Specificity of the mammalian glycolipid transfer proteins

Chemistry and Physics of Lipids 194 (2016) 72–78

Contents lists available at ScienceDirect

Chemistry and Physics of Lipids journal homepage: www.elsevier.com/locate/chemphyslip

Specificity of the mammalian glycolipid transfer proteins Peter Mattjus Biochemistry, Faculty of Science and Engineering, Åbo Akademi University, Turku, Finland

A R T I C L E I N F O

A B S T R A C T

Article history: Received 16 May 2015 Received in revised form 8 July 2015 Accepted 27 July 2015 Available online 30 July 2015

Structurally the glycolipid transfer protein (GLTP) fold differs from other proteins that recognize glycolipids, such as non-specific lipid transfer proteins and lysosomal lipid degradation assisting proteins, even though they act on the same class of lipids. Proteins with glycan binding domains, such as lectins and pulmonary surfactant proteins share no structural similarity with the GLTP family either. Currently the unique GLTP-fold specific for binding glycosphingolipids is found only in the founding member GLTP and the phosphoinositol 4-phosphate adapter protein 2, FAPP2. FAPP2 was originally characterized as a member eight of the pleckstrin homology domain-containing family A (PLEKHA8). This review summarizes what is structurally required by the glycosphingolipids in order for them to be transported by the GLTPs. ã 2015 Elsevier Ireland Ltd. All rights reserved.

Keywords: Glycosphingolipid Lipid transfer Transfer protein

Raymond J. Metz when working on his thesis projects1, under the supervision of Norman S. Radin, discovered2 the ‘CUP’ protein in the late 70s (Metz and Radin, 1980). The project was a follow-up study stimulated by the earlier Dawson and Sweeley paper from 1970. Dawson and Sweeley (1970) reported glucosylceramide (GlcCer) exchange between red blood cells and plasma. The objective by Metz and Radin (1980) was to study if a protein could in fact have mediated the exchange of GlcCer seen by Dawson and Sweeley. They were indeed “dealing with a new protein”, because they found that purified bovine spleen cytosol contained a protein that in vitro was accelerating a GlcCer exchange between rat erythrocytes and liposomes. Now we know the CUP protein (cerebroside uptake protein) as GLTP (glycolipid transfer protein). The acronym GLTP was later given by the Japanese team of Akira Abe, Keiko Yamada and Terukatsu Sasaki from the Sapporo Medical College. They extensively continued the study of GLTP in the late 80s (Sasaki, 1985, 1990). GLTP accelerates glycosphingolipid (GSL) intervesicular movement in vitro at least a thousand-fold compared to spontaneous transfer (Correa-Freire et al., 1982; Wong et al., 1984; Mattjus, 2009). This review will summarize what is structurally required by the ‘lipid ligands’ in order for GLTP to see them as substrates, based on the different transfer assays and experimental setups published in the last 35 years. A handful of all the studies that today lay as a

E-mail address: pmattjus@abo.fi (P. Mattjus). The demonstration, purification, and characterization of a glucosylceramide transfer protein. Raymond J. Metz, thesis defense January 1, 1982, University of Michigan. 2 Personal email communication, October 2005.

foundation for the physical and biochemical knowledge that we have about all the members in the glycolipid transfer protein superfamily3 , have ties to the Robert Bittman laboratory. Different structural modifications to sphingolipids have served as tools for us to better understand how the GLTPs work at membrane interfaces (Brown and Mattjus, 2007; Mattjus, 2009; Tuuf and Mattjus, 2013; Malinina et al., 2015). Lipid transfer events were indeed also familiar grounds for Bob. One of his earliest works were sterol and cholesterol transfer between different membranes (Rottem et al., 1978; Bittman et al., 1983). 1. Methods used to study lipid transfer protein activity The in vitro transfer properties of GLTP have been studied using radionuclide and fluorescently labeled lipids. The early transfer assays usually employed natural membranes, that were sooner replaced by chain pure and structurally defined matrix lipids (Brown and Mattjus, 2007; Mattjus, 2009). We also have label-free approaches to study not only lipid transfer, but also lipid binding events (Locatelli-Hoops et al., 2006; Ohvo-Rekilä and Mattjus, 2011; D'Angelo et al., 2013). 1.1. Charged vesicle transfer assays and radiolabeled lipids In this assay, donor vesicles are constructed so that they contain a negatively charged phospholipid (PL) while the acceptor vesicles are neutral. The charged donors can be separated from the neutral

1

http://dx.doi.org/10.1016/j.chemphyslip.2015.07.018 0009-3084/ ã 2015 Elsevier Ireland Ltd. All rights reserved.

3

http://supfam.org ID 110004.

P. Mattjus / Chemistry and Physics of Lipids 194 (2016) 72–78

73

Fig. 1. Schematic illustration of the ‘free transporter’ mechanism of glycolipid transfer protein. GLTP partition onto the membrane surface in a non-specific, non-perturbing manner (K1), regardless of ligand lipid presence or not (Rao et al., 2004). At the membrane leaflet aqueous interface GLTP interacts with the membrane. One of the parts of GLTP that is known to interact with the membrane is shown in purple. GLTP scans the membrane until a GLTP-lipid complex is formed, Ks. The rate-limiting step in the transfer is the formation of GLTP-lipid complex and the desorption from the membrane interface. In the next step conformational changes to the GLTP membrane interaction domains presumably cause the complex to be released into the aqueous environment, K2. It is not known how the mechanism of the release of the glycolipid from GLTP into the acceptor membrane occurs. The GLTP (1WBE) and the membrane phospholipid molecules are rendered in the van der Waals volume (MacPyMol).

acceptor vesicles by ion exchange chromatography. Usually sonicated donor vesicles are used containing a charged PL like 5–10 mol % phosphatidic acid (PA), together with the radiolabeled glycolipid, usually 3H-labeled and in a couple of mole percentage. To correct for spill over and vesicle fusion of donors into the acceptor vesicle pool, a trace of a non-exchangeable, usually a 14Clabeled lipid, is incorporated into the donors. At desired time intervals, acceptors vesicles are separated from donor vesicles by elution over mini columns, and the eluate is analysed by scintillation counting. The amount of radioactivity in the different fractions corresponds to the transfer of glycolipids mediated by the transfer protein. 1.2. Brominated lipid transfer assays Brominated acyl chains create heavier PLs that sediment faster compared to natural lipids, yet they form structurally equivalent bilayer lipid vesicles. This property has been used to separate donor and acceptor vesicles in transfer protein experiments (Wong et al., 1984; Brown et al., 1985). This assay allows the recovery of both the donor and acceptor vesicles with the use of density gradients and is both sensitive and reproducible. Tritiated glycolipids and a nonexchangeable 14C-labeled lipid (such as cholesteryl oleate, or triolein) are incorporated into either non-brominated or brominated vesicles. The donor and acceptor vesicles are co-mixed and incubated with the transfer protein to be studied. The transfer reaction is terminated by fast centrifugation of the two vesicle populations over a step gradient in an Airfuge (Beckman). The separation needs to be done quickly and at high centrifugal forces, therefore the Airfuge has proven to be one of the only options available. The brominated vesicles and non-brominated vesicles are quantitatively well

separated, and the transferred labeled glycolipids are finally measured by scintillation counting. 1.3. Fluorescence transfer assays By far the most common lipid transfer assays use fluorescently labeled lipids. The setup usually involves a fluorophore pair, and are often called RET or FRET assays. In the donor vesicles one fluorophore, the energy donor, usually the transferred lipid is excited and its emission is quenched by the second fluorophore, the energy acceptor. The second fluorophore is a non-transferrable lipid that stays in the donor vesicles, at least well beyond the experimental time frame. Once the transfer protein starts to move the transferrable lipid to the acceptor vesicles, its emission is no longer quenched. The increase in emission as a function of time describes the transfer rate. Commonly used pairs are, BODIPY4 and DiO-C16 (Nylund and Mattjus, 2005), anthrylvinyl and perylenoyl (Mattjus et al., 1999), NBD and rhodamine (Nichols and Pagano, 1983). Other fluorescence assay setups have also been used to analyse different transfer proteins. For instance, taking advantage of excimer/monomer fluorescence such as for pyrene (Brown et al., 1985; Somerharju et al., 1987) or self quenching of NBD (Nichols and Pagano, 1981) or parinaric acid (Somerharju et al., 1981). 1.4. Fluorescence competition assay To analyse unlabeled lipids as potential substrates for transfer proteins a competition assay (Dansen et al., 1999) can be used

4 BODIPY, dipyrrometheneboron difluoride; DiO, 3,3-dioctadecyloxacarbocyanine perchlorate; NBD, 7-nitro-2-1,3-benzoxadiazol-4-yl.

74

P. Mattjus / Chemistry and Physics of Lipids 194 (2016) 72–78

(Edqvist et al., 2004; Viitanen et al., 2006). After a steady transfer rate has been reached, of for instance BODIPY-GlcCer by GLTP using a fluorescence assay described above, unlabeled lipids incorporated in vesicles are added to the transfer reaction mixture. This would result in a decrease in the transfer rate of the fluorescently labeled lipid, if the added lipids were competing with the labeled lipid. If the added lipids were not substrates for the transfer protein, no deviation in the slope of the transfer rate would occur (Dansen et al., 1999; West et al., 2008). This approach could also be used for many types of lipids, inhibitors or activators that can be screened in automated lipid transfer assay setups. 1.5. Surface plasmon resonance assay To be able to analyse the binding and transfer of unlabeled GSLs by GLTP we developed a surface plasmon resonance (SPR) assay (D'Angelo et al., 2007, 2013; Ohvo-Rekilä and Mattjus, 2011). The advantage of this technique is that no labeled lipids are needed, and only small amounts of transfer protein are required (LocatelliHoops et al., 2006). The SPR instrument measures the change in refractive index by an optical method and displays the response signal as resonance units versus time. In the assay setup the lipid ligands are in vesicles (few mole percentage in a matrix lipid) that are captured to the surface of the sensor chip. If the lipid transfer protein that is flowed over the immobilized vesicles binds to the vesicles the mass on the chip surface increases and this leads to an increase in the refractive index of the chip surface. This first part of the SPR curve describes the protein binding to the vesicles, and is of course also very useful information. Once the transfer protein starts to remove lipids from the vesicles the mass starts to decrease as a function of time. The second part of the SPR curve now describes the lipid transfer event (Locatelli-Hoops et al., 2006; Ohvo-Rekilä and Mattjus, 2011). This technique is useful for studies of membrane–protein interactions and lipid-transfer rates, and it can easily be adapted to other membrane-interacting proteins. 2. Membrane properties that affect the GLTP action The lipid transfer of GLTP includes several steps. The initial GLTP membrane binding, membrane scanning and the lipid uptake, followed by the subsequent extraction of the lipid from the membrane (Fig. 1). How the binding of GLTP to the destination membrane and the final release of the GSL to the acceptor membrane occur is still not known. Other GLTP binding mechanisms such as a ‘bound transporter’ model, in which GLTP remains continuously bound to donor vesicles and mediates transfer by lowering the energy barrier for selective GSL desorption, or a ‘conduit’ model, in which GLTP would form a connecting conduit between donor are not supported by any data so far (Rao et al., 2004, 2005). Extraction of GSLs by GLTP only happens from one of the leaflets of the bilayer membrane. The GSLs do not have to be imbedded in a bilayer leaflet in order for GLTP to be able to access and extract the lipid. GSLs present in a monolayer on an air water interface are indeed also recognized by GLTP (Sasaki and Demel, 1985; Nylund et al., 2007; Zhai et al., 2013). However, bilayers with curvature stress stimulate much faster GSL intervesicular transfer than nonstressed planar bilayers (Nylund et al., 2007). The transfer of GSLs from planar POPC monolayers is nearly non-existent, however interestingly pure phosphatidylethanolamine (PE) monolayers allow for transfer of BODIPY-GSLs by GLTP at both high and low surface pressures (Zhai et al., 2013). Negatively charged pure PA and phosphatidylserine (PS) matrix monolayers also allow BODIPY-GSLs to be transferred by GLTP, as well as if these lipids were included as minor components in POPC (5–15%). These studies by the Brown laboratory, conclude that the matrix lipid

headgroups, rather than lateral lipid packing could strongly influence how GLTP works at different membrane interfaces (Zhai et al., 2013). As pointed out in earlier studies as well, electrostatic differences in the amino acids on the surface of the GLTP protein surrounding the sphingolipid-binding sites and the membrane interaction sites differ among known GLTP homologs (Brown and Mattjus, 2007; Simanshu et al., 2013; Tuuf and Mattjus, 2013). Different charged lipids in both the donor and acceptor membrane could therefore be of a significant physical ‘non-protein’ regulator of how various GLTP works. This has also been proposed as a potential regulatory mechanism for other lipid transfer proteins such as the ceramide transfer protein CERT (Tuuf et al., 2011). The transbilayer movement, flip-flop of GSLs is slow (Buton et al., 2002), meaning that the transfer of GSLs only takes place from the leaflet exposed to the transfer protein. This is experimentally evident in several studies showing that the endpoint of transfer from unilamellar vesicles (diameter of 50– 100 nm) plateaus out at about 50–60% of the total GSL concentration (Mattjus et al., 1999, 2000; Rao et al., 2004, 2005; Nylund and Mattjus, 2005; Brown and Mattjus, 2007; Mattjus, 2009). This mass distribution is expected in vesicles of this size (Huang and Thompson, 1974; Lichtenberg and Barenholz, 1988). However, the distribution between the inner and the outer leaflets of galactosylceramide (GalCer) in highly curved unilamellar vesicles (diameter of 25–30 nm) shows that only 30% of GalCer is localized to the outer leaflet (Mattjus et al., 2002). If the size of the vesicles increase, the curvature stress is reduced and the distribution becomes more even. A shift in the GalCer transbilayer distribution was also induced by addition of sphingomyelin (SM, <30%). The addition of SM caused GalCer to move to the outer leaflet to an equal distribution, even in highly curved 25 nm in diameter vesicles (Mattjus et al., 2002). The change in GalCer transbilayer distribution caused by SM when included in POPC vesicles was further analysed, and we confirmed that the 3-OH group, 4,5-trans double bond, and amide linkage all were required in SM (Malewicz et al., 2005). Both of these studies could not have been done without the collaboration with the Bittman laboratory. High curvature of membranes is of great interest in biology because of its importance in vesiculation events, as well as to better understand processes at membrane regions that are under curvature stress. At these locations GSLs are often involved (Ewers and Helenius, 2011). 3. Lipid properties that affect the GLTP action Three structural criteria for a lipid that needs to be fulfilled for optimum GLTP transferability are: (1) the lipid needs to have two acyl chains, (2) the lipid needs to have an amide-nitrogen linked chain in its backbone, (3) and the first saccharide needs to be bound to the hydrophobic ceramide with a glycosidic bond in a beta conformation. In other words, it needs to be a glycosphingolipid. 3.1. Two acyl chains, one amide-linked Many different GSLs have been assayed, both radiolabeled (Table 1) and labeled with a fluorescent group (Table 2). The data presented in these tables are from different experimental conditions. When possible, some of the transfer rates are compared, and then a normalized transfer rate is given. The author trusts that the experimental setup has been the same for the data sets published in the same figure or table, and are therefore comparable. The transfer rates from different data sets, figures or tables have not been compared. GLTP binds poorly to single chain substances that have one or more sugars (Zhai et al., 2009). The amide functional group of

P. Mattjus / Chemistry and Physics of Lipids 194 (2016) 72–78

ceramide orients the GLS entry of the hydrocarbon chain(s) through the cleft-like gate. Early studies claimed that GLTP was specific for both sphingoid-based and glycerol-based glycolipids in which the first sugar is beta-linked to the nonpolar lipid moiety. However this should now be revised in terms of GLTP’s capacity to

75

transfer glycerol-based glycolipids (Table 3). Based on the collective knowledge from both transfer experiments (West et al., 2008) as well as detailed structural studies clearly showing the importance of the amide-linked chain of the ceramide backbone (Malinina et al., 2006; Zhai et al., 2009; Malinina

Table 1 Transbilayer movement of glycosphingolipids by mammalian GLTPs. The table indicates transferred or non-transferred lipids by GLTP for human, bovine and porcine proteins. When rates are given, they are only comparable within the given set. Lipids

Rates

Chain pure or known acyl chain mixtures of glycosphingolipids Palmitoyl-GalCer, unlabeled

GalCer, bovine brain, 3H at C6 of galactose GlcCer, Gaucher spleen, 3H at 4,5-positions of sphingosine 1.3 1.0

GlcCer, Gaucher spleen, 3H at 4,5-positions of sphingosine GM1, bovine brainb SM, bovine brain

1.0

1.0 0.9 0.9 0.8 0.4 0.2

GlcCerd GM1,3H-headgroup label GalCerS, sulfatide, 3H-headgroup label GalCer, 3H-headgroup label LacCer, 3H-palmitic acid

1.0 0.9 0.9 0.6 0.6

a b

e f

Kenoth et al. (2010)

Gammon et al. (1987)

1.0 0.9 0.8 0.7 0.7 0.6 0.3 0.3

Yamada et al. (1985) 3 H-label in galactose

Brown et al. (1985) Yamada et al. (1985) 3 H-label with palmitic acid in lymphoblastoid cells

The asialo-GM1 contained C16:0 (0.7%), C18:0 (91.3%), C20:0 (6.4%), C22:0 (0.5%), and miscellaneous components (1.1%). The GM1, contained C16 (0.4%), C18:0 (91.6%), C20:0 (7.4%), C22:0 (0.2%), and miscellaneous components (0.4%). H-GlcCer, New England Nuclear, presumably 3H at 4,5-positions of sphingosine. American Radiochemical, 14C-octanoyl-GlcCer Schwarzmann (1978). A simple and novel method for tritium labeling of gangliosides and other sphingolipids. Biochim. Biophys. Acta 529, 106–114. Produces ketone and aldehyde groups on the headgroup saccharides

c 3 d

Yamada et al. (1985) 3 H-label in galactose

0.2 0.2

GM1,3H at C6 of galactose, or in the ceramide portione GA1, asioalo-GM1,3H at C6 of galactose, or in the ceramide portion LacCer, 3H at C6 of galactose, or in the ceramide portion GD1a GD1b GT1b

Non-transferrable assayed lipids 14 Cholesteryl oleate 3 H-SM (Egg) PC PI PE SM

Brown et al. (1985) All except GlcCer are 3H-labeled headgroup with galactose oxidase/sodium boro 3H-hydride method (Radin et al., 1969)

0.95 No transfer No transfer

GlcCerc GalCer, Gaucher spleen, non-hydroxy fatty acids LacCer, rat liver GalCer, Gaucher spleen, a-hydroxy fatty acids Gb3, globotriaosylceramide, human erythrocyte stroma Gb4, globotetraosylceramide, human erythrocyte stroma Gb5, globopentaosylceramide, equine kidney GM3, equine erythrocyte stroma

GalCer, Gaucher spleen LacCer, rat liver, periodate-oxidizedf LacCer-L1, periodate-oxidized GalCer-G1, periodate-oxidized LacCer-L2, periodate-oxidized GalCerS, sulfatide, monkey kidney epithelial GalCer-G2, periodate-oxidized LacCerS, monkey kidney epithelial

Nylund et al. (2007) (Transfer from monolayers) Ohvo-Rekilä and Mattjus (2011) Brown et al. (1990) Wong et al. (1984)

GA1, asialo-GM1, bovine braina LacCer, bovine brain

POPC

References

76

P. Mattjus / Chemistry and Physics of Lipids 194 (2016) 72–78 Table 2 Transbilayer movement of fluorescently labeled glycosphingolipids mediated by mammalian GLTPs. The table indicates transferred or non-transferred lipids by GLTP for human, bovine and porcine proteins. When rates are given, they are only comparable within the given set. Lipid

Rates

Reference

Transferrable fluorescent glycosphingolipids BODIPY-GlcCer (11-(dipyrrometheneboron difluoride)undecanoyl) BODIPY-LacCer BODIPY-GalCer

1.0 0.8 0.8

West et al. (2008)

BODIPY-GalCer (C15)

Zhai et al. (2013)

Anthrylvinyl-GalCer (12-(9-anthryl)-11-dodecenoyl) Anthrylvinyl-GalCerS, AV-sulfatide Pyrene-GlcCer (monopyrene, 10-[1-pyrenyl decanoyl]) Pyrene-GalCer

Mattjus et al. (1999) Samygina et al. (2011) Wong et al. (1984) Abe et al. (1984)

Non-transferrable assayed fluorescent lipids Anthrylvinyl-sphingomyelin Anthrylvinyl-phosphatidylcholine Anthrylvinyl-ceramide BODIPY-SM BODIPY-PC

Mattjus et al. (1999) Mattjus et al. (1999) Tuuf et al. (2011) West et al. (2008) Tuuf and Mattjus (2007)

et al., 2015). The early studies monitoring the GLTP activity towards glycoglycerolipids was very different from how the activity was measured for GSLs. Instead of using labeled lipids, Yamada et al. (1985) used unlabeled GlcDG and DiGlcDG and measured the transfer between vesicles and pig heart mitochondria. The amount of glycolipid transferred was determined by hydrolysis of the liposomal glycolipid and subsequent determination of glucose in the lysate. For the galactose-containing glycoglycerolipids they used an assay based on concanavalin A-reactive donors and concanavalin A-nonreactive acceptors (Abe et al., 1982). It is not known why a small transfer of GalDG and DiGalDG was detected, however it is possible that the observed transfer was in fact spontaneous transfer, something that cannot be excluded using these types of assays. No control experiments without GLTP were presented in the work (Yamada et al., 1985). In another substrate specificity analysis, contradicting data exists regarding the rate of GlcDG compared to GlcCer (Table 3) (Yamada et al., 1986). Speculatively assay conditions and differences in the acyl chain compositions of the assayed glycolipids might cause different rates. The temperature of 25
C used might also give more variations in the matrix lipid phase states, since natural PLs were used. We now know that the phase state of the PL membranes

harboring the GSLs greatly affect the rate of transfer. A higher temperature of 37
C usually used in the later studies represent a more physiological condition of the membranes, and a better miscibility and less phase separation of the assayed lipids in the donor and acceptor vesicles. Therefore it might be that the transfer seen with the glycerol-based glycolipids could be non-protein mediated, and simply spontaneous. 3.2. Initial saccharide in a beta-glycosidic bond The first saccharide residue of the GSL acts as a primary specificity determinant as a substrate for GLTP. Clearly the first residue with a glycosidic bond in a beta-conformation to the backbone is recognized much better by GLTP, if not exclusively (Table 3). Our laboratory has failed to conclusively assay alphaliked glycolipids for transfer, such as long chain alpha-galactosylceramide. However, if GLTP can transfer alpha-GalCer, the rate is very slow, and might be due to spontaneous transfer (data not published). Early on Yamada and Sasaki postulate an inverse relationship between the length of saccharide chains and the transfer rates (Yamada et al., 1985). This is to some extent still valid, monohexosylceramides, GlcCer and GalCer are transferred faster

Table 3 Transbilayer movement of glycoglycerolipids mediated by mammalian GLTPs. The table indicates transferred or non-transferred lipids by GLTP for human, bovine and porcine proteins. Lipid

Rates

Reference

GalDG, galactosyldiacylglycerol, wheat flour DiGalDG, wheat flour POPC

No transfer No transfer No transfer

West et al. (2008) Competition assay

GalCer, non-hydroxyl fatty acids GalDG, spinach leaves DiGalDG, wheat flour

Transfer Yamada et al. (1985) Marginal transfer 3H-labeled headgroup with galactose oxidase/sodium boro 3H-hydride method Marginal transfer

DiManDG, dimannosyldiacylglycerol (3Man-al ! 3Man-al),Micrococcus lysodeikticus

No transfer

Yamada et al. (1985) 2-3H-glycerol labeled

1.0 GlcCer ManGlcCer, (Man-bl ! 4Glc-bl), bivalve Corbiculu japonica 1.4 ManManGlcCer, (Man-al ! 4Man-bl ! 4Glc-bl), Corbiculu japonica 1.0

Yamada et al. (1986)

GlcDG, (Glc-al-sn-1,2-DG),Streptococci DiGlcDG, (Glc-al ! 2Glc-al-sn-1,2-DG),Streptococci GlcDG, (Glc-bl-rac-1,2-DG), synthetic

Yamada et al. (1986)

No transfer No transfer Transfer

P. Mattjus / Chemistry and Physics of Lipids 194 (2016) 72–78

than LacCer (Table 1). However, when charged residues, such as sialic acids in gangliosides or sulfonat in sulfatides are present in the saccharide chains the measured transfer rates differs. Charged GSLs interact with the matrix lipids, with each other and with the transfer protein differently compared to neutral GSLs. Depending on the donor and acceptor vesicle compositions as well as buffer conditions it becomes challenging to compare the rates.

77

charged amino acids, a lysine in position 60, arginine at 106 and at position 110 another arginine, this patch recognizes and binds the C1P phosphate headgroup. GLTP has in the same positions, aspartic acid, valine and serine, and for FAPP2 isoleucine, glycine and glutamic acid. These amino acids completely eliminates GLTPs and FAPP2s ability to recognize ceramide-1-phosphase as a substrate. 5. Concluding remarks

4. FAPP2 and CPTP mediated lipid transfer For recent reviews focusing on the GLTP-fold superfamily and the GLTP and FAPP2 and their regulation by lipid composition readers are referred to Malinina et al. (2015) and Tuuf and Mattjus (2013), respectively. The recently described ceramide-1-phosphate transfer protein, CPTP, is also part of a larger group of proteins that recognize sphingolipids (Fig. 2). With its GLTP-like fold CPTP is specific for ceramide-1-phosphate (Simanshu et al., 2013). Other proteins that act on sphingoid-based lipids are the ceramide transfer protein, CERT and the ganglioside GM2 activator protein and the four saposins involved in the degradation of GSLs in the lysosomes. The activator proteins and saposins assist and “lift up” lipid substrate from the membrane matrix, thus making it more accessible to the soluble degradative enzymes (Sandhoff, 2013). The selectivity of the FAPP2 appears to be more analogous to the fungal glycolipid transfer protein HET-C2 (Saupe et al., 1994) rather than human GLTP. Human GLTP can transfer sulfatide as well as various GSLs with charged and uncharged saccharide headgroups, where as FAPP2 transfers GSLs with uncharged simple headgroups, and not sulfatide or GM1 (Kamlekar et al., 2013). This difference can be structurally explained by homology modeling that shows that the negatively charged glutamic acid at position 403 of FAPP2 (nonpolar leucine in GLTP, position 92) may form a salt bridge with lysine 367. The positively charged lysine can also repel the negative groups of sulfatide, consequently interfering with the capacity of FAPP2 to bind and transfer charged GSLs. CPTP has three positively

Were FAPP2 and CPTP discovered much earlier? In 1985 Charles Gammon and Robert Ledeen reported two distinct glycolipid transfer activities (TPI and TPII, 20–22 kD) in bovine brain (Gammon and Ledeen, 1985). With the knowledge we have today it is likely that their two proteins activities were an effect of GLTP instability and breakdown during the lengthy purification procedures, and not the activity of two different GLTP types. It is unlikely to be the activity of the other family member FAPP2, since its size (58 kD) is much larger than the reported purified protein fractions used by Gammon and Ledeen. However, it cannot be excluded that the activity could be from the cleaved-off GLTP-like domain of FAPP2. CPTP, even though it has the approximate right size to the TPI and TPII, does not transfer GSLs, and was most likely not the other activity seen by Gammon and Ledeen. It is likely that the difference in the TPI and TPII activity towards GM1 and GA1 could be due to the difference in the charge between the full length (TPII) and apparently shorter TPI. GM1 is negatively charged, like all gangliosides, whereas the asialo-forms are more neutral. This is the case for GLTP and the less positively charged (at neutral pH) HETC2 protein (Mattjus et al., 2003), and therefore HET-C2 is not as sensitive to negatively charged membranes as GLTP. Acknowledgments This work was supported in part by the Åbo Akademi University and Sigrid Jusélius Foundation. References

Fig. 2. Phylogeny analysis of amino acid sequences of proteins known to act on sphingolipids. These sphingolipid recognizing proteins, SRPs bind, sense, present and transfer different classes of sphingolipids, and belong to different families. The analysis was performed using the online software at Phylogeny.fr (Dereeper et al., 2008, 2010). CERT, ceramide transfer protein, binds ceramide with its START domain and mediates the intracellular trafficking of ceramides in a non-vesicular manner. START domain stands for steroidogenic acute regulatory protein-related lipid-transfer. Saposins stimulate and aid in the degradation of different GSLs in the lysosomes. GM2A, ganglioside GM2 activator protein is an essential cofactor for lysosomal betahexosaminidase A in the enzymatic hydrolysis of GM2 ganglioside to GM3. CPTP is the ceramide-1-phosphate transfer protein and CPTP2 (currently annotated as the GLTPD2 gene in the databases) is a larger GLTP-fold protein that is predicted to have a signal peptide and more closely related to CPTP than GLTP. ACD11 is the Arabidopsis CPTP homolog. HET-C2 is the filamentous fungi GLTP homolog. The reference sequences used were as follows: GM2A (CAA43994), CERT (Q9Y5P4.1), SAP (NP_002769), AtGLTP1 (BAH19942.1), AtGLTP2 (Q6NLQ3.1), AtGLTP3 (Q9LU33.1), HET-C2 (AAA20542.1), GLTP (NP_057517), FAPP2 (NP_001183955), FAPP1 (AAH63575.1), ACD11 (NP_181016.1), CPTP (NP_001025056) and CPTP2 (NP_001014985).

Abe, A., Yamada, K., Sasaki, T., 1982. A protein purified from pig brain accelerates the intermembrane translocation of mono- and dihexoxylceramides, but not the translocation of phospholipids. Biochem. Biophys. Res. Commun. 104, 1386–1393. Abe, A., Yamada, K., Sakagami, T., Sasaki, T., 1984. A fluorimetric determination of the activity of glycolipid transfer protein and some properties of the protein purified from pig brain. Biochim. Biophys. Acta 778, 239–244. Bittman, R., Clejan, S., Rottem, S., 1983. Transbilayer distribution of sterols in mycoplasma membranes: a review. Yale J. Biol. Med. 56, 397–403. Brown, R.E., Mattjus, P., 2007. Glycolipid transfer protein—review. Biochim. Biophys. Acta 1771, 746–776. Brown, R.E., Stephenson, F.A., Markello, T., Barenholz, Y., Thompson, T.E., 1985. Properties of a specific glycolipid transfer protein from bovine brain. Chem. Phys. Lipids 38, 79–93. Brown, R.E., Jarvis, K.L., Hyland, K.J., 1990. Purification and characterization of glycolipid transfer protein from bovine brain. Biochim. Biophys. Acta 1044, 77–83. Buton, X., Herve, P., Kubelt, J., Tannert, A., Burger, K.N., Fellmann, P., Muller, P., Herrmann, A., Seigneuret, M., Devaux, P.F., 2002. Transbilayer movement of monohexosylsphingolipids in endoplasmic reticulum and Golgi membranes. Biochemistry 41, 13106–13115. Correa-Freire, M.C., Barenholz, Y., Thompson, T.E., 1982. Glucocerebroside transfer between phosphatidylcholine bilayers. Biochemistry 21, 1244–1248. D'Angelo, G., Polishchuk, E., Di Tullio, G., Santoro, M., Di Campli, A., Godi, A., West, G., Bielawski, J., Chuang, C.C., van der Spoel, A.C., Platt, F.M., Hannun, Y.A., Polishchuk, R., Mattjus, P., De Matteis, M.A., 2007. Glycosphingolipid synthesis requires FAPP2 transfer of glucosylceramide. Nature 449, 62–67. D'Angelo, G., Uemura, T., Chuang, C.C., Polishchuk, E., Santoro, M., Ohvo-Rekila, H., Sato, T., Di Tullio, G., Varriale, A., D'Auria, S., Daniele, T., Capuani, F., Johannes, L., Mattjus, P., Monti, M., Pucci, P., Williams, R.L., Burke, J.E., Platt, F.M., Harada, A., De Matteis, M.A., 2013. Vesicular and non-vesicular transport feed distinct glycosylation pathways in the Golgi. Nature 501, 116–120. Dansen, T.B., Westerman, J., Wouters, F.S., Wanders, R.J.A., van Hoek, A., Gadella, T.W. J., Wirtz, K.W.A., 1999. High-affinity binding of very-long-chain fatty acyl-CoA esters to the peroxisomal non-specific lipid-transfer protein (sterol carrier protein-2). Biochem. J. 339, 193–199. Dawson, G., Sweeley, C.C., 1970. In vivo studies on glycosphingolipid metabolism in porcine blood. J. Biol. Chem. 245, 410–416.

78

P. Mattjus / Chemistry and Physics of Lipids 194 (2016) 72–78

Dereeper, A., Guignon, V., Blanc, G., Audic, S., Buffet, S., Chevenet, F., Dufayard, J.F., Guindon, S., Lefort, V., Lescot, M., Claverie, J.M., Gascuel, O., 2008. Phylogeny.fr: robust phylogenetic analysis for the non-specialist. Nucleic Acids Res. 36, W465–W469. Dereeper, A., Audic, S., Claverie, J.M., Blanc, G., 2010. BLAST-EXPLORER helps you building datasets for phylogenetic analysis. BMC Evol. Biol. 10, 8. Edqvist, J., Rönnberg, E., Rosenquist, S., Blomqvist, K., Viitanen, L., Salminen, T.A., Nylund, M., Tuuf, J., Mattjus, P., 2004. Plants express a lipid transfer protein with high similarity to mammalian sterol carrier protein-2. J. Biol. Chem. 279, 53544–53553. Ewers, H., Helenius, A., 2011. Lipid-mediated endocytosis. Cold Spring Harb. Perspect. Biol. 3, a004721. Gammon, C.M., Ledeen, R.W., 1985. Evidence of the presence of a ganglioside transfer protein in brain. J. Neurochem. 44, 979–982. Gammon, C.M., Vaswani, K.D., Ledeen, R.W., 1987. Isolation of two glycolipid transfer proteins from bovine brain: reactivity toward gangliosides and neutral glycosphingolipids. Biochemistry 26, 6239–6243. Huang, C.H., Thompson, T.E., 1974. Preparation of homogeneous: single-walled phosphatidylcholine vesicles. Methods Enzymol. 32, 485–489. Kamlekar, R.K., Simanshu, D.K., Gao, Y.G., Kenoth, R., Pike, H.M., Prendergast, F.G., Malinina, L., Molotkovsky, J.G., Venyaminov, S.Y., Patel, D.J., Brown, R.E., 2013. The glycolipid transfer protein (GLTP) domain of phosphoinositol 4-phosphate adaptor protein-2 (FAPP2): structure drives preference for simple neutral glycosphingolipids. Biochim. Biophys. Acta 1831, 417–427. Kenoth, R., Simanshu, D.K., Kamlekar, R.K., Pike, H.M., Molotkovsky, J.G., Benson, L. M., Bergen 3rd, H.R., Prendergast, F.G., Malinina, L., Venyaminov, S.Y., Patel, D.J., Brown, R.E., 2010. Structural determination and tryptophan fluorescence of heterokaryon incompatibility C2 protein (HET-C2), a fungal glycolipid transfer protein (GLTP), provide novel insights into glycolipid specificity and membrane interaction by the GLTP fold. J. Biol. Chem. 285, 13066–13078. Lichtenberg, D., Barenholz, Y., 1988. Liposomes: preparation, characterization, and preservation. Methods Biochem. Anal. 33, 337–462. Locatelli-Hoops, S., Remmel, N., Klingenstein, R., Breiden, B., Rossocha, M., Schoeniger, M., Koenigs, C., Saenger, W., Sandhoff, K., 2006. Saposin a mobilizes lipids from low cholesterol and high bmp containing membranes. patient variant saposin a lacks lipid extraction capacity. J. Biol. Chem. 281, 32451– 32460. Malewicz, B., Valiyaveettil, J.T., Jacob, K., Byun, H.S., Mattjus, P., Baumann, W.J., Bittman, R., Brown, R.E., 2005. The 3-hydroxy group and 4,5-trans double bond of sphingomyelin are essential for modulation of galactosylceramide transmembrane asymmetry. Biophys. J. 88, 2670–2680. Malinina, L., Malakhova, M.L., Kanack, A.T., Lu, M., Abagyan, R., Brown, R.E., Patel, D. J., 2006. The liganding of glycolipid transfer protein is controlled by glycolipid acyl structure. PLoS Biol. 4, e362. Malinina, L., Simanshu, D.K., Zhai, X., Samygina, V.R., Kamlekar, R., Kenoth, R., Ochoa-Lizarralde, B., Malakhova, M.L., Molotkovsky, J.G., Patel, D.J., Brown, R.E., 2015. Sphingolipid transfer proteins defined by the GLTP-fold. Q. Rev. Biophys. 1–42. Mattjus, P., 2009. Glycolipid transfer proteins and membrane interaction. Biochim. Biophys. Acta 1788, 267–272. Mattjus, P., Molotkovsky, J.G., Smaby, J.M., Brown, R.E., 1999. A fluorescence resonance energy transfer approach for monitoring protein-mediated glycolipid transfer between vesicle membranes. Anal. Biochem. 268, 297–304. Mattjus, P., Pike, H.M., Molotkovsky, J.G., Brown, R.E., 2000. Charged membrane surfaces impede the protein mediated transfer of glycosphingolipids between phospholipid bilayers. Biochemistry 39, 1067–1075. Mattjus, P., Malewicz, B., Valiyaveettil, J.T., Baumann, W.J., Bittman, R., Brown, R.E., 2002. Sphingomyelin modulates the transbilayer distribution of galactosylceramide in phospholipid membranes. J. Biol. Chem. 277, 19476– 19481. Mattjus, P., Turcq, B., Pike, H.M., Molotkovsky, J.G., Brown, R.E., 2003. Glycolipid intermembrane transfer is accelerated by HET-C2, a filamentous fungus gene product involved in the cell–cell incompatibility response. Biochemistry 42, 535–542. Metz, R.J., Radin, N.S., 1980. Glucosylceramide uptake protein from spleen cytosol. J. Biol. Chem. 255, 4463–4467. Nichols, J.W., Pagano, R.E., 1981. Kinetics of soluble lipid monomer diffusion between vesicles. Biochemistry 20, 2783–2789. Nichols, J.W., Pagano, R.E., 1983. Resonance energy transfer assay of proteinmediated lipid transfer between vesicels. J. Biol. Chem. 258, 5368–5371. Nylund, M., Mattjus, P., 2005. Protein mediated glycolipid transfer is inhibited FROM sphingomyelin membranes but enhanced TO sphingomyelin containing raft like membranes. Biochim. Biophys. Acta 1669, 87–94. Nylund, M., Fortelius, C., Palonen, E.K., Molotkovsky, J.G., Mattjus, P., 2007. Membrane curvature effects on glycolipid transfer protein activity. Langmuir 23, 11726–11733.

Ohvo-Rekilä, H., Mattjus, P., 2011. Monitoring glycolipid transfer protein activity and membrane interaction with the surface plasmon resonance technique. Biochim. Biophys. Acta 1808, 47–54. Radin, N.S., Hof, L., Bradley, R.M., Brady, R.O., 1969. Lactosylceramide galactosidase: comparison with other sphingolipid hydrolases in developing rat brain. Brain Res. 14, 497–505. Rao, C.S., Lin, X., Pike, H.M., Molotkovsky, J.G., Brown, R.E., 2004. Glycolipid transfer protein mediated transfer of glycosphingolipids between membranes: a model for action based on kinetic and thermodynamic analyses. Biochemistry 43, 13805–13815. Rao, C.S., Chung, T., Pike, H.M., Brown, R.E., 2005. Glycolipid transfer protein interaction with bilayer vesicles: modulation by changing lipid composition. Biophys. J. 89, 4017–4028. Rottem, S., Slutzky, G.M., Bittman, R., 1978. Cholesterol distribution and movement in the Mycoplasma gallisepticum cell membrane. Biochemistry 17, 2723–2726. Samygina, V.R., Popov, A.N., Cabo-Bilbao, A., Ochoa-Lizarralde, B., Goni-de-Cerio, F., Zhai, X., Molotkovsky, J.G., Patel, D.J., Brown, R.E., Malinina, L., 2011. Enhanced selectivity for sulfatide by engineered human glycolipid transfer protein. Structure 19, 1644–1654. Sandhoff, K., 2013. Metabolic and cellular bases of sphingolipidoses. Biochem. Soc. Trans. 41, 1562–1568. Sasaki, T., 1985. Glycolipid-binding proteins. Chem. Phys. Lipids 38, 63–77. Sasaki, T., 1990. Glycolipid transfer protein and intracellular traffic of glucosylceramide. Experientia 46, 611–616. Sasaki, T., Demel, R.A., 1985. Net mass transfer of galactosylceramide facilitated by glycolipid transfer protein from pig brain: a monolayer study. Biochemistry 24, 1079–1083. Saupe, S.J., Descamps, C., Turcq, B., Begueret, J., 1994. Inactivation of the Podospora anserina vegetative incompatibility locus het-c, whose product resembles a glycolipid transfer protein, drastically impairs ascospore production. Proc. Natl. Acad. Sci. U. S. A. 91, 5927–5931. Simanshu, D.K., Kamlekar, R.K., Wijesinghe, D.S., Zou, X., Zhai, X., Mishra, S.K., Molotkovsky, J.G., Malinina, L., Hinchcliffe, E.H., Chalfant, C.E., Brown, R.E., Patel, D.J., 2013. Non-vesicular trafficking by a ceramide-1-phosphate transfer protein regulates eicosanoids. Nature 500, 463–467. Somerharju, P.J., Brockerhoff, H., Wirtz, K.W.A., 1981. A new fluorimetric method to measure protein-catalyzed phospholipid transfer using 1-acyl2parinaroylphosphatidylcholine. Biochim. Biophys. Acta 649, 521–528. Somerharju, P.J., van, L., oon, D., Wirtz, K.W., 1987. Determination of the acyl chain specificity of the bovine liver phosphatidylcholine transfer protein. Application of pyrene-labeled phosphatidylcholine species. Biochemistry 26, 7193–7199. Schwarzmann, G., 1978. A simple and novel method for tritium labeling of gangliosides and other sphingolipids. Biochim. Biophys. Acta 529, 106–114. Tuuf, J., Mattjus, P., 2007. Human glycolipid transfer protein-intracellular localization and effects on the sphingolipid synthesis. Biochim. Biophys. Acta 1771, 1353–1363. Tuuf, J., Mattjus, P., 2013. Membranes and mammalian glycolipid transferring proteins. Chem. Phys. Lipids . Tuuf, J., Kjellberg, M.A., Molotkovsky, J.G., Hanada, K., Mattjus, P., 2011. The intermembrane ceramide transport catalyzed by CERT is sensitive to the lipid environment. Biochim. Biophys. Acta 1808, 229–235. Viitanen, L., Nylund, M., Eklund, D.M., Alm, C., Eriksson, A.K., Tuuf, J., Salminen, T.A., Mattjus, P., Edqvist, J., 2006. Characterization of SCP-2 from Euphorbia lagascae reveals that a single Leu/Met exchange enhances sterol transfer activity. FEBS J. 273, 5641–5655. West, G., Viitanen, L., Alm, C., Mattjus, P., Salminen, T.A., Edqvist, J., 2008. Identification of a glycosphingolipid transfer protein GLTP1 in Arabidopsis thaliana. FEBS J. 275, 3421–3437. Wong, M., Brown, R.E., Barenholz, Y., Thompson, T.E., 1984. Glycolipid transfer protein from bovine brain. Biochemistry 23, 6498–6505. Yamada, K., Abe, A., Sasaki, T., 1985. Specificity of the glycolipid transfer protein from pig brain. J. Biol. Chem. 260, 4615–4621. Yamada, K., Abe, A., Sasaki, T., 1986. Glycolipid transfer protein from pig brain transfers glycolipids with beta-linked sugars but not with alpha-linked sugars at the sugar-lipid linkage. Biochim. Biophys. Acta 879, 345–349. Zhai, X., Malakhova, M.L., Pike, H.M., Benson, L.M., Bergen, H.R., Sugar, I.P., Malinina, L., Patel, D.J., Brown, R.E., 2009. Glycolipid acquisition by human glycolipid transfer protein dramatically alters intrinsic tryptophan fluorescence: insights into glycolipid binding affinity. J. Biol. Chem. 284, 13620–13628. Zhai, X., Momsen, W.E., Malakhov, D.A., Boldyrev, I.A., Momsen, M.M., Molotkovsky, J.G., Brockman, H.L., Brown, R.E., 2013. GLTP-fold interaction with planar phosphatidylcholine surfaces is synergistically stimulated by phosphatidic acid and phosphatidylethanolamine. J. Lipid Res. 54, 1103–1113.