Biochemical Journal

Research article

Mucolipin 1 channel activity is regulated by protein kinase A-mediated phosphorylation

Silvia Vergarajauregui, Ross Oberdick, Kirill Kiselyov, Rosa Puertollano

Abstract

Mucolipins constitute a family of cation channels with homology with the transient receptor potential family. Mutations in MCOLN1 (mucolipin 1) have been linked to mucolipidosis type IV, a recessive lysosomal storage disease characterized by severe neurological and ophthalmologic abnormalities. At present, little is known about the mechanisms that regulate MCOLN1 activity. In the present paper, we addressed whether MCOLN1 activity is regulated by phosphorylation. We identified two PKA (protein kinase A) consensus motifs in the C-terminal tail of MCOLN1, containing Ser557 and Ser559. Ser557 was the principal phosphorylation site, as mutation of this residue to alanine caused a greater than 75% reduction in the total levels of phosphorylated MCOLN1 C-terminal tail. Activation of PKA with forskolin promoted MCOLN1 phosphorylation, both in vitro and in vivo. In contrast, addition of the PKA inhibitor H89 abolished MCOLN1 phosphorylation. We also found that PKA-mediated phosphorylation regulates MCOLN1 channel activity. Forskolin treatment decreased MCOLN1 channel activity, whereas treatment with H89 increased MCOLN1 channel activity. The stimulatory effect of H89 on MCOLN1 function was not observed when Ser557 and Ser559 were mutated to alanine residues, indicating that these two residues are essential for PKA-mediated negative regulation of MCOLN1. This paper presents the first example of regulation of a member of the mucolipin family by phosphorylation.

  • lysosome
  • mucolipin 1 (MCOLN1)
  • mucolipidosis IV (MLIV)
  • protein kinase A (PKA)
  • transient receptor potential (TRP)

INTRODUCTION

MLIV (mucolipidosis type IV) is an autosomal recessive disease characterized by mental and psychomotor retardation, diminished muscle tone or hypotonia, achlorhydria and visual problems, including corneal clouding, retinal degeneration, sensitivity to light and strabismus [13]. This disorder is found at a relatively high frequency among Ashkenazi Jews, with 1:93 of the population estimated to be a genetic carrier [4]. Analysis of fibroblasts from MLIV patients by electron microscopy revealed the presence of enlarged vacuolar structures that accumulate mucopolysaccharides, phospholipids and gangliosides [58]. These enlarged vacuoles are present not only in fibroblasts, but also in many different cell types, indicating a general impairment of lysosomal function in MLIV patients. However, in contrast with other lysosomal storage diseases, lysosomal hydrolases do seem to be functional and correctly transported to lysosomes in MLIV, suggesting that the observed defect in lysosomal function may be the result of alterations in trafficking along the lysosomal pathway [9,10].

Mutations in MCOLN1 (mucolipin 1) have been shown to be the cause of MLIV [1114]. MCOLN1 is an ion channel with homology with the TRP (transient receptor potential) channel superfamily. The TRPs constitute a large family of cation channels that are expressed in most cell types and function as cellular sensors of various internal and external stimuli, including changes in temperature and osmolarity, light, pain, pheromones, odorants and membrane stretch (reviewed in [15]). The predicted topology of MCOLN1 is of six transmembrane-spanning domains, with the N- and C-terminal tails oriented within the cytosol. The region of homology with the other members of the TRP family consists of transmembrane domains 3–6 (amino acids 331–521), which includes the pore located between transmembrane segments 5 and 6 (amino acids 496–521). The selectivity of the MCOLN1 channel remains controversial, as different studies have suggested that the channel is permeable to Ca2+ [16]; Ca2+, K+ and Na+ [17]; and H+ [18]. In addition, MCOLN1 activity can be modulated by changes in pH and Ca2+ concentration [17,19].

The lysosomal defects observed in MLIV patients have led to the suggestion that MCOLN1 might play a role in the biogenesis of lysosomes [20]. This idea was supported by the observation that knockout of cup-5, the orthologue of MCOLN1 in Caenorhabditis elegans, resulted in formation of enlarged hybrid organelles that contained both late-endosomal and lysosomal markers [21]. Recently, several groups have suggested additional roles for MCOLN1 in different cellular processes, including regulation of lysosomal acidification [18], autophagy of mitochondria [22] and lysosomal secretion [23].

One way to gain information on the function of a protein is to analyse its distribution and trafficking into the cell. For example, consistent with its role in lysosomal function, MCOLN1 contains specific sorting signals that mediate its transport to lysosomes [2426]. These sorting signals include two di-leucine motifs located at the N- and C-terminal tails that mediate direct interaction with clathrin adaptors [24]. Post-translational modifications also play an important role in the regulation of protein function. We have shown previously that the C-terminal tail of MCOLN1 is palmitoylated, and that palmitoylation promotes efficient internalization of MCOLN1 from the plasma membrane [24].

Phosphorylation is another modification that critically regulates protein activity. The activity of many TRP channels is regulated by phosphorylation and dephosphorylation events. A comprehensive overview of the regulation of TRP channels, including phosphorylation and interaction with auxiliary proteins is reviewed in [26a,26b]. TRP phosphorylation may occur on serine, threonine and tyrosine residues, and is mediated by different kinases, including PKC (protein kinase C), PKA (protein kinase A), Src, Fyn, PKG (protein kinase G) and CaMKII (Ca2+/calmodulin-dependent protein kinase II) (reviewed in [27]).

In this present paper, we addressed whether the activity of MCOLN1 is regulated by phosphorylation. We report the identification of two consensus PKA phosphorylation sites in the C-terminal tail of MCOLN1 at Ser557 and Ser559. Activation of PKA by FSK (forskolin) induced MCOLN1 phosphorylation, both in vitro and in vivo, and negatively regulated MCOLN1 channel activity in vivo. Conversely, treatment with the PKA inhibitor H89 inhibited phosphorylation of the MCOLN1 C-terminal tail and caused an increase in MCOLN1 activity. Therefore our results show for the first time that PKA-mediated phosphorylation regulates MCOLN1 channel activity.

EXPERIMENTAL

Antibodies and reagents

Rabbit polyclonal anti-GFP (green fluorescent protein) antibody was purchased from MBL International (Woburn, MA, U.S.A.) and mouse monoclonal anti-FLAG M2 antibody from Sigma. Mouse monoclonal anti-GFP (ab1218) and rabbit anti-LAMP1 (lysosome-associated membrane protein 1) (ab24170) antibodies were from Abcam. Rabbit polyclonal anti-(MCOLN1 CTail) was raised against the C-terminal tail of MCOLN1 (amino acids 518–580) fused to GST. The specificity of the antibody was verified by recognition of transfected proteins and by the absence of a specific signal in fibroblasts from MLIV patients. Phosphatase inhibitor cocktails 1 and 2, genistein, FSK and protease inhibitor cocktail were obtained from Sigma. Serine/threonine kinase inhibitor set [containing BIS (bisindolylmaleimide I), H89, KN-93, ML-7 and staurosporine], Akt inhibitor IV and the purified catalytic fragment of PKA were supplied by Calbiochem. [γ-32P]ATP (3000 Ci/mmol) was obtained from GE Healthcare. [32P]Pi (orthophosphoric acid, 8500–9120 Ci/mmol) was obtained from PerkinElmer.

Plasmids

Full length MCOLN1–GFP, MCOLN1 L15A/L16A/L577A/L578A–GFP and GFP–MCOLN1-CTail (where CTail is C-terminal tail) were cloned as described previously [24]. The N-terminal (amino acids 1–66) and the C-terminal cytosolic tail (amino acids 518–580) of MCOLN1 were cloned into the EcoRI and SalI sites of pGEX-5X (Pharmacia Biotech). To generate MCOLN1–FLAG, full length MCOLN1 was cloned into the EcoRI and SalI sites of pCMV-Tag 4B (Stratagene). Mutations of residues in the cytosolic tails or the first extracellular loop of MCOLN1 were introduced using the QuikChange site-directed mutagenesis kit following the manufacturer's instructions (Stratagene). The incorporation of site-directed mutations was confirmed by subsequent DNA sequencing.

Cell lines and transfections

HeLa and NRK cells (normal rat kidney cells) were maintained in DMEM (Dulbecco's modified Eagle's medium, Life Technologies) supplemented with 100 units/ml penicillin (Gibco), 100 μg/ml streptomycin (Gibco) and 10% (w/v) FBS (fetal bovine serum). Transfection of HeLa and NRK cells was performed using FuGENE™ 6 (Roche) according to the manufacturer's recommendations. In transient transfection assays, analyses were conducted 24 h post-transfection.

In vitro phosphorylation

GST (glutathione transferase) fusion proteins were expressed in bacteria and affinity purified using standard methods as described previously [28]. Purified GST fusion proteins (10 μg) bound to glutathione–Sepharose (20 μl volume) were washed once in lysate kinase buffer [20 mM Tris (pH 7.5), 0.5 mM DTT (dithiothreitol) and 10 mM MgCl2], and then incubated with either 1 μl postnuclear supernatant (14 μg total protein) from HeLa cells or with 1000 units of the purified catalytic subunit of PKA (Calbiochem) for 20 min at 30 °C in lysate kinase buffer containing 200 μM ATP and [γ-32P]ATP (final specific radioactivity of 500 μCi/μmol). The reactions were stopped by washing with cold PBS containing 15 mM EDTA. Phosphorylated proteins were eluted by the addition of 25 μl Laemmli buffer before separation by electrophoresis on 4–20% gradient polyacrylamide gels. Gels were fixed and stained with Coomassie Brilliant Blue, and then dried and subjected to autoradiography. Preparation of the postnuclear supernatant was performed as described previously [28]. Briefly, cells were harvested using trypsin, resuspended in 10 mM Hepes/KOH (pH 7.4), 320 mM sucrose, 10 mM EDTA, 5 mM EGTA and 1 mM DTT, supplemented with protease inhibitors, and disrupted by sonication (15 pulses, intermediate power setting). Nuclei and cell debris were removed by centrifugation at 1300 g for 5 min at 4 °C, and the supernatant stored at −70 °C until required. For experiments using inhibitors, the postnuclear supernatant was pre-incubated with the indicated inhibitor for 30 min at room temperature (23 °C) prior to the phosphorylation reaction.

Metabolic labelling and immunoprecipitation

HeLa cells transfected with GFP–MCOLN1-CTail, GFP–MCOLN1-CTail S557A/S559A, MCOLN1–FLAG or MCOLN1 S557A/S559A–FLAG for 24 h were incubated with serum-free DMEM lacking phosphate for 4 h at 37 °C, and then radiolabelled for 3 h in the same medium in the presence of 0.15 mCi/ml [32P]Pi. For kinase stimulation, 100 μM FSK was added by rapid mixing during the last 30 min of radiolabelling. After radiolabelling, the cells were washed with ice-cold PBS and lysed in phosphorylation buffer [20 mM Tris/HCl (pH 7.4), 400 mM NaCl, 1 mM EDTA, 1 mM MgCl2 and 1 mM CaCl2] supplmented with 2% (v/v) Triton X-100 and protease and serine/threonine phosphatase inhibitors. The lysate was pre-cleared by centrifugation at 16000 g for 10 min at 4 °C and incubated for 4 h at 4 °C with 5 μl monoclonal anti-GFP antibody or 3 μl anti-FLAG antibody, followed by Protein G–Sepharose (20 μl of a 50% slurry; Amersham Biosciences) for 4 h at 4 °C. Immunoprecipitates were collected, washed three times with phosphorylation buffer with 1% Triton X-100, once with phosphorylation buffer with 0.1% SDS and once with PBS. The washed immunoprecipitates were then eluted using Laemmli sample buffer, analysed by SDS/PAGE (4–20% gradient gels) under reducing conditions, and transferred on to nitrocellulose. The nitrocellulose membrane was exposed to Kodak X-AR film for 1 day to visualize radiolabelled proteins and then subjected to Western blotting using polyclonal anti-GFP (1:1000 dilution) and anti-(MCOLN1 CTail) (1:1000 dilution) antibodies, which were incubated overnight at 4 °C.

Electrophysiology

HEK-293 (human embryonic kidney) cells were transfected with GFP–MCOLN1 constructs or with YFP (yellow fluorescent protein)- and HA-tagged wild-type human MCOLN1 and maintained in DMEM. Cells were seeded 24–48 h post transfection on to coverslips and used for experiments. The whole-cell mode of patch–clamp technique was executed as described previously [18,29]. The pipette solution contained 150 mM caesium aspartate, 2 mM MgCl2, 5 mM sodium ATP (pH 7.2) and 10 mM Hepes, with 0.1 μmol Ca2+ buffered at pCa 8 with 2 mM EGTA or 10 mM BAPTA [1,2-bis-(o-aminophenoxy)ethane-N,N,N′,N′-tetra-acetic acid]. The extracellular solution contained 10 mM Hepes, 150 mM NaCl (or sodium gluconate), 1 mM CaCl2, 1 mM MgCl2 and 5 mM KCl (pH 7.6). Drugs were applied by bath perfusion.

Pipette resistance was in the 2–7 MΩ range and seal resistance was always greater than 1 GΩ. Shortly after establishing the seal and breaking into the cells, cells were challenged with 20 mV voltage steps (250 ms long) between −100 and +100 mV, and the resulting current was recorded using a EPC-10 amplifier (HEKA Instruments, Lambrecht, Germany). Currents were recorded using the Patchmaster program (HEKA Instruments). The liquid junction potential was neutralized and currents were digitized at 1 kHz, filtered using a low-pass digital Bessel filter set at 300 Hz, and were otherwise not conditioned. Offline analysis was performed using the Origin 7 program (OriginLabs, Northampton, MA, U.S.A.).

RESULTS

The C-terminal tail of MCOLN1 is phosphorylated by serine/threonine protein kinases

To better understand the molecular mechanisms that regulate MCOLN1 activity, we analysed whether the N- and C-terminal tails of MCOLN1 are modified by phosphorylation. The soluble N-terminal tail [GST–MCOLN1-NTail (where NTail is the N-terminal tail), residues 1–66] and the C-terminal tail (GST–MCOLN1-CTail, residues 518–580) proteins were synthesized in bacteria as GST-tagged fusion proteins. Following purification, these fusion proteins were incubated with HeLa cell lysates and subjected to in vitro kinase assays. After 20 min, the reactions were terminated by adding PBS containing EDTA, the products were separated by SDS/PAGE, and the gel was dried and exposed to X-ray film to detect 32P incorporation. The samples were also stained with Coomassie Brilliant Blue to verify that equal amounts of protein were loaded in each lane. As seen in Figure 1(A), GST–MCOLN1-CTail was efficiently phosphorylated in vitro by protein kinase activities present in HeLa cell lysates. The N-terminal tail of MCOLN1 was also phosphorylated, although to a much lower extent than the C-terminal tail. Therefore we concentrated on the characterization and relevance of phosphorylation of the MCOLN1 C-terminal tail.

Figure 1 In vitro phosphorylation of the C-terminal tail of MCOLN1

(A) Phosphorylation of MCOLN1 N- and C-terminal tail fusion proteins. Fusion proteins containing GST fused to wild-type MCOLN1 N-terminal tail (GST–MCOLN1-NTail) or C-terminal tail (GST–MCOLN1-CTail) were incubated in vitro with [γ-32P]ATP and postnuclear HeLa cell supernatant for 20 min at 30 °C. The proteins were then separated by electrophoresis, stained with Coomassie Brilliant Blue (TOTAL, lower panel) and subjected to autoradiography (32P, upper panel). (B) Treatment with staurosporine causes a reduction in MCOLN1 phosphorylation. GST–MCOLN1-CTail was incubated as stated in (A) with postnuclear supernatant from HeLA cells that had been pre-incubated for 30 min at room temperature with the broad spectrum serine/threonine kinase inhibitor staurosporine (100 nM), the tyrosine kinase inhibitor genistein (100 μM), a serine/threonine phosphatase inhibitor (Ser Pase inh; phosphatase inhibitor cocktail 1, 1:100 dilution) or with a tyrosine phosphatase inhibitor (Tyr Pase inh; phosphatase inhibitor cocktail 2, 1:100 dilution). Samples were separated by electrophoresis, stained with Coomassie Brilliant Blue (TOTAL, lower panel) and subjected to autoradiography (32P, upper panel). Staining with Coomassie Brilliant Blue showed similar levels of fusion protein in each lane.

To determine whether phosphorylation of the GST–MCOLN1 C-terminal tail occurs on tyrosine or serine/threonine residues, we performed in vitro phosphorylation assays in the presence of specific inhibitors. Incubation with staurosporine, a broad-range serine/threonine kinase inhibitor, almost completely abolished GST–MCOLN1-CTail phosphorylation, whereas the addition of serine/threonine phosphatase inhibitors slightly increased the signal (Figure 1B). In contrast, incubation with tyrosine phosphatase inhibitors or with genistein, an inhibitor of tyrosine phosphorylation, did not have any effect on the phosphorylation of the MCOLN1 C-terminal tail (Figure 1B). Therefore our results suggest that a serine/threonine kinase present in HeLa cell lysates phosphorylates the C-terminal tail of MCOLN1.

Ser557 and Ser559 are two phosphorylation sites in MCOLN1 C-terminal tail

We then conducted a sequence search using the NetPhosK 2.0 program, which predicts potential phosphorylation sites at serine, threonine and tyrosine residues [30]. We found six residues that are potential candidate substrates for serine/threonine-dependent phosphorylation in the MCOLN1 C-terminal tail sequence: Thr523, Ser547, Ser550, Ser557, Ser559 and Ser572. To identify the residues that participate in MCOLN1 phosphorylation, several mutant proteins were generated (Figure 2A). When all the above detailed serine residues were mutated to alanine residues (S547A/S550A/S557A/S559A/S572A mutant), there was virtually no incorporation of 32P into GST–MCOLN1-CTail, indicating that phosphorylation occurs at serine residues. In agreement with this observation, the T523A mutant exhibited phosphorylation levels equivalent to those of the wild-type MCOLN1 C-terminal tail. Moreover, mutation of Ser574, Ser550, and Ser572 to alanine (S572A and S547A/S550A/S572A mutants) did not affect MCOLN1 C-terminal tail phosphorylation. In contrast, the proteins which were mutated on Ser557 and Ser559 (S547A/S550A/S557A/S559A/S572A and S557A/S559A/S572A mutants) were not phosphorylated (Figure 2B).

Figure 2 MCOLN1 is phosphorylated on Ser557 and Ser559

(A) The primary structure of the C-terminal tail of wild-type MCOLN1 (WT), with the putative phosphorylated residues in bold identified using the NetPhosK 2.0 Program, and the different mutants carrying alanine replacements at the indicated residues. 5S/A, S547A/S550A/S557A/S559A/S572A; S547,550,572/AAA, S547A/S550A/S572A; S557,559,572/AAA, S557A/S559A/S572A; S572/A, S572A; T523/A, T523A. (B) In vitro phosphorylation of wild-type GST–MCOLN1-CTail (WT) or GST–MCOLN1-CTail T523A (T523/A), S547A/S550A/S557A/S559A/S572A (5S/A), S572A (S572/A), S557A/S559A/S572A (S557,559,572/AAA) or S547A/S550A/S572 (S547,550,572/AAA) mutants. Samples were separated by electrophoresis, stained with Coomassie Brilliant Blue (TOTAL, lower panel) and subjected to autoradiography (32P, upper panel). Staining with Coomassie Brilliant Blue showed similar levels of fusion protein in each lane. (C) In vitro phosphorylation was performed using HeLa cell extracts to phosphorylate wild-type GST–MCOLN1-CTail (WT) or GST–MCOLN1-CTail S557A (S557/A), S559A (S559/A) or S557A/S559A (S557,559/AA) mutants. Samples were separated by electrophoresis, stained with Coomassie Brilliant Blue (TOTAL, lower panel) and subjected to autoradiography (32P, upper panel). Staining with Coomassie Brilliant Blue showed similar levels of fusion protein in each lane. The right-hand panel shows quantification of phosphorylation of the GST–MCOLN1-CTail S557A (S557/A), S559A (S559/A) or S557A/S559A (S557,559/AA) mutants, expressed as a percentage of wild-type GST–MCOLN1-CTail (WT) phosphorylation (y-axis). Results are means±S.D. (n=3).

To address the individual contribution of Ser557 and Ser559 in the C-terminal tail of MCOLN1, we mutated these residues separately. When the S557A mutant was used as a substrate, there was an >75% reduction in the extent of in vitro phosphorylation of GST–MCOLN1-CTail. In contrast, there was only a 25% difference in protein phosphorylation levels between the wild-type GST–MCOLN1-CTail and the S559A mutant. Finally, the S557A/S559A double mutant was not significant phosphorylated (Figure 2C). Taken together, the results indicate that Ser557 is the major in vitro phosphorylation site in the MCOLN1 C-terminal tail, whereas Ser559 is phosphorylated to a much lesser extent.

PKA induces phosphorylation of the MCOLN1 C-terminal tail

To determine the type of serine/threonine protein kinase responsible for phosphorylation of MCOLN1, we examined the effects of various serine/threonine protein kinase inhibitors. Using our in vitro kinase assay, we found that addition of H89 to HeLa cell lysates effectively inhibited phosphorylation at concentrations at which this inhibitor is known to inhibit PKA [31]. In contrast, there was no effect on phosphorylation with other inhibitors, including BIS, KN-93, Akt inhibitor IV or ML-7, which are known to inhibit PKC, CaMKII, Akt and MLCK (myosin light-chain kinase) respectively (Figure 3A).

Figure 3 Phosphorylation of GST–MCOLN1-CTail by PKA

(A) In vitro phosphorylation of GST–MCOLN1-CTail was performed using HeLa extracts pre-incubated with either 1 μM BIS (PKC inhibitor), 1 μM H89 (H-89; PKA inhibitor), 1 μM KN-93 (CaMKII inhibitor), 5 μM Akt inhibitor IV (AKT inh IV; protein kinase B inhibitor), 1 μM ML-7 [MLCK (myosin light-chain kinase) inhibitor] or 100 nM staurosporine (Staurosp; broad-range serine/threonine kinase inhibitor). Samples were separated by electrophoresis, stained with Coomassie Brilliant Blue to determine total protein levels (TOTAL, lower panel) and subjected to autoradiography (32P, upper panel). (B) GST–MCOLN1-CTail phosphorylation by HeLa cell extracts (14 μg total protein) treated with increasing concentrations of the PKA inhibitor H89 (0.1, 1 and 10 μM), or with 10 μM of the PKA activator FSK. Samples were separated by electrophoresis (left-hand panel), stained with Coomassie Brilliant Blue to determine total protein levels (TOTAL, lower panel) and subjected to autoradiography (32P, upper panel). Quantification of GST–MCOLN1-CTail phosphorylation, expressed as fold activation of samples without treatment, is shown in the right-hand panel. Results are means±S.D. (n=3) (C) Phosphorylation of GST–MCOLN1-CTail by purified PKA. GST, wild-type GST–MCOLN1-CTail (WT) or GST–MCOLN1-CTail S557A (S557/A), S559A (S559/A) or S557A/S559A (S557,559/AA) mutants were incubated with [γ-32P]ATP and the catalytic fragment of PKA for 20 min at 30 °C. Samples were separated by electrophoresis (left-hand panel), stained with Coomassie Brilliant Blue to determine total protein levels (TOTAL, lower panel) and subjected to autoradiography (32P, upper panel).Wild-type GST–MCOLN1-CTail shows robust phosphorylation, and mutation of Ser557 largely reduced this phosphorylation, whereas the double mutant eliminated phosphorylation.

Because H89 inhibition of MCOLN1 phosphorylation is indicative of PKA activity, we sought to determine whether MCOLN1 is phosphorylated after treatment of cells with FSK, a PKA activator. As expected, phosphorylation of the C-terminal tail of MCOLN1 was significantly increased after FSK treatment. Densitometric measurements indicated an >3-fold increase in phosphorylation after incubation with FSK (Figure 3B).

Finally, we tested the ability of GST–MCOLN1-CTail to serve as a substrate for purified PKA. As shown in Figure 3(C), GST–MCOLN1-CTail was substantially phosphorylated after incubation with purified PKA. Once again, mutation of Ser557 to an alanine residue resulted in a 90% reduction in PKA-mediated phosphorylation of GST–MCOLN1-CTail, whereas phosphorylation of the S557A/S559A double mutant was not detectable. In conclusion, our results demonstrate that MCOLN1 is directly phosphorylated by PKA.

FSK induces phosphorylation of MCOLN1 in vivo

Although our biochemical experiments showed that PKA incorporated 32P into MCOLN1, the conditions used for in vitro phosphorylation may not reflect the conditions found in the cellular environment. In addition, treatment of proteins with detergent for solubilization may expose residues that are typically not available for phosphorylation in vivo. Therefore we investigated whether a chimaera consisting of the C-terminal tail of MCOLN1 fused to GFP (GFP–MCOLN1-CTail) was also phosphorylated in vivo. To reduce basal phosphorylation, cells were maintained in serum-free DMEM for 4 h prior to treatment with FSK. Subsequent metabolic labelling with [32P]Pi and immunoprecipitation of HeLa cells expressing GFP–MCOLN1-CTail showed that incubation with FSK induced an >3-fold increase in the levels of phosphorylation of this chimaera, whereas it had no effect on the S557A/S559A double mutant (Figure 4A).

Figure 4 In vivo phosphorylation of MCOLN1

(A) HeLa cells transfected with wild-type GFP–MCOLN1-CTail (WT) or with a GFP–MCOLN1-CTail S557A/S559A double mutant (S557,559/AA) were starved for 4 h in DMEM and metabolically labelled with [32P]Pi for 3 h. FSK (100 μM) was added to the medium during the last 30 min of labelling. Cells were then lysed, immunoprecipitated with a monoclonal anti-GFP antibody, subjected to SDS/PAGE and analysed either by Western blotting (WB) using a polyclonal anti-GFP antibody (lower panel) or by autoradiography (32P, upper panel). Quantification of GFP–MCOLN1-CTail in vivo phosphorylation is expressed as fold activation of wild-type GFP–MCOLN1 CTail in the absence of FSK (right-hand panel). Results are means±S.D. (n=3). (B) Phosphorylation of full length MCOLN1–FLAG. HeLa cells transfected with full length MCOLN1–FLAG (MCOLN1 WT) or MCOLN1 S557A/S559A–FLAG (MCOLN1 S557,559/AA) were metabolically labelled in the presence of [32P]Pi for 3 h at 37 °C, and 100 μM FSK or 1 μM H89 was added for the last 30 min of labelling. Cells were then lysed, immunoprecipitated with a monoclonal anti-FLAG antibody, subjected to SDS/PAGE and analysed either by Western blotting using a polyclonal anti-(MCOLN1 CTail) antibody (WB, lower panel) or by autoradiography (32P, upper panel). The arrows correspond to the theoretical molecular mass of the monomeric and dimeric forms of the full length MCOLN1–FLAG fusion protein. Higher bands may correspond to oligomers of MCOLN1–FLAG. Quantification of full length MCOLN1–FLAG in vivo phosphorylation is expressed as fold activation of MCOLN1–FLAG in the absence of FSK (right-hand panel). Results are means±S.D. (n=3).

Finally, we expressed FLAG-tagged full length MCOLN1 (MCOLN1–FLAG) in HeLa cells. We have shown previously that the presence of epitopes at the C-terminal tail of MCOLN1 does not alter the function or cellular distribution of the protein, as these chimaeras are targeted to lysosomes and induce accumulation of late endosome/lysosome hybrid organelles [24]. As demonstrated in Figure 4(B), analysis of [32P]Pi-labelled HeLa cells revealed that the level of phosphorylation of full length MCOLN1 was increased after treatment with FSK and decreased by H89. In contrast, MCOLN1 S557A/S559A was not phosphorylated in response to FSK. Therefore we conclude that full length MCOLN1 is phosphorylated in vivo by PKA on Ser557 and Ser559.

Effect of PKA-mediated phosphorylation on MCOLN1 distribution

We have described previously that MCOLN1 reaches lysosomes through either a direct [from TGN (trans-Golgi network) to early endosomes and subsequently to lysosomes] or an indirect (from TGN to plasma membrane, followed by internalization to endosomes and transport to lysosomes) pathway [24]. To assess whether phosphorylation of the MCOLN1 C-terminal tail has an effect on the trafficking of the protein, we analysed the cellular distribution of two MCOLN1 mutants in which Ser557 and Ser559 were mutated to alanine residues (MCOLN1 S557A/S559A–GFP) or to an aspartic acid residue (MCOLN1 S557D–GFP) to prevent or mimic phosphorylation respectively. Both mutants extensively co-localized with the lysosomal marker LAMP1, indicating that phosphorylation of Ser557 and Ser559 did not affect trafficking of MCOLN1 to lysosomes (see Supplementary Figure 1 at http://www.BiochemJ.org/bj/410/bj4100417add.htm).

We also queried whether phosphorylation occurs once MCOLN1 reaches lysosomes, or in a pre-lysosomal compartment. To address this question, we used a plasma-membrane-retained MCOLN1 mutant. Generation of this mutant has been described previously [24] and was obtained by mutating two di-leucine motifs located in the N- and C-terminal tails of MCOLN1 (MCOLN1 L15A/L16A/L577A/L578A). As seen in Supplementary Figure 2 (http://www.BiochemJ.org/bj/410/bj4100417add.htm), MCOLN1 L15A/L16A/L577A/L578A was phosphorylated in vivo, and the level of phosphorylation was increased after treatment with FSK. Moreover, another MCOLN1 mutant that contains mutations at the first extracellular loop (MCOLN1 V195A/D196A/P197A/P198A) and is retained at the ER (endoplasmic reticulum) was also phosphorylated in vivo after incubation with FSK (see Supplementary Figure 2). Therefore these results suggest that MCOLN1 phosphorylation may occur in pre-lysosomal compartments.

PKA activation and inhibition affects MCOLN1 activity

In order to test the effect of MCOLN1 phosphorylation on its activity, we analysed changes in amplitudes of MCOLN1-dependent currents associated with inhibition or activation of PKA. Full length recombinant human MCOLN1 was transfected into HEK-293 cells, a model that has been shown previously to deposit some of the recombinant overexpressed MCOLN1 into the plasma membrane [18,29]. To facilitate MCOLN1 detection, we also used the plasma-membrane-retained MCOLN1 mutant MCOLN1 L15A/L16A/L577A/L578A–GFP. In these experiments, the pipette solution contained 150 mM caesium, and the outside medium contained 150 mM sodium as the main charge carrier. These conditions were used to preserve continuity with our previous publications [18,29]. Whole-cell current recordings in HEK-293 cells showed that expression of plasma-membrane-retained MCOLN1 resulted in characteristically outwardly rectifying currents, which are very similar to those of wild-type MCOLN1 localized to the plasma membrane under overexpression conditions (Figure 5), and to those reported previously in the plasma membrane under overexpression conditions [18,29] or in artificial lipid bilayers using cell-free synthesized protein [17,19]. Control YFP only transfected cells did not show such activity (Figure 5). The similarities include monovalent cation permeability and strong outward rectification. Similar to previously published reports [17,19], MCOLN1 rectification did not depend on the permeating cation species and was unaffected by omission of either Ca2+ or Mg2+ from the extracellular and intracellular solutions (results not shown). Rectification is a characteristic feature of many TRP channels and it is usually attributed to the dependence of the open probability of the channel on membrane potential. This exact phenomenon has been directly shown for MCOLN1: its open probability, but not the single channel current amplitude, is significantly decreased at negative membrane potentials [17,19]. The currents recorded using the membrane-retained construct were somewhat less variable than those detected with wild-type recombinant MCOLN1. Among other factors, fusion of wild-type MCOLN1 containing lysosomes with the plasma membrane or endocytosis of MCOLN1 may contribute to such variability, and thus we opted to use the plasma-membrane-targeted constructs alongside wild-type MCOLN1 to test the effect of FSK and H89 on MCOLN1 activity.

Figure 5 Monovalent cation permeability through wild-type and membrane-targeted MCOLN1

HEK-293 cells were transfected with MCOLN1 L15A/L16A/L577A/L578A–GFP (MCOLN1-GFP-L15L/AA-L577L/AA) or co-transfected with wild-type MCOLN1–HA (WT MCOLN1-HA) and YFP. Transfection with YFP only was performed as a control. Whole-cell currents were recorded with 150 mM caesium aspartate in the pipette solution and 150 mM NaCl in the extracellular solution. Similar outward rectification and reversal potentials of the monovalent currents were mediated by wild-type HA-tagged MCOLN1 and the plasma membrane-retained MCOLN1 L15A/L16A/L577A/L578A–GFP mutant.

Figure 6 shows examples of such experiments. In all experiments, the cells were allowed to stabilize for 6 to 9 min after the start of the experiment, during which time the basal MCOLN1 activity was recorded every 3 min. After application of the drugs, the current was recorded every 3 min for 15–24 min. Figures 6(A) and 6(B) (upper panels) show that cell treatment with 10 μM H89 induces an increase in amplitudes of both wild-type and membrane-retained MCOLN1. The increase was gradual, and in most cases lasted beyond the timeframe of our experiments. In Figure 6(E), we summarize the effect of H89 on MCOLN1 activity. Amplitudes of MCOLN1-dependent currents recorded 15 min after H89 application were normalized to current amplitudes recorded before H89 application, which were considered to be 100% in each separate experiment. Most of the control cells did not respond to H89 application; in the few cases that responded to H89, the current amplitude was significantly lower than that measured in MCOLN1-expressing cells (Figures 6D and 6E).

Figure 6 The effect of PKA activation and inhibition on MCOLN1 activity

(AD) Examples of whole-cell current recordings obtained from cells incubated with FSK (50 μM) (A, B) or H89 (10 μM) (AD) after co-transfection with YFP and MCOLN1 wild-type (WT MCOLN1) (A), MCOLN1 L15A/L16A/L577A/L578A–GFP (MCOLN1-GFP-L15L/AA-L577L/AA) (B), MCOLN1 S557A/S559A–GFP (MCOLN1-GFP-S557,559/AA) (C), and from YFP-only transfected cells treated with 10 μM H89 (control) (D). In (A) and (B), the time and current amplitude scales are shared between the upper and lower panels. Larger control currents were selected for FSK experiments to emphasize the effect. The pipette solution contained 150 mM caesium aspartate and the extracellular solution contained 150 mM NaCl as the main charge carrier. The current traces illustrate channel activity before and 15 min after application of the drugs. (E) A summary of the effects of FSK and H89 on MCOLN1 channel activity. To statistically analyse the effect of the drugs, currents at +100 mV were recorded 15 min after stimulation and expressed as a percentage of the current before stimulation, which was taken as 100%. Results are means±S.D. (n=4–6), in which MCOLN1-mediated currents were clearly detectable. All the differences are statistically significant, with the exception of the effect of H89 on the MCOLN1 S557A/S559A–GFP (MCOLN1-GFP-S557,559/AA) mutant, which did not demonstrate a statistically significant change.

Figures 6(A) and 6(B) (lower panels) show that treatment with 50 μM FSK inhibited both membrane-retained and wild-type MCOLN1 currents. As with H89, the effect was gradual and fully developed within 15 min following application of the drug. Neither H89 nor FSK seemed to have a noticeable effect on the S557A/S559A double mutant (MCOLN1 S557A/S559A–GFP) (Figures 6C and 6E). On the basis of these results, we conclude that inhibition of PKA and MCOLN1 dephosphorylation increases the activity of MCOLN1, whereas MCOLN1 phosphorylation inhibits the activity of MCOLN1.

DISCUSSION

In this present paper, we have shown that the activity of MCOLN1 is regulated by phosphorylation. We found that PKA activation by FSK induced MCOLN1 phosphorylation, both in vitro and in vivo, and reduced MCOLN1 channel activity. In contrast, inhibition of PKA by treatment with H89 dramatically increased MCOLN1 activity. Mutagenesis analysis identified two PKA consensus sites in the C-terminal of MCOLN1 that mediate PKA phosphorylation of MCOLN1. Ser557 seems to be the major phosphorylation site for MCOLN1, as mutation of this residue to alanine caused an >75% reduction in the levels of phosphorylation on the MCOLN1 C-terminal tail. Alternatively, Ser559 was shown to be a relatively minor phosphorylation site that accounts for approx. 25% of total phosphorylation of the C-terminal tail. As expected, activation of the MCOLN1 S557A/S559A double mutant was not observed in response to H89 treatment, indicating that these residues mediate PKA-induced regulation of MCOLN1.

Numerous studies have suggested an important role for protein kinases in the regulation of TRP channels. In most cases, phosphorylation leads to an increase in the activity of the channel. For example, phosphorylation of TRPM4 (TRP melastatin 4) by PKC increases the activity of the channel [32], whereas Src phosphorylation enhances the activity of TRPM7 [33]. Within the TRPV (TRP vanilloid) family, it has been demonstrated that PKA, PKC and CaMKII can activate TRPV1 [34]. In addition, the channel activity of TRPC3 (TRP canonical 3) and TRPC6 is increased after phosphorylation by Src and Fyn respectively [35]. However, there are also examples that show that phosphorylation inactivates some TRP channels. For example, the activity of TRPC4, TRPC5 and TRPC6 may be inhibited by PKC-mediated phosphorylation [36].

The mechanism by which phosphorylation regulates MCOLN1 is unclear. Some ion channels regulated by phosphorylation contain specific regulatory domains, the best known example of which is the R domain of CFTR (cystic fibrosis transmembrane conductance regulator) (reviewed in [37]). No such domains are evident in the region of MCOLN1 that we identified as containing the MCOLN1 phosphorylation site. The fact that the MCOLN1 phosphorylation domain is located far from the pore region makes it unlikely that phosphorylation directly affects ion permeation through the MCOLN1 pore. It is possible that phosphorylation induces a conformational change in the MCOLN1 C-terminal, which changes its interaction with other proteins, such as the MCOLN1 subunits, or perhaps with other TRP channels. It should be noted that several TRP channels are known to homomultimerize and heteromultimerize, and recent results from Montell and colleagues show that MCOLN1 is no exception [38]. MCOLN1 homomultimerization seems to be very dynamic as, for example, pH changes were shown to reorganize MCOLN1 conductance substates [17,19] and induce structural changes consistent with the disorganization of MCOLN1 homomultimeric complexes. Heteromultimerization of TRP channels have been shown to dramatically affect the selectivity and activity of TRP channels [39]. Therefore it is possible that MCOLN1 phosphorylation may affect the interaction of MCOLN1 subunits, or perhaps influence its interaction with other TRP channels.

In many cases, the activation of TRP channels is accompanied by changes in their cellular distribution such as translocation from internal vesicles to the plasma membrane (reviewed in [40]). However, mutation of Ser557 and Ser559 did not change the distribution of MCOLN1, indicating that PKA-mediated phosphorylation does not play a role in the regulation of MCOLN1 trafficking to lysosomes.

One attractive possibility is that phosphorylation at the N-terminal regulates MCOLN1 trafficking. Interestingly, sequence searches performed using the NetPhosK 2.0 program predicted a potential PKA phosphorylation site at Ser10. This residue is especially interesting as a result of its proximity to the acidic di-leucine motif (S10ETERLL16) that regulates direct trafficking of MCOLN1 from the TGN to late endosomes through interactions with clathrin adaptors [24]. However, mutation of Ser10 to an alanine residue did not affect the levels of phosphorylation of the N-terminal tail or distribution of the full length protein, suggesting that this residue does not regulate MCOLN1 trafficking (see Supplementary Figures 3 and 4 at http://www.BiochemJ.org/bj/410/bj4100417add.htm).

PKA plays an important regulatory role in several anterograde traffic pathways, including transportation between the ER and the Golgi, intra-Golgi transport and budding of constitutive vesicles from the TGN to the plasma membrane [41,42]. In addition, PKA has also been implicated in retrograde transport from endosomes to the Golgi [43] and from the Golgi to the ER [44]. Our results suggest that PKA could also be involved in the regulation of lysosomal function through its ability to modulate the activity of MCOLN1.

One of the characteristics of patients with MLIV is the presence of achlorhydria with elevated plasma gastrin levels [3,45], as well as the accumulation of enlarged vacuoles in the parietal cells of the gastric mucosa [46]. Treatment of animals with omeprazole, an inhibitor of the H+/K+–ATPase, causes similar abnormalities to those observed in MLIV patients in parietal cells, suggesting that the absence of MCOLN1 could affect the activity or distribution of the H+/K+–ATPase [47]. Interestingly, acid secretion into the lumen of the stomach requires translocation of the H+/K+–ATPase from vacuoles to the plasma membrane, and this process is dependent on PKA [48]. Lysosomal exocytosis has also been shown to be impaired in MLIV fibroblasts [22] and, at the same time, PKA is known to play a critical role in regulated exocytosis [49]. Therefore lysosomal exocytosis and acid secretion are examples of processes that may be dependent on both MCOLN1 and PKA. Further studies will help to determine the role of PKA-mediated modulation of MCOLN1 in those processes listed above and other traffic events.

Acknowledgments

We thank the NIH (National Institutes of Health) Fellows Editorial Board for their editorial assistance prior to acceptance of this manuscript. This project was supported by funding from the Intramural Research Program of the NHLBI (National Heart, Lung, and Blood Institute) to R. P., and by funding from the Mucolipidosis 4 Foundation to K. K.

Abbreviations: BIS, bisindolylmaleimide I; CaMKII, Ca2+/calmodulin-dependent protein kinase II; CTail, C-terminal tail; DMEM, Dulbecco's modified Eagle's medium; DTT, dithiothreitol; ER, endoplasmic reticulum; FSK, forskolin; GFP, green fluorescent protein; GST, glutathione transferase; HEK-293, human embryonic kidney; LAMP, lysosome-associated membrane protein; MCOLN1, mucolipin 1; MLIV, mucolipidosis type IV; NRK cell, normal rat kidney cell; NTail, N-terminal tail; PKA, protein kinase A; PKC, protein kinase C; TGN, trans-Golgi network; TRP, transient receptor potential; TRPC, TRP canonical; TRPM, TRP melastatin; TRPV, TRP vanilloid; YFP, yellow fluorescent protein

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 26a.
  28. 26b.
  29. 27.
  30. 28.
  31. 29.
  32. 30.
  33. 31.
  34. 32.
  35. 33.
  36. 34.
  37. 35.
  38. 36.
  39. 37.
  40. 38.
  41. 39.
  42. 40.
  43. 41.
  44. 42.
  45. 43.
  46. 44.
  47. 45.
  48. 46.
  49. 47.
  50. 48.
  51. 49.
View Abstract