Preliminary identification of the human macular pigment

C'ision Rrs.

Vol

25. So.

I I, pp.

Prmted ,n Grent Bnun.

1531-1535.

1985

Copynghr

All nghts reserved

PRELIMINARY

RICHARD Departments

of Physics

W-t:-6YXY ‘(5 Si i)o - 0 00 (I 19x5 Pergzmon Prcrs Ltd

IDENTIFICATION OF THE HUMAN MACULAR PIGMENT

A. BONE. JOHN T. LANDRU~~and

and Chemistry. (Receked

Florida

20 March

International

SARA

University,

1985; in rrtisedfornl

6 Jlrnr

L. TARSIS Miami,

FL 33199. U.S..\.

1985)

Abstract--The

human macular pigment has been found to be composed of two chromatographicall) separable components, which are tentatively identified as lutein [(3R,3’R,6’R)-P.E-Carotene-3.3’-diol] and zeaxanthin [(3R,3’R)-fi./?-Carotene-3,3’-diol]. Chromatograms of retinal extracts. obtained b) HPLC on three different stationary phases, were found to match those obtained with mixtures of lutein and zeaxanthin standards. Identical retention times were confirmed by coinjection of each of the isolated components with the appropriate standard. U.V.-visible spectra of the purified components uere identical in all respects with those of lutein and zeaxanthin. Further support for our identification was obtained by the preparation and chromatographic comparison of derivatives of the macular pigment and of the standards. Macular

pigment

Lutein

Zeaxanthin

Carotenoids

INTRODUflION

The existence of the yellow macular pigment, which characterizes the retinas of primates including man, has been known since the eighteenth century (Polyak, 1941). Yet a definitive characterization of the species responsible is still lacking. Early attempts to identify the pigment have been summarized by Walls and Mathews (1952), and the generally accepted identification of the pigment as lutein is based solely on its characteristic carotenoid absorption spectrum reported by Wald (1945, 1949). Since then, it has been suggested (Brown and Wald, 1963) that a mixture of the cis and trnns isomers of lutein might be involved, this producing a closer spectroscopic match with the pigment than the trans isomer alone. The macular pigment was shown by Wald (1949) to reduce the sensitivity of the macula region to blue light by behaving as a broad band filter. Consequently it influences color appearance, color matching and tests for color vision (Ruddock, 1972). An advantage to the visual system of its presence may be the improvement of visual acuity. This may be achieved through compensation for chromatic aberration in the eye’s refractive media and by reduction in the amount of atmospherically scattered blue light reaching the receptors (Walls, 1967; Reading and Weale, 1974). While the identity of the pigment is not crucial to these hypotheses, it is more relevant to the postulated function of the pigment of protecting retinal tissue against photosensitized reactions (Kirschfeld, 1982; Bone and Landrum, 1984). Further, a definitive characterization would be of interest to biochemists and nutritionists concerned with the metabolic and/or dietary origin of the pigment. The present investigation was initiated for the

, K ‘5 II-4

HPLC

purpose of establishing a preliminary identification through the powerful analytical potential of highperformance liquid chromatography. This technique. applied to macular extracts and their chemical derivatives, and supported by spectroscopic studies. has revealed the composition of the macular pigment as a mixture of two carotenoids, most probably lutein and zeaxanthin.

EXPERIMENTAL

Extraction

METHODS

of the maculur pigment

Human eyes were obtained frozen from the Florida Lions Eye Bank and the National Diabetes Research Interchange. Retinas were dissected in 0.9% saline solution, care being taken to minimize exposure to U.V. and visible light. The macular tissue, identified by its yellow coloration, was removed from each retina. In the few cases where no yellow pigment was discernible, the retinas were discarded. The macular pigment was extracted by grinding the tissue in redistilled acetone at 0°C. The resulting yellow solution was filtered (0.4 pm Millipore filter) and placed in a freezer at -20°C for several hours. During this period,

fractional

crystallization

of lipophilic

com-

pounds occurred, indicated by a white deposit appearing on the container walls. The remaining solution was transferred to another container and thoroughly dried under a stream of pure nitrogen. On redissolving the pigment in ethanol, a 3-peaked spectrum, characteristic of carotenoids, was obtained. Assuming an extinction coefficient of approximately 1.3 x 10’1 mol-’ cm-‘, typical of a wide range of carotenoids (Strain, 1938). an estimate was obtained of pigment available. on average. from of -long 1531

BONE rr uf

RICHARD A

0

lo

20

30

40

50

0

3 0

to

20

b.2

IO

20

30

40

Time, min

0

10 20 Time, min

30

0m

II30

0 5 10 15 Time, min

Fig. I. Chromatograms of the macular pigment components (upper diagrams) and lutein and zeaxanthin standards (lower diagrams) on 3 different stationary phases. (a) Column, reversed-phase 5pm ODS; eluent, 93% methanol, 7% water/acetonitrile (75:25, v/v); Row rate, 0.5 ml/min; sample volume, 20 pl. (b) Column, normal-phase 5 pm SiO,; eluent, 90% n-hexane, 10% acetone; flow rate, I mI/min; sample volume, 20 ~1. (c) Column, normal-phase 73% CaCO,, 15%Mg0, 12% Ca(OH), (< 25 pm); eluent, 55% n-hexane, 45% chloroform; Bow rate, 2.5 ml/min; sample volume, 20~1. In all cases, MPI and lutein eiuted with shorter retention times than MP2 and zeaxanthin.

each retina. In each of the experiments described below, between 10 and 20 retinas were used. High-performance

liquid chromatography (HPLC)

HPLC was carried out on a Laboratory Data Control unit with dual pump and gradient capability, using 3 mm (ID) x 25cm columns. All eiuting solvents were of HPLC quality but, nevertheless, were redistilled prior to use. Initial purification of the macular pigment was accomplished using a reversed-phase (5 pm Techsphere ODS) column. Chromatograms consistently indicated the presence of two major components, MPI and MP2, in similar, but somewhat varying proportions. Baseline separation was achieved [Fig. I(a.l)] and the two components could be collected individually. Further HPLC studies were conducted with a normal-phase column packed with 5 pm silica gel (Techsphere) and a noncommercial, normal-phase column containing 73% CaCO,, 15% MgO and 12% Ca(OH), (by weight), a combination known to produce good separation of lutein and zeaxanthin when applied to thin layer chromatography (Hager and Meyer-~ertenrath, 1966). Isolation of carotenoid standards

U.V.-visible spectra of the purified components MPI and MP2 were found to be consistent with those reported for lutein and zeaxanthin, respectively

(Strain, 1938). Also their HPLC retention times on the reversed-phase column were found to match closely those obtained for Iutein and zeaxanthin by Braumann and Grimme (1981) when using similar elution conditions. Isolation of these pigments from spinach and, in the case of teaxanthin from corn as well, was accomplished by the following procedures. Methanol extracts were treated with KOH to saponify the chlorophylls, and the carotenoids were separated by sotvent partition using diethyl ether (see, for example, Strain, 1938). Initial purification and identification of the individual carotenoids was achieved by thin layer chromato~aphy using the methods of Hager and Meyer-Bertenrath (1966). For large scale work, these methods were modified for Iiquid~oIumn chromatography. Final purification was by reversedphase HPLC. Further support for the identification of the standards was provided by the observed consistency with literature data on HPLC retention times (Braumann and Grimme, 1981), u.v.-visible spectra (Strain, 1938) and, most impo~antly, mass spectra (Heiler and Milne, 1978). Experimental details and results of the mass spectrometry are given in the Appendix. Derivatizatian

Derivatives of the two macular pigments and of the standards were prepared successfully on the nano-

Idenulication

of human

macular

I533

pigment

OH

HO

Fig. __ 7 Structures of lutein and zeaxanthin. gram scale by modifications of synthetic procedures reported in the literature for lutein and zeaxanthin. Acetate esters (Britton and Goodwin, 1971) were obtained in good yield by the addition of 0.1 ml of acetic anhydride to 0.2 ml pyridine solutions of each of the parent compounds. The reactions were allowed to proceed in the dark under a nitrogen atmosphere for I8 hr. Work-up was accomplished by the methods described by Britton and Goodwin (1971). Chromatography of the esterified pigments was carried out on the reversed-phase column, eluting with methanol at a flow rate of I ml,‘min. Lutein is distinguished from zeaxanthin by its single allylic hydroxyl group (Fig. 2). This may be exploited in the selective preparation of the allylic monomethyl ether of lutein by treatment of lutein with methanol and hydrochloric acid (Curl, 1956; Jensen and Hertzberg, 1966). Under the same conditions, zeaxanthin is unreactive. Solutions of MPt. MP2, lutein and zeaxanthin were dried under streams of pure nitrogen and the residues redissolved in 5-10 /tl of methanol,‘con. HCI (9: I by volume). The reactions were terminated after 7 min by the addition of saturated solutions of sodium bicarbonate. Chromatography of the products was conducted on the reversed-phase column (Fig. 5).

RESULTS

AND

DISCUSSION

Chromatography of extracts of the macular pigment has proven that it consists of two distinct components, MPl and MP2 (Fig. I). Absorption spectra in the u.v.-visible (see Figs 3 and 4) were

Wavelength

, nm

Fig. 3. Absorption spectra of the macular pigment component MPI (-) and lutein (---). Solvent, 93% methanol, 7% wateriacetonitrile (75:25. v/v), as used for HPLC elution. The III/II absorption ratio for each pigment is 0.91.

__. Wavelength,

nm

Fig. 4. Absorption spectra of the macular pigment component MP? (-) and zeaxanthin (---). Solvent. 93% methanol. 7% water acetonitrile (75:25, v’v). as used for HPLC elution. The III 11absorption ratio for each pigment is 0.89. found to be totally consistent with those of lutein and zeaxanthin respectively. Among the characteristic features, which emphasize the consistency. are the wavelengths of the absorption maxima and minima, the III/II absorption ratios (absorption at the long wavelength maximum (III) to that at the main peak (II)) and the definition of the short wavelength shoulders and long wavelength peaks. The previous suggestion (Brown and Wald, 1963), that both cis and tran~ lutein were present in the macular pigment. may now be ruled out as no cis bands were observed in the 320-340 nm range (Vetter er al.. 1971). Chromatography was performed on three different stationary phases (see Experimental Methods). In each case, the same retention times, within the limits of the equipment, were observed for MPI and lutein as well as for MP2 and zeaxanthin (Fig. I). Coinjections of MPl and lutein were carried out on all three chromatographic stationary phases, resulting in the elution of only single peaks. Similarly, MP2 and zeaxanthin coeluted as single peaks. during initial Chromatograms obtained purification of retinal extracts on the reversed-phase column indicated the presence of other weakly absorbing compounds [Fig. l(a.!)]. The peak which appears at about 22 min retention time showed some increase in height with the length of time of tissue storage and, in one experiment, was sufficient for a crude absorption spectrum of the collected compound to be recorded. This was characterized by a broad maximum in the 400-420 nm range and may, therefore, represent one of the subsidiary, nonmacular pigments (P410) found by Snodderly et al. (1984a). None of the other minor compounds which appear in the chromatogram of Fig. I(a.1) was sufficient for spectral analysis. Lutein and zeaxanthin are isomers having the structural formulas shown in Fig. 2. They differ only in the placement of one double bond. In lutein, one of the end group hydroxyl moieties is allylic while in zeaxanthin, neither hydroxyl is associated with the double bonds. Both hydroxyl groups in these two xanthophylls can be esterified to produce less polar

derivatives having distinctly ditrerent chromatographic properties when compared with the parent compounds. Upon reaction of MPI with excess acetic anhydride. we obtained a derivative with a retention time on the reversed-phase HPLC approximately twice that of the parent compound, and matching precisely that of lutein diacetate prepared under identical conditions. Coinjection of the two derivatives resulted in a single chromatogram peak. Similarly the MPZ acetate derivative and zeaxanthin diacetate eluted as a single peak when coinjected, the retention time again being roughly double that of zeaxanthin. Yiefds from these reactions were sufficient for a comparison of u.v.-visible spectra. Complete consistency was observed between the acetate derivatives of the macular pigments and the derivatives of the corresponding standards. The spectra were, in fact, essentially indistinguishable from those of the standards themselves (Figs 3 and 4). To further substantiate the assignment of MPI and MP2 as lutein and zeaxanthin respectively, we sought data which would distinguish clearly between these two carotenoids. The aflylic hydroxyl group of lutein provides the means by which such distinction may be made. Conversion of this group to a methyl ether has been reported to occur under conditions which do not after zeaxanthin (see Experimental Methods). However, reaction of lutein with HC~/methanoi consistently produced two new compounds, rather than one, with retention times on the reversed-phase column of approximately 30 and 32 min, compared with 15.2 min for iutein itself (Fig. 5). This observation suggests that rearrangement accompanies the reaction. Racemization at the 3’ carbon may be

a

A

I4

Fig. 5. Chromatograms of monomethyl ethers of (a) MPI and (b) lutein. The single peaks at _ t 5 min are tentatively identified as unreacted parent compounds, based on coinjection with the lutein standard; those at -30 and 32 min, the reaction products. Column, reversed-phase 5 pm ODS; eluent, 93% methanol, 7% water/acetonitrile (75:25, v/v): Row rate, 1ml/min; sample volume, 20 I( 1.

occurring, giving rise to the two derivarlvcs. This would be consistent hith the reaction proceedins bq a carbocation mechanism. .%PI likewise reacted to produce two chromato~~am peaks nr 30 and 32 min. While the exact nature of the two components is still under investigation, the observation of identical reactivity of MPI and lutein to produce indistinguishable products is compelling evidence that these two pigments are indeed identical.

Under the same experimental conditions, neither MPZ nor zeaxanthin reacted, the recovered products having identical retention times (16.4 min) to that of the untreated zeaxanthin. This result shows conclusively that the second component of the macufar pigment cannot be a cis isomer of lutein. as suggested in the earlier work (Brown and Wald. 1963).

While a definitive identification of the human macular pigment must await mass spectrometry. our present studies suggest that it is composed of two pigments that are similar to lutein and zeaxanthin. If the macular pigment is derived from dietary sources, as it almost certainly must be (Malinow et nl., 1980), the relative abundances of futein and zeaxanthin in the diet might be expected to be duplicated in the macular pigment. However, while fukin is vastly more abundant in green plants, our studies have led US to conclude that this is never the case with the macular pigment. In general, the two components were found to be present in similar amounts [see, for example, Fig. I(a.l)], neither exceeding the other by more than about 30%. Occasionally, though less frequently, zeaxanthin was in excess, yet the only common plant source where this is so is corn (Weedon, 1971). We plan to study this anomaty and its possible metabofic implications further, initially by examining the relative abundance of the two carotenoids in individual maculae. We have previously shown (Bone and Landrum, 1984) that Haidinger’s polarization figure is consistent with incorporation of the macular pigment, assumed to be lutein, into the membranes of Henle fibers, the inner segments of cone receptor cells. This hypothesis is supported by the results of a recent microdensitometry study (Snodderfy er al., 1984b). Owing to the close similarity of lutein and zeaxanthin, our explanation of the brushes is still valid, and suggests that both carotenoids may be involved in the protection of these membranes against photosensitized reactions (Kirschfefd, 1982). Ack~uw~ed~emenfs-We thank the Florida Lions Eye Bank, Inc. of Miami and the National Diabetes Research Interchange for suooiying the human eyes used in this study. Support was $o;ided by grants frbm the National Ins&tutes of Health. No. I R01 EYO5452-01.and the Florida International University Foundation. R. A. Yost and E. D. Britton kindly performed the mass spectroscopy of the

Identification of human macular pigment carotenoid standards. We acknowledge the laboratory assistance of A. Gonzalez and wish to thank J. Xl. E. Quirke for his helpful comments.

APPENDIX Mass

spectra of lutein and zeaxanthin standards were run on a triple-stage quadrupole mass spectrometer (Finnigan MAT TSQ45. INCOS data system) scanning m,‘z 35-950 in 3 sec. Source conditions for electron impact tonization were 70 eV electron energy, 0.3 mA emission current and 190X. Samples were volatilized from the solids probe at 4 330-C during a programmed linear temperature gradient of 5O‘C per min. The relative intensities of some characteristic peaks for the lutem standard were as follows (literature values (Heller and Milne, 1978) are in parentheses): m’z 568:28 (23); m/z 550:41 (31): m/z 476:not detected (8); m,z 119: 100 (70); m/z 91:53 (100). For the zeaxanthin standard. the values were: m/z 568: 100 (100): m/z 550:5 (12); m;z 476:43 (34); m/z Il9:85 (92).

REFERENCES Bone R. A. and Landrum J. T. (1984) Macular pigment in Henle fiber membranes: a model for Haidinger’s brushes. Vision Res. 24, 103-108. Braumann E. and Grimme H. L. (1981) Reversed-phase high-performance liquid chromatography of chlorophylls and carotenoids. Biochitrr. biophys. Acm 637, 8-17. Britton G. and Goodwin f. W. (1971) Biosynthesis of carotenoids. In Methods in En:~~o~o~~ (Edited by McCormick D. B. and Wright L. D.). Vol. 18, Part c. Academic Press, New York. Brown P. K. and Wald G. (1963) Visual pigments in human and monkey retinas. Nurure, Land. 200, 37-43. Curl A. L. (1956) On the structure of “Hydroxy-aCarotene” from orange juice. Food Res. 21, 689-693. Hager A. and Meyer-Bertenrath T. (1966) Die Isolierung und quantitative Bestimmung der Carotenoide und Chlorophylle von Blattern, Algen und isoiierten Chloroplasten mit Hilfe d~nn~hichtchromatograph~scher Methoden. Pfunta 69, 198-217.

1535

Heller F. R. and Milne G. W. A. (1375) ET.4 .YIH Ilu.vs Spectrul Dma Base, Vol. 4. National Bureau of Standards, Washington. Jensen S. L. and Hertzberg S. (1966) Selective preparation of the lutein monomethyl ethers. dcra them. sco~t
Polyak S. L. (1941) The Rerina. University of Chicago Press. Chicago. Reading V. M. and Weale R. A. (1974) Macular pigment and chromatic aberration. J. opt Sot. Am. 64, 331-23-t. Ruddock K. H. (1972) Light transmission through the ocular media and macular pigment and its signilicance for psychophysical investigation. In Handbook oJ Sensory Phvsioloav (Edited bv Jameson D. and Hurvich L. M.). Vol. 7(4).~Springer, Berlin. Snodderly D. M.. Brown P. K., Delori F. C. and Auran J. D. (l984a) The macular pigment. I. Absorbance spectra, localization, and discrimination from other yellow pigments in primate retinas. Inresr. Oph~irai.i-i.watSzi. 25, 66&673. Snodderly D. M., Auran J. D. and Delori F. C. (l98Jb) The macular pigment. II. Spatial distribution in primate retinas. Inucst. Ophrhal. oisual Sci. 25, 614685. Strain H. H. (1938) Leuf Xunfhophylls. Carnegie Inst. of Washington, Washington. D.C. Vetter W.; Englert G., Rigassi N. and Schwieter U. (1971) Spectroscopic methods. In Cororenoids (Edited by Isler 0.). Birkhauser Verlag, Basel. Wald G. (1945) Human vision and the spectrum. Scieftce 101, 653-658. Wald G. (1949) The photochemistry of vision. Docrtrnenra Ophrhnl. 3, 94-137.

Walls C. L. (1967) The Verlibrale Eye and irs Adopriw Radiarion. Hafner, New York. Walls G. L. and Mathews R. W. (1952) New means of studying color blindness and normal fovea1 color vision. Unio. Cal. Pubis. Psycho/. 7, I-172.

Weedon B. C. L. (1971) Occurrence. In Caru~enojff~ (Edited by Isler 0.). Birkhauser, Basei.