Regulation of Phosphatidylinositol-specific Phospholipase C at the Nuclear Envelope in Cardiac Myocytes

Abstract

Phosphatidylinositol 4,5 bisphosphate hydrolysis at the plasma membrane by phospholipase C is one of the major hormone regulated intracellular signaling systems. The system generates the diffusible second messenger IP₃ and the membrane bound messenger diacylglycerol. Spatial regulation of this system has been thought to be through specific subcellular distributions of the IP₃ receptor or PKC. As is becoming increasingly apparent, receptor-stimulated signaling systems are also found at intracellular membranes. As discussed in this issue, GPCRs have been identified at the nuclear envelope implying intracellular localization of the signaling systems that respond to GPCRs. Here we discuss the evidence for the existence of PLC signals that regulate nuclear processes, as well as the evidence for nuclear and nuclear envelope localization of PLC signaling components, and their implications for cardiac physiology and disease.

Introduction

Hormone-dependent regulation of phosphoinositide turnover is a fundamental process in regulation cell function and homeostasis. The classical paradigm that has developed over the past 40 years began with the original demonstration of hormone-stimulated turnover of phosphoinositides [1, 2] to the now well-understood stimulation-dependent hydrolysis of phosphatidylinositol 4,5 bisphosphate (PIP₂) by phospholipase C (PLC) to generate the second messengers inositol 1,4,5 trisphosphate (IP₃) and diacylglycerol (DAG). These second messengers drive cellular functions by binding to IP₃ receptors (IP3R) to release Ca²⁺ from the endoplasmic reticulum (ER) and through DAG dependent regulation of protein kinase signaling cascades through protein kinase C (PKC) activation [3]. In addition to the classical role of PIP₂ as a precursor for hydrolysis it is an important cellular regulator in its own right serving as a direct regulator of ion channel and cytoskeletal functions [4, 5]. PIP₂ is also a direct substrate for phosphorylation by phosphoinositide 3 kinases (PI3K) that to generate phosphatidylinositol 3,4,5 trisphosphate (PIP₃) an important regulator of Akt signaling and other downstream processes [6].

In most of these scenarios PIP₂ is localized and functions at the plasma membrane [7]. On the other hand PIP₂ has been found inside the nucleus associated with the nuclear matrix, along with all the enzymes and precursors required for its synthesis [8, 9]. Additionally PLCs have been found in the nucleus suggesting a role for PIP₂ hydrolysis in the nucleus [10]. Recently it has been described that the precursor for PI 4,5 P₂, PI4P can be a substrate for PLC in cardiac cells [11]. PI4P is found on the plasma membrane and intracellular membranes and may represent a novel intracellular substrate for PLC in multiple cell types [12, 13].

In this review we will discuss the some basic concepts in phosphoinositide signaling downstream of GPCRs and receptor tyrosine kinases, and how they may relate to signaling in the nucleus by PLC enzymes localized at, in, or near the nucleus in cardiac cells.

Phospholipase C enzymes

There are multiple isoforms and splice variants that comprise the family of enzymes that hydrolyze PIP₂. A basic outline of the PLC families and their regulation is presented here. For a detailed discussion of this topic please refer to several comprehensive reviews on the subject [14–17]. All PLC isoforms contain conserved domains that comprise the catalytic domain, an EF-hand domain, whose specific function is not well defined, and a C2 domain that binds phospholipids [14]. All PLCs require calcium for function but this does not involve the EF hands. Rather, calcium in the catalytic site is directly involved in catalysis [18]. All of the PLC isoforms except for the sperm specific PLCζ contain a pleckstrin homology (PH) domain that can either involved in phosphoinositide binding (PLCδ) [19] or binding of protein regulators such as Rac or Gβγ in the PLCβ isoforms [20]. The PLC β isoforms are activated directly downstream of GPCRs by virtue of direct binding of either G protein α_q or βγ subunits, or both [21–24]. Direct binding of the Rho family small G protein the, Rac, also regulates PLCβ’s [25]. The PLC γ isoforms are regulated by direct binding to phosphorylated tyrosine residues in receptor and non receptor tyrosine kinases [26]. PLCδ is highly responsive to calcium [27] and may respond to changes in calcium that occur secondarily to IP₃-dependent release initiated by PLC β or γ activity, or in response to calcium influx through other mechanisms [28]. PLCε has the potential to be downstream of multiple stimuli by virtue of its ability to respond to direct binding of multiple proteins of the Ras family, including Ras, Rho and Rap [17, 29]. PLCε can also be activated either directly or indirectly by Gβγ subunits [30].

Role of canonical PIP₂ hydrolysis in cardiac myocytes

Cardiac myocytes contain multiple isoforms described of PLC including PLCβ1 and β3, PLC δ1 and PLC γ1 and PLCε [31, 32]. In cardiac myocytes a role for IP₃ dependent calcium release downstream of PLC has been enigmatic because calcium levels in the cytosol control beat-to-beat contraction of the myocyte. With each heart beat depolarization of the sarcolemma tightly controls calcium- influx through L-type calcium channels coupled to release of calcium from the sarcoplasmic reticulum through the type 2 ryanodine receptor (Ryr2). How an IP₃-dependent increase in calcium would interact with this system is not as straightforward as in other cell types where IP₃ dramatically regulates global changes in cytosolic Ca²⁺. Type 2 IP₃ receptors are the predominant isoform found in cardiac myocytes [33, 34]. These IP₃ receptors are present at very low levels in the SR of ventricular myocytes and somewhat higher levels in atrial myocytes, but in both cases are greatly outnumbered by type 2 ryanodine receptors (Ryr2) [35]. In atrial myocytes IP₃ receptors may cooperate with Ryr2 to regulate the amplitude of calcium signals in response to stimuli that couple to PLC hydrolysis including angiotensin II (ATII), endothelin 1 (ET-1) and norepinephrine [36, 37]. Recently it has been shown that IP₃Rs in the SR in ventricular myocytes may contribute to action potential prolongation through local regulation of the Na⁺/Ca²⁺ exchanger in the sarcolemma [38].

In part because IP₃ receptors are present in relatively high levels in atrial myocytes and are present in the SR along with Ryr2, PLC activation by Gq/11-coupled GPCRs can enhance inotropic function but can also trigger arrhythmic responses [36, 37, 39, 40]. Normally Calcium release by Ryr2 is triggered by the rhythmic depolarization induced influx of Calcium through L-type calcium channels. The presence of IP₃ receptors in the SR can serve as an alternate calcium source that can trigger ectopic calcium release that is potentially arrythmogenic. In ventricular myocytes the level of IP₃ receptors in the SR is very low but increase during cardiac failure [41]. The levels of PLCβ and IP₃ generation can also increase during cardiac disease and may contribute to arrhythmias observed during failure [42].

Most relevant to this review, IP3R2 is enriched in the nuclear envelope (NE) of ventricular cardiac myocytes [33]. Other components of the calcium handling machinery are present on the nuclear envelope such as the Serca pump, and Ryr2 [43]. The presence of IP₃ receptors on the nuclear envelope, if localized facing the nuclear matrix, suggests that IP₃ can regulate nuclear calcium levels. Indeed it has been shown that ATII and ET- I treatment of either neonatal or adult cardiac myocytes results in release of calcium into the nuclear matrix without a significant associated rise in cytosolic calcium [44, 45]. This nuclear calcium increase is critical for cell growth and cardiac hypertrophy since blocking the calcium increase by blocking IP₃ receptors or blocking IP₃ itself with an IP₃ sponge blocks hypertrophy development [44–46]. Calcium in the nucleus regulates CamKII-dependent phosphorylation of histone deacetylase (HDAC) [47] and regulates calcineurin/NFAT activity in the heart [44]. Both HDAC and NFAT regulate expression of hypertrophic genes.

A widely accepted model for how the PLC signaling system regulates nuclear calcium is through PM GPCR-dependent activation of Gα_q, which in turn activates PLCβ and PIP₂ hydrolysis at the plasma membrane [48]. The IP₃ generated is diffusible in the cytoplasm [49] and can reach the nucleus to interact with IP₃ receptors to release calcium from the nuclear envelope, which is continuous with the SR. Evidence in favor of this hypothesis is that blocking PLC activation at the PM blocks hypertrophy [50] and blocking cytosolic IP₃ with a cytoplasmic IP₃ sponge blocks hypertrophy [46].

The other arm of the PLC pathway involves production of DAG and subsequent downstream activation of PKC and PKD. PKCs have been shown to play various roles in regulating cardiac function [51]. In the nucleus PKD in particular has been shown to be important for regulation of hypertrophy [52, 53]. Deletion of PKD decreases hypertrophy development in vivo and in vitro. A mechanism for PKD’s hypertrophic activity involves phosphorylation of HDAC leading to nuclear export and expression of hypertrophic genes. Unlike IP₃, DAG cannot diffuse through the cytoplasm but rather is retained in the membrane where it was generated. In order to achieve activation of PKD in the nucleus via the classical PIP₂ hydrolysis pathway, PKD would have to be activated at the PM and then diffuse to the nucleus. Indeed such translocation has been demonstrated in cardiac myocytes [54]. Recent evidence suggests that activation at the PM is not sufficient to maintain activation of PKD in the nucleus or may not be the only location where DAG-dependent PKD activation occurs [11].

Phosphoinositide signaling in the nucleus

Using a variety of biochemical approaches and recently developed fluorescent probes for detection of phosphoinositides, a consensus has emerged that PIP₂ is not present in significant quantity on intracellular membranes in most cells and is enriched in the PM of most cells [7, 12]. On the other hand PIP₂ is found inside the nucleus in an apparently non-membranous form associated with the matrix [8, 9, 55, 56]. This PIP₂ is not extractable with detergents indicating it is not associated with the nuclear membrane [56, 57]. Also found in the nucleus are PLCβ1 [10, 58], δ1 [59], DAG [10, 57], PI and PI4P [10, 57] indicating the presence of all of the components of the phosphoinositide cycle.

Nuclear matrix PIP₂ levels decrease as a result of PLC-dependent hydrolysis in response to stimulation with growth factors such as IGF-1 in Swiss 3T3 fibroblasts leading to nuclear DAG production and translocation of PKC to the nucleus [57, 60]. The mechanism for IGF-1 stimulation is through activation of nuclear MAPK, which phosphorylates PLC in the nucleus increasing its activity [57, 61]. Regulation of hydrolysis of the nuclear matrix pool of PIP₂ through GPCR driven signals has not been clearly demonstrated. However; as discussed in the next section, there is evidence for a GPCR stimulated IP₃ signaling pathway in the nucleus but the location for the PIP₂ substrate is not clear. The nuclear matrix phosphoinositide cycle has not been investigated in cardiac myocytes, but the data discussed above from other cell types, suggest one potential source of substrate for PLC pathways stimulated by GPCRs localized to the nuclear envelope. Theoretically this could represent a mechanism for GPCR-dependent DAG generation in the nucleus that could drive nuclear PKD activation. In this scenario GPCR activation in the inner nuclear envelope could activate PLCβ, but subsequent access of the matrix localized PIP₂ would involve an unconventional activity of this enzyme that would seem to require dissociation from the inner nuclear membrane. Classical G protein regulated activation of PLCβ isoforms occurs on a two-dimensional lipid bilayer surface where collisions between activated G protein and PLC can occur [20]. Additionally PLCβ isoforms prefer a two-dimensional lipid substrate for G protein-dependent activation [62]. The G protein-dependent enhancement of PLCβ catalytic activity requires the continuous engagement with the G protein. For G protein-activated PLC to access PIP₂ associated with in the nuclear matrix would require simultaneous dissociation of PLC and G protein from the NE membrane or access to nuclear PIP₂ while remaining bound to the NE. Additionally, the G protein-PLC complex would have to access PIP₂ in non-planar bilayer configuration.

Evidence for GPCR stimulated PIP₂ hydrolysis in the nucleus

Gq and Gs coupled GPCRs have been found at the nuclear envelope in cardiac myocytes [63–66] as well as other cell types [67] and specific roles for these receptors have been found in regulation of cell physiology [68–70]. A key question is whether Gq coupled receptors at this location couple to PLC to mediate local release of IP₃ and DAG. Evidence in favor demonstrating existence of this specific pathway in cardiac myocytes relies mainly of pharmacological evidence. Stimulation of nuclear ET1B receptors with caged ET-1 leads to a gradual increase in nuclear Ca²⁺ that is partially inhibited by the IP₃ receptor blocker 2-Aminoethoxydiphenyl borate (2-APB) or completely inhibited by ryanodine [71]. The only partial inhibition by 2-APB and the fact that 2-APB can inhibit Trp and Ori channels casts some doubt on whether the observed calcium release is IP₃ mediated. While the data support the idea that nuclear ET1B signals can cause a nuclear calcium increase, without further evidence it cannot be concluded that ET1B receptors are stimulating PLC-dependent IP₃ production. Indeed recent experiments from our laboratory failed to detect PIP₂ at the nuclear envelope in cardiac myocytes, although it is possible that very low levels of PIP₂ exist or the NE PIP₂ may not be accessible to the GFP-PLCδ-PH probe [11]. While, as discussed in the previous section, there is PIP₂ in the nucleus it is unclear whether a membrane associated PLC system could access this PIP₂.

In another study examining local nuclear calcium signaling through IGF-1 it was suggested that invaginations of the sarcolemma closely interact with the nuclear envelope to mediate Ca²⁺ signals [72]. While these detailed studies indicate that IGF-1 can initiate a local calcium signal in the nucleus and IGF-1 receptors are localized in invaginations that closely surround the nucleus, whether this system couples through the classical PLC system is not clear. The authors used the GFP-PLCδ-PH domain probe to follow apparent nuclear PIP₂ signaling but these data are not clear and directly contrast with PIP₂ localization studies from our own laboratory using the same cell preparation with GFP-PLCδ-PH domain or Tubby-GFP as cellular PIP₂ detectors.

In other cell types nuclear GPCR coupling to nuclear PLC signaling has been examined in some greater detail. As an example, in HEK293 cells and striatal neurons, overexpressed and endogenous, respectively, metabotropic glutamate receptor mGluR5 has been found at the nucleus [67]. Stimulation of nuclei isolated from mGluR5 transfected HEK cells with glutamate led to intranuclear calcium oscillations that were blocked by PLC inhibitors (U73122 and ET18OCH3), IP₃ receptor inhibitors 2-APB and Xestospongin, and by the ryanodine receptor antagonist, Ryanodine [73]. In intact striatal neurons, transfection of dominant negative Gα_q or PLCβ1 siRNA inhibited mGluR dependent increases in nuclear calcium [73]. Finally, the GFP-PLCδ-PH domain PIP₂ probe seemed to detect PIP₂ in the nuclear envelope in isolated nuclei from HEK cells and in intact striatal neurons. Stimulation of mGluR receptors led to an increase in nuclear fluorescence associated with GFP-PLCδ-PH domain interpreted as IP₃ production from the nuclear envelope [73]. These data support the existence of a mGluR-Gq-PLCβ1-PIP₂-IP₃ cascade in HEK293 cells.

As discussed earlier, the evidence for PIP₂ at membranes other than the PM is sparse. As with most experiments, lack of detection is a negative result, so it is difficult to definitively state that PIP₂ is not present on the NE. With the emergence of multiple GFP based phosphoinositide specific detector probes, most famously the GFP-PLCδ-PH domain as well as GFP-Tubby, multiple laboratories have examined the subcellular localization of PIP₂ in live intact cells [74–76]. The majority of these studies show prominent plasma membrane PIP₂ associated fluorescence and a lack of labeling of internal membranes. Additionally, antibodies raised against PIP₂ detect predominantly stain the plasma membrane [12, 77]. Some caveats to these studies are that GFP-lipid detection domains or PIP₂ antibodies may only detect enriched pools of lipid and while the overwhelming evidence supports enrichment of PIP₂ at the PM, lower levels of PIP₂ in other compartments may not be detected [78]. With GFP-lipid probes there may also be an issue with accessibility to the inner nuclear envelope, although tubby-GFP can be found in the nucleus but is not localized to the nuclear periphery as would be expected for NE association. Antibody based approaches require specific fixation and cell permeabilization protocols that may not preserve NE integrity.

The nuclear envelope is continuous with the endoplasmic reticulum which is enriched in phosphatidylinositol, the precursor for PI4P and PI4, 5P₂ synthesis. Higher level phosphorylated forms of PI have not been detected in the ER which may be in part due to the high levels of PI4P phosphatase activity associated with Sac1 that cycles between the ER and the cisGolgi [79]. As will be discussed later, this may function to preserve a gradient of PI4P between the cisGolgi and the ER that may drive lipid exchange between these compartments. If Sac1 is also present in the nuclear envelope it could explain a lack of PI4, 5P₂ in the NE.

On the other hand there are some reports using other methods that appear to detect NE PIP₂. One report used an immunostaining approach with isolated hepatic nuclei showing apparent peripheral staining consistent with NE localization [77]. Another approached used probed thin cryosections of cultured cells with GST-PLCδ-PH followed by immunoEM with anti-GST immunogold antibodies [80]. Here the majority of the labeling was at the PM but some labeling was observed at intracellular locations including the NE.

Thus while there is some evidence for PIP₂ in the nuclear envelope, more convincing evidence is needed to show that a sufficient pool of NE localized PIP₂ to support receptor-stimulated IP₃ production at this location. Improved nuclear-targeted PIP₂ detectors or improved imaging methods could be developed that might detect PIP₂ at low levels in the NE. Such methods that detect very low levels of lipids could always be subject to questions of specificity or background issues. One possible approach to resolve this more conclusively might be to target a PI5 phosphatase to the inner and/or outer nuclear envelope using a chemical dimerization approach FKBP-rapamycin binding domain(FRB)/FK506 binding protein (FKBP) fusion proteins. Here FRB might be fused to Nesprin or another NE specific protein and the PI5 phosphatase would be fused to FKBP. Rapamycin could be used to induce translocation of the PI5 phosphatase to NE and deplete the PIP₂ specifically in this compartment. This approach has been used to specifically deplete PIP₂ at the PM [75]. For this to be definitive it would have to be shown that detectable PIP₂ is removed at the NE by rapamycin treatment and that this eliminates receptor-stimulated responses.

Phosphatidylinositol 4-phosphate (PI4P) as an alternative PLC substrate at the nuclear envelope Golgi interface

Recent studies in our laboratory have focused on understanding the roles for PLCε relative to PLCβ isoforms in cardiac myocytes. We found that deletion of PLCε in cardiac myocytes prevents hypertrophy development in response to multiple stimuli [81]. In intact animals, PLCε deletion inhibits development of pressure overload hypertrophy [11]. In cardiac myocytes PLCε scaffolds to both Ryr2 and to muscle-specific A kinase anchoring protein (mAKAP) [11]. As discussed in this issue, mAKAP is highly enriched in the nuclear envelope and scaffolds to multiple signaling proteins associated with cardiac gene expression and hypertrophy to the nuclear envelope [82–85]. These include participants in the PLCε cascade such as Epac and PKC [86]. In cardiac myocytes disruption of PLC binding to mAKAP inhibits increases in cell size and hypertrophic gene expression driven by various extracellular stimuli [11]. Since PLCε is scaffolded to mAKAP at the nuclear envelope and regulates key myocyte functions that arise in the nucleus, it suggests that the local catalytic activity of PLCε is important at the nuclear envelope. As has been discussed, both products of PLC hydrolysis, IP₃ and DAG, regulate Ca²⁺ and protein kinase activity in the nucleus important for expression of hypertrophic genes.

Similarly to what was discussed in the previous section, in order for PLCε to produce local second messengers a substrate must be present. In our hands two different fluorescent PIP₂ probes, GFP-PLCδ-PH domain and GFP-Tubby domain, were unable to detect PIP₂ associated with the nuclear envelope of neonatal cardiac myocytes, adult cardiac myocytes or HEK 293 cells [11]. In contrast in AVMs and HEK293 cells clear PM localization of PIP₂ was observed. While it is still possible that PIP₂ is present in the membrane, the levels must be very low relative to the PM. On the other hand two different PH domain probes detected strong signals for the precursor to PIP₂, PI4P surrounding the nucleus. Stimulation of PLCε activity via direct Epac stimulation, or extracellular ET-1 led to a time dependent depletion of PI4P associated fluorescence that is dependent on PLCε activity. Pharmacologic or genetic depletion of PI4P, or disruption of PLCε-mAKAP scaffolding, block activation of a nuclear pool of PKD by ET-1, and inhibits hypertrophy development [11].

A surprising caveat to this model is that the PI4P fluorescence that surrounds the nuclear envelope in cardiac myocytes is associated with the Golgi apparatus, not the nuclear envelope [11]. This is based on: 1) Treatment with an ADP ribosylation factor 1 (ARF1) inhibitor that disrupts the Golgi apparatus, brefeldin A, eliminates Golgi PI4P associated fluorescence, but not nuclear envelope mAKAP staining. 2) PI4P associated fluorescence co-localizes with Golgi markers but not ER markers in myocytes. 3) PI4P in most cell types is enriched in the Golgi apparatus where it is plays an important role in vesicle trafficking and lipid transport [7]. In a standard confocal fluorescence microscope, Golgi associated PI4P fluorescence colocalizes with mAKAP staining (Malik and Smrcka, unpublished data), even though these are clearly in different membrane compartments. Thus the separation between these compartments in myocytes is at the limits or below the resolution limits of light microscopy. Indeed early EM ultrastructural analysis of skeletal and cardiac muscle myocytes indicates an unusual conformation of the Golgi apparatus where it tightly surrounds the outer nuclear envelope with a consistent distance of approximately 100–200 nm [87, 88].

Exactly how mAKAP-scaffolded PLCε on the nuclear envelope accesses Golgi PI4P is unclear. Some possibilities are diagrammed in Figure 1. Recent data suggests an active PI4P transport system at the ER-Golgi interface that transports PI4P from the trans Golgi to the ER at specialized interaction surfaces where the distance between the Golgi and the ER is 10–20 nm [89]. Since the nuclear envelope is an extension of the ER, the architecture of the Golgi in cardiac myocytes yields a large surface of ER-Golgi interactions with potential for formation of close contact sites. PI4P could be transported from the Golgi to the NE for where it could be hydrolyzed by PLCε (Fig. 1C), although this would have to occur at the cis face of the Golgi rather than trans, and no contact sites with dimensions less than 100 nm have yet been observed in myocytes. Another possibility is that PLCε in an extended conformation, bound to an extended conformation of mAKAP, which is in turn bound to an extended conformation of Nesprin 1α, could span NE-Golgi junctions to access PI4P for hydrolysis at the cis Golgi (Fig. 1A). This accessibility would depend on the dimensions of the proteins involved and the distance between the two membranes in cardiac myocytes. These dimensions should be carefully reexamined in cardiac myocytes to determine the viability of such models. A final possibility is that the K_d for PLCε binding to mAKAP could be in a range that could allow local transient dissociation from mAKAP such that it can access the local Golgi pool through a local diffusion based mechanism (Fig. 1B).

Figure 1.

Possible mechanisms for nuclear envelope scaffolded PLCε to access and hydrolyze PI4P found in the Golgi. A) An extended scaffold involving PLCε scaffolded to mAKAP, which is in turn scaffolded to nesprin-1α at the nuclear envelope may span the space between the outer nuclear envelope and the cis Golgi to hydrolyze PI4P in trans. B. Scaffolding of PLCε to mAKAP may bind PLCε in a way that allows dissociation and local action at the cis Golgi. Binding to of PLCε to mAKAP would serve to increase the local concentration of PLCε in the space between the NE and the cis Golgi. C. Phospholipid transfer proteins (PLTP) such as oxysterol binding protein (OSBP) have recently been shown to transport PI4P from the Golgi to the ER in exchange for cholesterol at Golgi-ER membrane contact sites. Since the NE is contiguous with the ER it is possible that this process could occur at the large Golgi-ER interface to supply PI4P to the NE where it could be accessed by mAKAP associated PLCε.

The consequences of PI4P as a substrate localized in close proximity to the NE are multiple. First, most obviously, IP₃ would not be generated from the hydrolysis reaction and thus this reaction would not be directly involved in regulation of nuclear calcium through IP₃ receptors. The only bioactive product produced would be DAG, which is consistent with a role for the PLCε in regulating nuclear PKD activation. This potentially makes sense because nuclear Ca²⁺ release in cardiac myocytes can be controlled by the restricted localization of the IP₃ receptor and diffusion of IP₃ from PLC dependent PIP₂ hydrolysis at the PM. DAG regulation of nuclear PKD is more problematic because DAG is not diffusible through the cytoplasm. It has been shown that some agonists stimulate transient PKD activation at the PM followed by subsequent diffusion to the nucleus [54]. On the other hand we showed that PLCε-dependent PI4P hydrolysis was required for nuclear PKD activity stimulated by ET-1 [11]. While there are some discrepancies to be resolved in these studies, a partial reconciliation can be proposed if initial PKD activation occurs at the PM but for this signal to be maintained in the nucleus, DAG has to be generated at or near the nucleus.

Implications of PI4P hydrolysis at the perinuclear Golgi for NE GPCR-stimulated PLC activation

As has been discussed here and in other articles in this issue, there is significant evidence for the existence of GPCRs in the nuclear envelope in cardiac myocytes. Caution should be taken with these results because GPCRs traffic through the Golgi in the biosynthetic maturation process as the receptors are transported to the cell surface. Since it is impossible to distinguish NE from Golgi by light microscopy in cardiac myocytes, and likely difficult to remove Golgi contamination from isolated nuclei, it remains a possibility that some apparently NE GPCRs in cardiac myocytes are indeed receptors in the process of being transported through the Golgi to the cell surface. Most of the cytochemical studies in cardiac myocytes demonstrating NE GPCRs have not examined potential Golgi localization.

Arguing in favor of GPCR localization at the NE are studies with α1-adrenergic receptor (α1-AR) showing that mutation of a nuclear localization sequence in the C terminal tail (α1A-AR-NLSmut) leads to localization away from the nucleus [90]. In myocytes isolated from α1AB-AR knockout mice, wild type α1A-AR rescues phenylephrine induced contraction while α1A-AR-NLS does not. These studies would strongly argue against Golgi localization for α1-AR. The authors go on to show that stimulation of α1-ARs in isolated nuclei from cardiac myocytes leads to activation of PKCδ. These results could be consistent with either PI4P or PIP₂ as a substrate for PLC pathway activation.

In another example, stimulation of intracellular ET1B receptors stimulates an elevation of intranuclear calcium, NO and alterations in gene expression in adult cardiac myocytes [69, 71]. As was discussed earlier, this could be the result of activation of PLCβ inside the nucleus to hydrolyze either PIP₂ in the nuclear matrix (Fig. 2A) or a difficult to detect PIP₂ substrate in the nuclear envelope (Fig. 2B). Alternatively one could imagine GPCRs at the NE stimulating PLCε scaffolded to mAKAP, or PLCβ that could be at the nuclear envelope to hydrolyze PI4P to generate local DAG (Fig. 2C and D). Such hydrolysis would not be expected to stimulate IP₃-dependent Ca²⁺ release but could play a very important role in local PKC or PKD activation. Indeed experiments with ET1B stimulated a slow Ca²⁺ release that is not typical of IP₃ dependent release and is only weakly blocked with an IP₃ receptor blocker [71] suggesting perhaps another mechanism for ET1B stimulated Ca²⁺ release.

Figure 2.

Scenarios for access to phosphoinositide substrates by PLCs downstream of GPCRs. A) GPCRs on the inner NE could activate PLCβ that could theoretically transiently dissociate to access PIP₂ in the nuclear matrix. PIP₂ has been found in the nuclear matrix. B) Similar to A but the PIP₂ substrate would be present in the nuclear envelope. This would be a preferred substrate for a PLC activated on the nuclear envelope but the evidence for the presence of PIP₂ at this location is lacking. C) As in Figure 1C, GPCR activated PLC on the outer nuclear envelope could access PI4P supplied to the nuclear envelope from the Golgi by PLTPs. D) PLCs activated by GPCRs on the outer nuclear envelope could transiently dissociate to access PI4P in the Golgi.

Whether PLCβ or PLCε is involved in NE GPCR-dependent regulation has not been demonstrated but either is theoretically possible. PLCε has been definitively identified as being involved in cardiac myocyte NE localized PI4P hydrolysis while PLCβ has not been clearly identified at this location. On the other hand, PLCε is not directly regulated by Gα_q but rather is regulated by small GTPases or Gβγ [91]. Thus Gq coupled receptors would rely on Gβγ released from Gq or a more complex cascade involving small GTPases to activate PLCε. Many Gq coupled receptors such as ET1 receptors couple to more than one G protein so Gi/Gβγ responses are also possible. Gs coupled receptors localized at the NE could also potentially regulate the PLCε PI4P/hydrolysis signaling system. The β-adrenergic receptor has been identified at the NE and adenylyl cyclase has been found associated with mAKAP [65, 92]. Locally generated cAMP could activate Epac, also found in the mAKAP signaling complex, which could activate PLCε through Rap activation as has been previously demonstrated [93, 94].

Summary

Clearly PLCs can regulate nuclear signaling processes. Mechanisms where PLC activity at the plasma membrane can generate signals that diffuse to the nucleus are well described. Emerging ideas based on the discovery of PLCs and GPCRs at or near the nucleus are changing our concepts about how PLC may regulate gene expression. Despite this emerging data, the details of how GPCRs may couple to PLCs at the nucleus remain to be dissected. One particular issue that must be better understood is what is the substrate and what are the products of PLC reactions that are stimulated by NE GPCRs or other receptors. This will require more sensitive probes to detect phosphoinositides at the nuclear envelope as well as more sophisticated imaging techniques to visualize possible PIP₂ in the NE assuming that there is actually PIP₂ present in the NE or that barely detectable levels of PIP₂ are able to support IP₃ production at a level that would activate IP₃ receptors.

The recent discovery that PLC at the nuclear envelope can access PI4P instead of PIP₂ suggests one possible alternative to the classical PIP₂ hydrolysis reactions that may occur at or near the nucleus. It is possible that the role of GPCR coupling to PLCs at the nuclear envelope is to generate local DAG rather than IP₃, obviating the need to find a PIP₂ substrate to support the physiological function of NE PLCs.