Research ArticleINTERORGANELLE SIGNALING

Polycystin 2 regulates mitochondrial Ca2+ signaling, bioenergetics, and dynamics through mitofusin 2

See allHide authors and affiliations

Science Signaling  07 May 2019:
Vol. 12, Issue 580, eaat7397
DOI: 10.1126/scisignal.aat7397

PC2 separates mitochondria from the ER

Patients with loss-of-function mutations in polycystin (PC) 1 or 2 develop fluid-filled cysts due to excessive proliferation of kidney epithelial cells. Kuo et al. found that loss of the ER cation channel PC2 led to increased abundance of the mitochondrial fusion factor MFN2 and enhanced tethering of mitochondria to the ER. The increased mitochondria-ER association resulted in greater mitochondrial Ca2+ influx, biogenesis, and respiration and cellular proliferation, which in cultured cells and mouse models of polycystic kidney disease was rescued by deficiency in MFN2. These results show that PC2 acts to restrict mitochondrial tethering to the ER in kidney cells to prevent inappropriate Ca2+-dependent increases in mitochondrial function and cellular proliferation.

Abstract

Mitochondria and the endoplasmic reticulum (ER) have an intimate functional relationship due to tethering proteins that bring their membranes in close (~30 nm) apposition. One function of this interorganellar junction is to increase the efficiency of Ca2+ transfer into mitochondria, thus stimulating mitochondrial respiration. Here, we showed that the ER cation-permeant channel polycystin 2 (PC2) functions to reduce mitochondria-ER contacts. In cell culture models, PC2 knockdown led to a 50% increase in mitofusin 2 (MFN2) expression, an outer mitochondrial membrane GTPase. Live-cell super-resolution and electron microscopy analyses revealed enhanced MFN2-dependent tethering between the ER and mitochondria in PC2 knockdown cells. PC2 knockdown also led to increased ER-mediated mitochondrial Ca2+ signaling, bioenergetic activation, and mitochondrial density. Mutation or deletion of the gene encoding for PC2 results in autosomal dominant polycystic kidney disease (ADPKD), a condition characterized by numerous fluid-filled cysts. In cell culture models and mice with kidney-specific PC2 knockout, knockdown of MFN2 rescued defective mitochondrial Ca2+ transfer and diminished cell proliferation in kidney cysts. Consistent with these results, cyst-lining epithelial cells from human ADPKD kidneys had a twofold increase in mitochondria and MFN2 expression. Our data suggest that PC2 normally serves to limit key mitochondrial proteins at the ER-mitochondrial interface and acts as a checkpoint for mitochondrial biogenesis and bioenergetics. Loss of this regulation may contribute to the increased oxidative metabolism and aberrant cell proliferation typical of kidney cysts in ADPKD.

INTRODUCTION

Mitochondrial energy production requires a continuous supply of Ca2+ to maintain oxidative metabolism through regulation of Ca2+-dependent enzymes in the tricarboxylic acid cycle (1, 2). However, the amount of Ca2+ entering mitochondria must be precisely regulated because mitochondrial Ca2+ overload can lead to apoptosis through increased mitochondrial membrane permeabilization, depolarization, and opening of the permeability transition pore (36).

Under nonpathological conditions, Ca2+ flux occurs between the endoplasmic reticulum (ER) and mitochondria at a region that forms close physical contacts (~30 nm) between these organelles, called the mitochondria-associated ER membrane (MAM) (3, 7, 8). The MAM effectively integrates signal transduction with metabolic pathways to regulate communication between the ER and mitochondria. Regulated transfer of Ca2+ from the ER to mitochondria occurs through opening of the inositol 1,4,5-trisphosphate receptor (InsP3R) on the ER membrane and the mitochondrial Ca2+ uniporter (MCU) complex on the inner mitochondrial membrane (3, 9, 10). Specifically, the type 3 InsP3R (InsP3R3), but not InsP3R1 (11), is reported to localize primarily at the MAM to coordinate Ca2+ transfer to mitochondria.

In many disease states, including Alzheimer’s disease (12) and diabetes (13), the contacts between the ER and mitochondria and Ca2+ transfer between these organelles are altered, ultimately affecting cellular metabolism. The polycystin proteins, particularly polycystin 1 (PC1), have emerged as modulators of mitochondrial metabolism. Loss-of-function mutations in the polycystin genes PKD1 and PKD2 result in autosomal dominant polycystic kidney disease (ADPKD), a disease for which there is no cure. There is a notable change in the metabolism of cystic epithelial cells (14, 15), and parallels between cystogenesis and cancer development have been noted (14). Loss of PC1 results in a shift to glycolytic metabolism, and treatment with 2-deoxyglucose, an inhibitor of glycolysis, reduces renal cysts in PC1-deficient mice (1517). In addition, the polycystin complex (composed of PC1 and PC2) is sensitive to oxygen concentrations, suggesting an ability to sense environmental conditions and regulate cellular metabolism (18). However, studies investigating whether PC2 depletion specifically alters mitochondrial function or Ca2+ signaling have been limited (19). Most of the protein product of PKD2, PC2, is found on the ER of epithelial cells, with the remainder localized to the primary cilium (20, 21). In contrast to the largely plasma membrane–localized PC1, PC2 can act as a Ca2+-dependent Ca2+ release channel in the ER membrane and can also interact with and regulate the release of ER Ca2+ via the InsP3R (2224). Loss-of-function mutations in PC2 disrupt downstream Ca2+ signaling, leading to decreased intracellular Ca2+ release and increased cyclic adenosine 3′,5′-monophosphate (cAMP) signaling, a consequence of ADPKD (25).

We hypothesized that a reduction of PC2 would diminish mitochondrial Ca2+ signals. However, in this study, we found that loss of PC2 increased mitochondrial Ca2+ transfer. We also observed decreased mitochondrial movement due to increased expression of the ER-mitochondrial tethering protein, mitofusin 2 (MFN2) (2628). Specific knockdown of MFN2 in the tubules of a PC2-deficient cystic mouse model restored mitochondrial Ca2+ signaling, morphological alterations, and reduced proliferation in cystic cells. Collectively, our results suggest that PC2 inhibits mitochondrial Ca2+ entry and that its absence alters the complement of Ca2+ transfer molecules at the mitochondrial-ER interface, ultimately affecting Ca2+-mediated mitochondrial bioenergetics that contribute to cell proliferation.

RESULTS

PC2 knockdown results in enhanced mitochondrial Ca2+ transients

As one of its main functions, PC2 modulates ER Ca2+ release. In light of work demonstrating metabolic alterations of ADPKD cystic cells, we examined mitochondrial Ca2+ signaling in an in vitro system comparing cells with PC2 knockdown (PC2 KD) to control (scrambled “SCR”) kidney cells (LLC-PK1) (fig. S1A). We used adenosine triphosphate (ATP) as an agonist of G protein (guanine nucleotide-binding)–coupled receptors that activate phospholipase C to produce inositol trisphosphate (InsP3) to activate the InsP3R. PC2 KD cells had a significant decrease in the cytoplasmic Ca2+ transient in response to ATP stimulation, as analyzed by the area under the curve (Fig. 1, A and B), consistent with previous reports (22, 29). We predicted that reduced cytoplasmic Ca2+ transients would be correlated with a reduction in mitochondrial Ca2+ uptake. To test this hypothesis, we used the mitochondrial-targeted ratiometric Ca2+ probe, pericam (30, 31), to interrogate mitochondrial Ca2+ signals (fig. S1, B and C). In contrast to our expectation, given the decreased cytoplasmic Ca2+ release after stimulation with ATP, the mitochondrial Ca2+ transient was significantly larger in PC2 KD cells compared to SCR cells (Fig. 1D). There was no alteration in baseline cytoplasmic Ca2+ or Ca2+ released from the ER stores after the addition of ionomycin, an ionophore causing release of all ER Ca2+ stores (fig. S1E).

Fig. 1 PC2 KD results in decreased cytoplasmic but increased mitochondrial Ca2+ signal.

(A) Measurements of the cytoplasmic Ca2+ signal in response to 5 μM ATP (measured using gCaMP6F) in PC2 KD cells compared to SCR cells. (B) Quantification of area under the curve of the gCaMP6F cytoplasmic Ca2+ amplitude. *P < 0.0001, determined by unpaired t test. n = 30 to 94 cells for each group from at least three independent experiments. (C) Example traces depicting mitochondrial Ca2+ measurements using the FRET sensor 4mitD3cpV in SCR cells (left) and PC2 KD cells (right) in response to 5 μM ATP. Em., emission. (D) Averaged data of the baseline mitochondrial Ca2+ levels measured with 4mitD3cpV. Differences were determined not significant by unpaired t test. (E) Averaged mitochondrial maximum Ca2+ response after 5 μM ATP stimulus measured with 4mitD3cpV. *P < 0.05, determined by unpaired t test. SCR: n = 3 individual experiments with 12 cells per experiment; PC2 KD: n = 6 individual experiments with 27 cells per experiment.

Because ratiometric pericam is sensitive to changes in pH (32, 33), these mitochondrial Ca2+ findings were corroborated using the pH-insensitive mitochondrial Ca2+ fluorescence resonance energy transfer (FRET) sensor, 4mitD3cpV (Fig. 1C) (34, 35). There was no significant difference in the baseline mitochondrial Ca2+ (Fig. 1D). However, consistent with the pericam measurements, mitochondrial Ca2+ influx in response to ATP was significantly greater in PC2 KD cells than in SCR cells (Fig. 1E).

To confirm that our findings remained consistent in other cell lines, we used stable mouse kidney epithelial (mIMCD3) cell lines stably expressing mutant Cas9 (IMCD3) or functional Cas9 and guide RNA (gRNA) targeting Pkd2 to genetically delete Pkd2 [PC2 knockout (KO); fig. S1F]. Cytoplasmic (gCaMP6F) and mitochondrial (mito-gCaMP6F) Ca2+ responses to ATP were measured in these cells, and similar to LLC-PK1 cells, PC2 KO cells had decreased cytoplasmic and increased mitochondrial Ca2+ responses compared to wild-type (WT) IMCD3 cells (fig. S1, G and H).

PC2 is located at the MAM and interacts with VDAC, an outer mitochondrial membrane channel

The change in mitochondrial Ca2+ could be explained if PC2 KD cells exhibited increased physical tethering between the ER and mitochondria. To determine whether tethering was influenced by the loss of PC2, we first sought to see whether PC2 was present in crude mitochondrial fractions, which contain resident MAM proteins. Co-immunoprecipitation experiments demonstrated that endogenous PC2 was enriched in the crude mitochondrial fraction and was associated with the voltage-dependent anion channel (VDAC; Fig. 2A), the outer mitochondrial membrane protein responsible for Ca2+ uptake into the intermembrane space. These data agree with a previous study that identified VDAC as an interacting protein of PC2 by immunoprecipitation mass spectrometry of PC2 in the vasculature (36).

Fig. 2 PC2 resides in ER membranes close to mitochondria.

(A) PC2 immunoprecipitates (IP) from cytoplasmic and MAM-containing crude mitochondrial fractions were immunoblotted for VDAC. n = 4 independent experiments. (B to D) Example images of live cells transfected with Mito-YFP and PC2-mCherry (B) or Mito-YFP and ER-mCherry in SCR (C) or PC2 KD cells (D) taken with STED super-resolution microscopy. Scale bars, 2.5 μm (left) and 1 μm (right). n = 12 to 21 cells for each group from at least three independent experiments. (E) Proposed schematic of Ca2+ handling and tethering proteins at the MAM. PC2 and InsP3R reside on the ER membrane in close apposition to mitochondria to regulate ER Ca2+ release. The cytoplasmic C-terminal tail of PC2 spans the distance between the ER and mitochondria to interact with VDAC on the outer mitochondrial membrane. The MCU complex resides on the inner mitochondrial membrane, and MFN2 is a tethering protein that physically links the ER and mitochondrial membranes at the MAM. TCA, tricarboxylic acid.

We then determined whether the localization of PC2 in the MAM was due to its presence within mitochondria. Conventional confocal microscopy revealed areas of apparent colocalization between the ER and mitochondria as detected by mitochondrial-targeted yellow fluorescence protein (Mito-YFP) and ER-targeted mCherry (ER-mCherry) and quantified using Mander’s overlap coefficient analyses (fig. S2, A and B). However, with super-resolution stimulated emission depletion (STED) microscopy, in which we obtained ~80-nm resolution with live-cell imaging, a distinct reticular expression pattern was observed with mCherry-tagged PC2 (PC2-mCherry), consistent with PC2 adopting ER tubule localization (Fig. 2B). There was no indication that PC2 colocalized with Mito-YFP, and there was no loss of ER patterning with PC2 KD (Fig. 2, B to D, and fig. S2C). These data suggest that intracellular PC2 is located on the ER membrane in close proximity to mitochondria but not within mitochondria themselves. This interpretation is consistent with the MitoCarta 2.0 database, which does not list PC2 as a resident mitochondrial protein (37). Together, these results present a picture of how these proteins of interest at the MAM coordinate to regulate Ca2+ transfer from the ER to mitochondria (Fig. 2E).

PC2 KD alters the distribution of ER Ca2+ release channelsin close proximity to mitochondria and increases mitochondrial Ca2+ uptake capacity

In addition to altering mitochondrial-ER tethering, redistribution of the Ca2+ channels at the MAM could also contribute to the differences between cytoplasmic and mitochondrial Ca2+. We therefore examined the MAM-containing mitochondria-enriched fraction for InsP3R isoform expression. The abundance of InsP3R3 in the mitochondrial fraction was significantly lower in PC2 KD cells than in SCR cells when normalized to VDAC, suggesting that PC2 modulates the composition of the Ca2+ release channels located at the MAM (Fig. 3, A and B). In contrast, there was no change in InsP3R1 amount (Fig. 3C). PC2 KD does not alter the overall amounts of InsP3R (38). The specific subtype of InsP3R is an important contributor to Ca2+ release dynamics because InsP3R1 is more sensitive than InsP3R3 to modulation by both InsP3 and ATP levels (39, 40). As a result, more Ca2+ will be released for a lower concentration of stimulus when InsP3R1 is present.

Fig. 3 PC2 KD modifies expression of ER Ca2+ release channels associated with the mitochondria.

(A) Crude MAM-containing mitochondrial fractions from PC2 KD and SCR cells were immunoblotted for InsP3R3, InsP3R1, and VDAC. Two different example preparations are shown. (B) Quantification of InsP3R3 levels in crude mitochondrial fractions from SCR and PC2 KD cells normalized to VDAC. **P < 0.01, determined by Mann-Whitney U test. n = 4 independent experiments. (C) Quantification of InsP3R1 levels in crude mitochondrial fractions from SCR and PC2 KD cells normalized to VDAC. Differences were determined not significant by Mann-Whitney U test. n = 4 independent experiments. (D) MCU protein abundance in SCR and PC2 KD cells. Three different example preparations are shown. (E) mRNA expression of MiCU2 and MCUb in SCR and PC2 KD cells. *P < 0.05 determined by unpaired t test. n = 4 independent experiments. (F) Representative trace of Ca2+ uptake into mitochondria measured with Calcium Green-5N. Digitonin (0.02%) was added ~1000 s after the start of the assay to permeabilize the cell membrane. Each spike is the addition of 200 nmol of calcium. (G) Quantification of mitochondrial Ca2+ uptake in permeabilized SCR and PC2 KD cells. *P < 0.0001 determined by unpaired t test. n = quadruplicate 96 wells for each group averaged from three separate experiments.

We then measured the expression levels of the MCU (9, 10) because MCU is the channel responsible for Ca2+ uptake into the mitochondrial matrix (41). Whereas MCU expression was unchanged in PC2 KD cells (Fig. 3D and fig. S3A), the expression of mRNAs encoding the MCU inhibitors MiCU2 and MCUb was significantly lower in PC2 KD cells than in SCR cells (Fig. 3E). The expression of the mRNAs encoding MiCU1, MiCU3, and EMRE were not affected by PC2 KD (fig. S3B). This expression pattern is consistent with an increase in mitochondrial Ca2+ uptake capacity.

Mitochondrial Ca2+ uptake experiments in permeabilized cells (Fig. 3F) confirmed that mitochondria from PC2 KD cells took up more Ca2+ (Fig. 3G), consistent with loss of inhibition due to decreased MiCU2 and MCUb. Together, the combined effects of the changes in the MCU complex coupled with the alterations in InsP3R distribution at the MAM help to explain the increased mitochondrial Ca2+ uptake observed in PC2 KD cells.

ER-mitochondrial tethering is increased with PC2 KD

Because our experiments indicated changes in the MCU complex and MAM-associated Ca2+ channels, we examined whether these changes were accompanied by structural changes in mitochondria. Transmission electron microscopy (TEM) analysis revealed a substantial increase in mitochondrial-ER tethering and mitochondrial morphological rearrangements in PC2 KD cells. Specifically, we found that the distance between the ER and mitochondria was decreased compared to SCR cells (Fig. 4, A and B). In addition, mitochondria in PC2 KD cells had a larger area and a more circular morphology (Fig. 4, C and D).

Fig. 4 Ultrastructural analysis of PC2 KD cells reveals closer ER-mitochondrial distances and increased mitochondrial fragmentation.

(A) Example TEM images from SCR and PC2 KD cells. Arrows denote example mitochondrial (m) distance to the ER. “n” denotes cell nucleus. Area outlined in black is shown at higher magnification (bottom). Scale bars, 500 nm. (B) Analysis of distance between mitochondria and their nearest ER, with each symbol representing a single calculated ER-mitochondrial distance. *P < 0.0001, determined by unpaired t test. (C) Histogram of individual mitochondrial area as assessed by TEM images. (D) Histogram of mitochondrial circularity as assessed by TEM images, where 1.0 represents a circle. n = 20 cells for each condition with four different regions imaged per cell.

PC2 KD cells have increased MFN2 expression and altered mitochondrial metabolism

Consistent with increased mitochondrial-ER tethering, we found increased abundance of MFN2, a mitochondrial-ER tethering protein, and increased overall mitochondrial content in PC2 KD cells (Fig. 5, A and B, and fig. S4A). This increase in mitochondrial density was also observed in three-dimensional (3D) cultured cysts (Fig. 5, C and D). We hypothesized that this increase in MFN2 contributed to the altered mitochondrial Ca2+ dynamics seen in PC2 KD cells because modulation of MFN2 levels directly influences mitochondrial Ca2+ uptake (42). To test this notion, we knocked down MFN2 in PC2 KD cells (fig. S4B). Mitochondrial Ca2+ was measured in response to ATP using mitochondrial-targeted gCaMP6F, and, consistent with our hypothesis, knockdown of MFN2 in PC2 KD cells rescued the mitochondrial Ca2+ response to levels similar to those in SCR cells (Fig. 5, E and F). In addition, knockdown of MFN2 in PC2 KD cells restored the cytoplasmic Ca2+ response to ATP (fig. S4C).

Fig. 5 PC2 KD increases MFN2 expression and mitochondrial metabolism through altered Ca2+ dynamics.

(A) Western blot of MFN2 protein abundance in SCR and PC2 KD cells. Four different preparations are shown. (B) Quantification of MFN2 protein abundance in SCR and PC2 KD cells, normalized to actin. *P < 0.05, determined by Mann-Whitney U test. n = 4 biological replicates. (C) 3D cultured SCR and PC2 KD cysts were stained for phalloidin (red), 4′,6-diamidino-2-phenylindole (DAPI; blue), and TOM20 (green), an outer mitochondrial membrane protein. (D) TOM20 intensity normalized to cell area was quantified. Scale bar, 5 μm. *P < 0.0001, determined by unpaired t test. (E) Example average trace of mitochondrial Ca2+ responses to 5 μM ATP using mito-gCaMP6F in SCR, PC2 KD, and PC2 KD MFN2 KD cells. (F) Quantification of maximum mitochondrial Ca2+ responses in SCR, PC2 KD, and PC2 KD MFN2 KD cells. *P < 0.0001, determined by unpaired t test. n = 30 or more cells for each group from three independent experiments. (G) OCRs and maximal respiration capacity as revealed by oxygen flux mitochondrial stress test analyses in SCR and PC2 KD cells. *P < 0.05 determined by unpaired t test. n = 3 independent experiments with 16 replicates for each group in each experiment. (H and I). Simultaneous measurements of mitochondrial Ca2+ using mito-GCaMP6F (green lines) and mitochondrial membrane potential (red line) in SCR (H) or PC2 KD cells (I) in response to 5 μM ATP, 2 μM oligomycin, and 2 μM FCCP. Data are averaged from three to six independent experiments representing 12 to 27 individual cells for each group.

In 2D culture, PC2 KD cells turned the growth medium acidic more quickly than SCR cells (fig. S4D), suggesting an effect of PC2 KD on the cells’ metabolic status. To compare the metabolic capacity of PC2 KD and SCR cells, we measured the oxygen consumption rate (OCR) of cells subjected to a mitochondrial stress test and found a substantially higher OCR in PC2 KD cells under basal conditions, but lower OCR after mitochondrial uncoupling with the ionophore carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (FCCP) (Fig. 5G). Similar results for basal conditions were obtained using an assay to measure ATP/ADP (adenosine diphosphate; fig. S4E) and a Clark electrode to measure OCR (fig. S4F). Collectively, these data suggested that, after PC2 KD, cells have an increased basal metabolic profile but an overall loss of maximal respiratory capacity. These results are consistent with our data demonstrating increased mitochondrial Ca2+ transients with PC2 KD cells, because increased mitochondrial Ca2+ uptake is associated with enhanced metabolism through Ca2+-dependent regulation of matrix dehydrogenases (43).

To confirm that altered mitochondrial Ca2+ and not mitochondrial membrane potential was responsible for these metabolic changes, we measured mitochondrial Ca2+ dynamics [using mitochondrial-targeted gCaMP6F (mito-gCaMP6F)] and inner mitochondrial membrane potential [using tetramethylrhodamine methyl ester (TMRE)] in PC2 KD and SCR cells under the same treatment conditions as the mitochondrial stress test. In agreement with the hypothesis that mitochondrial Ca2+ is responsible for the observed changes in OCR, we determined that PC2 KD cells had enhanced mitochondrial Ca2+ uptake after the addition of ATP and oligomycin, but the change in mitochondrial Ca2+ after FCCP addition was less pronounced in PC2 KD cells than in SCR cells (Fig. 5, H and I). Conversely, TMRE measurements showed no difference between SCR and PC2 KD cells.

Super-resolution microscopy analysis of mitochondrial dynamics reveals decreased mitochondrial movement with PC2 KD

Changes in mitochondrial-ER tethering contacts result in altered mitochondrial movement dynamics, which is closely correlated with changes in metabolism (44). Although TEM can obtain a snapshot of the ultrastructure, it does not provide information as to how stable these interorganellar contacts are. Using time-lapse super-resolution microscopy under conditions that did not induce differential mitochondrial reactive oxygen species (ROS) production (fig. S5), we tracked mitochondrial dynamics for 5 min (Fig. 6A and movie S1) and found that, although mitochondria in PC2 KD and SCR cells traveled similar distances at equivalent velocities (Fig. 6, B and C), PC2 KD mitochondria showed a significant decrease in vectorial displacement (Fig. 6D), suggesting reduced linearity in their movement. Examination of the PC2 KD cells revealed that many mitochondria adopted dynamic ring-like structures, which transitioned from a linear to circular conformation without directed movement along the cellular cytoskeleton (Fig. 6E, arrows). These data suggest that PC2 KD mitochondria, which spend more time in contact with particular MAM regions of the ER, enable privileged and prolonged Ca2+ communication between the ER and mitochondria.

Fig. 6 PC2 expression alters mitochondrial dynamics.

(A) Single–time point images of SCR (top left) or PC2 KD (bottom left) cells transfected with ER-mCherry and Mito-YFP. Scale bar, 2.5 μm. Tracks of individual mitochondrial movement from frame to frame are depicted, with each color representing a different individual mitochondrion over time. (B) Averaged total distance traveled by mitochondria over a 5-min period. (C) Averaged velocity of mitochondria over a 5-min period. Data in (B) and (C) are representative of at least 12 mitochondria per cell. Each symbol is representative of an individual cell from at least three different preparations. (D) Absolute direction in x and y axes traveled by each mitochondrion over a 5-min period. Each symbol is representative of 5 to 10 cells, with 87 to 267 mitochondria per cell. AU, arbitrary units. (E) Time-lapse example of static circular mitochondria found exclusively in PC2 KD cells (white arrows).

The PGC1α-CREB axis is activated in PC2 KD cells

We wanted to further explore the mechanism by which MFN2 expression is increased to affect these observed mitochondrial changes. We tested the levels of peroxisome proliferator-activated receptor gamma coactivator 1 alpha (PGC1α), an established upstream regulator of MFN2 expression, and found a >40-fold increase in the mRNA expression for this transcription factor in PC2 KD cells (Fig. 7A). Because PGC1α activates mitochondrial biogenesis, its up-regulation is also consistent with the increased mitochondrial density observed in both PC2 KD 3D and 2D cultures (Fig. 5, C and D, fig. S4A) and with the increased expression of MFN2 (Fig. 5, A and B). Increased expression of PGC1α is stimulated through cAMP response element–binding protein (CREB) activation (45), and in ADPKD kidneys, cAMP levels are elevated because of increased vasopressin receptor signaling and cAMP production (46, 47, 48). Consistent with this literature, phosphorylated CREB (pCREB), which is activated through cAMP signaling, was significantly increased in the PC2 KD samples (Fig. 7B).

Fig. 7 PC2 KD increases mitochondrial biogenesis through the PGC1α-CREB axis.

(A) mRNA expression of PGC1α in SCR and PC2 KD cells. ***P < 0.001, determined by unpaired t test. n = 4 independent experiments. DCT, distal convoluted tubule. (B) Immunoblot for phosphorylation of the transcription factor CREB in SCR and PC2 KD cells. Four different example preparations are shown. Right: Quantification of pCREB normalized to CREB in SCR and PC2 KD cells. *P < 0.05 determined by Mann-Whitney U test. n = 4 independent experiments.

Although PGC1α was up-regulated, there was no change in the expression of mRNAs encoding alternate mitochondrial fusion or other tethering proteins, OPA1 or Mitofusin 1 (fig. S6A). Similarly, sigma1 receptor, whose abundance is increased under conditions of ER stress and can act as a tethering protein between the ER and mitochondria (49), showed no significant change in protein expression (fig. S6B). These results suggest that the increased abundance of MFN2 facilitates increased tethering between the ER and mitochondria in PC2 KD cells. In addition, dysregulated levels of PGC1α can render cells susceptible to hyperfragmentation of mitochondria and altered mitochondrial dynamics (50), as seen in PC2 KD cells (Figs. 4D and 6, A, D, and E).

Tissue-specific targeting of MFN2 in cystic mice restores Ca2+ signaling and decreases proliferation

Because we observed that diminishing the tethering between the ER and mitochondria reversed the increased mitochondrial Ca2+ signaling in PC2 KD cells (Fig. 5, E and F, and fig. S4, B and C), we asked whether this manipulation would also be effective in cystogenic signaling pathways present in animal models. We therefore created MFN2-specific small interfering RNA (siRNA) expressed under a Cre promoter [pSico lentiviral vector (51)]. The addition of two siRNA sequences against MFN2 was effective in knocking down MFN2 in collecting duct cells from Pkd2F/F mice with adenoviral Cre recombinase added ex vivo (Fig. 8A and fig. S7A). Pkd2F/FPkhd1Cre mice, which develop cysts from birth due to PC2 KO in the collecting ducts (fig. S7B), were transduced with the lentivirus (LV) through retro-orbital injection at 3 weeks of age. MFN2 abundance was relatively uniform across tubules in Pkd2F/F control mice (Fig. 8B, left). MFN2 abundance was increased in the cystic kidneys of Pkd2F/FPkhd1Cre mice (Fig. 8B, middle) and was reduced by LV-siMFN2 transduction (Fig. 8B, right).

Fig. 8 Reduction of MFN2 expression in murine cystic cells restores mitochondrial Ca2+ signaling and hyperproliferation.

(A) PC2 and MFN2 protein abundance in Pkd2F/F collecting duct cells after the addition of adenoviral Cre ex vivo and lentiviral transduction of two different siRNAs against MFN2. Actin was used as a loading control. Representative of two independent experiments (see also fig. S7A). (B) Expression of MFN2 (green) in Pkd2F/F tubules (left), Pkd2F/FPkhd1Cre cystic tubules (middle), and Pkd2F/FPkhd1Cre cystic tubules 4 weeks after injection of siMFN2 (right). Representative of four different mice per group. Scale bar, 10 μm. (C) Mitochondrial Ca2+ changes upon addition of 5 μM ATP. Traces from a single coverslip as a representation of the data are shown. Data were collected from four separate coverslips with at least 20 cells per coverslip, representing two to three different mice per genotype. (D) Ki67 staining (green) for proliferating cells in kidney sections from Pkd2F/FPkhd1Cre mice (middle) and siMFN2-transfected Pkd2F/FPkhd1Cre mice (right) (see also fig. S8, C and D) compared to Pkd2F/F mice (left). Sections from additional mice are presented in fig. S8D. Representative of two to three different mice per group. Scale bar, 10 μm.

The cystic collecting duct cells (fig. S8A) from Pkd2F/FPkhd1Cre recapitulated several of the features observed in the LLC-PK1 PC2 KD cells. In cells transfected with Mito-YFP and ER-mCherry, the mitochondria were more circular in cystic cells, a phenotype that was corrected by LV-siMFN2 treatment (fig. S8B). Compared to Pkd2F/F control cells (Fig. 8C, left), mitochondrial Ca2+ signals in response to ATP were enhanced in cystic cells (Fig. 8C, middle) and restored to control values by LV-siMFN2 transduction (Fig. 8C, right).

We also examined cystic parameters of cell proliferation and apoptosis after siMFN2 KD. Three weeks of siMFN2 KD resulted in a marked reduction in cell proliferation in the LV-siMFN2–treated mice that was comparable to Pkd2F/F mice, as measured by Ki67 staining (Fig. 8D, and fig. S8, C and D). However, apoptosis measured through caspase-3 activation did not differ between cystic and noncystic cells (fig. S8E).

Cystic kidneys from patients with ADPKD have increased mitochondrial density and MFN2

We examined kidney sections from noncystic (normal) and ADPKD-diagnosed patients. In kidney sections from normal individuals, MFN2 and the mitochondrial protein translocase of the outer mitochondrial membrane (TOM20) were increased in the distal convoluted tubule and collecting ducts [stained by the lectin Dolichos biflorus agglutinin (DBA)] compared to other more proximal nephron regions [Fig. 9, A and B (gray bars compared to red bars), and fig. S9A]. Moreover, MFN2 and TOM20 staining were localized to the basolateral membrane in DBA-positive tubules [Fig. 9, A (bottom) and B (Ap. compared to BL gray bars), consistent with previous findings (52)]. In ADPKD patient samples, extensive fibrosis was found, and cysts were generally DBA positive (Fig. 9C). MFN2 and TOM20 staining were increased in cyst-lining cells (Fig. 9D, blue bars, and fig. S9B). Note that noncystic tubules in patients with ADPKD had MFN2 and TOM20 levels comparable to those in WT control samples [Fig. 9A (bottom) compared to Fig. 9C (bottom left); ratio of MFN2 to TOM20 is presented in fig. S9 (C and D)]. These data indicate that the mitochondrial density and the MAM-tethering protein MFN2 are increased in cysts from human subjects with ADPKD.

Fig. 9 Renal cyst-lining cells from human patients with ADPKD have increased MFN2 and mitochondria.

(A) Human kidney samples from normal patients were stained with antibodies against TOM20 (green) and MFN2 (red) and counterstained with the lectin DBA (blue) as a marker for collecting duct tubules. Scale bar, 50 μm (top) and 5 μm (bottom). (B) Quantification of MFN2 staining normalized to area in normal human kidney, separated into DBA lectin–positive tubules (gray bars) and DBA lectin–negative tubules (red bars). Within each group, MFN2 staining was further separated by its intensity in the whole tubule (W Tb.), apical localization (Ap.), and basolateral localization (BL). *P < 0.05; **P < 0.01; ****P < 0.0001, determined by Mann-Whitney U test. (C) Human kidney samples from patients with ADPKD were stained with antibodies against TOM20 (green) and MFN2 (red) and counterstained with lectin DBA (blue). Scale bar, 50 μm (top) and 5 μm (bottom left and bottom right). (D) Quantification of MFN2 staining in DBA-positive tubules (gray bars), DBA-negative tubules (red bars), and cyst lining epithelium (blue bars) from patients with ADPKD. *P < 0.05; **P < 0.01 determined by Mann-Whitney U test. Quantification of five images per group (five lectin-positive, five lectin-negative, and five cystic samples) and two tissues of each (two normal human kidney and two ADPKD). W Cy., whole cyst.

DISCUSSION

Here, we established that PC2 served to limit Ca2+ signaling to mitochondria by modulating proteins located at the MAM and by increasing mitochondrial biogenesis and mitochondrial-ER tethering through PGC1α. Loss of PC2 resulted in enhanced mitochondrial Ca2+ uptake, mitochondrial bioenergetics, and increased mitochondrial-ER tethering through increased MFN2 abundance. These findings were also present in cysts from human patients with ADPKD and kidney tubule cells from cystic mice. Knockdown of MFN2 restored mitochondrial Ca2+ signaling, morphological changes, and hyperproliferation in cystic cells. Collectively, these findings demonstrate that PC2 is an important regulator of mitochondrial function, which is relevant in pathological settings.

We propose that PC2 modulates the ER protein composition associated with mitochondria and can conduct this function by directly interacting with VDAC. The InsP3R is the main intracellular release channel that provides Ca2+ to the mitochondria, and our data do not point to PC2 being another intracellular release channel for the mitochondria, but rather, a regulator of Ca2+ entering the mitochondria. This function can be achieved in two ways: (i) by directly linking mitochondria to the ER and (ii) by modulating InsP3R subtype expression in the MAM. The extensive C-terminal tail of PC2 is long enough to span the 10- to 20-nm distance between the ER and mitochondria (30, 53) to interact with VDAC, and, in addition, increased MFN2 through CREB-mediated up-regulation helps to facilitate more sustained ER-mitochondrial contacts. Our finding that the expression of InsP3R3 was decreased in the PC2 KD MAM was notable because InsP3R3 is considered the dominant InsP3R at the mitochondrial-ER junction (11). Because PC2 is not only found in the kidney but also in other organs including the vasculature (36) and cardiomyocytes (54), our data suggest that PC2 expression in other cell types may affect cytoplasmic and mitochondrial Ca2+ signals independently of unrelated to cyst formation.

A second way by which PC2 loss modulates mitochondrial signaling is by increasing PGC1α expression through CREB phosphorylation and activation, which leads to enhanced mitochondrial biogenesis and increased abundance of the ER-mitochondrial tethering protein MFN2. The functional role of MFN2 is complex because loss of MFN2 has been associated with loss of ER-mitochondrial contacts (2628, 5560), whereas other reports suggest that loss of MFN2 results in increased ER-mitochondrial contacts (41, 61). Although static EM studies have been previously used to assess ER-mitochondrial contacts, here, we combined EM data with super-resolution live-cell imaging of organelles to demonstrate a reduction of vectorial movement of mitochondria upon PC2 KD, indicating enhanced long-term contacts between mitochondria and ER. Our data suggest that changes in MAM proteins, along with altered mitochondrial morphology and movement, enable permissive uptake of Ca2+ into mitochondria, which leads to a heightened metabolic state (Fig. 5G and fig. S4, D to F), as would be expected in cystic tissue. Consistent with this notion, human cysts also had increased MFN2 expression, and in our cell culture and murine cystic models, decreasing MFN2 expression restored both mitochondrial and cytoplasmic Ca2+ signaling. These data suggest that mitochondrial-ER tethering through MFN2 is critical to balance the amount of ER Ca2+ released into the cytoplasm versus into mitochondria. Because mitochondrial signaling was restored through MFN2 KD alone, the expression of MiCU2 and MCUb is likely reciprocally regulated by the altered intracellular Ca2+ dynamics that can then influence transcriptional regulation. Hence, cAMP-mediated pCREB/PGC1α activation in PC2 KD cells may ultimately affect mitochondrial function by increasing the abundance of MFN2, altering the path of ER Ca2+ release to preferentially be taken up into mitochondria, which is then exacerbated by transcriptional increases of MiCU2 and MCUb. Nevertheless, despite these corrections and the reduction in proliferation with MFN2 KD, no gross reduction in cysts was observed. However, because the nascent renal cysts had already developed by the time the LV-siMFN2 was introduced to the mice at 3 weeks of age, pretreatment may have a larger impact on cyst formation.

Although we have focused on PC2, further studies are needed to examine whether mitochondrial Ca2+ and dynamics are likewise disrupted in cysts caused by PKD1 mutations. Unlike PC2, PC1 does not appear to be present on the ER membrane, and thus, a direct perturbation of mitochondrial Ca2+ uptake by ER Ca2+ release would be likely to occur only if the ER-mitochondrial membrane trafficking of PC2 and/or InsP3R distribution is altered by PC1. Because PC2 regulates the abundance and trafficking of PC1 (6265), mutations in PC1 are likely to also reciprocally alter expression and function of PC2 at mitochondrial-ER contacts, as previously reported in the cilia (66). A cleavage product of the PC1 C-terminal tail can enhance mitochondrial respiration and alter morphology when overexpressed in heterologous systems (67). However, how this cleavage product is involved in cyst formation or mitochondrial dysfunction in ADPKD, when the polycystins are absent or mutated, remains to be discovered. Morphological differences seen in mitochondria from kidneys of patients with ADPKD and renal epithelial cells isolated from Pkd1ko/ko mice fit well with our finding that (i) PC2 KD cyst-modeling cells exhibit altered metabolic capacity and (ii) that mitochondria from these cells and from ADPKD kidneys show a high degree of fragmentation. This fragmentation and decreased mitochondrial network formation have been noted in cystic tissue derived from PC1-deficient mice and a cystic rat model as well (68). Our finding that mitochondria from PC2 KD cells have a decreased overall metabolic capacity accords well with this phenotype of hyperfragmentation and circularity, because mitochondria that are unable to form elongated intracellular networks become dysfunctional and consequent targets for mitophagy, thus necessitating increased mitochondrial biogenesis (69). However, although morphological differences appear to be a common phenotype between both PKD1- and PKD2-caused ADPKD development, PGC1α-mediated mitochondrial density differs depending on whether cystic cells derive from PKD1 or PKD2 loss of function. Whereas kidney tissues from PC1-deficient mice show decreased PGC1α levels (68), our data demonstrate a substantial increase in PGC1α in PC2 KD cells. These different phenotypes may be a result of the metabolic differences between PC1- and PC2-mediated cyst development because, in contrast to our results in PC2 KD cells, Pkd1−/− cells have decreased basal metabolic rates (18), which may also be influenced by the lack of the PC1 cleavage product to enhance mitochondrial function. In addition, our human data demonstrate altered ER-mitochondrial contacts and mitochondrial content in cystic tissue. Because PKD2 mutations contribute to a lower frequency of ADPKD (~20%) than PKD1 mutations, most of the human tissue samples are presumed to have PKD1 mutations. Therefore, our findings accord well with the idea that mitochondria-ER communication is a common mechanism contributing to altered metabolism observed in ADPKD (15).

Our finding that PC2 levels alter mitochondrial function has implications for cyst-promoting cell proliferation pathways associated with loss of polycystins, as evidenced by the decreased Ki67 staining in the MFN2 KD cyst samples. Cystic cell proliferation is due to a combination of elevated cAMP and decreased Ca2+ signaling, which collectively initiate a growth-promoting B-Raf signaling pathway (70, 71). Enhanced mitochondrial function could also aid in cellular proliferation when combined with B-Raf pathway activation, because cell proliferation requires high ATP production, which, under the hypoxic conditions present in cysts, would require a shift to anaerobic glycolytic pathways (15). Increased ATP production can inhibit adenosine 5′ monophosphate–activated protein kinase (AMPK), which may activate the cell survival–associated mammalian target of rapamycin (mTOR) pathway. A component of the mTOR signaling pathway, mammalian TOR complex 2, is physically present within the MAM (72, 73). These pathways, along with a cAMP signaling pathway, are enhanced in ADPKD, making them possible targetable therapeutic pathways (7376). Here, we found that, in a cystic mouse model, knockdown of MFN2 markedly reduces cell proliferation, uncovering this linker protein as a potential therapeutic target for patients with PKD2 mutations to restore homeostatic calcium and the resulting altered signaling pathways in these cells.

In conclusion, we describe a new functional role for PC2 in the regulation of physical contacts between the mitochondria and the ER. Moreover, these contacts are perturbed in ADPKD, similar to several other disease states (77, 78). The contribution of PC2 to mitochondrial function implicates a pathway involved in cystogenesis in ADPKD. Last, the wide tissue distribution of PC2 raises the possibility that PC2 may contribute to disease states in other tissues where mitochondria and cellular metabolism are disrupted.

MATERIALS AND METHODS

Cell culture and creation of stable knockdown cells

Short hairpin RNA (shRNA) against PC2 and MFN2 was designed to be specific to the pig sequence. The PC2 shRNA constructs were previously described and characterized (38, 79) and were put into a psiRNA vector (InvivoGen, San Diego, CA), which was stably transfected into LLC-PK1 cells. Cells were exposed to the selection agent G418, which was maintained in the medium at 2.5 mg/ml and removed 2 days before experimentation, as previously described (38, 79). LLC-PK1 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and kept at 37°C in 5% CO2.

Generation of Pkd2 KO mIMCD3 cell line

To maximize the effect of gene disruption, gRNA sequences near exon 2 and exon 3 of the mouse Pkd2 genomic DNA were used. Pkd2 single guide RNA (sgRNAs) were designed with the CRISPR Design Tool (http://crispr.mit.edu/). sgRNAs CACAGATGCACAAGACTACG and TACACGCCCTGCCCCTCTCG were used for Pkd2 exon 2 editing, and sgRNAs TACCTTCCAGAAGTCCTCCA and GTCTCTGTGAATTACTGACT were used for Pkd2 exon 3 editing. Two pairs of 20-nt (nucleotide) sgRNAs in a tail-to-tail orientation for Pkd2 knockout were cloned into the plasmid pGL3-U6-sgRNA-PGK-Hygromycin, which was obtained from the modified pGL3-U6-sgRNA-PGK-puromycin plasmid (Addgene plasmid no. 51133). sgRNA sequences were inserted between the promoters (U6 and/or H1) and the sgRNA scaffold.

The Cas9 D10A plasmid [CMV-hspCas9(D10A)-T2A-Puro; catalog no. CASLV100PA-1, System Biosciences, Inc.], pMDLg/pRRE, pRSV-Rev, and pMD2.G were transfected into the human embryonic kidney (HEK) 293 T cell line with Lipofectamine 2000 (Invitrogen) to generate Cas9 D10A LV. Cas9 D10A LV was infected into mIMCD3 cell lines, and the infected cells were selected with puromycin to obtain the stable Cas9 D10A mIMCD3 cell lines. pGL3-U6-sgRNA-PGK-Hygromycin with Pkd2-specific sgRNAs was transfected, and sgRNAs were introduced into Cas9 D10A mIMCD3 stable cells. Individual cells were transferred into 96-well plates after selection with hygromycin and puromycin. The cells were reseeded and cultured in the plates in duplicate after expansion for 2 to 3 weeks with antibiotics.

These cells were collected and digested overnight with lysis buffer. Polymerase chain reaction (PCR) products were amplified from the extracted and purified genomic DNA of the different clones. Final Pkd2 knockout mIMCD3 cell lines carrying frameshift insertions-deletions (indels) in exon 2 and exon 3 sequences of Pkd2 genomic DNA were determined by sequencing of the PCR products. Pkd2 knockout mIMCD3 cell lines were further confirmed with Western blot using an antibody toward the PC2 antibody [YCC2, as previously described (21)]. Confirmed Pkd2 knockout and Cas9 mutant control cells were maintained in DMEM/F-12 50:50 supplemented with 10% FBS and kept at 37°C in 5% CO2.

Extracellular flux analysis

The mitochondrial respiration capabilities of LLC-PK1 SCR and PC2 KD cells were tested using the Seahorse XF96 Extracellular Flux Analyzer (Agilent Technologies). Cells were plated in 96-well plates at a density of 8 × 104 cells per well 24 hours before the assay. Two hours before analysis, cells were washed twice with and then changed to Seahorse Assay Medium (Agilent Technologies) containing d-glucose (4.5 g/liter), 2 mM l-glutamine, and 1 mM sodium pyruvate before being placed in a CO2-free incubator at 37°C. Basal OCR measurements were taken before the sequential addition of 2 μM Oligomycin A, 2 μM FCCP, and 1 μM rotenone to measure ATP production and proton leak, maximal mitochondrial respiration, and nonmitochondrial respiration, respectively. After completion of Seahorse analysis, OCR results were normalized to cell number by lysing cells and quantifying the double-stranded DNA (dsDNA) in each well using the Quant-iT PicoGreen dsDNA Assay Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. Experiments were performed in quadruplicates or greater.

Clark electrode

Cells (1 × 106) were trypsinized and resuspended in respiration buffer [120 mM KCl, 5 mM KH2PO4, 10 mM tris-HCl, 3 mM MgCl2, and 1 mM EDTA (pH 7), modified from (80)) and permeabilized with 0.03% (w/v) digitonin]. Cells were placed in a Clark electrode oxygraph and allowed to equilibrate at 37°C. The following reagents were added sequentially with Hamilton syringes: ADP (2 to 8 mM), pyruvate (5 mM) and malate (10 mM), and oligomycin (6 μM, used to block ATP synthase).

ATP/ADP assay

Cells were plated in 96-well plates, and ATP/ADP ratios were measured according to the manufacturer’s instructions (ADP/ATP Ratio Assay Kit, Sigma-Aldrich). Experiments were performed in quadruplicates.

Mitochondrial enrichment, Western blot, and immunoprecipitation

Cells from two T75-cm flasks were combined and centrifuged. Crude mitochondrial extracts were obtained by using a mitochondrial extraction kit (Thermo Fisher Scientific, Waltham, MA). The final spin was conducted at 1000 rpm for 15 min at 4°C to obtain a cleaner mitochondrial fraction with reduced lysosome contamination. Mitochondria were lysed in 2% CHAPS with tris buffer, and the protein concentration was measured with the bicinchoninic acid assay (Thermo Fisher Scientific).

For immunoprecipitation experiments, 100 μg of protein was added to precleared A/G agarose beads (Santa Cruz Biotechnology, Santa Cruz, CA) and allowed to bind with 1 to 2 μg of antibody overnight at 4°C. Antibodies used were VDAC (Abcam, Cambridge, UK) and PC2 (a gift from Y. Cai, Yale University). The adhered protein was removed from the beads by boiling in loading buffer and reducing agent for 10 min after extensive washing of the beads (four times) with lysis buffer.

Samples for all Western blot analyses were loaded on 4 to 12% gradient gels (NuPage gels, Life Technologies) and run with either MES or MOPS buffer. Protein was transferred to polyvinylidene difluoride membranes by wet transfer. After blocking in 5% milk, primary antibodies were applied overnight.

Antibodies

Antibodies for Western blot and immunohistochemistry were as follows: InsP3R1 and InsP3R3 (BD Biosciences), PC2 (a gift from Y. Cai, Yale University), VDAC (Abcam), MCU (Cell Signaling Technology, Danvers, MA), MFN2 (Cell Signaling Technology), TOM20 (Santa Cruz Biotechnology), pCREB and CREB (Cell Signaling Technology), and tubulin (Abcam). VDAC and TOM20 were used to assess mitochondrial quality and purity. β-Actin (Abcam) was used as a loading control and to assess cytoplasmic contamination of mitochondrial samples. For Western blot analyses, horseradish peroxidase–conjugated biotin-labeled secondary antibodies were used and detected with West Dura Extended Duration Substrate (Thermo Fisher Scientific).

Cytosolic Ca2+, mitochondrial membrane potential, and mitochondrial ROS measurements

Cells were transfected with gCaMP6F [Addgene (81)] and imaged the next day. The mitochondrial membrane potential (Δψm) was assessed with 10 nM mitochondrial voltage-sensitive dye TMRE (Life Technologies) (82). After loading for 15 min at 37°C, the medium was exchanged to imaging buffer without TMRE, and cells were left to equilibrate at room temperature for 15 min. The imaging buffer consisted of 1.25 mM CaCl2, 19.7 mM Hepes, 4.7 mM KCl, 1.2 mM KH2PO4, 1 mM MgSO4, 130 mM NaCl, and 5 mM dextrose (pH 7.2) at room temperature. Cells were imaged using the excitation wavelengths of 488 nm (gCaMP6F) and 561 nm (TMRE), and average fluorescence intensities over 30 min were acquired. As a positive control for loss of Δψm, cells were treated with 50 μM FCCP (Life Technologies) to depolarize mitochondria at the end of the experiment. Cells were imaged at room temperature on a Leica SP5 microscope with a 63× 1.4 numerical aperture oil objective and 2.5× zoom. Fluorescence intensities were quantified with a combination of Leica TCS5 software and ImageJ [National Institutes of Health (NIH)].

For measurement of superoxides generated by mitochondria, cells were incubated with MitoSOX (Life Technologies) according to the manufacturer’s recommendations. Cells were then washed and imaged.

Mitochondrial Ca2+ imaging experiments

Cells were transiently transfected with the ratiometric indicator, pericam, which is directed to the mitochondria. Cells were imaged using dual excitation at both 380 and 480 nm (Lambda DG-4; Sutter Instruments, Novato, CA), and images were acquired with an Orca camera (Hamamatsu Photonics, Middlesex, NJ) with MetaMorph software (Axon Instruments, Sunnyvale CA). Mitochondrial Ca2+ transients were elicited by the addition of 1 μM ATP. Data were analyzed using PRISM software (GraphPad Software, La Jolla, CA).

For FRET imaging experiments of matrix Ca2+, cells were transiently transfected with pcDNA-4mitD3cpV [a gift from A. Palmer and R. Tsien; Addgene plasmid no. 36324 (35)]. FRET emission ratios (535 nm/485 nm; excitation, 440 nm) were acquired every 5 s using fluorescence ratio imaging systems running MetaFluor software (Molecular Devices). Cells were stimulated with 5 μM ATP followed by 5 μM ionomycin and imaged using Nikon TE200 or TE2000-U inverted fluorescence microscopes equipped with a QuantEM 512 camera (Photometrics, Tuscon, AZ) or an ORCA ER camera (Hamamatsu Photonics), respectively. Data were analyzed using MetaFluor software.

For mito-gCaMP6F experiments (a gift from D. Stefani), cells were transfected with mito-gCaMP6F. Cells were imaged using sequential excitation at 488 nm, and images were acquired with emission bandwidth of 501 to 555 nm. Images were acquired every second on a Leica SP5 confocal microscope with a 40× objective lens. ATP (5 μM) was added to stimulate mitochondrial Ca2+ signals and was followed by application of oligomycin and FCCP.

Super-resolution microscopy

Cells were plated in eight-well, 1.5-mm-thick cover glass chambers (LabTek, Rochester, NY) and transfected the day before the experiment. Cells were imaged on a Leica SP8 equipped with a STED depletion laser (Heidelberg, Germany). Cells were imaged with a white-light laser at 488 and 568 nm in sequential line mode, and a 740-nm STED depletion laser with a 100× oil objective and two to four times zoom was used. For time-lapse imaging, cells were imaged using bidirectional scanning, and each frame was acquired about 45 s apart to minimize photobleaching. Images were then subjected to deconvolution using Huygens software (Scientific Volume Imaging, Hilversum, The Netherlands).

mRNA analysis

mRNA was extracted from cells grown to confluency using an RNeasy kit and QiaShredder (Qiagen, Hilden, Germany) and reverse-transcribed to complementary DNA (cDNA) using Multiscribe (Life Technologies) and random oligomers as primers (Applied Biosystems, Foster City, CA). For real-time reverse transcription PCR, 20 ng of cDNA was used as transcript in a reaction with POWER SYBR Green MasterMix (Life Technologies) on a 7500 Fast machine (Applied Biosystems). Fold change in mRNA transcript levels was determined by using the 2-ΔΔCt method. Actin was used as a control.

Measurement of mitochondrial DNA

DNA was isolated from cultured cells using a DNeasy blood and tissue kit according to the manufacturer’s instructions (Qiagen). Twenty nanograms of DNA was used as transcript with POWER SYBR Green MasterMix (Life Technologies) on a 7500 Fast machine (Applied Biosystems). Primers against the porcine sequences of mitochondrial genes mtATP6 [5′:CTACCTATTGTCACCTTA (forward); 5′:GAGATTGTGCGGTTATTAATG (reverse)] and mtATP8 (5′:ATCTACATGATTCATTACAAT (forward); 5′:CTATGTTTTTGAGTTTTGAGTTCA (reverse)] were used. Primers against the DNA sequence of myosin were used to measure genomic DNA [5′:TTGTGCAAATCCTGAGACTCAT (forward); 5′:ATACCAGTCCCTGGGTTCAT (reverse)].

Image analysis

For the calculation of TOM20 or VDAC staining, ImageJ was used for image processing. The ImageJ plugin MTrackJ was used to calculate mitochondrial movement (83). For analysis of time-lapse images, a different investigator coded the image sequences so that the analysis was done in a blinded manner. For circularity measurements, a mask of the mitochondria was created in ImageJ (NIH). Huygens software package analysis was used to analyze for colocalization.

Transmission electron microscopy

SCR and PC2 KD cells were plated on glass coverslips and allowed to grow to 90% confluency before fixing with 4% glutaraldehyde. Cells were incubated in 1% osmium tetroxide and dehydrated in increasing ethanol concentrations and then embedded in Durcupan resin. Ultrathin sections were cut on a Leica Ultra-Microtome, collected on Formvar-coated single-slot grids, and analyzed with a Tecnai 12 Biotwin electron microscope (Field Electron and Ion Company, Hillsboro, OR). The investigator collecting the TEM images was blinded to the identity of the samples.

Mitochondrial Ca2+ uptake assay

Cells were plated in a 96-well plate and grown to confluency. The protocol broadly followed previously published studies (84). Briefly, the medium was changed to a potassium buffer [125 mM KCl, 2 mM K2HPO4, 1 mM MgCl2, 20 mM Hepes, 5 mM glutamate, and 5 mM malate (pH 7.0)] with 0.1 μM Calcium Green-5N. After baseline measurements, digitonin (0.02%) was added to permeabilize cells, followed by the serial addition of 200 nM CaCl2 solution in the same potassium buffer. Measurements were conducted on a BioTek spectrophotometer, and experiments were conducted in quadruplicate. Data were collected using Gen 5 software.

Human tissue

ADPKD and normal human kidney tissues were obtained by the PKD Biomaterials Core at the University of Kansas Medical Center. The use of these tissues for research was approved by the institutional review board. Tissues were fixed with 4% paraformaldehyde at 4°C overnight and embedded in paraffin, and 5-μm sections were cut for immunofluorescence. After deparaffinization and rehydration, antigen retrieval was performed by incubating the sections in a steamer for 20 min in sodium citrate buffer [10 mM trisodium citrate and 0.05% Tween 20 (pH 6.0)]. Sections were then quenched of autofluorescence, blocked and permeabilized, and incubated with TOM20 (Cell Signaling) and anti-MFN2 (Cell Signaling) antibodies. Tissues were mounted in ProLong Diamond Antifade Mountant (Life Technologies), counterstained with DBA, and imaged with 100× oil immersion lenses on a Leica SP8 STED microscope using Leica LAS X software (Heidelberg, Germany). To enable comparison between samples, the same laser power, gain, and acquisition settings were used.

Animal studies

Pkd2fl/flPkhd1Cre mice were obtained from the laboratory of S. Somlo (Yale University). Animals were housed under a 12-hour light/dark cycle and genotyped at 3 weeks of age with the primers described previously (85, 86). Pkd2fl/fl littermates were used as controls. All animal experiments were performed in a blinded manner. The Yale University Animal Ethics committee (Institutional Animal Care and Use Committee) approved the animal housing conditions and experimental procedures conducted in this study.

Primary culture of collecting duct cells

Kidneys were isolated from isoflurane-anesthetized mice after perfusion with 0.9% NaCl. The kidney capsule was removed, and the tissue was cut into small pieces. The tissue was incubated at 37°C under continuous agitation (220 rpm) for 1.5 to 2 hours with a mixture of hyaluronidase (1 mg/ml) and collagenase (2.2 mg/ml) for protease digestion. The cell suspension was washed three times with phosphate-buffered saline and incubated at room temperature for 30 min with magnetic beads (Dynabeads 11047, Life Technologies) previously coated with biotinylated DBA. The cells were washed three times with isolation buffer [Dulbecco’s phosphate-buffered saline supplemented with 0.1% bovine serum albumin and 2 mM EDTA (pH 7.4)] and suspended and plated in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin.

Lentiviral shRNA construction

Two different shRNAs against mouse MFN2 were constructed using Broad Institute software (one beginning at base pair 2009 and the other at 3355) and cloned into pSico (Addgene). A scrambled sequence was also constructed. The LV was then produced by cotransfecting the pSico vector harboring the siRNA sequence with the following packing, Res, and Pol plasmids: pRSV-Rev, pMDLg/pRRE, and pMD2.G, respectively. Transfection was achieved using polyethylenimine, and the supernatant containing the virus was harvested at 48 and 72 hours afterward. The effectiveness of the LV was tested on HEK293 cells to determine titer.

The virus was injected into 3-week-old mice by retro-orbital injection. Animals were euthanized 4 weeks after injection, the kidneys were removed for histological analysis, and the cells were isolated for experiments.

Statistical analysis

For cell-based experiments, data were calculated from individual cells and multiple preparations. Where appropriate, one-way analysis of variance (ANOVA) with multiple comparisons or Student’s t test was applied. For population-based experiments (e.g., Western blots), multiple experiments and samples were analyzed and subjected to nonparametric Mann-Whitney U tests. Data are presented as the averaged mean with SEM. In all experiments, P < 0.05 was considered to be statistically significant. *P < 0.05, **P < 0.01, and ***P < 0.001.

SUPPLEMENTARY MATERIALS

stke.sciencemag.org/cgi/content/full/12/580/eaat7397/DC1

Fig. S1. PC2 KD and PC2 KO result in increased mitochondrial Ca2+ signals.

Fig. S2. PC2 resides in ER membranes close to mitochondria.

Fig. S3. PC2 KD does not alter the expression of MCU or multiple MCU complex components.

Fig. S4. MFN2 KD rescues Ca2+ signaling in PC2 KD cells and PC2 KD alters mitochondrial metabolism and density.

Fig. S5. Super-resolution STED microscopy does not alter mitochondrial ROS production.

Fig. S6. PC2 KD does not alter the expression of alternative ER-mitochondrial linkers.

Fig. S7. LV-siMFN2 is effective at knocking down MFN2 in cells from Pkd2F/FPkhd1Cre mice.

Fig. S8. LV-siMFN2 treatment rescues multiple effects of PC2 deficiency in cystic cells.

Fig. S9. Mitochondrial density is increased in cyst-lining cells from ADPKD kidneys.

Movie S1. Super-resolution movie of organelle dynamics.

REFERENCES AND NOTES

Acknowledgments: We thank M. Forte (Vollum Institute, OHSU) for the ratiometric pericam construct and S. Somlo and M. Ma (Yale University) for the Pkd2F/F and Pkhd1Cre mice. We thank D. Stefani (University of Padua) for the mito-GCaMP6F construct. Helpful discussions with L. H. Young, V. G. Zaha, M. A. Carpio, S. G. Katz, J. Hwa, K. Min, A. M. Bennett, K. Martin, and G. Shadel (Yale University) are acknowledged. We thank L. Nguyen (Yale University) and S. Curci (VA, Harvard Medical School) for technical assistance and advice. We thank K. Szigeti-Buck (Yale University) for the TEM sample preparation and imaging and N. Mikush (Yale University) for technical expertise with echocardiograms. We thank S. Kaech and J. Low (Yale University) for the use of the Seahorse Flux analyzer. We thank J. Hwa and Z. Jaji (Yale University) for the use of the spectrometer and A. M. Bennett for the use of the Clark electrode. Funding: We acknowledge the use of the Yale Cell Biology Microscopy Core (NIH grants 5P30DK034989 and OD020142). Federal and NIH grant support is acknowledged: K99 DK101585 (I.Y.K.), 5P01DK057751 and P30DK090744 (B.E.E.), 1F31EB018718 (E.P.K.), VA-ORD 1 I01 BX000968-01 and UL1 TR001102 (A.M.H.), and R24DK106743 (D.L.K.). WT IMCD3 and PC2 KO cells were created by members of the George M. O’Brien Kidney Center at Yale (P30 DK079310). The PKD Biomaterials Core is part of the Kansas PKD Core Center, NIH P30 DK106912 (D.P.W.). Author contributions: I.Y.K. conceived the project. I.Y.K. and A.L.B. conducted most of the experiments and analyzed all data. F.O.L. performed the experiments shown in Fig. 8D and fig. S8 (A and C to E). J.Y.J. and J.L.F. performed the experiments shown in Fig. 1 (C to E). A.M.H., D.P.W., Y.C., K.D., D.L.K., and E.P.K. contributed reagents. I.Y.K. and A.L.B. drafted and edited the manuscript. B.E.E. and A.M.H. edited the manuscript. All authors agreed to and edited the final manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials. The PKD2-mCherry construct requires a material transfer agreement from VA Boston Healthcare System.

Stay Connected to Science Signaling

Navigate This Article