Location of macular xanthophylls in the most vulnerable regions of photoreceptor outer-segment membranes

Abstract

Lutein and zeaxanthin are two dietary carotenoids that compose the macular pigment of the primate retina. Another carotenoid, meso-zeaxanthin, is formed from lutein in the retina. A membrane location is one possible site where these dipolar, terminally dihydroxylated carotenoids, named macular xanthophylls, are accumulated in the nerve fibers and photoreceptor outer segments. Macular xanthophylls are oriented perpendicular to the membrane surface, which ensures their high solubility, stability, and significant effects on membrane properties. It was recently shown that they are selectively accumulated in membrane domains that contain unsaturated phospholipids, and thus are located in the most vulnerable regions of the membrane. This location is ideal if they are to act as lipid antioxidants, which is the most accepted mechanism through which lutein and zeaxanthin protect the retina from age-related macular degeneration. In this mini-review, we examine published data on carotenoid-membrane interactions and present our hypothesis that the specific orientation and location of macular xanthophylls maximize their protective action in membranes of the eye retina.

Introduction

Lutein and zeaxanthin—dipolar, terminally dihydroxylated carotenoids—selectively accumulate at an extremely high concentration in membranes of the primate eye retina [1–3] from blood plasma (where more than 20 other carotenoids are available [4]). These two carotenoids can impede the onset of age-related macular degeneration (AMD) [5–8] and have been recently added to the list of potentially beneficial nutrients provided by leafy greens [9]. What role can lutein and zeaxanthin play in protecting the retina? Blue-light filtration [10] and antioxidant functions [11] are often effects that are assumed for these carotenoids. Although, blue-light absorption can be considered an indirect antioxidant action because it prevents blue light from generating reactive oxygen species that can damage photoreceptor cells [12]. Reacting as antioxidants with free radicals and reactive oxygen species, macular xanthophylls protect the retina against peroxidation and photo-damage [13,14]. The ability of macular xanthophylls to quench singlet oxygen and triplet states of photoactive molecules is especially significant [15].

Why have lutein and zeaxanthin been selected from the more than 20 carotenoids present in human plasma? The ability of these xanthophylls to filter out blue light is not better than that of other carotenoids. Likewise, their abilities to quench singlet oxygen [15] and to scavenge free radicals [16] in organic solvents are not better. Therefore, it must be that some specific property or properties of these xanthophylls can help explain their presence in the primate retina. One such property is their behavior in biological membranes [17,18]. In this mini-review, we will present data that show the unique lutein-and zeaxanthin-membrane interaction that distinguishes these carotenoids from others available from blood plasma. We will also present data that support our hypothesis that the specific orientation and location of macular xanthophylls maximize their protective action in the membranes of the eye retina.

The precise location of macular xanthophylls in the Henle’s fiber layer of photoreceptor axons and in photoreceptor outer segments (POS) is not known. There are two major hypotheses about this localization. According to the first, macular xanthophylls transversely incorporate in the lipid-bilayer portion of membranes of the human retina [19,20]. High-energy, short-wave visible light promotes the formation of reactive oxygen species that can initiate lipid peroxidation in membranes of the human retina—a tissue that is abundantly illuminated, has large respiratory demands for oxygen, and is rich in long-chain polyunsaturated fatty acids (such as docosahexaenoic acid [DHA] [21]), which are quite vulnerable to lipid peroxidation. Macular xanthophylls are thought to combat light-induced damage mediated by reactive oxygen species by absorbing the most damaging incoming wavelength of light prior to the formation of reactive oxygen species (a function expected of carotenoids in Henle’s fibers) and by chemically and physically quenching reactive oxygen species once they are formed (a function expected of carotenoids in POS). The membrane location of macular xanthophylls is ideal for these actions. According to the second hypothesis, macular xanthophylls are protein-bound by membrane-associated, xanthophyll-binding proteins [22,23]. Bernstein’s group has identified and characterized a zeaxanthin-binding protein in human macula that enhances the antioxidant activity of zeaxanthin [24,25]. This membrane-associated, xanthophyll-binding protein binds zeaxanthin (but not lutein) with high specificity and affinity. Recently, the Bernstein group also reported the presence of a lutein-binding protein in human macula [26]. Lastly, there is the question of whether the amount of these proteins is sufficient to bind and to store all xanthophyll molecules, which accumulate in the retina in extremely high concentrations. Both interactions of lutein and zeaxanthin with lipid-bilayer membranes and specific proteins are significant. However, in this mini-review, we will focus on carotenoid-membrane interactions.

Transport of macular xanthophylls and their solubility and orientation in membranes

There are data that show that polar carotenoids are more efficiently transported from the lumen of the gastrointestinal tract into intestinal mucosal cells than nonpolar carotenes, which results in their better bioavailability [27]. Carotenoids appear to be absorbed by mucosal cells by a mechanism involving passive diffusion [27–29], although During et al. [30] suggest that intestinal transport of carotenoids might be facilitated by the participation of a specific epithelial transporter. Owing to their more polar nature, xanthophylls like lutein and zeaxanthin can more easily be incorporated into the outer portions of lipid micelles within the gastrointestinal tract and therefore can be more easily taken up by enterocyte membranes and, eventually, chylomicrones [27].

Carotenoids are transported in human blood plasma exclusively by lipoproteins. Nonpolar carotenoids are transported primarily in light density lipoproteins (LDL), whereas more polar carotenoids are more evenly distributed between LDL and high density lipoproteins (HDL) [23,28]. It is thought that most tissues obtain carotenoids via the LDL receptor route [28]. However, in the case of lutein and zeaxanthin transport, we believe that receptors for HDL should be involved instead. It has been suggested that this role can be played by receptors that are similar to those found in the central nervous system for HDL particles containing ApoE [23,31]. However, there is a lack of convincing evidence that lipoproteins containing ApoE are responsible for delivery and accumulation of macular xanthophylls in the retina. Nevertheless, data presented above suggest that the segregation of polar and nonpolar carotenoids already occurs on the level of carotenoid transport.

Macular xanthophylls (dipolar, terminally dihydroxylated carotenoids) are well soluble in lipid bilayers. The reported xanthophyll solubility thresholds (concentration of xanthophylls at which aggregation initiates) in fluid-phase model membranes lie in the area of 10 mol% [32]. Also, our results show that xanthophylls affect membrane properties at the concentration up to 10 mol% without indicating saturation of the observed effect [33–35]. However, lower values of xanthophyll solubility, such as 5 and 2 mol%, were reported for small unilamellar vesicles and lipid multibilayers [36–38]. Nonpolar β-carotene starts to aggregate at a concentration as low as 0.5 mol% [39], although values of the solubility threshold as high as 1 mol% have also been reported [40]. Monopolar β-cryptoxanthin is also less soluble in the lipid bilayer than macular xanthophylls [41]. Interestingly, the tendency of cis-isomers of xanthophylls to aggregate is usually much less than their all-trans counterparts [42,43], and they also affect membrane properties more strongly (the effect of zeaxanthin on membrane properties increases in the direction: all-trans < 9-cis ≤ 13-cis [44]). Cis-isomers are also more readily solubilized, absorbed, and transported [42]. It can be hypothesized that the high solubility of lutein and zeaxanthin in the membrane determines the transport and selective accumulation of macular xanthophylls in the retina of the eye. We hypothesize that the high membrane solubility of macular xanthophylls is one of the major characteristics that distinguishes them from other dietary carotenoids.

In the human retina, the concentration of carotenoids reaches a level between 0.1 and 1 mM in the central fovea [3], which is about 1000 times higher than in other tissues. The extremely high level of macular xanthophylls in the retina does not reflect their content in POS membranes. Macular xanthophylls in POS constitute about 10% of the amount in the entire retina [45], although values as high as 25% have also been reported in the outer segment [46]. Despite the lower percentage, the local concentration of macular xanthophylls in membranes of the rod outer segment is ~70% higher than in residual retina membranes [46]. As indicated in the introduction, we will focus on xanthophyll-membrane interactions. Based on measurements with model lipid-bilayer membranes, Bone and Landrum [19] concluded that lutein is located in the membrane and oriented perpendicular to the bilayer surface. This orientation in lipid-bilayer membranes was confirmed for zeaxanthin [36]. However, linear dichroism analysis of the mean orientation of the dipole transition moment of zeaxanthin and lutein incorporated to the oriented EYPC multibilayers revealed essentially different orientation of zeaxanthin and lutein in the membranes [47]. Zeaxanthin was found to adopt roughly vertical orientation with respect to the plane of the membrane, showing the 33° orientation angle between the transition dipole and the axis normal to the plane of the membrane. The relatively large orientation angle of 67° found in the case of lutein was interpreted as a representation of the existence of two orthogonally oriented pools of lutein, one following the orientation of zeaxanthin and the second parallel with respect to the plane of the membrane. We direct readers to a review by Gruszecki [48] where possible orientations of carotenoids in the lipid bilayer membranes are discussed in details and to a review by Krinski [18] and our paper [41] where the unusual results obtained for lutein are critically evaluated. Nevertheless, in this review, we will assume only the transmembrane localization of xanthophylls because the polar hydroxyl groups at each end of the xanthophyll molecule encourage a membrane-spanning configuration in the lipid bilayer. The presence of xanthophylls in biological membranes is ideal if they are to act as a lipid antioxidant, although lipid association alone cannot explain the extraordinary specific uptake of xanthophylls into the macula.

The transmembrane localization of a significant portion of macular xanthophylls in the retinal cells seems to be obvious. Such localization of macular xanthophylls can explain their very slow removal from the retina, observed after discontinuation of a lutein supplement given to healthy volunteers in a study by Landrum et al. [49]. After discontinuing a 140-day lutein supplement, Landrum et al. [49] observed a relatively fast decrease of lutein concentration in the serum, whereas the level of lutein in the retina remained unchanged and at a high concentration for up to six months. Similar effects were observed by Hammond et al. [50]. These observations suggest that anchoring xanthophyll molecules at opposite membrane surfaces is significant not only in enhancing their effects on membrane properties [41] (see also reviews by Gruszecki [51] and Gruszecki and Strzalka [52]), but also in stabilization of these molecules in membranes of the human retina. Thus, transmembrane orientation can also be included as a characteristic that distinguishes macular xanthophylls from other dietary carotenoids.

Antioxidant potency of macular xanthophylls in membranes

For organisms with a high carotenoid content in their membranes (for example, bacteria, and in some situations, plants, in which the local carotenoid concentration in the lipid bilayer can reach a value up to a few mol%), it is most important to understand how carotenoids affect the membrane’s physical properties, structure, and dynamics. It has been shown that membranes of halophyles and thermophylic bacteria contain a fairly large amount of polar carotenoids [53,54]. These bacteria, which live in extreme conditions, should possess stable membranes that provide a high barrier for nonspecific permeation of small molecules. Incorporation of dipolar carotenoids into the membrane serves this purpose well. Dipolar carotenoids stabilize both halves of the lipid bilayer like transmembrane “rivets”, increasing membrane rigidity by ordering the alkyl chains of lipids [33,34], and raising the membrane hydrophobic barrier for polar molecules and ions [35]. Our results support Rohmer’s hypothesis [55] that polar carotenoids regulate membrane properties in prokaryotes in a manner similar to cholesterol in eukaryotes (see also discussions in [35] and [56]). Polar carotenoids are also present transiently in the lipid-bilayer portion of thylakoid membranes at a high enough concentration to regulate membrane fluidity during the xanthophyll cycle [57].

In animals, the highest carotenoid concentration is found in the eye retina of primates, but even here the carotenoid concentration in the lipid-bilayer portion of the membrane is much lower than 1 mol% [19]. Therefore, it seems advisable that to act as antioxidants in vivo, carotenoids should be incorporated in tissues in the correct location and/or selectively accumulated to present a local concentration high enough to protect vulnerable molecules. Thus, for systems with a low carotenoid concentration, it is especially important to understand how the membrane itself—its composition, structure, and lateral organization—affects the organization of carotenoids in the lipid bilayer, including orientation (transmembrane vs. parallel) and localization (distribution between membrane domains).

Orientation of polar and nonpolar carotenoids (including their cis-forms) in the lipid-bilayer membrane has been investigated by many laboratories [19,44,47,58,59]. It is suggested that the presence of polar hydroxyl groups at the ends of carotenoid molecules and their transmembrane orientation (as in the case of zeaxanthin and lutein) enhance their antioxidant properties, as compared with the antioxidant properties of monopolar (β-cryptoxanthin) and nonpolar (β-carotene) carotenoids [40,60]. Although dipolar zeaxanthin and nonpolar β-carotene show similar antioxidant properties in organic solutions, they differ when incorporated into membranes. Zeaxanthin was shown to react with free radicals slightly more effectively than β-cryptoxanthin and much more effectively than β-carotene [40,61]. β-carotene and lycopene are able to react efficiently only with radicals generated in the inner part of the membrane, whereas zeaxanthin and lutein, with their end groups exposed to an aqueous environment, can also scavenge free radicals generated in the aqueous phase [42].

We hypothesize that the transmembrane orientation of xanthophylls may also enhance the antioxidant properties of these carotenoids indirectly—by changing membrane properties so that the membrane becomes less sensitive to oxidative damage. It has been shown that perpendicularly oriented dipolar carotenoids (but not nonpolar carotenoids [62]) significantly affect membrane properties, including membrane fluidity [33,34], vertical fluctuations of alkyl chains [63], the hydrophobicity of the membrane interior [35], and the oxygen transport rate within the membrane [64]. McNulty et al. [65] tried to relate the physical interactions of carotenoids with the membrane to carotenoids’ antioxidant effects. According to this group, an ordering effect of carotenoids is accompanied by a strong antioxidant action, as seen for the dipolar xanthophyll, astaxanthin. On the contrary, nonpolar carotenoids like β-carotene and lycopene disordered the membranes and acted as prooxidants. β-carotene—because of its low membrane solubility as a monomer and low incorporation efficiency, and, therefore, weak effects on membranes—does not protect membranes against lipid peroxidation. Moreover, at a high oxygen concentration, it may act as a prooxidant [66]. It has to be emphasized, however, that possible indirect antioxidant action of carotenoids via their alteration of membrane properties would require a high concentration of carotenoids in the membrane.

Distribution of macular xanthophylls between membrane domains

The lipid-bilayer portion of biological membranes is currently depicted not as a passive matrix in which membrane proteins are immersed, but as an active membrane component that controls a variety of biological functions through selective accumulation [67–70] or exclusion [71,72] of certain proteins and lipids [73] from specific membrane domains. Raft domains have been postulated to enhance signal transduction [71,72,74], and are also involved in lipid sorting [75] and protein trafficking/recycling [69,76]. It has been shown that in membranes of retinal pigment epithelium and photoreceptors, raft domains are present [77–80]. Rafts in membranes of photoreceptor cells are involved in regulation of the G-protein-mediated pathway of photo-transduction [79]. Aggregation of small, unstable rafts in bigger platforms (observed, for example, in retinal pigment epithelium cells) is supposed to enhance signal transduction to the cell interior and cause a specific reaction in the cell, such as apoptosis [80].

Membranes of the human retina are abundant in long-chain, polyunsaturated fatty acids, such as docosahexaenoic acid (DHA), with a six double-bond chain [21,45,82]. Additionally, POS discs contain nearly equimolar concentrations of unsaturated fatty acids (DHA), saturated fatty acids (myristoyl + palmitoyl + stearoyl), and cholesterol [83], which makes them very similar to raft-forming mixtures. Indeed, rafts were isolated as a detergent-resistant membrane (DRM) fraction from POS disc membranes [77–80]. Interestingly, the remaining detergent-soluble membrane (DSM) fraction, which is formed by the bulk lipid domain surrounding the raft domain in POS disc membranes, is rich in long-chain, polyunsaturated fatty acids. Additionally, rhodopsin, the main protein of POS membranes (comprising more than 90% of all proteins in these membranes) that is responsible for the first stages of visual signal transduction, is also located in the bulk domain of the POS membrane [84] and can be isolated mostly with the DSM fraction [77,78,80]. Also, Wang et al [85] using fluorescence recovery after photobleaching technique showed that in the dark and immediately after photoexcitation, rhodopsin diffuses freely in the bulk lipid phase. It is worth mentioning that rhodopsin requires the presence of polyunsaturated lipids (DHA) for its activity [86–88], and thus their co-localization is functionally justified. Surprisingly, in the model of POS membranes, macular xanthophylls were also about 14 times more concentrated in the unsaturated bulk domain (enriched in polyunsaturated DHA and isolated as DSM) and excluded from the raft domain (enriched in saturated lipids and cholesterol and isolated as DRM) [89] (Fig. 1B). Similar distribution was also found in membranes made of raft-forming mixture where macular xanthophylls lutein and zeaxanthin were about eight times more concentrated in the bulk, unsaturated domain than in the raft domain [90] (Fig. 1A). A similar distribution has been observed for monopolar β-cryptoxanthin, but not for nonpolar β-carotene, which is more uniformly distributed between DRM and DSM domains (Fig. 1A).

Fig. 1.

(Mole ratio of carotenoids and total lipids in DSM)/(mole ratio of carotenoids and total lipids in DRM) indicated on the y-axis as (Carotenoid in DSM)/(Carotenoid in DRM) in domains isolated from membranes made of raft-forming mixture (equimolar ternary mixture of dioleoylphosphatidylcholine/sphingomyelin/cholesterol) with 1 mol% carotenoid added (see Ref. [90] for more detail) (A). (Mole ratio of xanthophylls and total lipids in DSM)/(mole ratio of xanthophylls and total lipids in DRM) indicated on the y-axis as (Xanthophyll in DSM)/(Xanthophyll in DRM) in domains isolated from the model of POS membranes (equimolar ternary mixture of 1-palmitoyl-2-docosahexaenoylphosphatidylcholine/distearoylphosphatidylcholine/cholesterol) with 1 mol% lutein or zeaxanthin added (see Ref. [89] for more detail) (B).

Localization of macular xanthophylls in domains rich in unsaturated lipids is ideal if they are to act as a lipid antioxidant, which is the most accepted mechanism through which lutein and zeaxanthin protect the retina from age-related macular diseases [17,18,91]. Polyunsaturated DHA, localized in this domain and extremely susceptible to lipid peroxidation, needs to be protected in its direct proximity. Also, the oxygen transport parameter (oxygen diffusion-concentration product) is two to four times greater in the bulk domain than in the raft domain [89], which makes polyunsaturated lipids located in the former domain even more susceptible to oxidative damage. Finally, photoactivation of rhodopsin (also located in the unsaturated bulk domain) leads to isomerization of its chromophore, 11-cis-retinal to all-trans-retinal (ATR), which under certain conditions can act as a photosensitizer. During intense light illumination, upon regeneration of rhodopsin, ATR can be released from the opsin to the inner leaflet of the disc membrane before being reduced to all-trans-retinol [92]. Free ATR may absorb light and transfer energy from its excited triplet state to molecular oxygen, generating singlet oxygen [93]. Owing to close proximity of xanthophylls, which are able to effectively quench excited triplet states of molecules, the energy transfer from excited ATR to xanthophyll is possible, which prevents singlet oxygen generation by this photosensitizer [92]. The unsaturated bulk domain is fluid (more fluid than the raft domain [95,96]), so xanthophyll molecules can diffuse there faster and encounter a greater number of reactive species. The schematic drawing illustrating the distribution of macular xanthophylls between domains in the POS membrane is presented in Fig. 2. As indicated in the figure, the bulk domain contains rhodopsin surrounded by unsaturated (and polyunsaturated) lipids, which, during illumination, makes this domain even more vulnerable to oxidative damage. Thus, accumulation of xanthophylls directly in this domain is more than perfect.

Fig. 2.

Schematic drawing of the distribution of xanthophyll molecules between the saturated raft domain and the unsaturated bulk domain in the membrane of the photoreceptor outer segment. In this illustration, the integral membrane protein, rhodopsin, which is located in the unsaturated membrane domain, is also included to show its co-localization with unsaturated lipids and xanthophylls. As was demonstrated by X-ray diffraction and linear dichroism (in thin membranes), xanthophyll molecules are inclined with respect to the bilayer normal [59]. This inclination decreases with membrane thickness [33] (see Ref. [99] for membrane thicknesses). The thickness of the POS membranes, which contain not only long-chain fatty acids such as DHA [82] but also very-long-chain fatty acids [100,101], should be significantly greater than those investigated in Refs. [33] and [59]. Therefore, we depict xanthophylls as perpendicular to the membrane surface.

The presented data clearly demonstrate that macular xanthophylls are excluded from cholesterol-rich membrane domains. They are also poorly soluble in membranes with a high cholesterol content [97]. This has been confirmed by Wisniewska et al. [98] who showed that spin-labeled lutein was completely insoluble in saturated PC membranes containing 30 mol% cholesterol. All these suggest that the xanthophyll-cholesterol interaction is weaker than the xanthophyll-phospholipid interaction. In the lipid bilayer, the rigid bar-like xanthophyll molecule does not conform to the cholesterol molecule, which has a rigid, plate-like tetracyclic ring structure and flexible isooctyl chain. Two polar groups of the xanthophyll molecule interact with opposite surfaces of the membrane, and its rigid bar-like portion crosses the entire membrane. In contrast, the molecule of cholesterol is located in one leaflet of the bilayer, and its rigid plate-like portion extends to the depth of the seventh to ninth carbon in the lipid bilayer. The cross-section of the isooctyl chain of the cholesterol molecule is much smaller than the cross-section of the rigid steroid ring. When rigid xanthophyll and cholesterol molecules are located next to each other in the cholesterol-rich, lipid-bilayer domain, a free space is created in the membrane center (indicated by the ring in Fig. 3). Cholesterol molecules are forced to sink deeper into the bilayer, which is energetically unfavorable because it allows water to access the hydrophobic surface of alkyl chains. Thus, macular xanthophylls are excluded from cholesterol-rich domains, as illustrated in Figs. 2 and 3.

Fig. 3.

Schematic drawing showing the localization of xanthophyll molecules in the cholesterol-rich raft domain and the cholesterol-poor bulk domain. Unfavorable interaction of rigid xanthophylls with cholesterol, which creates a free space in the membrane center, is indicated by the circle. When cholesterol is surrounded by phospholipids, such as in the cholesterol-pure domain, the created free space is filled by the flexible hydrocarbon chains of phospholipids.

Concluding remarks

As presented above, dipolar macular xanthophylls are largely excluded from DRMs extracted from xanthophyll-containing membranes made of raft-forming mixtures [90] or from models of POS membranes [89] and remain concentrated in DSMs. DRMs and DSMs isolated from model and cell membranes, thought to be related to membrane domains, have similar lipid compositions as rafts and bulk domains, respectively [68,73]. All these suggest that in POS membranes macular xanthophylls should also be concentrated in the bulk domain enriched in polyunsaturated lipids, where rhodopsin is also located.

Co-localization of rhodopsin with polyunsaturated phospholipids has its functional purpose [86–88]. However, it creates a dangerous situation for both, especially during illumination when reactive oxygen species can be produced by photosensitizers. Thus, nature has used xanthophylls as an effective protector that can neutralize photosensitizers and reactive oxygen species. Co-localization of these molecules, together with polyunsaturated phospholipids and rhodopsin, should significantly enhance their effectiveness as protector, especially when the local concentration of xanthophylls in the membrane is not so high. This is possible because of the domain structure of the POS membrane and the ability of these domains to select and exclude specific classes of lipids and proteins. Thus, the membrane domain structure can also play a significant role in the protection of vulnerable molecules and processes by co-localizing them with protective molecules and processes. We consider this a function of membrane domains that has not yet been explored and that can be included with already accepted domain functions such as signal transduction, lipid sorting, and protein trafficking/recycling. Nevertheless, the domain structure allows location of macular xanthophylls in the most vulnerable regions of POS membranes, which should significantly enhance their ability to prevent AMD.