Fig 1.
Astrocytes lose fibroblast-like morphology in a neuron-glia coculture system.
(A) Astrocytes (DIV 22) grown in monoculture exhibit a fibroblast-like morphology and express Fibronectin, as illustrated by immunocytochemistry. Direct contact with neurons in coculture (DIV 22) led to increased process formation resembling a morphology more similar to protoplasmic astrocytes, and a drastically reduced expression of Fibronectin. Nuclei are stained with bis-Benzimide (bisBenz) and astrocytes or neurons are identified by either GFAP or MAP2 immunostainings, respectively (scale bar = 10 μm). (B) Immunoblot analysis of lysates derived from astrocytic monocultures and neuron-glia cocultures (DIV 22) demonstrate a loss of Fibronectin expression and a reduction in GFAP in cocultures.
Fig 2.
Up-regulation of BDNF-induced astroglial protein synthesis depends on the presence of neurons.
(A) Schematic drawing of the lentiviral expression vector LVGFAPEGFP that expresses EGFP cell type-specifically in astrocytes under control of the simplified GFAP promoter version gfaABC1D. (B) Astrocytes expressed EGFP cell type-specifically after infection with LVGFAPEGFP in a neuron-glia coculture (DIV 22). Cultures were treated for 4 h with either 50 ng/ml BDNF or 2 μM TTX in the presence of 4 mM AHA and processed for 'click reaction' with a TAMRA-alkyne tag in the FUNCAT procedure. BDNF application led to an upregulation of TAMRA signal intensities up to 150%. TAMRA intensities are color coded in the lower panel (scale bar = 10 μm). (C) GFAP-positive astrocytes in monoculture (DIV 20–22) expressing EGFP show fibroblast-like morphology. No differences in TAMRA signal intensities were observed when cultures were treated analogous to (B) (scale bar = 10 μm). (D) Cultures were manipulated as described in (B). Mean TAMRA signal intensities were determined within the EGFP mask. A significant increase in TAMRA signal intensities was detected after BDNF treatment. Numbers at the X-axis indicate the number of cells included in the quantification (3 independent experiments; represented data are mean +/- SEM; ONE-way ANOVA, ***: p<0.0001). (E) TAMRA signal intensities were quantified in GFAP-positive astrocytes expressing EGFP according to (D). No significant differences in signal intensities after either BDNF or TTX treatment were observed (3 independent experiments; represented data are mean +/- SEM; ONE-way ANOVA, p>0.05).
Fig 3.
Cell type-specific protein labeling with ANL and a mutated Mus musculus MetRS.
(A) ANL carries an enlarged side chain and therefore cannot fit into the methionine binding pocket of the wild type MetRS. An enlargement of the methionine binding pocket by a single amino acid exchange allows ANL binding and coupling to the Met-tRNA. (B) HEK293T cells overexpressing EGFP-tagged wild type mMetRSwt or the mutated versions mMetRSNLL, mMetRSPLL and mMetRSL274G were subjected to either 4 mM ANL, AHA or Met in Met-free medium for 1 h. Successful ANL integration into proteins was detected with a biotin-alkyne tag by performing a ‘click-reaction’ of cell lysates, followed by immunodetection of biotinylated proteins. Whereas the endogenous MetRS can process AHA, ANL integration is specific for cells carrying the mutated version EGFP-mMetRSL274G only. Blockade of protein synthesis with 100 μg/ml cycloheximide and 40 μM anisomycin (CA) prevents ANL integration and subsequent biotin tagging. Very weak to none-tagged biotin was observed for both EGFP-mMetRSNLL and EGFP-mMetRSPLL expressing cells. (C) HEK293T cells overexpressing either EGFP-mMetRSwt or EGFP-mMetRSL274G were incubated for 1 h with either 4 mM AHA or ANL in Met-free medium with or without 40 μM anisomycin (A). Cells were fixed and AHA or ANL labeled proteins were visualized with a TAMRA-alkyne tag in the ‘click reaction’. TAMRA-positive signals are specific for cells expressing EGFP-mMetRSL274G with comparable efficiency to AHA incorporation by the wild type MetRS. Notably, TAMRA signals are detectable within the entire cell. Moreover, metabolic incorporation of ANL can be blocked by anisomycin (A). Nuclei are stained with bis-Benzimide (bisBenz) (scale bar = 10 μm). (D) HEK293T cells overexpressing EGFP-mMetRSwt or EGFP-mMetRSL274G were incubated for 1 h with 4 mM Met, AHA or ANL either without Met or in combination with 4 mM Met in Met-free medium or standard cell culture medium containing 0.2 mM Met. Successful AHA or ANL integration was visualized by tagging with a biotin-alkyne tag. Low or equimolar Met concentrations reduced biotin-tagged proteins after AHA inoculation, whereas a biotin signal is observable after ANL integration. Although with reduced efficiency, ANL labeling is still possible under conditions where low concentrations of Met are present.
Fig 4.
(A) Scheme of the lentiviral construct driving the non-cell-selective expression of EGFP-mMetRSL274G under the ubiquitin promoter (LVEGFP-mMetRSL274G). (B) Primary neuron-glia cocultures (DIV 22), infected with LVEGFP-MetRSL274G were labeled for 4 h with either 4 mM ANL, AHA or Met. Subsequent immunostaining of MAP2 and Synaptophysin (Syphy) or visualization of ANL- or AHA-labeled proteins by ‘click reaction’ revealed an intact morphology of ANL-labeled cells indistinguishable from AHA or Met treated controls. TAMRA-positive signals are evident in neuronal dendrites and spines. Segments of dendrites in the detail were straightened and magnified (scale bar = 10 μm). (C) Neurons of a cortical neuron-glia coculture (DIV 22) that express EGFP-mMetRSL274G were labeled with 4 mM ANL for incubation times ranging from 0 h to 4 h. Newly synthesized proteins that integrated ANL were visualized with the TAMRA-alkyne tag by ‘click reaction’ and subsequent immunocytochemistry. Examples of representative 20 μm dendritic sections after the 'click reaction' display constant increases of TAMRA intensity in neuronal dendrites with ANL labeling duration ranging from 0 to 4 h. TAMRA stainings are intensity coded in the right panel. (D) Mean fluorescence intensities of ANL-labeled proteins were measured within the main dendrite in a section ranging from 20 μm–40 μm of the proximal end of the dendrite using the MAP2 staining as a mask. Mean fluorescence intensities in dendrites increased constantly over the observed time range suggesting a continuous integration of ANL into newly synthesized proteins over time. Numbers represent the number of analyzed dendrites (represented data are mean +/- SEM of 3 independent experiments).
Fig 5.
(A) Scheme of the lentivirus construct LVGFAPEGFP-mMetRSL274G driving the expression of EGFP-mMetRSL274G from the simplified GFAP promoter version gfaABC1D. (B) Primary neuron-glia cocultures (DIV 14), infected with the lentivirus LVGFAPEGFP-mMetRSL274G, expressed the enzyme cell type-specifically in astrocytes. Cultures were incubated with either 4 mM AHA, ANL or Met for 4 h in Met-free medium, and AHA or ANL incorporation was visualized using a TAMRA-alkyne tag in a subsequent ‘click-reaction’. Astrocytes are identified employing GFAP immunostaining. AHA labeling occurred in a non-cell-selective manner both in neurons and astrocytes. TAMRA-positive proteins are detected solely in GFAP-positive, EGFP-mMetRSL274G expressing astrocytes upon incubation with ANL. Comparable TAMRA signals are observable in ANL and AHA labeled cells with a high nuclear and cytoplasmatic intensity of tagged proteins (scale bar = 10 μm). (C) Primary neuron-glia cocultures were infected with LVGFAPEGFP-mMetRSL274G to express EGFP-mMetRSL274G in astrocytes. Cocultures (DIV 22) were subjected to either 4 mM AHA, ANL or Met in Met-free medium for 4 h. Live-staining with 2 μg/ml propidium iodide for dead or dying cells revealed a non-significant increase in the percentage of dead cells of the counted EGFP-mMetRSL274G-positive astrocyte population after ANL incubation. Numbers at the X-axis indicate the total number of EGFP-mMetRSL274G positive astrocytes counted (n = 5 independent experiments; represented data are mean +/- SEM; ONE-way ANOVA, p>0.05).
Fig 6.
Upregulation of candidate proteins upon BDNF treatment in astrocytes.
(A) Neuron-glia cocultures (DIV 22), infected with LVGFAPEGFP-mMetRSL274G, were incubated with 4 mM ANL or 4 mM Met (Met) for 4 h with (BDNF) or without 50 ng/ml BDNF (CNTR). Cell lysates were tagged with a biotin-alkyne tag via 'click-chemistry'. Biotin-tagged proteins were affinity purified with NeutrAvidin agarose. ANL incorporation and subsequent biotin tagging allows enrichment of de novo synthesized proteins compared to the Met incorporation control (Met Eluate). Overall protein levels of Cx43 and Rpl10a are comparable among the three different conditions in the lysates (Input) representing the joined pool of both pre-existing and de novo synthesized proteins as well as in the unbound fraction (Sup). In contrast, BDNF application resulted in increased levels of de novo synthesized and thus biotin-tagged Cx43 (43 kDa; both lower bands represent unphosphorylated and monophosphorylated Cx43) and Rpl10a (25 kDa) on immunoblots after purification (Eluate; 3 independent experiments). (B) Neuron-glia cocultures (DIV 22) were treated with 50 ng/ml BDNF for 4 h. A positive immunocytochemical GM130 staining for the Golgi apparatus (ICC) was used as a mask to quantify Cx43 signal intensity. No difference in Cx43 signal intensities was found. Numbers at the X-axis indicate the number of cells included in the quantification (3 independent experiments, represented data are mean +/- SEM, student's t-test, p>0.05). (C) Neuron-glia cocultures (DIV 22) were infected with LVGFAPEGFP and treated with 50 ng/ml BDNF for 4 h. Elevated Rpl10a positive signals were observed in GFAP-positive astrocytes by applying immunocytochemistry for Rpl10a and GFAP. Rpl10a signal intensities are color coded in the lower panel (scale bar = 10 μm). (D) Quantification of Rpl10a signal intensities, obtained as in (C), was done within the EGFP mask. A significant increase in Rpl10a signal intensity was observed. Numbers at the X-axis indicate the number of cells included in the quantification (5 independent experiments, represented data are mean +/- SEM, student's t-test, *: p<0.05). (E) Neuron-glia cocultures were infected with LVGFAPEGFP and treated with 50 ng/ml BDNF for 1 h. FISH procedure was conducted on fixed samples and revealed an increase of Cx43 mRNA-positive punctae in BDNF treated cells (scale bar = 10 μm). Images are depicted as 3D-reconstructions. FISH positive punctae were enlarged for representative purposes. (F) Quantification of Cx43 mRNA-positive punctae, obtained as described in (E), exhibited a significant increase in BDNF treated cultures. Numbers at the X-axis indicate the number of cells included in the quantification (3 independent experiments, represented data are mean +/- SEM, student's t-test, *: p<0.05).