With age, fluorescent granules called lipofuscin or age pigments accumulate in the retinal pigment epithelium (RPE). 1 These granules are believed to lead to cellular aging processes and related diseases, notably age-related macular degeneration (AMD), 2 the leading cause of blindness in elderly people for which no remedy exists. It is generally accepted that the pigments are formed as a consequence of accumulation of debris resulting from incomplete digestion of phagocytosed outer segment disks in lysosomes. Among the compounds that accumulate in lipofuscin, the orange fluorophores have attracted wide interest since they are considered to be the possible cause of age-related decline of cell functions. In this communication we report the structure of the major fluorophore. This orange pigment was first isolated from> 250 human donor eyes (age> 40 years) by a series of silica gel column chromatography and preparative TLC. 3, 4a Structural studies were mostly performed by analysis of the FAB-CAD MS/MS fragments of the hydrogenated pigment because of the limited amount,< 100 µg, and ill-defined 1H NMR peaks. This led to structure 1 for the major orange fluorophore. 4a Since it consisted of two retinal moieties and one ethanolamine moiety, retinal and ethanolamine in a 2: 1 molar ratio were reacted to give an orange pigment having the same TLC Rf values and UV and fluorescent spectra as the native fluorophore. 4a, 5 However, no further structural studies were performed due to the minuscule amount of material. The present studies were performed because of ambiguities4b present in structure 1; this has led to revised structure 2 for this fluorophore “A2-E”(2 mol of vitamin A aldehyde and 1 mol of ethanolamine).

Structural studies were performed with synthetic A2-E (5 mg) 6 prepared from all-trans-retinal (1 g) and ethanolamine. 4a The FAB-MS M+ peak at m/e 592 agrees with the HR-MS derived C42H58NO formula reported for the native pigment, 4a which had two hydrogens fewer than expected from the earlier structure 1; ie, another degree of unsaturation should have been present. Two retinoid moieties up to C-12 could be readily discerned by comparison of its 1H NMR spectrum with that of the quaternary iminium salt of retinal (Figure 1a). 7 COSY, longrange COSY, NOEs, and coupling constants enabled one to extend these units to two retinoid moieties: substructure I lacking the 20-methyl and substructure II lacking the 15′-H (Figure 1b). The remaining signals were assigned to an sp2 carbon (III) and an ethanolamine moiety (IV). HMBC correlations between the δ 7.39 proton signal in III and C-12, C-14, and C-15′ indicated the presence of an sp2 carbon at C-20 instead of methyl, thus leading to connections of moieties I/III/II through C-13/C-20/C-15′. Furthermore, HMBC between 15-H (δ 9.20) and C-15′ disclosed the connectivity of I/II through C-15/C-15′. Their chemical shifts implied the presence of nitrogen between C-15 and C-15′, thus forming a pyridine ring. The connectivities between I, II, and IV were clarified by the NOEs between 1′′-H’s and 14′-H/15-H, which revealed the attachment of the hydroxyethyl chain to the pyridinium nitrogen. The hydroxyl group was shown by the appearance of a hydroxyl proton (δ 5.20) in DMSO-d6, while the presence of a trifluoroacetate counteranion8 revealed by 19F NMR (δ-76.0) and