Stem cell receptor degradation by niche cells restricts signalling

The stem cell niche utilizes short-range signalling, such that only stem cells but not their differentiating progeny receive self-renewing signals1. A cellular projection, the microtubule-based nanotube (MT-nanotube), projects from Drosophila male germline stem cells (GSCs) into niche “hub” cells, ensuring that GSC-produced receptor Thickveins (Tkv) receives sufficient hub-produced ligand Decapentaplegic (Dpp) while excluding non-stem cells from self-renewing signal activation. Here we show that GSC-produced Tkv is taken up by hub cells from the MT-nanotube and degraded in lysosomes. Failure of the hub cells to take up Tkv, or perturbation of hub cell lysosomal function lead to excess Tkv within GSCs, elevated downstream signal activation, and GSC tumor formation in non-niche locations. We propose that down-regulation of the self-renewal factor receptor by niche cells restricts the ligand/receptor interaction to the surface of the MT-nanotube membrane, fine-tuning the span of the short-range niche-stem cell signalling, and that this may be a general feature of contact-dependent signalling.

2 indicates division orientation). Ectopic expression of the niche ligands leads to tumorous proliferation of undifferentiated germ cells outside the niche 2 5 6 7 8 9 . Therefore, the range of the signal must be tightly regulated to activate signalling only within the cells that are juxtaposed to the hub, without erroneously activating the signal in differentiating cells. However, our understanding of the mechanisms that defines the range of the niche remains limited.
Previously, we demonstrated that MT-nanotubes form specifically at the hub-GSC interface and protrude from GSCs into hub cells. The niche ligand, Dpp, and its receptor, Tkv, interact at the nanotube-hub membrane contact sites, resulting in efficient and specific signal activation 10 ( Figure   1A, right). We reported that a Tkv-GFP fusion protein expressed in germline cells localizes to the MT-nanotubes, as they are trafficked along the MT-nanotubes. Accordingly, suppression of MTnanotube formation by knockdown of intraflagellar transport (IFT)-B (osm6, and oseg2) lead to accumulation of Tkv on the GSC plasma membrane 10 ( Figure S1A and B recapitulate the data shown in previous publication). In addition to the localization of Tkv to the MT-nanotube membrane, we have noted Tkv-GFP 'puncta' within the hub cells 10 . By inducing GSC clones that co-express Tkv-mCherry and GFP-Tub, similar puncta were observed in the hub, but only in the vincinity of the clone, indicating that the Tkv observed within hub originated in the GSC clone ( Figure 1C). The Tkv-GFP punctae in the hub often colocalize with lysosomes ( Figure 1B), leading us to hypothesize that Tkv-GFP protein is degraded in hub lysosomes. To confirm the lysosomal localization of Tkv, we suppressed lysosomal-dependent degradation using chloroquine, a drug that raises lysosomal pH and inhibits lysosomal enzymes. Treatment of isolated testes with chloroquine significantly increased the size of Tkv-GFP punctae within the hub ( Figure 1D, E, arrowheads, and G). Abnormal distribution of Tkv-GFP on the GSC plasma membrane was also often observed ( Figure 1F, arrow), confirming that Tkv-GFP is degraded in hub lysosomes, and suggesting that lysosome-mediated degradation of Tkv within the hub contributes to downregulation of Tkv in the GSC, thus preventing Tkv localization outside the MT-nanotubes. A similar co-localization with lysosomes was also observed for Dpp ligand punctae ( Figure 1H), indicating that the Tkv-GFP downregulation might occur after engaging with the Dpp ligand. Domeless-GFP, the receptor for the ligand Upd, did not localize to lysosomes; this lack of localization indicates the selectivity of this regulation ( Figure 1I).
To examine whether defective lysosomal degradation alters the signalling strength of the Dpp-Tkv pathway, we measured the level of phosphorylated Mad (pMad) in GSCs, a downstream target 10 . Genes involved in lysosomal-dependent degradation were knocked down using GAL4-UAS system. Hub-(and/or germ cell-) specific promoter-driven GAL4 drivers were used to determine the 3 cell type in which these genes are functioning. Spinster (Spin) is a putative lysosomal H +carbohydrate transporter and a common regulator of lysosomal biogenesis, as well as a known regulator of Dpp signalling 11 12 . Lysosomal-associated membrane protein 1 (Lamp1) is an abundant protein in the lysosomal membrane and is required for lysosomes to fuse with endosomes/autophagosomes 13 . Germ cell-specific knockdown of these lysosomal genes did not alter pMad level ( Figure 2C, E, and K). In contrast, hub-specific downregulation of these genes led to a significant increase in pMad level, indicating that lysosomal activity in the hub cells is required to dampen Dpp signalling ( Figure 2D, F, and K).
We tested the effect of knockdown of genes required for membrane vesicle trafficking, which is potentially utilized for Tkv-containing vesicle formation and sorting. We found that these genes are required to regulate pMad level in GSCs but not in the hub cells ( SMAD ubiquitination regulatory factor (Smurf)-mediated Tkv ubiquitination is necessary for GSC differentiation both in testicular and ovarian GSCs 3 18 . Smurf is a HECT (Homologous to the E6-AP Carboxyl Terminus) domain protein with E3 ubiquitin ligase activity, and disruption of Smurf function enhances Dpp-Tkv signal activation 18 . Ubiquitination of membrane proteins is required for recognition by ESCRT, and thus is required for endocytosis, lysosomal fusion, and degradation 19 . To determine if ubiquitination-defective Tkv no longer localizes to hub cells we generated a UAS-tkv-S238A-GFP transgenic line that carries a phosphorylation site mutation required for Smurf-mediated ubiquitination 3 . Tkv-S238A-GFP protein was not detected in hub cell lysosomes and exhibited strong retention within GSCs compared to wildtype Tkv-GFP protein ( Figure 3A, B), indicating that Tkv-4 S238A protein is not transferred from GSCs to hub cells. A small fraction of Tkv-GFP foci were still observed in the hub area, often contacting lysosomes ( Figure 3B', D, E) but not precisely colocalized ( Figure 3C, E), suggesting that the fusion step of receptor-containing vesicles with lysosomes is also impaired. Moreover, expression of Tkv-S238A-GFP in GSC resulted in elevated signal activation as indicated by higher pMad levels ( Figure 3F, G, H). Thus, Tkv protein is the target of lysosomal degradation in hub cells, allowing hub lysosomes to negatively regulate Dpp signalling.
We wanted to understand the consequence of failure to traffic Tkv to the MT-nanotubes. We previously demonstrated that the MT-nanotube is required for signal reception, thus interfering with MT-nanotube-formation by knockdown of IFT-B (oseg2, osm6, and che-13) reduced pMad levels in GSCs 10 . Here we found that knockdown of IFT-B frequently caused GSC tumors in non-niche locations where Tkv-GFP is also expressed ( Figure 4A-C). Under these conditions, the over-expressed Tkv localized to the GSC plasma membrane, likely because GSCs were unable to process Tkv via nanotube transport for hub-mediated degradation. The GSC-like tumour phenotype was caused by a defect within GSCs and their immediate progeny rather than a defect in later stage spermatogonia, since use of the bamGal4 driver (active in 4-to 8-cell stage spermatogonia), instead of the nosGal4 driver, to mediate IFT-B knockdown, and Tkv-GFP overexpression, did not cause tumour formation ( Figure 4C, Figure S3). These data strongly suggest that receptor protein degradation within niche cell lysosomes contributes to restriction of niche size by inhibiting stem cell proliferation outside of the niche.
Most signalling pathways consist of many components, whose activity must be closely coordinated. Here we provide evidence that signal activation and inhibition occur via the same structure: MT-nanotubes. MT-nanotubes transport Tkv receptors into the hub to interact with the Dpp ligand pool, promoting signal reception. They also guide Tkv receptor transport from the GSC membrane to niche (hub) cell lysosomes for destruction ( Figure S4). Lysosomal localization-defective Tkv, and MT-nanotube loss, both cause Tkv retention along the entire GSC cortex, suggesting that the Tkv degradation limits the location for ligand/receptor interaction to the MT-nanotube membrane, rather than the entire GSC membrane, preventing the receptor from entering differentiating daughter cells during cell division.
Our study demonstrates that lysosomal degradation of a receptor can occur in neighbouring cells, rather than the receptor-producing cell. Components mediating lysosomal function (Lamp1 and Spin) are required in neighbouring hub cells ( Figure S4). In contrast, components mediating vesicle trafficking (ESCRTIII and VPS-C) are required in GSC itself. MT-nanotube may shed its membrane directly into hub cells, the process being initiated by outward budding rather than inward budding of the plasma membrane ( Figure S4). Consistent with this model, enhancing the MT-nanotube formation by knocking down klp10A, or overexpressing Tubulin, in GSCs substantially increased the appearance of double-membrane vesicles in hub cells ( Figure 5A, 5B, and 5C), with 3D reconstitution indicating that these vesicles are not part of the MT-nanotubes, and instead were likely shed from MTnanotubes. If GSCs endocytosed the receptor the cytoplasmic domain of Tkv would be facing the outside of endosomes and the signal could be transduced until it was degraded by fusion with a lysosome 20 . In our model, such endocytosis would not occur in GSCs. In hub cells the cytoplasmic domain would be sequestered inside the vesicle. Stem cells could employ this mechanism to rapidly shut down the signal without taking the receptor back into its own cells. However, Tkv is ubiquitinated on its cytoplasmic domain 21 , raising the question of how vesicle recognition by the lysosome occurs.
Alternatively, GSCs may endocytose receptor to form a multivesicular body prior to GSC to hub transfer.
Our previous discovery of MT-nanotube-mediated signalling revealed that niche and stem cells interact in a contact-dependent manner, enabling highly specific cell-cell interactions. To ensure specificity, "contact-independent" signal activation should not occur. This study demonstrates that niche cells carry out stem cell receptor digestion to restrict receptor localization to MT-nanotubes, preventing "contact-independent" ligand-receptor interactions outside of the MT-nanotubes.
Cytonemes, another type of actin-dependent signalling protrusion 22 23 , transfer ligand and receptor, allowing the interaction to occur in a contact-dependent manner between the cells at a distance. Ligandproducing and receptor-producing cells both form cytonemes and both cells have been observed to take up signalling proteins: receptor into the ligand-producing cells and ligand into the receptorproducing cells 23           Synthesized fragments were annealed and digested by NotI and BglII. Resultant two inserts (TkvS238A and GFP) were ligated to modified pPGW vector using NotI and AscI sites in the multiple cloning site.
Briefly, testes were dissected in phosphate-buffered saline (PBS) and fixed in 4% formaldehyde in PBS for 30-60 minutes. Next, testes were washed in PBST (PBS +0.3% TritonX-100) for at least 30 minutes, followed by incubation with primary antibody in 3% bovine serum albumin (BSA) in PBST at 4 °C overnight. Samples were washed for 60 minutes (three times for 20minutes each) in PBST, incubated with secondary antibody in 3% BSA in PBST at 4 °C overnight, and then washed for 60 minutes (three times for 20 minutes each) in PBST. Samples were then mounted using VECTASHIELD with 4′,6diamidino-2-phenylindole (DAPI) (Vector Lab, H-1200).
Images were taken using a Zeiss LSM800 confocal microscope with a 63 ×oil immersion objective (NA=1.4) and processed using Image J and Adobe Photoshop software. Three-dimensional rendering was performed by Imaris software.

In situ hybridization
In situ hybridization on adult ovaries was performed mostly as described previously 6 .
Briefly, testes were dissected in 1XPBS and then fixed in 4% formaldehyde/PBS for 45 min. After rinsed 2 times with 1XPBS, then resuspended in 70% EtOH, left overnight at 4°C. The next day, testes were washed briefly in the wash buffer containing 2XSSC and 10% deionized formamide, then incubated overnight at 37°C in the dark with the 50nM of Quasar® 570 labeled stellaris probe against dpp mRNA (LGC Biosearch Technologies) in the Hybridization Buffer containing 2XSSC, 10% Dextran sulfate (MilliporeSigma), 1μg/μl of yeast tRNA (MilliporeSigma), 2mM Vanadyl ribonucleoside complex (NEB), 0.02% RNAse free BSA (ThermoFisher) and 10% of deionized formamide. On the 3 rd day, testes were washed 2 times for 30 min each at 37°C in the dark in the prewarmed wash buffer and then resuspended in a drop of VECTASHIELD with DAPI (Vector Lab, H-1200).

Chloroquine or Lysotracker/LysoSensor treatment
Testes from newly eclosed flies were dissected into Schneider's Drosophila medium containing 10% fetal bovine serum. Then testes were incubated at room temperature with or without 100M Chloroquine in 1mL media for 4 hours prior to imaging. For the lysosome staining, testes were incubated with 50nM of LysoTracker Deep Red (ThermoFisher L12492) or 100nM of LysoSensor Green DND-189 (ThermoFisher L7535) probes in 1mL media for 10 min at room temperature then briefly rinsed with 1mL of media for 3 times prior to imaging. For Tkv-mCherry clonal expression, hs-flp; nos-FRTstop-FRT-gal4, UAS-GFP 7 with UAS-Tkv-mCherry, UAS-GFP-Tubulin flies were heat-shocked at 37°C. for 30 min. Testes were dissected 24 hour after heat shock.
These testes were placed onto Gold Seal™ Rite-On™ Micro Slides two etched rings with media, then covered with coverslips. An inverted Zeiss LSM800 confocal microscope with a 63 ×oil immersion objective (NA=1.4) was used for imaging.