Rapid Spine Delivery and Redistribution of AMPA Receptors After Synaptic NMDA Receptor Activation
Tóm tắt
Từ khóa
Tài liệu tham khảo
GFP (enhanced GFP; Clontech) was inserted between the third and fourth amino acids after the predicted signal peptide cleavage site of rat GluR1-flop cDNA with standard molecular biology techniques. HEK cells were transfected with plasmid-based mammalian expression vector with lipofectin (Gibco-BRL Life Technologies). GluR1-GFP and GluR2 were cotransfected in 1:4 ratio. Protein immunoblotting was carried out with antibodies to GluR1 COOH-terminal (0.1 μg/ml; Chemicon International). Whole-cell recordings were obtained 2 to 5 days after transfection in Hepes (10 mM)-buffered Hanks' solution in the presence of 100 μM cyclothiazide. Kainate (1 mM) was applied through a puffer pipette positioned close to recorded cell. Five to ten records were obtained at holding potentials of −80 to 60 mV (20-mV steps).
; R. Malinow et al. in Imaging Living Cells (Cold Spring Harbor Laboratory Press Cold Spring Harbor New York in press).
Whole-cell recordings from neurons infected with Sindbis virus for 1 to 4 days show normal passive membrane properties (for example input resistance: uninfected 276 ± 64 N = 13; GluR1-GFP infected 302 ± 51 N = 8; P = 0.78) (S. Shi Y. Hayashi R. Malinow unpublished results).
Dissociated cultured neurons were prepared as previously described (33).
To estimate the level of recombinant GluR1-GFP expression relative to endogenous GluR1 expression we performed immunohistochemistry on fixed (14) dissociated hippocampal neurons (33) with a GluR1 COOH-terminal antibody (1 μg/ml; Chemicon International) that recognizes both proteins (Fig. 1B) as primary and Texas Red coupled as secondary (which does not overlap in fluorescence with GFP). In a field of infected and noninfected cells dendrites (80 μm from cell body) of infected cells showed 2.7 ± 0.2 fold (mean ± SEM N = 8) immunolabel compared with similar regions of uninfected cells.
Tissue was fixed with freshly made 4% paraformaldehyde and 4% sucrose in phosphate-buffered saline (PBS) (dissociated neurons: 4°C for 30 min; organotypic slices (16): 4°C for 1 hour). With such fixation immunohistochemistry (see below) detected only surface epitopes (dissociated cells: Fig. 2C; organotypic slices: Fig. 3G). To detect intracellular epitopes tissue was further treated with 0.3% Triton X-100 in PBS (dissociated cells: 4°C for 10 min; organotypic slices: 4°C for 30 min). Immunohistochemistry: Cells were blocked in blocking solution (10% horse serum in PBS; 60 min) and then incubated with primary antibody (4°C overnight) in blocking solution. All primary rabbit polyclonal antibodies were visualized with biotin-conjugated antibody to rabbit immunoglobulin G (IgG) and Texas Red–avidin system. Mouse monoclonal antibodies were visualized with antibodies to mouse IgG labeled with Cy5 (Jackson ImmuoResearch). Images of dissociated neurons were collected with a cooled charge-coupled device (Photometrics) and analyzed with PMIS or V for Windows (Photometrics); organotypic slices were imaged with TPLSM (21). Fraction of GluR1-GFP molecules on surface for a given region can be quantified by computing [(antibody to GFP signal)/(GFP signal)] in nonpermeabilized conditions for that region and normalizing by [(antibody to GFP signal)/(GFP signal)] obtained from similar tissue in permeabilized conditions. Signals in above computation indicate background subtracted values.
For whole-cell recordings from dissociated neurons internal solution consisted of (in mM) 115 cesium methanesulfonate 20 CsCl 10 Hepes 2.5 MgCl 2 4 Na 2 -adenosine triphosphate 0.4 Na-guanosine triphosphate 10 Na-phosphocreatine and 0.6 EGTA (pH 7.2). Neurons were perfused with artificial cerebrospinal fluid (ACSF) of the following composition (in mM): 119 NaCl 2.5 KCl 4 CaCl 2 4 MgCl 2 26.2 NaHCO 3 1 NaH 2 PO 4 and 11 glucose; then the neurons were gassed with 5% CO 2 and 95% O 2 at 28°C. Responses to glutamate were evoked by adding (in μM) 40 l -glutamic acid γ-(α-carboxyl-2-nitrobenzyl) ester (Molecular Probes) 1 tetrodotoxin 100 picrotoxin 100 APV and 1 3-((R)-2-carboxypiperazin-4-yl)-propyl-1-phosphoric acid. Ten to 25 ms of flash (100 W of Hg; quartz optics Zeiss Axiovert 135) was delivered to a spot (∼50-μm diameter) positioned over dendrites.
Organotypic slice culture of hippocampus was prepared as described [
]. Hippocampal slices (400 μm) were cut from postnatal 6- to 8-day-old rats with tissue chopper and cultured in medium as described [
]. Slices were placed in a recording chamber and perfused with ACSF. Synaptic transmission was evoked (0.1 Hz 100 μs ∼10 V) with a glass stimulation electrode (1 to 2 megaohms filled with 3 M NaCl) positioned in CA1 stratum radiatum. Tetanic stimuli consisted of two 100-Hz trains of 1-s duration separated by 20 s.
The distribution of GluR1 and GluR1-GFP was measured quantitatively with immuno-gold labeling (18). Sections from dendritic regions ∼100 μm from the cell bodies were examined. To identify a cell infected with GluR1-GFP virus we identified a dendritic segment (∼1 μm in diameter) with immunolabeling. Comparable regions (with dendritic segments ∼1 μm in diameter) were chosen in tissue from uninfected P10 animals for analysis of endogenous GluR1. For each section the label was identified as being in one of the following compartments: A dendritic cytosol; B dendritic surface; C spine non-PSD; and D PSD. The total number of grains counted was 515 for endogenous GluR1 and 785 for GluR1-GFP. For background subtraction 50 presynaptic terminals were randomly chosen label was counted and density was calculated (0.17 gold per square micrometer). This background density constituted 10 to 20% of the signal found in dendritic cytosol (0.92 gold per square micrometer) and was subtracted to reach dendritic cytosol value (A above); this region is the only compartment with sufficient area to be affected by background labeling.
; Z. F. Mainen et al. Methods in press. Optical stacks (30 μm by 30 μm by 50 μm; step size 0.5 μm) were captured about every 15 min with a custom-built two-photon laser scanning microscope [Zeiss 63X objective; Ti:sapphire laser tuned to wavelength (λ) ∼ 900 nm]. Stimulation electrode was placed nearby dendrites expressing GluR1-GFP under visual guidance. Quantification of fluorescence from images was carried out with custom-made programs in IDL (Research Systems). Images were imported into Adobe Photoshop for figure presentation. Each displayed image is generally the average of three to five optical sections. In some cases an image is the composite of images captured at slightly different (∼1 μm) optical sections; this was necessary to offset tilting of structures over the course of hours of examination.
Single-spine Ca 2+ imaging studies with the same stimulus electrodes (tip resistance 1 to 2 megaohms) and parameters indicate that synaptic excitation in this preparation is sparse. Thus as in previous studies (21 37) the tip of the stimulating electrode must be placed around 5 to 15 μm from imaged postsynaptic regions.
Within 3 SD of the noise measured in background regions.
The random fluctuation of GluR1-GFP fluorescence was measured in six spines identified during control conditions. At these spines the GluR1-GFP intensity decreased 344 ± 259 AU (mean ± SD) from one observation period to the next (from 1755 ± 613 AU to 1411 ± 478 AU). Thus the increase in spine GluR1-GFP fluorescence after tetanus differs (by 6 SD for “active” spines and by 7 SD for “empty” spines) from the average changes seen in the absence of tetanus. This indicates that such events are unlikely to occur from chance alone. Furthermore the density of GluR1-GFP–containing spines detected after tetanus in the analyzed dendrites (0.93 ± 0.12 μm −1 over 10 dendritic segments 7.5 ± 4.7 μm in length) was ∼16 SD outside the mean density of GluR1-GFP–containing spines detected in the absence of tetanus (0.021 ± 0.055 μm −1 mean ± SD measured over 30 randomly chosen dendritic segments).
After selecting a region of interest (about 8 μm by 8 μm by 3 μm) an autocorrelation function A ( r ) of fluorescence intensity I ( x y z ) was computed as a function of distance r : A(r)=ΣxΣyΣzΣθr[I(x y z)*I(xθr yθr z)]/ [I(x y x)+I(xθr yθr z)]2where θ r indexes x and y over 72 locations about a circle centered at x y and of radius r. This function decayed smoothly and monotonically as a function of distance (Fig. 5B) reflecting the nonperiodic distribution of fluorescence. We used the distance at which the autocorrelation function reached 50% decay ( R 50% ) as an index of signal homogeneity.
This decrease is 3.6 SD beyond those changes seen in the absence of stimulation indicating a small likelihood that these changes occur from chance alone.
Lledo P.-M., et al., ibid. 279, 399 (1998).
; P. Osten et al. ibid. p. 99; I. Song et al. ibid. p. 393.
Maletic-Savatic M., Malinow R., ibid. 18, 6803 (1998);
; M. Maletic-Savatic et al. ibid. p. 6814.
Takechi H., et al., ibid. 396, 757 (1998).
We thank S. Schlesinger for Sindbis virus expression system; P. De Camilli for antibody against synapsin I; C. S. Zucker for antibody to GFP; J. Boulter S. F. Heinemann and P. H. Seeburg for cDNA clones; N. Dawkins-Pisani B. Burbach and P. O'Brian for technical assistance; and Y.-X. Wang for assistance in the immunogold studies. Y.H. is a recipient of research fellowships from Japan Society for the Promotion of Science and Uehara Memorial Foundation. S.H.Z. was a recipient of Wellcome Trust Fellowship. This study was supported by the NIH (R.M. and K.S.) the Mathers Foundation (R.M.) and the Human Frontier Science Program Pew and Whitaker Foundations (K.S).