Characterization of Peroxy-A2E and Furan-A2E Photooxidation Products and Detection in Human and Mouse Retinal Pigment Epithelial Cell Lipofuscin*

The nondegradable pigments that accumulate in retinal pigment epithelial (RPE) cells as lipofuscin constituents are considered to be responsible for the loss of RPE cells in recessive Stargardt disease, a blindness macular disorder of juvenile onset. This autofluorescent material may also contribute to the etiology of age-related macular degeneration. The best characterized of these fluorophores is A2E, a compound consisting of two retinoid-derived side arms extending from a pyridinium ring. Evidence indicates that photochemical mechanisms initiated by excitation from the blue region of the spectrum may contribute to the adverse effects of A2E accumulation, with the A2E photooxidation products being damaging intermediates. By studying the oxidation products (oxo-A2E) generated using oxidizing agents that add one or two oxygens at a time, together with structural analysis by heteronuclear single quantum correlation-NMR spectroscopy, we demonstrated that the oxygen-containing moieties generated within photooxidized A2E include a 5,8-monofuranoid and a cyclic 5,8-monoperoxide. We have shown that the oxidation sites can be assigned to the shorter arm of A2E, to the longer arm, or to both arms by analyzing changes in the UV-visible spectrum of A2E, and we have observed a preference for oxidation on the shorter arm. By liquid chromatography-mass spectrometry, we have also detected both monofuran-A2E and monoperoxy-A2E in aged human RPE and in eye cups of Abca4/Abcr–/– mice, a model of Stargardt disease. Because the cytotoxicity of endoperoxide moieties is well known, the production of endoperoxide-containing oxo-A2E may account, at least in part, for cellular damage ensuing from A2E photooxidation.

The bisretinoid fluorophores that accumulate in retinal pigment epithelial (RPE) 3 cells as lipofuscin constituents are considered to be responsible for the loss of RPE cells in recessive Stargardt disease (1)(2)(3), an early onset form of macular degeneration, and may also be involved in the etiology of age-related macular degeneration (4). The RPE lipo-fuscin fluorophores isolated thus far include A2E (5-7), iso-A2E (7), less abundant double bond isomers of A2E (8), and all-trans-retinal dimer conjugates (9, 10) (Fig. 1). The most intensely studied of the RPE lipofuscin constituents are A2E and related photoisomers, pigments that, when accumulated by RPE cells in culture, have been shown to bestow a sensitivity to light damage (11)(12)(13). Blue light produces the most pronounced effect (11). The augmentation of cell death under conditions that prolong the lifetime of singlet oxygen together with the protection provided by quenchers and scavengers of singlet oxygen has implicated singlet oxygen as having a role in the events leading to the death of the cells (14).
The propensity for A2E to undergo photooxidation was initially revealed by a tendency for fluorescence quenching of intracellular A2E upon blue light illumination (13). HPLC analysis later confirmed this observation, the absorbance of the A2E peak after 430 nm irradiation, exhibiting a corresponding reduction (13). Subsequent analysis by mass spectrometry showed that 430 nm irradiation of A2E, either in an acellular or cellular environment, yielded products that, starting from the M ϩ m/z 592 peak attributable to A2E, presented as consecutive peaks differing in m/z by 16 (15). Several lines of investigation suggested that singlet oxygen is involved in the photooxidation, with A2E serving as a sensitizer for the generation of singlet oxygen from triplet oxygen (13,15,16). For instance, the extent of photooxidation was found to increase in deuterium oxide (D 2 O), a solvent that prolongs the lifetime of singlet oxygen. Irradiation (430 nm) of A2E in chloroform also produces a luminescence at ϳ1270 nm, which is typical of the phosphorescence of singlet oxygen. In addition, experiments demonstrated that singlet oxygen generated by an endoperoxide of 1,4-dimethylnaphthalene could substitute for blue light in mediating A2E oxidation. Singlet oxygen quenchers were also found to be inhibitory. Nonetheless, oxidation by other reactive forms of oxygen may also occur (17,18).
Given that A2E has nine double bonds besides the pyridinium ring, together with the observation that nine m/z peaks (592 ϩ n (16)) culminating in the m/z 736 peak (592 ϩ 9(16)) ( Fig. 2) (15,18,19) appear following 430 nm irradiation of A2E, we previously suggested that A2E undergoes oxidation at all nine double bonds to form an unprecedented nonaoxirane structure. Because of the likelihood that numerous stereoisomers of the nonaoxirane species were present, together with its instability and the small amount of compound that was available, structural studies were not performed. As part of our effort to elucidate mechanisms involved in A2E oxidation, we also oxidized A2E with meta-chloroperoxybenzoic acid (MCPBA) (15). This approach led to the appearance of an m/z 624 peak (parent peak), to which was assigned a 7,8,7Ј,8Јbisoxido structure based on the nonaoxirane and mass spectrometry and 1 H NMR data (15). More recently, it has been reported that photooxidation products of A2E can include a mono-furanoid oxide, a bis-furanoid oxide (Fig. 3B), and a mono-furanoid oxide with a second oxygen attached to the cyclohexenyl ring (20). However, this conclusion was based on the fragmentation patterns generated by collision-induced dissociation and tandem mass spectrometry, an approach that cannot discriminate between a 5,8-furanoid and a 7,8-epoxide. Nor can the use of 1 H NMR spectra distinguish a furanoid from an epoxide moiety because the chemical shift of the 7,8-protons in 1 H NMR does not reveal the resonance of the 7,8-carbon. Here we provide definitive evidence from HSQC-NMR spectroscopy that the complex mixture of oxidized species resulting from A2E photooxidation includes a 5,8-monofuran-A2E. Further investigation has also uncovered a bisoxygenated photoproduct that has been structurally characterized as 5,8-monoperoxy-A2E. More importantly, through an analysis of chromatographic properties and UV-visible spectra together with mass spectrometry to provide molecular weight information, we have also detected a monofuran-A2E and monoperoxy-A2E in aged human RPE and in the eyecups of mice with null mutations in Abca4/Abcr, the gene responsible for recessive Stargardt disease.  Under the influence of light, A2E and iso-A2E undergo photoequilibrium to generate an ϳ4:1 mixture (A2E/iso-A2E). ESI mass spectrum of A2E (m/z 592) after irradiation at 430 nm is shown. The addition of oxygens is evidenced by a series of peaks that differ in m/z by 16 with the peak at m/z 736, indicating the formation of nonaoxo-A2E, a compound that is likely a complex mixture of stereoisomers.

EXPERIMENTAL PROCEDURES
Reagents-MCPBA, ethanolamine, and trifluoroacetic acid were purchased from Aldrich; HEPES was obtained from Sigma; acetonitrile was purchased from Fisher; and Dulbecco's phosphate-buffered saline was from Invitrogen. All of the other chemicals were from Sigma. A2E was synthesized as described previously (7).
MCPBA Oxidation of A2E-To a solution of A2E (20.0 mg) in methanol (1.0 ml) was added MCPBA (33.0 mg, 2 eq), and the mixture was stirred for 12 h at room temperature in the dark. After concentration in vacuo, the reaction mixture was subjected to HPLC analysis. A2E mono-and bisfuranoid (Fig. 4) were eluted at 31 and 21 min, respectively, using a Vydac C18 column ( Endoperoxide of 1,4-Dimethylnaphthalene and A2E Oxidation-1,4-Dimethylnaphthalene endoperoxide was synthesized as described previously (21). Subsequently, endoperoxide of 1,4-dimethylnaphthalene (48.0 mg, 10 eq of A2E) was added to a solution of A2E (15.0 mg) in CD 3 OD (1.0 ml), and the mixture was stirred overnight at room temperature in the dark. After removal of solvent, the oxidation products of A2E were isolated and purified by HPLC using a YMC C18 column (10 ϫ 250 mm) with the following solvent system: solvent A, CH 3  NMR-1 H NMR and heteronuclear singlet-quantum coherence spectroscopy (HSQC) spectra of 5,8,5Ј,8Ј-bisfuran-A2E were obtained at 500 MHz using a Bruker DMX-500 spectrometer. 1 H NMR and HSQC spectra of 5,8,5Ј,8Ј-bisperoxy-A2E were recorded at 400 MHz (Bruker DRX-400). All data were recorded in CD 3 OD, and TMS was used as internal standard.
NMR of Bisfuran-A2E: 1  Cell Culture and Illumination-Human adult RPE cells (ARPE-19, American Type Culture Collection, Manassas, VA) lacking endogenous A2E (22) were grown as described previously (22). To generate A2Eladen RPE, nonconfluent cultures were allowed to accumulate A2E from a 20 M concentration in medium. Cells were subsequently exposed to 430 nm illumination (0.36 milliwatt/mm 2 ) as described previously (13). Illuminated cells were harvested and added to a solution of chloroform and methanol at a 2:1 ratio. This mixture were homoge-nized in a glass tissue homogenizer and centrifuged at 10,000 ϫ g for 10 min. The supernatant was dried under argon, redissolved in methanol, and subjected to liquid chromatography-mass spectrometry (LC-MS) study.
Human and Mouse Tissue-Human donor eyes were obtained from the National Disease Research Interchange (Philadelphia). Abca4/Abcr null mutant mice (Rpe65 450Leu, pigmented, 129/SV ϫ C57BL/6J) were generated as described previously (23). The mice were raised under a 12-h on-off cycle lighting with an in-cage illuminance of ϳ250 lux. Human RPE (from four eyes, ages 58 -68) and murine posterior eye cups (six eye cups; age 17 months) were homogenized in phosphatebuffered saline using a tissue grinder. An equal volume of a mixture of chloroform/methanol (2:1) was added, and the sample was extracted three times. To remove insoluble material, extracts were filtered through cotton and passed through a reversed phase cartridge (C18 Sep-Pak, Millipore) with 0.1% trifluoroacetic acid in methanol. After removing solvent by evaporation under argon gas, the extract was dissolved in methanol containing 0.1% trifluoroacetic acid, for LC-MS analysis.
HPLC-A Waters TM 600 equipped with photodiode array detector (model 996) and operating with Empower software was used for HPLC analysis. A Cosmosil 5C18 column (4.6 ϫ 150 mm; Nacalai Tesque, Japan) was used for analytical scale HPLC.
Mass Spectrometry-Fast atom bombardment ionization-MS was performed on a JEOL JMS-HX110A/110A tandem MS (Akishima, Tokyo, Japan), using 10-kV acceleration voltage and fitted with a Xenon beam FAB gun (6 kV) on the MS-1 ion source. 3-Nitrobenzyl alcohol was used as matrix. HPLC separation combined with mass spectrometry for analysis of A2E oxidation products in cultured cells and tissues was performed on an Esquire 3000 (Bruker Daltonic Inc., Billerica, MA) with electrospray ionization (ESI) source. Samples were introduced with flow injection mode. A reversed phase column (Atlantis dC18, 3 m, 4.6 ϫ 150 mm; Waters) was used under the following mobile phase conditions: solvent A, CH 3

RESULTS
To generate a partially oxidized A2E (oxo-A2E) species, we began by using MCPBA as the oxidizing agent. Reaction of A2E with MCPBA (2 eq) in the dark followed by HPLC analysis revealed a bisoxo product with 1 H NMR and mass spectral profiles that were identical to those reported previously for MCPBA-oxidized A2E (15). We surmised that the oxygen-containing moiety could exhibit one of three possible structures: 5,6-epoxide, 5,8-furanoid, or 7,8-epoxide (Fig. 3). The 5,6-epoxide and 5,8-furanoid were considered to be candidates because MCPBA oxidation of ␤-carotene with the same ring moiety yields a 5,6-epoxide; moreover, the 5,6-epoxide readily rearranges to the 5,8-furanoid structure even under mild acidic conditions (20,24). However, the 5,6-epoxide was eliminated as a possibility because NMR revealed that the 7,8,7Ј,8Ј proton signals at 5.15-5.26 ppm are shifted upfield (see the NMR data under "Experimental Procedures") compared with those of A2E at 6.18 to 6.53 ppm (6). Because fragment ions generated using collision-induced dissociation tandem mass spectral analysis for structural determination of oxo-A2E can form by rearrangement during ionization, we sought confirmatory evidence of a furanoid ring by NMR analysis. Accordingly, by using HSQC-NMR spectroscopy to reveal correlations between carbon atoms and directly attached protons (hydrogen), the structure was shown to be that of a 5,8-furanoid (Fig. 4). Specifically, HSQC analysis revealed that the two 87 ppm sp 3 carbons are coupled to 5.15 (8-H)/5.19 (8Ј-H) ppm protons, whereas the two 117 ppm sp 2 carbons are coupled to 5.22 (7-H)/5.26 (7Ј-H) ppm protons. In addition, because both of the two UV-visible absorbance maxima of A2E were blue-shifted (see below) in this bisoxo-A2E, it was established that the bisoxo-A2E is 5,8,5Ј,8Ј-bisfuran-A2E (Fig. 4). We were also able to identify the monooxo-A2E as 5,8-monofuran-A2E.
Because A2E photooxidation is likely to occur, at least in part, via a singlet oxygen-mediated pathway (13,15), A2E was also reacted with 1,4-dimethylnaphthalene endoperoxide, an aromatic compound that decomposes to singlet oxygen and 1,4-dimethylnaphthalene with a convenient half-life of 5 h at 25°C. Analysis of the product by LC-MS revealed the presence of a peak that exhibited a UV-visible absorption spectrum suggestive of a bisoxo-A2E but a molecular mass (m/z 656) that was consistent with the addition of 4 oxygens. NMR studies, including 1 H and HSQC NMR, showed upfield-shifted protons at the 7, 8, and 7Ј, 8Ј positions, which together with the existence of sp 2 carbons evidenced by HSQC data (115.2 (C-7), 116.7 (C-7Ј)) confirmed this tetraoxo-A2E as 5,8,5Ј,8Ј-bisperoxy-A2E (Fig. 5). The presence of the 1,2dioxin moieties in this oxo-A2E species was corroborated by comparison with a previous investigation that demonstrated the production of a cyclic 5, 8-peroxide upon photolysis of A1E, a nonbiological single side arm counterpart to A2E (25).
We found that oxidation-associated changes in the UV-visible spectrum of A2E served as a means to determine the oxidation site (Fig. 6). This was possible because the absorption peaks of A2E at 337 nm (band S) and 439 nm (band L) could be assigned to the shorter and longer chains that extend from the pyridinium ring, respectively. Thus, whether the hypsochromic shift occurred in either band S or band L revealed the side arm on which the loss of conjugation had occurred. For instance, comparison of the UV-visible spectra of the mono-furanoid and mono-peroxide with that of A2E revealed that only band S was blue-shifted. In contrast, both bands were observed to undergo a hyps-ochromic shift in the bisfuranoid and bisperoxide. Both furanoid and peroxide formation resulted in the loss of two successive conjugation systems, hence the absorption maximum (either band S or L) of the affected side arm was shifted toward the blue region by ϳ40 nm.
HPLC analysis of products generated by photooxidation and chemical oxidation also revealed the presence of 335 and 400 nm absorbance peaks that corresponded to blue shifts in only the L band. This observation indicated oxidation exclusively on the longer arm. However, the minuscule amount of these compounds precluded further structural studies. Oxidation apparently occurs more readily on the shorter arm of A2E. Of the two polyenes extending ortho and para to the pyridinium nitrogen, electron delocalization is favored along the former ortho arm, which has one extra conjugation bond; this leads to relatively lower density of electrons in each of the sp 2 carbons, thus making it less nucleophilic and less susceptible to the MCPBA attack.
In order to examine the formation of the oxidative products of A2E in a cellular environment, ARPE-19 cells were allowed to incorporate A2E, and the A2E-laden ARPE cells were irradiated at 430 nm. Chloroform/ methanol extracts of the cells were analyzed by LC-MS. As shown in Fig.  7, two eluting components that were more polar than A2E and with m/z of 624 and 608 were resolved. The UV-visible spectra of these peaks had similar profiles; specifically, absorbance maxima occurred at 296 nm (band S) and ϳ435 nm (band L). Because only band S exhibited a hypsochromic shift, relative to A2E, it was apparent that in both cases two successive conjugate systems on the shorter arm of A2E were lost by oxidation. On the basis of the presence of two peaks, which by LC-MS exhibited molecular sizes that corresponded to a mono-oxo (m/z 608) and bisoxo-A2E (m/z 624), together with our prior NMR analysis that established the structures of these products, it was evident that in this cellular system the mono-furanoid and mono-peroxide formed.
We also detected oxo-A2E in chloroform/methanol extracts of RPE cells isolated from human donor eyes and in extracts of posterior eye cups of Abca4/Abcr Ϫ/Ϫ mice. The latter mice are considered to be a model of Stargardt macular degeneration. Again we used LC-MS for this analysis. It is our experience that the use of MS/MS fragmentation analysis with collision-induced dissociation is not a reliable approach to the assignment of oxo-A2E structure because the fragmentation patterns vary, probably because the oxygen-containing fragment ions can undergo complex intramolecular rearrangements. As shown in Fig. 8, examination of the human and mouse samples by LC-MS revealed the presence of reversed phase HPLC eluates with extracted ion signals indicative of m/z 592, 608, and 624. The corresponding compounds   DECEMBER 2, 2005 • VOLUME 280 • NUMBER 48 could be identified as A2E (m/z 592), monofuran-A2E (m/z 608), and monoperoxy-A2E (m/z 624) because they had the same retention time and UV-visible absorbance spectra as the species detected in A2E-laden ARPE-19 cells following 430 nm irradiation (Fig. 7).

Photooxidation Products of A2E
To determine whether these partially oxidized forms of A2E could serve as intermediates on a pathway to further oxidation of A2E, we incubated A2E with MCPBA and isolated the product 5,8-monofuran-A2E by HPLC. The monofuranoid in phosphate-buffered saline with 0.5% Me 2 SO was subsequently submitted to blue light irradiation (430 nm), and unirradiated and irradiated samples were analyzed by MS. As shown in Fig. 9, the 5,8-monofuran-A2E (m/z 608) underwent further oxidation to generate higher molecular species representing the incorporation of additional oxygens (i.e. m/z 640, 672, and 720 peaks corresponding respectively to addition of 2, 4, and 7 oxygens). Most interestingly, however, if irradiation of 5,8-monofuran-A2E was carried out with methanol as solvent, no additional oxidation occurred.
We also proceeded to revisit the issue of solvent effects on A2E photooxidation. To this end, we performed HPLC analysis of unirradiated samples of A2E and samples in which A2E was irradiated in an environment of Me 2 SO/H 2 O and samples irradiated in Me 2 SO/D 2 O (Fig. 10). Consistent with the ability of deuterium solvent to extend the lifetime of singlet oxygen (26), irradiation in D 2 O resulted in more pronounced FIGURE 6. Correlation between hypsochromic shifts in the UV-visible spectra and A2E oxidation. The UV-visible spectrum of A2E exhibits two major absorbance bands at 337 and 438 nm. These bands can be assigned to the shorter (S) and long (L) arms of A2E, respectively. Oxidation of a carbon-carbon double bond is accompanied by a hypsochromic shift. A hypsochromic shift in band S corresponds with oxidation on the short arm; a blue-shift in band L reflects oxidation on the long arm; hypsochromic shifts in both peaks indicates oxidation on both side arms.  photooxidation, the latter reflected in the greater diminution in the A2E peak relative to the sample irradiated in H 2 O. Specifically, quantitation of the peak area revealed that with irradiation in Me 2 SO/H 2 O A2E was decreased by 75%, whereas in Me 2 SO/D 2 O under the same conditions of irradiation the reduction was 90%. In these experiments, only the singlet oxygen that escaped to the medium would be affected by the deuterated solvent.

DISCUSSION
Based on the foregoing results, we propose that A2E oxidation occurs by means of multiple independent mechanisms. For instance, the addition of one atom of oxygen can occur at the 5,6 position in the cyclohexenyl ring to form an epoxide followed by, at least in some cases, a rearrangement to a 5,8-furanoid product. The initial epoxidation occurs at the 5,6 double bond as it is the most electron-rich carbon-carbon  DECEMBER 2, 2005 • VOLUME 280 • NUMBER 48 double bond because of a substantial methyl substitution effect. Rearrangement of the epoxide to a furanoid oxide structure would be expected to occur under acidic conditions, such as is present in RPE lysosomes. Peroxy compounds such as MCPBA, which are a source of an electrophilic oxygen, serve as a model for this mechanism, and one might expect that oxygen radicals could react in a similar fashion, adding one oxygen at a time. An alternative mechanism with singlet oxygen as intermediate would involve the reaction of molecular singlet oxygen at two carbon-carbon double bonds ([4 ϩ 2] cyclo-addition), adding two oxygen atoms to form a cyclic 5,8-peroxide product. Evidence for oxidation by molecular singlet oxygen is provided by experiments in which we incubated A2E with the singlet oxygen generator, 1,4-dimethylnaphthalene endoperoxide. Nevertheless, because of the participation of various reactive oxygen species and because of the possibility of trans-cis isomerization of A2E at various stages, the photooxidation of A2E in vivo is likely to be very complex. It has been observed repeatedly that as many as nine oxygens can be incorporated into the retinoid-derived side arms of A2E (15,18,19). The moieties constituting the nonaoxo-A2E and the mechanisms by which they are generated are as yet unknown.

Photooxidation Products of A2E
The most abundant species formed by reaction of molecular singlet oxygen with A2E is the product resulting from the addition of two oxygen atoms (m/z 624). The m/z 624 peak is also generated as a prominent peak during photooxidation, as are other peaks indicative of a pairwise incorporation of oxygens (13). Peaks reflecting an uneven number of oxygens are also present. Under conditions of photooxidation, individual oxygen atoms may also be incorporated by the addition of one atom of an oxygen radical (e.g. superoxide anion, hydroxyl radical). It is reported that a low yield of superoxide anion can be detected by EPR spectrometry when A2E is irradiated (Ͼ550 nm) in protic solvent (ethanol) (18) and when ESR spin trapping is used to monitor the 5,5dimethyl-1-pyrroline-N-oxide-OOH formation from 5,5-dimethyl-1pyrroline-N-oxide and superoxide after irradiation of A2E at 407 nm (17). Semi-oxidized and semi-reduced forms of A2E have been generated by interaction with bromide radical and formate, respectively, and the A2E radical was shown to react with oxygen to generate superoxide anion (27). However, the biological relevance of these experimental conditions is not clear.
The conjugated double bonds at the 5,6 and 7,8 carbons of A2E are excellent substrates for [4 ϩ 2] cycloaddition of singlet oxygen. Thus the formation of a cyclic peroxide following the photooxidation of intracellular A2E is likely to be a major product. Because of the weakness of the O-O bond, the major pathway from the endoperoxide is homolytic O-O bond cleavage to form an unstable diradical from which a number of reactive intermediates and products can form, including aldehydes, epoxides, and epoxyketones (28). In this way the endoperoxide moiety of oxo-A2E can be expected to promote reactivity (29). Indeed, the instability of A2E-endoperoxide may account for the observation that upon irradiation of A2E in cells, wherein the environment is conducive to reactivity, the peak at m/z 624 (monoperoxide) exhibits reduced intensity relative to the peak at m/z 608 (monofuranoid) (13). It is also important to note that the cytotoxicity of metabolites and synthetic endoperoxide-containing compounds is well known (29,30). The involvement of singlet oxygen in the formation of the peroxy-A2E photooxidation product is corroborated by the observation, in the present work, that the photochemical consumption of A2E was greater in D 2 O than in H 2 O. A similar effect of deuterium solvent was reported previously (15). The deuterium effect observed in the present work is also consistent with our previous observation that the singlet oxygen quencher 1,2,2,6,6-pentamethyl-4-piperidinol can protect against the photooxidation of A2E (13). It was also found that azide, histidine, and 1,4-diazabicyclo[2.2.2]octane all of which are efficient quenchers/scavengers of singlet oxygen, protected A2E-containing RPE from irradiation-induced cell death, whereas in deuterium-based media cell death was potentiated (13). These observations contrast with the report that photooxidation-associated loss of A2E occurred twice as fast in H 2 O than in D 2 O and that the rate of loss of A2E was also increased in the presence of azide, a singlet oxygen quencher (18).
We have shown previously that the monoretinoid A1E, a synthetic compound designed to model A2E except that it has a single side arm instead of two, undergoes photooxidation by mechanisms that are oxygen-and solvent-dependent. For instance, irradiation of A1E (425 nm) in methanol solution resulted in pericyclization to form pyridinium terpenoids, presumably secondary to photoisomerization of the C-7-C-8 double bond and ring closure. The quantum yield for this cyclization reaction was very low because of efficient photoisomerization of the 7-cis-isomer back to the all-trans-isomer of A1E. It is possible that irradiation-induced pericyclization of A2E could also occur, but because we have not observed this product, if present it must be very minor. Irradiation of A1E in air-saturated carbon tetrachloride or deuterated chloroform also generated a cyclic 5,8-peroxide, the same product as that observed in the current experiments with A2E. The yield of cyclic 5,8-peroxide formation on A1E was increased in environments that extended the singlet oxygen lifetime. Most interestingly, an increase in the concentration of A1E also increased the yield. We did not test the effect of A2E concentration on the yield of cyclic 5,8-peroxide in the present experiments, but we expect there to be a similar effect. Most interestingly, A1E photooxidation always ceases after the addition of two oxygens, whereas A2E oxidation can proceed until nine oxygens are added. These observations suggest that polyoxidation of A2E is enabled by the presence of two closely spaced side arms.
Photoreceptor cell degeneration in recessive Stargardt macular degeneration likely occurs as a result of the RPE cell dysfunctioning and death that is precipitated by the accumulation of lipofuscin pigments (31)(32)(33). Although RPE lipofuscin consists of a mixture of fluorophores, our understanding of the adverse effects of specific lipofuscin fluorophores has come largely from studies of A2E and its photoisomers. For instance, the photooxidative processes initiated in cultured RPE cells through blue light-induced sensitization of A2E can lead to DNA base lesions, modifications of protein, and changes in protein expression (14,34). Several areas of investigation suggest that the photooxidation products of A2E may be damaging agents that are biologically significant. For instance, cellular damage is realized in the presence of oxo-A2E even under conditions that eliminate singlet oxygen and other reactive forms of oxygen as the damaging intermediate (16). Additionally, fragmentation of photooxidized A2E with diffusion of the products may be the explanation for the observation that cellular damage can be observed at sites other than the lysosomal compartment in which A2E is housed, sites that are too distant for the damage to be accounted for by short lived singlet oxygen and other cytotoxic forms of oxygen generated by A2E photosensitization (35). A species corresponding to mono-oxo-A2E was detected previously in human retinal lipofuscin (36), whereas in the present work, monofuran-A2E and monoperoxy-A2E were detected in extracts of both human RPE and in eyecups of Abca4/ Abcr Ϫ/Ϫ mice. Also suggestive of the presence of blue-shifted photooxidation products are spectral studies of fundus autofluorescence, demonstrating that in patients with Stargardt disease the fluorescence emission is shifted toward lower wavelengths over hyperfluorescent lipofuscin-rich sites corresponding to ophthalmoscopically visible yellow flecks (37). Flecks are prognostically significant because they are a sign of a more severe disease (38). Although the spectral shift may indicate a difference in the relative levels of individual fluorophores in these areas, it is also conceivable that the blue shift reflects an increased content of oxo-A2E. In this regard, it may also be significant that fundus spectrophotometric measurements obtained in human subjects using 550 nm excitation indicate that RPE lipofuscin fluorescence declines after age 70 (39). Although there is potentially more than one explanation for the declining fluorescence at these excitation wavelengths (40), perhaps the spectral shifts associated with the photooxidation of A2E are a contributing factor.