Identification of Sidt2 as a lysosomal cation-conducting protein

Abstract

A screen to identify lysosomal-expressed ion channels led to the discovery of the human Sidt2 protein. Sidt2 is expressed within lysosomal organelles but as a result of heterologous overexpression the protein is also detectable within the plasma membrane of human embryonic kidney cells. The overexpressed protein leads to cell depolarization upon sodium addition. Accordingly in whole-cell patch clamp experiments a spontaneous noninactivating monovalent cation current can be detected in Sidt2-overexpressing cells. Strong overexpression of Sidt2 in HEK293 cells is attended by a significant reduction/loss of detectable lysosomes, indicating that the overexpressed protein leads to lysosomal dysfunction, a hallmark of Alzheimer's disease. Sidt2 is located on chromosome 11q23, a locus repeatedly found by chromosomal mapping of Alzheimer's disease-related genes.

Abbreviations

CFTR, cystic fibrosis transmembrane regulator

EST, expressed sequence tag

GFP, green fluorescent protein

mRFP, monomeric red fluorescent protein

NAADP, nicotinic acid adenine dinucleotide phosphate

The action of the vacuolar H⁺-ATPase leads to the acidification of lysosomes as a mandatory step for the activation of a number of hydrolytic enzymes which in turn contribute to the turnover of macromolecules [1-3]. The electrogenic action of the ATPase evokes a huge potential across the lysosomal membrane limiting its own activity and thus has to be counteracted by a transporter(s) or an ion channel(s) providing a counter ion flux [4-11]. Potential counter ions could be negatively charged ions if transferred into the lysosomes or positively charged ions transported from the lysosomal lumen into the cytosolic compartment of the cell. Based on knock-out studies the lysosomal proteins CLC-7, a Cl⁻/H⁺ antiporter, as well as cystic fibrosis transmembrane conductance regulator (CFTR) were already excluded as candidate proteins [12-16]. Furthermore, direct measurement of the lysosomal acidification in the absence of Cl⁻ suggests that the counter ion is a cation [15]. A screen to identify potential ion channels led to the identification of Sidt2, a multimembrane-spanning protein which we cloned from human placenta. Another aspect which underlines the importance of Sidt2 is the expression of Sidt2 within lysosomes [17, 18]. Jialin and coworkers demonstrated the localization of endogenously expressed Sidt2 in lysosomal membrane fractions of a number of different cell types. The multimembrane-passing structure taken together with the broad expression pattern and the localization within lysosomal membranes prompted us to clone the Sidt2 cDNA. The derived amino acid sequence reflects a putative membrane protein with 832 amino acid residues and a calculated mass of ~94 kDa and resembles the sequence published by Jialin and coworkers [17]. Most likely, the protein sequence starts with a signal peptide that is cleaved after amino acid 18 resulting in a mature protein of 814 amino acids. Occurrence of a signal peptide taken together with multiple transmembrane regions typically indicates that the mature protein is located within the plasma membrane or the endoplasmic reticulum [19].

Several protein topology prediction programs identified 8–11 membrane-spanning regions of the Sidt2 protein. The human Sidt2 protein is similar to human Sid1 in terms of sequence identity (57% identity) and membrane topology. Feinberg and Hunter presented experimental data of a Caenorhabditis elegans homolog Sid1 and showed that the protein exhibits most likely 11 transmembrane regions with an extracellular N terminus and the C terminus facing the cytoplasm [20]. The sequence identity (identical residues) of human Sidt2 and C. elegans Sid1 is only 21%, one would expect a similar membrane topology of the human Sidt2 protein from hydropathy analysis [21]. In this paper, we show that the Sidt2 protein conducts cations and may be involved in the counter ion flux in lysosomes.

Material and methods

Cloning of human Sidt2

Sidt2 was amplified from human placenta-derived first-strand cDNA with phusion polymerase (NEB, Ipswich, MA, USA) using the primer pair 5′ccgccgccaccatgcgcggctgcctgcg3′ and 5′tcagaagacagggatctgg3′ according to the manufactures protocol. The cDNA was cloned into the EcoRV site of the pCAGGS-IRES-GFP and the pCAGGS-IRES-mRFP expression vectors allowing simultaneous expression of Sidt2 and GFP/mRFP [22]. For subcellular expression Sidt2 cDNA was amplified without the stop codon and cloned into the EcoRV site of the pCAGGS vector fused to eGFP and mRFP allowing expression of Sidt2-eGFP and Sidt2-mRFP fusion proteins, respectively. The cloned Sidt2 cDNA is identical to the sequence with the accession number Q8NBJ9.

Lysosome staining

Lysosomes were visualized by overexpression of a fusion construct of lamp1-eGFP or the two-pore channel TPC1-eGFP in HEK293 cells. In addition, we used a stable cell line which expresses the lysosomal two-pore channel TPC2 fused to eGFP (TPC2-eGFP; [23]. Lysosomes were also stained with LysoTracker Red DND-99 (ThermoFisher, Dreieich, Germany).

Stable cell lines

Sidt2 was cloned in the pcDNA5 FRT/TO vector to create stable HEK293 Flp-In cells (Invitrogen, Carlsbad, CA, USA). The Sidt2 protein expression was induced due to addition of tetracycline (1 μg·μL⁻¹). Cells were harvested 20–24 h after induction.

Expression analysis

Northern blots and dot blots were probed overnight with a [α-32P]-dCTP-labeled 1500 bp XhoI fragment of the Sidt2 cDNA in the presence of 50% formamide at 42 °C as described earlier [22]. Filters were exposed to imager screens (Fuji, Japan) over night.

Confocal and conventional fluorescence microscopy

For confocal microscopy Sidt2 was C-terminally fused to eGFP or to monomeric (m)RFP and transiently expressed in HEK293 cells. Confocal pictures were taken using a Zeiss LSM780 confocal microscope equipped with a 63× 1.4 Plan Apochromat objective. eGFP and mRFP were excited with the 488 nm and 543 nm lines of an Argon and a HeNe laser, respectively. Pictures were analyzed with the zen or axiovision software (Zeiss, Oberkochen, Germany) and processed (merged) with ImageJ (NIH, Bethesda, MD, USA). Fluorescence pictures were taken at an Axio Observer Z1 microscope (Zeiss) equipped with a 63× Plan Apochromat objective (Zeiss), a HXP 120 C lamp (Zeiss) and an Axiocam color CCD camera (Zeiss), using either a GFP (excitation 470/40 nm, dichroic mirror 495 nm, emission >500 nm) or RFP (excitation 560/40 nm, dichroic mirror 590 nm, emission >590 nm) filter set (AHF Analysetechnik AG, Tübingen, Germany). Fluorescence pictures were analyzed and processed (merged) using the imagej software (NIH).

Fluorescent membrane potential measurements

HEK293 cells were transiently transfected with 4 μg of plasmid using the transfection reagent Fugene (Promega, Madison, WI, USA) and plated on poly-l-lysine coated cover slips. Cells were transfected with Sidt2-pCAGGS-IRES-GFP or as control with the empty pCAGGS-IRES-GFP vector and analyzed 48 h post-transfection. Transfected cells were incubated for 10 min in FLIPR membrane potential (FMP) dye (Molecular Devices, Sunnyvale, CA, USA) solution at room temperature. The membrane potential-sensitive dye FMP was dissolved in physiological buffer containing (in mm) 115 NaCl, 2 MgCl₂, 5 KCl 10 HEPES, pH 7.4. For measurements in sodium- and chloride-free conditions, NaCl was substituted by NMDG (N-Methyl-D Glucamine)-Cl and Na-Gluconate, respectively. FMP measurements were performed at an Axiovert S100 microscope (Zeiss) equipped with a 20× Fluar objective (Zeiss), a polychrome V monochromator (TILL Photonics, Martinsried, Germany), and a TILL IMAGO CCD camera (TILL Photonics). FMP was excited by 530 nm light and fluorescence was detected using a customized filter block with excitation filter 534/20 nm, dichroic mirror 565 nm, and emission filter 605/70 nm (AHF Analysentechnik AG) as described previously [24]. Solutions were applied by bath application. FMP fluorescence was measured using the tillvision software (TILL Photonics) and data were further analyzed with the igor pro 6 software (WaveMetrics, Portland, OR, USA). Results are given as mean (±SEM) absolute FMP fluorescence, plotted versus time, with x indicating the number of measured cells in y independent experiments (x/y).

Electrophysiology

Whole-cell currents from HEK cells transiently transfected for 24–48 h with either Sidt2 cDNA or empty vector (see above) were recorded at an Axiovert 135 M microscope (Zeiss) equipped with a 470 nm LED (Rapp OptoElectronics, Hamburg, Germany) and a GFP filter set (see above), using a computer-controlled EPC-9 patch clamp amplifier (HEKA Electronics, Lambrecht, Germany) and the patchmaster software (HEKA). Positively transfected cells were identified by the expression of GFP. Patch pipettes were pulled from glass capillaries GB150T-8P (Science Products, Hofheim, Germany) at a PC-10 micropipette puller (Narishige, Tokyo, Japan) and had resistances between 2 and 4 MΩ when filled with internal (pipette) solution containing (in mm) 120 Na-aspartate, 8 NaCl, 1 MgCl₂, 10 HEPES, 10 Na-BAPTA (1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid); pH adjusted to 7.2 with NaOH. 3.1 mm CaCl₂ were added to reach a free intracellular Ca²⁺ concentration of 100 nm (http://web.stanford.edu/~cpatton/webmaxcS.htm). The external solution comprised (in mm) 136 Na-aspartate, 4 NaCl, 2 MgCl₂, 1 CaCl₂, 10 HEPES, 10 glucose; pH adjusted to 7.2 with NaOH. For application of sodium-free saline Na⁺ was replaced by NMDG and administered directly onto the patch clamped cell via a wide-tipped patch pipette. For detection of cesium outward currents intra and extracellular Na-aspartate was replaced by Cs-glutamate and NaCl, respectively. Voltage ramps (400 ms) spanning from −100 to +100 mV were applied every 2 s from a holding potential (V_h) of 0 mV. Currents were filtered at 2.9 kHz and digitized at 400 μs intervals. Inward and outward currents were extracted at −80 mV and +80 mV, respectively, from the individual ramp current recordings and plotted versus time. Representative current–voltage relationships (IVs) were extracted at certain times as indicated.

All solutions had an osmolality of approximately 300 mosmol·kg⁻¹. Capacitive currents and series resistance were determined and corrected before each voltage ramp. Currents were normalized to the cell capacitance, which was extracted as a representative measure for the cell size, to calculate current densities (pA/pF). Results are given as mean (±SEM) inward and outward currents, respectively, plotted versus time, with n indicating the number of measured cells. The current voltage relationships represent means without SEM from n measured cells.

Biotinylation assay

Biotinylation was performed as described previously [25]. A Flp-In HEK293-derived cell line expressing Sidt2 under the control of a tetracycline-inducible promotor was established (see above). Seventy-five-milliliter flasks with confluent Sidt2-Flp-In cells, either induced with 1 μg·μL⁻¹ tetracycline or noninduced (control), were washed with saline and incubated with NHS-LC-biotin freshly dissolved in PBS buffer including 0.1% (w/v) bovine serum albumin, washed with PBS and incubated in the presence of a mixture of protease inhibitors. The protein solution was incubated with avidin-agarose beads for 2 h at 4 °C, washed and subjected to SDS page, blotted and analyzed using a commercial Sidt2 specific polyclonal antibody (N-20, Santa Cruz Biotech. Inc., Dallas, TX, USA).

Data analysis

An unpaired two-tailed Student′s t-test (Excel, Microsoft) was used to test two populations of data for a possible significant difference (***P < 0.001).

Results

Tissue expression of Sidt2

Lysosomal acidification is influenced by a yet unknown ion transporter(s) and/or ion channel(s) which possibly transfer cations from the luminal side of lysosomes to the cytosolic side of the cells. We tried to identify candidate proteins which could fulfill this function. Blast search using known protein sequences of ion channels from several families including TRP-, potassium-, and anoctamin-channels led to the identification of a putative protein which exhibits several transmembrane regions and which is called Sidt2 or Sid1 member 2 [26].

Sidt2 shows a broad expression pattern and could be detected in all tested tissues including spleen, thymus, prostate, testis, ovary, small intestine, colon, leukocytes, heart, brain, placenta, lung, liver, skeletal muscle, kidney, and pancreas as depicted in Fig. 1A. An independent analysis of the expressed sequence tag (EST) database also revealed the almost ubiquitous expression of Sidt2-related sequences. The broad expression pattern was also observed using a mRNA dot blot (Fig. 1B). Sidt2 expression is most abundant in mRNAs derived from spinal cord (B7), prostate (C7), stomach (C8), testis (D1), pancreas (D3), salivary gland (D7), and placenta (F4). Jialin and coworkers [17] could not detect Sidt2 in skeletal muscle and heart in rat tissue by western blotting although in the EST database from rat several skeletal muscle-derived Sidt2-related EST clones can be identified including CR756086. On the other hand we found Sidt2-related EST clones derived from heart in the human EST database like W68674 but no clones in the rat EST database. The discrepancy is not clear but in case of the heart we cannot exclude species-specific expression of Sidt2.

Figure 1

Expression pattern of Sidt2. (A) Northern blot from diverse murine tissues was probed with a 1.5 kb Xho I fragment of the Sidt2 cDNA. (B) Dot blot probed with same probe as in (A): A1–8 brain, amygdala, caudate nucleus, cerebellum, frontal lobe, hippocampus, medulla oblongata; B1–7 occipital lobe, putamen, substantia nigra, temporal lobe, thalamus, subthalamic nucleus, spinal cord; C1–8 heart, aorta, skeletal muscle, colon, bladder, uterus, prostate, stomach; D1–8 testis, ovary, pancreas, pituitary gland, adrenal gland, thyroid gland, salivary gland, mammary gland; E1–8 kidney, liver, small intestine, spleen, thymus, peripheral leukocytes, lymph node, bone marrow; F1–4 appendix, lung, trachea, placenta; G1–7 fetal tissues: brain, heart, kidney, liver, spleen, thymus, lung; H1–8 yeast RNA, yeast tRNA, Escherichia coli rRNA, E. coli DNA, poly r(A), human c₀t 1 DNA (100 ng), human DNA (100 ng), human DNA (500 ng). C Expression of Sidt2 in carcinoma tissue compared to the corresponding healthy tissue. In many carcinomas, Sidt2 seems to be downregulated—for a detailed description see Fig. S1.

Apparently, in many carcinomas Sidt2 appeared to be downregulated on the transcription level compared with the corresponding healthy tissue (Fig. 1C; see Fig. S1 for detailed description). Sidt2 seems to be downregulated in tumors derived for example from adrenal tissue (A3 tumor/A4 healthy), bladder (B1 tumor/B2 healthy), breast (B3, B5, C1 tumor/B4, B6, C2 healthy), prostate (I3 tumor/I4 healthy). In most of the tumors shown Sidt2 seems to be downregulated (for details see Fig. S1). In case of prostate cancer our data are in good agreement with data present in the GEO database. For example, the dataset GDS2546 contains tissue samples from normal prostate and metastatic prostate cancer and the data indicate that Sidt2 levels decrease in the metastatic samples of prostate cancer [27].

Sidt2 conducts sodium ions

We overexpressed Sidt2 in HEK293 cells using a bicistronic construct encoding the green fluorescent protein (GFP) in addition to Sidt2 and loaded the cells with the FLIPR membrane potential-sensitive fluorescent dye FMP. In physiological solution nominally free of calcium, Sidt2-expressing cells (red) exhibited a significantly higher basic FMP fluorescence suggesting a more depolarized membrane potential compared to just GFP-expressing HEK293 cells (black; Fig. 2A). The virtually identical traces of green (black) and nongreen (blue) cells from GFP-transfected control experiments show that the green fluorescence of GFP does not interfere with the fluorescent FMP measurement (Fig. 2A,B). The absolute FMP fluorescence depends on the concentration of the dye, thickness of the cells, and other parameters and is therefore not readily applicable as absolute indicator for the membrane potential [24]. We kept the conditions as consistent as possible to minimize the experimental differences. However, switching the extracellular condition to 25 mm potassium chloride leads to robust and reproducible fluorescence changes of both Sidt2 and control cells as a result of less potassium efflux (Fig. 2A), indicating the depolarization of these cells [24]. The already basically increased FMP fluorescence in Sidt2-expressing cells suggests that a constitutive active mechanism (ion channel or electrogenic transporter) leads to a more depolarized membrane potential. Such a depolarization can be mediated by efflux of negative ions and/or influx of positive ions. Changing the extracellular Cl⁻ concentration or using chloride channel inhibitors such as niflumic and flufenamic acid did not affect the FMP fluorescence, that is, membrane potential of the Sidt2-expressing cells (data not shown), thus, we next tested a sodium addition protocol. Sidt2-expressing cells were FMP-loaded and measured in NMDG-chloride-based Na⁺-free external solution (Fig. 2B). Sodium chloride was added at the indicated time point without changing the concentration of any other ion thereby keeping the osmolality constant. Sidt2-expressing cells show instantly an increase in the FMP fluorescence, that is, membrane depolarization after sodium exposure indicating that Sidt2-expressing cells depolarize by sodium influx (Fig. 2B, red). The experiment shows that Sidt2-expressing cells exhibit an active additional sodium permeability which could not be seen in just GFP-expressing control cells (Fig. 2B, black). A small but significant fluorescence increase after sodium addition was also detectable in non-green cells taken from Sidt2-transfected dishes (Fig. 2B, green). This increase may reflect that some of the nongreen cells already express Sidt2 before developing a detectable GFP fluorescence. Adding 25 mm potassium chloride to Sidt2-, GFP-, or nongreen cells of both transfections exhibited an increase in the fluorescence, that is,depolarization, which was very similar in amplitude in all cells independent of Sidt2 expression. Taken together the data indicate that Sidt2 permeates sodium ions.

Figure 2

Overexpression of Sidt2 in HEK cells leads to depolarized membrane potentials. A Sidt2-IRES-GFP construct was overexpressed in FMP-loaded HEK293 cells. FMP was excited by 530 nm light and (absolute) FMP fluorescence was measured above 605 nm using a FMP filter [24]. (A) Cells were incubated in Ca²⁺-free physiological solution containing (in mm) 115 NaCl, 2 MgCl₂, 5 KCl, 10 HEPES, pH 7.4, and 25 mm KCl was added as indicated. Sidt2-expressing cells (red) revealed a much higher basic fluorescence, that is, a more positive membrane potential, than just GFP-expressing HEK cells (black). (B) Cells were incubated in Ca²⁺- and Na⁺-free saline containing (in mm) 115 NMDG-Cl, 2 MgCl₂, 5 KCl, 10 HEPES, pH 7.4. Addition of 75 mm sodium resulted in a significant fluorescence increase (i.e. depolarization) in Sidt2 (red) but not in just IRES-GFP (black) -transfected cells. All cells showed a similar fluorescence increase after depolarization with 25 mm potassium. Nongreen cells from both transfections (nongreen Sidt2 – green, nongreen GFP – blue) behave similar as just GFP-expressing cells (A and B; blue traces behind black traces). Data represent means ± SEM of x cells from y experiments (x/y).

A bicistronic Sidt2-IRES-GFP construct was analyzed by whole-cell patch clamp in HEK293 cells 2 days after transfection. Voltage ramps from −100 to +100 mV were applied every 2 s from a holding potential of 0 mV. Cells were kept in Na-aspartate-based solution and the extracellular solution was switched to NMDG-aspartate (NMDG) as indicated (Fig. 3A). Already upon break-in Sidt2-expressing cells (red) revealed a constitutive higher inward and outward current compared to mock-transfected control cells (black). After extracellular sodium depletion (NMDG) the inward current significantly decreased. The corresponding voltage relationships are shown in Fig. 3B. The data reveal that in the presence of Sidt2, a noninactivating inward and outward current is present in HEK293 cells, which is not present in the control cells. This current is small (about 8 pA/pF at −80 and +80 mV) and the inward part is not present in the absence of sodium ions. In good agreement with the data from the FMP imaging (see above), the whole-cell patch clamp experiments also reveal that Sidt2-expressing HEK293 cells express a constitutive sodium conductance which is not present in control cells. To test whether Sidt2 also conducts other cations we replaced intracellular Na-aspartate by Cs-glutamate. As shown in Fig. 3F,G Sidt-2-expressing cells reveal also significant cesium-dependent outward currents.

Figure 3

Sidt2 overexpression in HEK cells generates a cation conductance. (A) Inward and outward currents measured at −80 and 80 mV, respectively, from 400 ms voltage ramps spanning from −100 to 100 mV (0.5 Hz; holding potential 0 mV) in HEK293 cells transiently transfected (48 h) with the bicistronic Sidt2-IRES-GFP construct (red) or just IRES-GFP (black). Experiments started with external saline based on Na-Aspartate and Na⁺-free NMDG-Aspartate (NMDG) was applied as indicated. (B) depicts the mean current voltage relationships (IVs) in sodium- and NMDG-based saline. Note, that Sidt2-expressing cells revealed a significant current already at break-in and that the inward part disappeared after removal of external sodium. (C) HEK293 cells were transiently transfected (48 h) with cDNA for Sidt2-mRFP (red) and lamp1-eGFP (green). (D) HEK293 cells transiently expressing (48 h) only lamp1-eGFP (green). Note that cells strongly expressing Sidt2 become more round shaped and smaller in size (circled cell; see also Figs 4 and 5), also shown by measurements of the cell capacitance of Sidt2- and just GFP-expressing HEK2903 cells in (E) (***P < 0.001). (F) Inward and outward currents in the presence of intracellular Cs-glutamate and extracellular NaCl in HEK293 cells transiently transfected (48 h) with the bicistronic Sidt2-IRES-GFP construct (red) or just IRES-GFP (black). (G) depicts the mean current voltage relationships (IVs) from the currents in (F). Data in (A), (E) and (F) represent means ± SEM from as many measured cells as indicated in brackets.

Sidt2 transfection affects cell shape and cell size

Since Sidt2 is shown to be endogenously expressed in lysosomes [17], we coexpressed the fusion constructs Sidt2-mRFP and lamp1-eGFP, a lysosomal membrane marker, to test whether Sidt2-mRFP also localizes in the lysosomes after heterologous expression. We noticed that low Sidt2-mRFP overexpression revealed a punctuate pattern, which obviously did not really colocalize with lamp1-eGFP (Fig. 3C). In addition, cells which strongly express Sidt2-mRFP seemed to reveal less punctuate lamp1-eGFP pattern (will be discussed below) and apparently become more round shaped and smaller (see circled cell, Fig. 3C). For comparison, Fig. 3D shows a HEK293 cell just expressing lamp1-eGFP with clear punctuate lysosomal staining. As a measure for the cell size, the cell capacitance of the Sidt2-transfected cells was detected in the whole-cell configuration and was found to be significantly lower than in control cells. The average cell capacitance of Sidt2 versus GFP-transfected HEK293 was calculated to 13.9 ± 0.5 (n = 67) versus 9.4 ± 0.5 (n = 91) pF indicating that the cell surface of Sidt2 cells is about 33% smaller than in just GFP-expressing HEK293 cell (Fig. 3E).

Sidt2 is present in the cell membrane after overexpression

Jialin et al. [17] showed a prominent localization of Sidt2 in lysosomal organelles of many different cell types. We could detect Sidt2-related membrane depolarization and membrane currents in Sidt2-expressing HEK293 cells by fluorescent FMP imaging and whole-cell patch clamp, respectively. Therefore, one would assume to find Sidt2 in the plasma membrane of the transfected cells. We established a stable cell line that expresses Sidt2 under the control of a tetracycline-inducible promotor. Using a biotinylation assay we could show, that overexpressed Sidt2 is indeed trafficked to the plasma membrane (Fig. 4A). Calnexin, which is an ER protein, was used as the control, and was not detectable within the membrane fraction. Thus, Sidt2, although endogenously expressed in lysosomes, was found in the plasma membrane after overexpression. To test if Sidt2 can be detected in the plasma membrane by optical methods we used the Sidt2 construct C-terminally fused to the monomeric red fluorescent protein (mRFP). Sidt2-mRFP-transfected HEK293 or COS cells showed little to no plasma membrane-associated fluorescence (Fig. 4B, C; see also Figs 3C and 5B,C). Most of the fluorescence signal was located within the cells. In contrast, TMEM16a (anoctamin 1) a calcium-activated chloride channel, fused to eGFP, was found predominantly in the plasma membrane (Fig. 4B). Although most of the Sidt2 protein is localized in intracellular compartments we were able to detect Sidt2 in the plasma membrane by biotinylation (see above, Fig. 4A). Thus, plasma membrane localization of Sidt2 seems to be an effect of heterologous overexpression.

Figure 4

After overexpression, some Sidt2 localize to the plasma membrane. (A) A stable HEK293 cell line was generated, which expresses Sidt2 under the control of a tetracycline-inducible promotor. Biotinylated proteins were bound to avidin and analyzed with a Sidt2-directed antibody. Lane 1—input noninduced cells; lane 2 –biotinylated membranes from noninduced cells; lane 3—input from induced cells; lane 4—biotinylated membranes from induced cells; below—incubation with calnexin (Cal, ER protein) antibody as control. (B) shows confocal images of HEK293 cells transiently coexpressing (48 h) Sidt2-mRFP (red) and TMEM16a (Anoctamin1)-eGFP (green) fusion constructs. (C) shows fluorescence images of the same constructs as shown in (B) transfected in COS cells. Note that some of the Sidt2-staining (red) can be localizes at the plasma membrane together with the membrane marker TMEM16a (merged image, right).

Figure 5

Reduced detection of lysosomes in HEK293 cells strongly overexpressing Sidt2. (A and B) show confocal pictures of HEK293 cells transiently expressing (48 h) Sidt2-eGFP (green) costained with LysoTracker Red (red, A) or transiently coexpressing (48 h) the fusion products Sidt2-mRFP (red) and the two-pore channel TPC1-eGFP (green, B). (C and D) show confocal images of HEK293 cells stably expressing the fusion product two-pore channel TPC2-eGFP (green) and transiently (48 h) the fusion product Sidt2-mRFP (red, C) or the bicistronic construct Sidt2-IRES-mRFP (red, D). The last column depicts the merges. Note that intensively green Sidt2-expressing cells exhibit low LysoTracker staining (see circled cell in A), that there is virtually no colocalization of Sidt2 and TPC1 (B), and that the punctuate pattern in TPC2-eGFP-expressing cells becomes fuzzy if the cells additionally express Sidt2-mRFP or Sidt2-IRES-mRFP (C and D). All pictures are taken with the same magnification (see scale bar in A).

Sidt2 overexpression leads to a decrease in lamp1 and LysoTracker detectable lysosomes

Recently, Jialin and coworkers published that the endogenously expressed Sidt2 protein is localized in membranes of lysosomes [17]. Therefore, plasma membrane localization of Sidt2 might be an effect of the heterologous overexpression in HEK293 cells. As shown in Fig. 3C overexpressed Sidt2 C-terminally fused to mRFP (Sidt2-mRFP) does not really localize with the lysosomes. However, to further test whether overexpressed Sidt2 is visible in lysosomes we transfected Sidt2-eGFP into HEK293 cells and stained these cells with LysoTracker-Red (DND-99, ThermoFisher). LysoTracker-Red stains only acidic compartments of the cell. However, LysoTracker staining was weak to nondetectable in cells expressing significant amounts of Sidt2 (Fig. 5A, see, for example, circled cell), whereas cells with low Sidt2 expression (weakly green) showed punctuate lysosomal LysoTracker staining as expected. Therefore, we asked whether the lysosomal pH in Sidt2 cells is not acidic anymore, but, as already shown in Fig. 3C, costaining of Sidt2-mRFP and lamp1-eGFP, a protein which is transported to lysosomal membranes independent of the luminal pH [28, 29], revealed the same results: No punctuate staining, that is, lysosomes were detectable in strongly Sidt2-expressing cells, whereas cells expressing low amounts of Sidt2-mRFP and control cells only transfected with lamp1-eGFP showed a number of lysosomes (Fig. 3C,D). The results show that negative LysoTracker staining of Sidt2-expressing cells does not depend on an inappropriate lysosomal pH but rather on dysfunction and loss of lysosomes in the Sidt2-transfected cells. Sidt2-mRFP and Sidt2-eGFP overexpressed in HEK293 cells does not punctually colocalize with the lysosomal markers lamp1-eGFP and LysoTracker-Red, respectively (Figs 3C and 5A). As a further control, no colocalization of the endosomal two-pore channel TPC1 [30] with Sidt2 could be detected, too (Fig. 5B).

To test whether the presence of overexpressed Sidt2 indeed leads to loss or partial loss of lysosomes we transfected Sidt2 into a cell line which stably expresses the lysosomal two-pore channel TPC2 fused to eGFP [23, 30]. This cell line shows a strong staining of the lysosomes visible as intensive punctuate pattern. This pattern became fuzzy or was not detectable in cells strongly overexpressing Sidt2 (Fig. 5C). Also transfection of a Sidt2-IRES-mRFP construct into the TPC2-eGFP stably expressing cell line resulted in a decrease of the lysosomal fluorescence in these cells showing that the wild-type construct also leads to similar results as the Sidt2-mRFP fusion construct (Fig. 5D).

The results show that Sidt2-mRFP and Sidt2-eGFP heterologously overexpressed in HEK293 cells reveal a punctuate pattern which does not significantly colocalize with typical lysosomal markers lamp1, LysoTracker Red, and TPC2 or the endosomal marker TPC1. However, strong overexpression of Sidt2 results in a dysfunction and loss of lysosomes. Whether heterologously overexpressed Sidt2-mRFP and Sidt2-eGFP only localizes in a subdetectable amount in lysosomes leading to their dysfunction and loss or whether Sidt2 overexpression affects the lysosomes from other cellular locations cannot be answered so far.

Evolutionary aspects of Sidt2 and its link to Alzheimer′s disease

Analysis of the NCBI database indicates that Sidt2 orthologous proteins are found in all sequenced mammalia. Sidt2 orthologs are not found in archaea bacteria, eubacteria, plants, and eukaryotic protozoa including Plasmodium, Euglena, Amoeba, and Giardia. In the green algae Chlamydomonas reinhardtii, several ion channel genes can be found like transient receptor potential TRP-type channels. TRP-type ion channels are typically found in animals and fungi, but not in bacteria, algae, and plants, however, in Chlamydomonas no Sidt2-related genes seem to be present which is consistent with the finding that these algae do not have lysosomes instead of a lytic vacuole. Furthermore, we did not observe Sidt2-related genes in most fungi including Yeast, Aspergillus, and Neurospora with the exception of the slime mold Dictyostelium discoideum. Jialin et al. [17] reported that Sidt2 is predominantly found in lysosomal membranes and accordingly Sidt2 homologous genes are only identifiable in organisms that have developed lysosomes or lysosome-like organelles as Dictyostelium discoideum [31]. The evolutionary distribution of Sidt2 genes reflects its involvement in lysosomal function.

Discussion

Lysosomal function depends on an acidic intralysosomal pH requiring a counter flux ion which in turn prevents the development of a huge lysosomal potential and thereby limiting the action of the lysomal H⁺/ATPase [4-11]. It has been shown that lysosomal membranes exhibit a Cl⁺/H⁺ antiport that is conducted by CLC-7, however, conflicting results leaving their role in the acidification process unclear [12, 14, 32]. Graves et al. [32] demonstrated a prominent role for CLC-7 in HeLa cells, whereas others could not find any modified lysosomal acidification in CLC-7-deficient background [12-15]. However, CLC-7 is a Cl⁻/H⁺ antiporter which appeared to be counterproductive in the acidification process [10, 32]. Other chloride channels of the same family (CLC-3, -4, -5, -6) were shown to be present in endosomes and responsible for Cl⁻ conductance in these organelles [33-39]. The role of another chloride channel present in lysosomal membranes, the cystic fibrosis transmembrane regulator (CFTR), in lysosomal acidification is also not clear. Di and coworkers [40] proposed a role for CFTR in alveolar macrophages, however, direct measurements of lysosomal pH in CFTR- and CLC-7-deficient background indicates that both channels do not contribute to lysosomal acidification [8]. The authors explained this contradiction by methodical pitfalls and suggested a role of cation channels supporting the lysosomal acidification [9]. In addition, it was demonstrated that lysosomes acidify normally in Cl⁻-free buffer pointing toward the possibility that the counter ions most probably are cations leaving the lysosomal lumen [15]. Steinberg and colleagues demonstrated that cations indeed support lysosomal acidification [8]. Endosomal/lysosomal organelles express the cation permeable two-pore channels, TPC1 and TPC2, but they were already excluded as counter ion candidates [41]. A possible candidate might be TRPML1, a member of the mucolipin TRP (transient receptor potential) ion channels, permeable to mono- and divalent cations, and primarily expressed in lysosomes [42]. However, mutations in TRPML1 are associated with type IV mucolipidosis and it is suggested that TRPML1 is important for Fe²⁺ release from late endosomes and lysosomes [43]. Here, we present another lysosomal protein, Sidt2, which, based on its properties, might be a further candidate for contributing to the H⁺-ATPase counter ion flux in lysosomes.

Several studies indicate that in addition to other factors, the dysfunction of lysosomes might have an impact on Alzheimer's disease [44-46]. Sidt2 is located on the human chromosome 11q23, a chromosomal locus that was identified in three independent studies as related to Alzheimer's disease [47]. In a meta-analysis Bertram and Tanzi analyzed 90 studies in which Alzheimer's disease-related genomic loci were published and recognized that three loci including 11q23 were repeatedly found in addition to the established apolipoprotein E and the presenilin loci. In this critical region there is also the Beta-site APP-cleaving enzyme 1 (BACE) located which exhibits a synonymous point mutation in exon 5, and which is an excellent candidate protein for Alzheimer′s disease based on its biochemical properties [45, 48]. Conflicting results have been published if the polymorphism of BACE is associated with Alzheimer's disease or not.

Surprisingly, a Sidt2-deficient mouse line showed only a mild phenotype characterized by increased glucose intolerance and decreased serum insulin level [49]. Sidt2-deficient mice express reduced levels of the SNARE proteins synap1 and synap3 indicating that Sidt2 and SNAREs are directly involved in insulin secretion [50]. Furthermore, the intracellular calcium signaling was changed in primary pancreatic β-cells, this effect could be reversed by bafilomycin connecting intracellular calcium signaling to lysosomes [51]. Application of nicotinic acid adenine dinucleotide phosphate (NAADP) rescued the calcium response although the islets of Sidt2-deficient mice synthesized NAADP normally. The authors concluded that in Sidt2-deficient islets the NAADP-dependent calcium traffic from the intracellular acidic compartment is impaired.

Recently it was published that Sidt2 is involved in the uptake of RNAs into lysosomes and seem to be part of the cellular RNA degradation system [52]. From that finding Sidt2 may act as a Na⁺/nucleic acid antiporter. However, from its function, cellular localization and evolutionary distribution Sidt2 is a candidate protein that potentially contributes to the lysosomal counter flux.

Acknowledgements

The authors thank Christine Wesely for western blotting, Dr. Norbert Klugbauer (Experimentelle und Klinische Pharmakologie und Toxikologie, Universität Freiburg, Freiburg, Germany) for the TPC1 cDNA, and Dr. Martin Biel and Dr. Chen-Chang Chen (both Fakultät für Chemie und Pharmazie, LMU München, München, Germany) for providing the stable TPC2-expressing cell line. The authors also thank Heidi Löhr, Ramona Grünewald, and Ute Soltek for cell culture and establishing the stable cell lines. This work was supported from the Deutsche Forschungsgemeinschaft (DFG) through grants to AB and VF (SFB894, GRK1326).

Author contributions

AB performed electrophysiological measurement, CF-T performed biotinylation experiments and western blots, KW cloned Sidt2 fusion constructs, SEP constructed stable cell lines, VF critically read the manuscript and UW cloned Sidt2, performed FMP measurements and wrote the manuscript.