PLD was first characterized in vegetables by Hanahan and Chaikoff (1947) and changes in its lipase activity have been reported in relation to lipid metabolism in seed germination and lipid turnover and lipid composition during plant development and in membrane deterioration as a result of stress injuries. PLD was first cloned from cabbage (Brassica oleracea), castor bean (Ricinus communis L.), rice (Oryza sativa L.), corn (Zea mays L.), and rockcress (Arabidopsis), which code for an 808-amino acid cytosolic protein of ∼90 kDa MW, as mentioned in the review by (Cockcroft 1996). Plant PLD is linked to membrane deterioration during plant senescence as a result of decreased membrane phospholipid content. Additionally, PLD is present in large quantities in bacteria (Streptomyces), yeast (Saccharomyces cerevisiae), and mammalian cells.
Within mammalian sources, PLD has been found in a variety of cell types including neutrophils, promyelocytic leukemia, hepatocytes, platelets, endothelial cells, and spermatozoa, and is predominant in three organs, placenta, brain, and lung. A majority of the primary literature on PLD or PLD activity has been centered on in vivo and in vitro studies referring to changes in cellular distribution, intrinsic molecular alterations, association with other proteins or regulators, and availability of substrate as pertaining to phospholipid turnover in cellular membranes or wherever the substrate might be localized in the cell. PLD is expressed in a wide variety of tissues and cell lines and its activity has been reported predominantly in the plasma membrane, as well as in cytoplasmic locations, the mitochondrial membrane, the Golgi endoplasmic reticulum (ER), the nucleus, the nuclear membrane, and subcellular compartments. Additionally, PLD is palmitoylated on conserved cysteine residues and contributes to localization to membranous environments (Foster and Xu 2003).
Development of potent isoform-specific small-molecule PLD inhibitors would be integral to the advancement of the PLD field. Until recently, many PLD inhibitors lacked isoform specificity and did not act directly on the lipase. Halopemide and its subsequent derivative 5-fluoro-2-indoyl des-chlorohalopemide (FIPI) have been found to be very effective inhibitors of PLD-mediated F-actin cytoskeleton reorganization, cell spreading, and chemotaxis (Su et al. 2009). Use of iterative analog library synthesis approaches coupled with biochemicals assays and mass spectrometric lipid profiling of cellular responses has given rise to the next generation halopemide derivatives, which have yielded the development of dual PLD1/2, PLD1 selective and PLD2 selective inhibitors (Lewis et al. 2009). Small molecules that either indirectly or directly inhibit PLD1 or PLD2 could represent novel approaches for the treatment of metastatic cancer and inflammatory diseases.
Characterization of PLD
The PLD1 gene has been localized to the long arm (q) of chromosome 3 (3q26) (Park et al. 1998a), covers 210 kb of genomic DNA that is defined by 31 exons, whereby 27 exons result in the expression of four splice variants (PLD1a, PLD1a2, PLD1b, and PLD1b2) (Hammond et al. 1997; Katayama et al. 1998). PLD1a and PLD1a2 mRNAs express exon 19 (113 bp) and 29 (166 bp), respectively, while PLD1b and PLD1b2 do not express exon 19. PLD1a is the longest PLD1 splice variant at 1,072 amino acids in length and yields a 120 kDa MW protein. PLD1 is for the most part associated with perinuclear, Golgi and heavy membrane fractions, as reiterated in the review by Foster and Xu (2003). PLD1 is PC specific, Mg2+ dependent, and Ca2+ insensitive (Hammond et al. 1997), inhibited by oleate and has a basal level that is virtually undetectable. Human PLD1 is regulated by the cytosolic GTP binding protein Arf (Arf 1 and Arf3) and by small GTPases (Rac, Cdc42 and RhoA) via GTPγS, while it is also regulated by PKC (α and β isoforms) via Ca2+ and DAG/PMA (Cockcroft 1996; Hammond et al. 1997). Evidence of synergy between Arf, Rho, and PKC as related to regulation/activation of PLD1 activity has been reported first in human HL-60 leukemic cells and then in human neutrophils and rat brain, as reiterated by Cockcroft (1996). These facts implicate a widespread and ubiquitous nature to Arf-dependent PLD activity and specifically implicates only one PLD isoform in this process of lipase activation instead of activation by multiple other PLD forms.
The mammalian PLD2 gene is found on the short arm (p) of chromosome 17 (17p13) (Park et al. 1998b), is defined by 25 known exons of a genomic region spanning 16.3 kb, and encodes for two splice variants (PLD2a and PLD2b) of 933 amino acids in length each (Steed et al. 1998), which yields functionally indistinguishable proteins of 106 kDa MW. PLD2 is for the most part localized on the plasma membrane in light membrane lipid rafts that also associate with caveolin, as restated in the subject review by Foster and Xu (2003). The first PLD2 gene exon (112 bp) encodes for the 5′-untranslated region, the initiation codon (A1TG) is located on the second bp of exon 2, whereas the stop codon (TAG2803) is located 568 bp downstream in exon 25. The PLD2b variant is the result of 33 bp being alternatively spliced from exon 23 of the originally described PLD2a. PLD2 requires PIP2 and is largely insensitive to PKCα, Arf, or Rho (unlike PLD1 which is dependent upon these three cofactors).
Although the DNA sequences of both PLD1 and PLD2 share about 50% homology, all members of the PLD superfamily possess two highly conserved phosphatidyltransferase HKD catalytic domains (HKD1 and HKD2) that are defined by the consensus peptide sequence HxK(x)4D(x)6GSxN, which are vital to the lipase activity, as well as the phox homology (PX) and pleckstrin homology (PH) domains and the phosphatidylinositol 4,5-bisphosphate [PIP2] binding site (Frohman et al. 1999). As stated in the review by Exton (2000), PLD HKD motifs are requisite for catalytic activity and possibly dimerize to form an active center and are also present in biologically diverse proteins represented by bacterial phospholipid synthases and endonucleases, a pox envelope protein and a Yersinia toxin. Lysine to arginine point mutations of the HKD2 domain of PLD1 at K860 or of either the HKD1 or HKD2 domains of PLD2 at K444 or K758, respectively, result in lipase-dead enzymes because these K→R mutations yield lipases catalytically incapable of synthesizing PA or PBut as the readout for PLD activity. It has been theorized that the histidine in one of the HKD domains of PLD acts as a nucleophile to degrade the phosphodiester bond and the histidine in the other HKD domain protonates the oxygen of the leaving group, as reiterated by Exton (2000). The PX domain has been heavily implicated in binding to certain regulatory factors (PIP) and proteins (growth factor receptor-bound protein 2 (Grb2) and epidermal growth factor receptor (EGF-R)), while the PH domains of PLD1 and PLD2 have been demonstrated to function as strong modulators of the membrane recycling machinery that results in regulated growth factor receptor endocytosis and also linked to binding to SH2/SH3-containing tyrosine kinases. Deletion of either the PX or the PH domains results in a gross relocalization of PLD from the plasma membrane back to endosomes and in vivo renders the lipase unable to be activated, which ultimately negatively affects the catalytic activity of these isoenzymes.
PLD has been associated with a variety of physiological cellular functions, such as cancer cell progression, intracellular protein trafficking, cytoskeletal dynamics, membrane remodeling and cell proliferation in mammalian cells and meiotic division and sporulation in yeast. PLD regulation in mammalian cells falls into two major signaling categories: growth factors/mitogens that implicate tyrosine kinases (Frohman et al. 1999; Min et al. 1998) and small GTPases (Cockcroft 1996; Hammond et al. 1997; Powner and Wakelam 2002).
Role of Tyrosine Kinases and Phosphatases in PLD Signaling
Although PLD2 can be phosphorylated by the serine/threonine kinase AKT at residue T175 which serves to upregulate DNA synthesis, more typically PLD is known as a substrate for many receptor (EGF-R and PDGF-R) and non-receptor tyrosine kinases (Src and JAK3). Reagents like hydrogen peroxide when in the presence of vanadate can activate PLD in many different cells via tyrosine phosphorylation (Exton 2000; Min et al. 1998). Use of phorbol esters (PMA or TPA) or the PKC inhibitor Ro31-8220 to deplete PKC in the cell resulted in a significant loss of PLD activation (Cockcroft 1996). Additionally, evidence of considerable synergy between GTPγS and tyrosine kinase–based mechanisms has been reported using permeabilized cells as mentioned in the review by Cockcroft (1996). The PLD1 isoform is phosphorylated on tyrosine residues, which does not lead to changes in lipase activity (Min et al. 1998). The PLD2 isoform is expressed as a constitutively active enzyme in many different cell types that is detected as a phosphotyrosine protein in vivo and in vitro. These two scenarios heavily implicate a role for phosphorylation/dephosphorylation of PLD by protein tyrosine kinases and phosphatases in the control of PLD activity in response to such signaling mechanisms as osmotic stress, de novo DNA synthesis, cell proliferation, differentiation, transformation, and degranulation of mast cells.
Phosphorylation/dephosphorylation of PLD2 at certain tyrosine residues dictates whether or not PLD2 is activated or suppressed by certain signaling molecules at other subsequent tyrosine residues. This dichotomy is achieved in part through a complex process of phosphorylation by tyrosine kinases and dephosphorylation by phosphatases, such as CD45 and protein tyrosine phosphatase 1b (PTP1b) (Fig. 3b). PTP1b is already known to dephosphorylate EGF-R substrates and regulate the kinase in vivo. Experiments with phosphatases indicate that both activator and inhibitory sites exist on the PLD2 molecule (Gomez-Cambronero 2010). Low concentrations of phosphatases or phosphatases that target inhibitory or ambivalent sites specifically result in positive regulation of PLD2 and high lipase activity. Contrarily, high concentrations of phosphatases or phosphatases that specifically target activator sites result in loss of lipase activity based on the degree of cellular dependence on activator or inhibitory sites.
Additionally, phosphorylated PLD2 forms a ternary complex with both PTP1b and Grb2 (Gomez-Cambronero 2010), a critical signal transducer of EGF-R, via two SH2 recognition sites (Y169 and Y179) expressed within the context of the consensus YXNX in the PX domain of PLD2, which occurs independent of the lipase activity. A recent report indicates that increased cell transformation in PLD2-overexpressing cells occurs as a result of increased de novo DNA synthesis induced by PLD2 with the specific tyrosine residues involved in these functions being Y179 and Y511 (Gomez-Cambronero 2010). PLD2 residue Y169 modulates lipase activity, while PLD2 residue Y179 regulates total tyrosine phosphorylation of PLD2 (Gomez-Cambronero 2010). Complete simultaneous removal or phenylalanine replacement of these two sites on PLD2 completely abrogates or reduces binding to Grb2, respectively. Although the kinase that phosphorylates these two PLD2 residues is still unknown, interaction occurs through the C-terminal proline-rich domain of the Ras guanine-nucleotide exchange factor, Sos, and links PLD2 via residue Y169 to cellular proliferation and the MAPK and Ras/Erk pathways.
Role of Small GTPases in PLD Signaling
A role for PLD and its product PA has been presumed in the regulation of actin (Porcelli et al. 2002) and leukoctye cell migration (Gomez-Cambronero 2010) because the formation of lamellipodia structures and membrane ruffles can be abolished if PLD is inhibited. It has also been surmised that Rho GTPase stimulation of PLD activity is key to actin stress fiber formation and the ultimate regulation of cell movement because PLD activity has already been shown to aid the formation of stress fibers, as restated by Foster and Xu (2003). These findings indicate a potential signaling feedback mechanism does in fact exist between the activation loop (Switch 1) of the Rho family of small GTPases (RhoA, RhoB, RhoC, Rac1, Rac2, Cdc42, and TC10) and PLD and potentially protein kinase C (PKC) (Cockcroft 1996). These GTPases must be in the active (GTP- or GTPγS-liganded) form to yield PLD stimulation/activation (Exton 2000). Preincubation of plasma membranes from liver cells with RhoGDI (a protein that extracts membrane-associated Rho) led to the removal of both RhoA and Cdc42 concomitant with a decrease in PLD activity, which was reversed in part with the addition of recombinant RhoA and Rac1 (Cockcroft 1996). Rho and Rac activate the synthesis of PIP2 via the PI4-P5 kinase (PI4-P5K), and PIP2 controls PLD activity in vivo and in vitro via mediation of nucleotide-binding interactions, such as GTPγS regulation and Mg.ATP. As reported in a mini review elsewhere (Powner et al. 2002), members of the Rho family of small GTPases physically bind to PLD1 between amino acid 984 and 1000. Therefore, taking into consideration these sets of facts, it is likely that Rac, PIP2, and PLD are involved in the same signaling pathways and collectively regulate a variety of cellular functions.
In neutrophils, Rac1 plays an important role during gradient detections and actin assembly via PI-3K and AKT and has been reported to directly activate PLD1 (Powner and Wakelam 2002). Rho GTPases indirectly regulate PLD1 lipase activity via stimulation of PI(4,5)P2 kinase, Rho kinase and intracellular translocation of PLD (Powner et al. 2002). RhoA, Rac1, Arf, and Cdc42 also directly interact with and stimulate PLD1 activity in the presence of GTPγS (Cockcroft 1996), because mutation of the Rho-binding site on PLD1 abrogates PLD1–Arf interaction (Du et al. 2000). PLD1 is a downstream target of the Ras/RalA small GTPase cascade that has been associated with mitogenic and oncogenic signaling (Foster and Xu 2003).
PLD2 can be activated in intact cells by agonists and possibly by PLD1 (Foster and Xu 2003) and can be regulated by small GTPases and certain PKC family members (Du et al. 2000). PLD2 and Rac2 physically interact and heterodimerize in vitro, and recently, the biphasic effect of a monomeric GTPase acting as a master switch has been shown to both promote and inhibit phospholipase activity as related to the timeline of chemotaxis (Peng et al. 2011). Macrophages that overexpressed both Rac2 and PLD2 experienced a strong initial response toward the chemoattractant that was significantly decreased at later time points. This initial positive response was attributed to the presence of a PLD2-Rac2 positive feedback loop, while the subsequent negative response of Rac2 on PLD2 was confirmed using cells from Rac2−/− mice that exhibited increased PLD2 enzymatic activity, which was reversed by PIP2. It has been hypothesized that this Rac2-mediated inhibition of PLD2 function occurs because of Rac2 sterical interference with the PH domain membrane-binding site of PLD2 and ensuing PIP2 deprivation. Rac2 localized in vivo to the leading edge of leukocyte pseudopodia with PLD2 being physically posterior to this wave of Rac2. Both PLD2 and PA signal to DOCK2, which mediates Rac activation and actin modeling (Nishikimi et al. 2009).
Role of PLD in Leukocyte Cell Adhesion and Migration
Leukocyte adhesion and migration are steps crucial to the antimicrobial and cytotoxic functions of leukocytes. PLD is expressed in monocytes, macrophages, basophils, eosinophils, dendritic cells, lymphocytes and NK cells and a variety of leukemic cells (U937, THP-1, HL-60, and PLD-985) and has been associated with tumor invasion, chemotaxis, adhesion, phagocytosis, degranulation, microbial killing, and leukocyte maturation. PLD is activated in human and murine myeloid-macrophage cell lines following adhesion to various extracellular matrix (ECM) proteins and plastic (Iyer et al. 2006). PLD concentrates at forming phagosomes, which occurs as a result of PA being concomitantly produced (Rossi et al. 1990) and demonstrates that PLD is in fact catalytically active during this process. PLD activation is an early event in neutrophil signal transduction following exposure of adherent cells to GM-CSF and is regulated by tyrosine phosphorylation, which can in turn be inhibited by tyrosine kinase inhibitors.
PLD1 activity is rapidly enhanced following cell adhesion, which serves to regulate the initial stages of neutrophil and macrophage adhesion. If PLD activity is inhibited, then a likewise inhibition in cell adhesion is evidenced. PLD activation plays a vital role in actin cytoskeleton formation, which stimulates the formation of actin stress fibers in cells, and use of lipase-dead mutants suggests this to be a PLD1-mediated process (Powner and Wakelam 2002). Immunofluorescence microscopy of human neutrophils has shown that both PLD isoforms were associated with cell polarity and directionality concomitant with adhesion and F-actin polymerization in response to IL-8 (Gomez-Cambronero 2010). It has been reported elsewhere (Powner and Wakelam 2002) that actin directly binds PLD2 with a concomitant decrease in lipase activity, which can be reversed by Arf1.
Chemokine receptors differentially regulate PLD, and while PLD1 has been implicated in other migration processes besides chemotaxis (rolling, adhesion, and diapedesis), PLD2 is more directly specialized for chemotactic processes (Gomez-Cambronero et al. 2007). PA is a second messenger in neutrophils that transduces signals to the cell interior upon agonist stimulation and results in development of polarized neutrophil morphology with focused distribution of F-actin, which is also partially dependent on basal PI3K activity and interaction with the C-terminal region of DOCK2 (Nishikimi et al. 2009). PLD-produced PA was able to sequester DOCK2 at the leading egde of migrating neutrophils from the cytocol via interaction with a DOCK2 polybasic amino acid cluster (Ser-Lys-Lys-Arg) that contributed to an increase in actin polymerization, which was also dependent on an intact lipase activity (Nishikimi et al. 2009). PIP3 was another cofactor manufactured during this process besides PA, which implicates both phospholipids in chemotactic translocation and stabilization.
Recently, PLD has been implicated as being needed to leukocyte, macrophage, and fibroblast movements. This was demonstrated through use of RNA interference–mediated depletion of PLD1 and PLD2, which resulted in impaired leukocyte adhesion and reduced chemokinesis and chemotaxis toward the chemokine gradient (Gomez-Cambronero 2010; Knapek et al. 2010). Overexpression of either active PLD1 or active PLD2 yielded cell migration capabilities that were elevated well beyond that of chemoattractant only negative controls. The mechanism for this enhancement in lipase activity is complex and involves two different pathways: one pathway is dependent on the lipase activity and signals directly through the product of this reaction, PA, and the other pathway involves protein–protein interactions.
First, PLD-mediated chemotaxis is mediated through extracellular PA, the pleiotropic lipid second messenger derived from PLD hydrolysis, which has been documented to act as a chemoattractant in human neutrophils and dHL-60 cells as membrane-soluble dioleoyl-PA (DOPA) elicited actin polymerization, cell spreading, pseudopodia formation, and chemotaxis (Frondorf et al. 2010). PLD’s involvement is directly implicated in these cell migration processes (1) as PC on the outer leaflet of the plasma membrane can be cleaved by PLD action that is secreted by microorganisms following interaction with a phagocyte and (2) via intracellular PLD-derived PA generated by PLD2. It has previously been shown that extracellular PA stimulates PLD and results in the generation of intracellular PA and ultimately amplifies the original signal. Exogenous PA or PA generated in situ by bacterial PLD (Streptomyces chromofuscus) enters the cell and results in S6K accumulation in vesicle-like cytoplasmic structures (Frondorf et al. 2010).
Additionally, another protein–protein interaction that positively affects PLD-mediated chemotaxis occurs with S6K via the p70 subunit of the ribosomal S6 kinase (p70S6K), which correlates well with immunofluorescent staining of S6K that translocates from perinuclear regions and colocalizes with PLD2 in the cytosol following chemokine stimulation (Gomez-Cambronero 2010). LR5/RAW264.7 macrophages also use a PLD2/S6K-dependent chemotactic pathway that signals through PLD2-Y296, which is already known to be phosphorylated by EGF-R (Knapek et al. 2010) (Fig. 4). Mutation of this tyrosine residue to phenylalanine completely abrogates chemotaxis to basal levels. Overexpression of PLD2 in dHL-60 leukemic cells results in an elevation of S6K activity, phosphorylation of p70S6K and chemokinesis, while both lipase-dead PLD mutants and si-RNA specific for PLD were inhibitory to this type of cell movement. A similar negative effect of the lipase-dead PLD2-K758R mutant on chemotaxis is also evidenced in LR5/RAW264.7 macrophages and through n-butanol treatment of cells.
PLD regulation in cells occurs via two different signaling pathways. One is via growth factors/mitogens, such as EGF, PDGF, insulin, and serum, and implicates tyrosine kinases. This pathway involves interactions with Grb2; Sos; and the kinases EGF-R, JAK3, and Src. The other pathway is via the small GTPases, such as Arf and Rho, and is directly related to chemotaxis, a process in which PLD plays a vital role. Even though the end results of PLD action as related to downstream signaling mechanisms are still currently being elucidated, adhesion and chemotaxis, which are both requisite for the inflammatory actions of leukocytes, are modulated directly by PLD. The functional consequences of receptor activation are not limited to leukocyte movement but also include degranulation, gene transcription, and mitogenic and apoptotic effects and are seen in angiogenesis, organogenesis, inflammation, and tumor development, growth, and metastasis.
- Hammond SM, Jenco JM, Nakashima S, Cadwallader K, Gu Q, Cook S, Nozawa Y, Prestwich GD, Frohman MA, Morris AJ. Characterization of two alternately spliced forms of phospholipase D1. Activation of the purified enzymes by phosphatidylinositol 4,5-bisphosphate, ADP-ribosylation factor, and Rho family monomeric GTP-binding proteins and protein kinase C-alpha. J Biol Chem. 1997;272:3860–8.PubMedCrossRefGoogle Scholar
- Lewis JA, Scott SA, Lavieri R, Buck JR, Selvy PE, Stoops SL, Armstrong MD, Brown HA, Lindsley CW. Design and synthesis of isoform-selective phospholipase D (PLD) inhibitors. Part I: impact of alternative halogenated privileged structures for PLD1 specificity. Bioorg Med Chem Lett. 2009;19:1916–20.PubMedPubMedCentralCrossRefGoogle Scholar
- Nishikimi A, Fukuhara H, Su W, Hongu T, Takasuga S, Mihara H, Cao Q, Sanematsu F, Kanai M, Hasegawa H, Tanaka Y, Shibasaki M, Kanaho Y, Sasaki T, Frohman MA, Fukui Y. Sequential regulation of DOCK2 dynamics by two phospholipids during neutrophil chemotaxis. Science. 2009;324:384–7.PubMedPubMedCentralCrossRefGoogle Scholar
- Peng HJ, Henkels KM, Mahankali M, Bubulya P, Dinauer MC, Gomez-Cambronero J. The dual effect of Rac2 on PLD2 regulation that explains both onset and termination of chemotaxis. Mol Cell Biol. 2011. doi: 10.1128/MCB.01348-10.