A practical guide to data management and sharing for biomedical laboratory researchers.

Effective data management and sharing have become increasingly crucial in biomedical research; however, many laboratory researchers lack the necessary tools and knowledge to address this challenge. This article provides an introductory guide into research data management (RDM), and the importance of FAIR (Findable, Accessible, Interoperable, and Reusable) data-sharing principles for laboratory researchers produced by practicing scientists. We explore the advantages of implementing organized data management strategies and introduce key concepts such as data standards, data documentation, and the distinction between machine and human-readable data formats. Furthermore, we offer practical guidance for creating a data management plan and establishing efficient data workflows within the laboratory setting, suitable for labs of all sizes. This includes an examination of requirements analysis, the development of a data dictionary for routine data elements, the implementation of unique subject identifiers, and the formulation of standard operating procedures (SOPs) for seamless data flow. To aid researchers in implementing these practices, we present a simple organizational system as an illustrative example, which can be tailored to suit individual needs and research requirements. By presenting a user-friendly approach, this guide serves as an introduction to the field of RDM and offers practical tips to help researchers effortlessly meet the common data management and sharing mandates rapidly becoming prevalent in biomedical research.

Computerized detection and segmentation of mitochondria on electron microscope images.

Mitochondrial function plays an important role in the regulation of cellular life and death, including disease states. Disturbance in mitochondrial function and distribution can be accompanied by significant morphological alterations. Electron microscopy tomography (EMT) is a powerful technique to study the 3D structure of mitochondria, but the automatic detection and segmentation of mitochondria in EMT volumes has been challenging due to the presence of subcellular structures and imaging artifacts. Therefore, the interpretation, measurement and analysis of mitochondrial distribution and features have been time consuming, and development of specialized software tools is very important for high-throughput analyses needed to expedite the myriad studies on cellular events. Typically, mitochondrial EMT volumes are segmented manually using special software tools. Automatic contour extraction on large images with multiple mitochondria and many other subcellular structures is still an unaddressed problem. The purpose of this work is to develop computer algorithms to detect and segment both fully and partially seen mitochondria on electron microscopy images. The detection method relies on mitochondria's approximately elliptical shape and double membrane boundary. Initial detection results are first refined using active contours. Then, our seed point selection method automatically selects reliable seed points along the contour, and segmentation is finalized by automatically incorporating a live-wire graph search algorithm between these seed points. In our evaluations on four images containing multiple mitochondria, 52 ellipses are detected among which 42 are true and 10 are false detections. After false ellipses are eliminated manually, 14 out of 15 fully seen mitochondria and 4 out of 7 partially seen mitochondria are successfully detected. When compared with the segmentation of a trained reader, 91% Dice similarity coefficient was achieved with an average 4.9 nm boundary error.

A hybrid human and machine resource curation pipeline for the Neuroscience Information Framework.

The breadth of information resources available to researchers on the Internet continues to expand, particularly in light of recently implemented data-sharing policies required by funding agencies. However, the nature of dense, multifaceted neuroscience data and the design of contemporary search engine systems makes efficient, reliable and relevant discovery of such information a significant challenge. This challenge is specifically pertinent for online databases, whose dynamic content is 'hidden' from search engines. The Neuroscience Information Framework (NIF; http://www.neuinfo.org) was funded by the NIH Blueprint for Neuroscience Research to address the problem of finding and utilizing neuroscience-relevant resources such as software tools, data sets, experimental animals and antibodies across the Internet. From the outset, NIF sought to provide an accounting of available resources, whereas developing technical solutions to finding, accessing and utilizing them. The curators therefore, are tasked with identifying and registering resources, examining data, writing configuration files to index and display data and keeping the contents current. In the initial phases of the project, all aspects of the registration and curation processes were manual. However, as the number of resources grew, manual curation became impractical. This report describes our experiences and successes with developing automated resource discovery and semiautomated type characterization with text-mining scripts that facilitate curation team efforts to discover, integrate and display new content. We also describe the DISCO framework, a suite of automated web services that significantly reduce manual curation efforts to periodically check for resource updates. Lastly, we discuss DOMEO, a semi-automated annotation tool that improves the discovery and curation of resources that are not necessarily website-based (i.e. reagents, software tools). Although the ultimate goal of automation was to reduce the workload of the curators, it has resulted in valuable analytic by-products that address accessibility, use and citation of resources that can now be shared with resource owners and the larger scientific community. DATABASE URL: http://neuinfo.org.

Prototype of Kepler Processing Workflows For Microscopy And Neuroinformatics.

We report on progress of employing the Kepler workflow engine to prototype "end-to-end" application integration workflows that concern data coming from microscopes deployed at the National Center for Microscopy Imaging Research (NCMIR). This system is built upon the mature code base of the Cell Centered Database (CCDB) and integrated rule-oriented data system (IRODS) for distributed storage. It provides integration with external projects such as the Whole Brain Catalog (WBC) and Neuroscience Information Framework (NIF), which benefit from NCMIR data. We also report on specific workflows which spawn from main workflows and perform data fusion and orchestration of Web services specific for the NIF project. This "Brain data flow" presents a user with categorized information about sources that have information on various brain regions.

Multiscale imaging characterization of dopamine transporter knockout mice reveals regional alterations in spine density of medium spiny neurons.

The dopamine transporter knockout (DAT KO) mouse is a model of chronic hyperdopaminergia used to study a wide range of neuropsychiatric disorders such as schizophrenia, attention deficit hyperactivity disorder (ADHD), drug abuse, depression, and Parkinson's disease (PD). Early studies characterizing this mouse model revealed a subtle, but significant, decrease in the anterior striatal volume of DAT KO mice accompanied by a decrease in neuronal cell body numbers (Cyr et al., 2005). The present studies were conducted to examine medium spiny neuron (MSN) morphology by extending these earlier reports to include multiscale imaging studies using correlated light microscopy (LM) and electron microscopy (EM) techniques. Specifically, we set out to determine if chronic hyperdopaminergia results in quantifiable or qualitative changes in DAT KO mouse MSNs relative to wild-type (WT) littermates. Using Neurolucida Explorer's morphometric analysis, we measured spine density, dendritic length and synapse number at ages that correspond with the previously reported changes in striatal volume and progressive cell loss. Light microscopic analysis using Neurolucida tracings of photoconverted striatal MSNs revealed a highly localized loss of dendritic spines on the proximal portion of the dendrite (30 µm from the soma) in the DAT KO group. Next, thick sections containing MSN dendritic segments located at a distance of 20-60 µm from the cell soma, a region of the dendrite where spine density is reported to be the highest, were analyzed using electron microscope tomography (EMT). Because of the resolution limits of LM, the EM analysis was an extra measure taken to assure that our analysis included nearly all spines. Spine density measurements collected from the EMT data revealed only a modest decrease in the DAT KO group (n=3 mice) compared to age-matched WT controls (n=3 mice), a trend that supports the LM findings. Finally, a synaptic quantification using unbiased stereology did not detect a difference between DAT KO mice (n=6 mice) and WT controls (n=7 mice) at the EM level, supporting the focal nature of the early synaptic loss. These findings suggest that DAT KO mice have MSNs with highly localized spine loss and not an overall morphologically distinct cell shape. The characterization of morphological changes in DAT KO mice may provide information about the neural substrates underlying altered behaviors in these mice, with relevance for human neurological disorders thought to involve altered dopaminergic homeostasis. Results from this study also indicate the difficulty in correlating structural changes across scales, as the results on fine structure revealed thus far are subtle and non-uniform across striatal MSNs. The complexities associated with multiscale studies are driving the development of shared online informatics resources by gaining access to data where it is being analyzed.

Automated microscopy system for mosaic acquisition and processing.

An automatic mosaic acquisition and processing system for a multiphoton microscope is described for imaging large expanses of biological specimens at or near the resolution limit of light microscopy. In a mosaic, a larger image is created from a series of smaller images individually acquired systematically across a specimen. Mosaics allow wide-field views of biological specimens to be acquired without sacrificing resolution, providing detailed views of biological specimens within context. The system is composed of a fast-scanning, multiphoton, confocal microscope fitted with a motorized, high-precision stage and custom-developed software programs for automatic image acquisition, image normalization, image alignment and stitching. Our current capabilities allow us to acquire data sets comprised of thousands to tens of thousands of individual images per mosaic. The large number of individual images involved in creating a single mosaic necessitated software development to automate both the mosaic acquisition and processing steps. In this report, we describe the methods and challenges involved in the routine creation of very large scale mosaics from brain tissue labelled with multiple fluorescent probes.

Differential subcellular localization of the two alternatively spliced isoforms of the Kv3.1 potassium channel subunit in brain.

Voltage-gated K(+) channels containing pore-forming subunits of the Kv3 subfamily have specific roles in the fast repolarization of action potentials and enable neurons to fire repetitively at high frequencies. Each of the four known Kv3 genes encode multiple products by alternative splicing of 3' ends resulting in the expression of K(+) channel subunits differing only in their C-terminal sequence. The alternative splicing does not affect the electrophysiological properties of the channels, and its physiological role is unknown. It has been proposed that one of the functions of the alternative splicing of Kv3 genes is to produce subunit isoforms with differential subcellular membrane localizations in neurons and differential modulation by signaling pathways. We investigated the role of the alternative splicing of Kv3 subunits in subcellular localization by examining the brain distribution of the two alternatively spliced versions of the Kv3.1 gene (Kv3.1a and Kv3.1b) with antibodies specific for the alternative spliced C-termini. Kv3.1b proteins were prominently expressed in the somatic and proximal dendritic membrane of specific neuronal populations in the mouse brain. The axons of most of these neurons also expressed Kv3.1b protein. In contrast, Kv3.1a proteins were prominently expressed in the axons of some of the same neuronal populations, but there was little to no Kv3.1a protein expression in somatodendritic membrane. Exceptions to this pattern were seen in two neuronal populations with unusual targeting of axonal proteins, mitral cells of the olfactory bulb, and mesencephalic trigeminal neurons, which expressed Kv3.1a protein in dendritic and somatic membrane, respectively. The results support the hypothesis that the alternative spliced C-termini of Kv3 subunits regulate their subcellular targeting in neurons.

Counting dendritic spines in brain tissue slices by image correlation spectroscopy analysis.

Growth of new micrometre sized projections called dendritic spines in neurones has been linked to the encoding of long-term memories in vertebrates. Numerous studies have been carried out at both the light and electron microscopy level to quantify dendritic spine densities in brain tissue in laboratory animals. Currently, such efforts using light microscopy have relied on manual counting of spines in confocal or two-photon optical slice images of tissue containing fluorescently labelled spines. This manual approach can be slow and tedious, especially for samples with high spine densities. We introduce an alternative way of performing spine counting that uses an applied image intensity threshold followed by spatial image correlation spectroscopy (ICS) analysis. We investigated the effect of particle sizes above the diffraction limit on the autocorrelation analysis as well as the influence of background fluorescence. Our results show that, for well labelled cerebellar tissue samples imaged with a signal-to-noise ratio of 5 or greater, ICS-based spine counts can be conducted with the same 15-20% precision as manual counting, but much more rapidly.

Filamentous actin is concentrated in specific subpopulations of neuronal and glial structures in rat central nervous system.

This paper is the second in a series of studies on the light and electron microscopic distribution of filamentous actin (F-actin) in the rat central nervous system (CNS) using phalloidin tagged with the fluorophore eosin followed by fluorescence photooxidation. A previous report described the selective localization of high concentrations of F-actin in subpopulations of dendritic spines in hippocampus, cerebellum and neostriatum. Dendritic spines were the most intensely stained structures in the CNS, but several other structures were notable for their consistent staining for F-actin. Although the majority of cell bodies, axons and large dendrites were unlabeled, mossy fibers and Schaffer collaterals in the hippocampal formation, basket cell axons in the cerebellar pinceau, and granule cell dendrites in the glomeruli of the cerebellar cortex routinely showed strong F-actin labeling. Staining was observed in all three glial cell types. Labeling was consistently observed in the astrocytic processes surrounding the Purkinje cell soma and primary dendrite. Intense but sporadic staining was observed in the perinodal glia of the Node of Ranvier. A few examples of labeled oligodendrocyte processes were also seen in the neostriatum. Labeling was observed in microglia in every brain region examined, although the labeling was present in the lumen of the endoplasmic reticulum and the nuclear membrane, leading to questions about its specificity. Perycites apposed to the blood vessels also showed very consistent labeling. Our results suggest that selected structures in the adult CNS in addition to dendritic spines are enriched in F-actin.

Ablation of Cypher, a PDZ-LIM domain Z-line protein, causes a severe form of congenital myopathy.

Cypher is a member of a recently emerging family of proteins containing a PDZ domain at their NH(2) terminus and one or three LIM domains at their COOH terminus. Cypher knockout mice display a severe form of congenital myopathy and die postnatally from functional failure in multiple striated muscles. Examination of striated muscle from the mutants revealed that Cypher is not required for sarcomerogenesis or Z-line assembly, but rather is required for maintenance of the Z-line during muscle function. In vitro studies demonstrated that individual domains within Cypher localize independently to the Z-line via interactions with alpha-actinin or other Z-line components. These results suggest that Cypher functions as a linker-strut to maintain cytoskeletal structure during contraction.

Phalloidin-eosin followed by photo-oxidation: a novel method for localizing F-actin at the light and electron microscopic levels.

We describe a novel high-resolution method to detect F-actin at the light and electron microscopic levels through the use of the actin-binding protein phalloidin conjugated to the fluorophore eosin, followed by photo-oxidation of diaminobenzidine. This method possesses several key advantages over antibody-based labeling and structural methods. First, phalloidin binding to F-actin can tolerate relatively high concentrations of glutaraldehyde (up to 1%) in the primary fixative, resulting in good ultrastructural preservation. Second, because both eosin and phalloidin are relatively small molecules, considerable penetration of reagents into aldehyde-fixed tissue was obtained without any permeabilization steps, allowing 3D reconstructions at the electron microscopic level. By employing a secondary fixation with tannic acid combined with low pH osmication, conditions known to stabilize actin filaments during preparation for electron microscopy, we were able to visualize individual actin filaments in some structures. Finally, we show that fluorescent phalloidin can be directly injected into neurons to label actin-rich structures such as dendritic spines. These results suggest that the fluorescent phalloidin is an excellent tool for the study of actin networks at high resolution.

Selective localization of high concentrations of F-actin in subpopulations of dendritic spines in rat central nervous system: a three-dimensional electron microscopic study.

Dendritic spines differ considerably in their size, shape, and internal organization between brain regions. We examined the actin cytoskeleton in dendritic spines in hippocampus (areas CA1, CA3, and dentate gyrus), neostriatum, and cerebellum at both light and electron microscopic levels by using a novel high-resolution photoconversion method based in the high affinity of phalloidin for filamentous (F)-actin. In all brain regions, labeling was strongest in the heads of dendritic spines, diminishing in the spine neck. The number of labeled spines varied by region. Compared with the cerebellar molecular layer and area CA3, where nearly every dendritic spine was labeled, less than half the spines were labeled in CA1, dentate gyrus, and neostriatum. Serial section reconstructions of spines in these areas indicated that phalloidin labeling was restricted to the largest and most morphologically diverse dendritic spines. The resolution of the photoconversion technique allowed us to examine the localization and organization of actin filaments in the spine. The most intense staining for actin was found in the postsynaptic density and associated with the spines internal membrane system. In mushroom-shaped spines, F-actin staining was particularly strong between the lamellae of the spine apparatus. Three-dimensional reconstruction of labeled spines by using electron tomography showed that the labeled dense material was in continuity with the postsynaptic density. These results highlight differences in the actin cytoskeleton between different spine populations and provide novel information on the organization of the actin cytoskeleton in vivo.

Specific labeling of connexin43 in NRK cells using tyramide-based signal amplification and fluorescence photooxidation.

Imaging of gap junction proteins, the connexins, has been performed in tissue culture cells both by labeling of connexins with immunocytochemical tags and by cloning and expressing chimeras of connexins and fluorescent proteins such as Green Fluorescent Protein. These two approaches have been used to gain information about protein localization or trafficking at light microscopic resolution. Electron microscopy provides higher resolution; however, analysis of electron micrographs of unlabeled connexins has been generally limited to recognition of gap junction structures. Immunolabeling of gap junction proteins in whole cells at the electron microscopic level has been difficult to achieve because of the fixation sensitivity of most gap junction antibodies. To obtain reasonable sensitivity, immunoperoxidase procedures are typically employed, and these suffer from relatively poor resolution. Here we describe the combination of tyramide signal amplification techniques and fluorescence photooxidation for higher resolution immunolocalization studies for correlative light and electron microscopic imaging. By using correlative microscopy, we can not only localize connexin pools or structures, but also discover what other cellular substructures interact with gap junction proteins. The use of tyramide signal amplification techniques is necessary to increase fluorescence levels that have decreased due to increased specimen fixation required to maintain cell ultrastructure. The fluorescence photooxidation technique provides a high-resolution method for staining of proteins in cells. Unlike colloidal gold-based methods, fluorescence photooxidation allows for three-dimensional localization using high-voltage electron microscopy.

Alterations of hippocampal postsynaptic densities following transient ischemia.

A transient interruption in cerebral blood flow can lead to delayed neuronal death in certain vulnerable cell populations several days after blood flow is restored. Among the most vulnerable cell populations in the forebrain are hippocampal CA1 pyramidal neurons, which die between 48-72 h after the ischemic insult. Neurons in the dentate gyrus and area CA3 are relatively resistant, and will recover from the same insult. Uncovering the factors that render some neuronal populations vulnerable to transient ischemia is key to understanding mechanisms leading to cell death and to developing therapeutic interventions. By applying selective staining and three-dimensional (3D) imaging with electron tomography, we uncovered dramatic structural modifications in postsynaptic densities in the postischemic brain. Postsynaptic densities in the postischemic brain appeared both thicker and less condensed than those from sham-operated controls. Although the class of synapse could not be determined with the methods used, most are likely to be glutamatergic synapses onto dendritic spines, because the majority of synapses in the region examined belong to this class. Further analysis using electron tomography to examine the 3D structure of postsynaptic densities revealed degenerative changes, as evidenced by an overall loosening of the normally compact structure. Synaptic modifications were particularly severe and persistent in hippocampal area CA1 compared to the dentate gyrus. These structural modifications correlate well with biochemical and physiological studies indicating that alterations in synaptic transmission occur in the postischemic brain. The combination of selective staining and 3D reconstruction provides a valuable tool for revealing aspects of synaptic morphology not apparent from standard electron microscopic evaluation.

Protein aggregation after transient cerebral ischemia.

Protein aggregates containing ubiquitinated proteins are commonly present in neurodegenerative disorders and have been considered to cause neuronal degeneration. Here, we report that transient cerebral ischemia caused severe protein aggregation in hippocampal CA1 neurons. By using ethanolic phosphotungstic acid electron microscopy (EM) and ubiquitin immunogold EM, we found that protein aggregates were accumulated in CA1 neurons destined to die 72 hr after 15 min of cerebral ischemia. Protein aggregates appeared as clumps of electron-dense materials that stained heavily for ubiquitin and were associated with various intracellular membranous structures. The protein aggregates appeared at 4 hr and progressively accumulated at 24 and 48 hr of reperfusion in CA1 dying neurons. However, they were rarely observed in dentate gyrus neurons that were resistant to ischemia. At 4 hr of reperfusion, protein aggregates were mainly associated with intracellular vesicles in the soma and dendrites, and the nuclear membrane. By 24 hr of reperfusion, the aggregates were also associated with mitochondria, the Golgi apparatus, and the dendritic plasmalemma. High-resolution confocal microscopy further demonstrated that protein aggregates containing ubiquitin were persistently and progressively accumulated in all CA1 dying neurons but not in neuronal populations that survive in this model. We conclude that proteins are severely aggregated in hippocampal neurons vulnerable to transient brain ischemia. We hypothesize that the accumulation of protein aggregates cause ischemic neuronal death.

Chronic phospholamban-sarcoplasmic reticulum calcium ATPase interaction is the critical calcium cycling defect in dilated cardiomyopathy.

Dilated cardiomyopathy and end-stage heart failure result in multiple defects in cardiac excitation-contraction coupling. Via complementation of a genetically based mouse model of dilated cardiomyopathy, we now provide evidence that progressive chamber dilation and heart failure are dependent on a Ca2+ cycling defect in the cardiac sarcoplasmic reticulum. The ablation of a muscle-specific sarcoplasmic reticulum Ca2+ ATPase (SERCA2a) inhibitor, phospholamban, rescued the spectrum of phenotypes that resemble human heart failure. Inhibition of phospholamban-SERCA2a interaction via in vivo expression of a phospholamban point mutant dominantly activated the contractility of ventricular muscle cells. Thus, interfering with phospholamban-SERCA2a interaction may provide a novel therapeutic approach for preventing the progression of dilated cardiomyopathy.

Survival- and death-promoting events after transient cerebral ischemia: phosphorylation of Akt, release of cytochrome C and Activation of caspase-like proteases.

Release of cytochrome c (cyt c) into cytoplasm initiates caspase-mediated apoptosis, whereas activation of Akt kinase by phosphorylation at serine-473 prevents apoptosis in several cell systems. To investigate cell death and cell survival pathways, the authors studied release of cyt c, activation of caspase, and changes in Akt phosphorylation in rat brains subjected to 15 minutes of ischemia followed by varying periods of reperfusion. The authors found by electron microscopic study that a portion of mitochondria was swollen and structurally altered, whereas the cell membrane and nuclei were intact in hippocampal CA1 neurons after 36 hours of reperfusion. In some neurons, the pattern of immunostaining for cyt c changed from a punctuate pattern, likely representing mitochondria, to a more diffuse cytoplasmic localization at 36 and 48 hours of reperfusion as examined by laser-scanning confocal microscopic study. Western blot analysis showed that cyt c was increased in the cytosolic fraction in the hippocampus after 36 and 48 hours of reperfusion. Consistently, caspase-3-like activity was increased in these hippocampal samples. As demonstrated by Western blot using phosphospecific Akt antibody, phosphorylation of Akt at serine-473 in the hippocampal region was highly increased during the first 24 hours but not at 48 hours of reperfusion. The authors conclude that transient cerebral ischemia activates both cell death and cell survival pathways after ischemia. The activation of Akt during the first 24 hours conceivably may be one of the factors responsible for the delay in neuronal death after global ischemia.

Cypher, a striated muscle-restricted PDZ and LIM domain-containing protein, binds to alpha-actinin-2 and protein kinase C.

We have cloned and characterized a novel striated muscle-restricted protein (Cypher) that has two mRNA splice variants, designated Cypher1 and Cypher2. Both proteins contain an amino-terminal PDZ domain. Cypher1, but not Cypher2, contains three carboxyl-terminal LIM domains and an amino acid repeat sequence that exhibits homology to a repeat sequence found in the largest subunit of RNA polymerase II. cypher1 and cypher2 mRNAs exhibited identical expression patterns. Both are exclusively expressed in cardiac and striated muscle in embryonic and adult stages. By biochemical assays, we have demonstrated that Cypher1 and Cypher2 bind to alpha-actinin-2 via their PDZ domains. This interaction has been further confirmed by immunohistochemical studies that demonstrated co-localization of Cypher and alpha-actinin at the Z-lines of cardiac muscle. We have also found that Cypher1 binds to protein kinase C through its LIM domains. Phosphorylation of Cypher by protein kinase C has demonstrated the functional significance of this interaction. Together, our data suggest that Cypher1 may function as an adaptor in striated muscle to couple protein kinase C-mediated signaling, via its LIM domains, to the cytoskeleton (alpha-actinin-2) through its PDZ domain.

Enteroviral protease 2A cleaves dystrophin: evidence of cytoskeletal disruption in an acquired cardiomyopathy.

Enteroviruses such as Coxsackievirus B3 can cause dilated cardiomyopathy, but the mechanism of this pathology is unknown. Mutations in cytoskeletal proteins such as dystrophin cause hereditary dilated cardiomyopathy, but it is unclear if similar mechanisms underlie acquired forms of heart failure. We demonstrate here that purified Coxsackievirus protease 2A cleaves dystrophin in vitro as predicted by computer analysis. Dystrophin is also cleaved during Coxsackievirus infection of cultured myocytes and in infected mouse hearts, leading to impaired dystrophin function. In vivo, dystrophin and the dystrophin-associated glycoproteins alpha-sarcoglycan and beta-dystroglycan are morphologically disrupted in infected myocytes. We suggest a molecular mechanism through which enteroviral infection contributes to the pathogenesis of acquired forms of dilated cardiomyopathy.

Modification of postsynaptic densities after transient cerebral ischemia: a quantitative and three-dimensional ultrastructural study.

Abnormal synaptic transmission has been hypothesized to be a cause of neuronal death resulting from transient ischemia, although the mechanisms are not fully understood. Here, we present evidence that synapses are markedly modified in the hippocampus after transient cerebral ischemia. Using both conventional and high-voltage electron microscopy, we performed two- and three-dimensional analyses of synapses selectively stained with ethanolic phosphotungstic acid in the hippocampus of rats subjected to 15 min of ischemia followed by various periods of reperfusion. Postsynaptic densities (PSDs) from both area CA1 and the dentate gyrus were thicker and fluffier in postischemic hippocampus than in controls. Three-dimensional reconstructions of selectively stained PSDs created using electron tomography indicated that postsynaptic densities became more irregular and loosely configured in postischemic brains compared with those in controls. A quantitative study based on thin sections of the time course of PSD modification indicated that the increase in thickness was both greater and more long-lived in area CA1 than in dentate gyrus. Whereas the magnitude of morphological change in dentate gyrus peaked at 4 hr of reperfusion (140% of control values) and declined thereafter, changes in area CA1 persisted and increased at 24 hr of reperfusion (191% of control values). We hypothesize that the degenerative ultrastructural alteration of PSDs may produce a toxic signal such as a greater calcium influx, which is integrated from the thousands of excitatory synapses onto dendrites, and is propagated to the neuronal somata where it causes or contributes to neuronal damage during the postischemic phase.