5430 research outputs found
Sort by
Benchmarking Reinforcement Learning and Off Policy Evaluation for Medical Decision Making
No full textHealthcare applications pose significant challenges to existing Reinforcement Learning (RL) methods due to implementation risks, low data availability, short treatment episodes, sparse re[1]wards, partial observations, and heterogeneous treatment effects (HTE). Despite significant interest in developing Dynamic Treatment Regimes (DTRs) for longitudinal patient care scenarios, no standardized benchmark has yet been developed. To address this gap, this thesis introduces Episodes of Care (EpiCare), a benchmark designed to mimic the challenges associated with applying RL to longitudinal healthcare settings. I leverage this benchmark to test seven state-of-the-art offline RL models as well as five common off-policy evaluation (OPE) techniques. My results suggest that while offline RL may be capable of improving upon existing standards of care given large data availability, its applicability does not appear to extend to the moderate to low data regimes typical of healthcare settings. Additionally, I demonstrate that several OPE techniques which have become standard in the medical RL literature fail to perform adequately under simulated conditions. These results suggest that the performance of RL models in DTRs may be difficult to meaningfully evaluate using current OPE methods, indicating that RL for this application may still be in its early stages. It is my hope that these findings, along with the EpiCare benchmark itself, will facilitate the comparison of existing methods and inspire further research into techniques that increase the practical applicability of medical RL
A Structural Perspective on Bacterial Transcription
No full textTranscription is a highly conserved process that lies at the very center of biology. As in the other two domains of the tree of life, bacterial transcription is a multistep process. In the first step, RNA polymerase (RNAP) contacts with the promoter DNA are established and then must be broken for the enzyme to transition into the elongation phase of RNA production; a process known as promoter escape. While single-molecule and biochemical observations report that promoter escape is a highly regulated and sometimes rate-limiting step in the transcription cycle, the structural details remain obscure. Promoter escape also serves as the target for the clinically important antibiotic rifampicin (Rif), used to treat tuberculosis. In this work, we first provide the structural details of M. tuberculosis RNAP escaping from a promoter using a de novo cryoelectron microscopy approach, revealing seven distinct intermediates. We reveal an unanticipated level of structural rearrangement that RNAP undergoes to clear the promoter, including those required to release the initiation factor, σ, providing mechanisms to decades of biochemical observations. These structures and supporting biochemistry are consistent with a model of promoter escape that includes unexpected conformations exploitable in the development of Rif-alternatives. After RNAP achieves promoter escape, it transitions into the elongation phase of transcription. This elongation phase is vulnerable to bulky helix-distorting DNA lesions capable of stalling elongating RNAP in its tracks. Transcription-coupled repair (TCR) is a sub-pathway of the nucleotide excision repair pathway that preferentially removes lesions from the DNA template-strand. These lesions stall RNAP elongation complexes (ECs). The SF2 MFD translocase mediates TCR in bacteria by removing stalled RNAP from DNA lesions and recruiting appropriate TCR factors. Previously, we used cryoelectron microscopy to visualize MFD engaging with and attempting to displace the EC, revealing seven MFD-EC complexes spanning the MFD loading and EC displacement pathway. However, the first MFD-EC loading intermediate (L1) was poorly resolved and the transition from L1 to the second loading intermediate (L2) was unclear. To investigate further, we pre-loaded MFD with ATP in the presence of a y-phosphate mimic, BeF3- , limiting rounds of ATP hydrolysis by MFD before being trapped by BeF3- binding. This biochemical strategy allowed us to improve L1 resolution, revealing the nucleotide occupancy (ATP). We also identified an additional loading intermediate between L1 and L2 (L1.5) that clarifies the transition from L1 to L2. After the elongation phase concludes, transcription terminates. Following transcript release during intrinsic termination, Escherichia coli (E. coli) RNAP often remains associated with DNA in a post-termination complex (PTC). RNAPs in PTCs are removed from the DNA by the Swi2/Snf2 ATPase RapA. In this work, we determined PTC structures on negatively-supercoiled DNA as well as of RapA engaged to dislodge the PTC. We found that core RNAP in the PTC can unwind DNA and initiate RNA synthesis but is prone to producing R-loops. We show that RapA helps to control cytotoxic R-loop formation invivo, likely by disrupting PTCs. Nucleotide binding to RapA triggers a conformational change that opens the RNAP clamp, allowing DNA in the RNAP cleft to reanneal and dissociate. We suggest that analogous ATPases acting on PTCs to suppress transcriptional noise and R-loop formation may be widespread. These results hold significance for the bacterial transcription cycle and highlight a role for RapA in maintaining genome stability. Overall, these studies come together to provide a deeper and ultimately revised perspective on the mechanisms underlying bacterial transcription
Spontaneously Regenerative Corticospinal Neurons in Mice
No full textThe mammalian spinal cord receives input from the motor cortex via the corticospinal tract (CST). The CST comprises corticospinal neuron (CSN) axons that begin in Layer 5b of motor cortex and can extend up to one meter to reach targets in the caudal spinal cord in humans. CSNs regulate the spinal cord\u27s sensorimotor function; however, the extent of their impact on motor control differs between species. In macaques, lesion of the pyramidal tract produces immediate limb paralysis. In rodents, their genetic ablation profoundly impacts skilled-reaching ability but does not produce paralysis. In almost all examined mammals, corticospinal axons primarily terminate contralateral to their cell bodies of origin. Despite the predominance of contralateral CSN termination, a minority of CSN axons terminate ipsilaterally. There has been considerable examination of ipsilateral axons in unilateral cortical injury because of evidence that they are a potential circuit for the uninjured cortex to compensate for the damaged one. However, little is known about their anatomy and function in healthy animals. Though sparse, these ipsilateral axons are a remarkable deviation from the norm. Brains have evolved only in bilaterally symmetric animals, so the midline is a foundational architectural principle of all central nervous systems. Accordingly, an intricate and redundant ensemble of molecular signals exerts tight control over midline crossing. The ipsilateral axons represent a consistent anomaly across many species at an evolutionarily ancient midline boundary that is under strict developmental control. To better understand these ipsilateral axons, we first sought to characterize their anatomy with a diverse array of emerging methods in mouse molecular neuroscience. Using anterograde tracing methods, tissue clearing, and Smart-seq3 single-nucleus RNA-sequencing (snRNA-seq), we found that ipsilateral CSN axons project to regions of the ventral horn, including directly to motor neurons. Barcode-based Multiplexed Analysis of Projections by Sequencing (MAPseq) of the CST revealed that the neurons contributing these axons primarily comprise a class of bilaterally projecting CSNs and represent a more substantial population of total CSNs than their sparse ipsilateral axons suggest. With retrograde tracing and tissue clearing, we found that the cell bodies of this neuronal population reside in distinct cortical regions, almost entirely absent from the caudo-lateral cortex. We deeply profiled their molecular characteristics using the viral implementation of Translating Ribosome Affinity Purification (vTRAP) and discovered a striking similarity to the embryonic-like molecular signature of regenerating corticospinal neurons. We noticed that the anatomical characteristics of IP-CSNs we had documented were shared with regenerating CSNs: in addition to molecular similarity, both share bilaterality, projection to ventral spinal cord regions, and motor neuronal connectivity. Given these similarities, we hypothesized that IP-CSNs might themselves have regenerative properties. Finally, we show that IP-CSNs are spontaneously regenerative. The discovery of a class of spontaneously regenerative CSNs may prove valuable to the study of spinal cord injury. Additionally, this work suggests that the retention of juvenile-like characteristics may be a widespread phenomenon in adult nervous systems
Molecular Mechanisms of Sensing and Synergy by Anti-Bacteriophage Immune Systems in Staphylococcus
Full text linkViruses parasitize every known life form on the planet for their propagation and spread. To deal with this constant assault, organisms across all domains of life have evolved immune strategies in the form of genetically encoded systems to defend themselves. Immunity can be conceptualized as occurring in three fundamental stages. First, the immune system must identify an invading pathogen— either directly via recognition of a pathogen-associated molecular pattern (PAMP), or indirectly by sensing the perturbation of homeostasis during the process of infection ( guard hypothesis ). Second, this information must be converted into molecular signals to be transmitted to the appropriate compartment within the cell or organism and to amplify downstream immune responses. Third, effector programs must be enacted to interfere with the virus\u27 ability to replicate and parasitize the host. Arising from these fundamental properties, immune systems are composed of multiple distinct components with specialized functions (e.g. the innate and adaptive arms of immunity in vertebrates) and involve a complex network of interactions. A viable immune response must therefore be able to optimally integrate multiple branches within this network in order to operate effectively—that is, to simultaneously enable robust immunity against any invading pathogen while limiting autoimmunity. Bacteriophages (phages) are viruses infecting bacteria that are thought to outnumber their hosts by a factor of ten to one in most environments. They are by far the most ubiquitous biological entities, estimated to exceed a staggering quantity of 1031 phage particles on the planet. As a result of this host-parasite conflict over billions of years, bacteria have evolved many diverse mechanisms of antibacteriophage defense, including many different innate systems as well as the adaptive CRISPR-Cas systems. In the last five years, dozens of new genes that confer bacterial immunity against bacteriophages have been identified through bioinformatic searches and more recently, functional screens. Many of these antibacteriophage defense systems have subsequently been validated and the details of their modes of activity have been revealed through experimental approaches. Although the mechanistic details of how these newly discovered defense genes interfere with bacteriophage propagation are rapidly being uncovered, there is a major gap in our understanding of how these systems sense bacteriophage infection in order to initiate immunity. Furthermore, it has only recently been recognized that within a single cell, bacteria commonly harbor many different defense systems—on average, five to seven distinct systems. Therefore, our understanding of how different defense systems interact with one another and the consequences of this crosstalk for the immune response and bacterial evolution remains in its infancy. In this thesis, I investigate these two fundamental, but nascent areas of the burgeoning field of bacterial immunology using staphylococci and staphylococcal phages as a powerful model system to study host-virus interactions. In the first part of my thesis (Chapters 1 and 2), I provide a broad introduction to the evolutionary conflict between bacteria and their phages, with a particular focus on staphylococci, a widespread taxon of bacteria with the utmost medical importance. I also discuss a conceptual framework for understanding how a successful immune response is generated in response to infection, with a focus on the current state of understanding of how bacterial defense systems are triggered by invading phages. In the second part of my thesis (Chapters 3 and 4), I demonstrate that a newly discovered defense system called CBASS, the bacterial homolog to the cGASSTING innate immune pathway in eukaryotes, confers anti-viral defense in staphylococci. Specifically, a minimal type I CBASS operon from Staphylococcus schleiferi that is also broadly present in other species, including Staphylococcus aureus, can provide immunity against some, but not all, staphylococcal phages. I then detail our discovery and characterization of the CBASS-activating bacteriophage RNA (cabRNA), which we propose to be a phage-specific cue during infection that triggers the initiation of CBASS-mediated immunity. In the third part of my thesis (Chapters 5, 6, and 7), I focus on the current state of knowledge on immune crosstalk in prokaryotes, with a focus on how these interactions impact the strength and durability of immunity. Specifically, I explore the complex tripartite interactions between bacteriophages, S. aureus pathogenicity islands (SaPIs), and CRISPR-Cas systems and detail our discovery that SaPI-mediated parasitism of phages can stimulate the development of CRISPR-Cas adaptive immunity through the production of immunizing defective viral particles. In the last part of my thesis (Chapter 8), I discuss the many unanswered questions that remain and the exciting avenues of inquiry that extend from my research. I further discuss the mechanistic insights gleaned from these experimental studies and the implications that this work has on understanding the interactions between bacteria and bacteriophages. In conclusion, the research described in this thesis collectively reveals novel molecular mechanisms underlying the sensing of bacteriophage infection by CBASS—the bacterial counterpart to cGAS—and the synergistic interactions between S. aureus pathogenicity islands and adaptive CRISPR-Cas immune systems in staphylococci. This work touches upon two fundamental features of every immune response across all domains of life: sensing and synergy. Furthermore, this work provides a foundation for future studies that will continue to uncover the details of how various anti-bacteriophage defense systems are activated during viral infection and additional modes of immune crosstalk in bacteria
Anterior Cingulate Function Across Multiple-Time Scales Shapes Learning and Memory
No full textWe are constantly learning from our everyday experiences about the people we are with, the things we do,where we are, and how we feel. If deemed significant enough,we may choose to remember what we learn for the rest of our lives.For my doctoral work, I studied how a part of the neocortex, the anterior cingulate cortex (ACC),has critical functions across multiple time-scales,from minutes to days-long learning,to weeks-long memory consolidation.The prefrontal cortex, and particularly the ACC portion, is known to be important for goal-directed learning, especially as the relationship between our actions (including effort) and their outcomes (including value) has to be continually adjusted to maximize rewards. To gain mechanistic insight not how ACC facilitates such goal-directed learning, I designed a behavioral task in which mice self-initiate trials to learn various cue- reward contingencies. By performing brain recordings as they perform this task, I found that ACC encodes and facilitates an extended motivational state, including trial history, reward outcome, and vigor to initiate the next trial, that together leads to maximization of rewards and improved learning. To determine how ACC inherits these signals, I recorded from inputs to ACC and identified a ramp in bulk neural activity in OFC-to-ACC that continued to rise as mice traversed non-rewarded rooms, which peaked when they finally reached a rewarded room, thus maintaining an extended motivational state. Cellular resolution imaging of OFC further confirmed these neural correlates of motivation, and further delineated separate ensembles of neurons that sequentially tiled the ramp. Together, these results identify a mechanism by which OFC maps out task structure to extend motivation and trains ACC in the development of a learned goal-directed behavior. As learned representations become episodic memories, their initial encoding is thought to occur in the hippocampus. However, overtime, it is thought that memories become less reliant on the hippocampus and consolidated in the neocortex for long-term stabilization. Despite extensive phenomenological study, we still lack mechanistic understanding of this brain-wide reorganization process. To provide insights, we developed a behavioral task where mice consolidate some (highly salient) memories, while forgetting others (less salient), and recorded brain activity in hippocampus, cortex, and intervening circuits throughout weeks-long memory consolidation. Initial bulk neural activity recordings during behavior identified a unique and significant neural correlate of memory in anterior thalamus that emerged in training and persisted for weeks. Inhibition of the anteromedial (AM) thalamic projection to ACC during training caused deficits in memory consolidation, whereas more strikingly, mild excitation was sufficient to enhance consolidation of otherwise unconsolidated (less salient) memories. We next developed a technique for imaging three brain regions simultaneously with single-cell resolution in the behaving mouse to gain mechanistic understanding into the role of anteromedial thalamus during consolidation. Using this technology, we found that while the hippocampus encodes low and high salient memories equally, the anteromedial thalamus forms preferential tuning to salient memories, and establishes long-ran synchrony with ACC that are causally required for stabilizing cortical representations to achieve successful memory consolidation. Thus, we extend previous models of memory consolidation to include AM as a critical mediator of long-term cortical storage. Overall, my thesis work has identified mechanisms, spanning diverse time-scales, by which ACC builds and sustains learned representations that can be used for long-term memory
Towards a Better Understanding of Facial Movements: Computational Models for Perception, Characterization, and Neural Production
No full textFacial movements are the primary medium for non-verbal communication and involve a complex orchestration of muscles controlled by the brain. The ability to interpret and produce these movements enables the expression of a wide range of emotions and social cues, all essential for social interactions. Despite their significance, the neural mechanisms governing facial movements remain poorly understood, hindered by the complexity of muscle coordination and the limitations of traditional, subjective, and labor-intensive analysis methods. To objectively understand facial motor control, this research adopts a three-pronged approach: 1) Developing and interpreting computational models within a novel multi-task training framework to simultaneously distinguish between facial expression and identity recognition, 2) Introducing a novel self-supervised Person-Specific Model (PSM) framework that extracts person-specific facial movements independently of other facial characteristics, enhancing facial muscle action characterization by leveraging individual differences, and 3) Utilizing data-driven computational models to analyze a unique dataset of single-cell recordings from sensorimotor cortex regions and behavioral video recordings of spontaneous, unconstrained, and naturalistic facial movements of macaques.This research first focused on developing computational models capable of separating facial movements from other characteristics like identity, which is challenging for computational models due to individual variations in facial movements and shapes. Utilizing Convolutional Neural Networks in a novel training framework, networks were trained simultaneously for facial identity and expression recognition, mirroring the human visual system\u27s face processing multi-tasking capability. This approach revealed functional segregation within the network, enabling the differentiation between a person\u27s identity and their expressions by dedicating different facial zones to solve each task and identifying task-specific facial features emerging from the network\u27s latest layers. Building on this, I introduced Person-Specific Model (PSM), an innovative self-supervised learning approach, which successfully extracts individual-specific facial movements independently of other facial characteristics. PSM stands out by leveraging individual differences, improving facial muscle action characterization. Its dual learning approach uniquely reveals a repertoire of facial movement primitives, capturing both universal patterns shared across individuals, and more complex, nuanced movements unique to each individual missed by other traditional methods. In parallel, this research explored the neural basis of facial movements, from larger movements like threats and chewing, to subtle, spontaneous movements in monkeys using naturalistic facial video recordings and single-cell neural data from sensorimotor cortex regions. A flexible computational framework was developed to analyze unconstrained continuous behavior. The analysis revealed that specialized neural patterns were linked to various movements; for example sensory area S1 was active during lipsmack, and primary motor cortex M1 was involved in actions like chewing and lipsmack. Distinct neural subspaces and neurons were associated with different social behaviors. The findings also revealed parallels between the neural dynamics of facial and well-studied arm motor behaviors; for example, threat expressions\u27 neural dynamics resembled reaching and sensory cortical areas like S1 exhibited unique dynamics during these expressions. Transitioning to continuous unconstrained behavior, this framework confirmed S1 and M1\u27s role in lower face movements, and importantly detected subtle neural-behavioral temporal patterns, like the role of anterior primary motor F4 in nose movements and ventral premotor PMV in eye movements, which traditional techniques failed to capture due to minimal movement variance in these facial locations. Particularly, distinct neural control strategies were identified, with regions like S1and M1 more active in larger expressive movements and PMV involved in more subtle movements, highlighting a sophisticated level of neural segregation between these different movement scales. Taken together, these results unveil a comprehensive picture of the intricate cortical control underlying facial movements, distinguishing between larger expressive motions and smaller, subtle actions. Crucially, this work underscores the need for standardized tools capable of analyzing spontaneous, unconstrained behaviors, beyond labeled expressions. I address this challenge in this analysis, promising a deeper understanding of natural behavior and its neural underpinnings
Fate, Phases, and Form in Vertebrate Organ Morphogenesis-Uncovering a Role of Morphogenesis at the Supracellular Scale
No full textHow morphology unfolds from fertilization to birth is one of the most fundamental questions in the life sciences. Self-organization in embryonic development, whereby organs robustly adopt forms through intrinsic processes, is a central feature that remains enigmatic. To achieve greater clarity of such processes, new conceptual and experimental approaches may be needed. Much of work in developmental biology in the past three to four decades has focused on cell and subcellular levels of organization. In line with work of some key conceptual thinkers, we revisit the notions of epigenetics. Rather than confining it to the molecular mechanisms of chromatin remodeling, we propose a broader understanding that includes processes beyond the cellular scale that possess their own generative power. Among the various scales at play, we propose to shift our focus to the mechanical and material processes at the supracellular scale. The hair or feather follicle in the skin represents one of the most apparent morphological features in development, and its spacing serves as a classic model system to study pattern formation. Mesenchymal-ECM relations are a key regulatory niche that remains poorly understood. By developing an ex vivo essay that reconstitutes the initiation of feather follicle pattern, we demonstrated that mesenchymal cell-ECM interplay can create supracellular structures independent of any morphogen activities. This challenges the classic chemical model where a morphogen pre-pattern dictates follicle patterning, prompting the question of what functional roles morphogens serve. found that morphogens enable the creation of membrane-less tissue compartments within the mesenchyme of the follicle with distinct biophysical properties, such as elasticity, viscosity and contractility. Specifically, fibroblast growth factor (FGF) promotes a stiff and solid hemispheric core compartment, whereas bone morphogenetic protein (BMP) promotes a surrounding fluid and active contractile margin compartment. Through their geometric arrangement, the two compartments are mechanically primed to break tissue symmetry and result in follicle budding. We also identified morphogen-enabled supracellular material property differences that were minimal or lost at cellular scales, which highlights the importance of shifting our focus to the supracellular scale in order to understand morphogenetic processes in development and diseases. This new paradigm redefines the role of morphogens, which requires distinguishing between the proximal effects of morphogens, such as gene regulation at the molecular and cellular scale, and their ultimate functional effects, which emerge at the supracellular scale
Investigating the Effect of Antibody-Mediated Feedback on Ongoing Germinal Center Responses
Full text linkAntibodies play a crucial role in protection against a wide range of pathogens. As such, the goal of most vaccinations is to induce immune memory in the form of long-lasting, high-affinity serum antibody titers, which requires the participation of B cells in the germinal center (GC) response. It is well-established that pre-existing antibodies can influence the outcome of future humoral responses; however, whether antibody from an ongoing response can feedback onto contemporaneous GCs to influence B cell clonal selection and affinity maturation in real-time remains poorly understood. Here we present a genetic mouse tool to interrogate the effect of antibody-mediated feedback on ongoing GC responses. We have designed an oligoclonal B cell transfer mouse model with B cell receptor specificities to non-overlapping epitopes on the same antigen, allowing us to distinguish between epitope-specific and epitope-non-specific effects of secreted antibody. In addition, we have genetically engineered a Blimp-1-DTR mouse that allows for temporally controlled ablation of plasma cells and, therefore, depletion of antibody titers. Combining these two models has revealed that high levels of antibody produced during B cell responses can modulate interclonal selection in the GC by suppressing GC B cells that bind the same epitopes. However, a traveling wave of antibody is not required to guide intraclonal selection and, therefore, is not a driver of affinity maturation. We propose antibody-mediated suppression via epitope masking may be an important mechanism in guiding the development of primary GCs towards epitopes that are poorly represented in the serum antibody compartment, thereby maintaining diversity within humoral responses
Mechanistic Insights into Direct Interactions that Mediate the Activity of DOT1L Complex in MLL-Rearranged Leukemia
No full textAcute leukemia can arise from the translocation-mediated fusion of the N terminus of the Mixed Lineage Leukemia (MLL) protein to various partner proteins. Although more than 60 different translocation-based fusion proteins have been identified, fusion partners such as AF9, AF10, and ENL constitute the majority of the MLL-rearranged leukemia cases; additionally, these fusion partners are also a part of the DOT1L (disruptor of telomeric silencing 1-like) complex. Aberrant histone H3 Lysine 79 (H3K79) methylation catalyzed by DOT1L was shown to be crucial for the maintenance of MLL-rearranged leukemia. DOT1L is the only known H3K79 methyltransferase, and further, a DOT1L inhibitor is currently being evaluated in clinical trials to treat Acute Myeloid Leukemia. Although cell-based studies have implicated the importance of interactions on the N terminus of the MLL fusion protein for leukemogenesis and disease maintenance, there have been very limited corresponding biochemical analyses demonstrating how these interactions of MLL regulate or contribute to abnormal levels of methylation of H3K79 by DOT1L. Additionally, both cell-based and in vitro studies suggested that H3K79 methylation is dependent upon histone H2B ubiquitylation (H2Bub). However, there is a lack of understanding of how these interactions stimulate MLL-fusion target genes. Besides its catalytic activity, DOT1L complex activity has been postulated to regulate transcriptional elongation on the grounds of its co-localization with Pol II, H3K79 methylation mark distribution on gene loci, and its association with subunits of SEC (Super Elongation Complex) such as AF4 (or AFF4), AF9 (or AF9-related ENL), and ELL1. However, any effect of DOT1L on other transcriptional steps is not known. Addressing these gaps, the present study employs a robust biochemical approach to underscore the importance of both MLL N terminus (MLLN) interactions and H2B ubiquitylation in the regulation of aberrant H3K79 methylation by MLL-fusion protein containing DOT1L complex in MLL-rearranged leukemia. Specifically, interactions between the PWWP (LEDGF) domain and histone H3 Lysine 36 tri-methylation; between the CXXC(MLLN) domain and unmethylated CpG of DNA; and between ubiquitylated H2B and DOT1L protein are found to stabilize the intrinsic nucleosome binding property of MLL. AF9, consequently augmenting the H3K79 methyltransferase activity of the MLL-fusion containing DOT1L complex. As a result, the leukemogenic MLL-fusion protein containing the DOT1L complex exhibits inherently higher methylation activity on nucleosome arrays compared to the natural DOT1L complex. Biochemically defined invitro assays helped to show the individual and synergistic effects of these interactions, offering valuable insights into the mechanisms underlying aberrant H3K79 methylation levels in MLL-fusion target loci. These interactions hold potential as therapeutic targets, paving the way for novel therapeutic strategies. Moreover, immobilized template assays reveal a direct interaction between TFIID and the DOT1L complex, with H2Bub and DOT1L protein enhancing the recruitment of TFIID to chromatin. These direct interactions implicate the DOT1L complex in transcriptional initiation. In summary, this study advances our understanding of the critical role played by the DOT1L complex in MLL-rearranged leukemia. By shedding light on the intricate biochemical mechanisms governing MLL fusion protein interactions and their impact on H3K79 methylation, the research opens new avenues for targeted interventions in the treatment of this disease
Recombination and Excision: DNA Repair Proteins in Prokaryotic Host-Virus Conflicts
Full text linkBacteriophages, or simply phages, are viruses that infect bacteria. They are the most abundant biological entity on our planet and outnumber bacteria 10:1 in the ocean. In response to this threat, bacteria have evolved a diverse battery of immune systems that prevent infection, which in turn has resulted in the development of numerous counter-defense mechanisms by phages. This evolutionary arms race drives molecular innovations and presents exciting avenues for the discovery of new molecular biology and new biotechnology tools, such as restriction enzymes and CRISPR-Cas9. My thesis investigates how mechanisms of DNA repair, specifically recombination and base excision, have been co-opted by phages and bacteria to execute non-canonical immune and counter-immune functions in prokaryotic host-virus genetic conflicts. CRISPR-Cas adaptive immune systems, found in nearly half of all bacteria, use sequence-specific guide RNAs to cleave the genetic material of infecting phages. Bacteria and some phages encode recombination systems that could repair the cleaved viral DNA. At the outset of my PhD thesis, it was unknown whether phages could counteract CRISPR-Cas cleavage of phage DNA by repairing CRISPR-induced DNA breaks. Bacteriophage λ, which infects Escherichia coli, encodes the Red system (gam-exo-bet) to promote recombination between related phages. Here, using molecular genetics and sequencing, I show that λ Red mediates evasion of CRISPR-Cas targeting in E. coli. Gam inhibits the host E. coli RecBCD recombination system, allowing recombination and repair of the cleaved DNA by the phage Exo-Beta exonuclease-recombinase. Repair by Exo-Beta promotes mutations, deletions, and genomic rearrangements within the target sequence in phage DNA to prevent recognition by CRISPR. I find that λ Red recombination is strikingly more efficient than the host\u27s RecBCD-RecA recombination pathway in the production of large numbers of phages that escape CRISPR targeting. These findings establish recombination-mediated DNA repair as a novel viral anti-CRISPR strategy that, rather than binding CRISPR-Cas nucleases and impeding their activity, provides a solution to evade the CRISPR-Cas immune response after it has been set off. While recombinases canonically function in DNA repair, my findings reveal an additional role for Red-like recombination systems in countering bacterial immunity, through the protection of phages against sequence-specific nucleases. Based on these findings, I speculate that the counter-immune advantage imparted by Red-like systems may facilitate their spread across bacteriophage genomes. For the second half of my thesis, I set out to discover novel defense systems in bacteria, specifically those that target viral DNA. To achieve this, I pioneered a new screening methodology to discover anti-phage defense systems from unculturable microorganisms using diverse bacterial metagenomic DNA libraries. These metagenomic libraries contain millions of DNA sequences from different microorganisms that are absent in available genetic databases. While bioinformatic mining has led to the discovery of many new bacterial immune systems, the genetic screening of DNA libraries has the advantage of: (i) examining unsequenced DNA, including the dark matter of microbial genomes, and (ii) discovering novel defense systems that cannot be predicted via computational analyses. By subjecting the metagenomic libraries (cloned in E. coli) to phage infection and isolating resistant colonies, I discovered a novel bacterial DNA glycosylase that I named Brig1 (bacteriophage replication inhibition DNAglycosylase1). Brig1 provides immunity against phages that carry hypermodified DNA nucleobases, specifically alpha-glucosyl-hydroxymethylcytosine nucleobases. Brig1 excises these nucleobases from the genome of T-even phages, such as coliphage T4, to generate abasic sites that inhibit DNA replication, which constitutes a novel anti-phage defense mechanism. Many phages have evolved to introduce DNA modifications to avoid recognition and cleavage by bacterial nucleases, including CRISPR-Cas and restriction endonucleases. Brig1 supplies the next step on the bacterial side of the arms race, reestablishing the restriction of phages with modified genomes. Structural predictions suggest that Brig1 is part of a novel family of bacterial anti-phage DNA glycosylases that evolved from uracil DNA glycosylases involved in base excision repair, a pathway that removes mis-incorporated uracil bases from DNA. Interestingly, Brig1homologs are present in multiple phage defense loci across distinct clades of bacteria, and these will be the focus of future studies. Future work will also employ the same metagenomic screening approach, but infecting with phages harboring different DNA modifications, to drive the discovery of new DNA glycosylases, restriction enzymes and DNA repair modules that target modified DNA. Overall, my thesis establishes DNA repair proteins as emerging players in prokaryotic host-virus warfare and will spur future work that explores the co-option of these proteins in varying immune and counter-immune contexts in both bacteria and their viruses