先端技術 1 Technique 1
O3-6-1-1 表現型を復帰することが可能なレトロウイルスベクターを用いた神経細胞の形態形成に必要な遺伝子の同定 Uncovering genes required for neuronal morphology by morphology-based gene trap screening with a revertible retrovirus vector ○原田彰宏1 ○Akihiro Harada1 大阪大学大学院　医学系研究科　細胞生物学1 Dept Cell Biol, Osaka Univ, Osaka1 The molecular mechanisms of neuronal morphology and synaptic vesicle transport have been largely elusive, and only a few of the molecules involved in these processes have been identified. Here, we developed a novel morphology-based gene trap method, which is theoretically applicable to all cell lines, to easily and rapidly identify the responsible genes. Using this method, we selected several gene-trapped clones of rat pheochromocytoma PC12 cells, which displayed abnormal morphology and distribution of synaptic vesicle-like microvesicles (SLMVs). We identified several genes responsible for the phenotypes, and analyzed three genes in more detail. The first gene was BTB/POZ domain-containing protein 9 (Btbd9), which is associated with restless legs syndrome. The second gene was cytokine receptor-like factor 3 (Crlf3), whose involvement in the nervous system remains unknown. The third gene was single-stranded DNA-binding protein 3 (Ssbp3), a gene known to regulate head morphogenesis. These results suggest that Btbd9, Crlf3 and Ssbp3 regulate neuronal morphology and the biogenesis/transport of synaptic vesicles. Because our novel morphology-based gene trap method is generally applicable, this method is promising for uncovering novel genes involved in the function of interest in any cell lines.	O3-6-1-2 神経突起形成関連酵素の情報抽出法 Identification of Key Kinases for Neurite Growth by Computational Classification ○池田和司2, 間島慶2, 山田達也2, 丸野由希2, 作村諭一3, 吉田裕司1 ○Yuji Yoshida1, Kei Majima2, Tatsuya Yamada2, Yuki Maruno2, Yuichi Sakumura3, Kazushi Ikeda2 奈良先端大バイオサイエンス1, 奈良先端大情報2, 愛知県立大3 Graduate School of Biology of Science Nara Institute of Science and Technology,Nara1, Graduate School of Information Science Nara Institute of Science and Technology,Nara, Nara2, School of Information Science and Technology, Aichi Prefectural University, Aichi3 Kinases play important roles in a developing neuron for forming neurite or axon by complex biochemical interactions among them. However, various kinds of the kinases prevent us from identifying key kinases and their combinations for neurite elongation because the number of experimental dataset is much smaller than that of kinases. To develop the computational methodology for estimating key kinases contributing neurite elongation from a limited dataset, in this work we applied machine learning methods to a synthetic dataset generated by a mathematical model of biochemical interactions. Four algorithms, ridge partial least squares (PPLS), naive Bayes classifier, support vector machine, and random forest classification, were used for kinase classification. Of all the algorithms, by reducing data dimension, RPLS has an advantage in classifying a dataset in which each element has a strong correlation with the others. Such a data property is shown in the kinase dataset in the sense that they interact with each other and frequently show a similar biological function. Therefore, it is expected that RPLS can show a better performance in classifying kinase dataset. To test the ability of RPLS, we applied all the four algorithms to the synthetic dataset. As we expected, the comparison result showed that the RPLS is the best classifier, suggesting that dimension reduction is important for classifying kinases.
O3-6-1-3 マウス深部脳機能の経頭蓋光学イメージング Optical imaging of deep brain activity in mice ○澁木克栄1, 塚野浩明1, 駒形成司1, 菱田竜一1 ○Katsuei Shibuki1, Hiroaki Tsukano1, Seiji Komagata1, Ryuichi Hishida1 新潟大学 脳研究所 システム脳生理学分野1 Dept Neurophysiol, Brain Res Inst, Niigata Univ, Niigata, Japan1 In the present study, we visualized neural activities in the deep brain structures of mice using transcranial macroconfocal miscroscopy. A macroconfocal microscope has an objective lense with a large depth of focus, and can visualize deep brain structures depending on the wave length. Using the wave length of 488 nm derived from an Ar lazer, we have succeeded to visualize stimulus-evoked neural activities in the deep cortical layers of S1. In this experiment, we used acitivity-dependent changes in engogenous fluorescence derived from mitochondrial flavoproteins. Using the wave length of 633 nm derived from a Xe lazer, we have succeeded to visualize neural activities elicited by visual stimuli in the lateral geniqulate body (LGB) locationg approximately 2.5 mm deep from the cortical surface. In this experiment, we used activity-dependent changes in blood flow, which are usually monitored as changes in the light absorbance of hemoglobin (Hb). However, Hb signals were reflected in the light reflection of brain tissue, so that they could not be monitored around the center of the optical field, where the light reflection of brain tissue are masked by the light reflection from the surface of optical devises in the macroconfocal microscope. Furthermore, the amplitudes of the activity-dependent Hb signals in LGB were as small as 0.5 %. To avoide these difficulties, we intravenously injected fluorescent albumin labeled with alexa fluor 633 to measure blood flow changes as fluorescence changes. Alternatively, we also used fluorescent red blood corpuscles (RBCs) labeled with DiD, one of lipophilic carbocyanine dyes. Using these methods, the amplitudes of activity-dependent optical changes in LGB were approximately 5-10 times lager than those of Hb signals. Since, fluorescent RBCs are known to be viable in vivo, macroconfocal microscopy of fluorescent RBCs is expected to stably and repeatedly monitor neural activities in deep brain structures of awake mice.	O3-6-1-4 Differential wiring of sister mitral cells in the olfactory bulb revealed by a novel optical clearing agent, SeeDB Differential wiring of sister mitral cells in the olfactory bulb revealed by a novel optical clearing agent, SeeDB ○今井猛1,2 ○Takeshi Imai1,2, Meng-tsen Ke1,3 理化学研究所 発生・再生科学総合研究センター1, JSTさきがけ2, 京都大学大学院生命科学研究科3 RIKEN Center for Developmental Biology1, JST PRESTO2, Graduate School of Biostudies, Kyoto University3 We developed a novel water-based optical clearing agent, SeeDB, which clears fixed mouse brain samples in a few days without quenching many of the fluorescent dyes, including fluorescent proteins and lipophilic neuronal tracers. Unlike previous methods, the sample volume remained constant during the clearing procedure, which is important not only to keep cellular morphology intact, but also to acquire deep tissue images using the limited working distance of objective lenses. We could, for example, image the Thy1-YFP-H mouse brain from dorsal surface to the bottom, using two-photon microscopy. Thus, our method is useful to image detailed morphological architecture, up to scales as large as the whole mouse brain. We used SeeDB to investigate the wiring specificity of the olfactory bulb circuits in the mouse. In the mouse olfactory system, olfactory sensory neurons expressing a given type of odorant receptor converge their axons to a pair of glomeruli in the olfactory bulb. Odor inputs to a single glomerulus are then relayed to 20-50 mitral and tufted cells through their primary dendrites. Although sister mitral cells connecting to the same glomerulus receive correlated excitatory inputs, it remains unknown to what extent the inhibitory modulations are diverse among sister mitral cells. Using single glomerulus labeling with fluorescent dextran tracers, we described the detailed wiring diagram of the sister mitral cells in the bulb. We found that cell bodies of sister mitral cells are widespread to an area corresponding to 20-30 glomeruli in the mitral cell layer. Furthermore, distribution of lateral dendrites was diverse and dissimilar among sister mitral cells. These results suggest that inhibitory inputs are diverse among sister mitral cells and support the idea that odor information is non-redundantly encoded by sister mitral cells.

上部に戻る

前に戻る