MOLECULAR, BIOPHYSICAL AND INTEGRATIVE MECHANISMS OF MEMORY IN BRAIN NETWORKS

LAS研究脑网络中联想记忆的分子生物物理和综合基础。LAS观察到活体动物的学习和记忆行为与神经元网络中的信号处理和亚细胞分子级联有关。我们的数据暗示了软体动物和哺乳动物物种中保守的分子和生物物理机制，因此可能与人类的学习和记忆有关。对Hermissenda蜗牛（巴甫洛夫/经典条件反射）、兔子（经典条件反射）和大鼠（空间迷宫学习、嗅觉辨别）的联想记忆的细胞分析揭示了记忆形成过程中的一系列细胞和亚细胞事件。这些事件包括：gaba能抑制向兴奋的长期突触转化（LTT）；细胞内钙和DAG升高；PKC易位；pkc介导的Ca2+和gtp结合蛋白cp20（也称为Calexcitin）的磷酸化：电压依赖性K+通道的失活；学习特异性基因转录调控；以及突触末端分支的重排。其他信号蛋白，如map激酶和ras，最近也被认为与记忆存储的长时间域有关。其他分子生物学工具，如反义，已经帮助识别参与长期记忆的神经元树突上的特定离子通道。最先进的分子生物学筛选技术在LAS最近暗示了新的生化步骤记忆存储。最近的观察发现，特定的晚期基因经历了长时间的激活，直到记忆巩固期。这些基因已经通过Northern blot分析、逆转录pcr和原位杂交技术得到证实。其中一种与记忆相关的基因编码II型ryanodine受体。ryanodine受体（RR）是一个450 k的钙通道，具有10个跨膜结构域。该受体负责内质网钙介导的钙释放（CICR）。最近，LAS研究发现了第一个已知的以钙依赖方式激活神经元RR的信号蛋白。这种蛋白是Calexcitin (cp20)，在联想学习和记忆过程中被PKC的α -同工酶磷酸化。钙激素最近也被LAS科学家证明能有效激活内质网上的Ca2+- atp酶。其他分析表明，一个编码核苷酸负责cal兴奋蛋白c端p环结构域的存在。p环导致钙激素的Ca2+- atp酶激活效应增加4倍。最近的其他研究已经将钙激素-红嘌呤受体级联与长期突触修饰如LTT（长期转化）和LTP联系起来。该级联的激活可诱导LTT持续数小时，但可阻止LTP的发生。后者的发现与LAS研究一致，该研究清楚地将大鼠空间迷宫学习与短期和长期LTP分离开来。RT-PCR和特异性抗体Western blot分析显示，将Kv1.4的反义寡脱氧核糖核苷酸微注射到大鼠脑室内，阻断了海马Kv1.4 mRNA，敲低了海马蛋白。这种反义敲除对大鼠空间迷宫学习、记忆和探索行为没有影响，但消除了早期和晚期LTP，降低了CA1锥体神经元的配对脉冲促进（一种突触前效应），而不影响齿状回LTP。突触前Kv1.4敲低和之前的突触后Kv1.1敲低表明，CA1 LTP对于大鼠空间记忆既不是必要的，也不是充分的。在过去的一年中，对大鼠空间迷宫的注意门控学习的突触基础的重大见解已经被发现。简而言之，在注意聚焦过程中，在大鼠海马中诱导了胆碱能介导的θ节律。θ节律伴随着海马锥体细胞接收的gaba能突触的长期转化（LTT）。阻断LTT的潜在电导（碳酸氢盐）会阻断θ节律并阻止迷宫学习的获得。在CA1子集或集合中，迷宫记忆伴随着gaba能突触逆转电位向碳酸氢盐的长期转移。这些发现和其他发现构成了在联想记忆巩固过程中细胞内钙反复和长时间动员的似是而非的分子级联的基础。最近，一种概念性的综合（TINS, 1998）出现在LAS发现的上述分子事件中，这些事件被证明发生在哺乳动物的大脑学习过程中。记忆的时间域与钙信号增强的时间域相对应。相关的训练刺激使PKC易位，激活Calexcitin，使外膜上电压依赖性的K+通道失活，激活ryanodine受体和Ca2+- atp酶，以放大神经元内，如LAS研究所暗示的，树突内钙波。因此，这些连续的分子事件可能参与了大脑中记忆表征的形成，使其在以后的回忆中更加持久。这些事件在不同物种的记忆中暗示了进化过程中的保护。这种跨物种的保护表明，人类类似的联想记忆机制可能是阿尔茨海默病功能障碍的靶点。最近在LAS开发的阿尔茨海默病诊断证实了上述Ca++信号级联与人类记忆的相关性。基于这种级联，LAS预测特定分子的功能，如PKC的a同工酶、电压依赖性IA和I Ca2+-K+通道、钙激素和内质网上的细胞内钙释放受体，将在阿尔茨海默病中受到损害。许多研究（包括最近在校外实验室进行的研究）一致地证实了这些预测，从而确定了阿尔茨海默病的生理诊断方法并获得了专利。因此，这种钙信号级联在人类记忆功能障碍中起重要作用，并可能成为阿尔茨海默病治疗的靶点。最后，从基于大脑的记忆网络衍生的理论结构也被数学描述并纳入基于计算机的人工网络，这些网络已经展示了显著的模式识别能力。未来的方向将包括对级联事件的进一步分子表征，如确定钙激素对红嘌呤受体和Ca2+- atp酶激活的精确活性位点，对哺乳动物钙激素的全面表征，通过钙激素-红嘌呤受体级联诱导树突、突触和形态转化，参与阿尔茨海默病和智力迟钝的级联步骤，以及相关的亚细胞事件与记忆储存的持续关联。

The LAS studies the molecular biophysical, and integrative bases of associative memory in brain networks. LAS observations have related learning and memory behavior of living animals to signal processing in neuronal networks and to subcellular molecular cascades. Our data have implicated molecular and biophysical mechanisms that are conserved in molluscan and mammalian species and thus could have relevance for human learning and memory. Cellular analyses of associative memory in the snail Hermissenda (Pavlovian/classical conditioning), the rabbit (classical conditioning), and the rat (spatial maze learning, olfactory discrimination) revealed a cascade of cellular and subcellular events during memory formation. These events include: long-term synaptic transformation of GABAergic inhibition into excitation (LTT); elevation of intracellular calcium and DAG; translocation of PKC; PKC-mediated phosphorylation of the Ca2+ and GTP-binding protein, cp20 (also called Calexcitin): inactivation of voltage-dependent K+ channels; learning-specific regulation of gene transcription; and rearrangement of synaptic terminal branches. Other signaling proteins such as map kinase and ras have also recently been implicated in longer time domains of memory storage. Other molecular biologic tools such as antisense have helped to identify specific ionic channels on neuronal dendrites that participate in long-term memory. State-of-the-art molecular biologic screening techniques in the LAS have recently implicated new biochemical steps in memory storage. Recent observations have uncovered specific late genes that undergo prolonged activation well into the period of memory consolidation. These genes have been confirmed with Northern blot analyses, reverse transcriptase-PCR, and in situ hybridization techniques. One of these memory-related genes encodes the type II ryanodine receptor. The ryanodine receptor (RR) is a 450 k calcium channel that has 10 membrane-spanning domains. This receptor is responsible for calcium-mediated calcium release (CICR) from the endoplasmic reticulum. Very recently LAS studies identified the first known signaling protein that activates the neuronal RR in a calcium-dependent manner. This protein is Calexcitin (cp20) which was previously shown to be phosphorylated by the alpha-isozyme of PKC during associative learning and memory. Calexcitin has also recently been shown by LAS scientists to potently activate the Ca2+-ATPase on the ER. Other analyses revealed that a single coding nucleotide is responsible for the presence of a P-loop domain on the C-terminal end of calexcitin. The P-loop causes a 4-fold increase of the Ca2+-ATPase activation effect of calexcitin. Other recent studies have linked the calexcitin-ryanodine receptor cascade to long term synaptic modifications such as LTT (long-term transformation) and LTP. Activation of this cascade can induce LTT for hours but prevent the occurrence of LTP. The latter finding is consistent with LAS research that has clearly dissociated rat spatial maze learning from short and long-term LTP. RT-PCR and Western blot analysis with specific antibodies showed that antisense oligodeoxyribonucleotide to Kv1.4 microinjected intraventricularly into rat brains obstructed hippocampal Kv1.4 mRNA, knocking-down the protein in the hippocampus. This antisense knockdown had no effect on rat spatial maze learning, memory or exploratory behavior, but eliminated both early and late phase LTP and reduced paired-pulse facilitation ( a pre-synaptic effect) in CA1 pyramidal neurons without affecting dentate gyrus LTP. This presynaptic Kv1.4 knockdown together with previous post-synaptic Kv1.1 knockdown demonstrates that CA1 LTP is neither necessary nor sufficient for rat spatial memory. Major insights into the synaptic basis of attention-gated learning of the rat spatial maze have been uncovered during the past year. In brief, during attentional focusing a cholinergic-mediated theta rhythm is induced in the rat hippocampus. The theta rhythm is accompanied by long-term transformation (LTT) of GABAergic synapses received by the hippocampal pyramidal cells. Blocking the underlying conductance (bicarbonate) of LTT blocks the theta rhythm and prevents acquisition of the maze learning. Within a CA1 subset or ensemble, memory of the maze is accompanied by long-term shifts of GABAergic synaptic reversal potentials toward that of bicarbonate. These and other findings form the basis of a plausible molecular cascade for repeated and prolonged mobilization of intracellular calcium during consolidation of associative memory. A conceptual synthesis (TINS, 1998) has emerged recently from LAS discovery of the above molecular events demonstrated to occur during learning in the mammalian brain. Time domains of memory correspond to time domains of enhanced calcium signaling. Associated training stimuli translocate PKC, activate Calexcitin, inactivate voltage-dependent K+ channels on the outer membrane and activate the ryanodine receptor and Ca2+-ATPase to amplify intraneuronal, and as LAS studies have implicated, intradendritic calcium waves. These sequential molecular events could participate, therefore, in making memory representations in the brain more permanent for later recall. Implication of these events in memory of diverse species suggests conservation during evolution. Such conservation across species suggests that comparable associative memory mechanisms in humans may provide targets of dysfunction in Alzheimers disease. Recent corroboration of Alzheimers diagnostics developed in the LAS supports the relevance of the above Ca++ signaling cascade for human memory. Based on this cascade, the LAS predicted that function of specific molecules such as the a isozyme of PKC, the voltage-dependent IA and I Ca2+-K+ channels, calexcitin, and intracellular calcium release receptors on the ER will be compromised in Alzheimers disease. These predictions have been borne out so consistently by numerous studies (including those recently conducted at extramural laboratories) that physiologic Alzheimers diagnostic measures have been identified and patented. Thus, this calcium signaling cascade is implicated as important for human memory dysfunction and possibly as a target for AD therapeutics. Finally, theoretical constructs derived from brain-based memory networks have also been mathematically described and incorporated into computer-based artificial networks which have demonstrated significant pattern recognition capabilities. Future directions will include further molecular characterization of events within this cascade such as determining the precise active sites in calexcitin for ryanodine receptor and Ca2+-ATP-ase activation, full characterization of mammalian calexcitin, induction of dendritic, synaptic, and morphologic transformations by the calexcitin-ryanodine receptor cascade, involvement of cascade steps in Alzheimers disease and mental retardation, as well as continued correlations of relevant subcellular events with memory storage.