ABSTRACT
During adverse conditions, mammalian cells regulate protein production by sequestering the translation machinery in membraneless organelles (i.e. stress granules) whose formation is carefully regulated. Here, we show a direct connection between G protein signaling and stress granule formation through phospholipase Cβ1 (PLCβ1). In cells, PLCβ1, the most prominent isoform of PLCβ in neuronal cells, localizes to both the cytoplasm and plasma membrane. Here, we show that a major population of cytosolic PLCβ1 binds to stress granule proteins, such as PABPC1, eIF5A and Ago2. PLCβ1 is activated by Gαq in response to hormones and neurotransmitters and we find that activation of Gαq shifts the cytosolic population of PLCβ1 to the plasma membrane, releasing stress granule proteins. This release is accompanied by the formation of intracellular particles containing Ago2 aggregates, an increase in the size and number of particles containing PABPC1 and Ago2, and a shift of cytosolic RNAs to larger sizes consistent with cessation of transcription. These particles are seen when the cytosolic level of PLCβ1 is lowered by siRNA, osmotic stress or Gαq stimulation by carbachol. These stresses, in addition to cold, heat, oxidative and arsenite stress produces particles that appear to have different molecular compositions. Our results fit a simple thermodynamic model in which cytosolic PLCβ1 solubilizes stress granule proteins and its movement to Gαq upon stimulation releases these particles to allow the formation of stress granules. Taken together, our studies show a link between Gαq-coupled signals and transcription through stress granule formation.
INTRODUCTION
When cells are subjected to environmental stress, they halt the production of many housekeeping proteins to preserve resources for the synthesis of proteins that will help the cell alleviate the particular stress. These stalled transcription complexes, called stress granules, are thought to act as junctures that protect non-translated mRNAs from degradation until the stress is removed while allowing the synthesis of other proteins (for review see (1, 2)). Stress granules are distinct from processing bodies or P-bodies that store and process mRNA, although they have also been observed under non-stress conditions. Depending on the cell conditions, the mRNA held in these stalled complexes may be degraded, translated or stored until needed. Additionally, studies in yeast subjected to hypo-osmotic stress found that P-bodies and stress granules may form hybrid structures (3). Physically, stress granules are phase-separated domains composed of non-translating mRNAs, translation initiation complexes, poly (A)-binding protein, and many additional mRNA-related proteins (4). They consist of a packed core with loosely associated peripheral proteins (5). Stress granules appear when cells are subjected to environmental conditions such as cold or heat shock, exposure to toxic molecules, oxidative stress, hypo- or hyper-osmolarity, UV irradiation and nutrient deprivation. The molecular mechanisms that transmit these different stresses into the cell interior remain largely unresolved.
Although stress granules appear in many cell systems, we focus here on those that form in mammalian cells. Stress granules have been implicated in the pathogenesis of various diseases such as cancer, neurodegeneration and viral infections (1, 6). Many stress granule proteins contain disordered domains and these regions play important roles in the liquid-like nature of stress granules. Neuronal cells, in particular, contain many proteins with disordered domains and so it is not surprising that some neurological diseases (e.g. ALS) have been attributed to abnormal stability of stress granules (see (7)). Thus, it is important that cells have mechanisms to prevent premature formation of stress granules, and to insure their reversible assembly and disassembly.
While stress granules primarily contain proteins associated with translation, it is notable that argonaute 2 (Ago2) can be found in these domains (see (8)). Ago2 is the main nuclease component of the RNA-induced silencing complex (9). Ago2 binds small, silencing RNAs in their double-stranded form, and holds the guide strand after the passenger strand is degraded to allow hybridization with a target mRNA. If pairing between the passenger strand and the mRNA is perfect, as is the case in exogenous siRNAs, then Ago2 will undergo conformational changes that result in mRNA degradation. Alternately, if pairing is imperfect, as is frequently the case for endogenous microRNAs, the conformational changes that allow Ago2 nuclease activity does not occur which results in a stalled complex. Thus, the formation and stability of these stalled complexes and their incorporation into stress granules will alter the local protein population, and ultimately functional properties of the cell.
The mechanisms through which environmental changes are communicated into the cell to promote stress granule formation are unclear and likely to differ with different types of stress. Here, we show that extracellular signals impact stress granule formation via G proteins to stall translation. Signaling through G proteins is initiated when external ligands bind to their target G protein coupled receptor that activates intracellular Gα subunits. The Gαq family of G proteins is activated by agents such as acetylcholine, dopamine, bradykinin, serotonin, histamine, etc. (10, 11). Activated Gαq in turn activates phospholipase Cβ (PLCβ) which catalyzes the hydrolysis of the signaling lipid phosphatidylinositol 4, 5 bisphosphate resulting in an increase in intracellular calcium. There are four known family members of PLCβ that differ in their tissue distribution and their ability to be activated by G proteins (11), and these studies focus on PLCβ1 which is the most highly activated by Gαq and is prominent in neuronal tissue. While the major population of PLCβ1 lies on the plasma membrane where it binds Gαq and accesses its substrate, PLCβ is also found in the cytosol in every cell type examined and under different conditions (12, 13).
Several years ago we found that a cytoplasmic PLCβ1 population binds to C3PO, the promoter of RNA-induced silencing, and this binding can reverse RNA-induced silencing of specific genes (13, 14). Reversal of silencing was independent of PLCβ1’s catalytic activity. Subsequent studies showed that the association of PLCβ and C3PO is critical for PC12 differentiation (15, 16), but little or no association is seen in non-differentiating cells leading to the question of whether cytosolic PLCβ1 has other binding partners. Recently, we found that PLCβ1 will bind to and inhibit a neuronal-specific enzyme required for proliferation, CDK16 (17, 18), and this association allows cells to cease proliferation and transition into a differentiated state (16). Again, this association is confined to a specific event that drives neuronal cells out of stemness, and suggests that under non-proliferating, non-differentiating conditions cytosolic PLCβ serves some other function. In this study, we show that a major population of cytosolic PLCβ is bound to stress granules proteins, and that this binding prevents premature stress granule formation. Removal of PLCβ1 from the cytoplasm by stress or by Gαq stimulation promotes particle assembly. The interaction between PLCβ1 and stress granule proteins shows a novel feedback mechanism between the external environment and the protein translation machinery.
RESULTS
PLCβ1 binds to stress granule-associated proteins
We began this work by carrying out experiments to determine novel binding partners to cytosolic PLCβ1 in PC12 cells under non-differentiating conditions. Our approach was to isolate the cytosolic fractions of unsynchronized, undifferentiated PC12 cells, and pull down proteins bound to PLCβ1 using a monoclonal antibody. We collected the PLCβ1-bound proteins and identified them by mass spectrometry. Unexpectedly, we found that ∼30% of the total proteins associated with cytosolic PLCβ1 are markers for stress granules (19), (supplemental). The most prevalent ones are listed in Fig.1A. In contrast, control studies using cytosolic fractions of cells with reduced PLCβ1 levels only and identified non-specific proteins such as tubulin, actin and mitochondrial proteins (supplemental).
Proteins associated with PLCβ1 in cytosolic fractions of PC12 cells were pulled down using a monoclonal antibody in A-unsynchronized cells primarily in the G1 phase, or B-PC12 cells arrested in the G2/M phase. Levels of proteins are expressed as the quantitative normalized iBAQ for each sample. Proteins in red font are phosphorylated. All proteins identified in panel (A) were also found in an identical study in which proteins associated with Ago2 were identified except for eIF5B, which is presented as a gray bar.
To see whether the binding of PLCβ1 to stress granule proteins occurs only in unsynchronized cells where ∼90% in the G1 phase (16), we repeated these studies in PC12 cells arrested in the G2/M phase, and these results are presented in Fig.1B. Again, we find that 32% of the proteins bound to PLCβ1 are markers for stress granules, with the most prominent being eukaryotic initiation factor 5A (eIF5A) and polyadenylate binding protein C1 (PABPC1) (Fig.1B). Additionally, other stress-granule and translation proteins appear at high or moderate levels in both groups such as FXR1/2 and other eukaryotic initiation factors. It is also notable that Ago2, which is associated with both RNA-induced silencing complexes and stress granules (8), appears in these cells.
While the proteomics studies described above are only indicative of potential protein partners of PLCβ1 since they are identified under non-physiological conditions. Therefore, we verified the binding of PLCβ1 to stress granule proteins by several methods. First, we again carried out pull-down studies using a monoclonal antibody and monitored the association of two stress granule proteins by western blotting. The first, PABPC1, is an established marker for stress granules (19). The second, Ago2 moves into granules under stress conditions (8). Using unsynchronized, undifferentiated PC12 cells, we verified that PABPC1 and Ago2 bind to PLCβ1 (Fig. 2A). We then determined whether the levels of these proteins changed when two of PLCβ1’s established binding partners, Gαq and C3PO, are over-expressed (Fig. 2B). We find that the level of Ago2 is reduced when the levels of either of these partners is raised, suggesting that Ago2 binds to similar regions of PLCβ1. However, the amount of PABPC1 pulled down with PLCβ1 does not significantly change with over-expression of either Gαq or C3PO (p=0.81 and p=0.54, respectively) but is lowered with Ago2 over-expression. The simplest interpretation of these data is that interactions between PLCβ1 and PABPC1 differ from those of Ago2, and that the reduction in PABPC1 levels with Ago2 over-expression is simply due to a redistribution of the PLCβ1 pool (see Discussion). Nevertheless, our studies support the idea that cytosolic PLCβ1 associates with stress granules proteins.
Using a monoclonal anti-PLCβ1 to isolate PLCβ1-associated proteins from the cytsolic fractions of undifferentiated PC12 cells, we monitored two stress-granule associated proteins, PABPC1 and Ago2. The figure presents a compilation of band intensities in mock transfected cells (control) and in cells over-expressing Ago2, C3PO or constituatively active Gαq) where * denotes p<0.001. Differences between C3PO and Gaq pulled down with PLCβ1/PABPC1 were not significant, p=0.81 and 0.54, respectively. Data are from n=8 independent experiments and S.D. is shown. We note that because TRAX and translin expression are linked (26), we over-expressed TRAX to increase levels of C3PO.
PLCβ1 and Ago2 associated in living cells
The above studies suggest that PLCβ1 and Ago2 interactions may be modulated by G protein stimulation and cellular events associated with C3PO. Keeping in mind that C3PO promotes RNA-induced silencing, we set-out to characterize the factors that regulate PLCβ-Ago2 association. We first isolated the cytosolic fractions of unsynchronized, undifferentiated PC12 cells, pulled-down proteins associated with Ago2 and identified them by mass spectrometry. We find that PLCβ is included in the set of data. Interestingly, all of the proteins listed in Fig. 1A were found in both PLCβ1 and Ago2 pull-downs (see supplemental material).
We measured the association between PLCβ1 and Ago2 in living PC12 cells by Förster resonance energy transfer (FRET) as monitored by fluorescence lifetime imaging microscopy (FLIM). In this method, the FRET efficiency is determined by the reduction in the time that the donor spends in the excited state (i.e., the fluorescence lifetime) before transferring energy to an acceptor fluorophore (see (20)). If we excite the donor with light that has modulated intensity, the lifetime can be determined by the reduction in modulated intensity (M) as well as the shift in phase (φ) of the emitted light. If FRET occurs when the donor is in the excited state, the fluorescence lifetime will be reduced as indicated by a reduced change in modulation and phase shift. The amount of FRET can then be directly determined from the raw data by plotting the lifetimes in each pixel in the image on a phasor plot (i.e. S versus G where S = M*sin (φ) and G = M*cos (φ) (see (21)). In these plots, the lifetimes from each pixel in a FLIM image will fall on the phasor arc for a single population. However, when two or more lifetimes are present, the points will be a linear combination of the fractions with the points inside the phase arc that move towards the right due to shortened lifetimes (i.e. FRET). We note that phasor representation is simply the Fourier transform of the lifetime decay curves but readily displays lifetimes directly from raw data without the need for model-dependent fitting of the lifetimes or other corrections.
In Fig. 3 we show the phasor plot and the corresponding image of a PC12 cell expressing eGFP-PLCβ1 where the phase and modulation lifetime of each pixel in the image is presented. As expected for a single lifetime species, all points fall on the phasor arc (Fig. 3A). When we repeat this study in cells expressing both eGFP-PLCβ1 and mCherry-Ago2 where mCherry is a FRET acceptor, the average donor lifetime drops from 2.5 to 1.7ns, and the points move inside the phasor (Fig. 3B). This reduction in lifetime and movement of the points into the phasor show the occurrence of FRET between the probes. Because the amount of FRET depends on the distance between the fluorophores to the sixth power, and since the distance at which 50% transfers for the eGFP / mCherry pair is ∼30Å (22), our results indicate direct interaction between PLCβ1 and Ago2 in the cytosol.
Examples of phasor plots in which PC12 cells were transfected with (A) eGFP-PLCβ1 or (B) eGFP-PLCβ1 and mCherry-Ago2 where the raw lifetimes are plotted as S versus G (see text). Each point in the phasor plots corresponds to the lifetime from the eGFP-PLCβ1 emission measured in each pixel from the corresponding cell image shown in the graph. C – a phasor diagram where the non-FRET and FRET points are selected and the pixels are shown in the cell images. We note that no significant FRET was detected in control studies of eGFP-Ago2 and mCherry-C3PO (supplemental).
We can select points in the phasor plots and visualize their localization in the cell image. We find that the points corresponding to FRET, and thus eGFP-PLCβ1 / mCherry-Ago2 complexes, are localized in the cytoplasm (Fig. 3C). In contrast, points corresponding eGFP-PLCβ1 alone are found both on the plasma membrane and the cytoplasm.
eIF5A binds to PLCβ1 competitively with C3PO and Gαq
To remain cytosolic, stress granule proteins must bind to PLCβ1 in a manner competitive with Gαq or it would localize to the plasma membrane. We previously have shown that PLCβ1 binds to C3PO in the same C-terminal region as Gαq (14) and that competition between C3PO/Gαq regulates PLCβ1’s ability to generate calcium signals through Gαq activation, or its ability to reverse siRNA, respectively (23). With this in mind, we searched the proteins identified in the mass spectrometry for stress granule proteins that could draw PLCβ1 away from Gαq. These data indicate a potentially strong association between PLCβ1 and eIF5A (Fig.1B). Additionally, eIF5A, which is a GTP-activating protein (24), has homologous regions to the GTPase region of Gαq (see Discussion) and so we chose it for further testing.
In an initial study, we purified PLCβ1 and covalently labeled it with a FRET donor, Alexa488. We then purified eIF5A, labeled it with a FRET acceptor, Alexa594, and measured the association between the two proteins in solution by fluorescence titrations similar to previous studies (25). We find that the two proteins bind with strong affinity (Kd = 27 + 5nM). However, when we formed the Alexa488-PLCβ1-C3PO complex and titrated Alexa594-eIF5A into the solution, we could not detect FRET (supplemental). This result suggests that eIF5A binds to the same region as C3PO, and similarly, to Gαq (see (14)).
To determine whether eIF5A competes with C3PO for PLCβ1 in cells, we transfected PC12 cells with eGFP-PLCβ1 and mCherry-TRAX, to produce fluorescent C3PO (see (26)). Cellular association between these proteins is easily detected by FLIM/FRET (Fig. 4A-B). We then microinjected purified eIF5A to increase its intracellular amount by ∼10nM and find that FRET is completely eliminated (Fig. 4C). This results shows that eIF5A displaces C3PO from PLCβ1 and suggests that both proteins bind to a similar region in PLCβ1’s C-terminus.
A - Example of a phasor plot and the corresponding image from a FLIM measurement of eGFP-PLCβ1 expressed in a PC12 cells. The heat map indicates eGFP signal intensity. B - A similar study as in (A) except the PC12 cell is cotransfeted with eGFP-PLCβ1 and mCherry-C3PO (TRAX). Note the movement of pixels into the interior of the phasor arc due to FRET. Purple dots indicate pixels represented in the phasor plot. C - The same study as in (B) except this cell was microinjected with purifed, unlabeled eIF5A. Note the movement of points back to the phasor arc showing a loss of FRET due to displacement of C3PO from PLCβ1. D - Compilation of eGFP-PLCβ1 lifetime results for 2 independent studies where n=10 cells for each study and where a negative control using free mCherry is included. Comparison of data before and after eIF5A microinjection is statistically significant p=<0.001 t-10.665 and f value for Anova test is 96.606.
We confirmed the idea that eIF5A binds to the C-terminal region of PLCβ1 using purified proteins in solution. In these studies, we formed the eIF5A-PLCβ1 complex, chemically cross-linked proteins, digested the samples, separated the fragments by electrophoresis and sequenced the peptides by mass spectrometry (supplemental). We find several interaction sites between the proteins but one of the most prominent is between residues 1085-1095 of PLCβ1and 97-103 of eIF5A which are expected to be close to the Gαq activation region. Taken together, these studies suggest that eIF5A competes with both Gαq and C3PO for PLCβ1 binding and directs PLCβ1 to complexes containing stress granule proteins.
Effect of osmotic stress on PLCβ1 isoforms
To determine whether PLCβ1 can impact stress granule formation, we started with mild, hypo-osmotic stress (300 to 150 mOsm where we have found has reversible effects on the Gαq/PLCβ signaling pathway in muscle cells (27, 28). We first determined whether hypo-osmotic stress affected the association between PLCβ1 and stress granule proteins within 5 minutes before adaptive changes in the cell occur.
PLCβ1 has two major subtypes (1a and 1b) where the 1a form is the best characterized, and most prevalent subtype, and has additional residues in its C terminus (1142–1216) (29). Although both forms are similarly activated by Gαq, some studies found differences in localization although these appear to be cell type specific (30–32). Surprisingly, we find that 5 minute exposure of PC12 cells to osmotic stress caused a dramatic reduction in PLCβ1a while the amount of PLCβ1b is unchanged (Fig. 5A). Gαq stimulation did not affect the level of either isoform. Tracking total PLCβ1 using an antibody that recognizes both isotypes, we found that the cytosolic population is preferentially reduced (Fig. 5B). Considering the half-life of total PLCβ1 in PC12 cells is 20 min (16), this data in Fig. 5 suggests that osmotic stress enhances PLCβ1a degradation. After 30 min of stress, the level of PLCβ1b increases as does the level of PABPC1, but PLCβ1a remains low as the cells adapt (Fig. 5B). Because the majority of our studies cannot distinguish between the a and b PLCβ1 isoforms, we will refer to total PLCβ1.
A- A western blot of the cytosolic fractions of PLCβ1 at 300 mOsm under control conditions where PLCβ1a is the upper band of the doublet and PLCβ1b is the lower band. While the intensities of these band are unchanged 10 minutes after Gαq stimulation with 5µM carbachol, PLCβ1a is not detected at 150 mOsm for 5 and 30 minutes. PABPC1 and β-tubulin bands are shown for comparison and n=3. B – Results of a study in which eGFP-PLCβ1 was transfected into undifferentiated PC12 cells and changes in cytosolic fluorescence intensity in the a slice in middle of the cell was quantified. Identical behavior was seen over 9 independent experiments, and we note that the plasma membrane population showed similar changes in intensity. C-A study showing the change in calcium release when PC12 cells labeled with Calcium Green are stimulated with 5 µM carbachol under basal conditions (300 mOsm) and hypo-osmotic stress (150 mOsm) for 10 minutes where n=12 and SD is shown, and where that basal points for basal and osmotic stress completely overlap.
We determined the ability of hypo-osmotic stress to affect the ability of Gαq to transduce calcium signals in response to 1µM carbachol. Using a fluorescent calcium indicator, Calcium Green (see methods), we first verified that lowering the osmolarity from 300 to 150 Osm does not affect the intracellular calcium levels. However, osmotic stress completely quenches an increase in calcium signals transduced through Gαq / PLCβ1 (Fig. 5C). This loss is consistent with reduced cellular PLCβ1 levels.
Cytosolic levels of PLCβ1 impact stress granule assembly
It is possible that PLCβ1 binds stress granule associated proteins to prevent premature assembly of stress particles. To test this idea, we followed stress granule formation in PC12 cells under hypo-osmotic stress by counting the number of particles microscopically in undifferentiated, unsynchronized PC12 cells using a 100x objective under normal (300 mOsm) and hypo-osmotic conditions (150 mOsm). We note that this resolution may not capture the formation of small particles (see (5)) and we might be viewing the assembly of primary particles as well as the fusion of small pre-formed ones. We also note that we analyzed particle number and sizes in 1.0 µ slices though several cells as noted in the figures. Therefore, we report particle sizes in area as seen for each slice since converting the particles into three dimensions would result in a loss in resolution and was not necessary for this analysis.
We fixed PC12 cells under normal and hypo-osmotic conditions and stained them with monoclonal antibodies to the stress granule marker, PABPC1. In control cells, PABPC1 antibody staining shows ∼750 particles below 25 µm2 (Fig. 6A). When PLCβ1 is down-regulated, we find a large increase in PABPC1 particles from 25 to 100 µm2 suggesting that PLCβ1 is promoting the formation of larger particles. When we apply osmotic stress we find an increase in the number of particles between 25 and 50 µm2 (Fig. 6B). However, osmotic stress does not change the size or number of particles in cells where PLCβ1a has been down-regulated suggesting that osmotic stress and loss of PLCβ1 are not additive effects. In another series of studies, we stimulated cells with carbachol to activate Gαq (Fig. 6C). We find that stimulation produces a high number of particles up to ∼150 µm2 and down-regulating PLCβ1 does not greatly impact the size or number of PABPC1 particles. Taken together, these studies suggest that loss of PLCβ1 either from down-regulation, osmotic stress or binding to activated Gαq promotes the incorporation of PABPC1 into larger aggregates.
The size and number of particles associated with monoclonal anti-PABPC1 in the cytosol of fixed and immunostained PC12 cells was measured on a 100x objective and analyzed using Image J (see methods). A – Treatment of cells with siRNA(PLCβ1) results in the formation of larger aggregates relative to mock-transfected controls. An enlarged version of the plot is shown directly below to allow better comparison. B- While osmotic stress (150 mOsm, 5 minutes) does not affect the size or number of PABPC1-associated particles, PLCβ1 down-regulation causes a significant increase in particle size and number. C- Stimulation of Gαq by treatment with 5 µM carbachol also impacts particle size. All measurements are an average of 3 independent experiments that sampled 10 cells, where SD shown and where the p values was determined using ANOVA.
We then tested the effect of PLCβ1 on the size and number of particles associated with Ago2 by immunofluorescence. For Ago2, the number of smaller particles substantially increased when PLCβ1 was down-regulated (Fig. 7A). Unlike PABPC1, the size and number of Ago2-associated particles were not affected by osmotic stress, although an increase in the number of small particles with PLCβ1 down-regulation was still seen (Fig. 7B). Additionally, carbachol stimulation of Gαq resulted in an increase in the number of small particles (Fig. 7C).
In similar study in Fig. 6, the size and number of particles associated with anti-Ago2 in the cytosol of PC12 cells was measured on a 100x objective and analyzed using Image J (see methods). A – Treatment of cells with siRNA(PLCβ1) results in the formation of numerous small aggregates relative to mock-transfected controls. B- Subjecting cells to 5 minutes of osmotic stress (150 mOsm) does not affect the size or number of Ago2-associated particles while siRNA(PLCβ1) causes a significant increase in the number of small aggregates. C- Stimulation of Gαq by treatment with 5 µM carbachol results in the formation of a large number of small aggregates. Similar behavior is observed when PLCβ1 is down-regulated. D- Analysis of particles associated with mCherry-Ago2 in live PC12 cells under control, 150 mOsm osmotic stress for 5 minutes and 5µ carbachol stimulation. E- Particles associated with eGFP-PLCβ1 under basal (300 mOsm) and stress (150 mOsm, 5 min) conditions. Measurements in A-C are an average of 3 independent experiments that sampled 10 cells, while measurements in D-E are an average of 3 independent experiments that sampled 5 cells. For A-E, SD is shown and the p values were determined using ANOVA.
The studies above were carried out in fixed, stained cells. We also followed Ago2 particle formation in live cells by transfecting PC12 cells with mCherry-Ago2. While the number and areas varied somewhat with the level of transfection, the results show the same trend as the immunostained samples (Fig. 7D), (i.e. Gαq stimulation or osmotic stress increases the number of smaller mCherry-Ago2 associated particles). These studies indicate that reduction of cellular PLCβ1 helps to drive Ago2 to small cellular particles.
While the increase in PABPC1 and Ago2 particle assemblies could simply be due to the loss of cellular PLCβ1, it may also be due to removal of PLCβ1 from pre-formed particles. To address this question, we transfected PC12 cells with eGFP-PLCβ1 and analyzed the particles (Fig. 7E). We could not detect particles below 400 µm2 after which the number climbed to ∼1000. No differences were found in cells subjected to osmotic stress. These data suggest that PLCβ1 does not associate with large particles in the cell. In support of this idea, we used fluorescence correlation spectroscopy to monitor changes in the diffusion of eYFP-PLCβ1 when osmotic stress is applied. We find the diffusion coefficient is unchanged at 4.2 +/-0.4 µm2/s during the first 10 minutes of exposure to osmotic stress.
The differences in the size and number of PABPC1 versus Ago2 particles suggest they partition into different types of granules. We tested this idea by monitoring the effect of PLCβ1 and osmotic stress on the colocalization between Ago2 and PABPC1 (Fig. 8). Under normal osmolarity, we find little colocalization between the proteins both at endogenous and knocked-down levels of PLCβ1. However, when the cells are subjected to osmotic stress, colocalization between the species increases, and this increase is more pronounced when PLCβ1 is down-regulated. These results suggest that PABPC1 and Ago2 form distinct particles that may begin to fuse or associate under high stress conditions, such as loss of cytosolic PLCβ1 and osmotic stress.
A- Images of PC12 cells immunstained for Ago-2 (green) and PABPC1 (red) under (top to bottom) basal conditions, when the osmolarity is lowered from 300 to 150 mOsm, when cells are treated with siRNA(PLCβ1) under normal conditions and under osmotic stress. The scale bar length is 20 µm. B- Graph of the resulting colocalization data were a significant change between 50% osmotic stress sample with wt-PLCβ1 versus siRNA-PLCβ1 samples is noted p=0.002 n=20.
Assembly of Ago2 stress granules depends on the type of environmental stress
It is probable that osmotic stress produces granules that are different in size, number and composition than those produced by other stresses. We compared the formation of Ago2 particles under different types of stress by Number & Brightness (N&B) analysis (see Methods). This method measures the number of fluorescent molecules associated with a diffusing particle in living cells (33) (see Methods). Thus, N&B measurements of cells expressing eGFP-Ago2 will indicate the conditions that promote the formation of Ago2 aggregates.
In Fig. 9 we present the N&B histograms for some example cell images (top panels), visualization of aggregates in the cells (middle panels), and the corresponding fluorescence images (bottom panels) where the purple areas correspond to higher intensities. To analyze these images, we selected the regions of N&B values that correspond to free eGFP (Fig. 9A) as outlined in red in the upper panels, and appear in red in the images in the middle panels. We then determined the number of pixels in each of the N&B images that had values outside these regions that correspond to small Ago2 aggregates (n=2-5) as shown in green, or large Ago2 aggregates shown in purple. The number of pixels corresponding to Ago2 aggregates were tabulated for 10-14 images and are given in Fig. 9.
The top panels (A-F) are the N&B data that display the pixels that correspond to different brightness versus intensity values. The pixels contained in the red boxes correspond to the same values found for free eGFP and correspond to monomeric eGFP-Ago2. Points outside the red boxes, shown in green and blue, correspond to higher order species. The pixels corresponding to monomeic (red) and higher order mEFGP-Ago2 aggregates (green and blue) can be seen in the cell images directly below (middle panels, G-L). The bottom panels (M-R) show the corresponding fluorescence microscopy images where the purple pixels denote Ago2 aggregates, and the scale bar length is 10 µm. Panels A, G, M are control cells; Panels B, H, N are cells subjected to hypo-osmotic stress (150 mOsm, 5 min); Panels C, I, O are cells subjected at cold shock 12°C for 1 hr; Panels D, J, P are cells subjected to heat shock at 40°C for 1 hr; Panels E, K, Q are cells subjected to arsenite stress (0.5mM, 10 min); Panels F, L, R are cells subjected to carbachol stimulation (5µm, 10 min). We note that while all cells showed similar aggregation behavior within the error reported in the figure, in the osmotic stress studies, 20% of the cell did not change. These cells were omitted from the error calculation.
We followed Ago2 aggregation in PC12 cells subjected to a variety of stress conditions (Fig. 9). Subjecting cells to osmotic stress resulted shift the distribution of eGFP-Ago2 particles to a point where ∼60% of the eGFP-Ago2 were significantly larger than a monomer (Fig. 9B). It is notable that these Ago2 stress granules formed throughout the cytoplasm, and is notable that only 75-80% of cells showed aggregation. We compared the aggregation of eGFP-Ago2 in cells subjected to other stresses: cold shock (12°C for 1 hour), heat shock (40°C for 1 hour) (Fig. 9C-D), arsenite treatment (0.5mM for 10 min) and oxidative stress (1mM CoCl2 8hrs, supplemental). Unlike osmotic stress, these other stresses produced changes in all cells observed but eGFP-Ago2 aggregates were seen in only a few large particles rather than evenly distributed through the cell.
In a final series of studies, we stimulated cells under normal conditions with carbachol to activate Gαq and promote relocalization of cytosolic PLCβ1 to the plasma membrane (Fig. 9F,L,R). Unlike other stresses, we find the formation of small eGFP-Ago2 aggregates distributed throughout the cytosol. This behavior was seen in every cell tested. These data show that different stresses, including Gαq activation, result in patterns of formation of stress granules containing multiple Ago2 molecules.
Cytosolic PLCβ1 levels impact the size of cytosolic RNAs
If PLCβ1 is inhibiting the ability of Ago2 to form stress granules, we would expect an increase in the size distribution of cytosolic RNA as mRNA accumulates due to the arrest of translation. We measured the sizes of cytosolic RNA by dynamic light scattering (DLS) (Fig. 10). Subjecting cells to osmotic stress caused a significant shift to larger sizes. Down-regulating PLCβ1 resulted in a small peak at low molecular weights followed by a broad peak at larger sizes that shifts to the right when compared to control. The small peak is consistent with enhanced C3PO activity due to the relief of inhibition by PLCβ1 (25). Over-expressing Gαq resulted in a similar behavior as treatment with siRNA(PLCβ1). Surprisingly, neither cold, heat, oxidative stress or arsenite (supplemental) increased the average size of cytosolic RNA. These data are consistent with the translation arrest and accumulation of higher molecular weight RNAs due to Ago2-associated particles when the cellular level of PLCβ1 is reduced.
A- The size distribution of cytosolic RNA isolated from PC12 cells was measured by dynamic light scattering for control conditions (black) and was found to shift to higher sizes with hypo-osmotic stress (150 mOsm 5 min, green), over-expression of constituatively active Gαq (blue), and down-regulation of PLCβ1 (pink). Note that a peak at small RNA sizes is seen for these latter two samples due to reversal of inhibition of C3PO activity by PLCβ1. B- An identical study study showing the size distribtion of control cells (black), cells subjected to hypo-osmotic stress (150 mOsm 5min, red), heat shock (40°C for 1 hour, green), cold shock (12°C for 1 hour, blue), arsenite treatment (0.5mM 10 min, pink). We note that no changes were observed in cells subjected to oxidatiave stress (1 mM CoCl2 8 hrs) (suplemental). Normalized data are shown. Each sample was scanned 3 times with 10 minutes per run. The number of independent samples were control samples 6, PLCβ1 knockdown 2, and Gαq over-expression, 1.
Cytosolic PLCβ1 levels affect stress granules in smooth muscle cells
Myocytes and other cell types may experience changes in osmotic conditions during their lifetime. With this in mind, we extended these studies to two different smooth muscle cell types, rat aortic smooth muscle (A10) and Wystar Kyoto rat 3M22 (WKO-3M22) cells. We first focused on A10 cells where we identified PLCβ1-associated proteins from a monoclonal antibody pull-down by mass spectrometry under basal conditions and 5 minutes after hypo-osmotic stress (supplemental). Interestingly, a large fraction of proteins pulled down with PLCβ1are associated with transcription which is most likely due to the nuclear population of PLCβ (see (34)). Stress granule proteins appear at lower levels. Many of the stress granule proteins were also found in PC12 cells, e.g. PABPC1 and eIF5A (Fig. 11A), while others, such as Ago2 and FXR, are not seen. Repeating these studies using A10 cells subjected to hypo-osmotic stress for 5 minutes before lysing, resulted in a loss in almost all of these transcription-associated proteins and most stress granule proteins (Fig. 11B). These results show that PLCβ1 binds to stress granule-associated proteins in A10 as well as PC12 cells, and that osmotic stress releases these proteins from PLCβ1. While the cellular amount of PLCβ1 in A10 cells is reduced with osmotic stress, the the effect is much lower than in PC12 cells (supplemental).
A- Mass spectrometry analysis for proteins associated with PLCβ1 in A10 cells under normal osmotic conditions and B- after 5 minutes of hypo-osmotic stress (150 mOsm). Shown below (C-E) are results for The size and number of particles associated with anti-PABPC1 in the cytosol of A10 cells was measured on a 100x objective and analyzed using Image J (see methods). C- Treatment of cells with siRNA(PLCβ1) results in the formation of larger aggregates relative to mock-transfected controls. D - Osmotic stress (150 mOsm) enhances the formation of PABPC1-associated particles and this behavior is promoted by PLCβ1 down-regulation. E- Stimulation of Gαq by treatment with 5 µM carbachol also impacts particle size. All measurements are an average of 3 independent experiments that sampled 10 cells, where S.D. is shown and p values were determined using ANOVA.
We measured the formation of PABPC1 particles in A10 cells when the cytosolic levels of PLCβ1 are perturbed, i.e. osmotic stress, siRNA down-regulation and Gαq stimulation (Fig. 11C-E). These cells showed particles that are ∼10 fold larger than in PC12 cells. Like PC12 cells, down-regulating PLCβ1 results in larger particles, and osmotic stress increases the formation of large particles. Also, carbachol stimulation results in a significant increase in particles. In summary, these studies show that the size and number of PABPC1 particles depend on the cytosolic level of PLCβ1.
Both osmotic stress and carbachol also promoted Ago2-related stress granules in WKO-3M22 cells whereas arensite had a much smaller effect. In Fig. 12A-L we show N&B results of these cells subjected to osmotic and arsenite stress, and carbachol stimulation. Unlike PC12 cells, these stress granules were large and distributed through the cytoplasm. Additionally, ∼20% of cells subjected to osmotic stress did not form granules whereas all cells subjected to arsenite stress or carbachol treatment formed granules. Down-regulating PLCβ1 promoted Ago2 aggregation in both control and arsenite stressed cells. Interestingly, down-regulating PLCβ1 did not result in higher levels of Ago2 particles in cells subjected to osmotic stress or carbachol stimulation which may indicate that in untransfected PC12 cells, these stresses lower PLCβ levels to give results similar to the down-regulated ones. In a final series of studies, we monitored shifts in the sizes of cytosolic RNA in WKO-3M22 cells with osmotic stress and carbachol stimulation and found significant increases in the RNA sizes under both conditions (Fig. 12M).
The top panels (A-D) show graphs of the brightness versus intensity with the pixels of the colored boxes corresponding to the specific regions in the cells (E-H). The bottom panels show the corresponding fluorescence microscopy images where the scale bar length is 10 µm (I-L). The red box corresponds to monomeric eGFP-Ago2 while points outside this box and in the green and blue boxes correspond to higher order species. Panels A, E, I are control cells; Panels B, F, J are cells subjected to hypo-osmotic stress (150 mOsm, 5 min); Panels C, G, K are cells subjected to arsenite stress (0.5mM, 10 min); Panels D, H, L are cells subjected to carbachol stimulation (5µm, 10 min); Panel M are cells treated with siRNA PLCβ1; Panel N are cells treated with siRNA PLCβ1 and arsenite (0.5mM 10 min); Panel O are cells treated with siRNA PLCβ1 and osmotic stress (150 mOsm); Panel P are cells treated with siRNA PLCβ1 and carbachol stimulation (5µM 10 min); Panel Q is a plot of the size distribution of cytosolic RNA of WKO-3M22 cells as measured by dynamic light scattering for control conditions (black), osmotic conditions (150 mOsm 5min, red) and stimulation of Gαq by treatment with carbachol (5 µM 10 min, green).
Dependence of stress granule formation on PLCβ1 levels
To understand the dependence of stress granule formation on the concentration of PLCβ, we assume that eIF5A is the primary contact between PLCβ1 and stress granule proteins. We can then express the partitioning of eIF5A from the cytosol (c) to the particulate phase (p) as:
where eIF5Ap is the stress granule phase also termed ‘G’. The total amount of eIF5A is:
We can express the association between PLCβ and eIF5A in terms of a bimolecular dissociation constant.
In this equation, PLCβ refers to cytosolic PLCβ. We only consider the cytosolic population and not the membrane-bound one in accord with our results showing that loss of cytosolic of PLCβ promotes stress granule formation. Thus, the total cytosolic amount of PLCβ is:
If we combine these equations to determine the relationship between and [PLCβ], we obtain an equation that is quadratic in G (i.e. eIF5Ap.
To give:
To determine the applicability of this model, we first need to estimate values for G. We find a linear dependence between the number of particles and the average area of the particles for PABPC1 (Fig. 13A-B) in PC12 and possibly in A10 cells. This linearity allows us to estimate G using either of these measurements. We note that this linearity does not occur for Ago2 particles where stress primarily increases the number of particles rather than the size (Figs. 6–7).
Relationship between the number of particles and average area of PABPC1 particles found in A- PC12 and B- A10 cells. C- shows the inverse relationship of the number of particles and the estimated concentraiotn of total cytosolic PLCβ1.
Under basal conditions, PLCβ1 is distibuted both on the plasma membrane and in the cytosol where it may interact with stress granules proteins. Activation of Gαq promotes movement of PLCβ1 to the plasma membrane releasing stress granule proteins and promoting particle formation.
We can estimate the total amount of cytosolic PLCβ using single molecule fluorescence measurements. Briefly, we measure the number of eYFP-PLCβ1 molecules diffusing in a specific confocal volume after calibration, and we note that we typically over-express ∼2 fold more protein as indicated by western blotting. Our measurements give a cytosolic eYFP-PLCβ1 concentration of ∼43 nM in PC12 cells and ∼49nM in A10 cells which is reduced to 10-15 nM under hypo-osmotic conditions. This decline can be compared with the ∼2.5 fold reduction in cytosolic PLCβ1 upon carbachol stimulation (e.g. Fig. 5B) and the 80-90% reduction in PLCβ1 levels with siRNA treatment. Although these values for PLCβ1 are approximate, we can use them to determine its dependence on stress granules as expressed as G2. The data in Fig. 13C show the sensitivity of stress granules in this concentration range of PLCβ1 (see Discussion)
DISCUSSION
The studies presented here support the intriguing idea that cells direct signals through the reversible sequestration of proteins in membraneless organelles. In some cases these structures promote protein interactions by reducing the local concentration and mobility of the proteins, whereas in other cases they effectively halt a functional pathway (35, 36). In this work, we show that the atypical cytosolic population of PLCβ1 helps to organize stress granules in response Gαq signals. The specific types of stress granules that result from Gαq/PLCβ1 signals have properties and compositions that may be distinct from other types of stress. As detailed below, our studies suggest that PLCβ1 regulates the entry of Ago2 and other stress granule proteins into particulates through a simple thermodynamic binding mechanism that is competitive with Gαq, and that is absolutely dependent on the cytosolic level of PLCβ.
The traditional function of PLCβ is to generate calcium signals through molecules that activate Gαq such as acetylcholine, dopamine, serotonin, melatonin, histamine and angiotensin II. However, it is clear that PLCβ1 plays multiple roles in cell function. Cocco and others have shown that PLCβ1 can localize to the nucleus to regulate cell growth and differentiation (see (37, 38)). Our lab has found that PLCβ1 has a stable cytosolic population that impacts various cell functions, such as RNA silencing, and neuronal cell differentiation and proliferation (39, 40). These alternate cytosolic functions of PLCβ only occur at specific and limited times in the cell cycle. For this reason, we set out to determine whether cytosolic PLCβ1 binds to other proteins in non-differentiating cells under non-stimulated conditions. We used a proteomics approach to uncover potential interacting proteins and validated some of these interactions in living cells. Our studies show that cytosolic PLCβ1 associates with stress granule proteins in intact cultured cells.
Stress granules are RNA/protein aggregates that allow cells to halt the translation of non-essential proteins when they are subjected to environmental stress. We find that many of the proteins that associate with PLCβ1 complexes directly contribute to RNA processing and ribosome assembly; and these proteins are found in PLCβ1 complexes isolated from PC12 cells as well as A10 cells. During the initiation stage of protein translation, polyadenylate binding protein PABP binds to the tail of mRNA which then associates to eIF4 allowing the mRNA-protein complex to bind the 40S subunit (41). A cytosolic form of PABP, i.e. PABPC1, was found in our PLCβ1 pull-downs, along with eIF4 subtypes, which binds to PABPC1 (see https://www.uniprot.org/uniprot/P11940). Several other eukaryotic initiation factors were also identified in our analysis.
An important step in the progression of translation is the hydrolysis of GTP on eIF2 catalyzed by the GTPase activity of eIF5. Our results indicate that eIF5A may be a primary binding partner for PLCβ1’s association with stress granule proteins. EIF5A is found at very high levels in cells arrested at the G2/M checkpoint where protein translation is expected to be low. EIF5A has several regions in its sequence that are homologous to Gαq including a region where Gαq directly contacts PLCβ (aa 147-162) as indicated by homologous sequence alignment and chemical cross-linking (see (42)). We directly tested the association between PLCβ1 and eIF5A using purified proteins. Not only was the PLCβ1-eIF5A affinity in the same range as the affinity between PLCβ1 and C3PO (14) but binding was competitive both in solution and in cells. Because the binding site of C3PO overlaps the binding site of Gαq, PLCβ1 association with eIF5A will depend on the level of Gαq activation, and this behavior is observed in our studies. Thus, in the absence of Gαq stimulation, a population of cytosolic PLCβ1 may associate to eIF5A until specific events such as differentiation, causes PLCβ1 binding to shift to C3PO and inhibit of RNA-induced silencing.
We find that cytosolic PLCβ1 also binds to Ago2 as seen in pull-down studies, co-immunoprecipitation and FRET/FLIM. Our ability to disrupt their interactions by the addition of purified eIF5A suggest that it is a primary association. Ago2 is the key nuclease component of the RNA-induced silencing complex (RISC) (see (43)) and our previous work hinted an association between these enzymes (13). Sequence alignment of Ago2 and the TRAX subunits of C3PO shows four homologous regions ranging from ∼20-40 aa in length and from 21-30% identity and 40-56% homology (2-54, 87-119, 202-228, 109-136 on C3PO and 788-826, 555-598, 188-202, 831-858 Ago2). It is notable that at least three of the C3PO regions are potential interaction sites for PLCβ1 binding and at least one of these may be available for PLCβ1-Ago2 binding (25). By this argument, it is possible that PLCβ1 directly binds Ago2 through similar interactions as C3PO.
We determined whether PLCβ1 impacts stress granule formation by monitoring the behavior of two established stress granule markers, Ago2 and PABPC1. We initially used mild osmotic stress that may occur physiologically. Hypo-osmotic stress initiates a series of cellular events to reduce the number of osmolytes in the cell, such as the synthesis of glycogen from glucose, and as well as ion flow (44). While we expected osmotic strength to change the ability of PLCβ1 to interact with stress granules proteins by changes in tertiary or quaternary structure, we were surprised to find a large reduction in PLCβ1a in PC12 cells when osmotic stress is initially applied, although this effect is far less pronounced in A10 cells. PLCβ1a and 1b isoforms differ by ∼20 amino acids in the C-terminal region but are similarly activated by Gαq. Cocco and colleagues have found that PLCβ1b, but not PLCβ1a, prevents cell death under oxidative conditions by impacting levels of key signaling proteins (45). Additionally, these two PLCβ1s may localize differently depending on the cell types (29–32, 46, 47). While our studies cannot adequately distinguish between these isozymes, it would be interesting to see the separate roles they may play in stress granule formation. We note that in conjunction with changes in PLCβ1 levels or properties that occur with osmotic stress, we also varied cytosolic PLCb1 levels by stimulating Gαq to drive PLCβ1 to the plasma membrane, and we down-regulated the enzyme using siRNA(PLCβ1). All these methods showed a connection between cytosolic PLCβ1 and the formation of particles.
Our studies indicate differences in size, distribution and composition of stress granules formed by different stresses. We find a dramatic assembly of Ago2 under osmotic stress to produce particles distributed throughout the cytosol. Alternately, arsenite, oxidation and temperature shock produce large particles containin primarily monomeric Ago2. Additionally, osmotic stress results in a large increase in the size distribution of cytosolic RNA whereas asenite, oxidation and temperature does not. Studies in S.cerevisiae (3) indicate that hypo-osmotic stress promotes particles composed of markers of both P-bodies and stress granules supporting our findings that subjecting mammalian cells to hypo-osmotic stress forms particles with compositions that differ from other types of stress. Our results also suggest that these latter granules, which have low Ago2 content and are rich in proteins associated with RNA processing, such as G3BP, are poised to prevent the transcription of genes whose protein products would not survive asensite stress or oxidation, such as those involved in phosphorylation (48). We speculate that granules mediated by the Gαq/PLCβ1 pathway may shift the transcription to genes whose protein products allow cells to better respond to external signals. Thus, unlike arsenite or other stresses, Gαq activation may give rise to more physiologically relevant particles.
We monitored the appearance of stress granules under hypo-osmotic conditions structurally by fluorescence imaging and functionally by the accumulation of large cytosolic RNAs. Parker and colleagues have shown that initially, stress granules are small and grow in size in a time-dependent manner (5). Here, we resolved particles over 10 µ2 that form in the cytoplasm and the size and number of particles did not vary between 5 and 10 minutes after induction of stress. Additionally, while a very small population of eGFP-PLCβ1 incorporated into particles ∼400 µ2 these were unchanged with osmotic stress suggesting that PLCβ1 delivers proteins into particles but does not incorporate into them. PABPC1 was associated with a high number of aggregates whose numbers were affected by the level of cytosolic PLCβ1 as determined by immunofluorescence. Formation of Ago2-associated particles, as monitored by both immunofluorescence and live cell imaging, was sensitive to Gαq stimulation but not other stresses. These data suggest that cells respond to Gαq activation by sequestering Ago2 into stress granules to halt the production of specific proteins. Additionally, we find that the stress granules generated by Gαq activation are similar, but have some distinction from other forms of stress. These particles have lower levels of Ago2 aggregation as compared to osmotic stress with uniformly distributed particles. Thus, we conclude that the type of stress regulates the composition of stress granules including the amount of mRNA-Ago2 complexes.
A loss in cytosolic PLCβ1 may arrest the translation of specific proteins by promoting the formation of mRNA-Ago2-associated stress granules. In a preliminary study, we monitored the levels of the heat-inducible chaperone Hsp90A (49) with PLCβ1 down-regulation since Hsp90 is not strongly regulated by PLCβ1-C3PO (13) and because its transcripts may be associated with stress granules formed by arsenite exposure stress (50). Western blotting shows a reduction in Hsp90 levels with siRNA(PLCβ1) consistent with increased formation of Ago2-associated stress granules (supplemental). The idea that sustained Gαq activation can regulate the production of specific proteins is intriguing and a comprehensive study of all of the transcripts affected by PLCβ1 is now underway.
Neurons and cardiomyocytes are long-lived cells and their health depends on reversible assembly of stress granules. We used PC12 cells as model for the role of stress granules in neurological diseases and A10 cells as a model for muscle cells regularly handle changes in osmolarity. We also used WYK-3M22 rat aortic smooth muscle cells as another model of muscle cells that is used as the nonmotensive control for spontaneously hypertensive rats, which are a common models of hypertension (51). We were pleased to find a similar set of stress related proteins in PLCβ1 complexes in the two cell lines with the exception of neural specific proteins and RNA-induced silencing proteins (i.e. Ago2 and C3PO). Thus, PLCβ1 may serve a similar role in many cell types by mediating stress granule formation but not in regulating RNA processing.
We constructed a simple thermodynamic model in which the partitioning of eIF5A into particulates is regulated by its association with PLCβ1 but note that eIF5A can easily be replaced with Ago2. The expression derived from this model shows the scope that PLCβ1 impacts stress granule formation. Specifically, if the total amount of eIF5A is much higher than PLCβ1, then stress granule formation will be independent of PLCβ1. Considering the high concentration of ribosomes in cells, it is difficult to estimate the amount of eIF5A available to bind PLCβ1. We know that microinjection studies that delivered ∼10nM of eIF5A into cells can displace C3PO from PLCβ1 which can give us a quantitative handle for future studies. Regardless of the specific nature of eIF5A and its associated proteins, our data show that there is a concentration range of PLCβ1 that is sensitizes cells to stress granule formation and that this range is under the control of Gαq activation. Additionally, endogenous levels of PLCβ1 help to control premature stress responses.
Our studies suggest a model in which cytosolic PLCβ1 binds to stress granules protein complexes, to keep these proteins disperse under basal conditions. Activation of Gαq shifts the cytosolic population of PLCβ1 to the plasma membrane displacing stress granules proteins and promoting the formation of particles. This dynamic nature of PLCβ1 fits in well with FCS studies showing rapid movement of the enzyme between the cytosol and plasma membrane (25). We propose a model (Fig. 15) in which cells use cytosolic PLCβ1 levels to regulate the formation and timing of protein synthesis, and to prevent the formation of irreversible aggregates. We note that the quenching of Gαq/PLCβ1 calcium signals in cells under osmotic stress suggests that the stress granules effectively block this signaling pathway. While the model in Fig. 15 explains our data, we emphasize that the importance of this study lies in the clear relationship between Gαq and stress granule formation.
We have previously found that PLCβ1 plays an important but unknown role in neuronal development. Specifically, the expression of PLCβ1a increases dramatically within the first 24 hrs of PC12 cells differentiation and the slowly decreases (15) leading to the question of why its expression is so variable. PLCβ1 is highly expressed in neuronal tissue where stress granule dysfunction has been implicated in disease (see (52, 53)). Our particle analysis data as well as DLS studies suggest that PLCβ1 may act as a chaperone to keep stress granule proteins disperse under basal conditions. It is notable that reduced PLCβ1 levels are associated with a host of neurological disorders that may result from disruptions in calcium signaling, neuronal cells proliferation and differentiation (e.g. (54–57). Schizophrenia and suicide has been shown to specifically involve varying levels of PLCβ1a and 1b in the prefrontal cortex (56). It is also notable that the PLCβ1-associated proteins identified here play important roles in neuronal function. FXR proteins are associated with the most common form of hereditary mental retardation (see (58, 59)) while eIF5A is associated with neuronal growth and brain development (60). It is interesting to speculate connections between PLCβ1 neurological disorders and those associated with FXR and eIF5Awhich may involve dysfunction in stress granule assembly / disassembly.
MATERIALS AND METHODS
Cell culture
Rat phenochromocytoma (PC12) and rat aortic smooth muscle (A10) cells were obtained from ATCC. Wystar Kyoto rat 3M22 (WKO-3M22) cells, originally obtained from ATCC, were a generous gift from Dr. Marsh Rolle. PC12 cells were cultured in high glucose DMEM (GIBCO) with 10% heat-inactivated horse serum (GIBCO) and 5% fetal bovine serum (Atlanta Biologicals). Rat aortic smooth muscle (A10) cell lines were cultured in high glucose DMEM with 10 % fetal bovine serum and 1% sodium pyruvate. WKY-3M22 cell lines were cultured in high glucose DMEM (Corning) without L-glutamine with 10% fetal bovine serum, 1% sodium pyruvate, 1% non-essential amino acids (VWR) and 1% L-glutamine (VWR). All cells were incubated at 37°C in 5% CO2. Cells were synchronized in the G2/M phase as described (16). Briefly, 2mM thymidine was added to cells for 24 hours, the medium was removed and replaced by fresh complete culture medium for 8 hours after, 40 ng/ml nocodazole was added.
Plasmids
EGFP-hArgonaute-2 (eGFP-Ago2) was purchased from (Addgene plasmid # 21981) and was prepared in the laboratory of Philip Sharp (MIT). MCherry-Ago2 was a gift from Alissa Weaver (Vanderbilt University). MCherry-TRAX-C1 plasmid was constructed by inserting TRAX gene between BamHI and EcoRI restriction sites in mCherry-C1 backbone using T4 DNA ligase (NEB). Plasmid transfections and siRNA knock-downs were done using Lipofectamine 3000 (Invitrogen) in antibiotic-free media. Medium was changed to one containing antibiotic (1% Penicillin/Streptomycin) 12-18 hours post-transfection. For every FLIM experiment, two separate samples were prepared: donor alone, donor in presence of acceptor.
Co-immunoprecipitation
PC12 cells were lysed in buffer containing 1% Triton X-100, 0.1% SDS, 1x protease inhibitor cocktail and 10 mM Tris, pH 7.4. After 200 μg of soluble protein was incubated with 2 μl of PLCβ1/Ago2 antibody overnight at 4 °C. After addition of 20 mg of protein A-Sepharose 4B beads (Invitrogen), the mixture was gently rotated for 4 h at 4 °C. Beads were washed three times with lysis buffer, and bound proteins were eluted with sample buffer for 5 min at 95 °C. Precipitated proteins were loaded onto two 10% polyacrylamide gel. After SDS-PAGE one gel was transferred to nitrocllulose membranes, proteins were detected by immunoblotting with anti-PLCβ1 (D-8, Santa Cruz) and anti-Ago2 (Abcam) antibody.
Application of stress conditions
For the hypo-osmotic stress conditions, the medium was diluted with 50% water for 5 minutes before it was removed and replaced with Hank’s Balanced Salt Solution (HBSS) for imaging. For the arsenite treatment, a stock solution of 100mM arsenite in water was prepared. Cells were exposed to a final concentration of 0.5mM arsenite for 10 minutes before the medium was removed and replaced by HBSS for imaging. For the heat shock, cells were incubated at 40°C for 1 hour whereas for the cold shock cells were incubated at 12°C for 1 hour. For the oxidative stress treatment, a stock solution of 1M CoCl2 was prepared and cells were exposed to a final concentration of 1mM CoCl2 for 12-16 hours (overnight) before the medium was removed.
FRET studies between purified PLCβ1 and eIF5A
PLCβ1 was purified by over-expression in HEK293 cells as previously described, see (26). Purification of eIF5A was based on the method described in (24). Purified eIF5A in a pET28-mhl vector was expressed in bacteria (Rosetta 2 DE3 plysS) by inoculating 100 mL of overnight culture grown in Luria-Bertani medium into a 2L of Terrific Broth medium in the presence of 50 μg/mL kanamycin and 25 μg/mL chloramphenicol at 37°C. When OD 600 reached ∼3.0, the temperature of the medium was lowered to 15°C and the culture was induced with 0.5 mM IPTG. The cells were allowed to grow overnight before harvesting and stored at −80°C. Frozen cells from 1.8L TB culture were thawed and resuspended in 150 mL lysis Buffer (20 mM HEPES pH 7.5, 300 mM NaCl, 5% glycerol, 2mM BME, 5mM imidazole, 0.5% CHAPS, protease inhibitor cocktail, 5 μL DNAase) and lysed using a panda homogenizer. The lysate was centrifuged at 15,000 rpm for 45 minutes, added to a cobalt column in binding (20 mM HEPES pH 7.5, 300 mM NaCl, 5% glycerol, 2mM BME, 5mM imidazole), equilibrated in a 4 x 1 mL 50% flurry of cobalt resin, passed the 150 ml of supernatant through each cobalt column at approx.. 0.5ml/min, washed with (1) 20 mM HEPES pH 7.5, 300 mM Knack, 5% glycerol, 2mM BME, 30 mm imidazole, and then (2) 20 mM HEPES pH 7.5, 300 mM NaCl, 5% glycerol, 2mM BME, 75 mM imidazole and eluted with 20 mM HEPES pH 7.5, 300 mM NaCl, 5% glycerol, 2mM BME, 300mM imidazole.
Protein associations were assessed by FRET using sensitized emission. Briefly, PLCβ1a and eIF5A were covalently labeled on their N-terminus with Alexa488 and Alexa594 (Invitrogen), respectively, and the increase in acceptor emission when exiting at the donor wavelength in the presence of Alexa488-PLCβ1 was noted. Studies were repeated using by pre-binding catalytically inactive C3PO with Alexa488-PLCβ1.
Mass Spectrometry
Mass spectrometry measurements were carried out as previously described at the University of Massachusetts Medical School (61). Cytosolic fractions were isolated from cells, and proteins bound to monoclonal anti-PLCβ1a (Santa Cruz, D-8) were separated by electrophoresis. Protein bands were isolated by cutting the gels into 1×1 mm pieces, placed in 1.5 mL eppendorf tubes with 1mL of water for 30 min. The water was removed and 200µl of 250 mM ammonium bicarbonate was added. Disulfide bonds were reduced by incubating with DTT at 50°C for 30 min followed by addition of 20 µl of a 100 mM iodoacetamide 30 min at room temperature. The gel slices were washed 2 X with 1 mL water aliquots. The water was removed and 1mL of 50:50 (50 mM ammonium bicarbonate: acetonitrile) was placed in each tube and samples were incubated at room temperature for 1hr. The solution was then removed and 200 µl of acetonitrile was added to each tube at which point the gels slices turned opaque white. The acetonitrile was removed and gel slices were further dried in a Speed Vac (Savant Instruments, Inc.). Gel slices were rehydrated in 100 µl of 4ng/µl of sequencing grade trypsin (Sigma) in 0.01% ProteaseMAX Surfactant (Promega): 50 mM ammonium bicarbonate. Additional bicarbonate buffer was added to ensure complete submersion of the gel slices. Samples were incubated at 37 C for 18 hrs. The supernatant of each sample was then removed and placed in a separate 1.5 mL eppendorf tube. Gel slices were further extracted with 200 µl of 80:20 (acetonitrile: 1% formic acid). The extracts were combined with the supernatants of each sample. The samples were then completely dried down in a Speed Vac.
Tryptic peptide digests were reconstituted in 25 µL 5% acetonitrile containing 0.1% (v/v) trifluoroacetic acid and separated on a NanoAcquity (Waters) UPLC. In brief, a 2.5 µL injection was loaded in 5% acetonitrile containing 0.1% formic acid at 4.0 µL/min for 4.0 min onto a 100 µm I.D. fused-silica pre-column packed with 2 cm of 5 µm (200Å) Magic C18AQ (Bruker-Michrom) and eluted using a gradient at 300 nL/min onto a 75 µm I.D. analytical column packed with 25 cm of 3 µm (100Å) Magic C18AQ particles to a gravity-pulled tip. The solvents were A, water (0.1% formic acid); and B, acetonitrile (0.1% formic acid). A linear gradient was developed from 5% solvent A to 35% solvent B in 90 minutes. Ions were introduced by positive electrospray ionization via liquid junction into a Q Exactive hybrid mass spectrometer (Thermo). Mass spectra were acquired over m/z 300-1750 at 70,000 resolution (m/z 200) and data-dependent acquisition selected the top 10 most abundant precursor ions for tandem mass spectrometry by HCD fragmentation using an isolation width of 1.6 Da, collision energy of 27, and a resolution of 17,500.
Raw data files were peak processed with Proteome Discoverer (version 2.1, Thermo) prior to database searching with Mascot Server (version 2.5) against the Uniprot_Rat database Search parameters included trypsin specificity with two missed cleavages or no enzymatic specificity. The variable modifications of oxidized methionine, pyroglutamic acid for N-terminal glutamine, phosphorylation of serine and threonine, N-terminal acetylation of the protein, and a fixed modification for carbamidomethyl cysteine were considered. The mass tolerances were 10 ppm for the precursor and 0.05Da for the fragments. Search results were then loaded into the Scaffold Viewer (Proteome Software, Inc.) for peptide/ protein validation and label free quantitation.
Fluorescence correlation (FCS) and
FCS measurements were performed on the dual-channel confocal fluorescence correlation spectrometer (Alba version 5, ISS Inc.) equipped with avalanche photodiodes and a Nikon Eclipse Ti-U inverted microscope as previously described (28).
Number and Brightness (N&B) measurements
N&B theory and measurement has been fully described (see (33). Experimentally, we collected ∼100 cell images viewing either free eGFP (control) or eGFP-Ago2, at a 66nm/pixel resolution and at a rate of 4 µs/pixel. Regions of interest (256×256 box) were analyzed from a 320×320 pixel image. Offset and noise were determined from the histograms of the dark counts performed every two measurements. Number and Brightness (N&B) data was analyzed using SimFC (www.lfd.uci.edu). The pixels corresponding to specific values of B and Intensity on the N&B plots were identified by SIM-FCS4 software (ISS, Inc).
N&B analysis
N&B defines the number of photons associated with a diffusing species by analyzing the variation of the fluorescence intensity in each pixel in the cell image. In this analysis, the apparent brightness, B, in each pixel is defined as the ratio of the variance, σ, over the average fluorescence intensity <K>:
and
where n is the number of fluorophores. The determination of the variance in each pixel is obtained by rescanning the cell image for ∼100 times as described above. The average fluorescence intensity, <K> is directly related to the molecular brightness, €, in units of photons per second per molecule, and n. B can also be expressed as
And the apparent number of molecules, N, as
Fluorescence lifetime imaging measurements (FLIM)
FLIM measurements were performed on the dual-channel confocal fast FLIM (Alba version 5, ISS Inc.) equipped with photomultipliers and a Nikon Eclipse Ti-U inverted microscope. A x60 Plan Apo (1.2 NA, water immersion) objective and a mode-locked two-photon titanium-sapphire laser (Tsunami; Spectra-Physics) was used in this study. The lifetime of the laser was calibrated each time before experiments by measuring the lifetime of Atto 435 in water with a lifetime of 3.61 ns (Ref) at 80MHz, 160MHz and 240MHz. The samples were excited at 800/850 nm, and emission spectra were collected through a 525/50 bandpass filter. For each measurement, the data was acquired until the photon count was greater than 300. Fluorescence lifetimes were calculated by allowing ω = 80 MHz:
Statistical analysis
Data was analyzed using Sigma Plot 13 statistical packages that included student’s t-test and one way analysis of variance (ANOVA).
Dynamic light scattering (DLS)
DLS measurements were carried out on a Malvern Panalytical Zetasizer Nano ZS instrument. For these experiments, mRNA from PC12 cells was extracted following the instructions from the Qiagen Mini Kit (Cat #: 74104). Cells were exposed to stress, treated with siRNA(PLCβ1a), or transfected constitutively-active GαqRC (62), before the mRNA was extracted. For these measurements, approximately 50µL of extracted mRNA in RNase free water was added in a Hellma Fluorescence Quartz Cuvette (QS-3.00mm). Each sample was run 3 times, 10 minutes per run. Control samples were repeated 6 times, PLCβ knockdown twice, and Gαq over-expression once.
Particle analysis
Samples were imaged using a 100X/1.49 oil TIRF objective to microscopically count the number of particles formed under different conditions per µm2. For each condition, 10-20 cells were randomly selected and z-stack measurements were taken (1.0 µ/frame). Analysis was performed using ImageJ where each measurement was thresholded before analyzing and averaging the number of particles per frame per measurement.
Microinjection studies
Microinjection of a solution of 100 nM eIF5A into PC12 cells was carried out using an Eppendorf Femtojet Microinjector mounted on Nikon Eclipse Ti-U inverted confocal microscope under 0.35 PSI pressure and 0.5 Seconds injection time per injection.
Immunofluorescence
Cells were samples were fixed with 3.7% formaldehyde and permeabilized using 0.1% triton X-100 in PBS then blocked using 10% goat serum, 5% BSA, 50mM glycine in PBS. Cells were then stained with primary antibodies from (Abcam), incubated for 1 hour, washed and treated with a secondary antibody for 1 hour. After another wash, the cells were viewed on either a Zeiss Meta 510 laser confocal microscope. Data were analyzed using Zeiss LSM software and Image J. The secondary antibodies used were Anti-rabbit Alexa-488 for Ago-2 and Anti-mouse Alexa 647 for PABPC1.
Calcium signal imaging
Single cell calcium measurements were carried out by labeling the cells with Calcium Green (Thermofisher), incubating in HBSS for 45 minutes and washing twice with HBSS. Release of intracellular calcium in live PC12 cells was initiated by the addition of 2 µM carbachol before imaging the time seirs on a Zeiss LSM 510 confocal microscope excitation at 488 using time series mode as previously described, see (63).
Western blotting
Samples were placed in 6 well plates and collected in 250μL of lysis buffer that included NP-40 and protease inhibitors as mentioned before, sample buffer is added at 20% of the total volume. After SDS-PAGE electrophoresis, protein bands were transfer to nitrocellulose membrane (Bio-Rad, California USA). Primary antibodies from (Santa Cruz) and (Abcam) are used. Membranes were treated with antibodies diluted 1:1000 in 0.5% milk, washed 3 times for 5 minutes, before applying secondary antibiotic (anti-mouse or anti-rabbit from Santa Cruz) at a concentration of 1:2000. Membranes were washed 3 times for 10 minutes before imaging on a BioRad chemi-doc imager to determine the band intensities. Bands were measured at several sensitivities and exposure times to insure the intensities were in a linear range. Data were analyzed using Image-J.
ACKNOWLEDGEMENTS
The authors are grateful for the support of NIH GM116187. Dr. Garwain was supported by a Richard Whitcomb Fellowship. The authors also would like to thank Dr. Elizabeth Bafaro for her help in cloning mCherry-TRAX and to Dr. Siddartha Yerramilli for help with N&B and for his helpful comments throughout this work.
