UCSF

Transient interactions between the Anaphase-Promoting Complex/Cyclosome (APC/C) and its activator subunit Cdc20 or Cdh1 generate oscillations in ubiquitylation activity necessary to maintain the order of cell cycle events. Activator binds the APC/C with high affinity and exhibits negligible dissociation kinetics in vitro, and it is not clear how the rapid turnover of APC/C-activator complexes is achieved in vivo. Here, we describe a mechanism that controls APC/C-activator interactions based on the availability of substrates. We find that APC/C-activator dissociation is stimulated by abundant cellular polyanions such as nucleic acids and polyphosphate. Polyanions also interfere with substrate ubiquitylation. However, engagement with high-affinity substrate blocks the inhibitory effects of polyanions on activator binding and APC/C activity. We propose that this mechanism amplifies the effects of substrate affinity on APC/C function, stimulating processive ubiquitylation of high-affinity substrates and suppressing ubiquitylation of low-affinity substrates.

Biological macromolecular assemblies play crucial roles in most cellular processes. The determination of their structures, thermodynamics, and kinetics is essential to understand their function, evolution, modulation, and design. Determining such models, however, remains challenging. One particularly powerful approach to constructing models in general is integrative modeling. Integrative modeling aims to maximize the accuracy, precision, and completeness of models, by simultaneously utilizing all available information, including experimental data, physical principles, statistical analyses, and other prior models. The goal of this thesis is to expand the scope of integrative modeling to the inference of spatial, thermodynamic, and kinetic aspects of macromolecular assemblies.

In Chapter I, I introduce the integrative modeling framework for spatiotemporal modeling of biological macromolecular assemblies. In Chapter II, I demonstrate how the synergy between multi-chemistry cross-linking mass spectrometry and integrative modeling can map the structural dynamics of macromolecular assemblies, by application to the human Cop9 signalosome complex. In Chapter III, I present a method for determining structures, free energy differences, and kinetic rates of macromolecular assemblies along their functional cycle, mainly from negative stain electron microscopy (EM). We apply the method to the yeast Hsp90 to estimate the free energy differences and kinetic parameters along its nucleotide hydrolysis cycle, which includes open and closed states of Hsp90. In Chapter IV, I describe a validation of stochastic sampling in integrative modeling. The remaining chapters describe applications of integrative modeling to assemblies of various sizes and scales, using various sources of information, thus illustrating the flexibility of the integrative modeling approach. Specifically, I apply integrative modeling to the human ECM29-Proteasome assembly under oxidative stress (Chapter V), the yeast nuclear pore complex (NPC) cytoplasmic mRNA export platform (Chapter VI), the major membrane ring component of the yeast NPC (Chapter VII), the entire yeast NPC (Chapter VIII), and the reconstruction of 3D structures of MET antibodies (Chapter IX).

Magnetic resonance imaging (MRI) is one of the widely used medical imaging modality, since it can provide both structural and functional assessment in a single imaging session.

However, two major challenges should be considered by using MRI for lung imaging. The first challenge is the intrinsic low SNR of H-1 lung MRI due to the low proton density as well as the fast decay of the lung parenchyma signal. And the second challenge is subject motion. To achieve high resolution structural image, MRI requires a long scan time, usually a few minutes or even longer, which make MRI sensitive to subject motion.

To address the first challenge, ultra-short echo time (UTE) MRI sequence is used to capture the lung parenchyma signal before decay.

As for subject motion, two major strategies are widely used. One strategy is fast breath-holding scan, the subjects are asked to hold their breaths for a short duration, and the fast 3D MR sequence would be used to acquire data within that duration. This dissertation proposes a new acquisition scheme based on the standard UTE sequence, which largely increases the encoding efficiency and improves the breath-holding scan images.

The other is free breathing scan with motion correction. The subjects are allowed to breathe during the MR acquisition. After the acquisition, the motion corrupted data would go through the motion correction step to reconstruct the motion free images. In this dissertation, two novel motion corrected reconstruction strategies are proposed to incorporate the motion modeling and compensation into the reconstruction to get high SNR motion corrected 3D and 4D images.

When translating the developed techniques to the clinical studies, specifically for pediatric and neonatal studies, more practical problems need to be considered, such as smaller but finer anatomy to image, the different respiratory patterns of the young subjects etc. This dissertation proposes a 5-minute free breathing UTE MRI strategy to achieve a 3D high resolution motion free lung image for pediatric and neonatal studies.

Protein binding to small molecules is fundamental to many biological processes, yet it remains challenging to predictively design this functionality de novo. Current state-of-the-art computational design methods typically rely on existing small molecule binding sites or protein scaffolds with existing shape complementarity for a target ligand. Here we introduce new methods that utilize pools of discrete contacts observed in the Protein Data Bank between protein residues and defined small molecule ligand substructures (ligand fragments). We use the Rosetta Molecular Modeling Suite to recombine protein residues in these contact pools to generate hundreds of thousands of energetically favorable binding sites for a target ligand. These composite binding sites are built into existing scaffold proteins matching the intended binding site geometry with high accuracy. In addition, we apply pools of rotamers interacting with the target ligand to augment Rosetta’s conventional design machinery and improve key metrics known to be predictive of design success. We demonstrate that our method reliably builds diverse binding sites into different scaffold proteins for a variety of target molecules. Our generalizable de novo ligand binding site design method will lay the foundation for versatile design of protein to interface previously unattainable molecules for applications in medical diagnostics and synthetic biology.

Bulk RNA degradation irreversibly removes an mRNA from the translating pool to regulate gene expression. Transcripts must be degraded in a coordinated manner during cell cycle, development and in response to stimuli. A dense network of proteins assembles on a transcript to ensure timely and specific destruction of the RNA. This results in trimming of the 3’ poly(A) tail followed by removal of the 5’ methyl-7 guanine (m7G) cap structure, leading to rapid exonucleolysis of the message. A key question has been understanding how protein factors at the 3’ end trigger decapping at the 5’ end. Pat1 is a central scaffolding protein that interacts with multiple decay factors to control distinct steps during RNA turnover. Previous work has demonstrated that Pat1 enhances binding of the heteroheptameric Lsm1-7 complex to the 3’ end and promotes decapping by the Dcp1/Dcp2 enzyme complex at the 5’ end. Due to the multifunctional nature of Pat1, however, we lack a mechanistic understanding of how it regulates RNA turnover. In this work, I have used a reconstituted system to understand how Pat1 interacts with and activates distinct factors during 5’-3’ degradation. First, I show how Pat1 interacts with and enhance the RNA binding of Lsm1-7. This increased affinity is selective for adenine-rich oligoRNAs, which in turn broadens the specificity of the Lsm1-7 complex. Second, I show how Pat1 interacts with short linear motifs in the disordered C-terminal tail of Dcp2 to activate decapping by either recruiting the enzyme complex to substrate or alleviating autoinhibition to promote catalysis. Both activation of Lsm1-7 and decapping require a bipartite interaction between two domains of Pat1 and involve distinct surfaces and motifs. Last, I uncover how different decay factors tune both the size and assembly of Pat1, which may be leveraged to organize an active decapping complex. This biochemically reconstituted system provides a framework for how Pat1 can regulate multiple protein cofactors and steps during bulk RNA turnover.

Experience-dependent myelination is hypothesized to shape neural circuit function and subsequent behavioral output. Using a contextual fear memory task in mice, we demonstrate that fear learning induces oligodendrocyte precursor cells (OPCs) to proliferate and differentiate into myelinating oligodendrocytes (OLs) in the medial prefrontal cortex (mPFC). Transgenic animals which cannot form new myelin exhibit deficient remote, but not recent, fear memory recall. Recording population calcium dynamics with fiber photometry, we observe that the neuronal response to conditioned context cues evolves over time in the mPFC, but not in animals that cannot form new myelin. Finally, we demonstrate that pharmacological induction of new myelin formation with clemastine fumarate improves remote memory recall and promotes fear generalization. Thus, bidirectional manipulation of myelin plasticity functionally impacts behavior and neurophysiology, suggesting that neural activity during fear learning instructs the formation of new myelin, which, in turn, supports the consolidation and/or retrieval of remote fear memories.

Primary care and healthcare providers can facilitate children’s timely referral to a dental home. However, there are few studies of providers’ oral health knowledge and clinical skills. This study aims to improve future healthcare providers’ knowledge, confidence, attitude and clinical competence in assessing children’s oral health. Sixty-five health professional students participated in a 10-week didactic and clinical curriculum on children’s oral health. They completed pre- and post-training questionnaire to assess changes in knowledge, confidence and attitude. Calibrated faculty graded students’ clinical skills on a 24-point grading criterion. Descriptive statistics, paired sample t-test and Pearson correlation were used in data analyses. Students were in dentistry (46%), nursing (28%), medicine (22%), and pharmacy (3%). Students significantly improved in knowledge (t=-7.71, p<.001), confidence (t=-10.30, p=<.001) and attitude (t=-4.24, p=<.001). Students on average scored 83% on clinical competence, with the highest average for fluoride varnish application (96%) and lowest for providing anticipatory guidance (69%). There was a moderate correlation between improvement in knowledge and their clinical skills (r=.39, p=.010). Interprofessional education improves students’ knowledge, confidence, attitude and clinical competence in assessing children’s oral health. Such education is necessary in guiding future providers to gain adequate competence in serving the children’s oral health needs.

Keywords: Pediatric Dentistry; Primary Care; Children’s Oral Health; Interprofessional Education; Oral Health Education; Public Health Dentistry; Oral Health Disparity; Access to Care; Clinical Competency; Oral Health Assessment

Throughout history, methodological innovations have resulted in breakthroughs in our understanding of biology. Methods for determining static protein structures, as well as those for probing protein dynamics, are well-established. Nonetheless, visualizing molecules as dynamic entities that respond to their environment is still an outstanding challenge. Specifically, it is challenging to measure the spatial position of all the atoms within a molecule as a function of time. That challenge is the broad focus of this dissertation.

In chapter one, I begin by diving into modern crystallographic techniques that enable one to solve protein structures from sub-micron-sized crystals. I compare and contrast two methods, serial crystallography and electron crystallography, asking how each technique affects the protein’s structure. A primary factor differentiating these two methods is the temperature of the sample during the experiment. Despite this difference, both methods enable one to solve high-resolution structures from small crystals. This is advantageous for time-resolved experiments. Since there are fewer molecules in a small crystal, the perturbation is more uniform, which provides a clearer time-resolved signal.

In chapter two, I investigate temperature-jumps as a generalized perturbation for resolving the energy landscape of proteins. In this work, I focus on solution scattering experiments, which allow one to examine large-scale perturbations to a protein, as well as changes in the solvent shell surrounding the molecule. By mutating selected residues, we inhibited specific protein motions. Comparing these mutants to the wild-type protein allowed us to resolve the motions driven by an infrared laser. Nonetheless, we wished to gain all-atom spatial resolution, which required us to perform a temperature-jump within the context of crystallography rather than solution scattering.

In chapter three, I expand upon the temperature-jump detection method described in chapter two. By adapting this method to accommodate X-ray diffraction images, I demonstrate that we can detect temperature-jumps within a crystalline context. This is a crucial step in the development of a generalized perturbation for time-resolved crystallography. Given the timescale of the measurements, reading out the temperature directly from the X-ray data is the only effective way to track the sample’s response. Thus, our method offers proof-of-principle that IR laser-based temperature-jumps are feasible for time-resolved crystallography. While measuring the diffuse scattering signal is useful for temperature-jump detection, the diffuse signal also holds the potential to inform our understanding of protein dynamics.

In chapter four, I review the field of macromolecular diffuse scattering, as of late 2017. I begin by considering data collection practices, which requires extremely careful and controlled measurements. Then I examine different group's approaches to processing the data, as well as their models of the disorder that drives it. Finally, I consider the broader impact of diffuse analysis upon the field, ranging from the improvement of molecular dynamics forcefields to improved phasing and resolution extension. While these impacts hold exciting implications, it is clear that collecting high-quality is the first challenge to solve.

In chapter five, I examine the challenges of collecting high-quality diffuse scattering from protein crystals. I describe how parasitic scattering can confound our ability to develop rigorous models of the crystalline disorder that gives rise to the diffuse signal. Then I work through experimental measures that we took to minimize parasitic scattering while maximizing diffuse scatter driven by protein motions.

For a cell to divide correctly, the spindle must connect to and align chromosomes and then generate force to move them into two daughter cells. The kinetochore is the macromolecular machine that connects chromosomes to a bundle of dynamic microtubules, the kinetochore-fiber (k-fiber). While we have a nearly complete parts list of kinetochore components and regulators, how they together give rise to the robust mechanics of the kinetochore-microtubule interface remains poorly understood. This is due to the fact that mammalian kinetochores and k-fibers cannot yet be reconstituted in vitro and there are few tools to perturb forces and measure the mechanics at this interface in vivo. In my thesis work I have addressed this gap using direct biophysical assays in mammalian cells to focus on two main questions about the kinetochore-microtubule interface. First, how do kinetochores hold on to microtubules that grow and shrink? Using live imaging to monitor spindle dynamics and laser ablation to challenge kinetochore grip, I show that regulation of the key microtubule binding protein Ndc80/Hec1 at the outer kinetochore by the kinase Aurora B specifically affects kinetochore movement on polymerizing microtubules without disrupting coupling to depolymerizing microtubules that generate force to move chromosomes. Second, at the other side of the interface, how do kinetochore-fibers remodel under force? I directly exert forces on individual mammalian k-fibers and find that even under high force for minutes they do not lose grip. Instead, k-fibers bend and elongate by polymerizing at normal rates at plus-ends and inhibiting depolymerization at minus-ends – thus ensuring robust connection to kinetochores. Altogether I find that robust kinetochore grip emerges from underlying properties of both the kinetochore and k-fiber microtubules – specialized regulation at the kinetochore allows the cell to adjust grip while still allowing force generation and dynamic mechanical feedback of k-fiber microtubules locally dissipates force, protecting spindle connections. These fundamental physical properties of the kinetochore-microtubule interface allow the spindle to faithfully segregate chromosomes and more broadly suggest a model for how force-generating cellular machines can also robustly maintain their structure.

Many mechanisms govern both how animals respond to stimuli and how these responses inform physical state and future behavior. In C. elegans, these mechanisms include the generation and spread of various species of small RNAs. Emerging as a potent regulator of gene expression in both a single animal and its progeny, small RNAs are reshaping the fields of behavior and transgenerational inheritance. The rising prominence of small RNA within the context of these fields highlights gaps in the knowledge of essential processes, such as the precise mechanisms of small RNA export. RNA import, thanks to earlier efforts characterizing systemic RNA interference (RNAi), are relatively well understood. However, much less is known about how RNAi and other small RNA exit cells. This manuscript details tools and approaches for identifying mobile small RNAs, members of RNA export pathways, and points of regulation in the export process.