Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Jun 8.
Published in final edited form as: Dev Cell. 2015 May 28;33(5):549–561. doi: 10.1016/j.devcel.2015.04.028

Phosphatidylinositol-Phosphatidic Acid Exchange by Nir2 at ER-PM Contact Sites Maintains Phosphoinositide Signaling Competence

Yeun Ju Kim 1,2, Maria-Luisa Guzman-Hernandez 1,2, Eva Wisniewski 1, Tamas Balla 1,*
PMCID: PMC4476625  NIHMSID: NIHMS695371  PMID: 26028218

SUMMARY

Sustained agonist-induced production of the second messengers InsP3 and diacylglycerol requires steady delivery of phosphatidylinositol (PtdIns) from its site of synthesis in the ER to the plasma membrane (PM) to maintain PtdIns(4,5)P2 levels. Similarly, phosphatidic acid (PtdOH), generated from diacylglycerol in the PM, has to reach the ER for PtdIns resynthesis. Here, we show that the Drosophila RdgB homolog, Nir2, a presumed PtdIns transfer protein, not only transfers PtdIns from the ER to the PM but also transfers PtdOH to the opposite direction at ER-PM contact sites. PtdOH delivery to the ER is impaired in Nir2-depleted cells, leading to limited PtdIns synthesis and ultimately to loss of signaling from phospholipase-C-coupled receptors. These studies reveal a unique feature of Nir2, namely its ability to serve as a highly localized lipid exchanger that ensures that PtdIns synthesis is matched with PtdIns(4,5)P2 utilization so that cells maintain their signaling competence.

Graphical Abstract

graphic file with name nihms-695371-f0001.jpg

In Brief

Kim et al. identify the Nir2 protein as a lipid transporter that moves phosphatidylinositol from the ER to the plasma membrane while transferring phosphatidic acid in the other direction. Therefore, Nir2 helps maintain plasma membrane lipid identity when phosphoinositides are rapidly consumed during receptor-mediated activation of phospholipase C enzymes.

INTRODUCTION

Non-vesicular lipid transfer between the membranes of different organelles has been increasingly recognized as a means by which cells maintain distinct lipid compositions of their different membranes in the face of intense vesicular membrane trafficking (Lev, 2012; Prinz, 2010). Such mechanisms have been recently identified for cholesterol transport between the ER and the Golgi both in yeast (de Saint-Jean et al., 2011) and mammalian cells (Mesmin et al., 2013) or for the transport of ceramide and glucosyl ceramide between the same two membrane compartments (D’Angelo et al., 2007; Hanada et al., 2003). Moreover, other lipid transfer proteins, such as Sec14 or its five homologs in yeast (SfH1–5), have also been found to play important roles in lipid homeostasis in yeast (Bankaitis et al., 2010). Importantly, in almost all cases, phosphoinositides have been implicated in the control of these non-vesicular lipid transfer processes (Kim et al., 2013b).

Ironically, one of the long-outstanding and unresolved questions in lipid signaling is related to the process by which phosphatidylinositol (PtdIns) gets to the plasma membrane (PM) from its site of synthesis in the ER. PtdIns in the PM is the ultimate precursor of PtdIns(4,5)P2, a critically important phosphoinositide that controls a multitude of processes in the PM (Balla, 2013). Consumption of PtdIns(4,5)P2 occurs when phospholipase C (PLC) is activated by cell-surface receptors to generate inositol 1,4,5-trisphosphate (Ins(1,4,5)P3) and diacylglycerol (DG), the two important messengers linked to a variety of cellular responses (Berridge, 1984). Without replenishment of the limited PM PtdIns(4,5)P2 pool from newly synthesized PtdIns, cells are unable to maintain their transmembrane signaling. Early studies of the 1950s and 1960s (Hokin and Hokin, 1955, 1964) have postulated that the products of PLC activation (at that time, it was believed that PLC acts directly on PtdIns) get recycled. DG is phosphorylated to phosphatidic acid (PtdOH), which is then converted back to PtdIns via two enzymatic steps, with CDP-DG being an intermediate (Agranoff et al., 1958) (Figure 1A). For this “cycle” to work efficiently, PtdIns has to be transferred from the ER to the PM and PtdOH has to be transferred from the PM to the ER. PtdIns transfer proteins (PITPs) have long been identified (Cockcroft and Garner, 2011; Ile et al., 2006), but it is still not clear which, if any, are important for PtdIns transfer between the ER and the PM. In contrast, PtdOH transfer proteins between the PM and the ER have not yet been described, neither is it understood how PtdIns and PtdOH transfer is coordinated between these two membranes.

Figure 1. Nir2 Depletion Affects Phosphoinositide Cycle Intermediates.

Figure 1

Open in a new tab

(A) Current models of the phosphoinositide cycle and its compartmental distribution.

(B) Labeling HEK293-AT1 cells with myo-[3H]inositol (24 hr) after 3-day knockdown of the indicated PITPs (see Figure S1A for analysis of knockdown efficiency). Lipids were extracted and separated by TLC and quantified by densitometry of exposed films from TLC plates treated with Enhancer. Grand means ± SEM of three experiments are shown, each performed in duplicate.

(C) After 3-day knockdown with control or Nir2 siRNAs, cells were labeled with 32P-phosphate for 1 hr followed by a 10-min treatment with AngII (10−7 M) or solvent. Lipids were extracted and separated by TLC, then subjected to phosphorimage analysis. Grand means ± SEM of five experiments are shown, each performed in duplicate.

(D) Calculated PtdOH/PtdIns activity ratios normalized to absolute controls (grand means ± SEM, n = 5; **p < 0.05, ***p < 0.001, using two-way ANOVA).

(E) Incorporation of [3H]cytidine into CDP-DG in HEK293-AT1 cells. Cell were treated with the indicated siRNAs for 3 days and inositol depleted for the last 16 hr. Cells were then stimulated with AngII (10−7 M) for 1 hr in the presence of the isotope with or without pretreatment with 10 mM Li+ for 30 min prior to stimulation. Top panel shows representative TLC, and the graph below shows quantification (means ± range) from two independent experiments. This result was reproduced in ten other experiments in the course of these studies. For rescue experiments, see Figure S1C.

(F) Effect of Nir2 knockdown on myo-[3H]inositol incorporation into phosphoinositides during a short labeling (2 hr) in the presence or absence of AngII (2 hr) (see Figure S1B for myo-[3H]inositol incorporation kinetics).

During studies of these questions, we observed that the Nir2 protein was important for CDP-DG and PtdIns synthesis during agonist stimulation. Nir2 and Nir3 (also called PITPnm1 and PITPnm2) are mammalian homologs of the Drosophila RdgB protein possessing PITP domains at their N termini (Trivedi and Padinjat, 2007). RdgB mutant flies undergo retinal degeneration, presumably because of a defect in the maintenance of PtdIns(4,5)P2 in their photoreceptor membranes (Milligan et al., 1997). According to current views, RdgB transfers PtdIns from the ER to the PM (Trivedi and Padinjat, 2007). Recent studies suggested that Nir2 supports the maintenance of PtdIns(4,5)P2 pools via its PtdIns transfer function at ER-PM contact sites (Chang et al., 2013; Kim et al., 2013a). Here, we report that Nir2 is an intricate lipid transfer device that transports PtdOH from the PM to the ER, by a process that is tightly coupled to PtdIns(4,5)P2 hydrolysis and PtdIns transport in the other direction. Nir2 is recruited to ER-PM junctions during PLC activation and is critically important for the utilization of PtdOH generated in the PM for PtdIns synthesis in the ER. Nir2 is, therefore, a critical component of the PI cycle that, via its lipid exchange function, coordinates PtdIns synthesis and transfer with PtdIns(4,5)P2 utilization to ensure maintenance of this important signaling lipid in the PM.

RESULTS

Nir2 Is Important for Both PtdIns and PtdOH Metabolism

To identify the transfer protein(s) responsible for PtdIns transfer between the ER and the PM, we performed RNAi experiments targeting the known mammalian PITP proteins. We measured [3H]-inositol incorporation into phosphoinositides in intact cells with the idea that inhibition of PtdIns transfer between the ER and PM would impede the flux of 3H from PtdIns made in the ER to PtdIns(4,5)P2 that is synthesized in the PM (Figure 1A). The expectation was that 3H-inositol labeling of PtdIns(4,5)P2 would be selectively impaired when the relevant PITP was eliminated. Inactivation of either or both class I PITPs by RNAi-mediated gene silencing in HEK293 cells stably expressing the AT1a angiotensin receptors (HEK293-AT1) (Figure S1A) had no major effect on a 24-hr 3H-inositol labeling of PtdIns or its phosphorylated forms; in fact, their labeling was slightly increased (Figure 1B). In contrast, knockdown of Nir2 had a profound inhibitory effect on 24-hr 3H-labeling, and this equally affected PtdIns and its phosphorylated forms, PtdIns4P and PtdIns(4,5)P2 (Figure 1B).

Since 24-hr labeling labels the inositide pools to ~70% of equilibrium (Figure S1B), these findings could be caused either by a decrease in the size of the PtdIns pools or by a defect in PtdIns synthesis manifested at the end of the 3-day knockdown period. To further analyze this question, we performed experiments with shorter-term 32P-phosphate labeling and agonist stimulation. The advantage of this approach is that a shorter labeling period informs on the metabolic fluxes between PI cycle intermediates rather than on their absolute levels. Cells were labeled with 32P-phosphate for 1 hr and stimulated with angiotensin II (AngII) for the last 10 min in the presence of the isotope. Phospholipid analysis showed that Nir2 downregulation decreased the 32P-labeling response to AngII in both PtdOH and PtdIns, but the 32P-labeling of PtdIns was more profoundly affected (Figure 1C). While the decrease in PtdOH 32P labeling could reflect a smaller DG response, the larger decrease in PtdIns labeling suggested an additional block between PtdOH formation and PtdIns synthesis. We calculated PtdOH/PtdIns ratios as a more sensitive indicator of a block in the flux of 32P from PtdOH to PtdIns. These ratios showed an increase after AngII stimulation, which was then significantly augmented in Nir2-knockdown cells (Figure 1D).

Nir2 Is Required for CDP-DG Synthesis during Agonist Stimulation

The decreased [3H]-Ins labeling of PtdIns together with a reduced [32P] flux between PtdOH and PtdIns in Nir2-depleted cells raised the possibility that some of the PtdOH may not have been used for PtdIns synthesis in AngII-stimulated cells. This would occur if the transfer of PtdOH between the PM and the ER was impaired. In this case, one would expect a defect in CDP-DG synthesis in the ER in Nir2-depleted cells after AngII stimulation. Therefore, we measured CDP-DG levels in cells labeled with [3H]-cytidine and knocked down Nir2 with different duplexes (Figure S1A). As observed earlier by other laboratories (Stubbs and Agranoff, 1993), agonist stimulation alone did not elicit a measurable increase in [3H]-CDP-DG levels. However, when cells were depleted in myo-inositol overnight and treated with 10 mM Li+ for 30 min, a procedure that blocks InsP monophosphatases (Berridge et al., 1982) and deprives PtdIns synthesis with one of its substrates, myo-inositol (Allison and Stewart, 1971), AngII stimulation led to a significant accumulation of [3H]-CDP-DG (Figure 1E). This confirmed conclusions from earlier studies that most of the CDP-DG produced as a result of phosphoinositide breakdown is recycled back to PtdIns via CDP-DG.

Importantly, Nir2 knockdown substantially decreased the AngII-stimulated [3H]-CDP-DG response, suggesting a defect in recycling the PLC-generated lipid products for CDP-DG synthesis in Nir2-depleted cells. Shorter (2-hr) 3H-inositol labeling (in this case, without inositol depletion and without Li+ treatment) also showed that Nir2 depletion resulted in decreased [3H]-Ins incorporation into PtdIns, consistent with a defect in CDP-DG formation (Figure 1F). Several Nir2 small interfering RNAs (siRNAs), some targeting the non-coding region, were tested to rule out off target effects. These yielded similar results, and the severity of their effects appeared to be proportional to the level of Nir2 knockdown (Figures 1C–1F and S1A).

Initial efforts to rescue the effects of Nir2 knockdown by over-expression of RNAi-resistant Nir2 constructs were unsuccessful, often showing even further inhibition rather than rescue in 3H-cytidine labeling experiments. This was caused by cytotoxicity of the overexpressed Nir2 protein during prolonged expression, the reason for which is still under investigation. Careful titration of the expression levels of Nir2 using a set of plasmid with increasing truncations within the cytomegalovirus (CMV) promoter (described in Morita et al., 2012) revealed that rescue was possible at low levels of Nir2 expression (Figure S1C). Because of the cytotoxic effects of both the Nir2 knockdown and prolonged Nir2 overexpression, rescue experiments were difficult to properly dose and evaluate.

Nir2 Is Important for Sustained PLC Signaling

To determine whether Nir2 knockdown had an impact on the ability of the cells to maintain signaling, we tested several cellular responses. Cytoplasmic Ca2+ measurements showed that cells depleted in Nir2 could not maintain increased [Ca2+]i levels in response to AngII stimulation (Figure 2A). Recent studies suggested that Nir2 depletion resulted in a decrease in PtdIns(4,5) P2 levels, which resulted in decreased signaling (Chang et al., 2013; Kim et al., 2013a). To measure the impact of Nir2 depletion on the level of PtdIns(4,5)P2 pool, we used fluorescence resonance energy transfer (FRET) measurements using the PtdIns(4,5)P2 binding domain of Tubby (Szentpetery et al., 2009). Lipid binding domains fused to two different fluorophores (mTq and Venus) can serve as donor and acceptor pairs for energy transfer when they are bound to the lipid in the PM (van Der Wal et al., 2001). Addition of ionomycin (10 μM) leads to a complete breakdown of PtdIns(4,5)P2, causing the Tubby domain to fall off the membrane and thereby decreasing the FRET signal. The size of the FRET change in response to complete elimination of PtdIns(4,5)P2 can inform about the size of the resting PtdIns(4,5)P2 pool. These experiments showed only a moderate but statistically significant reduction of the PtdIns(4,5)P2 pool in Nir2-depleted cells (Figure 2B).

Figure 2. Nir2 Is Needed for Sustained PLC Signaling.

Figure 2

Open in a new tab

(A) Cytoplasmic Ca2+ responses of Fura-2-loaded cells stimulated with AngII. Cells were treated with control siRNA or siRNA#1 against Nir2 for 3 days before Ca2+ measurements. Means ± SEM from 60 to 70 cells.

(B–D) FRET measurements to quantify lipid changes in live cells showing schematics of the principles. FRET signal was measured by expressing mTq and Venus-tagged fluorescent reporters that bind specifically to the lipids indicated. Cells were treated with Nir2 siRNA #1 or control siRNA for 3 days and FRET sensors were expressed for the last day. (B) The Tubby domain was used to assess resting levels of PtdIns(4,5)P2. The magnitude of FRET change after a complete elimination of PtdIns(4,5)P2 by treatment with 10 μM ionomycin is proportional to the resting PtdIns(4,5)P2 levels. Means ± SEM, n = 77 and 66 for control and Nir2 siRNA groups, respectively, obtained in three independent knockdown experiments (p < 0.05 by unpaired t test). (C and D) Kinetics of DG (C) and PtdOH (D) changes after AngII (10−7 M) stimulation in control and Nir2-depleted cells using FRET analysis. Means ± SEM are shown from a total of 150–200 cells obtained in two or three separate dishes from three independent experiments. The translocations of wild-type and mutated Spo20 probe are shown in Figure S2.

Next, we followed the DG and PtdOH responses of the cells, applying the same FRET principle, using a DG sensor based on the C1ab domain of PKD described previously (Kim et al., 2011) and a PtdOH sensor based on a short sequence of the yeast Spo20 protein (Kassas et al., 2012). The Spo20 sensor was modified for this analysis by adding a nuclear export signal (NES) to counter its high nuclear localization and mutating two of the cysteines that reacted with the PLC inhibitor, U73122. These modifications allowed the probe to be used as a faithful PtdOH sensor (Figure S2). These results showed that the DG response was smaller and more transient in the Nir2-depleted cells (Figure 2C). Unexpectedly, the PtdOH response was increased rather than decreased in the cells knocked down for Nir2 (Figure 2D).

The decreased Ca2+ and DG responses were consistent with diminishing PtdIns(4,5)P2-derived signals in Nir2-depleted cells, as observed in recent studies (Chang et al., 2013; Kim et al., 2013a). However, the enhanced PtdOH response of Nir2-depleted cells suggested either an increased DG kinase activity or a decreased PtdOH utilization at the PM. The results of the metabolic labeling studies (Figure 1) made the latter explanation more plausible.

Nir2 Is Enriched in ER-PM Contact Zones following PLC Activation

Recent studies showed Nir2 in PM in ER-PM junctions after receptor stimulation (Chang et al., 2013; Kim et al., 2013a). We observed a similar translocation after AngII stimulation in the HEK-AT1 cells (Figure 3A) and also in ATP-stimulated COS-7 cells or carbachol-stimulated CHO cells (Figure S3A). Although expressed GFP-Nir2 was mostly cytosolic, association of the protein with the ER membranous network could be observed in cells showing low expression levels (Figures 3A and S3A). The ER association was more prominent when the cytosolic fraction of the protein was released with saponin permeabilization. This also revealed association of Nir2 primarily with part of the ER that followed the microtubule architecture (Figure S3B). No localization to the Golgi was observed in live cells, even if a higher ER density was observed around the Golgi region (Figure S3C). The punctate appearance of Nir2 after PLC activation completely overlapped with that of the ER Ca2+ sensor protein, STIM1 (Liou et al., 2005) (Figure 3B). Importantly, however, depletion of the ER Ca2+ stores with thapsigargin, a treatment that causes a strong STIM1 translocation response, failed to elicit any change in Nir2 distribution (not shown). The latter was only induced after PLC activation and was quickly reversed after the addition of the PLC inhibitor, U73122 (Figure 3A). These results confirmed that Nir2 was an ER-associated peripheral protein in the quiescent state but after PLC activation it also formed contacts with the PM. To determine if the punctate appearance was due to the recruitment of the protein to specialized PM regions or was caused by its association with the ER-bound VAP-B protein (Amarilio et al., 2005; Peretti et al., 2008), the localization response of the FFAT mutant (FF350,351>AA) Nir2 was studied. This mutant protein was fully cytosolic but rapidly translocated uniformly to the PM (Figure 3C) (these images show the middle section of cells, so the puncta appear in the cell periphery as opposed to the whole surface that is seen when the bottom of the cell was imaged). This finding argued against the role of specialized regions in the PM as the cause of the punctate pattern after stimulation.

Figure 3. Nir2 Is ER-Associated in Resting Cells and Translocates to ER-PM Contact Sites following PLC Activation.

Figure 3

Open in a new tab

Confocal images of live cells are shown.

(A) GFP-tagged Nir2 expressed at a low level shows faint ER localization in quiescent cells that changes into a punctate pattern after stimulation, as best seen in the bottom section of the cells. Punctation is reversed by inhibition of PLC by U73122 (5 μM). Figure S3A shows Nir2 translocation in other cells stimulated by other PLC activating receptors.

(B) Nir2 puncta in stimulated cells co-localize with ER-PM contact sites as marked by Stim1 proteins in COS-7 cells. Top panels show localization of the two proteins before ATP (50 μM) stimulation and their merged image. Bottom panels show localization after ATP stimulation (5 min) and the enlarged overlay from the area indicated by the white box. Close inspection of these areas indicates that Nir2 occupies slightly larger areas than Stim1 in the ER-PM junctions. Figure S3 shows co-localization of Nir2 with microtubules in quiescent cells after permeabilization (Figure S3B) and lack of Golgi localization (Figure S3C).

(C) GFP-Nir2 mutated in the FFAT domain is fully cytoplasmic in quiescent cells and translocates evenly to the PM after stimulation. Note that here, the middle sections of the cells are shown.

(D) GFP-VAP-B and mCherry-Nir2 were transfected in COS-7 cells. Two proteins were co-localized in the ER before stimulation (upper rows), and after ATP (50 μM) stimulation for 5 min, both the Nir2 and VAP-B proteins show clustering (lower rows).

(E) COS-7 cells expressing mCherry-Nir2 and CDS2-GFP were imaged before and after stimulation with ATP (50 μM) for 5 min. Note the localization of Nir2 puncta at ER regions that also show some enrichment in CDS2, although these are not as prominent as with VAP-B. Bars are 10 μm. See Figures S3D and S3E for TIRF analysis of co-localization of Nir2 and CDS2 and the confocal analysis of the Nir2 and PIS proteins, respectively.

To determine whether VAP-B expression increases the ER-associated fraction of Nir2 and whether VAP-B also redistributes during Nir2 translocation, we co-expressed mCherry-tagged Nir2 and a GFP-VAP-B construct in HEK293-AT1 cells and observed the distribution of both proteins before and after AngII stimulation. As shown in Figure 3D, VAP-B expression increased the ER association of Nir2 and after stimulation both protein showed clustering and co-localization in the junctional compartment. It was also of interest that Nir2 and CDS showed co-localization during agonist stimulation (Figure 3E). This co-localization in the contact zones was even more prominent in total internal reflection fluorescent microscopy (TIRF) images (Figure S3D); here, Nir2 was not detected at the membrane in quiescent cells, but after stimulation, it rapidly moved to PM-adjacent puncta, where it showed close co-localization with the CDS2 protein. It was difficult to determine whether the CDS2 protein showed further enrichment in the same compartment in response to stimulation. Notably, Nir2 showed no co-localization with the rapidly moving particles of PI synthase (PIS) enzyme only with the fraction that was present in the tubular ER (Figure S3E).

Nir2 Facilitates the Clearance of PtdOH from the PM

All of the data above, together with the localization of Nir2 in the ER-PM contact zones, raised the question as to whether Nir2 facilitates the transfer of PtdOH between the PM and the ER. To test whether Nir2 affects the clearance of PtdOH from the PM, we stimulated cells with AngII for 10 min to allow the accumulation of PtdOH. At this point, we added a DG kinase inhibitor (50 μM R59022) to block further PtdOH production and examined the rate at which PtdOH is removed from the PM. Because of the light-induced inactivation of the inhibitor at 430 nm illumination, we used a red version of the PtdOH sensor (NES-mdsRed-Spo20) and measured its cytosolic intensity changes in confocal microscopy to follow its membrane association (the 568 laser line had a less damaging effect on the inhibitor). AngII stimulation rapidly increased DG at the PM, which was reflected in the recruitment of the DG probe to the PM with concomitant decrease in the cytosolic fluorescence (note that the decrease in cytosolic intensity is plotted upward to reflect the direction of lipid changes in the membrane). It took a few minutes before the PtdOH probe translocated to the PM, and its increase coincided with the dropping DG levels. Inhibition of DG kinases caused an immediate increase in PM DG levels and initiated a decline in PtdOH levels (Figure 4A). To study the effect of Nir2 on the disappearance of PtdOH from the PM, we expressed GFP-Nir2 and used this treatment protocol comparing the Spo20 probe translocation kinetics between cells expressing GFP-Nir2 and those expressing only the empty plasmid (Figure 4B). Remarkably, in cells expressing GFP-Nir2, the accumulation of PtdOH after AngII stimulation was reduced and more importantly, PtdOH disappeared from the PM faster and more completely after addition of the DG kinase inhibitor (Figure 4B).

Figure 4. Nir2 Expression Facilitates Removal of PtdOH from the PM.

Figure 4

Open in a new tab

(A) Kinetics of DG (black) and PtdOH (red) changes after AngII (10−7 M) stimulation in HEK293-AT1 cells expressing GFP- and mdsRed-fused DG and PtdOH sensors, respectively (see Experimental Procedures for details). Transfected cells were imaged live by confocal microscopy, and the cytoplasmic intensity was monitored in the cytosol in regions of interest outside the nuclear area. Translocation to the PM is associated with a decrease in cytoplasmic fluorescence. Decreased intensities are plotted upward for easier conceptualization of the direction of changes in membrane lipid concentrations. AngII (10−7 M) and the DG kinase inhibitor, R59022 (50 μM), were added at the indicated times. The detailed time course is shown as means ± SEM of 11–12 cells obtained in one representative experiment reproduced three times.

(B) Expression of GFP-Nir2 facilitates the removal of PtdOH from the PM. Representative pictures show a group of cells after AngII stimulation followed by DG kinase inhibition at times indicated by the dotted lines. Note that cells that do not express GFP-Nir2 (white arrows) still show PM localization of the PtdOH sensor, while the localization is completely eliminated in those expressing the GFP-Nir2 protein. Scale bar, 10 μm.

(C) Quantification of cytoplasmic fluorescence changes in similar experiments using various mutant Nir2 constructs or the RdgBβ protein. Means ± SEM from 20 to 40 cells obtained in three separate experiments are shown.

We next examined whether this effect was related to the PITP domain and performed similar experiments with mutant GFPNir2 proteins. Mutant Nir2 lacking the PITP domain was unable to facilitate PtdOH clearance (Figure 4C, top), and mutation of the FFAT domain (FF350,351>AA), which eliminates interaction with the ER-bound VAP-B protein, greatly reduced, but did not completely eliminate, the effect (Figure 4C, top). Point mutations at the T59 residue (T59A or T59E) within the Nir2 PITP domain that had previously been shown to affect PtdIns transfer in vitro (Milligan et al., 1997) also prevented the stimulatory effect on PtdOH clearance (Figure 4C, middle). Conversely, the isolated PITP domain of Nir2 was partially active (comparable to the FFAT mutant), while RdgBβ was without effect (Figure 4C, bottom). These results together suggested that Nir2 is able to facilitate the clearance of PtdOH from the PM and the next step was to assess where this PtdOH was utilized.

Importantly, PtdOH was not detected in the ER with the Spo20-based PtdOH sensor, even when the clearance of PtdOH from the PM was enhanced by Nir2. This most likely reflected the rapid utilization of PtdOH in the ER by the CDS enzymes. To test this possibility, we performed 32P-phosphate labeling experiments in cells expressing Nir2. Moderate overexpression of Nir2 (an untagged version was used in these studies), but not its PITP-domain-deleted mutant, showed an increased 32P-phosphate labeling of PtdIns after a 10-min AngII treatment (Figure 5A). Importantly, this was paralleled by a decreased 32P-phosphate labeling of PtdOH (Figure 5A). The increased conversion of PtdOH to PtdIns was most apparent as a significant decrease in the 32P-labeled PtdOH/PtdIns ratios in AngII-stimulated cells expressing Nir2, but not its PITP-less version (Figure 5B). In further experiments using CDP-DG labeling, moderate Nir2 overexpression also increased the amount of CDP-DG formed, and this effect was not reproduced by the PITP-domain-deleted Nir2 (Figure 5C). These results together were consistent with an enhanced PtdOH to PtdIns conversion by Nir2 as a consequence of enhanced transfer of PtdOH from the PM to the ER.

Figure 5. Effects of Nir2 Overexpression on PtdOH Labeling and CDP-DG Production.

Figure 5

Open in a new tab

(A) 32P-labeling with or without AngII (10−7 M) stimulation in HEK293-AT1 cells transfected with empty vector, Nir2, or ΔPITP-Nir2 for 1 day. Representative image of a TLC analysis and results of quantification of PtdOH (upper) and PtdIns (lower) from three experiments performed in duplicate (grand mean ± SEM).

(B) Calculated PtdOH/PtdIns ratios (normalized to absolute controls) were significantly decreased in cells expressing Nir2 compared to those transfected with vector or ΔPITP-Nir2 (grand mean ± SEM, n = 5 duplicates, **p < 0.01; ***p < 0.001 using two-way ANOVA).

(C) [3H]cytidine incorporation into CDP-DG in HEK293-AT1 cells transfected with empty vector, Nir2, or ΔPITP-Nir2. Values were normalized to those obtained with AngII-stimulated Nir2-expressing cells (grand mean ± SEM, n = 8 duplicates; ***p < 0.005, using two-way ANOVA).

Mechanism of the PM Association of Nir2

Nir2 has multiple domains that show homology to those found in various protein families (Figure S4A). Notable among these are the N-terminal PITP domain and an LNS2 domain toward the C terminus also found in lipins (Kim et al., 2013a). Nir2 also has a DDHD domain found in a family of proteins comprising PLA1/DDHD1, p125/Sec23-interacting protein, and KIAA0725p/DDHD2 (Inoue et al., 2012). This latter domain has been claimed to bind PtdIns4P (Inoue et al., 2012). We also noted that the LNS2 domain is preceded by a short sequence that is not homologous in lipins but is present in a group of bacterial and archaeal proteins called “PITPs” that also have LNS2 domains but lack PITP domains (Figure S4A). This small segment showed vague similarity to the DG binding part of C1 domains. We designated this conserved segment as DGBL (for DG-binding-like) (Figures S4A and S4B).

Although the FFAT-mediated association of Nir2 with the VAP proteins in the ER has been established (see above and Amarilio et al., 2005), the mechanism of PM association of the protein is less understood. Recent studies showed that Nir2 binds to PtdOH via its LNS2 domain (Kim et al., 2013a). However, analysis of the kinetics of Nir2 PM association after stimulation showed that it was significantly faster than the recruitment of the PtdOH sensor, Spo20, yet it was still slower than the DG increase (Figure 6A). Remarkably, a PITP-deleted construct showed a PM-association response that was significantly faster and greatly enhanced compared to Nir2 wild-type (Figure 6B), indicating that the PITP domain was not responsible for PM association and exerted a negative, possibly regulatory effect on membrane interaction.

Figure 6. Kinetic Analysis of the Membrane Association of Nir2 Constructs Relative to DG and PtdOH Changes.

Figure 6

Open in a new tab

HEK293-AT1 cells were co-transfected with the indicated constructs for one day and observed in confocal microscopy. Cytoplasmic fluorescence changes were recorder in regions of interest outside the nuclear area and expressed as percent of pre-stimulatory values. Translocation to the PM is associated with a decrease in cytoplasmic fluorescence. Decreased intensities are plotted upward for easier conceptualization of the direction of changes in membrane lipid concentrations.

(A) Translocation of GFP-Nir2 is faster than the PtdOH increase (measured by Spo20m) but slower than that of DG (measured by the GFP-C1ab domain of PKD;Kim et al., 2011). Means ± SEM from 19 to 33 cells obtained in three independent experiments.

(B) Removal of PITP greatly enhances the speed and extent of membrane association (means ± SEM of six to eight cells from a representative experiment reproduced four times). Both the basal ER association and the AngII-induced punctuation of this construct are enhanced relative to full-length Nir2 (lower panels).

(C) Schematic representation of the Nir2 constructs tested in this study for their membrane association in response to AngII or DiC8-DG. Black lines represent constructs with no membrane recruitment, and red and orange colors mean good and intermediate recruitments, respectively. Figure S4 shows homology of the various domains of Nir2 with other proteins and the sequence alignment of the DG-binding-like domain (DGBL) with C1 domains and bacterial “PITP” proteins.

(D) Rapid membrane association of the Nir2(420–1,181) and Nir2(816–1,181) constructs after AngII stimulation alone (brown and cyan traces) or followed by DGK inhibition (red and blue traces). Means ± SEM from eight to nine cells obtained in one representative experiment reproduced two times.

(E and G) Membrane association of the Nir2(420–1,181) construct and its mutant (V934,L935>AA) after DiC8-DG (30 μM) (E) or AngII (G) treatment. Means ± SEM from n = 36–94 cells obtained in three to five dishes from three independent experiments.

(F) Membrane association of the indicated full-length Nir2 constructs after AngII treatment followed by DGK inhibition. Means ± SEM from 19 to 34 cells obtained in four to six dishes from three independent experiments. See Figure S4 for localization of and translocation of Nir2(420–1,181) and Nir2-VL>AA mutants.

Serial truncations from the N and C termini were then generated (Figure 6C), and the PM translocation responses of these fragments were analyzed after AngII stimulation followed by DGK inhibition (Figure 6D). A small deletion from the C terminus (1182stop) had no effect on the localization response. However, a larger deletion (1118stop) completely eliminated membrane recruitment. Deletion of the entire N terminus up to residue 966 also eliminated membrane recruitment. These experiments showed that the narrowly defined LNS2 domain was necessary, but not sufficient, for membrane recruitment. Addition of the DGBL segment to the LNS2 domain (residues 816–1,181) restored the PM localization response, which was further enhanced by extension of the construct toward the N terminus (420–1,181) (Figure 6C). Importantly, the PM association of these constructs was not reversed after DGK inhibition (Figure 6D).

The fast membrane-association kinetics and the lack of its reversal by DG kinase inhibitors suggested that PtdOH might not be the sole lipid responsible for the membrane association of Nir2. The requirement for the DGBL segment and the fast localization response raised the possible role of DG as another lipid binding factor in addition to PtdOH. This was tested by using DiC8-DG added to the cells expressing the GFP-Nir2(420–1,181) or GFP-Nir2(816–1,181) constructs. This treatment caused rapid translocation of both the DG sensor (Kim et al., 2011) and the Nir2 construct (Figure 6E) to the membrane, even after inhibition of DGK (Figure S4F). However, truncations of this construct from the C terminus (420–1,117) or (420–978) completely prevented its membrane recruitment response to AngII stimulation. Mutation of the V934A/L935A residues within the DGBL domain in the GFP-Nir2(420–1,181) construct reduced, but did not eliminate, the response to DiC8-DG, and also reduced the response to AngII stimulation (Figure 6G).

To further analyze the role of the PITP domain and the DG-binding-like segment in the full-length Nir2 protein, we generated point mutations within these respective domains (T59A and T59E and V934A/L935A, respectively). The T59 mutants showed stronger membrane association with similar kinetics to the wild-type Nir2, while the VL>AA mutant showed a very muted membrane translocation response (Figure 6F). These results together suggested that the PM interaction is a more complex process mediated by a combined action of DG and PtdOH binding by the C-terminal segment, requiring both the LNS2 and the DGBL segment. Importantly, this process was negatively controlled by the PITP domain.

DISCUSSION

The present experiments were initiated to address the question of which, if any, of the PITP proteins contribute to the transfer of PtdIns between the ER and the PM. Mammalian PITP proteins have been thoroughly studied and well characterized mostly using in vitro systems, but it has been an unresolved question whether they, indeed, are the proteins responsible for the transfer of PtdIns between the ER and PM (Cockcroft and Garner, 2011; Ile et al., 2006). Our attention was drawn to the Nir2 protein because it was the only one that had profound effects on the labeling of phosphoinositides. Our studies revealed that the Nir2 protein functions in ER-PM contact zones and responds to stimulation with PLC-activating stimuli by translocation to these junctional sites. While our experiments approached their conclusions, three studies reported important findings on this topic. The paper by the Cockcroft group has reported that RdgBβ (a protein that has high homology to Nir2 PITP domain but lacks the other parts of the Nir2 molecule) is capable of PtdOH binding in vitro (Garner et al., 2012). Two recent studies from the Lev and Liu groups showed that Nir2 translocated to the PM in stimulated cells and that the Nir2 protein was important for the maintenance of PtdIns(4,5)P2 and PtdIns(3,4,5)P3 pools in the PM, presumably due to the protein’s PtdIns transfer function (Chang et al., 2013; Kim et al., 2013a).

Our studies are in good agreement with those findings. However, they also revealed an additional feature of Nir2; namely, its ability to function as a PtdOH transfer protein that operates within ER-PM contact zones. Our studies showed that Nir2-depleted cells accumulate PtdOH in their PM and are impaired in converting PtdOH to CDP-DG and PtdIns during PLC activation. The moderate decrease in PtdIns(4,5)P2 levels in Nir2-depleted but unstimulated cells suggest that cells can maintain this important PM lipid even when PtdIns transfer is impaired and PtdIns pools are shrinking in the ER. Our structural analysis of the membrane association of Nir2 in activated cells suggested a more complex recruitment process than previously described (Kim et al., 2013a). The inhibitory effect of the PITP domain on membrane localization indicates that lipid binding to the PITP domain is probably an important clue for the molecule to assume a more open conformation that allows the other domains to bind both the ER and the PM. None of our deletion constructs followed either DG or PtdOH kinetics; rather, their PM association kinetics suggested that both DG and PtdOH binding was important for the membrane localization of the C-terminal segments of the protein. Whether these two lipids bind to the same or different sites remains be established. It is important to note that all of our conclusions were reached based on experiments in intact living cells using an array of methods from biochemical analysis to dynamic confocal imaging.

These findings fill an important gap in our understanding of the so-called PI cycle, whereby the products of the PLC-mediated breakdown of PtdIns(4,5)P2 in the PM are recycled back into the ER. Functioning as a PtdOH transfer protein, Nir2 plays a pivotal role in localizing the agonist-induced facilitation of the PI-cycle to ER-PM contact zones establishing a remarkable organization. Since the transfer of PtdOH is “upstream” of the PtdIns transfer process in the PI cycle when accelerated by PLC activation, it is not possible to draw a conclusion about the role of Nir2 as a PtdIns transfer protein in the current study (and in the other aforementioned studies). However, generation of PtdOH by PLD2 in the PM failed to support CDP-DG production even though it did cause recruitment of the Nir2 protein to the contact zones (Y.J.K. et al., unpublished data). Only after PLC activation was the PtdOH delivered to the CDS enzyme(s). This strongly suggested that the PtdOH transfer was a feature that required some form of priming and was likely coupled to PtdIns transfer. The well-documented ability of the Nir2 PITP domain to transfer PtdIns (Garner et al., 2012) in vitro served as the basis of the accepted view that Nir2 transfers PtdIns from the ER to the PM, and nothing in our findings counters those conclusions. Taking all these data together, we propose that the Nir2 protein works essentially as a PtdOH-PtdIns exchanger at the ER-PM contact zones. These findings also explain several previous puzzling results describing elements of the whole PI cycle associated with both the PM (McPhee et al., 1991; Vaziri et al., 1993) and the ER (Helms et al., 1991) in cell fractionation studies. The ER-PM compartment could have been recovered in either the ER- or PM-enriched fractions in those studies.

The Drosophila RdgB protein, the fly ortholog of Nir2, is responsible for the inability of these mutant flies to terminate light-induced electrical signal in electroretinograms and for the light-induced retinal degeneration (Vihtelic et al., 1991). The presence of a PITP domain that is highly homologous to the class I PITPs, PITPα, and PITPβ (Vihtelic et al., 1993), together with the importance of the phosphoinositide cycle in invertebrate photosignal transduction, led to the assumption that RdgB functions as a PtdIns transfer protein. This was corroborated by the findings that the PITP domain of RdgB was necessary and sufficient to rescue the mutant phenotype (Milligan et al., 1997). However, several observations suggested that there was more to RdgB function than a simple PtdIns transfer. First, class I mammalian PITPα or a chimeric RdgB having the mammalian PITPα was unable to rescue the phenotype (Milligan et al., 1997). Second, RdgB proteins mutated within the PITP domain (T59E) showed a dominant-negative effect in spite of the ability of this particular mutant to transfer PtdIns (Milligan et al., 1997). Third, the problem with signal termination was inconsistent with a simple rundown of the PtdIns(4,5)P2 pools that was expected if the protein simply worked as a PtdIns transfer protein (Trivedi and Padinjat, 2007). Our results provide a possible explanation for all of these findings. Defective PtdOH removal from the PM could explain the signal termination defect. The PtdOH transfer function being upstream of the presumed PtdIns transfer function also explains the lack of correlation with simple PtdIns transfer. Nevertheless, it is most likely that the protein works in a dual capacity, transferring both PtdOH and PtdIns between the ER and PM. The localization of the RdgB protein in the fly retina in the subplasmalemmal portion of the ER called sub-rhabdomeric cisternae (Suzuki and Hirosawa, 1994; Vihtelic et al., 1993) is also in good agreement with the presence of the Nir2 protein in the ER-PM junctions in stimulated mammalian cells.

It has been suggested that lipid transfer proteins may not simply function as diffusible vehicles mediating lipid transfer between membranes in a way that happens in simple lipid transfer assays (Mousley et al., 2012). It was proposed that PITPs function as devices that assign PI lipids to various enzymatic reactions in a strictly regulated biological context. Our results are fully compatible with these views. Although we propose that Nir2 allows transfer of lipids between two adjacent membranes, it does so only in a specific biological context. Nir2 supplies PtdIns to the PM, but only when PtdIns(4,5)P2 is consumed by PLC enzymes producing both DG and PtdOH, and all this happens in restricted membrane domains formed by contacts between the ER and the PM. While our present studies and those conducted in flies suggest that a diffusible PITP domain of Nir2/RdgB is capable of restoring the lipid transfer function, the precise and metabolically isolated function of these proteins requires their localization to the contact sites.

It is also important to emphasize that Nir2 is unlikely to be the sole protein responsible for either PtdOH or PtdIns transfer or metabolic channeling. Other PtdIns and PtdOH transfer domains have been identified in eukaryotic cells, such as the classical PITPs, Nir3, RdgBβ, PRELI, and Sec14 domain-containing proteins (Bankaitis et al., 2010; Lev, 2012). Nir3-knockout mice have been generated but showed no obvious phenotype (Lu et al., 2001), and Nir3 knockdown in our hands did not produce the same effects observed with Nir2 depletion (Y.J.K. and T.B., unpublished data). Also, RdgBβ failed to enhance PtdOH clearance from the PM in the present study. Although it is quite likely that other proteins will be identified that contribute to the transfer and metabolic flux of these important lipids, Nir2 appears to be a key molecule when it comes to control the flux and recycling of these lipids during agonist-induced PLC activation.

In summary, the present study uncovers a unique feature of the Nir2 protein; namely, its function as a PtdOH transfer device. The agonist-induced localization of Nir2 to ER-PM junctions, together with its role in feeding the ER-localized CDS enzyme with PtdOH, establishes Nir2 as a central organizer of the phosphoinositide cycle in this unique functional compartment. The exchange function of Nir2 is a perfect design to match PtdIns synthesis and transfer with PtdIns(4,5)P2 utilization so that cells can maintain the level of PtdIns(4,5)P2 during prolonged PLC activation. These results can also explain the functional and metabolic isolation of the agonist-controlled PtdIns cycle from the rest of the ER-associated lipid metabolic pool.

EXPERIMENTAL PROCEDURES

Materials

Angiotensin II (human octapeptide) DG kinase inhibitor I (R59022) were from Bachem and EMD Millipore, respectively. Saponin and U73122 were from Sigma-Aldrich. All other chemicals were of the highest analytical grade. Antibodies against PITPα and PITPβ were generous gifts from Dr. Shamshad Cockcroft (University College London). The Nir2 antibody was raised against the peptide CRSRGPSQAEREGPG in rabbits and affinity purified (New England Peptide). Nir2 antibodies were also generously provided by Dr. Tiansen Li (National Eye Institute, NIH).

DNA Constructs

The human Nir2 (accession number BC022230, obtained as a full-length clone [4396491] from Open Biosystems) was fused to pEGFP-C1 using PCR amplification with the primer pairs listed in Supplemental Experimental Procedures. All point mutations for GFP-Nir2 were generated by the QuikChange mutagenesis kit (Agilent Technologies) (primers listed in Supplemental Experimental Procedures). For rescue experiments, GFP-Nir2 or Nir2 was subcloned to pcDNA3.1 vectors with CMV promoters of different strengths kindly provided by Dr. Wes Sundquist (Department of Biochemistry, University of Utah) (Morita et al., 2012). For this, GFP-Nir2 was amplified with different primer pairs (see Supplemental Experimental Procedures) and digested using SpeI and SalI. The digested fragment was then ligated to individual pcDNA3.1 vectors cut by XbaI and XhoI. The yeast Spo20 PtdOH-binding domain (Spo2051–90) provided in a monomeric dsRed vector was a generous gift from Dr. Nicolas Vitale (University of Strasbourg). An NES (ALQKKLEELELDE from MAPKK) was added to the N terminus of dsRed, but this construct was not fully excluded from the nucleus. Therefore, there were two more versions of the construct created (one having two NES signals at the N terminus, the other having one NES signal at the N terminus) and another between dsRed and the Spo20 sequence. Both of these constructs were equally excluded from the nuclei and behaved similarly as PtdOH sensors. In subsequent experiments using U73122, we found that the sensor was recruited to the PM by this reagent. To generate a version without this undesirable feature, we mutated both cysteine residues (Cys54 and Cys82) to serine. This made the probe compatible with the PLC inhibitor, U73122. For FRET measurements of DG, the previously described DG sensor (NES-GFP-PKDC1ab [W166A]; (Kim et al., 2011) was generated using mTq2 (a kind gift from Dr. Theodorus Gadella, University of Amsterdam) and Venus in place of GFP. For PtdOH FRET studies, the double-cysteine mutant Spo20-based sensor was used with an additional NES signal (ALQKKLEELELDEQ) placed between the fluorophore (mTq or Venus) and the Cys mutated Spo20 peptide. The GFP-tagged Tubby domain was described previously (Szentpetery et al., 2009). GFP-VAP-B and HA-PLD2 were generous gifts from Dr. Chris Stephan (Cornell University) and Dr. John H. Brumell (Cell Biology Program, Hospital for Sick Children), respectively.

RNAi

The target sequences for various Nir2 siRNA were as follows. #1: 5′-AAGGTCTCTGGCTTCTTCCTC-3′, #2: 5′-CCAGTGTGTATAAATCCATGA-3′, #3: 5′-CGGCTTTATCACTCGGTCA-3′, #4: 5′-TGAGAACAGCTCCGAGGAA-3′, and #5: 5′-CAGCCTGGACCTGGGTTATTT-3′. The target sequences of PITPα and PITPβ were 5′-AAGGATGGAAGAAGAGACGAA-3′ and 5′-CGGGAAGATGGTGCTGATCAA-3′, respectively. Oligonucleotide duplexes were obtained from QIAGEN.

Transfection of Cells for Microscopy or RNAi

COS-7 cells or HEK293-AT1 cells (a HEK293 cell line stably expressing the rat AT1a angiotensin receptor) were plated onto 25-mm-diameter poly-L-Lysine-coated circular glass coverslips in six-well plates (300,000 cells/well), and plasmid DNAs (0.5-1 μg/well) were transfected using the Lipofectamine2000 reagent (Invitrogen) and OPTI-MEM (Invitrogen) following the manufacturer’s instructions. The target sequences for various Nir2 siRNAs are described in Supplemental Experimental Procedures. Cells were cultured in either six-well dishes (for microscopy) or 12-well plates (for metabolic labeling studies) and treated with 100 nM siRNA using the RNAiMAX reagent (Invitrogen), and they were analyzed after 3 days treatment with the siRNA. For the rescue experiment, 1 day after siRNA treatment, HEK293-AT1 cells were transfected with 0.5 μg of various CMV-promoter-driven Nir2 constructs for 2 days.

Live-Cell Imaging

After 20–24 hr of transfection, cells were washed on the glass coverslips with a modified Krebs-Ringer solution (containing 120 mM NaCl, 4.7 mM KCl, 1.2 mM CaCl2, 0.7 mM MgSO4, 10 mM glucose, and 10 mM Na-HEPES [pH 7.4]), and the coverslip was placed into a metal chamber (Atto, Invitrogen). Cells were examined in inverted microscopes in 1 μl of the Krebs-Ringer buffer. Stimuli were dissolved and added in 200 μl buffer removed from the cells. Confocal images were obtained with a Zeiss LSM 510 or LSM 780 laser confocal microscope (Carl Zeiss MicroImaging) using a 63× 1.4-numerical-aperture plan-apochromatic objective. TIRF analysis was performed in an Olympus dual launch TIRF microscope system equipped with a Photometrics Cascade II camera and a PlanApo 60×/1.45 objective.

FRET Measurements

HEK-AT1 cells grown on poly-Lysine-coated glass coverslips were co-transfected with the mTq- and Venus-tagged versions of the DG or PtdOH sensors (0.4 μg DNA each) for 24 hr. Coverslips were mounted into Atto chambers and examined in an Olympus IX70 inverted microscope equipped with a Lambda DG4 illuminator (Sutter Instruments) and a beam splitter appropriate for CFP-YFP separation (Photometrics) and a MicroMax BFT camera (Photometrics). Data acquisition and processing was done with the Metafluor software of Molecular Devices. Data are presented as 530/475 emission ratios normalized to initial values in each cells. Several cells per dish were analyzed from multiple dishes per experiments, which were repeated three independent times with separate knockdown.

Measurements of Receptor-Stimulated DG or PtdOH Kinetics using DG or PtdOH Sensor

HEK293-AT1 cells were transfected with the NES-GFP-PKDC1ab (W166A) or NES-dsRed-NES-Spo20m construct. After 24 hr, cells were imaged in a Zeiss LSM 780 confocal microscope at room temperature. After stimulation with 100 nM AngII, translocation of the fluorescent probe was monitored in time-lapse imaging. The translocation of the construct from the membrane to the cytosol was quantified by measuring cytosolic fluorescent intensity in regions of interest outside the nucleus and plotted against time using the Zeiss image-processing software. Intensity curves were normalized to pre-stimulatory values, and they were averaged from the indicated number of cells in recordings obtained in several dishes in multiple independent experiments.

Analysis of myo-[3H]Inositol-, [32P]Phosphate-, or [3H]cytidine-Labeled Lipids

HEK293-AT1 cells plated on 12-well plates were labeled with myo-[3H]inositol (20 μCi/ml) in inositol-free DMEM supplemented with 2% dialyzed fetal bovine serum for 2 or 24 hr. When added, AngII (10−7 M) was present for the entire short-term labeling period. For 32P-phosphate experiments, 2 μCi/ml o-[32P] phosphate was used for 1-hr labeling in phosphate-free DMEM and AngII was added for 10 min. For [3H]cytidine labeling, cells were incubated in inositol-free medium overnight and labeled with 10 μCi/ml [3H]cytidine in a modified Krebs-Ringer solution for a total of 90 min. Li+ (10 mM) was added at 30 min and AngII at 60 min. In some cases, longer incubation with AngII (60 min) was used with accordingly extended labeling. The labeling was terminated by the addition of ice-cold perchloric acid (5% final concentration), and cells were kept on ice for 30 min. After scraping and freeze-thawing, cells were centrifuged and the cell pellet was processed to extract the lipids by an acidic chloroform/methanol extraction followed by thin layer chromatography (TLC) essentially as described previously (Nakanishi et al., 1995). For some experiments using [3H]cytidine labeling, TLC was not performed but the entire extract was evaporated and analyzed by scintillation counting, since a negligible amount of radioactivity was associated with lipids other than CDP-DG. 32P-labeled samples were analyzed with phosphoimager, while [3H]-labeled samples were sprayed with autoradiography Enhancer (PerkinElmer) and exposed to X-ray films with multiple exposures In some cases, 3H-labeled lipids were also scraped and counted in scintillation counters for comparison.

Supplementary Material

1
2

Highlights.

  • Nir2 protein supports phosphatidic acid transfer from the plasma membrane (PM) to the ER

  • Nir2 translocates to ER-PM contact sites during PLC activation

  • Nir2 delivers phosphatidic acid to the ER for PtdIns synthesis

  • Accelerated phosphoinositide turnover is confined to ER-PM contact zones

ACKNOWLEDGMENTS

We would like to thank Drs. John H. Brumell, Theodorus Gadella, Chris Stephan, and Nicolas Vitale for DNA constructs and Drs. Shamsad Cockcroft and Tiansen Li for PITP and Nir2 antibodies. We are also grateful for Drs. Gerald Hammond and Marko Jovic for fruitful discussions and reading the manuscript. Confocal imaging was performed at the Microscopy and Imaging Core of the National Institute of Child Health and Human Development, NIH with the kind assistance of Drs. Vincent Schram and James T. Russell. This research was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the NIH. E.W. was supported by a scholarship from the Hungarian American Enterprise Scholarship Fund (HAESF).

Footnotes

SUPPLEMENTAL INFORMATION

Supplemental Information includes Supplemental Experimental Procedures and four figures and can be found with this article online at http://dx.doi.org/10.1016/j.devcel.2015.04.028.

AUTHOR CONTRIBUTIONS

Y.J.K. designed, performed, and analyzed experiments and wrote parts of the manuscript. M.L.G.-H. generated DNA constructs and performed and analyzed experiments. E.W. generated DNA constructs and performed and analyzed FRET experiments. T.B. designed experiments, performed confocal microscopy, analyzed data, and wrote the manuscript.

REFERENCES

  1. Agranoff BW, Bradley RM, Brady RO. The enzymatic synthesis of inositol phosphatide. J. Biol. Chem. 1958;233:1077–1083. [PubMed] [Google Scholar]
  2. Allison JH, Stewart MA. Reduced brain inositol in lithium-treated rats. Nat. New Biol. 1971;233:267–268. doi: 10.1038/newbio233267a0. [DOI] [PubMed] [Google Scholar]
  3. Amarilio R, Ramachandran S, Sabanay H, Lev S. Differential regulation of endoplasmic reticulum structure through VAP-Nir protein interaction. J. Biol. Chem. 2005;280:5934–5944. doi: 10.1074/jbc.M409566200. [DOI] [PubMed] [Google Scholar]
  4. Balla T. Phosphoinositides: tiny lipids with giant impact on cell regulation. Physiol. Rev. 2013;93:1019–1137. doi: 10.1152/physrev.00028.2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bankaitis VA, Mousley CJ, Schaaf G. The Sec14 superfamily and mechanisms for crosstalk between lipid metabolism and lipid signaling. Trends Biochem. Sci. 2010;35:150–160. doi: 10.1016/j.tibs.2009.10.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Berridge MJ. Inositol trisphosphate and diacylglycerol as second messengers. Biochem. J. 1984;220:345–360. doi: 10.1042/bj2200345. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Berridge MJ, Downes CP, Hanley MR. Lithium amplifies agonist-dependent phosphatidylinositol responses in brain and salivary glands. Biochem. J. 1982;206:587–595. doi: 10.1042/bj2060587. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chang CL, Hsieh TS, Yang TT, Rothberg KG, Azizoglu DB, Volk E, Liao JC, Liou J. Feedback regulation of receptor-induced Ca2+ signaling mediated by E-Syt1 and Nir2 at endoplasmic reticulum-plasma membrane junctions. Cell Rep. 2013;5:813–825. doi: 10.1016/j.celrep.2013.09.038. [DOI] [PubMed] [Google Scholar]
  9. Cockcroft S, Garner K. Function of the phosphatidylinositol transfer protein gene family: is phosphatidylinositol transfer the mechanism of action? Crit. Rev. Biochem. Mol. Biol. 2011;46:89–117. doi: 10.3109/10409238.2010.538664. [DOI] [PubMed] [Google Scholar]
  10. D’Angelo G, Polishchuk E, Di Tullio G, Santoro M, Di Campli A, Godi A, West G, Bielawski J, Chuang CC, van der Spoel AC, et al. Glycosphingolipid synthesis requires FAPP2 transfer of glucosylceramide. Nature. 2007;449:62–67. doi: 10.1038/nature06097. [DOI] [PubMed] [Google Scholar]
  11. de Saint-Jean M, Delfosse V, Douguet D, Chicanne G, Payrastre B, Bourguet W, Antonny B, Drin G. Osh4p exchanges sterols for phosphatidylinositol 4-phosphate between lipid bilayers. J. Cell Biol. 2011;195:965–978. doi: 10.1083/jcb.201104062. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Garner K, Hunt AN, Koster G, Somerharju P, Groves E, Li M, Raghu P, Holic R, Cockcroft S. Phosphatidylinositol transfer protein, cytoplasmic 1 (PITPNC1) binds and transfers phosphatidic acid. J. Biol. Chem. 2012;287:32263–32276. doi: 10.1074/jbc.M112.375840. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Hanada K, Kumagai K, Yasuda S, Miura Y, Kawano M, Fukasawa M, Nishijima M. Molecular machinery for non-vesicular trafficking of ceramide. Nature. 2003;426:803–809. doi: 10.1038/nature02188. [DOI] [PubMed] [Google Scholar]
  14. Helms JB, de Vries KJ, Wirtz KW. Synthesis of phosphatidylinositol 4,5-bisphosphate in the endoplasmic reticulum of Chinese hamster ovary cells. J. Biol. Chem. 1991;266:21368–21374. [PubMed] [Google Scholar]
  15. Hokin LE, Hokin MR. Effects of acetylcholine on the turnover of phosphoryl units in individual phospholipids of pancreas slices and brain cortex slices. Biochim. Biophys. Acta. 1955;18:102–110. doi: 10.1016/0006-3002(55)90013-5. [DOI] [PubMed] [Google Scholar]
  16. Hokin LE, Hokin MR. The incorporation of 32P from[32P]ATP into polyphosphoinositides and phosphatidic acid in erythrocyte membranes. Biochim. Biophys. Acta. 1964;84:563–575. doi: 10.1016/0926-6542(64)90126-x. [DOI] [PubMed] [Google Scholar]
  17. Ile KE, Schaaf G, Bankaitis VA. Phosphatidylinositol transfer proteins and cellular nanoreactors for lipid signaling. Nat. Chem. Biol. 2006;2:576–583. doi: 10.1038/nchembio835. [DOI] [PubMed] [Google Scholar]
  18. Inoue H, Baba T, Sato S, Ohtsuki R, Takemori A, Watanabe T, Tagaya M, Tani K. Roles of SAM and DDHD domains in mammalian intracellular phospholipase A1 KIAA0725p. Biochim. Biophys. Acta. 2012;1823:930–939. doi: 10.1016/j.bbamcr.2012.02.002. [DOI] [PubMed] [Google Scholar]
  19. Kassas N, Tryoen-Tóth P, Corrotte M, Thahouly T, Bader MF, Grant NJ, Vitale N. Genetically encoded probes for phosphatidic acid. Methods Cell Biol. 2012;108:445–459. doi: 10.1016/B978-0-12-386487-1.00020-1. [DOI] [PubMed] [Google Scholar]
  20. Kim YJ, Guzman-Hernandez ML, Balla T. A highly dynamic ER-derived phosphatidylinositol-synthesizing organelle supplies phosphoinositides to cellular membranes. Dev. Cell. 2011;21:813–824. doi: 10.1016/j.devcel.2011.09.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kim S, Kedan A, Marom M, Gavert N, Keinan O, Selitrennik M, Laufman O, Lev S. The phosphatidylinositol-transfer protein Nir2 binds phosphatidic acid and positively regulates phosphoinositide signalling. EMBO Rep. 2013a;14:891–899. doi: 10.1038/embor.2013.113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kim YJ, Hernandez ML, Balla T. Inositol lipid regulation of lipid transfer in specialized membrane domains. Trends Cell Biol. 2013b;23:270–278. doi: 10.1016/j.tcb.2013.01.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lev S. Nonvesicular lipid transfer from the endoplasmic reticulum. Cold Spring Harb. Perspect. Biol. 2012;4:a013300. doi: 10.1101/cshperspect.a013300. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Liou J, Kim ML, Heo WD, Jones JT, Myers JW, Ferrell JE, Jr., Meyer T. STIM is a Ca2+ sensor essential for Ca2+-store-depletion-triggered Ca2+ influx. Curr. Biol. 2005;15:1235–1241. doi: 10.1016/j.cub.2005.05.055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lu C, Peng YW, Shang J, Pawlyk BS, Yu F, Li T. The mammalian retinal degeneration B2 gene is not required for photoreceptor function and survival. Neuroscience. 2001;107:35–41. doi: 10.1016/s0306-4522(01)00337-2. [DOI] [PubMed] [Google Scholar]
  26. McPhee F, Lowe G, Vaziri C, Downes CP. Phosphatidylinositol synthase and phosphatidylinositol/inositol exchange reactions in turkey erythrocyte membranes. Biochem. J. 1991;275:187–192. doi: 10.1042/bj2750187. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Mesmin B, Bigay J, Moser von Filseck J, Lacas-Gervais S, Drin G, Antonny B. A four-step cycle driven by PI(4)P hydrolysis directs sterol/PI(4)P exchange by the ER-Golgi tether OSBP. Cell. 2013;155:830–843. doi: 10.1016/j.cell.2013.09.056. [DOI] [PubMed] [Google Scholar]
  28. Milligan SC, Alb JG, Jr., Elagina RB, Bankaitis VA, Hyde DR. The phosphatidylinositol transfer protein domain of Drosophila retinal degeneration B protein is essential for photoreceptor cell survival and recovery from light stimulation. J. Cell Biol. 1997;139:351–363. doi: 10.1083/jcb.139.2.351. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Morita E, Arii J, Christensen D, Votteler J, Sundquist WI. Attenuated protein expression vectors for use in siRNA rescue experiments. Biotechniques. 2012:1–5. doi: 10.2144/000113909. 0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mousley CJ, Davison JM, Bankaitis VA. Sec14 like PITPs couple lipid metabolism with phosphoinositide synthesis to regulate Golgi functionality. Subcell. Biochem. 2012;59:271–287. doi: 10.1007/978-94-007-3015-1_9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Nakanishi S, Catt KJ, Balla T. A wortmannin-sensitive phosphatidylinositol 4-kinase that regulates hormone-sensitive pools of inositolphospholipids. Proc. Natl. Acad. Sci. USA. 1995;92:5317–5321. doi: 10.1073/pnas.92.12.5317. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Peretti D, Dahan N, Shimoni E, Hirschberg K, Lev S. Coordinated lipid transfer between the endoplasmic reticulum and the Golgi complex requires the VAP proteins and is essential for Golgi-mediated transport. Mol. Biol. Cell. 2008;19:3871–3884. doi: 10.1091/mbc.E08-05-0498. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Prinz WA. Lipid trafficking sans vesicles: where, why, how? Cell. 2010;143:870–874. doi: 10.1016/j.cell.2010.11.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Stubbs EB, Jr., Agranoff BW. Lithium enhances muscarinic receptor-stimulated CDP-diacylglycerol formation in inositol-depleted SK-N-SH neuroblastoma cells. J. Neurochem. 1993;60:1292–1299. doi: 10.1111/j.1471-4159.1993.tb03289.x. [DOI] [PubMed] [Google Scholar]
  35. Suzuki E, Hirosawa K. Immunolocalization of a Drosophila phosphatidylinositol transfer protein (rdgB) in normal and rdgA mutant photoreceptor cells with special reference to the subrhabdomeric cisternae. J. Electron Microsc. (Tokyo) 1994;43:183–189. [PubMed] [Google Scholar]
  36. Szentpetery Z, Balla A, Kim YJ, Lemmon MA, Balla T. Live cell imaging with protein domains capable of recognizing phosphatidylinositol 4,5-bisphosphate; a comparative study. BMC Cell Biol. 2009;10:67. doi: 10.1186/1471-2121-10-67. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Trivedi D, Padinjat R. RdgB proteins: functions in lipid homeostasis and signal transduction. Biochim. Biophys. Acta. 2007;1771:692–699. doi: 10.1016/j.bbalip.2007.04.014. [DOI] [PubMed] [Google Scholar]
  38. van Der Wal J, Habets R, Varnai P, Balla T, Jalink K. Monitoring agonist-induced phospholipase C activation in live cells by fluorescence resonance energy transfer. J. Biol. Chem. 2001;276:15337–15344. doi: 10.1074/jbc.M007194200. [DOI] [PubMed] [Google Scholar]
  39. Vaziri C, Downes CP, Macfarlane SC. Direct labelling of hormone-sensitive phosphoinositides by a plasma-membrane-associated PtdIns synthase in turkey erythrocytes. Biochem. J. 1993;294:793–799. doi: 10.1042/bj2940793. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Vihtelic TS, Hyde DR, O’Tousa JE. Isolation and characterization of the Drosophila retinal degeneration B (rdgB) gene. Genetics. 1991;127:761–768. doi: 10.1093/genetics/127.4.761. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Vihtelic TS, Goebl M, Milligan S, O’Tousa JE, Hyde DR. Localization of Drosophila retinal degeneration B, a membrane-associated phosphatidylinositol transfer protein. J. Cell Biol. 1993;122:1013–1022. doi: 10.1083/jcb.122.5.1013. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS695371-supplement-1.doc (23.5KB, doc)
2
NIHMS695371-supplement-2.pdf (750KB, pdf)

RESOURCES