During astrocyte maturation, LELs establish a spatial degradative gradient, with degradative activity concentrating in the soma and branch LEL motility becoming progressively attenuated.

(A) Maximum-intensity projections of z-stacks acquired of GFP-LC3 transgenic astrocytes labeled with DQ-Red-BSA and BSA-647 at DIV3 or DIV7-8 of coculture. To visualize LELs in astrocytes, the inverted grayscale GFP-LC3 image was merged with the pseudocolored DQ-Red-BSA or BSA-647 image. Image of entire astrocyte; scale bar, 20 µm. Boxed regions correspond to straightened segments of the astrocyte branch shown below; scale bar, 5 µm. Images within each channel are displayed with identical minimum and maximum intensity settings across DIV. Filled orange arrowheads denote BSA-647 puncta that are co-positive for DQ-Red-BSA. Open orange arrowheads denote BSA-647 puncta that are negative for DQ-Red-BSA. (B) Schematic of probes used in this study to label the spectrum of organelles in the endolysosomal system. (C-D) Corresponding quantification of (A) to measure (C) the total area occupied by DQ-Red-BSA-positive puncta normalized to branch area and (D) the percentage of total BSA-647-positive puncta area that is co-positive for DQ-Red-BSA. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model; N=25-26 astrocytes from 3 biological replicates. (E) Representative kymographs of SiR-Lysosome (SiR-Lyso) and LysoTracker Red (LysoT) motility along a primary-to-secondary astrocytic branch from GFP-LC3 transgenic astrocytes in coculture. SiR-Lysosome puncta that were manually tracked (shown in periwinkle) are overlaid onto the LysoTracker Red kymographs. Open orange arrowheads denote LysoTracker Red tracks that are negative for SiR-Lysosome. All quantification of LEL dynamics is derived from kymograph analysis. Horizontal bar, 5 µm; vertical bar, 30 sec. (F-G) Corresponding quantification of (E) to measure (F) SiR-Lysosome track density normalized to a 50 µm branch segment and (G) percentage of LysoTracker Red tracks that are co-positive for SiR-Lysosome. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model; N=33-34 branches (one branch per astrocyte) across 4 biological replicates. (H) Schematic depicting the spatial distribution of degradative activity in cocultured astrocytes at DIV3 versus DIV7-8. (I) Corresponding quantification of (E) to measure directionality of SiR-Lysosome motility at DIV3 and DIV7-8 of coculture. Shown are means ± SD; N=20-23 branches (one branch per astrocyte) across 3 biological replicates. (J-K) Corresponding quantification of (E) to measure (J) percentage of stationary (or immobile) tracks and (K) mean speed of SiR-Lysosome tracks. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model; N=30-33 branches (one branch per astrocyte) across 4 biological replicates. Data in E-G and I-K are derived from the DMSO control data at each timepoint shown in Figure S4A-D. (L-N) (L) Representative single-plane images of a non-transgenic astrocytic branch transduced with lentiviral eGFP and LAMP1-mCherry (both driven under a shortened GFAP promoter), cocultured for DIV7-8. Scale bar, 20 µm. (M) Corresponding kymographs of LAMP1-mCherry and SiR-Lysosome motility along the astrocyte branch that is boxed in (L). Closed blue arrowheads denote SiR-Lysosome tracks that are positive for LAMP1-mCherry. Open blue arrowheads denote SiR-Lysosome tracks that are negative for LAMP1-mCherry. Open pink arrowheads denote LAMP1-mCherry tracks that are negative for SiR-Lysosome. Horizontal bar, 5 µm; vertical bar, 30 sec. (N) Corresponding quantification of the percentage of SiR-Lysosome tracks that are co-positive for LAMP1-mCherry. N=1,267 tracks from 33 branches (one branch per astrocyte) across 5 biological replicates. Data in L-N are derived from the DMSO control data shown in Figure S4E-H. For all superplots in the paper, small dots indicate the measurements from individual cells (e.g., the technical replicates) from each of the independent experiments; large shapes (e.g. triangles and squares) indicate the corresponding mean from each independent experiment (e.g., the biological replicates).

TRPML1 activation dampens LEL motility in astrocytic branches.

(A) Representative single-plane image of a non-transgenic DIV7 cocultured astrocyte transduced with lentiviral GCaMP6s-TRPML1 (expressed under a shortened GFAP promoter) and co-labeled SiR-Lysosome; scale bar, 20 µm. Box denotes region of zoom-in of GCaMP6s-TRPML1 and SiR-Lysosome; scale bar, 10 µm. LELs co-positive for GCaMP6s-TRPML1 and SiR-Lysosome are denoted with closed arrowheads. (B) Schematic of GCaMP6s-TRPML1 fusion protein; TRPML1 forms a homotetrameric channel on LEL membranes (82). Noted are pharmacological tools used to modulate TRPML1 activity. (C-E) (C) Representative kymographs of GCaMP6s-TRPML1 and SiR-Lysosome along the astrocyte branch; GCaMP6s-TRPML1 was used as a space-fill to delineate the branch boundary. Arrows indicate the time of on-scope addition of DMSO (solvent control), ML-SA1 (60 µM), or co-addition of ML-SA1 (60 µM) and ML-SI3 (30 µM). Prior to imaging, astrocytes were pretreated for 1.5 hours with DMSO (paired with the on-scope addition of DMSO or ML-SA1) or ML-SI3 (30 µM; paired with the on-scope co-addition of ML-SA1 and ML-SI3); SiR-Lysosome was added during the final 30 minutes of pretreatment. Given that GCaMP6s-TRPML1 overexpression alone dampens LEL motility, imaging was performed at coculture DIV5, when baseline motility is higher and affords a greater dynamic range for detecting motility changes. Horizontal bar, 5 µm; vertical bar, 30 sec. (D) Quantification of the mean intensity of GCaMP6s-TRPML1 along the astrocyte branch, normalized to the average mean intensity before on-scope addition. Shown are means ± SEM; N=9-12 branches (one branch per astrocyte) across 3-4 biological replicates. (E) Quantification of SiR-Lysosome mean speed based on kymograph analysis. Small, transparent paired points indicate the average SiR-Lysosome mean speed per branch pre-versus post-treatment; large, opaque paired points indicate the biological replicate mean pre-versus post-treatment. Shown are p-values from a mixed-effects model with Tukey’s post-hoc correction applied to biological replicate values; N=9-12 branches (one branch per astrocyte) across 3-4 biological replicates. (F-I) (F) Kymographs of LysoTracker Red motility along a primary-to-secondary astrocytic branch; the GFP-LC3 transgene was used as a space-fill to delineate the branch boundary. Cocultures were treated for 30 minutes with DMSO (solvent control), ML-SA1 (60 µM), or ML-SI3 (30 µM) prior to live-cell imaging. LysoTracker Red puncta that were manually tracked (shown in periwinkle) are overlaid onto the kymographs. Horizontal bar, 5 µm; vertical bar, 30 sec. All quantification is derived from the kymograph analysis: (G) mean speed of LysoTracker Red tracks, (H) cumulative frequency of LysoTracker Red mean speed (derived from the technical replicates in G), and (I) percentage of stationary (or immobile) tracks. (G, I) Small dots indicate measurements from individual branches from individual astrocytes (e.g., technical replicates) from each of the independent experiments; larger triangles indicate the corresponding mean from each independent experiment (e.g., the biological replicates). Horizontal bars represent the mean of biological replicates ± SD. Shown are p-values from an LME model; N=20-21 branches (one branch per astrocyte) across 3 biological replicates; DIV7-8 of coculture.

TRPML1 knockdown increases lysosome motility and antagonizes the ML-SA1-induced arrest of LEL motility along astrocyte branches.

(A) Schematic of the miR30-shRNA system used to knockdown TRPML1 expression in cocultured astrocytes. shRNA sequences are embedded within a miR30 backbone to enable Drosha/Dicer-dependent processing from a Pol II-driven transcript, permitting astrocyte-specific knockdown when paired with a GFAP promoter. (B-E) (B) Representative maximum-intensity projections of z-stacks of DIV7 cocultured astrocytes co-expressing lentiviral eGFP (identifies astrocytes transduced with shRNAmiR) and mCherry-TRPML1 under a shortened GFAP promoter. The shRNA constructs included either a non-targeting shRNA control (denoted as NT) or a pool of two individual shRNAs targeting TRPML1 (denoted as pooled; individual shRNAs are denoted ML1-1 and ML1-2). mCherry-TRPML1 images are displayed with identical minimum and maximum intensity settings across treatment conditions; scale bar, 10 µm. Box indicates location of zoom-in of mCherry-TRPML1 in the knockdown condition. Zoom-in is adjusted individually to emphasize the sphericity of residual mCherry-TRPML1 puncta, consistent with lipid accumulation observed in MLIV (48, 49); scale bar, 10 µm. (C) Quantification of mCherry-TRPML1 puncta mean intensity. (D) Quantification of the percentage of total area occupied by mCherry-TRPML1 normalized to soma area. (E) Quantification of the sphericity of mCherry-TRPML1 puncta. (C-E) Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model; N=29-35 astrocytes across 3 biological replicates; DIV7-10 of coculture. (F-I) (F) Kymographs of SiR-Lysosome motility along primary-to-secondary astrocytic branches; eGFP (shRNAmiR) was used as a space-fill to delineate the branch boundary. Cocultures transduced with either the non-targeting control or the pool of two individual TRPML1-targeting shRNAs were treated for 30 minutes with DMSO (solvent control) or ML-SA1 (60 µM) prior to imaging. SiR-Lysosome puncta that were manually tracked (shown in periwinkle) are overlaid onto kymographs. Horizontal bar, 5 µm; vertical bar, 30 sec. All quantification is derived from kymograph analysis: (G) mean speed of SiR-Lysosome tracks, (H) cumulative frequency of SiR-Lysosome mean speed (derived from the technical replicates in G), and (I) percentage of stationary (or immobile) tracks. (G, I) Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model with Holm’s correction for multiple comparisons; N=34-45 branches (one branch per astrocyte) across 5-6 biological replicates; DIV6-8 of coculture.

TRPML1-induced LEL motility arrest is dependent on F-actin and myosin-Va.

(A) Schematic depicting the proposed model for TRPML1-induced LEL arrest. Motile LELs exhibit predominately bidirectional trajectories. Tethering to F-actin via myosin-Va may reduce LEL motility. Perturbations to F-actin or myosin-Va activity are hypothesized to disrupt LEL tethering and increase LEL motility. (B) Representative maximum-intensity projections of z-stacks acquired during live-cell imaging of non-transgenic cocultured astrocytes transduced with lentiviral LifeAct-mNeonGreen (LifeAct-mNG; expressed under a shortened GFAP promoter) at DIV7. Cocultures were treated for 30 minutes with DMSO (solvent control) or Latrunculin A (LatA; 5 µM). Cyan box denotes the area shown in the zoom-in. Orange arrowheads denote hot spots or patches of actin along the shaft of the branch, in filopodia, and in small lamellar structures. Scale bars, 20 µm (main) and 10 µm (zoom-in). (C-F) (C) Kymographs of SiR-Lysosome motility along astrocytic branches; the GFP-LC3 transgene was used as a space-fill to delineate the branch boundary. Cocultures were treated for 30 minutes with DMSO (solvent control), ML-SA1 (60 µM), LatA (5 µM), or a co-treatment of ML-SA1 and LatA. SiR-Lysosome puncta that were manually tracked (shown in periwinkle) are overlaid onto kymographs. Horizontal bar, 5 µm; vertical bar, 30 sec. All quantification is derived from kymograph analysis: (D) mean speed of SiR-Lysosome tracks, (E) cumulative frequency of SiR-Lysosome mean speed (derived from the technical replicates in D), and (F) percentage of stationary (or immobile) tracks. (D, F) Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model with Holm’s correction for multiple comparisons; N=34-36 branches (one branch per astrocyte) across 5 biological replicates; DIV7-8 of coculture. (G-J) (G) Kymographs of SiR-Lysosome motility along astrocytic branches expressing lentiviral eGFP or myosin-Va-DN (C-terminal tail domain fused to the C-terminus of eGFP) under a shortened GFAP promoter. Cocultures were treated for 30 minutes with DMSO (solvent control) or ML-SA1 (60 µM). SiR-Lysosome puncta that were manually tracked (shown in periwinkle) are overlaid onto kymographs. Horizontal bar, 5 µm; vertical bar, 30 sec. All quantification is derived from kymograph analysis: (H) mean speed of SiR-Lysosome tracks, (I) cumulative frequency of SiR-Lysosome mean speed (derived from the technical replicates in H), and (J) percentage of stationary (or immobile) tracks. (H, J) Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model with Holm’s correction for multiple comparisons; N=19-29 branches (one branch per astrocyte) across 3-4 biological replicates; DIV6-7 of coculture.

TRPML1 activity modulates the phosphorylation of PAP-enriched actin-membrane linkers ezrin-radixin-moesin.

(A) Maximum-intensity projections of z-stacks acquired of DIV7 non-transgenic cocultured astrocytes transduced with lentiviral LifeAct-mNeonGreen (mNG) to visualize F-actin. Cocultures were immunostained for p-ERM proteins (phospho-ezrin Thr567, phospho-radixin Thr564, and phospho-moesin Thr558) or the plasma membrane marker GLAST/EAAT1. Scale bar, 20 µm. Boxed region denotes location of zoom-ins. (Ai-ii) Corresponding zoom-ins of astrocytic branches. White arrows denote F-actin-enriched foci co-positive for PAP-enriched markers. Scale bar, 10 µm. (B) Straightened branches from maximum-intensity projections of z-stacks acquired of DIV8 non-transgenic cocultured astrocytes transduced with lentiviral LifeAct-mNeonGreen (mNG) to visualize F-actin. Cocultures were immunostained for PAP-enriched markers p-ERM or GLAST and LEL-associated LAMP1. Orange arrowheads denote LAMP1-positive organelles positioned proximal to F-actin-enriched foci and PAP markers. Scale bar, 5 µm. (C-Di) (C) Straightened branches from maximum-intensity projections of z-stacks acquired of DIV7-8 non-transgenic cocultured astrocytes identified by immunostain for S100β and co-labeled for p-ERM. Cocultures were treated with DMSO (solvent control) or ML-SA1 (60 µM) for 5 minutes. S100β and p-ERM channels are each displayed with identical minimum and maximum intensity settings across treatment conditions. Scale bar, 10 µm. (D-Di) Corresponding quantification of the total area occupied by p-ERM-positive structures within the (D) proximal or (Di) distal 50 µm segment of the branch; values were normalized to the mean of the DMSO control per biological replicate. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model; N=34 branches (one branch per astrocyte) across 4 biological replicates. (E-Fi) (E) Straightened branches from maximum-intensity projections of z-stacks acquired of DIV7-8 non-transgenic cocultured astrocytes identified by immunostain for S100β and co-labeled for p-ERM. Cocultures were treated with DMSO (solvent control) or ML-SI3 (30 µM) for 5 minutes. S100β and p-ERM channels are each displayed with identical minimum and maximum intensity settings across treatment conditions. Scale bar, 10 µm. (F-Fi) Corresponding quantification of the total area occupied by p-ERM-positive structures within the (F) proximal or (Fi) distal 50 µm segment of the branch; values were normalized to the mean of the DMSO control per biological replicate. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model; N=28-31 branches (one branch per astrocyte) across 4 biological replicates. (G-H) (G) Immunoblot analysis of lysates from non-transgenic monocultured astrocytes (DIV6-7) treated with DMSO (solvent control), ML-SA1 (60 µM), or ML-SI3 (30 µM) for 5 minutes. Samples are immunoblotted for total ERM proteins (appear as doublets) and GAPDH (loading control). Shown are three independent biological replicates (denoted as n=1-3). (H) Corresponding densitometric analysis of total ERM levels normalized to GAPDH, and reported as a fold change relative to the DMSO control. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a one-sample t-test; N=3 biological replicates.

Model for TRPML1-dependent coordination of LEL positioning and regulation of the PAP cytoskeleton in astrocytic branches.

LELs establish a spatial degradative gradient in astrocytic branches, with degradative activity concentrated in the soma and proximal branches. We have identified two key functions for lysosomal TRPML1. (1) TRPML1 activity promotes LEL motility arrest through a mechanism dependent on F-actin and myosin-Va. LEL tethering is coupled to glutamatergic signaling, suggesting that neuronal cues could influence LEL positioning within astrocytic branches. (2) TRPML1 signaling rapidly modulates ERM phosphorylation within PAPs. Changes in ERM phosphorylation may in turn influence PAP structural dynamics at the synapse.

A cortical murine neuron-astrocyte coculture system to study stellate astrocytes in vitro.

(A) Schematic of the protocol for neuron-astrocyte coculture. Cortical glia enriched for astrocytes are isolated from postnatal day 0-1 (P0-1) pups. To distinguish astrocytes by live-cell imaging, cortical glia are often isolated from GFP-LC3 transgenic mice (for subsequent coculture with non-transgenic neurons). The day prior to neuron isolation, acid-washed coverslips are coated with poly-L-lysine. Cortical neurons are isolated at embryonic day 15.5, plated onto PLL-coated coverslips, and maintained as a monoculture for 4-5 days in vitro (DIV), at which point cortical glia are plated onto the neuronal meshwork (designated as DIV0 of coculture). One to two days after addition of glia, cytosine arabinoside (AraC; 2 µM) is added to the coculture to restrict astrocyte proliferation. Cocultures are maintained for a total of 6-10 days (neurons are a total of DIV11-14). (B) Quantification of the purity of the neuron monoculture. Shown is the percentage of nuclei in non-transgenic cortical neuronal monocultures at DIV8 that are co-positive for markers of neurons (β3-Tubulin to label neuronal microtubules, MAP2 to label dendrites), astrocytes (GFAP), oligodendrocytes (O4), or microglia (Iba1). A total of 4,198 cells were counted across 2 biological replicates. (C) Representative maximum-intensity projections of z-stacks of the neuron-astrocyte coculture at DIV7 of coculture. Non-transgenic neurons (total age of DIV11) were cocultured with GFP-LC3 transgenic glia and immunostained for β3-Tubulin (neuron marker), GFP (to visualize LC3-expressing glia), and GLAST to confirm astrocyte identity. Scale bar, 20 µm. (D-E) Quantification of the identity of GFP-LC3-expressing glia in coculture with neurons. (D) Representative maximum-intensity projections of z-stacks of the neuron-astrocyte coculture at DIV7 of coculture. Cocultures of non-transgenic neurons and GFP-LC3 transgenic glia were immunostained for GFP (to label the GFP-LC3 transgene only present in glia) and a panel of astrocyte markers: p-ERM (Ezrin [Thr567]/Radixin [Thr564]/Moesin [Thr558]), S100β (S100 calcium-binding protein B), GLAST (excitatory amino acid transporter 1, EAAT1), AQP4 (aquaporin-4), and GFAP (glial fibrillary acidic protein). Scale bar, 40 µm. (E) Percentage of GFP-LC3 transgenic glia (cocultured for DIV7) co-positive for markers of astrocytes (p-ERM, S100β, GLAST, AQP4, GFAP), oligodendrocyte lineages (A2B5, O4), or microglia (Iba1). A total of 903 cells were counted across 2 biological replicates.

Validation of endolysosomal markers across astrocyte branch maturation.

(A) Uncropped single channel and merge of images shown in Fig. 1A to visualize LEL signal in neighboring neurons; scale bar 20 µm. (B) The graph in B corresponds to the experiment and primary data in Figure 1A, C-D. Quantification of the total area occupied by BSA-647-positive puncta normalized to branch area; GFP-LC3 was used as a space-fill to delineate the branch boundary. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model; N=25-26 branches (one branch per astrocyte) across 3 biological replicates. (C-E) (C) Maximum-intensity projections of live-cell z-stacks from monoculture GFP-LC3 transgenic astrocytes at DIV6 labeled with SiR-Lysosome. Astrocytes were treated for 3 hours with DMSO (solvent control) or a protease inhibitor cocktail (PI; 30 µM Pepstatin-A and 30 µM E64D). Images within each channel are displayed with identical minimum and maximum intensity settings across treatment conditions. Scale bar, 40 µm. (D) Corresponding quantification of the total area occupied by SiR-Lysosome-positive puncta normalized to cell area. (E) SiR-Lysosome puncta maximum intensity; values are normalized to the mean of the DMSO-treated group within each replicate. (D-E) N=40-65 astrocytes from 2 biological replicates. (F-G) Graphs in F-G are paired with the analysis in Figure 1E-G, and correspond to the experiment and primary data for the DMSO controls at each time point derived from the dataset in Figure S4A-D. (F) The density of LysoTracker Red tracks normalized to a 50 µm branch segment of coculture astrocytes; the GFP-LC3 transgene was used as a space-fill to delineate the branch boundary. (G) Percentage of SiR-Lysosome tracks that are co-positive for LysoTracker Red. (F-G) Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model; N=33-34 branches (one branch per astrocyte) across 4 biological replicates.

Microtubule polarity becomes more mixed as astrocyte branches develop in coculture with neurons.

(A) Schematic of a dynamic microtubule. To track the directionality of polymerizing microtubules in astrocyte branches, we transduced non-transgenic cocultures with EB3-mNeonGreen (mNG) under a truncated GFAP promoter. EB3 (End-binding protein 3) specifically binds the GTP-tubulin-rich cap at the growing plus ends of microtubules. EB3 rapidly dissociates upon GTP hydrolysis during catastrophe/depolymerization. Thus, EB3-NG produces a comet-like fluorescent signal that tracks the plus-end of polymerizing microtubules. (B) Representative single-plane images of non-transgenic astrocytes expressing lentiviral EB3-mNeonGreen. Scale bar, 20 µm. (C-I) (C) Representative kymographs of EB3-mNeonGreen dynamics along a primary-to-secondary branch trajectory in astrocytes at coculture DIV3-4 versus DIV7-8. At each time point, top panels are raw kymographs; EB3 comets that were manually tracked (anterograde comets shown in magenta and retrograde comets shown in green) are overlaid onto the respective kymographs and shown in the bottom panels. All quantification is derived from the kymograph analysis. Horizontal bar, 5 µm; vertical bar, 30 sec. (D) Percentage of anterograde EB3-mNeonGreen tracks. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model; N=27 astrocytes across 4 biological replicates; each astrocyte value represents the mean of two branches. (E) Corresponding probability density histogram for data in D. Solid curves show kernel density estimates. N=53-54 individual branches (not averaged per astrocyte) across 4 biological replicates. (F) Percentage of retrograde EB3-mNeonGreen tracks. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model; N=27 astrocytes across 4 biological replicates; each astrocyte value represents the mean of two branches. (G) Corresponding probability density histogram with kernel density estimates for data in F. N=53-54 individual branches (not averaged per astrocyte) across 4 biological replicates. (H) Early electron microscopy studies reported that nascent astrocyte processes are densely populated with microtubules (162, 163). As astrocytes mature, however, these processes become enriched in intermediate filaments and comparatively depleted of microtubules (162, 163). Similarly, we observed a decrease in EB3 tracks in mature astrocyte branches. H shows the density of EB3-mNeonGreen tracks normalized to a 50 µm branch segment. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model; N=27 astrocytes across 4 biological replicates; each astrocyte value represents the mean of two branches. (I) Corresponding probability density histogram for data in H. N=53-54 individual branches (not averaged per astrocyte) across 4 biological replicates.

Lower concentrations of TRPML1 pharmacological modulators are sufficient to alter LEL motility across different stages of astrocyte branch maturation.

(A) Representative kymographs of SiR-Lysosome motility along astrocytic branches at DIV3 versus DIV7-8 of coculture; the GFP-LC3 transgene was used as a space-fill to delineate the branch boundary. Cocultures were treated for 2 hours with DMSO (solvent control) or ML-SA1 (20 µM). SiR-Lysosome puncta that were manually tracked (shown in periwinkle) are overlaid onto their respective kymographs. Horizontal bar, 5 µm; vertical bar, 30 sec. All quantification is derived from kymograph analysis: (B) mean SiR-Lysosome track speed (µm/sec), (C) cumulative frequency of SiR-Lysosome mean speed (derived from the technical replicates in B), and (D) percentage of stationary (or immobile) tracks. (B, D) Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model with Holm’s correction for multiple comparisons; N=29-33 branches (one branch per astrocyte) across 4 biological replicates. Data from the DMSO condition in Figure S4A-D are also analyzed in Figure 1E-G and I-K. (E-H) (E) Representative kymographs of SiR-Lysosome and lentiviral LAMP1-mCherry (expressed under a truncated GFAP promoter) motility along astrocytic branches cocultured for DIV7-8; lentiviral eGFP (co-expressed under a truncated GFAP promoter) was used as a space-fill to delineate the branch boundary. Cocultures were treated for 2 hours with DMSO (solvent control), ML-SA1 (20 µM), or ML-SI3 (10 µM). SiR-Lysosome puncta that were manually tracked (shown in periwinkle) are overlaid onto their respective kymographs. Horizontal bar, 5 µm; vertical bar, 30 sec. All quantification is derived from kymograph analysis: (F) mean SiR-Lysosome track speed (µm/sec), (G) cumulative frequency of SiR-Lysosome mean speed (derived from the technical replicates in F), and (H) percentage of stationary (or immobile) tracks. (F, H) Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a LME model; N=24-32 branches (one branch per astrocyte) across 4-5 biological replicates. Data from the DMSO condition in Figure S4E-H are also analyzed in Figure 1L-N.

Neurons form mature synapses and astrocytes form elaborate PAP structures in coculture.

(A) Straightened branches from a single plane of non-transgenic cocultures at DIV3 versus DIV7; astrocytes were identified by immunostain for GFAP. Cocultures were co-stained for Bassoon to label presynaptic compartments and Homer to label postsynaptic compartments. Images within each channel are displayed with identical minimum and maximum intensity settings across DIV. Filled orange arrowheads denote bona fide synapses that are co-positive for both pre- and postsynaptic markers, consistent with synaptic contacts adjacent to the astrocyte branch. Scale bar, 5 µm. Corresponding quantification of total area occupied by (B) Bassoon puncta or (C) Homer puncta. Data are normalized to a 50 µm branch segment and expressed as a fold change relative to the mean of the DIV3 corresponding biological replicate. (B-C) Horizontal bars represent the mean of biological replicates ± SD; N=17-20 branches (one branch per astrocyte) across 2 biological replicates. (D-Di) (D) Non-transgenic coculture at DIV7 (neurons are a total of DIV11) loaded with Fluo-4 AM and treated with 50 µM 4-AP and 50 µM bicuculline added via on-scope delivery between 29-30 seconds of live cell imaging, followed by on-scope addition of 50 µM CNQX and 50 µM AP5 added between 89-90 seconds. Trace represents the fluorescence mean intensity measured of the neuronal network, normalized to the average value for mean intensity of frames prior to spike in (0-29 seconds). (Di) Images corresponding to frames 1 and 39 from the trace in (D). (E-H) (E) Brightfield image of a neuron at DIV10 of coculture (total neuron age of DIV14) used for patch-clamp recording; soma is outlined by a black dashed contour. (F) Fired action potentials upon current injection (−50 to +130 pA in 20 pA steps). Under current-clamp mode, the baseline membrane potential was held at −70 mV. (G) Firing rates plotted against current injection (n=11 neurons at DIV10 of coculture). (H) All neurons exhibit sIPSCs (upper row: 7 out of 7 neurons recorded; holding potential = −10 mV) and sEPSCs (lower row: 14 out of 14 neurons recorded; holding potential = −70 mV) under voltage-clamp mode. (I-K) (I) Representative maximum-intensity projections of z-stacks of cocultured non-transgenic neurons and GFP-LC3 transgenic astrocytes at DIV3 versus DIV7, immunostained for GFP(LC3) and p-ERM. Orange arrowheads denote fan-like lamellate structures. Scale bar, 20 µm. (J) Straightened branches from maximum-intensity projections of z-stacks of cocultured non-transgenic neurons and GFP-LC3 transgenic astrocytes at DIV3 versus DIV7. Cocultures were immunostained for GFP (LC3; labels astrocytes), p-ERM, and the postsynaptic marker PSD-95. Images within each channel are displayed with identical minimum and maximum intensity settings across time points. Scale bar, 5 µm. (Ji) Higher-magnification view of PSD-95 with display settings adjusted independently to visualize low-level signal at DIV3 of coculture. (Jii) Higher-magnification view of p-ERM and PSD-95 highlighting regions of close apposition between the two signals (filled orange arrowheads). Scale bar, 5 µm. (K) Corresponding quantification of total area occupied by p-ERM-positive structures per unit branch length and expressed as a fold change relative to the mean of the DIV3 corresponding biological replicate. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from an LME model; N=29 branches (one branch per astrocyte) across 3 biological replicates.

Glutamatergic signaling dampens LEL motility in astrocytic branches.

(A) Kymographs of SiR-Lysosome motility along astrocytic branches at DIV7 of coculture; the GFP-LC3 transgene was used as a space-fill to delineate branch boundaries of astrocytes in coculture with non-transgenic neurons. Cocultures were treated for 30 minutes with DMSO (solvent control) or CNQX/AP5 (50 µM/50µM). SiR-Lysosome puncta that were manually tracked (shown in periwinkle) are overlaid onto their respective kymographs. Horizontal bar, 5 µm; vertical bar, 30 sec. All quantification is derived from kymograph analysis: (B) mean SiR-Lysosome track speed (µm/sec), (C) cumulative frequency of SiR-Lysosome mean speed (derived from the technical replicates in B), (D) percentage of stationary (or immobile) tracks. (B, D) Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from an LME model; N=25-26 branches (one branch per astrocyte) across 3 biological replicates.

Modulating TRPML1 activity regulates ERM phosphorylation in astrocytes.

(A) Immunoblot analysis of lysates from non-transgenic monocultured astrocytes (DIV7-9) treated with DMSO (solvent control) or ML-SA1 (60 µM) for 5 minutes. Corresponding densitometric analysis of (B) p-ERM or (C) total ERM levels, normalized to a GAPDH loading control and reported as a fold change relative to the DMSO control. Note p-ERM and ERM proteins appear as doublets by immunoblot. (D) Ratio of p-ERM/GAPDH to total ERM/GAPDH, reported as a fold change relative to the DMSO control. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a one-sample t-test; N=3 biological replicates. (E-H) Immunoblot analysis of lysates from non-transgenic monocultured astrocytes (DIV4-10) treated with DMSO or MK6-83 (30 µM) for 5 minutes. Corresponding densitometric analysis of (F) p-ERM or (G) total ERM levels, normalized to a GAPDH loading control and reported as a fold change relative to the DMSO control. (H) Ratio of p-ERM/GAPDH to total ERM/GAPDH, reported as a fold change relative to the DMSO control. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from a one-sample t-test; N=3 biological replicates. (I-K) (I) Straightened branches from maximum-intensity projections of z-stacks acquired of non-transgenic cocultured astrocytes at DIV7-8; astrocytes were identified by immunostain for GFAP (or S100β, not shown) and counter stained with p-ERM. Cocultures were treated with DMSO (solvent control) for 120 minutes or ML-SI3 (30 µM) for 5, 30, 60, or 120 minutes. p-ERM signal is displayed with identical minimum and maximum intensity settings across treatment conditions. Scale bar, 5 µm. Corresponding quantification of the total area occupied by p-ERM-positive structures within the (J) proximal or (K) distal 50 µm branch segment, normalized to the DMSO control mean of the corresponding biological replicate. Horizontal bars represent the mean of biological replicates ± SD; shown are p-values from an LME model; N=21-32 branches (one branch per astrocyte) across 3-4 biological replicates.