The local transcriptome in the synaptic neuropil revealed by deep sequencing and high-resolution imaging

Abstract

In neurons, dendritic protein synthesis is required for many forms of long-term synaptic plasticity. The population of mRNAs that are localized to dendrites, however, remains sparsely identified. Here, we use deep sequencing to identify the mRNAs resident in the synaptic neuropil in the hippocampus. Analysis of a neuropil data set yielded a list of 8,379 transcripts of which 2,550 are localized in dendrites and/or axons. Using a fluorescent barcode strategy to label individual mRNAs, we show that their relative abundance in the neuropil varies over 3 orders of magnitude. High-resolution in situ hybridization validated the presence of mRNAs in both cultured neurons and hippocampal slices. Among the many mRNAs identified, we observed a large fraction of known synaptic proteins including signaling molecules, scaffolds and receptors. These results reveal a previously unappreciated enormous potential for the local protein synthesis machinery to supply, maintain and modify the dendritic and synaptic proteome.

Figure 1. Isolation of Hippocampal RNA for Deep Sequencing

(A) Venn diagram showing overlap between mRNAs identified in three recently published neuropil transcriptome data sets. (B) Photograph of a slice before and after microdissection with somatic (red) and neuropil (blue) segments outlined. Scale bar = 1 mm. (C) Scheme of an individual microdissected slice, showing the orientation of CA1 neuron dendrites (for slice preparation, see Experimental Procedures). (D) Microdissected regions containing the somatic layer of CA1 are enriched with the neuron-specific transcription factor NeuN, whereas neuropil microdissected slices are de-enriched for NeuN. Protein lysates were prepared from neuropil, somata, and whole CA1 (CA1) tissues and analyzed by western blot using antibodies against NeuN and β-actin. (E) Graph shows the enrichment of proteins in the different CA1 regions. (F) Total RNA was isolated from the tissue corresponding to the microdissected neuropil and the microdissected somatic layer. The purity of the sample was analyzed using an Agilent 2100 Bioanalyzer. mRNAs are distributed throughout the lanes.

Figure 2. Flowchart of Experiments and Analysis

Shown is a description of the experiments, beginning with sample preparation and concluding with the analysis and validation of the neuropil mRNA population.

Figure 3. Gene Ontology Analysis of Neuropil Transcriptome and Validation and Quantification of Target Transcripts Using Nanostring

(A) Bar graph illustrating some transcript families from Gene Ontology (GO) that are significantly overrepresented in our samples. (B) Example of Nanostring image showing unique fluorescent bar codes for individual mRNA transcripts. Bright red spots are fiducial markers used to register high magnification images into a montage. Scale bar = 2 μm. (C) Plot of individual mRNA species and their abundance from Nanostring experiments (n = 3 experiments). A selection of mRNAs representing synaptically relevant transcripts is indicated by red circles. For full data set with numbers, including many more synaptically relevant transcripts see Table S5. See also Figure S1.

Figure 4. Relative Enrichment of Transcripts in the Neuropil, Glia, and Somata

(A) Heat map of a glia versus neuropil ramp experiment. Transcript enrichment is encoded in the heat map from low (green) to high (red). Transcripts that show similar expression patterns are clustered together, as indicated by the colored groups to the left of the heat map. (B) Summary graph indicating the significant enrichment of transcripts in either the neuropil (blue) or glial samples (red), (n = 3). (C) Heat map of a somatic layer versus neuropil ramp experiment. Transcript enrichment is encoded in the heat map from low (green) to high (red). Transcripts that show similar expression patterns are clustered together, as indicated by the colored groups to the left of the heat map. In the heat map it is evident that a clusters of ~105 and ~150 transcripts are enriched in the neuropil and somata, respectively. (D) Graph indicating those transcripts which exhibited a significant enrichment in either the neuropil (blue; n = 123) or somata (red; n = 14) (n = 3 experiments). See also Figure S2.

Figure 5. A Conservative Estimate of mRNAs Present in Synaptic Regions

(A) Venn diagram of the neuropil data set and its overlap with transcripts enriched in glia (astrocytes and oligodendroctyes) interneurons, mitochondria and nucleus as well as the endothelium (blood vessels). The remaining transcripts number is 2,550. (B) Analysis of the filtered neuropil list, showing a significant enrichment of transcripts in categories related to synaptic function and cell biological processes. (C) Scheme of a postsynaptic compartment highlighting some of the transcript families for synaptically relevant proteins represented in the filtered neuropil list. See also Figures S3 and S7.

Figure 6. High-Resolution Fluorescent In Situ Hybridization Detection of mRNA Transcripts in the Dendrites of Cultured Hippocampal Neurons

(A) Fluorescent in situ hybridization (FISH) signal in cultured hippocampal neurons (DIV 21) showing the presence of Camk2a and Shank1 mRNA molecules (red particles) within the cell body and entire dendritic arbor. The neuron was immunostained with an antibody to MAP2 (green) to generate a mask that outlines the dendrites. Scale bar = 20 μm. (B) FISH signal in individual dendrites from cultured hippocampal neurons linearized for analysis. MAP2 immunostaining was used to generate a mask that outlines the dendritic structure. The red particles show the transcripts for each indicated gene within the dendrites. The transcripts vary in their abundance as well as their distribution along the length of the dendrite. Dendrites are shown in groups that display similar distribution patterns. Scale bar = 20 μm. (C) The transcripts are represented in groups identified by cluster analysis (see Experimental Procedures) based the distribution of mRNA particles along the dendritic axis. The cophenetic correlation coefficient for the dendrogram is 0.8058 and the generation of each of the four labeled clusters is significant: p < 0.001. See also Figures S4, S5, S6, and S7.

Figure 7. High-Resolution Fluorescent In Situ Hybridization Detection of mRNA Transcripts in the Synaptic Neuropil of Area CA1 of the Rat Hippocampus

(A) Raw image from an in situ hybridization experiment using antisense probes against the Psd (pleckstrin and Sec7 domain protein) mRNA (green) in area CA1 of rat hippocampal slices (left). Dendrites were immunostained using an anti-Map2 antibody (red). The nuclei were stained with DAPI (blue). There are several examples where mRNA puncta are clearly present within the dendrites (arrows). In order to increase visibility when displaying larger areas of area CA1 we processed the images by dilating the mRNA puncta in ImageJ after setting a threshold (middle). For printing purposes, images were converted for display on white background (right). The DAPI channel was converted to a binary image, inverted, and displayed in gray to provide orientation in the CA1 region, the mRNA signal was converted to red, and the two channels were merged using Adobe Photoshop. Scale bar = 4 μm. (B) Distribution of mRNA puncta in CA1 from in situ hybridization experiments in hippocampal slices displayed as indicated in (A) (right panel). In some control experiments, slices were not exposed to the probe (no probe) but to all subsequent amplification steps; in other controls, a sense probe followed by all subsequent amplification steps was used. Experiments were performed with the indicated probes. Scale bar = 25 μm.