Animals and their behavior have been shaped by evolution to succeed in their respective natural environments. The ability to learn, memorize and retrieve information has proven to be an invaluable property in the success of most species living on earth today. The brain, be it as small as 20,000 neurons in honeybees or as large as 257 billion neurons in African elephants, is generally considered to be the organ that enables animals to learn about and interact with their environments. One brain region in particular, the mammalian hippocampus, is indispensable for many types of learning and has drawn the attention of thousands of researches over the last decades. As with all other brain regions, the primary functional building blocks of the hippocampus are neurons, which communicate via synapses. Synaptic transmission can vary in efficacy depending on various factors. This variation in transmission efficacy, often called synaptic plasticity, is accompanied by structural and compositional synaptic changes facilitated by an increase, decrease, displacement and rearrangement of numerous synaptic proteins. Synaptic plasticity is generally considered the basis of learning and memory.
TRIM3, the subject of this thesis, is one such synaptic protein. It is a RING E3 ubiquitin ligase that is expressed at synapses in the hippocampus and the cerebellum and was supposed to regulate ubiquitin-proteasome-mediated degradation of particular synaptic protein substrates. However, at the start of the research described in this thesis, its precise role and function in neurons and especially at synaptic sites had remained elusive. In order to understand the role of TRIM3 in hippocampal and cerebellar function, we made use of a newly generated Trim3 knockout mouse and investigated it in a multidisciplinary manner.
Chapter 2 focuses on the descriptive characterization of TRIM3 and on the effect of its loss on brain morphology, learning behavior and synaptic function. We used immunohistochemistry, immunocytochemistry, sub-cellular fractionation and overexpression experiments to show that TRIM3 protein is highly expressed at synaptic sites of the hippocampus and cerebellum and that it is indeed an ubiquitin ligase. We found that knocking out Trim3 has no overall effect on gross brain morphology and cytoarchitecture, and that basal locomotor and anxiety-related behaviors are unaffected. Hippocampus-dependent fear memory, however, was significantly enhanced in Trim3-/- mice, in particular short-term memory measured at two hours after conditioning. By using an electrophysiological approach we showed that hippocampal LTP is enhanced in Trim3-/- mice, but that spontaneous activity of hippocampal pyramidal cells is indistinguishable from that of wildtype neurons. Finally, we showed that loss of TRIM3 causes an increase in spine density in hippocampal pyramidal cells. Together these findings support a role for TRIM3 in structural and functional hippocampal synaptic plasticity.
Chapter 3 explores the possibility that the behavioral, electrophysiological and structural abnormalities of Trim3-/- mice described in chapter 2 are caused by deficits in the trafficking properties of messenger ribonucleoprotein (mRNP) particles. We used biochemistry to demonstrate that TRIM3 is part of PURA-containing mRNP granules in hippocampus. By co-immunoprecipitation and overexpression experiments we showed that TRIM3 interacts with PURA, one of the main protein constituents of mRNPs, but that the interaction is mediated by RNA and not by direct contact of TRIM3 and PURA. We further demonstrated that PURA abundance is unaffected by the presence or absence of TRIM3, making PURA an unlikely ubiquitylation substrate. We furthermore used time lapse imaging to analyze mRNP trafficking and mobility in cultured neurons, to find that TRIM3 is not essential for mRNP trafficking, but that the loss of TRIM3 does result in a small but significant increase in travelled distance and velocity of a small subpopulation of mRNP particles.
Chapter 4 describes a novel strategy to identify TRIM3 ubiquitylation substrates. We applied Tandem Ubiquitin Binding Entities (TUBEs) affinity purification and immunoprecipitation with three different TRIM3 antibodies in parallel to Trim3-/- mice and wildtype controls. Using this universally applicable, two-pillared quantitative mass spectrometry based strategy we identified synaptic TRIM3 interactors on the one side and differentially ubiquitylated synaptic proteins on the other. We demonstrate the suitability and reproducibility of our approach and present γ-actin (ACTG1) as the most likely ubiquitylation substrate of TRIM3 in hippocampal synapses.
Chapter 5 focuses on the validation of ACTG1 as a TRIM3 substrate in the hippocampus. By co-immunoprecipitating ACTG1 with TRIM3 and TRIM3 with ACTG1 from hippocampal synaptic fractions and from HEK293 cells after TRIM3 overexpression, we were able to show that TRIM3 and ACTG1 interact directly. We further demonstrate that loss of TRIM3 leads to an increase in ACTG1 in hippocampal neurons, and that overexpressing TRIM3 in heterologous cells causes a decrease in endogenous ACTG1 levels.
This decrease is caused by ubiquitylation and subsequent proteasomal degradation mediated by TRIM3. Finally, the importance of ACTG1 in learning and memory is evidenced for the first time by generating a forebrain specific Actg1 knockout mouse. These mice are viable, and show similar memory deficits in a contextual fear conditioning paradigm as Trim3-/- mice. However, whereas Trim3-/- mice showed increased short-term memory at two hours after conditioning, conditional Actg1 mice do so at 24 - 72 hours after conditioning. Taken together our findings suggest that temporal control of ACTG1 levels by TRIM3 is required to regulate the timing of hippocampal plasticity.
Chapter 6 Summarizes and discusses obtained data, placing TRIM3 and ACTG1 into a working model in which TRIM3, by fine tuning the expression, degradation and potentially polymerization of ACTG1, keeps the acquisition, consolidation and later expression of a fear memory within physiological boundaries, allowing individuals to function and properly react to their environment.