL also contains plugins to assist in phenotype scoring and ROI recording. To download SIAL, open FIJI, go to “Help > Update…” and then update FIJI. After FIJI is finished downloading all updates, a window named “ImageJ Updater” will open. Select “Manage Update Sites > Add update site” and add this url: https://sites.imagej.net/D-tear/ Be sure to check the box next to this update site to ensure the FIJI adds SIAL to your FIJI Plugins folder. Select “Close > Apply changes”. FIJI will download SIAL.jar and associated dependencies. After successfully updating, FIJI will then ask to be closed and restarted. After doing this, SIAL can be accessed via the Plugins tab in FIJI. Note that SIAL will usually be installed towards the bottom of the available FIJI plugins. To randomize a directory of images, select “SIAL > File Randomizer” and type in the extension of the imaging data (i.e. tif, czi, ori), and then select the input directory where the images are located and an output directory where the randomly renamed images should be placed. Links to YouTube tutorials covering the installation of SIAL and the use of its File Randomizer, PhenoScoreKeeper, and ROI Recorder programs are 85 included in the supplementary materials. Code for SIAL can be found on GitHub: https://doi.org/10.5281/zenodo.3698249 Critical Points and Comments: After randomly renaming images, and before beginning any analysis, ensure that the randomly renamed images are arranged by name in your chosen output directory (this is the default in the Windows and Mac operating systems we’ve tested). Arranging the randomly renamed images by name ensures that the images are uniformly shuffled. Once the randomized images are arranged by name, we recommend proceeding with subsequent analysis by opening up each image in the order that they are listed in the output directory. In contrast, if the randomized images are arranged by date modified or size (or any other metric), there is no guarantee that the images are uniformly shuffled after randomization. Note that SIAL’s PhenoScoreKeeper and ROI Recorder plugins automatically open up images in numeric order, so these programs ensure that the images are presented in a uniformly shuffled fashion after randomization. 4.4.4.2 Visual determination of intracellular Tt After using the File Randomizer to randomize time lapse image stacks of cells undergoing a temperature ramp, the transition temperature of any given cell can be determined through visual inspection by identifying the first frame at which that cell produced fluorescent microdomains. Examples of fluorescent microdomain formation are seen in Figure 4.5. Critical Points and Comments: As previously mentioned, due to cell to cell variability in GFP-ELP protein expression, cells across a given study will commonly exhibit large differences in fluorescence intensity. Accordingly, during visual inspection to determine intracellular Tt, 86 experimenters will likely need to adjust image contrast to ensure that they are not neglecting cells with weaker signal from their analysis. 4.4.4.3 Microdomain analysis All microdomain analysis was performed on blinded and randomized data sets. To analyze each transitioned cell’s number of microdomain and average microdomain size, we selected either the final image from the temperature ramp for that cell, or the last clear image, if the final image was out of focus or otherwise unclear. To measure microdomain size and number per cell, we used the Analyze particles program within FIJI. This program can be accessed via “Analyze > Analyze particles…”. Note that in order to use the analyze particles program, experimenters will have to threshold their image and then convert it into a binary image: “Image > Adjust > Threshold”. An appropriate threshold can then either be manually adjusted or one of the built-in thresholding algorithms can be selected. In our work we manually adjusted the threshold to identify a setting which agreed with visual inspection. However, in our experience FIJI’s default thresholding algorithm or Otsu also typically perform very well. Once a threshold is selected, the image can be converted into a binary image by pressing apply. After thresholding, microdomain measurements were obtained on the binary images for individual cells by using the ROI manager ( “Analyze > Tools > ROI Manager…”) to outline individual cells. Analyze particles… measurements were restricted to individual cells by selecting that cell’s ROI in the ROI manager. Critical Points and Comments: Proper image analysis begins with good image acquisition, and the ability to accurately threshold and separate foreground from background largely depends on high SNR, ideally with empty background exhibiting near zero intensity levels. Accordingly, 87 experimenters should optimize image acquisition to obtain the highest SNR possible without toxic photobleaching. In cases where noise has contaminated an image, the use of filters can improve thresholding performance. Although we didn’t find the use of filters or pre-processing necessary in this dataset, in our experience, median filters perform well for removing “salt and pepper” noise from images. In FIJI, filters can be accessed via “Process > Filters”. The addition or subtraction of filtered images from original images can also improve thresholding performance. These types of operations can be performed in FIJI via “Process > Image Calculator”. Background intensity levels can also be directly subtracted from images by using the Math tab in FIJI, accessed via “Process > Math > Subtract…”. Even in images with high SNR, we recommend that experimenters restrict analyses to features larger than one pixel (or the equivalent size in microns 2 ). This size option can be directly set in the “Analyze Particles…” menu. In our images, we restricted analyses to particles >= 0.044micron 2 . This criterion was chosen as our images were 338.29 by 338.29 microns at 2272x2272 pixels, which equals a scale of 0.02217microns 2 /pixels 2 . To ensure selection was conservatively restricted to more than square pixel, 0.02217 was multiplied by 2, resulting in 0.044micron 2 . 4.4.4.4 Background corrected integrated density (BCID) To collect measurements of background corrected integrated density (BCID), clear images of cells from below their transition temperature were analyzed by first collecting at least three ROIs from empty regions of that image. The average of the mean intensities of these empty regions was computed to determine the overall mean intensity of the background of that image. This value was then subtracted from the image by using “Process > Math > Subtract…”. If the overall mean intensity of the background was less than 1, nothing was subtracted from the image. After 88 background was subtracted from the image, ROIs were drawn to outline individual cells from that image and measurements of integrated density were recorded. Critical Comments and Suggestions: Background corrected integrated density is also sometimes called corrected total cell fluorescence (CTCF), and it is a commonly applied method in quantitative fluorescent microscopy. 83-86 We previously developed an image processing pipeline to help partially automate CTCF measurements from images. The code and instructions for this pipeline are freely available. 123 4.4.4.5 Standard deviation determination of intracellular Tt In addition to visually determining the temperature at which intracellular ELPs phase transitioned, we also examined an alternative technique which relies on measuring the standard deviation of the pixel values inside cells as they undergo the temperature ramp. The rationale for this technique is based on the observation that below the Tt, GFP-V60 fluorescence is homogenous throughout the body of the cell. Accordingly, the standard deviation of the pixel values inside the cell are low. However, as temperature is increased and GFP-V60 begins to phase separate, fluorescence becomes concentrated into discrete areas of the cell (Figure 4.4). This produces a shift upwards in the standard deviation, as some areas inside the cell become much darker while others become much brighter. To explore this approach, time lapse image stacks were first stabilized by using the Template Matching plugin (This plugin can be installed by adding this URL to the ImageJ updater: http://sites.imagej.net/Template_Matching/ ). This step is crucial as temperature changes often induce substantial lateral drift in the focal plane as the 35mm dish contracts and expands. After stabilizing the stacks with this plugin, ROIs were then drawn around each cell at the beginning of 89 the temperature ramp (i.e. the first frame of the image stack). To automatically collect standard deviation measurements of each cell/ROI on each image in the stack, we wrote an ImageJ macro which reapplies these initially drawn ROIs on each subsequent image on the stack. This macro is called measureROIsonStack.ijm and it is provided in the supplementary materials. To analyze the standard deviation data and determine the temperature at which phase transition occurred, we wrote an R function named batch_plot which calculates the baseline standard deviation from the initial <n> images of the temperature ramp (this <n> is specified by the user, but the default is 20); batch_plot then determines the temperature at which the standard deviation shifts above a specified threshold beyond this baseline standard deviation. This threshold is again specified by the user, but the default is 3, reflecting three standard deviations away from the baseline standard deviation (Another suitable threshold could be 2, reflecting two standard deviations away from the baseline, which mimics the 0.05 threshold often used for p-values. Relatedly, users could also select higher thresholds, 5 for example, to account for the increased probability of false positives due to testing multiple cells). The temperature/image slice at which the standard deviation shifts above the specified threshold is recorded as the intracellular Tt. In addition to recording these measurements for each cell, batch_plot also plots the standard deviation data and indicates where on the graph the threshold was passed (Figure 4.4). The code for batch_plot and some example data is provided in an R notebook on GitHub: https://doi.org/10.5281/zenodo.3950576 The standard deviation approach is attractive because it relies on a quantifiable feature of the image. However, in practice we found this method to be more time consuming, unreliable, and tedious than visually scoring blinded and randomized data. For example, scoring an entire dataset of 56 cells could be accomplished in approximately one workday. In contrast, collecting standard deviation measurements took approximately three workdays. This increase in time was mainly due 90 to the fact that hand drawn ROIs did not closely follow the cells during the time lapse imaging due to lateral drift. Even after stabilization with the Template Matching plugin, it was not uncommon for several frames, out of the approximately 90 frames in a single time lapse image stack, to be shifted with respect to the rest of the frames. These shifted frames meant that the ROIs were no longer measuring the cells, but instead were measuring background, or partial background, which produced false positives as the standard deviation measurements would shift dramatically. As a workaround, the ROIs were manually redrawn at the problematic frames, but this task was time consuming and tedious. In contrast, the human eye is not fooled by lateral drift and can easily track the same cell even if it shifts around during imaging. For these reasons, we recommend visual determination of intracellular Tt from blinded and randomized data, although the use of standard deviation data could be used to help improve accuracy in images where visual determination is ambiguous. 91 Figure 4.4. The standard deviation of intracellular pixel values during a temperature ramp provides an alternative method for determination of intracellular transition temperature. Each circle is the standard deviation of the pixels inside a single GFP-V60 transfected cell at that image slice and corresponding temperature. The “baseline standard deviation” is averaged from the first twenty points, and then the data is normalized to a z-score by using the mean and standard deviation of these twenty points. The horizontal red dashed line is placed at three standard deviations away from the baseline. The vertical red line indicates the first data point which crosses the threshold, and this is the estimated T t for the cell. Both the number of baseline data points and the threshold can be specified by the user in the batch_plot function. The title of the plot, “200225.GFP-V60.C”, is the ID for this particular cell (cell IDs must be provided in the raw data and can be anything, as long as they are unique to each cell in the dataset). Using these cell IDs, the batch_plot function produces transition temperatures and plots for each cell in the dataset. 92 4.4.5 Fixed Cell Imaging to confirm GFP-V60 colocalization with non-fluorescent ELP fusion 4.4.5.1 Transfection Fixed cell immunofluorescence was used to confirm GFP-V60 colocalization with non- fluorescent CAV1-V96. The steps are similar to transfection for live cell imaging, except that 12- well plates were used instead of 35mm MatTek dishes. Before transfection, coverslips were placed individually into the wells of 12-well plate. Coverslips were then immersed with 1ml of Poly-D- lysine solution and incubated for five minutes and then washed with dPBS, as described above. 1ml of resuspended cells at a density of 0.5 x 10 6 /ml were then seeded into each well before being placed back into the 37C, 5% CO2 incubator for 24hrs. After the 24hrs, cells were transfected with the Lipofectamine kit. The steps are the same as for live cell imaging transfection, except that the volumes of transfection reagents and DNA were downscaled. The recipe for a single well involved 1) 1µg of total plasmid DNA (i.e. either 1µg of GFP-V60 DNA for single transfected wells, or 500ng of GFP-V60 DNA + 500ng of CAV1-V96 DNA for dual transfected wells) 2) transfection mix consisting of 50 µl of diluted DNA/P300 mixture (containing 1µg total of plasmid DNA, the P3000 Reagent at a concentration of 2µg/µl of DNA, and Opti-MEM up to 50µl) + 50µl of the Lipofectamine mixture (47.5µl of Opti-MEM + 2.5µl of Lipofectamine 3000). Each individual mixture was thoroughly mixed before being added together to create the 100µl master mix. This master mix was then incubated for 15 minutes at room temperature. During this time, the cell seeded MatTek dishes were removed from the incubator, washed one time with room temperature dPBS, and were then replenished with 900µl of Opti-MEM. After the 15-minute incubation, the 100µl master mix was pipetted drop-by-drop into the 900µl of Opti-MEM media. Cells were then placed in a 30˚C incubator with 5% CO2 and were incubated for 72 hrs before subsequent steps. 93 4.4.5.2 Temperature incubation, fixation, and immunostaining After the 72hr incubation, the 12-well plates were removed the incubator and placed on ice for 50 mins before being a placed a heating block (IC25, Torrey Pines Scientific, Carlsbad, CA) for 50 mins at 37°C. Media was then removed from each well and the cells were fixed with a 4% solution of room temperature 4% PFA and incubated on the lab bench for fifteen minutes. After this incubation, cells were rinsed with 50mM NH4Cl (in PBS) for 5 min before being washed with 1X PBS for three minutes, three times (3min x 3) on an orbital shaker. Note that all PBS washing steps were performed on an orbital shaker. Cells were then permeabilized with 0.1% Triton-X (in PBS) and then washed with 1X PBS (3min x 3) before blocking with a 90-minute room temperature incubation in 1% BSA (dissolved in PBS). Note that all primary and secondary antibodies were also diluted in this 1% BSA solution. After the incubation, coverslips were removed the wells and placed face down on a 35µl droplet of primary antibody solution, which was placed on parafilm. The primary antibodies used were myc-tag mouse mAb [1:1000] for CAV1-V96, and rabbit GFP pAb antibody [1:500] for GFP-V60. Primary antibodies were incubated overnight at 4°C, then coverslips were removed from the parafilm, flipped over, and placed back into the 12-well plate before being washed with 1X PBS (5min x 5). Coverslips were then removed the wells and placed faced down on a 35ul droplet of secondary antibody solution, which was placed on parafilm. The secondary antibodies used were Chicken anti-Mouse IgG Alexa Fluor 647 [1:500] and Donkey anti-Rabbit IgG Alexa Fluor 488 [1:500]. Secondary antibodies were incubated at room temperature for 1hr, then coverslips were removed from the parafilm, flipped over, and placed back into the 12-well plate before being incubated with DAPI for 5 mins at room temperature on an orbital shaker. Cells were than washed with 1X PBS (5min 94 x 6) and were then mounted on a glass slide, using ProLong™ Glass Antifade Mountant. Slides were cured for 48hrs in the dark at room temperature before imaging. Critical Comments and Suggestions: “No primary-antibody controls”, consisting of labeling with secondary antibody without primary antibody incubation, are necessary to determine non-specific antibody binding and subsequent imaging settings. We also recommend the use of “bleed through controls”, consisting of only one secondary antibody, to determine non-specific bleed through of each fluorophore. The exact immunofluorescence (IF) procedure outlined is likely not critical, as many labs have their own successful routines for IF staining. However, we do recommend that experimenters pre-incubate on ice to resolubilize any ELPs which may have assembled during the 72hr transfection incubation. Relatedly, we recommend that subsequent incubation steps, after the ice pre-incubation, be carried out by placing samples directly in contact with a heated surface. This ensures more direct transfer of heat to the samples. Also ensure that the samples are protected from light once the fluorescent secondary antibodies are applied. 4.4.5.3 Imaging Slides were imaged on an LSM800 laser scanning confocal microscope equipped with an airy scan detector. All images were acquired with the Airy scan acquisition mode using a plan Apochromat 63x/1.40 numerical aperture Oil DIC M27 objective. Immersol 518F was used as the immersion media. AF647, AF488, and DAPI channels were scanned sequentially. To excite AF647/CAV1-V96 the 640nm laser line was used and wavelengths from 647nm-700nm were collected. To excite AF488/GFP-V60, the 488nm laser line was used and wavelengths from 487nm-533nm were collected. To excite DAPI, the 405nm laser line was used and wavelengths of 400nm-508nm were collected. All images were 16-bit with a pixel dwell time of 3.14µs, a lateral sampling of 0.029µm, and no averaging. 95 4.5 VALIDATION AND EXPECTED RESULTS 4.5.1 Visual determination of intracellular transition temperature of CAV1-ELP library For the CAV1-ELP library, live cell imaging confirmed that CAV1-V96 + GFP-V60 dual transfection produced a substantial downward shift in intracellular Tt compared to GFP-V60 single transfection as well as CAV1-A96 + GFP-V60 dual transfection (Figures 4.5 and 4.6, Table 4.1). Intracellular transition temperatures for each cell in the study were determined by eye, after image stacks were blinded and randomized using the File Randomizer program within the FIJI plugin SIAL. GFP-V60 displayed a mean Tt of 35.5°C, while CAV1-V96 + GFP-V60 displayed a mean Tt of 29.2°C before filtering and 28.8°C after filtering. In contrast, dual transfection of temperature insensitive CAV1-A96 + GFP-V60 displayed a mean Tt of 37.3°C. This reported Tt for CAV1- V96 + GFP-V60 is several degrees lower than our previous reported value of 35.9°C. 1 This difference is almost certainly due to the fact that, compared with our previous imaging equipment, our current imaging system provided superior resolution, SNR, automatic image acquisition every 20 seconds, and automatic focus stabilization. These improvements (especially the automated functions) allowed for more rapid and consistent acquisition of high-quality images than what was possible with the older manually operated imaging equipment we previously employed. Image analysis with the Analyze particles program within FIJI confirmed that CAV1-V96 + GFP-V60 produced microdomains which were larger and fewer number than those produced by GFP-V60 single transfection or CAV1-A96 + GFP-V60 (Figure 4.6B). The max microdomain 96 size recorded for all GFP-V60 single transfected cells was 1.81µm 2 , and this parameter was used as the filter to separate truly dual transfected cells from those which were only (or predominantly) expressing GFP-V60 (Figure 4.6B, C). 97 Figure 4.5. Cells transfected with CAV1-V96 + GFP-V60 exhibit a reduced transition temperature and large microdomains compared to transfection with GFP-V60 alone or CAV1-A96 + GFP-V60. Images were taken from live cells undergoing a temperature ramp. CAV1-V96 is temperature sensitive and its self-assembly induces co-transfected GFP-V60 to assemble together with it. In contrast, GFP-V60 single transfection phase separates at a higher temperature and also produces cytosolic microdomains which are smaller and more numerous than the membrane anchored microdomains of CAV1-V96 + GFP-V60. CAV1-A96 (middle column) is not temperature sensitive and does not produce the characteristic large microdomains of CAV1-V96 when co-transfected with GFP-V60. Instead, phase separation of GFP-V60 occurs but appears to be largely restricted to the plasma membrane, where CAV1-A96 is located, rather than being distributed through the cytosol. Although the only visible fluorescence is from GFP-V60, the downward shift of transition temperature combined with the dramatic differences in microdomain size and number allow experimenters to filter truly dual transfected CAV1-V96 + GFP-V60 cells from those expressing GFP-V60 alone. Scale bar is 10µm. 98 Figure 4.6. Changes in intracellular transition temperature (T t), microdomain size, and microdomain number can be quantified from live cells transfected with different ELP fusions. In each panel, a circle is an individual cell. A) The raw transition temperatures for all cells in the study. CAV1-V96 + GFP-V60 displays a significant difference in T t compared to CAV1-A96 + GFP-V60 or GFP-V60 alone: p- value < 0.00001 in both comparisons. In contrast, GFP-V60 vs. CAV1-A96 + GFP-V60 is not significant: p-value = 0.29. The three red dots are cells from the CAV1-V96 + GFP-V60 group which exhibited average microdomain sizes consistent with GFP-V60 single transfection, which is quantified in panel B. B) Different ELP fusions display differences in average microdomain size and number per cell. Although CAV1-A96 + GFP-V60 and GFP-V60 alone display numerous small microdomains, CAV1- V96 + GFP-V60 tends to produce only a few large microdomains per cell. The black dotted line represents the maximum average microdomain size recorded among all GFP-V60 cells (1.81µm 2 ). This size was used as the filter to separate dual transfected CAV1-V96 + GFP-V60 cells from those which may only be expressing GFP-V60. Note the three purple circles below the dotted line. These correspond to the three red circles in panel A and are the CAV1-V96 + GFP-V60 cells which did not pass the filter. C) Transition temperature data after filtering out the three CAV1-V96 + GFP-V60 cells which did not pass the microdomain size filter. CAV1-V96 + GFP-V60 still displays a significant difference in T t compared to CAV1-A96 + GFP-V60 or GFP-V60 alone: p-value < 0.00001 in both comparisons. In contrast, GFP-V60 vs. CAV1-A96 + GFP-V60 is not significant (p-value was raised from 0.29 to 0.32, due to a larger residual mean squared error from removing the three CAV1-V96 + GFP-V60 cells from the dataset). All p-values were calculated from a Tukey’s post-hoc test following a statistically significant one-way ANOVA. 99 4.5.2 The effect of intracellular GFP-V60 concentration on the intracellular Tt of CAV1-ELP fusion proteins Measurements of background corrected integrated density (BCID), a proxy for GFP-V60 intracellular concentration, were collected to determine the effect of GFP-V60 concentration on intracellular transition temperature (Figure 4.7). Linear regressions of Intracellular Tt ~ log10(BCID) for both the GFP-V60 single transfection and CAV1-A96 + GFP-V60 dual transfection groups displayed a significant linear trend of decreasing Tt with increasing log10(BCID) (i.e. increasing GFP-V60 concentration). For GFP-V60 single transfection, a 10-fold increase in BCID resulted in average decrease in Tt of 3.89°C (p-value = 3.52e-08, R 2 = 0.68). For CAV1-A96 + GFP-V60, a 10-fold increase in BCID resulted in average decrease in Tt of 3.98°C (p-value = 1.92e-07, R 2 = 0.81). In contrast, CAV1-V96 + GFP-V60 did not display a significant linear trend with log10(BCID) of GFP-V60 signal on either the unfiltered or filtered data (unfiltered p-value = 0.92, unfiltered R 2 = -0.12; filtered p-value = 0.35, filtered R 2 = 0.01). However, the BCID measurements in Figure 4.7 do reveal that given the same the same intracellular concentration of GFP-V60 (i.e. the same BCID), CAV1-V96 + GFP-V60 cells will typically display a Tt several degrees lower than those seen in GFP-V60 or CAV1-A96 + GFP- V60 cells. 100 Figure 4.7. Intracellular transition temperature (T t) and GFP-V60 integrated density follow a log- linear relationship [CAV1-ELP dataset]. Each dot is a cell from the same data shown in Figure 4.6. The purple diamonds are the three CAV1-V96 + GFP-V60 cells which were filtered out based on their average microdomain size. The x-axis is the background corrected integrated density (BCID) of GFP-V60 signal from that cell. Both CAV1-A96 + GFP-V60 (p-value = 1.92e-07 ***, R 2 = 0.81) and GFP-V60 alone (p- value = 3.52e-08 ***, R 2 = 0.68) display a significant linear trend in decreasing T t with increasing log 10(BCID). In contrast, CAV1-V96 + GFP-V60 does not display a significant trend, either on the unfiltered data or the filtered data (unfiltered p-value = 0.92, unfiltered R 2 = -0.12; filtered p-value = 0.35, filtered R 2 = 0.01). 101 4.5.3 Fixed cell imaging to confirm GFP-V60 colocalization with non-fluorescent CAV1-V96 To confirm that GFP-V60 does indeed colocalize with phase separated CAV1-V96, transfected cells were pre-incubated on ice to resolubilize any assembled ELPs and were then incubated at 37°C for 50 minutes before fixation and immunostaining with antibodies for myc (CAV1-V96) and GFP (GFP-V60). As expected, phase transitioned CAV1-V96 assembled large microdomains which overlapped with GFP-V60 signal (Figure 4.8). In contrast, single transfected GFP-V60 cells displayed numerous small microdomains throughout the cytosol and nucleus. Figure 4.8. Indirect staining in fixed cells confirms that CAV1-V96 co-assembles with GFP-V60. Cells were either transfected with CAV1-V96 + GFP-V60 (top row) or GFP-V60 alone (bottom row). 72hrs after transfection, cells were pre-incubated for 50 minutes at 4°C and were then incubated at 37°C for 50 minutes before fixation and immunostaining. In dual transfected cells, CAV1-V96 self-assembles to form large intracellular microdomains which overlap with co-assembled GFP-V60. In contrast, GFP-V60 alone transfection assembles small numerous microdomains throughout the cytosol and nucleus. CAV1-V96 was detected with a myc antibody. GFP-V60 was detected with a GFP antibody. Scale bar is 10um. 102 4.5.4 Visual determination of intracellular transition temperature of ELP-CLC library The overall interpretation and analysis of the ELP-CLC dataset was more complicated than for the CAV1-ELP dataset. This is because i) as a preliminary dataset, cells were collected under a variety of image acquisition parameters, with laser power, master gain, and pixel dwell time not being consistent across all images. ii) dual and single transfected treatments displayed no actionable difference sin microdomain size or number. This prevented any filtering like that applied with the CAV1-ELP library to separate dual from single transfected cells. To increase statistical power, GFP-V60 cells from the CAV1-ELP and ELP-CLC datasets were combined as they did not display a significant difference in Tt (p-value = 0.15), even though they were imaged with different acquisition parameters and by different personnel. For the ELP- CLC library, live cell imaging confirmed that V96-CLC + GFP-V60 dual transfection produced a significant downward shift in intracellular Tt compared to the combined GFP-V60 single transfection group (Figure 4.10, Table 4.1). The GFP-V60 data from the ELP-CLC dataset displayed a mean Tt of 36.6°C while the combined GFP-V60 data from the CAV1-ELP and ELP- CLC datasets displayed a mean Tt of 36.0°C. In contrast, V96-CLC + GFP-V60 displayed a mean Tt of 33.7°C. Surprisingly, A96-CLC + GFP-V60 displayed a mean Tt of 34.3°C, nearly two degrees lower than GFP-V60, although this difference was not significant (p = 0.23). A96-CLC + GFP-V60 vs. V96-CLC + GFP-V60 also did not produce a significant p-value (p = 0.89). The reasons underlying these results are not clear and warrant further investigation with a larger sample size and verification with indirect immunofluorescence in fixed cells to confirm any association of the A96-CLC and V96-CLC constructs with GFP-V60. Although the indirect 103 immunofluorescence experiments should be straightforward, increasing the sample size for the A96-CLC + GFP-V60 group may be especially challenging, as we found dual transfection of A96- CLC + GFP-V60 led to low overall transfection efficiencies. In contrast to the CAV1-ELP library (and EGFR-ELP library), dual transfected treatments for the ELP-CLC library did not produce microdomains which varied in size or number compared to single transfected GFP-V60 (Figures 4.9 and 4.10). This result may be due to the fact that like GFP-ELPs, CLC-ELPs are cytosolic proteins. The co-assembly of these two cytosolic ELP fusion proteins simply produces cytosolic microdomains which are similar to GFP-V60 microdomains. In contrast, CAV1-ELPs and EGFR-ELPs are membrane bound proteins. The co-assembly of GFP-V60 with membrane embedded ELPs is likely responsible for the large membrane localized fluorescent microdomains seen with these libraries that is so distinct from the cytosolic microdomains produced from GFP-V60 alone. 104 Figure 4.9. Cells transfected with ELP-CLC + GFP-V60 display microdomains which are similar to GFP-V60 alone. Images were taken from live cells undergoing a temperature ramp. Note that while the GFP-V60 cell (third column) does display microdomains before the V96-CLC + GFP-V60 cells (first column), on average the GFP-V60 group displayed a higher T t. Scale bar is 10µm. 105 Figure 4.10. Changes in intracellular transition temperature (T t) can be quantified from live cells transfected with different ELP-CLC fusions. However, note that cytosolic ELP-CLC fusion proteins do not permit filtering based on microdomain size or number, as dual-transfected and single transfected GFP- V60 treatments produce microdomains with similar characteristics. Note that the GFP-V60 data contains all cells from both the CAV1-ELP and CLC-ELP library. These datasets were combined as they did not display a significant difference in T t (p-value = 0.15), even though they were imaged with different acquisition parameters and by different personnel. A) The raw transition temperatures for all cells in the study. V96-CLC + GFP-V60 displays a significant difference in T t compared to GFP-V60 alone: p-value = 0.015. In contrast, GFP-V60 vs. A96-CLC + GFP-V60 did not display a significant: p-value = 0.23. However, A96-CLC + GFP-V60 vs. V96-CLC + GFP-V60 also did not display a significant difference: p- value = 0.89. B) In contrast to CAV1-V96 + GFP-V60, cells transfected with cytosolic ELP-CLCs + GFP- V60 do not display filterable differences in average microdomain size and number per cell compared to cells transfected with GFP-V60 alone. The black dotted line represents the maximum average microdomain size recorded among all GFP-V60 cells from both the CAV1-ELP and ELP-CLC datasets (2.91µm 2 ). 106 4.5.5 The effect of intracellular GFP-V60 concentration on the intracellular Tt of ELP-CLC fusion proteins Figure 4.11 displays the linear relationship between log10(BCID) and Tt for the ELP-CLC data and the GFP-V60 data from both the ELP-CLC and CAV1-ELP datasets. Note that these two GFP-V60 groups have different BCID profiles but relatively equal transition temperatures. The horizontal shift in BCID is due to the fact that different imaging acquisition parameters were used between the two groups. The GFP-V60 cells from the ELP-CLC dataset displayed a non- significant linear trend in decreasing Tt with increasing log10(BCID): p-value = 0.0617, R2 = 0.10. However, when Date was added as a covariate [i.e. Tt ~ Date + log10(BCID)], log10(BCID) became a significant predictor: p-value = 0.00103 **, R 2 = 0.58. Date was a significant explanatory covariate because imaging acquisition parameters were changed during the course of this dataset. The V96-CLC + GFP-V60 group also did not display a significant trend with log10(BCID): p-value = 0.29, R 2 = 0.019. Including Date as a covariate improved the R 2 and lowered the p-value for log10(BCID) but it was still not significant: p-value = 0.10, R 2 = 0.72. However, Figure 4.11 does reveal that given the same log10(BCID), V96-CLC + GFP-V60 cells tend to display lower transition temperatures than A96-CLC + GFP-V60 or GFP-V60 alone cells. A96-CLC + GFP-V60 displayed a significant p-value for log10(BCID) even without adding Date as a covariate: p-value = 0.023*, R 2 = 0.61. This result can be attributed to the fact that this group was imaged with constant imaging acquisition parameters over all dates. 107 Figure 4.11. Intracellular transition temperature (T t) and GFP-V60 integrated density follow a log- linear relationship [ELP-CLC dataset]. Each dot is a cell from the same data shown in Figure 4.10. The GFP-V60 cells from ELP-CLC dataset are shown in grey, while the GFP-V60 cells from the CAV1-ELP dataset, which were analyzed in Figure 4.7, are shown in black. Note that these two groups have different BCID profiles but relatively equal transition temperatures. This horizontal shift is due to the fact that different imaging acquisition parameters were used between the two groups. The GFP-V60 cells from the ELP-CLC dataset displayed a non-significant linear trend in decreasing T t with increasing log 10(BCID): p- value = 0.0617, R 2 = 0.10. However, when Date was added as a covariate, T t ~ Date + log 10(BCID), log 10(BCID) became a significant predictor: p-value = 0.00103 **, R 2 = 0.58. This is because imaging acquisition parameters were changed during the course of this dataset. The V96-CLC + GFP-V60 group also did not display a significant trend with log 10(BCID): p-value = 0.29, R 2 = 0.019. Including Date as a covariate improved the R 2 and lowered the p-value for log 10(BCID) but it was still not significant: p-value = 0.10, R 2 = 0.72. In contrast, A96-CLC + GFP-V60 displayed a significant p-value for log 10(BCID) even without adding Date as a covariate: p-value = 0.023*, R 2 = 0.61. The insignificance of Date in the model is due to the fact that this group was imaged with non-changing imaging acquisition parameters. 108 4.6 CONCLUSION In this chapter, I presented a simple live cell imaging technique to estimate the intracellular Tt of non-fluorescent ELP fusion proteins. As a case study, I primarily focused on the recently developed CAV1-ELP library but also presented some preliminary data from the ELP-CLC library. The intracellular Tt of a temperature sensitive fusion, such as CAV1-V96, can be estimated by co-transfection with a less temperature sensitive fluorescent ELP fusion, like GFP-V60. Intracellular assembly can then be visualized in live cells through the co-assembly of the non- fluorescent and fluorescent ELP fusion proteins. Critically, if the two different ELP species exhibit differences in temperature sensitivity, the resulting mixture of ELPs exhibit a decreased Tt which corresponds to the Tt of the more temperature sensitive ELP. In addition to changes in Tt, co- assembled ELP microdomains may also exhibit pronounced differences in size or number compared to single transfected treatments. However, these particular differences may be restricted to membrane bound fusions like CAV1-ELPs and EGFR-ELPs, not cytosolic fusions such as the ELP-CLCs. In addition to providing a detailed protocol, I also developed an easily installed FIJI plugin named SIAL, which contains programs for image randomization and blinding, phenotype scoring, and ROI selection. These tasks are important parts of the protocol detailed here but are also commonly employed in other image analysis workflows. 109 Chapter 5 Mitochondrial DNA Deletions and Copy Number in Whole Genome Sequencing (WGS) Data: Analyses of Aging and Parkinson’s Disease 5.1 Abstract Mitochondria are membrane bound organelles which provide every nucleated cell in the body with ATP. An individual cell may contain dozens to thousands of mitochondria, depending on the energy demands of the tissue. In addition to producing ATP, mitochondria contain their own 16.6kb circular genome (mtDNA), which codes for the proteins involved in ATP production. Deletions in the mitochondrial genome, where hundreds to thousands of base pairs are missing, are correlated with age and a number of degenerative illnesses, such as Parkinson’s Disease, diabetes, macular degeneration, and liver disease. Although some deletions can be maternally inherited, many deletions occur spontaneously. For unknown reasons, mutant mtDNA often clonally expands and eventually outcompetes wild type mtDNA, eventually producing a cellular and potentially tissue level disruption in mitochondrial functions. Despite their importance, traditional methods for identifying deletions are biased and low throughput. In this chapter I describe a novel high-throughput method for identifying and quantifying mtDNA deletions from whole genome sequencing data obtained from the North American Brain Expression Consortium (NABEC). I report that in non-diseased donors, mtDNA deletions and mtDNA copy number are elevated in the frontal cortex compared to the cerebellum. In addition, I show that mtDNA copy number is elevated in the cerebellums of Parkinson’s Disease donors, compared to age matched non-diseased donors. These results yield insight on biological aging and add to the growing body of evidence linking mitochondria to Parkinson’s Disease. 110 5.2 Introduction Alterations in mitochondrial biology, including mtDNA deletions and changes in mtDNA copy number, are associated with biological aging and disease, including NASH/NAFLD, heart disease, macular degeneration, Alzheimer’s, and Parkinson’s Disease. 124-128 Although the link between mitochondrial function and tissue specific complex disease is not well understood, a unifying theme is that these tissues are mitochondria rich, and disease progression is often characterized by oxidative stress and damage to the tissue. While some mtDNA deletions can be maternally inherited, many deletions appear to occur spontaneously over the course of aging. 129-131 Because cells contain dozens to thousands of mitochondria, deleted mtDNA exists alongside wild type mtDNA in a condition known as heteroplasmy. The existence of deleted mutant mtDNA is not in of itself detrimental, as long as there is a sufficient pool of wild type mtDNA. However, for reasons that are not well understood mutant mtDNA often clonally expands and eventually outcompetes wild type mtDNA, thereby producing a biochemical defect within the cell. The reasons underlying this clonal expansion are unknown, but one reasonable hypothesis is that deleted mtDNA are smaller and thus can replicate faster. However, a variety of observational studies, elegant animal models, and computer simulations have shown this to be an unsatisfactory explanation. For example, by studying the skeletal muscle fibers from donors afflicted with mtDNA maintenance disorders, which produce mtDNA deletions, Campbell and colleagues determined that deletion size had no effect on the clonal expansion and prevalence of a deletion within the muscle fibers. 132 By studying two different C. elegans models, one harboring a large mtDNA deletion and the other harboring small deletion, Gitschlag et al. found that both mtDNA deletions exhibited remarkably similar 111 replication dynamics. 133 In contrast to these results, Fukui and colleagues found that large deletions accumulate faster than small deletions in the neurons of their transgenic mouse model which harbors restriction enzyme cut sites in its mtDNA. 134 These conflicting results could potentially be ascribed to the use of two vastly different approaches for studying mtDNA deletions, but they also may highlight species differences and how lifespan plays an important role in mitochondrial dynamics. 135 Indeed, mtDNA replication computer simulations performed by Kowald and colleagues show that larger mtDNA deletions do provide a replicative advantage, but only in long lived species like humans, not rodents or nematodes. 136 Adding to this complexity, Yoneda et al. reported that in cultured human cells even a single point mutation, not a deletion, exhibited a replicative advantage over wild type mtDNA. 137 Taken together, these conflicting results suggest that mtDNA replication dynamics are likely governed by a still an unknown mechanism. The traditional method for identifying mtDNA deletions is based on designing primers that flank a hypothesized deletion site. PCR amplification of this designated region is then followed by sanger sequencing to confirm the existence of a deletion. Due to its requirement of an a priori hypothesis of a deletion site and the additional requirement of successful PCR amplification of that region, the traditional method suffers from bias and is largely unsuitable for discovery, in addition to being low throughput. To circumvent these limitations, Hjelm and colleagues recently reported a novel high- throughput method, termed Splice Break, to identify and map mtDNA deletions. Splice Break employs long-range mtDNA PCR, next generation sequencing (NGS), and bioinformatics using an RNA-seq alignment algorithm to identify reads containing a deletion. 138 Simulations with 112 artificial sequencing reads and verification with the traditional method show that this new approach exhibits high specificity and sensitivity in identifying both previously reported and novel deletions. Although Splice Break yields significant improvements over traditional PCR and sanger sequencing, the potential to apply the bioinformatics portion of the technique to whole genome sequencing (WGS) to identify mtDNA deletions has of yet remained unexplored. The ability to identify deletions from WGS instead of using long range PCR (LR-PCR) of mtDNA offers several advantages: 1) LR-PCR is not necessary. In WGS all DNA, autosomal and mitochondrial, is retrieved and sequenced. So, the additional step of LR-PCR is unnecessary. 2) WGS may be less biased than LR-PCR. This is because WGS involves the collection of all DNA followed by fragmentation and the subsequent addition of universal adapters and PCR primers to the ends of all DNA fragments. In contrast, LR-PCR requires primers that bind specifically to a mtDNA sequence. In addition, LR-PCR may favor amplification of smaller mtDNA templates; these smaller mtDNA templates represent mtDNA with large deletions. Accordingly, results employing LR-PCR of mtDNA may overrepresent large deletions. In contrast, WGS involves fragmentation of all DNA into a uniform size, which mitigates the potential for PCR to favor smaller templates. 3) Calculating mtDNA copy number from NGS data requires information about autosomal sequencing coverage. WGS provides this information. In contrast, LR-PCR of mtDNA followed by NGS is not suitable for calculations of mtDNA copy number. 4) Due to the abundance of publicly available WGS datasets, wet lab may not even be required. In this chapter I report that Splice Break can identify mtDNA deletions from WGS data. By using WGS obtained from non-diseased frontal cortex and cerebellum tissue from the North 113 American Brain Expression Consortium (NABEC), I show that mtDNA deletions significantly increase with age in both tissues, with the frontal cortex accumulating deletions at a greater rate and also accumulating more overall unique deletions than the cerebellum. By mapping deletion start and end sites back to the mitochondrial genome, I report that most deletions fall within the major arc, consistent with previous literature. By using fastMitoCalc, a program for calculating mtDNA copy number from WGS, I report that frontal cortex has significantly higher mtDNA copy number than cerebellum. In addition, mtDNA copy number is elevated in the cerebellum of donors afflicted with Parkinson’s Disease (PD), compared to age matched non-diseased controls. Furthermore, cerebellum mtDNA copy is positively correlated with severity of PD associated pathology in the Substantia Nigra. These results yield insight on biological aging in different brain regions and add to the growing body of evidence linking mitochondria to Parkinson’s disease. 5.3 Materials and methods 5.3.1 General bioinformatics and statistical software • SAMtools version 1.9 (using htslib 1.9) • BEDtools version 2.28.0 • R version 3.6.1 • RStudio 5.3.2 Whole genome sequencing (WGS) WGS data, in compressed alignment map (CRAM) format, was obtained from the North American Brain Expression Consortium (NABEC). To generate this WGS data, paired-end 150bp FATSQ sequencing reads were aligned to the hg38 human reference genome using the BWA- MEM algorithm. 114 5.3.3 Splice-Break analysis of WGS To analyze mtDNA deletions in WGS using Splice Break, unmapped reads and mtDNA reads were pulled from the WGS sequencing data (i.e. the non-sorted CRAM files) and converted into individual BAM files using the samtools view command. These BAM files were then sorted using samtools sort. After sorting, the unmapped and mtDNA BAM files were merged using samtools merge. Finally, the merged BAM files were split into two FASTQ files (indicating read1 and read2, as samples were paired end) using bedtools bamtofastq. The Splice Break pipeline aligned these paired-end FASTQ files to the revised Cambridge Reference Sequence (rCRS) of human mitochondrial DNA (version NC_012920.1). The Splice Break pipeline is available from GitHub: https://github.com/brookehjelm/Splice-Break/ Note that the Splice Break pipeline uses MapSplice, an RNA-Seq splice junction algorithm, for read alignment and mtDNA deletion junction calling. 139 5.3.4 mtDNA copy number determination with fastMitoCalc Before analysis with fastMitoCalc, CRAM files were converted to BAM files and then sorted and indexed with SAMtools. To determine mtDNA copy number, I used a modified version of fastMitoCalc. This modification involved removing the sequencing coverage limit from inside fastMitoCalc.pl (fastMitoCalc.pl uses the samtools depth command to calculate sequencing coverage, and the default limit for samtools depth is 8,000X). In my modified version, the samtools depth command is set to samtools depth -d 0, which removes this 8,000X limit. This modification is necessary because in mitochondrial rich tissues, like the brain, mtDNA sequencing coverage will almost always exceed 8,000X. 115 5.3.5 Unique deletions per 10K coverage To calculate unique deletions per 10K coverage for each sample (Figures 5.1 and 5.4), the number of unique deletions for that sample were divided by the mtDNA benchmark coverage for that sample and this number was then multiplied by 10,000. An example calculation is shown below. Sample has three unique deletions and a benchmark coverage of 100,000: Unique Deletions (per 10K coverage) = (3/100,000)*10,000 = 0.3 5.3.6 Analysis of mtDNA deletions in small open reading frames (sORFs) encoding mitochondrial derived peptides Table 5.1 below contains the coordinates I used for the analysis of mtDNA deletions affecting mitochondrial derived peptides encoded by sORFs (Figure 5.4). For each sample, a deletion was identified as affecting a sORF if either its 5’ or 3’ end point matched those of a sORF or if the deletion entirely or partially removed any or all of the sORFs listed below (e.g. some large deletions could remove all 8 of these sORFs). Deletions occurring entirely between sORFs were not counted. Table 5.1. mtDNA (rCRS NC_012920.1) coordinates of sORFs in the mitochondrial genome Gene Symbol Position 5’-start 3’-end HN 2634-2797 2634 2797 SHLP1 2490-2561 2490 2561 SHLP2 2092-2170 2092 2170 SHLP3 1707-1821 1707 1821 SHLP4 2446-2524 2446 2524 SHLP5 2785-2856 2785 2856 SHLP6 2992-3051 2992 3051 MOTS-C 642-1607 642 1607 116 5.3.7 Statistics All statistics were calculated using R. General linear models (i.e. linear regressions and multiple regressions) were constructed using the lm function. These linear models were analyzed with both the summary and anova functions, and the reported p-values for predictors were taken from the summary function. All tests were two-sided and an alpha of 0.05 was pre-determined as the significance level. 5.4 Results 5.4.1 Description of NABEC control cohort Whole genome sequencing (WGS) data was obtained from four brain banks participating in the North American Brain Expression Consortium (NABEC). This data set contained 292 WGS samples, with 149 samples from the cerebellum and 143 from the frontal cortex. Samples were not paired (i.e. samples either came from the cerebellum of a donor or from the frontal cortex but not both). The data set ranges across the human life span (0.4 - 100yrs. old at time of death). Importantly, samples were obtained from non-diseased donors who were not diagnosed with neurogenerative diseases at the time of death. Because of these characteristics, this cohort served as a healthy control group to study how mtDNA changes with age. Summarized descriptions of the cohort are listed in tables 5.2 and 5.3. 117 Table 5.2. Summary of the NABEC control cohort across brain regions. Total Samples min (yrs.) mean (yrs.) median (yrs.) max (yrs.) All 292 0.4 50.1 45 100 Female 103 0.4 54.2 51 100 Male 189 0.9 47.9 43 97 Total Samples min (yrs.) mean (yrs.) median (yrs.) max (yrs.) Cerebellum 149 0.4 49.7 44 95 Female 61 0.4 49.2 44 95 Male 88 0.9 50.1 44 95 Total Samples min (yrs.) mean (yrs.) median (yrs.) max (yrs.) Frontal Cortex 143 1.0 50.5 46 100 Female 42 16.0 61.5 70 100 Male 101 1.0 46.0 43 97 Table 5.3. Summary of the NABEC control cohort sample sizes across brain banks. BANNER a UKY b UMARY c MIAMI d Total 63 32 174 23 Female 24 17 54 8 Male 39 15 120 15 a) Banner Health Brain and Tissue Bank, Sun City, AZ, USA b) University of Kentucky School of Medicine, Lexington, KY, USA c) University of Maryland School of Medicine, Baltimore, MD, USA d) University of Miami Brain Endowment Bank, Miami, FL, USA 118 5.4.2 mtDNA deletions are positively correlated with age in both Frontal Cortex and Cerebellum, with the Frontal Cortex accumulating deletions at a greater rate. Previous literature has shown a strong correlation between mtDNA deletions and age, especially in brain tissue. 140 However, many of these studies relied on the traditional low- throughput method of analyzing specific deletions. Accordingly, I determined whether Splice Break with WGS would reveal that unique deletions accumulate with age in either the frontal cortex or cerebellum in samples obtained from the NABEC control cohort. Figure 5.1 shows that unique deletions significantly increase with age in both tissues, with the frontal cortex (Figure 5.1A) accumulating deletions at a greater rate than the cerebellum (Figure 5.1B). 119 Figure 5.1. The number of unique mtDNA deletions increases with age in both Frontal Cortex and Cerebellum. Each dot represents a sample from the NABEC control cohort. A) Unique deletions per 10K coverage in the frontal cortex increase with age: age (regression coef., p-value) = (1.605e-01, 1.1e-10 ***). B) Unique deletions per 10K coverage in the cerebellum increase with age: age (regression coef., p-value) = (3.206e-02, 0.03217 *). P-values and regression coefficients were obtained from a multiple regression after correcting for mitochondrial benchmark coverage and potential batch effects contributed by different brain banks: Unique Deletions per 10K Coverage ~ Benchmark Coverage + Brain Bank + Age 120 5.4.3 Splice Break with WGS detected 1,728 unique deletions, with the great majority attributed to the frontal cortex. Splice Break applied to WGS from the NABEC control group reported 8,314 unique mtDNA deletions. After filtering to retain deletions that appeared in at least two samples (i.e. >= 2), this number was reduced to 1,728 mtDNA deletions. Out of this number, 1,369 (~79%) were unique to the frontal cortex. Only 6 deletions (~0.035%) were unique to the cerebellum, and 353 (~20%) were common to both tissues. Figure 5.2. After filtering, Splice Break with WGS identified 1,728 unique mtDNA deletions. Deletions were obtained from the 292 samples from the NABEC control cohort. Filtering retained deletions occurring in >= 2 samples. The great majority of these deletions are unique to the frontal cortex. 121 5.4.4 The majority of mtDNA deletions fall within the major arc of the mitochondrial genome. The mitochondrial genome contains two origins of replication, one for each strand. The origin of replication for the lighter stand is named OL, and likewise, the origin of replication for the heavier strand is named OH. These two origins of replication divide the circular mitochondrial genome into two arcs. The major arc spans the largest distance between OL and OH, and it is inside this approximately 8000bp region where the vast majority of reported mtDNA deletions have been discovered. 134,141 Accordingly, I sought to determine where the 1,728 filtered mtDNA deletions occurred. By mapping each deletion’s 5’-start-site and 3’-end-site, I discovered that most deletions do begin and end within the major arc, consistent with existing literature (Figure 5.3). The largest mtDNA gene NADH dehydrogenase 5 (ND5), was the site of most deletion start and end sites. However, the single most common deletion was a small 250bp deletion (15490bp – 15740bp) found within Cytochrome b (CYTB). This deletion was found in 97% of the 143 frontal cortex samples, 69% of the 149 cerebellum samples, and 83% of the overall 292 samples. Consistent with previous literature showing that most deletions are flanked by perfect (or near perfect) repeat sequences, this deletion was flanked by a perfect 7bp repeat of 5’-acctcct-3’. Interestingly, even though this deletion was the most common in the dataset it has not been reported in MitoBreak, a repository for discovered mtDNA deletions. 142 I also note that this deletion was not reported in the original Splice Break using LR-PCR, because primer accumulation did not permit accurate quantification of deletions where either the 5′ breakpoint or 3′ breakpoint fell within 500 bp of the 16 kb long-range PCR primer start positions (NC_012920.1 356-15926; modified NC_012920.1 16070-500). 122 Figure 5.3. The majority of the 1,728 mtDNA deletions fall within the major arc, between O L and O H. In this figure, the circular mitochondrial genome has been linearized, and the major arc is indicated by a gray band. A) Most deletions begin inside the major arc. B) Most deletions end inside the major arc. Asterisks indicate the position of CYTB, where the single most common deletion was found. 123 5.4.5 mtDNA deletions accumulate with age inside regions coding for mitochondrial-derived peptides In recent years, it was discovered that mitochondria produce small peptides with cytoprotective effects. 143,144 These mitochondrial-derived peptides (MDPs) are encoded by small open reading frames (sORFs) within other genes of the mitochondrial genome. There are currently 8 known MDPs: Humanin, MOTS-c, and small humanin-like peptides 1-6 (SHLP1-6). Although these small peptides possess cytoprotective effects, their expression level in the body decreases with age. 145,146 Accordingly, I determined whether mtDNA deletions affecting the sORFs encoding these peptides increase with age. Because the frontal cortex is more affected by mtDNA deletions with age and because it contains more deletions overall (Figures 5.1 and 5.2), I focused my analysis on the 143 frontal cortex samples. Figure 5.4 shows that mtDNA deletions affecting the sORFs encoding these peptides increase with age (regression p-value = 3.98e-15 ***). 124 Figure 5.4. mtDNA deletions accumulate with age inside regions coding for mitochondrial-derived peptides. Each dot is a frontal cortex sample (n = 143). All frontal cortex mtDNA deletions were filtered to retain only those that either entirely or partially encompassed one of the 8 sORFs. Age p-value = 3.98e- 15 ***, Age regression coefficient = 6.509e-03. P-value and regression coefficient obtained from a multiple regression after correcting for mitochondrial benchmark coverage: Unique Deletions per 10K Coverage ~ Benchmark Coverage + Age 125 5.4.6 Splice Break WGS vs. Splice Break LR-PCR exhibit biases in discovered mtDNA deletions As mentioned in the introduction, one of the motivations for applying the Splice Break bioinformatic pipeline to WGS was to avoid some of the potential biases associated with LR-PCR. Two notable biases are the potential for LR-PCR to favor amplification of small mtDNA templates (i.e. mtDNA with large deletions). In addition, LR-PCR primer accumulation does not permit accurate quantification of deletions where either the 5′ breakpoint or 3′ breakpoint is within 500 bp of the 16 kb long-range PCR primer start positions (NC_012920.1 356-15926; modified NC_012920.1 16070-500). Accordingly, I determined whether the WGS and LR-PCR datasets exhibited differences in the top 30 most common deletions they reported. However, because the original Splice Break paper using LR-PCR included 76 non-diseased frontal cortex samples and no cerebellum tissue, I compared the mtDNA deletions in these 76 samples with the mtDNA deletions obtained from the 143 frontal cortex samples of the NABEC control group. Figure 5.5 reveals substantial overlap in reported common deletions, but also substantial biases between the two methods. Splice Break with WGS (yellow circles) reported several small deletions (top right corner) as being among the top 30 most common deletions in the frontal cortex of the NABEC cohort. In contrast, Splice Break with LR-PCR (blue circles) failed to report any of these deletions as being in the top 30 in its respective samples, largely for the reason that these deletions occur to close to the LR-PCR primers to be accurately quantified. In addition, Splice Break with LR-PCR reported several enormous mtDNA deletions as being among the most common (bottom right corner). Notably, these deletions actually begin before the OL, which means they remove this origin of replication. It seems unlikely that mtDNA deletions that remove an origin of replication would be able to successfully clonally expand to be among the most common deletions, and so the 126 prevalence of these deletions in the LR-PCR data is likely a reflection of the bias that LR-PCR amplification cycles display towards smaller DNA templates. Figure 5.5. Splice Break WGS vs. LR-PCR exhibit biases towards small and large deletions, respectively. The top 30 most common frontal cortex deletions from the original Splice Break dataset using LR-PCR are in light blue. The top 30 most common frontal cortex deletions from the NABEC control cohort using Splice Break with WGS are in yellow. Each bubble represents a mtDNA deletion. Color represents the dataset, and the size of the bubble is the proportion of frontal cortex samples from that dataset that have the deletion. The dashed red line indicates the light strand origin of replication, O L. Note that both datasets exhibit overlaps in terms of the most common mtDNA deletions discovered. However, Splice Break with LR-PCR displays bias toward large mtDNA deletions (bottom right square). These are deletions that actually remove the O L. In contrast, Splice Break with WGS displays bias towards small mtDNA deletions (top right square). 127 5.4.7 Measuring mtDNA copy number from WGS with fastMitoCalc One of the added benefits of applying the Splice Break bioinformatic pipeline to WGS is the potential to estimate mtDNA copy number directly from the sequencing data. This cannot be accomplished if LR-PCR of mtDNA is used because sequencing coverage from both the autosomal genome and mitochondrial genome is required to estimate mtDNA copy number (Equation 5.1). Equation 5.1 is based on the logic that as long as mtDNA and autosomal DNA are sequenced at the same efficiency, average sequencing coverage should be proportional to copy number. Autosomal DNA copy number is equal to 2, and both mtDNA and autosomal DNA average coverage can be measured from the sequencing data. With this in mind, the only unsolved variable is mtDNA copy number. By rearranging the terms of Equation 5.1, we can solve for mtDNA copy number (Equation 5.2). Because the autosomal genome is so much larger than mitochondrial genome, the rate limiting step in solving Equation 5.2 is the determination of the autosomal DNA average coverage across approximately 3 billion bases. fastMitoCalc is an NIH developed computer program which solves this dilemma by randomly sampling only 0.1% of the autosomal genome. 147 Surprisingly, the average coverage calculated from this small random sample displays a correlation > 0.999 with the average coverage calculated from the entire autosomal genome. fastMitoCalc was originally validated with WGS data obtained from blood, which is a low mtDNA copy number tissue. Accordingly, I sought to examine fastMitoCalc’s performance on brain tissue, which is abundant in mitochondria, and also determine fastMitoCalc’s performance 128 with varying levels of sequencing coverage. To give a preliminary answer to these questions, I selected one frontal cortex and one cerebellum sample from the NABEC cohort. For the cerebellum I selected sample UMARY-177, a 16yr old male with an average autosomal sequencing coverage of 30 and average mtDNA coverage of 8.6865 x 10 4 . For the frontal cortex I selected KEN-1229, a 92yr old female with an average autosomal sequencing coverage of 36 and an average mtDNA coverage of 9.0345 x 10 4 . Figure 5.6 displays the autosomal and mitochondrial coverage across the NABEC cohort and highlights these two samples. Figure 5.6 also reveals that frontal cortex and cerebellum display different levels of mtDNA coverage, despite having relatively similar autosomal DNA coverage. This result suggests that the two tissues exhibit differences in mtDNA copy number. To determine the effect of sequencing coverage on fastMitoCalc’s performance, we randomly sampled decreasing proportions of the original BAM (binary sequence alignment map) files from UMARY-177 and KEN-1229 to obtain bam files with overall sequencing coverage ranging from the original amounts down to one millionth of the original amounts. These bam files were then run through fastMitoCalc three times each to obtain average mtDNA copy number estimates for the sample at that level of sequencing data (i.e. at that level of autosomal DNA and mtDNA coverage). Figure 5.7 reveals that even at a proportion of 0.001 (i.e. 0.1%) of the original sequencing data, fastMitoCalc displays low variability and high concordance with the original data. However, fastMitoCalc became unstable at the one millionth sampling proportion (1e-06) and reported infinite mtDNA copy number for both KEN-1229 and UMARY-177. This result can be understood in light of Equation 5.2; when autosomal coverage becomes very low, there is division by a number close to zero and thus estimated mtDNA copy number is erroneously inflated. 129 Equation 5.1 01234 5678597 :;678597 01234 :;<= >?0@78 = 5?1;A;05B 234 5678597 :;678597 5?1;A;05B 234 :;<= >?0@78 Equation 5.2 01234 :;<= >?0@78 = 01234 5678597 :;678597×2 5?1;A;05B 234 5678597 :;678597 Figure 5.6. Frontal Cortex and Cerebellum display similar levels of autosomal DNA coverage but different levels of mtDNA coverage. Each dot is a sample from the NABEC control cohort. KEN-1228 and UMARY-177 are enlarged with respect to the other samples. These two samples display similar levels of mtDNA coverage but different levels of autosomal DNA coverage. Note that frontal cortex and cerebellum samples cluster along the x-axis, revealing that these two tissues likely exhibit differences in mtDNA copy number. 130 Figure 5.7. fastMitoCalc performs well even at low autosomal sequencing coverage. On the x-axis, a proportion of 1 corresponds to the original bam files which had ~ 35X autosomal DNA coverage and ~ 87,000X mtDNA coverage. Each dot corresponds to one of three runs through fastMitoCalc at that sampling proportion and the height of the bars represents the average of the three runs. The error bars are standard deviation. Even at a proportion of 0.001 (i.e. 0.1%) of the original sequencing data, fastMitoCalc displays low variability and high concordance with the original data. Note that fastMitoCalc becomes unstable at the one millionth sampling proportion (1e-06). This result can be understood in light of Equation 5.2; when autosomal coverage becomes very low, there is division by a number close to zero and thus estimated mtDNA copy number is erroneously inflated. 131 5.4.8 The Frontal Cortex and Cerebellum display significant differences in mtDNA copy number One reasonable explanation for the elevated levels of mtDNA deletions in the frontal cortex compared to the cerebellum is that the frontal cortex simply contains more mitochondria and thus more mtDNA. Holding all other variables constant, increased mtDNA copy number should provide more opportunity for deletions to occur. To examine this hypothesis, I applied fastMitoCalc to all 292 samples of the NABEC control cohort. Figure 5.8 reveals that the cerebellum and frontal cortex exhibit significant differences (p-value < 2e-16) in mtDNA copy number after correcting for age, with the frontal cortex containing far more mtDNA than the cerebellum: frontal cortex mtDNA average = 4,869; cerebellum mtDNA average = 1,694. I then examined the effect of age on mtDNA copy number within each brain region, as previous literature has reported negative correlations between mtDNA copy number and biological aging. 148-150 Figure 5.9 reveals that the frontal cortex exhibits a significant decrease in mtDNA copy number with age (p-value = 0.0113*). In contrast, the cerebellum does not exhibit a significant correlation with age (p-value = 0.09). 132 Figure 5.8. Frontal Cortex displays elevated mtDNA copy number compared with Cerebellum. mtDNA copy number from the 149 cerebellum and 143 frontal cortex samples of the NABEC control cohort was analyzed with fastMitoCalc. The average mtDNA copy number for cerebellum was 1,694. In contrast, the average for frontal cortex was 4,869. The p-value for brain region was obtained from a linear regression after correcting for age: mtDNA copy number ~ Age + Brain Region 133 Figure 5.9. mtDNA copy number significantly decreases with age in the frontal cortex but not the cerebellum. Each dot is either a frontal cortex or cerebellum sample. This is the same data as shown in Figure 5.8 but plotted over Age. Linear regressions of mtDNA copy number ~ Age were made for the frontal cortex (Age p-value = 0.0113*) and the Cerebellum (Age p-value = 0.09) 134 5.4.9 Description of the Parkinson’s Disease versus age matched healthy control (PDVHC) cohort In addition to the 292 samples of NABEC control cohort, I also obtained whole genome sequencing (WGS) from the cerebellums of 341 donors who were diagnosed with Parkinson’s Disease (PD). These samples were collected across five different brain banks and contained donors ranging in age from 45 - 101yrs old at time of death. To compare mtDNA copy number in these PD cerebellums with healthy cerebellums, I filtered the NABEC control cohort to only include cerebellum samples >= 45yrs old at time of death, which was the youngest donor among the PD cerebellums. This filtering produced an age-matched healthy cohort of 74 samples. I combined these 74 healthy samples with the 341 PD samples to obtain the Parkinson’s Disease versus age matched healthy control (PDVHC) cohort. Summarized descriptions of the PDVHC cohort are listed in tables 5.4 and 5.5. 135 Table 5.4. Summary of the PDVHC cohort across disease status Total Samples min (yrs.) mean (yrs.) median (yrs.) max (yrs.) All 415 45 78.1 80.0 101 Female 144 45 79.4 82.9 95 Male 271 45 77.5 78.0 101 Total Samples min (yrs.) mean (yrs.) median (yrs.) max (yrs.) Healthy 74 45 75.6 82.0 95 Female 30 47 76.1 84.0 95 Male 44 45 75.3 81.5 95 Total Samples min (yrs.) mean (yrs.) median (yrs.) max (yrs.) PD a 341 45 78.7 78.7 101 Female 114 45 80.2 81.5 95 Male 227 49 77.9 78.0 101 a) Parkinson’s Disease Table 5.5. Summary of the PDVHC cohort sample sizes across brain banks BANNER a Harvard b HBSF c JHU d UMARY e MIAMI f All 212 16 47 91 47 2 Female 73 7 13 34 17 0 Male 139 9 34 57 30 2 BANNER a Harvard b HBSF c JHU d UMARY e MIAMI f Healthy 53 0 0 0 19 2 Female 22 0 0 0 8 0 Male 31 0 0 0 11 2 BANNER a Harvard b HBSF c JHU d UMARY e MIAMI f PD 159 16 47 91 28 0 Female 51 7 13 34 9 0 Male 108 9 34 57 19 0 a) Banner Health Brain and Tissue Bank, Sun City, AZ, USA b) Harvard University, Cambridge, MA, USA c) Sepulveda Research Corporation, Los Angeles, CA, USA d) Johns Hopkins University, Baltimore, MD, USA e) University of Maryland School of Medicine, Baltimore, MD, USA f) University of Miami Brain Endowment Bank, Miami, FL, USA 136 5.4.10 PD cerebellums display higher mtDNA copy number than age-matched healthy controls Previous literature has reported that the cerebellum is hyperactivated in parkinsonian animal models and human patients afflicted with PD. 151-156 The reasons underlying this hyperactivity are not well understood, but a reasonable hypothesis is that the cerebellum attempts to compensate for the loss of motor control due to the death of dopaminergic neurons within the midbrain. Hyperactivity also suggests that mtDNA copy number may be elevated in PD cerebellums. To examine whether PD cerebellums display elevated mtDNA copy number with age-matched healthy controls, I applied fastMitoCalc to the WGS data of the PDVHC cohort. Figure 5.10 reveals that PD cerebellums exhibit significantly elevated mtDNA copy number compared with age-matched healthy controls (p-value = 2.86e-06 ***). 137 Figure 5.10. PD cerebellums display higher mtDNA copy number than age-matched healthy controls. mtDNA copy number from the cerebellums of the 74 age-matched controls (Female = 30, Male = 44) and 341 samples (F = 114, M = 227) of the PDVHC cohort were analyzed with fastMitoCalc. The average mtDNA copy number for the age-matched controls was 1,300. In contrast, the average for the PD group was 2,054. The displayed p-value for disease status was obtained from a multiple linear regression after correcting for age: mtDNA copy number ~ Age + Disease Status 138 5.4.11 PD cerebellums appear to maintain copy number with age As discussed in section 5.4.7, mtDNA copy number has been reported to display a negative correlation with age. Although the cerebellum samples of NABEC control cohort did not display a significant correlation between mtDNA copy number and age (p-value = 0.09, Figure 5.9), this result examined the entire age range from 0.4-100 years old. It is reasonable to assume that changes in mtDNA copy number within a tissue may not be linear over the entire range of the human lifespan, and thus an examined linear trend may not be significant. For example, mtDNA may increase during development and adolescence, retain homeostasis during adulthood, and then decline during older adulthood. Thus, an examination of mtDNA copy number versus age in older individuals may display a significant negative correlation between mtDNA copy number and age. Because the PDVHC cohort contained older donors ranging in age from 45-101 years old (mean of 78.1 years old; table 5.4), I sought to determine whether the cerebellums of donors with PD displayed differences in mtDNA copy number correlation with age compared to the cerebellums from age-matched controls. Figure 5.11 displays the same data as Figure 5.10 but plotted over age. Although the slopes for age in both the PD and age-matched controls are non- significant, compared to the age-matched controls, the slope for the PD group is more level, suggesting a relatively constant amount of mtDNA with age. This observation is bolstered by observing the differences in the magnitude of the p-values for age, the differences in magnitude of the slopes for age, the differences in the sign of these slopes (negative for healthy, positive for PD), and the confidence intervals for these slopes (largely below zero for healthy, balanced around zero for PD): Healthy age slope = -7.63, p-value = 0.119, 95% CI = (-17.2, +2.0). PD age slope = 139 2.19, p-value = 0.81, 95% CI = (-15.9, +20.3). Also note that the PD group is several times larger than the healthy group (341 vs. 74). Taken together, these results suggest that a larger cohort of healthy age-matched controls may display a significant negative correlation with age. In contrast, it is likely that a larger cohort of similarly aged PD samples would still display a non-significant correlation with age. Figure 5.11. PD cerebellums appear to maintain copy number with age. Each circle is a cerebellum sample from either a PD donor or an age-matched control. This is the same data as shown in Figure 5.10 but plotted over age. Healthy age slope = -7.63, p-value = 0.119, 95% CI = (-17.2, +2.0). PD age slope = 2.19, p-value = 0.81, 95% CI = (-15.9, +20.3). Also note that the PD group is several times larger than the healthy group (341 vs. 74). P-values for each group were obtained from a linear regression of mtDNA Copy Number ~ Age 140 5.4.12 Cerebellum mtDNA copy number is increased in samples with more severe subtsantia nigra depigmentation PD pathology is largely characterized by the loss of dopamine producing neurons within the substantia nigra (SN), a small region located in the midbrain. Under normal physiological conditions, these dopaminergic neurons are pigmented with neuromelanin. However, during cell death these neurons become depigmented, a phenomenon which can be seen in histopathological slides from deceased donors and magnetic resonance imaging of living patients. SN depigmentation is a hallmark of PD pathology and the severity of this depigmentation is positively correlated with symptom progression. 157-159 Based on this knowledge, the literature reports of hyperactivation in parkinsonian cerebellums, and the elevated levels of mtDNA seen in parkinsonian cerebellums (Figures 5.10 and 5.11), I sought to determine whether mtDNA copy number is correlated with SN depigmentation. Fortunately, a subset of the PDVHC cohort included SN depigmentation scores. This scoring system is an integer scale containing 4 levels (i.e. 0,1,2,3), with severity of SN depigmentation increasing with the magnitude of the score. To determine whether there were significant differences in mtDNA copy number between these levels, I combined samples with levels 0-1 (Low SN Depigmentation) into one bin and combined samples with levels 2-3 (High SN Depigmentation) into another bin. I then compared mtDNA copy number between these bins. Figure 5.12 levels reveals that mtDNA copy number is significantly different between these SN depigmentation bins (p-value = 1.15e-05 ***), with the High SN Depigmentation bin displaying elevated mtDNA copy number. 141 Figure 5.12. Cerebellum mtDNA copy number is increased in samples with more severe subtsantia nigra depigmentation. The average cerebellum mtDNA copy number for the “Low SN Depigmentation” bin was 1,328. In contrast, the average cerebellum mtDNA copy number for the “High SN Depigmentation” bin was 2,194. The p-value for the difference in mtDNA copy number between the two bins was calculated from a multiple linear regression after correcting for age: mtDNA Copy Number ~ Age + SN Depigmentation Bin 142 5.5 Discussion The existence of deleterious mtDNA deletions were first reported in 1988 by Holt and colleagues. 160 In the decades since, researchers discovered that mtDNA deletions exist in a wide variety of tissues in both humans and animal models. Despite this knowledge, traditional methods for detecting mtDNA deletions have largely prevented an unbiased and high-throughput examination of the role they play in biological aging and complex disease. The Splice-Break method developed by Hjelm and colleagues now allows researchers to discover and quantify the relative abundance of mtDNA deletions in a high-throughput fashion, and this chapter shows that the bioinformatic portion of Splice-Break can be applied to WGS. This result means that from the same sequencing file researchers can now identify and quantify mtDNA deletions, estimate mtDNA copy number, and discover autosomal DNA variants. With this newfound functionality for WGS data, researchers can begin to uncover how mtDNA deletions and copy number contribute to disease and aging, and how autosomal DNA variation controls mtDNA deletions and copy number in different tissues. The differences in mtDNA copy number and deletions between frontal cortex and cerebellum, and how these differences are magnified by age, is a striking example of how mitochondrial dynamics vary widely between different tissues, even within the same organ (Figures 5.1-5.2, 5.8). The reasons underlying why frontal cortex mitochondria are more affected by aging than the cerebellum isn’t entirely clear. More work needs to be done and bioinformatic data should be paired with data from histopathology and molecular biology. However, it’s intriguing to think about how the metabolic demands of a tissue may dictate its response to aging, 143 especially in light of recent studies suggesting that caloric restriction imparts age and mito- protective effects. 161,162 The fact that PD affected cerebellums display elevated mtDNA copy number compared to age-matched controls (Figures 5.10 and 5.11) supports previous reports of elevated activity in the cerebellums of patients afflicted with PD. 156 This elevated activity may indicate that the cerebellum attempts to compensate for the loss of motor control due to the death of dopaminergic neurons in the substantia nigra. This hypothesis is supported by the fact that mtDNA copy correlates with severity of PD associated pathology in the substantia nigra (Figure 5.12). Although the palliative effects of this increased cerebral activity aren’t yet clear, these data suggest that PD associated motor symptoms could potentially be alleviated by improving mitochondrial function in the cerebellum. 144 5.6 Conclusion This chapter demonstrates for the first time that mtDNA deletions can be detected and quantified from WGS. This result can be achieved with the bioinformatic portion of Splice-Break, which uses the MapSplice algorithm for read alignment and mtDNA deletion junction calling. In healthy non-diseased donors, mtDNA deletions and copy number are elevated in the frontal cortex compared to the cerebellum. In addition, mtDNA copy number is elevated in the cerebellum of donors afflicted with Parkinson’s Disease (PD), compared to age matched non-diseased controls. Furthermore, cerebellum mtDNA copy is positively correlated with severity of PD associated pathology in the Substantia Nigra. These results yield insight on how mitochondrial dynamics change with biological aging in different brain regions and add to the growing body of evidence linking mitochondria to Parkinson’s Disease. 145 CONCLUSION AND FUTURE DIRECTIONS To summarize, Chapters 2 and 3 describe new advancements in the control of protein assembly inside of biological systems using thermo-responsive Elastin-like polypeptides (ELPs). Chapter 2 demonstrates the first report of thermo-responsive Caveolin 1 (CAV1) ELP fusion proteins that assemble CAV1 into large microdomains that internalize Cholera Toxin Subunit B. This microdomain formation and internalization of ligand is reminiscent of caveolae mediated endocytosis. Given the importance of caveolae (and CAV1) in endocytosis, cell signaling, and mechanosensing, CAV1-ELPs could be applied to study the endocytosis of various ligands, to control cell signaling, or to study the effects of caveolar ELP microdomain formation on cell survival and the cell’s ability to respond to mechanical or osmotic pressure. Chapter 3 is the first report demonstrating that ELPs can be applied for tunable protein assembly in vivo. Based on these studies, an exciting future direction would be the application of functional ELP fusion proteins, such as CAV1-ELPs, ELP-CLCs, or EGFR-ELPs to control cellular pathways in vivo. Chapter 4 describes the live cell imaging technique I developed to determine the intracellular Tt of non- fluorescent ELP fusions proteins. One unanswered question from this method is how to determine the intracellular concentration of the non-fluorescent ELP fusion proteins (only the concentration of the fluorescent GFP-ELP can be directly determined from the live cells). The requirement of tracking the Tt from each individual cell and then matching the non-fluorescent ELP concentration of each cell, possibly determined through fixation and ELISA or through fixation and immunofluorescence with FACS or microscopy, does seem challenging. To avoid the complexity of these experiments, one could envision the initial creation of stable cell lines that express constant levels of non-fluorescent ELP fusion protein. Given that viral transfection has previously been 146 used to create stable ELP fusion protein cell lines, it is likely that this method could be applied to CAV1-ELPs and ELP-CLCs. 104 Chapter 5 describes a new bioinformatic technique for detecting deletions in the mitochondrial genome (mtDNA) from whole genome sequencing data. This chapter also details how these mtDNA deletions and mtDNA copy number change with age and Parkinson’s Disease status. Future directions could examine the ability of this technique to identify mtDNA deletions from whole exome sequencing, which is substantially cheaper than whole genome sequencing. In addition, this technique could be applied to identify mtDNA deletions from sequencing data obtained from other brain tissues, such as the substantia nigra, or even non- brain tissues, like the liver, retina, or heart. Such studies could elucidate the pathological impact of mtDNA deletions in a variety of mitochondrial rich tissue. 147 REFERENCES 1 Tyrpak, D. R. et al. Caveolin Elastin-Like Polypeptide Fusions Mediate Temperature- Dependent Assembly of Caveolar Microdomains. Acs Biomater Sci Eng 6, 198-204, doi:10.1021/acsbiomaterials.9b01331 (2020). 2 Li, Z., Tyrpak, D. R., Lien, C. L. & MacKay, J. A. Tunable Assembly of Protein- Microdomains in Living Vertebrate Embryos. Adv Biosyst 2, doi:ARTN 1800112 10.1002/adbi.201800112 (2018). 3 Venters, B. J. & Pugh, B. F. How eukaryotic genes are transcribed. Crit Rev Biochem Mol Biol 44, 117-141, doi:10.1080/10409230902858785 (2009). 4 Kaksonen, M. & Roux, A. Mechanisms of clathrin-mediated endocytosis. Nat Rev Mol Cell Biol 19, 313-326, doi:10.1038/nrm.2017.132 (2018). 5 Sepulveda, J. L., Gkretsi, V. & Wu, C. Assembly and signaling of adhesion complexes. Curr Top Dev Biol 68, 183-225, doi:10.1016/S0070-2153(05)68007-6 (2005). 6 Dutta, D. & Donaldson, J. G. Search for inhibitors of endocytosis: Intended specificity and unintended consequences. Cell Logist 2, 203-208, doi:10.4161/cl.23967 (2012). 7 Ivanov, A. I. Pharmacological inhibition of endocytic pathways: is it specific enough to be useful? Methods Mol Biol 440, 15-33, doi:10.1007/978-1-59745-178-9_2 (2008). 8 Schenone, M., Dancik, V., Wagner, B. K. & Clemons, P. A. Target identification and mechanism of action in chemical biology and drug discovery. Nat Chem Biol 9, 232-240, doi:10.1038/nchembio.1199 (2013). 9 Miyazawa, K. Encountering unpredicted off-target effects of pharmacological inhibitors. J Biochem 150, 1-3, doi:10.1093/jb/mvr053 (2011). 10 Zhang, X. H., Tee, L. Y., Wang, X. G., Huang, Q. S. & Yang, S. H. Off-target Effects in CRISPR/Cas9-mediated Genome Engineering. Mol Ther Nucleic Acids 4, e264, doi:10.1038/mtna.2015.37 (2015). 11 Jackson, A. L. & Linsley, P. S. Recognizing and avoiding siRNA off-target effects for target identification and therapeutic application. Nat Rev Drug Discov 9, 57-67, doi:10.1038/nrd3010 (2010). 148 12 Lino, C. A., Harper, J. C., Carney, J. P. & Timlin, J. A. Delivering CRISPR: a review of the challenges and approaches. Drug Deliv 25, 1234-1257, doi:10.1080/10717544.2018.1474964 (2018). 13 El-Brolosy, M. A. & Stainier, D. Y. R. Genetic compensation: A phenomenon in search of mechanisms. PLoS Genet 13, e1006780, doi:10.1371/journal.pgen.1006780 (2017). 14 DeRose, R., Miyamoto, T. & Inoue, T. Manipulating signaling at will: chemically- inducible dimerization (CID) techniques resolve problems in cell biology. Pflugers Arch 465, 409-417, doi:10.1007/s00424-012-1208-6 (2013). 15 Athavankar, S. & Peterson, B. R. Control of gene expression with small molecules: biotin- mediated acylation of targeted lysine residues in recombinant yeast. Chem Biol 10, 1245- 1253, doi:10.1016/j.chembiol.2003.11.007 (2003). 16 Kopytek, S. J., Standaert, R. F., Dyer, J. C. & Hu, J. C. Chemically induced dimerization of dihydrofolate reductase by a homobifunctional dimer of methotrexate. Chem Biol 7, 313-321, doi:10.1016/s1074-5521(00)00109-5 (2000). 17 Spencer, D. M., Wandless, T. J., Schreiber, S. L. & Crabtree, G. R. Controlling signal transduction with synthetic ligands. Science 262, 1019-1024, doi:10.1126/science.7694365 (1993). 18 Rollins, C. T. et al. A ligand-reversible dimerization system for controlling protein-protein interactions. Proc Natl Acad Sci U S A 97, 7096-7101, doi:10.1073/pnas.100101997 (2000). 19 Barrero, J. J., Papanikou, E., Casler, J. C., Day, K. J. & Glick, B. S. An improved reversibly dimerizing mutant of the FK506-binding protein FKBP. Cell Logist 6, e1204848, doi:10.1080/21592799.2016.1204848 (2016). 20 Lin, Y. C. et al. Rapidly reversible manipulation of molecular activity with dual chemical dimerizers. Angew Chem Int Ed Engl 52, 6450-6454, doi:10.1002/anie.201301219 (2013). 21 Liu, P. et al. A bioorthogonal small-molecule-switch system for controlling protein function in live cells. Angew Chem Int Ed Engl 53, 10049-10055, doi:10.1002/anie.201403463 (2014). 149 22 Voss, S., Klewer, L. & Wu, Y. W. Chemically induced dimerization: reversible and spatiotemporal control of protein function in cells. Curr Opin Chem Biol 28, 194-201, doi:10.1016/j.cbpa.2015.09.003 (2015). 23 Chen, X. & Wu, Y. W. Tunable and Photoswitchable Chemically Induced Dimerization for Chemo-optogenetic Control of Protein and Organelle Positioning. Angew Chem Int Ed Engl 57, 6796-6799, doi:10.1002/anie.201800140 (2018). 24 Crick, F. The impact of molecular biology on neuroscience. Philos Trans R Soc Lond B Biol Sci 354, 2021-2025, doi:10.1098/rstb.1999.0541 (1999). 25 Crick, F. H. Thinking about the brain. Sci Am 241, 219-232, doi:10.1038/scientificamerican0979-219 (1979). 26 Zemelman, B. V., Lee, G. A., Ng, M. & Miesenbock, G. Selective photostimulation of genetically chARGed neurons. Neuron 33, 15-22, doi:10.1016/s0896-6273(01)00574-8 (2002). 27 Matsuno-Yagi, A. & Mukohata, Y. Two possible roles of bacteriorhodopsin; a comparative study of strains of Halobacterium halobium differing in pigmentation. Biochem Biophys Res Commun 78, 237-243, doi:10.1016/0006-291x(77)91245-1 (1977). 28 Govorunova, E. G., Sineshchekov, O. A., Li, H. & Spudich, J. L. Microbial Rhodopsins: Diversity, Mechanisms, and Optogenetic Applications. Annu Rev Biochem 86, 845-872, doi:10.1146/annurev-biochem-101910-144233 (2017). 29 Deisseroth, K. Optogenetics. Nat Methods 8, 26-29, doi:10.1038/nmeth.f.324 (2011). 30 Nagel, G. et al. Channelrhodopsin-1: a light-gated proton channel in green algae. Science 296, 2395-2398, doi:10.1126/science.1072068 (2002). 31 Boyden, E. S., Zhang, F., Bamberg, E., Nagel, G. & Deisseroth, K. Millisecond-timescale, genetically targeted optical control of neural activity. Nat Neurosci 8, 1263-1268, doi:10.1038/nn1525 (2005). 32 Tsai, H. C. et al. Phasic firing in dopaminergic neurons is sufficient for behavioral conditioning. Science 324, 1080-1084, doi:10.1126/science.1168878 (2009). 150 33 Kim, J. M. et al. Light-driven activation of beta 2-adrenergic receptor signaling by a chimeric rhodopsin containing the beta 2-adrenergic receptor cytoplasmic loops. Biochemistry 44, 2284-2292, doi:10.1021/bi048328i (2005). 34 Airan, R. D., Thompson, K. R., Fenno, L. E., Bernstein, H. & Deisseroth, K. Temporally precise in vivo control of intracellular signalling. Nature 458, 1025-1029, doi:10.1038/nature07926 (2009). 35 Shimizu-Sato, S., Huq, E., Tepperman, J. M. & Quail, P. H. A light-switchable gene promoter system. Nat Biotechnol 20, 1041-1044, doi:10.1038/nbt734 (2002). 36 Wang, X., Chen, X. & Yang, Y. Spatiotemporal control of gene expression by a light- switchable transgene system. Nat Methods 9, 266-269, doi:10.1038/nmeth.1892 (2012). 37 Kennedy, M. J. et al. Rapid blue-light-mediated induction of protein interactions in living cells. Nat Methods 7, 973-975, doi:10.1038/nmeth.1524 (2010). 38 Levskaya, A., Weiner, O. D., Lim, W. A. & Voigt, C. A. Spatiotemporal control of cell signalling using a light-switchable protein interaction. Nature 461, 997-1001, doi:10.1038/nature08446 (2009). 39 Strickland, D. et al. TULIPs: tunable, light-controlled interacting protein tags for cell biology. Nat Methods 9, 379-384, doi:10.1038/nmeth.1904 (2012). 40 Yazawa, M., Sadaghiani, A. M., Hsueh, B. & Dolmetsch, R. E. Induction of protein-protein interactions in live cells using light. Nat Biotechnol 27, 941-945, doi:10.1038/nbt.1569 (2009). 41 Bugaj, L. J., Choksi, A. T., Mesuda, C. K., Kane, R. S. & Schaffer, D. V. Optogenetic protein clustering and signaling activation in mammalian cells. Nat Methods 10, 249-252, doi:10.1038/nmeth.2360 (2013). 42 Mas, P., Devlin, P. F., Panda, S. & Kay, S. A. Functional interaction of phytochrome B and cryptochrome 2. Nature 408, 207-211, doi:10.1038/35041583 (2000). 43 Taslimi, A. et al. An optimized optogenetic clustering tool for probing protein interaction and function. Nat Commun 5, 4925, doi:10.1038/ncomms5925 (2014). 151 44 Lee, S. et al. Reversible protein inactivation by optogenetic trapping in cells. Nat Methods 11, 633-636, doi:10.1038/nmeth.2940 (2014). 45 Mannix, R. J. et al. Nanomagnetic actuation of receptor-mediated signal transduction. Nat Nanotechnol 3, 36-40, doi:10.1038/nnano.2007.418 (2008). 46 Bharde, A. A. et al. Magnetic nanoparticles as mediators of ligand-free activation of EGFR signaling. PLoS One 8, e68879, doi:10.1371/journal.pone.0068879 (2013). 47 Cho, M. H. et al. A magnetic switch for the control of cell death signalling in in vitro and in vivo systems. Nat Mater 11, 1038-1043, doi:10.1038/nmat3430 (2012). 48 Steketee, M. B. et al. Nanoparticle-mediated signaling endosome localization regulates growth cone motility and neurite growth. Proc Natl Acad Sci U S A 108, 19042-19047, doi:10.1073/pnas.1019624108 (2011). 49 Etoc, F. et al. Subcellular control of Rac-GTPase signalling by magnetogenetic manipulation inside living cells. Nat Nanotechnol 8, 193-198, doi:10.1038/nnano.2013.23 (2013). 50 Hoffmann, C. et al. Spatiotemporal control of microtubule nucleation and assembly using magnetic nanoparticles. Nat Nanotechnol 8, 199-205, doi:10.1038/nnano.2012.246 (2013). 51 Li, T. L. et al. Engineering a Genetically Encoded Magnetic Protein Crystal. Nano Lett 19, 6955-6963, doi:10.1021/acs.nanolett.9b02266 (2019). 52 Pastuszka, M. K. et al. A tunable and reversible platform for the intracellular formation of genetically engineered protein microdomains. Biomacromolecules 13, 3439-3444, doi:10.1021/bm301090x (2012). 53 Despanie, J., Dhandhukia, J. P., Hamm-Alvarez, S. F. & MacKay, J. A. Elastin-like polypeptides: Therapeutic applications for an emerging class of nanomedicines. J Control Release 240, 93-108, doi:10.1016/j.jconrel.2015.11.010 (2016). 54 MacEwan, S. R. & Chilkoti, A. Elastin-like polypeptides: biomedical applications of tunable biopolymers. Biopolymers 94, 60-77, doi:10.1002/bip.21327 (2010). 152 55 Pastuszka, M. K., Okamoto, C. T., Hamm-Alvarez, S. F. & MacKay, J. A. Flipping the Switch on Clathrin-Mediated Endocytosis using Thermally Responsive Protein Microdomains. Adv Funct Mater 24, 5340-5347, doi:10.1002/adfm.201400715 (2014). 56 Li, Z., Tyrpak, D. R., Park, M., Okamoto, C. T. & MacKay, J. A. A new temperature- dependent strategy to modulate the epidermal growth factor receptor. Biomaterials 183, 319-330, doi:10.1016/j.biomaterials.2018.07.063 (2018). 57 Walser, P. J. et al. Constitutive formation of caveolae in a bacterium. Cell 150, 752-763, doi:10.1016/j.cell.2012.06.042 (2012). 58 Parton, R. G. & del Pozo, M. A. Caveolae as plasma membrane sensors, protectors and organizers. Nat Rev Mol Cell Biol 14, 98-112, doi:10.1038/nrm3512 (2013). 59 Parton, R. G. Caveolae: Structure, Function, and Relationship to Disease. Annu Rev Cell Dev Biol 34, 111-136, doi:10.1146/annurev-cellbio-100617-062737 (2018). 60 Hill, M. M. et al. PTRF-Cavin, a conserved cytoplasmic protein required for Caveola formation and function. Cell 132, 113-124, doi:10.1016/j.cell.2007.11.042 (2008). 61 Kovtun, O., Tillu, V. A., Ariotti, N., Parton, R. G. & Collins, B. M. Cavin family proteins and the assembly of caveolae. J Cell Sci 128, 1269-1278, doi:10.1242/jcs.167866 (2015). 62 Liu, L. B. et al. Deletion of Cavin/PTRF Causes Global Loss of Caveolae, Dyslipidemia, and Glucose Intolerance. Cell Metab 8, 310-317, doi:10.1016/j.cmet.2008.07.008 (2008). 63 Schlegel, A. & Lisanti, M. P. A molecular dissection of caveolin-1 membrane attachment and oligomerization. Two separate regions of the caveolin-1 C-terminal domain mediate membrane binding and oligomer/oligomer interactions in vivo. J Biol Chem 275, 21605- 21617, doi:10.1074/jbc.M002558200 (2000). 64 Thorn, H. et al. Cell surface orifices of caveolae and localization of caveolin to the necks of caveolae in adipocytes. Mol Biol Cell 14, 3967-3976, doi:10.1091/mbc.e03-01-0050 (2003). 65 Echarri, A. & Del Pozo, M. A. Caveolae - mechanosensitive membrane invaginations linked to actin filaments. J Cell Sci 128, 2747-2758, doi:10.1242/jcs.153940 (2015). 153 66 Kiss, A. L. & Botos, E. Ocadaic acid retains caveolae in multicaveolar clusters. Pathol Oncol Res 15, 479-486, doi:10.1007/s12253-008-9139-4 (2009). 67 Gabor, K. A. et al. Super resolution microscopy reveals that caveolin-1 is required for spatial organization of CRFB1 and subsequent antiviral signaling in zebrafish. PLoS One 8, e68759, doi:10.1371/journal.pone.0068759 (2013). 68 Patel, H. H., Murray, F. & Insel, P. A. Caveolae as organizers of pharmacologically relevant signal transduction molecules. Annu Rev Pharmacol Toxicol 48, 359-391, doi:10.1146/annurev.pharmtox.48.121506.124841 (2008). 69 Okamoto, T., Schlegel, A., Scherer, P. E. & Lisanti, M. P. Caveolins, a family of scaffolding proteins for organizing "preassembled signaling complexes" at the plasma membrane. J Biol Chem 273, 5419-5422, doi:10.1074/jbc.273.10.5419 (1998). 70 Cheng, J. P. et al. Caveolae protect endothelial cells from membrane rupture during increased cardiac output. J Cell Biol 211, 53-61, doi:10.1083/jcb.201504042 (2015). 71 Cheng, J. P. X. & Nichols, B. J. Caveolae: One Function or Many? Trends Cell Biol 26, 177-189, doi:10.1016/j.tcb.2015.10.010 (2016). 72 Golani, G., Ariotti, N., Parton, R. G. & Kozlov, M. M. Membrane Curvature and Tension Control the Formation and Collapse of Caveolar Superstructures. Dev Cell 48, 523-538 e524, doi:10.1016/j.devcel.2018.12.005 (2019). 73 Kiss, A. L. & Botos, E. Endocytosis via caveolae: alternative pathway with distinct cellular compartments to avoid lysosomal degradation? J Cell Mol Med 13, 1228-1237, doi:10.1111/j.1582-4934.2009.00754.x (2009). 74 Kiss, A. L. Caveolae and the regulation of endocytosis. Adv Exp Med Biol 729, 14-28, doi:10.1007/978-1-4614-1222-9_2 (2012). 75 Mayor, S., Parton, R. G. & Donaldson, J. G. Clathrin-independent pathways of endocytosis. Cold Spring Harb Perspect Biol 6, doi:10.1101/cshperspect.a016758 (2014). 76 Ge, X., Conley, A. J., Brandle, J. E., Truant, R. & Filipe, C. D. In vivo formation of protein based aqueous microcompartments. J Am Chem Soc 131, 9094-9099, doi:10.1021/ja902890r (2009). 154 77 Jang, Y. et al. Engineering Globular Protein Vesicles through Tunable Self-Assembly of Recombinant Fusion Proteins. Small 13, doi:10.1002/smll.201700399 (2017). 78 Vogele, K. et al. Towards synthetic cells using peptide-based reaction compartments. Nat Commun 9, 3862, doi:10.1038/s41467-018-06379-8 (2018). 79 McDaniel, J. R., Mackay, J. A., Quiroz, F. G. & Chilkoti, A. Recursive directional ligation by plasmid reconstruction allows rapid and seamless cloning of oligomeric genes. Biomacromolecules 11, 944-952, doi:10.1021/bm901387t (2010). 80 Corrected-Total-Cell-Fluorescence v. v1.0 (github, github, 2018). 81 Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat Methods 9, 676-682, doi:10.1038/nmeth.2019 (2012). 82 Costes, S. V. et al. Automatic and quantitative measurement of protein-protein colocalization in live cells. Biophys J 86, 3993-4003, doi:10.1529/biophysj.103.038422 (2004). 83 Potapova, T. A., Sivakumar, S., Flynn, J. N., Li, R. & Gorbsky, G. J. Mitotic progression becomes irreversible in prometaphase and collapses when Wee1 and Cdc25 are inhibited. Mol Biol Cell 22, 1191-1206, doi:10.1091/mbc.E10-07-0599 (2011). 84 Gavet, O. & Pines, J. Progressive activation of CyclinB1-Cdk1 coordinates entry to mitosis. Dev Cell 18, 533-543, doi:10.1016/j.devcel.2010.02.013 (2010). 85 Burgess, A. et al. Loss of human Greatwall results in G2 arrest and multiple mitotic defects due to deregulation of the cyclin B-Cdc2/PP2A balance. Proc Natl Acad Sci U S A 107, 12564-12569, doi:10.1073/pnas.0914191107 (2010). 86 McCloy, R. A. et al. Partial inhibition of Cdk1 in G 2 phase overrides the SAC and decouples mitotic events. Cell Cycle 13, 1400-1412, doi:10.4161/cc.28401 (2014). 87 Bland, J. M. & Altman, D. G. Transformations, means, and confidence intervals. BMJ 312, 1079, doi:10.1136/bmj.312.7038.1079 (1996). 88 Bland, J. M. & Altman, D. G. Statistics notes. Logarithms. BMJ 312, 700, doi:10.1136/bmj.312.7032.700 (1996). 155 89 A language and environment for statistical computing (R Foundation for Statistical Computing, Vienna, Austria, 2018). 90 Jose Pinheiro, D. B., Saikat DebRoy, Deepayan Sarkar, R Core Team. {nlme}: Linear and Nonlinear Mixed Effects Models. (2018). 91 West, B. T., Welch, K. B., Gałecki, A. T. & Gillespie, B. W. Linear mixed models : a practical guide using statistical software. Second edition. edn. 92 Han, B., Tiwari, A. & Kenworthy, A. K. Tagging strategies strongly affect the fate of overexpressed caveolin-1. Traffic 16, 417-438, doi:10.1111/tra.12254 (2015). 93 Chinnapen, D. J., Chinnapen, H., Saslowsky, D. & Lencer, W. I. Rafting with cholera toxin: endocytosis and trafficking from plasma membrane to ER. FEMS Microbiol Lett 266, 129-137, doi:10.1111/j.1574-6968.2006.00545.x (2007). 94 Pang, H., Le, P. U. & Nabi, I. R. Ganglioside GM1 levels are a determinant of the extent of caveolae/raft-dependent endocytosis of cholera toxin to the Golgi apparatus. J Cell Sci 117, 1421-1430, doi:10.1242/jcs.01009 (2004). 95 Torgersen, M. L., Skretting, G., van Deurs, B. & Sandvig, K. Internalization of cholera toxin by different endocytic mechanisms. J Cell Sci 114, 3737-3747 (2001). 96 Liu, L. & Pilch, P. F. A critical role of cavin (polymerase I and transcript release factor) in caveolae formation and organization. J Biol Chem 283, 4314-4322, doi:10.1074/jbc.M707890200 (2008). 97 Tsai, M. F. et al. Au nanorod design as light-absorber in the first and second biological near-infrared windows for in vivo photothermal therapy. ACS Nano 7, 5330-5342, doi:10.1021/nn401187c (2013). 98 Shirata, C. et al. Near-infrared photothermal/photodynamic therapy with indocyanine green induces apoptosis of hepatocellular carcinoma cells through oxidative stress. Sci Rep 7, 13958, doi:10.1038/s41598-017-14401-0 (2017). 99 Papapostolou, D. & Howorka, S. Engineering and exploiting protein assemblies in synthetic biology. Mol Biosyst 5, 723-732, doi:10.1039/b902440a (2009). 156 100 Ng, D. Y. W. et al. Directing intracellular supramolecular assembly with N-heteroaromatic quaterthiophene analogues. Nat Commun 8, 1850, doi:10.1038/s41467-017-02020-2 (2017). 101 Benner, S. A. & Sismour, A. M. Synthetic biology. Nature Reviews Genetics 6, 533 (2005). 102 Cameron, D. E., Bashor, C. J. & Collins, J. J. A brief history of synthetic biology. Nature Reviews Microbiology 12, 381 (2014). 103 Rawls, R. L. 'Synthetic biology'makes its debut. Chemical & Engineering News 78, 49-49 (2000). 104 Shi, P., Lin, Y. A., Pastuszka, M., Cui, H. & Mackay, J. A. Triggered sorting and co- assembly of genetically engineered protein microdomains in the cytoplasm. Adv Mater 26, 449-454, doi:10.1002/adma.201303356 (2014). 105 Shah, M. et al. Biodegradation of elastin-like polypeptide nanoparticles. Protein Sci 21, 743-750, doi:10.1002/pro.2063 (2012). 106 Despanie, J., Dhandhukia, J. P., Hamm-Alvarez, S. F. & MacKay, J. A. Elastin-like polypeptides: Therapeutic applications for an emerging class of nanomedicines. Journal of Controlled Release 240, 93-108 (2016). 107 Dooley, K. & Zon, L. I. Zebrafish: a model system for the study of human disease. Curr Opin Genet Dev 10, 252-256 (2000). 108 Lieschke, G. J. & Currie, P. D. Animal models of human disease: zebrafish swim into view. Nat Rev Genet 8, 353-367, doi:10.1038/nrg2091 (2007). 109 Becker, C. G. & Becker, T. Adult zebrafish as a model for successful central nervous system regeneration. Restorative neurology and neuroscience 26, 71-80 (2008). 110 Mackay, E. W., Apschner, A. & Schulte-Merker, S. A bone to pick with zebrafish. BoneKEy reports 2 (2013). 111 Bakkers, J. Zebrafish as a model to study cardiac development and human cardiac disease. Cardiovascular research 91, 279-288 (2011). 157 112 Lahiri, K. et al. Temperature regulates transcription in the zebrafish circadian clock. PLoS Biol 3, e351, doi:10.1371/journal.pbio.0030351 (2005). 113 Brand, M., Granato, M. & Nüsslein-Volhard, C. Keeping and raising zebrafish. Zebrafish 261, 7-37 (2002). 114 Aranda, P. S., LaJoie, D. M. & Jorcyk, C. L. Bleach gel: a simple agarose gel for analyzing RNA quality. Electrophoresis 33, 366-369 (2012). 115 Meyer, D. E. & Chilkoti, A. Purification of recombinant proteins by fusion with thermally- responsive polypeptides. Nat Biotechnol 17, 1112-1115, doi:10.1038/15100 (1999). 116 Guryev, V. et al. Genetic variation in the zebrafish. Genome research 16, 491-497 (2006). 117 Balik-Meisner, M., Truong, L., Scholl, E. H., Tanguay, R. L. & Reif, D. M. Population genetic diversity in zebrafish lines. Mammalian Genome 29, 90-100 (2018). 118 Holden, L. A. & Brown, K. H. Baseline mRNA expression differs widely between common laboratory strains of zebrafish. Scientific reports 8, 4780 (2018). 119 Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B. & Schilling, T. F. Stages of embryonic development of the zebrafish. Developmental dynamics 203, 253-310 (1995). 120 Avdesh, A. et al. Regular care and maintenance of a zebrafish (Danio rerio) laboratory: an introduction. Journal of visualized experiments: JoVE (2012). 121 Janib, S. M. et al. A quantitative recipe for engineering protein polymer nanoparticles. Polymer chemistry 5, 1614-1625 (2014). 122 Le, D. H. T. & Sugawara-Narutaki, A. Elastin-like polypeptides as building motifs toward designing functional nanobiomaterials. Mol Syst Des Eng 4, 545-565, doi:10.1039/c9me00002j (2019). 123 Tyrpak, D. d-tear/Corrected-Total-Cell-Fluorescence: Initial release of CTCF pipeline, <https://github.com/d-tear/Corrected-Total-Cell-Fluorescence/tree/v1.0> (2019). 124 Nassir, F. & Ibdah, J. A. Role of mitochondria in nonalcoholic fatty liver disease. Int J Mol Sci 15, 8713-8742, doi:10.3390/ijms15058713 (2014). 158 125 Siasos, G. et al. Mitochondria and cardiovascular diseases-from pathophysiology to treatment. Ann Transl Med 6, 256, doi:10.21037/atm.2018.06.21 (2018). 126 Terluk, M. R. et al. Investigating mitochondria as a target for treating age-related macular degeneration. J Neurosci 35, 7304-7311, doi:10.1523/JNEUROSCI.0190-15.2015 (2015). 127 Ridge, P. G. & Kauwe, J. S. K. Mitochondria and Alzheimer's Disease: the Role of Mitochondrial Genetic Variation. Curr Genet Med Rep 6, 1-10, doi:10.1007/s40142-018- 0132-2 (2018). 128 Chen, C., Turnbull, D. M. & Reeve, A. K. Mitochondrial Dysfunction in Parkinson's Disease-Cause or Consequence? Biology (Basel) 8, doi:10.3390/biology8020038 (2019). 129 Greaves, L. C., Barron, M. J., Campbell-Shiel, G., Kirkwood, T. B. & Turnbull, D. M. Differences in the accumulation of mitochondrial defects with age in mice and humans. Mech Ageing Dev 132, 588-591, doi:10.1016/j.mad.2011.10.004 (2011). 130 Khrapko, K. et al. Cell-by-cell scanning of whole mitochondrial genomes in aged human heart reveals a significant fraction of myocytes with clonally expanded deletions. Nucleic Acids Res 27, 2434-2441, doi:10.1093/nar/27.11.2434 (1999). 131 Brierley, E. J., Johnson, M. A., Lightowlers, R. N., James, O. F. & Turnbull, D. M. Role of mitochondrial DNA mutations in human aging: implications for the central nervous system and muscle. Ann Neurol 43, 217-223, doi:10.1002/ana.410430212 (1998). 132 Campbell, G., Krishnan, K. J., Deschauer, M., Taylor, R. W. & Turnbull, D. M. Dissecting the mechanisms underlying the accumulation of mitochondrial DNA deletions in human skeletal muscle. Hum Mol Genet 23, 4612-4620, doi:10.1093/hmg/ddu176 (2014). 133 Gitschlag, B. L. et al. Homeostatic Responses Regulate Selfish Mitochondrial Genome Dynamics in C. elegans. Cell Metab 24, 91-103, doi:10.1016/j.cmet.2016.06.008 (2016). 134 Fukui, H. & Moraes, C. T. Mechanisms of formation and accumulation of mitochondrial DNA deletions in aging neurons. Hum Mol Genet 18, 1028-1036, doi:10.1093/hmg/ddn437 (2009). 135 Kowald, A. & Kirkwood, T. B. L. Resolving the Enigma of the Clonal Expansion of mtDNA Deletions. Genes (Basel) 9, doi:10.3390/genes9030126 (2018). 159 136 Kowald, A., Dawson, M. & Kirkwood, T. B. Mitochondrial mutations and ageing: can mitochondrial deletion mutants accumulate via a size based replication advantage? J Theor Biol 340, 111-118, doi:10.1016/j.jtbi.2013.09.009 (2014). 137 Yoneda, M., Chomyn, A., Martinuzzi, A., Hurko, O. & Attardi, G. Marked replicative advantage of human mtDNA carrying a point mutation that causes the MELAS encephalomyopathy. Proc Natl Acad Sci U S A 89, 11164-11168, doi:10.1073/pnas.89.23.11164 (1992). 138 Hjelm, B. E. et al. Splice-Break: exploiting an RNA-seq splice junction algorithm to discover mitochondrial DNA deletion breakpoints and analyses of psychiatric disorders. Nucleic Acids Res 47, e59, doi:10.1093/nar/gkz164 (2019). 139 Wang, K. et al. MapSplice: accurate mapping of RNA-seq reads for splice junction discovery. Nucleic Acids Res 38, e178, doi:10.1093/nar/gkq622 (2010). 140 Kazachkova, N., Ramos, A., Santos, C. & Lima, M. Mitochondrial DNA damage patterns and aging: revising the evidences for humans and mice. Aging Dis 4, 337-350, doi:10.14336/AD.2013.0400337 (2013). 141 Phillips, N. R., Sprouse, M. L. & Roby, R. K. Simultaneous quantification of mitochondrial DNA copy number and deletion ratio: a multiplex real-time PCR assay. Sci Rep 4, 3887, doi:10.1038/srep03887 (2014). 142 Damas, J., Carneiro, J., Amorim, A. & Pereira, F. MitoBreak: the mitochondrial DNA breakpoints database. Nucleic Acids Res 42, D1261-1268, doi:10.1093/nar/gkt982 (2014). 143 Cobb, L. J. et al. Naturally occurring mitochondrial-derived peptides are age-dependent regulators of apoptosis, insulin sensitivity, and inflammatory markers. Aging (Albany NY) 8, 796-809, doi:10.18632/aging.100943 (2016). 144 Yang, Y. et al. The role of mitochondria-derived peptides in cardiovascular disease: Recent updates. Biomed Pharmacother 117, 109075, doi:10.1016/j.biopha.2019.109075 (2019). 145 Muzumdar, R. H. et al. Humanin: a novel central regulator of peripheral insulin action. PLoS One 4, e6334, doi:10.1371/journal.pone.0006334 (2009). 146 Zarate, S. C., Traetta, M. E., Codagnone, M. G., Seilicovich, A. & Reines, A. G. Humanin, a Mitochondrial-Derived Peptide Released by Astrocytes, Prevents Synapse Loss in 160 Hippocampal Neurons. Front Aging Neurosci 11, 123, doi:10.3389/fnagi.2019.00123 (2019). 147 Qian, Y. et al. fastMitoCalc: an ultra-fast program to estimate mitochondrial DNA copy number from whole-genome sequences. Bioinformatics 33, 1399-1401, doi:10.1093/bioinformatics/btw835 (2017). 148 Zhang, R., Wang, Y., Ye, K., Picard, M. & Gu, Z. Independent impacts of aging on mitochondrial DNA quantity and quality in humans. BMC Genomics 18, 890, doi:10.1186/s12864-017-4287-0 (2017). 149 Xia, C. Y. et al. Reference Intervals of Mitochondrial DNA Copy Number in Peripheral Blood for Chinese Minors and Adults. Chin Med J (Engl) 130, 2435-2440, doi:10.4103/0366-6999.216395 (2017). 150 Praveen, G., Shalini, T., Sivaprasad, M. & Reddy, G. B. Relative telomere length and mitochondrial DNA copy number variation with age: Association with plasma folate and vitamin B12. Mitochondrion 51, 79-87, doi:10.1016/j.mito.2020.01.007 (2020). 151 Heman, P. et al. Nigral degeneration correlates with persistent activation of cerebellar Purkinje cells in MPTP-treated monkeys. Histol Histopathol 27, 89-94, doi:10.14670/HH- 27.89 (2012). 152 Wu, T. & Hallett, M. A functional MRI study of automatic movements in patients with Parkinson's disease. Brain 128, 2250-2259, doi:10.1093/brain/awh569 (2005). 153 Wu, T., Wang, L., Hallett, M., Li, K. & Chan, P. Neural correlates of bimanual anti-phase and in-phase movements in Parkinson's disease. Brain 133, 2394-2409, doi:10.1093/brain/awq151 (2010). 154 Jahanshahi, M. et al. Dopaminergic modulation of striato-frontal connectivity during motor timing in Parkinson's disease. Brain 133, 727-745, doi:10.1093/brain/awq012 (2010). 155 Rascol, O. et al. The ipsilateral cerebellar hemisphere is overactive during hand movements in akinetic parkinsonian patients. Brain 120 ( Pt 1), 103-110, doi:10.1093/brain/120.1.103 (1997). 156 Wu, T. & Hallett, M. The cerebellum in Parkinson's disease. Brain 136, 696-709, doi:10.1093/brain/aws360 (2013). 161 157 Schwarz, S. T., Xing, Y., Tomar, P., Bajaj, N. & Auer, D. P. In Vivo Assessment of Brainstem Depigmentation in Parkinson Disease: Potential as a Severity Marker for Multicenter Studies. Radiology 283, 789-798, doi:10.1148/radiol.2016160662 (2017). 158 Sulzer, D. et al. Neuromelanin detection by magnetic resonance imaging (MRI) and its promise as a biomarker for Parkinson's disease. NPJ Parkinsons Dis 4, 11, doi:10.1038/s41531-018-0047-3 (2018). 159 Giguere, N., Burke Nanni, S. & Trudeau, L. E. On Cell Loss and Selective Vulnerability of Neuronal Populations in Parkinson's Disease. Front Neurol 9, 455, doi:10.3389/fneur.2018.00455 (2018). 160 Holt, I. J., Harding, A. E. & Morgan-Hughes, J. A. Deletions of muscle mitochondrial DNA in patients with mitochondrial myopathies. Nature 331, 717-719, doi:10.1038/331717a0 (1988). 161 Lettieri-Barbato, D., Cannata, S. M., Casagrande, V., Ciriolo, M. R. & Aquilano, K. Time- controlled fasting prevents aging-like mitochondrial changes induced by persistent dietary fat overload in skeletal muscle. PLoS One 13, e0195912, doi:10.1371/journal.pone.0195912 (2018). 162 Sergi, D. et al. Mitochondrial (Dys)function and Insulin Resistance: From Pathophysiological Molecular Mechanisms to the Impact of Diet. Front Physiol 10, 532, doi:10.3389/fphys.2019.00532 (2019).

Abstract (if available)

Abstract This dissertation describes original work in two different scientific fields: 1) Synthetic biology, which is concerned with controlling and mimicking biological systems. 2) Bioinformatics, which is focused on the use and development of computer programs to analyze complex biological data. The work described in this thesis is mainly concerned with innovative methodologies for use in biomedical research. However, in Chapter 5 some novel biological findings regarding mtDNA dynamics in regard to aging and Parkinson’s Disease status are discussed.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Synthetic biopolymers modulate cell signaling

PDF

Tubulin-based fusion proteins as multifunctional tools

PDF

Targeting LFA-1/ICAM-1 interaction using ICAM-1 binding elastin-like polypeptide

PDF

Development of novel immunosuppressant-based therapies to treat dacryoadenitis in a Sjögren’s syndrome mouse model

PDF

Temperature triggered protein assembly enables signaling switching and peptide drug delivery

PDF

Integrin-mediated targeting of protein polymer nanoparticles carrying a cytostatic macrolide

PDF

Characterization of changes in metabolism and inflammation in the lacrimal gland in dry eye disorders

PDF

Expression and purification of different elastin like polypeptides (ELPs) constructs for therapeutic applications

PDF

The modulation of dynamin and receptor endocytosis machinery using elastin-like polypeptides

PDF

Trafficking of targeted elastin‐like polypeptide nanoparticles in the lacrimal gland

PDF

Stable cell lines expressing elastin-like polypeptide fusions with epidermal growth factor receptor modulate gene expression in a heat dependent manner

PDF

Methods and protocols for detecting the intracellular assembly of elastin-like polypeptides

PDF

Temperature-mediated induction of caveolin-mediated endocytosis via elastin-like polypeptides

PDF

Development of protein polymer therapeutics for the eye

PDF

Development and therapeutic assessment of multivalent protein polymers for cancer and eye diseases

PDF

Cognate receptor fusion proteins enable parenteral delivery of challenging rapalogues

PDF

Elastin like polypeptides as thermoresponsive modulators in reversible bioadhesives for fragile skin

PDF

Multivalent smart elastin-like polypeptide therapeutics with drug delivery and biosensing applications.

PDF

Sustained and targeted delivery of rapamycin using FKBP-elastin like polypeptide fusion proteins

PDF

Physical chemistry and polymeric modification of elastin-like polypeptides

Asset Metadata

Creator Tyrpak, David Ryan (author)

Core Title Three advancements in biotechnology: new tools for synthetic biology and next generation sequencing

School School of Pharmacy

Degree Doctor of Philosophy

Degree Program Pharmaceutical Sciences

Publication Date 09/18/2020

Defense Date 07/23/2020

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag caveolin 1,elastin-like polypeptides,mtDNA,mtDNA deletions,OAI-PMH Harvest,Parkinson's disease,SIAL

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor MacKay, Andrew (committee chair), Hamm-Alvarez, Sarah (committee member), Okamoto, Curtis (committee member), Stiles, Bangyan (committee member)

Creator Email dtyrpak@g.clemson.edu,tyrpak@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-373060

Unique identifier UC11666477

Identifier etd-TyrpakDavi-8975.pdf (filename),usctheses-c89-373060 (legacy record id)

Legacy Identifier etd-TyrpakDavi-8975.pdf

Dmrecord 373060

Document Type Dissertation

Rights Tyrpak, David Ryan

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA