Metabolism and function of mitochondrial cardiolipin

Progress in Lipid Research 55 (2014) 1–16

Contents lists available at ScienceDirect

Progress in Lipid Research journal homepage: www.elsevier.com/locate/plipres

Review

Metabolism and function of mitochondrial cardiolipin Mindong Ren a,c, Colin K.L. Phoon b, Michael Schlame a,c,⇑ a

Department of Anesthesiology, New York University School of Medicine, New York, USA Department of Pediatrics, New York University School of Medicine, New York, USA c Department of Cell Biology, New York University School of Medicine, New York, USA b

a r t i c l e

i n f o

Article history: Received 4 March 2014 Received in revised form 4 April 2014 Accepted 14 April 2014 Available online 24 April 2014 Keywords: Cardiolipin Disease Fatty acids Membranes Mitochondria Phospholipids

a b s t r a c t Since it has been recognized that mitochondria are crucial not only for energy metabolism but also for other cellular functions, there has been a growing interest in cardiolipin, the specific phospholipid of mitochondrial membranes. Indeed, cardiolipin is a universal component of mitochondria in all eukaryotes. It has a unique dimeric structure comprised of two phosphatidic acid residues linked by a glycerol bridge, which gives rise to unique physicochemical properties. Cardiolipin plays an important role in the structural organization and the function of mitochondrial membranes. In this article, we review the literature on cardiolipin biology, focusing on the most important discoveries of the past decade. Specifically, we describe the formation, the migration, and the degradation of cardiolipin and we discuss how cardiolipin affects mitochondrial function. We also give an overview of the various phenotypes of cardiolipin deficiency in different organisms. Ó 2014 Elsevier Ltd. All rights reserved.

Contents 1. 2.

3.

4.

5.

6.

7.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 The CL pathway of mitochondria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2.1. Supply of PA for CL formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2.2. Conversion of PA to CL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2.3. Remodeling of CL. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Trafficking of mitochondrial CL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 3.1. Intramembrane translocation (flip-flopping) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 3.2. Intermembrane translocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 3.3. Exposure of CL at the surface of mitochondria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 3.4. CL trafficking beyond mitochondria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Degradation of mitochondrial CL and apoptosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 4.1. CL hydrolysis by phospholipases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 4.2. Peroxidation of CL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Function of CL in mitochondrial membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 5.1. Physicochemical properties of CL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 5.2. CL in specific membrane domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 5.3. Interaction of CL with proteins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Models of CL deficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 6.1. Experimental models of CL deficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 6.2. Barth syndrome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 6.3. Abnormalities of CL implicated in other human diseases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Conflict of Interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

⇑ Corresponding author at: Department of Anesthesiology, NYU Langone Medical Center, 550 First Avenue, New York, NY 10016, USA. Tel.: +1 212 263 5072. E-mail address: [email protected] (M. Schlame). http://dx.doi.org/10.1016/j.plipres.2014.04.001 0163-7827/Ó 2014 Elsevier Ltd. All rights reserved.

2

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

1. Introduction Cardiolipin (CL) is a unique phospholipid dimer consisting of two phosphatidyl residues linked by a glycerol bridge, i.e. 1’,3’bis(1,2-diacylglycero-3-phospho-)glycerol. CL occurs in various ATP-producing membranes of prokaryotes and eukaryotes, but we will limit this review to CL of mitochondria because lately, mitochondria have become a central aspect of research in cell biology. Along with the surging interest in mitochondria, there has been an explosion of papers addressing the role of CL. In a previous review on CL in this journal more than a decade ago, most if not all papers published on the subject were included [1]. Today, such an approach would be all but impractical. Instead, we decided to be selective and to include only those references that we felt are essential to the topics we wanted to discuss. As a result, we have omitted many papers, including many of our own, not to ignore them and certainly not to insult the authors but to focus on the key issues in this field that has become rather large. Furthermore, we have made an effort to cite mostly original articles rather than other reviews, except when the multitude and the convoluted nature of the original papers made it difficult to maintain an economic writing style. 2. The CL pathway of mitochondria During the evolution of mitochondria, a large portion of their ancestral genome migrated into the nuclear compartment, which made mitochondria dependent on the import of proteins and

lipids. However, mitochondria have maintained the ability to synthesize CL from its basic building blocks, glycerol-3-phosphate and fatty acids. The CL pathway can be divided into three parts, namely (i) the formation of phosphatidic acid (PA) and its translocation from the outer to the inner membrane, (ii) the conversion of PA to CL on the matrix side of the inner membrane, and (iii) the remodeling of CL, which in yeast takes place in membrane leaflets facing the intermembrane space (Fig. 1, Table 1). Thus, the overall process involves all mitochondrial compartments. 2.1. Supply of PA for CL formation PA, the central precursor for the biosynthesis of neutral glycerolipids and glycerophospholipids, is formed by sequential acylation of glycerol-3-phosphate. In mammals, four proteins have been identified to carry glycerol-3-phosphate acyltransferase activity, two of which are localized in mitochondria (GPAT 1 and 2) and two of which in the endoplasmic reticulum (GPAT 3 and 4) [2]. The product of these enzymes is lyso-phosphatidic acid (LPA) that can be further acylated to PA. In mammalian mitochondria, acylation of LPA is catalyzed by AGPAT 5, a member of a large family of lysophospholipid acyltransferases [3]. The significance of having two branches of PA formation, one in mitochondria and one in the endoplasmic reticulum, has not been established. While the endoplasmic reticulum has traditionally been viewed as the principal site of lipid biosynthesis, mitochondrial glycerol-3-phosphate acylation has been associated with essential functions of its own, such as the synthesis of triglycerides [4] and mitochondrial fusion [5]. The frequently made assumption that it is the mitochondrial branch that supplies PA to the CL pathway has also not been proven. Mitochondrial PA is formed on the outer leaflet of the outer membrane [6], which is localized near the mitochondria-associated endoplasmic reticulum membranes. Thus, the two branches of PA formation are in close proximity to each other. Whatever the true source of PA is, it has to migrate across the outer membrane, the intermembrane space, and the inner membrane in order to reach the compartment where CL biosynthesis takes place. In yeast, Ups1 was identified as the lipid transfer protein that facilitates the transfer of PA across the intermembrane space [7]. 2.2. Conversion of PA to CL

Fig. 1. Overview of CL biosynthesis in mitochondria. Glycero-3-phosphate (G3P) is acylated to lysophosphatidic acid (LPA) and then to phosphatidic acid (PA). PA formation occurs on the outer face of the outer mitochondrial membrane (OM) and in the endoplasmic reticulum. PA is transferred from the outer to the inner mitochondrial membrane (IM) via the intermembrane space (IMS) and is converted to CL on the matrix face of the IM via the intermedates CDP-diacylglycerol (CDPDG), phosphatidylglycerophosphate (PGP), and phosphatidylglycerol (PG). The final step of the pathway is the remodeling of nascent cardiolipin (CLn) to mature cardiolipin (CLm). Tafazzin, the remodeling enzyme requires disturbance of the bilayer packing order, such as high membrane curvature, in order to be active. In yeast, tafazzin faces the intermembrane space. Mammalian genes are shown for the enzymes of PA synthesis and yeast genes are shown for all other enzymes.

The biosynthesis of CL takes place on the matrix face of the inner mitochondrial membrane by a sequence of four reactions (Fig. 1). All enzymes have been unequivocally identified in yeast, a task that took until 2013 to complete (Table 1). Since most proteins of the inner membrane are arranged in various kinds of complexes, it seems likely that the four enzymes cluster around each other and are not scattered throughout the membrane. Localization of the pathway in the inner leaflet of the inner membrane has been demonstrated in liver mitochondria [8] but any more specific sub-compartmentalization, i.e. whether the enzymes are

Table 1 Enzymes of the mitochondrial CL pathway. Step

Enzyme

Gene

References

G3P + acyl-CoA ? LPA + CoA LPA + acyl-CoA ? PA + CoA PA transport PA + CTP ? CDPDG + P  P CDPDG + G3P ? PGP + CMP PGP + H2O ? PG + P PG + CDPDG ? CL + CMP CL transacylation

Glycerophospate acyltransferase Acylglycerophosphate acyltransferase Lipid transfer protein CDP-diacylglycerol synthase Phosphatiylglycerophosphate synthase Phosphatidylglycerophosphate phosphatase Cardiolipin synthase Tafazzin

GPAT1⁄, GPAT2⁄ AGPAT5⁄ Ups1 Tam41 Pgs1 Gep4, PTPMT1⁄ Crd1, CLS⁄ Taz1, TAZ⁄

[188,189] [3] [7] [10] [14] [16,17] [18] [32]

The yeast nomenclature of the genes was used except for the ones marked by an asterisk.

3

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

associated with the cristae, the cristae junctions, or the inner boundary membrane, has not been documented. Given the emerging view of the architecture of the inner membrane [9], such sub-compartmentalization should be suspected though. The first reaction on the CL pathway is catalyzed by Tam41 [10]. Tam41 was originally discovered as an activity required for the assembly and the maintenance of protein translocases of the inner membrane [11,12]. However, it is not an integral component of the translocator complex but attaches peripherally to the matrix side of the inner membrane. Later, Tam41 was found to be essential for CL biosynthesis [13] and eventually its catalytic function was identified as CDP-diacylglycerol (CDP-DG) synthase [10]. CDP-DG can also be synthesized in the endoplasmic reticulum and Tamura et al. provided some evidence that mitochondria can import CDPDG if Tam41 is deleted, although this import pathway seems to be rather inefficient [10]. The presence of two different CDP-DG synthases, one in mitochondria (Tam41) and one in the endoplasmic reticulum (Cds1), underscores not only the importance of CDPDG for phospholipid biosynthesis but also the relative independence of the two organelles with regard to phospholipid biosynthesis. Sequence analysis of the two CDP-DG synthases, suggests that the endoplasmic reticulum enzyme is derived from ancient eukaryotes while the mitochondrial enzyme is derived from the prokaryotic ancestors of mitochondria [10]. The second reaction on the CL pathway is catalyzed by Pgs1 [14]. The enzyme forms phosphatidylglycerophosphate (PGP) by transferring a phosphatidyl group from CDP-DG to the sn-1 hydroxyl group of glycerol-3-phosphate. The conversion of an anhydride bond into an ester bond produces a large drop in free energy, which makes this reaction a suitable target for biological rate control. It has been shown that phosphorylation of Pgs1 reduces its catalytic activity and causes a decrease of the CL concentration in yeast [15]. The third reaction on the CL pathway, is catalyzed by Gep4 [16] or PTPMT1 [17]. These enzymes remove the terminal phosphate group from PGP to form phosphatidylglycerol (PG). Establishing their activity has been challenging because PGP is a relatively hydrophilic phospholipid, that may be missed by standard lipid analysis. Gep4 and PTPMT1 have very different primary structures. While Gep4 orthologs exist in fungi and some plants, PTPMT1 orthologs catalyzes PGP hydrolysis in higher eukaryotes. Interestingly, PTPMT1 belongs to the protein tyrosine phosphatase family [17]. The fourth and final reaction of CL biosynthesis is catalyzed by Crd1 [18]. This enzyme, referred to as cardiolipin synthase, uses two phospholipid substrates, PG and CDP-DG, to form CL. The identity of the yeast gene Crd1, has been confirmed by different groups [19,20] and eventually became an important clue for the identification of cardiolipin synthases in humans [21–23] and plants [24,25]. CL biosynthesis in mitochondria is different from CL biosynthesis in prokaryotes, which proceeds through reversible phosphatidyl transfer from one PG molecule to another (for a review, see Ref. [26]) or, as recently discovered, by phosphatidyl transfer from phosphatidylethanolamine to PG [27]. The clear separation between a eukaryotic and a prokaryotic mechanism of CL formation has been called into question though because some unicellular eukaryotes contain a bacterial-type rather than a eukaryotic cardiolipin synthase [28,29]. The evolution of eukaryotes may have proceeded through a stage, in which two different enzymes coexisted, one that employs the prokaryotic mechanism (derived from the mitochondrial ancestor) and one that employs the eukaryotic mechanism (derived from the protoeukaryote). According to this theory, the eukaryotic enzyme was eventually lost on the evolutionary path towards protozoae, and the prokaryotic enzyme was lost on the evolutionary path towards higher eukaryotes [30].

2.3. Remodeling of CL Subsequent to its biosynthesis, CL acquires a new set of fatty acids. The specific fatty acids vary between organisms but common to all is that only one or two types of unsaturated acyl groups dominate in mature CL. The post-synthetic exchange of acyl groups, commonly called remodeling, has attracted much attention, partly because it has been linked to human Barth syndrome and partly because it has provided novel insights into the inner workings of mitochondrial membranes. A detailed review of CL remodeling has recently been published [31]; thus we will limit ourselves here to three issues, which are the mechanism, the topology, and the biological function of the process. Our understanding of the mechanism of CL remodeling has undergone several revisions over the past twenty years. Now it is generally accepted that CL remodeling critically depends on tafazzin, a phospholipid–lysophospholipid transacylase [32]. The participation of other enzymes has been invoked and some controversy still exists with regard to how fatty acids are removed and how they are re-attached, but the involvement of tafazzin has been clearly established [33]. It is important to realize that tafazzin can catalyze both the removal and the re-attachment of fatty acids, which results in an exchange of acyl groups between CL and other phospholipids (Fig. 2). The question then becomes how does this process direct specific fatty acids into CL? Experiments with purified tafazzin demonstrated strong specificity in the acyl exchange reaction as a result of the lipid phase state, driven by membrane curvature or other properties related to the packing of lipids [34]. For instance, the characteristic molecular species of mitochondrial CL were observed in this experiment only if the lipids formed an aqueous interface with high negative curvature. Thus, tafazzin alone can account for all features of the remodeling process, including acyl removal, acyl re-attachment, and acyl specificity (Fig. 2). Nevertheless, other remodeling enzymes have been discussed in the literature, such as the CL hydrolase Cld1 [35], or enzymes with lysophospholipid:acyl-CoA acyltransferase activity, such as ALCAT1 [36] and MLCLAT1 [37]. In particular, the yeast enzyme Cld1 has clearly an effect on the molecular species

A

CL + LPL

B

MLCL + PL

PL

LPL

PL MLCL

CL

LPL

CL

CL LPL PL

LPL

MLCL MLCL

PL

CL PL LPL

C

CL + 4 PL

LPL

PL CL + 4 PL

Fig. 2. Mechanism of CL remodeling by tafazzin. A: The remodeling process is based on a single type of reaction, which is the transacylation between CL and a lysophospholipid (LPL), producing monolyso-cardiolipin (MLCL) and a phospholipid (PL). B: Because the transacylation is reversible, it can remove fatty acids from CL and re-attach them. The ensuing fatty acid exchange can lead to a shift in the composition of molecular species. Fatty acids are represented by color-coded balls. C: The net reaction of tafazzin is the exchange of fatty acids between CL and other PLs.

4

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

composition of CL [35,38]. More work is necessary to determine the role these enzymes play, either in CL remodeling or in other mitochondrial functions. The topology of CL remodeling has to be considered in light of the localization of CL, which is present on both sides of the inner mitochondrial membrane [39–41]. Since CL is formed on the matrix side, it must traverse the inner membrane and, given its presence on the outer face of the outer membrane [42–44], must be able to migrate further into the periphery (see Section 3). In yeast, it has been established that tafazzin, the key enzyme of CL remodeling, is localized on the outer side of the inner membrane and on the inner side of the outer membrane, i.e. it is facing the intermembrane space [45]. Thus, CL has to be translocated in order to be remodeled and the question arises whether CL remodeling is mechanistically linked to CL translocation? Although this question remains to be answered, it should be pointed out that CL remodeling requires disruption of the bilayer state [34] and non-bilayer phases provide a potential mechanism for CL translocation. Remodeling is not a continuous process but occurs only in newly synthesized CL [46] and remodeling is limited to specific membrane domains [34]. Thus, the current data suggest that CL synthesis and remodeling are a combined process that is confined to a specific site within the mitochondria. What is the function of CL remodeling? Naturally, we and others have suspected that remodeling produces a ‘‘better’’ type of CL, i.e. a CL that is more functional for mitochondrial membranes. However, another possibility is that CL remodeling, by virtue of re-arranging acyl groups between lipid species, minimizes the energy constraints of dynamical events, such as membrane fusion, the formation of membrane folds, or the assembly of membrane domains [34]. For instance, tafazzin can support the formation of high negative curvature at the lipid-water interface (Fig. 3). In that sense, it is possible that the changes in CL composition brought about by tafazzin, are merely the byproduct of the transacylation process, the actual function of which lies in supporting mitochondrial membrane dynamics. The idea that the CL composition may not be as important as the remodeling itself, is supported by the observation that abnormal CL compositions may be compatible with normal oxidative phosphorylation and normal mitochondrial morphology [38,47]. 3. Trafficking of mitochondrial CL Although CL is synthesized in the inner leaflet of the inner mitochondrial membrane, it can be detected in other locations, e.g. the outer leaflet of the inner membrane, the outer mitochondrial membrane, and even extra-mitochondrial locations, thus necessitating its translocation between biological membranes and also between leaflets within these membranes. Trafficking of CL from the inner leaflet of the inner membrane to the surface of the mitochondrion requires at least three translocations: (i) from the inner to the outer leaflet of the inner membrane, (ii) from the outer leaflet of the inner membrane to the inner leaflet of the outer membrane and, finally (iii) from the inner to the outer leaflet of the outer membrane. While recent studies have revealed the physiological significance of CL translocation, little is known about the molecular mechanisms involved in CL trafficking. 3.1. Intramembrane translocation (flip-flopping) Since CL is synthesized in the matrix-facing leaflet of the inner membrane but undergoes acyl remodeling in the leaflet facing the intermembrane space, either CL or monolyso-cardiolipin (MLCL) must flip across the inner membrane (see Section 2.3). Obviously, to carry out its function, mature CL has to be able to flip back to

A

B

C

+

TAZ

+

Fig. 3. The function of tafazzin. The figure demonstrates how tafazzin can lower the energy requirement for membrane bending by forming new molecular species. A: In a flat membrane, lipids form stable arrangements of tightly packed molecules. Head groups are depicted as circles, and fatty acids are shown as color-coded rectangles. Blue and cyan represent saturated residues, and yellow represents unsaturated residues. Only one leaflet of the bilayer is shown. B: Bending of the membrane will disturb the packing order of lipids, which is thermodynamically unfavorable. C: Tafazzin reshuffles acyl residues to optimize lipid packing. In this example, diunsaturated phospholipids (yellow-yellow) are formed because they promote negative curvature as a result of their cone-like shape.

the matrix-facing leaflet of the inner membrane, as it is generally believed that in normally functioning mitochondria CL is found primarily in the inner leaflet of the inner membrane [48,49]. How CL or MLCL flip-flop between leaflets of the inner membrane is not clear. It may involve CL interactions with integral membrane proteins, for example, the ADP/ATP carrier, the most abundant protein in the inner membrane [50]. Alternatively, accumulation of CL at the site of its synthesis may result in non-bilayer arrangements [51,52], facilitating CL flip-flop between leaflets of the inner membrane. 3.2. Intermembrane translocation The mechanism for the second stage of CL redistribution, its transfer from the inner to the outer mitochondrial membrane, has also not been established. It is thought to occur mostly at the contact sites between outer and inner membranes, whose existence has been observed by electron microscopy in negatively stained samples since 1960 [53] and more recently by electron tomography [54,55]. The idea that CL transport may occur at these sites was first inferred from studies in which transported lipids

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

were found to accumulate in membrane fractions that contain contact sites [56] and from studies that showed contact site-enriched membrane fractions contain more CL than generally found in the outer or inner membrane [57,58]. Recently, three types of protein complexes have been identified that could physically connect inner with outer membrane and potentially provide the structural basis of these contacts. First, a group of homo-oligomeric kinases located in the intermembrane space and enriched in contact sites, namely nucleoside diphosphate kinase D (NDPK-D) and mitochondrial creatine kinases (MtCK), are capable of cross-linking two opposite membranes [59]. Second, a large hetero-oligomeric protein complex anchored in the inner membrane and first described to maintain cristae junctions and morphology, the MINOS/MICOS/MitOS complex, is also able to bind outer membrane proteins (reviewed in Ref. [60]). Third, the ATPase family AAA Domain-containing protein 3 (ATAD3) can bridge inner and outer membrane (reviewed in Ref. [61]). The following is a brief discussion of their potential role in intermembrane translocation of CL, more details can be found in two excellent recent reviews on this topic [62,63]. NDPK-D and MtCK. NDPK-D and MtCK are not phylogenetically related but share several important properties. They both use mitochondrially generated ATP to maintain proper nucleotide pools, are located in the intermembrane space, form large symmetrical homooligomeric structures, display high affinity for CL, and promote lipid transfer between two membranes. In the mitochondrial intermembrane space, NDPK-D forms symmetrical hexameric complexes, in which CL binding sites have been identified by molecular modeling [64]. NDPK-D can induce intermembrane contacts between CL-containing liposomes in vitro. Importantly, formation of such contact sites promotes the transfer of fluorescently labeled model lipids from donor to acceptor liposomes without inducing liposome fusion [65]. Thus, in its fully CL-bound form, a NDPK-D hexamer could physically bridge the inner and the outer membrane [66]. In Hela cells, overexpression of wild-type NDPK-D, but not CL-binding deficient mutants, selectively increases the CL content of the outer membrane that renders the cells more sensitive to rotenone-induced apoptosis [64]. Similar to NDPK-D, the octameric MtCK complexes also have CL binding motifs and are able to cross-link two different membranes [67,68] and transfer lipids between the two membranes [65]. Interestingly, the tissue expression patterns of MtCKs and NDPK-D are complementary [69], suggesting that they serve equivalent roles in intermembrane CL transfer [59]. MINOS/MICOS/MitOS. Anchored in the inner membrane and enriched at cristae junctions, this large hetero-oligomeric protein complex contains at least 6 different proteins in yeast (mammalian orthologues in brackets): the core components Fcj1 (mitofilin) and Mio10 (MINOS1), as well as Aim5, Aim13 (CHCHD4/MINOS3) and Aim37-Mio27 (MOMA1) [60]. This complex can interact with various integral outer membrane proteins to form contact sites, including the translocase of the outer membrane (TOM), the sorting and assembly machinery (SAM), the voltage-dependent anion channel (VDAC), and the outer membrane fusion protein Ugo1. It was originally discovered as a protein complex that maintains mitochondrial morphology by supporting the cristae junctions that connect the inner boundary membrane to the cristae membrane. There is evidence that MINOS complexes are involved in intermembrane lipid movements. Apolipoprotein O and apolipoprotein O-like protein physically interact with the MINOS complex and show CL binding activity in vitro [70]. MitOS genes show a strong genetic interaction with genes in the CL biosynthetic pathway [71]. However, MINOS mutants with aberrant mitochondrial morphology did not display changes in mitochondrial phospholipid distribution [72]. Thus, it remains to be established whether this complex is active in CL transport.

5

ATAD3. ATAD3 (ATPase family AAA domain-containing protein 3) is a mitochondrial membrane bound ATPase with its C-terminal ATPase domain located in the matrix and its N-terminus exposed to the cytosol, thus physically bridging the inner and outer membrane (reviewed in Ref. [61]). Like many other ATPases, ATAD3 can form homo-oligomeric structures up to hexamers [73]. Knock-down studies in primary and immortalized cultured cells have shown that ATAD3 is not only essential for maintaining the mitochondrial network [74,75], but also important for lipid-requiring mitochondrial biogenesis [76] and cholesterol transfer from ER to mitochondria [74,77,78]. How ATAD3 functions in cholesterol transfer and whether it plays a role in CL trafficking are yet to be clarified. 3.3. Exposure of CL at the surface of mitochondria Although CL is found mostly in the inner membrane of normally functioning mitochondria, upon mitochondrial injury and depolarization, a significant portion of CL becomes exposed on the mitochondrial surface, where it serves as either pro-mitophagic or pro-apoptotic signals, depending on the extent of mitochondrial injury [79,80]. The appearance of CL at the mitochondrial surface may also explain why ALCAT 1, an acyltransferase of the mitochondria-associated endoplasmic reticulum [36], may be actively involved in CL metabolism of injured mitochondria [81]. Exposure of CL at the surface of mitochondria obviously requires its translocation from the inner to the outer leaflet of the outer membrane. Recent studies indicate that this process requires one of the phospholipid scramblases, PLS3 [80,82,83]. Phospholipid scramblases (PLS1–4) are a group of homologous proteins originally thought to be directly involved in destroying membrane phospholipid asymmetry at critical cellular events like cell activation, injury and apoptosis [84]. But despite what their names suggest, recent studies have shown that these proteins are not phospholipid scramblases in vivo and instead, may have a role in cell signaling [85]. As little if any of PLS3 has been found in mitochondria [86], it remains unclear whether PLS3 is the actual vehicle of CL translocation. Alternatively, CL translocation from the inner to the outer leaflet of the outer membrane may involve integral outer membrane proteins capable of binding CL, such as members of the Bcl-2 family [87] or VDAC [88]. 3.4. CL trafficking beyond mitochondria Although CL is the signature phospholipid of mitochondria, it has been detected in other subcellular fractions, e.g. peroxisomes [89,90] and plasma membrane [91,92]. While it is notoriously difficult to rule out cross-contaminations in such studies, it seems at least conceivable that CL traffics beyond mitochondria. Trafficking to organelles within the endomembrane system may be explained by mitophagy and the membrane fusion/fission within the endomembrane system (including plasma membrane). CL trafficking to peroxisomes may be mediated by a class of recently discovered mitochondria-derived vesicles which bud from mitochondria, traffic through cytoplasm, and fuse with peroxisome [93]. Mitochondria-derived vesicles may also fuse with lysosomes, which may selectively remove proteins, complexes and lipids (e.g. CL) from actively respiring mitochondria and subject them to degradation. This form of degradation is distinct from mitophagy [94]. 4. Degradation of mitochondrial CL and apoptosis Although CL has been shown to have a much slower metabolism than other phospholipids [46,95], it is degraded to some

6

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

extent either by phospholipases or by peroxidation. CL degradation does not only maintain turnover but is also involved in the regulation of cell death. Specifically, the peroxidation of CL, which in itself may destroy CL molecules or make them more susceptible to phospholipase-mediated hydrolysis, has been linked to apoptosis [96]. In order to complete the apoptotic program, CL has to re-distribute from the inner to the outer membrane [97]. This suggests that the same mechanisms involved in CL externalization and signaling in mitophagy are also required for the apoptotic death pathway. An interesting question is how intracellular regulatory mechanisms switch from mitophagic to apoptotic pathways. 4.1. CL hydrolysis by phospholipases Like any other phospholipid, CL can be hydrolyzed by a number of phospholipases. Some of these enzymes have been demonstrated to act on CL only in vitro but others have been implicated in CL hydrolysis in vivo (Table 2). Cld1p. The yeast enzyme Cld1p is a CL-specific deacylase but does not belong to any of the established phospholipase families. Since deletion of Cld1p leads to large changes in the composition of molecular species of CL, this enzyme has been suspected to play a role in CL remodeling [35]. Cld1p is associated with the matrixfacing leaflet of the inner membrane; thus, in relation to tafazzin, it is localized on the opposite side of the inner membrane [98]. There are no orthologs of Cld1p in higher eukaryotes. iPLA2b. A calcium independent phospholipase A2, iPLA2b has been shown to localize to mitochondria in many cell types [99,100]. It has been suggested to be partially responsible for CL depletion and MLCL accumulation in tafazzin-deficient Drosophila [101]. There is also evidence that iPLA2b is important in repairing oxidized mitochondrial lipids, such as oxidized CL, and that this prevents cytochrome c release in response to stimuli that otherwise induce apoptosis [100,102]. In an in vitro CL hydrolysis assay, iPLA2b caused an accumulation of MLCL [103]. iPLA2c. The iPLA2c is a membrane-bound calcium independent phospholipase A2 with dual mitochondrial and peroxisomal localization signals. Genetic ablation of iPLA2c in mice leads to multiple tissue specific biochemical, ultrastructural, and functional alterations [104,105]. For example, the observed increase in hippocampal CL content and specifically the increase in shorter saturated acyl species could be direct consequences of decreased hydrolysis of CL [105]. Loss of iPLA2c function may either directly affect the turnover of CL acyl groups or affect the turnover of acyl groups of other phospholipids, which may then be transmitted to CL via transacylations. The iPLA2c–knockout myocardium, on the other hand, contains reduced CL levels with an increased relative proportion of molecular species of arachidonic and docosahexenoic acids

Table 2 Putative phospholipases involved in CL degradation. Phospholipase

Product

Localization

Evidence

Cld1p iPLA2b

MLCL, FA MLCL, FA

In vitro, yeast In vitro, flies, mice

iPLA2c cPLA2 sPLA2

MLCL, FA MLCL, FA MLCL, DLCL, FA PA, PG

IMM (matrix side) Mitochondria and cytosol Mitochondria Cytosol Secreted OMM (cytosolic side)

In vitro, NIH3T3 cells

MitoPLDa

In mice, rats In vitro In vitro

Abbreviations: DLCL, dilyso-cardiolipin; FA, free fatty acid; IMM, inner mitochondrial membrane; MLCL, monolyso-cardiolipin; PA, phosphatidic acid; PG, phosphatidylgycerol; OMM, outer mitochondrial membrane. a MitoPLD was recently shown to act as an endoribonuclease but not a phospholipase [109–111].

[104]. The lack of accumulation of CL in myocardial mitochondria suggests that compensatory mechanisms for hydrolyzing CL are present in myocardial mitochondria that are not effective in hippocampus. Kinetic studies investigating CL remodeling in cell culture of isolated myocytes also suggest that mitochondrial iPLA2c has an effect on the acyl composition of CL [106]. cPLA2. CL has been shown to be a substrate for the cytosolic phospholipase A2 cPLA2 by using fluorescent substrate [107]. This conclusion was reaffirmed by a recent mass spectrometric assay without fluorescent labeling [103]. As cPLA2 is cytosolic, hypothetically it may hydrolyze the externalized CL on the mitochondrial surface. sPLA2. CL is also a substrate for the secreted phospholipase A2 (sPLA2) [107]. In vitro assays documented that sPLA2 causes an accumulation of dilyso-cardiolipin (DLCL), in contrast to iPLA2 that causes an accumulation of MLCL [103]. sPLA2 purified from snake venom showed the highest activity among all PLA2s tested. When added to tissues, sPLA2 in the snake venom can cause excessive inflammation through hydrolysis of phospholipid membranes. MitoPLD. MitoPLD, a divergent member of the phospholipase D (PLD) superfamily, was found to localize to the mitochondrial surface via insertion of a short N-terminal anchor into the outer membrane where it functions as a homodimer [108]. The MitoPLD protein sequence is most similar to the bacterial endonuclease Nuc and the bacterial CL synthase. Neither nuclease activity nor CL synthase activity were detected for MitoPLD in the original report; however, when CL was supplied as a substrate in vitro, hydrolysis to PA was achieved, suggesting phospholipase D activity [108]. Recently, several groups independently solved the MitoPLD crystal structures and carried out careful enzymatic studies using purified recombinant MitoPLD, and reached the conclusion that MitoPLD acts as a single-strand-specific endoribonuclease but not a phospholipase in vitro [109–111].

4.2. Peroxidation of CL During apoptosis initiation, CL is the only phospholipid in mitochondria that undergoes peroxidation, catalyzed by a cardiolipin-specific peroxidase activity of CL-bound cytochrome c [96]. Cytochrome c-catalyzed peroxidation of CL utilizes polyunsaturated molecular species, whereas saturated and monounsaturated CL molecules do not undergo peroxidation [112]. In its native structure, cytochrome c functions as an electron shuttle between respiratory complexes III and IV in mitochondria. Upon binding of CL, cytochrome c undergoes a structural reconfiguration and transforms into a CL-specific peroxidase [113]. The structural reconfiguration of cytochrome c appears to involve both electrostatic interactions of one or more positively charged amino acid (lysine) residues with negatively charged phosphate groups of CL as well as hydrophobic interactions of one or more of CL’s fatty acid residues with nonpolar sites of the protein [114,115]. The peroxidase function of cytochrome c requires its direct physical interaction with CL. Normally, CL is present primarily in the inner leaflet of the inner membrane, whereas cytochrome c is confined to the intermembrane space. Thus, binding of cytochrome c to CL depends on the availability of the latter in the outer leaflet of the inner membrane. Moreover, significant demand for high-affinity CL binding by other mitochondrial proteins such as mitochondrial respiratory complexes I, III, and IV, and ADP/ATP carriers, also limits access of cytochrome c to CL. It appears that mitochondrial injuries (metabolic or chemical) generate reactive oxygen species, such as superoxide radicals, which cause (through an as yet unknown mechanism) a significant amount of CL to flip to the outer leaflet facing the intermembrane space where cytochrome c is located. Upon binding of CL, cytochrome c is transformed into a CL-specific

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

peroxidase, which catalyzes CL peroxidation in the presence of sufficient supply of H2O2, a product of superoxide dismutation. The initial generation of peroxidized CL is believed to trigger a further surge of peroxidase activity by cytochrome c unfolding, which promotes further peroxidation of CL [116]. It is well documented that CL peroxidation is essential for the release of pro-apoptotic factors from mitochondria into the cytosol [117,118], but how peroxidized CLs facilitate this release has yet to be clarified [87]. Since CL peroxidation may surge in a positive-feedback chain reaction once initiated and culminate in apoptosis, an injury threshold may exist for mitochondria to respond appropriately to insults of varying degrees. Recent findings from Kagan’s laboratory have demonstrated the existence of such a threshold. They showed that mitochondria respond to a sub-lethal insult by redistributing significant amounts of CL from the inner leaflet of the inner membrane to the mitochondrial surface without CL peroxidation. Then the exposed CL initiates mitophagy, not apoptosis [80]. This could happen if the level of superoxide radicals generated by the sub-lethal insult is enough to cause CL translocation, but not enough to provide sufficient amounts of H2O2 for CL peroxidation. 5. Function of CL in mitochondrial membranes The function of CL in mitochondria has to be understood in the context of its unique physicochemical properties. What is unique about CL in contrast to other phospholipids can be traced back to a single structural feature, namely the bonding of two phosphatidyl moieties with a single glycerol group. This feature results in a small, relatively immobile head group, which in turn promotes negative curvature, cohesive effects between hydrocarbon chains, and electrostatic interactions [119]. The reason why such properties are of particular importance to mitochondria may be related to the high protein density of cristae membranes, the need for

7

extensive membrane folds, or the need for continuous fission and fusion, which requires non-bilayer lipid phases with high negative curvature. Thus, the likely role of CL in mitochondria is (i) to support membrane dynamics and (ii) to stabilize the lateral organization of protein-rich membranes. 5.1. Physicochemical properties of CL An excellent review of the physicochemical properties of CL has been published [119] and the key points of this article are summarized in Fig. 4. The central aspect of the CL structure, one that explains many of its physical characteristics, is the presence of a small head group that is largely immobile because it is tethered to two phosphatidyl groups [120]. An NMR study of deuteriumlabeled CL demonstrated that the glycerol head group is less flexible than the head groups of other phospholipids and is oriented parallel to the membrane surface, whereas the two backbone glycerols are oriented parallel to the membrane normal [121]. The two phosphatidyl groups of CL are stereochemically non-equivalent [122], which may result in different reaction rates or binding affinities, in particular if the difference is amplified by unequal acyl composition [123]. The stereochemistry is even more important for monolyso-CL where four rather different species will emerge depending on whether the acyl group is missing from the sn-10 or the sn-30 phosphatidyl group and whether it is missing from the sn-1 or the sn-2 carbon position. By virtue of having two phosphate groups, CL is an acid that can form salts in the presence of cations. There has been some controversy about engagement of the phosphates in intramolecular hydrogen bonds, which would affect the net charge of CL, but the majority of data seem to support that each phosphate carries one negative charge at physiologic pH [119]. The four acyl groups of CL form a compact hydrophobic moiety that is large compared to the size of the head group. The presence of four acyl groups instead of two together with motional

Backbone Glycerol Groups Posioned parallel to membrane normal; liming the head group mobility

Headgroup Glycerol Group Posioned parallel to membrane surface; small size relave to total molecule; restricted moonal flexibility; H-bonding via free hydroxyl group

Phosphate Groups Net charge: -2 (pH dependent); ionic bonding and H-bonding Membrane Surface

Acyl Groups Posioned parallel to membrane normal; large size relave to total molecule

Fig. 4. Physicochemical properties of CL. The molecule consists of 3 glycerol groups, 2 phosphate groups, and 4 acyl groups. Overall, CL has a small head group and a large and rigid hydrophobic tail; it has intrinsic negative curvature and exerts a strong ionic force [119].

8

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

constraints resulting from the fixed position of the head group, are likely to encourage hydrophobic interactions between the acyl groups [119]. The small head group of CL and its large hydrophobic tail produce a conical shape of the molecule, which tends to impose negative curvature on the lipid-water interface. As a result, CL is prone to form non-bilayer phases with high negative curvature, such as inverted hexagonal or inverted cubic phases. CL has been demonstrated to accumulate in negatively curved regions of Escherichia coli membranes [124]. However, a force that counteracts the formation of inverted non-bilayer phases is repulsion by the charged phosphate groups, simply because the surface density of those groups increases in inverted phases. As a result, CL forms bilayers unless the surface charges are screened by low pH [125], high ionic strength [126], divalent cations [127], cationic drugs [51], proteins [128], or positively charged lipids [129,130]. The number of acyl groups also plays a role in the formation of nonbilayer phases, i.e. adding an acyl group increases the tendency towards non-bilayer phases whereas removing an acyl group decreases it [131]. The phase polymorphism of CL could be partially simulated by molecular dynamics calculations [132], which has inspired confidence in the ability of computational methods to predict the physical behavior of CL. The relative rigidity of the CL molecule is likely to make the negatively charged phosphates more prone to ionic interactions with neighboring molecules, such as proteins [119]. Rigidity may also explain why CL increases the packing order in mixed monolayers [120,133,134]. Thus, CL exerts strong ionic and hydrophobic forces on neighboring molecules. 5.2. CL in specific membrane domains The majority of biophysical studies have shown that the principal effect of CL on biological membranes is to increase the order of the lateral distribution of lipids and proteins. A possible outcome of this effect is the segregation of membrane domains, driven either by the limited miscibility of CL with other lipids [135] or by the tendency of CL to induce the clustering of proteins (see Section 5.3). Specific domains of the inner mitochondrial membrane, such as the inner boundary membrane, cristae junctions, cristae, and cristae tips, have been identified based on their morphology and their protein composition [9] but the concentration of CL in these domains has not been determined. However, CL is apparently enriched in contact sites between the inner and the outer membrane [57,58]. Membrane fusion sites, cristae junctions, and cristae tips contain leaflets with high negative curvature and could potentially benefit from the ability of CL to stabilize negative curvature. CL also confers the capacity to form cristae-like membrane invaginations in response to changes in local pH in giant unilamellar vesicles [136]. This phenomenon, which is based on surface expansion of the protonized leaflet of the membrane, is certainly different from the way cristae are formed in mitochondria; however, it highlights the inherent potential of CL to stimulate domain formation and dynamical behavior in biological membranes. 5.3. Interaction of CL with proteins CL interacts strongly with a number of mitochondrial proteins. There is little specificity in these interactions, i.e. CL also binds to

extramitochondrial and even extracellular proteins in vitro. Nevertheless, the interactions involve strong binding forces. In fact, high but promiscuous protein affinity is one of the most prominent characteristics of CL, the foundation of which probably lies in the specific molecular geometry and the presence of unshielded negative charges (see Section 5.1). CL exerts its influence on different levels of the structural hierarchy ranging from monomers to large protein complexes. Effects may be seen on the tertiary structure, the catalytic function, the propensity of protein association, and the supramolecular order in protein clusters up to a micron in size (Table 3). The interactions of CL with proteins have been well documented, in particular for cytochrome c and the ADP/ATP carrier. Cytochrome c is a small water-soluble protein whereas the ADP/ ATP carrier is a transmembrane protein with six membrane-spanning helices. The fact that CL binds strongly to such different proteins is perhaps the best evidence for the versatile nature of CL–protein interactions. In the case of cytochrome c, CL induces changes in the tertiary structure that alter the environment of the heme catalytic center and ultimately convert the enzyme into a peroxidase [96]. The CL–cytochrome c binding is primarily electrostatic in nature but involves also hydrophobic interactions [137]. Likewise, the ADP/ATP carrier binds CL with high affinity. This binding was first detected by line broadening of the NMR resonances of the CL phosphorus atoms in preparations of the purified protein [138]. CL stabilizes the tertiary structure of the ADP/ATP carrier and can only be released by denaturation of the protein. Not surprisingly, CL was detected in the crystal structure of the carrier [139]. The first published structure contained 2 molecules of CL per carrier monomer but this was later revised to 3 molecules of CL per carrier monomer [140], which is in agreement with the NMR study [138]. Very recently it was reported that the phosphate groups of CL do not bind electrostatically to positively charged amino acid side chains, as previously believed, but interact with the amide moieties of conserved glycines [141]. CL acyl groups on the other hand adopt flexible conformations within the crystal structures and do not require any conserved protein surface [141]. In the crystals, CL–CL interactions contribute to the interface between neighboring monomers, which is consistent with a role of CL to facilitate protein association (Fig. 5). Further supporting this idea, CL was shown to keep the ADP/ATP carrier integrated into a large heterogeneous protein supercomplex [142]. The ability to promote the clustering of membrane proteins is an important property of CL, for which there are many examples. This was first recognized in respiratory supercomplexes of yeast mitochondria [143] but also applies to other large protein assemblies, such as dynamin-related proteins involved in membrane fusion [144]. CL is a ‘‘soft’’ component of supercomplexes without fixed stochiometry but its presence is crucial to shift the association–dissociation equilibrium of these structures towards association, both inside the membrane [145] as well as in the solubilized state [146,147]. The number of CL molecules per supercomplex was estimated to be 50 in yeast III2IV2 [148] and 200–400 in bovine heart I1III2IV1 [149]. Finally, there are supramolecular structures that are even larger than supercomplexes, such as the dimer ribbons of complex V at the tip of mitochondrial cristae [150]. CL has an effect on both the degree of dimerization and the structural order within the dimer ribbon [151]. Thus, CL exerts a strong ‘‘organizing’’ effect on all kinds of protein assemblies.

Table 3 Effect of CL on protein assemblies. Hierarchy

Mass (Da)

Example

Effect of CL

References

Protein Protein complex Supercomplex Complex array

104–105 105–106 106–107 >107

ADP/ATP carrier Complex III Respiratory supercomplexes ATP synthase dimer ribbon

Stabilization of the tertiary structure Support of proton conduction Increase of supercomplex association Increase of order in the arrangement of protein complexes

[138] [152] [143] [151]

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

9

Fig. 5. Schematic showing the localization of CL in the crystal structure of the ADP/ATP carrier. Orange circles represent the surface of the protein monomers viewed from the matrix. Blue sticks represent CL molecules. Each carrier monomer binds three molecules of CL. CL molecules contribute to the protein–protein interface in the crystal and may potentially promote oligomerization of carrier monomers in vivo [140].

CL has also been found in some bona fide protein complexes, i.e. those that rely on specific protein–protein interactions. For instance, CL is present in crystals of respiratory complexes III [152] and IV [153] and there is NMR evidence for its tight association with complex V [154]. The topology of CL in complex III, suggests that it is directly involved in proton conduction, which essentially makes it a prosthetic group of the enzyme [152]. Another example for protein complexes that interact with CL is the water-soluble phosphotransferases of the intermembrane space. Both MtCK and NDPK-D form symmetrical oligomers that bind to CL-containing membranes peripherally and induce the segregation of CL domains (for a review see Ref. [59]). Preliminary evidence suggests that these oligomers may link the inner and the outer membrane and may provide a mechanism for CL transfer between the two membranes [64,65] (see Seection 3.2). In summary, CL binds tightly to a large variety of proteins, which can affect their tertiary structure and their catalytic function. Few molecules of CL are typically immobilized at the protein surface. If more CL is available, proteins universally become more prone to form clusters.

6. Models of CL deficiency Because the CL biosynthetic pathway is evolutionarily conserved across species from yeast to human beings, disruption at various points in the pathway have led to insights into the cellular

and organismal phenotypes associated with CL deficiency. A number of model organisms are available to study the phenotypic consequences of CL deficiency, each with its own advantages and disadvantages (Table 4). Furthermore, CL deficiency is associated with human diseases, although the role CL plays in such diseases has remained poorly understood. Caused by mutations in the tafazzin gene with consequences on CL remodeling, Barth syndrome was the first human disease shown to result from a primary deficiency in CL. In addition, CL has been implicated to play a role in a range of other disorders. In this section, we will focus on models of and diseases associated with, CL deficiency.

6.1. Experimental models of CL deficiency CL’s central role in mitochondrial functioning is evolutionarily conserved across a broad range of species. Thus, experimental model systems have yielded important insights into CL biology (Table 5). Not surprisingly, the ablation of CL causes abnormalities in mitochondrial structure and function, such as disorganized cristae and dissociated supercomplexes along with the expected defects in membrane potential and oxidative phosphorylation. These changes are the apparent cause for reduced growth and viability. However, the specific phenotypes are rather different in various organisms, likely reflecting different importance of mitochondria in their development. As a general rule, ablation of CL seems to be tolerated to some degree by unicellular cultures, albeit

Table 4 Comparison of model organisms used to study CL biology. Organism

Advantages

Disadvantages

Yeast S. cerevisiae

Highly tractable organism All enzymes in the CL biosynthetic pathway have been ablated genetically to determine downstream effects Simple organism that resembles higher-order organisms Well described anatomy and physiology Powerful genetic tools Invariant cell number and anatomy Certain aspects of striated muscle may model cardiac muscle Powerful genetic tools

Unicellular organism Muscle physiology cannot be studied

Worm C. elegans

Fly D. melanogaster

Zebrafish D. rerio

Mouse M. musculus

Flight muscles can model striated muscles Heart tube is a model for conserved early cardiac development in higher organisms Vertebrate organism Larval stages not dependent on heart function Rapid development Taxonomic bridge between worm/fly and mouse Mammalian heart closely resembling human heart

Heart and circulatory system not essential

Heart tube does not resemble a mature or even embryonic heart in higher organisms Cardiomyocytes exhibit significant ultrastructural differences from vertebrates Heart has only a single ventricular chamber

May differ from human clinical conditions, including responses to potential therapies

10

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

Table 5 Models of CL deficiency. Organism

Mutation

Enzyme

Features

References

S. cerevisiae

ups1pD

Ups1p (lipid transfer protein)

[190]

Tam41 D pgs1D

CDP-DG synthase PGP synthase

gep4D

PGP phosphatase

crd1D

CL synthase

cld1D taz1D

CL deacylase Tafazzin

CHO cells

pgs-s

PGP synthase

T. brucei

cls

CL synthase(bacterial type)

A. thaliana

cls mutants

CL synthase

C. elegans

pgs-1

PGP synthase

crls-1

CL synthase

DCLS

CL synthase

DTAZ

Tafazzin

D. rerio

TAZ knockdown

Tafazzin

M. musculus

Ptpmt -/-

PGP phosphatase

TAZ knockdown

Tafazzin

TAZ mutations

Tafazzin

Decreased CL Abnormal mitochondrial ultrastructure when grown on fermentable media Growth defects Temperature-sensitive growth defect Unable to grow on non-fermentable carbon sources Loss of mtDNA converts pgs1D mutant from viable to a lethal phenotype Temperature-sensitive growth defect Destabilization of respiratory supercomplexes Loss of mtDNA exacerbates gep4D growth defects Temperature-sensitive growth defect Decrease in mitochondrial membrane potential Impaired iron homeostasis Cell cycle defects Normal growth but altered acyl chain composition of CL Poor growth Accumulation of MLCL at the expense of CL Increased oxidative stress Thermolability Increased glycolysis, reduced oxygen consumption, defective respiratory chain activity Growth defects Reduced motility Abnormal mitochondrial morphology Loss of membrane potential Reduced viability and infertility Abnormal mitochondrial ultrastructure Reduced membrane potential Abnormal mitochondrial ultrastructure and reduced membrane potential only in germ cells not in somatic cells Abnormal mitochondrial ultrastructure and reduced membrane potential only in germ cells not in somatic cells Decreased life span Profound muscle weakness Bradycardia Increased MLCL, decreased CL Abnormal mitochondrial ultrastructure Motor weakness Male sterility Embryonic lethality Unlooped, dysmorphic hearts, with pericardial edema and absent blood circulation Bradycardia Tail & eye abnormalities Whole-body deletion leads to early embryonic lethality Cre-mediated ablation in MEF’s causes slow growth, reduced oxygen consumption, reduced complex I activity, abnormal mitochondrial morphology Increased MLCL:CL ratio Fetal LV noncompaction cardiomyopathy or adult-onset cardiomyopathy Prenatal-perinatal lethality Altered cardiomyocyte proliferation in trabecular and compact zones Abnormal mitochondria Reduced respiratory activity Barth syndrome

D. melanogaster

H. sapiens

with strongly reduced viability. Sensitivity to CL deficiency increases though in more developed species. Among multicellular organisms, only Caenorhabditis elegans [155], Drosophila melanogaster [151], and interestingly also the plant Arabidopsis thaliana [156], can reach maturity in the absence of CL synthase, but they all show severe pathologies. Mammals, in contrast, suffer embryonic lethality when CL synthase is not present. In the following, we will discuss the most important model organisms. Yeast. A highly tractable organism, yeast has provided much of our knowledge of the biology of CL. All enzymes in the CL biosynthetic pathway have been ablated genetically to determine

[10] [14,191]

[16]

[192–194]

[35,38,47] [159,195,196]

[197]

[28]

[156]

[155]

[155]

[151]

[158]

[157]

[17]

[164–167,198]

[169–172,174,175,199]

downstream effects (Table 5). In general, yeast deficient in CL exhibit abnormalities in growth that are thought to relate to mitochondrial dysfunction. Typically, the more upstream the biosynthetic enzyme ablated, the more profound the phenotype. The yeast experiments in the logarithmic, early stationary, and stationary phases of growth, coupled with developmental data in mice (see below), suggest organisms exhibit different sensitivities to CL deficiencies at different stages. Worm. In worm, germ cells of a CL synthase mutant exhibited abnormal mitochondrial structure and function but somatic cells appeared unaffected [155]. Germ cell mitosis was greatly

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

decreased, an interesting finding given the male sterility reported in the tafazzin mutant fly model that preceded this model organism by several years [101]. Zebrafish. An antisense morpholino oligonucleotide-mediated knockdown of tafazzin led to early embryonic lethality in zebrafish, pointing to a critical role of CL in the early development of this organism [157]. The hearts remained un-looped and so never progressed beyond a very primitive stage. Fly. Drosophila mutants of CL synthase [151] and tafazzin [158] have been studied. Consistent with observations made in yeast [159], the phenotype of the former was more severe than the phenotype of the latter. Mutation of CL synthase caused nearly complete loss of CL, reduced life span, and severe muscle and cardiac dysfunction [151]. The tafazzin mutant of Drosophila provided some of the first detailed information on the effects of tafazzin knockout on organ functioning in a multicellular, higher-order organism [158,160]. This model exhibited a normal lifespan with a grossly normal heart rate, chamber size, and cardiac function. It should be noted, however, that the fly heart is a primitive tube contained within an open circulatory system and its ultrastructure exhibits notable differences from vertebrate hearts [161–163]. Two phenotypic features of the tafazzin deletion included abnormal flight muscles and male sterility. These organ-specific dysfunctions were accompanied by the expected abnormalities in mitochondrial structure and function. Genetic inactivation of

11

iPLA2b rescued the sterility [101], a finding that suggested that either CL depletion or MLCL accumulation led to the disease phenotype. Mouse. Novel genetic technologies have made it possible to manipulate the CL pathway in mice, specifically to knock out PTPMT1, a member of the protein tyrosine phosphatase family that was shown to be the mitochondrial phosphatidylglycerophosphate phosphatase [17]. Whole-body ablation of PTPMT1 led to early embryonic lethality post-implantation, while Cre-mediated conditional deletion in mouse embryonic fibroblasts demonstrated many consequences to be expected of cardiolipin deficiency in isolated cells (Table 5). A tafazzin knock-down mouse model has been commissioned by the Barth Syndrome Foundation after attempts at creating a tafazzin-knockout mouse have been unsuccessful. Initial reports by Acehan et al. [164] and Soustek et al. [165] showed a mild, late-onset cardiomyopathy and skeletal myopathy at the age of 7–8 months but no perinatal or prenatal lethality. Powers et al. [166] confirmed skeletal myopathy coupled with abnormal mitochondrial respiration and ultrastructure. Employing the same inducible knockdown model, Phoon et al. [167] unveiled a developmental cardiomyopathy, including left ventricular noncompaction–hypertrabeculation (Fig. 6). The difference between this study [167] and the initial studies [164,165] was a higher doxycycline dose that likely induced a greater degree of knockdown of

Fig. 6. Left ventricular noncompaction with hypertrabeculation in Barth syndrome. A: Echocardiogram of an infant with left ventricular noncompaction (adapted with permission from Ref. [200]). B: Echocardiogram of a child with left ventricular noncompaction (adapted with permission from Ref. [175]). C: Histologic section through a normal newborn mouse heart showing a well-developed left ventricular compact zone. D: Histologic section through a newborn heart from the tafazzin-knockdown mouse model showing deeper trabecular recesses and more abundant trabeculae (boxed area), thinner myocardium (arrowheads), and a less well-developed and more perforate interventricular septum (arrows). E: Coronal section of a normal E13.5 embryonic mouse heart. F: Coronal section of E13.5 embryonic heart from a tafazzin-knockdown mouse, demonstrating a thinner myocardium with more prominent trabeculae that occupy more of the left ventricular lumen; again, the septum appears more porous (C–F were adapted from Ref. [167]).

12

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

tafazzin in the embryos. Cardiomyopathy could be induced only during a narrow developmental window, consistent with the timing of the myocardial compaction process. Although systolic function appeared to be preserved, diastolic function was abnormal and preceded universal prenatal and perinatal lethality. The myocardium showed the typically increased MLCL:CL ratio and several mitochondrial ultrastructural abnormalities. However, mitochondrial ultrastructural abnormalities such as mitochondrial number and cristae density were dependent on the stage of development. Data from tafazzin-deleted mouse embryonic stem cells and Drosophila also indicated the importance of differentiation in the manifestations of tafazzin deficiency [160]. Altered cellular proliferation, with a differential between the trabecular and compact zones, accompanied the noncompaction and suggested a possible mechanism for this developmental cardiomyopathy. Thus, the tafazzin knockdown mouse model appears to recapitulate many of the clinical, biochemical, and cellular features of Barth syndrome. 6.2. Barth syndrome Barth syndrome is caused by mutations in tafazzin, the enzyme responsible for CL remodeling. The biochemical consequences of tafazzin deficiency include a decrease in CL, an increase in MLCL, and an altered composition of CL molecular species [168]. The clinical presentation of Barth syndrome includes cardiomyopathy, skeletal myopathy with muscle wasting, cyclical neutropenia, and growth delay [169,170]. Often diagnosed in infancy, the cardiomyopathy is a prominent feature of Barth syndrome patients, and may be of the dilated, hypertrophic, or noncompaction types [171]. Left ventricular noncompaction is considered a hallmark of Barth syndrome (Fig. 6). With increased awareness of Barth syndrome and advancements in fetal monitoring has come the realization that many Barth patients are lost in utero [172], a result also suggested in animal models of tafazzin and cardiolipin deficiency [17,157,167]. Barth syndrome exhibits remarkably little involvement of the central nervous system, whereas other mitochondrial disorders are frequently characterized by encephalopathy [173]. However, developmental delays are common [174,175]. Cardiac failure is a leading cause of death and disability in patients with Barth syndrome. Patients are also prone to life-threatening arrhythmias, and the risk of arrhythmias appears to be independent of the degree of cardiomyopathy. The cyclic neutropenia may also lead to sepsis, the other leading cause of death in this population. Notably, there is no known genotype–phenotype correlation, and the same gene mutation can lead to vastly different clinical courses even within the same family [176]. The highly variable clinical spectrum in Barth syndrome and the lack of genotype–phenotype correlation supports the notion that modifying factors play an important role in disease manifestations. Such modifying factors may include environmental influences or genetic background, or modifier genes. It is also possible there are compensatory mechanisms in play that may be stronger in some patients than in others. One has the impression that the phenotypes are more than just a reflection of energy deficiency from mitochondrial dysfunction. Much of what we know about the mitochondrial pathophysiology in Barth patients comes from studies of human fibroblasts and immortalized lymphoblast cell lines. These studies have demonstrated an increased MLCL:CL ratio and abnormal mitochondrial ultrastructure and function, including a decreased membrane potential while total ATP production is preserved due to a compensatory increase in mitochondrial mass [177–180]. A relatively recent advance has been the use of inducible human pluripotent stem cells (iPSC’s). Dudek et al. [181] confirmed respiratory chain abnormalities associated with structural changes in respiratory chain supercomplexes, leading to massive increases

in the generation of reactive oxygen species, in iPSC derived from fibroblasts from Barth patients. The MLCL:CL ratio was increased as expected with a total decrease in CL levels. The respiratory deficiencies appeared to correlate with phenotypic severity, in this very small cohort. Finally, Barth syndrome resembles another rare disease caused by mutations in DNAJC19 that is involved in mitochondrial protein import [182]. These parallels lend further clues to the pathophysiology of CL deficiency and the most promising pathways to be targeted for therapy. Nevertheless, while the critical role of CL is clear, how its deficiency leads to the pathology in Barth syndrome is not. 6.3. Abnormalities of CL implicated in other human diseases In addition to the rare Barth syndrome, many common human diseases have now been linked to abnormalities in CL. Since the CL deficiency appears to be secondary to the primary disease process, we will only briefly review this broad spectrum of diseases. Chicco and Sparagna [183] placed diseases into three general categories: (i) loss of cardiolipin content; (ii) cardiolipin peroxidation; and (iii) change in the cardiolipin acyl chain composition. Changes in cardiolipin content may result from reduced synthesis or increased degradation. Cardiolipin itself appears to be particularly sensitive to oxidative damage from reactive oxygen species, since it is rich in polyunsaturated fatty acids. Reactive oxygen species contribute to mitochondrial damage through CL peroxidation with loss of functional CL (reviewed in Ref. [184]). The potential role of CL oxidation in conditions such as aging, heart failure, ischemia– reperfusion injury, thyroid diseases, and neurodegenerative diseases (including Parkinson’s disease and perhaps Alzheimer’s disease) have been discussed [185]. One of the most common diseases in which CL is affected is diabetes. Both drug-induced and genetically induced diabetes in mice, are associated with rather profound alterations in the molecular species composition of CL in myocardial tissue [186,187]. The significance of this observation and how it relates to the molecular pathogenesis of the disease is not clear, although it has been speculated that it may be involved in diabetic cardiomyopathy. 7. Conclusions CL research has become an active field, in which new concepts rapidly emerge and in which review articles, such as the present one, may become outdated very quickly. Nevertheless, we have tried to describe the current state of affairs as accurately as possible and hope this article will not only serve as an overview but also help to map out areas of future research. Many of the gaps in our knowledge of CL biology have been filled in the past decade. For instance, all enzymes of the biosynthetic pathway have been identified. What is less clear though is the control of CL biosynthesis. Is CL synthesis a continuous process or is it occurring in bursts and how is it integrated in the overall biogenesis of mitochondria? In addition to such temporal questions, there are questions of spatial distribution. Where precisely does CL synthesis occur and which micro-compartment(s) does CL subsequently move to? The discovery that CL does not only accumulate in the inner membrane but can also migrate to the surface of mitochondria is an important milestone in our understanding of the CL dynamics [80]. Another area that is poorly understood is how CL is assembled into the working units of mitochondrial membranes, i.e. protein complexes and supercomplexes. Before CL is assembled or perhaps during the assembly process the fatty acids of CL are being exchanged. Acyl remodeling is one of the most intriguing steps as it selects molecular species that support negative membrane

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

curvature, thus creating a link between CL formation and membrane geometry and potentially membrane assembly [34]. However, the exact function and even the precise localization of CL remodeling have remained elusive. Once CL is assembled in the membrane, it remains remarkably stable, i.e. its turnover is much slower than the turnover of other phospholipids [46]. Whatever degradation occurs is either by enzymatic hydrolysis or by peroxidation. Degradation seems to be accelerated though in certain pathologic situations. The enzyme(s) involved in CL degradation remain to be identified as are the potential effects of oxidized CL species. The function of CL in mitochondrial membranes can be largely deduced from its unique physicochemical properties. Perhaps the most important effect of CL is that it promotes the assembly of protein complexes into supercomplexes [143]. In the absence of CL, the lateral distribution of proteins in the inner membrane is less ordered and mitochondria become morphologically abnormal and the energy metabolism becomes less efficient. Other cellular functions that directly or indirectly require mitochondrial activity become vulnerable as well. However, CL deficiency is associated with rather different phenotypes in different organisms. The investigation of inter-species differences in the response to CL deficiency promises new insight into the role of mitochondria in tissue differentiation and patterning. In that regard, Barth syndrome has become an important field of investigations. It is the only human disease that is specifically caused by a defect in CL metabolism, more precisely in CL remodeling. Research in this area will help to further dissect the mechanism of CL remodeling, the biology of CL in general, and perhaps answer questions, such as how does CL deficiency lead to abnormal myocardial patterning? Conflict of Interest The authors declare that there are no conflicts of interest. Acknowledgements The author’s research has been supported by the Barth Syndrome Foundation, the National Institutes of Health, the United Mitochondrial Disease Foundation, and the American Heart Association. References [1] Schlame M, Rua D, Greenberg ML. The biosynthesis and functional role of cardiolipin. Progr Lipid Res 2000;39:257–88. [2] Gimeno RE, Cao J. Mammalian glycerol-3-phosphate acyltransferases: new genes for an old activity. J Lipid Res 2008;49:2079–88. [3] Prasad SS, Abhimanyu G, Agarwal AK. Enzymatic activities of the human AGPAT isoform 3 and isoform 5: Localization of AGPAT5 to mitochondria. J Lipid Res 2011;52:451–62. [4] Igal RA, Wang S, Gonzalez-Baro M, Coleman RA. Mitochondrial glycerol phosphate acyltransferase directs the incorporation of exogenous fatty acids into triacylglycerol. J Biol Chem 2001;276:42205–12. [5] Ohba Y, Sakuragi T, Kage-Nakadi E, Tomioka NH, Kono N, Imae R, et al. Mitochondria-type GPAT is required for mitochondrial fusion. EMBO J 2013;32:1265–79. [6] Chakraborty TR, Vancura A, Balija VS, Haldar D. Phosphatidic acid synthesis in mitochondria. Topography of formation and transmembrane migration. J Biol Chem 1999;274:29786–90. [7] Connerth M, Tatsuta T, Haag M, Klecker T, Westermann B, Langer T. Intramitochondrial transport of phosphatidic acid in yeast by a lipid transfer protein. Science 2012;338:815–8. [8] Schlame M, Haldar D. Cardiolipin is synthesized on the matrix side of the inner membrane in rat liver mitochondria. J Biol Chem 1993;268:74–9. [9] Rabl R, Soubannier V, Scholz R, Vogel F, Mendl N, Vasiljev-Neumeyer A, et al. Formation of cristae and cristae junctions in mitochondria depends on antagonism between Fcj1 and Su e/g. J Cell Biol 2009;185:1047–63. [10] Tamura Y, Harada Y, Nishikawa S, Yamano K, Kamiya M, Shiota T, et al. Tam41 is a CDP-diacylglycerol sunthase required for cardiolipin biosynthesis in mitochondria. Cell Metab 2013;17:709–18.

13

[11] Tamura Y, Harada Y, Yamano K, Watanabe K, Ishikawa D, Ohshima C, et al. Identification of Tam41 maintaining integrity of the TIM23 protein translocator complex in mitochondria. J Cell Biol 2006;174:631–7. [12] Gallas MR, Dienhart MK, Stuart RA, Long RM. Characterization of Mmp37p, a Saccharomyces cerevisiae mitochondrial matrix protein with a role in mitochondrial protein import. Mol Biol Cell 2006;17:4051–62. [13] Kutik S, Rissler M, Guan XL, Guiard B, Shui G, Gebert N, et al. The translocator maintenance protein Tam41 is required for mitochondrial cardiolipin biosynthesis. J Cell Biol 2008;183:1213–21. [14] Chang SC, Heacock PN, Clancey CJ, Dowhan W. The PEL1 gene (renamed PGS1) encodes phosphatidylglycero-phosphate synthase of Saccharomyces verevisiae. J Biol Chem 1998;273:9829–36. [15] He Q, Greenberg ML. Post-translational regulation of phosphatidylglycerophosphate synthase in response to inositol. Mol Microbiol 2004;53:1243–9. [16] Osman C, Haag M, Wieland FT, Brügger B, Langer T. A mitochondrial phosphatase required for cardiolipin biosynthesis: the PGP phosphatase Gep4. EMBO J 2010;29:1976–87. [17] Zhang J, Guan Z, Murphy AN, Wiley SE, Perkins GA, Worby CA, et al. Mitochondrial phosphatase PTPMT1 is essential for cardiolipin biosynthesis. Cell Metab 2011;13:690–700. [18] Jiang F, Rizavi HS, Greenberg ML. Cardiolipin is not essential for the growth of Saccharomyces cerevisiae on fermentable or non-fermentable carbon sources. Mol Microbiol 1997;26:481–91. [19] Chang SC, Heacock PN, Mileykovskaya E, Voelker DR, Dowhan W. Isolation and characterization of the gene (CLS1) encoding cardiolipin synthase in Saccharomyces cerevisiae. J Biol Chem 1998;273:14933–41. [20] Tuller G, Hrastnik C, Achleitner G, Schiefthaler U, Klein F, Daum G. YDL142c encodes cardiolipin synthase (Cls1p) and is non-essential for aerobic growth of Saccharomyces cerevisiae. FEBS Lett 1998;421:15–8. [21] Chen D, Zhang XY, Shi Y. Identification and functional characterization of hCLS1, a human cardiolipin synthase localized in mitochondria. Biochem J 2006;398:169–76. [22] Lu B, Xu FY, Jiang YJ, Choy PC, Hatch GM, Grunfeld C, et al. Cloning and characterization of a cDNA encoding human cardiolipin synthase (hCLS1). J Lipid Res 2006;47:1140–5. [23] Houtkooper RH, Akbari H, van Lenthe H, Kulik W, Wanders RJA, Frentzen M, et al. Identification and characterization of human cardiolipin synthase. FEBS Lett 2006;580:3059–64. [24] Katayama K, Sakurai I, Wada H. Identification of an Arabidopsis thaliana gene for cardiolipin synthase located in mitochondria. FEBS Lett 2004;577: 193–8. [25] Nowicki M, Müller F, Frentzen M. Cardiolipin synthase of Arabidopsis thaliana. FEBS Lett 2005;579:2161–5. [26] Schlame M. Cardiolipin synthesis for the assembly of bacterial and mitochondrial membranes. J Lipid Res 2008;49:1607–20. [27] Tan BK, Bogdanov M, Zhao J, Dowhan W, Raetz CRH, Guan Z. Discovery of a cardiolipin synthase utilizing phosphatidylethanolamine and phosphatidylglycerol as substrates. Proc Natl Acad Sci USA 2012;109:16504–9. [28] Serricchio M, Bütikofer P. An essential bacterial-type cardiolipin synthase mediates cardiolipin formation in a eukaryote. Proc Natl Acad Sci USA 2012;109:E954–61. [29] Lykidis A. Comparative genomics and evolution of eukaryotic phospholipid biosynthesis. Progr Lipid Res 2007;46:171–99. [30] Tian HF, Feng JM, Wen JF. The evolution of cardiolipin biosynthesis and maturation pathways and its implications for the evolution of eukaryotes. BMC Evolut Biol 2012;12:32. [31] Schlame M. Cardiolipin remodeling and the function of tafazzin. Biochim Biophys Acta 2013;1831:582–8. [32] Xu Y, Malhotra A, Ren M, Schlame M. The enzymatic function of tafazzin. J Biol Chem 2006;281:39217–24. [33] Claypool SM, Koehler CM. The complexity of cardiolipin in health and disease. Trends Biochem Sci 2012;37:32–41. [34] Schlame M, Acehan D, Berno B, Xu Y, Valvo S, Ren M, et al. The physical state of lipid substrates provides transacylation specificity for tafazzin. Nat Chem Biol 2012;8:862–9. [35] Beranek A, Rechberger G, Knauer H, Wolinski H, Kohlwein SD, Leber R. Identification of a cardiolipin-specific phospholipase encoded by the gene CLD1 (YGR110W) in yeast. J Biol Chem 2009;284:11572–8. [36] Cao J, Liu Y, Lockwood J, Burn P, Shi Y. A novel cardiolipin-remodeling pathway revealed by a gene encoding an endoplasmic reticulum-associated acyl-CoA:lysocardiolipin acyltransferase (ALCAT1) in mouse. J Biol Chem 2004;279:31727–34. [37] Taylor WA, Hatch GM. Identification of the human mitochondrial linoleoylcoenzyme A monolysocardiolipin acyltransferase (MLCL AT-1). J Biol Chem 2009;284:30360–71. [38] Baile MG, Sathappa M, Lu YW, Pryce E, Whited K, McCaffery JM, et al. Unremodeled and remodeled cardiolipin are functionally indistinguishable in yeast. J Biol Chem 2014;289:1768–78. [39] Chenval D, Müller M, Toni R, Ruetz S, Carafoli E. Adriamycin as a probe for the transversal distribution of cardiolipin in the inner mitochondrial membrane. J Biol Chem 1985;260:13003–7. [40] Gallet F, Petit JM, Maftah A, Zachowski A, Julien R. Asymmetrical distribution of cardiolipin in yeast inner mitochondrial membrane triggered by carbon catabolite repression. Biochem J 1997;324:627–34.

14

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

[41] Petit JM, Huet O, Gallet PF, Maftah A, Ratinaud MH, Julien R. Direct analysis and significance of cardiolipin transverse distribution in mitochondrial inner membranes. Eur J Biochem 1994;220:871–9. [42] Hovius R, Thijssen J, van der Linden P, Nicolay K, de Kruijff B. Phospholipid asymmetry of the outer membrane of rat liver mitochondria. Evidence for the presence of cardiolipin on the outside of the outer membrane. FEBS Lett 1993;330:71–6. [43] Gonzalvez F, Schug ZT, Houtkooper RH, MacKenzie ED, Brooks DG, Wanders RJA, et al. Cardiolipin provides an essential activating platform for caspase-8 on mitochondria. J Cell Biol 2008;183:681–96. [44] Gebert N, Joshi AS, Kutk S, Becker T, McKenzie M, Guan XL, et al. Mitochondrial cardiolipin involved in outer-membrane protein biogenesis: implications for Barth syndrome. Curr Biol 2009;19:2133–9. [45] Claypool SM, McCaffery JM, Koehler CM. Mitochondrial mislocalization and altered assembly of a cluster of Barth syndrome mutant tafazzins. J Cell Biol 2006;174:379–90. [46] Xu Y, Schlame M. The turnover of glycerol and acyl moieties of cardiolipin. Chem Phys Lipids 2014;179:17–24. [47] Ye C, Lou W, Li Y, Chatzispyrou IA, Huttemann M, Lee I, et al. Deletion of the cardiolipin-specific phospholipase Cld1 rescues growth and lifespan defects in the tafazzin mutant: Implications for Barth syndrome. J Biol Chem 2014;289:3114–25. [48] Krebs JJ, Hauser H, Carafoli E. Asymmetric distribution of phospholipids in the inner membrane of beef heart mitochondria. J Biol Chem 1979;254(12 (June)):5308–16. [49] Harb JS, Comte J, Gautheron DC. Asymmetrical orientation of phospholipids and their interactions with marker enzymes in pig heart mitochondrial inner membrane. Arch Biochem Biophys 1981;208(1 (April)):305–18. [50] Klingenberg M. Cardiolipin and mitochondrial carriers. Biochim Biophys Acta 2009;1788(10 (October)):2048–58. [51] Cullis PR, Verkleij AJ, Ververgaert PHJT. Polymorphic phase behavior of cardiolipin as detected by 31P-NMR and freeze-fracture techniques – effects of calcium, dibucaine, and chlorpromazine. Biochim Biophys Acta 1978;513:11–20. [52] Van Venetië R, Verkleij AJ. Possible role of non-bilayer lipids in the structure of mitochondria. A freeze-fracture electron microscopy study. Biochim Biophys Acta 1982;692(3):397–405. [53] Robertson JD. The molecular structure and contact relationships of cell membranes. Prog Biophys Mol Biol 1960;10:343–418. [54] Csordás G, Renken C, Várnai P, Walter L, Weaver D, Buttle KF, et al. Structural and functional features and significance of the physical linkage between ER and mitochondria. J Cell Biol 2006;174:915–21. [55] Mannella CA. Structural diversity of mitochondria: functional implications. Ann N Y Acad Sci 2008;1147:171–9. [56] Ardail D, Lerme F, Louisot P. Involvement of contact sites in phosphatidylserine import into liver mitochondria. J Biol Chem 1991;266(13):7978–81. [57] Simbeni R, Pon L, Zinser E, Paltauf F, Daum G. Mitochondrial membrane contact sites of yeast. Characterization of lipid components and possible incolvement in intramitochondrial translocation of phospholipids. J Biol Chem 1991;266:10047–9. [58] Ardail D, Privat JP, Egret-Chalier M, Levrat C, Lerme F, Louisot PJ. Mitochondrial contact sites. Lipid composition and dynamics. J Biol Chem 1990;265:18797–802. [59] Schlattner U, Tokarska-Schlattner M, Ramirez S, Brückner A, Kay L, Polge C, et al. Mitochondrial kinases and their molecular interaction with cardiolipin. Biochim Biophys Acta 2009;1788:2032–47. [60] van der Laan M, Bohnert M, Wiedemann N, Pfanner N. Role of MINOS in mitochondrial membrane architecture and biogenesis. Trends Cell Biol 2012;22:185–92. [61] Li S, Rousseau D. ATAD3, a vital membrane bound mitochondrial ATPase involved in tumor progression. J Bioenerg Biomembr 2012;44:189–97. [62] Schlattner U, Tokarska-Schlattner M, Rousseau D, Boissan M, Mannella C, Epand R, et al. Mitochondrial cardiolipin/phospholipid trafficking: the role of membrane contact site complexes and lipid transfer proteins. Chem Phys Lipids 2014;179:32–41. [63] Tatsuta T, Scharwey M, Langer T. Mitochondrial lipid trafficking. Trends Cell Biol 2014;24:44–52. [64] Schlattner U, Tokarska-Schlattner M, Ramirez S, Tyurina YY, Amoscato AA, Mohammadsanyi D, et al. Dual function of mitochondrial Nm23-H4 in phosphotransfer and intermembrane lipid transfer: a cardiolipin-dependent switch. J Biol Chem 2013;288:111–21. [65] Epand RF, Schlattner U, Wallimann T, Lacombe ML, Epand RM. Novel lipid transfer property of two mitochondrial proteins that bridge the inner and outer membranes. Biophys J 2007;92:126–37. [66] Tokarska-Schlattner M, Boissan M, Munier A, Borot C, Mailleau C, Speer O, et al. The nucleoside diphosphate kinase D (NM23-H4) binds the inner mitochondrial membrane with high affinity to cardiolipin and couples nucleotide transfer with respiration. J Biol Chem 2008;283:26198–207. [67] Schlattner U, Wallimann T. Octamers of mitochondrial creatine kinase isoenzymes differ in stability and membrane binding. J Biol Chem 2000;275:17314–20. [68] Eder M, Fritz-Wolf K, Kabsch W, Wallimann T, Schlattner U. Crystal structure of human ubiquitous mitochondrial creatine kinase. Proteins 2000;39:216–25.

[69] Lacombe ML, Tokarska-Schlattner M, Epand RF, Boissan M, Epand RM, Schlattner U. Interaction of NDPK-D with cardiolipin-containing membranes: structural basis and implications for mitochondrial physiology. Biochimie 2009;91:779–83. [70] Weber TA, Koob S, Heide H, Wittig I, Head B, van der Bliek A, et al. APOOL is a cardiolipin-binding constituent of the mitofilin/MINOS protein complex determining cristae morphology in mammalian mitochondria. PLOS One 2013;8:e63683. [71] Hoppins S, Collins SR, Cassidy-Stone A, Hummel E, Devay RM, Lackner LL, et al. A mitochondrial-focused genetic interaction map reveals a scaffold-like complex required for inner membrane organization in mitochondria. J Cell Biol 2011;195:323–40. [72] Harner M, Körner C, Walther D, Mokranjac D, Kaesmacher J, Welsch U, et al. The mitochondrial contact site complex, a determinant of mitochondrial architecture. EMBO J 2011;30:4356–70. [73] Gilquin B, Cannon BR, Hubstenberger A, Moulouel B, Falk E, Merle N, et al. The calcium-dependent interaction between S100B and the mitochondrial AAA ATPase ATAD3A and the role of this complex in the cytoplasmic processing of ATAD3A. Mol Cell Biol 2010;30:2724–36. [74] Gilquin B, Taillebourg E, Cherradi N, Hubstenberger A, Gay O, Merle N, et al. The AAA+ ATPase ATAD3A controls mitochondrial dynamics at the interface of the inner and outer membranes. Mol Cell Biol 2010;30:1984–96. [75] Hoffmann M, Bellance N, Rossignol R, Koopman WJ, Willems PH, Mayatepek E, et al. C. elegans ATAD-3 is essential for mitochondrial activity and development. PLoS One 2009;4:e7644. [76] He J, Cooper HM, Reyes A, Di Re M, Sembongi H, Litwin TR, et al. Mitochondrial nucleoid interacting proteins support mitochondrial protein synthesis. Nucleic Acids Res 2012;40:6109–21. [77] Rone MB, Midzak AS, Issop L, Rammouz G, Jagannathan S, Fan J, et al. Identification of a dynamic mitochondrial protein complex driving cholesterol import, trafficking, and metabolism to steroid hormones. Mol Endocrinol 2012;26:1868–82. [78] Issop L, Rone MB, Papadopoulos V. Organelle plasticity and interactions in cholesterol transport and steroid biosynthesis. Mol Cell Endocrinol 2013;371:34–46. [79] Heit B, Yeung T, Grinstein S. Changes in mitochondrial surface charge mediate recruitment of signaling molecules during apoptosis. Am J Physiol Cell Physiol. 2011;300:C33–41. [80] Chu CT, Ji J, Dagda RK, Jiang JF, Tyurina YY, Kapralov AA, et al. Cardiolipin externalization to the outer mitochondrial membrane acts as an elimination signal for mitophagy in neuronal cells. Nat Cell Biol 2013;15:1197–207. [81] Li J, Romestaing C, Han X, Li Y, Hao X, Wu Y, et al. Cardiolipin remodeling by ALCAT1 links oxidative stress and mitochondrial dysfunction to obesity. Cell Metab 2010;12:154–65. [82] Liu J, Dai Q, Chen J, Durrant D, Freeman A, Liu T, et al. Phospholipid scramblase 3 controls mitochondrial structure, function, and apoptotic response. Mol Cancer Res 2003;1:892–902. [83] Van Q, Liu J, Lu B, Feingold KR, Shi Y, Lee RM, et al. Phospholipid scramblase-3 regulates cardiolipin de novo biosynthesis and its resynthesis in growing HeLa cells. Biochem J 2007;401:103–9. [84] Sahu SK, Gummadi SN, Manoj N, Aradhyam GK. Phospholipid scramblases: an overview. Arch Biochem Biophys 2007;462:103–14. [85] Bevers EM, Williamson PL. Phospholipid scramblase: an update. FEBS Lett 2010;584:2724–30. [86] Inuzuka T, Inokawa A, Chen C, Kizu K, Narita H, Shibata H, et al. ALG-2interacting Tubby-like protein superfamily member PLSCR3 is secreted by an exosomal pathway and taken up by recipient cultured cells. Biosci Rep 2013;33:e00026. [87] Raemy E, Martinou JC. Involvement of cardiolipin in tBID-induced activation of BAX during apoptosis. Chem Phys Lipids 2014;179:70–4. [88] Rostovtseva TK, Bezrukov SM. VDAC regulation: role of cytosolic proteins and mitochondrial lipids. J Bioenerg Biomembr 2008;40:163–70. [89] Zinser E, Sperka-Gottlieb CD, Fasch EV, Kohlwein SD, Paltauf F, Daum G. Phospholipid synthesis and lipid composition of subcellular membranes in the unicellular eukaryote Saccharomyces cerevisiae. J Bacteriol 1991;173:2026–34. [90] Wriessnegger T, Gübitz G, Leitner E, Ingolic E, Cregg J, de la Cruz BJ, et al. Lipid composition of peroxisomes from the yeast Pichia pastoris grown on different carbon sources. Biochim Biophys Acta 2007;1771:455–61. [91] Sorice M, Circella A, Misasi R, Pittoni V, Garofalo T, Cirelli A, et al. Cardiolipin on the surface of apoptotic cells as possible trigger for antiphospholipid antibodies. Clin Exp Immunol 2000;122:277–84. [92] Sorice M, Circella A, Cristea IM, Garofalo T, Di Renzo L, Alessandri C, et al. Cardiolipin and its metabolites move from mitochondria to other cellular membranes during death receptor-mediated apoptosis. Cell Death Differ 2004;11:1133–45. [93] Neuspiel M, Schauss AC, Braschi E, Zunino R, Rippstein P, Rachubinski RA, et al. Cargo-selected transport from the mitochondria to peroxisomes is mediated by vesicular carriers. Curr Biol 2008;18:102–8. [94] Soubannier V, McLelland GL, Zunino R, Braschi E, Rippstein P, Fon EA, et al. A vesicular transport pathway shuttles cargo from mitochondria to lysosomes. Curr Biol 2012;22:135–41. [95] Landriscina C, Megli FM, Quagliariello E. Turnover of fatty acids in rat liver cardiolipin: comparison with other mitochondrial phospholipids. Lipids 1976;11:61–6.

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16 [96] Kagan VE, Tyurin VA, Jiang J, Tyurina YY, Ritov VB, Amoscato AA, et al. Cytochrome c acts as a cardiolipin oxygenase required for release of proapoptotic factors. Nat Chem Biol 2005;1:223–32. [97] Gonzalvez F, Gottlieb E. Cardiolipin: setting the beat of apoptosis. Apoptosis 2007;12:877–85. [98] Baile MG, Whited K, Claypool SM. Deacylation on the matrix side of the mitochondrial inner membrane regulates cardiolipin remodeling. Mol Biol Cell 2013;24:2008–20. [99] Gadd ME, Broekemeier KM, Crouser ED, Kumar J, Graff G, Pfeiffer DR. Mitochondrial iPLA2 activity modulates the release of cytochrome c from mitochondria and influences the permeability transition. J Biol Chem 2006;281:6931–9. [100] Seleznev K, Zhao C, Zhang XH, Song K, Ma ZA. Calcium-independent phospholipase A2 localizes in and protects mitochondria during apoptotic induction by staurosporine. J Biol Chem 2006;281:22275–88. [101] Malhotra A, Edelman-Novemsky I, Xu Y, Plesken H, Ma J, Schlame M, et al. Role of calcium-independent phopholipase A2 in the pathogenesis of Barth syndrome. Proc Natl Acad Sci 2009;106:2337–41. [102] Zhao Z, Zhang X, Zhao C, Choi J, Shi J, Song K, et al. Protection of pancreatic beta-cells by group VIA phospholipase A(2)-mediated repair of mitochondrial membrane peroxidation. Endocrinology 2010;151:3038–48. [103] Hsu YH, Dumlao DS, Cao J, Dennis EA. Assessing phospholipase A2 activity toward cardiolipin by mass spectrometry. PLoS One 2013;8:e59267. [104] Mancuso DJ, Sims HF, Han X, Jenkins CM, Guan SP, Yang K, et al. Genetic ablation of calcium-independent phospholipase A2gamma leads to alterations in mitochondrial lipid metabolism and function resulting in a deficient mitochondrial bioenergetic phenotype. J Biol Chem 2007;282:34611–22. [105] Mancuso DJ, Kotzbauer P, Wozniak DF, Sims HF, Jenkins CM, Guan S, et al. Genetic ablation of calcium-independent phospholipase A2{gamma} leads to alterations in hippocampal cardiolipin content and molecular species distribution, mitochondrial degeneration, autophagy, and cognitive dysfunction. J Biol Chem 2009;284:35632–44. [106] Zachman DK, Chicco AJ, McCune SA, Murphy RC, Moore RL, Sparagna GC. The role of calcium-independent phospholipase A2 in cardiolipin remodeling in the spontaneously hypertensive heart failure rat heart. J Lipid Res 2010;51:525–34. [107] Buckland AG, Kinkaid AR, Wilton DC. Cardiolipin hydrolysis by human phospholipases A2. The multiple enzymatic activities of human cytosolic phospholipase A2. Biochim Biophys Acta 1998;1390:65–72. [108] Choi SY, Huang P, Jenkins GM, Chan DC, Schiller J, Frohman MA. A common lipid links Mfn-mediated mitochondrial fusion and SNARE-regulated exocytosis. Nat Cell Biol 2006;8:1255–62. [109] Ipsaro JJ, Haase AD, Knott SR, Joshua-Tor L, Hannon GJ. The structural biochemistry of Zucchini implicates it as a nuclease in piRNA biogenesis. Nature 2012;491:279–83. [110] Nishimasu H, Ishizu H, Saito K, Fukuhara S, Kamatani MK, Bonnefond L, et al. Structure and function of Zucchini endoribonuclease in piRNA biogenesis. Nature 2012;491:284–7. [111] Voigt F, Reuter M, Kasaruho A, Schulz EC, Pillai RS, Barabas O. Crystal structure of the primary piRNA biogenesis factor Zucchini reveals similarity to the bacterial PLD endonuclease Nuc. RNA 2012;18:2128–34. [112] Tyurina YY, Kini V, Tyurin VA, Vlasova II, Jiang J, Kapralov AA, et al. Mechanisms of cardiolipin oxidation by cytochrome c: relevance to pro- and antiapoptotic functions of etoposide. Mol Pharmacol 2006;70:706–17. [113] Kagan VE, Bayir HA, Belikova NA, Kapralov O, Tyurina YY, Tyurin VA, et al. Cytochrome c/cardiolipin relations in mitochondria: a kiss of death. Free Radic Biol Med 2009;46:1439–53. [114] Kapralov AA, Kurnikov IV, Vlasova II, Belikova NA, Tyurin VA, Basova LV, et al. The hierarchy of structural transitions induced in cytochrome c by anionic phospholipids determines its peroxidase activation and selective peroxidation during apoptosis in cells. Biochemistry 2007;46:14232–44. [115] Muenzner J, Toffey JR, Hong Y, Pletneva EV. Becoming a peroxidase: cardiolipin-induced unfolding of cytochrome c. J Phys Chem B 2013;117:12878–86. [116] Belikova NA, Tyurina YY, Borisenko G, Tyurin V, Samhan Arias AK, Yanamala N, et al. Heterolytic reduction of fatty acid hydroperoxides by cytochrome c/ cardiolipin complexes: antioxidant function in mitochondria. J Am Chem Soc 2009;131:11288–9. [117] Belikova NA, Jiang J, Tyurina YY, Zhao Q, Epperly MW, Greenberger J, et al. Cardiolipin-specific peroxidase reactions of cytochrome C in mitochondria during irradiation-induced apoptosis. Int J Radiat Oncol Biol Phys 2007;69:176–86. [118] Paradies G, Petrosillo G, Paradies V, Ruggiero FM. Role of cardiolipin peroxidation and Ca2+ in mitochondrial dysfunction and disease. Cell Calcium 2009;45:643–50. [119] Lewis RN, McElhaney RN. The physicochemical properties of cardiolipin bilayers and cardiolipin-containing lipid membranes. Biochim Biophys Acta 2009;1788:2069–79. [120] Lewis RN, Zweytick D, Pabst G, Lohner K, McElhaney RN. Calorimetric, x-ray diffraction, and spectroscopic studies of the thermotropic phase behavior and organization of tetramyristoyl cardiolipin membranes. Biophys J 2007;92:3166–77. [121] Allegrini PR, Pluschke G, Seelig J. Cardiolipin conformation and dynamics in bilayer membranes as seen by deuterium magnetic resonance. Biochemistry 1984;23:6452–8.

15

[122] Powell GL, Jacobus J. The nonequivalence of the phosphorous atoms in cardiolipin. Biochemistry 1974;13:4024–6. [123] Schlame M, Ren M, Xu Y, Greenberg ML, Haller I. Molecular symmetry in mitochondrial cardiolipins. Chem Phys Lipids 2005;138:38–49. [124] Renner LD, Weibel DB. Cardiolipin microdomains localize to negatively curved regions of Escherichia coli membranes. Proc Natl Acad Sci USA 2011;108:6264–9. [125] Seddon JM, Kaye RD, Marsh D. Induction of lamellar-inverted hexagonal phase-transition in cardiolipin by protons and monovalent cations. Biochim Biophys Acta 1983;734:347–52. [126] Sankaram MB, Powell GL, Marsh D. Effect of acyl chain composition on saltinduced lamellar to inverted hexagonal phase transitions in cardiolipin. Biochim Biophys Acta 1989;980:389–92. [127] Rand RP, Sengupta S. Cardiolipin forms hexagonal structures with divalent cations. Biochim Biophys Acta 1972;225:484–92. [128] De Kruijff B, Cullis PR. Cytochrome-c specifically induces nonlamellar structures in cardiolipin-containing membranes. Biochim Biophys Acta 1980;602:477–90. [129] Lewis RN, McElhaney RN. Surface charge markedly attenuates the nonlamellar phase-forming propensities of lipid bilayer membranes: Calorimetric and 31P-nuclear magnetic resonance studies of mixtures of cationic, anionic, and zwitterionic lipids. Biophys J 2000;79:1455–64. [130] Tarahovsky YS, Arsenault AL, MacDonald RC, McIntosh TJ, Epand RM. Electrostatic control of phospholipid polymorphism. Biophys J 2000;79:3193–200. [131] Powell GL, Marsh D. Polymorphic phase behavior of cardiolipin derivatives studied by 31P-NMR and X-ray diffraction. Biochemistry 1985;24:2902–8. [132] Dahlberg M. Polymorphic phase behavior of cardiolipin derivatives studied by coarse-grained molecular dynamics. J Phys Chem B 2007;111:7194–200. [133] Etienne F, Roche Y, Peretti P, Bernard S. Cardiolipin packing ability studied by grazing incidence X-ray diffraction. Chem Phys Lipids 2008;152:13–23. [134] Nichols-Smith S, The SY, Kuhl TL. Thermodynamic and mechanical properties of model mitochondrial membranes. Biochim Biophys Acta 2004;1663:82–8. [135] Frias M, Benesch MG, Lewis RN, McElhaney RN. On the miscibility of cardiolipin with 1,2-diacyl phosphoglycerides: binary mixtures of dimyristoylphosphatidylethanolamine and tetramyristoylcardiolipin. Biochim Biophys Acta 2011;1808:774–83. [136] Khalifat N, Puff N, Bonneau S, Fournier JB, Angelova M. Membrane deformation under local pH gradient: mimicking mitochondrial cristae dynamics. Biophys J 2008;95:4924–33. [137] Belikova NA, Vladimirov JA, Osipor AN, Kapralov AA, Tyurin VA, Potapovich MV, et al. Peroxidase activity and structural transitions of cytochrome c bound to cardiolipin-containing membranes. Biochemistry 2006;45:4998–5009. [138] Beyer K, Klingenberg M. ADP/ATP carrier protein from beef heart mitochondria has high amounts of tightly bound cardiolipin, as revealed by 31P nuclear magnetic resonance. Biochemistry 1985;24:3821–6. [139] Pebay-Peyroula E, Dahout-Gonzalez C, Kahn R, Trezeguet V, Lauquin GJ, Brandolin G. Structure of mitochondrial ADP/ATP carrier in complex with carboxyatractyloside. Nature 2003;426:39–44. [140] Nury H, Dahout-Gonzalez C, Trezeguet V, Lauquin G, Brandolin G, PebayPeyroula E. Structural basis for lipid-mediated interactions between mitochondrial ADP/ATP carrier monomers. FEBS Lett 2005;579:6031–6. [141] Ruprecht JJ, Hellawell AM, Harding M, Crichton PG, McCoy AJ, Kunji ERS. Structures of yeast mitochondrial ADP/ATP carriers support a domain-based alternating-access transport mechanism. Proc Natl Acad Sci USA 2014;111:E426–34. [142] Claypool SM, Oktay Y, Boontheung P, Loo JA, Koehler CM. Cardiolipin defines the interactome of the major ADP/ATP carrier protein of the mitochondrial inner membrane. J Cell Biol 2008;182:937–50. [143] Zhang M, Mileykovskaya E, Dowhan W. Gluing the respiratory chain together. Cardiolipin is required for supercomplex formation in the inner mitochondrial membrane. J Biol Chem 2002;277:43553–6. [144] DeVay RM, Dominguez-Ramirez L, Lackner LL, Hoppins S, Stahlberg H, Nunnari J. Coassembly of Mgm1 isoforms requires cardiolipin and mediates mitochondrial inner membrane fusion. J Cell Biol 2009;186:793–803. [145] Zhang M, Mileykovskaya E, Dowhan W. Cardiolipin is essential for organization of complexes III and IV into a supercomplex in intect yeast mitochondria. J Biol Chem 2005;280:29403–8. [146] Pfeiffer K, Gohil V, Stuart RA, Hunte C, Brandt U, Greenberg ML, et al. Cardiolipin stabilizes respiratory chain supercomplexes. J Biol Chem 2003;278:52873–80. [147] Bazán S, Mileykovskaya E, Mallampalli VK, Heacock P, Sparagna GC, Dowhan W. Cardiolipin-dependent reconstitution of respiratory supercomplexes from purified Saccharomyces cerevisiae complexes III and IV. J Biol Chem 2013;288:401–11. [148] Mileykovskaya E, Penczek PA, Fang J, Mallampalli VK, Sparagna GC, Dowhan W. Arrangement of the respiratory chain complexes in Saccharomyces cerevisiae supercomplex III2IV2 revealed by single particle cryo-electron microscopy. J Biol Chem 2012;287:23095–103. [149] Althoff T, Mills DJ, Popot JL, Kühlbrandt W. Arrangement of electron transport chain components in bovine mitochondrial supercomplex I1III2IV1. EMBO J 2011;30:4652–64. [150] Strauss M, Hofhaus G, Schröder RR, Kühlbrandt W. Dimer ribbons of ATP synthase shape the inner mitochondrial membrane. EMBO J 2008;27:1154–60.

16

M. Ren et al. / Progress in Lipid Research 55 (2014) 1–16

[151] Acehan D, Malhotra A, Xu Y, Ren M, Stokes DL, Schlame M. Cardiolipin affects the supramolecular organization of ATP synthase in mitochondria. Biophys J 2011;100:2184–92. [152] Lange C, Nett JH, Trumpower BL, Hunte C. Specific roles of protein– phospholipid interactions in the yeast cytochrome bc1 complex structure. EMBO J 2001;20:6591–600. [153] Shinzawa-Itoh K, Aoyama H, Muramoto K, Terada H, Kurauchi T, Tadehara Y, et al. Structures and physiological roles of 13 integral lipids of bovine heart cytochrome c oxidase. EMBO J 2007;26:1713–25. [154] Eble KS, Coleman WB, Hantgan RR, Cunningham CC. Tightly associated cardiolipin in the bovine heart mitochondrial ATP synthase as analyzed by 31 P nuclear magnetic resonance spectroscopy. J Biol Chem 1990;265:19434–40. [155] Sakamoto T, Inoue T, Otomo Y, Yokomori N, Ohno M, Arai H, et al. Deficiency of cardiolipin synthase causes abnormal mitochondrial function and morphology in germ cells of Caenorhabditis elegans. J Biol Chem 2012;287:4590–601. [156] Pineau B, Bourge M, Marion J, Mauve C, Gilard F, Maneta-Peyret L, et al. The importance of cardiolipin synthase for mitochondrial ultrastructure, respiratory function, plant development, and stress responses in Arabidopsis. Plant Cell 2013;25:4195–208. [157] Khuchua Z, Yue Z, Batts L, Strauss AW. A zebrafish model of human Barth syndrome reveals the essential role of tafazzin in cardiac development and function. Circ Res 2006;99:201–8. [158] Xu Y, Condell M, Plesken H, Edelman-Novemsky I, Ma J, Ren M, et al. A Drosophila model of Barth syndrome. Proc Natl Acad Sci 2006;103:11584–8. [159] Ma L, Vaz FM, Gu Z, Wanders RJ, Greenberg ML. The human TAZ gene complements mitochondrial dysfunction in the yeast taz1D mutant. Implications for Barth syndrome. J Biol Chem 2004;279:44394–9. [160] Acehan D, Khuchua Z, Houtkooper RH, Malhotra A, Kaufman J, Vaz FM, et al. Distinct effects of tafazzin deletion in differentiated and undifferentiated mitochondria. Mitochondrion 2009;9:86–95. [161] Curtis NJ, Ringo JM, Dowse HB. Morphology of the pupal heart, adult heart, and associated tissues in the fruit fly, Drosophila melanogaster. J Morphol 1999;240:225–35. [162] Wolf MJ, Amrein H, Izatt JA, Choma MA, Reedy MC, Rockman HA. Drosophila as a model for the identification of genes causing adult human heart disease. Proc Natl Acad Sci USA 2006;103:1394–9. [163] Lehmacher C, Abeln B, Paululat A. The ultrastructure of Drosophila heart cells. Arthropod Struct Dev 2012;41:459–74. [164] Acehan D, Vaz F, Houtkooper RH, James J, Moore V, Tokunaga C, et al. Cardiac and skeletal muscle defects in a mouse model of human Barth syndrome. J Biol Chem 2011;286:899–908. [165] Soustek MS, Falk D, Mah C, Toth M, Schlame M, Lewin A, et al. Characterization of a transgenic shRNA induced murine model of tafazzin deficiency. Hum Gene Therapy 2011;22:865–71. [166] Powers C, Huang Y, Strauss A, Khuchua Z. Diminished exercise capacity and mitochondrial bc1 complex deficiency in tafazzin-knockdown mice. Front Physiol 2013;4:74. [167] Phoon CKL, Acehan D, Schlame M, Stokes DL, Edelman-Novemsky I, Yu D, et al. Tafazzin knockdown in mice leads to a developmental cardiomyopathy with early diastolic dysfunction preceding myocardial noncompaction. J Am Heart Assoc 2012;1. http://dx.doi.org/10.1161/JAHA.111.000455. [168] Schlame M, Ren M. Barth syndrome, a human disorder of cardiolipin metabolism. FEBS Lett 2006;580:5450–5. [169] Barth PG, Scholte HR, Berden JA, Van der Klei-Van Moorsel JM, Muyt-Houwen IE, Van’t Veer-Korthof ET, et al. An X-linked mitochondrial disease affecting cardiac muscle, skeletal muscle and neutrophil leucocytes. J Neurol 1983;62:327–55. [170] Kelley RI, Cheatham JP, Clark BJ, Nigro MA, Powell BR, Sherwood GW, et al. Xlinked dilated cardiomyopathy with neutropenia, growth-retardation, and 3methylglutaconic aciduria. J Pediatr 1991;119:738–47. [171] Spencer CT, Bryant RM, Day J, Gonzalez IL, Colan SD, Thompson WR, et al. Cardiac and clinical phenotype in Barth syndrome. Pediatrics 2006;118:e337. [172] Steward CG, Newbury-Ecob RA, Hastings R, Smithson SF, Tsai-Goodman B, Quarrell OW, et al. Barth syndrome: an X-linked cause of fetal cardiomyopathy and stillbirth. Prenatal Diagn 2010;30:970–6. [173] Scaglia F, Towbin JA, Craigen WJ, Belmont JW, Smith EO, Neish SR, et al. Clinical spectrum, morbidity, and mortality in 113 pediatric patients with mitochondrial disease. Pediatrics 2004;114:925–31. [174] Clarke SLN, Bowron A, Gonzalez IL, Groves SJ, Newbury-Ecob R, Clayton N, et al. Barth syndrome. Orphanet J Rare Dis 2013;8:23. [175] Jefferies JL. Barth syndrome. Am J Med Genet C Semin Med Genet 2013;163C:198–205. [176] Ronvelia D, Greenwood J, Platt J, Hakim S, Zaragoza MV. Intrafamilial variability for novel TAZ gene mutation: Barth syndrome with dilated cardiomyopathy and heart failure in an infant and left ventricular noncompaction in his great-uncle. Mol Genet Metab 2012;107:428–32.

[177] Xu Y, Sutachan JJ, Plesken H, Kelley RI, Schlame M. Characterization of lymphoblast mitochondria from patients with Barth syndrome. Lab Invest 2005;85:823–30. [178] Acehan D, Xu Y, Stokes DL, Schlame M. Comparison of lymphoblast mitochondria from normal dubjects and patients with Barth syndrome using electron microscopic tomography. Lab Invest 2007;87:40–8. [179] He Q, Wang M, Harris N, Han X. Tafazzin knockdown interrupts cell cycle progression in cultured neonatal ventricular fibroblasts. Am J Physiol Heart Circ Physiol. 2013;305:H1332–43. [180] Valianpour F, Wanders RJ, Overmars H, Vaz FM, Barth PG, van Gennip AH. Cardiolipin deficiency in X-linked cardioskeletal myopathy and neutropenia (Barth syndrome, MIM 302060): a study in cultured skin fibroblasts. J Pediatr 2002;141:729–33. [181] Dudek J, Cheng I-F, Balleininger M, Vaz FM, Streckfuss-Bömeke K, Hübscher D, et al. Cardiolipin deficiency affects respiratory chain function and organization in an induced pluripotent stem cell model of Barth syndrome. Stem Cell Res 2013;11:806–19. [182] Ojala T, Polinati P, Manninen T, Hiippala A, Rajantie J, Karikoski R, et al. New mutation of mitochondrial DNAJC19 causing dilated and noncompaction cardiomyopathy, anemia, ataxia, and male genital anomalies. Pediatr Res 2012;72:432–7. [183] Chicco AJ, Sparagna GC. Role of cardiolipin alterations in mitochondrial dysfunction and disease. Am J Physiol Cell Physiol 2007;292:33–44. [184] Shi Y. Emerging roles of cardiolipin remodeling in mitochondrial dysfunction associated with diabetes, obesity, and cardiovascular diseases. J Biomed Res 2010;24:6–15. [185] Yin H, Zhu M. Free radical oxidation of cardiolipin: chemical mechanisms, detection and implication in apoptosis, mitochondrial dysfunction and human diseases. Free Rad Res 2012;46:959–74. [186] Han X, Yang J, Cheng H, Yang K, Abendschein DR, Gross RW. Shot-gun lipidomics identifies cardiolipin depletion in diabetic myocardium linking altered substrate utilization with mitochondrial dysfunction. Biochemistry 2005;44:16684–94. [187] Han X, Yang J, Yang K, Zhao Z, Abendschein DR, Gross RW. Alterations in myocardial cardiolipin content and composition occur at the very earliest stages of diabetes: a shotgun lipidomics study. Biochemistry 2007;46:6417–28. [188] Lewin TM, Schwerbrock NM, Lee DP, Coleman RA. Identification of a new glycerol-3-phosphate acyltransferase isoenzyme, mtGPAT2, in mitochondria. J Biol Chem 2004;279:13488–95. [189] Yet SF, Lee S, Hahm YT, Sul HS. Expression and identification of p90 as the murine mitochondrial glycerol-3-phosphate acyltransferase. Biochemistry 1993;32:9486–91. [190] Tamura Y, Onguka O, Aiken Hobbs AE, Jensen RE, Iijima M, Claypool SM, et al. Role of two conserved intermembrane space proteins, Ups1p and Ups2p, in intra-mitochondrial phospholipid trafficking. J Biol Chem 2012;287:15205–18. [191] Dzˇugasova V, Obernauerová, Horváthová K, Vachová M, Zˇáková M, Šubik J. Phosphatidylglycerolphosphate synthase encoded by the PEL1/PGS1 gene in Saccharomyces cerevisiae is localized in mitochondria and its expression is regulated by phospholipid precursors. Curr Genet 1998;34:297–302. [192] Zhong Q, Gohil VM, Ma L, Greenberg ML. Absence of cardiolipin results in temperature sensitivity, respiratory defects, and mitochondrial DNA instability independent of pet56. J Biol Chem 2004;279:32294–300. [193] Patil VA, Fox JL, Gohil VM, Winge DR, Greenberg ML. Loss of cardiolipin leads to perturbation of mitochondrial and cellular iron homeostasis. J Biol Chem 2013;288:1696–705. [194] Chen S, Liu D, Finley Jr RL, Greenberg ML. Loss of mitochondrial DNA in the yeast cardiolipin synthase crd1 mutant leads to up-regulation of the protein kinase Swe1p that regulates the G2/M transition. J Biol Chem 2010;285:10397–103407. [195] Gu Z, Valianpour F, Chen S, Vaz FM, Hakkaart GA, Wanders RJA, et al. Aberrant cardiolipin metabolism in the yeast taz1 mutant: a model for Barth syndrome. Mol Microbiol 2004;51:149–58. [196] Chen S, He Q, Greenberg ML. Loss of tafazzin in yeast leads to increased oxidative stress during respiratory growth. Mol Microbiol 2008;68:1061–72. [197] Ohtsuka T, Nishijima M, Suzuki K, Akamatsu Y. Mitochondrial dysfunction of a cultured Chinese hamster ovary cell mutant deficient in cardiolipin. J Biol Chem 1993;268:22914–9. [198] Kiebish MA, Yang K, Liu X, Mancuso DJ, Guan S, Zhao Z, et al. Dysfunctional cardiac mitochondrial bioenergetic, lipidomic, and signaling in a murine model of Barth syndrome. J Lipid Res 2013;54:1312–25. [199] Rigaud C, Lebre A-S, Touraine R, Beaupain B, Ottolenghi C, Chabli A, et al. Natural history of Barth syndrome: a national cohort study of 22 patients. Orphanet J Rare Dis 2013;8:70. [200] Yen TY, Hwu WL, Chien YH, Wu MH, Lin MT, Tsao LY, et al. Acute metabolic decompensation and sudden death in Barth syndrome: report of a family and a literature review. Eur J Pediatr 2008;167:941–4.