Introduction

It is a well-established fact that the inner and outer leaflets of cellular and organellar membranes show differences in lipid composition, which is popularly called lipid asymmetry. In eukaryotic cell membranes, sphingomyelin (SM) and phosphatidylcholine (PC) along with other choline-containing headgroups are predominantly located in the outer or exoplasmic leaflet of the bilayer membrane. Phosphatidylethanolamine (PE) and phosphatidylserine (PS), the amine-containing phospholipids, on the other hand, are largely confined to the inner or cytoplasmic leaflet, observed in numerous studies on red blood cells and other cell types (Bretscher 1972; Verkleji et al. 1973; Op den Kamp 1981; Devaux 1991; Williamson and Schlegel 1994; Devaux and Morris 2004; Son and London 2013; Shin and Takatsu 2019).

Two energy-dependent processes involving the enzymes—aminophospholipid translocase and ATP-dependent floppase, are found to be responsible for the maintenance of this asymmetric transbilayer distribution. Inhibition of either or both of these enzymes does not destroy the non-random distribution immediately. It rather takes several days in vitro for the complete loss of membrane asymmetry (Zwaal and Schroit 1997). There is also a third, Ca2+-dependent enzyme—lipid scramblase, which mediates rapid transbilayer mixing of phospholipids across the membrane bilayer (Connors et al. 1992; Smeets et al. 1994). In an important MD simulation work, it has been shown that membrane electrostatics could affect asymmetric distribution in lipid membranes composed of zwitterionic PC or PE and the anionic PS resulting in a nonzero potential difference between the two leaflets (Gurtovenko and Vattulainen 2008).

Available literature on the role of other proteins, such as those of the membrane skeletal network in maintaining asymmetric distribution of PS and PE, is inconclusive and provides contradictory evidences. However, this is still noteworthy and perhaps unresolved since skeletal proteins like spectrin and band 4.1 are known to specifically bind to aminophospholipids (Haest et al. 1983; Dressler et al. 1984; Mohandas et al. 1985; Pradhan et al. 1991; Salomao et. al. 2008; Basu and Chakrabarti 2015). The asymmetric transbilayer distribution of phospholipids is crucial to the functioning of proteins embedded in or associated with the membrane and plays a key role in various physiological functions such as cell signalling, transport across membranes, cell–cell interactions and cellular adhesion. Loss of asymmetry, particularly the appearance of PS on the cell surface, is strongly associated with many physiological and pathological conditions (Zwaal and Schroit 1997; Balasubramanian and Schroit 2003; Fadeel and Xue 2009; Tan et al. 2017).

It is also worth noting that the existing literature on various aspects of membrane lipid asymmetry is heavily biased in favour of PS. Loss of PS asymmetry and appearance of this anionic lipid on the cell surface have received enormous attention, whereas the neutral aminophospholipid PE with a smaller headgroup size and a natural propensity to reside in the inner cytoplasmic leaflet, is much less well understood and has drawn little attention. The preferential localization of PE in the inner leaflet of the highly curved membranes of small unilamellar vesicles (SUV), microparticles from malignant cells, apical plasma membranes and red cell spicules in sickle cell disease have been demonstrated (Nordlund et al. 1981; Larson et al. 2012; Julien et al. 1993; Choe et al. 1986). The importance of the biological effects of PE distribution is seen in the membrane associations of protein kinase C and other proteins, which show selectivity for membranes containing PE over PS (Bazzi et al. 1992). Variation of the PS:PE ratio shows that membranes containing about 20% PS:60% PE provide optimum conditions for binding and are as effective as membranes composed of 100% PS (Bazzi et al. 1992; Ray and Chakrabarti 2004). In recent times, PE has also gained importance as a potential chemotherapeutic target due to its higher abundance in cancer cells (Tan et al. 2017).

This commentary aims to concentrate on the status of current knowledge on the membrane asymmetry of PE, which is potentially one of the best candidates for transbilayer movement across cell membranes, guided by lateral associations with other membrane lipids such as cholesterol and long-chain SM molecules (Vance 2008; Steck and Lange 2018). An updated list of literature regarding the status of PE asymmetry in eukaryotic cells is presented in Table 1, along with remarks about their major conclusions.

Table 1 Phosphatidylethanolamine asymmetry in eukaryotic cells
Full size table

Why to Study PE?

The role of lipids as an important structure-forming and/or function-inducing environment for membrane proteins is well known, and PE plays its part in this. PE is asymmetrically distributed like PS in the membrane with about one third in the outer leaflet and two thirds in the inner leaflet. The size of the PE headgroup is one of the smallest among the phospholipids and can form a stable bilayer, particularly in presence of PC, SM and PS. PE with unsaturated fatty acyl chains is also capable of forming non-bilayer hexagonal phases (Gennis 1989a). PE is the second most abundant phospholipid in eukaryotic cell membranes, accounting for about 20% of the total phospholipids (Spector and Yorek 1985; Devaux 1991). PE participates in many important pathophysiological processes (Vance 2008; Calzada et al. 2016). Distribution and differential localization of PE play an important role in cell division, cell death and cytokinesis (Emoto et al. 1996, 1997; Emoto and Umeda 2000). The existence of PE at the cleavage sites during cell division may be attributed to the small headgroup of PE, supporting the seeding of local non-bilayer structures with uncontrolled transbilayer movement. Asymmetric distribution of sarcolemmal PE in neonatal rat cardiomyocytes was found to cause membrane damage after a prolonged period of ischemia (Musters et al. 1993). Many such diverse molecular properties PE, thus, represent a rich target of study among the phospholipids.

PE has been also established as a non-protein molecular chaperone in the folding and maturation process of some proteins in prokaryotes (Bogdanov et al. 1999; Bogdanov and Dowhan 1998, 1999). PE acts as an endogenous cofactor for prion propagation in vitro (Supattapone 2012). All these indicate the potential of PE to play an important role in pathogenic and/or proteopathic disease progression in eukaryotes and could even act as an anti-cancer target (Tan et al. 2017).

Measurement of PS & PE Asymmetry

The rate of phospholipid transbilayer diffusion is slow, taking time from hours to days; thus, steady state measurements are possible for phospholipids—following the order PS > PE ≫ PC > SM. Classical biochemical techniques are still used for the study of lipid asymmetry in membranes (Op den Kamp 1979, 1981; Etemadi 1980; Gennis 1989b).

In a hallmark study, both PE and PS were estimated in platelet plasma membrane using 2,4,6-trinitrobenzenesulphonate (TNBS) which does not penetrate intact cells. Results indicated that PE was partly and PS was completely inaccessible to TNBS in intact platelets (Schick et al. 1976). They concluded that “PS and probably PE are located primarily in the inner lipid bilayer of the platelet plasma membrane”. Schick et al. could estimate a total amount of 26.5% PS and 71% PE in the extracted phospholipids; however, only 6.9% PE of platelet membranes was accessible to TNBS after 30 min. The levels of PS and PE in different eukaryotic membranes are assessed to be around 10% and 20% of the total phospholipids, respectively (Stuart et al. 1998; Spector and Yorek 1985). A sizable fraction of both PS and PE exists in the outer leaflet; amounts were shown to vary among different cell types in the ranges of 0–40% for PS and 0–70% for PE, respectively (Devaux 1991; Zachowski 1993).

Similar observations were made in intact rat liver mitochondria and derived sealed vesicles where 55% and 77% of PC and PE, respectively, were localized in the outer membrane (Hoviusa et al. 1993). However, questions were raised on the accuracy of those measurements (Op den Kamp 1979; Etemadi 1980; Zachowski 1993; Fujimoto and Parmryd 2017). But the divergent results could not be explained in terms of methodological inaccuracy alone, indicating the presence of residual, non-negligible amounts of both PS and PE on the outer leaflet of the plasma membrane in most cell types.

Annexin V (ANV) binding in the presence of Ca2+ has remained the most popular flow cytometry-based method of studying PS asymmetry (Koopman et al. 1994). ANV belongs to the family of proteins that binds with high affinity, almost solely to PS, preferably in the presence of Ca2+. Early work showed that ANV bound to model membranes containing PS:PC (20:80)% with an estimated binding dissociation constant (Kd) of < 10−10 M at physiological concentration of Ca2+ (Tait et al. 1989; Andree et al. 1990). Subsequent studies of binding of ANV conjugates to platelets and erythrocytes produced widely different Kd values ranging from 10−11 to 10−8 M (Thiagarajan and Tait 1990; Yen et al. 2010).

In an earlier work, binding of ANV to phospholipid bilayers adsorbed onto glass beads was found to increase with increasing PS concentrations, only up to 6 mol% PS (Stuart et al. 1998). Calcium concentrations below 3 mM were found to reduce the efficiency of the ANV binding. Interestingly, the addition of 30 mol% PE in the presence of 1–4 mol% PS in the bilayer, significantly increased the maximum binding of ANV over that in the absence of PS, thus, indicating the binding of ANV to PE at lower PS concentrations. Binding of ANV to phospholipid bilayers containing PE has also been reported by others (Meers and Mealy 1994). Validity of ANV as a probe for loss of membrane asymmetry or exposure of PS is thought to be apparently unhampered in the presence of up to 10 mol% PE in the outer membrane (Stuart et al. 1998). However, with a further increase in the PE content of the outer leaflet in combination with even a marginal increase in the PS level, a strong difference in ANV binding could result. This is shown in Fig. 1 as a schematic model.

Fig. 1
figure 1

The schematic diagram of membranes shows the binding of the Annexin V (ANV), to a phospholipid bilayer. In all three panels, the top leaflet represents the extracellular face, and the bottom leaflet represents the intracellular face. Panel a shows the binding of ANV to the PS headgroups in a PC/PS bilayer and Panel b shows the same in the presence of PE, below 10 mol%. Panel c shows the scenario when the content of PE is greater than 10 mol% showing ANV binding to PE along with PS, particularly in the absence of Ca2+ ions

Full size image

Cell surface localization of PE was studied in dividing Chinese hamster ovary cells using cinnamycin (Ro 09-0198), a tetracyclic peptide antibiotic that binds specifically to PE with a 1:1 stoichiometry. PE was found exposed on the cell surface, at the cleavage furrow, only during late telophase. PE distribution was otherwise not found to be altered in the plasma membrane (Emoto et al. 1996). Kd values for Ro 09-0198-PE complex ranged from 10–7 to 10–8 M in liposome membranes (Machaidze and Seelig 2003). It has been shown by a few groups, including ours, which the membrane skeletal protein spectrin binds both PS and PE and possesses a high affinity-binding site for PE at its ankyrin-binding domain (Ray and Chakrabarti 2004; Grzybek et al. 2006; Giri et al. 2017; Bose and Chakrabarti 2019).

Conclusion

Classically phospholipids were only treated as the matrix that provides the right orientation to the more functional and active membrane proteins. Importance started to be attributed to membrane phospholipid asymmetry, after the discovery of its loss as a marker of cellular apoptosis—using Ca2+-dependent high affinity binding of ANV as a tool to assess the level of surface-exposed PS. However, like in other phospholipid-binding proteins, binding of ANV is also found to be associated with PE as well as PS. The size of the PE headgroup, its higher abundance, zwitter ionic nature, natural propensity for faster transbilayer diffusion, and its unique functional properties in maturation of prokaryotic membrane proteins, indicate that adequate attentions are to be paid in studying the asymmetric distribution of PE.

Only a few reports are available till date on the asymmetric distribution of PE alone. It is envisioned that this commentary could help to establish that most of the research related to membrane asymmetry is a bit biased towards PS, when considerable evidence is available in the literature in favour of PE being at least a co-partner of PS in driving the biological functions due to membrane phospholipid asymmetry. As such PE must receive greater attention in the ever-growing literature on membrane phospholipid asymmetry and its pathophysiological outcome.