MINIREVIEW The lipid translocase, ABCA4: seeing is believing Naomi Laura Pollock and Richard Callaghan Nuffield Department of Clinical Laboratory Science, University of Oxford, UK Introduction Many members of the A subfamily of ATP binding cassette (ABC) transporters have crucial roles in lipid metabolism. Their importance is demonstrated by the severe consequences of their absence or inability to function normally. For example, mutations to the glu- cosylceramide transporter ABCA12 can cause harle- quin ichthyosis, a potentially lethal condition in which the epidermal layer of skin is abnormally thickened and lacks integrity, leaving sufferers vulnerable to excessive water loss and recurrent infection through the skin [1,2]. The lack of functional ABCA1 also has serious clinical implications, namely Tangier disease, characterized by deposits of cholesterol in peripheral tissues, resulting from inhibition of the reverse choles- terol pathway [3–5]. The focus of this review is the protein ABCA4. Mutations affecting the function of this ABC trans- porter also lead to the formation of lipid-rich deposits, but in this case they are limited to a specific region of one tissue: the macular region of the retina. Malfunc- tion of ABCA4 can lead to juvenile-onset macular degeneration, notably the condition Stargardt disease (SD) [6]. ABCA4 and heritable disorders of vision SD is recognized as the most common heritable macu- lar degenerative disorder, with a prevalence of up to 1 in 8000 [6]. Additional recessively inherited juvenile- onset retinal degenerative conditions have been described, including retinitis pigmentosa, cone-rod dys- trophy [7–10] and age-related macular degeneration (AMD) [11]. Symptoms shared by these conditions Keywords ABC transporter; all-trans-retinal; phospholipid translocase; Stargardt disease Correspondence R. Callaghan, Nuffield Department of Clinical Laboratory Science, University of Oxford, UK Fax: +44 1865 221 834 Tel: +44 1865 221 110 E-mail: richard.callaghan@ndcls.ox.ac.uk (Received 21 December 2010, revised 28 February 2011, accepted 6 May 2011) doi:10.1111/j.1742-4658.2011.08169.x Mutations to members of the A subfamily of ATP binding cassette (ABC) proteins are responsible for a number of diseases; typically they are associ- ated with aberrant cellular lipid transport processes. Mutations to the ABCA4 protein are linked to a number of visual disorders including Stargardt’s disease and retinitis pigmentosa. Over 500 disease-associated mutations in ABCA4 have been demonstrated; however, the genotype– phenotype link has not been firmly established. This shortfall is primarily because the function of ABCA4 in the visual cycle is not yet fully under- stood. One hypothesis suggests that ABCA4 mediates the trans-bilayer translocation of retinal-phosphatidylethanolamine conjugates to facilitate the retinal regeneration process in the visual cycle. This review examines the evidence to support, or refute, this working hypothesis on the function of this clinically important protein. Abbreviations ABC, ATP-binding cassette; AMD, age-related macular degeneration; ATR, all-trans-retinal; ATRol, all-trans-retinol; ECD, extracellular domain; ER, endoplasmic reticulum; NBD, nucleotide binding domain; NrPE, N-retinylidene-phosphatidylethanolamine; OS, outer segment; PE, phosphatidylethanolamine; PRC, photoreceptor cell; RDH, retinal dehydrogenase; RPE, retinal pigment epithelium; SD, Stargardt disease; TMD, transmembrane domain; WT, wild-type. 3204 FEBS Journal 278 (2011) 3204–3214 ª 2011 The Authors Journal compilation ª 2011 FEBS include loss of visual acuity, development of yellow pigmentation in the retina and loss of central vision [12]. In 1997, a gene common to SD and some cases of cone-rod dystrophy and retinitis pigmentosa was iden- tified [13]. The protein it encoded was homologous to the Rim protein that had previously been isolated from Xenopus laevis photoreceptor cells (PRCs) [14–16]. This large membrane protein comprised 2273 amino acids, with a predicted molecular weight of 220–250 kDa. From its primary amino acid sequence, the protein was identified as an ABC transporter [14,17]. Its topol- ogy was predicted to include the core ABC transporter domains of two nucleotide binding domains (NBDs) and two bundles of six transmembrane helices (TMDs). Like other proteins in the A subfamily, ABCA4 has two large extracytoplasmic domains (ECDs), consisting of almost 40% of its amino acid residues [18]. Over 500 mutations to the ABCR gene are now associated with macular degenerative disorders and extensive screening is available to identify families at risk from these diseases [8,19–21]. However, we have relatively little insight into the biochemical conse- quences of these mutations. In this review we seek to summarize research to date on the ABCA4 protein, identify some of the outstand- ing questions regarding its activity, and set this in the context of the visual system. For example, what role does ABCA4 fulfil in the visual cycle? What is the mechanism which links ABCA4 dysfunction to macu- lar degeneration? What is the substrate specificity of this transporter and how does it transport its substrates? ABCA4 is involved in the visual cycle Specialized cell types coordinate vertebrate vision PRCs (Fig. 1A) are a major constituent of the retina. There are two types of PRCs – rods and cones – which are adapted to detecting different intensities of light. Detection of light by PRCs relies on opsin proteins, localized to the outer segments (OSs) of the cells, which contain a covalently bound retinoid chromo- phore [22]. The highest concentration of PRCs is found in the macula, an oval-shaped region surround- ing the optic nerve [23]. Loss of photoreceptors from this region results in the loss of central vision that is characteristic of SD [24]. Apical to the OSs of the PRCs is the retinal pigment epithelium (RPE) (Fig. 1B). It is underpinned by a basement membrane, the Bruch’s membrane, and a capillary bed, which supplies oxygen and nutrients, including the precursor of 11-cis-retinal, vitamin A, to the retina [25]. Another vital function of the RPE is the engulfment and digestion of old disc membranes. As new discs bud from the PRC plasma membrane, older discs are displaced towards the RPE and shed for phagocytosis by the RPE cells [22]. Compounds that cannot be digested in this way may accumulate, either in the RPE or the Bruch’s membrane below it [26]. These by-products of disc membrane phagocyto- sis, including cholesterol, cholesteryl-esters and other lipids, are collectively known as lipofuscin [26–28]. The build-up of lipofuscin deposits, and the toxic com- pounds within them, impair the function of the RPE cells and prevent their metabolic support of PRCs [29]. Fig. 1. Schematic diagrams of PRCs. ABCA4 is expressed exclusively in the disc membranes of rod and cone PRCs (A). Villi extending from the RPE cells intercalate with PRCs (B). N. L. Pollock and R. Callaghan ABCA4: seeing is believing FEBS Journal 278 (2011) 3204–3214 ª 2011 The Authors Journal compilation ª 2011 FEBS 3205 Therefore function of the PRCs is dependent on the RPE cells. Trafficking and regeneration of retinoids is essential to maintain vision Healthy RPE cells cooperate with PRCs to recycle all- trans-retinal (ATR) in a process known as the retinoid cycle (Fig. 2) [25,30,31]. This cycle involves the release of ATR, a highly reactive molecule, from rhodopsin. The high concentration of rhodopsin in the disc mem- branes, up to 3 mm [25], means that in conditions of high light intensity it is possible that the rate of ATR release may outstrip the rate of its reduction to all- trans-retinol (ATRol), necessitating alternative means of processing or sequestering ATR [32,33]. The aldehyde group of ATR has the potential to create reactive oxygen species, which can initiate the oxidation of lipids and induce apoptosis [34]. In addition, ATR is known to react with phosphatidyleth- anolamine (PE) to form N-retinylidene-phosphatidyl- ethanolamine (NrPE) [35,36], which can react with a further ATR molecule to form toxic bisretinoid com- pounds [37,38]. The latter cannot be catabolized in the RPE, accumulate in lipofuscin and cause degeneration of the RPE [12,27,37,38]. Therefore it is vital for the PRCs to process ATR as quickly as possible. It has been suggested that each retinoid in this path- way has a specific chaperone to prevent unwanted reactions [25]. For instance rhodopsin has a total of three binding sites for retinoids, allowing one 11-cis- retinal to bind an entrance site and another to bind the active site while ATR remains covalently bound at an exit site, where it can be reduced to ATRol or released [25,33,39]. The reversible formation of NrPE allows PE in the disc membranes to act as a temporary sink for ATR; subsequent hydrolysis enables ATR to re-enter the retinoid cycle [40]. However, the reversible formation of NrPe is the first step towards the forma- tion of bisretinoids [28,38,41], which makes it a high- risk strategy for the chaperonage of ATR and unlikely to be a principal pathway for ATR in the retinoid cycle. After ATR has left the OS discs, the remaining steps of the retinoid cycle occur in the RPE cells (Fig. 2). Disc membrane composition modulates the visual cycle The lipid composition of OS discs is distinct from that of the plasma membrane from which they are derived, providing a highly fluid membrane environment to enable rapid signalling from rhodopsin to the brain [25,42,43]. Creating this distinct lipid composition necessitates extensive sorting of phospholipids when the discs are created, the details of which are not well understood. Certain phospholipids and cholesterol associate with rhodopsin to modulate its activity [44,45], although cholesterol is progressively lost from the Fig. 2. Overview of the retinoid cycle. (1) ATR moves out of the active site of rhodop- sin into the OS disc (2), where it may be transported into the PRC cytoplasm by ABCA4. It is reduced to ATRol by an all- trans-retinol dehydrogenase. (3) ATRol moves from the OS disc of the PRC into the RPE cell layer. (4) Lecithin retinol acyl- transferase, (5) retinal-pigment-epithelium- specific 65 kDa protein and (6) 11-cis-retinol dehydrogenase regenerate ATRol into 11-cis-retinal. (7) 11-cis-retinal moves into the OS disc, where it (8) binds to rhodopsin for photoisomerization. ABCA4: seeing is believing N. L. Pollock and R. Callaghan 3206 FEBS Journal 278 (2011) 3204–3214 ª 2011 The Authors Journal compilation ª 2011 FEBS ageing OS discs [46,47]. Whether this is a symptom or a cause of disc ageing and what significance it has in the visual cycle is unclear. The investment in creat- ing the unique lipid composition of the OS discs, and the existence of other visual disorders caused by aber- rant lipid sorting [19,48], indicate that this is a critical process to enable vision. As ABCA4 is localized exclusively to this membrane, it is likely that its activ- ity is also influenced by the unique membrane envi- ronment of the OS discs, although there is currently no evidence for direct involvement of ABCA4 in its creation. Experimental evidence for the role of ABCA4 Retinoid transport by ABCA4 was first proposed in 1997, shortly after the ABCA4 gene was identified [11,49,50]. This hypothesis was deduced from the local- ization of the protein to the disc membranes of PRCs, its ability to bind ATP and its homology with the ABC transporter family [17,51]. Studies on purified ABCA4 have enabled this hypothesis to be tested in some detail, while the creation of ABCA4 ) ⁄ ) mice has provided an in vivo model for macular degenerative disorders [52–55]. Studies on knockout mice ABCA4 ) ⁄ ) mice enabled detailed characterization of changes in the retina caused by a lack of ABCA4 activity. Electroretinography, the measurement of the electrical response of the eye to light, and analysis of tissue samples taken from eyes have been used to examine the ABCA4 ) ⁄ ) phenotype [52,54]. The first study on ABCA4 ) ⁄ ) mice [54] reported delayed adaptation to dark and delayed clearance of ATR after photobleaching (the conversion of 11-cis- retinal to ATR within rhodopsin). The levels of rho- dopsin and 11-cis-retinal in ABCA4 ) ⁄ ) mice were similar to wild-type (WT) mice, indicating that ABCA4 is not an essential protein in the retinoid cycle, nor does its absence alter the availability of rhodopsin. Rather, the accelerated accumulation of ATR in the disc membranes of ABCA4 ) ⁄ ) mice pro- vided strong evidence that ABCA4 mediates the pro- cessing or transport of ATR following its dissociation from rhodopsin. ABCA4 ) ⁄ ) mice also had an increased rate of lipofuscin accumulation at their reti- nas and the Bruch’s membrane underlying the RPE was thicker than in their WT counterparts [54]. This corresponds to observations of the retinas of human subjects with retinal degenerative disorders [56], although this seems to be occur in AMD rather than in SD. Finally, ABCA4 ) ⁄ ) mice contained at least 10-fold more A2E, or isoA2E, in retinal extracts than WT mice of the same age [54]. A2E was detectable only in RPE extracts, not at the OS discs, suggesting that A2E was formed rapidly in the RPE of ABCA4 ) ⁄ ) mice, despite the localization of ABCA4 to the OS disc membranes. This highlights the crux of the ABCA4 question: how is loss of ABCA4 activity in the OS discs related to changes in the RPE cells, and how is this effect propagated back to the PRCs to cause macular degeneration? Biochemical analyses of purified ABCA4 Some biochemical evidence supports the case for ABCA4 acting as an ATP-powered retinoid trans- porter. The first observation relating to this was the release of ATR from purified ABCA4 upon the addi- tion of ATP or GTP [57]. ATR appeared to remain bound to ABCA4 during purification from rod OS discs, but binding or hydrolysis of ATP altered the affinity of the protein for ATR, leading to its release. The rate of ATP hydrolysis by purified, reconsti- tuted ABCA4 has also been measured [58–60] to exam- ine its physiological function. Many ABC transporters have a background or basal rate of ATPase activity, which is stimulated when the protein interacts with its specific transport substrate(s) [61]. In the case of ABCA4, ATR but no other retinoid compound was observed to stimulate the ATPase activity of the pro- tein to an appreciable extent, which led to the conclu- sion that ATR could be the retinoid substrate transported by ABCA4 in vivo [60,62]. However, the presence of PE in the reconstituted proteoliposomes also enhanced the basal activity of ABCA4. On this basis, it was proposed that the substrate of ABCA4 could be NrPE, the product of an equilibrium reaction between ATR and the amine group of PE [57,60,63]. It has also been shown that ATR can quench the intrinsic tryptophan fluorescence of isolated WT ECD2, suggesting that ATR binds to ABCA4 at the ECDs [64]. A dissociation constant (K D ) of 0.17 lm for ATR binding to WT ECD2 was inferred from the data. Moreover, specific mutations to the ECDs, which are linked to SD, were shown to increase K D , indicat- ing that the binding affinity was lower in the mutant ECDs. This could account for the poor function of some mutant forms of ABCA4, which result in the accumulation of ATR in the OS discs and ultimately in loss of vision. N. L. Pollock and R. Callaghan ABCA4: seeing is believing FEBS Journal 278 (2011) 3204–3214 ª 2011 The Authors Journal compilation ª 2011 FEBS 3207 Defining the orientation of ABCA4 in the disc mem- brane is also fundamental to our understanding of its activity. The current topological model [18], based on analysis of the amino acid sequence and biochemical data, predicts that the ECDs are located in the disc lumen, while the NBDs reside in the OS cytoplasm (Fig. 3A). There is good evidence to support this model. For example, the endoplasmic reticulum (ER) lumen and disc lumen are topologically equivalent and the abundance of glycosylation sites in the ECDs indi- cates that these domains are located within the ER lumen during protein synthesis. In addition positively charged residues at the N-terminal suggest a cytoplas- mic localization for this region [18]. Combined with the predicted topology of 12 transmembrane helices, this gives us a model with the ECDs and NBDs on opposite sides of the membrane, with the former located within the disc lumen. ABCA4 activity – the NrPE flippase model Based on data from the ABCA4 ) ⁄ ) mice and biochemi- cal experiments, a hypothesis has been proposed indicat- ing that ABCA4 acts as a transporter of NrPE [63]. Following regeneration of rhodopsin with 11-cis-retinal, ATR is released from the ‘exit site’ of opsin into the disc lumen, where a proportion of it reacts with a PE mole- cule to form NrPE. The selectivity of ECD2 for ATR suggests that the role of the ECDs is recognition of the substrate NrPE, which is structurally related to ATR. Following interaction with the ECDs, the substrate is flipped or transported across the disc membrane into the cytoplasmic leaflet, or directly into the cytoplasm. Translocation is powered by hydrolysis of one or two ATP molecules at the NBDs, which reside in the cytoplasm ensuring ready provision of nucleotides. Fol- lowing release, NrPE can be hydrolysed to PE and ATR. At this cytosolic location the latter is more acces- sible to retinal dehydrogenase (RDH), thereby re-form- ing ATRol and returning to the retinoid cycle. Loss of function of ABCA4 leads to the accumula- tion of NrPE in the disc lumen. A subsequent reaction between NrPE and ATR leads to the irreversible for- mation of A2E. When discs are shed from the PRCs and phagocytosed by the RPE cells, A2E cannot be degraded. Instead it is deposited as lipofuscin in Bruch’s membrane where it causes RPE cell death and PRC degeneration, giving rise to the symptoms of SD and other retinal degenerative disorders. The role of ABCA4 in disc membranes: insight or oversight? The model described above provides a plausible expla- nation for most of the evidence that we have about the activity of ABCA4. However, gaps in our understand- ing of the protein in particular, and the visual cycle in general, pose a number of intriguing puzzles. Is ABCA4 really a flippase? In the absence of a direct functional assay, the fre- quent assertion that ABCA4 acts as a flippase of NrPE remains speculative. A number of ABC transporters have been proposed to act as flippases [65–67], and although in some cases, for instance the human phos- phatidylcholine transporter and the Escherichia coli MsbA protein [68,69], there is reasonable evidence to support this we have yet to conclusively demonstrate flippase activity for ABCA4. Fig. 3. Orientation of ABCA1 and ABCA4 in the membrane. ABCA4 (A) exists within the disc membranes in PRCs. The ECDs are located within the disc lumen (L) and substrate (ATR) is hypothesized to travel (arrow A) from the lumen into the cytoplasm. ABCA1 (B) is located in the plasma membrane and oriented with NBDs in the cytoplasm and the ECDs located extracellularly (EC). Substrates including cholesterol (Chol) are transported (arrow B) from the cytoplasm to an extracellular acceptor (e.g. apoA-1). ABCA4: seeing is believing N. L. Pollock and R. Callaghan 3208 FEBS Journal 278 (2011) 3204–3214 ª 2011 The Authors Journal compilation ª 2011 FEBS Several alternative mechanisms to substrate flipping have been proposed to describe the protein-mediated passage of lipids across the bilayer [70]. For example ABCA1, the protein with most homology to ABCA4, is thought to mediate transport of cholesterol directly from the cytoplasmic leaflet to an extracellular binding protein (Fig. 3B). This is at odds with the classical flip- pase activity that has been suggested for ABCA4, in which the substrate would be flipped from the luminal to the cytoplasmic leaflet of the membrane. Whether two such closely related proteins could operate by dif- ferent mechanisms remains an open question. What is the role of the ECDs? Accepting the flippase mechanism presents us with another puzzle with respect to the role of the ECDs. These domains have been shown to selectively bind ATR in vitro [71], yet this molecule can also react with PE to form NrPE in the luminal leaflet of the mem- brane. If the protein acts as an NrPE flippase, it must bind NrPE from the luminal leaflet. However, this undermines the role of the ECDs in ATR binding, as the ECD region of the protein is soluble and resides in the disc lumen (Fig. 3). If ABCA4 acts as a flippase, it would be more logical for an NrPE recognition site to exist in the TMD of the protein. Even so, the strict evolutionary conservation within the ECDs [72] and the grave consequences of mutations in this region indicate a vital functional role [8,71]. For ABCA1 there is good evidence that the ECDs interact with lipoproteins to facilitate transport, deliv- ering substrate from the ABC protein to the soluble lipoproteins, apoA-1 and apoE-1 [3,70,73] These lipo- proteins are essential for the efflux of lipids and their assembly into high-density lipoproteins [70]. Conserved Cys residues in ECD1 and ECD2 have been shown to form a disulfide bridge, which is vital for apoA-1 bind- ing and lipid unloading [74]. This highlights the ques- tion of whether ABCA4 would behave in the opposite way, with substrate recognition occurring at the ECDs [71] (Fig. 3). Phylogenetic analysis of the ABCA transporters in Amphioxus, an organism often used as a model of early vertebrate lineages, has uncovered a close evolu- tionary relationship between ABCA1, ABCA7 and ABCA4 [75]. All three are thought to derive from the same ancestral gene through gene duplication events. Therefore, it is logical to infer that function as well as structure of the ECDs may be conserved between these three proteins [76]. Furthermore, the ECDs comprise nearly 40% of the molecular mass of ABCA4 and mutations associated with SD map to amino acid substitutions in the ECDs, indicating that loss of function here does affect the function of the protein as a whole [20]. The scarcity of experimental data describing the ECDs of ABCA4 ren- ders this a subject for speculation. The role of these domains requires extensive investigation in order to fully understand the functional and mechanistic details of ABCA4. How significant is the activity of ABCA4 in the retinoid cycle? It is generally accepted that the majority of ATR is processed back to ATRol by an RDH enzyme, possi- bly while ATR remains bound in the ‘exit site’ of rho- dopsin [12,33,39]. Both ATR and ATRol can diffuse through the disc membrane [77], which enables them to move into the PRC cytoplasm and then to the RPE cells for conversion back to 11-cis-retinal [40,78]. Based on this ease of diffusion across the disc mem- brane, one obvious question is whether ATR requires a transporter at all. Possibly, a specific transporter is required not to facilitate pigment regeneration but to facilitate the reversible sequestration of ATR in a less reactive form, namely as NrPE. This seems plausible due to the rapid diffusibility of ATR within the disc. Studies in ABCA4 ) ⁄ ) mice have also estimated that just 30% of ATR leaves the OS discs as NrPE [25,52]. Hence some regard ABCA4, although vital, as a minor mechanism for ATR processing [25,52]: if the formation of NrPE is inevitable, removing it to the cytoplasm where it may hydrolyse back to PE and ATR could reduce the probability of bisretinoid formation. This would also correspond with the fact that the pathologies connected to ABCA4 deficiency are degen- erative. The slow decline of the PRCs and the RPE layer would actually be the cumulative effect over many years of relatively small A2E. Even in individu- als with fully functional ABCA4, lipofuscin deposits are common in later years [27]. Is disc membrane lipid homeostasis linked to the function of ABCA4? The unique lipid composition of OS disc membranes is achieved by extensive redistribution of lipids after the creation of the OS discs, but it is not yet clear why, or even how, this is effected [22]. Flip-flop of lipids between the leaflets of the disc membranes is rapid [79,80] and new flippases are still being identified [81]. The similarity between ABCA4 and ABCA1 [75,82], a cholesterol and phospholipase (PL) efflux pump, N. L. Pollock and R. Callaghan ABCA4: seeing is believing FEBS Journal 278 (2011) 3204–3214 ª 2011 The Authors Journal compilation ª 2011 FEBS 3209 suggests that ABCA4 could have a role in this lipid sorting. Could ATR be exploiting the activity of ABCA4 as a PE flippase by ‘piggy-backing’ onto the transbilayer PE movement? Or has ABCA4 evolved specifically to fulfil this niche role of removing ATR from the disc lumen? One of the major observations in ABCA4 ) ⁄ ) mice was the abnormally elevated levels of PE in the disc membrane [54], which contained 1.6 times the amount of PE found in WT OS discs. It was assumed that the lack of ABCA4 disrupted NrPE transport, so the rate of PE movement to the cytoplasm was reduced and both this phospholipid and ATR accumulated at the luminal leaflet of the disc membrane. In WT cells, PE turnover by a specific phospholipase is relatively rapid in the cytoplasmically oriented PE [54]. One hypothesis to account for the elevated level of PE in ABCA4 ) ⁄ ) mice is that, without ABCA4, PE does not reach the cytoplasmic leaflet of the disc membrane and is trapped in the luminal leaflet. The functional consequences of the change in the phospholipid composition of OS discs in ABCA4 ) ⁄ ) mice are unknown. Given the sensitivity of rhodopsin to cholesterol [44,45], it seems likely that an altered lipid composition (i.e. increase in PE, loss of choles- terol over time) could affect the kinetics of the visual cycle in other ways [39,42,83], which may also affect the kinetics of photobleach recovery. One comparison between WT and ABCA4 ) ⁄ ) mice noted that, under conditions in which the exposure to light of the ABCA4 ) ⁄ ) mice was varied resulting in different extents of rhodopsin photobleaching ( 1 · 10 )4 % to 30%), WT mice actually had a slower recovery than ABCA4 ) ⁄ ) mice [52]. The excess of PE in the disc membranes of ABCA4 ) ⁄ ) mice was suggested as an explanation for this effect: the phospholipid acts as a sink for ATR allowing more rapid dissociation from rhodop- sin, despite the lack of functional ABCA4. Alterna- tively, the ability of ABCA4 to bind 11-cis-retinal [60] could lead to competition between rhodopsin and ABCA4 for 11-cis-retinal binding. In the absence of ABCA4 this competition would be removed, increasing the availability of 11-cis-retinal to rhodop- sin and increasing the speed of the photobleach recovery. Hence, this study suggested that the role of ABCA4 in OS discs, rather than acting as a major pathway for ATR reprocessing, is a minor route for ATR out of the disc membranes, with diffusion playing the signifi- cant role. ABCA4 would be essential for the removal of residual amounts of ATR from the OS discs [52]. The slightly reduced efficiency of rhodopsin regenera- tion would be a small trade-off for ensuring the lon- gevity of the PRCs. Therefore, although ABCA4 ) ⁄ ) mice have greatly increased our understanding of the effects of ABCA4 deficiency, they have not provided conclusive evidence for the exclusive role of ABCA4 as an NrPE trans- porter. Rather, the implications of lipid sorting in the discs, and the possible consequences of this, have been highlighted by these studies. Lipid homeostasis is clearly vitally important to maintaining vision; lipofuscin deposits contain not only the retinoid by-product A2E, but also lipids and cholesterol derivatives [26–28]. In fact, there is evi- dence that ABCA4 is not the only ABC transporter that plays a role in lipofuscin accumulation. Polymor- phisms in ABCA1 and its partner lipoprotein apoE-1 have recently been linked to an increased risk of AMD [84,85] and ABCA1 is known to mediate cholesterol efflux from lysosomes in RPE cells [38]. Inhibition of this process by A2E has been linked to increased lipo- fuscin deposits. This is one of the first hints of the mechanism directly linking A2E accumulation, due to ABCA4 dysfunction, to the lipofuscin accumulation which causes macular degeneration. Import or export? Finally, perhaps the most intriguing of the conun- drums about the activity of ABCA4 is the direction of transport. In the NrPE flippase model of ABCA4 activity, the substrate is transported out of the disc lumen and into the cytoplasm of the PRC [63]. In terms of the postulated role of ABCA4 in the visual cycle, this is a logical suggestion. However, in terms of our understanding of the mechanism of ABC trans- porters, this represents a huge departure from the accepted canon. All eukaryotic ABC transporters are thought to function in the export direction, with the exception of Arabidopsis ABCB14, which may act as an importer [86]. The phenotypical consequence of deleting ABCB14 was examined in these experiments, which is analogous to the use of ABCA4 ) ⁄ ) mice. Hence a direct observation of eukaryotic ABC-medi- ated import has yet to be made. Most eukaryotic ABC proteins are believed to trans- port their substrates by an alternating access mecha- nism [87]: the inward facing protein conformation has a high affinity binding site to bind the substrate; the outward facing conformation has a low affinity site, enabling its release [88,89]. In the case of the prokary- otic importer ABC proteins, the hypothesis of alternat- ing access is retained but the high and low affinity binding sites are reversed [90,91]. The closest ABCA4: seeing is believing N. L. Pollock and R. Callaghan 3210 FEBS Journal 278 (2011) 3204–3214 ª 2011 The Authors Journal compilation ª 2011 FEBS homologues of ABCA4 are both believed to transport phospholipids and cholesterol by an alternating access mechanism [4,76], while ABCA4 is proposed to act as a flippase of retinal and PE. The similarity of its sub- strate to those of ABCA1 and ABCA7 indicates that the substrate binding sites of all these proteins have features in common. If we invoke sequence homology between ABCA1 and ABCA4 as evidence that ABCA4 has capabilities as a lipid transporter [64,92], can we overlook the fact that transport occurs in opposite directions? Identifying an importer amongst the human ABC transporters would present a major step forward in our understanding of these proteins. For instance, it would imply that the distinction between importers and exporters is more subtle than we assume at pres- ent, since ABCA1 and ABCA4 have 40% homology yet are postulated to act in opposite directions. Phylo- genetic analysis of the ABC transporter superfamily indicates that importer and exporter function diverged long before the prokaryotic ⁄ eukaryotic division [93]. The existence of a eukaryotic importer would represent a new subclass of ABC protein. Conclusion and perspectives Despite having posed so many questions about our understanding of ABCA4, it is important to emphasize that, at present, all the evidence we have points towards ABCA4 acting as an import-directed flippase. None of the questions discussed above necessarily con- tradicts this model, but we hope that they highlight gaps in our knowledge which justify investigation. In this review, we have attempted to describe the biochemistry of ABCA4 in the context of the visual cycle. Our current understanding is that ABCA4 is a transporter of ATR or NrPE. However, its existence in a specialized lipid environment and its close relation- ship to other PL and cholesterol transporters also implicate it in lipid transport. The complexity of inter- preting recovery after photobleaching and the relative importance of ABCA4 in ATR regeneration plus lack of data on the role of ECDs leave some distance until we fully understand the molecule and its role in main- taining vision. Purification and reconstitution of ABCA4 to estab- lish a full functional assay must be a priority to con- firm the substrate of this protein. With such a system in place, one approach is to fluorescently label putative substrates and measure translocation across a bilayer. Though technically difficult, this is a more reliable means of identifying the substrate of an ABC trans- porter than simply observing stimulation in its rate of ATP hydrolysis in the presence of the putative sub- strate. This approach has been successful with other ABC transporters [68,69] and hence may enable the categorical classification of ABCA4 as an importer. With the substrate(s) and direction of transport firmly established, identifying the location of the binding site and the mechanism of transport may also be more straightforward. References 1 Annilo T et al. (2002) Identification and characteriza- tion of a novel ABCA subfamily member, ABCA12, located in the lamellar ichthyosis region on 2q34. Cyto- genet Genome Res 98, 169–176. 2 Kelsell DP et al. (2005) Mutations in ABCA12 underlie the severe congenital skin disease harlequin ichthyosis. Am J Hum Genet 76, 794–803. 3 Oram JF, Lawn RM, Garvin MR & Wade DP (2000) ABCA1 is the cAMP-inducible apolipoprotein receptor that mediates cholesterol secretion from macrophages. J Biol Chem 275, 34508–34511. 4 Oram JF & Vaughan AM (2000) ABCA1-mediated transport of cellular cholesterol and phospholipids to HDL apolipoproteins. Curr Opin Lipidol 11, 253–260. 5 Rust S et al. (1999) Tangier disease is caused by muta- tions in the gene encoding ATP-binding cassette trans- porter 1. Nat Genet 22, 352–355. 6 Allikmets R et al. (1997) A photoreceptor cell-specific ATP-binding transporter gene (ABCR) is mutated in recessive Stargardt macular dystrophy. Nat Genet 15, 236–246. 7 Cremers FP et al. (1998) Autosomal recessive retinitis pigmentosa and cone-rod dystrophy caused by splice site mutations in the Stargardt’s disease gene ABCR. Hum Mol Genet 7, 355–362. 8 Klevering BJ, Deutman AF, Maugeri A, Cremers FP & Hoyng CB (2005) The spectrum of retinal phenotypes caused by mutations in the ABCA4 gene. Graefes Arch Clin Exp Ophthalmol 243, 90–100. 9 Martinez-Mir A, Paloma E, Allikmets R, Ayuso C, del Rio T, Dean M, Vilageliu L, Gonzalez-Duarte R & Bal- cells S (1998) Retinitis pigmentosa caused by a homozy- gous mutation in the Stargardt disease gene ABCR. Nat Genet 18, 11–12. 10 van Driel MA, Maugeri A, Klevering BJ, Hoyng CB & Cremers FP (1998) ABCR unites what ophthalmologists divide(s). Ophthalmic Genet 19, 117–122. 11 Allikmets R et al. (1997) Mutation of the Stargardt dis- ease gene (ABCR) in age-related macular degeneration. Science 277, 1805–1807. 12 Rozanowska M & Sarna T (2005) Light-induced damage to the retina: role of rhodopsin chromophore revisited. Photochem Photobiol 81, 1305–1330. N. L. Pollock and R. Callaghan ABCA4: seeing is believing FEBS Journal 278 (2011) 3204–3214 ª 2011 The Authors Journal compilation ª 2011 FEBS 3211 13 Allikmets R (1997) A photoreceptor cell-specific ATP-binding transporter gene (ABCR) is mutated in recessive Stargardt macular dystrophy. Nat Genet 17, 122. 14 Papermaster DS, Converse CA & Zorn M (1976) Bio- synthetic and immunochemical characterization of large protein in frog and cattle rod outer segment mem- branes. Exp Eye Res 23, 105–115. 15 Papermaster DS, Schneider BG, Zorn MA & Kraehen- buhl JP (1978) Immunocytochemical localization of a large intrinsic membrane protein to the incisures and margins of frog rod outer segment disks. J Cell Biol 78, 415–425. 16 Szuts EZ (1985) Light stimulates phosphorylation of two large membrane proteins in frog photoreceptors. Biochemistry 24, 4176–4184. 17 Illing M, Molday LL & Molday RS (1997) The 220- kDa rim protein of retinal rod outer segments is a member of the ABC transporter superfamily. J Biol Chem 272, 10303–10310. 18 Bungert S, Molday LL & Molday RS (2001) Membrane topology of the ATP binding cassette transporter ABCR and its relationship to ABC1 and related ABCA transporters: identification of N-linked glycosylation sites. J Biol Chem 276, 23539–23546. 19 Molday RS & Zhang K (2010) Defective lipid trans- port and biosynthesis in recessive and dominant Stargardt macular degeneration. Prog Lipid Res 49, 476–492. 20 Sun H, Smallwood PM & Nathans J (2000) Biochemical defects in ABCR protein variants associated with human retinopathies. Nat Genet 26, 242–246. 21 Rivera A et al. (2000) A comprehensive survey of sequence variation in the ABCA4 (ABCR) gene in Star- gardt disease and age-related macular degeneration. Am J Hum Genet 67, 800–813. 22 Boesze-Battaglia K & Schimmel R (1997) Cell mem- brane lipid composition and distribution: implications for cell function and lessons learned from photorecep- tors and platelets. J Exp Biol 200, 2927–2936. 23 Provis JM, Penfold PL, Cornish EE, Sandercoe TM & Madigan MC (2005) Anatomy and development of the macula: specialisation and the vulnerability to macular degeneration. Clin Exp Optom 88, 269–281. 24 Rotenstreich Y, Fishman GA & Anderson RJ (2003) Visual acuity loss and clinical observations in a large series of patients with Stargardt disease. Ophthalmology 110, 1151–1158. 25 Lamb TD & Pugh EN Jr (2004) Dark adaptation and the retinoid cycle of vision. Prog Retin Eye Res 23 , 307–380. 26 Holz FG, Sheraidah G, Pauleikhoff D & Bird AC (1994) Analysis of lipid deposits extracted from human macular and peripheral Bruch’s membrane. Arch Ophthalmol 112, 402–406. 27 Sparrow JR & Boulton M (2005) RPE lipofuscin and its role in retinal pathobiology. Exp Eye Res 80, 595– 606. 28 Eldred GE & Lasky MR (1993) Retinal age pigments generated by self-assembling lysosomotropic detergents. Nature 361, 724–726. 29 Lakkaraju A, Finnemann SC & Rodriguez-Boulan E (2007) The lipofuscin fluorophore A2E perturbs choles- terol metabolism in retinal pigment epithelial cells. Proc Natl Acad Sci USA 104, 11026–11031. 30 McBee JK, Palczewski K, Baehr W & Pepperberg DR (2001) Confronting complexity: the interlink of photo- transduction and retinoid metabolism in the vertebrate retina. Prog Retin Eye Res 20 , 469–529. 31 Rando RR (2001) The biochemistry of the visual cycle. Chem Rev 101, 1881–1896. 32 Saari JC, Garwin GG, Van Hooser JP & Palczewski K (1998) Reduction of all-trans-retinal limits regeneration of visual pigment in mice. Vision Res 38, 1325–1333. 33 Schadel SA, Heck M, Maretzki D, Filipek S, Teller DC, Palczewski K & Hofmann KP (2003) Ligand chan- neling within a G-protein-coupled receptor. The entry and exit of retinals in native opsin. J Biol Chem 278, 24896–24903. 34 Sparrow JR, Zhou J, Ben-Shabat S, Vollmer H, Itagaki Y & Nakanishi K (2002) Involvement of oxidative mechanisms in blue-light-induced damage to A2E-laden RPE. Invest Ophthalmol Vis Sci 43, 1222–1227. 35 Anderson RE & Maude MB (1970) Phospholipids of bovine outer segments. Biochemistry 9, 3624–3628. 36 Mata NL, Weng J & Travis GH (2000) Biosynthesis of a major lipofuscin fluorophore in mice and humans with ABCR-mediated retinal and macular degeneration. Proc Natl Acad Sci USA 97, 7154–7159. 37 Sparrow JR, Fishkin N, Zhou J, Cai B, Jang YP, Krane S, Itagaki Y & Nakanishi K (2003) A2E, a byproduct of the visual cycle. Vision Res 43, 2983– 2990. 38 Sparrow JR, Wu Y, Kim CY & Zhou J (2010) Phos- pholipid meets all-trans-retinal: the making of RPE bis- retinoids. J Lipid Res 51, 247–261. 39 Heck M, Schadel SA, Maretzki D, Bartl FJ, Ritter E, Palczewski K & Hofmann KP (2003) Signaling states of rhodopsin. Formation of the storage form, metarhodop- sin III, from active metarhodopsin II. J Biol Chem 278, 3162–3169. 40 Rattner A, Smallwood PM & Nathans J (2000) Identifi- cation and characterization of all-trans-retinol dehydro- genase from photoreceptor outer segments, the visual cycle enzyme that reduces all-trans-retinal to all-trans- retinol. J Biol Chem 275, 11034–11043. 41 Ben-Shabat S, Parish CA, Vollmer HR, Itagaki Y, Fish- kin N, Nakanishi K & Sparrow JR (2002) Biosynthetic studies of A2E, a major fluorophore of retinal pigment epithelial lipofuscin. J Biol Chem 277, 7183–7190. ABCA4: seeing is believing N. L. Pollock and R. Callaghan 3212 FEBS Journal 278 (2011) 3204–3214 ª 2011 The Authors Journal compilation ª 2011 FEBS 42 Niu SL, Mitchell DC, Lim SY, Wen ZM, Kim HY, Salem N Jr & Litman BJ (2004) Reduced G protein- coupled signaling efficiency in retinal rod outer seg- ments in response to n-3 fatty acid deficiency. J Biol Chem 279, 31098–31104. 43 Niu SL, Mitchell DC & Litman BJ (2002) Manipulation of cholesterol levels in rod disk membranes by methyl- beta-cyclodextrin: effects on receptor activation. J Biol Chem 277, 20139–20145. 44 Albert AD, Boesze-Battaglia K, Paw Z, Watts A & Epand RM (1996) Effect of cholesterol on rhodopsin stability in disk membranes. Biochim Biophys Acta 1297, 77–82. 45 Albert AD, Young JE & Yeagle PL (1996) Rhodopsin- cholesterol interactions in bovine rod outer segment disk membranes. Biochim Biophys Acta 1285, 47–55. 46 Boesze-Battaglia K & Albert AD (1989) Fatty acid composition of bovine rod outer segment plasma mem- brane. Exp Eye Res 49, 699–701. 47 Boesze-Battaglia K, Hennessey T & Albert AD (1989) Cholesterol heterogeneity in bovine rod outer segment disk membranes. J Biol Chem 264, 8151–8155. 48 Boesze-Battaglia K, Organisciak DT & Albert AD (1994) RCS rat retinal rod outer segment membranes exhibit different cholesterol distributions than those of normal rats. Exp Eye Res 58, 293–300. 49 Sun H & Nathans J (1997) Stargardt’s ABCR is loca- lized to the disc membrane of retinal rod outer seg- ments. Nat Genet 17 , 15–16. 50 Travis GH & Bennett J (1997) The ABCs of AMD. Nat Med 3, 1196–1197. 51 Travis GH (1997) Insights from a lost visual pigment. Nat Genet 15, 115–117. 52 Pawar AS, Qtaishat NM, Little DM & Pepperberg DR (2008) Recovery of rod photoresponses in ABCR- deficient mice. Invest Ophthalmol Vis Sci 49, 2743– 2755. 53 Radu RA, Yuan Q, Hu J, Peng JH, Lloyd M, Nusino- witz S, Bok D & Travis GH (2008) Accelerated accu- mulation of lipofuscin pigments in the RPE of a mouse model for ABCA4-mediated retinal dystrophies follow- ing Vitamin A supplementation. Invest Ophthalmol Vis Sci 49, 3821–3829. 54 Weng J, Mata NL, Azarian SM, Tzekov RT, Birch DG & Travis GH (1999) Insights into the function of Rim protein in photoreceptors and etiology of Stargardt’s disease from the phenotype in abcr knockout mice. Cell 98, 13–23. 55 Wu L, Nagasaki T & Sparrow JR (2010) Photoreceptor cell degeneration in abcr (- ⁄ -) mice. Adv Exp Med Biol 664, 533–539. 56 Chong NH, Keonin J, Luthert PJ, Frennesson CI, Weingeist DM, Wolf RL, Mullins RF & Hageman GS (2005) Decreased thickness and integrity of the macular elastic layer of Bruch’s membrane correspond to the distribution of lesions associated with age-related macu- lar degeneration. Am J Pathol 166, 241–251. 57 Beharry S, Zhong M & Molday RS (2004) N-retinyli- dene-phosphatidylethanolamine is the preferred retinoid substrate for the photoreceptor-specific ABC transpor- ter ABCA4 (ABCR). J Biol Chem 279, 53972–53979. 58 Ahn J & Molday RS (2000) Purification and characteri- zation of ABCR from bovine rod outer segments. Methods Enzymol 315, 864–879. 59 Ahn J, Wong JT & Molday RS (2000) The effect of lipid environment and retinoids on the ATPase activity of ABCR, the photoreceptor ABC transporter responsi- ble for Stargardt macular dystrophy. J Biol Chem 275, 20399–20405. 60 Sun H, Molday RS & Nathans J (1999) Retinal stimu- lates ATP hydrolysis by purified and reconstituted ABCR, the photoreceptor-specific ATP-binding cassette transporter responsible for Stargardt disease. J Biol Chem 274, 8269–8281. 61 Senior AE, al-Shawi MK & Urbatsch IL (1998) ATPase activity of Chinese hamster P-glycoprotein. Methods Enzymol 292, 514–523. 62 Sun H & Nathans J (2001) ABCR, the ATP-binding cassette transporter responsible for Stargardt macular dystrophy, is an efficient target of all-trans-retinal- mediated photooxidative damage in vitro. Implications for retinal disease. J Biol Chem 276, 11766–11774. 63 Molday RS, Beharry S, Ahn J & Zhong M (2006) Binding of N-retinylidene-PE to ABCA4 and a model for its trans- port across membranes. Adv Exp Med Biol 572, 465–470. 64 Zhong M & Molday RS (2010) Binding of retinoids to ABCA4, the photoreceptor ABC transporter associated with stargardt macular degeneration. Methods Mol Biol 652, 163–176. 65 Doerrler WT, Reedy MC & Raetz CR (2001) An Escherichia coli mutant defective in lipid export. J Biol Chem 276, 11461–11464. 66 Romsicki Y & Sharom FJ (2001) Phospholipid flippase activity of the reconstituted P-glycoprotein multidrug transporter. Biochemistry 40, 6937–6947. 67 Ruetz S & Gros P (1994) Phosphatidylcholine translo- case: a physiological role for the mdr2 gene. Cell 77, 1071–1081. 68 Eckford PD & Sharom FJ (2010) The reconstituted Escherichia coli MsbA protein displays lipid flippase activity. Biochem J 429, 195–203. 69 Smith AJ, Timmermans-Hereijgers JL, Roelofsen B, Wirtz KW, van Blitterswijk WJ, Smit JJ, Schinkel AH & Borst P (1994) The human MDR3 P-glycoprotein promotes translocation of phosphatidylcholine through the plasma membrane of fibroblasts from transgenic mice. FEBS Lett 354, 263–266. 70 Nagao K, Kimura Y, Mastuo M & Ueda K (2010) Lipid outward translocation by ABC proteins. FEBS Lett 584, 2717–2723. N. L. Pollock and R. Callaghan ABCA4: seeing is believing FEBS Journal 278 (2011) 3204–3214 ª 2011 The Authors Journal compilation ª 2011 FEBS 3213 [...].. .ABCA4: seeing is believing N L Pollock and R Callaghan 71 Biswas-Fiss EE, Kurpad DS, Joshi K & Biswas SB (2010) Interaction of extracellular domain 2 of the human retina-specific ATP-binding cassette transporter (ABCA4) with all-trans-retinal J Biol Chem 285, 19372–19383 72 Yatsenko AN, Wiszniewski W, Zaremba CM, Jamrich M & Lupski JR (2005) Evolution... outer segments Biochemistry 46, 8669–8679 79 Hessel E, Herrmann A, Muller P, Schnetkamp PP & Hofmann KP (2000) The transbilayer distribution of phospholipids in disc membranes is a dynamic equilibrium evidence for rapid flip and flop movement Eur J Biochem 267, 1473–1483 80 Wu G & Hubbell WL (1993) Phospholipid asymmetry and transmembrane diffusion in photoreceptor disc membranes Biochemistry 32, 879–888... purification, and functional reconstitution of the P4-ATPase Atp8a2, a phosphatidylserine flippase in photoreceptor disc membranes J Biol Chem 284, 32670–32679 82 Dean M & Annilo T (2005) Evolution of the ATP-binding cassette (ABC) transporter superfamily in vertebrates Annu Rev Genomics Hum Genet 6, 123–142 3214 83 Gibson NJ & Brown MF (1993) Lipid headgroup and acyl chain composition modulate the MI-MII... understanding of eukaryotic P-glycoproteins and homologues FEBS J 277, 550–563 92 Molday RS, Zhong M & Quazi F (2009) The role of the photoreceptor ABC transporter ABCA4 in lipid transport and Stargardt macular degeneration Biochim Biophys Acta 1791, 573–583 93 Saurin W, Hofnung M & Dassa E (1999) Getting in or out: early segregation between importers and exporters in the evolution of ATP-binding cassette... access of the putative substrate-binding chamber in the ABC transporter MsbA J Mol Biol 393, 574–585 90 Oldham ML, Davidson AL & Chen J (2008) Structural insights into ABC transporter mechanism Curr Opin Struct Biol 18, 726–733 91 Kerr ID, Jones PM & George AM (2010) Multidrug efflux pumps: the structures of prokaryotic ATP-binding cassette transporter efflux pumps and implications for our understanding... equilibrium of rhodopsin in recombinant membranes Biochemistry 32, 2438–2454 84 Duncan KG, Hosseini K, Bailey KR, Yang H, Lowe RJ, Matthes MT, Kane JP, LaVail MM, Schwartz DM & Duncan JL (2009) Expression of reverse cholesterol transport proteins ATP-binding cassette A1 (ABCA1) and scavenger receptor BI (SR-BI) in the retina and retinal pigment epithelium Br J Ophthalmol 93, 1116– 1120 85 Neale BM et... vertebrates J Mol Evol 60, 72–80 73 Oram JF (2000) Tangier disease and ABCA1 Biochim Biophys Acta 1529, 321–330 74 Hozoji M, Kimura Y, Kioka N & Ueda K (2009) Formation of two intramolecular disulfide bonds is necessary for ApoA-I-dependent cholesterol efflux mediated by ABCA1 J Biol Chem 284, 11293–11300 75 Li G, Zhang QJ, Ji ZL & Wang YQ (2007) Origin and evolution of vertebrate ABCA genes: a story from amphioxus... cellular cholesterol and phospholipid to generate high density lipoprotein J Biol Chem 279, 604–611 77 Wu Q, Chen C & Koutalos Y (2006) All-trans retinol in rod photoreceptor outer segments moves unrestrictedly by passive diffusion Biophys J 91, 4678–4689 78 Wu Q, Blakeley LR, Cornwall MC, Crouch RK, Wiggert BN & Koutalos Y (2007) Interphotoreceptor retinoid-binding protein is the physiologically relevant... (2010) Genome-wide association study of advanced age-related macular degeneration identifies a role of the hepatic lipase gene (LIPC) Proc Natl Acad Sci USA 107, 7395–7400 86 Lee M, Choi Y, Burla B, Kim YY, Jeon B, Maeshima M, Yoo JY, Martinoia E & Lee Y (2008) The ABC transporter AtABCB14 is a malate importer and modulates stomatal response to CO2 Nat Cell Biol 10, 1217– 1223 87 Jardetzky O (1966) Simple... E (1999) Getting in or out: early segregation between importers and exporters in the evolution of ATP-binding cassette (ABC) transporters J Mol Evol 48, 22–41 FEBS Journal 278 (2011) 3204–3214 ª 2011 The Authors Journal compilation ª 2011 FEBS . MINIREVIEW The lipid translocase, ABCA4: seeing is believing Naomi Laura Pollock and Richard Callaghan Nuffield Department of Clinical. the unique lipid composition of the OS discs, and the existence of other visual disorders caused by aber- rant lipid sorting [19,48], indicate that this