Douglas Allan

Next-generation sequencing has made genetic variant discovery routine clinical practice. To assess a variant’s pathogenicity classical clinical genetics methods, require familial transmission or the identification of the variant in multiple affected individuals. However, most identified variants are rare, often de novo, or present in few individuals. Therefore, classical genetics methods such as case-control or co-segregation studies cannot be applied. Computational methods can be used to predict the function of thousands of gene variants, but these methods are not reliable. Experimental assays to assess the relative function of variants can provide a path to interpretation of variant activity. Drosophila melanogaster offers a robust platform for variant functional testing at scale, but remains unvalidated for accurate assessment of the function of high numbers of variants in well-established assays.The first section of my thesis describes my role in a multi-model organism variant functionalization project characterizing 100 human PTEN variants. In flies, ubiquitous overexpression of PTEN-WT resulted in developmental delay. By comparing eclosion delay for overexpressed PTEN variants, we categorized variants as wildtype, gain of function or loss of function. This work demonstrates the utility of Drosophila as a powerful platform for high volume screening for the function of human gene variants from healthy and diseased individuals.However, clinically interpreting the functional consequence of the identified variation is challenging. Therefore, my third chapter aimed to establish a structured ‘well-established’ assay calibrated against known pathogenic and benign PTEN variants to convert variant function in an assay into a clinically-relevant interpretation in adherence to ClinGen Sequence Variant Interpretation (SVI) Working Group recommended guidelines. We screened ~100 human PTEN variants for suppression of PI3K/AKT signaling dependent cellular proliferation in Drosophila, a pathway that underlies the primary cause of PTEN’s contribution to cancer. The assay correctly assigned the function of known pathogenic and benign variants and exhibited a high correlation with human cell line functional data. We also showed that PTEN functionally replaces its Drosophila ortholog in developmental growth. Our work provides evidence that well- established assays, directly testing disease-relevant protein activity in Drosophila, can be used to generate reliable functional data appropriate for clinical variant interpretation.

View record

Bone morphogenetic proteins (BMPs) are a group of phylogenetically conserved signaling proteins, first identified to play important roles in bone formation. Since their discovery, they have been recognized to contribute to embryonic development and adult homeostasis in a multitude of tissues, by regulating cellular lineage commitment, morphogenesis, differentiation, proliferation, and apoptosis. BMPs transduce their signals through intracellular downstream effectors, primarily the Smad transcription factors, many of which bind to genomic BMP-responsive cis-regulatory elements (BMP-CREs) to direct gene expression. Despite their importance in cellular processes and maintenance, BMP-CREs remain largely unidentified at a genomic level for most BMP-dependent cellular processes.The overall objectives of this thesis were to experimentally characterize the widespread function of a novel low-affinity BMP-CRE motif in the Drosophila nervous system and to identify the BMP-driven regulatory network underlying mammalian chondrogenesis.To address the first goal, we used computational methods to identify this novel BMP-CRE through the Drosophila genome and used in vivo transgenic reporters to determine their function in the Drosophila nervous system. Our results show that this BMP-CRE is used within multiple enhancers to mediate their BMP-dependent activity. For our second goal, we used poly-A transcriptome sequencing (RNA-seq) to characterize differentially expressed genes (DEGs) during chondrogenesis in primary murine cells. Amongst these DEGs, we identified transcription factors/cofactors with previously unknown roles in chondrogenesis that are of interest for further study. Further, we used histone modification ChIP-seq to identify more than 2000 candidate regulatory regions in the vicinity of BMP-responsive DEGs. Using computational tools, we examined these candidate regulatory regions for Smad-binding sites using BMP-CRE motifs identified in Drosophila. We then applied multiple selection criteria to prioritize likely BMP-responsive regulatory regions and assessed four novel regions for BMP-responsive reporter expression, using mouse primary limb mesenchymal (PLM) cells. Among these, we identified two BMP-responsive regulatory regions, including one within 50kb of the transcription factor Jdp2, a gene with previously unknown roles in chondrogenesis.The genomic mapping of BMP-CREs remains incomplete. Mutations in these cis-regulatory sites and BMP-regulated genes could potentially result in disease, and therefore their identification is of critical importance to help further our understanding of disorders in various human tissues.

View record

This thesis describes two projects that examined protein function in neural plasticity and cancer development. The first project created assays that tested the pathogenicity of human variants identified in SMAD4 and BMPR1A, genes that are associated with juvenile polyposis syndrome. As exome sequencing becomes easier and more cost-effective, many human variants are being identified. However, for most variants, their impact on protein function and their ability to cause disease are unknown. Drosophila melanogaster offers an efficient system for testing human variant protein functionality in a panel of assays to screen through many variants. I have used simple overexpression assays in Drosophila to test human SMAD4 and BMPR1A variants. I developed two assays in which wildtype SMAD4, but not loss of function variants, caused either lethality or wing vein defects. I screened through seven human SMAD4 variants implicated in disease to assess their relative function and identified four that exhibit functional differences to wildtype. I also tested human BMPR1A but found that overexpression of this gene in Drosophila had no effect. I postulated this is due to a lack of ligand binding. Therefore, I created reagents for alternative methods to screen BMPR1A variant function. First, I generated mimetic mutations in the orthologous tkv gene. Second, I created a chimeric gene comprising the extracellular domain of Tkv and the intracellular domain of BMPR1A. I postulate that this chimera should bind Drosophila BMP ligands and activate canonical BMP signaling, allowing for assays of BMPR1A variants in the intracellular domain. These reagents and assays are important for experimentally determining ariant activity and for improving our understanding of structure/function relationships for SMAD4 and BMPR1A. Going forward, functionally testing large numbers of variants will inform personalized medicine approaches and improve computer models for projecting pathogenicity of human variants.In the second project, I created CalpA and CalpB double mutants to test whether a reduction of calpain activity could stimulate de novo neurite formation. Also, I overexpressed a proteolytic target of Calpain, Cortactin, and created a Calpain proteolysis resistant version of Cortactin. Surprisingly, I was unable to identify any phenotype in the nervous system.

View record

Terminal differentiation of neurons often requires retrograde signals from the target cells they innervate, which trigger neural subtype-specific gene expression upon target contact. Target-derived BMP signaling and transcription factors including the LIM-Homeodomain transcription factor Apterous are required for FMRFa neuropeptide gene expression in Drosophila Tv4 neurons. We modeled the integrative mechanism of these extrinsic and intrinsic inputs at a Tv4 neuron-specific FMRFa enhancer. We show that Tv4-specific FMRFa expression requires two separable cis-elements, a BMP-response element (BMP-RE) that binds Mad, and a homeodomain response element (HD-RE) that binds Apterous. Strikingly, we find that concatemers of these two short (~30bp) cis-elements each independently drive spatial and temporal expression appropriate for Tv4-specific FMRFa. Thus, specific and robust expression is generated from the synergy of two low-activity heterotypic cis-elements that encode the same output from distinct inputs. We further examined the timing mechanism of FMRFa initiation, which models predict would be solely based on target contact. In contrast, we find that the timed downregulation of the COUP-TF I/II nuclear receptor Seven up functions to de-repress HD-RE and BMP-RE activity immediately prior to target contact. Thus, we reveal that the active suppression of neurotransmitter identity, prior to target contact, is an innate component of the target-dependent mechanism for timed gene activation.Further examination of the FMRFa BMP-RE shows that the sequence of the Mad and Medea binding sites to be “perfectly wrong.” It displays sequence similarity to both the previously characterized BMP Silencing Element (SE) and Activating Elements (AE) but with base substitutions that should abrogate Smad binding. Biochemical and reporter construct analysis demonstrate that the FMRFa BMP-RE is an atypical activating element that is specifically attenuated in its ability to interact with the co-Smad Medea. In vivo reporter assays show that this attenuation is required to drive appropriate expression in the CNS. These findings represent only the third verified type of BMP-RE found in Drosophila leading us to call this class an AE2 element.

View record

Over the past years, numerous studies have advanced our understanding of the generation and function of the sex-specific neuronal populations that control sex-specific behaviours. Prior to the work presented in this thesis, no female-specific subsets of neurons had been identified in Drosophila; thus, all models and studies of sex-specific neurons have had a male bias.This thesis describes the first identification and characterization of a female-specific neuronal population in the central nervous system of Drosophila, the Ilp7-motoneurons. These neurons innervate the oviduct and are required for egg-laying. We further identified cellular and genetic mechanisms that direct the dimorphic generation of these female-specific neurons. Programmed cell death of post-mitotic nascent Ilp7-motoneurons in males accounts for their female-specific generation in a process regulated by a non-canonical and dosage-sensitive pro-apoptotic role for the male fruitless isoform (fruM). Thus, we find that analysis of female-specific neuron generation unveils novel mechanisms of dimorphic nervous system construction.Our characterization of Ilp7-motoneurons led to a collaboration with Eric Lai (Sloan Kettering, USA), to study the neuronal basis of the female sterility phenotype of the ∆mir mutant, a deficiency in the bidirectional mir-iab-4 and mir-iab-8 miRNA locus of the Bithorax-Complex. We find that female sterility arises from derepression of mir-iab-4/8 targets, Ultrabithorax and homothorax, in fru-expressing neuronal populations of the posterior abdominal segments of the ventral nerve cord. This results in numerous phenotypes that each likely contribute to sterility. ∆mir females have reduced Ilp7-motor innervation of the oviduct. ∆mir virgin females are constitutively unreceptive to males; however, if mated, they fail to increase egg production. Our data suggests a novel mechanism that may explain this phenotype; after mating, sex peptide from the male seminal fluid is retained in the female reproductive tract, rather than being transferred to the hemolymph, where it is believed to effect the increase in egg production. Ongoing work aims to identify the neuronal populations that are disrupted in ∆mir mutants. Taken together, this thesis provides novel insight and models to further our understandingof female-specific neuronal differentiation, a field that has long been under-represented in theliterature.

View record

Appropriate integration of neurons into functioning networks is the ultimate goal ofneuronal differentiation. My thesis examined the mechanism and timing of neuronaldifferentiation in Drosophila melanogaster in relation to the functional requirements of adeveloping neuronal network.Specifically, my thesis aimed to address the role of extrinsic signaling in inducing andmaintaining the expression of genes important to the function of neurons within their network.CCAP-neurons were chosen as a model because: 1) GAL4 drivers are available for cell-specificgenetic manipulation. 2) The critical role of CCAP-neurons in the behavior, ecdysis, provides foran easily assayed phenotype if these neurons fail to function properly. 3) Four peptide hormonesare selectively expressed in differentiated CCAP-neurons that are essential for the normalfunction of CCAP-neurons in ecdysis; this provides a direct link between gene expression andbehavior. 4) Ecdysis is reiterated at multiple developmental steps, thus the CCAP-neuronalpopulation functions throughout development. Together, these factors allow my work to relateneuronal subtype-specific differentiation to the regulation of gene expression and then directly tobehavior.Larval Drosophila CCAP-neurons comprise ~46 neurons [~36 interneurons (CCAP-INs)and 10 efferent-neurons (CCAP-ENs)] that express a number of terminal differentiation genes(TDGs; such as neuropeptides). To begin, we delineated mechanisms underlying the expressionof four TDGs, the peptide hormones CCAP, MIP, Bursicon-α and Bursicon-β, which togethermediate the functional output of those neurons. Importantly, my studies found that a specificsubset of CCAP-neurons, the CCAP-ENs, is both necessary and sufficient for ecdysis, and thattheir function in ecdysis is mediated by extrinsic BMP-dependent peptide hormone expression.Additionally, we found that the change in the ecdysis behavioral sequence from larval to pupalecdysis is supported by the recruitment of a ‘late’ subset of CCAP-neurons that are born in theembryo but undergo extrinsic ecdysone-triggered, temporally-tuned differentiation immediatelyprior to pupal ecdysis.

View record

The diversification of cellular subtype during development is directed by combinatorially acting transcription factors and signaling pathways that act to regulate subtype specific gene expression profiles in post-mitotic cells. These key transcription factors and signaling pathways operate in a transcriptional network, which act to establish cellular subtype identity over the course of a developing cellular lineage. Lineage progression towards ever increasing cellular diversity is often viewed as a ratchet mechanism of irreversible steps resulting in the specification and then terminal differentiation of cell subtype identities. From this viewpoint, terminally differentiated cells have long been considered as irreversibly locked into their identity. In a landmark article, Blau and Baltimore (Blau and Baltimore, 1991) postulated that a cell’s identity, or differentiated status, requires persistent active regulation, rather than lapsing into a passive ‘locked-in’ state. While little genetic evidence was available at the time, sufficient evidence has since accumulated to propose that the terminally differentiated state, or identity, of a cell subtype indeed requires active maintenance. Currently, however, we have only the most rudimentary understanding of the regulatory mechanisms that maintain neuronal identity. This thesis presents a systematic effort to characterize the role of the transcriptional networks that differentiate neuronal identity in the mature neurons of the adult nervous system. Using the Drosophila Tv cluster neurons I show the persistent requirement of 1) target derived signals and 2) networks of transcription factors for the maintenance of the cellular subtype specific expression profiles of terminal differentiation genes, genes that define these neuron’s function and identity. This work establishes one of the most comprehensive transcriptional models for maintenance of cell identity to date. It also provides novel mechanistic insights showing that cellular differentiation is a persistent process that requires active maintenance, rather than being passively ‘locked-in’ or unalterable. As such, the work of this thesis provides critical insight that provides a strong foundation for further efforts to determine how neuronal identity is maintained.

View record