Accuracy and Reliability of Chatbot Responses to Physician Questions

Importance

Natural language processing tools, such as ChatGPT (generative pretrained transformer, hereafter referred to as chatbot), have the potential to radically enhance the accessibility of medical information for health professionals and patients. Assessing the safety and efficacy of these tools in answering physician-generated questions is critical to determining their suitability in clinical settings, facilitating complex decision-making, and optimizing health care efficiency.

Objective

To assess the accuracy and comprehensiveness of chatbot-generated responses to physician-developed medical queries, highlighting the reliability and limitations of artificial intelligence-generated medical information.

Design, Setting, and Participants

Thirty-three physicians across 17 specialties generated 284 medical questions that they subjectively classified as easy, medium, or hard with either binary (yes or no) or descriptive answers. The physicians then graded the chatbot-generated answers to these questions for accuracy (6-point Likert scale with 1 being completely incorrect and 6 being completely correct) and completeness (3-point Likert scale, with 1 being incomplete and 3 being complete plus additional context). Scores were summarized with descriptive statistics and compared using the Mann-Whitney U test or the Kruskal-Wallis test. The study (including data analysis) was conducted from January to May 2023.

Main Outcomes and Measures

Accuracy, completeness, and consistency over time and between 2 different versions (GPT-3.5 and GPT-4) of chatbot-generated medical responses.

Results

Across all questions (n = 284) generated by 33 physicians (31 faculty members and 2 recent graduates from residency or fellowship programs) across 17 specialties, the median accuracy score was 5.5 (IQR, 4.0-6.0) (between almost completely and complete correct) with a mean (SD) score of 4.8 (1.6) (between mostly and almost completely correct). The median completeness score was 3.0 (IQR, 2.0-3.0) (complete and comprehensive) with a mean (SD) score of 2.5 (0.7). For questions rated easy, medium, and hard, the median accuracy scores were 6.0 (IQR, 5.0-6.0), 5.5 (IQR, 5.0-6.0), and 5.0 (IQR, 4.0-6.0), respectively (mean [SD] scores were 5.0 [1.5], 4.7 [1.7], and 4.6 [1.6], respectively; P = .05). Accuracy scores for binary and descriptive questions were similar (median score, 6.0 [IQR, 4.0-6.0] vs 5.0 [IQR, 3.4-6.0]; mean [SD] score, 4.9 [1.6] vs 4.7 [1.6]; P = .07). Of 36 questions with scores of 1.0 to 2.0, 34 were requeried or regraded 8 to 17 days later with substantial improvement (median score 2.0 [IQR, 1.0-3.0] vs 4.0 [IQR, 2.0-5.3]; P < .01). A subset of questions, regardless of initial scores (version 3.5), were regenerated and rescored using version 4 with improvement (mean accuracy [SD] score, 5.2 [1.5] vs 5.7 [0.8]; median score, 6.0 [IQR, 5.0-6.0] for original and 6.0 [IQR, 6.0-6.0] for rescored; P = .002).

Conclusions and Relevance

In this cross-sectional study, chatbot generated largely accurate information to diverse medical queries as judged by academic physician specialists with improvement over time, although it had important limitations. Further research and model development are needed to correct inaccuracies and for validation.

Predicted expression of genes involved in the thiopurine metabolic pathway and azathioprine discontinuation due to myelotoxicity

TPMT and NUDT15 variants explain less than 25% of azathioprine-associated myelotoxicity. There are 25 additional genes in the thiopurine pathway that could also contribute to azathioprine myelotoxicity. We hypothesized that among TPMT and NUDT15 normal metabolizers, a score combining the genetically predicted expression of other proteins in the thiopurine pathway would be associated with a higher risk for azathioprine discontinuation due to myelotoxicity. We conducted a retrospective cohort study of new users of azathioprine who were normal TPMT and NUDT15 metabolizers. In 1201 White patients receiving azathioprine for an inflammatory disease, we used relaxed Least Absolute Shrinkage and Selection Operator (LASSO) regression to select genes that built a score for discontinuing azathioprine due to myelotoxicity. The score incorporated the predicted expression of AOX1 and NME1. Patients in the highest score tertile had a higher risk of discontinuing azathioprine compared to those in the lowest tertile (hazard ratio [HR] = 2.15, 95% confidence interval [CI] = 1.11-4.19, p = 0.024). Results remained significant after adjusting for a propensity score, including sex, tertile of calendar year at initial dose, initial dose, age at baseline, indication, prior TPMT testing, and the first 10 principal components of the genetic data (HR = 2.11, 95% CI = 1.08-4.13, p = 0.030). We validated the results in a cohort (N = 517 non-White patients and those receiving azathioprine to prevent transplant rejection) that included all other patients receiving azathioprine (HR = 2.00, (95% CI = 1.09-3.65, p = 0.024). In conclusion, among patients who were TPMT and NUDT15 normal metabolizers, a score combining the predicted expression of AOX1 and NME1 was associated with an increased risk for discontinuing azathioprine due to myelotoxicity.

Syntaxin 4 is concentrated on plasma membrane of astrocytes

Syntaxins are a family of transmembrane proteins that participate in SNARE complexes to mediate membrane fusion events including exocytosis. Different syntaxins are thought to participate in exocytosis in different compartments of the nervous system such as the axon, the soma/dendrites or astrocytes. It is well known that exocytosis of synaptic vesicles at axonal presynaptic terminals involves syntaxin 1 but distributions of syntaxins on neuronal somal and dendritic, postsynaptic or astroglial plasma membranes are less well characterized. Here, we use pre-embedding immunogold labeling to compare the distribution of two plasma membrane-enriched syntaxins (1 and 4) in dissociated rat hippocampal cultures as well as in perfusion-fixed mouse brains. Comparison of Western blots of neuronal cultures, consisting of a mixture of hippocampal neurons and glia, with glial cultures, consisting of mostly astrocytes, shows that syntaxin 1 is enriched in neuronal cultures, whereas syntaxin 4 is enriched in glial cultures. Electron microscopy (EM)-immunogold labeling shows that syntaxin 1 is most abundant at the plasma membranes of axons and terminals, while syntaxin 4 is most abundant at astroglial plasma membranes. This differential distribution was evident even at close appositions of membranes at synapses, where syntaxin 1 was localized to the plasma membrane of the presynaptic terminal, including that at the active zone, while syntaxin 4 was localized to nearby peri-synaptic astroglial processes. These results show that syntaxin 4 is available to support exocytosis in astroglia.

Homer is concentrated at the postsynaptic density and does not redistribute after acute synaptic stimulation

Homer is a postsynaptic density (PSD) scaffold protein that is involved in synaptic plasticity, calcium signaling and neurological disorders. Here, we use pre-embedding immunogold electron microscopy to illustrate the differential localization of three Homer gene products (Homer 1, 2, and 3) in different regions of the mouse brain. In cross-sectioned PSDs, Homer occupies a layer ∼30-100nm from the postsynaptic membrane lying just beyond the dense material that defines the PSD core (∼30-nm-thick). Homer is evenly distributed within the PSD area along the lateral axis, but not at the peri-PSD locations within 60nm from the edge of the PSD, where type I-metabotropic glutamate receptors (mGluR1 and 5) are concentrated. This distribution of Homer matches that of Shank, another major PSD scaffold protein, but differs from those of other two major binding partners of Homer, type I mGluR and IP3 receptors. Many PSD proteins rapidly redistribute upon acute (2min) stimulation. To determine whether Homer distribution is affected by acute stimulation, we examined its distribution in dissociated hippocampal cultures under different conditions. Both the pattern and density of label for Homer 1, the isoform that is ubiquitous in hippocampus, remained unchanged under high K(+) depolarization (90mM for 2-5min), N-methyl-d-asparic acid (NMDA) treatment (50μM for 2min), and calcium-free conditions (EGTA at 1mM for 2min). In contrast, Shank and calcium/calmodulin-dependent kinase II (CaMKII) accumulate at the PSD upon NMDA treatment, and CaMKII is excluded from the PSD complex under low calcium conditions.

Effects of CaMKII inhibitor tatCN21 on activity-dependent redistribution of CaMKII in hippocampal neurons

TatCN21 is a membrane permeable calcium/calmodulin-dependent protein kinase II (CaMKII) inhibitor derived from the inhibitor protein CaMKIIN. TatCN21 has been used to demonstrate the involvement of CaMKII in a variety of physiological and pathological phenomena, and it also limits excitotoxic damage in neurons. Here we use preembedding immunogold electron microscopy to examine the effect of tatCN21 on the redistribution of CaMKII in cultured hippocampal neurons. Incubation of cultures with tatCN21 (20 μM for 20 min) prior to exposure to N-methyl-d-asparic acid (NMDA) (50 μM for 2 min) inhibited both the accumulation of CaMKII at postsynaptic densities (PSDs) and CaMKII clustering in the dendrites. Under these conditions, CaMKII also formed morphologically distinct aggregates with polyribosomes near the PSD and in dendrites. Formation of these CaMKII-polyribosome aggregates requires the presence of both tatCN21 and calcium, and was augmented upon exposure to high K(+) or NMDA. CaMKII-polyribosome aggregates formed consistently with 20 μM tatCN21, but minimally or not at all with 5 μM. However, these aggregates are not induced by another CaMKII inhibitor, KN93. Formation of CaMKII-polyribosome aggregates was completely reversible within 1h after washout of tatCN21. Effects of tatCN21 were largely restricted to dendrites, with minimal effect in the soma. The effects of tatCN21 on CaMKII distribution can be used to dissect the mechanism of CaMKII involvement in cellular events.

OLFACTORY CILIA IN THE FROG

Olfactory epithelium from the frog was examined in the living state by light microscopy and in the fixed state by electron microscopy. Particular attention was paid to the layer of cilia and mucus which covers the surface of the epithelium. The olfactory cilia differed from typical cilia in that they (a) arose from bipolar neurons and had centrioles near their basal bodies, (b) were up to 200 microns in length, of which the greater part was a distal segment containing an atypical array of ciliary fibers, (c) were often immotile, (d) had their distal segments arranged in parallel rows near the surface of the mucus, and (e) had many vesicles along their shafts and had splits in the array of fibers in their distal segments. These specializations make the olfactory cilia similar to cilia found on other sensory cells and support the theory that they are the locus where electrical excitation in the olfactory organ is initiated by contact with odorous substances.

Rapid turnover of spinules at synaptic terminals

Spinules found in brain consist of small invaginations of plasma membranes which enclose membrane evaginations from adjacent cells. Here, we focus on the dynamic properties of the most common type, synaptic spinules, which reside in synaptic terminals. In order to test whether depolarization triggers synaptic spinule formation, hippocampal slice cultures (7-day-old rats, 10-14 days in culture) were exposed to high K+ for 0.5-5 min, and examined by electron microscopy. Virtually no synaptic spinules were found in control slices representing a basal state, but numerous spinules appeared at both excitatory and inhibitory synapses after treatment with high K+. Spinule formation peaked with approximately 1 min treatment at 37 degrees C, decreased with prolonged treatment, and disappeared after 1-2 min of washout in normal medium. The rate of disappearance of spinules was substantially slower at 4 degrees C. N-methyl-D-aspartic acid (NMDA) treatment also induced synaptic spinule formation, but to a lesser extent than high K+ depolarization. In acute brain slices prepared from adult mice, synaptic spinules were abundant immediately after dissection at 4 degrees C, extremely rare in slices allowed to recover at 28 degrees C, but frequent after high K(+) depolarization. High pressure freezing of acute brain slices followed by freeze-substitution demonstrated that synaptic spinules are not induced by chemical fixation. These results indicate that spinules are absent in synapses at low levels of activity, but form and disappear quickly during sustained synaptic activity. The rapid turnover of synaptic spinules may represent an aspect of membrane retrieval during synaptic activity.

Inhibition of phosphatase activity facilitates the formation and maintenance of NMDA-induced calcium/calmodulin-dependent protein kinase II clusters in hippocampal neurons

The majority of hippocampal neurons in dissociated cultures and in intact brain exhibit clustering of calcium/calmodulin-dependent protein kinase II (CaMKII) into spherical structures with an average diameter of 110 nm when subjected to conditions that mimic ischemia and excitotoxicity [Neuroscience 106 (2001) 69]. Because clustering of CaMKII would reduce its effective concentration within the neuron, it may represent a cellular strategy to prevent excessive CaMKII-mediated phosphorylation during episodes of Ca2+ overload. Here we employ a relatively mild excitatory stimulus to promote sub-maximal clustering for the purpose of studying the conditions for the formation and disappearance of CaMKII clusters. Treatment with 30 microM N-methyl-D-aspartic acid (NMDA) for 2 min produced CaMKII clustering in approximately 15% of dissociated hippocampal neurons in culture, as observed by pre-embedding immunogold electron microscopy. These CaMKII clusters could be labeled with antibodies specific to the phospho form (Thr286) of CaMKII, suggesting that at least some of the CaMKII molecules in clusters are autophosphorylated. To test whether phosphorylation is involved in the formation and maintenance of CaMKII clusters, the phosphatase inhibitors calyculin A (5 nM) or okadaic acid (1 microM) were included in the incubation medium. With inhibitors more neurons exhibited CaMKII clusters in response to 2 min NMDA treatment. Furthermore, 5 min after the removal of NMDA and Ca2+, CaMKII clusters remained and could still be labeled with the phospho-specific antibody. In contrast, in the absence of phosphatase inhibitors, no clusters were detected 5 min after the removal of NMDA and Ca2+ from the medium. These results suggest that phosphatases type 1 and/or 2A regulate the formation and disappearance of CaMKII clusters.

Inhibition of phosphatase activity prolongs NMDA-induced modification of the postsynaptic density

NMDA-induced modification of postsynaptic densities (PSDs) was studied by immunoelectron microscopy. Treatment of cultured hippocampal neurons with NMDA for 2 min promotes a 2.3 fold thickening of the PSD and a 4 fold increase in PSD-associated CaMKII immunolabel. These changes are reversed 5 min after the removal of NMDA and Ca2+ from the medium. In addition, following NMDA treatment, PSDs exhibit a 7.5 fold increase in labeling with an antibody specific to the (Thr286) phospho-form of CaMKII, indicating that CaMKII translocated to the PSD is phosphorylated. When the phosphatase inhibitors, calyculin A or okadaic acid, are included in the medium, the NMDA-induced thickening of the PSD as well as the increase in PSD-associated CaMKII immunolabeling are largely maintained (75% and 88% of the peak values respectively) at 5 min after removal of NMDA and Ca2+ from the medium. These results imply that NMDA receptors can mediate activity-induced changes in the PSD and that phosphatases of type 1 and/or 2A are involved in the reversal of these changes.

Calcium/calmodulin-dependent protein kinase II clusters in adult rat hippocampal slices

We have previously reported the formation of calcium/calmodulin-dependent protein kinase II (CaMKII) clusters approximately 110 nm in diameter in hippocampal neurons in culture and in the intact adult brain, under conditions that simulate ischemic stress and increase [Ca(2+)](i) [Dosemeci et al. (2000) J. Neurosci. 20, 3076-3084; Tao-Cheng et al. (2001) Neuroscience 106, 69-78]. These observations suggest that ischemia-like conditions that prevail during the dissection of brain tissue for the preparation of hippocampal slices could lead to the formation of CaMKII clusters. We now show by pre-embedding immuno-electron microscopy that, indeed, CaMKII clusters are present in the CA1 pyramidal neurons in hippocampal slices from adult rats fixed immediately after dissection, and that the number of CaMKII clusters increases with the delay time between decapitation and fixation. Moreover, CaMKII clusters are typically localized near the endoplasmic reticulum. When acute slices are allowed to recover in oxygenated medium for 2 h, CaMKII clusters mostly disappear, indicating that clustering is reversible. Also, the postsynaptic density, another site for CaMKII accumulation under excitatory conditions, becomes thinner upon recovery. Treatment of recovered slices with high potassium for 90 s causes the re-appearance of CaMKII clusters in nearly all CA1 pyramidal cells examined. On the other hand, when dissociated hippocampal neurons in primary culture are exposed to the same depolarizing conditions, only approximately 25% of neurons exhibit CaMKII clusters, indicating a difference in the susceptibility of the neurons in culture and in acute slices to excitatory stimuli. Altogether these observations indicate that the effect of CaMKII clustering should be considered when interpreting experimental results obtained with hippocampal slices.

Sustained elevation of calcium induces Ca(2+)/calmodulin-dependent protein kinase II clusters in hippocampal neurons

Treatment of cultured hippocampal neurons with the mitochondrial uncoupler carbonyl cyanide m-chlorophenylhydrazone (CCCP) in the absence of glucose mimics ischemic energy depletion and induces formation of Ca(2+)/calmodulin-dependent protein kinase II (CaMKII) clusters, spherical structures with diameters of 75-175 nm [Dosemeci et al., J. Neurosci. 20 (2000) 3076-3084]. The demonstration that CaMKII clustering occurs in the intact, adult rat brain upon interruption of blood flow indicates that clustering is not confined to cell cultures. Application of N-methyl-D-aspartate (250 microM, 15 min) to hippocampal cultures also induces cluster formation, suggesting a role for Ca(2+). Indeed, intracellular Ca(2+) monitored with Fluo3-AM by confocal microscopy reaches a sustained high level within 5 min of CCCP treatment. The appearance of immunolabeled CaMKII clusters, detected by electron microscopy, follows the onset of the sustained increase in intracellular Ca(2+). Moreover, CaMKII does not cluster when the rise in intracellular Ca(2+) is prevented by the omission of extracellular Ca(2+) during CCCP treatment, confirming that clustering is Ca(2+)-dependent. A lag period of 1-2 min between the onset of high intracellular Ca(2+) levels and the formation of CaMKII clusters suggests that a sustained increase in Ca(2+) level is necessary for the clustering. CaMKII clusters disappear within 2 h of returning the cultures to normal incubation conditions, at which time no significant cell death is detected. These results indicate that pathological conditions that promote sustained episodes of Ca(2+) overload result in a transitory clustering of CaMKII into spherical structures. CaMKII clustering may represent a cellular defense mechanism to sequester a portion of the CaMKII pool, thereby preventing excessive protein phosphorylation.

Glutamate-induced transient modification of the postsynaptic density

Depolarization of rat hippocampal neurons with a high concentration of external potassium induces a thickening of postsynaptic densities (PSDs) within 1.5-3 min. After high-potassium treatment, PSDs thicken 2.1-fold in cultured neurons and 1.4-fold in hippocampal slices compared with their respective controls. Thin-section immunoelectron microscopy of hippocampal cultures indicates that at least part of the observed thickening of PSDs can be accounted for by an accumulation of Ca(2+)/calmodulin-dependent protein kinase II (CaMKII) on their cytoplasmic faces. Indeed, PSD-associated gold label for CaMKII increases 5-fold after depolarization with potassium. The effects of high-potassium treatment on the composition and structure of the PSDs are mimicked by direct application of glutamate. In cultures, glutamate-induced thickening of PSDs and the accumulation of CaMKII on PSDs are reversed within 5 min of removal of glutamate and Ca(2+) from the extracellular medium. These results suggest that PSDs are dynamic structures whose thickness and composition are subject to rapid and transient changes during synaptic activity.

Activation of calpain may alter the postsynaptic density structure and modulate anchoring of NMDA receptors

Elevation of calcium during sustained synaptic activity may lead to the activation of the postsynaptic calcium-dependent protease calpain and thus could alter the integrity and localization of endogenous proteins. The distribution of anchoring proteins for neuroreceptors is an important determinant of the efficacy of neuronal transmission. Many of these anchoring proteins are concentrated within the postsynaptic density (PSD). In the present study, we examined the effects of calpain II on isolated PSDs using biochemical and electron microscopic techniques. Biochemical analysis reveals that PSD-95, a clustering molecule which anchors NMDA receptors by interaction with their NR2 subunits, as well as the NR2 subunits themselves, are cleaved by calpain. On the other hand, under conditions where all the PSD-95 protein is cleaved, actin and alpha-actinin-a protein thought to anchor NMDA receptors to actin filaments-remain intact. For analysis by electron microscopy, PSDs were adsorbed on glass, immunogold-labeled with an antibody to PSD-95, slam frozen, freeze dried, and rotary shadowed. Electron micrographs of replicas indicate that PSDs are disc-shaped and are composed of a lattice-like structure which labels with PSD-95 immunogold. After calpain treatment, PSDs adsorbed on glass become thinner overall and the lattice becomes fragmented. Altogether, these results suggest that calpain activity could produce changes in the organization of the PSD and, by cleaving PSD-95 associated with the PSD lattice, could modify the anchoring of NMDA receptors.

Multiple local contact sites are induced by GPI-linked influenza hemagglutinin during hemifusion and flickering pore formation

Membrane fusion intermediates induced by the glycosylphosphatidylinositol-linked ectodomain of influenza hemagglutinin (GPI-HA) were investigated by rapid freeze, freeze-substitution, thin section electron microscopy, and with simultaneous recordings of whole-cell admittance and fluorescence. Upon triggering, the previously separated membranes developed numerous hourglass shaped points of membrane contact (approximately 10-130 nm waist) when viewed by electron microscopy. Stereo pairs showed close membrane contact at peaks of complementary protrusions, arising from each membrane. With HA, there were fewer contacts, but wide fusion pores. Physiological measurements showed fast lipid dye mixing between cells after acidification, and either fusion pore formation or the lack thereof (true hemifusion). For the earliest pores, a similar conductance distribution and frequency of flickering pores were detected for both HA and GPI-HA. For GPI-HA, lipid mixing was detected prior to, during, or after pore opening, whereas for HA, lipid mixing is seen only after pore opening. Our findings are consistent with a pathway wherein conformational changes in the ectodomain of HA pull membranes towards each other to form a contact site, then hemifusion and pore formation initiate in a small percentage of these contact sites. Finally, the transmembrane domain of HA is needed to complete membrane fusion for macromolecular content mixing.

Retrograde axonal transport of herpes simplex virus: evidence for a single mechanism and a role for tegument

Herpes simplex virus type I (HSV) typically enters peripheral nerve terminals and then travels back along the nerve to reach the neuronal cell body, where it replicates or enters latency. To monitor axoplasmic transport of HSV, we used the giant axon of the squid, Loligo pealei, a well known system for the study of axoplasmic transport. To deliver HSV into the axoplasm, viral particles stripped of their envelopes by detergent were injected into the giant axon, thereby bypassing the infective process. Labeling the viral tegument protein, VP16, with green fluorescent protein allowed viral particles moving inside the axon to be imaged by confocal microscopy. Viral particles moved 2.2 +/- 0.26 micrometer/sec in the retrograde direction, a rate comparable to that of the transport of endogenous organelles and of virus in mammalian neurons in culture. Electron microscopy confirmed that 96% of motile (stripped) viral particles had lost their envelope but retained tegument, and Western blot analysis revealed that these particles had retained protein from capsid but not envelope. We conclude that (i) HSV recruits the squid retrograde transport machinery; (ii) viral tegument and capsid but not envelope are sufficient for this recruitment; and (iii) the giant axon of the squid provides a unique system to dissect the viral components required for transport and to identify the cellular transport mechanisms they recruit.

A novel particulate form of Ca(2+)/calmodulin-dependent [correction of Ca(2+)/CaMKII-dependent] protein kinase II in neurons

Cytoskeletal and postsynaptic density (PSD) fractions from forebrain contain discrete spherical structures that are immunopositive for Ca(2+)/calmodulin-dependent protein kinase II (CaMKII). Spherical structures viewed by rotary shadow electron microscopy have an average diameter of approximately 100 nm and, in distinction to postsynaptic densities, do not immunolabel for PSD-95. These structures were purified to near homogeneity by extraction with the detergent N-lauryl sarcosinate. Biochemical analysis revealed that CaMKII accounts for virtually all of the protein in the purified preparation, suggesting that spherical structures are clusters of self-associated CaMKII. Exposure of cultured hippocampal neurons to a mitochondrial uncoupler in glucose-free medium promotes the formation of numerous CaMKII-immunopositive structures identical in size and shape to the CaMKII clusters observed in subcellular fractions. Clustering of CaMKII would reduce its kinase function by preventing its access to fixed substrates. On the other hand, clustering would not affect the ability of the large cellular pool of CaMKII to act as a calmodulin sink, as demonstrated by the Ca(2+)-dependent binding of gold-conjugated calmodulin to CaMKII clusters. We propose that the observed clustering of CaMKII into spherical structures is a protective mechanism preventing excessive protein phosphorylation upon loss of Ca(2+) homeostasis, without compromising calmodulin regulation.

Localization of the linker domain of Ca2+/calmodulin-dependent protein kinase II

Electron micrographs of rotary shadowed replicas of alpha-Ca2+/calmodulin-dependent protein kinase II reveal a flower-shaped multimeric molecule with a central particle surrounded by 8-10 smaller peripheral particles. Peripheral particles are attached to the central particle by thin arms or "linkers." Movement of peripheral particles to contact each other for autophosphorylation is likely to involve these linkers. It has generally been accepted that the segment 317-328 of the alpha-subunit constitutes the linker domain. In the present study we test this assumption by generating a mutant lacking the proposed sequence. The mutant has biochemical and morphological properties similar to those of the wild type, and a thin linker is occasionally observed in replicas from either type. The results indicate that the deleted sequence does not correspond to the linker domain. This conclusion, combined with observations from two recent studies which identify the C-terminal domain involved in oligomerization, narrows down the location of the linker domain within the sequence 330-354.

Slow transport of unpolymerized tubulin and polymerized neurofilament in the squid giant axon

A major issue in the slow transport of cytoskeletal proteins is the form in which they are transported. We have investigated the possibility that unpolymerized as well as polymerized cytoskeletal proteins can be actively transported in axons. We report the active transport of highly diffusible tubulin oligomers, as well as transport of the less diffusible neurofilament polymers. After injection into the squid giant axon, tubulin was transported in an anterograde direction at an average rate of 2.3 mm/day, whereas neurofilament was moved at 1.1 mm/day. Addition of the metabolic poisons cyanide or dinitrophenol reduced the active transport of both proteins to less than 10% of control values, whereas disruption of microtubules by treatment of the axon with cold in the presence of nocodazole reduced transport of both proteins to approximately 20% of control levels. Passive diffusion of these proteins occurred in parallel with transport. The diffusion coefficient of the moving tubulin in axoplasm was 8.6 micrometer(2)/s compared with only 0.43 micrometer(2)/s for neurofilament. These results suggest that the tubulin was transported in the unpolymerized state and that the neurofilament was transported in the polymerized state by an energy-dependent nocodazole/cold-sensitive transport mechanism.

Correlated measurements of free and total intracellular calcium concentration in central nervous system neurons

Transient changes in the intracellular concentration of free calcium ([Ca2+])i) act as a trigger or modulator for a large number of important neuronal processes. Such transients can originate from voltage- or ligand-gated fluxes of Ca2+ into the cytoplasm from the extracellular space, or by ligand- or Ca2+(-)gated release from intracellular stores. Characterizing the sources and spatio-temporal patterns of [Ca2+]i transients is critical for understanding the role of different neuronal compartments in dendritic integration and synaptic plasticity. Optical imaging of fluorescent indicators sensitive to free Ca2+ is especially suited to studying such phenomena because this approach offers simultaneous monitoring of large regions of the dendritic tree in individual living central nervous system neurons. In contrast, energy-dispersive X-ray (EDX) microanalysis provides quantitative information on the amount and location of intracellular total, i.e., free plus bound, calcium (Ca) within specific subcellular dendritic compartments as a function of the activity state of the neuron. When optical measurements of [Ca2+]i transients and parallel EDX measurements of Ca content are used in tandem, and correlated simultaneously with electrophysiological measurements of neuronal activity, the combined information provides a relatively general picture of spatio-temporal neuronal total Ca fluctuations. To illustrate the kinds of information available with this approach, we review here results from our ongoing work aimed at evaluating the role of various Ca uptake, release, sequestration, and extrusion mechanisms in the generation and termination of [Ca2+]i transients in dendrites of pyramidal neurons in hippocampal slices during and after synaptic activity. Our observations support the long-standing speculation that the dendritic endoplasmic reticulum acts not only as an intracellular Ca2+ source that can be mobilized by a signal cascade originating at activated synapses, but also as a major intracellular Ca sink involved in active clearance mechanisms after voltage- and ligand-gated Ca2+ influx.

Association of actin filaments with axonal microtubule tracts

Axoplasmic organelles move on actin as well as microtubules in vitro and axons contain a large amount of actin, but little is known about the organization and distribution of actin filaments within the axon. Here we undertake to define the relationship of the microtubule bundles typically found in axons to actin filaments by applying three microscopic techniques: laser-scanning confocal microscopy of immuno-labeled squid axoplasm; electronmicroscopy of conventionally prepared thin sections; and electronmicroscopy of touch preparations-a thin layer of axoplasm transferred to a specimen grid and negatively stained. Light microscopy shows that longitudinal actin filaments are abundant and usually coincide with longitudinal microtubule bundles. Electron microscopy shows that microfilaments are interwoven with the longitudinal bundles of microtubules. These bundles maintain their integrity when neurofilaments are extracted. Some, though not all microfilaments decorate with the S1 fragment of myosin, and some also act as nucleation sites for polymerization of exogenous actin, and hence are definitively identified as actin filaments. These actin filaments range in minimum length from 0.5 to 1.5 microm with some at least as long as 3.5 microm. We conclude that the microtubule-based tracks for fast organelle transport also include actin filaments. These actin filaments are sufficiently long and abundant to be ancillary or supportive of fast transport along microtubules within bundles, or to extend transport outside of the bundle. These actin filaments could also be essential for maintaining the structural integrity of the microtubule bundles.

Architectural features of the Salmonella typhimurium flagellar motor switch revealed by disrupted C-rings

The three-dimensional surface topology of rapid-frozen Salmonella typhimurium flagellar hook basal body complexes was studied by stereo-examination of thin-film metal replicas. The complexes contained the extended cytoplasmic structure, composed of the switch complex proteins; FliG, FliM, and FliN. Distinct nanometer-scale element arrays, separated by grooves, defined the outer surface of the cytoplasmic (C-) ring. The number of array elements was comparable to previously determined FliG and FliM copy numbers in the basal body. In addition to basal body complexes lacking C-rings, complexes containing incomplete C-rings were identified. The incomplete C-rings had lost segments of the proximal array. Basal bodies with the distal C-ring array alone were not found. These findings are compatible with the spatial organization of the flagellar switch suggested by previous biochemical data.

Immunoelectronmicroscopy of soluble and membrane proteins with a sensitive postembedding method

The application of immunoelectronmicroscopy to soluble proteins is limited because soluble proteins can redistribute during fixation. Fixation may also adversely affect the recognition of proteins associated with membranes. We show here how displacements of soluble proteins can be prevented and antigen sensitivity improved by freeze-substitution immunocytochemistry. The usefulness of this method for soluble cytoplasmic proteins is demonstrated for the twitchin protein in Aplysia muscle and the kinesin motor proteins in squid giant axons, in which the sizes of various cytoplasmic pools of kinesins are estimated. The utility for membrane proteins present in small numbers of copies is demonstrated by labeling a glutamate receptor subunit in mouse cerebellar cortex and the ZO-1 protein in tight junctions between MDCK cells. Thus, freeze-substitution immunocytochemistry can show the native distribution of both soluble and membrane proteins labeled with polyclonal antibodies and, at the same time, can reveal structural features comparable to those in chemically fixed or osmium freeze-substituted samples.

Activity-dependent calcium sequestration in dendrites of hippocampal neurons in brain slices

Synaptic activity-dependent changes in the spatio-temporal distribution of calcium ions regulate important neuronal functions such as dendritic integration and synaptic plasticity, but the processes that terminate the free Ca2+ transients associated with these changes remain unclear. We have characterized at the electron microscopic level the intracellular compartments involved in buffering free Ca2+ transients in dendritic cytoplasm of CA3 neurons by measuring the larger changes in the concentrations of total Ca that persist for several minutes after neuronal activity. Quantitative energy-dispersive x-ray microanalysis of cryosections from hippocampal slice cultures rapidly frozen 3 min after afferent synaptic activity identified a subset of dendritic endoplasmic reticulum (ER) as a high-capacity Ca2+ buffer. Calcium sequestration by cisterns of this subset of ER was graded, reversible, and dependent on a thapsigargin-sensitive Ca2+-ATPase. Sequestration was so robust that after repetitive high-frequency stimulation the Ca content of responsive ER cisterns increased as much as 20-fold. These results demonstrate that a subpopulation of ER is the major dendritic Ca sequestration compartment in the minutes after neuronal activity.

Organization of the cortical endoplasmic reticulum in the squid giant axon

The organization of the cortical endoplasmic reticulum in the squid giant axon was investigated by rapid freeze and freeze-substitution electron microscopy, thereby eliminating the effects of fixatives on this potentially labile structure. Juvenile squid, which have thinner Schwann sheaths, were used in order to achieve freezing deep enough to include the entire axonal cortex. The smooth endoplasmic reticulum is composed of subaxolemmal and deeper cisternae, tubules, tethers and vesicles. The subaxolemmal cisternae make junctional contacts with the axolemma which are characterized by filamentous-granular bridging structures approximately 3 nm in diameter. The subaxolemmal junctions with the axolemma resemble the coupling junctions between the sarcoplasmic reticulum and the T-tubules in muscle. Reconstruction of short series of sections showed that a number of the elements of the endoplasmic reticulum were continuous but numerous separate vesicles were present as well. The morphology of endoplasmic reticulum as described here suggests that it is a highly dynamic entity as well as a Ca2+ sequestering organelle.

Phosphorylation and subunit organization of axonal neurofilaments determined by scanning transmission electron microscopy

Phosphorylation plays a critical role in controlling the function of cytoskeletal assemblies but no direct method yet exists to measure the phosphorylation state of proteins at the level of individual molecules and assemblies. Herein, we apply scanning transmission electron microscopy in combination with electron energy loss spectroscopy to measure the distributions of mass and phosphorus in neurofilaments (NFs) isolated from the squid giant axon. We find that native squid NFs, in contrast to typical reconstituted intermediate filaments, are a relatively homogeneous population containing only eight coiled-coil dimers per cross section. The measured stoichiometry of approximately 1:1 for light/heavy peptides strongly suggests that squid NFs are composed of heterodimers. Furthermore, each heavy chain of the dimers carries at least 100 phosphate groups and is, therefore, near-maximally phosphorylated. These results also demonstrate that scanning transmission electron microscopy combined with electron energy loss spectroscopy at the nanometer scale is capable of characterizing the level and distribution of phosphorylation in individual mass-mapped protein assemblies.

Plus-end motors override minus-end motors during transport of squid axon vesicles on microtubules

Plus- and minus-end vesicle populations from squid axoplasm were isolated from each other by selective extraction of the minus-end vesicle motor followed by 5'-adenylyl imidodiphosphate (AMP-PNP)-induced microtubule affinity purification of the plus-end vesicles. In the presence of cytosol containing both plus- and minus-end motors, the isolated populations moved strictly in opposite directions along microtubules in vitro. Remarkably, when treated with trypsin before incubation with cytosol, purified plus-end vesicles moved exclusively to microtubule minus ends instead of moving in the normal plus-end direction. This reversal in the direction of movement of trypsinized plus-end vesicles, in light of further observation that cytosol promotes primarily minus-end movement of liposomes, suggests that the machinery for cytoplasmic dynein-driven, minus-end vesicle movement can establish a functional interaction with the lipid bilayers of both vesicle populations. The additional finding that kinesin overrides cytoplasmic dynein when both are bound to bead surfaces indicates that the direction of vesicle movement could be regulated simply by the presence or absence of a tightly bound, plus-end kinesin motor; being processive and tightly bound, the kinesin motor would override the activity of cytoplasmic dynein because the latter is weakly bound to vesicles and less processive. In support of this model, it was found that (a) only plus-end vesicles copurified with tightly bound kinesin motors; and (b) both plus- and minus-end vesicles bound cytoplasmic dynein from cytosol.

FliN is a major structural protein of the C-ring in the Salmonella typhimurium flagellar basal body

The Salmonella typhimurium FliN protein has been proposed to form a mutually interacting complex with FliG and FliM, the switch complex, that is required for flagellar morphogenesis and function. We have used affinity chromatography for purification of extended flagellar basal bodies sufficient for quantitative analysis of their protein composition. The belled, extended structure is predominantly comprised of the switch complex proteins; with FliN present in the most copies (111 +/- 13). This explains why single, missense fliN, fliG or fliM mutations, found in many non-motile strains, can alter the belled morphology. Cell lysates from these strains contained the wild-type complement of FliG, FliM and FliN; but the basal bodies lacked the outer, cytoplasmic(C)-ring of the bell and were separated by sedimentation from FliM and FliN. The amount of FliG present in basal bodies from wild-type and one such mutant, FliN100LP, was comparable. These data show that: (1) the mutations define a FliG and FliMFliN multiple contact interface important for motility. (2) FliG is responsible for the increased size of the membrane-embedded MS-ring complex of belled relative to acid-treated basal bodies. (3) FliN, together with FliM, account for most of the C-ring. As a major component of the C-ring, FliN is distinct from the other proteins implicated in axial flagellar protein export. Inner, cytoplasmic rod basal substructure, seen by negative-stain and quick-freeze replica electron microscopy, may gate such export. Lack of connectivity between the cytoplasmic rod and ring substructures places contacts between FliG and FliMFliN at the periphery of the basal body, proximal to the flagellar intramembrane ring particles. This topology is consistent with models where torque results from interaction of circumferential arrays of the switch complex proteins with the ring particles.

An axoplasmic myosin with a calmodulin-like light chain

Organelles in the axoplasm from the squid giant axon move along exogenous actin filaments toward their barbed ends. An approximately 235-kDa protein, the only band recognized by a pan-myosin antibody in Western blots of isolated axoplasmic organelles, has been previously proposed to be a motor for these movements. Here, we purify this approximately 235-kDa protein (p235) from axoplasm and demonstrate that it is a myosin, because it is recognized by a pan-myosin antibody and has an actin-activated Mg-ATPase activity per mg of protein 40-fold higher than that of axoplasm. By low-angle rotary shadowing, p235 differs from myosin II and it does not form bipolar filaments in low salt. The amino acid sequence of a 17-kDa protein that copurifies with p235 shows that it is a squid optic lobe calcium-binding protein, which is more similar by amino acid sequence to calmodulin (69% identity) than to the light chains of myosin II (33% identity). A polyclonal antibody to this light chain was raised by using a synthetic peptide representing the calcium binding domain least similar to calmodulin. We then cloned this light chain by reverse transcriptase-PCR and showed that this antibody recognizes the bacterially expressed protein but not brain calmodulin. In Western blots of sucrose gradient fractions, the 17-kDa protein is found in the organelle fraction, suggesting that it is a light chain of the p235 myosin that is also associated with organelles.

Direct visualization of a transmembrane channel associated with ribosomes

Examination of directly frozen rough endoplasmic reticulum (ER) of retinal pigment epithelial cells by freeze-fracture and freeze-substitution revealed distinct paired transmembrane proteins associated with membrane ribosomes. Ribosomal subunits on intact ER membrane are directly visualized for the first time, providing a global view of the structure of the ribosome and the corresponding structures on the ER membrane. The ribosomal intersubunit cleft appears to be continuous with a cleft between paired transmembrane proteins that extends into the lumen of the ER. This continuous cleft may be the path taken by nascent polypeptides.

Freeze-substitution as a preparative technique for immunoelectronmicroscopy: evaluation by atomic force microscopy

Cryofixation followed by freeze substitution in osmium tetroxide was evaluated as a method for preparing biological specimens for immunoelectronmicroscopy. Samples were rapidly frozen by impact onto a sapphire block cooled with liquid nitrogen, substituted at -80 degrees C in acetone containing osmium tetroxide, and embedded in epoxy resin. With this protocol, excellent ultrastructure can be combined with localization of antigens that otherwise would be inactivated by the osmium, but labeling may need to be enhanced by chemically etching the sections prior to staining. The effects of etching on various structures in the sections were investigated by examining the sections with atomic force microscopy, an approach that yields three-dimensional views of the surface of the section. A considerable part of the section was removed or collapsed by the etching, and these effects occurred differentially in several components of the tissue and with different etching protocols. Nevertheless, the results suggest that the partial removal of the plastic by etching of freeze-substituted tissue can be explored as a method for exposing fine biological structures for observation with atomic force microscopy.

Actin-based motility of isolated axoplasmic organelles

We previously showed that axoplasmic organelles from the squid giant axon move toward the barbed ends of actin filaments and that KI-washed organelles separated from soluble proteins by sucrose density fractionation retain a 235-kDa putative myosin. Here, we examine the myosin-like activities of KI-washed organelles after sucrose density fractionation to address the question whether the myosin on these organelles is functional. By electron microscopy KI-washed organelles bound to actin filaments in the absence of ATP but not in its presence. Analysis of organelle-dependent ATPase activity over time and with varying amounts of organelles revealed a basal activity of 350 (range: 315-384) nmoles Pi/mg/min and an actin-activated activity of 774 (range: 560-988) nmoles/mg/min, a higher specific activity than for the other fractions. By video microscopy washed organelles moved in only one direction on actin filaments with a net velocity of 1.11 +/- .03 microns/s and an instantaneous velocity of 1.63 +/- 0.29 microns/s. By immunogold electronmicroscopy, 7% of KI-washed organelles were decorated with an anti-myosin antibody as compared to 0.5% with non-immune serum. Thus, some axoplasmic organelles have a tightly associated myosin-like activity.

Transport of cytoskeletal elements in the squid giant axon

In order to explore how cytoskeletal proteins are moved by axonal transport, we injected fluorescent microtubules and actin filaments as well as exogenous particulates into squid giant axons and observed their movements by confocal microscopy. The squid giant axon is large enough to allow even cytoskeletal assemblies to be injected without damaging the axon or its transport mechanisms. Negatively charged, 10- to 500-nm beads and large dextrans moved down the axon, whereas small (70 kDa) dextrans diffused in all directions and 1000-nm beads did not move. Only particles with negative charge were transported. Microtubules and actin filaments, which have net negative charges, made saltatory movements down the axon, resulting in a net rate approximating that previously shown for slow transport of cytoskeletal elements. The present observations suggest that particle size and charge determine which materials are transported down the axon.

Cytoplasmic constriction and vesiculation after axotomy in the squid giant axon

The squid giant axon responded to a transection injury by producing a gradient of cytoplasmic and vesicular changes at the cut end. At the immediate opening of the cut axon the cytoplasm was fragmented and dispersed and the vesicles in this region were in rapid Brownian movement. Approximately 0.1 mm further in, at the site of maximal axonal constriction, the axoplasm was condensed into a compact, constricted mass containing many large vesicles. The axoplasm was normal a few millimetres beyond this constricted, vesiculated end. It appears that transection triggered the transformation of normal axoplasm into a tightly constricted, highly vesiculated structure. This modified axoplasm at the cut end may slow the spread of damage and degeneration by preventing the bulk outflow of axoplasm, by slowing down the loss of intracellular molecules and by slowing down the influx of destructive extracellular ions (like calcium and chloride).

Effect of calpain on the composition and structure of postsynaptic densities

Endogenous calcium-activated proteases, the calpains, are thought to play a role in the regulation of postsynaptic function. Here we characterize some biochemical and morphological effects of calpain on isolated postsynaptic densities (PSDs). When a PSD preparation from rat forebrain was treated with exogenous calpain, many proteins, including spectrin, tubulin and the alpha-subunit of calcium calmodulin-dependent protein kinase (alpha-CaM kinase), were proteolyzed at varying rates, while another major protein, actin, remained intact. The selectivity of calpain action became more apparent in experiments designed to achieve limited proteolysis by using a lower calpain-to-protein ratio; it was possible to obtain extensive breakdown of spectrin with no decrease in the levels of either tubulin, alpha-CaM kinase, or actin. Electron microscopy of freeze-substituted preparations showed that limited calpain action caused a partial unraveling of the PSD, in which the characteristic central dense lamina became wider and less dense. We interpret these changes as due to calpain-mediated breakdown of cross-bridging elements, leading to a partial unraveling of the central PSD lamina. Opening up of the PSD structure following limited calpain action could facilitate exposure of previously occluded functional sites within the PSD and contribute to the modification of the synaptic function.

Conserved machinery of the bacterial flagellar motor

Novel periplasmic and cytoplasmic structural modules of the bases of bacterial flagella have been observed in situ and isolated using new biochemical protocols. Flagellar rotation may depend upon interactions of these modules with the intramembrane particle rings, a ubiquitous feature of flagellar bases necessary for torque generation. The outer membrane-associated basal disk of the Wolinella succinogenes polar flagellum has architecture well suited for interaction with the ring particles. However, antibody against the main W. succinogenes basal disk protein did not cross-react with flagella-enriched fractions from Salmonella typhimurium and Bacillus firmus; nor have such structures been observed in these species thus far. Antibodies against two S. typhimurium proteins, FliG and FliM, known to be involved in motor function and part of the cytoplasmic module in this species cross-reacted with flagella-enriched fractions from both W. succinogenes and B. firmus. In addition, flagellar cytoplasmic structure could be isolated from B. firmus. The basal disk may anchor the flagellar motor to the cell wall in some polar bacteria, but this does not seem to be a unique strategy. In contrast, the data indicate that the cytoplasmic module is conserved.

Continuous network of endoplasmic reticulum in cerebellar Purkinje neurons

Purkinje neurons in rat cerebellar slices injected with an oil drop saturated with 1,1'-dihexadecyl-3,3,3',3'-tetramethylindocarbocyanine perchlorate [DiIC16(3) or DiI] to label the endoplasmic reticulum were observed by confocal microscopy. DiI spread throughout the cell body and dendrites and into the axon. DiI spreading is due to diffusion in a continuous bilayer and is not due to membrane trafficking because it also spreads in fixed neurons. DiI stained such features of the endoplasmic reticulum as densities at branch points, reticular networks in the cell body and dendrites, nuclear envelope, spines, and aggregates formed during anoxia nuclear envelope, spines, and aggregates formed during anoxia in low extracellular Ca2+. In cultured rat hippocampal neurons, where optical conditions provide more detail, DiI labeled a clearly delineated network of endoplasmic reticulum in the cell body. We conclude that there is a continuous compartment of endoplasmic reticulum extending from the cell body throughout the dendrites. This compartment may coordinate and integrate neuronal functions.

Interactions among endoplasmic reticulum, microtubules, and retrograde movements of the cell surface

Relationships among the endoplasmic reticulum (ER), microtubules, and bead movements on the cell surface were investigated in the thin peripheral region of A6 cells, a frog kidney cell line. ER tubules were often aligned with microtubules, as shown by double-labeling with DiOC6(3) and anti-tubulin in fixed cells. In living cells stained with DiOC6(3) and observed in time lapse, there were frequent extensions, but few retractions, of ER tubules. In addition, there was a steady retrograde (towards the cell center) movement of all of the ER at approximately 0.3 microns/min. Since microtubules are often aligned with the ER, microtubules must also be moving retrogradely. By simultaneous imaging, it was found that the ER moves retrogradely at the same rate as aminated latex beads on the cell surface. This indicates that the mechanisms for ER and bead movement are closely related. Cytochalasin B stopped bead and ER movement in most of the cells, providing evidence that actin is involved in both retrograde movements. The ER retracted towards the cell center in nocodazole while both ER and microtubules retracted in taxol. Time lapse observations showed that for both drugs, the retraction of the ER is the result of retrograde movement in the absence of new ER extensions. Presumably, ER extensions do not occur in nocodazole because of the absence of microtubules, and do not occur in taxol because taxol-stabilized microtubules move retrogradely and there is no polymerization of new microtubule tracks for ER elongation.

Evidence for myosin motors on organelles in squid axoplasm

Squid axoplasm has proved a rich source for the identification of motors involved in organelle transport. Recently, squid axoplasmic organelles have been shown to move on invisible tracks that are sensitive to cytochalasin, suggesting that these tracks are actin filaments. Here, an assay is described that permits observation of organelles moving on unipolar actin bundles. This assay is used to demonstrate that axoplasmic organelles move on actin filaments in the barbed-end direction, suggesting the presence of a myosin motor on axoplasmic organelles. Indeed, axoplasm contains actin-dependent ATPase activity, and a pan-myosin antibody recognized at least four bands in Western blots of axoplasm. An approximately 235-kDa band copurified in sucrose gradients with KI-extracted axoplasmic organelles, and the myosin antibody stained the organelle surfaces by immunogold electron microscopy. The myosin is present on the surface of at least some axoplasmic organelles and thus may be involved in their transport through the axoplasm, their movement through the cortical actin in the synapse, or some other aspect of axonal function.

Multiple structural types of gap junctions in mouse lens

Gap junctions in the epithelium and superficial fiber cells from young mice were examined in lenses prepared by rapid-freezing, and processed for freeze-substitution and freeze-fracture electron microscopy. There appeared to be three structural types of gap junction: one type between epithelial cells and two types between fiber cells. Epithelial gap junctions seen by freeze-substitution were approximately 20 nm thick and consistently associated with layers of dense material lying along both cytoplasmic surfaces. Fiber gap junctions, in contrast, were 15–16 nm (type 1) or 17–18 nm thick (type 2), and had little associated cytoplasmic material. Type 1 fiber gap junctions were extensive in flat expanses of cell membrane and had a thin, discontinuous central lamina, whereas type 2 fiber gap junctions were associated with the ball-and-socket domains and exhibited a dense, continuous central lamina. Both types of fiber gap junction had a diffuse arrangement of junctional intramembrane particles, whereas particles and pits of epithelial gap junctions were in a tight, hexagonal configuration. The type 2 fiber gap junctions, however, had a larger particle size (approximately 9 nm) than the type 1 (approximately 7.5 nm). In addition, a large number of junctional particles typified the E-faces of both fiber types but not the epithelial type of gap junction. Gap junctions between fiber and epithelial cells had structural features of type 1 fiber gap junctions.(ABSTRACT TRUNCATED AT 250 WORDS)

Inhibition of endogenous phosphatase in a postsynaptic density fraction allows extensive phosphorylation of the major postsynaptic density protein

The major postsynaptic density protein, proposed to be a calcium/calmodulin-dependent protein kinase, becomes phosphorylated when a postsynaptic density preparation from rat cerebral cortex is incubated in medium containing calcium and calmodulin. Upon longer incubation, however, the level of phosphorylation declines, suggesting the presence of a phosphatase activity. When Microcystin-LR, a phosphatase inhibitor, is included in the phosphorylation medium, the decline in phosphorylation is prevented and a higher maximal level of phosphorylation can be achieved. Under these conditions, the maximal phosphorylation of major postsynaptic density protein is accompanied by a nearly complete shift in its electrophoretic mobility from 50 kDa to 54 kDa, similar to that described for the alpha subunit of the soluble calcium/calmodulin-dependent protein kinase II. Of the four major groups of serine/threonine protein phosphatases, the enzyme responsible for the dephosphorylation of major postsynaptic density protein is neither type 2C, which is insensitive to Microcystin-LR, nor type 2B, which is calcium-dependent. As Microcystin-LR is much more potent than okadaic acid in inhibiting the dephosphorylation of major postsynaptic density protein, it is likely that the postsynaptic density-associated phosphatase is a type 1. The above results indicate that the relatively low level of phosphorylation of the major postsynaptic density protein observed in preparations containing postsynaptic densities is not due to a difference between the cytoplasmic and postsynaptic density-associated calcium/calmodulin-dependent kinases as previously proposed, but to a phosphatase activity, presumably belonging to the type 1 group.

Single kinesin molecules crossbridge microtubules in vitro

Kinesin is a cytoplasmic motor protein that moves along microtubules and can induce microtubule bundling and sliding in vitro. To determine how kinesin mediates microtubule interactions, we determined the shapes and mass distributions of squid brain kinesin, taxol-stabilized microtubules (squid and bovine), and adenosine 5'-[beta, gamma-imido]triphosphate-stabilized kinesin-microtubule complexes by high-resolution metal replication and by low-temperature, low-dose dark-field scanning transmission electron microscopy of unfixed, directly frozen preparations. Mass mapping by electron microscopy revealed kinesins loosely attached to the carbon support as asymmetrical dumbbell-shaped molecules, 40-52 nm long, with a mass of 379 +/- 15 kDa. The mass distribution and shape of these molecules suggest that these images represent kinesin in a shortened conformation. Kinesin-microtubule complexes were organized as bundles of linearly arrayed microtubules, stitched together at irregular intervals by cross-bridges typically < or = 25 nm long. The crossbridges had a mass of 360 +/- 15 kDa, consistent with one kinesin per crossbridge. These results suggest that kinesin has a second microtubule binding site in addition to the known site on the motor domain of the heavy chain; this second site may be located near the C terminus of the heavy chains or on the associated light chains. Thus, kinesin could play a role in either crosslinking or sliding microtubules.

Quantitative scanning transmission electron microscopy of ultrathin cryosections: subcellular organelles in rapidly frozen liver and cerebellar cortex

Freeze-dried, ultrathin cryosections of directly frozen mouse liver and brain have been prepared and characterized by low-dose dark-field scanning transmission electron microscopy (STEM). These improved cryosections gave images comparable to those from conventional plastic sections. They were thin enough (< < 1.0 elastic mean free path) to use established dark-field techniques, modified for thickness-dependent nonlinearities, to measure the dry mass fraction of individual organelles, and hence to deduce their water content. Digital STEM imaging in combination with electron and X-ray spectroscopy has important biological applications, as illustrated by studies on calcium regulation in Purkinje neurons. Calcium concentrations per unit dry weight of dendritic compartments were determined by the peak/continuum method of energy-dispersive X-ray spectroscopy (EDXS), which necessarily overstates elemental concentrations because of beam-induced mass loss. The dry mass content of organelles at low dose and the percentage of dry mass retained after analysis at high dose were as follows: mitochondria (46.0 g dry mass/100 g hydrated mass, 67% mass retained); endoplasmic reticulum (27.9 g/100 g, 57%); and cytoplasm (16.3 g/100 g, 41%). These values were used to correct elemental concentrations for mass loss. Results indicated that the major calcium storage organelle in Purkinje cell dendrites is the endoplasmic reticulum, of which there are two types distinguished by their levels of calcium. Parallel electron energy loss spectroscopy of dendritic organelles corroborated EDXS measurements, with an improved sensitivity that indicates the feasibility of quantitative calcium mapping.

Kinesin is bound with high affinity to squid axon organelles that move to the plus-end of microtubules

This paper addresses the question of whether microtubule-directed transport of vesicular organelles depends on the presence of a pool of cytosolic factors, including soluble motor proteins and accessory factors. Earlier studies with squid axon organelles (Schroer et al., 1988) suggested that the presence of cytosol induces a > 20-fold increase in the number of organelles moving per unit time on microtubules in vitro. These earlier studies, however, did not consider that cytosol might nonspecifically increase the numbers of moving organelles, i.e., by blocking adsorption of organelles to the coverglass. Here we report that treatment of the coverglass with casein, in the absence of cytosol, blocks adsorption of organelles to the coverglass and results in vigorous movement of vesicular organelles in the complete absence of soluble proteins. This technical improvement makes it possible, for the first time, to perform quantitative studies of organelle movement in the absence of cytosol. These new studies show that organelle movement activity (numbers of moving organelles/min/micron microtubule) of unextracted organelles is not increased by cytosol. Unextracted organelles move in single directions, approximately two thirds toward the plus-end and one third toward the minus-end of microtubules. Extraction of organelles with 600 mM KI completely inhibits minus-end, but not plus-end directed organelle movement. Upon addition of cytosol, minus-end directed movement of KI organelles is restored, while plus--end directed movement is unaffected. Biochemical studies indicate that KI-extracted organelles attach to microtubules in the presence of AMP-PNP and copurify with tightly bound kinesin. The bound kinesin is not extracted from organelles by 1 M KI, 1 M NaCl or carbonate (pH 11.3). These results suggest that kinesin is irreversibly bound to organelles that move to the plus-end of microtubules and that the presence of soluble kinesin and accessory factors is not required for movement of plus-end organelles in squid axons.

The cytoplasmic component of the bacterial flagellar motor

We have used electron microscopy to examine freshly isolated Salmonella typhimurium and Escherichia coli basal flagellar fragments, purified without resort to extremes of pH or ionic strength. Such fragments contain the large bell-like basal structures visualized recently in freeze-substituted or fixed preparations. We have found mot (non-motile) mutants produced by lesions in fli genes (G, M, N) in which the bell structures do not coisolate with the flagellar basal body. The coisolation of the bell with the flagellar basal body was unaffected in strains lacking the genes for the motility-associated Mot proteins or for the Che family of proteins, which are necessary for chemotaxis. Proper assembly and interaction of the cytoplasmically located bell with the membrane-associated flagellar basal structures appears to be necessary for motor function. The FliG, FliM, and FliN proteins are thought to form a structural complex responsible for energization and switching of the flagellar motor. Our findings are consistent with the existence of such a complex and imply that it forms part of the flagellar bell.

Characterization of endoplasmic reticulum by co-localization of BiP and dicarbocyanine dyes

The original concept of endoplasmic reticulum derived from the observation of a reticular network in cultured fibroblasts by electron microscopy of whole cells. It was previously reported that the fluorescent dye, DiOC6(3), stains a similar network as well as mitochondria and other organelles in living cells. Here, we investigate the significance of the structures labeled by DiO6(3) in CV-1 cells, a monkey epithelial cell line. First, we show that the network stained in living CV-1 cells is preserved by glutaraldehyde fixation and then we co-label it with an antibody against BiP (immunoglobulin binding protein), a protein commonly accepted to be present in the endoplasmic reticulum. Anti-BiP labeled the same network as that labeled by DiOC6(3), so this network now is identified as being part of the endoplasmic reticulum. DiOC6(3) labels many other membrane compartments in addition to the endoplasmic reticulum. This, along with its lipophilic properties, suggests that DiOC6(3) stains all intracellular membranes. However, the extensive reticular network in the thin peripheral regions of cultured cells is easily distinguished from these other membranes. Thus, staining by DiOC6(3) is a useful method for localizing the endoplasmic reticulum, particularly in thin peripheral regions of cultured cells.

Structural domains of the tight junctional intramembrane fibrils

Freeze-fracture reveals intramembrane fibrils lying along the intermembrane contacts that characterize tight junctions. Tight junctions from a variety of species are reexamined here by rapid freezing prior to freeze-fracture. The tight junction fibril is uprooted alternatively from either the cytoplasmic or the exoplasmic hemibilayer during freeze-cleavage, exposing two distinct but complementary views of its hybrid structure within the same replica. When the transmembrane fibril is uprooted from the exoplasmic hemibilayer it appears on the P-fracture face as a smooth-surfaced cylinder which is sometimes resolved into periodic globular structures. The lack of indication that the P-face cylinder has been pulled out through the opposite membrane half indicates that this domain of the fibril is, in large part, buried in the hydrophobic interior of the membrane. However, when the transmembrane fibril is uprooted from the cytosolic hemibilayer it appears on the E-fracture face as a row of irregular intramembrane particles. The irregular particles on the E-face aspect of the fibril are interpreted as corresponding to transmembrane protein segments that may very well make projections onto the cytosolic surface of the bilayer. En face views of the outermost junction strand between adjacent epithelial cells show periodic lines on the bilayer on each side of the junction which are interpreted as periodic transmembrane protein segments arising from the core structure of the tight junction fibril. If the backbone of the tight junction strand is an inverted cylindrical micelle, it must typically include proteins, which might anchor it to structures outside the membrane bilayer.

New structural features of the flagellar base in Salmonella typhimurium revealed by rapid-freeze electron microscopy

The structure of the flagellar base in Salmonella typhimurium has been studied by rapid-freeze techniques. Freeze-substituted thin sections and freeze-etched replicas of cell envelope preparations have provided complementary information about the flagellar base. The flagellar base has a bell-shaped extension reaching as far as 50 nm into the bacterial cytoplasm. This structure can be recognized in intact bacteria but was studied in detail in cell envelopes, where some flagella lacking parts of the bell were helpful in understanding its substructure. Structural relationships may be inferred between this cytoplasmic component of the flagellum and the recently described flagellar intramembrane particle rings as well as the structures associated with the basal body in isolated, chemically fixed flagella.

Pertussis toxin-sensitive G proteins are transported toward synaptic terminals by fast axonal transport

We find that half of the pertussis toxin-sensitive guanine nucleotide-binding protein (G protein) in the squid (Loligo pealei) giant axon is cytoplasmic and that this species of G protein is intermediate in size between the two forms present in axolemma. This G protein is transported toward synaptic terminals at 44 mm/day. Moreover, these data are consistent with there being two additional steps leading to the maturation of G proteins: (i) association with and transport on intracellular organelles and (ii) modification at the time of transfer to the plasmalemma resulting in a molecular weight shift. Since the other two components of G protein-mediated signal transduction pathways, receptors and effector enzymes, are known to be delivered to the synaptic terminals by fast axonal transport, our findings introduce the possibility that these three macromolecules are assembled as a complex in the cell body and delivered together to the plasma membrane of the axon and synaptic terminals.