Journal list menu

Volume 591, Issue 18 p. 2720-2729
Review Article
Free Access

Plasmalogen homeostasis – regulation of plasmalogen biosynthesis and its physiological consequence in mammals

Masanori Honsho

Masanori Honsho

Medical Institute of Bioregulation, Kyushu University, Fukuoka, Japan

Search for more papers by this author
Yukio Fujiki

Corresponding Author

Yukio Fujiki

Medical Institute of Bioregulation, Kyushu University, Fukuoka, Japan

Correspondence

Y. Fujiki, Medical Institute of Bioregulation, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan

Fax: +81 92 642 4233

Tel: +81 92 642 4232

E-mail: [email protected]

Search for more papers by this author
First published: 07 July 2017
Citations: 77
Edited by Wilhelm Just

Data Accessibility:

Research data pertaining to this article is located at figshare.com: https://dx.doi.org/10.6084/m9.figshare.5311087

Abstract

Plasmalogens, mostly ethanolamine-containing alkenyl ether phospholipids, are a major subclass of glycerophospholipids. Plasmalogen synthesis is initiated in peroxisomes and completed in the endoplasmic reticulum. The absence of plasmalogens in several organs of peroxisome biogenesis-defective patients suggests that the de novo synthesis of plasmalogens plays a pivotal role in its homeostasis in tissues. Plasmalogen synthesis is regulated by modulating the stability of fatty acyl-CoA reductase 1 on peroxisomal membranes, a rate-limiting enzyme in plasmalogen synthesis, by sensing plasmalogens in the inner leaflet of plasma membranes. Dysregulation of plasmalogen homeostasis impairs cholesterol biosynthesis by altering the stability of squalene monooxygenase, a key enzyme in cholesterol biosynthesis, implying physiological consequences of plasmalogen homeostasis with respect to cholesterol metabolism in cells, as well as in organs such as the liver.

Abbreviations

24,25EC, 24,25-epoxycholesterol

ABC, ATP-binding cassette

AG, alkylglycerol

AGPS, alkylglycerone phosphate synthase

DHAPAT, dihydroxyacetonephosphate acyltransferase

ER, endoplasmic reticulum

Etn, ethanolamine

FAO, fatty alcohol:NAD oxidoreductase

Far1, fatty acyl-CoA reductase 1

MARCH6, membrane-associated RING finger 6

PlsCho, choline plasmalogen

PlsEtn, ethanolamine plasmalogen

PtdEtn, phosphatidylethanolamine

RCDP, rhizomelic chondrodysplasia punctata

SM, squalene monooxygenase

Phospholipids constitute a critical component of the cellular membranes. In mammalian cells, phosphatidylcholine is the most abundant phospholipid. Ethanolamine (Etn)-containing phospholipid is the second most abundant phospholipid, comprising approximately 15–25% of total phospholipids [1], in which phosphatidylethanolamine (PtdEtn), a diacyl glycerophospholipid, and ethanolamine plasmalogen (PlsEtn), an alkenylacylglycerophospholipid, are major constituents [2]. In mammals, PlsEtn is found in several organs, such as the brain, kidney and heart, in which PlsEtn constitutes approximately 50% of Etn-containing phospholipids, whereas, in the liver, plasmalogens are only slightly detectable [2-4]. Choline plasmalogen (PlsCho) is abundant in both the heart and skeletal muscle [2, 3].

A deficiency of plasmalogen evoked by dysfunction of peroxisomal enzymes essential for the plasmalogen synthesis causes rhizomelic chondrodysplasia punctata (RCDP), a fatal genetic disease. To date, five distinct genotypes of RCDP have been identified [5, 6]. Inherited mutations in the genes encoding the enzymes essential for the synthesis of plasmalogens, including dihydroxyacetonephosphate acyltransferase (DHAPAT), alkylglycerone phosphate synthase (AGPS) and fatty acyl-CoA reductase 1 (Far1), cause RCDP types 2 [7], 3 [8] and 4 [9], respectively, whereas dysfunction of PEX7 and PEX5L, which are indispensable cytosolic factors for the transport of AGPS and other peroxisomal matrix proteins harboring peroxisome targeting signal type 2 [10], causes RCDP types 1 [11-13] and 5 [14], respectively. These patients show a significantly reduced level of plasmalogens in erythrocytes [15-17]. In addition, such a severely reduced level of plasmalogens is also reported in peroxisome biogenesis disorders where the plasmalogen level in organs such as brain, heart and kidney from patients is lower than 10% of that in non-peroxisome biogenesis disorder controls [18]. Similarly, PEX7-knockout mice show a severe decline of PlsEtn in several tissues [19, 20]. Taken together, these studies strongly suggest that de novo synthesis of plasmalogens plays a pivotal role in the homeostasis of tissue plasmalogens.

The biosynthesis of PlsEtn is initiated in peroxisomes and completed in the endoplasmic reticulum (ER) via seven-step reactions [3]. The initial two steps of PlsEtn synthesis are catalyzed by peroxisomal matrix enzymes, DHAPAT and AGPS, where AGPS generates alkyl-DHAP by replacing the acyl chain of acyl-DHAP, the product of DHAPAT, with a long chain fatty alcohol (Fig. 1). Far1, a peroxisomal C-tail anchored protein, is the enzyme responsible for the synthesis of C16:0, C18:0 and C18:1 fatty alcohols [21-23]. Alkyl-DHAP is further reduced and PlsEtn is synthesized via the remaining four steps in the ER. The biosynthetic pathway of PlsEtn has been comprehensively reviewed [2, 3, 24, 25]. In the present review, we summarize current knowledge about the regulation of plasmalogen synthesis and discuss the physiological consequences of plasmalogen homeostasis in health and disease.

Details are in the caption following the image
Schematic view of regulation of PlsEtn synthesis and PlsEtn-mediated regulation of cholesterol synthesis. (A) Structure of PlsEtn. Vinyl-ether bond in the sn-1 position of PlsEtn is indicated by a red dashed box. (B) The initial two steps of PlsEtn synthesis are catalyzed by peroxisomal matrix enzymes, DHAPAT and AGPS (steps 1 and 2), where AGPS catalyzes the formation of alkyl-DHAP by replacing the acyl chain of acyl-DHAP with a long chain alcohol. Far1, a peroxisomal C-tail anchored protein, is the enzyme responsible for the synthesis of long chain alcohols. Alkyl-DHAP is further reduced (step 3) and PlsEtn is synthesized via the remaining four steps in the ER (steps 4–7). PlsEtn is transported to the post-Golgi compartment, including endosomes and plasma membranes in a manner dependent on ATP but not vesicular transport [33] (stage a). PlsEtn is preferentially localized in the inner leaflet of plasma membrane in a P-type ATPase-dependent manner. The biosynthesis of PlsEtn is regulated by modulating the stability of Far1 in a manner dependent on the total cellular level of PlsEtn. PlsEtn localized in the inner leaflet of plasma membrane is sensed (stage b) and the signal indicating the cellular PlsEtn level is conveyed to peroxisomes (stage c), inducing the regulation of the stability of Far1 [32] (stage d). Biosynthesis of cholesterol is regulated by the homeostasis of PlsEtn, where PlsEtn regulates the stability of SM, the proposed second rate-limiting enzyme in cholesterol synthesis (stage e). A detailed mechanism of the PlsEtn-mediated regulation of cholesterol synthesis is provided in Fig. 2.

Homeostasis of PlsEtn in cells

The level of PlsEtn in plasmalogen-deficient mutant cells is lower than 10% of that in wild-type cells [3], implying that de novo synthesis of PlsEtn contributes greatly to the homeostasis of PlsEtn in cells. The finding that de novo synthesis of PlsEtn in CHO-K1 cells is reduced by the elevation of cellular PlsEtn but not PtdEtn and PlsCho indicates that the synthesis of PlsEtn is regulated by a feedback mechanism as a result of sensing PlsEtn and/or its metabolites in CHO-K1 cells (Fig. 1) [22].

The finding of increased activity of fatty acyl-CoA reductase in plasmalogen-deficient cells [26, 27] suggests that fatty acyl-CoA reductase is a potential rate-limiting enzyme in plasmalogen synthesis. Indeed, the elevated activity of Far1 in plasmalogen-deficient cells is reduced to normal by restoring the cellular level of plasmalogens [22]. Conversely, the activity of DHAPAT, AGPS and acyl/alkyl-DHAP reductase, which are responsible for the initial three steps in plasmalogen biosynthesis, is not elevated in plasmalogen-deficient cells [28-30]. The activity of Far1 is not regulated at the transcription level; rather, it is post-translationally regulated by modulating the stability of Far1 in response to the amount of cellular plasmalogens [22, 23] but not hexadecanol [31]. Degradation of Far1 is inhibited in plasmalogen-deficient cells, whereas elevation of cellular PlsEtn enhances the degradation of Far1, as reported in several cell lines such as CHO-K1, HeLa and HepG2 [22, 32], thereby reducing the synthesis of PlsEtn. Collectively, such experimental evidence suggests that modulation of the stability of the rate-limiting enzyme Far1 as a result of sensing the cellular level of plasmalogens regulates the synthesis of PlsEtn (Fig. 1).

Two acyl-CoA reductases, Far1 and Far2, have been identified [21]. The amino acid sequence of human Far2 has 59% identity and 78% similarity to that of human Far1. Far1 and Far2 both target peroxisomes in a manner dependent on Pex19p, a cytosolic receptor for peroxisomal membrane proteins, and localize to peroxisomes as a C-tail anchored protein [23]. Unlike Far1, Far2 is not degraded upon restoring the plasmalogen level in plasmalogen-deficient CHO mutant and MCF7 cells [23]. Domain swapping of the two isoforms suggests that the transmembrane-flanking region of Far1 plays a major role in its degradation [23]. Far1 is expressed widely in murine tissues [21]. A similar distribution in human tissues is suggested by analyzing databases such as the Human Protein Atlas [25]. By contrast, FAR2 mRNA is largely restricted to eyelid containing wax-laden meibomian glands, skin and brain in mouse [21], suggesting that Far1 plays a major role in the generation of fatty alcohols for the synthesis of plasmalogens. The severely reduced level of plasmalogens in red blood cells from FAR1-deficient patients further supports the suggestion that the fatty alcohols required for the synthesis of plasmalogens are mainly provided by Far1 [9] and that Far2 appears to contribute less to the reduction of acyl-CoA with respect to the synthesis of plasmalogens.

Mechanisms of the regulation of plasmalogen synthesis

Sensing of plasmalogens is a crucial step for the regulation of plasmalogen synthesis. Although Far1 is localized to peroxisomes where the plasmalogen synthesis is initiated, the sensing of plasmalogens is unlikely on peroxisomes because PlsEtn is below a detectable level in peroxisomes [33-35]. Recent studies report that PlsEtn located in the inner leaflet of plasma membranes is most likely sensed (Fig. 1) [32]. Interestingly, two distinct inhibitors including dynasore and myristyl trimethyl ammonium bromide of dynamin, a GTPase required for membrane scission during endocytosis, and knockdown of flotillin1, a protein enriched in lipid rafts of the plasma membrane [36, 37], enhance the degradation of Far1, thereby reducing plasmalogen synthesis [32]. In flotillin1-knocked down HeLa cells, the amount of PlsEtn in the Triton X-100-insoluble fraction is augmented, probably as a result of the inhibition of the endocytosis of PlsEtn, whereas the total cellular level of PlsEtn is not increased. Therefore, PlsEtn enriched in the inner leaflet of plasma membranes [32, 38, 39] is most likely sensed, followed by accelerated degradation of Far1 [32]. This asymmetric distribution of PlsEtn mediated by a flippase belonging to the P4 subfamily of P-type ATPases is also important for sensing PlsEtn [32]. Therefore, the biosynthesis of PlsEtn is spatiotemporally regulated by sensing PlsEtn in the inner leaflet of plasma membranes. This is followed by a signal conveying infromation that arises from sensing the plasmalogen level to peroxisomes and modulation of the stability of Far1.

The elevation of PlsEtn in the Triton X-100-insoluble fraction upon knocking down of flotillin1 suggests that lipid rafts, a cholesterol and sphingomyelin enriched domain of the plasma membrane, might be important for sensing PlsEtn [32] because flotillin1 is shown to be enriched in lipid rafts of the plasma membranes [36, 37]. Indeed, PlsEtn is enriched in rafts [33, 40] and nystatin, a cholesterol-chelating agent, inducing distortion of the structure and function of cholesterol-rich membrane domain [41], inhibits the degradation of Far1 without reducing the cholesterol level. To clarify whether PlsEtn located in the inner leaflet of plasma membrane lipid rafts is sensed, the more precise manipulation of the membrane property in lipid rafts and the identification of a PlsEtn sensor are required. The molecular mechanism underlying plasmalogen-mediated degradation of Far1 on peroxisomes remains unknown. Peroxisomal matrix proteinases such as trypsin domain-containing 1 [42, 43] and peroxisomal Lon protease [43, 44] are less likely involved in the PlsEtn-mediated degradation of Far1 because Far1, a tail-anchored membrane protein, is degraded upon the restoration of PlsEtn in PEX1-defective CHO cells [22]. Hence, several issues, including the sensing of PlsEtn, transfer of the signal indicating the plasmalogen level and the modulation of Far1 stability on peroxisomes, remain to be addressed to clarify the molecular mechanisms of the spatiotemporal regulation of plasmalogen synthesis.

Regulation of plasmalogen synthesis in tissue

The regulation of plasmalogen synthesis in tissues has not been addressed. However, data from the administration of alkylglycerol (AG) to the PEX7-knockout mouse and wild-type rat provide some idea how the plasmalogen synthesis is regulated in animals [20, 45]. AG restores the synthesis of PlsEtn in cultured cells by bypassing the first three steps of plasmalogen synthesis [3, 33, 46] and induces the degradation of Far1 [22, 23]. Feeding AG (i.e. 1-O-octadecyl-sn-glycerol) to the PEX7-knockout mouse successfully elevates the level of plasmalogen containing C18:0 fatty alcohol in its peripheral tissues to levels in the AG-fed wild-type mouse [20]. Similarly, the administration of 1-O-heptadecyl-sn-glycerol, an uncommon ether lipid, to young rats results in its incorporation to a high extent into PlsEtn in all tissues except the brain. However, the total plasmalogen content of the peripheral tissues is not increased, which is most likely a result of the compensatory reduction of endogenous plasmalogens harboring C16:0, C18:0 and C18:1 fatty alcohols [47]. Collectively, these studies suggest that there is some mechanism regulating the synthesis of PlsEtn by sensing cellular plasmalogens in most peripheral tissues.

However, it remains unknown how the synthesis of PlsEtn is regulated in the brain where PlsEtn is highly abundant. Feeding AG to the PEX7-knockout mouse does not increase PlsEtn in the brain to the levels found in a normal mouse, despite the elevation of PlsEtn in peripheral tissues [20]. The plasmalogen level in the brain is severely reduced in mutant mice with a brain-specific deficiency of PEX13 and PEX5 [48, 49]. Conversely, brain-selective restoration of peroxisomes in the PEX5-knockout mouse complements the normal plasmalogen level in the brain [50]. These studies may indicate that plasmalogens essential for the function of brain are more likely supplied by PlsEtn synthesis in the brain rather than by the transport of peripherally synthesized PlsEtn crossing the blood–brain barrier. In the brain, PlsEtn is relatively static and functions as a component of myelin in white matter [51]. By contrast, PlsEtn is abundant in gray matter and actively metabolized with a half-life of 10–30 min [51], suggesting that the synthesis of plasmalogens in gray matter is tightly regulated by sensing the cellular level of plasmalogens.

Currently, the molecular mechanisms underlying the regulation of plasmalogen synthesis in tissues remain unknown. The finding that the Far1 protein level in the kidney is markedly increased in the PEX7-knockout mouse [52] compared to that in the wild-type mouse suggests that plasmalogen synthesis is regulated by modulating the stability of Far1 in the kidney as well, similar to the mechanism reported in cultured cells [22]. Given the findings regarding the regulation of plasmalogen synthesis in cultured cells, a key regulatory factor for plasmalogen synthesis appears to rely on the availability of fatty alcohols that are synthesized by Far or derived from the diets [53]. However, fatty alcohols are also required for the synthesis of wax esters [54] and excess fatty alcohol is oxidized to fatty acids by the fatty alcohol:NAD oxidoreductase (FAO) complex [25, 55, 56] that constitutes the fatty alcohol cycle together with Far. Interestingly, the brain contains higher amounts of plasmalogens and, correspondingly, shows a higher activity of Far, whereas FAO activity in brain is low [57]. By contrast, in the liver [57], plasmalogens are much less abundant [2, 58] and Far activity is extremely low [59], whereas FAO activity is high [58] and thereby excess fatty alcohols are rapidly oxidized to fatty acids. However, it remains unknown whether plasmalogen synthesis in tissues with a lower level of plasmalogens is likewise regulated by the fatty alcohols generated by Far1 in a manner dependent on the level of plasmalogens. In this context, the fact that the elevated level of fatty alcohols in keratinocytes from a patient deficient in the FAO activity does not significantly increase the plasmalogen synthesis [54] may suggest an alternative mechanism for the regulation of plasmalogen synthesis. Similarly, the regulation of plasmalogen synthesis in the liver has not yet been defined. Cellular levels of plasmalogens and Far activity are low in the liver [59], whereas PlsEtn is efficiently exported with lipoproteins [60]. Interestingly, PlsEtn secretion from keratinocytes is also reported [61]. Based on these studies, it is tempting to speculate that the synthesis of PlsEtn in the liver and keratinocytes is regulated by a distinct mechanism devoid of any sensing of plasmalogens.

Regarding the homeostasis of PlsCho, the regulatory mechanism of PlsCho synthesis remains unknown. PlsCho is considered to be synthesized as a result of the hydrolysis of PlsEtn and subsequent modification of the resultant alkenylglycerol with phosphocholine catalyzed by choline phosphotransferase [2, 3, 62, 63]. PlsCho is highly abundant in the heart and muscle, where PlsCho comprises approximately 50% of plasmalogens [2]. Feeding 1-O-octadecyl-sn-glycerol to the wild-type mouse increases the total PlsEtn level in the heart with a reduced level of PlsEtn containing C16:0 fatty alcohol. However, a reduction of PlsCho containing C16:0 fatty alcohol is not evident despite elevation of PlsCho containing C18:0 fatty alcohol [64]. Given such evidence, the synthesis of PlsEtn and PlsCho is most likely regulated by a mutually distinct mechanism in the heart.

Physiological consequences of plasmalogen homeostasis

Several physiological roles of plasmalogens are suggested from analyses of defects in plasmalogen-deficient or -reduced cells [30, 65-71] and animals [19, 61, 72-78]. Recent reviews on the functions of plasmalogens are also recommended [2, 5, 62, 79-82].

Restoration of plasmalogens in the PEX7-knockout mouse by feeding AG stops the progression of pathogenesis in testis, adipose tissue and the harderian gland [20]. Similarly, complementation of PlsEtn with AG by culturing CHO cell mutants defective in peroxisome assembly and plasmalogen-deficient murine cells derived from RAW 264.7 recovers the impaired resistance against reactive oxygen species [65, 83]. Recent studies using the plasmalogen-deficient mouse and Caenorhabditis elegans are less supportive of the protective function of PlsEtn as a scavenger of reactive oxygen species [84, 85]. Moreover, plasmalogen deficiency causes an impaired high-density lipoprotein-mediated cholesterol efflux in murine macrophage-like cells [86], the delayed transport of internalized cholesterol to the ER [87] and reduced cholesterol synthesis [88]. Importantly, these impairments are complemented by restoring the level of plasmalogens, suggesting that plasmalogens play an important role at multiple steps in cholesterol homeostasis. Our recent findings demonstrate that plasmalogen homeostasis links to cholesterol synthesis by regulating the stability of squalene monooxygenase (SM) [88], the proposed second rate-limiting enzyme in cholesterol synthesis [89] (Fig. 2). Elevation of PlsEtn enhances the degradation of SM by an E3 ubiquitin ligase, membrane-associated RING finger 6 (MARCH6). Conversely, the absence of PlsEtn stabilizes SM by lowering the interaction of SM with MARCH6, resulting in an elevation of the protein level of SM, thereby preferentially synthesizing diepoxysqualene followed by 24,25-epoxycholesterol (24,25EC) in plasmalogen-deficient cells, similar to cells highly expressing SM [90]. Consequently, this metabolic flow reduces the synthesis of monoepoxysqualene utilized for the synthesis of cholesterol.

Details are in the caption following the image
Synthesis of cholesterol in cells where PlsEtn is elevated or deficient. Cholesterol is synthesized from acetyl-CoA via the isoprenoid biosynthetic pathway, followed by conversion of two farnesyl pyrophosphates to squalene. Squalene is oxidized by SM to generate 2,3-epoxysqualene, then the first sterol intermediate, lanosterol, is produced by cyclization of 2,3-epoxysqualene. Lanosterol is then converted to cholesterol through a series additional 19-step reactions. SM oxidizes 2,3-epoxysqualene (monoepoxysqualene) to generate 2,3;22,23-diepoxysqualene (diepoxysqualene), followed by generating epoxylanosterol and its conversion to 24,25EC. Arrows indicate the level of metabolites or expression level of SM. (A) Elevation of cellular PlsEtn stimulates the interaction of SM with MARCH6, an E3 ubiquitin ligase termed the membrane-associated RING finger (C3HC4) 6, thereby inducing degradation of SM, which lowers the synthesis of cholesterol and 24,25EC. (B) Degradation of SM is inhibited by lowering the interaction of SM with MARCH6 in plasmalogen-deficient cells. Under such condition, SM preferentially synthesizes diepoxysqualene and its metabolite, 24,25EC, concomitantly decreasing the synthesis of monoepoxysqualene and its metabolite cholesterol. Activity of HMG-CoA reductase, a rate-limiting enzyme of cholesterol synthesis, is not altered by the level of cellular PlsEtn. Note that cholesterol synthesis is reduced in both PlsEtn-elevated and reduced cases, suggesting that homeostasis of PlsEtn is tightly linked to the synthesis of cholesterol.

The physiological consequence of a plasmalogen-mediated regulation of cholesterol synthesis is not shown. Similar to plasmalogen-deficient cells, plasmalogen levels are very low and they constitute < 5% of the total phospholipids in the liver [2], where the activity of SM is higher than that in non-cholesterogenic tissues [91]. Therefore, it is plausible that the synthesis of diepoxysqualene is preferred in the liver, thereby stimulating the synthesis of 24,25EC. Consequently, 24,25EC activates liver X receptors for which it is a natural ligand [92], followed by transcriptional induction of the expression of target genes, including ATP-binding cassette (ABC) transporters such as ABCA1, ABCG5 and ABCG8 [93, 94]. These transporters play important roles in a reverse cholesterol transport, a complex process ensuring the efflux of cholesterol from peripheral cells and its transport back to liver for metabolism and biliary excretion [95]. Consequently, the lower plasmalogen level in the liver likely possesses the potential to maintain the activity of reverse cholesterol transport as a result of the efficiently elevated synthesis of 24,25EC.

Conclusions

Plasmalogen homeostasis is maintained by a combination of the regulation in biosynthesis and degradation of PlsEtn. In cultured cells, the synthesis of PlsEtn is regulated in a manner dependent on the level of PlsEtn located in the inner leaflet of the plasma membrane. However, the way in which biosynthesis of PlsEtn is controlled in tissues remains unknown. Although PlsEtn synthesis in tissues may be regulated by a feedback mechanism similar to that in cultured cells, other mechanisms, including the regulation of PlsEtn synthesis without sensing PlsEtn, may be considered in PlsEtn-effluxing tissues.

The turnover of PlsEtn needs to be considered for the regulation of PlsEtn homeostasis. PlsEtn is hydrolyzed by plasmalogen-specific phospholipaseA2 [62, 96, 97] and polyunsaturated fatty acids are released. The remaining lysoplasmalogen is either reacylated back to PlsEtn or further hydrolyzed by lysoplasmalogenase, which is abundant in the liver and small intestine [2, 62, 98]. However, the degradation mechanism of PlsEtn in other tissues or cell types remains largely unknown. Further investigations are required to address such issues.

Acknowledgements

We thank K. Shimizu for preparing Figs 1 and 2, as well as the other members of our laboratory for discussions. This work was supported in part by Grants-in-Aid for Scientific Research 23570236, 26440102 and 17K07337 (to MH) and 24247038, 25112518, 25116717, 26116007, 15K14511, 15K21743 and 17H03675 (to YF), as well as grants from the Takeda Science Foundation, the Naito Foundation, the Japan Foundation for Applied Enzymology and the Novartis Foundation (Japan) for the Promotion of Science (to YF).

    Data Accessibility