Phosphatidylinositol (PI) and phosphatidylinositol phosphates (PIPs)
Phosphatidylinositol phosphates (PIPs) possess an inositol ring facing the cytosol that can be phosphorylated and dephosphorylated in position 3, 4 and 5 by appropriate kinases or phosphatases. This property can give rise up to seven PIP species including phosphatidylinositol monophosphate (PI(3)P, PI(4)P, PI(5)P), phosphatidylinositol biphosphate (PI(3,4)P2, PI(3,5)P2, PI(4,5)P2) and phosphatidylinositol triphosphate PI(3,4,5)P3. PIPs derive from phosphatidylinositol (PI) that is generated facing the cytosol in the endoplasmic reticulum (ER) for de novo synthesis by phosphatidylinositol synthase (PIS) from CDP-diacylglycerol and L-myo-inositol or by sac1 phosphatase from PI(4)P, in yeast and mammals (A) (Figure 7A) (Bochud and Conzelmann, 2015). The PIS enzyme localizes in the ER and in an ER-derived highly mobile “organelle” that may serve as a dynamic PI distribution device to several organelles (Kim et al., 2011). PI is then distributed throughout the cell presumably by several PI transfer proteins (PITPs) and possibly via vesicular trafficking (Figure 7B). In particular, PI is extracted from the ER at membrane contact sites by PI transfer protein such as secretory protein 14 (sec14) or Pyk2 N-terminal domain-interacting receptor 2 (Nir2) to reach trans-golgi network (TGN) or the plasma membrane, respectively (Bankaitis et al., 2010; Kim et al., 2015). Sec14 exchanges PI from the ER to trans-golgi network and in counterpart exchanges PC located at the TGN to the ER, while Nir2 exchange PI from the ER to the plasma membrane and phosphatidic acid (PA) in the way back (PM -> ER)12,13 (Figure 7C; Bankaitis et al., 2010; Kim et al., 2015). PIPs can be transported through the cellular membrane by regular trafficking such as endocytosis or exocytosis (Balla, 2013). Depending on the location and enzymatic specificity for a given PIP.
Anionic lipids are required to maintain the plasma membrane electrostatic field
The first study to analyze the role of anionic lipids in MSC-establishment was published by the group of Sergio Grinstein in 2006 (Yeung et al., 2006). In this seminal paper, Yeung et al., described and validated the first set of MSC-reporters and showed that the cytosolic leaflet of the PM is highly electrostatics. They then perturbed anionic phospholipid pools using pharmacological approaches to determine their roles for the generation of the PM electrostatic field.
Ionomycin elevates cytosolic calcium, which induces PI(4,5)P2 hydrolysis through activation of phospholipase C (PLC). At the same time, the increase of cytosolic calcium activates lipid scramblase, which results in the translocation of PS from the inner (cytosolic facing) leaflet of the PM to the outer leaflet. Therefore, ionomycin induces the concomitant loss of PI(4,5)P2 and PS in the inner PM leaflet. Ionomycin treatment delocalized the cationic probes (K-pre, Krphy, K-myr), while membrane integrity sensors (GPI, GT46 and Palmitoylation) were not affected (Figure 11A). A more recent study (Ma et al., 2017) using the MCS+ based-FRET sensor verified this observation. Indeed, the concomitant depletion of PI(4,5)P2 and PS following ionomycin treatment induced a decrease of the FRET signal Figure 11B). Dibucaine promotes PS flipping from the inner to the outer leaflet independently of PIPs metabolism and induced a delocalization of cationic probes into the cytosol and a decrease FRET-ratio of the MSC+ probe. In addition, the drug fendiline sollubilizes PS and impacts PM MSC (Ma et al., 2017) (Figure 11A-B). Taken together, these results suggest that a decrease of PI(4,5)P2 and PS concentration at the PM inner leaflet affects membrane surface charge.
Involvement of PIPs in the plasma membrane surface charge
The drugs mentioned above are expected to have pleiotropic effects. It is therefore impossible to exclude that they might induce a large-scale remodeling of global cell physiology and membrane lipids thereby affecting the localization of MSC sensors. In parallele to pharmacological approaches, a number of genetic strategies were implemented to modify anionic lipid pools. Most of these tools are based on the targeting of lipid phosphatase activities at the PM. For example, overexpression of Inp54p, a 5-specific phosphatidylinositol 4,5-bisphosphate phosphatase (5-phosphatase) induced a significant reduction in the FRET efficiency of the MCS+ reporter. However, in this case, Inp54 is constitutively overexpressed, leading to chronic PI(4,5)P2 depletion. Such chronic depletion may also have side effects and therefore it is difficult to pin point the change in the FRET efficiency of the MCS+ probe to the sole depletion of PI(4,5)P2 (Ma et al., 2017).
To overcome these drawbacks, (Heo et al. 2006), designed an elegant method based on the inducible recruitment of phosphoinositide phosphatases at the PM. This inducible phosphatase recruitment is built on genetically encoded PM-localized FK506-binding protein (FKBP12)-rapamycin-binding (FRB) construct and a cytosolic inositol polyphosphate 5-phosphase (Inp54p) enzyme conjugated with FKBP12 (CF-Inp; Figure 12). Rapamycin treatment induces CF-Inp translocation from the cytosol to the PM by chemical heterodimerization that triggers Inp54p activity at the PM and thereby inducible and rapid (i.e. minutes) depletion of PI(4,5)P2 specifically to this membrane. By contrast to chronic depletion of PI(4,5)P2, rapamycin-induced PM-Inp54p recruitment did not affect cationic probes localization (MARCKS-ED, Rin and Rit tail; Figure 13A-B). This result suggests that PI(4,5)P2 is not required, by itself, for MSC and that another anionic phospholipid(s) might act redundantly with PI(4,5)P2 to control PM MSC.
A limitation of the approach from Heo et al., was that dephosphorylation of PI(4,5)P2 by Inp54p induces the production of PI4P, another PM phosphoinositide which is also anionic, albeit to a lesses extent (roughly -3 vs -5 net charges, for PI4P and PI(4,5)P2, respectively). This could explain the absence of effect of PI(4,5)P2 depletion on cationic probes, since only part of the charges carried by PI(4,5)P2 are depleted with this technique. Hammond et al., in 2012 found that PM PI4P is not only a precursor of PI(4,5)P2 biosynthesis but also an important regulator of PM identity. Gerry Hammond indeed built a rapamycin–triggered system, which allows the inducible recruitment at the PM of a chimeric synthetic enzyme, composed of 4- and 5-phosphatase catalytic activities. He named this enzyme “pseudojanine” by analogy to the protein synaptojanin, which naturaly carries 4- and 5-phosphatase catalytic activities. The 4- and 5-phosphatases catalytic domains of pseudojanin came from the yeast Sac1p and human INPP5E proteins, respectively. Point mutations within the Sac1p or INPP5E catalytic domains shut down either the 4-phosphatase or 5-phosphatase activity or both and are used as a control. Depending on the mutations, this system allows altering either PI4P, PI(4,5)P2, both, or none (when both phosphatase domains are mutated). Using an optimized immunolocalization protocol (Hammond et al., 2009), they validated their system and showed that the rapamycin-inducible PM-targeted 5-phosphatase had no effect on PI4P, but depleted PI(4,5)P2 from the PM. Similarly, decreasing the PM-PI4P pool by PM-targeted 4-phosphatase had no effect on the PM PI(4,5)P2 abundance. This unexpected observation suggested a relative independence of the PI4P and PI(4,5)P2 pools at the PM, even though PI(4,5)P2 is made from PI4P. Depletion of either PI4P or PI(4,5)P2 had no effect on the PM targeting of various MSC probes (including Kras-tail, MARCKS-ED, Rit tail and the KA1 domain; Figure 13A-B). Therefore, PI4P and PI(4,5)P2 are not required (on their own) to maintain PM surface charge. However, the concomitent depletion of both PI4P and PI(4,5)P2 altered the PM localization of MSC reporters but did not affect membrane integrity.
Anionic lipids in the maintenance of the PM electrostatics field in yeast
By contrast to mammals, yeast has a single PS synthase gene, called cho1. Cho1p has a different catalytic activity than the mamalian PSS1/PSS2 enzymes as it produces PS via a CDP-diacylglycerol:l-serine O-phosphatidyltransferase activity (Figure 15A). In laboratory conditions, when grown in rich medium, Cho1p is not critical for yeast viability. However, biochemically, the cho1 mutant does not contain PS and the PS sensor C2LACT becomes cytosolic when expressed in cho1. While the catalytic activity of Cho1p is different than PSS1/PSS2, there are extensive paralellism between PS synthesis in yeast and animals: PS is produced in the ER and then transferred to the PM at membrane contact sites by evolutionary conserved proteins19,20 (Maeda et al., 2013; Moser von Filseck et al., 2015). However, the site of PS subcellular accumulation are different in yeast and animals. In mammals, C2LACT localizes at the PM but also in PM-derived organelles along the endocytic pathway. In yeast however, C2LACT is exclusively localized at the PM and virtualy no intracellular compartments are labelled by this probe (Yeung et al., 2008; Moravcevic et al., 2010; Fairn et al., 2011b; Filseck et al., 2015; Maeda et al., 2013) (Figure 15B). This suggests that by contrast to animal cells, PS is predominantly accumulated at the PM in yeast, and therefore might have a predominant role for PM electrostatics.
The K-Ras based MSC probe is not restricted to the PM in S. cerevisiae. This prevented Yeung et al., to obtain direct comparison of MSC reporter localization in WT vs cho1 mutant. However, in 2010, Moravcevic et al. identified a new MSC reporter by characterizing the Kinase associated 1 (KA1) domain. KA1 domains have been identified in both yeast and mammalian proteins involved in kinases regulation. Biochemical assay, crystallography and in vivo experiments define the KA1 domain as a membrane-associated domain that binds all acidic phospholipids, regardless of their respective head group. Similarly, to the PS probe C2LACT, KA1 domains localized strictly to the plasmalemma in yeast. Next, the authors addressed, whether PS, PIP or both participate in KA1 domains localization. In cho1, KA1 lost its specific PM localization (i.e. it became soluble and associated with a much broader membrane domain including both PM and intracellular compartments; Figure 15C). This results suggested a major contribution of PS in setting up PM surface charges in yeast. Thermo-sensitive mutations in the genes encoding the PI4-kinases that generate PI4P at the plasma membrane (Stt4p) and Golgi (Pik1p) deplete PI4P from these mutants at restrictive temperature. In the other hand, thermo-sensitive mutation in the only gene that codes for PI4P-5-kinase (Mss4p) inhibits PI(4,5)P2 production. Mss4p mutant is depleted of PI(4,5)P2 but not PI4P at restrictive temperature, while the double mutant stt4p;pik1p (that lack a PI4-kinase), lack both PI4P and PI(4,5)P2. Surprisingly, KA1 domains remains strictly localized at the PM in all these yeast mutant strains (Figure 15D), suggesting that unlike in animals, PIPs do not play a major role in PM MSC and that PS is the major anionic lipids of the yeast plasmalemma inner leaflet (importantly, the yeast S. cerevisiae does not produce any PI(3,4)P2 and PI(3,4,5)P3, making PI4P and PI(4,5)P2 the only phosphoinositide at the cell surface). Altogether, these results suggest either no or minor role of PIPs in plasmalemma surface charge, while PS is the main anionic lipid driving the PM electrostatic potential in yeast.
High performance thin layer chromatography (HPTLC)
Lipids were deposited on HPTLC plates (Silica gel 60G F₂₅₄ glass plates Merck Millipore) together with external pure lipid standards (Avanti lipids). Plates were developed according to Heape et al (1985). Following chromatography, the lipids were charred for densitometry according to Macala et al. (1983). Briefly, plates were dipped into a 3% cupric acetate (w/v)-8% orthophosphoric acid (v/v) solution in H2O and heated at 110°C for 30min. Plates were scanned at 366 nm using a CAMAG TLC scanner 3. 8 independent samples were quantified for pss1-3 and pss1-4 and 6 samples for col0.
For the analysis of phospholipids by LC-MS/MS, phospholipid extracts were dissolved in 100 µL of eluent A (isopropanol/methanol/water 5/1/4 + 0.2% formic acid + 0.028% NH3) containing synthetic internal lipid standards (PS 17:0/17:0; PE 17:0/17:0; PI 17:0/14:1 and PC 17:0/14:1 from Avanti Polar Lipids). LC-MS/MS (multiple reaction monitoring mode) analyses were performed with a model QTRAP 5500 (ABSciex) mass spectrometer coupled to a liquid chromatography system (Ultimate 3000; Dionex).
Analyses were performed in the negative (PS, PS, PI) and positive (PC) modes with fast polarity switching (50 ms); nitrogen was used for the curtain gas (set to 15), gas 1 (set to 20), and gas 2 (set to 0). Needle voltage was at -4500 or +5500 V without needle heating; the declustering potential was adjusted between -160 and -85 V or set at +40 V. The collision gas was also nitrogen; collision energy varied from -48 to -62 eV and +47 eV on a compound-dependent basis. Reverse-phase separations were performed at 50°C on a Luna C8 150×1 mm column with 100-Å pore size and 5-µm particles (Phenomenex). The gradient elution program was as follows: 0min, 30%B (isopropanol + 0.2%formic acid + 0.028%NH3); 5min, 50% B; 30 min, 80% B; 31 to 41 min, 95% B. The flow rate was set at 40 mL/min, and 3mL sample volumes were injected. The areas of LC peaks were determined using MultiQuant software (version 2.1; ABSciex) for relative phospholipid quantification. Quantification of molecular phospholipids species were performed on five independent samples for pss1-3 and pss1-4 and ten independent samples for Col0.
QUANTIFICATION AND STATISTICAL ANALYSES.
Quantitative co-localization results were statistically compared using the Kruskal-Wallis bilateral test (p-value=0.05) using XLstat software (http://www.xlstat.com/). Pairwise comparisons between groups were performed according to Steel-Dwass-Critchlow-Fligner procedure (different letters indicate statistical difference between samples) (Hollander and Wolfe, 1999). For quantitative co-localization results of Figure 5E we used the bilateral test Kruskal-Wallis test (p-value=0.15) using XLstat software (http://www.xlstat.com/). Pairwise comparisons between groups was performed according to Dunn procedure (different letters indicate statistical difference between samples). Statistical analyses between two samples were performed using the non-parametric Wilcoxon-Mann-Whitney test (p-value=0.05).
Table of contents :
I. General introduction on the electrostatic field
a. Definition of the electrostatic field
b. The membrane surface charge (MSC)
ii. Tools to investigate PM surface charge
c. The plasma membrane is the most ionic compartment in eukaryotes
d. Anionic phospholipids present in the inner leaflet of cellular membrane
i. Phosphatidylinositol phosphates, phosphatidic acid and phosphatidylserine are anionic lipids presenting different charges and concentration in cellular membrane
ii. Regulation, turnover and localization of anionic phospholipids in yeast and mammals
II. Anionic lipids in the maintenance of the PM electrostatic field in mammals
a. Anionic lipids are required to maintain the plasma membrane electrostatic field
b. Involvement of PIPs in the plasma membrane surface charge
c. Evidence that PS contributes to the PM electrostatic field
III. Anionic lipids in the maintenance of the PM electrostatics field in yeast
IV. Membrane surface charges defines an electrostatic territory corresponding to PM-derived organelles
a. In mammals, endocytic compartments are electrostatic.
b. Membrane electrostatic of intracellular compartments in yeast
V. Why electrostatism?
a. Tuning the electrostatic field by adjusting the spatio-temporal control of anionic phospholipid enrichment
i. Example 1: Anionic lipid remodeling controls protein dynamics during phagocytosis
ii. Example 2: Phosphatidylserine generates a charge gradient along the plasma membrane to coordinates proper polarity in yeast
b. Modulation of protein cationic regions to drive its own localization
i. Encoding cationic regions with a variety of charges
ii. Postranslational modification on proteins and the electrostatic switch hypothesis
c. The ion shielding effect links membrane potential and membrane electrostatics and localy organize the plasma membrane electrostatic field
VI. The limits of the electrostatic framework
a. Membrane targetting: a combination of electrostatic, hydrophobic interactions and trafficking
b. Polybasic regions may have specificity for lipid head group
VII. General conclusion and problematic
I. Chapter I: A PtdIns(4)P-driven electrostatic field controls cell membrane identity and signalling in plants
II. Chapter II: Combinatorial PA, PS and PI4P cooperation determines the electrostatic territory in plants
III. Chapter III: A tunable lipid rheostat steers Rho-mediated auxin signaling
a. General introduction on Rho of plant GTPases.
b. ROPGTPase as a central regulator of the ‘non-genomic’ auxin signaling pathway155
i. Auxin influence the interdigitating status of leaf pavement cells
ii. Auxin activates ROPs to orchestrate cytoskeletal rearrangement required to establish pavement cell polarity.
iii. TRANSMEMBRANE RECEPTOR KINASEs (TMKs) acts upstream of ROP signaling in the “non-genomic” auxin signaling pathway
iv. “Non-genomic” auxin signaling pathway regulates the gravitropic response through ROPGTPase
v. Auxin controls microtubule organization during the gravitropic response, which may impact differential cell elongation
vi. Activated ROPs are localized in the so called detergent resistant membrane (DRM) to trigger proper signaling
c. A tunable lipid rheostat steers Rho-mediated auxin signaling
C. DISCUSSION AND PERSPECTIVES
I. The membrane electrostatic field
a. Maintenance of the electrostatic field
i. By phosphatidylinositol 4-phosphate
ii. By phosphatidylserine
iii. By phosphatidic acid
b. Why does a cooperation of three anionic lipids sustain the plasma membrane electrostatic field in plants?
c. PS and PI(4)P sustain the gradient of electrostatic all along the endocytic pathway
II. ROPs nanoclustering
a. How are PS platforms up set up?
b. How is ROPs nanoclustering regulated?
c. Why are nanoclusters required for signaling?