Introduction

Lysosomes are dynamic organelles primarily associated with the degradation of long-lived proteins, lipids, and cytoplasmic organelles delivered from the endocytic and autophagic pathways and contains ~60 types of acidic hydrolases1,2. The majority of these hydrolases are modified with mannose 6-phosphate (M6P) residues, ensuring their recognition by M6P receptors in the Golgi complex and proper trafficking to the lysosome3,4. While two M6P receptors, cation-dependent M6P receptor (CD-M6P receptor) and insulin-like growth factor-II/cation-independent mannose 6-phosphate receptor (IGF-II/M6P receptor), have been identified, IGF-II/M6P receptor is the primary receptor responsible for transporting these acidic hydrolases to lysosomes, as it has a higher affinity for M6P-tagged cargo than does CD-M6P receptor3,5,6. Upon delivery to the endosome, the IGF-II/M6P receptor-hydrolase complexes dissociate due to the acidic pH, and hydrolases are released to the lumen, whereas unoccupied IGF-II/M6P receptors are retrieved to the trans-Golgi network (TGN) via the retrograde trafficking pathways to initiate further rounds of enzyme delivery. Defects in IGF-II/M6P receptor trafficking itinerary lead to the inappropriate sorting and delivery of hydrolases and, therefore, impair the degradative capacity of lysosomes3.

Although the mechanism underlying the retrograde trafficking pathways is still ambiguous, several different protein complexes have been implicated in this process, including AP17, the retromer8,9, phosphofurin acidic cluster sorting protein 1 (PACS-1)10. PACS-1, an adapter that directs the movement of proteins containing acidic sorting motifs within the molecules such as furin and IGF-II/M6P receptor between endosomes and the TGN10,11. These acidic residues often contain serine or threonine residues that can be phosphorylated by casein kinase 2 (CK2)12,13,14 and the regulation of sorting of these from the endosomes to the TGN by CK2 is also implicated12,14.

S1P is a phosphorylated product of sphingosine catalyzed by sphingosine kinase 1 (SphK1) or sphingosine kinase 2 (SphK2) depending on cell types and functions and has emerged as a potent lipid mediator with diverse effects on multiple biological processes including cell growth, survival, differentiation, motility, cytoskeletal organization, neurotransmitter release, endosome maturation and surface transport carrier formation at the TGN15,16,17,18,19. Most of these processes are mediated by five S1P-specific G-protein-coupled receptors (S1P1-5Rs) that show distinct expression in tissues and cells, and also unique G-protein-coupling patterns suggesting distinctive cellular functions20,21,22.

We have recently demonstrated that continuous ongoing activation of S1P receptors in the cellular organelles through an intracrine system, i.e., an end product of sphingolipid catabolism, S1P is transported into the luminal side of organelles and activates S1P receptors in situ which keeps transmitting G-protein signal necessary for vesicular trafficking and maturation. For instance, in the regulation of maturation of multivesicular endosomes SphK2/S1P1R and/or S1P3R axis plays an important role for switching the lysosome/exosome maturation: inhibition of the S1P signal resulted in the formation of exosomal cargo-depleted exosomes18. More recently, continuous activation of S1P receptors at the TGN is crucial for the surface transport carrier formation: continuous activation of S1PR-coupled Gi/Go proteins enables constant supply of the free Gβγ at the TGN, which is a prerequisite for the transport carrier formation and trafficking from the TGN to the plasma membranes (PM)19. These observations have led to the hypothesis that a similar S1P signaling paradigm may govern the regulation of trafficking of proteins critical for lysosomal functions such as IGF-II/M6P receptor. Several lines of evidence indicate that the human IGF-II/M6P receptor has two distinct signaling functions that regulate the activity of Gi type 2 in response to the binding of IGF-II or M6P23 and IGF-II/M6P receptor stimulated with IGF-II utilizes the transactivation of S1PR for G protein-dependent extracellular signal-regulated kinases 1 and 224, however, the regulation of intracellular trafficking of IGF-II/M6P receptor, particularly the retrograde transport, i.e., from the endosomes to the TGN is largely unclear. Recently, we have found that S1P signaling is required for the maintenance of basal lysosomal proteolytic activity as assessed by cellular cathepsin D activity. In the present study, we have analyzed the molecular mechanisms underlying the retrograde trafficking of IGF-II/M6P receptor from the endosome to the TGN and uncovered for the first time that S1P/S1PR/Gq signaling is vital for IGF-II/M6P receptor retrograde trafficking by regulating CK2-mediated phosphorylation of PACS-1, a connector molecule to AP1 required for the formation of transport carriers toward the TGN. The physiological importance of S1P signaling in the maintenance of lysosome function in terms of vesicular trafficking of IGF-II/M6P receptor trafficking and pathogenesis of lysosomal insufficiency due to sphingolipid metabolism dysfunction is discussed herein.

Results

Requirement of receptor-mediated S1P signal for the maintenance of basal lysosomal cathepsin D activity

Cathepsin D activity was measured in cell lysates prepared from HeLa cells, which are known to express endogenous S1P1R and S1P3R25, maintained under basal conditions using a synthetic peptide substrate for the enzyme. The enzyme activity was significantly reduced by a SphK inhibitor, HACPT or S1P1/3R blocker, VPC23019 as compared with vehicle control. S1P1R-selective blocker, W146 showed little or no significant effect (Fig. 1A). Considering the ineffectiveness of W146, the effect of VPC23019 may mainly be attributable to S1P3R, suggesting that basal lysosomal cathepsin D activity may be regulated by a S1P3R-mediated S1P signal. Among several reasons we will address later, as for the S1PRs, we focused on S1P3R in the following studies. The importance of SphK1/S1P3R axis in the regulation of cathepsin D activity was further supported by the experiments using siRNA-mediated inhibition of the respective gene expression. As for SphKs, SphK1 downregulation caused a profound decrease in the cathepsin D activity, while SphK2 knockdown had little or no significant effect (Fig. 1B). S1P3R knockdown had a profound inhibitory effect. Knockdown of VPS35, a vital element of the retromer complex necessary for the retrograde transport of IGF-II/M6P receptor trafficking9 and whose disfunction is reported to cause an aberrant maturation of cathepsin D26, was also included as a positive control. These results suggest that S1P3R mediated S1P signal plays an important role in the regulation of basal cathepsin D activity with an extent comparable to VPS35 although the precise molecular mechanism needs to be explored. Consistent with the lower cathepsin D activity (Fig. 1A, B), siRNA-mediated interference either by SphK1 or S1P3R siRNA showed an increase in immature pro-cathepsin D secretion into the culture media compared with that of control siRNA treatment to an extent same as VPS35 siRNA used as positive control26 (Fig. 1C, D). Corresponding to the increased secretion of immature pro-cathepsin D in the culture media, there was a clear decrease in mature cathepsin D in the cell lysates from the cells in which SphK1, S1P3R or VPS35 was downregulated by siRNA-mediated interference, suggesting that maturation of cathepsin D may be disturbed during vesicular trafficking processes where SphK1/ S1P3R axis was impaired.

Fig. 1: Requirement of S1P signal for basal cathepsin D activity.
figure 1

HeLa cells maintained under standard conditions were treated with 5 µM HACPT, 5 µM W146, 5 µM VPC23019 or DMSO as a vehicle control for 24 h in a serum-depleted medium (A). The cells were also transfected with control, SphK1-, SphK2-, S1P3R- or VPS35-siRNA and cultured for 48 h (B). The cells were lysed in 1% Triton X-100 containing buffer and subjected to cathepsin D in vitro assay. The mean activities were expressed as % of control on scatter-dot plots (**p < 0.01; *p < 0.05 versus vehicle treatment in (A), versus control siRNA in (B); NS not significant; Welch’s t-test). Bars represent the mean ± s.e.m. C HeLa cells transfected with control, SphK1-, S1P3R-, or VPS35-siRNA were cultured for 48 h in DMEM, and for the last 3 h, cells were incubated in serum-free DMEM containing 100 µg/mL cycloheximide at 37 °C. The culture media were collected and treated with TCA. The precipitated proteins were washed with ice-cold acetone and subjected to SDS-PAGE followed by immunoblot with anti-cathepsin D antibody (culture media). Recombinant GST protein (5 µg) was added before TCA precipitation as protein input and detected by anti-GST antibody. Cells were lysed by 1% Triton X-100-containing buffer, cleared by centrifugation and subjected to SDS-PAGE (cell lysates). Three experiments were conducted, and representative images are shown. D Quantitative analysis of pro-cathepsin D in culture media and mature cathepsin D in cell lysate (**p < 0.01; *p < 0.05 versus control siRNA; Welch’s t-test, n = 3).

Full size image

Regulation of endogenous IGF-II/M6P receptor distribution by an S1P-activated S1PR signal

Given that an S1P3R-mediated S1P signal regulates cathepsin D activity under basal conditions, it may be reasonable to assume that cellular distribution of IGF-II/M6P receptor, a major receptor for lysosomal hydrolases, may change by manipulating S1P signal. Next experiments were performed to confirm the involvement of S1P signal in the regulation of IGF-II/M6P receptor distribution. Under basal conditions majority of endogenous IGF-II/M6P receptor-positive structures were co-stained with a TGN marker, golgin-97 (Fig. 2A) as previously reported3. When cells were treated with a SphK inhibitor, HACPT, the extent of co-localization of IGF-II/M6P receptor with the TGN was apparently decreased. Similarly, treatment of cells with S1P1/3R blocker VPC23019 caused a reduction in the co-localization compared with vehicle treatment. S1P1R blocker W146 had no clear effects. Quantitative analysis showed that both HACPT and VPC23019 treatment reduced the TGN localization of IGF-II/M6P receptor as judged by Pearson’s coefficient (Fig. 2B). Knockdown experiment confirmed that SphK1 and S1P3R, but not SphK2 or S1P1R, were involved in the regulation of IGF-II/M6P receptor trafficking (Fig. 2C, Supplementary Fig. 2). When VPC23019-treated cells were stained with anti-Rab5, Rab7 or Rab11 antibody, some IGF-II/M6P receptor-positive endosomes outside the TGN were co-localized clearly with Rab5, but only slightly with Rab11 and hardly with Rab7 (Fig. 2D), suggesting the possibility that IGF-II/M6P receptor trafficking was disturbed in Rab5-positive early endosomes or their intermediate structures during maturation. These results suggest that SphK1/S1P3R signaling axis may be involved in the trafficking of IGF-II/M6P receptor among endosomes, the TGN and presumably the PM.

Fig. 2: Requirement of S1P signal for TGN localization of endogenous IGF-II/M6P receptor.
figure 2

A HeLa cells were incubated with 5 µM HACPT, 5 µM W146, 5 µM VPC23019 or DMSO as a vehicle in the presence of cycloheximide containing standard medium for 3 h and fixed for immunofluorescence study with anti-IGF-II/M6P receptor and anti-golgin-97 antibodies. Cells were analyzed by confocal microscopy. Cell boundaries are indicated as dotted lines. B The 30 cell images were subjected for Pearson’s coefficient analysis. Results were expressed as the mean ± s.e.m. (**p < 0.01 versus vehicle treatment; Welch’s t-test) on scatter-dot plots. C HeLa cells were transfected with SphK1-, SphK2-, S1P1R-, S1P3R- or VPS35-siRNA and fixed, stained and analyzed as in (B). Microscopic data were shown in Supplementary Fig. 2. Bars represent the mean ± s.e.m. (*p < 0.05; **p < 0.01 versus control siRNA; Welch’s t-test). D HeLa cells were incubated with 5 µM VPC23019 in the presence of cycloheximide containing standard medium for 3 h and fixed for immunofluorescence study with anti-IGF-II/M6P receptor, anti-Golgin97 and anti-Rab5, anti-Rab7 or anti-Rab11 antibodies. Note that some IGF-II/M6P receptor-positive endosomes outside TGN were co-localized clearly with Rab5 (see arrows), but only slightly with Rab11 and hardly with Rab7. The cell boundaries are indicated as dotted lines. Scale bars, 10 µm.

Full size image

Regulation of the retrograde trafficking of IGF-II/M6P receptor by an S1P-activated S1P3R signal

We next investigated the retrograde (from the endosomes to the TGN) trafficking of IGF-II/M6P receptor. In these experiments, we used a CD8-IGF-II/M6P chimeric protein, which is constructed with the ectodomain of CD8, which will be exposed on the cell surface and can be recognized by the extracellular CD8 antibody, followed by the transmembrane and cytoplasmic domains of M6P receptor, which regulate the endocytosis and targeting of these chimeras to intracellular compartments, including the TGN27. HeLa cells stably expressing CD8-IGF-II/M6P receptor were incubated with antibodies against CD8 on ice for 30 min. After wash, CD8-IGF-II/M6P receptor internalization was then chased at 37 °C every 5 min for up to 15 min, fixed, and immuno-labeled with antibody against golgin-97, a TGN marker, or Rab5, an early endosome marker. As revealed by confocal microscopy and colocalization analysis, the majority of antibodies bound to CD8-IGF-II/M6P receptor were internalized following endocytic pathways and delivered to the golgin-97-positive TGN compartment at 15 min post-chase (Supplementary Fig. 3A, B). When cells were pretreated with HACPT, internalization of the chimeric receptor was not influenced but retrograde transport of CD8-IGF-II/M6P receptor to the TGN was inhibited and the receptor was retained in the dispersed vesicular structures at 15 min post-chase (Fig. 3A, B). Similarly, treatment of the cells with VPC23019, a S1P1/3R blocker suppressed the receptor transport to the TGN potently. These results suggest that S1P3R may regulate the retrograde transport of CD8-IGF-II/M6P receptor to the TGN. To get further information on the molecular mechanism underlying the regulation of the endosomes to the TGN trafficking of CD8-IGF-II/M6P receptor, the role of individual molecules involved in the signaling axis was analyzed by siRNA-mediated gene silencing techniques. For SphKs, SphK1 knockdown had a significant inhibitory effect whereas SphK2 depletion showed no clear changes as compared with that of control siRNA (Fig. 3C, D). Downregulation of S1P3R showed significant inhibitory effects on the CD8-IGF-II/M6P receptor retrograde transport to the TGN as compared with the control siRNA-treated cells (Fig. 3D). S1P1R siRNA did not show any significant inhibition of CD8-IGF-II/M6P receptor retrograde transport (Supplementary Fig. 4). These results strengthen the notion that S1P3R -mediated S1P signal regulates the retrograde transport of CD8-IGF-II/M6P receptor to the TGN. Interestingly, reduced retrograde transport of CD8-IGF-II/M6P receptor to the TGN induced by downregulation by RNA interference (Fig. 3C, D) was correlated with the significantly enhanced co-distribution of the chimeric receptor with Rab5-positive early endosomal compartments at 15 min post-chase (Fig. 3E, F). In contrast, the colocalization of CD8-IGF-II/M6P receptor and Rab7 or Rab11 was not increased, actually decreased, by SphK1 or S1P3R knockdown (Supplementary Fig. 5A–D), reinforcing our hypothesis that SphK/S1P axis is working at retrograde transport from the Rab5-positive compartments to the TGN. These results suggest that the receptor-mediated S1P signal might regulate transport carrier formation at the early endosomes destined to the TGN and the inhibition of this signal results in the retention of retrograde transport cargoes such as CD8-IGF-II/M6P receptor in the early endosome compartments. To emphasize further the importance of agonist-dependent activation of the S1P3R in the retrograde transport of CD8-IGF-II/M6P receptor, we attempted to rescue the inhibition of the retrograde transport of CD8-IGF-II/M6P receptor seen in the gene-targeting downregulation of SphK1 (Fig. 3D) by an siRNA-resistant SphK1 expression. Importantly, only wild type SphK1, SphK1(WT), but not catalytically inactive mutant, SphK1(KD) could rescue the inhibition of the retrograde transport (Supplementary Fig. 6), suggesting the importance of S1P production and the ligand-dependent activation of S1P3R in this phenomenon. To prove the hypothesis that the endocytosed antibody-labeled CD8-IGF-II/M6P receptor follows the retrograde trafficking to the TGN under the regulation by the receptor-mediated S1P signal, it is important to show that individual key proteins essential to exert the signal reside in the early endosomal compartments. To show this, HeLa cells stably expressing both CD8-IGF-II/M6P receptor and either transiently expressing GFP-SphK1 or S1P3R-GFP were treated with antibodies against CD8 on ice, chased at 37 °C for 5 min to maximize the co-localization with Rab5 (see Supplementary Fig. 3), fixed, immuno-labeled with antibodies against Rab5 and analyzed by confocal microscopy. Indeed, an S1P-producing enzyme, SphK1, necessary for the maintenance of basal cathepsin D activity (Fig. 1) and for the retrograde transport of CD8-IGF-II/M6P receptor (Fig. 3C, D), co-localized at least partly with CD8-IGF-II/M6P receptor and Rab5-positive early endosomal compartments (Supplementary Fig. 7A). Similarly, S1P3R partly co-distributed with CD8-IGF-II/M6P receptor and Rab5-positive early endosomal compartments. It is noteworthy that in the same incubation time (5 min), some SphK1 and CD8-IGF-II/M6P receptor-positive and S1P3R and CD8-IGF-II/M6P receptor-positive endosomes are also co-localized with VPS35 (Supplementary Fig 7B). These results strengthen the hypothesis that S1P3R-mediated S1P signal can regulate the retrograde transport of CD8-IGF-II/M6P receptor presumably with the aid of retromer complex to the TGN.

Fig. 3: S1P signaling contribution for retrograde transport of CD8-IGF-II/M6P receptor.
figure 3

HeLa cells stably expressing CD8-IGF-II/M6P receptor were incubated with 5 µM HACPT, 5 µM VPC23019 or DMSO as a vehicle in serum-depleted medium for 1 h (A, B) or transfected with control, SphK1-, SphK2-, or S1P3R-siRNA (C–F). After transfection, cells were cultured for 48 h at 37 °C and then incubated with anti-CD8α antibody (5 μg/mL) on ice for 30 min in DMEM without serum or bicarbonate. After two times wash with the same medium without the antibody, the cells were incubated for 15 min at 37 °C in serum- and bicarbonate-free medium. Cells were fixed, permeabilized, stained with an antibody against golgin-97, a TGN resident marker or an antibody against Rab5, an early endosome marker, and analyzed by confocal microscopy. Note that SphK1- or S1P3R-knockdown decreased the co-localization between CD8-IGF-II/M6P receptor and Golgin97 (C, D) while increased that between CD8-IGF-II/M6P receptor and Rab5 which were clearly seen in the inset of (E). The 30 cell images each were subjected for Pearson’s coefficient analysis (B, D, F). Results were expressed as the mean on scatter-dot plots (**p < 0.01 versus vehicle treatment in (B), versus control siRNA in (D) and (F); Welch’s t-test). Bars represent the mean ± s.e.m. The cell boundaries are indicated as dotted lines. Scale bars, 10 µm.

Full size image

Regulation of the retrograde trafficking of cholera toxin subunit B by an S1P-activated S1P3R signal

To ask whether the regulation of retrograde transport of IGF-II/M6P receptor by an S1P3R-mediated S1P signal is a general mechanism for the retrograde trafficking or limited to this hydrolase receptor, a possibility for regulation of the retrograde transport of a well-known bacterial toxin, cholera toxin subunit B (CtxB) by this S1P signal was addressed next. CtxB binding to its ganglioside receptor GM1 initiates retrograde trafficking from the cell surface through endosomes and the TGN into the endoplasmic reticulum28. Twenty-five minutes after the addition of fluorescence-conjugated CtxB to HeLa cells, the majority of CtxB fluorescence distributed in the golgin-97-positive structures in the control siRNA transfected cells (Fig. 4A, B). Intriguingly, SphK1- but not SphK2 knockdown showed that CtxB-positive structures were distributed in small punctate structures around the cell periphery. S1P3R-downregulation caused a similar effect, suggesting that the retrograde transport of CtxB to the TGN is also regulated by S1P3R-mediated S1P signal and that the retrograde transport of cargoes from endosomes to the TGN may be regulated by a S1P3R-mediated S1P signal in general.

Fig. 4: S1P signaling contribution for retrograde transport of CTxB to the TGN.
figure 4

HeLa cells were transfected with control, SphK1-, SphK2- or S1P3R-siRNAs and cultured under standard conditions for 48 h at 37 °C. The cells were then treated with 1 µg/mL CTxB conjugated with Alexa 555 on ice in DMEM in the absence of serum and bicarbonate for 30 min. After wash, the cells were incubated for 25 min at 37 °C in serum- and bicarbonate-free medium. After fix and permeabilization, the cells were stained by anti-golgin-97 antibody and analyzed by confocal microscopy (A). The cell boundaries are indicated as dotted lines. Scale bars, 10 µm. In each experiment, 30 cell images were subjected to Pearson’s coefficient quantification (B). Results were expressed as the mean on scatter-dot plots (**p < 0.01 versus control siRNA; Welch’s t-test). Bars represent the mean ± s.e.m.

Full size image

Regulation of the retrograde trafficking of IGF-II/M6P receptor by an S1P3R/Gq signal

Next experiments were performed to map the downstream signal after S1P3R activation. Since S1P3R is coupled with various G proteins, e.g., Gi/o, G12/13 and Gq depending on cells and tissues20, several G protein inhibitors or toxins were used to identify the G protein subtype involved in the retrograde transport of CD8-IGF-II/M6P receptor to the TGN. Cathepsin D activities were measured in cell lysates from HeLa cells pretreated with various inhibitors or pertussis toxin (PTX) using a synthetic peptide substrate for the enzyme. Treatment of cells with PTX, which is known to ADP-ribosylate the Giα subunit and to make Gi-coupled receptors refractory to agonists29, showed no clear changes as compared with that of vehicle-treated cells. VPC23019 results were also added as a positive control (Fig. 5A). Importantly, treatment of cells with YM-254890, a Gq-selective inhibitor known to suppress GDP/GTP exchange of Gqα subunit30, significantly suppressed the protease activities. Treatment with Gallein, a Gβγ subunit inhibitor, showed only slight inhibition. Suppose that these lysosomal protease activities are correlated with the amounts of IGF-II/M6P receptor trafficked from the endosome to the TGN, the Gq inhibitor should have some effects on the retrograde transport of IGF-II/M6P receptor, a receptor for many acid hydrolases. Expectedly, consistent with the result of cathepsin D activity (Fig. 5A), YM-254890 treatment caused an inhibition of the retrograde transport of CD8-IGF-II/M6P receptor to the TGN and resulted in a retention of the receptor in the dispersed vesicular structures at 15 min post-chase, which is comparable with the results of cells treated with VPC23019, a S1P1/3R blocker (Fig. 5B). A quantitative colocalization analysis also supports the effectiveness the inhibitor on the retrograde transport (Fig. 5C). To get more substantial information about the involvement of Gq in the regulation of the retrograde transport of IGF-II/M6P receptor, a dominant negative mutant Gq(Q209L/D277N), which lacks GTP binding capacity, were transiently overexpressed and observed the retrograde transport of CD8-IGF-II/M6P receptor to the TGN. Importantly, Gq(Q209L/D277N) mutant expression caused an inhibition in the retrograde transport of CD8-IGF-II/M6P receptor (Fig. 5D, E), suggesting the involvement of Gq signal in this phenomenon.

Fig. 5: Involvement of Gq in the retrograde transport of CD8-IGF-II/M6P receptor to the TGN.
figure 5

A HeLa cells were treated with 100 ng/mL PTX, 10 µM Gallein, 10 µM YM-254890, 5 µM VPC23019 or DMSO as a vehicle in serum-depleted DMEM for 24 h (in the case of PTX for overnight) at 37 °C were lysed in 1% Triton X-100 containing Tris buffer and subjected for cathepsin D activity assay. The mean activities were expressed as % of control and on scatter-dot plots (**p < 0.01; *p < 0.05 versus vehicle treatment; NS not significant versus vehicle treatment; Welch’s t-test). Bars represent mean ± s.e.m. B HeLa cells treated with various inhibitors were fixed, permeabilized, stained with an antibody against golgin-97 and analyzed by confocal microscopy. The cell boundaries are indicated as dotted lines. Scale bars, 10 µm. C Thirty cell images each were subjected to Pearson’s coefficient analysis (**p < 0.01 versus vehicle treatment; NS not significant versus vehicle treatment; Welch’s t-test). D HeLa cells stably expressing CD8-IGF-II/M6P receptor were transfected with cDNA encoding mCherry and wild type Gq (Gq WT) or dominant negative Gq (Gq(Q209L/D277N)) at a ratio of 1:3 and cultured for 24 h and incubated with anti-CD8α antibody on ice for 30 min in DMEM without serum or bicarbonate. After washing twice with the same medium, cells were incubated for 15 min at 37 °C in serum- and bicarbonate-free medium. Cells were fixed, permeabilized, stained with an antibody against golgin-97 and analyzed by confocal microscopy. The cell boundaries are indicated as dotted lines. Scale bars, 10 µm (D). The 30 cell images each were subjected for Pearson’s coefficient analysis. The cells expressing mCherry were considered to be transfected with Gq(WT) or Gq(Q209L/D277N) (E). Results were expressed as the mean on scatter-dot plots (**p < 0.01 versus Gq WT; Welch’s t-test). Bars represent the mean ± s.e.m.

Full size image

Regulation of the retrograde trafficking of IGF-II/M6P receptor by a Gq/CK2/PACS1 signal axis

It has been reported that some of the cargo proteins such as IGF-II/M6P receptor require additional adapter proteins such as PACS1 for the retrograde trafficking to the TGN11. To assess whether Gq is involved in the regulation of PACS1-mediated retrograde transport of IGF-II/M6P receptor to the TGN, first we tested the ability of Gq to regulate the association of PACS1 with CD8-IGF-II/M6P receptor in a pull-down assay in the lysates from HeLa cells co-expressing FLAG-PACS1 and CD8-IGF-II/M6P receptor. When FLAG-PACS1 was collected by anti-FLAG antibody-coated beads CD8-IGF-II/M6P receptor was efficiently pulled-down (Fig. 6A), showing a decent association of both proteins in consistent with a previous report11. Importantly, in the cell lysates from cells treated with a Gq inhibitor, YM-254890 resulted in strong reduction of PACS1 association with CD8-IGF-II/M6P receptor. Since the retrograde transport of CD8-IGF-II/M6P receptor is regulated by both Gq and S1P3R (Fig. 5C), it is reasonable to surmise that S1P/S1P3R/Gq axis regulates PACS1 association with CD8-IGF-II/M6P receptor. To get more insight into the molecular mechanism underlying how Gq regulates PACS1 association with CD8-IGF-II/M6P receptor, the capacity of PACS1 association with this chimeric protein was compared in the presence or absence of a constitutively active Gq mutant, Gq(Q209L). In the absence of Gq(Q209L), VPC23029 treatment of the cells resulted in a robust inhibition of PACS1 association with CD8-IGF-II/M6P receptor under vehicle control (Fig. 6B). As S1P3R is known to couple with Gq, the result indicates that Gq is activated in coupled with S1P-dependent S1P3R activation. Since PACS1 was identified from yeast two-hybrid screening for the cytoplasmic proteins that have an ability to bind to the CK2-phosphorylated acidic cluster-based motifs on the furin cytosolic domain10, involvement of CK2 in the regulation of PACS1 association with CD8-IGF-II/M6P receptor was assessed next. Treatment of cells with apigenin, a selective inhibitor for CK2, caused a strong inhibition of CD8-IGF-II/M6P receptor association with PACS1 to an extent similar to VPC23019 treatment (Fig. 6B). In the presence of influenza hemagglutinin (HA)-Gq(Q209L) CD8-IGF-II/M6P receptor was efficiently pulled-down with anti-FLAG antibody-coated beads with a level comparable with that in an endogenous Gq (Fig. 6C Vehicle, compare with Fig. 6B Vehicle). With a striking contrast, however, VPC23019 treatment became ineffective in the presence of HA-Gq(Q209L) (Fig. 6C). These results suggest that under the physiological conditions activation of Gq occurs by an S1P-dependent S1P3R activation-coupled mechanism in endosomes and the constitutively active Gq mutant bypasses the S1P receptor activation, which makes the receptor blocker, VPC23019, insensitive. Remarkably, treatment of cells with apigenin almost overcame the effect of the constitutively active Gq mutant and strongly inhibited the PACS1 association with CD8-IGF-II/M6P receptor. Knockdown of S1P3R or CK2α, a catalytic subunit of CK2, showed almost similar effects as VPC23019 or apigenin, respectively (Fig. 6D, E). Furthermore, CK2α-siRNA treatment caused the inhibition of CD8-IGF-II/M6P receptor transport to the TGN (Fig. 6F) as well as maturation of cathepsin D (Fig. 6G), implying that as downstream of the Gq signal CK2-mediated phosphorylation of PACS1 may play an important role in the regulation of PACS1 association with IGF-II/M6P receptor, thereby regulate retrograde transport of IGF-II/M6P receptor to the TGN presumably with the aid of other proteins such as AP1 and retromer complex.

Fig. 6: Gq signal downstream of S1P3R activation regulates PACS1 association with CD8-IGF-II/M6P receptor by CK2-catalyzed phosphorylation.
figure 6

A HeLa cells stably expressing CD8-IGF-II/M6P receptor were transfected with cDNA encoding FLAG-PACS1 and cultured for 48 h under standard conditions and treated without (DMSO) or with 10 µM YM-254890 in serum-depleted medium for 1 h, lysed with lysis buffer containing protease inhibitor cocktail and immunoprecipitated by anti-FLAG beads, followed by immunoblotting analysis using ant-IGF-II/M6P receptor. HeLa cells stably expressing CD8-IGF-II/M6P receptor were transfected with FLAG-PACS1 in the absence (B) or presence (C) of constitutively active HA-Gq(Q209L), cultured for 48 h and treated with 5 µM VPC23019 or 20 µM apigenin in serum-depleted medium for 1 h. Cells were lysed and subjected to pull-down assay as in (A) and detected by anti-FLAG, anti-IGF-II/M6P receptor. The expression of HA-Gq(Q209L) was detected by anti-HA antibody in the lysates. HeLa cells stably expressing CD8-IGF-II/M6P receptor were transfected with control, S1P3R- or CK2α-siRNA and FLAG-PACS1 in the absence (D) or presence (E) of constitutively active HA-Gq(Q209L) and cultured for 48 h. Cells were lysed and subjected to pull-down assay as in (B, C) and detected by anti-FLAG or anti-IGF-II/M6P receptor antibody. The expression of HA-Gq(Q209L) was detected by anti-HA antibody in the lysates. The data presented are a typical representative of three independent experiments. F HeLa cells stably expressing CD8-IGF-II/M6P receptor were transfected with control or CK2α-siRNA. Cells were cultured for 48 h at 37 °C and retrograde transport of CD8-IGF-II/M6P receptor was assayed as Fig. 3C, D (**p < 0.01 versus control siRNA; Welch’s t-test). G HeLa cells transfected with control or CK2α-siRNA were cultured for 48 h in DMEM, and for the last 3 h, cells were incubated in serum-free DMEM containing 100 µg/mL cycloheximide at 37 °C. Pro-cathepsin D in culture media and mature cathepsin in cell lysates were detected by immunoblotting as in Fig. 1C. Cropped images from the experiment in Fig. 1C were shown. Uncropped immunoblots were shown in the Supplementary Figure file. Three experiments were conducted, and the representative images are shown.

Full size image

Regulation of the CK2-mediated phosphorylation of PACS1 by a S1P3R/Gq/CK2 signal axis

To obtain substantial evidence that S1P3R/Gq/CK2 axis regulates retrograde transport of IGF-II/M6P receptor to the TGN through CK2-mediated phosphorylation of PACS1, FLAG-PACS1 was immunoprecipitated from HeLa cells transiently expressing this fusion protein and its phosphorylation was assessed by using Zn2+-phos-tag-biotin after SDS-PAGE and transferred to PVDF membrane31. The phos-tag probe detected phosphorylation signal on the FLAG-PACS1 as verified by complete disappearance of the bands after treatment with alkaline phosphatase for 30 min (Fig. 7A). The phosphorylation of PACS1 was reduced by CK2α knockdown and enhanced by GFP-CK2α overexpression, consistent with the previous report showing the involvement of CK2 in the phosphorylation of PACS111 (Fig. 7A). The phosphorylation level of FLAG-PACS1(S278A), where phosphorylatable serine residue 278 by CK2 was mutated to alanine, was 67 ± 9.6% (mean ± s.e.m.) compared with FLAG-PACS1(WT) (p < 0.05, n = 3), which has a similar tendency with a previous report using 32P autoradiography method32 (Fig. 7B). The phosphorylation of FLAG-PACS1 was significantly reduced by apigenin, YM-254890, HACPT or VPC23019 (73.9 ± 3.9%, 80.3 ± 3.1%, 79.0 ± 2.6% or 75.0 ± 3.0% compared to Vehicle control, respectively, mean ± s.e.m., p < 0.01 for all, n = 3), confirming the involvement of Gq, SphK and S1P3R for the PACS1 phosphorylation (Fig. 7C). Importantly, VPC23019 had almost no inhibitory effect on FLAG-PACS1 phosphorylation, whereas apigenin still did, in Gq(Q209L)-expressing cells, which was consistent with the co-immunoprecipitation assay in Fig. 6B, C (Fig. 7D). Lastly, VPC23019-induced inhibition of CD8-IGF-II/M6P receptor retrograde transport was fully rescued by phosphorylation-mimicking mutant PACS1(S278D) but not by PACS1(WT) (Fig. 7E). These data strongly suggest that SphK1/S1P3R/Gq/CK2 signal axis is critical for the phosphorylation of PACS1, which facilitates the association of PACS1 and IGF-II/M6P receptor necessary for trafficking them to the TGN.

Fig. 7: Regulation of the CK2-mediated phosphorylation of PACS1 by an S1P3R/Gq/CK2 signal axis.
figure 7

A HeLa cells were transfected with FLAG-PACS1 only (left two lanes), FLAG-PACS1 and control- or CK2α-siRNA (middle two lanes), or FLAG-PACS1 and GFP or GFP-CK2α (right two lanes). After incubation for 48 h cells were lysed and FLAG-PACS1 was immunoprecipitated by anti-FLAG beads. In an experiment, immunoprecipitants were treated with alkaline phosphatase for 30 min (indicated as Phosphatase). The phosphorylation of FLAG-PACS1 was detected by Zn2+-phos-tag-biotin and total FLAG-PACS1 was detected by anti-FLAG antibody. B FLAG-PACS1 or FLAG-PACS1(S278A) expressed in HeLa cells was immunoprecipitated and the phosphorylation level was analyzed as in (A). C HeLa cells transfected with FLAG-PACS1 were treated with 20 µM apigenin, 10 µM YM-254890, 5 µM HACPT, 5 µM VPC23019 or DMSO as vehicle for 1 h and the phosphorylation of FLAG-PACS1 was analyzed as in (A). D HeLa cells were transfected with FLAG-PACS1 in the absence or presence of constitutively active HA-Gq(Q209L), cultured for 48 h and treated with 5 µM VPC23019 or 20 µM apigenin in serum-depleted medium for 1 h. Cells were lysed and the phosphorylation of FLAG-PACS1 was analyzed as in (A). The expression of HA-Gq(Q209L) was detected by anti-HA antibody in the lysates. E HeLa cells stably expressing CD8-IGF-II/M6P receptor were transfected with cDNA encoding mCherry and wild type FLAG-PACS1(PACS1(WT)) or phospho-mimicking mutant of PACS1 (PACS1(S278D)) at a ratio of 1:3 and cultured for 48 h and incubated with 5 µM VPC23019 or DMSO as vehicle in serum-depleted medium for 1 h, then retrograde transport of CD8-IGF-II/M6P receptor was assayed as in Fig. 3B. Results were expressed as the mean on scatter-dot plots (**p < 0.01; Welch’s t-test). Bars represent the mean ± s.e.m. The cell images and the expression level of FLAG-PACS1(WT) and FLAG-PACS1(S278D) were shown in Supplementary Fig. 9.

Full size image

Discussion

IGF-II/M6P receptor follows a highly regulated sorting itinerary to deliver hydrolases between the endo/lysosomes and the TGN (Fig. 8). In the present study, we have shown that an S1P3R-mediated S1P signal regulates lysosomal cathepsin D activity (Fig. 1) and endogenous IGF-II/M6P receptor distribution under basal conditions (Fig. 2). To dissect signaling pathway we used CD8-IGF-II/M6P receptor as a tool to study retrograde transport of IGF-II/M6P receptor from the endosomes to the TGN. We have found that S1P/S1P3R/Gq/CK2 signaling axis plays a central role to induce PACS1-mediated retrograde transport of IGF-II/M6P receptor to the TGN (Figs. 3, 5, 6, 7). In this phenomenon SphK1 but not SphK2 is involved (Fig. 3). It is interesting to note that SphK2 is involved in the maturation of late endosomes, i.e., SphK2/S1P1/3R/Gi axis regulates exosomal cargo sorting into intraluminal vesicles of multivesicular endosomes18. In the case of surface transport carrier formation at the TGN, SphK1/S1P1/2R axis at the TGN keeps transmitting G-protein signals including the Gβγ subunit supply enables transport carrier formation at the TGN destined for the PM19. Although SphK1 and SphK2 catalyze the same reaction, i.e., formation of S1P, these isozymes show different intracellular distribution and less functional redundancy. We have also shown that treatment of cells with Gq-selective inhibitor YM-254890 or CK2 inhibitor apigenin resulted in reduction of PACS1 phosphorylation and its association with CD8-IGF-II/M6P receptor and that the effect of constitutive active mutant of Gq(Q209L) on phosphorylation of PACS1 and its association with CD8-IGF-II/M6P receptor was insensitive to S1P1/3R blocker, VPC23019, but canceled by CK2 inhibitor, apigenin (Figs. 6 and 7). Although CK2 was assumed to be a constitutively active enzyme and had a passive role in signaling33, it is also reported that Wnt3a signal-stimulated Gqɑ activated CK234. It is plausible to surmise that S1P3R-coupled Gqɑ may activate CK2 directly or indirectly to phosphorylate PACS1, which facilitates its association with IGF-II/M6P receptor for the retrograde transport to the TGN. Recent reports suggest that frequent somatic mutations in the heterotrimeric G protein ɑ-subunit, GNAQ, in ocular melanoma of the uvea. The mutations occur exclusively in codon 209, used in the present study (Figs. 6 and 7), turning GNAQ into a dominant-acting oncogene35. Further studies on S1P-mediated cascade leading to Gq signaling pathways may be important to reveal the molecular mechanism underlying oncogenesis of uvea melanoma.

Fig. 8: Current model for S1P signal-regulated retrograde transport of IGF-II/M6P receptor from the endosomes to the TGN.
figure 8

During maturation of endosomes membrane sphingolipids including sphingomyelin and glycosphingolipids are metabolized finally to S1P through multiple stepwise catabolic reactions via sphingosine39. Ongoing in situ activation of Gq-coupled S1P3R on endosomal membranes by an SphK1-involved intracrine system causes Gq subunit dissociation. Released Gqα subunit presumably activates CK2 directly or indirectly34. Active CK2 phosphorylates PACS1 which allows association with cargo-unloaded IGF-II/M6P receptor. This process may be necessary for the retrograde transport of IGF-II/M6P receptor with the aid of other proteins such as AP110,40 and retromer complex8,9. Sph sphingosine, CK2 casein kinase 2.

Full size image

The lysosomal storage diseases are a group of rare inherited metabolic disorders, which result from lysosomal dysfunction that stems from mutations or deficiency of lysosomal hydrolases or lipid-binding proteins. For example, Gaucher disease is caused by defects in the activity of the lysosomal glucocerebrosidase resulting in the accumulation of glucosylceramide in lysosomes, which in turn causes an impairment in digestion of other macromolecules such as sphingomyelin and results in lysosomal insufficiency36. The simple question as to why one lysosomal hydrolase defect triggers an impairment in the digestion of other macromolecules in the lysosomes remains unanswered. Lysosomal glycosphingolipids are mainly catabolized by stepwise removal of terminal sugar moieties by sequential action of glycosidases, assisted by sphingolipid activator proteins37. Ceramide, the common lipid product of glycosphingolipids and sphingomyelin catabolism, is hydrolyzed by the lysosomal acid ceramidase into free fatty acid and sphingosine, which is finally metabolized by SphK1 and SphK2 to produce S1P38. We have shown that S1P/S1P3R/Gq/CK2 signaling axis plays a central role to induce PACS1-mediated retrograde transport of IGF-II/M6P receptor to the TGN (Figs. 3, 5, 6, 7), which implies that overall catabolic systems including the regulation of distribution of IGF-II/M6P receptor among cellular organelles are finally regulated by an end product, S1P. It is tempting to speculate that if sphingolipid catabolism is blocked in any degradation steps due to an inherited deficiency of lysosomal hydrolases, the end product S1P may be decreased in some local endo/lysosomal areas and the decrease has a negative effect on S1P3R/Gq/PACS1 axis signaling cascade to regulate the retrograde transport of IGF-II/M6P receptor to the TGN, which in turn affect to deliver hydrolases to the lysosomes. Further studies are necessary to support this hypothesis by showing improvement of lysosomal insufficiency by increasing cellular content of S1P, which may provide a novel molecular target therapy for the lysosomal storage diseases.

During endosomal maturation, i.e., from early endosomes to late/multivesicular endosomes, lipid composition changes drastically, e.g., catabolism of sphingomyelin and glycosphingolipids in the limiting membranes into S1P via sphingosine39. Concurrently, continuous ongoing activation of S1P receptors on each organelle via intracrine mechanisms, i.e., ligand-dependent activation of the S1P receptors within each organelle, may be a general phenomenon for constitutive vesicular transport thereby keeping transmitting G-protein signals to maintain cellular function. In the present study, we have shown that SphK inhibitor treatment or knockdown of SphK1 caused a decrease in the endogenous IGF-II/M6P receptor content at the TGN (Fig. 2) and inhibited the retrograde transport of CD8-IGF-II/M6P receptor to the TGN (Fig. 3), suggesting continuous SphK1/S1P-mediated activation of S1P3R on the endosomes, which keeps transmitting Gq signal (Figs. 5, 6, 7). The Gqɑ signal in turn activates CK2 by a mechanism needs to be clarified. Activated CK2 phosphorylates serine residues including Ser278 in the acidic cluster region of PACS1 and facilitates its association with IGF-II/M6P receptor. This process may be necessary for the retrograde transport of IGF-II/M6P receptor presumably with the aid of other proteins such as AP110,40 and retromer complex8,9. Interestingly, our pulse-chase assay suggested that SphK1/S1P3R and VPS35 are working on the same compartment (Supplementary Fig. 7A, B). Although we do not have any direct evidence about the relationship between SphK1/S1PR/Gq/CK2/PACS1 and VPS26/VPS35/VPS29 retromer pathways, it is tempting to hypothesize that the co-operation of PACS1 and retromers brings about an efficient exit of cargoes from the endosome. Our present study warrants further investigation into the functional relationship between PACS1 and retromer pathways. In the present study, we mainly focused on the retrograde transport of IGF-II/M6P receptor from the endosomes to the TGN. Since PACS1 has been reported to bind many proteins10, the importance of SphK1/S1P3R/Gq/CK2/PACS1 axis in retrograde transport may be more general rather than restricted to IGF-II/M6P receptor, as we showed for CTxB in this study. In terms of S1P signaling further studies are necessary to clarify the anterograde transport of IGF-II/M6P receptor from the TGN to the endosomes or to the PM, since SphK1 is abundant in the TGN19.

Methods

Antibodies & reagents

Anti-Cathepsin D antibody (clone 892CT11.1.1) was obtained from Abcepta. Anti-CD8a antibody (Clone RPA-T8) was from BioLegend; rabbit monoclonal anti-CD8a antibody (clone 1B19) was from Merck; rabbit anti-golgin-97 antibody from Proteintech; mouse anti-golgin97 antibody (clone CDF4) from Molecular Probes; anti-Rab5 (clone C8B1), anti-Rab7 (clone D95F2) and anti-Rab11 (clone D4F5) antibodies from Cell Signaling Technologies; anti-DDDDK-tag antibody and anti-α-tubulin antibody from Medical & biological laboratories; anti-VPS35 (clone B-5), anti-CK2α (clone E-7) and anti-SphK1 (clone G-11) from Santa Cruz; rabbit anti-M6PR antibody (clone EPR6599) from Abcam; mouse anti-M6PR antibody (clone 2G11) from Fitzgerald; anti-β-actin antibody from WAKO. Cathepsin D substrate (MOCAc-Gly-Lys-Pro-Ile-Leu-Phe-Phe- Arg-Leu-Lys(Dnp)-D-Arg-NH2) was purchased from Peptide Institute; HACPT (2-(p-hydroxyanilino)-4-(p-chlorophenyl)thiazole) from Calbiochem; W146, VPC23019 from Cayman Chemical; Gallein from Tokyo Chemical Industry; PTX, YM-254890, Cycloheximide, apigenin, Phos-tag Biotin, Anti DYKDDDDK tag antibody beads from Fujifilm Wako Pure Chemical; CtxB (Recombinant), Alexa Fluor™ 555 Conjugate from Thermo Fisher Scientific. Other reagents and chemicals were of analytical grade.

Plasmids and mutations

CD8-IGF-II/M6P receptor chimera was constructed by insertion of cDNA fragment of extracellular domain of human CD8a (sense primer, 5′-ATAGAATTCGCCACCATGGCCTTACCAGTGACCGC-3′; antisense primer, 5′-ATAAAGCTTACGTCTTCGGTTCCTGTGG-3′) and intracellular domain of human IGF-II/M6P receptor (sense primer, 5′-ATAAAGCTTAAGAAGGAGAGGAGGGAAAC-3′; antisense primer, 5′-ATAGGATCCTCAGATGTGTAAGAGGTCCTC-3′) into pCMV5 by using EcoRI, HindIII and BamHI. S1P3R-GFP was constructed as described18. N-terminally green fluorescent protein (GFP)-tagged mouse SphK1 and its catalytically inactive mutant, mSphK1G81D were constructed as described41. Human Gq cDNA was amplified by using primer sets 5′-ATAGGTACCATGACTCTGGAGTCATCATCGCG-3′ and 5′-ATAACGCGTTTAGACCAGATTGTACTCCTTCAGGTTCAAC-3′ to make N-terminally HA-tagged construct in pCMV5. Constitutively active Gq(Q209L) was constructed by QuikChange site-directed mutagenesis protocol using 5′-GATGTAGGGGGCCTAAGGTCAGAGAG-3′ and 5′-CTCTCTGACCTTAGGCCCCCTACATC-3′ and dominant negative Gq(Q209L/D277N) was constructed by 5′-CTGTTCTTAAACAAGAAAAATCTTCTAGAGGAG-3′ and 5′-CTCCTCTAGAAGATTTTTCTTGTTTAAGAACAG-3′ using Gq(Q209L) as a template. Human PACS1 cDNA was amplified by using primer sets 5’-ATAGGTACCATGGCGGAACGCGGAGGGGC-3′ and 5′-ATAACGCGTTCAGGTGGCCTTGCTGCCACTGAAG-3′ to make N-terminally FLAG-tagged construct in pCMV5. PACS1(S278A) or PACS1(S278D) was constructed by using 5′-GATATTGACAATTATGCCGAGGAAGAGGAAGAG-3′ and 5′-CTCTTCCTCTTCCTCGGCATAATTGTCAATATC-3′ or 5′-GATATTGACAATTATGACGAGGAAGAGGAAGAG-3′ and 5′- CTCTTCCTCTTCCTCGTCATAATTGTCAATATC-3′, respectively. Human CK2α cDNA was amplified by using primer sets 5′-ATAGAATTCTATGTCGGGACCCGTGCCAAGCAG-3′ and 5′-ATAGGTACCTTACTGCTGAGCGCCAGCGGCAG-3′ to make N-terminally GFP-tagged construct in pEGFP-C1.

siRNA

For RNA interference following oligonucleotides (Japan Bio Services, Saitama, Japan) were used: Sense 5′- GGGCAAGGCCUUGCAGCUCdTdT-3′ and antisense 5′-GAGCUGCAAGGCCUUGCCCdTdT-3′ for human SphK1; sense 5′- GCUGGGCUGUCCUUCAACCUdTdT-3′ and antisense 5′-AGGUUGAAGGACAGCCCAGCdTdT-3′ for human SphK2; sense 5′-AUAUCAUUUAGAUAAAGUGdGdG-3′ and antisense 5′-CACUUUAUCUAAAUGAUAUdTdA-3′ for human S1P1R; sense 5′-GGUCAACAUUCUGAUGUCUdTdT and antisense 5′-AGACAUCAGAAUGUUGUCCdTdT-3′ for human S1P3R; sense 5′-UAAAUGAGUUGGAUUUAGGdGdT-3′ and antisense 5′-CCUAAAUCCAACUCAUUUAdTdA-3′ for human CK2α; sense 5′-CUGGACAUAUUUAUCAAUAUAdTdT-3′ and antisense 5′-UAUAUUGAUAAAUAUGUCCAGdTdT-3′ for human VPS35. The knockdown effect of the siRNA was confirmed by quantitative PCR for SphK1-, SphK2-, S1P1R- or S1P3R-siRNA and western blotting for CK2α- or VPS35-siRNA (Supplementary Fig. 1).

Quantitative PCR

Total RNA was extracted from HeLa cells transfected with siRNA using NucleoSpin RNA II (Macherey-Nagel) according to the manufacturer’s instructions. One microgram of RNA was used for reverse transcription (ReverTra Ace qPCR RT kit, TOYOBO). Quantitative PCR was performed with SYBR Premix (Takara) on ABI Prism 7000. The expression level of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as the internal control. The primer sequences (sense and antisense) were as follows: for human S1P1R, 5′-TTCCTGGTGTTAGCTGTGCTCAAC-3′ and 5′-TCGCTTGAATTTGCCAGCAGAGTC-3′; for human S1P3R, 5′-AGCAGCAACAATAGCAGCCACTC-3′ and 5′-AGT- GCTGCGTTCTTGTCCATGATG-3′; for human SphK1, 5′-AGCTTCCTTGAACCATTAT- GCTG-3′ and 5′-AGGTCTTCATTGGTGACCTGCT-3′; for human SphK2, 5′-ATGAATGGACACCTTGAAGCAG-3′ and 5′-CATGGCCTTAGCCCTGACCAG-3′; for human GAPDH, 5′-GCCATCAATGACCCCTTCATT-3′ and 5′-TCTCGCTCCTGGAAGATGG-3′.

Cell culture and transfection

Human cell line derived from cervical cancer, HeLa (RCB0007), was provided by the RIKEN BRC through the National Bio-Resource Project of the MEXT, Japan, and tested for mycoplasma contamination occasionally by 4′,6-diamidino-2-phenylindole (DAPI) staining. HeLa cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Wako) containing 10% fetal bovine serum and 1% penicillin/streptomycin at 37 °C in 5% CO2. Transient transfection of plasmid DNA was carried out using FuGENE HD (Promega) according to the manufacturer’s instructions. Co-transfection of plasmid DNA and siRNA was done by Lipofectamine 3000 (Thermo Fisher Scientific) and siRNA was transfected by using Lipofectamine RNAiMAX (Thermo Fisher Scientific) according to the manufacturer’s instructions. For establishing HeLa cells stably expressing CD8-IGF-II/M6P receptor, HeLa cells were transfected with CD8-IGF-II/M6P receptor plasmid DNA and blasticidin-resistant gene-harboring plasmid DNA and selected by 10 µg/mL blasticidin for 3 weeks. Antibiotics-resistant cells were mildly trypsinized, stained by CoraLite Plus 647-conjugated anti-CD8 antibody (clone RPA-T8, Proteintech) and CD8-IGF-II/M6P receptor-positive cells were sorted by using FACSMelody (BD Biosciences).

Cathepsin D activity assay

Cathepsin D activity assay was performed as described in the previous report42. HeLa cells cultured in 24-well plates were incubated for 48 h after transfection with siRNA or 24 h after the addition of inhibitors without serum. Cells were then solubilized using lysis buffer (20 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% Triton X-100). The reaction mixture contained ~40 µg cell lysate and 10 µM fluorescent peptide substrate in 50 µL of assay buffer (50 mM acetate, 100 mM NaCl, pH 4.0). Mixtures were incubated at 37 °C for 60 min, and then the increase in fluorescence intensity produced by substrate cleavage during incubation was measured at an emission wavelength of 393 nm with excitation at 328 nm using EnSpire multimode plate reader (PerkinElmer), normalized by signals in the blank wells.

Cathepsin D secretion assay

Pro-cathepsin D secretion assay was performed as previously described43. HeLa cells grown on 6 well plates were transfected with siRNAs and incubated for 48 h in 10% serum-containing DMEM, and for the last 3 h, in serum-free DMEM containing 100 µg/mL of cycloheximide at 37 °C. Medium samples were treated with TCA to a final concentration of 17.2% after adding 5 µg of recombinant glutathione S-transferase (GST) protein as an internal standard and the acid precipitates were collected by centrifugation at 10,000 × g for 30 min and washed with ice-cold acetone and subjected to immunoblot analysis by using anti-IGF-II/M6P receptor antibody or anti-GST antibody. Cell lysates were prepared in the same experiments after harvesting culture media and subjected to the immunoblot analysis to measure the level of mature cathepsin D.

CD8-IGF-II/M6P receptor internalization assay and endogenous IGF-II/M6P receptor staining

HeLa cells stably or transiently expressing CD8-IGF-II/M6P receptor grown on glass-bottom dishes were transfected with siRNAs for 48 h or plasmid DNAs for 24 h. Cells were incubated in serum-free medium with inhibitors (5 µM HACPT, 5 µM W146, 5 µM JTE013 and 5 µM VPC23019, 100 ng/mL PTX and 10 µM Gallein or 10 µM YM-254890) in some experiments for 1 h before the assay except for PTX treatment being overnight. After washed with the medium twice, cells were incubated in DMEM without serum or bicarbonate with anti-CD8 antibody (5 µg/mL) on ice for 30 min. The cells were washed with DMEM without bicarbonate twice and incubated at 37 °C for 15 min in serum- and bicarbonate-free medium. Cells were fixed by 4% paraformaldehyde/PBS solution for 15 min and permeabilized by 0.2% Triton X-100-containing PBS for 10 min. Cells were blocked by 1% BSA/PBS for 10 min and then internalized mouse anti-CD8 was visualized by anti-mouse IgG Alexa 488 and golgin-97, Rab5, Rab7 or Rab11 was co-stained by the respective antibodies and Alexa 594-conjugated secondary antibody. To show the distribution of endogenous IGF-II/M6P receptor, cells were stained by anti-IGF-II/M6P receptor antibody and anti-Golgin97, anti-Rab5, anti-Rab7 or anti-Rab11 antibody after fixation and permeabilization. In some experiments, FlexAble CoraLite Plus 647 Antibody Labeling Kit for IgG1 (Proteintech) was used to label mouse anti-Golgin97 antibody for triple staining. The fluorescent images were obtained using a confocal laser scanning microscope (LSM700, Carl Zeiss) using an x63 oil plan-apochromat objective (numerical aperture, 1.4; Carl Zeiss). In each experiment, 30 cell images were subjected for Pearson’s coefficient quantification.

CTxB internalization assay

HeLa cells grown on glass-bottom dishes were transfected with siRNAs for 48 h. After washed with medium twice, cells were incubated in DMEM without serum or bicarbonate with 1 µg/mL of CtxB (Recombinant), Alexa Fluor™ 555 Conjugate (Thermo Fisher Scientific) on ice for 30 min. The cells were washed with DMEM without bicarbonate twice and incubated at 37 °C for 25 min in serum- and bicarbonate-free medium. Cells were fixed by 4% paraformaldehyde/PBS solution for 15 min and permeabilized by 0.2% Triton X-100- containing PBS for 10 min. Cells were blocked by 1% BSA/PBS for 10 min and then stained by anti-golgin-97 antibody. The fluorescent images were obtained using a confocal laser scanning microscope (LSM700, Carl Zeiss) using an x63 Oil Plan-Apochromat Objective (numerical aperture, 1.4; Carl Zeiss). In each experiment, 30 cell images were subjected for Pearson’s coefficient quantification.

Co-immunoprecipitation assay

HeLa cells stably expressing CD8-IGF-II/M6P receptor were transfected with FLAG-PACS1 with or without HA-Gq(Q209L) or siRNAs and incubated for 48 h at 37 °C, then treated with inhibitors and solubilized by lysis buffer (20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 5 mM NaF and protease inhibitor cocktail), followed by centrifugation at 20,000 × g for 10 min. FLAG-PACS1 was immunoprecipitated by anti-FLAG beads and separated by SDS-PAGE gel, blotted onto PVDF membranes (Millipore), and probed with anti-IGF-II/M6P receptor antibody or anti-FLAG antibody. Expression of HA-Gq(Q209L) was checked by anti-HA antibody in the cell lysate where anti-IGF-II/M6P receptor antibody was used as loading control.

Detection of PACS1 phosphorylation by Phos-tag technology

Phos-tag solution (Zn2+-phos-tag-biotin-streptavidin-HRP complex) was prepared by incubating 5 µL of Pierce High Sensitivity Streptavidin-HRP (Thermo Fisher Scientific) and 5 µL of 1 mM phos-tag biotin (WAKO) in the presence of 20 µM Zn(NO3)2 in 500 µL of Tris-buffered saline with 0.1% Tween 20 (TTBS) for 30 min at room temperature, followed by filter concentration to 10 µL using Nanosep with 30 K Omega (Pall corporation). FLAG-PACS1 or FLAG-PACS1(S278A) expressed in HeLa cells was immunoprecipitated by anti-FLAG beads, separated by SDS-PAGE, and blotted onto PVDF membrane as mentioned above. The blots were washed with TTBS and incubated with 30 mL TTBS containing Zn2+-phos-tag-biotin-streptavidin-HRP complex for 30 min to detect PACS1 phosphorylation. In some experiments, immunoprecipitants were incubated with 0.5 unit of alkaline phosphatase (Escherichia coli C75, Takara) for 30 min at 37 °C before SDS-PAGE analysis.

Pearson’s coefficient

Images were processed using ImageJ software. Colocalization analysis was performed using the ImageJ Coloc 2 plugin. For the analysis cell boundaries were identified by increasing the image contrast of anti-CD8 or anti-IGF-II/M6P receptor staining. Pearson’s R value was tabulated on GraphPad Prism software. The individual dots in the bar plot represent PCC value of each cell and the data for one graph were from one representative experiment from several replicated experiments, because absolute values of PCC varied slightly from experiment to experiment, probably dependent on the intensity of staining and other factors.

Statistics and reproducibility

Results are expressed as means ± s.e.m. Data were analyzed by Welch’s t-test using Prism 6 for Mac OS X (GraphPad Software). p values < 0.05 were considered significant. Statistical significance is shown as follows: *p < 0.05, **p < 0.01. p > 0.05 was considered not significant (NS). Error bars indicate s.e.m. Each bar graph shows a representative data in an experiment from at least three independent experiments except Fig. 1D where three independent experiments data were combined. Each dot in the bar graphs represents each well (in Figs. 1A, B, D; 5A; S1A–D) or each cell (in Figs. 2B, C; 3B, D, F; 4B; 5C, E; 6F; 7E; S3B; S4B; S5B, D; S6A). The number of replicates was at least 30 for the calculation of Pearson’s coefficient and 3 for the cathepsin D biochemical assay.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.